repo_name
stringlengths 4
136
| issue_id
stringlengths 5
10
| text
stringlengths 37
4.84M
|
|---|---|---|
klaussilveira/gitter | 343389950 | Title: PHP Fatal error: Cannot use 'Object' as class name
Question:
username_0: PHP Fatal error: Cannot use 'Object' as class name as it is reserved in /usr/local/share/gitlist/vendor/klaussilveira/gitter/lib/Gitter/Model/Object.php on line 14
See https://github.com/klaussilveira/gitlist/issues/765 |
readsoftware/ReadIssues | 406020190 | Title: Segment Clipping Service - Load Baseline
Question:
username_0: As our workflow involves automated build of db’s (rebuild is not an issue but generating new -where we import segments- and merged db is) we have found an issue with the way READ kicks off the segment clipping service upon attempting to open a db. On large texts we have to wait until the service completes saving hundreds or even thousands of images before it will allow the text to open.
Is there a compelling reason why the service runs on open rather than only doing this when one loads the palaeography report?
Answers:
username_0: As discussed with SW, 'There is no reason to clip and re-clip before loading a baseline, this should be restricted to paleography report.'
username_0: SW
In reviewing the code, indeed, the system gets segment information and if not currently in cache the current default is to crop the segment and store it to segment_cache. This was changed due to a bug filed about the length of time for loading the paleography report. There are only two scenarios that would require you to crop:
1. First time segmentation or erasure of segment_cache files
2. Changing segment IDs without changing the labeling on cached cropped images
FYI there is a service reclipSegmentImageToCache.php that takes a list of blnIDs or a single blnID and a force (0 or 1)
where 0 will only clip if the segment cache file doesn't exist and 1 forces all segments to be re-cropped.
username_0: Understand the conditions under which the service is kicked off and as discussed, as we are generating new dbs with segmentation but sans cache, we are indeed triggering the service.
Question remains, is there a compelling reason why this is triggered upon opening a baseline rather than triggering this upon opening a paleography report,
username_1: choice was made for first time entry as front loading the calculation
juxtaposed to the same time at paleography report 1st load. Seemed more
recoverable. The notion is that this doesn't happen for the scenario where
the user creates segments as these should be created as they go (needs
testing).
Is there a way to include the reclipSegmentImageToCache when creating the
db? The alternative is to config this and shut of caching, which would
mean calling cropping each time the paleography report is loaded.
username_0: Yes we can certainly include reclipSegmentImageToCache when creating the db but this might need to run for a significant amount of time on large texts and collections. If we can config the reclipSegmentImageToCache service to either run on open baseline or on open paleography report then I think we have a more complete solution.
username_0: Thanks Steve. We'll instead rename the cached segments upon merge as you suggest.
Status: Issue closed
|
digitalbazaar/jsonld-signatures | 891938812 | Title: Verification fails for `BbsBlsSignatureProof2020` for jsonld-signatures >8.0.0
Question:
username_0: I'm using [@mattrglobal/jsonld-signatures-bbs](https://github.com/mattrglobal/jsonld-signatures-bbs) for issuing credentials and creating a zero-knowledge proof for a selectively disclosed credential. However the `verify` method of `jsonld-signatures` fails when the version of `jsonld-signatures` is >8.0.0.
I've shared an example code in [this gist](https://gist.github.com/username_0/444d6ef70c29596a343e71c66ce12532) with the keypair and the used context hardcoded inside the document. To reproduce `npm init` a new project, paste the code and install the packages via `npm install`. This should install the latest `jsonld-signatures` version which at the time of writing is `9.0.2`. Runing the file with `node index.js` throws the following error:
```js
{
verified: false,
error: VerificationError: Verification error(s).
at verify (/home/username_0/Desktop/repos/jsonld-bug-test/node_modules/jsonld-signatures/lib/jsonld-signatures.js:125:22)
at async verifySelectiveDisclosedCredential (/home/username_0/Desktop/repos/jsonld-bug-test/index.js:168:20) {
errors: [
Error: Could not verify any proofs; no proofs matched the required suite and purpose.
at ProofSet.verify (/home/username_0/Desktop/repos/jsonld-bug-test/node_modules/jsonld-signatures/lib/ProofSet.js:152:15)
at async verify (/home/username_0/Desktop/repos/jsonld-bug-test/node_modules/jsonld-signatures/lib/jsonld-signatures.js:114:18)
at async verifySelectiveDisclosedCredential (/home/username_0/Desktop/repos/jsonld-bug-test/index.js:168:20)
]
}
}
```
However when revered to an older version of `jsonld-signatures` the document gets successfully verified.
```
npm install jsonld-signatures@7^
```
and running `node index.js` outputs:
```js
Found cached https://raw.githubusercontent.com/username_0/peer-review-verifiable-credentials-thesis/main/code/PeerReview.json, not requesting
Found cached did:example:489398593#test, not requesting
Found cached did:example:489398593, not requesting
{ verified: true, results: [ { proof: [Object], verified: true } ] }
```
OS: Ubuntu 20.04.2 LTS
node version: 14.17.0
npm version: 7.13.0
Answers:
username_1: Experiencing the same at https://github.com/mattrglobal/jsonld-signatures-bbs/issues/139 and I do think I have pinned it down to a line. But I dont know what the solution shall be. Awaitng reply from the creators as well
username_0: I am using version 7.0.0 as a workaround @username_1. Only fails with version >8.0.0
username_1: @username_0 cool, thanks! I have use the one bundled with this lib, which currently is only 5, but works for now. I will look to test out 7.0.0
username_2: I took a look at this today...
The issue is caused by changes in the way suites assume contexts... recently, this package was updated to care less about sec-v2... "sec:"... those changes appear to have had some kind of sad interaction with inheritance model in the mattr library.
All that `suite.matchProof` is trying to do is string equality check really...
In our fork of this library, we were able to resolve this issue by just not using it...
```
for (const s of suites) {
const matchFound = s.type.replace("sec:", "") === proof.type;
if (matchFound) {
return s
.verifyProof({
proof,
document,
purpose,
documentLoader,
expansionMap,
compactProof,
})
.catch((error: any) => ({ verified: false, error }));
}
}
```
This is very hacky... ideally, we would update the suite and account for https://github.com/mattrglobal/jsonld-signatures-bbs/issues/142 which is related.
username_1: @username_2 nice find, so the process for this would be to solve that attached issue?
username_2: yes, I think so... IMO the mattr suite should be updated, since it now has several breaking issues with the latest version of jsonld-signatures that assumes new suite methods, such as `ensureContext` |
Zrips/CMI | 831022792 | Title: Schedule sending things to next line??
Question:
username_0: XpBoost:
Enabled: true
MinPlayers: 1
Repeat: true
PerformOn:
'1':
Day: saturday
Commands:
- check:%xpboost_hasboost%==no! check:%mvdw_world%==hardcore! xpboost give %player_name% 2 259200
- check:%localtime_time_u%==1! xpboost clear %player_name%
I have this setup in the configuration file. When I reload it, it moves the "2 259200" section to a new line and then it will not run the command. Example below:
- check:%xpboost_hasboost%==no! check:%mvdw_world%==hardcore! xpboost give %player_name%
2 259200
Answers:
username_0: Actually I did a bit more testing and none of the "check:" are working.
I had them working once setup the same way and now they're not. Does something look wrong?
Status: Issue closed
|
Einsteiny/einsteiny | 220843460 | Title: Milestone 1 feedback !
Question:
username_0: :+1: Nice work creating tasks for sprint 1. The goal of this sprint is to have the basic screens and workflows ready as much as possible. A few notes
* Tasks are well defined and seem achievable in the given timeframe
* Good work dividing tasks among team members. This is critical in working as a team and also gives equal learning opportunity to everyone involved
* Nice work creating and distributing tasks to create database schema and models.
* Consider creating a project wiki so you can have a thorough documentation of all design decisions made during the project. You have a good start on this by listing down required and advanced features on ReadMe.
* Depending on the strategy your group has come up with, consider creating regular check ins to sync up your code to avoid having multiple errors at the end.
* Consider creating tasks to QA the app a few times to check the basic flow. Integration testing will especially be important for this project since each team member is working on a separate module. The best part is seeing it all come together into a polished app.
Consider incorporating as many screens as possible in this sprint so you can have enough time to add UI polish and advanced features later on. It is those little extra things you do that make everything feel more polished and desirable to use. Look forward to what you all come up with. |
OpenNeuroOrg/openneuro | 525958104 | Title: Upgrade EKS nodes to Kubernetes 1.14
Question:
username_0: Priority on this might be lower but 1.14 has been stable on EKS for a while now and 1.12 support is discontinued upstream. This looks unlikely to require changes to our configuration but we should spin up a test cluster and deploy a staging environment to be sure.
Answers:
username_0: This has been deployed for production now, we're on 1.15 the latest EKS compatible release.
Status: Issue closed
|
postmanlabs/newman | 283443394 | Title: Basic Auth fails with Collection v2.1+Newman
Question:
username_0: 1. Newman Version (can be found via `newman -v`): 3.8.1
2. OS details (type, version, and architecture): Win 7, 64 Bit
3. Are you using Newman as a library, or via the CLI? CLI
3. Did you encounter this recently, or has this bug always been there: After the v2.1
4. Expected behaviour: Basic Auth should work in both v2.0 and v2.1
5. Command / script used to run Newman: newman run Test.postman_collection.json -e SGP_SIT.postman_environment.json -g globals.postman_globals.json -r cli,html -k --ignore-redirects
6. Sample collection, and auxiliary files (minus the sensitive details):
7. Screenshots (if applicable):
When run as Collection v2
When run as Collection v2.1
I tried setting auth at collection level and using Inherit Auth from Parent at step level, still ended up with 401 error for v2.1
Note: Everything works fine in Postman, issue is when running through Newman
<!--
Steps to reproduce the problem:
1. Export the collection as v2, trigger through Newman --> 200 Response Code
2. Re-export the same collection as v2.1, trigger through Newman --> 401 Unauthorised
-->
<!--
Before proceeding, please consider the following:
1. If this is a Postman Client App issue, go to https://github.com/postmanlabs/postman-app-support/issues/new
2. Report security issues directly to <EMAIL>. Please refrain from publicly disclosing any details about security issues in any form, until a fix has been rolled out.
3. For Postman Cloud / personal information related issues, send an email to: [<EMAIL>](mailto:<EMAIL>)
4. Verify that you have checked the documentation at https://www.getpostman.com/docs/newman_intro, and searched through previously closed issues.
5. Please ensure that all supporting files (collections or screenshots) do not contain any sensitive information.
-->
Answers:
username_1: @username_2 Could you confirm that this bug is present in Newman v3.9.1 as well?
username_2: Ahh, why didnt I think of updating Newman.
Yes, It works well with v3.9.1.
Thanks @username_1 !!
Status: Issue closed
|
farin/JCloisterZone | 608911819 | Title: Load savegame.jcz in multiplayer online
Question:
username_0: Please, can you add the possibility to load and play saved .jcz files also in multiplayer?
As always, thanks a lot
Answers:
username_1: @username_0 it's possible. Just load your game and let other to join when page with expansions is open. Just like when you creating a new game.
username_0: ehm... you mean also if we play in public server?
username_1: public server is different story, there is no "load", but anyone can continue a game he participated before
(as long as his clientId and secret in config remains unchanged)
(well there as some issues with reconnecting to game but this should be now ok)
username_0: yes, with games played in a public server, I can continue a game... but only if I controll all the other players, it isn't?
username_1: no, on public server players are still assigned to original clientId. You can control only your own players.
Did you share you config with others? Because in such case, all parties will be use same clientId and after reconnect all offline player will be assigned to a new session.
It's good idea keep clientId/secret unique for all players.
username_0: My only purpose is to end with my friends some multiplayer games in the public server that arrested for some errors (that you said that are now fixed)... so i wanted to load my backup files and tell to my friends to re-connect to that session to finish.
username_1: @username_0 unfortunately that's not easily possible. You can arrange direct connection and load game here.
Status: Issue closed
username_0: ok thanks... i did not found how to do, but ok.
Can be possible that my provider disabled this service?
username_1: @username_0 you need public IP address (which is often not true) or use tunnel (eg Hamachi VPN)
username_0: very kind... i know hamachi only for name, i will investigate
thanks |
Seterator/seterator.ru | 599016123 | Title: В проде не работает vue
Question:
username_0: **Что случилось?**
При запуске приложения в продакшн не происходит трансляция исходного кода `.vue` в `.js`.
**Как воспроизвести?**
1. Перейти на [seterator.ru](http://www.seterator.ru/)
2. Увидеть, что нет vue компонентов.
**Сейчас пока что все работает, так как был ручной перезапуск сервера, но при следующей его перезагрузке опять все навернется. Проблему нужно срочно исправить. Тем более сделать это не сложно.**
**Описание**
После решения #90 был обнаружен баг, при котором прекращается поддержка vue.
При запуске проекта в продакшн происходит трансляция исходного кода `.vue` в `wwwroot/dist/main.js`. Однако, _travi/ci_ выводит ошибку, так как происходит незарегистрированное изменение файла `wwwroot/dist/main.js`.
**Как исправить?**
Нужно добавить файл `wwwroot/dist/main.js` в `.gitignore`<issue_closed>
Status: Issue closed |
YCP-Swarm-Robotics-Capstone-2020-2021/swarm-website-visualization | 710606749 | Title: Instanced draw calls
Question:
username_0: Since a lot of what's being rendered will be the same model but with different transformations, it would improv performance to use instanced draw calls.
Answers:
username_0: This might be difficult to do with the current scene graph/render system since this would likely require the same models to be grouped together, which would remove the scene graph hierarchy and accumulative transformations. |
freakboy3742/pyxero | 495306334 | Title: Support for paymentsOnly parameter in Journals endpoint
Question:
username_0: I'm trying to pull data from the Journals endpoint on a cash basis, but when I send the request using the following:
xero.journals.filter(paymentsOnly=True)
I get the following error: "QueryParseException: No property or field 'paymentsOnly' exists in type Journal."
Does the filter method support paymentsOnly or is there a different way of pulling journals on a cash basis in the SDK?
Thanks! |
kennell/schiene | 411954669 | Title: "self.data.update & self.only_direct" in update
Question:
username_0: i use this package with home-assistant.
Since i upgraded i have following error:
"/usr/local/lib/python3.5/dist-packages/homeassistant/components/sensor/deutsche_bahn.py", line 82, in update
"/usr/local/lib/python3.5/dist-packages/homeassistant/components/sensor/deutsche_bahn.py", line 105, in update
Here is the same error reported what i had a couple times in log:
https://github.com/home-assistant/home-assistant/issues/21118
self.data.update self.only_direct
Could you please fix it? :-)
Thanks a lot!
Answers:
username_1: Duplicate of #18
Status: Issue closed
username_1: FYI: Fixed in https://github.com/username_1/schiene/pull/20, released on PyPI as version `0.23` |
Zabolekar/jezik | 286589349 | Title: Yat variants and different long yat accentuation
Question:
username_0: e ~ је
е ~ и
е̄ ~ ије
е̄ ~ је̄
е̄ ~ ӣ
Also:
ри̏јеч ~ ри̯је̑ч
вијѐнац ~ ви̯је́нац
по̏вијест ~ по̏ви̯је̄ст
All these correspondences between ekavian, two-syllable ijekavian and one-syllable jekavian pronunciations should be foreseen and supported.
Maybe a checkbox e/je/ije.
Answers:
username_1: Чекбокса пока нет, ије vs. и̯је пока нет, про тједан лучше в комментарий, прочее готово.
username_0: Yat works.
Status: Issue closed
|
bazelbuild/bazel | 730111943 | Title: `--noshow_loading_progress` still shows loading progress
Question:
username_0: ### Description of the problem / feature request:
I noticed that despite setting `--noshow_loading_progress` I still get messages like "Loading: 0 packages loaded". This sure *seems* a lot like loading progress, but it appears necessary to add `--noshow_progress` to silence these for some reason.
Taking a peak at the code, these messages appear to come from [UIStateTracker.java](https://github.com/bazelbuild/bazel/blob/3f0917a2cf0cb676598a5b21ab4c8d58b7e8d7a9/src/main/java/com/google/devtools/build/lib/runtime/UiStateTracker.java#L1082) whereas `--noshow_loading_progress` is used in [PackageFunction.java](https://github.com/bazelbuild/bazel/blob/3f0917a2cf0cb676598a5b21ab4c8d58b7e8d7a9/src/main/java/com/google/devtools/build/lib/skyframe/PackageFunction.java#L1180). Is my understanding off that it should also silence these messages?
### Feature requests: what underlying problem are you trying to solve with this feature?
Silencing Bazel spammy log messages when called as part of scripting
### What operating system are you running Bazel on?
Linux Debian
### What's the output of `bazel info release`?
3.3.1
### Have you found anything relevant by searching the web?
https://docs.bazel.build/versions/master/user-manual.html#flag--show_loading_progress
https://docs.bazel.build/versions/3.3.0/user-manual.html#flag--show_loading_progress
https://docs.bazel.build/versions/master/command-line-reference.html#flag--show_loading_progress |
CosyClub/ProjectTempo | 313806701 | Title: Congo Style Boss
Question:
username_0: A Snake-like boss where if a player kills a section then the boss splits up into smaller snakes.
Billboards will be an issue as 'connecting' them top to bottom may look very bad/strange.
One alternative is to make a mesh instead of billboards for the boss.
Another alternative is to have a Congo line of different enemies which follow the same mechanics. |
cosmos/cosmos-sdk | 508431907 | Title: ICS 05: Transfering port spec
Question:
username_0: Currently in the ICS 05 "Transferring ownership of a port" it specifies that the intention of the transfer is to transfer the ownership of a port from one module (aka owner) to another. However, in the reference code it simply copies the Port and therefore both modules ends up with ownership of the same port.
Is this intended for two or more modules to share the same port? Or we actually want to fully transfer the ownership from one owner to another?
Answers:
username_1: (transferred to the SDK as this is an implementation issue)
username_2: Closing this as it's already covered on #5191
Status: Issue closed
|
SlimeKnights/TinkersConstruct | 855206131 | Title: I'm sure you probably know this issue but there's a render issue with Chests in Station.
Question:
username_0: Crafting station renders Chests as Purple/Black boxes
Versions: Java
Minecraft: 1.12.2
Forge: 14.23.2855
Mantle: 1.12-1.3.3.55
Tinkers Construct: 1.12.2-2.13.0.183
Can it be reproduced with *just* Tinkers Construct? Honestly, I haven't tried.
If not, list the other mods *required* to reproduce the issue. The issue happens on FTB Academy, Direwolf20 mod pack
, and my custom pack.
Answers:
username_1: Duplicate of #2111. Won't be fixed in 1.12, already fixed in 1.16.
Status: Issue closed
username_0: OK, I thought so, sorry for the inconvenience... |
ether/etherpad-lite | 150759609 | Title: [Deprecated LIBS] Several warnings on deprecated libs.
Question:
username_0: How to test:
1. Download [docker file](https://github.com/ether/etherpad-docker).
2. docker build -t etherpad .
```
npm WARN deprecated [email protected]: graceful-fs version 3 and before will fail on newer node releases. Please update to graceful-fs@^4.0.0 as soon as possible.
npm WARN deprecated [email protected]: 'native-or-bluebird' is deprecated. Please use 'any-promise' instead.
npm WARN deprecated [email protected]: Jade has been renamed to pug, please install the latest version of pug instead of jade
npm WARN deprecated [email protected]: graceful-fs version 3 and before will fail on newer node releases. Please update to graceful-fs@^4.0.0 as soon as possible.
npm WARN deprecated [email protected]: lodash@<3.0.0 is no longer maintained. Upgrade to lodash@^4.0.0.
```
Answers:
username_1: These warnings appear on every etherpad instance. I'm currently updating the dependencies
username_1: See #2957
Status: Issue closed
username_1: Should be fixed |
abedra/libvault | 861400647 | Title: Vault::AWS access scope issue
Question:
username_0: I ran into a scope issue when attempting to access / construct Vault::AWS::Auth and Vault::AWS:Secrets in VaultClient.h ... while all of the methods for the nested classes (Auth and Secrets) are public, the nested classes themselves are not, and there is no typedef or factory method at the Vault::AWS level to provide a way to instantiate.
I worked around this by adding a "public" scope in Vault::AWS to make the nested classes accessible.
```
class AWS {
public: // <== Added this
class Auth {
```
Am I overlooking something on usage, or is this a bug?
Answers:
username_1: Hi @username_0, there's a chance this is a bug. I'll spend some time adding an example to the collection of examples. That's typically the fastest way to spot and fix things. I haven't had to use the AWS stuff in anger with libvault yet, so I will assume it's a mistake in the implementation for now and examine.
Status: Issue closed
username_1: This is fixed in 0.41.0 or in the latest master |
bcgov/name-examination | 301573853 | Title: Search
Question:
username_0: ----
----
#### **Epic**
##### _**Detailed Description**_
The ability to search outside the context of examining a specific name request.
##### _**Inputs**_
##### _**Outputs**_
##### _**Impacts**_
##### _**Users**_
##### _**Definition of Done**_ (:one:-Mandatory to add to the Backlog, :two:-Mandatory to add to the Sprint Backlog)
- [ ] Detailed Description :one:
- [ ] Rough Inputs :two:
- [ ] Rough Outputs :two:
- [ ] Impacts :two:
- [ ] Users :two::
- [ ] Priority Label :one:
- [ ] Attached Journey Map(s)
- [ ] Attached Relevant Screen Shots from the Existing Application :two:
----
----
#### **Story**
##### _**Story Sentence**_ (As a "user type" I want a "goal" so that "benefit".)
##### _**Inputs(Data)**_
##### _**Outputs(Data)**_
##### _**User Research Test Results**_
##### _**Definition of Done**_ (:one:-Mandatory to add to the Backlog, :two:-Mandatory to add to the Sprint Backlog)
- [ ] User Story :one:
- [ ] Priority Label :one:
- [ ] Story Label :one:
- [ ] Linked EPIC :two:
- [ ] Linked Scenarios :two:
- [ ] Inputs(Data) :one:
- [ ] Outputs(Data) :one:
- [ ] Attached Sketch/Mock-up
- [ ] User Research Test Results
----
----
#### **Scenario**
##### BDD Pseudo-Code (GIVEN "context" ... AND "more context" ... WHEN "this action occurs" ... THEN ..."result")
##### _**Definition of Done**_ (:one:-Mandatory to add to the Backlog, :two:-Mandatory to add to the Sprint Backlog)
- [ ] BDD Pseudo-Code :two:
- [ ] Estimate :two:
- [ ] Priority Label :one:
- [ ] Scenario Label :one:
- [ ] Assignee :one:
- [ ] Linked Story :one:
----
----
#### **Task** (Use for Work not Directly related to a Story but supports the Sprint Goals)
##### _**Detailed Description**_
##### _**Sprint Goal**_
##### _**Acceptance Criteria**_
##### _**Definition of Done**_ (:one:-Mandatory to add to the Backlog, :two:-Mandatory to add to the Sprint Backlog)
- [ ] Acceptance Criteria Defined :one:
- [ ] Estimate :two:
- [ ] Priority Label :one:
- [ ] Task Label :one:
- [ ] Assignee :two:
- [ ] Sprint Goal (in line with the goal of the sprint) :two:
----
----<issue_closed>
Status: Issue closed |
dv-here/randomname | 463228451 | Title: Login system is unable to contain special symbols
Question:
username_0: ## How to replicate
- Create a new account on locally hosted service with the master branch.
- blah blah
- something
## Suggested Improvements
- why to do this.
- how to do this.
Answers:
username_0: May i work on this issue.My suggested workflow would be |
microsoft/PowerToys | 650884977 | Title: UI hangs when power toys start up on boot
Question:
username_0: <!--
**Important: When reporting BSODs or security issues, DO NOT attach memory dumps, logs, or traces to Github issues**.
Instead, send dumps/traces to <EMAIL>, referencing this GitHub issue.
-->
# Environment
```
Windows build number: [run "ver" at a command prompt]
PowerToys version:
PowerToy module for which you are reporting the bug (if applicable):
```
# Steps to reproduce
<!-- A description of how to trigger this bug. -->
Set to start power toys on bootup
# Expected behavior
<!-- A description of what you're expecting, possibly containing screenshots or reference material. -->
UI should not hang and you can change the settings
# Actual behavior
<!-- What's actually happening? -->
Ui hangs and you can't switch tabs change settings or do anything
# Screenshots
<!-- If applicable, add screenshots to help explain your problem. -->
Answers:
username_1: Which UI hangs? The PowerToys settings application?
username_0: Yes, and it does not respond until you exit the power toys application and start it again.
username_1: Even on .20.1? Can you take a screenshot of what it looks like?
Anything in event logs?
Right now this isn't much to go on
username_2: Ok I am in |
elvispt/zeteticelench | 802961654 | Title: CI failing
Question:
username_0: **Describe the bug**
CI with github runner is failing
**To Reproduce**
Steps to reproduce the behavior:
1. Create a PR
2. Check Github actions.
**Expected behavior**
Should be able to run tests.
Answers:
username_0: Update to composer 2.
Status: Issue closed
|
jlippold/tweakCompatible | 569290980 | Title: `Sylph` working on iOS 13.3
Question:
username_0: ```
{
"packageId": "ch.mdaus.sylph",
"action": "working",
"userInfo": {
"arch32": false,
"packageId": "ch.mdaus.sylph",
"deviceId": "iPhone11,8",
"url": "http://cydia.saurik.com/package/ch.mdaus.sylph/",
"iOSVersion": "13.3",
"packageVersionIndexed": true,
"packageName": "Sylph",
"category": "Tweaks",
"repository": "Maxwell Dausch's Repo",
"name": "Sylph",
"installed": "1.3~Beta.1",
"packageIndexed": true,
"packageStatusExplaination": "This package version has been marked as Working based on feedback from users in the community. The current positive rating is 100% with 2 working reports.",
"id": "ch.mdaus.sylph",
"commercial": false,
"packageInstalled": true,
"tweakCompatVersion": "0.1.5",
"shortDescription": "Customize your LS music player",
"latest": "1.3~Beta.1",
"author": "<NAME>",
"packageStatus": "Working"
},
"base64": "<KEY>
"chosenStatus": "working",
"notes": ""
}
```<issue_closed>
Status: Issue closed |
shapesecurity/salvation | 138109737 | Title: local scheme (GUID) matching should be case-insensitive
Question:
username_0: Example:
```
Policy p = Parser.parse("script-src *.example.com data:", "http://example.com");
assertTrue(p.allowsScriptFromSource(new GUID("data:")));
assertTrue(p.allowsScriptFromSource(new GUID("DATA:"))); // fails now
```
Answers:
username_0: https://w3c.github.io/webappsec-csp/#match-url-to-source-expression
Status: Issue closed
|
google/closure-templates | 488279691 | Title: Can we get a release?
Question:
username_0: Sorry to file an issue for this, but it would be sweet if we could get a release cut of Soy.
Upstream, I'm working on integrating `SoySauce`/`SoyTofu` into Micronaut (see micronaut-projects/micronaut-views#13, micronaut-projects/micronaut-views#16). However, the integration needs patches here in Soy, specifically, #185 and #186, which add the `hasTemplate` method.
In any case, those patches are merged, but we don't yet have a release that contains those (it barely missed the `2019-08-22` release), so we have to ship a JAR with Micronaut, which is preventing merge of that change. Thank you in advance for considering my request
Answers:
username_1: Ok I'll release tonight
username_0: @username_1 thank you! you guys rock. happy labor day if you're in the US.
Status: Issue closed
username_1: I believe I released just now.
username_0: @username_1 thank you! |
barretlee/weibo | 118748714 | Title: #淘宝前端团队新博文# 使用 JS 构建跨平台的原生应用二React Native for Android 调试技术剖析http://t.cn/RUBc2GS 本篇文章跟大家一起深入了解了 react-native run-android 背后的执行逻辑1. RN JS 层面的打包流程 2.RN android 应用的打包流程和技术背景By Taobao FED-梧忌
Question:
username_0: <img src="http://ww1.sinaimg.cn/large/6c0378f8gw1eyd0zuvlq8j20sb0ci404.jpg"> #淘宝前端团队新博文# 《使用 JS 构建跨平台的原生应用（二）：React Native for Android 调试技术剖析》http://t.cn/RUBc2GS 本篇文章跟大家一起深入了解了 react-native run-android 背后的执行逻辑：1. RN JS 层面的打包流程 2.RN android 应用的打包流程和技术背景。By Taobao FED-梧忌 November 25, 2015 at 10:47AM via Weibo http://api.weibo.com/2/statuses/go?uid=1812166904&id=3913017894748183<issue_closed>
Status: Issue closed |
benzait27/Dauphine-Open-Data | 286821115 | Title: Projet, second
Question:
username_0: Il faut protéger beaucoup mieux vos objets contre les pointeurs null et clarifier votre API. Exemple, Person (mais aussi Planning, …). FirstName peut être null, contrairement à ce qui est indiqué dans le commentaire ! (Exemple, à l’initialisation.) emails doit être une liste, éventuellement vide, et non null. Une information String manquante peut généralement être représentée par un string vide plutôt que par null. Etc. Voir BP Null.
Encore des erreurs détectées par Eclipse (tq Redundant specification of type).
VCal
Je déconseille une méthode pour une seule ligne de code (DateUtils).
Ne pas utiliser le fuseau horaire "local", qui dépend du serveur sur lequel votre code sera installé !
planning getLectures, pas getTeachings
Votre code serait plus clair, et plus flexible, avec une méthode transformant une "lecture" en "event" (plutôt que l’imbriquer dans un lambda). Si vous aimez les streams, vous pouvez ensuite transformer le stream de "lecture" en un stream de "event" (bien qu’on n’y gagne probablement pas en lisibilité, voir https://github.com/google/guava/wiki/FunctionalExplained#functions).
Si votre liste de teachings est supposée accessible en écriture (préciser dans la javadoc de planning), votre test peut être simplifié !
Indiquer dans la javadoc (actuellement incorrecte !) si le paramètre peut être null, ou quelles sont les contraintes.
Ne pas stocker la date d’une "lecture" avec une localdatetime, car c’est un événement situé précisément dans le temps !
Test utilise packages incorrects.
VCard mêmes remarques
ObjectsXML: BP Exceptions !
SetDB1: Bon début, mais code non fonctionnel (pas de persistence.xml). Votre code source doit être encodé en UTF8. Coder en anglais.
Answers:
username_1: most of these notes were taken into consideration
Status: Issue closed
|
PyPSA/PyPSA | 870765186 | Title: Use property decorator for snapshots
Question:
username_0: <!-- Please do not post usage questions here. Ask them on the PyPSA mailing list: https://groups.google.com/forum/#!forum/pypsa -->
At the moment we are requiring to set the snapshots via the `set_snapshots` function. However the user can change the snapshots series by calling `n.snapshots = ...` which will then lead to a crash later on. A property decorator as explained in [this nice example](https://www.programiz.com/python-programming/property) prevents this.
We should also add the same for components, but there comes again the issue with flexible components...<issue_closed>
Status: Issue closed |
hovancik/stretchly | 679497817 | Title: Add an option to change the opacity of the breaks windows
Question:
username_0: ### Prerequisites
- [x] I'm using latest version
- [x] I've checked [Known issues](https://github.com/username_1/stretchly#known-issues)
- [x] I've checked [Open issues](https://github.com/username_1/stretchly/issues) (and [Closed issues](https://github.com/username_1/stretchly/issues?q=is%3Aissue+is%3Aclosed))
### Description
Sometimes a break start while I'm writing, and since the window is not transparent, I cannot finish writing down my sentence/line of code before I take my break because I cannot see what I'm writing. In this case, I don't really want to postpone my break because I only need a few seconds to finish writing.
So it could be really nice to have an option to add transparency, with the possibility to change the opacity percentage (maybe a Contributor Preferences only ?).
I already tried changing the opacity and it looks like this with an opacity at 0.5 (with Green Clouds theme):
However, depending on the background, it may be hard to read due to the white font color:
Answers:
username_1: Looks nice! Is it supported in all OSes?
username_2: Looks like something that will tempt me to continue working. What about gradually changing opacity, for example: start with 40%, than increase linearly to 100% in, say, 30 seconds.
username_0: @username_2 That's a nice idea, but I don't think we can do something like that with Electron. I will do some research tho.
username_0: I forgot something:
- Using `transparent: true` + `opacity: number` is the easiest solution, but it doesn't work on Linux.
- Using `transparent: true` + `backgroundColor: #AARRGGBB` works on every OS, but it's a more tricky solution because we need to set the alpha of backgroundColor
Altough I already have a function that can calculate the right backgroundColor:
```javascript
function calculateBackgroundColor () {
const opacity = settings.get('opacityPercent') / 100
const themeColor = settings.get('mainColor').substr(1) // remove #
return '#' + Math.round(opacity * 255).toString(16) + themeColor
}
```
username_1: So I think there are 2 possible stories here:
- break window opacity is going down as break is progressing; user is able to turn it on/off
- break window has opacity; user is able to set numeric value that represents it
My gut feeling is that it's not good idea to combine those 2. Everybody likes something else, so it would be
- either too many options to set (resulting in noone using it)
- less options, but then people would be asking me to change how it works all the time
Maybe I am wrong, but I would go with simple opacity for now :)
There is a big dream of mine, that in the future people will be able to write/script their own themes (there are few issues opened around that), so I would then outsource those things (like changing opacity with time) there
username_0: Yeah I agree, there would be too many options, so we can simply go with letting the user choose the opacity of the window.
And in case you would want to implement the fading in in the futur, there shouldn't be any need to refactor anything.
As for the properties for changing the opacity, easy solution with the `opacity` property (but doesn't work on Linux) or 'trickier' solution with the `backgroundColor` property (working on every OS) ?
username_1: Let's go with version that works everywhere :)
Could you also use `settings.get('opacity')` instead of `settings.get('opacityPercent')` so we could have standard `opacity` from `0` to `1`?
username_0: Yes of course, I was thinking that percentage could have been easier for the users to understand, but at the end of the day, it's probably the same.
Status: Issue closed
|
EvanGeissler/ENSE-400-Project-Night-Terror | 403655449 | Title: Setting Up Objectives
Question:
username_0: Need to set up the list of objectives (and side objectives) so that the player can get to them at different parts in the game. Should be actually very similar to how the collectables are set up BUT only one be displayed at a time and they will be displayed only at certain parts.
This will be split across both milestones 7 and 8 since much of the UI can be created now and some of it may not be readily available until later in the month (Feb)
Answers:
username_0: Did not see I already created this. Will close this and finish it in the other one.
Status: Issue closed
|
google/eleventy-high-performance-blog | 700271034 | Title: Guidance for base of the blog other than root
Question:
username_0: Most blogs will be under the subdirectory blog on their domain, but the repo assumes the blog will be at the root folder.
Is there any guidance/advice for folks with the above scenario? Thanks.
Answers:
username_0: I had to go into the rabbit hole of finding whats causing this, so read up the upstream eleventy starter blog code and started debugging it : his is a bug inherited from the upstream
https://github.com/11ty/eleventy-base-blog/issues/74
But after solving for this, another issue remains the srcset of the image processing account does not take into account of a pathprefix @cramforce
username_0: To have this blog template work as a subdirectory, there are quite a few changes. Just one of the code commit . We wont be posting any PR for this since the blog is not engineered for this.
https://github.com/HackerspaceMumbai/eleventy-high-performance-blog/commit/89eac6a332fe2f83084db424f40910b8927203fe
Status: Issue closed
|
Seeed-Studio/pi-hats | 520338348 | Title: Heavy traffic will
Question:
username_0: It seems to be the new driver's fault, in Aug and Sept I used to run my project with 100 Hz multiple 8 messages, it was good during 3 hour+ test. The hardware was same.
How can I roll back to the old driver?
Answers:
username_1: I'm having the same problem, but without heavy writes - It seems to happen after around 5-10 minutes of testing. The error reported is:
```
[ 1909.422685] mcp25xxfd spi0.0 can0: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
[ 1909.422758] mcp25xxfd spi0.0 can0: tefif: fifo 4 not pending - tef data: id: 0013d803 flags: 00000817, ts: 6991dd58 - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced
[ 1914.423354] mcp25xxfd spi0.0 can0: tefif: fifo 1 not pending - tef data: id: 0013e803 flags: 00000217, ts: 6991eb0f - this may be a problem with spi signal quality- try reducing spi-clock speed if this can get reproduced
```
Restarting the interface (ip link down and then up) restores functionality for a similar duration, but eventually the same error as above happens.
username_0: It seems to be the new driver's fault, in Aug and Sept I used to run my project with 100 Hz multiple 8 messages, it was good during 3 hour+ test. The hardware was same.
How can I roll back to the old driver?
username_1: I just tested with an older version of the kernel: `Linux raspberrypi 4.19.57-v7+ #1244 SMP Thu Jul 4 18:45:25 BST 2019 armv7l GNU/Linux` from the previous release of Raspbian, but the problem is also present there. I think it might be some kind of bug in the `mcp25xxfd` driver itself. Bringing the network interface (can0 / can1) down and then up is a workaround for now, but not ideal.
username_0: I've tested another board mounting MCP2515(driver version 2017) with 1300 fps message rate, 500kbps baud rate, for 17 hours, and it's OK.
username_1: I've tried the driver here: https://github.com/GBert/misc/tree/master/RPi-MCP2517 (changing the oscillator setting to 40 MHz), and that worked for around 15 hours before failing hard. After that, nothing but a full power cycle would bring it online again. At least the driver in this repository works after a software reset, if only for a short while.
username_2: Same issue with the CAN-HAT MCP2517 on latest raspbian 4.19.75-v7l, on Raspberry Pi4. It works for a while, and the suddenly stops. It seems the higher the load, the higher likelihood of getting the error.
username_2: It seems related to the fact that the MCP2517FD have now been discontinued due to a quite severe silicon bug, described in this errata document:
http://ww1.microchip.com/downloads/en/DeviceDoc/MCP2517FD-Data-Sheet-Errata-DS80000792B.pdf
They specifically mention this to be related to slower linux (raspberry pi) and a too long delay between spi bytes:
```
TX MAB underflow/RX MAB overflow due to
long delays between SPI bytes:
The SPI Interface can block the CAN FD Controller
module from accessing RAM in between SPI bytes
and between the last byte and the rising edge of
the nCS line during an SPI READ or SPI READ
CRC instruction while accessing RAM.
If the CAN FD Controller module is blocked for
more than TSPIMAXDLY, a TX MAB underflow or an
RX MAB overflow can occur.
Work around:
Keep the delay between two SPI bytes and
between the last SPI byte and the rising edge of
nCS shorter than TSPIMAXDLY; see Figure 1.
The maximum allowed delay between two bytes
depends on which CAN message frame is
transmitted and on the selected Nominal Bit Time
(NBT) and Data Bit Time (DBT). Table 1 lists
TSPIMAXDLY for the worst-case scenarios.
For example: TSPIMAXDLY is 8.5 µs for a CAN FD
frame at 500 kbps/2 Mbps. In comparison, an SPI
byte takes 0.67 µs at 12 MHz SCK. A delay of 10
times the duration of one SPI byte could cause a
TX MAB underflow. It is highly unlikely for an MCU
application to introduce such a long delay, but this
error could occur when running an operation
system like Linux on a slower MPU.
In case of a TX MAB underflow, the device will
notify the application by setting SERRIF and
MODIF and by transitioning to Restricted
Operation or Listen Only mode (depending on
CiCON.SERR2LOM). After the application
requests Normal mode, the CAN FD Controller
module will automatically attempt to retransmit the
message that caused the TX MAB underflow. It is
not necessary to reset the device.
In case of an RX MAB overflow, the device will
notify the application by setting SERRIF. The
device will remain in Normal mode. The message
that caused the RX MAB overflow will be
discarded.
```
username_2: We tested with a board using the MCP251**8**FD where the errata of the MCP2517 have been fixed. This did not solve the issue with the "Something went wrong - we got a TEF..." error.
username_3: Note, this appears to be this issue: https://github.com/msperl/linux-rpi/issues/6
username_4: We will also upgrade this product to ***MCP2518FD*** in the future
username_5: I'm bitten by this too, a real show stopper. @username_4 do you know who might resolve the PI kernel bug (if that is indeed the cause of this issue and more importantly do you know if whoever might resolve the PI kernel bug is aware of the bug?
username_4: @username_5 https://github.com/msperl/linux-rpi/issues/6 Here are some updates; I don't know if they are what you want. Besides, we are experiencing the Chinese New Year, and we will not have enough energy to devote to this matter until February 1, 2020.
username_5: Hi Baozhu,
Thanks for the prompt response. I did read all the posts on github and even tried reversing the TEf optimisations as some people seem to think these might have introduced the problem but so far have no luck getting the hardware to behave.
Wish you a great festive season and hope to hear from you re this in Feb.
kind regards,
Jan
>
username_5: Reverting the commits as suggested by @username_8 [here](https://github.com/msperl/linux-rpi/issues/6) makes seems to resolve the problem for me too on a Raspberry Pi 4. It's more or less stable after changing MCP25XXFD_SCLK_DIVIDER to 4 in mcp25xxfd_regs.h
However I do see these appear in syslog as well:
test kernel: [ 1538.699553] mcp25xxfd spi1.0: ECC double bit error at c00
test kernel: [ 1538.699695] mcp25xxfd spi1.0: unidentified system interrupt - intf = 391a1108
test kernel: [ 1538.699846] mcp25xxfd spi1.0: ECC double bit error at c00
test kernel: [ 2222.497103] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a1118
test kernel: [ 2222.497225] mcp25xxfd spi1.0: ECC double bit error at 800
This problem was resolved by applying the changes suggested by @jrm06c [here](https://github.com/msperl/linux-rpi/issues/8)
test kernel: [ 1913.359886] mcp25xxfd spi0.0: found IVMIF situation not supported by driver - bdiag = [0x00008400, 0x00005ed7]
It would be great to get this driver stable; happy to help in anyway I can.
username_5: Further debug info. I've written a simple program that forwards packets from can0 (called b) to can1 (called v) and vice versa. This runs fine for about 7 minutes (having processed ~70K frames), then single bit ECC errors are showing up. These are (most likely) corrected and all continues to run fine.
Then after ~150K frames, I start to see double ECC errors with subsequent transmit errors.
See my previous post for the changes I made to the driver.
Any suggestions?
20200126-142108: Frames received on can0 (b) will be forwarded to can1 (v)
20200126-142108: Frames received on can1 (v) will be forwarded to can0 (b)
20200126-142108: Starting notifiers
20200126-142108: Starting test
20200126-142142: Processing v queue: 25 messages
20200126-142142: Processing b queue: 8 messages
20200126-142336: Saving statistics to /data/logs/20200126-142336-test-v-stats.log
20200126-142336: Saving statistics to /data/logs/20200126-142336-test-b-stats.log
20200126-142336: Processing v queue: 24579 messages
20200126-142340: Processing b queue: 2493 messages
20200126-142707: Saving statistics to /data/logs/20200126-142707-test-v-stats.log
20200126-142707: Saving statistics to /data/logs/20200126-142707-test-b-stats.log
20200126-142707: Processing v queue: 44220 messages
20200126-142717: Processing b queue: 6007 messages
kernel: [ 7050.997314] mcp25xxfd spi1.0: ECC single bit error at c00
20200126-143028: Saving statistics to /data/logs/20200126-143028-test-v-stats.log
20200126-143028: Saving statistics to /data/logs/20200126-143028-test-b-stats.log
20200126-143028: Processing v queue: 40758 messages
20200126-143029: Processing b queue: 914 messages
kernel: [ 7297.112660] mcp25xxfd spi1.0: ECC single bit error at 800
kernel: [ 7304.042530] mcp25xxfd spi1.0: ECC single bit error at 800
20200126-143404: Saving statistics to /data/logs/20200126-143404-test-v-stats.log
20200126-143404: Saving statistics to /data/logs/20200126-143404-test-b-stats.log
20200126-143404: Processing v queue: 45941 messages
20200126-143405: Processing b queue: 643 messages
kernel: [ 7446.778260] mcp25xxfd spi1.0: ECC double bit error at c00
kernel: [ 7446.778557] mcp25xxfd spi1.0: ECC double bit error at c00
20200126-143504: b->v: CanError('Transmit buffer full')
username_6: @username_5 can you send me your test program?
username_5: @username_6, yeah sure no worries: [cantxrx.py.txt](https://github.com/Seeed-Studio/pi-hats/files/4125121/cantxrx.py.txt). Note tabstop=3 makes for easy reading.
I see you're working with Victron equipment too :-) ?
username_6: @username_5 I'm on a rapi3, using a [PiCAN FD Duo](http://skpang.co.uk/catalog/pican-fd-duo-board-with-real-time-clock-for-raspberry-pi-3-p-1568.html), which has two `mcp2517fd`, both connected to `spi0` and is probably similar to the Seed Studio's. Both lines are connected and terminated.
- What bitrate are you using in your test?
- How do you inject CAN messages into the loop?
Can you improve your python test, that it doesn't loopback the error frames that are received on a interface:
```
(2020-01-29 16:46:00.992598) can0 RX - - 002 [1] 00
(2020-01-29 16:46:00.993560) can1 RX - - 20000004 [8] 00 08 00 00 00 00 00 00 ERRORFRAME
controller-problem{tx-error-warning}
(2020-01-29 16:46:00.993705) can0 RX - - 20000004 [8] 00 08 00 00 00 00 00 00 ERRORFRAME
controller-problem{tx-error-warning}
(2020-01-29 16:46:00.993868) can1 RX - - 20000004 [8] 00 20 00 00 00 00 00 00 ERRORFRAME
controller-problem{tx-error-passive}
(2020-01-29 16:46:00.993899) can0 RX - - 20000004 [8] 00 20 00 00 00 00 00 00 ERRORFRAME
controller-problem{tx-error-passive}
(2020-01-29 16:46:00.994153) can0 RX - - 002 [1] 08
(2020-01-29 16:46:00.994184) can1 TX - - 002 [1] 08
[...]
(2020-01-29 16:46:01.001755) can0 TX - - 002 [1] 08
(2020-01-29 16:46:01.002159) can1 RX - - 002 [1] 09
(2020-01-29 16:46:01.002306) can0 TX - - 002 [1] 09
(2020-01-29 16:46:01.002798) can1 RX - - 004 [8] 00 08 00 00 00 00 00 00
```
username_5: @username_6 I'm using a bitrate of 500kbps. I've been doing two tests. For the first test I connect can0 to can1 with two wires and use cangen and candump. This test appears to run fine, at least for a day or so. Hence the need for another test that reliably reproduces the issue(s).
This other test uses my Python script on a live CAN bus with plenty of traffic (@ 500kbps) that normally connects two nodes. For my test, both nodes are connected to the Pi HAT instead of directly to each other. The Pi running my Python script functions as the wire mentioned in the first test. If this can run for an extended period of time without causing issues (either showing up in the Pi kernel log or my Python script) or in the live CAN system, we can hopefully conclude that the issues with the driver have been resolved.
I've added a test for error frames in the code although I do not see them appear (ie. makes no difference). Maybe these are not passed on by the Python CAN library.
The Seeed Studio Pi Hat also has two mcp2517fd's, but those are connected to spi0 and spi1.
After a while (sometimes 15 minutes, sometimes an hour), I see:
[47007.453669] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a1118
[47007.453792] mcp25xxfd spi1.0: ECC double bit error at 800
20200130-113800: b->v: CanError('Transmit buffer full')
This suggests a timing issue / bug in the driver causing a data corruption.
from which the driver does not recover. I can't restart it either:
# ip link set can1 type can restart
RTNETLINK answers: Invalid argument
but have to reload the kernel module.
I'm bringing the devices up at boot with:
set_can () {
ip link set $1 up type can bitrate 500000 restart-ms 100 fd off
ifconfig $1 txqueuelen 65536
ifconfig $1 up
}
username_5: Hi folks,
Back at testing the driver after having to rebuild the test system due to a failing SD card.
I've started at the beginning; ie with the driver as currently published by Seeed Studio.
Now getting a slightly different message after [this](https://github.com/Seeed-Studio/pi-hats/commit/0db233c054cdd1b37fe7d6f6c3c6ecd019742449) commit.
Kernel log filling up with heaps of these (possibly one message for each forwarded frame while running my 2nd test see above):
Feb 12 08:40:28 mitm kernel: [43531.353724] mcp25xxfd spi0.0 can0: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
Feb 12 08:40:28 mitm kernel: [43531.353831] mcp25xxfd spi1.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
Feb 12 08:40:28 mitm kernel: [43531.354085] mcp25xxfd spi1.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
Feb 12 08:40:28 mitm kernel: [43531.354318] mcp25xxfd spi1.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
Feb 12 08:40:28 mitm kernel: [43531.354561] mcp25xxfd spi1.0 can1: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
Feb 12 08:40:28 mitm kernel: [43531.354695] mcp25xxfd spi0.0 can0: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
username_6: Hey @username_5, good to see you back!
In the mean time, I've squashed my driver into a nice patch stack, which is available here:
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-2
I'd really appreciate if you give it a go.
username_5: Hey @username_6, glad to be back :-)
Still very busy with other stuff so working on this in between other things. Do you want me to test the complete kernel with driver from your repo or just the driver, ie. are there changes to other parts of the kernel that are relevant as well?
username_5: Interesting point to note: despite the kernel messages mentioned above:
Feb 12 08:40:28 mitm kernel: [43531.353724] mcp25xxfd spi0.0 can0: Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo
the live CAN system (so far) seems to be happy. I'll let it run and expect to see ECC errors showing up after a while; after all nothing has changed yet in that department.
username_6: @username_5 The driver is based on the latest v4.19-rpi kernel and includes some backports for the spi-aux driver. Using the latest stable kernel is always a good idea. So compile the whole kernel.
username_5: Update (still on current Seeed Studio driver and 4.19.97-v7l+ kernel):
Feb 12 09:16:13 test kernel: [45675.865861] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:03:43 test kernel: [59326.835049] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:03:43 test kernel: [59326.835426] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 3 to 6
Feb 12 13:14:35 test kernel: [59978.328018] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:18:49 test kernel: [60232.023712] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:19:32 test kernel: [60275.729571] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:20:58 test kernel: [60361.052802] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:21:16 test kernel: [60379.242129] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:26:37 test kernel: [60700.705055] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:26:37 test kernel: [60700.705152] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 3 to 6
Feb 12 13:32:31 test kernel: [61054.898096] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:32:31 test kernel: [61054.898234] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 3 to 6
and after increasing the traffic a bit:
Feb 12 13:53:01 test kernel: [62284.237189] mcp25xxfd spi1.0: ECC single bit error at 000
Otherwise everything still running...
username_5: As expected, the driver didn't recover after the first double bit ECC error. The "mode switches" are new... haven't seen those before. Kernel 4.19.97-v7l+ with current Seeed Studio driver.
@username_6, will now compile your kernel + driver and give that a go.
Feb 12 09:16:13 test kernel: [45675.865861] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:03:43 test kernel: [59326.835049] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:03:43 test kernel: [59326.835426] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 3 to 6
Feb 12 13:14:35 test kernel: [59978.328018] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:18:49 test kernel: [60232.023712] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:19:32 test kernel: [60275.729571] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:20:58 test kernel: [60361.052802] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:21:16 test kernel: [60379.242129] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 7 to 6
Feb 12 13:26:37 test kernel: [60700.705055] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:26:37 test kernel: [60700.705152] mcp25xxfd spi0.0: Controller unexpectedly switched from mode 3 to 6
Feb 12 13:32:31 test kernel: [61054.898096] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 13:32:31 test kernel: [61054.898234] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 3 to 6
Feb 12 13:53:01 test kernel: [62284.237189] mcp25xxfd spi1.0: ECC single bit error at 000
Feb 12 15:00:23 test kernel: [66326.870662] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.870799] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a1108
Feb 12 15:00:23 test kernel: [66326.870819] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 6 to 3
Feb 12 15:00:23 test kernel: [66326.870839] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.870979] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.871023] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.871043] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.871186] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.871230] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.871250] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.871389] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.871433] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.871452] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.871591] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.871635] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.871655] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.871793] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.871837] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.871857] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.871996] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.872040] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.872059] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.872198] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.872242] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.872261] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.872402] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.872446] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.872466] mcp25xxfd spi1.0: ECC double bit error at 000
Feb 12 15:00:23 test kernel: [66326.872605] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.872650] mcp25xxfd spi1.0: unidentified system interrupt - intf = b91a9100
Feb 12 15:00:23 test kernel: [66326.872805] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.872997] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.873190] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.873382] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.873574] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.873766] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.873958] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.874276] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
Feb 12 15:00:23 test kernel: [66326.874603] mcp25xxfd spi1.0: Controller unexpectedly switched from mode 7 to 3
username_6: @username_5 Due to the [errata](http://ww1.microchip.com/downloads/en/DeviceDoc/MCP2517FD-Silicon-Errata-and-Data-Sheet-Clarification-DS80000792A.pdf) the mcp2517fd can experience TX MAB underflows. With this error it will switch from CAN-2.0 Mode (6) to Restricted Mode (7). Then the driver brings the device back into CAN-2.0 Mode. However from the driver's point of view the transition from Restricted to CAN-2.0 (6 -> 7) is unexpected.
For yet unknown reasons, in contradiction to the errata the devive will sometimes switch to Listen Only Mode (3) instead of Restricted. This (6 -> 3) and the recovery (3 -> 6) are also not expected by the driver.
A debug version of my driver read the mode value several times and it seems the device sometimes changes first into Listen Only and then into Restricted:
`CAN-2.0 ---<TX MAB Underflow>---> Listen Only --> Restricted`
I don't know if the Listen Only -> Restricted change creates an interrupt.
I cannot explain the 7 -> 3 transition. Might be a recovery 7 -> 6, followed immediately by a TX MAB error, where the device changes into Listen Only. The interrupt handler's delay might be so high, that the device already changed to Listen Only.
Long story short, the driver doesn't take all these mode changes into account.
```
#define MCP25XXFD_CAN_CON_MODE_LISTENONLY 3
#define MCP25XXFD_CAN_CON_MODE_CAN2_0 6
#define MCP25XXFD_CAN_CON_MODE_RESTRICTED 7
```
username_5: @username_6 I wonder why I haven't seen those mode switches before in all my tests... the patch authored by @username_1 should have no effect on this. Or maybe it does...
I have compiled your kernel and driver - I presume. This is what I did:
- git clone --depth=1 --branch v4.19-rpi/mcp25xxfd-20200211-2 https://github.com/username_6/linux.git
- KERNEL=kernel7l
- make bcm2711_defconfig
- make -j4 zImage modules dtbs
- sudo make modules_install
- sudo cp arch/arm/boot/dts/*.dtb /boot/
- sudo cp arch/arm/boot/dts/overlays/*.dtb* /boot/overlays/
- sudo cp arch/arm/boot/dts/overlays/README /boot/overlays/
- sudo cp arch/arm/boot/zImage /boot/$KERNEL.img
- reboot
- uname -a
Linux test 4.19.102-v7l+ #1 SMP Wed Feb 12 18:16:40 NZDT 2020 armv7l GNU/Linux
test@test: $ sudo ./reload-interfaces.sh
rmmod: ERROR: Module mcp25xxfd_can is not currently loaded
modprobe: FATAL: Module mcp25xxfd_can not found in directory /lib/modules/4.19.102-v7l+
Cannot find device "can0"
Cannot find device "can0"
Cannot find device "can0"
Cannot find device "can1"
Cannot find device "can1"
Cannot find device "can1"
test@test: $ find /lib/modules/4.19.102-v7l+/ -iname 'mcp*' -print
/lib/modules/4.19.102-v7l+/kernel/drivers/net/can/spi/mcp251x.ko
/lib/modules/4.19.102-v7l+/kernel/drivers/iio/adc/mcp3422.ko
/lib/modules/4.19.102-v7l+/kernel/drivers/iio/adc/mcp320x.ko
It's getting late here and brain refuses to dig deeper into this after a long day :-)
What am I missing?
username_6: I think the seedstudio's overlay should work with my driver, but I haven't tested it. Post the output of `dtc -I fs /proc/device-tree` if the driver doesn't find the device.
I'm using this snippet in my `/boot/config.txt`:
```
# Seedstudio CAN-BUS FD
dtoverlay=mcp25xxfd-spi0-0,interrupt=25
dtoverlay=mcp25xxfd-spi1-0,interrupt=24
```
username_5: @username_6 Thanks for helping my poor brain :-) I'll start my test and will report back to you tomorrow morning my time, tonight your time. The standard overlay works fine.
username_5: @username_6 Test running. I can report this at this stage:
Feb 12 21:30:13 test kernel: [ 2517.408097] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
username_6: This means at least one lost CAN frame during RX. In CAN-2.0 mode the RX FIFO holds 32 CAN frames. But I've plans to support a second FIFO. Have you decreased the `spimaxfrequency` via `config.txt`?
username_5: @username_6 A bit more info from the log (omitting more RX-FIFO overflow in FIFO 2 messages; these happen for both spi1.0 (less frequent) and spi0.0 (more frequent)):
Feb 12 21:39:56 test kernel: [ 3099.960780] mcp25xxfd spi1.0 can1: Stopping tx-queue (tx_head=0x00057754, tx_tail=0x0005774c, len=8).
Feb 12 22:10:13 test kernel: [ 4917.527332] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04c8.
I have not decreased spimaxfrequency in config.txt; default Raspbian system running 4.19.102-v7l+.
I expect the double bit ECC error will show up at some time... but for now the live CAN system is happy (ie. hasn't yet detected a fault with the data).
username_5: And here's the config for the interfaces:
#!/bin/bash
set_can () {
ip link set $1 type can bitrate 500000 restart-ms 100 fd off
ip link set $1 txqueuelen 65535
ip link set $1 up
}
rmmod mcp25xxfd_can
modprobe mcp25xxfd_can
set_can can0
set_can can1
username_6: There is a race condition in the driver, TX-Complete vs. TX wakeup, fixing this properly costs 20% of performance on my rpi1, working around is much cheaper. When this happens `Stopping tx-queue` is printed. Should document this in the driver properly and make it a `netdev_dbg()`.
username_6: The interface config looks good.
username_5: Running on a RPi4 here.
username_6: I haven't seen any RX-Overflows on my rpi3. Should test on rpi4.
username_5: Just a sample from the log:
Feb 12 22:07:43 test kernel: [ 4767.274098] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:08:26 test kernel: [ 4810.320286] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:09:08 test kernel: [ 4852.307645] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:09:19 test kernel: [ 4863.316216] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:11:05 test kernel: [ 4969.581118] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:11:07 test kernel: [ 4971.384544] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:11:25 test kernel: [ 4989.101337] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:14:09 test kernel: [ 5153.567410] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:15:09 test kernel: [ 5213.427739] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:15:41 test kernel: [ 5245.817137] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:17:10 test kernel: [ 5334.505669] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:17:33 test kernel: [ 5357.172694] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:17:50 test kernel: [ 5374.217732] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:19:14 test kernel: [ 5458.775201] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:20:32 test kernel: [ 5536.241862] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:20:34 test kernel: [ 5538.943836] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:20:35 test kernel: [ 5539.743918] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:23:24 test kernel: [ 5708.776282] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:23:59 test kernel: [ 5743.762741] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:25:21 test kernel: [ 5825.246199] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:25:21 test kernel: [ 5825.545287] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:26:45 test kernel: [ 5909.529351] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:26:48 test kernel: [ 5912.633137] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:29:00 test kernel: [ 6044.136954] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:29:10 test kernel: [ 6054.345870] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 12 22:29:28 test kernel: [ 6072.093586] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 12 22:31:29 test kernel: [ 6193.554320] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
username_6: Oh, that's quite often :(
username_5: I think the driver now lost the plot, the following message is repeated many times per sec. Bedtime for me, have a good day :-)
Feb 12 22:36:46 test kernel: [ 6510.614345] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.614587] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.614758] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.614860] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.614961] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615063] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615164] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615266] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615367] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615475] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615725] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615838] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.615945] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
Feb 12 22:36:46 test kernel: [ 6510.616081] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0488.
username_5: Oh, and the system just crashed as well. I stopped the Python script but the messages in the log kept coming; maybe these were buffered but it could also be that the driver got itself in a spin from which it would not recover.
username_6: I test on my rpi4 and come back to you. gn8
username_5: @username_6 Just wondering how it's going. Anything I can do to help?
username_6: While working with a rpi3 I figured out that sometimes the 1st byte or the CAN-ID of CAN frames read from the controller has a bit flip. Here the CAN-ID should be `0x77`, but it was `0xf7`:
```
Feb 22 10:01:51 rpi3 kernel: mcp25xxfd spi1.0 can1: RX-0: offset=7, len=1, object=7/0x07, canid=000000f7
Feb 22 10:01:51 rpi3 kernel: data = f7 00 00 00 08 00 00 00
Feb 22 10:01:51 rpi3 kernel: 0d 92 91 c9 c7 c8 c9 ca
Feb 22 10:01:51 rpi3 kernel: cb cc cd ce
```
Here the 1st data byte should be `0xf5` but it was `0x75`
```
Feb 23 19:49:16 rpi3 kernel: mcp25xxfd spi0.0 can0: RX-0: data expected=f5, got=75
Feb 23 19:49:16 rpi3 kernel: data = 78 00 00 00 08 00 00 00
Feb 23 19:49:16 rpi3 kernel: 3f ef 7f ca 75 f6 f7 f8
Feb 23 19:49:16 rpi3 kernel: f9 fa fb fc
```
I eliminated that class of errors by using the CRC transfer feature on the mcp2517fd. In the above mentioned cases the CRC is broken and the transfer is retried. Sometimes the CRC is broken, but the CAN-ID and the data looks good, this seems to be the problem that is described in the mcp2517fd's errata sheet.
Further the driver now checks if the thread associated with the SPI Controller is scheduled with real time priorities - usually it isn't. The driver will print a message to the logs in the form of (where PID is replaced by the actual PID of the thread):
```
SPI thread not scheduled with RT priority, expect poor performance. Reconfigure with: 'chrt -f -p 1 <PID>'",
```
To make this permanent add this to your startup scripts:
```
chrt -f -p 1 `pidof spi0`
chrt -f -p 1 `pidof spi1`
```
Next I'll update the TX-PATH to use CRC transfers as well, maybe this helps even more.
Can you give it another short test. The current driver is available in `mcp25xxfd-20200211-3` branch:
https://github.com/username_6/linux
username_5: @username_6 Thanks for that. Test is running, so far so good (as in driver continues to run, live CAN system happy, no ECC errors). But:
Feb 28 19:21:06 test kernel: [ 1470.418732] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:08 test kernel: [ 1472.520786] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:09 test kernel: [ 1473.020464] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:14 test kernel: [ 1478.525127] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:20 test kernel: [ 1483.631749] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:25 test kernel: [ 1489.537997] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:33 test kernel: [ 1496.753161] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:46 test kernel: [ 1509.584868] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:46 test kernel: [ 1510.185041] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:49 test kernel: [ 1512.591132] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:54 test kernel: [ 1517.791889] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:55 test kernel: [ 1518.692150] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:21:59 test kernel: [ 1522.800577] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:00 test kernel: [ 1524.400587] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:22 test kernel: [ 1546.028679] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:28 test kernel: [ 1552.538172] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:33 test kernel: [ 1557.458603] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:37 test kernel: [ 1561.461031] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:56 test kernel: [ 1579.784572] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:22:56 test kernel: [ 1580.186095] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:02 test kernel: [ 1586.392456] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:11 test kernel: [ 1595.404363] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:21 test kernel: [ 1604.720108] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:52 test kernel: [ 1636.081787] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:53 test kernel: [ 1636.784686] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:55 test kernel: [ 1638.888977] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:23:59 test kernel: [ 1643.393622] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:02 test kernel: [ 1646.497012] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:06 test kernel: [ 1650.499197] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:08 test kernel: [ 1652.202024] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:16 test kernel: [ 1660.008380] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:18 test kernel: [ 1662.008685] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:18 test kernel: [ 1662.210749] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:26 test kernel: [ 1669.619292] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:42 test kernel: [ 1686.258962] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:47 test kernel: [ 1690.663138] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:48 test kernel: [ 1691.563521] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:48 test kernel: [ 1692.164678] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:24:52 test kernel: [ 1696.464965] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:00 test kernel: [ 1704.172320] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:11 test kernel: [ 1714.780924] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:12 test kernel: [ 1716.482184] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:12 test kernel: [ 1716.482276] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:40 test kernel: [ 1744.129699] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:25:48 test kernel: [ 1752.146231] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:02 test kernel: [ 1765.766387] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:02 test kernel: [ 1766.168039] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:32 test kernel: [ 1796.215899] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:39 test kernel: [ 1803.436067] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:40 test kernel: [ 1803.638517] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:52 test kernel: [ 1815.759818] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:56 test kernel: [ 1820.463122] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:26:57 test kernel: [ 1820.963154] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:27:01 test kernel: [ 1825.465106] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:27:10 test kernel: [ 1834.473137] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:27:11 test kernel: [ 1834.673665] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
Feb 28 19:27:17 test kernel: [ 1841.179927] mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2.
Feb 28 19:27:21 test kernel: [ 1844.881424] mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2.
[Truncated]
Feb 28 19:23:30 test /testd: 20200228-192330: Processing b queue: 13040 messages
Feb 28 19:23:30 test /testd: 20200228-192330: Processing v queue: 33606 messages
Feb 28 19:24:31 test /testd: 20200228-192431: Processing b queue: 13048 messages
Feb 28 19:24:32 test /testd: 20200228-192432: Processing v queue: 33621 messages
Feb 28 19:25:32 test /testd: 20200228-192532: Processing b queue: 13037 messages
Feb 28 19:25:33 test /testd: 20200228-192533: Processing v queue: 33602 messages
Feb 28 19:26:34 test /testd: 20200228-192634: Processing b queue: 13040 messages
Feb 28 19:26:34 test /testd: 20200228-192634: Processing v queue: 33634 messages
Feb 28 19:27:35 test /testd: 20200228-192735: Processing b queue: 13037 messages
Feb 28 19:27:35 test /testd: 20200228-192735: Processing v queue: 33623 messages
Feb 28 19:28:36 test /testd: 20200228-192836: Processing b queue: 13044 messages
Feb 28 19:28:37 test /testd: 20200228-192837: Processing v queue: 33636 messages
Feb 28 19:29:38 test /testd: 20200228-192938: Processing b queue: 13037 messages
Feb 28 19:29:38 test /testd: 20200228-192938: Processing v queue: 33935 messages
Feb 28 19:30:39 test /testd: 20200228-193039: Processing b queue: 13038 messages
Feb 28 19:30:39 test /testd: 20200228-193039: Processing v queue: 33636 messages
Feb 28 19:31:40 test /testd: 20200228-193140: Processing b queue: 13035 messages
Feb 28 19:31:41 test /testd: 20200228-193141: Processing v queue: 33623 messages
I'll let it run and will report back later with more.
username_5: @username_6 I read somewhere that the (Seeed Studio) driver can take a parameter to increase the number of FIFOs (but couldn't find the syntax of that parameter). Is that something worth trying? Which parameter would I pass to modprobe to do this for your driver?
username_5: @username_6 Driver crashed the system after 391825 of these in just over two minutes:
First one:
Feb 28 19:49:23 test kernel: [ 3167.256306] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
Last one:
Feb 28 19:51:36 test kernel: [ 3299.919855] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
System crashed.
username_6: How exactly? Just locked up? Or any other error message?
username_6: @username_5
The next iteration of the driver: https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-4
username_5: @username_6 Yes, I did set the SPI priorities as mentioned in your message. All 391825 messages are the same (ie. can1 and 0x0478). It looks like the driver gets stuck processing the same frame as the log messages are being generated at a very high frequency until something somewhere gets exhausted. I am not sure (I don't have a serial console connected to the Pi) what exactly happened on the RPi. My ssh sessions terminated (as if I would have logged off) and the RPi no longer responded to pings.
I'll run another test with -4 and report back.
Do you think it's somehow possible to work around the RX-MAB overflow or should we consider the MCP2517FD not fit for purpose?
username_5: @username_6 With -4, the repeated error message (1237517 in 3 mins 40 secs) in the log is: mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04a4". I also continue to see "mcp25xxfd spi0.0 can0: RX-FIFO overflow in FIFO 2" and "mcp25xxfd spi1.0 can1: RX-FIFO overflow in FIFO 2" at irregular intervals varying from 1 to 12 seconds.
The return error code is 0 for both single and double bit ECC errors. I assume mcp25xxfd_handle_eccif is repeatedly called for the same CAN frame but why and from where eludes me at the moment. The driver seizes to transmit frames once these (single bit) ECC errors are happening. But we should expect the driver to continue to transmit frames, right? Does the silicon perhaps turn the bus off after a single (or double bit for that matter) ECC error? The interface is configured with "restart-ms 100" but that might not be happening. Or perhaps once a ECC error occurs, the hardware continues to think there are ECC errors for subsequent CAN frames even if that's not the case?
Would it be possible for debugging purposes to print the contents of the offending CAN frame to syslog?
I also observe that I do not see mode changes with your driver.
The system did not crash this time. I did kill the CANTXRX process (which didn't stop the flooding of syslog but that could be a tail -f /var/log/syslog backlog), then I remove and reloaded the modules, which stopped the flood.
username_5: but the message I see is still the same. Does that mean we're not hitting the RX-MAB overflow?
username_5: Ran another test without flipping the bits of ecc_stat. I know it's wild guess ;-)
But any easy test (I changed line 1547 from "~ecc_stat" to "ecc_stat"):
`err = regmap_update_bits(priv->map, MCP25XXFD_ECCSTAT, MCP25XXFD_ECCSTAT_IF_MASK, ~ecc_stat);
`
` err = regmap_update_bits(priv->map, MCP25XXFD_ECCSTAT, MCP25XXFD_ECCSTAT_IF_MASK, ecc_stat);
`
Result: 730691 times "mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478" in the logs (and well as the usual RX-FIFO overflow in FIFO 2".
username_6: Are you sure, you're using the correct version of the driver? In https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-4, I've clarified the text in the error message.
https://github.com/username_6/linux/blob/v4.19-rpi/mcp25xxfd-20200211-4/drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c#L1336 :
```
/* If SERRIF is active, there was a RX MAB overflow. */
if (priv->regs_status.intf & MCP25XXFD_CAN_INT_SERRIF) {
if (priv->devtype_data->quirks &
MCP25XXFD_QUIRK_MAB_ERROR_NO_WARN)
netdev_dbg(priv->ndev,
"RX-%d: MAB overflow detected.\n",
ring->nr);
else
netdev_info(priv->ndev,
"RX-%d: MAB overflow detected.\n",
ring->nr);
} else {
netdev_info(priv->ndev,
"RX-%d: FIFO overflow.\n", ring->nr);
}
```
username_6: I've adjusted the text of the RX FIFO and RX MAB overflow messages, so if you still see the exact same text, make sure you've compiled and bootet the correct branch.
username_5: @username_6 Now I'm a bit confused:
$ git branch
mcp25xxfd-20200211-3
\* mcp25xxfd-20200211-4
v4.19-rpi/mcp25xxfd-20200211-2
$
$ grep "MAB overflow" drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c
/* RX MAB overflow
$
so it seems I'm on the right branch but the code change you're referring to is not there?
username_5: Note sure what happened with the branches there, so started from scratch and now I can see the changes in the source. Will run the test and report back.
username_5: Hi @username_6 so far not a single "RX-FIFO overflow in FIFO 2" messages in the log. I guess somehow mcp25xxfd-20200211-3 didn't stick either. Test running without a single error message for 30 mins now :-)
username_6: I think we have some slight progress :)
However with `mcp25xxfd-20200211-4` I've muted the RX MAB overflow and TX MAB underflow, as well as the change to Restricted Mode Error messages for the mcp2517fd. A true RX-FIFO overflow error is still printed to syslog. Memory ECC and SPI CRC errors are always shown.
If you want to see _all_ error messages, add a `#define DEBUG` before all `#include` statements in the `mcp25xxfd-core.c`, so that it looks like this. With this, the `netdev_dbg()` prints, otherwise it's quiet.
```
// CAN bus driver for Microchip 25XXFD CAN Controller with SPI Interface
//
// Copyright (c) 2019 <NAME> <<EMAIL>>
//
#define DEBUG
#include <linux/bitfield.h>
#include <linux/clk.h>
#include <linux/device.h>
#include <linux/module.h>
```
Or you can fiddle around with the `quirks` that are defined for you controller, i.e. the mcp2517:
Have a look at https://github.com/username_6/linux/blob/v4.19-rpi/mcp25xxfd-20200211-4/drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c#L36
```
/* Silence RX/TX MAB over/underlow warnings */
#define MCP25XXFD_QUIRK_MAB_ERROR_NO_WARN BIT(0)
/* Use CRC in RX-PATH */
#define MCP25XXFD_QUIRK_RX_CRC BIT(1)
/* Use CRC in TX-PATH */
#define MCP25XXFD_QUIRK_TX_CRC BIT(2)
static const struct mcp25xxfd_devtype_data mcp25xxfd_devtype_data_mcp2517fd = {
.quirks = MCP25XXFD_QUIRK_MAB_ERROR_NO_WARN | MCP25XXFD_QUIRK_RX_CRC,
.model = MCP25XXFD_MODEL_MCP2517FD,
};
```
With the mcp2517fd, errors related to MAB problems are not printed and the RX path uses a SPI transfer secured with CRC, you'll see a error message if the transfer is repeated due to a CRC error.
In my tests I see that sometimes the controller sends out broken CAN frames. I still have to debug where the data gets broken.
username_6: `/* Silence RX/TX MAB over/underlow warnings */`
... and I have to fix some typos :)
username_5: @username_6 Thanks Marc. Please see [syslog.txt](https://github.com/Seeed-Studio/pi-hats/files/4271009/syslog.txt) for the detailed syslog output of this morning's test. Note: no RX-FIFO overflows of any kind (ie. also no the RX MAB overflows). The test started running at approx 10:00.
username_5: @username_6 Our messages crossed :-) I only see now that you've muted the RX MAB overflow and TX MAB underflow. Hope the detailed dump after the unhandled interrupt will shed more light on this!
username_5: And one additional (new) message after this: Mar 1 10:58:02 test kernel: [ 3734.111341] mcp25xxfd spi0.0 can0: CRC read error: reg=0x05a8 len=20. From the output of the CANTXRX I can tell that the driver has stopped transmitting frames after the unhandled interrupt on the affected CAN bus.
username_6: `CRC read error: reg=0x05a8 len=20.` This is a CRC error while reading a mailbox. The driver will retry.
username_6: `Unhandled System Error Interrupt!` This is probably a RX MAB overflow or a TX MAB underflow. But as you see in https://github.com/username_6/linux/blob/v4.19-rpi/mcp25xxfd-20200211-4/drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c#L1603 I do quite strict test if all the usual errors bits and interrupts that come with the MAB error is active. And if it's something just slightly unknown the driver bails out for now. I'll look into the dump and try to see what's going on there.
I'll should probably remove the `return -EINVAL` in https://github.com/username_6/linux/blob/v4.19-rpi/mcp25xxfd-20200211-4/drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c#L1648 .
username_5: Ah yes, I see now that this is a CRC on the other CAN bus (can0). The driver has stopped transmitting on can1.
username_6: Was with the `mcp25xxfd-20200211-4` version?
Let's have a look at the logs:
We see a bunch of ECC errors, the problem is always at address `0x04b8`:
```
Mar 1 10:48:15 test kernel: [ 3146.956287] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04b8.
```
What's at `0x04b8`?
Luckily, due to the later unhandled SERRIF the driver dumped out the chip's contents. So this is probably _not_ the state when the ECC error showed up, but we cannot trust what we read from the RAM location `0x04b8` anyways if there is a ECC error....
```
Mar 1 10:48:15 test kernel: [ 3146.959344] TX Object: 0x05 (0x4b0) chip-HEAD priv-TAIL
Mar 1 10:48:15 test kernel: [ 3146.959349] id = 0x000001da
Mar 1 10:48:15 test kernel: [ 3146.959354] flags = 0x2af0ca08
Mar 1 10:48:15 test kernel: [ 3146.959359] SEQ_MCP2517FD = 0x000065 Sequence (MCP2517)
Mar 1 10:48:15 test kernel: [ 3146.959365] SEQ_MCP2518FD = 0x157865 Sequence (MCP2518)
Mar 1 10:48:15 test kernel: [ 3146.959368] data = 03 64 18 00 00 01 03 2b
```
So TX Object `5` starts at address `0x4b0`. So `0x4b8` will be the the first 32 bit of the payload, here `03 64 18 00`. Do you know if this makes any sense?
Unfortunately, none of the documentation tells you what to do in case of an ECC error. :(
username_5: Unfortunately I don't have the documentation explaining the contents of that CAN frame so I don't know if the data makes sense :-(
username_5: I also wonder why the driver repeats itself in the log regarding (presumably) the same ECC error / same CAN frame.
username_5: Maybe @username_7 is able to help with that? I'll ask him [here](https://github.com/msperl/linux-rpi/issues/6)
username_6: I managed to get the driver into this state:
```
Mar 1 00:17:21 rpi4 kernel: [ 2562.220387] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.220520] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
Mar 1 00:17:21 rpi4 kernel: [ 2562.220554] mcp25xxfd spi1.0 can0: Single ECC Error corrected at address 0x04d4.
Mar 1 00:17:21 rpi4 kernel: [ 2562.220603] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.220634] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
Mar 1 00:17:21 rpi4 kernel: [ 2562.220785] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.220899] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
Mar 1 00:17:21 rpi4 kernel: [ 2562.220933] mcp25xxfd spi1.0 can0: Single ECC Error corrected at address 0x04d4.
Mar 1 00:17:21 rpi4 kernel: [ 2562.220986] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.221017] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
Mar 1 00:17:21 rpi4 kernel: [ 2562.221071] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.221102] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
Mar 1 00:17:21 rpi4 kernel: [ 2562.221139] mcp25xxfd spi1.0 can0: Single ECC Error corrected at address 0x04d4.
Mar 1 00:17:21 rpi4 kernel: [ 2562.221201] mcp25xxfd spi1.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 1 00:17:21 rpi4 kernel: [ 2562.221232] mcp25xxfd spi1.0 can0: TX MAB underflow detected.
```
I've added a `#define DEBUG`, so I see the TX MAB underflow and Mode change messages. What about detecting this situation and try some sort of recovery procedure?
username_6: According to the documentation both the `DEDIF` and the `SECIF` bits (double error detection and single error correction) are not writeable, just clearable...by writing a `0`. All other are neither write nor clearable.
username_5: I'll now run another test with DEBUG and report back.
username_5: Mar 2 08:55:06 test kernel: [ 5094.083593] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0474.
Mar 2 08:55:06 test kernel: [ 5094.083749] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 2 08:55:06 test kernel: [ 5094.083781] mcp25xxfd spi1.0 can1: TX MAB underflow detected.
Mar 2 08:55:06 test kernel: [ 5094.083826] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0474.
Mar 2 08:55:06 test kernel: [ 5094.083937] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt!
Mar 2 08:55:06 test kernel: [ 5094.083943] mcp25xxfd spi1.0 can1: IRQ handler mcp25xxfd_handle_serrif() returned -22.
Mar 2 08:55:06 test kernel: [ 5094.083948] mcp25xxfd spi1.0 can1: IRQ handler returned -22.
Mar 2 08:55:06 test kernel: [ 5094.085416] mcp25xxfd spi1.0 can1: -------------------- register dump --------------------
Full log (to preserve formatting) [here](https://github.com/Seeed-Studio/pi-hats/files/4272384/syslog.txt).
# grep "08:55:06" /var/log/syslog | head -n1
Mar 2 08:55:06 test kernel: [ 5093.433509] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
# grep "08:55:06" /var/log/syslog | tail -n 1
Mar 2 08:55:06 test kernel: [ 5094.088328] 0063: mcp25xxfd_start_xmit: id=11a --- --- --- --- tx_h=00284eb0/00 ---
# grep "08:55:06" /var/log/syslog | wc -l
15608
so ... 10000 messages all printed in ~0.7 sec of which 3049 were Single bit ECC error messages. On the CAN bus I see approx 600 CAN frames per second. This, I think, suggests that the silicon and/or driver are getting stuck internally (as in not related to new CAN frames being received and thus forwarded) possibly processing the same CAN frame over and over again.
Would you agree?
username_5: Mar 2 12:45:43 test kernel: [ 180.651910] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:45:43 test kernel: [ 180.751883] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:45:52 test kernel: [ 190.361691] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:45:59 test kernel: [ 196.853164] mcp25xxfd spi0.0 can0: CRC read error: reg=0x05a8 len=20
Mar 2 12:46:05 test kernel: [ 203.389870] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:06 test kernel: [ 203.917816] mcp25xxfd spi1.0 can1: CRC read error: reg=0x06fc len=20
Mar 2 12:46:16 test kernel: [ 214.309612] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:21 test kernel: [ 219.217844] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:46:24 test kernel: [ 221.918748] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:34 test kernel: [ 231.701240] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0558 len=20
Mar 2 12:46:37 test kernel: [ 234.836867] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:39 test kernel: [ 236.740013] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:40 test kernel: [ 238.441360] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:40 test kernel: [ 238.541338] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:41 test kernel: [ 239.341493] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:43 test kernel: [ 241.044219] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:44 test kernel: [ 241.945196] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:45 test kernel: [ 242.645734] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:55 test kernel: [ 253.191045] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0620 len=20
Mar 2 12:46:55 test kernel: [ 253.356078] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 2 12:46:55 test kernel: [ 253.356288] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
Mar 2 12:46:57 test kernel: [ 255.458593] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:46:58 test kernel: [ 255.658518] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:07 test kernel: [ 264.967046] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:11 test kernel: [ 269.284865] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:23 test kernel: [ 281.003294] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:23 test kernel: [ 281.503470] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:27 test kernel: [ 285.499442] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 2 12:47:27 test kernel: [ 285.499504] mcp25xxfd spi1.0 can1: TX MAB underflow detected.
Mar 2 12:47:30 test kernel: [ 288.112074] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:31 test kernel: [ 288.911754] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 2 12:47:31 test kernel: [ 288.911817] mcp25xxfd spi1.0 can1: TX MAB underflow detected.
Mar 2 12:47:31 test kernel: [ 289.206639] mcp25xxfd spi0.0 can0: CRC read error: reg=0x065c len=20
Mar 2 12:47:38 test kernel: [ 296.521488] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:40 test kernel: [ 298.222180] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:42 test kernel: [ 299.922526] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:43 test kernel: [ 301.325543] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:45 test kernel: [ 303.027210] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:51 test kernel: [ 309.530963] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:47:59 test kernel: [ 316.739420] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:47:59 test kernel: [ 317.339474] mcp25xxfd spi1.0 can1: RX-0: MAB overflow detected.
Mar 2 12:48:03 test kernel: [ 320.940475] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0710 len=80
Mar 2 12:48:12 test kernel: [ 330.527637] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0558 len=20
Mar 2 12:48:25 test kernel: [ 343.087288] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 2 12:48:25 test kernel: [ 343.087491] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:25 test kernel: [ 343.087570] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
Mar 2 12:48:26 test kernel: [ 343.987737] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:26 test kernel: [ 344.089720] mcp25xxfd spi0.0 can0: CRC read error: reg=0x065c len=20
Mar 2 12:48:33 test kernel: [ 351.496821] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
Mar 2 12:48:33 test kernel: [ 351.496870] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
Mar 2 12:48:38 test kernel: [ 355.605365] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 2 12:48:38 test kernel: [ 355.605427] mcp25xxfd spi1.0 can1: TX MAB underflow detected.
Mar 2 12:48:43 test kernel: [ 361.493576] mcp25xxfd spi1.0 can1: CRC read error: reg=0x0648 len=20
Mar 2 12:48:46 test kernel: [ 363.961908] mcp25xxfd spi0.0 can0: CRC read error: reg=0x05a8 len=20
Mar 2 12:48:48 test kernel: [ 366.218896] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:50 test kernel: [ 367.920238] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:50 test kernel: [ 368.120165] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:51 test kernel: [ 368.920291] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:48:53 test kernel: [ 371.523089] mcp25xxfd spi0.0 can0: RX-0: MAB overflow detected.
Mar 2 12:49:05 test kernel: [ 383.310983] mcp25xxfd spi0.0 can0: CRC read error: reg=0x05e4 len=20
Only after `echo "performance" > /sys/devices/system/cpu/cpufreq/policy0/scaling_governor` did these messages stop.
The test ran for about an hour and then ended the same as the earlier test (see above for those results).
It makes me think though: could a SPI clock being too slow (TBC) for the CAN bus running at 500kbps cause CRC errors?
username_6: I think these are new errors, see above. Even if we don't clear the ECC interrupts the chip should continue to send, we would get flooded by ECC interrupts though.
username_6: For debugging purposes add `#define DEBUG` as mentioned above to my driver.
username_6: Using the `performance` governor and scheduling the SPI thread with rt-prio (that's what the `chrt` does) decrease the likelihood that the controller will run into the problem (If you want to be sure desolder the mcp2517fd and use a mcp2518fd instead.) As you will run into the problem earlier or later anyways I want to have a driver that can cope with the errors.
username_5: Does your driver report all transition changes, ie. also from Restricted back to Normal (CAN2.0) mode?
Do you think it's feasible to implement some of the (broken?) controllers functions in the driver and reset the controller (RAM, ECC bits etc) when the driver detects a TX MAB underflow to give the controller a clean start?
username_6: [No](https://github.com/username_6/linux/blob/v4.19-rpi/mcp25xxfd-20200211-4/drivers/net/can/spi/mcp25xxfd/mcp25xxfd-core.c#L1571), if the driver changes into the normal mode, no message is printed.
```
static int mcp25xxfd_handle_modif(const struct mcp25xxfd_priv *priv)
{
const u8 mode_reference = mcp25xxfd_get_normal_mode(priv);
u8 mode;
int err;
err = mcp25xxfd_chip_get_mode(priv, &mode);
if (err)
return err;
if (mode == mode_reference)
return 0;
[...]
}
```
username_6: Yes, that's the last resort. But I want to try different things first.
username_5: Thanks @username_6 Good luck. Standing by for more tests :-)
username_6: Hey @username_5,
the current recovery strategy is to stop the transmission, re-submit the TX frame to the controller and restart transmission. The problem is that this has to be coordinated with the TX path of the networking stack.
From my current tests:
```
rpi4 kernel: mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04b4.
```
```
TX Object: 0x04 (0x4a0)
id = 0x00000078
flags = 0x11edb808
SEQ_MCP2517FD = 0x00005c Sequence (MCP2517)
SEQ_MCP2518FD = 0x08f6dc Sequence (MCP2518)
data = dd de df e0 e1 e2 e3 e4
TX Object: 0x05 (0x4b0) chip-HEAD chip-TAIL chip-FIFO-full
id = 0x00000078
flags = 0x11edaac8
SEQ_MCP2517FD = 0x000055 Sequence (MCP2517)
SEQ_MCP2518FD = 0x08f6d5 Sequence (MCP2518)
data = d6 d7 d8 d9 da db dc dd
TX Object: 0x06 (0x4c0)
id = 0x00000078
flags = 0x11edac08
SEQ_MCP2517FD = 0x000056 Sequence (MCP2517)
SEQ_MCP2518FD = 0x08f6d6 Sequence (MCP2518)
data = d7 d8 d9 da db dc dd de
```
The `flags = 0x11edaac8` of TX object 5 are broken. The 0x000000c0 doesn't belong there.
username_7: @username_6 So the FDF and BRS bits are set even though you sent a CAN classic frame? Just for the record: your entire bus is CAN classic and all messages that are being sent on the bus are CAN classic? Is that the state after the correction?
username_6: Yes. What happens here is:
1. driver write CAN frame into RAM for tx
2. driver sets TXREQ+UINC
...
3. Controller signals serrif, modif, ivmif and changes to restricted operation
```
Mar 01 17:29:37 rpi4 kernel: mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 01 17:29:37 rpi4 kernel: mcp25xxfd spi1.0 can1: TX MAB underflow detected.
```
4. driver switches controller back into can-2.0 mode
5. Controller signals serrif, modif, ivmif, eccif (sec at 0x4c4) and changes to restricted operation
```
Mar 01 17:29:37 rpi4 kernel: mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
Mar 01 17:29:37 rpi4 kernel: mcp25xxfd spi1.0 can1: TX MAB underflow detected.
Mar 01 17:29:37 rpi4 kernel: mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04c4.
```
6. driver switches controller back into can-2.0 mode
7. goto 3.
After several of these loops in a certain time the driver is instructed to bail out. Then I analyse the RAM and it has the above shown contents. Each time I read this register the ECC SEC error is indicated.
username_7: @username_6
How do you provoke the error? Are you using the python script from Jan? I think I'll have to get a pi4 with your setup to reproduce this. I haven't seen it yet on mine.
Can you dump the content of the TEFs when you bail and see if everything in there is fine?
Another idea would be to read back the TXOBJ right after it was written. (Potentially with ECC disabled if errors show up)
username_6: I have to test if we can read from RAM with ECC disabled, if we have written with ECC enabled.
username_7: Hi @username_6 ,
did you have a chance to test these two?
Another idea would be to read back the TXOBJ right after it was written.
Yes, that would be best, but I'm using async spi, which needs a bit more work :)
(Potentially with ECC disabled if errors show up)
I have to test if we can read from RAM with ECC disabled, if we have written with ECC enabled.
Reading a value with ECC disabled that was written with ECC enabled "should" work for testing.
username_7: One more thing I noticed
mcp25xxfd_chip_interrupts_disable() doesn't only disable interrupts, it disables ECC completely. So anything coming in while ECC is disabled will mess up the ECC.
Why is the ECC disabled in that function?
username_6: No, I currently don't have a $CUSTOMER who's interested in the mcp2517fd, so I'm doing this in my spare time.
username_6: It's a bug, I'll fix this.
username_6: Fixed in https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-6
username_5: @username_6 Thanks Marc, for all your work on this. I'm in the same position: unpaid R&D that might pay in the future. At least it's challenging and interesting! I just received alternative hardware (CAN-USB interfaces) for an Orange Pi and will test those soon. At least it will give us an idea of whether the mcp2517fd (and maybe mcp2518fd) hardware are fit for purpose; it seems to me we're trying to work around an issue/issues in the silicon.
And of course I'm happy to test new iterations of the driver for the mcp2517fd. From what I understand though we're not expecting different results from 20200211-6, so no need for me to test at this stage?
username_6: Which CAN-USB interface are you using? The `mcp2518fd` works quite well compared to the `mcp2517fd`. No need to test the `20200211-6` it just fixes the bug found.
On the bright side, I think I've found a way to get the controller out of the ECC-in-TX-RAM situation. I'm cleaning up the patches....
username_5: I got the one shown [here](https://www.tindie.com/products/lll7/can-ethernet-converter-can-logger-linux-can) for use with an Orange Pi Zero I still had laying around although in hindsight I should probably have gotten two of [these](https://www.tindie.com/products/lll7/can-usb-converter-uccb/) as I can use those with the existing RPi. Oh well :-)
If the mcp2517fd is not fit for purpose (we got them in January this year and as far as I know the issue with the silicon had be identified by then), then I hope Seeed Studio will do the right thing and swap the mcp2517fd units for mcp2518fd units. Having said that, a workaround for the issue would be way better of course: less e-Waste.
username_5: Unfortunately the CAN-USB devices won't do the job either and trip much quicker. The driver however seem to be sustain the Transmit buffer full event better as it doesn't seize straight away (but still ends up seizing later). I'm not really surprised as this is quite likely caused by limitations off the USB implementation on the Orange Pi. Maybe I should test the single CAN-USB modules with the Rpi4 using the USB-3 ports...
What still puzzles me is that my test Python script can run for a while without issue on the RPi4 with mcp2517fd based HAT happily passing CAN frames backwards and forwards and only after having processed quite a number of frames, we're hitting this problem.
One possibiilty is that the race condition (can we call it that?) is triggered by specific data although that would be quite a worrying possibility: the driver and silicon should of course be totally data agnostic.
It could perhaps also be temperature related but I recall having run a few tests in succession so the 2nd test should crash much sooner than the 1st and it doesn't.
If not triggered by specific data nor temperature, what could it possibly be? I'll write that on a piece of paper for under my pillow, sometimes that helps :-)
username_6: Hey @username_5,
feel free to test my current branch:
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-9
You should an output like this:
```
spi1.0 can0: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +RX_CRC -TX_CRC +ECC) successfully initialized.
spi0.0 can1: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +RX_CRC -TX_CRC +ECC) successfully initialized.
```
Feel free to add the additional `#define DEBUG` and/or tweak the quirks. The current active quirks settings are reflected in the above shown output.
username_5: @username_6 Thanks :-) Got everything ready for a test tomorrow morning my time.
username_5: @username_6 test running, log shows:
test kernel: [ 776.332279] mcp25xxfd spi0.0 can0: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +RX_CRC -TX_CRC +ECC) successfully initialized.
test kernel: [ 776.345064] mcp25xxfd spi1.0 can1: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +RX_CRC -TX_CRC +ECC) successfully initialized.
test kernel: [ 975.559698] mcp25xxfd spi1.0 can1: CRC read error: reg=0x05f8 len=20
test kernel: [ 1095.827933] mcp25xxfd spi1.0 can1: CRC read error: reg=0x0648 len=20
test kernel: [ 1103.777207] mcp25xxfd spi1.0 can1: CRC read error: reg=0x06c0 len=20
test kernel: [ 1104.667691] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0698 len=20
test kernel: [ 1116.437676] mcp25xxfd spi0.0 can0: CRC read error: reg=0x0544 len=20
test kernel: [ 1126.365129] mcp25xxfd spi1.0 can1: CRC read error: reg=0x06c0 len=20
After
echo "performance" >/sys/devices/system/cpu/cpufreq/policy0/scaling_governor
the under and overflow messages become much less frequent and the CRC read error no longer shows (so far). I think I understand the relationship between the performance governor and the under and overflows but don't know what to think of the relationship between the performance governor and CRC errors.
test kernel: [ 1227.575969] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
test kernel: [ 1227.576004] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
test kernel: [ 1227.576208] mcp25xxfd spi0.0 can0: Controller changed into CAN 2.0 Mode (6).
test kernel: [ 1233.385417] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
test kernel: [ 1233.385911] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
Will let the test continue to run and report back later.
username_5: @username_6 Please see the log [here](https://github.com/Seeed-Studio/pi-hats/files/4347352/mcp25xxfd-9.log). Log entries mentioning ECC [here](https://github.com/Seeed-Studio/pi-hats/files/4347360/mcp25xxfd-ecc-9.log).
username_5: @username_6 Result of reloading the interfaces [here](https://github.com/Seeed-Studio/pi-hats/files/4347377/mcp25xxfd-reload-9.log). Note `Controller failed to reset. osc=0x00000064, reference value=0x00000460`. Perhaps to be expected after the driver bails though.
username_6: When there is an error in the IRQ handler like this:
```
Mar 18 12:03:36 test kernel: [12278.968619] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt!
Mar 18 12:03:36 test kernel: [12278.968625] mcp25xxfd spi1.0 can1: IRQ handler mcp25xxfd_handle_serrif() returned -22.
Mar 18 12:03:36 test kernel: [12278.968630] mcp25xxfd spi1.0 can1: IRQ handler returned -22.
```
... the driver dumps it's own and the chips state and shuts the chip down. During the subsequent reset, e.g. after interface down/up, the chip doesn't come up with the reset defaults.
```
Mar 18 19:39:30 test kernel: [39633.905525] mcp25xxfd spi1.0 (unnamed net_device) (uninitialized): Controller failed to reset. osc=0x00000064, reference value=0x00000460
Mar 18 19:39:30 test kernel: [39633.905532] mcp25xxfd spi1.0 (unnamed net_device) (uninitialized): Retrying to reset Controller.
```
Here the driver retries and successfully resets the chip. So this is no problem.
username_6: This all boils down to better performance, i.e. the SPI and the IRQ threads get more CPU time (because there is more) and the chip select doesn't stay asserted as long after a SPI transfer.
On the bright side, on another SoC (allwinner H3), the situation is much worse, I even see broken RAM contents on the mcp2518fd.
username_6: BTW: this is the driver recovering from the ECC-in-TX-RAM error:
```
Mar 18 09:41:40 test kernel: [ 3763.137059] mcp25xxfd spi1.0 can1: Recovering ECC Error in TX-RAM (nr=2).
```
Without this the driver would be stuck in the ECC error at the same address problem.
username_6: I've created a new branch [mcp25xxfd-20200211-10](https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-10). A ` Unhandled System Error ` isn't fatal any more, additionally the `intf` register is printed.
```
Mar 18 12:03:36 test kernel: [12278.968619] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt!
Mar 18 12:03:36 test kernel: [12278.968625] mcp25xxfd spi1.0 can1: IRQ handler mcp25xxfd_handle_serrif() returned -22.
Mar 18 12:03:36 test kernel: [12278.968630] mcp25xxfd spi1.0 can1: IRQ handler returned -22.
```
Please send me the logs if you run into a `Unhandled System Error`.
username_5: That all makes perfect sense for under and overflows. But why would we see multiple (5 in 1 minute) `CRC read error: reg=0x05f8 len=20' and much less (once every 10/20/30 mins) after switching to the performance governor? I guess a CRC error is different from an ECC error ...but perhaps there is a correlation.
There's not much else running on the RPi, CPU load very low (0.00 if cantxrx.py is not running). So it seems there is plenty of CPU time available but it seems the ondemand governor (default if I'm not mistaken) is not able to allocate that CPU time quickly enough?
username_5: @username_6 will run the test with -10 tomorrow morning my time.
username_6: I'm not that familiar with the frequency scaling and power saving in general and especially on the various rpi models.
But if a SoC is in some kind of power save mode, the main frequency might have been scaled down and the supply voltages might have been lowered, too. It all takes times to rise up the voltage again and get the SoC back to full speed. The governor controls if and when the CPU cores change their frequencies. As far as I know with the `performance` governor the CPU stays at nominal speed, unless throttled for thermal reasons. With `ondemand` the CPU frequency is scaled down and scaled up "on demand", I don't know if immediately or after some time. This probably explains the difference the governor makes.
Additionally there is the spi-clock frequency issue (at least on the rpi4), where one of the other clocks in the SoC (that is subject to CPU frequency scaling) affects the SPI clock.
On the rpi4 in addition there is this problem with spi clock frequency that you found.
username_5: @username_6 I am not sure what I'm seeing in the logs... but the script almost immediately starts complaining about "Transmit buffer full". No errors in the log as the default is "nodebug". It might be useful while we're testing to add the "#define DEBUG"... what do you think? I'm recompiling the driver now with DEBUG and will rerun the test.
username_5: @username_6 Progress! The -10 driver has been running for the whole day. 200 ECC errors, 28 CRC errors. But no transmit errors, no error frames, live CAN system has not detected any faults. Logs [here](https://github.com/Seeed-Studio/pi-hats/files/4352951/mcp25xxfd-10.log).
username_6: Looks good so far! \o/
Can you upload the full kernel log output and/or look for `Unhandled System Error Interrupt`?
username_5: @username_6 Rest assured I would have included `Unhandled System Error Interrupt` log entries if they would have occurred, which is not the case :-)
I haven't had the time to look at the changes between -9 and -10 ... but if you only made a change so the driver doesn't bail after a unhandled interrupt I wonder why the test with -9 failed and the one with -10 did not.
I am wondering if the problem is perhaps temperature related. What other reason could there be for the RPi to happily process 50000 frames/min for a number of hours before something goes wrong?
Autumn is here and today was a little bit cooler (it was still sunny for most of the day mind you). I've added CPU temp, load and memory info to the logs but unfortunately that is not revealing much, except that we have confirmation that the RPi is not throttling.
I did see one (just the one) interesting log entry though:
Mar 19 14:56:41 test kernel: [20144.670381] mcp25xxfd spi1.0 can1: Stopping tx-queue (tx_head=0x00a9e900, tx_tail=0x00a9e8f8, len=8).
There is no accompanying "Starting tx-queue" log entry but that's perhaps implicit as everything continued to run just fine.
Full log [here](https://github.com/Seeed-Studio/pi-hats/files/4353364/mcp25xxfd-full-10.log)
username_6: Additionally I did some cosmetic changes and fixed the check against `MCP25XXFD_ECC_CNT_MAX`.
This means the driver will recover after `MCP25XXFD_ECC_CNT_MAX` (by default 3) consecutive ECC errors in the same location in the TX-RAM instead of `MCP25XXFD_ECC_CNT_MAX + 1`.
username_6: Good to know.
username_6: ACK
username_5: Would it be an idea to simulate that condition in the driver to confirm that hypothesis?
Meanwhile, I'll keep running tests and will report back with anything relevant.
username_6: Keep me posted.
username_5: Mar 19 09:22:06 test kernel: [ 68.948609] mcp25xxfd spi1.0 can1: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +RX_CRC -TX_CRC +ECC) successfully initialized.
This looks PDG! Not sure if you drink beer, but you surely have earned a few craft beers :-). Wish I could forward some $ as well but we're still far away from any earnings with our project.
username_6: Yes, I enjoy craft beer, but due to the current situation here in Germany it's not appropriate to go to pubs, they are closed anyways now.
username_5: Pubs and restaurants are not closed yet here in NZ but if what's happening else where in the world is any indication of what will eventually happen here too, they should IMHO. Maybe we have to revert to having beer delivered by mail...
username_7: @username_5 could you try printing C1VEC at the time we get the unhandled system error? From reading the code it should be enough to add that register to regs_status and then it gets updated automatically in mcp25xxfd_irq(), correct @username_6 ?
username_7: The mode change is only for TXMAB underflows, how is the RX configured here, could a RXMAB overflow happen on this channel?
From what I understood this is a simple gateway forwarding everything from CAN0 to CAN1, correct?
@username_6 One more thing - you mentioned you saw the "same problems" on the MCP2518FD. Which of the problems did you mean? The SPI should be fixed on that so I wouldn't expect the problems with the timeout.
The pi hat is currently out of Stock - checking the schematic there doesn't seem to be anything sophisticated on the PCB though, so connecting other devboards to the PI4 should be just fine. Do you use any other features of the board @username_5 ?
username_6: I've ordered mine here: https://coolcomponents.co.uk/products/2-channel-can-busfd-shield-for-raspberry-pi
username_5: I am not sure what features you're referring to... can you elaborate a bit on this?
username_6: I've updated my driver to print the vec during an unhandled system error interrupt.
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-11
username_5: Thanks Marc. Will let the test run tomorrow.
username_7: @username_6 Thanks for the link - that worked
@username_5 I was planning to build up my own version of the board as it's out of stock in the official shop. The board has some more components VREGs etc. and I wanted to know whether they are needed to supply anything on the pi. Anyway, the shop that Marc pointed out still has a couple on stock so I ordered there. So, nothing left to be done.
username_5: @username_7 here you go:
[11443.455528] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt (vec=0x40400045, intf=0x3f1a9104)!
Output produced with Marc's latest driver (-11), which, so far, has been able to "self-heal" without fail.
username_6: This is consistent with the previous analyzed Unhandled System Error:
```
VEC: vec(0x018)=0x40400045
rxcode: No Interrupt (0x40)
txcode: No Interrupt (0x40)
icode: RX/TX MAB Overflow/Underflow (0x45)
INT: intf(0x01c)=0x3f1a9104
IE IF IE & IF
IVMI x Invalid Message Interrupt
WAKI Bus Wake Up Interrupt
CERRI x CAN Bus Error Interrupt
SERRI x x x System Error Interrupt
RXOVI x Receive FIFO Overflow Interrupt
TXATI x Transmit Attempt Interrupt
SPICRCI x SPI CRC Error Interrupt
ECCI x x x ECC Error Interrupt
TEFI x Transmit Event FIFO Interrupt
MODI x Mode Change Interrupt
TBCI x Time Base Counter Interrupt
RXI x Receive FIFO Interrupt
TXI Transmit FIFO Interrupt
```
Again, in contrary to the errata there is no mode change interrupt.
username_7: @username_5 thanks for testing!
@username_6 I agree, handling it like this should be fine for the moment. I'll hopefully receive my hardware soon so I can start debugging.
username_6: Hey @username_7,
See https://github.com/Seeed-Studio/pi-hats/issues/7#issuecomment-593967309 for details on my setup to reproduce the issue.
username_6: I updated the driver to accept `SERRIF` with `ECCIF` without `MODIF` to be a valid system error. There should be no Unhandled System Error Interrupts any more.
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-12
username_5: Full log [here](https://github.com/Seeed-Studio/pi-hats/files/4400772/20200330-spi-errors.txt). The driver was not able to recover from this but got stuck printing these error messages.
username_6: I have no idea what happened here. Have to look into the logs. Thanks for keep testing!
username_6: I had a quick look into the allwinner H3's spi driver. It's not as optimized as the spi drivers of the rpi, but it shows the problem of the current `mcp25xxfd` driver's SPI interface.
In Linux we have SPI messages. One SPI message consist of at least one SPI transfers. By default, when sending a SPI message, the Chip Select is asserted, all transfers are done, and at the end the Chip Select is de-asserted again.
The raspi and the H3 provide the `set_cs()` (set chip select) and `transfer_one` callbacks to the SPI core. In the `transfer_one()` callback the driver has to handle a single SPI transfer. The raspi drivers are optimized and poll for completion, if the transfer is short (otherwise an interrupt). The H3 driver always sets up an interrupt per transfer to wait for the completion. This is quite a overhead for small SPI transfers.
In the `mcp25xxfd`, I've split a single SPI message into several transfers, as this makes the CAN driver simpler, without knowing what that means for the the SPI host drivers. I'll rework the SPI access to have only one transfer per message.
username_5: Me neither :-) While the driver is getting pretty close to be useable, I hope we can iron out the exceptions that have become very rare indeed. Good luck hunting!
Maybe you can make SPI access conditional on a #define so you can choose the behaviour at compile time? Just an idea ...
username_6: Currently the driver makes only half duplex SPI transfers, meaning per transfer either RX or TX is `NULL`.
There are some half duplex only SPI controllers out there, so maybe it's a good idea to keep them.
All but the TX-Path uses regmap. So I can just create another regmap backend, one using half duplex (but having more than one transfer per message), and one using full duplex with a single transfer per message. The TX-Path is just used in one location, so it's should be not so complicated to have both variants there. It already supports "normal" and CRC'ed SPI write.
username_5: Hey Marc. Great. Do you think it would make sense to implement these conditionals as module parameters instead of compile time #defines? Same perhaps for (some of) the quirks? It would facilitate running different tests without having to recompile in between.
username_5: I caught one today:
Apr 4 11:09:44 test kernel: [407771.180240] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt (intf=0x3f1a9004)!
I'm pretty sure I'm running with the -12 driver, but here's the md5sum so you can verify that:
md5sum /opt/kernel/mkb.12/linux/drivers/net/can/spi/mcp25xxfd/mcp25xxfd.ko
fd1b28a1c1df7f8c9a2bb426d0b58fd8 /opt/kernel/mkb.12/linux/drivers/net/can/spi/mcp25xxfd/mcp25xxfd.ko
Output of "grep spi /var/log/syslog" [here](https://github.com/Seeed-Studio/pi-hats/files/4431104/unhandled-system-error.txt)
username_6: Ohh, that's only a system error:
```
INT: intf(0x01c)=0x3f1a9004
IE IF IE & IF
IVMI x Invalid Message Interrupt
WAKI Bus Wake Up Interrupt
CERRI x CAN Bus Error Interrupt
SERRI x x x System Error Interrupt
RXOVI x Receive FIFO Overflow Interrupt
TXATI x Transmit Attempt Interrupt
SPICRCI x SPI CRC Error Interrupt
ECCI x ECC Error Interrupt
TEFI x Transmit Event FIFO Interrupt
MODI x Mode Change Interrupt
TBCI x Time Base Counter Interrupt
RXI x Receive FIFO Interrupt
TXI Transmit FIFO Interrupt
```
...but directly after a ECC error in the TX-RAM:
```
Apr 4 11:09:44 test kernel: [407771.179929] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
Apr 4 11:09:44 test kernel: [407771.180146] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
Apr 4 11:09:44 test kernel: [407771.180240] mcp25xxfd spi1.0 can1: Unhandled System Error Interrupt (intf=0x3f1a9004)!
Apr 4 11:09:44 test kernel: [407771.181027] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x0478.
Apr 4 11:09:44 test kernel: [407771.181049] mcp25xxfd spi1.0 can1: Recovering ECC Error in TX-RAM (nr=1).
```
I'll update the driver to accept this.
Thanks for keeping testing.
username_6: Should be fixed in https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-13
username_5: @username_6 Thanks Marc. Will continue to run tests :-)
username_7: @username_6 I'm still trying to reproduce the issue but currently I'm not able to do so. My test setup is slightly different. I'm using a custom tool to generate a steady 100% can classic random busload on CAN0. What I see on CAN1 is oscillating around 70% busload. I'm running the python script from above and did configure the SPI thread priorities - anything else I should look for? All that against the -11 kernel.
username_6: I have the feeling that intermixing RX and TX (or rather the TEF) on the `mcp25xxfd` brings the problems. Try `canfdtest` as described in https://github.com/Seeed-Studio/pi-hats/issues/7#issuecomment-593967309.
username_7: Mhm, I think I wasn't clear enough - the 100% bus load come from an external tool, the RPI4 is only running the cantrx as described in your comment. When I slow down everything's fine and the PI is able to replicate everything on the bus. Are you generating 100% bus load with canfdtest or running it at a slower pace. So far I'm only seeing single ECC errors (which I still don't understand but I'll keep digging)
username_5: @username_7 My tests involve two CAN nodes that are both transmitting and receiving on the CAN bus. Normally they communicate directly with each other. Now CAN frames received on can0 are transmitted on can1 while CAN frames received on can1 are transmitted on can0.
username_7: @username_5 I currently have one direction only - but that's apparently too much for the pi4 to handle already, I see a few (20 in >5M RX errors) but CAN1 shows only ~70%(verified with a scope) the amount of TX'ed frames than the RX'ed on CAN0 (also with <30 TX errors). So somwhere frames get lost but the logs don't show anything (or I'm looking in the wrong location)
Do you know how much bus load your nodes approximately create?
username_5: @username_7 See below for an extract of the log here. Please note you need to add a #define DEBUG to mcp25xxfd-core.c so see all debug messages.
[39690.312976] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
[39690.313010] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
[39690.313103] mcp25xxfd spi0.0 can0: Controller changed into CAN 2.0 Mode (6).
[39879.866165] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
[39879.866199] mcp25xxfd spi1.0 can1: TX MAB underflow due to ECC error detected.
[39879.866292] mcp25xxfd spi1.0 can1: Controller changed into CAN 2.0 Mode (6).
[40128.072369] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04b8 (in TX-RAM).
[40128.072556] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04b8 (in TX-RAM).
[40128.072579] mcp25xxfd spi1.0 can1: Recovering Single ECC Error at address 0x04b8 (in TX-RAM), tx_obj=5.
[40153.660757] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
[40153.660792] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
[40153.660842] mcp25xxfd spi0.0 can0: Controller changed into CAN 2.0 Mode (6).
[40240.595249] mcp25xxfd spi1.0 can1: Single ECC Error corrected at address 0x04b8 (in TX-RAM).
[40339.643885] mcp25xxfd spi0.0 can0: Controller changed into Restricted Operation Mode (7).
[40339.643919] mcp25xxfd spi0.0 can0: TX MAB underflow detected.
[40339.643970] mcp25xxfd spi0.0 can0: Controller changed into CAN 2.0 Mode (6).
[40393.064781] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
[40393.064815] mcp25xxfd spi1.0 can1: TX MAB underflow due to ECC error detected.
[40393.064865] mcp25xxfd spi1.0 can1: Controller changed into CAN 2.0 Mode (6).
[40396.834861] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
[40396.834895] mcp25xxfd spi1.0 can1: TX MAB underflow due to ECC error detected.
[40396.834945] mcp25xxfd spi1.0 can1: Controller changed into CAN 2.0 Mode (6).
[40399.344907] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
[40399.344942] mcp25xxfd spi1.0 can1: TX MAB underflow due to ECC error detected.
[40399.344992] mcp25xxfd spi1.0 can1: Controller changed into CAN 2.0 Mode (6).
[40501.776626] mcp25xxfd spi1.0 can1: Controller changed into Restricted Operation Mode (7).
[40501.776660] mcp25xxfd spi1.0 can1: TX MAB underflow due to ECC error detected.
[40501.776710] mcp25xxfd spi1.0 can1: Controller changed into CAN 2.0 Mode (6).
[40549.862010] mcp25xxfd spi1.0 can1: Stopping tx-queue (tx_head=0x0014dc75, tx_tail=0x0014dc6d, len=8).
As it stands Marc's driver is pretty darn good. There's still once race condition (see above "274589.159543] mcp25xxfd spi1.0 can1: Transmit Event FIFO buffer empty. (seq=0x0000007f, tef_tail=0x02db08fe, tef_head=0x02db0901, tx_head=0x02db0901)") that only happens very occasionally it seems, so not so easy to reproduce.
username_5: #MODULE=mcp25xxfd_can # MS Perl / Seeed Studio
BITRATE=500000
set_can () {
ip link set $1 type can bitrate $BITRATE restart-ms 100 fd off
ip link set $1 txqueuelen 65535
ip link set $1 up
}
chrt -f -p 1 $(pidof spi0)
chrt -f -p 1 $(pidof spi1)
rmmod $MODULE
modprobe $MODULE
set_can can0
set_can can1
echo "performance" >/sys/devices/system/cpu/cpufreq/policy0/scaling_governor
username_7: So far I've only encountered Single bit ECC errors and interestingly only on CAN1 I haven't seen a single ECC error on CAN0 - looking at your logs I see the same @username_5 or did I miss anything?
I don't see any obvious differences between SPI1 and SPI0 though.
For the Busload, am I using canfdtest in a wrong way or does it generate a rather low busload? @username_6 I tried "canfdtest -vg can1" that generates only 1k frames/sec
username_6: @username_7 Yes, `canfdtest` doesn't generate a high bus load, but it interleaves RX and TX.
username_6: Regarding the problems I'm hunting on the mcp2518fd, try lowering the SPI frequency.
One test board I have has a 20 MHz quartz attached to a mcp2518fd and it shows all kinds of read-from-RAM and write-to-RAM errors that I can trigger with `canfdtest`. Lowering the frequency from 10 MHz to 5 MHz (change device tree property to `spi-max-frequency = <5000000>`) seems to help. More details to follow, once I have reliable information. I'll keep you updated.
username_5: I suspect that the interface transmitting the most traffic is showing the (most) errors. If interface can0 receives traffic at 215fps, then interface can1 transmits it at 215fps and can1 then receives traffic at 560fps which is then transmitted by can0 at 215fps.
username_5: I suspect that the interface transmitting the most traffic is showing the (most) errors. If interface can0 receives traffic at 215fps, then interface can1 transmits it at 215fps and can1 then receives traffic at 560fps which is then transmitted by can0 at 560fps.
username_6: Hello @username_7 and @username_5,
My newest driver is available here: https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-16.
It turns out the controller isn't stable when operated with the maximum allowed SPI clock in certain use cases, especially when it comes to back-to-back RX and TX. With a 20 MHz clock on the H3 it was unstable with a SPI clock of 9.375 MHz but stable with 9.09 MHz. See commit
https://github.com/username_6/linux/commit/497d7395f8ba1723cddde429f5145a545f43f13a for details. (Although the table says 10 MHz external clock, it's 20 MHz. I have a local patch fixing this.) To work around this issue the driver limits the max SPI clock to 92.5%of the external clock / 2. The 92.5% is mostly pulled out of thin air, as the H3 doesn't allow more fine grained control of the SPI clock than 9.09 MHz and 9.375 MHz.
Next I'd like to extend the table for the boards I have here: rpi1/rpi3/rpi4 and imx6/imx8mm.
username_6: update: -16 is broken, use -17
username_7: @username_6 Thanks for that - I'll test it. Quick question: Did you actually measure that the Allwinner does generate a frequency that's limited to clk/2 (just to be sure that it doesn't automatically round it up or anything)
username_5: @username_6 Thanks Marc, great work! Are both mcp2517fd and mcp2518fd unstable if SPI clock > 0.925 * ext clock/2? I recall playing with MCP25XXFD_SCLK_DIVIDER in @msperl's driver [a while ago](https://github.com/Seeed-Studio/pi-hats/issues/7#issuecomment-578457547) a while ago and did notice an improvement for values > 2. Interesting.
I'll test -17 tomorrow on the rpi4 here.
username_6: So far I just had time to test the mcp2518fd on the H3, as shown in the table in the code. Next I'll test the mcp2517fd on an imx8mm.
I'll look into the `MCP25XXFD_SCLK_DIVIDER` later. According to the datasheet, dividing the external clock doesn't fix this issue, but makes it worse, as the SPI clock (`SCK`) has to be < `SYSCLK / 2`.
username_8: @username_6: Thanks for your new driver. I was just trying it out on Allwinner H5 with an MCP2517FD and it seems to be running well until this commit https://github.com/username_6/linux/commit/34f075b349dbbf707d13c8608b4ec0c16d0b9250
From this commit on I'm not getting it to run properly, so maybe it has something to do with the introduction of full duplex SPI transfers?
The H5 is using the spi-sun6i SPI driver. I don't know if the H3 uses a different SPI driver.
Do you have any plans to reintroduce half duplex support, at least for testing purposes?
username_5: @username_6 I ran the test for most of the day including some moments with increased traffic that in the past would trip the driver. I have not seen a single ECC or CRC error, just "TX MAB underflow detected." as we'd expect. I think you may have nailed the issue, at least for the RPi4, Marc :-) ! See [here](https://github.com/Seeed-Studio/pi-hats/files/4480617/mcp2517fd-no-ecc-crc-errors.txt) for the full log.
I'll continue to try to reproduce the error as reported [here](https://github.com/Seeed-Studio/pi-hats/issues/7#issuecomment-605776739) as it would be nice to increase the robustness of the driver so it never gets into a tailspin. But of course it could well be true that it won't get there anymore when running with a slightly reduced SPI clock.
username_6: Hey @username_8,
the sunxi-6i SPI driver has a clock problem, please try https://github.com/username_6/linux/tree/v4.19-sunxi/mcp25xxfd-wip. I'll send the patches upstream soonish.
username_6: issue you mentioned in https://github.com/Seeed-Studio/pi-hats/issues/7#issuecomment-605776739.
It seems due to the issue with the too high SPI-Clock a bit flipped in the sequence number field in the RAM. (Details: The driver writes the to-be-send CAN frame and a sequence number into the TX-RAM, the chip sends the CAN frame and copies the sequence number into the TEF-RAM, and signals that the frame has been send.)
In your case the driver was expecting the sequence number `0x7e`, which are the lowest 7 bit of the `tef_tail=0x02db08fe`. But in the controller it found `seq=0x0000007f`. So I expect the lowest bit flipped while the driver was writing into the TX-RAM.
Looking at the TX-RAM here I see:
```
TX Object: 0x00 (0x460) priv-HEAD priv-TAIL priv-FIFO-empty
id = 0x00000078
flags = 0x0a851008
SEQ_MCP2517FD = 0x000008 Sequence (MCP2517)
SEQ_MCP2518FD = 0x054288 Sequence (MCP2518)
data = b8 b9 ba bb bc bd be bf
TX Object: 0x01 (0x470) chip-HEAD chip-TAIL chip-FIFO-empty
id = 0x00000078
flags = 0x0a8500a8
SEQ_MCP2517FD = 0x000000 Sequence (MCP2517)
SEQ_MCP2518FD = 0x054280 Sequence (MCP2518)
data = b1 b2 b3 b4 b5 b6 b7 b8
TX Object: 0x02 (0x480)
id = 0x00000078
flags = 0x0a850408
SEQ_MCP2517FD = 0x000002 Sequence (MCP2517)
SEQ_MCP2518FD = 0x054282 Sequence (MCP2518)
data = b2 b3 b4 b5 b6 b7 b8 b9
TX Object: 0x03 (0x490)
id = 0x00000078
flags = 0x0a850608
SEQ_MCP2517FD = 0x000003 Sequence (MCP2517)
SEQ_MCP2518FD = 0x054283 Sequence (MCP2518)
data = b3 b4 b5 b6 b7 b8 b9 ba
```
Here the TX Object `0x01` contains the wrong sequence number `0x054280`, while it should contain `0x054281`. This means the lowest bit flipped from `1` to `0`, while the driver was writing into the TX-RAM.
So with the lowered SPI-clock rate consider this problem solved. \o/
username_5: @username_6 Thanks :-) !
That all makes sense. Great work. I'll contact you via Das Geschäft to sort out some craft beer ;-)
username_6: Update https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-22
- fixed problems during ifup due to race of IRQ handler and chip reset/init
- added back half duplex mode
username_6: Update https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-23
username_6: Update https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-25
* added known good/bad SPI-clk for 40 MHz mcp2518fd on allwinner H3
```
MCP ext-clk SoC SPI SPI-clk max-clk parent-clk Status
2518 20 MHz allwinner,sun8i-h3 allwinner,sun8i-h3-spi 9090909 Hz 90.90% 600000000 Hz good
2518 20 MHz allwinner,sun8i-h3 allwinner,sun8i-h3-spi 9375000 Hz 93.75% 600000000 Hz bad
2518 40 MHz allwinner,sun8i-h3 allwinner,sun8i-h3-spi 17647058 Hz 88.24% 600000000 Hz good
2518 40 MHz allwinner,sun8i-h3 allwinner,sun8i-h3-spi 18750000 Hz 93.75% 600000000 Hz bad
2517 20 MHz fsl,imx8mm fsl,imx51-ecspi 9090909 Hz 90.09% 18181819 Hz good
2517 20 MHz fsl,imx8mm fsl,imx51-ecspi 9523809 Hz 95.34% 28571429 Hz bad
```
Turns out my 92.5% guess was not that bad :)
username_9: Hi all,
I'am new to the CAN on Raspberry Pi. I had the same Issue as mentioned in the first post. What I can see there was some code removed for the optimized transmission that was not removed completely. The was a logging statement that was originally intended to execute only if the optimized transmission failed. This statement was not removed and therfore execute on every tx that consumes a lot CPU (99% on my RasPi 4). Removing this line and the unecessary declation of two variables.
This is done in the file: mcp25xxfd_can_tx.c
```
int mcp25xxfd_can_tx_handle_int_tefif(struct mcp25xxfd_can_priv *cpriv)
{
/* unsigned long flags;
u32 finished; */
if (!(cpriv->status.intf & MCP25XXFD_CAN_INT_TEFIF))
return 0;
MCP25XXFD_DEBUGFS_STATS_INCR(cpriv, int_tef_count);
spin_lock_irqsave(&cpriv->fifos.tx_queue->lock, flags);
/* compute finished fifos and clear them immediately */
finished = (cpriv->fifos.tx_queue->in_can_transfer ^
cpriv->status.txreq) &
cpriv->fifos.tx_queue->in_can_transfer;
spin_unlock_irqrestore(&cpriv->fifos.tx_queue->lock, flags);
/* run in optimized mode if possible */
if (finished)
return mcp25xxfd_can_tx_handle_int_tefif_optimized(cpriv,
finished);
/* otherwise play it safe */
/*
netdev_warn(cpriv->can.dev,
"Something is wrong - we got a TEF interrupt but we were not able to detect a finished fifo\n");*/
return mcp25xxfd_can_tx_handle_int_tefif_conservative(cpriv);
}
```
Best regards,
Volker
username_6: Hey @username_9,
thanks for you input. That driver is not developed active any more. Please use the driver available in my repo: https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-25 (or in the future the one with the newwest timestamp and highest number).
username_6: Update https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200211-27
* use CRC secured register access by default (tnx @username_7)
username_2: @username_6 i would like to test your driver. But I don't know how to install it. Do you have a short description, that maybe does not involve building your entire kernel tree on a raspberry pi.
username_5: git clone --depth=1 --branch v4.19-rpi/mcp25xxfd-20200211-27 https://github.com/username_6/linux.git
cd linux
KERNEL=kernel7l
make bcm2711_defconfig
make -j4 zImage modules dtbs
sudo make modules_install
sudo cp arch/arm/boot/dts/*.dtb /boot/
sudo cp arch/arm/boot/dts/overlays/*.dtb* /boot/overlays/
sudo cp arch/arm/boot/dts/overlays/README /boot/overlays/
sudo cp arch/arm/boot/zImage /boot/$KERNEL.img
Now reboot and you're running the kernel with the required/recommended SPI patches as well as Marc's driver.
username_5: @username_6 Is half-duplex mode enabled or disabled by default? In other words, would I see any difference in behaviour between -17 and -27 in that sense?
username_5: @username_6 Driver -27 fails to load on the RPi4 with:
Apr 30 11:15:53 test kernel: [ 131.464702] mcp25xxfd: probe of spi1.0 failed with error -22
Apr 30 11:15:53 test kernel: [ 131.504698] mcp25xxfd: probe of spi0.0 failed with error -22
Reverted back to -21, which works fine.
username_6: Hey @username_5,
thanks for the report, the problem is fixed. Meanwhile I've rebased to the latest raspi v4.19 kernel (_note_: the branch has a new timestamp) and backported some SPI improvements, reducing the overall time of a TX message.
https://github.com/username_6/linux/tree/v4.19-sunxi/mcp25xxfd-20200429-34
username_5: @username_6 Somehow you lost bcm2711_defconfig in this tree:
*** Can't find default configuration "arch/arm/configs/bcm2711_defconfig"!
No problem; I copied it from the previous version but you may want to add it back in :-)
username_6: Hey @username_5,
handling too many trees right now, the raspi tree is:
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200429-35
The above wrong link was for the `sunxi` platform.
Sorry for the noise :)
username_5: Hey @username_6,
I guess I could have spotted that. Too quick with my copy / paste ;-)
-35 up & running now.
Thanks for all your work on this!
username_6: Hello,
some updates for you:
https://github.com/username_6/linux/tree/v5.6-rpi/mcp25xxfd-20200429-36
https://github.com/username_6/linux/tree/v4.19-rpi/mcp25xxfd-20200429-36
* fixed mem leak with enabled berr reporting
* reduced time spend in TX SPI transfer
username_4: @username_6 Thank you. We will update it as soon as possible, so that users can use your latest software directly. Let's test and perfect it together.
username_7: @username_6 I tested -36 for a bit now and it looks really stable - I didn't see any issues in either classic or FD mode. Will continue testing.
username_10: Hi @username_6 . I wasn't sure if it was appropriate to comment here or on your github, but seeing as this is active I will comment here. I am currently using the MCP2515 on a Allwinner V3S (essentially a single core H3, identical SPI driver). Currently we are working on a revision and the new board design accepts either MCP2515 or MCP2518FD - with the hope of utilizing a higher SPI frequency (to save time for other devices on the SPI bus. e.g. display)
From reading your notes above, if I have a 40Mhz clock on the MCP2518FD, the appropriate SPI frequency would be ~17.64MHz? What are the increments on the H3 driver, and what is the limitation here?
I am currently using buildroot with mainline 5.3.13, its it possible I can package your driver as an out of tree driver (e.g. build patch against mainline /drivers)? Is there anything that I need to be aware of here, I guess I need to use the sunxi branch in your repository? any other advice?
Thankyou
username_6: Hey @username_10,
I hijacked this issue while writing my driver, so better ask on my github next time :)
There are two limitations on the usable SPI clock speed. One is a yet unknown issue with the mcp25xxfd: it doesn't properly work with `SYSCLOCK / 2` speed. Which would be `40 MHz / 2 = 20 MHz` in your case. The other is the limitation of the `sun6i` SPI IP core and the H3's clock tree. The clock tree can provide 600 MHz and you can use multiple-of-two as a divider. With experiments I figured out that `600 MHz / 32 = 18.75 MHz` is unstable and `600 MHz / 34 = 17.65 MHz` seems to be stable.
Yes, you can package my driver as an external one, or port the patches to your branch. Note: you'll need the sun6i spi patches, too. Drop me a note if you need commercial support for this.
username_11: Hey @username_6,
We have problem when using your version of kernel:
```
[ 6.799566] mcp25xxfd spi0.0 (unnamed net_device) (uninitialized): CRC read error at address 0x0000, length 4.
[ 6.799693] mcp25xxfd spi0.0 (unnamed net_device) (uninitialized): CRC read error at address 0x0000, length 4.
[ 6.866262] mcp25xxfd: probe of spi0.0 failed with error -74
[ 6.877261] mcp25xxfd spi1.0 can0: MCP2517FD rev0.0 (-RX_INT +MAB_NO_WARN +CRC_REG +CRC_RX +CRC_TX +ECC -HD m:20.00MHz r:18.50MHz e:17.85MHz) successfully initialized.
[ 6.894772] mcp25xxfd spi1.0 can1: renamed from can0
```
Any advice on this? We are using RPi 3B+ and SeedStudio.
username_6: Hey @username_11,
which kernel are you exactly using?
username_11: I have used this branch: `v4.19-rpi/mcp25xxfd-20200429-36`
`uname -a`:
```
Linux seamount-201 4.19.118-v7+ #1 SMP Tue Jun 30 13:00:09 CEST 2020 armv7l GNU/Linux
```
username_6: Use something more recent, like https://github.com/username_6/linux/tree/v5.7-rpi/mcp25xxfd-2020618-44
The error message `probe of spi0.0 failed with error -74` indicates that there is a communication problem with one of the mcp2517fd attached to `spi0.0`. Please check the electrical connection.
username_11: @username_6 Thanks for reply.
I've also checked `v4.19-rpi/mcp25xxfd-20200429-41` but there was the same error as above. And this is no electrical connection problem because in version `v4.19-rpi/mcp25xxfd-20200211-21` it works good.
However with v21 there is a problem that when CAN receiver is not connected and we send frames (~1000/sec) RPi hangs after some time (20-30seconds). Do you have any ideas what can we check?
Also could you provide some instructions how can I compile kernel v5.7? (unfortunately instructions that works for kernel 4.19 does not work here)
username_8: @username_6: I tried the latest sunxi -46 branch but still can't get it to run reliable on my hardware.
The strange thing is, it works fine for a few minutes (with a few thousand frames received and transmitted) and then the driver suddenly stops receiving and transmitting frames.
Even with debug printing enabled and also the additional logging option in Kconfig I don't see anything in dmesg.
And also no error counters going up in ifconfig.
On the other hand, msperls original driver runs just fine (with his queue optimizations removed).
Any idea what I can do to diagnose the issue further?
username_6: Hey @username_8,
are you using the proper IRQ type? The correct one is probably `IRQ_TYPE_LEVEL_LOW`. See:
https://github.com/username_6/linux/commit/fdce9f56124dc3bb108f8d73861107677337cbdf#diff-b0098f47cd6958a4951df22fb1074161R60
username_6: Hey @username_11,
the driver fails with CRC error, you can disable CRC check by removing:
`MCP25XXFD_QUIRK_CRC_REG | MCP25XXFD_QUIRK_CRC_RX | MCP25XXFD_QUIRK_CRC_TX` for all three `mcp25xxfd_devtype_data_mcp2517fd` `mcp25xxfd_devtype_data_mcp2518fd` and `mcp25xxfd_devtype_data_mcp25xxfd` at
https://github.com/username_6/linux/commit/b5dc9fe1cc5e64f81af468316135137b14d09515#diff-e32dfa5339ecf2e91cb25021a6d7dd0eR30
username_8: @username_6: Seems like this did the trick. I still had this set on falling edge. Thanks!
username_10: I am using @username_6's -47 release that was pushed to linux-can-next for mainline:
https://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next.git/log/?h=mcp25xxfd-47
After a few patching issues due to the driver being broken up and some changed flags in device tree I thought I was getting some success:
`[ 1.255123] CAN device driver interface
[ 1.259309] spi_master spi0: will run message pump with realtime priority
[ 1.304566] mcp25xxfd spi0.1 can0: MCP2518FD rev0.0 (-RX_INT -MAB_NO_WARN +CRC_REG +CRC_RX +CRC_TX +ECC -HD m:8.82MHz r:8.82MHz e:8.33MHz) successfully initialized.`
However that was short lived; I couldn't receive a single message (candump 500k) without getting a crc error:
`[ 48.469759] mcp25xxfd spi0.1 can0: CRC read error at address 0x001c (length=4, data=00 00 1a 3f, CRC=0x1e7c).
[ 48.479730] mcp25xxfd spi0.1 can0: IRQ handler returned -74 (intf=0x3f1a0002).`
I used the advice and disabled the CRC flags which got me some results (I was able to load-test once at around 15% bus utilisation at 500k), however now I am getting persistent errors after a few packets:
`[ 382.629705] mcp25xxfd spi0.1 can0: RX tail of chip (26) and ours (27) inconsistent.
[ 382.637422] mcp25xxfd spi0.1 can0: IRQ handler mcp25xxfd_handle_rxif() returned -84.
[ 382.645189] mcp25xxfd spi0.1 can0: IRQ handler returned -84 (intf=0x3e1a0002).
`
It seems horribly inconsistent, sometimes high busload is ok, sometimes anything more than 250ms in-between will fault the driver. This seems related to ECC, next build I think I might try and disable this and see if I have any more luck.
Anyone got any more thoughts?
username_10: Just to follow up: building the driver without EEC support didn't help. Still get RX tail errors.
again, wildly inconsistent and an ip link down / up might only keep it going for a few more frames.
username_12: I'm using https://github.com/username_6/linux/tree/v5.4-rpi/mcp251xfd-20201022-55, and the canbus fails under heavy load (a few thousands messages/s ). This is what I read on dmesg:
```
[ 2338.525710] mcp251xfd spi1.0 can1: CRC error
[ 2338.525728] mcp251xfd spi1.0 can1: Stuff error
[ 2345.150009] mcp251xfd spi1.0 can1: CRC error
[ 2345.150028] mcp251xfd spi1.0 can1: Format error
[ 2347.707524] mcp251xfd spi1.0 can1: Stuff error
[ 2361.197609] mcp251xfd spi1.0 can1: Bit0 error
[ 2361.197768] mcp251xfd spi1.0 can1: New error state: 2
[ 2361.198138] mcp251xfd spi1.0 can1: Bit0 error
[ 2361.198526] mcp251xfd spi1.0 can1: Bit0 error
[ 2361.198910] mcp251xfd spi1.0 can1: Bit0 error
[ 2361.199060] mcp251xfd spi1.0 can1: New error state: 3
[ 2361.199384] mcp251xfd spi1.0 can1: bus-off
```
```
pi@raspberrypi:~/canvi/linux $ ip link show can1
4: can1: <NO-CARRIER,NOARP,UP,ECHO> mtu 16 qdisc pfifo_fast state DOWN mode DEFAULT group default qlen 100
link/can
```
Should I use some other version of the driver? Did anyone manage to get it working reliably on some other version? Could I be missing something really stupid? (I am not familiar at all with low level stuff)
username_6: Hey @username_12,
meanwhile you can use the official rpi-5.4.y branch https://github.com/raspberrypi/linux/tree/rpi-5.4.y, however the official overlay is a bit different now. However the driver is the same.
Please describe your test setup, used CAN controllers and raspberries, who is sender/reciever and send messages more closely. Also please check the cabling of the CAN bus. Keep in mind that you cannot send the same message from two CAN controllers over the same bus (at the same time), you'll see these strange messages.
username_12: I got my hands on a brand new raspberry pi 4, a brand new seed can shield and after a clean install everything works perfectly!
I guess the problem was on how I configured my older raspberry, which unfortunately I don't have access to at the moment. Thank you for your patience. |
dotnet/orleans | 118157525 | Title: Can't use Null Propagation Operator (?.)
Question:
username_0: `var myObject = State.MyObjects?.FirstOrDefault(i => i.Id == e.Id);`
Answers:
username_1: For this to work you should use the beta and the new codegen created by @username_3
A new beta should be released soon with fixes as well.
username_2: I don't get it, are you trying to use null propagation in Orleans source itself or in your own project using Orleans?
username_0: On my own project
username_2: @username_0
Wouldn't that be a VS problem? I'm using C# 6.0 just fine in my custom stream providers.
username_0: hmm.. why @username_1 mentioned a new codegen then?
username_2: that's why I'm confused
username_3: Orleans used to generate code and compile it during build using csc.exe. The new codegen compiles that code using MSBuild, using your current C# version.
username_0: so.. Any clues on why I'm seeing such an error?
username_3: Are you using the new beta? The reason you're seeing a problem is that directly invoking csc.exe as in the current stable (1.0.10) will compile for C# 5, not C# 6.
The new beta will compile using whatever the rest of your project is set to use.
username_0: Ahhh, yeah.. that makes sense.. I'm not using the beta yet. Any ETA on the RTM?
username_3: We will roll out a second beta with some important fixes to serialization generation (plus others), determine the stability, and then possibly RTM.
Is this issue resolved?
username_0: Yeha, thanks!
Status: Issue closed
|
google/mediapipe | 804249649 | Title: cannot able to run the holistic js demo in codepen.
Question:
username_0: when I run the holistic js demo (https://codepen.io/mediapipe/pen/LYRRYEw), I got errors like below. Is it possible to fix it?
Thanks
`E0209 06:18:00.916000 1 solutions_wasm.embind.cc:266] Calculator ConstantSidePacketCalculatorUpperBodyOnly not found. holistic_solution_wasm_bin.js:9:80637
E0209 06:18:00.918000 1 solutions_wasm.embind.cc:276] List of calculators holistic_solution_wasm_bin.js:9:80637
-------------------------- holistic_solution_wasm_bin.js:9:80637
holistic_solution_wasm_bin.js:9:80637
ConstantSidePacketCalculatorSmoothLandmarks holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__ConstantSidePacketCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__ImagePropertiesCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__PreviousLoopbackCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__GateCalculator_1 holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__PacketPresenceCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__GateCalculator_2 holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__facelandmarksfromposegpu__ImagePropertiesCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__ImageToTensorCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__InferenceCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__SsdAnchorsCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__TensorsToDetectionsCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__NonMaxSuppressionCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__posedetectiongpu__DetectionLetterboxRemovalCalculator holistic_solution_wasm_bin.js:9:80637
holisticlandmarkgpu__poselandmarkgpu__SplitDetectionVectorCalculator`
Answers:
username_1: it's also broken on the mediapipe hosted demo: https://mediapipe.dev/demo/holistic
username_1: @eknight7 any updates on this?
username_2: This should be fixed in the next update; will close issue at that time.
username_1: Thank you @username_2!
username_3: hi, could the problem be in the chrome browser? Does anyone know what is happening? I need a solution for this please
username_2: It should now be fixed with the latest push; see if it works for you now? I left additional comments on #1621 .
username_4: @username_0,
Can you please respond to the above comment? Thanks!
Status: Issue closed
username_5: Closing this as it is resolved |
sara-sabr/ITStrategy | 732624481 | Title: ABB v0.067 - Increase connections to other strategies, frameworks and programmes
Question:
username_0: Next iteration should increase connections to:
- Interoperability (increase opening of APIs as a mandatory set of requirements, no matter what tech is brought in or for any existing solutions to which changes will be introduced)
- Tech Debt (help decompose large monoliths into reusable structures - see Maryanne's blogs about how to do so)
- SDM (model to operate in a loosely coupled architecture),
- Market Watch initiative (broadly looking at tech to match capabilities needs, taking ownership of tech choices vs choosing in pool of submitted tech)
Etc.
_Originally posted by @username_0 in https://github.com/sara-sabr/ITStrategy/pull/1325#issuecomment-718992376_<issue_closed>
Status: Issue closed |
DIYgod/RSSHub | 678086711 | Title: Bilibili的rss突然无法打开
Question:
username_0: - 完整路由地址，包含所有必选与可选参数
rsshub.app/bilibili/user/video/25522213
- 预期是什么
返回数据
- 实际发生了什么
```
Looks like something went wrong
Error: this route is empty, please check the original site or create an issue
at module.exports (/app/lib/middleware/parameter.js:19:19)
at processTicksAndRejections (internal/process/task_queues.js:97:5)
at async module.exports (/app/lib/middleware/anti-hotlink.js:33:5)
at async module.exports (/app/lib/middleware/template.js:15:5)
at async /app/lib/middleware/api-response-handler.js:143:9
at async module.exports (/app/lib/middleware/api-template.js:2:5)
at async module.exports (/app/lib/middleware/utf8.js:4:5)
at async module.exports (/app/lib/middleware/debug.js:14:5)
at async grant (/app/lib/middleware/access-control.js:23:13)
at async module.exports (/app/lib/middleware/access-control.js:31:20)
at async module.exports (/app/lib/middleware/header.js:20:5)
at async module.exports (/app/lib/middleware/onerror.js:20:9)
```
- 部署相关信息
官方和自己在heroku上的部署都尝试了一下，都出现此类错误。
- 额外信息（日志、报错等）
Answers:
username_1: 有相同问题。
部署方式为自建。
username_1: 已解决。
git pull 一下就好了。
Status: Issue closed
|
mislav/will_paginate | 267486044 | Title: "Will Paginate" string rendered instead of page gap separator - missing reasonable default
Question:
username_0: `will_paginate-3.1.6` renders "Will Paginate" in the absence of `page_gap:` translation.
Actual result: as on the attached screenshot picture ("Will Paginate" string rendered)
Expected result: in the absence of translated string a reasonable default is rendered.
Screenshot:
 |
gradle/guides | 542002114 | Title: Handle case where remote sites don't support HEAD verb.
Question:
username_0: @ysb33r commented on [Tue Jun 13 2017](https://github.com/gradle-guides/gradle-guides-plugin/issues/4)
`checkLinks`: In some cases, like www.scalatest.org, using HEAD on a URL will return 404, but GET on the same URL will work. It might be worthwhile to have a failover where when HEAD fails, then falling back to GET is an option. In this case the call should not retrieve full content at the URL, but probably only the first 100 bytes. |
vslavik/poedit | 58730939 | Title: Bing Translations
Question:
username_0: Please also add Bing Translator for internet translation suggestions.
It often gives better results than Google does.
http://www.microsoft.com/web/post/using-the-free-bing-translation-apis
Answers:
username_1: Poedit doesn’t use Google Translate’s suggestions, but rather the [MyMemory engine](https://mymemory.translated.net). As far as I know, it sources Microsoft Translator too. In any case, while Translator may provide better suggestions for some specific languages and texts, I believe MyMemory is better overall, across all languages — generally speaking, Translated.net are *the* guys to go to for CAT expertise.
It’s also worth noting that Microsoft Translator API isn’t free, but uses volume-based monthly subscription and it’s (at its current pricing anyway) financially not feasible to fund its usage without making Poedit *itself* subscription-based, which I’d really rather not.
Status: Issue closed
username_1: Times change. Some engines improve, some less so, and some things that made sense in 2015 no longer do in 2018.
Starting with 2.1 (now [in beta](https://poedit.net/download)), Poedit ditches MM for machine translation and uses a combination of [DeepL](https://deepl.com) (if your language is supported, that's what you want to use — [better](https://www.deepl.com/quality.html) than Google's, even w/ NMT), Google Translate and Microsoft Translator. |
cooperka/react-native-snackbar | 591691315 | Title: Can we have an option to re-position SnackBar ?
Question:
username_0: Hi,
Can we have an option to re-position SnackBar, currently it is painted at the very bottom of the screen, if we can include an option to change it or add some `bottomMargin` offset value, it would be nice.
Project Settings:
RN: 0.61.5
Platform: iOS and Android
Answers:
username_1: Indeed! Please see #15 for discussion of positioning, along with open PR #137; that PR is blocked on feedback that needs to be addressed by the submitter. If you'd like to address my feedback and submit an updated PR for margins, feel free! Thanks.
Status: Issue closed
|
graphql-python/graphene | 169933702 | Title: Use UUIDs instead of Relay IDs
Question:
username_0: Hey Guys,
Awesome library 👍
We are looking for some best practises regarding the use of Relay IDs together with UUIDs.
**Our Setup**
We are using Django Graphene with two different setups:
1. Redux & GraphQL (no Relay)
2. Relay
We are using a postgresql for DB. Our django models generate UUIDs for all our entities. We really like the UUIDs :)
Based on the following issue https://github.com/facebook/relay/issues/191 and the fact that we want the client to use a single ID throughout the app, we have decided that the client side should only know about the Relay ID.
**The Problem**
This raises three issues:
1. Mutations and advanced resolves with foreign keys need to be translated from Relay IDs to our UUIDs.
2. More importantly, it makes it much harder to debug ids all the way from client to server.
3. It makes it impossible to use non graphql APIs.
**Current Solution**
We have written a middleware that solves problem (1), by translating IDs to our UUIDs.
However, this does not solve (2) and (3), and we still feel our solution is kind of patchy :)
**Questions**
1. Can we use our UUIDs instead of the base64 IDs generated by graphene? Would it still work fine with Relay?
2.Any other suggestions or other best practises we should use?
Thanks!
Nir
Answers:
username_1: Hi @ username_0, it's possible to use the `UUID`s as long as you can map from a UUID to a type/model.
So, given a UUID get the corresponding model.
With the current implementation this is not as easy to achieve but with the upcoming version `1.0` it will be trivial:
https://github.com/graphql-python/graphene/blob/1.0/graphene/relay/tests/test_node_custom.py
username_0: Got you. Thanks for the answer.
Our UUID is generated by an external system, and is completely random, so it would be kind of hard to follow that logic.
Do you think it would be possible to somehow figure out the mapping from the query when using Django? Any other suggestions?
Thanks :)
Status: Issue closed
username_1: A quick update on this issue!
The `1.0.dev` version is live on PyPI, I recommend you to give it a try having a similar approach to this: https://github.com/graphql-python/graphene/blob/master/graphene/relay/tests/test_node_custom.py
Usually for mapping this random id's to type+id is useful to have a "utility table" that stores what is the table/type corresponding to each "UUID".
Hope that helps! |
biglotteryfund/blf-alpha | 239713238 | Title: Make various changes to the site to improve complaints experience
Question:
username_0: Builds up #97
[ ] Implement #66 in order to prevent Google from indexing https://biglotteryfund.org.uk/england/about-big/customer-service/making-a-complaint
[ ] Turn the two PDFs that appear when you google for 'Big lottery fund complaint' into HTML pages, or replace them entirely<issue_closed>
Status: Issue closed |
pandas-dev/pandas | 421429262 | Title: pd.merge combine two df was so huge!
Question:
username_0: My Code
```python
quote = pro.daily(ts_code=code, end_date=end_date.strftime("%Y%m%d"))
basic = pro.daily_basic(ts_code=code, end_date=end_date.strftime("%Y%m%d"))
combine_quote = pd.merge(left=quote, right=basic, on='ts_code', how='inner')
```
#### Problem description
I have two dataframe which named quote and basic,let me see two tables detail
```python
quote.info(memory_usage='deep')
<class 'pandas.core.frame.DataFrame'>
Index: 2869 entries, 20190315 to 20070427
Data columns (total 10 columns):
ts_code 2869 non-null object
open 2869 non-null float64
high 2869 non-null float64
low 2869 non-null float64
close 2869 non-null float64
pre_close 2869 non-null float64
change 2869 non-null float64
pct_chg 2869 non-null float64
vol 2869 non-null float64
amount 2869 non-null float64
dtypes: float64(9), object(1)
memory usage: 568.8 KB
```
```python
basic.info(memory_usage='deep')
<class 'pandas.core.frame.DataFrame'>
Index: 2869 entries, 20190315 to 20070427
Data columns (total 15 columns):
ts_code 2869 non-null object
close 2869 non-null float64
turnover_rate 2869 non-null float64
turnover_rate_f 2869 non-null float64
volume_ratio 2864 non-null float64
pe 2869 non-null float64
pe_ttm 2869 non-null float64
pb 2869 non-null float64
ps 2869 non-null float64
ps_ttm 2869 non-null float64
total_share 2869 non-null float64
float_share 2869 non-null float64
free_share 2869 non-null float64
total_mv 2869 non-null float64
circ_mv 2869 non-null float64
dtypes: float64(14), object(1)
memory usage: 680.8 KB
```
then I will merge two to one use pd.merge,this procedure is not fast,about 1.5-2 second,I was shocked when I check memory usage,see that
```python
combine_quote.info(memory_usage='deep')
<class 'pandas.core.frame.DataFrame'>
Int64Index: 8231161 entries, 0 to 8231160
[Truncated]
feather: None
matplotlib: 3.0.3
openpyxl: 2.6.1
xlrd: 1.2.0
xlwt: None
xlsxwriter: None
lxml.etree: 4.3.2
bs4: 4.7.1
html5lib: None
sqlalchemy: 1.3.1
pymysql: 0.9.3
psycopg2: None
jinja2: 2.10
s3fs: None
fastparquet: None
pandas_gbq: None
pandas_datareader: None
gcsfs: None
</details>
Answers:
username_1: Can you provide a minimal example? http://matthewrocklin.com/blog/work/2018/02/28/minimal-bug-reports
Status: Issue closed
username_2: Looking at the numbers here you have 2,869 entries in the two frames to begin with and 8,231,161 entries in the frame resulting from the merge operation. Coincidentally enough, 8,231,161 is 2,869**2 so I would gather that your dataset is simply generating a cartesian product.
Based off of that I don't think this is a pandas issue so closing. However, if you disagree and can provide the minimal example above feel free to reopen |
libertyernie/LoopingAudioConverter | 1032069900 | Title: Request. *.wem Audio Format.
Question:
username_0: detail.
https://residentevilmodding.boards.net/thread/10415/modding-changing-points-adding-longer
<File Id="************" Language="SFX>
<ShortName>**********.wav</ShortName>
<Path>SFX/*********/***********.wem</Path>
</File>
Alternatively, it is described in any text file in the following format. (TAB delimited)
Streamed Audio ID Name Audio source file Generated audio file (wem) Wwise Object Path Notes |
mdolab/OpenAeroStruct | 349990001 | Title: Redo derivatives for lift_coeff_2D to be vectorized
Question:
username_0: We currently have analytic derivs for this comp, but they are dense and use multiple Python for loops.
Instead, we want to use a sparse formulation and use vectorized math to get the derivatives.
Answers:
username_0: https://github.com/mdolab/OpenAeroStruct/blob/master/openaerostruct/aerodynamics/lift_coeff_2D.py
Status: Issue closed
username_0: Fixed by #105 |
RTRGitHub/RTRRedux | 185463744 | Title: Yellorite Ore
Question:
username_0: I was wondering if the yellorite ore was supposed to spawn in the over world only or if it the mod it self isn't allowing it to spawn in the other dimensions. I can't find any Yellorite ore in the mining world or the deep dark I even used the digital miner to see if I can find anything. It isn't picking any up. So far the only place it spawning is in the overworld.
Thanks!
Cody,
Answers:
username_1: Potential world around could be the dimensional miner or the mining laser setup.
username_2: cant replace ore in the overworld at this time, use the mining laser for now
Status: Issue closed
|
vlang/v | 774864052 | Title: Check version of installed v module
Question:
username_0: We need a way to get the current version of the installed V module.
like
```
v version mymod
```
Answers:
username_0: and a check if a newer version is available would be nice. as well as an update command.
username_1: `v update modname` works, as well as `v install modname` when the module is already installed.
username_1: Also just `v update` updates everything.
username_1: There is `v outdated` that checks if some modules have newer versions.
For more details, see cmd/tools/vpm.v .
username_1: Extending `v version mymod` to show the version of `mymod` would be nice indeed.
username_0: strange that both have the same behaviour. could we clean that up?
username_1: That is expected behaviour, not a bug - `update` only updates, while `install` can install and update, depending on whether the module is present.
username_0: What is the `update` for when `install` does both? Any use case?
username_2: i prefer `upgrade` over `update`
username_3: Or we can have `v install -u` where `-u` means to update if needed.
username_4: `v install` already does that without the extra switch.
username_1: `v update package` will not install a package, if it is not already installed. `v install package` will. Both are useful.
username_0: `v show <module>` shows the version.
Status: Issue closed
|
ValveSoftware/steam-for-linux | 23394884 | Title: Request
Question:
username_0: Dear people of valve, please, please, please! add support to nvidia legacy driver 173.xx.xx, counter strike runs awesome in a fx5500 even better then windows, the only thing why i can't switch entirely to linux is the steam client, i can't believe i can play cs in xp, vista or windows 7 whit this card but no in linux i mean c'mon
Answers:
username_1: While the steam client does not explicitly check for driver versions anymore, it generally expects GL 2 capable hardware. Your mileage will vary with hardware before that. There's no reasonable expectation for pre-GL2 hardware to gain support in the future.
Status: Issue closed
|
BeresIvan/HM1_Refactor | 713998173 | Title: Screenshot
Question:
username_0: https://github.com/username_0/HM1_Refactor/blob/master/assets/images/Screenshot%20(9).png
Answers:
username_0: https://github.com/username_0/HM1_Refactor/blob/master/assets/images/Screenshot%20(10).png
username_0: https://github.com/username_0/HM1_Refactor/blob/master/assets/images/Screenshot%20(11).png
username_0: https://github.com/username_0/HM1_Refactor/blob/master/assets/images/Screenshot%20(12).png |
charleshendrickson/eds221-day8-activities | 975050888 | Title: New Issue
Question:
username_0: ``` r
library(tidyverse)
df <- tribble(
~name, ~height, ~hair_color,
"Teddy", 2.2, "brown",
"Khora", 3.1, "white",
"Waffle", 1.0, "tan"
)
df %>%
filter(name == "Waffle")
#> # A tibble: 1 × 3
#> name height hair_color
#> <chr> <dbl> <chr>
#> 1 Waffle 1 tan
```
<sup>Created on 2021-08-19 by the [reprex package](https://reprex.tidyverse.org) (v2.0.0)</sup> |
SDM-TIB/SDM-RDFizer | 831145992 | Title: Base URI in Subject of N-Triples RDF File
Question:
username_0: **Describe the bug**
When generating data from a CSV throught an RML mapping, the subject of the N-Triples contains the base URI
**To Reproduce**
Steps to reproduce the behavior (and resources):
1. Get CSV File
2. Get an RML mapping file without self-join and child/parent relations
3. Use RDFizer to generate the data
**Expected behavior**
N-Triples should only contain the URI specified for the subject
**Screenshots**
If applicable, add screenshots to help explain your problem.
**Desktop (please complete the following information):**
- OS: Windows 10
- Browser: Chrome
Answers:
username_1: @username_0 could you send us the link to the mapping and csv data?
username_0: Mapping: https://github.com/username_0/TFG-OpenCityData-JuanBernal/blob/main/data/mappings/ruido2021-RML.rml
CSV: https://github.com/username_0/TFG-OpenCityData-JuanBernal/blob/main/data/preprocessed/RuidoDiario2021/RuidoDiario2021_v2.csv
username_2: Dear @username_0,
Thank you for using the SDM-RDFizer.
I fixed the issue. I ran it with the data and mapping provided and didn't get the error anymore. Please test it out, so we can close this issue. Please note that the changes I made are not in the library yet.
Thank you again.
Enrique
Status: Issue closed
username_0: Perfect, that work!
Thank you |
dotnet/roslyn | 55794854 | Title: Language Supported Contracts (`require`)
Question:
username_0: As mentioned in #98 it'd be nice to have dedicated syntax and compiler support for pre-post conditions.
I like the use of the `require *expr*`. I'd like to see it done in a future-compatible, extendable way. For the simple case `require name!=null` will just compile to something like:
if (name!=null) throw new ContractViolation()
`require` should be allowed at the beginning of the method (perhaps even before the `{`) signifying it's a pre-condition, anywhere inside the method (in which case it's a condition that basically confirms that some code that just run acted like it should) or after the method (in which case it's wrapped in `try{} finally{}`) for a post-condition.
Pre/Post conditions (as well as the inline ones) are really only powerful when combined with static analysis tools.
Since determining when a given pre/post condition will be violated is a non-trivial task, the compiler should perhaps reduce itself to only checking the very trivial ones (when a constant violates the pre-condition for example). Language services can be used to provide warnings and errors for more involved cases, since they can be changed and updated much more frequently.
The conditions should also be eliminated when it can be statically proven that they are not needed at run-time. (There should also be a compile time flag to remove all if performance is desired over proven correctness, or if they trust the static analyzers enough). These need not be specified in the language spec (and hence doesn't force mono or others to implement them) but rather should be treated as an optimization detail (since if they can be statically proven to be not needed, then it does not change the semantics of the program).
This should allow the `require` tags to be forwards compatible, and not require breaking changes in the future when it's decided that more analysis can be done (as would have to be done now if more local type inference wanted to be introduced).
Answers:
username_1: Thanks, @username_0. We've posted the proposal we were discussing in #98 as #119.
username_2: As an aside, I believe your example should be:
`if (name==null) throw new ContractViolation()`
username_0: Okay I'm closing this issue because it looks like everything I wanted is included in the proposal. Just wanted to try and get the discussion out of #98 but you beat me to it.
Status: Issue closed
|
verdaccio/verdaccio | 346672167 | Title: NewRelic Agent throwing errors after update to 3.3.0 version
Question:
username_0: **Describe the bug**
NewRelic Agent throwing errors after update to 3.3.0 version.
Error is following:
```
{"v":0,"level":50,"name":"newrelic","hostname":"5538dce32c23","pid":75,"time":"2018-08-01T14:20:12.118Z","msg":"Unable to serialize payload for method connect.","component":"remote_method","stack":"TypeError: this is not a Date object.\n at Object.valueOf (<anonymous>)\n at Object.[Symbol.toPrimitive] (<anonymous>)\n at Object.toJSON (<anonymous>)\n at JSON.stringify (<anonymous>)\n at stringify (/www/docker/node_modules/json-stringify-safe/stringify.js:5:15)\n at RemoteMethod.serialize (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:42:15)\n at RemoteMethod.call [as invoke] (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:64:8)\n at CollectorAPI._connect (/www/docker/node_modules/newrelic/lib/collector/api.js:186:21)\n at getEnvDict (/www/docker/node_modules/newrelic/lib/collector/api.js:178:10)\n at factMapCb (/www/docker/node_modules/newrelic/lib/collector/facts.js:67:12)","message":"this is not a Date object."}
```
As result - empty NewRelic logs
**After some investigation**
After some investigation in diffs:
https://github.com/verdaccio/verdaccio/compare/v3.2.0...v3.3.0
I've found a PR https://github.com/verdaccio/verdaccio/pull/764.
After that PR merge it breaks NewRelic logic.
It works well after the commit https://github.com/verdaccio/verdaccio/commit/825f5a57eaa04a895fbf9673481fcf647b0a0d68
But after the https://github.com/verdaccio/verdaccio/commit/29bb57ad5f5d762e8f1f4e3267f5bd11f8caa272 commit it is broken.
**Screenshots**
Answers:
username_1: Adding formatted stack trace:
```
{
"v":0,
"level":50,
"name":"newrelic",
"hostname":"5538dce32c23",
"pid":75,
"time":"2018-08-01T14:20:12.118Z",
"msg":"Unable to serialize payload for method connect.",
"component":"remote_method",
"stack":"TypeError: this is not a Date object.\n
at Object.valueOf (<anonymous>)\n
at Object.[Symbol.toPrimitive] (<anonymous>)\n
at Object.toJSON (<anonymous>)\n
at JSON.stringify (<anonymous>)\n
at stringify (/www/docker/node_modules/json-stringify-safe/stringify.js:5:15)\n
at RemoteMethod.serialize (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:42:15)\n
at RemoteMethod.call [as invoke] (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:64:8)\n
at CollectorAPI._connect (/www/docker/node_modules/newrelic/lib/collector/api.js:186:21)\n
at getEnvDict (/www/docker/node_modules/newrelic/lib/collector/api.js:178:10)\n
at factMapCb (/www/docker/node_modules/newrelic/lib/collector/facts.js:67:12)",
"message":"this is not a Date object."
}
```
username_2: @username_0 please we need some log from the server or client side, otherwise, we cannot track any issue.
username_0: Here you go:
The first one is a stdout logs on start of the verdaccio 3.2.0 server:
[3.2.0-OK.txt](https://github.com/verdaccio/verdaccio/files/2252515/3.2.0-OK.txt)
This one is with verdaccio 3.4.1 (errors are the same for version 3.3.0-3.4.1)
[3.3.0-3.4.1-Errors.txt](https://github.com/verdaccio/verdaccio/files/2252514/3.3.0-3.4.1-Errors.txt)
I hope this could help you to find an issue and fix it as soon as possible.
Thanks.
username_3: @username_0 we have analyzed the log and the error seems to come from YOUR plugin.
If the problem is due to the library conflict, as you said, we would like you to prove it to us.. Please feel free to create a bug simulation and post the link here ... Thanks!
username_0: Hey @username_3 .
What kind of simulation do you need?
I've already sent you logs from working(...-3.2.0 verdaccio version) and not working (3.3.0-... verdaccio versions).
I've already found a commit after what NewRelic NodeJS Agent starts to fail. (Please note that it is NOT my plugin - it is a global and official NewRelic plugin designed for work with NodeJS applications)
https://www.npmjs.com/package/newrelic
I've already tried to decrease the version of the NewRelic Agent - no results.
The only thing to have it working correct is to use 3.2.0 version or to revert "asciidoctor.js" https://github.com/verdaccio/verdaccio/commit/29bb57ad5f5d762e8f1f4e3267f5bd11f8caa272 changes.
In case you need to provide you a manual how to implement a NewRelic NodeJS Agent - in the verdaccio engine - just let me know.
Thanks!
username_3: Sorry but I would prefer your full example.
Thank you!
username_1: @username_0 Please reopen the issue if the problem persists.
Status: Issue closed
username_0: Hi @username_1 . Yes this issue still exists.
Here is the same issue that i've found of "opal-runtime" module (that is using by "asciidoctor.js" package)
https://github.com/opal/opal/issues/1684#issuecomment-370683649
username_1: @username_0 Thanks for reporting. I am investigating it again.
username_0: @username_1 In case you need some help with NewRelic agent integration - just let me know
username_0: @username_1 .
Seams like i've found a solution - could you please update "asciidoctor.js" package to the latest "1.5.7-rc.1" version? (it already includes the "opal-runtime" latest 1.0.7 version)
https://www.npmjs.com/package/asciidoctor.js?activeTab=versions
Current 1.5.6 version of "asciidoctor.js" requires only "1.0.3" version of the "opal-runtime" module.
I've changed it manually in the verdaccio's package.json, rebuilt it and newRelic become alive.)))
username_0: @username_1 could you please review this PR?
https://github.com/verdaccio/verdaccio/pull/934
username_0: For now I've restored the registry update using `npm-shrinkwrap.json` file
```
{
"dependencies": {
"verdaccio":{
"version": "3.5.1",
"from": "[email protected]",
"dependencies": {
"asciidoctor.js": {
"version": "1.5.7-rc.1",
"from": "[email protected]",
"dependencies": {
"opal-runtime": {
"version": "1.0.6",
"from": "opal-runtime@~1.0.7"
}
}
}
}
}
}
}
```
I think have this here would be useful.
username_2: **Describe the bug**
NewRelic Agent throwing errors after update to 3.3.0 version.
Error is following:
```
{"v":0,"level":50,"name":"newrelic","hostname":"5538dce32c23","pid":75,"time":"2018-08-01T14:20:12.118Z","msg":"Unable to serialize payload for method connect.","component":"remote_method","stack":"TypeError: this is not a Date object.\n at Object.valueOf (<anonymous>)\n at Object.[Symbol.toPrimitive] (<anonymous>)\n at Object.toJSON (<anonymous>)\n at JSON.stringify (<anonymous>)\n at stringify (/www/docker/node_modules/json-stringify-safe/stringify.js:5:15)\n at RemoteMethod.serialize (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:42:15)\n at RemoteMethod.call [as invoke] (/www/docker/node_modules/newrelic/lib/collector/remote-method.js:64:8)\n at CollectorAPI._connect (/www/docker/node_modules/newrelic/lib/collector/api.js:186:21)\n at getEnvDict (/www/docker/node_modules/newrelic/lib/collector/api.js:178:10)\n at factMapCb (/www/docker/node_modules/newrelic/lib/collector/facts.js:67:12)","message":"this is not a Date object."}
```
As result - empty NewRelic logs
**After some investigation**
After some investigation in diffs:
https://github.com/verdaccio/verdaccio/compare/v3.2.0...v3.3.0
I've found a PR https://github.com/verdaccio/verdaccio/pull/764.
After that PR merge it breaks NewRelic logic.
It works well after the commit https://github.com/verdaccio/verdaccio/commit/825f5a57eaa04a895fbf9673481fcf647b0a0d68
But after the https://github.com/verdaccio/verdaccio/commit/29bb57ad5f5d762e8f1f4e3267f5bd11f8caa272 commit it is broken.
**Screenshots**
Status: Issue closed
username_4: Are there plans to put AsciiDoc support back it?
username_2: @username_4 I'll be glad to add it again if someone finds a way using something different as the library `asciidoctor.js` which does some sort of monkey patching to native libraries. |
dotnet/wpf | 640978170 | Title: XamlXmlWriter is including the namespace for each element
Question:
username_0: * .NET Core Version: (e.g. 3.0 Preview1, or daily build number, use `dotnet --info`)
Preview6. Regression from Preview5.
* Does the bug reproduce also in WPF for .NET Framework 4.8?: Yes/No
No.
**Problem description:**
XamlXmlWriter is including the namespace where it wasn't before.
**Minimal repro:**
Run this code:
```csharp
XamlSchemaContext xsc = new XamlSchemaContext();
var generated = new StringBuilder();
var xmlSettings = new XmlWriterSettings { Indent = true, OmitXmlDeclaration = true };
var hashSetAssemblyName = typeof(HashSet<List<BigContainer>>).GetAssemblyName();
var listAssemblyName = typeof(List<Dictionary<int, HashSet<List<BigContainer>>>>).GetAssemblyName();
using (XamlXmlWriter writer = new XamlXmlWriter(XmlWriter.Create(new StringWriter(generated), xmlSettings), xsc))
{
XamlType list = xsc.GetXamlType(typeof(List<Dictionary<int, HashSet<List<BigContainer>>>>));
var ns1 = $"clr-namespace:System.Collections.Generic;assembly={hashSetAssemblyName}";
var ns2 = "clr-namespace:Test.Elements;assembly=XamlTestClasses";
var ns3 = "http://schemas.microsoft.com/winfx/2006/xaml";
writer.WriteNamespace(new NamespaceDeclaration(ns1, "a"));
writer.WriteNamespace(new NamespaceDeclaration(ns2, "b"));
writer.WriteNamespace(new NamespaceDeclaration(ns3, "c"));
writer.WriteStartObject(list);
writer.WriteEndObject();
}
var expected = @"<List c:TypeArguments=""Dictionary(c:Int32, a:HashSet(List(b:BigContainer)))"" xmlns=""clr-namespace:System.Collections.Generic;assembly={0}"" xmlns:a=""clr-namespace:System.Collections.Generic;assembly={1}"" xmlns:b=""clr-namespace:Test.Elements;assembly=XamlTestClasses"" xmlns:c=""http://schemas.microsoft.com/winfx/2006/xaml"" />";
expected = string.Format(expected, listAssemblyName, hashSetAssemblyName);
Assert.AreEqual(expected, generated.ToString());
```
```xml
--
| | - | | "<List c:TypeArguments=\"Dictionary(c:Int32, a:HashSet(List(b:BigContainer)))\" xmlns=\"clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib\" xmlns:a=\"clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib\" xmlns:b=\"clr-namespace:Test.Elements;assembly=XamlTestClasses\" xmlns:c=\"http://schemas.microsoft.com/winfx/2006/xaml\" />"\r\n
1 | | + | | {<a:List c:TypeArguments="a:Dictionary(c:Int32, a:HashSet(a:List(b:BigContainer)))" xmlns:a="clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib" xmlns:b="clr-namespace:Test.Elements;assembly=XamlTestClasses" xmlns:c="http://schemas.microsoft.com/winfx/2006/xaml" />}
COMMENTS:
```
Answers:
username_1: Is this a true bug, or a change in formatting that’s both syntactically legal and semantically correct but perhaps happens to break a test etc?
username_0: The output is legally and semantically correct, and it actually looks like an improvement to me. Prior to Preview6, the first namespace is written out twice: once without an name prefix, as the default namespace, and once with a name prefix.
```xml
xmlns=\"clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib\" xmlns:a=\"clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib\"
```
In Preview6, there is no default namespace, and the first namespace is only written once. The absence of the default namespace requires that each element also get a name prefix.
```xml
xmlns:a="clr-namespace:System.Collections.Generic;assembly=System.Private.CoreLib"
```
```xml
<a:List c:TypeArguments=...>
```
This is what I would expect the line below, rather than having the writer give a default namespace _and_ a name prefix for the same namespace.
```csharp
writer.WriteNamespace(new NamespaceDeclaration(ns1, "a"));
```
So to answer your question, yes, this is syntactically legal and semantically correct and actually looks better than before, but the writer output is changed from previous releases.
And this is not a bug unless WPF provides a guarantee about the consistency of the output of the XamlXmlWriter, release to release.
username_1: We don’t provide preview-to-preview compat.
Serialization compatibility is something you should take into consideration between netcoreapp3.1 & net48, net5.0 & netcoreapp3.1, net5.0 & net4.8.
Serialization compat doesn’t mean that the output of XamlXmlWriter remains static; it just means that it remains compatible within reason, and in the context of safe (ie secure) [de]serialization actions.
I strongly suspect that a valid XML would remain compatible.
username_0: In that case, it sounds the DRT should be updated.
username_2: Sounds like tests need to be updated here. |
brookhong/Surfingkeys | 507901618 | Title: 按 `t` 列出标签页的速度有点慢
Question:
username_0: 按下 `t` 列出所有标签页的速度比较慢, 搜索框出现后基本都需要再等上 0.5 秒才会列出标签, 但是 `T` 的表现就很快.
是不是因为 `t` 的列表中有热门网站的缘故?
Status: Issue closed
Answers:
username_1: 是的，`t`需要搜索更多URL数据。
你可以在设置中增加
map("t", "T");
如果不需要overlay这种显示方式，希望跟`t`的显示方式一样，可以再加入
settings.tabsThreshold = 0;
username_0: @username_1
谢谢!!
还有一个问题, 既然 `t` 是按最近使用的顺序来列出标签页, 可不可以考虑默认选中下面的第一个候选项? 也就是 `map("t", "t ↓")` 这样的感觉.
虽然有 `<Ctrl-6>` , 但是我往往并不知道上一个标签到底是哪个, 所以大多数时候都是按下 `t` 之后还需要再按一下 ↓ 然后回车, 如果可以省掉按 ↓ 这一下就很人性化了, 而且默认光标下移一格也不会影响搜索框内的输入或是继续向下选择.
username_2: 非常赞同这个建议, 默认第一个结果是最好的, 和 alfred 一样,这样就很舒服了
username_2: 目前macOS 选择也无法用 ctrl + 数字呀 |
logstash-plugins/logstash-output-elasticsearch | 193158285 | Title: Don't include sprintf'ed fields if the operation fails
Question:
username_0: currently this plugin can be configured to pull certain information from the event like document_id, index or document_type. This is implemented using `event.sprintf("%{[@metadata][id]})` for the document_id for example.
If this sprintf fails, the key is still created in the event with the raw string value of `%{[@metadata][id]}`. this created problems as documents will be indexed in elasticsearch with incorrect data.
My suggestion is that we add an extra condition after each sprintf and only modify the event if the sprintf succeeds.
Answers:
username_0: closing in favor of https://github.com/elastic/logstash/issues/6363
Status: Issue closed
|
GetStream/stream-js | 100557032 | Title: How to get all activities?
Question:
username_0: Hello, could you help me with this problem:
- I try to get all activities ( command: feed.get() ). When i get the response, i can't parse it ( can see it in console, but when i try to parse it in JS, the response is undefined )
Answers:
username_1: feed.get() requires a callback function
```
feed.get(function(error, response, body) { /* callback */ })
```
username_0: Thanks a lot! A little more question:
- How to change a number of activities in response? In callback, when person do post(for example), or i do feed.get(), i have an array for 15 elements, how can i change the number of activities in array?
username_1: have a look at the examples here: https://github.com/GetStream/stream-js/blob/master/README.md
Status: Issue closed
|
Coursemology/coursemology2 | 119898513 | Title: Implement Leaderboard
Question:
username_0: Think we will introduce groups to leaderboard in v2 ? That means we can see the leaderboard of a group.
### Tasks
- [ ] Rank students by level, and by achievements
- [ ] Leaderboard Settings
Answers:
username_1: Completed, pending small fixes for Course Leaderboard.
Do we need to do for groups, or no need?
username_2: How is it done in v1?
username_1: Comparing to v1, I implemented guilds the other time, which is a leaderboard for a collective group (using average experience points of members in that guild/group). That has not been done in v2 yet; and I can do that next.
username_1: Comparing to v1, I implemented guilds the other time.
I suspect guilds are slightly similar to groups now, and what I also did was have a leaderboard for each guild. The leaderboard will rank guilds based on average experience points of members in that guild/group. That has not been done in v2 yet; and I can do that next.
username_2: I think by group is good.
username_1: I will need to fix up the leaderboard stuff first, otherwise I should be able to get the group leaderboard based on the existing features.
Status: Issue closed
username_1: This issue is completed, to close. |
JuliaChae/M3D-RPN-Waymo | 875722054 | Title: Great Work!
Question:
username_0: Hi ! This is really good repo and I am able to train the warmup on waymo.
However there is one issue I dont understand.
For depth aware , we need the prev_0 images right?
But the waymo Kitti adapter does not provide that?
How did you get it?
Status: Issue closed
Answers:
username_0: It appears you do not need this folder. Sorry! |
dbs-leipzig/gradoop | 217273978 | Title: Create Nesting operator
Question:
username_0: 1. new interface: GraphGraphCollectionToGraphOperator
2. unit tests
3. GraLa example
Answers:
username_1: x Reminder: No benchmarks
username_2: Implementations might be outdated. Please create new issue if you are still working on this.
Status: Issue closed
username_1: That depends if you are still interested in a nested implementation of
Gradoop.
-- Giacomo
Il giorno ven 5 ott 2018 alle ore 09:12 <NAME> < |
SeleniumHQ/selenium | 897936482 | Title: --session-timeout does not close the browser, if the browser is still trying to load the resource
Question:
username_0: ## 🐛 Bug Report
If a browser is still trying to load a resource and the timeout specified by `--session-timeout` is reached, the closing of the browser will fail, but the selenium server will reopen the slot potentially leading to many still opened stale browsers.
(Same issue first opened in the docker repository: https://github.com/SeleniumHQ/docker-selenium/issues/1283)
## To Reproduce
- Start selenium: `java -jar selenium-server-4.0.0-beta-3.jar standalone --session-timeout 5` (needs chrome and chrome driver in path, probably other browsers are affected too)
- Start a web server that will take a long time to answer
- Run a selenium script that tries to load a resource of the slow web server
- Wait for the timeout
- Result: closing fails (browser still opened), selenium has reopened the slot (http://localhost:4444/ui/index.html#/)
## Expected behavior
Selenium closes the browser correctly after the session times out.
## Test script or set of commands reproducing this issue
`server.py`:
```python
from http.server import BaseHTTPRequestHandler,HTTPServer
import time
class Handler(BaseHTTPRequestHandler):
def do_GET( self ):
time.sleep(6000)
self.send_response(200)
self.send_header( 'Content-type', 'text/html' )
self.end_headers()
httpd = HTTPServer( ('127.0.0.1', 8080), Handler )
httpd.serve_forever()
```
`test.py`:
```python
import os
from selenium import webdriver
grid_url = "http://localhost:4444/wd/hub"
def get_driver():
return webdriver.Remote(
command_executor=grid_url,
options=webdriver.ChromeOptions())
try:
driver = get_driver()
driver.get("http://127.0.0.1:8080")
except Exception as e:
print(e)
```
- Run `java -jar selenium-server-4.0.0-beta-3.jar standalone --session-timeout 5`
- Run `python3 server.py`
- Run `python3 test.py`
## Environment
OS: Linux
Browser: Chrome
Browser version: 90.0.4430.212
Browser Driver version: ChromeDriver 90.0.4430.24
Language Bindings version: python-4.0.0.b3
Selenium Grid version (if applicable): 4.0.0-beta-3
Answers:
username_0: - Logs:
- `test.py`: [out.txt](https://github.com/SeleniumHQ/selenium/files/6522052/out.txt)
- `selenium server`: [out-server.txt](https://github.com/SeleniumHQ/selenium/files/6522063/out-server.txt)
username_1: @username_0, thank you for providing enough information to reproduce the issue.
I was able to see the same behaviour you describe, and after double checking the code and going through the code flow, I can conclude that this is one of the reasons why the session timeout is configurable.
Let me describe the code flow:
Session starts -> load url command sent by client -> Grid forwards command to ChromeDriver (or GeckoDriver, browser is not relevant in this case) -> ChromeDriver tells Chrome to load url -> Long wait due to the page loading slowly
At this point, the request is waiting for the page to load. The session timeout triggers and the session is [stopped](https://github.com/SeleniumHQ/selenium/blob/47824b377592e4949af94327fb559483830176f6/java/server/src/org/openqa/selenium/grid/node/local/LocalNode.java#L161). When the session is stopped, one of the steps is to [shutdown](https://github.com/SeleniumHQ/selenium/blob/49f706f436b7c97d5ac39596d141d345aef58b79/java/client/src/org/openqa/selenium/remote/service/DriverService.java#L276) the browser driver.
That shutdown is an http request to the browser driver, but the browser driver is blocked since it is still loading the page, therefore the shutdown will most likely timeout (which is what you see in the logs), and the browser & browser driver will remain open.
This scenario can happen with any interaction that eventually needs a long time to process. Which is why the session timeout is configurable, to allow a proper timeout depending on your environment. This also helps to spot issues on your infrastructure or potential bugs in the application under test, as it should not be normal for a request to take longer than the session timeout. So, if you see often browsers left open, session timeout needs to be adjusted or the network connections need to be double checked or the application under test is having issues.
I will close this issue as there it no needed change in the Grid, but if you want to propose an alternative that still keeps the Grid code simple and maintanable, feel free to comment and we can see if there is a way to add your suggestion.
Status: Issue closed
|
vim/vim | 476501453 | Title: [RFC] I hope '->' operator can be applied to all built-in functions, funcref
Question:
username_0: @username_1
I really love the recently added `->` operator! ac92e25a33c37ec5becbfffeccda136c73b761ac
we can write like `list->map(...)->filter(...)->join(...)`
I'm very pleased to write method chaining in vim script! 😄
But it seems it is limited to these functions. why?
https://github.com/vim/vim/blob/ac92e25a33c37ec5becbfffeccda136c73b761ac/src/evalfunc.c#L997-L1020
I hope `lhs->rhs()` operator can be applied to when rhs is:
* Other all built-in functions
* Funcref (Lambda/Partial/User-defined function)
Answers:
username_0: These are examples for above proposal.
How do you think?
## Other built-in functions
```vim
echo {'key': 42}->has_key('key')
```
## User-defined function
```vim
let s:Module = {}
function! s:Module.flatmap(list, func)
...
endfunction
function! s:fold(list, init, func)
...
endfunction
echo range(1,10)->s:Module.flatmap({n -> n % 2 == 0 ? ['even'] : []})
echo range(1,10)->s:fold(0, {acc,n -> acc + n}
```
## Lambda
```vim
call obj->{o -> execute('echo o')}
```
Lambda is a special case.
because rhs is not function call (`func()`).
if only function call is allowed, it must be like this.
```vim
" obj is passed to the first argument of function call
call obj->{o -> execute('echo o')}()
```
`->` operator can define a flow of reading order.
One can read in very natural order (left to right).
but in above example, (1) "obj" (left) (2) "()" (right) (3) lambda (left)
so I think "()" is not necessary.
## Similar operator in other languages
Vim script `->` operator looks like method call.
However other languages seems to call this operator "pipeline operator".
* [ECMACcript pipeline operator](https://github.com/tc39/proposal-pipeline-operator)
* [Elixir pipeline operator](https://stackoverflow.com/a/41585450/6321251)
You may feel strange `->` accepts user defined function like this.
(rhs is not only alphabet)
```vim
echo range(1,10)->s:Module.flatmap({n -> n % 2 == 0 ? ['even'] : []})
echo range(1,10)->s:fold(0, {acc,n -> acc + n}
```
but if this is pipeline operator, I think this is not strange.
```vim
echo range(1,10) |> s:Module.flatmap({n -> n % 2 == 0 ? ['even'] : []})
echo range(1,10) |> s:fold(0, {acc,n -> acc + n}
```
username_0: Sorry, @thinca told me that `|>` cannot be used in Vim script.
It breaks compatibility because `|` is used to separate each Ex command (`:h :bar`),
and `>` Ex command exists (`:h :>`).
I don't mean `->` operator should be changed.
I just say, hmm, different ways of thinking?
username_0: Added more examples to above comment:
```vim
function! GlobalFunc(list)
...
endfunction
function! Main()
let LocalFunc = function('s:fold')
echo range(1,10)->LocalFunc(0, {acc,n -> acc + n})
echo range(1,10)->GlobalFunc()
endfunction
```
username_0: I have created patches, but I want to hear @username_1 and other users' voices other than me :)
* Allow all built-in functions (`dict->has_key('key')`)
* https://github.com/vim/vim/compare/master...username_0:chain-op-allow-builtin-func?expand=1
* Allow user-defined function (Both `GlobalFunc` and `s:func`)
* https://github.com/vim/vim/compare/master...username_0:chain-op-allow-userfunc?expand=1
* Allow lambda (`echo obj->{o -> execute('echo o')}`)
* https://github.com/vim/vim/compare/master...username_0:chain-op-allow-lambda?expand=1
* derived from `username_0:chain-op-allow-userfunc` branch
username_1: Several other functions can be used as a method. I just added a few more, and now notice that has_key() was missing. For some the base is not the first argument, and also I want to test each one we add, thus we'll do a few at a time.
I'm not sure if we want to use this for simple values, such as number and Float. E.g. it looks unnatural to do 1.234->asin()
I think Lambda makes most sense when defining it separately:
let percent = {x -> x * 100}
echo Fraction()->percent()
It doesn't look good in a -> chain, also because it has -> inside in another meaning.
username_1: For user defined functions, we have a few options:
- Just allow any user defined function to be used after -> and pass the base as the first argument.
- When defining a function specify on what types it can be used as a method. This way you can define an Operate() function on a list and a separate Operate() function on a dict.
- When the base is a dictionary, use functions defined in that dictionary (a bit like a class).
The third one might not be needed, since we already have dictionary functions, and also supporting -> will lead to name conflicts. The example at ":help Dictionary-function", defining a len() method already shows this.
The first two might be both done, I don't see a problem with that yet.
username_0: How about adding type to argument, and "method" attribute?
```vim
function! Percent(n: Number) method
...
endfunction
```
if type is not specified, any value is accepted.
username_0: Lambda is funcref, variable of funcref must start with uppercase.
```vim
let Percent = {x -> x * 100}
echo Fraction()->Percent()
```
And this is little off-topic,
currently if global function `Percent()` exists, `E705` error occurs.
so receiving funcref as variable `Percent` may be dangerous, not `l:Percent`.
```vim
function! Percent(x)
return a:x * 100
endfunction
function! s:main()
" E705: Variable name conflicts with existing function: Percent
let Percent = {x -> x * 100}
" This is OK
let l:Percent = {x -> x * 100}
endfunction
call s:main()
```
username_0: In [8.1.1816](https://github.com/vim/vim/commit/fcfe1a9b8950b8b211ab3b24d84b17c6847ea43f), these cases seems to be supported. thanks!
```vim
echo {'foo': 42}->HasKey('foo')
echo {'foo': 42}->s:has_key('foo')
```
however, I want lambda style as well.
certainly your example works like this.
but the advantage of lambda is that it is not necessary to define a function to pass function to map(), filter(), and so on.
```vim
let l:Percent = {x -> x * 100}
echo (10.0 / 100.0)->l:Percent()
```
so I want
```vim
echo (10.0 / 100.0)->{x -> x * 100}
```
username_2: I also think that being able to use lambdas like this would be very very useful.
Not necessarily (for me) in actual scripts, where I would probably spread things out into several lines for readability. But when just testing things out or checking variable values etc., I find it much easier to write long one-liners in the command line than first defining a lambda, then using it in a new command, then needing to redefine the lambda and then using it again etc.
Once using methods, connecting them with lambdas feels very natural to me, and gives an awful lot of flexibility.
username_3: Shouldn't it be as follows?
```vim
echo (10.0 / 100.0)->{x -> x * 100}()
```
username_4: This lambda support seems really good!
I want this!!
(Personally I like `call obj->{x -> x}()` (Needing `()`) because I maybe confused :smiley:)
username_1: We need the () for consistency.
let Lambda = {x -> x * 100}
echo (10.0 / 100.0)->Lambda()
Does the same as:
echo (10.0 / 100.0)->{x -> x * 100}()
It's more obvious if you have an argument:
let Lambda = {x, y -> x * y}
echo (10.0 / 100.0)->Lambda(100)
Does the same as:
echo (10.0 / 100.0)->{x, y -> x * y}(100)
That last line is not something you would actually use though.
I'm still looking for a good example where the inline Lambda is useful.
Perhaps for modifying a value:
echo GetResult()->{x -> floor(x + 0.5)}->Process()
Line continuation can help for long chains:
echo GetResult()
\ ->{x -> floor(x + 0.5)}
\ ->Process()
\ ->DoSomeMore()
--
"Space is big. Really big. You just won't believe how vastly hugely mind-
bogglingly big it is. I mean, you may think it's a long way down the
road to the chemist, but that's just peanuts to space."
-- <NAME>, "The Hitchhiker's Guide to the Galaxy"
/// <NAME> -- <EMAIL> -- http://www.Moolenaar.net \\\
/// sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\ an exciting new programming language -- http://www.Zimbu.org ///
\\\ help me help AIDS victims -- http://ICCF-Holland.org ///
username_4: I have an example.
I'm always thinking how to write cool codes.
(code readabilities)
```vim
let x = foo
echo x
\ ->very(very_complex, arguments)
\ ->z(too_many_chars)
```
I think this is very cool.
I write below instead above if the inline lambda is nothing.
```vim
let x = foo
let x = very(very_complex, arguments)
let x = z(too_many_chars)
echo x
```
But this is fail if I want to use `const x`.
And in my opinion, the first example is very very cooler than the below basic writing.
```vim
const x = foo
echo z(
\ very(x, very_complex, arguments),
\ too_many_chars
\ )
```
Thanks :sunglasses::+1:
username_0: The combination with `:const` is certainly reasonable case to support lambda.
It encourages programming style without side-effect.
Status: Issue closed
|
tencentyun/cos-nodejs-sdk-v5 | 1105022181 | Title: 使用pnpm安装出现类型报错
Question:
username_0: ```
[下午7:01:59] Starting compilation in watch mode...
node_modules/.pnpm/[email protected]/node_modules/cos-nodejs-sdk-v5/index.d.ts:1857:12 - error TS2304: Cannot find name 'Body'.
1857 Body?: Body | string,
~~~~
[下午7:02:04] Found 1 error. Watching for file changes.
``` |
valor-software/ngx-bootstrap | 295085184 | Title: How to show only month in bsDatepicker popup instead of date?
Question:
username_0: How to show only month in bsDatepicker popup instead of date?
Status: Issue closed
Answers:
username_1: There's no need to open an issue here if you want to ask general question, use [StackOverflow](https://stackoverflow.com/questions/tagged/ngx-bootstrap) or [Slack](https://ngx-slack.herokuapp.com/) instead
Closed as a duplicate of #2627
username_0: <input type="text" name="startDate" class="form-control" bsDatepicker [datepickerMode]="'year'" [minMode]="'month'" ngModel/>
#2627 this issue is not understandable. can you please help me in above code?
username_2: On View:
<input
bsDatepicker
(onShown)="onOpenCalendar($event)">
On Component:
onOpenCalendar(e) {
if (this.formatoMesAnio) {
e.setViewMode('month');
e.monthSelectHandler = (event: CalendarCellViewModel): void => {
e.value = event.date;
return;
};
}
}
It´s manually and work for me !!! |
Radseq/Gold-Chat | 492297649 | Title: Horrible margins, duplicated controls, connection with server
Question:
username_0: 
This is no how you should write your xaml code. You should use Grids to make your controls responsive.
I don't really get why every single control is splited to user control and window...
Writing your connecting data in .cs file is huge mistake because user can not change IP without rebuilding whole application. You should write it in some configuration file. |
gyulanemeth/game-of-life-js | 138330487 | Title: új pull request hiba
Question:
username_0: Szia!
Nem tudok új pull request-et létrehozni, nincs hozzá külső felhasználónak jogosultsága.
Egy fél óra után azt is meguntam, hogy próbáljak publikálni valahogy a github-ra saját felhasználónévvel :D.
Feltettem users-re a következő linkre:
http://users.itk.ppke.hu/~kohkr/game_of_life_js
Változás: width/height átadás tervezési hiba javítás, play/pause, reset, design (mérethez kicsit igazodó) és inicializálások kerültek bele (kezdeti kirajzolások, kezdőállapot). Memory leak elkerülésére figyeltem (elvileg). Némi átstrukturálás is történt, pl. a kirajzolás vagy az állapotmátrix random értékekkel feltöltése függvénybe lett kiemelve, új paraméterek és eseménykezelő függvények.
<NAME>f
Answers:
username_1: Az én repómat forkoltad? Úgy pedig mennie kéne... Megnéztem a tiédet és üresnek tűnik, ezért merült fel bennem a kérdés. Amúgy jól néz ki nagyon, amit csináltál, úgyhogy kéne az a pull request mindenképp ;)
username_0: Áááá, "that's where the magic happens" :). Így már sikerült, köszönöm!
Status: Issue closed
|
pingcap/tispark | 950102487 | Title: [BUG] Tipsark omit the privileges of one user when write in a table
Question:
username_0: **Describe the bug**
I created a user without privileges and I pass the following:
- Case 1 : the user was able to write to the tables that he should not be able to write when using the TiBatchWrite method.
- Case 2 : The user could not write when using normal method (df.write)
- Case 3 : I gave the user the select privilege and he was able to write with the normal method (df.write)
**What did you do**
<!--
1. Create of tables and user :
```
create database db_test;
CREATE TABLE db_test.table_test (
`id` varchar(36) COLLATE utf8_general_ci NOT NULL,
`name` varchar(36) COLLATE utf8_general_ci DEFAULT NULL,
`school` varchar(36) COLLATE utf8_general_ci NOT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `test_unique_1` (`name`,`school`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;
CREATE TABLE db_test.table_test2 (
`id` varchar(36) COLLATE utf8_general_ci NOT NULL,
`name` varchar(36) COLLATE utf8_general_ci DEFAULT NULL,
`school` varchar(36) COLLATE utf8_general_ci NOT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `test_unique_1` (`name`,`school`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8 COLLATE=utf8_general_ci;
CREATE USER 'test1' IDENTIFIED BY 'test1';
```
2. run this code and the user can write in this tables
```
val sparkConf = new SparkConf()
.setIfMissing("spark.master", "local[*]")
.setIfMissing("spark.sql.extensions", "org.apache.spark.sql.TiExtensions")
.setIfMissing("spark.tispark.pd.addresses", "127.0.0.1:2379")
.setIfMissing("spark.tispark.isolation_read_engines", "tiflash")
val spark = SparkSession.builder.config(sparkConf).getOrCreate()
import spark.implicits._
val df = Seq(
("10", "n5", "n13"),
("11", "n7", "n11")
).toDF("id", "name", "school")
val data = Map(DBTable("db_test", "table_test") -> df, DBTable("db_test", "table_test2") -> df)
TiBatchWrite.write(
data,
spark,
Map(
"tidb.addr" -> "127.0.0.1",
"tidb.port" -> "4000",
"tidb.user"-> "test1",
"tidb.password" -> "<PASSWORD>",
"replace" -> "true")
)
```
3. run this code , and the user can't write
```
val sparkConf = new SparkConf()
[Truncated]
"replace" -> "true"
)
df.write
.format("tidb")
.options(tidbOptions)
.mode("append")
.save()
```
-->
**What do you expect**
**What happens instead**
The user can write when he should not do it since he does not have enough privileges
**Spark and TiSpark version info**
spark : 2.4.0
tispark : 2.3.16
Answers:
username_1: Hi @username_0
Thanks for bring up this issue. The authorization mechanism has been implemented in big data components yet. We know there are other users having concern about this and we definitely want to get it implemented at some point. Will update this issue once the roadmap about security is clear.
username_0: Hi @username_1 , I have a doubt ,can I use tispark with spark 3 and java 11 ?
username_1: Master version has spark 3 support, however it's not been widely used yet.
Talking about java 11, unfortunately we still compile the package with java 8.
username_0: Ok, thank you for the answer |
dentemple/jest-invert | 560126199 | Title: Improve handling of React components / Create an "intuitive" inversion of a React component
Question:
username_0: Currently, `jest-invert` shows a rather arcane error when dealing with React components. Which, considering that the library simply inverts object keys, makes sense considering how React v16 works under the hood.
This issue is for considering what it would mean to "invert" a React component, and to update the library to execute it.
Answers:
username_0: An object can be checked to see if it's a React component, but what to do about transforming it? Replace it with `null`? Replace it with a Fragment? Nest it under a custom component that just happens to be named `Inverted`?
Furthermore, it's not ideal to have to require users to download the `react` library for this, considering that `jest-invert` has a more general scope.
More than likely, a companion library (e.g., `jest-invert-react`) should be created to handle this specific use case. |
LRNWebComponents/hax-body | 288708251 | Title: abstract hax-source-search from hax-source
Question:
username_0: hax-source-search
- all the visual side of hax-source
- make the template stamp in via templatizer
hax-source
- all the wiring of BREAD https://gist.github.com/username_0/f6e10c6bbfa3846be8952c48ad47f622
- try and map what we have currently to be in this
hax wide changes
- source selected needs to just be set and then hax-source-search needs notified of this wiring change (perhaps via some kind of active-connection object or something)
- we need global registration for hax-source-search as far as the source-search object attaching to the hax-store. This way people could swap it out
Status: Issue closed
Answers:
username_0: it now does this |
rana01645/android-push-notification | 652597029 | Title: onMessageReceived() method not called
Question:
username_0: Hi recently I'm learning push notification.. I've configured with FCM while testing from **the FCM `onMessageReceived()` method is called. But I can't receive any push notification from App server.** I've checked the device_token sent to server it was same as I tested in FCM.
I Can't even find out what was my mistake.
If you could help me plz. |
giampaolo/psutil | 298723844 | Title: wait() function could throw TimeoutExpired exception when it shouldn't
Question:
username_0: This issue is about function wait() https://github.com/username_1/psutil/blob/dd19ba7b5bbf5ea67a4d386baa3fabd13b118533/psutil/_pswindows.py#L791.
My test scenario is listed below, although it doesn't repro 100%, it happens quite often. I didn't measure how frequently it happens though.
1) Have a process object, proc, call proc.kill().
2) Call proc.wait(30)
3) proc.wait() throws TimeoutExpired almost instantly.
Speculation:
1) cext.proc_wait() returns quickly because process exited (killed) and returns a value not equal to WAIT_TIMEOUT.
2) Somehow, the process still exists, and thus pid_exists() returns True. (Zombie status? Other processes, threads still have references to the process handle?)
3) Since timeout value is not None, TimeoutExpired throws.
This behavior is different from the previous version. In the previous version, wait() will not throw TimeoutExpired, but just return the value returned from proc_wait().
Answers:
username_1: Mmmm, can you paste the traceback?
username_0: This is our code snippet that calls psutil:
--------------------------------
# timeout = 30
377 logging.info('Terminating process id: %r', process_id)
378 try:
379 proc = psutil.Process(process_id)
380 proc.terminate()
381 proc.wait(timeout)
382 logging.info('Process %d was terminated successfully', process_id)
383 return True
384 except psutil.TimeoutExpired as e:
385 logging.exception(e)
386 logging.error('Failed to terminate process %d in %d seconds',
387 process_id, timeout)
388 return False
389 except psutil.AccessDenied:
390 logging.error(
391 'Unable to terminate process %d due to insufficient privileges',
392 process_id)
393 return False
394 except psutil.NoSuchProcess:
395 logging.info('Process %d was not running', process_id)
396 return True
--------------------------------
This is the log got:
--------------------------------
[INFO 14:31:30.256 5056:516 test_lib.py:377]Terminating process id: 5472
[ERROR 14:31:30.367 5056:516 test_lib.py:385]psutil.TimeoutExpired timeout after 30 seconds (pid=5472)
Traceback (most recent call last):
File "c:\TestAutomation\lib\test_lib.py", line 381, in _TerminateProcess
proc.wait(timeout)
File "c:\TestAutomation\third_party\py\psutil\__init__.py", line 1264, in wait
return self._proc.wait(timeout)
File "c:\TestAutomation\third_party\py\psutil\_pswindows.py", line 635, in wrapper
return fun(self, *args, **kwargs)
File "c:\TestAutomation\third_party\py\psutil\_pswindows.py", line 805, in wait
raise TimeoutExpired(timeout, self.pid, self._name)
TimeoutExpired: psutil.TimeoutExpired timeout after 30 seconds (pid=5472)
--------------------------------
Please see the log timestamp, they happened roughly within .1 second. This is on Windows platform.
I tried to terminate notepad.exe and wait, and failed to repro. So it maybe it is harder to repro when terminating a light-weight process. This is just speculation though.
username_1: I briefly investigated this. It may be due to `WaitForSingleObject` returning `WAIT_ABANDONED` or `WAIT_OBJECT_0`:
https://msdn.microsoft.com/en-us/library/windows/desktop/ms687032(v=vs.85).aspx
username_0: I've created a simple script that could repro the issue occasionally. Originally I said the issue happens quite often. This assertion seems to be corrected. We ran tests in parallel on lots of machines, that increase the repro rate a lot. And the process we kill may also play a factor. Please run the script multiple times if needed.
[psutil_test.zip](https://github.com/username_1/psutil/files/1835141/psutil_test.zip)
username_1: OK, I think I fixed this in https://github.com/username_1/psutil/pull/1253 but unfortunately I cannot reproduce the issue. Can you try? If you don't have a C compiler setup just tell me what Python version you have and I can attach a wheel file for you to try.
username_1: Wheel for python 2.7 64bit:
https://ci.appveyor.com/project/username_1/psutil/build/1463/job/56s14mi6p932cycv/artifacts
username_1: I'm assuming this is fixed.
Status: Issue closed
|
mapbox/mapbox-gl-js | 126836605 | Title: 0.12.3 drawCircles error?
Question:
username_0: My map that was working fine with 0.12.2 is now freezing and throwing the attached error in both Safari and Chrome.
Working 0.12.2 version here: http://parallel.co.uk/os-openmap-gl/
Answers:
username_1: Maybe related to 5739dd09a1216770984d512a390074046fa2c7d9, which may contain a fix?
cc @username_2 @lucaswoj
Status: Issue closed
username_2: Yes, very likely a duplicate of #1933. @lucaswoj is on it. |
enthought/traits | 600321752 | Title: HasRequiredTraits documentation is garbled
Question:
username_0: The rendered version of the `HasRequiredTraits` documentation doesn't look like wot it oughta.
<img width="1107" alt="Screenshot 2020-04-14 at 17 30 06" src="https://user-images.githubusercontent.com/662003/79346268-1079e780-7f2a-11ea-849a-8ef0278668d8.png">
Answers:
username_1: This is because of `Usage` section which isn't in Numpy docstring format. Best equivalent is most likely `Examples`:
I see a lot more rendering issues in documentation, mainly with the use of `Description` section (should be just blocks of text after one line description) and class/function mixups (e.g. Property). Should I go through docs and fix all of these as well?
username_0: If you have the energy, yes please! Though I'm also fine with a fixing-things-as-we-notice-them approach.
username_0: And yes, `Examples` sounds good here.
Status: Issue closed
|
google/go-jsonnet | 260362887 | Title: Weird stuff in autogenerated doc
Question:
username_0: Where does it come from?
https://imgur.com/a/oJMNy
Answers:
username_1: Those functions are functions from github.com/mjibson/esc, which is used to generate std.go. But I don't know why they are included in the GoDoc.
By the way, how std.go can be generated, can you include the generation script in the repo? And I can't run tests.sh in my laptop, due to missing the tests_path.source files.
username_0: Thanks.
To generate std.go it's enough to run `go get github.com/mjibson/esc && go generate`. No special script needed, just the standard go stuff.
Regarding tests, there are two test suites. One can be run just by running `go test`. This one was created specifically for go version and it should just work. But there is also an older and more authoritative test suite in the stable C++ version that can be used to find the missing parts. It can be run using `./tests.sh` and it needs to be pointed at C++ repo. Hence `tests_path.source`. If you want to run these tests please clone C++ repo (https://github.com/google/jsonnet). Then create `tests_path.source` which contains `TESTS_PATH="/full/path/to/the/root/of/C++/repo"`. Then you should be able to run these tests (some of them still fail due to missing features). FYI there are some plans to unify and improve the way tests work. |
lfdelta/PitGameMaker | 213505732 | Title: Generalize player controls
Question:
username_0: Directional keys are already implemented in this way, but moving forward, replace expressions such as "if(mouse_check_held(mb_left))" with a single variable like "if(grappleButton)" where grappleButton = mouse_check_held(mb_left). This allows for greater generality and flexibility, and may also have a small effect on computation, since the check is performed only once per cycle rather than multiple times throughout the code.
Status: Issue closed
Answers:
username_0: grappleKey defined alongside other player inputs |
lpil/mix-test.watch | 161812117 | Title: Does not work on 1.3.0
Question:
username_0: with elixir 1.3.0, mix_test_watch 0.2.6, I get:
```
⇒ mix test.watch
Running tests...
** (TokenMissingError) nofile:1: missing terminator: ) (for "(" starting at line 1)
(elixir) lib/code.ex:168: Code.eval_string/3
(elixir) lib/enum.ex:651: Enum."-each/2-lists^foreach/1-0-"/2
(elixir) lib/enum.ex:651: Enum.each/2
(mix) lib/mix/tasks/run.ex:67: Mix.Tasks.Run.run/1
(mix) lib/mix/task.ex:296: Mix.Task.run_task/3
(elixir) lib/enum.ex:651: Enum."-each/2-lists^foreach/1-0-"/2
```
when `mix test` runs and passes.
Normally I'd try to do some digging to figure out what's going on, but I'm not sure where to look with that anonymous stack trace. Let me know if there's additional information about my project I can provide (if this is not a general problem)
Answers:
username_1: I see this too
username_2: Well this looks rather grim. :(
username_2: I think this is a bug in mix. https://github.com/elixir-lang/elixir/issues/4861
username_3: fyi - I just confirmed this is fixed by upgrading to elixir 1.3.1.
```
brew update
brew upgrade elixir
```
username_2: Thanks!
username_4: Confirmed, saw this on Elixir 1.3.0, upgrading to 1.3.1 fixed it.
Status: Issue closed
|
phoenixframework/phoenix | 394941895 | Title: New static files not found on first request, but working on following requests.
Question:
username_0: <!--
### Precheck
* For help and support, use Stack Overflow or the Phoenix Talk mailing list
* For proposing a new feature, start a discussion on the Phoenix Core mailing list
* For bugs, do a quick search and make sure the bug has not yet been reported
* Ensure that this issue is related to the Phoenix library and not one of the dependencies listed in mix.exs (Ecto, Plug, etc.)
* All checked? Be nice and have fun!
-->
### Environment
* Elixir version (elixir -v): Elixir 1.7.4
* Phoenix version (mix deps): 1.4.0 - 1.3 also
* NodeJS version (node -v): v8.10.0
* NPM version (npm -v): 5.6.0
* Operating system: Windows 10
### Expected behavior
After adding new file to priv/static/images it should be immediately available to domain/images/file.jpg requests.
### Actual behavior
When the new file is requested for the first time, it gives a 404 error in the browser and a "(Phoenix.Router.NoRouteError) no route found for GET /images/file.gif (appWeb.Router)" error on the server. A second request for the same file, makes the request work, for some reason. First request always fails, even if I wait 5 minutes before making a request for the new file, and following requests are successful.
Answers:
username_1: if you recompile the app (kill the server and re-run phx.server), does the first request work? Thanks!
username_0: Request is working when app is recompiled.
I also found out that any request for a non-existing file, that results in an error, will make a following request for the new file to work.
If I add test2.jpg to the folder, request notthere.gif, results in error, then request test2.jpg, the request works, otherwise it fails.
I also want to mention that I reinstalled windows, changed the Antivirus, added folders to exceptions, disabled AV, reinstalled Elixir and so on.
The only thing that works so far is recompiling the app or making an error request or requesting the new file twice.
username_0: It may be Windows related because there's no issue running the app on Ubuntu server.
Status: Issue closed
|
naser44/1 | 118330604 | Title: اللهم في هذا الصبـــــــــاح
Question:
username_0: <a href="http://ift.tt/1SdRAyo">اللهم في هذا الصبـــــــــاح</a> |
mojolicious/minion | 686570770 | Title: [ Feature request ] An event for progress?
Question:
username_0: * Minion version: 10.08
* Perl version: 5.32
* Operating system: FreeBSD 11.3
So you provide this nice thing called 'progress' in notes. That's great and it works. However, on my first (always naive) attempt to use Minion I find myself polling notes periodically because there's no event to subscribe to.
Is it possible to fire an event when a job updates notes (or notes->progress)? Failing that, is it possible to document how to use progress to actually deliver progress data in an event-driven way?
Apologies if I missed anything in the documentation that covers this. Thanks in advance.
Answers:
username_1: Theoretically this would be possible to implement with PostgreSQL listen/notify, but the SQLite and MySQL backends could not provide the feature. So we are not going to do it. And it's too complex to include as a documentation example.
username_0: Ok, so could you at least mention that using progress requires polling the database? :)
username_1: How about you point out specifically the sections of the documentation that you think are too ambiguously worded?
username_0: Well I'm not sure it's an issue of ambiguity as much as omission. You mention progress at first in the Minion docs here:
```
Minion is a high performance job queue for the Perl programming
language, with support for multiple named queues, priorities, delayed
jobs, job dependencies, job progress,...
```
Then you hint at the idea here further down in the same page:
```
job
my $job = $minion->job($id);
Get Minion::Job object without making any changes to the actual job or
return "undef" if job does not exist.
# Check job state
my $state = $minion->job($id)->info->{state};
# Get job metadata
my $progress = $minion->$job($id)->info->{notes}{progress};
```
I guess that can be generalized as metadata but as far as I can tell there's really two metadatum ... notes and results. Progress is really a subset of notes.
The word progress is not mentioned further. Digging into the documentation, there appears to be confusion in Minion::Job's docs because in the same page you first say this:
```
info
my $info = $job->info;
Get job information.
# Check job state
my $state = $job->info->{state};
# Get job metadata
my $progress = $job->info->{notes}{progress};
```
but further down you say this:
```
note
my $bool = $job->note(mojo => 'rocks', minion => 'too');
Change one or more metadata fields for this job. ...
# Share progress information
$job->note(progress => 95);
```
Perhaps I'm mistaken, but I really see 'note' as the metadata since it's just an arbitrary hash reference conceptually.
Nowhere else that I have found do I see information or examples as to how to retrieve this progress information when it changes. Maybe you could add some blurb that says "getting the notes requires polling the database for them so you can detect when they change"? I just threw that out there. ;)
username_2: I think you read into the example too much. Progress is an arbitrary
meta-data field, not an inherent part of Minion.
username_0: I get that. However Minion explicitly states that it supports job progress, so it's natural to me to go looking for where that support is.
username_1: Yes, you're reading too much into the example. The notes feature is job progress information, there is no subset.
username_1: I think what it all comes down to is that the Minion guide explaining all the features with examples is still missing.
username_1: The guide is now a planned feature, but i have no idea when i'll get around to actually writing it. https://github.com/mojolicious/minion/issues/100
username_0: A guide which explains all the features with examples would address my concerns. IMHO if I have to look at your code to understand what something actually does, the documentation is by definition lacking in that area.
Not that I don't like looking at your code either, you have a uniquely terse style that can be fun to figure out...and not that people who are using your code shouldn't look at it anyway...but this is strictly a measure of documentation quality.
username_0: I've addressed the event issue myself because I use the Pg backend. What would it take to get support for listen/notify in SQLite or MySQL? Can you do listen/notify without Pg? I'm just curious here.
username_3: It's not possible with SQLite because there is no server.
username_2: You would need to use something which has a pub/sub interface. Options
include a websocket hub (like Mercury) or Redis (if you want to stay
strictly in the Mojo camp).
username_0: I'm just covering all the bases here. I recognize this, but maybe someone has written something that does pubsub for SQLite or even for nothing at all (username_2's comment ... redis for example). I'm trying to work within the bounds of what has been stated to get the feature I feel should likely be a part of Minion.
So to my original request ... not a wanted feature then? I can live with that as long as every other option has been exhausted. Maybe you could make pubsub event an optional feature of Minion::Backend::Pg?
username_2: You're working with too blunt of a hammer. You're assuming that EVERY item
put in notes is related to progress, which isn't necessarily true.
For example, I personally use notes to cache progress in a potentially
flaky job, for example. No need for any pubsub there.
So to my original request ... not a wanted feature then? I can live with
> that as long as every other option has been exhausted. Maybe you could make
> pubsub event an optional feature of Minion::Backend::Pg?
>
No, username_1 said that he won't add features unique to any specific backend.
You can write a role which wraps `note` and does the pubsub there. That
should be fairly trivial. It just can't be a builtin feature.
username_0: Not sure what you are perceiving here. :)
username_2: Notes is a hashref. I don't know what you mean by subset.
username_0: So if notes is a hashref and progress can be a hashref (it always is in my case) and progress/hashref is a key/value pair in notes... in that case is it not valid to say that progress is a subset of notes?
username_1: You might want to continue that discussion in the Google Group or on IRC. |
343max/TabHibernation | 539552274 | Title: Fix "Not allowed to navigate top frame to data URL"
Question:
username_0: Hello,
it seems, that Chrome has implemented a security feature that prevents to execute data URLs from JavaScript, which causes TabHibernation to not function anymore: if you try to hibernate a tab, the console shows a message like "Not allowed to navigate top frame to data URL: data: ...".
It seems, it is possible to use the _src_ attribute of an <iframe>, though.
Maybe one could replace all the content with an <iframe> or use about:blank and add the <iframe> ? (just guessing; obviously, I'm not much into Extension programming, yet ^^).
Resources:
https://ourcodeworld.com/articles/read/682/what-does-the-not-allowed-to-navigate-top-frame-to-data-url-javascript-exception-means-in-google-chrome
https://groups.google.com/a/chromium.org/forum/#!topic/blink-dev/GbVcuwg_QjM%5B1-25%5D
https://stackoverflow.com/questions/45493234/jspdf-not-allowed-to-navigate-top-frame-to-data-url |
DigitalInBlue/Celero | 93616935 | Title: CMake fails
Question:
username_0: Hi,
I get the following error on the latest *master* branch:
[sgwood@mstdev31 Celero (master)]cmake .
-- The C compiler identification is GNU 4.8.3
-- The CXX compiler identification is GNU 4.8.3
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- SYSTEM: Linux
CMAKE_CXX_COMPILER: /usr/bin/c++
CMake Error at experiments/CMakeLists.txt:6 (ADD_SUBDIRECTORY):
The source directory
/nis_home/sgwood/programs/Celero/experiments/CMakeFiles
does not contain a CMakeLists.txt file.
-- Configuring incomplete, errors occurred!
Answers:
username_1: You shouldn't have a CMakeFiles directory. When configuring CMake, ensure that you have a different "where is the code" directory from the "where to build the binary" directory. Hopefully that helps.
username_0: Yep, that did it. Thanks.
Status: Issue closed
|
discordjs/discord.js | 750131893 | Title: GuildMember.send docs actually TextChannel.send docs
Question:
username_0: <!--
If you need help with discord.js installation or usage, please go to the discord.js Discord server instead:
https://discord.gg/bRCvFy9
This issue tracker is only for bug reports and enhancement suggestions.
You won't receive any basic help here.
-->
**Please describe the problem you are having in as much detail as possible:**
The GuildChannel.send method documentation is actually the documentation for the TextChannel.send method.
[Link to relevant section](https://discord.js.org/#/docs/main/stable/class/GuildMember?scrollTo=send)
Answers:
username_1: 
That is not `GuildChannel`. That is `GuildMember` which has a `send` method implemented from the `TextBasedChannel` class, as the page heading suggests:
Status: Issue closed
username_0: @username_1 Sorry, i mixed myself up. @username_2 why close this? shouldn't the example have `GuildMember.send` rather than `channel.send`?
username_3: It's closed because the documentation comes from a single place, because the method is only written in a single place in the source-code: `TextBasedChannel#send()`
GuildMember does not have its own different send method, and therefore does not get its own docs.
username_0: ah ok, thanks! |
beagle-dev/beagle-lib | 819989577 | Title: Add CONFIGURE parameter to specify NVIDIA cards generation upon config
Question:
username_0: So it’s a known issue that the config file currently has “compute_30” and that if you have an NVIDIA card that won’t work.
The compute number changes depending on what generation of card you’re running. For example, my GTX 1070 is compute_61 or sm_61. The 20x0 are 70 and so on. What is for sure wrong is compute_30. I think it’s so old that it’s not supported anymore.
I don’t know how to probe the system to find out what card is installed and pick the number accordingly, so I was wondering if we could add a —with-NVIDIA=Pascal/etc option to the configure script to select the correct number and avoid this issue.
I want to pick this up but I wanted some input first to know if this is the way to go
Answers:
username_1: My suggestion would be to remove required compute level specification. Current `nvcc`compilers all support by default double-precision.
So, please make a `pull request` and I'll merge.
Also, I personally have switched over to `cmake` (in the `hmc-clock` branch) for current `beagle` development and will merge that in when finished (a couple of months).
username_2: For what it's worth, I'm doing this for an NVIDIA Tesla V100 that supports compute capability 7.0:
```console
$ sed -i 's/compute_30/compute_70/' configure.ac
$ ./autogen.sh
$ ./configure
$ make
``` |
bkiers/Liqp | 469605269 | Title: Support for properties (dot notation)
Question:
username_0: Not sure if this should work or not. Given an object:
```
public class Resource {
private String title;
public Resource(String title) { this.title = title; }
public String getTitle() { return title; }
```
Currently this will resolve to an object:
```
<html>
<head><title>{{resource}}</title></head>
<body>...</body>
</html>
```
But this will not (`liqp.exceptions.VariableNotExistException: Variable 'resource.title' does not exists`):
```
<html>
<head><title>{{resource.title}}</title></head>
<body>...</body>
</html>
```
Answers:
username_1: There is a way to do that. Have a look at https://github.com/username_1/Liqp/blob/master/src/main/java/liqp/Template.java#L307 and the unit tests in https://github.com/username_1/Liqp/blob/master/src/test/java/liqp/TemplateTest.java#L28
In your case, do this:
```java
String template = "{{resource.title}}";
String rendered = Template.parse(template).render(true, "resource", new Resource("MU"));
```
and `rendered` mill now be `"MU"`
Status: Issue closed
username_1: As a `Map` is also possible, btw:
```java
String template = "{{resource.title}}";
Map<String, Object> map = new HashMap<String, Object>(){{
put("resource", new Resource("MU"));
}};
String rendered = Template.parse(template).render(map);
```
username_0: Ok, I am actually trying the Map version and it doesn't work for me. The complete example is actually using a javax.script.Bindings object as the Map, as I'm trying to make a simple liquid template engine:
https://github.com/micronode/whistlepost/blob/develop/extensions/whistlepost-scripting-liquid/src/main/java/org/whistlepost/scripting/liquid/LiquidScriptEngine.java#L29
Result of this input is the error mentioned above:
```
<html>
<head>
<title>{{resource.valueMap}}</title>
</head>
<body>
</body>
</html>
```
I'll test some more if I get time but meanwhile will focus on other template engines.
username_1: If you could post code so that I could reproduce this in a unit test, I could have a look at it. |
sensu-plugins/sensu-plugins-selinux | 124769994 | Title: Push gem
Question:
username_0: Hi,
would it be possible for you to push the gem to rubygems.org?
This way the plugin can be installed through the standard sensu plugin installation.
Cheers,
Matteo
Answers:
username_1: Hi Matteo, I've pushed [0.0.4](https://rubygems.org/gems/sensu-plugins-selinux/versions/0.0.4) to rubygems. Thanks for noticing!
Status: Issue closed
|
oppia/oppia-android | 1168932855 | Title: Fix SVG image performance & scaling
Question:
username_0: The fix in #4027 introduces DP-based scaling for images, and fixes some alignment/sizing issues. Unfortunately, these exposes a few problems:
1. Sometimes upscaling SVGs on Android seems to result in a blurry image (which suggests rasterization happening under the surface). Since we use AndroidX for SVG rendering internally, it seems likely that this is happening (though I didn't confirm). Trying different scaling techniques didn't work, but I suspect that changing the Picture's dimensions during creation time in AndroidSVG to something large (e.g. by adding a configurable scalar that we can set from Oppia Android) might ensure the rasterization is correct. Note that blurry images will also be a problem for bitmaps since there's no upscaling going on, so we need to add that, too.
2. SVG rendering generally seems slow, and using bitmaps (like we do for KitKat & Lollipop) doesn't actually fully fix the problem since it recomputes the bitmap on rebinding, and upscaling results in very large bitmaps that use up far too much app memory.
We want both upscaling and performant rendering for SVGs. I think that involves:
1. Computing the smallest size necessary to render the image without it resulting in blurriness (which probably requires better understanding all the different scenarios an SVG can be rendered)
2. Precomputing the bitmap and caching it via Glide (similar to what we do for LaTeX in #4068) which should significantly improve performance. We do, however, need to make sure that any dependencies on display density are reflected in the caching message digest since the user changing the system's display zoom should result in recomputing the images (to ensure they don't waste space and/or end up blurry due to upscaling). |
strapi/strapi | 804433747 | Title: strapi.log.debug({ someObject }) logs 'undefined'
Question:
username_0: ## Bug report
### Describe the bug
`strapi.log.debug()` cannot be used to print objects unless you pass a string as a first argument first. This is counter intuitive, undocumented, and illogical.
### Steps to reproduce the behavior
1. Put `strapi.log.debug({works: false})` in your code
2. Run your code
3. See `undefined` printed rather than your passed object
4. Put `strapi.log.debug("", {works: false})` in your code
5. Run your code
6. See your passed object printed
### Expected behavior
I expect `strapi.log.debug()` to print exactly what I passed, prefixed with a timestamp.
### Screenshots
If applicable, add screenshots to help explain your problem.
### Code snippets
If applicable, add code samples to help explain your problem.
### System
- Strapi version: 3.4.5
Answers:
username_1: Going to move this over to documentation as we use pino for logging and we need to expand the documentation around it.
https://strapi.io/documentation/developer-docs/latest/global-strapi/api-reference.html#strapi-log |
PhonologicalCorpusTools/CorpusTools | 801019360 | Title: Mutual Information on Buckeye Corpus
Question:
username_0: cf. we have a separate branch for this: 'pronunciation_variants'
Answers:
username_1: Hmmm I think this is caused by the mixing of attribute _transcription, Transcription, _transcription_name, and the transcription getter and setter... This issue is related to #756, and we'll need to do some fundamental refactoring of the Word class to solve this type of problems once and for all.
username_0: cf. we have a separate branch for this: 'pronunciation_variants'
username_0: The recent commit, the one that forces the column name (92113c5b02b884e74b6f78a131bcd50cf35553b2) gets ride of the 'Unwanted transcription column' issue (i.e., no 'canonical' column to start with). However, for an independent reason, "List pronunciation variants" is acting out again 😳.
I have added an example file csv_pron_var.txt to the 'csv_sample' folder. |
htmlacademy/yomoyo | 495713467 | Title: JS1. Пропущена запятая в "Код и магия"
Question:
username_0: Это игра, где главного героя, которым вам предстоит управлять и изменять заклинаниями окружающий мир`,` зовут <NAME>. Вместе с ним вас ждет увлекательное приключение…
https://up.htmlacademy.ru/javascript/18/project/code-and-magick
index.html
Answers:
username_1: Спасибо. Поправим.
Status: Issue closed
|
axios/axios | 369649885 | Title: Axios Post Works but Responds with Network Error
Question:
username_0: I searched for this here, around the internet and I went and signed up in the chat but no one's there. No response in a day.
#### Summary
I'm learning so excuse my newb code. I'm using async functions with axios that all work except for one. This one did work before as I never saw this error come up when I implemented it. The `updateItem(item)` method is the one giving an error. What's weird is that I was using `axios.put` no problem but was getting an invalid verb error. Switched to `axios.post` and it does work, it hits the method and sends the item and when the method is finished an Ok(200) response is returned. But an error is thrown on the `axios.post` call of a network error. I'll show the code and error below.
```js
async function deleteSelectedItemsAsync(removedItems) {
try {
let response = await axios.post('/Items/DeleteItems', removedItems);
console.log('Delete response: ' + response.status);
} catch (e) {
console.log(e);
}
}
async function updateItem(item) {
var response;
try {
item.editButtonMessage = "Edit";
response = await axios.post('/Items/UpdateItems', item);
} catch (e) {
console.log(response);
console.log(e);
}
}
async function addItemToDatabaseAsync(item) {
try {
const response = await axios.post('/Items/PostItem', item);
} catch (e) {
alert(e);
}
}
```
#### Context
- axios version: v0.18.0
- Environment: Windows 10, Edge 41.16299.666.0, Chrome Dev 71.0.3573.0 (Official Build) dev (64-bit), and Regular Chrome latest version
Answers:
username_1: Okay! Where does this axios Instance Point?
username_0: I never did get a response in the provided chat room. Any idea here?
username_0: @username_1 Anything else I can provide? Axios is included with a `<script>` reference and then I just use `axios.post` or `get` or whatever.
username_0: :(
username_2: @username_0 I get the error when my API is not accepting CORS.
Do you POST from localhost to an API?
username_0: I will try the suggested steps tomorrow. To answer @username_2, I'm posting from localhost when developing. I'll check and confirm that localhost is the only time I'm getting the error, not on the published site.
Thanks for the responses. Will report what happens.
username_0: I haven't been able to work on what I've wanted to recently but now I've been able to get back to this. I've changed the `try/catch` block situation to use the regular `.then().catch()` and I'm getting the same `Network Error` issue still except developer tools is showing me something different than above. I'm not POSTing to an API but I am running from localhost. I'm posting to an MVC endpoint that's in the same application so it has the same URL which I suppose can be called an API but there should be no cross-origin there.
Below is my code, I tried `error.response.data` @heyamykate but I'm told it's undefined. If I just log out `console.log(error)` I get `Error: Network Error` exactly. But browser developer tools will tell me `net::ERR_INCOMPLETE_CHUNKED_ENCODING 200 (OK)` which is weird that it has the `200 (OK)` part in it.
So the `axios.post()` does work, the endpoint it calls does get hit and the data is sent, the task is performed and then the endpoint returns an `Ok(200)` result. But I still get this error which the catch block catches.
```
axios
.post('/Items/UpdateItems', item)
.then(function(result) {
console.log('Logging result ' + result);
})
.catch(function(error) {
console.log('What happened? ' + error.response.data);
});
```
username_3: do one thing. change url to the page where is just show string and then check still it comes in catch block. if not then there is some problem here '/Items/UpdateItems'.
username_0: I'm not sure that I follow, @username_3.
username_4: @username_0 Not sure if this helps, but I had a similar issue that was caused by axios calling a different port on localhost.
I had a selenium test running on http://localhost:51641, but erroneously POSTing to a development server on http://localhost:8000. :8000 returned a `200`, but axios threw a `Error: Network Error`.
username_5: Having the same error on two local host ports. When looking in dev tools I can see the request returns correct contents but showing CORS error and Network Error
username_6: Having the same error :(
username_7: Having the same issue. Postman does work, axios doesn't
username_8: This sounds a lot like a mixture of errors. I doubt the error lies within Axios though from what I have read. If someone can give an example where it is shown to be a definitive Axios issue please open a new ticket for that issue.
Status: Issue closed
username_9: I ran into an error similar to this. Very baffling. It turned out to be cors related. The options works, and the post works, or at least gets a 200 in dev tools, but you still get a "Network Error..." thrown. It's because the "Access-Control-Allow-Origin" header is missing on the post request. Works fine if you add that.
username_10: In my case, after adding the cors in server, at the php code, I've made this error disappear. Here is the different language system to enable cors -
https://enable-cors.org/server.html
username_11: Hello, could you please describe how you added "Access-Control-Allow-Origin" to the axios post ? Thank you
username_11: Hello, please help me :
'NETWORK ERROR':
https://stackoverflow.com/questions/58717673/axios-sending-network-error-while-status-is-200-ok
username_12: Any luck?
username_13: Not sure how helpful this is for everyone else, but I was experiencing the exact same issue with the login route of my own project. The request succeeded and login worked, but `axios` threw a `networkError`. I also tried the request in Postman, where it worked flawlessly.
For me the issue ended up being in the API itself, _not_ `axios`. Turns out I had a misspelled function within the login route in the backend that I was calling with `await`, and for whatever reason it failed completely silently. It was pretty deeply nested within other `async/await` functions, which could explain why it was so hard to track down. Either way, fixing the misspelling also fixed the issue I was having with `axios` on the frontend.
TL;DR: your issue might be in the route you're calling, not `axios`. Check the backend code for any silly mistakes if you haven't already.
username_14: Maybe... Disable your AD blockers in your browser? ....
https://github.com/axios/axios/issues/383#issuecomment-575408350
username_15: Having same issue, any solutions yet?
username_16: Having the same issue. Postman does work, axios doesn't. I wrote both server side and client side.
Server reduced to this dead simple code;
const httpTrigger: AzureFunction = async function(context: Context, req: HttpRequest): Promise<void> {
if (req.body && req.body.name) {
context.res = {
body: { hello: 'bruce' }
}
} else {
context.res = { status: 400, body: '{"Error":"Bad input."}',}
}
}
Client doing this request:
let objJson = JSON.stringify(planObject)
let url = 'http://localhost:7071/api/MoneyPlan'
let res = await axios.post(url, objJson)
Both client and server running in debug mode (VSCode) on my machine. Client sends Post, Server receives message and returns 200 ok. Client alerts Network Error. This was working a few weeks ago.
username_17: I've found the solution here. If you are testing on android device at localhost, localhost is 10.0.2.2
https://github.com/axios/axios/issues/973#issuecomment-437221047
username_16: Well in my case the problem persisted even when running client and server online with both the web site and server function hosted on Azure. I fixed that problem by using fetch instead of Axios. But even with fetch, I still had the problem running local tests. Right now I'm busy on other things, but I'll try your suggestion later.
username_18: Having same issue axios version: `0.19.0`, any solutions yet?
username_19: Same issue here :[
username_20: 
Then later in the stack:
(last line is related to my throw after i have not found any response in the error object)
Request is correct and works from with client (tried requests.posts in python)
Any hints?
username_21: Same issue here!!
username_22: Try with
`<?php
header('Access-Control-Allow-Origin: *');
header("Access-Control-Allow-Headers: X-API-KEY, Origin, X-Requested-With, Content-Type, Accept, Access-Control-Request-Method");
header("Access-Control-Allow-Methods: GET, POST, OPTIONS, PUT, DELETE");
header("Allow: GET, POST, OPTIONS, PUT, DELETE");`
username_23: 
I have same problem
It works in Postman, but not when a request is sent to axios.
axios version `^0.19.2`
```javascript
const handleUpload = event => {
const data = new FormData();
data.append("image", event.target.files[0])
api.post("images", data)
.then(response => console.log(response))
.catch(errors => console.log(errors));
}
```
username_24: I was getting **Network Error** when making an API request using AXIOS only in Window Server Firefox.Following are the steps to fix the issue.I am using NodeJS server
1- npm install cors
2-const cors = require("cors");
3- app.use(cors());
4- app.use(function(req, res, next) {
res.setHeader("Access-Control-Allow-Origin", "*");
res.setHeader("Access-Control-Allow-Credentials", "true");
res.setHeader(
"Access-Control-Allow-Methods",
"GET,HEAD,OPTIONS,POST,PUT,DELETE"
);
res.setHeader(
"Access-Control-Allow-Headers",
"Origin,Cache-Control,Accept,X-Access-Token ,X-Requested-With, Content-Type, Access-Control-Request-Method"
);
if (req.method === "OPTIONS") {
return res.status(200).end();
}
next();
});
username_25: @username_24 I was having this issue using Mockoon adding the headers defined above resolved the issue for me, thank you.
username_26: if youre trying to reach localhost, just replace the localhost with `10.0.2.2`
https://stackoverflow.com/questions/5528850/how-do-you-connect-localhost-in-the-android-emulator
username_20: So in my case the problem was the following:
1. user auth token expired, and browser sent a regular request to backend (OPTIONS + POST)
2. backend responded with 200 correctly on OPTIONS request, with correct CORS headers
3. then axios sent a second request to main method (in my case it is POST)
4. since token have expired - backend returned 401, but **axios expects CORS headers to be present with this 401 response as well** (see [here ](https://stackoverflow.com/questions/49886315/axios-interceptors-response-undefined) and [here](https://github.com/serverless/serverless/issues/3191)), but backend have not been providing them (i did not know i need to configure CORS headers on AWS API Gateway 4XX responses in addition to regular CORS configuration for every REST Endpoint)
After i added CORS headers to 4xx processing - everything worked well.
username_27: It would be nice if axios said something about CORS if the error stems from CORS.
username_28: It does say that it is a CORS issue, but in Firefox, in Chrome yes it does not show what is wrong. |
angelitomg/QimageComponent | 58052904 | Title: Not generating thumb
Question:
username_0: I am trying but no results and no error in log...
$data['file'] = 'http://fc07.deviantart.net/fs71/f/2012/212/5/2/tumblr_background___loki___avengers_edition_by_miratheracusia-d57swys.gif';
$data['width'] = '75';
$data['output'] = WWW_ROOT.'/files/';
echo 'img is : '.$this->Qimage->resize($data);
Answers:
username_1: The source file must be on the local file system.
Good luck!
username_0: Well. I have write some code for your component and It is working file for url. :)
You can add this function into your component for url
/**
* Method to getFileByUrl
*
* $data['file'] -> web url of image file
* */
public function getFileByUrl($data=''){
#$file = 'http://31.media.tumblr.com/fcb6f027cab033aaf594c76ecad62307/tumblr_myg30z1d2H1t4nrlno1_400.gif';
//debug($data);
if(isset($data['file']))
$isUrl = explode(':/',$data['file']);
else
return false;
if(isset($isUrl[0]))
$img = file_get_contents($data['file']);
if(!empty($img)){
$fileName = explode('/',strrev($data['file']));
$newfile = WWW_ROOT.'/img/'.strrev($fileName[0]);
if(file_put_contents($newfile, $img)){
chmod ($newfile , 0777 );
$data['file'] = $newfile;
}else
return false;
return $data;
}
}
and then modify this code
if (!is_file($data['file'])){
$data = $this->getFileByUrl($data);
if (!is_file($data['file'])){
$this->errors[] = 'Invalid file!';
return false;
}
}
It is 100% working for url :)
Status: Issue closed
username_1: You can create a fork and apply this patch.
Thank you for your contribution.
Cheers. ;) |
gpac/ComplianceWarden | 553275826 | Title: MIAF spec should also test HEIF spec which should test ISO BMFF spec
Question:
username_0: It is possible that I read the code incorrectly, but it looks like when testing the MIAF spec, it is only testing the rules that are related to MIAF. So, to fully verify MIAF compliance one would have to run the program three times: Once for MIAF, once for HEIF and once for ISO BMFF.
Maybe this is intentional, but I think it would be more convenient to have the MIAF spec handler also include all the HEIF test cases, and the HEIF test cases should include all the ISO BMFF test cases.
Answers:
username_1: Agreed. My only concern is that MIAF checks are almost complete whereas HEIF and ISOBMF still require a lot of work (looking for contributors or sponsors!).
Status: Issue closed
username_1: Fixed by 41b4e1197f81dd2fc449f9f28bbc027a214057bc |
bootstrap-vue/bootstrap-vue | 596549107 | Title: b-form-checkbox styled wrong
Question:
username_0: Label of checkbox should styled using 'form-check-label'
https://getbootstrap.com/docs/4.1/components/forms/#checkboxes-and-radios
Answers:
username_1: Class `'form-check-label'` is used in `plain` mode checkboxes, while `'custom-control-label'` is used on the custom styled checkboxes/radios
From Bootstrap v4.4 docs:
- Plain style radios/checks: https://getbootstrap.com/docs/4.4/components/forms/#checkboxes-and-radios
- Custom style radios/checks: https://getbootstrap.com/docs/4.4/components/forms/#checkboxes-and-radios-1
BootstrapVue renders the custom style radios/checks by default. to get the non-custom styled, set the `plain` prop on the radio/check: https://bootstrap-vue.js.org/docs/components/form-checkbox#non-custom-check-inputs-plain
Status: Issue closed
username_0: Thanks! |
barncastle/Battle.Net-Installer | 860569460 | Title: i get this error
Question:
username_0: D:\d2r>BNetInstaller.exe --prod osib --uid osi_beta --lang enus --dir "D:\D2R"
Unhandled exception. System.ComponentModel.Win32Exception (2): The system cannot find the file specified.
at System.Diagnostics.Process.StartWithCreateProcess(ProcessStartInfo startInfo)
at System.Diagnostics.Process.StartCore(ProcessStartInfo startInfo)
at System.Diagnostics.Process.Start()
at System.Diagnostics.Process.Start(ProcessStartInfo startInfo)
at System.Diagnostics.Process.Start(String fileName, String arguments)
at BNetInstaller.AgentApp.StartProcess()
at BNetInstaller.AgentApp..ctor()
at BNetInstaller.Program.Run(Options options)
at BNetInstaller.Program.Main(String[] args)
at BNetInstaller.Program.<Main>(String[] args)
Answers:
username_1: The tool is failing to start Battle.net for some reason. Do you have
Battle.net installed (link is in the README)? If you do could you try
running as admin?
Also out of curiosity, what version of the tool are you using? Is it 1.3 or
1.4?
Status: Issue closed
username_0: thank for reply, i forgot open batle.net |
gmusic-utils/gmusic-theme.js | 147935500 | Title: Google just rethemed Google Play Music
Question:
username_0: If you guys head over to GPM in your browser (https://play.google.com/music/listen?u=0#/now) you should see that Google has changed a few things, most notably their standard orange color has changed and their icon/logo has changed.
I'm currently in the progress of doing a test to ensure that the theme still works for everything else but in particular that orange color will need to be updated
/cc @gmusic-utils/gmusic-js
Answers:
username_1: Tangent: I'm going to push a test branch on `gmusic.js` to make sure Travis CI is still green there
username_1: Woot, no errors in `gmusic.js` https://travis-ci.org/gmusic-utils/gmusic.js/builds/122689800
Status: Issue closed
|
NuGet/NuGetGallery | 757335032 | Title: Cannot convert account to organization
Question:
username_0: ### Describe the bug
A clear and concise description of what the bug is.
### To Reproduce
Steps to reproduce the behavior:
1. Log with my account `losol`, go to this URL 'https://www.nuget.org/account/transform.', add username_0 as new admin
1. Log in as username_0, go to `https://www.nuget.org/account/Organizations`
2. Perform this operation: accept joining organization by clicking the V
3. See error: Oops! Error 500 - Internal Server Error. An error occurred while processing your request. We really messed up this time...
### Expected behavior
Joining losol organization
Answers:
username_1: Hey @username_0 and @username_2 (from https://github.com/NuGet/NuGetGallery/issues/8349), this is a legitimate bug on nuget.org right now. Thanks for bringing this to our attention. We are working quickly to get a fix out to unblock your experience.
username_2: As a work around ... individual packages can be reassigned successfully.
username_1: @username_0 and @username_2, a fix has been shipped to nuget.org with a fix. You should be able to convert your account into an organization now.
Please go ahead and give it a try again and let me know if you're still blocked.
username_0: Thanks, @username_1. Confirming that this bug is fixed, and that the organization conversion succeeded. cc @username_2
Status: Issue closed
|
DevExpress/testcafe | 227957155 | Title: Test hangs if browser redirected to non-page
Question:
username_0: ### Are you requesting a feature or reporting a bug?
Bug
### What is the current behavior?
If you are redirected by tested page to some non-HTML document (e.g. SVG, PNG, or PDF), or use such document in a fixture as the test page, the document will be loaded, but test will hang because we don't proxy such documents.
### What is the expected behavior?
Show an error with a message that will explain that we don't support such documents as test pages.
### How would you reproduce the current behavior (if this is a bug)?
Run test
#### Provide the test code and the tested page URL (if applicable)
Tested page URL:
Test code
```js
fixture `Redirect`
.page `http://localhost:81/test.html`;
test('1',async t => {
await t
.click('a')
.expect(true).ok();
});
```
### Specify your
* operating system: Windows 10
* testcafe version: 0.15.0
* node.js version: 7.4.0
Status: Issue closed
Answers:
username_1: It's synthetic scenario. We will wait until a real-user case.
username_2: I have run into this today while trying to use testcafe to test an application which heavily uses PDF export functionality. I'm not expecting to be able to do assertions on the content of a PDF or anything like that, but it would be essential for my use case to be able to test that a PDF is generated and displayed in the browser.
username_1: ### Are you requesting a feature or reporting a bug?
Bug
### What is the current behavior?
If you are redirected by tested page to some non-HTML document (e.g. SVG, PNG, or PDF), or use such document in a fixture as a test page, the document will be loaded, but the test will hang because we don't proxy such documents.
### What is the expected behavior?
Show an error with a message that will explain that we don't support such documents as test pages.
### How would you reproduce the current behavior (if this is a bug)?
Run test
#### Provide the test code and the tested page URL (if applicable)
Tested page URL: https://andreybelym.github.io/test-pages/gh-1471.html
Test code
```js
fixture `Redirect`
.page `https://andreybelym.github.io/test-pages/gh-1471.html`;
test('1',async t => {
await t
.click('a')
.expect(true).ok();
});
```
### Specify your
* operating system: Windows 10
* testcafe version: 0.15.0
* node.js version: 7.4.0
username_3: +1 for this.
I'm behind a corporate proxy which prevents certain third-party scripts from loading properly (such as Google Analytics, Google reCAPTCHA, etc) - since the initial response contains a login/redirect page.
The only way around this is to visit each script URL we need directly, log in and/or wait for it to redirect, and then go back to the original site we're developing for. After that the script URLs all load properly for the duration of the browser session.
If TestCafe could allow me to do something like the following it would be incredibly useful:
```
await t
.navigateTo("some/path/to/file.js")
.navigateTo("another/path/to/file.js")
.navigateTo("mysite.com")
```
TestCafe is an amazing tool btw. Keep up the great work!
username_4: Yes, this is what I'm experiencing as well. I hope you will be able to add support to Non Html Files.
https://stackoverflow.com/questions/52361644/testcafe-cant-take-screenshot-when-opening-pdf
username_5: Hi, having a similar problem with this. Some of our error messages come with type "text/plain" so when testcafe visits a broken page, instead of marking it as failed and continuing, it breaks the entire test. Is there any workaround to this? I've tried setting the timeout, but it doesn't help.
username_6: @username_5
Hi,
There is no workaround. This issue is in the [Enhancement processing](https://github.com/DevExpress/testcafe/projects/11#card-18273858) project. Please stay tuned to our updates here.
username_7: +1 for it, I also need to go through user flow, where in one step is generating PDF in another window (but due to fact, that TestCafe didn't allow for multi tabs, it's loaded in main window) so it crash test execution.
Although you only want go through and do `navigate(url)` to another place, still didn't work. 💔
username_8: @username_7
Thank you for your feedback.
Also I want to mention that we have plans to implement multiple window support according to the [testcafe roadmap](https://devexpress.github.io/testcafe/roadmap/)
username_9: I also have need of this feature.
username_10: yes please, this is a deal breaker for some projects.
username_11: s same situation like in testcase described by @username_7 so please consider it faster ;-)
username_0: Also, we have a problem with file downloading dialogs in Firefox & IE11: https://github.com/DevExpress/testcafe/issues/2741. When we fix this issue, it's better to cover #2741 too.
username_12: Any updates on this? run into this problem a couple times as well.
username_1: Our developers will look into this issue in one of the next springs. You will be automatically notified once we make progress. Your patience is greatly appreciated.
username_13: We would appreciate it too, as we are handling with some long URL during multiple redirects.
username_14: When this issue will be solved? A sprint is common 7-14 days. In my eyes it is not. bug this is definitely an incident because it blocks your customer. So if this will take a year more, it was reported in 2018 as I read, I have to look for another solution. Be fair and please give an estimation.
username_15: Sorry, we cannot give any estimates. This issue is not planned for the current sprint, because we are focused on other tasks right now. And of course, your PR would be highly appreciated.
username_16: I have the same problem, after login I got redirect to a page which shows some JSON. I am successfully logged in but everything freezes so I can't navigate to the html page.
username_17: Hello,
We have not implemented this enhancement yet. Please design your test so that you don't end up on non-html pages.
username_18: For this issue when browser hangs, testcafe restarts it but it only does it 3 times and then stops everything. Maybe if you allow an option to make testcafe restart a set number of times or infinite restarts, that should work for me.
Also, is there an interrupt handler that will notify if testcafe is about to go in non-html page so I can avoid going there? That would also help since I have a lot of tests and manually checking /updating them will take a very long time.
username_19: Hello @username_18 ,
Thank you for sharing your idea with us. We will take it into account.
As a workaround, try to use the [RequestMock](https://devexpress.github.io/testcafe/documentation/reference/test-api/requestmock/) API. It will allow you to check if a requested page is non-HTML and redirect to another page. See the following test example:
```js
import { RequestMock } from 'testcafe';
fixture `Redirect`
.page `https://andreybelym.github.io/test-pages/gh-1471.html`;
test('1', async t => {
const mock = RequestMock()
.onRequestTo(/\.svg$/)
.respond('<script>window.location="https://google.com"');
await t
.addRequestHooks(mock)
.click('a')
.expect(true).ok();
});
```
username_18: Hello @username_19,
Thank you very much for your reply. I will try your suggestion.
However, due to the high number of occurrences of non html pages in the test suite, it will take quite a bit of time to figure out all the URIs that return non-html pages even with regex since they can be all sorts of format right?
So, I wanted to ask if there is a way to modify testcafe code instead to increase the number of allowed restarts from 3 to infinite?
For example, where in the code can I update that number 3, then build it and install it for my system? I could not find the section of code I need to update...
username_20: Hello @username_18,
Yes, the workaround implies that you know in advance which non-HTML pages TestCafe will be redirected to. I think this solution is more stable than relying on multiple TestCafe restarts when your browser hangs. If you still want to try that, you can change the threshold value [here](https://github.com/DevExpress/testcafe/blob/faec4a93682d51a6f51909514b437236899b08c1/src/runner/test-run-controller.ts#L17) and build a custom version of TestCafe using [Build Instructions](https://github.com/DevExpress/testcafe/blob/master/CONTRIBUTING.md#build-instructions) from our contributing guide.
username_18: Hello @username_20,
Thank you so much for this! That helped perfectly.
username_18: Hello @username_20,
Regarding this issue, I was also wondering how to modify the timeout value such that it only waits a short amount of time before restarting the browser. I tried setting the --ajax-request-timeout and --page-request-timeout but that does not affect the default 2 mins wait time for the browser restart case...
Thank you in advance!
username_21: Hi all,
I am trying to reduce flakiness and optimize my tests where we currently create a job via UI and after the job is created we verify it was created with correct values. Verification can be done using UI but page is opening slow and is affected by other functionality that we don't want to test in the same test. The best for optimization is to split creation via UI and checking created values via opening API endpoint, e.g _/api/jobKey_ that has **Content-Type: application/json;charset=UTF-8**. Currently TestCafe hands when we open such page with json content and I can't reuse browser's authorization to just open /api/jobKey and read response as json.
What do you think is this improvement will be worked on some time soon? Maybe someone can suggest what I could miss and have a solution for opening in TestCafe test **Content-Type: application/json;charset=UTF-8** page and reusing authorization from browser to open it?
username_19: Hi @username_21,
You can programmatically make a request to your API endpoint from the TestCafe test and then check the received JSON data. Please take a look at the answer in the following thread on StackOverflow: https://stackoverflow.com/questions/62271173/make-a-real-http-request-using-testcafe
Status: Issue closed
|
rParslow/TeamWhisky | 237815237 | Title: NEW ZEALAND WHISKY (THE) 21 ans High Wheeler43%
Question:
username_0: NEW ZEALAND WHISKY (THE) 21 ans High Wheeler 43%<br>
http://ift.tt/2sXlO6y<br>
#TeamWhisky NEW ZEALAND WHISKY (THE) 21 ans High Wheeler 43% Single Malt Nouvelle Zelande LMDW http://ift.tt/2sXlO6y 82 € <img src="http://ift.tt/2twyNZJ"><br><br>
via Fishing Reports http://ift.tt/2dm5cfF<br>
June 22, 2017 at 01:26PM |
flutter/flutter | 679667621 | Title: FileSystemException: FileSystemException: Deletion failed,
Question:
username_0: Flutter crash report.
Please report a bug at https://github.com/flutter/flutter/issues.
## command
flutter pub get
## exception
FileSystemException: FileSystemException: Deletion failed, path = '/Users/xuhaiqing/Desktop/MyKey/mykey_flutter/.ios' (OS Error: Permission denied, errno = 13)
```
#0 _Directory._deleteSync (dart:io/directory_impl.dart:201:7)
#1 FileSystemEntity.deleteSync (dart:io/file_system_entity.dart:455:47)
#2 ForwardingFileSystemEntity.deleteSync (package:file/src/forwarding/forwarding_file_system_entity.dart:72:16)
#3 ForwardingFileSystemEntity.deleteSync (package:file/src/forwarding/forwarding_file_system_entity.dart:72:16)
#4 _deleteIfExistsSync (package:flutter_tools/src/project.dart:886:15)
#5 IosProject._regenerateFromTemplateIfNeeded (package:flutter_tools/src/project.dart:596:5)
#6 IosProject.ensureReadyForPlatformSpecificTooling (package:flutter_tools/src/project.dart:547:11)
#7 FlutterProject.ensureReadyForPlatformSpecificTooling (package:flutter_tools/src/project.dart:246:17)
<asynchronous suspension>
#8 PackagesGetCommand.runCommand (package:flutter_tools/src/commands/packages.dart:128:23)
<asynchronous suspension>
#9 FlutterCommand.verifyThenRunCommand (package:flutter_tools/src/runner/flutter_command.dart:907:18)
#10 _rootRunUnary (dart:async/zone.dart:1198:47)
#11 _CustomZone.runUnary (dart:async/zone.dart:1100:19)
#12 _FutureListener.handleValue (dart:async/future_impl.dart:143:18)
#13 Future._propagateToListeners.handleValueCallback (dart:async/future_impl.dart:696:45)
#14 Future._propagateToListeners (dart:async/future_impl.dart:725:32)
#15 Future._completeWithValue (dart:async/future_impl.dart:529:5)
#16 Future._asyncCompleteWithValue.<anonymous closure> (dart:async/future_impl.dart:567:7)
#17 _rootRun (dart:async/zone.dart:1190:13)
#18 _CustomZone.run (dart:async/zone.dart:1093:19)
#19 _CustomZone.runGuarded (dart:async/zone.dart:997:7)
#20 _CustomZone.bindCallbackGuarded.<anonymous closure> (dart:async/zone.dart:1037:23)
#21 _microtaskLoop (dart:async/schedule_microtask.dart:41:21)
#22 _startMicrotaskLoop (dart:async/schedule_microtask.dart:50:5)
#23 _runPendingImmediateCallback (dart:isolate-patch/isolate_patch.dart:118:13)
#24 _RawReceivePortImpl._handleMessage (dart:isolate-patch/isolate_patch.dart:169:5)
```
## flutter doctor
```
[✓] Flutter (Channel master, 1.21.0-10.0.pre.110, on Mac OS X 10.15.6 19G2021, locale zh-Hans-CN)
• Flutter version 1.21.0-10.0.pre.110 at /Users/xuhaiqing/FlutterGit/flutter
• Framework revision 2b25e29192 (25 hours ago), 2020-08-15 01:06:03 +0100
• Engine revision 7571e7c380
• Dart version 2.10.0 (build 2.10.0-11.0.dev)
• Pub download mirror https://pub.flutter-io.cn
• Flutter download mirror https://storage.flutter-io.cn
[✓] Android toolchain - develop for Android devices (Android SDK version 29.0.3)
• Android SDK at /Users/xuhaiqing/Library/Android/sdk
• Platform android-29, build-tools 29.0.3
• Java binary at: /Applications/Android Studio.app/Contents/jre/jdk/Contents/Home/bin/java
• Java version OpenJDK Runtime Environment (build 1.8.0_242-release-1644-b3-6222593)
• All Android licenses accepted.
[✓] Xcode - develop for iOS and macOS (Xcode 11.6)
• Xcode at /Applications/Xcode.app/Contents/Developer
• Xcode 11.6, Build version 11E708
• CocoaPods version 1.9.3
[✓] Android Studio (version 4.0)
• Android Studio at /Applications/Android Studio.app/Contents
• Flutter plugin version 46.0.2
• Dart plugin version 193.7361
• Java version OpenJDK Runtime Environment (build 1.8.0_242-release-1644-b3-6222593)
[✓] VS Code (version 1.48.0)
• VS Code at /Applications/Visual Studio Code.app/Contents
• Flutter extension version 3.13.2
[✓] Connected device (1 available)
• xu的 iPhoneg (mobile) • 00008030-00022C58229A802E • ios • iOS 14.0
! Error: SH is not connected. Xcode will continue when SH is connected. (code -13)
• No issues found!
```
Answers:
username_1: **Errno13** is a ownership error i beleive.
Try runnning with sudo instead,```sudo flutter pub get```
username_2: @username_0 Is this an add-to-app module?
Status: Issue closed
username_3: Fixed via https://github.com/flutter/flutter/commit/9248fda41005ce0af2e522fabd0e9fedf2b3c863 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.