Datasets:

---

mano-wii

/

blender_duplicates

like 0

[ "Weird VSE crash\nOperating system: Win11\nGraphics card: RX580\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen) blender 4.0.0 alpha\nWorked: (newest version of Blender that worked as expected) no clue\n\nBlender crashes on opening file. Was trying to make an alternative caption that gets around the Multiply color issue #112267 but came across a new bug.\n\nBased on the default startup or an attached .blend file (as simple as possible).\nOpen the file I've attached\n\n", "MultiRes+Solidify while trying to use Sculpt Mode cause Blender to crash\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.30\n\nBroken: version: 2.81 (sub 12)\nWorked: (optional)\n\nI tried deactivating one modifier while keeping the other one active to see if Blender would crash, but this was not the case. Only when I used both simultaneously did the program reliably crash every time, regardless of brush.\n\nAdd the modifiers that are active in the video and follow along. \n[Desktop 2019.09.23 - 00.49.39.02.mp4](Desktop_2019.09.23_-_00.49.39.02.mp4)\n", "Blender 3.3 crashing unnecessarily/ unexpectedly/ frequently/ no apparent reason until now\nOperating system: Windows 10 Pro\nGraphics card: Intel i3\n\nBlender 3.3\n\n\n\nWhenever I open blender with some projects...whenever I enter material preview mode or try to render (sometimes even in solid mode...still slow) Whenever I try to apply a material and the crashes even when the project is simple and not finished or being rendered....it was working just fine and started to act weirdly during the last couple of days...I even tried, manually, to allocate more RAM to the program but that just crashed the whole thing. \n\n\nExact steps for others to reproduce the error:\n\nopen the blender file I attached as an example (The project that blender doesn't tolerate) try to apply a texture or render \n\nNOTE: I'm looking for a software solution...I don't think reproducing the error on a RTX 3090 is fair...so...we'll probably look for software solution or tweak the settings.\n\n", "Crashes when Switching to Material Viewport / Rendered Viewport / Rendering\nOperating system: Windows 10\nGraphics card: 2 Nvidia GTX 1660 Super's\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: blender 2.8 / 3.4\n\n\nBlender runs completely fine when editing and building a model but once rendering is involved, even in preview, it immediately closes with no explanation. This crash will also happen with the default cube. This had not happened before and have tried Updating blender, reinstalling, simplifying, etc but no change and can reproduce the issue on my computer. i can have it in material mode if i have 0 assets in the scene but once i add something it crashes, i am suspecting a bad pathing issue as i had recently refreshed my computer but do not know how to fix it or give blender a hard reset.\n\n{[F14170509](blender.crash.txt)}[blender_debug_output.txt](https://archive.blender.org/developer/F14170500/blender_debug_output.txt)\nOpen Blender, have any Physical asset inside the scene, switch to rendered / Material Viewport, Crash\n\nI have included the project I was working on and the default cube as they create the same issue.\nPlease note that I have installed the most recent Blender program onto my laptop and opened the files attached and did not have any crash happen so it seems the issue is solely between my computer and the blender program.\n\n[Default Cube.blend](Default_Cube.blend) [Skell Rig base.blend](Skell_Rig_base.blend)", "GPencil: Blender crash when file open\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.4.0 Beta\n\nAs soon as I open attached file Blender crashes. Attached also crash log file.\n\n- Open attached file\n- Change radius (say 0.6)\n- Save file\n- reopen it (Crash)\n[#102930.blend](T102930.blend)", "Use draw manager for clip editor drawing\nCould be stalled to 2.93 as there isn't an real issue here as far as we are aware of. The compositor issues are more tangable.", "Crash with animated displacement modifier in Texture paint mode while playing\nAMD Ryzen 5 2600, 16GB ram\nWindows 7 64bit\nNvidia GTX 1060 6gb (driver version 416.34)\n\n2.79b\n\nI intend to paint a heightmap interactively. So i keyframed a displacement modifier (2 frames) on a subdivided plane and assigned a texture.\nPlaying the timeline should now refresh the modifier every frame. \nPainting in the UV-editor this way works fine, but switching to texture paint mode and painting on the mesh directly crashes blender after about 0.5seconds.\n[HeightmapPaint.blend](HeightmapPaint.blend)\nI attached an example file.\nHit play and in the uv editor paint on the noise texture. This works okay.\nNow switch to Texture painting in the 3d viewport. The moment a stroke is applied to the mesh it should crash (it does consistently for me). ", "Crash when selecting line art\nOperating system: Linux mint 21. Blender installed using flatpak.\nGraphics card: See attached system_information.txt\n\nBroken: version: 3.5.1\n\nBlender Crashes.\n```\nbacktrace\n\nPython backtrace\nFile \"/app/blender/3.5/scripts/startup/bl_ui/space_view3d.py\", line 704 in draw\nFile \"/app/blender/3.5/scripts/modules/bpy_types.py\", line 903 in draw_ls\n```\n\n```Py\nbpy.ops.outliner.item_activate(deselect_all=True) # Operator\nbpy.ops.outliner.item_activate(deselect_all=True) # Operator\n```\nWith attached blend file, in the outliner, select cameras, line-art then try to change the color of the background.\n\n", "Regression: Image rendering fails while viewport rendering is in progress\nOperating system: OSX Ventura Version 13.3\nGraphics card:AMD Radeon Pro Vega II Duo 32 GB\n\nBroken: 3.5.0 Date 2023-03-29 02:56 Hash: 1be25cfff18b \nWorked: 3.4.1 Date 2022-12-19 17:00 Hash 55485cb379f7\nCommandBuffer Failed: cycles_metal_integrator_compact_shadow_states\n\n* Load file provided\n* Set ViewPortShading to MaterialPreview\n* Then select ViewPortShading to Rendered\n* While the viewport is rendering Select Render Image.\n\nThis will often cause the error. It may take an attempt or two but more likely to happen after at least one attempt.\n\nIn some case blender will throw the error in both the viewport and in the image window. In some cases...Blender will hang and require you to force quit. At other times it will crash the OS. \n\nThe larger the scene the more drastic the failure. In this simplified file it typically will recover.\n\n", "Crash when playing animation in rendered view (cycles)\nOperating system: macOS-12.0.1-arm64-arm-64bit 64 Bits\nGraphics card: Apple M1 Max Apple 4.1 Metal - 76.1\n\nBroken: version: 3.0.0, 3.5\nWorked: unknown\n\nPlaying animations in cycles rendered view frequently crash blender. With the attached file I consistently get a crash within 1-2 minutes of starting the animation.\n\nThe crashes do not occur when running the intel version through rosetta. They occur much faster/more often when using the CPU than when using the GPU.\n\nOpen attached file\nSwitch to rendered view (Crashes with and without Metal preference)\nPress spacebar to play\n\n[Boat boating.blend](Boat_boating.blend)\n[Animation crash.txt](Animation_crash.txt)\n```\nThread 17 Crashed:\n0 Blender \t 0x10319354c tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::process_bypass_loop(tbb::internal::context_guard_helper<false>&, tbb::task*, long) + 340\n1 Blender \t 0x103193644 tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::process_bypass_loop(tbb::internal::context_guard_helper<false>&, tbb::task*, long) + 588\n2 Blender \t 0x103192c60 tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::local_wait_for_all(tbb::task&, tbb::task*) + 184\n3 Blender \t 0x103183300 tbb::internal::arena::process(tbb::internal::generic_scheduler&) + 252\n4 Blender \t 0x10318c3b8 tbb::internal::market::process(rml::job&) + 40\n5 Blender \t 0x10318d6b0 tbb::internal::rml::private_worker::run() + 288\n6 Blender \t 0x10318d584 tbb::internal::rml::private_worker::thread_routine(void*) + 12\n7 libsystem_pthread.dylib \t 0x1b39cd4ec _pthread_start + 148\n8 libsystem_pthread.dylib \t 0x1b39c82d0 thread_start + 8```", "Suspected asset library crash by `json_sax_dom_parser`\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.5.1\n\nI dont completely know why but blender crashes very often for me.\nIts doing something no idea what it is doing.\nSorry for being vague, but it seems to mostly happen when i have the viewport set to cycles render and i am trying to modify materials, in the shader editor.\n\n", "Blender 3.6.1 crashes when switching to Viewport Shading to Rendered on MacBook Pro M1\nOperating system: macOS-13.5-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Max 1.2\n\nBroken: version: 3.6.1\n\nEnabling objects in the scene causes blender to no longer render. It just says: Sample 0/32 (Using optimized kernels) and if i try and change the Viewport Shading back to something else it crashes. This issue seems to only happen on a specific file.\n\nOpen file, switch Viewport Shading to Rendered, try and switch it back to anything else.\n\n", "Crash when auto script execution is enabled\nOperating system: Windows 10 64-bit operating system, x64-based processor\nGraphics card: NVIDIA GeForce RTX 2060\n\nBroken: Blender 4.0.0 7372b3bdda35 and later\nWorked: never? 2.93 fails too\n\nBlender daily build crashes immediately on launch. The blender start screen isn't shown at all, only a grey window which then immediately closes. The crash log indicates that it is a memory EXCEPTION_ACCESS_VIOLATION (see log file attached).\n\n- Open preferences, enable `auto script execution`\n- Save prefs and close blender\n- Rename attached file and replace it with existing startup file\n- Launch blender (crash)\n\n- - -\n<details>\n <summary> Original report </summary>\nStart Blender through Steam, Blender Launcher, blender_debug_log.cmd, blender_debug_gpu.cmd, or blender_debug_gpu_glitchworkaround.cmd.\nA grey window opens with the Blender logo in the top bar.\nThe window closes as Blender crashes.\n\n[ Crashing_4-0-0_daily.blend](https://projects.blender.orgattachment)\n</details>", "Crash when baking Dynamic Paint Image Sequence multiple times in loop\nBroken: 2.90.0 Alpha\nWorked: blender-2.80-windows64\n\nBlender crashes when baking dynamic paint image sequence in for loop. Depending on wait time crash will have differens stack traces.\n\n[#74052.blend](T74052.blend)\n\n- Open file\n- Run script", "Crash when trying to show a Workspace using a deleted Screen.\nIf a Screen ID is deleted, and one then try to show a Worspace using it, Blender will crash.\n\nSetting to low prio since deleting a Screen ID is not easy to do for end user currently (you'll have to use py console, e.g. `bpy.data.batch_remove(ids=(bpy.data.screens- [x],))`), but still, this is violating general design and expected sane behavior of ID handling.\n\nThink solution should be either:\n* Deleting Screens IDs should be forbidden (not ideal, since it kind of breaks general ID design, but should be doable).\n* Deleting Screens IDs should also delete matching Workspace (not ideal either, there should not be that kind of relationships between IDs, but we already have similar things between objects and obdata...).\n* Or (probably the sanest thing to do) a Workspace should handle gracefully the case where it has no assigned Screen." ]

[ "Blender 2.80 crash after changing slot on image editor\nOperating system: Windows 10 Pro\nGraphics card: GTX 650 TI\n\nBroken: 2.80, 26d5a3625ed, blender2.8, 2018-11-29\n\nWorked: 2.80, 565de7750b9, blender2.8, 2018-11-28\n\nAfter hit the render button, on the UV/Image Editor, the Blender 2.80 crashes and closes himself.\n\nBased on the default startup or an attached .blend file (as simple as possible). Simply press F12, and on the image editor (with the \"Render Results\" active) press to change image slot, it closes the Blender." ]

[ "Render problems\nOperating system: Windows 10 Pro Ameliorated\nGraphics card: RTX 3070 driver version 472.12\n\nBroken: v3.6 and v3.61\nWorked: v3.51 <- Using the same file created in v3.61\n\nWhile rendering a frames of an animated character, the current rendered animation frame usually shows in the 3D viewport in v3.51. However, in the broken version, it fails to show the current frame being rendered, nothing gets updated visually. Then eventually in the 3D viewport the character goes into its T pose for a few frames then the character starts losing visibility of its layers, at that point Blender crashes. \nI will create a .blend file if I am the only one getting this bug.\n\n", "unable to use 'grid fill' or 'fill' on screwed object\nOperating system: Window 10\nGraphics card: NVIDIA GeForce RTX 2060 SUPER\n\nBroken: Blender 2.83\nWorked: Blender 2.76 \n\n\nI am trying to create a solid spring using Screw modifier. After applied grid fill, I expected a full capped end of the screwed object, but it turns out to be like figure below. The complete file also attached in this post. Hope to get your response soon. \n\n[screw.blend](screw.blend)\n\nBased on the default startup or an attached .blend file (as simple as possible).", "race_spaceship.blend demo file renders empty result (no renderlayer selected in Render Layers node dropdown)\nOperating system: Windows 11\nGraphics card: Radeon Graphic\n\nBroken: 3.6\n\nScene not display well as final render (demo file).\n\n1. Start Blender File.\n2. Check if Viewport working (all great).\n3. Render (seems going well).\n4. Final image (no image, trasparent background).\n\n", "Freeze when using Object.ray_cast in a driver script\nOperating system: Linux-5.10.15-1-MANJARO-x86_64-with-arch-Manjaro-Linux 64 Bits\nGraphics card: Radeon RX 580 Series (POLARIS10, DRM 3.40.0, 5.10.15-1-MANJARO, LLVM 11.0.1) AMD 4.6 (Core Profile) Mesa 20.3.4\n\nBroken: version: 2.91.2, 8dc95f61f3\nWorked: Unknown, fails on at least as old as 2.83.10.\n\nRunning a driver that includes a call to Object.ray_cast will freeze the interface entirely. Still exits normally when sent a SIGTERM. It does not create a crash.txt. Blender crashes with a segmentation fault in version 2.83.10 (hash: `b439a155442b`), and does create a crash.txt (see attached). Blender version 2.93.0 Alpha (hash: `7a608f88a713`) also freezes in the same way as 2.91.2. Using the ray_cast function in a non-driver script appears to function as expected.\n\nIn the attached blend file, run the text block `TurretRayCasting.py`, to put the script into the driver namespace. Next, click \"update dependencies\" in the driver panel. The application should freeze entirely. If `return target.ray_cast(turretTargetSpace, direction)- [x]` is commented out and replaced with `return True`, the application does not freeze.\n\n[driver_freeze2.blend](driver_freeze2.blend)\n\n[driver_freeze2.crash.txt](driver_freeze2.crash.txt)\n\n[system-info-2.91.2.txt](system-info-2.91.2.txt)", "Python Operator.report does not show in status bar when overriding context\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.09\n\nBroken: version: 3.2.0 Alpha\nAlso broken all the way back to 2.79.\n\nAn operator's `report()` call does not show a message in the status bar or info editor when the operator is not invoked by the user. This is expected.\nHowever, overriding the execution context to `'INVOKE_DEFAULT'` or even a specific area does not make it work either. This part seems like a bug.\n\n**Simple reproduction file**\n[no_reports.blend](no_reports.blend)\n- Run the script and notice Warning reports in the stdout but not in the status bar\n- Same when saving the .blend file\n- As expected, running the operator manually (F3 search) works as intended\n - Alternatively, modify the `invoke` function to require user input (such as via `context.window_manager.invoke_confirm`), then the timer/save handler calls also work\n- The override_context makes no difference here\n\n**Reproducing from a blank startup file**\n- Create an operator that calls `self.report` on execute\n- Call the operator from e.g. a timer or application handler\n- Even when overriding the context, the report only shows in stdout, but not in the user interface\n\n", "Saving single image of image sequence: image.save_render(): Could not acquire buffer from image. \nOperating system: Linux-5.15.5-76051505-generic-x86_64-with-glibc2.33 64 Bits\nGraphics card: Radeon RX 590 Series (POLARIS10, DRM 3.42.0, 5.15.5-76051505-generic, LLVM 12.0.0) AMD 4.6 (Core Profile) Mesa 21.2.2\n\nBroken: version: 3.1.0 Alpha\n\nWhen trying to call the .save_render() function via the PythonAPI on an image datablock that holds an image sequence Blender throws the \"Could not acquire buffer from image. \" error.\n\n1. Load any image sequence in image editor. \n2. Call bpy,data.images- [ ].save_render()\n\nI appended a zip with an already prepared blend file and image sequence. You can use that for convenience. [bug_report_bundle.zip](bug_report_bundle.zip)\n\n", "Viewport don`t updated, when UDIM tiles is created\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.6.1\n\n[video](-PMixtP5f3c)\n\n- Create a model where the UVs are spread across several UDIM tiles\n- Create a UDIM texture and use it within the material -- but do not create all the tiles necessary to cover the UVs\n- Go into Material Preview mode in the Viewport -- pink sections of the model should appear\n- Create a UDIM tile to cover the UVs\n- The Viewport will not be updated until some interaction occurs (the bug; an update can be triggered by orbiting the viewport or making a selection for example)", "Startup.blend file is not loaded on startup if userpref.blend does not exist.\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 26.20.100.7000\n\nBroken: version: 2.81 (sub 3)\nWorked: (optional)\n\nStartup.blend file is not loaded on startup if userpref.blend does not exist.\n\n- Remove userpref.blend file from it's usual location\n- Start blender\n- Dismiss splash screen by clicking on 3D viewport\n- Make some change to the file (move the cube)\n- Save as startup file\n- Close Blender\n# Open blender\n\nStartup file is not loaded", "Render Result not applying the compositor if the region is cropped\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.91.0\n\nWhen you render just a small region (to save time) and make the composite in full size (fullHD), it isn't taken into account in the final result. \nHaving the resulting file (in this case full size - fullHD) without the compositing taken into account, makes no sense.\n\nThis really limits the use of both *render region* and *Compositor*\n\nWorkaround is *File Output* node but that doesnt support video output.\n\n\nThanks and regards\n\n\n\n- Open attached file\n- render it and compare the compositor preview with the render result output\n\n[Test.blend](Test.blend)\n\n", "multiview: overwrite and file naming \n\n\"not overwrite\" option doesn't work with multiview when some views already exist on disk, could be a minor problem but rendering one single view with views option activated create a 0 byte \"myfile_L.jpg\" file but save only myfile.jpg\n\n\n\n\n\n", "--render-output command line argument and backslash\n## Commandline Rendering\n\nOperating system: Windows 11\nGraphics card: NVIDIA GTX 3060 Ti\n\n3.4.1\n\n\nIf the value for \"-o\" or \"--render-output\" ends with a backslash no rendering occur and Blender exits clean (0) with no error.\n\n### Example\n\nRender blender-file in current folder to subfolder with a space.\n\n>blender -b myfile.blend -o \"my folder\\\\\" -F PNG -f 1\n\n```\nBlender 3.4.1 (hash 55485cb379f7 built 2022-12-20 01:51:19)\nRead prefs: C:\\Users\\***\\AppData\\Roaming\\Blender Foundation\\Blender\\3.4\\config\\userpref.blend\nRead blend: c:\\blender\\myfile.blend\n\nBlender quit\n```\n\n**Observed behavior**\nNo render. No errors. No information of what went wrong.\n\n**Expected behavior**\nBlender should render to generic file \"0001.png\" in folder \"c:\\blender\\my folder\".\nBlender should provide some information why nothing got rendered.\n\n\n", "Avoid graphical glitches during threaded preview generation\nWhen previews are generated using an OpenGL engine, there can be a number of glitches until the preview is done. That's because some data is accessed/modified that is not thread safe. E.g. `object_preview_render()` modifies `U.pixelsize`. I think other glitches are caused by overriding the active theme in `ED_view3d_draw_offscreen()` - although when I avoided this some time ago for similar work on threaded offscreen drawing for VR, this didn't solve the color flickering.\nThe active theme should probably use thread local storage.\n\nBesides that, the preview generation should be thread safe with OpenGL:\n* It uses the draw-manager's own context, which is mutex protected.\n* The non-thread-safe immediate mode API is not used inside the draw-manager.\n* The draw-manager already has workarounds for GPU driver issues with threaded drawing.\n\nA note: We could also speed up Eevee preview generation significantly by using viewport render quality, not final render quality. The `_imbuf` offscreen rendering functions don't expose the `is_image_render` option for that, which they probably should.", "Dicrepency between requested mesh data layer and eval mesh data layer\nWhile fixing #91838 it became apparent that the data layers requested by `mesh_cd_calc_used_gpu_layers` differs from the ones when creating the batches. Which lead in this case to a crash.\n\nThe fix was just a workaround and fixing this would mean to have the same layers at request time. Or maybe, not assuming they will be available during batch creation.", " EXCEPTION_ACCESS_VIOLATION when entering Edit Mode\nOperating system: Windows 10 1903 (18362.239)\nGraphics card: Intel HD Graphics 4000\n\n***Broken:***\nBlender 2.80 (sub 75) / 2019-07-29 09:44 AM Windows / Stable\nBlender 2.81 (sub 1), branch: master, commit date: 2019-08-13 11:31, hash: 6f9cbbc8ec4f\n***Worked:***\nBlender 2.79 (copied mesh over through copybuffer.blend)\n\n\nWhen entering edit mode of a specific object Blender crashes with EXCEPTION_ACCESS_VIOLATION. Logs and example .blend provided.\n\nThe issue shouldn't be in my GPU, since the same issue persists even with Software Rendering.\n\n1. Open up example file\n[window_broken.blend](window_broken.blend)\n\n2. Select `window_panes` and enter Edit Mode\n3. Get EXCEPTION_ACCESS_VIOLATION\n\n**Logs (2.80 `blender_factory_startup`)**\n[blender_system_info.txt](blender_system_info.txt)\n[blender_debug_output.txt](blender_debug_output.txt)\n\n**Logs (2.81 `blender_factory_startup`)**\n[blender_system_info.txt](blender_system_info.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)", "Cycles destroys an existing texture image while baking to a new image\nOperating system: Windows 11\nGraphics card: rtx 2070\n\nBroken: 51bc44420b89\n\n\nBlender seems to destroy or empty an existing texture in a multi material setup while baking to a new image.\n\n\n\n\nload the attached file\n\nselect the BAKE material\n\nselect the BAKE image node\n\nRender Props -> Bake ( it is already setup)\n\nPay attention to the viewport to see that one of the textures turn black, and the BAKE texture is still black\n\nYou can try variations as in putting the BAKE node in other materials however the result seems to be the same failed bake.\n\n" ]

[ "Batch generate data-block previews broken\nOperating system: Kubuntu 18.10\nGraphics card: Nvidia Geforce GTX970\n\nBroken: 2.80, a070815d127e, blender2.8, 2019-03-06\nWorked: (optional)\n\nBatch generation of data-block previews fails with an error code.[empty.blend](empty.blend)\n\nI have attached empty.blend, which is a file I saved immediately after File->New->General.\nIf I open Blender 2.80 and File->Data Previews->Batch-Generate Previews and then choose empty.blend, the generation of the preview fails with the following Error in the terminal:\n - [begin]-------------\nERROR: bpy_struct: item.attr = val: enum \"\" not found in ('None', 'Agfa Agfacolor Futura 100', 'Agfa Agfacolor Futura 200', 'Agfa Agfacolor Futura 400', 'Agfa Agfacolor Futura II 100', 'Agfa Agfacolor Futura II 200', 'Agfa Agfacolor Futura II 400', 'Agfa Agfacolor HDC 100 plus', 'Agfa Agfacolor HDC 400 plus', 'Agfa Agfacolor HDC 200 plus', 'Agfa Agfacolor Optima II 100', 'Agfa Agfacolor Optima II 200', 'Agfa Agfacolor Ultra 050', 'Agfa Agfacolor Vista 100', 'Agfa Agfacolor Vista 200', 'Agfa Agfacolor Vista 400', 'Agfa Agfacolor Vista 800', 'Agfa Agfachrome CT Precisa 100', 'Agfa Agfachrome CT Precisa 200', 'Agfa Agfachrome RSX2 050', 'Agfa Agfachrome RSX2 100', 'Agfa Agfachrome RSX2 200', 'Agfa Advantix 100', 'Agfa Advantix 200', 'Agfa Advantix 400', 'Kodak Gold 100', 'Kodak Gold 200', 'Kodak Max Zoom 800', 'Kodak Portra 100T', 'Kodak Portra 160NC', 'Kodak Portra 160VC', 'Kodak Portra 800', 'Kodak Portra 400VC', 'Kodak Portra 400NC', 'Kodak Ektachrome 100 plus', 'Kodak Ektachrome 320T', 'Kodak Ektachrome 400X', 'Kodak Ektachrome 64', 'Kodak Ektachrome 64T', 'Kodak Ektachrome E100S', 'Kodak Ektachrome 100', 'Kodak Kodachrome 200', 'Kodak Kodachrome 25', 'Kodak Kodachrome 64', 'Kodak DSCS 3151', 'Kodak DSCS 3152', 'Kodak DSCS 3153', 'Kodak DSCS 3154', 'Kodak DSCS 3155', 'Kodak DSCS 3156', 'Kodak KAI-0311', 'Kodak KAF-2001', 'Kodak KAF-3000', 'Kodak KAI-0372', 'Kodak KAI-1010', 'Eastman Double X Neg 12min', 'Eastman Double X Neg 6min', 'Eastman Double X Neg 5min', 'Eastman Double X Neg 4min', 'Fujifilm F-125', 'Fujifilm F-250', 'Fujifilm F-400', 'Fujifilm FCI', 'Fujifilm FP2900Z', 'Canon Optura 981111', 'Canon Optura 981113', 'Canon Optura 981114', 'Canon Optura 981111.SLRR')\n*NOT* Saving /media/raid/stationary/blender_assets/empty.blend, because some error(s) happened while deleting temp render data...\nError: Not freed memory blocks: 130, total unfreed memory 0.057281 MB\n\nBlender quit\n - [end]---------------" ]

[ "VSE: Ensure all operators in the Strip menu are working on the selected and not only on active strip (+unselected) \nAs it is confusing and inconsistent to have functions in the Strip menu which only works on properties of the active strip even if it is unselected, this design list is for ensuring all functions in the strip menu to work on selection. \n\n- - [x] Swap Strip: #91191 Fix: [D12441](D12441)\n- - [x] Effect menu. Fix: [D12451](D12451)\n- - [x] Movie(Set Render Size + Deinterlace) and Image(Set Render Size + Separate Images) menu entries should be moved into the source sidebar with +alt for selection. Fix: [D12464](D12464)\n- - [ ] Add Modifier in the menu needs to work on selection and is missing +alt for selected, when executed in the sidebar.\n- - [ ] Inputs menu: Swap data should go into the swap menu and the rest into Strip Source sidebar with +alt for selected. Bug report: #91350\n- - [ ] Toggle Meta and UnMeta should go into the strip sidebar, as they are only working on the active strip. Old patch: #74130\n(A checked checkbox means a fix has been submitted, but not necessarily been committed)\n\n", "Tracking graph not showing as expected (Not a bug)\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 960/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 398.82\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional) Not sure this is a bug as I went back to Blender 2.79 and noticed similar behavior. It is however unexpected behavior.\n\n\nUnexpected/confusing Linked behavior between areas.\nMovie Clip Editor graph mode shows unexpected behavior.\n\n\nOpen a new general workspace. In my case I have a 3D Viewport as the main area, below it a Timeline. To the right of the 3D Viewport and the Timeline I have a Outliner and below the Outliner a Properties area.\n\nChange the main 3D Viewport to a Movie Clip Editor.\nLoad a image clip sequence. (for this sequence it does not seem to matter if you make any changes to the tracking points, focal length or sensor size)\nSelect any number of tracking points and track them. (Does not seem to matter if you track forward or backwards).\n\nChange the timeline to a movie clip editor.\nChange the type from a Clip to a Graph.\nEven though you have tracked several markers the graph is empty.\n\nRemove the Timeline/Movie Clip Editor area by letting the top Movie Clip Area expand over it.\nThen split/duplicate the Movie Clip Editor area downwards so there is a duplicate above and below.\nNow take the bottom Movie Clip Editor and change its type from clip to graph. You will now see the tracking graph with green and red graphs for the tracking markers.\n\nNow split the original Movie Clip Editor in two horizontal areas. So above the graph you will have two Movie Clip Editors, one left and the other right.\nTake any one of the two and either load a new clip, image sequence or simply unlink the data block and the graph will disappear.\n\nThis behavior is unexpected. It does not throw away or corrupt any data but did cause me to have to spend quite a bit of time trying to find out how to get the graph to actually appear and not disappear. I'm hoping you can find a more consistent and intuitive way of handling this.\n\nThank you for your awesome work on Blender!\n", "VSE Linked external scene with audio with input Sequencer only works if the scene is NOT also linked with input Camera\nOperating system: Linux-4.9.0-9-amd64-x86_64-with-debian-9.9 64 Bits\nGraphics card: Mesa DRI Intel(R) HD Graphics 520 (Skylake GT2) Intel Open Source Technology Center 4.5 (Core Profile) Mesa 13.0.6\n\nBroken: version: 2.81 (sub 3)\nWorked: Haven't yet tried other versions.\n\nVSE: Linking external scene with audio with input Sequencer only works if the scene is NOT also linked with input Camera.\n\n1. Create file A with Scene-A and add audio to it on the VSE.\n2. Create file B\n3. From the VSE in file B, link the Scene-A from file A. Set it's input to \"Sequencer\". On file B, playing the strip will correctly play the sound from Scene-A.\n3. Duplicate (or re-instance) Scene-A on the VSE sequencer. If the Input is still \"Sequencer\", both clips will play the audio, as expected.\n4. Switch one of the Scene-A instances to use the Camera as input. \n\nExpected behaviour: \n\nWe should be able to both see the Camera strip and hear the Sequence strip for Scene-A\n\nActual behaviour:\n\nFor all the strips of Scene-A that are set to Input \"Sequencer\", the audio will stop playing.\n\nNote: I've determined that it is not that the audio crashes: other clips with sound on the sequencer keep playing normally.\n\nNote: If I delete the strip with Input Camera and leave only the input Sequencer strips, the problem will not go away, unless I save the file like this and \"revert\" to it.\n\nI cannot find a suitable work-around. In the video that I'm producing I need to edit local sound effects on the shot sequence, and incorporate it on the larger edit.\nI cannot, therefore, both see and hear my sequence on the host sequencer.\n\n\nIf a workaround is possible (that is not reconstruct and re-import the sounds on the original scene) I'd be a happy man :-)\nThanks!\n\n\n", "Sound stops playing after undo\nOperating system: Darwin-18.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: Intel Iris OpenGL Engine Intel Inc. 4.1 INTEL-12.9.22\n\nLinux has this issue as well, no problem on windows.\n\nBroken: 2.92.0 Alpha\n\nSound stops playing after undoing all actions back to original file state.\n\n[#84052.blend](T84052.blend)\n\n- Open file\n- Select strip\n- Undo selection, so file is in original state\n- Play animation\n\nThere will be no sound.\n", "Audio sync issue with speed control and pitch in VSE\nOperating system: Linux-5.7.0-2-amd64-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.80.02\n\nBroken: version: 2.92.0 Alpha\n(The bug also occurs in 2.90 from Debian Bullseye)\n\nThere is a sync issue when using the speed control effect on a video clip as well as the pitch setting in an audio clip in the video sequencer. If you configure the speed control and the audio pitch to be anything but 1 and then start playing the video in the VSE from anywhere but the beginning, the audio will not be in sync at all with the video.\n\n- Add a movie clip.\n- Add speed control to the video part.\n- Configure the speed control as follows:\n - Uncheck \"Stretch to input strip length\".\n - Check \"Use as speed\".\n - Set speed factor to 3.\n - Set multiply speed to 1.\n- Configure the audio that was added as part of your movie clip as follows:\n - Change pitch to 3.\n- Start playing the movie from anywhere but the beginning.\n- The audio is now way out of sync with the video.\n\nYou can download an example project that reproduces the problem here:\n[test-blender-speed-control.zip](test-blender-speed-control.zip)\n\n\n", "Custom presets but with the name of builtin presets cannot be deleted\n\nBroken: Blender 2.83.5, blender-2.90.0-21cb6f09ffa8-windows64, blender-2.91.0-396d39c6b904-windows64\n\nThe original sampling presets in Cycles can't be deleted. However, it is possible to create presets of the same name. These presets become undeletable, because Blender assumes them to be the default presets.\n\n\n\n\nCreate sampling preset with name \"Preview\" or \"Final\". Try deleting new presets. \n", "Bug while moving vertices with mirror & subdivision modifier\nOperating system: Windows 10 64-bit\nGraphics card: ATI Radeon HD 5670\n\nBroken: 2.82\nWorked: at least 2.8\n\nWhen I try to move a vertice it seems like the mirror modifier gets applied in an odd way. The result even changes when moving the camera or hide one of the images I use for reference. In fact it becomes distorted when hiding the images.\nThe following link is a recording of my screen:\nview?usp=sharing\nI encountered this bug only recently. The file itself is older and I do not remember which version I used (since I use it via Steam I did not pay attention to the automatic patches).\n\n1. Open this file [Bugged Project.blend](Bugged_Project.blend)\n2. Disable visibility of Subdivision modifier for Edit Mode (which should already cause the bug)\n3. Move a vertice\n\n[system-info.txt](system-info.txt)", "UI bug in preset window after repeated remove\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.29\n\nBroken: version: 3.2.0\n\nWhen you remove a preset the textfield with the +/- moves one entry up\n[2022-07-15 09-17-07.mp4](2022-07-15_09-17-07.mp4)\n\n1. Go to the scene tab\n2. Create several new presets\n3. Remove all the new created presets in one go without moving the mouse away.\nI think the textfield with the +/- moves should just stay at the bottom and the window should shrink from the top.", "VSE 2.0: Performance, Cache System\n# Video Sequencer Cache\n\nNOTE: This is the first pass of the design. It will be worked a bit more after discussion within the module, and presentation and diagrams will become more clear.\n\nThis document describes caching system which design and implementation for the `VSE 2.0 project` (#78986).\nThere is some intersection of performance topics listed in the `VSE 2.0: Performance ` (#78992). \n\n## User level design\n\nOn user level cache system should follow zero configuration principle: the video playback and tweaking should be as fast as possible without user spending time on fine-tuning per-project settings.\n\nThe only settings which user should be interacting with are in the `User Preferences`:\n* In-memory cache size limit\n* Disk cache size limit\n* Disk cache folder\n\nThese settings are set once in the User Preferences and are used by all sequencer projects. The default values are based on the minimal [[ requirements/ | hardware\nrequirements ]].\n\n## Code design\n\n### Levels of caching\n\nFor best editing and playback performance multiple levels of cache are needed.\nThe most important ones are:\n\n* Cache of strip input (colormanaged using general Blender color space rules)\n\n```\nThis allows to faster, without lag, move image strip in time, adjust strip settings like transform, by avoiding need to re-read file on every modification. Lets call this cache level `STRIP_INPUT`.\n```\n\n* Cache of the final sequencer result.\n\n```\nThis allows to have realtime playback after the sequencer got cached. Lets call this cache level `SEQUENCER_FINAL`.\n```\n\nThe simplified flow of the image from disk to the artist is presented in the following diagram:\n\n\n\n---\n**NOTE**\n\nNeed to think about whether having strip output cache is helpful. If the stack rendering is fast, having extra levels of cache will have negative affect due to less final frames fitting into the memory.\n\n---\n\n### Cache referencing\n\nCache levels are to utilize reference counting as much as possible. For example, when having single Image Strip without modifications set up in the strip the final sequencer frame in the `SEQUENCER_FINAL` cache is to reference the image from `STRIP_INPUT` cache. This allows to minimize memory footprint, playback performance for the story boarding type of tasks performed in the sequencer.\n\nThe following example visualizes cache frame referencing in the following scenario:\n\n* Sequencer have single Image Strip using `HappyFroggo.png` as an input. The strip has length of 4.\n\n\n\nIn Blender terms, the cache contains a single copy of `ImBuf` created for `HappyForggo.png`. All the sequencer cache entries are referencing this `ImBuf` for lowest possible memory footprint.\n\n### Cache resolution\n\nIn a typical video editing scenario an artist views the sequencer result in a rather small area of the screen layout:\n\n\n\nThis behavior can be exploited in the following way: the sequencer processing and caching can happen in the lower resolution. This is something what current proxies design is solving, but does in the fully manual manner.\n\nThere is a possibility to make proxies behavior more automatic, by performing downscale on an image after it was read from disk, but before it gets put to the `STRIP_INPUT` cache. Default behavior could be something like:\n\n* Use closest power-of-two downscaling (similar to mipmaps)\n* The target resolution is 50% of the window resolution, but no more than 1080p.\n\nIn order to support workflows where an artist needs to investigate in a close-up manner the final result, there will be a display option `Best Quality` (but defaulting to `Best Performance`). This could fit into existing `Proxy Display Size` menu.\n\nIn the future more automated input resolution selection is possible to be implemented. For example, it is possible to automatically switch to the\n`Best Quality` mode when zoom-in is detected.\n\nImage scaling with a power-of-two scale factor can be implemented very efficiently using threading and vectorization.\n\nOn a performance aspect, for image sequences such scale down will be an extra computation, which will only pay off if effects/transformation is used.\n\nFor the movie files, this step will actually make things faster because it is required to convert color spaces (happening in `sws_scale`), which is not\nthreadable. The scale down will be done together with color space conversion, which is expected to give better performance compared to the current state of the sequencer playback.\n\n---\n\n## Task progress\n\n- [x] Cache referencing\n - d837923a56\n - c74086376f\n\n- [x] User level design\n- [x] Levels of caching\n - f448ff2afe\n - 445ebcaa30\n\n- [ ] Cache resolution\n - [D9414: VSE: Render in size nearest to preview image](D9414)\n - Update/followup above patch to work on images", "Particle display problem in rendering when using the Time remapping option\nOperating system: Mac Os Mojave 10.14.6 (18G95) and Windows 10 (Last update)\nGraphics card: AMD Radeon R9 M295X 4 Go & Nvidia GTX 1070\n\nBlender 2.80\n\nThere is a problem of particle rendering if we use the Time remapping function which makes it possible to make a film slower by multiplying the number of images of an animation. The particle animation is not continuous, some disappear and then reappear creating a jerky effect.\nThe problem does not appear when launching the animation directly in the software but only by doing a rendering calculation whether it is in film or frame-by-frame mode or movie (with Cycle or Eevee rendering)\n\nIf I leave the time remapping by default (old 100, new 100) there is no problem.[torus particules.blend](torus_particules.blend)\nIf I change the time remapping to have a movie with 4 times more image (old 25, new 100), the problem appears: the particle animation is not continuous, some disappear and then reappear creating a jerky effect.", "Sequencer: Driver on any strip property does not invalidate its cache [prefetch frames]\nOperating system: Darwin-18.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: Intel Iris OpenGL Engine Intel Inc. 4.1 INTEL-12.9.22\n\nBroken: version: 2.83.9\n\nSequencer: Driver on any strip property does not invalidate its cache [prefetch frames]\nThis can be any driver on any driven strip property:\nDoes not matter which property is the driving property (could be x Location of default cube as well)\n\n\n[#82776.blend](T82776.blend)\n- Open file\n- `driven` strips opacity is driven by Cube 's X location\n- change Cube 's X location --> driven strips opacity value changes, but the cache is not invalidated\n\n---\n\nOriginal report:\n\n Driver on `wipe.effect_fader` doesn't invalidate cache for driven wipe sequences. It's only invalidates the cache of the currently displayed wipe. This leads to display errors.\n\nNB: The issue is the same for any other parameter (e.g angle, blur...)\n\n- Open this file: [Bug_Wipe_Driver.blend](Bug_Wipe_Driver.blend)\n- There are two wipes, the first is the driver, the second one has the `effect_fader` value driven. Prefetch is enabled.\n- Move the `effect_fader` value of the driving wipe, its cache is reloaded, but not the cache of the driven one.\n\n", "NLA: Moving Strips Auto Changes Between Hold and Hold_Forward\nBug Report #42808 (NLA: Moving one strip resets extrapolation on other strip) #45854 (Hold forward is reseted in NLA)\n\n**Source of Problem**: #42808#843296 \n\n**Solution:** I think the code enforcing this can be removed. I don't think this restriction was ever necessary. When an animator uses a Full Replace Hold strip above other strips, it's expected behavior that it occludes the lower strips. That's what a Full Replace strip is supposed to do. Full means strip influence = 1. For consistency, we'll have to add support when a strip has extrapolation `None` leading to a strip with `Hold`. We'll also have to add support for `Hold_Backwards` extrapolation for completeness. In the case of conflicting adjacent strips (`Hold` into a `Hold`), we let the first strip have priority which leads to the next strip as behaving as `Hold_Forward`.\n\n**Potential Problems with Solution:** \n- When two adjacent strips have `Hold` extrapolation, which should evaluate? It seems natural to let only the preceding strip evaluate.\n- When a strip has `Hold_Forwards` extrapolation and the next has `Hold_Backwards`, which should evaluate? Here, I think we should still let only the preceding strip evaluate, for consistency with the previous point. \n\n____\n\n**Implementation/Patches**\n[D9942: NLA: Remove Hold resetting between Hold_Forward Behavior](D9942) \n[D9943: NLA: Strip Evaluate Held Strips Even When Not First Strip](D9943) (depends on previous patch)\n(todo) Patch3: Add `Hold_Backwards` support. Unsure about whether this is actually useful. Maybe we should wait till there is a concrete use-case for it? \n", "Crash playing animation with rigid body\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.83.1\n\nAfter opening a blend file (ver BLENDER-v280RENDH). The frame starts at 25 upon opening it. I changed the frame to 0. I pressed space to play then the application closes.\n\n[Whiskers Away.blend](Whiskers_Away.blend) \n- Open file\n- Rewind to frame 1 {key Shift Left}\n- Seek to frame 0 {key Left}\n- Play animation\n\nCrash doesn't happen every time. Redo all steps in such case. There is about 30% probability to reproduce successfully.", "alpha does not contribute to adaptive sampling noise treshold\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51\n\nBroken: version: 2.91.0\nWorked: i dont think it ever worked\n\nwhen rendering with adaptive sampling enabled, the noise treshold does not seem to be affected by a noisy alpha channel.\n\nthis is most visible on shadow catchers: since their rgb is always black, they are always rendered with \"Min Samples\", as soon as adaptive sampling is enabled.\neven though the alpha channel is still very noisy, it stops rendering additional samples.\n\n[shadowCatcherNoise.blend](shadowCatcherNoise.blend)\n- open file\n- render samples at 64-512 samples\n- when adaptive sampling is enabled, the shadow will remain very noisy. when adaptive sampling is disabled, the shadow will be smooth\n\n", "Line Art isn't updating active camera\nOperating system: GNU/Linux mint 19\nGraphics card: gforce RTX 2060\n\nBroken: 2.93.4 \n\nLine Art isn't updating active camera when it's changed in a 'Scene' strip (VSE)\n\n\n- Create a scene, a camera and a LineArt object.\n- Now, create a 2nd camera (Cam.001) (not the active one)\n- Create an empty scene in order to be used for video creation.\n- In `Video Sequencer`, add a Scene strip and choose the Cam.001 camera ->it changes ok, but lineart goes dependent on the first one.\n\nTest File:\n[#91971.blend](T91971.blend)" ]

[ "VSE: Audio comes from the soft-cut & removed portion\nOperating system: Ubuntu 18.04\nGraphics card: nVidia GTX 1050i\n\nBroken: 2.83.0, branch: master, commit date: 2020-06-03 14:38, hash: 211b6c29f771, type: Release\nWorked: 2.82\n\nIn the Video Editing tool, after cutting some audio and video strips for a while, I started listening to muffled sounds coming from previously deleted sound strips. If I save and reopen the file, the ghost audio disappears for a while, coming back when I start to edit again. \n\nEdit some video file, make a soft cut, remove a piece, save it, close the project. Open project again, everything should be normal so far. Start soft-cutting again and play the video. You should hear the sound coming not only from the remaining piece of the strip but also from the piece that you cut. \n\nIt is very random. I used 2.83 for two days without that and now it seems to be happening in every project I open. I'm not sure what triggered it." ]

[ "Grease Pencil: Better render Anti-Aliasing\n## Motivations:\nThin lines are correctly rendered using grease pencil with the currently implemented post process Anti-Aliasing. This is done by making sure the line width is not less than 1.3 and by modulating its opacity.\nHowever, there is cases where this is not enough. For instance really dense drawings with small lines with little to no curvature still tends to flicker a lot.\nAnother case is when the gpencil strokes intersects the scene geometry. This makes the line less thick to the post process which makes it too small.\n\n## Proposed solution:\nWe propose to implement a Temporal AntiAliasing solution. It is robust and will match Cycles & EEVEE's film filtering if using the same filtering function. This only adds one temporary buffer (low VRAM usage compared to rendering at higher resolution) and is completely self contained (works with transparency and don't interfere with other features).\n\nI propose to only implement this for final rendering since it might introduce flickering that might be annoying while drawing.\n\nHowever this wouldn't solve the issue of gpencil strokes intersecting geometry.\nFor this I propose to erode or reproject the render's depth to match the render sample of the new AA scheme. This should fix the second problem completely.\n\n## Technical details:\n- Implementation just need to be ported from Workbench which has a lighter form of TAA than EEVEE. Also [D10414](D10414) can be of great help to reduce render/convergence time.\n- Depth buffer modification can be implemented after new AA is in.", "Context menus for Grease Pencil layer visibility/lock icons are inconsistent between editors\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 26.20.100.7262\n\nBroken: version: 2.93.0 Beta\nWorked: (unknown)\n\nWhen selecting a Grease Pencil object, the \"Object Data\" tab of the Properties editor shows the object's layers, where each layer has (among others) an eye icon for visibility and a padlock icon for locking. These icons can not only be left-clicked to toggle their state, but also right-clicked to open a context menu for animating them (through keyframes or drivers).\n\nThe Grease Pencil mode of the Dope Sheet editor also lists the layers and also has these icons, but while left-clicking them does the same as in the Properties editor, right-clicking does not: it opens the context menu for the channel instead.\n\nFor the other Dope Sheet icons (\"Use Mask\" and \"Onion Skinning\"), the context menu works as expected.", "Snap support for Grease Pencil objects\n**Description**\n**Big picture:** Make it possible for the snap options: `Vertices, Edges and Face`, to work on Grease Pencil objects\n\n**Use cases**:\n* Precisely positioning a Grease Pencil object on a Mesh in the 3DView\n* More flexibility in positioning the 3D Cursor\n\n**Design:**\n* Will work in conjunction with 3DView's snap system (no new interface required)\n\n**Engineer plan:**\n* Since Grease Pencil objects can be 3D, a BVHTree system may have to be implemented to snap to faces and occlusion test.\n* Another option is to work with a depth buffer.", "Selected object color is weird in wireframe mode if the object color isn't single.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 465.89\n\nBroken: version: 3.0.0 Alpha\n\nOutline overlay seems to be drawn under wireframe.\n\n[#91619.blend](T91619.blend)\n\nImmediately visible in provided file.", "Pivot point \"Active Element\" not functional for grease pencil strokes\nBroken: 2.83.10\n\nPivot point setting \"Active Element\" is not functional when used for grease pencil strokes.\nThis mode is nice for foreward kinematic rotations of say legs.\n\n- Open attached file\n- In edit mode, select pressing shift to have selected points and one active.\nEven with pivot point set to \"Active Element\", notice that the Transform Gizmo does not jump to the last selected point (the active one).\n[pivot bug.blend](pivot_bug.blend)", "Texture painting does blend modes like overlay incorrectly when using 16/32bit images\nOperating system: Linux-5.17.15-76051715-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: Quadro RTX 6000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.48.07\n\nBroken: version: 3.4.0 Alpha\nWorked: unknown\n\n[texture_paint_overlay_16F_bug-2022-09-28_15.37.38.mp4](texture_paint_overlay_16F_bug-2022-09-28_15.37.38.mp4)\nHere you can see that even though the paint color is white, the overlay blend mode results in darkening of the gray base color, when painting with a low strength/at the edge of the brush falloff. This happens only for 16/32bit images and independently of the color space of the texture. It seems like brush falloff and strength are applied before the blend mode.\n\n- create 32bit/16bit image texture\n- draw with white, set to Overlay on shade of gray\n[16F_texture_paint_overlay_bug.blend](16F_texture_paint_overlay_bug.blend)\n\n", "Grease Pencil problem (especially soft pencil) \nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.09\n\nBroken: version: 3.0.0\n\nSince couple version of Blender I cannot use a soft pencil (mainly) brush in Grease Pencil 2D animation mode. The issue depends on how much You zoom. When You zoom in and use same size it's not visible. When You zoom out and draw with same size dots are visible. \nIt's behavior is very strange now and produces bad results (visible dotted lines or other which are visible on attached file. Some of these issues occur when the brush is small (line type set as dots). This is much more more visible when stroke stabilization is on. The bug was not there about 2 years ago and I was able to use the soft pencil without such problems. When line type is set to line the pencil stroke has strange contour around and when self overlap is off it produces artifacts on edges. \n\n\n\nJust choose a soft pencil and make it small. Check it with stroke stabilization as well. Then change the line type to line with self overlap on and off. \n\n", "Crash after using Grease Pencil object as object font\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.1\n\nWhen I use a grease pencil object for the object font of a text, blender crashes. I have had the same issue on a different computer with the version 3.1.0.[Test.blend](Test.blend)\n\n1. Add a text object\n2. Add a grease pencil object, for example a stroke, and put it into a seperate collection\n3. Put the instancing of the text to vertices (first image) and change the object font to \"Strok\", so it references the grease pencil object (second image)\n4. Repeatedly click the \"exclude from view layer\" symbol of the collection the grease pencil object is in\n\nThank you for your effort, I hope I filled everything out correctly\n{[F14114132](1.PNG), size=full}\n{[F14114131](2.PNG), size=full}\n\n\n", "Dynamic Paint sub-steps doesn't work with particles\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51\n\nBroken: version: 2.93.0 Alpha\n\nWhen using a particle system as a dynamic paint brush, the sub-steps setting of the canvas doesn't work. The sub-steps appear to work as intended with all other brush types. Whether the painting is done with vertex or texture maps doesn't seem to have an impact on the bug.\n\n* Open attached .blend\n* Bake the dynamic paint cache with a sub-steps value > 0\n* The result shows no sub-steps\n[Bug2.blend](Bug2.blend)\n", "Subdivision modifier artifacts in Blender 2.80\nOperating system: Windows 10\nGraphics card: 1080ti\n\nBroken: \n2.80 downloaded on the 3rd of Feb, 2019 as found on the download beta section\nWorked: (optional)\n\n\nSome artifacts around the out circle in the blend file, this is not present in Blender 2.79\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\nload attached file, doesn't matter if it is Cycles or Blender the artifact is present.\n\n[ukelele.blend](ukelele.blend)", "Decentralize .blend I/O for grease pencil modifiers.\nThis is a follow up task for #76372.\n\nFor normal modifiers this has been done already. Have a look at `ModifierTypeInfo.blendWrite` and `ModifierTypeInfo.blendRead` for more details.", "Wireframe opacity slider does not work.\nOperating system: Windows 10\nGraphics card: gtx1060\n\nBroken: 3.1\nWorked: \n\nWireframe opacity slider does not work if smooth wires is turned off.\n\nopen preferences>viewport\ndisable smooth wire overlay\nsave preference\nset viewport to wireframe shading\nin the overlay panel in viewport drag the wireframe opacity slider, the wire color will fade to black, the opacity won't change.", "GPencil: Texture Image Alpha pass through grease pencil strokes\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.83.1\n\nContext: Using texture image to make some kind of 2d fur on top of a GPencil Object\n\nThe Image Texture used to make some fur is mostly transparent and the brush used to make the colored part is a texture paint brush with Cloud Mask (it was made entirely in Blender with Blender tools).\n\nWhen on top of the GPencil object, the part of the Image Texture that is only half transparent(between 99%-1% transparent), makes the GPencil color disappear, even when moving the center point behind the one of the fur images.\n\nOn the contrary, when on top of another Image Texture (imported to be used as camera mapping), the Image Texture of the fur doesn't impact the transparency of that one.\nNotice that the Image texture boundary, even from the imported one, has problems with clipping and always makes a white line appear.\n\n- Open attached .blend file\n- Zoom in and out to notice the alpha transparency influencing the GP object\n\nThanks for your help!\n\n\n\n\n\n\n\n[bug_alpha.blend](bug_alpha.blend)", "Grease Pencil: Radius property misses blue slider background to visualize the percentage between soft min and max\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.81\n\nBroken: 2.80 -> 2.92.0 Alpha\nWorked: I believe it's been like this since initial GPencil implementation\n\nIn the GPencil editor the slider number button for \"Radius\" located in the topbar does not draw a percentage fill as is done in other editors such as Vertex Paint, Sculpt, Texture Paint and Weight Paint. \n\nStart blender with default settings\nAdd->Grease Pencil ->Blank\nSelect the Grease Pencil Object if not already selected\nSwitch from Object mode to Draw mode\nNote that the \"Radius\" topbar slider does not contain a fill color as does the \"Strength\" topbar slider \n\n[GreasePencilRadiusSliders.mov](GreasePencilRadiusSliders.mov)\n\nI've created [D9948](D9948) as a potential fix", "GPencil: Tint modifier in Fill mode, don't get restricted by vertex group influence\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.1\n\nTint modifer in Fill mode, don't get restricted by vertex group as influence. in other hand, it works well on stroke mode,i really need this to be fixed really soon as possible please, see video\n\n" ]

[ "Grease Pencil: Subtract mode is not working as expected (regression)\nThe substract mode when using masking is not working as expected." ]

[ "GPencil: Stroke not working with Holdout\nOperating system: Windows 10\nGraphics card: GeForce GTX 970\n\nBroken: 2.82a, as found on the splash screen\nWorked: Never\n\nSetting a collection as holdout in view layer options in cycles works as expected on grease pencil objects in the viewport render preview, but does not work on grease pencil objects when rendering. The entire grease pencil object is visible in the renders, whereas in the viewport render preview the parts of the grease pencil obscured by foreground objects set to holdout is alpha. \n\nCreate a grease pencil empty behind the startup cube. \nDraw some stuff. \nPut the startup cube in a collection and set the engine to cycles. \nFrom the view layers options for the new collection containing the startup cube, set it to holdout. \nTurn on render preview from the viewport shading options. Notice that the cube is invisible, along with the parts of the grease pencil that had been obscured by the cube.\nPress F12\nNotice in the rendered image that all of the grease pencil is visible, even the parts we had expected to be alpha masked from the startup cube holdout.\n\nSee the attached blend file.\n[HoldoutTrouble.blend](HoldoutTrouble.blend)\n", "Grease Pencil - drawing preview ignores canvas settings\nOperating system: Windows 10 Pro 64-bit operating system, x64-based processor\nGraphics card: NVIDIA GeForce RTX 2080 Ti\n\nBroken: 2.83 and 3.5 latest\n\nWhen previewing grease pencil (before pen release), the Stroke Placement settings do not render properly. It looks like the preview is stuck on \"View\" but renders correctly on pen release. In my GIF you can see the strokes clipping, but appear correct after pen release. Can this be fixed? It's easily repro'd and consistent in multiple versions\n\n1. Create basic Mesh.\n2. Create Grease Pencil object above Mesh.\n3. Set Stroke Placement to \"3D Cursor\" or \"Origin\" and set Drawing Plane to anything but View.", "Port Temporal AntiAliasing to Grease Pencil\nPort should be as simple as copying what Workbench does, with the addition of [D10414](D10414).\n\nSee parent task for more detail.", "Grease Pencil Tint / Also Vertex Paint on Grease Pencil Objects very slow with Cintiq Pro 24\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.4.1\n\nWhen i use Grease Pencil Tint with the a Cintiq 24 Pro it's very very slow. I have to wait after each stroke to continue. Drawing with the mouse is no problem.\nSame with vertex paint on a grease pencil object.\n\nWindows Ink and Wintab made no difference\n\nJust tinting with the cintiq stylus in the attached scene on the current layer.\n\n", "Bisect Tool Todo\n## Bisect\n * - [ ] {icon circle color=yellow} We should add an updated Gizmo here \n\n *Comment: should this gizmo's plane be an outline (design looks opaque), also, it's quite different to the previous design, see: #54661 ~ ideasman42*\n\n *Answer: This design is more similar to what we have now, but makes it easier to rotate and move the bisect plane along specific axes. The blue plane here should be transparent, not opaque. ~ billreynish*\n\n * - [x] {icon circle color=yellow} We should add these options to the tool settings: Fill, Clear Outer, Clear Inner. Otherwise, users have to keep setting these settings for every bisect operation.\n * - [ ] {icon circle color=yellow} We should add an option called 'Only Selected' just like the Knife tool. Currently it always only applies to selected elements, but that makes many successive bisects annoying because you constantly have to select all. \n", "Clicking on Driven decorator should open the Edit Driver popup\nIt would be nice if single-click on the driver icon would open the edit driver popover and double-clicking open driver editor. \n\nIn the keyframe icons by double clicking you could open in floating window the curve editor would be good, or ctrl + left click.\n\n\n\n", "Grease Pencil TimeOffset modifier work incorrectly\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: Radeon RX 590 Series ATI Technologies Inc. 4.5.0 Core Profile Context 22.20.27.07.221020\n\nBroken: version: 3.4.1\nWorked: (version: 3.2.2)\n\nAfter applying Time Offset modifier, the animation starts to twitch a little. This problem is observed when the animation starts from the zero frame.\n\nOpen the attached file. Start the animation (Time Offset modifier is enabled), we will see a slight twitch of the animation. Turn off the Time Offset modifier and we will notice that the animation has become smoother. If you set the animation range from 1 to 20, and move the animation frames of the Grease Pencil object to the right by one frame, the problem disappears.\nIn version 3.2.2, the modifier works correctly\n\n", "UI: Cursor flickering when hovering over properties\nBroken: version: 3.0.0 Beta\n\nWhen hovering from one property to another in property area, sometimes cursor changes from <--> to standard cursor. There are no instances where you stop moving it and it stays as standard one - it always changes to <--> after a fraction of a second. UI scaling does not affect it.\n\nNote that when tried without recording problem was more apparent, almost always cursor was flickering.\n[UI Cursor flickering take 2 2021-11-27 10-58-42-112.mp4](UI_Cursor_flickering_take_2_2021-11-27_10-58-42-112.mp4)", "GPencil: Modifiers don't work after separate strokes and use hide in outliner [issue with multiuser gpencil objects and modifiers]\nOperating system: Ubuntu 20.04\nGraphics card: Mesa Intel® HD Graphics 620 (KBL GT2\n\nBroken: (2.83.3)\n\n\nThe mirroring of a grease pencil object briefly stops being visible after performing a stroke separation.\n\n1. Boot up Blender's default 2D animation workspace\n2. Draw a stroke on the GP object \"Stroke\"\n3. Add a mirror modifier to the object.\n3. Draw another stroke\n4. Change to edit mode and select the new stroke\n5. separate said stroke with P, creating \"Stroke.001\"\n6. Still on edit mode, click on the 'eye' icon on the outliner to make \"Stroke\" invisible\n7. Click on the eye icon again, the mirroring will have stopped being visible.\n8. Clicking twice on the \"Display modifier on viewport\" button in the modifier tab fixes the issue.\n\n**Status**\nThe \"wrong usercount when separating\" is fixed, see 4251a87bf6, however the underlying issues with multiuser gpencil and modifiers still remains.\nTo reproduce:\n- add a gpencil stroke\n- {key Alt + D} duplicate linked\n- add a Thickness modifier.\n- Hide/Unhide the object in outliner.\n", "Pie menu sub menu cascade\nOperating system: Windows 10\nGraphics card: GeForce 3070 8GB\n\nBroken: 3.1.2 and 3.5 and 3.6 and 4.0.0(827baed610b8)\nWorked: didn't find\n\nWhen adding a multiple column sub .menu() in a pie menu, it doesn't draw properly, it offsets every row.\nIt does this with rows in a column, or columns in a row, and even with grid_flow(row_major=True).\nBut, when drawing .menu_contents() instead, it draw properly.\n\n\nExecuting the appended pie menu python script.\n\n", "Tool with Dragmenu is hard to select using Stylus\nOperating system: Linux-5.15.0-2-amd64-x86_64-with-glibc2.32 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.86\n\nBroken: version: 3.1.0 Alpha\nWorked: (long time ago if ever)\n\nThe tools in the tools panel on the left, which have mutliple tools assgined to on button, are really hard to select using a tablet. This is because they open the menu on a timer and on even tiny drag input. (The timer seems to interact strangely with the tooltip timer, causing it to open immediatly sometimes, but that is not part of this report). The drag threshold for detecting a drag seems to be not influenced by the settings in the preferences (#92995, #84726). Also the drag causes the selection to be canceled on release leaving it not selected, while on normal tool buttons it is finished on release, meaning they will be selected. That is IMO the real bug here.\n\nI checked that the correct drag threshold is used by `WM_event_drag_threshold` on my device and it is! But the tool in the toolbar which is explicitly mentioned here seems to be not using it.\n\n\n\nself explanatory, just drag the toolbar buttons and observe. Also play with the setting \"Tablet/Mouse Drag Threshold\" to see that it in fact doesn't change anything.", "Grease Pencil: draw pixels don't scale to pixels.\nCurrently grease pencil uses pixels in a very confusing way, since it doesn't actually relate to pixels.\n\nEither:\n\n- The size of a brush should use the current zoom level and use pixels as we do for all other draw tools.\n- We should use blender units for grease pencil brushes.", "Color swatch in grease pencil in the topbar is half transparant, showing whats underneat\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.79\n\nBroken: version: 3.5.1\n\nThe color swatch in the top bar in Grease Pencil draw mode is half transparant, showing what is in the viewport underneath. I don't want my color swatch to show anything but the solid color that I've chosen. \n\n\n\nPS\nThis is also present in the latest 3.6 beta", "Multi-Object Properties Editing\nIn Blender, we want to make it easier to set values for multiple objects, bones, strips, keys and other items at once. \n\nCurrently you can do this in limited circumstances, but you have to right-click on a value after you've changed it and pick Copy to Selected from the menu or hold Alt while editing values. This not only slow, but also very hidden, and the Alt-key conflicts with Emulate 3 Button Mouse. Lastly, it isn't clear at all to users which values work for multi-editing, and which values don't.\n\nHere's how we would like to solve it:\n\n\n## Selection vs Active\n\nThe concept of ***active*** in Blender means that we only show properties and only affect the active object or item. However, if we want to better support multi-item editing, we must start by making it so the Properties reflects the ***selection*** rather than the ***active*** item, of which there can only be one. Going forward, by default the Properties will be set to reflect the selection instead. This could be controlled in the Properties like so:\n\n{[F8363009](Screenshot_2020-02-22_at_17.34.54.png), size=full}\n*(The search box is added here to reflect how the Properties will be updated when that item is also added, but is a separate feature)*\n\nIf set to ***Active***, the Properties will continue to function as it does today, only showing properties for the active item.\n\n## When there are multiple selected items\n\nWhen the user has selected multiple items, a few things change:\n\nInstead of showing the name of the object at the top, Blender tells the user how many items are selected:\n\n\nIn the list of tabs, only properties sections are shown that are common among the selected items. For example, if you select a text and a mesh object, we don't show Modifier, Mesh or Curve Properties, since these properties only exist on a subset of the selected objects:\n\n\nThis rule also goes for individual panels. Here, both a Sun and Spot light are selected, and we only see properties that are shared between them. A note at the bottom clarifies that you are only seeing the shared subset:\n\n{[F8364914](Screenshot_2020-02-23_at_18.49.33.png), size=full}\n\n## When values differ\nWhen multiple items are selected, Blender will compare values to check if they are the same, or if they are different. If they differ, we change the way properties are displayed, like so:\n\n{[F8363037](Screenshot_2020-02-22_at_17.52.51.png), size=full}\n\n\n\n## Manipulating values across multiple items\n\n\n\nFor simply setting a value to be the same. you would simply type a new value, hit Return, and that value is then propagated to the selection.\n\nFor manipulating numbers and sliders relatively, you can just drag on them (like Alt-dragging today), or type the character '=', so the user can type '= -3' to subtract 3 from the value on all selected items.", "GPencil: Low hardness values always draws yellow\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.91.0 Alpha\n\nGrease Pencil creates yellow transparency with lower hardness values on brush.\n\n - Open in 2D Animation mode\n - Set to a \"Dots\" brush (eg. Dots Brush) its most noticeably shown there. It happens for the other ones too but this one is most noticeable\n - Set the Radius to something largish so we can see whats happening (eg 2000px), or zoom in \n - Draw a circle and you'll be able to see it\n\nIt looks like there's a faint yellow tinge transparency which builds up when putting the grease pencil strokes on top of each other.\nWhat it looks like:\n\n\nExample blend\n[yellowtransparency.blend](yellowtransparency.blend)\n\nBasic repro gif\n\n\nMe shading a ball\n\n\n\n\n\n" ]

[ "Move all Grease Pencil Settings to the Toolbar Tab\n## Current Situation\nThe grease pencil is presently split between the toolbar and the properties in the 3D View. In other editors it's only in the properties panel and there's no access to the draw tools except with hotkeys. \n\n**This has a couple of problems:**\n- There's a bad visual disconnect between the tool actions and the tool settings. I believe most users expect the settings for their present tool to be located next to the tool its self. This makes the learning curve worse than it should be and also forces the user to move back and forth from one side of the screen to the other when trying to draw with the grease pencil and then make changes (assuming they're not using hotkeys). \n- The grease pencil panels are inconsistent across editors. The 3D View has both the toolbar panel and the properties panel. Whereas the Image Editor, Node Editor, Movie Clip Editor, and probably others only have the properties panel. \n\n## Proposed Improvement\nTo fix this inconsistency and disconnect between the grease pencil tools and it's settings, I suggest we merge both the Grease Pencil Draw panel and the Grease Pencil properties panel into a single section, under the **Grease Pencil Tab** in the Toolbar. \n\n\n\nThis was discussed a fair bit in [D84](D84) and seems to have general consensus.\n\n### Advantages\nCombining the Grease Pencil panels in this way improves a few key things:\n- The tools and settings are both accessible in the same place, making it much easier for users to find and utilize. This is particularly true for new users. \n- Save space by utilizing the new toolbar tabs, keeping everything in one place and giving more room to the properties region (which is overwhelming right now).\n# Can easily be made consistent across all editors without affects existing layouts and organization. Simply add the Grease Pencil tab to each applicable editor.\n\n### Disadvantages\nThe main point I can think of for disadvantages is that doing this breaks the Tool / Settings separation paradigm that was put in place for 2.5. However, as discussed in [D84](D84), this has proven to have some impractical consequences (like this exact issue) that don't really make sense to keep so far separated. \n\n## Development Tasks\nIf accepted I'm happy to implement these changes myself. \n\n" ]

[ "overlays extras option being ignored\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.1.0 Beta\nWorked: (unsure, definitely 2.8 possibly 3.0)\n\nbounding box representation of particle instanced geometry (hair) where the object being instanced is set to bounding box is not being hidden from the viewport when the 'extras' option is checked. \n\n\n", "Simulation Nodes: show bakes in outliner\n_No response_\n\n", "2.8 - Particles in multi-level nested collections not visible\nOperating system: win10x64\nGraphics card: AMD Radeon Pro WX 7100\n\nBroken: 2.80, 54ffc4e19dc4, 2019-02-25\n\n\n\nIn nested collections, objects with particles work only in two levels of nesting. Particles in further levels of nesting seem to ignore collections instancing.\n\n\nRelated file: [particle-instances.blend](particle-instances.blend)\n\n- create cones with hair particles on a collection called **branches** \n- create a cube on a collection called **tree** \n- add a collection instance of branches within the collection called **tree** (1st level of nesting of branches) \n- create a new collection called **trees** and add a few collection instances of tree. \n- for all objects in trees, the particles are visible ( 2nd level of nesting) \n- create a new collection called **instances** and add a collection instance of trees. Here, the particle settings are not replicated. (3rd level of nesting) .\n- For the sake of testing this further, create a new collection called **instances2** and add a collection instance of instances . Also, here the particle settings are not replicated. (4th level of nesting) \n\n\nIn collection Instances, only the initial particles are visible and their location seems to be reset based on the initial location\n\n\n", "Improvements to \"Bake Action...\" Operator\nWith the developments happening in the blender animation module more robust tools and feature are being needed.\n\nThe \"Bake Action...\" operator is a rather useful tool that feels incomplete:\n\nHere are some options that would take it to the next level.\n\n\n - Options to set between Current frame range, preview range and custom range.\n - Frame step should only output the desired frame step, currently it merges the current keys with the frame step so in example: if I have keys on ones and I bake on 2's the current result is ones as there are already keys there. It should remove the keys (in the baked range) and replace them with the new keys on 2's.\n - When baking, the resulting keys should obey the default keyframe interpolation type set in the animation preferences. (if its set to constant, all the resulting keys should be constant, if its linear the resulting keys should be linear)\n - A \"smart bake\" which should only perform the baking action where key frames already exist in the selected objects/bones (up for discussion but maybe going up one level up in the chain and transfer those keys to the object would be good too.\n - Include \"custom properties\" in the result of the bake, currently custom properties lose their keys. (and anything that can be stored stored in the action)\n - option to \"merge actions\" that will bake the animation data from all the available actions and put it in the main \"non pushed\" action and delete the other actions (without needing to go into the nla and get rid of them manually)\n\n\n\nThese seem from my ignorance very minor features that would make the operator much more useful in production.\n", "Reflection of smoke on object surface has white box surrounding \nOperating system: win10\nGraphics card: GTX TITAN X\n\nBroken: (2.82, 2.83.10, 2.9, 2.91)\nWorked: (none)\n\nReflection of smoke on object surface has white box surrounding when a back light is added.\n\nLooks fine with front light only\n\n\n**With back light added:**\n2.83.10: Domain Empty Space = 0\n\n\n2.83.10: Domain Empty Spcace = 0.000001\n\n\n2.91: Domain Empty Spcace = 0\n\n\n2.91: Domain Empty Spcace = 0.000001\n\n\nSet shot with back light highlighted\n\n\nbake domain and render at around frame 200+ (when smoke touches the black object)\n\n[TEST.blend](TEST.blend)", "Unable to get object to stay hidden while new mesh tool is active.\nWindows 10, 1060\n\nBroken: 2.80 master (also reproduced in a month-old daily build, so not new behavior).\n\nIf you have an active tool like Add Cylinder going, and you hide another object using the eye icon in the outliner, then as soon as you touch the cylinder parameters, the hidden object becomes unhidden.\n\n1) From the default scene, Add a Cylinder to the scene and expand the tool properties in the lower left of the 3dview.\n2) Because the cylinder is hidden inside the cube, we can't easily see the effect of changing the tool parameters, so go to the outliner and click the eyeball icon associated with the cube to hide it in the viewport.\n3) Change any parameter in the cylinder tool properties pop-up in the 3dview. The cube reappears and the eyeball indicates it is no longer hidden.\n\nBasically any changes made to the Outliner visibility icons gets reverted as soon as you change any active tool property, (at least for the Cylinder tool). This at least unexpected behavior.", "Lifetime-Keyframes for particle system are ignored when rendering a baked particle simulation\nOperating system:Windows 10\nGraphics card:GTX1070\n\nBroken: Blender 3.5\nWorking:\n\n\nKeyframing the Lifetime of particles inside of a particle-system and baking the system (Both bake/bake all dynamics) leads to only the first keyframe being utilized when rendering an animation. For both the viewport and single-frame renders it works as expected.\n\n1. Create a simple particle system.\n2. Keyframe the Lifetime value at **1** on frame 30, and value **50** on frame 31.\n3. Bake all particle dynamics.\n4. Observe how the particles disappear after 1 frame up until frame 31, where they then last for 50 frames.\n5. Render an animation of the particle system.\n6. The particle system will only utilize the first keyframe for the duration of the animation, in this case 1. So even after frame 31 the particles will despawn after 1 frame.\n\nThe blendfile has a particle system setup, you merely need to look at the behavior in the viewport vs rendering an animation.\n\nThank you.\n\n", "Keymap changes are not reflected for the local view setting\nOperating system: Linux-5.18.10-arch1-1-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.57\n\nBroken: version: 3.3.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `rBUnknown`\nWorked:\n\nWhen I turn off frame selected in keymap -> local view and restore it to its default it keeps the old behavior until blender is restarted.\n\nI tested changing delete in object mode (x) key to disable confirm and when I restored it it worked as expected so I think it is related to local view specifically.\n\n- Open default scene\n- In Preferences > Keymap: search for local view and uncheck `Frame selected` checkbox\n- use the shortcut in viewport. Notice frame select operation is not performed (which is expected)\n- In keymap, restore changes\n- Again use shortcut to switch to local view (frame selected operation is expected to execute but it doesn't)\n\n\n", "Hair Dynamics: Reduced animation frame count leads to loss of Hair Particles on Collection visibility updates\nOperating system: Windows 10 Home 64\nProcessor: AMD Ryzen 5 4600H with Radeon Graphics, 3.00 GHz\nRAM: 16.0 GB (15.4 GB usable)\nGraphics card: NVIDIA GeForce GTX 1650 Ti\n\n\nBroken: 3.4.1, 3.3.5\nWorked: 3.3.x (before 3.3.5)\n\n\nWhen applying Hair Dynamics with a reduced simulation frame count, the Hair Particles \"disappear\" on updating the visibility of any Collection in the Outliner after successfully baking.\n\n\nBased on the file hairbaketest_72frames_notOK.blend (attached).\n\n* Add hair to default cube\n* Enable hair dynamics\n* Set cache end frame to 10\n* Bake\n* Hide and unhide collection\n\n**Expected Behaviour**\nThe Hair Particles are still visible, the simulation can be played. After closing and re-opening the file, the current saved frame will show the Hair Particles at that frame as in the bake. \n\n\n**Actual Behaviour**\nThe Hair Particles are \"gone\", not visible anymore. After closing and re-opening no Hair Particles are shown. \n\n\n**Additional Information**\nThe file hairbaketest_250frames_OK.blend (attached) shows the successful workflow, it saved the baked state successfully and I could toggle the visibility of Collections in the Outliner without any effect on the Hair Particles. The workflow described in the steps is as it worked for me for a long time. I tried to break down the issue and the reduced frame count seems to make the difference for all I can tell.\n\n\n**Windows Update**\n* Last time I used Hair Dynamics workflow successfully: 3/5/2023 (March 5th) with Blender v3.3.x\n* First time I wanted to use it again but the bug occured: 3/25/2023 (March 25th) with Blender v3.3.5\n\nBetween these dates a Windows Update was made. See attached Screenshot of my Windows Update history. The reason I think it's worth mentioning, is that I did a lot of testing and as I have to use Hair Dynamics for a project and it's not possible, I tried and installed older versions of Blender that definitely worked without any issues (back to 2.93) but even the older versions now have this behavior when I try to make new files with Hair Dynamics.", "Alpha Blend material has inconsistent behavior with Geometry nodes\nOperating system: win 10\nGraphics card: 1050ti\n\nBroken: 3.4.1\n\nAlpha Blend material renders differently on the object, which is referenced inside Geometry Nodes tree of another object, despite this another object is already hidden and has muted GN modifier.\n\nIn attached file:\n* Remove/Select *Shell* object inside GN tree of object *Mesh*\n* Activate/Deactivate *Viewport Disable* for *Collection* or object *Mesh*\nIn both cases material becomes transparent or not\n\n", "Strange shadow when rendering with GPU device\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.96\n\nBroken: version: 3.2.0\n\nThere's a shadow in the scene that doesn't make sense.\n\nI've attached a Blender file that I've tried to strip down as much as possible. Here's what I get (notice the strange shadow bottom middle)...\n\nThe odd thing is that if I delete/hide random objects in the scene, the shadow disappears.\n\nExample... hide any of the following objects:\nMISTY_MOUNTAINS_39_2k_c.001\nPlane.001\nAny object in the atmos folder\nObject_4.001\nthe whole quixel folder\n\nI know the shadow shape is cast from the volumetric cloud container shape (a spherical shape) but it's behaving very oddly and, as I use volumetrics in every single scene I ever create, I'm familiar with them behaving oddly. This is just a new one.\n\n[shadow error.zip](shadow_error.zip)\n", "3D Paint Mode stencil disappears when panning viewport.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.76\n\nBroken: version: 2.93.4\n\nWhen panning viewport in 3D paint mode with stencil mask enabled, the stencil mask disappears when moving the viewport, however does not disappear when zooming it, which indicates it's not a technical limitation.\n\n1. Enter 3D paint mode\n2. Enable texture mask, select some texture and set mapping to stencil\n3. Navigate the viewport\nResult: The stencil disappears when moving the viewport, which makes it very difficult to line up stencil projection with the 3D geometry\nExpected: The stencil does not disappear when navigating the view\n\n", "Fire does not render if flow object had disabled Show in Renders\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.19\n\nBroken: version: 2.83 (sub 2)\nWorked: (optional)\n\nFire does not render if flow object had disabled Show in Renders\n\n[F_F_I_R.blend](F_F_I_R.blend)\n\n- Open file\n- Bake domain cache\n- Render\n", "Workbench: Cull before batch creation during viewport animation rendering\n\n## Current Situation\n\nWhen drawing an object (eg sending draw calls to the GPU) the object boundary sphere is tested against the active view). When the test fails the object isn't send to the GPU. Before this happens the GPUBatches were already generated.\n\nWhen doing animation playback this could lead to the situation that the GPU Batches are calculated, but not used.\n\n## Target Situation\n\nWhen using the Overlay engine/Workbench engine we could test the objects to the default view (e.g viewport). When this fails we could skip this object at all. This feature would be disabled when the shadows are turned on and only enabled when doing an image/animation render.\n\nA consistent framerate is more important to animators then having some frames faster.\n", "Cycles Bake Selected to Active with transparency gives undesirable results\nSystem Information:\nOperating system: Windows-8.1-6.3.9600 64 Bits\nGraphics card: GeForce GT 635M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\nBroken: version: 2.81 (sub 16)\n\n\nBug or unintended behavior:\nBaking using \"Selected to Active\" does not take into account transparency \n\nHow to reproduce:\nCreate two objects with materials and textures assigned to each;\n```\n One texture to be baked to\n One texture with an alpha channel (optionally a normal map)\n```\n\nIn the shader editor, modify the material of the object with a transparency channel\n```\n Join the transparent textures' alpha into the alpha of the shader that is being used\n (optionally join the color of a normal map)\n```\n\nIn render properties, enable Selected to Active and use an appropriate ray distance\n```\n Finally bake.\n```\n\nResult:\nWhen baking combined or most channels, the transparent areas will assume the color as if it were completely opaque,\nWhen baking normals with transparency, the transparent areas will mix with the normal of the geometry that is being baked to the active selection.\nBoth outcomes are unfavorable when working with transparent cards/ribbons.\n\n[bugreport.blend](bugreport.blend)\n\nI packed the textures edited the project to make the issues that I am having more clear.\n\n{[F8286028](image.png), size=full}\nThere are now 3 overlapping objects with transparency.\nAfter baking normals or color, none of the colors or normals behind the transparent objects appear because the object appears to draw over top without transparency.\nThis example also shows that with full normal transparency, the normal will then be the geometry normal of the object rather than a flat normal or anything that is below the object." ]

[ "Hidden objects reappear after rendering.\nBlender Version: blender-2.80-ecdd1864d162-linux-glibc224-x86_64\nI have some objects which are hidden in viewport. After rendering and closing the render window the hidden objects appear while they are still depicted as hidden objects in Outliner panel. After clicking on the eye icon double times the objects become invisible." ]

[ "Interface. Sculpt mode, checkbox title not react to click\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: since 2.83.9, checked with version: 2.93.0\n3.0 alpha broken too\n{[F10178598](изображение.png) size=full} — here\n\n\n[2021-06-18_00-08-00.mp4](2021-06-18_00-08-00.mp4)\n", "Remesh modifier causes splitting wrongly\nWindows 7 Professional Service Pack 1 64-bit\nIntel Core i7-2600 3.40GHz\n8GB RAM\nNVIDIA GTX560\n\nBroken: 2.77 22a2853\n\nRemesh modifier causes splitting wrongly.\nIt happens when remesh modifier's scale value is 3/4, 1/2, 3/8, 1/4, 3/16 etc.\n\n[RemeshGlitch.blend](RemeshGlitch.blend)", "Texture Mapping issues on large brush radius\nOperating system: Linux-6.4.6-gentoo-x86_64-AMD_Ryzen_Threadripper_1950X_16-Core_Processor-with-glibc2.37 64 Bits, X11 UI\nGraphics card: AMD Radeon RX 7900 XT (gfx1100, LLVM 15.0.7, DRM 3.52, 6.4.6-gentoo) AMD 4.6 (Core Profile) Mesa 23.3.0-devel (git-50c29e1ffa)\n\nBroken: version: 4.0.0 Alpha\nWorked: --\n\n\nWhen a large brush radius for textured brushes in sculpt mode, the outer-most affected geometry doesn't apply the texture fully. This is especially noticeble on the Clay Strips and Paint brushes when the stroke method is set to anchored.\n\n\n\n\n\nBlend file = [bug.blend](attachment)\n\n- Open the File\n- Use the provided paint brush and clay strips brush on a subdivided cube or suzanne head like shown\n- Notice how the outer edges of the falloff is not updating the geoemtry fully\n\n", "Right clicking on 3D Viewport Header not working as expected\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 850M/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 537.13\n\nBroken: version: 4.0.0 Alpha\nWorked: (3.6 and older)\n\n\nIf you Right click on empty space of 3D Viewport Header it will not bring Header Menu for disabling or enabling Header. it was possible in version 3.6\nThere are just 2 options to bring menu:\n1- Right click very close to Header items but not inside items(it's really hard to reach)\n2- disable region overlap (many prefer not to turn it off)\n\nI know it is now available to see and select objects through header but I think old right click should back to header because:\n1- constancy( when you right click on all other editor's header that menu will pop up)\n2- beginner friendly (imagine someone new to blender or blender 4 wants to hide header and he does not know new region toggle pie menu so it is hard to hide it)\n\n", "Crosshair cursor is too small when sculpting on 4k resolution.\nOperating system: Linux-5.4.0-48-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.91.0 Alpha\n\nOn 4k monitor, while sculpting, the cursor is barely visible.\nIf the mesh is grey/white, you cant see the cursor.\nSame on Vertex Paint and Weight Paint modes.\n\nwatch?v=fpADGvD47hE\n\n- On Edit Mode however the cursor is big and quite visible.\n", "Mantaflow replay length change doesn't take into account effectors.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.92\n\nBroken: version: 2.83 (sub 15)\n\nWhen you run a mantaflow simulation with replay, then increase the length of the simulation, collision effectors fail\n\nOpen file, Shift + right arrow to go to initial frame and make folder. play to bake till frame 20. (it'll be quick)\nPause, change frame end for baking to something bigger: 50. \nShift + right arrow to go to first frame. Play. After 20th frame, see effector being ignored & fluid rushing in.\n[#76075.blend](T76075.blend)", "Regression: Armatures [in front] overlay doesn't update properly (toggling overlay ON and OFF)\nOperating system: Linux-5.15.0-52-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.103.01\n\nBroken: version: 3.4.0 Alpha\nBroken: 2.92.0\nWorked: 2.91.2\n\nBisecting gives 0f95f51361 as the culprit commit\n\nif the armature viewport mode is set to \"in front\" it doesn't properly update when the overlays are turned off and on\n[Peek 2022-11-01 05-35.mp4](Peek_2022-11-01_05-35.mp4)\n\n1-add an armature\n2-set the \"viewport display\" to in front\n3-turn off and on overlays\n4-if you can confirm the bug the armature should appear after changing anything in the 3d view to update as in the video, move rotate or zoom \n\n", "VSE Sequencer timecodes display almost correct values\nOperating system: Linux-5.4.0-94-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.86\n\nBroken: version: 3.0.0\n\nsometimes the timeline displays 1m60s instead of 2m. this is kind of correct but actually not correct. I have no idea what could cause this, this appears on various zoom levels and only on specific positions that seem to be fixed positions.\n[superstious.mp4](superstious.mp4)\n[2022-01-14 20-43-02.mp4](2022-01-14_20-43-02.mp4)\n\n\n[#94920.blend](T94920.blend)\n", "Use Scene Spacing Distance by default\nNOTE: This design is still being investigated and discussed. \n\n# Issue\n\nThe default for brushes in every mode has always been to use view based spacing for each stroke.\nThis wasn't always an issue since brushes used to work via a projected falloff.\n\nBut when using a spherical falloff on curved surfaces this can lead to inconsistent spacing artefacts.\nSince more painting and sculpting modes are implementing brushes with a spherical falloff by default, it is a reasonable adjustment to make the spacing also scene based.\n\n# First Steps\n\nBefore changing any defaults we should investigate any remaining issues, downsides or bugs if this behaviour would be implemented by default for all brushes.\nOnce it is clear that this change would be an overall improvement, and how feature might need to change.\n\n### Fix remaining bugs and issues\n\nFurther issues from using Scene Spacing need to be investigated.\nThe currently known reports are:\n- #98111 (Sculpt mode: Draw Sharp brush artefacts when using Scene Spacing in stroke settings)\n- #99638 (Line stroke method not working with Scene spacing)\n- #99642 (Clay strips brush freeze when coming from outside the horizon of surface curvature)\n- #99644 (Draw Sharp brush stops on anchored during the stroke)\n\n# Proposal\n\nScene spacing currently accomplishes two distinct behaviours:\n1. The spacing is consistent in 3 dimensions\n2. The spacing is consistent in world space (especially when zooming in/out)\n\nInstead of just making Scene Spacing Distance the default, we should change the \"View\" scene spacing to work 3 dimensionally for a better brush behaviour in general.\nThis could be accomplished by associating these two different behaviours with distinct settings:\n\n1. Setting the falloff to be Spherical or Projected = 3D or projected spacing distance\n2. Switching between View & Scene spacing distance = The spacing is a percentage of the relative brush size diameter or an absolute world size\n\nThe UI should then also reflect that change, by changing the spacing slider value:\n- View = Percentage\n- Scene = Distance in units\n\nThis updated View spacing then needs to be integrated in every mode for 3D painting/sculpting.", "Lookdev HDRIs are interpreted incorrectly when using different colour spaces\nOperating system: Any\nGraphics card: Any\n\nBroken: 2.82 (sub 1), type: Release\n\nHDRIs installed into Blender for use in Lookdev Mode don't offer an option to set their colour space.\n- Blender probably assumes they are encoded in Rec.709, but this doesn't hold up if they're not.\n- Alternatively, when using a different OCIO config, the Rec.709 colour space is almost never listed as just 'Linear', so Blender picks a different space from the config.\nBoth scenarios cause undesired hue shifts. This is most probably also true for MatCaps, but to be fair I haven't looked yet into how they work.\n\n\n- Install a differently encoded HDRI into Blender, for example one of the EXRs from the imageworks/OpenColorIO-Configs repo on GitHub.\n- Switch the viewport to lookdev mode and activate the newly installed HDRI. The colours will have shifted.\n\nor:\n\n- Start Blender using an alternate OCIO config, such as one from said imageworks/OpenColorIO-Configs repo.\n- Switch the viewport to lookdev mode and try the HDRIs that ship with Blender by default. Those colours will also have shifted.", "Misuse of 'U.pixelsize' in code\nOperating system: macOS-12.6.3-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon Pro 455 1.2\n\nBroken: version: 4.0.0 Alpha\nWorked: -\n\n\nThe `U.pixelsize` is the user preference for line width in the UI. Go to \"Preference\" > \"Interfece\" > \"Line Width\".\nChanging the \"Line Width\" has an undesirable effect in some cases.\n\n- [ ] Zooming value with MMB-drag in Clip and Image editors\n- [ ] Dynamic-topology Detail Size\n- [ ] Width of tooltips block\n\nThere's an old list of places that need to be re-checked:\nD9604\n\n", "Cycles: Bump node doesn't distort normals\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.59\n\nBroken: version: 2.83 (sub 10)\nWorked: Never\n\nHeight input have no effect on bump node output in cycles.\nThe main problem, that Generated texture cordinates give seams with bump. That is why I desided to try Normal input.\n\nEevee:\n\n\nCycles:\n\n\nHere the example file.\n[WhereNormals.blend](WhereNormals.blend)\n", "Geometry Node: Changing modifier input values with python no cause update\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.72\n\nBroken: version: 2.93.0 Alpha\n\n\nChanging modifier input values with python no cause update.\nBy code:\n\n```\nbpy.data.objects[\"Cube\"].modifiers[\"GeometryNodes\"][\"Input_1\"][0] = 13.8\n```\n\nValue in modifier property was changed. User interface no updated before modifiers panel are not redrawn. Result of modifier didn't updated.\n\n\n\n- Open attached file\n- Run the code `bpy.data.objects[\"Cube\"].modifiers[\"GeometryNodes\"][\"Input_7\"] = 1.0` in the python console\nThe modifier UI may change, but the Viewport does not.\n[geom_node_py_bug.blend](geom_node_py_bug.blend)\n\n", "Blender 2.8 Defaults\nThis is a list of things we should change about Blender's default settings and behaviour.\n\nThe list will grow and change over time.\n\n\nCommits:\n\n- 4d5b7696cb - prefs\n- bcf6cc1f6b - startup\n- Update: default preferences are now maintained in: `./release/datafiles/userdef/userdef_default.c`\n\n## Preferences\n\n- - [x] Python tooltips OFF\n- - [x] Auto Perspective ON ***\n- - [x] Navigation Manipulator ON ***\n- - [x] Region Overlap ON (Could be removed completely, now that it's optimized) ***\n - To make transparent toolbars work better, we would like to:\n - Make the toolbar region height be limited to the contents.\n - Make the toolbar region a separate theme\n\n## Default Settings\n\n- - [x] Default Shader should be set to Principled BSDF\n- - [x] Default Lamp increased strength (10x stronger)\n- - [ ] Playback set to AV Sync (reverted, breaks physics caching)\n- - [x] 3D View Lens = 50mm\n- - [x] Camera film size = 36x24mm Full Frame\n- - [x] Camera Focal Length = 50mm\n- - [x] Render Size Percentage = 100%\n- - [x] Render Display = New Window\n- - [x] Scene Units = Metric\n- - [x] Color Management View = Filmic\n- - [x] Workbench Object Overlap = ON\n- - [x] Absolute Shape Keys Interpolation = Linear\n- - [x] Curve Fill Mode = Full\n\n## Default Renderer\n\n- - [x] Viewport = Workbench + Studio Lighting\n\n\n## Layout\n- - [x] Headers on top for all editors, except the Timeline at the bottom\n- - [x] Default Properties tab = Object Properties\n\n## New Objects\n- - [x] Generate UV's = ON\n\n## Default Theme\n\n- - [x] Flat\n- - [x] Dark UI Chrome\n- - [ ] Universal widget radius\n- - [x] Transparent header in 3D View\n- - [x] Transparent toolbar area in 3D View\n\n## Tools\nVertex Slide:\n- - [x] Correct UV's = ON\n\nExtrude Along Normals:\n- - [x] Even Thickness = ON\n\nLaplacian Smooth:\n- - [x] Lambda Factor = 1.000\n\n## UV/Image Editor\n- - [x] Normalized Coordinates\n\n## Weight Painting\n- - [ ] Auto-Normalize = ON\n\n\n## Cycles\n- - [x] Dithering = 1\n- - [x] Image Sequence Auto Refresh = ON\n\n\n## New Objects\n- - [ ] SubdivisionSurfaces = ON (via OpenSubdiv)\n- - [ ] Shade Smooth = ON (Wait till we have OpenSubdiv)\n\n## Preferences\n - - [ ] Compute Devices ON (Postpone until we can detect crappy GPU's)\n - - [x] Release Confirms Transform\n\n## Text Editor\n - - [x] Line Numbers ON\n\n## Addons\n - - [ ] All exporters should be set to Only Selected by default", "Zooming in and out of the node viewport, text keeps changing its size\nOperating system: Arch Linux\nGraphics card: NVIDIA GTX 980\n\nBroken: Blender 2.90.0 Stable Release\nWorked: I'm not sure if this has ever worked correctly but it looks ugly.\n\n\nWhen zooming in and out of the node viewport, the text keeps changing its size making it look like it's jumping all around. I think the view should only zoom in and out without ever changing the text's size.\n\n\nAdd a principled BSDF node and then zoom in and out.\n\n[nodetitle.webm](nodetitle.webm)\n\n" ]

[ "sculpt mode SHIFT+R shortcut and unit incosistentcy\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51\n\nBroken: version: 2.91.0\n\n\nIn sculpt mode SHIFT+R shortcut doen't respect the Unit system and Length\nAs you can see in the attached image the value is still displayed in meters, while on the remesh tab is correctly displayed as centimeters.\nThis can cause problem when working on objects smaller than 1 meter, in fact to get enough resolution we have to scale down the voxel size to 0,0x or even 0,00x values, which are hard to control using the shift+r shortcut.\n\n\nIn a new file set a different type of Unit System or even unit length to see the incosistentcy,\n\n" ]

[ "Assigning shortcut to Sequencer data is wrongly 'hooked' the Scene instead of Scene's VSE data\nThis error shows all the time when I right click on the viewscreen, and I have to right click and press ESC, so the message disappears and I keep working this way, right click, a message appears, ESC.\nIt suddenly started to happen and even deleting and downloading blender again, even with the other 2.8 version it happens.\n\nThank you, I love your work guys\nJust trying to help\n\nOperating system: Windows 10 Home, Core i7 - 7500U @ 2.70GHz, 8.00 GB RAM\nGraphics card: AMD Radeon R7 M440 32GB, 4.00 Gbps\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\nWorked: (optional) ", "Transform Node Messes Up Alpha Channel Manipulation\nArch Linux\nNvidia 750ti\n\nBroken: 2.79 (Tried 2.8 as well and I have the same issue)\n\nAs referenced here:\n4\n\nWhen I take an image in the compositor and scale it up 2x or whatever and then separate the RGBA and dilate/erode the alpha and use that to set the alpha on the image, it only works within the size of the original image and not the scaled up image. Also even if I use the R, G, or B, it still only uses the size of the original image and not that of the scaled up one.\n\n\n\nCropped if I try to mess with the alpha after a transform.\n\n\nUncropped if I mute my changes to the alpha.\n\n\nAlso I believe that if you scale down\n\n[dilatecropissue.blend](dilatecropissue.blend)\n", "Driven values do not change with units\nOperating system: Windows-10-10.0.20206-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.90.0\nWorked: No working versions (at least >= 2.79)\n\nWhen creating a driver from an unlabeled value, pasting it into a field that accepts units of measurement will always use meters rather than the Scene unit.\nThis means that for a scene using inches, a driver with a value of 1 will translate to approx. 39.\nWhile this prevents things from changing when units change, this is not ideal as exactly re-converting in an expression is impossible.\n(It is highly unlikely someone working specifically in inches will want to input their values in meters.)\n\n- Create a new project, and change the units. (in this case, to Inches)\n- Copy a new driver from an unlabeled value. (ex. array count)\n- Paste the driver into a field with a unit label. (ex. constant offset value)\n\n[#80649.blend](T80649.blend)", "Auto Render doesnt work with scene animations\n{[F8575831](autorender.gif)}Operating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.92\n\nBroken: version: 2.83.0 Beta\n\nIt should render when the scene animation updates. In general auto render also seems a bit unreliable. eg. Sometimes I turn it off and it keeps rendering, and it doesn't render even if some changes are made to the scene. Anyway in this case it doesn't update when there're scene animations.\n\nFrom startup:\n1. Set \"Use nodes\" in compositor\n2. Turn on \"Auto render\"\n3. Animate default cube (I used rotate here)\n4. Scrub timeline\nIt doesn't auto render with new changes.\n\nWith attached blend:\n- Scrub timeline\n\n[autorender_animation.blend](autorender_animation.blend)\n\n\n", "Cycles: overlapping VDBs cause rendering artifacts\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.79\n\nBroken: version: 3.1.2\nWorked: version: 3.0.1\n\nFirst reported here: #97062\nOverlapping VDBs cause black rendering artifacts.\nThis is likely a known issue or related to a known issue.\nFrom the release notes for 3.1: [Ray Tracing Precision](Cycles#Ray_Tracing_Precision)\n>Many artifacts from rendering small, large and far away objects have been eliminated. There will always be object scales where numerical precision becomes a problem, but it's further out now.\n>\n>There can still be artifacts with rendering overlapping geometry, in some cases more severe than before. Such overlapping geometry should be removed, or have a small distance added in between.\n\n\n|3.0.1|3.1.2 CPU|3.1.2 GPU|\n| -- | -- | -- |\n||||\n\n1) Import a VDB volume.\n2) Duplicate the volume without moving it.\n3) Render with Cycles.\n\nHere is the test file:\n[#97094.zip](T97094.zip)", "Cloth physics cache gets reset by changing object materials\nOperating system: Darwin-19.4.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.8.24\n\nBroken: version: 2.83 (sub 11)\nWorked: unknown\n\nWhen using a cloth physics simulation, if you add or remove a material from an object, the cloth system cache is reset (\"cache is outdated\"). I don't think the material affects the cloth physics in any way (as it seems you can change any parameter within the material without the cache resetting), and therefore this should probably not happen. Please correct me if I am wrong and the material does affect the physics somehow.\n\nIn the default scene, enable cloth physics for the cube. Let the playback run for a few frames, building up the cache. Add a material to the cube; now the cache gets reset.\n\n", "View zoom (scale mode in particular) doesn't work well when dragging horizontally, or crossing the middle of the screen\nBroken: 2.79, 2.80, 2.81 HEAD (8c9e6f123a2).\nWorked: ?\n\nNavigation Mode Scale Is Broken - it should never be calculated from the vector zero.\n\n- Set Blender Navigation to Scale\n- Try to press Alt + click and drag in viewport to make it zooming (or click and hold the \"zoom\" icon in the viewport).\n- Try to cross the mouse arrow in the vector zero (i.e., cross the mouse around the middle of the viewport).\n- As you can see the tool start to work weirdly inverting the zooming as soon you cross the scene vector zero.\n\nNote: The other zoom modes (e.g., dolly) also present a strange behaviour when zooming horizontally. But the zoom seems to reset when going over the origin regardless of the mouse direction.\n", "Linked data-block has \"pointer\" property still active\nBroken: version: 3.4.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `rBUnknown`\nWorked: Probably never\n\n\n\n[torus.zip](torus.zip)\n\nOpen torus.blend. It has a linked collection containing the selected object `Torus`.\n\nAs you can see in the image, the \"Object\" rna properties (parent, and the array modifier one) are both active:\n\n{[F13627034](image.png), size=full}\n\nIn fact you can even click on the property to change it (the popup show up, but clicking on it at least doesn't change it):\n\n{[F13627039](image.png), size=full}", "Armature Object's \"Display As\" does nothing\nOperating system: Linux-5.3.0-26-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 435.21\n\nBroken: version: 2.82 (sub 6)\nWorked: version: 2.79 release.\n\nWhen I add an armature and change the Viewport Display Mode, nothing happens.\n\n\nOpen the blender;\nDelete the default objects;\nAdd an Armature, single bone;\nIn the Properties Panel, Object Properties, Viewport Display:\nChange 'Display As' to anything, it doesn't work.\n\n\n\n", "Python Operators that add/remove ID data without an undo step crash Blender\nOperating system: Win 10\nGraphics card: GTX 1070\n\nBroken: 2.8.3, 2.9 aed11c673efe\nWorked: 2.8.0\n\nUndoing after object linking to the scene crashes blender. I used the latest daily build\n\n1. Load the crash blend file [crash_modal.blend](crash_modal.blend)\n2. Run the script attached\n3. Click in viewport to add a new object, and then press ESC\n4. Press Ctrl+Z to undo\n5. Crash..\n\n", "VSE thumbnails improvements\n- [x] Check if reusing strip hash used by rendering job speeds up initialization with large number of strips. if it doesn't, fix it anyway. (b534806ecb) Note: it wasn't initialization that caused freezing, it was rather strip `anim` freeing, reusing ghash seems to have no impact on performance.\n- [x] To speed up loading, Check strategy to render \"guaranteed\" set of thumbnails. Especially if cache can be shared this could help a lot. (3a898db3638d) Note: better strategy seems to be other way around as far as performance goes.\n- [x] Make first (and last) thumbnail reflect actual handle position. (29b9187b327c) Note: Only first image is updated, though it's possilbe to update last one too. \n\n- Change cache so thumbnail (and possibly raw) types are keyed by path to file and disregard various context settings and Sequence pointer. Front-end should stay same. ([D12771](D12771))\n - There are rabbitholes like effect strips producing raw images and scene strips rendering at \"host\" scene resolution so simple string key is probably not enough. \n - Added complexity is, that all cache types per frame should be invalidated at once, otherwise prefetch doesn't work and issues like that.\n - Above points don't quite apply to thumbnails, but there are issues of this type in general. For example if we would implement downscaling for movies, first point will apply.\n\n", "Lattice Object Data Properties fields do not update on Undo in Edit mode\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.4.0 Alpha\n\nIn Lattice Object Data Properties, when a field is changed to a different value and then the User performs an undo, the undo can be seen to take effect in the 3d viewer but the fields do not revert. This bug happens in Edit mode. It seems to work as expected in Object mode, and the changes do revert when switching back to Object mode.\n\n1. Start Blender. The General option is good.\n2. Add a Lattice.\n3. switch to edit mode.\n4. In object data properties, change a field.\n5. perform an undo.\n6. note that the changes took effect in the 3d viewer, but the fields did not revert in Lattice Object Data Properties.\n7. Switch to object mode.\n8. note that the fields do revert now.\n\n\n\n", "EEVEE viewport rendering loop bug when using walk or fly navigation\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GT 1030/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 3.0.0 Alpha\nWorked: As far as I know, this bug was always in Blender\n\nEEVEE viewport rendering loop bug when using walk or fly navigation:\nwhile using fly or walk navigation and staying still, the sampling always just jumps back to 0 when finished sampling and just starts sampling over and over again thus creating an irritating loop of constant resampling\n\n1. Use EEVEE\n2. For the purpose of this example, change the scene filter size to 50px so that the effect is easily noticeable \n3. Press shift + ` to go to walk navigation (fly navigation has this bug too)\n4. Stay still so that the viewport could start sampling to the specified amount\n\nNotice when the sample amount reaches the specified viewport samples amount, rendering starts again as in if you have moved\n\n[EEVEE viewport navigation loop bug.blend](EEVEE_viewport_navigation_loop_bug.blend)\n\n\n\n\n\n", "Low unit scale values show wrong Stepsize in sliders\nOperating system: Linux-5.8.0-2-amd64-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.91.0 Alpha\nWorked: not in 2.80 or 2.83\n\nWhen the unit scale is too low (<= 0.002 for imperial, <= 0.00009 for metric) the stepsize is too big and sliding bound sliders (like thickness in solidify) will bounce them between their limits. At almost the same time seperate units stop showing the correct unit and instead switches to the default unit. But that seems to don't quite align as I was able to get the unit issue without the stepsize issue (might related to #81663).\n\nThe stepsize issue does not appear when Continuous Grab is disabled in the User Preferences.\n\nBased on default startup.\n1. In Scene Settings: enable seperate units\n2. set unit system to imperial with a unitscale of 0.001\n3. add a solidify modifier and slide the thickness value\n4. set unit system to metric with a unitscale of 0.00009\n5. slide the thickness value of the solidify modifier again (same behavior)\n6. disable Continuous Grab in the User Preferences\n7. slide the thickness value of the solidify modifier again (normal behavior)\n\n", "`action.frame_range` span always >=1 when there's only 1 key frame in it.\nOperating system: Mint\nGraphics card: Nvidia\n\nBroken: 3.5 & main branch\n\n\nframe_range wrong value when only 1 keyframe\n\n- with default cube, go to frame 6\n- Insert a keyframe (for example, on location)\n- bpy.context.object.animation_data.action.frame_range returns (6.0, 7.0)\n\nMy expectation was it returns (6.0, 6.0)\n\nIf you then insert a keyframe at frame 7, it still returns (6.0, 7.0)\n\n" ]

[ "Volume Object Data Render Step Size cannot be reset back to 0 once changed.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 470/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 388.13\n\nBroken: version: 2.83 (sub 9)\nWorked: Never, it was added in 7537cad5761e\n\nIf the Step Size setting in the Render sub panel of the Volume Object Data tab is changed from it's default of `0`, it cannot be changed back to `0`\n\n\n\n\n - Start with the default scene. \n - Switch the render engine to Cycles. \n - Add a Volume object (even an empty one will do.) \n - Change the Step Size setting in the Render sub panel of the Volume Object Data tab to anything other than the default of `0`.\n # Now try to change it back to `0`\n\nIn the file below I have already done the first 4 steps for you.\n\n[step_size_bug.blend](step_size_bug.blend)\n\nIn rna_gen_volume.c, I believe it is line 1070 where the minimum value is set to `0.0010000000f` or `0.0000100000f` based upon my experimenting.\n\n```FloatPropertyRNA rna_VolumeRender_step_size = {\n\t{(PropertyRNA *)&rna_VolumeRender_clipping, (PropertyRNA *)&rna_VolumeRender_space,\n\t- 1, \"step_size\", 1, 0, 0, 4, 0, \"Step Size\",\n\t\"Distance between volume samples. Higher values render more detail at the cost of performance. If set to zero, the step size is automatically determined based on voxel size\",\n\t0, \"*\",\n\tPROP_FLOAT, PROP_DISTANCE | PROP_UNIT_LENGTH, NULL, 0, {0, 0, 0}, 0,\n\trna_Volume_update_display, 0, NULL, NULL, rna_property_override_diff_default, rna_property_override_store_default, rna_property_override_apply_default,\n\toffsetof(VolumeRender, step_size), 5, NULL},\n\tVolumeRender_step_size_get, VolumeRender_step_size_set, NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL, 0.0010000000f, 100.0f, 0.0000100000f, FLT_MAX, 1.0f, 3, 0.0f, NULL\n};\n```" ]

[ "Optimized per-datablock global undo\nGlobal undo is slow on scenes with many objects. A big reason is that undo writes and reads all datablocks in the file (except from linked libraries). If we can only handle the ones that actually changed, performance would be much improved in common scenarios.\n\nThis proposal outlines how to achieve this in 3 incremental steps. Only 1 or 2 of these steps would already provide a significant speedup.\n\nAt the time of writing there is no concrete plan to implement this, but I've had this idea for a while and it seemed useful to write down in case someone likes to pick it up.\n\n## ~~1. Read only changed datablocks~~ DONE\n\nLinked library datablocks are already preserved on undo pop. We can do the same thing for all non-changed datablocks, and only read the changed ones.\n\nWe can detect which datablocks were changed by checking that their diff is empty. Addition and removal of datablocks also needs to be detected. We need to be careful to exclude runtime data like `LIB_TAG_DOIT` from writing, to avoid detecting too many datablocks as changed. Most changes to runtime data go along with an actual change to the datablock though.\n\nSince unchanged datablocks may point do changed datablocks, we need to reload datablocks in the exact same memory location. We can safely assume the required memory size stays the same, because if it didn't those other datablocks' pointers would be changed as well. In principle other datablocks should not point to anything but the datablock itself, but there may be exceptions that need to be taken care of *(e.g. pointers to [armature] bones hold by [object] pose data...)*.\n\n`Main` fields other than datablock lists may need some attention too.\n\n## ~~2. Optimize post undo/redo updates~~ DONE\n\nThe most expensive update is likely the dependency graph rebuild. We can detect if any depsgraph rebuild was performed between undo pushes, and if so mark that undo step as needing a depsgraph rebuild. This means the undo step is no more expensive than the associated operation.\n\nThis requires the dependency graph to be preserved through undo in scene datablocks, similar to how images preserve image buffers. This would then also preserve GPU data stored in the dependency graph. If the dependency graph supports partial rebuilds in the future, that could be taken advantage of.\n\nThere may be other expensive updates, to be revealed by profiling.\n\n## 3. Rethink how we handle cache preservation across undo steps.\n\n#76989 seems to point to some pointer collision issue, similar to the reason that lead us to add session uuid to IDs when we started re-using existing pointers in undo.\n\nNote that according to the report, there were already issues before undo refactor - hard to tell though if they had the same root cause?\n\nThis needs to be investigated further.\n\n## 4. Write only changed datablocks\n\nWhile undo/redo is usually the most expensive, the undo push that happens after each operation also slows down as scene complexity increases. If we know which datablocks changed, we can only save & diff those.\n\nThis could be deduced from dependency graph update tags. However there are likely still many operations where tags are missing. If it is not done correctly, then instead of a missing refresh there is a more serious bug of not undoing all changes. So this needs verification throughout the Blender code for correct tagging, before it can be enabled.\n\nDebug builds could verify correctness by testing the diff is empty whenever there was no tag.\n\n\n-----------------------------\n\n\nCurrent Status and Plan\n****\n\nThere is an experimental option in 2.83 to address step 1 and (part of) step 2. This was developed in branches and is now in master. Implementation of step 3 has not started.\n\n- [x] [id-ensure-unique-memory-address]() implements the idea of never re-using a same memory address twice for IDs. This is required to allow usage of ID pointers as uids of datablocks across undo history.\n - [x] Add storage of addresses history for each ID (has to be stored in blend memfile, can be ignored/reset when doing an actual blendfile reading).\n - [x] Discuss how to best implement the 'never get same allocated address again' process, current code here is basic, we could probably be much smarter and efficient with some changes to our MEM allocator itself.\n - [x] We are going to try and use session-wise uuids for data-blocks instead.\n - [x] Make new undo behavior controllable from Experimental preferences area.\n- [x] [undo-experiments]() Implements detection of unchanged datablocks at undo, and re-using them instead of re-reading them.\n - [x] Switch to using id addresses instead of names to find back a given data-block across undo history (depends on completion of work in `id-ensure-unique-memory-address` branch).\n- [x] [undo-experiments-swap-reread-datablocks]() implements re-using old ID address even for changed IDs that need actual reading from memfile. Also adds support for more refined depsgraph update flags handling from undo. Both changes combined allow to reduce dramatically depsgraph work needed during most undo steps.\n - [x] Switch to using id addresses instead of names to find back a given data-block across undo history (depends on completion of work in `id-ensure-unique-memory-address` branch).\n- [x] Investigate crashes due to invalid or used-after-free memory in depsgraph after undo step.\n- [x] Investigate why some undo steps are still using full reload of the memfile. Some may be necessary, but think we can improve here.\n - [x] In particular, the absolute last undo step (the first to be undone) is always seen as 'all data are changed', for some reasons...\n- [x] Investigate how to discard runtime data from 'data is changed' detection (probably by nullifying those pointers at write time?).\n- [x] Fix slow undo of adding/duplicating/deleting datalocks. The undo diffing does not appear to be smart to match up the right datablocks in such cases, causing many unrelated datablocks to be marked as changed, which also means excessive memory usage. The `session_uuid` could be used to find matching datablocks.", "Undo in Armature Edit Mode does not undo Custom Properties\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.5 Beta\n\nUndo in Armature Edit Mode does not undo changes to Armature Custom Properties\n\n- [optionally] In default scene delete all objects\n- Add an \"Armature\" object\n- Go to `Edit Mode`\n- In `Object Data Properties` add a `Custom Property`\n- Change the value of the property (to for example `0.0`) in order to add an Undo step\n- Try undo (`Ctrl+Z`)\n\nThe step is not undone in edit mode.\nInterestingly, changes to the Custom Properties of Bones are correctly undone in Edit mode.\nThis makes it seem that changes to Armature Custom Properties were perhaps missed in the Armature Edit Mode Undo Context.\n\n---\n**Original Report**\n[Video](3RZNWd5tRQI)\nNote when changing the **head size** in the redo panel, you can observe how the number of elements in in the Collection Property `bones_for_delete` grows.\n\n1. Execute the script generate_rig.py from the file, if you do not have automatic script execution enabled\n2. After execution, the Generate Rig panel will appear in the Rig category\n3. Select Empty.021 first, then Armature.001. Switch to armature editing mode, and click Generate Rig on the panel\n4. The bones will appear, and also the text on the panel will notify that there are elements in bones_for_delete. After that, you can scroll the **head size** on the redo panel, or simply undo the operator. the label in the panel will notify you that the number of items will not change\n\n", "Toggle to Edit Mode call error when the Blender use linked a object data from another file\nOperating system: Linux 5.10.2\nGraphics card: Intel Haswel (4-th Gen ) HD Graphics 4600\n\nBroken: 2.92\n\n\nToggle to Edit Mode call error when the Blender use linked a object data from another file\n\n\n1) Create any object in a scene and save this file \n2) Create a new blend file and link the object from previous saved a file with `File` → `Link..` and then select this object.\n3) Then convert linked object to local object with `Object` → `Relations` → `Make Local...` → `Selected Objects`\n4) Select the object and press Tab that go to Edit Mode. \n5) Error - Unable to execute 'Toggle Editmode', error changing modes\n\nor download both files and then open linked.blend:\n\n[source.blend](source.blend)\n[linked.blend](linked.blend)\n\n", "Undo Makes Completed Renders Disappear From View Render Window Slots\nOperating system: Windows 10 Build 19043\nGraphics card: Nvidia RTX 3090\n\nBroken: 2.92\nWorked: Can't remember, but I don't recall this happening before\n\nUndo makes completed renders disappear from other slots in the View Render window. Some of them are preserved, but the more you undo, the more disappear from other slots.\n\n1. Load default scene (with the cube)\n2. Render scene\n3. Go to the cube's shader and change its base color\n4. Switch to next unoccupied slot in the View Render window\n5. Render scene\n6. (optional) Repeat steps 3-5 as many times as you like\n7. Go to View Render window and scroll through slots to verify all renders are there (with J key)\n8. Activate undo (ctrl+z)\n9. Go to View Render window and scroll through slots. Some will be missing.\n10. Activate undo again and more will be missing.\n\nThis doesn't seem like intended behavior, and I've replicated this issue on a different machine. Hope this helps!\n\nEdit: I've replicated this issue in 2.91, 2.90 and the latest LTS release. I assume intended behavior is for undo to change the shaders back, but for renders you've completed to remain in their slots.", "Pie Menus re-call when using operator bound key that is the same as the Pie Menu Call\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.90.0\n\nName: 3D Viewport Pie Menus (1, 2, 9)\nAuthor: meta-androcto\n\nI have encountered some maybe unexpected behaviors with the Pie Menu system.\n\nSome context, pie menus create automatic key bind's for all operators within the pie (represented by underscores). In some cases the key bind for that operator can be the same as the call bind for the pie menu. While this doesn't present the following issue if you wait ~1-2 seconds after calling the menu, it does re-call the menu if you tap the corresponding key bind twice quickly.\n\nFor my example I used the Blender included '3D Viewport Pie Menus` add-on, but this issue is **not** restricted to add-ons. Realistically I feel it will only happen with add-ons, but it is also possible on built-in pie menus assuming you butcher your keymap's intentionally to break it:\n\n1. Open a new scene\n2. Turn on '3D Viewport Pie Menus` within blenders add-on tab\n3. Press 'A' in the 3d viewport quickly twice, quickly is important. It needs to be around before the Pie animation is finished\n\nWhat happens is the 'Select All Toggle' operator is ran, as it should be, but the pie menu is also called a second time. This can happen with any pie assuming there is a key bind within the pie that is the same as the menus call bind.\n", "Undo eliminates brush texture\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: version: 3.0.0 Alpha\n\nI think it shouldn`t be undoable as other brush settings now.\nIt is much easier to delete texture manually, than restore it and tune it again.\n\n[2021-05-26_14-00-22.mp4](2021-05-26_14-00-22.mp4)\n\n\n", "Uncheck Save Copy freezes on next render - Mutex is still locked\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.91.0 Alpha\n2.90.1 and 2.83.5 have the same behavior\n\nWith default scene :\nWhenever I try to save a render, I uncheck both save as render and save copy, the next time I hit render, blender will freeze.\n- Unticking save as render only does not seem to create the freezes, however the option is not saved, I have to manually uncheck it every time.\n\nNew file : \n- Save file\n- Hit render\n- Save as\n- Uncheck Save as Render and Save Copy\n- Save.\n- Hit Render again = freeze\n\n\n", "Mirror Modifier On Cage Selection Papercut \nBroken: 2.93.4\nWorked: Maybe never.\n\nWhen in edit mode, using the \"On Cage\" option for the mirror modifier results in some unintuitive behavior when selecting vertices created by the bisect feature. The vertices created by the by the bisection are mapped to the original mesh data in a way that appears to be based on blenders internal representation of the mesh and not based on something observable to the user.\nThe behavior when selecting should feel consistent to the user. To fix this, I suggest that the vertices created by the bisection should always map to the vertices on the original mesh that are on the removed side of the bisection plane. This way users would be able to tweak the centerline of the mirrored mesh in while using \"On Cage.\" As it is currently implemented, there are vertices that affect the mesh that cannot be accessed through normal click selection unless \"On Cage\" is off.\n\n**Example**\n\n\nSelect the +X or -X face of the default cube and scale it by 0.5.\nAdd the mirror modifier and select bisect for the X axis so that the object is mirrored and bisected about the X axis.\nTurn on \"On cage\" for the mirror modifier then enter edit mode for the object.\nWhen selecting the vertices on plane X=0, you should notice that two of the vertices correspond to visible mirrored vertices and two correspond to vertices removed by the modifier. \n\nSimple scene showing the behavior. \n[MirrorSelectOnCagePapercut.blend](MirrorSelectOnCagePapercut.blend)\n", "Operator macros [e.g. ARMATURE_OT_duplicate_move] have non-functional 'Proportional editing' option in Adjust Last Operation panel\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.90.0 Alpha\n\nDuplicating a bone in edit mode. In the Adjust Last Operation panel there is a proportional edit checkbox. Which can't be activated.\n\n\nOpen Blender\nCreate armature\nSwitch to Edit Mode\nDuplicate the bone\nIn Adjust Last Operation panel try to activate Proportional editing.\n\n\n\n", "Texture paint undo sometimes changes texture slot\nOperating system: Windows 10\nGraphics card: GTX 1060\n\nBroken: 2.80 (sub 48), 2019-03-12 13:53, hash: 936014ea8c74\n\nWhen texture painting on a particular texture slot, then undoing that painting, sometimes the selected paint slot will change.\n\nI couldn't reproduce from the default startup file, so I've stripped down the file that's having problems as much as I can.\n\n1. Open attached file\n2. Select paint slot \"Skin mask\" from either topbar or properties panel.\n3. Paint some strokes on the model or image editor\n4. Press Ctrl-Z\n5. Notice the paint slot is no longer set to \"Skin mask\"\n\n- - -\n\n- Change to the Skin Mask slot.\n- Paint a single stroke in the image editor.\n- Undo two times.\n- Paint another single stroke in the image editor.\n- Undo one time.\n- Notice that the slot changed.\n[texturepaintbug.blend](texturepaintbug.blend)", "Undo when sculpting changes brush type\nOperating system: Windows 10 Pro - 20H2 - 19042.867\nGraphics card: RTX 2070 SUPER - Driver version 497.09\n\nBroken: 3.0.0, f1cca3055776, master, 2021-12-02\n\nWhen sculpting, pressing undo will change the brush type and the brush settings to the previous selection.\n\nThe issue has existed for me for several years across multiple PCs, and countless versions of Blender. It has been reported here by many people but I've never seen anyone mention the following which I think could be helpful. I always work with a second window on another screen with just a UV Editor open which I think has something to do with the issue.\n\nFor example, in a new scene, start with the Draw brush active and change to the Grab brush then make three brushstrokes on the corner of the cube with it to pull it out. Pressing undo once will only remove the last Grab brushstroke as expected but also revert the active brush to the Draw brush which is not expected.\n\nAfter loading factory settings, the problem doesn't exist. What causes it to happen is if you add an additional New Main Window and set it to the UV Editing workspace and combine the UV Editor area and the 3D Viewport area to make just one large UV Editor area. Repeating the previous steps mentioned in the Sculpting workspace then causes the issue to happen.\n\nThere is some inconsistency with what happens afterwards with regards to changing the windows, workspaces and areas around. It seems various combinations can cause this to happen but the following steps should consistently replicate the issue for anyone on factory settings.\n\n - File > Defaults > Load Factory Settings\n - Window > New Main Window\n\n - On the New Main Window, select the UV Editing workspace\n - Combine the two main areas so only the UV Editor area remains, covering most of the workspace, and the 3D Viewport area is removed\n\n - On the original Main Window, select the Sculpting workspace\n - Select the Grab brush\n - Make three brushstrokes\n - Press undo once\n - Observe as one brushstroke is removed as expected but the brush type is also changed back to the default Draw brush\n\nAdditional observation:\n - Close the New Main Window that has the UV Editor open\n\n - On the original Main Window, select the Sculpting workspace\n - Select the Grab brush\n - Make three brushstrokes\n - Press undo once\n - Observe as one brushstroke is removed as expected and the brush type does not change", "Text editor undo system, undo step limiting behavior\nCurrently having a linear undo stack has a limitation which is most obvious with text undo.\n\nTyping a character uses a single undo step, meaning if you are working in 3D sculpting for eg, then type a short paragraph over 32 characters - you'll loose all your undo steps.\n\nProposed solution is for undo systems to support counting their undo steps separately, so text editing operations wouldn't use up undo steps for edits in other modes.\n", "Vertex paint mode breaks permanently under certain conditions (Global undo involved)\nOperating system: Linux-6.1.18-200.fc37.x86_64-x86_64-with-glibc2.36 64 Bits, WAYLAND UI\nGraphics card: AMD Radeon Graphics (renoir, LLVM 15.0.7, DRM 3.49, 6.1.18-200.fc37.x86_64) AMD 4.6 (Core Profile) Mesa 22.3.7\n\nBroken: version: 3.6.0 Alpha\n\nCaused by 575ade22d4de (vertex paint now does `SCULPT_undo_push_begin_ex` in wpaint_invoke [which leads to `BKE_undosys_step_push` using sculpt undo])\n\nVertex paint mode breaks permanently under certain conditions\n\n- turn OFF Global undo in Preferences\n- enter vertex paint mode\n- make two strokes\n- Undo (prior to 575ade22d4de nothing was undone, now stroke is actually undone)\n- Undo again (this now takes you out of vertexpaint mode into object mode prior to 575ade22d4de you would never leave vertexpaint mode)\n- enter vertex paint mode\n- painting is not possible anymore (strokes have no effect)\n- this is also true for other objects\n\n", "Unable to undo subdivision with Multiresolution modifier\nOperating system: Linux-5.4.0-91-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.86\n\nBroken: version: 2.93.6\n\n**Bug**: If i subdivide with Multiresolution modifier after a sculpt action it doesn't undo subdivision but it needs to undo the sculpt action to undo that subdivision done with Multisolution modifier.\n**Note**: If no sculpt action is done before subdivision with Multiresolution modifier, the undo for subdivision just works (No need to undo the action done just before the subdivision to get the result back).\n**Mode** : Sculpt\n\n**Steps**:\n1. File> New> Sculpting. \n2. Add Multiresolution modifier.\n3. Sculpt with Grab brush.\n4. Subdivide once at Multiresolution modifier.\n5. Undo (Doesn't work).\n\n**File**:[Multiresolution_subdivision_bug.blend](Multiresolution_subdivision_bug.blend)\n\n", "Clicking on a pie menu item doesn't work when holding the pie menu key.\nOperating system: Windows 10\n\nBroken: 2.79, 2.92\n\nClicking on a pie menu item doesn't work when holding the pie menu key. This is a problem because sometimes the pie menu key release isn't detected by blender and then the it locks the UI.\n\n1. On the 3D Viewport hold Z for the shading pie menu.\n2. While still holding Z try clicking on the pie menu items. Doesn't work.\n3. While still holding Z, left click on Windows's Taskbar then release z.\nAfter that, clicking on the pie menu items doesn't work even though the user isn't holding the pie menu key anymore, blender still considers that the key is held. Then Blender gets locked to the user unless he knows to press esc or right click. The solution is making the pie menu clickable even when the pie menu key is held." ]

[ "Crash when undo a loop cut\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 750/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.60\n\nBroken: version: 2.80 (sub 75)\n\nI was working on Blender and it crashed when I undoed a Loop Cut. I've tried to recreate the bug on a default file but I coudn't, so I deleated everything on the scene and left just a square.\n\n[loop.blend](loop.blend)\n\nIt happens randomly, just do a loop cut and undo it. Some times it happens on the first try, but I also get it with 30 tries.\n\nDebug output:\n[blender_debug_output.txt](blender_debug_output.txt)\n\nI don't know if it's related to [#67454 ](T67454), but it just happens with Loop Cut, I tried the same with Extrude and Bevel and it didn't crash.\n" ]

[ "Pie menu breaks if spawned near window edge\nPie menus are sort of \"clamped\" in that if they are spawned closer to the window edge than their spread radius (UserPreferencesView.pie_menu_radius), instead of spawning partly offscreen, instead they will spawn stuck to the edge and will automatically validate the menu item that's in that direction. To make it clearer, here is a capture.\n\n\n\nThis is a problem because pies can hardly be spawned in cramped workspaces - for instance inside a graph editor occupying the lower third of a standard 16:9 monitor, unless the radius is set super small.\nThis happens on Windows.", "Regression: Sculpt mode Annotation tool cursor conflict\nOperating system: Win 10\n\nBroken: 2.91, 3.4\nWorked: 2.90\nCaused by b9e4696641\n\nThe Annotation tool in sculpt mode is not using it's proper cursor. It's using the regular sculpting cursor instead.\nWhen you start drawing, the original cursor shows up on top of the regular sculpting cursor. \nThe end result is two cursors displaying at the same time and the original one flickering sometimes...\n\nTry to use the Annotation tool in sculpt mode....\n\nHere 2.90 (working fine)\n\n\nHere 2.91 (broken)\n\n", "Ctrl+Click for Nurbs Surfaces\nWith Meshes and Curves, you can select a point, press Ctrl+LMB, and the point is extruded to the mouse.\n\nThis however does not work for Nurbs, even though you can extrude a point, Ctrl+LMB does not work.\n\nTry:\n- Add -> surface -> nurbs curve\n- Enter editmode\n- Select an end-point\n- Ctrl+Click... *nothing happens*\n\nThe same works for a curve or a mesh.\n\nThis could be split into 2 tasks\n- Get Ctrl+Click working for single nurbs curves.\n- Get Ctrl+Click working for nurbs surfaces (using the extrude code), This should extrude and place the newly extruded vertices at the mouse position, as is done with mesh data.", "Expand face sets freeze blender when cursor is on mesh boundaries\nOperating system: Linux (Mint 21) \nGraphics card: RTX 3070 Ti\n\nBroken: 4.0.0 alpha, 12f51bd1c77b, master, 2023-09-21, as found on the splash screen)\nWorked: 3.6\n\nTry to create new face sets from first or last face loops on a cylinder.\nBut expand Face Set from boundaries is frozen in 4.0 alpha.\n\n\n- Open the blend file\n\n- Switch to Sculpt mode\n\n- Move your cursor outside boundaries and press Expand Face Set (Shift +W). (See on pictures)\n\nAt this point, Blender freeze, not crash. No more info in console.\n\nBug :\n\n\nExpected (3.6 version): \n\n\nThanks for your help.\n\n", "Keyframing Corner pinning in compositing is inconsistent\nOperating system: windows 10\nGraphics card: Nvidia geforce GTX 1050 Ti\n\nBroken: 3.3.0\n\n\nWhen using the background X symbols to drag and drop corners with corner pinning, only the topmost value of each vector is keyframed\nExample image(wanted to edit a meme):\n\n\n- Open attached file\n- Single Render to Update Compositor\n- Move the Corner Pins\nor\n- Arrange nodes as seen.\n```\n{F13601679}\n```\n- Select the corner pin node\n- Try to keyframe the location of the second image/video, using auto keyframes and the background viewport manipulators.\nOnly the topmost (X) value gets keyframed\n\n[Corner pin.blend](Corner_pin.blend)\n", "Broken/weird Continuous Grab behaviour\nOperating system: Linux-5.8.0-43-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: SVGA3D; build: RELEASE; LLVM; VMware, Inc. 3.3 (Core Profile) Mesa 20.2.6\n\nBroken: version: 2.93.0 Alpha\n\nbroken/weird behaviour under VMWare Player/Workstation 14.1.2+, Ubuntu 20.04.2 LTS\nEdit -> Preferences -> Input -> Continuous Grab\n|[2021-02-27 03-11-16.mp4](2021-02-27_03-11-16.mp4)|[cursorbug.mp4](https://archive.blender.org/developer/F13560997/cursorbug.mp4)\n| -- | -- |\n\n- activate grab mode (G-key) in the viewport or in any blender's area which support 'Continuous Grab'\n- drag mouse toward area top or bottom edge\n\nAlternatively:\n- Active walk navigation and quickly move the mouse outside of the Blender window. \n", "Vertex slide (GG) acts weird with zoom (steppy, jumping, not slowing down with shift)\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.19\n\nBroken: version: 2.82 (sub 7)\n\n\n[2020-02-22_06-30-18.mp4](2020-02-22_06-30-18.mp4)\n\n[gg.blend](gg.blend)\n\n", "Tooltip gets triggered when leaving modal dialog\nOperating system: macOS-11.4-arm64-arm-64bit 64 Bits\nGraphics card: Apple M1 Apple 4.1 Metal - 71.6.4\n\nBroken: version: 2.93.0 Beta\nWorked: -\n\nWhen moving the mouse cursor outside of a modal File Dialog, sometimes a tooltip is triggered when it shouldn't. Probably an edge detecting bug.\n\nI've included a GIF recording with this bug report that hopefully explains the behavior I noticed. Here are some steps to reproduce:\n1. Start Blender and select `File -> Open` to display the File Dialog.\n2. Ensure the \"Recent\" section on the left of the dialog has multiple entries.\n3. Resize the File Dialog so that the \"Recent\" section is truncated by the bottom border of the File Dialog. It should be truncated so that it displays only the top few recent folders (i.e. the other folders, the little disclosure triangle and the section's bottom drag bar shouldn't be visible).\n4. Hover the cursor over the File Dialog and move it outside of the dialog, making sure you exit the dialog through the bottom part of the \"Recent\" section.\n5. Watch as the \"Active Recent Folder\" tooltip gets triggered/displayed while the mouse cursor is outside of the dialog. This shouldn't happen!\n6. Now enlarge the bottom part of the File Dialog, so that the \"Recent\" section is no longer truncated and fully displayed (including its disclosure triangle and drag bar).\n7. Repeat step 4 and notice the tooltip doesn't get triggered anymore! This is the desired behavior.\n8. Resize the bottom part of the File Dialog so that the \"Recent\" section is again truncated, but this time don't truncate its bottom part (i.e. the disclosure triangle should be visible).\n9. Repeat step 4 and notice the tooltip again doesn't get triggered anymore! This is again the desired behavior.\n\n\n\n**Conclusion**\nIt looks like the the File dialog somehow doesn't properly detect the mouse has exited the dialog when it exits directly from a section's contents, thereby not crossing over the section's border/other UI parts.\n\nI hope this makes some sense ;-)\n", "Error with paint curves and drawing with mouse buttons\n[system-info.txt](system-info.txt)\n\nBroken: blender-2.80.0-git.cc600de6695a-windows64 (and probably above)\n\nWhen using paint curves, they don't draw upon clicking the right mouse button if specified as the button to select with in the preferences. But they do if you set the mouse button to select with to be the left one. So if you have selected the LMB in the preferences then you can draw the paint curve (I don't mean creating the curve) with the LMB, but if you have chosen it to be the RMB then it doesn't work when pressing the RMB, neither does the LMB. (Also, the UI button for drawing the curve in the stroke panel does not work. But you can draw it when you type draw curve in the search panel)\n \nGo to the *Sculpting* screen layout and create a paint curve on an object (if it helps, I used the the normal *Draw* brush). Open the *Stroke* panel in the *Properties Editor* and set the *Stroke Method* to *Curve*. If you have chosen the left mouse button in the *Keymap* section in the *Blender Preferences* under *Select With:* , then everything works fine. You can draw the paint curve on the object by doing a left click after creating it by doing multiple Ctrl-RMB clicks. But if you now set the mouse button with which you want to select to *Right*, then you can create the paint curve with Ctrl-LMB clicks, but when you try to sculpt/draw the paint curve with either the left mouse button or the right mouse button, none of them work. (And again, if you try to draw the paint curve with the *Draw Curve* UI button in the *Stroke* panel, it will also do nothing but turn blue because you clicked on it...)", "Material Preview (EEVEE) can cause access violations when switching to edit mode\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.6.3\nWorked: N/A\n\nSwitching to edit mode with material preview (EEVEE) enabled results in a crash \n(Log from my machine produced with the joined blender file also joined to this report)\n\nusing the file joined in this report:\n- Open the scene\n- Select \"Circle.039\"\n- Switch to material preview\n- Attempt to enter edit mode on \"Circle.039\"\n\n", "The Tool Selector Window close one time every two times, when opened too closed to the bottom border of the screen\nOperating system: Windows 10\nGraphics card: GTX 1080\n\nBroken: 2.80\nWorked: (optional)\n\nThe Tool Window, open and close immediately, choosing the tool under the mouse cursor, if the tool window is opened too close from the bottom of the screen.\n\n1) In Edit > Preferences > Keymap, set the \"Space Bar action\" to \"Tools\"\n2) On the 3D viewport, place your mouse cursor in the top area of the viewport\n3) press the spacebar, the tool menu open without any problem, and you can choose a tool with a shortcut key (for example press B to choose the Box selection)\n4) You can repeat that several times in a row without any problems.\n5) Now move your mouse cursor much more closer to the bottom border of the screen, so that you know that the tool window will have to be shifted upward to be displayed correctly.\n6) Press the spacebar a first time. This is ok, you can select your tool\n7) Now press the spacebar a second time. This time, the tool window open and intermediately closed, and the tool selected is the one that is under your mouse cursor.\n8) Press the spacebar again: the tool window open normally\n9) Press the spacebar again: the bug reappear\n10) Now on, every one time over two, the bug appear.\n\n**Expected behavior**\nThe Tool Window should open and stay opened consistently, no matter where the mouse cursor is placed on the screen, when the spacebar is pressed.", "GRAB_CURSOR bl_option stops working as soon as modal operator returns PASS_THROUGH the first time\nOperating system: Linux-4.13.10-041310-generic-x86_64-with-debian-stretch-sid 64 Bits\nGraphics card: GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.67\n\n\nBroken: version: 2.80 (sub 57)\n\n\nUsing the `GRAB_CURSOR` bl_option in a modal operator, works fine until the operator returns `PASS_THROUGH` the first time - for instance, to adjust the viewport using `MIDDLEMOUSE`, while the modal is active.\nAfter the first `PASS_THROUGH` the cursor is no longer warped to the opposite side.\n\n\n* open attached blend file\n* run the script\n* in the 3d view, look for 'GrabCursorBug' operator in the search menu and run it\n* move the mouse horizontally across the region borders, notice how it warps to the opposite side accordingly\n* hold down `MIDDLEMOUSE` and move the mouse to adjust the viewport, release again\n* try moving the mouse across the region borders again, it won't warp to the other side anymore\n\n[video demo ](watch?v=liPmg5Z_hCk)\n[grab_cursor.blend](grab_cursor.blend)\n\n\n", "Nodes: Socket and title click detection is broken for multiple monitors\nOperating system: Windows 11 Home, Version 22H2, OS Build 22621.1265, 64-bit\nGraphics card: NVIDIA GeForce RTX 3060 Ti, NVIDIA GeForce RTX 2070 SUPER\n\nBroken: Blender 3.4.1, Blender 3.5.0 Beta (2cd7e70c18a8), Blender 3.6.0 Alpha (ef60b13c1f29)\n\n\nThe click detection for node sockets and titles does not work when using main windows in two separate monitors.\n\nThis issue is similar to 96255 , but I'm facing this issue even with UI scale = 1.0.\n\n\n1. Create a geometry node tree. Node socket/title click detection works properly.\n2. Create a new Main Window. Click detection still works.\n3. Move a main window to another screen. Click detection fails. Clicking and dragging sockets/titles triggers box selection. \n\n(See attached video)\n\nMain monitor: Dell U2720Q\nSecondary monitor: LG DualUp 28MQ780\n\n", "Box Select (B) changed to Left Mouse unable to Deselect\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.18\n\nBroken: 3.0, 3.5\nWorked: 2.93.9 (31712ce77a6a), 3.0.0 broke it\n\nBox Select (standard B) changed to Left Mouse ignores Gesture Box Deselect. Instead Deselect now selects Objects.\n\n(See attached Image. These preferences work different until 3.0.0. The Image and Full Version below is basically an alternative to Blenders standard \"Left Mouse Drag Box Select\" which removes the need to hold the mouse button)\n\nPreferences > 3D View > Box Select:\n\n-SHORT VERSION-\n1. Set Left Mouse as Event (Press)\n2. Go to Gesture Box\n3. Change \"Begin - Left Mouse\" to \"Release\"\n4. Use a Left Mouse click + release in 3D View to begin the Box.\n\nProblem: Middle Mouse (Deselect) only selects Objects.\n\n-FULL VERSION-\n1. Set Left Mouse as Event (Press)\n2. Set Mode to Extend\n3. Go to Gesture Box\n3.1 Change \"Begin - Left Mouse\" to \"Release\"\n3.2 Change \"Select - Left Mouse\" to \"Press\"\n3.3 Deactivate \"Begin - Middle Mouse\"\n3.4 Change \"DeSelect - Middle Mouse\" to \"Press\"\n4. Use a Left Mouse click + release in 3D View to begin the Box.\n\nExpected: Left Mouse will select, Middle Mouse will deselect.\nProblem: Middle Mouse (Deselect) only selects Objects.\n\n------\n\nI could not find a way to have it Deselect or if there is a conflict with other settings. Can you take a look at this? Thank you.\n\n", "Blender 3.5: Cursor warping usable on windows remote desktop (RDP)\nOperating system: Windows 10 21H1 19043.2130\nGraphics card: GTX1650 4GB\n\nBroken: 3.5.0 Alpha, hash: 508fd044b44a\nWorked: 3.4.1, hash: ef9ca44dee7f\n\nI'm using Blender through Windows Remote Desktop (RDP) for quite a long time.\nNow in the new alpha 3.5 the Blender UI acts weird and non controllable anymore when using mouse middle click to pan viewport or left clicking into number fields to drag values up or down.\n\nAt this point for me Version 3.5 alpha is not usable anymore through an RDP connection - which is quite important for me! - while at least in Blender Version 3.4.1 everything worked fine!\n\n- Open blender from a Windows computer on another Windows computer via remote session\n- use {key MMB} to pan viewport or {key LMB} into number fields to drag values up or down.\n\n\n" ]

[ "Conflect between Emulate3button and sliding extending\nSelect a point, double press \"G\" to enter slide mode. Now, if you press ALT, you get to extend sliding.\n\nThe bug is that if you activate the \"Emulate 3 button mouse\" option, you can't left click to confirm the action." ]

[ "2D transform operations show options and use a modal keymap intended for 3D transform\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.92.0 Alpha\n\ntransform.seq_slide and transform.transform are linked to the transform keymap of the 3D transform. This results in the wrong shortcuts hints #83307 in the status bar and also wrong options in the Edit Last Operation: \n{[F9508650](transform_last_operation.gif),size=full} \n\n- Use `Move/Extend from Current Frame` to move the handle.\n- Check the `Edit Last Operation` options\n\n", "can't use logitech mx master 3 mouse shortcut keys in blender 3.6 LTS\nOperating system: Win 11\nGraphics card: RTX 3070\n\nBroken: (3.6 LTS)\n\n**I can't use the logitech mx master 3 mouse shortcut keys in blender 3.6 LTS.**\n\n\n", "The Object Mode drop down menu in the 3D viewports is not reactive to larger font point settings.\nOperating system: Windows 10 1803\nGraphics card: nVidia RTX 2070\n\nBroken: v 2.80 as found on the splash screen\n\nThe Object Mode drop down menu in the 3D viewports does not properly resize itself if the user picks a larger font point size in the Themes section of the preferences. Instead of resizing the control truncates the overrunning text with a square character.\n\nStart Blender\nIn Preferences \\Themes \\ Text Style change the Widget Points setting from the default of 11 to a larger value. I used 15.\n\n", "The button \"Draw Curve\" for Stroke mode Curve probably doesn't work.\nOperating system: Windows 10\nGraphics card: AMD Ryzen 7, 2700, 8-core\n\nBlender Version: 3.0.1\n\nThe button \"Draw Curve\" from properties editor doesn't work.\n\nMy steps:\n1) Add any mesh object.\n2) Change Layout to Texture Paint (Default)\n3) Press + button below \"Mode: Material\". Add \"Base Color\". Press Ok.\n4) Scroll Down parameters to Stroke section. Expand Stroke section.\n4) Change Stroke Method to Curve.\n5) Press Ctrl+RMB button on mesh some times.\n6) Press \"Draw Curve\" button in properties editor - nothing happens!\nBut if I press Enter on my keyboard or Ctrl+LMB on mesh its drawing stroke.", "Inconsistent Grease Pencil auto-keying behavior\nOperating system: Linux 64-bit\nGraphics card: GTX 1080\n\n3.1 and prior\n\nWhen auto-keying is enabled and you're transforming a GP stroke in edit mode on a new frame, a keyframe is added at the beginning of the operation before it is confirmed. If you right-click to cancel the operation, the redundant keyframe remains. This is inconsistent with auto-keying in every other area, in which Blender waits for the user to confirm the operation before adding the keyframe. \n\nAlso, deleting a stroke fails to add a new keyframe. I apologize if these are two different bugs but they seem related.\n[GP_autokey_bug.blend](GP_autokey_bug.blend)\n - Open the attached file.\n - Grab (G) a stroke, then right-click to cancel moving it. The keyframe was inserted, even though no change was made.\n # Go to another frame, then delete a stroke. No keyframe was inserted this time, even though the GP drawing has definitely changed!\n\n", "There is a delay when dragging the value sliders while using a Wacom pen.\nOperating system: Windows 10\n\n\nBroken: Blender 3.0\n\n\n\nTrying to drag value input sliders always has an initial friction which makes value dragging not an instant interaction. I do not get this initial friction when I use a mouse instead. \n\nPlease see the movie below. As you can see the values start changing just a little bit after dragging starts with the pen. You cant see the mouse pointer in the recording but you can see that the value box activated then the value changing starts happening.\n\nThis issue happens with all of the value boxes, not just the transform inputs as in the movie.\n\n\n[pzJM3EGql3.mp4](pzJM3EGql3.mp4)\n\n\n\nMake sure you are using a Wacom pen with Windows Ink (although Wintab does not seem to make a diff)\n\nAdd a basic object\n\nOpen the transform panel\n\nTry to slide the values in the boxes.\n", "modifier offset slider step value being driven by the mouse looses one order of magnitude worth of accuracy when units are metric vs none\nwin 10\n2080ti\n\nBroken: 2.80+\n\nSince blender 2.8 has switched to real world units the modifier slider step(when driven by the mouse) value has lost one order of magnitude worth of accuracy, 0.001 instead of 0.0001 \nIt is possible to switch the unit system back to none to get the higher accuracy. \nThey should be consistent and both work to the higher level of accuracy as this level is essential when working on real work scale small items e.g. a watch\n\n\n- with scene unit type set to \"metric\" add a bevel modifier the standard cube\n- press shift while clicking and dragging the bevel modifier offset slider and notice how the smallest amount possible to move is 0.001 m\n- switch scene unit type to \"none\"\n- press shift and click drag the modifier offset slider and notice the smallest amount possible to move is 0.0001 m\n", "Pie Menus initiated by tweak events will disable the mouse button used in the tweak event. Works fine for regular menus, operators.\nWindows 10\ngeForce GTXTitan\n\nBroken: blender-2.77.0-git.94e18c5-AMD64\nWorked: (optional)\n\nI've set up some UI shortcuts that use 'tweak' events (or directional gestures) for various pie menus, regular menus, and operators. This work fine for regular menus and operators. After calling a pie, however, the tweak button defined in the input editor (left, right, middle, action) becomes unusable. \n\n1) Open User Prefs, select the add-ons tab, and enable Official Pie menu addon\n2) Select the 'Input' tab in User Prefs, and in the keymap filter (set to name), type the the word 'pie' without quotes\n3) The first filtered pie is 'Frames'. change the hotkey as follows: \nClick the event mapping dropdown (its set to keyboard by default), and change to Tweak. two new options appear next to it: 'Left', and 'Any'.\nChange the type of event, or leave it as 'Left' \nChange 'Any' to a direction.\nSelect a modifier key such as Shift\n\n4) go to the 3d view and execute the new shortcut you set up. The playback options should appear in the pie menu but are not selectable. Hitting the escape key will get out of the pie, but the defined tweak button is not usable.\n\nNOTE: If you use the default right click select, you can still place the cursor, but can't trigger the shortcut again. If left click select is set, then normal tweak events like moving selected objects are also not possible.\n\nNOTE: Toggling Object/Edit mode will make the button usable again.", "Box Select (B) changed to Left Mouse unable to Deselect\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.18\n\nBroken: 3.0, 3.5\nWorked: 2.93.9 (31712ce77a6a), 3.0.0 broke it\n\nBox Select (standard B) changed to Left Mouse ignores Gesture Box Deselect. Instead Deselect now selects Objects.\n\n(See attached Image. These preferences work different until 3.0.0. The Image and Full Version below is basically an alternative to Blenders standard \"Left Mouse Drag Box Select\" which removes the need to hold the mouse button)\n\nPreferences > 3D View > Box Select:\n\n-SHORT VERSION-\n1. Set Left Mouse as Event (Press)\n2. Go to Gesture Box\n3. Change \"Begin - Left Mouse\" to \"Release\"\n4. Use a Left Mouse click + release in 3D View to begin the Box.\n\nProblem: Middle Mouse (Deselect) only selects Objects.\n\n-FULL VERSION-\n1. Set Left Mouse as Event (Press)\n2. Set Mode to Extend\n3. Go to Gesture Box\n3.1 Change \"Begin - Left Mouse\" to \"Release\"\n3.2 Change \"Select - Left Mouse\" to \"Press\"\n3.3 Deactivate \"Begin - Middle Mouse\"\n3.4 Change \"DeSelect - Middle Mouse\" to \"Press\"\n4. Use a Left Mouse click + release in 3D View to begin the Box.\n\nExpected: Left Mouse will select, Middle Mouse will deselect.\nProblem: Middle Mouse (Deselect) only selects Objects.\n\n------\n\nI could not find a way to have it Deselect or if there is a conflict with other settings. Can you take a look at this? Thank you.\n\n", "Wrong status bar keymap for release confirms and right click select\nOperating system: Manjaro x64\nGraphics card: HD5770\n\nBroken:\n2.80, 81ea815dcb6, blender2.8, 2018-12-12\nWorked: 2.79b\n\nWhen grabbing something(object, vertex, etc) by right clicking on it and then releasing/clicking on the same button the action is not performed, object doesn't return to its initial position. \nHowever if you use G or a Tool to grab instead of mice click Cancel by Right clicking option works.\n\nRight click on a Cube(in 2.79 thats enough to start moving an object but not here here you have to hold the button) AND HOLD the button\nDrag mice somewhere to move the Cube\nRight Click again to Cancel movement\nNothing happens.\n[blender_bug-2018-12-13_22.18.53.mkv](blender_bug-2018-12-13_22.18.53.mkv)\nI'm also attaching video file for it to be more clear.", "Support expanded operator menus in the right click button context menu\nEdit, made this a TODO, since support for these kinds of buttons has not been added.\n\nCurrently menu items defined by `uiItemMenuEnumO_ptr` don't have right click menus, this is because they reference the operator from a `MenuItemLevel` struct.\n\n----\n\nOperating system: Windows-10-10.0.14393 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.86\n\nBroken: version: 2.82 (sub 1)\nWorked: (optional)\n\n\nAfter noticing that right click context menu was a bit broken on some menu items, I proceeded to test it on every single drop down menu in Blender and came up with this list below.\nAll the items here fail to display right click context menu, instead it simply opens the window the same way as a left click\n\n\n\n3D Viewport\n     Object mode\n          Select\n                Select All by Type...\n                Select Grouped\n                Select Linked\n\n          Add\n                Grease Pencil\n                Empty\n                Force Field\n\n          Object\n                Set Origin\n                Convert to\n\n     Edit mode\n          Mesh\n                Sort Elements...\n                Separate\n\nUV Editor\n          UV\n                Snap to Pixels\n\nVideo Sequencer\n     Sequencer\n          View\n                Waveform Displaying\n          Select\n                Grouped\n          Marker\n                Duplicate Marker to Scene\n          Add\n                Fade\n\nDope Sheet\n     Dope Sheet (and all other editor subtypes)\n          Marker\n                Duplicate Marker to Scene\n          Channel\n                Toggle Channel Setting\n                Enable Channel Setting\n                Disable Channel Setting\n                Extrapolation Mode\n                Move\n          Key\n                Snap\n                Mirror\n                Keyframe Type\n                Handle Type\n                Interpolation Mode\nTimeline\n          Marker\n                Duplicate Marker to Scene\n\nGraph Editor/Drivers\n          Marker\n                Duplicate Marker to Scene\n          Channel\n                Toggle Channel Setting\n                Enable Channel Setting\n                Disable Channel Setting\n                Extrapolation Mode\n                Move\n          Key\n                Snap\n                Mirror\n                Insert Keyframes\n                Add F-Curve Modifier\n                Handle Type\n                Interpolation Mode\n\nNon Linear Animation\n          Marker\n                Duplicate Marker to Scene\n          Edit\n                Snap\n                Track Ordering...\nFile Browser\n          View\n                Display Size\n                Recursion\n\n\n", "Selection problem on new 'add modifier' menu\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 850M/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 537.13\n\nBroken: version: 4.0.0 Alpha\n\nin new 'add modifier menu' which introduced in Blender 4.0 if you hover your mouse over a main menu items like 'edit' to open edit's submenu then move your mouse diagonally from top right to bottom left on main menu, other main menu items not highlighting or selecting.\nI attach a video take look at it, in this video I move my mouse from edit(while it's open) to physics , non of menus open, not even physics which I stop mouse there no matter how much time my mouse stay there it will not open.\nalso I moved my mouse from edit to generate and also menu did not open.\n\n", "Changing the Orientation of Modifiers\nCurrently, LineArt is created in 3D space, however, the Modifiers can be oriented incorrectly.\n\nIn this example, the Noise Modifier is moving the Grease Pencil points on the Normal but not up and down on Z. There is currently no way to change it.\n\n[Noise - Position.mp4](Noise_-_Position.mp4)\n\nThis is a request for a way to change the orientation of the Modifiers.", "Edge/Vert Slide fail to properly \"redo\"\nCurrently vertex & edge-slide use the mouse input to calculate which edge to slide, the direction to slide, and the **Even** toggle is set at runtime.\n\nUsing operator redo is not re-running with the same options, so redoing the action fails.\n\nWe could simply disable redo with this operator, however being able to toggle **Correct UV** is important.\n\nAdding this report to merge in duplicates, since there are various reports on this, which all have the same root-cause.", "Lasso Select Broken if Select key is changed from Default Left to Right\nOperating system: Windows 10 (Home)\nGraphics card: Nvidia 1070 \n\n**Blender Version** \nBroken: 2.90.0 (Steam)\nWorked: Not sure, Has been like this since before 2.79b\n\nWhen the default select mode is changed from Left to Right Mouse button Lasso Select no longer works with Tweak + RMB Any. \n\nSplash Screen Change on initial set up or Select button or Edit - Preferences - KeyMap - Select With Mouse Button\n\nAffected: any View Mode that has the Lasso option. (Edit mode, Object Mode, Grease Pencil, etc)\n\n\nSee attached MP4.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\nStart Blender and use Start up Cube. [BlendBug_2.mp4](BlendBug_2.mp4)\n\nTemporary fix if the Select key is changed from Left to Right Mouse Button is to walk through the KeyBindings and set anything that uses Lasso to Tweak Right and anything that was using Tweak Left to Tweak Right where Lasso is was changed. " ]

[ "Cant confirm clamped edge slides with left click when Emulate 3 button mouse is on.\nOperating system: Windows 10 Version\t10.0.17763\nGraphics card: NVIDIA GeForce GTX 980 Ti\n\nBroken: Latest 2.81 and 2.7\n\nWhen \"emulate 3 button mouse\" is turned on, you are unable to confirm clamped edge slides with left mouse button.\n\n\nTurn on \"emulate 3 button mouse\" in preferences.\nCreate an any object.\nEdge slide with ALT held down.\nyou cannot then confirm with left mouse button\n" ]

[ "Crash when moving to Shading workspace\nOperating system: Darwin-16.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-10.17.5 355.10.05.45f01\n\n|Broken:|version: 2.83.0, commit date: 2020-06-03,|hash: `211b6c29f7`\n| -- | -- | -- |\n||version: 2.90, commit date: June 08,|hash: `9f7d84b656`\n|Worked:|version: 2.82|-\n\nWhen I open Blender, without changing anything in the default scene, if I move to the 'Shading' workspace my system freezes for a few seconds and crashes (it closes).\n\n- Default startup scene\n- click on 'Shading' workspace from the top bar\n", "Selecting vertices and centering the model crashes Blender\nOperating system: Windows 10\nGraphics card: RTX 2070 \n\nBroken: 03015a9b222e\n\n\nSelecting some vertices and trying to center the selection crashes blender. See the crash log\n\n[crash.txt](crash.txt)\n\nOpen the attached .blend\nSelect the object\nGo into edit mode\nSelect all vertices\nCenter (press Numpad .)\n\n\n\n\n\n[_13122021_2147_15.blend](_13122021_2147_15.blend)", "Blender crashes: OpenVDB math::Transform::read throws exception. \nOperating system: Darwin-19.5.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.9.15\n\nBroken: version: 2.83.2\n\nBlender keeps crashing.\n\nI have a project were I created an animation of modified arrays. Everytime I play it it crashes\n\nCrash report: P1536\n\ncrashed thread: \n```\n\tThread 35 Crashed:\n0 libsystem_kernel.dylib \t0x00007fff726c733a __pthread_kill + 10\n1 libsystem_pthread.dylib \t0x00007fff72783e60 pthread_kill + 430\n2 libsystem_c.dylib \t0x00007fff7264e8a2 __abort + 139\n3 libsystem_c.dylib \t0x00007fff7264e817 abort + 135\n4 libc++abi.dylib \t0x00007fff6f8ae458 abort_message + 231\n5 libc++abi.dylib \t0x00007fff6f89f8bf demangling_terminate_handler() + 262\n6 libobjc.A.dylib \t0x00007fff713da5b1 _objc_terminate() + 104\n7 libc++abi.dylib \t0x00007fff6f8ad887 std::__terminate(void (*)()) + 8\n8 libc++abi.dylib \t0x00007fff6f8b01a2 __cxxabiv1::failed_throw(__cxxabiv1::__cxa_exception*) + 27\n9 libc++abi.dylib \t0x00007fff6f8b0169 __cxa_throw + 113\n10 org.blenderfoundation.blender \t0x0000000110291c61 openvdb::v7_0::math::Transform::read(std::__1::basic_istream<char, std::__1::char_traits<char> >&) + 4801\n11 org.blenderfoundation.blender \t0x000000011026082b openvdb::v7_0::io::Archive::readGrid(std::__1::shared_ptr<openvdb::v7_0::GridBase>, openvdb::v7_0::io::GridDescriptor const&, std::__1::basic_istream<char, std::__1::char_traits<char> >&) + 1787\n12 org.blenderfoundation.blender \t0x0000000110286589 openvdb::v7_0::io::File::Impl::unarchive(openvdb::v7_0::io::File const&, std::__1::shared_ptr<openvdb::v7_0::GridBase>&, openvdb::v7_0::io::GridDescriptor const&, openvdb::v7_0::io::File::Impl::NoBBox) + 57\n13 org.blenderfoundation.blender \t0x0000000110282c91 openvdb::v7_0::io::File::readGridByName(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, openvdb::v7_0::math::BBox<openvdb::v7_0::math::Vec3<double> > const&) + 465\n14 org.blenderfoundation.blender \t0x0000000110282ab6 openvdb::v7_0::io::File::readGrid(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 54\n15 org.blenderfoundation.blender \t0x00000001128a2cc9 MANTA::updateGridFromVDB(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, float*, bool) + 361\n16 org.blenderfoundation.blender \t0x0000000112890985 MANTA::updateGridFromFile(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, float*, bool) + 869\n17 org.blenderfoundation.blender \t0x0000000112890cee MANTA::updateNoiseStructures(FluidModifierData*, int) + 558\n18 org.blenderfoundation.blender \t0x000000011287e2b5 manta_update_noise_structures + 21\n19 org.blenderfoundation.blender \t0x000000010df34cd0 BKE_fluid_modifier_processDomain + 3168\n20 org.blenderfoundation.blender \t0x000000010df316ce BKE_fluid_modifier_do + 142\n21 org.blenderfoundation.blender \t0x000000010e3ad4f6 applyModifier + 54\n22 org.blenderfoundation.blender \t0x000000010dea5b0a mesh_calc_modifiers + 2778\n23 org.blenderfoundation.blender \t0x000000010dea4d6b mesh_build_data + 123\n24 org.blenderfoundation.blender \t0x000000010dea4035 makeDerivedMesh + 133\n25 org.blenderfoundation.blender \t0x000000010dfe1281 BKE_object_handle_data_update + 1089\n26 org.blenderfoundation.blender \t0x000000010dfe1704 BKE_object_eval_uber_data + 52\n27 org.blenderfoundation.blender \t0x000000010e9f36ba DEG::(anonymous namespace)::deg_task_run_func(TaskPool*, void*, int) + 138\n28 org.blenderfoundation.blender \t0x00000001142cadbc task_scheduler_thread_run(void*) + 364\n29 libsystem_pthread.dylib \t0x00007fff72784109 _pthread_start + 148\n30 libsystem_pthread.dylib \t0x00007fff7277fb8b thread_start + 15\n```\n", "Quick Smoke crash guide February 3, 2023\nOperating system: macOS-13.2-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon Pro 575 1.2\n\nBroken: version: 3.5.0 Alpha\n\nBake ALL button within Sample 128 from Frame 1 to 220 cause CRASH GUIDE\n\nBake ALL button within Sample 128 from Frame 1 to 220 cause CRASH GUIDE\n\n[image.png](image.png)\n[image.png](image.png)\n[Blender-2023-02-03-113746.ips](Blender-2023-02-03-113746.ips)\nTest File: [3.5.zip](3.5.zip)\n```\nCall Stack: Main thread\n__psynch_cvwait\n_pthread_cond_wait\ntake_gil\nPyEval_RestoreThread\nPyGILState_Ensure\npy_timer_execute\nBLI_timer_execute\nwm_event_do_notifiers\nWM_main\nmain\nstart```", "addon preferences changed by code (not by user in UI) are not saved\nOperating system: Ubuntu\nGraphics card: RTX\n\nBroken: master\nWorked: ?\n\naddon preferences changed by code (not by user in UI) are not saved at Blender close.\n\n- Open Blender. Load Factory Settings, close Blender. \n- Open Blender\n- In console, change an addon preferences. For example, change device of Cycles : bpy.context.preferences.addons['cycles'].preferences['compute_device_type'] = 1 (this change to CUDA)\n- Close Blender. No log saying preferences are saved.\n- Open Blender again, check that device is still None. pref save is KO\n- Change preferences to \"CUDA\" by clicking with your mouse\n- Close Blender\n- See that a log saying preferences are saved\n- Open again, you can see that preferences are set to CUDA, pref save was OK", "UI: Discussion Topics & Possible To-Do's for 3.0\n3.0 is the next chance to do breaking changes which should not go into normal releases. That may be: higher level design changes, muscle memory breaking, documentation breaking (e.g. changed operator names) or non-trivial visual changes. There should still be no .blend file compatibility breaking changes.\n\nIn this task we can collect such changes for consideration. This is not a \"wish a feature\" list. These are things we discussed already but couldn't do earlier because they would be too disruptive.\n\n* Active vs. selected (#82359 (Stronger binding of Active and Selected))\n* Keymap polish\n\n - Consistent selection with modifier keys #78253 (Formalize Default Selection Behavior and Key-map)\n - {key 1}/{key 2}/{key 3} for changing into vertex/edge/face edit mode #88071 (Keymap: use number keys for mode switching (2nd Iteration))\n - {key D} for view-navigation sub-menu #88092 (Keymap: use D-key for view navigation pie-menu (2nd iteration)).\n - {key Alt} usage #83689 (Default Keymap Update: Improve tool access using Alt-LMB)\n\n* Toolsystem wrap-up (change toolsystem defaults if needed)\n* Theme polish (#74358)\n* General defaults (Preferences, startup configuration, bundled workspaces, etc)\n - [ ] Set default subdivision modifier levels to be 1 for viewport and render.\n* More dynamic file selection (e.g. show import/export settings based on file selection)\n* Reevaluate closing popups on mouse moves.\n* Improve naming:\n**Avoid \"link\" for anything unrelated to link/append (e.g. *Link Materials* -> *Share Materials*).** Correct usages of \"copy\" (e.g. *Copy Modifiers* -> *Replace Modifiers*).\n* 3D cursor default orientation method to \"Surface\"\n* #89266 (Investigate renaming the \"Video Sequencer\" editor)\n* #83612 (UI: Tracking: Migrate to active tools based toolbar)", "Intel GPU: switching between applications when in full screen mode causes glitches or screen flickering\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: Intel(R) HD Graphics 530 Intel 4.5.0 - Build 30.0.101.1340\n\nBroken: version: 3.1.2\n\nWhen in Fullscreen, switching to other program and then returning to blender, glitch(black screen or something else) occurs suddenly and disappears.\nThis is not critical bug, but it would be good if it is fixed.\n\n\n- Open default blender file\n- {nav Window > Toggle Window full screen}\n- Switch to another application and again back to Blender", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Weird VSE crash\nOperating system: Win11\nGraphics card: RX580\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen) blender 4.0.0 alpha\nWorked: (newest version of Blender that worked as expected) no clue\n\nBlender crashes on opening file. Was trying to make an alternative caption that gets around the Multiply color issue #112267 but came across a new bug.\n\nBased on the default startup or an attached .blend file (as simple as possible).\nOpen the file I've attached\n\n", "Modifying keymap won't tag prefs dirty\nWindows 7 Professional 64-bit service pack 1\nIntel Core i7-2600\n8GB RAM\nNVIDIA GTX 1060 6GB\n\nBroken: blender-2.80.0-git.aa003c73245f-windows64\n\n\"Save Preferences\" button, shown when \"Auto-Save Preferences\" is turned off, stays unavailable after modifying keymap.\n\n1. Make sure \"Auto-Save Preference\" is turned off. If it isn't, turn off this and press \"Save Preferences\" button.\n2. Modify keymap, editing the settings used to activate an operator (example: search for `Clear Rotation` operator, changing the `Clear Delta` option will not tag prefs dirty).", "Random crashes with NVidia + GPU Subdivision (sometimes when activating X-Ray or Wireframe overlay)\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\n| Graphics card: | Driver : | User:\n| --- | --- | ---\n| NVIDIA GeForce RTX 3090/PCIe/SSE2 | 4.5.0 NVIDIA 516.94 | @GabrielMoro\n| NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 | 4.5.0 NVIDIA 512.15 | @STYXtheMUTT\n| NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 | 4.5.0 NVIDIA 516.94 | @Jojolix\n|NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2|NVIDIA 516.59|@MassimilianoPuliero\n|NVIDIA GeForce RTX 3090|NVIDIA 516.94|@The5\n\nBroken: version: 3.2.2\n\nOn some seemingly random occasions, Blender crashes with the following stack:\n```lines=10\nStack trace:\nblender.exe :0x00007FF77C1458B0 BKE_subdiv_free\nblender.exe :0x00007FF77C147880 BKE_subsurf_modifier_subdiv_descriptor_ensure\nblender.exe :0x00007FF77BFFC3D0 <lambda_c000e58b9a8caf72c54e0e02a46effdb>::operator()\ntbb.dll :0x00007FFDBAF64FD0 tbb::interface7::internal::isolate_within_arena\nblender.exe :0x00007FF77BFFC6F0 BKE_mesh_wrapper_ensure_subdivision\nblender.exe :0x00007FF77CB38FA0 raycast_obj_fn\nblender.exe :0x00007FF77CB375E0 iter_snap_objects\nblender.exe :0x00007FF77CB3E530 transform_snap_context_project_view3d_mixed_impl\nblender.exe :0x00007FF77CB3F0E0 ED_transform_snap_object_project_view3d\nblender.exe :0x00007FF77CB5E7C0 applyFaceProject\nblender.exe :0x00007FF77CB5EE00 applySnappingIndividual\nblender.exe :0x00007FF77CB72F70 recalcData_objects\nblender.exe :0x00007FF77CB81010 applyTranslation\nblender.exe :0x00007FF77CB46BE0 transformApply\nblender.exe :0x00007FF77CB41920 transform_modal\nblender.exe :0x00007FF77C229D30 wm_macro_modal\nblender.exe :0x00007FF77C20F9E0 wm_handler_operator_call\nblender.exe :0x00007FF77C210BD0 wm_handlers_do_intern\nblender.exe :0x00007FF77C210000 wm_handlers_do\nblender.exe :0x00007FF77C216590 wm_event_do_handlers\nblender.exe :0x00007FF77C1FC380 WM_main\nblender.exe :0x00007FF77B2F12A0 main\nblender.exe :0x00007FF7814128B0 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFDC7B67020 BaseThreadInitThunk\nntdll.dll :0x00007FFDC85A2680 RtlUserThreadStart\n```\n```lines=10\nStack trace:\nblender.exe :0x00007FF7A50C5B10 BKE_subdiv_update_from_mesh\nblender.exe :0x00007FF7A50C7880 BKE_subsurf_modifier_subdiv_descriptor_ensure\nblender.exe :0x00007FF7A52A1110 draw_subdiv_create_requested_buffers\nblender.exe :0x00007FF7A52A2080 DRW_create_subdivision\nblender.exe :0x00007FF7A52D1870 DRW_mesh_batch_cache_create_requested\nblender.exe :0x00007FF7A52C6000 drw_batch_cache_generate_requested\nblender.exe :0x00007FF7A529F1A0 drw_engines_cache_populate\nblender.exe :0x00007FF7A529B8C0 DRW_draw_render_loop_ex\nblender.exe :0x00007FF7A529C9B0 DRW_draw_view\nblender.exe :0x00007FF7A5D3FA00 view3d_main_region_draw\nblender.exe :0x00007FF7A5551070 ED_region_do_draw\nblender.exe :0x00007FF7A51A3700 wm_draw_window_offscreen\nblender.exe :0x00007FF7A51A3560 wm_draw_window\nblender.exe :0x00007FF7A51A2FB0 wm_draw_update\nblender.exe :0x00007FF7A517C380 WM_main\nblender.exe :0x00007FF7A42712A0 main\nblender.exe :0x00007FF7AA3928B0 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFD1C4A54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFD1D524830 RtlUserThreadStart\n```\n\nOccasions observed:\n- try to snap objects on faces (the object that shall be snapped to is usually a complex mesh with modifiers like subdivisions, mirror, solidify, bevel and booleans)\n- toggle the shading mode between shaded and wireframe or between X-Ray and Wireframe Overlay.\n[intro.crash.txt](intro.crash.txt)\n\n\n- open attached file\n- disable and enable X-Ray multiple times\nIt happens specially when working fast.\n\n[BLENDER_Youre_Alive_v67.blend](BLENDER_Youre_Alive_v67.blend)\n\n", "Keyed Particle Duration Crash\nOperating system: Windows 7 64\nGraphics card: nVidia GeForce GTX 970\n\nv.2.80.60\nDate: 2019-05-10 22:21\nHash: de9d846353bd\n\nBlender will sometimes crash when a keyed physics type particle system references more than one particle system. This usually occurs when attempting to change the duration setting of the second particle system.\n\nIn the provided file, change the duration setting in the particle system settings. Blender will crash.\n\n[Keyed Particle Duration Crash.blend](Keyed_Particle_Duration_Crash.blend)", "GPencil: Blender crash when file open\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.4.0 Beta\n\nAs soon as I open attached file Blender crashes. Attached also crash log file.\n\n- Open attached file\n- Change radius (say 0.6)\n- Save file\n- reopen it (Crash)\n[#102930.blend](T102930.blend)", "crash when x-ray and wireframe are enabled\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.1 Release Candidate\n\nWhen I am working and periodically switch to x-ray and wireframe mode, the program crashes. It just shuts down and that's it. I thought it was related to GPU-subdiv, but today it was not confirmed, I turned off GPU-subdiv and literally after about 20 minutes of work when switching on X-ray and wireframe mode the program crashes. I turn on mode the shift+z hotkeys.\nThis happens randomly!!!\n\n\n\n", "The program crashes when texturing \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1660/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.72\n\nBroken: version: 2.93.0\n\nWhen try texture object in texture paint mode, and there are another map in shader tab. Program crashes. \n\n1. In default (or other) scene go to Texture paint window, create new texture for paint. \n2. Apply this texture to diffuse color object in shader editor. (or in object properties)\n3. Go back to texture window. We make sure that everything is painted correctly\n4. Go to shading window and drag&drop any other texture to node window. (NOT APPLY TO DIFFUSE COLOR (or APPLY))\n5. Go back to texture paint window. For some reason, the dragged texture itself was assigned to the duffuse color.\n6. Change linced texture in image editor (in left window)\n7. Try paint on RIGHT window (3D viewport)\n8. Click on left texture editor (try paint in 2D ) program crash. \n\nVideo steps: 0Ophy4Lmx0A\n\nDebug logs:\n[blender.crash.txt](blender.crash.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[Blender_crash.webm](Blender_crash.webm)\n\n" ]

[ "Crash close window renderer or window users preference\nHi\n\nWindows MACHINE:X64 - Gforce 940M\n\nBroken:version:\n2.80 (sub 20), branch: blender2.8, commit date: 2018-07-09 22:47, hash: 1112f49, type: Release\nbuild date: Tue 07/10/2018, 04:28 AM\nplatform: Windows\n\n\nClose Windows user preference or close window rendered image \"F12\" -> Crash windows\n\n1/ Start Blender 2.8 Alpha\n2/ Edit>User Preferences\n3/ Close Windows cross at the top right -> Crash windows\n\n1/ Start Blender 2.8 Alpha\n2/ EEVEE test texture and rander F12\n3/ Close Windows cross at the top right -> Crash windows\n\nBest regard" ]

[ "Cycles crashes on viewport rendering in CUDA driver\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.0\nWorked: 3.2\n\nThe crash happens while navigating the viewport in shading mode, using Cycles.\n\nThis crash reappeared in the latest 3.3 release. It happened in older releases, then it got stable, but now it's back again.\nApparently, it has to do with Cuda.\nIt's not happening in one particular scene only, but rather randomly regardless of the complexity of the scene.\n\n[Cave-CloseUp-v006.crash.txt](Cave-CloseUp-v006.crash.txt)\n\n[Cave-CloseUp-v005.crash.txt](Cave-CloseUp-v005.crash.txt)\n\n[blender.crash.txt](blender.crash.txt)", "GHOST:: m_windows not clearing list of fullscreen windows \nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.59\n\nBroken: version: 2.83 (sub 8)\n\nI just wanted to warn you about a potential bug in GHOST when we close a fullscreen window. The problem is that I don't know how to reproduce it in blender official.\nThen this is just if someone is interested...\n\nWhen we close a fullscreen window, GHOST::m_windows list doesn't erase the fullscreen window. Then m_windows list size increases.\nWhen I erase the fullscreen window from m_windows list, it is fixing a crash with our fork of blender (upbge)\n\nI describe the steps to reproduce the bug here: 4fc32990010d5765ba1abd523d57b252be73a657\n\nThis needs to compile restart_hacks branch from upbge github: upbge\n\nThere's a link to the testfile in the commit in mentionned.\n\nJust in case:\n\n[ledi.blend](ledi.blend)\n\n", "Closing System Console can lead to loss of work\nOperating system: Windows\n\nAny\n\n* Start Blender\n* Toggle the System Console (2.83 and newer `Windows > Toggle System Console`\n* Press the - [x] button on the system console", "Collada export crashes (when relative \"//\" is in `Preferences` > `File Paths` > `Temporary Files`)\nOperating system: macOS 12.1\nGraphics card: M1 MBP\n\nBroken: 3.0 with the attached userpref.blend\nBroken: 2.93.8\nBroken: 2.83.19\n\n\n\nfrom factory-startup:\n\n- - [x] type `//` in `Preferences` > `File Paths` > `Temporary Files`\n- - [x] `Export` > `Collada`\n- crash\n\nnote: in a debug build, the following assert is hit after step [1]\n```\n./bin/blender(BLI_system_backtrace+0x26) [0xf4a5f1b]\n./bin/blender(_BLI_assert_print_backtrace+0x16) [0xf344596]\n./bin/blender(BLI_exists+0x2a) [0xf497c6c]\n./bin/blender(BLI_is_dir+0x18) [0xf497d21]\n./bin/blender() [0x3c2a38d]\n./bin/blender(BKE_tempdir_init+0x27) [0x3c2a5e2]\n./bin/blender() [0x4c1ec17]\n./bin/blender() [0x4a2e592]\n./bin/blender(RNA_property_update+0x50) [0x4a2e7ca]\n./bin/blender() [0x55d728b]\n./bin/blender() [0x55f3b4a]\n./bin/blender() [0x435810e]\n./bin/blender() [0x435e018]\n./bin/blender() [0x435e5cc]\n./bin/blender(wm_event_do_handlers+0x7a0) [0x435fd81]\n./bin/blender(WM_main+0x30) [0x4351d2e]\n./bin/blender() [0x3c236af]\n/lib64/libc.so.6(+0x2d590) [0x7fb0c4842590]\n/lib64/libc.so.6(__libc_start_main+0x89) [0x7fb0c4842649]\n./bin/blender(_start+0x25) [0x3c23095]\nBLI_assert failed: source/blender/blenlib/intern/storage.c:348, BLI_exists(), at '!BLI_path_is_rel(path)'\n```", "Background scene cause a crash in render animation\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.55\n\nBroken: version: \n - 2.92.0 Alpha\n - 2.91.0\n - 2.90.1\nWorked: 2.83.9\n\nBackground scene makes blender unstable and cause a crash in render animation\n\n- Open default startup file.\n- Create a new scene named \"Scene.001\".\n- Assign default scene \"Scene\" as the background scene of \"Scene.001\" in scene properties.\n- Assign default camera as the camera of \"Scene.001\" in scene properties.\n- Hit render animation from render menu.\n- After a couple of frames rendered, blender crashes.\n\nTested in different situations i had some problems.\nI think assigning a background scene makes blender unstable in general.\nOccasionally crashes when opening a file which has background scene.\nThe worst case is render animation.\n\n[Background_Scene_Crash.mp4](Background_Scene_Crash.mp4)\n\nReproduction steps with this file:\n1. Open file.\n2. Start Render Animation.\n[#83852.blend](T83852.blend)\n", "Blender stops updating display after GPU spike\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Plus Graphics Intel 4.5.0 - Build 27.20.100.8681\n\nBroken: version: 2.92.0 Beta\nWorked: I have not found a version that works as expected on this computer. I have tried 2.79, 2.8, 2.91\n\nI was premature in saying the problem is gone for [#85061 ](T85061).\n>After a few minutes of using Blender, the window freezes. Windows does not detect that it is not responding, but Blender does not respond to any input, including clicking on the close button. All elements that are drawn by Blender remain as they were at the moment of the crash except if I resize the window smaller.\n\nIf I watch the performance tab on Windows Task Manager, the freeze coincides with a 100% spike on the GPU.\nAs long as that spike doesn't happen, there is no problem, but once it does, Blender stops responding to any input, but it does not crash.\nI have updated my graphics drivers.\n[blender_debug_output.txt](blender_debug_output.txt)\n\nI have not determined exact steps. It has occurred on multiple files doing multiple different tasks. As far as I can tell, there is no correlation between what I am working on or what task I am performing, and the error occurring.\n", "Eevee randomly causes system to crash when rendering\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.23 (8GB memory)\nCPU: Intel Core i9 9900k\nRAM: 32GB DDR4\n\nBroken: 3.0.0\n\n\nI'm new to Blender and currently trying to switch from my usual C4D and Redshift workflow, just to let you know.\n\nI have a fairly simple scene with a 250 frame animation. Everytime I render the animation with Eevee as a .png sequence with an .exr file for the layers, the process causes my system to crash.\nThe amount of samples seems to correlate with the amount of frames I'm able to render before failure. So I guess it might have something to do with memory.\n\nI'm not sure how that can be as more complex scenes never failed me before using Redshift as a render engine.\n[ocean_final.blend](ocean_final.blend)\nThe scene makes use of a lot of volumetric light and I also calculate the bloom.\n\nThere aren't a lot of vertices involved in the scene, but I'm not sure how I can display a value to give you here.\n\nAll drivers are up to date and I also reinstalled Blender.\n\nNot sure, but I attached the .blend file as well as logs created in debug mode.\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)", "Crash when auto script execution is enabled\nOperating system: Windows 10 64-bit operating system, x64-based processor\nGraphics card: NVIDIA GeForce RTX 2060\n\nBroken: Blender 4.0.0 7372b3bdda35 and later\nWorked: never? 2.93 fails too\n\nBlender daily build crashes immediately on launch. The blender start screen isn't shown at all, only a grey window which then immediately closes. The crash log indicates that it is a memory EXCEPTION_ACCESS_VIOLATION (see log file attached).\n\n- Open preferences, enable `auto script execution`\n- Save prefs and close blender\n- Rename attached file and replace it with existing startup file\n- Launch blender (crash)\n\n- - -\n<details>\n <summary> Original report </summary>\nStart Blender through Steam, Blender Launcher, blender_debug_log.cmd, blender_debug_gpu.cmd, or blender_debug_gpu_glitchworkaround.cmd.\nA grey window opens with the Blender logo in the top bar.\nThe window closes as Blender crashes.\n\n[ Crashing_4-0-0_daily.blend](https://projects.blender.orgattachment)\n</details>", "Crash Guide during render complete 2 December 9, 2022\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.4.0\n\nCrash Guide during render 2\n\nSame crash guide during render before and today.\nSee youtube: ZeCA2p5wISY\n\nHistory previous.\nT100691\n\n\n\n\n[Blender-2022-12-09-161546.ips](Blender-2022-12-09-161546.ips)\n\n[Horizontal Video.zip](Horizontal_Video.zip)", "EXCEPTION_ACCESS_VIOLATION when changing shaders\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.79\n\nBroken: version: 3.1.0\n\nThis also seems to crash 2.93 and 3.0. \n\nEXCEPTION_ACCESS_VIOLATION causes crash when some shaders are applied.\n\nOpen the file. In the shader editor, set the material of the cube to rim_gradient. Shift-click on the specified shader node in the picture to turn it into the material output. Observe crash.\n\n\n[crash.blend](crash.blend)\n\n\n", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "stack_overflow exception while rendering with viewport render on\nOperating system: win10\nGraphics card: 2070super\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: 2.83.4, 2.91 6fe609d4c3\n\n\nwhen pressing f12 in render mode, blender crashes,\nwhen pressing render in solid mode, blender doesnt crash\n\n[blender_debug_output.txt](blender_debug_output.txt)\n```\nError : EXCEPTION_STACK_OVERFLOW\nAddress : 0x00007FFF42963FDB\nModule : C:\\Windows\\system32\\DriverStore\\FileRepository\\nv_dispsi.inf_amd64_6af46268c19e4980\\nvcuda64.dll\n```\n[2.blend](2.blend) \n1. switch to render mode\n2. press f12 & watch task manager.\n3. blender will close", "Blender Freezes and Crashes on AMD Laptops with White and Black boxes\nOperating system: Windows 10 & 11\nGraphics card: Radeon RX 6600M ATI Technologies Inc. 4.5.14800 Core Profile Context 22.4.2 30.0.15021.7000\n\nI've been using Blender with many versions. From 2.8, to 3.1 and the laptops have issues with all of them.\n\nBlender will freeze, white and black boxes will appear, and then Blender will crash and Windows will close it.\n\nI don't have exact steps, because it happens infrequently. It can happen as quickly as opening Blender, or take an hour of work.\n\nI am a teacher using HP Omen laptops with AMD CPU's and graphics cards. I have 21 laptops, and about 5-7 of them have issues with Blender. I have done a clean install of the graphics drivers, updated Blender to the most recent version, and tried everything I can think of. My personal laptop rarely ever crashes, so it's hard for me to replicate the issue.\n\nI have attached log files from crashes from multiple laptops. Half the laptops have been upgraded to Windows 11, and that doesn't seem to make it more or less likely to crash. My hope is these logs can pinpoint what's happening so I can fix it. I\n\nLog files created with `blender_debug_gpu_glitchworkaround.cmd`\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\nLog files created with `blender_debug_gpu.cmd`\n\n[blender_debug_output (1).txt](blender_debug_output__1_.txt)\n\n[blender_system_info (1).txt](blender_system_info__1_.txt)\n\nWarning and errors:\n```lines=10\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_multi_rect_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_multi_rect_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_text VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_text FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_uniform_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_uniform_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_INST_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_INST_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_rect_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_rect_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_flat_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_flat_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_LINE_DASHED_UNIFORM_COLOR_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_LINE_DASHED_UNIFORM_COLOR_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_3D_uniform_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_3D_uniform_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_desaturate_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_desaturate_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_composite_studio VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_composite_studio FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_merge_infront VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_merge_infront FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transparent_resolve VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transparent_resolve FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_effect_outline VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_effect_outline FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_taa VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_taa FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_1 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_1 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_2 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_2 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_antialiasing_263 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_antialiasing_263 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_414 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_414 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_449 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_449 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_401 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_401 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 GeomShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n | \n 6 | #define DFDY_SIGN 1.0\n | ^\n | (#310) Inputs that are integers should be qualified with the interpolation qualifier \"flat\"\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_wire_469 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_wire_469 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_wire_538 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_wire_538 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_solid_559 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_solid_559 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_stick_518 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_stick_518 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_499 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_499 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_489 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_489 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_wire_579 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_wire_579 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_background_276 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_background_276 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_uniform_color_1465 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_uniform_color_1465 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_grid_905 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_grid_905 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_885 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_885 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_groundline_924 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_groundline_924 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_loose_point_983 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_loose_point_983 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_point_1004 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_point_1004 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_facing_1022 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_facing_1022 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_grid_1055 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_grid_1055 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 GeomShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_vert_1170 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_vert_1170 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_1204 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_1204 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_pointcloud_1257 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_pointcloud_1257 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_gpencil_1230 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_gpencil_1230 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_detect_1272 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_detect_1272 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_dot_1410 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_dot_1410 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_shape_1429 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_shape_1429 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_wireframe_1589 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_wireframe_1589 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nERROR (gpu.debug): : GL_INVALID_VALUE : generated before glGenTextures\n```\n\nStack trace:\n```lines=10\nblender.exe :0x00007FF6F22F7D90 bli_windows_system_backtrace_stack_thread\nblender.exe :0x00007FF6F22F7010 BLI_system_backtrace\nblender.exe :0x00007FF6F228C530 blender::gpu::debug::debug_callback\nblender.exe :0x00007FF6F228C3F0 blender::gpu::debug::check_gl_error\nblender.exe :0x00007FF6F220EE50 blender::gpu::GLTexture::GLTexture\nblender.exe :0x00007FF6F21F8420 blender::gpu::GLBackend::texture_alloc\nblender.exe :0x00007FF6F21F2E20 GPU_texture_create_2d\nblender.exe :0x00007FF6ED267400 DRW_texture_create_2d\nblender.exe :0x00007FF6ED267810 DRW_texture_ensure_fullscreen_2d\nblender.exe :0x00007FF6ED2A4F90 workbench_cache_finish\nblender.exe :0x00007FF6ED25E4D0 DRW_draw_render_loop_ex\nblender.exe :0x00007FF6ED25F5A0 DRW_draw_view\nblender.exe :0x00007FF6EDBC7820 view3d_main_region_draw\nblender.exe :0x00007FF6ED56D940 ED_region_do_draw\nblender.exe :0x00007FF6ED1660C0 wm_draw_window_offscreen\nblender.exe :0x00007FF6ED165F10 wm_draw_window\nblender.exe :0x00007FF6ED165A10 wm_draw_update\nblender.exe :0x00007FF6ED1488F0 ghost_event_proc\nblender.exe :0x00007FF6EDCB1E10 GHOST_CallbackEventConsumer::processEvent\nblender.exe :0x00007FF6EDCFD680 GHOST_EventManager::dispatchEvents\nblender.exe :0x00007FF6EDCFE2E0 GHOST_System::dispatchEvents\nblender.exe :0x00007FF6ED14A980 wm_window_process_events\nblender.exe :0x00007FF6ED140110 WM_main\nblender.exe :0x00007FF6ECDA03C0 main\nblender.exe :0x00007FF6F246FE74 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFF7DCC7020 BaseThreadInitThunk\nntdll.dll :0x00007FFF7E582630 RtlUserThreadStart\n```\n\n----\n\n[blender_system_info_1.txt](blender_system_info_1.txt)\n\n[blender_system_info_2.txt](blender_system_info_2.txt)\n\n[blender_system_info_3.txt](blender_system_info_3.txt)\n\n[blender_debug_output_2.txt](blender_debug_output_2.txt)\n\n[blender_debug_output_1.txt](blender_debug_output_1.txt)\n\n[blender_debug_output_3.txt](blender_debug_output_3.txt)", "Old blend file corrupts to crash blender, I guess\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: AMD Radeon R9 200 Series ATI Technologies Inc. 4.5.14761 Core Profile Context 20.11.2 27.20.14540.4003\n\nBroken: version: 3.0.0\n\nCrashes here and there. RN I witness crush form select/deselect. help plz ;c\n\nI do, tho, have installed two PROBABLY outdated addons BUT in new file crash doesn't seem to happen.\n\nThe problem is - Idk exact steps, but overall it takes to select/deselect all/smth at all few times\n\nSteps\n- Open the provided .blend below\n- Enter edit mode on the main SN4K3Mesh\n- Select-All, De-Select-All (Repeat)\n\n[#94213-asan.txt](T94213-asan.txt)\n\n```\n# Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nbpy.ops.object.editmode_toggle() # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF715993808\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters[0] : 0x0000000000000000\n\tParameters[1] : 0xFFFFFFFFFFFFFFFF\n```\n```\n# Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nbpy.ops.object.editmode_toggle() # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF71599380C\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters[0] : 0x0000000000000000\n\tParameters[1] : 0xFFFFFFFFFFFFFFFF\n```\n\n[SN4K3.blend](SN4K3.blend)\n\n[SN4K3.crash.txt](SN4K3.crash.txt)", "Blender 3.5 and 3.5.1 are crashing while using Metal as backend\nOperating system: macOS 12.6.3\nGraphics card: Radeon Pro 580 8GB\n\nBroken: 3.50, 3.51 (Metall Backend)\nWorked: 3.50, 3.51 (OpenGL Backend)\n\nBlender simply hang. I use eevee, if you switch to rendered viewport shading it crashed complete system. I use Metall as backend, if I choose OpenGL it works fine. \n\n \nI attached an .blend file for example.\n\n" ]

[ "User Preferences Crash\nWindows 10 (1803)\nNvidia Geforce 860m (4gb)\n\n2.80 alpha\n\n\nCrashes / Closes when closing separate window with User Preferences\n\nEdit > User Preferences *Opens separate Window* > closes both windows\n\ni dont think this is wanted", "Crash caused by closing 'Blender Render' preview child windows\nOperating system and graphics card\n$ uname -a\nLinux Inspiron-15-7000 4.10.0-32-generic #36~16.04.1-Ubuntu SMP Wed Aug 9 09:19:02 UTC 2017 x86_64 x86_64 x86_64 GNU/Linux\n\n\n$ lsb_release -a\nNo LSB modules are available.\nDistributor ID:\tUbuntu\nDescription:\tUbuntu 16.04.4 LTS\nRelease:\t16.04\nCodename:\txenial\n\n$ glxinfo -B\nname of display: :0.0\ndisplay: :0 screen: 0\ndirect rendering: Yes\nOpenGL vendor string: NVIDIA Corporation\nOpenGL renderer string: GeForce GTX 1050 Ti/PCIe/SSE2\nOpenGL core profile version string: 4.5.0 NVIDIA 378.13\nOpenGL core profile shading language version string: 4.50 NVIDIA\nOpenGL core profile context flags: (none)\nOpenGL core profile profile mask: core profile\n\nOpenGL version string: 4.5.0 NVIDIA 378.13\nOpenGL shading language version string: 4.50 NVIDIA\nOpenGL context flags: (none)\nOpenGL profile mask: (none)\n\nOpenGL ES profile version string: OpenGL ES 3.2 NVIDIA 378.13\nOpenGL ES profile shading language version string: OpenGL ES GLSL ES 3.20\n\n[system-info.txt](system-info.txt)\n\n\nBroken: blender-2.80-1112f49a14b-linux-glibc219-x86_64\nWorked: ?\n\n\nCrash randonly occured when closing 'Blender Render' preview child windows.\nIt does not occur with 'Cycle' engine selected.\n\n[blender.crash.txt](blender.crash.txt)\n\n1- Open the following attached blend file\n[4200.blend](4200.blend)\n2- Jump to a random keyframe\n3- Hit F12 to render a preview\n4- Wait or not rendering completion\n5- Close the rendering preview child window\n6- Crash may occur, if not, go back to step 2\n\n**Side note**\nIn the given blend file,\nSphere shader coordinate mapping (rotation X) is keyed.\nWith 'Eevee' engine is selected and viewport 'Render preview' enabled, jumping to random keyframe shows us wrong shader previewing. (purpulish material)\n" ]

[ "Animation Editors: Modifier keys cancel out dragging handles \nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.4.0 Alpha\n\nIn recent releases db4313610cab was implemented and the animation editors are still not adapted to this behaviour. This fix allows the use to press any keyboard key, like `Shift` or `X`, etc, before the drag threshold of the current action is reached, and it will cause the operator to execute.\n\nFor example: Click and hold on a move gizmo in object mode. Either drag 3 pixels to start the operation and then press `X` and hold `Shift` to slowly move along the x axis.\nOR immediately press `X` and hold `Shift` without reaching the drag threshold of 3 pixels to already start the operator.\n\nBefore this fix, any pressed keys would cancel the click & drag and prevent the operator from running.\nThis is still the case with the graph and dopesheet editors.\n\n- Create a simple animation to the default cube\n- click and drag on the keyframes or handles in the animation editors\n- Now click and hold them, hold `Shift`, and then continue to drag. Nothing happens.", "GPencil: Shortcuts in Draw Mode are hardcoded\nMultiple keys in the `gpencil_draw_modal` function are hardcoded and (some) don't show up in the info bar. This makes it impossible for the user to know what certain keys do as well as impossible to change them in the keymap. All of those keys should be in the keymap with the default keys set as they are now.\n\nHere is a list of these (hardcoded) shortcuts:\n\n - {key Enter}, {key Pad_Enter}, {key Esc}, {key Space}, {key Ctrl Z} (?): Confirm\n - {key Mousewheel_Up}, {key Pad_Plus}: Increase size of eraser\n - {key Mousewheel_Down}, {key Pad_Minus}: Decrease size of eraser\n - {key O}: Set center of guide\n - {key V}: Turn off speed guide\n - {key M}: Alternate or flip guide direction\n - {key L}: Turn on line guide; {key Ctrl L}: Turn on line guide with angle = 0; {key Alt L}: Turn on line guide with last angle\n - {key C}: Turn on circular guide\n - {key K}: Turn line guide clockwise\n - {key J}: Turn line guide anti-clockwise\n", "Pinning data (mesh, curve, grease pencil, lattice, particle system) makes certain tabs/options disappear (vertex groups, shapekeys facemaps)\nOperating system: Microsoft 10\nGraphics card:GeForce GTX 10606GB\n\nBroken: 2.93\nWorked: don't know\n\n**in the properties editor, in object data properties, the pin toggle hides vertex groups and shapekeys**\n\n- select default cube\n- go to properties editor\n- press object data properties tab\n- press the pin button\n", "Wrong SubSurf modifier results with inconsistent normals\nOperating system: Windows 10 Enterprise Version 1803\nGraphics card: 2 x GTX1080 ti\n\nBroken:\nblender2.8, 2018-12-05\n\nFrom latest builds subsurf modifier make a wrong subdivision of geometry as displayed in the attached files.\nBefore 18/20 November the behaviour of subsurf modifier was ok.\nI am sure because I did an animation on that date and the subdivided vase was ok.\n\nBest regards,\n\nUmberto\n\n\n\n\n\n[ceramic.blend](ceramic.blend)", "Scaling keys in dope sheet is different from graph editor (ignores handles, esp, noticable when scaling negative)\nOperating system:\nGraphics card:\n\nBroken: 3.4.0\n\nSeems like if handles are not auto-clamped then *Dope Sheet* messes things up.\n\n\nIn attached file duplicate and translate both keys further in timeline, also put timeline cursor between them:\n - With *Graph Editor* press **S-1** to reverse new keys -> result is OK\n - **CTRL Z** to undo and repeat reversing in *Dope Sheet* -> bezier handles are processed incorrectly\n[dopesheet_reverse.blend](dopesheet_reverse.blend)\n\n\n", "Mitigate the risk of wrong shape key selection in sculpt and edit mode.\n# The Issue\n\nA major weakness of any layered editing system is that it matters which layer is selected for editing operations, but this selection is not apparent in the actual editing space, because layer effects are by definition combined into a single result for display.\n\nFor instance, unless the single key mode is turned on, the effects of all relative shape keys are summed up into a final shape, which doesn't change no matter which one is selected.\n\nThis means that it is easy to apply editing operations to the wrong selected layer, including a possibility that the undo stack will be exhausted before the mistake is noticed.\n\nAn additional risk factor is that shape influence sliders are located in the same list, and a hurried misclick on a slider can change selection.\n\n# Proposed Solution\n\nThe industry standard way to mitigate this risk is layer locking: layers that are not currently being worked on can be locked by the user, in which case all editing operations that would affect that layer will instead abort with an error.\n\nBlender already supports this with Vertex Groups and weight painting - even though with groups it is more obvious which one is selected, and locking has additional effects related to normalization.\n\nI propose adding this feature for shape keys (implemented in #104463):\n\n\n\nSince it is a standard approach that is used by any advanced image editor, most users would likely be already familiar with the concept. However there are some details in its application that need discussing.\n\n## Scope Issue\n\nThe primary design question is which operations should react to layers being locked, balancing mainly implementation complexity and the effectiveness and convenience of the feature.\n\nIn implementing the prototype I decided that the main criteria should be:\n\n* Operations that affect vertex positions, and thus shape keys.\n\n* Operations where the shape key selection matters:\n\n For example, applying object mode transformation updates all shape keys. Thus the selection doesn't really matter, and respecting locks would inconvenience the user that is working on the scene layout, rather than individual object details.\n\n* Deform-only operations:\n\n Even though selection may affect the final vertex positions, operations that change topology necessarily change all shape keys. For that reason, topology manipulation also belongs in an earlier stage of the design pipeline than shape key creation, even though it may be necessary to do it later sometimes for various reasons. Finally, topology change is seldom not accompanied by using transformation operators that respect the locks.\n\n### Current Scope\n\nThis is the current scope of the checks in the prototype:\n\n* All Edit Mode <u>transformation</u> operations.\n\n Specifically, everything invoking the generic transform system initialization code.\n\n* All <u>Sculpt brushes</u> that affect vertices, and not colors or masks.\n\n Specifically, everything invoking the generic brush stroke initialization code.\n\n* <u>Smooth</u> and <u>Laplacian Smooth</u> operators in Edit Mode.\n* <u>Blend From Shape</u> operator in Edit Mode.\n* <u>Propagate To All Shapes</u> operator in Edit Mode.\n\n This one is a bit of a problem, because technically shape selection doesn't matter here: it snaps all shape keys to the current position. However currently I implemented a check to abort if any shape key is locked on the principle that it is a shape key operation. Maybe it should only check the main Basis, or work in some other way.\n\n* <u>Mirror Shape</u> and deleting shape keys.\n\n* A few other operators...\n\nThis [spreadsheet](edit?usp=sharing) containts an exact listing of which operators should and actually do check for locks.\n\n# Discussion\n\nIt is necessary to discuss whether the main criteria, or this specific list needs changing. For instance, I'm not confident that I remembered all operations that match the criteria for the checks.\n\nAnother minor design point is whether attempting to edit a locked layer should be an error (shows a message under the cursor) or a warning (appears in the status line).\n\nSince the checks are done by operators rather than low-level code, add-ons would have to implement them too where relevant. I wonder how many would be affected, and whether some API additions to help them do that easily would be necessary.\n\nThere have also been some questions raised in #104463, like:\n\n> - Are shape keys a special case? (is there other data we might want to lock).\n> - Is the current implementation disallowing editing in all anticipated situations? There seems to be quite a few ways to transform vertices that aren't included here, applying transform, moving object origins, ... to name a couple, probably other mesh tools too. It may feel like an incomplete feature if these are left out, yet, this means we need to add check in many more places.\n> - Are there ways this can be implemented in a way that doesn't frustrate users. For example, you load someone elses file, select objects and want to apply transform. If they have locked a shepe-key it could mean having to manually select and disable the the locks. Users may request a ways to disable locks on all selected objects, maybe going along in this direction is fine, but we can anticipate some of these issues and consider functionality that doesn't add unnecessary overhead. \n", "I have glitch when i turn on my Motion blur with a shape keys animate\nHello, \n\nOperating system: Darwin-19.4.0-x86_64-i38{[F8688055](Capture_d_e_cran_2020-07-12_a__01.21.11.png)}\n\n{[F8688058](Capture_d_e_cran_2020-07-12_a__01.21.11.png)}6-64bit 64 Bits\nGraphics card: AMD Radeon Pro 560 OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.8.24\n\nBroken: version: 2.90.0 Alpha\n\nI tried with Eevee i don't have the problem only with cycles. \nIs only when i render animation or image. I can't see this glitch on viewport.\nI tried with blender version on mac and window and multiples different graphics card but the problem is still append.\nWith or without modifiers or movement of camera still in. \n\n\n\n\nOpen blender with love.\nMake a beautiful animation with shape keys.\nTurn on motion blur.\nProblem when i render :( \nClose blender but with love to help. :D \n\nThanks you \n\npacpac\n\n{[F8688074](Capture_d_e_cran_2020-07-12_a__13.46.43.png), size=full}\n\n\n\n\n\n", "bevel (Miter Outer) + mirror modifier gives inconsistent result.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51\n\nBroken: version: 2.91.0\n\n\nMirror + Bevel modifier sometimes give some inconsistent result when using the \"Miter Outer\" options.\n\nOpen the file to see the error.\nThank you\n\n\n[bevel.blend](bevel.blend)\n\n\n\n\n\n", "Splitting two modifier keys between two input devices doesn't work\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.5.0\n\n\nPressing modifier keys from two keyboard device doesn't correctly register in blender.\n\n\n- Have two physical \"Keyboard\" device that both have modifier keys\n- Press Ctrl on keyboard 1 and Alt on keyboard 2\n- Blender doesn't recognize that both Ctrl and Alt are pressed.\n\n-----\n\nOriginal description:\n\nHi - I have modifier keys (shift, alt ctrl) assigned on mouse and keyboard and razer keypad. \nI use alt to select edge/face loops without issue. \nIf I press shift+alt on keyboard then I can select as many loops as I like. Same if I press shift+alt both on mouse. \nHowever, if I use modifier keys combined between deviced (shift on keyboard and alt assigned on mouse, or vice versa) then it lets me add only one additional loop to selection only. \nI initially thought it was accessory issue, but Its true with my default keyboard, razer keypad, logitech G600 and Razer Naga mouse, I suspect the issue is within blender. \n\nIn summary combining modifier keys between keyboard/mouse doesn't seem to work, atleast with loop selections. Any advice? Thanks\n", "Modifier are still evaluated even when not required.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 3.0.0\nWorked: never\n\nModifiers are evaluated, even when blank, for example a lattice modifier without any lattice, would still make depth graph calculation or something like that, that make each action (select, move, rotate, etc.) verry slower.\n\nCreate/ import a large object (60M poly mesh for example).\nAdd lattice modifier, let all option blank. \nDisable viewport visibility of this modifier.\nTry to move/rotate/scale the object in the viewport, the performance is really low.\n\nIn order to transform your mesh, you need to remove all modifiers, and add then back once transform is convenient.\n", "Surface Deform modifier always updates when selecting objects\nOperating system: Windows 10\nGraphics card: NVidia GTX1080\n\n2.93.1, 2.93.3\n\nSurface Deform modifier causes click lag even if object or/and the modifier is disabled in viewport. It doesn't affect camera panning or rotation.\nWhile I'm at it, it also seems to keep unnecessary data in saved file. I wanted to provide an example file, but it turned out to be over 1 GB. I don't see why it keeps the deformed mesh in the file rather than recalculating it when we enable it or open the file.\nIt seemed to me to be the best way of making a chainmail, but I can't work with the lag.\n\n1. Subdivide the default cube with 7-10 cuts (7 enough for my PC, if in the end it doesn't lag, try more divisions).\n2. Duplicate the cube.\n3. Add subdivision modifier (level=6) to 2nd Cube, then add Surface Deform, set the source to 1st Cube, bind.\n4. Disable both objects in viewport, you can even disable the modifiers.\n5. Add a new Cube to the scene. Try selecting and deselecting in in the viewport. There's a lag. Camera panning and rotation works fine. The only way to fix the lag is to apply or remove the Surface Deform modifier.\n6. Save the file. Our file is just 396 faces with modifiers applied. It's over 1GB in size.\n\nVideo: hEQnVVvPEGU\n\nor\n- Open .blend file\n- Bind mesh to the target\n- Try to select the Cube.001 in viewport\n[#90985.blend](T90985.blend)\n", "Next/Previous Keyframe breaks when \"Show Only Selected\" is disabled\nOperating system: Linux-5.4.0-91-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.91.03\n\nBroken: version: 3.1.0 Alpha, branch: master, hash: `60a9703de8`\nWorked: never, 2.79 also had this problem in a way (but slightly different due to limitations of its dope sheet). 2.83 has the exact same problem described here.\n\nNext/Previous Keyframe ( {key up} and {key down} in the dope sheet) do not jump to the next/previous key **shown**, but to the next/previous key of the **active object**.\n\nIn other words, it sort of works as if \"Show Only Selected\" is always enabled, but then even more strict, only responding to the keys of the active object. This can probably be tracked down to the dope sheet only showing animation data of the active object back in the old ≤2.79 days.\n\n1. Open the attached blend file.\n2. See that the current frame is set to frame 1.\n3. While hovering with the mouse over the dope sheet, press the up arrow {key up}.\n4. See the message \"No more keyframes to jump to in this direction\". *Expected behaviour:* it would jump to the next keyframe shown, which is at frame 5.\n5. Click on the Cube to make it active + selected.\n6. While hovering with the mouse over the dope sheet, press the up arrow {key up}.\n7. See that it jumps to frame 10, which is the next keyframe in the Cube's animation. *Expected behaviour:* it would jump to the next keyframe shown, which is at frame 5.\n\n\n\n[next-prev-keyframe-selection-issue.blend](next-prev-keyframe-selection-issue.blend)", "Bevel modifier. Incorrect behavior on vertexes where three or more edges with a bevel converge.\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 960/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.1\n\nThe bug is noticeable when the number of segments is greater than 1 and the profile value is greater than 0.5. There is a violation of flatness on convex vertexes. The outer loops of the bevel are correct, but the filling of the angle inside is incorrect.\nThe same bug appears if you use the bevel operator instead of the modifier.\nThe bug is also observed in previous versions of the Blender, at least after 2.93.\nOn concave vertexes, the bug is also present, but less noticeable.\n\nExtrude a cube face, scale it and shift it to form convex and concave vertexes. Add a bevel modifier with 2 or more segments and a profile from about 0.5 to 0.99(9).\n\n", "Using subdiv modifier affects normal baking performace of all objects in scene\nOperating system: Windows 10\nGraphics card: GTX 1080Ti\n\nBroken: version: 2.83 (sub 10)\nWorked: Never (2.8+)\n\n[#60428.blend](T60428.blend)\n- Open file\n- Press bake, watch memory usage\n# Remove subsurf modifier on `door_frame_low.001` object and repeat step 2\n\nBaking is much quicker and doesn't require too much RAM\n", "Modifier re-evaluation makes editing dimensions unreliable\nCurrently editing dimensions uses the bounding box that's recalculated while dragging.\n\nThis can cause a feed-back loop as the objects scale is used by some modifiers which calculate a new bounding based on the scale which is then used in dimensions.\n\nIn the case of armature modifier - any zero scaled axis makes causes the mesh to scale into a single point, making any further changes to the dimensions not have any effect.\n\n----\n\nOriginal report:\n\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GT 620/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\n\nBroken: version: 2.80 (sub 74)\nWorked: Blender 2.79\n\n\nIf object is rigged scrolling its dimension value toward zero causes object to disappear, even if right mouse button has been clicked in order to cancel a scaling. So blender accepts a scale of zero even if left mouse button hasn't been released yet, Is it an expected behavior in Blender 2.8?\n\n\n\n\n - Select the cube.\n - Scroll any of its dimensions value toward zero without releasing the left mouse button\n - Cancel scrolling by right mouse button click.\n\n\n\n[Bug_report.blend](Bug_report.blend)\n\n" ]

[ "some modifier doesn't take shape keys information correctly\nOperating system: Linux\nGraphics card: GeForce GTX 780/PCIe/SSE2\n\nBroken:\nversion: 2.80 (sub 39), branch: blender2.7, commit date: 2019-01-01 15:44, hash: 4dd0a90f4213, type: Release\n\nSome modifiers doesn't update correctly if the mesh has shape keys and the shape keys values are updated.\n\nSome modifiers that don't work correctly : Bevel, Wireframe, EdgeSplit \n\nSome modifiers that work : Subsurf, Triangulate, solidify, SimpleDeform, Screw\n\nI haven't checked every modifier but I can test them all if that can be helpful.\n\nOpen the attached .blend :\n- try to change the value of the shape key, nothing should append.\n- remove or disable the bevel modifier,\n- changing the value of the shape key should change the cube's shape.\n[bevel_and_shapes_keys.blend](bevel_and_shapes_keys.blend)\n" ]

[ "Sculpting laggy and weak from some directions.\nOperating system: Windows 10\nGraphics card: Nvidia GTX 970\n\nBroken: 3.6.1\nWorked:Never(2.80+)\n\nWhen i try to sculpt from from below, using radius unit: scene, sculptings get laggy and the stroke is weak.\n\nI'm trying to sculpt stones for a wall in place.\nThe sculpting works fine if i first move the stone i'm working on to the scene origin. This adds a lot of work as i want to for example put bevels on all objects first, then go on with cracks and other features!\n\nTo reproduce:\nOpen file, sculpt from top and bottom\n\nFile is default sculpt template with sphere moved up by 5 and using scene radius.\n\n", "Curves sculptmode: brush Strength (and Strength Pressure) UI inconsistencies\nOperating system: Linux-5.19.6-200.fc36.x86_64-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.65.01\n\nBroken: version: 3.4.0 Alpha\nWorked: -\n\nSculpt curves: brush Strength (and Strength Pressure) UI has inconsistencies between the Toolsettings in the header and the sidebar\n\nfrom the default startp:\n- add Curves\n- enter sculptmode\n- `Add` brush: does not have `Strength` in the header toolsettings (correct, not supported) yet the Brush Settings in the sidebar shows Strength (wrong)\n- `Delete` brush: same as above\n- `Slide` brush: shows `Pressure Sensitivity Strength` in the header toolsettings (has a bug, see #100324, should be removed)\n- all other brushes: shows `Pressure Sensitivity Strength` in the header toolsettings (correct) yet the Brush Settings in the sidebar dont show it (wrong)\n\n{[F13666040](image.png) size=full}\n\n{[F13666059](image.png) size=full}\n\n", "Dynamic paint sub-steps don't work when the brush position is set in geometry nodes\n\nI believe this **may** be the same issue as T40597\n\nIn order to have a brush follow a potentially very large collection of curves in sequence, I'm using geometry nodes to set the position, which avoids the pain of having to join the curve paths or painstakingly tweak the offsets and influence keyframes for each and every one.\n\nThis works, but with sub-steps enabled it seems that the nodes are still only being updated once/frame. Or, if they are being updated, then the `seconds` input from the `Scene Time` node is being derived from the `frames` value which is clamped to an integer.\n\n\n[text_path_test.blend](text_path_test.blend)\n", "Cursor size for \"Use Cursor for Depth\" in Sculpt doesn't reset properly on undo\nOperating system: MacOS 11.6\nGraphics card: Radeon RX 580 8 GB\n\nBroken: 2.93.4 (2.93.4 2021-09-01)\nBroken: 2.93.5 (2.93.5 2021-10-05)\nBroken: 3.0.0 (3.0.0 2021-10-05)\nWorked: don't know\n\n\nCursor size for \"Use Cursor for Depth\" (Sculpt -> Box/Lasso Trim -> Trim Mode: Join) doesn't reset properly on undo.\n\nBased on the default startup\n\n- File -> New -> Sculpting\n- Select Box Trim (or Lasso Trim)\n- In tool options, select Trim Mode: Join \n- In tool options, select Use Cursor for Depth\n\n- Use tool on the outside of the object (Note that the geometry created has a volume)\n- Undo (if you used the tool more times, repeat the undo each time).\n- Use the tool again. \n\nExpected result:\nTool works normally after all undo.\n\nActual result:\nAfter undoing all tool uses, the tool disregards the cursor size and the next shape created is flat/collapsed. \n\nAdditional notes:\nIt only happens when the file starts in Sculpt mode (hence the New -> Sculpting). If started in Object mode, and then switched to Sculpt, it is not occurring. \nAfter switching tools (even without actually using another tool), it seems to work correctly. \nAlso if switched from Sculpt to Object mode and back, it's working properly. Unless you undo from before switching to the Object mode, and undo all the tool uses, it happens again.", "Performance test and notes regarding multires\nDaniel Bystedt and Dalai Felinto had a catch up meeting 11:th of May. During that meeting we talked a bit about sculpting. Dalai asked me to create this task and to add a test file and some more info which is detailed below.\n\n**Computer and Blender specs**\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 465.89\n\nversion: 2.92.0\n\n\n**Blend file**\n[orc_head_5_subd_v001.blend](orc_head_5_subd_v001.blend)\n\n**Switching symmetry on/off**\nSwitching symmetry on/off is something a user does quite often in order to not make human faces (for example) look totally mirrored when sculpting. It’s a bit surprising that it is so slow when the vertex count is high. \n\n@PabloDobarro explained via chat: *The lag from turning on/off symmetry is because of a BVH rebuild. Symmetry is a property of meshes, so each time a property of the mesh changes, it considers it as a change in geometry and rebuilds the BVH again. It should be possible to do a more proper tagging of these changes to avoid this, which will fix this multires and any other future system we can make.*\n\n|Vertex count| Time to switch symmetry on/off\n| -- | -- |\n|10 million| 3 s 7ms|\n\n**Subdivide performance**\nSubdividing when vertex count will be above 1-3 million is usually rather slow. The user only have to subdivide the model once when sculpting a model so it’s not a frequent thing that you do. If it could be made faster it would be a nice improvement though\n\n|Vertex count| Time subdivide with multires|\n| -- | -- |\n|subdiv 2.6 million > 10 million| 16 s 5 ms|\n\n**Switching sculpt levels performance**\nSwitching sculpt subdivision levels is something the user does quite often, so if it is fast it would be great. The performance is not terrible though.\n\n|Vertex count| Time subdivide with multires|\n| -- | -- |\n|2.6 million > 10 million| 1 s 8 ms|\n\n\n**Performance investigation starting point**\n- Run a profiler to see if we do not have obvious bottlenecks and see that all threads are occupied\n- Measure exact memory footprint of OpenSubdiv topology. If it is not terribly big, can cache those on different levels while being in the sculpt mode\n- Look into sparse patch tables. This seems to be a solution to a specific of OpenSubdiv which is related on a single-threaded nature of topology construction\n- Run quick and dirty check whether using GPU evaluation brings 10x speedup (quick and dirty is because doing proper integration is not trivial and you don't want to do it if the speedup is not that great)\n- Check whether all tools are properly multi-threaded\n- Make ID_RECALC_PARAMETERS to work for non-geometry properties of meshes (to mitigate the symmetry lag)", "Subdivision Surface modifier performance issues\nOperating system: Xubuntu 18.04\nGraphics card: Nvidia Titan X (Pascal)\n\nBroken: 2.80 26d5a3625ed\nWorked: Before the new \"Quality\" setting was introduced in the Subdivision Surface modifier\n\nThe Subdivision surface modifier has slowed down significantly, especially noticable with multiple Subdivision Surface modifiers on the same object.\nCreating the same setup in the builds before the new setting is as fast as 2.79.\n\n- Create Monkey object\n- Add Subdivision Surface modifier\n- Add Wireframe Modifier\n- Add second Subdivision Surface modifer\n[slow-subsurf.blend](slow-subsurf.blend)", "Sculpt mode. Switching back to tool with brushes from tool without brushes rolls to next brush.\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken 2.83.*, 2.91,\nBroken: version: 2.92.0 Alpha\n\nSculpt mode\nCreate 2 grab brushes 1 and 2, that can be switched with G\nSelect Grab 1\nSwitch to mesh filter.\nPress G.\nNow you switched to Grab2 instead Grab1, that was selected before.\nExpected: return to same brush, that was selected before.\n\nIssue appears with:\nLine project, [Box] mask, Box hide, [Box] faceset, [Box] trim, Mesh/Cloth/Color filters, Mask by color, Edit faceset, all transform tools, annotation.\nAffects all tools, who can have few brushes.\nTested with factory startup too.\n\n[2020-12-10_03-13-39.mp4](2020-12-10_03-13-39.mp4)\n[grabroll.blend](grabroll.blend)", "Texture paint undo sometimes changes texture slot\nOperating system: Windows 10\nGraphics card: GTX 1060\n\nBroken: 2.80 (sub 48), 2019-03-12 13:53, hash: 936014ea8c74\n\nWhen texture painting on a particular texture slot, then undoing that painting, sometimes the selected paint slot will change.\n\nI couldn't reproduce from the default startup file, so I've stripped down the file that's having problems as much as I can.\n\n1. Open attached file\n2. Select paint slot \"Skin mask\" from either topbar or properties panel.\n3. Paint some strokes on the model or image editor\n4. Press Ctrl-Z\n5. Notice the paint slot is no longer set to \"Skin mask\"\n\n- - -\n\n- Change to the Skin Mask slot.\n- Paint a single stroke in the image editor.\n- Undo two times.\n- Paint another single stroke in the image editor.\n- Undo one time.\n- Notice that the slot changed.\n[texturepaintbug.blend](texturepaintbug.blend)", "Object transforms affecting Sculpt Mode Brush behaviour\nThis is a design proposal to change some underlying behaviour of the brush system in Sculpt Mode.\n\n**The issue:**\nWhen scaling or rotating an object and then proceeding to sculpt on it in Sculpt Mode, the brush behaviour is generally affected by the transforms of the object.\nSo if the scaling is 1 on X & Y but 3 on Z, the brush radius will be 3 times as high. \nIf the object is scaled down in general, the brush strength will be lower in total.\nIf the object is also rotated, the brush radius will be as well and rake effects like for the Snake Hook brush will be badly affected since the object rotation is added to the rotation of the brush.\n\nIn recent versions the brush radius is at least reflecting the object transforms, which helps spotting these negative influences and reminding the user to apply the object transforms before sculpting.\nBut there is not benefit to having the Sculpt Mode being affected by the object transforms, so this is just a limitation of the current system.\nIt's also good to add that on the other hand Edit Mode is not negatively affected by the transforms of the object.\n\n[2020-03-02 11-42-04.mp4](2020-03-02_11-42-04.mp4)\n\n**The proposal:**\n\nTo make the sculpting workflow easier on the user it would be ideal if the brushes are not affected by the local transforms of the object but are instead based on the world scale & orientation.\nOnly symmetry options should be directly influenced bu the local transforms.", "Mesh box stl sizing issue when exporting\nOperating system: Windows 11\nGraphics card: Intel UHD Graphics 630\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen): 2.9.1 and 3.5.1\nWorked: (newest version of Blender that worked as expected). Unknown\n\nI started a brand new design/project from design. Nothing imported at all and this was a simple design, nothing hard. I've have done others, but those involved mesh cylinder's. For this, i was starting with a mesh cube. I've started from scratch at least 10 times. Before starting I have units set to Metric with measurements set to millimeters. For this issue, if I click on Add -> Mesh -> Cube. Then size it to what I need, in this case between 170-200mm ( aprox 7 inches). Once I size it, I export to an stl file and import into superslicer the cube shows up as a spec of dirt that you can't even see. Even if I import it using imperial units, it's slightly bigger, but only around 2 or 3mm in length.\nWhat is odd is, I do NOT have this problem if I'm dealing with a mesh cone or a Mesh cylinder. I have found a partial work around. When I start a new design from scratch, and my first object is a cylinder, i size that, then add a mesh cube... the sizing works perfectly. Now, once I do that though, if I add a second cube, then it's 100% totally messed up. Size is totally off and not even close. I do not have this behavior with cone or cylinder.\nI also get the same result if I take the original cube that is correctly sized, and I do a Shift+D to create a duplicate cube. \n\nBased on the default startup or an attached .blend file (as simple as possible).\nYou can start with the existing Cube when it starts up, or you can choose File -> New -> General\nsize the cube so it's... 200mm x 200mm or anything like that. Then choose File -> Export - > Stl\nNow, for this project i'm printing on a Rat Rig printer so I'm using Superslicer for that. This is when it shows up like a spec of dust.\nNow, the work around, if you start new, delete the initial existing cube, then click on File -> Add -> Mesh -> Cylinder so that is the first and ONLY object on the screen. Size the cylinder to 300mm\nNow, if you click on File -> Add -> Mesh -> Cube and size that for example 200mmx200mm\nthen export as Stl, and import into SuperSlicer... it shows up EXACTLY the correct size you specify in Blender.\nSo far that is the only work around, but again, i haven't found a way to add a second cube and not mess up the sizing.\nThanks\n\nJim\n\n", "GPencil: Inconsistent brush profile using Line/Shape tool\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 750/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 2.93.0\n\n- When certain brushes with unique characteristics like \"Ink Pen\" or \"Marker Chisel\" brush is selected, and *Draw tool* is used, the brush characteristics (thickness, frequency of effect) are appealing (image: green). \n- But when those brushes are selected, and any *Line/Curve/Shape* tool is used, the brush characteristics have undesired effect (image: red). The peaks are extreme, and the frequency of the extremes along the stroke is high (not as smooth as drawing).\n- (A) The Thickness Profile can reduce the maximum peaks, but not even-out the minimum valleys. (B) The frequency of the brush effect can't be distributed evenly across the stroke. It seems the only way to edit it is through manual sculpting. \n- The number of points (few/many) also doesn't smooth out the effect. \n\n*Also just an additional suggestion, when using line/shape tool in its edit mode (before confirmed with Enter) it would be helpful to be able to adjust its settings (Subdivision, Radius, Strength, Thickness Profile) in real-time to preview the changes - without having to confirm the stroke, undo, change settings, and apply stroke again. Only mouse-wheel scrolling, that controls subdivisions, allows preview of settings changes in real-time.\n\n\n\n", "Workbench refactor causes significant performance regression with workbench shadows\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.75\n\nBroken: version: 2.83.0\nCaused by c476c36e40\n\nWorked: 2.82a\n\n2.83 solid viewport shading work slowly than 2.82 when shadows is on\n\nReduction of performance in viewport very high in heavy scenes as example scene with sculpts 6 millions of triangles\n\n", "Max brush radius is too small for 4k screens\nThe maximum brush radius (500 pixels) is too small when working on 4K monitors. Possible solutions:\n\n - Increase maximum to 1000. Not a generic solution and could make the radius slider less usable (Krita's radius slider suffers from this).\n - Scale brush radius by DPI. Need to investigate if the existing DPI can be used or if we need a separate one for the 3D viewport.\n", "Sculpting: Vector Displacement option not enabled on appended brushes\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.41\n\nBroken: version: 3.5.0\n\n\nWhen sculpting with imported brushes in 3.5 when applying an anchored brush (settings the same as in 3.3.5 LTS), the result looks lifted up from the surface.\n\n**Caused by** disabled \"Vector Displacement\" option when brushes from 3.3 were appended into 3.5, likely a versioning issue.\n\nThe option is added with #104481.\n\n\n\n\nSteps used in both releases:\n- Add Cube\n- Add Multisres modifier\n- Subdivide times 6\n- Apply modifier\n- Ctrl-A, origing to Geometry\n- Append brushes\n- Go to sculptmode\n- Choose a brush\n- Apply brush on object\n\nUsing 3.3.5 The brush is nicely put on the object, in 3.5 (with the same settings) you get a pimple. Lowering strength doesn't do the trick in 3.5.\n\n", "VSE thumbnails improvements\n- [x] Check if reusing strip hash used by rendering job speeds up initialization with large number of strips. if it doesn't, fix it anyway. (b534806ecb) Note: it wasn't initialization that caused freezing, it was rather strip `anim` freeing, reusing ghash seems to have no impact on performance.\n- [x] To speed up loading, Check strategy to render \"guaranteed\" set of thumbnails. Especially if cache can be shared this could help a lot. (3a898db3638d) Note: better strategy seems to be other way around as far as performance goes.\n- [x] Make first (and last) thumbnail reflect actual handle position. (29b9187b327c) Note: Only first image is updated, though it's possilbe to update last one too. \n\n- Change cache so thumbnail (and possibly raw) types are keyed by path to file and disregard various context settings and Sequence pointer. Front-end should stay same. ([D12771](D12771))\n - There are rabbitholes like effect strips producing raw images and scene strips rendering at \"host\" scene resolution so simple string key is probably not enough. \n - Added complexity is, that all cache types per frame should be invalidated at once, otherwise prefetch doesn't work and issues like that.\n - Above points don't quite apply to thumbnails, but there are issues of this type in general. For example if we would implement downscaling for movies, first point will apply.\n\n" ]

[ "Sculpt Performance Regression in 2.81\nOperating system: Windows 10 64 bit\nGraphics card: Nvidia GTX 750 Ti\n\nBroken: blender-2.81 eefd806afc15\nWorked: blender-2.81 aff6446e064f\n\n\nSculpting Performance is noticeably worse.\nEspecially switching brushes lags.\n\n[slow_sculpt.blend](slow_sculpt.blend)\n1. Open attached file.\n2. Apply the modifiers on the mesh to get a dense mesh.\n3. Enter sculpt mode.\n4. Start sculpting and switch between brushes with keyboard shortcuts.\n5. Experience lag with the newer version." ]

[ "Don't show catalog path when dragging assets over root level catalogs\nWe show the name, and the full path of a catalog in parentheses, when dragging assets over it. For root level items this will just result in showing the catalog name twice. This is confusing and should be avoided.\n\n", "Unbaked Dynamic Paint Not Displayed in Rendered Output but Shows in Viewport\nOperating system: Manjaro 18.0.1 Illyria / x86_64 Linux 4.14.90-1-MANJARO\nGraphics card: Nvidia 1080 Ti\n\nBroken: 2.80 (sub 39), Commit date: 2018-12-29 18:10, Hash 266b1e2cdbc1\nWorked: 2.79 (sub 6), Commit date: 2018-10-22 16:17, Hash 86dbbd156fa\n\nUnless baked, dynamic paint does not affect the final rendered image but is properly displayed in viewport.\nThe default and expected behavior is to be able to render the effect regardless if baked or not. This is how 2.7x operated in the past.\nMoreover this problem, which I assume has something to do with vertex colors [has caused other issues ]] with different [[ https:*github.com/hsab/tensionmap | addons ](https:*github.com/JacquesLucke/animation_nodes/issues/989) that depend on similar workflows of modifying the vertex color using the python API (eg. proximity, stress mapping etc.) \nNot sure if #51214 is related since it's from April 2017. If so please close as a duplicate.\n\nTwo blend files attached, one for 2.8 and one for 2.79.\n2.79 file works as expected.\n2.8 file is broken unless baked.\n\n2.79.6 Render:\n\n\n2.80.0 render:\n\n\n[dynamicpaint2.8.blend](dynamicpaint2.8.blend)\n[dynamicpaint2.79.blend](dynamicpaint2.79.blend)", "render lagging behind tiles\nOperating system: win10\nGraphics card:rts 2070 super\n\n2.81\nwhen rendering, the picture doesnt appear for a while after the render tile already finished rendering that region, not sure if its a bug. didn't see anything like this online. here is a video\n\n\n[bug (2).blend](bug__2_.blend)\n\n- Open file and render\n\n[gnVknluHJZ.mp4](gnVknluHJZ.mp4)", "Viewport display: in front should be listed under visibility as it also affects final render\nOperating system: macOS-11.4-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.5.14\n\nBroken: version: 3.0.0 Alpha\nWorked: unknown\n\nUnder object properties, viewport display: in front affects both the visibility in the viewport and in the final render, and as such should probably be listed under visibility rather than viewport display.\n\n-", "Erratic behaviour of materials/material slots for text objects\nOperating system: Linux-5.19.0-45-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce GTX 980M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.108.03\n\nBroken: version: 3.6.0\n\nMaterial assignment for text objects doesn't behave as expected. Removing an unused material slot will cause changes to the materials used by the object. The same happens if you rearrange the material slots.\n\n1. Add a new text object. Leave the default string \"Text\".\n2. Add 3 new material slots with different materials to the object. Use different viewport display colors to be able to see changes.\n(You will now have Material.001, Material.002 and Material.003)\n3. In edit mode, select the last two letters and assign Material.002.\n(You should now have \"Te\" using Material.001 and \"xt\" using Material.002)\n4. In object mode, remove the last material slot containing Material.003.\n(This material was not assigned to any character, and removing it should not cause any change.)\n5. All 4 letters are now using Material.001. \"xt\" should still be using Material.002.\n\nAlternatively, execute step 1-3 above and then.\n4. Move the last material slot up to the first place.\n(This should not cause any change. Normally in blender, the materials used by the geometry will not change if you rearrange the material slots.)\n5. You will now have \"Te\" using Material.004 and \"xt\" using Material.002. \"Te\" should still be using Material.001.\n\nI have attached a blend file where you just need to remove the last (unused) material slot to see the problem, or move the last material slot up to first position.\n\n", "Some particles do not appear in the render when the 'Vector' pass is enabled\nOperating system: Linux-5.19.17-2-MANJARO-x86_64-with-glibc2.36 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.141.03\n\nBroken: 3.2.2 - 3.3.1\nWorked: 2.79\n\nParticles in Render don't match particles in viewport. Some particles do not appear. When they move very fast the issue is more visible.\nIt happens when we enable the pass `Vector`\n\n- Open attached file or:\n - In the default scene (with the Cube), add a Plane\n - Add a ParticleSettings toe the plane\n - In {nav Particle Properties > Render} set `Render As` to `Object`\n - For the `Instance Object` choose the Cube\n - In {nav Particle Properties > Field Weights} set `Gravity` to `0.0`\n - Optionally bake for cache\n - Set `Render Engine` to `Cycles`\n - In {nav View Layer Properties > Passes > Data} enable `Vector`\n - Move Cube out of camera view\n- Choose a frame in the middle and render\nNote that the particles closest to the plane disappear\n\n[particles displaced.blend](particles_displaced.blend)\n\n", "Transparent Object intersecting shows artefacts in Render, but not in VP preview render\nOperating system: Win10\nGraphics card: GeForce980\n\n\nBroken: 2.93, 2.93, only in Cycles\nNote this was always visible, but there seems to be a change between 0f61b27e93 (where this was a lot less noticable) and then e14894aad2 (which already shows the current behavior).\n\n2 Objects imported from a CAD via .PLY, they intersect slightly. Gave one of them a fully transparent material, it shows up in render in the form of shadows (or dark reflections?) on the other one. Solving the intersection, disabling ray visibility for camera both solves this, but it still feels like a bug.\n\n\nOpen .blend, press render. Move up solid2_solid2, disable its ray visibility or use vp preview render for comparison.\nNormal Render (object is fully transparent)\n\nobject hidden for camera (still some artefacts)\n\nobject excluded from render - no artefacts\n\n\n[OverlapBug.blend](OverlapBug.blend)", "Material Library VX causes permanent freeze with EEVEE rendering\nOperating system: Windows 10, 64 bit\nGraphics card: Intel HD Graphics\n\nBroken: 2.81\n\nUsing materials from Material Library VX addon, particularly Floor Wood and Rusty Metal, causes EEVEE to freeze all of Blender. They work fine with Cycles, so it's not the addon. Not even the partial \"developer view\" render can take the materials.\n\nOpen default blend file.\nEnable Material Library VX addon.\nApply \"Floor Wood\" to the cube.\nSwitch to \"render view\" without changing the engine off of EEVEE. Blender 2.81 will freeze.\n\nNote: If you try to view the material preview and EEVEE is the render engine even *that* will freeze Blender 2.81.", "\"Material preview\" shows always eevee mat branch / never cycles\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.82 (sub 6)\nWorked: never, I think\n\n\"Material Preview\" shows always the material output assigned to eevee, but I want the cycles branch. Changing renderengine or the active material output node doesn't change anything.\n\nOpen the blendfile and try to preview the material with \"cycles\" image - without going to rendered viewport. The material preview seems to always use the node tree going to the Eevee material output node, even when the render enging is set to Cycles.\n[specific_material_output.blend](specific_material_output.blend)\n\n\n", "Workbench: Random object color disables transparency in solid view\nOperating system: Windows 10 64bit\nGraphics card: Nvidia RTX 2080ti\n\nBroken: 2.83.1 official\n\n\n\n\nWhen setting the color of the material in the viewport to be semitransparent (Alpha value = 0.1) it works if the color of the viewport objects is set to \"Material\" it works, while if the color is set to \"Random\" the object is not semitransparent anymore.\n\nBlend file to test:\n[bug.blend](bug.blend)", "Tooltips doesn't display when play animation\nOperating system: Linux-4.15.0-64-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\nBroken: version: 2.81 (sub 12)\nWorked: 2.8\n\nTooltips doesn't display when play animation, work otherwise\n\nDefault startup:\n \n* Play animation\n* Tooltips doesn't show \n\n", "Workbench shadows disappear and reappear depending on what's selected.\nOperating system: Linux-6.5.3-arch1-1-x86_64-with-glibc2.38 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 535.104.05\n\nBroken: version: 4.0.0 Alpha\n\nWorkbench shadows disappear and reappear depending on what's selected. Makes no real sense as to why. This behaviour seems odd and I don't know if it was ever supposed to be like this.\n\n\n1. Open a new default scene and raise the default cube to sit on the floor.\n2. Add a mesh plane and scale it up to create a floor underneath it to display the cube's shadow.\n3. Change the render engine to Workbench and activate shadows in viewport shading popover menu.\n4. Optionally, toggle shadows in workbench render properties, as well.\n5. Select cube and watch the shadow disappear.\n6. Select the plane and watch the shadow reappear.\n\n", "Add Asset Browser to Animation Workspace?\nShould the Asset Browser, or at least the 3D View Sidebar with the pose library asset view be added to the default *Animation* workspace?", "Viewport shading solid mode not displaying geometry nodes color attributes correctly\nOperating system: Win10\nGraphics card: 960m\n\nBroken: 3.5\nWorked: never\n\n\nWhen using geometry nodes to add color attributes, there are several situations where the viewport solid mode attribute option does not display the color attributes correctly.\n\nAs can be seen by this example. The first image shows a cube with a color attribute added. The geometry nodes setup adds a red color to the \"Color\" attribute. This all works as expected as can be seen in the viewport and the spreadsheet. \n\n\n\nIf a GN cube is added instead of the standard mesh cube (which has been bypassed by the switch node) then the color attribute is added as expected. This can be seen in the spreadsheet as to be the same as the previous screenshot for the color attribute. However, it is not being displayed in the viewport.\n\n\n\nThis is a problem seen with any procedural geometry made with just nodes. As a workaround, you can have a single vertex as the only \"real\" geometry and use a join node to join with any nodal procedural geometry. Strangely though, this has to be the first connection in the join node, and won't work if it is second. This can be seen by the next 2 images:\n\n\n\n\n\nA similar related case is that some mesh nodes will also make the viewport lose the displayed attribute. For instance, if you create a cube, add a GN modifier and add a dual mesh node to the chain, the color attribute is still there, but no longer displayed in the viewport. This happens with some nodes (like mesh to volume and back to mesh) but not others (like subdivide or extrude).\n\n\n\nThere is also a related issue that when exporting to obj or gltf the color attributes are not exported if they are not displayed in the viewport. If apply the GN modfier, you have to actually click on the color attribute in the properties panel to see it in the viewport or to be able to export it. \n\nOddly, the fbx exporter seems fine without applying the modifier. It also doesn't matter if the color attribute is on the vertices, or face corners, or if the color is float or byte color. The same issues happen regardless.\n\n", "Broken textures and shading\nOperating system: Linux-5.15.0-40-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce MX130/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.73.05\n\nBroken: version: 3.2.0\n\nBroken textures when starting up blender and loading object textures and shading issues\n\n[Blender Texture Bug.mp4](Blender_Texture_Bug.mp4)\n\nPossible causes:\n * Not loading the parent asset blend file before the asset in another scene (opening it first seems to fix the issue in my testing) \n * 4k textures (i've heavily used 4k textures in the asset)\n * Viewport Texture size limit setting (Changed it to 2k)\n\n\n[untitled.blend](untitled.blend)\n" ]

[ "New Materials doesn't appear collada preview\nMacOS 10.14.3\n2017\n27\"\n5K\n\nPlease do use Macintosh. Do not use Windows.\n\nBlender 2.80 latest version February 1, 2019\n\nCreate new materials and work on texture I show you picture.\ninYexxa \n\nExport doesn't appear preview for Mac.\nL2uON8e\n\nMaterials need to fix bug support collada.dae should appear preview on Mac. Blender 2.79b it work as well but 2.80 some bug.\n\n[test.blend](test.blend)\n\n[Test.dae](Test.dae)" ]

[ "Cycles: Persistent Data option corrupt geometry nodes instance meshes with changed normals\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.6.1\n\nThe instance mesh changes if render cycles use persistent data.\n\n1. Open attached file.\n2. Render few frame.\n\n", "Bone Transformation - Make Individual Origins work with groups of linked bones\n**Short description**\nWhen the pivot point is set to Individual Origins, and multiple edit bones are selected, when rotating they are expected that each group of linked bones has its own rotation pivot.\nBut instead these bones pivot from the head of their parent, which seems unintentional, considering this works as I would expect in pose mode.\n\n[UncommonGlaringGilamonster.webm](UncommonGlaringGilamonster.webm)\n\nIn the video, is expected Individual Origins to behave the same way as Active Element, since in the case of a single selected(therefore the active) element, those two do the same thing in other cases in Blender.", "Wrong UV after Vertex Slide or moving mesh with Correct Face Attributes\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.5.1\nWorked: unknown\n\nAfter Vertex Slide with \"Correct UV\" checked, or after moving mesh with \"Correct Face Attributes\" UV mapping is wrong. Vertices in UV get split and wrongly offset.\n\n\nExample 1\n1) Open attached Blend file\n[wrong_uv.blend](attachment)\n2) Locate objects plane01_small and plane01_larger. They are identical models with the same UVs. Just one is slightly smaller and one larger and they have different position.\n\n3) Now we will be vertex sliding this vertex to the diagonal one with snapping on vertex on (see picture) with \"Correct UV\" checked. See that the larger model has UV correct, but the smaller one has wrong UV after the vertex slide\n\n4) now go back (Undo) and lets try grab/move with snapping instead of vertex slide. Be sure to have checked ON \"Correct Face Attributes\" and \"Keep Connected\" in Tool Settings/Option - see picture. Now UVs are wrong in both cases\n\n\nExample 2:\n1) Lets look at plane02_01 and plane02_02. They are totally identical duplicated objects (not instances) just with different position.\n\n2) Vertex slide works OK in both cases here, so lets try Grab/move with vertex snapping and \"Correct Face Attributes\" + \"Keep Connected\" ON again. Here result UVs are OK in case of plane01_02 but wrong in case of plane01_01\n\n\n", "Changes to a PropertyGroup in the preferences of an addon do not trigger the flag indicating changes to save\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.81\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\nUser preferences and add-on preferences are normally auto-saved when Blender exits, and it is triggered when the user modifies preferences data through the UI.\nBut there appears to be a bug that prevents this from happening when the preference data is defined inside of a property group.\n[2020-01-08 16-39-50.mp4](2020-01-08_16-39-50.mp4)\n\n- Install the simplified addon attached - (simple script that adds/draws two preference variables for an add-on: One directly inside of a class derived from `bpy.types.AddonPreferences`, and one inside of a class derived from `bpy.types.PropertyGroup`)\n- In the addon UI, modify any the group preferences - (neither value will be saved on exit)\n- In the addon UI, modify the root preference - (dirty flag or auto-save will be triggered)\nA similar issue occurs when preference values are programmatically altered (auto-save is not triggered), but I'm not sure if this is a bug, or lack of a feature.\n[user_pref_addon.py](user_pref_addon.py)", "vertex group re-appears, after it was previously removed via the API\nOperating system: Linux-4.15.0-175-generic-x86_64-with-glibc2.27 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.103.01\n\nBroken: version: 3.3.1\n\nI have an object with 2 bevel modifiers. Each one uses it's own vertex group to control the beveling.\nI have a script that applies the first mod in the stack, and then removes its vertex group as well.\nIf you run this script 2 times, both mods, and both vertex groups should be removed. \nCuriously, while the modifiers are both removed successfully, the first vgroup will reappear after the second run of the script, so after the 2nd mod and vgroup have been removed.\n\nTo workaround this, you can toggle into edit mode and back into object mode at the beginning of the script.\n\nPossibly related to the commits mentioned here: T93896 ? \n\n[video demo ](BNwfPiyczdg)\n[bevel_vgroups.blend](bevel_vgroups.blend)\n\n* open the attached blend file\n* verify how each bevel mod is related to a vertex group sharing its name\n* run the script two times, and notice how the first vgroup reappears at the end\n\n", "NLA is editable even when the animation data is linked\nBroken: 2.80 to current main branch (`96abaae9ac57e3833f862bc0b562d2fe7a95f702`)\nWorked: probably never\n\nThe NLA is editable, even when the animation data is linked and thus should be read-only.\n\n\n- Extract the attached ZIP\n- Open `nla-link-test.blend`\n- Move the strip in the NLA\n- Save and reload the file\n- See that the strip edit has not been saved to disk.\n\nObserved behaviour: the NLA strip can be moved.\n\nExpected behaviour: the NLA data of the linked object should be read-only. Just like the cube itself (you cannot move it).\n\n", "Eevee Group Node ignores socket type conversion (eg. Vector to Float)\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 527.56\n\nBroken: version: 3.5.0 Alpha, branch: main\nWorked: N/A\n\nIf the input of a group node is `Float`, but you connect a `Vector`, the group will read the input as a vector instead of a float.\n\n1. Download the Blender file below\n[test.blend](test.blend)\n2. Switch to Eevee, it looks like this:\n\n3. Switch to Cycles, it looks like this:\n\n\nIt seems to be node group evaluation. When ungrouped it looks identical.", "Mismatch between nodetrees when creating a link does not throw error\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.41\n\nBroken: version: 3.6.0 Beta\nWorked: Tested as far back as 3.3.4 and no error was thrown\n\nWhen creating a node-link an error is not thrown if the node trees of `links`, `input socket`, and `output socket` are not all the same node tree or even the same type! It makes troubleshooting code sometimes difficult when no errors are thrown but with unexpected results.\n\nPaste the following into the text editor in Blender and Run it:\n\n```py\nimport bpy\n\nD = bpy.data\n\nnt1 = D.node_groups.new(\"NT1\", \"GeometryNodeTree\")\nnt2 = D.node_groups.new(\"NT2\", \"ShaderNodeTree\")\nnt3 = D.node_groups.new(\"NT3\", \"CompositorNodeTree\")\n\nnd1 = nt1.nodes.new(\"NodeReroute\")\nnd2 = nt2.nodes.new(\"NodeReroute\")\n\nnt3.links.new(nd1.outputs[0], nd2.inputs[0])\n```\n\n", "Design discussion: Add object instancing to support transforming linked objects?\nThis is an initial task to start discussion on this topic. More details would have to be looked into for an actual design.\n\n### Motivation\n\nLinking objects in Blender has the inherent design limitation that it also links the object's transform, which means you can't transform an object that was linked. This is a very basic and important operation. With library overrides, that issue is mitigated; however IMO we should look at this as a feature for production use-cases, not as something for more informal, low-overhead usages. This would be a complex solution for the user to manage for a simple problem. They need rather deep understanding of how Blender data-management works to manage library overrides.\n\nEspecially for the Asset Browser, it would be good to have a simple solution that behaves as users expect. Apparently what they do expect is that they can simply link objects, while being able to transform them still. Not being able to edit other object properties seems fine.\n\n### Possible Solutions\n\nFor collections we have actually addressed this issue by supporting linked collection instances. That means, while the collection and its contents are linked, the collection is placed relative to an emtpy that is local to this file. So by transforming the empty the entire collection is transformed. The idea of instancing can be brought to objects so transforming linked objects becomes possible in effect.\n\nSome ideas on how object instancing could be supported on linking:\n* **Object is made local, the mesh, materials, etc stay linked** Simple to implement - downside is that object properties, modifiers, constraints, physics and the like would not be linked anymore.\n* **Instancing via empty.** This would be very similar to collections. An empty can instance an object which means it acts as a parent of the object.\n* **Wrap into collection instance.** Effectively the same as instancing via an empty, but the user would see an intermediate collection. It's simple to add this, but kind of forces an implementation detail onto users.\n\n### Other Approaches\n\nInstead of trying to support object instancing, other things could be done on linking:\n* **Automatically create library overrides for transforms**. Another simple solution technically and at least saves the user from having to know about or setup the library overrides when linking. But obviously pushes users towards using library overrides, which they sooner or later will still have to deal with.", "Geometry Nodes: Curve primitives issues with object material\nBroken: version: 3.2.0 Beta, 3.3.0 Alpha\nWorked: n/a\n\nMaterials assign to the object don't work in some cases for curve primitives created in the geometry nodes tree.\n\nOpen file, you will see that the star doesn't get the red object material.\n\n[curve_object_material.blend](curve_object_material.blend)\n\nUnhiding any of the hidden nodes (Grid or Set Material) will make the object material show up.", "Some particles do not appear in the render when the 'Vector' pass is enabled\nOperating system: Linux-5.19.17-2-MANJARO-x86_64-with-glibc2.36 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.141.03\n\nBroken: 3.2.2 - 3.3.1\nWorked: 2.79\n\nParticles in Render don't match particles in viewport. Some particles do not appear. When they move very fast the issue is more visible.\nIt happens when we enable the pass `Vector`\n\n- Open attached file or:\n - In the default scene (with the Cube), add a Plane\n - Add a ParticleSettings toe the plane\n - In {nav Particle Properties > Render} set `Render As` to `Object`\n - For the `Instance Object` choose the Cube\n - In {nav Particle Properties > Field Weights} set `Gravity` to `0.0`\n - Optionally bake for cache\n - Set `Render Engine` to `Cycles`\n - In {nav View Layer Properties > Passes > Data} enable `Vector`\n - Move Cube out of camera view\n- Choose a frame in the middle and render\nNote that the particles closest to the plane disappear\n\n[particles displaced.blend](particles_displaced.blend)\n\n", "Redo Panel changes the state of the last active Vertex Group\nOperating system: macOS-12.4-arm64-arm-64bit 64 Bits\nGraphics card: Apple M1 Apple 4.1 Metal - 76.3\n\nBroken: version: 3.4.0 Alpha\n\nActivating a certain Vertex Group before calling the Smooth Vertex Weights (or another weight paint operator) won't necessarly respect the choices even when the Subset Active Group is enabled. As soon as you interact with the pop-up operator's parameters seems Blender will pick up another Vertex Group instead of the current activated one.\n\nVideo:\n[WightPaint-SmoothOperator-Bug.mp4](WightPaint-SmoothOperator-Bug.mp4)\n\nSample file:\n[Smooth-Operator_Bug.blend](Smooth-Operator_Bug.blend)\n\n\n1. Paint some vertex groups on your mesh.\n2. Click on one of your groups to set it active.\n3. Call the Operator from Weight Paint > Weights > Smooth\n4. Set operator's Submenu to Active Group\n\n*Note that is might switches to another Vertex Group as soon as you interact with the pop-up's operator parameters.\n_ \nWorkaround: after setting the desired Vertex Group active, you must perform another action like clicking with you brush on an empty area of the 3D view. This way it will respect the active Vertex Group selected.\nIt seems some kind of residue living in the memory of undo buffer. I don't know... could you please investigate?\n\nThanks you!\n\n", "Including node group attribute into NodeTreePath\nAt this moment `SpaceNodeEditorPath` has method append with parameters `node_tree` and `node`. This method is adding `NodeTreePath` into path. But event if `node` parameter was added into `append` method the `NodeTreePath` object does not have any information about root node of group tree.\n[bpy.types.SpaceNodeEditorPath](bpy.types.SpaceNodeEditorPath.html#bpy.types.SpaceNodeEditorPath.append), [bpy.types.NodeTreePath](bpy.types.NodeTreePath.html)\n\n\n\nAccording the fact that one group tree can be used by many group nodes it quite important to have information about via which group node the group tree is eddited. Sverchok addon is used to color problem nodes. Without information about group nodes in path it is impossible to mark such nodes.\n\nI would not bother if there was any workaround. What I actually was trying:\n```\nnode_tree_path.name = 'group_node'\nAttributeError: 'NodeTreePath' object has no attribute 'name'\n\nbpy.types.NodeTreePath.node = bpy.props.StringProperty()\nnode_tree_path.node = 'name'\nAttributeError: 'NodeTreePath' object attribute 'node' is read-only\n\nnode_tree_path['node_name'] = 'name'\nTypeError: bpy_struct[key] = val: id properties not supported for this type\n\narea_names = {}\narea_names[node_tree_path] = 'my_name'\n```\nEven last example would not work for several reasons. ID of path can be changed on undo event for example. If area is splitting into two I'm not able to update second area in the dictionary.\n\nI think, for API consistency, either `node` parameter should be removed out from the `update` method or `node` attribute should be added to `NodeTreePath` object. I would prefer second choice.", "Node Tools Overview\n## Background\nWith node group assets becoming more prevalent and geometry nodes becoming essential\nfor manipulating geometry in Blender, we want to add the ability to implement operators\nwith geometry node groups.\n\nRelevant node groups would need a geometry input and output, and could have other inputs (and possibly outputs), with similar mappings as the geometry nodes modifier.\n\nAdditionally, specific inputs could be marked as editable modally, for example, a \"Factor\" input could be controlled by moving the mouse cursor left or right.\n\nThe specific data design still needs to be settled on (i.e. is the node group an operator, or does an operator reference a node group?).\n\n## Design\n- The node group itself is the operator asset\n- Geometry nodes gets new input nodes\n\t- Selection\n\t- Mouse coordinates\n - View transform\n - 3D cursor location\n - etc.\n- Geometry node groups with these input nodes can only be used as operators\n\t- This design is similar to the shared node groups between shader and geometry nodes, for example.\n- The group outputs are used to adjust the output geometry\n\t- The geometry output will adjust the \"active\" geometry\n - We _might_ need to store a specific mapping for outputs, but relying on the name is a start\n- The project is \"completely finished\" when any existing edit mode operator could be implemented with nodes\n\n## Implementation\n\n_Organized by priority in each category_\n\n### 4.0 Targets\n\n_Changes needed to move the feature out of experimental and enable it in the release_\n\n- [x] Initial basic feature for curves sculpt mode e4cc91a6113f14418b5d8f2a0aace252dc3949d8\n- [x] Allow controlling which modes assets are available in fd9d22ba57d75b278a0350492c5eab2d39d39070\n- [x] Update 3D view menus dynamically (avoid need to restart Blender) cc83951951efc28411960d950435224e8d06606c\n- [ ] Make object info node geometry output work properly\n- [x] #111529 Show menus for unassigned node group assets d2d4de8c710c4e8e648a86f904e6436133664836\n- [x] Tool specific nodes\n - [x] Selection input e3e6fb8ecf8252344e9bfc9dc255bf7f722aa5f0\n - [x] Set selection e3e6fb8ecf8252344e9bfc9dc255bf7f722aa5f0\n - [x] 3D cursor e3e6fb8ecf8252344e9bfc9dc255bf7f722aa5f0\n - [x] Sculpt face sets input e3e6fb8ecf8252344e9bfc9dc255bf7f722aa5f0\n- [ ] Node group inputs displayed in the redo panel\n - [x] Initial input drawing bc393c403325c956ec4e83eeb072a9bdb2e7824d\n - [x] Display node group name in panel title f0467b4615b366f34f0daae81ef717c342e8f520 ee1ee584ded3b76fdd9b661244274af850b08fa9\n - [ ] Add attribute search\n - [x] Don't draw redo panel when there are no inputs c59d998ce524e37c2449694602fb6d57b6d292fa\n- [x] #111523 Improved default traits b4dc16f32d92a2fbc140abfe6eb438ff11b5d5e8\n- [ ] #111524 Finalize trait popover locations\n- [ ] Add support for evaluating output attributes\n- [x] Shortcut support 65591b2c8cc83b024bc0a283bae1bbeb7ee85132\n- [x] Quick favorite support 65591b2c8cc83b024bc0a283bae1bbeb7ee85132\n- [ ] #111526 Filter data-block selector based on tool/modifier option #112389\n- [ ] Publish the user manual\n\n### Polishing\n\n_Not necessarily required for a release in main_\n\n- [ ] Show visual difference for tool nodes in modifier context\n- [ ] Show visual difference for viewer and simulation nodes in tool context\n- [ ] Resolve flickering in asset browser when using redo panel\n- [ ] Warnings in node editor header\n - [ ] When node group is not an asset (with button to mark as asset) #110813\n - [ ] When the node group is not marked as a tool #110813\n - [ ] When asset is not in a catalog (and not in a menu) #110813\n\n### Milestone 2\n\n_The second phase of the project, after the fundamentals are working well_\n\n- [ ] Option to show settings in popup before operation starts\n- [ ] Tool specific nodes\n - [ ] Sculpt \"soft selection\" input and set\n - [ ] View transform\n - [ ] Mouse position\n - [ ] Deformed (\"cage\") positions\n- [ ] Modal operator support\n- [ ] \"Wait for cursor\" option\n- [ ] Store operator properties for the next execution\n\n\n## Use Cases and Ideas\n- [ ] Mesh sculpt \"face set extract\"\n- [ ] Mesh sculpt \"mask extract\"\n- [ ] Lasso selection operator", "Unable to fix broken linked IDs when renamed in libfile\nWindows 10\nWindows 7 x64\n\nWorked: 2.82\nBroken: 2.83.0\n\nCant rename linked collection. Error: Cannot edit external library data\n\nDefault Blender setup (first file lets call it assets.blend)\nCreate a collection and name it \"Test\" \nCreate a cube and move it to \"Test\" collection\nSave the file\n\nCreate another file (lets call it level.blend)\nLink the \"test\" collection from the assets.blend\nCheck all linking options before hiting link\nSave the file\n\nNow open assets.blend again and rename \"Test\" collection to \"Cube\"\nAnd save it again.\n\nOpen level.blend\nA message appears that link is broken\nGo to Outliner and change Display mode to Blender File\nNow go down to the linked collection\nDouble click broken collection to rename it\nError message appears \"Cannot edit external library data\"\n\nRenaming linked collections works fine in 2.82\n" ]

[ "Vertex group information corrupt when changing mesh data for an object with vertex groups\nnot applicable\n\nBroken: (master, blender 2.8)\nremark: I have not tested in 2.8 but christian_ told me in IRC that blender 2.8 crashes in this situation.\n\n\nAssign the mesh data of an object that has vertex groups assigned to an object without vertex groups or with different vertex groups. Then the vertex group information becomes a total mess\n\n\nSimple way to reproduce:\n\n- create 2 cubes A and B\n- for cube A define a couple of vertex groups and assign vertex group values to the vertices\n- then select cube B and change its mesh data to the Cuba A Mesh\n\nNow inspect the bpy.context.object.data.vertices- [ ].groups There you find group assignments but they all point to nowhere now.\n\nSee for example in the attached blend file [vgroup_bug.blend](vgroup_bug.blend)" ]

[ "Missing overlays in the viewport when viewing through transparent shader\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.6.1\n\nwhen viewing an LGT-shadow_caster.004 object, all overlays disappear from a certain side of it\n\n\nSee LGT-shadow_caster.004 in blender 3.6 splash screen\n\n", "The viewport lock modes are super glitch (in general) with zooms and centering\nOperating system: Windows 10\n\n\nBroken: blender-3.0.0-alpha+master.eae4e2251839-windows.amd64-release\n\n\n\nThe viewport lock modes like 3d cursor or object locking is super glitchy in general which creates a disorienting experience. The main issue is sudden jumps in zooms also sometimes the center or rotation becomes jumpy as well. Some jumps in the video is worse than the others. Bear in mind that I am not zooming intentionally when there is a sudden jump.\n\nThis gets even worse in different scene scales and scene layouts. Here I am demonstrating it only with the cursor but this erratic behavior happens with object locking too. This becomes even worse with a Wacom tablet since it is harder to keep the pen movement as static as the mouse while zooming.\n\n\n- Load the attached .blend file\n- Make sure the viewlock is set to the cursor, and have similar preferences as shown in the pictures\n- Try repositioning the cursor, just move the cursor to different objects.\n- Move the cursor to the surface of another object then try ctrl+single middle click to activate the zoom\n\n\n\n\n\n\n\nTest File:\n[glitchy_lock_modes_28082021_2053_04.blend](glitchy_lock_modes_28082021_2053_04.blend)\n\n[A7ysxb6GVc.mp4](A7ysxb6GVc.mp4)", "Invisible object in preview mode\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: AMD Radeon HD 8600/8700M ATI Technologies Inc. 4.5.14831 Core Profile Context 21.5.2 27.20.20903.8001\n\nBroken: version: 3.1.0\n\nWhen I enable Material Preview mode in Blender, the object becomes invisible regardless of the Render Engine. Also the object is invisible with Rendered mode enabled when I use the Render Engine - Eevee. \n\nI don't know how it happened but I can't repair Blender. I reset all settings to factory, reinstalled the application, but it did not help. Can you help me, please?\n\n\n\n\n\n\n\n\n\n", "Regression: broken normal shading, sharp edges appear on smooth surfaces.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: 2.93, version: 3.0.0 Alpha\nWorked: 2.83.9, 2,92\n\nCaused by 1c22b551d0\nTo fix this, 5c4d24e1fd got reverted\nProblem still persists though...\n\n\nIn newer versions normal mapping looks broken, sharp edges appear.\nEevee and Cycles have same bug.\n\nupdated: [untitled.blend](untitled.blend)\n\ndifference in shading\n**2.93 on top, 2.8 down below:**\n\n\nThis effect appears even with very small angle of clean surface (without normal maps)\n\n\n\n<video src=\"2021-05-18_16-17-05.mp4\" title=\"F10122604/2021-05-18_16-17-05.mp4\" controls></video>", "Overlay: Edit-Mode: Anti-Alliasing not working with constructive modifiers\n```\nOperating system: Linux-4.15.0-50-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce 610M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\n\nBroken: version: 2.80 (sub 69)\nWorked: (optional)\n\n```\n\nthe bug is really intresting so... when you add mirror modifier to object.. and u enter edit mode and u select face or edges etc... the AA goes crazy it is so pixelised like even worse than tottaly without AA.\n\n\n\n\n\n\n\n", "Edge overlay seems to do nothing expected\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.23\n\nBroken: version: 3.1.0\nWorked: 2.79\n\nBack in 2.79, we could disable the \"edge\" overlay in edit mode. This was useful I.E. in cases where highlighting edges in vertex selection mode made things hard to read :\n\n\nIn the current stable version, this option seems to do nothing:\n\n\n\nActually, when in solid mode, it seems to display another layer of wireframe on top of the already existing one, which definitely looks weird and pointless. GIF demo:\n\n\n\nOn a mesh in edit mode, try toggling the edge overlay on and off.\n\n", "Mantaflow - Flickering Square pattern of the particles\nOperating system: Windows 10\nGraphics card: Quadro RTX 4000\nCPU: Ryzen Threadripper 2950x\nRAM: 64GB\n\nBroken: (2.90.0 // 2.83.5)\n\n\nI got the Problem both in the 2.83 LTS & 2.9 Version:\nIf i bake the fluid simulation in Mantaflow the particles of Bubble & Foam result in flickering squares: In one frame the particles are present and in the next frame they are abscent (in a square pattern). If I render an animation sequence out of this the video is flickering. I added the .blend-File for the used Settings and also two screenshots to demonstrate the square Pattern. Furthermore a rendered animation of another scene of mine shows the flickering result when rendered. \n\n\n[Mantaflow Flickering.avi](Mantaflow_Flickering.avi)\n\n[2020_09_19 - Fluid Test.blend](2020_09_19_-_Fluid_Test.blend)\n\n\n\n\n\nJust bake the attached .blend File (You need to change the saving location)", "Mesh selection in viewport broken with scenes using a high number of models\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.02\n\nBroken: version: 3.5.0 Alpha\nWorked: Never (2.8+)\nWe are working on a commercial project with imported SketchUp scenes using 55K+ models and running more and more into broken scenes with selection issues in the viewport.\n\n`What we figured out so far:`\n- Selection in the outliner works\n- Some models can be selected by a point and click\n- Other models can't be selected by a point and click\n- Models at a different location will be selected by a point and click and not at the mouse pointer location\n- Box select is messed up and doesn't work overall (selects just parts of the model or fails completely)\n- Lasso select seems to work as intended\n- Circle select seems to work as intended (but can freeze Blender)\n- Some random models can cause a hard crash of Blender when selecting them by a point and click\n**When we delete all my other scenes in the Blender file and purge the data the selection modes work again.**\n\n\n\n\n[T104008_Selection_Bug_350_v006_CUBES_REPORT.blend](T104008_Selection_Bug_350_v006_CUBES_REPORT.blend)\n\nTry to select cubes.\n\n", "Crash with certain file. Sculpt mode. Vertex color fill + edge.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.93.0 Alpha\n\nWhen I try to apply vertex color Fill (Sculpt mode) on edge in this particular file, I get crash\nAnd all fine for shapes/triangle.\n\n[VC + edge crush.blend](VC___edge_crush.blend)\n\n[debug_logs.7z](debug_logs.7z)\n", "Metal Viewport flickering on interaction\nOperating system: Mac OS Monterey, Version 12.6.2\nGraphics card: Apple M1 Max, Mac Book Pro 16 inch 64 GB, LG Ultra Fine Display as secondary monitor\n\nBroken: 3.5.1 release candidate\nWorked: 3.4 with Open GL does not have these issues\n\n\nWhen interacting in the viewport say moving a camera the viewport flickers. Also when adjusting shaders in a cycles window the viewport also flickers depending on the speed of the interaction. Evee viewport does not have this issue, but solid render does\n\nBased on supplied file Look through camera and split the viewport in two. Left side has cycles rendering turned on. Change the default renderer from Eevee to Cycles and change to GPU compute. Rotate camera with the empty and cycles view on left flickers. Next select cube and change the shader colour from pink to any other colour you like. Again viewport flickers like crazy.\n\n", " Freestyle renders extra lines\n%%%--- Operating System, Graphics card ---\n1st PC: Win7 64bit & Quattro 2000\n2nd PC: Win7 64bit & Nvidia 560ti\n\n\n\n - Blender version with error, and version that worked ---\n2.68a\n\n - Short description of error ---\nFreestyle produces lines through objects as they are transparent\nI build a moving staircase and some lines of the glass roof appear \"randomly\" through the walls.\n\n - Steps for others to reproduce the error (preferably based on attached .blend file) ---\n\n\nI just created objects + render. %%%", "Animated curves flickers when Motion Blur and Persistent Data are enabled\nIf there is an animated curve and Motion Blur + Persistent Data are enabled, the curves will randomly appear and disappear. Here is a simplified file that reproduces the issue on master. Render the animation an notice how the curves flicker.\n\n[curvesAnimatedCyclesIssue.blend](curvesAnimatedCyclesIssue.blend)\n\n### Original Report\n\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nDisplay Graphics card: Radeon RX 580 Series 4.5.14800 Core Profile Context 22.3.1 30.0.15002.1004\nCUDA enabled Graphics card: NVIDIA GeForce GTX 1650 Driver version. DCH 472.12\nCPU: Ryzen 7 2700 (16 threads, 8 cores) \n\n\nBroken: version: 3.2.0\n\n**Description of error**\nObjects disappear from what seems to be random frames close to the beginning of keyframed object transformations in renders produced with Cycles (only).\nThis appears to be the same bug as archived [#40313 ](T40313).\n\n\n\nBug is more prevalent in blender version: 3.1.2 however is still present.\nRender Examples:\n3.1: \n[3.1 build.mkv](3.1_build.mkv)\n3.2: \n[3.2 build.mkv](3.2_build.mkv)\n\nThis Project was using AnimAll addon to animate some of the objects featuring the glitch, these appare to be the only ones still glitching in 3.2. However testing this Addon in a controlled environment does not feature this behaviour, which might suggest its cause is elsewhere, but might be aggravated by animating individual vertices. \n\n* Open Attached Blend file\n* Render Animation in Cycles using files settings.\n\nThis render was made using GPU Compute with CUDA and CPU enabled in system preferences.\nEdit: works on CPU only also. Updated .Blend file with packed textures.\n\nFile removed now simplified is available\n", "Viewport Z-Bias & retopology overlay related issues\nSince this is being reported frequently and retopology overly has caused some weird issue with mask display as well previously, it's best we keep track of them in a central place so we could check if any particular fix is working for all situations:\n\n------\n\nProblem still presents:\n\n- Retopology overlay affecting other parts of the drawing/picking operation: #109115 #107768 ~~#109277~~ ~~#107411~~\n- (Potentially related) Selecting when X-Ray turned on is not visually effective and then switching to face mode crashes: #110136\n\n------\n\nFixed:\n\n- ~~Wireframe visible in close distances: #110048 #109040 #109514 #109499 #109685~~\n- ~~Retopology overlay 0 z offset problem: #109662~~\n\n-------\n\nProbable cause:\n\nAs suggested in 107768#issuecomment-937545 , I believe the problem here is that\n\n1. `DRW_object_is_in_edit_mode` does not exclusively check if object is in edit mode, but it combines other logics that affects drawing and stuff.\n2. `XRAY_ENABLED` marco is not always accompanied with `RETOPOLOGY_ENABLED`, this leads to inconsistencies in the user interaction and the display.", "Fix preview flickering when using \"Adjust Last Operation\"\nTweaking properties in the \"Adjust Last Operation\" panel removes all previews for a moment, causing quite drastic and annoying flickering.", "Occasional render glitches when using simulation zones on an object without a material.\nThis bug has an easy workaround, mostly posting this so anyone running into the same issue knows the solution.\n\nOperating system: Linux-5.15.0-79-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 530.30.02\n\nBroken: version: 3.6.2\nWorked: N/A, 3.6 is the first version with simulation nodes.\n\nIn Eevee, rendering an object with a simple simulation zone applied in geometry nodes and no material assigned to the object results in some rendered frames being blank or displaying odd glitches.\n\nThe workaround is to simply assign a material to the object.\n\nCheck out the attached .blend file for reproduction.\n\n" ]

[ "Trembling when selected cube in edit mode\nOperating system: MacOS Sierra - Versione 10.12.6\nGraphics card: NVIDIA GeForce GTX 680MX 2048 MB\n\nBroken: 2.83.0 (2020-06-03), 2.90 (2020-06-12)\nWorked: 2.82a\n\nA strange flicker happens when I select the faces of the cube.\n\nI select the cube.\nI change the visualization in Edit Mode.\nI set select mode to Face.\nI select the faces of the cube but they do not remain selected instead, a strange flicker of the object occurs. Here a video example:\n\n[blender_283_bug.mp4](blender_283_bug.mp4)", "Selected faces rendering incorrect\n(Sorry if this is a duplicate, the search thing works really bad, and the \"How to report a bug\" video is not up to date with the current website).\n\nOperating system: OSX\nGraphics card: Geforce GT 650M\n\n2.83.9, 2.83.8\n(it works fine in blender 2.82.7 for me, I tested one older version, 2.83.8, it also happens there, I don't have the time or disk space to test them all (unless I'm the only one having this bug....)\n\nSelected faces are rendered incorrect, see video:\nwatch?v=fjfIG9cUy9c&feature=youtu.be\n\nOpen blender, select the default cube, go to edit mode, select all faces.\n\n" ]

[ "Cannot Apply Multi Objects\nOperating system: Linux-5.8.0-53-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.14735 Core Profile Context 21.10\n\nBroken: version: 2.92.0\n\nWhen I try to apply scaling, rotation or location to a linked object (Alt+D) I get this error message:\n\"Cannot apply to a multi user: Object \"Cube.001\", Mesh \"Cube\", aborting\"\n \n- Scale the cube \n- Link it with Alt+D,\n- Apply scaling, location or rotation to one of the two cubes.\n\n[Blender 2021-05-30 00-10-18.mp4](Blender_2021-05-30_00-10-18.mp4)", "Regression: Image Editor Texture Draw Stops Updating\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 3GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.79\n\nBroken: version: 3.1.0 (probably introduced by {f663a1dc09}?)\nWorked: 3.0\n\nthis was actually exposed with 507a4deef1 (but we have seen before that that commit is not the real cause for a couple of issues -- @lichtwerk)\n\nImage on Image Editor -> Paint does not update when being drawn on at certain close view levels until the viewport is moved \n\nLoad default scene\nSwitch to Texture Paint tab\nCreate new image\nZoom in 7 or 8 times (for reference I am using a scroll wheel that has steps)\nDraw on the image\nIt will sometimes update fine, sometimes stop updating when it passes over a what looks like a UV boundry, that extends off the screen\n\nSee gif where I added a material to it\nSorry if this has already been reported, haven't been able to find a public bug tracker\n[image_paint_bug.blend](image_paint_bug.blend)\n\nThis is the file I (-- @lichtwerk) used on linux to bisect/reproduce on a 1920x1080 display (since this seems to depend on monitor/zoom etc):\n[image_paint_bug_repro.blend](image_paint_bug_repro.blend) ", "Constrained [on an axis] transform of an object far out of view can result in insanely high numbers\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.09\n\nBroken: version: 3.0.0\nWorked: never\n\nConstrained [on an axis] transform of an object far out of view can result in insanely high numbers\n\nThis file just contains a single cube 250m away from the origin:\n[#94105.blend](T94105.blend)\n- open file\n- hit {key G Y}\n- start moving the mouse\n- see values of 1000000000 m in the header\n\n**Original report**\n\nYour camera has to be close enough as in the attached video. Press G+Y and do a little move. The Cube starts to ignore modifiers and drastically increase Y-axis numeric value.\n[bug_10.mkv](bug_10.mkv)\n[bug_10.blend](bug_10.blend)", "Click selection and box selection in 3D view do not work properly in scenes with many objects\nOperating system: Darwin-19.2.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 580X OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.4.19\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional)\n\n[In my environment, if there are close to 100,000 objects, the objects can not be selected normally]\n[select_bug.mp4](select_bug.mp4)\n\nUploaded blend file has nearly 100,000 simple cubes\nClick on an object in the 3D view to select a distant object\nEven a box selection can make a crazy choice\nBut lasso selection and circle selection work fine\n[test_file.blend](test_file.blend)\n\n", "Wrong status bar keymap for release confirms and right click select\nOperating system: Manjaro x64\nGraphics card: HD5770\n\nBroken:\n2.80, 81ea815dcb6, blender2.8, 2018-12-12\nWorked: 2.79b\n\nWhen grabbing something(object, vertex, etc) by right clicking on it and then releasing/clicking on the same button the action is not performed, object doesn't return to its initial position. \nHowever if you use G or a Tool to grab instead of mice click Cancel by Right clicking option works.\n\nRight click on a Cube(in 2.79 thats enough to start moving an object but not here here you have to hold the button) AND HOLD the button\nDrag mice somewhere to move the Cube\nRight Click again to Cancel movement\nNothing happens.\n[blender_bug-2018-12-13_22.18.53.mkv](blender_bug-2018-12-13_22.18.53.mkv)\nI'm also attaching video file for it to be more clear.", "Animation Editors: Modifier keys cancel out dragging handles \nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.4.0 Alpha\n\nIn recent releases db4313610cab was implemented and the animation editors are still not adapted to this behaviour. This fix allows the use to press any keyboard key, like `Shift` or `X`, etc, before the drag threshold of the current action is reached, and it will cause the operator to execute.\n\nFor example: Click and hold on a move gizmo in object mode. Either drag 3 pixels to start the operation and then press `X` and hold `Shift` to slowly move along the x axis.\nOR immediately press `X` and hold `Shift` without reaching the drag threshold of 3 pixels to already start the operator.\n\nBefore this fix, any pressed keys would cancel the click & drag and prevent the operator from running.\nThis is still the case with the graph and dopesheet editors.\n\n- Create a simple animation to the default cube\n- click and drag on the keyframes or handles in the animation editors\n- Now click and hold them, hold `Shift`, and then continue to drag. Nothing happens.", "Knife tool fails on a non-planar n-gon\nBroken: 2.74 up to and including 2.92 master\nWorked: 2.73 (yes, I bisected that far back ;) )\n\n\nKnife tool fails to cut between verts in n-gons of certain configuration.\n\n\nLoad the attached .blend and try to make a knife cut between the two selected verts (i.e. try to create an edge connecting them).\n\n[knife_fail.blend](knife_fail.blend)\n\n(If you want to test in 2.73 yourself, copy the object from a running instance of 2.9x into 2.73 using ctrl+c/ctrl+v, then replace the mesh in another existing mesh object).", "Zoom stops at certain level and won't go further\nOperating system: Windows 10\nGraphics card: GeForce GTX 1070\n\nBroken: 2.92, 3.0 alpha\n\nIn this simple model, I cannot zoom my view closer to mesh than I already did. I saved the closest achievable view position in the .blend.\nI can reach further close-ups using some more advanced navigation tools, like dolly zoom, but standard zoom won't let me.\nThe issue is linked to this particular model - if I copy the object (there's only one in the scene) to a new model, the issue is gone.\nThe model was originally saved in Blender 3.0 alpha, only then in stable 2.92. Could it be the cause of this bug?\n\nBased on the default startup or an attached .blend file:\n[zoom issue.blend](zoom_issue.blend)", "UI: activate_init doesn't work with Int and Float type properties\nOperating system: Windows & Linux\n\nBroken: 3.2\nWorked: Unknown\n\nUsing activate_init to allow users to directly edit a property doesn't seem to work when the property is of type IntProperty or FloatProperty.\n\n1. Open the attached script in blenders text editor\n2. Run the script\n3. Go to the 3d viewport\n4. Ensure an object is selected/active\n5. Run the simple operator, either through the menu or operator search\n6. Notice that the property in the popup is not directly editable\n\n[test_activate_init.py](test_activate_init.py)", "Auto perspective sometimes fails\n**System Information** \nOperating system: Windows 10\nGraphics card: 3070\n\nBroken: 3.2.0 and 3.2.1\nWorked: 3.1\n\nAuto perspective setting doesn't work in this case: if you are in camera view or any side view and then you are switching to another side view you have this smooth transition animation between views. If during this animation you press middle mouse button (rotation) to go from side view it doesn't change to perspective. Moreover it \"shifts\" viewport like it was aligned to a bad quad. \n\n1. Turn on auto perspective.\n2. Go to top view.\n3. Change top view to any other side view.\n4. During smooth transition between view press middle mouse button and rotate the viewport.\n5. Now viewport is shifted and stays in orthographic instead perspective.", "Python API - Scale parameter isn't taken into account when adding certain primitives\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 3.0.0 Alpha\n\nIn text editor run this:\n\n```\nimport bpy\nbpy.ops.mesh.primitive_plane_add(scale=(5.0, 1.0, 0.0))\n```\n\n... and observe that plane has wrong scale(always default (1,1,1)).\n\ngoing over the primitives, the following need attention:\n\n```\n- add_primitive_plane_exec\n- add_primitive_grid_exec\n- add_primitive_monkey_exec\n- add_primitive_circle_exec\n- curvesurf_prim_add\n\n- NOPE object_add_exec\n\n- TODO effector_add_exec\n\n- these end up in object scale? (or dont make sense):\n- check sense lightprobe_add_exec\n- check object_camera_add_exec\n- object_metaball_add_exec\n\n- object_armature_add_exec\n- object_empty_add_exec\n- object_gpencil_add_exec\n- object_add_text_exec\n\n- object_light_add_exec\n- collection_instance_add_exec\n- object_speaker_add_exec\n- object_hair_add_exec\n- object_pointcloud_add_exec\n- object_volume_add\n```", "Importing a keymap with a default name makes a non deletable copy\nBroken: version: 3.2.1\n\nAfter importing a keymap configuration with the name \"Blender\" it created a copy of the default setting that is impossible to remove from the UI at least.\n\n\n\n\n\nImport a keymap setting with a default name\n\n\n**Fix**\nDelete keymap configuration from the user folder directly \n\n> \\AppData\\Roaming\\Blender Foundation\\Blender\\3.1\\scripts\\presets\\keyconfig\n\n\n", "View zoom (scale mode in particular) doesn't work well when dragging horizontally, or crossing the middle of the screen\nBroken: 2.79, 2.80, 2.81 HEAD (8c9e6f123a2).\nWorked: ?\n\nNavigation Mode Scale Is Broken - it should never be calculated from the vector zero.\n\n- Set Blender Navigation to Scale\n- Try to press Alt + click and drag in viewport to make it zooming (or click and hold the \"zoom\" icon in the viewport).\n- Try to cross the mouse arrow in the vector zero (i.e., cross the mouse around the middle of the viewport).\n- As you can see the tool start to work weirdly inverting the zooming as soon you cross the scene vector zero.\n\nNote: The other zoom modes (e.g., dolly) also present a strange behaviour when zooming horizontally. But the zoom seems to reset when going over the origin regardless of the mouse direction.\n", "In orthogonal view, XYZ axis overlay toggle doesn't work with axis aligned views\nBroken: 2.8x\n\nIn orthogonal view, XYZ axis overlay toggle doesn't work\n\n- press numpad 1/3/7\n- toggle xyz axis from overlays panel, they won't hide.\n\nVideo: [F8571714](2020-06-02_20-18-29.mp4)", "Python - operator called by gizmo always called the invoke() from the combination ALT + LEFTMOUSE (also asserts)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.83.0\n\nWhen an operator is called by a gizmo, the combination of the event ALT or CTRL or SHIFT + LEFTMOUSE always initialize the operator by calling the invoke(). \nOther combination of keys + LEFTMOUSE does not call the invoke()\nThis problem makes impossible to take into account the industry compatible keymap in the 3dview with the target_set_operator in gizmo.\n\nOpen the file gizmo_operator_invoke.blend\n- Toggle the system console to view the prints\n- Run the script :\n - In the 3dView, click once on the gizmo (gizmo now follows you mouse)\n - now press {key Alt} and start clicking -- using the combination ALT+LEFTMOUSE will call invoke() [inspect in the console] each time\n - use RIGHTCLICK or ESC repeatedly to stop the modal (if you clicked multiple times you have to do {key RMB} multiple times as well )\n - now press {key S} and start clicking -- use the combination \"S\"+LEFTMOUSE will **not** call invoke() [inspect in the console] each time\n - use RIGHTCLICK or ESC a single time to stop the modal\n\nNote: in a Debug build, an Assert should be triggered when pressing {key Alt} and start clicking\n\n```\nBLI_assert failed: source/blender/windowmanager/gizmo/intern/wm_gizmo_map.c:1068, wm_gizmomap_modal_set(), at 'gzmap->gzmap_context.modal == ((void *)0)'\n```\n\n\nThanks for your help\n\n[gizmo_operator_invoke_phi.blend](gizmo_operator_invoke_phi.blend)\n" ]

[ "Regression: Gizmos became significantly harder to use (in orthographic view)\nOperating system: Fedora 35 Linux (Xfce)\nGraphics card: GTX 750 Ti\n\nBroken: 3.1. Caused by {0cb5eae9d0617abedf745753c23061ddfcfd1416}\nSituation for the lines improved in 49fc4449e7 again, but the boxes/spheres/... on the end of the gizmo are still bad\nWorked: 2.93\n\nI've already reported this bug before and we archived it after I reset the settings to factory settings but now I realized that the problem is still there even with the default settings. It's just that it only happens in orthographic view, for some reason.\nUsing gizmos on a fixed axis only works from certain angles. Sometimes it's impossible, sometimes pixel perfect, sometimes works fine.\n[2022-04-06 20-10-20.mp4](2022-04-06_20-10-20.mp4)\n\n- Open attached file or:\n - With the default Cube go to Edit Mode\n - Enable orthographic view.\n - Try to use the Scale Tool or Extrude Tool with XYZ setting.\n - Try various different angles. Some of them won't work\n- Try to use the Blue axis\n[select_gizmo_ortho_bug.blend](select_gizmo_ortho_bug.blend)" ]

[ "blender 3.4 and 3.5 hangs during compositing when rendering an animation\nOperating system: Linux-5.4.0-132-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.141.03\n\nBroken: version: 3.4.0 Beta\n\nwhen rendering a longer animation the beta and alpha versions tend to hang on the \"initializing execution\" part of the compositing. No compositing happens on the render in the render window, you see frame n while the playhead already is a frame n+1.\n\nWhen rendering the same animation it happens randomly after 5 frames or after 150 frames, but mostly with 200+ samples and at least HD resolution, smaller res and less samples went through mostly. I compared the same animation file with 3.3.1 and it just worked and the very same file with 3.4 got stuck.\n\n", "Linked libraries+Cloth objects+LineArt -> Blender crashes\nOperating system: \"Linux Mint 20.1 Ulyssa\"\n\nGraphics card: NVIDIA Corporation TU106 [GeForce RTX 2060]\n\n**Blender Version** Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\n\nLinked libraries+Cloth objects+LineArt -> Blender crashes\n\nI'm in kind complex project. There're characters (roughly meshed, and bad rigged, but working), with 'clothed' hair, and all of them LineArted.\nMany of the scene files work ok, so they've not much characters. But I don't know why, some of them can't render more of a bunch of frames, even they crash when calculating something (i don't know if's about either cloth, collisions or LineArt modifiers, or whatever.\nMemory looks ok, a little is used.\nI attach both files: a scene, and the characters source (personajes.blend).\n\n - Load 's4 pl1 coleAna.blend'\n - Just 'Play' and wait.\n\n For me, it crashes after 5 frames or so.\n\nA simpler backtrace:\n\n# backtrace\n/home/klopes/soft/blender3/blender(BLI_system_backtrace+0x20) [0xb0b3460]\n/home/klopes/soft/blender3/blender() [0x10babdb]\n/lib/x86_64-linux-gnu/libc.so.6(+0x46210) [0x7fdfaf178210]\n/home/klopes/soft/blender3/blender() [0x2349478]\n/home/klopes/soft/blender3/blender() [0x234a454]\n/home/klopes/soft/blender3/blender() [0xb0b71dd]\n/home/klopes/soft/blender3/blender() [0x14b5b05]\n/home/klopes/soft/blender3/blender() [0x14b5dbb]\n/home/klopes/soft/blender3/blender() [0x14a49b7]\n/home/klopes/soft/blender3/blender() [0x14af7f0]\n/home/klopes/soft/blender3/blender() [0x14b182c]\n/home/klopes/soft/blender3/blender() [0x14b1a29]\n/lib/x86_64-linux-gnu/libpthread.so.0(+0x9609) [0x7fdfaf874609]\n/lib/x86_64-linux-gnu/libc.so.6(clone+0x43) [0x7fdfaf254293]\n\n- Python backtrace", "Render can not be cancelled and crashes\n\n\n\n\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: AMD Radeon R9 200 Series ATI Technologies Inc. 4.5.13570 Core Profile Context 19.7.2 26.20.13001.16003\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\n\nHello,\n\nI wanted to export some scene using the Eevee render. When I render a frame, the image processing statut is still at 0%. Thus, when I try to cancel it, Blender hangs / crashes.\n\nI launched it using the GPU debug tool\n\nI'm assuming it comes from the hardware.\n\n\nThanks in advance.\n\n\n\n[eevee_logs.txt](eevee_logs.txt)\n\n\n[blender2.8.blend](blender2.8.blend)\n\n\n\n\n\n- Open the file\n- Render image {key F12}\n- Progress will hang at 0% for too long time\n- Cancel rendering\n\nRender job is not cancelled immediately / Blender crashes with low VRAM\n\n", "AMD - Blender Crashes When I Start Render - Cycles GPU Compute\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: AMD Radeon RX 6800 XT ATI Technologies Inc. 4.5.0 Core Profile Context 23.7.1.230626\n\nBroken: version: 3.6.1\n\n[Whenever I start Rendering, It crashes when using cycles on GPU compute. Doesn't happen on CPU option.]\n\n[On render engine, select cycles. Then set it to GPU Compute. Then click on render on top left corner. Then select render image.]\n\n", "Blender crashing while baking Smoke Domain\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.0.0 & 3.2.0 Alpha\nWorked: NA\n\n\nBlender Crashes while baking smoke domain\n\n- open Project \n- Select smooth domain from outliner\n- {nav Physic properties > Bake All} (crash between frame 60 & 70)\n\n\n\n\n", "Fractured object crashes Render\nOperating system: macOS-13.3-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 560X OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.10.12\n\nBroken: version: 3.4.1\n\nRendering image crashes Blender\n\nThe attached file is the output of a fractured cube using RBDLabs, nothing else. It's a series of 965 simple static meshes. When displaying rendered shading in viewport, all is well. When doing Render > Render Image, then Blender crashes. Happens on my Mac M1 and Mac Intel, on Blender 3.3.4, 3.4.1, 3.5 and 3.6 alpha. If I delete the 115 last cells or the 134 first ones, then the render will properly happen. Thanks !\n\n", "Selecting vertices and centering the model crashes Blender\nOperating system: Windows 10\nGraphics card: RTX 2070 \n\nBroken: 03015a9b222e\n\n\nSelecting some vertices and trying to center the selection crashes blender. See the crash log\n\n[crash.txt](crash.txt)\n\nOpen the attached .blend\nSelect the object\nGo into edit mode\nSelect all vertices\nCenter (press Numpad .)\n\n\n\n\n\n[_13122021_2147_15.blend](_13122021_2147_15.blend)", "Crash when baking Dynamic Paint Image Sequence multiple times in loop\nBroken: 2.90.0 Alpha\nWorked: blender-2.80-windows64\n\nBlender crashes when baking dynamic paint image sequence in for loop. Depending on wait time crash will have differens stack traces.\n\n[#74052.blend](T74052.blend)\n\n- Open file\n- Run script", "Crash when rendering animation\nOperating system: macOS-13.4.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Pro 1.2\n\nBroken: version: 3.6.0\n\n\nI was rendering a simple camera animation overnight, Blender crashed about 2/3rd the way through. Here's the crash log.\n", "Crash using blender_debug_log.cmd and \"Reload Scripts\" in Cycles preview render\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.91\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\nLoading Blender with blender_debug_log.cmd, it crash immediately when I try to reload the scripts if Cycles is running. This is not happening when launching blender.exe \n\nLaunch blender 2.8 using blender_debug_log.cmd.\nIn the view 3d, choose Cycles as render preview.\nIn the search bar, look for \"Reload Scripts\" and click on it.\nBlender Crash.\n[blender_system_info.txt](blender_system_info.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n", "Background scene cause a crash in render animation\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.55\n\nBroken: version: \n - 2.92.0 Alpha\n - 2.91.0\n - 2.90.1\nWorked: 2.83.9\n\nBackground scene makes blender unstable and cause a crash in render animation\n\n- Open default startup file.\n- Create a new scene named \"Scene.001\".\n- Assign default scene \"Scene\" as the background scene of \"Scene.001\" in scene properties.\n- Assign default camera as the camera of \"Scene.001\" in scene properties.\n- Hit render animation from render menu.\n- After a couple of frames rendered, blender crashes.\n\nTested in different situations i had some problems.\nI think assigning a background scene makes blender unstable in general.\nOccasionally crashes when opening a file which has background scene.\nThe worst case is render animation.\n\n[Background_Scene_Crash.mp4](Background_Scene_Crash.mp4)\n\nReproduction steps with this file:\n1. Open file.\n2. Start Render Animation.\n[#83852.blend](T83852.blend)\n", "Motion Blur causes visual glitch on point cloud object moving\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.5.0 Alpha\nWorked: not known\n\nPoint cloud object + Motion Blur + animation. The first frame of animated movement will incorrectly render compared to before and after the animation starts. \nConvert Cube to point cloud\nAdd a location keyframe on frame 2\nmove the object and add another keyframe on frame 3\nRender with motion blur turned on.\nCompare the rendered frames.\nframe 1 looks correct, frame 2 is too dim even though it is not moving yet.\n\nNote: CPU renders frame 3 incorrectly as well. not sure if this requires another bug report. \n\nFrame 1:\n\nFrame 2:\n\n\nFrame 3 on GPU:\n\nFrame 3 on CPU:\n\n[MotionBlur_PointCloud.blend](MotionBlur_PointCloud.blend)\n", "EOL GPU problems tracking report\nThis is a task to keep track of End Of Life GPU problems with blender 2.8\n\n(Note that closed issues at time of writing is not because they are resolved but because of lack of respons from reporter)\n\n**AMD**\nHD 8670M #64469 (UI and 3D viewport doesn't update visually unless mouse hovers over other UI elements)\nHD 8550G #58544 (White Viewport in [Object Mode])\nHD 7730M {#63243\nHD 7600 {#70091}}\nHD 7600 #70146 (hair eevee rendering broken.)\nHD 7450M #61890 (AMD Radeon HD 7450M. Exit before splash screen. 4-component vector of vec4 to: 4-component vector of uvec4)\nHD 7420G #63149 (Interface and vertices flickering)\nHD 7160M #61165 (Blender 2.8 beta doesnt start)\nHD 6870 #64081 (Eevee glitching [Linux, Radeon GPU])\nHD 5700 #62572 (unwrapping bug )\nHD 5700 #62701 (Hair edit mode crashes with AMD Radeon HD 5700)\nHD4870 #65721 (Crash on start)\nHD 4850 #61206 (Linux and AMD RV770 interface glitches)\nHD 4600 #59738 (Buggy GUI) #61896 (UI graphic glich) #63850 (Icons bug)\nHD 4550 #60104 (Blender's visuals go transparent and spasms )\nHD 4500M #63661 (Unstable back color in menues/ titlescreen. shifting Transparency /black )\nHD 4350 #59635 (blender 2.80.0 beta wont open on windows machine)\nHD 4300 #65075 (blender interface background of menus and every element in ui flickers (transparent and not))\nHD 3650 #60681 (Blinking menu and other menu)\nHD 3450 #63751 (when i open blender beta it chashes at start)\nHD 3200 #62543 (Crash at startup)\nHD 2600 #59738 (Buggy GUI)\n\n**NVIDIA**\nGTX 9800 #60425 (EEVEE objects black when using shadows, on 9800 GTX)\nGTX 9800 #65250 (Slackware, Blender 2.8 UI fonts glitch with nouveu driver.)\n\n**Intel**\n\nHD GFX (Baytrail) #65138 (Intel HD Graphics bone selection problems in edit mode/pose mode)\nHD 4000 #58938 (Crash when using EEVEE and Intel HD 4000) #65666 (Blender not responding a lot of time, especialy when working with texture and materials)\n\n**macOS**\n(AMD)\nHD 6750 #62072 (Princioled BSDF and Indirect Lighting BUG ) #65453 (Bake indirect lighting turns irradiance volumes into black areas)\nHD 4870 #62982 (Blender 2D draw mode shows as corrupted screen)\nHD 4670 #66014 (Blender 2.8 using Mac OS, all objects turn black in Lookdev mode)\nHD 2400 XT #60979 (2D Animation view goes crazy when trying to use brush)\n(NVIDIA)\nGS 8800 #59955 (Blender starts with totally corrupted graphics)\nGT 9600M #61844 (User interface looks very weird when starting Blender 2.8 with older MacBook Pro (Mid 2009))\nGT 130 #61551 (macOS 10.11 and NVIDIA user interface display broken)\nGT 320M #60179 (Bad render of default scene Mac OSX)\n780M #62578 (Eevee Screen Space Reflections crashes on OSX 10.9 / 10.11)\n640M #63018 (macOS 10.10 / NVIDIA: missing letters in 2.80 UI)\nQuadro 4000 #62766 (Eevee DOF cut plane)\n(Intel)\nHD 3000 #65551 (Bad shading on LockDev on Macbook 11\")\n", "Blender 3.4 Random Crashing Since GPU Replacement\nOperating System: Windows-10-10.0.19044-SP0\nGPU: NVIDIA GeForce RTX 3060/PCIe/SSE2 4.5.0 NVIDIA 527.56\n\nBroken: version: 3.4.0, branch: blender-v3.4-release\n\n**Short Description**\nI have replaced my RX580 with an RTX3060 earlier and ever since that replacement, I have been having constant crashes whenever I attempt to use Blender, do note that I've been in edit/object mode during these crashes as I'm working on a subd model, I have tried a handful of fixes such as upping TDR values but it didn't seem to fix the issue.\n\nRyzen 5 2600, RTX3060, 16gb RAM, no temperature exceeds 60 degrees Celsius and nothing is active besides Blender 3.4 along with addons I have installed.\n\nHere are the logs I get when I run Blender via the gpu debug cmd: \n\nEdit: I did try disabling UV Packmaster as per the error logs, but that didn't stop the crashing.\n\n[blender_system_info.txt](blender_system_info.txt)\n[blender_debug_output.txt](blender_debug_output.txt)\n[gunflipped.crash.txt](gunflipped.crash.txt)", "Regression: Child hair particles set to \"Interpolated\" causes jittering when the based vertex changes position\nOperating system:win10\nGraphics card:RTX3060ti\n\nBroken: c8c6f62cf3a3,fca8df9415\nBroken: 3.5\nWorked: 3.4\n\nCaused by d20f9923224d2637374dd4390ae84c5041e3f9d3\n\nChild hair particles set to \"Interpolated\" causes jittering when the based vertex changes position\nThis problem seems to have been solved before,in #63534 and #56408\n\nOpen the file and render animation\n\n" ]

[ "Blender crash when tracking motion\nwhen i try to track a videoblender just crash.\nim using version 3.0.0\n\n\n\n> Quoted Text\n\n# Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nbpy.context.space_data.system_bookmarks_active = 1 # Property\nbpy.ops.clip.open(directory=\"C:\\\\Users\\\\Daniel\\\\Desktop\\\\\", files=[{\"name\":\"IMG_1956.MOV\", \"name\":\"IMG_1956.MOV\"}], relative_path=True) # Operator\nbpy.context.scene.frame_end = 46 # Property\nbpy.ops.clip.detect_features() # Operator\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF7D18B4560\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters- [x] : 0x0000000000000000\n\tParameters- [x] : 0x0000000000000010\n\n\nStack trace:\nblender.exe :0x00007FF7D18B4560 MEM_CacheLimiter_unmanage\nblender.exe :0x00007FF7D0CE9BD0 moviecache_valfree\nblender.exe :0x00007FF7D5B7A750 ghash_remove_ex\nblender.exe :0x00007FF7D5B79950 BLI_ghash_remove\nblender.exe :0x00007FF7D0CE98A0 check_unused_keys\nblender.exe :0x00007FF7D0CE9940 do_moviecache_put\nblender.exe :0x00007FF7D09C6CC0 put_imbuf_cache\nblender.exe :0x00007FF7D09C63F0 movieclip_get_postprocessed_ibuf\nblender.exe :0x00007FF7D09C4D30 BKE_movieclip_get_ibuf\nblender.exe :0x00007FF7D0B202D0 accessor_get_ibuf\nblender.exe :0x00007FF7D0B20760 accessor_get_image_callback\nblender.exe :0x00007FF7D1724130 `anonymous namespace'::LibmvFrameAccessor::GetImage\nblender.exe :0x00007FF7D5DFF3A0 mv::`anonymous namespace'::GetImageForMarker\nblender.exe :0x00007FF7D5DFF640 mv::AutoTrack::TrackMarker\nblender.exe :0x00007FF7D5DE1880 libmv_autoTrackMarker\nblender.exe :0x00007FF7D5DACEF0 autotrack_context_step_cb\nblender.exe :0x00007FF7D5BAB0B0 tbb::interface9::internal::dynamic_grainsize_mode<tbb::interface9::internal::adaptive_mode<tbb::int\nblender.exe :0x00007FF7D5BAB8A0 tbb::interface9::internal::start_reduce<tbb::blocked_range<int>,RangeTask,tbb::auto_partitioner con\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AE51B0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FFF13AEA470 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThreads:\nThread : 000006f0\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D5DB4DD0 pthread_mutex_lock\nblender.exe :0x00007FF7D09C63F0 movieclip_get_postprocessed_ibuf\nblender.exe :0x00007FF7D09C4D50 BKE_movieclip_get_ibuf_flag\nblender.exe :0x00007FF7D14F9220 uiTemplateMovieclipInformation\nblender.exe :0x00007FF7D0FA0340 RNA_function_call\nblender.exe :0x00007FF7D10C80F0 pyrna_func_call\npython39.dll :0x00007FFEF047FE60 PyObject_MakeTpCall\npython39.dll :0x00007FFEF0587920 PyEval_GetFuncDesc\npython39.dll :0x00007FFEF0581700 PyEval_EvalFrameDefault\npython39.dll :0x00007FFEF04802C0 PyCFunction_Call\npython39.dll :0x00007FFEF04803D0 PyFunction_Vectorcall\nblender.exe :0x00007FF7D10C9A50 bpy_class_call\nblender.exe :0x00007FF7D1071C60 panel_draw\nblender.exe :0x00007FF7D10ED680 ed_panel_draw\nblender.exe :0x00007FF7D10EBA30 ED_region_panels_layout_ex\nblender.exe :0x00007FF7D10EB7D0 ED_region_panels\nblender.exe :0x00007FF7D10EA3A0 ED_region_do_draw\nblender.exe :0x00007FF7D0B63310 wm_draw_window_offscreen\nblender.exe :0x00007FF7D0B63160 wm_draw_window\nblender.exe :0x00007FF7D0B62C60 wm_draw_update\nblender.exe :0x00007FF7D0B3D330 WM_main\nblender.exe :0x00007FF7D07BF190 main\nblender.exe :0x00007FF7D5CD7208 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00001c64\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D568F130 IlmThread_2_5::Semaphore::wait\nblender.exe :0x00007FF7D568E900 IlmThread_2_5::ThreadPool::numThreads\nblender.exe :0x00007FF7D568F5A0 std::_LaunchPad<std::unique_ptr<std::tuple<void (__cdecl IlmThread_2_5::Thread::*)(void) __ptr64,Il\nblender.exe :0x00007FF7D37CE950 std::_Pad::_Call_func\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000036e4\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D568F130 IlmThread_2_5::Semaphore::wait\nblender.exe :0x00007FF7D568E900 IlmThread_2_5::ThreadPool::numThreads\nblender.exe :0x00007FF7D568F5A0 std::_LaunchPad<std::unique_ptr<std::tuple<void (__cdecl IlmThread_2_5::Thread::*)(void) __ptr64,Il\nblender.exe :0x00007FF7D37CE950 std::_Pad::_Call_func\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000024a8\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D568F130 IlmThread_2_5::Semaphore::wait\nblender.exe :0x00007FF7D568E900 IlmThread_2_5::ThreadPool::numThreads\nblender.exe :0x00007FF7D568F5A0 std::_LaunchPad<std::unique_ptr<std::tuple<void (__cdecl IlmThread_2_5::Thread::*)(void) __ptr64,Il\nblender.exe :0x00007FF7D37CE950 std::_Pad::_Call_func\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000036fc\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D568F130 IlmThread_2_5::Semaphore::wait\nblender.exe :0x00007FF7D568E900 IlmThread_2_5::ThreadPool::numThreads\nblender.exe :0x00007FF7D568F5A0 std::_LaunchPad<std::unique_ptr<std::tuple<void (__cdecl IlmThread_2_5::Thread::*)(void) __ptr64,Il\nblender.exe :0x00007FF7D37CE950 std::_Pad::_Call_func\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 0000314c\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\natio6axx.dll :0x0000000000ADC870 DrvPresentBuffers\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00001314\nntdll.dll :0x00007FFF2CAE3EA0 NtWaitForMultipleObjects\nKERNELBASE.dll :0x00007FFF2A2DFA20 WaitForMultipleObjectsEx\ncombase.dll :0x00007FFF2AA8F020 CoGetCurrentLogicalThreadId\ncombase.dll :0x00007FFF2AA8F020 CoGetCurrentLogicalThreadId\ncombase.dll :0x00007FFF2AA8F020 CoGetCurrentLogicalThreadId\ncombase.dll :0x00007FFF2AA8F020 CoGetCurrentLogicalThreadId\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000037a8\nntdll.dll :0x00007FFF2CAE6F00 ZwWaitForWorkViaWorkerFactory\nntdll.dll :0x00007FFF2CA565B0 EtwNotificationRegister\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 0000105c\nntdll.dll :0x00007FFF2CAE6F00 ZwWaitForWorkViaWorkerFactory\nntdll.dll :0x00007FFF2CA565B0 EtwNotificationRegister\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 0000085c\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\natio6axx.dll :0x0000000000ADC870 DrvPresentBuffers\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00002214\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D5DB4DD0 pthread_mutex_lock\nblender.exe :0x00007FF7D09C63F0 movieclip_get_postprocessed_ibuf\nblender.exe :0x00007FF7D09C4D30 BKE_movieclip_get_ibuf\nblender.exe :0x00007FF7D0B202D0 accessor_get_ibuf\nblender.exe :0x00007FF7D0B20760 accessor_get_image_callback\nblender.exe :0x00007FF7D1724130 `anonymous namespace'::LibmvFrameAccessor::GetImage\nblender.exe :0x00007FF7D5DFF3A0 mv::`anonymous namespace'::GetImageForMarker\nblender.exe :0x00007FF7D5DFF640 mv::AutoTrack::TrackMarker\nblender.exe :0x00007FF7D5DE1880 libmv_autoTrackMarker\nblender.exe :0x00007FF7D5DACEF0 autotrack_context_step_cb\nblender.exe :0x00007FF7D5BAB0B0 tbb::interface9::internal::dynamic_grainsize_mode<tbb::interface9::internal::adaptive_mode<tbb::int\nblender.exe :0x00007FF7D5BAB8A0 tbb::interface9::internal::start_reduce<tbb::blocked_range<int>,RangeTask,tbb::auto_partitioner con\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AE51B0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FFF13AEA470 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00002880\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D5DB4DD0 pthread_mutex_lock\nblender.exe :0x00007FF7D09C63F0 movieclip_get_postprocessed_ibuf\nblender.exe :0x00007FF7D09C4D30 BKE_movieclip_get_ibuf\nblender.exe :0x00007FF7D0B202D0 accessor_get_ibuf\nblender.exe :0x00007FF7D0B20760 accessor_get_image_callback\nblender.exe :0x00007FF7D1724130 `anonymous namespace'::LibmvFrameAccessor::GetImage\nblender.exe :0x00007FF7D5DFF3A0 mv::`anonymous namespace'::GetImageForMarker\nblender.exe :0x00007FF7D5DFF640 mv::AutoTrack::TrackMarker\nblender.exe :0x00007FF7D5DE1880 libmv_autoTrackMarker\nblender.exe :0x00007FF7D5DACEF0 autotrack_context_step_cb\nblender.exe :0x00007FF7D5BAB0B0 tbb::interface9::internal::dynamic_grainsize_mode<tbb::interface9::internal::adaptive_mode<tbb::int\nblender.exe :0x00007FF7D5BAB8A0 tbb::interface9::internal::start_reduce<tbb::blocked_range<int>,RangeTask,tbb::auto_partitioner con\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AE51B0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FFF13AEA470 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FFF13AEDBD0 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 0000332c\nntdll.dll :0x00007FFF2CAE6F00 ZwWaitForWorkViaWorkerFactory\nntdll.dll :0x00007FFF2CA565B0 EtwNotificationRegister\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00002e60\nntdll.dll :0x00007FFF2CAE6F00 ZwWaitForWorkViaWorkerFactory\nntdll.dll :0x00007FFF2CA565B0 EtwNotificationRegister\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00003660\nntdll.dll :0x00007FFF2CAE6F00 ZwWaitForWorkViaWorkerFactory\nntdll.dll :0x00007FFF2CA565B0 EtwNotificationRegister\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000004d8\nntdll.dll :0x00007FFF2CAE6EA0 NtWaitForAlertByThreadId\nntdll.dll :0x00007FFF2CAAA560 RtlSleepConditionVariableSRW\nKERNELBASE.dll :0x00007FFF2A2ED4F0 SleepConditionVariableSRW\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\nmsvcrt.dll :0x00007FFF2C98DE80 beginthreadex\nmsvcrt.dll :0x00007FFF2C98DFE0 endthreadex\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00002ff0\nntdll.dll :0x00007FFF2CAE6EA0 NtWaitForAlertByThreadId\nntdll.dll :0x00007FFF2CAAA560 RtlSleepConditionVariableSRW\nKERNELBASE.dll :0x00007FFF2A2ED4F0 SleepConditionVariableSRW\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\nmsvcrt.dll :0x00007FFF2C98DE80 beginthreadex\nmsvcrt.dll :0x00007FFF2C98DFE0 endthreadex\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00003480\nntdll.dll :0x00007FFF2CAE6EA0 NtWaitForAlertByThreadId\nntdll.dll :0x00007FFF2CAAA560 RtlSleepConditionVariableSRW\nKERNELBASE.dll :0x00007FFF2A2ED4F0 SleepConditionVariableSRW\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\nmsvcrt.dll :0x00007FFF2C98DE80 beginthreadex\nmsvcrt.dll :0x00007FFF2C98DFE0 endthreadex\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 00001cc4\nntdll.dll :0x00007FFF2CAE6EA0 NtWaitForAlertByThreadId\nntdll.dll :0x00007FFF2CAAA560 RtlSleepConditionVariableSRW\nKERNELBASE.dll :0x00007FFF2A2ED4F0 SleepConditionVariableSRW\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\navcodec-58.dll :0x00007FFEBBA76D70 av_parser_iterate\nmsvcrt.dll :0x00007FFF2C98DE80 beginthreadex\nmsvcrt.dll :0x00007FFF2C98DFE0 endthreadex\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nThread : 000037f8\nntdll.dll :0x00007FFF2CAE33D0 ZwWaitForSingleObject\nKERNELBASE.dll :0x00007FFF2A2D29C0 WaitForSingleObjectEx\nblender.exe :0x00007FF7D5DB4DD0 pthread_mutex_lock\nblender.exe :0x00007FF7D09C63F0 movieclip_get_postprocessed_ibuf\nblender.exe :0x00007FF7D09C4D30 BKE_movieclip_get_ibuf\nblender.exe :0x00007FF7D0B202D0 accessor_get_ibuf\nblender.exe :0x00007FF7D0B20760 accessor_get_image_callback\nblender.exe :0x00007FF7D1724130 `anonymous namespace'::LibmvFrameAccessor::GetImage\nblender.exe :0x00007FF7D5DFF3A0 mv::`anonymous namespace'::GetImageForMarker\nblender.exe :0x00007FF7D5DFF640 mv::AutoTrack::TrackMarker\nblender.exe :0x00007FF7D5DE1880 libmv_autoTrackMarker\nblender.exe :0x00007FF7D5DACEF0 autotrack_context_step_cb\nblender.exe :0x00007FF7D5BAB0B0 tbb::interface9::internal::dynamic_grainsize_mode<tbb::interface9::internal::adaptive_mode<tbb::int\nblender.exe :0x00007FF7D5BAB8A0 tbb::interface9::internal::start_reduce<tbb::blocked_range<int>,RangeTask,tbb::auto_partitioner con\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFF13AEF690 tbb::recursive_mutex::scoped_lock::internal_try_acquire\nblender.exe :0x00007FF7D5BABB00 BLI_task_parallel_range\nblender.exe :0x00007FF7D5DAC820 BKE_autotrack_context_step\nblender.exe :0x00007FF7D150B710 track_markers_startjob\nblender.exe :0x00007FF7D0B67860 do_job_thread\nblender.exe :0x00007FF7D5DB39C0 _ptw32_threadStart\nucrtbase.dll :0x00007FFF2A066BB0 recalloc\nKERNEL32.DLL :0x00007FFF2C3D54D0 BaseThreadInitThunk\nntdll.dll :0x00007FFF2CA44830 RtlUserThreadStart\n\n\nLoaded Modules :\n0x00007FF7D0640000 3.0.0.0 blender.exe C:\\Program Files\\Blender Foundation\\Blender 3.0\\blender.pdb \n0x00007FFF2CA40000 10.0.22000.348 ntdll.dll \n0x00007FFF2C3C0000 10.0.22000.318 KERNEL32.DLL \n0x00007FFF2A290000 10.0.22000.348 KERNELBASE.dll \n0x00007FFF274F0000 10.0.22000.282 apphelp.dll \n0x00007FFF2AE60000 10.0.22000.282 USER32.dll \n0x00007FFF2A8A0000 10.0.22000.348 win32u.dll \n0x00007FFF1AEA0000 tbbmalloc.dll \n0x00007FFF13AE0000 tbb.dll \n0x00007FFF2A960000 10.0.22000.1 GDI32.dll \n0x00007FFF29F20000 10.0.22000.71 gdi32full.dll \n0x00007FFF2A800000 10.0.22000.1 msvcp_win.dll \n0x00007FFF2A040000 10.0.22000.1 ucrtbase.dll \n0x00007FFF2B540000 10.0.22000.348 SHELL32.dll \n0x00007FFF1BC00000 14.28.29910.0 VCRUNTIME140.dll \n0x00007FFF0B520000 14.28.29910.0 MSVCP140.dll \n0x00007FFF2A8F0000 10.0.22000.1 WS2_32.dll \n0x00007FFF2C5F0000 10.0.22000.258 RPCRT4.dll \n0x00007FFF26570000 14.28.29910.0 VCRUNTIME140_1.dll \n0x00007FFF2A990000 10.0.22000.282 ADVAPI32.dll \n0x00007FFF2C950000 7.0.22000.1 msvcrt.dll \n0x00007FFF2ADC0000 10.0.22000.1 sechost.dll \n0x00007FFF03070000 8.0.1.0 openvdb.dll \n0x00007FFF01C70000 10.0.22000.41 OPENGL32.dll \n0x00007FFF2BCF0000 10.0.22000.120 ole32.dll \n0x00007FFF2AA40000 10.0.22000.282 combase.dll \n0x00007FFF2A8D0000 10.0.22000.1 PSAPI.DLL \n0x00007FFF2C720000 10.0.22000.1 SHLWAPI.dll \n0x00007FFF2C500000 10.0.22000.71 shcore.dll \n0x00007FFEBB590000 58.134.100.0 avcodec-58.dll \n0x00007FFF2B4A0000 10.0.22000.1 IMM32.dll \n0x00007FFF1A3E0000 58.13.100.0 avdevice-58.dll \n0x00007FFF2C870000 10.0.22000.1 OLEAUT32.dll \n0x00007FFEFE630000 58.76.100.0 avformat-58.dll \n0x00007FFEF0350000 3.9.7150.1013 python39.dll \n0x00007FFEE4ED0000 56.70.100.0 avutil-56.dll \n0x00007FFF0B480000 5.9.100.0 swscale-5.dll \n0x0000000070440000 1.0.28.0 libsndfile-1.dll \n0x000000006ACC0000 libgmp-10.dll \n0x00007FFF010B0000 1.20.1.0 OpenAL32.dll \n0x00007FFEFDFF0000 2.0.12.0 SDL2.dll \n0x00007FFF26490000 tbbmalloc_proxy.dll \n0x00007FFF29B50000 10.0.22000.1 CFGMGR32.dll \n0x00007FFF2BEA0000 10.0.22000.194 SETUPAPI.dll \n0x00007FFF1BEB0000 6.10.22000.120 COMCTL32.dll \n0x00007FFF13AB0000 10.0.22000.1 AVIFIL32.dll \n0x00007FFF21540000 10.0.22000.1 VERSION.dll \n0x00007FFF17E00000 10.0.22000.1 dbghelp.dll \n0x00007FFF13A30000 10.0.22000.41 GLU32.dll \n0x00007FFF131B0000 3.9.100.0 swresample-3.dll \n0x0000000070680000 libfftw3-3.dll \n0x00007FFF1A380000 10.0.22000.1 AVICAP32.dll \n0x00007FFF29810000 10.0.22000.1 bcrypt.dll \n0x00007FFF15300000 10.0.22000.1 Secur32.dll \n0x00007FFF21500000 10.0.22000.1 WINMM.dll \n0x00007FFF0D420000 10.0.22000.1 MSACM32.dll \n0x00007FFF12FC0000 10.0.22000.1 MSVFW32.dll \n0x00007FFF277A0000 10.0.22000.1 dxcore.dll \n0x00007FFF29260000 10.0.22000.282 SSPICLI.DLL \n0x00007FFF29040000 10.0.22000.71 kernel.appcore.dll \n0x00007FFF280C0000 10.0.22000.348 windows.storage.dll \n0x00007FFF27F50000 10.0.22000.41 wintypes.dll \n0x00007FFF276A0000 10.0.22000.120 uxtheme.dll \n0x00007FFF2A780000 10.0.22000.318 bcryptPrimitives.dll \n0x00007FFF29E50000 10.0.22000.1 profapi.dll \n0x00007FFF2C310000 2001.12.10941.16384 clbcatq.dll \n0x00007FFF1C9F0000 10.0.22000.1 MMDevApi.dll \n0x00007FFF29B20000 10.0.22000.1 DEVOBJ.dll \n0x00007FFF10C50000 10.0.22000.348 AUDIOSES.DLL \n0x00007FFF29DF0000 10.0.22000.1 powrprof.dll \n0x00007FFF29DD0000 10.0.22000.1 UMPDC.dll \n0x00007FFF2B350000 10.0.22000.348 MSCTF.dll \n0x00007FFF23250000 10.0.22000.258 AppXDeploymentClient.dll \n0x00007FFF22B60000 11.0.22000.282 urlmon.dll \n0x00007FFF22080000 11.0.22000.348 iertutil.dll \n0x00007FFF22050000 10.0.22000.1 srvcli.dll \n0x00007FFF28B70000 10.0.22000.1 netutils.dll \n0x00007FFF14370000 8.14.1.6489 atig6pxx.dll \n0x00007FFF05CC0000 7.16.10.1301 atiadlxx.dll \n0x00007FFF26580000 7.0.22000.37 PROPSYS.dll \n0x00007FFF29540000 10.0.22000.1 USERENV.dll \n0x00007FFF26680000 10.0.22000.1 WTSAPI32.dll \n0x00007FFF28B80000 10.0.22000.282 IPHLPAPI.DLL \n0x00007FFF2A220000 10.0.22000.194 WINTRUST.dll \n0x00007FFF2A610000 10.0.22000.348 CRYPT32.dll \n0x00007FFF296F0000 10.0.22000.1 MSASN1.dll \n0x0000000000990000 6.14.10.13417 atio6axx.dll \n0x00007FFF27960000 10.0.22000.41 dwmapi.dll \n0x00007FFF11660000 8.14.1.6489 atig6txx.dll \n0x00007FFF09B60000 10.0.22000.1 dataexchange.dll \n0x00007FFF22D50000 10.0.22000.318 twinapi.appcore.dll \n0x00007FFF1C5D0000 10.0.22000.282 textinputframework.dll \n0x00007FFF270B0000 10.0.22000.71 CoreMessaging.dll \n0x00007FFF244C0000 10.0.22000.132 CoreUIComponents.dll \n0x00007FFF296B0000 10.0.22000.1 CRYPTBASE.DLL \n0x00007FFF0A560000 10.0.22000.348 explorerframe.dll \n0x00007FFF29690000 10.0.22000.1 CRYPTSP.dll \n0x00007FFF28FA0000 10.0.22000.282 rsaenh.dll \n0x00007FFF26450000 3.9.7150.1013 python3.DLL \n0x00007FFF0F3D0000 3.9.7150.1013 _bz2.pyd \n0x00007FFF0CFC0000 3.9.7150.1013 _lzma.pyd \n0x00007FFF0A360000 sketchup.cp39-win_amd64.pyd \n0x00007FFEDDC70000 21.0.339.0 SketchUpAPI.dll \n0x00007FFF08E90000 21.0.339.121 SketchUpCommonPreferences.dll \n0x00007FFF0D500000 3.9.7150.1013 _socket.pyd \n0x00007FFF26430000 3.9.7150.1013 select.pyd \n0x00007FFF21A60000 3.9.7150.1013 _queue.pyd \n0x00007FFEE4BE0000 _multiarray_umath.cp39-win_amd64.pyd \n0x00007FFF0D390000 _multiarray_tests.cp39-win_amd64.pyd \n0x00007FFF0CD10000 3.9.7150.1013 _ctypes.pyd \n0x00007FFF20640000 libffi-7.dll \n0x00007FFF08170000 lapack_lite.cp39-win_amd64.pyd \n0x00007FFEEFF40000 _umath_linalg.cp39-win_amd64.pyd \n0x00007FFF0C930000 _pocketfft_internal.cp39-win_amd64.pyd \n0x00007FFF06FE0000 mtrand.cp39-win_amd64.pyd \n0x00007FFF0C140000 bit_generator.cp39-win_amd64.pyd \n0x00007FFF0B3C0000 _common.cp39-win_amd64.pyd \n0x00007FFF0B6E0000 3.9.7150.1013 _hashlib.pyd \n0x00007FFEE1C70000 1.1.1.12 libcrypto-1_1.dll \n0x00007FFF0A310000 _bounded_integers.cp39-win_amd64.pyd \n0x00007FFF0AED0000 _mt19937.cp39-win_amd64.pyd \n0x00007FFF0A2F0000 _philox.cp39-win_amd64.pyd \n0x00007FFF08E70000 _pcg64.cp39-win_amd64.pyd \n0x00007FFF08E50000 _sfc64.cp39-win_amd64.pyd \n0x00007FFF01BC0000 _generator.cp39-win_amd64.pyd \n0x00007FFF1E100000 3.9.7150.1013 _uuid.pyd \n0x00007FFF083F0000 3.9.7150.1013 _ssl.pyd \n0x00007FFF01000000 1.1.1.12 libssl-1_1.dll \n0x00007FFF083D0000 3.9.7150.1013 _sqlite3.pyd \n0x00007FFEFC190000 3.35.5.0 sqlite3.dll \n0x00007FFF29450000 10.0.22000.1 mswsock.dll \n0x00007FFEFD8A0000 3.9.7150.1013 unicodedata.pyd \n0x00007FFF13E10000 10.0.22000.1 LINKINFO.dll \n0x00007FFF11BB0000 10.0.22000.51 NetworkExplorer.dll \n0x00007FFF174D0000 10.0.22000.1 MPR.dll \n0x00007FFF09A70000 10.0.22000.1 p9np.dll \n0x00007FFF13150000 10.0.22000.1 drprov.dll \n0x00007FFF11720000 10.0.22000.282 ntlanman.dll \n0x00007FFF11090000 10.0.22000.1 davclnt.dll \n0x00007FFF18D90000 10.0.22000.1 Windows.System.Launcher.dll \n0x00007FFF1D080000 10.0.22000.1 msvcp110_win.dll \n0x00007FFF1AF40000 10.0.22000.65 windows.staterepositorycore.dll \n\n# Python backtrace\n\n", "Blender 3.0 Motion tracking footage just crashes blender\nHaving troubles tracking footage inside of the latest official blender, tracking footage works for a few seconds before the program just closes, footage works just fine in Blender 2.93.\n\n**Footage Information**\n35.2MB, 3840x2160 29.97 FPS, mp4\n\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.09\n\nBroken: version: 3.0.0\n\n\n\n", "Motion Tracking very ez to crush.\nOperating system:\nGraphics card:\n\nBroken: (example: 3.0.0, f1cca3055776, master, 2021-12-02)\n\nMotion Tracking very ez to crush.\nMaybe 4K video easy to let blender shutdown.\n\nFilm conditions : 4K 3840*2160 & 5s .\n\n1.Open blender.\n2.Drag video file in blender \"motion Tracking\".\n3.\"Set SceneFreames \" -> \"Prefetch\" -> Motion Model select \"Perspective\".\n4.\"Detect Features\" Margin:16 Threshold : 0.5 Distance : 80.\n5.Ctrl+T .\n6.wait blender crush." ]

[ "Set as Background crashes blender instantly.\nOperating system: Windows 10 Pro\nGraphics card: RTX 2080 Ti and RTX 3060\n\nBroken: 3.62\n\nI'm trying to do object tracking. Have tracked an object and clicked Solve Object Motion, then when I click on Set as Background Blender crashes instantly. Trying to load in a 400 frame 1920x1080 png sequence. Blender also crashes if I try to load the background image sequence via the camera properties (Background Images > Add Image).\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n", "Rendering object in Cycles X preview hard crashes whole system\nOperating system: Windows 10 x64\nGraphics card: nVidia RTX 8000\n\nBroken: blender-3.5.0-alpha+universal-scene-description.1f9e90bb1cf3-windows.amd64-release\n\nRendering of scene in Cycles preview hard crashes whole system. System screen goes to black and the fans go to max.\n\nLoad blend file at link: brutes_set1C_trimmed.blend?dl=0\n\nSet screen to Render preview and navigate around.\n\n", "Geometry Nodes related memory leak and exception access violation\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.5.1\nWorked: \n\nBlender freezes, memory usage spikes and eventually you get exception access violation crash.\n\nOpen attached file and enable Layer 3. Blender will freeze, memory usage will spike and eventually lead to exception access violation crash. Crash log is generated.\n\n[gnodes bug1.blend](attachment)\n\nTo stop crashing disconnect Twig Object from Object Info node. I think this started happening after I deleted the \"fir_twig\" object, but there was still a hidden reference in the Object info. However, I was not able to replicate this in a new file and the crash stops happening if I delete / modify some of the node groups, even in unrelated ways. I am pushing Blender to the limits and this is a heavily reduced file.\n\n", "Geometry Node Crash\nOperating system: macOS-11.6-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon HD - FirePro D500 1.2\n\nBroken: version: 3.6.2\n\nGeometry Node crash when I try to add a node to the default Cube.\n| Node Setup: | Crash Log: |\n| --- | --- |\n|||\n\n- open blender\n- open the geometry node workspace\n- With the default cube selected, add a new Geo Node setup\n- add a node using the menu or Shift A\n\nBlender crashes.\n\nIf I delete the cube, add a plane, and add the geo node to the plane, blender does not crash.\n\n", "Crash when importing / working with big or multiple VDB's in Shaded or Rendered 3dView (GPU & CPU)\nOperating system: macOS-13.3.1-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon Pro 5700 XT 1.2\n\nBroken: version: 3.6.0 Alpha\n\n\nBlender crashes when using big VDB files or scattered VDB's.(clouds)\nIn the viewport, when scattering VDB files, it is only possible in wireframe view mode.\n- When switching to only GPU shaded or rendered mode blender crashes.\n- When switching to only CPU you are able to use wirframe and rendered 3dview as well as rendering the image.\nBut not shaded view!\n\nMacOS 13.3.1 CPU --> 64GB \nAMD Radeon Pro 5700XT GPU --> 16GB\n\nSingle VDB files up to 450MB are OK. \nBigger cloud VDB used = 2.8GB crashes right after or while importing.\nThe scattered clouds are smaller in filesize, but more of them used.\n\nWhen opening the files, it's then only possible in wireframe mode.\n\n3DViewport used without realtime compositor !\n\nSince this is to big to upload here, the VDB file is here for free to test: clouds/\n\nOpen latest MacOS Blender version, switch to wireframe mode and CPU.\nImport the VDB from the link above.\nNow try shaded & rendered 3dview with CPU and then GPU. \n\n\nCRASHLOG:\n```\nRead prefs: \"/Users/NURB/Library/Application Support/Blender/3.6/config/userpref.blend\"\nDevice with name AMD Radeon Pro 5700 XT supports metal minimum requirements\nMETAL API - DETECTED GPU: AMD Radeon Pro 5700 XT\nregister\n\nWARN (bpy.rna): source/blender/python/intern/bpy_rna.c:1342 pyrna_enum_to_py: current value '17' matches no enum in 'View3DShading', '(null)', 'studio_light'\nWARN (bpy.rna): source/blender/python/intern/bpy_rna.c:1342 pyrna_enum_to_py: current value '17' matches no enum in 'View3DShading', '(null)', 'studio_light'\n\n[MTLTextureDescriptorInternal validateWithDevice:]:1344: failed assertion `Texture Descriptor Validation\nMTLTextureDescriptor has depth (2088) greater than the maximum allowed size of 2048.'\n\n[1] 42841 abort /Users/NURB/blender-git/build_darwin/bin/Blender.app/Contents/MacOS/Blender\n```\n", "Dragging the cursor (shift-right-click-drag) locks it's rotation\nOperating system: ~~Windows-10-10.0.22621-SP0 64 Bits~~ Windows 11 22H2\n> ***Meta Bug: this doesn't seem to recognize I'm on Windows 11.***Graphics card: NVIDIA RTX A2000 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.24\nBroken: version: 3.4.1\nWorked: /\n\nWhen setting the 3D cursor to project to surface and align its rotation to the geometry, it doesn't update when dragging. It only sets the aligned rotation on first click.\n\n1. Change Keybinding setting for shift right-click\n\n2. Add UV sphere\n3. Shift right-click-drag the 3D cursor over different face of the sphere\n4. Notice it only sets the rotation once initially\n\n", "Crash to desktop when trying to bake (Blender 3.3.6)\nOperating system: Windows 11 (64-bit)\nGraphics card: NVIDIA RTX A2000\n\nBroken: (3.3.5, 3.3.6, blender-3.3.6-stable+v33.948f8298b982-windows.amd64-release)\nWorking (2.91.2)\n\nWhen trying to bake scene produced in script, Blender crashed.\n\nOpen attached blend file and press Bake.", "3D Cursor don´t take into actcount the face orenration with keyboard shortcut ( Shift + RClick)\nversion: 2.80 (sub 51), branch: blender2.7, commit date: 2019-03-25 00:55, hash: 07f6be87a97e, type: Release\nbuild date: 24/03/2019, 18:03\nplatform: Windows\n\n3D Cursor don´t take into account the Geometry orentation (topbar options ) with keyboard shortcut\n\n- Open blender and select 3D cursor tool\n- Set \"Geometry\" in the topbar 3D Cursor options\n- Use Shift + RClick to set your 3D cursor on top of default cube\n\nResult:\n\n3D Cursor don´t take into actcount the face orenration \n", "Cursor frozen when switching to walk/fly navigation while moving the mouse\nOperating system: Windows 10\n\nBroken: 2.82 (sub 3)\n\nSwitching to walk/fly navigation through the shortcut while moving the mouse, causes the cursor to become stuck. The cursor remains stuck until the mouse isn't moved anymore for a short moment.\n\n\n - Move the mouse around in 3D view.\n - While moving the mouse press the shortcut to enable walk/fly navigation.\n - Continue to move the mouse and observe the cursor not following the mouse movement.\n - Stop moving the mouse.\n - Move the mouse again and the cursor should follow as expected.\n", "Regression: Crash when animation is playing\nOperating system: Windows 10\nGraphics card: 2x RTX3080\n\nBroken: 3.6 and 3.6.1 release canditate\n\nPress the space bar, the animation starts and Blender closes. The console says : EXCEPTION_ACCESS_VIOLATION\nI also test the file with an independent portable version of Blender, but the crash still occurs.\n\nPress space bar to play the animation and wait the frame 48.\n\n[Fluent_Testing_7_for_blend_dev.blend](https://projects.blender.orgattachment)", "Crash with Boolean modifier\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 3.4.0 Alpha\nWorked: 3.1.2 (appears to have been broken since 3.2.0)\n\nBlender crashes when performing an Intersection Boolean between a cube and a bunch of loose faces. (Granted, this is probably not something you should do in the first place, but at least it should be limited to producing a bad result, not a crash.)\n\nI didn't find a crash log, but it's easy to reproduce.\n\n* Load the attached .blend file.\n* Scrub the timeline to frame 209. Blender crashes.\n\n[crash.blend](crash.blend)\n", "Regression: Crash on deleting NLA strip in tweak mode\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.0, 3.6.0\nWorked: 2.93\n\nWhen using the 'Previous Layer' function of the Action Editor to tweak NLA strips, trying to delete a (green) strip in tweak mode crashes Blender.\n\n\n- Open `109430.blend`\n- In the NLA area, press X to delete the green strip.\n- Blender crashes\n\nOriginal/from fresh file:\n\n- In the default startup file, open 2 areas; a Dope Sheet set to Action Editor, and a Nonlinear Animation.\n- Keyframe any property of the default cube.\n- In the Action Editor, click Push Down on the newly created action.\n- Also in the Action Editor, click the 'Previous Layer' button (the small down arrow). This activates tweaking for the NLA strip.\n- In the NLA area, click the green strip so it's highlighted.\n- Press X to delete it. Blender crashes.\n\n", "Blender 3D shift+RMB \"set 3D cursor\" not sensing strokes made with the grease pencil nor strokes made with the annotation tool\nOperating system: Windows 10\nGraphics card: GTX 1060 6gb (Idk if this is very relevant lol)\n\nBroken: 2.82 BETA, 689a873029b9, master, 2020-01-15 21:07\nWorked: -\n\nBlender 3D shift+RMB \"set 3D cursor\" not sensing strokes made with the grease pencil nor strokes made with the annotation tool\n\nAdd a stroke, hover it with your mouse and press your \"Set 3D cursor\" hotkey combination (by default Shift RMB).\n\n**Expected result**\nWhen doing just this, but when hovering your cursor/mouse over a mesh object instead, the 3D cursor will be set on the surface of the mesh object right where the cursor/mouse is hovering over that mesh/object. As the grease pencil is rather new, Im guessing making this react with the grease pencil strokes were just something that was forgotten when implementing it.", "simulation in new scene causes crash in original scene's simulation\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.19\n\nBroken: version: 2.83 (sub 6)\nWorked: (optional)\n\n**steps**\n- make cube an active rigid body. \n- play simulation\n- then click on new scene button and on full copy.\n- in the new scene, rewind playhead to frame 0\n- run the simulation.\nBlender should crash.\n\nThere are more ways to cause crash when full scene copy is done (see comments), but I guess these will stem from single issue.", "Regression: Attempt to move object with Copy Location (with Offset) crashes\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.67\n\nBroken: version: 3.4.0 Alpha\nWorked: version: 3.4.0 Alpha\n\nA consistent crash when trying to move an object with COPY LOC w/ Offset enabled.\n\n\n1. Open the attached blend file\n\n2. Move the pre-selected object in any direction, you may have to do this two or three times\n\n3. program will crash with EXCEPTION_ACCESS_VIOLATION\n\nThis is happening every time.\n\n```\nStack trace:\nblender.exe :0x00007FF73E03C510 blender::bounds::min_max<blender::VecBase<float,3> >\nblender.exe :0x00007FF73E03BD30 blender::FunctionRef<void __cdecl(blender::Bounds<blender::VecBase<float,3> > &)>::callback_fn<<lam\ntbb.dll :0x00007FF9998C4FD0 tbb::interface7::internal::isolate_within_arena\nblender.exe :0x00007FF7425212F0 blender::CacheMutex::ensure\nblender.exe :0x00007FF73E042390 BKE_mesh_minmax\nblender.exe :0x00007FF73E0409F0 BKE_mesh_boundbox_get\nblender.exe :0x00007FF73E02E940 BKE_object_boundbox_get\nblender.exe :0x00007FF73E02F410 BKE_object_dimensions_get\nblender.exe :0x00007FF73E613520 RNA_property_float_get_index\nblender.exe :0x00007FF73E235050 dtar_get_prop_val\nblender.exe :0x00007FF73E235410 evaluate_driver\nblender.exe :0x00007FF73E168590 calculate_fcurve\nblender.exe :0x00007FF73E156270 BKE_animsys_eval_driver\nblender.exe :0x00007FF73E4464F0 blender::deg::`anonymous namespace'::evaluate_node\nblender.exe :0x00007FF73E446320 blender::deg::`anonymous namespace'::deg_task_run_func\nblender.exe :0x00007FF7425175C0 tbb::internal::function_task<Task>::execute\ntbb.dll :0x00007FF9998CF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FF9998CF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FF9998C4FD0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FF9998CA120 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FF9998CD800 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FF9998CD800 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FF9A5D192C0 recalloc\nKERNEL32.DLL :0x00007FF9A6F72690 BaseThreadInitThunk\nntdll.dll :0x00007FF9A882AA40 RtlUserThreadStart\n```\n\n" ]

[ "Critical crash in Edit mode u sing cursor tool\n[system-info.txt](system-info.txt)\n\nWindows 10 Pro 64 bit\n\nDescription:\nRepeating crash\n\n1.Open blender.\n2.Switch work mode to \"Edit mode\".\n3.Select cursor tool (default selected)\n4.While holding right mouse button click \"S\", and then blender crash.", "Blender 2.8 crashes when trying to Scale the 3D cursor\nFedora 28, Intel CPU, Nvidia 1050 mobile GPU\n\nBroken: 2.8 0c955ed6ea\n\nPressing S to scale while moving the 3D cursor results in a crash\n\n- start blender\n- click on the top toolbar icon (the 3D cursor)\n- drag 3d Cursor around with left mouse button, status bar shows options G,R,S among others\n- click S while dragging\n- (sometimes) continue to drag for a bit\n- CRASH\n\nIt's possible this happens because the 3D pointer doesn't have a scale value? in which case, the hotkey and the prompt for it should just be removed (I was testing to see what happens)\nIt's also possible this is not implemented yet, in which case, nothing to see here, please close this bug!\n", "Crash dragging Cursor tool and hitting S (scale)\nwindows 10, NIVDIA gtx 1050\n\n2.80-dc3b3d94538-win64\n\nblender crash\nuse the cursor tools, drag him along the windos(click&drag), simultaneously\npress the S key.\n(i tried it just becuse the option apear in the info panel.\n\nthank very much for yours hard work", "3D cursor moving Crash\nOperating system and graphics card\nWindows 10 64 Bit NVidia 970\n\nBroken: 2.80 alpha a1689fb091a\n\n\nBlender closed\n\nSwitch to Edit mode.\nKlick left mousebutton and move the 3D cursor arround.\nPress Shift+S and blender closed", "3d Cursor Crash\nBroken: 2.80.21 1a7837596cb\n\nWhile dragging the cursor the lower bar shows that you can press \"s\" to resize it (I was curious because I didn't new that resizing the 3d cursor was a feature), so I tried and it crashed.\n\nOpen new file\nClick and hold to translate 3d cursor\nWhile holding press \"s\".", "Blender 2.80 CRASH\nWindows\n\nrenderer:\t'GeForce GTX 950/PCIe/SSE2'\nvendor:\t\t'NVIDIA Corporation'\nversion:\t'4.5.0 NVIDIA 388.13'\n\n\nBroken: version: 2.80 (sub 21), branch: blender2.8, commit date: 2018-07-30 13:01, hash: 27567a6c760, type: Release build date: 30/07/2018, 07:05\nWorked: (optional)\n\nCRASH\n\n1. Create a Cube\n2. Enter Edit Mode\n3. Right-Click to move the 3dCursor and KEEP Rightclick Pressed while moving the mouse\n4. With the RightClick Pressed - Press S\n5. This is when it crashes .\n", "(Blender 2.8) Left Click, Drag, then Scale crashes Blender\nWindows 10 64-bit\nZOTAC Nvidia GTX 1070\n\n\nBlender 2.8 - September 10th Build\n\n\n\nOpen a new project. Click and hold left click anywhere. Drag while holding. Press S. Software will crash. If you don't drag it won't crash. If you don't hold it won't crash. You must both hold and drag then press S without letting go of the left mouse button.\n", "Crash Using 3D Cursor\nWindows 10, Nvidia 880m\n\nBlender 2.8 Alpha 2\n\nBlender crashes when holding left click with 3d cursor tool activated and pressing 's' key.\n\n\nOn any .blend file, while in object mode click in blank area of modeling space with 3d cursor tool selected. While dragging around the 3d cursor, pres 's' key. This will cause Blender to crash." ]

[ "The Select Hierarchy keyboard shortcut fails to select nested meshes\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: Quadro #2000 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.0\n\nIn the Outliner, Select Hierarchy on the context menu works as expected. The corresponding keyboard shortcut (Shift-RightBracket) fails to select meshes that are nested in an empty.\n\n1. Add an empty, called \"Parent empty\".\n2. In \"Parent empty\" add a mesh (\"Mesh A\") and an empty (\"Nested empty\")\n3. In \"Nested empty\" add a mesh called \"Nested mesh\".\n4. In the Outliner, right-click \"Parent empty\" and select Select Hierarchy. \"Mesh A\" and \"Nested mesh\" are selected, as expected.\n5. In the Outliner, select \"Parent empty\". In the Viewport, press Shift-RightBracket. **\"Nested mesh\" is not selected.**\n\n\n\n[Select_Hierarchy_keyboard_shortcut.blend](Select_Hierarchy_keyboard_shortcut.blend)\n\n", "Scene pre/post handlers are called for each main loop iteration\nTitle says it pretty much everything. just execute this in py console:\n\n```\ndef cb(sce): print(\"foo\")\n\nbpy.app.handlers.scene_update_post.append(cb)\n```\n\n…and you'll get endless 'foo' lines in your console, even when Blender is hidden and all.\n\nIssue is, main loop calls (through `wm_event_do_notifiers()`) `BKE_scene_update_tagged()`, which unconditionally fires those pre/post scene update handlers.\n\nNot quite sure what's to be done here really, not familiar at all with this area of the code.", "Blender crashes when using an EnumProperty to set Space\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.0 Alpha\n\nIn this script, I'm using `context.area.type` with a property to change the Space type. It caused Blender to crash.\n\n1. Running this script from the Text Editor (below)\n2. Switch to Graph Editor or Dope Sheet\n3. Click new property buttons on right of Header\n\n```\nimport bpy\nfrom bpy.types import PropertyGroup\nfrom bpy.props import EnumProperty\n\ndef set_space(self,value):\n context = bpy.context\n\n for area in context.screen.areas:\n if area != context.area:\n break\n\n override = {'region': area.regions[0]}\n\n if value == 0:\n context.area.type = \"GRAPH_EDITOR\"\n elif value == 1:\n context.area.type = \"DOPESHEET_EDITOR\"\n\nclass SWITCH_PG_switcher_props(PropertyGroup):\n space: EnumProperty(\n name=\"Space Target\",\n description=\"Switch to space\",\n items=(\n ('GRAPH_EDITOR', \"\", \"Graph Editor\", \"GRAPH\", 0),\n ('DOPESHEET_EDITOR', \"\", \"Dopesheet Editor\", \"ACTION\", 1),\n ),\n default='GRAPH_EDITOR',\n set = set_space\n )\n\ndef draw_enum(self, context):\n layout = self.layout\n scene = context.scene\n layout.prop(scene.switcher, \"space\", icon_only=True, expand=True)\n\ndef register():\n bpy.utils.register_class(SWITCH_PG_switcher_props)\n bpy.types.Scene.switcher = bpy.props.PointerProperty(type=SWITCH_PG_switcher_props)\n bpy.types.DOPESHEET_HT_header.append(draw_enum)\n bpy.types.GRAPH_HT_header.append(draw_enum)\n\ndef unregister():\n del bpy.types.Scene.switcher\n bpy.types.DOPESHEET_HT_header.remove(draw_enum)\n bpy.types.GRAPH_HT_header.remove(draw_enum)\n bpy.utils.unregister_class(SWITCH_PG_switcher_props)\n\nif __name__ == \"__main__\":\n register()\n```\n\nSetting the `context.area.type` in Python seems to cause others' scripts to crash:\n[Performing a Translate Operation after switching the context area crashes Blender ](1164763)\n", "Poly Build Tool: Editing poly build keymap causes poly build tool to not work properly\nWindows 10\nGeForce GTX 1080\n\nv2.81 Alpha\nDate: 2019-09-10 15:07\nHash: 8f55794c0e80\n\nWith a default blender open, edit the poly build keymap. The tool doesn't work properly with the new key assignment. \n\nFor example, switch 'face at cursor move' to shift + left mouse, and 'poly build delete at cursor' to ctrl + left mouse.", "Usability issues with old group operators\nBroken: 2.79, 2.8x\n\n\nWhen using linked objects, all the Group commands [Remove From Group / Remove From All Groups / Add Selected To Active Group / Remove Selected From Active Group] except [Create New Group (Ctrl G)], do nothing\n\n\nIn my particular usage, I link a scene from one file into my main scene, afterwards I create a [New scene with Link Objects] from the Linked scene. \n\n [GroupsFail.blend](GroupsFail.blend)\n [GroupsFail-MainScene.blend](GroupsFail-MainScene.blend)\n\nI can Create Groups by selected objects, and I can remove the group from the object in the Object Properties tab but the only way to remove multiple groups is through python script or deleting the group itself.\n\nSince this report is used to more generally handle issues with the old group operators and other reports have already been merged here, will also list these:\n- Ctrl G is placing new collections inside \"Orphan Data\"", "Node Operators: Add Selection input node\n_No response_\n\n", "Non Uniform Scaling possible when Camera is parented\nOperating system: Win 10\nGraphics card: Geforce 2070\n\nBroken: 2.82-3.2\n\n\nIf you parent a Camera to an Object and scale the parent non uniformly, the Camera View gets scaled and distorted too.\n\n\nCreate an Empty, parent the Camera to an Empty, scale the Empty. The Camera View gets distorted.\n\n[blender_camera_scaling_bug.blend](blender_camera_scaling_bug.blend)", "Crash when trying to delete particle system or undo CTRL+Z\nBroken: version: 2.82 (sub 7)\nWorked: 2.79\n\nAfter updating from 2.79 to 2.8 this file file crashes when I try to delete the Plane. When I go to Edit mode, trying to rotate, grab or do anything it crashes. Plane has a particle system on it. When i try to delete it crashes too. Basically if i touch the Plane in any way it will crash and there is no way to get rid off it. Cannot delete the particle system either. Also, when i try to undo the last step CTRL it also makes it crash.\n\n- Open the attached file\n- Try to delete the plane or its particle system \nCrash.\n\n[T74786_2-83.blend](T74786_2-83.blend)\n\n\n\n", "Auto keying not keying for child bones and not transformed\nOperating system: Windows 10\nGraphics card: RTX 3060 Ti\n\nBroken: (3.1.0 c77597cd0e15 master 2022-03-08 18:16)\n\nWhen I tried to auto key multiple bones by select these and hit G > Mouse1 only the top of parented bone is keyed and nothing for anything else \n\n・Add armature\n・In edit mode -> extrude bone\n・Enable Auto keying\n・In pose mode -> Select All (A) ->Move/Rotate/Scale(but not make any transform, just apply) to add a keyframe\n\nTest file with first 3 steps included:\n[#96473.blend](T96473.blend)", "Keyed Particle Duration Crash\nOperating system: Windows 7 64\nGraphics card: nVidia GeForce GTX 970\n\nv.2.80.60\nDate: 2019-05-10 22:21\nHash: de9d846353bd\n\nBlender will sometimes crash when a keyed physics type particle system references more than one particle system. This usually occurs when attempting to change the duration setting of the second particle system.\n\nIn the provided file, change the duration setting in the particle system settings. Blender will crash.\n\n[Keyed Particle Duration Crash.blend](Keyed_Particle_Duration_Crash.blend)", "\"Emulate 3 Button Mouse\" conflicts with ALT left click, for instance to disable multipe modifiers on multiple objects\nOperating system: mac os ventura 13.4\nGraphics card: Apple M1 Max GPU 32cores Metal 3\n\n3.6\n\nEnabilng setting \"emulate 3 button mouse\" conflicts with ALT key left click operations, for instance when interacting with a panel. \nIn this case in the properties panel when ALT left clicking on an icon to disable multipe modifiers on multiple objects.\nthe cursor turns into a navigation cross and doesnt execute a click.\n\n\n\nGo to the settings menu/input - Enable \"Emulate 3 Button Mouse\" option.\n\nAdd a modifier to an object.\nDuplicate the object multiple times.\nSelect all the duplicated objects.\n\nGo to the modifiers panel.\n\nWhile holding down the ALT key, attempt to disable the modifier on all the selected objects by left clicking on the \"Display Modifier in Viewport\" icon or any other relevant modifier option.\n\nYou will find the cursor overtaken by the navigation mode, the cursor turning into navigation arrows and disabling a regular click operation.\n\n", "Error message and unable to edit a skeleton when using a library override armature\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: version: 3.5.0\nWorked: No idea.\n\nBlender generates an error when you attempt to enter edit mode on a library overriden armature. There's no description of what the error is or if there is some extra step to make it work. Hopefully this isn't one of those, \"works as intended issues,\" because then there's no way to make multiple characters share an updateable base rig.\n\n\n- Add an armature, enter edit mode, extrude out a few bones.\n- Save the file and open a new scene.\n- Link in the armature you just created.\n- Make the armature a library override.\n- Try to enter edit mode. Should get an error.\n\n", "Allow menus to be added referenced to another menu \n## Motivation\nAddon developers cannot add menus on specified place. \nThey are always added at the end (append), or at begin (prepend)\n\n## Analysis\nAfter a look at current code, i think a good approach should be allow to define a reference label on append function.\n\nWhen drawing the menu layout, a new RNA call should be done on a new function class, that should then execute the referenced drawing calls on py scope.\n\nIssue with this development is that for any but menu drawn two python calls are performed to allow addons draw before and after itself. Maybe this could be solved checking in C scope if it has post/pre functions added.\n\n## Development steps\n* {icon check-square-o color=green} Draw after label\n* {icon check-square-o color=green} Draw before label\n* {icon check-square-o color=green} Give warning if not used (label not found)\n\n## Know Issues\n* Issues with enums, if shown as submenu it's not working. Is shown on self menu, the label is the operator name, and considered a block\n\nFeedback is welcome!\n\n\n\n\n", "Intersect Knife: options 'self interesect' + 'cut' result in too many loose parts\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: Intel(R) HD Graphics 630 Intel 4.5.0 - Build 26.20.100.6912\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\n[2 intersected objects should do an intersection on each one but keep 2 objects not 4 ]\n\n[[2 intersected objects>ctrl+J>'EDIT'>select all verts>intersect (knife)>options self intersected and cut] [which should be equivalent to 1 object selected>interect (knife)>options select/unsel and cut >>given 1 intersection but still 2 initial objetcs]]\n\n", "pan gesture behaves randomly\nOperating system: macOS-12.5-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.101\n\nBroken: version: 3.4.1\n\nHi,\n\nOn external monitor, two fingered pan gesture on a Mac touchpad behaves randomly between orbit and zoom\n\nThis happens when you use the default Orbit & Pan navigation settings. Most of the time it works correctly: You make a two fingered pan gesture and the camera view orbits. But once a while a two fingered pan gesture will zoom the camera view (instead of obiting). As if you are making a pinch gesture (which you are not). I have disabled the pinch gesture on the system level to rule out the possibility that the pinch gesture interferes with the pan gesture. It is not a pinch gesture conflict.\n\nTo me it looks like following is happening. It seems like that although you have multi-touch gestures enabled, the two fingered pan gesture is sometimes handled as if multi-touch gestures is disabled. (When you go to Input > Touchpad preferences and disable Multi-touch Gestures then the camera view always zooms when you make a two fingered pan gesture.) \n\nThis happens about once every minute, and sometimes even a couple times a minute. It's random. When it happens and you don't end the gesture (you keep your fingers down on the touchpad) then you can continue to zoom in and out by moving two fingers left and right. So once the pan gesture handler chooses to zoom instead of to orbit then it is locked in this zoom behaviour until you end the pan gesture.\n\n\nWorkaround: 103973#issuecomment-79513\n\n" ]

[ "Instant crash upon parenting two separate armatures in the edit mode.\nOperating system: Windows 10, 64 bit\nGraphics card: Nvidia Geforce GTX 1070 8 gb\n\nBroken: \n2.8 b936d7b16c62 2019-03-31\n2.8 a813e259d630 2019-04-02\n\nWorked: (optional)\n\nHello there, wonderful people of Blender development community! \n\nI think the ability to edit multiple objects in the edit mode presented a set of new bugs. For example, all the time I create two separate Armatures with bones in them I can select them in edit mode and try parenting them. Which will cause an instant crash with no notification or errors. Here's the video of the bug: \n\n[2019-04-03_14-28-10.flv](2019-04-03_14-28-10.flv)\n\nI guess if I'm not allowed to do that I must get the error message and not crash instantly.\n\n\n\n1) - Create a scene \n2) - Create two separate Armatures sets with one bone in each\n2) - Select both Armatures \n3) - Go to edit mode with both Armatures selected\n4) - Select both bones from separate Armatures\n5) - Press Ctrl P and parent them\n6) - Crash with no error or notification\n\nThank you for your hard work!" ]

[ "Cycles: Caustics don't work when volume boundary between casting and receiving surfaces\nOperating system: MacOS, Windows\nGraphics card: Apple Silicon, Nvidia RTX3090\n\nBroken: 3.5.1, 3.6 Alpha\nWorked: -\n\nWhen there is a volume boundary between a caustic casting and receiving surface, the caustics vanish. \n\nOpen the attached .blend file. There is a sun, an ocean surface, and a sea floor. There are 4 sections: \n- On the right, there is no volume and the caustics appear\n- the two middle sections contain volumes where either the top or bottom boundary of the volume is between ocean and floor, and no caustics appear\n- on the left the volume boundaries are above the ocean, and below the floor, and caustics are visible\n\n", "\"Join Meshes\" changes the material slot assigned to faces of the original mesh\nOperating system: MacOs BigSur\n\nBroken: 2.91.2\n\nSee video:\n[Screen Recording 2021-02-12 at 12.04.24.mov](Screen_Recording_2021-02-12_at_12.04.24.mov)\n\n1) Create Mesh ONE with 2 material slots, with the same material, each slot assigned to different mesh island in mesh. \nSo faces on mesh islands have different material_indexes, although both use the same material (in different material slots)\n2) Create Mesh TWO without any materials\n3) Join Mesh TWO into Mesh ONE\n4) Check material indexes in resulted mesh. They wil ALL point to first material slots.\nFaces, that was linked to second material slot initially will be \"relinked\" to first slot, so theier material_index attributes are altered/messed \n\n**Expected behaviour:**\nFaces linked to second material slot initially should stay with the same material_index, join action has no reason to change them\n\n", "Creating a new image texture in shader editor does not appear among the texture slots\nWindows 10 Pro 64bit, AMD Ryzen 5 2600, 16Gb RAM\nNVidia GTX 1060 6Gb\n\nWindows 10 Pro 64bit, AMD Ryzen 5 1600, 8Gb RAM\nNVidia GTX 1050Ti 4 Gb\n\nBlender 2.8 Beta 91a155833e59\n\nIf you are texture painting in one window and you modify the shader by adding/removing an image texture to it in another, it won't appear/disappear from the texture slot list until you change to edit mode (or anything else) and back. Another issue is that this way you can paint onto an image texture that is no more connected to the shader and it doesn't disappear even if you do the \"edit mode and back\" (TAB x2) trick.\n\nSteps: Add an image texture to the shader in the attached file, connect to the roughness for example, create a new texture for. It won't be in the texture slots list. Now TAB x2 in the 3D view, slot appears. Paint on it, delete the node from the shader. It is still in the list. TAB x2. It is no more in the list, but the 3D view still shows the texture.\n\n[test1.zip](test1.zip)", "Cycles shading nodes missing updates when editing muted nodes\nOperating system: Linux-5.13.0-28-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.103.01\n\nBroken: version: 3.2.0 Alpha\n\nConnecting, desconnecting or muting separate node doesn't properly update \n[cycles update.m4v](cycles_update.m4v)\n\n[cycles update bug.blend](cycles_update_bug.blend)\n\n1- Switch viewport to render preview\n2-Connect the separate node. \n3-mute unmute, connect disconnect as its shown in the video description. ", "Cycles: improve subsurface scattering for some meshes shapes\nIn cycles-x, the Christensen-Burley dipole approximation was fully replaced by random walk.\n\nHowever for some mesh shapes results are not as good, or could be improved regardless.\n* Self-intersection meshes give dark borders. Counting back/front facing intersections could help determine inside/outside better to avoid such artifacts.\n* Sharp corners \"glow\", which may be physically correct but is not necessarily desired. It may be possible to add user control with an option somehow, using some minimum throughput\n\nIf these can't be solved, we may consider adding back the dipole approximation, but since path splitting and the memory usage is a difficult fit for the new kernels we'd really like to avoid it.", "Color Attribute not affecting \"bump\" shader node\nOperating system: Windows 10\nGraphics card: RTX 3080\n\nBroken: 3.50 beta (03 March 23)\n\nUsing a mesh's color attribute as an input to the \"bump\" shader node does not result in the expected behavior. Using the color attribute for practically every other value (including the normal map node, roughness, metalness etc.) works as expected.\n\nCreate a mesh and paint a color attribute with, for example, a black-and-white pattern.\nGive the mesh a material and add the color attribute.\nPlug the color attribute into the \"height\" input of the bump node, and view the result. There is no difference compared to the \"height\" input being left disconnected.\n\nOther inputs (such as a noise texture) result in the expected behavior.\n\n", "Cycles viewport unneccessary update calls\nOperating system: Ubuntu 20.04\nGraphics card: Nvidia MX150\n\nBroken: 2.83.17, 8a8c61190b06, Beta and 2.90.2, 8f21f867e715, Alpha\nWorked: 2.82a official\n\n[#76970.blend](T76970.blend)\n\n* set scene renderer to Cycles\n* configure two viewports and a shader editor in a single screen\n* set one viewport to cycles preview render\n* set second viewport to workbench texture color\n* add two image texture nodes and load them with different textures\n* by clicking on the texture node both viewports are redrawn.", "shadow catcher data missing from diffuse passes when transparency is disabled, and shadow catcher object is masking environment when transparency is enabled (can't recombine moving scene envronment)\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 526.98\n\nBroken: version: 3.5.1 Release Candidate\n\nshadow catcher data missing from diffuse passes when transparency is disabled, and shadow catcher object is masking environment when transparency is enabled (can't recombine moving scene envronment). This makes it impossible to rebuild the image unless the shadow catcher pass is used, which is not suitable when using vector motion blur.\n\nOpen below blend file (choose to load ui)\ngo to scene 'no transparency'\nin the compositor press f12 to render a frame\ncompare the combined image and the rebuilt image\n\ngo to the second scene 'transparency'\nin the compositor hit f12 to render again\nthis time the environment is masked out by the shadow catcher, making it impossible to use the rendered environment (which is necessary if the camera is animated)\n\nrequired behaviour - \n\nshadows appear in diffuse passes when transparency is disabled, so the rebuilt image matches the viewport.\n\nshadow catcher object doesn't mask out the environment pass when transparency is enabled (so the environment can be added back to the image when the camera is moving (backplate cant be used here))\n\n\n\nThe only possible solution currently is to always render with transparency enabled, and then render your scene a second time with only the environment visible, which causes issues when rendering via scripts.\n\n", "Sculpt mode: hide masked do not work when modifier enabled, but hidden part works with modifier.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.0.0 Alpha\n\n\nAs title said. Hiding works with modifier but not work simultaneously as «hide masked»\nFor example, edit mode work well with mirror and hiding, hiding by faceset works well too...\n\n**Steps to Reproduce**\n- Open .blend file\n- Paint with mask brush\n- {nav Sculpt > Hide Masked} \n\nTest File:\n[#92372.blend](T92372.blend)\n[2021-10-20_18-49-00.mp4](2021-10-20_18-49-00.mp4)\n", "Backface culling is not working for mask in edit mode\n2.80 (sub 45), 893fa598319e, 2019-02-24 22:02\n\nAnd wireframe too heavy.\n\n\n\nIdk if it is bug or intended, but it will be nice, if it will work like on the 2nd or 3rd picture: much more readable.\nAnd when x-ray is on, wireframe on the background will be faded too, accordingly slider-setting.\n\n\n\n\n", "Toon shader glossy removes direct and indirect light of diffuse shaders at low sizes with path tracing integrator\nOperating system: Ubuntu 18.04.1\nGraphics card: RX 480\n\nBroken: blender-2.80-a1ae04d15a9f\nEither CPU or GPU as rendering device\n\nWorked: set integrator to Branched Path Tracing\n\n\nToon BSDF glossy component at pretty low sizes nullifies(or removes) diffuse shader and toon shader diffuse direct lighting, at low or medium sizes it removes indirect lighting. This only occurs when the integrator is set to Path Tracing\n\nIn Cycles render engine\n1) subdivide initial cube, then set to smooth shading, set light to 1000 strength\n2) add a plane (to cast indirect light)\n3) add material to the cube, add mix shader with a fac of 0.1, in the first slot use a diffuse shader (or toon shader with diffuse component) with a color, now for the second slot set a toon shader with glossy component, set its size to 0.001\n4) set to render mode, the sphere should be colored but is dark with only the glossy component, when you change the glossy toon size to 0.003 the diffuse part appears but pretty dark, at 0.010 only the direct part of the diffuse shader is fully illuminated, at 0.030 the indirect part is still dark and at around 0.400 the indirect part is fully illuminated, all these were compared to branched path tracing\n[Glossy_toon_shader_bug.blend](Glossy_toon_shader_bug.blend)\n", "Workbench: Shadow Bug\nOperating system: MacOS Catalina\nGraphics card: Intel Iris Pro 1536 MB\n\nBroken: 2.83\n\nWhen trying to add a shadow, all screen \"cracks\" into different-shadow pieces\n\n\n\nViewport Shading> \"checkbox\" Shadow ", "Wrong SubSurf modifier results with inconsistent normals\nOperating system: Windows 10 Enterprise Version 1803\nGraphics card: 2 x GTX1080 ti\n\nBroken:\nblender2.8, 2018-12-05\n\nFrom latest builds subsurf modifier make a wrong subdivision of geometry as displayed in the attached files.\nBefore 18/20 November the behaviour of subsurf modifier was ok.\nI am sure because I did an animation on that date and the subdivided vase was ok.\n\nBest regards,\n\nUmberto\n\n\n\n\n\n[ceramic.blend](ceramic.blend)", "Inconsistency of some tools/operators (e.g. Subdivide) in regards to hidden geometry\nBroken: all or none\n\nI'm not really sure this is even an issue, though it feels like it. As far as I understand it, mesh editing tools are not supposed to affect hidden geometry. But the Subdivide tool does. Consider this mesh:\n\n\nIf you hide the middle face...\n\n\n...and then Subdivide the two remaining faces...\n\n\n...Subdivide will create new edges cutting through the hidden face, also creating inconsistent selection (i.e. selected edges in face select mode). Shouldn't the operator check that that face is hidden and not create edges through it?", "Subtype for input value nodes\n\nThe user may be at a loss when entering a rotation radians value into some built-in node, copies it into the input value node, and now these are degrees.\nThe explicit way of specifying the sub type must be an option for all value input nodes that it can have.\nThe type list option should be available in the N-Panel so as not to clutter up the node space.\n\n" ]

[ "Eevee: Second SSS shader does not work in Mix node\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.93.0 Alpha\nWorked: ?\n\nUsing Eevee, when combining two Subsurface Scattering shaders in a Mix node, the second SSS shader does not work.\n\n- In a default Blender scene, delete the cube, add a Suzanne monkey and add a few subdivs for a better look.\n- Add a material to Suzanne and delete the default Principled Shader.\n- Add a Subsurface Scattering shader and copy it.\n- Plug both SSS shaders into a Mix Shader node.\n- Change the viewport shading mode to Rendered.\n- Change the Mix node's value to 0. The first SSS shader becomes fully visible.\n# Change the Mix node's value to 1. The second SSS shader does not become visible.\n\n\n\nNote 1: If you change the first SSS shader to a non-SSS shader, the second shader works, so it appears that two SSS nodes can not be mixed when using Eevee.\n\nNote 2: The same issue seems to occur when using an Add node in stead of a Mix node.\n\nNote 3: it does work in Cycles." ]

[ "Eevee Alpha Blend artifacts in lookdev and VR with Viewport Denoising enabled\nOperating system: Windows 10\nGraphics card: Nvidia GTX 970\n\n\nBroken: version: 3.0.0 Alpha, branch: master (modified), commit date: 2021-07-26 13:54, hash: 3059853732d7\nBroken: version: 2.93.1, branch: master, commit date: 2021-06-22 05:57, hash: 1b8d33b18c2f, type: release\nBroken: version: 2.83.16, branch: master, commit date: 2021-06-15 08:46, hash: 051567553d29, type: release\nWorked: Unknown\n\n\n\nNavigation within a viewport (such as panning / orbiting the scene) with objects that have a material with Alpha Blend enabled results in visual artifacts for those materials, seemingly regardless of the alpha channel itself (i.e. it affects materials with an alpha of one).\n\nIn my case I first saw the artifacts while using the VR mode in blender where these artifacts feel extremely exaggerated in some cases. I have an environment with foliage billboards that appear like they are slightly head-locked and they sway around a lot when moving in VR. I later realised I could reproduce similar artifacts outside of VR and then realised that disabling denoising avoided the artifacts in the editor and in VR.\n\nIt seems that Eevee's denoising introduces some kind of latency related to how the alpha blending is handled which results in temporal, visual artifacts while the viewport's perspective is changing.\n\nI'd say the reproduction below probably doesn't represent how significant the artifacts feel in VR but I think it highlights the same underlying issue (the same two workarounds that affect the editor, also help in VR).\n\nTurn on Viewport Shading: Material Preview or Rendered\n\nDelete cube\nAdd plane\nAdd material (default principled BSDF is fine, though in my case I'm working with emissive shaders and I have a modified build of Blender supporting equirectangular image textures that let me access their alpha channel)\nSet material Base Color to an Image Texture.\nCreate a new image and generate a Color Grid to set as the base color. (with or without Alpha)\nSet material blend mode to 'Alpha Blend'\n\nGyrate/pan/orbit the viewport view. (Note the artifacts are related to the viewport navigation that needs to move around, and just moving/rotating the plane itself won't show artifacts)\n\n(For the Color Grid the artifacts mainly look like shimmering, and glints on the white letters. That's quite different to the original artifacts I saw in VR for my use case but still it seems to represent the same underlying issue.)\n\nUnder render properties, toggle Viewport Denoising and perform the same viewport movements to compare.\nWith Viewport Denoising back on, switch to using Alpha Clip instead of Alpha Blend and again compare.\n\n[eevee-denoise-artifacts.blend](eevee-denoise-artifacts.blend)", "EEVEE: Mixing between a fully transparent and holdout shader yields a half transparent result\nOperating system: macOS-10.13.6-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 580 OpenGL Engine ATI Technologies Inc. 4.1 ATI-1.68.25\n\nBroken: version: 3.2.0 Release Candidate\nWorked: I thinck the issue is for \"all\" versions of blender\n\nHoldout shader having issue with alpha (semi transparency).\n\nPut a image plane with exr image that contain alpha. When using a holdout shader mixed with another shader (transparency shader for exemple) the alpha of the image sequence is used for the mix of the 2 shaders. It should have a complete transparent image with the holdout of other object in the scene. The issue is that there is a dark outline for the semi transparent. Holdout have issue dealing with semi tranparency ? or maybe something else.\nSee the attached image to understand the issue.\n\n\n", "EEVEE and Workbench missing animatable properties\nOperating system:\tWindows 7\nGPU:\t\t\tRadeon Vega 64, Divider version: 19.12.2\nCPU:\t\t\tIntel Core i7-5960X\n\n\nWorking-ish*:\t\t`blender-2.79b-f4dc9f9d68b-windows64` (official release) \nBroken:\t\t\t`blender-2.80-f6cb5f54494e-windows64` (official release) and later\n\n\nEEVEE and Workbench's render properties (`Properties Editor --> Render Properties`) are, for the most part, not animatable.\nThis is in contrast to Cycles which, *for the most part*, has all settings set to be animatable.\nBack in 2.79b the OpenGL \"Workbench\" Render options were supported:\n\n\n\nSome of the panels like Film, Simplify and Colour Management (which exist in all render engines) work as expected from 2.7X to 2.8X.", "Eevee Attribute Rendering is Having Problem with Scene Referred/Open Domain Data\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce MX130/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.2.0 Alpha\nWorked: None\n\nEevee attribute rendering fails to render scene referred/open domain data\n\n\nAnd here is Cycles that does it properly:\n\n\nBased on the attached .blend file\n\n1. Turn on Rendered view, see how Eevee clips all values above 1 and the sweep end up with the six colors on the right side and it doesn't go to white at all (The Notorious Six)\n2. Switch to Cycles and see how Cycles renders open domain data properly\n\n[Eevee Attribute Rendering Open Domian Bug.blend](Eevee_Attribute_Rendering_Open_Domian_Bug.blend)\n\n[Eevee Attribute Rendering Open Domain Bug.mp4](Eevee_Attribute_Rendering_Open_Domain_Bug.mp4)", "Eevee render display error in orthographic camera mode: Mesh material color changes during scaling \nOperating system: Win7 Home Premium 64-bit, SP1\nGraphics card: Nvidia GeForce GTX 750\n\nBroken: 2.80, c9df453ce70b, blender2.8, 2018-12-21\n2.80, d96596de294e blender2.80, 2018-12-31\n\nWorkspace: Layout\nObject mode\nDisplay rendered view: Eevee (other rendered views with Cycles, etc. do not have this problem)\nCamera orthographic view (!! in perspective or panoramic view, problem is not there)\nCamera type: orthographic\nLight: Sun shining perpendicular onto objects\n\nThree cube mesh objects with individual materials\n- Principled BSDF\n- Base Color (either Red, Green or Blue)\n\nThe *scaling** of one or more objects changes the color appearance in the camera view *and* in rendering.\n\n**scaling*can happen either:\n- pressing 's' key and move mouse\n- changing scale values in transform tab\n\n - > I thought there seemed to be an dependency on scale values. For example, scaling to 1.0 would always be a lighter tone than in a different scaling factor. But simply pressing 's' to scale and escaping can change the color tone of the mesh as well.\n\nThe direction of the sunlight makes of course a difference when viewing the objects, but after changing sunlight angles there is still the color changes during scaling.\n\nBlend file attached.\n\n\n - Select Display render preview \n - Simply choose one or more cubes \n # scale these slowly and the color will change from a fade color to a strong color infrequently\n\nIt seems the problem is the camera orthographic mode!?\n\n[Scale-Color-error.blend](Scale-Color-error.blend)\n", " Render Face Sets and Mask as contour lines\nFace Sets were added to control the visibility of all paint modes and currently, they are rendered as colors (which is the easiest way to debug them and work with them from sculpt mode). This won't work for vertex paint and texture paint (and any other mode that works with color), so a different rendering needs to be implemented that does not affect the color of the surface as an option these painting modes. \n\nIf possible, this should be implemented for the current sculpt mask (add an option to render it as a contour line, similar to the lasso selection) as the same mask implementation and rendering will also be used for vertex paint.", "Crash when selecting line art\nOperating system: Linux mint 21. Blender installed using flatpak.\nGraphics card: See attached system_information.txt\n\nBroken: version: 3.5.1\n\nBlender Crashes.\n```\nbacktrace\n\nPython backtrace\nFile \"/app/blender/3.5/scripts/startup/bl_ui/space_view3d.py\", line 704 in draw\nFile \"/app/blender/3.5/scripts/modules/bpy_types.py\", line 903 in draw_ls\n```\n\n```Py\nbpy.ops.outliner.item_activate(deselect_all=True) # Operator\nbpy.ops.outliner.item_activate(deselect_all=True) # Operator\n```\nWith attached blend file, in the outliner, select cameras, line-art then try to change the color of the background.\n\n", "Intel Mac Metal: UV Display Stretch Angle Fault Colors\nOperating system:Mac OS 13.3.1\nGraphics card:Radeon Pro Vega 48 8GB\n\nBroken: 3.5.0 - 3.5.1 - 3.6.0 alpha\nWorked: NA\n\nUsing Intel Apple, with GPU Backend Metal enabled\nUV Editing, when you turn on Display Stretch: Angle\nCorrect angles will show faulty colors.\nWhen I switched to OpenGL and restart.\nThe Stretching shows properly as it should.\nI tested using default blender with no addons. Clean install with Default Cube and the error persisted on the versions noted above.\n\n\n\n- Activate blender with GPU Backend: Metal\n- Use a default Cube, within the UV Editing tab\n- In Overlays menu, choose Display Stretch, and pick Angle.\n- Select all the Faces. The colors will appear faulty and extremely inaccurate.\n\n\n", "Sculpt: Paint tool cavity custom curve crash\nOperating system: Windows 10\nGraphics card: NVDIA RTX 2060 Super\n\nBroken: 3.4, 3.5\n\nSteps to reproduce the issue:\n - Switch to Sculpt Mode or open default sculpting template file\n - Choose the Paint Tool\n - Click on Brush dropdown\n - Enable Cavity by clicking on Advanced >> Cavity in the \"Brush\" dropdown\n # Enable Custom Curve under Advanced >> Cavity in the \"Brush\" dropdown\n\n\n\nResult:\nNo **Custom Curve** is shown as in other Sculpting Tools & after trying to paint with the **Paint** tool setup this way blender instantly crashes.", "Default Value for Viewport Display Color's Alpha Wrong.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.09\n\nBroken: version: 2.92.0 Alpha\nAlso tested on 2.83.8 and same problem.\nWorked: It do not think this has ever worked\n\nWhen reverting to the Default Value for Viewport Display Color the alpha gets down to 0.8 making the object semi transparent.\n\nOpen up a new scene.\nSelect the cube.\nGo to the materials tab.\nScroll down to \"Viewport Display\"\nRight Click on \"Color\"\n\"Reset to Default Value\"\nResult: RGB values go back to the default, but Alpha goes to 0.8\n\n\nThank you so much !", "Broken edge lines with boolean and freestyle\nOperating system: Linux-5.0.7-200.fc29.x86_64-x86_64-with-fedora-29-Twenty_Nine 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.56\n\n\nBroken: version: 2.80 (sub 58)\nWorked: (optional)\n\n\nI'm seeing two different issues:\n(1) There is a broken line in the edge (see attached render). Rotating the object for a different angle, changing the boolean by moving one of the cylinders, etc. still causes broken lines, sometimes on multiple edges. Also, using the simple surface (without the toon shading setup of Shader-to-RGB and gradient) still causes the broken line. \n(2) The second possible issue is that while Eeevee renders correctly otherwise, Cycles also has the broken line and in addition seems to be rendering incorrectly for the toon shade effect - i.e., all solid color vs. multiple colors expected, see attached render. \n\nCreate a cube and two cylinders. Add two boolean modifiers to the cube to subtract the two cylinders. Add a toon shading setup (shader to RGB and gradient) and free style to render lines for edges. Render with Eevee and Cycles (see above). File attached. [sekhar-04-20.blend](sekhar-04-20.blend)\n\n", "Crash when switching to Sculpt mode\nOperating system: \nLinux bu-dvl-2 5.13.0-51-generic #58~20.04.1-Ubuntu SMP Tue Jun 14 11:29:12 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux\n\nGraphics card:\n0:02.0 VGA compatible controller: Intel Corporation UHD Graphics 620 (rev 07)\n\nBroken: 3.2.0\n\ncrashed when switching to Sculpt mode\n\n- Open alien.blend (attached) \n- Switch to Sculpt mode\n\n[alien.blend](alien.blend)\n\ncrash.txt: \n[alien.crash.txt](alien.crash.txt)", "The Hair Particles interprets the color from the newer vertex domain and float color types wrong\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.3.0\n\nI am trying to modify hair color using color attribute (called as vertex color in previous versions). But hair particles simply ignores the color attribute. I just plugged in a musgrave texture and that works as expected but color attribute does not. I have attached a screenshot for a better understanding.\n\n\n\n\n", "Cycles vertex baking seams\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: GeForce GTX 560 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 3.1.0\n\nThere are seams between the vertices when I try to bake lighting into vertex colors using Cycles, even when the vertices are merged.\n\nHere's my .blend file:\n[baking test.blend](baking_test.blend)", "Texture Paint: Color Palette operator issues with Undo/Redo -- Vertex Paint is fine though\nOperating system: Darwin-15.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: ATI Radeon HD 5770 OpenGL Engine ATI Technologies Inc. 4.1 ATI-1.42.15\n\nBroken: version: 2.83.0\n\n\nIn Texture Paint mode (contrary to vertexpaint mode or gpencil), the palette operators dont work with Undo/Redo\n\nFrom the default startup file:\n- go to vertexpaint mode\n- (a) in the Active Tool tab in the Properties Editor, under Brush Settings > Color Palette, add a palette\n- (b) add a color to the palette\n- (c) Undo --> should be working [color is removed again]\n- so far everything is working\n- go to texture paint and do the same a, b, c --> color is not removed again, Undo is not working\n\nThis is what is probably the cause for the following more specialized original report (but also there the same thing works fine in vertex paint mode)\n\n## original report\n\nWhile in texture mode users press S to activate a color picker (not sure why this isnt the GUI nor the tools palette?)\nWhen its active, we have the options menu (used to be F6 i believe) There is this option \"Add to Palette\".\nThis however always adds colors. When you uncheck it, it will enable itself the next time your use the S key and pick a color.\n\n01. Open the attached scene.\n02. Start texture paint mode\n03. Use S key to enable color picker\n04. Options menu (bottom left) shows \"Add option to palette\" > Deactivate it.\n05. Use S key again > Notice colors are still being added but option is unchecked.\n06. When you let go of the S key, its checked again\n\n\n\n\n[IMport Palette - Blank.blend](IMport_Palette_-_Blank.blend)" ]

[ "Vertex Color doesn't show up (in Eevee using the Vertex Color shader node or in workbench vertex shading mode)\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.09\n\nBroken: version: 3.0.1 Release Candidate\nWorked: 2.93.7\n\nThe fail in `Material Preview` viewport shading mode as well as Eevee `Rendered` mode was caused by 03013d19d1 \nThe fail in `Solid` viewport shading `Vertex` mode (both object mode and editmode) is probably not considered a \"fail\", see #83891 (Vertex Colors not displayed in object/edit/sculpt mode (if Sculpt Vertex Colors are enabled from Experimental))\n\nIf I create a simple shader using only vertex color input plugged in the surface output in 3.01 the result is black in Eevee., but works properly in 2.93.7\n\n\n\n\n\n\n\n[VertexColorEevee.blend](VertexColorEevee.blend)\n- open blender and enable `Experimental` > `Sculpt Vertex Colors`\n- open .blend\n- change `Solid` viewport shading mode to `Vertex` : no vertex colors\n- change to `Material Preview` viewport shading : black\n- change renderer to Eevee and go to rendered viewport : black\n\nYou can open this scene in both 2.93.7 and 3.01 and switch from Cycles to Eevee to see the difference\n\nThanks in advance ;)\n\n", "Vertex Paint is not showing when you set viewport mode color to vertex\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.29\n\nBroken: version: 3.2.0 Alpha\nWorked: version: 3.0.0\n\n\nwhen Viewport shading color is set to vertex it is not displayed correctly for objects\n\n3.0:\n\n\n3.2 Alpha:\n\n\n\nset viewport shading mode to vertex\nvertex paint an object\nwhen you go back to object mode vertex paint is not displayed anymore\n\n\n" ]

[ "Dyntopo detail size drawing with Shift + D seems broken\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.79\n\nBroken: version: 2.92.0 Alpha\nWorked: At least a few alpha versions ago\n\nDyntopo size control with Shift + D seems broken.\nThe first time I tried, the gizmo simply froze and the Resolution value didn't change after mouse movement.\nAfter installing the latest GeForce Studio Driver — version 457.30 — I tried again. The gizmo casts infinite lines, but I noticed that the Resolution value does change after mouse movement.\n\n1. Enter Sculpt Mode.\n2. Activate Dyntopo.\n3. Switch to Constant Detail mode.\n4. Press Shift + D.", "Adjusting spotlight blend using gizmo does not register undo\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.0 Core Profile Context 22.11.2.221130\n\nBroken: version: 3.5.0 Alpha\nWorked: n/a\n\nThe gizmo control is introduced in 3c8c0f1094.\n\nAdjusting spotlight blend using gizmo does not register undo\n\n[SpotlightBlendGizmoBrokenUndo.mp4](SpotlightBlendGizmoBrokenUndo.mp4)\n", "Multiple calls of 'view3d.select' with LMB selection keymap\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.14760 Core Profile Context 20.45.37.01 27.20.14537.1001\n\nBroken: 2.83, 3.2.0\nWorked: Apparently never\n\nIn the left-click selection settings, considering only `Object Mode` in the keymap, there are currently 3 places setting the {key LMB} selection shortcut:\n1. {nav 3D View > 3D View (Global) > Selection}\n2. {nav 3D View > Object Mode > 3D View Tool: Tweak > Selection}\n3. {nav 3D View > Object Mode > 3D View Tool: Tweak (fallback) > Selection}\n\nRespectively each of them can be tested by pressing {key LMB} in the `3D View` under the following conditions:\n1. Always active\n2. Active when `Tweak Selection Tool` is enabled (between selection tools)\n3. Active when fallback selection is `Tweak` and the `Selection Tool` is not enabled\n\nThe problem is that, in the cases `2` and `3`, the key item of `1` is also mapped resulting in `view3d.select` being called 2 times (since this operator is pass though).\n\nThis can be a problem for users editing the settings as they need to make sure the properties of both calls are \"compatible\".\n\nNOTE: for undo, with the default settings, this is not a problem as an undo step is only added when something changes.\n\n - Make sure to use keymap with LMB selection\n\n **Testing the Selection Tool Shortcut:**\n - In the 3D view, activate the `Tweak` selection tool\n - In the keymap, go to {nav 3D View > Object Mode > 3D View Tool: Tweak > Selection} and disable the option `Deselect On Nothing`\n - In the 3d View, select the Cube and then click on an empty region - Note that the Cube has been deselected which is wrong\n - In the keymap, go to {nav 3D View > 3D View (Global) > Selection} and disable the option `Deselect On Nothing`\n - In the 3d View, select the Cube and then click on an empty region - Note that the Cube has **not** been deselected which is correct.\n\n **Testing the Selection Fallback Shortcut:**\n - Re-enable all `Deselect On Nothing` options and now disable the option located in {nav 3D View > Object Mode > 3D View Tool: Tweak (fallback) > Selection}\n - In the 3D View, activate a tool other than the selection tool (so the selection fallback will be used)\n - select the Cube and then click on an empty region - Note that the Cube has been deselected which is wrong\nRepeat the last two steps of the selection test with the tool to get the expected result.", "Bisect Tool Snap does not Work\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.67\n\nBroken: version: 3.6.1\n\nWhen using the bisect tool and dragging the yellow arrow, snapping does not work when holding the CTRL key despite it being an option in the context tool bar. With the rotate gizmo it works fine. \n\n1. Add for example Suzanne.\n2. Select the whole mesh.\n3. Use the bisect tool.\n4. Try first the rotate gizmo and hold Ctrl, snapping works fine.\n5. Now use the yellow arrow instead, no snapping at all regardless of the snapping settings.", "\"RENDER_OT_render\" called with \"EXEC_DEFAULT\" freezes Blender with Eevee motion blur enabled\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.90.0 Beta\n\nbpy.ops.render.render() called without 'INVOKE_DEFAULT' from the Python Console or script freezes Blender when Eevee motion blur is enabled.\nAnd the same issue is in the 2.91 Alpha version.\n\n----\n\n- Open Blender\n- Make sure context scene render engine is Eevee\n- Turn on motion blur\n- Open Python console\n- type in bpy.ops.render.render() and hit enter\n\n----\n\n*Edit* (added method that doesn't activate via Python).\n\n**Exact steps to reproduce the error without activating from Python.**\n- Open Blender\n- Make sure context scene render engine is Eevee\n- Turn on motion blur\n- Run this script in the text editor\n\n\n```\nimport bpy\ndef menu_func_import(self, context):\n layout = self.layout\n layout.operator_context = 'EXEC_DEFAULT'\n layout.operator(\"render.render\", text=\"Blocking Render\")\nbpy.types.TOPBAR_MT_render.append(menu_func_import)\n```\n\n- Run \"Render -> Blocking Render\" from the top-bar.", "Horizontal mouse wheel support for MS-Windows\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.81 (sub 16)\nWorked: (optional)\n\nLogitech MX Master 3 horizontal scrollwheel button works in mac os x, but not in windows 10 although both have the same settings in Preferences in Blender and Logitech Options App.\n\n\nSTEPS TO REPRODUCE ISSUE:\n\n1. Rez object, select it and orbit around it in Mac OS High Sierra and Blender 2.81 using MX Master 3.\n2. Do the same, but in Windows 10.\n\nBased on the default startup:\nA. Performance in Mac OS is the expected: when using the horizontal scrollwheel button (thumb button mouse) orbits. Works in a similar way than when Orbit with the Magic Mouse.\n\nB. In Windows 10 the thumb button doesn't work, even when the settings are the same in both computers for Blender and the mouse. \n\nWhy this happens and how can I get to make the thumb button to work in Blender for Windows?\n\n\n\n\n", "Switching windows does not register as click\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 465.89\n\nBroken: version: 2.93.0\n\nI use industry compatible keymaps.\nWhen switching between windows, no matter whether they are 2 Blender windows or Blender and any other, any camera adjustment is not registered and, depending on the keymap preset you use, you either rotate the camera or move the object instead(see attached capture).\n[switchingWindowsBug.mp4](switchingWindowsBug.mp4)\n\n1. Shift LMB Press on the corner where the viewport meets the outliner and drag to the left, this creates a new viewport window\n2. Create a cube and press on the move tool to show it's gizmo\n3. Pan the camera in one window\n4. Then without clicking into the second viewport try to pan the camera in the second viewport, depending on the keymap preset you use you either rotate the camera instead of panning or move the object\n\nBlender does not register when you have switched to the new window without first clicking with the LMB into the window.\nAs you can guess it makes it pretty nifty to switch between windows, either 2 Blender viewports, Blender and the tutorial you are watching or any window else.", "RigidBody explodes\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11\n\nBroken: version: 2.93.2 Release Candidate\n\nWhen adding RigidBody, the simulation \"explodes\".\nI noticed that it only happens when the objects are near the center of the scene.\nIf we move the objects in the +Y direction, the problem disappear\n\nIn the video example I show that the rigidbody is working correctly by adding an icosphere and coliding it to the wall\n\nAdd rigidbody near the center of the scene\n\n[ZzbBGXn6sP.mp4](ZzbBGXn6sP.mp4)\n[rigidbody_problem_example.blend](rigidbody_problem_example.blend)\n\n", "Extended Controls for Camera Gizmo\nCurrently, when switching to the camera view, it’s not immediately clear to users how to manipulate it. The camera must be selected, and it requires some difficult multi-step actions to be able to rotate or move the camera. \n\nAdditionally, from a camera operation perspective, users usually need a more optimized and precise way to only affect the pitch, yaw, panning, tracking and zooming, separately. \n\nWe can make this a lot easier by surfacing optimized camera manipulation gizmos right in the viewport when viewing the scene through the camera view:\n\n\n", "Outliner: Sync Selection off not working when clicking LMB on data object name\nBroken: version: 3.0.0\nBroken: version: 2.91.2 (works correctly on meshes for some reason till worked version)\nWorked: version: 2.82\n\nWhen `Sync Selection` is off it's still possible to sync selection by clicking on object data name (+ small margin). Additionaly clicking on object data name will sync it with relevant slot in Properties editor. This bugged area becomes slightly longer to the left when active, see video.\n\n1. Open default scene.\n2. Turn off `Sync Selection` on Outliner.\n3. Deselect onjects in your scene by clicking on empty space in 3D Viewport.\n4. Select objects or object data in outliner without clicking on their name - see that `Sync Selection` is off\n5. Select objects or object data in outliner by clicking in their name - see that now parent onject is selected in Outliner aswell as in 3D Viewport.\n[Outliner Sync Selection off bug 2021-12-15 20-45-35-213.mp4](Outliner_Sync_Selection_off_bug_2021-12-15_20-45-35-213.mp4)", "Gizmo and transform direction does not align with Normal Orientation and Individual Origins\nOperating system: Linux-5.15.0-41-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: Quadro GP100/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.73.05\n\nBroken: version: 3.4.0 Alpha\nWorked: 2.83\n\nThe gizmo axes orientation should give a preview of the transform direction of the selected elements.\nBut the behaviour is breaking when using **Normal** orientation and **Individual Origins** for the pivot point.\n\nIn Object Mode when selecting a single object it behaves like expected. But when selecting multiple objects, the gizmo and transforms will use the Median Point pivot point instead. I don't know if this is the intended behaviour.\n\nMore striking is the behaviour in Edit Mode. If a single vertex is selected, the behaviour is as expected. The gizmo and transforms will use the normal orientation of the individual single vert.\nBut once more verts are selected the gizmo no longer matches the transforms.\nThe gizmo seems to match the median point orientation. But once a gizmo handle is dragged, it will use something else.\nThis often leads to flipped directions on transforms.\n\n**Note on behaviour in Blender 2.83**\nWhile this behaviour worked as expected in that version and earlier, it still failed to show the correct orientation when more than 2 verts were selected.\nOf course not all individual orientations can be displayed with 1 axis. But when the selected individual elements are connected, they will move as one. This behaviour should be reflected in the gizmo orientation as well.\nIf the selected individual elements would not move in unison, the gizmo could show the combined average as a good enough estimation.\n\n- On the default cube, enter edit mode\n- Set orientation to Normal and pivot point to Individual\n- Make the Move Tool active in the Toolbar\n- Select 2 vertices that form an edge and drag along the X handle\n- If it moves as expected, select the 2 vertices again but in the reverse order\n- Then repeat", "Formalize Default Selection Behavior and Key-map\nSelection should behave as consistent as possible throughout Blender. There should be a formal description for that to help avoiding inconsistencies.\n\nNOTE: This task is WIP, all content is tentative.\n\nWe *may* have to draw distinctions between right-click vs. left-click and 2D vs. 3D editors. For the latter, there are rather different requirements.\n\n### 2D Editors\n\n| Key-map item | Action\n| ---------------- | --------\n| {key Click} | Select item (and activate) an item. Deselect others.\n| {key Shift Click} | Toggle selection (and activate):\n| | * Select (and activate) if item was *not* selected\n| | * (Activate it item was selected)\n| | * Deselect item (if already active, or) if selected.\n| ... | ...\n\n\n[TODO make table per-tool?]", "Addon: 3d Viewport Pie Menus. Press-move-release behaviour inconsistency\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\n\n\nBroken: version: 2.92.0 Alpha\n\nName: 3D Viewport Pie Menus (1, 2, 9)\nAuthor: meta-androcto\n\nIn pie-menus Option, who have additional menus do not react to fast action P-M-R or local hotkeys\n[2020-12-16_20-50-47.mp4](2020-12-16_20-50-47.mp4)\n\n- In sculp mode press W and then 2 or M\n", "Assigning pie menus to mouse clicks using the tweak function causes these menu shortcuts to stop working properly\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.39\n\n\nBroken: version: 2.80 (sub 71)\n\n\nBeen experimenting with creating keybinds and one of those experiments were to try and assign several menus to R click (using L click for select) by using the tweak function in the keymap settings. The idea was to have as many menus as possible on the same keybind to reduce the amount of keys needed, but using this method has proven difficult. \n\n\nWhile I can make standard menus work with tweaks, the pie menus don't work at all. If you manage to get the pie menu to show itself, it will eventually stop all the tweak commands from working unless you open up a menu on a keybind that is not using a tweak function. The second issue is that even when you do get the pie menus to show themselves, nothing can be selected on them, therefore making the keybind completely useless.\n\n\n1) Assign a standard menu and a pie menu to a mouse click (i.e. R click) using the tweak function, which can be set to north and south (just as an example).\n2) Use the new shortcuts to click and drag your mouse in the directions you assigned to to open up each menu (standard menu should be working correctly).\n3) Open the pie menu with this method and try to select any option on said menu, which should cause the pie menu to fail.\n4) Try opening the menus again, but this time none of the assigned menus will show up.\n5) To fix the previous error, open up a menu on a different shortcut (like something on your keyboard) to make the tweak shortcuts be able to appear again.\n6) Repeat the previous steps, which should cause the same error every time.", "Support Undo for tweaking Gizmos\nSupport Undo for tweak gizmos, such as the Spin gizmo. Currently, the user may have performed many adjustments using adjustment gizmos, but if the user performs one undo, everything is lost.\n" ]

[ "Gizmos blink when selecting with right mouse button\nWindows 10 64 Bits.\nGraphics card GeForce GTX 1080 Ti.\n\nBroken: version: 2.82.6\nWorked: 2.79b, 2.80 & 2.81.\n\nWhenever you select or extend select the gizmos do a quick blink.\n\n - I have a custom keymap that changed the type of event from Click to Press for the default.\n - This issue can easily be replicated if you use the Right Click Select.\n - Default Cube subdivided then Select few Vertices in a normal manner like shown in the Gif. \n\n\n\n{[F8246599](Select.gif) size = full}\n" ]

[ "Blender 4 Alpha Won't Open\nOperating system: Linux-5.15.0-79-generic-x86_64-with-glibc2.31 64 Bits, X11 UI\nGraphics card: AMD JUNIPER (DRM 2.50.0 / 5.15.0-79-generic, LLVM 12.0.0) X.Org 3.3 (Core Profile) Mesa 21.2.6\n\nBroken: I tried opening Blender 4.0.0 - Alpha\n August 30, 13:25:10\n e071288ab290\n tar.xz\n 250.91MB\n\nWorked: Blender 4 Alpha from July 15 opens, though.\nBlender 4.0.0 - Alpha\n July 15, 02:35:02\n d8e4fe320784\n tar.xz\n 249.01MB\n\n\n\n- I tried opening Blender 4 Alpha, latest version, and nothing happens.\n- I also tried an earlier version from August 15, and another from July 15.\n- August 15 wouldn't open. I made sure it was set to \"run as executable\".\n- July 15 opens, though.\n- I might check other versions. They take a while to download, though (I live in Canada).\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n - Not sure how others can reproduce this unless they have my pc, or are having the same problem.\n - If there's some kind of terminal window thingy that writes down info while trying to open Blender, let me know how to do that, and I'll copy and paste whatever it says.\n- Note: sometimes Blender 3.6 doesn't open at first. But if I try again, it opens just fine.\n\n", "closest_point_on_mesh() freezes Blender when called from a driver\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 580/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 2.82 (sub 6)\nWorked: 2.79\n\nCalling closest_point_on_mesh() from within a driver causes Blender to freeze.\n\n[test2.blend](test2.blend)\nOpen the file allowing scripts -> Freeze", "Eyedropper for props does not work in dialog boxes\nBlender 2.81, 71f2229b0d82\nfor example, in Texture Paint Context Menu\nI wanted to implement something using Operator and context.window_manager.invoke_props_dialog()\nThis is not a bug, but why not do it at the dialog level.\n", "Render problems\nOperating system: Windows 10 Pro Ameliorated\nGraphics card: RTX 3070 driver version 472.12\n\nBroken: v3.6 and v3.61\nWorked: v3.51 <- Using the same file created in v3.61\n\nWhile rendering a frames of an animated character, the current rendered animation frame usually shows in the 3D viewport in v3.51. However, in the broken version, it fails to show the current frame being rendered, nothing gets updated visually. Then eventually in the 3D viewport the character goes into its T pose for a few frames then the character starts losing visibility of its layers, at that point Blender crashes. \nI will create a .blend file if I am the only one getting this bug.\n\n", "Intel (R) HD graphics 4600 also has the problem of cubemaps shows black\nIntel (R) HD graphics 4600 also has the problem of cubemaps shows black\n\nSystem Information\nOperating system: windows7 64 Bits\nGraphics card: Intel(R) HD Graphics 4600 Intel 3.4.0 - Build 10.18.14.5117\n\nBlender Version\nBroken: since 2.83.3 branch: master\nCurrent broken: 2.93.7\nWorked: 2.82a\n\nShort description of error\nLike #75943 and #79158, HDRI cannot be displayed normally\n\nCan be resolved adding \"**HD Graphics 4600**\" to the blacklist\n\nComment from #79158: \"*I'm not sure whether Intel's graphics cards are like this. Can you put this repair method in preferences as an option?*\"\nOr perhaps adding command line flag like: **--disable_texture_cube_map_array**", "Cycles viewport render doesn't resize if an animation is playing.\nOperating system: macOS-13.0-arm64-arm-64bit 64 Bits\nGraphics card: Apple `M1` Pro Apple 4.1 Metal - 83\n\nBroken version: 3.5.0 Alpha\nBroken version: 3.3.1\n\nIf a .blend file contains a animation, and the animation is playing, and you're in the Cycles rendered viewport mode, then the viewport resolution/size will not update as you adjust it's size.\n\n1. Create a scene that has a colour that differs from the standard background colour of the 3D viewport.\n2. Animate something in the scene (E.G. Animate the movement of an object)\n3. Change the render engine to Cycles (This issue only appears in Cycles for me.)\n4. Enter the rendered viewport mode so you're now seeing a Cycles viewport.\n5. Start playing the animation.\n6. While the animation is playing, start adjusting the size of the 3D viewport editor and notice how the Cycles rendered view isn't updating with the change in size. This is the issue. Once you stop playing the animation, it will update.\n\nHere is a video demonstrating the issue:\n\n[Cycles preview doesn't resize.mp4](Cycles_preview_doesn_t_resize.mp4)\n", "Asset Browser: Material preview with black background when Cycles is disabled\nBroken: version: 3.0.0, 3.4.0\nWorked: Never\n\nWhen a scene was saved with Cycles but I disable the Cycles add-on the material previews get a black background.\n\n* Disable Cycles add-on.\n* Click on any (e.g., the Yellow) material asset and update its preview.\n[black_preview.blend](black_preview.blend)\n", "Modal keymap for snapping in the `Corner Pin` widget has no effect\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\nThe bottom bar states that pressing ^ will enable snapping, but it seems it does nothing.\n\n\n\n1. Create a new default file;\n2. Go to the compositor;\n3. Enable nodes;\n4. Add a viewer node (this allows you to resize and test different sizes for the `Corner Pin` Gizmo);\n5. Add a `Corner Pin` node;\n6. Edit the `Corner Pin` widget and try to snap.\n", "Viewport: anti-alising on normal lines\nNormals on faces/vertices are not antialiased:\n\n\n\nTo see them, enable them at the bottom of the 3D Viewport Overlay popover:\n\n\nThis issue is similar to what was covered in #107394", "--render-output command line argument and backslash\n## Commandline Rendering\n\nOperating system: Windows 11\nGraphics card: NVIDIA GTX 3060 Ti\n\n3.4.1\n\n\nIf the value for \"-o\" or \"--render-output\" ends with a backslash no rendering occur and Blender exits clean (0) with no error.\n\n### Example\n\nRender blender-file in current folder to subfolder with a space.\n\n>blender -b myfile.blend -o \"my folder\\\\\" -F PNG -f 1\n\n```\nBlender 3.4.1 (hash 55485cb379f7 built 2022-12-20 01:51:19)\nRead prefs: C:\\Users\\***\\AppData\\Roaming\\Blender Foundation\\Blender\\3.4\\config\\userpref.blend\nRead blend: c:\\blender\\myfile.blend\n\nBlender quit\n```\n\n**Observed behavior**\nNo render. No errors. No information of what went wrong.\n\n**Expected behavior**\nBlender should render to generic file \"0001.png\" in folder \"c:\\blender\\my folder\".\nBlender should provide some information why nothing got rendered.\n\n\n", "Resetting default value does not work in popovers or menus\nOperating system: Windows 10, 64 bits, \nGraphics card: Nvidia GTX 1080 8 gb, Nvidia GTX1070 8 gb\n\nBroken:\nblender-2.81.0-git.ce2368cf69f5-windows64\n\nWorked: (optional)\n\nResetting options in the context menu (right click) of things like brush stroke strength, Autosmooth, plane offset and other parameters does not work at all, but does work through active Tool side panes and top bar. Probably it should work everywhere? If I'm working full screen I would want to be able to reset it in the context menu with my stylus as I sculpt.\n\n\n\n1 Create any mesh\n2 Enable Sculpt mode\n3 Select any brush\n4 Right click to invoke context menu\n5 Try resetting any of the parameter to default values\n", "GPencil: Square type strokes do not react to Rotation option with Solid Shading\nOperating system: Linux-5.10.0-6-amd64-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.73.01\n\nBroken: version: 3.0.0 Alpha\nWorked: ?\n\nThe stroke type Squares draw little squares at the points of the stroke. The rotation is controlled by the alignment option and the rotation slider, which in this case does nothing if solid shading mode is selected. Note that textures are not displayed in solid shading mode and even if used the squares don't rotate as long as I am in solid shading mode.\n\n1. Add default Stroke Grease Pencil Object\n2. Navigate to the Material and change Line Type to Squares\n3. Make sure you are in Solid Shading Mode with \"Color\" not set to \"Texture\"\n4. Change the Rotation slider from the same panel, which says in the tooltip it should rotate the squares\n5. observe how the squares do nothing at all.\n\n", "Missing note in the documentation about contexts in application handlers\nOperating system: Linux-5.15.0-41-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: AMD Radeon RX 580 Series (polaris10, LLVM 14.0.6, DRM 3.42, 5.15.0-41-generic) AMD 4.6 (Core Profile) Mesa 22.1.4 - kisak-mesa PPA\n\nBroken: version: 3.2.0\nWorked: Could not find one\n\nWhen running `bpy.ops.render.render` with 'INVOKE_DEFAULT' as a result of a render_complete handler throws an exception, stating that `Missing 'window' in context`\n\n1. Create new file\n2. Add script to run render on render finish\n```\nimport bpy\n\ndef render_complete(a, b):\n bpy.ops.render.render('INVOKE_DEFAULT', write_still=True, use_viewport=False)\n\n\nbpy.app.handlers.render_complete.append(render_complete)\n\n```\n3. Run script and start render\n4. Exception gets thrown in Terminal:\n```\nTraceback (most recent call last):\n File \"/Text\", line 6, in render_complete\n File \"/home/richards/Documents/BlenderDev/blender-3.2.0-linux-x64/3.2/scripts/modules/bpy/ops.py\", line 113, in __call__\n ret = _op_call(self.idname_py(), C_dict, kw, C_exec, C_undo)\nRuntimeError: Operator bpy.ops.render.render.poll() Missing 'window' in context\n```\n\nI am unsure if this is intended behavior, as invoking a render in other circumstances reopens the render view. If it is intentional, it would be a useful thing to mention in the handler documentation, that inside the handler, the context does not contain 'window'", "Rendering animation freezes with particle system\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.91.2\nWorked: 2.83LTS (not sure, this is said in most of the comments)\n\nI am trying to render out an animation based on the CGBoost Blender Launchpad Tutorial. Every time I try to render, it will render around 50-400 frames and then it will just stop rendering new frames. My CPU usage goes straight up to 100% when that happens and my GPU usage drops down to 0%. The Blender UI is still responsive, it is just that the render seems to get stuck on a frame. If I cancel the render, then Blender becomes totally unresponsive and I have to force quit via the Task Manager.\n\nIssue happens in both EEVEE and Cycles\n\n**Workarounds (picked up from comments)**\n- Render animation from command line\n- Check lock interface: {nav Render > Lock interface}\n\n1. Open attached Blend file \n2. Press Ctrl-F12 to render out animation.\n3. Wait for render to freeze on a frame. This can happen within the first few frames or after 400-500 frames. It seems random.\n4. Examine CPU and GPU usage.\n\n[Blender Launchpad.blend](Blender_Launchpad.blend)\n[ING BG.blend](ING_BG.blend) (File provided by @Megaseppl: #85267#1122053)", "modifier display/render driver not responding.\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 527.56\n\nBroken: version: 3.4.0\nWorked: unknown\n\nIf I have my Mask display/render driver from another objects display/render setting, changing controlling obj disp/render doesn't take effect without coaxing.\n\n1. Open the attached file. \n2. In outliner, disable the viewport display of the Cube object (drives Mask visibility of part of Sphere object)\n3. For the Mask modifier, slide the Threshold a little. Display effect now takes place.\n\nThis all works without a problem if I animate the Cubes visibility. This appears to be a problem for any/all modifier display drivers (e.g., Subdiviison). Sometimes I just need to hover over a value like Threshold, or just over time it responds to the change.\n\n" ]

[ "Snap version of Blender v2.83.0 doesn't accept command line arguments\nOperating system: Ubuntu 20.04 LTS x86_64 \nGraphics card: VIDIA GeForce RTX 2080 Mobile, Intel UHD Graphics 630 \n\nBroken: v2.83.0 (from blender, e.g. the Steam version is not affected)\nWorked: v2.83.0 from Steam; I'm also sure there was a v2.82.x from the snapstore but I think it was replaced\n\nOn Ubuntu/Linux when installing/updating blender through the built-in snap store, the latest version doesn't accept any [command line arguments ](arguments.html) anymore (e.g. `--background`). It used to work before (2.82) and it still does with the Steam install of v.2.83.0. It seems something in the \"snap\" package broke.\n\n\n\n1. Install blender via the snapstore. E.g. through the command-line `snap install blender --classic`. (If already installed you can verify with `snap info blender` if was automatically upgraded to the latest. The last line printed should be `installed: 2.83.0`)\n\n2. start blender in background mode via the command line `blender --background`. Notice the it opens regularly since the argument is ignored. Any other arguments like python related are as well.\n\nOptionally:\n\n3. Install blender via Steam for Linux\n\n4. Run the that install and notice the arguments work. Default path should be: `$HOME/.local/share/Steam/steamapps/common/Blender/blender --background`\n\n\n----\n\n```\n$ snap info blender\nname: blender\nsummary: Blender is the free and open source 3D creation suite.\npublisher: Blender Foundation (blenderfoundation✓)\nstore-url: blender\nlicense: GPL-3.0\ndescription: |\n Blender is the free and open source 3D creation suite. It supports the\n entirety of the 3D pipeline—modeling, rigging, animation, simulation,\n rendering, compositing and motion tracking, even video editing and game\n creation.\n \n Blender is a public project, made by hundreds of people from around the\n world; by studios and individual artists, professionals and hobbyists,\n scientists, students, VFX experts, animators, game artists, modders, and\n the list goes on.\ncommands:\n - blender\nsnap-id: ZWyHJJiPBBfMKpaaKKrpSgAcupWkHWPz\ntracking: latest/stable\nrefresh-date: yesterday at 08:58 CEST\nchannels:\n latest/stable: 2.83.0 2020-06-03 (38) 170MB classic\n latest/candidate: ↑ \n latest/beta: ↑ \n latest/edge: ↑ \n 2.83lts/stable: 2.83.0 2020-06-04 (38) 170MB classic\n 2.83lts/candidate: ↑ \n 2.83lts/beta: ↑ \n 2.83lts/edge: ↑ \n 2.79/stable: 2.79b 2019-07-30 (20) 129MB classic\n 2.79/candidate: ↑ \n 2.79/beta: ↑ \n 2.79/edge: ↑ \ninstalled: 2.83.0 (38) 170MB classic\n```", "blender 2.83 not reading command line\nOperating system: Ubuntu 18.04.4 LTS (Ubuntu Studio, Linux 5.3.0-53-lowlatency #47~18.04.1-Ubuntu SMP PREEMPT)\nGraphics card: NVidia GeForce GTX 1050 Ti\n\nBroken: official snap install of Blender 2.83.0 (snap revision 38, tracking latest/stable, publisher blenderfoundation, notes classic)\nWorked: official snap install of Blender 2.82a (snap revision 37, tracking latest/stable, publisher blenderfoundation, notes classic)\n\nBlender doesn't seem to be listening to or parsing command line options in this installation, including the loading of .blend files.\n\nexecute `blender <blendfile>` from the command line, where `<blendfile>` is an edited `.blend` file. The error will manifest as the loading of the default, empty `.blend` file.\n\nexecute `blender -b` from the command line. The error will manifest as blender opening with the default `.blend` file rather than exiting with `Blender quit`.\n", "Blender ignores commandline options\nOperating system: Linux-5.6.11-050611-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX TITAN X/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.64\n\nBroken: version: 2.83.0\nWorked: 2.82a\n\nIt seems, that 2.83 ignores commandline options.\n\nI have tried to run Blender as usually with blender - b [file] -S [something] -f 1 but Blender started full GUI with default scene. The same result I get if I try to run blender --help.\n\nI have Blender installed from snap and it upgraded automatically. With 2.82 it worked fine.\n" ]

[ "UV Editor Snap Active to Vertex is Not Working\nOperating system: Linux-5.19.0-38-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.105.01\n\nBroken: version: 3.5.0\n\nI can not snap active to vertex, no matter what I try it acts like it is in median mode.\n\n\n\n\nIn the startup file snapping is enabled. In the uv editor try to move the selected vertex to snap to another vertex. Or try to snap a face instead, same issue. Sync mode on/off doesn't change the behavior.\n\n", "Sculpt Mode: Expand Mask not properly snapping to face sets\nOperating system: Linux-5.15.0-56-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: Quadro GP100/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.108.03\n\nBroken: version: 3.5.0 Alpha\nWorked: never\n\nWhen using mask expand via `Shift A` it is common to hold `Ctrl` to snap the expansion to face sets.\nThis is the most direct way of filling face sets with a mask.\n\nThe issue is that expand seems to pre-determine which vertex is assigned to which face set for snapping.\nThis is not ideal, since it can lead to inconsistent results along the face set boundaries.\nIf the face set is small enough it can also lead to the expand operation having no effect.\n\n\nNotice in the image how the pink face set is unmasked on the left boundary but masked on the right boundary.\n\nThe ideal result would be that the expansion is always fully snapping to any vertex that is part of the face set.\n\n- Add a monkey primitive\n- Create some face sets with the Draw Face Set brush \n- Use `Shift A` to expand a mask\n- Hold Ctrl to snap to any of the face sets\n", "Knife not snapping on backfacing elements (wireframe, xray mode)\nSteps to recreate:\n1. Select object and goto edit mode\n2. Enable X-ray Mode.\n3. Select \"Knife\" Tool.\n4. Now, when i want to snap my knife to the vertices(\"Back Side\") for cut it didnt snap. Although It snaps to vertices on front part", "snap to face broken with mirror modifier\nOperating system: Linux-4.18.0-25-generic-x86_64-with-debian-buster-sid 64 Bits\n(same behavior on Windows 8.1)\n\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\nBroken: version: 2.81 (sub 11)\nAlso broken in 2.80 stable\n\n* open demonstration blend file with no config dir\n* note that snap-to-face (target:closest, affect:move) is already enabled\n* select small cuboidal in lower left\n* G, then X, to move along X, and attempt to snap to broad side of nearby horizontal cuboidal\n* see that the object flashes erratically, alternately assuming its original position, and alternately assuming an incorrectly snapped position\n* if instead the cuboidal we are trying to move is moved a bit first, without snapping, and then we again try to snap to the face, it doesn't flicker any more, but it still snaps incorrectly (and actually still does flicker on the overlapping face but doesn't jump around in position)\n* if the mirror modifier on the cuboidal we are trying to move is disabled, it snaps correctly\n\nStarting position: \n\nIncorrect snap result: \n\nCorrectly snapped result: \n\nDemo blend file: [snapbug.blend](snapbug.blend)", "Some transform modes don't apply absolute grid snaping \nOperating system: Windows 10 \nGraphics card: GTX 1080 ti \n\nBroken:\nBlender 2.8 January 31, 01:48:40 - 8c8979067490 Beta\nBlender 2.79 (It worked)\nBlender 2.77 (It worked)\n\nAbsolut grid snaping dont work with \"Face Inset\" and neither with loop cuts. \n\nTurn on the snaping (increment) -> turn on absolut grid snaping -> spawn a plane -> try to snap while loop cutting or inseting faces.", "Regression: Texture slots thumbnails in Texture painting mode are missing\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.3.0 Alpha\nCaused by 8b7cd1ed2a\n\nTexture slots thumbnails in Texture painting mode are missing\n\nIn Texture Paint mode, create texture, begin to paint, save texture and blender file. In the active tools panel>Texture Slots, the texture appears as a generic icon and its name, in 2.90 there was a thumbnail of the created texture instead of a generic icon.\n\n\nimages for reference:\n\n\n", "Weird behavior with the set spline type node selection\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.6.1\n\nThe selection on the *set spline type* isn't working right for bezier and NURBS modes.\nAlso adding another node before the *set spline type* node will manipulate the weird behavior, even though the node doesn't do anything or is muted etc.\n\n- add a GN setup\n- add more that one curve\n- add a *set spline type* \n- try adding a selection that doesn't select all splines\n- see the weirdness.\n(it is more perceivable if the splines have more than two points, though the bug is still there)\n\nFile:\n\n[GN_set_spline_type_test.blend](attachment)\n\nVideo:\n\n<video src=\"attachment\" title=\"2023-07-31 21-15-45.mp4\" controls></video>\n\n", "Geometry Nodes Volume Features Implementation\n**Status:** *In planning phase*\n\n---\n\n## Team\n**Commissioner:** @HooglyBoogly\n**Project leader:** `?`\n**Project members:** @erik85 \n\n## Description\n**Big picture:** *Address the large opportunities for procedural volumes while fitting into geometry nodes design, especially using existing openVDB tools*\n\n**Use cases**:\n* Procedural modeling\n* Fluid simulation\n* Motion graphics\n* ...\n\n**Design:**\n* Volume grids are like attributes conceptually\n* SDF Volumes are openVDB level sets and can be accessed directly with new nodes\n* Volumes have one built-in grid, \"density\", which is an SDF grid\n* Volumes can be adjusted and created with fields\n\n**Engineer plan:**\n* Besides field evaluation, openVDB tools will be used for most functionality\n* Initially fields will just be used to reference grids directly\n* Field evaluation will be very basic, just flattening out all data in the context's voxel topology\n\n\n## Work plan\n### Milestone 1 - Very basic SDF volume nodes\n*Expose SDF volumes as a basic data type, provide a few fundamental nodes for future development*\n\n- [x] SDF Sphere input node (creates a sphere with openVDB function, puts it in \"distance\" grid\n- [x] Add \"Mesh to SDF Volume\" and \"Points to SDF Volume\" nodes\n- [x] Add a SDF volume \"filter\" node using the openVDB level set tools functions (no mask input yet). Just acts on \"distance\" grid\n\n### Milestone 2 - Basic support for fields as references to grids\n*Allow referring to grids with the named attribute node, but not any modification*\n- [x] The named attribute node (`AttributeFieldInput`) should work for volumes too, referencing grids by name\n- [x] Add a \"Sample Volume\" node that gives the value of a grid at a location, with an `AttributeFieldInput` input. Any other field can be rejected for now. It just looks up the grid by name\n- [x] Add the \"Signed Distance\" node which is a field input node for the \"distance\" attribute\n\n### Milestone 3 - Grids as anonymous attributes\n*Creation of and references to grids without a name, to allow referencing them with node links*\n- [ ] Add the ability to store grids with an anonymous attribute ID rather than a name\n- [ ] Add \"SDF to Fog Volume\" and \"SDF to Mask Volume\" nodes that create anonymous grids\n- [ ] Support anonymous grids in \"Sample Volume\" node\n- [ ] Add basic support to \"Store Named Attribute\" node. For now it can only give a name to an anonymous grid\n- [ ] Change Volume Cube node to output an anonymous grid, use \"Store Named Attribute\" for versioning\n- [ ] Add field inputs to \"Volume to Mesh\" and \"Distribute Points in Volume\" nodes. Like above, only supports `AttributeFieldInput` for now.\n\n### Milestone 4 - Evaluating fields on grids\n*Support basic volume advection*\n- [ ] Add a `VolumeGridFieldContext` that can provide the position field at each voxel (by index for now, so voxels probably have to be flattened)\n- [ ] Add an \"Advect Volume\" node with an \"Offset\" field input.\n- [ ] Give the grid field context the ability to sample other volumes (special VArray output for attribute field input)\n- [ ] Add the ability to create grids matching the exact topology of existing grids with capture attribute node and store named attribute node\n\n### Milestone 5 - Changing grid values with fields\n- [ ] Allow function nodes on arrays of flattened voxel values provided by the field inputs from milestone 4\n\n### Milestone 6 - More advanced volume nodes\n*More advanced use cases building on the fundamentals above*\n- [ ] \"Fracture SDF Volume Node\" corresponding to fracture level set openVDB tool\n- [ ] Volume Boolean node including CSG operations (this could come earlier, but the design isn't totally clear yet)\n- [ ] More...?\n\n\n---\n\n### Relevant links\n* #91668 (How to deal with multiple volume grids in a geometry)\n* [D12100: Geometry Nodes: Level Set Nodes (WIP)](D12100)", "Unpredictable 3D cursor positioning with vertex snapping and active target\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.63\n\nBroken: version: 3.2.0\n\nUnpredictable 3D cursor positioning when using active + vertex as snapping option\n\nUse snap with vertex and snap set to active.\nPosition 3D cursor by either having it as the active tool or with right + shift (and ctrl if snapping is not currently activated)\nSometimes it goes under the position under the mouse, sometimes it goes somewhere apart and depending on the direction you hover over it gets a different direction offset.\nThis is very annoying as I have to change snapping to closest everytime I want to accurately position the cursor, I use active most of the time and the cursor should allways go there, at least when only vertex is the snapping option that is used.\nAlso when I don't have any object selected it works as intended.\n\n[20220620-160415.mp4](20220620-160415.mp4)", "GPencil: Fill tool Ignore Transparent strokes setting broken\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.5.0 Beta\n\nFill tool not ignoring transparent strokes when the settings is enabled. When you enable Show boundary lines, it shows that transparent strokes are not a boundary but the fill still uses them as boundaries.\n\n- Open attached file and try the fill tool. Under the Advanced menu 'Ignore Transparent' and 'Show Lines' are enabled.\n\n", "Workspace tabs click-dragging only works when starting from current tab.\nSystem Information\nOperating system: Windows 10\nGraphics card: Geforce GTX 1070\n\nBlender Version\nBroken: 2.80 2019-02-20 20:10\n\nShort description of error\nClick-dragging on tabs works fine in the Preference's tabs now, but fails to work on the Workspace tabs *unless*you start from the current one (so to rephrase, for clarity: if you start on one that's not the currently selected one, it will simply not go through the tabs)\n\n- Create workspace (either duplicate or append) so there's more than one\n- Click on one of the workspaces that isn't currently displayed\n- Drag left and right, you'll notice nothing happens.\n- Click on the currently selected workspace's tab\n- drag left and right: this should switch between workspaces as you pass over their tabs", "Transform cursor (patch included)\nRecently LMB-Drag was made to move the cursor *(edit, this has since been reverted)*.\n\nHowever this just moves the cursor under the mouse,\n\nit _could_ be useful to use transform here, it allows.\n\n- Drag, XKey to lock to an axis\n- ... Shift+Y to constrain to a plane.\n- ... Hold Ctrl to snap (snap to verts for example is handy)\n- ... Shift for more precise placement...\n... most other transform keys.\n\n\nOne reservation I have about this, is that you probably would often want to be able to snap to vertices and edges, and not have to depend on the current snap options in the 3d view header.\n\nSo we could override the snap options that transform uses.... but this seems a bit crappy too, since you may want to snap to faces in some cases, or even snap to grid.\n\nAs an alternative, we could just not bother attempting to integrate this with transform, and add snapping to the current modal operator (similar to how we have with the Ruler operator, which shares low level code with transform snapping)\nBut in that case we miss the axis constraints which seem like they would be nice to support.\n\n\nHere is a test patch to show how it works: [P87](P87.txt)\nNote, it crashes if you try to scale and its hard coded to LMB for cursor place, so its mostly a proof of concept.", "New geometry nodes viewer overlay makes mesh edit mode overlays invisible\nOperating system: Linux-5.19.10-200.fc36.x86_64-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.65.01\n\nBroken: version: 3.4.0 Alpha, current master\n\n\n| {[F13588717](image.png) size=full} | {[F13588720](image.png) size=full} |\n| -- | -- |\n", "Snap vertices with \"Closest\" is not acting as expected\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.83.3\nWorked: 2.82a at least. Didn't check any 2.83 releases. \n\nSnapping to vertex when moving with \"Closest\" works as if \"Center\" is chosen when moving several vertices at once.[cubicle.blend](cubicle.blend)\n\n- In the attached file, go to edit mode with the selected cubic shape.\n- 3 vertices already selected, try to move them down on Z axis with snap to vertices, \"Closest\".\n- The expected behavior is that the vertex at the bottom among the 3 selected vertices, should be at the same position of the vertex I'm snapping to.\n\n#46892 (Snapping closest vertex with Pivot point set to Individual Origins)\n\n", "Wrong interaction by extrude region with 2 cube's faces\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce 940MX/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.22\n\nBroken: version: 3.1.0\n\nWhen extrude the \"z\" and \"-x\" faces of cube they extrude up \n\n* Create a cube\n* Choose edit mode\n* Select \"z\" face(top of cube) and \"-x\" face\n* Use Extrude Region or press E\n\n{[F12930516](image.png), size=full}\n\nThanks!" ]

[ "Snap cursor gets too small sometimes\nBroken: version: 2.93.6 Release Candidate\nWorked: Never\n\nWhen the snap cursor doesn't snap anything (or is snapped to grid) and you zoom in on the scene, it gets too small.\n|||\n| -- | -- | -- |\n\nNOTE: the cursor displays increments that are not actually used for snapping. This can be confusing when trying to increment over small distances.\n\n- Enable the `Add Cube` tool (to display the cursor)\n- Point to a region of the Grid further in the background.\n- Zoom In\n\n" ]

[ "Hair particles crashing when baking or animating\nOperating system: windows 10 64 , windows 8 64\nGraphics card: intel, nvidia\n\nBroken: blender-2.81-9bd0d8b5505a-windows64\nWorked: (optional)\n\nwhen trying to bake a hair particles the system hangs giving this message\n\nError : EXCEPTION_ACCESS_VIOLATION\nAddress : 0x00007FF7CEBD24A4\nModule : C:\\Users\\dribeiro\\Desktop\\blender-2.81-9bd0d8b5505a-windows64\\blender.exe\n\n\nopen the file\nclick 'play' on timeline or bake on hair particles\n\nNEW INFO\nremoving the hair particles the animation runs normally\n[giselle.blend](giselle.blend)", "Blender stops updating display after GPU spike\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Plus Graphics Intel 4.5.0 - Build 27.20.100.8681\n\nBroken: version: 2.92.0 Beta\nWorked: I have not found a version that works as expected on this computer. I have tried 2.79, 2.8, 2.91\n\nI was premature in saying the problem is gone for [#85061 ](T85061).\n>After a few minutes of using Blender, the window freezes. Windows does not detect that it is not responding, but Blender does not respond to any input, including clicking on the close button. All elements that are drawn by Blender remain as they were at the moment of the crash except if I resize the window smaller.\n\nIf I watch the performance tab on Windows Task Manager, the freeze coincides with a 100% spike on the GPU.\nAs long as that spike doesn't happen, there is no problem, but once it does, Blender stops responding to any input, but it does not crash.\nI have updated my graphics drivers.\n[blender_debug_output.txt](blender_debug_output.txt)\n\nI have not determined exact steps. It has occurred on multiple files doing multiple different tasks. As far as I can tell, there is no correlation between what I am working on or what task I am performing, and the error occurring.\n", "Render Error using the HIP backend\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: AMD Radeon RX 7900 XTX ATI Technologies Inc. 4.5.0 Core Profile Context 22.40.51.06.230425\n\nBroken: version: 3.6.0 Beta\nWorked: Unknown\n\nMessed up renders in multiple of the common Blender Demo files. Classroom being the most obvious.\n\nStarting blender and running the render with the HIP or HIPRT backend using the 7900 XTX GPU causes the render errors in the attached \"broken\" images. Running and finishing a HIPRT render targeting the CPU results in a proper render, and fixes this for subsequent GPU renders with both HIP and HIPRT, see \"non-broken\" images.\n\n", "Blender 2.91 crashes when enabling Archipack addon\nOperating system: Linux-5.8.0-23-generic-x86_64-with-debian-bullseye-sid 64 Bits\nUbuntu 20.10 Groovy Gorilla (development branch)\nGraphics card: Mesa Intel(R) UHD Graphics 620 (KBL GT2) Intel 4.6 (Core Profile) Mesa 20.2.0\n\nBroken: version: 2.91.0 Alpha\n\nBlender crashes after enabling Archipack addon. Terminal says 'Aborted (core dumped).' Also, Blender doesn't record a crash.txt (that I can find). \n\nStart Blender>Edit>Preferences>Addons>Enable Archipack addon.\n\n[CoreDump.mp4](CoreDump.mp4)\n", "Mantaflow: Crash when setting a Guide Fluid Effector\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.96\n\nBroken: version: 3.3.0\n\nAfter setting a mesh as a \"Guide\" fluid effector, checking the \"Guides\" box in the fluid domain settings crashes Blender. \n\n- Open attached file,\n- go to the fluid domain settings\n- check the \"Guides\" box. \n[Fluid_Guide_Crash_01.blend](Fluid_Guide_Crash_01.blend)\n\n", "Vulkan: Better handling of GPU debug when not having vulkan SDK installed.\nBlender libraries only contain a subset of the vulkan SDK. Most important is\nthat validation layers aren't added. These validation layers will be enabled\nwhen starting blender with `--debug-gpu` (or test cases).\n\nIn this case there is an error printed to the console. This should not be an\nerror, but we could add a log line when not found.\n\n", "Crash when updating image pixel data through python API\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.83.0, 2.90\nWorked: N/A\n\nBlender crashes with apparent memory corruption, when image pixel data and previous are updated frequently with python script.\nStack traces give no obvious answer, i have seen it happen in depsgraph eval, node graph code, imbuf, etc.\n\n*Note: i could not reproduce this issue in debug mode so far, only release or RelWithDebInfo seems to work.*\n\n 1. Open attacehd file\n - Run the script\n - Make sure the scene properties are visible (it seems to only crash when displaying image previews).\n - Grab the default cube and wiggle it around to force depsgraph updates.\n - If it doesn't want to crash, re-open Blender, it seems to be quite random.\n\nI'm using a node group to access image data blocks, so they can be displayed in a panel as pointers. The code generates data to replace the image pixels as well as the preview.\nI have also tried running a timer and deferring pixel updates using a queue as suggested [here ](bpy.app.timers.html#use-a-timer-to-react-to-events-in-another-thread), just in case the depsgraph post update is somehow not safe, but this doesn't change anything substantially. The bug happens in both cases.\n\n[crashtest.blend](crashtest.blend)", "Blender 3.6.3 SegFault Crash at startup on MacOS Ventura 13.5.2\nOperating system: MacOS Ventura 13.5.2 (22G91)\nGraphics card: Intel Iris Plus Graphics 650 1536 Mo\n\nBroken: 3.6.3 (3.6.3 2023-09-21)\nWorked: None tried on this machine\n\nCrash on startup with `Segmentation fault: 11`\n\nEither:\n- Double-click on Blender application icon\nOr:\n- command line : `/Applications/Blender.app/Contents/MacOS/Blender` with or without option `--gpu-backend`\n\n", "GPU Compute is not available\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: GeForce GTX 660/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 3.6.2\n\nGPU Compute is not available in render device untill recheck\n\n\nAn interesting circumstance was discovered:\nIf open `preferences/System/CUDA` and just uncheck and check the checkbox of GPU Device, the problem is solved, then GPU Compute is accessible and fully functional.\n\nhad a similar problem: 94699\nThe solution is supposedly installing a newer version of nvidia driver. However, in fact, Blender works fine with the existing version.\n\n", "Extremely frequent crashing when using the compositor. Especially when using the glare node.\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\nWorked: I believe it was far more stable before the realtime compositor implementation. (Blender 3.3-ish)\n\nBlender crashes extremely frequently when using the compositor on my system, especially when the glare node is being used at all. Does not seem related to vram or system specs, and looking through logs has not come up with any errors - it just closes.\n\nAdjust compositor glare node values with the viewer enabled. It seems to crash within a minute of doing so, usually sooner.\n\n", "Problems due to leftover Nvidia drivers\nOperating system: Windows 11 Home\nGraphics card: Intel Arc a770 16gb Limited Edition (latest drivers)\n\nBroken: (Blender 3.3, 3.5 and 4.0) Trying to work out if it is just my GPU that is the issues\n\n\nBlender 3.5.1 will not allow me to access editor preferences, and won't allow me to render in Cycles. CPU or GPU. Eevee works.\n\nBlender 3.3 allows me to access editor preferences, but still won't allow me to render in Cycles. CPU or GPU. Eevee works.\n\nBlender 4.0 allows for editor preferences, same issue with Cycles. Eevee works fine.\n\nBy not allowing me to access editor preferences or Cycles I mean it crashes.\n\nI have tried to converse with Intel engineers with a video of the crashes but they seem incompetent with knowing how to enable GPU rendering in their own tests.\n\n", "F12 render all 600 frames complete cause crash.\nOperating system: macOS-13.5-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.14.1\n\nBroken: version: 3.6.3 Release Candidate\n\nF12 render all 600 frames complete cause crash.\n\n\n\n\nBlender 3.6.0 crash\nBlender 4.0.0 is fine\n\n", "[winerror 233] .glb, .gltf\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce MX130/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.42\n\nBroken: version: 3.6.2\n\n\n\ncant import .glb or .gltf files\n\n", "Blender Instantly Crashes When Rendering Cycles Or Clicking Wrong Setting\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nCPU: 12th Gen Intel(R) Core(TM) i7-12700KF, 3600 Mhz, 12 Core(s), 20 Logical Processor(s)\nGraphics card: AMD Radeon RX 6800 XT ATI Technologies Inc. 4.5.0 Core Profile Context 23.9.1.230821\n\nBroken: version: 3.6.2\n\nWhenever Blender attempts to render in cycles mode it exits immediately. It doesn't try for a few seconds, there's no loading, just click and instantly the process dies. This happens whether Preferences>System>Cycles Render Devices is set to \"None\", \"HIP\" with my GPU selected, or \"HIP\" with my CPU selected. Regardless of the cycles render device setting when I turn on viewport shading it works as it defaults to Eevee. However, if I try to change it to cycles, the moment I click cycles Blender crashes.\n\nThis same instant crash happens when I try to click OptiX or CUDA in Preferences>System>Cycles Render Devices instead on None or HIP.\n\n\n- Start Blender with a default project\n- Press Z then 8 to go into rendering mode\n- On the sidebar click render properties\n- Click the dropdown to change render engine\n- Click Cycles\n- Should see it crash instantly\n\nTo see the settings crash\n- Start Blender with a default project\n- Navigate to Edit>Preferences>System>Cycles Render Devices\n- Click between None, HIP, and OneAPI without issue\n- Click CUDA or OptiX and see the instant crash\n\n", "Subdivision Surface: Out of memory exception in in post 2.79 versions\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.90.0\nWorked: (2.79)\n\nWhen rendering a displacement heavy blend file, We consistently get out of memory exception in blender 2.8x and 2.9.\n\n - Open the attached blend file.\n - Click F12 (Render Image).\n # Wait till it crashes (approximately 6min).\n\nMore info about this blend file:\n```\nResolution: 650px by 650px\nCycles: 10\nAll polygons are triangles.\nOne material with a displacement map on all objects.\nSubdivision modifier with Catmull Clark with 1.0 adaptive value, on all 3 objects.\nSimplify->Render->Maximum Subdivion is equal to 2.\nSimplify->Render->Texture Limit is equal to 8192.\nSimplify->Viewport->Maximum Subdivion is equal to 0.\n```\n\n[Bug-Upload(2.79).blend](Bug-Upload_2.79_.blend)" ]

[ "GPU subdivision: Crash on adding subdivision surface modifier\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 30.0.101.1660\n\nBroken: version: 976c91cd770f\nWorked: 5035fbdd23ba\n\nCaused by 374cb32d4a\n\nCrash occurs at least on my intel integrated GPU after adding subdivision modifier\n\n- Open default blender scene\n- Enable GPU subdivision\n- Add subdivision surface modifier on default cube\n\n```ERROR (gpu.shader): subdiv patch evaluation ComputeShader:\n```\n |\n```\n 7 | #define GPU_OPENGL\n```\n | ^\n | Error: 'array' : failed to expand macro```\n```\n\n```lines=10\nStack trace:\nblender.exe :0x00007FF6C4B95BF0 GPU_shader_bind C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\gpu\\intern\\gpu_shader.cc:499\nblender.exe :0x00007FF6C41FF270 draw_subdiv_extract_pos_nor C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache_impl_subdivision.cc:1256\nblender.exe :0x00007FF6C427EEA0 blender::draw::extract_pos_nor_init_subdiv C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\mesh_extractors\\extract_mesh_vbo_pos_nor.cc:232\nblender.exe :0x00007FF6C42377D0 blender::draw::mesh_buffer_cache_create_requested_subdiv C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache_extract_mesh.cc:872\nblender.exe :0x00007FF6C41FCFA0 draw_subdiv_create_requested_buffers C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache_impl_subdivision.cc:1971\nblender.exe :0x00007FF6C41FDE50 DRW_create_subdivision C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache_impl_subdivision.cc:2124\nblender.exe :0x00007FF6C421EEA0 DRW_mesh_batch_cache_create_requested C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache_impl_mesh.c:2178\nblender.exe :0x00007FF6C420E340 drw_batch_cache_generate_requested C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_cache.c:3358\nblender.exe :0x00007FF6C41FB160 drw_engines_cache_populate C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_manager.c:1073\nblender.exe :0x00007FF6C41F79B0 DRW_draw_render_loop_ex C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_manager.c:1670\nblender.exe :0x00007FF6C41F8AA0 DRW_draw_view C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\draw\\intern\\draw_manager.c:1592\nblender.exe :0x00007FF6C48A7B40 view3d_main_region_draw C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\editors\\space_view3d\\view3d_draw.c:1564\nblender.exe :0x00007FF6C3DA8480 ED_region_do_draw C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\editors\\screen\\area.c:547\nblender.exe :0x00007FF6C3D8FD60 wm_draw_window_offscreen C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\windowmanager\\intern\\wm_draw.c:730\nblender.exe :0x00007FF6C3D8FBB0 wm_draw_window C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\windowmanager\\intern\\wm_draw.c:884\nblender.exe :0x00007FF6C3D8F6E0 wm_draw_update C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\windowmanager\\intern\\wm_draw.c:1087\nblender.exe :0x00007FF6C3D69C20 WM_main C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\blender\\windowmanager\\intern\\wm.c:630\nblender.exe :0x00007FF6C3D653A0 main C:\\Users\\Pratik\\Desktop\\BlenderOSP\\blender\\source\\creator\\creator.c:552\nblender.exe :0x00007FF6C4DA8784 __scrt_common_main_seh d:\\a01\\_work\\12\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl:288\nKERNEL32.DLL :0x00007FF93FD27020 BaseThreadInitThunk\nntdll.dll :0x00007FF941D02630 RtlUserThreadStart```" ]

[ "Gradient weight tool confusing (using previous tool brush)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.82 (sub 7)\nWorked: (optional)\n\nGradient Tool in its current state is not very user friendly:\n\n- generally this accesses previously selected tools brush as it doesnt have its own brush (so it looks like it has its own settings for strength/weight, but those are from an 'alien' brush) when `Unified Weight` is turned ON (default) the tool actually works with the `Unified Weight` (set on some other brush, e.g. the Draw Brush), but doesnt present you with that value in the UI (bug!)\n- when `Unified Weight` is turned OFF you are actually tweaking the weight setting of the previously selected tools brush (the 'alien' one) -- this causes confusion (it doesnt affect e.g. `Blur`, `Average` or `Smear` since these dont have weight capabilities)\n- this is even worse when tweaking the Strength, you are actually tweaking the weight setting of the previously selected tools brush (the 'alien' one), this affects `Blur`, `Average` or `Smear` (and Unified Strength is OFF by default!)\n\nSo not sure how to solve the issue with the \"tweaking setting of a seemingly unrelated brush\" issue, what I think we should do though is at least have the values in the UI respect Unified Weight / Unified Strength since this is what the tool will use internally...\n\nSo the main issue is fixed in f38c54d56e: Fix Vertex weight gradient tool show wrong weight/strength values in the UI.\n\n", "GPencil: Randomize brush's 'Position' attribute doesn't work when camera is too close to Grease Pencil object\nOperating system: Win10 Home 64\nGraphics card: rtx3070\n\nBlender Version\nBroken: 2.93, 3.1.0\nWorked: N/A\n\nExact steps to reproduce the error:\n\n- Open file\n- try to use Randomize brush\n- Strength and Thickness work, but not Position \n- Switch to the other camera that's farther away from the GP object, and randomizing the position works fine.\n\n[randomize_not_working_close_NEW.blend](randomize_not_working_close_NEW.blend)\n\n\nwatch?v=4pTrEm5HqVU", "Weight painting with Zero Weights overlay method Active, once painted vertices remains blue\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.33\n\nBroken: version: 3.5.0 Alpha\n\nI don't think it has ever worked.\n\n\nWhen you choose the Zero Weights overlay method Active, then vertices with a weight of 0 are displayed as black. But once painted vertices remains blue when you paint with a value of 0 across it. You cannot paint them to zero again. Not without lots of tries. I managed to get it back to black once after five minutes permanently painting across it.\n\n\n- Switch to weight paint\n- In the Viewport Overlay settings change the Zero Weights behaviour to Active.\n- Paint a vertice \n- Now change the value back to zero and try to paint it back to zero again. So that it displays black like the other vertices with a weight of zero.\n\nSometimes, in very rare cases, i really managed to paint it back to black again. And i can misuse the Gradient tool here, which is a bit more reliable. But also this method needs lots of tries. Which made me report this as a bug then. I would expect that i can paint it to zero with a few strokes like i can paint it to 1.\n\n\n\n[2022-11-18 18-43-59.mp4](2022-11-18_18-43-59.mp4)\n\n\n", "Grease pencil eraser settings not loaded on startup when using ctrl+LMB with the Draw tool\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Intel(R) HD Graphics 620 Intel 4.5.0 - Build 26.20.100.7925\n\nBroken: version: 2.90.0\n\nWhen first opening a blend file, your eraser settings are set to default when using ctrl+LMB with the draw tool, and only update after switching to the eraser tool and using it.\n\n\n- Open attached file (the Stroke object is selected and the Draw tool activated in Draw mode).\n- Draw a line\n- Erase part of the line using the same tool but with Ctrl selected. Notice how it behaves as if the `\"Eraser Soft\"` brush is used. \n- Switch to the Eraser tool, erase part of the line. Notice how it behaves according to the specified Brush: `\"Eraser Point\"`. \n- Switch back to Draw tool, and hold down ctrl to erase another part of the line, it should now start behaving like the `\"Eraser Point\"` brush.\n\n[eraser bug.blend](eraser_bug.blend)\n\n", "Grease pencil. Artefacts at beginning of strokes\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 3.0.0 Alpha\n\nGrease pencil strokes drawn with a stylus have unwanted artefacts at their beginning. Especially noticeable at high radius settings. Related to this, the start of a stroke is often narrower than the artist's intention.\n\n\n\n* Open new document, 2d animation\n* Set brush strength to 1 (not using pressure)\n* Set brush radius to 200 (using pressure)\n* draw several lines with a stylus\n* zoom in to inspect the start of the lines. There are artefacts, a 'lumpy' line close to the beginning of the stoke (see screenshot)\n\n**Notes**\nWhen you draw with deformable physical natural media (e.g. a paint brush, felt-tip) adding pressure increases the size of the mark, even when the tip is not moving over the paper. Other drawing programs also work this way. Currently blender doesn't work this way.\n\nI looked at the source code (gpencil_paint.c). Here's what i think is going wrong with the grease pencil strokes:\n\nWhen the pressure difference between two close adjacent vertices is large this can cause ugly 'lumps' in the line.\n\nCurrently, stylus pressure can be very different between the first and second vertex recorded in a grease pencil stroke. This is because there can be a large time difference between the recording of the two points; The first vertex stores a record of the pressure only at the moment the stylus first touches the tablet, subsequent changes of pressure before the stylus begins to move across the tablet are not recorded in the stroke. Only after the stylus starts moving is the second point recorded with the current pressure.\n\n**Proposed solution**\nAny time the stationary stylus tip is touching the tablet, the drawing routine polls the changing tablet pressure continuously. When the detected pressure is higher than the pressure of the last recorded point, the pressure of the last recorded point is overwritten by the newly detected (higher) pressure. This way, the pressure of the first and second points in a stroke is much more likely to be similar, the width of the start of the line will more closely match the artists intention, and the likelihood of artefacts is reduced.", "GP kills bloom\nOperating system: Linux-5.15.0-83-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: GeForce GT 630/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.157\n\nBroken: version: 3.6.3 Release Candidate\nWorked: \n\nWhen you have mesh objects emitting in front of grease pencil objects, bloom doesn't work there.\n\n\n", "overlay.show_paint_wire doesn't work in vertex paint modes depending on edit mode selection\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 3.0.0 Alpha\nWorked: 2.80?\n\nThis overlay option seems to take into account edit-mode selection\n\n\nHere's what it's supposed to do:\n[viIeK7Tkv0.mp4](viIeK7Tkv0.mp4)\n\nBut on 2.93 and 3.0 it appears to do nothing in Weight/Vertex Paint mode if the geometry is selected in edit mode. This seems like a bug, unless it's some odd design choice?\n\n- Create a cube and select all its verts in edit mode.\n- Enter weight paint mode\n- Toggle the Show Wire setting under Viewport Overlays->Weight Paint (NOT \"Wireframe\" under Viewport Overlays->Geometry)\n- Nothing happens\n- Switch to edit mode\n- Deselect all with Alt+A\n- Go back to weight paint mode\n- Wireframe display now works as expected.\n\n\nNote: This now changed again in 44d2479dc3 (now it does not draw **at all** -- reported separately in #89880 (overlay.show_paint_wire doesn't work anymore))", "Impossible to move with G key an object scale to 0 and parent to an empty\nOperating system: Windows 7\nGraphics card: GTX 780\n\nBroken: 3.5.1\nWorked: \n\nWhen an object is parented to an empty, and this object is scale to 0, you cannot move it anymore with the G key. But you can move it in the transform menu location.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n1. Open blend.\n2. Try to move object with G key.\n3. It is possible with Transfom menu.\n\nThanks a lot !\ndono", "Weight Paint Vertex Selection Display issue\nOperating system: Linux Mint\nGraphics card: RTX 3070\n\nBroken: latest master\nWorked: never really, more on that below\n\nWhen in weight paint mode with\n* vertex selection enabled\n* show wire enabled\n\nedges will not be drawn within the selection\n\n\n**The odd part**\nClicking in an empty space will get the wires to show again, but keep the vertices selected. \nThat's a behaviour that has changed from 3.1.2 to 3.2.2\nPreviously it deselected the vertices but didn't draw the edges\n\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n* Open the attached blender file\n* you can see that edges are not drawn between selected vertices\n* click in an empty space (important to click, not click drag)\n* vertices stay selected but wires are drawn\n\n", "Grease Pencil Arrange Stroke operator does not function with Curve Editing\nOperating system: Linux 64-bit\nGraphics card: GTX 1080 ti\n\n2.93.5 and 3.0.0 alpha\n\nIn edit mode, \"Arrange Stroke\" operator does not work when use_curve_edit is enabled.\n\n[GP_arrange_bug.blend](GP_arrange_bug.blend)\n\n - Open attached file\n - Use any of the Arrange Stroke operators on the selected stroke\n # Disable Curve Editing and try again to see that it otherwise works as intended\n\n", "Grease pencil eyedropper doesn't sample correct vertex color.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken:\n```\nversion: 2.90.0\nversion: 2.92.0 Alpha\nversion: 2.83.9\n```\n\nWorked: (unknown maybe never)\n\nWhen using the grease pencil eye dropper with filmic color management selected; the sampled color is different than the original vertex stroke color. \nWorks as expected with standard color management selected\n(Note this is confusing to new users as they expect to color sample and return the previous color of their previous stroke.)\n\n\n - Start with a new file\n - Set color management to filmic.\n - Add grease pencil object\n - Enter Draw Mode\n - Paint in vertex color mode using a lighter color.\n - With a blank palette use the color eyedropper tool on the stroke you just made\n - now paint with the sampled color in the palette\n - result should be slightly darker than the first stroke.\n\nvideo reference:\n**with filmic enabled:**\n[Desktop 2020.11.25 - 02.35.06.03.mp4](Desktop_2020.11.25_-_02.35.06.03.mp4)\n\n**with filmic disabled:**\n[Desktop 2020.11.25 - 02.36.30.06.mp4](Desktop_2020.11.25_-_02.36.30.06.mp4)\n\n\n", "Grease Pencil Tint / Also Vertex Paint on Grease Pencil Objects very slow with Cintiq Pro 24\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.4.1\n\nWhen i use Grease Pencil Tint with the a Cintiq 24 Pro it's very very slow. I have to wait after each stroke to continue. Drawing with the mouse is no problem.\nSame with vertex paint on a grease pencil object.\n\nWindows Ink and Wintab made no difference\n\nJust tinting with the cintiq stylus in the attached scene on the current layer.\n\n", "Weight paint without weight leaks particles\nOperating system: Windows 10\nGraphics card: GTX1070\n\nBroken: 2.83\nWorked: ?\n\nIf I create a vertex group/weight paint without assigning any faces to it and then use it as a \"density\" in the vertex groups for particles it still creates particles.\nYou might not think this is a bug, but together with \"Dynamic Paint\" it gets to be one.\nThis because there you can use type \"weight\" on particles. \nIn most cases you will then get frames of unused weight paints when particles are not affecting the weight paint.\nIn those cases there is a leakage of particles in the center of the weight paint, which creates a lot of disturbance in almost all cases when you are trying to do some stuff where you mix weight paint and particles.\n\nYou can do this in many ways. I will show two;\nWay no 1.\n1) Create a plane as emitter.\n2) Create a vertex group on that plane\n3) Add a particle system to that plane.\n4) Add the vertex group you created in step 2 to \"Density\" in \"Vertex groups\" section for particles.\nRun an animation... the particles are coming even if the weight paint now is unused and \"blue\".\n[T75859_easy.blend.zip](T75859_easy.blend.zip)\n\nWay no 2 using mask modifier and dynamic paint. \"The hard\" way.\n1) Create a plane as an emitter and call it \"emitter\".\n2) Create an icosphere called \"Particle\" and move it a bit so it is away from the plane.\n3) Now create another plane called \"canvas\" below the first plane.\n4) Subdivide \"canvas\" like 10 times so you get a lot of faces on it.\n5) Create a vertex group called \"MyVertexGroup\" on the \"Canvas\"\n6) In \"Emitter\", just create a default particle system.\n7) Set \"render\" to object and Instance Object to \"Particle\"\n8) Bake it.\n9) Now add dynamic paint of type \"Brush\" to the \"Emitter\" plane.\n10) Set the brush to \"Particlesystem\" and select the particlesystem you created\n11) Select the \"Canvas\" and do a dynamic paint of type \"Canvas\" on that one.\n12) Set Surface type to \"Weight\", check \"Dissolve\" and set it to 5 frames and uncheck \"Slow\".\n13) Save your .blend file.\n14) In \"Output\" of the canvas for the dynamic paint, select the vertex group \"MyVertexGroup\"\n15) Bake.\n16) Run an animation in \"Weight paint\" mode and see so it works (creates red areas when a particle hits the weight paint)\n17) Go back to object mode (if 16 works, which it should).\n18) Select the \"Canvas\" and add modifier \"mask\". In mask select vertex group \"MyVertexGroup\"\n19) Run animation and see so the mask seems to work.\n20) Select \"Canvas\" and create a particlesystem.\n21) In \"Source\" for that system, check \"Use Modifier stack\".\n22) Run animation.\nNow the particles are created when the mask is visible... but if the mask is NOT visible, the particles are still created in the center of the object instead of being \"off\".\nIf you use \"density\" or not on that particlesystem does not matter. Still same behavior.\n[T75859_hard.zip](T75859_hard.zip) \n\n\nin console: \n```\nCan't create particles with the current modifier stack, disable destructive modifiers\n```\n", "Dynamic vertex groups (Dynamic Paint, Weight Proximity modifier, Geometry Nodes, ...) do not update density of particles in viewport display\nOperating system: windows 10\nGraphics card: Quadro M620\n\nBroken: 2.90\nWorked: Never\n\nI want to use my dynamic vertex group to generate particles on a plane.\nIt just works when I change parameters during the playback but it's not realtime when I hit play.\n\n- Open attached file\n- Play animation\n- Change any property of the particle during the playback\n[dynamic_paint_test.blend](dynamic_paint_test.blend)", "State of hidden geometry not synced between edit mode and vertex paint/weight paint mode\nOperating system: Linux-5.8.0-7630-generic-x86_64-with-glibc2.32 64 Bits\nGraphics card: Quadro RTX 6000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.32.03\n\nBroken: version: 3.0.0 Alpha\n\nWhen vertex selection (or face selection) is used in either vertex paint or weightpaint mode the state of hidden geometry is not synchronized but overwritten by some specific state on this object. That makes it impossible to paint on using vertex selection. Joining the mesh into a new object did not fix the issue.\n[hidden_geo-2021-04-29_20.57.11.mp4](hidden_geo-2021-04-29_20.57.11.mp4)\n\nSee file:\n[hidden_geo.blend](hidden_geo.blend)\n\n" ]

[ "Regression: Select tool does not work on bones in weight paint grease pencil\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.5.0\n\nCaused by 2eeec49640b3df3f975d947936998cf05443424a\n\n[Previously, in blender 3.3 when regify bones with greese pensil you could in the weght paint mode click on the bone with the left mouse button pressed and the required bone was selected, but now in version 3.5 and alpha 3.6 this function does not work which complicates the rigify process]\n\n[1. Select new 2D animation\n2. Draw a simplified arm\n3.Create on top of the bone\nCombine with the empty groups. \n5. Hold down the shifter and left click first on the bone then on the greese pensil\n6. Go to Weight paint mode\n7. and try holding down the shifter and click on the desired bone, it does not work and in blender version 3.3 it worked]\n\n" ]

[ "Blender sometimes crashes when loading a file with unregistered nodes\nOperating system: Windows 10 64 Bit\nGraphics card: Nvidia Geforce GTX 750 Ti\n\nBroken: 2.83.7 release\n\nSometimes when opening a file with unregistered custom nodes, Blender will crash with an `EXCEPTION_ACCESS_VIOLATION` error.\n\n**Stack trace from Visual Studio:**\n```\nblender.exe!node_free_node(bNodeTree * ntree, bNode * node) Zeile 2033\n\tunter C:\\blender-git\\blender\\source\\blender\\blenkernel\\intern\\node.c (2033)\nblender.exe!ntree_free_data(ID * id) Zeile 227\n\tunter C:\\blender-git\\blender\\source\\blender\\blenkernel\\intern\\node.c (227)\nblender.exe!BKE_libblock_free_datablock(ID * id, const int UNUSED_flag) Zeile 136\n\tunter C:\\blender-git\\blender\\source\\blender\\blenkernel\\intern\\lib_id_delete.c (136)\nblender.exe!DEG::deg_free_copy_on_write_datablock(ID * id_cow) Zeile 1073\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval_copy_on_write.cc (1073)\nblender.exe!DEG::deg_update_copy_on_write_datablock(const DEG::Depsgraph * depsgraph, const DEG::IDNode * id_node) Zeile 952\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval_copy_on_write.cc (952)\nblender.exe!DEG::deg_evaluate_copy_on_write(Depsgraph * graph, const DEG::IDNode * id_node) Zeile 1088\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval_copy_on_write.cc (1088)\nblender.exe!std::invoke<void (__cdecl*&)(Depsgraph *,DEG::IDNode const *),Depsgraph *,DEG::IDNode * &>(void(*)(Depsgraph *, const DEG::IDNode *) & _Obj, Depsgraph * && _Arg1, DEG::IDNode * & <_Args2_0>) Zeile 1626\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\type_traits (1626)\nblender.exe!std::_Invoker_ret<std::_Unforced,0>::_Call<void (__cdecl*&)(Depsgraph *,DEG::IDNode const *),Depsgraph *,DEG::IDNode * &>(void(*)(Depsgraph *, const DEG::IDNode *) & <_Vals_0>, Depsgraph * && <_Vals_1>, DEG::IDNode * & <_Vals_2>) Zeile 1658\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\type_traits (1658)\nblender.exe!std::_Call_binder<std::_Unforced,0,1,void (__cdecl*)(Depsgraph *,DEG::IDNode const *),std::tuple<std::_Ph<1>,DEG::IDNode *>,std::tuple<Depsgraph * &&>>(std::_Invoker_ret<std::_Unforced,0> __formal, std::integer_sequence<unsigned __int64,0,1> __formal, void(*)(Depsgraph *, const DEG::IDNode *) & _Obj, std::tuple<std::_Ph<1>,DEG::IDNode *> & _Tpl, std::tuple<Depsgraph * &&> && _Ut) Zeile 1388\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\functional (1388)\nblender.exe!std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &>::operator()<Depsgraph *>(Depsgraph * && <_Unbargs_0>) Zeile 1427\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\functional (1427)\nblender.exe!std::invoke<std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &> &,Depsgraph *>(std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &> & _Obj, Depsgraph * && _Arg1) Zeile 1626\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\type_traits (1626)\nblender.exe!std::_Invoker_ret<void,1>::_Call<std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &> &,Depsgraph *>(std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &> & <_Vals_0>, Depsgraph * && <_Vals_1>) Zeile 1641\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\type_traits (1641)\nblender.exe!std::_Func_impl_no_alloc<std::_Binder<std::_Unforced,void (__cdecl&)(Depsgraph *,DEG::IDNode const *),std::_Ph<1> const &,DEG::IDNode * &>,void,Depsgraph *>::_Do_call(Depsgraph * && <_Args_0>) Zeile 904\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\functional (904)\nblender.exe!std::_Func_class<void,Depsgraph *>::operator()(Depsgraph * <_Args_0>) Zeile 952\n\tunter A:\\Programs\\Microsoft Visual Studio\\2019\\Community\\VC\\Tools\\MSVC\\14.27.29110\\include\\functional (952)\nblender.exe!DEG::`anonymous namespace'::evaluate_node(const DEG::`anonymous-namespace'::DepsgraphEvalState * state, DEG::OperationNode * operation_node) Zeile 117\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval.cc (117)\nblender.exe!DEG::`anonymous namespace'::deg_task_run_func(TaskPool * pool, void * taskdata, int thread_id) Zeile 129\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval.cc (129)\nblender.exe!BLI_task_pool_work_and_wait(TaskPool * pool) Zeile 939\n\tunter C:\\blender-git\\blender\\source\\blender\\blenlib\\intern\\task_pool.cc (939)\nblender.exe!BLI_task_pool_work_wait_and_reset(TaskPool * pool) Zeile 970\n\tunter C:\\blender-git\\blender\\source\\blender\\blenlib\\intern\\task_pool.cc (970)\nblender.exe!DEG::deg_evaluate_on_refresh(DEG::Depsgraph * graph) Zeile 403\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\eval\\deg_eval.cc (403)\nblender.exe!DEG_evaluate_on_refresh(Main * bmain, Depsgraph * graph) Zeile 65\n\tunter C:\\blender-git\\blender\\source\\blender\\depsgraph\\intern\\depsgraph_eval.cc (65)\nblender.exe!scene_graph_update_tagged(Depsgraph * depsgraph, Main * bmain, bool only_if_tagged) Zeile 1329\n\tunter C:\\blender-git\\blender\\source\\blender\\blenkernel\\intern\\scene.c (1329)\nblender.exe!BKE_scene_graph_update_tagged(Depsgraph * depsgraph, Main * bmain) Zeile 1367\n\tunter C:\\blender-git\\blender\\source\\blender\\blenkernel\\intern\\scene.c (1367)\nblender.exe!wm_event_do_depsgraph(bContext * C, bool is_after_open_file) Zeile 360\n\tunter C:\\blender-git\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c (360)\nblender.exe!wm_event_do_refresh_wm_and_depsgraph(bContext * C) Zeile 387\n\tunter C:\\blender-git\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c (387)\nblender.exe!wm_event_do_notifiers(bContext * C) Zeile 552\n\tunter C:\\blender-git\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c (552)\nblender.exe!WM_main(bContext * C) Zeile 456\n\tunter C:\\blender-git\\blender\\source\\blender\\windowmanager\\intern\\wm.c (456)\nblender.exe!main(int argc, const unsigned char * * UNUSED_argv_c) Zeile 530\n\tunter C:\\blender-git\\blender\\source\\creator\\creator.c (530)\nblender.exe!invoke_main() Zeile 79\n\tunter D:\\agent\\_work\\9\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl (79)\nblender.exe!__scrt_common_main_seh() Zeile 288\n\tunter D:\\agent\\_work\\9\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl (288)\nblender.exe!__scrt_common_main() Zeile 331\n\tunter D:\\agent\\_work\\9\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl (331)\nblender.exe!mainCRTStartup() Zeile 17\n\tunter D:\\agent\\_work\\9\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_main.cpp (17)\nkernel32.dll!00007ffca1667bd4()\nntdll.dll!00007ffca1e0ce51()\n```\n\nThe exception in line 2033 of node.c says that there is an access violation when reading at position `0xFFFFFFFFFFFFFFFF`. `node->typeinfo` does exist (is not null), but `freefunc` (and actually most of the other attributes) point to `0xdddd...`. Because of that, I think that this bug might be caused by a similar issue as described in #72833.\n\nI'm not sure if a core dump helps you, it's quite large (1GB + another GB for the .pdb file) so if you need it I can look for a way to make it available to you. I'm not a C or C++ dev so I don't know what is helpful.\n\n**EDIT**: I think this crash is caused by opening files files with nodes that were registered while the file was originaly saved but not anymore.\n\nIt's a bit complicated because it doesn't happen all the time. The particular file(s) where I observed this issue use nodes from an addon (Armory3D) which supports per-project custom nodes which get registered via a `on_load_post()` handler (so there is a delay between opening the file and registering the nodes). I'm not entirely sure, but it seems like the crash occurs when the nodes are registered again. The crash even occurs when unregistering the nodes before saving the file.\n\nBecause of that complicated setup (addon required + project setup + custom node package) I will go without an example file. If you aren't able to find the cause without one, please tell me so that I can prepare a minimal project setup for you.\n\nThank you very much!\n\n*Original issue*: 1890", "Line Art Thickness doesn't consider Render Resolution Percentage\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 30.0.101.1122\n\nBroken: version: 3.2.2\nBroken: version: 3.4.0 Alpha\nBroken\n\nWhen using Line Art Modifier, on a Grease Pencil Object in Screen Space Mode, Line Thickness doesn't scale with resolution as expected. \n\n1. Open new General Startup File\n2. Create New Object>Grease Pencil>Scene Line Art\n3. Under Object Data Properties>Strokes>Stroke Thickness and set to World Space\n4. Set Render Resolution % to a factor. line thickness will be different than expected\n\n\n\n\n\n", "Crash when updating image pixel data through python API\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.83.0, 2.90\nWorked: N/A\n\nBlender crashes with apparent memory corruption, when image pixel data and previous are updated frequently with python script.\nStack traces give no obvious answer, i have seen it happen in depsgraph eval, node graph code, imbuf, etc.\n\n*Note: i could not reproduce this issue in debug mode so far, only release or RelWithDebInfo seems to work.*\n\n 1. Open attacehd file\n - Run the script\n - Make sure the scene properties are visible (it seems to only crash when displaying image previews).\n - Grab the default cube and wiggle it around to force depsgraph updates.\n - If it doesn't want to crash, re-open Blender, it seems to be quite random.\n\nI'm using a node group to access image data blocks, so they can be displayed in a panel as pointers. The code generates data to replace the image pixels as well as the preview.\nI have also tried running a timer and deferring pixel updates using a queue as suggested [here ](bpy.app.timers.html#use-a-timer-to-react-to-events-in-another-thread), just in case the depsgraph post update is somehow not safe, but this doesn't change anything substantially. The bug happens in both cases.\n\n[crashtest.blend](crashtest.blend)", "Regression: Crash with boolean modifier\nOperating system: Fedora Linux 36 (Workstation Edition)\nGraphics card: Nvidia Geforce RTX 3080\n\nBroken: 3.2.1, 3.3 alpha\nWorked: 3.1\n\nCrash with boolean modifier\n\n- Open .blend file\n- In boolean modifier tab -> select `dot cutter` for boolean operation\n[yin_yang-v1.blend](yin_yang-v1.blend)\noriginal file: [F13292564](yin_yang.blend)", "stack_overflow exception while rendering with viewport render on\nOperating system: win10\nGraphics card: 2070super\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: 2.83.4, 2.91 6fe609d4c3\n\n\nwhen pressing f12 in render mode, blender crashes,\nwhen pressing render in solid mode, blender doesnt crash\n\n[blender_debug_output.txt](blender_debug_output.txt)\n```\nError : EXCEPTION_STACK_OVERFLOW\nAddress : 0x00007FFF42963FDB\nModule : C:\\Windows\\system32\\DriverStore\\FileRepository\\nv_dispsi.inf_amd64_6af46268c19e4980\\nvcuda64.dll\n```\n[2.blend](2.blend) \n1. switch to render mode\n2. press f12 & watch task manager.\n3. blender will close", "Holdout Fill not working as expected with Grease Pencil\nOperating system: Win 10\nGraphics card: Nvidia GTX 1060\n\nBroken: 2.93\nWorked: unknown\n\nI would expect holdout fills to render as 100% transparent. In Blender's image viewer, they don't - their opacity seems to change based on how bright they are, so white is 100% opaque, but grey is semi-transparent.\n\nSaving the same image to a PNG file, the holdouts are fully transparent, but with aliased artefacts around the edge of the surrounding black strokes, in the colour of the holdout fill. If there is another grease pencil object positioned behind the holdout fill, instead of vanishing, its strokes are tinted the colour of the holdout fill.\n\n\n(The artefacts are more noticeable against a dark or checkered background.)\n\nTry rendering the simple drawing in the attached .blend, exporting a PNG and examining it in an image editor. [holdouttest.blend](holdouttest.blend)", "Snap support for Grease Pencil objects\n**Description**\n**Big picture:** Make it possible for the snap options: `Vertices, Edges and Face`, to work on Grease Pencil objects\n\n**Use cases**:\n* Precisely positioning a Grease Pencil object on a Mesh in the 3DView\n* More flexibility in positioning the 3D Cursor\n\n**Design:**\n* Will work in conjunction with 3DView's snap system (no new interface required)\n\n**Engineer plan:**\n* Since Grease Pencil objects can be 3D, a BVHTree system may have to be implemented to snap to faces and occlusion test.\n* Another option is to work with a depth buffer.", "Extremely frequent crashing when using the compositor. Especially when using the glare node.\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\nWorked: I believe it was far more stable before the realtime compositor implementation. (Blender 3.3-ish)\n\nBlender crashes extremely frequently when using the compositor on my system, especially when the glare node is being used at all. Does not seem related to vram or system specs, and looking through logs has not come up with any errors - it just closes.\n\nAdjust compositor glare node values with the viewer enabled. It seems to crash within a minute of doing so, usually sooner.\n\n", "Grease Pencil Fill Tool Breaks After Scroll-Wheel Zooming Window\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.91.0 Beta\nAlso tested in 2.83.9 & 2.92 Alpha on Windows, and 2.91 Beta on Mac with same results.\n\nIf an 'unclosed' area made with grease pencil lines on its own layer overlaps and is 'closed' with grease pencil lines on a second layer, zooming the 2d workspace in and out will cause the fill tool to 'break' and fill the whole screen; even when the fill tool is set to: Advanced-Layers-Visible. \n\nCreate closed shape on one layer. Create 'open' shape on second layer that overlaps and is closed by grease pencil lines from the first layer. Zoom in and out in workspace while using the fill tool in a separate 'fill' layer. \nPlease see attached .blend file and screen capture: \n\n[FillTest.blend](FillTest.blend)\n\n[FillTest.mp4](FillTest.mp4)\n\n\nI really hope this isn't just user error. Thank you for taking a look!\n", "Grease Pencil: various issues\n*Note: this tasks can be split up once we get feedback on issues, for now this is to discuss changes to make.*\n\n----\n\n\n- Brush pixels are not pixels,\n```\nshouldn't we use a meaningful value?\n(pixels or blender units).\n```\n\n- Grease pencil brushes have material pointers which are globally overwritten.\n\n```\nWhen you set a material, all brushes are looped over and have their material values clobbered.\n```\n\n```\nWhy have a material attached to the brush at all? This could be a scene level setting (or at least not stored per-brush).\n```\n\n```\nSee: `BKE_brush_update_material` called by rna `rna_Brush_material_update` when setting any material.\n```\n\n- Grease pencil sculpt uses it's own brush system,\n```\nis there a good reason for this?\n```\n\n```\nSince these brushes are fairly advanced (pressure & curve mapping options)\nwouldn't it be better to use the existing brush system?\n```\n\n *Note: while particle system for eg: has 'simple' brush types like this, grease-pencil sculpt brushes are more advanced and duplicate more of what we have for regular brushes used elsewhere.* \n\n- Grease pencil weight brush seems to have no way to paint a weight value (only add weight).\n\n```\nAs with mesh weight paint we should be able to paint a weight.\n```\n\n- In edit-mode, vertex selection exists but is hidden by default (0.0 opacity).\n\n```\nShouldn't they be visible?\n```\n\n----\n\n### Lower Priority\n\n- Internally grease pencil sculp is called:\n\n - grease pencil edit (`GP_BrushEdit_Settings`, `GP_EditBrush_Data`, `GP_BRUSHEDIT_FLAG_APPLY_UV`)\n\n```\nThis is highly confusing since we have grease pencil edit-mode, can't we rename these `GP_Sculpt*`. ?\n```\n\n - grease pencil paint (`GPENCIL_OT_brush_paint`)\n\n```\nPaint is easily confused with draw, seems this is for sculpt/weight painting.", "Blender crashes when using bpy.ops.view3d.view_orbit with a custom context\nOperating system: Linux-5.15.11-arch2-1-x86_64-with-glibc2.33 64 Bits\nGraphics card: AMD RENOIR (DRM 3.42.0, 5.15.11-arch2-1, LLVM 13.0.0) AMD 4.6 (Core Profile) Mesa 21.3.2\n\nAlso tested on WIndows:\n\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: AMD Radeon(TM) Graphics ATI Technologies Inc. 4.5.14756 Core Profile Context 20.40.32 27.20.14032.8\n\nBroken (Linux): version: 3.0.0\nBroken (Windows): version: 3.0.0\nWorked: Never, the crash happens in 2.7 and 2.6 doesn't seem to support the override thing (I haven't tested subversions)\n\nBlender crashes when using `bpy.ops.view3d.view_orbit` with a custom context\n\nRun this in the Python console:\n```\narea = [n for n in bpy.context.screen.areas if n.type == \"VIEW_3D\"][0]\noverride = bpy.context.copy()\noverride['area'] = area\nbpy.ops.view3d.view_orbit(override, type='ORBITLEFT')\n```\n\n[blender.crash.txt](blender.crash.txt)", "GP transparent layer doesn't blend with background object \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 3.0.0 Alpha\n\nI have a foreground grease pencil object, and a background grease pencil object (yellow blob).\nThe foreground object contains a layer set to 0 opacity.\nWhere this layer overlaps the background object in the viewport, the background object is invisible.\nExpected behaviour: the visibility of the background grease pencil object isn't affected by transparent layers in a foreground object.\n\n\n\n\n\n[#89286.blend](T89286.blend)\n\nForeground object \"makes hole\" in background, doesn't blend. Rotate viewport camera, suddenly it blends correctly.\nWhen foreground location is set to 0,0,0 it blends correctly too.", "Crash when adding materials\nOperating system: Linux-5.18.9-200.fc36.x86_64-x86_64-with-glibc2.35 64 Bits\nGraphics card: Mesa Intel(R) UHD Graphics 630 (CFL GT2) Intel 4.6 (Core Profile) Mesa 22.1.3\n\nBroken: version: 3.2.0\n\nBlender crashes when adding a second material (sometimes you need to add some more to crash it)\n\n1. Open the appended .blend file.\n2. Select the object.\n3. Go to material tab.\n4. Try to add a second material.\n\n[crash.blend](crash.blend)", "Crash when sculpting with a texture, deleting the image of the texture and exiting sculpt mode\nOperating system: Windows 10\nGraphics card: RTX 2070 Super\n\nBroken: version: 3.4, 3.5, 3.6\nBroken: version: 3.3.3\nBroken: 3.1.0\nWorked: 3.0.1\nWorked: 2.93 LTS\n\nCaused by 9111ea78ac\n\nBlender crashes when a texture image is deleted and you exit sculpt mode.\n\nYou can test it inside the attached blend file. I basically just attached an image to a sculpting brush.\n\n1. Open the blend file below\n2. draw on the surface of the sphere\n3. delete the image inside the outliner\n4. exit sculpt mode\n\nCrash occurs because `item->image->runtime.cache_mutex` is null inside `BKE_image_pool_free`\n```\nStack trace:\nblender.exe :0x00007FF713D04C10 pthread_mutex_lock\nblender.exe :0x00007FF70E6EE880 BKE_image_pool_free\nblender.exe :0x00007FF70E6B5CF0 BKE_sculptsession_free\nblender.exe :0x00007FF70F1B67A0 ED_object_sculptmode_exit_ex\nblender.exe :0x00007FF70F1B85F0 sculpt_mode_toggle_exec\nblender.exe :0x00007FF70E953220 wm_operator_invoke\nblender.exe :0x00007FF70E9528D0 wm_operator_call_internal\nblender.exe :0x00007FF70E955880 WM_operator_name_call_ptr\nblender.exe :0x00007FF70F171800 ED_object_mode_set_ex\nblender.exe :0x00007FF70F148710 object_mode_set_exec\nblender.exe :0x00007FF70E953220 wm_operator_invoke\n```", "blender crash itself while importing a dae (collide) file\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1650 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.1\nBroken: version: 2.82 and all versions in between then and now\nWorked: n/a\n\nHi, I'm using blender version 3.3 on a windows 10 downloaded on steam.\n\nwhen i try to import dae files, blender crashes itself and shuts down\n(this only happens with some dae files, not all and the ones who crash are not to big ).\n\nI tried to search with powershell to find the problem and I got this:\n\nError: EXCEPTION_ACCESS_VIOLATION\n[bgr_darryls_ship_ground.dae](bgr_darryls_ship_ground.dae)\n\n```\nThe gh->hashfp function pointer is NULL (hasn't been initialized yet)\n\nblender.exe!ghash_keyhash(const GHash * gh, const void * key) Line 133\tC\nblender.exe!ghash_lookup_entry(const GHash * gh, const void * key) Line 411\tC\nblender.exe!BLI_ghash_lookup(const GHash * gh, const void * key) Line 736\tC\nblender.exe!BKE_armature_find_bone_name(bArmature * arm, const unsigned char * name) Line 601\tC\nblender.exe!ArmatureImporter::create_armature_bones(Main * bmain, std::vector<Object *,std::allocator<Object *>> & arm_objs) Line 478\tC++\nblender.exe!ArmatureImporter::make_armatures(bContext * C, std::vector<Object *,std::allocator<Object *>> & objects_to_scale) Line 826\tC++\nblender.exe!DocumentImporter::finish() Line 228\tC++\n[External Code]\t\nblender.exe!DocumentImporter::import() Line 127\tC++\n```\n" ]

[ "Can't open file with modifiers to a grease pencil library override\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 650 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.92.0\n\nBlender can't open files that have a library link to a grease pencil object in another file, if the linking file has a modifier on the grease pencil object.\n\n - Create a new blender file.\n - Delete the default cube, light and camera.\n - Create a grease pencil Suzanne.\n - Create a new blender file.\n - In the new file create a library link to the collection where the Suzanne object is.\n - Create a library override to that collection.\n - Add a thickness modifier to the Suzanne object.\n - Close blender.\n # Try opening the second file (the one with the library override) - blender gets stuck.\n\nExample files:\n[LibraryBug.zip](LibraryBug.zip)\n\nI've found some workarounds to get to the file that blender can't open:\n - Change the name of the original file so that the second file can't find it. This way blender will be able to open the second file.\n - Create a new blender file and append the contents from the second blender file.\n" ]

[ "Interpolated children on object rendered particles not working\nOperating system: Manjaro Linux\nGraphics card: AMD Radeon rx570\n\nBroken: 2.60, 2.79, 2.80, 2.81, 2.82, 2.82a, 2.83, 2.90Alpha\nWorking: Couldn't find any\n\nCan't see any particles when there set to render as an object with an object selected and interpolated children\n\n\n1. Open this file: [particle_child_interp_bug_2_79.blend](particle_child_interp_bug_2_79.blend)\n2. Play animation\n3. In the particle children settings, set Interpolate \nor \n1. Create a particle system on a uv sphere,\n2. set the render type to object,\n3. select a 2nd uv sphere as the render object\n4. set the children to interpolated\n[Daily Challenge 80(particles galore).blend](Daily_Challenge_80_particles_galore_.blend)", "Hair particles rendered as an object do not follow parent object when the parent object is falling under gravity as a rigid body when animating\nOperating system: Microsoft Windows 10 Home, Version 10.0.18362 Build 18362\nGraphics card: GeForce GTX 1080\n\nversion: 2.82 (sub 7), branch: master, commit date: 2020-03-12 05:06, hash: 375c7dc4caf4, type: Release\nbuild date: 2020-03-12, 15:41:08\nplatform: Windows\n\nI have an issue where i have a cube mesh, I create hair particles, and set the hair to render as an object, say a cylinder. I then enable rigid body physics for the cube. I also have a passive plane below the cube. When I run an animation, the cube falls as if under gravity and stops at the plane, but the cylinder particles remain in their original position?\n\nIf I add soft body physics to the cube and order the soft body physics before the particle settings, then the cylinders fall, but lag behind the cube. The cube falls through the plane below, but the cylinders are stopped by the plane?\n\nWith soft body added (to top of stack). Particles do fall but have a lower acceleration rate, and the cube no longer detects the floor and passes through the plane?!\n\nThanks!!\n\nSee attached blend file [StackExchangeParticleProblem-001.blend](StackExchangeParticleProblem-001.blend)", "Killed particle's position is reset to the initial position when rendered\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.92.0\n\nI am encountering a strange particle behavior.\nWhen a particle collides with a plane where collision is enabled and \"kill particles\" is checked, the particle position seems to be reset to its initial position. This can not be observed in the viewport. This unexpected position reset is observed only when rendering the images.\nAs shown in this video:\n[0001-0050.mp4](0001-0050.mp4)\n\n[debug.blend](debug.blend)\nBake particle simulation and render images. You will get the same video as shown above.\n", "Rendering animation using \"Generate\" type modifiers after animated displacement texture crashes Blender\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.79\n\nBroken: version: 2.92.0 Alpha\nWorked: 2.83.10\n\nIf 2 high poly (~100k faces per obj) objects in modifiers properties have after Displacement modifier (with animated texture) \"Generate\" type modifiers (I've tried Boolean and Decimate), Blender cant run Animation render (but still can run Image render). \nCrash happens on 2.9 and higher. I've tested this on 2.83.10 LTS (from steam) and everything worked fine there.\nCrashes with Decimate a little bit hard to recreate. So I will go with steps for \"Boolean crash\"\n\n- create Cude, Empty and Icosphere\n- Add Subdivision mod. to Cube: set it to \"Simple\" and 7 levels\n- Add Displacement mod. to Cube and set texture (Clouds) coordinates to object - Empty\n- Add Boolean mod. (difference) and select object Icosphere\n- Select Cude with Icosphere and duplicate them\n- Save and press Render Animation\n[bug.blend](bug.blend)\n[bug.crash.txt](bug.crash.txt)\n\n```lines\n## 3883==ERROR: AddressSanitizer: heap-use-after-free on address 0x61500034ec08 at pc 0x000016fc795d bp 0x7fffc19dd6c0 sp 0x7fffc19dd6b0\nREAD of size 8 at 0x61500034ec08 thread #15\n - 0 0x16fc795c in blender::gpu::GLBatch::verts_(int) const /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.hh:112\n - 1 0x17013501 in blender::gpu::GLVertArray::update_bindings(unsigned int, GPUBatch const*, blender::gpu::ShaderInterface const*, int) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_vertex_array.cc:124\n - 2 0x16fa900b in blender::gpu::GLVaoCache::vao_get(GPUBatch*) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.cc:273\n - 3 0x16fa84d8 in blender::gpu::GLVaoCache::base_instance_vao_get(GPUBatch*, int) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.cc:234\n - 4 0x16fa9d3e in blender::gpu::GLBatch::bind(int) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.cc:317\n - 5 0x16fa9f3f in blender::gpu::GLBatch::draw(int, int, int, int) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.cc:328\n - 6 0x16eca2be in GPU_batch_draw_advanced /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_batch.cc:282\n - 7 0x16fc1207 in blender::gpu::GLDrawList::append(GPUBatch*, int, int) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_drawlist.cc:122\n - 8 0x16efa9f5 in GPU_draw_list_append /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_drawlist.cc:52\n - 9 0x5978acc in draw_indirect_call /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:598\n - 10 0x5978acc in draw_call_batching_flush /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:795\n - 11 0x597c5f0 in draw_call_batching_finish /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:891\n - 12 0x597eae7 in draw_shgroup /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:1056\n - 13 0x597f89f in drw_draw_pass_ex /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:1113\n - 14 0x597fe50 in DRW_draw_pass /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager_exec.c:1153\n - 15 0x5c51bc0 in EEVEE_velocity_resolve /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_effects.c:466\n - 16 0x5c5202b in EEVEE_draw_effects /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_effects.c:486\n - 17 0x5a21f42 in EEVEE_render_draw /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_render.c:671\n - 18 0x59b1dbc in eevee_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_engine.c:559\n - 19 0x5948820 in DRW_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager.c:1909\n - 20 0x14ff1d77 in engine_render_view_layer /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:786\n - 21 0x14ff40a3 in RE_engine_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:914\n - 22 0x1501aafb in do_render_3d /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1132\n - 23 0x1501c24a in do_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1209\n - 24 0x1501e036 in do_render_composite /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1349\n - 25 0x15022789 in do_render_all_options /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1615\n - 26 0x15027839 in RE_RenderFrame /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:2015\n - 27 0xd7b42cf in render_startjob /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/render/render_internal.c:630\n - 28 0x4fed4f1 in do_job_thread /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:389\n - 29 0x18de408d in tslot_thread_start /home/dev/01-data/01-git/blender-git/blender/source/blender/blenlib/intern/threads.cc:231\n - 30 0x7ffff7566608 in start_thread /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_create.c:477\n - 31 0x7ffff6f46292 in __clone (/lib/x86_64-linux-gnu/libc.so.6+0x122292)\n\n0x61500034ec08 is located 8 bytes inside of 512-byte region [0x61500034ec00,0x61500034ee00)\nfreed by thread #15 here:\n - 0 0x7ffff76927cf in __interceptor_free (/lib/x86_64-linux-gnu/libasan.so.5+0x10d7cf)\n - 1 0x18e0eab0 in MEM_lockfree_freeN /home/dev/01-data/01-git/blender-git/blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:129\n - 2 0x16ef56c6 in blender::gpu::GLVertBuf::operator delete(void*) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_vertex_buffer.hh:57\n - 3 0x17018021 in blender::gpu::GLVertBuf::~GLVertBuf() /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_vertex_buffer.hh:35\n - 4 0x16f8bb19 in blender::gpu::VertBuf::reference_remove() /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_vertex_buffer_private.hh:94\n - 5 0x16f877f5 in GPU_vertbuf_discard /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_vertex_buffer.cc:161\n - 6 0x5a0eab8 in EEVEE_motion_blur_cache_finish /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_motion_blur.c:449\n - 7 0x59b1bef in eevee_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_engine.c:542\n - 8 0x5948820 in DRW_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager.c:1909\n - 9 0x14ff1d77 in engine_render_view_layer /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:786\n - 10 0x14ff40a3 in RE_engine_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:914\n - 11 0x1501aafb in do_render_3d /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1132\n - 12 0x1501c24a in do_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1209\n - 13 0x1501e036 in do_render_composite /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1349\n - 14 0x15022789 in do_render_all_options /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1615\n - 15 0x15027839 in RE_RenderFrame /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:2015\n - 16 0xd7b42cf in render_startjob /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/render/render_internal.c:630\n - 17 0x4fed4f1 in do_job_thread /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:389\n - 18 0x18de408d in tslot_thread_start /home/dev/01-data/01-git/blender-git/blender/source/blender/blenlib/intern/threads.cc:231\n - 19 0x7ffff7566608 in start_thread /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_create.c:477\n\npreviously allocated by thread #15 here:\n - 0 0x7ffff7692bc8 in malloc (/lib/x86_64-linux-gnu/libasan.so.5+0x10dbc8)\n - 1 0x18e0f510 in MEM_lockfree_mallocN /home/dev/01-data/01-git/blender-git/blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:276\n - 2 0x16ef5676 in blender::gpu::GLVertBuf::operator new(unsigned long) /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_vertex_buffer.hh:57\n - 3 0x16ef70a2 in blender::gpu::GLBackend::vertbuf_alloc() /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_backend.hh:121\n - 4 0x16f85d10 in blender::gpu::VertBuf::duplicate() /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_vertex_buffer.cc:77\n - 5 0x16f8750d in GPU_vertbuf_duplicate /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/intern/gpu_vertex_buffer.cc:149\n - 6 0x5a0ee06 in EEVEE_motion_blur_cache_finish /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_motion_blur.c:467\n - 7 0x59b1953 in eevee_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/engines/eevee/eevee_engine.c:507\n - 8 0x5948820 in DRW_render_to_image /home/dev/01-data/01-git/blender-git/blender/source/blender/draw/intern/draw_manager.c:1909\n - 9 0x14ff1d77 in engine_render_view_layer /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:786\n - 10 0x14ff40a3 in RE_engine_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/engine.c:914\n - 11 0x1501aafb in do_render_3d /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1132\n - 12 0x1501c24a in do_render /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1209\n - 13 0x1501e036 in do_render_composite /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1349\n - 14 0x15022789 in do_render_all_options /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:1615\n - 15 0x15027839 in RE_RenderFrame /home/dev/01-data/01-git/blender-git/blender/source/blender/render/intern/pipeline.c:2015\n - 16 0xd7b42cf in render_startjob /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/render/render_internal.c:630\n - 17 0x4fed4f1 in do_job_thread /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:389\n - 18 0x18de408d in tslot_thread_start /home/dev/01-data/01-git/blender-git/blender/source/blender/blenlib/intern/threads.cc:231\n - 19 0x7ffff7566608 in start_thread /build/glibc-ZN95T4/glibc-2.31/nptl/pthread_create.c:477\n\nThread #15 created by T0 here:\n - 0 0x7ffff75bf805 in pthread_create (/lib/x86_64-linux-gnu/libasan.so.5+0x3a805)\n - 1 0x18de42b9 in BLI_threadpool_insert /home/dev/01-data/01-git/blender-git/blender/source/blender/blenlib/intern/threads.cc:245\n - 2 0x4fee516 in WM_jobs_start /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:484\n - 3 0xd7b88bd in screen_render_invoke /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/render/render_internal.c:1041\n - 4 0x4f8f3ba in wm_operator_invoke /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1293\n - 5 0x4f912f2 in wm_operator_call_internal /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1500\n - 6 0x4f91551 in WM_operator_name_call_ptr /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1548\n - 7 0x90f1c95 in ui_apply_but_funcs_after /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/interface/interface_handlers.c:932\n - 8 0x917b8d0 in ui_handler_region_menu /home/dev/01-data/01-git/blender-git/blender/source/blender/editors/interface/interface_handlers.c:10809\n - 9 0x4f88a9b in wm_handler_ui_call /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:636\n - 10 0x4f9ec36 in wm_handlers_do_intern /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:2757\n - 11 0x4f9fc78 in wm_handlers_do /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:2865\n - 12 0x4fa537f in wm_event_do_handlers /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:3288\n - 13 0x4f732cf in WM_main /home/dev/01-data/01-git/blender-git/blender/source/blender/windowmanager/intern/wm.c:635\n - 14 0x34f58fa in main /home/dev/01-data/01-git/blender-git/blender/source/creator/creator.c:522\n - 15 0x7ffff6e4b0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x270b2)\n\nSUMMARY: AddressSanitizer: heap-use-after-free /home/dev/01-data/01-git/blender-git/blender/source/blender/gpu/opengl/gl_batch.hh:112 in blender::gpu::GLBatch::verts_(int) const\nShadow bytes around the buggy address:\n 0x0c2a80061d30: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061d40: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061d50: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061d60: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fa\n 0x0c2a80061d70: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa\n# >0x0c2a80061d80: fd[fd]fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061d90: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061da0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061db0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c2a80061dc0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa\n 0x0c2a80061dd0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\nShadow byte legend (one shadow byte represents 8 application bytes):\n```\nAddressable: 00\nPartially addressable: 01 02 03 04 05 06 07 \nHeap left redzone: fa\nFreed heap region: fd\nStack left redzone: f1\nStack mid redzone: f2\nStack right redzone: f3\nStack after return: f5\nStack use after scope: f8\nGlobal redzone: f9\nGlobal init order: f6\nPoisoned by user: f7\nContainer overflow: fc\nArray cookie: ac\nIntra object redzone: bb\nASan internal: fe\nLeft alloca redzone: ca\nRight alloca redzone: cb\nShadow gap: cc\n```\n", "Particle systems render wrong on frame 0\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 2.93.5 Release Candidate\n\nA particle system renders moving particles in the wrong place on frame 0, if particles are spawned (=Frame Start) before frame 0.\n\n1. Create a particle system\n2. Set Frame Start to something smaller than 0 (with bigger negative numbers the effect is more obvious, so f.e. Frame Start = -20) (the effect is even more obvious if \"End\" is set to 0 or close to 0)\n3. Set particle object to render (Render -> Render As -> Object; Select Instance Object)\n4. Setting timeline start to 0 and rendering the frames, the wrong particle placement in frame 0 is clearly observable in relation to the next few frames. \n\nMoving the whole project/particle system timing to later frames manually, fixes this location jump. This shows, the problem is only the frame being 0.\n\n[particle_system_bug.blend](particle_system_bug.blend)\n", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Crash when deleting one rigidbody from constraint during simulation\nOperating system: Windows 11\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU\n\nBroken: master branch\nWorked: Don't know if it ever worked\n\nDeleting one rigidbody that is part of a constraint during simulation causes a crash.\n\n**Steps to Reproduce**\n- Add two cubes to a scene. Add a rigidbody to both cubes.\n- Add an empty object to the scene. Add a fixed rigidbody constraint to the cube. Set the cubes as the first and second object in the constraint.\n- Play the simulation for a few frames and then pause the simulation.\n- Delete either cube, then continue playing the simulation.\n- Crash", "Kill Particles does not make them really dead\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.47\n\nBroken: version: 2.93.6\nWorked: -\n\nIf you give a mesh a physics property of collision and check the \"Kill Particles\" Checkbox, they are not becoming dead, they just stop moving. If you add a Particle Instance to another object and only show dead, the dead objects show correctly in viewport but not in render (eevee/cycles).\n\nI created a plane, put it about 10m higher and turned it 180° on x axis. Then i added a particle system to that plane with default settings. I added another plane and scaled it up a few times and applied all transforms. I then added a collision physics property with \"Kill Particles\" enabled to that plan. Now i added an Icosphere, scaled it down, applied all transforms and added the Particle Instance modifier. As Object i chose the first created plane and only left show dead active. After baking the Particle system, you see in viewport the Icosphere as soon as the particles hit the Plane with the collison physics property. But if you then want to render the image they are gone.\n\nWith the attached .blend file you just have to click bake in the particle system properties of the ParticleEmitter object and you should see the dead ParticleInstance object at Frame 40. But if you render the image then they are gone.[ParticlesNotDead.blend](ParticlesNotDead.blend)", "Particle System switching between Emitter/Hair \"Viewport Display\" sticks to Hair-Viewport-Display-Settings\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.92.0 Alpha\n\nIn the particle system, when you change the particle type form Emitter to Hair and then back to Emitter , the \"Viewport Display\" panel does not change back accordingly - it remains as if the Hair-type emitter is still present.\n\nit would be just as easy to create a new particle system with the right settings, to remedy this situation\n\n- create a particle system on a object\n- Observe the properties of the \"Viewport Display\" panel\n- change the particle type form Emitter to Hair (properties have changed accordingly)\n- change it back to emitter\nViewport Display won't show Options Particles have, instead it will show Hair Options like \"Strand Steps\"\n\n", "Particle collisions aren't simulated before frame zero for negative frame simulations\nBroken: 2.79, 2.81\nWorked: (optional)\n\nMaking particles simulation start from negative frames don't allow for those particles to have collisions with meshes. Unless I set the initial frame to be negative.\n\n\nOpen the file and look at the already baked simulation. The first particles went through while the later ones collide perfectly. You can also rebake the cache and see the same result\n\n[ParticleBug.blend](ParticleBug.blend)\n\n**Known workaround**\nChange initial frame to match the first simulated frame before baking the simulation.\n\n", "bake hair dynamics issues\nOperating system: wind 10\nGraphics card: gtx 1070\n\nBroken: blender-2.80-79f67acccb36-win64\nWorked: idk\n\nunable to bake hair dynamics. i click the button but nothing happens. in particle edit mode particle hair still looks as if it was not simulated.\n\n", "Crash after Disk Cache Particle bake while set to render as path\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.0.1\nsame results on daily 3.1 build\n\nBlender crashes after finishing baking the particles to disk.\nThen proceeds to delete the baked files upon reopening the file. \n(don't know if the deletion is intended behaviour or not, but it stops you from being able to salvage the bake unless you rename / duplicate the cache folder beforehand)\n\nSeems to only happen with disk cache bake and not baking to memory. \nChanging the 'render as' setting to anything but 'path' seems to stop the crashing.\n\nOpen file, bake the particle system\n[Particle Bake to DiskCache Crash.blend](Particle_Bake_to_DiskCache_Crash.blend)", "Regression: Some particles have darker faces\nOperating system: Windows 10 1909\nGraphics card: GTX 1070 512.15\n\nBroken: 3.1 (also 3.1.1 RC)\nWorked: 3.0.1\n\nSome object instances, in a hair particle system, have darker faces\n\n3.0.1\n\n\n3.1 and 3.1.1 RC\n\n\nPress F12 in this test file:\n[Dark Particles.blend](Dark_Particles.blend)", "Blender crashes if simplify is turned on while baking multires\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 XT ATI Technologies Inc. 4.5.14742 Core Profile Context 21.7.2 27.20.22021.1002\n\nBroken: version: 3.0.0 Alpha\nWorked: --\n\n\nBlender crashes if simplify is turned on while baking multires data\n\n\n- Grab any object and add multiresolution to if, subdivide at least once.\n- Activate simplify, and turn the max subdivisions on viewport to 0\n- Create a material and try to bake the multiresolution. Blender will crash instead of giving an error message or just baking normally \n\nTest File:\n[multires_crash.blend](multires_crash.blend) ", "Regression: Rendering hair instances size in animation bug with animated emitter object scale\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.47\n\nBroken: version: 3.1.2\nBroken: 2.93\nWorked: 2.92\n\nRegression from 50782df425.\n\nIncorrect rendered instanced object scale if rendering hair animation with animated object.\n\n1 Create Cube and object for instancing\n2 Set cube scale animation for some frame range\n3 Apply patricle system to cube (hair / rendered instances as object or collection)\n4 Try to render single frame and render animation, compare difference\n\n[hair_render.mp4](hair_render.mp4)\n[test_hair.blend](test_hair.blend)\n\nWill set to high prio to raise awareness this is a regression, but this can probably be lowered again since also involves the EOL particle system.\nNote: if you really need to render this, consider using Cycles (where this still works when the `Persistent Data` option is OFF." ]

[ "Blender crashes when the particle system is turned off in particle editing mode\nOperating system: Linux-4.15.0-47-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\nBroken: version: 2.80 (sub 58)\n\nBlender crashes when the particle system is turned off in particle editing mode. The specific configuration of particle systems cause this error, I tried to create a simpler example, but without success.\n\n[hairtest.blend](hairtest.blend)\n1. Open file.\n2. Switch to particle edit mode\n3. Disable first particle system in modifier stack - Blender shuts down immediately \n\n" ]

[ "Brushes: Smooth algorith is not consistent.\nCurrent smooth algotithem overwrites input data that can then be used during the same iteration. Meaning that the smooth operation is not consistent.\nDuring usage this is not that noticeable as smoothing is done per brush evaluation step. But it could be improved.\n\nThe memory could be allocated per thread and not per work item to reduce the overhead introduced.", "GPencil: Average tool in Vertex Paint mode not working as expected\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.6.0 Beta\n\nThe tool paints colors that exist in points far from the brush's area of influence and sometimes paints unexisting colors.\n\n1. Draw simple strokes with the Solid Stroke material.\n2. In VP mode paint half of it orange and half blue.\n3. try the the Average tool in the half blue area then in the other half.\n4. Now do Paint > Reset Vertex Color.\n5. Try the Average tool again.\n\n<video src=\"attachment\" title=\"Vertex Paint Bug-1.mp4\" controls></video>\n\n", "Lag Adjusting Value Sliders & Navigating Nested Menus with a Wacom Tablet\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics Intel 4.5.0 - Build 27.20.100.8681\n\nBroken: version: 3.0.0 Alpha\nWorked: 2.93.2\n\nTrying to change the value of shapekeys or shader sliders with my stylus doesn't get detected the first few seconds, then picks up, but in a non responsive way. Also while using my stylus, if I navigate to a nested menu like say \"Object > Transform\" There's a Lag when moving to another option in the \"Object\" Menu.\n\nThe steps above should apply. I did notice though, that if I change my Tablet API from 'Automatic' to 'wintab' in the the input preferences, it helps, but its still not as responsive as it is in 2.93\n\nAll these issues though, were non existent, if I tried using my laptop's trackpad. Also, my tablet is a Wacom Intuous Touch 5 with a stylus that supports up to 2048 levels if pressure sensitivity and has tilt support, just incase it could be any of these causing the issue. \n\nThanks.\n\n", "GPU Subdivision not optimized in Weight Paint mode\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1650 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.95\n\nBroken: version: 3.4.0 Alpha\n\nI'm not sure if it's unimplemented/unplanned or just overlooked, so I'm here to report in case.\nWhen you have a mesh using an Armature modifier and Subdivision modifier, the Subdivision will affect performance while posing the armature but enabling GPU Subdivision will noticeably improve performance during posing and animation. \n\nHowever, if you select the armature and the mesh and go into weight paint mode, allowing you to pose the rig, the GPU Subdivision will not boost performance with the Subdivision modifier. Weight Paint Pose is slower regardless as to the Subdivision modifier, just GPU Subdivision seems to be disabled here. \n\n- Open attached .blend file\n- Make sure `GPU Subdivision` is enabled in Preferences\n- in Weight Paint Mode, pose the rig (move any bone)\n- Go to object mode, select the rig and go to Pose mode\n- Compare performance when posing the rig\n\n[#101485.blend](T101485.blend)\n", "GPencil: Randomize brush's 'Position' attribute doesn't work when camera is too close to Grease Pencil object\nOperating system: Win10 Home 64\nGraphics card: rtx3070\n\nBlender Version\nBroken: 2.93, 3.1.0\nWorked: N/A\n\nExact steps to reproduce the error:\n\n- Open file\n- try to use Randomize brush\n- Strength and Thickness work, but not Position \n- Switch to the other camera that's farther away from the GP object, and randomizing the position works fine.\n\n[randomize_not_working_close_NEW.blend](randomize_not_working_close_NEW.blend)\n\n\nwatch?v=4pTrEm5HqVU", "Mask brush stops working in tiling sculpting mode after changing its size through \"F\" hotkey\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 2.93, 3.3.0\nWorked: -\n\nWhen an object has the Multires modifier, mask brush stops working in tiling sculpting mode after. (See video below)\n[Desktop 2022.09.17 - 21.18.10.07.mp4](Desktop_2022.09.17_-_21.18.10.07.mp4)\n\n- Open attached file or\n - Add a grid object\n - Add a multires modifier\n - Subdivide Simple a few times\n - Select mask brush in sculting mode\n - set tiling on x and y axis to 1 or 2m \n- try paint mask (sometimes the tiling doesn't work at this stage, you need to paint again and undo)\n- change brush size through F hotkey\n- try paint mask (tiling stops working)\n- press undo\n- try paint mask (tiling works again)\n\n[tiling_sculpt_bug.blend](tiling_sculpt_bug.blend)\n\n", "Loading a brush alpha/texture via file browser on dense meshes takes very long.\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.0\nWorked: never i think\nLoading a brush alpha/texture on dense meshes takes very long. Especially when you do a brush before swapping out the brush texture again. If you don't do a brush stroke after loading a texture and directly load another one it will be done quicker.\n\nMy machine: AMD 5900x, 32gb ddr4 ram, RTX 3060 12gb\n\n[2022-09-20 14-13-15.mp4](2022-09-20_14-13-15.mp4)\n1. Add a standard uv sphere and apply 7 subdivisions. You will end up with over 8 million faces. \n2. Go to sculpt mode and add an image texture to the standard brush. -> rather quick\n3. Do a stroke and load another texture/alpha. -> long loading time (several seconds) and this will go higher and higher with more polygons\n4. Now don't do a brush stroke and load another texture/alpha. -> quick loading time\n\nHere are the textures i've used in the video to test. [Organic.7z](Organic.7z)\n\n\n", "Regression: lag from camera view when using mouse scroll\nOperating system: Windows 10\nGraphics card: Nvidia Titan XP\n\nBroken: 3.3.1-3.5\nWorked: 3.2\nCaused by befe38fe1d\n\nMassive lag in camera view mode with objects about 20 mil triangles when using mouse scroll with view locked to camera even in Solid preview mode. When using Ctrl Alt mouse move seems no lag, scroll works as usual. Only scrolling is not working properly, rotation works nomally.\nIn this video: Blender 3.3.1, two separate cubes subdivided to level 10, first trying to scroll with mouse wheel - it lags, then using Ctrl Alt Mouse - no lag, then again with mousewheel - lag. No matter if GPU subdivision is on or off.\n\n- Open attached .blend file\n- first trying to scroll with mouse wheel - it lags\n- using Ctrl Alt Mouse - no lag\n[Blender Mouse scroll issue.blend](Blender_Mouse_scroll_issue.blend) \n[Lag when in camera view Blender 332.mp4](Lag_when_in_camera_view_Blender_332.mp4)", "Context Menus for Paint Modes\nThis is a follow on from a discussion on IRC with @campbellbarton.\n\nRecently, contextual menus menus were added to the paint modes. However, inside the paint modes we hardly have any useful operators to add here.\n\nInstead, we could add common brush controls here, as many painting apps do.\n\nTo start, we could just add a simple list of sliders:\n\n\n\nLater on, when the brush picker is more fleshed out, we could add it here too, like so:\n\n", "Dynamic paint sub-steps don't work when the brush position is set in geometry nodes\n\nI believe this **may** be the same issue as T40597\n\nIn order to have a brush follow a potentially very large collection of curves in sequence, I'm using geometry nodes to set the position, which avoids the pain of having to join the curve paths or painstakingly tweak the offsets and influence keyframes for each and every one.\n\nThis works, but with sub-steps enabled it seems that the nodes are still only being updated once/frame. Or, if they are being updated, then the `seconds` input from the `Scene Time` node is being derived from the `frames` value which is clamped to an integer.\n\n\n[text_path_test.blend](text_path_test.blend)\n", "Modifier are still evaluated even when not required.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 3.0.0\nWorked: never\n\nModifiers are evaluated, even when blank, for example a lattice modifier without any lattice, would still make depth graph calculation or something like that, that make each action (select, move, rotate, etc.) verry slower.\n\nCreate/ import a large object (60M poly mesh for example).\nAdd lattice modifier, let all option blank. \nDisable viewport visibility of this modifier.\nTry to move/rotate/scale the object in the viewport, the performance is really low.\n\nIn order to transform your mesh, you need to remove all modifiers, and add then back once transform is convenient.\n", "GPencil: Canvas Overlay becomes inconsistent when both Object and Layer are transformed\nOperating system: Windows\nGraphics card:\n\nBroken: 3.4.1\nWorked: never?\n\nModifying Grease Pencil Layer transforms affects the Canvas overlay, but when both the Object and Layer have transforms the overlay does not describe the drawing plane actually being used by tools.\n\n- Open attached .blend\n- Draw new strokes into the Lines layer\n- Rotate the viewport to view the stroke placement.\n\n", "Crash with animated displacement modifier in Texture paint mode while playing\nAMD Ryzen 5 2600, 16GB ram\nWindows 7 64bit\nNvidia GTX 1060 6gb (driver version 416.34)\n\n2.79b\n\nI intend to paint a heightmap interactively. So i keyframed a displacement modifier (2 frames) on a subdivided plane and assigned a texture.\nPlaying the timeline should now refresh the modifier every frame. \nPainting in the UV-editor this way works fine, but switching to texture paint mode and painting on the mesh directly crashes blender after about 0.5seconds.\n[HeightmapPaint.blend](HeightmapPaint.blend)\nI attached an example file.\nHit play and in the uv editor paint on the noise texture. This works okay.\nNow switch to Texture painting in the 3d viewport. The moment a stroke is applied to the mesh it should crash (it does consistently for me). ", "Camera projection mapping for brush texture\n**Status:** Task design and engineer plan review\n\n---\n\n**Team**\n**Code Review:** @Jeroen-Bakker @ideasman42\n**Project leader:** @ivpe \n**Project members:** @ssh4 \n\n\n**Preamble**\n\nGoal of this project are add code support and tools for Camera Projection Painting to Blender.\nSuch workflows used in 3D CG from early days for matte painting, 3D and video integration, etc.\nIn all workflows most logical UX is projection preview as it seen from camera and projection preview overlay on 3D model or scene that visible from any point.\n\nWhen matte painting is one of oldest workflows. Last years we can see huge progress in 3D scanning/Photogrammetry methods, that now one of main parts of 3D assets or scenes creation process. And in same time a lot of artists use photo modeling workflows with photo or video data as a source for textures.\nAnd again, most logical and fast way to work with textures for that models is using 3D paint feature directly in 3D app, instead of use external 3D painters or 2D editors to work with UVed textures.\n\n-----------------\n\n**Who can use it**\n\nThis project improvements in used UX from 3D assets creators, Photogrammetrists (3D scanners) and VFX/Matte painters point of view.\nFor that we made a list of current limitations and propose improvements for wide type of 3D CG users.\n\n* 3D assets creators:\n\nProbably one of biggest group. Standard photo modeling workflow use photo or images as a backgrounds. Final model UVed in human readable form and textures are making with standard for 2D editors or texture painter's workflows with using layers, masks, etc.\n\nPaint tool that can allow project image or photo from specific camera view can speedup time for mock-ups iterations in works with clients and art/creative directors. Due to real-time preview of results in 3D. In same moment using real photo, especially in architectural models limiting possible source of images. Due to lens distortions in all photo, required additional attention to used only high-quality professional photo equipment. Or applying undistortion to all sources before using them for texturing. And when using video in that moment required lot of additional disk space to store that temporary data.\n\nIf 3D assets are low-polygon than this also required creation of “proxy” objects with higher polygons count, because current Blender UV Projection give visible distortions on textures if used low-poly mesh or mesh with n-gons.\n\n* 3D sculptors:\n\nAnother and also big group of 3D content creators. And if in early CG days that was mostly artists that working with organic shapes or characters. In current time a lot of artists doing hard-surface modeling in 3D sculpting editors.\nThat group of artists also need the way to project photo data to textures and using paint editor in same app, speedup sketching or work in progress iterations.\n\nAll models made in that workflows in time of work can have complicated and non-optimized UVs. And using external 2D editors or texture painters not usable until model are retopologized and UVed in human readable form.\nCurrent Blender UV projection have less issues with such meshes. But polygons count already have a big impact on CPU time required for work. Computing new UV Projection from new camera/view for 5-10 million poly meshes already far from real-time.\n\nThis user group also have demand on works with photo or video sources, and lens distortion in source data required to use of undistorted images.\n\n* Photogrammetry and 3D scanning:\n\nLast years one of most growing groups of 3D content creators. That group work with a huge poly count meshes. And it more narrows to 3D sculpting by used workflows in post-processing scan data. With one but important difference, 3D scans are often monolith or split to chunks 3D meshes. And in the middle of process this meshes have completely non optimized UV's with thousands UV islands made by UV parametrization algorithms optimized for speed and works with thousands or even hundreds of millions polygons meshes.\n\nPhotogrammetry with all its power have a limitations in used algorithms due to a huge amount of data used in time of creation of 3D meshes and their texturing. And photogrammetry user not always have way to easily control process of choosing best source for texture in specific part of model. And result texture can use source data from image (camera) with worse depth of field or changed environment lights.\n\nLot of clients of 3D scan studios prefer to work with raw meshes and original textures to be sure that all possible details from 3D scan was used in most efficient way for their project.\nBut photogrammetrist still need fast way to quickly fix raw mesh and raw texture from specific image (specific camera) with better details, or correct light, etc.\n\nAs with 3D sculptors, mesh polygon count is more than enough to avoid UV projection deformations on textures. But used polygon count around 10-20 million of polygons in raw mesh is not a rare use case for high-quality high-resolution 3D scan.\n\nOne of biggest differences for other groups, is an image (camera count) used for 3D model and texture creation. 50-60 images/cameras are average small photogrammetry scan.\nBig or precise scans can easily have 1000 or even 10000 images with 8K or bigger resolution.\nGood point that using all these images/cameras for final texture is not a mandatory. But workflow that allow quickly switching between images/cameras are big time saver.\n\nAnd as soon as source for 3D model and texture are real photo, that always have lens distortions. But if other groups probably can work with some small distortions. Photogrammetry due to it high precision required using mathematically correct lens distortion models. That allow precise copy of source texture from scanned objects with sub millimeter resolution.\n\nUsed image count and size of this images required a lot of disc space for any additional temporary files like undistorted images.\nSo support for using lens undistortion in projection painting “must have” feature.\n\n* Matte painting/VFX:\nThat group can combine workflows and issues from all groups mentioned above. But in addition, this group often work with video sources. That mean this can be hundreds or thousands of frames that also can have real lens distortions or even rolling shutter issue (visible as a skew image deformation).\n3D models can be a combination of low-polygon models or blocks or n-gons for quickly build different distance planes on scene.\nAnd all of this in production or sketching time required quick integration with projection mapping painting.\n\n----------------\n\n**Current Blender workflow limitations and issues**\n\nSummarizing the above limitations of Blender:\n\nBlender currently has a way to do camera projection painting, with \"Clone\" brush and \"UV Project\" modifier. This method is enough labor intensive on the part of the user experience, requires to choose a lot of options even just to change camera from which projection act, which negatively affects the speed of the workflow. The main disadvantage of this method is the requirement for the density of the geometry. Low poly models (for example, after retopology) will have unnecessary distortions when drawing. Also, an extra step in this process is the export undistorted images/movie clips, because datasets can be large enough to create a lot of unnecessary files. It is also desirable to have control over brightness/contrast, ect, this method do not allow this.\n\n-----------------\n\n**User workflow expectations**\n\nThe ultimate goal of this project is to build a workflow that eliminates these shortcomings:\n\n**Scene setup**\nAssumed that the user has a scene with mesh objects, created by himself or imported from third-party software. There are also a number of cameras and original images/movie clips.\n- Automatically attach images/movie clips to specific cameras, they have similar names in any software. It should also be possible to do this manually.\n- Distortion parameters can be obtained through Movie Tracking Solver, but it should be possible to set parameters separately.\n\n**Canvas texture correction/painting process**\nAssumed that user is in Texture Paint Mode.\n- Select a camera. It is assumed that when you select a camera, the corresponding image/movie clip will also be selected for painting.\n- Paint. It is assumed that texture projection will occur from the camera with distortion correction according to the distortion parameters.\n\nMore globally, a major refactoring of the Texture Paint module is needed (like this one [#73935 ](T73935)), all users expect smooth workflow with high-polygon objects and high-resolution images/movie clips.\nAlso, an important part is to give the user a visual representation in the viewport of exactly how the texture will project on the object and from which camera the projection is act (the most obvious is to project from the active camera of the scene).\n\n------------------\n\n**Suggestions**\n\nHere described only a suggestions, each of them can be an open question.\n\n- For both \"Movie Tracking\" and \"Texture Paint\" modules image/movie clip as well as distortion parameters should be part of a particular camera (That should be coordinated with Movie Tracking module). About the Brush Management project. The described workflows have a tight relationship with a specific scene and objects in it, hence from the point the user's view of the photogrammetrist or VFX artist, it is expected that the work done by him at the scene setup stage will be saved specifically for this scene and objects in it. Therefore, it is important to understand that the camera with which the projection is act, its distortion parameters should not have relations with a specific brush. Since the dataset of images/movie clips can be large and also associated with specific cameras of a specific scene, it is also cannot relate to the brush.\n- Automate the process of attaching images/movie clips to specific cameras. This is most often possible by the name matching of the camera and the image/movie clip file.\n- Add \"Camera\" brush texture projection mode. It should use data (image/movie clip) from the active scene camera for the brush texture and also to determine the current parameters of the distortion.\n- Refactoring the brush preview. At the moment it is not informative and has limited presentation of the end result (for example, for \"3D\" texture map mode it's not presented for many years). In the context of this project the preview should be an overlay that takes into account the geometry of the object and gives an actual representation at any view.\n\n**All this but more visually**\nPreviously, we (@ssh4, @ivpe) did it as an addon, it was more photogrammetry-oriented so UI/UX should be different in some cases. In this short video described basic principles of workflow with large dataset of raw (distorted) images, viewport preview. Non-commercial scene is used.\n[Camera Projection Mapping.mov](Camera_Projection_Mapping.mov)\n\n-------------------\n**Work plan**\n> This section will proceed only after final design is ready:\n\n- ...\n\n**Open questions**\nOpen questions according to the suggestions below. Everyone can have their own opinion, so we need a general solution to be approved by everyone. Therefore, I leave a link to the original post, as well as its summary. After design document update some suggestions are included into it and some are still questionable:\n\n- [Original post by]] @brecht [[ https:*developer.blender.org/T81662#1038426 | Original post by](https:*developer.blender.org/T81662#1037146) @Sergey Where to store the distortion parameters of images/movie clips as well as the relationship between a specific camera with a specific image as given the current Movie Tracking design and in the case of its modification to more unified?\n- [Original post by](T81662#1037214) @brecht In case if camera store image or movie clip as background image. Should we have option for automation switch brush texture image / movie enabled by default? Or it should be the only way? \n- [Original post by](T81662#1046815) @sebastian_k Camera background image is enough for visual preview or it should be shader based?\n- [Original post by](T81662#1046815) @sebastian_k VFX scenes has already set background movie clips on cameras. It can be used for any workflow as well, I think. In terms of current project - to store image or movie clip relative to specific camera. But if we have more than one background image on single camera? In this case background images/movies are blended. The oldest principle: \"You get what you see\" should work here too. So how can we handle this? In my opinion the simple way - remove list of background images and leave only one. As I remember in older releases (< 2.8x) we had this as a list to use from different view angles (from left, right, top, camera). So, I think it is more Blender's design questionable too.\n- [Original post by](T81662#1047507) @ssh4 should be camera intrinsics animated for support videos that captured using zoom lenses as well as using optical or sensor stabilization?\n\n- Refactoring of the Texture Paint mode. Is it planned and is it worth counting on help in its implementation?\n- ...\n\n**Relevant links**:\nA small research on YouTube about current limitations:\n[CGMatter: Blender 2.8 Panorama stitching projection painting (part 2)](s_x62z1T9tk?t=467)\n[CGMatter: 2D To 3D](nWCWtojeM14)\n[CGMatter: Realistic 3D? EASY](uWkKV0HxwFQ?t=919)\n[IanHubert: Wild Tricks for Greenscreen in Blender](RxD6H3ri8RI?t=131)\n\nPrevious implementation using UV-projection method in separate boost-module:\n[Camera projection painter addon](watch?v=lasdJIIAv70&t=13s)\nIndustry-compatible lens distortion models:\n[Brown-Conrady]], [[ https:*developer.blender.org/D7484 | Nuke ](https:*developer.blender.org/D9037)", "Texture paint: Bleedind around thin cuts do not fill texture\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.95\n\nBroken: version: 3.6.0 Alpha\nWorked: never?\n\n\n1. Wider distance allow to overlap color\n2. closer distance between faces do not\n3. closer distance between islands allow to overflow to neighbour island…\n\n\n[bleeding test.blend](attachment)\nopen file, try to paint as seen in the [video 2023-03-24_05-09-32.mp4](attachment)\n\n" ]

[ "Stroke lag in texture paint with Geforce GTX 1080 ti\nOperating system: Windows 64 bits\nGraphics card: GTX 1080 ti\n\nBroken:\n2.79 and 2.8\n\n**\nWhen painting in texture paint I´m having lag that gives me straight lines instead soft curves.\nThis is happening to me and a friend, and both have a 1080 ti.\nI tested the same problem in a pc with a Geforce 820m and in other pc with a RTX 2080 and worked perfectly, so appareantly the problem is only with the 1080 ti **\n\n**\n- Select the default cube and select the texture paint workspace\n- Create a new Image and select this image in the texture slots (single image mode)\n- With the brush draw some curves in the cube and you will notice **" ]

[ "Rendering object in Cycles X preview hard crashes whole system\nOperating system: Windows 10 x64\nGraphics card: nVidia RTX 8000\n\nBroken: blender-3.5.0-alpha+universal-scene-description.1f9e90bb1cf3-windows.amd64-release\n\nRendering of scene in Cycles preview hard crashes whole system. System screen goes to black and the fans go to max.\n\nLoad blend file at link: brutes_set1C_trimmed.blend?dl=0\n\nSet screen to Render preview and navigate around.\n\n", "Running multiple instances of Cycles HIP ends in segfault\nOperating system: Windows 10\nGraphics card: 2x Radeon W6800 32GB\n\nBroken: 3.0.0, 3788003cdaeb, master, 2021-11-30\nWorked: Never\n\nWhen I have 2 instances of Cycles rendering using HIP at the same time on a particular production scene, one instance will crash with the following error: \n\n\n```\n...\nFra:441 Mem:8965.99M (Peak 8965.99M) | Time:00:21.57 | Mem:1469.81M, Peak:1469.81M | Scene, View Layer | Updating Images | Loading Scene2_Garage_Furniture_Roughness.png\nFra:441 Mem:8965.99M (Peak 8965.99M) | Time:00:21.70 | Mem:1533.81M, Peak:1533.81M | Scene, View Layer | Updating Images | Loading Metal02_nrm.jpg\nFra:441 Mem:8965.99M (Peak 8965.99M) | Time:00:21.73 | Mem:1569.81M, Peak:1569.81M | Scene, View Layer | Updating Images | Loading Scene2_Garage_Objects_normal.png\nFra:441 Mem:9030.99M (Peak 9030.99M) | Time:00:21.73 | Mem:1570.69M, Peak:1570.69M | Scene, View Layer | Updating Images | Loading Tires_rough.png\nMalloc returns null: len=37748736 in Cycles Aligned Alloc, total 1149156740\nError : EXCEPTION_ACCESS_VIOLATION\nAddress : 0x00007FFED98C853D\nModule : amdhip64.dll\nThread : 00000734\n\n```\nI was unable to reproduce it in a simple scene or any commonly available demo scenes such as classroom or Barbarshop. And unfortunately I cannot share the particular scene that's causing the issue. But the scale of the scene that's crashing should easily fit onto the GPU vram three times.\n\nI've attached the crash report. If more info is needed I will try to create a dummy scene that show the same behavior.\n\n[Anim-A-Garage.crash.txt](Anim-A-Garage.crash.txt)", "2.9.1 Close without error message\nOperating system: Windows 10, 10gb ram, ssd\nGraphics card: Nvidia GTX 650 1gb vram\n\nBroken: (2.83 LTE and 2.90 current available version)\nWorked: (2.82)\n\nWhen rendering with GPU AND CPU, the app crashes, works perfectly fine on CPU only. On GPU only too but slower than older versions.\n\n\n```\nblender.exe :0x00007FF7A5B76440 ccl::kernel_cpu_sse41_filter_finalize\nblender.exe :0x00007FF7A57A9220 ccl::CPUDevice::denoising_solve\nblender.exe :0x00007FF7A57DA470 ccl::DenoisingTask::reconstruct\nblender.exe :0x00007FF7A57D8ED0 ccl::DenoisingTask::run_denoising\nblender.exe :0x00007FF7A57A7240 ccl::CPUDevice::denoise_nlm\nblender.exe :0x00007FF7A57AB590 ccl::CPUDevice::thread_render\nblender.exe :0x00007FF7A57ABA30 ccl::CPUDevice::thread_run\nblender.exe :0x00007FF7A57A60B0 ?_Do_call@?$_Func_impl_no_alloc@V<lambda_d884e08a7877f415f05ce8ffda8f97b4>@@X$$V@std@@EEAAXXZ\nblender.exe :0x00007FF7A7D5F520 tbb::internal::function_task<std::function<void __cdecl(void)> >::execute\ntbb.dll :0x00007FFBF5F037A0 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFBF5F037A0 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFBF5EF51D0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FFBF5EFA490 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FFBF5F019C0 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FFBF5F019C0 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FFC08611FA0 o_exp\nKERNEL32.DLL :0x00007FFC0B647960 BaseThreadInitThunk\nntdll.dll :0x00007FFC0B7AA250 RtlUserThreadStart\n```\n\nBased on the default startup or an attached .blend file (as simple as possible).", "Blender Crashing When Painting Stroke into a \"Moving Animated Character\" (heap-use-after-free)\nBroken: 4.0 alpha\nBroken: version: 2.79 - 2.82 (sub 7)\nWorked: Never\n\nI have this experiment of making LIVE talking head, via OSC. And I noticed that if I tried painting a stroke on a talking moving animated head, it will crash. Even with timeline not running.\n\nI got a video (Might need iPhone X with MOM addon to test this....) :\njhHdG6ZXFDU\n\n- Open attached file\n- Run animation\n- Paint over the animated object\n[paint_animated_shapekey_bug.blend](paint_animated_shapekey_bug.blend)\n\n\n", "Crash while using several rigid bodies and some force fields.\nOperating system: Darwin-18.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon R9 M370X OpenGL Engine ATI Technologies Inc. 4.1 ATI-2.11.20\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional)\n\n\nThe scene is pretty simple, a house with some elements used as rigid bodies and some force field, vortex and wind force are applied.\n\nWhen I play the animation, it starts and around frame 140-200 it crashes.\n\nAttached blend file. The only thing to do is to play the animation and wait for the crash.\nAttached crash report.[Blender_2019-08-26-215345_MacBook-Pro-de-Kote.crash](Blender_2019-08-26-215345_MacBook-Pro-de-Kote.crash)\n\n[casaFalla.blend](casaFalla.blend)\n\n", "Crash when moving to Shading workspace\nOperating system: Darwin-16.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-10.17.5 355.10.05.45f01\n\n|Broken:|version: 2.83.0, commit date: 2020-06-03,|hash: `211b6c29f7`\n| -- | -- | -- |\n||version: 2.90, commit date: June 08,|hash: `9f7d84b656`\n|Worked:|version: 2.82|-\n\nWhen I open Blender, without changing anything in the default scene, if I move to the 'Shading' workspace my system freezes for a few seconds and crashes (it closes).\n\n- Default startup scene\n- click on 'Shading' workspace from the top bar\n", "Crash when baking Dynamic Paint Image Sequence multiple times in loop\nBroken: 2.90.0 Alpha\nWorked: blender-2.80-windows64\n\nBlender crashes when baking dynamic paint image sequence in for loop. Depending on wait time crash will have differens stack traces.\n\n[#74052.blend](T74052.blend)\n\n- Open file\n- Run script", "Crash when rendering animation\nOperating system: macOS-13.4.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Pro 1.2\n\nBroken: version: 3.6.0\n\n\nI was rendering a simple camera animation overnight, Blender crashed about 2/3rd the way through. Here's the crash log.\n", "Crash on render\n**System Information** MSI BRAVO 15\nOperating system: WINDOW 11\nGraphics card: RX 6600\n\n**Blender Version**3.6\n\nWhen attempting to render a scene in Blender, the software consistently crashes, causing frustration and hindrance to the rendering process. This issue is hindering users from efficiently creating their projects and requires immediate attention to maintain a stable and reliable rendering environment.\n\n[debug-log](https://projects.blender.orgattachment)\n[system info](https://projects.blender.orgattachment)", "Crash with animated displacement modifier in Texture paint mode while playing\nAMD Ryzen 5 2600, 16GB ram\nWindows 7 64bit\nNvidia GTX 1060 6gb (driver version 416.34)\n\n2.79b\n\nI intend to paint a heightmap interactively. So i keyframed a displacement modifier (2 frames) on a subdivided plane and assigned a texture.\nPlaying the timeline should now refresh the modifier every frame. \nPainting in the UV-editor this way works fine, but switching to texture paint mode and painting on the mesh directly crashes blender after about 0.5seconds.\n[HeightmapPaint.blend](HeightmapPaint.blend)\nI attached an example file.\nHit play and in the uv editor paint on the noise texture. This works okay.\nNow switch to Texture painting in the 3d viewport. The moment a stroke is applied to the mesh it should crash (it does consistently for me). ", "GPencil: Selecting textured brushes will crash Blender\nOperating system: Linux-5.8.0-7642-generic-x86_64-with-glibc2.32 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.39\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.92.0\n\nWhen appending the textured brushes from the [Blender Cloud](5ccfe64353b85e279cf72acd) into a file, Blender will crash when trying to preview the brushes in the tool settings.\n\n\n - Open Blender 2.93 and select the \"2D Animation\" template.\n - Go to *File*>*Append* and browse to the `pepeland_GP_brush_pack_V1.blend` file. Go into the *Brush* folder and append the brushes `GBrush_001` through `GBrush_010`.\n # Click the brush icon in the tool settings. Blender crashes.\n\n\n\n", "Sculpt: Paint tool cavity custom curve crash\nOperating system: Windows 10\nGraphics card: NVDIA RTX 2060 Super\n\nBroken: 3.4, 3.5\n\nSteps to reproduce the issue:\n - Switch to Sculpt Mode or open default sculpting template file\n - Choose the Paint Tool\n - Click on Brush dropdown\n - Enable Cavity by clicking on Advanced >> Cavity in the \"Brush\" dropdown\n # Enable Custom Curve under Advanced >> Cavity in the \"Brush\" dropdown\n\n\n\nResult:\nNo **Custom Curve** is shown as in other Sculpting Tools & after trying to paint with the **Paint** tool setup this way blender instantly crashes.", "Cycles HIP device crash on rendering on latest AMD Adrenaline Driver\nOperating system: Windows 11 Pro for Workstation\nGraphics card: AMD Radeon RX 6750XT (Driver: AMD Adrenaline 23.7.1, 23.7.2 23.8.1 23.9.1)\n\nBroken: 3.6.1 LTS (On latest driver)\nWorked: 3.6.1 LTS (On old driver: AMD Adrenaline 23.5.2)\n\nAfter synchronizing objects, Blender crash\nBut render the default cube scene Blender works well even using 23.7.1 driver\n\nBased on the default startup or an attached .blend file (as simple as possible).\nScene is from Blender official demo file: \nbarbershop_interior/\nCrash log file: \n\n", "Application templates crashes if the startup.blend of the application template contains scene named \"Asset\"\nOperating system: Windows-10-10.0.22621-SP0'\nGraphics card: NVIDIA GeForce RTX 2070 Super with Max-Q Design/PCIe/SSE2\n\nBroken: 3.5.1 2023-04-24 18:00 e1ccd9d4a1d3\nWorked: Never\n\nApplication templates that contain certain pattern in startup.blend lead to Blender to crash.\nApparently occurs when the scene is created with the name \"Asset\" and used as a application template.\n\n- Download attached [MCE.zip](attachment) file\n- Install the attached application template file (`Blender` menu > `Install Application Template...`).\n- Load it (choose the new \"MCE\" option in `File` > `New`).\n\nBlender crashes.\n\nAttached also the `blender.crash.txt` and another template.\n\n[MCE.zip](attachment)", "Smoke and Fire simulation bug\nOperating system: Mac osx 10.15.6\nGraphics card: AMD Radeon Pro 5500M 8 GB\nIntel UHD Graphics 630 1536 MB\n\nBroken: 2.90.0 (202-08-31)\n\n\nBlender crashes while trying to bake a fire + smoke simulation in 128 resolution divisions... Sometimes it works. Mostly it doesn't.\n\nAttaching the file.\n\n[Blender Bug copy.rtf](Blender_Bug_copy.rtf)\n\n[fire.blend](fire.blend)" ]

[ "Blender 2.91 multi scene and engine render crash\nOperating system: Linux-5.4.0-56-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.80.02\n\nBroken: version: 2.91.0\nWorked: 2.83.5\n\nBlender crashes every time when eevee scene rendering starts on Blender 2.91.\n\n1. Open provided blend file\n2. Press F12 to render\n[b291_crash_bug.blend](b291_crash_bug.blend)\n\n```\n# Blender 2.92.0, Commit date: 2020-12-22 20:58, Hash 8d3d4c884084\n\n# backtrace\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(BLI_system_backtrace+0x20) [0x862ae20]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0xea90aa]\n/lib/x86_64-linux-gnu/libc.so.6(+0x46210) [0x7f3e239e5210]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(GPU_texture_mipmap_mode+0) [0x721a7b0]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0xee5fe4]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0x1384133]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0x1384476]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0x13762b6]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(DRW_draw_render_loop_2d_ex+0x38d) [0x1377f5d]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0x1b7fa81]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(ED_region_do_draw+0x7e1) [0x17958a1]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(wm_draw_update+0x52e) [0x122e3ee]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(WM_main+0x30) [0x122c290]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender(main+0x31e) [0xde340e]\n/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf3) [0x7f3e239c60b3]\n/home/xard/Downloads/blender-2.92.0-8d3d4c884084-linux64/blender() [0xea5aa3]```\n\n\n" ]

[ "Render cause crash guide many time August 10, 2022\nOperating system: macOS-13.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.39\n\nBroken: version: 3.2.2\n\nRender crash too many\n\nFor Mac only. Do not use Windows please.\n\n\n\n\n\nSee blender 3.2.2 on Cycle-x picture\n\n\nCommand + F12 then render for a while few minutes cause crash many times.\n\n[Blender report a ticket August 10 2022.zip](Blender_report_a_ticket_August_10_2022.zip)", "Blender crashes after GPU spike (Specially when we have a character with Rigify)\nOperating system: Windows 10\nGraphics card: RTX 2070 Super\n\nBroken: 3.4.1\nWorked: 3.3\n\nEvery time I'm working in Blender, after 2 minutes if I start to move around on a rig made based on Rigify, or work on geometry nodes, Blender Crashes. It happens even faster if I'm working in eeve.\n\n- Open a blender file\n- Import a character using rigfy with Linked Library\n- Library override\n- Turn Eeve on\n- Start working\n- Eventually, blender will crash.\n\nI noticed that if I don't import a character with rigfy, it doesn't have the GPU spike\n[Christmas.crash.txt](Christmas.crash.txt)\n[KalyAndAkrasiel 3.5.crash.txt](KalyAndAkrasiel_3.5.crash.txt)", "Suspected asset library crash by `json_sax_dom_parser`\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.5.1\n\nI dont completely know why but blender crashes very often for me.\nIts doing something no idea what it is doing.\nSorry for being vague, but it seems to mostly happen when i have the viewport set to cycles render and i am trying to modify materials, in the shader editor.\n\n", "crash when rendering with optics\nOperating system: windows\nGraphics card: 3060\n\n**Error**\n\n```\n# Blender 3.4.1, Commit date: 2022-12-19 17:00, Hash 55485cb379f7\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF7B98A5710\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters[0] : 0x0000000000000001\n\tParameters[1] : 0x0000000000000000\n\n\nStack trace:\nblender.exe :0x00007FF7B98A56F0 blender::default_construct_n<blender::SimpleMapSlot<std::pair<int,int>,blender::meshintersect::ITT_\nblender.exe :0x00007FF7B98B6730 blender::Array<blender::SimpleMapSlot<std::pair<int,int>,blender::meshintersect::ITT_value>,1,blend\nblender.exe :0x00007FF7B98B5770 blender::Map<std::pair<int,int>,blender::meshintersect::ITT_value,0,blender::PythonProbingStrategy<\nblender.exe :0x00007FF7B98B7570 blender::meshintersect::trimesh_nary_intersect\nblender.exe :0x00007FF7B98BE4D0 blender::meshintersect::boolean_trimesh\nblender.exe :0x00007FF7B98BE240 blender::meshintersect::boolean_mesh\nblender.exe :0x00007FF7B94D43D0 blender::meshintersect::direct_mesh_boolean\nblender.exe :0x00007FF7B3ACF380 exact_boolean_mesh\nblender.exe :0x00007FF7B3ACE760 modifyMesh\nblender.exe :0x00007FF7B37C78F0 modifier_modify_mesh_and_geometry_set\nblender.exe :0x00007FF7B37C6190 mesh_calc_modifiers\nblender.exe :0x00007FF7B37C5F80 mesh_build_data\nblender.exe :0x00007FF7B37C8560 makeDerivedMesh\nblender.exe :0x00007FF7B3759B30 BKE_object_handle_data_update\nblender.exe :0x00007FF7B3759990 BKE_object_eval_uber_data\nblender.exe :0x00007FF7B39AA920 blender::deg::`anonymous namespace'::evaluate_node\nblender.exe :0x00007FF7B39AA6C0 blender::deg::`anonymous namespace'::deg_task_run_func\nblender.exe :0x00007FF7B92A2780 tbb::internal::function_task<Task>::execute\ntbb.dll :0x00007FFA475AF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFA475AF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FFA475A4FD0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FFA475AA120 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FFA475AD800 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FFA475AD800 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FFA60E41B20 configthreadlocale\nKERNEL32.DLL :0x00007FFA61A27600 BaseThreadInitThunk\nntdll.dll :0x00007FFA63722680 RtlUserThreadStart\n\n\nThreads:\nThread : 00004b34\nntdll.dll :0x00007FFA6376D6D0 NtDelayExecution\nblender.exe :0x00007FF7B9267160 BLI_thread_is_main\n```\n\nuse a lot of boolean modifiers(4-6) and high subdivision modifier level(3) and set it to display the changes only on the render and use optix\n\n", "Viewport animation render crashes when collapsing the viewport.\nBroken: version: 2.91.0 Beta (and 2.83, 2.80, ...)\nWorked: never\n\nViewport animation render crashes when collapsing the viewport.\n\n[0001-0485.mp4](0001-0485.mp4)\n[Viewport render crash.blend](Viewport_render_crash.blend)\n\n", "Crash when moving to Shading workspace\nOperating system: Darwin-16.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-10.17.5 355.10.05.45f01\n\n|Broken:|version: 2.83.0, commit date: 2020-06-03,|hash: `211b6c29f7`\n| -- | -- | -- |\n||version: 2.90, commit date: June 08,|hash: `9f7d84b656`\n|Worked:|version: 2.82|-\n\nWhen I open Blender, without changing anything in the default scene, if I move to the 'Shading' workspace my system freezes for a few seconds and crashes (it closes).\n\n- Default startup scene\n- click on 'Shading' workspace from the top bar\n", "Cycle gpu rendering crashes randomly\nOperating system: Win11\nGraphics card: Nvidia RTX 3070 Ti\n\nbroken: all 3.0 versions\n\nRenders keep crashing randomly when I’m using my gpu with cycles, I updated my drivers, my memory use is okay, I tried to not use denoiser, lower my samples, resolution etc but nothing works, sometimes it render 10 frames sometimes 100 but it can't render my whole animation without crashing (also for animations with low poly count and small textures tiles) \nDon't know how to read the crash log file.\n\n", "Cycles Rendering of project causes crash of program.\nOperating system: Windows-10-10.0.22621-SP0 64 Bits (Windows 11)\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 537.13\n\nBroken: version: 3.6.2\n\nTrying to render, pressing F12 of the file to render project it crashes. I'm using Blender 3.6.2\n\nTrying to render, pressing F12 of the file to render project it crashes. I'm using Blender 3.6.2\nNot sure if you need textures, can provide if needed.", "Blender Freezes and Crashes on AMD Laptops with White and Black boxes\nOperating system: Windows 10 & 11\nGraphics card: Radeon RX 6600M ATI Technologies Inc. 4.5.14800 Core Profile Context 22.4.2 30.0.15021.7000\n\nI've been using Blender with many versions. From 2.8, to 3.1 and the laptops have issues with all of them.\n\nBlender will freeze, white and black boxes will appear, and then Blender will crash and Windows will close it.\n\nI don't have exact steps, because it happens infrequently. It can happen as quickly as opening Blender, or take an hour of work.\n\nI am a teacher using HP Omen laptops with AMD CPU's and graphics cards. I have 21 laptops, and about 5-7 of them have issues with Blender. I have done a clean install of the graphics drivers, updated Blender to the most recent version, and tried everything I can think of. My personal laptop rarely ever crashes, so it's hard for me to replicate the issue.\n\nI have attached log files from crashes from multiple laptops. Half the laptops have been upgraded to Windows 11, and that doesn't seem to make it more or less likely to crash. My hope is these logs can pinpoint what's happening so I can fix it. I\n\nLog files created with `blender_debug_gpu_glitchworkaround.cmd`\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\nLog files created with `blender_debug_gpu.cmd`\n\n[blender_debug_output (1).txt](blender_debug_output__1_.txt)\n\n[blender_system_info (1).txt](blender_system_info__1_.txt)\n\nWarning and errors:\n```lines=10\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_multi_rect_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_multi_rect_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_text VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_text FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_uniform_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_uniform_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_INST_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_WIDGET_BASE_INST_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_rect_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_rect_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_flat_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_flat_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_LINE_DASHED_UNIFORM_COLOR_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): GPU_SHADER_2D_LINE_DASHED_UNIFORM_COLOR_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_3D_uniform_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_3D_uniform_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_desaturate_color VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): gpu_shader_2D_image_desaturate_color FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_mesh_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_hair_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_opaque_ptcloud_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_composite_studio VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_composite_studio FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_merge_infront VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_merge_infront FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_mesh_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_hair_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_none_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_none_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_single_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_single_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_tile_no_clip VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transp_studio_ptcloud_tex_tile_no_clip FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transparent_resolve VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_transparent_resolve FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_effect_outline VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_effect_outline FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_taa VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_taa FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_0 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_0 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_1 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_1 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_2 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): workbench_smaa_stage_2 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_antialiasing_263 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_antialiasing_263 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_414 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_414 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_449 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_449 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_401 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_sphere_401 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_439 GeomShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n | \n 6 | #define DFDY_SIGN 1.0\n | ^\n | (#310) Inputs that are integers should be qualified with the interpolation qualifier \"flat\"\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_wire_469 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_shape_wire_469 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_wire_538 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_wire_538 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_solid_559 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_degrees_of_freedom_solid_559 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_stick_518 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_stick_518 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_499 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_499 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_489 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_envelope_489 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_wire_579 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_armature_wire_579 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_background_276 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_background_276 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_uniform_color_1465 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_uniform_color_1465 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_grid_905 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_grid_905 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_885 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_885 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_groundline_924 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_groundline_924 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_wire_962 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_loose_point_983 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_loose_point_983 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_point_1004 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_extra_point_1004 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_facing_1022 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_facing_1022 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_grid_1055 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_grid_1055 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_line_1151 GeomShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_vert_1170 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_motion_path_vert_1170 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_1204 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_1204 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_pointcloud_1257 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_pointcloud_1257 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_gpencil_1230 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_prepass_gpencil_1230 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_detect_1272 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_outline_detect_1272 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_dot_1410 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_dot_1410 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_shape_1429 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_particle_shape_1429 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_wireframe_1589 VertShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nWARN (gpu.shader): OVERLAY_shader_wireframe_1589 FragShader: \n | \n | (#62) enable/warn/disable extension is not found. Extension \"GL_ARB_conservative_depth\" is not supported\n\nERROR (gpu.debug): : GL_INVALID_VALUE : generated before glGenTextures\n```\n\nStack trace:\n```lines=10\nblender.exe :0x00007FF6F22F7D90 bli_windows_system_backtrace_stack_thread\nblender.exe :0x00007FF6F22F7010 BLI_system_backtrace\nblender.exe :0x00007FF6F228C530 blender::gpu::debug::debug_callback\nblender.exe :0x00007FF6F228C3F0 blender::gpu::debug::check_gl_error\nblender.exe :0x00007FF6F220EE50 blender::gpu::GLTexture::GLTexture\nblender.exe :0x00007FF6F21F8420 blender::gpu::GLBackend::texture_alloc\nblender.exe :0x00007FF6F21F2E20 GPU_texture_create_2d\nblender.exe :0x00007FF6ED267400 DRW_texture_create_2d\nblender.exe :0x00007FF6ED267810 DRW_texture_ensure_fullscreen_2d\nblender.exe :0x00007FF6ED2A4F90 workbench_cache_finish\nblender.exe :0x00007FF6ED25E4D0 DRW_draw_render_loop_ex\nblender.exe :0x00007FF6ED25F5A0 DRW_draw_view\nblender.exe :0x00007FF6EDBC7820 view3d_main_region_draw\nblender.exe :0x00007FF6ED56D940 ED_region_do_draw\nblender.exe :0x00007FF6ED1660C0 wm_draw_window_offscreen\nblender.exe :0x00007FF6ED165F10 wm_draw_window\nblender.exe :0x00007FF6ED165A10 wm_draw_update\nblender.exe :0x00007FF6ED1488F0 ghost_event_proc\nblender.exe :0x00007FF6EDCB1E10 GHOST_CallbackEventConsumer::processEvent\nblender.exe :0x00007FF6EDCFD680 GHOST_EventManager::dispatchEvents\nblender.exe :0x00007FF6EDCFE2E0 GHOST_System::dispatchEvents\nblender.exe :0x00007FF6ED14A980 wm_window_process_events\nblender.exe :0x00007FF6ED140110 WM_main\nblender.exe :0x00007FF6ECDA03C0 main\nblender.exe :0x00007FF6F246FE74 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFF7DCC7020 BaseThreadInitThunk\nntdll.dll :0x00007FFF7E582630 RtlUserThreadStart\n```\n\n----\n\n[blender_system_info_1.txt](blender_system_info_1.txt)\n\n[blender_system_info_2.txt](blender_system_info_2.txt)\n\n[blender_system_info_3.txt](blender_system_info_3.txt)\n\n[blender_debug_output_2.txt](blender_debug_output_2.txt)\n\n[blender_debug_output_1.txt](blender_debug_output_1.txt)\n\n[blender_debug_output_3.txt](blender_debug_output_3.txt)", "Blender crashes: OpenVDB math::Transform::read throws exception. \nOperating system: Darwin-19.5.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.9.15\n\nBroken: version: 2.83.2\n\nBlender keeps crashing.\n\nI have a project were I created an animation of modified arrays. Everytime I play it it crashes\n\nCrash report: P1536\n\ncrashed thread: \n```\n\tThread 35 Crashed:\n0 libsystem_kernel.dylib \t0x00007fff726c733a __pthread_kill + 10\n1 libsystem_pthread.dylib \t0x00007fff72783e60 pthread_kill + 430\n2 libsystem_c.dylib \t0x00007fff7264e8a2 __abort + 139\n3 libsystem_c.dylib \t0x00007fff7264e817 abort + 135\n4 libc++abi.dylib \t0x00007fff6f8ae458 abort_message + 231\n5 libc++abi.dylib \t0x00007fff6f89f8bf demangling_terminate_handler() + 262\n6 libobjc.A.dylib \t0x00007fff713da5b1 _objc_terminate() + 104\n7 libc++abi.dylib \t0x00007fff6f8ad887 std::__terminate(void (*)()) + 8\n8 libc++abi.dylib \t0x00007fff6f8b01a2 __cxxabiv1::failed_throw(__cxxabiv1::__cxa_exception*) + 27\n9 libc++abi.dylib \t0x00007fff6f8b0169 __cxa_throw + 113\n10 org.blenderfoundation.blender \t0x0000000110291c61 openvdb::v7_0::math::Transform::read(std::__1::basic_istream<char, std::__1::char_traits<char> >&) + 4801\n11 org.blenderfoundation.blender \t0x000000011026082b openvdb::v7_0::io::Archive::readGrid(std::__1::shared_ptr<openvdb::v7_0::GridBase>, openvdb::v7_0::io::GridDescriptor const&, std::__1::basic_istream<char, std::__1::char_traits<char> >&) + 1787\n12 org.blenderfoundation.blender \t0x0000000110286589 openvdb::v7_0::io::File::Impl::unarchive(openvdb::v7_0::io::File const&, std::__1::shared_ptr<openvdb::v7_0::GridBase>&, openvdb::v7_0::io::GridDescriptor const&, openvdb::v7_0::io::File::Impl::NoBBox) + 57\n13 org.blenderfoundation.blender \t0x0000000110282c91 openvdb::v7_0::io::File::readGridByName(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, openvdb::v7_0::math::BBox<openvdb::v7_0::math::Vec3<double> > const&) + 465\n14 org.blenderfoundation.blender \t0x0000000110282ab6 openvdb::v7_0::io::File::readGrid(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&) + 54\n15 org.blenderfoundation.blender \t0x00000001128a2cc9 MANTA::updateGridFromVDB(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, float*, bool) + 361\n16 org.blenderfoundation.blender \t0x0000000112890985 MANTA::updateGridFromFile(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> >, float*, bool) + 869\n17 org.blenderfoundation.blender \t0x0000000112890cee MANTA::updateNoiseStructures(FluidModifierData*, int) + 558\n18 org.blenderfoundation.blender \t0x000000011287e2b5 manta_update_noise_structures + 21\n19 org.blenderfoundation.blender \t0x000000010df34cd0 BKE_fluid_modifier_processDomain + 3168\n20 org.blenderfoundation.blender \t0x000000010df316ce BKE_fluid_modifier_do + 142\n21 org.blenderfoundation.blender \t0x000000010e3ad4f6 applyModifier + 54\n22 org.blenderfoundation.blender \t0x000000010dea5b0a mesh_calc_modifiers + 2778\n23 org.blenderfoundation.blender \t0x000000010dea4d6b mesh_build_data + 123\n24 org.blenderfoundation.blender \t0x000000010dea4035 makeDerivedMesh + 133\n25 org.blenderfoundation.blender \t0x000000010dfe1281 BKE_object_handle_data_update + 1089\n26 org.blenderfoundation.blender \t0x000000010dfe1704 BKE_object_eval_uber_data + 52\n27 org.blenderfoundation.blender \t0x000000010e9f36ba DEG::(anonymous namespace)::deg_task_run_func(TaskPool*, void*, int) + 138\n28 org.blenderfoundation.blender \t0x00000001142cadbc task_scheduler_thread_run(void*) + 364\n29 libsystem_pthread.dylib \t0x00007fff72784109 _pthread_start + 148\n30 libsystem_pthread.dylib \t0x00007fff7277fb8b thread_start + 15\n```\n", "Blender 3.5 and 3.5.1 are crashing while using Metal as backend\nOperating system: macOS 12.6.3\nGraphics card: Radeon Pro 580 8GB\n\nBroken: 3.50, 3.51 (Metall Backend)\nWorked: 3.50, 3.51 (OpenGL Backend)\n\nBlender simply hang. I use eevee, if you switch to rendered viewport shading it crashed complete system. I use Metall as backend, if I choose OpenGL it works fine. \n\n \nI attached an .blend file for example.\n\n", "Particle Edit mode for non hair particles - point select mode crash\nOperating system: Linux-5.6.0-050600-lowlatency-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.64\n\nBroken: version: 2.90 (sub 3)\n\nBlender crashes when entering into the Particule Edit mode\n\n- Delete all objects from the default scene\n- Create a plane\n- Add a particle system\n- Set end frame to 50\n- Bake the particle system\n- Switch to the Particle Edit mode\n- Select Point select mode\n\n- > Crash\n\n[blender.crash.txt](blender.crash.txt)\n\n\n", "Crash guide during render complete August 29, 2022\nOperating system: macOS-13.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.44\n\nBroken: version: 3.3.0 Beta\n\nBlender 3.3 Beta crash guide during render August 29, 2022\n\n\nNot matter all version 3.0 to 3.4 keep me crash. Do not use Windows. Only for Macintosh.\n\n\n\n[Blender-2022-08-29-092340.ips](Blender-2022-08-29-092340.ips)\n\n[Armor wear 1.0.0.blend](Armor_wear_1.0.0.blend)\n\n\n\nT100328\n@Nurb2Kea\nRender 600 (complete render cause crash guide)\n", "Crash when deleting ViewLayer while rendering\nOperating system: Linux-5.4.0-66-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.92.0\n\nA crash will happen when deleting a ViewLayer while rendering stills or animations, for any render_display_type\n\n- In the default scene, add a new ViewLayer\n- Set the renderer to Cycles so that you can see the render progress. In Eevee you could set up a little animation to get the same problem.\n- Start a normal render\n- While Blender is still rendering, delete the new ViewLayer that you just added and that was being in use when the render process was called\n- A crash happens with a Segfault, even after some time.\n\nThis problem looks to me as just an unhandled context change, because I noticed that if you delete the current scene or the current camera used while rendering animations, those context changes are appropriately handled and the render process is effectively stopped without a crash.\n", "Crash with GPU Render\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1660 Ti with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.02\n\nBroken: version: 3.4.0\nWorked: unknown\n\nOn cycles, using the GPU renderer blender crashes. This only happens when I am using Optix with GPU and CPU. If I disable my CPU, Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz, then the render works, using only CPU rendering also works.\n\nMy Memory is not fully used, both GPU and ram, and I have the newest GPU driver, GameReady version 528.02\nHope the blend file works without the video files.\nThanks for helping\n[obj_tracking_test_2.crash.txt](obj_tracking_test_2.crash.txt)\n\n[obj_tracking_test_2.blend](obj_tracking_test_2.blend)\n\n" ]

[ "Crashing when compositing 2 scenes\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.79 **&**GTX 970Broken: version: 2.91.0\nWorked: 2.90\n\nWhen compositing 2 scenes together Blender crashes during the animation render. This seems to affect both Eevee and Cycles, although Eevee seems to crash sooner. \n\n1. Open provided file of mobius strip\n2. Render out animation with compositing (crashes)\n3. Render out animation with \"use node,\" for compositing turned off and it will finish. \n\n[mobius.blend](mobius.blend)\n\nEDIT: Here are the debug and crash logs created using factory_startup\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[mobius.crash.txt](mobius.crash.txt)\n\n```lines\nStack trace:\nblender.exe :0x00007FF6D6BD2A80 GPU_texture_mipmap_mode\nblender.exe :0x00007FF6D2EE3F20 image_get_gpu_texture\nblender.exe :0x00007FF6D310DCC0 image_cache_image\nblender.exe :0x00007FF6D310DA20 IMAGE_cache_init\nblender.exe :0x00007FF6D30D4220 drw_engines_cache_init\nblender.exe :0x00007FF6D30D0980 DRW_draw_render_loop_2d_ex\nblender.exe :0x00007FF6D3830520 image_main_region_draw\nblender.exe :0x00007FF6D34763F0 ED_region_do_draw\nblender.exe :0x00007FF6D3019E70 wm_draw_window_offscreen\nblender.exe :0x00007FF6D3019CC0 wm_draw_window\nblender.exe :0x00007FF6D30197C0 wm_draw_update\nblender.exe :0x00007FF6D2FF6990 WM_main\nblender.exe :0x00007FF6D2CE9C60 main\nblender.exe :0x00007FF6D71E86D8 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FF81E967020 BaseThreadInitThunk\nntdll.dll :0x00007FF81F4FD0B0 RtlUserThreadStart\n```" ]

[ "Regression: Blender 3.1 UI editor resize /border dragging performance\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.1.2\nWorked: 3.0.1\n\nCaused by {6738ecb64e8b6d0c1f5a79f1e63f468b4138de19}\n\nResizing UI editor panels in Blender 3.1 is significantly slower (framerate) and laggier than it was in 3.0. This is not related to the UV Editor performance regression as it happens even when there's no UV editor displayed in the given workspace. \n\nHere's a video of the issue, recorded at 60FPS and slowed down to 50% so the issue is more apparent: watch?v=a0wzLgkf6mk\n\n1. Open blender with a new empty file \n2. Drag the top of the Timeline Editor UI area around\nResult: The UI redraw framerate and input response (lag) are worse than in Blender 3.0\nExpected: The UI redraw framerate and input response (lag) are same or better than in Blender 3.0\n\n", "The 3d cursor looks very dim on AMD cards with the Mesa driver.\nOperating system: Gentoo Linux\nGraphics card: AMD Vega64\n\nBroken: All current Blender versions (2.93 to 3.5)\n\nThe 3D cursor in the viewport seems to be drawn with transparency.\nHowever this is only an issue on AMD cards it seems.\nIt started happening 4-5 months ago, so this is most certainly a driver issue.\n\nOpen the default cube scene and observe that the 3D cursor looks very dim:\n", "VSE thumbnails improvements\n- [x] Check if reusing strip hash used by rendering job speeds up initialization with large number of strips. if it doesn't, fix it anyway. (b534806ecb) Note: it wasn't initialization that caused freezing, it was rather strip `anim` freeing, reusing ghash seems to have no impact on performance.\n- [x] To speed up loading, Check strategy to render \"guaranteed\" set of thumbnails. Especially if cache can be shared this could help a lot. (3a898db3638d) Note: better strategy seems to be other way around as far as performance goes.\n- [x] Make first (and last) thumbnail reflect actual handle position. (29b9187b327c) Note: Only first image is updated, though it's possilbe to update last one too. \n\n- Change cache so thumbnail (and possibly raw) types are keyed by path to file and disregard various context settings and Sequence pointer. Front-end should stay same. ([D12771](D12771))\n - There are rabbitholes like effect strips producing raw images and scene strips rendering at \"host\" scene resolution so simple string key is probably not enough. \n - Added complexity is, that all cache types per frame should be invalidated at once, otherwise prefetch doesn't work and issues like that.\n - Above points don't quite apply to thumbnails, but there are issues of this type in general. For example if we would implement downscaling for movies, first point will apply.\n\n", "Cycles : Massive OSL performance deficit compared to SVM using AcesCG studio profile\nOperating system: Ubuntu 23.04\nGraphics card: 6900XT (not used here)\nCPU: Threadripper 3990X\nRAM: 256GB\n\nCycles Standalone 3.5.0 (under Gaffer though)\nOSL version 1.14.6 (vs Blender 3.5.1's `1a7670600c8b08c2443a78d03c8c27e9a1149140` - specific commit I believe for OptiX support?)\nOIIO version 2.4.11.0 (vs Blender 3.5.1's 2.4.9.0)\nOCIO version 2.2.1 (vs Blender 3.5.1's 2.2.0)\n\nMassive 10x OSL vs SVM performance deficit, using a photo-realistic human asset with a lot of hair + SSS random-walk + normal map, moderate texture usage using mip-mapped/tiled tx.\n\nSwitch between OSL and SVM and compare the performance, as well as htop to see green bars and 99+% CPU utilisation under SVM vs OSL which is mostly red and 20-25% green CPU utilisation (all up 99+% CPU usage?). Also check render-times, SVM takes 7-8 mins/frame vs OSL taking 1hr 10 mins so roughly 10x the time.\n\nAll textures are maketx'd tiled+mip-mapped .tx files.\n\nApologies I cannot share the render due to NDA. I am re-creating something in Blender which just has a sphere with random walk + SSS + normal map, and I get similar results but not as severe. Roughly 7-8 mins for SVM but 15-20 mins for OSL so 2x which I assume is something to be expected? Htop has all the bars as green. Further investigation I believe it might be something to do with OIIO's texture cache getting thrashed, although I am pretty sure I am not hitting its built-in limit here:\n\nosl.cpp#L257\n\nRam usage in both OSL/SVM is about 3.5GB.\n\nWill keep digging...\n\n", "LightGroups & big render resolution results in Hardware failure & macOS crash incl. reboot\nOperating system: macOS-12.3.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.15\n\n**Blender VersionS**\nBroken: version: 3.2.0 Alpha\nBroken: version: 3.2.0 Alpha\n\n\n(GPU only)\nI usually render with a resolution of 5120x2048 on 50% as check resolution. \nWhen I scale it up with the slider to 100% and render I get a stripe pattern as we seen in other bugreports related to M1 Mac's.\nThen when I try double the resolution on 50% I get the same.\n\nThe crash of macOS with a hardware failure and reboot happened when I tried my final resolution 10240x2048 on 100%. (blender calculated a sec. and when the render via GPU only started it hanged for two sec. and then completely crashed)\nTo be honest, a crash like this I never had with 3 imac's the last 6 years. Only with a hackintosh not related to blender...\n\nDespite the fact of the total crash, I decided not to try to reproduce the total crash,\nBUT...\nthe stripes I could with the resolution I used. \n- Open attached file\n- Render once in 50% from 5120x2048,\n- and then render on 100%, or try the higher resolution.\n\n(The scene was: just a plane, monkey and 3 lights, to check the lightgroups and compositing options (worked all fine, exept the render...))\n(see attached blendfile)\n\nSo just 3 pointlights, each into a one of the 3 lightgroups a plane, monkey...(something to get lit),\nresolution 5120x2048 on 50% and render. Then render on 100% and you should get stripes.\n\n`***When there, just delete the lightgroups under Scene -> View Layer Tab and the stripes are gone.***`\nOperating system: (INTEL) macOS-12.3.1-x86_64-i386-64bit 64 Bits \nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.15\n[untitled.blend](untitled.blend)\n", "Object rotation increases render time more than 10 times.\nTested with windows 7 64bit and windows 10 64bit.\nCPU ryzen 5 1600\nGPU GTX 780\nRAM 8GB GDDR4\n(bug appears on other computers too)\n\ntested with 2.79 and 2.8\nWorked: (optional)\n\nIf I rotate object for about 45 degrees around Z axis render time increases more than 10 times.\nBut in preview mode there is no difference in render time before and after rotation.[test.blend](test.blend)\njust rotate the object on attached blend file around Z axis about 45 degrees and compare render times before and after.\n10seconds for 2000 samples before rotation and about 10 minutes after.\nBased on a (as simple as possible) attached .blend file with minimum amount of steps\n", "Alembic mesh cache viewport performance affected even when disabled for viewport\nOperating system: manjaro-kde-21.0.3-210428-linux510 switched to linux512 kernel\nGraphics card: rx6800\n\nBroken: 2.93\n\n\nViewport animation playback performance is affected even when alembic modifier is disabled for viewport and no visible changes are visible.\n\n[blend and abc.zip](blend_and_abc.zip)\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\nAdd mesh sequence cache modifier with alembic file, select an object from abc, hit play in viewport, watch fps go down, disable modifier, watch fps go up, remove modifier or change selection to none, watch fps go up even more. ", "Automated performance testing\nSee here for documentation on the performance testing framework:\nPerformance\n\nThis system is separate from [Open Data ]() by design, for developers rather than users.\n\n**Tasks**\n\n- [ ] Automate running on buildbot\n - [ ] Nightly builds tracking performance over time\n - [ ] GPU support\n - [ ] Publish HTML results on builder.blender.org/download\n- [ ] Tests\n - [x] Cycles\n - [ ] Animation playback\n - [x] Geometry nodes\n - [ ] Object and mesh editing operators\n - [ ] I/O", "Mantaflow - Enabling \"Noise\" breaks fast-moving smoke\nOperating system: Windows 10\nGraphics card: Nvidia GeForce GTX 650 Ti\nProcessor: Intel® Core™ i7-3770\n\nBroken: 2.83.0\nWorked: Never\n\nEnabling noise on fast-moving smoke produces a \"grainy\" result that doesn't match original simulation.\n[C - High Speed Noise - Demo.mp4](C_-_High_Speed_Noise_-_Demo.mp4)\n\nAlso, this problem is less profound when the cache type is set to \"Final\", though still clearly visible.\n\n**Exact steps to reproduce the error**\n1. In any smoke simulation, dramatically increase the velocity of the smoke, either in the emitter settings, or with a force-field.\n2. Bake Data\n3. Bake Noise\n\n[Noise on High-Speed Smoke.blend](Noise_on_High-Speed_Smoke.blend)", "Compositor improvement plan\n# Overview\n\nThis is an initial pass on formalizing proposal how to improve performance and user experience of Compositor.\n\nThere are three aspects which are aimed to be addressed:\n* Ease of use\n* Performance\n* Memory consumption\n\n## Ease of use\n\nCurrently there are many settings which needs to be set and tweak to have best performance: tile size, OpenCL, buffer use and so on. Settings also have some implicit dependencies under the hood: for example, OpenCL needs big tile size, but big tile size might make some nodes measurably slower, and bigger tile size \"breaks\" the initial intention of compositor design to show tiles appearing as quick as possible.\n\nFor OpenCL case it's also not clear when it's actually being engaged: it will just fail silently, falling back to CPU, giving false impression of having GPU accelerated compute.\n\n# Performance\n\nPerformance of compositor is not up to date. Partially due to its scheduler design, partially due to technical implementation which is per-pixel virtual call (which ruins all sort of coherency).\n\n# Memory usage\n\nThis is something what is absolutely out of artists control: some operations require memory buffers before/after the node and those are being created automatically. This makes it hard to predict how much memory node setup requires, and how much extra memory is needed when increasing final resolution.\n\n# Solution\n\nProposed end goal is: deliver final image as fast as possible.\n\nThis is different from being tile-based, where goal was to have first tiles to appear as quick as possible, with giving gradual updates. The downside of this is that overall frame time is higher than giving an entire frame at once. Additionally, tile-based nature complicates task scheduler a lot, and makes it required to keep track of more memory at a time.\n\nIt should be possible to transform current design to proposed one in incremental steps:\n\n* [Temporarily] Remove code which unnecessarily complicates scheduler and memory manager which is GPU support.\n* Convert all operations to operate in a relative space rather than pixel space (basically, make it possible to change final render resolution without changing compositor network setup).\n* Make compositor to operate on final resolution which closely matches resolution of the current \"viewer\": there is no need to do full 8K compositing when final result is viewed as a tiny backdrop on Full HD monitor.\n* Modify operations to operate on an entire frame (or on a given area).\n* Modify scheduler to do bottom-to-top scheduling, operating on the entire image.\n* Modify memory manager to allocate buffers once is needed and discard them as soon as possible.\n* Vectorize (SIMD) all operations where possible.\n\nLook into GPU support with the following requirements:\n* Minimize memory throughput, which implies the following point.\n* Have all operations implemented on GPU, which again implies following point.\n* Share implementation between CPU and GPU as much as possible.\n\nThe steps can be gradual and formulated well-enough to happens as code quality days in #73586.", "Clang on windows\nWhile we have supported building with clang on windows for a while, support has deteriorated a little due to it not getting regular exercise also not all features were supported this is a list of outstanding issues with clang on windows.\n\nOrder of importance:\n\n- {icon circle color=red} **Very Important** - *These we should handle before the next release.*\n- {icon circle color=yellow} **Somewhat Important** - *These issues would be nice to do as soon as possible*\n- {icon circle color=green} **Less Important** - *Extra polish, nice to have*\n- `?` **Incomplete** - *Tasks needing more details before implementing.*\n \n- {icon circle color=green} Better documentation on how to build, [the initial commit ](rB81060ff6b26c22331d2066bb83d492bbe61e936c) had instructions but some of those instructions are no longer applicable \n- {icon circle color=green} [D5962: Fix: Building quadraflow with clang on windows.](D5962)\n- {icon circle color=green} ~~Building with clang+ninja is broken~~ fixed:34bd23dbb7\n- {icon circle color=gray} VS2019 IDE support is broken due to the llvm integration addon, VS2017 IDE works, ninja work for both 2017/2019 \n- {icon circle color=green} [D5976: Add support for building with OpenMP support for clang on windows.](D5976)\n- {icon circle color=green} [D6012: Fix: BLI_expr_pylike_eval_test test failing with clang on windows](D6012)\n- {icon circle color=green} Cycles had a broken AVX2 kernel when running the benchmarks on EC2 \n- {icon circle color=green} Too many warnings a full build log is over 1GB \n- {icon circle color=green} cycles kernels get fed clang flags which it does not recognize `-fno-trapping-math`, `-fno-math-errno` , `-fno-signed-zeros`", "Slow camera view navigation in Cycles rendered shading mode with many instances\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 3.0.0 Alpha\nWorked: 2.93\n\nWhen looking through camera and navigating the viewport (with Lock Camera to View for example), the movement is much slower than a regular perspective view. This happens in rendered shading mode with Cycles, when many object instances are in the scene. Instances can be generated with Geometry Nodes or the old particle system. Also they don't even have to be visible on screen, you could be looking at a single cube somewhere to the side, and the camera navigation will still be slow.\n[camera navigation bug.blend](camera_navigation_bug.blend)\n\n\n - Use the provided file\n - Switch to rendered mode and disable overlays\n - Try looking around\n - Switch to camera view\n # Navigation should be much slower and unresponsive\n\n\n\n", "Mem leak when rendering on GPU\nOperating system: Windows 10 20H2 Build 19042.985\nGraphics card: Gigabyte RTX 30870\n\nBroken: 2.92 - 3.0\n\n\nRender 2.92 splash screen on GPU 60337d495677e942564cce76\n\nOn CPU, memory usage was 18GB, on GPU it went to 60GB and I had to terminate the process.", "Edge Sliding with subdivision surface is extremely laggy (caused by \"bad\" UVs and UV Smoothing in the modifier)\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 527.56\n\nBroken: version: 3.4.0\n\nSo I've been modelling a raptor the last week and I noticed that viewport performance in edit mode was extremely laggy with subdivision surface modifier enabled. I checked the face count and it was just 150k faces which is not a lot at all. It's definitely a modifier problem because when I apply the subd modifier the lag goes away completely even when sliding twice the amount of edges around. I thought that was maybe a hardware issue so I opened task manager and while sliding edges with subd on I get like 4 fps with severe stuttering to top it off. All that while the cpu was at 10% usage and 5% gpu usage with memory barely touched (6gb used out of 64gb). I have nothing else running in the background like an ativirus or music app, I have the latest version of windows 10, latest gpu and chipset drivers, latest bios etc... I even have gpu subdivision enabled without autosmooth to make sure it actually works properly. I even run an sfc/scannow to check for other kinds of issues but nothing changed. Subd in edit mode however is not only slow when sliding edges and vertices but also in adding loop cuts or insetting faces. It is extremely frustrating so it would be great if you could fix this. I know edit mode was not made to handle massive polycounts but 150k is again not a whole lot and definitely comfortable to run on my PC. Here is the file to see for yourselves.\n\n- Open attached .blend\n- Execute Edge slide/loop cuts/inset faces\n[Broken Subd.blend](Broken_Subd.blend)\n\n", "Regression: Importing of collada model with instances became a LOT slower since 2.80 (practically freezing)\nOperating system: windows 10\nGraphics card: geforce gtx 1070\n\nBroken: blender 3.4.1\nWorked: 2.80\n\nImporting of collada model makes Blender not respond anymore with high CPU usage in windows task manager (and never finishes no matter how long you wait)\n\n- import attached collada file\n- Blender not respond anymore with high CPU usage in windows task manager (and never finishes no matter how long you wait)\n\nLink to model file:\n\nCruisn-Exotica-arcade-game-Dual\n\n[doubleexotica.zip](doubleexotica.zip)" ]

[ "ROCm 2.2 OpenCL kernels don't compile\nOperating system: Kubuntu 18.04\nGraphics card: Vega 56, RX 480 8gb\n\nBroken:\n2.8 cc5bdf029324, 23:34\nWorked:\nUse AMDGPU-pro drivers but they only recognize RX480 8gb, or disconnect RX480 to use Vega 56 (it's not a blender issue)\n\nROCm 2.2 OpenCL kernels don't compile and hang, in blender it says is compiling but it never ends.\n\nPS: AMDGPUpro 18.50 drivers recognize Vega 56 and RX480 8gb as different OpenCL platforms, but instead ROCm 2.2 and Windows 10 drivers put these GPUs correctly in the same OpenCL platform. 237937\nPS2: ROCm 2.2 OpenCL kernels compile in Luxrender though" ]

[ "EEVEE: Contact shadows problem on layered hashed materials\nOperating system: linux open source drivers\nGraphics card: rx 6800\n\nBroken: 2.93\nWorked: Never (2.83+)\n\ncontact shadows dont look correct on certain angles for layered alpha hashed materials(hair cards for example). iissue is more noticable in viewport then in final render, but it is noticable in both.\n\nmake a sphere make it alpha hashed and slightly transparent then duplicate and scale geometry with slight offset to make layers then rotate around object with contact shadows enabled on the lamp\n\n\n[contact shadows problem.blend](contact_shadows_problem.blend)", "Improved way to display weight painting colors\nWeight group colors in vertices blend between their neighbors by going through the entire range of values between the 2 points\n\n\nFor example if a vertex has a 1.0 (red) value and one of it's neighbors has a 0.0 (blue) value, the face will show a rainbow interpolation going trough orange, yellow, green, cyan and finally blue\n\n\nIn certain situations where each of the vertices of a face have different values (like 1.0, 0.5, 0.75, 0.0) it will create weird intersections in the middle that look like there is a vertex there that doesn't exist\n\n\nThe problem is even worst when the wireframe is not displayed and we only see gradients of colors indicating vertex values that do not exist\n\n\nThe solution is to fix the color blending mode to simply **mix** between the 2 neighboring colors. This way it is very clear that vertices have values of 1.0, 0.75 and 0.5 and nothing in between.\n", "GP transparent layer doesn't blend with background object \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 3.0.0 Alpha\n\nI have a foreground grease pencil object, and a background grease pencil object (yellow blob).\nThe foreground object contains a layer set to 0 opacity.\nWhere this layer overlaps the background object in the viewport, the background object is invisible.\nExpected behaviour: the visibility of the background grease pencil object isn't affected by transparent layers in a foreground object.\n\n\n\n\n\n[#89286.blend](T89286.blend)\n\nForeground object \"makes hole\" in background, doesn't blend. Rotate viewport camera, suddenly it blends correctly.\nWhen foreground location is set to 0,0,0 it blends correctly too.", "Bug - Refraction BSDF in Eevee interferes with Render properties > Film > Transparency - it blocks the transparency so to speak.\nBest demonstrated with a video (please note that the refraction bsdf is transparent, so objects can se seen through, so can the world background - when the world is not transparent that is) :\n\n[DE3VFV8aIO.mp4](DE3VFV8aIO.mp4)\n\n", "Inconsistent Limitations of Instances in Geometry Nodes\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.2.0 Alpha\n\nIt's not clear what operations can or can't be done to an instance, eg. Extrude and Scale Element, Dual Mesh, Triangulate etc all work fine but Set Position doesn't. \n\nSorry for the complex file but I wanted to leave in all the nodes I'd been through before realising that I was working an Instance still.\nThe Set Position node was the first one in that line that gave me any indication that I was still on an instance.\n\n[12 Marching Squares.blend](12_Marching_Squares.blend)\n", "Erase Alpha in vertex paint not working\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.83.0\nWorked: latest 2.7 nightly build date 2019/06/27\n\nErase Alpha not showing in viewport.\n\ngo to vertex paint set mode to erase alpha and when painting nothing happens.\nalso found no ways to use the alpha data.\n\nVertex Alpha painting are needed for asset creation on some game engines like creation engine.\n", "Right clicking on 3D Viewport Header not working as expected\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 850M/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 537.13\n\nBroken: version: 4.0.0 Alpha\nWorked: (3.6 and older)\n\n\nIf you Right click on empty space of 3D Viewport Header it will not bring Header Menu for disabling or enabling Header. it was possible in version 3.6\nThere are just 2 options to bring menu:\n1- Right click very close to Header items but not inside items(it's really hard to reach)\n2- disable region overlap (many prefer not to turn it off)\n\nI know it is now available to see and select objects through header but I think old right click should back to header because:\n1- constancy( when you right click on all other editor's header that menu will pop up)\n2- beginner friendly (imagine someone new to blender or blender 4 wants to hide header and he does not know new region toggle pie menu so it is hard to hide it)\n\n", "UV: Modified Edges UV overlay don't show on generated Mesh from Geometry Nodes\nBroken: version: 3.6.0\n\n\"Modified Edges\" UV Option does not work when there is Join Node or primitives is used to generate geometry.\n\nIn setup like this, where we generate UVs in GN, and have \"Modified Edges\" overlay work correctly.\n\n\n\nWhen we generate geometry eg. by using Join Node, or Primitives. Overlay no longer works (even in UV sync mode where all UVs should be shown)\n\n\n\n\nI'm not sure if its desired behavior because eg. some setup do work when we modify and not generate new geometry.\n\n\n--\n\n(that's the overlay if someone is not aware of its existence :p)\n\n\n- Open attached .blend file. (Note that the UV for \"Modified Edges\" is showing as expected)\n- Unmute `Join Geometry` node. (UV for \"Modified Edges\" no longer appears)", "GPU subdivision option with subdivision surface modifier collapses geometry (AMD GPU)\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: AMD Radeon HD 7800 Series ATI Technologies Inc. 4.5.14831 Core Profile Context 21.5.2 27.20.20903.8001\n\nBroken: version: 3.1.0 Alpha\n\nSubdivision surface modifier with GPU Subdivision enabled produce geometry that appear to be lost\n\n1. Open default scene.\n2. Add subdivision surface modifier to an object.\n3. Make sure GPU Subdivision option is enabled.\n4. ~~Uncheck Use Limit Surface~~\n5. ~~Disable GPU Subdivision to see correct result.~~\nLast two steps are not necessary for recreation of the bug since this change- 118a219e9d\n[Subdiv GPU Use Limit Surface 2022-01-15 22-13-33.mp4](Subdiv_GPU_Use_Limit_Surface_2022-01-15_22-13-33.mp4)\n\n------\n\nNOTE: This bug seems to only occur on AMD cards on Windows.\n\n**Reproducible on:**\n\n| Operating System | GPU | Driver |\n| ----------------|-------------|---------|\n| Windows 10 | AMD Radeon HD 7800 Series | 4.5.14831 Core Profile Context 21.5.2 27.20.20903.8001 |\n| Windows 10 | AMD Radeon (TM) RX 480 | 4.5.14760 Core Profile Context 20.45.37.01 27.20.14537.1001 |\n| Windows 10 | AMD Radeon(TM) 535 | 4.5.14831 Core Profile Context 21.5.2 27.20.21003.8013 |\n\n**Not reproducible on:**\n\n| Operating System | GPU | Driver Version |\n| -----------------|-------------|---------|\n| Linux Mint 20 | AMD Radeon RX550 | 4.5.14800 Core Profile Context 21.40 |\n| Linux Mint 20 | AMD Radeon RX550 | 4.6 (Core Profile) Mesa 21.3.0-devel |\n| Windows 10 | Intel(R) UHD Graphics 620 | |\n| Windows 10 | NVIDIA GeForce GTX 1050 | 472.39 |\n| Linux Mint 20 | NVIDIA RTX 2070 | 510.47.03 |\n| Linux | NVIDIA GeForce GTX 970M | 495.44\n|Windows-10-10.0.19043-SP0 64 Bits|Radeon (TM) RX 480 Graphics| 21.30.23.01, 22.2.3, 21.10.2 |\n", "Vertex Paint Viewport Overlays stencil mask not working\nplatform: Windows-10-10.0.19042-SP0\nGraphics card: 1080\n\nBroken: 3.1.0 Alpha \n```\n commit date: 2021-11-14 00:26 hash: 7e82c840b7a4\n build date: 2021-11-15, 00:51:48\n```\n\nWorked: Has'nt worked for ages\n\nGo to vertex paint mode, select some polys on a model.\nGo to overlay options/vertexpaint/stencil mask opacity the slider does nothing to unselected polys overlay.\n\n\n\n\n", "Selected object color is weird in wireframe mode if the object color isn't single.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 465.89\n\nBroken: version: 3.0.0 Alpha\n\nOutline overlay seems to be drawn under wireframe.\n\n[#91619.blend](T91619.blend)\n\nImmediately visible in provided file.", "Wireframe opacity doesn't work if object is \"in front\" and viewport shading is set to material preview or rendered.\nBroken: 3.0\nWorked: \n\nWireframe opacity slider doesn't work if the object is \"in front\" and \"texture space\" is checked and viewport shading is set to material preview or rendered.\n\nSelect Cube\nObject properties-> viewport display -> in front ON\nObject properties-> viewport display -> texture space ON\nIn the viewport overlay panel, drag the wireframe opacity to lower the value.\nchange the viewport shading to preview or rendered, wire will have full opacity.", "Alpha doesn't work as a mask \nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.79\n\nBroken: version: 3.0.0\n\nI use an alpha channel as a mask for mix node. If I put an alpha channel from the same node where I take a color channel I get a strange result but if I grab a mask from an alpha channel a copy of the same texture node I get the proper result. \n\n\n\n\n[bug_11.blend](bug_11.blend)\n\n", "Collada import/export: transparent textures don't work\nOperating system: Linux\n\nBroken: blender-3.3.0-alpha+master.dfb8c90324c0, probably all versions since 2.8\nWorked: 2.79\n\nCollada cannot import or export transparent (alpha) textures.\n\nImport this .dae\n\n[BlenderLogoDae.zip](BlenderLogoDae.zip)\n\nThe result (with 2.79 for comparison).\n\n\n\nManually hook up the alpha in the node graph and export. Open the .dae in a text editor and confirm there is no <transparent> element (which sets the alpha texture). You can confirm exporting works in 2.79 by exporting and reimporting.\n\n**Other**\nNote that OpenCollada's parsing of transparency is weird and buggy, see [D11259](D11259) and 651.", "Wireframe opacity slider does not work.\nOperating system: Windows 10\nGraphics card: gtx1060\n\nBroken: 3.1\nWorked: \n\nWireframe opacity slider does not work if smooth wires is turned off.\n\nopen preferences>viewport\ndisable smooth wire overlay\nsave preference\nset viewport to wireframe shading\nin the overlay panel in viewport drag the wireframe opacity slider, the wire color will fade to black, the opacity won't change." ]

[ "viewport material transparency faces goes through objects with in front option enabled\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: GeForce GTX 960/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.36\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional)\n\n\nwhile Changing the alpha of the the material in viewport setting the faces goes through the objects instead of becoming transparent\n\n\n\n\n1>enable the In Front in the object properties \n2> in material Viewport display change the alpha to 80%\nsuddenly the \n" ]

[ "2.8 - Dynamic Paint and Rigid body does not work\nOperating system: Linux-5.0.1-050001-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 960/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.43\n\n\nBroken: version: 2.80 (sub 51)\nWorked: 2.79\n\n\nThis file is for 2.79.\n[dynamic_paint_paint-Rigid Body-2.79.blend](dynamic_paint_paint-Rigid_Body-2.79.blend)\nI can not see paint results in LookDev/Rendered mode in 2.8. I also could not recreate the scene from scratch in 2.8", "Grease Pencil: various issues\n*Note: this tasks can be split up once we get feedback on issues, for now this is to discuss changes to make.*\n\n----\n\n\n- Brush pixels are not pixels,\n```\nshouldn't we use a meaningful value?\n(pixels or blender units).\n```\n\n- Grease pencil brushes have material pointers which are globally overwritten.\n\n```\nWhen you set a material, all brushes are looped over and have their material values clobbered.\n```\n\n```\nWhy have a material attached to the brush at all? This could be a scene level setting (or at least not stored per-brush).\n```\n\n```\nSee: `BKE_brush_update_material` called by rna `rna_Brush_material_update` when setting any material.\n```\n\n- Grease pencil sculpt uses it's own brush system,\n```\nis there a good reason for this?\n```\n\n```\nSince these brushes are fairly advanced (pressure & curve mapping options)\nwouldn't it be better to use the existing brush system?\n```\n\n *Note: while particle system for eg: has 'simple' brush types like this, grease-pencil sculpt brushes are more advanced and duplicate more of what we have for regular brushes used elsewhere.* \n\n- Grease pencil weight brush seems to have no way to paint a weight value (only add weight).\n\n```\nAs with mesh weight paint we should be able to paint a weight.\n```\n\n- In edit-mode, vertex selection exists but is hidden by default (0.0 opacity).\n\n```\nShouldn't they be visible?\n```\n\n----\n\n### Lower Priority\n\n- Internally grease pencil sculp is called:\n\n - grease pencil edit (`GP_BrushEdit_Settings`, `GP_EditBrush_Data`, `GP_BRUSHEDIT_FLAG_APPLY_UV`)\n\n```\nThis is highly confusing since we have grease pencil edit-mode, can't we rename these `GP_Sculpt*`. ?\n```\n\n - grease pencil paint (`GPENCIL_OT_brush_paint`)\n\n```\nPaint is easily confused with draw, seems this is for sculpt/weight painting.", "Use Scene Spacing Distance by default\nNOTE: This design is still being investigated and discussed. \n\n# Issue\n\nThe default for brushes in every mode has always been to use view based spacing for each stroke.\nThis wasn't always an issue since brushes used to work via a projected falloff.\n\nBut when using a spherical falloff on curved surfaces this can lead to inconsistent spacing artefacts.\nSince more painting and sculpting modes are implementing brushes with a spherical falloff by default, it is a reasonable adjustment to make the spacing also scene based.\n\n# First Steps\n\nBefore changing any defaults we should investigate any remaining issues, downsides or bugs if this behaviour would be implemented by default for all brushes.\nOnce it is clear that this change would be an overall improvement, and how feature might need to change.\n\n### Fix remaining bugs and issues\n\nFurther issues from using Scene Spacing need to be investigated.\nThe currently known reports are:\n- #98111 (Sculpt mode: Draw Sharp brush artefacts when using Scene Spacing in stroke settings)\n- #99638 (Line stroke method not working with Scene spacing)\n- #99642 (Clay strips brush freeze when coming from outside the horizon of surface curvature)\n- #99644 (Draw Sharp brush stops on anchored during the stroke)\n\n# Proposal\n\nScene spacing currently accomplishes two distinct behaviours:\n1. The spacing is consistent in 3 dimensions\n2. The spacing is consistent in world space (especially when zooming in/out)\n\nInstead of just making Scene Spacing Distance the default, we should change the \"View\" scene spacing to work 3 dimensionally for a better brush behaviour in general.\nThis could be accomplished by associating these two different behaviours with distinct settings:\n\n1. Setting the falloff to be Spherical or Projected = 3D or projected spacing distance\n2. Switching between View & Scene spacing distance = The spacing is a percentage of the relative brush size diameter or an absolute world size\n\nThe UI should then also reflect that change, by changing the spacing slider value:\n- View = Percentage\n- Scene = Distance in units\n\nThis updated View spacing then needs to be integrated in every mode for 3D painting/sculpting.", "Greasepencil z fighting alpha issue\nHi,\nnot too sure if this is known issue or any fix but I am getting this alpha issue from certain angle when I have 2 different grease object. \n\n", "creating custom (combined ) texture brush is not visually updated in texture brush tab\nWin 10 \nIntel i5\namd radeon \n\nblender 2.8 \n8996e26116f0\n\nGo into the texture for the brush settings and enable ''úse nodes'' \nGive the new texture brush a name \nMix two textures together , the output result is not reflected in the text.brush tabe\nThis was not the case wiith blender 2.7 \n\n\n\n\nAlso , when rotating the brush (ctr-F) ,the alpha overlay of the brush texture is nor", "Rim on objects when shadow catcher is in play\n开启阴影捕捉后边缘有硬边BUG \n\n\n\n", "Sculpt Paint: Smoothing Strength\nNOTE: This design is still under discussion and potentially incomplete.\n\n# Issues\n\n### Hard to access smoothing strength\n\nIn sculpt mode it is possible to smooth geometry or colors (blurring) directly by holding `Shift`.\nAdjusting the smoothing strength should be very accessible.\nUnfortunately the current design in sculpt mode is less than ideal.\n\nFor sculpting brushes the strength of the smooth brush is used as the global smoothing strength.\nSo the workflow has been to switch to the smooth brush and adjust the brush strength there.\nThis is not as intuitive as changing the strength of the currently active brush.\n\n### Bloating keymap & toolbar\n\nSince Blender 3.2 we also have a paint brush within the same mode. The issue here is that the paint brush smoothing strength is hard-coded. We must add a way to control the color smoothing strength.\nWe could change that wit the same solution as the regular sculpting brushes, but that would mean adding another smoothing brush and another shortcut. \nThis unnecessarily bloats the keymap and toolbar further. We shouldn't add any more smoothing brushes unless really necessary.\n\n# Proposal:\n\nA possible solution for this would be to make the smoothing strength a brush setting for every brush.\nThis would have the following advantages:\n\n- We could add a shortcut operator on `Alt F` to control the smoothing strength of the currently active brush. This makes the setting much more accessible.\n- Depending on if you are using a sculpting or a smoothing brush, the shortcut operator would control the related smoothing strength\n- The smoothing strength can have a \"Use Unified Strength\" toggle (enabled by default), to optionally give each brush the smoothing strength it needs (This would need further investigation into the toggle behaviour)", "Use draw manager for clip editor drawing\nCould be stalled to 2.93 as there isn't an real issue here as far as we are aware of. The compositor issues are more tangable.", "Resetting default value does not work in popovers or menus\nOperating system: Windows 10, 64 bits, \nGraphics card: Nvidia GTX 1080 8 gb, Nvidia GTX1070 8 gb\n\nBroken:\nblender-2.81.0-git.ce2368cf69f5-windows64\n\nWorked: (optional)\n\nResetting options in the context menu (right click) of things like brush stroke strength, Autosmooth, plane offset and other parameters does not work at all, but does work through active Tool side panes and top bar. Probably it should work everywhere? If I'm working full screen I would want to be able to reset it in the context menu with my stylus as I sculpt.\n\n\n\n1 Create any mesh\n2 Enable Sculpt mode\n3 Select any brush\n4 Right click to invoke context menu\n5 Try resetting any of the parameter to default values\n", "Sculpt mode. Layer brush. Persistent base breaks with local mode.\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.92.0 Alpha\n\n\nIf I set persistent base in default mode and go Local, i loose persistent base. Same for Local→Main\nI can't rememeber if persistent base or sililar thing used in another area. But I assume, it may be broken too.\n\n[2020-12-16_17-15-04.mp4](2020-12-16_17-15-04.mp4)\n\nsummon @PabloDobarro ", "Update sculpt mode symmetry options as properties of meshes\n(NOTE) This task is part of the community & sub tasks of the workboard. Anyone is free to pick up this task and contribute. For any questions or needed reviewers, please tag @JosephEagar and @JulienKaspar .\n\nThere have been multiple patches to improve the symmetry behavior on painting/sculpt modes:\n- 5502517c3c\n- fa9b05149c2ca3915a4fb2670c87a648d927336c\n\nSome `bpy.data.scenes` properties have been left out of this and still need to be unified as a mesh property instead:\n- `tool_settings.sculpt.lock_x`\n- `tool_settings.sculpt.lock_y`\n- `tool_settings.sculpt.lock_z`\n- `tool_settings.sculpt.use_symmetry_feather`\n- `tool_settings.sculpt.radial_symmetry[0]`\n- `tool_settings.sculpt.radial_symmetry[1]`\n- `tool_settings.sculpt.radial_symmetry[2]`\n- `tool_settings.sculpt.tile_offset[0]`\n- `tool_settings.sculpt.tile_offset[1]`\n- `tool_settings.sculpt.tile_offset[2]`\n- `tool_settings.sculpt.symmetrize_direction`\n\n", "Knife Tool produces bad geometry in certain cases (disconnected edges/vertices, overlapping faces)\nif i use tools knife i see this bug.\n\n\n\n", "2.91 - Undo in Sculpt Mode switches tool to inflate brush (specific to attached file)\nOperating system: Win 10 64-bit\nGraphics card: Nvidia GFX 1060 6GB\n\nBroken: \n - 2.90.1\n - 2.91.0, from steam, seems *file specific*\n - 2.92.0\nWorked: 2.91.0, same version, fresh file\n\nWith attached file: In sculpt mode calling undo changes the tool to \"Inflate\" besides undoing.\n\nAdditionally, when undo is called via menu, the undo operation happens only when undoing so far back sculpt mode was entered. On reentering Sculpt Mode, tool is set to \"Inflate\". \n\nThis issues seems to be tied to this specific file, as it's not reproducible in a fresh file. The file was created with blender 2.91.0. I sculpted multiple objects using both Dynamesh and Remesh in Sculpt Mode. I deleted the sculpts to keep the file size small and since the same behavior persisted with newly created geometry (in this case a cube). I don't know how to debug this any further, any hints appreciated. As a workaround I will try to reimport the geometry into a new file.\n\nBased on the attached .blend file [SculptModeBug.blend](SculptModeBug.blend)\n\nUndoing with shortcut:\n - Select object \n - enter sculpt mode\n - select any other tool than \"Inflate\"\n - modify mesh with tool\n - hit ctrl + z\n - tool switched to \"Inflate\"\n\nUndoing via menu:\n - Select object \n - enter sculpt mode\n - select any other tool than \"Inflate\"\n - modify mesh with tool\n - modify mesh with tool\n - call \"undo\" from menu \"edit/undo\" - this has no effect\n - call \"undo\" from menu \"edit/undo\" until the point object goes back to Object Mode - changes get undone\n - On entering Sculpt Mode again, the active tool has switched to \"Inflate\"\n", "Sculpt brush with texture breaks sculpt tiling\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Radeon (TM) RX 470 Graphics ATI Technologies Inc. 4.5.14757 Core Profile Context 20.11.2 27.20.14501.18003\n\nBroken: version: 2.93.0 Alpha\n\nIf the sculpt brush has a texture enabled with mapping set to *Area Plane*, then tiled strokes break up and do not match properly at the seams.\nIn this example the original stroke is on the right.\n\n\n1 Add texture to a brush\n2 Set mapping to area plane\n3 Enable tiling in the Symmetry drop-down menu\n4 Try sculpting at the seams\n\nOr use the file\n[tiling_bug.blend](tiling_bug.blend)\n\n", "Intersecting Meshes are not marked as overlapping anymore when selected\nOperating system: Linux-5.4.0-48-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.90.1\nWorked: 2.83.x\n\nIn blender 2.83.x Solid View intersecting objects had, if selected, the otherwise hidden parts shown in a more muted color when overlapping. with 2.90.x this is no longer the case. \nI hope this did not get removed intentional but is a missing feature.\n\nDuplicate the default Cube, move it a bit so that is still intersects with the first cube and confirm movement. \nWatch the outline in Solid View be the same color and thickness all around the mesh. \nin 2.83 while moving the white outline turned light grey when intersecting and the selected object outline turned a darker orange. This is very helpful for positioning elements.\n\n" ]

[ "Clay Strips Brush Doesn't work correctly when the size brush is bigger and the thickness is thin.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.90.0\n\nClay Strips Brush Doesn't work correctly when the size brush is bigger and the thickness is thin. The brush works correctly when I use the key control. I was working on sculpting 3D character y this problem occurred when I sculpted on the foot. And Later I simplified the geometry in a new file but it was the same result.\n\n1-Sculpt mode 2-Clay strips brush 3-Bigger size 4-Sculpt on the thinnest surface\n[Clay strips Brush.blend](Clay_strips_Brush.blend)\n[Video Clay strips Brush.mp4](Video_Clay_strips_Brush.mp4)\n\n" ]

[ "Blender 2.90.1 crashes after rendering the first few frames of animation.\n**System Information** \nOperating system: windows 10 64 bit\nGraphics card: Nvidia GeForce GTX 1650\n\nBroken: (2.90.1 3e85bb34d0d7; master)\nWorked(didn't): (only tried on 2.8.3 and did not work at all)\n\nWhen I try to render the animation, after about 50 or so frames(varies randomly), blender crashes with no alerts or anything that tells me what the error is, it simply closes.\nI've rendered another project, and that did indeed render, so I'm left confused why this short animation would cause a crash.\n\nEdit: I disabled motion blur, and it rendered fine. However, I would really like to know why blender won't render if motion blur is enabled. The renderer is Eevee and while my laptop isn't a super computer, it should be able to handle it. Thank you all in advance!\n\n- Open attached file\n- render the animation\n[cube-loop-tut-V1.0.blend](cube-loop-tut-V1.0.blend)", "Blender freezes with big custom data\nmacOS 10.14\nBlender 2.92, 2.93.0, \n\nI found if I store big data (like 5 millions of points) then Blender will freeze a bit. This is not a huge bug but it would be nice not to freeze blender.\n[tmp1.blend](tmp1.blend)\n\n\n\nFor example, a console looks like this if I do \"print(test2)\". And there is no freeze in the console.\n\n\n- Run the script\n- Hover over the custom properties section. (try to rotate the cube to observe the freeze, beachball may take time to show up)/\n\n\n\nMemory usage rises and falls on every hover.", "Blender 3.2 crashes on opening a specific file\nOperating system: Windows 11\nGraphics card: RTX 3060\n\nBroken: 3.2.0\nWorked: unknown\n\n**problem**\nBlender immediately crashes when opening the file steam_engine.blend\n\nThe latest object I was working on, was UV-unwrapping a mesh which I think I called 'greenthingy' or something like that. There was no crash prior to re-opening the file. I attach the file (hope it's not too big and apologies for the mess...). crash report and system_info.\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[steam_engine.crash.txt](steam_engine.crash.txt)\n\n**Steps for others to reproduce the error**\n[steam_engine.blend](steam_engine.blend)\n[#99105.blend](T99105.blend)\n\nAfter opening `steam_engine.blend` file Blender crashes. \n`#99105.blend` is same file as `steam_engine.blend`, but material preview panel has beeen collapsed. This file does not crash. Interestingly it does not crash even when panel is re-opened. Tested with GPU and CPU\n", "Texturing with anchored type stroke\nOperating system: Linux-5.4.0-47-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.51.06\n\nBroken: version: 2.91.0 Alpha\n\nTexture paint with stroke set to anchored while using texture significantly slows down blender up to non responsive state and crash while paint texture is larger than 2k.\n\n[Texture pain stroke ty[e anchored.blend](Texture_pain_stroke_ty_e_anchored.blend)\n", "Crash when switching from solid to rendered view\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.6.0 Alpha\n\nI have a scene about 1.7GB in size with 4K textures but renders in 2K using simplify, blender crashes when I open the file > hit F12 and render the scene then close the render windows and switch to render view > it starts compiling then blender closes, in the past it would show out of memory this report: \n\n105522#issuecomment-904934 \n\nThis issue happens only when persistent Data is active\n\nunfortunately I can't share the file at the moment. \n\n\n\n\n", "Geometry node - setting lacunarity to 0 crashes Blender (musgrave texture)\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.93\n\nBroken: version: 3.2.2\n\nIn a fairly simple geonode tree, trying to set lacunarity to 0 will crash Blender instantly.\n\n- Use attached Blend.file\n- In the 'set position' branch, try to set 'musgrave texture' lacunarity to 0\n- Crashes everytime for me\n\n[Geonode attribute.blend](Geonode_attribute.blend)", "Rendering object in Cycles X preview hard crashes whole system\nOperating system: Windows 10 x64\nGraphics card: nVidia RTX 8000\n\nBroken: blender-3.5.0-alpha+universal-scene-description.1f9e90bb1cf3-windows.amd64-release\n\nRendering of scene in Cycles preview hard crashes whole system. System screen goes to black and the fans go to max.\n\nLoad blend file at link: brutes_set1C_trimmed.blend?dl=0\n\nSet screen to Render preview and navigate around.\n\n", "Redo panel does not work for the 3d cursor\nUbuntu 18.04\n\nBroken: 7d38654\n\nRedo panel for moving the 3d cursor does not work as expected.\n\nStart moving the 3d cursor and then change the value in the redo panel.\n\n\n\nMy guess is that this happens because the 3d cursor is not effected by the normal undo.", "Crush while baking a cloth simulation\nOperating system: macOS-13.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Pro 1.2\n\nBroken: version: 3.6.2\n\nI was performing the baking of a cloth simulation. During the initial and subsequent baking attempts, everything worked fine. However, when I tried to re-bake the simulation after moving the cursor in the timeline, the program crashes. I tries to bake that cloth in previous versions. It didn't work as well\n\nTo trigger the crash, you simply need to move the cursor in the timeline and then initiate the baking process.\n\nUnfortunately, i cannot add the file (size is 200mb and i can't make it smaller ), so than i added the crash txt file and a link to the file on google drive\nview?usp=sharing\nThank you!\n\n", "Segmentation Fault Crashing Blender when Rendering\nOperating system: Linux-5.19.0-46-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.54.03\n\nBroken: version: 3.6.1\n\nDuring rendering, Blender will often crash with segmentation fault \"Memoryblock MORE THAN 1 MEMORYBLOCK CORRUPT: error in header\" message.\n\nNeed guidance to help provide a blend file to reproduce the issue.\n\nThe issue occurs 100% of the time with an old version of a large blend file on certain frames. I have attached the debug and crash reports. \n\nMy workaround was to turn on subframes and render subframes either slightly earlier or later than the main frame. For example, if the frame I wanted to render was 187, I would render frame 187.1 or 187.2 until blender would render the frame.\n\nI believe the problem relates to a model I had commissioned, but I cannot recreate the issue in other blend files with that model.\n\nGPU memory usage is about 7.6/8 Gb, so there is headroom, and there is plenty of RAM. It does not matter if it is rendered on GPU/CPU, GPU only or CPU only it will always crash.\n\nI have tested on 2 other Windows computers with the same results, so it is not Ubuntu related.\n\nOther frames render perfectly.\n\nThere is no issue with viewport render mode. Often I get \"out of memory\", but that is only because Blender does not dump some things from memory before viewport rendering an updated scene. It's bound to be a known issue and easy to work around.\n\nI have prepared the file to upload here, but it's 1.9 Gb zipped. Before I send it there maybe some advice about troubleshooting my end someone might know and can help me with. I have spent 3 solid weeks trying to troubleshoot this and the only reliable way I know of reproducing the issue is with the full file.\n\nN.B. The update to 3.6.1 saw me return to this with an updated and improved version of the file. Background rendering will SOMETIMES render problem frames. I have included debug files for this as well.\n\nIs there any advice I can get as to how to narrow down this problem?\n\n", "Blender 3.3 crashing unnecessarily/ unexpectedly/ frequently/ no apparent reason until now\nOperating system: Windows 10 Pro\nGraphics card: Intel i3\n\nBlender 3.3\n\n\n\nWhenever I open blender with some projects...whenever I enter material preview mode or try to render (sometimes even in solid mode...still slow) Whenever I try to apply a material and the crashes even when the project is simple and not finished or being rendered....it was working just fine and started to act weirdly during the last couple of days...I even tried, manually, to allocate more RAM to the program but that just crashed the whole thing. \n\n\nExact steps for others to reproduce the error:\n\nopen the blender file I attached as an example (The project that blender doesn't tolerate) try to apply a texture or render \n\nNOTE: I'm looking for a software solution...I don't think reproducing the error on a RTX 3090 is fair...so...we'll probably look for software solution or tweak the settings.\n\n", "Cursor Warped on Scale Cage Cancel Transform\nOperating system: Windows 11, Linux Zorin\nGraphics card: GTX 1060 MSI OC 3GB\n\nBroken: 3.3.x - 3.6.2 (Latest)\n\nCursor warp when i cancel resize operator using Scale Cage tool.\n\n1. Open Blender\n2. Activate Scale Cage tool from the toolbar\n3. scale anything, but cancel operation with RMB\n\n", "Blender crashes right after loading up a blend file or when creasing a specific mesh on a slightly older version of said blend file\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.77\nCPU: Ryzen 7 3700X\nRAM: 16GB, roughly 7GB available when loading\n\nBroken: version: 3.2.0\n\nFollowing a sculpting course i'm currently trying to crease the claws of a model in the attched blend file.\nBlender keeps crashing without further notice when doing so. I managed to save in the specific state that causes the crash, but can also reproduce it by creasing the front-left hoof/claw-thing in the older state\n\nLoad up file, change to sculpt mode without hiding linked duplicates (feet/hooves)\n[dog!.blend](dog_.blend)\n\nOld:\nLoad up attached blend file -> Crash immediately after apparently successfully loading the scene\n[dog!.blend](dog_.blend)\nLoad up attached blend1 file, crease said claw region with Dyntopo active -> Crash soon after doing so\n[dog!.blend1](dog_.blend1)", "Error: Illegal address in CUDA queue copy_from_device\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.6.0\n\nWhen rendering in GPU mode, Render crashes.\n\nAttached Blend File\nAny frame after 1150 crashes on render with the title error for us.\nTry rendering frame 1160 - This one crashes for me all the time.\n\n", "Cycle crash\n[System Information](https://projects.blender.orgattachment)\n\nBlender version: 3.4.1\n\nCrash when switching from Material Preview to Rendered in Viewport Shading and GPU Compute was set for Cycles. In CPU mode it's OK.\n\n\nOpen project.\nSelect CPU in Cycles settings.\nSwitch from Material preview to Rendered in Viewport shading.\nNow it works.\nChange to GPU Compute in Cycles settings.\nCrash...\n\n" ]

[ "(Blender 2.8) Left Click, Drag, then Scale crashes Blender\nWindows 10 64-bit\nZOTAC Nvidia GTX 1070\n\n\nBlender 2.8 - September 10th Build\n\n\n\nOpen a new project. Click and hold left click anywhere. Drag while holding. Press S. Software will crash. If you don't drag it won't crash. If you don't hold it won't crash. You must both hold and drag then press S without letting go of the left mouse button.\n" ]

[ "Shadow catcher ignored in AO pass\nOperating system: Windows 10\nGraphics card: Nvidia RTX 2080 Ti\n\nBroken: 3.1.2 (also tested in 3.2 Beta and 3.3 Alpha on June 7th)\nWorked: (I recall this working as expected in pre-3.0, before the new shadow catcher was introduced)\n\nThe AO pass doesn't work as expected when using a shadow catcher. Shadow catcher's AO is not visible in AO pass. In this sample render, the floor is a shadow catcher, and as seen, AO is ignored in that area:\n\n\nThis is the expected result (created with the same exact same scene and settings, just with a floor plane that isn't a shadow catcher):\n\n\n1. Render the file.\n2. Check the AO pass.\n\nThis attached file is a very simple version of the same scene with just a sphere and a plane.\n[AOTest.blend](AOTest.blend)", "Cycles - AO Shader node causes infinite \"loading render kernels\" and crashes on return to solid mode view\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.09\n\nBroken: version: 3.2.0\n\nAO Shader node causes infinite \"loading render kernels\" and crashes on return to solid mode view\n[AO BUG Test.blend](AO_BUG_Test.blend)\nAttaching Sample file.\nJust render the scene in viewport (doesnt render, just \"loading render kernel\" message)\nTry to go back to solid view - results in freeze/crash\n\nregards\n", "Render will not load anything but a cube silhouette in a new untouched file\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: AMD Radeon(TM) R4 Graphics ATI Technologies Inc. 4.5.14830 Core Profile Context 21.09.04 27.20.20904.1\n\nBroken: version: 3.6.2\n\nSo, when I open in a new file, new to blender and following a tutorial, I go to render the cube immediately, but instead of a gray cube with shading, it is a transparent silhouette. I have touched nothing, just installed from steam, and opened. Then that, IDK what's going on with it.\n\nIDK what to put here, since I didn't do anything.\n\n\n", "Eevee's AO shades even outside of the distance\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.90.0\n\nEevee's AO creates certain amount of gradients and shading even outside the Distance area. I'm not refering to the typical screen space oclussion problems. \n\nFor example, if you turn Distance to 0, instead of getting the same result as having AO turned off, you get a gradient. Here's how it looks with a flat lit, diffuse scene:\n\nDistance 0, Factor 0:\n\nDistance 0, with a Factor of 10 to exagerate the effect:\n\n\nEven if the gradient is subtle with Factor 1, it can be very visible in certain workflows. Example: There are no other screen space effects in this image, other than AO, which has small distance anyway but this gradient spans the entire image (I'm aware that 360 is not officially supported but still, the gradient itself is not a fault of the 360 stitching and can be a problem for regular images too \n\nIf this is a limitation of the method, may I suggest a clamping slider of some sort for very light values? Thank you❤\n\nOpen the file and turn AO on and off\n[AOBUG.blend](AOBUG.blend)\n\n", "Alpha bug on Principaled BSDF with GPU\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA RTX A2000 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.92\n\nBroken: version: 3.2.1\n\nDecreasing alpha makes material black in the Principaled BSDF shader when using GPU compared to CPU \n\n\n\n - Decrease \"Alpha\" on the Principaled BSDF shader\n - Change to device to either \"GPU Compute\" or \"CPU\" \n - Adjust alpha value to discern difference if not apparent when switching Device property\n\n\n**Notes**\nCPU appears to work correctly with this value, it's only on GPU on my machine that causes alpha to make the material black. Bug appears in rendered images too and not just viewport, and is present in both CUDA and Optix. \n\nMake my alpha 0% if duplicate.\n\n[Blender GPU Alpha Bug.blend](Blender_GPU_Alpha_Bug.blend)", "Cycles/Compositor: The denoiser returns a black image if not all requirements are fulfilled\nOperating system: Default KVM (Proxmox, Linux)\nGraphics card: None\n\nBroken: Blender 3.0\nWorked: Before Blender 3.0 the frame was just not denoised\n\nIf you enable OIDN/Optix denoising without having the needed CPU flags you will get a black frame back.\nThe desired outcome would be an aborted render with a helpful error.\n\n\n1. Set up a default KVM \n2. Enable denoising \n3. Render over the CLI one Frame\n4. It will be black\n\n**Further information**\nThe config of a default KVM:\n```\nArchitecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 40 bits physical, 48 bits virtual\nCPU(s): 6\nOn-line CPU(s) list: 0-5\nThread(s) per core: 1\nCore(s) per socket: 6\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: AuthenticAMD\nCPU family: 15\nModel: 6\nModel name: Common KVM processor\nStepping: 1\nCPU MHz: 3399.998\nBogoMIPS: 6799.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 384 KiB\nL1i cache: 384 KiB\nL2 cache: 3 MiB\nL3 cache: 16 MiB\nNUMA node0 CPU(s): 0-5\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Spec store bypass: Not affected\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Full AMD retpoline, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx lm rep_good nopl cpuid extd_apicid tsc_known_freq pni cx16 x2apic hypervisor cmp_legacy 3dnowprefetch vmmcal\n l\n```\nThe command that I used to render one frame: \n```\n./blender -b --debug --debug-cycles --python-expr 'import _cycles;print(_cycles.with_openimagedenoise);print(_cycles.with_embree);print(_cycles.with_osl)' \"your path to the .blend\" -f 1 \n```\n\n[oidn_disabled.blend](oidn_disabled.blend)\n[oidn_disabled.txt](oidn_disabled.txt)\n\n[oidn_enabled.blend](oidn_enabled.blend)\n[oidn_enabled.txt](oidn_enabled.txt)\n\n[optix_enabled.blend](optix_enabled.blend)\n[optix_enabled.txt](optix_enabled.txt)\n\n\nIf I set up the KVM to use HOST at least OIDN works\nHOST config:\n```\nVirtualization: AMD-V\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 384 KiB\nL1i cache: 384 KiB\nL2 cache: 3 MiB\nL3 cache: 16 MiB\nNUMA node0 CPU(s): 0-5\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Full AMD retpoline, IBPB conditional, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx1\n 6 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core ssbd ibpb vmmcall fsgsbase tsc_adjust bm\n i1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero xsaveerptr virt_ssbd arat npt nrip_save arch_capabilities\n```\n**Why it would be nice to have the bug fixed:**\nWe have a renderfarm that consist out of random machines and it is always a problem if random frames are black/not denoised/strange colored/broken/out of order. ", "GPencil: Hardlight blend issue\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 960M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.66\n\nBroken: version: 2.83.0\n\nThe semi-transparent textures of Grease Pencil material turn black when their layer is in HardLight blend mode.\n\n- open the provided blend file\n- see, wrong alpha blending\n- NB: change the blend mode and it works as expected\n\n[GP-hardlight-blend-issue.blend](GP-hardlight-blend-issue.blend)\n\n", "Default Value for Viewport Display Color's Alpha Wrong.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.09\n\nBroken: version: 2.92.0 Alpha\nAlso tested on 2.83.8 and same problem.\nWorked: It do not think this has ever worked\n\nWhen reverting to the Default Value for Viewport Display Color the alpha gets down to 0.8 making the object semi transparent.\n\nOpen up a new scene.\nSelect the cube.\nGo to the materials tab.\nScroll down to \"Viewport Display\"\nRight Click on \"Color\"\n\"Reset to Default Value\"\nResult: RGB values go back to the default, but Alpha goes to 0.8\n\n\nThank you so much !", "EEVEE: Mixing between a fully transparent and holdout shader yields a half transparent result\nOperating system: macOS-10.13.6-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 580 OpenGL Engine ATI Technologies Inc. 4.1 ATI-1.68.25\n\nBroken: version: 3.2.0 Release Candidate\nWorked: I thinck the issue is for \"all\" versions of blender\n\nHoldout shader having issue with alpha (semi transparency).\n\nPut a image plane with exr image that contain alpha. When using a holdout shader mixed with another shader (transparency shader for exemple) the alpha of the image sequence is used for the mix of the 2 shaders. It should have a complete transparent image with the holdout of other object in the scene. The issue is that there is a dark outline for the semi transparent. Holdout have issue dealing with semi tranparency ? or maybe something else.\nSee the attached image to understand the issue.\n\n\n", "VSE: The blend and opacity settings that strips are created with does not match the default values they revert back to\nOperating system: Linux-5.4.0-80-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce 9800 GT/PCIe/SSE2 NVIDIA Corporation 3.3.0 NVIDIA 340.108\n\nBroken: version: 3.0.0 Alpha\nWorked: Unsure\n\nStrips are created by default with blend method \"Cross\" and blend opacity \"1\", however reverting to default values results in \"Replace\" and \"0\".\n\n- Open VSE\n- Add/create any strip\n- Under strip properties in sidebar, note that blend method is \"Cross\", and opacity value is \"1\"\n- Choose \"Reset to default value\" from context menu for both of these settings\n- Blend method will change to \"Replace\"\n- Opacity value will change to \"0\"", "The 3d cursor looks very dim on AMD cards with the Mesa driver.\nOperating system: Gentoo Linux\nGraphics card: AMD Vega64\n\nBroken: All current Blender versions (2.93 to 3.5)\n\nThe 3D cursor in the viewport seems to be drawn with transparency.\nHowever this is only an issue on AMD cards it seems.\nIt started happening 4-5 months ago, so this is most certainly a driver issue.\n\nOpen the default cube scene and observe that the 3D cursor looks very dim:\n", "Blocky artifacts in various render passes when rendering Mantaflow volumes in Cycles\n## Simplified Report\n\n- Open the attached file and set the VDB path to the attached VDB volume.\n- The VDB volume file is the result of a Mantaflow simulation generated in Blender using Quick smoke.\n- Render the scene and look at the shadow pass.\n- The shadow pass contains blocky artifacts as shown in the attached image.\n\n[blockyVolumeShadowPass.blend](blockyVolumeShadowPass.blend)\n\n[fluid_noise_0060.vdb](fluid_noise_0060.vdb)\n\n\n\n## Original Report\n\n[system-info.txt](system-info.txt)\n\nBroken: 3.1, 3.4.0 Alpha, branch: master, commit date: 2022-09-27 23:42, hash: 5beaecb33e74\nWorked: most likely never\n\nRendering imported vdb generated with Mantaflow in Cycles gives blocky artifacts in various passes. \nMost visible artifacts are in shadow pass and denoising albedo. Less visible can be seen in volume direct and indirect.\n\nThis bug is particularly nasty as the artifacts are mostly invisible in beauty pass. \nHowever when volume have high density and is illuminated with intense light the artifacts starts to appear in combine pass.\nArtifacts in denoising albedo pass can ruin seemingly good volume render.\n\nI wasn't able to replicate the artifacts in EEVEE.\n\nHere is how the bug looks with Quick smoke on default cube (scene in attached blend file):\n\nshadow pass\n\n\ndenoise albedo pass\n\n\nAnd here is an example how this bug might affect production scene:\n\nshadow pass\n\n\n\ndenoise albedo pass\n\n\n\nvolume direct\n\n\n\nvolume indirect\n\n\n\nbeauty (cropped)\n\n\n\n\nReproduce render artifacts:\n- open first attached blend file\n- hit `F12`\n- observe shadow pass\n- change `Render Layer` output in `Compositor` if needed\n\nRecreate vdb (optional):\n- open second attached blend file\n- (optional) in `Properties` -> `Physics Properties` -> `Fluid` -> `Cache` set path for cache\n- in the same panel hit `Bake All` and wait until bake is completed\n- navigate to cache directory and open `noise` folder, pick one file (better to pick one from the end where the smoke have more complicated shape)\n- import vdb to the first file and place it so the camera can see it\n\n[volume_bug_render.blend](volume_bug_render.blend)\n[volume_bug_sim.blend](volume_bug_sim.blend)", "GP transparent layer doesn't blend with background object \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 3.0.0 Alpha\n\nI have a foreground grease pencil object, and a background grease pencil object (yellow blob).\nThe foreground object contains a layer set to 0 opacity.\nWhere this layer overlaps the background object in the viewport, the background object is invisible.\nExpected behaviour: the visibility of the background grease pencil object isn't affected by transparent layers in a foreground object.\n\n\n\n\n\n[#89286.blend](T89286.blend)\n\nForeground object \"makes hole\" in background, doesn't blend. Rotate viewport camera, suddenly it blends correctly.\nWhen foreground location is set to 0,0,0 it blends correctly too.", "Invisible object in preview mode\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: AMD Radeon HD 8600/8700M ATI Technologies Inc. 4.5.14831 Core Profile Context 21.5.2 27.20.20903.8001\n\nBroken: version: 3.1.0\n\nWhen I enable Material Preview mode in Blender, the object becomes invisible regardless of the Render Engine. Also the object is invisible with Rendered mode enabled when I use the Render Engine - Eevee. \n\nI don't know how it happened but I can't repair Blender. I reset all settings to factory, reinstalled the application, but it did not help. Can you help me, please?\n\n\n\n\n\n\n\n\n\n", "Images loaded from palletized TGA files lack alpha channels\nOperating system: Windows 11 Home\nGraphics card: Intel ARC A750 8GB\n\nBroken: 3.6.1 and 4.0.0 2ed108cb84cf\nWorked: None tested\n\nWhen a palletized TGA file is loaded as an image, any transparent pixels come through as opaque black.\nThis does not occur for unpalletized TGA files; they come through transparent as expected.\n\nAn example of the same file in both palletized and unpalletized formats is attached. The files look identical in Aseprite.\n\n- UV Editing\n- Image -> Open...\n- Select palletized TGA file\n- Note that the area surrounding the vegetation at the bottom of the image is black: \n- Image -> Open...\n- Select unpalletized TGA file.\n- Note that the area surrounding the vegetation at the bottom of the image is transparent: \n\nThis can additionally be confirmed by looking at the histogram, with the unpalletized file having a small spike at 0% for alpha while the palletized file is flat until 100%.\n\n" ]

[ "Regression: Shader AOV doesn't seem to work for World on Cycles\nOperating system: Windows 11\nGraphics card: Geforce 1070Ti\n\nBroken: 3.1.2\nWorked: 2.93\n\nCaused by 1df3b51988\n\nShader AOV attached to World shader produces black image when rendered with Cycles (GPU and CPU). It's ok with Eevee.\n\nOpen default cube scene. Add AOV output to world, then add Shader AOV with same name in View Layer tab.\nOr just load my file and render using Eevee and then switch into Cycles. Cycles render will be completely black.\n[blender_env_shader_aov_test.blend](blender_env_shader_aov_test.blend)" ]

[ "VSE: audio animations lost in scene strips\nOperating system: Windows-10-10.0.18363-SP0 64 Bits\nGraphics card: AMD Radeon HD 6800 Series ATI Technologies Inc. 4.5.13399 Core Profile Context 15.201.1151.1008\n\nBroken: version: 3.6.0\n\nIn a video project that consists of a main scene that links in several sub-scenes, animations of the audio volume in the sub-scenes are gone in the main scene. The audio will remain at the volume that was set in frame 1 of the sub-scene.\n\n(1) Open the attched .blend file.\n(2) Go to Scene 1 and play it. You will hear that the volume of the audio strip will go from 3.000 to 0.000\n(3) Go to the Main scene and play it. You will hear that the audio remains at volume 3.000 (the initial volume) from beginning to end.\n\nNOTE: If you select Scene 1 in the Main scene, go to a certain frame number (e.g. 600) and type \"1\" in the volume box of the scene strip, the audio will switch to the right volume at that particular frame, but will stay at that volume for the remainder of the frames.\n\n", "Proposal for sequencer tool interactions \nThis task is intended to address design issues necessary for adding tool-system tools to the sequencer.\n\nCurrent state:\n\n- Immediately activating actions with keys operates on the selection.\n- The only tool that isn't selection (Blade), operates on the sequence strip directly under the cursor (not the selection).\n\nThe patch proposed to add the Slip tool, was also isolating the strip under the cursor, which I think is overly limiting since it means tools which behave this way can't operate on more than one item at a time. see: [D7063: VSE: Add slip tool](D7063)\n\n----\n\n## Proposes Solution\n\nTools should behave like the \"Tweak\" tool, unless there is a good reason not to *(The blade tool could be kept an exception - although cutting multiple strips vertically might be useful to support too)*.\n\n- A single click selects (with Shift-Click to toggle... etc).\n- A click-drag action runs the tool on the selected strips.\n\nThis means the Slip tool (in this case), would behave much the same as the tweak tool. Where you could quickly operate on many different sequence strips by clicking and dragging on them. With the ability to select multiple then click-drag on any one of them to operate on all strips.", "Mask Opacity keyframes not visible anywhere.\nUbuntu 14.04 64x\nnVidia GT620\n\nBroken: 2.78a\n\nKeyframes in LayerMask \"Opacity\" or \"Invert the mask B/W\" can be created. These keyframes exist in an existing Action (can be seen in Outliner - Actions and has datapath), but they don't appear in DopeSheet - DopeSheet nor DopeSheet - Mask. They can't be modified as the other keyframes. Nor changing its type (keyframe, interpolation, handle).\n\n- Open this file: [101-mask-keyframes.blend](101-mask-keyframes.blend)\n- Check the MaskLayer Opacity animation by going from 10 to 20 and back. Keyframes are visible only in the slider (as yellow slider).\n- The same with \"Invert the mask B/W\". Frame 30 and 40 have keyframes.\n\nEDIT: file uploaded.", "VSE 2.0: Performance\nThis task is to keep track of performance-related topics which needs to be addressed in order to achieve pleasant video editing experience. If any of the topics needs any deeper consideration and discussion it is to become a sub-task of this one.\n\nFlat assorted list of performance and memory-usage related topics:\n\n* - [x] Seek in video files is rather slow, feels measurably (2x) slower compared to seek in VLC with the same video file.\n\n* - [x] Proxy file generation is slower than realtime. Sometimes only time code (TC) is needed, so, probably, the solution is to allow generating TC without scaled-down video. Encoding of different resolution should be possible to do in separate threads.\n\n* - [ ] Color strip puts ever frame in cache. Simple to reproduce: crate new 4K project, add black color strip with the duration of entire 250 frames, playback. In theory, the final VSE frame cache should be able to re-use final strip frame.\n\n* - [x] Image strip caches every frame. Similar to the color case: adding and stretching image to occupy multiple frames will lead to way higher memory usage. This makes memory usage way higher in projects like story board, where edit consists of handful of image strips, and every image is stretched to occupy a frame range.\n\n* - [x] Image strips performs read-from-disk on every frame. Again, leads to worse performance for story-board like edit.\n\n* - [x] Moving image and cached video strips while being under the playhead is very slow. In 2.79 VSE the strip cache was attempted to be re-used as much as possible. Image strips should not be doing re-reads at all, video strips should reuse frames which are already in the cache.\n\n* - [ ] Operate in preview resolution. Operating and caching on original 4K images just to make the result be 320x240px in preview is very wasteful. Operation and cache should use the preview resolution. This will lower memory usage for all types of projects, and will make VSE faster when a lot of effects are used.\n\nFor the caching related topics see #80278 as well.", "Grid doesnt show\nOperating system:\nGraphics card:\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n", "VSE displaying png metadata/artifacts\nOperating system: Linux-5.11.2-1-MANJARO-x86_64-with-glibc2.33 64 Bits\nGraphics card: AMD VERDE (DRM 2.50.0, 5.11.2-1-MANJARO, LLVM 11.1.0) AMD 4.5 (Core Profile) Mesa 20.3.4\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.91, 2.92\n\nAfter loading a png created with inkscape, some hidden metadata of that png file (that I could not see using image viewers but only if I open the png in a text editor: line 4 \"tEXtSoftware www.inkscape.org\") www.inkscape.org is displayed below the image together with artifacts.\n\nexport something from inkscape to png\nin the VSE, load that png using Add Image/Sequence\ntext/artifacts appear below the image\n\npng:\n\n\nresult:\n\n\n", "VSE thumbnails improvements\n- [x] Check if reusing strip hash used by rendering job speeds up initialization with large number of strips. if it doesn't, fix it anyway. (b534806ecb) Note: it wasn't initialization that caused freezing, it was rather strip `anim` freeing, reusing ghash seems to have no impact on performance.\n- [x] To speed up loading, Check strategy to render \"guaranteed\" set of thumbnails. Especially if cache can be shared this could help a lot. (3a898db3638d) Note: better strategy seems to be other way around as far as performance goes.\n- [x] Make first (and last) thumbnail reflect actual handle position. (29b9187b327c) Note: Only first image is updated, though it's possilbe to update last one too. \n\n- Change cache so thumbnail (and possibly raw) types are keyed by path to file and disregard various context settings and Sequence pointer. Front-end should stay same. ([D12771](D12771))\n - There are rabbitholes like effect strips producing raw images and scene strips rendering at \"host\" scene resolution so simple string key is probably not enough. \n - Added complexity is, that all cache types per frame should be invalidated at once, otherwise prefetch doesn't work and issues like that.\n - Above points don't quite apply to thumbnails, but there are issues of this type in general. For example if we would implement downscaling for movies, first point will apply.\n\n", "Tooltips doesn't display when play animation\nOperating system: Linux-4.15.0-64-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\nBroken: version: 2.81 (sub 12)\nWorked: 2.8\n\nTooltips doesn't display when play animation, work otherwise\n\nDefault startup:\n \n* Play animation\n* Tooltips doesn't show \n\n", "The display will be damaged when the node editor in Fullscreen Area mode (ctrl alt spacebar)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon(TM) R7 Graphics ATI Technologies Inc. 4.5.14736 Core Profile Context 20.11.1 27.20.12033.2007\n\nBroken: 2.93 as well as version: 3.0.0 Beta\nWorked: Haven't tested before 2.93.\n\nThe display will be damaged when the node editor in Fullscreen Area mode (ctrl alt spacebar).\n\n* Switch workspace to Geometry Nodes\n* Add a new Geometry Nodes modifier/datablock\n* Ctrl+Alt+Space to go fullscreen (not Ctrl + Space)\n* Select or move a node\n\nFile:\n[bug.blend](bug.blend)\n\nGIF：\n", "Pinning Bones/Constraint Tabs in Properties Editor doesn't work\nBroken: 2.76 - master (1283c66)\nWorked: Unknown\n\nThe context pinning functionality in the Properties Editor doesn't work for Bones and Constraints. There may be a few others too that I haven't noticed yet, but these are the biggest ones.\n", "VSE View menu - missing hotkeys for sidebar and toolbar in Preview mode\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.09\n\nBroken: version: 3.0.0 Alpha\n\n\"Caused\" by 6aef124e7d\n\nThe VSE is missing the default hotkeys to open the sidebar and toolbar in Preview mode. T and N\n\nWhen you add them by yourself, then these hotkeys gets added in the Window section. Which can lead to override conflicts.\n\n\n- Open Blender\n- Switch to VSE\n- Watch the View menu in the upper Preview window\n\n\n\n\n\n\n\n\n", "Scrollbar of Clip Editor's Graph and Dopesheet View does not update\nOperating system: elementaryOS 0.4.1\nGraphics card: GTX 1080\n\nBroken: 2.80, 58bfe93db614\n\nWorked: apparently never...\n\nIf you change the endframe of your scene, the scrollbar in the Clip Editor's Graph and Dopesheet view does not reflect the changes.\n\nOpen startup .blend\nGo to Motion Tracking Layout.\nSet Endframe to 1000.\nNotice how the scrollbar down in the Graph View of the Clip Editor does not change its length.\nNow try the same with any other Animation Editor, for example in Animation workspace\nIf you set the endframe to anything that makes the scene longer, the scrollbar will update accordingly and become shorter, indicating that you do not see the full range of your scene.", "VSE: The blend and opacity settings that strips are created with does not match the default values they revert back to\nOperating system: Linux-5.4.0-80-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce 9800 GT/PCIe/SSE2 NVIDIA Corporation 3.3.0 NVIDIA 340.108\n\nBroken: version: 3.0.0 Alpha\nWorked: Unsure\n\nStrips are created by default with blend method \"Cross\" and blend opacity \"1\", however reverting to default values results in \"Replace\" and \"0\".\n\n- Open VSE\n- Add/create any strip\n- Under strip properties in sidebar, note that blend method is \"Cross\", and opacity value is \"1\"\n- Choose \"Reset to default value\" from context menu for both of these settings\n- Blend method will change to \"Replace\"\n- Opacity value will change to \"0\"", "VSE: Separate images does not handle resolution correctly\nOperating system: Windows 11 Home 64 bit\nGraphics card: NVidia GeForce 2060 RTX SUPER\n\nBroken: 3.0.1\n\n- Add image sequence to VSE where each image has different resolution\n- Separate images to discrete strips\n\nCheck resolution of separated strips - they are all same, but they should not be.\n\n\n\n\n\n\n\n", "BLF: Support GPU driven character stream\nFor overlays such as edit-mesh edge length, the screen placement and values are all done on CPU.\n\nThis also is an issue with geometry nodes needing debug attribute values displayed efficiently.\n\nThis also has other benefits such as better debug printing.\n\nThe BLF module should be modified to allow such needs:\n- [ ] Add a way to preload a set of characters into the glyph cache (avoid readback of drawn characters to CPU).\n- [ ] Design an API to record BLF calls in a threadsafe manner, in GLSL.\n- [ ] Allow to feed the recorded GPU BLF calls into the BLF drawing pipeline.\n- [ ] Convert the recorded BLF calls into Glyph stream.\n\n" ]

[ "New Text effect strip in sequence editor not working\nOperating system and graphics card\nWindows 10, Onboard Graphics\n\nBroken: blender-2.76-testbuild2-win64\n\nIn the sequence editor the new text effect strip is not rendered.\n\nBased on a (as simple as possible) attached .blend file with minimum amount of steps\n1. Switch to Video Editing\n2. Add -> Effect Strip -> Text\n3. Render sequence\nI would expect that the text is rendered somewhere. It also does not work when using an image and then alpha over the text strip.", "Text from Effect Strip is invisible in VSE Preview.\nubuntu 14.04 64 bits Nvidia Titan black\n\nBroken: master 26041f9\n\nIn release notes, new text strip is described with another design than in current master.\nSequencer \n\nText auto-alignement was redesigned.\nAs explained in release notes, disabling auto-center option was the condition to allow display of text in VSE preview.\nBut it was removed in new design and there is no way to see and to synchronize text while playback.\n\n\nAdd a text strip in Video Sequence Editor.\nYou should see a white text in Prevew. It stays blank as in attached.blend.\n[Text_Effect_Strip_Invisible.blend](Text_Effect_Strip_Invisible.blend)" ]

[ "Blender 2.93 crashes randomly\nHardware: MacBook Pro 13\", 8 GB 2016, as well MacMini 8GB 2020\nOperating system: MacOS\nGraphics card:\n\nBroken: version: 2.93.1\nWorked: ?\n\nI work with files exported from CAD-software packages imported into Blender showing technical installations: vacuum chambers and related technical infrastructure. I process the objects in the files by sorting them into collections representing subsystems (vacuum chambers, electrodes, magnets, ...) joining parts of larger systems, assigning materials from a material library file. I use hierarchies of files containing standard parts occurring at various places, materials and larger structures. In total, the files have a size of about 1 GB of data.\n\nWhen working with the files, Blender repeatedly and randomly crashes without obvious relation to specific actions.\n\nCannot provide this kind of information: random crashes.\n\nI am willing to give access to a git repo containing the 1 GB files\n\nSystem_info.txt:\n\n```lines\n> =====================================\n> = Blender 2.93.1 System Information =\n> =====================================\n> \n> \n> Blender:\n> =====================================\n> \n> version: 2.93.1, branch: master, commit date: 2021-06-22 05:57, hash: 1b8d33b18c2f, type: release\n> build date: 2021-06-22, 23:47:10\n> platform: Darwin\n> binary path: '/Applications/Blender.app/Contents/MacOS/Blender'\n> build cflags: -Wall -Werror=implicit-function-declaration -Werror=return-type -Wno-tautological-compare -Wno-unknown-pragmas -Wno-char-subscripts -Wstrict-prototypes -Wmissing-prototypes -Wunused-parameter -mmacosx-version-min=10.13 -Xclang -fopenmp -I'/Users/blender/git/blender-v293/blender.git/../lib/darwin/openmp/include' -std=gnu11 -msse -pipe -funsigned-char -fno-strict-aliasing -msse2 \n> build cxxflags: -Wall -Wno-tautological-compare -Wno-unknown-pragmas -Wno-char-subscripts -Wno-overloaded-virtual -Wno-sign-compare -Wno-invalid-offsetof -mmacosx-version-min=10.13 -ftemplate-depth=1024 -stdlib=libc++ -Xclang -fopenmp -I'/Users/blender/git/blender-v293/blender.git/../lib/darwin/openmp/include' -std=c++17 -msse -pipe -funsigned-char -fno-strict-aliasing -msse2 \n> build linkflags: -fexceptions -framework CoreServices -framework Foundation -framework IOKit -framework AppKit -framework Cocoa -framework Carbon -framework AudioUnit -framework AudioToolbox -framework CoreAudio -framework Metal -framework QuartzCore -framework ForceFeedback -liconv -Wl,-stack_size,0x100000 -Wl,-unexported_symbols_list,'/Users/blender/git/blender-v293/blender.git/source/creator/osx_locals.map' -stdlib=libc++\n> build system: CMake\n> \n> Python:\n> =====================================\n> \n> version: 3.9.2 (default, Mar 16 2021, 14:53:03) [Clang 10.0.0 (clang-1000.11.45.5)]\n> file system encoding: utf-8:surrogateescape\n> paths:\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/startup'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/modules'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/python/lib/python39.zip'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/python/lib/python3.9'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/python/lib/python3.9/lib-dynload'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/python/lib/python3.9/site-packages'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/freestyle/modules'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/addons/modules'\n> \t'/Users/wgeithner/Library/Application Support/Blender/2.93/scripts/addons/modules'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/addons'\n> \t'/Users/wgeithner/Library/Application Support/Blender/2.93/scripts/addons'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/addons_contrib'\n> \t'/Users/wgeithner/.local/lib/python3.9/site-packages'\n> \n> Python (External Binary):\n> =====================================\n> \n> binary path: '/Applications/Blender.app/Contents/Resources/2.93/python/bin/python3.9'\n> version: Python 3.9.2\n> \n> Directories:\n> =====================================\n> \n> scripts:\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts/modules'\n> \t'/Applications/Blender.app/Contents/Resources/2.93/scripts'\n> \t'/Users/wgeithner/Library/Application Support/Blender/2.93/scripts'\n> user scripts: '/Users/wgeithner/Library/Application Support/Blender/2.93/scripts'\n> pref scripts: None\n> datafiles: '/Users/wgeithner/Library/Application Support/Blender/2.93/datafiles'\n> config: '/Users/wgeithner/Library/Application Support/Blender/2.93/config'\n> scripts : '/Users/wgeithner/Library/Application Support/Blender/2.93/scripts'\n> autosave: '/Users/wgeithner/Library/Application Support/Blender/2.93/autosave'\n> tempdir: '/var/folders/qc/ph84tvmx4h19sn6wl8nnn8hw0000gn/T/blender_iWUimt/'\n> \n> FFmpeg:\n> =====================================\n> \n> avcodec: '58, 54, 100'\n> avdevice: '58, 8, 100'\n> avformat: '58, 29, 100'\n> avutil: '56, 31, 100'\n> swscale: ' 5, 5, 100'\n> \n> SDL:\n> =====================================\n> \n> Version: 2.0.12\n> Loading method: linked (WITH_SDL_DYNLOAD=OFF)\n> \n> Other Libraries:\n> =====================================\n> \n> OpenColorIO: 2, 0, 0\n> OpenImageIO: 2, 1, 15\n> OpenShadingLanguage: 1, 11, 10\n> OpenSubdiv: 0, 0, 0\n> OpenVDB: 8, 0, 0\n> Alembic: 1, 7, 16\n> USD: 0, 21, 2\n> \n> OpenGL:\n> =====================================\n> \n> renderer:\t'Intel(R) Iris(TM) Plus Graphics 640'\n> vendor:\t\t'Intel Inc.'\n> version:\t'4.1 INTEL-16.5.2'\n> extensions:\n> \tGL_APPLE_client_storage\n> \tGL_APPLE_container_object_shareable\n> \tGL_APPLE_flush_render\n> \tGL_APPLE_object_purgeable\n> \tGL_APPLE_rgb_422\n> \tGL_APPLE_row_bytes\n> \tGL_APPLE_texture_range\n> \tGL_ARB_ES2_compatibility\n> \tGL_ARB_blend_func_extended\n> \tGL_ARB_draw_buffers_blend\n> \tGL_ARB_draw_indirect\n> \tGL_ARB_explicit_attrib_location\n> \tGL_ARB_gpu_shader5\n> \tGL_ARB_gpu_shader_fp64\n> \tGL_ARB_instanced_arrays\n> \tGL_ARB_internalformat_query\n> \tGL_ARB_occlusion_query2\n> \tGL_ARB_sample_shading\n> \tGL_ARB_sampler_objects\n> \tGL_ARB_separate_shader_objects\n> \tGL_ARB_shader_bit_encoding\n> \tGL_ARB_shader_subroutine\n> \tGL_ARB_shading_language_include\n> \tGL_ARB_tessellation_shader\n> \tGL_ARB_texture_buffer_object_rgb32\n> \tGL_ARB_texture_cube_map_array\n> \tGL_ARB_texture_gather\n> \tGL_ARB_texture_query_lod\n> \tGL_ARB_texture_rgb10_a2ui\n> \tGL_ARB_texture_storage\n> \tGL_ARB_texture_swizzle\n> \tGL_ARB_timer_query\n> \tGL_ARB_transform_feedback2\n> \tGL_ARB_transform_feedback3\n> \tGL_ARB_vertex_attrib_64bit\n> \tGL_ARB_vertex_type_2_10_10_10_rev\n> \tGL_ARB_viewport_array\n> \tGL_ATI_texture_mirror_once\n> \tGL_EXT_debug_label\n> \tGL_EXT_debug_marker\n> \tGL_EXT_framebuffer_multisample_blit_scaled\n> \tGL_EXT_texture_compression_s3tc\n> \tGL_EXT_texture_filter_anisotropic\n> \tGL_EXT_texture_sRGB_decode\n> \tGL_NV_texture_barrier\n> \n> Implementation Dependent OpenGL Limits:\n> =====================================\n> \n> Maximum DrawElements Vertices:\t1048575\n> Maximum DrawElements Indices:\t150000\n> \n> GLSL:\n> Maximum Varying Floats:\t150000\n> Maximum Vertex Attributes:\t16\n> Maximum Vertex Uniform Components:\t4096\n> Maximum Fragment Uniform Components:\t4096\n> Maximum Vertex Image Units:\t16\n> Maximum Fragment Image Units:\t16\n> Maximum Pipeline Image Units:\t80\n> \n> Cycles:\n> =====================================\n> \n> \n> CPU device capabilities: SSE2 SSE3 SSE41 AVX AVX2\n> \n> Enabled add-ons:\n> =====================================\n> \n> io_anim_bvh (version: (1, 0, 0), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_anim_bvh/__init__.py)\n> io_curve_svg (version: UNKNOWN, path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_curve_svg/__init__.py)\n> io_mesh_ply (version: UNKNOWN, path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_mesh_ply/__init__.py)\n> io_mesh_stl (version: (1, 1, 3), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_mesh_stl/__init__.py)\n> io_mesh_uv_layout (version: (1, 1, 1), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_mesh_uv_layout/__init__.py)\n> io_scene_fbx (version: (4, 22, 0), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_scene_fbx/__init__.py)\n> io_scene_gltf2 (version: (1, 6, 16), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_scene_gltf2/__init__.py)\n> io_scene_obj (version: (3, 8, 1), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_scene_obj/__init__.py)\n> io_scene_x3d (version: (2, 2, 5), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/io_scene_x3d/__init__.py)\n> cycles (version: UNKNOWN, path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/cycles/__init__.py)\n> GTry (version: (1, 3, 7), path: /Users/wgeithner/Library/Application Support/Blender/2.93/scripts/addons/GTry/__init__.py)\n> add_curve_sapling (version: (0, 3, 4), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/add_curve_sapling/__init__.py)\n> materials_utils (version: (2, 2, 0), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/materials_utils/__init__.py)\n> materials_library_vx (version: (0, 6, 0), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/materials_library_vx/__init__.py)\n> render_auto_tile_size (version: (3, 1, 3), path: /Applications/Blender.app/Contents/Resources/2.93/scripts/addons/render_auto_tile_size.py)\n> ```\n\n\n[BlenderCrash_20210730-01](BlenderCrash_20210730-01)\n[BlenderCrash_20210801-01](BlenderCrash_20210801-01)", "macOS todo list\nNOTE: this is a direct dump of Install-OS\n\n- [ ] Users have to manually change macOS defaults:\n * Map middle mouse to regular \"Mouse button 2\"\n * Make the Function keys use modifiers (so Blender sees these)\n * Change the Shift+NumPad options to not use Shift (change spaces)\n * The latter two could become part of a new standard macOS compliant map in the future?\n- [ ] ~~Finder aliases or~~(1f223b9a1f afb1a64ccb) hidden files are not being recognized by Blender\n- [ ] macOS swallows some key events, like e.g. emulated numpad - on Norvegian keyboard (see #38273).\n- [ ] Blender prevents macOS from sleep #38702", "Regression: Crash when sculpting shape keys in 3.6.2 release\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\nWorked: 3.5 did not crash, I'm told.\n\n**Description**\nA user in a Discord community shared a file which was crashing for them while in Sculpt mode. I could reproduce the crash, so I simplified the file and found repro steps.\n<video src=\"attachment\" title=\"2023-09-14 00-35-34.mp4\" controls></video>\n\nAs you can see, it takes a few seconds of aggressively dragging the shape key value to get the crash to occur.\nAnother way to get the crash is to just make sculpt brush strokes. I found the shape key slider dragging easier, but I could get it to crash with sculpt strokes too, which was the original complaint from the user.\nMaybe worth noting that, as far as I can tell, if the shape key is not this extreme in its effect, then it seems even harder or impossible to produce the crash.\n\n- File: [Vessa_clean5.blend](attachment)\n- Make sure you're in sculpt mode (I don't think it can crash otherwise)\n- Drag the shape key value around for several seconds, or make a few dozen sculpt strokes\n- Should crash eventually.\n\n", "Sculpting laggy and weak from some directions.\nOperating system: Windows 10\nGraphics card: Nvidia GTX 970\n\nBroken: 3.6.1\nWorked:Never(2.80+)\n\nWhen i try to sculpt from from below, using radius unit: scene, sculptings get laggy and the stroke is weak.\n\nI'm trying to sculpt stones for a wall in place.\nThe sculpting works fine if i first move the stone i'm working on to the scene origin. This adds a lot of work as i want to for example put bevels on all objects first, then go on with cracks and other features!\n\nTo reproduce:\nOpen file, sculpt from top and bottom\n\nFile is default sculpt template with sphere moved up by 5 and using scene radius.\n\n", "Crash creating \"position\" attribute with the wrong type while in Edit mode\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.41\n\nBroken: 3.5.0 and 3.6.0 Alpha\nWorked:\n\nWhile in Edit mode, attempting to create an attribute called `\"position\"` that has type not set to `\"FLOAT_VECTOR\"` will crash Blender.\n\nI couldn't get any other attributes to crash when added while in Edit mode.\nCreating the `\"position\"` attribute with the correct type, but incorrect domain does not cause a crash either.\n\nI can't imagine a use case for users wanting to do this, but it came up while I was determining the extent of 107228 and isn't really the same issue.\n\n1. Select a mesh object\n1. Enter Edit Mode\n1. Either:\n 1. Run `bpy.context.object.data.attributes.new(\"position\", \"FLOAT\", \"POINT\")` from the Python Console\n 1. Create a new attribute called \"position\" using the `bpy.ops.geometry.attribute_add` operator and set the `Data Type` to anything other than `Vector`: ", "Blender Opens Window then Closes/crashes\nOperating system: Windows 10\nGraphics card: AMD FirePro W2100\n\nBroken: 3.5.1 2023-04-24 18:11 Hash:e1ccd9d4a123 Branch: Blender -v3.5-release\nWorked: first version used on this computer\n\n-First attempted to run blender, found error by running the debug, error:\n\n_EXCEPTION_ACCESS_VIOLATION_\n\n- Did specific driver updates for AMD FirePro W2100 and Intel CPU\n\n- Then ran dell command update\n\n- New error appeared after driver updates and running blender:\n\n_A graphics card and driver with support for OpenGL 3.3 or higher is required.\nInstalling the latest driver for your graphics card might resolve the issue._\n\n- Then downloaded GPU-Z to figure out OpenGL version\n- Open GL version was 1.1 when monitor DisplayPort was plugged into the motherboard\n- Open GL version was 4.5 when monitor DisplayPort was plugged into GPU\n\n- Now blender window would appear but closes immediately\n- running blender_debug_gpu_glitchworkaround.cmd in the Blender file worked to make Blender run but is NOT the overall fix\n- updated python\n- did windows updates\n\nunsure\n\n", "GPencil: Selecting textured brushes will crash Blender\nOperating system: Linux-5.8.0-7642-generic-x86_64-with-glibc2.32 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.39\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.92.0\n\nWhen appending the textured brushes from the [Blender Cloud](5ccfe64353b85e279cf72acd) into a file, Blender will crash when trying to preview the brushes in the tool settings.\n\n\n - Open Blender 2.93 and select the \"2D Animation\" template.\n - Go to *File*>*Append* and browse to the `pepeland_GP_brush_pack_V1.blend` file. Go into the *Brush* folder and append the brushes `GBrush_001` through `GBrush_010`.\n # Click the brush icon in the tool settings. Blender crashes.\n\n\n\n", "Blender crashes as soon as an image texture is plugged in the material BSDF node\nOperating system: Windows 10\nGraphics card: NVIDIA GeForce GTX 1080 Ti\nSee attached log for more detail\n\nBroken: 3.5.1, but also 4.0.0 Alpha, did not try intermediate versions\nWorked: ??? The earliest available build, 2.93.18, also causes the same crash\n\nThe second any kind of image texture is plugged in a material, Blender crashes without any warning. Occurs when creating a new image texture, when plugging an existing file, or even when trying to plug a UV grid for unwraping purposes\n\n**Similar tickets**\n#76193\n#76377\n\n1. Start Blender.\n2. Create new default file.\n3. Select the default cube.\n4. Switch to Shading workspace.\n5. Add Image Texture node.\n6. Create new image or select an existing one.\n7. Plug Color output to the Base Color input of the Principled BSDF node.\n8. Watch Blender shut down.", "Particle Edit mode for non hair particles - point select mode crash\nOperating system: Linux-5.6.0-050600-lowlatency-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.64\n\nBroken: version: 2.90 (sub 3)\n\nBlender crashes when entering into the Particule Edit mode\n\n- Delete all objects from the default scene\n- Create a plane\n- Add a particle system\n- Set end frame to 50\n- Bake the particle system\n- Switch to the Particle Edit mode\n- Select Point select mode\n\n- > Crash\n\n[blender.crash.txt](blender.crash.txt)\n\n\n", "Crash Guide during render complete 2 December 9, 2022\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.4.0\n\nCrash Guide during render 2\n\nSame crash guide during render before and today.\nSee youtube: ZeCA2p5wISY\n\nHistory previous.\nT100691\n\n\n\n\n[Blender-2022-12-09-161546.ips](Blender-2022-12-09-161546.ips)\n\n[Horizontal Video.zip](Horizontal_Video.zip)", "Metal Viewport \n**Status:** `?` Initial stages of development\n\n---\n\n**Description**\nBringing Metal support to the Blender viewport for optimal performance on macOS. This will be implemented as a new GPU Backend within Blender's GPU Module.\n\n**GHOST**\n\n* - [x] Encapsulate use of OpenGL within GHOST_ContextCGL\n* - [x] Add support for Metal Backend to GHOST_ContextCGL\n* - [ ] EDR (Extended Dynamic Range) support for GHOST swapchain\n\n**EEVEE Next - Workbench Next - Overlay Next**\n\nInitial bring-up\n\n* - [ ] Add support for 2D Texture Atomics via Buffer + Implement metal path\n* - [ ] Add support for virtual frame buffers (No attachments, but specified dimensions, currently allocating temporary texture for enablement). \n* - [ ] EEVEE + GPU Unit tests 100% pastorate\n* - [ ] EEVEE irradiance cache baking - Metal support\n* - [ ] EEVEE Next shadows - Metal support\n\nOptimisation + Later:\n\n* - [ ] Optimize compute kernel workgroup sizes for all passes\n* - [ ] Investigate kernels wherein threadgroup size requirement exceeds optimal sizing heuristic. Aim to avoid using maxTotalThreadsPerThreadgroup.\n* - [ ] Ensure resource usage is optimal (compression etc)\n\n**GPU Module METAL Backend**\n * - [x] MTL_backend\n * - [x] MTL_batch\n * - [x] -> Metal indirect draw support\n * - [x] MTL_capabilities\n * - [x] MTL_command_buffer\n * - [x] MTL_context\n * - [x] MTL_debug\n * - [x] MTL_drawlist\n * - [x] MTL_element\n * - [x] MTL_extensions (via backend)\n * - [x] MTL_framebuffer\n * - [x] MTL_immediate\n * - [x] MTL_memory_manager\n * - [x] MTL_query\n * - [x] MTL_shader\n * - [x] MTL_shader_interface\n * - [x] MTL_state\n * - [x] MTL_storage_buf\n * - [x] MTL_texture\n * - [x] MTL_uniformbuffer\n * - [x] MTL_vertex_buffer\n * - [x] Compute support for Metal\n\nMore items to be added as necessary.\n\n**GPU Feature coverage**\n\nA number of features available in Blender require API functionality which is either not directly available in the Metal API, such as geometry shaders, or requires a higher degree of explicitness from the high-level renderer. These tasks represent non-invasive changes required to enable full Metal support:\n\n * - [x] GPU - Geometry shader workflow alternative\n * - [x] GPU - Multi-layered rendering support in Vertex Shader\n * - [x] GPU - Shading language translation layer from GLSL to MSL (Metal Shading Language).\n * - [x] GPU - Alternatives for unsupported GPU Primitive types: `GPU_PRIM_TRI_FAN`/`GPU_PRIM_LINE_LOOP` (mtlprimitivetype?language=objc)\n * - [x] GPU - Shader uniform struct alignment validation for Metal (`GPU_uniformbuf_create_from_list`)\n * - [x] GPU - Shader barycentric support in Metal.\n * - [x] GPU - default depth format support checks (Apple Silicon does not support `GPU_DEPTH24_STENCIL8`)\n * - [x] GPU - Lightweight GLSL shader requirements for MSL compatibility.\n * - [ ] GPU - Extended Dynamic Range (EDR) support for OCIO shaders and GPUViewport.\n * - [x] DRW/GPU/COM - Support for Realtime compositor (T99210)\n * - [ ] #104025 Offline Metal shader compilation - Startup compilation time and system caching.\n * - [ ] Metal/DRW - EEVEE/Material PSO (Pipeline State Object) parallel compilation.\n\n> Shader libraries compile (Front-end compilation) in parallel while displaying the default EEVEE material. For Metal (and likely Vulkan too), PSO compilation (backend/low-level compilation) happens on demand, depending on the current render pipeline state in combination with a given shader, and this can take time, introducing stalls and stutters, such as when an object is first selected and the select-engine/depth-only permutation is generated for that particular mesh.\n>\n> Ideally we would want to avoid these one-off PSO stalls, which for OpenGL, are generally hidden behind driver processing. Propose a few options:\n> - Option to add a \"default/fallback shader to a **GPUShader** such that if it needs a fresh PSO, and a default shader/fallback is provided, it can instead bind that shader for the draw. In essence, continue using the EEVEE default material, if PSO compilation is needed, not only when the shader library is compiling.\n> - Shader PSO cache warming: Generate the render pipeline state configurations/permutations up-front to allow background parallel PSO compilation. This could be done by performing a dry-run of Blender and packaging a data-file containing the serialized input structure for the PSO compilation function, and performing background warming after start-up, using this data. This avoids the need to manually specify the configs.\n> - Prepare certain configurations up-front in background by performing dummy renders in different states, and skipping draws which trigger a PSO compilation. This allows the dynamic render state to be set, without performing rendering.\n\n\n * - [x] Metal viewport/Metal Cycles shader compiler load distribution manager. \n\n> As both backends use Metal, there is only a fixed pool of shader compilation threads available. A mechanism should be in place to ensure compilation is not stalled, which could cause the viewport to become unresponsive for synchronous shader compilations. e.g. PSO backend compilations.\n\n```\n\n```\n * - [x] Python -- Use of OpenGL within Python API - Ensure safe fallback (T103863)\n * - [x] Cycles - Live viewport display -- Currently uses OpenGL directly\n\n---\n\n**Relevant links**:\n* `-` Metal-Shading-Language-Specification.pdf\n* `-` metal/\n* `-` Metal-Feature-Set-Tables.pdf", "Cycles GPU crash when switching to Rendered 3DViewport --MacOS 13.1--\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.5.0 Alpha\nWorked: 3.5.0 Alpha (works fine)\n\nInstant crash of blender when switching to rendered 3DView.\n\nCompile latest 3.5 version as of 2022-12-18 20:06, hash: `50c7eb14f4`\nopen empty scene provided below and switch to rendered 3DView.\n[gpu.blend](gpu.blend)\n\n\nxcode 14.1\nlatest brew updates\n(cycles-cache & build-darwin folders deleted before compiling )\n\nPS: When switching to CPU there is no crash. Only GPU !", "Circle Select Tool Todo\n## Circle Select Tool Todo\n * - [x] {icon circle color=red} Add Radius control to the tool settings, to set the radius of the select brush\n * - [x] {icon circle color=red} There should be a brush circle size preview cursor always visible, just like a paint tool. Currently the radius is only visible while painting your selection.\n * - [ ] {icon circle color=yellow} Currently, Circle Select only selects object origins. Add enum option to select normally too.\n * - [ ] {icon circle color=yellow} Add options to create new selections, add to selections or subtract from selections.\n\n * - [x] 3D view.\n * - [x] UV editor.\n * - [ ] Action editor.\n * - [ ] Annotations.\n * - [ ] Tracking.\n * - [ ] Node.\n * - [ ] Graph.\n\n\n\n\n", "Mesh related crash in add-on\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 780 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.27\n\nBroken: version: 3.0.0 Alpha\n\nUsing an the operation connect spread in an addon is causing an EXCEPTION_ACCESS_VIOLATION. \n\n[blender.crash.txt](blender.crash.txt)\n\n```\nStack trace:\nblender.exe :0x00007FF7A109EC60 blender::deg::deg_graph_flush_updates\nblender.exe :0x00007FF7A1084AA0 DEG_evaluate_on_refresh\nblender.exe :0x00007FF79C983790 scene_graph_update_tagged\nblender.exe :0x00007FF79CB61510 wm_event_do_notifiers\nblender.exe :0x00007FF79CB4B550 WM_main\nblender.exe :0x00007FF79C7E2ED0 main\nblender.exe :0x00007FF7A16421A8 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFEEB937020 BaseThreadInitThunk\nntdll.dll :0x00007FFEECA62630 RtlUserThreadStart\n```\n\nDownload [Forgotten Tools ](forgotten_tools_0_5_0.zip) use Connect spread on 2+ edges go to operation popup menu increase the cuts.\n\n[2021-06-20_00-26-23.mp4](2021-06-20_00-26-23.mp4)\n\n", "\"Slice String\" field input shows as error, but works (lack of string fields)\nOperating system: macOS-13.3.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Max 1.2\n\nBroken: version: 3.5.1\nBroken: version: 4.0.0 Alpha\nWorked: /\n\nA field can be connected to \"Slice String\" node \"Position\" input, and while it shows as red error, it does work. In the attached files, the connection between \"Duplicate Elements\" and \"Slice String\" is shown in red, despite node tree working as expected (tested in Blender 3.5 + current 4.0 daily-build).\n\nAlso note the lack of fields between the \"Slice String\" and string \"Equal\" nodes.\n\nSee attached files.\n\n---\n\nIn the example file, each symbol in a string is an instruction for creating an object at a specified position.\nThe instructions implemented here are \"T(op)\", \"C(enter)\" and \"B(ottom)\".\nI therefore take the input string and create an object for each symbol in the string, then use the geometry index to find the corresponding string symbol to do modifications on the geometry.\n\nFor this to work, \"Slice String\" needs to accept the field input (from \"Duplicate Elements\" field \"Duplicate Index\") so each geometry duplicate gets its own symbol from the input string. This works, but the connection is shown as error in red.\n\nIdeally \"Slice String\" would accept a field for \"Position\" and \"Length\", so arbitrary parts of the string can be extracted for each geometry duplicate. The output of \"Slice String\" should also be a field.\nI'm not sure if there's a way to handle this without an error in the geometry nodes.\n\nOverall, it seems Blender is lacking string-fields, because the \"Equal\" node also defaults to fields for all types except strings. Yet, fields also seem to work here, too?\n\nNot needed here, but the \"String Length\" node also doesn't seem to explicitly support fields.\n\n" ]

[ "Grease Pencil Fill Tool crashes with Metal GPU Backend\nOperating system: macOS Monteray 12.6.2\nModel: Macbook Pro 15,1 6-core i9\nGraphics card: Radeon Pro 560X 4 GB\n\nBroken: 3.5.0 Beta (2023-02-19 and 2023-02-21)\nWorked: Works with OpenGL backend, Metal GPU Backend not available in 3.4\n\nFilling a grease pencil shape immediately quits blender and produces a crash report.\n\n1. Create a new 2D animation project.\n2. Go to prefs and make sure you have Metal selected for GPU Backend. (restart if you had to change it)\n2. Draw four overlapping edges of a rectangle\n4. Switch to fill tool.\n5. Select a fill material (optional)\n6. Click inside the rectangle and click again to fill it.\n\nBlender quits with the attached crash report.\nvideo at: ScreenRecord-BlenderGPCrash.mov (I had crash report turned off so the included CR is from a separate session - but it does pop up. It's too large to attach here.)\n\n\n\n" ]

[ "Blender crashes when calling ArgumentParser.parse_args() if sys.argv is anything other than [\"path/to/blender.exe\"]\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 980 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.18\n\nBroken: version: 3.5.1\n\nBlender crashes when calling `ArgumentParser.parse_args()` if `sys.argv` is anything other than `[\"path/to/blender.exe\"]`\nI discovered this bug while trying to run a script using the `exec()` function. Any script that uses argparse makes blender crash.\n\nYou can run this code\n```Py\nimport argparse\nimport sys\nimport bpy\n\nsys.argv = [bpy.app.binary_path]\nsys.argv.append(\"--some_arg\") # Comment this line and everything works fine\n\nparser = argparse.ArgumentParser()\nargs = parser.parse_args()\n```\n\n", "Crash When Adding Hair Subdivision upon Viewer node\nOperating system: Windows-10-10.0.17134-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.11\n\nBroken: version: 3.4.0 Alpha\nWorked: Never\n\nCaused by: c55d38f00b\n\n\n\nAdding hair subdivision will cause crash immediately when you are using viewer node to visualize field values in the viewport.\nNote no crash will occur if the viewer node is inactive.\n\n[Hair Crash.crash.txt](Hair_Crash.crash.txt)\n\n1. select default cube\n2. shift+A to add empty hair\n3. in the node tree add viewer node.\n4. Crash if adding subdivision\n[Hair Crash.blend](Hair_Crash.blend)", "Connect failed in certain case\nWin8 64bit | GTX680 \n\nBroken: 2.71.6 a0b3604\n\nWhen select two vertices and press J to connect, it fails to add vertices to edges in between.\n\nPlease see the attached case for example. Thanks for checking.\n\n\n\n[vertex_connect.blend](vertex_connect.blend)", "Shading or Lighting Discrepancy issue\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.97\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional)\n\n\nSo I’ve been having this problem since working in 2.8 and I don’t have a clue what could be causing this issue but I went and stripped down a project as simple as possible trying to isolate the root cause. It seems its a mesh that has an emissive material and when I disconnect the node or hide the mesh the problem disappears. Weird thing is that the brighter version is when hiding mesh number 2. I am attaching the file.\n\nYou will see how the tone changes. This never happened to me after so many years using Blender until now already in a few projects in 2.8. \n\n- Open attached file.\n- Render once using slot 1.\n- Hide mesh 2 or disconnect the emissive node.\n- Render again using slot 2.\n# Toggle with J\n\nNotice the different lighting between each result.\n\n[Shading Issue - Simplied.blend](Shading_Issue_-_Simplied.blend)", "Blender just crashes for no reason, and the crash report is empty. (Segmentation fault)\nOperating system: Arch Linux x86_64\nGraphics card: AMD Radeon RX 6900 XT\n\nBroken: (example: 3.6.2, ffe93138751e, master, 2023-08-31, as found on the splash screen)\nWorked: (No idea when this error started happening.)\n\nBlender does not start up. When I start it via the command line, I get this error:\n```\nWriting: /tmp/blender.crash.txt\nSegmentation fault (core dumped)\n```\n\n`/tmp/blender.crash.txt`:\n```\n# Blender 3.6.2, Commit date: 2023-08-14 14:05, Hash ffe93138751e\n\n# backtrace\n\n# Python backtrace\n```\n\nTrying to open Blender.\n\n", "GPencil: Blender crash when file open\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.4.0 Beta\n\nAs soon as I open attached file Blender crashes. Attached also crash log file.\n\n- Open attached file\n- Change radius (say 0.6)\n- Save file\n- reopen it (Crash)\n[#102930.blend](T102930.blend)", "EEVEE crash while rendering animation, cycles works.\nOperating system: Win10\nGraphics card: GTX1070MAX-Q\nCPU : i7-6700HQ\nRAM : 32GB DDR4\n\nBroken: 2.90\nWorked: -\n\nBlender crashes w/o any errors when I started to rendering animation in Eevee. \nIt happens on 2 laptops and it always happen in 154 or 155 frame while rendering.\nIn attachment you can download the project which gets this error. \nWhen I use cycles rendering it works perfect.\n\n[#80740.blend](T80740.blend)\n\n- Open file\n- Render frames 154 and 155\n", "Weird VSE crash\nOperating system: Win11\nGraphics card: RX580\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen) blender 4.0.0 alpha\nWorked: (newest version of Blender that worked as expected) no clue\n\nBlender crashes on opening file. Was trying to make an alternative caption that gets around the Multiply color issue #112267 but came across a new bug.\n\nBased on the default startup or an attached .blend file (as simple as possible).\nOpen the file I've attached\n\n", "Crash on Mac when starting with `--debug-gpu` flag\nOperating system: Mac OS Ventura 13.1\nGraphics card: Apple M1 Max\n[system-info.txt](system-info.txt)\n\nBroken: 3.4.1\nWorked: 3.3.2\n\nBlender won't start on a Mac when starting with `--debug-gpu` flag.\nThe terminal output is this:\n\n```\n/Applications/Blender\\ 3.4.1.app/Contents/MacOS/Blender --debug-gpu\nRead prefs: /Users/vollstock/Library/Application Support/Blender/3.4/config/userpref.blend\nWARN (gpu.debug): source/blender/gpu/opengl/gl_debug.cc:165 init_gl_callbacks: Failed to hook OpenGL debug callback. Use fallback debug layer.\nzsh: segmentation fault /Applications/Blender\\ 3.4.1.app/Contents/MacOS/Blender --debug-gpu\n```\n\n\nStart with the flag\n\n```\n/Applications/Blender.app/Contents/MacOS/Blender --debug-gpu\n```", "Ctest: Crash in cycles_image_colorspace_cpu\nOperating system: Windows\nGraphics card: Irrelevant\n\nBroken: current master (c32c6a5f85) \nWorked: Unknown\n\n\nDuring testing of [D10212: Cmake/deps: Update OSL to 1.11.10.0](D10212) I noticed a single test failing... once... by repeating the test over and over I got it to misbehave once every 5-45 minutes , so... that's not gonna be super fun to track down.\n\nI attached 3 stack traces which all point to the same location, a null check there just mask the issue, personally I'd take a crash over a random bad frame being output. \n\n\n`ctest -C Release -VV -R cycles_image_colorspace_cpu --repeat-until-fail 9999 -O Fail.log`\n\n\n[image_non_color.crash.txt](image_non_color.crash.txt)\n\n[image_log_osl.crash.txt](image_log_osl.crash.txt)\n\n[image_alpha_blend.crash.txt](image_alpha_blend.crash.txt)\n\nI mentioned this on chat earlier today, but tracking ticket is probably nicer cc: @brecht ", "crash when i try to make instances real for nodes\nhi , in attached files all info about the crash , its when i try to make instances real for nodes modifier addon scifi pro + flex\nso can i fix it ?\n\n", "Sculpt: crashes when using the Edit Face Set in Delete Geometry mode on an object with 0 faces\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Xe Graphics Intel 4.5.0 - Build 30.0.100.9864\n\nBroken: version: 3.3.0 Alpha\n\nBlender crashes when trying to delete non existing geometry with the Edit Face Set tool\n\n[blender.crash.txt](blender.crash.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n\n1) select the default cube, go to edit mode and delete all vertices\n2) switch to the sculpting tab\n3) select the \"edit face set\" tool and switch the mode to \"Delete Geometry\"\n4) left click anywhere in the viewport and Blender will crash\n\nThere are many things going wrong in sculptmode with faceless meshes.\n\nSee also reports:\n- #86940 (Crash with certain file. Sculpt mode. Vertex color fill + edge.)\n\nAnd (possible but partially abandoned) solutions:\n- [D10851: Fix #86940: sculpt automasking facesets option crashes on meshes without faces](D10851)\n- [D10966: Prevent changing to sculptmode on faceless mesh objects](D10966)\n\n**Stack trace**\n\nblender.exe :0x00007FF603BDA400 sculpt_face_set_edit_invoke\nblender.exe :0x00007FF6033688F0 wm_operator_invoke\nblender.exe :0x00007FF603366280 wm_handler_operator_call\nblender.exe :0x00007FF603367CA0 wm_handlers_do_keymap_with_keymap_handler\nblender.exe :0x00007FF603367440 wm_handlers_do_intern\nblender.exe :0x00007FF6033668A0 wm_handlers_do\nblender.exe :0x00007FF603363430 wm_event_do_handlers\nblender.exe :0x00007FF60334D530 WM_main\nblender.exe :0x00007FF602431210 main\nblender.exe :0x00007FF608464500 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFF3D887020 BaseThreadInitThunk\nntdll.dll :0x00007FFF3EAE2630 RtlUserThreadStart", "Blender crash on pose library N-panel\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 850M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.5.1\n\nEvery time I click on invert button (which invert filtering) inside pose library N-panel blender crashes under this condition that your active asset library set to all.\n\n1. First be sure animation pose library add-on is enable.\n2. Create new scene.\n3. Create an armature.\n4. Go to pose mode.\n5. Click on Animation tab on side bar(N-panel).\n6. Change active pose library from \"Current File\" to \"All\".\n7. Click on double arrow icon which indicate invert filtering one or two times.\n8. Blender will crash.\n\nI also tested on blender 3.6 Beta , It still exist but in v 3.6 beta you should click on view port shading options after step 6 to perform crash.\n\nAnyway I upload crash log.\nThanks a lot.", "Crash Guide during render complete 2 December 9, 2022\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.4.0\n\nCrash Guide during render 2\n\nSame crash guide during render before and today.\nSee youtube: ZeCA2p5wISY\n\nHistory previous.\nT100691\n\n\n\n\n[Blender-2022-12-09-161546.ips](Blender-2022-12-09-161546.ips)\n\n[Horizontal Video.zip](Horizontal_Video.zip)", "Crash when deleting ViewLayer while rendering\nOperating system: Linux-5.4.0-66-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.92.0\n\nA crash will happen when deleting a ViewLayer while rendering stills or animations, for any render_display_type\n\n- In the default scene, add a new ViewLayer\n- Set the renderer to Cycles so that you can see the render progress. In Eevee you could set up a little animation to get the same problem.\n- Start a normal render\n- While Blender is still rendering, delete the new ViewLayer that you just added and that was being in use when the render process was called\n- A crash happens with a Segfault, even after some time.\n\nThis problem looks to me as just an unhandled context change, because I noticed that if you delete the current scene or the current camera used while rendering animations, those context changes are appropriately handled and the render process is effectively stopped without a crash.\n" ]

[ "2.80 Beta Eevee Crash\nOperating system: Windows 8.1\nGraphics card: Nvidia GeForce GT 820M\n\nBroken: blender-2.80.0-git.33993c056a5-windows64 2018 - 12 -13\nWorked: -\n\nWhen either trying to switch to LookDev mode or Rendered mode with Eevee, Blender immediately crashes.\n\nOn default startup file, switch to LookDev or Rendered mode while Eevee is enabled." ]

[ "MacOS, Metal: Banding artefacts in viewport when using curves in color management\nOperating system: macOS-13.2.1-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon Pro Vega 64 1.2\nScreen: iMac 27 inch Retina 30bit\n\nBroken: version: 3.5.0\nWorked: same version with openGL viewport API\n\nIn the \"wanderer\" demo file you get white lines in the bright parts of the scene. They are visible in rendered view as well as in the final render (but only inside blender). \nWhen \"Use Curves\" in \"Color Management\" is turned off, they dissapear. \nWhen switching back to openGL api everything looks ok.\n\nOpen the demo file wanderer wanderer.blend\n\n", "Preview Texture Blend\nOperating system: Arch Linux\nGraphics card: GTX 1060 6GB\n\nBroken:\n(example: 2.79b release)\n(example: 2.80, edbf15d3c044, blender2.8, 2018-11-28, as found on the splash screen)\nWorked: (optional)\n\n\nTexture gradient is not displayed correctly. Texture Blend.\nWhen you configure the texture, the preview displays the texture view.\nThe preview is more white than black.\n\n\n\n\nIf the texture to use as brush, you can see that the boundary particle moves wrong.\n\nIf in 2.8 open the file created in 2.7.\nThe texture is displayed differently.\nBut it works the same way as in 2.7.\n\n\n\nRun the attached files, see how the particles behave.\n[Blend.zip](Blend.zip)\n", "Missing visual representation of multiple selected color channels (`Blender Light` theme)\nOperating system: win10 64\nGraphics card: 1050ti\n\nBroken: (example: 2.91, default light theme)\nWorked: (2.79b)\n\nThis is how it was in 2.7x when selecting multiple RGB channels:\n\n In recent versions it misses visual representation of how many channels are selected until changing values to see if slider has been updated (left - nonselected, right - all channels are selected):\n\n\n\n - Open default scene\n - Go to material tab\n - Left mouse click any color slot\n - Left mouse click first color channel and hold it\n # While holding, slide cursor below to cover all channels\n\n\n", "Vertex Paint ➔ Paint menu ➔ Smooth Vertex Colors could be improved\nOperating system: Windows 10 Pro 64-bit\nGraphics card: NVIDIA GeForce RTX 2080 Ti\n\nTested: Blender 2.81 Alpha, September 18, 14:29:45 - b962aca8003d\nWorked: never\n\nVertex Paint ➔ Paint menu ➔ Smooth Vertex Colors tool does not smooth other than direct loop neighbors.\n\n1. Divide the default cube a few times (apply if subdivision is done with modifier).\n2. Enter Vertex Paint mode.\n3. Paint an area with black.\n4. Apply Paint menu ➔ Smooth Vertex Colors.\n5. The vertex paint is not smoothed other than direct loop neighbors.\n\n\nHere is some clarification as the descriptions above are not very clear (I work with vertex colors very regularly and still came out confused by reading this post).\n\n- The Smooth Vertex Colors command will have visible results *if* it is run on vertex colors that have been applied per-face (by first selecting the desired faces in Edit mode, and then filling them with colors in Vertex color mode with the face filter enabled). Think of these as \"harsh\" vertex colors. When such vcols exist, then the smooth operation will indeed smooth out the vcol data so that the harsh separation between painted faces will be smoothed out, blending it with the surrounding faces.\n- But if the command is run on vcol data that has been painted on using the paint brush or on vcol generated by the \"dirty vertex colors\" command then nothing will happen as the vcols will already be point-based.\n\nTherefore there is no bug here, just some very confusing naming of the command and a lack of clarity when the command is run but nothing happens. For any user working with painted-on vertex colors, it will indeed seem like nothing is happening at all since one would expect some sort of global smoothing to be applied.\n\nPart of the confusion is also coming from the fact that this command sits just below \"Dirty Vertex Colors\" in the menu, which generates smooth, per-vertex vcol data. Meaning that for a user exploring vcol mode by trying one feature after the next, \"smooth vertex colors\" will almost always feel like it is doing nothing at all (unless the user first applied some vcol using face masking, which is extremely unlikely).\n\nIf anything the tooltip and/or the menu entry could be changed for clarity :\n\n- The menu item name could be renamed from \"Smooth vertex colors\" to \"Smooth harsh per-face vertex colors\"\n- The tooltip could be changed from \"Smooth colors across vertices\" to \"Smooth colors across vertices. Only operates on harsh, per-face vertex colors\".\n\nAnd then of course it would be great to have a new feature altogether to blur vcols globally, similarly to what the blur brush does.", "Color picker samples merged (display output) colors on a certain object regardless of \"Sample Merged\" setting\nOperating system: Windows 10 v1909\nGraphics card: NVIDIA GeForce GTX 1660 super\n\nThe bug occured on the latest Steam release of Blender (2.90.1). I also tried to open the project file with the said bug in 2.83 LTS steam beta branch but it happened there too.\n\nWhen trying to texture paint a certain project it ignores the \"Sample Merged\" setting and always picks the display output color, while other objects don't share this problem. \nIt happens regardless of the material or texture I'm editing, it seems to be tied to a single object. For example, If I create a cube and assign the material which the bugged object uses color picking works fine with it, and if I assign any other material to the bugged object it still samples display output color regardless of the settings.\n\nI'm sadly not sure how to reproduce it. Though I appended the bugged object to a fresh file: [color_picker_thing.blend](color_picker_thing.blend)\nIt's system info:\n[system-info.txt](system-info.txt)\nI separated it so you wouldn't have to dig through the mess that is my main project file :D. It retained it's weird inabillity to pick colors though.\n\nI couldn't find a solution to this when I tried to google for it, so my only assumption left here is it being a bug. I'm sorry in advance if something I said isn't clear. If needed, I can provide a whole project or try and provide more specific information if what I wrote here isn't enough as soon as possible.\nCheers!", "Meshes are invisible if any objects are set to display 'In Front'\nOperating system: Darwin-18.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: Intel(R) HD Graphics 6000 Intel Inc. 4.1 INTEL-12.10.14\n\nBroken: version: 2.83.0\nWorked: 2.82\n\nIf any objects are set to display as 'in front' in a project, all meshes will display only as faint outlines. In edit mode, meshes display similar to how they would if they were set to display as wireframe.\n\nAdd a new mesh, curve, or armature to the scene.\nObject properties -> Viewport display -> Enable 'In Front'\n\n", "Sculpt Mode Border Hide Tool Todo\n## Sculpt Mode Border Hide Tool Todo\n * - [x] {icon circle color=red} When this tool is selected, it currently shows the wrong tool settings which affect the brushes, not this tool. These should not be visible when this tool is active.\n * - [ ] {icon circle color=yellow} We should add a tool setting to add/remove the mask as you drag. Currently, the only way to remove from border hide, is very hidden: You drag with RMB. This could become more consistent and more discoverable if we added the same Add/Remove buttons as with other Sculpt tools", "Bevel Bug\nblender-2.80-07765499abee-win64 \n\n**cylinder :**\non the cylinder with a lot of segments, bevel doesn't merge in one point. Create nice see n-gon(normal we must see the only verts in one point)\n\n\n**Cube :**\nThe middle loop has deformed when we bevel sharp corner for small mesh. On the large mesh, we cant see this deform .\n\n\n[Cylinder_Bevel_Bug.blend](Cylinder_Bevel_Bug.blend)\n\n[Cube_Bevel_Symmetry_Bug.blend](Cube_Bevel_Symmetry_Bug.blend)\n ", "Dashed geometry nodes field link too hard to see on macOS\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.5.0 Alpha\n\nBreak line (I don't know word) see picture\n\nOnly for Mac. Do not use Windows.\n\n\n\n\nSee my picture on my iMac there.\n\n\nBrowser to show picture.\n0q6NTJs\n\n\n\nTheme on Nodes \"Dashed Lines Opacity\"\n", "Spreadsheet not reading evaluated `bevel_weight_edge` attribute\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 4.0.0 Alpha\n\nthe `bevel_weight_edge` isn't being displayed with the correct value in the spreadsheet, both for GN geometry[^1] and original[^2] geometry [not made in GN]; both when the bevel weight is set in GN[^1], and when it's set in edit mode[^2].\nalthough, when using the *viewer node* the correct value is displayed[^1].\n\nThis could be happeding with other attribute, but I haven't tested.\n\nset the bevel weight (any way you'd like) and look in the spreadsheet.\n\n[^1]: <video src=\"attachment\" title=\"2023-07-08 22-47-22.mp4\" controls></video>. \n[^2]:<video src=\"attachment\" title=\"2023-07-08 23-02-16.mp4\" controls></video>\n\n", "Dissolving edge doesn't work as expected\nOperating system: Windows 10\nGraphics card: RTX 2070 SUPER\n\nBroken: 3.2.1\nWorked: not sure\n\n\nDissolving edges doesn't work as expected.\n\n\n\n\n\n[untitled.blend](untitled.blend)", "Random Per Island working incorrectly with sharp edges and volumes\nOperating system: Linux-5.0.0-36-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: TITAN X (Pascal)/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.50\n\nBroken: version: 2.82 (sub 4)\nWorked: (optional)\n\nRandom Per Island not working correctly with sharp edges or volumes.\n\n\nSharp Edges:\n\n- Create Cube\n- Add Bevel Modifier with Harden Normals\n- Set Smooth Shading\n- Set Auto Smooth\n- Random Per Island is affected incorrectly by sharp edges, even after applying Bevel modifier\n\nVolumes:\n\n- Create cube (or array of objects)\n- Create volume shader\n- Assign Random Per Island to volume shader\n- No random colors\n\n[random_per_island.blend](random_per_island.blend)", "Proposal: New Bone Color Presets\n\n\n| | Edge Shapes Deselected | Edge Shapes Selected | Polygon Shapes Deselected| Polygon Shapes Selected | Theme UI |\n|---|---|---|---|---|---|\n| Before |  |  |  |  |  |\n| After |  |  |  |  |  |\n\nMost rigs are forced to use Custom Colors very often, because not enough of the preset colors feel nice to use. This is a shame, because Custom Colors cannot be customized on a user-level, via the theme settings. I think if the Preset Colors were better, Custom Colors wouldn't even be necessary; The 20 slots that we have are plenty enough to cover the entire color wheel with distinct enough colors. At that point, even if you wanted to add a custom color, you'd be hard pressed to find a color that isn't very similar to one of the presets.\n\n### Issues with current bone color presets:\n- Slots 16-20 are blank, immediately not useful.\n- Most colors are too dark, and do not contrast well against Blender's default viewport background color, which is a dark gray in the case of all shipped themes.\n\n### The colors in this proposal:\n- Base colors are much brighter, with a minimum value of 0.75.\n- Colors are spread out sort of evenly across the color wheel.\n- Hues are fairly similar to the old ones, so the color language of existing rigs that used presets shouldn't change much.\n- Proposal is for **ALL themes**. Previously, lighter and darker themes had slightly different presets, but they all had the issues described above. The Modo theme had the additional issue of no visual distinction between selected and active bones.\n\nThe file I used to create the screenshots, with the script to place bones based on their theme color, and to output the code for the [PR](112639): [bone_colors.blend](attachment)\n\n### Questions: \n- Are the colors... good?\n- Should we re-order the presets by hue?\n - Pro: Helps avoid cases where you assign 2 colors that were far from each other in the preset list, but are actually quite similar, so you don't get the contrast you were hoping for.\n - Con: Would make existing rigs using the old presets look very different, potentially with less distinct colors (although still much brighter!).\n- Anti-Versioning: Julian Eisel tells me `do_versions_theme()` can be used to force reset theme colors, but I may need some help with that part.", "Intersecting Meshes are not marked as overlapping anymore when selected\nOperating system: Linux-5.4.0-48-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.90.1\nWorked: 2.83.x\n\nIn blender 2.83.x Solid View intersecting objects had, if selected, the otherwise hidden parts shown in a more muted color when overlapping. with 2.90.x this is no longer the case. \nI hope this did not get removed intentional but is a missing feature.\n\nDuplicate the default Cube, move it a bit so that is still intersects with the first cube and confirm movement. \nWatch the outline in Solid View be the same color and thickness all around the mesh. \nin 2.83 while moving the white outline turned light grey when intersecting and the selected object outline turned a darker orange. This is very helpful for positioning elements.\n\n", "\"Regression\" : Using normal or bump maps in cycles produces flat washed-out shading (caused by terminator softening)\nOperating system: Windows 10 x64\nGraphics card: GeForce GTX 1660 Dual OC\n\nBroken: 2.81, 2.82, 2.83 from 2020-03-06 22:46\nWorked: 2.80\n\nCaused by c3be14e151\n\nUsing normal or bump maps in cycles produces flat washed-out results comparing to EEVEE.\nThe bug also appears in the latest 2.83 build, it affects both Diffuse and Principled Shader but not Glossy.\nNormal Map and Bump nodes seem to produce identical results with the 2.80 version, so the problem lies with the Shader node itself.\n\n\n[bug_report.blend](bug_report.blend)\n1. Open file\n2. Switch to Rendered viewport shading" ]

[ "Edge Seam not displayed\nOperating system: Mac OS 10.14.3\nGraphics card: AMD Vega64\n\nBroken: 2.8, 08ab09cf04d9, 2019-02-05 15:52\nWorked: 2.8, a3bf8da8758f, 2019-02-01 22:37\n\nUse the startup cube:\nIn Edit mode when marking an edge as seam, the edge does not change the color. \nWhen the object is unwrapped the seam is recognized tough.\n" ]

[ "Animation Editors: Modifier keys cancel out dragging handles \nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.4.0 Alpha\n\nIn recent releases db4313610cab was implemented and the animation editors are still not adapted to this behaviour. This fix allows the use to press any keyboard key, like `Shift` or `X`, etc, before the drag threshold of the current action is reached, and it will cause the operator to execute.\n\nFor example: Click and hold on a move gizmo in object mode. Either drag 3 pixels to start the operation and then press `X` and hold `Shift` to slowly move along the x axis.\nOR immediately press `X` and hold `Shift` without reaching the drag threshold of 3 pixels to already start the operator.\n\nBefore this fix, any pressed keys would cancel the click & drag and prevent the operator from running.\nThis is still the case with the graph and dopesheet editors.\n\n- Create a simple animation to the default cube\n- click and drag on the keyframes or handles in the animation editors\n- Now click and hold them, hold `Shift`, and then continue to drag. Nothing happens.", "Move To Line Begin and Reset View operators in the console editor conflict when pressing Home key\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1650 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.95\n\nBroken: version: 3.4.0 Alpha\n\n\nIn the python console editor, both the Move To Line Begin and the Reset View operators have the same shortcut \"Home\", so it is not possible to invoke the Move To Line Begin operator.", "Sequencer: Can't toggle Channels sidebar when VSE view type is \"Sequencer & Preview\"\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: AMD Radeon (TM) R9 380 Series ATI Technologies Inc. 4.5.13596 Core Profile Context 20.10.35.02 27.20.1034.6\n\nversion: 3.2.0\n\nAll in the title.\n\n1) Be in Video Sequencer editor\n2) Switch view type to Sequencer & Preview\n3) Try to toggle Channels sidebar : either with the little arrow (that doesn't show up), or by going to View > Channels (it doesn't toggle the sidebar, its checkbox doesn't get checked)\n\n", "Menu label does not update after \"Remove Shortcut\" from context menu\nOperating system: Darwin-19.5.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 455 OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.9.15\n\nBroken: version: all\n\nAfter \"Remove Shortcut\", the menu label is not updated properly.\nJust for a moment, but it's noticeable.\n\n**Only in the case when the \"Remove Shortcut\" menu item (mouse cursor actually)\nis located above the deleted item or above separator.**\n\n", "Entering names in Object pointer parameter and confirming with the mouse in the right part of the button will clear user input immediately again\nWin 8.1x64, 3x gtx580\n\n2.79a\n\nIn a node's Object parameter, on entering the text manually, after pressing enter to confirm, if the mouse happened to be located where normally the \"X\" button to clear the parameter is, the field gets cleared.\nAlso applies to other object pointer fields.\n\nWhen entering names in Node Editor, (at least for me) it is common to move the mouse slightly sideways, so that the cursor is not over the field being typed in. As a result it happens to be over said clear button quite often. Given that it is currently impossible to undo a single name change #54580 , it turns into either loosing the name and having to search/type it again, or more often - by automatically pressing Undo - losing a whole row of changed fields.\n\n[rename.blend](rename.blend)\n - Click the Object parameter to type manually,\n - Move the mouse to the right side of the field, where the \"X\" was located before starting to type,\n - Enter a new object and press enter,\n - The Object parameter field is empty.", "NLA Strip influence slider doesn't respond to autokey correctly.\nOperating system: Win 10\nGraphics card: GeForce GTX 1070\n\n\nBroken: 2.82 Release\n\n\nThe Influence property of the NLA's strip, has some strange behavior and something that might be a bug.\nThis was reported before in #63865, which was supposedly resolved, but the problem still exists\n\nWhen changing the value of the influence slider with the autokey turned on, it jumps back to its previous value and does not update to the new one. After repeating it 2 or 3 times it applies the value. When applying the value numerically it works fine, it seems to be a problem with the update of the slider.\nMoreover, when turning influence on it always adds a keyframe to the first frame of the action (where the action is on the timeline ). It sets a keyframe with the value 0, instead of the current value.\n\n- Open attached file\n- Select one of the Action strips.\n- In the Sidebar, go to {nav Strip > Active Strip} and turn on `Animated Influence`. (Note that a keyframe is added to the first frame of the action).\n- Go to a new frame with the Autokey turned on, change the value of the influence slider.\n- Go to a new frame - the value should jump back to the previous one\n\nI've made a short video which describes the errors\nwatch?v=Sh5MwYCH9T4\n\nSample file:\n\n[Influence_bugreport.blend](Influence_bugreport.blend)", "Can't undo the keyframe change in edit mode\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.1.2\n\nWhile in edit mode, when you hit Undo in the graph editor after moving a keyframe, the graph doesn't change.\n\n- Animate anything using a couple of keyframes.\n- Change to edit mode in viewport\n- Switch timeline to the graph editor (not necessary, you can keep timeline as it is)\n- Select a keyframe and move it\n# Hit undo\n\nResult: The graph editor stays the same\n\nExpected: The move you just made in the graph editor should be the only thing being undone!\n\nTest file with first four steps included\n[#97700.blend](T97700.blend)", "3D View - removing Hotkeys for Toolbar/Sidebar throws empty keymap 3D View Generic error\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 419.35\n\n\nBroken: version: 2.80 (sub 72)\nWorked: (optional)\n\n\n\n3D View - removing Hotkeys for Toolbar/Sidebar in the View menu throws empty keymap 3D View Generic error in the console.\n\n\nOpen Blender. In the View menu of the 3D View remove the hotkeys for Toolbar and Sidebar. Open the Console ...\n\n\n\n", "Native addons lose the \"Report a bug\" button when using a custom Config folder\nOperating system: Windows\n\nBroken: 4.0\nWorked: -\n\nReport by @Todor-CGDive (moved from 104635).\nPre-installed addons have a \"Report a bug\" button. However, if you create a custom Config folder, that button changes to \"Remove\". I am not sure if that's by design but it looks suspicious.\n\n\n- On Windows, go to directory `%USERPROFILE%\\AppData\\Roaming\\Blender Foundation\\Blender\\4.0\\` and copy de `config` folder\n- Past it in the your blender installation directory `/4.0` folder\n\n", "Non-keyboard shortcuts allowed to be set via context menu, but cannot be unset without going to user-preferences\nBroken: 2.79 or latest 2.80 (fa59346c13).\n\nAlthough I can easily (and often accidentally) assign non-keyboard shortcuts to menu items, I cannot easily undo it.\n\nFrom the viewport: View > Frame All [RMB] > Change Shortcut.\n\nAlthough the UI says \"Press a Key\", if you mouse click it (e.g., LBM) will register as the new shortcut.\nNow if you try to change the shortcut again it says it is a Non-Keyboard shortcut and cannot be changed.\n\nTooltip: *\"Only keyboard shortcuts can be changed that way (...)\"*", "Package manager design\nAs the GSoC period comes to a close, I thought it would be a good idea to see where the UI stands.\nI already got some design feedback from my [prior progress video ](watch?v=b_iBVNmeBdM&lc=z13hfndjmzqmtpe0p234vroyex3fgbovp):\n\n> * The Install/Documentation/Uninstall/Disable buttons should have the same height as a regular button\n> * The position of these buttons should stay on the top-right when the add-on is expanded (to avoid jumping position of buttons)\n> * Update to XX could perhaps just be called \"Update\". Number versions (and conventions) vary between add-ons, making the button different widths and making a bit of visual noise. We could display the version number available when expanded, and its changes (are we supporting changelog? That'd be awesome!).\n> * The Enable button does the same as the checkbox in the regular add-ons list? Maybe we should just use the same familiar UI for that?\n> * Refresh repositories should be triggered when adding a new repo maybe?\n\n\n\nSo, the current design of the package manager appears thusly:\n\n\n\nAs it is now, I already have a few concerns:\n\n* Not a lot of space for repository names, and no way for the user to resize that area (not counting resizing the window) \n * Perhaps the sidebar should be put in a region, so as to at least be user-resizeable?\n* Enable/disable checkbox make names of installed packages misalign with names of uninstalled packages.\n * This could be avoided by only showing uninstalled packages in \"available\", so that uninstalled and installed packages are never displayed in the same list. However, this has a side effect of removing any way to search **all** packages at once. \n* Not a lot of distinction between installed and uninstalled packages.\n * The checkbox only appears on installed packages, but isn't consistent; it can appear unchecked or checked.\n * The button text (\"Install\" vs \"Uninstall\") looks rather similar.\n\nAny other observations and suggestions/recommendations?", "Link-drag-search unsupported in texture node editor\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.3.1\nWorked: 3.0.0, since there is no such feature.\n\n\"Auto-completion/search\" is not working as intended with Texture Node Editor. For some reason, only reroutes are being displayed. [2022-11-23 22-42-21_1.mp4](2022-11-23_22-42-21_1.mp4)\n\nOpen Texture Node Editor -> Drag from any socket.\n\nNote: sorry for the incorrect spelling of word \"reroute\"\n", "Small improvements to 3D View Transform Manipulators\nAs per feedback from the Spring team, we would like to make some smaller improvements to the 3D View transform manipulators:\n\n\n - Increase the treshold for hiding the planar transform widgets, so that users don't accidentally drag the plane when they meant to drag the arrow\n - Increase arrow click threshold. Currently it's very unforgiving. Even if you click outside the arrow by one pixel, it doesn't register as an arrow click. Give it a few more pixels all the way around.\n - Separate the view-aligned rotation handle from the X, Y and Z rotation handles, so that users don't accidentally click the wrong one. We can make the view-aligned handle offset, like it was in 2.79\n", "File browser buttons disappear on undo in another window\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.79\n\nBroken: version: 2.92.0 Alpha\n- All versions since the new file browser was introduced are affected by this\nWorked: Never\n\nI've tried to save file with CTRL+S and by accident pressed my LMB on a viewport. Because of that mistake, I pasted object in my scene, instead of pasting file name in \"Blender File View\" window. So I tapped CTRL+Z (while still being active in a viewport) and whole panel with buttons \"Save file\", \"Cancel\" in \"Blender File View\" window disappeared. I've tested this on a completely new empty blender file. Pressing on a viewport didn't recreate that effect (probably, because nothing was there). But pressing on outliner or timeline have the same issue.\n\n- Open Blender\n- CTRL+S\n- Press LMB on an Outliner\n- CTRL+Z\n\n[UI bug.blend](UI_bug.blend)", "Menu Editor for Quick Favorites\nCurrently users can only manage menu items in the quick-favorites menu by adding/removing items.\n\nThis task proposes to: \n\n- Create a generic menu editor in the user preferences.\n\n - Add/remove menu items\n - Re-ordering menu items.\n - Operators options should be able to be changed or unset - as we have with the keymap editor.\n\n**Implementation**\n\n- This could share UI code with the keymap editor for adjusting operator settings, since internally they are very similar.\n- We may support multiple kinds of menus, besides \"Quick Favorites\", internally there is support for this. Even if this isn't exposed to users initially.\n\n```\nThis would be useful for declaring user defined toolbars for example.\n```\n\n**User Interface**\nHere's how the UI could be organized: \n\n{[F8282863](Screenshot_2020-01-17_at_11.08.58.png), size=full}" ]

[ "NLA Sidebar tab doesn’t refresh on keyframe creation and there is no UI change to indicate action selection in the track list\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: Quadro RTX 3000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.88\n\nBroken: version: 3.1.2\nWorked: Never\n\nNLA Sidebar tab doesn’t refresh on keyframe creation when first using the NLA and there is no UI change to indicate action selection in the track list so a user doesn't know they need to select it to get to the settings they need. Once they do select it, they have no idea of the current selected state, as there is no UI change to show it is selected/active or not.\n\n\n Creating from scratch workflow 1 When using the NLA for the first time\n - Open Blender and change the timeline view to the NLA editor\n - Hit N or Menu- view-sidebar to show an empty sidebar area\n - Select the default cube (if you don't have it create a cube) and keyframe the cube. at frame 1 and frame 20\n - Notice that keyframes are created and selected in the timeline after creation yet the NLA UI side panel remains blank. No update has happened to make the action selected or active.\n - Also notice that the ACTION is always colored and looks like it is selected but the Theme NLA setting \"Source List Text Highlight\" doesn't seem to effect the Text on the Action display.\n - Next click on the object name (Cube or the action name (Cube action) in the left track list of the NLA, Notice that once you do it refreshes and the sidebar panel now shows settings for the current action. Notice that there NO UI change to indicate that after clicking on the Action name in the track list that it has been selected.\n #Once you have clicked on it and caused the sidebar to refresh, it never goes away, it doesn't go back to being blank, it always shows the settings.\n\nSuggested fix. The sidebar should refresh to with creation of first key or action, it should never be blank when an object is selected and it has keyframes inserted, starting a new action. The theme should indicate the selected state of the default action track.\n\nThere are other issues around the UI but they are more design while this is a UI bug.\n\n\n\n\n\n" ]

[ "Assigning shortcuts: Lots of keyboard keys unknown to Blender (international keyboard layout)\nOperating system: Ubuntu 18.04\nGraphics card: GeForce GTX 1060\n\nBroken: 2.81a, 2.79, 2.81\nWorked: (optional)\n\n\nI want to use '<' key (grless in windows) as shortcut, like I have in Windows but it doesn't work with \"unsupported key: unknown\" message. The same behavior append with '^', '$', ':', '!' keys (maybe some more, didn't test) . All these keys work in all other apps.\nI have a french azerty keyboard and tried multiple keyboard configurations without success.\n\nInstall Ubuntu 18.04, install Blender 2.81a, plug a french azerty keyboard, try to change a shortcut with the mentioned keys.\n\nAdding some more info, the keys works in Blender console and text editor, it's only the shortcut assignation that doesn't work.\n\n\nI'm open to make some more testing if needed.", "Keep Alt Key not work in the F-Curve properties\nwin10 64bit, gtx1060\n\nBroken: 2.79a 8928d99270f\nWorked: (optional)\n\nKeep alt key for multiple objects not work in F-curve property ( N panel )\n\nOpen my .blend file, when I select multiple curves and keep alt key to twice the Interpolation or Easing option, it's just worked in one cure,not for all selected, if this is bug, please fix, it very useful for lots curves, else I must one by one to change those, thanks.\n\n\n\n[bug.blend](bug.blend)", "ALT and ALT GR are not the same in Spanish Spain keyboard distribution\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 560 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 388.13\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional)\n\n\nIn english keyboard distribution, ALT and ALT GR is the same key, for example, ALT+M to Merge is the same for the ALT GR+M, so much comfortable to use with the left hand, but if I change they Windows key distribution to Spanish (Espanol Espana) ALT GR + M won't work, only works ALT + M, and If I change into Blender preferences the keymap for ALT + M to ALT GR + M, it works, but it shows as CTRL + ALT + M and left ALT + M won't work again only works ALT GR + M.\n\nI'm working with Spanish key combination in windows 10\n\nUsing Spanish Spain Windows keyboard distribution, select 2 vertices in edit mode, press ALT GR + M, it won't work, go to Preferences, Keymap, search for ALT M Key-Binding and change Merge under Mesh to ALT GR + M.\n\nIt will work, but ALT + M won't work again and in keymap will show as CTRL + ALT + M.\n\n", "\"Whole Character\" keying set not working properly with autokey enabled\nOperating system: macOS-11.5.2-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575X OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.6.20\n\nBroken: version: 2.93.4\n\nUsing the keying set \"Whole character\" with autokey enabled creates a keyframe only for the selected bone and not for all bones and properties.\n\n- Open attached file or create a new armature with more than one bone, add a few bone properties to each bone.\n- Enable autokey and set \"Whole Character\" as the keying set.\n- Move one bone and you can see that only the transform values are keyed.\n- If you enable \"Only active keying set\" in the autokey options and move the bone, a keyframe for the custom properties of the selected bones, at least, will be created, but other bones and properties will be unaffected.\n[bugreport.blend](bugreport.blend)\n", "The X or Delete button does not work in the outliner (Blender File)\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 2.93.6\n\n[The X or Delete button does not work in the outliner (Blender File)]\n\n- Open default startup file\n- {nav Outliner >Display Mode > Blender File}\n- In outliner {nav Brush > RMB > Delete > RMB > Assign Shortcut > X}\n- Attempt to delete any brush with hotkey (will not delete brush)\n- Change hotkey (eg {key Shift X}) and now attempt to delete brush with hotkey\n\nwhen you press X items in:\n| Broken | Brushes, Cameras, Images, Lights, Line Styles, Materials, Meshes, Palettes, Workspaces, Worlds |\n| -- | -- |\n| worked | Collections, Objects, Scenes |\n\n\n\n\n[Delete _X_ Button.blend.py](Delete__X__Button.blend.py)\n\n{[F12271213](Delete__X__Button.blend)}]\n\n", "Fallback Fonts\nBlender is currently shipping with four font files: variable width and fixed-pitch in small and (larger) \"international\" versions. It is in the plans to [reduce these to just two and use the big international fonts always ](D6854). This would mean that a large proportion of the glyphs of the all the major languages would always be available to all users, regardless of what language is chosen.\n\nAfter we do so it would be great to add two more: fallballback fonts.\n\nThe two fallbacks (variable width and fixed-pitch) would be small font files (less than 50k each) that would be loaded in RAM on startup and would be used to load glyphs that are not found in the main fonts. \n\nThese are needed because we have always allowed users to replace the main fonts (Edit / Preferences / Interface / Text Rendering) with any of their choosing. This might be because they prefer the appearance, or because the font handles their language better. But the danger is that the user will do so and replace the main font with one that does not contain characters needed for Blender to work correctly.\n\nFor example, [Google Noto fonts ](https:*www.google.com/get/noto/) can be found to work very well with almost every known language. However, many of these fonts do not even include the Basic Latin set (A-Z, etc). A user selecting such a font would surely file a bug report which would require someone to examine the font and then report back to the user that the font file is incomplete for the purpose. Not only would this waste time but it is also something the user cannot fix themselves. With a*built-in// **non-replaceable** set of fallback fonts Blender would continue to work correctly no matter what fonts the user selelected.\n\nFollowing are fallback versions of our two fonts, each made from our current versions (and also with Cyrillic corrections in D6960). Each of these contain only the Basic Latin set U+0000 - U+007E plus a minimal number of symbols that we use in the interface or might possibly use.\n\n[fallback.ttf](fallback.ttf)\n\n[fallbackmono.ttf](fallbackmono.ttf)\n\n\n", "Latest Blender 4.0 not opening\nOperating system: Linux-5.15.0-83-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: GeForce GT 630/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.157\n\nBroken: version: blender-4.0.0-alpha+main.b5c89822ac40-linux.x86_64-release\nWorked: 3.6.3 Release Candidate\n\nAfter double-clicking on the \"blender\" starter-folder, it shows what you can see on the picture, before canceling the session\n\nJust try to start 4.0. I guess it's my drivers and noone can confirm, right?\n\n", "Splitting two modifier keys between two input devices doesn't work\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.5.0\n\n\nPressing modifier keys from two keyboard device doesn't correctly register in blender.\n\n\n- Have two physical \"Keyboard\" device that both have modifier keys\n- Press Ctrl on keyboard 1 and Alt on keyboard 2\n- Blender doesn't recognize that both Ctrl and Alt are pressed.\n\n-----\n\nOriginal description:\n\nHi - I have modifier keys (shift, alt ctrl) assigned on mouse and keyboard and razer keypad. \nI use alt to select edge/face loops without issue. \nIf I press shift+alt on keyboard then I can select as many loops as I like. Same if I press shift+alt both on mouse. \nHowever, if I use modifier keys combined between deviced (shift on keyboard and alt assigned on mouse, or vice versa) then it lets me add only one additional loop to selection only. \nI initially thought it was accessory issue, but Its true with my default keyboard, razer keypad, logitech G600 and Razer Naga mouse, I suspect the issue is within blender. \n\nIn summary combining modifier keys between keyboard/mouse doesn't seem to work, atleast with loop selections. Any advice? Thanks\n", "Crash (memory use after free) on freeing RigidBody world\nReproducible with current master, probably there since forever.\n\n# Steps\nWith an ASAN build:\n\n- Open `rigid.anim.blend`\n- Close Blender\n\n[rigid.anim.blend](rigid.anim.blend) \n\n(Case found while working on #86379 (When using \"Append\" rigid body objects + empty constraint, none of its constraints work.))\n", "VSE Linked external scene with audio with input Sequencer only works if the scene is NOT also linked with input Camera\nOperating system: Linux-4.9.0-9-amd64-x86_64-with-debian-9.9 64 Bits\nGraphics card: Mesa DRI Intel(R) HD Graphics 520 (Skylake GT2) Intel Open Source Technology Center 4.5 (Core Profile) Mesa 13.0.6\n\nBroken: version: 2.81 (sub 3)\nWorked: Haven't yet tried other versions.\n\nVSE: Linking external scene with audio with input Sequencer only works if the scene is NOT also linked with input Camera.\n\n1. Create file A with Scene-A and add audio to it on the VSE.\n2. Create file B\n3. From the VSE in file B, link the Scene-A from file A. Set it's input to \"Sequencer\". On file B, playing the strip will correctly play the sound from Scene-A.\n3. Duplicate (or re-instance) Scene-A on the VSE sequencer. If the Input is still \"Sequencer\", both clips will play the audio, as expected.\n4. Switch one of the Scene-A instances to use the Camera as input. \n\nExpected behaviour: \n\nWe should be able to both see the Camera strip and hear the Sequence strip for Scene-A\n\nActual behaviour:\n\nFor all the strips of Scene-A that are set to Input \"Sequencer\", the audio will stop playing.\n\nNote: I've determined that it is not that the audio crashes: other clips with sound on the sequencer keep playing normally.\n\nNote: If I delete the strip with Input Camera and leave only the input Sequencer strips, the problem will not go away, unless I save the file like this and \"revert\" to it.\n\nI cannot find a suitable work-around. In the video that I'm producing I need to edit local sound effects on the shot sequence, and incorporate it on the larger edit.\nI cannot, therefore, both see and hear my sequence on the host sequencer.\n\n\nIf a workaround is possible (that is not reconstruct and re-import the sounds on the original scene) I'd be a happy man :-)\nThanks!\n\n\n", "use-after-free crash after setting scene's render.engine inside custom render engine context\nOperating system: Ubuntu 20.04.2LTS\n\nBroken: 2.93.0 Beta, branch: master, commit date: 2021-05-19 10:14, hash: a294670bacb1, type: release\nAlso Broken: 0efb627bbd3ac3e53afb95591ae4e244bd1f8af4\n\nBlender crashes when setting some scene's render.engine property inside `__init__` or the `view_update` context of a custom render engine to a different one than the custom engine.\n\n[T88404_asan_report.txt](T88404_asan_report.txt)\n**Code to reproduce the error**\n```\nimport bpy\nimport bgl\n\nclass CustomRenderEngine(bpy.types.RenderEngine):\n bl_idname = \"CUSTOM\"\n bl_label = \"Custom\"\n bl_use_preview = True\n \n def __init__(self):\n return\n scene = bpy.data.scenes.new(\"foo\")\n scene.render.engine = \"CYCLES\"\n bpy.data.scenes.remove(scene)\n\n def __del__(self):\n pass\n\n def view_update(self, context, depsgraph):\n scene = bpy.data.scenes.new(\"foo\")\n scene.render.engine = \"CYCLES\"\n bpy.data.scenes.remove(scene)\n\n def view_draw(self, context, depsgraph):\n pass\n\ndef get_panels():\n exclude_panels = {\n 'VIEWLAYER_PT_filter',\n 'VIEWLAYER_PT_layer_passes',\n }\n\n panels = []\n for panel in bpy.types.Panel.__subclasses__():\n if hasattr(panel, 'COMPAT_ENGINES') and 'BLENDER_RENDER' in panel.COMPAT_ENGINES:\n if panel.__name__ not in exclude_panels:\n panels.append(panel)\n\n return panels\n\n\ndef register():\n bpy.utils.register_class(CustomRenderEngine)\n\n for panel in get_panels():\n panel.COMPAT_ENGINES.add('CUSTOM')\n\n\ndef unregister():\n bpy.utils.unregister_class(CustomRenderEngine)\n\n for panel in get_panels():\n if 'CUSTOM' in panel.COMPAT_ENGINES:\n panel.COMPAT_ENGINES.remove('CUSTOM')\n\nif __name__ == \"__main__\":\n register()\n```\n\n1) Copy code from above into a new script file inside of Blender\n2) Run the script\n3) Select \"Custom\" as Render Engine inside Render Properties\n4) Select \"Material Preview\" as display method in the viewport.\n\n[crash.blend](crash.blend)\n\nCrash log:\n```\n# Blender 2.93.0, Commit date: 2021-05-19 10:14, Hash a294670bacb1\nEngine 'CUSTOM' not available for scene 'Scene' (an add-on may need to be installed or enabled) # Error\nbpy.ops.text.run_script() # Operator\n\n# backtrace\n./blender(BLI_system_backtrace+0x20) [0xa63c950]\n./blender() [0xf6082a]\n/lib/x86_64-linux-gnu/libc.so.6(+0x46210) [0x7f7ce17be210]\n./blender(GPU_framebuffer_blit+0x2f) [0x8664faf]\n./blender() [0x1485fc3]\n./blender() [0x14597fe]\n./blender(DRW_draw_render_loop_ex+0x242) [0x1459a72]\n./blender(view3d_main_region_draw+0x8f) [0x1d3570f]\n./blender(ED_region_do_draw+0x851) [0x189c771]\n./blender(wm_draw_update+0x52e) [0x130cc9e]\n./blender(WM_main+0x30) [0x130aa20]\n./blender(main+0x319) [0xe881c9]\n/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf3) [0x7f7ce179f0b3]\n./blender() [0xf5d1bc]\n\n# Python backtrace\n```", "Included USD libraries cause symbol conflict with addons using USD binding\nOperating system: macOS\nGraphics card: N/A\n\nBroken: 2.83 beta\nWorked: Never\n\nIncluded USD libraries cause symbol conflict with addons using USD python bindings. That is, if an addon wants to use USD in it (as a python c-extension most likely), there is an error:\n\n\n```\n\"FATAL ERROR: [TF_DEBUG_ENVIRONMENT_SYMBOL] multiple symbol definitions. This is usually due to software misconfiguration. Contact the build team for assistance. (duplicate 'TF_SCRIPT_MODULE_LOADER')\nin _Add at line 96 of /Users/brecht/dev/build_darwin/deps/build/usd/src/external_usd/pxr/base/lib/tf/debug.cpp\"\n```\n\n1. Download and extract a build of USD. These ones have python3 bindings (I built my own). usd#bin\n2. Open python editor in blender.\n3. Run commands:\n\n```\nimport sys\nsys.path.append(\"PATH_TO_USD/lib/python\") # this is to get the pythonpath right\nfrom pxr import Usd\n```\n\n\nSo I can see two possible solutions to this. \n1. include python bindings with Blender\n2. somehow fix the symbol names in the blender side or in the external USD build so they don't clash. I tried to do this with specifying a different namespace name than \"pxr\" in the USD build but this didn't seem to help.\n3, link python bindings of USD against blender.exe (Not sure how/if this would work)\n\nWilling to work on the solution, but wanted input from Blender developers.\n", "Can't bind shortcut Alt+Numbers 0-9\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\nKeyboard: UK ISO (tenlesskey) w/ UK Layout\n\nBroken: version: 2.93.4\nBroken: version: 3.0.0 Alpha\nWorked: None (tried 1.8, 2.28, 2.34, 2.58, 2.66, 2.79, 2.80, 2.82)\n\nUnable to use Alt modifier with Numbers 0-9\n\nTry to rebind something (e.g Hide Collections to Alt+Number key).\nThe binds will fail to execute.\n\n{[F10604609](210926_blender_10b48.png), size=full}\n\nBroken:\n• Alt+Numbers 0-9\n• Alt+Numpad Numbers 0-9\n• Using 'Left Alt' as a custom modifier key in the keymap plus Numbers 0-9\n\nWorks:\nAlt+F Keys 1-24 work (unfortunately Alt+F4 is hardcoded to exit blender.exe)\nAlt+Qwerty/Punctuation keys work.. just seems like its Alt+Numbers that fail\n\nThank you for your time and consideration :3", "Proportional editing lag\nWin7 x64, GTX 460\n\nBroken: 2.73, 2.74\nWorked: 2.63 \n\nWith proportional editing on, I get a few seconds freeze when activate translating (G); than after freeze, \neverything goes ok. I thought that the problem was a large mesh, but 2.63 works great and without lags in the same file. \n\nMy \"bugged\" file is the one, I am working with. If it is possible, can i sent it in private and not give it to the world here?\n\nThank you!", "Menu search is case sensitive for non-ASCII characters\nOperating system: Linux-6.2.0-32-generic-x86_64-with-glibc2.37 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 535.86.05\n\nBroken: version: 4.0.0 Alpha\nBroken: version: 2.80 (sub 75)\n\nWhen using the fuzzy search feature of some menus, the result is usually case-insensitive. But this only applies to ASCII characters.\nThis applies to the F3 menu, and those affected by #110855.\n\n- Set the translation to French in the User Preferences\n- Open a Geometry Nodes editor\n- Shift+A to bring up the Add Node menu\n- Start typing “éval” (the first E has an acute accent)\n - If you can’t input this, just copy and paste it\n- The expected results start with “Évaluer” (evaluate), but what comes up is “Décaler” (offset), probably because of the fuzzy search\n- Start typing “Éval”, with a capital É\n- The expected results are now displayed.\n\n<video src=\"attachment\" title=\"case_sensitive_search_na.mp4\" controls></video>" ]

[ "Keyboard shortcut doesn't work after switching applicationss\nOperating system: windows 11\nGraphics card: Intel(R) UHD Graphics 630 & NVIDIA GeForce GTX 1060 with Max-Q Design\n\nBroken: Blender 3.1, a solid update to an already outstanding release. Released March 9th, 2022\n\nWhen using different Keyboard shortcut for every window, keyboard shortcut doesn't work when we switch between applications\nCTRL, ALT, Shift, tab are fine, but all of the characters are no response, such as G, E, Shift D, etc.\n\n- Open Windows Settings tab\n- Enable: `Let me use a different input method for each app window` {nav Time & Language > Typing > Advanced Keyboard Settings}\n- Open chrome or any other application and change input method (Chinese for example)\n- Open blender and keep default input method (English for example)\n- Type something in other application window\n- Switch to Blender and use hotkey" ]

[ "GPencil: Add multiframe support to Move to Layer\n", "Render Artifacts with SSS + Motion Blur + Use Spacial Splits\nWhen you have the \"perfect storm\" of Subsurface Scattering, Motion Blur, and \"Use Spacial Splits\" (Render>Performance) all applied to an animated mesh, the rendered image will exhibit some pretty severe artifacts and mesh cracking on certain frames of an animation.\n\nThe main issue appears to be related to \"Use Spacial Splits\"\n\nOpen the attached zip - watch the comparison video. Open the .blend file, hit F12 to render (frame 34 is a good example)[Spacial Splits +Motion Blur Artifact Bug.zip](Spacial_Splits__Motion_Blur_Artifact_Bug.zip)\n\n[Render Bug - Spacial Splits + Motion Blur + Sss-1.m4v](Render_Bug_-_Spacial_Splits___Motion_Blur___Sss-1.m4v)", "Omitting a leading zero on transform commands complains on the terminal\nOperating system: Kubuntu 20.10\n\nBroken: 2.83 - master\n\nTyping transform commands with a number lacking a leading zero complains on the terminal but works otherwise. For example, when inserting commands like scale 90% as `s.9` there is output on the terminal saying:\n\n```\n File \"<string>\", line 1\n .\n ^\n SyntaxError: unexpected EOF while parsing\n```\n\nThat is a bit worrisome because I thought there was something wrong with the blend file I'm working on.\n\n1. blender --factory-startup\n2. select the default cube\n3. insert `s.9`\n", "Isolate File Encoding Colour Management\nCurrently, a pixel pusher with the correct view set for an `Apple Display [P3](P3.txt) Display`, ends up baking in the `Display [P3](P3.txt)` transform into file encodes due to the way Blender encodes the file based off of the currently selected view.\n\nSolution is to isolate the file encoding from the other output contexts such as displays and views.", "Improve MemFile Undo (for general use)\n**Motivation:**\n\n- Resolve bugs where changes to data outside a 'mode' are ignored by the undo system.\n- Unify logic and reduce code duplication.\n\n**Description:**\n\nCurrently we have a blend file based undo system (internally called 'MemFile').\n\nWe could use this almost everywhere (edit-modes for example), it would be sub-optimal at the moment.\n\nThis project would involve optimizing the case when only a limited set of data-blocks change - so we can skip writing all other data-blocks into the memfile.\n\nIt may also involve using better binary diffing for `MemFile` undo (as edit-modes currently use).\n\nThis would most likely not cover image painting and sculpting.\n\n**Estimated Time:** possibly 2-4 weeks *(hard to know)*.", "Support MinGW\nWe have some patches in MINGW-packages to make blender build on MinGW.\nWe already have it for Windows ARM64.\nIs it OK to upstream those patches?", "Add support for more gettext flags to UI translation tools\nAs reported in 101830#issuecomment-1023470 , Weblate `C-format` is not very smart, and generates a lot of false-positives.\n\nThe solution is likely to not enforce these checks over the whole messages, but selectively enable them for messages that actually need it (e.g. (almost?) all RNA-extracted message will not have such formatting).\n\nThis will require updates in the python modules and add-on code generating these translation files.\n\n----------------\n\nParent task: infrastructure/blender-projects-platform#65", "'Solve Camera Motion' operator sometimes prevents Auto Save from running\nBroken: 3.5.1, 3.6\nWorked: maybe 3.4\n\nInvestigated in #107421 by @3di.\nIt's still unclear when this happens, but it could be replicated with the attached file.\nThe function `void wm_autosave_timer(Main *bmain, wmWindowManager *wm, wmTimer * /*wt*/)` in `wm_files.cc` skips AutoSave if any modal operators are running.\nIn this case, this can happen with the `\"CLIP_OT_solve_camera\"` operator even after the operation is theoretically completed.\n\n- Unzip attached .blend file\n- Go to \"Motion Tracking\" workspace\n- In the solve tab, press `Solve Camera Motion`\n\nAuto Save stops saving or saves very late.\n\nno%20autosave.zip?dl=0\n\n", "File Output node outputs incorrect dimensions in some cases\nOperating system: Linux 64-bit\n\n2.92\n\nWhen connected to certain composite node trees, the File Output node will output images with incorrect and unexpected dimensions. This appears to have something to do with the blur node, but I suspect the issue might occur with other nodes that alter the image buffer dimensions.\n\n[composite bug.blend](composite_bug.blend)\n\n - Open and inspect attached file\n - Adjust output file paths to whatever is convenient\n - Render a frame\n - Review the outputted image\n\n", "Add operator to export assets with catalog definition\nSince the catalog definition is stored in separate files inside the asset library, sharing individual .blend files will not bring along the catalog info. The contained assets will be unassigned (the catalog ID is still stored in the asset metadata, but there is no matching catalog).\n\nWe discussed having a way to export assets together with the necessary catalog definition. This could simply create an archive file containing .blends (or just a single .blend?) and a catalog definition file.\nOr maybe we should add support for packing catalog definition files? Interesting idea in fact.", "Port Temporal AntiAliasing to Grease Pencil\nPort should be as simple as copying what Workbench does, with the addition of [D10414](D10414).\n\nSee parent task for more detail.", "2.90 Not opening on Linux\nOperating system: Linux Mint 18.1\nGraphics card: Intel HD 620\n\nBroken: \nBlender 2.90 won't open up fully, it appears to open up and immediately closes\nWorked: \n\nTried a few different ways to open, won't open.\n\nOne of the methods I tried opening with was through the Terminal and this was the message I had returned back:\n\nRead prefs: /home/john/.config/blender/2.90/config/userpref.blend\nfound bundled python: /home/john/Desktop/3D Related/1BlenderVersions/blender-2.90.0-linux64/2.90/python\nGPUShader: linking error:\nerror: Input block `ShaderStageInterface' is not an output of the previous stage\n\nGPUShader: linking error:\nerror: Input block `ShaderStageInterface' is not an output of the previous stage\n\nWriting: /tmp/blender.crash.txt\nSegmentation fault", "persisting issue: MacOS, Blender's system device usage rights (info.plist) add camera & mic usage descriptions\nOperating system: MacOS (13.2.1) \nGraphics card: any\n\nBroken: (All Blender addons that would need external peripheral like camera or microphone access, will cause Blender to CRASH, because of MacOS)\nWorked: (it's a MacOS dependent issue (since maybe macOS 12?)\n\n*\n\"Termination Reason: Namespace TCC, Code 0\nThis app has crashed because it attempted to access privacy-sensitive data without a usage description. The app's Info.plist must contain an com.apple.security.device.camera key with a string value explaining to the user how the app uses this data.\"*\n\n\nInstall an anddon that would need camera / microphone access. After checking the checkbox in blender addon menu, Blender crashes (or at the point that addon is running code that would need such peripheral. Nowadays you cannot add it manually and this needs to be addressed globally.\n\nFIX:\nglobally add info.plist usage rights for camera, mic and other widely used peripherals for blender's MacOS version, since the current state of MacOS, doesn't allow addon's to access any of those, and it really hinders the blender as a core, those of course aren't used if not needed, but at the moment addon developer cannot have any modules that use any of those without MACOS causing the blender to literally crash. I hope this could be corrected asap. Before it was still possible to manually grant access to applications via macos settings, but alas, its not anymore possible.\n\n", "Support Multiple Catalog Definition Files\n", "Tasks after BI removal\n* Code cleanup\n * Preview render code could be simplified, not used old preview.blend** Customdata mask handling still assumes old draw types\n * Move camera matrix code out of Render and into Freestyle, the only user\n * ~~Remove BLENDER_RENDER compat engine, find better system~~\n * Simplify GLSL shader coordinates, originally designed for camera space shading\n\n* Bigger tasks\n * #54696 (Blender Internal displacement baking move to Cycles)\n * [x] #54697 (Vertex color baking for Cycles)\n * #54656 (Cycles/Eevee texture nodes for modifiers, painting and more)" ]

[ "Unify blend file thumbnail extraction\nCurrently there are two utilities for extracting thumbnails from Blend files.\n\n- `./source/blender/blendthumb/src/BlenderThumb.cpp`\n- `./release/bin/blender-thumbnailer.py`\n\nThis is causing a problem for adding a new compression method [D4402: LZ4 Compression for blend file IO](D4402) because we need to support a new kind of decompression for both.\n\nBesides this, Python doesn't support LZ4, making it likely we would need to write a new utility in C/C++ anyway.\n\n----\n\nThis task proposes to share use a single thumbnail extraction utility (probably based on `BlenderThumb.cpp`) which all platforms can share." ]

[ "Cycles: Motion blur artifacts\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.88\n\nBroken: version: 3.2.2\n\n\n**description of error**\nWhen rendering an animated rigged model on cycle **with motion blur** ; if the model has **auto smooth enable AND a subdivision modifier**, then it is possible that strange artefact will appear during the render. (doesn't happen in the viewport) \n- This bug do not happen if auto smooth is disable or doesn't have a subdivision modifier. \n- The bug also only appear during some specfic poses. So this is mainly noticiable when rendering an animation.\n- Those artefact often appear near sharp edge.\n- Only happen on Cycle, Eevee render seem fine.\n- [Edit] changing the shutter values of the blur will sometime remove the bug on some frames OR make the bug appear on frame that was fine before.\n\nHere 2 Render exemple with and without autosmooth , both has motion blur enabled :\n \nAnd here an animation where you can notice the bug.\n[Animated0035-0318.mp4](Animated0035-0318.mp4)\n\n\n[MotionBlurCycleArtefact.blend](MotionBlurCycleArtefact.blend)\nIn this blender file, rendering the frame 302 will make the bug happen. \n- If you disable auto smooth on the model (the main body and not the cloth) then the render will be fine.\n- If you disable the subdivision modifier on the model then the render will be fine.\n- If you disable the subdivision modifier AND the auto smooth on the model then the render will be fine.\n\nThank\n", "Cycles: Banding when using SSS Random walk and SS Anisotropy\nOperating system: Windows 10\nGraphics card: RTX 2070 Super\n\nBroken: Blender 3.3.0\n\nHello,\n\nI'm getting render artifacts when render a small object with a low SSS value and SS Anisotropy greater than 0. It seems like the edges of the mesh are visible in the SSS effect.\nI've tried to increase render samples and turn off denoise and no luck. If a remove SSS it works fine, setting anisotropy to zero they become barely visible.\nI uploaded an example scene file and a screenshot of the artifacts.\n\nThanks\n\n\n\n[system-info.txt](system-info.txt)\n\nA simple scene for reproduction. Just a sun light on top and an area light behind.\n \n\n \n[ArtifactsExample_v2.blend](ArtifactsExample_v2.blend)", "Occasional render glitches when using simulation zones on an object without a material.\nThis bug has an easy workaround, mostly posting this so anyone running into the same issue knows the solution.\n\nOperating system: Linux-5.15.0-79-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 530.30.02\n\nBroken: version: 3.6.2\nWorked: N/A, 3.6 is the first version with simulation nodes.\n\nIn Eevee, rendering an object with a simple simulation zone applied in geometry nodes and no material assigned to the object results in some rendered frames being blank or displaying odd glitches.\n\nThe workaround is to simply assign a material to the object.\n\nCheck out the attached .blend file for reproduction.\n\n", "Poor stroke quality when a model is far from the world origin\nOperating system: Win 10\nGraphics card: Nvidia Rtx 2070\n\nBroken: 2.83, master\nWorked: perhaps never?\n\njittery strokes when mesh is far from the world origin in perspective mode\n\nHow to reproduce this bug : \n- Create a grid (set the location to x : 2000 m and y : 1000 m)\n- Create a multires modifier, with 5 subdiv.\n- Sculpt with the clay brush\n- In Perspective mode (left stroke) the stroke is ugly\n- In ortho mode (right stroke) the stroke is clean\nTo fix the stroke poor quality, i had to fine tune the Clip start and the clip end values of the 3d view. \n[#102747.blend](T102747.blend)\n", "Artifact with cycles (interaction with moving spotlight and motion blur)\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.1\n\nI wouldnt describe it as a error, more as a artifact in cycles, basically when you parent a light to an object with motion blur, it doesnt behave as expected, it clips into objects and deforms due to the motion blur, there is no easy way to explain this, its better to show it, keep in mind that the spot light in these examples is completely Static! its just parented to the empty that also is parented to the camera, what is moving is the empty, light in real life doesn't behave this way!\n\ni first noticed this in a render i did for my animation \nrender: view\nproject file: view?usp=sharing\n\nso i did a test render and the problem pressists\nrender: view?usp=sharing\nproject file: view?usp=sharing\n\n\n\nAll the best! If this report is invalid please inform me why, because i cant find a work around this issue and there is no information online regarding this problem. Thank you.\n\n\n\n", "Custom Normals Render Artifact in Cycles \nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.99\n\nBroken: version: 3.0.1\nWorked: Unknown\n\nWhen rendering a cube with custom normals in cycles the render result in cycles has artifacts. This is an issue for us because we also use cycles to bake Ambient Occlusion data to the vertex colors. Those artifacts are than also visible in the vertex colors. The same artifact occurs in previous versions of blender aswell (LTS Blender 2.93.7)\n\n\nBased on startup file:\n\n- Shade the default cube smooth\n- Apply Normals Auto Smooth\n- Add a single bevel using a bevel modifyer and apply it\n- Add a Weighted normal modifyer with weight 100 to the cube and appy it aswell\n# Render in cycles and behold render artifacts near the beveled edges\n\n[custom-normals-cycles.blend](custom-normals-cycles.blend)\n\n\n", "Cycles HIP RT geometry artifacts\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 6600 ATI Technologies Inc. 4.5.0 Core Profile Context 23.5.2.230523\n\nBroken: version: 3.6.0\nWorked: 3.5.1\n\nMesh in blender cycles looks weird, some faces in the geometry don't show correctly in Blender 3.6 with the new HIP RT\n\nJust installed and opened blender, turned on HIP RT in the settings tab and then opened a file with a model i made with blender 3.5.1 and switched to cycles.\n\ni attach a screenshot of how it looks with HIP RT and how it's supposed to look (how it looks with HIP RT deactivated)\n\n", "Eevee refraction regression in orthographic view\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.91\nWorked: 2.92\n\nIn orthographic view, with the view or camera camera perpendicular to a refractive surface, screen space refractions appears blurry, as if it had roughness, even with refraction shaders with 0 roughness\n\n2.91: \n\n2.93: \n\nEdit: In perspective it's also noticeable in the center. It looks like a small ball or dot that distorts the image slightly\n\n\nOpen file in different versions\n[bugRefractionEevee.blend](bugRefractionEevee.blend)\n\n\n", "World Light Extraction\nEevee world probe is not suited for very high frequency details and even less for sunlight. For this we want to introduce a threshold. All light under the threshold would be pushed to the world diffuse/reflection probe. All light above the threshold would be moved to an importance sample map and used to position a (virtual) sun light that represent the missing energy from the probe.\n\nWe can position the sun in 2 ways: randomly per sample using the importance map, or fixed using a Spherical harmonic decomposition.\nThe first one offers more precise lighting (can work with multiple light source) when the second one offers more stable lighting (good for a sun).\n\nThe good thing is that this virtual sunlight can be shadowed! Thus making most of the light coming from the environment shadowed.\n\nThis should fix this sort discrepancy between cycles and eevee:\n\n\nEstimate : 2 weeks", "Unwanted behavior when modifying the normal vector of the refraction node\nOperating system: Windows 10\nGraphics card: RTX 3090\n\nBroken: Blender 3.3\n\nA subtle but sharp bending effect can be seen on refractive material with a modified normal vector.\n\nWhen modifying the \"normal\" vector of the refraction node (or any node with refraction), when viewing at a certain angle, a \"sharp\" bend appears.\n\nA series of bends in the red circled area can be seen where it should be a smooth or straight line.\n\nIt's more obvious if the normal vector is not normalized, but it should be normalized anyway, or the refraction/reflection calculation will not work correctly:\n\n\n\n**Possible causes**\nThe \"ensure_valid_reflection\" could be the cause.\n\nThe commit that added \"ensure_valid_reflection\" is to fix the reflection black spot issue when using a normal map.\nrBd6e769d32e7939e3bbd1986cdc4abd2b13135eab\nWhere the normal vector is being regularized so the reflection ray will not penetrate the surface.\n\nThe code \"ensure_valid_reflection\" is currently here:\nbsdf_util.h#L109\n\nIt is being used in many places where \"refraction\" is present and modifies the normal vector:\nclosure.h#L105\n\nI couldn't track down where exactly causes the issue, but I can show the bending effect aligns where \"ensure_valid_reflection\" kicks in.\n\nThe sharp line that appears on the \"reflection\" side is expected; without it, the ray will penetrate the surface.\nWhile on the \"refraction\" side, the ray should not need to be \"ensured.\"\n\n**Expected behavior**\nThe ray should behave the same as there is a surface with the same normal vector.\n\nNote:\nIn most cases, this doesn't make a significant appearance difference.\nBut this could open a lot of creative uses where light rays can be directed to any angle procedurally;\nFor example, a procedural fresnel lens surface can be used to create something like magnifiers.\n\n- Open attached blend file\n- Enable `Viewport Shading` `Rendered`\n[RefractionBendingEffect.blend](RefractionBendingEffect.blend)\n\nThe bending effect can be seen in the middle\n\nI have provided two mix nodes for quickly examining normalize/non-normalized normal vector, and refraction/reflection behavior.\nNote that when switching to a non-normalized vector, the bending effect will appear on a different viewing angle.", "Black artefact when mixing transparency and glass at a certain ratio\nOperating system: WIn 10\nGraphics card: GTX 1080\n\n2.92\n\nBlack artefact appear on flatter surfaces when mixing 99.9% transparent and 0.1% glass shaders. Glass should be rough. \n\nMake a cube, switch to Cycles, add material, add Glass and Transparent Shaders, Mix them together, put factor on 0.999, increase Glass roughness, depending on roughness a entirely black circle will appear on the cube when viewed straight on. \n\n[TransBug.blend](TransBug.blend)", "Surface smooth brush produces artifacts when set to high strength on dense mesh\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: AMD Radeon R7 200 Series ATI Technologies Inc. 4.5.13587 Core Profile Context 20.4.2 26.20.15029.27017\n\nBroken: version: 2.90.0 Alpha\n\nSurface smooth brush produces artifacts if set to high strength, 0 shape preservation and 0 per vertex displacement when used on the very edges of a dense mesh i.e not at dead center or head on.\n\n1) Apply subdivision surface modifier set to level 5 Catmull-Clark on the default cube\n2) Switch to sculpt mode\n3) Switch to Smooth brush and switch deformation from 'Laplacian' to 'Surface'\n4) Set brush strength to 1 and set both 'Shape preservation' and 'Per vertex displacement' to 0\n5) Use brush, 'edge on' i.e not at the center, in orthographic 'Front' view this translates to using the brush the furthest away on the X-axis from the world origin while still being on the mesh.\n\nI have had much more success in reproducing this bug by using brush on the left and right side of the mesh (+ or - X) as opposed to up and down (+ or - Z). In the file linked below i have already set everything up, to reproduce the bug you just have to use the brush at the right place.\n\n[SurfaceSmooth.blend](SurfaceSmooth.blend)\n\n\n", "Normal map node worked is not correct\nOperating system: win10 x64\nGraphics card: gtx 1650\n\nBroken: 3.5.0, 1be25cfff18b, 2023-03-29\nBroken: 2.93\nWorked: 2.92\n\nCaused by 1c22b551d0\n\n\nI do not think that this is an error or a bug, rather it is not a refinement. \nIn some cases, the normal map node does not display them as you would like. \n\n\nI created a Hight poly model and bake it into a plane. Then I applied the resulting texture to the Low poly base mesh. It didn't display correctly there, only part of it smoothed out, but the other part showed sharp edges, very similar to the Fresnel effect\n\n", "Dynamic Indirect light cache\nThe idea would be to have a multiple level cache that stores a voxelized version of the scene to render the irradiance points.\n\nEach voxel would be linked to a set of surfel (surface elements) that stores color and normal of the surface.\nEach time the lights in the scene changes, the voxelized version is updated and propagates the change to the irradiance points.\n\nThis means having a list of surfel per voxel (say 6, 8) and a list of voxel per irradiance points.\nThis would be a mix of The Division dynamic indirect lighting [link]] and Mc Guire raytraced irradiance probes [[http:*jcgt.org/published/0008/02/01/|link](https:*www.gdcvault.com/play/1023273/Global-Illumination-in-Tom-Clancy).\nThis feature might be limited to hardware that supports compute shaders.\nThis is not a trivial feature and needs a lot of R&D.\n\nEstimate : 4 weeks\n", "EEVEE: Contact shadows problem on layered hashed materials\nOperating system: linux open source drivers\nGraphics card: rx 6800\n\nBroken: 2.93\nWorked: Never (2.83+)\n\ncontact shadows dont look correct on certain angles for layered alpha hashed materials(hair cards for example). iissue is more noticable in viewport then in final render, but it is noticable in both.\n\nmake a sphere make it alpha hashed and slightly transparent then duplicate and scale geometry with slight offset to make layers then rotate around object with contact shadows enabled on the lamp\n\n\n[contact shadows problem.blend](contact_shadows_problem.blend)" ]

[ "Random Walk SSS in Principled BSDF gives very dark result for small subsurface inputs\nOperating system: NixOS 19.03, Linux 4.19.36\nGraphics card: NVIDIA 970 GTX\n\nBroken: 2.79, 2.80 git (93cabce3a156a07a76c26d7f3e78c3bc3212f623)\n\n\nUsing the Principled BSDF's \"Random Walk\" subsurface scattering method, if the \"Subsurface\" input is set to a low, non-zero value (e.g. < 0.01), you get a very dark looking object. For example, the following image shows three Suzannes with the same material except a different \"Subsurface\" value, from left to right 0, 0.0001, and 0.1:\n\n\n\nThe Blender manual (principled.html) somewhat cryptically states that the Subsurface input acts both as a mix factor between Diffuse and SSS *and* as a multiplier for the subsurface radius, but either way this doesn't seem like expected behaviour. It doesn't occur with the Christensen-Burley method.\n\n\nBlendfile: [random-walk.blend](random-walk.blend)" ]

[ "Looks like an eevee limitations with UV maps and attributes\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.5.0 Alpha\n\nWhen working with multiple UVs and attributes, eevee disables some normal/bump maps.\nThere is an error in the console:\n```\nGPU_material_compile: Attribute \"nor\" overlap one or more index from another attribute. Note that mat4 takes up 4 indices. While merging eevee_legacy_material_surface_frag_opaque\n```\nIn 3.4, the shader cannot compile at all (pink texture).\n\nThis is not documented anywhere and it is not clear to me how it works, because until I connect nodes with attributes in Base Color, all normals are displayed.\n\n1. Open .blend file.\n2. Turn on the eevee render.\n3. Look, all normals are displayed.\n4. Connect nodes with attributes to Base Color.\n5. One normal map is missing. An error appeared in the console.\n\n", "Intersect Boolean does not work properly when mesh has a degenerate face\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.14760 Core Profile Context 20.45.37.01 27.20.14537.1001\n\nBroken: version: 2.91 - 3.0.0\nWorked: n/a\n\nThe `Intersect Boolean` Exact mode does not work correctly when the mesh has at least one degenerate face.\nIn this case, it was tested on a mesh with a triangle whose vertices are collinear\n\n- Open attached file - (The edited mesh has a degenerate triangle)\n- Go to `Face` -> `Intersect (Boolean)` - (The operation will not work because of that triangle)\n\n- Undo the operation\n- Move the triangle vertex a little so that the face is no longer degenerate.\n- Reselect that loose geometry ({key L})\n- Redo Boolean operation (now it works)\n[intersect_w_degenerated_tri.blend](intersect_w_degenerated_tri.blend)\n\n", "Cloth simulation collision differs between viewport and render when collider is hidden\nOperating system: Fedora Linux\nGraphics card: AMD\n\nBroken: 3.3.0 - 3.6.1\nWorked: \n\nCloth simulation collision differs between viewport and render when collider is hidden in both.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n- create a cloth that will collide with the default cube\n- hide the default cube in both render and viewport\n- press play\n- the cloth will collide with the hidden default cube\n- press pause\n- go to frame 1\n- enter and exit edit mode on the cloth mesh to reset the cloth simulation cache\n- render the animation\n- the cloth will not collide with the hidden default cube\n\nAdditionally:\n\n- If you play with cube hidden, up to when cloth collided the cube, and render animation, the animation will be like: cloth collided the cube, and after the cache runs out, it starts falling as if the cube has not been there.\n\n", "Viewport Z-Bias & retopology overlay related issues\nSince this is being reported frequently and retopology overly has caused some weird issue with mask display as well previously, it's best we keep track of them in a central place so we could check if any particular fix is working for all situations:\n\n------\n\nProblem still presents:\n\n- Retopology overlay affecting other parts of the drawing/picking operation: #109115 #107768 ~~#109277~~ ~~#107411~~\n- (Potentially related) Selecting when X-Ray turned on is not visually effective and then switching to face mode crashes: #110136\n\n------\n\nFixed:\n\n- ~~Wireframe visible in close distances: #110048 #109040 #109514 #109499 #109685~~\n- ~~Retopology overlay 0 z offset problem: #109662~~\n\n-------\n\nProbable cause:\n\nAs suggested in 107768#issuecomment-937545 , I believe the problem here is that\n\n1. `DRW_object_is_in_edit_mode` does not exclusively check if object is in edit mode, but it combines other logics that affects drawing and stuff.\n2. `XRAY_ENABLED` marco is not always accompanied with `RETOPOLOGY_ENABLED`, this leads to inconsistencies in the user interaction and the display.", "In cycles, ignore shadow of volume object will cost more time.\nOperating system: windows 10 64\nGraphics card: rtx 2080\nCpu: AMD Ryzen 7 2700\n\nBroken: 2.83 2.82 .....\n\nIn cycles, ignore shadow of volume object will cost more time.\nUse texture in volume material.\nVolume object cross other object.\n\n\n[vol_bug.blend](vol_bug.blend)", "Exaggerated Ambient Occlusion Effect with Subsurface Scattering Materials in Cycles\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 2.93.2\n\nWhen rendering with Cycles, whenever meshes with materials that have subsurface scattering intersect, a very obvious ambient occlusion effect is generated around the intersections. It functions as expected with Eevee, but in Cycles, there's always a very unattractive dark seam line.\n\nApply any subsurface scattering materials to any two intersecting meshes, and render out with Cycles.\n\nHere are my results, I used a procedural snow shader, but I get similar results with any subsurface scattering materials.\n[subsurface scattering.blend](subsurface_scattering.blend)\nResults with Subsurface Scattering in Cycles\n\nSame shader in Cycles with Subsurface scattering turned off\n\nSame shader in Eevee with baked indirect lighting and subsurface scattering turned on\n\n\n", "A combination of shader nodes causes lack of material emission at any brightness (Cycles)\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: version: 3.4.0(tested as well on 3.3.0 and 3.5.0 alpha)\nWorked: (none)\n\nA difference of bevel node and normal coordinates produces no light at any emission level. Clipping the result due to suspicion about negative brightness has no result.\nNote: adding any value to the result adds corresponding emission, but despite visuals, brightness from combination above does not.\n\n\n- Open attached file\n[cycles_where_is_emission__bug.blend](attachment)\n\n\n", "Intersect (Boolean) in edit mode produces inconsistent selection\nBroken: 2.92 master\nWorked: N/D\n\n\nThis is how selection looks like after performing a Difference operation in edit mode using an enlarged copy of the same cube:\n\n\n\n4 verts and edges selected, yet 0 faces (should be 1). The operation was performed in vertex mode. If you do it in edge mode, again 4 verts and 4 edges, 0 faces (should be 1). In face mode though - nothing gets selected at all.\n\n\n1. Enter edit mode on the cube.\n2. Duplicate the geometry, move it to the side and scale it slightly.\n3. Invoke a Face -> Intersect (Boolean).", "Geometry Nodes: Curve primitives issues with object material\nBroken: version: 3.2.0 Beta, 3.3.0 Alpha\nWorked: n/a\n\nMaterials assign to the object don't work in some cases for curve primitives created in the geometry nodes tree.\n\nOpen file, you will see that the star doesn't get the red object material.\n\n[curve_object_material.blend](curve_object_material.blend)\n\nUnhiding any of the hidden nodes (Grid or Set Material) will make the object material show up.", "Shadow caustics visible through colored glass has intensive, solid color for directional/sunlight and are probably too dark for spotlight/pointlight\nOperating system: Windows 11\nGraphics card: GTX 1070 Ti\n\nBroken: 3.22 master\nWorked: none\n\nThis happens when transparent object intersects with diffuse one and caustics are visible though volume. I know that MNEE doesn't work with volumes, but I think this might be a bug.\nFor directional light and sunsky system caustics are colored to pure red, green or blue or are just white. No other colors are possible. Spotlight and pointlight work, but color of intersected primitives differs substantially from not intersecting ones when looking thought them. This is especially visible on the left upper sphere that has only slight tint. Hope this is useful.\n\nmnee nishita sunsky:\n\n\nmnee directional light:\n\n\nmnee spotlight:\n\n\n\n1. Create an object casting caustics like sphere over default cube.\n2. Create transparent material for sphere that consists of two shaders, glass and volume, adjust volume color to something darker then white\n3. Turn on caustics for light and for objects, render in viewport.\n4. Move the sphere so it intersects with the cube, see the result.\n5. Switch light into directional mode and try to adjust volume absorption color.\n\n...or just open attached file and render. Lighting is set to nishita sky, but both spotlight and directional light are added and hidden.\n[mnee_test_master.blend](mnee_test_master.blend)", "Intersecting Meshes are not marked as overlapping anymore when selected\nOperating system: Linux-5.4.0-48-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.90.1\nWorked: 2.83.x\n\nIn blender 2.83.x Solid View intersecting objects had, if selected, the otherwise hidden parts shown in a more muted color when overlapping. with 2.90.x this is no longer the case. \nI hope this did not get removed intentional but is a missing feature.\n\nDuplicate the default Cube, move it a bit so that is still intersects with the first cube and confirm movement. \nWatch the outline in Solid View be the same color and thickness all around the mesh. \nin 2.83 while moving the white outline turned light grey when intersecting and the selected object outline turned a darker orange. This is very helpful for positioning elements.\n\n", "Retopology: Active Object Geometry gets Occluded by Semi-Transparent Reference Object in Edit Mode\nOperating system: Linux Ubuntu 16.04\nGraphics card: Nvidia GeForce MX130\n\nBroken: 2.83.0\nWorked: 2.82.a\n\nActive mesh is occluded by high poly or another reference mesh in Edit mode. This is: I want to edit mesh A. I need mesh B to get a reference. So I make B to have a semi-transparent material for Viewport view. I select A and go to Edit mode to edit it. However, its edges and vertexes are occluded by B. We can see a comparison of 2.83.0 (first image) with 2.82a (second image): version: 2.83.0 does not show active object geometry so that we can not edit it. On the other hand 2.82a does show required geometry.\n\n \n\n1. Open Blender 2.83.0\n2. Create a Sphere mesh in the world origin so that it fits inside the default cube.\n3. Assign a semi-transparent viewport display material to the Cube so that we can see the Sphere inside. Change color also to have a better view.\n4. Select Sphere and go to Edit mode.\n5. Press A to select all vertexes.\n6. We can see that Edges and Vertexes are occluded by the Cube so that we can not edit the Sphere properly.\n\n\n[bug-report.blend](bug-report.blend)\n", "Blender Cycles \"Launch failed in CUDA queue copy from device (integrator_intersect_shadow)\"\nOperating system: Windows 10 - 22H2\nGraphics card: 3090ti 31.0.15.3667\n\nBroken: (3.5, 3.5.1, 3.6.1)\nWorked: None\n\nBlender Cycles \"Launch failed in CUDA queue copy from device (integrator_intersect_shadow)\"\n\nI have a volumetric setup in the scene. To show up as haze on the horizon. Error only occurs when Volumetric is active.\n\nEnsure the WorldSphere is enabled for render and just render animation for 1279. \nIt will do 1 - 10 frames before throws the error attached. \n\nFile is too big for the attacher, please get blend file from google drive link.\n\nview?usp=drive_link\n\nSimplified: view?usp=drive_link\n\n", "Cycles Shadow Catcher Creating Multiple Shadows\nHi,\n\nThis .blend file will demonstrate two shadow catchers overlapping each other and creating double shadows. It would perhaps be interesting to add a termination value for the amount of shadows cast or just removed entirely.\n\n\nThanks!\nBeau\n\nSystem:\nWindows 10 64 bit\n\nBlender 2.79b f4dc9f9d68b\n\n\n\n\n\n[shadow_catcher_double_shadows.blend](shadow_catcher_double_shadows.blend)", "CYCLES: Surfaces inside volumes are Not Lit or Shadowed by the volume if the volume is not visible to Camera Rays\nLinux Mint Mate\nGTX 1050\n\n\nBlender 2.81 Alpha\n2019-09-27 22:44\nHash: 1c1a3198af9d\n\n\nDiscovered while trying to catch shadows from volumes on surfaces for compositing.\n\nSurfaces inside of volume meshes are not shaded (lit/shadowed) by the volume if the volume is invisible to camera rays\n(checkbox in Object Properties -> Visibility -> Ray Visibility -> Camera)\nAs far as i can tell this also affects Holdout.\n\n\n\n\nCurrent workaround: make sure volume IS visible to camera rays, and instead bypass it in the volume shader using the Camera Ray socket on the Light Path node\n\n\nTo Reproduce:\n\nOpen attached .blend , press render\nNote the missing shadow/light where the volume cubes intersect the default cube.\n\nI have in the .blend also included the \"expected behaviour\" by using the workaround, switch the Scene from \"Bug\" to \"Workaround\" and press render again.\nNote the shadow/light now covers the default cube properly." ]

[ "Mesh to Volume intersection artifacts\nOperating system: Ubuntu18x64, Ubuntu20x64, Win8x64.\nGraphics card: Mesa DRI Intel(R) HD Graphics (BYT) Intel Open Source Technology Center 4.2 (Core Profile) Mesa 20.0.8\n\nBroken: version: 3.0, 3.1\n\nIntersecting Mesh to Volume produces artifacts (on CPU. GPU not tested).\n\n[volume-intersection.blend](volume-intersection.blend)\n - Open,\n - Run viewport Render Preview on the left side." ]

[ "Blender 3.5.1 keeps crashing on occasion\nOperating system: Mac OS Ventura 13.3.1 (a) (22E772610a)\nGraphics card: Apple M1 Ultra (64-Core GPU)\n\nBroken: Version 3.5.1 (3.5.1 2023-04-25)\nWorked: N/A\n\nBlender occasionally crashes on my new Mac Studio (I used to work on a Windows PC beforehand and never had that problem). I'm guessing it is a saving error.\n\n- Download and open the demo-file `Blender 3.4 – “Charge” Open Movie` from demo-files/\n- Change the engine from Eevee to Cycles GPU\n- Set \"Viewport Shading\" to \"Rendered\" - nothing happens\n- Go back to Solid View - blender freezes\n\nAttached is is a video demonstrating what is happening, but in a few words:\n\nThe same happens with some of my heavy(ish) project in 3.6.0 beta\n\nIn 3.5.1 the same project responds in render view as it should be.\n\nWhen I try to reopen the same project blender doesn’t load the project but either freezes or crashes ( one of these crash logs is also attached)\nI manage to reopen the project after a few crashes.\n\nI hope it helps! \n\n\n", "Cycles crashes on viewport rendering in CUDA driver\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.0\nWorked: 3.2\n\nThe crash happens while navigating the viewport in shading mode, using Cycles.\n\nThis crash reappeared in the latest 3.3 release. It happened in older releases, then it got stable, but now it's back again.\nApparently, it has to do with Cuda.\nIt's not happening in one particular scene only, but rather randomly regardless of the complexity of the scene.\n\n[Cave-CloseUp-v006.crash.txt](Cave-CloseUp-v006.crash.txt)\n\n[Cave-CloseUp-v005.crash.txt](Cave-CloseUp-v005.crash.txt)\n\n[blender.crash.txt](blender.crash.txt)", "The version over 3.5 (3.6 and 3.6.2) crashes for me after using control of view.\nOperating system: Linux-6.2.0-32-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: Mesa Intel(R) Iris(R) Plus Graphics 650 (Kaby Lake GT3e) (KBL GT3) Intel 4.6 (Core Profile) Mesa 23.0.4-0ubuntu1~22.04.1\n\nBroken: 3.6 and 3.6.2\nWorked: version: 3.5.0\n\n\nI have quite simple scene: 1 table, 1 chair. I change selection from chair to table f.x., then deselection or I make zoom out, the screen freezes, Blender proposes me to closing because not responding\n\n", "Crash when run the script code follow\nOperating system: Mac 11.4\nGraphics card: Intel Iris 1536 MB\n\nBroken: 3.0.1(2022-01-26)\n\nBlender crashed when running the script followed in \"Scripting Window\":\n\n```\nimport bpy\n\nfor obj in bpy.data.collections['Bg_objects'].all_objects:\n obj.is_holdout = True\n```\n\nAddressSanitizer error log:\n```lines=20\n==34714==ERROR: AddressSanitizer: heap-use-after-free on address 0x00015d792c38 at pc 0x00010523645c bp 0x00016fdfa1d0 sp 0x00016fdfa1c8\nREAD of size 8 at 0x00015d792c38 thread T0\n - 0 0x105236458 in rna_iterator_listbase_next rna_access.c:4677\n - 1 0x1053cc940 in Collection_all_objects_next rna_collection_gen.c:209\n - 2 0x105234920 in RNA_property_collection_next rna_access.c:3690\n - 3 0x105f6c330 in pyrna_prop_collection_iter_next bpy_rna.c:7107\n - 4 0x11366ad40 in _PyEval_EvalFrameDefault ceval.c:4001\n - 5 0x113664fe0 in _PyEval_Vector ceval.c:5065\n - 6 0x113664f38 in PyEval_EvalCode ceval.c:1134\n - 7 0x105f1812c in python_script_exec bpy_interface_run.c:121\n - 8 0x105f18d5c in BPY_run_text bpy_interface_run.c:216\n - 9 0x10a275444 in text_run_script text_ops.c:773\n - 10 0x10a275284 in text_run_script_exec text_ops.c:812\n - 11 0x103035044 in wm_operator_invoke wm_event_system.c:1335\n - 12 0x103033cf4 in wm_operator_call_internal wm_event_system.c:1530\n - 13 0x10303ad6c in WM_operator_name_call_ptr wm_event_system.c:1577\n - 14 0x10303b978 in WM_operator_name_call_ptr_with_depends_on_cursor wm_event_system.c:1765\n - 15 0x109722ff8 in ui_apply_but_funcs_after interface_handlers.c:1008\n - 16 0x10971b1d4 in ui_handler_region_menu interface_handlers.c:11383\n - 17 0x10304ab2c in wm_handler_ui_call wm_event_system.c:698\n - 18 0x103047864 in wm_handlers_do_intern wm_event_system.c:3025\n - 19 0x1030426f4 in wm_handlers_do wm_event_system.c:3145\n - 20 0x1030406f0 in wm_event_do_handlers wm_event_system.c:3675\n - 21 0x1030043ec in WM_main wm.c:623\n - 22 0x10000ba94 in main creator.c:544\n - 23 0x1454710f0 in start+0x204 (dyld:arm64e+0x50f0)\n\n0x00015d792c38 is located 8 bytes inside of 80-byte region [0x00015d792c30,0x00015d792c80)\nfreed by thread T0 here:\n - 0 0x145a890f0 in wrap_free+0x98 (libclang_rt.asan_osx_dynamic.dylib:arm64e+0x3d0f0)\n - 1 0x1151b28e8 in MEM_lockfree_freeN mallocn_lockfree_impl.c:116\n - 2 0x114b67940 in BLI_freelistN listbase.c:463\n - 3 0x1003acbf8 in collection_object_cache_free collection.c:816\n - 4 0x1003aca74 in BKE_collection_object_cache_free collection.c:826\n - 5 0x100e4b080 in BKE_main_collection_sync_remap layer.c:1366\n - 6 0x1059bcf0c in rna_Object_hide_update rna_object.c:370\n - 7 0x105257fa8 in rna_property_update rna_access.c:2069\n - 8 0x105257bac in RNA_property_update rna_access.c:2138\n - 9 0x105f50268 in pyrna_py_to_prop bpy_rna.c:2083\n - 10 0x105f54308 in pyrna_struct_setattro bpy_rna.c:4453\n - 11 0x1135de558 in PyObject_SetAttr object.c\n - 12 0x113667dd0 in _PyEval_EvalFrameDefault ceval.c:2850\n - 13 0x113664fe0 in _PyEval_Vector ceval.c:5065\n - 14 0x113664f38 in PyEval_EvalCode ceval.c:1134\n - 15 0x105f1812c in python_script_exec bpy_interface_run.c:121\n - 16 0x105f18d5c in BPY_run_text bpy_interface_run.c:216\n - 17 0x10a275444 in text_run_script text_ops.c:773\n - 18 0x10a275284 in text_run_script_exec text_ops.c:812\n - 19 0x103035044 in wm_operator_invoke wm_event_system.c:1335\n - 20 0x103033cf4 in wm_operator_call_internal wm_event_system.c:1530\n - 21 0x10303ad6c in WM_operator_name_call_ptr wm_event_system.c:1577\n - 22 0x10303b978 in WM_operator_name_call_ptr_with_depends_on_cursor wm_event_system.c:1765\n - 23 0x109722ff8 in ui_apply_but_funcs_after interface_handlers.c:1008\n - 24 0x10971b1d4 in ui_handler_region_menu interface_handlers.c:11383\n - 25 0x10304ab2c in wm_handler_ui_call wm_event_system.c:698\n - 26 0x103047864 in wm_handlers_do_intern wm_event_system.c:3025\n - 27 0x1030426f4 in wm_handlers_do wm_event_system.c:3145\n - 28 0x1030406f0 in wm_event_do_handlers wm_event_system.c:3675\n - 29 0x1030043ec in WM_main wm.c:623\n\npreviously allocated by thread T0 here:\n - 0 0x145a89380 in wrap_calloc+0x9c (libclang_rt.asan_osx_dynamic.dylib:arm64e+0x3d380)\n - 1 0x1151b13b8 in MEM_lockfree_callocN mallocn_lockfree_impl.c:222\n - 2 0x1003b5754 in collection_object_cache_fill collection.c:754\n - 3 0x1003b5d20 in collection_object_cache_fill collection.c:775\n - 4 0x1003b53b0 in BKE_collection_object_cache_get collection.c:786\n - 5 0x1053cc5d8 in rna_Collection_all_objects_begin rna_collection.c:54\n - 6 0x1053cc2d0 in Collection_all_objects_begin rna_collection_gen.c:200\n - 7 0x105233a88 in RNA_property_collection_begin rna_access.c:3674\n - 8 0x105f6c158 in pyrna_prop_collection_iter_CreatePyObject bpy_rna.c:7076\n - 9 0x105f67ef0 in pyrna_prop_collection_iter bpy_rna.c:7083\n - 10 0x11357f400 in PyObject_GetIter abstract.c:2813\n - 11 0x11366ac34 in _PyEval_EvalFrameDefault ceval.c:3958\n - 12 0x113664fe0 in _PyEval_Vector ceval.c:5065\n - 13 0x113664f38 in PyEval_EvalCode ceval.c:1134\n - 14 0x105f1812c in python_script_exec bpy_interface_run.c:121\n - 15 0x105f18d5c in BPY_run_text bpy_interface_run.c:216\n - 16 0x10a275444 in text_run_script text_ops.c:773\n - 17 0x10a275284 in text_run_script_exec text_ops.c:812\n - 18 0x103035044 in wm_operator_invoke wm_event_system.c:1335\n - 19 0x103033cf4 in wm_operator_call_internal wm_event_system.c:1530\n - 20 0x10303ad6c in WM_operator_name_call_ptr wm_event_system.c:1577\n - 21 0x10303b978 in WM_operator_name_call_ptr_with_depends_on_cursor wm_event_system.c:1765\n - 22 0x109722ff8 in ui_apply_but_funcs_after interface_handlers.c:1008\n - 23 0x10971b1d4 in ui_handler_region_menu interface_handlers.c:11383\n - 24 0x10304ab2c in wm_handler_ui_call wm_event_system.c:698\n - 25 0x103047864 in wm_handlers_do_intern wm_event_system.c:3025\n - 26 0x1030426f4 in wm_handlers_do wm_event_system.c:3145\n - 27 0x1030406f0 in wm_event_do_handlers wm_event_system.c:3675\n - 28 0x1030043ec in WM_main wm.c:623\n - 29 0x10000ba94 in main creator.c:544\n```\n\n- Open attached file\n- Go to Scripting workspace\n- Run the above script\n[bug.zip](bug.zip)", "System UI Freeze and blender UI blinks\nOperating system: Linux-6.4.12-zen1-1-zen-x86_64-with-glibc2.38 64 Bits, WAYLAND UI\nGraphics card: Mesa Intel(R) HD Graphics 630 (KBL GT2) Intel 4.6 (Core Profile) Mesa 23.1.6-arch1.4\n\nBroken: version: 3.6.2\n\nblender often freezes (with the entire UI of the System), sometimes blender comes back to control, but, there are times when doesn't, the blender UI blinks to come entirely black and unresponsive.\n\nThis happens when I was using a graphic tablet to sculpt something.\nJust opened blender and started to sculpt, eventually the situation occur.\n\n", "Sculpt crash in certain scenario (PBVH vert number does not match)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.76\n\nBroken: version: 2.93.6\n\nCrash upon altering the model in any way.\n\n1. Open file\n2. Use a tool/brush to alter the model.\n3. Instant Crash\n\nI have updated GPU drivers and the latest Blender version as of today, still crashes.\n\nCould not find any duplicate bugs of crashes when interacting with models, only crash on startup and others. \n\nI expect my GPU is just too trash.\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[Goblin.crash.txt](Goblin.crash.txt)\n\n[Goblin.blend](Goblin.blend)\n", "Set as Background crashes blender instantly.\nOperating system: Windows 10 Pro\nGraphics card: RTX 2080 Ti and RTX 3060\n\nBroken: 3.62\n\nI'm trying to do object tracking. Have tracked an object and clicked Solve Object Motion, then when I click on Set as Background Blender crashes instantly. Trying to load in a 400 frame 1920x1080 png sequence. Blender also crashes if I try to load the background image sequence via the camera properties (Background Images > Add Image).\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n", "Blender development todo list -- Render\nNOTE: this is a direct dump of Render (without cherrypicking the valid items)\n\n### Render & UI\n\n- [ ] While rendering, manipulating images (like reload, resizing generated) can still crash Blender. Better image \"owner\" locking is required for 2.5 (#28477)\n- [ ] While rendering, several UI changes can cause crashes (#24758)\n- [ ] Rendering can't really use normal scene data if ui is interactive without causing all kinds of trouble, for examples and related issues see the item above this one.. and the reports (#26535)\n- [ ] Renaming scene leads to missing rendered result in image editor and compositor (see (#30222))\n- [ ] Rendering with File Output nodes could produce unwanted output images, see (#38605)\n- [ ] Shaded draw mode (using render code for vertex colors) is most certainly not coming back\n- [ ] Color Management can be communicated better, like when sampling colors in Image Window it confuses (could indicate such). Color Management also wasn't added backward compatible, disabling it gives dark results.\n- [ ] Compositor \"file out\" node doesn't work for FSA, and as color management troubles too.\n- [ ] FSA render uses temp files with names preventing other Blender instances to render too. (#28314)\n- [ ] Texture nodes have been implemented by recursively (backwards) calling parent nodes, this gives issues for viewers, previews and threads in general.We will try to keep it stable, but the functionality for this is going to replaced with a better shader system in 2.6. (compositor and material nodes call nodes forward, sorted on dependency. Each node is only called once this way).\n- [ ] Render display in Image window, for a window previously showing UV editor, doesn't restore correctly. (#26828). Code for this is too much spread around and fuzzy, needs full revision.\n- [ ] Object transform is not possible during rendering #37019\n- [ ] Don't render and warn when a driver or pre-render script can't be executed, a texture is missing, physics are not baked. Needed e.g. when renderfarms have scripts blocked.\n\n### Node Textures\n- [ ] Grouping texture nodes affects on data range #45696\n\n### Baking\n- [ ] add option to smooth vertex colors as a post process after baking #35066.\n- [ ] Baking to a low-res textures might miss pixels which lies on the UV island boundary #37681\n- [ ] Vertex baking might cause AO artifacts #44178\n- [ ] Support dupligroups #43553\n- [ ] Modifier produces non deterministic effects #42015\n- [ ] Update Blender side baking to support different topologies needed for adaptive subdivision #49405", "Crash when switching from solid to rendered view\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.6.0 Alpha\n\nI have a scene about 1.7GB in size with 4K textures but renders in 2K using simplify, blender crashes when I open the file > hit F12 and render the scene then close the render windows and switch to render view > it starts compiling then blender closes, in the past it would show out of memory this report: \n\n105522#issuecomment-904934 \n\nThis issue happens only when persistent Data is active\n\nunfortunately I can't share the file at the moment. \n\n\n\n\n", "Latest Blender 4.0 not opening\nOperating system: Linux-5.15.0-83-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: GeForce GT 630/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.157\n\nBroken: version: blender-4.0.0-alpha+main.b5c89822ac40-linux.x86_64-release\nWorked: 3.6.3 Release Candidate\n\nAfter double-clicking on the \"blender\" starter-folder, it shows what you can see on the picture, before canceling the session\n\nJust try to start 4.0. I guess it's my drivers and noone can confirm, right?\n\n", "AMD - Blender Crashes When I Start Render - Cycles GPU Compute\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: AMD Radeon RX 6800 XT ATI Technologies Inc. 4.5.0 Core Profile Context 23.7.1.230626\n\nBroken: version: 3.6.1\n\n[Whenever I start Rendering, It crashes when using cycles on GPU compute. Doesn't happen on CPU option.]\n\n[On render engine, select cycles. Then set it to GPU Compute. Then click on render on top left corner. Then select render image.]\n\n", "Blender crash when selecting and moving vertex using API\nOperating system: Linux Mint 20 Ulyana\nGraphics card: NVIDIA GeForce RTX 2060 Super\n\nBroken: 2.91.2, 5be9ef417703, master, 2021-01-19\n\n\nBlender crashes when opening a new instance pasting the below script in a new document in the \"Scripting\" editor and clicking play:\n\n```\n\nimport bpy\n\ncontext = bpy.context\n\n# Assume default cube exists and select it\nobj = bpy.data.objects[\"Cube\"]\nobj.select_set(state=True)\ncontext.view_layer.objects.active = obj\n\n# Select a vertex and move it\nobj = bpy.context.active_object\nobj.data.vertices[0].select = True\nbpy.ops.object.mode_set(mode = 'EDIT')\nbpy.ops.transform.translate('INVOKE_DEFAULT',\n constraint_axis=(True, True, False),\n orient_type='GLOBAL',\n release_confirm=True)\nbpy.ops.object.mode_set(mode='OBJECT')\n```\n", "Cycles is not working in 3.5.1 on a MacBook M1 Max—causes the program to immediately freeze!\nOperating system: macOS-13.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Max 1.2\n\nBroken: version: 3.5.1\nWorked: (This started happening two days ago, and I only updated to 3.5.1 today. But in the days/weeks prior it was 3.4)\n\nOn my MacBook M1 Max (with all upgrades possible), I have had to force quit Blender countless times in the last 48 hours after turning on viewport Cycles view. Upgrading to 3.5.1 has seemingly made it a constant problem. Even now opening a brand new project, then turning on Cycles render, it starts saying it’s loading kernels (very slowly), and never renders, freezing the viewport, forcing me to force quit. So it seems I no longer can turn on Cycles in 3.5.1.\n\nThere is no further steps really, this is happening on a new project for me. Hopefully you all can see this and figure out something! I’m happy to provide further details if you need anything. I use Blender for my day-to-day at work so fingers crossed this isn’t the new normal for the foreseeable future. Also, love you all!\n\n", "Blender 2.8 and 2.81 (ab519b91b2c4) has high chance of crashing when using Python to create a new node-tree link.\nOperating system: Windows 10 (x64) v1903 Build 18362.356\nGraphics card: Nvidia GeForce GTX 1060 6GB\nCPU: Intel Core i5 7600K\n\nBroken: 2.81-ab519b91b2c4 **AND** 2.80 release\n\nCreating a new node-tree link via Python sometimes causes Blender to crash with the error \"EXCEPTION_ACCESS_VIOLATION\".\nI have attached a .blend file that will cause the crash, just by clicking \"Run Script\" once or twice. \nWhen I found the bug, I spent some time making the script as small as possible while still causing the crash, so the script is less than 30 lines.\n\nI also installed Blender 2.81-ab519b91b2c4 on a Windows 10 laptop that had never seen Blender before, and the bug was present on there too, so I'm confident the issue isn't specifically due to my hardware.\nI've also included blender_debug_output.txt and blender_system_info.txt, generated by the blender_debug_log.cmd file included in the Blender 2.81 download.\n\n1) Open the attached .blend file.\n2) Click \"Run Script\" in the bottom right of the text editor pane.\n3) Repeat 2 until crash (should only take 1 or 2 tries)\n\n[Crashy crash.blend](Crashy_crash.blend)\n[blender_system_info.txt](blender_system_info.txt)\n[blender_debug_output.txt](blender_debug_output.txt)", "Blender crashes when using an EnumProperty to set Space\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.0 Alpha\n\nIn this script, I'm using `context.area.type` with a property to change the Space type. It caused Blender to crash.\n\n1. Running this script from the Text Editor (below)\n2. Switch to Graph Editor or Dope Sheet\n3. Click new property buttons on right of Header\n\n```\nimport bpy\nfrom bpy.types import PropertyGroup\nfrom bpy.props import EnumProperty\n\ndef set_space(self,value):\n context = bpy.context\n\n for area in context.screen.areas:\n if area != context.area:\n break\n\n override = {'region': area.regions[0]}\n\n if value == 0:\n context.area.type = \"GRAPH_EDITOR\"\n elif value == 1:\n context.area.type = \"DOPESHEET_EDITOR\"\n\nclass SWITCH_PG_switcher_props(PropertyGroup):\n space: EnumProperty(\n name=\"Space Target\",\n description=\"Switch to space\",\n items=(\n ('GRAPH_EDITOR', \"\", \"Graph Editor\", \"GRAPH\", 0),\n ('DOPESHEET_EDITOR', \"\", \"Dopesheet Editor\", \"ACTION\", 1),\n ),\n default='GRAPH_EDITOR',\n set = set_space\n )\n\ndef draw_enum(self, context):\n layout = self.layout\n scene = context.scene\n layout.prop(scene.switcher, \"space\", icon_only=True, expand=True)\n\ndef register():\n bpy.utils.register_class(SWITCH_PG_switcher_props)\n bpy.types.Scene.switcher = bpy.props.PointerProperty(type=SWITCH_PG_switcher_props)\n bpy.types.DOPESHEET_HT_header.append(draw_enum)\n bpy.types.GRAPH_HT_header.append(draw_enum)\n\ndef unregister():\n del bpy.types.Scene.switcher\n bpy.types.DOPESHEET_HT_header.remove(draw_enum)\n bpy.types.GRAPH_HT_header.remove(draw_enum)\n bpy.utils.unregister_class(SWITCH_PG_switcher_props)\n\nif __name__ == \"__main__\":\n register()\n```\n\nSetting the `context.area.type` in Python seems to cause others' scripts to crash:\n[Performing a Translate Operation after switching the context area crashes Blender ](1164763)\n" ]

[ "Blender Crash -on Launch- / accessing System tab / Cycles Render mode\nWindows 10\nNVidia Geforce 10 GTX 1050\nIntel i7-7700HQ\n\nBroken: 2.79b\nWorked: 2.79b from the Blender Builder\n\nLast I used Blender a couple weeks/months back and it worked, now when launching blender it immediately crashes. Sometimes it takes a second to crash, but there's never any UI. I can't think of a reason other than Windows updates or some driver, but would that not crash the Blender Builder version (which works)?\n\n**Update and Summary:**\nAfter moving out igdrclneo64.dll and igdrcl64.dll from the system32 folder the instant crashes stopped, same effect as using blender in compatibility mode. Only if I move nvopencl.dll and nvcuda.dll out of system32 can i use 3d programs. That disables all CUDA functions though. The error message I got was can't access nvcuda.dll and nvopencl.dll\n\n**Final Update and Summary**\nThe root of the issue is some Error with Windows, likely to be introduced by a software update. It causes all Graphics Drivers to not function correctly - CUDA and OpenCL become unavailable as well.\nThe only way I was able to solve it was to reinstall Windows.\n\n\n\n\n\nI can't give exact steps to recreate it.\n\nThings I tried:\n- sampled some older Nvidia Drivers\n- Used Intel Integrated Graphics to launch. (didn't work)\n- Updating all my drivers.\n- Using some sort of blender.bat file from another forum. I didn't know how to do the step of linking .blend.bat files though.\n- Using the blender builder version, which worked - the installed version still doesn't. What does this mean? (Update: it means blender builder doesn't use CUDA)" ]

[ "Modifiers added in edit mode are undone out of order if you go back to object mode\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.66\n\nBroken: version: 2.83 (sub 13)\nWorked: Never (2.8+)\n\nModifiers added in edit mode are undone out of order if you go back to object mode and then start undoing.\n\nAffects both old and new undo system. There's often mentions of edit<->object transitions being problematic but just in passing; I think this is a long standing issue. Just wanted to note it down as I test out the undo systems.\n\nHoping the new undo system can account for this now.\n\n- Default scene\n- Go to edit mode on the default cube\n- Add any modifier (bevel, subdivision, etc.)\n- Add any loop cut 1\n- Add another loop cut 2\n\n- Go to object mode\n- Undo --> Modifier is removed first (out of order)\n- Undo --> Loop cut 2 is removed\n- Undo --> Loop cut 1 is removed\n\n\n", "2.93 texture coordinate display bug\nOperating system: win10\nGraphics card: 2070 super\n\nBroken: 2.93.0-321eef6a0c0f 4/2\n\n\n\ntexture coordinate display bug differently in viewport compared to render\n\ni used a generated coordinate. in the viewport it seems that blender applied the coordinate after the curve modifier (maybe)\n\nbut in render it will apply the coordinate before the curve modifier, so the gradient is able to follow the curve which is my desired look.\n\nsometimes it will display the coordinate correctly but I don't know how I achieved it\n[bug.blend](bug.blend)\n\n", "Move cache visualization to draw overlay\nCurrenlty the cache visualization isn't updated during playback, only when the screen is redrawn.\nThis needs an optimization of the drawing of the cache lines.", "When you clear measurements through View > Annotations, they don't vanish immediately\nOperating system: Windows\nGraphics card: GTX-2080\n\nBroken: 3.4.1\n\n\nDeleting measurement annotations appears to be broken because the UI is not refreshed until you change tool.\n\n\n1. Create some measurements\n2. Open View settings Tab\n3. Remove ruler Annotations\n\nExpected behavior: Measurements are removed and action can be undone.\n\nCurrent behavior: The Annotation is no longer visible in the annotation list, but still shown in edit view.\n\nHere is a screen recording: 185530\n\n", "Broken simulation nodes cache indicator in timeline\nOperating system: ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.6 Beta\n\nWhen Simuation nodes caching is disabled, indicator from timeline is not disappear, and behaves incorrectly:\n1. Under the widget of the current frame there is a piece of the indicator \n2. when moving the current frame to the left, a piece of the indicator remains in place, and after the operation is terminated, it returns under the widget of the current frame\n[2023-05-20 17-30-40.mp4](attachment)\n\n1. Open blend file\n2. perform the operations specified in the description\n\n", "Image sequence playhead for particle system textures never advances past first frame of sequence when particles are baked\nOperating system: Windows 10/64\nGraphics card: NVIDIA GeForce GTX 1060\n\nBroken: 2.93, 3.0.0\nWorked: n/a\n\nWhen I create a particle system that relies on an image texture containing an image sequence to control density, the playhead for that image sequence remains \"stuck\" on the first frame of the image sequence when I bake the particle system, regardless of the output length of the composition. This is true regardless of what type of image sequence is used (i.e. frames in a container such as .mp4 or .mov, or sequences of images such as .exr, .tif, or .png) or what playback options for the sequence are set (i.e. Cyclic and Auto Refresh can both be checked; it won't matter).\nRenders of an un-baked particle system using a moving texture to control particle density in Cycles are unusable.\nRenders of an un-baked particle system using a moving texture to control particle density in Eevee are occasionally usable (although unreliable: one must make absolutely certain that no frames are cached before attempting a render).\n\n**Expected behavior**\nMy expectation is that the playhead for an image sequence using a texture to control particle density would advance when the particle system is baked. It should appear both in stage preview and final rendered output that the source of particle emission appears to move, which is to say it should accurately track the black and white areas in the image sequence.\n\nCreate an empty Blender stage;\nAdd a plane to the stage;\nCreate a texture based on any black-and-white-only image sequence (i.e. a white circle whose position changes with time, superimposed on a black frame); the sequence can be either containered-image or discrete-image;\nCreate a particle system for the plane;\nAssign the texture to the particle system, leaving only the \"Density\" influence in the texture checked;\nMove the stage playhead to frame 0 of the composition;\nClick \"Delete All Bakes\" to clear any baked frames, also change number of particles in order to delete frames in memory;\nClick Play on the stage;\nHalo objects should be being emitted from the white areas in the texture as they move around in the texture (the emitter will appear to change position);\nMove the stage playhead back to frame 0 of the composition;\nClick \"Delete All Bakes\" to clear any baked frames;\nClick \"Bake All Dynamics\" to bake the particles;\nClick Play on the stage;\n\nSteps done in following file:\n[particle-density-demo.rar](particle-density-demo.rar)\n\nBoth on the stage and in rendered output, particles will be emitted only from the location on the plane that corresponds to the white area as it was on the first frame of the image sequence.\nAnd if someone could show me where in the code this stuff is handled, I'd be more than happy to take a look.", "Performance: EditMesh to Mesh on every modifier run\nThis is a part of performance regression investigation task.\n\nEvery movement of vertex will request modifier stack to be re-run. This involves conversion of EditMesh to Mesh, no matter whether there are modifiers or not.\n\nSince 84cf670d1d it seems that we can simply delay creation of `mesh_cage` in `editbmesh_calc_modifiers` and only do it if there is actual modifier which is to be applied. We can then have `runtime.is_original` set to truth for all Mesh IDs which are covered by Copy-on-Write, and set it to false when modifier runs.\n\nThis will solve about 50% of slowdown in the attached file (where EditMesh to Mesh takes almost same amount of time as batch cache creation).\n\n[megaheavy_mesh_edit.blend](megaheavy_mesh_edit.blend)", "Cache loading via animation system\nCurrently the Cache Sequence modifier is used for loading geometry data from Alembic files, and the Transform Cache constraint is used for loading transformation data. This approach poses a few limitations:\n- A modifier is either generative or deforming, whereas the Cache Sequence modifier is both (depending on the particular frame-to-frame changes in the Alembic file).\n- Non-mesh geometry data (like curves) are troublesome to load, as the modifier system can only produce an evaluated mesh (#51311).\n- Non-geometry, non-transform properties, if loaded at all, are loaded at import time, but after that are static (#54050, #50725). This can be solved by having per-property F-curve modifiers ([D2324](D2324)), but this is cumbersome and unintuitive to handle.\n\nThis task proposes a new approach, adding a 'cache' step to the `AnimData` struct. If set, cached data would be loaded by the animation system first, before applying F-curves and drivers as usual. This would allow the cache loader to influence properties in a more general way:\n- It's a new system, so no pre-existing separation between 'generative' and 'deforming'.\n- Any ID datablock that supports animation can use this, so no distinction between mesh and non-mesh geometry.\n- Any property can be animated.\n\nI'll study the current animation system, and then update this task with more technical details on how to approach this.\n", "Geometry Nodes: simulation crashes when \"Cache\" option is disabled\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: AMD Radeon(TM) Graphics ATI Technologies Inc. 4.5.14802 Core Profile Context 21.40.64.06 30.0.14064.6002\n\nBroken: version: 4.0.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `Unknown`\n\nDisabled \"Cache\" option causes a crash when jumping back to start frame.\n\nAdd a simulation zone, output the geometry result.\n\nGo to the Physics tab and disable the __Cache__ option in the Simulation Nodes\n\nSimulation a few frames, then jump back to the start frame.\n\n", "Particle instance modifier can crash when create from Children is checked and Normal is not\nOSX 10.11.6, xeon w3680 6core 12thread \n[system-info.txt](system-info.txt)\n2.69.7 5bd8ac9\nalso Nightly 2.79 34fe3f9c069\n\n\nUsing a particle instance modifier with Children checked and Normal not checked and then attempting to change certain parameters of the particle system such as in the velocity or children tabs causes a crash. Enabling Normal in the particle instance modifier prevents this from happening.\nCreate a Plane with a hair type particle system.\nCreate a cube with a particle instance modifier and set the object to the Plane with the particle system.\nCheck the box for Children and uncheck the box for Normal.\nGo to the particle system on the Plane and check Advanced. \nDrag any of the values in Velocity or Children and soon it will crash.\n[Crash_ParticleInstance.zip](Crash_ParticleInstance.zip)", "Joining meshes or realizing instances in geometry nodes removes custom normals data\nOperating system: macOS Big Sur 11.5.2\nGraphics card: Apple M1\n\nBroken: 2.93 and 3.0 Alpha\n\n\nUsing the collection info node in geometry nodes destroys custom normals data when the geometry node modifier is applied (or when the project is exported).\n\n\nApply the geometry node modifier on the \"Dummy\" object in the attached blender project. I've also included a video that shows the issue.[collection_info_test.blend](collection_info_test.blend)\n\n[collection_info_bug.mov](collection_info_bug.mov)", "Undo Makes Completed Renders Disappear From View Render Window Slots\nOperating system: Windows 10 Build 19043\nGraphics card: Nvidia RTX 3090\n\nBroken: 2.92\nWorked: Can't remember, but I don't recall this happening before\n\nUndo makes completed renders disappear from other slots in the View Render window. Some of them are preserved, but the more you undo, the more disappear from other slots.\n\n1. Load default scene (with the cube)\n2. Render scene\n3. Go to the cube's shader and change its base color\n4. Switch to next unoccupied slot in the View Render window\n5. Render scene\n6. (optional) Repeat steps 3-5 as many times as you like\n7. Go to View Render window and scroll through slots to verify all renders are there (with J key)\n8. Activate undo (ctrl+z)\n9. Go to View Render window and scroll through slots. Some will be missing.\n10. Activate undo again and more will be missing.\n\nThis doesn't seem like intended behavior, and I've replicated this issue on a different machine. Hope this helps!\n\nEdit: I've replicated this issue in 2.91, 2.90 and the latest LTS release. I assume intended behavior is for undo to change the shaders back, but for renders you've completed to remain in their slots.", "Some mesh operations (e.g. smooth, randomize) are not using mirror modifier mirror object transform correctly for clipping\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 950M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 3.0.1 Release Candidate\n\nWhen using mirror modify with clipping, Smooth Vertices ignore clipping and smoothing all, and this is no options\n\n\nTools to check:\n`Transform` > `Randomize`: does not clip (even without mirror object usage)\n`Transform` > `Warp`: does not clip (even without mirror object usage)\n`MESH_OT_vertices_smooth`\n\nnote: all of the above are not part of the transform system (randomize is even written in python), so it might be a bit cumbersome to implement", "Clicking \"Bake All Dynamics\" causing various crashes.\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\n\nUsing \"Bake All Dynamics\" is beyond finicky and crashes Blender in many different use cases. Many little \"tricks\" are needed to make it work. I am using Disk Cache for all below situations, but I believe the crashes happen with or without this.\n\nSituations where it crashes:\nClicking \"Bake All Dynamics\" within the Cloth cache settings. It seems it's required to click \"Bake All Dynamics\" button within the context of the Rigid Body system, or else it crashes instantly.\n\nClicking \"Bake All Dynamics\" in the context of hair dynamics seems to do nothing. Same goes for \"Update All to Frame\"\n\nIf a rigid body simulation has been baked, clicking \"Delete All Bakes\", then clicking \"Bake All Dynamics\" immediately will also crash blender. The only workaround for this is to \"move\" an object even if using a zero value, before clicking \"Bake All Dynamics\" again. This prevents the crash. \n\n\nCreate a Rigid Body Sim, create a Cloth Sim. Check \"Disk Cache\" as an option. Try the various methods listed above.\n\n", "Overlay: Edit-Mode: Anti-Alliasing not working with constructive modifiers\n```\nOperating system: Linux-4.15.0-50-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce 610M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.116\n\n\nBroken: version: 2.80 (sub 69)\nWorked: (optional)\n\n```\n\nthe bug is really intresting so... when you add mirror modifier to object.. and u enter edit mode and u select face or edges etc... the AA goes crazy it is so pixelised like even worse than tottaly without AA.\n\n\n\n\n\n\n\n" ]

[ "Modifiers after physics ones shouldn't invalidate cache\nOperating system: macOS-10.14.6-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 560 OpenGL Engine ATI Technologies Inc. 4.1 ATI-2.11.21\n\nBroken: version: 2.93.0\n\nIn current Blender, any change on modifiers invalidates.\nModifiers *after* physics modifiers won't affect the result of physical simulation, therefore they shouldn't invalidate the cache. \n\n\n1. Create a plane\n2. Give it Cloth and Subdivision modifiers, in this order\n3. Set the end frame to 10\n4. Play animation until the end\n5. Stop animation\n\nNow the simulation is cached.\n\n6. Change the viewport division level on Subdivision modifier\n\nNow the cache is invalidated.\n\n" ]

[ "Regression: GPU Subdivision: EEVEE doesn't refresh viewport correctly with Armature modifier\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3LTS, master\nWorked: 3.2.2\n\nGPU Subdivision doestn't refresh viewport correctly an object when Armature and Subdivision Surface Modifiers are used.\n\n1) Start Blender.\n2) Add Armature object\n3) Switch to \"Shading\" workspace.\n4) Select the Cube and add `armature modifier` then `Subdivision Surface Modifier`.\n5) Switch to Properties Editor to \"Modifier Properties\"\n6) Add armature in armature modifier's object property. \n7) Switch between object-edit mode\nIt works fine when \"GPU Subdivion option\" is turned OFF (Prefrences > Viewport > Subdivision > GPU Subdivision)\n\nI added \"BeforeGPUSubDBug.blend\". You can reproduce this issue by loading the blend file and doing step 7)\n", "Cycles crashes on viewport rendering in CUDA driver\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.0\nWorked: 3.2\n\nThe crash happens while navigating the viewport in shading mode, using Cycles.\n\nThis crash reappeared in the latest 3.3 release. It happened in older releases, then it got stable, but now it's back again.\nApparently, it has to do with Cuda.\nIt's not happening in one particular scene only, but rather randomly regardless of the complexity of the scene.\n\n[Cave-CloseUp-v006.crash.txt](Cave-CloseUp-v006.crash.txt)\n\n[Cave-CloseUp-v005.crash.txt](Cave-CloseUp-v005.crash.txt)\n\n[blender.crash.txt](blender.crash.txt)", "Regression: GPU subdivision - Mesh does not immediately count viewport statistics\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.14800 Core Profile Context 22.3.1 30.0.15002.1004\n\nBroken: version: 3.3.0 Alpha\nWorked: blender-3.3.0-alpha+daily.`d095fcd6b417`\n\nI didn't do rigorous bisecting but I suspected this is caused by 163f6a17e4 so I tested with a daily build before it was committed.\n\nMesh does not immediately count in viewport statistics with GPU subdivision enabled.\n\n\n\n- Open startup file.\n- Add Subdivision Surface modifier to the cube\n- See the mesh is not counted towards viewport stats\n- Deselect the cube (or do some other action related to viewport display, for example, ticking viewport display options), and see now it is counted towards viewport stats\n", "Lattices don't display selected vertices with modifier \"Edit Mode\" enabled\nSince 2.8x, lattices have not been displaying selected vertices when the modifier option \"Edit Mode\" id enabled.\n\nThis is easy to test:\n\n- add a new lattice object\n- add a wave deform modifier.\n- enter edit-mode.\n- enable \"Edit Mode\" toggle for the modifier.\n\nThis test file shows the problem too, open it and toggle the \"Edit Mode\" modifier button.\n\n[#77184.blend](T77184.blend)\n\n----\n\nReporting this to keep track of the issue: `BKE_lattice_modifiers_calc` is making a copy of the lattice `lt_eval` which doesn't contain an edit-lattice, which is then used for drawing.", "Fake Dependency Cycle: 'Disable in Viewport' depends on 'Disable in Renders'\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.2.2\n\nDriving hide_render from hide_viewport causes a dependency cycle\nwhile the error occurs on all objects, the driver does seem to work on some objects, but not others. In the default scene for example, the driver works with the camera, but not the cube or the light\n\n- in the outliner, edit the filter to display the `Disable in Viewport` option.\n- Right click the `Disable in Viewports` icon on any given object\n- Copy as New Driver\n- Right click the `Disable in Render` icon of the same object\n- Paste Driver\n[render visibility driver issue.blend](render_visibility_driver_issue.blend)\n\n", "OpenEXR can't be used to save results during the render process\nOperating system: XUbuntu 20.04\nGraphics card: Not applicable\n\nBroken: Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nWorked: Blender 2.93.7\n\nSeems OpenEXR now can't be used to save intermediate results during rendering using Cycles. The other formats (like PNG) seems works just fine. Blender starting v3 just crashing if you trying to save as single or multi-layer EXR. When the render is completed the `save_render` function is working correctly.\n\nThis bug doesn't allow to prepare the proper distributed preview for user outside of Blender UI (EXR needed to merge multiple preview samples from the remote workers).\n\nThe default cube blender scene is just enough to reproduce. Please switch to script console and paste the next code:\n```\nbpy.context.scene.render.engine = 'CYCLES'\nbpy.context.scene.render.image_settings.file_format = 'OPEN_EXR' # OPEN_EXR_MULTILAYER here will fail too, but PNG for example will work just fine.\nbpy.context.scene.cycles.use_adaptive_sampling = False\nbpy.context.scene.cycles.samples = 512\n\nbpy.ops.render.render('INVOKE_DEFAULT', use_viewport=True)\n\nimport time\ntime.sleep(2)\nbpy.data.images['Render Result'].save_render('_preview.exr')\n\n### Crash will be right here ###\n\nimport os\nprint(\"Preview file exists:\", os.path.exists('_preview.exr'))\n```\n\nCrash report: [blender.crash.txt](blender.crash.txt)\n", "Blender Wont run.\nOperating system: Windows\nGraphics card: AMD Radeon HD 8570\n\nBroken: 2.92\nWorked: Nothing has worked. But blender worked before I did this. \n\nWhen I open blender it closes immediately.\nAnd I get this \n\nwatch?v=3rugIsYwl7s).\nSo Out internet was down, And I wanted to get blender On my other pc account, So I went to the main account And Thinking that the installer required internet, I took the system files, and copied them to a thumb drive and pasted them Into the other account when I did that this first started happening, so when The internet came back On. I just got rid of the files I pasted in and then Installed it. With the same result as last time. But this time the bug/glitch was on all accounts. and I tried uninstalling and reinstalling it from the apps and features and from the control panel. I also installed the app from the windows app store with no luck. If you guys know how to fix this issue then please email me at [email protected]\n Ps, My CPU wasn't hot.\n- - -\nAble to run blender from `blender_debug_gpu_glitchworkaround.cmd` batch file", "AMD GPU Denoising via \"Radeon Image Filters\" library.\n**Overview**\nThis task is to enable GPU denoising on AMD GPUs. We have a denoising library (RadeonImageFilter) publically available which operates similarly to OIDN, Optix, etc. We would like to use this to enable faster GPU denoising for users on AMD hardware. It should be noted, this library actually works on other GPUs and CPU (the latter via OIDN) but on Nvidia GPUs, the Optix denoiser will most likely be more optimal.\n\n**GPL compliance**\nWe offer the library free under a permissive Apache2 license at the link above. However, it is not open source. A shortly upcoming version of the AMD driver on windows will include this RIF library. Thus we propose to use it similarly as Optix, as a closed source \"system device\" library exception to the GPL.\n\n**Code design** \nWe added a new denoiser type, AMD Radeon Image Filter, with similar options to Optix Denoising to enable usage.\n\nPatch coming soon.", "Grid doesnt show\nOperating system:\nGraphics card:\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n", "OptiX Denoiser fails with large output resolution (final rendering completely black)\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: version: 3.4.1\n\nDuring my final rendering, in 100% resolution, the final rendering is completely black. When calculating the image everything looks normal, I can see the tiles generating normally. Then the final render appears black. The other layers (shadow, OA, etc.) are well rendered. \n\nThis looks to be caused by OptiX failing to denoise a very large final image.\n\nThe example here will attempt to render a 6144x6144px image on an 8gb 1070 and fail during the Denoising step (leaving a completely black image too).\n\nThe original report attempted to render at >10240px (so adjust accordingly if it doesn't happen at first with your GPU).\n\n- Load attached `denoise.blend` file.\n- Attempt to F12 render\n\n```\nI0301 12:53:47.768775 15952 path_trace.cpp:971] Processing full frame buffer file T:\\Temp\\blender_a18956\\cycles-tile-buffer-15716-2377920588664-0-0.exr\nI0301 12:53:49.619309 15952 denoiser_gpu.cpp:51] Creating temporary buffer on denoiser device.\nI0301 12:53:50.388029 15952 device.cpp:669] Buffer allocate: RenderBuffers, 2,868,903,936 bytes. (2.67G) in device memory\nI0301 12:53:50.995292 15952 device_impl.cpp:80] using cuda device \"NVIDIA GeForce GTX 1070 with Max-Q Design\" (6.1), buffers: fp16, Winograd convolution, rt v11060\nI0301 12:53:52.356184 15952 device.cpp:669] Buffer allocate: __denoiser_state, 2,999,350,016 bytes. (2.79G) in host memory\nI0301 12:53:52.356184 15952 device_impl.cpp:80] layers created for resolution 4352 4352, inp 8, outp 3\nE0301 12:53:52.356184 15952 device_impl.cpp:74] copy of converted weights failed (CUDA error string: invalid argument, CUDA error code: 1)\nFailed to set up OptiX denoiser\n\nRefer to the Cycles GPU rendering documentation for possible solutions:\ngpu_rendering.html\n\nE0301 12:53:52.356184 15952 denoiser_optix.cpp:577] OptiX denoiser configuration has failed.\n\n```\n\nVerbose 5 logs attached.\n\n", "Some particles are not being rendered during the first render\nOperating system: Windows 10\nGraphics card: I tried on two different computers with different GPUs - Nvidia RTX3070 and Mobile Nvidia RTX2060, same behaviour.\n\nBroken: 2.93.6, 3.0.0 Release Candidate, 3.1.0 Alpha\n\n\nBlender is not rendering some particles (in my case this is a grass) during the first render.\nIt does after re-rendering the scene, but this is quite confusing and inconvenient, especially in the moments when I have to resume render of animation.\nApparently this only happens when we add a force field to control the particles.\n\nOn top of this, 3x does not render trees (particles).\n\nOpen the blend file and press F12 to render. You should have similar result as on the screenshot:\n\n\n\nRender again, and the grass will appear:\n\n\n\nI attach results also from 3.0:\n\n\n\n\n\nand 3.1:\n\n\n\n\n\n- Open attached file\n- Render (F12) - See the particles issue\n- Render again - Now it is ok\n\n[turbulence_particle_eevee_bug.blend](turbulence_particle_eevee_bug.blend)\n\n---\nOriginal File:\nBecause file is big (260M) I have decided to share the link: view?usp=sharing <- if you would prefer me to upload it anyway here, please just let me know.", "stack_overflow exception while rendering with viewport render on\nOperating system: win10\nGraphics card: 2070super\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: 2.83.4, 2.91 6fe609d4c3\n\n\nwhen pressing f12 in render mode, blender crashes,\nwhen pressing render in solid mode, blender doesnt crash\n\n[blender_debug_output.txt](blender_debug_output.txt)\n```\nError : EXCEPTION_STACK_OVERFLOW\nAddress : 0x00007FFF42963FDB\nModule : C:\\Windows\\system32\\DriverStore\\FileRepository\\nv_dispsi.inf_amd64_6af46268c19e4980\\nvcuda64.dll\n```\n[2.blend](2.blend) \n1. switch to render mode\n2. press f12 & watch task manager.\n3. blender will close", "Vulkan: Better handling of GPU debug when not having vulkan SDK installed.\nBlender libraries only contain a subset of the vulkan SDK. Most important is\nthat validation layers aren't added. These validation layers will be enabled\nwhen starting blender with `--debug-gpu` (or test cases).\n\nIn this case there is an error printed to the console. This should not be an\nerror, but we could add a log line when not found.\n\n", "16 bit transparent Image renders color data on CPU Cycles\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.02\nCPU: Intel i7-4770k\n\nBroken: version: 2.81 (sub 12)\nWorked: (optional)\n\nRendering a transparent Filmic Log 16bit PNG image plane (meaning node color space set to filmic log) with Cycles the CPU renders color data where there should not be any. The GPU works as expected.\n\nevery thing with Cycles (have not tested eevee)\n1) render something with transparent background and save as PNG (RGBA, 16bit)\n2) import image as plane ( shadeless )\n3) set image node color space to Filmic log and background to transparent \n4) rendering with CPU or CPU+GPU results in the tiles rendered by the CPU to have color data (see image)\n\nSwitch Display Channels to color to see this issue\n\n[test.blend](test.blend)\n\nRedner: \n\n- on the left is no image plane, only the transparent background\n\nUsed image:\n\n\n", "No error prompted when Cuda fails during GPU bake\nOperating system: Linux Debian 10 (buster) - kernel 4.19.0-19-amd64\nGraphics card: GeForce GTX 1060 Mobile\n\nBroken: 3.1.0\nWorked: -\n\nWhen you try to bake, having GPU set as compute device, and CUDA has an error, no error is shown.\nYou only notice the error when you try to render and get \"*Failed to create CUDA context (unknown error)*\"\n\nIt is useful to show an error, in the same way that, for example, \"Circular dependency for image...\" is shown during the bake.\n\nNote that this report is not about the error itself. It is about Blender not reporting any errors during baking, which causes user confusion.\n\nThank you,\nRiccardo" ]

[ "RTX 3090 not working with cycles rendering in blender 2.83 LTS (Also checked with the latest version of 2.83 LTS)\nOperating system: Windows 10\nGraphics card: RTX 3090\n\n\nBroken: Blender 2.83 LTS (also checked with latest 2.83 LTS version)\nWorked: 2.91\n\nRTX 3090 graphic card rendering issue with blender 2.83 LTS. I am trying to use RTX 3090 graphic card with blender 2.83 LTS. I have also tried the latest version of LTS - 2.83.12 but it is still not working. I am getting following error while rendering with cycles - \n\nCUDA binary kernel for this graphics card compute capability (8.6) not found\n\nHowever it is working fine with blender 2.91\n\n\nUse blender 2.83lts with RTX 3090 graphic card", "Cycles render engine not working with RTX 3070\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.83.5\n\nAs I stated in the title, Cycles rendering engine is not producing a rendered output when using GPU compute. I am using an RTX 3070 graphics card with the latest drivers installed.\n\nGPU compute on. Doesn't matter if it's CUDA or OptiX, same outcome.\n\nCycles with GPU computing on:\n\n\nCycles with CPU computing on:\n\n\nMy preferences:\n\n\n\n", "Gpu render doesn't work\nOperating system: Windows 10 pro\nGraphics card: RTX 3070\nCPU : AMD Ryzen 9 3900xt\nRAM: 64gb\n\nBlender version: 2.83\n\nWhen i set render engin to cycles and device to GPU Compute, my objects in the viewport are grey. Cycles works on CPU device. I have already ativated my graphics card in the system preferences under CUDA. Blender instaltion is brand new, hardware is brand new. I think the problem ocured when i installed Nvidia's studio drivers.\n\n" ]

[ "Bisect Tool rotation handle does not work from some preset orthographic views\nOperating system: Windows 10\nGraphics card: RTX 2070\n\nBroken: 2.91.2\n\n\nBisect Tool rotation handle does not work from some preset orthographic views.\n\n* Select default cube\n* Enter edit mode\n* Press numpad 1 to enter front orthographic view\n* Select all verts\n* Click bisect tool\n* Draw line across cube\n* Click circle widget to rotate bisection <- fails\n\nWorks fine from top orthographic view and from arbitrarily orientated perspective views.", "Autokey keys disabled driver values\nOperating system: Darwin-17.2.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-10.27.6 378.10.10.10.20.109\n\nBroken: version: 2.83 (sub 2)\nWorked: (optional)\n\nWhen you have autokeying turned on for LocRotScale, then disable a driver (for say a shape key) and then manually tweak the value. Blender will key this change in the driver editor.\n\n- turn on autokey (keying set LocRotScale will do)\n- create a property on the default cube\n- copy as new driver\n- create a shape key\n- paste this driver to this shape key\n- then disable the curve for this driver in the driver editor (check box)\n- manually tweak the driver value and this will be keyed in the driver editor.\n\n[autokey_disabled_drivers.blend](autokey_disabled_drivers.blend)\n\n[auto_keying_disabled_drivers.mp4](auto_keying_disabled_drivers.mp4)\n\n", "Solidify Node\nWe will drop support for Simple mode since the difference is hard to tell and performance-wise better to optimize the Complex mode than maintaining both.\n\nFill and Only options are a bit confusing in the modifier since fill refers to the fill of the rim and has to be checked on to make Rim Only work.\nInstead, I suggest making Fill refer to the part that is not the rim and Rim refer to the rim and make them independent from each other and both true by default.\n\nThe modifier has options for a vertex group as a factor for the thickness and two vertex groups to store the fill and rim geometry.\nvertex groups will be replaced by two attribute string fields for the factor and to tag the rim geometry\n\n", "Layer Switching Shortcut Inconsistency\nThere is some improvement concept for collections shortcuts.\nCurrently, 1234 buttons shows a collection, but sending objects requires name in menu (M - \"Collection 1\" - \"Collection 3\" for example)\nThat would be nice if moving to collection will also support their number.\n\nFor example, \"5\" button shows collection called \"Lighting Collection\", so pressing \"M5\" will send selected objects to it.\n\n----\n\n*See: #55162#619317 - originally from @1D_Inc*", "Click selection and box selection in 3D view do not work properly in scenes with many objects\nOperating system: Darwin-19.2.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 580X OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.4.19\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional)\n\n[In my environment, if there are close to 100,000 objects, the objects can not be selected normally]\n[select_bug.mp4](select_bug.mp4)\n\nUploaded blend file has nearly 100,000 simple cubes\nClick on an object in the 3D view to select a distant object\nEven a box selection can make a crazy choice\nBut lasso selection and circle selection work fine\n[test_file.blend](test_file.blend)\n\n", "Grease Pencil wrong Cursor Size Smooth Brush when accesed with the [SHIFT] key\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 460M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 3.3.1\nWorked: \n\nThe On Screen **Smooth Brush Cursor Size** is inaccurate when using the Smooth Brush with the **[SHIFT] **key while using any other brush of a different radius setting.Grease Pencil > Draw any curve \nGrease Pencil > Sculpt > **Smooth Brush**> Make the Smooth **Brush Radius VERY SMALL** (10).\nGrease Pencil > Sculpt > **Grab Brush** (or **ANY OTHER BRUSH**) > Make the **Brush Radius HUGE**(150). \nUse the **Grab Brush** as you wish.\nPress **[SHIFT]**to activate**Smooth Brush**while in Grab Brush (or**ANY OTHER BRUSH**)\nThe on screen cursor Radius while the [SHIFT] key is pressed does not reflect the VERY SMALL Brush Radius (10) that we set earlier. This is constantly confusing. Can you please make the radius display correctly based on the actual Smooth Brush Setting?\nattached scene: GreasePencil_SmoothBrush_CursorSize_bug.blend\n[GreasePencil_SmoothBrush_CursorSize_bug.blend](GreasePencil_SmoothBrush_CursorSize_bug.blend)\nThank you!", "Touchscreen 'Continuous Grab'\nOperating system: Linux-4.15.0-54-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: Mesa DRI Intel(R) Haswell Mobile Intel Open Source Technology Center 4.5 (Core Profile) Mesa 19.0.2\n(bug also tested on Windows 10)\n\n\nBroken: version: 2.80 (sub 74)\nWorked:\n\n\n* Activate a gizmo using the **touchscreen** on a computer. (The 3D navigation gizmo for example)\n* Drag your finger outside the original editor\n* The gizmo is manipulated incorrectly\nThis bug disappears if \"*continuous grab*\" is disabled in the User Preferences.\nI think it's a matter of auto-disabling *continuous grab* while receiving touch input, like how pen tablets are handled.", "GPencil: Shortcuts in Draw Mode are hardcoded\nMultiple keys in the `gpencil_draw_modal` function are hardcoded and (some) don't show up in the info bar. This makes it impossible for the user to know what certain keys do as well as impossible to change them in the keymap. All of those keys should be in the keymap with the default keys set as they are now.\n\nHere is a list of these (hardcoded) shortcuts:\n\n - {key Enter}, {key Pad_Enter}, {key Esc}, {key Space}, {key Ctrl Z} (?): Confirm\n - {key Mousewheel_Up}, {key Pad_Plus}: Increase size of eraser\n - {key Mousewheel_Down}, {key Pad_Minus}: Decrease size of eraser\n - {key O}: Set center of guide\n - {key V}: Turn off speed guide\n - {key M}: Alternate or flip guide direction\n - {key L}: Turn on line guide; {key Ctrl L}: Turn on line guide with angle = 0; {key Alt L}: Turn on line guide with last angle\n - {key C}: Turn on circular guide\n - {key K}: Turn line guide clockwise\n - {key J}: Turn line guide anti-clockwise\n", "How to handle edit-mode face/edge select-modes without any face/edges.\nBased on this report #41562,\n\n----\n\nSimple example of the problem:\n\n- Add a Circle\n- Enter EditMode\n- Select Face Mode\n- Select All (AKey)\n- Grab *(Nothing happens)*\n\n----\n\nIf you have a mesh without faces in edit-mode and use face-select for example,\nQuite a lot of tools don't work usefully.\n*(same for edge select-mode on a mesh containing only vertices, but for purpose of discussion - just talking about faces for now).*\n\nLists of tools which fail when edges/vertices are selected but no faces exist.\n\n- Transform Grab/Rotate/Scale does nothing.\n- Extrude does nothing.\n- Inconsistent de-select-all verts and edges (AKey) - *(but border/circle/mouse select dont work).*\n- *...Likely others exist, off hand these are the ones I can find...*\n\nThere are some possible solutions.\n\n- Automatically switch out of face mode *(when entering edit-mode - doesn't solve problem completely since you can remove all faces while editing)*\n- Make tools support this configuration *(need to check how to support for each tool).*\n# Make tools report a warning when its not supported (or add support when it makes sense), *otherwise - assume users know what they're doing and not attempt to second-guess their actions.*\n\nFor this design issue, Id like to come to some agreement on a good way to handle this case.\n\n----\n\nSome things to consider regarding automatic select-mode switching:\n\n- Users may temporarily remove all faces, so changing select-modes automatic for the users could disrupt their workflow.\n- Switching modes automatically could break scripts (and generally give unpredictable behavior).\n- Checking if the mesh has faces isn't really reliable since you may have hidden faces which is like having no faces from the users POV.", "Pie menu design improvements\n\nNow that we are adding more pie menus, we could make some small but effective improvements to them.\n\n## Wedge Highlighting\nFirst, the wedge could be highlighted as you rollover it, to make it clearer how big the area is, like so:\n\n\n## Holding, Then Releasing Key Should Always Dismiss Pie\nThere's a slight quirk in the way we handle input: Normally, if you hold a key to invoke a pie menu and then let go, the pie menu is dismissed. This works if you move your cursor to an item, or to an empty area. But, if you don't move the cursor at all, this doesn't work. This makes the behaviour inconsistent, and makes the pies 'stick' in an annoying way if you regret opening one.\n\n## Letter Input While Pie is Active\nWhile number input works reasonably well, in Blender we also use letters to invoke options in menus. For pies, letters work better, because the order makes less direct sense. There's nothing intuitive that says left=1, right=2, and so on. We should enable letter input to activate pie options. This way, user can press Z, S to activate Solid Mode.\n\n----\n\nThese three changes should make pie menus more delightful to use.", "Drag-to-toggle visible/selectable/renderable in the Group view.\nWindows 64\n\nBroken: 2.70a\n\nThe drag-to-toggle visible/selectable/renderable in the Outliner does not work in the Groups view.\n", "Annotation: The d-key for annotation not correctly bind in edit mode when using: knife tool, loopcut tool, polybuild, curso tool\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 550 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 376.51\n\nBroken: version: 2.82 (sub 1)\nWorked: (optional)\n\nAnnotation: The d-key for annotation not correctly bind in edit mode when using: knife tool, loopcut tool, polybuild, curso tool\n\n- Go to edit mode try to draw with the d-key down + left mouse drag while one of this tool is selected: knife tool, loopcut tool, polybuild, curso tool \nNotice while drawing ( d-key down + left mouse drag )is broken, erase ( d-key down + right mouse drag) is still working properly.\n", "Circle select does not select all vertices within the same pixel\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.90.1\n\nSelecting closely-arranged vertices with circle select tool doesn't consistently select all vertices.\n\nCreate plane geometry with three vertices located close together (e.g., the end of a sword or a blade of grass). Type 'C' to use the circle select tool and select the vertices. Zoom in to verify that all vertices have been selected. De-select vertices. Change zoom level. Repeat selection. Repeat verification. See animated GIF, attached, using simple .blend file, attached.\n\n\n\n[C3DE_Ch4_L12a.blend](C3DE_Ch4_L12a.blend)\n\n", "Outliner: Alt+disable is work for all selected, but Alt+Enable is not\nBroken: version: 2.82 (sub 6)\nWorked: (optional)\n\n{key Alt LMB} generally enables a feature that repeats the operation on all selected items in the outliner.\nBut that seems to work only if the object is selected in 3d View.\n[2020-01-12_20-44-23.mp4](2020-01-12_20-44-23.mp4)\n\nOpen:\n[untitled4.blend](untitled4.blend)\n- select the three objects (either in 3d view or outliner).\n- {key Alt LMB} click on the icon that disables object in viewports.\n- {key Alt LMB} again to try to enable them all.\nOnly one object is enabled, because the others are not selected in 3dview.\n\n", "Accessed thru Parsec software, with left hand mouse, some operations need a right click instead of a left click to confirm\nOperating system: ?\nGraphics card: ?\n\nBroken: version: 3.6.1 Release Candidate, branch: blender-v3.6-release, commit date: 2023-07-16 02:43, hash: edc1a09966a7, type: release\nbuild date: 2023-07-16, 23:35:12\n\n\n\nI m using parsec to acces to my workstation, but through parsec I have issues,\n\nI m left hand so my left click is a right one. I notice some operations needs a right click instead of a left click to confirm. For validating a scale I need to right click, for validating a rotation too.\nfor grab it's ok.\n\n\n- Access Blender through Parsec\n- I'mm a left hand guy I set my mouse with the master click as the right button instead of the left one, \nso everything should work with a right click instead of a left one...\n- Select an object with right click and grab, and can confirm grabbing with right click.\n- Repeat for Rotate/Scale, but have to confirm with left click.\n\n" ]

[ "Circle Select tool: status bar info doesn't match with the actual operations\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 25.20.100.6444\n\nBroken: version: 3.0.0 Alpha\nWorked: NA\n\n\nThe \"Select circle\" command is riddled with problems:\n\n**Issue 1**\n\nWhen we choose the \"Circle Select\" tool, the status line does not provide the relevant information at all.\n\n - There is no information about how to change the diameter of the circle\n - The status line shows that \"LMB+drag\" moves the item, which is completely erroneous.\n \nOnce we press the LMB down, the status line changes to-\n\n\n**Issue 2**\nEven this is totally faulty: \n - There is no difference between ESC (Cancel) and ENTER (confirm). The selection remains in both cases.\n - The WhUp increases the diameter of the circle. It does NOT \"add\" anything to the selection.\n - The WhDown decreases the diameter of the circle. It does NOT \"Subtract\" anything from the selection.\n - The meaning of \"MSPan\" is not clear at all. I could not find its meaning on the whole Internet!\n - The SHIFT+LMB does NOT deselect an object (or toggle its selection). In fact, there is no way to deselect any object at all!\n - The role of RMB (to cancel) is not listed at all.\n\nI found experimentally that when we roll the mouse wheel while the LMB is pressed down, the diameter changes.\n\n**Issue 3**\nAt this time, the mouse shows two different circles simultaneously: The new circle is in dotted line. The default circle is in solid line. Showing two circles at a time is totally unnecessary, and does not serve any purpose.\n\nEven in the light of this decoded behavior, I cannot guess what \"MsPan\" means!\n\nSee above\n\n**Suggested solutions:**\n\n - Correct the status line to show the correct hints\n - In the sculpting brushes, the diameter can be adjusted by pressing F and then dragging the mouse. Use the same method here too, and have consistency!\n - Please provide a way to deselect items." ]

[ "VSE Channel headers TODO\nImprovements for channel headers in general:\n- - [ ] 2-way `v2d` scroll synchronization\n- - [ ] Investigate why changing RNA property doesn't update timeline region\n- - [ ] Cleanup loading code for `SeqTimelineChannel`, looks like it could be optimized a bit more.\n- - [ ] Use better integrated drawing engine like list-views\n\nOther related tasks:\n - Decide on [D14263: VSE: Limit timeline view height](D14263)", "Display issue in model corner when loop cutting\nBroken: 2.79, 2.81\nWorked: -\n\nWhen creating and edge loop around de outside faces it creates in one of the cornes an extrange artifact\n\nCreate an edge loop across the outside faces of the model.\n\nBlender file:\n[loop cut.blend](loop_cut.blend)\n\nArtifact:\n\n", "Proposal: Remove bottom headers\nFollowing on from [D4302](D4302) here. \n\nIt may be good to remove the ability for headers to be at the bottom of editors. Here's why:\n\n - Popovers don't work well when opened from the bottom.\nWhen the height varies, the contents gets pushed around:\n\n\n - Opening old blend files is now a pain, because users have to go ahead and manually move all the headers to the top\n\n\n - It makes it possible, or at least easier, to add footers to some areas, such as in the Text Editor to better give space for the file path:\n\n\n - We can then remove the header position preference, which is a small simplification.\n", "Sequencer: transforming strips with overwrite enabled sometimes shuffles\nThe overwrite option sometimes shuffles the strips.\n\nWith `838b6ec48af6fb767089784193d8525c5f5faf13` (current master), and the attached file, the following error can be observed.\n\n- `G` key.\n- Type in `654`.\n- Press Enter to confirm.\n\n... if this doesn't shuffle the strips (so one strip is placed above the others), press Ctrl-Z and repeat the steps above.\n\n[sequencer_overwrite_sometimes_shuffles.blend](sequencer_overwrite_sometimes_shuffles.blend)\n\n", "Transform cursor (patch included)\nRecently LMB-Drag was made to move the cursor *(edit, this has since been reverted)*.\n\nHowever this just moves the cursor under the mouse,\n\nit _could_ be useful to use transform here, it allows.\n\n- Drag, XKey to lock to an axis\n- ... Shift+Y to constrain to a plane.\n- ... Hold Ctrl to snap (snap to verts for example is handy)\n- ... Shift for more precise placement...\n... most other transform keys.\n\n\nOne reservation I have about this, is that you probably would often want to be able to snap to vertices and edges, and not have to depend on the current snap options in the 3d view header.\n\nSo we could override the snap options that transform uses.... but this seems a bit crappy too, since you may want to snap to faces in some cases, or even snap to grid.\n\nAs an alternative, we could just not bother attempting to integrate this with transform, and add snapping to the current modal operator (similar to how we have with the Ruler operator, which shares low level code with transform snapping)\nBut in that case we miss the axis constraints which seem like they would be nice to support.\n\n\nHere is a test patch to show how it works: [P87](P87.txt)\nNote, it crashes if you try to scale and its hard coded to LMB for cursor place, so its mostly a proof of concept.", "\"Z\" changed to \"depth\" breaks consistency in compositing and in Render layers\n\nwin7 gtx 550ti\nOperating system and graphics card\n\nBroken by: 4cf7fc3b3a\n\n\"Z\" changed to \"depth\" breaks consistency in compositing and in Render layers\n\n\n- Make sure that the **z**-pass is checked in the Render Layers tab\n- Open node editor, notice that the**z**-pass is now **depth**-pass\n- Notice as well that everywhere else it is still consider as**z**-pass even in the manual z_combine.html\n\n\n\nAlthough **depth** is more meaningful name, it is breaking consistency. Suggestion: we bring back the old behavior or we change everywhere to z-depth or depth.\n", "Adjust Last Operation: Other changes outside that panel are reverted back automatically\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.90.0 Alpha\n\nHere the demo and file\n[2020-06-23_20-30-11.mp4](2020-06-23_20-30-11.mp4)\n[Scene state reset.blend](Scene_state_reset.blend)\n\n1. Add some objects, add last one\n2. Change visibility for other objects via outliner\n3. Change property in adjustment popup\n4. Scene settings resetted to previous", "3D Texturing: When changing the texture resolution, the pixel data isn't recreated\n* Start painting\n* Change image resolution\n* Continue painting.\n\nWe have a safe guard for this, but it doesn't seem to work anymore.\n", "Path Guiding: Cycles in viewport produce inconsistent results\nOperating system: Windows 11\nGraphics card: Geforce 1660S\n\nBroken: 3.5\n\nJust moving the camera around the viewport gives inconsistent cycles renders. I am not sure if it's directly related to path guiding as the issue is only visible in the mirror caustics.\n\n\n\n\n", "Front view Y axis keybinding is reversed\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: version: 3.5.0\nWorked: Never?\n\nFront view key binding should give us +Y view, now it gives us -Y view. The negative Y axis is now key binded for the front view instead of the back view.\n\n- Press numpad 1, viewport switched to -Y view. (Should be +Y)\n- Press cntrl + numpad 1, viewport switched to +Y view. (Should be -Y)\n\n", "Regression: panels in the properties panel lose their place when scrolling one of the other property panels.\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.96\n\nBroken: version: 3.4.0 Alpha\nBroken: 2.91\nWorked: 2.90\n\nThis is caused by 940b239ad4\n\nWhen we scroll down in any panel and switch to another panel then changing back to previous panel will reset its position (i.e. to the top)\n\n\n[2022-10-18 22-54-24.mp4](2022-10-18_22-54-24.mp4)\n\n", "Inconsistency: Vector math node flips large list of options/operations.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 473.04\n\nBroken: version: 3.1.2\n\nWhen direction of opening changes, direction/order of list changes too.\n\n[2022-04-12_09-48-55.mp4](2022-04-12_09-48-55.mp4)\nIt is fun, it was the same in 2.79. But there was very short list with 1 column And it wasn`t so annoying. \nNow, with 2 columns the order should be fixed same way as in Math node.\nBecause user remember the relative position of option too. Not just «how far from start»\nAlso the same flipping exists for other lists, like Transfer Attributes node. But as I said, for one short column it is not the issue at all\n", "NLA strip unexpectedly auto-switching from Hold to Hold Forward\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 770/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.64\n\n\nBroken: version: 2.80 (sub 74)\n\n\nWhen creating a new action using the HOLD_FORWARD extrapolation, frames to the left of the keyframes will keep the value of the first keyframe, instead of ignore them like they did before and do with NOTHING extrapolation. They now behave like HOLD.\nThey work like normal in strips, just not as the *(orange)* action.\n\n[nla_hold_back.blend](nla_hold_back.blend)\n[2019-07-14 19-08-18.mp4](2019-07-14_19-08-18.mp4)", "Scene strips don't include passes\nOperating system: Linux-5.8.11-1-MANJARO-x86_64-with-glibc2.32 64 Bits\nGraphics card: AMD VERDE (DRM 2.50.0, 5.8.11-1-MANJARO, LLVM 10.0.1) X.Org 4.5 (Core Profile) Mesa 20.1.8\n\nBroken: from 2.79 to version: 2.92.0 Alpha\nWorked: -\n\nThe render passes of another scene (as scene strip) are ignored in the master scene.\nThis will most probably be categorized as limitation. Since it is a severe one, it should at least be documented here:\nPagename: `video_editing/sequencer/strips/scene`\n\nName current scene \"A\", create new scene, in the sequencer add scene strip for \"A\", hit F12\nThe content of scene \"A\" will be rendered, but the Z channel (enabled by default / other channels if activated) of scene \"A\" will not be rendered, the Select Pass dropdown is not there and the layers are not saved to multilayerEXR.\n\n", "Regression: Bevel issue when its percent/absolute mode is operated on edge loop along sharp angle\nBroken: 2.91~4.0\nWorked: 2.90.2\n\nCaused by 48c484a22e0d105ed21a4f18b0cb455af6aac661\n\nIf bevel's percent and absolute modes are operated on an edge loop along sharp angle, the sharp angle part does not get beveled.\n\nTest file: [bevel_sharp.blend](https://projects.blender.orgattachment) ~~[bevel_issue_2.blend](attachment)~~\n\n### Test mesh\nEdge loop at the center has bevel weight of 1.0 and runs along a sharp angle.\n\n\n\nI'm using bevel modifier here for convenience but this also happens with edit mode bevel operator.\n\n| | 2.83.20 | 2.93.16 | 4.0 alpha 2023-06-26 `82d9eb09dbbe` |\n|--|--|--|--|\n| Percent mode |  |  |  |\n| Absolute mode | - | - |  |\n\nEdit: Reproduce:\n\n- Open the updated [bevel_sharp.blend](https://projects.blender.orgattachment)\n- Bevel\n- Select to Absolute/Percent mode and adjust depth\n- observe two cubes behaves differently \n" ]

[ "Inability to move the pasted strip (image, video) towards at the lower channel\n%%%If you put two strips (e.g, first - channel 1, second - channel 2) and you move the first strip at the channel 3 (not with drag and drop but manually by clicking on \"position of the sequence strip\" in the edit strip panel) then isn't possible to do it in the opposite direction (from the channel 3 back at the channel 1).\n___________\n(Vista 32 bit, nVidia GF 9300, Blender 2.57b)%%%" ]

[ "GN: Delete Geometry node hides edges in some cases\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.23\n\nBroken: version: 3.2.0 Alpha\n\nHard to describe this one more abstract..\n\nI start with a mesh line primitive (0,0,0) (0,1,0) and extrude it **n** times along **x** to get a grid.\nNow I try to delete all points with a distance from the **y** axis greater than **d** \nWhenever **d** lies between the first and the last extrusion, the edges of the starting primitive are not displayed.\nIt shows up in the spreadsheet however and reappears if I apply the modifier and tab into edit mode and back \n\n[2022-02-05 19-27-50.mp4](2022-02-05_19-27-50.mp4)\n[del_geo_node.blend](del_geo_node.blend)\n\n", "Geometry nodes. Viewer node data late update (delete node and undo)\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.1.0 Alpha\n\n**Short Description**\nSpread Sheet editor is not updating when I undo the action of viewer node deletion\n\n**Steps to Reproduce**\n- Open .blend file\n- Delete viewer node\n- undo (You will not see any data in the spreadsheet until click on the viewer node or hovering mouse over spreadsheet)\n\n[#94641.blend](T94641.blend)\n[2022-01-04_20-09-05.mp4](2022-01-04_20-09-05.mp4)\n", "Animated Values in Node Trees and Geometry Nodes Modifier are not distinguishable \nOperating system: all\nGraphics card: all\n\nBroken: blender-3.0.0-alpha+master.12e8c783535c and below\n\nHello dear Blender devs,\nHello dear Geometry Nodes developers and maintainers,\n\nI did wait some time before reporting this as a \"bug\" as from some perspective this might sound like a feature request. But as there are some, to users unidentifiable names used, I guess this is may very well count as a bug.\n\nWhen working in Geometry nodes it is important to be able to keyframe some values and properties to achieve more controlled and art directable Animation of the Asset. While the setting of keyframes works really well, it gets really complicated to tell the different animation channels apart as they neither show the nodes name nor the nodes Label.\n\n\nOpen attached file, try to figure out which graph in the editor belongs to which Node.\n[geonodes_01.blend](geonodes_01.blend)\n\nFor simplification I added three screenshots each showing one of the three cases I could Isolate where this bug appears:\n\n1.) Animating Node Values in the Geometry Nodes Tree\n\n\n\n2.) Animating Exposed Custom Properties of the Geometry Nodes Modifier\n\n\n\n3.) Animating Node Values inside the Shader Graph.\n\n\nThank you in advance for your evaluation of this bug report\n\nAlex / zuggamasta", "Hotkeys conflict between Shader Nodes, Compositing Nodes, GeoNodes and Texture Node Editor.\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 2.93.6\n\nI have hotkeys on frequently used nodes in shader editor. When I switch to Compositing, first of all I have hotkeys already assigned to similar nodes (I did it in shader editor, but they are automatically assigned in compositing). Secondly, I get an error in Compositing - Not a shader or geometry node tree.\n\n1) Apply the \"I\" button to the \"Invert\" node in the shader editor\n2) Switch to Compositing or Texture Node Editor\n3) Call the invert node with the I button\n\nYou get the following error \n```\nPython: Traceback (most recent call last):\n\nFile \"D:\\BF_daily_build\\2.93.0\\blender-2.93.7-candidate+v293.4ec6cc412d7f-windows.amd64-release\\2.93\\scripts\\startup\\bl_operators\\node.py\", line 138, in invoke\n result = self.execute(context)\nFile \"D:\\BF_daily_build\\2.93.0\\blender-2.93.7-candidate+v293.4ec6cc412d7f-windows.amd64-release\\2.93\\scripts\\startup\\bl_operators\\node.py\", line 129, in execute\n self.create_node(context)\nFile \"D:\\BF_daily_build\\2.93.0\\blender-2.93.7-candidate+v293.4ec6cc412d7f-windows.amd64-release\\2.93\\scripts\\startup\\bl_operators\\node.py\", line 92, in create_node\n node = tree.nodes.new(type=node_type)\n\nRuntimeError: Error: Cannot add node of type ShaderNodeInvert to node tree 'Geometry Nodes'\n\nNot a shader node tree\n\nlocation: <unknown location>:-1\n```\nbandicam_2021-11-26_16-53-56-020.mp4\nNode_Conflict.blend\n\n", "Copy/Paste object operator copies collections specified somewhere in that object (e.g. Geometry Node Collection Info node, particle system instance collection, ...)\nBroken: Current master.\nWorked: Never.\n\nThe copy operator treats the `Collection Info` node as the fact that the object belongs to the collection.\n\n| Ctrl+C | Ctrl+V | Expected |\n| -- | -- | -- |\n|  |  |  |\n\n1. Create collections `Collection_1`, `Collection_2`.\n2. Add Cube in `Collection_1`.\n3. Add Cone in `Collection_2`.\n4. Create geometry node modifier on Cube. Go in node tree:\n a. Create `Collection Info` node.\n b. Choose a `Collection_2`.\n5. Select `Collection_1`, Cube.\n6. Copy object in 3d view.\nExpected: Cube object will copied.\nBug: Copyring object, collection and attach that in collection that chosen in `Collection Info` node.\n\nThe same happens when you:\n\n- Create collections Collection_1, Collection_2.\n-- Add Cube in Collection_1.\n-- Add Cone in Collection_2.\n- Create a particle system on the Cube\n-- Render As Collection\n-- specify Collection_2 as the Instance Collection\n- Select Collection_1, Cube.\n-- Copy object in 3d view.\n-- Expected: Cube object will copied.\n-- Bug: Copying object, collection and attach that in collection that chosen in the Particle Instance Collection.", "Texture Node Editor Using nodes causes weird issues in sculpt mode.\nOperating system: Linux-5.15.0-48-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.76\n\nBroken: version: 3.3.1\nWorked: perhaps never? (can redo in 2.83)\n\nAdding an image from the nodes of the texture editor for example, a sculpting brush, causes really weird issues. Disabling nodes and using the texture properties does not have this issue.\n\n- Check the blend file. Try sculpting, you will see weird noise being displaced with the texture.\n- Now disable the nodes and set the texture type from the texture properties to the same image. Sculpting again you see the issue is not happening.\n[bug.blend](bug.blend)\n\n\n\n", "Curve Deform Node\nNOTE: The design below is outdated. A curve deform node should be a field node now, with a vector input and output, and a curve input. Arguably it should just be a node group too.\n\n\n| {[F10152887](image.png) size=full}| {[F10152878](image.png) size=full} | {[F10152882](image.png) size=full} |\n| -- | -- | -- |\n\n\nThis is task is for a port of the \"Curve\" modifier to a node.\n{[F10150643](image.png), size=50%}\n\n\n**Details**\n* Any geometry component with a position attribute is supported by this node.\n* Only the first spline of the curve is used for deformation.\n* Two input modes modes are available. \n**The position mode uses point positions for the curve parameters.*** The position input is provided for convenience. In most common cases, a single axis of the position is used as a parameter for the curve.\n***The position is mapped to the curve based on length. So \"x=5m\" means, transform by traveling 5m along the spline. This is how the modifier works.** The attribute input mode uses any float attribute between zero and one, as a factor along the length of the curve.\n** For convenience, a \"Float\" option is available for the attribute when the whole geometry should have the same parameter.\n* Extrapolation should be handled by a separate node that can extrapolate a spline.\n \n\n**Further Questions**\n* Is an \"influence\" attribute input really necessary? The same thing could be achieved by mixing a copied position attribute before starting the operation.\n** The answer to this question will apply to other operations as well.\n* Which of the three \"Curve Deform\" settings are necessary in this node? \"Use Radius\" is likely not necessary, the attribute could just be filled beforehand.\n\n**Related Nodes**\n* A \"Curve Sample\" node could do a similar operation, but instead of deforming positions, it would sample an attribute along the length of a curve based on the same mapping controls.\n** This could be useful to have more flexible control of attributes with curves.\n\n\n**Implementation**\nThe implementation should be straightforward. Sampling arbitrary points along the curve is an natural extension of the uniform sampling that is already implemented.\nCurve parameters should be sorted before, and then changed into evaluated point indices and factors.\nThen, interpolation between the transform of neighboring evaluated points should be quite simple.\n", "Weird behavior with the set spline type node selection\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.6.1\n\nThe selection on the *set spline type* isn't working right for bezier and NURBS modes.\nAlso adding another node before the *set spline type* node will manipulate the weird behavior, even though the node doesn't do anything or is muted etc.\n\n- add a GN setup\n- add more that one curve\n- add a *set spline type* \n- try adding a selection that doesn't select all splines\n- see the weirdness.\n(it is more perceivable if the splines have more than two points, though the bug is still there)\n\nFile:\n\n[GN_set_spline_type_test.blend](attachment)\n\nVideo:\n\n<video src=\"attachment\" title=\"2023-07-31 21-15-45.mp4\" controls></video>\n\n", "Node Tools not exposed in the Preferences > Keymap\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: AMD Radeon (TM) R9 380 Series ATI Technologies Inc. 4.5.13596 Core Profile Context 20.10.35.02 27.20.1034.6\n\nBroken: version: 3.3.1\n\n**Short description of problem**\nFor the Node Editor keymap, there are hidden Node Tool keymaps that exist within Blender (they are only visible via search filter) that override the keymaps that are exposed, and can't be customized properly.\n\n\n\nThis report stems from this bug report:\nT102422 ", "Array modifier is crazy, when using \"object offset\" pointed to empty, which is parented to curve\nOperating system: Mac OS Catalina 10.15.4 (latest)\nGraphics card: Intel HD 6000\n\nBroken: 2.83 beta as of 8-th of May\n\nArray modifier is crazy, when using \"object offset\" pointed to empty, which is parented to curve. And when the CURVE CONVERTED TO MESH -- it just destroys scale of object!\n\n1) Open .blend\n[plants.blend](plants.blend)\n2) Try to move empty -- everything works fine\n3) Select \"empty\" and \"petals\" (with shift) then parent them (Ctrl + P -- to object) to \"bezier curve\"\n4) Select bezier curve and \"convert to mesh\"\n5) Now try to move empty\n\nBefore:\n\nAfter:\n", "Add Signed Noise node to Shader and GN\nThe existing Noise texture was originally designed for textures where values in the range 0-1 are useful.\n\nFor 3D space used in geometry and volumes it is useful to have signed noise.\n\nAfter exploring adding a signed noise output to the existing Noise Texture I created the following patches.\n\n\n\n[D12820: Nodes: Add Vector Distortion node (WIP)](D12820)\n\n[D15947: Nodes: Add Signed output value to the Noise Texture node](D15947)\n\n[D15944: Nodes: Add Signed Noise node](D15944) \n\nSeparate sockets:  \n\nEnum (current patch): \n\nI'm currently in favour of [D15944](D15944) and adding the function as a simple node. \n\nDistortion using just the Signed Noise patch.\n\n\n\nThe advantage of exposing the base noise function is that it allows users to create their own noise based node groups. It decouples the function from the existing Noise texture and only exposes the noise function for a simple UI.\n\nIt is possible to use the existing noise texture for this but it needs the correct values setting plus additional nodes. The underlying code is less efficient.\n\n\n\n\n```\n/* Distortion. */\n\nBLI_INLINE float3 perlin_distortion(float3 position, float strength)\n{\n return float3(perlin_signed(position + random_float3_offset(0.0f)) * strength,\n perlin_signed(position + random_float3_offset(1.0f)) * strength,\n perlin_signed(position + random_float3_offset(2.0f)) * strength);\n}\n\n/* Positive fractal perlin noise. */\n\ntemplate<typename T> float perlin_fractal_template(T position, float octaves, float roughness)\n{\n float fscale = 1.0f;\n float amp = 1.0f;\n float maxamp = 0.0f;\n float sum = 0.0f;\n octaves = CLAMPIS(octaves, 0.0f, 15.0f);\n int n = static_cast<int>(octaves);\n for (int i = 0; i <= n; i++) {\n float t = perlin(fscale * position);\n sum += t * amp;\n maxamp += amp;\n amp *= CLAMPIS(roughness, 0.0f, 1.0f);\n fscale *= 2.0f;\n }\n float rmd = octaves - std::floor(octaves);\n if (rmd == 0.0f) {\n return sum / maxamp;\n }\n\n float t = perlin(fscale * position);\n float sum2 = sum + t * amp;\n sum /= maxamp;\n sum2 /= maxamp + amp;\n return (1.0f - rmd) * sum + rmd * sum2;\n}\n```\n\nvs\n\n\n```\n/* Perlin signed noise */\nr_value = noise::perlin_signed(position);\n```\n", "Wireframe overlay opacity does not work correctly with viewer node\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.0 Core Profile Context 23.3.1.230305\n\nBroken: version: 3.6.0 Alpha\nWorked: n/a\n\nAlso confirmed in 3.4 (which introduced the Viewer Node).\n\nWireframe overlay opacity does not work correctly if Geometry Nodes viewer node is activated and it's showing attribute.\n\nIt looks as if there's extra wireframe overlay under the usual overlay. Remaining wireframe does not respect smooth wires option from the preferences.\n\n- From the startup file, add Geometry Nodes to the cube\n- From Geometry Nodes editor, add a node that outputs attribute (Random Value node for example)\n- Connect Group Input and the attribute node to Viewer node\n- From viewport overlay option, enable wireframe. Try opacity slider\n\nVideo:\n<video src=\"attachment\" controls></video>\n\nFile for quick test:\n[ViewerNodeWireframe.blend](attachment)\n\n", "Nodes: Node Name are overlap Icons in header\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 527.56\n\nBroken: version: 4.0.0 Alpha\nWorked: 2.78...\n\nCropping text of node Name sometimes allows a collision of text and icons:\n\n\n1. Any node editor.\n2. Create node group.\n3. Change node scale to see effect for narrow header.\n\n", "Dependency Graph: Assert for DEG_get_evaluated_scene\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 527.56\n\nBroken: version: 4.0.0 Alpha\n\nSee attached stack trace.\n\nDebug build:\n1. Default cube.\n2. Add Geometry Nodes modifier. In node tree:\n a. Add Cube primitive node.\n b. Add Scene Time node.\n c. Putt Cube geometry to Group Output node.\n d. Set some large values in Vertices inputs of Cube node. (~0.5s to compute)\n3. Start playing of animation.\n4. Click on some place of time line\n5. Press undo/redo few times.\n\n", "Node edge scroll affects cancel position\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 XT ATI Technologies Inc. 4.5.14830 Core Profile Context 21.4.1 27.20.21002.112\n\nBroken: version: 3.0.0 Alpha\nWorked: Unknown\n\nWhen starting a translate transform and using edge scroll this will affect the position after cancelling the translation.\n\n- Select node and start translate (use hotkey instead of LMB drag)\n- Scroll the screen\n- Cancel the transform (right click)\n\nNode will end up at start position relative to current viewport, but not where the transform actually began\n\n(Confirmed by @PratikPB2123 in related issue [#90391](T90391))" ]

[ "Regerssion: Search menu not working correctly in node editors\nOperating system: MacOS 12.6 Monterey\nGraphics card: Radeon Pro Vega 56 8 GB\n\nBroken: version: 3.4.0 Alpha\nWorked: 3.4.0 Alpha\n\n\nIn Blender 3.4 beta and 3.5 alpha, when using the menu > add > search option in any node editor the new node appears about half a screen width right of the position of the mouse cursor and about half a screen height under the position of the mouse cursor\n\nWhen using the shortcut shift-A > search, Blender 3.4 and 3.5 behave normally and the new node appears under the mouse cursor\n\n\nUse the menu > add > search option (not the shortcut shift-A) to place a node in any node editor." ]

[ "Wrong width returned from bpy.types.Window.width (depending on native pixel size)\n\nOS: macOS Mojave (10.14.6)\nGraphic card: Radeon Pro 450 2GB\n\n\n\nBroken version: 2.91.0\n\n\n**Description of error**\n\nbpy.types.Window.width does not return correct value in my environment.\n\nI tried to get window size of Blender via bpy.context.window_manager.windows- [x].width.\nBut the returned value differs from the actual window width (Actual window width: 2880, Returned value: 1680).\nAlso, I found that bpy.context.window_manager.windows- [x].x also returned invalid value.\n\nOther related values such as region.width, area.width seems fine in my environment.", "Wrong add selection to active collection operator description\nOperating system: Arch Linux\nGraphics card: GTX 970\n\nBroken: 036e95bb21fa\n\n\nWhen left-clicking on a collection in the outliner it is highlighted in white and its name is also shown in the 3D viewport and status bar. This to me indicates it is the \"active collection\", analogous to the active object. \n\nHowever, the tooltip (and collections in the list) for Add selection to active collection (Ctrl-Shift-G) says \"Add the object to the collection **that contains the active object**\". This indicates a different meaning of \"active collection\". \n\nSo two questions:\n- Which is the correct definition of active collection?\n- What is the use of the highlighted collection in the outliner?\n\n\n", "Stabilize boolean's epsilon use, write tests\n**Motivation:** resolve bugs in boolean code.\n\n**Description:**\n\nCurrently booleans use epsilon comparisons in a non-uniform way,\ncausing near edge-cases for fall *in-between* detecting a segment and an end-point intersection.\n\nThis would resolve some of the open bug reports we have now.\n\n**Estimated Time:** 2 weeks\n\n- 1 week writing tests, make sure existing bug reports are covered by these tests.\n\n *This is very important since currently there aren't any good ways to know if a change to boolean code breaks edge cases.*\n\n- 1 week (max) to go over current boolean code and make all comparisons uniform\n\n *(may be a day or two as well, just needs to be done carefully, possibly re-working intersection code)*\n\n\n```\n", "Hair BSDF\nThis is a tough topic, I need to investigate it more.\nIdea: Dual Scattering using shadowmaps.\n\nEstimate : 2 weeks\n", "GP: Time offset and Armature modifier causes glitches\nOperating system:WINDOWS10\nGraphics card:GTX1060\n\nBroken: 2.93LTS\n\n\n[#89433.blend](T89433.blend)\n\n- Open file\n- Move selected bone to left until it hits constraint, and then a bit randomly\n\nGP object will keep rotating in one direction and can be reset only by changing frame.\nSame happens when scrubbing in negative frame range.\n", "2D transform operations show options and use a modal keymap intended for 3D transform\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.92.0 Alpha\n\ntransform.seq_slide and transform.transform are linked to the transform keymap of the 3D transform. This results in the wrong shortcuts hints #83307 in the status bar and also wrong options in the Edit Last Operation: \n{[F9508650](transform_last_operation.gif),size=full} \n\n- Use `Move/Extend from Current Frame` to move the handle.\n- Check the `Edit Last Operation` options\n\n", "Texture UV/Image Editor coordinates slightly inaccurate\n**System**\nGentoo with old Advanced Micro Devices, Inc. [AMD/ATI] RV710 [Radeon HD 4350/4550]\n\nSlightly broken: media-gfx/blender-2.79b-r1\n\nWhen zooming in as much as possible on a texture in the UV/Image Editor, there seems to be a pixel border at y≈235.994 instead of exactly at y=236 like I'd expect.\n\nAssuming the error was due to 32-bit floating-point rounding, the border coordinates should have at least 2 more digits of precision.\n\n", "Volumes look much much different in master compared to 2.79b\nWindows 10 pro x64\nGTX 970\nThread Ripper 1950x\n\n\n\nBasic volume shader with noise texture from build bot today doesnt look the same as 2.79b\n\n\n\n[untitled.blend](untitled.blend)\n", "Adding new hook while in Edit mode breaks when trying to undo/redo the operator\nWin7 (32bit)\n\nBroken: 2.78a\n\n\n- Add Path from Curve menu, add new Armature\n- Assign Bone to control point as Hook\n- Blank Hook modifier appears along with real one\n\nScreencast: [screencast0001-0250.mp4](screencast0001-0250.mp4)\n", "Knife not snapping on backfacing elements (wireframe, xray mode)\nSteps to recreate:\n1. Select object and goto edit mode\n2. Enable X-ray Mode.\n3. Select \"Knife\" Tool.\n4. Now, when i want to snap my knife to the vertices(\"Back Side\") for cut it didnt snap. Although It snaps to vertices on front part", "Support MinGW\nWe have some patches in MINGW-packages to make blender build on MinGW.\nWe already have it for Windows ARM64.\nIs it OK to upstream those patches?", "Radius in place of Size at Empty viewport menu\nI{[F6596506](RIGHT.JPG)}\n\n am new to blender and not know where to post this but plz check it.It seem wrong to me\nthanks", "Regression - depth buffer drawing go bonkers when selecting object on second collection\nOperating system: Linux-5.8.18-1-MANJARO-x86_64-with-arch-Manjaro-Linux 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.100\n\nBroken: version: 2.91.0\nWorked: 2.90\n\nAfter drawing simple 3d lines with python script - depth buffer is drawn incorrectly when deselecting object that is placed on **second collection**. \nScript itself is very generic - from blender gpu docs examples. It worked ok in all previous blender versions.\n\n[Linde_draw_depth_bug.mp4](Linde_draw_depth_bug.mp4)\n\n- run script - it will draw 2 lines in 3d view for 8 seconds\n- if you select/deselect box which is placed on second layer/collection you will see those lines are drawn incorrectly on object deselect.\n\n[3d_view_lines_bug.blend](3d_view_lines_bug.blend)\n\n\n\n", "UI: Show invalid cache as stripped lines\nAt the moment we use a faded color to indicate that a cache is invalid. The difference is a too subtle at times, even when users customize their theme colors.\n\nThe proposal here is to draw the invalid caches with a darker stripped diagonal line (see image). In this case I used the original color with a 50% value.\n\nIn this mockup you see the Simulation cache (pink) as invalid, and the Cloth cache (blue) as valid.\n\n\n\n_Design done with [Pablo Vazquez](pablovazquez)._", "Partial UV selection working incorrectly with Metal backend\nOperating system: macOS-13.2-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Max 1.2\n\nBroken: version: 3.5.0 Beta\n\nUsing hide unselected in the UV editor seems to randomly 'explode' the UVs around the map.\n\nOpen [file](https://projects.blender.orgattachment), go to edit mode.\n\nUV is very messy, but when everything is selected, it's OK. Works fine with OpenGL backend\n" ]

[ "Knife Project Resolution is View Dependent \nWindows 7, x64, gtx 580\n\nBroken: 85c38e7\n\nKnife project resolution seems to be based on the camera zoom level. Artists are expecting that it wouldnt operate like this, ending up with meshes which look inaccurate.\n\n1. Open attached file\n2. knife project\n3. zoom in a lot\n4. notice how horrible it is\n5. undo\n6. knife project\n7. notice the difference\n\n[knifebug.blend](knifebug.blend)" ]

[ "Mulitres: Grab brush + 'Grab Active Vertex' option seems not intuitive or broken 3.2\nOperating system: Windows-11-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.95\n\nBroken: version: 3.2.0\nWorked: (In 3.1 sort of?)\n\nI am not sure if its a bug but here we go:\nWhen you add a mulitres modifier to an object and try to sculpt on it with the grab brush it works.\nIf you enable Grab active vertex in the Brush settings it looks broken to me. Even in 3.1 the vertex are displaced from the original and impossible to grab. \n\n1. Start blender\n2. Add a sphere\n3. Add a multires on lvl 1 or 2\n4. Using the grab brush in sculpt mode works until you enable the grab active vertex option.\n\nThe blend file has the setup just try in sculpt mode the grab brush and the options.\n\n\nOn a side note the bug report said win 10 but i run win 11. It seems not to recognized my system.[GrabActiveVertexBUG.blend](GrabActiveVertexBUG.blend)", "Poor stroke quality when a model is far from the world origin\nOperating system: Win 10\nGraphics card: Nvidia Rtx 2070\n\nBroken: 2.83, master\nWorked: perhaps never?\n\njittery strokes when mesh is far from the world origin in perspective mode\n\nHow to reproduce this bug : \n- Create a grid (set the location to x : 2000 m and y : 1000 m)\n- Create a multires modifier, with 5 subdiv.\n- Sculpt with the clay brush\n- In Perspective mode (left stroke) the stroke is ugly\n- In ortho mode (right stroke) the stroke is clean\nTo fix the stroke poor quality, i had to fine tune the Clip start and the clip end values of the 3d view. \n[#102747.blend](T102747.blend)\n", "Wrong particle distribution\nOperating system: Ubuntu 18.04 LTS\nGraphics card: NVidia Quadro K2000D\n\nversion: 2.80 (sub 75), branch: master, commit date: 2019-07-29 14:47, hash: f6cb5f54494e, type: Release\n\nThe density of randomly generated particles inside of a deformed mesh is not uniform.\n\nThis behaviour is unwanted for a random distribution and leads to visual artifacts as well as inconsistencies in simulations. ('even distribution' does not solve this, the distribution probably has to be normalized on the actual mesh volume)\n\nSet up a particle system in a cube volume and set the distribution to 'random' and generate all particles in the first frame. Take one edge of the cube pointing in z-direction and pull it further away. Press 7 on the numpad to see the effect in the distribution. The effect becomes prominent with around 5000 particles.\n\nThis is the top view of a cube with one vertical edge dragged out:\n\n\n\n[#71731-particleVolume.blend](T71731-particleVolume.blend)\n", "Bug while moving vertices with mirror & subdivision modifier\nOperating system: Windows 10 64-bit\nGraphics card: ATI Radeon HD 5670\n\nBroken: 2.82\nWorked: at least 2.8\n\nWhen I try to move a vertice it seems like the mirror modifier gets applied in an odd way. The result even changes when moving the camera or hide one of the images I use for reference. In fact it becomes distorted when hiding the images.\nThe following link is a recording of my screen:\nview?usp=sharing\nI encountered this bug only recently. The file itself is older and I do not remember which version I used (since I use it via Steam I did not pay attention to the automatic patches).\n\n1. Open this file [Bugged Project.blend](Bugged_Project.blend)\n2. Disable visibility of Subdivision modifier for Edit Mode (which should already cause the bug)\n3. Move a vertice\n\n[system-info.txt](system-info.txt)", "Unify Interpolation In Vertex Paint And Sculpt Mode\nVertex paint squares colors when it interpolates, while the paint tools in sculpt mode do not. This should really be made consistent across both modes. \n\n\n", "The Principled (transparency=1)/Glass BSDF converge incorrectly when using Multiscatter GGX on non manifold meshes.\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.07\n\nBroken: version: 3.0.0\n\nThe Principled (transparency=1)/Glass BSDF converge incorrectly when using Multiscatter GGX on non manifold meshes. This issue is only present with Multiscatter GGX and does not occur with other microfacet distribution models. It also seems to only affect transmission as the Glossy BSDF does not have this issue. Additionally it also seems to only occur in combination with non manifold meshes.\n\nNon manifold mesh:\nGGX+non manifold mesh\n\nMultiscatter GGX+non manifold mesh\n\nThe problem is the sudden incorrect apperance of light in within the red circle in the case of Multiscatter GGX.\n\nManifold mesh:\nGGX+manifold mesh\n\nMultiscatter GGX+manifold mesh\n\nNo problem here.\n\n[Multitest4.blend](Multitest4.blend)\n\nIn this test file the gradient going into the roughness input goes from 0 at the top to 1 at the bottom. There is another hidden manifold cylinder. Unhiding the manifold cylinder shows that this issue does not occur with manifold meshes.\n\nThank you and have a great day!\n", "Incorrect twist in bevel/extrusion using curve tangent mode\n\n\n\nOperating system:\nGraphics card:\n\nBroken: present in 2.83.3, 2.91.0, and 3.0.0 alpha\nWorked: ?\n\n\nBoth open and closed curves exhibit a \"sausage effect\" when a bevel is applied. This is not visible when the curve is closed and 'ResolutionU' is below 4. The issue is somehow attenuated by applying large amounts of smoothing (twist_smooth).\nBelow is a flat and symmetric curve showing this issue.\n\n\n[blender_sausage_bug.mp4](blender_sausage_bug.mp4)\n\n\nDelete default cube. Create a Bèzier Circle. Enter Edit Mode, select all vertices, subdivide once.\nIn Object Data Properties, select Shape->Twist Method-> Tangent, and Geometry->Bevel->Depth: 0.2.\nToggle the Active Spline->Cyclic->U, and change the Active Spline->Resolution U between 4-5.\nYour smooth symmetric curve will turn into sausages!\n\nThe effect is slightly different depending on the type of handles.\n[sausage_bug.blend](sausage_bug.blend)", "Edge/Vert Slide fail to properly \"redo\"\nCurrently vertex & edge-slide use the mouse input to calculate which edge to slide, the direction to slide, and the **Even** toggle is set at runtime.\n\nUsing operator redo is not re-running with the same options, so redoing the action fails.\n\nWe could simply disable redo with this operator, however being able to toggle **Correct UV** is important.\n\nAdding this report to merge in duplicates, since there are various reports on this, which all have the same root-cause.", "Transform Tools influence on Hidden Face Sets with Multires Modifier\nOperating system: Linux-5.3.0-62-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: Quadro GP100/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.100\n\nBroken: version: 2.91.0 Alpha\n\nThere seem to be multiple strange effects that the Transfrom Tools (Move, Rotate, Scale, Transform) have on an object with Multires in Sculpt Mode.\nWhen hiding face sets the boundaries are still being effected but are lagging behind.\nIf the operation is cancelled (Esc) the boundary edges will move half of the way anyway.\nIf the operation is accepted it will unhide all face sets.\n\n[2020-07-27 14-06-58.mp4](2020-07-27_14-06-58.mp4)\n\n- In the default cube, add the Multires modifier\n- Subdivided the mesh a few times.\n- In sculpt mode, draw a Face Set for one of the faces.\n- Hide the rest\n- Enable the Transform tool and move the object.\n", "Pose Mode X-Axis Mirror is not working properly with IK when skeleton is moved\nIntel Core i7, 16 Gb RAM, NVidia GeForce 2070 RTX\nWindows 10 x64\n\nBroken: 2.83 cfdb5d9a8b07 from 2020-02-19\nWorked: 2.83 5df494d75c9a from 2020-01-27 although it seems that this feature was introduced recently so technically it wasn't working before because symmetry didn't work with IK at all.\n\nPose Mode X-Axis Mirror is not working properly with IK when skeleton is moved or rotated from zero in Pose Mode.\n\n1. Open provided test file and try moving selected bone - opposite side will twitch weirdly instead of moving symmetrically.\n2. Now reset the skeleton to its default position and move the same bone - symmetry will work properly.\n3. Rotate first bone and try again - weird twitching again.\n[IK_Symmetry_Error.blend](IK_Symmetry_Error.blend)", "Multires: Editing base mesh creates random results on higher subdiv levels\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.90.1\n\nI applied a multiresolution modifier on this mesh, and it adds geometry that I didn't sculpt on the mesh (bumps, cracks...). It isn't always symmetric, as it happened even after I made more progress in my project. Deleting a face of the mesh and recreating it is enough to generate more of that random geometry. I don't know how to reproduce the problem. Removing and reapplying the multiresolution modifier works, but all the other sculpting is lost.\n[multiresolution.rar](multiresolution.rar)\n[multires_problem.blend](multires_problem.blend)\n", "Exact Boolean with Incorrect Result - multiprecision/float round trips not exact\nOperating system: Windows 64 bit\nGraphics card: Nvidia GTX 750 Ti\n\nBroken: 2.91.0 Beta dca36a8ec922\n\nAnother case, where the exact boolean does not work as expected. (Sorry @howardt, I had to break it. I have to admit, that I have not run into this problem in a real scene.)\nAgain, this manifests after applying the same difference operation several times.\n\n[boolean_rotated_box3.blend](boolean_rotated_box3.blend)\nLook at the scene.\nThe first run of the modifier works as expected, the second creates overlapping faces, the third removes most of the mesh.\nThis \"bug\" is very erratic, and if the transform of the objects is slightly changed, it does not occur.\n\nThis is what I see:\n\n\nThe expected behavior is that the second and third modifier do nothing.", "Greasepencil z fighting alpha issue\nHi,\nnot too sure if this is known issue or any fix but I am getting this alpha issue from certain angle when I have 2 different grease object. \n\n", "Make \"offset edge loop\" slide evenly\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.86\n\n\nBroken: version: 2.80 (sub 75)\n\n\n\n[2019-09-17_15-04-45.mp4](2019-09-17_15-04-45.mp4)\n\nJust repeat the video example\n", "Cycles does not render geo with Multires modifier while in Sculpt mode\nOperating system: Windows 10, Linux\nGraphics card: Nvidia GTX 960\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\nWorked: (optional)\n\nAdding a Multires modifier with the Sculpt parameter > 0 (to further develop the model I'm working on) makes the geo disappear when using Cycles in the viewport, while in Sculpt Mode. As a first time Blender user, I'm not sure if this is the expected behaviour for this. I have checked with other users and they can reproduce the error. \n\nIt would be useful to render in viewport without leaving Sculpt Mode.\n\nCreate a geometry. Add a Multires modifier. Set Sculpt to a value greater than 0. Enter Sculpt Mode. Change the viewport render from Eevee to Cycles. The geometry disappears on rendering. I have a sample of the problem in this tweet I published several hours ago: 1150657198374080512\n\nThanks!" ]

[ "The problem with the symmetry in the mode of sculpting\n%%%--- Operating System, Graphics card ---\nWindows 7 x64 (SP1), GeForce GTX 660\n\n\n - Blender version with error, and version that worked ---\n2.66.1 \nr55078\n\n - Short description of error ---\nIf you use a mode of symmetry along the X axis and try to apply the brush mode Smooth, in the center of a symmetrical object, its central vertex shift relative to its center.\n\n\n - Steps for others to reproduce the error (preferably based on attached .blend file) ---\n1. We need to create any symmetrical object in which the central peaks will be located exactly on the axis X. \n2. Switch to sculpting and activate the symmetry axis X \n3. Spend some time in the brush mode Smooth (shift LMB) in the center of the object. \n4. Press Z and look closely at the central peak in zoom mode. \n5. It can be seen that some of the peaks previously located on the X axis shifted\n%%%" ]

[ "Adjust modifier values when using \"Apply scale\"\nThis feature request was discussed with Campbell Barton during meeting 2021-02-23\n\nWhen using \"Apply scale\", the modifier values are not adjusted if the object scale is not [1,1,1]. I would like to suggest that the operator \"Apply scale\" iterates through all modifiers of the selected objects and adjusts all values related to scale\n\nFor example\nBevel modifier - Amount\nDisplace - Strength\nData transfer - max distance, ray radius\nUV project - Scale X, Scale Y, \netc\n\n---\n\nCurrent behaviour\n{[F9849514](image.png), height = 700}\n\n---\n\nDesired behaviour\n{[F9849529](image.png), height = 700}\n\n---\n\nsuggested Operator mockup\n\n{[F9852183](image.png), height = 300}", "Perfomance drop if Texture Limit Size enabled \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.92.0\n\n[2021-04-02 12-29-11.mp4](2021-04-02_12-29-11.mp4)\n\n1. Add 4k cheker map to default cube\n2. Go to material preview\n3. Go to UV edit, manipulate UVs\n4. Toggle Texture Size Limit \n\n[#87133.blend](T87133.blend)\n", "Modifiers header misalignment on change Resolution Scale [pixel perfect thing]\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) HD Graphics 4000 Intel 4.0.0 - Build 10.18.10.5161\n\nBroken: version: 2.92.0; blender-2.90.0-d7c4e9649320\nWorked: don't know\n\nModifiers/Constraints header buttons misalignment on different (especially smaller) Resolution Scale value\n\nSimple as that. Tried set Resolution Scale down on a notebook, and saw the modifier name input \"jump\" out of the modifier box.\nSo checked that any value bellow 1, varnishes the \"top margin\" of it. The same on constraints, bones included.\nThan i suppose it has something to do with the dragging feature.\n\nOpen Blender.\nAdd any modifier/constraint an object.\nGo to Preferences -> Interface -> Display -> Resolution scale, and click the input's left arrow.\nSe the bug on the modifier header\n\n\n\n", " Greasepencil Hue/Saturation modifier\nOperating system:windows 10\nGraphics card:GTX 1060\n\nBroken:2.93beta\n\nHue/Saturation modifier can not take effect on fill-gradient", "Bone Heat weighting: failed (due to none of the vertices in an island being visible to any bones)\n**Win 10 64. gtx 970**\n\n**Blender 2.75a Hash c27589e**\n\n**Bone Heat weighting: failed to find solution for one or more bones**\n\nI attach the blend file and two prints..\n\nAlso the same broblems apper when i try to \"assign automatic from bones\" by individual bones in Weight Paint. Like one by one. Please see print2.\n\n\n\n\n\n[cat.blend](cat.blend)\n", "Sculpt Paint: Accumulative Painting/Sculpting\nNOTE: This design is still under discussion and is incomplete.\n\nThe recent addition of painting tools in sculpt mode also added an alternative design for accumulative painting.\nBecause of this there are now two different ways of handling accumulative strength.\nThis needs to be polished and properly integrated into the UI and painting workflow.\n\n# Strength and Accumulate\n\nSo far brushes always had an \"Accumulate\" setting for removing the maximum strength cap. Brushes would then keep adding their effect as long as the stroke continues.\nOtherwise they would either accumulate up to a hard-coded upper strength cap (sculpt mode) or up to the set strength (painting modes).\n\nThe issues with this implementation are:\n- It is easy to accidentally clip the colors while painting (There is no upper cap with accumulate)\n- The setting is hidden in \"Advanced\" brush settings even though it is regularly used, and even preferred for painting.\n\n# Flow\n\nThe new painting brush in sculpt mode instead uses a \"Flow\" slider, which defines how strong each step of the stroke is. \nThe stroke will then keep accumulating more slowly towards the set Strength.\n\nThere are a few issues with this implementation:\n- The Flow setting is less accessible, even though it can sometimes to be the preferred way of changing the stroke intensity.\n- The relationship between Strength and Flow becomes very confusing, especially when comparing sculpting & painting brushes or when enabling pressure sensitivity settings\n\n# Proposal\n\nHow to make the settings and behaviours more predictable and accessible needs to be further discussed.", "Edges generated by Bevel Modifier do not inherit correct crease values\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.02\n\nBroken: version: 2.81 (sub 14)\nWorked: (optional)\n\nThe Bevel Modifier ignores the crease values of connected edges, producing unexpected/undesired results.\n\nExample image of the bug in production asset (the rear of a flatbed car):\n\nPlease observe how the top is rounded toward the center of the car, despite the vertical corner crease and the horizontal beveling. This roundness goes against the artist's intention.\n\n\nOpen this file: [#70587-bevel-edges-bug.blend](T70587-bevel-edges-bug.blend)\n\nThat file was created by following these steps:\n\n1. Create a cube\n2. Set Edge Crease of vertical edges to 1.0\n3. Set Edge Bevel Weight of top horizontal edges to 1.0\n4. Add Bevel Modifier in Weight Mode\n5. Add Subdivision Modifier\n\nThis produces undesired result 1:\n\nPlease observe how the top of the cube becomes circular because the creasing on the vertical edges does not extend to the newly generated loop of faces.\n\nThis can be seen in image 2:\n\n\nOnly after manually adding the crease value to the new edges do we get the expected result, image 3:\n\n", "Mesh Deform modifier doesn't deform after binding\nBlender 3.3.0 (hash: 0759f671ce1f)\n\nMesh Deform modifier simply doesn't work. The cage mesh doesn't deform the effected mesh.\n\nI modeled my base mesh (a simple array of duplicated rectangles). Then i added a cube, as a separate object, entered Edit Mode, adjusted the size to fit my base mesh, added some loop cuts, exited Edit Mode. Selected my base mesh, added the Mesh Deform modifier, chose the object to deform with, clicked on Bind. \n\n[mesh deform.blend](mesh_deform.blend)\nThe deformation mesh is a solid geometry. I just changed the Viewport Display option to Wire. ", "modifier offset slider step value being driven by the mouse looses one order of magnitude worth of accuracy when units are metric vs none\nwin 10\n2080ti\n\nBroken: 2.80+\n\nSince blender 2.8 has switched to real world units the modifier slider step(when driven by the mouse) value has lost one order of magnitude worth of accuracy, 0.001 instead of 0.0001 \nIt is possible to switch the unit system back to none to get the higher accuracy. \nThey should be consistent and both work to the higher level of accuracy as this level is essential when working on real work scale small items e.g. a watch\n\n\n- with scene unit type set to \"metric\" add a bevel modifier the standard cube\n- press shift while clicking and dragging the bevel modifier offset slider and notice how the smallest amount possible to move is 0.001 m\n- switch scene unit type to \"none\"\n- press shift and click drag the modifier offset slider and notice the smallest amount possible to move is 0.0001 m\n", "Annotations do not scale in the Shader Editor with the Resolution Scale or Windows Display Scaling\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1050 Ti with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.13\n\nBroken: version: 2.83.3\nWorked: Couldn't find one. Tried 2.90 beta too.\n\nAnnotations made in the Shader Editor do not scale with the interface, either with the Preferences resolution scaling or with Windows Display scaling. This is problematic when reviewing annotations made on a computer using a different screen scale as notes do not line up.\n\nGo to Shader Editor - Add default material/nodes for scale reference.\nT - for left toolbar. Select Annotation tool.\nAdd annotation to the screen, e.g. draw around a node.\nOn the Menu > Edit > Preferences > Interface \nChange the resolution scale.\nNodes will scale. The annotation marks will not.\n\nSame behaviour is observed by changing the Windows Scaling % in the Windows Display 'Scale & Layout' menu.\n\n\n\n", "Hair with material using a non-existing attribute looks different between Eevee and Cycles\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.0 Core Profile Context 22.20.00.20.220516\n\nBroken: version: 3.4.0 Alpha\nWorked: -\n\nHair can be affected by materials that use attributes such as mesh.\nThe color of the mesh at the root position of the strand usually defines the color.\nBut that's not the case in Eevee if an attribute doesn't exist.\n\n\nThe actual color can be kind of undefined, but there should be a convention between the renders.\n\n- Open attached .blend file\n- Compare Cycles and Eevee\n[hair_material_with_non-existing_attribute__bug.blend](hair_material_with_non-existing_attribute__bug.blend)\n\n", "Vertex Weight Transfer in LineArt may not work when Remove Double option is enabled\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Radeon(TM) RX 460 Graphics ATI Technologies Inc. 4.5.13596 Core Profile Context 20.10.35.02 27.20.1034.6\n\nBroken: version: 2.93.0 Beta\nWorked: (the newest version of Blender that worked as expected)\n\nVertex Weight Transfer in LineArt may not work when Remove Double option is enabled. Following is a gif.\n\n\n1. Open the following blend file. \n1. Toggle Remove Double option.\n[remove-double-bug.blend](remove-double-bug.blend)\n", "Resetting default value does not work in popovers or menus\nOperating system: Windows 10, 64 bits, \nGraphics card: Nvidia GTX 1080 8 gb, Nvidia GTX1070 8 gb\n\nBroken:\nblender-2.81.0-git.ce2368cf69f5-windows64\n\nWorked: (optional)\n\nResetting options in the context menu (right click) of things like brush stroke strength, Autosmooth, plane offset and other parameters does not work at all, but does work through active Tool side panes and top bar. Probably it should work everywhere? If I'm working full screen I would want to be able to reset it in the context menu with my stylus as I sculpt.\n\n\n\n1 Create any mesh\n2 Enable Sculpt mode\n3 Select any brush\n4 Right click to invoke context menu\n5 Try resetting any of the parameter to default values\n", "Resetting singular values in the Themes disables them\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 2.93.5\n\nInstead of globally resetting the default themes, if I use the \"Reset to default\" for a color or a value, it simply puts it to black/0.0.\n[2021-10-27 14-12-52.mp4](2021-10-27_14-12-52.mp4)\n\nGo to themes and right click a color or value.\n[Based on the default startup]\n\n", "Boolean Modifier not working \nOperating system: Windows \nGraphics card: Nvidia Geforce GTX 770\n\nBroken: 2.93.4\nWorked: 2.90\n\nIf one tries to apply the boolean modifier \"Difference\" it does nothing. In previous version it functions.\n\n1. Download the Blender File from above \n2. There a 3 objects in the collections. The objects are in the positive xy-quardant. (Sorry that they are not centered)\n3. If one tries to apply the modifier on the Frame object there is no error message but the frame object doesnt get modified.\n\nGreetings Philip \n\nTest File:\n[#71360.blend](T71360.blend)" ]

[ "Weight Paint Gradient does not reflect settings instead relies on Brush Settings for Weight-value and/or Falloff-Settings of Brush\nI noticed that when weigthpainting a Gradient with the Gradient-tool it sometimes not works and/or works differently then expected - when i switch back to weight-paint-Brush and edit certain settings like weight-value and Falloff-Settings then the Gradient behaves as expected - while doing the same in gradient-tool has no effect. Leading me to think that the Weightpaint-Brushsettings are relevant for the Gradient settings (i expect them to have separate settings).\nThey seem connected but in the Gradient-tool-settings i dont have control over weight-value - i have to switch back to brushsettings.\nFalloff-settings for the Gradient-tool are essential - this should imo reflect in the Gradient-tool-settings also as the brush-falloff-settings govern the weight-paint-gradient as it stands now. sorry to bother :/\n\n\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.66\n\nBroken: version: 2.83 (sub 4)\nWorked: (optional)\n\n\n\n", "Weight painting: Gradient tool does not have its own settings such as falloff adjustment or blend mode settings\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.90.0\n\nWeight paint: Gradient tool does not have its own setting for its functionality such as falloff or blend mode. These settings are copied from Draw mode and there are no indication that this is happening for the user.\n\n\n1. Select or create cube or any other primitive object.\n2. Go to weight paint mode with selected object.\n3. Select Gradient tool and try to create gradient onto the model - try to change falloff curve or blend mode setting - can't do that since there is no such settings when this tool is selected.\n4. Select Draw tool and modify brush settings: falloff curve and blend mode setting. \n5. Select Gradient tool and try to create weight gradient with it onto the model - see how the gradient is affected by the blend mode and falloff curve of different tool.\n\n\n", "Gradient weight tool confusing (using previous tool brush)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.82 (sub 7)\nWorked: (optional)\n\nGradient Tool in its current state is not very user friendly:\n\n- generally this accesses previously selected tools brush as it doesnt have its own brush (so it looks like it has its own settings for strength/weight, but those are from an 'alien' brush) when `Unified Weight` is turned ON (default) the tool actually works with the `Unified Weight` (set on some other brush, e.g. the Draw Brush), but doesnt present you with that value in the UI (bug!)\n- when `Unified Weight` is turned OFF you are actually tweaking the weight setting of the previously selected tools brush (the 'alien' one) -- this causes confusion (it doesnt affect e.g. `Blur`, `Average` or `Smear` since these dont have weight capabilities)\n- this is even worse when tweaking the Strength, you are actually tweaking the weight setting of the previously selected tools brush (the 'alien' one), this affects `Blur`, `Average` or `Smear` (and Unified Strength is OFF by default!)\n\nSo not sure how to solve the issue with the \"tweaking setting of a seemingly unrelated brush\" issue, what I think we should do though is at least have the values in the UI respect Unified Weight / Unified Strength since this is what the tool will use internally...\n\nSo the main issue is fixed in f38c54d56e: Fix Vertex weight gradient tool show wrong weight/strength values in the UI.\n\n" ]

[ "Mask brush stops working in tiling sculpting mode after changing its size through \"F\" hotkey\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 2.93, 3.3.0\nWorked: -\n\nWhen an object has the Multires modifier, mask brush stops working in tiling sculpting mode after. (See video below)\n[Desktop 2022.09.17 - 21.18.10.07.mp4](Desktop_2022.09.17_-_21.18.10.07.mp4)\n\n- Open attached file or\n - Add a grid object\n - Add a multires modifier\n - Subdivide Simple a few times\n - Select mask brush in sculting mode\n - set tiling on x and y axis to 1 or 2m \n- try paint mask (sometimes the tiling doesn't work at this stage, you need to paint again and undo)\n- change brush size through F hotkey\n- try paint mask (tiling stops working)\n- press undo\n- try paint mask (tiling works again)\n\n[tiling_sculpt_bug.blend](tiling_sculpt_bug.blend)\n\n", "Regression: Crash when animation is playing\nOperating system: Windows 10\nGraphics card: 2x RTX3080\n\nBroken: 3.6 and 3.6.1 release canditate\n\nPress the space bar, the animation starts and Blender closes. The console says : EXCEPTION_ACCESS_VIOLATION\nI also test the file with an independent portable version of Blender, but the crash still occurs.\n\nPress space bar to play the animation and wait the frame 48.\n\n[Fluent_Testing_7_for_blend_dev.blend](https://projects.blender.orgattachment)", "Cloth simulation randomly breaks down when scrubbing timeline\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.2.1\n\nWhen scrubbing the timeline , the cloth simulation may cause the mesh to become distorted even with the existence of cache. The video below will show a quick way to break the simulation, however, I first stumbled upon the problem when I set the keys for the bones.\n\n[2022-07-22 15-48-43.mp4](2022-07-22_15-48-43.mp4)\n\n1. Create cloth simulation.\n2. Scrub the timeline. After some time, a situation will arise where the mesh will be distorted.", "Voxel Bug Holes after Remesh\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.95\n\nBroken: version: 3.2.0\n\nVoxel Bug Holes after Remesh\n\n\nClick on Voxel Remesh, I think any voxel size produces the same result.\n\n\nI tried on default sculpt spear it works fine that way but not on the model I am using.\n\nThe model was originally produced from Reality Capture software, then I cleaned it up using decimate and smooth modifier. I fixed the details and did some more cleaning in Sculpt Mode. Unfortunately when I tried Remesh this issue happened.\n\nThank you in advance.\n\n[VeniceBaseSculptingVoxelBug.blend](VeniceBaseSculptingVoxelBug.blend)\n\n", "Outliner crash/assert when deleting an override on a mesh\nOperating system: Linux-6.1.18-200.fc37.x86_64-x86_64-with-glibc2.36 64 Bits, WAYLAND UI\nGraphics card: AMD Radeon Graphics (renoir, LLVM 15.0.7, DRM 3.49, 6.1.18-200.fc37.x86_64) AMD 4.6 (Core Profile) Mesa 22.3.7\n\nBroken: version: 4.0.0 Alpha\nWorked: never\n\nOutliner crash/assert when deleting an override on a mesh\n\nIn a debug build:\n- open `main.blend`\n- clear the override on the mesh in the Outliner\n\n\nThere seems to be something wrong with updating the tree, it still wants to draw the mesh override?\n(tried some additional stuff with `DEG_relations_tag_update`, `outliner_cleanup_tree`, `BKE_view_layer_synced_ensure` in `id_override_library_clear_single_process` to no avail...)\n\n```\n==1574045==ERROR: AddressSanitizer: heap-use-after-free on address 0x61c00030490c at pc 0x00000b189b87 bp 0x7ffe8b70fea0 sp 0x7ffe8b70fe98\nREAD of size 4 at 0x61c00030490c thread T0\n #0 0xb189b86 in outliner_but_identity_cmp_context_id_fn /blender/source/blender/editors/space_outliner/outliner_draw.cc:1973\n #1 0xa1946d9 in ui_but_equals_old /blender/source/blender/editors/interface/interface.cc:732\n #2 0xa19822c in ui_but_update_from_old_block /blender/source/blender/editors/interface/interface.cc:967\n #3 0xa1a07ab in UI_block_update_from_old(bContext const*, uiBlock*) /blender/source/blender/editors/interface/interface.cc:1869\n #4 0xa1a1cea in UI_block_end_ex(bContext const*, uiBlock*, int const*, int*) /blender/source/blender/editors/interface/interface.cc:1964\n #5 0xa1a2cd5 in UI_block_end(bContext const*, uiBlock*) /blender/source/blender/editors/interface/interface.cc:2058\n #6 0xb1a58dc in blender::ed::outliner::draw_outliner(bContext const*) /blender/source/blender/editors/space_outliner/outliner_draw.cc:4097\n #7 0xb210d95 in outliner_main_region_draw /blender/source/blender/editors/space_outliner/space_outliner.cc:88\n #8 0x7025c43 in ED_region_do_draw(bContext*, ARegion*) /blender/source/blender/editors/screen/area.cc:539\n #9 0x42a3126 in wm_draw_window_offscreen /blender/source/blender/windowmanager/intern/wm_draw.cc:1000\n #10 0x42a482a in wm_draw_window /blender/source/blender/windowmanager/intern/wm_draw.cc:1166\n #11 0x42a78e4 in wm_draw_update(bContext*) /blender/source/blender/windowmanager/intern/wm_draw.cc:1564\n #12 0x428bfd8 in WM_main(bContext*) /blender/source/blender/windowmanager/intern/wm.cc:645\n #13 0x8ef1a5 in main /blender/source/creator/creator.cc:593\n #14 0x7f56eb04a50f in __libc_start_call_main (/lib64/libc.so.6+0x2750f)\n #15 0x7f56eb04a5c8 in __libc_start_main@GLIBC_2.2.5 (/lib64/libc.so.6+0x275c8)\n #16 0x8ee494 in _start (/build_linux_debug/bin/blender+0x8ee494)\n\n0x61c00030490c is located 140 bytes inside of 1712-byte region [0x61c000304880,0x61c000304f30)\nfreed by thread T0 here:\n #0 0x7f56f4cb9388 in __interceptor_free.part.0 (/lib64/libasan.so.8+0xb9388)\n #1 0x4278fad in MEM_lockfree_freeN /blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:104\n #2 0xdd817a in BKE_id_free_ex /blender/source/blender/blenkernel/intern/lib_id_delete.cc:170\n #3 0xdd9f4d in id_delete /blender/source/blender/blenkernel/intern/lib_id_delete.cc:366\n #4 0xdda2a9 in BKE_id_delete_ex /blender/source/blender/blenkernel/intern/lib_id_delete.cc:384\n #5 0xdda304 in BKE_id_delete /blender/source/blender/blenkernel/intern/lib_id_delete.cc:389\n #6 0xb1d5a80 in id_override_library_clear_single_process /blender/source/blender/editors/space_outliner/outliner_tools.cc:1489\n #7 0xb1da009 in outliner_liboverride_operation_exec /blender/source/blender/editors/space_outliner/outliner_tools.cc:1959\n #8 0x43624d5 in WM_menu_invoke_ex(bContext*, wmOperator*, wmOperatorCallContext) /blender/source/blender/windowmanager/intern/wm_operators.cc:1046\n #9 0x4362a0b in WM_menu_invoke(bContext*, wmOperator*, wmEvent const*) /blender/source/blender/windowmanager/intern/wm_operators.cc:1071\n #10 0x42b76d9 in wm_operator_invoke /blender/source/blender/windowmanager/intern/wm_event_system.cc:1530\n #11 0x42b943e in wm_operator_call_internal /blender/source/blender/windowmanager/intern/wm_event_system.cc:1731\n #12 0x42b96a7 in WM_operator_name_call_ptr(bContext*, wmOperatorType*, wmOperatorCallContext, PointerRNA*, wmEvent const*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:1779\n #13 0x42babf9 in WM_operator_name_call_ptr_with_depends_on_cursor(bContext*, wmOperatorType*, wmOperatorCallContext, PointerRNA*, wmEvent const*, char const*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:1972\n #14 0xa2348a7 in ui_apply_but_funcs_after /blender/source/blender/editors/interface/interface_handlers.cc:1038\n #15 0xa2c8f30 in ui_popup_handler /blender/source/blender/editors/interface/interface_handlers.cc:11657\n #16 0x42b0485 in wm_handler_ui_call /blender/source/blender/windowmanager/intern/wm_event_system.cc:818\n #17 0x42d0410 in wm_handlers_do_intern /blender/source/blender/windowmanager/intern/wm_event_system.cc:3319\n #18 0x42d17cf in wm_handlers_do /blender/source/blender/windowmanager/intern/wm_event_system.cc:3436\n #19 0x42dbd54 in wm_event_do_handlers(bContext*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:4063\n #20 0x428bfc0 in WM_main(bContext*) /blender/source/blender/windowmanager/intern/wm.cc:639\n #21 0x8ef1a5 in main /blender/source/creator/creator.cc:593\n #22 0x7f56eb04a50f in __libc_start_call_main (/lib64/libc.so.6+0x2750f)\n\npreviously allocated by thread T0 here:\n #0 0x7f56f4cba097 in calloc (/lib64/libasan.so.8+0xba097)\n #1 0x42796e2 in MEM_lockfree_callocN /blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:210\n #2 0x42646c0 in DNA_struct_reconstruct /blender/source/blender/makesdna/intern/dna_genfile.cc:1236\n #3 0x44e9d47 in read_struct /blender/source/blender/blenloader/intern/readfile.cc:1810\n #4 0x44f74ff in read_libblock /blender/source/blender/blenloader/intern/readfile.cc:2955\n #5 0x450125d in blo_read_file_internal /blender/source/blender/blenloader/intern/readfile.cc:3624\n #6 0x44d538b in BLO_read_from_file /blender/source/blender/blenloader/intern/readblenentry.cc:414\n #7 0x913295 in BKE_blendfile_read /blender/source/blender/blenkernel/intern/blendfile.cc:1014\n #8 0x4301733 in WM_file_read(bContext*, char const*, ReportList*) /blender/source/blender/windowmanager/intern/wm_files.cc:1024\n #9 0x430e599 in wm_file_read_opwrap /blender/source/blender/windowmanager/intern/wm_files.cc:2718\n #10 0x430f626 in wm_open_mainfile__open /blender/source/blender/windowmanager/intern/wm_files.cc:2841\n #11 0x430e957 in operator_state_dispatch /blender/source/blender/windowmanager/intern/wm_files.cc:2754\n #12 0x430f889 in wm_open_mainfile_dispatch /blender/source/blender/windowmanager/intern/wm_files.cc:2865\n #13 0x430ec2d in wm_open_mainfile__discard_changes /blender/source/blender/windowmanager/intern/wm_files.cc:2793\n #14 0x430e957 in operator_state_dispatch /blender/source/blender/windowmanager/intern/wm_files.cc:2754\n #15 0x430f889 in wm_open_mainfile_dispatch /blender/source/blender/windowmanager/intern/wm_files.cc:2865\n #16 0x430f8b6 in wm_open_mainfile_invoke /blender/source/blender/windowmanager/intern/wm_files.cc:2870\n #17 0x42b76d9 in wm_operator_invoke /blender/source/blender/windowmanager/intern/wm_event_system.cc:1530\n #18 0x42b959e in wm_operator_call_internal /blender/source/blender/windowmanager/intern/wm_event_system.cc:1765\n #19 0x42b96a7 in WM_operator_name_call_ptr(bContext*, wmOperatorType*, wmOperatorCallContext, PointerRNA*, wmEvent const*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:1779\n #20 0x42babf9 in WM_operator_name_call_ptr_with_depends_on_cursor(bContext*, wmOperatorType*, wmOperatorCallContext, PointerRNA*, wmEvent const*, char const*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:1972\n #21 0xa2348a7 in ui_apply_but_funcs_after /blender/source/blender/editors/interface/interface_handlers.cc:1038\n #22 0xa2c8f30 in ui_popup_handler /blender/source/blender/editors/interface/interface_handlers.cc:11657\n #23 0x42b0485 in wm_handler_ui_call /blender/source/blender/windowmanager/intern/wm_event_system.cc:818\n #24 0x42d0410 in wm_handlers_do_intern /blender/source/blender/windowmanager/intern/wm_event_system.cc:3319\n #25 0x42d17cf in wm_handlers_do /blender/source/blender/windowmanager/intern/wm_event_system.cc:3436\n #26 0x42dbd54 in wm_event_do_handlers(bContext*) /blender/source/blender/windowmanager/intern/wm_event_system.cc:4063\n #27 0x428bfc0 in WM_main(bContext*) /blender/source/blender/windowmanager/intern/wm.cc:639\n #28 0x8ef1a5 in main /blender/source/creator/creator.cc:593\n #29 0x7f56eb04a50f in __libc_start_call_main (/lib64/libc.so.6+0x2750f)\n\nSUMMARY: AddressSanitizer: heap-use-after-free /blender/source/blender/editors/space_outliner/outliner_draw.cc:1973 in outliner_but_identity_cmp_context_id_fn\nShadow bytes around the buggy address:\n 0x0c38800588d0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00\n 0x0c38800588e0: 00 00 00 00 00 00 00 00 fa fa fa fa fa fa fa fa\n 0x0c38800588f0: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa\n 0x0c3880058900: fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa fa\n 0x0c3880058910: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n=>0x0c3880058920: fd[fd]fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c3880058930: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c3880058940: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c3880058950: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c3880058960: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c3880058970: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\nShadow byte legend (one shadow byte represents 8 application bytes):\n Addressable: 00\n Partially addressable: 01 02 03 04 05 06 07 \n Heap left redzone: fa\n Freed heap region: fd\n Stack left redzone: f1\n Stack mid redzone: f2\n Stack right redzone: f3\n Stack after return: f5\n Stack use after scope: f8\n Global redzone: f9\n Global init order: f6\n Poisoned by user: f7\n Container overflow: fc\n Array cookie: ac\n Intra object redzone: bb\n ASan internal: fe\n Left alloca redzone: ca\n Right alloca redzone: cb\n==1574045==ABORTING\n```\n\n", "Regression: Crash when Sculpting on a mesh from a GP Lineart Modifier\nOperating system: Windows-10-10.0.22621-SP0 64 Bits (W11)\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.2, 4.0.0 Alpha\nWorked: 3.1\n\n\nWhen trying to sculpt a mesh in an assigned collection that a grease pencil lineart modifier references and the layer it is drawing to contains Grease Pencil strokes - it will instantly crash Blender. \n\n\n1. Download and extract zip to get the test blend file\n2. Load attached file with Load UI off\n3. Select the bust, it is in a GP lineart modifier collection \n4. Go to sculpt mode\n5. Try sculpt\n6. Crash\n\nAlternative steps:\n1. Create new scene\n2. Create a sphere\n3. Create a Grease Pencil object\n4. Add sphere to a new collection\n5. Select the GP object and add a lineart modifier\n6. Reference the sphere collection in the modifier\n7. Assign layer and material to the modifier\n8. Draw in the same layer in Draw mode of the GP\n9. switch to sculpt mode\n10. Try sculpt over sphere surface\n11. Instant crash\n\n", "Render cloth animation does not resets cache\nBroken: 2.82.7\nWorked: 2.79b\n\nWhen modifier Subdivision is before the cloth physics modifier,but with different subdivision settings for rendering and viewport, the cache remains active even when rendering.\nPossibly related to #74523.\n\n- Open attached file\n- Run animation (without baking)\n- Render animation\nNotice that the cloth is deformed with each rendered frame.\nThis is because rendering reuses the cache.\n\n[bug.blend](bug.blend)", "LineArt; Crahes when I reopen a file\nAn old HP laptop with\nOperating system: Linux Ubuntu 18.04.5 LTS\nGraphics card: Mesa DRI Intel(R) HD Graphics 5500 (BDW GT2)\n\nBroken: Blender 2.93.0 Alpha, 3ea1779365b5, master, 2021-03-23 18:34\n\n\nI added two GP LineArt on the scene and it worked and then I saved the scene, with new grease pencil modifiers activated, but when I try to reopen in Blender the same scene crashes every times and I'm not able to see anythings after waiting the processing until crashes without any report. the file is attached.\n\n[yach.tar.xz](yach.tar.xz)", "Combine crash\nOperating system: win 10\nGraphics card: 3090\n\nBlender Version 3.5.1 \n\nWhen using ctrl+j to combine components, blender will crash.\n\n", "Particle hair simulation not working\nOperating system: Windows 10\nGraphics card: GTX 1070ti\n\nBroken: 2.93.4\nWorked: unknown\n\nI modeled particle-hair and try to make it dynamic, for whatever reason I cannot make it to work and sometimes Blender crashes.\n\n[Test.blend](Test.blend)\n\n[Test.crash.txt](Test.crash.txt) \n\nIf I try another simple particle-hair system on the-same object it does work. \n\nTry to get the hair in attached file to simulate.", "Sculpt undo bug, mixing sculpt brush and multi-res modifier changes.\nWith current master (also 2.83).\n\nThe following steps show and error when redoing sculpt mode undo, when mixing multi-res subdivisions.\n\n- Startup file\n- Add multi-res modifier.\n- Enter sculpt mode.\n- *Repeat this 3-5 times.*\n - Multi-res subdivide.\n - Brush Stroke.\n\n- Now undo until the very beginning.\n- Now redo until the very end.\n\nWhen redoing, there are glitches where the multi-res has tearing/artifacts.\n\nSee video: [sculpt_undo_bug.mp4](sculpt_undo_bug.mp4)", "Particle Edit mode for non hair particles - 2 Crashes\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 760/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.65\n\n\nBroken: version: 2.80 (sub 69)\nWorked: (optional)\n\n\nParticle Edit: 2 Crashes\n\nParticle Edit: 2 Crashes\n\n- Open the blender file and click on the buttons \"Particle select and display point\"\n- Crash 1 (Screenshot)\n\n- Open the blender file and click on the tool \"Weight\" and click into the particles\n- Crash 2 (Screenshot)\n\n- Also had a crash with \"Puff\" \n- file not included\n\n- There are some tools that do nothing while others work\n- User gets no feedback / error message but give no feedback\n\n\n\n[2019-05-19-particle-edit-crash.blend](2019-05-19-particle-edit-crash.blend)\n\n", "Crash with mantaflow\nOperating system: Windows-10-10.0.18363-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.0.0 Alpha\n\nBaking simulation causes crash\n\n[splash.blend](splash.blend)\n\nOpen file and bake liquid\n", "Crash when rendering a specific 4k scene with Cycles Metal\nOperating system: macOS-13.3.1-arm64-arm-64bit 64 Bits\nGraphics card: Apple M2 Pro Apple 4.1 Metal - 83.1\n\nBroken: version: 3.5.1\n\nWhen I render with Blender 3.5.1 on a MacBook 16\" with an M2 Pro chip at a 4K quality, Blender closes before finishing the rendering process.\"\n\n- Open attached .blend file\n- Make sure you have enabled Metal GPU+CPU in the Preferences\n- Render the scene\n\nIt should crash about halfway through the rendering process.\n\n\n", "Segmentation fault when selecting context.area.type before applying operator (scripting)\nOS: Ubuntu 16.04.4 LTS\nHW: Dell XPS 9550 notebook, using Intel HD GPU (nvidia deactivted)\n\nBroken: 2.83\n\nExecuting the operator `bpy.ops.transform.translate` causes a crash if `bpy.context.area.type` is set manually before.\nCrash in `setTransformViewMatrices` because `t->ar->regiondata` is `NULL`\n\nRunning the following script causes the crash:\n```\nimport bpy\nimport bmesh\n\n# Create object\nbpy.ops.mesh.primitive_plane_add(size=1, enter_editmode=False, location=(0, 0, 0))\n\n# Enter edit mode\nbpy.context.area.type = 'VIEW_3D' # If this is not set, crash does not occur\nbpy.ops.object.mode_set(mode='EDIT', toggle=False)\n\n# Modify vertices\nmesh = bmesh.from_edit_mesh(bpy.context.active_object.data)\nmesh.verts.ensure_lookup_table()\nmesh.verts[0].select = True\n\n# Next call causes crash:\nbpy.ops.transform.translate(value=(0, 0, 1), constraint_axis=(False, False, True),\n mirror=False, use_proportional_edit=True,\n proportional_edit_falloff='SHARP')\n```\n" ]

[ "Mask after remesh causes crash\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.92.0 Alpha\n\nIf you voxel remesh an object in sculpt mode and immediately apply a mask (regular or box), Blender crashes. If you go to object mode after the remesh and then go back to sculpt mode, there is no crash.\n\n1) Go to sculpt mode with default cube\n2) Apply voxel remesh with Ctrl+R\n3) Try applying a box mask, Blender crashes\n\n", "Blender crash when using the mask brush on a recently remeshed sphere\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.09\n\nBroken: version: 2.92.0 Alpha\nWorked: 2.91\n\nWhen I tried to paint a mask on a remeshed (Ctrl+R) model Blender crashes. It works fine on objects that haven't been remeshed or if I close and re-open the file.\n\n- Select any of the spheres, go to sculpt mode and paint a mask on it. It should work fine.\n- Now remesh the object with Ctrl+R and try to paint the mask again, Blender will crash immediately.\n- Close and open the file again and there will be no issues to paint the mask.\n\n[Sphere_SculptMask_Crash.blend](Sphere_SculptMask_Crash.blend)\n\n", "Sculpt: crash in Sculpt Mode when using mask operators after voxel remeshing\nOperating system: Linux-5.4.0-53-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce GTX 670/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 435.21\n\nBroken: version: 2.92.0 Alpha\n\nCrash in Sculpt Mode when trying to use (almost?) any mask operator **after** using voxel remesh (line mask, box mask, lasso mask, invert mask, clear mask...).\n\n\n1. Load factory setting\n2. Select default cube, enter Sculpt Mode\n3. Remesh geometry with Voxel Remesh \n4. Use Box Mask, lasso mask, etc on the voxel remeshed default cube \n\nImportant note: it only crashes when using the voxel remesh **inside** Sculpt Mode and **then** using mask operators (without exiting Sculpt Mode). \nSo it doesn't crash in cases like:\n1. In Object Mode use Voxel Remesh > Enter Sculpt Mode > Use any mask operator\n2. In Sculpt Mode use Voxel Remesh > Exit Sculpt Mode > Enter Sculpt Mode > use any mask operator\n", "Sculpt crash after remesh (mask operators, cloth brush, ...)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.92.0 Alpha\nWorked: 2.91\n\nBlender crash in Sculpt Mode when Lasso Mask is used after performing remesh (Crtl_R).\n\n\n1. Create new default scene\n2. Delete default cube and lights\n3. Create a UV Sphere\n4. In Sculpting workspace use Lasso Mask\n5. Result - Lasso Mask works as expected\n6. Perform a remesh (Ctrl+R) with default settings\n7. Result - Blender crash", "Sculpt: Activate Expand Mask on mesh w/ Voxel Remesh Crashes\nOperating system: Manjaro/Arch Linux (w/ XFCE & KDE)\nGraphics card: nVidia 1060 or Intel graphics\n\nBroken: Blender 2.92.0, Commit date: 2021-01-03 01:04, Hash 1e8f2665916c\nWorked: 2.91\nWhile in sculpt mode, working on a subdivided mesh, pressing Shift+A to Expand Mask by Topology (or Curveature), Blender crashes and dumps the following\n\n```\n# Blender 2.92.0, Commit date: 2021-01-03 01:04, Hash 1e8f2665916c\nbpy.ops.sculpt.sculptmode_toggle() # Operator\nbpy.ops.object.voxel_remesh() # Operator\n\n# backtrace\n./blender(BLI_system_backtrace+0x20) [0x8642740]\n./blender() [0xeac81a]\n/usr/lib/libc.so.6(+0x3d6a0) [0x7f4f035bc6a0]\n./blender() [0x1a5647f]\n./blender(SCULPT_undo_push_node+0x7c7) [0x1a59297]\n./blender() [0x1a48a18]\n./blender() [0x1234114]\n./blender() [0x1237014]\n./blender() [0x12382ba]\n./blender() [0x12385cc]\n./blender(wm_event_do_handlers+0x60a) [0x1238faa]\n./blender(WM_main+0x20) [0x122f9d0]\n./blender(main+0x31e) [0xde69de]\n/usr/lib/libc.so.6(__libc_start_main+0xf2) [0x7f4f035a7152]\n./blender() [0xea9213]\n\n# Python backtrace\n```\n\n1. new scene\n2. select cube (or sphere) and go into sculpt mode\n3. shift + r and voxel remesh to default 0.1\n4. ctrl + r to apply new voxel mesh\n5. shift + a\n6. crashes on both my desktop (nVidia) and laptop (Intel)\n\n*note: If performing this on the default cube (or sphere) without voxel remesh, it did not crash.", "Crash. Mask after remesh causes Blender`s crash\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.92.0 Alpha\nLast worked i have: blender-2.92.0-4797c13e8f3a-windows64\nFirst broken i have: blender-2.92.0-8491e4ab8636-windows64\n\nTo reproduce: add object, enter sculpt mode, remesh with hotkey, touch with mask brush.\nIf you exit to object mode and back to sculpt, mask will work fine again.\n\n[2021-01-03_13-47-41.mp4](2021-01-03_13-47-41.mp4)\n\n[debug_logs.zip](debug_logs.zip)\n" ]

[ "Improve auto offset when inserting new node\nIn the example below the Noise Texture node should be moved automatically as well, not just the Set Position node.\n\n[screen_recording-2021-12-13_17.34.37.mp4](screen_recording-2021-12-13_17.34.37.mp4)\n\nThe exact behavior in more complex trees still has to be designed.", "Geometry nodes. Viewer node data late update (delete node and undo)\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.1.0 Alpha\n\n**Short Description**\nSpread Sheet editor is not updating when I undo the action of viewer node deletion\n\n**Steps to Reproduce**\n- Open .blend file\n- Delete viewer node\n- undo (You will not see any data in the spreadsheet until click on the viewer node or hovering mouse over spreadsheet)\n\n[#94641.blend](T94641.blend)\n[2022-01-04_20-09-05.mp4](2022-01-04_20-09-05.mp4)\n", "Move cache visualization to draw overlay\nCurrenlty the cache visualization isn't updated during playback, only when the screen is redrawn.\nThis needs an optimization of the drawing of the cache lines.", "2D transform operations show options and use a modal keymap intended for 3D transform\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.92.0 Alpha\n\ntransform.seq_slide and transform.transform are linked to the transform keymap of the 3D transform. This results in the wrong shortcuts hints #83307 in the status bar and also wrong options in the Edit Last Operation: \n{[F9508650](transform_last_operation.gif),size=full} \n\n- Use `Move/Extend from Current Frame` to move the handle.\n- Check the `Edit Last Operation` options\n\n", "Linked collection with smoke simulation cached doesn't work\nOperating system: Windows 10\nGraphics card: RTX 2070 laptop\n\nBroken: 2.82\nWorked: IDK\n\n Linked collection with smoke doesn't seems to load the cache properly or something.\n\n- Create a scene with a quick effect smoke,\n- Save the file,\n- Bake the simulation.\n- Save the file again,\n- File -> New -> General\n- File -> Link the collection of the another file\nThe domain doesn't load the cache", "Motion Paths not updating when working on the graph editor\nOperating system: Windows 10\nGraphics card: 1050 TI\n\nBroken: 3.5\nWorked: Supposedly never worked due to the way the code is set up.\n\n\nMoving the F-curve for key-framed object will not update motion paths.\n\n\n- Do some location key frames on the default cube.\n- Calculate the cube's motion path in the object properties tab.\n- Move the f-curve key frames on the default cube, motion path does not update.\n - Move the object directly, only the motion path of the current frame is updated.\n\n", "Bone's Transformation constraint numbers won't be symmetrized correctly.\nOperating system:Windows 10\nGraphics card:RTX 2070\n\nBroken: 3.3.1\n\n[Use this file](https://projects.blender.orgattachment)\n\nThis will break:\n\n- open the blend file\n- delete the selected Left bone\n- select Right bone and symmetrize it\n- check both Right and Left bones' Transformation constraint setting. Numbers on the Left one will be messed up.\n\nBut this works:\n\n- delete the selected Left bone\n- go to pose mode\n- select the Right bone\n- go to edit mode again and symmetrize it\n- Transformation numbers are both fine. \n\n------\n\nOriginal description:\n\nWhen Bones have Transformation constraint and are symmetrized in edit mode, its numbers are symmetrized correctly too.(Map from numbers like Location X min , Location X max and Map to numbers like Location min max etc) but if Transformation constraint's Target and Owner space are set as not Local those numbers are totally messed up and broken.\n\n", "Geometry Node: Viewer output for instance domain ignore if value nearest to 0.01\nOperating system: win 10\nGraphics card: 1050ti\n\nBroken: 3.4, 3.5\n\nGeometry Node: Viewer output for instance domain ignore if value nearest to 0.01\n\nIf you explicitly specify the instance domain, then with a value nearest to 0.01, the viewer will ignore the value input\n\n[2022-11-11 23-41-29.mp4](2022-11-11_23-41-29.mp4)\n\n\n- Put some intance to Viewer\n- Connect Float Value to Viewer set to instance domain\n- Test various floats (0.01 and 0.3)\n\nOr in attached file switch between couple of Viewer nodes\n[gn_viewer_value.blend](gn_viewer_value.blend)", "adding outflow object removes smoke from viewport\nOperating system: Windows 10\nGraphics card: Geforce RTX 2060\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\nversion: 2.90.1, branch: master, commit date: 2020-09-23 06:43, hash: 3e85bb34d0d7\nversion: 2.91.0 Alpha, branch: master, commit date: 2020-10-02 22:11, hash: 743eca0f316d, \n\nWorked: (newest version of Blender that worked as expected): \nunknown\n\nA simple smoke simulation with a plane inflow works fine. When a torus shaped outflow mesh is added the smoke is not calculated.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n1. Start Blender\n2. Add Fluid Domain to default cube\n3. Add 0.5m plane mesh\n4. Add Flow->Inflow to plane\n5. Cube-Fluid->cache->type->Modular\n6. Cube-Fluid->cache->Frame end->50\n7. Bake data\n8. Verify smoke is seen rising in view port\n9. Free data\n10. Add torus mesh object, major radius 0.8, minor radius 0.1\n11. Add Flow->Outflow to torus\n12. Free data\n13. Bake data\n14. Verify that smoke is not seen in view port (this is assumed a bug)\n\n[#81423.blend](T81423.blend)", "Multiple Smoke Flow Colors Rendering Black when Flow Behavior is set to Geometry\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 670/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.48\n\nBroken: version: 2.83.1\nWorked: 2.81 -- N/A, newer versions use fluid settings, but the old smoke sim worked as it should\n\nBaking a smoke domain multiple smoke flow objects with flow set to \"geometry\" and each set to a color results in all of the smoke rendering black.\n\nCreate a smoke domain on any object and place two smoke flow objects inside the domain. Set the smoke color of the smoke flow objects to any color, and the flow behavior to \"geometry\" vs \"inflow\" or \"outflow\". Change the material on the domain object to a material without a surface bsdf but with a volume bsdf whose color is set to the color output of an attribute node with the Name property set to \"color\". Bake the sim and render, and you should observe that all smoke is black, as opposed to the colors set on the flow objects. This works correctly when the flow behavior is set to inflow (except on the first frame), so I'm guessing this has something to do with the initialization of the smoke.\n\n[smoke_multi_color_bug.blend](smoke_multi_color_bug.blend)", "Incorrect Areas affected by Timeline Playback settings\nVersion: 2.80 rc2 (Linux 64)\n\n[Timeline_bug.blend](Timeline_bug.blend)\n\nTo reproduce:\n\n - Open the example file (enable \"Load UI\" too)\n - With the mouse cursor in the 3d view, play the animation using Shift + Space (or Alt + A, whatever your keymap is)\n - Ignore the 3d view. Notice which animation editors show a moving playhead.\n - Inspect the playback settings in the timeline editor.\n - Disable \"All 3d Viewports\" and repeat step 2.\n - Now Enable \"Active Editor Only\" and repeat step 2.\n\nWith certain settings, the Dope Sheet and Timeline show playback even when the Graph Editor and NLA do not. It's as if these windows think they are 3d viewports!", "Geometry Nodes: Modifier input properties don't cause update for screen statistic\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.24\n\nBroken: version: 3.4.1\n\nModifier input properties of geometry node don't cause update for screen statistic.\n\n1. Create object with geometry nodes.\n2. Add any Mesh Primitive node in modifier node tree.\nPut mesh primitive on group output. Updating the detail of a primitive affects the statistics of the scene.\n3. Put group input dummy socket in detail of a primitive node to create connection.\nChanging the details through the modifier settings is not reflected in the statistics.\n", "Cycles: modifying properties through the Python console does not trigger updates.\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.1.0 Beta\n\n\nLoad a scene with a 16k hdri\nwith viewport rendering happening, try and rotate the hdri environment using a mapping node:\n\n\nnow change the world surface setting sampling method to manual, and the map resolution to 1024\n\nmessage will appear:\n\n\nnow rotate the hdri again. It'll be much more responsive.\n\nnow in the python console enter:\nbpy.data.scenes['Scene'].world.cycles.sample_map_resolution = 15360\n\nNotice the message 'updating lights' didn't appear in the viewport.\n\n rotate again, it's still responsive because cycles is still using the 1024 manually entered.\n\nNow restart the viewport render by clicking on solid viewport mode and back to rendered viewport mode\n\nRotate again. Still responsive. Even though the viewport has been restarted, cycles is still using the last manually entered value of 1024.\n\n-------\n\nNOTE: no updates seem to be triggered for the properties. The only update triggered seem to be for the console's current line which is modified after executing the entered line.\n", "flow simulation element viewport size indicator strange after applying resize\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.83 (sub 17)\nWorked: not known\n\nElement size preview in viewport behaves unexpected.\n\n1) Open Blender\n2) Apply Quick Smoke to cube\n3) scale up Smoke Domain by 20 in x and y direction while keep z scale unchanged (S+Shift Z+20). Notice that preview of element size in corner of Smoke Domain gets resized and becomes more flat. Maybe that is bug because I would have expected that fluid elements are always cubic. But maybe that is intended?\n4) Apply All transforms to Smoke Domain\n5) I would now expect that viewport showed a resized preview of element size for Smoke Domain that was cubic. Instead size preview completely disappears from Smoke Domain. Instead a preview element is shown in the Cube. See . \n\nWhy does a flow object, that is not a domain object, need a element size preview? I think this is a bug.\nAlso, why doesn't the Smoke Domain get a updated cubic element preview when the Smoke Domain object changes size? I think this is also a bug.\n\n\n\n", "Hair particle distribution changes randomly\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.41\n\nBroken: version: 3.5.0\n\nHair particle distribution changes without the user changing parameters.\n\nOpen the attached Blender file.\n\nIn the particle settings of the active object, click on the Emission -> Number field to edit it, and then without actually changing it, press enter.\n\nNotice that even though no parameter was changed, the distribution pattern on the emitter changes. \n\nThis keeps happening for 5 repeats of the previous step, until it seems to reach an equilibrium. \nBut you can start it again, by clicking onto the Emission -> Seed field, and hitting enter without changing it. This will allow you to click on the Number field again and cause changes.\n\nFor this to function, we seem to need both Interpolated Children to be active, and a texture to drive the distribution.\n\nAs a side note here, I'm not sure if this should be a separate bug report, but the texture is set to influence Density, but also influences the size of the scattered objects. Whereas if you try to have it influence the Size field, nothing changes.\n\n" ]

[ "Mantaflow smoke preview will not update till domain settings are changed\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.90 (sub 4)\nWorked: (2.7)\n\nMantaflow somke preview will not update till setting on the domain are changed.\n\nbased on default cube\napply quick smoke and move the cube inside the domain.\n[2020-05-29 17-35-45.mp4](2020-05-29_17-35-45.mp4)\n\n\n", "Fluid Domain replay mode is completely broken in Blender 2.90\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.90.0\n\nFluid Domain Replay mode is completely broken and unusable in Blender 2.90, with majority of the changes affecting the outcome of simulation not resetting the domain state, therefore user can end up wasting tons of time making changes that seemingly do not do anything, just to realize that the domain result is not getting updated. The only way to update the result is to select the domain, and re-select the Replay mode as shown on the video below:\n[2020-09-11_21-11-43.mp4](2020-09-11_21-11-43.mp4)\n\nChange of pretty much any aspect of the scene which affects the result of the simulation but is not a property of the domain object object itself is not registered as change, therefore in vast majority of time, in Replay mode, user is viewing outdated, incorrect version of the simulation. This makes any attempts to use the smoke simulation for any serious work impossible.\n\n1. Open the attached .blend file: [FX_SmokeTrail_V01.zip](FX_SmokeTrail_V01.zip)\n2. Start the animation playback\n4. Pause the animation playback once it goes through and simulation caches\n5. Perform any changes to the inflow object, such as changing the inflow parameters of even moving the inflow object in viewport\n6. Start playblack again\n\nResult: The displayed simulation result in the domain is in previous invalid state\n\nExpected: It should **never ever**be possible to see an outdated version of simulation when the domain mode is set to Replay under any circumstances.\n\n" ]

[ "Blender development todo list -- Scripting\n| Notes: |\n| -- |\n| - This is a direct dump of Scripting (without cherrypicking the [still] valid items)\n| - Changed the broken python documentation links |\n\n\n**Unstable or poorly tested**\n- [ ] BMesh operators currently use int's for enums that have undocumented values. blender/blender-addons#34914\n\nFixes that still need doing.\n**Internal API Problems**\n- [ ] Class registration needs to be reworked. as well as some internal bugs fixed.\n- [ ] RNA api has a bug where 'raw_type' defaults to INT but is not set for MFace verts in makesrna.\n- [ ] ID data returned from functions is not included unless the type is an ID itself.Returning a mesh will work but not a meshes face for eg. See bpy_rna.c - /- [ ] XXX this is missing the ID part! []/\n- [ ] Dynamic enums currently refer to invalid memory when accessed from python, we need a way for python to hold the memory until its done with the dynamic enum.\n\n**Low Priority Py API**\n- [ ] Re-Registering classes leaks memory, would be good to resolve this. (see comments in bpy_rna.c)\n- [ ] 'Safe Mode' - where writes to invalid memory are detected.\n- [ ] Implement import hooks rather than replacing import/reload pep-0302/\n\n### Python/RNA API Design TODO\n- [ ] Add a way to cancel running scripts from the user interface, with no console this becomes very problematic (Esc --> KeyboardInterrupt for example).\n- [ ] Currently there is no way to write packed images from blend files, support for this should be added to exporters. eg: image.pack_write(\"somepath.png\")\n- [ ] Addressing operators/panels/menus, should we translate all names fromSOME_MT_menu --> some.menu, ONE_PT_panel --> one.panel ? would mean the UI code could reference them by this more pythonic syntax.\n- [ ] need to decide how 2.4x api's will fit in with new apis. PyNodes, PyConstraints.\n- [ ] convenience function to get selected verts/edges/faces to prevent expensive iteration over meshes. Something like: me.verts.selected\n- [ ] Design a method to allow custom UIs to filter properties/options. Useful for Blender GE too (example: which constraints are supported (#26804). Complex matter for which rna needs to be extended?\n- [ ] Zero bytes are not supported in RNA byte string types, see blender/blender-addons#32581\n- [ ] Better python API and undo stack integration, see #38386\n\n**Inconsistencies**\n- [ ] bpy.ops.mode_set(...) has different mode-names than bpy.context.mode. E.g. TEXTURE_PAINT vs PAINT_TEXTURE''CoDEmanX:'' Getting mode from bpy.context.mode for mode_set() is inappropriate, use '''bpy.context.object.mode''' instead!\n- [ ] bpy.ops.object.move_to_layer() takes a boolean array of 20 elements (like in Blender GUI), but most functions require 32 elements. bpy.ops.object.add() for example.\n- [ ] bone.matrix returns a matrix, but requires a list when assigning to it. This also goes for some of the other matrix attributes of bone, editbone and posebone. [appears to be resolved as of 2.67?]\n- [ ] the term 'size' is used incorrectly in some cases, where it should be 'scale' instead. Review and correct: search.html?q=size&check_keywords=yes&area=default\n- [ ] keyframe_insert(), options for insert or replacing existing keyframes, threshold for detecting doubles see: bpy.types.bpy_struct.html#bpy.types.bpy_struct.keyframe_insert\n- [ ] The ID data block path returned by RNA_path_full_ID_py (rna_access.c) for node trees only works for node groups, which are stored as true ID data blocks in bpy.data.node_groups. For material/compo/texture nodes which are part of Material/Scene/Texture respectively the function can not determine the correct path. Best solution would be to get rid of nested ID data blocks altogether blender/blender-addons#35720\n\n### Python RNA access\n- [ ] Remove mesh vertices / edges / polygons (without bmesh module)\n- [ ] Shorthand for active vertex / active edge (bmesh module)\n- [ ] Active vertex / edge index for meshes (like Mesh.polygons.active)\n- [ ] Remove Shape Key\n- [ ] Remove Curve Points\n- [ ] Particle edit mode\n- [ ] Start/stop the physics sim\n- [ ] Sculpt mask custom data\n- [ ] Add / remove rigid body + constraints + world\n- [ ] Add / remove transform orientation\n- [ ] Add / remove Text curve object Textboxes\n- [ ] Add / remove Particle Systems\n- [ ] Sequencer, ability to load and images blender/blender-addons#40663\n- [ ] Pixels and resolution are not available with multilayer EXR or Render Result images #53768\n\n### Python Console\n- [ ] Append to (and remove from?) SpaceConsole .history and .scrollback\n- [ ] input() is not supported #40592\n\n### General todo items\n- [ ] Operator redo (toolbar) could also change the last InfoWindow print of the python operator equivalent.\n- [ ] mathutil.geometry.intersect_point_quad_2d works inpredictable for non-concave or quads with singularity edges #30831\n- [ ] modules/rna_xml.py: BoolVectorProperty and CollectionProperty issue blender/blender-addons#32882\n- [ ] Drivers should be able to access '''self''' for bones and objects.\n- [ ] Add the ability to report() without an operator, internally use CTX_wm_reports(C), See: operator-report-outside-operators\n", "Indices order is inconsistent with duplicate operation\nOperating system: win10\nGraphics card: 1050ti\n\nBroken: 3.2.0\n\nIndices on copied geometry (in edit mode) are sometimes reversed (verts) or misplaced (edges). Not sure when it happens exactly, but it can happen, if you duplicate and delete new geo couple times.\n\n\n\n\n - Enable *indices* overlay to see results\n - Create clockwise mesh loop (for example draw it in *Edit mode* from single vert via *ctrl + mouse click*)\n - Select mesh and press *shift + D* and *Enter*\n - Check verts indices order, it should match initial clockwise direction\n # Delete copied geometry and repeat steps 3 & 4 -> indices should be reversed. (Interesting thing that if you change *Mode* option in operation's menu from 1 to 0 it preserves original order)\n\nOr in attached file perform steps from 3 to 5\n\n[indices_order.blend](indices_order.blend)\n\n", "Multi-object mode support (parent task)\nThis is a parent task to manage to todo's for multi object editing.\n\nThere are many small todo's.\n\n### Reference Patches:\n\nFor anyone interested in helping out, here are some example commits which can be used as a reference.\n\n- Edit Mesh Face Poke - 3a9669bfff *Committed without indentation, so you can see lines changed more easily.*\n- Edit Mesh Flip Normals - 4704f2687a\n- Pose Mode: Flip Quaternions - 7560aabf71\n\n\nOther commits may be useful too.\n\n### Object Loops\n\nThere is some duplicate code for looping over objects.\n\nWe'll keep the option of having a way of simplifying the looping open.\n\n----\n\nNote: to collect operator ID's.\n\n- this script helps copy them from menus: [P656](P656.txt)\n- this is a more efficient version that watches the x11 clipboard so you can hover over items and press Ctrl-C: [P657](P657.txt)", "Asset Browser - Append (Reuse Data) without instancing collections doesn't add more objects\nOperating system: Windows 10 Home 64bit (10.0, Build 19044)\nGraphics card: Nvidia RTX 3080\n\n\nBroken: 3.2.0\n\nWhen dragging a collection item from the Asset Browser with \"Append (Reuse Data)\" and \"Instance Off\" it doesn't add the collection again, it moves one object and leaves the rest where they were.\n[Asset Browser Reuse data issue.mp4](Asset_Browser_Reuse_data_issue.mp4)\n\n1. Create a collection asset with at least two items inside\n2. Save that collection to your library (save file inside the correct folder)\n3. Open a new scene\n4. Drag your new collection asset into your scene with Append (Reuse Data) active\n5. In the tool window (not sure about its name) lower left turn off \"Instance\"\n6. Drag out another of the same collection asset\n7. It should now have moved one of the two items", "Support for tools with interactive geometry preview\nSome of the current or planned tools (add object tool, insert mesh brushes...) and other functionality requested project members require adding new geometry to the scene. \n\nAs far as I know, these tools can't render new geometry without creating a new object or adding that geometry to an existing mesh. This makes the gizmo previews of the add object tool not ideal and some of the requested functionality for sculpt mode is not possible to implement. It would require to modify the mesh and rebuild the PBVH in real time each time geometry changes. \n\nThe only tool so far that requires this functionality in sculpt mode is a limited version of the insert mesh brush in [D9380](D9380). That tool joins the insert mesh to the active mesh using BMesh when the stroke starts and then rebuilds the PBVH so the new geometry can be rendered. This makes the tool lag each time a stroke starts in a high poly mesh, and it limits the functionality to only an insert operation per stroke. (The ideal behavior should be that the insert mesh brush should insert a new mesh per stroke step, so you can insert meshes with spacing over the mesh). \n\nThe main goal of this design of to have a fast way to create previews for tools that modify geometry which look like the mesh is actually being modified. This is just a implementation detail, so the user should not know if the mesh geometry is actually being modified or if it is just a preview, as both options should ideally look the same. \n\n## Proposed solution\n\nThe ideal solution would be to have a generic function(s) that can render arbitrary geometry using the current scene shaders and render mode (as it was a new mesh object). This function should be possible to use from any operator and mode. \n\nIn the case of the insert mesh brush, the tool can manage a list of triangles and indices created by combining the geometry of the insert mesh and use this function to render them in real time without modifying the PBVH or the active object geometry. Then after the stroke finishes, all the new geometry can be joined to the active mesh in a single operation.\n\n## Workarounds\nIf implemented that functionality is not possible, some workaround for sculpt mode could be:\n- When PBVH rendering is used, add a \"fake\" node to the PBVH and add the new geometry there. If this fake node exists, render it as a normal node. When the operation finishes, delete this node, join the meshes and rebuild the tree. This approach won't work when PBVH rendering is disabled, so these tools will still lag.\n- Exit sculpt mode when the operation starts and create a new temporally object to do the geometry preview there. After the operation finishes, join the geometry, delete the preview object and enter sculpt mode in the initial object. Not sure what potential problems this approach could have (like undo). \n", "particle system locations not as expected\nOperating system: macOS-11.4-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5500M OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.5.14\n\nBroken: version: 3.0.0 Alpha\n\nif i parent an animated cube vertex to a vertex with a particle system and rotate that cube, it works fine. \nIf i copy that vertex with the particle system to another position - the locations get a bit weird.\n\n- tab -> edit mode\n- select vertex\n- Shift-D -> double vertex\n- Mesh -> Separate -> selection\n- rename that to vertex (not important, just for me)\n- add a default particle system to vertex\n- select vertex, ctrl select cube -> edit mode -> ctrl-p\n- select vertex -> origin to geometry -> run\n- works fine\n- select cube, insert keyframe scale/rot/loc\n- go to frame 100\n- location.z: 8 rotation.z: 360*8 -> keyframe both\n- run, works fine, all locations as expected\n\n- select the vertex\n\n- shift-d X -2\n\nrun -> weird location \"disorders\"\n\n\n[particle_bug.blend](particle_bug.blend)\n\n---\nor\n\n- open .blend file\n- select Vertex from outliner\n- {key Shift D} and place the duplicate Vertex anywhere\n- Play animation\n\n[#90494.blend](T90494.blend)\n\n---\nVideo:\n[particle_bug.mov](particle_bug.mov)", "Add \"Link (Library Override)\" option when dragging assets into the scene\nAs a user, I want to be able to drag an object asset into a scene and have it have the following properties:\n\n1. The data block is not copied into the new scene, it should be linked.\n2. The underlying *data block* cannot be edited (i.e., you cannot go into edit mode on a mesh object)\n3. The new object should be able to be modified (i.e., translating, scaling, rotating etc.)\n\nThis is not possible with [the options provided](asset_browser.html#using-assets) at the moment.\n\nI propose that a new option be added, called **Link (Library Override)**.\n\nThis would essentially be the same as using \"Link\", followed by calling the \"Make Library Override\" operator on the linked object.", "'Show Instancer' option for Viewport hides the mesh in edit mode\nOperating system: Linux-5.7.8-200.fc32.x86_64-x86_64-with-fedora-32-Thirty_Two 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.100\n\nBroken: version: 2.90.0 Alpha\nWorked: Never\n\nWhen using either Face or Vertex duplication on meshes, turning off `Show Instancer` in the Instancing settings (Object Properties) hides the parent object.\nThe problem is when you enter edit mode. Both the instancer and the instances are invisible.\nOne solution could be to unhide the instancer.\n\n - Open attached file\n - Uncheck 'Show Instancer' for Viewport\n # In the 3D View, press tab or otherwise enter editmode\n\nYou now are editing an invisible mesh - since we can select and edit the mesh, it would be much, much friendlier to allow editing, as the user probably expects this, by showing the mesh just in edit mode - it could be wireframe only or semitransparent, to indicate that it's hidden in object mode.\n\n(This may be considered an enhancement or a bug - at the very least it feels like a lack of polish in the interface, but it would be extremely befuddling to some users should they find themselves in this state.)\n[instancer_bug.blend](instancer_bug.blend)\n\n", "Duplicating a node with a driver loses the driver\nWindows 10 Pro, GTX 1060 6 GB\n\nBroken: 2.79b\nWorked: (optional)\n\nDuplicating a node with a driver (e.g. an RGB mix with a driver for the factor) causes the driver to be lost from the duplicate. This makes node setup more tedious than it needs to be.\n\nAdd an empty to the scene.\nSet up an RGB mix node with two color inputs.\nCreate a driver on the factor that uses the empty Y world position.\nWith the node selected, shift-D to duplicate.\n\nThe duplicate will have lost the driver.\n", "Copy Datablock button has different behaviour in Graph as Modifier\nOperating system: Windows-10-10.0.18363-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.77\n\nBroken: version: 3.3.0\n\nThe button to duplicate the current node tree works as such in the node graph but on the modifier, it creates a new node tree.\nIn other areas of Blender (eg materials) the same button duplicates current datablock.\n\n\n\nAdd a new geometry node tree\nAdd a node\nClick the Copy Node group button in graph and see graph is duplicated\nClick the Copy Node group button in modifier and see graph is replaced\n[2022-09-15 11-14-53.mp4](2022-09-15_11-14-53.mp4)\n", "Inconsistency of some tools/operators (e.g. Subdivide) in regards to hidden geometry\nBroken: all or none\n\nI'm not really sure this is even an issue, though it feels like it. As far as I understand it, mesh editing tools are not supposed to affect hidden geometry. But the Subdivide tool does. Consider this mesh:\n\n\nIf you hide the middle face...\n\n\n...and then Subdivide the two remaining faces...\n\n\n...Subdivide will create new edges cutting through the hidden face, also creating inconsistent selection (i.e. selected edges in face select mode). Shouldn't the operator check that that face is hidden and not create edges through it?", "full scene copy does not duplicate obdata of objects in Edit mode\nOperating system:\nGraphics card:\n\nBroken: 3.3 LTS and 3.3.1 LTS\n\n\nif you make a full screen copy of a 2D blend file with some objects in edit mode, the objects in edit mode are linked to the original scene i.e.not independent.\n\nCreate a new 2D project and add a blank grease pencil object\nSelect draw mode and add a stroke say a square\nSelect edit mode and select the stroke for point editing\nCreate a full copy of the scene still in edit mode\nMove a point on the selected stroke in the new scene and observe that the same point moves on the original scene. This means the two scenes are linked.\nCarry out the same steps but before creating the new full copy change the mode to object mode.\nagain move one of the points of the stroke in the second scene and note that the object in the original scene does not change i.e. a full copy.\n\nSame is true at least for meshes and curves, so presumably all obdata.\n\nBased on the default startup or an attached .blend file (as simple as possible).", "Joining objects can create more customdata layers (e.g. UVMaps) than allowed, can eat huge amounts of RAM, bloats the filesize\nOperating system: Linux-6.1.18-200.fc37.x86_64-x86_64-with-glibc2.36 64 Bits, WAYLAND UI\nGraphics card: AMD Radeon Graphics (renoir, LLVM 15.0.7, DRM 3.49, 6.1.18-200.fc37.x86_64) AMD 4.6 (Core Profile) Mesa 22.3.7\n\nBroken: version: 4.0.0 Alpha\nWorked: never\n\nJoining objects can create more customdata layers (e.g. UVMaps) than allowed.\nWhile there is no hard limit for UV layers (yet), we usually restrict this to 8 from the UI.\nWhy can this be a problem?\n\nSee reports #111564, #71823, #96926\n\nJoining many (hundreds or thousands -- which is not uncommon for CAD data) small objects with uniquely named UV layers will result in many, many UV layers (not only UVs, other customdata layers / attributes may be affected as well, but encountered most frequently for those), each then taking up memory/space for the whole -- now possibly very large -- mesh.\n\nThis will - unexpectedly - bloat up the filesize tremendously, also take up huge amounts of RAM for the joining operation, leading to possible crashes.\nThis is also not actually supported for UVMaps (the UI limit/restriction of 8 is there for a reason since e.g. Eevee, probably Cycles, too, dont handle more then 8).\n\n- open `uv_maps_on_merge_single_user_mesh_unique_uv_layer_names.blend` (1.6MB)\n- join all objects, save (file is now 3.8MB, the joined object has 121 UVmaps)\n- the problem gets much worse if the resulting object is larger and/or if object count increases\n\n", "Appending grease pencil brushes creates \"fake user\" unintentionally\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: GeForce GTX 960M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.55\n\nBroken: version: 3.0.0 Beta\nWorked: 2.93.4\nCaused by 3be5ce4aad\n\nWhen appending grease pencil brushes, they will be added to the file as a \"fake user\", even if this option is unchecked in the append window.\n\n- Open default 2D animation file\n- {nav File > append > fakeuserbug.blend > Brush > Shift select the custom brushes (start with \"pp_\")}\n# Change brush from `Active tool & Workspace settings` (notice it has been added as a fake user)\nappend brushes from another file\nThey will be added as a fake user, whether or not the \"fake user\" box is checked in the append window\n\n[fakeuserbug.blend](fakeuserbug.blend)", "Visualized face normals are duplicated when Edit Cage is enabled in subdivision modifier (GPU subdivision turned OFF)\n### Simplified Report\n\nHaving Edit Cage on subdivision modifier enabled causes visualized face normals to be duplicated started from the origin. Here is a simplified file with a single face:\n\n\n\n[editCageWrongNormalDisplay.blend](editCageWrongNormalDisplay.blend)\n\nThis issue exists back to 2.83.\n\n### Original Report\n\nOperating system: Ubuntu 22.04 LTS\nGraphics card: Nvidia Quadro M4000\n\nBroken: 2.93.10, 2.83\nWorked: Unknown.\n\n\"Flying\" normals everywhere\n\nI have not discovered a correlation. The alignment of the normal is random, it changes when you exit edit mode and return.\n\n\n\n\n\n\n\n\n\n" ]

[ "Datablocks can't be duplicated everywhere, because trick with fake user no longer works\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: Intel(R) HD Graphics 4600 Intel 4.3.0 - Build 20.19.15.4454\n\n\nBroken: version: 2.80 (sub 72)\nWorked: (optional) 2.79b\n\n\nThe MCE or tracker, should allow you to duplicate the clip asset so that you can attach other tracker bundles. In 2.8 it does not allow you to create more instances.\n\nHowever I was able to make a duplicate of a Movie Clip which is reflected in its datablock. I opened a legacy tracking project from 2.79b. It had some duplicated clips already. Once this function (multiple instances) is activated I am able to duplicate them more by using the counter buton next to the Clip name in the MCE.\n\n[Please describe the exact steps needed to reproduce the issue] Make a project in 2.79 with multiple instances of a Movie Clip then open it in 2.8. The number indicator appears. Otherwise opening a movie clip in 2.8 it does not.\n[Based on the default startup or an attached .blend file (as simple as possible)] Find the .blend of a project with extra Movie Clip instances attached{[F7091277](broken_MCE_clip_users.blend)}\n\n" ]

[ "The tools panel suddenly disappears in engraving mode.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Quadro K620/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 425.31\n\n\n\nBroken: version: 2.93.0\n\n\n\n\nIn the engraving mode, use the blank key to call the tools panel may suddenly disappear.\n\nMy native language text↓↓↓↓\n\n在雕刻模式，使用空白鍵呼叫　tools 面板會有突然消失的可能。\n\n\n\n\n\n1. Set KeyMap>Spacebar Action>Tools\n2. Select any object and enter sculpt mode.\n3. Press the blank key to call the tools panel. At this time, the panel will suddenly disappear. (Error generation)\n4. Move the mouse arbitrarily\n5. Press the space button again to call the tools panel. At this time, the panel will suddenly disappear. (Error occurs again)\n\nOther additional tests:\n\n1. The panel disappears (when an error occurs), do not move the mouse, press the blank key again, and it will appear normally\n\n2. When it appears normally, as long as you move the mouse and press the blank key, an error will occur, but there is one exception, which is simply moving the mouse horizontally so that the panel appears in the second position, the selected brush and slide The overlapping position of the mouse is exactly the same as the last time, the error will not occur (image)\n\n3. In the object mode, everything is normal, I think it should be operated like the object mode.\n\nMy native language text↓↓↓↓\n\n1.設定 KeyMap >Spacebar Action>Tools\n2.選擇任意物體，進入sculpt模式。\n3.按空白鍵呼叫tools面板。這時候面板會突然消失。(錯誤產生)\n4.任意移動滑鼠\n5.再次按空白鍵呼叫tools面板。這時候面板會突然消失。(再次錯誤產生)\n\n其他額外測試 :\n\n1.面板消失(錯誤產生時)，滑鼠不要動，再按一次空白鍵，就可以正常出現\n\n2.當正常出現後，只要有移動滑鼠，再按空白鍵就會發生錯誤，但有一個時候例外，就是單純水平移動滑鼠，讓第二個位置出現面板時，選取的筆刷與滑鼠重疊的位置和上一次一模一樣，錯誤就不會發生 (圖片)\n\n3.在物體模式，一切正常，我認為應該操作起來要像物體模式這樣。\n\n\n\n", "Editing redo panel properties of torus in edit mode reverts some steps\nOperating system: macOS-10.14.6-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-12.0.24 355.11.10.50.10.103\n\nBroken: version: 3.4.1\nWorked: (never worked as expected)\n\nAdding and then adjusting a Torus or any Add Mesh add-on in edit mode reverts edits made to nodes, video sequencer, and more, since entering edit mode on the mesh object.\n\n1. Go to the Geometry Nodes workspace\n2. Select the default cube\n3. Click the New button to add the Geometry Nodes modifier and create and assign a new node group\n4. In the 3D Viewport switch to Edit mode\n5. In the Node Editor, pour your heart and soul into crafting an elaborate node tree (lol, jk 👀💔)\n6. In the 3D Viewport add a Torus\n7. Tweak any setting in the Adjust Last Operation panel\n8. Observe Geometry Nodes edits reverted (all edits since entering edit mode in the 3D Viewport)\n\nNo other built-in mesh primitives (nor even primatives!) have this effect, but the Add Mesh add-ons all do.\n\nI have checked some of the editors to see what data is reverted and what is preserved. This is not yet an exhaustive review.\nThe following data is reverted:\n- Video Sequencer\n- Properties\n- every Node editor\n- Timeline\n\nThe following data is preserved:\n- Preferences\n- Text Editor\n- Python Console\n- Movie Clip Editor\n- UV Editor\n\nThe Outliner has mixed results:\n- Hierarchy expansion reverts in Blender File, Data API, and Orphan Data but not in Scenes nor View Layer.\n\nSome editors have special behavior:\n- Image editor\n - Paint mode\n - {Add torus, paint image, tweak torus} reverts the paint\n - {Paint, add, tweak} kicks out of Edit mode into Object mode\n - Switching Image Editor modes does not close the ALO panel. The following sequence of steps produces a Python error:\n - Add torus\n - Switch Image Editor modes\n - Adjust torus\n (If this merits its own tracker issue, tell me and I will write one up. This all seems so tied to one root culprit that it is not my initial inclination to do so)\n\n", "3D Paint Mode stencil disappears when panning viewport.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.76\n\nBroken: version: 2.93.4\n\nWhen panning viewport in 3D paint mode with stencil mask enabled, the stencil mask disappears when moving the viewport, however does not disappear when zooming it, which indicates it's not a technical limitation.\n\n1. Enter 3D paint mode\n2. Enable texture mask, select some texture and set mapping to stencil\n3. Navigate the viewport\nResult: The stencil disappears when moving the viewport, which makes it very difficult to line up stencil projection with the 3D geometry\nExpected: The stencil does not disappear when navigating the view\n\n", "Multiple pop-ups overlayed with same functioning on Right Clicking `Online Manual` button\nOperating system: Windows 10 Pro\nGraphics card: GeForce GTX 1650\n\nBroken: version: 3.5.0 Alpha, branch: master, commit date: 2022-12-31, hash: 614704f90c, type: release\n\nRight Click on any operation/property on interface until `Online Manual` button is available and keep right clicking on `Online Manual` to get prompted again to right on `Online Manual` and keep going to create an infinite error pop-ups type screen. \n\n- Keep right clicking on `Online Manual` option in context menu\n", "Grease Pencil Proportional Edit \"Connected Only\" fails to delimit unconnected strokes if they are selected\nVersion 2.80 rc3 (Linux 54)\n\n[GP_proportional_edit_bug.blend](GP_proportional_edit_bug.blend)\n\nHow to reproduce:\n\n - Open attached file, remain in edit mode.\n - Note the proportional edit settings and selected stroke vertices.\n # Move around the whole blue stroke, along with one vertex at the tip of the black stroke.\n\nObserve how the black stroke is affected by the selected vertices of the blue stroke, even though the two strokes are not connected and are on different layers. This is not the expected behavior of the \"Connected Only\" setting. The strokes should only be affected by their own vertices, but as long as any vertex of either stroke is selected, they will affect each other as though the \"Connected Only\" setting is ignored!", "Emulate 3 Button Mouse conflicting with Add Primitive tool\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.91.0 Alpha\n\nAs the title suggests, emulate 3 button mouse which works with Alt key conflicts with the alt key of the Add primitive tool where you get to extrude on both sides. I can't commit the extrude while holding alt.\n\n[2020-08-30 21-56-30.mp4](2020-08-30_21-56-30.mp4)\n1. Turn on Emulate 3 button mouse\n2. Select Add Primitive tool, drag out a shape.\n3. Hold Alt to turn on double side extrude and try to click to commit and it won't.\n\n", "Annotations (D) do not work with 3d cursor tool (keymap issue)\nMacBook Pro (Retina, 15-inch, Mid 2015)\n2.5 GHz Intel Core i7\n16 GB 1600 MHz DDR3\nAMD Radeon R9 M370X 2048 MB\nIntel Iris Pro 1536 MB\n\nBlender 2.80 Alpha 2\nHash: c94f806dc73\n\ndate: 2018-11-14 22:51\n\nAnnotation tool does not work when pressing D and LEFT MOUSE BUTTON.\n\nAnnotation tool will work when searching with CMD+F; but also within the search, if you click ENTER the annotation tool will pop up for a split second and disappear. \n\nAttaching a short video.\n[Annotation Bug.mov](Annotation_Bug.mov)\n", "Outliner selection not working correctly after undoing object deletion\nOperating system: Linux-5.17.15-1-MANJARO-x86_64-with-glibc2.35 64 Bits\nGraphics card: AMD VERDE (LLVM 13.0.1, DRM 2.50, 5.17.15-1-MANJARO) AMD 4.5 (Core Profile) Mesa 22.1.2\n\nBroken: 2.91.2, 3.2.0, version: 3.3.0 Alpha\nWorked: prior to b077de086e\n\nAn object that is linked in another collection can usually be removed from just that collection by using unlink, while using delete removes it from both collections. But if you delete the object and undo, unlink does delete the object in both collections. \nRenaming the cube or creating a new object \"breaks\" the cubes apart so they can be selected/unlinked individually again.\n\nIn the default scene,\n- create a new collection\n- in the outliner, ctrl-drag the default cube object to that collection to link it\n- delete the cube\n- undo\n- in the outliner, it is now impossible to select the cube in the new collection separately and using unlink will delete both cubes", "Non animatable properties\nBroken: version: 3.3.0 Alpha\n\nWorking on refreshing the in-built AnimAll addon by adding the latest properties supported in Blender.\n@ideasman42 asked me to file a task\n\nCertain properties are missing animation capabilities:\n\n - Lattice Vertex Groups\n - Mesh Vertex Creases", "Multiresolution Issues\nThis is a temporary parent task to gather long standing and highly problematic issues with the Multiresolution modifier.\nThis mostly concerns bugs, known issues and badly supported existing features and gives an overview of the most pressing issues.\nNew features are not part of this task and don't need to be discussed.\n\n# Goal\nThe intent behind the task is to incrementally fix long standing issues to fix the current experience of subdivision based sculpting.\n\n# List of Issues\n\n## Crashes\n\n- [ ] #99887 - Multires Unsubdivide freezes Blender with hidden geoemtry\n- [ ] #84551 - Multiresolution - Crash when Linear Subdivision level is above 1\n- [ ] #100985 - Crash on sculpt mask brush when Multiresolution modifier is copied from another object\n- [ ] #85065 - Sculpt Mode - Using Mask while Multiress active = Crash\n- [ ] #86114 - Multires Displacement Smear brush + Dyntopo Error\n\n## Long Standing Issues\n\n### Visibility\nOne of the most important badly supported features is face visibility. \n\n- [ ] #87821 - Sculpt mode: annotation and unhide conflicts with Multiresolution modifier\n- [ ] #92372 - Sculpt mode: hide masked do not work when modifier enabled, but hidden part works with modifier\n- [ ] #95419 - sculpt mode: invert visible hides all faces with Multires modifier\n\n### Propagation\nAlso a high prioprity. Spikes and artefacts of details is making switching levels a bad experience.\n\n- [ ] #76374 - Multires: Artefacts from lower subdiv level changes\n- [ ] #94769 - Multires spikes\n\n### Undo\nBroken sculpted levels and other undo issues are very severe, since they are easy to trigger and cannot be easily recovered from. This also makes switching levels worse.\n\n- [ ] #93511 - Unable to undo subdivision with Multiresolution modifier\n- [ ] #79487 - Multires Level switching undo bug\n- [ ] #76311 - Redo causes pointy mesh artifacts in Multires\n- [ ] #76313 - Redo in Multires doesn't bring back stroke made after a subdivide\n\n### Rebuilding Subdivs\nA vital feature that is not reliable and surprisingly destructive.\n\n- [ ] #99857 - Rebuilding subdivisions with multiresolution deletes vertex groups\n- [ ] #81645 - Rebuilding multires subdivisions sometimes puts some vertices at the origin\n- [ ] #85792 - unsubdivide of Multires bug\n- [ ] #95022 - Using \"Rebuild Subdivisions\" in Multires modifier on hires mesh returns an error: Not valid subdivisions found to rebuild lower levels\n\n### Face corner inaccuracies\nVisibility and masking can lead is having different results in multires.\n\n- [ ] #79319 - Transform Tools influence on Hidden Face Sets with Multires Modifier\n- [ ] #79317 - Hidden Face Set Boundaries unmasked after inverting the mask on Multires Modifier\n\n### Grids consistency\nSome common workflows operators lead to severe issues with the multires grids.\n\n- [ ] #76682 - Applying negative scale inverts multires displacement\n- [ ] #76735 - Sculpt: Artifact in Multires when mirror modifier is applied\n\n### Smooth Normals\nSmooth faces is overall a worse expereince and is often avoided as a result.\n\n- [ ] #80641 - Sharp normals along base mesh wires in multires sculpting with smooth shading\n\n### Performance\nSome performance issues were investigated but never fixed.\n\n- [ ] #79488 - Multires clay brushes performance issue\n\n### Other\nVarious issues, arranged by importance.\n\n- [ ] #63581 - Textures not displayed in sculpt mode due to missing UVs (PBVH rendering, Multires)\n- [ ] #94631 - Multires: Duplicating objects sets the viewport level to sculpt level\n- [ ] #107102 - No visible mesh on Viewport Screenshot in sculpt mode with Multires\n- [ ] #73355 - Multires: Missing updates are not communicated in the UI\n- [ ] #90089 - Multires the Apply Base add to Quick Favorites can't change mesh\n- [ ] #76769 - MultiRes cannot \"Apply Base\" when ShapeKey have been added\n- [ ] #81497 - Multires: Editing base mesh creates random results on higher subdiv levels\n- [ ] Reshape is unreliable (Needs report)\n\n## Unsupported\n\n- [ ] Geodesic & Diagonals Expand (Needs report)\n- [ ] Fairing operators (Needs report)\n- [ ] Mask Extract (Needs report)\n- [ ] #95665 - Grab Silhouette broken with Multires \n- [ ] #88999 - Sculpt mode: Pose brush with Face sets FK mode do not work with Multiresolution\n\n## UI\n\n- Graying out/disabling intentionally unsupported features (Like with Remesh)\n - Dyntopo\n - Trimming operators/tools\n - Mask Slice\n\n## Needs Investigation\n\n- [ ] #95557 - Multiresolution sculpting broken after change in base geometry\n\n", "Sculpt: crashes when using the Edit Face Set in Delete Geometry mode on an object with 0 faces\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Xe Graphics Intel 4.5.0 - Build 30.0.100.9864\n\nBroken: version: 3.3.0 Alpha\n\nBlender crashes when trying to delete non existing geometry with the Edit Face Set tool\n\n[blender.crash.txt](blender.crash.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n\n1) select the default cube, go to edit mode and delete all vertices\n2) switch to the sculpting tab\n3) select the \"edit face set\" tool and switch the mode to \"Delete Geometry\"\n4) left click anywhere in the viewport and Blender will crash\n\nThere are many things going wrong in sculptmode with faceless meshes.\n\nSee also reports:\n- #86940 (Crash with certain file. Sculpt mode. Vertex color fill + edge.)\n\nAnd (possible but partially abandoned) solutions:\n- [D10851: Fix #86940: sculpt automasking facesets option crashes on meshes without faces](D10851)\n- [D10966: Prevent changing to sculptmode on faceless mesh objects](D10966)\n\n**Stack trace**\n\nblender.exe :0x00007FF603BDA400 sculpt_face_set_edit_invoke\nblender.exe :0x00007FF6033688F0 wm_operator_invoke\nblender.exe :0x00007FF603366280 wm_handler_operator_call\nblender.exe :0x00007FF603367CA0 wm_handlers_do_keymap_with_keymap_handler\nblender.exe :0x00007FF603367440 wm_handlers_do_intern\nblender.exe :0x00007FF6033668A0 wm_handlers_do\nblender.exe :0x00007FF603363430 wm_event_do_handlers\nblender.exe :0x00007FF60334D530 WM_main\nblender.exe :0x00007FF602431210 main\nblender.exe :0x00007FF608464500 __scrt_common_main_seh\nKERNEL32.DLL :0x00007FFF3D887020 BaseThreadInitThunk\nntdll.dll :0x00007FFF3EAE2630 RtlUserThreadStart", "Trembling when selected cube in edit mode\nOperating system: MacOS Sierra - Versione 10.12.6\nGraphics card: NVIDIA GeForce GTX 680MX 2048 MB\n\nBroken: 2.83.0 (2020-06-03), 2.90 (2020-06-12)\nWorked: 2.82a\n\nA strange flicker happens when I select the faces of the cube.\n\nI select the cube.\nI change the visualization in Edit Mode.\nI set select mode to Face.\nI select the faces of the cube but they do not remain selected instead, a strange flicker of the object occurs. Here a video example:\n\n[blender_283_bug.mp4](blender_283_bug.mp4)", "Fix preview flickering when using \"Adjust Last Operation\"\nTweaking properties in the \"Adjust Last Operation\" panel removes all previews for a moment, causing quite drastic and annoying flickering.", "Material slots undo is inconsistent in non-Object modes\nBroken: 3.0\nMaterial slots undo is inconsistent. Sometimes multiple slots are added/removed in one undo, and sometimes undo won't change back mode.\n\n\ntab to edit mode \nadd a material slot\nundo twice\nadd another material slot\nundo\nboth material slots will be undone. Also undo won't go back into edit mode.\nif you redo both material slots are redone at the same time.\n", "Quadriflow Remesh: Undoing the next operator also resets the mesh\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 960M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.83.4\n\nUndoing the operation after the remesh also undoes the remesh.\n\n- Quadriflow remesh the default cube\n- Execute any other operator (switch to edit mode, add curve, etc..)\n- Undo\n\nThe mesh will reset to its state before the remesh.\n\n" ]

[ "Annotation tool: wrong undo order and different in different modes\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: 2.93.2\nBroken: 3.0.0 Alpha, `c1f5ac7cfe`, 2021-07-25 18:36\nBroken: 2.83.17\nWorked: never\n\n\n| Object mode | works correct, as expected. Each undo action undoes 1 stroke |\n| -- | -- |\n| Weight paint | works properly |\n| Vertex paint | works properly most of the time, sometimes as Texture paint |\n| Texture paint | do not undoes each stroke but after number of {key Ctrl+Z} (n=strokes) undoes all strokes simultaneously or disables em at n=strokes+1 and undo previous action (mode switch in my case) |\n| Edit mode | same as Texture paint |\n| Sculpt mode | Undoes strokes one by one, but on the last stroke it do not react to {key Ctrl+Z} once, and then on the secon disables stroke and previous action (mode change) |\n\n**Steps to Reproduce**\n- Open default scene\n- Switch to Edit or texture paint mode\n- Draw 4 strokes with annotation tool\n- Undo 4 times (nothing happens to any stroke)\n- Undo again: Clears all strokes\n[2021-08-15_15-46-27.mp4](2021-08-15_15-46-27.mp4)\n\n", "Wrong undo order with annotation tool in edit mode.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 473.04\n\nBroken: version: 3.1.2\nReminds me #90291 (Undo, Sculpt mode options reverts back with wrong order and simultaneously)\n\nIn edit mode undo with annotation strokes do not work until full geometry undo.\n\nOpen default scene, select cube, enter edit mode\nMake 1 annotation stroke\nMove 1 vertex\nRepeat previous two actions few times.\nPress {key Ctrl Z} few times\n\n[2022-04-18_10-30-49.mp4](2022-04-18_10-30-49.mp4)\n", "Annotations cannot be undone if in edit mode-ish\nOperating system:\nGraphics card:\n\nBroken: 2.83.5-linux64\nWorked: unknown\n\npressing ctrl-z with annotate tool while in edit mode does not undo last annotation stroke as expected.\n\nNOTE: the \"-ish\" is because it is not exactly clear if to me if there is an exact relation between Edit Mode and the last annotation stroke not being undone.\n\n0. (start in Object Mode)\n1. press tab to enable Edit Mode\n2. draw something with the annotation tool\n3. press ctrl-z (nothing happens)\n4. press tab to go back to Object Mode\n5. press ctrl-z again (now the last annotation stroke gets undone)" ]

[ "Image Editor in Mask mode: Toolbar broken \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 462.59\n\nBroken: version: 3.0.0 Alpha\n\nImage Editor in Mask mode: Toolbar broken \n\n- Open the Image Editor\n- Change the UI mode to Mask.\n- Add a new mask.\n- Toggle Toolbar(T).\nNotice that the Adjust Last Operation pop up i moving but no Toolbar becomes visible.\n{[F11089235](mask.gif),size=full}\n\n\n", "Volumetrics: Add Light Grid support\nVolumetrics are currently not affected by lightgrids. FIX THAT.", "Eevee: Rendering particle hair results in plain white pixels \nOperating system: GENTOO Linux, updated daily\nGraphics card: NVIDIA Corporation TU106 High Definition Audio Controller (rev a1)\nGraphic driver: x11-drivers/nvidia-drivers-460.67\n\n**Blender Version*\"\nBroken: after blender-2.93.0-8f2c3c63e970-linux64\nWorked: before and including blender-2.93.0-8f2c3c63e970-linux64\n\nPlain white pixels when rendering particle hair (Eevee viewport/render)\n\nLoad attached blend file. Make sure to be in orthographic mode. Viewport-render \nwith Eevee. Render using camera of scene.\nI both cases a lot of plain white pixels appear in the hair.\nWhen changing to perspective mode, only pixels perpendicular to\nthe view (viewport or camera) are effected.\n[HairRoughnessProblem.blend](HairRoughnessProblem.blend)\n\n\n\n\n\n", "EEVEE: Add support for extracting alpha channels from a color\nI mistakenly filed a bug report --> #72856 EEVEE transparency problem with mtpaint PNG\n\nI didn't fully understand the bug and attached a blender file that didn't show the bug properly.\nPacking the image file with blender command 'Pack' makes the image\nopaque so it is worthless to include a blender document. I will\ninclude a mtpaint PNG image instead and you can load it into\nblender 2.81 yourself and see how it renders in EEVEE versus cycles.\n\nI noticed that your forum converts the image to opaque if I try to attach to this post\nso I'll use a shareable link that preserves the transparency:\n\nopen?id=1TjfSjzabYEmVHXTODPFymGVNwuL02SnJ\n\nHere is how I created the mtpaint image to reproduce the problem...\n\n1) Launch GIMP -->new gimp document -no alpha channel - paint with brush\n2) save as PNG format with no alpha channel\n3) Launch mtpaint ---> open the PNG document\n4) 'Save as' to output the document - one color chosen to be transparent - PNG format\n5) Launch blender 2.81 (linux) - images > Images as planes\n\nEEVEE rendering doesn't respect the transparent region.\nCycles rendering does respect the transparent region.\n\n\nNot a high priorty.", "mask visibility in sculpt mode using shape keys or generative modifiers\nOperating system: Windows 10\nGraphics card: nvidia 1060 Ti\n\n2.8 RC2\n\nIf you have a mesh with a mask and create a shape key, the mask will be no longer visible, still active but not visible.\n\n\n- Add UV Sphere\n- Sculpt mode\n- Paint a mask\n- Create shape keys\n\n\nOr from the default startup:\n- in sculptmode, `Mask` > `Invert Mask` (default cube should turn dark)\n- strokes dont have an effect now (this is expected)\n- add **any** generative modifier to the default cube (e.g. triangulate)\n- back in objectmode, save the file\n- reload the file\n- enter sculptmode\n- no mask is visible\n- strokes dont have an effect now (this is confusing since you are not aware masks prevent an effect since you dont even see the masks)\n\n(note we have the option `Use Deform Only` in sculptmode which might be exactly for that purpose, still not being able to use masks with generative modifiers should be more clear)", "With enabled Modifier On Cage option crashes on entering Edit Mode if in Render Viewport Shading\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.3.0\n\nIn Eevee, in Render Viewport Shading, an object having a Geometry Nodes modifier with enabled On Cage Modifier Result, that affects a UV Attribute which is used in Shader, crashes upon entering Edit Mode.\n\n - Set Viewport Shading Mode to Render\n - Enter Edit Mode ->Crash\n[displace-preview.blend](displace-preview.blend)", "Eevee Volume Artifacting\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 517.40\n\nBroken: version: 3.3.1\n\nI'm rendering some nebulae in Eevee and despite having 256 render samples, 256 volumetric samples and 128 volumetric shadow samples with 0.9 distribution and 2px tile size with proper start and end clipping distance both in camera and volume settings I'm getting really ugly flickering where the whole volume just artifacts which is really irritating. I'd love it if you could fix this issue in the upcoming blender release because even though I have a good computer it still takes 30 minutes to render the exact same scene in Cycles (where the issue is absent). watch?v=8ImqpM4zOEU\n\n[#101826.blend](T101826.blend)\n\n- Open file \n- Render animation\n\nIssue is visible after rendering ~10 frames when animation is inspected.\n\n", "Cycles: Volumes in the world shader behave strangle with some textures\n**System Information:**\nOperating system: Linux-5.14.0-1-amd64-x86_64-with-glibc2.32 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.63.01\n\nBroken: version: 3.0.0 Alpha, and every version between master at 2.81 (That is how far I tested).\nWorked: Unsure\n\n**Short description of error:**\nWhen using certain procedural textures in the volume output of the world shader in Cycles, you get unexpected results.\n\n**Exact steps for others to reproduce the error:**\n1. Create a scene in Cycles.\n2. Change to the rendered viewport.\n3. Open a shader editor and change to the `World` shader and delete the `Background` node.\n4. In the `World Shader`add a `Magic Texture` node and connect one of the outputs to the `Volume` socket on the `World Output` node.\n5. You will noticed that in the viewport you can see the `Magic Texture` is projecting light onto objects in the scene, but you can't see the `Magic Texture` directly.\n6. Place a `Color Ramp` node between the `Magic Texture` and the `World Output` nodes. Now you can see the `Magic Texture` in the viewport.\n\nAnother test:\n1. Create a scene in Cycles.\n2. Change to the rendered viewport.\n3. Open a shader editor and change to the `World` shader and delete the `Background` node.\n4. Add a `Noise Texture` node and connect an output to the `Volume` socket on the `World Output` node.\n5. Notice that you can see the impact the `Noise Texture` is having on the scene, but you can't see the `Noise Texture`\n6. Place a `Color Ramp` between the `Noise Texture` and the `World Output` nodes. You still can't see the `Nosie Texture` in the scene.\n7. Place a `Volume Scatter` node in the `World Shader`, connecting the output to the `Volume` socket on the `World Output` node and connect the output of the `Color Ramp` node to the `Density` input of the `Volume Scatter` node. You can now see the noise texture in the volume if you have a light in your scene.\n8. Select the left most handle (the handle that is black) in the `Color Ramp` node and drag it closer to the position `0.5`. You will see the results you expect until you reach the position `0.5` where the volume will seem to disappear from your scene. Yet when you look at objects in the scene, the light reaching them is still affected by a volume shader you can't see.\n\nHere's a video demonstrating this:\n[Volumetrics not visibile in viewport in certain sitautions.mp4](Volumetrics_not_visibile_in_viewport_in_certain_sitautions.mp4)", "Empty Images react with Viewport World Opacity\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.83.8 through 2.93 master.\nWorked: version: 2.82a\n\nThere is a problem in EEVEE where Empty Images Depth set to \"Back\" will also draw them behind the world HDRI background. So the Image will fade into the background when changing the Viewport Shadings World Opacity.\n\n(From the image you can see that it only reacts with values below 1.0. Opacity maxed out will make Empty images non transparent again. That's where I started to think this isn't fully intended to work like that, because consequential the value 1.0 would go full transparent.)\n\nMy issue here is that working with images shouldn't be relevant to this setting at all (as its shown for value 1.0).\nThe current behavior discourages to use world opacity in order to keep a clear image or if I want to use the color picker in the image for example.\n\n[EmptyImageOpacityBug.blend](EmptyImageOpacityBug.blend)\n\n1. Create an Empty, set to Image display.\n2. Set its Depth Setting to \"Back\"\n3. Change the Viewport Shading > World Opacity\n\n", "\"No Render\" volume of space with motion blur and GPU render\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA RTX A5000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 527.27\n\nBroken: version: 3.4.1\nWorked: N/A\n\nThere appears to be a volume of virtual space where objects do not render.\n\nLoad attached file. Set render frame to 110 and render the scene. Note the progress it makes as the large cube slowly vanishes with each render pass, while the smaller thin cube only partially renders. Set the frame to 50 and it renders fine. The problem starts with frame 106 and gets progressively worse for the large cube, while it seems to somewhat recover for the thin cube in later frames.\n\nThe problem does not repro with CPU rendering, and did not repro with my older graphics card way back when I first made this scene (GeForce 1080Ti). It repros with GPU rendering on both 3.4.1 and 3.1.0. It does not repro with motion blur disabled (hence the viewport render is also fine).\n[GPU Motion Blur Bug.blend](GPU_Motion_Blur_Bug.blend)\n", "Using python handlers introduces artifacts with EEVEE motion blur\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.90.0 Beta\n\n\nRendered image has \"wrong\" motion blur\n\n0. Open attached .blend file\n[motion_blur.blend](motion_blur.blend)\n1. It contains a script, which sets the cube position via a follow path constraint based on the current frame:\n\n2. Set motion blur steps to 1, render:\n\n3. Set motion blur steps to 8, render:\n\n", "EEVEE Next - 'Alpha Blend' - distant objects - shadow buffer overflow\n### Reported as per @fclem [request](111809#issuecomment-1014998)\n\n--\n\nLinux Mint 21.2 Cinnamon kernel 6.2.0-32\n3400G\n\n- Broken\nblender-4.0.0-alpha+main.b5c89822ac40-linux.x86_64-release.tar.xz\n\n- Worked\nN/A (new feature).\n\n- Short description of error:\nWhen an object's shader uses Alpha, and (Material tab) 'Blend Mode' uses 'Alpha Blend', distant objects cause a shadow buffer overflow.\n\n- Exact steps\n1. Open following .blend file and zoom (far) out / back in.\n[EEVEE Next - 'Alpha Blend' shadow buffer overflow.blend](attachment)\n\n2. Appears no issue while using 'Alpha Clip' or 'Alpha Hashed'.\n\n'Alpha Blend' preferred for preventing Z fighting in Orthographic view.\n\nKind regards.\n\n**Screenshot**\n\n\n", "Eevee refraction regression in orthographic view\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.91\nWorked: 2.92\n\nIn orthographic view, with the view or camera camera perpendicular to a refractive surface, screen space refractions appears blurry, as if it had roughness, even with refraction shaders with 0 roughness\n\n2.91: \n\n2.93: \n\nEdit: In perspective it's also noticeable in the center. It looks like a small ball or dot that distorts the image slightly\n\n\nOpen file in different versions\n[bugRefractionEevee.blend](bugRefractionEevee.blend)\n\n\n", "Two materials: volumetric being in second slot causes error in EEVEE\nOperating system: Windows 10 64 bit\nGraphics card: Intel HD Graphics\n\nBroken: 2.90.1\nWorked: IDK, but I've been using blender since 2.79 and have not encountered this problem when I have done this kind of thing.\n\nWhen I have a mesh object that has both solid and Volume Scatter materials on it (in different slots), in EEVEE, the volumetric material only works if it is not in the second slot. If it is in the second slot (top down), it causes the black \"shader missing\" to be applied. Moving the volume slot up to the top with the solid material in the slot at the bottom causes normal operation.\n\nGo into edit mode on the default cube.\nDuplicate the mesh, but leave it at 0 0 0. Scale the duplicate up a little.\nCreate a second material slot, and assign the duplicate larger cube to it. Leave edit mode.\nCreate a \"Volume Scatter\" material for the second slot. The outer cube will turn black.\nMove volumetric material up in list of materials. The outer cube will change to the proper fuzzy grey.\n\n[#83538.blend](T83538.blend)\n", "Light of 'Area' type lamp presents artifacts when placed at great distances\nOperating system: macOS-13.4.1-x86_64-i386-64bit 64 Bits\nGraphics card: Metal API AMD Radeon Pro 5700 XT 1.2\n\nBroken: versions: 3.1 to 4.0a \n\nArtefact's appearing with bigger Arean Light distances and spread angle 0.1 to 20~ \n(Like for space scenes with bigger distance Area lights for better sun effects...)\n\nJust open one of the files or follow the image below to set up the light.\n\n" ]

[ "Eevee spotlight clipping through solid objects while using Principled Volume\n - AMD® Ryzen 7 1700x eight-core processor × 16\n- 16 GB\nOperating system:\n - Either Windows 10 or Debian 10\nGraphics card:\n- GeForce GTX 1060 6GB/PCIe/SSE2\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\nWorked: (optional)\n\n[system-info.txt](system-info.txt)\n[Volumetric_with_spotLight.blend](Volumetric_with_spotLight.blend)\n\n\nIf you use a spotlight inside a solid object (Simulating a lamp reflector) and you use principled volume on your scene when you render,\nyou can see the light in the other side of object. Like as an interference.\n\nYou can reproduce in the blender file attached.\n\n\n" ]

[ "Cycles doesn't render smoke properly with adjusted clipping distances\nWindows 10 / AMD Radeon 6800 Series \nBroken: 2.79.1 1802d14\n\nI need to render a cross section of a smoke simulation. I set up the simulation and adjust the start/end clipping distances of the active camera to the desired values and in OpenGL viewport it works fine, renders exactly the cross section I want.  \nHowever, when I go into cycles and render it out, I get nothing: [*The second camera next to the active one is just for the convenience sake, mimicking the clipping values. Cycles behaves the same without it.*] \nI know that the materials are working fine because I can render it outside the camera's view: \nIf it helps, here's something else that I've found. When the end clipping distance goes beyond the Flow mesh (emitter) I get this weird result: \nAnd when it (end clipping distance) goes even outside the adaptive domain box, then it renders the smoke: \n[smoke.blend](smoke.blend)", "EEVEE-Next Irradiance Baking\nDue to time constraints it isn't possible to start implementing a solution that leverage raytracing. So the implementation will try to minimize the amount of changes to the current system while cleaning up the implementation and changing the baking algorithm.\n\nHere are the reference papers the new solution will be based on:\nSIGGRAPH2022-Advances-Enemies-Ciardi%20et%20al.pdf (Probe-based lighting section)\nHobson_Josh_The_Indirect_Lighting.pdf\n\nOne remaining unknown is whether or not we keep the light-grids behavior as it was in the original EEVEE or if we go for a single large light-grid structure.\n\nHere is a breakdown of the tasks:\n- [x] #105802 Add debug visuals for surfels.\n- [x] Port Irradiance Cache display.\n- [x] Port Spherical Harmonics lib & BSDF lib and add unit tests.\n\n#### Milestone: Dummy irradiance grid displaying.\n- [x] Rewrite cache structure load / store.\n- [x] Rewrite light-cache baking in C++ with new draw-manager and unit-tests.\n\n#### Milestone: Baking, Saving to and Loading from disk should work.\n- [x] Make shader variation to output surfel.\n- [x] Bin surfels to ray tiles.\n- [x] Sort surfels in ray tiles, dead simple sorting for now.\n- [x] Traverse the rays and add contribution to nearest surfels.\n- [x] Insert irradiance samples to ray list.\n- [x] Project result to irradiance grid.\n\n#### Milestone: Irradiance baking fully works.\n- [x] Add indirect diffuse contribution to BSDFs\n- [x] #110386 Add dilation for invalid lightprobes\n- [x] #110312 Add sampling biases\n- [x] #110851 Add leak prevention\n- [ ] #106449 Finalize the design and implement the behavior of multiple irradiance grid.\n\n#### Bugs / Cleanups\n- [x] Add memory budget option\n- [ ] Add safety check for out of memory cases\n- [x] #110858 Add back intensity option\n- [x] #110858 Add back clamp option to avoid fireflies\n- [ ] Add smooth transition between irradiance volumes\n- [ ] Fix append / link of baked data\n- [ ] Add back capture bounds parameters\n- [ ] Add option on light and object to exclude from irradiance cache\n- [ ] Add option on material to be considered double sided by irradiance bake\n\n#### Milestone: Irradiance bake minimum-viable-product for 4.0 release.\n\n#### Extended targets\n- [ ] Support transparency\n- [ ] Support translucency\n- [ ] Support volumes during baking\n- [ ] Sky visibility baking\n- [ ] Auto Sample Resolution based on density\n- [ ] Sparse grid\n- [ ] Light injection\n- [ ] Multi-LOD grid\n\n#### Dropped features (not considered regressions)\n- Capture Clip Near / Far distances: Capture is done for the whole scene.\n- Bounce Count: Light transport is done using an infinite number of bounce and converge to the correct solution with more added samples.\n", "Eevee Motion Blur not applied to AOVs\nOperating system: Ubuntu 18.04\nGraphics card: GTX 2080 Super\n\n2.92 Official\n\nWhen rendering with eevee, motion blur is only applied to the beauty pass. Other passes such as cryptomatte or normal dont have mblur applied, but you can see the steps of the accumulation blur in some passes. This is a huge problem for compositing, it makes all the passes mostly useless.\n\n- animate an object\n- turn on passes\n- render with eevee with mblur on\n- check the aovs, there will be no mblur on them\n\n[aov_motion_blur.blend](aov_motion_blur.blend)", "Non Uniform Scaling possible when Camera is parented\nOperating system: Win 10\nGraphics card: Geforce 2070\n\nBroken: 2.82-3.2\n\n\nIf you parent a Camera to an Object and scale the parent non uniformly, the Camera View gets scaled and distorted too.\n\n\nCreate an Empty, parent the Camera to an Empty, scale the Empty. The Camera View gets distorted.\n\n[blender_camera_scaling_bug.blend](blender_camera_scaling_bug.blend)", "EEVEE is ghosting with viewport denoising enabled.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.0.0 Alpha\n\nHere's a viewport render:\n[EEVEE ghosting0001-0090.mp4](EEVEE_ghosting0001-0090.mp4)\n\nHere's what it looks like during playback if viewport denoising is enabled:\n[2021-10-23 17-07-44.mp4](2021-10-23_17-07-44.mp4)", "Moving a keyframed object starts moving from wrong location when linked scene open in separate window\nOperating system: linux\nGraphics card: nvidia\n\nBroken: 3.2\nWorked: didn't check\n\n- Create a keyframed object that moves, i.e. frame 1 (object.x=0) and frame 30 (object.x=10)\n- Create a linked copy scene (Scene.001) and open a new main window set to that scene, and set timeline to frame 1.\n- In original window in original scene, scrub timeline to anywhere after frame 1.\n- Press \"g\" to move object in original window. Notice the object jumps to where it is in the other main window (Scene.001) instead of starting to move from where it is visually in the current window and scene.\n\nExpected:\n- Moving respects the current window/scene.\n\nRelated issues:\n- The keyframe color indicators stay orange (instead of turning yellow) when scrubbing to the frame with the keyframe and the location values don't update.\n\nTest file (first 3 steps included)\n[#99597.blend](T99597.blend)", "Slow EEVEE timeline scrubbing when using Scene World material\nOperating system: Arch Linux\nGraphics card: Intel UHD 620 (`iris` with `INTEL_DEBUG=reemit`)\n\nBroken: `0781c22ceedc`, `2.90`\nWorked: \n\n\nScrubbing through the timeline on an empty scene is slow when viewing through EEVEE and Scene World enabled, even though the node graph for it is a simple Background Shader node.\n\nThis also happens when *nothing* is connected to World's material output.\n\n[slow_scrub.mp4](slow_scrub.mp4)\n\nYou can see in the first part of the video, when using LookDev and have Scene World disabled, the scrubbing is relatively fluid and very close to the cursor. After enabling Scene World you can see that the scrubbing is lagging behind the cursor.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n1. Create a new scene\n2. Delete everything (not really necessary)\n3. Try scrubbing the timeline without Scene World in LookDev\n4. Try scrubbing the timeline *with* Scene World in LookDev\n5. Notice that the latter is less responsive\n\n**Note**\n\nI have only experienced this on this GPU (probably due to its low-performance), however I still feel this is a weird bug.", "Using View Navigation in Camera View Causes Viewport View to Focus on World Origin\nOperating system: Windows-10 64 Bits\nGraphics card: GeForce GTX 1070 NVIDIA 442.19\n\nBroken: version: 2.83 (sub 6)\nWorked: Never (2.79+)\n\nUsing view navigation (Shift+\") in camera view (Numpad 0) causes to lose the previous viewport view (Numpad 0 again) and it focuses world origin. \n\n1- Launch default Blender.\n2- Change your view to anywhere (For example zoom in to a vertex of the cube).\n3- Go in camera view (Numpad 0), use view navigation (Shift+\") and right click to cancel it (you can change view with left click as well, it doesn't matter).\n4- Go in viewport view (hit Numpad 0 again), notice that your previous view changed (it now focuses world origin)\n\nBelow gif demonstrates the issue:\n\nView navigation that I used:\n\n\n", "HDRI animation bug in eevee\nThe HDRI illumination shouldn't change on the cube, because every thing including the HDRI is rotating at the same speed. \nHowever, this went wrong in EEVEE... It seems that in Cycles everything went on so well.\n\nThis is a simplified bug presenting file of a complex scene which I found this bug, in which I have to rotate an object, a HDRI and camera together, and I'm struggling to get rid of this bug...\n\n[BugReport.blend](BugReport.blend)\n\n[0001-0096.mkv](0001-0096.mkv)\n\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11\n\nBroken: version: 2.93.1\n\n\n\n", "The viewport lock modes are super glitch (in general) with zooms and centering\nOperating system: Windows 10\n\n\nBroken: blender-3.0.0-alpha+master.eae4e2251839-windows.amd64-release\n\n\n\nThe viewport lock modes like 3d cursor or object locking is super glitchy in general which creates a disorienting experience. The main issue is sudden jumps in zooms also sometimes the center or rotation becomes jumpy as well. Some jumps in the video is worse than the others. Bear in mind that I am not zooming intentionally when there is a sudden jump.\n\nThis gets even worse in different scene scales and scene layouts. Here I am demonstrating it only with the cursor but this erratic behavior happens with object locking too. This becomes even worse with a Wacom tablet since it is harder to keep the pen movement as static as the mouse while zooming.\n\n\n- Load the attached .blend file\n- Make sure the viewlock is set to the cursor, and have similar preferences as shown in the pictures\n- Try repositioning the cursor, just move the cursor to different objects.\n- Move the cursor to the surface of another object then try ctrl+single middle click to activate the zoom\n\n\n\n\n\n\n\nTest File:\n[glitchy_lock_modes_28082021_2053_04.blend](glitchy_lock_modes_28082021_2053_04.blend)\n\n[A7ysxb6GVc.mp4](A7ysxb6GVc.mp4)", "Eevee renders full frame and cut it instead region rendering\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.82 (sub 6)\n\n\nI tested with same settings and I have the same time for whole frame and region.\n\n\n\nIt makes problem, when I want to render small part zoomed. Now the workaround — is making new camera with smaller angle and render from it.\nIf it is limitation, may be render should create temporary camera that cuts region.\n\nRender same scene with Eevee through {key F12} and it`s region and compare render time.", "GP transparent layer doesn't blend with background object \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 3.0.0 Alpha\n\nI have a foreground grease pencil object, and a background grease pencil object (yellow blob).\nThe foreground object contains a layer set to 0 opacity.\nWhere this layer overlaps the background object in the viewport, the background object is invisible.\nExpected behaviour: the visibility of the background grease pencil object isn't affected by transparent layers in a foreground object.\n\n\n\n\n\n[#89286.blend](T89286.blend)\n\nForeground object \"makes hole\" in background, doesn't blend. Rotate viewport camera, suddenly it blends correctly.\nWhen foreground location is set to 0,0,0 it blends correctly too.", "Render automatically sets scene camera when there is none\nOperating system: Arch Linux\nGraphics card: GTX TITAN\n\nBroken: blender 2.80.68 (dfbbc5067e40)\n\n\nThe different ways in the active camera is shown isn't updating consistently, nor always correct\n\n\n- start with default scene (cube, light, cam)\n- add a 2nd camera, make it active and switch to cam view with Ctrl+Numpad0\n- use something like fly mode to get a good view of the cube\n- switch to Scene properties, note Camera is set correctly\n- delete the active camera with X\n- there's no active camera anymore, shown consistently:\n - Camera is empty in Scene properties\n - direction triangle of remaining cam is not filled\n - outliner doesn't show outline rectangle on camera objectdata\n- however, press F12, render works using the remaining camera\n- escape image view\n- select the remaining camera\n- move it a bit in 3D view\n- press F12, note that render is using new view\n- in Scene properties the Camera selection is now sometimes correct, sometimes still empty. Similar for camera objectdata icon in the outlienr. However, the direction triangle on the camera stays empty\n", "Particle systems on a moving emitter with particle settings start/end outside the animation frame range showing up in wrong place intermittently\nOperating system: Windows 10\nGraphics card: RTX 3080\n\nBroken: 3.0.1\nWorked: xx\n\n[Desktop 2022.02.10 - 14.44.35.08.mp4](Desktop_2022.02.10_-_14.44.35.08.mp4)\n\nRepro steps with particle_bug.blend (attached)\n[particle_bug.blend](particle_bug.blend)\n1. Press play, observe:\n - on the initial frame of the animation, a cubic volume of particles spawn at the initial position of the cube\n - when the animation loops, the trail of particles disappears\n**This is the behavior I believe is a bug.**\n\n2. Change the animation start frame from 40 to 0, and let the animation play through at least one full loop.\n3. Change the animation start frame back to 40, and observe:\n - on the start frame of the animation, no cube of particles appears at the initial position of the cube\n - when the animation loops, the trail persists with each particle seeming to remain in its existing location.\n**This is the behavior I believe is intended when using duplicate particle systems with non-overlapping timespans to create the illusion of looping particles.**\n\n4. In the particle properties tab on the right, select the 'during' particle system (instead of the currently selected 'pre' particle system), and observe:\n - the behavior from step 1 immediately resumes: the particle trail disappears and the cube of particles appears at the initial position of the cube on the first frame of the animation.\n**This is the easiest way that the bug can be reproduced predictably.**", "Geometry Nodes fouling up motion blur\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA RTX A5000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 526.86\n\nBroken: version: 3.1.0\nWorked: N/A\n\nGeometry Nodes applied to a spline which affect its shape over time do not blur when Motion Blur is enabled.\n\nLoad attached simple file. Render frame zero. There is no blur at all. Render frame 1, *everything* is blurred, despite only part of the spline is in motion. It would appear the entire spline is blurred according to how the base object is moving, not how the spline is deformed as a result of the applied Geometry Nodes.\n\nNote that same file loaded into 3.3.1 causes Blender to crash (GPU rendering only - it's a stack corruption bug in the nVidia driver, but 3.1 doesn't cause a problem). I previously reported that crash but it was rejected due to possibly being related to the Photographer add-on, which I have uninstalled. It still crashes.[Motion Blur Bug.blend](Motion_Blur_Bug.blend)" ]

[ "No motion blur on moving particles in Eevee\nOperating system: Windows 10\nGraphics card: Nvidia GTX 1050\n\nBroken: 2.90 beta, 9de5adc6a1a9, August 03, 2020\n\nThere is no motion blur in Eevee for moving particles.\n\nExact steps:\n[20200804particlenomotionblureevee.blend](20200804particlenomotionblureevee.blend)\nOpen blend file and hit render with the particles visible in the shot. Even though they are moving away from the camera, they are not blurred." ]

[ "\"Reset to Default Value\" sets alpha incorrectly in [Material Properties > Viewport Display > Color]\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.09\n\nBroken: version: 3.5.0\n\nRight-clicking \"Reset to Default Value\" on [Material Properties > Viewport Display > Color] sets both the color's \"value\" and \"alpha\" bars to 0.8, when the alpha one specifically should reset to 1. Alpha resets incorrectly.\n\nRight-click \"Reset to Default Value\" on [Material Properties > Viewport Display > Color] on any material.\n\n", "Metal Viewport flickering on interaction\nOperating system: Mac OS Monterey, Version 12.6.2\nGraphics card: Apple M1 Max, Mac Book Pro 16 inch 64 GB, LG Ultra Fine Display as secondary monitor\n\nBroken: 3.5.1 release candidate\nWorked: 3.4 with Open GL does not have these issues\n\n\nWhen interacting in the viewport say moving a camera the viewport flickers. Also when adjusting shaders in a cycles window the viewport also flickers depending on the speed of the interaction. Evee viewport does not have this issue, but solid render does\n\nBased on supplied file Look through camera and split the viewport in two. Left side has cycles rendering turned on. Change the default renderer from Eevee to Cycles and change to GPU compute. Rotate camera with the empty and cycles view on left flickers. Next select cube and change the shader colour from pink to any other colour you like. Again viewport flickers like crazy.\n\n", "Lights with radius 0 produce a dark ring in EEVEE\nOperating system: Microsoft Windows 10 Pro\nGraphics card: Nvidia GeForce GTX 1050\nProcessor: Intel(R) Core(TM) i5-2500 CPU @ 3.30GHz, 3301 Mhz, 4 Core(s), 4 Logical Processor(s)\n\nBroken: blender-2.91.0-02677ec4e0fa-windows64\n\nIn EEVEE, objects with certain roughness values have a ring inside the specular highlight in which there is no light, when the light source has a radius of 0. Changing the radius to have a higher value makes the ring go away. Setting the value to near zero values such as 0.002 will still have the ring, but smaller. Both distributions GGX and Multiscatter GGX have the problem. Only the lights with a radius parameter have the ring (point and spot lights). Area lights use a size parameter and behave differently, so they do not have this issue. Sun lights have an angle parameter instead, and the fact that they are treated as having an infinite distance may also play a role.\n\n\n[specular_ring.blend](specular_ring.blend)\nOpen blend file\nSet mode to rendered view\nNotice a dark ring around the specular highlight\nChange radius to value over 0.005\nThe ring goes away", "Cycles: Baking to an externally loaded image fails to update the image buffer results in Cycle's Rendered Viewport Shading\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.59\n\nBroken: version: 3.3.0, 2.93\nWorked: probably never\n\nBaking to an externally loaded image will not changed the baked image as displayed in the Viewport Shading's Rendered Mode.\nThe newly baked image buffer is shown correctly in the image editor and (EEVEE's) Material Preview, but wrong in Cycles's render view.\n\n- Open the attached .blend file.\n- Open the attached baked image in the image texture node.\n- Use it to bake any type of pass to the loaded image.\n- Observe that the image looks unchanged in the Rendered View by connecting the image texture node to the material output.\n[#98063.blend](T98063.blend)\n\n[2022-05-12 09-12-20.mp4](2022-05-12_09-12-20.mp4)", "The colors of the vertices do not correspond to the values in which I paint them with vertex paint.\nOperating system: Windows-10-10.0.14393-SP0 64 Bits\nGraphics card: GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.92.0\n\nthe vertex color values that I get from the vertices in the shader editor node called \"vertex_color\" do not correspond to the values that I write there using vertex painting.\n\n\nSet the viewport mode to render or material preview. And change the inputs from the manual value to the value taken from vertex color. As you can see, the values themselves should not differ, but the color change.\n\n[vertex_color_error.blend](vertex_color_error.blend)\n\nPerhaps this is due to this issue D10956, but I want to say that we then need the ability to convert from one color space to another.\nBecause I use a blender to create game models and would like to use vertex colors to mix materials. But I can't, because it doesn't correspond to the colors and values that I enter. And it will not match what I will see in the game.\nIn addition, I enter specific values using the color wheel and sliders. And I expect to get them unchanged. That would be logical.\n", "Image Editor: Viewing a value buffer could use incorrect channel selector\nThis is a long outstanding issue in the image editor that should be fixed.\nWhen rendering a value based AOV and switching in the render result to the AOV would render incorrect as the wrong channel selector is selected (Still Color + Alpha).\n\nThis is also visible when reading single channel images or OpenEXR files.\n\n**Solution**\nAdd an option to view `Value` option to the channel selector and change RNA get/set/itemf + add support of this to the image engine.", "EEVEE Shading: string sockets node group interface value doesn't work\nBroken: version: 3.0.0 Alpha\nWorked: n/a\n\nin shading, String sockets exposed via node group interface don't use the value from the node group interface\n\nSteps to illustrate the problem\n\n* Use a quick smoke effect on a cube\n* Add a principled volume and connect it to material volume output\n* Change the Density Attribute to other text\n* Group the principled volume\n* Change the Density Attribute back to density\n\n[volume.blend](volume.blend)\n\n", "Black Artefacts in EEVEE on Transparent BSDF when using Alpha Hashed for Blend mode and Camera Depth Of Field\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.92.0\n\nBlack Artefacts in EEVEE on Transparent BSDF when using Alpha Hashed for Blend mode and Camera Depth Of Field, if Shader Color is not pure white (similar problem with Layer Weight)\n\n**Exact steps for others to reproduce the error**\n- Open attached file (Just a simple box (can be a plane) with Transparent BSDF. Camera' distance or depth of focus influences the result. Attached pictures in Image Editor from a real case).\n- Press F12 to render\n[transprarent_shader_artifacts_with_camera_DOF1.blend](transprarent_shader_artifacts_with_camera_DOF1.blend)\n\n\n", "White Noise Glitch using Value Attribute of Particle Instances\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.93.0 Alpha\nWorked: (?)\n\nThe White Noise Texture (in 1D) does not display as expected using the Value Attribute of Particle Instances in both EEVEE and Cycles. Instead of showing a single random value or color on the surface of an instance there's a fuzzy mess of different values similar to the look of overlapping faces:\n\nThe issue also appears on non flat geometry not only on planes. The using the Index attribute shows the same result. It's the same on the current stable release (that being 2.91.2) and previous releases. I tried the same setup with a Noise Texture and it seems to work just fine:\n\n- Choose any object as the emitter of a particle hair system.\n- Adjust the settings of the particle system to display the desired quantity of hairs\n- Add another object and add a Particle Instance Modifier to it\n - under 'Object' set the emitter object\n - check 'Create Along Paths'\n - in the object data properties under vertex color create a new vertex color\n - choose the same vertex color as the 'Value' data of the particle instance modifier under 'Layers'\n- use the created data as the name in the Attribute Node\n- connect any output of the Attribute Node to a White Noise Node set to '1D'\n- display the results\n(same last 5 steps for Index Attribute)\n[ParticleInstanc_ValueAttributeBug.blend](ParticleInstanc_ValueAttributeBug.blend)\n\nThank you so much for looking into it! And keep up the great work.", "Eevee crashes when rendering some of the Blender Studio rigs\nOperating system: Windows 10 Home\nGraphics card: Intel(R) HD Graphics 4600\n\nAlso tested on Debian 11 on both GNOME and MATE. It didn't make much of a difference, the GNOME session crashes and takes me back to the login manager meanwhile on MATE the program gets killed after it freezes for too long. I'd say that this is just a difference in the behaviour of the two desktop environments, but I can't say for sure.\n\nAlso asked a friend of mine to test it on their MacBook Pro, they claim to have seen the same results as I have. Just keep in mind that they're busy and I don't have access to their hardware.\n\nTried opening it on multiple versions:\n3.0.1 (stable), 3.5.1 (stable) and 4.0.0 (alpha) with [66d20c8723e6](66d20c8723e6) being the latest commit.\n\n(Windows/Linux/Mac)\nWhen opening it from the list of recent files I see the Blender project being displayed on the screen, but apparently for only one frame then freezes. I can see the prompt that asks for script execution to be enabled, the model itself, the pose library, etc. It just freezes afterwards.\nSame thing happens for when opening it via File -> Open, but I only see the default cube project before the crash.\nGot it from .\n\n1. Run Blender\n2. Open `ellie_animation.blend`\n\nSomething else worth noting is that apparently a group of people on the Blender Artists forums are experiencing the same issue: 1342795\n\n`blender-stdout.txt` and `blender-stderr.txt` were generated like so:\n`C:\\Users\\Mustafa\\blender-4.0.0-alpha+main.66d20c8723e6-windows.amd64-release>blender --factory-startup --debug-all > blender-stdout.txt 2> blender-stderr.txt`\nI've even provided a file (`gdb-blender.txt`) containing the output of interactively debugging blender with the GNU Debuger, in case you'll need it.\nAll debug logs and related files were created on Windows.\n\nThanks in advance! It'd take a miracle to get such a huge open source project going and you're doing a great job, I appreciate your hard work on Blender.\n", "Eevee viewport soft shadows are broken on higher sample count\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.92\n\nBroken: version: 2.93.0\nWorked: ---\n\nEevee soft shadows in viewport have color spills. But it works fine on final render (F12)\n\n\n\n- Open .blend file\n\n[bug_soft_shadows.blend](bug_soft_shadows.blend)\n\n- Switch to Rendered Mode of Viewport Shading\n- Wait half a min and see the problem(", "Different padding between eevee and cycle properties\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: GeForce GT 630M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 3.0.0 Beta\nWorked: ???\n\ndifferent padding see image below:\n\n\n\ni dont know is this because new ui design or not.\n\n\n", "Blender 3.5 and 3.5.1 are crashing while using Metal as backend\nOperating system: macOS 12.6.3\nGraphics card: Radeon Pro 580 8GB\n\nBroken: 3.50, 3.51 (Metall Backend)\nWorked: 3.50, 3.51 (OpenGL Backend)\n\nBlender simply hang. I use eevee, if you switch to rendered viewport shading it crashed complete system. I use Metall as backend, if I choose OpenGL it works fine. \n\n \nI attached an .blend file for example.\n\n", "Two materials: volumetric being in second slot causes error in EEVEE\nOperating system: Windows 10 64 bit\nGraphics card: Intel HD Graphics\n\nBroken: 2.90.1\nWorked: IDK, but I've been using blender since 2.79 and have not encountered this problem when I have done this kind of thing.\n\nWhen I have a mesh object that has both solid and Volume Scatter materials on it (in different slots), in EEVEE, the volumetric material only works if it is not in the second slot. If it is in the second slot (top down), it causes the black \"shader missing\" to be applied. Moving the volume slot up to the top with the solid material in the slot at the bottom causes normal operation.\n\nGo into edit mode on the default cube.\nDuplicate the mesh, but leave it at 0 0 0. Scale the duplicate up a little.\nCreate a second material slot, and assign the duplicate larger cube to it. Leave edit mode.\nCreate a \"Volume Scatter\" material for the second slot. The outer cube will turn black.\nMove volumetric material up in list of materials. The outer cube will change to the proper fuzzy grey.\n\n[#83538.blend](T83538.blend)\n", "Support Face color attributes for painting\nFace color attributes should be supported for sculpt painting and vertex paint mode.\n\nThey should also be listed under the Color Attributes UI List.\n\n" ]

[ "Hair with material using a non-existing attribute looks different between Eevee and Cycles\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.0 Core Profile Context 22.20.00.20.220516\n\nBroken: version: 3.4.0 Alpha\nWorked: -\n\nHair can be affected by materials that use attributes such as mesh.\nThe color of the mesh at the root position of the strand usually defines the color.\nBut that's not the case in Eevee if an attribute doesn't exist.\n\n\nThe actual color can be kind of undefined, but there should be a convention between the renders.\n\n- Open attached .blend file\n- Compare Cycles and Eevee\n[hair_material_with_non-existing_attribute__bug.blend](hair_material_with_non-existing_attribute__bug.blend)\n\n" ]

[ "The program crashes when texturing \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1660/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.72\n\nBroken: version: 2.93.0\n\nWhen try texture object in texture paint mode, and there are another map in shader tab. Program crashes. \n\n1. In default (or other) scene go to Texture paint window, create new texture for paint. \n2. Apply this texture to diffuse color object in shader editor. (or in object properties)\n3. Go back to texture window. We make sure that everything is painted correctly\n4. Go to shading window and drag&drop any other texture to node window. (NOT APPLY TO DIFFUSE COLOR (or APPLY))\n5. Go back to texture paint window. For some reason, the dragged texture itself was assigned to the duffuse color.\n6. Change linced texture in image editor (in left window)\n7. Try paint on RIGHT window (3D viewport)\n8. Click on left texture editor (try paint in 2D ) program crash. \n\nVideo steps: 0Ophy4Lmx0A\n\nDebug logs:\n[blender.crash.txt](blender.crash.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[Blender_crash.webm](Blender_crash.webm)\n\n", "Segfault in CustomData_blend_read\nOperating system: Debian Linux, Stable\nGraphics card: GTX970\n\nBroken: 3.6.0 Alpha, branch: main (modified), commit date: 2023-04-05 18:55, hash: 92c9c1d4003e, type: Release\nWorked: Unsure. Old blend I loaded and worked on for a while. Errors on open now.\n\nSegfault on load.\n\n1. Load blend, or open blender and load blend\n\n(No idea what the gvfs error is.)\n```\nFailed to open dir (No such file or directory): /run/user/6001/gvfs/\n[New Thread 0x7fffe3ae9700 (LWP 2310937)]\n...\n[New Thread 0x7fffdb9fe700 (LWP 2310945)]\nRead blend: \"/tmp/segfault.blend\"\n\nThread 1 \"blender\" received signal SIGSEGV, Segmentation fault.\n0x0000000001009af9 in CustomData_blend_read (reader=0x7fffffffd498, data=0x7fffd9c471f0, \n count=5002)\n at /opt/src/blender-git/blender/source/blender/blenkernel/intern/customdata.cc:5194\n5194 if (layer->flag & CD_FLAG_EXTERNAL) {\n(gdb) l\n5189\n5190 int i = 0;\n5191 while (i < data->totlayer) {\n5192 CustomDataLayer *layer = &data->layers[i];\n5193\n5194 if (layer->flag & CD_FLAG_EXTERNAL) {\n5195 layer->flag &= ~CD_FLAG_IN_MEMORY;\n5196 }\n5197\n5198 layer->flag &= ~CD_FLAG_NOFREE;\n(gdb)\n```\n\n", "Segfault pasting stale driver.\nOperating system: Linux\nGraphics card: Intel\n\nBroken: 2.92\n\n\nDrives copied and pasted with significant change in file data in between can trigger segfaults. \n\n```\n# Blender 2.92.0, Commit date: 2021-02-24 16:25, Hash 02948a2cab44\n\n# backtrace\nblender(BLI_system_backtrace+0x34) [0x55dfd70fd084]\nblender(+0xe151bd) [0x55dfd4bd41bd]\n/usr/lib/libc.so.6(+0x3cf80) [0x7f9b26870f80]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder26build_object_data_geometryEP6Objectb+0x170) [0x55dfd7000380]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder17build_object_dataEP6Objectb+0x54) [0x55dfd6ffb4b4]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder12build_objectEiP6ObjectNS0_26eDepsNode_LinkedState_TypeEb+0x1d3) [0x55dfd6ffe2c3]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder22build_driver_variablesEP2IDP6FCurve+0x77) [0x55dfd6ffad87]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder14build_animdataEP2ID+0x1f9) [0x55dfd6ffe8b9]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder12build_objectEiP6ObjectNS0_26eDepsNode_LinkedState_TypeEb+0x228) [0x55dfd6ffe318]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder16build_view_layerEP5SceneP9ViewLayerNS0_26eDepsNode_LinkedState_TypeE+0xb2) [0x55dfd7003892]\nblender(_ZN7blender3deg23AbstractBuilderPipeline16build_step_nodesEv+0x76) [0x55dfd6fea296]\nblender(_ZN7blender3deg23AbstractBuilderPipeline5buildEv+0x42) [0x55dfd6fea4b2]\nblender(DEG_graph_build_from_view_layer+0x33) [0x55dfd6fd24e3]\nblender(+0xef501a) [0x55dfd4cb401a]\nblender(wm_event_do_depsgraph+0xc5) [0x55dfd4f59245]\nblender(wm_event_do_refresh_wm_and_depsgraph+0x92) [0x55dfd4f593b2]\nblender(wm_event_do_notifiers+0x36a) [0x55dfd4f6055a]\nblender(WM_main+0x2b) [0x55dfd4f5532b]\nblender(main+0x367) [0x55dfd4b9da67]\n/usr/lib/libc.so.6(__libc_start_main+0xd5) [0x7f9b2685bb25]\nblender(_start+0x2e) [0x55dfd4bd056e]\n\n# Python backtrace\n```\n\n\n* Add a custom property to the Scene.\n\n* Right click on its value field, and run \"Copy as New Driver\".\n\n* Run `CTRL`+`N` or \"File\" —> \"New\" —> \"General\".\n\n* Right click on any value field, and run \"Paste Driver\".", "Disk Cache combined with Cache Step does not work\nOperating system:\nGraphics card:\n\nBroken: blender-2.80.0-git.8d95dcd87dbb-windows64\nWorked: Blender 2.79 b f4dc9f9d68b\n\nBaking a particle system with *Disk Cache* checked and *Cache Step* set to greater than 1 results in the baking failing.\nNo error is reported, but the bake is obviously corrupt.\nIt worked in 2.79.\n\n[bake.zip](bake.zip)\n1. Extract the archive and open the *.blend* file. (in 2.80)\n2. Playback works at this point. (Particles fall down from the plane.)\n3. Click *Delete Cache* and then *Bake*.\n4. Try playback again. It's not incorrect. (Particles appear at the bottom and don't move. If you check the cache folder, you will see it cached every second frame---instead of every 10th as configured, except for the first step.)", "Glossy Direct/Indirect pass multiplied by the Glossy Color pass using Principled BSDF (Cycles and EEVEE)\nOperating system: W10\nGraphics card: Titan Xp \n\nBroken: since the principled BSDF was introduced in blender (2.79)\nWorked: -\n\nGlossy Direct and Indirect pass are already multiplied by Glossy Color. Glossy color pass is always = 1 (white)\n\n- Download this file :\n[PRINCIPLED_BUG.blend](PRINCIPLED_BUG.blend)\n- Render image\n- Compare Glossy pass in the compositor (Left : old glossy shader, right : principled)\n\nas you can see, in the glossy direct and indirect pass, the glossy color is already multiplied :\n \n\nand glossy color pass is white :\n\n\nWhen everything is combined in compositing, the result is correct (because the Glossy color is already multiplied and his pass is white)\nbut we can't actually work with it with any software.", "Geometry Nodes: Crashes when the curve resampling node length is 0\nOperating system: macOS-13.4.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M2 Pro 1.2\n\nBroken: version: 3.6.0\n\n\n\n\nThis is an irregular operation, but I hope it is not a crash but a warning.\nI don't know if this is a bug or a feature request. \nThere is also a possibility that if the length is close to zero, it will cause the computer to have 100% CPU.\n\n", "Eternal compilation of Eevee shaders\nOperating system: macOS-12.0.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro Vega 64 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.7.29\n\nBroken: version: 3.0.0 Beta\n\nIn layout, clicking shading workspace ta brings spinning when of death.\n\n- Open attached file\n- click in the Shading workspace\n[tudor-bedroom 2.blend](tudor-bedroom_2.blend) \n\n", "Material-based displacement on source geometry fails to bake\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.91.0 Beta\nWorked: ?\n\nWhile baking in Selected to Active mode, applying a displacement shader to source geometry fails to actually displace the geometry, resulting in an incorrect bake.\n\n1. Open attached .blend file.\n2. Bake the texture.\n3. Compare to viewport render of the scene.\n[displacement_bake.blend](displacement_bake.blend)\n\n", "Blender development todo list -- Render\nNOTE: this is a direct dump of Render (without cherrypicking the valid items)\n\n### Render & UI\n\n- [ ] While rendering, manipulating images (like reload, resizing generated) can still crash Blender. Better image \"owner\" locking is required for 2.5 (#28477)\n- [ ] While rendering, several UI changes can cause crashes (#24758)\n- [ ] Rendering can't really use normal scene data if ui is interactive without causing all kinds of trouble, for examples and related issues see the item above this one.. and the reports (#26535)\n- [ ] Renaming scene leads to missing rendered result in image editor and compositor (see (#30222))\n- [ ] Rendering with File Output nodes could produce unwanted output images, see (#38605)\n- [ ] Shaded draw mode (using render code for vertex colors) is most certainly not coming back\n- [ ] Color Management can be communicated better, like when sampling colors in Image Window it confuses (could indicate such). Color Management also wasn't added backward compatible, disabling it gives dark results.\n- [ ] Compositor \"file out\" node doesn't work for FSA, and as color management troubles too.\n- [ ] FSA render uses temp files with names preventing other Blender instances to render too. (#28314)\n- [ ] Texture nodes have been implemented by recursively (backwards) calling parent nodes, this gives issues for viewers, previews and threads in general.We will try to keep it stable, but the functionality for this is going to replaced with a better shader system in 2.6. (compositor and material nodes call nodes forward, sorted on dependency. Each node is only called once this way).\n- [ ] Render display in Image window, for a window previously showing UV editor, doesn't restore correctly. (#26828). Code for this is too much spread around and fuzzy, needs full revision.\n- [ ] Object transform is not possible during rendering #37019\n- [ ] Don't render and warn when a driver or pre-render script can't be executed, a texture is missing, physics are not baked. Needed e.g. when renderfarms have scripts blocked.\n\n### Node Textures\n- [ ] Grouping texture nodes affects on data range #45696\n\n### Baking\n- [ ] add option to smooth vertex colors as a post process after baking #35066.\n- [ ] Baking to a low-res textures might miss pixels which lies on the UV island boundary #37681\n- [ ] Vertex baking might cause AO artifacts #44178\n- [ ] Support dupligroups #43553\n- [ ] Modifier produces non deterministic effects #42015\n- [ ] Update Blender side baking to support different topologies needed for adaptive subdivision #49405", "Shadow buffers\nIn order to reduce Shader register pressure during shading phase, we need to transfer the shadow sampling of opaque surfaces to a fullscreen pass.\nWe would store 1 or 2 bits per lights (limiting number of shadowed lights per pixels to 32 or 16) and use noise to dither the shadow shape.\nThe noise would disappear with accumulation.\nThis idea was explored in this presentation [link](view).\nThe issue is that this technique will not work for alpha-blended surfaces that will still need the standard shadow code.\nContact shadow and cascade shadows cost could be reduced. This could be the opportunity to remove support for contact shadows for alpha blend surfaces.\nThis also will permit to have raytraced shadows in the future.\n\nEstimate : 2 weeks\n", "Crash with animated displacement modifier in Texture paint mode while playing\nAMD Ryzen 5 2600, 16GB ram\nWindows 7 64bit\nNvidia GTX 1060 6gb (driver version 416.34)\n\n2.79b\n\nI intend to paint a heightmap interactively. So i keyframed a displacement modifier (2 frames) on a subdivided plane and assigned a texture.\nPlaying the timeline should now refresh the modifier every frame. \nPainting in the UV-editor this way works fine, but switching to texture paint mode and painting on the mesh directly crashes blender after about 0.5seconds.\n[HeightmapPaint.blend](HeightmapPaint.blend)\nI attached an example file.\nHit play and in the uv editor paint on the noise texture. This works okay.\nNow switch to Texture painting in the 3d viewport. The moment a stroke is applied to the mesh it should crash (it does consistently for me). ", "Light Nodetree support\nWhen rendering shadowmaps, we could run a postprocess shader (using the depth texture to retreive position) to output a Mono or Poly-chromatic intensity texture.\n\nThis would enable the use of custom light shaders and IES light profiles.\n\nThis will have to use another sampler since the shadow texture use another format.\n\nEstimate : 2 weeks", "Eevee: Occasional render glitches when using attribute node\nOperating system: Linux, specifically KUbuntu with X11\nGraphics card: RTX 3080 (Driver Version: 530.30.02, CUDA Version: 12.1)\n\nBroken: 2.93 - 4.0.0 Alpha, 2023-06-29, f47eed749eaf\nWorked: 2.92 (as available on Steam)\n\nUsing an attribute in an Eevee material causes occasional but severe render glitches. This does not happen consistently on any particular frame. It does not happen when rendering from the command line with `--background`.\n\nTo reproduce, render an animation with the attached Blender file. One in every 50 frames or so are corrupted. Occasionally it gets most of the way through the animation before ever producing a bad frame.\n\nThis file is basically the startup file with two changes:\n1. An attribute node with the name set to \"Color\" is plugged into the base color on the default material.\n2. The sample count is lowered to one. (The issue persists without this, but with a higher sample count, some samples glitch and others do not. The final image is a mix of the glitched and regular samples.)\n\nI have ensured that this is not a problem with my particular Blender settings by deleting everything in `~/.config/blender/4.0/config` and selecting \"Save New Settings\" in the splash screen before opening and rendering the file in the 4.0.0 alpha.\n\nNormal frame:\n\n\nCorrupted frame (as rendered in 2.93, some other versions simply display a transparent hole where the cube was without affecting the background):\n\n\n", "Intel GPUs: 3.6.2 crash on initializing Eevee rendered view\nOperating system: Linux-6.5.3-300.fc39.x86_64-x86_64-with-glibc2.38 64 Bits, WAYLAND UI\nGraphics card: Mesa Intel(R) Arc(tm) A770 Graphics (DG2) Intel 4.6 (Core Profile) Mesa 23.2.0-rc3\n\nBroken: version: 3.6.2\nWorked: 6 months ago :)\n\nBlender crashes when you opened the attached project and switch to rendered view.\n\n", "\"Illegal address in CUDA queue\" error occurs when I use both Principled BSDF and Adaptive Subdivision\nOperating system: Windows 11 Pro 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\n\n-**\"Illegal address in CUDA queue\"** error occurs when I activate both Principled BSDF and Adaptive Subdivision at the same time.\n-It would work fine if I don't use both at the same time.\n-Here are the possibilities of the error occurring when I use different settings:\n1. Principled BSDF + Adaptive Subdivision = **error**\n2. Texture Coordinate surface + Adaptive Subdivision = **no error**\n3. Principled BSDF + NO Adaptive Subdivision = **no error**\n\n-I can use a high poly count without Adaptive Subdivision *(e.g. Level viewport = 7)*, works fine. But as soon as I turn it on *(even with low poly count, e.g. Dicing scale = 1)* combined with Principled BSDF, the error occurs.\n\nIn the file I sent, I've already activated Principled BSDF + Adaptive Subdivision, you need to turn on the \"Rendered Viewport Shading\" and then slide the Scale parameter of the Noise Texture very quickly and the error should occur.\n" ]

[ "Missing sss_blur pass in EEVEE crashing indirect light bake\nOperating system: Win10\nGraphics card: Nvidia 980gx\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen) master\n\nLoad the following file and click \"bake indirect lighting.\" It will crash in eevee_subsurface.c:EEVEE_subsurface_add_pass. \"psl->sss_blur_ps\" is NULL. If you add a simple \"if (!psl->sss_blur_ps) return;\" it won't crash; interestingly enough, the compiler optimizes out the entire function. Anyway, that probably isn't the right fix.\n\n[cycles room test.blend](cycles_room_test.blend)\n\nSee above." ]

[ "Bone rotation results in different roll (editmode vs. posemode + apply as restpose vs. setting tail position). \nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.65\n\nBroken: version: 3.0.1\nWorked: unknown, possibly never.\n\nBone rotation in editmode using the transform system introduces roll.\n\n[#95560.blend](T95560.blend)\n- Open .blend\n- a single bone in editmode (with its tail at 0/1/1\n- set its tail to 1/0/1 in the sidebar (bone roll remains at zero)\n- undo\n- rotate on global Z axis -90 degrees (head/tail are now exactly like above example)\n- bone roll is now at 26 degrees\n- undo\n- enter posemode\n- rotate on global Z axis -90 degrees\n- `Pose` > `Apply` > `Apply Pose As Rest Pose`\n- check in editmode again (head/tail are now exactly like above example): bone roll is now at -45 degrees\n\nSo this might be expected behavior (not sure it is very predictable, there is some bone roll correction going on?), but\n- it is undocumented\n- if this is the desired behavior, shouldnt above methods result in the same bone roll?\n\n\n**Original Report**\n\nRotating a bone around the 3d cursor in edit mode, using Global Z axis, introduces roll into the bone (IFF the head and tail of the bone are not perfectly in line). This behaviour is unintuitive and does not occur if the head and tail of the bone are perfectly in line with the Z axis. See Armature 1 in attached image. Performing the same action in pose mode has different behaviour, in line with expectations.\n\nOpen the attached blend file. Select one of the two example bones (Armature 4 will produce the error, Armature 5 will not. Enter edit mode, select the bone body in edit mode. Hit R -> Z -> 90 (rotate it around the Z axis 90 degrees). Observe the roll of the bone. \nFor comparison, perform the same steps in pose mode.\n\n\n\n\n[Bone rotation confusion example.blend](Bone_rotation_confusion_example.blend)\n\n", "Connected child bone's nominal location values affected when transformed about 3D cursor\nBroken: 2.79, 2.90.1\nWorked: N/a\n\nAs expected, when moving a single connected child bone, its location is not affected, only the rotation.\nBut it is still possible to affect them accidentally if transforming about the 3D cursor.\nThis is a minor bug, but it's a definite inconsistency.\n\n - Open attached file.\n - Move/Rotate/Scale the child bone about the 3D cursor.\nNote the bone's nominal location values in the side panel.\n\n[Connected Bone Pose Bug.blend](Connected_Bone_Pose_Bug.blend)", "Toggling edit mode messes with active attribute\nOperating system: Linux-5.17.15-76051715-generic-x86_64-with-glibc2.35 64 Bits, X11 UI\nGraphics card: Quadro RTX 6000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.48.07\n\nBroken: version: 3.6.0 Alpha\n\nToggling edit mode does not properly preserve the active attribute.\n\n\n\nFrom default startup:\n- go to attribute panel\n- select UVMap as active\n- go to edit mode\n- nothing is active\n\n", "Bone Override Transform affects bone local axis display.\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.3.0\n\nWhen a bone gets assigned a display transform override, it's viewport display axis gets transformed, which probably shouldn't happen as its purpose is to display the Bone's not it's Shape's local axes.\n\n[override-transform.blend](override-transform.blend)\n - Remove Override Transform from Bone's Viewport Display's Custom Shape -> Displayed Bone's axes change location.", "Object rotation increases render time more than 10 times.\nTested with windows 7 64bit and windows 10 64bit.\nCPU ryzen 5 1600\nGPU GTX 780\nRAM 8GB GDDR4\n(bug appears on other computers too)\n\ntested with 2.79 and 2.8\nWorked: (optional)\n\nIf I rotate object for about 45 degrees around Z axis render time increases more than 10 times.\nBut in preview mode there is no difference in render time before and after rotation.[test.blend](test.blend)\njust rotate the object on attached blend file around Z axis about 45 degrees and compare render times before and after.\n10seconds for 2000 samples before rotation and about 10 minutes after.\nBased on a (as simple as possible) attached .blend file with minimum amount of steps\n", "Blender 2.8. In edit mode when you have hard scene (slow down processing) all your actions will applyes not at right moment (not at click but at moment when blender finn. caclucations)\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: GeForce GTX 560/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\n\nBroken: version: 2.80 (sub 53)\nWorked: (optional)\n\n\nI have scene where I put a lot boolean(and not only) modificators on one objects and when I enter into edit mode all processing erally slow (each action: movement, extrusion, etc..) And problem here in next: when you doing any action (and it took time to process it) and then press mouse button to apply action affect, but! it will not memorize position of the mouse when you pressed, and still continue doing (move to the cursor) And this really anoing because you need repeat all twice or more time. Or just seating and waiting with no mouse movement at all, when it will apply action effect. \nSo core problem is that: in edit mode any affect will applyes not at moment when you decide but by moment when CPU will finished calculetion of movement/action and will change your action to position where is going be mouse positioned at the moment of the end calculation... something like that. \n\nCreate some worcload/heavy scene. For example by a lot subdivisions and booleans at object. When each your action will be like \"under water gravity\" slowed down. And then try to change(extrude/move vertext / anything) boolean part that will be used as cut area in the edit mode. \n\n[2019-04-01 17-39-35.mov](2019-04-01_17-39-35.mov)", "Impossible to move with G key an object scale to 0 and parent to an empty\nOperating system: Windows 7\nGraphics card: GTX 780\n\nBroken: 3.5.1\nWorked: \n\nWhen an object is parented to an empty, and this object is scale to 0, you cannot move it anymore with the G key. But you can move it in the transform menu location.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n1. Open blend.\n2. Try to move object with G key.\n3. It is possible with Transfom menu.\n\nThanks a lot !\ndono", "Context override of bpy.ops.object.mode_set does not work\nOperating system: Ubuntu 18.04.5 LTS\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 440.100\n\nBroken: version: 2.93.0 Beta, branch: master, commit date: 2021-05-04 23:07, hash: 96abe5ebbc55\n\nSetting the object whose edit mode you want to enter via context overrides does not work.\n\nAlthough the active object is changed via the context override, the following line still switches into edit mode of the object selected in the UI:\n\n\n\n```\nbpy.ops.object.mode_set({\"active_object\": bpy.context.scene.objects[\"Cube\"]}, mode='EDIT')\n```\n\nI also tested:\n```\nbpy.ops.object.mode_set({\n\"active_object\": bpy.context.scene.objects[\"Cube\"]\n\"object\": bpy.context.scene.objects[\"Cube\"]\n\"selected_objects\": [bpy.context.scene.objects[\"Cube\"]]\n\"selected_editable_objects\": [bpy.context.scene.objects[\"Cube\"]]\n}, mode='EDIT')\n```\n\nAnd \n\n\n```\ncontext = bpy.context.copy()\ncontext[\"active_object\"] = bpy.context.scene.objects[\"Cube\"]\ncontext[\"object\"] = bpy.context.scene.objects[\"Cube\"]\ncontext[\"selected_objects\"] = [bpy.context.scene.objects[\"Cube\"]]\ncontext[\"selected_editable_objects\"] = [bpy.context.scene.objects[\"Cube\"]]\nbpy.ops.object.mode_set(context, mode='EDIT')\n```\n\nBoth do not work.\n\nThe only thing that works is:\n\n\n```\nbpy.ops.object.select_all(action='DESELECT')\nbpy.context.view_layer.objects.active = bpy.context.scene.objects[\"Cube\"]\nbpy.context.scene.objects[\"Cube\"].select_set(True)\nbpy.ops.object.mode_set(mode='EDIT')\n```\n\nBut it would be great to be able to do that simply via context overrides, as it also can be done for other operators.\n\n- Open Startup Scene\n- Duplicate Cube\n- Select Cube.001\n- Enter \"bpy.ops.object.mode_set({\"active_object\": bpy.context.scene.objects[\"Cube\"]}, mode='EDIT')\" in console\n- > The edit mode of Cube.001 is activated, however it should be the one of Cube", "Vertex Transformation, Number Rounding Problem, versions 2.79 and above!\nOperating system: Windows 7 x64 Ultimate SP1\nGraphics card: onboard Intel HD 4600\n\nBroken:\n2.79 (latest) AND 2.80 latest beta\n\nWorked: (optional)\n2.77a works correctly!\n\n\nThis problem is happening on blender versions 2.79 and 2.80! It is NOT present in 2.77a.\nHow to replicate: Selecting a vertex and clicking on its axis transformation on the right pane (N), for example on the Y-axis, enter the number 47.15426 for example.\nSo far, so good. But after clicking on that same field again (to edit or just copy the value to the clipboard), it gets automatically rounded off (just by clicking on it) to 47.1543 for no reason.\nBasically, when I want to select the coordinate of that vertex (47.15426 is its correct value) and just copy it to my clipboard so I can paste it into another program, that number gets changed automatically to 47.1543 inside Blender and it stays that way after de-selecting the Y-axis number field in the Transformation tab!\n\nIn 2.77a version, no such thing can happen, I have been using it since 2016 and there weren't any problems with that.\nI have tested both the 32-bit and 64-bit versions of the affected 2.79 and 2.80 releases, all have this issue.\n\nOpen the attached .blend file: [vertex_coordinate_rounding_problem.blend](vertex_coordinate_rounding_problem.blend)\nThe vertex is already selected. Just navigate with your mouse to the Y: axis transformation and click on the field with the number. It will change from 47.15426 to 47.1543 (which literally lowers the precision of the float by a decimal point). If you try that in Blender 2.77a, this will not happen.", "Snapping: with cursor as pivot and orientation blender still use object`s local orientation for rotation.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: version: 3.0.0 Alpha\n\nAs title says. I expect, blender should use cursor`s orientation for «align rotation to target» and allow to rotate object in any position without changing its local orientation.\n\n[2021-07-11_23-03-49.mp4](2021-07-11_23-03-49.mp4)\n\n[untitled.blend](untitled.blend)\n\n\n", "Bug when working with the Child Of constraint and the Selection to Cursor command\nBroken: version: 3.0.0 Alpha\nWorked: Never\n\nWhen working with the Child Of constraint the `Select To Cursor` command does not work correctly.\n\n- Open attached file\n- In the 3d View press {key Shift S} and `Select To Cursor`\n[Bug_ChildOf_Selection2Cursor.blend](Bug_ChildOf_Selection2Cursor.blend)\n\n---\n**Original Report:**\nWhen working with the Child Of constraint, it is impossible to use it several times by enabling and disabling the influence of the constraint. \nHe second time the offset is different and the child object jumps away from the desired position. \nIt's hard to put it in the right place. I tried to use a 3D cursor for this, but the Select To Cursor command does not work correctly when working with this the Child Of constraints.", "pressing z key to switch between extrusion \n intel i5 ,win ,amd radeon \nblender 2.8 \n490a385c8124\n\n\n\n\nThis happens when transform orientation set to Global \nCreate a sphere , select a face and extrude \nBy default it will extrude along the normal , when pressing z it will now extrude along the global but it won't go back to normal when pressing the 'z'key again \n \n\n\n ", "UV map inconsistency after bevel modifier\nOperating system: win10\nGraphics card: 1050ti\n\nBroken: 3.4.1\n\nMesh position in edit mode affects how bevel processes UV map\n\nIn attached file translate selected part of mesh along world Z axis -> UV is changing position\n[bevel_uv_issue.blend](bevel_uv_issue.blend)", "Rotation/Move support for Transform Cage\nThe idea is have a Transform Cage to move, rotate and scale in the 3D viewport (similar to the gizmo that exists for 2D UV editing).\nThis is will be realy useful for Grease Pencil and 2D animation workflow and is common in other 2D animation programs.\n\nPlease, check this video for more info:\n[Tramsform Cage2.mp4](Tramsform_Cage2.mp4)\n\n\nDaniel\n\n----\n\n\n****is this just the current scale cage with rotation handles added?***\nYes, basically. It would be nice if you grab an edge of the cage you have a skew transform (see the video)\n\n[2020-02-13 01-53-56.mp4](2020-02-13_01-53-56.mp4)\n\n***Would the pivot point be used only for rotation or scale as well?***Yes, for all tranformations***Would it be possible not to use this pivot point, and use the 3D cursor instead?***It´s is possible but I think that is more convenient and easy for the user***How this is accessed (As a sub-tool of transform? like scale cage?)***\nYes, as a sub-tool of transform\n\n", "Incorrect gizmo orientation \nOperating system: MacOS Monterey v12.3.1\nGraphics card: AMD Radeon Pro 5300M 4 GB Intel UHD Graphics 630 1536 MB\n\nBroken: 3.1.2, cc66d1020c3b, master, 2022-03-31\nWorked: ?\n\nWhen changing orientation of gizmo, previous (incorrect) orientation is used with select tool.\n\n1. New General project\n2. Rotate object/cube so local orientation is different to global (say, rotate 45 degrees on x-axis).\n3. In Gizmo dropdown, under Object Gizmos section:\n - Activate 'Move' checkbox\n - Select Local orientation\n4. Activate Select tool on the toolbar - the gizmo will appear using Local orientation (expected)\n5. Select Move tool on toolbar\n4. Change gizmo orientation to Global\n11. Click Select tool on toolbar\n\nObserve the gizmo orientation is Local, while it should be Global" ]

[ "Weird scaling with Cursor location (optionally orientation)\nOperating system:\nGraphics card: Intel\n\nversion: 2.80 (sub 74)\n\nThe scaling on major axis is pretty bizarre given that the object's scaling does not scale from where it is supposed to be, which is the cursor in this case. \n\n\n\n\nOpen the attached filea and scale the torus in Y, using the Gizmo. See how the scaling behaves. \n\nThe pivot and the orientation is set to the Cursor. However orientation type does not make a difference in the scaling issue.\n\nAlso checkout the attached video\n\n[weird_scale.blend](weird_scale.blend)\n\n[3lGjFYIjNC.mp4](3lGjFYIjNC.mp4)" ]

[ "Smooth shading usabilty\nWhen adding a mesh primitive in Blender, it is by default flat-shaded without subdivision surfaces. This means that after adding a new object with a smooth surface (like a sphere or monkey mesh), the user almost always has to change some settings.\n\nWe could change the default mesh settings to avoid that. However there are two distinct use cases which require different settings.\n\n* Meshes with autosmooth and/or custom normals. This could be either a low poly model for games, or a high poly model imported from CAD software.\n* Subdivision surfaces with creases to define sharp areas of the model. Custom normals are lost in the subdivision process, and autosmooth can't be combined with GPU acceleration.\n\nOnce subdivision surface settings become part of the mesh (#68891 (Subdivision surface settings part of the Mesh)), we could handle these as more distinct cases in the user interface and implementation.\n* If a mesh is a subdivision surface, no autosmooth or custom normals would be available.\n* Subdivision surface mesh primitives could be in a separate submenu of the Add menu, with all faces smooth shaded by default.\n* Mesh primitives could have smooth shaded faces and autosmooth enabled by default when it makes sense.\n\nEnabling autosmooth by default will negatively impact performance, particularly in edit mode. It's not obvious this is something we want to do. Further optimizations in this area are possible, but there will always be a cost.", "EEVEE baking\n", "Cycles freeze on render and texture baking (with packed textures)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.72\n\nBroken: version: 2.92.0\nWorked: 2.83\n\nThis broke somewhere between b21ba5e579 and f2b5f731d5\nAssume 99436acde8\n\nBlender freezes on render\nBlender also freezes on texture bake and rendered viewport\nTried on 3 different PCs, all the same\n\nIf textures are unpacked, rendering is fine.\n\nSimply open attached .blend file and hit F12\n\n[render_test2b_delmesh3_simplified.blend](render_test2b_delmesh3_simplified.blend)", "Prepare custom data API for attribute sharing.\nThe custom data api in `BKE_customdata.h` has multiple methods to get a specific layer. For example `CustomData_get_layer`. The main issue is that these functions are not const correct. They always return a non-const `void *` and have no idea whether the caller wants to modify the data or not.\n\nThat is problematic when we want to use attribute sharing, because we have to make a copy of shared attributes when the caller wants to modify the attribute. \n\nThe solution is to split every `CustomData_get_*` method in two: \n* A `*_for_read` method that returns a const pointer, that the caller must never change.\n* A `*_for_write` method that returns a non-const pointer and internally makes a copy of the data if necessary.\n\nWith this refactor, the `CustomData_duplicate_referenced_layer_*` functions become redundant, because the `*_for_write` methods will make the copy already.\n\nUnfortunately, the \"convenience pointers\" like `Mesh.mvert` make this system quite annoying and error prone:\n* The convenience pointers are not const. On a const-correctness level that implies that everyone with a non-const pointer to a mesh can also modify the vertex positions of that mesh. This is not true, because the array might be shared.\n* A first step in the right direction would be to make these pointers const. However, even with that, there is the issue that one would have to pass the mesh into `*_for_write` methods, so that the pointer can be updated if necessary.\n* A possibly better alternative would be to remove these convenience pointers completely and to provide a simple convenience api instead. E.g. there could be a `MVert *BKE_mesh_mverts_for_write(Mesh *)` and a `const MVert *BKE_mesh_mvert_for_read(const Mesh *)` method. This refactor is a bit ugly because the direct data access through convenience pointers is very common, but it might be worth it regardless. ", "Mantaflow simulation stops and crashes at the end of the bake (or when closing blender): The CG solver diverged, residual norm > 1e30, stopping.\nBroken: version: 3.0.0 Alpha\nWorked: -\n\nCrash baking fluid simulation.\n\n\nOpen the file and bake all its frames (or only frame 181 to 183).\n\n[beach.blend](beach.blend)\n\n```\nTraceback (most recent call last):\n File \"<string>\", line 1, in <module>\n File \"<string>\", line 169, in bake_fluid_data_5\n File \"<string>\", line 164, in bake_fluid_process_data_5\n File \"<string>\", line 136, in liquid_adaptive_step_5\n File \"<string>\", line 221, in liquid_step_5\nRuntimeError: GridCg::iterate: The CG solver diverged, residual norm > 1e30, stopping.\nError raised in extern/mantaflow/preprocessed/conjugategrad.cpp:562\n(...)\nterminate called after throwing an instance of 'Manta::Error'\n what(): can't clean grid cache, some grids are still in use\nError raised in extern/mantaflow/preprocessed/fluidsolver.cpp:33\nAborted (core dumped)\n```\n\nComplete output: [P2568](P2568.txt)", "Apply base results in oversized mesh using Catmull-Clark.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.50\n\nBroken: version: 2.83 (sub 13)\nWorked: Never\n\n\n\n[#75463.blend](T75463.blend)\n\n- Open file\n- Examine how subdivided object `original` looks like\n- Change number of subdivisions in Preview to 0\n- Click on Apply Base\n# Compare to form of subdivided mesh\n\nModified base mesh is slightly oversized.", "Sculpt: Paint tool cavity custom curve crash\nOperating system: Windows 10\nGraphics card: NVDIA RTX 2060 Super\n\nBroken: 3.4, 3.5\n\nSteps to reproduce the issue:\n - Switch to Sculpt Mode or open default sculpting template file\n - Choose the Paint Tool\n - Click on Brush dropdown\n - Enable Cavity by clicking on Advanced >> Cavity in the \"Brush\" dropdown\n # Enable Custom Curve under Advanced >> Cavity in the \"Brush\" dropdown\n\n\n\nResult:\nNo **Custom Curve** is shown as in other Sculpting Tools & after trying to paint with the **Paint** tool setup this way blender instantly crashes.", "Material-based displacement on source geometry fails to bake\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.91.0 Beta\nWorked: ?\n\nWhile baking in Selected to Active mode, applying a displacement shader to source geometry fails to actually displace the geometry, resulting in an incorrect bake.\n\n1. Open attached .blend file.\n2. Bake the texture.\n3. Compare to viewport render of the scene.\n[displacement_bake.blend](displacement_bake.blend)\n\n", "Normal Baking Inconsistence\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 526.98\n\nBroken: version: 3.5.0 Alpha\nWorked: unknown\n\nNormal baking result is different from original.\nAs below, a texture is baked to be used as normal map. Using \"Normal\" bake type the result is inconsistent with the original normal. A baking result using \"Emit\" type is used for comparison.\n\nDiffuse shader :\n[Correct (Diffuse).mp4](Correct__Diffuse_.mp4)\n[Incorrect (Diffuse).mp4](Incorrect__Diffuse_.mp4)\n\"Emit\" bake type has identical result to the original one, while \"Normal\" bake type has a little dimmer color.\n\nGlossy shader :\n[Correct (Glossy).mp4](Correct__Glossy_.mp4)\n[Incorrect (Glossy).mp4](Incorrect__Glossy_.mp4)\nDue to image resolution, the result of \"Emit\" type has acceptable differences at the edges of the squares. The result of \"Normal\" type has areas that shouldn't be brighter/dimmer (check from multiple view angles in the file below).\n\n[Bug Report.7z](Bug_Report.7z)\n\n\n - This is the node setup to generate the texture : \nThese nodes simulate the kind of texture that will connect into \"Normal Map\" node so that it can also be baked using \"Emit\" type.\n\n\n\n - Use the \"Material Output\" nodes to view the differences : \n\n\n\n - If re-bake, use 16-bit color depth can help to get the similar result from original nodes.", "Eevee 'Principled Normal' connected to 'Normal Map' node is disabling 'Displacement'\nOperating system: Win 10\nGraphics card: 2x1080Ti\n\nBroken: blender-2.91.2-windows64, blender-2.93.0-420f538fadfd-windows64\nNever worked as long as I remember\n\nBasically connecting normal map to principled shader is disabling any displacement map connected to the output. Just unplugging normal from principled turning it on. I don't know if its a bug or is it intentional \n\nNormal and displacement plugged in - displacement not working\n\nNormal unplugged - displacement working\n\n\n- Open attached file\n- Compare the result in Eevee with Cycles\n[bump_test2.blend](bump_test2.blend)", "Bmesh module does not support smooth shading / split normals\nBroken: 2.75 RC2\n\nBMLoop.calc_normal() returns normal for flat-shading, regardless of applied shading.\n\nBased on a (as simple as possible) attached .blend file with minimum amount of steps\n\nRun the following script, both APIs should give the same normals, but Bmesh module returns the same for all 3 objects.\n\n```\nimport bpy\nimport bmesh\n\nscene = bpy.context.scene\nfor ob in scene.objects:\n ob.select = True\nbpy.ops.object.delete()\n\nbpy.ops.mesh.primitive_cube_add(location=(0,3,0))\ncube1 = bpy.context.object\n\nbpy.ops.mesh.primitive_cube_add(location=(0,0,0))\nbpy.ops.object.shade_smooth()\ncube2 = bpy.context.object\n\nbpy.ops.mesh.primitive_cube_add(location=(0,-3,0))\nbpy.ops.object.shade_smooth()\nbpy.ops.mesh.customdata_custom_splitnormals_add()\ncube3 = bpy.context.object\n\n# no difference\ncube1.data.use_auto_smooth = True\ncube2.data.use_auto_smooth = True\ncube3.data.use_auto_smooth = True\n\ncube1.data.calc_normals_split()\ncube2.data.calc_normals_split()\ncube3.data.calc_normals_split()\n\nbm1 = bmesh.new()\nbm1.from_mesh(cube1.data)\nbm1.faces.ensure_lookup_table()\n\nbm2 = bmesh.new()\nbm2.from_mesh(cube2.data)\nbm2.faces.ensure_lookup_table()\n\nbm3 = bmesh.new()\nbm3.from_mesh(cube3.data)\nbm3.faces.ensure_lookup_table()\nbm3.normal_update()\n\nprint(\"Standard API:\")\nprint(cube1.data.loops[0].normal, \"Flat\")\nprint(cube2.data.loops[0].normal, \"Smooth\")\nprint(cube3.data.loops[0].normal, \"Smooth\")\nprint(\"---\")\nprint(\"Bmesh module:\")\nprint(bm1.faces[0].loops[0].calc_normal(), \"Flat\")\nprint(bm2.faces[0].loops[0].calc_normal(), \"Smooth\")\nprint(bm3.faces[0].loops[0].calc_normal(), \"Smooth\")\nprint(\"------\")\n```\n\nNote: Standard API could be used instead, but there's no RNA method for triangulation (only modifier and operator). Mixing Bmesh module and standard API seems unsafe (and there's no direct access to loops in bm). Object.to_mesh() could use support for triangulation, hence quad_mathod / ngon_method to improve the situation.", "NLA: Support for Merge Strips\n**Preface:**\nI know this isn't the place for feature requests. However I plan on implementing this myself soon. This is more of a heads up and acts a way to get feedback before or while I'm working on it. Maybe I'm overlooking something or maybe the idea can be developed further. People can discuss problems and improvements here instead of within the patch comments itself. That way the important discussion can remain here for easier reference.\n\nEdit: Diff: [D8867: Feature: NLA Merge Strips](D8867)\n\n**Problem:**\nFor the animation layer workflow, the only way to combine strips into a single strip is to use the Bake Operator. However, it will always bake all active NLA strips to a single full replace strip. Effectively, this is a \"Merge All\" operator. There is no way to merge a selected set of strips. \n", "Use after Free (according to asan) while baking.\nOperating system: Linux-5.4.0-90-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.82.00\n\nBroken: version: 3.1.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `rBUnknown`\nWorked:Don't know\n\nWhen pressing 'bake' in the attached (simple) blend file blender crashes when compiled with address sanitizer. Strangely only when in the UV-editor workspace, and not in the shading workspace.\n\n\n- build blender in debugmode with developer options\n- open the attached .blend\n- press bake\n\n\n[printuvs.blend](printuvs.blend)\n\nOr else from default startup:\n- build blender in debugmode with developer options\n- add a noise texture + colorramp to the base color input of the default material\n- add an image texture (sized 128x128) to the material and make it active\n- switch to uv editor workspace\n- switch renderer to cycles\n- open bake tab, choose diffuse bake\n- press bake\n\nASAN output:\n```\n\n==30632==ERROR: AddressSanitizer: heap-use-after-free on address 0x613000566c90 at pc 0x000006232e48 bp 0x7fff79d57550 sp 0x7fff79d57540\nREAD of size 4 at 0x613000566c90 thread T0\n - 0 0x6232e47 in paint_object_is_rendered_transparent /home/martijn/progs/blender-git/blender/source/blender/draw/engines/overlay/overlay_paint.c:52\n - 1 0x62333f5 in OVERLAY_paint_init /home/martijn/progs/blender-git/blender/source/blender/draw/engines/overlay/overlay_paint.c:75\n - 2 0x61e20f1 in OVERLAY_engine_init /home/martijn/progs/blender-git/blender/source/blender/draw/engines/overlay/overlay_engine.c:138\n - 3 0x60336a7 in drw_engines_init /home/martijn/progs/blender-git/blender/source/blender/draw/intern/draw_manager.c:1037\n - 4 0x60394be in DRW_draw_render_loop_ex /home/martijn/progs/blender-git/blender/source/blender/draw/intern/draw_manager.c:1641\n - 5 0x6038abd in DRW_draw_view /home/martijn/progs/blender-git/blender/source/blender/draw/intern/draw_manager.c:1571\n - 6 0xadf0436 in view3d_draw_view /home/martijn/progs/blender-git/blender/source/blender/editors/space_view3d/view3d_draw.c:1573\n - 7 0xadf05bf in view3d_main_region_draw /home/martijn/progs/blender-git/blender/source/blender/editors/space_view3d/view3d_draw.c:1595\n - 8 0x7f2353c in ED_region_do_draw /home/martijn/progs/blender-git/blender/source/blender/editors/screen/area.c:564\n - 9 0x53d6339 in wm_draw_window_offscreen /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_draw.c:730\n - 10 0x53d7569 in wm_draw_window /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_draw.c:880\n - 11 0x53d89d8 in wm_draw_update /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_draw.c:1081\n - 12 0x53c5120 in WM_main /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm.c:658\n - 13 0x34c4239 in main /home/martijn/progs/blender-git/blender/source/creator/creator.c:558\n - 14 0x7f41af1ae0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x270b2)\n - 15 0x34c35bd in _start (/mnt/fast/martijn/progs/blender-git/build_linux_debug/bin/blender+0x34c35bd)\n\n0x613000566c90 is located 208 bytes inside of 360-byte region [0x613000566bc0,0x613000566d28)\nfreed by thread T20 here:\n - 0 0x7f41af9ef1c7 in __interceptor_free ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:123\n - 1 0x1bb10832 in MEM_lockfree_freeN /home/martijn/progs/blender-git/blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:130\n - 2 0x5f86383 in blender::deg::IDNode::destroy() /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/node/deg_node_id.cc:143\n - 3 0x5e21387 in clear_id_nodes_conditional<blender::deg::Depsgraph::clear_id_nodes()::<lambda(ID_Type)> > /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph.cc:162\n - 4 0x5e1e4c0 in blender::deg::Depsgraph::clear_id_nodes() /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph.cc:173\n - 5 0x5e1cdf8 in blender::deg::Depsgraph::~Depsgraph() /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph.cc:89\n - 6 0x5e20080 in DEG_graph_free /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph.cc:325\n - 7 0x9ca04e1 in bake /home/martijn/progs/blender-git/blender/source/blender/editors/object/object_bake_api.c:1608\n - 8 0x9ca3d40 in bake_startjob /home/martijn/progs/blender-git/blender/source/blender/editors/object/object_bake_api.c:1775\n - 9 0x5461946 in do_job_thread /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:393\n - 10 0x1bae287e in tslot_thread_start /home/martijn/progs/blender-git/blender/source/blender/blenlib/intern/threads.cc:231\n - 11 0x7f41af920608 in start_thread /build/glibc-YbNSs7/glibc-2.31/nptl/pthread_create.c:477\n\npreviously allocated by thread T20 here:\n - 0 0x7f41af9ef6f7 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:154\n - 1 0x1bb10f20 in MEM_lockfree_callocN /home/martijn/progs/blender-git/blender/intern/guardedalloc/intern/mallocn_lockfree_impl.c:236\n - 2 0x370cccf in BKE_libblock_alloc_notest /home/martijn/progs/blender-git/blender/source/blender/blenkernel/intern/lib_id.c:1095\n - 3 0x5f84eff in blender::deg::IDNode::init_copy_on_write(ID*) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/node/deg_node_id.cc:114\n - 4 0x5e1e02c in blender::deg::Depsgraph::add_id_node(ID*, ID*) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph.cc:127\n - 5 0x5fa14e5 in blender::deg::DepsgraphNodeBuilder::add_id_node(ID*) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:178\n - 6 0x5fcd141 in blender::deg::DepsgraphNodeBuilder::build_material(Material*) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:1797\n - 7 0x5fcdee5 in blender::deg::DepsgraphNodeBuilder::build_materials(Material**, int) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:1819\n - 8 0x5fb47b3 in blender::deg::DepsgraphNodeBuilder::build_object_data(Object*, bool) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:941\n - 9 0x5fb0010 in blender::deg::DepsgraphNodeBuilder::build_object(int, Object*, blender::deg::eDepsNode_LinkedState_Type, bool) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:787\n - 10 0x601516d in blender::deg::DepsgraphNodeBuilder::build_view_layer(Scene*, ViewLayer*, blender::deg::eDepsNode_LinkedState_Type) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/deg_builder_nodes_view_layer.cc:117\n - 11 0x5f19f8e in blender::deg::ViewLayerBuilderPipeline::build_nodes(blender::deg::DepsgraphNodeBuilder&) /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/pipeline_view_layer.cc:35\n - 12 0x5f06808 in blender::deg::AbstractBuilderPipeline::build_step_nodes() /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/pipeline.cc:72\n - 13 0x5f058af in blender::deg::AbstractBuilderPipeline::build() /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/builder/pipeline.cc:51\n - 14 0x5e2ff1d in DEG_graph_build_from_view_layer /home/martijn/progs/blender-git/blender/source/blender/depsgraph/intern/depsgraph_build.cc:259\n - 15 0x9c9c19e in bake /home/martijn/progs/blender-git/blender/source/blender/editors/object/object_bake_api.c:1236\n - 16 0x9ca3d40 in bake_startjob /home/martijn/progs/blender-git/blender/source/blender/editors/object/object_bake_api.c:1775\n - 17 0x5461946 in do_job_thread /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:393\n - 18 0x1bae287e in tslot_thread_start /home/martijn/progs/blender-git/blender/source/blender/blenlib/intern/threads.cc:231\n - 19 0x7f41af920608 in start_thread /build/glibc-YbNSs7/glibc-2.31/nptl/pthread_create.c:477\n\nThread T20 created by T0 here:\n - 0 0x7f41af996a65 in __interceptor_pthread_create ../../../../src/libsanitizer/asan/asan_interceptors.cpp:216\n - 1 0x1bae2aa8 in BLI_threadpool_insert /home/martijn/progs/blender-git/blender/source/blender/blenlib/intern/threads.cc:245\n - 2 0x546296a in WM_jobs_start /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_jobs.c:488\n - 3 0x9ca69f8 in bake_invoke /home/martijn/progs/blender-git/blender/source/blender/editors/object/object_bake_api.c:1938\n - 4 0x53e6d36 in wm_operator_invoke /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1347\n - 5 0x53e8c9e in wm_operator_call_internal /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1554\n - 6 0x53e8ef9 in WM_operator_name_call_ptr /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1602\n - 7 0x53ea323 in WM_operator_name_call_ptr_with_depends_on_cursor /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:1790\n - 8 0xa4d665e in ui_apply_but_funcs_after /home/martijn/progs/blender-git/blender/source/blender/editors/interface/interface_handlers.c:1030\n - 9 0xa565bfd in ui_handler_region_menu /home/martijn/progs/blender-git/blender/source/blender/editors/interface/interface_handlers.c:11435\n - 10 0x53e0497 in wm_handler_ui_call /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:690\n - 11 0x53f88dc in wm_handlers_do_intern /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:3052\n - 12 0x53f99d3 in wm_handlers_do /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:3166\n - 13 0x53ffb0e in wm_event_do_handlers /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm_event_system.c:3656\n - 14 0x53c5108 in WM_main /home/martijn/progs/blender-git/blender/source/blender/windowmanager/intern/wm.c:652\n - 15 0x34c4239 in main /home/martijn/progs/blender-git/blender/source/creator/creator.c:558\n #16 0x7f41af1ae0b2 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x270b2)\n\nSUMMARY: AddressSanitizer: heap-use-after-free /home/martijn/progs/blender-git/blender/source/blender/draw/engines/overlay/overlay_paint.c:52 in paint_object_is_rendered_transparent\nShadow bytes around the buggy address:\n 0x0c26800a4d40: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4d50: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4d60: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4d70: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd\n 0x0c26800a4d80: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n=>0x0c26800a4d90: fd fd[fd]fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4da0: fd fd fd fd fd fa fa fa fa fa fa fa fa fa fa fa\n 0x0c26800a4db0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4dc0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4dd0: fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd fd\n 0x0c26800a4de0: fa fa fa fa fa fa fa fa fd fd fd fd fd fd fd fd\nShadow byte legend (one shadow byte represents 8 application bytes):\n Addressable: 00\n Partially addressable: 01 02 03 04 05 06 07 \n Heap left redzone: fa\n Freed heap region: fd\n Stack left redzone: f1\n Stack mid redzone: f2\n Stack right redzone: f3\n Stack after return: f5\n Stack use after scope: f8\n Global redzone: f9\n Global init order: f6\n Poisoned by user: f7\n Container overflow: fc\n Array cookie: ac\n Intra object redzone: bb\n ASan internal: fe\n Left alloca redzone: ca\n Right alloca redzone: cb\n Shadow gap: cc\n==30632==ABORTING\n```\n\n\n", "The result of Dynamic paint Image sequence is brighter than intended when using initial color.\nOperating system:Windows 10 64bit\nGraphics card: GTX 1060\n\nBroken: 2.79b, 2.80\nWorked: \n\nthe result of baked Image is birghter than it displays on properties panel. \nNot sure but seems like it is gamma correcting the color but the input color is already in sRGB space, not in linear space!\n\n\nset any color that is not too bright or dark to \nDynamic paint - Canvas - Image Sequence - Initial Color : color.\nthen press \"Bake Image sequence\".\nThe color of baked Image will be brighter than the color you set.\n[dynamic paint color.blend](dynamic_paint_color.blend)", "Cycles strands always use rounded normals even with custom Normal input to BSDF\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: Intel(R) Arc(TM) A770 Graphics Intel 4.5.0 - Build 31.0.101.4090\n\nThis behavior has been observed on nVidia cards as well, on both Linux and Windows, as early as Blender 3.0. It is not an Intel Arc issue like most of my bugs :)\n\nBroken: version: 3.4.1\n\n\n*TL:DR*\n\n**Expected:**Using a custom Normal input in a BSDF for hair strands (such as \"True Normal\" from the Geometry Info, or overwriting with another source like an attribute or texture) should allow complete control over the rendered hair strand in both Eevee and Cycles.\n\n**Actual:** Only Eevee renders fur as desired. Cycles appears to force some sort of rounded/inverted cylinder normal at some point after the BSDF, leading to shading artifacts.\n\n\n\n\nNote the line down the middle of the strand/curves on the Cycles version. This is usually at the point where the (default) normal is facing fully away from the light source.\n\nThis happens on both the legacy particle system as well as the newer Curve based hair. It happens on both Rounded Ribbons and 3D Curve primitives.\n\n**Longer version**\n\nI want to render stylized fur. The approach I use is to copy the normals from the underlying mesh to the particle system or hair curves, and disable shadows. This makes a very smooth, toon-like shader.\n\n\n\n\nThis is based off of my old addon, encode-normals but normals passed as attributes from geometry nodes have the same issue, as do normals set right in the material itself, so it's not an issue with my addon or node tree.\n\nIn the example blender file, I have two planes facing upward. One has a legacy particle system and the other has a new hair curve. Both sets of strands/curves have a material where the normal is forced to be (0, 0, 1), or straight up, to show the base case. \n\nIn Eevee I get the toony flat shading, where each strand is basically a solid color. \n\nIn Cycles, I end up with part of the strand in shadow, and if you look closely the emulated rounded normals appear to be an \"inside out\" cylinder, so the line tends to show up where the light direction and normal of the fake \"cylinder\" are perpendicular (i.e., where light is grazing the outermost part of the cylinder)\n\nI can't turn this rounded normal off or override it. It's coming from something outside the BSDF. This completely breaks my stylized shader in Cycles and is an unintuitive behavior. Normals can be overriden on a material in any other situation/primitive that comes to mind.\n\nSteps I've tried:\n1. Setting normals from encoded vertex colors (addon)\n2. Setting normals from geometry nodes\n3. Setting normals outright in the shader (this example)\n4. Disabling shadow ray visibility\n5. Making the strand material transparent to shadows in shader\n\nOpen provided .blend and check both curve systems for rendering differences between Eevee and Cycles. \n\n[TestParticles1.blend](TestParticles1.blend)" ]

[ "Custom vertex normals ignored when baking normals from cycles\nWindows 10, Geforce 1070\n\nBroken: 2.79.4 2018-03-28 e3F1d98 \n\n\nCustom vertex normals of an object are ignored when baking normal maps, tangent space, from cycles.\nWhen baking with Blender Internal, everything works as it should.\n\n\nOpen test scene, press bake. See result. Switch engine to Cycles, and repeat. Compare results. \n\n[customvertexnormals_bake_test.blend](customvertexnormals_bake_test.blend)\n", "baking normal map using cycles does not support custom vertex normals on the baking target mesh\nWindows 10\n\nblender-2.79.0 2bf3825\n\n**Bug Description:**\n\nWhen baking normal maps from a high poly mesh to a low poly mesh using cycles, if the low poly mesh has customized vertex normals, cycles will not use them. \n\nI uploaded a test file here which contains a simple highpoly (using a material with the new blend node) and lowpoly with custom normals (created using the Y.A.V.N.E. plugin):\n\ncycles_baketest.blend\n\nUsing cycles for baking I consistently get incorrect normals baked down.\n\nCustom normals are very often used in game development to get higher quality bakes and more flexible normal maps. Getting this feature to work correctly in Blender, used together with the new bevel node will make Blender a very powerful tool for producing PBR ready game art. \nIf anyone has time to take a look at this, I'd be very grateful! Hit me up if you need more test cases to work with and I'll gladly set them up.\n\n" ]

[ "Sculpt undo bug, mixing sculpt brush and multi-res modifier changes.\nWith current master (also 2.83).\n\nThe following steps show and error when redoing sculpt mode undo, when mixing multi-res subdivisions.\n\n- Startup file\n- Add multi-res modifier.\n- Enter sculpt mode.\n- *Repeat this 3-5 times.*\n - Multi-res subdivide.\n - Brush Stroke.\n\n- Now undo until the very beginning.\n- Now redo until the very end.\n\nWhen redoing, there are glitches where the multi-res has tearing/artifacts.\n\nSee video: [sculpt_undo_bug.mp4](sculpt_undo_bug.mp4)", "The tools panel suddenly disappears in engraving mode.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Quadro K620/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 425.31\n\n\n\nBroken: version: 2.93.0\n\n\n\n\nIn the engraving mode, use the blank key to call the tools panel may suddenly disappear.\n\nMy native language text↓↓↓↓\n\n在雕刻模式，使用空白鍵呼叫　tools 面板會有突然消失的可能。\n\n\n\n\n\n1. Set KeyMap>Spacebar Action>Tools\n2. Select any object and enter sculpt mode.\n3. Press the blank key to call the tools panel. At this time, the panel will suddenly disappear. (Error generation)\n4. Move the mouse arbitrarily\n5. Press the space button again to call the tools panel. At this time, the panel will suddenly disappear. (Error occurs again)\n\nOther additional tests:\n\n1. The panel disappears (when an error occurs), do not move the mouse, press the blank key again, and it will appear normally\n\n2. When it appears normally, as long as you move the mouse and press the blank key, an error will occur, but there is one exception, which is simply moving the mouse horizontally so that the panel appears in the second position, the selected brush and slide The overlapping position of the mouse is exactly the same as the last time, the error will not occur (image)\n\n3. In the object mode, everything is normal, I think it should be operated like the object mode.\n\nMy native language text↓↓↓↓\n\n1.設定 KeyMap >Spacebar Action>Tools\n2.選擇任意物體，進入sculpt模式。\n3.按空白鍵呼叫tools面板。這時候面板會突然消失。(錯誤產生)\n4.任意移動滑鼠\n5.再次按空白鍵呼叫tools面板。這時候面板會突然消失。(再次錯誤產生)\n\n其他額外測試 :\n\n1.面板消失(錯誤產生時)，滑鼠不要動，再按一次空白鍵，就可以正常出現\n\n2.當正常出現後，只要有移動滑鼠，再按空白鍵就會發生錯誤，但有一個時候例外，就是單純水平移動滑鼠，讓第二個位置出現面板時，選取的筆刷與滑鼠重疊的位置和上一次一模一樣，錯誤就不會發生 (圖片)\n\n3.在物體模式，一切正常，我認為應該操作起來要像物體模式這樣。\n\n\n\n", "\"Radius of the brush\" field missing in box trim and lasso trim sculpt tools.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.91.0\nWorked: Never (tool was introduced in 2.91.0)\n\n\"Radius of the brush\" field missing in box trim and lasso trim sculpt tools after checking use cursor location and radius for trimming shape. The tool uses the brush radius for the depth but the field is not visible.\n\nUse Trim Lasso and Box Trim tools in sculpt mode.\n\n", "Black artifacts on smooth surface in 2.79 64bit daily build\nOperating system: Windows 10\nGraphics card: RTX 2080Ti\n\nBroken: blender-2.79.0-git.d7cecc2ba3f6-windows64\nWorked: 2.79b\n\n\nThe software renders black artifacts on a smooth surface. This happens both in the rendered preview and in the final render. Artifacts disappear when smooth is not applied.\n\n\n\n\n\n\n\nJust go in rendered preview mode in the editor, or just F12 to see the artifacts in the render.\n\n[bug report.7z](bug_report.7z)", "Brush Management: setting a brushes tool [vertex_tool, weight_tool, ...] may leave you with unaccesible tools (when no corresponding brush is left)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GT 630M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 2.83.2\n\nSee attached gif. Scene is the default scene without camera and light. Once I start switching tools in Weight Paint Mode through the different methods shown in the gif, I eventually reach a point where I lose access to some tools. Notice that even though the brush menus initially show the expected brush as the one currently selected, it isn't the one performing the painting (shown in the gif as the desync between \"Active Tool\" and the brush that appears as selected). As the gif also shows, the side menu doesn't select the expected brush either. \n\nAs a side question: why is the Smear brush named \"Smear.001\"? instead of just \"Smear\" like the other brushes?\n\nSee gif. \n\nThanks.\n\n\n\n\n", "Small Custom Bone Shapes have some Wireframes Artifact.\nWindows 10 64 Bits.\nGraphics card GeForce GTX 1080 Ti \n\nBroken: version: 2.82.6, hash: a854840e76ae\nWorked: (optional)\n\nSmall custom bone shape has wireframe glitching, the Scale is based on real world but if you resize it in/out u can see it happen.\n\n{[F8257584](Shape.gif) size = full}\n\n\n\n[Bug.blend](Bug.blend)\n", "Freestyle Lines Stopping Short of Vertices & Partially Failing\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) HD Graphics 620 Intel 4.5.0 - Build 27.20.100.8682\n\nBroken: version: 2.92.0\n\nSee attached PNG file: Freestyle lines sometimes do not reach vertices - sometimes by a long way; and on some edges they are completely absent when they should be present.\n\n[The PNG file also shows a shadow problem which has been submitted separately: T88013 ]\n\nOpen the attached .blend file and press F12 ('Render').\n\n[BUG_FreestlyeLinesStoppingShort-of-Vertices-01.blend](BUG_FreestlyeLinesStoppingShort-of-Vertices-01.blend)\n\n\n\n", "Sharp normals along base mesh wires in multires sculpting with smooth shading\nOperating system: Linux-5.4.0-47-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: Quadro GP100/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 440.100\n\nBroken: version: 2.91.0 Alpha\nWorked: As far as I know in 2.79 before the inclusion of OpenSubdiv\n\nThis has been a long standing issue in multires sculpting with shading artefacts in smooth shading.\nThese will only start appearing when sculpting on the surface and may flicker between different spots when sculpting directly over them.\n\n- Create a suzanne head\n- Add a multires modifier with 4 subdivisions\n- Change the object to smooth shading\n- Enter sculpt mode and smooth on the surface of the object\n\nOn these spots the sharp shading should consistently appear among others. A high contrast Matcap can help with spotting them.\nIn this example they are very subtle but they can be extremely noticeable on very curved surfaces.\n\n\n\n\n\n", "Texture on active object is not visible in solid view with sculpt mode activated\nOperating system: Windows 10, Ubuntu 16.04\nGraphics card: GeForce GTX 760\n\nBroken: 2.80\n\nTexture is not visible in solid view with sculpt mode activated\n\nCreate a texture, apply it to an object, like start cube, make sure you're in solid view mode with textured color active, you'll see the texture in object mode. After switching to sculpt mode, object becomes gray.\nIf you change view mode to LookDev or Material, everything is fine.\n\n**Extra info**\nIf you go back to Solid view and add a Shape Key, texture is visible again, but mask brush doesn't work. The problem with mask brush and shape keys is already mentioned #67303\nI made a video about those problems Mk2w41C87AI", "Sculpting: Ctrl+S reverts recent progress. \nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.92.0,\nBroken too: b67423f806\n\nWhen saving with ctrl+S the last changes applied to sculpt with \"grab\" brush gets reverted. This should be observed by reopening the file. \n\n- Have a sculpt with dynamic topology. Disable dynamic topology for the next step.\n- Move a part of the mesh with the grab brush.\n- Save your last progress. \n- Reopen the file using file browser. All changes done with the grab brush will be undone.\n\n\n\n[face sculpt broken.blend](face_sculpt_broken.blend)\n\nThanks, have a great day and stay healthy!", "MultiRes+Solidify while trying to use Sculpt Mode cause Blender to crash\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.30\n\nBroken: version: 2.81 (sub 12)\nWorked: (optional)\n\nI tried deactivating one modifier while keeping the other one active to see if Blender would crash, but this was not the case. Only when I used both simultaneously did the program reliably crash every time, regardless of brush.\n\nAdd the modifiers that are active in the video and follow along. \n[Desktop 2019.09.23 - 00.49.39.02.mp4](Desktop_2019.09.23_-_00.49.39.02.mp4)\n", "Black Artifacts While Using Adaptive Subdivision\n{[F7718097](bug.png)}Operating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.15\n\n\nBroken: version: 2.80 (sub 75)\n\n\nBlack artifacts on edges of geometry while using adaptive subdivision. I was trying to duplicate my workflow from other 3d software. Extruding geometry from 2d curves creates extra edges on the model. And extra edges create weird shadows on displacement.\n\nImport curve from dxf file. Use extrude Object Data > Geometry. Fill mode: Both. Use subdivision and displacement.\n\n[1.blend](1.blend)\n", "Edges, generated with Triangulate modifier looks differend and do not disappear with hiding.\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GT 640/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.90.0 Alpha\n2.83 release\n\n\n[2020-06-18_15-14-00.mp4](2020-06-18_15-14-00.mp4)\n[untitled.blend](untitled.blend)\n\nDefault cube with Triangulate modifier, wireframe mode, edit mode.\nSelect all, hide. You will have this.\n\n\nThere is no way to part of mesh with generated edges and hide another part.\nAlso weird shading for polygons.\n\n\n", "Grease pencil stroke partially invisible until mouse-up\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.30\n\nBroken: version: 2.93.0 Alpha\n\nWhile drawing in the attached scene the stroke is only partially visible. The full line becomes visible only on mouse-up\n\n* open the attached .blend\n* draw over the area with the 3d geometry behind it, only part of the line will show until you end the stroke (see video below)\n\nScreen recording of the problem: Jehz9YAXmE0\n\nBlend file: [gpdrawbug.blend](gpdrawbug.blend)\n\n", "Grease Pencil - A corrupt Sculpt > Push brush is created when undoing a mode change\nOperating system: Darwin-19.5.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-14.0.32 355.11.11.10.10.143\n\nBroken: version: 2.90.0 Alpha\n\nA corrupt Push brush is created when I undo a Draw -> Sculpt mode change.\n\n* File > New > 2D animation\n* Change from draw to sculpt mode\n\n* Edit > Undo (the mode changes back to draw mode)\n* Change again to sculpt mode as before\n* A new corrupt Push brush has appeared called F Push (instead of F Push Stroke), and is selected. It has no thumbnail. It is missing the usual UI elements of the Push brush, like strength. See screenshot.\n \n\n" ]

[ "Sculpt mode: cloth brush breaking up mesh\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 550 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 376.54\n\nBroken: version: 2.83 (sub 9)\nWorked: Never\n\nSculpt mode: cloth brush breaking up mesh\n\n\n- Load factory settings\n- Go to sculpt mode\n- Ctrl+r 3 times\n- Get the cloth brush \n- Give some strokes on the default Cube.\n\n [cloth_brush.blend](cloth_brush.blend)\nWith this blend file sculpt some on the Cube\n" ]

[ "Automatic moving of the timeline when the current frame reach the region border\n\nAnimation and keyframes are visible through multiple spaces : nla, graph, sequencer, timeline, etc...\nWhen it comes complex animation with a lot of keyframes / clips / actions, we often zoom in the timeline to better edit these things. But when we start the playback, we can not see what's happening when the current frame is offscreen.\n\nIn my opinion, it is inconvenient for some uses, like video editing, keyframing.\n\n### Solutions\n\n------------------------\n\n**Fixed cursor**\nInstead of having the time cursor moving, it could be the timeline that moves :\n[Screencast_2021-10-30_16_42_15-converted.mp4](Screencast_2021-10-30_16_42_15-converted.mp4)\n**Floating cursor inside the area**\nThe time cursor is moving, but it can not go outside the view, so the timeline moves to keep the cursor in view :\n[Screencast_2021-10-30_16_38_32-converted.mp4](Screencast_2021-10-30_16_38_32-converted.mp4)\n\n### Issues\n\n------------------------\n\nThe first solution is maybe a bit to limiting for viewing around the current frame. I prefer the second approach.\n\n\nThe main problem with the 2 solution is the performance of redrawing the area for each frame.\nDo it really impact the performances or is it minor ?\n\nI have a patch about some changes, if it is something interesting to look at.\n", "\"Reset to Default Value\" sets alpha incorrectly in [Material Properties > Viewport Display > Color]\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.09\n\nBroken: version: 3.5.0\n\nRight-clicking \"Reset to Default Value\" on [Material Properties > Viewport Display > Color] sets both the color's \"value\" and \"alpha\" bars to 0.8, when the alpha one specifically should reset to 1. Alpha resets incorrectly.\n\nRight-click \"Reset to Default Value\" on [Material Properties > Viewport Display > Color] on any material.\n\n", "Displacement Animation works only with keyframed node present\nOperating system: Linux-5.4.0-107-generic-x86_64-with-glibc2.27 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.60.02\n\nBroken: version: 3.1.2\nBroken: version: 3.2.0 Alpha\nWorked: unkown\n\nDisplacement Animation works only with a keyframed node present, even if it is not connected!\nI was trying to animate displacement using the object color, finding that it doesn't work.\nWhen trying a keyframed mapping node instead I found that it works but even if the node is not connected!\n\n - Open the attached blender file, switch to render mode and play the animation\n - The displacement will be animated.\n - Now delete the not connected mapping node or clear its keyframes.\n - The displacement will not be animated anymore. Even though I would expect it to be animated by the object color.\n\n[displace-test-1-1.blend](displace-test-1-1.blend) \n\n[displace-test-1-1.mp4](displace-test-1-1.mp4)\n", "Keyframing off Animated property of rigid body on second frame of rigid body simulation\nOS: Ubuntu 13.10\nLinux kernel: 3.11.0-18-generic\n\nBroken: 2.70-b64bdb2-linux-glibc211-x86_64\n\n\nNormally, keyframing off the Animated property of a rigid body will cause rigid body to maintain its velocity and rotation that it has at the last animated keyframe. However, if the Animated property is keyframed off on the second keyframe of the rigid body simulation, the object will continue the animation from a stand-still, i.e. it loses its initial velocity / rotation.\n\n\n\nCreate bug:\n\n - Open new Blender file\n - Keyframe location of the default cube\n - Go to frame 20, move cube along the y-axis by 20 units and keyframe location.\n - Enable Rigid Body for the cube\n - Go to frame 10, enable Animated (Physics context -> Rigid Body panel) and keyframe it\n - Go to frame 11, disable Animated and keyframe it\n - Go to Scene context -> Rigid Body Cache panel, and set Start to 10 (and say, End to 50)\n - Go to 3D View and hit ALT A to start animation. Notice that when the cube reaches frame 10, it loses its y-velocity and just falls straight down.\n\nResolution:\n\n - Set the start of Rigid Body Cache to 9. \n - Update cache by unlinking the RigidBodyWorld group from Rigid Body World panel, then choose it again.\n # Play animation again, and it's fine.\n\nAlthough, I have figured out how to get it working, I think this is unexpected behaviour (which took me a long time to figure out).\n\nSteps 1-7 have already been completed in this file: [rigidBody_bug.blend](rigidBody_bug.blend)\n", "\"Disable in Viewport\" option results in UI for \"Disable in Render\" not updating properly when it is keyframed\nOperating system: Linux-5.10.0-8-amd64-x86_64-with-glibc2.31 64 Bits\nGraphics card: NA\n\nBroken versions: 2.82-3.4\nWorked: Unsure\n\n**Short description of error:**\n\"Disable in Viewport\" option results in UI for \"Disable in Render\" not updating properly when it is keyframed.\n\n**Exact steps for others to reproduce the error:**\nFollow the steps shown in the video above or follow the steps below:\n1. Start with a new file.\n2. Select an object.\n3. In the `Object Properties` tab of the `Properties Editor`, expand the `Visibility` panel.\n4. Inside the `Visibility panel`, animate the the `Show in -> Renders` property. You can do this by inserting a keyframe for it being enabled on frame 1, then go to frame 2 and disable it and keyframe it again.\n5. Once the `Show in -> Renders` property is animated, disable `Show in -> Viewports`. Now if you scrub through your time line, you will notice that the UI for `Show in -> Renders` does not update the `Tick` as you go back to a frame you had keyframes to be enabled. This can also be observed in the `Outliner` as the `Camera` icon does not change from disabled to enabled or vice-versa.\n6. If you enable `Show in -> Viewports`, you will notice the `Tick` and `Camera` icon start updating again as you scrub through the time line.\nHere is a video demonstrating the issue:\n[Disable in viewport breaking disabled in render.mp4](Disable_in_viewport_breaking_disabled_in_render.mp4)", "Render cloth animation does not resets cache\nBroken: 2.82.7\nWorked: 2.79b\n\nWhen modifier Subdivision is before the cloth physics modifier,but with different subdivision settings for rendering and viewport, the cache remains active even when rendering.\nPossibly related to #74523.\n\n- Open attached file\n- Run animation (without baking)\n- Render animation\nNotice that the cloth is deformed with each rendered frame.\nThis is because rendering reuses the cache.\n\n[bug.blend](bug.blend)", "MNEE work only with 1st refractive shader node and ignore other inside one material\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 473.04\n\nBroken: version: 3.2.0 Beta\n\n\n**Short description**\nMNEE just ignores any refractive shader except the 1st one:\n\n\nFile:\n[test caustic.blend](test_caustic.blend)", "VSE autokeying transforms in preview dont respect keyingsets\nOperating system: Linux-5.13.0-0.rc6.45.fc35.x86_64-x86_64-with-glibc2.34.9000 64 Bits\nGraphics card: NVIDIA GeForce GTX 970M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 495.44\n\nBroken: version: 3.3.0 Alpha\nWorked: never since autokeying came in ffd1a7d8c8\n\nVSE autokeying transforms in preview dont respect keyingsets\n\n(this report was split from #98015 (VSE strip transforms do not animate correctly and persist after clearing the animation) since multiple issues came up in that report)\n\n[VSE_keyframing.blend](VSE_keyframing.blend)\n\n- open .blend\n- create a keyframes for Position X/Y on frame 1\n- for **Position X** only: {key RMB} Add To `Keying Set`\n- in Scene Properties, Keying Sets panel, verify only `transform.offset_x` is in the Keying Set\n- in the Timeline, Keying popover, verify that that Keying Set is used\n- turn ON autokeying\n- in the Timeline, autokeying popover, enable `Only Active Keying Set`\n- go to frame 10\n- translate the image in the `Preview` again and confirm the translation\n- notice keyframes are created for **Position X as well as Position Y** (only Position X should be keyed though)\n\n\n\n", "Cycles and Eevee Significant Difference in Texture Output\nOperating system: Ubuntu 20.04\nGraphics card: GTX 1060 3Gb\n\nBroken: 2.80 and on (3.0.0 is still affected)\nWorked: 2.79\n\nFrom Blender 2.80, Cycles produces incorrect output for some procedural textures, while Eevee produces correct ones.\n{[F12696991](Cycles.png)}\n\nTest file (I assume there might be other files affected too). [Carbon_90.blend](Carbon_90.blend)", "Eevee: image sequence data bloc displays different offset attributes when set from different locations (texture editor, shader editor, image locator)\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.5.0\n\nif I change the offset (or start frame) of the sequence in the shader editor, I would expect it to also change in the image editor or wherever I am using the same image sequence, otherwise it causes discrepancies and buggy viewport display (since we dont know which offset is displayed in the viewport). \n\nJust change the offset of 1 of the 3 displayed image sequence settings displayed... the other 2 won't follow.\n\nCycles does this fine btw.", "Eevee: Procedural material very slow\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.85\n\nBroken: version: 2.83.3 / 2.90.0 Beta\nWorked: -\n\nFor some reason in eevee my procedural asphalt material works pretty slow.\nI don't see any problems with Cycle and the same scene works as fine on a Radeon RX 480, so it seens to be a bug.\n\n- Open attached file\n- go to material view \n- move around the plane (it's getting worse if you get closer)\n\n[Eevee_RTX_material_bug.blend](Eevee_RTX_material_bug.blend)\n", "Update property when adding or removing keyframe\nRemoving keyframe can cause issues in all editors and adding keyframe needs to update VSE.\n\nWhen removing keyframe, RNA update function should be called, since property value may change. This is reproducible with simple 3 keyframe animation, place CFRA on middle keyframe, remove (press Shift + Alt +I on propery in side panel) it and no update will happen.\n\nWhen adding keyframe, sequencer prefetch is not notified about this change. So it works on outdated data until keyframe is added. Simply notifying prefetch is not enough and cache must be cleared too which is currently done from RNA update function.\nSimilar case is speed effect, where it calculates input frame based on animation, which is integrated over time. Array mapping current to input frames is cached and recalculated in RNA update function.", "Blender 2.80.* Video Editor: Timeline breaks upon time-remapping.\nOperating system: Linux Mint 19.1 Tessa 64-bit\nGraphics card: Radeon 7570M 1Gb\n\nBroken: 2.79\nWorked: -\n\nChanging the \"Time Remapping\" setting to\n**25 Old**\n**100 New**\nCauses issues with the timeline in the Video Editor window. It shows items as they were, without visually scaling the timeline, however the timeline itself looks unchanged. The seek bubble moves separately from the seek bar as well.\n\nWhen clicking where you'd expect frame 25 to be, it actually moves the cursor to the wrong location. See attached screenshot and .blend files below.\n\n\nOpen a video in the video editor, set the time remapping, observe the broken timeline.\n\n\n[issue.blend](issue.blend)", "EEVEE: Crash if computation is too heavy for the system (TDR issue)\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.64\n\n\nBroken: version: 2.80 (sub 72)\nWorked: (optional)\n\n\nAny time I try to render an image sequence in any format in Eevee, it constantly crashes the whole program on the first frame. I was able to render before until I changed some camera location animations and adding a focusing animation. As well as changing the camera sensor size. I am rendering in 4K I have attached the log file.\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[NebulaV2.blend](NebulaV2.blend)", "Eevee \"Shutter\" setting in motion blur render settings not animatable\nOperating system: Win 10\nGraphics card: RTX 3090\n\nBroken: 2.91.0\nWorked: never worked (as far as i know)\n\n\nEevee \"Shutter\" in motion blur render settings not animatable\nIn Cycles it ***is***animatable.N/A" ]

[ "EEVEE image sequence not updating in animation render\nSystem Information\nOperating system: ubuntu 18.04 lts 64 bit\nGraphics card: nvidia geforce gt 840 (in this video i turned it off and used CPU graphics because of power saving)\n\nBlender Version\nBroken: blender-2.80-72ede30d0f97-linux-glibc224-x86_64\n\n**Problem to render every frames in animation in eevee**\n\n**There is a problem that eevee doesnt render every frames and make mistakes. here is a VERY SHORT video to explain problem. Thankkss**\n[EEVEE bug.mp4](EEVEE_bug.mp4)" ]

[ "X-Ray mode shows flickering edges of planar mesh\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.40\n\nBroken: version: 2.80, 2.83.20, 2.93.9, 3.2.0 release, 3.3.0 Alpha\nWorked: never, from what I can see. Didn't test older than 2.80.\n\nX-Ray solid shading of a mesh shows flickering edges when rotating the camera. \n\nThis could have something to do with the mesh being planar.\n\n- Open the attached file.\n- Rotate the viewport camera with {key MMB}.\n- See that some edges flicker into and out of existence.\n\n[bugreport.blend](bugreport.blend)\n\n[Desktop 2022.06.25 - 14.38.23.03.mp4](Desktop_2022.06.25_-_14.38.23.03.mp4)", "Self-referencing driver works in viewport, but not in render.\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: GeForce GT 730/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional) 2.79b\n\nRotation property of object is updated in viewport, but not in render.\nIf run same script in 2.79 - all works fine both in viewport and render.\n\n```\ndef rotator():\n var1=bpy.data.objects[\"Suzanne\"].rotation_euler.z\n var3=bpy.data.scenes[\"Scene\"].render.fps\n\n new_value=var1+math.radians(var3)\n\n return new_value\n```\n\n\n[Script_bug_2.8.2.blend](Script_bug_2.8.2.blend)\n\n- run the file Script_bug_2.8.2 ('Run script', 'Update Dependecies')\n- play in the viewport - the animation works correctly\n- render (Ctrl + F12) - the animation does not work;\n\nViewport:\n[bug_video_render0001-0050.mkv](bug_video_render0001-0050.mkv)\n\nRender:\n[bug_video_render_20001-0050.mkv](bug_video_render_20001-0050.mkv)\n\n2.79 file:\n[Script_work_2.79.blend](Script_work_2.79.blend)\n\n", "Blender 2.83 smoke or liquid simulation not showing up when switching to Modular mode\nOperating system: Windows 10pro 64 bit\nGraphics card: Nvidia GTX 960\n\nBroken: 2.83, v2.83.0 \nWorked: 2.82\n\nWhen I'm simulating smoke or liquid. If I switch to modular mode, the bake the simulation. I do not see any smoke or liquid after it has run the bake. I can see the domain moving as if particles are filling the domain, but during play back nothing is there.\nIf I change the mode to replay mode. I can see the smoke or liquid as it is simulating. But I don't get this in modular mode\nI have a youtube video link that shows the problem. \n\nIoqbkzs7h5w\n\nI like modular mode, because baking the simulation this was takes less resources by not updating the view port.\n\n- Create a box.\n- Duplicate the box. \n- In physics tab... Make one for the emitter and the other as domain.\n- Scale domain as necessary.\n- By default Blender is in Replay mode. Change it to modular to get the Bake buttons. \n- Bake simulation. \n- Then change scale of emitter and bake simulation again.\nor\n- open attached file\n- Try see simulation with Replay (it should work)\n- Then try with Modular\n[Name_animationV2.blend](Name_animationV2.blend)", "Cycles freeze on render and texture baking (with packed textures)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.72\n\nBroken: version: 2.92.0\nWorked: 2.83\n\nThis broke somewhere between b21ba5e579 and f2b5f731d5\nAssume 99436acde8\n\nBlender freezes on render\nBlender also freezes on texture bake and rendered viewport\nTried on 3 different PCs, all the same\n\nIf textures are unpacked, rendering is fine.\n\nSimply open attached .blend file and hit F12\n\n[render_test2b_delmesh3_simplified.blend](render_test2b_delmesh3_simplified.blend)", "Compositor seam after rotation\nOperating system: windows\nGraphics card: gtx1050\n\nBroken: 2.93, 3.0, 4.0\nWorked: \n\nThere is a seam after the rotate node, which isnt present before it. Also If you set the rotation to 0 degree there will be no seam.\n\ncheck in the provided file: [compositor seam bug.blend](compositor_seam_bug.blend)", "Simulation Nodes: show bakes in outliner\n_No response_\n\n", "Burned Out Screen Space Reflections in Eevee Viewport - Nvidia GT 750M\nOperating system: macOS 10.14.3\nGraphics card: NVIDIA GeForce GT 750M 1024 MB\n\nBroken: 2.90.0\nWorked: 2.79\n\nIn some cases I detected that the Screen Space Reflections in Eevee Viewport gets out of control.\nA few seconds after switching to the viewport (even when viewport samples are set to 1) the issue kicks in and – as the sampling continues – an increasing percentage of the reflections get totally white.\n\n\n\nThe renders are correct.\nSurfaces affected by reflection planes are correct.\nHiding the chandeliers meshes solves the problem. Switch tab and back for this to work.\n\nI was able to strip down the scene and identify the source of the problem: Originally I set up the glass material with \"multiply\" blending. This was converted (when opened with belnder 2.82 alpha) to a \"Shader to RGB\" and a \"Transparent BSDF\" node. [1]\n\nThe \"Shader to RGB\" causes the problem in 2.82.\nThe \"multiply\" blending of the material causes it in 2.8.\n\nOpen the attached file.\nSwitch to LookDev or Rendered viewport.\nWait.\n\n[Chandelier.blend](Chandelier.blend)", "Bug while moving vertices with mirror & subdivision modifier\nOperating system: Windows 10 64-bit\nGraphics card: ATI Radeon HD 5670\n\nBroken: 2.82\nWorked: at least 2.8\n\nWhen I try to move a vertice it seems like the mirror modifier gets applied in an odd way. The result even changes when moving the camera or hide one of the images I use for reference. In fact it becomes distorted when hiding the images.\nThe following link is a recording of my screen:\nview?usp=sharing\nI encountered this bug only recently. The file itself is older and I do not remember which version I used (since I use it via Steam I did not pay attention to the automatic patches).\n\n1. Open this file [Bugged Project.blend](Bugged_Project.blend)\n2. Disable visibility of Subdivision modifier for Edit Mode (which should already cause the bug)\n3. Move a vertice\n\n[system-info.txt](system-info.txt)", "Cycles Ray Visibility trick for motion blur on objects with animated render-ability is no longer working\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 398.11\n\n\nBroken: version: 2.80 (sub 74)\nWorked: 2.79b, date: 2018-03-22 14:10, hash: `f4dc9f9d68b`\n\n\nCycles motion blur artifacting is a known issue with objects that become renderable while the shutter is open.\nA workaround is to transfer the Restrict Render animation to all the Cycles Ray Visibility options at the object level.\nThis solution worked for us in production on Next Gen using 2.78, but appears to no longer work in 2.80.\n\nThe attached file contains 4 planes linked to an animated camera. The 2 on the right are just static. The one on the upper left has animated Restrict Render to reveal the plane on frame 6. The plane on the lower left has Cycles Ray Visibility revealing it on frame 6.\nSince all planes are linked to the camera and have no transformation animation, none of them should show any motion blur at all.\n\nOpen the attached file in 2.79b and render frame 6 to see that the top left plane appears blurred (it shouldn't) and the lower left plane appears correct.\nOpen the attached file in 2.80 and render frame 6 to see that both left side planes are now showing the blur artifact.[rnd.motion_blur_bug.v0003.blend](rnd.motion_blur_bug.v0003.blend)\n\n", "Image Editor Blank After Switching Slots, Even If There Is A Render In That Slot\nOperating system: Windows 10 64 bit\nGraphics card: Nvidia GTX 750 Ti\n\nBroken: (2.80, 79f67acccb36, blender2.7, 2019-03-24)\n\nSwitching the slots in the image editor for the render result (hotkey \"j\") can show a blank image, even if there is an image available in that slot.\nSelecting the render layer in the dropdown menu brings it back.\nThis is irritating, especially since the image is shown right after rendering.\nThis only happens if the compositor is disabled.\n\n- Load [bug.blend](bug.blend).\n- Press F12 to render. The image editor comes up.\n- Press \"j\" to switch to switch the slot.\n- Press F12 to render again.\n- Press \"j\" to switch back to the previous slot. It now appears blank.", "Selected Objects render state (hidden in renders) causes baking of black images (also these are visible in renders after baking, setting is not as it was before)\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 497.09\n\nBroken: version: 3.1.0\nWorked: Cannot find any - tested back to 2.81\n\nIf rendering of the selected object is not enabled, then the baking will fail with a black image created.\nStrangely the rendering icon for this object will be automatically turned-on in the outliner - if you hover your mouse over.\nBaking will still fail - even though the icon now shows that rendering is enabled.\nTurning off render, then turning it back on again will fix the problem so that baking will now work.\n\nLoad the blend file attached (or use your own scene set-up for baking).\n[Bake selected to active has black image.blend](Bake_selected_to_active_has_black_image.blend)\nClick on Bake to confirm that the baking is successful.\n\nDisable render for the Selected object.\n\nClick on Bake again, and a black image will be created.\n\nNote that the Selected object still has rendering disabled.. Move your mouse into the Outliner above the render icon, and observe that rendering for the Selected object automatically becomes enabled.\n\nClick on Bake again and observe a black image is still created.\nDisable rendering for the Selected object, then enable it again.\nClick on Bake again and the baking will be successful.\n\n\nObservations:\n1) If rendering is disabled for the active object, then this error is shown.\n\n2) The bake duration is the same whether the rendering is enabled for the selected object or not. (Mine was taking a 2.5 minutes, and it took a long time to figure out what the problem was.)\n3) Saving the blend file and reloading fixes the problem - the state is persisted as render==enabled.", "Regression: broken normal shading, sharp edges appear on smooth surfaces.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: 2.93, version: 3.0.0 Alpha\nWorked: 2.83.9, 2,92\n\nCaused by 1c22b551d0\nTo fix this, 5c4d24e1fd got reverted\nProblem still persists though...\n\n\nIn newer versions normal mapping looks broken, sharp edges appear.\nEevee and Cycles have same bug.\n\nupdated: [untitled.blend](untitled.blend)\n\ndifference in shading\n**2.93 on top, 2.8 down below:**\n\n\nThis effect appears even with very small angle of clean surface (without normal maps)\n\n\n\n<video src=\"2021-05-18_16-17-05.mp4\" title=\"F10122604/2021-05-18_16-17-05.mp4\" controls></video>", "Render glitches on cycles when 'Motion Blur' and 'Use Limit Surface' are turned on. \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.92.0\nAlso broken in: 2.93.0 Alpha, branch: master, commit date: 2021-03-20 21:00, hash: 9ca67c47a152, type: Release\nbuild date: 2021-03-20, 23:00:48\nWorked: unknown. This file was created within 2.92\n\nWhen rendering this file I get (for me) random glitches in the render, when Motion Blur is active in Cycles.\nIn the attached file you can see this in for example frames 13,46,77 and 153 (amongst others) for the figure on the left and in frames 57,65 and 82 for the figure on the right.\nTurning off motion blur makes the glitches go away. It happens - for me - both with CPU only and with GPU renders.\n\nThe only objects - as far as I can tell - that produce these glitches are the two 'jasjes' with the material called 'overhemd'. (changing 'overhemd' to a different color also changes the glitches to that same color) Apart from the very obvious triangular glitches, you can also see some sort of wrinkly render errors near the ties on the shirt that only appear when the glitch is visible.\n\nFrame (13) with a glitch: \nFrame (14) without a glitch: \n\nThe objects started out as duplicates of each other.\n\n- Open attached file\n- Render frame 153 (or any of these: 13, 46, 77)\n\nI'm sorry the file is as large as it is, but I haven't had this happen in other renders/files yet, so I couldn't reproduce with another file.\nPlease let me know if you prefer to get Dropbox links instead of an actual upload.\n\n[Statler_en_Waldorf_animatie_013.blend](Statler_en_Waldorf_animatie_013.blend)", "\"Regression\" : Using normal or bump maps in cycles produces flat washed-out shading (caused by terminator softening)\nOperating system: Windows 10 x64\nGraphics card: GeForce GTX 1660 Dual OC\n\nBroken: 2.81, 2.82, 2.83 from 2020-03-06 22:46\nWorked: 2.80\n\nCaused by c3be14e151\n\nUsing normal or bump maps in cycles produces flat washed-out results comparing to EEVEE.\nThe bug also appears in the latest 2.83 build, it affects both Diffuse and Principled Shader but not Glossy.\nNormal Map and Bump nodes seem to produce identical results with the 2.80 version, so the problem lies with the Shader node itself.\n\n\n[bug_report.blend](bug_report.blend)\n1. Open file\n2. Switch to Rendered viewport shading", "FADE INNACTIVE GEOMETRY MODE doesn't work properly with COLLECTION INSTANCES\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.13587 Core Profile Context 20.4.2 26.20.15029.27017\n\nBroken: version: 2.91.0 Alpha\nBroken: version: 2.91.0 Alpha\n\nI really like this mode, but it feels likeit not completely done. It's not work with collection instances at all.\nHere is was my expectations and what it looks like currently: \n\nAnd even instances from other scenes ignore this mode too. Will be cool if it(instances) will work in stream ideology of this mode\n\n\nOpen this file [FADE INNACTIVE GEOMETRY DOESN'T WORK PROPERLY WITH INSTANCES.blend](FADE_INNACTIVE_GEOMETRY_DOESN_T_WORK_PROPERLY_WITH_INSTANCES.blend) and select (one by one) buttom figures (cone/torus/cube) and after selecting each one - press Tab.\nYou will see that instances ignores this mode\n\n" ]

[ "Eevee. Reflections with reflection plane broken (regression)\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GT 640/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.90.0 Alpha\nWorked: 2.83 release and b18c2a3c413b^ (caused by b18c2a3c41)\n\n[untitled.blend2](untitled.blend2)\n\nHere the comparison.\n[2020-06-09_12-04-11.mp4](2020-06-09_12-04-11.mp4)\n", "Eevee reflection plane glitching\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: AMD Radeon RX 580 2048SP ATI Technologies Inc. 4.5.13587 Core Profile Context 20.4.2 26.20.15029.27016\n\nBroken: 2.8 to 2.82 (sub 7)\nWorked: None\n\nI have a floor with a reflection plane on it, first renders were perfectly fine but as I do more renders, reflection planes begin misbehaving randomly, it starts out fine but gets progressively worse, here are two consecutive frames one with good reflections, the second glitched, with a perfect mirror for reference:\n\n||\n| -- | -- |\n\nIt only happens with reflection planes, not with plain SSR or cubemaps.\n\nWhat I have tried: All the SSR options, reflection plane settings, increasing samples, shadow size and soft shadow, different materials, AO and bloom.\n\nI created a new scene with only a perfect mirror and a sphere to reflect, same thing, it's not something to do with my scene:\n\nGIF:\n\n\nThe only thing that momentarily helps letting Blender \"rest\", then it gets worse over time (clean to messy frames ratio).\n\n1. Open attached file\n2. Go to render preview\n\n[Reflection plane.blend](Reflection_plane.blend) " ]

[ "Texture painting with mirroring and viewport clipping does not mirror\nOperating system: Linux-5.8.4-200.fc32.x86_64-x86_64-with-fedora-32-Thirty_Two 64 Bits\nGraphics card: GeForce GTX 970M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.66\n\nBroken: version: 2.91.0 Alpha\nWorked: never, 2.79 also did not do this\n\nTexture painting with mirroring and viewport clipping does not mirror\n\n[texturepaint_mirror_clipping.blend](texturepaint_mirror_clipping.blend)\n- Open .blend\n- Do a stroke on the right side --> updates on the left as well because x-mirroring is turned ON\n- (Undo)\n- {key Alt B} a clipping region in the viewport\n- Do a stroke again on the right side --> updates on the right side\n- {key Alt B} to clear the clipping region in the viewport, see no updates have taken place on the clipped left side\n\nExpected: stroke should update the clipped area as well [as we do for vertexpaint, weightpaint and sculpting], everything else would be inconsistent\n\n", "Display issue in model corner when loop cutting\nBroken: 2.79, 2.81\nWorked: -\n\nWhen creating and edge loop around de outside faces it creates in one of the cornes an extrange artifact\n\nCreate an edge loop across the outside faces of the model.\n\nBlender file:\n[loop cut.blend](loop_cut.blend)\n\nArtifact:\n\n", "Texture Mapping issues on large brush radius\nOperating system: Linux-6.4.6-gentoo-x86_64-AMD_Ryzen_Threadripper_1950X_16-Core_Processor-with-glibc2.37 64 Bits, X11 UI\nGraphics card: AMD Radeon RX 7900 XT (gfx1100, LLVM 15.0.7, DRM 3.52, 6.4.6-gentoo) AMD 4.6 (Core Profile) Mesa 23.3.0-devel (git-50c29e1ffa)\n\nBroken: version: 4.0.0 Alpha\nWorked: --\n\n\nWhen a large brush radius for textured brushes in sculpt mode, the outer-most affected geometry doesn't apply the texture fully. This is especially noticeble on the Clay Strips and Paint brushes when the stroke method is set to anchored.\n\n\n\n\n\nBlend file = [bug.blend](attachment)\n\n- Open the File\n- Use the provided paint brush and clay strips brush on a subdivided cube or suzanne head like shown\n- Notice how the outer edges of the falloff is not updating the geoemtry fully\n\n", "Grease Pencil mask error\nOperating system:windows10\nGraphics card:GTX1060ti\n\nBroken: 2.83&2.9\nWorked:greasepencil-Mask error\n1.When 2.83 or later, creates two objects, the masks collide, causing a strange phenomenon, and in 2.82 there are no errors.\n2.And creating two masks under the same object can cause problems too。\niJlqkf556sA\n[GP剪贴测试.blend](GP剪贴测试.blend)", "Black artifacts on smooth surface in 2.79 64bit daily build\nOperating system: Windows 10\nGraphics card: RTX 2080Ti\n\nBroken: blender-2.79.0-git.d7cecc2ba3f6-windows64\nWorked: 2.79b\n\n\nThe software renders black artifacts on a smooth surface. This happens both in the rendered preview and in the final render. Artifacts disappear when smooth is not applied.\n\n\n\n\n\n\n\nJust go in rendered preview mode in the editor, or just F12 to see the artifacts in the render.\n\n[bug report.7z](bug_report.7z)", "Texture paint Backface culling option does not work if symmetry is enabled\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11\n\nBroken: version: 3.6.1\nWorked: never since implemented in e2d60d180ed8 afaict\n\nWhile in the texture paint mode, the backface culling option of the brush is sometimes ignored, allowing painting onto backfaces. It only seems to affect certain objects. I could not discern why. See the attached blend file.\n\n1. Open attached blend file.\n2. Paint on the backfaces of the mesh; observe bug.\n3. Select cube\n4. Paint on the backfaces of the cube; observe no bug.\n\n", "Intersect Boolean does not work properly when mesh has a degenerate face\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.14760 Core Profile Context 20.45.37.01 27.20.14537.1001\n\nBroken: version: 2.91 - 3.0.0\nWorked: n/a\n\nThe `Intersect Boolean` Exact mode does not work correctly when the mesh has at least one degenerate face.\nIn this case, it was tested on a mesh with a triangle whose vertices are collinear\n\n- Open attached file - (The edited mesh has a degenerate triangle)\n- Go to `Face` -> `Intersect (Boolean)` - (The operation will not work because of that triangle)\n\n- Undo the operation\n- Move the triangle vertex a little so that the face is no longer degenerate.\n- Reselect that loose geometry ({key L})\n- Redo Boolean operation (now it works)\n[intersect_w_degenerated_tri.blend](intersect_w_degenerated_tri.blend)\n\n", "3D Texturing: When changing the texture resolution, the pixel data isn't recreated\n* Start painting\n* Change image resolution\n* Continue painting.\n\nWe have a safe guard for this, but it doesn't seem to work anymore.\n", "VSE: Separate images does not handle resolution correctly\nOperating system: Windows 11 Home 64 bit\nGraphics card: NVidia GeForce 2060 RTX SUPER\n\nBroken: 3.0.1\n\n- Add image sequence to VSE where each image has different resolution\n- Separate images to discrete strips\n\nCheck resolution of separated strips - they are all same, but they should not be.\n\n\n\n\n\n\n\n", "Line Art Thickness doesn't consider Render Resolution Percentage\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 30.0.101.1122\n\nBroken: version: 3.2.2\nBroken: version: 3.4.0 Alpha\nBroken\n\nWhen using Line Art Modifier, on a Grease Pencil Object in Screen Space Mode, Line Thickness doesn't scale with resolution as expected. \n\n1. Open new General Startup File\n2. Create New Object>Grease Pencil>Scene Line Art\n3. Under Object Data Properties>Strokes>Stroke Thickness and set to World Space\n4. Set Render Resolution % to a factor. line thickness will be different than expected\n\n\n\n\n\n", "Compositor: Viewport issues when Passepartout is off\nBroken: version: 4.0.0 Alpha\nWorked: never\n\nComposition Mask is \"rendered\" wrongly in the viewport when the camera `Passepartout` is off.\nNote that the Compositor is set to Camera.\n\n\n* Open this file: [zoomed_out.blend](attachment)\n* Notice that the passepartout is disabled in this camera, but the viewport compositor is set to Camera.\n\nCurrent buggy result:\n\n\nThe mask only matches when the camera has `Passepartout` on:\n\n\n\nRelated: #111344\n\n", "Sculpting: Draw Sharp brush has sharp point at the beginning of the line\n**Broken:** \nBlender version: 2.93.7, 3.0, 3.1 (27.12.2021)\n\n**Short Description**\nDraw Sharp brush line is uneven at the beginning of the line there is a sharp point. It can be seen with large radius of the brush.\n\n**Steps to reproduce:**\n1. Default scene\n2. Remesh cube, Voxel Size 0.01.\n3. Brush \"Draw Sharp\", Radius 200-500 px.\n\n\nTest File:\n[#94463.blend](T94463.blend)", "GPencil: Randomize brush's 'Position' attribute doesn't work when camera is too close to Grease Pencil object\nOperating system: Win10 Home 64\nGraphics card: rtx3070\n\nBlender Version\nBroken: 2.93, 3.1.0\nWorked: N/A\n\nExact steps to reproduce the error:\n\n- Open file\n- try to use Randomize brush\n- Strength and Thickness work, but not Position \n- Switch to the other camera that's farther away from the GP object, and randomizing the position works fine.\n\n[randomize_not_working_close_NEW.blend](randomize_not_working_close_NEW.blend)\n\n\nwatch?v=4pTrEm5HqVU", "alt brush doesn't work in image editor -> paint\ni think this should be fairly easy to reproduce so i wont bother sharing system info. i was able to reproduce this on 3.3.0 and 3.3.1.\n\nalt brush doesn't work in image editor -> paint\n\ni have a different colour on my brush but when i press ctrl and try to paint the primary colour is painted onto the canvas", "Line stroke method not working with Scene spacing\nOperating system: Linux-5.4.0-121-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce GTX 670MX/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.151\n\nBroken: version: 3.2.1\n\nLine stroke method not working with Scene spacing\n\nOpen blender, then open a new sculpt file, select any brush change the stroke method to line and chose scene spacing distance. \nIt only works when started and ended inside the object, but when coming from outside the object or end up outside the object it doesn't work, sometimes it gives an initial effect when it comes in contact with the object but only one sample.\nI the video you can see how it works with view spacing and then how behaves with scene spacing.\n\n[vokoscreen-2022-07-12_14-43-22.mp4](vokoscreen-2022-07-12_14-43-22.mp4)\n\n" ]

[ "Move brush with topology enabled does not work correctly\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 2.93.0\n\n\nMove brush with topology enabled on creates incorrect deformation if you start moving outside the model\n\n\n[Move_Brush_Topology_bug.mp4](Move_Brush_Topology_bug.mp4)\n\n[Move_Brush_Topology.blend](Move_Brush_Topology.blend)" ]

[ "Remove/Update \"Emulate Middle Mouse\" preference\nThis task is to check if it is reasonable to remove the emulate middle mouse option.\n\nBlender 2.7x didn't allow left-click-select and emulate middle mouse to be enabled at once, this never worked - it's just that in 2.8x we allow both options to be enabled at once with key-conflicts.\n\nSince I don't use a touch-pad on macOS / ms-windows, I'm not sure how common it is that middle-click gestures are enabled without having to manually configure it.\n\nAre there systems where using a middle-click still isn't practical?\n\nOptions could be:\n\n- Remove this preference.\n- Remove the keys that use Alt-LMB *(means re-working the keymap in some places)*.\n- Use an alternative shortcut that doesn't conflict, eg: OSKey (Cmd / Super / Windows-Key).\n\n *(By default this conflicts on Linux with many window-managers, these would need to be re-configured not to use Super-LMB).*\n\nUnless this is given some attention it's likely Alt-LMB will be used more, to the point Emulate-Middle-Mouse preference isn't useful and we'll need to remove or document all it's known limitations and accept Blender isn't fully usable with this enabled. \n\n----\n\nThere are currently conflicts using with emulate middle mouse\n\n- Number Button Alt-Drag to change all other selected items #69053 (Emulate 3 Buttons doesn't work with ALT Modifier key on UI buttons)\n- Edit-Mesh: Edge-ring de-select isn't possible.\n- Alt-MMB-drag to change views.\n- Un-clamped vertex slide can't be confirmed.\n- Grease Pencil, Circle Primitive\n- Weight Paint Gradient\n- Loop select *(Worked around with double-click, but not very convenient since selecting boundaries requires quad-click).*\n\nHere is a full list of potential conflicts in the default keymap.\n\nAre we serious about resolving these or would it be better to drop this option? \n\n----\n\nFrom a search of the code:\n\n```\ncat release/scripts/presets/keyconfig/keymap_data/blender_default.py | \\\n pcregrep \"(LEFTMOUSE|TWEAK_L|params.select_mouse|params.action_tweak|params.action_mouse).*alt\\\":\"\n```\n\n```\n (\"uv.select_loop\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"alt\": True},\n (\"uv.select_loop\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"shift\": True, \"alt\": True},\n (\"uv.select_edge_ring\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"ctrl\": True, \"alt\": True},\n (\"uv.select_edge_ring\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"ctrl\": True, \"shift\": True, \"alt\": True},\n (\"uv.select_loop\", {\"type\": params.select_mouse, \"value\": 'DOUBLE_CLICK', \"alt\": True},\n (\"mask.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True},\n (\"mask.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"graph.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"alt\": True},\n (\"graph.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"shift\": True, \"alt\": True},\n (\"graph.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"ctrl\": True, \"alt\": True},\n (\"graph.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"node.select_lasso\", {\"type\": 'EVT_TWEAK_L', \"value\": 'ANY', \"ctrl\": True, \"alt\": True},\n (\"node.select_lasso\", {\"type\": 'EVT_TWEAK_L', \"value\": 'ANY', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"node.backimage_sample\", {\"type\": params.action_mouse, \"value\": 'PRESS', \"alt\": True}, None),\n (\"node.move_detach_links_release\", {\"type\": params.action_tweak, \"value\": 'ANY', \"alt\": True}, None),\n (\"action.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"alt\": True},\n (\"action.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"shift\": True, \"alt\": True},\n (\"action.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"ctrl\": True, \"alt\": True},\n (\"action.clickselect\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"sequencer.select\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"alt\": True},\n (\"sequencer.select\", {\"type\": params.select_mouse, \"value\": 'PRESS', \"shift\": True, \"alt\": True},\n (\"clip.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True},\n (\"clip.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"gpencil.annotate\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True, \"key_modifier\": 'D'},\n (\"gpencil.annotate\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"shift\": True, \"alt\": True, \"key_modifier\": 'D'},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"gpencil.select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"alt\": True},\n (\"gpencil.select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"shift\": True, \"alt\": True},\n (\"gpencil.draw\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.draw\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"paint.weight_gradient\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"paint.weight_gradient\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"ctrl\": True, \"alt\": True},\n (\"mesh.loop_select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"alt\": True},\n (\"mesh.loop_select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"shift\": True, \"alt\": True},\n (\"mesh.edgering_select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"ctrl\": True, \"alt\": True},\n (\"mesh.edgering_select\", {\"type\": params.select_mouse, \"value\": params.select_mouse_value, \"shift\": True, \"ctrl\": True, \"alt\": True},\n (\"mesh.loop_select\", {\"type\": params.select_mouse, \"value\": 'DOUBLE_CLICK', \"alt\": True},\n (\"gpencil.primitive_line\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.primitive_box\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.primitive_circle\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.primitive_curve\", {\"type\": 'LEFTMOUSE', \"value\": 'PRESS', \"alt\": True},\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n (\"gpencil.select_lasso\", {\"type\": params.action_tweak, \"value\": 'ANY', \"ctrl\": True, \"alt\": True}, None),\n```\n\n", "Seemingly no way to remap \"use tool\" button?\nBroken: 2.80+\nWorked: 2.79\n\nSee 599\n\nThere doesn't seem to be a way to remap or disable the button which activates the tools selected by the buttons found in the T panel. Sometimes this button conflicts with the 2.7x keymap. For example, in the node editor, left tweak *always* applies the selected tool (box select by default), which prevents all the 2.7x Left Tweak shortcuts from working\n\n\n[27_28_keymap.blend](27_28_keymap.blend)\n\n1. Open the file with --factory-startup\n2. Go to the keymap section in the preferences\n3. Select the 27X preset\n4. Try to use the Shift+Left Tweak shortcut to add a reroute node\n5. Box select will be invoked instead", "Sculpt Mode Border Hide Tool Todo\n## Sculpt Mode Border Hide Tool Todo\n * - [x] {icon circle color=red} When this tool is selected, it currently shows the wrong tool settings which affect the brushes, not this tool. These should not be visible when this tool is active.\n * - [ ] {icon circle color=yellow} We should add a tool setting to add/remove the mask as you drag. Currently, the only way to remove from border hide, is very hidden: You drag with RMB. This could become more consistent and more discoverable if we added the same Add/Remove buttons as with other Sculpt tools", "VSE/UI: Operators missing shortcut keys\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 462.59\n\nBroken: version: 3.0.0 Alpha\n\nSet Preview Range to Strips & Set Frame Range to Strips:\n{[F10276617](image.png),size=full}\n\nAnimation Render options are missing keys:\n{[F10276638](image.png),size=full}\n\nSide of Frame... Current Frame - is also suffering from a bad name and ctrl+left click on Playhead to select is missing.\nLeft & Right: Brackets do not work on non English keyboards.\n{[F10276621](image.png),size=full}\n\nAll of the handle selection options are missing keys:\n{[F10276625](image.png),size=full}\n\nAll of the Channel selection options are missing keys:\n{[F10276627](image.png),size=full}\n\nJump to previous/next marker are missing keys:\n{[F10276631](image.png),size=full}\n\n\n\n\n", "File browser buttons disappear on undo in another window\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.79\n\nBroken: version: 2.92.0 Alpha\n- All versions since the new file browser was introduced are affected by this\nWorked: Never\n\nI've tried to save file with CTRL+S and by accident pressed my LMB on a viewport. Because of that mistake, I pasted object in my scene, instead of pasting file name in \"Blender File View\" window. So I tapped CTRL+Z (while still being active in a viewport) and whole panel with buttons \"Save file\", \"Cancel\" in \"Blender File View\" window disappeared. I've tested this on a completely new empty blender file. Pressing on a viewport didn't recreate that effect (probably, because nothing was there). But pressing on outliner or timeline have the same issue.\n\n- Open Blender\n- CTRL+S\n- Press LMB on an Outliner\n- CTRL+Z\n\n[UI bug.blend](UI_bug.blend)", "How to keep asset viewer options and the search field in view?\n\n\nIt's annoying that animators have to scroll up all the time to access these (or down for the search field).", "Multi-Button Drag: not working properly for Object Dimensions\nWindows 7\n\nBroken: 2014-02-11\n\nMulti-Button Drag - D270 - doesn't work properly in N panel for Dimensions\n\nIf you drag over all 3, only X and Z change, Y remains unchanged. As mouse button is released, the first selected property will reset, only the last hovered is kept (and scale changed)", "Intermittent: Control + MMB to zoom non responsive or sluggish vs mouse speed\nOperating system: MacOS Ventura 13.4.1\nGraphics card: M1\n\nBroken: 3.5.1, Hash: e1ccd9d4a1d3, Branch: blender-v3.5-release\n\nZoom via Control + MMB becomes very sluggish to the point where excessive movement of the mouse is required. Does not match the mouse pointer speed for the rest of the application. \n\nAny model in perspective mode. Attempt to zoom view to the object.\n\n", "Float slider in the redo panel of the \"add object\" operator does not hide the mouse cursor\nBroken: 2.79b, 2.83\nWorked: (None)\n\nThe problem is not seen in editing mode, but in object mode when adding an object\nand changing the properties in the redo panel, the mouse cursor does not disappear.\n\n\n", "NLA strip unexpectedly auto-switching from Hold to Hold Forward\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 770/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.64\n\n\nBroken: version: 2.80 (sub 74)\n\n\nWhen creating a new action using the HOLD_FORWARD extrapolation, frames to the left of the keyframes will keep the value of the first keyframe, instead of ignore them like they did before and do with NOTHING extrapolation. They now behave like HOLD.\nThey work like normal in strips, just not as the *(orange)* action.\n\n[nla_hold_back.blend](nla_hold_back.blend)\n[2019-07-14 19-08-18.mp4](2019-07-14_19-08-18.mp4)", "Design: More frame snapping options\n## Problem\nWhen dragging the frame cursor around you often want to hit keyframes.\nCurrently this is hard to do, especially when zoomed out. (also because you are not sure when you've hit the key)\n\nFrame snapping already exists, but it only snaps to seconds.\nThis is activated with `CTRL`\n\n## Design Question\n* Where in the UI should this functionality live\n* How can this feature be discoverable and understandable for users\n* What are the snapping options required and how do they work\n\n## Things to consider\n* there is different snapping behaviour in the VSE, so different time editors can have different snapping functions. This needs to be clear to the user\n* In the graph editor and the dope sheet there is already a snapping menu, but that is for snapping the keys while moving them. How do we differentiate the two\n\n## Proposed Solution\nThe idea is to add an option to the way the snapping is performed, e.g. either to seconds or to keys (or another snap type)\n[2021-10-02 19-50-20.mp4](2021-10-02_19-50-20.mp4)\n\nThe snap type would be chosen in a dropdown menu\n\n\nThe subframe setting would be moved to a scene setting\n", "GPencil: Stroke not working with Holdout\nOperating system: Windows 10\nGraphics card: GeForce GTX 970\n\nBroken: 2.82a, as found on the splash screen\nWorked: Never\n\nSetting a collection as holdout in view layer options in cycles works as expected on grease pencil objects in the viewport render preview, but does not work on grease pencil objects when rendering. The entire grease pencil object is visible in the renders, whereas in the viewport render preview the parts of the grease pencil obscured by foreground objects set to holdout is alpha. \n\nCreate a grease pencil empty behind the startup cube. \nDraw some stuff. \nPut the startup cube in a collection and set the engine to cycles. \nFrom the view layers options for the new collection containing the startup cube, set it to holdout. \nTurn on render preview from the viewport shading options. Notice that the cube is invisible, along with the parts of the grease pencil that had been obscured by the cube.\nPress F12\nNotice in the rendered image that all of the grease pencil is visible, even the parts we had expected to be alpha masked from the startup cube holdout.\n\nSee the attached blend file.\n[HoldoutTrouble.blend](HoldoutTrouble.blend)\n", "Text Object -> Edit Mode, Many Short-cuts Showing Here & There, But Not Working \nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: AMD Radeon HD 8400 / R3 Series ATI Technologies Inc. 4.5.13399 Core Profile Context 15.200.1065.0\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (Probably Never)\n\n\nFor Text Object -> Many Short-cuts are showing in e.g. menus but not working. e.g. T (Tool Shelf), N (Numeric Pane) e.t.c\n\n\n\n\n*Wait:* I am checking a list the all short-cuts which are not working...", "Python API - Modal operator losing context.area after focus is stolen\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.77\n\n\nBroken: version: 2.80 (sub 58)\nWorked: (optional)\n\n\nWhen focus is stolen (for example, when the user uses Ctrl+S and the file browser dialog pops up), a running modal operator will lose the context.area and it never returns even after the dialog has been closed.\n\nRun the attached modal operator from a VIEW_3D area, without pressing RMB or Escape (which will cancel the operator), bring up the Save File dialog. Watch the console output- it will change from VIEW_3D to \"no context area\" because the dialog has focus and the area is no longer valid, but when you Cancel the file save operation, you would expect the context.area to return but it never does. You must cancel the operator and restart it.[modal_area_focus_bug.py](modal_area_focus_bug.py)\n\n", "Insert keyframe wrong with cycle-aware keying after deleting end frame\nOperating system: Linux-6.4.6-arch1-1-x86_64-with-glibc2.37 64 Bits, X11 UI\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.86.05\n\nBroken: version: 3.6.0\n\nI have an action with manual frame range, 1 to 11, with cyclic animation checked. I have cycle-aware keying on.\n\nI'm not sure exactly when the bug starts, but eventually adding a frame at time 9 creates the keyframe at frame 4 again.\n\n\nI'm not sure exactly when the issue started, but roughly\n1. Cube\n2. Insert frame at time 1\n3. Turn overlay action into layer (stack?) action with the down icon button\n4. Set up manual frame range 1-11, cyclic\n5. Turn on cycle-aware keying\n6. Insert another frame at time 6, blender automagically adds a frame at time 11 to match the frame at time 1\n7. Delete the frame at time 11\n8. Move to time 9 and insert frame\n\n7 seems to be the key. Additionally, it's impossible to add a new frame at time 11. The issue persists after closing and re-opening. \n The issue disappears when cycle-aware keying is off, but comes back if I turn it back on.\n\nI didn't try a daily build, but I searched issues to see if anything similar was reported and found nothing.\n\n" ]

[ "Press and hold W key in edit mode makes menu appears and disappear fast\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\n\nBroken: version: 2.92.0 Alpha\nWorked: 2.90.1 release\n\nUsing the \"Blender\" default keymap and the option to select with the right mouse button, if you hold the {key W} key in edit mode the menu appears and disappear fast and the \"Edge Bevel Weigth\" operation is triggered.\n\n\nW calls this option. So Blender reads holding {key W} as 2 hits, even if repeat is off.\n\n- With \"Blender\" default keymap + RMB selection **Go to edit mode**\n- Set selection to edges\n- Press and hold W key" ]

[ "Particle Texture influence broken when using two particle system on the same object\nOperating system: Ubuntu 18.04\nGraphics card: GeForce 970GTX\n\nBroken: 2.90.1\nWorked: Never (2.8+)\n\n\nWhen using two particles system on the same object with an influencing texture, one particle system has a wrong influence (e.g. Particle System shrink to quickly)\n\nBlend File: [particle_texture_bug.blend](particle_texture_bug.blend)\n\n0. Set the animation start and end to 100 - 200\n1. Create an Emitter and a Particle Object\n2. Add a Particle System to the Emitter and use the Particle Object for rendering\n3. Set the Start and End to 100 - 200\n4. Give the Particle System a Blend Texture that influences the size with \"Strand/Particle\" as Coordinates\n5. Add another System and Duplicate the previous one to get the exact same setting\n6. Change the Start and End of the new System to 0 - 100\n7. Bake all Dynamics\n8. Render Frame 199 and compare to rendering of 101. The particles in 199 are tiny in comparison (or invisible). They should look near identical\n\nOptional: \n9. Remove the 0-100 Particle and Redo Baking\n10. Frame 199 now has normal sized particles\n", "Crash when loading a geometry nodes asset into itself\nOperating system: Win 11\nGraphics card: 2070\n\nBroken: 3.4, 4.0\nWorked: none\n\nWhen working on an geometry nodes asset in an asset library - if you search for the name and adding it to the graph it causes a crash. \n\nDrag n drop from the asset browser behaves correctly here - so I think its just the node search code which is missing the check for self referencing.\n\n1. Download the blend file.\n2. Put into a folder.\n3. Add folder as asset library in prefs -> file paths -> asset libraries.\n4. Open the blend file from that folder.\n5. Go to the geometry node graph (should be open already).\n6. `Shift` + `A` to add new nodes.\n6. Search \"noop\" and press enter.\n\n", "Blender just crashes for no reason, and the crash report is empty. (Segmentation fault)\nOperating system: Arch Linux x86_64\nGraphics card: AMD Radeon RX 6900 XT\n\nBroken: (example: 3.6.2, ffe93138751e, master, 2023-08-31, as found on the splash screen)\nWorked: (No idea when this error started happening.)\n\nBlender does not start up. When I start it via the command line, I get this error:\n```\nWriting: /tmp/blender.crash.txt\nSegmentation fault (core dumped)\n```\n\n`/tmp/blender.crash.txt`:\n```\n# Blender 3.6.2, Commit date: 2023-08-14 14:05, Hash ffe93138751e\n\n# backtrace\n\n# Python backtrace\n```\n\nTrying to open Blender.\n\n", "Crash when auto script execution is enabled\nOperating system: Windows 10 64-bit operating system, x64-based processor\nGraphics card: NVIDIA GeForce RTX 2060\n\nBroken: Blender 4.0.0 7372b3bdda35 and later\nWorked: never? 2.93 fails too\n\nBlender daily build crashes immediately on launch. The blender start screen isn't shown at all, only a grey window which then immediately closes. The crash log indicates that it is a memory EXCEPTION_ACCESS_VIOLATION (see log file attached).\n\n- Open preferences, enable `auto script execution`\n- Save prefs and close blender\n- Rename attached file and replace it with existing startup file\n- Launch blender (crash)\n\n- - -\n<details>\n <summary> Original report </summary>\nStart Blender through Steam, Blender Launcher, blender_debug_log.cmd, blender_debug_gpu.cmd, or blender_debug_gpu_glitchworkaround.cmd.\nA grey window opens with the Blender logo in the top bar.\nThe window closes as Blender crashes.\n\n[ Crashing_4-0-0_daily.blend](https://projects.blender.orgattachment)\n</details>", "Clicking \"Particle Edit\" mode causes Blender to crash\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: Quadro RTX 3000 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.89\n\nBroken: version: 3.5.0\nWorked: I dunno\n\nCan't enter into the Particle Edit mode.\n\nThis issue is probably related to: 106927#issue-96303\n\nIn the sample file, in the Curves Sculpt mode, after converting my Curves (name of the layer: \"my_test_4\") to a particle system, then select the newly converted particle system and enter into the Particle Edit mode, and Blender crashes.\n\n\n\nA sample file is attached. Thank you!\n\n", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Crash on render\n**System Information** MSI BRAVO 15\nOperating system: WINDOW 11\nGraphics card: RX 6600\n\n**Blender Version**3.6\n\nWhen attempting to render a scene in Blender, the software consistently crashes, causing frustration and hindrance to the rendering process. This issue is hindering users from efficiently creating their projects and requires immediate attention to maintain a stable and reliable rendering environment.\n\n[debug-log](https://projects.blender.orgattachment)\n[system info](https://projects.blender.orgattachment)", "AMD - Blender Crashes When I Start Render - Cycles GPU Compute\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: AMD Radeon RX 6800 XT ATI Technologies Inc. 4.5.0 Core Profile Context 23.7.1.230626\n\nBroken: version: 3.6.1\n\n[Whenever I start Rendering, It crashes when using cycles on GPU compute. Doesn't happen on CPU option.]\n\n[On render engine, select cycles. Then set it to GPU Compute. Then click on render on top left corner. Then select render image.]\n\n", "Regression: Crash when Sculpting on a mesh from a GP Lineart Modifier\nOperating system: Windows-10-10.0.22621-SP0 64 Bits (W11)\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.2, 4.0.0 Alpha\nWorked: 3.1\n\n\nWhen trying to sculpt a mesh in an assigned collection that a grease pencil lineart modifier references and the layer it is drawing to contains Grease Pencil strokes - it will instantly crash Blender. \n\n\n1. Download and extract zip to get the test blend file\n2. Load attached file with Load UI off\n3. Select the bust, it is in a GP lineart modifier collection \n4. Go to sculpt mode\n5. Try sculpt\n6. Crash\n\nAlternative steps:\n1. Create new scene\n2. Create a sphere\n3. Create a Grease Pencil object\n4. Add sphere to a new collection\n5. Select the GP object and add a lineart modifier\n6. Reference the sphere collection in the modifier\n7. Assign layer and material to the modifier\n8. Draw in the same layer in Draw mode of the GP\n9. switch to sculpt mode\n10. Try sculpt over sphere surface\n11. Instant crash\n\n", "blender crash if boolean + very very big scale\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.2.1, 3.2\n\n[2022-07-21 19-15-24.mp4](2022-07-21_19-15-24.mp4)\n\n1. add boolean modifier\n2. scale any object on very very big value", "Hair particle system altering display percentage crashes Blender.\nOperating system: Windows-10-10.0.17134 64 Bits\nGraphics card: GeForce GTX 760/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.36\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\nChanging the percentage value in the Viewport Display section causes Blender to instantly crash.\n\nSet the display percentage with the currently active particle system\n[field_3400_autosave.blend](field_3400_autosave.blend)", "Dae parsing issue\nOperating system:Windows 10\nGraphics card: NVidia gforce GTX660\n\n 2.79\n\n\nIt is a common error, when trying to upload an object to Second-life or Opensim, to be unable to upload a model,created in Blender, due to a \"DAE parsing issue\"\nThis object is one that is effected.\n\n[lowpol.blend](lowpol.blend)\nExport as .DAE file and upload to Opensim or Second-life", "Instancing a collection with an object that has been Disabled in Viewport fails to correctly link in disabled objects breaking \"Make Instances Real\" operator\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.91.0\n\nInstancing a collection into a scene fails to properly link in objects that have been disabled in viewport (henceforth disabled object) using either the \"Disable in Viewport\" icon in the outliner or by setting obj.disable_viewport = True in a script or console. This bug doesn't appear when you hide an object using the H key or obj.hide_set(True)\n\nThis means that if you instance a collection and then \"Make Instances Real\" that the disabled objects will not be transformed with the parent object even if they have been parented to it in the original file and Keep Hierarchy is checked. \n\nThis is a problem because if you have used the disabled object as the operand in e.g. a boolean modifier then as soon as you make a collection real your collection breaks. This is doubly a problem since the default behaviour in the new asset browser is to instance collections into the scene rather than appending them. This bug also affects instancing a collection in the same scene as well as instancing in a collection from an external file. \n\nStart from default file. \nAdd a second cube.\nTransform second cube and place on corner of default cube.\nMake second cube a child of default cube.\nAdd a boolean difference modifier to default cube.\nSet second cube as Object in boolean modifier\nDisable second cube in viewport in the outliner\nInstance collection into scene by dragging and dropping collection from outliner somewhere which isn't the origin.\nEverything appears as it should.\nSelect instanced collection\nSelect Object > Apply > Make Instances Real\nWhile the none disabled cube has been made real the second cube hasn't and if we look in the modifier you can see that the boolean object is still set to the original second cube.\n\nIf instead of disabling the second cube we instead hide it and then instance the collection into the scene the second cube is linked in (although it's also unhidden which is annoying). If we the make this collection real everything works as it should (i.e. the second cube used as a boolean is also made real)\n[link bug.blend](link_bug.blend)", "Use after free: Baking Rigid Body physics causes instant crash\nOperating system: macOS 10.15.7\nGraphics card: AMD Radeon Pro 5500M 4 GB / Intel UHD Graphics 630 1536 MB\n\nBroken: 2.92.0, branch: master, commit date: 2021-02-24 16:25, 2.93.0 Beta, branch: master, commit date: 2021-04-28 22:25, \nbroken: f7b22fc3d27113715e8726c69ab3264e1d487439\nWorked: \n\nBaking the Rigid Body world physics causes Blender to crash either instantly, or after baking less than 100 frames (it varies). Interestingly, a friend with the PC version of 2.92 says the project works fine, no crashes whatsoever.\n\nOpen the project. Make sure Allow Negative Frames is on under Preferences/Animation. Go to the Rigid Body World and hit bake.\n\n[physics test b.blend](physics_test_b.blend)\nAsan report: [T88113_asan_report.txt](T88113_asan_report.txt)", "Pinning Bones/Constraint Tabs in Properties Editor doesn't work\nBroken: 2.76 - master (1283c66)\nWorked: Unknown\n\nThe context pinning functionality in the Properties Editor doesn't work for Bones and Constraints. There may be a few others too that I haven't noticed yet, but these are the biggest ones.\n" ]

[ "Make single user crashes blender\n\n\nOperating system: Windows 10\nGraphics card: Titan Black\n\nversion: 2.80 (sub 35), branch: blender2.8, commit date: 2018-12-02 04:09, hash: 925380050d0, type: Release\nbuild date: 02/12/2018, 01:16\nWorked: (optional)\n\nWhile using \"make Single User\" and taping on Object Data, Blender crashes.\n\nBased on the default startup or an attached .blend file (as simple as possible).", "Crash on Make Single User\n[system-info.txt](system-info.txt)\n\nDefault cube - Ctrl F - Make Single User - dialog - Check 'Animation' - Crash!\n\nActually - 'All Objects' seems fine, but 'Selected Objects' will crash on many of the check boxes?\n\nThanks\nWayne", "Make Single User crashes\nOperating system: Windows 10 Pro 64bit\nGraphics card: GTX 960\n\n2.80 Beta Hash 4c31bed6b46\n\n\nWhen I select an object and access Object > Relations > Make Single User > Object and Data > Selected Objects, Blender crashes and closes.", "Crash with Make Single User > Selected\nOperating system: Windows 10 Home Edition, Build 17134\nGraphics card: AMD A12-9720P RADEON R7\n\n2.80, a205493426b, 2018-12-03 00:27\n\nWhen attempting to use Make Single User on Selected Objects, the program crashes instantly. This does not occur when picking All Objects.\n\nIn Object Mode: Object > Relations > Make Single User > (Any option) > Selected Objects.", "Make Single User - Object Data CRASH\nOperating system: Windows 10\nGraphics card: Nividia Geforce 940mx\n\nBroken: blender-2.80.0-git.26d5a3625ed-windows64\n(example: 2.79b release)\n(example: 2.80, edbf15d3c044, blender2.8, 2018-11-28, as found on the splash screen)\nWorked: (optional)\n\nIf you have a duplicate of an object with the same object data and then want to make single user Blender crashes. It happens exactly when you check \"object data\" box on the pop up make single user window.\n\nDuplicate and object with alt+D - make single user on any of them by checking the box \"object data - CRASH!\n", "Blender crashes due to errors making materials single user\nOperating system: Win 10 64bit - Intel i7 16GB ram\nGraphics card: MSI 1080 Gaming X\n\nversion: 2.80 (sub 37), branch: blender2.8, commit date: 2018-12-12 01:18, hash: 16fc62e15f0, type: Release\nbuild date: 11/12/2018, 17:55\n\ncreate an object with a material. Duplicate it with Alt-D a few times. Select an object and then make it single user Object&Data&material from the Object/Relations menu. Blender will crash\n\nI found this problem because I have been using an earlier version (24th Nov - there were big performance problems after this) all my animated items had been created using Alt-D but I wanted the materials on each character to be changed separately. With this build I could make single user but for some reason what ever I did the Shader Node Tree curve in the graph editor stayed with the material no matter how I tried to recreate the material as a different material. Even deleting and recreating the animated shader node made no difference. I believe these two problems might be related and might possibly be contributary to the performance slowdowns experienced by some with eeVee after Nov 25th.\n" ]

[ "Macro operators [e.g. Loop cut and slide]: \"adjust last operation\" fails completely [when called from the menu -- F9 is fine]\nBroken: 2.8x\nWorked: 2.79b\n\nThis might fall into the same boat as anything related to #45964, although it does seem to be \"new\" in 2.8x.\nAfter performing a loop cut and slide operator, you can adjust the properties in the operator panel (albeit with caveats from #45964). However, if you instead open the Adjust Last Operation panel, it will list a different operator name (Loop Cut, not Loop Cut and Slide), and while you can change properties there, those changes have no effect, even though they are reflected by the operator panel.\nThis didn't use to happen in 2.79b, though IIUC the \"Adjust...\" panel is different in 2.8x from the \"Redo Last\" of the 2.7x era.\n\n1. Load the default scene.\n2. Go into edit mode on the cube.\n3. Add a loop cut, e.g. via the shortcut or using the active tool\n4. Expand the operator panel in the bottom left and adjust properties. Note how those adjustments do have an effect on the result.\n5. Go Edit -> Adjust Last Operation\n6. Note the different operator name in that newly opened panel (Loop Cut vs. Loop Cut and Slide).\n6. Try adjusting any properties in the newly opened panel. Note how the operator panel in the bottom left picks up these changes, but at the same time they have no effect on the operation.\n\n[Demo video ](eAlL08M8AZA)", "Sequencer: preview TODO's\nThis tasks is to track TODO's for the sequencer preview which can now select sequencer strips directly.\n\n\n**Unimplemented Operators**\n\n**Display Settings**\n\n- - [x] Option to toggle gizmos (matching the 3D view). ebe216f532\n- - [ ] Option to display Origins & Origins (all).\n\n**View**\n\n- - [ ] Frame Selected\n\n**Animation**\n\n- - [ ] Keyframe menu (location, rotation.. etc).\n\n**Selection**\n\n- - [x] All/None/Invert f0d20198b2\n- - [ ] Hide/Reveal\n\n- - [ ] Lasso Select *(low priority)*\n- - [ ] Circle Select *(low priority)*\n- - [ ] Menu picking (match 3D view Alt-Select). *(low priority)*\n\n**Transform**\n\n- - [ ] Snap Menu *(low priority)*\n- - [ ] Transform Origin *(low priority)*\n\n**Other**\n\n- - [x] Delete f0d20198b2\n- - [ ] Copy/Paste\n\n----\n\n**Gizmos**\n\n- - [ ] Crop gizmo. *(low priority)*\n\n-----\n\n\n**Bugs/Limitations**\n\n- Non 1:1 aspect currently don't work properly (when the render output aspect is not `1:1`).\n", "VSE: 'Clear fade' function clears every opacity/volume keyframe and resets non-default baseline value\nOperating system: Linux-5.4.0-84-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: GeForce 9800 GT/PCIe/SSE2 NVIDIA Corporation 3.3.0 NVIDIA 340.108\n\nBroken: version: 3.0.0 Alpha\nWorked: Unsure\n\nThe 'Clear fade' function used on either a visual or audio strip does not only clear the start/end fade(s), but any other set opacity/volume keyframes as well. It will also reset the baseline value to 1, even if the fades in the strip clearly operate on a baseline set to other than 1.\n\n[clear-fade-function-clears-internal-value-changes-and-does-not-preserve-non-default-baseline-value.blend](clear-fade-function-clears-internal-value-changes-and-does-not-preserve-non-default-baseline-value.blend)\n(NOTE: In blend file, choose 'Clear fade' from the context menu for the top strip)\n\n- Add sound or visual strip\n- Change opacity/volume to 0.5\n- Choose 'Fade in and out' from the context menu, and note that the changed 0.5 baseline has been respected\n- Add a manual keyframe change/fluctuation somewhere in the middle of the strip\n- Choose 'Clear fade' from the context menu\n- Notice that all keyframes including the middle ones are now gone, as well as the 0.5 baseline value (replaced by default 1.0)\n\nIt is not expected that any fades apart from in and out will be cleared. In particular, it is quite common to have volume keyframes toyed with throughout the strip, so I do think targeting just the ins and outs is useful, certainly much more so than destroying all the keyframes (which one can achieve by choosing \"Clear keyframes\" directly on the property).\n\nAlso all the options to add fades are for adding fade-ins and fade-outs, so it's expected that the \"Clear fade\" will remove fade ins/outs as well. Either way, the description should be made clearer to reflect what will really happen.\n\nA user-set, non-default baseline should also be preserved, if it clearly exists.", "Video reference doesn't load at the frame where it was saved if the mode was not object mode\n## Simplified Report\n\n- Add a movie reference image object.\n- Set the scene frame to 10.\n- Add a cube and enter Sculpt mode\n- Change the frame number to 20.\n- Save the file and reload.\n- The reference movie displays frame 10 even though the scene is at frame 20.\n- This seems to happen because the scene frame was changed and saved in non Object mode\n\nSeems to objects `ImageUser` that is used for saving is not the same as the one used for drawing, and the `framenr` is only updated for the ob used in drawing?\n\n## Original Report\n\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.0\n\nVideo reference does not load at same place left when closing and reopening blender.\n\nLoad a video file as a reference or a background. Scrub to a random location not at 0 for the video's duration. Save and close blender. Reopen blender. You will notice that the timeline is on frame 22 (for example) but the reference video is displaying frame 0. It would be nice if since you are on or saved on frame 22 that the reference video would load in the startup of blender on frame 22 of its playback as well.\n[Linework Finished.blend](Linework_Finished.blend)\n\n", "Animation Editors: Modifier keys cancel out dragging handles \nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.4.0 Alpha\n\nIn recent releases db4313610cab was implemented and the animation editors are still not adapted to this behaviour. This fix allows the use to press any keyboard key, like `Shift` or `X`, etc, before the drag threshold of the current action is reached, and it will cause the operator to execute.\n\nFor example: Click and hold on a move gizmo in object mode. Either drag 3 pixels to start the operation and then press `X` and hold `Shift` to slowly move along the x axis.\nOR immediately press `X` and hold `Shift` without reaching the drag threshold of 3 pixels to already start the operator.\n\nBefore this fix, any pressed keys would cancel the click & drag and prevent the operator from running.\nThis is still the case with the graph and dopesheet editors.\n\n- Create a simple animation to the default cube\n- click and drag on the keyframes or handles in the animation editors\n- Now click and hold them, hold `Shift`, and then continue to drag. Nothing happens.", "VSE: issue moving strips with changed speed factor\nOperating system: Linux-5.18.10-76051810-generic-x86_64-with-glibc2.34 64 Bits\nGraphics card: AMD Radeon Vega Frontier Edition (VEGA10, DRM 3.46.0, 5.18.10-76051810-generic, LLVM 12.0.1) AMD 4.6 (Core Profile) Mesa 21.3.5\n\nBroken: version: 3.5.0 Alpha\n\nIn the Video Sequence Editor, when moving strips that have the speed factor different than default 1.0 the strips seem to glitch so the start and end of each strip fluctuates.\n\nNew file\nGo to VSE\nAdd a movie strip\nUse \"K\" to cut the movie strip a few times (so you have maybe 10-15 strips)\nSelect a strip, go to the Time panel and change the speed factor to 1.5\nOne by one, do the same with all the strips (perhaps giving them various values between 1.1 and 1.5)\nMove each strip up against each other, so there's no dead space between\nSelect all of them and move them around...\n{[F14115112](vse_speedup_movement_glitch.mp4)}you will see the length of each strip fluctuates\n\n", "VSE prefetch not showing clip's keyframe changes\nOperating system: Ubuntu Linux 18.04.5 amd64\nGraphics card: Nvidia GTX 1070\n\nBroken: blender-2.90.1, blender-2.92.0-9d6d5b9bebd1-linux64\nWorked: unknown\n\nWhen VSE \"Prefetch Frames\" is turned on and you edit a clip's keyframes, the prefetch will continue rendering frames into the cache without your changes.\n\n\n1. Launch Blender\n2. Choose \"Video Editing\" under New File\n3. Drag a still image (any PNG should do) into the timeline.\n4. Drag clip extents so that it starts from frame 0 and goes to frame 2000.\n5. In video preview, enable \"Prefetch Frames\"\n6. Go to frame 20 in the timeline and select the clip.\n7. Create an Opacity keyframe by clicking the dot next to Opacity.\n8. Go to frame 80 in the timeline.\n9. Change the clip Opacity to 0.4 and click the dot to create a new keyframe.\n10. Start playback and see that the clip shows 100% opacity, even though the timeline shows that it should be 40%. Changing the proxy render size will clear the cache and show correct results.\n\nIf you clear the cache and do steps 8-10 quickly enough (i.e. before the prefetch finishes), you will also see that even though you modify the clip while prefetch is not done, the prefetching will continue to render incorrect frames. E.g. if you create the opacity keyframe while the prefetching is at frame 500, prefetched frames after 500 will still show as 100% opaque.", "Movie clip playback FPS is reduced in 3DView.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.59\n\nBroken: version: 2.83 (sub 9)\nWorked: ～2.79\n\nWhen a movie clip is read into a background image and displayed on a 3D view, the FPS decreases in a sequence of 2K or more.Even in full HD, FPS decreases a little.\nThere is no problem if you play only with the movie clip editor.\nRegardless of file format or file size, resolution seems to be involved.\nThere was no problem if the resolution was dropped to 2k or less.\nEven if the sequence is full HD, the FPS will drop slightly. There is no problem if it is less than Full HD.\n\nBelow is a comparison with 2.79 under the same conditions.\n\n2.83\n[Blender 2020-03-18 21-57-40.mp4](Blender_2020-03-18_21-57-40.mp4)\n\n2.79\n[Blender 2020-03-18 21-56-03.mp4](Blender_2020-03-18_21-56-03.mp4)\n\n\n[#74814.blend](T74814.blend)\n\n\n- Open file\n- Play animation (and make sure it is not able to play at 1200FPS)\n# In properties editor, toggle Show Background Image\n\nFPS rate will drop significantly", "Can't undo the keyframe change in edit mode\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.15\n\nBroken: version: 3.1.2\n\nWhile in edit mode, when you hit Undo in the graph editor after moving a keyframe, the graph doesn't change.\n\n- Animate anything using a couple of keyframes.\n- Change to edit mode in viewport\n- Switch timeline to the graph editor (not necessary, you can keep timeline as it is)\n- Select a keyframe and move it\n# Hit undo\n\nResult: The graph editor stays the same\n\nExpected: The move you just made in the graph editor should be the only thing being undone!\n\nTest file with first four steps included\n[#97700.blend](T97700.blend)", "Multiple Frame Rates/ Playback/ VSE scene strips\nOperating system: windows 10\nGraphics card: rtx 2060\n\nBroken: 2.90.0\nWorked: Never\n\nI'm using the VSE and I have 2 scene strips from two different scenes inside the VSE.\nThe first scene strip references a scene that is set to 12 FPS.\nThe second scene strip references a scene that is set to 24 FPS.\nThe 3rd scene is where I'm using the VSE, and this scene is set to 12 FPS.\n\nThe problem is, all the scene strips are playing back at 12 FPS by default in the scene where I'm using the VSE. In other words, the scene that is 24 FPS is playing back at 12 FPS.\nIf I'm not mistaken, this wouldn't happen with real footage. Two recordings with two different FPS dropped into a project would play back at different speeds.\nWhat I'm trying to do is have a video at 12 fps, and the put a 24 fps scene strip in the VSE so that i can put it in slow motion and have it play back with the same quality. \n\n- Open attached file\n- Run VSE\nNote that the cube shown on each strip has the same speed in the animation.\nIt was expected that one was twice as fast than the other.\n[_VSE scene strips.blend](_VSE_scene_strips.blend)", "Re-render after saving a render result as image with 'Copy' option turned OFF makes blender unresponsive\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.86\n\nBroken: versions 2.79 & 2.81 (sub 16)\nWorked: (optional)\n\nRender with F12, save the rendered image with 'Copy' turned off, and render again. This makes (parts of) Blender unresponsive.\n\n - Render the default cube.\n - Image - Save as, and uncheck Copy in the options.\n - Click Save as image.\n - Re-render the image.\n - Press {key Escape} to close the render window (it doesn't).\n - Press {key F12} to re-render (it doesn't).\n\nStrangely enough some parts of Blender are still responsive. For example {key Shift F12} still switches to the timeline.\n", "Crash when switching from solid to rendered view\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.6.0 Alpha\n\nI have a scene about 1.7GB in size with 4K textures but renders in 2K using simplify, blender crashes when I open the file > hit F12 and render the scene then close the render windows and switch to render view > it starts compiling then blender closes, in the past it would show out of memory this report: \n\n105522#issuecomment-904934 \n\nThis issue happens only when persistent Data is active\n\nunfortunately I can't share the file at the moment. \n\n\n\n\n", "Animation. Keyframes. Copy - move to the current frame.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.77\n\nBroken: version: 2.93.0\n\n[Animation. Keyframes. Copy - move to the current frame for replacing the previous one. If you copy keyframes and move them to the current frame it won't work properly. All will seem good but the last frame will play an animation relay on the previous parameters. ]\n\n\n[bug.blend](bug.blend)\n\n\n\n", "Crash playing animation with rigid body\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.83.1\n\nAfter opening a blend file (ver BLENDER-v280RENDH). The frame starts at 25 upon opening it. I changed the frame to 0. I pressed space to play then the application closes.\n\n[Whiskers Away.blend](Whiskers_Away.blend) \n- Open file\n- Rewind to frame 1 {key Shift Left}\n- Seek to frame 0 {key Left}\n- Play animation\n\nCrash doesn't happen every time. Redo all steps in such case. There is about 30% probability to reproduce successfully.", "Cycles: Keyframing motion blur shutter value fails to apply it properly when generating animation\nOperating system: Windows 10\nGraphics card: Nvidia RTX 3090\n\nBroken: 3.6, 4.0\n\n\nIf we keyframe shutter property of motion blur, the updated shutter value is not considered during the final render.\nSuppose shutter value is 0 at the start of render, it will remain 0 thorught the animation even if we've keyframed the value.\n\n- Open attached file\n- select cycles render Engine\n- Start animation render (Ctrl-F12)\n" ]

[ "Audio strip - SHIFT K crashes Blender\nSystem Information\nOperating system: Windows 10 Pro\nGraphics card: GTX 1080 Ti x2 [SLE] drivers 430.64 [latest]\n\nBroken: blender-2.80.0-git.81b68f7279cd-windows64\nWorked: unknown\n\nShortcut crashes Blender\n\nAdd Movie with Audio\nAdvance a few frames\nWith Audio selected press SHIFT K and Blender crashes." ]

[ "Intersect Knife produce wrong results when the faces are coplanar\nOperating system: windows 10\nGraphics card: 1070X2\n\nBroken: 2.83.5\nWorked: 2.79\n\n\nso all the points are on the plane, and I'd make a grid of faces. So i use the knife (self intersect) tool, which i used to use in 2.79, and it produces results that are not correct\n\n\n\n\nFile :\n[untitled.blend](untitled.blend)", "Multi-Object-Mode: EditMesh Tools\nNote, while these follow menus, duplicates have been removed.\n\n## Remaining\n\n- MESH_OT_select_similar_region ([D3672](D3672))\n- MESH_OT_knife_project (leave aside for now, too complex - see [D3637](D3637))\n- MESH_OT_knife_tool (leave aside for now, too complex - see [D3637](D3637))\n\n\n## Mesh Menu:\n\n- ~~MESH_OT_duplicate_move~~ 1ade071052\n- ~~MESH_OT_extrude_edges_indiv~~ d8f70a26fd\n- ~~MESH_OT_extrude_faces_indiv~~ 07774d4860\n- ~~MESH_OT_extrude_verts_indiv~~ 7c415e6fd7\n- ~~MESH_OT_extrude_edges_move~~ 1d7bdbd273\n- ~~MESH_OT_extrude_vertices_move~~ 1d7bdbd273\n- ~~MESH_OT_delete_loose~~ bfc9d426bb\n- ~~MESH_OT_decimate~~ 5b45d32f3a\n- ~~MESH_OT_dissolve_degenerate~~ (initial comsmit) bfc9d426bb\n- ~~MESH_OT_dissolve_limited~~ 09e14cb5bc\n- ~~MESH_OT_face_make_planar~~ 0e5a4f927a\n- ~~MESH_OT_vert_connect_nonplanar~~ b942106c73\n- ~~MESH_OT_vert_connect_concave~~4043e4b820\n- ~~MESH_OT_fill_holes~~ 69b95b45f4\n- ~~MESH_OT_symmetrize~~ 03a916e5b5\n- ~~MESH_OT_symmetry_snap~~ 7f9768cb80\n- ~~MESH_OT_bisect~~ f9eb2f29fe\n- ~~MESH_OT_sort_elements~~ ffb424c85f\n\n## Snap Menu:\n\n\n- ~~VIEW3D_OT_snap_selected_to_grid~~ 016ba0f38b\n- ~~VIEW3D_OT_snap_selected_to_cursor~~ ec3357e03a\n- ~~VIEW3D_OT_snap_selected_to_active~~ ec3357e03a\n- ~~VIEW3D_OT_snap_cursor_to_selected~~ ec3357e03a\n- ~~VIEW3D_OT_snap_cursor_to_center~~ (no change required)\n- ~~VIEW3D_OT_snap_cursor_to_grid~~ (no change required)\n- ~~VIEW3D_OT_snap_cursor_to_active~~ (no change required)\n\n\n## Vertex Menu:\n\n- ~~MESH_OT_merge~~ b73c7381cc\n- ~~MESH_OT_remove_doubles~~ 22b2bab702\n- ~~MESH_OT_rip~~ 9d0d7c375e\n- ~~MESH_OT_rip_edge~~ 6539c1fc41\n- ~~MESH_OT_split~~ fc8d903000\n- ~~MESH_OT_separate~~ 5e915baec4\n- ~~MESH_OT_mark_sharp~~ 0ac3d5f7db\n- ~~MESH_OT_vertices_smooth~~ d915aa57f8\n- ~~MESH_OT_vertices_smooth_laplacian~~ 1f30232aa7\n- ~~MESH_OT_blend_from_shape~~ 6ae09dfcad\n- ~~OBJECT_OT_vertex_group_smooth~~ 3dee8b4a12\n- ~~MESH_OT_shape_propagate_to_all~~ 3a43528274\n\n## Edge Menu:\n\n- ~~MESH_OT_edge_face_add~~ de2d642c84\n- ~~MESH_OT_subdivide~~ (initial patch)\n- ~~MESH_OT_subdivide_edgering~~ 853e55b043\n- ~~MESH_OT_unsubdivide~~ e245f7c09c\n- ~~TRANSFORM_OT_edge_crease~~ (initial patch)\n- ~~MESH_OT_mark_seam~~ 61f0608b4a\n- ~~MESH_OT_mark_sharp~~ 0ac3d5f7db\n- ~~MESH_OT_edge_rotate~~ 66ffbf9b7d\n- ~~MESH_OT_edge_split~~ 207b549d40\n- ~~MESH_OT_bridge_edge_loops~~ eb2226dde2\n- ~~TRANSFORM_OT_edge_slide~~ (wrapper to other operator, working)\n- ~~MESH_OT_loop_multi_select~~ df49155cfb\n- ~~MESH_OT_loop_to_region~~ 8b41999ea0\n- ~~MESH_OT_region_to_loop~~ 56d65719b3\n\n\n## Face Menu:\n\n- ~~MESH_OT_flip_normals~~ 4704f2687a\n- ~~MESH_OT_edge_face_add~~ de2d642c84\n- ~~MESH_OT_fill~~ b726ff4fc4\n- ~~MESH_OT_beautify_fill~~ e42b8c32ab\n- ~~MESH_OT_inset~~ 570455fb83\n- ~~MESH_OT_bevel~~ d9e04cb594\n- ~~MESH_OT_solidify~~ f309c34cfe\n- ~~MESH_OT_intersect~~ dffd37877d\n- ~~MESH_OT_intersect_boolean~~ 46aec45b2a\n- ~~MESH_OT_wireframe~~ a0a78f6da4\n- ~~MESH_OT_poke~~ 3a9669bfff\n- ~~MESH_OT_quads_convert_to_tris~~ (initial patch bfc9d426bb\n- ~~MESH_OT_tris_convert_to_quads~~ (initial patch bfc9d426bb\n- ~~MESH_OT_face_split_by_edges~~ b6a822d726\n- ~~MESH_OT_faces_shade_smooth~~ e4c4e8566f\n- ~~MESH_OT_faces_shade_flat~~ 9fc0c0c3de\n- ~~MESH_OT_uvs_reverse~~ 63951bc987\n- ~~MESH_OT_uvs_rotate~~ faaffc4011\n- ~~MESH_OT_colors_rotate~~ 42c5c20e54\n- ~~MESH_OT_colors_reverse~~ 753be0a2a5\n\n## Mesh Toolbar: (should be added to menu)\n\n- ~~MESH_OT_loopcut~~ e87dd9aa00\n- ~~MESH_OT_convex_hull~~ e1cee0ac71\n- ~~MESH_OT_delete~~ (initial patch bfc9d426bb)\n- ~~MESH_OT_delete_edgeloop~~ 49be803ad6\n- ~~MESH_OT_dissolve_edges~~ 7f5c469716\n- ~~MESH_OT_dissolve_faces~~ 848d9cda0e\n- ~~MESH_OT_dissolve_mode~~ (wrapper of vert/edge/face dissolve)\n- ~~MESH_OT_dissolve_verts~~ 5f59a8bb11\n- ~~MESH_OT_drop_named_image~~ b4d23818c8\n- ~~MESH_OT_dupli_extrude_cursor~~ 68a719f1ea\n- ~~MESH_OT_duplicate~~ 1ade071052\n- ~~MESH_OT_edge_collapse~~ 2349b9777a\n- ~~MESH_OT_edgering_select~~ 7fa77cd31d\n- ~~MESH_OT_edges_select_sharp~~ dd93c80747\n- ~~MESH_OT_extrude_region~~ 1d7bdbd273\n- ~~MESH_OT_extrude_repeat~~ 9a125ca5fe\n- ~~MESH_OT_faces_select_linked_flat~~ 1bb9ccf887\n- ~~MESH_OT_fill_grid~~ 1cbe3b6e03\n- ~~MESH_OT_hide~~ 5d8937e8ba\n- ~~MESH_OT_loop_select~~ 7fa77cd31d\n- ~~MESH_OT_normals_make_consistent~~ 266638d783\n- ~~MESH_OT_offset_edge_loops~~ efd843b51c\n- ~~MESH_OT_reveal~~ f4c1685a16\n- ~~MESH_OT_screw~~ badd88ca02\n- ~~MESH_OT_select_all~~ (initial patch bfc9d426bb)\n- ~~MESH_OT_select_axis~~ c899f21800\n- ~~MESH_OT_select_face_by_sides~~ 90c9458b2f\n- ~~MESH_OT_select_interior_faces~~ e161c51e7f\n- ~~MESH_OT_select_less~~ b6a7fa1880\n- ~~MESH_OT_select_linked~~ (initial commit) bfc9d426bb\n- ~~MESH_OT_select_linked_pick~~ (initial commit) bfc9d426bb\n- ~~MESH_OT_select_loose~~ eacb2eb423\n- ~~MESH_OT_select_mirror~~ 3ebe389605\n- ~~MESH_OT_select_mode~~ (initial patch bfc9d426bb)\n- ~~MESH_OT_select_more~~ 2309131b4d\n- ~~MESH_OT_select_non_manifold~~ 4da2aec2bb\n- ~~MESH_OT_select_nth~~ 63c2397171\n- ~~MESH_OT_select_random~~ 847f028b71\n- ~~MESH_OT_select_similar~~ 75b2091d42\n- ~~MESH_OT_select_ungrouped~~ 1d8e6c5cd0\n- ~~MESH_OT_shortest_path_pick~~ c32cc3e43e\n- ~~MESH_OT_shortest_path_select~~ 5a431be629\n- ~~MESH_OT_spin~~ ab8dded330\n- ~~MESH_OT_vert_connect~~ 944054fbb6\n- ~~MESH_OT_vert_connect_path~~ fc7316fb54\n", "Geometry instances cannot mix with other instancing\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.1.0 Alpha\nWorked: 3.0 alpha\n\n\n\n[cant scatter.blend](cant_scatter.blend)\n[2021-11-02 19-08-07.mp4](2021-11-02_19-08-07.mp4)\n\n", "Old blend file corrupts to crash blender, I guess\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: AMD Radeon R9 200 Series ATI Technologies Inc. 4.5.14761 Core Profile Context 20.11.2 27.20.14540.4003\n\nBroken: version: 3.0.0\n\nCrashes here and there. RN I witness crush form select/deselect. help plz ;c\n\nI do, tho, have installed two PROBABLY outdated addons BUT in new file crash doesn't seem to happen.\n\nThe problem is - Idk exact steps, but overall it takes to select/deselect all/smth at all few times\n\nSteps\n- Open the provided .blend below\n- Enter edit mode on the main SN4K3Mesh\n- Select-All, De-Select-All (Repeat)\n\n[#94213-asan.txt](T94213-asan.txt)\n\n```\n# Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nbpy.ops.object.editmode_toggle() # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF715993808\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters[0] : 0x0000000000000000\n\tParameters[1] : 0xFFFFFFFFFFFFFFFF\n```\n```\n# Blender 3.0.0, Commit date: 2021-12-02 18:35, Hash f1cca3055776\nbpy.ops.object.editmode_toggle() # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\nbpy.ops.mesh.select_all(action='SELECT') # Operator\nbpy.ops.mesh.select_all(action='DESELECT') # Operator\n\n# backtrace\nException Record:\n\nExceptionCode : EXCEPTION_ACCESS_VIOLATION\nException Address : 0x00007FF71599380C\nException Module : blender.exe\nException Flags : 0x00000000\nException Parameters : 0x2\n\tParameters[0] : 0x0000000000000000\n\tParameters[1] : 0xFFFFFFFFFFFFFFFF\n```\n\n[SN4K3.blend](SN4K3.blend)\n\n[SN4K3.crash.txt](SN4K3.crash.txt)", "Dissolving edge doesn't work as expected\nOperating system: Windows 10\nGraphics card: RTX 2070 SUPER\n\nBroken: 3.2.1\nWorked: not sure\n\n\nDissolving edges doesn't work as expected.\n\n\n\n\n\n[untitled.blend](untitled.blend)", "Blender crashes when calling ArgumentParser.parse_args() if sys.argv is anything other than [\"path/to/blender.exe\"]\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 980 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.18\n\nBroken: version: 3.5.1\n\nBlender crashes when calling `ArgumentParser.parse_args()` if `sys.argv` is anything other than `[\"path/to/blender.exe\"]`\nI discovered this bug while trying to run a script using the `exec()` function. Any script that uses argparse makes blender crash.\n\nYou can run this code\n```Py\nimport argparse\nimport sys\nimport bpy\n\nsys.argv = [bpy.app.binary_path]\nsys.argv.append(\"--some_arg\") # Comment this line and everything works fine\n\nparser = argparse.ArgumentParser()\nargs = parser.parse_args()\n```\n\n", "Mesh is not being deformated correctly via Mesh Deform Modifier (quads, tris behave better)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 2.93.4\n\nI have an Object 1 that I use as a Mesh Deformer target to on Object 2. Both are meshes, of course. Object 1 is parented to an Armature with just 1 Bone.\nWhen I rotate the bone, the Object 2 is deformed incorrectly. It looks like \"waves\" are going through it.\n\nBind Object 2 to and Object 1 with Mesh Deform Modifier. I use Precision of 6 which I think must be enough for such a simple construction. Then Rotate the Main Bone.\nI also added the Camera parented to the Bone, so that you could use view from camera to see changes, relative to the rotation movement itself.\n\nNote: triangulating the cage mesh with a modifier during bind seems to eliminate most of the wiggeling.\n\n[Mesh_Deform_Bug.blend](Mesh_Deform_Bug.blend)\n\n[Mesh_Deform_Bug.mp4](Mesh_Deform_Bug.mp4)", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Blender 2.91 alpha - Debug menu, instant crash\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX TITAN X/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.55\n\nBroken: version: 2.91.0 Alpha\nWorked: 2.90.1\n\nShader compilation is causing an error when the debug menu is invoked on a blank scene or working scene.\n\nOpen the .blend attached file, press F3>search: Debug. Type 22 (eevee GPU speed benchmarking verbose),\nBlender will crash immediately. The same is true on a blank new scene in the general application template.\n\nFile to test: \n44cb6eb461af31ad1209dedb738d76c0.blend\n\n", "Cut brush - curve sculpting\nSee #96259.\n\nThis initial implementation doesn't need yet to have the Knife overlay control (similar to mesh edit knife).", "Regression: Crash with boolean modifier\nOperating system: Fedora Linux 36 (Workstation Edition)\nGraphics card: Nvidia Geforce RTX 3080\n\nBroken: 3.2.1, 3.3 alpha\nWorked: 3.1\n\nCrash with boolean modifier\n\n- Open .blend file\n- In boolean modifier tab -> select `dot cutter` for boolean operation\n[yin_yang-v1.blend](yin_yang-v1.blend)\noriginal file: [F13292564](yin_yang.blend)", "Display issue in model corner when loop cutting\nBroken: 2.79, 2.81\nWorked: -\n\nWhen creating and edge loop around de outside faces it creates in one of the cornes an extrange artifact\n\nCreate an edge loop across the outside faces of the model.\n\nBlender file:\n[loop cut.blend](loop_cut.blend)\n\nArtifact:\n\n", "Blender 3.5.0 Win11 RTX3090 Fatal Crash on activating Cycles\n@deadpin thanks for replying.\n\nDriver Version: 31.0.15.3161 (NVIDA 528.24) DCH Win1164 installed\n\nCycles now working without issue thank u.\n\n_Originally posted by @Jaymi-McManus in /blender/blender/issues/106953#issuecomment-923094_\n\n>>>>>\n\nUpdate: \n\n1. Closed the ticket too soon, after adding some simple color, specular, metallic, changes in BSDF and then navigating between layout and other tabs, fatal crash occurs again\n\n2. Updated NVIDIA drivers to 31.0.15.3161 (NVIDIA 531.61) DHC / Win11 64 the issue still persists\n\nUpdate logs attached.", "Trim tools doesn't work with Dyntopo enabled\nBroken: blender-2.92.0-ef5d6e9c457a-windows64\n\nEnable the experimental \"tools with missing icons\"\nFile ´/ New / Sculpting\nEnable Dyntopo\nChoose the box or trim lasso tool and try to trim the mesh\nDoesn't work", "Mirror threshold too high, values less than 0.00002 are snapped to center\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX TITAN X/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.60\n\nBroken: version: 2.83.3\nWorked: 2.79b\n\nBlender 2.8x has a bug that forces it to snap to the YZ plane with a value less than 0.00002.\nThe above is occurring with the snap off.\nThe bug occurs even if the vertex's X coordinate is moved to more than 0.00002.\nDuplicating the object does not eliminate the bug.\n\n3D Viewport / Edit Mode.\nCreate an object with the following vertices:\n```\nObject-1 = {\n Vertex-A x = 0.00002, y = 0.00005, z = 0.0\n Vertex-a x = -0.00002, y = 0.00005, z = 0.0\n Vertex-B x = 0.000019, y = -0.00005, z = 0.0\n Vertex-b x = -0.000019, y = -0.00005, z = 0.0\n}\n```\n3D Viewport / Edit Mode / Tool / Options / Mirror X [enable].\n3D Viewport / Edit Mode / Snap [disable].\nSelect Vertex-A and Vertex-B.\nMove Vertex-A and Vertex-B.\nBlender 2.8x has a bug that forces it to snap to the YZ plane with a value less than 0.00002.\n\n3D Viewport / Object Mode.\nDuplicate the Object-1 to obtain the Object-2.\n3D Viewport / Edit Mode.\nMirror the vertices of object 2 as if they were object 1. This causes the same problem.\n\nEnlarge the Object-2 sideways.\nMirror the vertices of Object-2. The same problem still occurs. The problem occurs even if the Vertex-B x-coordinate is greater than 0.00002 after magnification.\n\n\n[X-Mirror_bug.blend](X-Mirror_bug.blend)\nThe objects in the X-Mirror_bug.blend file are very small, so zoom in on the view to the maximum size.\n\n\nThis is a 2.79b file with the X-Mirror working correctly:\n[X-Mirror_2.79b.blend](X-Mirror_2.79b.blend)\n\n\nThe difference in x-mirror behavior between 2.8x and 2.79b is shown in the following video:\n[Blender 2.8x BUG_X-Mirror.mp4](Blender_2.8x_BUG_X-Mirror.mp4)\n\n[Blender 2.79b X-Mirror.mp4](Blender_2.79b_X-Mirror.mp4)\n\n\nIf it is unintentionally aligned to the YZ plane like 2.8x, it will not be possible to edit it (if this behavior is a specification, it will be useful if you can implement it as a function separate from the mirror, align it to the YZ plane, and set a threshold to align).\n\n**When importing model data from the outside and editing a mirror, all vertices with an X coordinate of less than 0.00002 are already implicitly aligned to the YZ plane, so editing with an X mirror is no longer possible at all. Hence, until this bug is fixed, edits involving X mirrors should use 2.79b.**\n" ]

[ "Knife tool fails on a non-planar n-gon\nBroken: 2.74 up to and including 2.92 master\nWorked: 2.73 (yes, I bisected that far back ;) )\n\n\nKnife tool fails to cut between verts in n-gons of certain configuration.\n\n\nLoad the attached .blend and try to make a knife cut between the two selected verts (i.e. try to create an edge connecting them).\n\n[knife_fail.blend](knife_fail.blend)\n\n(If you want to test in 2.73 yourself, copy the object from a running instance of 2.9x into 2.73 using ctrl+c/ctrl+v, then replace the mesh in another existing mesh object)." ]

[ "Custom Property list not updating in UI when editing properties from Python\nOperating system: Arch Linux\nGraphics card: GTX 970\n\nBroken: 2.80.60 (1c5860407068)\n\n\nSetting a custom property from Python does not update the corresponding entry under Custom Properties until mouse over forces a refresh.\n\n\nUsing default blend file.\n\n1. Add a Python Console panel somewhere\n2. Select the cube object\n3. Switch to the object properties tab, unfold the Custom Properties panel\n4. In the python console:\n```\no = C.active_object\no['value'] = 456\n```\n5. No new entry is shown in the Custom Properties panel\n6. Move the mouse to the Custom Properties panel -> new value prop appears\n7. The same thing happens when updating a property value, a mouse over is also needed to refresh the property list being displayed\n", "\"Disable in Viewport\" option results in UI for \"Disable in Render\" not updating properly when it is keyframed\nOperating system: Linux-5.10.0-8-amd64-x86_64-with-glibc2.31 64 Bits\nGraphics card: NA\n\nBroken versions: 2.82-3.4\nWorked: Unsure\n\n**Short description of error:**\n\"Disable in Viewport\" option results in UI for \"Disable in Render\" not updating properly when it is keyframed.\n\n**Exact steps for others to reproduce the error:**\nFollow the steps shown in the video above or follow the steps below:\n1. Start with a new file.\n2. Select an object.\n3. In the `Object Properties` tab of the `Properties Editor`, expand the `Visibility` panel.\n4. Inside the `Visibility panel`, animate the the `Show in -> Renders` property. You can do this by inserting a keyframe for it being enabled on frame 1, then go to frame 2 and disable it and keyframe it again.\n5. Once the `Show in -> Renders` property is animated, disable `Show in -> Viewports`. Now if you scrub through your time line, you will notice that the UI for `Show in -> Renders` does not update the `Tick` as you go back to a frame you had keyframes to be enabled. This can also be observed in the `Outliner` as the `Camera` icon does not change from disabled to enabled or vice-versa.\n6. If you enable `Show in -> Viewports`, you will notice the `Tick` and `Camera` icon start updating again as you scrub through the time line.\nHere is a video demonstrating the issue:\n[Disable in viewport breaking disabled in render.mp4](Disable_in_viewport_breaking_disabled_in_render.mp4)", "Segmentation fault when selecting context.area.type before applying operator (scripting)\nOS: Ubuntu 16.04.4 LTS\nHW: Dell XPS 9550 notebook, using Intel HD GPU (nvidia deactivted)\n\nBroken: 2.83\n\nExecuting the operator `bpy.ops.transform.translate` causes a crash if `bpy.context.area.type` is set manually before.\nCrash in `setTransformViewMatrices` because `t->ar->regiondata` is `NULL`\n\nRunning the following script causes the crash:\n```\nimport bpy\nimport bmesh\n\n# Create object\nbpy.ops.mesh.primitive_plane_add(size=1, enter_editmode=False, location=(0, 0, 0))\n\n# Enter edit mode\nbpy.context.area.type = 'VIEW_3D' # If this is not set, crash does not occur\nbpy.ops.object.mode_set(mode='EDIT', toggle=False)\n\n# Modify vertices\nmesh = bmesh.from_edit_mesh(bpy.context.active_object.data)\nmesh.verts.ensure_lookup_table()\nmesh.verts[0].select = True\n\n# Next call causes crash:\nbpy.ops.transform.translate(value=(0, 0, 1), constraint_axis=(False, False, True),\n mirror=False, use_proportional_edit=True,\n proportional_edit_falloff='SHARP')\n```\n", "Cycles: modifying properties through the Python console does not trigger updates.\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.1.0 Beta\n\n\nLoad a scene with a 16k hdri\nwith viewport rendering happening, try and rotate the hdri environment using a mapping node:\n\n\nnow change the world surface setting sampling method to manual, and the map resolution to 1024\n\nmessage will appear:\n\n\nnow rotate the hdri again. It'll be much more responsive.\n\nnow in the python console enter:\nbpy.data.scenes['Scene'].world.cycles.sample_map_resolution = 15360\n\nNotice the message 'updating lights' didn't appear in the viewport.\n\n rotate again, it's still responsive because cycles is still using the 1024 manually entered.\n\nNow restart the viewport render by clicking on solid viewport mode and back to rendered viewport mode\n\nRotate again. Still responsive. Even though the viewport has been restarted, cycles is still using the last manually entered value of 1024.\n\n-------\n\nNOTE: no updates seem to be triggered for the properties. The only update triggered seem to be for the console's current line which is modified after executing the entered line.\n", "Windows Lock Screen (Windows+L) writes a single character 'l' in the Text Editor\nOperating system: Windows 10 Pro 64-bit (10.0, Build 18363)\nGraphics card: NVIDIA GeForce 920M\n\nBroken: 2.83.3\n\nWindows Lock Screen (Windows+L) writes a single character 'l' in the Text Editor.\n\n- Open Blender.\n- Open a Text Editor.\n- Create a new text by pressing the \"+ New\" button.\n- Move your mouse on the Text Editor area.\n- Press Windows + L\n- Login by entering your password if you have.\n- Check that there is a letter l (lowercase of L) in the Text Editor.\n\nI'm a coder with python script and I can no longer use Windows + L, it's dangerous. The thing that makes me confuse is I used Blender Script since a few years, I always press Windows+L all the time, since Blender 2.79 which also have the bug, but I've never remark that bug, but I remark it today. If I understand, my code is full of letters 'l'. I tried to restart Windows but the bug is still there. I updated Windows recently about a couple weeks ago.", "Scene pre/post handlers are called for each main loop iteration\nTitle says it pretty much everything. just execute this in py console:\n\n```\ndef cb(sce): print(\"foo\")\n\nbpy.app.handlers.scene_update_post.append(cb)\n```\n\n…and you'll get endless 'foo' lines in your console, even when Blender is hidden and all.\n\nIssue is, main loop calls (through `wm_event_do_notifiers()`) `BKE_scene_update_tagged()`, which unconditionally fires those pre/post scene update handlers.\n\nNot quite sure what's to be done here really, not familiar at all with this area of the code.", "Viewport display broken for fire- solid and rendered mode (Cycles & Eevee)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.86\n\nBroken: version: 2.83 (sub 0)\nWorked: (optional)\n\n[Fire seems to get 'stuck' in the viewport (as in it does not disappear), it does however behave correctly when F12 rendered.\nThis happens in 2.83Alpha and 2.82Beta (both downloaded as of today, approx 16:00 Sydney time 15 Jan 2020)]\n\n[Bake the attached and observe.\nhere's what it looks like-\n[ViewportFire_2-82+2-83.mp4](ViewportFire_2-82_2-83.mp4)\n\nIt works Ok when F12 rendered as seen here-\n[Fire26Nov20190001-0050.mp4](Fire26Nov20190001-0050.mp4)\nI couldn't be bothered to render out the clip in Cycles but stills show the error is only in the viewport-\n\n\nHere's the .blend-\n[Fire2.83A.blend](Fire2.83A.blend)\n\nThis works properly in the 26November2019 build, here's what that looks like-\n{[F8279386](ViewportFireMantaflow_26-11-19Build.mp4)}]\n\n", "Error when denoising (Optix) large image\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090, 3060 Driver 522.30\n\nBroken: version: 3.3.1\nWorked: Never (3.0 failed, 2.90 did not start rendering)\n\nDenoising large image will fail and may result in crash.\n\nShortened log:\n```\nIllegal address in CUDA queue synchronize (film_convert_float3 filter_guiding_preprocess)\nE1122 18:16:01.843204 7088 denoiser_optix.cpp:470] Error converting denoising passes to RGB buffer.\nIllegal address in CUDA queue copy_from_device ()\n```\n\n[#102515.blend](T102515.blend)\n\nOpen file and render image.", "Crash when rendering animations (GPU Subdivision ON, OFF is fine)\nSystem Information\nOperating system: WIndows 11 22H2\nGraphics card: Nvidia GeForce RTX 4080\n\nBlender Version\nBroken: 3.4.1, 3.5, 3.6\n\nForce closes when rendering animation after a number of frames and when opening larger blend files. Happens with all files. All drivers up to date and tested memory with no issues. Worked without issue for 6 months, crashing started 1 week ago.\n\nRendering animation with GPU Subdivision unchecked did not trigger a crash", "Geometry Nodes related memory leak and exception access violation\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.5.1\nWorked: \n\nBlender freezes, memory usage spikes and eventually you get exception access violation crash.\n\nOpen attached file and enable Layer 3. Blender will freeze, memory usage will spike and eventually lead to exception access violation crash. Crash log is generated.\n\n[gnodes bug1.blend](attachment)\n\nTo stop crashing disconnect Twig Object from Object Info node. I think this started happening after I deleted the \"fir_twig\" object, but there was still a hidden reference in the Object info. However, I was not able to replicate this in a new file and the crash stops happening if I delete / modify some of the node groups, even in unrelated ways. I am pushing Blender to the limits and this is a heavily reduced file.\n\n", "Crash when switching from solid to rendered view\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.6.0 Alpha\n\nI have a scene about 1.7GB in size with 4K textures but renders in 2K using simplify, blender crashes when I open the file > hit F12 and render the scene then close the render windows and switch to render view > it starts compiling then blender closes, in the past it would show out of memory this report: \n\n105522#issuecomment-904934 \n\nThis issue happens only when persistent Data is active\n\nunfortunately I can't share the file at the moment. \n\n\n\n\n", "driver expression not updated in command line rendering and render with lock interface\nOperating system: Macos\nGraphics card: M1\n\nBroken: 3.1.0 release version\nWorked: don't know\n\n\nWhen assign a property with driver expression (such as `#frame`), then write a script to print out the property value at each frame in `bpy.app.handers.frame_change_post`, then render animation with `lock interface`, or command line rendering (which needs to register the script), then the value printed from python code is not updated, it always prints the same value.\n\n- Open attached file\n- Run the script\n- Render animation (check output in console: frame numbers stays 0)\n[untitled.blend](untitled.blend)\n- - -\nTo repro from scratch:\n- Use the default cube, then change its `location.x` to `#frame`, then write a script\n```\nimport bpy\n\ndef fun(scene, despgraph):\n print(\"from python: \"+str(bpy.data.objects['Cube'].location.x))\n \n \nbpy.app.handlers.frame_change_post.append(fun)\n```\n- then render animation with `lock interface`\nThe value printed from python code is **incorrect** and always the same value\n\nFrame value can be **correctly printed**, when render without `lock interface`. \n\nBecause I mostly need it to work in command line rendering, but I found it has the same problem when render with `lock interface`\n\nThe file `untitled.blend` does the same thing as I talked above, once you open it, run the script, then render the animation, you can see the output in console.", "dual monitor will sometimes cause blender to freeze\nOperating system:win10\nGraphics card:2080s\n\nBroken: 3.4.1\nWorked: \n\ni have two monitors (4k and 2k) scrolling timeline or doing anything on one monitor and displaying cycles render viewport on the other monitor will sometimes cause blender to freeze forever\n\nit is not specific to this scene it will freeze on all kinds of scenes, might be related to the muti-monitor bug i have been reporting T96248\nhere is a screenshot of what it looks like when freezing \n\nmake some animation\nhave 2 monitors and use cycles rendered view on one monitor and scrolling timeline on the other monitor", "USD: Crash when loading the Intel Moore Lane demo asset (related to Color Attributes)\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 528.24\n\nBroken: version: 4.0.0 Alpha\nBroken: version: 4.0.0 Alpha\nWorked: version: 4.0.0 Alpha\n\nCrashes during import of the Intel 4004 Moore Lane asset. Seemingly related to Color Attribute processing though the stack traces for the crash are not related.\n\n- Download the \"4004 Moore Lane v 1.1.0\" asset: 4004-moore-lane/\n- Attempt to import the `USD\\4004MooreLane_III_03_ASWF.usd` file with default settings --> observe crash\n- Attempt to import the same file but with the \"Color Attributes\" option unselected --> works\n\nOddly, the crashing stack occurs inside tbbmalloc and while seemingly dealing with mesh normals...\n```\n...\ntbbmalloc.dll!00007ffd7fa29f4e()\tUnknown\nblender.exe!MEM_lockfree_mallocN_aligned(unsigned __int64 len, unsigned __int64 alignment, const unsigned char * str) Line 310\tC\n[Inline Frame] blender.exe!blender::GuardedAllocator::allocate(unsigned __int64) Line 47\tC++\n[Inline Frame] blender.exe!blender::Array<blender::VecBase<float,3>,4,blender::GuardedAllocator>::allocate(__int64) Line 425\tC++\n[Inline Frame] blender.exe!blender::Array<blender::VecBase<float,3>,4,blender::GuardedAllocator>::get_buffer_for_size(__int64) Line 419\tC++\n[Inline Frame] blender.exe!blender::Array<blender::VecBase<float,3>,4,blender::GuardedAllocator>::{ctor}(__int64) Line 122\tC++\nblender.exe!blender::bke::mesh::mesh_normals_loop_custom_set(blender::Span<blender::VecBase<float,3>> positions, blender::Span<blender::VecBase<int,2>> edges, const blender::offset_indices::OffsetIndices<int> faces, blender::Span<int> corner_verts, blender::Span<int> corner_edges, blender::Span<blender::VecBase<float,3>> vert_normals, blender::Span<blender::VecBase<float,3>> face_normals, const bool * sharp_faces, const bool use_vertices, blender::MutableSpan<blender::VecBase<float,3>> r_custom_loop_normals, blender::MutableSpan<bool> sharp_edges, blender::MutableSpan<blender::VecBase<short,2>> r_clnors_data) Line 1368\tC++\nblender.exe!blender::bke::mesh::mesh_set_custom_normals(Mesh * mesh, float[3] * r_custom_nors, const bool use_vertices) Line 1642\tC++\nblender.exe!blender::io::usd::USDMeshReader::process_normals_face_varying(Mesh * mesh) Line 797\tC++\nblender.exe!blender::io::usd::USDMeshReader::read_mesh_sample(blender::io::usd::ImportSettings * settings, Mesh * mesh, double motionSampleTime, bool new_mesh) Line 853\tC++\nblender.exe!blender::io::usd::USDMeshReader::read_mesh(Mesh * existing_mesh, USDMeshReadParams params, const char * * __formal) Line 1094\tC++\nblender.exe!blender::io::usd::USDMeshReader::read_object_data(Main * bmain, double motionSampleTime) Line 249\tC++\nblender.exe!blender::io::usd::import_startjob(void * customdata, bool * stop, bool * do_update, float * progress) Line 288\tC++\n\n```\n\n", "Error: Illegal address in CUDA queue copy_from_device\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.6.0\n\nWhen rendering in GPU mode, Render crashes.\n\nAttached Blend File\nAny frame after 1150 crashes on render with the title error for us.\nTry rendering frame 1160 - This one crashes for me all the time.\n\n" ]

[ "Thread race condition with lock interface enabled\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.92\n\nBroken: version: 2.93.2 Release Candidate\n\n**Description**\nUI redraw is still taking place in the render_init and render_complete handlers when use_lock_interface = True. This lock was only implemented for 3D viewport, but this is not mentioned in any of manuals.\n\nAdding lock to node editor seems to resolve the issue:\n\n```\ndiff --git a/source/blender/editors/space_node/space_node.c b/source/blender/editors/space_node/space_node.c\nindex ff848a7bb95..9813b8aaa86 100644\n--- a/source/blender/editors/space_node/space_node.c\n+++ b/source/blender/editors/space_node/space_node.c\n@@ -1073,6 +1073,7 @@ void ED_spacetype_node(void)\n art->cursor = node_cursor;\n art->event_cursor = true;\n art->clip_gizmo_events_by_ui = true;\n+ art->lock = 1;\n\n BLI_addhead(&st->regiontypes, art);\n```\n\nTrace: \n```\n>\tblender.exe!UI_block_end_ex(const bContext * C, uiBlock * block, const int * xy, int * r_xy) Line 1873\tC\n\n \tblender.exe!node_draw_basis(const bContext * C, const View2D * v2d, const SpaceNode * snode, bNodeTree * ntree, bNode * node, bNodeInstanceKey key) Line 1733\tC++\n \tblender.exe!node_draw_default(const bContext * C, ARegion * region, SpaceNode * snode, bNodeTree * ntree, bNode * node, bNodeInstanceKey key) Line 1943\tC++\n \t[Inline Frame] blender.exe!node_draw(const bContext *) Line 2002\tC++\n \t[Inline Frame] blender.exe!node_draw_nodetree(const bContext *) Line 2059\tC++\n \tblender.exe!draw_nodetree(const bContext * C, ARegion * region, bNodeTree * ntree, bNodeInstanceKey parent_key) Line 2097\tC++\n \tblender.exe!node_draw_space(const bContext * C, ARegion * region) Line 2235\tC++\n \tblender.exe!ED_region_do_draw(bContext * C, ARegion * region) Line 569\tC\n \tblender.exe!wm_draw_window_offscreen(bContext * C, wmWindow * win, bool stereo) Line 726\tC\n \tblender.exe!wm_draw_window(bContext * C, wmWindow * win) Line 868\tC\n \tblender.exe!wm_draw_update(bContext * C) Line 1067\tC\n \tblender.exe!WM_main(bContext * C) Line 654\tC\n \tblender.exe!main(int argc, const unsigned char * * UNUSED_argv_c) Line 563\tC\n\n```\n\nOpen below file, run the script and then press f12.\n\n[ui lock ignored on some handlers which are not on the main thread.blend](ui_lock_ignored_on_some_handlers_which_are_not_on_the_main_thread.blend)\n\n" ]

[ "Optix Viewport Denoiser Makes Objects Glow When Using High Exposure Even When There is no Light in the Scene\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce MX130/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.90.0 Alpha\n\nThe latest build added the feature to let non-RTX GPUs use the denoiser, so I tested it on my scene right away. However, the objects in the scene start glowing when the denoiser starts working. The scene was lit using the Physical Lights feature that the Photographer 3 addon has provided, which offers units like Lumen and Candela, etc. And since physically correct light values can sometimes be very dim (like a candle = 1 candela), sometimes be very bright (like the Sun at noon 111,000 lux), Photographer offers real life camera's settings for modifying the exposure in the color management in the render settings to adjust the overall brightness. And in my scene, I used a candle, which is very dim. Therefore the exposure has been set very high. It has been fine until the denoiser came. \n\nBased on the default startup\n1. Delete the default point light\n2. Turn the World Background Strength to 0\n3. Turn the Exposure in the Render Settings under Color Management to 10\n4. Change the render engine to Cycles\n5. Turn on render view, you should see nothing but complete darkness\n6. Turn on the Optix Viewport denoiser\n7. Wait for a while and see the default cube glowing\n\nI don't know if this is a bug or not, if it is not, please at least get some work around. Like making the denoiser refresh when the exposure changes or something?\n\n", "Animation: Exiting NLA tweak mode on empty action corrupts entire NLA track\nOperating system: Windows 11\nGraphics card: NVIDIA RTX 3060 Ti\n\nBroken: 3.3.0, 3.5.0\nWorked: Unknown\n\n\nWhen tweaking a strip on the NLA, if the user deletes all the keyframes of the action being tweaked, when the user exits tweak mode, all other NLA strips on that track will be moved in an unexpected and undesirable way.\n\n[tweakbug.mp4](tweakbug.mp4)\n\nBased on the default startup or an attached .blend file (as simple as possible).\n1. Open the attached .blend file\n2. Delete the keyframes on the track being tweaked.\n3. Exit tweak mode by pressing Tab.\n4. Observe that the NLA strips have been moved.\n\n[tweakbug.blend](tweakbug.blend)", "Baking animation of bones affected by a curve (spline IK) is wrong\n**System Information** \nOperating system: windows 10\nGraphics card: gforce 1060 6gb\n\nBroken 2.92 alpha\nBroken 2.83\nWorked: likely never\n\nI have a row of bones, I add the curve modifier, I make an animation with that modifier on, I bake the animation of the bones to be able to send it to unreal or unity and the animation breaks completely\n\n[CurveBreackExample.blend](CurveBreackExample.blend)\n- Open file\n- inspect wrong bake of bone action\n- to reproduce:\n- go to object mode\n- select `SkeletonBase` object\n- go to pose mode (all bones should be selected)\n- Pose > Animation > Bake Action... with following settings:\n", "Displacement baking with Cycles\n", "Alpha doesn't work as a mask \nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.79\n\nBroken: version: 3.0.0\n\nI use an alpha channel as a mask for mix node. If I put an alpha channel from the same node where I take a color channel I get a strange result but if I grab a mask from an alpha channel a copy of the same texture node I get the proper result. \n\n\n\n\n[bug_11.blend](bug_11.blend)\n\n", "Cycles Render: Volume (Before Baked Simulation Starts) still trying to be rendered\nOperating system:windows 11\nGraphics card:nvidia geforce rtx 2060\n\nBroken: 3.4.1\n\nThe cycles render tries to use the volume cache to render if the frame is less than the start of the simulation. See video example.\n\n1. Default cube: Add Quick Smoke.\n2. Cache settings:\n a. Type: All.\n3. Start frame: 50, end: 100.\n4. Bake.\n5. Start render on 30~ frames.\nExpected: Simulation chunks will appear in different parts of the domain.", "When using mask modifier, particles get also generated from the origin\nOperating system: macOS-10.14.4-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 750M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-12.0.23 355.11.10.50.10.103\n\nBroken: version: 3.0.0\n\nI'm emiting particles from an object. I use the mask modifier so that the particles are generated only in some areas. Particles will be generated from the origin, even if there's no geometry there.\n\nPlay the animation from the attached file. You will see a column of particles coming from the origin. \n\n[Bug.blend](Bug.blend)\n", "Disk Cache combined with Cache Step does not work\nOperating system:\nGraphics card:\n\nBroken: blender-2.80.0-git.8d95dcd87dbb-windows64\nWorked: Blender 2.79 b f4dc9f9d68b\n\nBaking a particle system with *Disk Cache* checked and *Cache Step* set to greater than 1 results in the baking failing.\nNo error is reported, but the bake is obviously corrupt.\nIt worked in 2.79.\n\n[bake.zip](bake.zip)\n1. Extract the archive and open the *.blend* file. (in 2.80)\n2. Playback works at this point. (Particles fall down from the plane.)\n3. Click *Delete Cache* and then *Bake*.\n4. Try playback again. It's not incorrect. (Particles appear at the bottom and don't move. If you check the cache folder, you will see it cached every second frame---instead of every 10th as configured, except for the first step.)", "Cycles: Baking to an externally loaded image fails to update the image buffer results in Cycle's Rendered Viewport Shading\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.59\n\nBroken: version: 3.3.0, 2.93\nWorked: probably never\n\nBaking to an externally loaded image will not changed the baked image as displayed in the Viewport Shading's Rendered Mode.\nThe newly baked image buffer is shown correctly in the image editor and (EEVEE's) Material Preview, but wrong in Cycles's render view.\n\n- Open the attached .blend file.\n- Open the attached baked image in the image texture node.\n- Use it to bake any type of pass to the loaded image.\n- Observe that the image looks unchanged in the Rendered View by connecting the image texture node to the material output.\n[#98063.blend](T98063.blend)\n\n[2022-05-12 09-12-20.mp4](2022-05-12_09-12-20.mp4)", "Blender baking's margin overlap if multiple meshes are selected\nBasically just a duplicate of T51134 that Aaron Carlisle asked me to send as a new report.\n\n\nOperating system: Linux-5.4.0-58-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: AMD RAVEN (DRM 3.35.0, 5.4.0-58-generic, LLVM 10.0.0) X.Org 4.6 (Core Profile) Mesa 20.0.8\n\nBroken: version: 2.90.1\nWorked: never with Cycles as far as I know.\n\nCycles overlap the region margin on other islands in the final result.\n\nCycles will bake an object/island, apply the margin, then proceed to do the same to the next one on top of the previous and they will overlap if the margin is bigger than the distance of the UVs, instead of applying to the finished bake and having the margins having an equal distance between islands, as Blender Internal used to do.\n\nJust press bake on the file. Or any multiple objects with a margin bigger than the UV islands.\n\nAs far as I know this has always been a problem with the Cycles baking implementation, besides some other issues of performance and usability. For instance if baked with multiple objects selected it is much, much slower than if I collapse them and merge, as I mentioned in the previous bug report.\n\nI understand it should be slower if they are separated, but I don't imagine it should be that much slower.\n\n\n\n", "Baking diffuse color takes a lot of time in version 3.0\nOperating system: Windows 10 64bits\nGraphics card: NVIDIA GeForce RTX 2060\n\nBroken: 3.0, 3.0.1, 3.1\nWorked: 2.93.8\n\nIt takes longer time to complete bake diffuse color in version 3.0 compared to 2.93\n\n\n\n[Bake 2.93.blend](Bake_2.93.blend)\n\n[Slow_Bake 3.0.blend](Slow_Bake_3.0.blend)\n", "Film -> Pixel Filter settings are not used when baking textures\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: AMD Radeon R9 200 Series ATI Technologies Inc. 4.5.14831 Core Profile Context 21.5.2 27.20.20903.8001\n\nBroken: version: 3.5.0\n\n\nPixel Filter settings are not used when Baking a Texture, changing them does not effect Baking in any way. The width seems to be fixed at a value of around 0.1\nHad a quick look at the github and I'm guessing they aren't passed to the render when Baking.\n\n\nRender any texture with sub pixel details to a texture, see attached file as example\n\n- Enable viewport render preview\n- Adjust Film->Pixel Filter value, it's effective\n- Bake the texture\n- No matter how you adjust Film->Pixel Filter value, it's not effective on the baked result\n", "Baking of normals show different results between versions\nOperating system: Win 10\nGraphics card: Geforce 3060\n\nBroken: 3.5.1\nWorked: 3.4.1\n\nThis changed in 75ad8da1ea6799e3983e4146f4b20dc4bfbb8415\n\n\nBaking of normal map differnet between Blender versions\n\nThe Attached blend file will output differnt result when trying to bake the normals.\nThe Mesh is part of a comercial mesh I use, therefore I cannot share the whole mesh, but in the blendfile you just need to bake the normals in blender 3.4.1 and then in 3.5.1\nThe results are differnet.\n\nBlender 3.5.1 seams to have some bug in the baking or handles the baking differnt.\n\nEdit:\nSeams like the Auto-Smooth of meshes bauses this difference between the versions", "Blender Crash. Code marked as unreachable has been executed.\nOperating system: Windows-11-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 536.67\n\nBroken: version: 3.6.1, 4.0.0 Alpha\nWorked: didn't test\n\nOpen the .blend and go to frame 60. Crash\n\nThis is a .blend from a random user. So, sadly, I don't know the exact steps to reproduce.\nIt just said: \"Code marked as unreachable has been executed. Please report this as a bug.\"\n\n", "Crash with Boolean modifier\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.13\n\nBroken: version: 3.4.0 Alpha\nWorked: 3.1.2 (appears to have been broken since 3.2.0)\n\nBlender crashes when performing an Intersection Boolean between a cube and a bunch of loose faces. (Granted, this is probably not something you should do in the first place, but at least it should be limited to producing a bad result, not a crash.)\n\nI didn't find a crash log, but it's easy to reproduce.\n\n* Load the attached .blend file.\n* Scrub the timeline to frame 209. Blender crashes.\n\n[crash.blend](crash.blend)\n" ]

[ "Regression: black seams in my Cycles bakes (caused by denoising) due to margin pixel extension darkening\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.92\n\nBroken: version: 3.0.0\nWorked: 2.93.7\n\nI get black seams when I bake textures that I never used to have before upgrading to 3.0. They occur on any corner that I UV as shown here: 4773f8c541316a64af42092a919c3e64.png\n\n[Black Seams.rar](Black_Seams.rar)\nUV a mesh with only 1 split on the corners leaving the other sides still attached. Bake a combined texture to 2048x2048 resolution with Cycles. Apply to mesh. Observe the black seams on the corners.\nNow open that same file and bake that same mesh in 2.93.7. No black seams. Re-save file in 2.93, and re-open that file in Blender 3.0, and it now bakes without a black seam too.\n\n" ]

[ "2.80 Pose Breakdowner doesn't work in Add NLA Strips\nOperating system: ubuntu 18.04\nGraphics card: gtx1080ti\n\nBroken: (2.80)\n\nPose Breakdowner doesnt work when editing NLA Tracks \n\n\nOpen attached file and use Pose Breakdowner\n[broken .blend](broken_.blend)", "Performance issues when moving bones of an armature when the Outliner is set to Blender file\nBroken: 2.93\n\nThis happens on heavy scenes and with many linked files, if you have an Outliner opened and set to \"Blender file\".\nThe fps drops dramatically only when trying to move a bone of an armature. It doesn't matter the complexity of said armature.\nThe problem stops happening after closing the Blender file Outliner.\nCould it be related to this other fix? T87842", "Crash on deleting driver\nOperating system: Linux neo 5.10.68-1-MANJARO #1 SMP PREEMPT Wed Sep 22 12:29:47 UTC 2021 x86_64 GNU/Linux\nGraphics card: nvidia GTX 970\n\nBroken: (\n```\n3.0.0-alpha+master.a3027fb09416-linux.x86_64\n```\n)\nWorked: (?)\n\nI duplicated a Text strip via <shift><D>. I don't know where that driver on Location X came from. The original didn't have one. I then wanted to delete that driver. After deleting it and moving the current frame blender crashed.\n\n- Load attached file\n- Delete driver of 'Location X'\n- Press Cursor keys Left/Right a few times till crash.\n[crash.blend](crash.blend)", "Redo panel disappears during edits in Tweak tool\nOperating system: Win10 x64\nGraphics card: Nvidia 1070\n\n**Blender Version** \nblender-2.80.0-git.26d5a3625ed-windows64\n\nThe info panel for move disappears in edit mode if tweak mode is in action. This does not seem to happen if one uses \"g\" for grab instead.\n\n\n- Enable left click select, optionally select all toggles\n\n- In your space bar menu turn the select mode (W) on\n\n- Create a mesh and move into edit mode \n\n- Start grab and drag of verts and watch the move info panel at the left bottom\n\n- Now try the same thing with selecting a vertex then press \"g\" to move.\n\n", "Eevee Depth of field creates white outline\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.11\n\nBroken: version: 2.93.2 Release Candidate\n\nwhite outlines appear on out of focus objects\n\n\n- open attached file and turn to rendered mode\n\n[dof bug recovered.blend](dof_bug_recovered.blend)\n", "Mode locked and crash if objects become disabled in viewport while not in object mode\nOperating system: Linux-5.15.0-2-amd64-x86_64-with-glibc2.33 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.86\n\nBroken: version: 3.1.0 Alpha\nWorked: didn't crash but showed an error message in 2.91.2\n\nObjects that are in a mode which is not Edit Mode nor Object Mode, will lock the mode setting (so the mode options are greyed out and unusable) if they get disabled in the viewport. Hiding will not cause the this bug, it needs to be disabled in viewport using that display icon in the outliner, not the eye.\n\nThe solution for this is probably to make the disable in viewport action switch the mode to object mode if the active objects gets removed (much like delete from outliner). Another solution would be to still allow the mode switch.\n\nMuch worse though! Blender crashes sometimes when it gets into this state. I can reliably crash blender as described below.\n\nGo into weightpaint mode of the default cube and disable it in the viewport using the outliner (you need to enable the option first). Now you are stuck and have to enable it in the viewport first to be able to edit anything.\n\nNow restart blender etc. and try this:\n1. Go into Vertex Paint Mode of the default cube\n2. Hit Tab to go into Edit Mode\n3. Disable the cube in the viewport using the outliner as above\n4. Hit Tab to switch out of Edit Mode -> Crash", "Delayed update of outliner ui on change of active collection\nBlender 3.6\n\nWhen you change active collection outliner UI updates only after you leave the outliner.\nAttached short demo gif.\n\nWay to reproduce the issue:\n1) Start default blender scene.\n2) Add new collection \"Example\"\n3) Run the script below in the text editor.\n4) Note that above the outliner you can see currently active collection, by default it's \"Collection\"\n5) Try to select \"Example\" instead of \"Collection\" - it won't update in the ui above.\n6) If you leave the outliner it will update the active collection.\n7) You can also try to choose some other collection and run operator to `.tag_redraw()` - F3 - Update Outliner. Note that info message that says that active collection didn't changed, it's just that outliner area wasn't updated.\n\nWhat expected - outliner ui to update as you change active collection.\n\n```python\nimport bpy\n\ndef get_outliner_area(context):\n screen = context.screen\n for area in screen.areas:\n if area.type == \"OUTLINER\":\n return area\n\nclass EXAMPLE_OT_custom_collection_new(bpy.types.Operator):\n bl_idname = \"example.redraw_outliner\"\n bl_label = \"Update outliner\"\n def execute(self, context):\n before = context.collection.name\n area = get_outliner_area(context)\n area.tag_redraw()\n self.report({\"INFO\"}, f\"{before} -> {context.collection.name}\")\n\n return {'FINISHED'}\n\ndef context_draw_outliner_foobar_buttons(self, context):\n layout = self.layout\n layout.label(text=context.collection.name)\n layout.label(text=context.view_layer.active_layer_collection.name)\n layout.operator(\"example.redraw_outliner\")\n\ndef register():\n bpy.utils.register_class(EXAMPLE_OT_custom_collection_new)\n bpy.types.OUTLINER_HT_header.append(context_draw_outliner_foobar_buttons)\n\ndef unregister():\n bpy.utils.unregister_class(EXAMPLE_OT_custom_collection_new)\n bpy.types.OUTLINER_HT_header.remove(context_draw_outliner_foobar_buttons)\n\nif __name__ == \"__main__\":\n register()\n```\n\n", "Wireframe opacity slider does not work.\nOperating system: Windows 10\nGraphics card: gtx1060\n\nBroken: 3.1\nWorked: \n\nWireframe opacity slider does not work if smooth wires is turned off.\n\nopen preferences>viewport\ndisable smooth wire overlay\nsave preference\nset viewport to wireframe shading\nin the overlay panel in viewport drag the wireframe opacity slider, the wire color will fade to black, the opacity won't change.", "Hitting 'Escape' while editing a shortcut enters that key on Windows, while on linux it cancels edition\nWindows 7 64bits, nvidia gt540m CUDA 2GB, Intel i7 2GHz, 8GB ram\n\nBroken: 2.75 c6b042b\n\n\nWhen i do a search of a shortcut, and then if i want to assign a new value and then cancel, the tab disappear.\n\nBlender User Preferences > Input > Search by key-binding \"Ctrl Tab\" > Show key map and property of any tab > Clic on \"type of event\" to change the value > Press Esc to cancel > the tab disappear.\n", "Sculpt Mode Border Mask Tool Todo\n## Sculpt Mode Border Mask Tool Todo\n * - [x] {icon circle color=red} When this tool is selected, it currently shows the wrong tool settings which affect the brushes, not this tool. These should not be visible when this tool is active.\n * - [ ] {icon circle color=yellow} We should add a tool setting to add/remove the mask as you drag. Currently, the only way to remove from mask, is very hidden: You drag with RMB. This could become more consistent and more discoverable if we added the same Add/Remove buttons as with other Sculpt tools", "Outliner doesnt preserve selected or active state when clicking mesh twice (objectmode -> editmode -> objectmode)\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.86\n\nBroken: version: 2.81 (sub 3)\nWorked: (optional)\n\nExport selected object export empty file if object have only mesh selected in outliner.\n\nProbably this is more UX bug than exporters problem.\nCreate default object: Cube or Suzanne.\nExpand Object tree and select mesh in outliner. Blender will switch to mesh edit mode.\nClick one more time to mesh in outliner. BLender will switch to Object editing mode.\nViusually Object still looks selected because mesh still highligted but in 3D viewport object is unselected.\nNow try export as OBJ, enable Export Selected only.\nBlender correctly will export empty OBJ file because there is no any object actually selected.\n\n\n\nThis UX bug quite confusing, and i catch it almost every time i need export files.\n", "When Object is not selectable it can still be switched to edit mode\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.24\n\nBroken: version: 3.5.0\nWorked: dont know\n\nWhen a object is disabled for selection in the outliner and the object is then selected in the outliner its possible to enter any of the interaction modes. \nI would assume if a object is not selectable this should not be happening. \nNote: it only happens with single object selections. \nMaybe this is by design, it just felt like a strange thing and i thought i report it just in case.\n\n1. Create a object\n2. disable the object selection in the outliner \n\n3. Choose your interaction mode \n\n4. The mode changes to the one you selected \n\n", "Click selecting a keyframe in the Dope Sheet deselects all channels\nOperating system: Linux Mint\nGraphics card: RTX 3070\n\nBroken: latest master\nWorked: checked until 2.93 and it still was broken\n\nWhen click selecting a keyframe on a summary channel in the Dopesheet (and the Timeline which is basically a Dopesheet),\nit deselects all channels.\nIf you drag select it, all channels of the keys are selected.\n\nClick selecting a keyframe that is not on a summary channel does select the channel as well.\n\nI would expect click and drag select to behave the same on summary channels.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n* open the provided blend file (channels are already selected)\n* click select a keyframe on a summary channel in either the Timeline or the Dope Sheet\n* note the channels are all deselected\n\n", "Operators to add to the Outliner\nCurrently the Outliner allows users to fully manage Collections. You can add them, remove them and manage them fully.\n\nHowever, for other data this is very limited currently. Many users would like to be able to use the Outliner to do many more basic management tasks. I think we should add many more basic things here that are related to adding, removing, copying, duplicating and so on.\n\n\n\n\nThese features we can add to the Outliner context menu system:\n\n**Objects:**\nCurrently you can perform these operations in the Outliner for objects.\n - - [x] Delete\n - - [x] Rename\n - - [x] Copy/Paste\n\nIn addition, it would be useful to add these commands too:\n\n - - [ ] Add New (with full list of primitive types)\n - - [ ] Duplicate\n - - [ ] Make Local\n - - [ ] Make Proxy\n - - [ ] Add Constraint\n - - [ ] Add Modifier\n\n\n**Constraints & Modifiers**\nCurrently you can:\n - - [x] Delete\n - - [x] Rename\n - - [x] Toggle visibility, but only in the context menu, strangely\n\nIn addition, we should add:\n - - [ ] Visibility toggles\n - - [ ] Re-ordering (by dragging)\n - - [ ] Transfer modifiers & constraints from one object to another (by dragging)", "Design issue: Icon next to overridden property does not perform the function it strongly suggests\n3.0.0\n\nWhen a propertry is overridden, the \"animate property\" button next to it is replaced by this \"reset\" icon:\n\nThese really seem to suggest that clicking them will remove the override. Instead it just adds a keyframe as usual. This seems misleading.\n\n - Override a property of some library override asset\n - Click the button to the right of that property\n\n" ]

[ "Outliner: Delete or X key don't work on objects\nOperating system: Linux-3.10.0-957.10.1.el7.x86_64-x86_64-with-centos-7.6.1810-Core 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.56\n\nBroken: version: 2.81 (sub 8)\n\nHitting delete or x key on objects does nothing, but works with collections.\nEven if I assign a shortcut for it.\nMay be it's a TODO though...\n\nTry to delete an object in the outliner with a key.", "Outliner: support delete in the keymap for different data-types\nOperating system:\nWindows 10 Enterprise 1903\n\nGraphics card:\n - RTX 2080 TI 11GB GDDR6\n - GTX 1080 TI 11GB GDDR5\n\n```\n Nvidia Driver 430.86, Studio Driver, SLI Disabled\n```\n\nBroken: blender-2.80rc2-windows64\nWorked: (optional)\n\nI want to be able to delete and item in the outline by hitting delete or backspace. But it does not work for some reason. I end up having to use the right click menu to delete. I even tried to make sure the delete shortcut keys were enabled for the outliner and when I re-setup the shortcut it sees the delete key in prefrences. But it still does not work in the outliner. Deleting with the delete key works fine in the 3D window.\n\nNote... I am using the Industry Compatible key map.\nCreate a cube. Click on cube in outliner and press delete. Cube does not delete as expected.\n\nI admit that I am very new to blender... but this seems like it should work similar to other 3D apps I use.", "IC Keymap: Delete Objects from Outliner Shortcut broken\nOperating system: Windows-10 - 64 Bits\nGraphics card: Radeon RX Vega ATI Technologies Inc. 4.5.13586 Core Profile Context 20.1.3 26.20.15015.1007\n\nBroken: version: 2.83 (sub 5)\nWorked: perhaps long time ago.\n\nHaving chosen the IC Keymap there are shortcuts mapped to \n*outliner.object_operation* - *Delete* to backspace and delete with press. These can't be triggered.\n\nSo deleting collections work but deleting objects isn't reachable\n\n**Possible Solution:**\n\nChanging their triggering method to \"Release\" works\n\n" ]

[ "Sculpt crash in certain scenario (PBVH vert number does not match)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.76\n\nBroken: version: 2.93.6\n\nCrash upon altering the model in any way.\n\n1. Open file\n2. Use a tool/brush to alter the model.\n3. Instant Crash\n\nI have updated GPU drivers and the latest Blender version as of today, still crashes.\n\nCould not find any duplicate bugs of crashes when interacting with models, only crash on startup and others. \n\nI expect my GPU is just too trash.\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n[blender_system_info.txt](blender_system_info.txt)\n\n[Goblin.crash.txt](Goblin.crash.txt)\n\n[Goblin.blend](Goblin.blend)\n", "Geometry nodes crash involving point translation and points to volume\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 XT ATI Technologies Inc. 4.5.14830 Core Profile Context 21.4.1 27.20.21002.112\n\nBroken: version: 3.0.0 Alpha\nWorked: Unknown.\nCannot open reference file in stable (crashes immediately). \nMy oldest 3.0 from 2021-08-01 (`836aeebf7077`) also has the issue.\n\nBlender crashes in certain setups involving point translation and points to volume.\n\n- Open .blend file\n- Connect \"Crash\" node with Reroute\n\nUpon some more testing I found that there are more specific values or ranges that cause the problem, when using the power node one is around 2 to the power of 24.003.\nKnowing values that crash, the attribute fill node can be used:\n\n\n[attribute-fill.blend](attribute-fill.blend)\n[attribute-power.crash.txt](attribute-power.crash.txt)", "Blender 3.6 problem by rendering scene on EEEVEE (Freeze and crash GL_INVALID_FRAMEBUFFER_OPERATION : Cubemap Workaround End9)\nOperating system: Windows 11 Home\nGraphics card: Radeon Graphics\n\nBroken: (example: 3.6, as found on the splash screen)\n\n\nBlender 3.6 doesn't render the scene as blender demo file in EEVEE\nOpen Blender splash fox of blender 2.90\nRender the scene\nFreeze and crash of blender when rendering on EEVEE\n\n", "Blender just crashes for no reason, and the crash report is empty. (Segmentation fault)\nOperating system: Arch Linux x86_64\nGraphics card: AMD Radeon RX 6900 XT\n\nBroken: (example: 3.6.2, ffe93138751e, master, 2023-08-31, as found on the splash screen)\nWorked: (No idea when this error started happening.)\n\nBlender does not start up. When I start it via the command line, I get this error:\n```\nWriting: /tmp/blender.crash.txt\nSegmentation fault (core dumped)\n```\n\n`/tmp/blender.crash.txt`:\n```\n# Blender 3.6.2, Commit date: 2023-08-14 14:05, Hash ffe93138751e\n\n# backtrace\n\n# Python backtrace\n```\n\nTrying to open Blender.\n\n", "Blender crashes with Noise Texture Node + Hair Particle System + Eevee\nOperating system: Darwin-19.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon RX 480 OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.10.18\n\nBroken: 8dc95f61f3\nBroken: version: 2.92.0\nBroken: version: 2.91.2\nWorked: version: 2.83.12\n\nBlender 2.92 with AMD crashes with **Noise Texture Node**and **Hair Particle System** with render engine Eevee.\nWith Blender 2.91.2 with AMD it doesn't crash but colors are incorrect.\n\nWith 8dc95f61f3 with Intel GPU, there's no crash but colors are incorrect in the same way as in AMD - 2.91.2.\n\n1. Open attached .blend file\n2. Viewport shading to: Render \n3. Play animation by tapping space bar\n\n[particle_loop.blend](particle_loop.blend)\n[blender_2_92_crash.txt](blender_2_92_crash.txt)\n\n**Blender 2.92**\n[blender_2_92.mp4](blender_2_92.mp4)\n\n**Blender 2.91**\n[blender_2_91_2.mp4](blender_2_91_2.mp4)\n\n**Blender 2.83.12**\n[blender_2_83_12.mp4](blender_2_83_12.mp4)", "Blender crashes as soon as an image texture is plugged in the material BSDF node\nOperating system: Windows 10\nGraphics card: NVIDIA GeForce GTX 1080 Ti\nSee attached log for more detail\n\nBroken: 3.5.1, but also 4.0.0 Alpha, did not try intermediate versions\nWorked: ??? The earliest available build, 2.93.18, also causes the same crash\n\nThe second any kind of image texture is plugged in a material, Blender crashes without any warning. Occurs when creating a new image texture, when plugging an existing file, or even when trying to plug a UV grid for unwraping purposes\n\n**Similar tickets**\n#76193\n#76377\n\n1. Start Blender.\n2. Create new default file.\n3. Select the default cube.\n4. Switch to Shading workspace.\n5. Add Image Texture node.\n6. Create new image or select an existing one.\n7. Plug Color output to the Base Color input of the Principled BSDF node.\n8. Watch Blender shut down.", "Blender crashes when click on shading tab \nOperating system: Windows 10, 64-bit\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 4.5.0 NVIDIA 445.87\n\nBroken: \n - version: 2.81 (sub 16), branch: v2.81_octane2019 (modified), commit date: 2020-02-11 11:57, hash: 81ff79b8cdb7, type: build date: 2020-02-11, 12:22:28\n - version: 2.90 (sub 0), branch: master, commit date: 2020-04-18 16:48, hash: 1f1520a04503, type: Release\n\nBlender crashes when clicking on the shading tab. It closes and goes back to the desktop.\n\nSee short description.\n\n[blender_system_info.txt](blender_system_info.txt)\n[blender_debug_output.txt](blender_debug_output.txt)", "\"RENDER_OT_render\" called with \"EXEC_DEFAULT\" freezes Blender with Eevee motion blur enabled\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.90.0 Beta\n\nbpy.ops.render.render() called without 'INVOKE_DEFAULT' from the Python Console or script freezes Blender when Eevee motion blur is enabled.\nAnd the same issue is in the 2.91 Alpha version.\n\n----\n\n- Open Blender\n- Make sure context scene render engine is Eevee\n- Turn on motion blur\n- Open Python console\n- type in bpy.ops.render.render() and hit enter\n\n----\n\n*Edit* (added method that doesn't activate via Python).\n\n**Exact steps to reproduce the error without activating from Python.**\n- Open Blender\n- Make sure context scene render engine is Eevee\n- Turn on motion blur\n- Run this script in the text editor\n\n\n```\nimport bpy\ndef menu_func_import(self, context):\n layout = self.layout\n layout.operator_context = 'EXEC_DEFAULT'\n layout.operator(\"render.render\", text=\"Blocking Render\")\nbpy.types.TOPBAR_MT_render.append(menu_func_import)\n```\n\n- Run \"Render -> Blocking Render\" from the top-bar.", "Cycles - AO Shader node causes infinite \"loading render kernels\" and crashes on return to solid mode view\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 511.09\n\nBroken: version: 3.2.0\n\nAO Shader node causes infinite \"loading render kernels\" and crashes on return to solid mode view\n[AO BUG Test.blend](AO_BUG_Test.blend)\nAttaching Sample file.\nJust render the scene in viewport (doesnt render, just \"loading render kernel\" message)\nTry to go back to solid view - results in freeze/crash\n\nregards\n", "Eevee \"Shutter\" setting in motion blur render settings not animatable\nOperating system: Win 10\nGraphics card: RTX 3090\n\nBroken: 2.91.0\nWorked: never worked (as far as i know)\n\n\nEevee \"Shutter\" in motion blur render settings not animatable\nIn Cycles it ***is***animatable.N/A", "Crash using blender_debug_log.cmd and \"Reload Scripts\" in Cycles preview render\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.91\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\nLoading Blender with blender_debug_log.cmd, it crash immediately when I try to reload the scripts if Cycles is running. This is not happening when launching blender.exe \n\nLaunch blender 2.8 using blender_debug_log.cmd.\nIn the view 3d, choose Cycles as render preview.\nIn the search bar, look for \"Reload Scripts\" and click on it.\nBlender Crash.\n[blender_system_info.txt](blender_system_info.txt)\n\n[blender_debug_output.txt](blender_debug_output.txt)\n\n", "Regressoin: Bisect: Unstable display when UV editor is open\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.49\n\nBroken: version: 3.4.0 Alpha\nWorked: version: 2.90.0 Beta\n\nflickering started with 51568030e9\n\n\nBisect unstable displayed when UV editor is open\n\nThis happens both with a convex polygon (example file) and with a regular cube.\n\n| 3.4 | 3.0 | 2.9 |\n| -- | -- | -- |\n| [2022-10-22 13-35-32.mp4](2022-10-22_13-35-32.mp4) | [2022-10-22 13-39-35.mp4](2022-10-22_13-39-35.mp4) | [2022-10-22 13-41-48.mp4](2022-10-22_13-41-48.mp4) |\n\n\n1. Open the example file\n2. Try to divide the polygon into parts using a bisect in edit mode\n\n[bisect_bug.blend](bisect_bug.blend)\n", "Crashes when Switching to Material Viewport / Rendered Viewport / Rendering\nOperating system: Windows 10\nGraphics card: 2 Nvidia GTX 1660 Super's\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: blender 2.8 / 3.4\n\n\nBlender runs completely fine when editing and building a model but once rendering is involved, even in preview, it immediately closes with no explanation. This crash will also happen with the default cube. This had not happened before and have tried Updating blender, reinstalling, simplifying, etc but no change and can reproduce the issue on my computer. i can have it in material mode if i have 0 assets in the scene but once i add something it crashes, i am suspecting a bad pathing issue as i had recently refreshed my computer but do not know how to fix it or give blender a hard reset.\n\n{[F14170509](blender.crash.txt)}[blender_debug_output.txt](https://archive.blender.org/developer/F14170500/blender_debug_output.txt)\nOpen Blender, have any Physical asset inside the scene, switch to rendered / Material Viewport, Crash\n\nI have included the project I was working on and the default cube as they create the same issue.\nPlease note that I have installed the most recent Blender program onto my laptop and opened the files attached and did not have any crash happen so it seems the issue is solely between my computer and the blender program.\n\n[Default Cube.blend](Default_Cube.blend) [Skell Rig base.blend](Skell_Rig_base.blend)", "Cycles shaders show up in EEVEE list (Properties > Material tab > Surface panel)\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.60\n\nBroken: version: 2.80 (sub 75)\nAlso confirmed broken in 2.80 release candidates 2 and 3\nWorked: ----\n\nSome shaders not yet available in EEVEE, such as the Toon shader, are included in the list of shaders in the materials tab of the properties window even when EEVEE is selected as the render engine. When such shaders are selected, adjusting most of their parameters has no effect on the viewport/render. Currently, the only way to know which shaders are supported is either by trial and error or by using the node editor.\n\n1. With a fresh blender scene, check that EEVEE is the selected render engine, and go to materials tab in the properties window\n2. You should already have a material if it's the default cube scene. In the surface panel, click on the shader selector (set to principled BSDF by default)\n3. The menu that shows up includes Toon shader among other unsupported ones.\n\n[Based on the default startup]\n\n**Screenshot**\n\n\n", "MultiRes+Solidify while trying to use Sculpt Mode cause Blender to crash\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.30\n\nBroken: version: 2.81 (sub 12)\nWorked: (optional)\n\nI tried deactivating one modifier while keeping the other one active to see if Blender would crash, but this was not the case. Only when I used both simultaneously did the program reliably crash every time, regardless of brush.\n\nAdd the modifiers that are active in the video and follow along. \n[Desktop 2019.09.23 - 00.49.39.02.mp4](Desktop_2019.09.23_-_00.49.39.02.mp4)\n" ]

[ "EEVEE: Crash caused by out of memory when using volumetrics\nOperating system: Linux-4.15.0-64-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: GeForce GTX 970/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.26\n\n\nBroken: version: 2.80 (sub 75)\nWorked: (optional)\n\n\nAny kind of volumetric rendering, and on occasion some simulations, causes Blender to crash, it also crashes my OS as well and forces the computer to reboot. The daily build from 29/09/19 also crashes but not as often as the 2.8 version.\nI have recently upgraded my computer hardware, I was not having any crashes before this upgrade. I have the latest available graphics driver and this has been reinstalled several times. The upgrade consisted of a AMD Ryzen 5 2600, a MSI X470 motherboard and 16 Gb of DDR4 RAM. [system-info.txt](system-info.txt) \n\nSet up some volumetric rendering using the principle volume shader. Switch to the eevee render engine, set tile size to 8px or lower and the samples to 100 or greater.[Test neb.blend](Test_neb.blend)\n" ]

[ "Light clamping from Normal or Bump maps near shadow terminator in Cycles\nOperating system : Windows 11\nCPU : AMD Ryzen 9 3900X\nGraphics card : RTX 3090\nRAM : 32GB\nHas also been tested on a Ryzen 7 Laptop with Ubuntu.\n\nHas been tested on every major release post 2.9, and on 3.6 build d5fc1b9ba4b8, identical behavior on all of them.\nAll provided screenshots of the test file (and the test file itself) were made with 3.5.1 Stable (build e1ccd9d4a1d3)\n\nAll normal information from a Height map passed through a Bump Node or from a Normal Map passed through a Normal Map node introduces some sort of light-clamping around the shadow terminator. The effect starts being visible at an angle of around 66° from the light source.\nThis is especially visible when the target object is Cylindrical and/or Spherical, and lit by a single Light source (Sun Lamp).\nIt looks like this behavior doesn't affect the Displacement node, since using it in Bump Only mode instead of the Bump node gives different results (they should both give identical results, as I understand it).\n\nFrom looking at other renderers (and aesthetically speaking) it looks like the way the Displacement node does it is supposed to be the correct one.\n\nA test Blend file is provided.\n\nSimplest way to reproduce it from scratch is to put a simple repeating height pattern on a smooth cylinder, using a bump node and a simple diffuse node, and lighting that cylinder from the side with a single Sun lamp.\n\nThe behavior of the Displacement node can be tested (and compared with the Bump/Normal) by plugging the same Height pattern into a Displacement node with the Shader Option set to \"Bump Only\", and plugging that displacement node into the Material Output.\n\nI was initially thinking it to be related to (or the same issue as) this one :\n95729\nBut it has apparently been fixed (105776), while the clamping is still clearly visible even in very recent builds.\n\n", "Cannot disable shadows if an object is set to holdout.\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: Confirmed on Version 3.4 to version: 3.5.0\nWorked: Never (Goes as far as 3.0, since the addition of the \"Object Properties -> Visibility -> Holdout\"option)\n\nCannot control an objects shadow visibility if the object is set to holdout. Having a shadow on a holdout object does not work as expected, especially when trying to use the holdout object as a 3d mask.\n\n\n- Set scene renderer to Eevee.\n- Create an object, set \"Shadow Mode\" to none in the material shader of said object.\n- The object will no longer have a shadow, as expected. \n- Set the object to holdout via \"Object Properties -> Visibility -> Holdout\" option, and the shadow reappears. Can not turn off shadow as long as this option is on.\n - However, if use Holdout shader in the nodes and connect it directly to the Surface output, no shadow will appear (this matches expectation).\n\nAdded a file that directly shows the issue.", "3D Paint Mode stencil disappears when panning viewport.\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.76\n\nBroken: version: 2.93.4\n\nWhen panning viewport in 3D paint mode with stencil mask enabled, the stencil mask disappears when moving the viewport, however does not disappear when zooming it, which indicates it's not a technical limitation.\n\n1. Enter 3D paint mode\n2. Enable texture mask, select some texture and set mapping to stencil\n3. Navigate the viewport\nResult: The stencil disappears when moving the viewport, which makes it very difficult to line up stencil projection with the 3D geometry\nExpected: The stencil does not disappear when navigating the view\n\n", "Modifiers don't work on particle instances of a hidden object\nOperating system: Linux Mint 20.3 Cinnamon kernel 5.13.0-35\nGraphics card: 3400G\n\nBroken: 2.83\nWorking: apparently never\n\nRelated to #96367 (Crash when trying to snap to instances on an object)\nTo replicate bug, the instantiated Object (the one with a `Modifier`) has to meet the following conditions:\n* Must have a modifier (obviously)\n* Must be instantiated through a particle system. (instantiated through Geom Nodes is OK and through Dupli Verts/Faces hides the `Object` as well)\n* Must be hidden inside a collection with the `Excluded from View Layer` option enabled\n\nAnd the instantiating Object (the one with `Particle System`) has to meet the following conditions:\n* Must be referenced by a `Camera` in the \"`Focus on Object`\" field\n* In the outliner, the `Camera` object must be placed in a `Collection` before the instantiating Object\n\n|||\n| -- | -- |\n|Camera placed before |Camera placed after\n\n # Open the attached file\nUnhide collection or switch `Camera` and `Plane with particle system` collections order (via Drag and Drop) get around the problem.\n[Camera 'Focus on Object' - particle system issue.blend](Camera__Focus_on_Object__-_particle_system_issue.blend)", "HDRI animation bug in eevee\nThe HDRI illumination shouldn't change on the cube, because every thing including the HDRI is rotating at the same speed. \nHowever, this went wrong in EEVEE... It seems that in Cycles everything went on so well.\n\nThis is a simplified bug presenting file of a complex scene which I found this bug, in which I have to rotate an object, a HDRI and camera together, and I'm struggling to get rid of this bug...\n\n[BugReport.blend](BugReport.blend)\n\n[0001-0096.mkv](0001-0096.mkv)\n\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11\n\nBroken: version: 2.93.1\n\n\n\n", "Unwanted camera movement\nOperating system:\nGraphics card: NVIDIA GeForce RTX 3060 Ti\n\nBroken: 3.6.1, 2023, non of the versions I have tried has worked \n\n\nI have enabled the feauture to \"fly around\" by pressing shift + f. When you press those buttons you will be able to fly using W D A S keys. Sometimes when I use this feauture it works fine but after a while it starts bugging. For example I use this tool for zooming in on objects really close and when this bug comes in this happens: I go really near the object and everything looks fine but when i left click like your supposed to do the camera moves places. It usally starts just moving a little, but the closer you get to an object the more it moves. Like it snaps out of place.\n\n\n- click shift + F \n- move near an object (cube) \n- left click to settle the camera\n- and see how it moves\n\n", "Ray Visibility (glossy, diffuse, transmission) on light sources generate a different result when using hit by the BSDF or next-event estimation \nOperating system:\nGraphics card:\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\n\nWhen changing/disabling the ray visibilities on a light source (glossy, diffuse, or transmission), the result becomes\ninconsistent when using only BSDF sampling, next-event estimation, or MIS between both.\n\nThe reason for that is that the contribution of a light source is completely neglected when the type of the BSDF closure used for sampling the direction is disabled on the light source.\nDuring next-event estimation, on the other hand, the light source ALWAYS contributes. Only the BSDF evaluation components multiplied with the light contribution differ [ eg. instead: L*glossy + L*diffuse -> L*diffuse if glossy is deactivated].\n \nThe example shows a scene with one big rectangular light at the ceiling, where glossy rays are disabled on. \nAs you can see, the BSDF image is darker compared to MIS, and the NEE version is a little bit brighter.\n\n\nThe MIS image is more or less a blended version of BSDF and NEE.\nBecause both BSDF and NEE do not converge to the same result, the MIS one is inconsistent/biased.\nSince the blending factors/mis weights depend on the sampling strategies, the result changes when using path guiding or the light tree.\nThis makes these optimized strategies unpredictable and can confuse the artists.\nIn the worst case, they report bugs on for the light tree and path guiding.\n\nA WIP fix is proposed in [D16965](D16965)\n\nBased on the default startup or an attached .blend file (as simple as possible).", "Object cryptomatte is not respecting the camera Dof\nOperating system: Windows 11\n \n\nBroken: 3.0, 3.6\nWorked: never?\n\n\n\nCycles object cryptomatte is not respecting the camera depth of field while the material one respects it properly. See the image for comparison\n\n\n\n\nload the attached .blend (it already has the object and the material cryptomattes enabled)\n\nrender\n\ncompare the current cryptomatte types in the compositor\n\n", "Eevee viewport soft shadows are broken on higher sample count\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.92\n\nBroken: version: 2.93.0\nWorked: ---\n\nEevee soft shadows in viewport have color spills. But it works fine on final render (F12)\n\n\n\n- Open .blend file\n\n[bug_soft_shadows.blend](bug_soft_shadows.blend)\n\n- Switch to Rendered Mode of Viewport Shading\n- Wait half a min and see the problem(", "Panoramic Render in Cycles. Bump bug.\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 4090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.02\n\nBroken: version: 3.4.1\n\nWhen using a material with fine bump and a panoramic camera the bump appears inncorrectly in rendered view.\n\n- Give an object a material with a fine bump (e.g. noise texture 15 detail, .8 roughness. Bump strength 1, distance 1)\n- Bring camera in really close.\n- Set to cycles.\n- Switch camera between panoramic and perspective mode.\n\nPlease let me know if there are any work-arounds, this bug is causing me problems.\nThank you.\n\n", "Dynamic Indirect light cache\nThe idea would be to have a multiple level cache that stores a voxelized version of the scene to render the irradiance points.\n\nEach voxel would be linked to a set of surfel (surface elements) that stores color and normal of the surface.\nEach time the lights in the scene changes, the voxelized version is updated and propagates the change to the irradiance points.\n\nThis means having a list of surfel per voxel (say 6, 8) and a list of voxel per irradiance points.\nThis would be a mix of The Division dynamic indirect lighting [link]] and Mc Guire raytraced irradiance probes [[http:*jcgt.org/published/0008/02/01/|link](https:*www.gdcvault.com/play/1023273/Global-Illumination-in-Tom-Clancy).\nThis feature might be limited to hardware that supports compute shaders.\nThis is not a trivial feature and needs a lot of R&D.\n\nEstimate : 4 weeks\n", "Shadow catcher causes artifacts in PNG exports with denoising\nOperating system: Manjaro Linux\nGraphics card: AMD Radeon 535\n\nBroken: 3.1.0\n\n\nIn Cycles, rendering a scene with transparent world background and shadow being projected into a shadow catcher causes a ring of pixels with very little opacity to be created around the render's edges when exporting it as a PNG. These artifacts go away when re-rendering the scene with the shadow catcher turned back into a normal plane.\n\nI'm aware of #93865 (Artifacts around shadow catcher object) but created this one because it wasn't clear to me if this was the same issue, as that one seems to be considered a property of the 'Fast' denoiser. These artifacts were recorded on the 'Accurate' one.\n\n- In Cycles, make the world background transparent\n- Move the default cube slightly up and add a plane, scale it up a bit. \n- Point the camera so that your render captures the cube projecting a shadow into the plane, but still has the transparent world background showing around the edges\n- Turn the plane into a shadow catcher and render the scene with denoising on. Export it as a PNG; color depth and compression don't affect this issue.\n- Save the render and open it in Krita (other image editing applications may have the following feature too)\n- On the Select menu, use Select Opaque to highlight all pixels with some opacity. You will notice pixels very close to the image border all around the edges. These are normally invisible, being of very low alpha, but if you stack many renders produced like this they will be revealed as being grayscale.\n- Now, change nothing else, but turn the shadow catcher into a normal plane. Render the scene again, save it with the same settings. The ring of noise around the image edges will have disappeared.\n\nScreenshot of Krita showing the opaque pixels on a normal render:\n\n\n\nOpaque pixels on the same scene rendered with the plane made into a shadow catcher and denoising set to Accurate. Note the barely visible artifacts selected around the plane and image edge. These can be observed no matter what shape the render output takes.\n\n\n\nAnd now the same scene, denoising turned off. No ring around the edges.\n\n\n\nBlender file used for the renders above: [democatcherbug.blend](democatcherbug.blend)", "SpaceView3D.draw_handler_add with PRE_VIEW draw type does not result in visible draws\nOperating system: MacOS 11.6.3, Ubuntu 18.04 LTS (issue exists in both)\nGraphics card: Intel UHD 630, Nvidia Quadro RTX 400 (issue exists in both)\n\nBroken: 2.93 LTS, latest master (b626edd751d64b269da12ef8e774ced2836c3c9e)\nBroken: 2.83.0\nWorked: 2.82\n\nBroke between good 189e96ba68 and bad d95e9c7cf8\n\n1. Startup Blender\n2. Go to scripting tab and create a new script\n3. Copy \"Triangle with custom shader\" code from: gpu.html\n4. Change draw_type param in draw_handler_add from\n```\n'POST_VIEW'\n```\nto\n\n```\n'PRE_VIEW'\n```\n5. Run script.\n\nExpected behavior:\nThe drawn triangle is visible.\n\nActual behavior:\nThe drawn triangle is not visible.\n", "2.93 EEVEE ScreenSpace Reflections - Trace Precision Factor - Faded Reflections (Reflections Not Fully Reaching The Upper Surface of Objects ''\nOperating system: Windows-10\nGraphics card: GeForce GTX 960M\nBroken: version: 2.93.0 Beta\n\n\n\n\n\n\n\n\n\n\n\n\nThe new 2.93 screenspace reflection implementation' s trace precision is not enough to put reflections on all the surface of object (especially near the top parts of surface near upper edge) compared to the screenspace reflections with same or less trace precision in old 2.92 implementation. Its like reflection always fades to top and increasing trace precision doest fix it.\nNormally trace precision allows for reflections go all the way up on (for example cube surface) to make its surface fully reflected, but new 2.93 Trace precision 1 doesnt allow for fully making surface reflective and give darker less reflective areas near upper side edges or upper corners\n\nFor instance, in most cases: 2.92' SSR with trace precision 0.25 is almost = (almost equal to ) = 2.93 SSR with trace precision 1 (maxed)\n\n(Even though camera is further away enough to get all the reflections from screenspace , new SSRI 2.93 trace precision is inadequate and kind of gives the effect of ''as if'' we make zoom in so screenspace doesnt see all objects and the top part of cube didnt get any reflections and get darker and faded into weak reflections) \n\nMoreover, when its compared with cycles, with trace precision of 1 ob both , 2.92 one looks closer to cycles while 2.93 has faded reflection on upper pars of surface (maybe even though new reflection implementation uses cycles implementation, new implementations inadequate trace precision might be making it differerent from cycles)\n\n\n1. Create a plane scale by 5 and move -1 (down)\n2. Edit the material of default cube and previously created plane by using same material and give it a red color FF0000 with max saturation 1 \n3. Optional step: make base color saturation of material lower to see faded darkening effect more Plus Make it metallic if you want to see the result more with metallic or just leave it as it is\n4. Ensure viewport shading settings do not use scene world but use the forest HDRI and maximize world opacity and remove blur\n5. In properties rendering tab, Enable screenspace reflections and make edge fading zero. Also uncheck half-res trace\n6. Go camera view and dont zoom too much on viewport so screenspace reflection can see everything around\n7. !!Do same processes in 2.92 and 2.93 releases to compare \n8. Then, set 2.92 and 2.93 have trace precision to 0.25 - and observe\n- In 2.93 U will notice the reflected surface of the cube as u approach top edges/as u go up, have darker/unsaturated or ''Faded'' screenspace reflections caused by the trace preciison is not enough to reach there\n- however in 2.92, the reflection reaches better to all the surface and it has almost no faded or darker look\n9. Make 2.93 SSR trace precision 1 while 2.92 SSR is still 0.25 and observe:\n- then u will notice, ''the result of 2.93 SSR Trace precision 1 '' is more similar to ''2.92 SSR with trace precision of 0.25 ''\n10. When 2.93 SSR Trace precision is 1, make 2.92 SSR trace precision 1 and observe agan\n- Then u will notice 2.92 SSR Trace precision is the ideal to fully trace all reflection to surface perfectly without having any faded area with less reflection. On the other hand, 2.93 Trace precision 1 is still not adequate to put reflection to upper parts of cubes surface near top edges with full power and precision\n\n'!! Compare these on both 2.93 and older 2.92 version with old SSR implementation\n\nIn summary in 2.92 when we increase trace precision from 0 to 1, the tracing or visibility of reflections go all the way up to top edge of cubes surface making it fully reflected as trace precision maxed\nOn the other hand, in 2.93 if we increase trace precision from 0 to 1, the reflections still cannot reach all the way up to cubes surface near top edge which gives a faded and partially traced upper surface area.\n\nMaybe, To solve this maybe, the screenspace trace precision factor can be multiplied by 4-6 or maybe the it should be square cubed so the 2.93 trace precision of 2.92.\n\n[2.93&92 SSR-TRACE PRECISION 1 - CAMERA VIEW .blend](2.93_92_SSR-TRACE_PRECISION_1__-_CAMERA_VIEW_.blend)\n\n[2.93&92 SSR-TRACE PRECISION 0.25 - VIEWPORT VIEW POSITIONED .blend](2.93_92_SSR-TRACE_PRECISION_0.25_-_VIEWPORT_VIEW_POSITIONED_.blend)\n\n[2.93&92 SSR-TRACE PRECISION 1 - VIEWPORT VIEW POSITIONED .blend](2.93_92_SSR-TRACE_PRECISION_1_-_VIEWPORT_VIEW_POSITIONED_.blend)\n\n[2.93&92 SSR-TRACE PRECISION 0.25 - CAMERA VIEW.blend](2.93_92_SSR-TRACE_PRECISION_0.25__-_CAMERA_VIEW.blend)", "In cycles, ignore shadow of volume object will cost more time.\nOperating system: windows 10 64\nGraphics card: rtx 2080\nCpu: AMD Ryzen 7 2700\n\nBroken: 2.83 2.82 .....\n\nIn cycles, ignore shadow of volume object will cost more time.\nUse texture in volume material.\nVolume object cross other object.\n\n\n[vol_bug.blend](vol_bug.blend)" ]

[ "CYCLES: Surfaces inside volumes are Not Lit or Shadowed by the volume if the volume is not visible to Camera Rays\nLinux Mint Mate\nGTX 1050\n\n\nBlender 2.81 Alpha\n2019-09-27 22:44\nHash: 1c1a3198af9d\n\n\nDiscovered while trying to catch shadows from volumes on surfaces for compositing.\n\nSurfaces inside of volume meshes are not shaded (lit/shadowed) by the volume if the volume is invisible to camera rays\n(checkbox in Object Properties -> Visibility -> Ray Visibility -> Camera)\nAs far as i can tell this also affects Holdout.\n\n\n\n\nCurrent workaround: make sure volume IS visible to camera rays, and instead bypass it in the volume shader using the Camera Ray socket on the Light Path node\n\n\nTo Reproduce:\n\nOpen attached .blend , press render\nNote the missing shadow/light where the volume cubes intersect the default cube.\n\nI have in the .blend also included the \"expected behaviour\" by using the workaround, switch the Scene from \"Bug\" to \"Workaround\" and press render again.\nNote the shadow/light now covers the default cube properly." ]

[ "Weird Motion Blur with Simulation Nodes\nBroken: 3.6.2\n\nMotion Blur doesn't work properly with Geometry Nodes instances.\nAs shown in the rendered video, the motion blur seems to give weird results, the motion trails of the instanced particles from the Geometry Nodes (and Simulation Nodes) setup don't seem to follow their direction properly.\n\n<video src=\"attachment\" title=\"animation.mp4\" controls></video>\n\n[Test.blend](attachment)\n- Open file\n- Hit play to bake simulation until let's say frame 45\n- Render frame 45\n\n", "Texturing with anchored type stroke\nOperating system: Linux-5.4.0-47-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 450.51.06\n\nBroken: version: 2.91.0 Alpha\n\nTexture paint with stroke set to anchored while using texture significantly slows down blender up to non responsive state and crash while paint texture is larger than 2k.\n\n[Texture pain stroke ty[e anchored.blend](Texture_pain_stroke_ty_e_anchored.blend)\n", "Shader compilation freeze Blender on Intel Arc770\nOperating system: Windows 10\nGraphics card: Intel Arc770 16Gb (Resizable Bar enabled)\n\n3.6.0\nDate: 2023-06-27 08:08\nHash: c7fc78b81ecb\nBranch: blender-v3.6-release\n\n**Intel driver version**\n31.0.101.4665\nReleased 11.08.2023 (latest on moment of this post)\n\nAll modes of viewport on new scene (default, for example) works fine: solid, material preview, display render. But, if I open my scene, works only 1 and 3 option. Building a shaders freezes Blender. It's around 800+ materials on scene, and GeForce cards works fine with that. The video card works at full power, with enabled Resizable Bar function. Cycles render (oneAPI) also works fine, but i can't switch the viewport on material preview. Blender freezes when shaders start compile.\n\n", "Uncheck Save Copy freezes on next render - Mutex is still locked\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.91.0 Alpha\n2.90.1 and 2.83.5 have the same behavior\n\nWith default scene :\nWhenever I try to save a render, I uncheck both save as render and save copy, the next time I hit render, blender will freeze.\n- Unticking save as render only does not seem to create the freezes, however the option is not saved, I have to manually uncheck it every time.\n\nNew file : \n- Save file\n- Hit render\n- Save as\n- Uncheck Save as Render and Save Copy\n- Save.\n- Hit Render again = freeze\n\n\n", "Crash When Adding Hair Subdivision upon Viewer node\nOperating system: Windows-10-10.0.17134-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 970M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.11\n\nBroken: version: 3.4.0 Alpha\nWorked: Never\n\nCaused by: c55d38f00b\n\n\n\nAdding hair subdivision will cause crash immediately when you are using viewer node to visualize field values in the viewport.\nNote no crash will occur if the viewer node is inactive.\n\n[Hair Crash.crash.txt](Hair_Crash.crash.txt)\n\n1. select default cube\n2. shift+A to add empty hair\n3. in the node tree add viewer node.\n4. Crash if adding subdivision\n[Hair Crash.blend](Hair_Crash.blend)", "Subdivision Surface: Out of memory exception in in post 2.79 versions\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.90.0\nWorked: (2.79)\n\nWhen rendering a displacement heavy blend file, We consistently get out of memory exception in blender 2.8x and 2.9.\n\n - Open the attached blend file.\n - Click F12 (Render Image).\n # Wait till it crashes (approximately 6min).\n\nMore info about this blend file:\n```\nResolution: 650px by 650px\nCycles: 10\nAll polygons are triangles.\nOne material with a displacement map on all objects.\nSubdivision modifier with Catmull Clark with 1.0 adaptive value, on all 3 objects.\nSimplify->Render->Maximum Subdivion is equal to 2.\nSimplify->Render->Texture Limit is equal to 8192.\nSimplify->Viewport->Maximum Subdivion is equal to 0.\n```\n\n[Bug-Upload(2.79).blend](Bug-Upload_2.79_.blend)", "Cycles: Pointiness has artefacts on extremely small mesh size\nOperating system:\nGraphics card:\n\nBroken: version: 3.6.0 Alpha\nWorked: Never\n\nPointiness has artefacts on extremely small mesh size. \n\n| Normal size | Small size | Smaller size |\n| -- | -- | -- |\n|  |  |  |\n\n1. Create cube with beveling edges. Create small copies on vertices (as on example pictures).\n2. Set shader to cube. Shader with:\n a. Inputs: Geometry -> Pointiness.\n b. Map range: From Min: 0.5, From Max: 0.54.\n c. Material Output.\n3. Render that cube with applied different sizes.", "Toggling Autosmooth changes evaluated bbox `obj.bound_box` when SUBSURF mod (using GPU subd) is present\nOperating system: Linux-6.2.0-76060200-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Ti Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 525.89.02\n\nBroken: version: 3.5.0\nWorked: -\n\nOn a mesh object with a SUBSURF mod active (using GPU Subdivision), the (evaluated) bounding box at `obj.bound_box` will only produce the correct coordinates when `obj.data.use_auto_smooth` is enabled. Otherwise it will be produce the coordinates of the original mesh bounding box, even though the evaluated mesh is clearly smaller, and so the bbox should be too.\n\nThat makes relying on `obj.bound_box` impossible, when working with the bounding box and assuming it represents the evaluated mesh object.\n\nInterestingly, boolean mods don't share this behavior, the bbox created using booleans, works properly with either auto smooth setting.\n\nSee watch?v=KKnWbRL5h7o for demonstration\n\n\n* open the attached blend file\n* toggle Autosmooth and see the bound box change, even though the evaluated mesh displayed does not change at all.\n\n", "Remesh Resolution strange behaviour (Shift-R)\nOperating system: MacOS 10.14.6\nGraphics card: AMD Radeon R9\n\nBroken: Daily Build\nWorked: Don't know\n\n**Description**\nSetting the remesh resolution with Shift-R during sculpt when object is unproportional in size makes the grid fade (too quickly) and when zoomed in the numeric display of scale is upside down\n\nIn this small video clip, first part is correct behaviour.. second with disproportional object\n<video src=\"remesh_behaviour.mov\" controls></video>\n\n**Steps**\n- Use default cube\n- In edit mode scale the object 10x in Z only\n- Enter Sculpt mode\n- zoom in\n- Shift-R\n- observe grid fading and display upside down..\n\n[remesh_behaviour.mov](remesh_behaviour.mov)\n\n\n", "closest_point_on_mesh() freezes Blender when called from a driver\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 580/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 2.82 (sub 6)\nWorked: 2.79\n\nCalling closest_point_on_mesh() from within a driver causes Blender to freeze.\n\n[test2.blend](test2.blend)\nOpen the file allowing scripts -> Freeze", "Unable to zoom/scale graph editor with [ALT] middle mouse button after frame selected of lowest point\nOperating system: Linux-5.4.0-107-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.60.02\n\nBroken: version: 3.2.0 Alpha\nWorked: Likely never\n\nGraph editor goes into locked cycle when frame selecting certain points\n\n[Bug.blend](Bug.blend)\nOn attached file. Box select lowest point of y location on frame 25. Hit Num Pad . to frame the selected point\nAttempt to scale or move graph up and down with [Alt} middle mouse button\n\n", "Thread race condition with lock interface enabled\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.92\n\nBroken: version: 2.93.2 Release Candidate\n\n**Description**\nUI redraw is still taking place in the render_init and render_complete handlers when use_lock_interface = True. This lock was only implemented for 3D viewport, but this is not mentioned in any of manuals.\n\nAdding lock to node editor seems to resolve the issue:\n\n```\ndiff --git a/source/blender/editors/space_node/space_node.c b/source/blender/editors/space_node/space_node.c\nindex ff848a7bb95..9813b8aaa86 100644\n--- a/source/blender/editors/space_node/space_node.c\n+++ b/source/blender/editors/space_node/space_node.c\n@@ -1073,6 +1073,7 @@ void ED_spacetype_node(void)\n art->cursor = node_cursor;\n art->event_cursor = true;\n art->clip_gizmo_events_by_ui = true;\n+ art->lock = 1;\n\n BLI_addhead(&st->regiontypes, art);\n```\n\nTrace: \n```\n>\tblender.exe!UI_block_end_ex(const bContext * C, uiBlock * block, const int * xy, int * r_xy) Line 1873\tC\n\n \tblender.exe!node_draw_basis(const bContext * C, const View2D * v2d, const SpaceNode * snode, bNodeTree * ntree, bNode * node, bNodeInstanceKey key) Line 1733\tC++\n \tblender.exe!node_draw_default(const bContext * C, ARegion * region, SpaceNode * snode, bNodeTree * ntree, bNode * node, bNodeInstanceKey key) Line 1943\tC++\n \t[Inline Frame] blender.exe!node_draw(const bContext *) Line 2002\tC++\n \t[Inline Frame] blender.exe!node_draw_nodetree(const bContext *) Line 2059\tC++\n \tblender.exe!draw_nodetree(const bContext * C, ARegion * region, bNodeTree * ntree, bNodeInstanceKey parent_key) Line 2097\tC++\n \tblender.exe!node_draw_space(const bContext * C, ARegion * region) Line 2235\tC++\n \tblender.exe!ED_region_do_draw(bContext * C, ARegion * region) Line 569\tC\n \tblender.exe!wm_draw_window_offscreen(bContext * C, wmWindow * win, bool stereo) Line 726\tC\n \tblender.exe!wm_draw_window(bContext * C, wmWindow * win) Line 868\tC\n \tblender.exe!wm_draw_update(bContext * C) Line 1067\tC\n \tblender.exe!WM_main(bContext * C) Line 654\tC\n \tblender.exe!main(int argc, const unsigned char * * UNUSED_argv_c) Line 563\tC\n\n```\n\nOpen below file, run the script and then press f12.\n\n[ui lock ignored on some handlers which are not on the main thread.blend](ui_lock_ignored_on_some_handlers_which_are_not_on_the_main_thread.blend)\n\n", " Mantaflow: Adaptive Domain breaks Dissolve if 'slow' is checked\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.28\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional)\n\n[Adaptive domain seems to completely ignore the 'dissolve' setting on the Domain object, or at least alter it massively.]\n\n[Bake the attached .blend\nHere's the results-\n\nThe left hand Domain has dissolve enabled but not Adaptive. The right hand one is a duplicate of the left (and the left hand emitter), nothing at all changed apart from enabling Adaptive domain.\nDoesn't look to me like the 'dissolve' parameter is doing terribly much at all on the right hand side, when 'Adaptive Domain' is enabled.]\n[Adaptive_Dissolve.blend](Adaptive_Dissolve.blend)\n\nI discovered this issue half way through a 400 frame render, when looking at the rendered stills; so I thought stuff-it! Carry on...\nSo here is what this issue does to an actual sim I wasted several hours on creating/tweaking, then the best part of a half a day rendering-\n[Domain_Jitter_Dissolve.mp4](Domain_Jitter_Dissolve.mp4)\n\nIt looks worse than the .blend I uploaded as it has noise baked as well, and this seems to generally expand the volume filled by the smoke from what I've experienced, therefore exacerbating the jitter where the smoke hits the Adaptive Domain boundary. \nAs you can see it's quite ugly when the boundary of the smoke gets close to the camera, and as I hope you can see the jitter at the edges of the smoke look suspiciously like the jitter in the size of the domain seen in the viewport capture I made (obviously I matched the speed in Davinci, my workstation could only sustain 2FPS playback).\nIf anyone cares for my two-pennys-worth, whatever coded-wizardry controls the size of the Adaptive Domain, need's to implement some sort of 'damping' for the domain size changes, at least several (10-15?) frames worth. As well as taking 'dissolve' into account.\n\n", "Mesh selection in viewport broken with scenes using a high number of models\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.02\n\nBroken: version: 3.5.0 Alpha\nWorked: Never (2.8+)\nWe are working on a commercial project with imported SketchUp scenes using 55K+ models and running more and more into broken scenes with selection issues in the viewport.\n\n`What we figured out so far:`\n- Selection in the outliner works\n- Some models can be selected by a point and click\n- Other models can't be selected by a point and click\n- Models at a different location will be selected by a point and click and not at the mouse pointer location\n- Box select is messed up and doesn't work overall (selects just parts of the model or fails completely)\n- Lasso select seems to work as intended\n- Circle select seems to work as intended (but can freeze Blender)\n- Some random models can cause a hard crash of Blender when selecting them by a point and click\n**When we delete all my other scenes in the Blender file and purge the data the selection modes work again.**\n\n\n\n\n[T104008_Selection_Bug_350_v006_CUBES_REPORT.blend](T104008_Selection_Bug_350_v006_CUBES_REPORT.blend)\n\nTry to select cubes.\n\n", "Vertex/Edge Slide stuttering with modifier.\nOperating system: Windows 11\nGraphics card: RTX 3070\nCPU: Ryzen 5900x\n\nBroken: 2.83 and all other.\nWorked: 2.79b\n\nVertex/Edge Slide freezes a lot with subdiv modifier in a relatively simple object.\n\n| With `Use Limit Surface` | Without `Use Limit Surface` |\n| -- | -- |\n| https://projects.blender.orgattachment | https://projects.blender.orgattachment |\n\n1. Import attached fbx.\n2. Add subdivision modifier set to 1 subdivision.\n3. Select the loop in the middle.\n4. Edge slide.\n\n" ]

[ "Cycles preview lockup / memory allocation loop while changing adaptive displacement parameters.\nWindows 10, Nvidia 1060 6GB, also reproduced on a second Windows 10 system with Nvidia 1070.\n\nBroken:\n1/9 daily build blender-2.80-91a155833e59-win64.zip\n\nCycles preview (Workspace Rendered mode) of object with displacement goes into a loop allocating memory.\n\n\nCAUTION: You may lose control of the machine and require a hard system reset! I have reproduced this issue on two separate machines, and on one I was unable to get the Task Manager or CTRL-ALT-DEL to respond after locking up Cycles. (Edit: actually this was probably just because there was another minimized GPU intensive application running as well and the combination of that and Blender looping probably just prevented getting any resources to get at the task manager).\n\n1) Open the attached .blend which contains the default box with an adaptive subdiv modifier.\n2) Switch to Rendered mode in the 3dview. Object edges look a bit rough but otherwise ok.\n3) In the Render Subdivision properties panel, change the Preview Dicing Rate to 2. Note that this does not trigger a preview refresh (it probably should I think?)\n4) Click the wrench tab to switch to the Modifiers Properties tab.\n5) Click on the Dicing Scale, type in 0.5 and hit enter. The preview refreshes correctly with the new settings.\n6) Now Click the increase (>) button at the right-hand end of the Dicing Scale a few times (about a half second between clicks say). After a couple changes (value up to 0.56 usually) the object in the display will pixelate and the render thread will go 100% CPU busy and will start allocating memory without limit until you kill Blender.\n\nThis isn't the only way I've gotten it into this state while messing around with Displacement in this file, but the above sequence is 100% repeatable where otherwise it's pretty random. I've gotten it to go berserk like this while twiddling displacement options including view displacement level, the render Subdivision options, and switching back and forth between the different Workspace display modes (lookdev<->rendered etc.).\n\nIt will also sometimes do something else weird, where instead of the smooth, connected displaced surface, it seems to be displacing each face of the subdivided mesh independently. I haven't been able to reproduce this reliably.[disptest.blend](disptest.blend)" ]

[ "Blender crash on combining textures (intel)\nOperating system: Windows 10\nGraphics card: Intel HD Graphics 4400\n\nBroken: 2.93.2\nWorked:\n\nWhen I tried to combine 2 textures like this blender is crashing.\n  \n\n[Test.blend](Test.blend)\n[Test.blend1](Test.blend1)\n[system-info.txt](system-info.txt)\n[Test.crash.txt](Test.crash.txt)\n[2021-11-12 09-11-32.mp4](2021-11-12_09-11-32.mp4)\n", "Cycles Render: Volume (Before Baked Simulation Starts) still trying to be rendered\nOperating system:windows 11\nGraphics card:nvidia geforce rtx 2060\n\nBroken: 3.4.1\n\nThe cycles render tries to use the volume cache to render if the frame is less than the start of the simulation. See video example.\n\n1. Default cube: Add Quick Smoke.\n2. Cache settings:\n a. Type: All.\n3. Start frame: 50, end: 100.\n4. Bake.\n5. Start render on 30~ frames.\nExpected: Simulation chunks will appear in different parts of the domain.", "Debug information not generated upon crash.\nOperating system: Arch Linux 6.4.12-arch1-1\nGraphics card: AMD ATI Radeon 680M\n\nBroken: 3.6.2\nWorked: Could not determine (tried up to 3.5.1, which does not work without a full system downgrade)\n\nBlender crashes immediately upon launch.\n\nLaunching from the command line, I get this:\n```\nWriting: /tmp/blender.crash.txt\nSegmentation fault (core dumped)\n```\n\nThe contents of /tmp/blender.crash.txt are this:\n```\n# Blender 3.6.2, Commit date: 2023-08-14 14:05, Hash ffe93138751e\n\n# backtrace\n\n# Python backtrace\n```\n\nPassing the --debug-all flag, I get no errors. I attached the output of this command as a text file.\n\nWhile Blender crashing on startup is not in and of itself an issue worthy of reporting, the lack of any debug information immediately makes the problem impossible to tackle independently. I have no idea what my next steps should be here.\n\nLaunch blender.\n\n", "Blander crash without ray shadows Nvidia GTX 950M CUDA\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 950M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 411.63\n\nBroken: version: 2.81 (sub 16)\nWorked: (optional)\n\nWhen turn off ray shadow visiblity render becomes slow and the blender crashes.\n[2019-10-28_03-53-23.mp4](2019-10-28_03-53-23.mp4)\n\n------------\nI tested again on GTX 750 Ti and 2.92.0 Alpha. If I turn shadows off first, then switch to rendered view, it gives bad render:\n\nIt disappears when you do something. Typical when there's something wrong in Cycles. I need to restart Blender to be able to render again.\nWhen I turn rendered view first, then turn shadows off, it sometimes crashes.\nThis is all Cuda. Optix renders nothing at all. I presume Optix doesn't do volumes yet.\nAlso noticed that deleting sun lamp and leaving only area lamps fixed the issue.\n\n**Steps to reproduce**\n[облака в стекле.blend](облака_в_стекле.blend)\n- Open file\n- Disable shadow ray invisibility on object\n- Switch to rendered shading\n", "Blender 3.1 CPU/GPU rendering denoise is broken on M1 Mac / Monterey\nOperating system: MacOS Monterey (M1 Max)\nGraphics card: Integrated GPU\n\n**Blender Version** 3.1.0 release for M1 Macs\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\nWorked: (newest version of Blender that worked as expected) 2.93.3\n\nOn all my projects, when rendering via the CPU or GPU, the render works as expected, but on completion the entire screen blacks out to a solid color and the render is LOST! I suspected it must be a post render step, and I have confirmed that it is the DENOISER that is broken in this version of blender. When I go into the scene and remove the denoiser, the render completes, and I can save the image!\n\nBased on the default startup or an attached .blend file (as simple as possible).\nYou need a project that uses denoising as a final compositing step. Attached is a simple matt object set to denoise at the end. Render it, and at the end you get a gray screen. Go into compositing and remove the denoise node, and you get the render.[TestRenderWithDenoise.blend](TestRenderWithDenoise.blend)\n", "Blender freezes when using Texture paint or Weight Paint\nOperating system: Windows 10 Insider build19592\nGraphics card: GTX 1660Ti Max Q 6GB\nProcessor: i7-9750H\nRAM: 32GB DDR4-2666Mhz\n\nBroken: version: 2.83 (sub 11)\nWorked: Never (2.79, 2.8+)\n\nBlender freezes or stops responding when using Texture Paint or Weight Paint. Tried game ready and Studio drivers, still the same issue.\nMoreover, RAM consumption spikes up to more than 20GB. This happens when the brush size is increased and starts painting. \n\nAssigned GPU in Preferences>System>CUDA, but the GPU is been barely used.\n\n[Blender.rar](Blender.rar)", "Incorrect shading on Intel HD 6000\nOperating system: Mac OS Catalina 10.15.3 (latest as of 13/02/2020), Windows 10 1909 (Bootcamp)\nGraphics card: Intel HD 6000\nCPU: Intel Core i5-5350U\nHardware: MacBook Air 2017\n\nBroken: 2.81a, 2.82\n\nProblems with shader of Auto-Eye add-on on Intel HD 6000, although on creator's machine works fine (I emailed him and he suggested it's a blender bug).\nBUT it's not only Mac OS problem! Also tested on Windows (via bootcamp) -- it crashed (on Mac OS it continue working at least).\n\nCreator's machine specs (where it works fine):\nIntel Core i7-4790 CPU @ 3.60GHz × 4\nGeForce GTX 1060 6GB\n\nJust open .blend file, make sure it's EEVEE render/look dev mode, test it with Mac OS or Windows and Intel HD card.\n\nAttaching file for testing and screenshots -- first two is how shaders looks broken, and third how it should approximately look, but with shader (right now it's only color output or flat texture output)\nBroken 1\n\nBroken 2\n\nHow it should like like(with shader enabled (BUT right now it's only color/flat texture output)\n\n\n[iris_only.blend](iris_only.blend)", "Cycles/Compositor: The denoiser returns a black image if not all requirements are fulfilled\nOperating system: Default KVM (Proxmox, Linux)\nGraphics card: None\n\nBroken: Blender 3.0\nWorked: Before Blender 3.0 the frame was just not denoised\n\nIf you enable OIDN/Optix denoising without having the needed CPU flags you will get a black frame back.\nThe desired outcome would be an aborted render with a helpful error.\n\n\n1. Set up a default KVM \n2. Enable denoising \n3. Render over the CLI one Frame\n4. It will be black\n\n**Further information**\nThe config of a default KVM:\n```\nArchitecture: x86_64\nCPU op-mode(s): 32-bit, 64-bit\nByte Order: Little Endian\nAddress sizes: 40 bits physical, 48 bits virtual\nCPU(s): 6\nOn-line CPU(s) list: 0-5\nThread(s) per core: 1\nCore(s) per socket: 6\nSocket(s): 1\nNUMA node(s): 1\nVendor ID: AuthenticAMD\nCPU family: 15\nModel: 6\nModel name: Common KVM processor\nStepping: 1\nCPU MHz: 3399.998\nBogoMIPS: 6799.99\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 384 KiB\nL1i cache: 384 KiB\nL2 cache: 3 MiB\nL3 cache: 16 MiB\nNUMA node0 CPU(s): 0-5\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Spec store bypass: Not affected\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Full AMD retpoline, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx lm rep_good nopl cpuid extd_apicid tsc_known_freq pni cx16 x2apic hypervisor cmp_legacy 3dnowprefetch vmmcal\n l\n```\nThe command that I used to render one frame: \n```\n./blender -b --debug --debug-cycles --python-expr 'import _cycles;print(_cycles.with_openimagedenoise);print(_cycles.with_embree);print(_cycles.with_osl)' \"your path to the .blend\" -f 1 \n```\n\n[oidn_disabled.blend](oidn_disabled.blend)\n[oidn_disabled.txt](oidn_disabled.txt)\n\n[oidn_enabled.blend](oidn_enabled.blend)\n[oidn_enabled.txt](oidn_enabled.txt)\n\n[optix_enabled.blend](optix_enabled.blend)\n[optix_enabled.txt](optix_enabled.txt)\n\n\nIf I set up the KVM to use HOST at least OIDN works\nHOST config:\n```\nVirtualization: AMD-V\nHypervisor vendor: KVM\nVirtualization type: full\nL1d cache: 384 KiB\nL1i cache: 384 KiB\nL2 cache: 3 MiB\nL3 cache: 16 MiB\nNUMA node0 CPU(s): 0-5\nVulnerability Itlb multihit: Not affected\nVulnerability L1tf: Not affected\nVulnerability Mds: Not affected\nVulnerability Meltdown: Not affected\nVulnerability Spec store bypass: Mitigation; Speculative Store Bypass disabled via prctl and seccomp\nVulnerability Spectre v1: Mitigation; usercopy/swapgs barriers and __user pointer sanitization\nVulnerability Spectre v2: Mitigation; Full AMD retpoline, IBPB conditional, STIBP disabled, RSB filling\nVulnerability Srbds: Not affected\nVulnerability Tsx async abort: Not affected\nFlags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid tsc_known_freq pni pclmulqdq ssse3 fma cx1\n 6 sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy svm cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw perfctr_core ssbd ibpb vmmcall fsgsbase tsc_adjust bm\n i1 avx2 smep bmi2 rdseed adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero xsaveerptr virt_ssbd arat npt nrip_save arch_capabilities\n```\n**Why it would be nice to have the bug fixed:**\nWe have a renderfarm that consist out of random machines and it is always a problem if random frames are black/not denoised/strange colored/broken/out of order. ", "EEVEE crash while rendering animation, cycles works.\nOperating system: Win10\nGraphics card: GTX1070MAX-Q\nCPU : i7-6700HQ\nRAM : 32GB DDR4\n\nBroken: 2.90\nWorked: -\n\nBlender crashes w/o any errors when I started to rendering animation in Eevee. \nIt happens on 2 laptops and it always happen in 154 or 155 frame while rendering.\nIn attachment you can download the project which gets this error. \nWhen I use cycles rendering it works perfect.\n\n[#80740.blend](T80740.blend)\n\n- Open file\n- Render frames 154 and 155\n", "Blender 3.5.1 keeps crashing on occasion\nOperating system: Mac OS Ventura 13.3.1 (a) (22E772610a)\nGraphics card: Apple M1 Ultra (64-Core GPU)\n\nBroken: Version 3.5.1 (3.5.1 2023-04-25)\nWorked: N/A\n\nBlender occasionally crashes on my new Mac Studio (I used to work on a Windows PC beforehand and never had that problem). I'm guessing it is a saving error.\n\n- Download and open the demo-file `Blender 3.4 – “Charge” Open Movie` from demo-files/\n- Change the engine from Eevee to Cycles GPU\n- Set \"Viewport Shading\" to \"Rendered\" - nothing happens\n- Go back to Solid View - blender freezes\n\nAttached is is a video demonstrating what is happening, but in a few words:\n\nThe same happens with some of my heavy(ish) project in 3.6.0 beta\n\nIn 3.5.1 the same project responds in render view as it should be.\n\nWhen I try to reopen the same project blender doesn’t load the project but either freezes or crashes ( one of these crash logs is also attached)\nI manage to reopen the project after a few crashes.\n\nI hope it helps! \n\n\n", "Cycles viewport rerender when changing any color management setting if a rigidbody is on any object\nOperating system: WIN 10\nGraphics card: GTX1660\n\nBroken: 3.0.0, 3.1 BETA, 3.2 ALPHA\nWorked: Never\n\nWhen changing any of the color management settings in blender, a viewport rerender will be triggered if you have a rigidbody on any object.\n\nOpen blender\nput renderer to cycles\nput a rigidbody on the default cube (or any object)\ngo into viewport rendered preview mode\nchange any setting in colormanagement (eg. exposure, gamma, false color, medium contrast to high contrast)\nwatch the viewport rerender\n\n[BugTest.blend](BugTest.blend)", "Blender freezes with big custom data\nmacOS 10.14\nBlender 2.92, 2.93.0, \n\nI found if I store big data (like 5 millions of points) then Blender will freeze a bit. This is not a huge bug but it would be nice not to freeze blender.\n[tmp1.blend](tmp1.blend)\n\n\n\nFor example, a console looks like this if I do \"print(test2)\". And there is no freeze in the console.\n\n\n- Run the script\n- Hover over the custom properties section. (try to rotate the cube to observe the freeze, beachball may take time to show up)/\n\n\n\nMemory usage rises and falls on every hover.", "object monitor and menu bar flickers when i move move my cursor\nOperating system: Windows-11-22H2\nGraphics card: NVIDIA GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.92\n\nBroken: version: 3.6.2\n\n\n\n", "VSE strip can't be seen with frame all\nOperating system:\t\tWindows 7\nGPU:\t\t\t\tRadeon Vega 64\nGPU Software Version:\t20.4.2 (2020.0515.1537.28108)\nCPU:\t\t\t\tIntel Core i7-5960X\n\n**Blender Version(s)**\n*(listed are versions tested)*\n\nWorking:\t\t\t\t`2.79 (sub 00), branch: master, commit date: 2018-03-22 14:10, hash: f4dc9f9d68b, type: Release, build date: 2018-03-22, 09:59 AM`\nBroken:\t\t\t\t` blender-2.90.0-ba2d1c889819-windows64` (BF Build Bot)\n\n\nChannel 32 can't been seen with Frame All\n\n\n|  |  |\n| -- | -- |\n\n\nHave at least 2.00 Display Scaling\n\n[Channel_32.blend](Channel_32.blend)", "Blender 3.5 and 3.5.1 are crashing while using Metal as backend\nOperating system: macOS 12.6.3\nGraphics card: Radeon Pro 580 8GB\n\nBroken: 3.50, 3.51 (Metall Backend)\nWorked: 3.50, 3.51 (OpenGL Backend)\n\nBlender simply hang. I use eevee, if you switch to rendered viewport shading it crashed complete system. I use Metall as backend, if I choose OpenGL it works fine. \n\n \nI attached an .blend file for example.\n\n" ]

[ "Blender open OSX can never go to sleep\nOperating system: MAC OS 10.13.6\nGraphics card: AMD HD 6970m 2048mb\n\nBroken: (2.80 Beta macOS 64 bit\nMay 25, 23:44:29 - 36ae49502bbe)\nWorked: (optional)\n\n**Blender open OSX can never go to sleep, see screnshot**\n\n\n\n", "Blender Prevents OSX Mavericks From Sleeping\nOSX 10.9\nIntel HD4000\n\nBroken: (example: 2.69.10 who can remember a hash, revision numbers were more human friendly)\nWorked: (optional)\n\nI am new to OSX but when I have Blender open OSX can never go to sleep or enter power management mode.\n\nOpen Blender on OSX 10.9 or better.\nHide Blender so it is no longer visible on the screen.\nWalk away and get coffee and talk to people at the water cooler.\nGo to sleep or wait and come back to the computer and you will find that the OSX machine has never entered sleep or screen saver mode. NOTE: Blender is not rendering, it is just sitting there idle.\n\nI notice Adobe products do not have this problem and the system works fine with them.\n" ]

[ "Render can not be cancelled and crashes\n\n\n\n\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: AMD Radeon R9 200 Series ATI Technologies Inc. 4.5.13570 Core Profile Context 19.7.2 26.20.13001.16003\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\n\nHello,\n\nI wanted to export some scene using the Eevee render. When I render a frame, the image processing statut is still at 0%. Thus, when I try to cancel it, Blender hangs / crashes.\n\nI launched it using the GPU debug tool\n\nI'm assuming it comes from the hardware.\n\n\nThanks in advance.\n\n\n\n[eevee_logs.txt](eevee_logs.txt)\n\n\n[blender2.8.blend](blender2.8.blend)\n\n\n\n\n\n- Open the file\n- Render image {key F12}\n- Progress will hang at 0% for too long time\n- Cancel rendering\n\nRender job is not cancelled immediately / Blender crashes with low VRAM\n\n", "Cycles GPU/CPU render difference with more than 64 transparent bounces and coplanar faces\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.66\n\nBroken: version: 2.81 (sub 16)\n\n\n\nWhen using particle plane with transparency ( to fake volume scatter ), we have a color mismatch between CPU and GPU rendering. \n\n\n see attach file\n\n\n\n[cpu_gpu_issue.blend](cpu_gpu_issue.blend)\n \n", "Freeze when using Object.ray_cast in a driver script\nOperating system: Linux-5.10.15-1-MANJARO-x86_64-with-arch-Manjaro-Linux 64 Bits\nGraphics card: Radeon RX 580 Series (POLARIS10, DRM 3.40.0, 5.10.15-1-MANJARO, LLVM 11.0.1) AMD 4.6 (Core Profile) Mesa 20.3.4\n\nBroken: version: 2.91.2, 8dc95f61f3\nWorked: Unknown, fails on at least as old as 2.83.10.\n\nRunning a driver that includes a call to Object.ray_cast will freeze the interface entirely. Still exits normally when sent a SIGTERM. It does not create a crash.txt. Blender crashes with a segmentation fault in version 2.83.10 (hash: `b439a155442b`), and does create a crash.txt (see attached). Blender version 2.93.0 Alpha (hash: `7a608f88a713`) also freezes in the same way as 2.91.2. Using the ray_cast function in a non-driver script appears to function as expected.\n\nIn the attached blend file, run the text block `TurretRayCasting.py`, to put the script into the driver namespace. Next, click \"update dependencies\" in the driver panel. The application should freeze entirely. If `return target.ray_cast(turretTargetSpace, direction)- [x]` is commented out and replaced with `return True`, the application does not freeze.\n\n[driver_freeze2.blend](driver_freeze2.blend)\n\n[driver_freeze2.crash.txt](driver_freeze2.crash.txt)\n\n[system-info-2.91.2.txt](system-info-2.91.2.txt)", "VSE: Cycles displayed as Rendered in Scene strip makes Blender hang - and other Scene strip bugs\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 462.59\n\nBroken: version: 2.93.0\n\nThe Render modes of Scene Strips are either missing or broken:\n- When Cycles is set as render and Rendered set as display, it'll make Blender hang.\n- Wireframe and Rendered doesn't work with any Render. \n- The only thing working is Eevee with either solid or Material Preview. \n- Workbench and Cycles settings seems to just use the Eevee render, since these are producing the same results in Solid and Material Preview.\n\n{[F10177251](scenestrip.gif),size=full}\n\nTest file:\nCycles is set up as render in the 3D Footage scene. In the Master Edit scene, switch display mode to Rendered. Blender will hang.\nTry the various other display modes and renders(by changing the render in the 3D Footage scene), to see wireframe and rendered doesn't work with any render. \n[multicam_3d_c.blend](multicam_3d_c.blend)\n\n\n", "Incorrect VRAM quantity in the Status Bar ( GNU/Linux AMDGPU )\nHello, first bug report :)\nA minor bug about the VRAM displayed in the status bar in my GNU/Linux AMDGPU system.\n\nOperating system: Linux-5.4.0-7642-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: Radeon RX 580 Series (POLARIS10, DRM 3.35.0, 5.4.0-7642-generic, LLVM 10.0.0) X.Org 4.6 (Core Profile) Mesa 20.0.8\n\nBroken: version: 2.90.1\nBroken: 2.91.0 alpha\nWorked: No\n\nDisplayed VRAM quantity in the status bar is wrong: 16 GB instead of 8 GB.\n\nIn a similar system enable the Video Monitor option to show the GPU usage in the status bar.\nUploaded an image and system info file:\n\n[system-info.txt](system-info.txt)", "Freezing with Scale tool\nOperating system: Ubuntu 23.04, Ubuntu 22.10, Ubuntu 22.04\nGraphics card: NVIDIA Quadro K620, Intel(R) UHD Graphic 630 (CFL GT2), Intel HD Graphics 520, Intel HD Graphics 615, Intel UHD Graphics 620\n\nBroken: Latest Stable 3.5.1, Stable 3.4.1, Lts Stable 3.3.8, Stable 3.2.2, Stable 3.1.2\nWorked: Stable 3.0.1 (I have not tested other inferior versions)\n\nBlender freezes completely when working with the \"Scale\" tool in Ubuntu. It also froze with the \"Transform\" tool, in Ubuntu.\nCurrently I use Ubuntu 23.04, with latest updates, but it happened with previous versions.\n\n1. Start Blender\n2. Delete all stage elements (cube, lights) (Also freezes with the cube and other elements)\n3. Add > Mesh > Plane\n4. Select Scale Tool\n5. Just pass the mouse over the plane, between the upper and lower circle (with the function to increase and decrease the plane) repeatedly and after some time the blender freezes.\n\n**Notes:**\n- In Blender version 3.1.2 it took a little longer to freeze.\n- Try the versions with Deb, Snap and Flatpak.\n- The problem is the same using Wayland or X11", "Blender crashes after GPU spike (Specially when we have a character with Rigify)\nOperating system: Windows 10\nGraphics card: RTX 2070 Super\n\nBroken: 3.4.1\nWorked: 3.3\n\nEvery time I'm working in Blender, after 2 minutes if I start to move around on a rig made based on Rigify, or work on geometry nodes, Blender Crashes. It happens even faster if I'm working in eeve.\n\n- Open a blender file\n- Import a character using rigfy with Linked Library\n- Library override\n- Turn Eeve on\n- Start working\n- Eventually, blender will crash.\n\nI noticed that if I don't import a character with rigfy, it doesn't have the GPU spike\n[Christmas.crash.txt](Christmas.crash.txt)\n[KalyAndAkrasiel 3.5.crash.txt](KalyAndAkrasiel_3.5.crash.txt)", "Vertex/Edge Slide stuttering with modifier.\nOperating system: Windows 11\nGraphics card: RTX 3070\nCPU: Ryzen 5900x\n\nBroken: 2.83 and all other.\nWorked: 2.79b\n\nVertex/Edge Slide freezes a lot with subdiv modifier in a relatively simple object.\n\n| With `Use Limit Surface` | Without `Use Limit Surface` |\n| -- | -- |\n| https://projects.blender.orgattachment | https://projects.blender.orgattachment |\n\n1. Import attached fbx.\n2. Add subdivision modifier set to 1 subdivision.\n3. Select the loop in the middle.\n4. Edge slide.\n\n", "System Hangs while Interacting with the Viewport\nOperating system: Manjaro KDE Plasma\nGraphics card: Intel HD Graphics (Cherry Trail) Atom X7-Z8700\n\nBroken: 3.5.0, 3.4.0, 3.4.1\nWorked: 3.3.2 (Haven't tried newer versions in the 3.3.x series)\n\nBlender causes the entire system to hang periodically while performing any task. Simply rotating the camera or zooming in or out results in the system freezing for minutes at a time. The mouse cursor responds to mouse and touchpad movement, but the system does not respond to mouse button clicks or keyboard input for long periods of time. Neither Blender nor the desktop respond.\n\nWhen rotating, moving or zooming the camera text overlays in the viewport flicker, sometimes the entire viewport flickers, then the system hangs. There are no errors or warnings printed to the terminal\n\nDoesn't appear to be related to a specific file. The issue is present in all blend files and when starting with the default setup.\n\n", "Black render artifact with CPU+GPU rendering in viewport\nWindows 7 Pro\nNvidia Quadro P6000\nIntel Xeon E5-2690 v4\n\nBroken: daily builds 2.79 since at least 25.10.2017.\nWorked: -\n\nIf I enable combined CPU+GPU rendering, a black rectangle shows up on one of my objects. It shows up in viewport rendering, and also in the final render. This seems to be independent of the material used. I have provided a stripped down version of the scene where it is still visible. This time only in the viewport, the final render works for this simplified scene. \n\nThe artifact seems to be \"locked in space\", I can move or rotate the object away, so that it disappears. In some rotation angles it is not a black rectangle, but dense black lines. \nWhat is also interesting is that the surface of the object in not only black, but partly transparent. I have put a lamp and a cube inside the object in the demo scene, these should not be visible from outside, but they are.\n\n- Open file. In the render settings GPU compute should be turned on, and in the user preferences CUDA should have both the CPU and GPU selected.\n- Turn on viewport rendering. --> black box visible.\n- Then switch to CPU rendering. No more black box.\n\n[black_artifact_bug3.blend](black_artifact_bug3.blend)\n\n", "GPencil: Blender crash when GP3 enabled & you open the default 2D animation file\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 536.23\n\nBroken: version: 4.0.0 Alpha\nWorked: 3.6\n\nBlender crash when GP3 enabled & you open the default 2D animation file & use the Draw tool\n\n- enable GP3 under the experimental features\n- save preferences and restart Blender\n- do File > New > 2D Animation\n- try to draw something.\n\nSince my 2D animation file is customized, I included it below.\n\n", "Bake diffuse causes exits the program and core dump\nOperating system: Ubuntu 20.04.1\nGraphics card: Sapphire RX 5700 XT with latest drivers 20.45\n\nBroken: 2.83.10, 2.83.9, 2.83,8, 2.90.1\nWorked: \n\nI have cube with simple material. When I hit bake (diffuse), Blender exits immediately.\nInfo from terminal: \n\"Memory access fault by GPU node-1 (Agent handle: 0x7f9faac09b00) on address 0x7f9e1782c000. Reason: Page not present or supervisor privilege.\n./apps.sh: line 34: 84735 Aborted (core dumped) /home/karol/port_apps/blender-2.83.10-linux64/blender\"\n\n", "Blender stops updating display after GPU spike\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Plus Graphics Intel 4.5.0 - Build 27.20.100.8681\n\nBroken: version: 2.92.0 Beta\nWorked: I have not found a version that works as expected on this computer. I have tried 2.79, 2.8, 2.91\n\nI was premature in saying the problem is gone for [#85061 ](T85061).\n>After a few minutes of using Blender, the window freezes. Windows does not detect that it is not responding, but Blender does not respond to any input, including clicking on the close button. All elements that are drawn by Blender remain as they were at the moment of the crash except if I resize the window smaller.\n\nIf I watch the performance tab on Windows Task Manager, the freeze coincides with a 100% spike on the GPU.\nAs long as that spike doesn't happen, there is no problem, but once it does, Blender stops responding to any input, but it does not crash.\nI have updated my graphics drivers.\n[blender_debug_output.txt](blender_debug_output.txt)\n\nI have not determined exact steps. It has occurred on multiple files doing multiple different tasks. As far as I can tell, there is no correlation between what I am working on or what task I am performing, and the error occurring.\n", "Mem leak when rendering on GPU\nOperating system: Windows 10 20H2 Build 19042.985\nGraphics card: Gigabyte RTX 30870\n\nBroken: 2.92 - 3.0\n\n\nRender 2.92 splash screen on GPU 60337d495677e942564cce76\n\nOn CPU, memory usage was 18GB, on GPU it went to 60GB and I had to terminate the process.", "GPU: Workarounds and Device Features checks becoming obsolete\nBlender 4.0 bumped its OpenGL version from 3.3 Core profile to 4.3 Core profile. Some workarounds and optional device features were added to the GPU Backend in order to emulate features that were not available on all platforms.\n\nThis issue contains a list of features, workarounds that we should revisit.\n\n* `shader_draw_parameters` In Vulkan these seems to be supported even on low end devices (displayreport.php?id=23923#features_core_11) we might want to consider making this a requirement, although it is not required by vulkan.\n* `compute_shader_support` is required for workbench-next\n* `shader_storage_buffer_objects_support` is required (#112572)\n* `shader_image_load_store_support` is required. \n* `clear_viewport_workaround` should be solved by the editor, not by the gpu backend. IIRC the VSE has fixed this some months ago.\n* remove `--debug-gpu-disable-ssbo` as SSBO's are required. (#112571)\n\n" ]

[ "GPU open CL Cycles rendering not working\nOperating system: Linux-5.3.0-7625-generic-x86_64-with-debian-buster-sid 64 Bits\nGraphics card: Radeon RX 580 Series ATI Technologies Inc. 4.5.13572 Core Profile Context\n\nBroken: version: 2.82 (sub 6)\nWorked: 2.81a works\n\nThis issue occurs only on the alpha/beta versions of Blender. In Blender 2.81a (stable) rendering with openCL goes through the usual procces of compiling the render kernel then everything is working perfectly, however on any release of the 2.82+ when switching to GPU OpenCL rendering blender kinda freezes, the ui seems responsive and interactible but there is nothing happending, can't close, click on anything, no info is displayed either compiling or rendering etc.\nOn the default opening scene i`ve left it for about an hour and nothing changed blender is frozen.\n" ]

[ "UI Accessibility: The color of grid emphasis lines is not light/dark enough to see easily. A color setting for accessibility would fix this.\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.96\n\nBroken: version: 3.3.0 Alpha\n\n\nSometimes the colors of grid lines and the grid emphasis lines are very similar. Especially when zooming in or rotating. Grid color can be controlled in the settings area but not grid emphasis color.\n\n\nJust open a file and zoom so the grid fills the screen. When you are close it is almost impossible to see the difference between grid line and emphasis line.\n\n\n\n\n\n", "Cycles Image Texture Alpha conversion inconsistency with EEVEE and Image Editor\nOperating system: Windows 10 Pro 64-bit (10.0, Build 18363)\nGraphics card: NVIDIA GeForce 920M (Laptop)\n\nBroken: v2.83.13 Beta, 2020-04-21\n\nIf an image is packed, it will get displayed differently in Cycles and EEVEE. It changes in EEVEE when\nAlpha Mode is changed from Straight to Premultiplied to match Cycles, but it does not change in Cycles\nby changing that option. If the image is unpacked Cycles will look like EEVEE with Straight Alpha Mode\nand again it will not be changed by the Alpha Mode option like EEVEE is.\n\n1. Open the provided file.\n2. Select the Image Texture Node\n3. Go into the right panel in the Shader Editor (N-panel) into the Item tab and Properties subsection\n4. Render Image\n5. try changing Alpha Mode and rerendering.\n6. try unpacking the image and rerendering.\n7. try changing the Alpha Mode on unpacked image and rerender.\n[packed_image_cycles_display_inconsistency.blend](packed_image_cycles_display_inconsistency.blend)\n", "Data Transfer modifier works only for Byte Color & Face Corner color attributes\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.12\n\nBroken: version: 3.2.0\n\nIf a target color attribute Domain is *Vertex* or it's Data Type is *Color*, the data are not transferred.\n\n[data-transfer-color-attr-bug.blend](data-transfer-color-attr-bug.blend)\nYou can see in the attached .blend that the Data Transfer modifier is set correctly, but the Col attribute isn't updated.\nAfter doing Convert Attribute -> Domain: *Face Corner*, Data Type: *Byte Color*, it starts to work correctly.\n\n\n", "Design: Shading: Extend Clamp and Map Range nodes to work with Vectors\nI'd like to make these nodes work with vector inputs and outputs.\n\nIt's unclear to me as the the best way of doing this. \n\n\n\nSome options:\n\n - **Separate Nodes** : Create separate versions of these nodes with Vector sockets.\n\n - **Dynamic** : Use a dynamic socket. This is better from a UX perspective but how do we show this? For these two nodes, they are pretty pointless without a value input so the edit control could be hidden. The node then operates accordingly depending on the input link. \n\n\n\n- **Mode Switch** : Add vector input and output sockets and add a mode switch to hide the unwanted input socket. \n\n- **Double Sockets** : Add vector input and output sockets and depending what is connected. For vector only input, `Vector` > `Float & Vector`. For float only input, `Float` > `Float & Vector`. If both inputs are connected do `Float` > `Float` and `Vector` > `Vector`.\n\nFor the combined cases, the node would output and show both Value and Vector.\n\n\n\n\n\n\n\n", "Mix Node Modes\nThe new mix node has different options according to data type.\n\nThe idea of this task is to look at adding the mix modes to float and vector data types.\n\nThe color blending modes could be used but a number of these modes assume that the\ndata is in the range of 0-1.\n\n[D16024: Node: Add mix type to Mix Node](D16024) \nThis a patch adds non-color mix types to float and vector data types.\n\nAdditionally, it would be good to look at how different interpolation modes could also be added. ", "Color picker samples merged (display output) colors on a certain object regardless of \"Sample Merged\" setting\nOperating system: Windows 10 v1909\nGraphics card: NVIDIA GeForce GTX 1660 super\n\nThe bug occured on the latest Steam release of Blender (2.90.1). I also tried to open the project file with the said bug in 2.83 LTS steam beta branch but it happened there too.\n\nWhen trying to texture paint a certain project it ignores the \"Sample Merged\" setting and always picks the display output color, while other objects don't share this problem. \nIt happens regardless of the material or texture I'm editing, it seems to be tied to a single object. For example, If I create a cube and assign the material which the bugged object uses color picking works fine with it, and if I assign any other material to the bugged object it still samples display output color regardless of the settings.\n\nI'm sadly not sure how to reproduce it. Though I appended the bugged object to a fresh file: [color_picker_thing.blend](color_picker_thing.blend)\nIt's system info:\n[system-info.txt](system-info.txt)\nI separated it so you wouldn't have to dig through the mess that is my main project file :D. It retained it's weird inabillity to pick colors though.\n\nI couldn't find a solution to this when I tried to google for it, so my only assumption left here is it being a bug. I'm sorry in advance if something I said isn't clear. If needed, I can provide a whole project or try and provide more specific information if what I wrote here isn't enough as soon as possible.\nCheers!", "Applying Subdivision/Multires Modifier De-merges UV's\n### Update\n\nThe issue of custom-data (UV's specifically) becoming detached remains, this problem existed before b88dd3b8e7, it was just less obvious since selection connected near-by UV's. Now UV's are merged by default at a very low threshold when applying modifiers to mitigate the problem. Keeping this report open as it would be good if subdivision-surface didn't separate UV's, replies to this report contain some notes that could lead to a solution. In practice users should not experience problems caused by this behavior, so marking low priority. ~ @campbellbarton.\n\n----\n\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.14736 Core Profile Context 20.8.3 27.20.12027.1001\n\nBroken: version: 2.91.0 Alpha \nWorked: working in 2.90 \n\nCaused by b88dd3b8e7\n\nWhen applying a subdivision modifier, certain UV's points disconnect from their neighbors. This does not seem to occur with the F3 menu subdivide operation.\n\n- Open default blender scene.\n- Subdivide cube - 1 level and apply it. Do this step 4x\n- In the UV editor use \"Alt select\" to select face loops at random.\n- Alt click around until a face loop is not fully completed, this will reveal unmerged vertices.\n- Check area that did not complete by moving faces, vertices, edges etc to confirm unmerged UV's.\nThis does not seem to be an issue with the newly added 9a52b30cc0 in my eyes, rather an issue with subdividing.\n[UV Disconnect.blend](UV_Disconnect.blend)\n\n", "Vertex paint + x-ray + polygon selection mode = weird shading\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.20\n\nBroken: version: 2.82 (sub 6)\n\n\nFor me it looks like this\n\n\n\n[vertex paint x-ray strange shading.blend](vertex_paint_x-ray_strange_shading.blend)", "OpenColorIO-Config-ACES 0.2.0 breaks in sRGB view transform\nOperating system: macOS-12.3-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.13\n\nBroken: version: 3.2.0 Alpha\n\nI tested the newest OCIO v2 config from Github (v0.2.0)\nin Blender 3.2 Alpha. It was first time I tried that.\nI had to change the OCIO 2.1 version to 2.0 to make it actually work in Blender. I was assured that this shouldn't be a problem.\n\nI rendered a test scene and saw some strange shading errors with the default sRGB view transform.\nSelecting a different view transform like Rec.709 fixes the issue.\n\nI posted about the issue on my blog ?page_id=5517 and on ACESCentral.com \n4354\n\nThe rendered EXR file is fine of course. The problem only occurs in the render view or when I try to export the rendered image.\nJPG works fine, PNG breaks in a different way and creates a strangely bright image.\n\n[Blender_32alpha_OCIOv2_sRGB_bug.blend](Blender_32alpha_OCIOv2_sRGB_bug.blend)\n\n\n\n\n\n\nI also tested the same .blend file on a new M1Max MBP. There I cannot replicate the error.", "Node Socket Name Change results in Cycles Material Rendering Black (name is also identifier, should be read-only?)\nWindows 10\nIntel Core i7-5820k 4.1GHz\n2x Nvidia GTX 970\n\nBroken: 2.79 5bd8ac9\n\nIf you change the name of a Node Socket (not the socket's identifier) of a Cycles Shader Node during execution of a python script, Cycles will render the Material black. \n\nDownload the blend file below:\n[socket_naming_bug.blend](socket_naming_bug.blend)\n\nRun the code in the text editor. \nSwitch to viewport preview.\n\nTo fix the problem, simply reset the two nodes, or add a new Material Output Node and a new Principled Node to the Node Tree. Connect the new nodes together and delete the old ones. \n", "Geometry Nodes: Mismatching data types connections work differently from other node editors\nOperating system: win10\nGraphics card: 1050ti\n\nBroken: 3.5.0\n\nFloat math node overrides color data, even though it is muted. Same setup works in shader editor without problems.\n\nOpen file and activate *Viewer* in GN. Expected result: colored output (judging by shader area behavior)\n\n", "Find solution to unify color picking methods with options for sample space\nReopened as, though the initial issue is invalid, the current solution, as it is, is problematic.\nThe sample_color paint operator {key S} has a problem with discoverability and is separate from both the general UI color picker and the sample tool. This needs a better design solution.\n\n**Old Description** (resolved by sample_color operator)\n\nRight now all color picking is using the display space, which usually makes it (almost) impossible to accurately pick colors from an image that is not sRGB for texture painting.\nSimilarly, for texture painting in the 3D viewport the picked color is that after shading and display transformation. This can be useful, but it should be optional. In a lot of cases the user wants to pick and draw a color from within the same image texture, without any additional color space conversion.\n\nThe View mode of the image editor already has a sample tool, which has the functionality to show the color data in image space without additional color management. So simply using this tool in paint mode for picking color with the option to select the color space would already be a huge benefit.\n\nFurther, having a similar tool for 3D texture painting would be even better, but I assume that this might have more difficult implications for the implementation. Tackling this for the image editor first using the Sample tool, might me a good trade-off for investing a small amount of work for maximum impact.\n\n[color_picker_behaviour-2021-03-02_15.50.34.mp4](color_picker_behaviour-2021-03-02_15.50.34.mp4)\n[non-color_picker.blend](non-color_picker.blend)", "Cycles Volume Scatter Density broken in \"complex\" node setup\nOperating system: Linux-4.12.14-lp151.28.87-default-x86_64-with-glibc2.17 64 Bits\nGraphics card: GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.27.04\n\nBroken: version: 2.91.0\n\n\nWhen constructing this node setup, the density-slider of the second volume scatter node seems to suddenly only act in a binary way, making the volume either transparent or taking on the density-value of the other volume scatter node. The point at which this happens can be influenced by changing either the colour of the node or the factor-value of the second mix shader. The settings of the other nodes seem to have no effect on this.\n*Edit: After further testing, the Volume Absorbtion and first Mix Shader Nodes are not needed.*\n\n\n*Load attached .blend\n*change density of the second (marked in image) Volume Scatter node (here, the switch happens somewhere between .003 and .004)\n\n[volume_bug_simplified.blend](volume_bug_simplified.blend)\n\n{[F9550433](blender_volume_bug.png), size=full}\n\nThanks for taking a look!\n", "Viewpor render includes color management, even if the view shading mode does not\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.77\n\nBroken: version: 3.1.2\n\nA viewport render should look exactly as what's seen in the viewport, especially if that viewport is set to, for example, solid shading. Right now, it does not, because color management is taken into account even in shading modes which do not use it.\n\n1. Set viewport to shaded mode\n2. Set viewport background to a white color\n\n3. Set something in the color management panel to something crazy\n4. Observe how the viewport does not change.\n5. Make a viewport render\n\nResult:\n\nThe viewport render includes the crazy values from color management.\n\nExpected:\n\nIt should look like the viewport and nothing else. It's *technically* not a viewport render currently.", "Alt-Drag (multi-editing) fails for HSV color buttons\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce RTX 2070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.28\n\nBroken: version: 2.83 (sub 1)\nSeems to be a long-lasting issue, as even Blender 2.81a shows the same behaviour.\n\nPretty much the title. \nTrying to assign an object color through the Object Properties tab (Viewport Display > Color) to multiple Objects doesn't work, neither by picking the Color on the Color Wheel, nor by moving the HSV sliders, while Alt-Clicking. Only the last selected Object will update.\nHowever:\n\n\n - If sliders are set to RGB, editing them will correctly affect all selected Objects.\n - The Alpha slider, which is shared between all three Color Wheel modes, works as intended and affects all selected Objects.\n\n\nFrom the Info console I've noticed a tuple is set everytime a Color is picked from the Color Wheel, but doesn't show anything when using the HSV Sliders.\nMoving the the RGB and Alpha sliders correctly sets each Object's property to the selected value.\n\n\n\n\n - Open Blender, default startup file\n\n - Add>Mesh - Select any kind of Mesh Object (have at least 2 items in the scene, or more)\n\n - Go to the Shading Panel in the top right and Select \"Object\" down the Color section, to see Object Colors in the viewport\n\n - Press A to select All Objects, or alternatively, select them one by one by Shift+Right Click\n\n - Go to Object Properties > Viewport Display > Color\n\n - Hold ALT either while clicking on the Color Wheel or while moving the HSV Sliders. Only the last selected Object will update.\n\n\n\n\n - While holding ALT, try moving the Alpha slider. All selected Object will update.\n\n - In the Color Wheel, select the RGB tab. While holding ALT, try moving the R, G or B slider. All selected Object will update.\n\n\n\n\nHere's a quick demonstration:\n\n[2020-01-20_22-01-14.mp4](2020-01-20_22-01-14.mp4)\n\nAnd the Info panel readings:\n[2020-01-20_22-02-46.mp4](2020-01-20_22-02-46.mp4)\n" ]

[ "Hue correct node does not match other hsv nodes\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 ATI Technologies Inc. 4.5.0 Core Profile Context 23.4.3.230420\n\nBroken: version: 3.6.0 Alpha\n\nunmodified hue correct node giving diff results than hsv node. The color should not be modified by this node if no settings have been changed and both node should have the same output util modified. \n\nwill reference this post 107583\n\nplugin the viewer to the hue correct and hsv node and the colors will be diff. there is no modification on these nodes\n\n\n" ]

[ "No way to distinguish enum buttons that support shift-click to select multiple\nBlender has some buttons that allow shift-click to select multiple, but no way to show the difference. eg:\n\nTransform Tool, Gizmo's (Move / Rotate / Scale)", "Custom properties without a name cannot be edited/deleted\nOperating system: macOS-12.0.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 560X OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.7.29\n\nBroken: version: 3.1.0 Alpha\nWorked: 2.93.6 Release Candidate\n\nA custom property without a name cannot be edited or deleted (gives an error), however in the worked version as stated above, the unnamed property can still be manipulated without triggering an error.\n\n1. In the default scene, create a new custom property for the cube and remove its name completely.\n\n2. Press OK\n3. Blender gives an error, as shown below:\n\n4. Dismiss the pop up and click on either the gear/cross icons:\n\n\n[unnamed_prop.blend](unnamed_prop.blend)\n", "Blender Crashes when deselecting Name-fields\nOperating system: Windows-10\nGraphics card: NVIDIA GeForce RTX 3070\n\nalso tested on:\nOperating system: Windows-11 Home - Microsoft Surface\nGraphics card: - / 11th Gen Intel(R) Core(TM) i7-1185G7\n\nBroken: (tested with)\n- version: 3.6.2, branch: blender-v3.6-release, commit date: 2023-08-16 16:43, hash: e53e55951e7a, type: release\n- version: 3.5.1, branch: blender-v3.5-release, Date: 2023-04-24 18:11, hash: e1ccd9d4a1d3\n\nWorked: -\n\nFor a specific project, if I select any Text field that contains the name of an asset (texture, object, scene, ...) and click anywhere else within the Blender window, Blender will crash. \n\nThis error does not appear for other projects of mine or a new one.\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n1. Open file\n2. Navigate to Texture\n3. Create a new Texture\n4. Select Texture name (no need to alter the name to reproduce the error)\n5. Click into e.g. Viewport or hit Enter\n\nI created a copy of the problematic project, deleted all of its content and used the Blender cleanup feature to remove any remaining data blocks. Still the issue persists. Except for the missing shutdown logs due to the crash, the Debug Log doesn't show any differences to a normal file AFAIK.\n\nBlender does NOT crash when I use \"Escape\" to exit the field.\nTo be precise, the crash happens when deselecting \"Name Fields\" in the outliner, the Properties Panel or similar.\nThis doesn't happen in Search fields or Value fields that take e.g. floats for properties like \"Strength\" etc.\n\nThe project (used as asset library) worked fine yesterday and as far as I'm aware no updates were installed. I tried updating my NVIDIA driver but to no avail.\n\nNext, I'll try to copy my assets to a new file.", "Multi-object-Mode: EditMode transform properties\nEdit-mode buttons currently only show median values for the active edit-object.\n\nThe Mean Crease value of the N panel doesn't work for multiple objects:\nHere i have 2 objects selected\n\nIn Edit Mode changing the Mean Crease value does change its effect only on the selected object:\n", "Node Group usability tasks for Node assets\n* Enum socket\n* Custom warnings and errors\n* Panels / Sub-panels\n* Dynamic UI changes\n* Surface choice for inputs in modifiers\n\n", "Sculpt Mode: Edit Voxel Size hotkey conflict\nOperating system: Win 10\n\nBroken: blender-2.90-51ac411ce80c-windows64\n\nThe \"Edit Voxel Size\" and the \"Repeat Last\" functions are using the same hotkey (Shift+R), and as a result, the \"Repeat Last\" command is not working in sculpt mode.\n", " Annotations are blended incorrectly (looks like added?) on viewport backgrounds in viewport renders\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER\n\nBroken: version: 3.0.0\n\nBlender renders colorful annotations as white while using the viewport render function, when the viewport background is set to be completely white.\n\n- Draw an annotation (e.g. blue)\n- Set the viewport background (via Viewport Shading) to \"Viewport\" and set it to be red (could be world as well if you like).\n- 3D Viewport > View > Viewport Render Image.\n- notice the blue annotation being added to red background (resulting in pink)\n- if you do this with a white background, you will think the annotations are gone (all strokes add to white... giving.... white :)\n\n[#93748.blend](T93748.blend)\n\n\n", "multiview: overwrite and file naming \n\n\"not overwrite\" option doesn't work with multiview when some views already exist on disk, could be a minor problem but rendering one single view with views option activated create a 0 byte \"myfile_L.jpg\" file but save only myfile.jpg\n\n\n\n\n\n", "Undo in Armature Edit Mode does not undo Custom Properties\nOperating system: Ubuntu 20.04\nGraphics card: Geforce MX230\n\nBroken: 3.5 Beta\n\nUndo in Armature Edit Mode does not undo changes to Armature Custom Properties\n\n- [optionally] In default scene delete all objects\n- Add an \"Armature\" object\n- Go to `Edit Mode`\n- In `Object Data Properties` add a `Custom Property`\n- Change the value of the property (to for example `0.0`) in order to add an Undo step\n- Try undo (`Ctrl+Z`)\n\nThe step is not undone in edit mode.\nInterestingly, changes to the Custom Properties of Bones are correctly undone in Edit mode.\nThis makes it seem that changes to Armature Custom Properties were perhaps missed in the Armature Edit Mode Undo Context.\n\n---\n**Original Report**\n[Video](3RZNWd5tRQI)\nNote when changing the **head size** in the redo panel, you can observe how the number of elements in in the Collection Property `bones_for_delete` grows.\n\n1. Execute the script generate_rig.py from the file, if you do not have automatic script execution enabled\n2. After execution, the Generate Rig panel will appear in the Rig category\n3. Select Empty.021 first, then Armature.001. Switch to armature editing mode, and click Generate Rig on the panel\n4. The bones will appear, and also the text on the panel will notify that there are elements in bones_for_delete. After that, you can scroll the **head size** on the redo panel, or simply undo the operator. the label in the panel will notify you that the number of items will not change\n\n", "Crash upon renaming collection or clicking on Blenderkit after creating said collection (without renaming)\nOperating system: Win 10\nGraphics card: Nvidia Quadro M5000 \n\nBroken: 3.4.1\nWorked: 3.4.1\n\nCrash upon renaming collection or clicking on Blenderkit after creating said collection (without renaming). \n\n- Download file through the link from Google Drive attached (I leave the file as it is since I don't know what caused the problem)\n- Click in the collection \"Möblierung\",\n- create some new collection in it,\n- rename the new collection (with double click)\n- hit enter - crash\n\nOr create the collection, don't rename it and then click something in the interface (Blenderkit \"eye\" symbol in my case).\n\nReload didn't do anything, just kept crashing.\n\nLoaded a recent autosave, seems to work. Already hat to kill startup file and all settings yesterday, got permanent exception access violation crashes upon file load, after enabling some addon (TrueAssets) it worked.\n\nLink to the blend: view?usp=sharing\n\n", "Custom Property list not updating in UI when editing properties from Python\nOperating system: Arch Linux\nGraphics card: GTX 970\n\nBroken: 2.80.60 (1c5860407068)\n\n\nSetting a custom property from Python does not update the corresponding entry under Custom Properties until mouse over forces a refresh.\n\n\nUsing default blend file.\n\n1. Add a Python Console panel somewhere\n2. Select the cube object\n3. Switch to the object properties tab, unfold the Custom Properties panel\n4. In the python console:\n```\no = C.active_object\no['value'] = 456\n```\n5. No new entry is shown in the Custom Properties panel\n6. Move the mouse to the Custom Properties panel -> new value prop appears\n7. The same thing happens when updating a property value, a mouse over is also needed to refresh the property list being displayed\n", "Breaking UI API changes for 4.0\n- [ ] Remove grid display for UI lists - Never worked well and is badly maintained. May have been added for an oudated asset UI design even?\n- [ ] Remove asset-view UI template - Should be replaced by the [asset shelf](102879).\n- [ ] Replace/deprecate UI lists? - Probably too much work, would require UI views to be exposed to Python. But can at least be converted to use UI views internally and benefit from its features.\n\n", "Multi-Object Properties Editing\nIn Blender, we want to make it easier to set values for multiple objects, bones, strips, keys and other items at once. \n\nCurrently you can do this in limited circumstances, but you have to right-click on a value after you've changed it and pick Copy to Selected from the menu or hold Alt while editing values. This not only slow, but also very hidden, and the Alt-key conflicts with Emulate 3 Button Mouse. Lastly, it isn't clear at all to users which values work for multi-editing, and which values don't.\n\nHere's how we would like to solve it:\n\n\n## Selection vs Active\n\nThe concept of ***active*** in Blender means that we only show properties and only affect the active object or item. However, if we want to better support multi-item editing, we must start by making it so the Properties reflects the ***selection*** rather than the ***active*** item, of which there can only be one. Going forward, by default the Properties will be set to reflect the selection instead. This could be controlled in the Properties like so:\n\n{[F8363009](Screenshot_2020-02-22_at_17.34.54.png), size=full}\n*(The search box is added here to reflect how the Properties will be updated when that item is also added, but is a separate feature)*\n\nIf set to ***Active***, the Properties will continue to function as it does today, only showing properties for the active item.\n\n## When there are multiple selected items\n\nWhen the user has selected multiple items, a few things change:\n\nInstead of showing the name of the object at the top, Blender tells the user how many items are selected:\n\n\nIn the list of tabs, only properties sections are shown that are common among the selected items. For example, if you select a text and a mesh object, we don't show Modifier, Mesh or Curve Properties, since these properties only exist on a subset of the selected objects:\n\n\nThis rule also goes for individual panels. Here, both a Sun and Spot light are selected, and we only see properties that are shared between them. A note at the bottom clarifies that you are only seeing the shared subset:\n\n{[F8364914](Screenshot_2020-02-23_at_18.49.33.png), size=full}\n\n## When values differ\nWhen multiple items are selected, Blender will compare values to check if they are the same, or if they are different. If they differ, we change the way properties are displayed, like so:\n\n{[F8363037](Screenshot_2020-02-22_at_17.52.51.png), size=full}\n\n\n\n## Manipulating values across multiple items\n\n\n\nFor simply setting a value to be the same. you would simply type a new value, hit Return, and that value is then propagated to the selection.\n\nFor manipulating numbers and sliders relatively, you can just drag on them (like Alt-dragging today), or type the character '=', so the user can type '= -3' to subtract 3 from the value on all selected items.", "Can't bind shortcut Alt+Numbers 0-9\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\nKeyboard: UK ISO (tenlesskey) w/ UK Layout\n\nBroken: version: 2.93.4\nBroken: version: 3.0.0 Alpha\nWorked: None (tried 1.8, 2.28, 2.34, 2.58, 2.66, 2.79, 2.80, 2.82)\n\nUnable to use Alt modifier with Numbers 0-9\n\nTry to rebind something (e.g Hide Collections to Alt+Number key).\nThe binds will fail to execute.\n\n{[F10604609](210926_blender_10b48.png), size=full}\n\nBroken:\n• Alt+Numbers 0-9\n• Alt+Numpad Numbers 0-9\n• Using 'Left Alt' as a custom modifier key in the keymap plus Numbers 0-9\n\nWorks:\nAlt+F Keys 1-24 work (unfortunately Alt+F4 is hardcoded to exit blender.exe)\nAlt+Qwerty/Punctuation keys work.. just seems like its Alt+Numbers that fail\n\nThank you for your time and consideration :3", "File Browser unable to rename with Upper/Lower case\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.83.5\nversion 2.90.0 as well.\n\nWhen renaming a file/folder case changes are not taken into account.\n\nOn default startup : \nFile -> Save As -> New Folder -> Type \"TEst\" as name, click to release focus.\nRight Click -> Rename (or F2); Type \"Test\", click to release focus.\nThe name changes back to \"TEst\".\nWorkaround : \nRename to \"Test 2\", then back to \"Test\".\n\n\n\n" ]

[ "Can't rename Annotation Layers\nOperating system: Windows 10, latest updates\nGraphics card: NVIDIA GeForce RTX 2070/PCIe/SSE2\n\nBroken: 3.1.2 shipping; blender-3.3.0-alpha+master.9531eb24b317-windows.amd64-release\nWorked: Unknown\n\nThe Annotations system lets you create annotations on named layers. The hover hint for each layer says \"Double click to rename.\" but that doesn't work. Nor does F2.\n[CantRenameAnnotationLayer.blend](CantRenameAnnotationLayer.blend)\n\n- Open the attached CantRenameAnnotationLayer.blend\n- Activate one of the Annotate tools\n- in the sidebar Tool menu, hover over the layer name field to the right of the Color swatch, which causes the layers info to display\n- DBLCLICK on Note.001\n\n**Expected/Actual**\nExpect the name text to enter edit mode so you can rename.\nActual: nothing happens. Cannot edit the layer name.", "Doubleclick not working in UI_Lists in popovers (e.g., grease pencil /annotations layers renaming)\nBlender 2.81, 2.82, ...\nWorked: ?\n\nUnable to rename grease pencil layer when tooltip says: 'Active layer index, Double click to rename.'\n\n\n- Go to either '2D Full Canvas' or ' 2D Animation', \n- Add a blank Grease Pencil, \n- Change mode to Draw\n- Drag the header bar to left to review 'Layer:'\n- Click the down arrow to show the layer's panel\n- Press + there to add a new layer\n- Select either old or new layer and while leaving mouse pointer hover over it, you should see a tooltip 'Active layer index, Double click to rename.' Try to double click on it and 'NOTHING HAPPENS'. \n- In 2D Animation layout you can access the 'Properties' panel on the left and rename there, in 2D Full canvas, you'll have to split window, creating a new one, change it to Properties before you can do the 'RENAME'.\n\nScreenshot for ' 2D Full canvas' layout, rename is IMPOSSIBLE:\n\n\nScreenshot for '2D Animation' layout, rename is IMPOSSIBLE:\n\n\nScreenshot for the same control in Properties panel, where RENAME is possible:\n", "Impossible to rename annotations notes\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.2.1\n\n\nAfter creating an annotations note, it is simply impossible to rename it (by double clicking) although the tool-tip proposes it. \n\n\nCreate an annotations note, go to the annotation list, and try to rename the created note by double clicking.\n\n", "Unable to rename annotation layer via topbar\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: Quadro RTX 3000/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.69\n\nBroken: version: 3.1.2\nWorked: ?\n\nUnable to rename annotation layer via topbar by double clic as expected\n\n- Draw an annotation.\n- Go to the topbar popover with annotation layers.\n- Try to double clic in a layer to rename : nothing happens.\n\n\n\nNota : no issue with view panel in sidebar.\n\n" ]

[ "Cycles viewport render doesn't resize if an animation is playing.\nOperating system: macOS-13.0-arm64-arm-64bit 64 Bits\nGraphics card: Apple `M1` Pro Apple 4.1 Metal - 83\n\nBroken version: 3.5.0 Alpha\nBroken version: 3.3.1\n\nIf a .blend file contains a animation, and the animation is playing, and you're in the Cycles rendered viewport mode, then the viewport resolution/size will not update as you adjust it's size.\n\n1. Create a scene that has a colour that differs from the standard background colour of the 3D viewport.\n2. Animate something in the scene (E.G. Animate the movement of an object)\n3. Change the render engine to Cycles (This issue only appears in Cycles for me.)\n4. Enter the rendered viewport mode so you're now seeing a Cycles viewport.\n5. Start playing the animation.\n6. While the animation is playing, start adjusting the size of the 3D viewport editor and notice how the Cycles rendered view isn't updating with the change in size. This is the issue. Once you stop playing the animation, it will update.\n\nHere is a video demonstrating the issue:\n\n[Cycles preview doesn't resize.mp4](Cycles_preview_doesn_t_resize.mp4)\n", "When changing the width of frame, it also changes frame location.\nIf you create frame and set it into (0,0) location. Then grab right side of the frame node and make frames width bigger. Now using python read the frame \nlocation again and it is not (0,0) anymore. \n\n", "Animation Player crashes on MacOS\n\n```\nOperating system: macOS-11.6.1-x86_64-i386-64bit 64 Bits or OSX 12.2.1\nGraphics card: AMD Radeon Pro 580 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.6.20\n```\n\n```\nmacOS 12.2.1 Intel HD Graphics 6000 1536 MB and Radeon Pro 5300M, macOS 11.4\n```\n\nBroken: version: 3.0.0 Beta\nBroken: blender-3.2.0-alpha+master.d3b1cce4000b-darwin.x86_64-release.dmg, \n\nI open a fresh Blender, Standard Cube scene.\nGo to View-> Viewport Render Animation\nIt freezes (Beachball of death) around frame 20 or 30 . Have to kill Blender, fan goes crazy.\nSame with the default cube scene. But the the crash doesn't come every time, but the image it way too dark.\n\n\nProblem 1 \n\nFreeze during \" Viewport Animation Render\"\n- Default scene, \"View\" menu -> \"Viewport Animation Render\".\n- after a couple dozen frames, Blender freezes, full CPU usage, and frames don't progress.\n\nProblem 2\nAssuming the above problem doesn't appear, \n- Default scene, \"View\" menu -> \"Viewport Animation Render\".\n- Render-> View Animation\n- Playback crashes at the end of caching.\n\n\nProblem 3\n- The image is much too dark compared to the Viewport.\n", "Push/Relax/Breakdown pose operators may flip rotation of quaternion bones\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 2.93.8\n\nPush/Relax/Breakdown pose operators may flip rotation of quaternion bones\nWhen the animation plays, this bone suddenly rotates unnaturally.\n\n - Open the attached blend file\n - Use Push/Relax/Breakdown to the bone and set a keyframe.\n - Check the bone animation.\n[flip_when_use_breakdown.blend](flip_when_use_breakdown.blend)", "Ability to cancel viewport operations (rotate/pan/zoom)\nSince most transforming view-port operations can be cancelled, it would be useful if viewport operations such as rotate/pan/zoom could be cancelled by pressing Escape. [See request ](showthread.php?392619-Call-for-Quick-Hack-ideas-(for-beginner-students-to-complate-in-a-day)&p=3010312&viewfull=1#post3010312).\n\nWhile not high priority, this is nice to support for consistency with other parts of Blender.", "ascending/descending for List View not working correctly\nOperating system: Linux-5.11.0-40-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.82.00\n\nBroken: version: blender-3.0.0-beta+v30.7b530c609611-linux.x86_64-release AND blender-3.1.0-alpha+master.f56562043521-linux.x86_64-release\nWorked: unsure\n\nWhen alphabetical order is selected for a list view, the descending order and ascending order button does not work correctly. \n\n1. add a variety of groups or entries into the list view\n2. scramble their order so it is not alphabetical\n2. select alphabetical ordering\n3. select descending \nHere is a video of the behavior, queued up ...\nwatch?v=gyqZB2qZ3R0&t=266s\n\n", "Quad view toggle changes \"main view\" orientation with lock rotation is unchecked\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: AMD Radeon HD 6700 Series ATI Technologies Inc. 4.5.13417 Core Profile Context 15.301.1901.0\n\nBroken: version: 2.82 (sub 7)\nWorked: Never (2.79+)\n\nQuad view toggle changes main view orientation with lock rotation is unchecked\n[Project 121.mpg](Project_121.mpg)\n\n- Open general template\n- ctrl + alt + q to toggle quad view\n- In side panel > Quad View uncheck Lock\n- Move mouse to any dedicated ortho views\n# Hit ctrl + alt + q twice\n\nTop right view - \"main\" view has orientation of view where mouse cursor was.\n", "Unable to flip side with \"Bridge Edge Loops\" operator\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 3.6.0 Beta\n\nwhen bridging 2 opposite faces that face away from each other, the bridge inversely connects or chooses the shortest path and there is no way to adjust the smoothness profile to invert the connection to the positive side, the image below shows the difference between blenders bridge tool on the left and Quick Bridge Addon connection on the right, I think this is an important problem as there are no easy way around this to achieve the same result, I hope you can fix it , maybe allow a negative input in smoothness or something, Thank you\n\n\n", "Multiresolution displacement baking do not work properly\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.93.0 Beta\n\nOriginal on the left produces map, thad displace surface like on the right.\nI suppose, there is wrong gamut/ color conversion or baking ray shoots from Hipoly to lowpoly instead measuring height difference from lowpoly to hipoly. Displacement baking is not equal to normal baking!\n\n{[F10039179](изображение.png) size=full}\n\n[mr bake.blend](mr_bake.blend)\n\n[2021-04-26_15-44-07.mp4](2021-04-26_15-44-07.mp4)", "View Pan not working for smaller clip start value\nOperating system: win 10 pro\nGraphics card: rtx 2060\n\nBroken: 3.5, 2.83\nWorked: never?\n\nFor smaller `clip start` value when we zoom out, `View pan` does not work\n\n- Open .blend file\n- Pan around by any method\nIncrease `clip start` value, this will fix panning problem\n[Sand_003.blend](Sand_003.blend)", "With FPS change, VSE does not change clip start frame\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.4.1\nAlso problem in 2.92, 3.3.2, 3.5 alpha\n\nWith a FPS change, VSE changes the duration of an audio clip but not the start frame and the clips become misaligned.\n\nBest seen with two clips starting at different frames. See images. In my example the audio clips are mono at 11025 Hz. (Seems video clips act the same way.)\n\n{[F14133157](24.png),size=full}\n{[F14133163](6.png),size=full}\n\nI know there are several somewhat related reports re: VSE and FPS. Hope not a duplicate.\n\nThanks!!! (for all the great work!)\n\n", "Cycles GPU crash when switching to Rendered 3DViewport --MacOS 13.1--\nOperating system: macOS-13.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.9.51\n\nBroken: version: 3.5.0 Alpha\nWorked: 3.5.0 Alpha (works fine)\n\nInstant crash of blender when switching to rendered 3DView.\n\nCompile latest 3.5 version as of 2022-12-18 20:06, hash: `50c7eb14f4`\nopen empty scene provided below and switch to rendered 3DView.\n[gpu.blend](gpu.blend)\n\n\nxcode 14.1\nlatest brew updates\n(cycles-cache & build-darwin folders deleted before compiling )\n\nPS: When switching to CPU there is no crash. Only GPU !", "LightGroups & big render resolution results in Hardware failure & macOS crash incl. reboot\nOperating system: macOS-12.3.1-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.15\n\n**Blender VersionS**\nBroken: version: 3.2.0 Alpha\nBroken: version: 3.2.0 Alpha\n\n\n(GPU only)\nI usually render with a resolution of 5120x2048 on 50% as check resolution. \nWhen I scale it up with the slider to 100% and render I get a stripe pattern as we seen in other bugreports related to M1 Mac's.\nThen when I try double the resolution on 50% I get the same.\n\nThe crash of macOS with a hardware failure and reboot happened when I tried my final resolution 10240x2048 on 100%. (blender calculated a sec. and when the render via GPU only started it hanged for two sec. and then completely crashed)\nTo be honest, a crash like this I never had with 3 imac's the last 6 years. Only with a hackintosh not related to blender...\n\nDespite the fact of the total crash, I decided not to try to reproduce the total crash,\nBUT...\nthe stripes I could with the resolution I used. \n- Open attached file\n- Render once in 50% from 5120x2048,\n- and then render on 100%, or try the higher resolution.\n\n(The scene was: just a plane, monkey and 3 lights, to check the lightgroups and compositing options (worked all fine, exept the render...))\n(see attached blendfile)\n\nSo just 3 pointlights, each into a one of the 3 lightgroups a plane, monkey...(something to get lit),\nresolution 5120x2048 on 50% and render. Then render on 100% and you should get stripes.\n\n`***When there, just delete the lightgroups under Scene -> View Layer Tab and the stripes are gone.***`\nOperating system: (INTEL) macOS-12.3.1-x86_64-i386-64bit 64 Bits \nGraphics card: AMD Radeon Pro 5700 XT OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.15\n[untitled.blend](untitled.blend)\n", "Suzanne (monkey) UV map is not symmetrical and has bad stretching around the beginning of the ear\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: NVIDIA GeForce 940MX/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.11\n\nBroken: has never worked as expected\nWorked: never worked, looking at the original task for default uv map for monkey primitive, it's wrong there also.\nT47488\n\nThe current UV map for the monkey default model is not symmetrical and has bad stretching around the beginning of the ear. \n\nNot symmetrical\n\n\n\nBad stretching\n\nExpected result\n\n\nThis also effects the way monkey was uv unwrapped since the uv islands cannot be unfolded correctly in a straight and symmetrical line due to the seams being misplaced.\ncurrent\n\nexpected\n\nexpected result .blend\n[Fix_Monkey_UVMap.blend](Fix_Monkey_UVMap.blend)\nmade a quick proposal for an improved version also for the eyes, currently they are rotated in a way which makes it bad for texturing since the margins are so thin in some places that the texturer can bleed over due to midmaping in some rare cases, this version fixes it. Also looks nicer. \n\n[Improved_Monkey_UVMap.blend](Improved_Monkey_UVMap.blend)\nshift a, add mesh, monkey\n\nenter edit mode and select all uvs in uv editor, UV, seams from islands\n\nAdd uv cheacker grid and look closely at where the ears start and notice that the seams are not symmetrical and some bad stretching around the beginning of the ear. \n\nbug .blend:\n[Suzanne (monkey) UV map Bug.blend](Suzanne__monkey__UV_map_Bug.blend)\nexpected result .blend:\n[Fix_Monkey_UVMap.blend](Fix_Monkey_UVMap.blend)\nimproved version .blend\n[Improved_Monkey_UVMap.blend](Improved_Monkey_UVMap.blend)\n", "Crash when rendering animation\nOperating system: macOS-13.4.1-arm64-arm-64bit 64 Bits\nGraphics card: Metal API Apple M1 Pro 1.2\n\nBroken: version: 3.6.0\n\n\nI was rendering a simple camera animation overnight, Blender crashed about 2/3rd the way through. Here's the crash log.\n" ]

[ "Imported video is wrong orientation on Mac OS Arm.\nI'm importing video on Mac OS Arm (3.0 stable) and the video is flipped and rotated. The video looks fine on other platforms.[bug.zip](bug.zip)\n\n\n\n" ]

[ "Vertex slide (GG) acts weird with zoom (steppy, jumping, not slowing down with shift)\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.19\n\nBroken: version: 2.82 (sub 7)\n\n\n[2020-02-22_06-30-18.mp4](2020-02-22_06-30-18.mp4)\n\n[gg.blend](gg.blend)\n\n", "Mesh edit, Shift+V vertex slide adjustment produces unexpected result\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.92.0\n\nInstead adjustment movement with slider i have broken shape:\n[2021-04-30_15-00-38.mp4](2021-04-30_15-00-38.mp4)\n\n", "Simulation Nodes: show bakes in outliner\n_No response_\n\n", "Regression: Guiding in manta is broken\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.29\n\nBroken: version: 3.5.0\nWorked: It was working in 3.2.1\n\nIt looks like this was caused by 05952aa94d33eeb504fa63618ba35c2bcc8bd19b\n(there is a bit of uncertainty to this though since with that commit, I dont get mentioned error in the console, bake is broken nonetheless with that commit)\n\nIf you us a smoke domain to guid a liquid you get no result because there is an error in the console that shows that two variables are undefined.\n\n```\nTraceback (most recent call last):\n File \"<string>\", line 1, in <module>\n File \"<string>\", line 250, in bake_guiding_166\n File \"<string>\", line 235, in bake_guiding_process_166\nNameError: name 'x_guidevel_s166' is not defined\nTraceback (most recent call last):\n File \"<string>\", line 1, in <module>\n File \"<string>\", line 169, in bake_fluid_data_166\n File \"<string>\", line 164, in bake_fluid_process_data_166\n File \"<string>\", line 80, in liquid_adaptive_step_166\n File \"<string>\", line 334, in fluid_pre_step_166\nNameError: name 'weightGuide_s166' is not defined\n```\n\n1. Open the file.\n2. Configure the cache folders in both domains.\n3. Bake the smoke domain.\n4. Bake the liquid domain (you will see the error in the console and no liquid in the scene).", "Vector displacement on top of adaptive subdivision breaks normal\nOperating system: Windows 10 / Ubuntu 22.04\nGraphics card: NVidia A6000 / K5100M\n\nBroken: 4.0\nWorked: never\n\n\nVector displacement on top of adaptive subdivision breaks normal.\n\n\n- Open attached file, switch to render preview in viewport.\n- It's in adaptive subdivision, where the vector displacement is connected.\n - Wrong normal in the \"hanging over\" side (in black).\n- If unmute the Geometry \"True Normal\" node, the normal seems to be correct but it lacks smooth interpolation (observable upon close inspection.\n- Add a displacement modifier (thus disabling the adaptive subdivision), and remove the displacement output link in the material.\n- Set the displacement modifier to the displacement texture (use RGB to XYZ mode).\n - Vector displacement renders normal correctly.\n\n| VD on adaptive subdiv | VD on adaptive subdiv (with true normal) | VD modifier on regular subdiv (correct) |\n|---|---|---|\n|  |  |  |\n\n", "Geometry Node: Viewer output for instance domain ignore if value nearest to 0.01\nOperating system: win 10\nGraphics card: 1050ti\n\nBroken: 3.4, 3.5\n\nGeometry Node: Viewer output for instance domain ignore if value nearest to 0.01\n\nIf you explicitly specify the instance domain, then with a value nearest to 0.01, the viewer will ignore the value input\n\n[2022-11-11 23-41-29.mp4](2022-11-11_23-41-29.mp4)\n\n\n- Put some intance to Viewer\n- Connect Float Value to Viewer set to instance domain\n- Test various floats (0.01 and 0.3)\n\nOr in attached file switch between couple of Viewer nodes\n[gn_viewer_value.blend](gn_viewer_value.blend)", "Glitches on Video Sequence timeline\nOperating system: Linux 5.9.11\nGraphics card: Intel Haswell\n\n[system-info.txt](system-info.txt)\n\nBroken: 2.92\n\nShowing glitches on video sequence timeline when area of Video Sequencer resizing\n\nSteps shown on video\n\n[2020-12-02_15-19-26.mov](2020-12-02_15-19-26.mov)", "Cycles: adaptive sampling improvements\n- [x] Optimize PMJ pattern generation, and generate more samples\n- [x] Solve issues with combined CPU and GPU not using the same amount of samples. Possibly by using a higher adaptive step for both CPU and GPU.\n- [x] Render multiple small times at the same time for GPU, to allow adaptive sampling to stop and match CPU without too much performance impact.\n- [x] Denoising can detect patterns due to adaptive sampling only working per tile. (#79190)\n\n- [ ] Detect errors in larger windows or using some hierarchical methods, as the simple 3x3 window fails for noisy scenes \n- [ ] Analyze different error metrics, and consider adding some user control over this", "Animation workspace: Switching between tabs causes viewport to change\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.59\n\nBroken: version: 3.2.0\nWorked: Not sure\n\nChanging between the Texture Paint tab and Animation tab causes the Animation viewport to take on properties of the Texture Paint viewport. If this is working as expected it's not ideal.\n\nOpen a new General 3D project. (Or use attached blend)\nSelect the Animation tab. The viewport should look normal.\nNow go to the Texture Paint tab.\nGo back to the Animation tab.\nThe viewport has taken on the properties of the Texture Paint tab.\n\n(A similar issue with the UV editor and Animation tab also occurs.)\n\n[TexPaintToAnimationTab.blend](TexPaintToAnimationTab.blend)\n\n\n", "Blender editing performance with many datablocks\n**Status:** 1st milestone almost complete, need final optimizations.\n\n---\n**Team**\n**Commissioner:** @brecht \n**Project leader:** @mont29\n**Project members:** @ZedDB \n**Big picture:** In heavy scenes, new features in Blender 2.8x are making the inherently linear (`O(N)`), when not quadratic (`O(N²)`), complexity of most operations over data-blocks more problematic than in 2.7x era. The general goal is to bring those operations back to a constant (`O(1)`) or logarithmic (`O(log(N))`) complexity.\n\n**Description**\n**Use cases:**\n*Need specific test cases.*\n* Undo changes in object mode in a heavy scene.\n* Undo changes in pose mode in a heavy scene.\n* Duplicating objects in a heavy scene.\n* Adding many objects in a scene.\n\n*Note:* There are two types of “heavyness” in scenes (that can be combined of course):\n * Scenes that have many objects and/or collections.\n * Scenes that have very heavy geometry, either in the geometry data itself, or generated from modifiers like subdivision surface.\n\n**Design:**\n # Dependency graph should not have to re-evaluate objects that did not change.\n ** This is mainly a problem currently during global undo, as all data-blocks are replaced by new ones on each undo step.\n # Handling naming of data-blocks should be `O(log(N))` (it is currently close to `O(N²)`, or `O(N log(N))` in best cases).\n **This will require caching currently used names.*** We will probably also have to handle smartly the numbering in that cache, if we really want to be efficient in typical “worst case” scenarii (addition of thousands of objects with same base name e.g.).\n - Caches/runtime data helpers should be always either:\n - Lazily-built/updated on-demand (code affecting related data should only ensure that the 'dirty' flag is set then).\n *** This approach can usually be easily managed in a mostly lock-less way in a threaded context (proper locking is only needed when actually rebuilding a dirty cache).\n ## Kept valid/in sync all the time (code affecting related data should take care of updating it itself, code using the cache can always assume it is valid and up-to-date).\n ***Such cache should be easy to incrementally update (it should almost never have to be rebuilt from scratch).****This implies that the cache is highly local (changes on a data-block only ever affect that data-block and a well-defined, easy to reach small number of “neighbors”).*** In general keeping this kind of cache valid will always be harder and more error-prone than approach #A.\n *** This approach often needs proper complete locking (mutexes & co) in a threaded context.\n \n *Note: The ViewLayer's collections/objects cache e.g. currently uses the worst mix of #A and #B - it is always assumed valid, but updating it requires a complete rebuild from scratch almost all the time.*\n\n *Note: Approach #B is usually interesting only if a complete rebuild of the cache is **very** costly, and/or if the cache is invalidated very often while being used.*\n\n**Engineer plan:** `-`\n\n**Work plan**\n\n**Milestone 1 - Optimized per-datablock global undo**\nTime estimate: `6 months`\n\n* Re-use as much as possible existing data, and try to swap newly read (modified) data to the original memory, during 'memfile' undo step. Updating the depsgraph should then be much less work.\n** #60695 (Optimized per-datablock global undo)\n** [D6580: WIP/Demonstration patch about undo speedup project.](D6580)\n\n**Milestone 2 - Lazy collection synchronizations**\nTime estimate: `1 month`\n* Fix view layer collection syncing (to not require looping over all objects, or to be ran much less often...).\n** #73411 (Fix ViewLayers cache building)\n\n**Milestone 3 - Data-blocks management performances with many of them**\nTime estimate: `5 months`\n\n# Investigate how to best handle naming conflicts issues.\n**Most obvious idea would be store all used base-names in a GHash (one per ID type), along with used indices. Needs further design though.** #73412 (Improve name conflict handling in ID management)\n# Investigate the best way to cache bi-directional relationships info between data-blocks.\n**We already have a start of this with `BKE_main_relations_create()` & co, but it is heavily under-used and needs better design to be usable more widely.** #73413 (Cache data-block relationships info )\n# Fix/Improve handling of the Bone pointers (which are sub-data of `Armature` ID) stored in poses (which are sub-data of `Object` ID).\n**Ideally such horrible pointers should not exist ever. They are a very solid recurrent source of bugs in Blender.** Not sure how to best deal with them, at the very least we could add some generic ID API to ensure those kind of caches are up-to-date?\n\n\nSee also #68938 (Blender editing performance with many datablocks).\n\n**Later**\n\n* Dependency graph rebuild to be `O(1)` or `O(log(N))` (it is also `O(N)` at the moment). *???? would assume linear would already be nice performance for such a code?*\n\n**Notes:** -\n\n---\n", "Alembic: Camera properties (focal length, focus distance, ...) animation not importing\nMacOS 10.13.3\n\nBroken: 2.79 5bd8ac9\n\n\n\nWhen importing camera in FBX or ABC file (exported from Nuke), the animation of focal length is not importing. It's being set to a constant value according to frame 1. Other values are imported correctly.\nIf I reimport to Nuke, the focal length is correctly animated.\n\nCamera files attached.[camera.fbx](camera.fbx)\n\n[camera.abc](camera.abc)\n", "Regression: Child hair particles set to \"Interpolated\" causes jittering when the based vertex changes position\nOperating system:win10\nGraphics card:RTX3060ti\n\nBroken: c8c6f62cf3a3,fca8df9415\nBroken: 3.5\nWorked: 3.4\n\nCaused by d20f9923224d2637374dd4390ae84c5041e3f9d3\n\nChild hair particles set to \"Interpolated\" causes jittering when the based vertex changes position\nThis problem seems to have been solved before,in #63534 and #56408\n\nOpen the file and render animation\n\n", "Baked cache files should not depend on blend file name\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 535.98\n\nBroken: version: 4.0.0 Alpha\n\nBaking eg. Dynamic Paint will create a folder with .bphys files. When I save the scene under a new name and reload the scene the baked files will be ignored. I need to delete the cache and Bake All Dynamics again.\n\nThe baked result should be independent of the scene filename.\n\nMaking the baked files internally independent of the scene filename is important for our render management with CGRU because CGRU will save a specific filename for the clients. Then the clients can't render the baked result.\n\nWorkaround in my case:\n\n- Submitting job\n- Stopping job\n- Loading scene\n- Delete All Bakes\n- Bake All Dynamics\n- Save scene\n- Restart the job\n\nBake something with Dynamic Paint. Save a new scene, reload and check if the baked result will be rendered.\n\n", "Domain cache (Gas) will not show Paste Driver RMB menu item\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 2.93.4\nWorked: 2.83.17 (`fb67495c7f`)\n\nIf I try to \"Copy as New Driver\" and then RMB in a Gas Domain cache Frame Start or End, I do not get a Paste Driver menu item, unlike other property fields in the Domain.\n\nSelect any object and menu: Object -> Quick Effects -> Quick Smoke.\nHover over Playback, End Frame, RMB -> Copy As New Driver\nIn Domain, expand Cache, hover over Frame End, RMB. No \"Paste Driver\" menu item appears.\n\n", "Performance bottleneck with many nested nodes\nOperating system: win 10\nGraphics card: 1050 ti\n\nBroken: 3.2.0 Alpha\n\nSeems like *Domain Size* performance drops if using nested groups.\n \nIn attached file mute *Float to Integer* node (yellow color) and look how *Domain Size*'s executing time is drastically increasing.\n[domain_size_performance.blend](domain_size_performance.blend)" ]

[ "Mantaflow openvdb cache moving around with adaptive domain enabled\nOperating system: WIN 10 \nGraphics card: 2080 Ti\n\nBlender 2.9.0 Master 01.09.2020\n\nMantaflow openvdb cache has wrong transforms when using adaptive domain\n\nMantaflow OpenVDB Caches produce problems when using the adaptive domain option. When importing the cached VDB into Blender u can see it \"moving\" and \"scaling\" as soon as the original adaptive domain gets bigger and transforms. It appears that the origin of the openvdb is the first vertex of the domain or the small indicator box for domain subdivision resolution. \n\nThe VDB should stay at the same place and don't scale or do other weird stuff. At the moment adaptive domain is not usable at all. Maybe it's a feature but it seems like a bug to me. ", "VDB - adaptive Domain - Reimport - Positionjitter \nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.75\n\nBroken: version: 2.91.0\n\nI was using Mantaflow to make a simple smoke simulation. I baked the simulation using \"modular\".\nAfter reimporting the VDB File sequence the position of the whole simulation is jittery. \n\n\nQuick Smoke/Addaptive Domain/Emitter is Inflow/Bake simulation using modular/Reimport it by shift A \"Import Volume\"/ Replay the VDB File/ The Issue accures/ \n\n", "Mantaflow: OpenVDB glitches due to Adaptive Domain\nOperating system: Windows 10\nGraphics card: Nvidia GeForce GTX 650 Ti\nProcessor: Intel® Core™ i7-3770\n\nBroken: Blender 2.91.0 (b522e834ec30)\nWorked: never\n\nThe Adaptive Domain causes the smoke to shift when imported as a volume object (Import OpenVDB)\n[VDB_Ad_D-vid-C.mp4](VDB_Ad_D-vid-C.mp4)\n\n**Exact steps to reproduce the error (based on the .blend file)**\n[VDB_Adaptive_Domain.blend](VDB_Adaptive_Domain.blend)\n1. Enable Adaptive Domain.\n2. Bake the smoke simulation.\n3. The file uses relative paths, so it should import the cache automatically after the first step, but if it doesn't, select fluid_data_0001 and load the cached files.\n4. Play the animation." ]

[ "Object drivers driven by outliner visibility not updating automatically\nOperating system: Windows 10\nGraphics card: GTX 1060 6gb\n\nBroken: 2.90.0-874af5301cdf (June 20)\n\n\nDrivers on object modifiers that are driven by outliner settings (Hide in viewport, hide in render, etc.) won't work when toggling said outliner settings and has to be forced manually with Update Dependencies.\n\n\n\n\n\n\n - Have 2 cubes in a scene, Cube 1 and Cube 2.\n - Add a modifier on Cube 2 (In my case, a mask, in which Cube 2 is subdivided twice and has a vertex group)\n - Add a driver on the Realtime button on said mask modifier.\n - Have it be driven by Cube 1's \"Disable in Viewports\" button with the following settings:\n\n\nSingle Property\nProperty: Object Cube 1\nPath: hide_viewport\n\n\n\n 6. Disabling Cube_1 from the viewport in the Outliner will not enable the driver on Cube_2's mask modifier.\n 7. Updating Dependencies in the driver editor will correct this.\n\n\n\nAfter updating dependency, enabling Cube_1 will disable the mask in the viewport, although the mask is still enabled in the modifier tab. Clicking once in the viewport will correct this.\n\nExample .Blend File \n[BrokenDriverTest.blend](BrokenDriverTest.blend)", "Pie menu breaks if spawned near window edge\nPie menus are sort of \"clamped\" in that if they are spawned closer to the window edge than their spread radius (UserPreferencesView.pie_menu_radius), instead of spawning partly offscreen, instead they will spawn stuck to the edge and will automatically validate the menu item that's in that direction. To make it clearer, here is a capture.\n\n\n\nThis is a problem because pies can hardly be spawned in cramped workspaces - for instance inside a graph editor occupying the lower third of a standard 16:9 monitor, unless the radius is set super small.\nThis happens on Windows.", "Disabled Selectability: it is still possible to enter other modes (The new Switch Objects operator, py API, ...) - should this be prevented?\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.91.0 Alpha\n\nAfter trying the new Switch Objects operator in Sculpt Mode and Edit Mode, I noticed that the operator completely ignores if an object has Disable Selection active or not in the Outliner.\n[Desktop 2020.09.19 - 22.58.16.01.mp4](Desktop_2020.09.19_-_22.58.16.01.mp4)\n\n1. Have two separate objects in your scene. Go into Sculpt Mode or Edit Mode with one object.\n2. Use Disable Selection in the Outliner on the object you are not selecting in Sculpt Mode/Edit Mode.\n3. Click on the object that uses Disable Selection. You will be able to select it regardless.", "Metaball fixes\n- [ ] Dependency graph integration\n- [ ] Naming based design is weak\n\nThere are many issues with the current version of metaballs. If we won't fix them we may want to consider to deprecate/remove this in future releases.", "Keymap changes are not reflected for the local view setting\nOperating system: Linux-5.18.10-arch1-1-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.57\n\nBroken: version: 3.3.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `rBUnknown`\nWorked:\n\nWhen I turn off frame selected in keymap -> local view and restore it to its default it keeps the old behavior until blender is restarted.\n\nI tested changing delete in object mode (x) key to disable confirm and when I restored it it worked as expected so I think it is related to local view specifically.\n\n- Open default scene\n- In Preferences > Keymap: search for local view and uncheck `Frame selected` checkbox\n- use the shortcut in viewport. Notice frame select operation is not performed (which is expected)\n- In keymap, restore changes\n- Again use shortcut to switch to local view (frame selected operation is expected to execute but it doesn't)\n\n\n", "Apply Parent inverse dose not take scale in to account\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1650/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.61\n\nBroken: version: 3.5.1\n3.6 alpha\n\nApply Parent inverse dose not take scale in to account while using alt+g , alt+r , alt+s to restore the coordinate\n\nnot sure if this is design intended.\n\n**Exact steps to reproduce the error**\n\nBased on the default startup\n1. create 2 cubes in same size ( display in wire frame mode for easier visual)\n2. Scale 2 cube independently in any size\n3. Parent them in either way , using any following parenting method below.\n\nParent object \nParent object Keep transform\nParent object (keep transform without inverse)\n\n*** Parent object (without inverse) works fine as intended there is no need to try on this ***\n\n4. Select child cube Apply Parent inverse\n5. Select both cube alt+g , alt+r , alt+s restore the coordinate , you will see child cube is not at the same size as parent\nbut the transform panel is displaying 1,1,1 both on the child and parent cube\n\n--------------------------------------------------------------\nrelated issue been reported \n97648\n\n", "Apply scale on armature breaks position of child elements with relationship 'Bone'\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.68\n\nBroken: version: 2.93.1\nAlso Broken: blender-2.93.4-candidate+v293.7ec351c0d531-windows.amd64-release\nWorked: Never (2.80+)\n\nIf I have an armature with non (1,1,1) scale and if I create an object that is parented to one of the armature bones, applying the scale to the armature moves the object.\n\nIf I then select my arrow/cube and remove the 'Parent Bone', it is aligned correctly again, if I then put the 'Parent Bone' bone back, the arrow/cube is moved to a completely new position.\n\nI can also see the following error in terminal when I unparent the bone (I don't think this is relevant and clearing the bone always prints this, but just in case)\n\n```\nadd_relation(Bone Parent) - Could not find op_from (ComponentKey(OBArmature, BONE))\nadd_relation(Bone Parent) - Failed, but op_to (ComponentKey(OBEmpty, TRANSFORM)) was ok\nERROR (bke.object): C:\\Users\\blender\\git\\blender-v293\\blender.git\\source\\blender\\blenkernel\\intern\\object.c:3330 ob_parbone: Object Empty with Bone parent: bone doesn't exist\n```\nMy theory is that Blender is trying to reposition the object without taking the bone parent into consideration.\n\nI've recorded a small video to illustrate: watch?v=BGNckTOfBQQ\n\nI was also able to create a script that fixes the position by detaching the parent and reattaching with keep_transform (also note the commented out line)\n\n\n```\nimport bpy\n\nprint ('Bone fixer')\nfor obj in bpy.data.objects:\n if (obj.parent_type == 'BONE' and obj.parent_bone):\n bpy.ops.object.select_all(action='DESELECT')\n \n print (f'Parent={obj.parent} parent_={obj.parent_type} parent_bone={obj.parent_bone}')\n print (obj.name, obj.scale, obj.location, obj.rotation_euler)\n \n # obj.parent_bone = obj.parent_bone # This line should not have any effect but it does move the object to a new position\n\n parent_bone = obj.parent_bone\n obj.parent_bone = '' # We are now correctly positioned but not parented \n\n obj.select_set(state=True)\n obj.parent.select_set(state=True)\n obj.parent.data.bones[parent_bone].select = True\n bpy.ops.object.parent_set(type='BONE', keep_transform=True)\n```\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n[test.blend](test.blend)\n\n1. Create armature ('Armature') and set a non 1,1,1 scale\n2. Add bones to armature\n3. Create a new object (I tried empty/arrow or cube)\n4. Set relationship on the cube as parent = 'Armature', Parent Type = Bone, Parent Bone = 'Bone.X' (last bone in the sample file)\n(^ Already done in sample file)\n5. Select armature and \"Apply scale\"\n\nExpected - the empty remains at the 'end' of the bone\nObserved - the empty is moved to a new location (I am unsure what is the logic)\n\n\n\n", "Regression: unhide now is global even in local mode\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.0.0 Alpha\nWorked: 2.79b\n\nAlt+H in local view unhides objects all over the scene instead only local context, as it work in 2.79\nPlease fix. It is very handy feature and we have lost it.\n\n**Step to Reproduce**\n- Open .blend file\n- Select a sphere and Hide: {key H}\n- Select all three cubes and enter in Local view: {key /}\n- Hide cubes\n- Now unhide cubes in Local view: {key Alt H}\n- sphere is also visible (Either check outliner or Switch back from Local view)\n\nTest File:\n[#92527.blend](T92527.blend)\n[2021-10-27_12-23-23.mp4](2021-10-27_12-23-23.mp4)", "Linked Text Object Data Bug\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Super with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.42\n\nBroken: version: 2.93, 3.0.0\nWorked: -\n\nText objects with linked object data behave strangely. Some letters change their shape, which may vary from instance to instance. (see uploaded image)\nMaybe i´ve got the bug is because i use the Small Caps format.\n\n\n1. Create several same text Objects (you maybe should use Small Caps format in them)\n2. Link their Object Data\n3. Try to switch several times between Object and Edit mode and you will see the bug. It occurs everytime on different letter. And every instance of the text looks often different then the others.\nI am attaching the Blend file\n[bug.blend](bug.blend)\n", "No snapping for objects referenced in modifiers (even if disabled)\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.67\n\nBroken: version: 2.91.0 Alpha\nWorked: probably never; hopefully it's just an oversight. Even 2.79b is broken\n\nI've experienced this a handful of times but finally sat down to investigate from the user perspective here today (sorry, haven't looked at the code). But apparently the object you use as Simple Deform's \"origin\" will not be snap-able to the object that has Simple Deform on it. It doesn't matter if the modifier is enabled or disabled or if the settings actually produces no deformation, it always breaks snapping for the origin object.\n\n[snap-issue.blend](snap-issue.blend)\n\n- Open the attached .blend file OR follow next instructions to setup\n - Start with default scene and add an Empty to the scene\n - Add a Simple Deform modifier to the Cube\n - Set the Simple Deform's \"origin\" object to the Empty, the mode to \"Bend\" and the \"Angle\" to 0deg (to prevent any deformation)\n - Set the scene snapping mode to vertex\n\n- Grab the Empty and attempt to snap it to verts of the Cube -- it cannot snap (bug)\n\n- Disable the modifier and repeat the snap attempt -- it still cannot snap to the Cube (bug)\n\n- Create a new object and attempt to snap the empty to that new object -- it succeeds (the Sphere in the attached .blend file)\n\n- Delete the modifier and repeat the snap -- finally, it is able to snap to the Cube\n\n", "Knife Tool produces bad geometry in certain cases (disconnected edges/vertices, overlapping faces)\nif i use tools knife i see this bug.\n\n\n\n", "Geometry Nodes: Evaluate node tree for get dependencies\n/ in a progress /", "Changing Resolution Preset doesn't allow to Undo\nBroken: 3.6.0\n\nChanging Resolution Preset and then trying to undo doesn't work, like if the action is not registered.\n\n\n- Open Blender\n- Click on the Format panel on the right icon to change preset\n- Select a preset\n- Undo\n\n", "Impossible to move with G key an object scale to 0 and parent to an empty\nOperating system: Windows 7\nGraphics card: GTX 780\n\nBroken: 3.5.1\nWorked: \n\nWhen an object is parented to an empty, and this object is scale to 0, you cannot move it anymore with the G key. But you can move it in the transform menu location.\n\nBased on the default startup or an attached .blend file (as simple as possible).\n1. Open blend.\n2. Try to move object with G key.\n3. It is possible with Transfom menu.\n\nThanks a lot !\ndono", "Outliner: Linking object hierachies to a collection is buggy\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Radeon RX Vega ATI Technologies Inc. 4.5.14830 Core Profile Context 21.4.1 27.20.21002.112\n\nBroken: version: 2.93.2 Release Candidate\nWorked: Unknown\nLinking objects with parent child relationship has inconsistent display behaviour in outliner\n\nThis bugreport is about linking an object in the outliner to an additional collection. \n\nA linked object can have children. Normally they are displayed in a greyed out state until they are also linked to the same collection as the parent. But there is a bug or at least an inconsistent behaviour if these childs are displayed in a greyed out state or not at all.\n\n\nThe following script adds 3 cubes ,creates a collection and links the cubes to it. \nI deleted everything from the Scene Collection and did run the script. \n\nThis will work: \n\nBut just one more child-level in the tree and it starts to behave inconsistent as can be seen in the outliner shown in the screenshot of some reruns of the same script on a totally empty scene. \n\n\n\n\n\n\n\n\n\n```\n\nimport bpy\n\n# create some cubes\nbpy.ops.mesh.primitive_cube_add(location=(0,0,0))\ncube0= bpy.context.active_object\nbpy.ops.mesh.primitive_cube_add(location=(5,0,0))\ncube1= bpy.context.active_object\nbpy.ops.mesh.primitive_cube_add(location=(10,0,0))\ncube2= bpy.context.active_object\nbpy.ops.mesh.primitive_cube_add(location=(10,0,0))\ncube3= bpy.context.active_object\n\n#create a new coll\ncollection = bpy.data.collections.new(\"Test\")\nbpy.context.scene.collection.children.link(collection)\n\n# establish parent child relationships\n\ncube1.parent = cube0\ncube2.parent = cube1\ncube3.parent = cube2\n\n# link root cube to new coll\ncollection.objects.link(cube0)\n\n\n```\n\n\n" ]

[ "linked objects un-parent after save and load.\nwindows 8.1, Geforce 660\n\nBroken: #5ebae1c\nWorked: i couldn't test as how far this existed as old versions won't compile against the current SVN.\n\nlinked objects un-parent after save and load.\n\ni included 3 files:\nfile 1: the parented objects.\nfile 2: the objects linked.\nfile 3: the parent object changed to local, everything is fine till it is saved it will un-parent!\n\n[1.blend](1.blend)\n\n[2.blend](2.blend)\n\n[3.blend](3.blend)" ]

[ "modifier offset slider step value being driven by the mouse looses one order of magnitude worth of accuracy when units are metric vs none\nwin 10\n2080ti\n\nBroken: 2.80+\n\nSince blender 2.8 has switched to real world units the modifier slider step(when driven by the mouse) value has lost one order of magnitude worth of accuracy, 0.001 instead of 0.0001 \nIt is possible to switch the unit system back to none to get the higher accuracy. \nThey should be consistent and both work to the higher level of accuracy as this level is essential when working on real work scale small items e.g. a watch\n\n\n- with scene unit type set to \"metric\" add a bevel modifier the standard cube\n- press shift while clicking and dragging the bevel modifier offset slider and notice how the smallest amount possible to move is 0.001 m\n- switch scene unit type to \"none\"\n- press shift and click drag the modifier offset slider and notice the smallest amount possible to move is 0.0001 m\n", "MantaFlow Smoke simulation \"breaks\" after 110 frames\nOperating system: Windows 10 \nGraphics card: NVidia RTX 2080 Ti\n\nBroken: (example: 2.83 Beta, d8133b3eb99a, 2020-05-04)\nWorked: Not that I know of\n\nWhen baking the Smoke Simulation of the attached File, the smoke 'explodes' at frame 110. It looks as if an 'Explosion Force Field' was placed in the middle of the domain that applies its forces at frame 110.\n\nTo Reproduce: Download attached file, Bake the simutaltion and look at frame 110+.\n\nBlendFile: [Blender_ [D__03_personal_MartinFroehlich_florianDubois_Laminar_7.blend] 2020-05-06 17-17-04.mp4](Blender___D__03_personal_MartinFroehlich_florianDubois_Laminar_7.blend__2020-05-06_17-17-04.mp4)\nVideo: [Laminar_7.zip](Laminar_7.zip)", "Mantaflow liquid collisions not working properly\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.90.1\n\nUnless a liquid collision effector's Surface Thickness is set to at least 0.5 or set to \"Is Planar\" it is not used.\n\nI reference two examples for liquid and smoke+fire [examples ]] and there are several issues logged in [[ https:*blender.stackexchange.com/search?q=liquid+collision | Blender stackexchange ]], along with my [[ 111042 | investigation ](http:*optionpeer.com/jimtpersonal/collisions.zip).\n\nA collision effector is ignored unless Surface Thickness is set to at least 0.5 or set to \"Is Planar\". However as [docs ](effector.html) state Surface Thickness is \"Additional area around the effector\".\n\nIn my first example \"liquid collision.blend\" the results (show in \"liquid collision.png\") shows how the extra Surface Thickness effects the liquid that becomes more confined in the container, whereas setting it to less than 0.5 will cause the liquid to fall through the container.\n\nThe same is true in the \"muzzle*.blend\" file where the flash & smoke comes through the collision object unless Surface Thickness is set (incorrectly) to at least 0.5\n\n[liquid collision.blend](liquid_collision.blend)\n[muzzle flash no projectile.blend](muzzle_flash_no_projectile.blend)\n", "Mantaflow - Bake not showing (when noise is used)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 26.20.100.7262\n\nBroken: version: 2.90.0 Alpha\nWorked: 2.83\n\nEnabling Noise and then Baking Noise, the preview no longer shows. \n\nFrom Attached Blend\n\n\n - Bake Data\n - Play Animation (see preview)\n - Enable Noise\n - Bake Noise\n - Play Animation (No preview is shown)\n - Disable Noise \n - Play Animation (see preview)\n\n [smoke_noise.blend](smoke_noise.blend)\n", "Rigid body: Animated body collision geometry have unapplied scale for the first frame.\nOperating system: macOS-12.6.5-x86_64-i386-64bit 64 Bits\nGraphics card: Intel(R) Iris(TM) Plus Graphics 645 Intel Inc. 4.1 INTEL-18.8.6\n\nBroken: version: 3.5.1\n\n\nIf an rigid body object has a scale that's not 1 , when physics simulation begins, the collision on the first frame will be calculated as if the object (collision geometry) is scaled back to 1, but subsequent frames will have correct scale.\n\n-----\n\nI have been following this tutorial\n\nwatch?v=2Dnks7C2q9U\n\nand as soon as I set the 'handle' object 'Animated' and press 'Play' on timeline, the bullet physics immediately kicks-in,\n\n\n\nThis is the file I'm using:\n\n[20230618_rigid_body_tipping_over_stacks_of_cubes_animated_wrong_activation_0001.blend](attachment)\n\n", "Multiple Smoke Flow Colors Rendering Black when Flow Behavior is set to Geometry\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 670/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.48\n\nBroken: version: 2.83.1\nWorked: 2.81 -- N/A, newer versions use fluid settings, but the old smoke sim worked as it should\n\nBaking a smoke domain multiple smoke flow objects with flow set to \"geometry\" and each set to a color results in all of the smoke rendering black.\n\nCreate a smoke domain on any object and place two smoke flow objects inside the domain. Set the smoke color of the smoke flow objects to any color, and the flow behavior to \"geometry\" vs \"inflow\" or \"outflow\". Change the material on the domain object to a material without a surface bsdf but with a volume bsdf whose color is set to the color output of an attribute node with the Name property set to \"color\". Bake the sim and render, and you should observe that all smoke is black, as opposed to the colors set on the flow objects. This works correctly when the flow behavior is set to inflow (except on the first frame), so I'm guessing this has something to do with the initialization of the smoke.\n\n[smoke_multi_color_bug.blend](smoke_multi_color_bug.blend)", "Adjust last operation: can't reset properties to default value\nBroken: 2.8x\n\nIn the Adjust Last Operation panel, the context menu for any property has the Reset to Default Value options disabled, although they are enabled in the operator panel.\n\n\n1. Load the default scene\n2. Enter edit mode on the cube\n3. Bevel an edge\n4. Expand the operator panel in the bottom left\n5. Right click any property, note that Reset to Default Value option is enabled and works\n6. Go Edit -> Adjust Last Operation\n7. Right click the same property in that panel, note that Reset to Default Value option is disabled", "Hair particles rendered as an object do not follow parent object when the parent object is falling under gravity as a rigid body when animating\nOperating system: Microsoft Windows 10 Home, Version 10.0.18362 Build 18362\nGraphics card: GeForce GTX 1080\n\nversion: 2.82 (sub 7), branch: master, commit date: 2020-03-12 05:06, hash: 375c7dc4caf4, type: Release\nbuild date: 2020-03-12, 15:41:08\nplatform: Windows\n\nI have an issue where i have a cube mesh, I create hair particles, and set the hair to render as an object, say a cylinder. I then enable rigid body physics for the cube. I also have a passive plane below the cube. When I run an animation, the cube falls as if under gravity and stops at the plane, but the cylinder particles remain in their original position?\n\nIf I add soft body physics to the cube and order the soft body physics before the particle settings, then the cylinders fall, but lag behind the cube. The cube falls through the plane below, but the cylinders are stopped by the plane?\n\nWith soft body added (to top of stack). Particles do fall but have a lower acceleration rate, and the cube no longer detects the floor and passes through the plane?!\n\nThanks!!\n\nSee attached blend file [StackExchangeParticleProblem-001.blend](StackExchangeParticleProblem-001.blend)", "Lattice UVW input Glitch\nOperating system: Darwin-18.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro Vega 64 OpenGL Engine ATI Technologies Inc. 4.1 ATI-2.11.21\n\nBroken: version: 2.81 (sub 16)\nWorked: (optional)\n\nClick dragging vertically through the UVW parameters of a Lattice will let you select all 3 parameters (as expected).\nBut, entering a new number will only change the 1st of the 3 selected parameters.\nWeirdly, click drag selecting all three parameters and entering nothing-at-all and then clicking away will somehow enter in the correct number that was entered one attempt back in time.\n\nStart a new General file and delete all objects.\nAdd a Lattice.\nGo to the Lattice Properties.\nClick and drag down through the U, V and W fields.\nEnter a new number.\nNotice that only 1 of the parameters changed (when all 3 should have changed).\nClick and drag down through the U, V and W fields again.\nClick away.\nNotice how the other 2 parameters \"catch up\" to the first 1.\n\n", "Menus open on upward mouse movement\nOperating system: Linux (Debian bullseye)\nGraphics card: Nvidia GTX 1070\n\nBroken: current master 2.90.0 (fc672ce8e24e)\nWorked: Never\n\nWhen there are a lot of modifiers in the stack which make the evaluation slow there is a weird effect with the dropdown menu. It somehow glides when dragging to the next modifier.\n\n\n\nTo the default cube, add array modifiers until your memory usage is at roughly 300MiB. Then drag over the drop down arrows like shown in the gif.", "Smoke does not leave through open domain boundary\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 451.48\n\nBroken: version: 2.90.1\nunknown\n\nSmoke does not leave through open domain boundary\n\n\n\n1. start blender\n2. add fluid domain to default cube\n3. add border collision to bottom\n4. Change cache type to Modular\n5. Change gas bouyancy density to 0 and Heat to 0\n6. Add 0.5m mesh plane\n7. add fluid->inflow to mesh plane\n8. set inflow initial temperature to 0\n9. enable initial velocity to mesh plane\n10. set initial z-velocity to -1m/s\n11. bake fluid domain\n12. See how smoke is crawling up along side walls despite that border collision is not enable for these walls. I see this as a bug.\n\n[#81442.blend](T81442.blend)\n \n", "Mantaflow: Noise doesn't follow smoke\nOperating system: Windows 10\nGraphics card: Nvidia GeForce GTX 650 Ti\nProcessor: Intel® Core™ i7-3770\n\nBroken: Blender 2.90\nWorked: Blender 2.81\n\nAfter enabling \"Noise\", the Wavelet turbulence doesn't follow the smoke. It seems to only mildly adapt to certain movements; but, unlike in 2.81 it doesn't get \"dragged\" by the velocity of the smoke.\n[NoiseDifference-C.mp4](NoiseDifference-C.mp4)\n\n**Exact steps to reproduce the error**\nIn any smoke simulation (preferably one where smoke is being constantly emitted), enable \"Noise\" and bake.\n\n2.9 Project file: [2.9 Example.blend](2.9_Example.blend)\n\n2.81 Project file for comparison: [2.81 Example.blend](2.81_Example.blend)", "Animated particle emitter refuses transforms within simulation range\nOperating system: Win8.1x64\nGraphics card: 3xgtx580\n\nBroken: 2.79, 2.82\n\nAn animated object with a particle system it is not properly transformed in the frames within the simulation range.\nIn **Blender 2.79** the object jumps back to it's animated location on confirmation on hotkey or widget transformation.\nIn **Blender 2.82** the object doesn't even transform, but its particles does.\n\nTemporarily setting particle system Subframes parameter to 0 seems to be a workaround, eliminating the behavior.\n\nHere is a gif showing the bug in 2.82 and 2.79:\n\n\n- Open attached file\n- (G)grab move the emitter\n[particle-subframes.blend](particle-subframes.blend)\n\nIn blender 2.82 you must change the visibility of objects. If you want to skip this step, use this file:\n[particle-subframes-283b.blend](particle-subframes-283b.blend)", "Fluid inside Effector / Fluid Inside Closed Object not Working\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060 Laptop GPU/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 512.96\n\nBroken: version: 3.2.0\nWorked: 2.83 (#104938)\n\nFluid simulation stops flowing while inside effector whether the effector its one mesh or separate mesh.\nThis also works, when I created a holes in the mesh, but I really need to close it completely for my actual project.\n\n\n[Fluid Inside Effector.mp4](Fluid_Inside_Effector.mp4)\n\n\n\n[Fluid Inside Effector.blend](Fluid_Inside_Effector.blend)\n", "Shader node float precision issue cause different results on CPU and GPU\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 1070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.86\nCPU: Intel(R) Core(TM) i7-8750H CPU @ 2.20GHz\n\nBroken: version: 2.80rc2 (2.80 (sub 74))\n\nWhen rendering a mandelbulb volumetric thing on cycles using CPU and GPU, the two have very different ideas on how the scene should look like, making the individual tiles clearly visible. \n\n**Render**\n \n(most tiles are GPU-rendered, note the stray red CPU-rendered tile to the left)\n\n**Blend**\n[mandelbrot.blend](mandelbrot.blend)\n\n**Potentially useful information**\n* #50193 may be the same bug; I opened a new one because that bug doesn't report differences within a single frame, and is from 2016.\n* Since most of this object is endless math nodes, it may be some sort of floating-point thing that works differently on CPUs and GPUs. \n* I was following this tutorial: watch?v=WSQFt1Nruns\n* I noticed that some operations for which I expected single inputs (namely the sines) had two inputs. In the tutorial, the dude said the second input doesn't do anything, but then, that's from the alpha. Maybe taking a 0.5th sine of a value does something strange?" ]

[ "[Mantaflow] initial velocity ignored in the emission object\nOperating system: Windows-10-10.0.17134-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 431.86\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional)\n\ninitial velocity ignored in the mantraflow emission object\n\n- Open file and run simulation.\n[liquid.blend](liquid.blend)\n(In the file the initial velocity is pointed to normal).\n\n" ]

[ "Subdivision surface settings part of the Mesh\n**Description**\n\nWe want to make subdivision surfaces a property of the Mesh datablock. The modifier would remain, but would likely be used less.\n\nThere are a few reasons for this:\n* OpenSubdiv GPU acceleration in the viewport and adaptive subdivision in Cycles need the modifier to be last in the stack. The available settings, behavior and implementation would be more clear if this was controlled outside of the modifier stack.\n* When we add modifier nodes support, the concept of a last modifier in the stack becomes fuzzy. Being able to pass along a subdivision surface Mesh to another node without actually subdividing is important for performance and to leave subdivision to the renderer.\n* For best performance with GPU acceleration, we don't want Blender to also subdivide the mesh on the CPU at all. But operations like hair positioning, particle emission and snapping still need to take into account the subdivision surface. If the Mesh datablock contains all the subdivision surface information, these could potentially work on the limit surface without subdividing the mesh on the CPU.\n* Simpler smooth normals handling, see #68893 (Smooth shading usabilty).\n* File formats like USD have native subdivision surface primitives. By making it part of the mesh, we are more compatible with such file formats. \n\n\n**TODOs**\n\n- Modeling\n - [ ] conjugate the presence of a subdivision surface modifier with the mesh settings if it is the last in the stack\n - [ ] ensure creases survive the modifier stack (unless consumed by a subsurf modifier)\n - [x] (follow up to [D10145: Subdivision: add support for vertex creasing](D10145)) move edge crease to a custom data layer\n\n- Cycles/EEVEE\n - [x] detect subdivision using the mesh settings instead of the modifier\n\n- Normals (work on some aspects of #68893 (Smooth shading usabilty))\n - [ ] remove autosmooth and custom normals options if the mesh is a subdivision surface\n - [ ] always render with smooth normals (flat areas can also be defined trough creasing)\n\n- Alembic/USD\n - [ ] set mesh settings during imports\n - [ ] export as SubD Object when subdivision is used\n\n- Workflow\n - [ ] option to set the subdivision settings when creating a mesh, or add a sub-menu with meshes with subdivision enabled\n - [ ] shortcuts to subdivide (?)\n - [ ] script/addon to update old files to replace last subdivision modifier in the stack with mesh settings (doing this in version code might have undesirable side effects)\n\n- Other\n - [ ] Update test files", "Modifiers header misalignment on change Resolution Scale [pixel perfect thing]\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Intel(R) HD Graphics 4000 Intel 4.0.0 - Build 10.18.10.5161\n\nBroken: version: 2.92.0; blender-2.90.0-d7c4e9649320\nWorked: don't know\n\nModifiers/Constraints header buttons misalignment on different (especially smaller) Resolution Scale value\n\nSimple as that. Tried set Resolution Scale down on a notebook, and saw the modifier name input \"jump\" out of the modifier box.\nSo checked that any value bellow 1, varnishes the \"top margin\" of it. The same on constraints, bones included.\nThan i suppose it has something to do with the dragging feature.\n\nOpen Blender.\nAdd any modifier/constraint an object.\nGo to Preferences -> Interface -> Display -> Resolution scale, and click the input's left arrow.\nSe the bug on the modifier header\n\n\n\n", "bmesh.ops.subdivide_edges doesn't report all inner edges in 'geom_inner'\nBroken: 2.91, 2.92 master\nWorked: never?\n\n\nI'm not sure if it's a bug or if there's some underlying reason for this behavior. The [BMesh subdivide edges op ](bmesh.ops.html#bmesh.ops.subdivide_edges) returns three lists of geometry, one of which is 'geom_inner'. It appears that this list doesn't always contain all new 'inner' edges. This depends on the topology being passed to the op. Example image, created from the attached blend file:\n\n\n\nOn the left is the original selection of edges being passed in to the op, and on the right - selection of edges from the resulting 'geom_inner' list. As you can see, some of the edges aren't selected (i.e. they were not present in the list). We can work around this by iterating all **vertices** from 'geom_inner' list, and taking all edges linked to them that are not also part of the 'geom_split' list returned by the op, but it feels like this information should just be provided in the 'geom_inner' list.\n\nLoad the attached file and run the script.\n\n[subdivide_geom_inner_issue.blend](subdivide_geom_inner_issue.blend)", "Measures are referred to as annotation layers but are not deleted when removing the layer.\nBroken: version: 2.83 (sub 0)\n\nUnable to delete measure through annotation layers, only after getting into edit mode on an object the measure annotations dissapears.\n\n- In the default blender scene select the `Measure` tool and create a measure in 3dview.\n- Go to Sidebar(N) -> View -> Annotations and remove `RulerData3D`.\n- Note that the ruler remains in 3dview.\n", "Mesh tool to 'unsplit' while preserving custom normals\nSome exporter generate split geometry to preserve 'sharp edges' (like Blender used to do not so long ago). However, now that we support custom normals, it would be nice to have e.g. an object-mode 'unsplit' modifier to remove doubles while preserving custom normals.\n\nSee #47540 for a usecase.", "Remove shapekey as a datablock\nShape keys are currently stored in a `Key` datablock, separate from the `Mesh`, `Lattice` and `Curve` datablocks.\n\nThis is invisible to the user, and each geometry datablock can be associated with exactly one `Key` datablock.\n\nThis makes datablock management code unnecessarily complicated. To simplify the internal implementation, we should embed the shape key data in the geometry datablock. Potentially shape keys could be stored as customdata layers.", "Geometry Nodes: Mesh Boolean destroys attributes\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\n\nUsing Mesh Boolean with Geometry Nodes destroys attributes from input meshes. This drastically reduces usability and node network clarity as user is required to surround the Mesh Boolean node with 4 more nodes for every single attribute they want to survive the boolean operation:\n\n\n1. Open the attached file: [BoolDeleteAttributeBug.zip](attachment)\n2. Hold Ctrl+Shift and left click the Geometry output sockets of Group Input and Object Info nodes. Notice the attribute being present.\n3. Hold Ctrl+Shift and left click the Mesh output socket of Mesh Boolean node. Notice the attribute destroyed.\nResult: The attributes are destroyed by Mesh Boolean node.\nExpected: The attributes on the parts of the topology unaffected by Mesh Boolean operation remain preserved.\n\n", "Unlinked/deleted objects reappear, when deleted at startup in a python script (i.e. ./blender --python delete.py)\nDebian 7.4 64bit, also occurs on OSX\n\nBroken: 2.63, 2.69.0\n\n\nUnder certain conditions objects which have been unlinked from a scene do reappear.\n\n\n\n\n - Start blender with `./blender --python delete.py` (see below for the script) (blender just shows an empty scene)\n - Once the GUI is loaded, do: **Add -> Mesh -> Plane**\n # Immediately afterwards hit **F6** and adjust one Property of the newly added object (e.g. the radius)\n\nResult: the objects from the start scene (camera, lamp, cube) reappear.\n\n**Doesn't occur when...**\n - ... A second Mesh is inserted before hitting **F6** to change its properties\n - ... The script is executed in the Blender Editor (and not at startup with option `--python`)\n\n\n**Python script `delete.py`**\n\t\n\n\n\n```\n import bpy\n \n```\n - IMPORTANT: run as: ./blender --python delete.py\n - once blender started, here is what I did in Blender GUI:\n - - add a new mesh (Add -> Mesh -> Plane)\n - - Hit F6 immediately afterwards and modify radius\n # Result: Blender start scene (cube, light & camera) is visible again\n```\n \n```\n # the described bahaviour and result is the same for deleteScene1() - deleteScene3()\n```\n \n def deleteScene1():\n bpy.ops.object.select_all(action='SELECT')\n bpy.ops.object.delete()\n \n def deleteScene2():\n bpy.ops.object.select_by_type(type='MESH')\n bpy.ops.object.delete()\n bpy.ops.object.select_by_type(type='LAMP')\n bpy.ops.object.delete()\n bpy.ops.object.select_by_type(type='CAMERA')\n bpy.ops.object.delete()\n \n def deleteScene3():\n scn = bpy.context.scene\n for obj in scn.objects:\n scn.objects.unlink(obj)\n \n```\n - deleteScene1()\n - deleteScene2()\n```\n deleteScene3()\n\n```\n\n\n\n", "Meshes with the remesh modifier disappear when they are edited at the same time\n**System Information** :\nOperating system :Windows10, CPU:3950X\nDisplayCard: RTX4000\n\n **Blender version**: \nBlender:2.92\n\nWhen editing multiple objects in edit mode, if you edit more than one at the same time, some modifiers (such as remesh) cause the meshes to disappear.\n[2021-04-13 07-02-23.mp4](2021-04-13_07-02-23.mp4)\n\n- Open attached file\n- In edit mode select the face of an object\n- Scale (modifier updates correctly)\n- Deselect all and select the face of the other object\n- Scale (modifier updates correctly)\n- Select both faces\n- Scale (Meshes disappear)\n\n[UNTD_SIMPLIFIED.blend](UNTD_SIMPLIFIED.blend)\n\n---\n**Original description:**\n\nI added a lot of modifiers to these two objects.\n\nSelect one. Edit mode scale is fine. \n\nSelect two edit modes scale will disappear\n\nTest file : \n[UNTD.blend](UNTD.blend)", "Removal of an attribute with same name as Vertex Group removes the Vertex Group instead of the Attribute\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 517.40\n\nBroken: version: 3.4.0 (Also 3.5 Alpha hash: `50c7eb14f455`)\nWorked: Never, 3.3 both the attribute and vertex group are removed\n\nWhen removing an attribute that has the same name as an existing vertex group using either the UI button or the Python API the vertex group is removed instead of the attribute\n\n1. Open Blender\n2. Add an attribute named 'Group' to the default cube\n3. Add a vertex group (automatically name 'Group')\n4. Remove the attribute named 'Group' using either the - button or the following API command `C.object.data.attributes.remove(C.object.data.attributes['Group'])`\n5. The vertex group is removed instead of the attribute\n\nIn 3.3 LTS both the attribute and vertex group are removed when trying to remove the attribute.\n\n[removal_of_attributes_vg_bug.webm](removal_of_attributes_vg_bug.webm)", "Unexpected behavior when applying Geometry Nodes modifier on top of the Multires.\nThis is a more comprehensive report on a bug that was recently posted and closed [HERE ](T97603)\n\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.84\n\nBroken: version: 3.0.1+\n\nApplying Geometry Nodes on top of the Multires modifier does not give expected results. Multires is an exception from the \"Applying a modifier will apply it as if it was first in the stack\" rule, when a modifier is applied on top of the Multires, the mesh changes will be applied to the currently visible subdivision level. Quick video below.\n[Multires GNodes Bug.mp4](Multires_GNodes_Bug.mp4)\nAnd the Example .blend file > [Multires GN Bug.blend](Multires_GN_Bug.blend)\n\nHow should applying the Geometry Nodes behave ? When the Geometry Nodes modifier doesn't change the vertex count and simply deforms the mesh, it should behave just like any other modifier, see example video with the Shrinkwrap modifier. \nSo what when Geometry Nodes does change the vertex count ? I say we keep the same behavior that solidify modifier has, it skips Multires and applies it as it would be first in the stack. While of course this will bring broken results, this is a already an existing behavior so i think it would be completely fine.\n\nSeries of images presenting how applying a Solidify modifier on top of Multires behaves. (Intended behavior)\n\n\n\n\n\n\n\n", "Data block name not unique when use gpu subdivision option turned on\nOperating system: Windows 11\nGraphics card: NVIDIA Quadro P6000\n\nBroken: 3.1 all the way to 3.6\nWorked: Anything before 3.1\n\nWhen the option Viewport->Subdivision->GPU Subdivision is turned on in the preferences, the data block names for meshes that have the subdivision modifier attached are the same.\n\n1. Start with the default scene.\n2. Create a second cube.\n3. Attach a subdivision modifier to each cube.\n4. Run the following script.\n\n```Py\nimport bpy\n\ndepsgraph = bpy.context.evaluated_depsgraph_get()\n\nfor i, ob_inst in enumerate(depsgraph.object_instances):\n \n if ob_inst.object.type == \"MESH\":\n print(\"%s (%s)\" % (ob_inst.object.data.name, ob_inst.object.name))\n```\nThe output you should see is something like this:\n\n```\nMesh (Cube)\nMesh (Cube.001) \n```\nIf you run the script again with the option turned off, you should see this:\n\n```\nCube (Cube)\nCube.001 (Cube.001)\n```\n\nI'm not sure if this is expected, but it's inconsistent behavior. As far as I can tell, it only happens on Windows and Linux. It doesn't happen on my Intel Mac running running macOS Ventura (13.4.1).", "Remove unused slots fails with non-mesh objects\nOperating system: all\nGraphics card: all\n\nBroken: 2.83 2.93 3.0daily \n\nwhen object type is Volume or metaball click remove unused slots button.\nWill delete all material slot whether the material been used.\n\nbut if object type is mesh, this action will keep the used \n\n- Open .blend file\n- Material properties tab -> materials specials -> remove unused slots button\n\nTest File:\n[#90338.blend](T90338.blend)", "Subdivide with smoothness not symmetrical (not deterministic, depends of order of vertices in face)\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GT 430/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 388.13\n\nBroken: version: 2.83.4\nWorked: Never (2.8+)\n\nMesh is no longer centered along axis after using Subdivide with Smoothness greater zero. Projection of mesh is not exactly rectangular when looking form top.\nThe position of vertices, created with Subdivide with applied Smoothness, is wrong/arbitrary only if the vertex is placed inside the original face/polygon not along the edges of it. (Quads have one vertex inside face after one subdivision, Triangles need atleast two subdivisions for the error to appear because first subdivision creates new vertices only on edges)\n\nFirst example:\n1. Open subdivide_smoothness_plane.blend\n2. Apply Subdivide with Smoothness to the mesh\n3. Mesh is no longer centered along y-Axis. Compare position of vertices created inside the two quads.\n[subdivide_smoothness_plane.blend](subdivide_smoothness_plane.blend)\n\nSecond example:\n1. Open subdivide_smoothness_multiple.blend\n2. Apply Subdivide with Smoothness to the mesh with MORE then one Number of Cuts\n3. Mesh is no longer centered along y-Axis\n[subdivide_smoothness_multiple.blend](subdivide_smoothness_multiple.blend)\n", "Pinning data (mesh, curve, grease pencil, lattice, particle system) makes certain tabs/options disappear (vertex groups, shapekeys facemaps)\nOperating system: Microsoft 10\nGraphics card:GeForce GTX 10606GB\n\nBroken: 2.93\nWorked: don't know\n\n**in the properties editor, in object data properties, the pin toggle hides vertex groups and shapekeys**\n\n- select default cube\n- go to properties editor\n- press object data properties tab\n- press the pin button\n" ]

[ "Regression: Multires Unsubdivide destroys the vertex groups\nOperating system: Windows 10\n\n\nBroken: 3.0 release, 3.0 (b42354672e)\nWorked: 2.93LTS, 3.0 (cec103d5a9) \n\n\nIt seems like unsubdividing using the Multires Modifier destroys the vertex groups, interesthingly enough other kind of mesh data stays intact (like the vertex colors) after the same operation.\n\nThe Decimate modifier's Unsubdivide does not destroy the data, but using the Decimate modifiers not option in these kinds of usage because the multires one is about getting the prior subd levels for sculpting purposes.\n\n[byUQjUlY19.mp4](byUQjUlY19.mp4)\n\n\nLoad the .blend file\nSelect the torus\nMultires->Unsibdivide\nMove your mouse to the 3d viewport to get the UI update properly\nThe Vgroups are gone\n\n[multires_09122021_1051_52.blend](multires_09122021_1051_52.blend)\n" ]

[ "Blender 2.71 (rc2) crashes when a UV sphere is added\nMac OS X 10.10 (beta) / AMD Radeon HD 6750M 512 MB\n\nBroken: 2.71 (release candidate 2)\nWorked: -\n\nBlender 2.71 (rc2) crashes when a UV Sphere is added to a existent project in Edit Mode.\n\nUsing the attached file, select the figure (stickman), go to Edit Mode and add a UV Sphere. After executing these steps, Blender stops responding and it's necessary to force quit and reopen. \n\nIt seems that's something to do with the modifiers (in the attached file, I'm using Mirror, Skin and Sub-surf). One or more of them are causing the crash when the UV Sphere is added. If I create a new project with no modifiers, I'm able to add UV Sphere. \n", "Changing outputs from Separate RGB node into Displacement (using microdisplacement) does not trigger geometry update.\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.83.0\n\nChanging outputs from Separate RGB node into Displacement (using microdisplacement) does not trigger geometry update. Have to tab into edit mode to force mesh update.\n[rgb_disp_bug.blend](rgb_disp_bug.blend)\n\nAdd plane with microdisplacement enabled.\nConnect Image node to Separate RGB node and one of the channels into Displacement node.\nChange view to Rendered mode.\nChange channels from R to G to B to R and notice no geometry updates happen in the view port. (Make sure to use an image with different channels)", "Multi-Object Properties Editing\nIn Blender, we want to make it easier to set values for multiple objects, bones, strips, keys and other items at once. \n\nCurrently you can do this in limited circumstances, but you have to right-click on a value after you've changed it and pick Copy to Selected from the menu or hold Alt while editing values. This not only slow, but also very hidden, and the Alt-key conflicts with Emulate 3 Button Mouse. Lastly, it isn't clear at all to users which values work for multi-editing, and which values don't.\n\nHere's how we would like to solve it:\n\n\n## Selection vs Active\n\nThe concept of ***active*** in Blender means that we only show properties and only affect the active object or item. However, if we want to better support multi-item editing, we must start by making it so the Properties reflects the ***selection*** rather than the ***active*** item, of which there can only be one. Going forward, by default the Properties will be set to reflect the selection instead. This could be controlled in the Properties like so:\n\n{[F8363009](Screenshot_2020-02-22_at_17.34.54.png), size=full}\n*(The search box is added here to reflect how the Properties will be updated when that item is also added, but is a separate feature)*\n\nIf set to ***Active***, the Properties will continue to function as it does today, only showing properties for the active item.\n\n## When there are multiple selected items\n\nWhen the user has selected multiple items, a few things change:\n\nInstead of showing the name of the object at the top, Blender tells the user how many items are selected:\n\n\nIn the list of tabs, only properties sections are shown that are common among the selected items. For example, if you select a text and a mesh object, we don't show Modifier, Mesh or Curve Properties, since these properties only exist on a subset of the selected objects:\n\n\nThis rule also goes for individual panels. Here, both a Sun and Spot light are selected, and we only see properties that are shared between them. A note at the bottom clarifies that you are only seeing the shared subset:\n\n{[F8364914](Screenshot_2020-02-23_at_18.49.33.png), size=full}\n\n## When values differ\nWhen multiple items are selected, Blender will compare values to check if they are the same, or if they are different. If they differ, we change the way properties are displayed, like so:\n\n{[F8363037](Screenshot_2020-02-22_at_17.52.51.png), size=full}\n\n\n\n## Manipulating values across multiple items\n\n\n\nFor simply setting a value to be the same. you would simply type a new value, hit Return, and that value is then propagated to the selection.\n\nFor manipulating numbers and sliders relatively, you can just drag on them (like Alt-dragging today), or type the character '=', so the user can type '= -3' to subtract 3 from the value on all selected items.", "EXCEPTION_ACCESS_VIOLATION error when trying to change object viewport color with the viewport compositing enabled in depsgraph_update_post\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 680/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.27\n\nBroken: version: 3.5.0\nWorked: 3.4\n\nThis error occurs only if the viewport compositor is enabled - rendered mode of the viewport and Compositor = Always.\nWhen trying to change the viewport color of the object by script on depsgraph update event, this error occurs.\n\nThe short script example:\n```\nimport bpy\n\ndef on_depsgraph_update_post(scene, depsgraph):\n if depsgraph.id_type_updated('SCENE'):\n for obj in depsgraph.updates:\n if isinstance(obj.id, bpy.types.Object):\n print('change object color')\n bpy.data.objects[obj.id.name].color = (1, 0, 0, 0)\n\nbpy.app.handlers.depsgraph_update_post.append(\n on_depsgraph_update_post\n)\n```\nAfter executing this script in the Text editor, switching to the \"Rendered\" mode of the 3D Viewport and enabling viewport compositing - when adding or copying the object - Blender throws an error.\n\nThis occurs only if the viewport compositing is enabled, if it is not used - code works as expected.\n\nError occurs on 3.5 and 3.6 alpha.\n\n1. I attached .blend file with the example\n2. Open .blend file\n3. Switch the viewport to the \"Rendered\" mode (z - Rendered)\n4. Execute script in the Text editor\n5. Try to copy cube (shift + d)\n6. Error throws\n\n", "Keyframed parameters are not preserved when frame_change_post handler is used\nOperating system: Win 10 Pro\nGraphics card: GTX 1080\n\nBroken: v2.81, master, 2018-11-26\nWorked: (optional)\n\n---\n\n\n**EDIT** by @EAW \n\n\n> In the .blend file setup, the 'mesh_cache' object would instance a cube over the vertices for display in rendering. An ocean modifier is added to the object with 'Geometry' set to Displace and the 'time' parameter is keyframed over a frame range. \n> \n> Playback in the viewport shows the ocean modifier evolving shape during the animation. If the animation is rendered, the ocean modifier does not evolve over time and the render result is static.\n> \n\n\n\n> Here is a script that generates a 2d grid of vertices in the frame_change_post handler:\n> \n> ```\n> import bpy\n> \n> def frame_change_post(scene, depsgraph):\n> # Generate a 2d grid of vertices. A cube object is set to be instanced over the vertices for rendering\n> vertices = []\n> for j in range(100):\n> for i in range(100):\n> vertices.append((i * 0.05, j * 0.05, 0.0))\n> \n> mesh_cache_object = bpy.data.objects.get(\"mesh_cache\")\n> mesh_cache_object.data.clear_geometry()\n> mesh_cache_object.data.from_pydata(vertices, [], [])\n> \n> depsgraph.update()\n> \n> \n> bpy.app.handlers.frame_change_post.append(frame_change_post)\n> ```\n> \n\n\n\n\n> Here is the .blend file including the script: [keyframed_modifier_bug.blend](keyframed_modifier_bug.blend)\n> \n> How to reproduce:\n> 1. Open .blend file\n> 2. Press 'Run Script'\n> 3. Blender > Render > Render Animation\n> \n> The result is a static animation.\n\n\n**End Edit**\n\n---\n\n\n\nI have created a FlipFluids simulation in Blender 2.81 ( 2019-11-25 21:45) and added an Ocean modifier in displacement mode to the fluid_surface. Then I have baked the complete sim from within FlipFluids.\nWhen I move from keyframe to keyframe in the default timeline everything is moving as desired in the default 3D window:\nOcean modifier waves, FlipFluids waves & wake, flags, radars...\n\nWhen I render the complete animation everything EXCEPT the ocean modifier waves are moving. The ocean modifier waves remain static but the FLIPFLUIDS waves move on that surface.\nALL modifiers and elements ARE enabled for renderings. D-NOISE: AI Denoiser is on (switching denoiser off did not help).\nComputer is a HP Z800 with dual Xeons, 72 GB of vRAM and a GTX 1060.\n\nIf I render & save the keyframes manually individually by hand the ocean surface moves! Which means I get the proper result.\n\nAs its working when rendering&saving each frame manually as image (Png) and its not giving the same result when saving out the same images (Png) when I hit render animation, I do suppose this might be a bug in 2.81.\nRunning indiviual frames via Renderpilot did not yield the proper result either. So I do suspect it might be that there is not enough time gap between saving the last frame and rendering the next. Or, something does not get purged properly.\n\nFor propriatary reasons I can not share the blend file.", "Generated Texture Coordinates Don't Update in Edit Mode.\nOperating system: Linux-5.15.0-52-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.65.01\n\nBroken: version: 3.3.1\nWorked: N/a\n\nGenerated texture coordinates don't update in edit mode, example when translating the vertices. I tried changing other coordinates, (uv, object), they all update fine in edit mode.\n\nBased on the default startup or an attached .blend file \n[Generated_editMode.blend](Generated_editMode.blend)\n1. Preview generated texture coordinates.\n2. Go into edit mode.\n3. Select everything and move it around in one direction (ex the left/-x).\nExpected result is the generated coordinates are identical to before moving the cube. Tab out of object mode and then it will update. \n\n", "Pinning data (mesh, curve, grease pencil, lattice, particle system) makes certain tabs/options disappear (vertex groups, shapekeys facemaps)\nOperating system: Microsoft 10\nGraphics card:GeForce GTX 10606GB\n\nBroken: 2.93\nWorked: don't know\n\n**in the properties editor, in object data properties, the pin toggle hides vertex groups and shapekeys**\n\n- select default cube\n- go to properties editor\n- press object data properties tab\n- press the pin button\n", "Python API: Mesh editing: Editing uvs after normals with a previously assigned variable doesn't work\nBroken: 3.5.1\nWorked: 3.4.1\n\nCaused by 6c774feba2\n\nHolding on a variable referencing the uv layer data for later uv editing, doesn't work when normals are edited in between\n\nRun the attached python script in Blender 3.5.1\nThis creates three objects, all should have all their loop uvs as 0,0:\n\n- no_set_normals (should have default normals, sideways)\n- with_set_normals_and_reaccess_uv (should have upwards +z normals)\n- with_set_normals_but_no_reaccess_uv (should have upwards +z normals)\n\nGo in edit mode on each generated object, look at the normals and uvs\nThe results are only correct for no_set_normals and with_set_normals_and_reaccess_uv\nFor with_set_normals_but_no_reaccess_uv, neither the uvs nor normals are correct. Both appear to be defaults\n\nIn Blender 3.4.1 this works as expected, i.e. with_set_normals_and_reaccess_uv and with_set_normals_but_no_reaccess_uv are identical\n\nAttached are also screenshots in 3.5.1 of the normals and uvs of each object\nAnd a screenshot in 3.4.1 showing the expected results\n\n**Guess at what's happening**\nOn more complicated cases (actual meshes with hundreds of triangles), the normals are messed up (seemingly random instead of defaults) and Blender is prone to crashing. This leads me to believe holding on the uv layer data after normal editing accesses freed memory or something like that?\n\n", "Allow changing the active modifier on linked objects\nUnzip the files from [active-modifier-linked.zip](active-modifier-linked.zip)\n\nOpen `linked.blend` and try to change the active modifer. You can't because it is linked.\n\nIn order to allow users to inspect their nodetrees (and eventually see the evaluated result of the modifier) I suggest we treat this as UI data and allow users to change that even for linked objects.\n\nWe do something similar already for the modifier \"Realtime - Display in viewport\" option.\n", "Hair: changing children related settings in objectmode dont show when switching to particle editmode\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.87\n\nBroken: version: 2.83.7\nWorked: 2.79\n\nWhen I change additional options like Clumping, children not updates and their update stuck for edit mode for any try except 1st. \nDemonstration and details:\n[2020-10-15_01-46-01.mp4](2020-10-15_01-46-01.mp4)\n\n#69000 (Hair: changing children related settings not updating in particle editmode) is very much related", "GPUOffscreen missing update when deforming geometry in Sculpt mode\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: Radeon RX 580 Series ATI Technologies Inc. 4.5.14830 Core Profile Context 21.4.1 27.20.21002.112\n\nBroken: version: 2.93.1 and since 2.82 (sub 7)\nWorked: 2.81, 2.80 (and even 2.82 alpha)\n\nGPUOffscreen missing update, cause when creating some from the 3d view it is not updated when performing deformations on the mesh in Sculpt mode.\n\nYou can try the '[Rendering the 3D View into a Texture ](gpu.html#rendering-the-3d-view-into-a-texture)' example in the Python API docs and see that it works in 2.80 and 2.81 but stop working on 2.82 while in Sculpt Mode (you deform the mesh with any brush and the offscreen is not updated).\n\nAlso there is no mention about this breaking change in the Python API section of the release notes of 2.82 and as it worked in previous releases I suppose this is a bug.\n\n\n\nThe easy way of testing is via the GPU Offcreen example in the API docs --> '[Rendering the 3D View into a Texture ](gpu.html#rendering-the-3d-view-into-a-texture)'\n - Open Blender 2.81, and run that script\n - Go to edit mode, move some verts, confirm it works (you can see the offscreen is updated)\n # Go to sculpt mode, deform the mesh, confirm it works...\n\n # Repeat steps 1-3 but in Blender 2.82 or latest release at this time (2.93 RC1)\n\n", "Tweaking redo values in extend vertices operator causes extended vertex to dissolve\nBroken in 2.80 (1d908bffddb4 from 2019-01-24) as well as 2.79b release\n\nTweaking redo values in extend vertices operator causes extended vertex (the one from which the operator is executed) to dissolve", "Can't edit new driver on an overriden property on a linked object with no animation data\n## Simplified Report\n\n- The following file contains an object with no animation data.\n- Link the object to a new file and make a library override from it.\n- Add a new driver on one of its location channels.\n- Notice that driver options are grayed out and can't be edited.\n- One expects to be able to edit the driver since the object has no animation data. \n\n[libraryFileWithNoAnimationData.blend](libraryFileWithNoAnimationData.blend)\n\n## Original Report\n\n\nOperating system: Linux-5.15.60-1-MANJARO-x86_64-with-glibc2.36 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.65.01\n\nBroken: version: 3.3.0\nWorked: ??? I could have sworn this worked in 3.2 but it seems to have the same behavior. Same in 2.93.\n\nUsing drivers with overridden properties is not really usable. I can add a driver, but then I cannot edit it.\n\nLink a collection into a file and make a library override. Right-click on a property and choose \"Add Driver\".\nRight click on the property (which should turn purple) and then choose \"Edit Driver\" Notice that the\ndriver's controls are all blurred out! You can't edit them in the UI.\nIt gets worse!\nOpen a driver-editor window so you can watch what happens in the next step, and select the driver you\njust created- open the \"Drivers\" tab in the window. Look at the expression.\nIn the main window (not the driver window), click on the driven property and type ~~\"#1\" - when you\nprepend a hash symbol, that tells Blender to edit the expression of a driver (it's a handy way to make\nnew drivers with simple expressions. Try it at home!)~~ **edit: it looks like you don't need the hash when it is already driven.**\nNow, notice the expression change in the driver window that you aren't normally able to edit!\nI have also confirmed that you can edit some properties with the Python API that you cannot edit in\nthe user interface.\n\n**My Thoughts** (If you're interested)\nI do not know if the design for overrides is intended to support using drivers like this. It feels a little squishy\nbecause I don't know if Blender will keep the driver around when I resync the override. On the other hand, it's\nreally handy when doing animation that is easier to do programmatically, such as causing a bone to oscillate\nwith sine-waves (e.g. a plant waving with the wind). Sometimes I use a driver to bake keyframes - so I set up the\ndriver and then bake its motion to an action, and get rid of the driver. It is my opinion that driver support is\nnecessary for overrides, because it's something I expect to be able to use, as an animator. I don't know if the\nold Proxy system supported drivers, though.\nAt any rate, if it is blocked from editing in one place, it should be blocked from editing through the API, too.\n\nI apologize if this is a known issue. I don't see any similar bug reports but it appears things have been like this\nfor a while.", "Undo doesn't consider changes in nodetree when object is in edit and texture paint mode\nOperating system: Windows-10-10.0.18362 64 Bits\nGraphics card: GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 436.48\n\nBroken: version: 2.79-3.4\n\nFor reference: this is due to this commit 4a08b974f4\nFrom the commit message:\n\"don't store undo steps for property changes in paint & edit modes.\"\n\nWhen in edit and texture paint mode, I can't undo changes, that I made in the shader editor. \n\n\n - Open .blend\n - Move some verts.\n - Change a shader node setting.\n - Press {key Ctrl Z}.\n # The verts move back, but the node setting remains.\n[#71895-undo-not-undoing-shader-nodes.blend](T71895-undo-not-undoing-shader-nodes.blend)", "Causing other objects to be named .001 doesn't immediately update the Viewport Name Display\nI found this in 3.0.0 but happens the same in 3.1.0.\n\nWhen you turn on Viewport Name Display:\n\n\nAnd then you rename a new object the same as object that was already in the scene:\n\n\nThe name of the original doesn't update until you somehow \"touch\" the object by moving it, entering Edit Mode etc.\n\nIt doesn't get fixed by simply affecting something else.\n\nIn the .blend you need to select the object with name \"Plane.001\" and rename it to \"Plane\" then the object on the left stays named as Plane until you somehow affect it.\n[Bug-ViewportDisplayName.blend](Bug-ViewportDisplayName.blend)" ]

[ "Changing vertex colors doesn't work for objects parented to armatures\n**Category**: Conversion\n\n\n%%%Vertex colours can normally be changed through Python.\nWhen an object is parented to a bone AND has one or more vertex groups, the colours will no longer change.\n\nSee attached example.%%%" ]

[ "Color > Multiply doesn't multiply alpha channels\nOperating system: Win11\nGraphics card: RX580\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen) 4.0.0 alpha\nWorked: (newest version of Blender that worked as expected) 3.6.2\n\nThe Color > Multiply section in a video sequencer strip no longer multiplies the alpha channel when that feature is pretty important for creating outlines using the glow effect. Perhaps adding an option whether to multiply alpha or not would be beneficial?\n\nStart up blender 4.0.0\nSwitch to Video Editing workspace\nAdd a white color strip with 0.1 opacity\nAdjust the Color > Multiply value and notice how it doesn't affect alpha anymore\n\n", "Improve upon the automatic weight paint assignment by using Bone Glow\nBlender currently uses the *Bone Heat*algorithm to automatically assign bone weights to a mesh.\n\nThe proposed *Bone Glow* algorithm improves over *Bone Heat* by fixing cases where it behaves poorly.\n\ndownload?doi=10.1.1.156.4787&rep=rep1&type=pdf\n\nThe provided example images should look very familiar to anyone that has used Blender's Bone Heat.\n\n", "Color attributes in sculpt mode not showing in instance objects\nOperating system: Linux-5.15.0-47-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: GeForce GTX 670MX/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 418.226.00\n\nBroken: version: 3.4.0 Alpha\n\nWhen sculpting an object its instance reflect the changes after the stroke, but not with color attributes, you have to go out of sculpt mode to see the color attributes.\n\n1. Open a new sculpt file.\n2. Make and instance in object mode with Alt D\n3. Enter sculpt mode of the original object.\n4. Paint some color attributes\n5. Sculpt with clay brush to see the expected behavior.\n\n[Grabación de pantalla desde 17-09-22 16:08:53.webm](Grabación_de_pantalla_desde_17-09-22_16_08_53.webm)\n\n\nOne funny thing is that if you give the original object a mirror modifier, the changes to the instance are in real-time, not after the stroke, both for sculpt and for paint. But not when you sculpt or paint in the one without the modifier.\n\n\n[Grabación de pantalla desde 17-09-22 16:22:16.webm](Grabación_de_pantalla_desde_17-09-22_16_22_16.webm)\n\n", "Viewport shading solid mode not displaying geometry nodes color attributes correctly\nOperating system: Win10\nGraphics card: 960m\n\nBroken: 3.5\nWorked: never\n\n\nWhen using geometry nodes to add color attributes, there are several situations where the viewport solid mode attribute option does not display the color attributes correctly.\n\nAs can be seen by this example. The first image shows a cube with a color attribute added. The geometry nodes setup adds a red color to the \"Color\" attribute. This all works as expected as can be seen in the viewport and the spreadsheet. \n\n\n\nIf a GN cube is added instead of the standard mesh cube (which has been bypassed by the switch node) then the color attribute is added as expected. This can be seen in the spreadsheet as to be the same as the previous screenshot for the color attribute. However, it is not being displayed in the viewport.\n\n\n\nThis is a problem seen with any procedural geometry made with just nodes. As a workaround, you can have a single vertex as the only \"real\" geometry and use a join node to join with any nodal procedural geometry. Strangely though, this has to be the first connection in the join node, and won't work if it is second. This can be seen by the next 2 images:\n\n\n\n\n\nA similar related case is that some mesh nodes will also make the viewport lose the displayed attribute. For instance, if you create a cube, add a GN modifier and add a dual mesh node to the chain, the color attribute is still there, but no longer displayed in the viewport. This happens with some nodes (like mesh to volume and back to mesh) but not others (like subdivide or extrude).\n\n\n\nThere is also a related issue that when exporting to obj or gltf the color attributes are not exported if they are not displayed in the viewport. If apply the GN modfier, you have to actually click on the color attribute in the properties panel to see it in the viewport or to be able to export it. \n\nOddly, the fbx exporter seems fine without applying the modifier. It also doesn't matter if the color attribute is on the vertices, or face corners, or if the color is float or byte color. The same issues happen regardless.\n\n", "Xray doesn't work in wireframe shading for objects set to display as wire.\nBroken: 3.0\nWorked: \n\nXray doens't work in wireframe shading mode for objects set to display as wire.\n\n- Open .blend file\n- Change the viewport shading to wireframe.\n- Set one of the cubes to \"Display As : Wire\" in object properties->viewport display.\n- change X-ray value from `Viewport Shading` popover\n- compare the two, only one has xray effect.\n\nTest File:\n[#94802.blend](T94802.blend)", "Texture Paint: Color Palette operator issues with Undo/Redo -- Vertex Paint is fine though\nOperating system: Darwin-15.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: ATI Radeon HD 5770 OpenGL Engine ATI Technologies Inc. 4.1 ATI-1.42.15\n\nBroken: version: 2.83.0\n\n\nIn Texture Paint mode (contrary to vertexpaint mode or gpencil), the palette operators dont work with Undo/Redo\n\nFrom the default startup file:\n- go to vertexpaint mode\n- (a) in the Active Tool tab in the Properties Editor, under Brush Settings > Color Palette, add a palette\n- (b) add a color to the palette\n- (c) Undo --> should be working [color is removed again]\n- so far everything is working\n- go to texture paint and do the same a, b, c --> color is not removed again, Undo is not working\n\nThis is what is probably the cause for the following more specialized original report (but also there the same thing works fine in vertex paint mode)\n\n## original report\n\nWhile in texture mode users press S to activate a color picker (not sure why this isnt the GUI nor the tools palette?)\nWhen its active, we have the options menu (used to be F6 i believe) There is this option \"Add to Palette\".\nThis however always adds colors. When you uncheck it, it will enable itself the next time your use the S key and pick a color.\n\n01. Open the attached scene.\n02. Start texture paint mode\n03. Use S key to enable color picker\n04. Options menu (bottom left) shows \"Add option to palette\" > Deactivate it.\n05. Use S key again > Notice colors are still being added but option is unchecked.\n06. When you let go of the S key, its checked again\n\n\n\n\n[IMport Palette - Blank.blend](IMport_Palette_-_Blank.blend)", "object display as wire , invisible when overlay is turned off \nwin 10\nintel i5\namd radeon \n\nblender 2.8 \n6b39dc7672eb\n\n\nNot sure if it's abug or intended behaviour \n\nWhen an object is set to wireframe ( object properties , viewport display ) , when overlays is turned off , the object is not shown ", "Geometry Node: Input Color node: Custom color and value color is the same property\nOperating system: Linux-5.15.0-60-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 3GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.86.01\n\nBroken: version: 3.4.1\nWorked: Never.\n\n\nCustom color and value color is the same property. Impossible to set custom color to node, always edited value color.\n\n1. Create Geometry Node tree.\n2. Create `Input Color` node\n3. In `N` panel, `Node` page:\n a. Enable `Custom Color` button.\n b. Try to change custom color.", "Vertex Painting with data type \"Color\": Painting pure Red or Green creating non Zero in Blue Channel\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 466.63\n\nBroken: version: 3.5.0 Alpha\n\n**Short description of error** \nIn Geometry Nodes, I've tried scattering objects with vertex painting (data type \"Color\") using pure red (1,0,0), green (0,1,0) and blue (0,0,1). With a math node set to \"Greater than\", the different color channels are tested whether their values are bigger than 0. However, I discovered that I needed actually a \"Threshold\" value slightly bigger than 0 (I used 0.0001, just below the \"Spreadsheet's\" precision) because my objects for the blue painted areas got also displayed in red and green areas with a \"Threshold\" value of exactly 0 for testing the blue color channel. The red and green channels could be tested with a threshold of exactly 0. \n\nWith vertex painting using the data type \"Byte\", I could also set the blue channel's testing threshold to exactly 0.\n\n**Steps to Reproduce**\n- Open attached file\n- Set Threshold value exact to 0 for blue color channel\n[Geometry-Nodes_Instancing_with_Vertex-Colors_1Jan2023_7.blend](Geometry-Nodes_Instancing_with_Vertex-Colors_1Jan2023_7.blend)\n\n", "Support for snapping to Lattice objects\nOperating system: Darwin-17.2.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-10.27.6 378.10.10.10.20.109\n\nBroken: version: 2.82 (sub 0)\nWorked: (optional)\n\nIt is not possible to snap any type of vertex (mesh, curve or lattice vertex) to a lattice vertex.\n\nsee video\n\n[latice_verts_not snapping.blend](latice_verts_not_snapping.blend)\n\n[lattice_verts_not_snapping.mp4](lattice_verts_not_snapping.mp4)\n\n", "Cycles: sRGB texture compression happens after interpolation \nBroken: 2.81\n\nCycles does sRGB->linear texture conversion after texture interpolation instead of before. This leads to incorrect results.\n\nExample scene is attached. Two squares with a fine checkerboard texture, one with linear interpolation, the other with closest interpolation.\n\nExpected result: Both textures should render to an identical shade of gray. When switching the texture color space to \"linear\" instead of \"sRGB\", the result should not change as the texture contains only pure black and white pixels.\nActual result: The textures show different shades of gray. Switching the color space changes the result.\n[checkerboard.blend](checkerboard.blend)\n\n", "Unable to accurately sample colors in Vertex Paint mode\nOperating system: Linux-5.10.11-arch1-1-x86_64-with-arch 64 Bits\nGraphics card: GeForce GTX 660M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.39\n\nBroken: version: 2.92.0\nBroken version: 2.83.1\nWorked: 2.79b\n\nSampling colors in Vertex Paint mode is unreliable even in flat shading or when utilizing an emission shader and standard color management (although this reports the correct color some of the time) in both Cycles and Eevee.\n\n1. With the default cube selected, enter Vertex Paint mode.\n2. Switch the viewport shading to Solid, the lighting to Flat, and the color to Vertex.\n3. Change the brush color to #000003.\n4. Go to the Paint menu and select Set Vertex Colors (or press Shift+K)\n5. Sample the color on the cube by either using the eyedropper from a color picker, or pressing S with the mouse hovering over the cube (doesn't matter if Sample Merged is checked or not)\n6. Check the brush color and see that it has changed to some other color (#010103 on my machine)\n", "Animation: Replacing bone layers + groups with bone collections\nThis design describes the replacement of armature _bone layers_ & _bone groups_ with a new _bone collections_ system.\n\nImplementation is in #109976 and #112271.\n\n## Goals\n\n- Named collections, instead of numbered layers, should give a better indication as to what the collection contains / is for.\n- Increased flexibility in number of collections.\n- More similarity to the scene object collections, including features like 'not selectable' which can increase clarity of how to use a rig by animators.\n- Compatibility with library overrides, so that new shot-specific bone collections can be added.\n- Unification of numbered bone layers (for organisation) and named bone groups (for coloring).\n\n## Properties of the System\n\n- Armatures have one or more bone collections.\n - If there is only one collection it cannot be removed.\n- Every bone is in zero or more bone collections.\n- Bone collections **cannot be nested**. At least not for 4.0, can be added later.\n- Bone collections must have a **unique name** within the armature.\n- Library Overrides:\n - New bone collections can be added.\n - Existing ones cannot be reordered / removed.\n- Bone collections have the following properties:\n - Name\n - Selectable: 'set' wins over 'cleared'\n - Visible: 'set' wins over 'cleared'\n - `IDProperties` (aka Custom Properties) for use by add-ons.\n- Bones will get individual color properties.\n - The color system of bone groups will be used, which allows selection of 'default', 20 theme-defined colors, and a custom RGB value.\n - Bones have a color defined as part of the armature data. Here 'default' is the default grey color.\n - Pose Bones have a color defined as part of the object/pose data. Here 'default' means 'uncolored', and will thus show the edit bone color.\n\n## Operations\n\nThese operations can be performed on/with bone collections:\n\n- (De)select one or more collections themselves\n- (De)select all bones in active collection\n- (De)select all bones in collections containing the currently selected bones\n- Create a new collection\n- Delete a collection. Bones in the collection that are now 'orphaned' are assigned to the 'default' collection.\n- Move collection up/down in the list.\n- Move selected bone(s) to a collection.\n- Select Bones of Same Color.\n\n## User Interface\n\nFor Blender 4.0 the UI is kept intentionally simple. I'm thinking of reusing the Bone Groups interface, with the addition of a few per-collection icons (visible, selectable, color palette) as well as an indicator of which collection(s) contain the selected bone(s).\n\n## Conversion from Old to New\n\n- Layers converted to `Layer {NUM}` collections.\n- Bone groups converted to `{name}` collections.\n- Selection Sets converted to `SelSet {name}` collections.\n- Maybe even automatically merge these when they contain the same bones?\n\n## Uses of the Current System\n\n### Bone Layers\n\n- Visibility\n- Logical grouping\n\n### Bone Groups\n\n- Coloring of logical groups\n- Set by rigger\n\n### Edit Mode vs. Pose Mode\n\nThis describes Blender's **current** behavior:\n\n| Property / Action | Edit | Pose |\n|---------------------------------------|------|------|\n| Name Bone | ✅ | ✅ |\n| Set Bone Layers | ✅ | ✅ |\n| See Bone in Group Color | ❌ ¹⁾ | ✅ |\n| Create / Assign Bone Group | ❌ ¹⁾ | ✅ |\n| Select Bones in Group | ❌ ¹⁾ | ✅ |\n| Set Bone Group Color | ✅ | ✅ |\n| Set Armature Viewport Display options | ✅ | ✅ |\n| Set Bone Viewport Display options | ❌ ²⁾ | ✅ |\n\n¹⁾ The new system should allow these in Edit mode too.\n²⁾ Bones can only be hidden, which is separated between edit mode & pose mode.\n\n## Future Work\n\nThese are ideas for future improvements / extensions:\n\n- Nestable bone collections.\n- Integration with the Outliner.\n- Per-collection option whether bones are deforming or not (rather than setting that per bone)\n- Per-collection viewport display mode (octo, line, bbone, etc)\n- Possibly unification of Selection Sets as well.\n\n## Code Location\n\nThis work will kickstart a new module in `source/blender/animrig`. This will become the place for new animation/rigging related code. The intention is that also a subset of the already-existing animation-related code from both `blenkernel` and `editors` is going to move there, to help clean up the API and make it more consistent.\n\n-------------------------\n\n## Ongoing Work\n\n**Targets for 4.0-bcon1: tracked in #109976**\n\n**Targets that could be pushed back to 4.0-bcon2:**\n\n- [x] Add ID Properties support to bone collections, including a UI.\n- [x] Removal of bone group RNA access & UI\n- [x] Ensure that library overrides can add new bone collections & assign bones to them (#112271)\n- [x] Removal of bone groups from the UI, including the Ctrl+G menu\n- [x] Add bone collections to the outliner (replacing the current list of bone groups)\n- [x] Operators to move/add selected bones to a collection:\n - [x] move to collection (similar to 'm' for objects & collections, from viewport)\n - [x] add to collection (similar to 'shift+m' for objects & collections, from viewport)\n- [x] Select Similar operator (by collection, by color) / Shift+G menu\n- [ ] Fixing the Collada I/O, as that also handles bone layers.\n- [x] Make bone collections on linked-but-not-overridden armatures usable. The object itself may be overridden, so changing the active bone collection should work (even when it cannot be saved), and then 'Select Bones' and 'Deselect Bones' operators should work.\n\n**Potential push back to later:**\n\n- [ ] Add selectability toggles to bone collections (only for pose mode)\n- [ ] Add armature-global override to selectability, so that a rigger can still opt to select all the bones, then lock them down for animators.\n- [ ] Managing bone assignments to collections from the outliner.\n", "Curves don't have 'All Edges' option for wireframe\nBroken: 2.80\n\nMesh objects have the toggle for its viewport display wireframe to render all edges, but curve objects don't.\n\n\nOpen the attached file and look for the 'All Edges' option.\n[Wiref.blend](Wiref.blend)\n", "Geometry Nodes: writing attributes (from procedurally created geometry) to vertex groups not working\nOperating system: macOS Monterey 12.4\nGraphics card: RX 6800 XT\n\nBroken: 3.1 & newer\nWorked: 3.0.1\n\n\nIn Blender 3.0.1, it was possible to write attributes from procedurally created geometry (created by geometry nodes) to vertex groups, as long as the \"Group Input\" node was joined with the procedural geometry. However, no version newer than 3.0.1 seems to have this capability. From what I understand, this may have been done intentionally in anticipation of attributes replacing vertex groups entirely... but it seems odd to kill this feature before that transition is complete, so I'm reporting this as a bug.\n\n\nOpen the attached .blend file in Blender 3.0.1:\n[geonodes_attrib_to_vgroup.blend](geonodes_attrib_to_vgroup.blend)\nThis file contains a geonodes setup where half of a procedurally generated grid is extruded using a solidify modifier. The \"geonodes\" object should look exactly the same as the \"Expected Result\" object. \n\nNext, open the file in Blender 3.3 (or 3.1, or 3.2) and notice how the \"Expected Result\" object no longer matches the \"geonodes\" object.\n" ]

[ "Unable to update mesh vertex properties via Python when mesh is influenced by armature\n\nWindows 7 64-bit\nAMD Radeon HD 7700\n\n\nBroken: Blender 2.7 3cf39fc\nWorked: Unsure if it ever worked correctly\n\n\nI reported this bug already before the bug tracker switched sites, but I guess it was lost in the move. So here's the report again.\n\nBasically, if an object has an armature modifier on it and is parented to an armature (to be deformed by an armature), you can no longer alter its mesh's vertex properties by Python. They update in Python, but the changes aren't applied visually. I would understand if you were just unable to update a vertex's position and perhaps its normals, as those properties are being altered by the armature. \n\nHowever, I think one should be able to alter the vertex's UV values and color properties (as well as any others that they wish), if possible, since the armature shouldn't have any influence on these properties.\n\n\n1) Just press P to start the game engine and see the cube sit there, unmoving (the bug).\n\n2) Select and delete the armature.\n\n3) Press P to start the game engine and see the vertices of the cube move along slowly, quickly altering their colors and UV values (the correct behavior).\n\n[ArmatureParentedUpdateBug.blend](ArmatureParentedUpdateBug.blend)" ]

[ "Orbit inverted on Mac with Magic Mouse\nOperating system: Darwin-18.7.0-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 750M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-12.0.24 355.11.10.50.10.103\n\nBroken: version: 2.92.0\nWorked: 2.91.0\n\nIn 2.91, when i moved my fingers on the touch-sensitive Magic Mouse to orbit, it would move in the opposite direction, which is what i, and most likely most Magic Mouse users are used to; however in 2.92.0, the orbit is the same as using a regular scroll wheel. Very disorienting, and there are no settings to invert this navigation, as far as i am aware.\n\nDefault Startup > Orbit with Apple Magic Mouse, touch sensitive surface.\n\n", "Push/Relax/Breakdown pose operators may flip rotation of quaternion bones\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 2.93.8\n\nPush/Relax/Breakdown pose operators may flip rotation of quaternion bones\nWhen the animation plays, this bone suddenly rotates unnaturally.\n\n - Open the attached blend file\n - Use Push/Relax/Breakdown to the bone and set a keyframe.\n - Check the bone animation.\n[flip_when_use_breakdown.blend](flip_when_use_breakdown.blend)", "Expose symmetry options for operators that use them in weight paint\nCurrently there are quite a few weight paint operators that use the symmetry/mirror setting under the hood.\n\nThis is great, but we should expose this in the operator settings so that is might be easier for the user to realize that the mirror/symmetry options also influence operators if they are on.\n\nThe new exposed operator settings should use the current symmetry options from the mesh/object.\nBut the user can then see and set the symmetry value manually in the operator if they wish.\nThis can be used to create custom shortcuts/key binds that always has symmetry on or off.\n\nThe operators use `ME_SYMMETRY_X` or `ME_EDIT_VERTEX_GROUPS_X_SYMMETRY` and are in `source/blender/editors/object/object_vgroup.c` and `blender/editors/sculpt_paint/paint_vertex_weight_ops.c`.", "Simple Deform Modifier - Bend Mode\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 980/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 445.75\n\nBroken: version: 2.82 (sub 7)\nWorked: (optional) possibly never?\n\n'Simple Deform' modifier - 'Bend' mode performs as expected if mesh extends along X axis or Z axis. Performs unexpectedly (similar to Taper mode) if mesh extends along Y axis\n\nStart with new 'general' scene.\nSelect default cube\nAdd 'Array' modifier with values for relative (or constant) offset as: X=2, Y=0, Z=2 (so it extends along X and Z axis)\nSet 'Count' to 6 to have 6 cubes along the X axis\nAdd 'Simple Deform' modifier\nSelect 'Bend' mode\nDefault values for Bend mode will be Axis: X and Angle: 45\nMesh will bend as expected around X axis - each cube is rotated around a vector from origin along X axis keeping the expected shape\n\n\nChange Axis to Z\nMesh will bend as expected around Z axis - each cube is rotated around a vector from origin along Z axis keeping the expected shape\n\n\nChange Axis to Y\nMesh rotates along vector from origin along Y axis, but tapers. \n\n\nThis behaviour is very similar to how the Taper mode works\n\n\nPlaying with the angle values for each axis will clearly illustrate the issue. Changing angles on X and Z axes bends as expected. Changing angles on Y axis tapers and stretches.\n\nI've seen similar reports here that may be talking about the same thing, but they've been closed down as expected but unintuitive behaviour. I hope this shows that this is more likely a bug rather than simply unintuitive, as bending on both other axes works as expected.\n\nThanks all.\n\n[bend_X.blend](bend_X.blend)\n", "Flip the Render\nI proposed this idea to Campbell via email and he said it sounded reasonable enough to add here...\n\nThe idea: In the image editor, when the Rendered Result is displayed, pressing F will flip the render horizontally. This is a common action that many digital painters use in Photoshop to refresh their eyes and see new flaws in their work.\n\nSome notes:\n - It will should only work in the Rendered Result, not on regular images\n - All tools that normally work in Rendered Result, should work here as well. (rendered border, sampling image color, painting)\n - Only Flip X. I've tried Flipping Y before, but I haven't found it to be that helpful in the artistic process. Could be added later though if people want it?\n - Should apply the Flip state to all Slots. Eg. If flipped in Slot 1, when you change to slot 2 it will also be flipped. If you turn Flip off in Slot 2, it will also be off in Slot 1 and all slots.\n - To remove the potential confusion for users who flipped it by accidentally pressing F but don't know the hotkey, a new button should display in the toolbar. Here: ", "Self-referencing driver works in viewport, but not in render.\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: GeForce GT 730/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 391.35\n\nBroken: version: 2.82 (sub 6)\nWorked: (optional) 2.79b\n\nRotation property of object is updated in viewport, but not in render.\nIf run same script in 2.79 - all works fine both in viewport and render.\n\n```\ndef rotator():\n var1=bpy.data.objects[\"Suzanne\"].rotation_euler.z\n var3=bpy.data.scenes[\"Scene\"].render.fps\n\n new_value=var1+math.radians(var3)\n\n return new_value\n```\n\n\n[Script_bug_2.8.2.blend](Script_bug_2.8.2.blend)\n\n- run the file Script_bug_2.8.2 ('Run script', 'Update Dependecies')\n- play in the viewport - the animation works correctly\n- render (Ctrl + F12) - the animation does not work;\n\nViewport:\n[bug_video_render0001-0050.mkv](bug_video_render0001-0050.mkv)\n\nRender:\n[bug_video_render_20001-0050.mkv](bug_video_render_20001-0050.mkv)\n\n2.79 file:\n[Script_work_2.79.blend](Script_work_2.79.blend)\n\n", "Subtype for input value nodes\n\nThe user may be at a loss when entering a rotation radians value into some built-in node, copies it into the input value node, and now these are degrees.\nThe explicit way of specifying the sub type must be an option for all value input nodes that it can have.\nThe type list option should be available in the N-Panel so as not to clutter up the node space.\n\n", " LineArt: (Collection) Issue for Cameras in Orthographic View\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.3.1\nChecked in Blender version 3.4.0 Apha, commit date: 2022-10-08 02:26 - same issue\n\nFor orthographic Camera without any shifts - Line art modifier (for a Collection) showing lines for objects behind other objects - which otherwise should not be visible. And some lines are not visible which should be.\nHappens with contours option in the Line Art modifier. Slight rotation on z axis solves the issue.\n\nNoticed issue with some specific frames on my complete blend file.\n\nOpen attached blend file and go to front orthographic view.\n\n[BugReport01.blend](BugReport01.blend)\n\n", "Bone Override Transform affects bone local axis display.\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 522.30\n\nBroken: version: 3.3.0\n\nWhen a bone gets assigned a display transform override, it's viewport display axis gets transformed, which probably shouldn't happen as its purpose is to display the Bone's not it's Shape's local axes.\n\n[override-transform.blend](override-transform.blend)\n - Remove Override Transform from Bone's Viewport Display's Custom Shape -> Displayed Bone's axes change location.", "Fisheye lens results in Window projected textures being mirrored across X-axis\nOperating system: Linux x64\nGraphics card: GTX 860m\n\nBroken:\nAny of the 2.80 daily builds I tried since several months ago.\n\nUsing Cycles with a fisheye lens camera, both Equisolid and Equidistant, and some texture with Window projection from the Texture Coordinate node flips the X-axis of the texture.\nNone of the other lens and projection types behaved like this as far as I saw.\n\nHere's a render with perspective (top) and fisheye (bottom) illustrating the problem.\n\n\nAnd here's a small .blend file. Set the view to Rendered and the camera lens to Panoramic.\n[fisheye bug.blend](fisheye_bug.blend)", "Curve Extrude Tool Todo\n## Curve Extrude\n * - [x] {icon circle color=red} Add Extrude tool to Curve Edit toolbar \n * - [x] {icon circle color=yellow} Add Extrude manipulator when using this tool\n * - [ ] {icon circle color=green} The direction of the extrude handle is only correct on one of the extremes:\n\nCorrect:\n\nGizmo points away from the curve\n\nIncorrect:\n\nGizmo points towards own curve\n\nThe gizmo should always point outwards at each end. ", "Grease Pencil Shear Tool Todo\n## Grease Pencil Shear Tool Todo\n * - [x] {icon circle color=red} The tool should execute on release, so when you click and drag, it stops. Currently you need to click an extra time, which is inconsistent and annoying.\n * - [ ] {icon circle color=green} Add Tool Setting to switch between horizontal and vertical shearing\n * - [ ] {icon circle color=green} Shear should have a manipulator to shear in any direction, the orientation on which is controlled by the Orientation setting, just like the Transform manipulators. Dragging any of the handles shears in that direction.", "Exact Boolean with Incorrect Result - multiprecision/float round trips not exact\nOperating system: Windows 64 bit\nGraphics card: Nvidia GTX 750 Ti\n\nBroken: 2.91.0 Beta dca36a8ec922\n\nAnother case, where the exact boolean does not work as expected. (Sorry @howardt, I had to break it. I have to admit, that I have not run into this problem in a real scene.)\nAgain, this manifests after applying the same difference operation several times.\n\n[boolean_rotated_box3.blend](boolean_rotated_box3.blend)\nLook at the scene.\nThe first run of the modifier works as expected, the second creates overlapping faces, the third removes most of the mesh.\nThis \"bug\" is very erratic, and if the transform of the objects is slightly changed, it does not occur.\n\nThis is what I see:\n\n\nThe expected behavior is that the second and third modifier do nothing.", "Some particle hairs that emitted from the vertices do not rotate with the armature.\nOperating system:MacOS\nGraphics card:Radeon Pro\n\nBroken: 2.93.2\n\n\nSome particle hairs that emitted from the vertices do not rotate with the armature.\n\n\n\n\n\n\n\n\n\n[Hair Rotation bug.blend](Hair_Rotation_bug.blend)\n\nSimple example wit a cube:\n[#89749.blend](T89749.blend)\n\nIssue only happens when emitting from Vertices (emitting from faces is fine)", "Regression: Extrude Tool XYZ Handles overlap on orthographic view and small world scale\nOperating system: Linux-5.15.0-43-generic-x86_64-with-glibc2.35 64 Bits\nGraphics card: Quadro GP100/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.85.02\n\nBroken: version: 3.4.0, 3.3 , 3.2\nWorked: 3.1\nCaused by 5045968f24\n\nThis bug might be related to #99670.\nWhen using the Extrude Tool, the handle type can be changed to show XYZ axes. When the object is way smaller than 1 Blender Unit, it becomes impossible to grab the Y and X handles in orthographic view.\n\n- Scale down the cube to 0.01 percent\n- Enter edit mode\n- Make the Extrude Tool active and switch the tool settings to XYZ\n- Try dragging the Z or Y handles\n\nFirst 3 steps are included in file\n[#100321.blend](T100321.blend)" ]

[ "Rotating Objects\nOperating system: Both MacOS and Windows\nGraphics card: Both AMD Radeon M9 and Nvidia 2080Ti\n\nBroken: 2.91.0\nWorked: 2.83LTS\n\nIf I open Blender 2.91.0 on either Mac or PC and rotate the default cube (by pressing R and rotating clockwise) the rotation is NEGATIVE. It still rotates clockwise in view, however the input is negative. This means the input (read in top left corner) is negative clockwise but positive anti-clockwise. So if I hit R and type 20, previously that would rotate clockwise 20 degrees, now it rotates anti-clockwise 20 degrees. This must be wrong, it has been clockwise-positive forever in Blender, why the change?\n\nIn 2.83LTS the rotation is as expected. Default cube rotation (press R then type 20) rotates clockwise as expected. Not anti-clockwise as it is in Blender 2.91.0.\n\nPlease confirm this with me.", "Rotate with \"R\" key or using bpy.ops.transform.rotate produces inverted rotation. Effects Object, Mesh Edit and Pose Mode\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.38\n\nBroken: version: 2.91.0\nWorked: version: 2.90.1\n\n\nIf a numeric figure is entered after pressing \"R\" to rotate an object by a set amount in object mode or if bpy.ops.transform.rotate is called via a script or the console, objects rotate in the opposite direction to intended. \n\nA similar bug was present in 2.80 T66527 but this had a workaround that if you set the orient matrix in a script it would force the rotation to be in the correct direction. This workaround no longer works. \n\nThis bug also effects mesh edit mode and pose mode!!!\n\n\nLaunch startup file.\nGo into top view\nGo into orthographic mode.\nSelect default cube\nPress \"R\" on the keyboard and enter 15\n\nResult:\nCube rotates -15 degrees rather than +15 degrees.\n\nAlternatively:\n\nLaunch startup file.\nGo into top view\nGo into orthographic mode.\nSelect default cube\n\nthen in the console:\n>>> bpy.ops.transform.rotate(value=0.261799)\n\n\nResult:\nCube rotates -15 degrees rather than +15 degrees.\n" ]

[ "Particle Edit mode for non hair particles - 2 Crashes\nOperating system: Windows-7-6.1.7601-SP1 64 Bits\nGraphics card: GeForce GTX 760/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 390.65\n\n\nBroken: version: 2.80 (sub 69)\nWorked: (optional)\n\n\nParticle Edit: 2 Crashes\n\nParticle Edit: 2 Crashes\n\n- Open the blender file and click on the buttons \"Particle select and display point\"\n- Crash 1 (Screenshot)\n\n- Open the blender file and click on the tool \"Weight\" and click into the particles\n- Crash 2 (Screenshot)\n\n- Also had a crash with \"Puff\" \n- file not included\n\n- There are some tools that do nothing while others work\n- User gets no feedback / error message but give no feedback\n\n\n\n[2019-05-19-particle-edit-crash.blend](2019-05-19-particle-edit-crash.blend)\n\n", "Crash opening file using Clump Curves on Particle Children \n****\n\nversion: 2.81 (sub 16), branch: master, commit date: 2019-12-04 11:32, hash: f1aa4d18d49d, type: Release\nbuild date: 2019-12-04, 14:30:40\nplatform: Windows\n\n***********************************************************\n\nWhen I worked on a project using particles, after a short break I wasn´t able to reopen the project. The project crashed right after opening. When starting Blender from console the only error I could find was:\nError : EXCEPTION_ACCESS_VIOLATION\nAddress : 0x00007FF706DC271E\n\nno Blender log file was written!\n\nEven in debug mode there were no more hints but his one. I opened the project in an Ubuntu VM and it also crashed there. No chance to recover the file.\n\n\nUsing Particle following a nurbspath with Force Field \"Curve\".\nParticles rendered as objects.\nUsed Simple Children.\nWhen ticking the box: Use Clump Curve it crashes.\n\nEither Blender crashes right after ticking the box or when reopen the project. The latter then means that I´m not able to recover my work at all.\nI might have to do with particles already baked or not. Not sure. \n\n[T72763_crash_log.txt](T72763_crash_log.txt)\n\nfile written from c2a37867f2 that crashes on file read:\n[#72763.blend](T72763.blend)\nfile written from 2.79b that crashes on file read:\n[T72763_279.blend](T72763_279.blend)", "Regression: Attempt to move object with Copy Location (with Offset) crashes\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.67\n\nBroken: version: 3.4.0 Alpha\nWorked: version: 3.4.0 Alpha\n\nA consistent crash when trying to move an object with COPY LOC w/ Offset enabled.\n\n\n1. Open the attached blend file\n\n2. Move the pre-selected object in any direction, you may have to do this two or three times\n\n3. program will crash with EXCEPTION_ACCESS_VIOLATION\n\nThis is happening every time.\n\n```\nStack trace:\nblender.exe :0x00007FF73E03C510 blender::bounds::min_max<blender::VecBase<float,3> >\nblender.exe :0x00007FF73E03BD30 blender::FunctionRef<void __cdecl(blender::Bounds<blender::VecBase<float,3> > &)>::callback_fn<<lam\ntbb.dll :0x00007FF9998C4FD0 tbb::interface7::internal::isolate_within_arena\nblender.exe :0x00007FF7425212F0 blender::CacheMutex::ensure\nblender.exe :0x00007FF73E042390 BKE_mesh_minmax\nblender.exe :0x00007FF73E0409F0 BKE_mesh_boundbox_get\nblender.exe :0x00007FF73E02E940 BKE_object_boundbox_get\nblender.exe :0x00007FF73E02F410 BKE_object_dimensions_get\nblender.exe :0x00007FF73E613520 RNA_property_float_get_index\nblender.exe :0x00007FF73E235050 dtar_get_prop_val\nblender.exe :0x00007FF73E235410 evaluate_driver\nblender.exe :0x00007FF73E168590 calculate_fcurve\nblender.exe :0x00007FF73E156270 BKE_animsys_eval_driver\nblender.exe :0x00007FF73E4464F0 blender::deg::`anonymous namespace'::evaluate_node\nblender.exe :0x00007FF73E446320 blender::deg::`anonymous namespace'::deg_task_run_func\nblender.exe :0x00007FF7425175C0 tbb::internal::function_task<Task>::execute\ntbb.dll :0x00007FF9998CF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FF9998CF220 tbb::recursive_mutex::scoped_lock::internal_try_acquire\ntbb.dll :0x00007FF9998C4FD0 tbb::interface7::internal::isolate_within_arena\ntbb.dll :0x00007FF9998CA120 tbb::task_scheduler_init::terminate\ntbb.dll :0x00007FF9998CD800 tbb::thread_bound_filter::try_process_item\ntbb.dll :0x00007FF9998CD800 tbb::thread_bound_filter::try_process_item\nucrtbase.dll :0x00007FF9A5D192C0 recalloc\nKERNEL32.DLL :0x00007FF9A6F72690 BaseThreadInitThunk\nntdll.dll :0x00007FF9A882AA40 RtlUserThreadStart\n```\n\n", "Sculpt Mode - Using Mask while Multiress active = Crash\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1660 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.91.2\n```\n Tried also with\n```\n 2.83LTS\n 2.92beta 77f73a928439\n 2.93.0 ae1e68f5146d\n\nWorked: None\n\nWas working in Sculpt mode, tried to hide the arms and the program crashed.\nI tried with another objects with multires and it doesn't happen\n\nHere the .Blend with the problem\n [Multires prueba.blend](Multires_prueba.blend)\n\nGo to Sculpt Mode -> Try to mask with mask brush, Mask Lasso or Mask Box", "crash when changing viewport shading to wireframe mode\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 XT ATI Technologies Inc. 4.5.14760 Core Profile Context 21.2.3 27.20.14535.3005\n\nBroken: version: 3.0.0 Alpha, branch: master, commit date: 2021-04-23 20:25, hash: 17fca62fff9b, type: Release\nBroken: 2.81\nWorked: 2.80\n\n\nif you have a high poly mesh and try change the shading mode to wireframe. depending on the model it constantly crashes.\n\n\nI noticed this trying to make a retopo of a character. I tried to check what was causing de crash by deleting each object and testing the wireframe. \n\nI deleted everything but the main body mesh, without any uv map, vertex color or any other thing that could make it heavier, and also I tried reseting factory settings.\n\nI would send the file but its 200mb, the main mesh has about 4million vertices.\n\n\nthe crash doesn't happen when changing to eevee or solid view, it only happens when changing to wireframe.\n\nImportant note: if you DON'T move the camera after opening the file before changing to wireframe it doesn't crash even if you go to solid and wireframe back and forth as shown in the video bellow \n[2021-05-01 14-18-16.mp4](2021-05-01_14-18-16.mp4)\n\n[blender_system_info.txt](blender_system_info.txt)\n[blender_debug_output.txt](blender_debug_output.txt)\n[untitled.blend](untitled.blend)\n- open the file above\n- rotate it about 180degrees\n- activate wireframe", "Crash on render\n**System Information** MSI BRAVO 15\nOperating system: WINDOW 11\nGraphics card: RX 6600\n\n**Blender Version**3.6\n\nWhen attempting to render a scene in Blender, the software consistently crashes, causing frustration and hindrance to the rendering process. This issue is hindering users from efficiently creating their projects and requires immediate attention to maintain a stable and reliable rendering environment.\n\n[debug-log](https://projects.blender.orgattachment)\n[system info](https://projects.blender.orgattachment)", "Crash on calling bpy.ops.mesh.loopcut in python script\nOperation system: macOS BigSur 11.6 (20G165) \nApple M1\n\nVersion 3.4.1 (3.4.1 2022-12-20)\n\nBlender crashes in background mode when you are calling bpy.ops.mesh.loopcut with overridden context\n\nRun Blender from cli with args:\n/path/to/bin --background --python bug.py\n[bug.py](bug.py)\n\n", "Crash at the end of preview generation for a deleted datablock\nOperating system: Windows-10-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\nWorked: -\n\nBlender crashes if you trigger the generation of a preview for a data block and then delete the data block itself. Obviously, the preview data is generated, but since the datablock itself no longer exists, I get an `EXCEPTION_ACCESS_VIOLATION` and a crash.\n\n\nRun simple script in attached *.blend file.\n\n", "Applying Subdivision/Multires Modifier De-merges UV's\n### Update\n\nThe issue of custom-data (UV's specifically) becoming detached remains, this problem existed before b88dd3b8e7, it was just less obvious since selection connected near-by UV's. Now UV's are merged by default at a very low threshold when applying modifiers to mitigate the problem. Keeping this report open as it would be good if subdivision-surface didn't separate UV's, replies to this report contain some notes that could lead to a solution. In practice users should not experience problems caused by this behavior, so marking low priority. ~ @campbellbarton.\n\n----\n\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Radeon (TM) RX 480 Graphics ATI Technologies Inc. 4.5.14736 Core Profile Context 20.8.3 27.20.12027.1001\n\nBroken: version: 2.91.0 Alpha \nWorked: working in 2.90 \n\nCaused by b88dd3b8e7\n\nWhen applying a subdivision modifier, certain UV's points disconnect from their neighbors. This does not seem to occur with the F3 menu subdivide operation.\n\n- Open default blender scene.\n- Subdivide cube - 1 level and apply it. Do this step 4x\n- In the UV editor use \"Alt select\" to select face loops at random.\n- Alt click around until a face loop is not fully completed, this will reveal unmerged vertices.\n- Check area that did not complete by moving faces, vertices, edges etc to confirm unmerged UV's.\nThis does not seem to be an issue with the newly added 9a52b30cc0 in my eyes, rather an issue with subdividing.\n[UV Disconnect.blend](UV_Disconnect.blend)\n\n", "Particle instance modifier can crash when create from Children is checked and Normal is not\nOSX 10.11.6, xeon w3680 6core 12thread \n[system-info.txt](system-info.txt)\n2.69.7 5bd8ac9\nalso Nightly 2.79 34fe3f9c069\n\n\nUsing a particle instance modifier with Children checked and Normal not checked and then attempting to change certain parameters of the particle system such as in the velocity or children tabs causes a crash. Enabling Normal in the particle instance modifier prevents this from happening.\nCreate a Plane with a hair type particle system.\nCreate a cube with a particle instance modifier and set the object to the Plane with the particle system.\nCheck the box for Children and uncheck the box for Normal.\nGo to the particle system on the Plane and check Advanced. \nDrag any of the values in Velocity or Children and soon it will crash.\n[Crash_ParticleInstance.zip](Crash_ParticleInstance.zip)", "Render crash when using Python API to modify object Attribute data in frame_change_pre handler\nOperating system: Windows 10 Home\nGraphics card: GTX 1070 8GB\n\nBroken: Blender 2.93.0 84da05a8b806 (2021-06-02 11:21)\nWorked: n/a\n\n\nThis issue was discovered during development of an addon that generates procedural meshes in a `frame_change_pre` handler. When adding attributes to these meshes in the `frame_change_pre` handler, Blender can crash during render on random frames.\n\nThe `Render → Lock Interface` option is enabled as per the note in the [handlers documentation ](bpy.app.handlers.html#note-on-altering-data).\n\nI have created a simplified script that reproduces this crash during render, typically within a minute of starting the render. This script randomly generates triangles and attribute values in a frame change handler:\n\n```lines=5\nimport bpy, random\n\n\ndef frame_change_pre(scene):\n n_triangles = 10000 # number of triangles to generate\n n_attributes = 10 # number of attribute 'layers' to generate\n \n tsize = 0.1\n vertices = []\n triangles = []\n for i in range(n_triangles):\n xshift, yshift, zshift = random.uniform(-3, 3), random.uniform(-3, 3), random.uniform(-3, 3)\n vertices.append((-tsize + xshift, -tsize + yshift, zshift))\n vertices.append((tsize + xshift, -tsize + yshift, zshift))\n vertices.append((0 + xshift, tsize + yshift, zshift))\n triangles.append((3 * i, 3 * i + 1, 3 * i + 2))\n\n object = bpy.data.objects[\"Cube\"]\n object.data.clear_geometry()\n object.data.from_pydata(vertices, [], triangles)\n \n for i in range(n_attributes):\n attribute_name = str(i)\n try:\n object.data.attributes.remove(object.data.attributes.get(attribute_name))\n except:\n pass\n \n attribute = object.data.attributes.new(attribute_name, \"FLOAT\", \"POINT\")\n for value in attribute.data:\n value.value = random.uniform(0, 1)\n \n \nbpy.app.handlers.frame_change_pre.append(frame_change_pre)\n```\n\nIncreasing the number of triangles and/or number of attribute layers seems to increase the frequency of crashes. The crashes do not seem to occur when rendering from the command line.\n\nThis is the error reported after the crash:\n\n```\nError : EXCEPTION_ACCESS_VIOLATION\nAddress : 0x00007FF7736C6490\nModule : blender.exe\nThread : 00000814\n```\n\nAttached is the crash log: [attribute_crash_example.crash.txt](attribute_crash_example.crash.txt)\n\n\nAttached is a .blend file including the script that reproduces this issue.\n\n1. Open the .blend file\n2. Press 'Run Script'\n3. Begin rendering the animation (Blender > Render > Render Animation)\n\n[attribute_crash_example.blend](attribute_crash_example.blend)\n\n-----\nSeems easier to reproduce on a debug build since it hits an assert (although i'm not sure who is at fault there, could be the script, could be blender, @HooglyBoogly will likely know since he worked in this area recently) \n\n1) Make a debug build\n2) Load up the .blend above\n3) Hit the Play button on the Python script\n4) Rather than render, just his play on the timeline\n5) Boom!\n\n```lines=10\nBLI_assert failed: K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\mesh_normals.cc:361, BKE_mesh_vertex_normals_ensure(), at 'CustomData_has_layer(&mesh->vdata, CD_NORMAL) || mesh->totvert == 0'\n\nStack trace:\nblender.exe :0x00007FF783355EA0 bli_windows_system_backtrace_stack_thread K:\\BlenderGit\\blender\\source\\blender\\blenlib\\intern\\system_win32.c:233\nblender.exe :0x00007FF783354F30 BLI_windows_system_backtrace_stack K:\\BlenderGit\\blender\\source\\blender\\blenlib\\intern\\system_win32.c:320\nblender.exe :0x00007FF783354900 BLI_system_backtrace K:\\BlenderGit\\blender\\source\\blender\\blenlib\\intern\\system_win32.c:386\nblender.exe :0x00007FF783359730 _BLI_assert_print_backtrace K:\\BlenderGit\\blender\\source\\blender\\blenlib\\intern\\BLI_assert.c:50\nblender.exe :0x00007FF777848300 BKE_mesh_vertex_normals_ensure K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\mesh_normals.cc:361\nblender.exe :0x00007FF7778467E0 BKE_mesh_calc_normals K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\mesh_normals.cc:475\nblender.exe :0x00007FF77951DC80 ED_mesh_update K:\\BlenderGit\\blender\\source\\blender\\editors\\mesh\\mesh_data.c:1106\nblender.exe :0x00007FF77885B520 Mesh_update_call K:\\BlenderGit\\2022_full_with_tests\\source\\blender\\makesrna\\intern\\rna_mesh_gen.c:7350\nblender.exe :0x00007FF778528720 RNA_function_call K:\\BlenderGit\\blender\\source\\blender\\makesrna\\intern\\rna_access.c:7336\nblender.exe :0x00007FF778A024F0 pyrna_func_call K:\\BlenderGit\\blender\\source\\blender\\python\\intern\\bpy_rna.c:6304\npython310_d.dll :0x00007FFA32E09B40 _PyObject_MakeTpCall E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:215\npython310_d.dll :0x00007FFA330F0690 _PyObject_VectorcallTstate E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Include\\cpython\\abstract.h:112\npython310_d.dll :0x00007FFA330F0C90 call_function E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:5870\npython310_d.dll :0x00007FFA330E37B0 _PyEval_EvalFrameDefault E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:4232\npython310_d.dll :0x00007FFA330F0350 _PyEval_Vector E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:5072\npython310_d.dll :0x00007FFA32E08510 _PyFunction_Vectorcall E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:342\npython310_d.dll :0x00007FFA32E0EAC0 _PyObject_VectorcallTstate E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Include\\cpython\\abstract.h:114\npython310_d.dll :0x00007FFA32E0ED10 method_vectorcall E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\classobject.c:53\npython310_d.dll :0x00007FFA330F0690 _PyObject_VectorcallTstate E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Include\\cpython\\abstract.h:115\npython310_d.dll :0x00007FFA330F0C90 call_function E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:5870\npython310_d.dll :0x00007FFA330E37B0 _PyEval_EvalFrameDefault E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:4183\npython310_d.dll :0x00007FFA330F0350 _PyEval_Vector E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Python\\ceval.c:5072\npython310_d.dll :0x00007FFA32E08510 _PyFunction_Vectorcall E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:342\npython310_d.dll :0x00007FFA32E07FD0 PyVectorcall_Call E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:255\npython310_d.dll :0x00007FFA32E086F0 _PyObject_Call E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:290\npython310_d.dll :0x00007FFA32E07680 PyObject_Call E:\\db23\\build\\S\\VS1564D\\build\\python\\src\\external_python\\Objects\\call.c:318\nblender.exe :0x00007FF778A12F20 bpy_app_generic_callback K:\\BlenderGit\\blender\\source\\blender\\python\\intern\\bpy_app_handlers.c:354\nblender.exe :0x00007FF776ECF220 BKE_callback_exec K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\callbacks.c:51\nblender.exe :0x00007FF776ECF400 BKE_callback_exec_id K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\callbacks.c:67\nblender.exe :0x00007FF7776D5C60 BKE_scene_graph_update_for_newframe_ex K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\scene.c:2648\nblender.exe :0x00007FF7776D5C10 BKE_scene_graph_update_for_newframe K:\\BlenderGit\\blender\\source\\blender\\blenkernel\\intern\\scene.c:2712\nblender.exe :0x00007FF778A46A10 ED_update_for_newframe K:\\BlenderGit\\blender\\source\\blender\\editors\\screen\\screen_edit.c:1729\nblender.exe :0x00007FF778A70B30 screen_animation_step_invoke K:\\BlenderGit\\blender\\source\\blender\\editors\\screen\\screen_ops.c:4754\nblender.exe :0x00007FF777D74A30 wm_operator_invoke K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:1339\nblender.exe :0x00007FF777D6FE10 wm_handler_operator_call K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:2340\nblender.exe :0x00007FF777D72F10 wm_handlers_do_keymap_with_keymap_handler K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:2693\nblender.exe :0x00007FF777D720D0 wm_handlers_do_intern K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:3020\nblender.exe :0x00007FF777D70B70 wm_handlers_do K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:3161\nblender.exe :0x00007FF777D6A440 wm_event_do_handlers K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm_event_system.c:3775\nblender.exe :0x00007FF777D3EB60 WM_main K:\\BlenderGit\\blender\\source\\blender\\windowmanager\\intern\\wm.c:642\nblender.exe :0x00007FF776EB2C60 main K:\\BlenderGit\\blender\\source\\creator\\creator.c:564\nblender.exe :0x00007FF7837A1B00 invoke_main d:\\a01\\_work\\20\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl:79\nblender.exe :0x00007FF7837A18B0 __scrt_common_main_seh d:\\a01\\_work\\20\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl:288\nblender.exe :0x00007FF7837A1890 __scrt_common_main d:\\a01\\_work\\20\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_common.inl:331\nblender.exe :0x00007FF7837A1BC0 mainCRTStartup d:\\a01\\_work\\20\\s\\src\\vctools\\crt\\vcstartup\\src\\startup\\exe_main.cpp:17\nKERNEL32.DLL :0x00007FFAE8F27020 BaseThreadInitThunk\nntdll.dll :0x00007FFAE9062630 RtlUserThreadStart\n\n```\n\n", "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Suspected asset library crash by `json_sax_dom_parser`\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.68\n\nBroken: version: 3.5.1\n\nI dont completely know why but blender crashes very often for me.\nIts doing something no idea what it is doing.\nSorry for being vague, but it seems to mostly happen when i have the viewport set to cycles render and i am trying to modify materials, in the shader editor.\n\n", "\"Updating Objects Flags\" Stalling Blender\n{[F10130762](22_Nautilus_through_City_-_Cycles___Eevee.zip)}Operating system: Linux-5.8.0-53-generic-x86_64-with-debian-bullseye-sid 64 Bits\nGraphics card: GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.73.01\n\nBroken: version: 2.92.0\nWorked: Not working down to 2.83 LTS\n\nThis file will not render the Scene named \"Cycles\" with the Cycles render engine. \n\n\n1. Open attached blend file. Note everything except camera and lights are turned off.\n\n2. Using command line, attempt to render a single frame (say 130).\n\n3. Result = Everything initialises but the render stalls on \"Updating Objects Flags\".\n\n4. Open the file in UI\n\n5. Render image renders a blank image (expected).\n\n6. Turn on Buildings and Landscape, 9 collections from the top, and Nautilus and Submarine Search Lights at the bottom. \n\n7. Render a single frame, or the cycles render layer in the compositor.\n\n8. Result = A single image may render in cycles (followed by the second scene in Eevee that composites).\n\n9. Render animation\n\n10. Result = Blender hangs on \"updating objects flags\" and it does not render, and it is not possible to cancel the render.\n\nIt may be possible to get back rendering by saving the file as a new file. But as soon as rendering an animation as attempted, \"updating objects flags\" hangs Blender either in command line or in UI mode.\n\nNotes: \n\n1. I have tried the same file on multiple machines and multiple Operating Systems with the same result.\n\n2. Quite a number of files have similar kinds of problems. Most commonly, the files will not render in UI mode. Usually though, rendering in command line works. In this case it does not.\n\n3. I have tried every permutation of trouble-shooting I can think of. I have added meshes back in one by one, re-baking ocean modifer, particles, everything. Trying recently, I had to re-bake the particles after which rendering stopped working in any configuration. The particle systems have been baked to a disk cache.\n\n4. Reconstructing from a new blend file is pointless - new blend files are fine and I do not know what is happening to bring about the bug.\n\n5. It may be necessary to re-send the file in its state when it cannot render. Please let me know if anymore information is required.", "Crash on Changing Repeat value to -1 in Corrective Smooth modifier\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 SUPER/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 496.49\n\nBroken: version: 2.93.5\n\nNot Responding when I change the Repeat value to -1\n\n{[F11810617](image.png), size=full}\n\n" ]

[ " Crash: Multi-object edit containing a (multiuser) mesh that is also used by an object in a hidden collection destroys CustomData layer\nOperating system: Windows 10\nGraphics card: Intel® HD Graphics 620 (Kaby Lake GT2)\nProcessor: Intel® Core™ i3-7020U CPU @ 2.30GHz × 4 (64-bit)\nMemory: 8 GB\n\nv 2.80 (Current Stable)\nv 2.81 (8f55794c0e80)\n\nEvery single time I try to triangulate the faces of all my meshes at once, returning to object mode crashes Blender if it's in \"Material Preview\" or \"Rendered\" Eevee shading. I can also do this in \"Solid\" shading and it won't crash after returning, but as soon as I switch shading to \"Material Preview\" shading, it crashes. I've reproduced this several times in the stable 2.80 so I tried my build of 2.81 from four days ago. Still crashes without failure to reproduce.\n\n[Undead Fields Remaster (BUG report for Blender Developers).blend](Undead_Fields_Remaster__BUG_report_for_Blender_Developers_.blend)\n\n - Open that file\n - Click the \"View Object Types\" button in 3D View and hide all but meshes.\n - Select all the objects with the `A` key and switch to edit mode.\n - Press `A` in edit mode to select all the geometry.\n - Press `Ctrl`+`T` to triangulate all the faces.\n - Enter object mode. If you're in \"Rendered\" or \"Material Preview\" shading, it will crash.\n # If you did steps 1-6 in \"Solid\" shading first and it didn't crash, try repeating it all while in \"Rendered\" or \"Material Preview\" shading beforehand.\n\n**Note**\nIf I can/cannot reproduce it in the build from today (13th as of writing) I'll edit this post with my findings. I'm fairly certain that this is just a bug that is so specific that it has not been found/reported yet.\n\nEDIT:\n\nDigging further this can be fixed by:\n- enable the Levels of Detail collection for the viewport\n- OR keep the Levels of Detail disabled in the viewport, but remove tree oak joined, tree pine joined, tree small joined from this collection\n- OR by making tree oak joined, tree pine joined, tree small joined so that they have single user meshes...\n\nminimal repro file based on the report file:\n[repro.blend](repro.blend)\n\nAnalysing this it boils down to:\n- if two objects share a mesh\n- if this mesh uses a material that makes use of UVs\n- if one of the objects is in a collection hidden from the viewport\n- if you are in lookdev/rendered (or go there later)\n- if you enter multi-object editing\n- then back to object mode --> crash\n\nmost minimal file I could come up with:\n[minimal.blend](minimal.blend)" ]

[ "Scaled up images are incorrectly interpolated\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 526.98\n\nBroken: version: 3.5.0 Alpha, branch: Unknown, commit date: Unknown Unknown, hash: `rBUnknown`\n\nScaling up strips will result in sharp edge on left and bottom side. This is caused by discarder when `IMB_TRANSFORM_MODE_CROP_SRC` is used\nAdditionally image is not centered correctly. which is caused by incorrect `uv_start` coordinate\n\nPatch [D13124](D13124) resolves incorrect image coordinate. Also uses `IMB_TRANSFORM_MODE_REGULAR` if no cropping is used, but with cropping image will still have sharp edge issue.\n\n\n\n[#103980.blend](T103980.blend)\n\nIssue visible immediately when file is opened\n", "Viewport: Auto-Depth rotation ignore transparent image reference empties\nOperating system: Windows-10-10.0.17763-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 442.92\n\nBroken: version: 2.90 (sub 0)\n\ncant rotate around partially occluded image objects.\n\n\nopen file, make sure rotate around mouse is enabled in preferences. hover mouse over either side images and hold in the mouse wheel, move the mouse, the view will rotate around the mouse. Now do the same for the middle image, the view no longer rotates around the mouse.\n\n[rotate around mouse bug with image objects that are occluded.blend](rotate_around_mouse_bug_with_image_objects_that_are_occluded.blend) \n\n", "weird unremovable noise which looks like white or semitransparent squares\nWindows GTX NVIDIA 860M Lenovo Y50 laptop 16GB RAM and 2GB VRAM\n\nBroken: (2.77, weird unremovable noise which looks like white or semitransparent squares)\n\n\n**I have horrible flickering noise during my animation and it flickers like a sparks. What causes this noice is alpha images with holes or models of the grids with holes + method of reducing it by using Z and normal passes combined with bilaterural blur (advised by Barteck Skorupa). But i still wanna use it since it still better and only way to properly reduce other types of noice presented in my scene. So this noice looks like SQUARE PIXELS which are flickering during animation. Change of iterations of bilatural blur helps a bit but not completely and not on all frames. \n\nI will give you more info now about this strange things -\n\n\n1- At most frames there is no such noise (because i removed it by using bilateral blur and determinator) but at other frames they are still appear and not despeckle not bilatural blur helps and the weird thing - THEY BECOME MORE VISIBLE AT HIGHER SAMPLES!!!!\n\n\n2 - At some frames this squares are transparent at others - semitransparent or completly white like you can see on the screenshot\n\n\n3 - If i render same image at small samples - like 125 for example - i may not see them at all. But if i change to 150 - they become very visible!!!. \n\n4 - They mostly appear around grids holes models or grids holes alpha images.\n\n\n5- Even if camera not moving they may still appear and look at different frames - differently (white or semitransparent )\n**\n\n\n\n\n\n\n\n\n\n**To reproduce the error you whould probably need to render scene when camera moves and looks directly on grids png images with holes or even models of grids with a holes and also try to use bilatural blur node with determinator + z pass and normal pass combined ro remove noise - since most of the issues appearing when you using this nodes combinations (but its necessarily to use them)**\n\n\nIf you will need a blend file please leave me a comment on [email protected]. Thanks\n", "Regression: Pixels in render result appear blurry during rendering\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.94\n\nBroken: version: 3.5.0 Alpha\nWorked: 3.4.1\n\nPixels in the render result can appear to be stretched into 1x2, 2x1 or 2x2 pixels depending on window resolution. \nThis is only the case while the render is still going; when it finishes the pixels appear as they should.\n\n\n1. Start rendering any scene, most noticeable in scenes with easily visible noise like this one: \n[noisebox.blend](noisebox.blend)\n2. Resize the window while the render is running\n3. Observe pixels changing size\n[blender_2023-02-07_04-11-02.mp4](blender_2023-02-07_04-11-02.mp4)\n", "Holding Ctrl disables Object Lock Modes\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 516.59\n\nBroken: version: 3.2.1\n\n\nObject Lock Mode have no effect when Ctrl is held, for instance, while using Shortest Path Selection as shown in the following video:\n\n[bug-object-lock-mode.mp4](bug-object-lock-mode.mp4)\n\nThis also happens in other cases like selecting Reference and Background images:\n\n[lockobject.mp4](lockobject.mp4)\n\n", "Booleans causing shadow artifacts in solid render mode\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: AMD Radeon RX 5700 XT ATI Technologies Inc. 4.5.14830 Core Profile Context 21.4.1 27.20.21002.112\n\nBroken: version: 3.0.0 Alpha\nWorked: Unknown\n\nI built up a random shape using a bunch of booleans via the box cutter addon.\nThis seems to have introduced artifacts when the shape is viewed from certain angles.\n\nThe artifacts remain when applying all the modifiers. I also tried to reset various edge data like bevel weight and crease without any effect.\nAdding a triangulate modifier however cleans it up, so maybe something about the automatic triangulation is off.\n\nI compared the output with the stable release: It also shows artifacts but they look slightly different, at least with my current configuration.\n\nv.2.93.1\n\n\nv.3.0.0 (see version info above)\n\n\nIn motion from another angle, showing convergence on a point and something that looks like an [even/odd fill rule](Even%E2%80%93odd_rule) of a shape being used to draw the shadow:\n[streaks.mp4](streaks.mp4)\n\n\nI cannot offer concrete steps; my default file also has certain settings already applied.\nHere is the file that may help show the issue: [shadow-streaks.blend](shadow-streaks.blend)\n", "Workbench: Random object color disables transparency in solid view\nOperating system: Windows 10 64bit\nGraphics card: Nvidia RTX 2080ti\n\nBroken: 2.83.1 official\n\n\n\n\nWhen setting the color of the material in the viewport to be semitransparent (Alpha value = 0.1) it works if the color of the viewport objects is set to \"Material\" it works, while if the color is set to \"Random\" the object is not semitransparent anymore.\n\nBlend file to test:\n[bug.blend](bug.blend)", "Incorrect interaction glass with visibility of other objects\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1060 6GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.96\n\nBroken: version: 2.93.3\n\nIncorrect interaction glass with visibility of other objects. I have objects which disappear if I look at them at a certain angle through the glass.\n\nGif of the trouble \n\n\nTest file:\n[#91424.blend](T91424.blend)\n", "Clipping Region in blender v3.2 and up is causing a werid artifact\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: AMD Radeon RX 6600M ATI Technologies Inc. 4.5.0 Core Profile Context 22.10.1.221003\n\nBroken: version: 3.3.1\n\nweird colorful noise when using Clipping region ALT+B\n\nit happens by simply using the clipping region, it happens in both object mode and edit mode\n\n**Diagnosis that I tried** (USEFUL NOTES{[F13703419](Untitled.png)} FOR FIXING THE PROBLEM)\nwhat is causing it?: \n- I noticed that the bug is only on the Viewport Overlays it seems like it's smth related to the graphical smoothing done by \"viewport overlay smooth wires\" from blender preferences -> viewport -> Quality -> Smooth Wires. I can turn of that for Overlay and for Edit mode, but it's just hiding the problem not fixing it, also the selected objects highlight (active element and not active element) are both still showing the same artifact and I don't know how to turn off smoothing for this part.\n\nis it gpu driver problem?\n- it most likely is a conflict between blender and my gpu, I don't think it's a driver problem tho because I tried the latest two gpu drivers available from amd, and both of them are showing the same problem. I tried running blender using my integrated gpu and tried my RX6600M gpu and in both cases it happend, I don't know if the smoothing is done on the gpu level or cpu level tho so it could be not related to which gpu I use\n\nWhat versions of blender does it happen in?\n- I mainly use blender from steam with auto updates, I noticed the problem when I was using v3.3.1 then I tried few older versions by manually installing them from the offical blender website, the bug doesn't show up in v3.1.2 and v2.79b however it's in showing up in v3.2.0 and v3.3.0 and v3.3.1 -not steam version- I didn't test more versions tho\n\nP.S. my device -Legion 5- has a mux switch, when I did the bug report the mux switch is turned off which means the integrated GPU is disconnected thus it's not showing up in the bug report, however when I tried to check the bug \n\n\n\n\n\n\n\n[untitled.blend](untitled.blend)", "Video Editing (outline on the border)\nOperating system: Windows-10 64 Bits\nGraphics card: GeForce GTX 750 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.89\n\nBroken: version: 2.93.0 Alpha\n\nSome images with an alpha channel have grey outline when scaled up and different bicubic or bilinear interpolation is used.\n\nIssue is visible immediately after opening .blend file\n[#86861.zip](T86861.zip)", "Snapping issue while rotating and 'Align Rotation to Target' is ON - can't set correct origon rotation for simple object \nOperating system: Windows 10\n\n\nBroken 2.81 new feature of origon \n\nsnaping rotation origon bug - can't set correct origon rotation for simple object \n\nrotation snaping of origon is tha same important as moveing origon to spec place\n \nI need to specyfy Z direction also x and y direction. To working with this propably I need to have 2 directions snaping z and x or z and y to work with this.\nso for example Z with snaping and then hold shift seting Y snaping... \n\nThis will help with orientation objects in space by orientation of another objects like aligment rotation object to another object ect. \n\nThanks for great work.\n\n\n[origon snaping rotation video.mp4](origon_snaping_rotation_video.mp4)", "3D View Selection Occlusion (Select Through) \nCreating this design task as there are unresolved issues with [D6322: Mesh: Select through](D6322) which are more design issues.\n\nThis is a place holder, some topics to address:\n\n- Where this option is stored, accessed?\n- Which selection tools/operators should use this?\n- Which modes this should be supported in (edit, pose, grease pencil... etc)\n- How to display this option when in wire-frame display (when the user will always select-through)\n- How to access this from both active-tools & key-bindings?\n- How to make it obvious selecting-trough is enabled *(draw style, cursor - this might help avoid confusion)*.\n- Should this be supported in object mode too?(the reverse since objects already select-through), IIRC we discussed having this option for box select too.\n\n\n### Related Topics\n\nNote that there are topics being raised which relate to select-through, but might be best treated as separate topics since we could change behavior here with or without select-through.\n\n- Current behavior of X-ray.\n- Current usability toggling X-ray (Z-drag or Alt-Z).\n- Selecting faces by their centers.\n- Lasso select using object centers #64281 (Option to lasso-select using geometry instead of object centers.).\n", "Regression: broken normal shading, sharp edges appear on smooth surfaces.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 375.70\n\nBroken: 2.93, version: 3.0.0 Alpha\nWorked: 2.83.9, 2,92\n\nCaused by 1c22b551d0\nTo fix this, 5c4d24e1fd got reverted\nProblem still persists though...\n\n\nIn newer versions normal mapping looks broken, sharp edges appear.\nEevee and Cycles have same bug.\n\nupdated: [untitled.blend](untitled.blend)\n\ndifference in shading\n**2.93 on top, 2.8 down below:**\n\n\nThis effect appears even with very small angle of clean surface (without normal maps)\n\n\n\n<video src=\"2021-05-18_16-17-05.mp4\" title=\"F10122604/2021-05-18_16-17-05.mp4\" controls></video>", "Texture Node Editor Using nodes causes weird issues in sculpt mode.\nOperating system: Linux-5.15.0-48-generic-x86_64-with-glibc2.31 64 Bits\nGraphics card: NVIDIA GeForce RTX 2060/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 515.76\n\nBroken: version: 3.3.1\nWorked: perhaps never? (can redo in 2.83)\n\nAdding an image from the nodes of the texture editor for example, a sculpting brush, causes really weird issues. Disabling nodes and using the texture properties does not have this issue.\n\n- Check the blend file. Try sculpting, you will see weird noise being displaced with the texture.\n- Now disable the nodes and set the texture type from the texture properties to the same image. Sculpting again you see the issue is not happening.\n[bug.blend](bug.blend)\n\n\n\n", "Boolean modifier artifacts\nOperating system: macOS-12.4-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.8.54\n\nBroken: version: 3.3.0 Alpha\n\nDifference on Brush Boolean is error code and lot bug.\nSee youtube tell you.JaCQrMN0quk\n\n**Steps to reproduce**\n- Open .blend file\n- Move target object over the object which has boolean modifier on it\n\n[3D print baby tall chair belt.blend](3D_print_baby_tall_chair_belt.blend)\n" ]

[ "Orthographic view is blurry - rB36814ddc94b5 regression\n**System Information:**\nOperating system: Linux-5.10.0-3-amd64-x86_64 64 Bits\nGraphics card: GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.39\n\n**Blender Version:**\nBroken version: 2.93.0 Alpha\nWorked: Prior to 2.93 36814ddc94\n\n**Short description of error:**\nWhen in orthographic mode, the 3D viewport appears \"blurrier\" than usual when compared to non-orthographic mode or older versions of Blender. Here's a few screenshots:\n\n|||||\n| -- | -- | -- | -- |\n|Prior to 36814ddc94| Current Master|Prior to 36814ddc94 (enlarged)|Current Master (enlarged)|\n\n**Exact steps for others to reproduce the error:**\n1. Open a scene with a model that's at least semi-complex, I used a scene with Suzanne in it.\n2. Enter orthographic mode `View -> Perspective/Orthographic`, or `Numpad 5`.\n3. Observe how the scene appears blurry when compared to older versions. Changing the anti-aliasing option in `User Preferences -> Viewport -> Viewport Anti-aliasing` can either make the effect worse or better. Decreasing anti-aliasing sample count makes the issue better while increasing the anti-aliasing sample count makes it worse. This is inline with the issue being related to a change in the TAA used by the workbench render engine introduced in 36814ddc94.\n\nNote: I am using a 1920x1080 monitor. I assume this issue is less noticeable on higher resolution monitors.\n", "Workbench. Viewport anti-aliasing (AA) on solid surface became blurry\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n\nBroken: version: 2.93.0 Alpha\nWorked: 2.93 few days ago.\n\nHere is 5 samples AA, and 32 samples go even worse\n{[F9815147](AA.png) size = full}\nIt appears only with surface, wireframe looks as good as before.\n\nThere is no blurry effect, when I move viewport. But it appears, when I stop and release mouse.\nYou can just look at shape and click {key MMB} to see the difference\n\n", "Blurry Workbench AA\n\nTested on 2 systems\n\nOperating system: Linux-5.8.0-7642-generic-x86_64-with-glibc2.32 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 460.39\n[system-info.txt](system-info.txt)\n\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce GTX 1080 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 461.40\n[system-info.txt](system-info.txt)\n\nBroken: version: 2.93.0 Alpha\n\nAA on Workbench is very blurry\n\nThis only happens for me on the workbench engine. I had to switch to single pass AA to get rid of the blurred effect\n\n\n\n" ]

[ "Crashes when Switching to Material Viewport / Rendered Viewport / Rendering\nOperating system: Windows 10\nGraphics card: 2 Nvidia GTX 1660 Super's\n\n[blender_system_info.txt](blender_system_info.txt)\n\nBroken: blender 2.8 / 3.4\n\n\nBlender runs completely fine when editing and building a model but once rendering is involved, even in preview, it immediately closes with no explanation. This crash will also happen with the default cube. This had not happened before and have tried Updating blender, reinstalling, simplifying, etc but no change and can reproduce the issue on my computer. i can have it in material mode if i have 0 assets in the scene but once i add something it crashes, i am suspecting a bad pathing issue as i had recently refreshed my computer but do not know how to fix it or give blender a hard reset.\n\n{[F14170509](blender.crash.txt)}[blender_debug_output.txt](https://archive.blender.org/developer/F14170500/blender_debug_output.txt)\nOpen Blender, have any Physical asset inside the scene, switch to rendered / Material Viewport, Crash\n\nI have included the project I was working on and the default cube as they create the same issue.\nPlease note that I have installed the most recent Blender program onto my laptop and opened the files attached and did not have any crash happen so it seems the issue is solely between my computer and the blender program.\n\n[Default Cube.blend](Default_Cube.blend) [Skell Rig base.blend](Skell_Rig_base.blend)", "Blender crash after maximizing/unmaximizing an editor (debug build only)\n\nBroken: master - debug build\nWorks: master - release build\nOS: Tested on both Windows and Linux\n\nWith a debug build, Blender crashes after maximizing an editor + moving mouse + unmaximizing.\nIt does not seem to be reproducible with a release build.\n\n\n- Open Blender default scene\n- Maximize the 3D viewport with CTRL+SPACE\n- Move the mouse\n- Unmaximize with CTRL+SPACE\n\n**Debug info**\nThe crash happens in `wm_event_do_handlers` when calling `wm_region_tag_draw_on_gizmo_delay_refresh_for_tweak(wmWindow *win, bScreen *screen)`.\nIn that case, `screen` points to invalid (freed?) memory.", "Sculpt mode: hide masked do not work when modifier enabled, but hidden part works with modifier.\nOperating system: Windows-8.1-6.3.9600-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 660 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 3.0.0 Alpha\n\n\nAs title said. Hiding works with modifier but not work simultaneously as «hide masked»\nFor example, edit mode work well with mirror and hiding, hiding by faceset works well too...\n\n**Steps to Reproduce**\n- Open .blend file\n- Paint with mask brush\n- {nav Sculpt > Hide Masked} \n\nTest File:\n[#92372.blend](T92372.blend)\n[2021-10-20_18-49-00.mp4](2021-10-20_18-49-00.mp4)\n", "Crash going to material preview or rendered shading mode Blender 2.83.0\nMacBook Pro 15 inch Early 2013\n2.7 GHz Intel Core i&\n16 GB 1600 Mhz DDR3\n|Operating system:|MacOS 10.12.5 - @Tramfabriek\n| -- | -- |\n||Mac OS X 10.12.6 (16G2136) - @johan-eds\n|| macOS 10.12.6 - @UmbertoOldani\n\n|Graphics card:|NVIDIA GeForce GT 650M 1024 MB + Intel HD Graphics 4000 1536 MB - @Tramfabriek\n| -- | -- |\n||NVIDIA GeForce GTX 775M, PCIe, 2048 MB - @johan-eds\n|| NVIDIA GeForce GTX 680MX 2048 MB - @UmbertoOldani\n\nBisected to find the bad commit: 804e90b42d\n\nCrash when clicking on Viewport Shading >Material Preview or >Rendered\nThis did happen already in the 2.83 betas, but none of the 2.82 and earlier version had this problem on this computer running this same OS at least since Blender 2.79.\n\n**System crash report 1** (crash in `UI_draw_roundbox_aa` -> `GPU_batch_draw `. Probably the material render icon):\n```lines=10\nThread 0 Crashed:: Dispatch queue: com.apple.main-thread\n0 libsystem_kernel.dylib \t0x00007fff909add42 __pthread_kill + 10\n1 libsystem_pthread.dylib \t0x00007fff90a9b457 pthread_kill + 90\n2 libsystem_c.dylib \t0x00007fff909134bb __abort + 140\n3 libsystem_c.dylib \t0x00007fff9091342f abort + 144\n4 libGPUSupportMercury.dylib \t0x00007fff88cdbfbf gpusGenerateCrashLog + 158\n5 com.apple.GeForceGLDriver \t0x00007fff76297390 0x7fff7607a000 + 2216848\n6 libGPUSupportMercury.dylib \t0x00007fff88cdd524 gpusSubmitDataBuffers + 560\n7 com.apple.GeForceGLDriver \t0x00007fff76385867 0x7fff7607a000 + 3192935\n8 com.apple.GeForceGLDriver \t0x00007fff763554fd 0x7fff7607a000 + 2995453\n9 com.apple.GeForceGLDriver \t0x00007fff76282a2b 0x7fff7607a000 + 2132523\n10 com.apple.GeForceGLDriver \t0x00007fff7626ee9a 0x7fff7607a000 + 2051738\n11 com.apple.GeForceGLDriver \t0x00007fff76289654 0x7fff7607a000 + 2160212\n12 com.apple.GeForceGLDriver \t0x00007fff7626af97 0x7fff7607a000 + 2035607\n13 com.apple.GeForceGLDriver \t0x00007fff763851ad gldUpdateDispatch + 1041\n14 GLEngine \t0x00007fff7ff1d32a gleDoDrawDispatchCoreGL3 + 520\n15 GLEngine \t0x00007fff7fecebd9 gleDrawArraysOrElements_Entries_Body + 128\n16 GLEngine \t0x00007fff7fecaab4 glDrawArraysInstanced_GL3Exec + 220\n17 org.blenderfoundation.blender \t0x0000000109614f2c GPU_batch_draw_advanced + 428\n18 org.blenderfoundation.blender \t0x0000000109614d6c GPU_batch_draw + 124\n19 org.blenderfoundation.blender \t0x0000000105218c51 UI_draw_roundbox_aa + 353\n20 org.blenderfoundation.blender \t0x0000000104f3d96c ED_region_do_draw + 1372\n21 org.blenderfoundation.blender \t0x0000000104b6d49c wm_draw_update + 2092\n22 org.blenderfoundation.blender \t0x0000000104b6b010 WM_main + 48\n23 org.blenderfoundation.blender \t0x0000000104827762 main + 962\n24 libdyld.dylib \t0x00007fff9087f235 start + 1\n\n\nThread 0 crashed with X86 Thread State (64-bit):\n rax: 0x0000000000000000 rbx: 0x0000000000000006 rcx: 0x00007fff5b521f78 rdx: 0x0000000000000000\n rdi: 0x0000000000000307 rsi: 0x0000000000000006 rbp: 0x00007fff5b521fa0 rsp: 0x00007fff5b521f78\n r8: 0x0000000000000000 r9: 0x0000000000989680 r10: 0x0000000008000000 r11: 0x0000000000000206\n r12: 0x000000011ab85cf8 r13: 0x0000000000000000 r14: 0x00007fff998033c0 r15: 0x000000011ab85000\n rip: 0x00007fff909add42 rfl: 0x0000000000000206 cr2: 0x00007fff960d0b58\n \nLogical CPU: 0\nError Code: 0x02000148\nTrap Number: 133\n```\n\n\n**System crash report 2** (crash in `blf_batch_draw ` -> `GPU_batch_draw `):\n```lines=10\nThread 0 Crashed:: Dispatch queue: com.apple.main-thread\n0 libsystem_kernel.dylib \t0x00007fff8b1dfd42 __pthread_kill + 10\n1 libsystem_pthread.dylib \t0x00007fff8b2cd457 pthread_kill + 90\n2 libsystem_c.dylib \t0x00007fff8b1454bb __abort + 140\n3 libsystem_c.dylib \t0x00007fff8b14542f abort + 144\n4 libGPUSupportMercury.dylib \t0x00007fff83587fbf gpusGenerateCrashLog + 158\n5 com.apple.GeForceGLDriver \t0x00007fff71040390 0x7fff70e23000 + 2216848\n6 libGPUSupportMercury.dylib \t0x00007fff83589524 gpusSubmitDataBuffers + 560\n7 com.apple.GeForceGLDriver \t0x00007fff7112e867 0x7fff70e23000 + 3192935\n8 com.apple.GeForceGLDriver \t0x00007fff710fe4fd 0x7fff70e23000 + 2995453\n9 com.apple.GeForceGLDriver \t0x00007fff7102ba2b 0x7fff70e23000 + 2132523\n10 com.apple.GeForceGLDriver \t0x00007fff71017f8e 0x7fff70e23000 + 2051982\n11 com.apple.GeForceGLDriver \t0x00007fff7100b6bc 0x7fff70e23000 + 2000572\n12 com.apple.GeForceGLDriver \t0x00007fff7100bec2 0x7fff70e23000 + 2002626\n13 com.apple.GeForceGLDriver \t0x00007fff7101113f 0x7fff70e23000 + 2023743\n14 com.apple.GeForceGLDriver \t0x00007fff7143f25e 0x7fff70e23000 + 6406750\n15 com.apple.GeForceGLDriver \t0x00007fff7143f4e6 0x7fff70e23000 + 6407398\n16 com.apple.GeForceGLDriver \t0x00007fff7143f5d7 0x7fff70e23000 + 6407639\n17 com.apple.GeForceGLDriver \t0x00007fff71013d3d 0x7fff70e23000 + 2035005\n18 com.apple.GeForceGLDriver \t0x00007fff7112e1ad gldUpdateDispatch + 1041\n19 GLEngine \t0x000000011925232a gleDoDrawDispatchCoreGL3 + 520\n20 GLEngine \t0x0000000119203bd9 gleDrawArraysOrElements_Entries_Body + 128\n21 GLEngine \t0x00000001191ffab4 glDrawArraysInstanced_GL3Exec + 220\n22 org.blenderfoundation.blender \t0x000000010b288f2c GPU_batch_draw_advanced + 428\n23 org.blenderfoundation.blender \t0x000000010b288d6c GPU_batch_draw + 124\n24 org.blenderfoundation.blender \t0x000000010b283d82 blf_batch_draw + 418\n25 org.blenderfoundation.blender \t0x000000010b282359 BLF_batch_draw_end + 9\n26 org.blenderfoundation.blender \t0x0000000106fc2163 view3d_draw_region_info + 8515\n27 org.blenderfoundation.blender \t0x0000000106936dbd DRW_draw_callbacks_post_scene + 445\n28 org.blenderfoundation.blender \t0x00000001069376cd DRW_draw_render_loop_ex + 1757\n29 org.blenderfoundation.blender \t0x0000000106fc2668 view3d_main_region_draw + 136\n30 org.blenderfoundation.blender \t0x0000000106bb157e ED_region_do_draw + 366\n31 org.blenderfoundation.blender \t0x00000001067e149c wm_draw_update + 2092\n32 org.blenderfoundation.blender \t0x00000001067df010 WM_main + 48\n33 org.blenderfoundation.blender \t0x000000010649b762 main + 962\n34 libdyld.dylib \t0x00007fff8b0b1235 start + 1\n\nThread 0 crashed with X86 Thread State (64-bit):\n rax: 0x0000000000000000 rbx: 0x0000000000000006 rcx: 0x00007fff598ab4b8 rdx: 0x0000000000000000\n rdi: 0x0000000000000307 rsi: 0x0000000000000006 rbp: 0x00007fff598ab4e0 rsp: 0x00007fff598ab4b8\n r8: 0x0000000000000000 r9: 0x0000000000989680 r10: 0x0000000008000000 r11: 0x0000000000000206\n r12: 0x0000000120bebcf8 r13: 0x0000000000000000 r14: 0x00007fff93fe13c0 r15: 0x0000000120beb000\n rip: 0x00007fff8b1dfd42 rfl: 0x0000000000000206 cr2: 0x00007fff8b1c727d\n \nLogical CPU: 0\nError Code: 0x02000148\nTrap Number: 133\n```\n\n1. Open Blender\n2. Click to remove start up information panel\n3. Click on Viewport Shading >Material Preview or >Rendered\n\nResult:\nCrash", "inset and bevel edits crashing blender 3.x\nThreadripper 1950x 128 gb ram 2tb SSD (home)\nNVIDA 3080 updated to latest drivers\n\nIntel i7 64 gb ram 2 tb SSD (work)\nAMD 6900xt updated to latest drivers\n\n\nHappens on all 3.x versions of blender.\n\nIssue happens on both my home PC (the thread ripper) and my work PC.\n\nStart editing default cube add bevels, insets place some cuts... crashes within a couple of minutes. Happens in all display modes.. mesh, solid, renders. Happens with a smooth display mesh and hard edges. Super frustrating need to save just before doing any of these procedures on any modeling task. Always update video drivers which probably cant be the issue because two different families of video cards in two different machines have the same issue. Been doing this ever since I updated to Blender 3. I assume it's a plugin I installed because I can't imagine why this would happen on two completely different computers. Is there a dump file or something I can send? Here is a video of a typical crash. No dialog box... just blender vanishes in the the middle of editing.\n\n", "Crash Python API - GPUOffScreen.draw_view3d() crashes blender with draw_type = POST_VIEW but works with draw_type = POST_PIXEL\nOperating system: Windows-10-10.0.22000-SP0 64 Bits\nGraphics card: GeForce RTX 3070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 457.51\n\nBroken: version: 3.3.1\n\nWhen using [GPUOffScreen.draw_view3d() ]] in a [[ https:*docs.blender.org/api/current/bpy.types.Space.html?highlight=post_view#bpy.types.Space.draw_handler_add | draw_handler ]], blender crashes when running the [[ gpu.shader.html?highlight=gpu+shader+from_builtin#built-in-shaders | 3D_IMAGE builtin shader ](https:*docs.blender.org/api/current/gpu.types.html?highlight=draw_view3d#gpu.types.GPUOffScreen.draw_view3d) if the draw_type was set to POST_VIEW but works if the draw_type was set to POST_PIXEL.\n\n\n\n - Open CrashScriptDrawView3D.blend\n - Run the script Crash3D and blender will crash\n\n\n\n\n\n - Open CrashScriptDrawView3D.blend\n - Run the script NoCrash2D, and it will draw the camera view in the viewport without crashing\n\nAttachments:\n\n[CrashScriptDrawView3D.blend](CrashScriptDrawView3D.blend)\n[blender.crash.txt](blender.crash.txt)\n\n\n\n\n\n\n", "Line art modifier clips with object on (MBP 16\")\nOperating system: macOS Big Sur\nGraphics card: 5500M \n\nBroken: 2.93.0 LTS\nWorked: -\n\nLine art modifier (grease pencil) seems to clip into / z-fighting the object. \n\n\nJust add a new line art modifier, possibly a mac problem?\n\nTest File:\n[line_sample.blend](line_sample.blend)", "Blender crashes with Noise Texture Node + Hair Particle System + Eevee\nOperating system: Darwin-19.6.0-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon RX 480 OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.10.18\n\nBroken: 8dc95f61f3\nBroken: version: 2.92.0\nBroken: version: 2.91.2\nWorked: version: 2.83.12\n\nBlender 2.92 with AMD crashes with **Noise Texture Node**and **Hair Particle System** with render engine Eevee.\nWith Blender 2.91.2 with AMD it doesn't crash but colors are incorrect.\n\nWith 8dc95f61f3 with Intel GPU, there's no crash but colors are incorrect in the same way as in AMD - 2.91.2.\n\n1. Open attached .blend file\n2. Viewport shading to: Render \n3. Play animation by tapping space bar\n\n[particle_loop.blend](particle_loop.blend)\n[blender_2_92_crash.txt](blender_2_92_crash.txt)\n\n**Blender 2.92**\n[blender_2_92.mp4](blender_2_92.mp4)\n\n**Blender 2.91**\n[blender_2_91_2.mp4](blender_2_91_2.mp4)\n\n**Blender 2.83.12**\n[blender_2_83_12.mp4](blender_2_83_12.mp4)", "ops.paint.image_paint() crashes blender with brush texture\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2070 with Max-Q Design/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 471.41\n\nBroken: version: 2.93.1\nWorked: unknown\n\n\nWhen using bpy.ops.paint.image_paint() in a addon and use a brush texture, blender crashes on the first stroke with a EXCEPTION_ACCESS_VIOLATION\n\n\n\n1. load the paintstroke addon \n```\n {F10253953}\n```\n\n2. load the blend file\n[paint_crash.blend](paint_crash.blend) \n3. go into texture paint mode\n3. make sure the brush texture is assigned\n\n4. paint on the object with (LEFTMOUSE)\n5. crash\n\n\n**addon code:**\n\n```\nbl_info = {\n 'name': 'paintstroke',\n 'description': 'Paint Material image channels at once',\n 'author': 'Daniel Grauer',\n 'version': (2, 0, 0),\n 'blender': (2, 92, 0),\n 'location': \"View3D > Sidebar > Edit Tab\",\n 'category': 'Image Paint',\n 'wiki_url': 'blender_Material_Brush',\n}\n\nimport bpy\nfrom bpy.utils import register_class, unregister_class\nfrom bpy.types import Operator, Brush\n\n \nclass PaintStroke(Operator):\n '''paintstroke'''\n bl_idname = \"paint.paintstroke\"\n bl_label = \"paintstroke\"\n bl_options = {'REGISTER', 'UNDO'}\n \n \n @classmethod\n def poll(cls, context):\n return bpy.ops.paint.image_paint.poll()\n \n def modal(self, context, event): \n #print(\"MB keybinding test modal\")\n if event.type in {'MOUSEMOVE'}: \n brushstroke = {\n \"name\": \"defaultStroke\",\n \"pen_flip\": False,\n \"is_start\": False,\n \"location\": (0,0,0),\n \"mouse\": (event.mouse_region_x, event.mouse_region_y),\n \"pressure\": 1,\n \"size\": bpy.context.tool_settings.unified_paint_settings.size,\n \"time\": 0,\n \"mouse_event\": (0.0, 0.0),\n \"x_tilt\": 0,\n \"y_tilt\": 0,\n }\n stroke = [brushstroke] \n bpy.ops.paint.image_paint(stroke=stroke, mode='NORMAL') \n\n elif event.value in {'RELEASE'} or event.type in {'ESC'}:\n return {'FINISHED'} \n return {'PASS_THROUGH'}\n \n\n def invoke(self, context, event):\n if event.type in {'LEFTMOUSE'}: \n #self.paint(event) \n context.window_manager.modal_handler_add(self)\n return {'RUNNING_MODAL'}\n \n\nclasses = (\n PaintStroke, \n )\n\n\naddon_keymaps = []\n\ndef register(): \n [register_class(c) for c in classes]\n \n # Keymapping\n kc = bpy.context.window_manager.keyconfigs.addon\n km = kc.keymaps.new(name=\"Image Paint\", space_type='EMPTY', modal=False)\n kmi = km.keymap_items.new(\"paint.paintstroke\", type='LEFTMOUSE', value='PRESS', shift=False, ctrl=False , alt=False)\n kmi.active = True\n addon_keymaps.append((km, kmi))\n \n\ndef unregister(): \n # remove keymaps\n for km, kmi in addon_keymaps:\n km.keymap_items.remove(kmi)\n addon_keymaps.clear() \n \n [unregister_class(c) for c in classes]\n \n\nif __name__ == \"__main__\":\n register() \n```\n", "Blender freezes when using Texture paint or Weight Paint\nOperating system: Windows 10 Insider build19592\nGraphics card: GTX 1660Ti Max Q 6GB\nProcessor: i7-9750H\nRAM: 32GB DDR4-2666Mhz\n\nBroken: version: 2.83 (sub 11)\nWorked: Never (2.79, 2.8+)\n\nBlender freezes or stops responding when using Texture Paint or Weight Paint. Tried game ready and Studio drivers, still the same issue.\nMoreover, RAM consumption spikes up to more than 20GB. This happens when the brush size is increased and starts painting. \n\nAssigned GPU in Preferences>System>CUDA, but the GPU is been barely used.\n\n[Blender.rar](Blender.rar)", "Crash when scrubbing the timeline when using the line art modifier\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.41\n\nBroken: version: 3.5.0\nWorked: 3.4.1\n\nWhen scrubbing the timeline, blender will crash. This happens frequently when I have a project of sufficient complexity that uses the line art modifier. The provided blend file has been simplified as much as possible while still reproducing the crashing behavior.\n\n1. Open line_art_crash.blend\n2. Scrub the timeline. Jumping between frames makes it less likely to crash, but holding onto the playhead and slowly moving should have it crash in a few seconds.\n\n", "Crash with lineArt modifier and Armature\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: GeForce RTX 2080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.71\n\nBroken: version: 3.0.0 Alpha\nExperienced with 2.93.2\n\nHaving a line art object in the scene applied to an object 'A' when there is an armature on object 'B' makes blender very unstable. A simple scroll in the timeline will crash instantly.\n\n\nIn the attached file, what I did : \n- Create a cube, subdivided along Y, add a simple armature, keyframe it with a simple motion.\n- Duplicate the cube, apply the armature.\n- Select the second cube, shift-a -> Object Line Art\nFrom there : \n- Scroll the timeline... Crash.\n\nIf I delete the animated cube, scrolling works fine. Disabling it in viewport as well.\n\n\nI also tried adding a simple cube, instead of duplicating the initial \"armatured\" cube, and adding the Object Line Art to it, but the same crash occurs.\n\n[LineArtCrash.blend](LineArtCrash.blend)\n\n", "Grease Pencil: Line Art › Light Contour causes crash\nOperating system: macOS-11.6.5-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon R9 M295X OpenGL Engine ATI Technologies Inc. 4.1 ATI-4.6.21\n\nBroken: version: 3.6.0 Alpha\nAlso reproduced with 3.5 and 3.4.1\n\nWhile using the Grease Pencil Line Art modifier, enabling the Light Contour feature causes a crash. The problem does not appear with a very simple scene, so a certain amount of geometry may be needed, hence the 6mb sample file. While trying to reproduce with smaller scenes, I sometimes had to toggle the Light Contour checkbox on and off a few times to trigger the crash, and the number of toggles actually varied from one launch to the next.\n\n1. Open provided file, GP crash.blend\n2. Select Grease Pencil object and expand its Line Art modifier\n3. Click Light Contour checkbox to enable it\n\nThis reliably causes a crash on my Intel Mac\n\n", "Adding new hook while in Edit mode breaks when trying to undo/redo the operator\nWin7 (32bit)\n\nBroken: 2.78a\n\n\n- Add Path from Curve menu, add new Armature\n- Assign Bone to control point as Hook\n- Blank Hook modifier appears along with real one\n\nScreencast: [screencast0001-0250.mp4](screencast0001-0250.mp4)\n", "Crash while switching to edit mode with EEVEE + Geometry Nodes\nOperating system: macOS-10.15.7-x86_64-i386-64bit 64 Bits\nGraphics card: NVIDIA GeForce GT 650M OpenGL Engine NVIDIA Corporation 4.1 NVIDIA-14.0.32 355.11.11.10.10.143\n\nBroken: version: 3.0.0\n\nWhen switching to edit mode in while the viewport's set to render it crashes (see attached video)\n[blender crash 2021-12-17 21-54-00.mp4](blender_crash_2021-12-17_21-54-00.mp4)\n\n1 load the attached file\n2 switch to rendered view in the viewport\n3 switch to edit vertex mode\n4 that's a crash!\n\n[transitions_001_blender3.blend](transitions_001_blender3.blend)" ]

[ "Adding masks layers crashes blender \nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1060 3GB/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.64\n\n\nBroken: version: 2.80 (sub 74)\nWorked: (optional)\n\n\nSeems that adding a mask layer into a mask in the image editor crashes blender \n\n \nOpen new scene\nchange to compositing work space\nchange editor to image editor\nchange to mask mode\ncreate a mask \nadd a mask layer in the N panel\nclick left click over the editor\nCrash\n" ]

[ "Crash/Assert with two modifiers that require normals data [e.g. DataTransfer]\nOperating system: Windows-10-10.0.16299-SP0 64 Bits\nGraphics card: GeForce MX150/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 441.87\n\nBroken: version: 2.90.0\n\nCrash/Assert with two modifiers that require custom normals data [e.g. DataTransfer]\n\nIn a Debug build:\n\n[#84664.blend](T84664.blend)\n- Open File\n- add another modifier that requires normal data (e.g. Displace), or just duplicate the existing modifier\n\n[un2.blend](un2.blend)\n- Try to go to edit mode of \"Vert.011\"\n- Blender will crash.\n\n", "Rigid Body Physics forces no longer restricted to specific layers/collections\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Plus Graphics Intel 4.5.0 - Build 26.20.100.7641\n\nBroken: version: 2.93.0 Alpha\n\nPrior to Blender version 2.8, forces would only affect rigid body objects within the same layers. With later versions the scene is managed via Collections but despite forces being able to be placed in any number of collections, all forces seem to affect rigid bodies in all collections.\n\nNote that saving a file with mulitple forces in different layers in Blender version 2.79b and loading into 2.93 Alpha does seem to retain the layer structure by creating an additional collection for the objects in different layers but the forces no longer work the same.\n\nAdd Rigid Body Physics to the default cube and duplicate it to a second cube. Move the second cube into a different collection. Switch off Scene Gravity. Create two forces that would affect the two cubes in different ways (eg, place one above one cube and the other below the other cube) and move the forces to the respective collection relating to its associated cube. This should result in each force affecting only one of the cubes (since they are in different collections). However, running the simulation results in both cubes being affected by both forces - rather than just the force in its own collection. \n\n", "Support Scenes Overrides\nCurrently Scenes liboverrides are not really working, changes to its master collection will be lost, resync of scene itself is not possible when the override is the active scene, etc.\n\nSee e.g. #89963 (crash when library override is applied to an object)", "Alpha hashed material produces shifting artifacts when animated in EEVEE\nWhen an object with material that has alpha hashed transparency is animated in EEVEE, the noise pattern may look either unstable or shifting as shown below: \n[alpha-hashed-transparency-problem.mp4](alpha-hashed-transparency-problem.mp4)\n\n\n\nYou can observe that the artifacts on top staying in place rather than rotating with the cube while those on the sides keep flickering, thus producing shifting and unstable appearance.\n\nThe symptom is also noticeable when changing the camera position in rendered viewport mode in Blender. And it's particularly problematic in UPBGE (a fork of the old BGE based on the latest Blender source) because a game engine must animate objects in realtime with relatively small number of samples per frame.\n\nUpon an examination from the developers of UPBGE, it is found that the problem is due to the fact that the current implementation of alpha hashed transparency uses the world coordinate instead of the object space, as it is often recommended (e.g. [2017 ACM SIGGRAPH Symposium on Interactive 3D Graphics and Games](p4TYf5DDpbQ)).\n\nIt appears that a simple change from the world to object coordinate fixes the problem as shown in this [test commit (see line 80) ](e0a8638d0b255f2fc6ce778d10dd0d12a467b4d3#diff-2b8bf4f9e10009510d02c14e3659ee032c73f829de5453d6abcdf6a3f884fdfdR80), which produces the following result:\n[alpha-hashed-problem-solved.mp4](alpha-hashed-problem-solved.mp4)\n\nIt's my understanding that the [original author ]() decided to use the world matrix to avoid the difficulties with certain corner cases, like when the object is scaled by an extreme factor. However, while such a problem may be circumvented relatively easily (e.g. discard the scale information from the matrix) such a problem described above may pose more substantial issues with making an animation or rendering in realtime.\n\nSo, I hope this problem and the suggested fix may be evaluated again since we have more people interested in using Blender for such purposes now, thanks to the gaining popularity of EEVEE.", "FADE INNACTIVE GEOMETRY MODE doesn't work properly with COLLECTION INSTANCES\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.13587 Core Profile Context 20.4.2 26.20.15029.27017\n\nBroken: version: 2.91.0 Alpha\nBroken: version: 2.91.0 Alpha\n\nI really like this mode, but it feels likeit not completely done. It's not work with collection instances at all.\nHere is was my expectations and what it looks like currently: \n\nAnd even instances from other scenes ignore this mode too. Will be cool if it(instances) will work in stream ideology of this mode\n\n\nOpen this file [FADE INNACTIVE GEOMETRY DOESN'T WORK PROPERLY WITH INSTANCES.blend](FADE_INNACTIVE_GEOMETRY_DOESN_T_WORK_PROPERLY_WITH_INSTANCES.blend) and select (one by one) buttom figures (cone/torus/cube) and after selecting each one - press Tab.\nYou will see that instances ignores this mode\n\n", "After duplicating a collection with rigidbody objects, the simulation doesn't apply to the new rigidbody objects\nOperating system: Windows 11\nGraphics card: NVIDIA GeForce RTX 3080 Laptop GPU\n\n\nBroken: `master` branch\nWorked: Don't know if it ever worked\n\nAfter duplicating a collection with rigidbody objects, the simulation doesn't apply to the new rigidbody objects\n\n**Steps to Reproduce**\n\nIn a new blender project, add a new collection. Add a cube to the collection. Add a rigidbody to the cube. Verify that playing the simulation causes the object to fall under gravity.\n\nCopy and paste the cube to create a new cube inside the collection. Move the new cube so both are visible. Verify that playing the simulation causes both cubes to fall under gravity.\n\nIn the outliner right-click on the collection and select \"Duplicate Collection\". Move the two new cubes so all four cubes are visible. Run the simulation.\n- The two cubes in the original collection will fall under gravity.\n- The two cubes in the duplicate collection do not move and are not affected by the simulation.\n\n \n\nIt appears the two cubes in the duplicated collection are not added to the RigidBodyWorld collection.\n\n", "particle instancer modifier 20 years BUG \nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 452.06\n\nBroken: version: 2.91.0\n\n\n\nlook at particle instancer rotation of some parts. you can`t rotate this particles and can`t fix\n\n\n", "Colection instances can't render if duplicated while instanced on a mesh with array\nOperating system: Windows-10-10.0.19043-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.23\n\nBroken: version: 3.6.1\n\ncol instances disapear from render view if duplicated when instanced on an array\n\nmake a plane, add aray modifier. \nput anythng in a collection, instance it \nput the instance as child of the aray, toggle face instances on.\nduplicate the col instance: the previous one(s) will disapear from render.\n\nin this blend file, just turn render view and you'll see only the lastest col instance renders.\n\n", "Rigid body doesn't work when collision is set to mesh, deform and deforming\nOperating system: Windows-10-10.0.19042-SP0 64 Bits\nGraphics card: Radeon RX Vega M GH Graphics ATI Technologies Inc. 4.5.14761 Core Profile Context 21.30.37.04 30.0.13037.4001\n\nBroken: version: 3.2.1\n\nThe rigid body simulation doesn't work with objects using collision with mesh, deform and deforming.\n\n* Make a plane, subdivided\n* Give it a wave modifier\n* Set up the rigid body type: dynamic, active, animated; collision: mesh, deform and deforming.\n* Use another object make it rigid body active\n* Run simulation\n* Sometimes works and when changing rigid body substeps per frame doesn't work at all (well sometimes works but when it does things jump out of control not like in Blender 2.79)\n\nIn Blender 2.79 works perfectly\n\n\n[Wave simulation 3.2.blend](Wave_simulation_3.2.blend)\n\n[Wave simulation 2.79.blend](Wave_simulation_2.79.blend)\n\nBlender 3.2\n\n\n\nBlender 2.79\n\n\n", "Rigid body bugs\nOperating system: macOS-10.15.7-x86_64-i386-64bit 64 Bits\nGraphics card: AMD Radeon Pro 575 OpenGL Engine ATI Technologies Inc. 4.1 ATI-3.10.21\n\nBroken: version: 2.93.1\n\nSimulation somehow broken for small (less than 1 cm) or light (less then 100 grams) objects. In first example below object stay still between bouncing and small shivering. And that is just one example. Another examples are – reaching speed of light instantly (one frame), partial simulation when one object (or some objects) acts like it suppose to, but others act completely inadequate, two or three similar parts in one simulation (example 2). I tried different options but nothing helped. \"Substeps per frame\" and \"Solver iterations\" seems not working because there are no actual changes with different values or simulation just goes nuts.\n\n[r_body-bug.blend](r_body-bug.blend)\n[r_body-bug2.blend](r_body-bug2.blend)\n", "Lots of animation-related operators do not check if the data is editable or not\nMany animation-related operators do not do any check whether the data-block they operate on is linked, liboverride or fully local data. This seems to affect in particular operations from the 3DView (like key insertions), and in the NLA (like adding new tracks, duplicating strips, etc.).\n\nE.g. one can insert keyframes on a linked object (ie create animdata for that linked object, a new local action, and assign that local action to the linked object!).\n\nRelated to issue reported in #98686 (Cant rename local NLA tracks within a local library override data-block.).", "Option to Apply multi-user\n### Motivation\n\nCurrently in Blender when you try to apply the transforms of an object that has its object data shared across multiple objects it will give an error message and abort the operation.\n\n\n\nIn order for #83927 to get implemented this case needs to be tackled. So instead of failing the operation there needs to be an option to apply the transforms of the current or all affected objects anyway.\n\n### Proposal\n\nWhen trying to apply the transforms of an object with shared objetct data it will promt the user with a p[opup message with the text:\n\" Warning\nThe selected objects have shared data. Do you want to continue to apply?\"\n\nThere should be 3 options presented:\n\"Make single user and apply\", \"Apply all affected\" and \"Cancel\"\n\nThis should cover all desirable outcomes.\n\n### Unknowns\n\nMore possible issues by including this feature need to be discussed.", "Force field + hair dynamics not working in (linked) instanced collections\nOperating system: Windows 10 64 bit\nGraphics card: GeForce GTX 1070\n\nBroken: 2.80 3d8cbb534f82 (2019-01-25)\nWorked: 2.79b release\n\nA surface force field does not affect hair dynamics when it is part of a collection that is linked into another file.\n\nSteps to reproduce:\n1. Create a hair particle system, with hairs along x axis. \n2. Enable hair dynamics on hair\n3. Create a surface force field underneath it.\n4. Press play button to verify hair is affected by force field.\n5. Save file\n6. Link the collection from the first file into another file. (I have 'Instance Collections' turned on)\n7. Press play. hair will fall and not be affected by the force field.\n\nExample blend file attached.\n[hair_force_test.blend](hair_force_test.blend)\n", "Can't move collections and objects to another collection at the same time\nOperating system: Linux-5.13.0-7614-generic-x86_64-with-glibc2.33 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 470.63.01\n\nBroken: version: 3.0.0 Alpha\n\ncant move collections and objects to another collection in the same time\n\nNew file\nCreate a cube\nIn the Outliner \nCreate tow collections\nSelect the Cube and one Collection and move them to the other Collection by dragging \nJust the last selected before dragging will be moved\n", "Assert duplicating a collection in a linked scene\nOperating system: Linux-5.13.0-0.rc6.45.fc35.x86_64-x86_64-with-glibc2.34.9000 64 Bits\nGraphics card: NVIDIA GeForce GTX 970M/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 465.31\n\nBroken: version: 3.1.0 Alpha\n\nAssert duplicating a collection in a linked scene\n\nfrom the default startup:\n- 1) `Scene` > `Linked Copy`\n- 2) select the collection `Collection` in the Outliner\n- 3) context menu > `Duplicate Collection` or `Duplicate Linked` (does not matter which)\n\n```\nBLI_assert failed: source/blender/blenkernel/intern/collection.c:184, collection_owner_get(), at '(id->tag & LIB_TAG_NO_MAIN) == 0'\n```\n\nIf the file is saved (and reloaded) before step 3), the above assert does not happen" ]

[ "Lattice objects evaluation is not properly decoupled when they share the same Lattice obdata.\nOperating system: Ubuntu 20.04 Linux-5.4.0-39-generic-x86_64-with-debian-bullseye-sid 64 Bits \nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 435.21\n\nBroken: version: 2.90.0 Alpha\nWorked: Never\n\nOpen the file below and see how scaling one bone from one of the armature affects both lattice objects, which are sharing the same obdata.\n\n[#78650.blend](T78650.blend)\n\n------------------------------------\n\n*original report:*\n\nSo I have a model of book page with lattice modifier for blending corners. Then using Library Override, Lattice modifier gets object from one and second linked blocks, constantly jumping over each other.\n\n[2020-07-05 21-25-55.mp4](2020-07-05_21-25-55.mp4)\n\n\n\n\n1. Extract this 2 files in same folder.\n\n[Library Override Modifer Test.zip](Library_Override_Modifer_Test.zip)\n\n2. Open test.blend and play animation.\n\n\nPS. I post this information here T70319, but it is not been resolved for my case, so I decided to open new bug report \n \n\n", "Overrides: Lattices modifiers \"fight each other\" between different copies when duplicated\nBroken: version: 2.93.0 Alpha\nWorked: This case hasn't come up much so far, so it probably never worked. It will be a common case during production now that we have some Forest Creature rigs, which will be duplicated a lot and use lattices.\n\nIt seems when a lattice which is attached to an armature via Hook modifier is overridden and then duplicated, it starts to \"flicker\".\n[07_lattice_override_flicker-2021-03-19_16.06.32.mp4](07_lattice_override_flicker-2021-03-19_16.06.32.mp4)\n\n- Lib file: [lattice_boy.blend](lattice_boy.blend)\n- Linking file: [lattice_boy_linked.blend](lattice_boy_linked.blend)\n- Put them next to each other and open the Linking file (lattice_boy_linked.blend)\n- Try moving the lattice controls in pose mode, it should flicker between the two states that now exist in the file.\n- If you delete one of them, it works fine again.\n- Duplicate it again, it flickers again.\n\nThe lattice is hooked up to the armature using a Hook modifier. You can add a Cast modifier with an Empty as the target object and the same issue occurs, so the issue is not specific to Hook modifiers, but probably all lattice modifiers.\n\nCC @mont29 " ]

[ "Crash when auto script execution is enabled\nOperating system: Windows 10 64-bit operating system, x64-based processor\nGraphics card: NVIDIA GeForce RTX 2060\n\nBroken: Blender 4.0.0 7372b3bdda35 and later\nWorked: never? 2.93 fails too\n\nBlender daily build crashes immediately on launch. The blender start screen isn't shown at all, only a grey window which then immediately closes. The crash log indicates that it is a memory EXCEPTION_ACCESS_VIOLATION (see log file attached).\n\n- Open preferences, enable `auto script execution`\n- Save prefs and close blender\n- Rename attached file and replace it with existing startup file\n- Launch blender (crash)\n\n- - -\n<details>\n <summary> Original report </summary>\nStart Blender through Steam, Blender Launcher, blender_debug_log.cmd, blender_debug_gpu.cmd, or blender_debug_gpu_glitchworkaround.cmd.\nA grey window opens with the Blender logo in the top bar.\nThe window closes as Blender crashes.\n\n[ Crashing_4-0-0_daily.blend](https://projects.blender.orgattachment)\n</details>", "just activating osl, without using it, makes any render 3 to 4 times slower\nWin7 x64\n\nBroken: 2.76b\n\nTake any prodution render with CPU, just activate osl in the render tab, it will render \" to 4 time slower, without using it in the scene. Compared to GPU render wich is already 5 times faster than a cpu render, it is = 15 to 20 times slower. OSL is unusable with this bug. Note that other renderer with OSL don't have this speed hit.\n\nWith CPU and 16x16 tiles, render this scene: pabellon_barcelona_v1.scene_.zip\nthen just activate osl, render again and compare times. although there isn't any use of OSL in the scene, the render times crawls down.\n(you can also lower the samples to 10 to make it faster for both render)", "BLI_path_util_test.cc TODO's\nThis is an area we have some tests setup, so its just a matter of adding remaining tests.\n*This is a relatively easy place to start and good for newer developers*\n\nNote, with these string functions typically the common cases are well tested, but the corner cases can trip us up, so be sure to include tests for empty strings, utf-8 with multi-byte characters, strings containing all path separators - the kinds of cases that may be overlooked.\n\nDone:\n\n- BLI_cleanup_path\n- BLI_path_frame\n- BLI_split_dirfile\n\n\nTODO:\n\n- BLI_split_name_num\n- BLI_make_file_string\n- BLI_path_suffix\n- BLI_path_frame\n- BLI_path_frame_range\n- BLI_stringdec, BLI_stringenc --- check they are compatible\n- BLI_path_abs\n- BLI_path_rel\n- BLI_join_dirfile\n\n- BLI_replace_extension\n- BLI_ensure_extension\n- BLI_ensure_filename\n\n*Note, its probably unreasonable to attempt to test all these functions as a single patch, so fine to submit tests for only a few*", "Assertion reloading space_headers.blend (UI test file)\nLinux-6.3.3\n\nBroken: `b567b249d40488e5d90235e1a89c98853c5fd099` (4.0 main branch)\nWorked: (needs investigation)\n\n\nLoading `lib/tests/ui_tests/space_headers.blend` twice asserts.\n\n\n1. Run `blender ../lib/tests/ui_tests/space_headers.blend`.\n2. Ctrl-O (and open the file again).\n\nBlender exits with:\n```\nBLI_assert failed: source/blender/editors/space_file/filesel.cc:1094, ED_fileselect_init_layout(), at 'layout->rows != 0'\n```", "Depsgraph - Dependency Cycle - concurrency issue leading to memory after use errors.\nDisclaimer: This is a very elusive 'bug'. I stumbled on it after making changes in a completely unrelated area (liboverrides diffing code in fcddb7cda7). Artificially raising the amount of threads seems to make it more easily reproducible though.\n\nReproducible fairly consistently on my machine with Sprites production file `110_0010_A.anim.blend` version `r2747`. This file contains dependency loops, which as far as I understand makes depsgraph fail to fully handle dependencies between updates, leading to some data being freed while/before it is used from another ID.\n\nNot sure how much we support these dependency loops cases, but think crashes at least should not be accepted here?\n\n--------------------------------\n\nHere is a small patch I used to debug:\n\n```\ndiff --git a/source/blender/blenkernel/intern/object_update.cc b/source/blender/blenkernel/intern/object_update.cc\nindex 625aa666023..7f1f4a1b4d2 100644\n--- a/source/blender/blenkernel/intern/object_update.cc\n+++ b/source/blender/blenkernel/intern/object_update.cc\n@@ -376,6 +376,20 @@ void BKE_object_select_update(Depsgraph *depsgraph, Object *object)\n if (object->type == OB_MESH && !object->runtime.is_data_eval_owned) {\n Mesh *mesh_input = (Mesh *)object->runtime.data_orig;\n std::lock_guard lock{mesh_input->runtime->eval_mutex};\n+\n+ BLI_assert(mesh_input->runtime->mesh_eval != nullptr);\n+ BLI_assert(mesh_input->runtime->mesh_eval == static_cast<Mesh *>(object->data));\n+ if (mesh_input->runtime->mesh_eval != static_cast<Mesh *>(object->data)) {\n+ printf(\n+ \"%s: mesh_input->runtime->mesh_eval (%p) of %s does not match object->data (%p) of \"\n+ \"%s\\n \",\n+ __func__,\n+ mesh_input->runtime->mesh_eval,\n+ mesh_input->id.name,\n+ object->data,\n+ object->id.name);\n+ }\n+\n BKE_object_data_select_update(depsgraph, static_cast<ID *>(object->data));\n }\n else {\n```\n\nThe interesting part of the error log with an ASAN build:\n\n```lines=10\nDependency cycle detected:\n OBGEO-log_moss/MODIFIER(GeometryNodes) depends on\n GRfallen_log/GEOMETRY_EVAL_DONE() via 'Nodes Modifier'\n OBGEO-log_moss/GEOMETRY_EVAL() via 'Collection Geometry'\n OBGEO-log_moss/MODIFIER(GeometryNodes) via 'modifier stack order'\nDependency cycle detected:\n OBGEO-oak_slingshot-moss/GEOMETRY_EVAL() depends on\n OBGEO-oak_slingshot-moss/MODIFIER(GeometryNodes) via 'modifier stack order'\n GRPR-slingshot_tree/GEOMETRY_EVAL_DONE() via 'Nodes Modifier'\n OBGEO-oak_slingshot-moss/GEOMETRY_EVAL() via 'Collection Geometry'\nDetected 2 dependency cycles\n\nBLI_assert failed: source/blender/blenkernel/intern/object_update.cc:380, BKE_object_select_update(), at 'mesh_input->runtime->mesh_eval != nullptr'\nBLI_assert failed: source/blender/blenkernel/intern/object_update.cc:381, BKE_object_select_update(), at 'mesh_input->runtime->mesh_eval == static_cast<Mesh *>(object->data)'\nBKE_object_select_update: mesh_input->runtime->mesh_eval ((nil)) of MEPlane.029 does not match object->data (0x61c000286888) of OBGEO-oak_slingshot-moss\n\n =================================================================\n==404553==ERROR: AddressSanitizer: heap-use-after-free on address 0x61c0002868b0 at pc 0x000013bb005c bp 0x7f394ec88070 sp 0x7f394ec88068\nREAD of size 2 at 0x61c0002868b0 thread T137\n - 0 0x13bb005b in BKE_object_data_select_update /home/bastien/blender/src/source/blender/blenkernel/intern/object_update.cc:357\n - 1 0x13bb0a4d in BKE_object_select_update /home/bastien/blender/src/source/blender/blenkernel/intern/object_update.cc:393\n - 2 0x168103db in operator() /home/bastien/blender/src/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:1532\n - 3 0x1684125a in __invoke_impl<void, blender::deg::DepsgraphNodeBuilder::build_object_data_geometry(Object*)::<lambda(Depsgraph*)>&, Depsgraph*> /usr/include/c++/12/bits/invoke.h:61\n - 4 0x1683a429 in __invoke_r<void, blender::deg::DepsgraphNodeBuilder::build_object_data_geometry(Object*)::<lambda(Depsgraph*)>&, Depsgraph*> /usr/include/c++/12/bits/invoke.h:111\n - 5 0x1683289f in _M_invoke /usr/include/c++/12/bits/std_function.h:290\n - 6 0x1696ceb0 in std::function<void (Depsgraph*)>::operator()(Depsgraph*) const /usr/include/c++/12/bits/std_function.h:591\n - 7 0x1696787c in evaluate_node /home/bastien/blender/src/source/blender/depsgraph/intern/eval/deg_eval.cc:101\n - 8 0x16967caf in deg_task_run_func /home/bastien/blender/src/source/blender/depsgraph/intern/eval/deg_eval.cc:119\n - 9 0x2f3f586c in Task::operator()() const /home/bastien/blender/src/source/blender/blenlib/intern/task_pool.cc:164\n - 10 0x2f3f9022 in tbb::internal::function_task<Task>::execute() /opt/lib/tbb/include/tbb/task.h:1059\n - 11 0x7f39bce4cff4 in tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::process_bypass_loop(tbb::internal::context_guard_helper<false>&, tbb::task*, long) (/opt/lib/tbb/lib/libtbb.so+0x2cff4)\n - 12 0x7f39bce4d382 in tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::local_wait_for_all(tbb::task&, tbb::task*) (/opt/lib/tbb/lib/libtbb.so+0x2d382)\n - 13 0x7f39bce32fb6 in tbb::internal::arena::process(tbb::internal::generic_scheduler&) (/opt/lib/tbb/lib/libtbb.so+0x12fb6)\n - 14 0x7f39bce4206f in tbb::internal::market::process(rml::job&) (/opt/lib/tbb/lib/libtbb.so+0x2206f)\n - 15 0x7f39bce4605b in tbb::internal::rml::private_worker::run() (/opt/lib/tbb/lib/libtbb.so+0x2605b)\n - 16 0x7f39bce46298 in tbb::internal::rml::private_worker::thread_routine(void*) (/opt/lib/tbb/lib/libtbb.so+0x26298)\n - 17 0x7f39ab0a7fd3 in start_thread nptl/pthread_create.c:442\n #18 0x7f39ab12866b in clone3 ../sysdeps/unix/sysv/linux/x86_64/clone3.S:81\n\n0x61c0002868b0 is located 48 bytes inside of 1704-byte region [0x61c000286880,0x61c000286f28)\nfreed by thread T137 here:\n - 0 0x7f39bc6b76a8 in __interceptor_free ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:52\n - 1 0x2f43bbfa in MEM_lockfree_freeN /home/bastien/blender/src/intern/guardedalloc/intern/mallocn_lockfree_impl.c:117\n - 2 0x134f9b18 in BKE_id_free_ex /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id_delete.c:168\n - 3 0x134f9b49 in BKE_id_free /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id_delete.c:174\n - 4 0x1364372f in BKE_mesh_eval_geometry /home/bastien/blender/src/source/blender/blenkernel/intern/mesh.cc:1876\n - 5 0x1681190d in operator() /home/bastien/blender/src/source/blender/depsgraph/intern/builder/deg_builder_nodes.cc:1561\n - 6 0x16841442 in __invoke_impl<void, blender::deg::DepsgraphNodeBuilder::build_object_data_geometry_datablock(ID*)::<lambda(Depsgraph*)>&, Depsgraph*> /usr/include/c++/12/bits/invoke.h:61\n - 7 0x1683a7bb in __invoke_r<void, blender::deg::DepsgraphNodeBuilder::build_object_data_geometry_datablock(ID*)::<lambda(Depsgraph*)>&, Depsgraph*> /usr/include/c++/12/bits/invoke.h:111\n - 8 0x16832ad5 in _M_invoke /usr/include/c++/12/bits/std_function.h:290\n - 9 0x1696ceb0 in std::function<void (Depsgraph*)>::operator()(Depsgraph*) const /usr/include/c++/12/bits/std_function.h:591\n - 10 0x1696787c in evaluate_node /home/bastien/blender/src/source/blender/depsgraph/intern/eval/deg_eval.cc:101\n - 11 0x16967caf in deg_task_run_func /home/bastien/blender/src/source/blender/depsgraph/intern/eval/deg_eval.cc:119\n - 12 0x2f3f586c in Task::operator()() const /home/bastien/blender/src/source/blender/blenlib/intern/task_pool.cc:164\n - 13 0x2f3f9022 in tbb::internal::function_task<Task>::execute() /opt/lib/tbb/include/tbb/task.h:1059\n #14 0x7f39bce4cff4 in tbb::internal::custom_scheduler<tbb::internal::IntelSchedulerTraits>::process_bypass_loop(tbb::internal::context_guard_helper<false>&, tbb::task*, long) (/opt/lib/tbb/lib/libtbb.so+0x2cff4)\n\npreviously allocated by thread T0 here:\n - 0 0x7f39bc6b83b7 in __interceptor_calloc ../../../../src/libsanitizer/asan/asan_malloc_linux.cpp:77\n - 1 0x2f43c32f in MEM_lockfree_callocN /home/bastien/blender/src/intern/guardedalloc/intern/mallocn_lockfree_impl.c:223\n - 2 0x134ede74 in BKE_libblock_alloc_notest /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id.c:1101\n - 3 0x134edfdc in BKE_libblock_alloc /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id.c:1113\n - 4 0x134ef71d in BKE_libblock_copy_ex /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id.c:1274\n - 5 0x134e9788 in BKE_id_copy_ex /home/bastien/blender/src/source/blender/blenkernel/intern/lib_id.c:637\n - 6 0x13632b83 in BKE_mesh_copy_for_eval /home/bastien/blender/src/source/blender/blenkernel/intern/mesh.cc:1109\n - 7 0x122b0c00 in operator() /home/bastien/blender/src/source/blender/blenkernel/intern/DerivedMesh.cc:1130\n - 8 0x122c264c in operator() /opt/lib/tbb/include/tbb/task_arena.h:96\n - 9 0x7f39bce332b4 in tbb::interface7::internal::isolate_within_arena(tbb::interface7::internal::delegate_base&, long) (/opt/lib/tbb/lib/libtbb.so+0x132b4)\n #10 0x7fff66c6047f ([stack]+0x4b47f)\n```\n\nThis trace makes it pretty clear that `BKE_object_select_update` from object evaluation code accesses data that has been freed by Mesh evaluation code (`BKE_mesh_eval_geometry`).\n\nNote that skipping (returning) here does not help, since code will crash in another place then (like inside the `BKE_library_foreach_ID_link` call in `deg_validate_copy_on_write_datablock`).\n\n----------------------------\n\nPS: Also seems weird to me that freeing runtime `mesh_eval` in `BKE_mesh_eval_geometry` is not protected by a lock on `mesh->runtime->eval_mutex`?\n", "Cycles: Enabling/disabling OSL changes the look of pre-existing shader nodes\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3090/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 528.49\n\nBroken: version: 3.5.0\n\nI can't find much info on osl within blender. As i understand it, the point is to be able to write your own shaders with scripts, it shouldn't change the functionality of preexisting shader nodes.\n\nOptix without OSL and CPU (with & without OSL) device renders darker droplets.\n\nIn rendered view. Enable/disable OSL and see the water droplets look drasticly different (OSL looks more correct, but that should just be the default behavior then)\n\n", "Optix not available after Manjaro Update\nOperating system: Linux-5.17.1-3-MANJARO-x86_64-with-glibc2.35 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 510.60.02\n\nBroken: version: 3.1.2\nWorked: both 3.1.0 as well as 3.1.2 was working before the Manjaro update. Version 3.0.0 is still working, regardless of the OS update.\n\nSince the last Manjaro Update I get an error message when I try to switch to rendered view in Blender 3.1.2 :\n```\nFailed to load Optix Kernel from: '/_DATA_WK/Blender/blender-3.1.2-linux-x64/3.1/scripts/addons/cycles/lib/kernel_optix.ptx' (OPTIX_ERROR_INVALID_FUNCTION_USE)\n```\n\nI tested it with Blender installed from the Manjaro Repository as well with Blender from Blender.org (extracted to the above location).\nBoth of them gives the same error message (with different kernel_optix.ptx location of course.)\nThey were working perfectly before the OS update. Blender 3.0.0 still works fine. \n\n- Start Blender. (No need to load any file or content).\n- switch to Rendered Viewport Shading. \n\n", "Viewport animation render crashes when collapsing the viewport.\nBroken: version: 2.91.0 Beta (and 2.83, 2.80, ...)\nWorked: never\n\nViewport animation render crashes when collapsing the viewport.\n\n[0001-0485.mp4](0001-0485.mp4)\n[Viewport render crash.blend](Viewport_render_crash.blend)\n\n", "Custom render engine is unable to create gpu context\nOperating system: Linux-5.19.17-2-MANJARO-x86_64-with-glibc2.36 64 Bits\nGraphics card: AMD Radeon Graphics (renoir, LLVM 14.0.6, DRM 3.47, 5.19.17-2-MANJARO) AMD 4.6 (Core Profile) Mesa 22.3.1\n\nBroken: version: 3.5.0 Alpha\nWorked: Blender 3.4\n\nUnable to perform background rendering using `gpu` module in custom render engine.\n\nRun the script in background mode: `blender -b -P background-gpu.py`. Blender throws: `SystemError: GPU API is not available in background mode`, despite the line `bl_use_gpu_context = True` in custom render engine. Worked perfectly fine in blender 3.3 and 3.4. Not in 3.5 though.\n[background-gpu.py](background-gpu.py)\n\n", "rendering crashed\nOperating system:\nGraphics card:\n\nBroken: (example: 2.80, edbf15d3c044, master, 2018-11-28, as found on the splash screen)\n\n\nBased on the default startup or an attached .blend file (as simple as possible).\n\n", "Unable to fix broken linked IDs when renamed in libfile\nWindows 10\nWindows 7 x64\n\nWorked: 2.82\nBroken: 2.83.0\n\nCant rename linked collection. Error: Cannot edit external library data\n\nDefault Blender setup (first file lets call it assets.blend)\nCreate a collection and name it \"Test\" \nCreate a cube and move it to \"Test\" collection\nSave the file\n\nCreate another file (lets call it level.blend)\nLink the \"test\" collection from the assets.blend\nCheck all linking options before hiting link\nSave the file\n\nNow open assets.blend again and rename \"Test\" collection to \"Cube\"\nAnd save it again.\n\nOpen level.blend\nA message appears that link is broken\nGo to Outliner and change Display mode to Blender File\nNow go down to the linked collection\nDouble click broken collection to rename it\nError message appears \"Cannot edit external library data\"\n\nRenaming linked collections works fine in 2.82\n", "LineArt; Crahes when I reopen a file\nAn old HP laptop with\nOperating system: Linux Ubuntu 18.04.5 LTS\nGraphics card: Mesa DRI Intel(R) HD Graphics 5500 (BDW GT2)\n\nBroken: Blender 2.93.0 Alpha, 3ea1779365b5, master, 2021-03-23 18:34\n\n\nI added two GP LineArt on the scene and it worked and then I saved the scene, with new grease pencil modifiers activated, but when I try to reopen in Blender the same scene crashes every times and I'm not able to see anythings after waiting the processing until crashes without any report. the file is attached.\n\n[yach.tar.xz](yach.tar.xz)", "Blender 3.0.0 breaks compatibility of rigs with the previous version\nOperating system: Windows 11 Pro 21H2\nGraphics card: Nvidia RTX 3070\n\nBroken: 3.0.0\nWorked: 2.93.6\nBlender 3.0.0 breaks compatibility of rigs with the previous version.\n\n1. Open Agent 327 shot 10_03_B-agent_dodges animation file (10_03_B.anim.blend)\n2. Observe that the rigs are completely broken\n\n\n", "Microsoft store: Blender cannot be used in terminal\nWindows 11\n\nBroken: 3.4\nWorked: 3.3\n\n**Short Description**\nEver since blender 3.4 released, I have been unable to use blender in the command line. Running just the command \"blender\" gives error 0xc0000022 and whenever I try to run any actual command like \"blender -b untitled.blend -a\" nothing happens. I have never had any issues like this in versions prior to 3.4. I am running the Microsoft store version.\n\nIssue only happens with blender from Microsoft store\n\n**Steps to Reproduce**\n * Assuming that blender is already in your system path, which it should be by default,\n * open a terminal in absolutely any folder;\n * run absolutely any blender command;\n * for example, run \"blender -E help\"\n\n- - -\n!!For now, workaround is to type \"blender-launcher\" instead of \"blender\" to launch blender from command line (but that won't accept any arguments)!!", "Impossible to copy/paste data between different Blender instances that are using different temp directories\n|Operating system:|Linux-5.8.16-2-MANJARO-x86_64-with-glibc2.2.5 64 Bits\n| -- | -- |\n||Windows 10\nGraphics card: AMD VERDE (DRM 2.50.0, 5.8.16-2-MANJARO, LLVM 10.0.1) X.Org 4.5 (Core Profile) Mesa 20.1.8\n\nBroken: version: 2.92.0 Alpha\nWorked: Never\n\nIt is not possible to copy/paste data between two Blender instances that use a different `Temporary Files` path.\n\n- Open two instances of Blender (regardless of whether they are the same version or different versions)\n- In the destination instance, delete all objects and set a custom path for Temporary Files (in Blender Preferences - File Paths)\n- In the other instance, Ctrl-C to copy default cube, \n- In the destination instance, Ctrl-V to paste,\nIt is not pasted, message \"no objects to paste\"\n\n" ]

[ "Error: OSL script compilation failed\nOperating system: Linux-4.15.0-136-generic-x86_64-with-glibc2.27 64 Bits\nGraphics card: Mesa DRI Intel(R) HD Graphics 5500 (BDW GT2) Intel Open Source Technology Center 4.6 (Core Profile) Mesa 20.0.8\n\nBroken: version: 2.93.0 Alpha\nWorked: (2.92.0)\n\nWhen I try to compile any OSL shader I get an error below\n\n```\nblender-2.93.0/2.93/scripts/addons/cycles/shader/stdcycles.h:31:10: fatal error: 'stdosl.h' file not found\n#include \"stdosl.h\"\n ^~~~~~~~~~\nError: OSL script compilation failed, see console for errors\n```\n\nI tryed with this OSL shader on manual\n\n\n```\nshader simple_material(\n color Diffuse_Color = color(0.6, 0.8, 0.6),\n float Noise_Factor = 0.5,\n output closure color BSDF = diffuse(N))\n{\n color material_color = Diffuse_Color * mix(1.0, noise(P * 10.0), Noise_Factor);\n BSDF = material_color * diffuse(N);\n}\n```\n\nafter I added a new text file inside Blender with code of simple OSL I added a script node with the OSL shader but this isn't compiled.\nattached .blend file\n[Simple_material_OSL_293_version.blend](Simple_material_OSL_293_version.blend)\n\n" ]

[ "Segfault pasting stale driver.\nOperating system: Linux\nGraphics card: Intel\n\nBroken: 2.92\n\n\nDrives copied and pasted with significant change in file data in between can trigger segfaults. \n\n```\n# Blender 2.92.0, Commit date: 2021-02-24 16:25, Hash 02948a2cab44\n\n# backtrace\nblender(BLI_system_backtrace+0x34) [0x55dfd70fd084]\nblender(+0xe151bd) [0x55dfd4bd41bd]\n/usr/lib/libc.so.6(+0x3cf80) [0x7f9b26870f80]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder26build_object_data_geometryEP6Objectb+0x170) [0x55dfd7000380]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder17build_object_dataEP6Objectb+0x54) [0x55dfd6ffb4b4]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder12build_objectEiP6ObjectNS0_26eDepsNode_LinkedState_TypeEb+0x1d3) [0x55dfd6ffe2c3]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder22build_driver_variablesEP2IDP6FCurve+0x77) [0x55dfd6ffad87]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder14build_animdataEP2ID+0x1f9) [0x55dfd6ffe8b9]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder12build_objectEiP6ObjectNS0_26eDepsNode_LinkedState_TypeEb+0x228) [0x55dfd6ffe318]\nblender(_ZN7blender3deg20DepsgraphNodeBuilder16build_view_layerEP5SceneP9ViewLayerNS0_26eDepsNode_LinkedState_TypeE+0xb2) [0x55dfd7003892]\nblender(_ZN7blender3deg23AbstractBuilderPipeline16build_step_nodesEv+0x76) [0x55dfd6fea296]\nblender(_ZN7blender3deg23AbstractBuilderPipeline5buildEv+0x42) [0x55dfd6fea4b2]\nblender(DEG_graph_build_from_view_layer+0x33) [0x55dfd6fd24e3]\nblender(+0xef501a) [0x55dfd4cb401a]\nblender(wm_event_do_depsgraph+0xc5) [0x55dfd4f59245]\nblender(wm_event_do_refresh_wm_and_depsgraph+0x92) [0x55dfd4f593b2]\nblender(wm_event_do_notifiers+0x36a) [0x55dfd4f6055a]\nblender(WM_main+0x2b) [0x55dfd4f5532b]\nblender(main+0x367) [0x55dfd4b9da67]\n/usr/lib/libc.so.6(__libc_start_main+0xd5) [0x7f9b2685bb25]\nblender(_start+0x2e) [0x55dfd4bd056e]\n\n# Python backtrace\n```\n\n\n* Add a custom property to the Scene.\n\n* Right click on its value field, and run \"Copy as New Driver\".\n\n* Run `CTRL`+`N` or \"File\" —> \"New\" —> \"General\".\n\n* Right click on any value field, and run \"Paste Driver\".", "Brushes Assets Viewport User Interface\n{[F13331829](image.png), size=full}\n\nAdditional Notes:\n\n{[F13331831](image.png), size=full}", "[Parent Task] Fix UBSan issues\nAt @ideasman42's suggestion, I'm creating this as a parent task to resolving all the low-hanging fruit issues caught by Undefined Behavior Sanitizer. Most of these that I have personally found have been alignment issues but UBSan covers a wide range of conditions.\n\nBeyond making Blender possibly more performant and/or stable, the primary goal is to reduce noise from UBSan output for easier debugging. **This is a low priority task!**\n\nCurrently know related tasks include:\n- [x] [D14798](T97691: RNA function call args alignment issue (resolved)\n- [x] #81340: UBSan: several runtime errors at launch (found in issue search, status unknown.)\n- [x] [D15390](D15390): GL offset pointer calculated as index into null array pointer (fix approved pending merge)\n- [ ] #97693: Unexplained `sizeof(void*)*2` alignment requirement on `blender::deg::OperationNode` instances (Archived but unresolved)\n- [ ] [D15262](D15262): Alignment issue with `Icon_Geom->colors` buffer\n- [ ] #99385 (Ubsan invalid vptr errors in TBB)\n", "\"Illegal address in CUDA queue\" error occurs when I use both Principled BSDF and Adaptive Subdivision\nOperating system: Windows 11 Pro 64 Bits\nGraphics card: NVIDIA GeForce GTX 1050 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 536.99\n\nBroken: version: 3.6.2\n\n-**\"Illegal address in CUDA queue\"** error occurs when I activate both Principled BSDF and Adaptive Subdivision at the same time.\n-It would work fine if I don't use both at the same time.\n-Here are the possibilities of the error occurring when I use different settings:\n1. Principled BSDF + Adaptive Subdivision = **error**\n2. Texture Coordinate surface + Adaptive Subdivision = **no error**\n3. Principled BSDF + NO Adaptive Subdivision = **no error**\n\n-I can use a high poly count without Adaptive Subdivision *(e.g. Level viewport = 7)*, works fine. But as soon as I turn it on *(even with low poly count, e.g. Dicing scale = 1)* combined with Principled BSDF, the error occurs.\n\nIn the file I sent, I've already activated Principled BSDF + Adaptive Subdivision, you need to turn on the \"Rendered Viewport Shading\" and then slide the Scale parameter of the Noise Texture very quickly and the error should occur.\n", "smaller startup.blend\nNot sure if this is technically a bug, but...\n\nI notice that `startup.blend` includes 64 blocks of type `MTex` that don’t seem to be referenced anywhere. Loading the startup file into Blender and then saving it without making any changes removes these blocks.\n\nEnclosed [startup.blend](startup.blend) is a copy of `startup.blend` from the current Blender source tree after processing through my experimental [blendhack ](blendhack) script to remove the unreferenced blocks. I also saved it with 32-bit rather than 64-bit pointers, which knocks about another 50K off the size.\n\nI also note the potential for information leaks in the current procedure for saving `.blend` files; for example, fixed-length fields containing null-terminated strings retain traces of previous values after the terminating null. And the saving of in-memory addresses can give hints about the configuration of the system on which the file was saved.", "[winerror 233] .glb, .gltf\nOperating system: Windows-10-10.0.19041-SP0 64 Bits\nGraphics card: NVIDIA GeForce MX130/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 472.42\n\nBroken: version: 3.6.2\n\n\n\ncant import .glb or .gltf files\n\n", "2.90 Not opening on Linux\nOperating system: Linux Mint 18.1\nGraphics card: Intel HD 620\n\nBroken: \nBlender 2.90 won't open up fully, it appears to open up and immediately closes\nWorked: \n\nTried a few different ways to open, won't open.\n\nOne of the methods I tried opening with was through the Terminal and this was the message I had returned back:\n\nRead prefs: /home/john/.config/blender/2.90/config/userpref.blend\nfound bundled python: /home/john/Desktop/3D Related/1BlenderVersions/blender-2.90.0-linux64/2.90/python\nGPUShader: linking error:\nerror: Input block `ShaderStageInterface' is not an output of the previous stage\n\nGPUShader: linking error:\nerror: Input block `ShaderStageInterface' is not an output of the previous stage\n\nWriting: /tmp/blender.crash.txt\nSegmentation fault", "Optix fails when geometry nodes instance has animated translation and scale 0 and motion blur is on\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: NVIDIA GeForce RTX 3080 Ti/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 531.18\n\nBroken: version: 3.6.0 Alpha\nAlso broken in 3.4.1\n\nRendering will not start. It stops at \"Building OptiX acceleration structure\"\n\n\n\nThe culprit seems to be a combination of animated Translation and Scale 0 together with animated Rotation of the object and motion blur.\n\n\n\nDoesn't work in OptiX. Cuda and CPU both work.\n\nOpen the .blend file and render the frame.\n\n", "Edit Source operator missing the \"see XXX.py in the text editor\" info in the status bar\nOperating system: Linux-6.4.13-200.fc38.x86_64-x86_64-with-glibc2.37 64 Bits, WAYLAND UI\nGraphics card: AMD Radeon Graphics (renoir, LLVM 16.0.6, DRM 3.52, 6.4.13-200.fc38.x86_64) AMD 4.6 (Core Profile) Mesa 23.1.6\n\nBroken: version: 4.0.0 Alpha\nWorked: 3.6 LTS\n\nCaused by 1251b7c4001b3f7de158c089256b04c887d6c5bb\n\nEdit Source operator missing the \"see XXX.py in the text editor\" info in the status bar\n\n- use the context menu on a button in the UI and call Edit Source\n- this is missing the following info in the status bar:\n\n\n", "Show warning when saving a .blend file with that has been saved with a newer version before\nThe following scenario is a big difficult with our versioning code:\n1. Save file in new version of Blender.\n2. Load and save the same file in an older version of Blender.\n3. Load the file again in the new version.\n\nNot all our versioning code is idempotent and it can lead to hard to detect bugs when some versioning code runs twice. Showing a warning when saving a .blend file in an older version of Blender might help mitigate that issue a bit.\n\nE.g. #93252.", "bake operator reports uninitialised image that isn't even in use in the shader..\nOperating system: Windows-10-10.0.19044-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 526.98\n\nBroken: version: 3.5.1 Release Candidate\n\ncan't bake due to strange error\n\nopen blend file\nselect object\nselect the baked image texture node in the shader editor\nclick bake in the render setting (leave settings as is)\n\nstrange error pops up saying \n\"uninitialised image \"nightmare0134.png\" from object 'pavement.002'\n\nThis image isn't even in use by the shader.\n\nPS, it seems intermittently not possible to paste screengrabs into this new bug reporter. Also, how to do bullet points and embed video?\n\n", "Startup.blend file is not loaded on startup if userpref.blend does not exist.\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: Intel(R) UHD Graphics 620 Intel 4.5.0 - Build 26.20.100.7000\n\nBroken: version: 2.81 (sub 3)\nWorked: (optional)\n\nStartup.blend file is not loaded on startup if userpref.blend does not exist.\n\n- Remove userpref.blend file from it's usual location\n- Start blender\n- Dismiss splash screen by clicking on 3D viewport\n- Make some change to the file (move the cube)\n- Save as startup file\n- Close Blender\n# Open blender\n\nStartup file is not loaded", "Blender-3.1 (git) boolean in Geomenty Nodes issue\nOperating system:\nGraphics card:\n\nBroken: All\n\nBulean issue like T91889\nbut now in GN\n[2021-11-25_22-29-42.mp4](2021-11-25_22-29-42.mp4)\nTest Scene\n[parktronic1-Blender-3.blend](parktronic1-Blender-3.blend)", "Dicrepency between requested mesh data layer and eval mesh data layer\nWhile fixing #91838 it became apparent that the data layers requested by `mesh_cd_calc_used_gpu_layers` differs from the ones when creating the batches. Which lead in this case to a crash.\n\nThe fix was just a workaround and fixing this would mean to have the same layers at request time. Or maybe, not assuming they will be available during batch creation.", "Blender Crash. Code marked as unreachable has been executed.\nOperating system: Windows-11-10.0.22621-SP0 64 Bits\nGraphics card: NVIDIA GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.6.0 NVIDIA 536.67\n\nBroken: version: 3.6.1, 4.0.0 Alpha\nWorked: didn't test\n\nOpen the .blend and go to frame 60. Crash\n\nThis is a .blend from a random user. So, sadly, I don't know the exact steps to reproduce.\nIt just said: \"Code marked as unreachable has been executed. Please report this as a bug.\"\n\n" ]

[ "not export to fbx\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 950/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 446.14\n\nBroken: version: 2.83.0\n\nnot export to fbx\n\nTraceback (most recent call last):\n```\nFile \"C:\\Program Files\\Blender Foundation\\Blender 2.83\\2.83\\scripts\\modules\\bpy_extras\\io_utils.py\", line 76, in invoke\n blend_filepath = context.blend_data.filepath\n```\nUnicodeDecodeError: 'utf-8' codec can't decode byte 0xc4 in position 24: invalid continuation byte\n\nlocation: <unknown location>:-1\n\n\n", "Various Addons using filesystem properties (PROP_FILEPATH etc): UnicodeDecodeError: 'utf-8' codec can't decode byte 0xe0 in position 24: invalid continuation byte\nOperating system: Window 10\nGraphics card: NVIDIA GeForce GT 1030\n\nBroken: 2.80\n\nI wanted to export model to .obj, but i can't its giving me\n```\nTraceback (most recent call last):\n File \"D:\\Steam\\steamapps\\common\\Blender\\2.82\\scripts\\modules\\bpy_extras\\io_utils.py\", line 76, in invoke\n blend_filepath = context.blend_data.filepath\nUnicodeDecodeError: 'utf-8' codec can't decode byte 0xe0 in position 24: invalid continuation byte\n\nlocation: <unknown location>:-1\n```\nAny idea how i can fix?\n\nMake a model from plane with 2 picture, 5 material. [Sword4(finish1).blend](Sword4_finish1_.blend)" ]

[ "Some transform modes don't apply absolute grid snaping \nOperating system: Windows 10 \nGraphics card: GTX 1080 ti \n\nBroken:\nBlender 2.8 January 31, 01:48:40 - 8c8979067490 Beta\nBlender 2.79 (It worked)\nBlender 2.77 (It worked)\n\nAbsolut grid snaping dont work with \"Face Inset\" and neither with loop cuts. \n\nTurn on the snaping (increment) -> turn on absolut grid snaping -> spawn a plane -> try to snap while loop cutting or inseting faces.", "Cycles: UV render pass corrupted in viewport\nOperating system: Windows-10-10.0.18363-SP0 64 Bits\nGraphics card: Intel(R) Iris(R) Plus Graphics Intel 4.5.0 - Build 27.20.100.9664\n\nBroken: version: 3.1.0 Alpha\nWorked: 2.93 LTS\n\nWhen switching the render passfrom normal to UV the UVs are not displayed correctly.\n\n\nDownload bmw demo file.\nin viewport switch to render mode\nEnable the normal render pass. After that enable the uv render pass.\n\n{[F11785393](uv_render_pass.png) width=100%}\n\n`Film::update_passes` sets `shader->need_update_uvs`. The problem is that in 3.0 this is called as part of scene update after all syncing, instead of immediately after updating the render passes. That means this information is not available when we sync the geometry.\n\nNote: you can get the proper UVs back if you click on the UVMap in the Property Editor `UV Maps` panel", "pressing z key to switch between extrusion \n intel i5 ,win ,amd radeon \nblender 2.8 \n490a385c8124\n\n\n\n\nThis happens when transform orientation set to Global \nCreate a sphere , select a face and extrude \nBy default it will extrude along the normal , when pressing z it will now extrude along the global but it won't go back to normal when pressing the 'z'key again \n \n\n\n ", "Click-and-Drag the Cursor with Snapping enabled and an Active Object selected missplaces the Cursor\nOperating system: Linux-5.2.9-125.current-x86_64-with-glibc2.9 64 Bits\nGraphics card: GeForce GTX 1050/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.40\n\n\nBroken: version: 2.80 (sub 75)\nWorked: None so far\n\n\nDragging the cursor and snapping it to a Vertex while having the Target set to Active makes the cursor snap to a point in space that is directly related to where you start your *click-and-drag* instruction in reference to the selected object's origin. \n\n1) Change snapping to Vertex, Edge or Face\n2) Change snapping target to Active\n3) Set active tool to Cursor\n4) Activate snapping during transform\n5) Select default cube\n6) Drag Cursor over cube\n\n**Important notes**\n\n- If you deselect everything it works, but while there's an object selected the snapping won't work.\n\n- The problem also happens in Edit Mode, but only if there's an object selected in Object Mode. If you deselect everything in Object Mode and then go back to Edit Mode it works as intended.\n\n- As far as I can tell the _offset_ of the snapping is directly related to where you start your click-and-drag in regards to the Objects Origin.\n\n- There's a BlenderArtist Topic in which some have tested the behavior in Blender 2.81: 1177897\n\n- I recorded a video showing the problem: [blender_bug_report_complete.mkv](blender_bug_report_complete.mkv)\n\n- I also uploaded the blend file I used to record the video: [bug_report_cursor_snapping.blend](bug_report_cursor_snapping.blend)", "Unify Interpolation In Vertex Paint And Sculpt Mode\nVertex paint squares colors when it interpolates, while the paint tools in sculpt mode do not. This should really be made consistent across both modes. \n\n\n", "Inconsistencies in GPU subdivision info message\nOperating system: Windows-10-10.0.19045-SP0 64 Bits\nGraphics card: Radeon RX 570 Series ATI Technologies Inc. 4.5.0 Core Profile Context 22.11.2.221130\n\nBroken: version: 3.5.0 Alpha\nWorked: n/a\n\nThis feature is introduced with 4d3bfb3f4195ac879b7330e981829122b2633e93\n\n1. Message does not update accordingly when...\n\n1-1. enabling and disabling GPU Subdivision in the preferences\n\n\n\n1-2. changing modifier visibility\n\n\n\n1-3. enabling and disabling auto smooth\n\n\n\n2. I don't know the technical details, but I feel that \"Using both CPU and GPU Subdivision\" contradicts to what the tooltip for GPU Subdivision says (\"Enable (...) for the last subdivision subsurface modifiers in the stack\").\n\n(On a related note, I think [documentation](viewport.html#subdivision) could be more descriptive about GPU Subdivision than just \"Under certain circumstances\".)\n\n", "Tweaking redo values in extend vertices operator causes extended vertex to dissolve\nBroken in 2.80 (1d908bffddb4 from 2019-01-24) as well as 2.79b release\n\nTweaking redo values in extend vertices operator causes extended vertex (the one from which the operator is executed) to dissolve", "Texture UV/Image Editor coordinates slightly inaccurate\n**System**\nGentoo with old Advanced Micro Devices, Inc. [AMD/ATI] RV710 [Radeon HD 4350/4550]\n\nSlightly broken: media-gfx/blender-2.79b-r1\n\nWhen zooming in as much as possible on a texture in the UV/Image Editor, there seems to be a pixel border at y≈235.994 instead of exactly at y=236 like I'd expect.\n\nAssuming the error was due to 32-bit floating-point rounding, the border coordinates should have at least 2 more digits of precision.\n\n", "Edge/Vert Slide fail to properly \"redo\"\nCurrently vertex & edge-slide use the mouse input to calculate which edge to slide, the direction to slide, and the **Even** toggle is set at runtime.\n\nUsing operator redo is not re-running with the same options, so redoing the action fails.\n\nWe could simply disable redo with this operator, however being able to toggle **Correct UV** is important.\n\nAdding this report to merge in duplicates, since there are various reports on this, which all have the same root-cause.", "Support dragging to change tabs in the side-bar\nNote, this was reported as a bug however it's a TODO (wasn't supported in 2.7x).\n\n----\n\n\nOperating system: Windows-10-10.0.17763 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 430.39\n\n\nBroken: version: 2.80 (sub 60)\nWorked: (version before the click transparent passthrought)\n\n\nCannot drag on the tabs to select other tab like normal tab behavior on Blender.\n\n- Factory defaults\n- Press 'N' on the 3D View.\n- Try to drag on the tabs on the \"WM_OT_context\", the result is the box selection.\n\nSee attached gif (ignore the mouse cursor offset, that is due to 4k display with only 125% scale applied.)\n\n\n\n\n", "Multi-Button Drag: not working properly for Object Dimensions\nWindows 7\n\nBroken: 2014-02-11\n\nMulti-Button Drag - D270 - doesn't work properly in N panel for Dimensions\n\nIf you drag over all 3, only X and Z change, Y remains unchanged. As mouse button is released, the first selected property will reset, only the last hovered is kept (and scale changed)", "Snap vertices with \"Closest\" is not acting as expected\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1080/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 432.00\n\nBroken: version: 2.83.3\nWorked: 2.82a at least. Didn't check any 2.83 releases. \n\nSnapping to vertex when moving with \"Closest\" works as if \"Center\" is chosen when moving several vertices at once.[cubicle.blend](cubicle.blend)\n\n- In the attached file, go to edit mode with the selected cubic shape.\n- 3 vertices already selected, try to move them down on Z axis with snap to vertices, \"Closest\".\n- The expected behavior is that the vertex at the bottom among the 3 selected vertices, should be at the same position of the vertex I'm snapping to.\n\n#46892 (Snapping closest vertex with Pivot point set to Individual Origins)\n\n", "Adjusting multiple properties at the same time doesn't work for all property types\nOperating system: Windows-10-10.0.18362-SP0 64 Bits\nGraphics card: GeForce GTX 1070/PCIe/SSE2 NVIDIA Corporation 4.5.0 NVIDIA 456.55\n\nBroken: version: 2.90.1\n\nMost properties in blender can be changed for all selected objects by holding ALT while changing it. However, this doesn't work for all properties. For instance, the 'Rig Type' property in the Rigify add-on doesn't work. We can make it work by right click and selecting 'Copy to selected' but it should be possible with the ALT shortcut for consistency.\n\n**Steps to reproduce**\nWhile in Pose Mode select multiple bones\ngo into Bone Properties -> Rigify Type -> Rig type\ntry to change the Rig Type for all of them by holding ALT", "Adjust modifier values when using \"Apply scale\"\nThis feature request was discussed with Campbell Barton during meeting 2021-02-23\n\nWhen using \"Apply scale\", the modifier values are not adjusted if the object scale is not [1,1,1]. I would like to suggest that the operator \"Apply scale\" iterates through all modifiers of the selected objects and adjusts all values related to scale\n\nFor example\nBevel modifier - Amount\nDisplace - Strength\nData transfer - max distance, ray radius\nUV project - Scale X, Scale Y, \netc\n\n---\n\nCurrent behaviour\n{[F9849514](image.png), height = 700}\n\n---\n\nDesired behaviour\n{[F9849529](image.png), height = 700}\n\n---\n\nsuggested Operator mockup\n\n{[F9852183](image.png), height = 300}", "Asset drag & drop snapping: Limitations\nLimitations for the drag & drop snapping implemented in the `asset-browser-snap-dragging` branch.\n\n* - [x] The bounding box drawing uses bounding box based snapping, actually dropping the object uses the origin. (279f271a0a)\n* - [ ] No way to select an up-axis (e.g. for putting frames on a wall).\n* - [ ] No way to rotate the object while or after dragging.\n* - [ ] Rotation of the dragged object is not applied.\n* - [ ] Snapping uses rotation of the snapped to object.\n* - [ ] Snapping uses normals of the snapped to surface, which are often flipped. Should probably work in both positive and negative normal direction?\n* - [ ] Not implemented for collections yet." ]

[ "Snap affect settings (Move, Rotate Scale) are used in the UV Editor, but are only accessible in the 3d view.\nVersion: 2.80 Daily Build May 22nd (OS: Linux 64-bit, Hash: 8532...)\n\nThis requires no example file to reproduce.\n\n\n\nThe parameters use_snap_translate, use_snap_rotate, and use_snap_scale belong to the Scene datablock, and affect snapping behavior in both the UV editor and the 3D view. However, the snap menu in the UV editor is incomplete and does not show these settings:\n\n\n\nThis is inconsistent and confusing for users. Settings that affect tools in one window should not require going to a different window. Parameters like this should either be localized to the UV editor or made accessible within both windows." ]