showvikdbz/code-reviews · Datasets at Hugging Face

text stringlengths 1 5.2k	labels listlengths 1 1
this is a bit expensive at the moment because it is called twice on serialization the way to get around this is to use setupcontext teardowncontext this seems to keep coming up maybe we should just fix	[ 0 ]
This is not really a valid protection for pointers those bastards get cast all the time :smile:. Even in this PR there are a few castPtr	[ 1 ]
(in particular as the efi var is actually defined as part of https://systemd.io/BOOT_LOADER_INTERFACE/ and the relation to TSC is not documented there. Hence assuming the variable is rubbish is a bit of a stretch, given that it might or might not be backed by TSC, could be backed by anything else really, we shouldn't care about it)	[ 1 ]
keystone admin port	[ 0 ]
dumb nit should this be logged before the exception i guess i am fine as it is because it shows the error right after the traceback and you do not need to scroll up much to see what the problem is in a ideal world	[ 0 ]
Yuck! The handling of the replies to "outdated" conversations seems pretty horrible. In "conversation" mode they remain attached to the original change, while in "files changed" mode they don't show up at all. Sorry this seems to be kind of a mess. I also have no idea why these got tagged as "outdated".	[ 1 ]
Holy shit all it's missing is a "And then everyone in the classroom clapped."	[ 1 ]
Thank you so much for this; this is a really nice little feature for when shit really hits the fan.	[ 1 ]
you re going to want to kill me but redrobot said these should be in the format assertequal expected returned he dinged me on that for mine i do not know if worth ask redrobot	[ 0 ]
condiser using httplib instead of immediate values	[ 0 ]
elif nit consider using end message not printed instead saves the negation nit consider flipping the order of the conjuncts boolean is cheaper	[ 0 ]
don't do world.time SECONDS, just do world.time * 10. This looks incredibly confusing to someone who might not understand it, and it looks really ugly in general.	[ 1 ]
you should not import individual functions from freezer import winutils winutils is windows would be better approach imports	[ 0 ]
this decorator should be removed	[ 0 ]
changed resp self ssh lsfcconsistgrp fc map id keep this function the same	[ 0 ]
oops i made a stupid mistake i am discussing with jay pipes in another patch sorry for this	[ 0 ]
it s the idea we want to monitor infinitely for proc pid to die	[ 0 ]
could you wrap in a function defer func cmd kill cmd wait	[ 0 ]
Oh, if you think 9223372036854775807 is a type mismatch... It is not, it is just a nonsense stride to expose these kinds of issues :)	[ 1 ]
The way I detected data loss is I summed the data being pushed and data being received and it was exactly half, however this might have been an issue caused by electron's node integration, rather than node itself, this definitely fixed it however, reproducing is a bitch however.	[ 1 ]
ridiculous that this super important feature is not even answered and I am paying for this	[ 1 ]
stupid question maybe but is there a reason this health check is sent over the message bus	[ 0 ]
it is an useless class	[ 0 ]
interesting does exit bypass atexit registered hooks i guess you do not get the dumps written unless you explicitly call llvm profile write file	[ 0 ]
i hoped to be able to output log messages which is not possible at the moment it seems a bit cleaner to kill off your goroutines before falling off the end of main i am not sure whether go does anything there anyway i have removed for now	[ 0 ]
this if term is active is now useless please remove	[ 0 ]
nit starting c this could be done directly in the declaration bool have token false	[ 0 ]
what the hell is this comment	[ 1 ]
test attach interface has a mock decorator calling it with like this will override the mocked argument call with key value arguments instead	[ 0 ]
tenant id is a kwarg let us use debtcollector removals removed kwarg here or oslo log versionutils report deprecated feature	[ 0 ]
this test failed on my saio because i already had a container in the test account is it possible to perform the check and be tolerant of existing containers	[ 0 ]
Ah. It's github fucking up. Sorry.	[ 1 ]
you can use get template function here	[ 0 ]
sometimes directories turn into byte files when there is xfs filesystem corruption mostly it is just paranoia you get enough clusters with enough different operators and someone s gonna do something dumb in here	[ 0 ]
Damn, you're right. Good catch. :(	[ 1 ]
added the mocks to the bug even though the mocks are only available on desktop and do not include managed prefs this change in particular will make the experience on android consistent with what we have today on desktop the page is accessible but controls are disabled the specific change has been recommended by rfeng our uxer	[ 0 ]
because no separator (space) should be no separator not an empty string. Everything in dm is nullable so I don't get the "typefucking" comment.	[ 1 ]
Could probably do it with a precompile step, but that starts getting into "clusterfuck of a buildsystem" territory.	[ 1 ]
hardcoded urls should be removed	[ 0 ]
I copy-pasta'd some code and now I feel bad. That said, I took off the default return ..() so help me god if I fucked up everything.	[ 1 ]
i was actually removing that while you wrote this, ain't that some shit	[ 1 ]
how? the old one is while attached_hand charge_counter = 0 and sle- oh oh oh wait fuck woops	[ 1 ]
enable security gruop should be enable security group	[ 0 ]
Supervolcano means evacuating more than just the impact area.... at least a quarter of the continental US is screwed by the Yellowstone caldera, and I can't imagine this would be different.	[ 1 ]
oh god why did that slip by me	[ 1 ]
i wish we could get rid of that automagic thing but i guess we have to live with this forever	[ 0 ]
use patcher for these mocks you have a plenty of methods with such mocks so it is better to initialize them in setup method	[ 0 ]
I didn't do this in my branch when I worked on it but holy shit this snowflakey shit for windows really should be done differently. This doesn't even work with railings atm.	[ 1 ]
ter/furn ids are weird like that. I think we'll need a header for this crap, it's messy. For now, it's no big deal.	[ 1 ]
missing field drive size initializer werror wmissing field initializers cgptaddparams params	[ 0 ]
Why a fix should be tied to a vote, this is ridiculous... There are lots of us with this issue, are any docs there describing how to fix it?	[ 1 ]
Sure, and if your code did a bunch of obfuscated or wacky things, I wouldn't have merged it (though I was probably bad to not allow PEP8 to finish - ahem).	[ 1 ]
good point how silly of me	[ 0 ]
@ninjanomnom adding all this brewery shit is gonna make the edible component incompatible. Shouldn't this brew shit be in a component?	[ 1 ]
ifdef config cgroup cpuacct else static inline int blah endif	[ 0 ]
hot damn lol. are you sure we can't do this more like `background: platformColors[platform].background`?	[ 1 ]
Who is newline, and why did the shit my pants????	[ 1 ]
i am not sure what you mean by reference error how about from line if lock file prefix if lock file prefix endswith lock file name s s lock file prefix safe name else lock file name s s lock file prefix safe name else lock file name safe name	[ 0 ]
add expect false using a path id that does not exist	[ 0 ]
Also, do me a solid and swap it to this: ``\t\tplaysound(A, 'sound/machines/twobeep.ogg', 50, 1) \t\tA.audible_message("[icon2html(A, hearers(A))] [fuck]", null, 1)``	[ 1 ]
english is not my native language but it looks like the message should be refined for example ipset utility does not exist on the host that you are running please	[ 0 ]
Well well, didn't find this issue when searching for it before. But now that I created a new one, there it "suddenly" :) is: #5788. Closing this one as duplicate of #5788. Solution: Increase precision. This sucks though, as it applies to all measurements, causing the database to eat up more disk space.	[ 1 ]
He's saying you're making people suicidal by wearing a mask. You've set the number too high. Also the message is fucking terrible.	[ 1 ]
then let us fix nailgun agent i really don like to see that data assumptions so deep in the code boolean update by agent must die or used at least somewhere really near the http handler not so deep in the code	[ 0 ]
lol this was dumb this exchange is hardcoded for rabbit t t i will add a config option like we have for other nodes	[ 0 ]
this code moved with some modification to test image util sh	[ 0 ]
damn, this `'...'` thing is magic	[ 1 ]
os brick is not usable in kilo so this comment should probably be removed	[ 0 ]
Blend is for fucking fags, stop using it	[ 1 ]
hi marios so far this patch is just a test to check whether this can mitigate lock wait timeouts the aim of this patch is to validate the hypothesis regarding these lock wait timeout failures honestly i have never been a fan of this approach and i would push for this patch to be merged only if we are unable to do anything better	[ 0 ]
The INL network is ridiculous. I'm not afraid; I'm merging	[ 1 ]
hmm, what precisely fails there, I have trouble reading the output...It appears to fail in the umount loop? Unfortunately the log data is all suppressed because it contains weird chars apparently. Do you get that data from journalctl? if so, could you add "--all" to its invocation, so that it prints even the rubbish it got?	[ 1 ]
There seems to be an awful lot of non-include path related activity in that commit..	[ 1 ]
prefer you use delete key as there is already a delete group delete host etc	[ 0 ]
Learn to interop with existing toolset folks! This is a ridiculous debate.	[ 1 ]
that sounds like it is blocking the user right my concern is it is pretty easy for users to type a dumb search e g search for the letter a accidentally and then be stuck in a frozen browser what do you think	[ 0 ]
this info seems useless	[ 0 ]
Same here. Its ridiculous! Is it so hard to filter scammers? I think People who have github accounts can protect their data...	[ 1 ]
name	[ 0 ]
if extract thread list is enabled it splits the raw data and the ps dump as well as the procfs dump previously it returned just the raw data and the mappings from pid to cmdline i went ahead and modified it so it returned the rest of the text dump as well which contains our proc dump and any future dumps that get added later any future text dumps will have to implement two cases either extraxt cmdline extracts the raw data or it doesn t then it gets passed to extract tgid which will either extract the raw data or it wont as you can see this is going to get cumbersome i will fix this now and have the raw data be extracted before any sort of text extraction happens not sure why this was not done in the first case	[ 0 ]
You need emag to make blood bro, so cost is upped, other way you get stupid clones that basically do nothing but impersonating some random dude they are clones of.	[ 1 ]
Still not quite right. It's now very easy to forget the `&` and pass random crap to `free`. Using `omf_free` should look _exactly_ like `free` to avoid this.	[ 1 ]
Jesus Christ. Monthly remember that we need outfit datums to fix this kind of shit.	[ 1 ]
move browser view into this block or inline getbrowserviewforbrowser call	[ 0 ]
Just to avoid multiple calls on the same frame, yeah. Can be removed if you think it's ugly.	[ 1 ]
maybe can we go for non admin examples or is that silly	[ 0 ]
And rest of request is nonsense mostly	[ 1 ]
need to add the secondary index vars otherwise puppet will complain so should add secondary index false secondary index type lucene	[ 0 ]
<shakes fist> darn you markdown!	[ 1 ]
useless empty line at end of file removal json dump does not create empty line at end of file json specification simply does not care	[ 0 ]
this is copy pasted verbatim from the neutron patch but it is silly just do for event in self pending events	[ 0 ]
please move it before putils execute invoke	[ 0 ]
It's been almost three years since opencv got java bindings. How the hell is the maven support still missing? This is ridiculous.	[ 1 ]
ideally it should be le agent cache inconsistency port id is not stored for s port mac	[ 0 ]
why admin creds are used for share creation	[ 0 ]
>Razor Component component model Please no! That just looks ridiculous \U0001f61d.	[ 1 ]
This is going to introduce a delay to using the status bar, god no. This is also going to be absurdly expensive, every time your mouse fucking MOVES it calls an extra proc for no reason.	[ 1 ]
Good point, I'll add some handling shit for that to end the fulton attempt there if it gets anchored before the fulton starts	[ 1 ]
this should be just using trans instead of blocktrans	[ 0 ]
seriously what the fuck are you doing	[ 1 ]

this is a bit expensive at the moment because it is called twice on serialization the way to get around this is to use setupcontext teardowncontext this seems to keep coming up maybe we should just fix

[ 0 ]

This is not really a valid protection for pointers those bastards get cast all the time :smile:. Even in this PR there are a few castPtr

[ 1 ]

(in particular as the efi var is actually defined as part of https://systemd.io/BOOT_LOADER_INTERFACE/ and the relation to TSC is not documented there. Hence assuming the variable is rubbish is a bit of a stretch, given that it might or might not be backed by TSC, could be backed by anything else really, we shouldn't care about it)

[ 1 ]

keystone admin port

[ 0 ]

dumb nit should this be logged before the exception i guess i am fine as it is because it shows the error right after the traceback and you do not need to scroll up much to see what the problem is in a ideal world

[ 0 ]

Yuck! The handling of the replies to "outdated" conversations seems pretty horrible. In "conversation" mode they remain attached to the original change, while in "files changed" mode they don't show up at all. Sorry this seems to be kind of a mess. I also have no idea why these got tagged as "outdated".

[ 1 ]

Holy shit all it's missing is a "And then everyone in the classroom clapped."

[ 1 ]

Thank you so much for this; this is a really nice little feature for when shit really hits the fan.

[ 1 ]

you re going to want to kill me but redrobot said these should be in the format assertequal expected returned he dinged me on that for mine i do not know if worth ask redrobot

[ 0 ]

condiser using httplib instead of immediate values

[ 0 ]

elif nit consider using end message not printed instead saves the negation nit consider flipping the order of the conjuncts boolean is cheaper

[ 0 ]

don't do world.time SECONDS, just do world.time * 10. This looks incredibly confusing to someone who might not understand it, and it looks really ugly in general.

[ 1 ]

you should not import individual functions from freezer import winutils winutils is windows would be better approach imports

[ 0 ]

this decorator should be removed

[ 0 ]

changed resp self ssh lsfcconsistgrp fc map id keep this function the same

[ 0 ]

oops i made a stupid mistake i am discussing with jay pipes in another patch sorry for this

[ 0 ]

it s the idea we want to monitor infinitely for proc pid to die

[ 0 ]

could you wrap in a function defer func cmd kill cmd wait

[ 0 ]

Oh, if you think 9223372036854775807 is a type mismatch... It is not, it is just a nonsense stride to expose these kinds of issues :)

[ 1 ]

The way I detected data loss is I summed the data being pushed and data being received and it was exactly half, however this might have been an issue caused by electron's node integration, rather than node itself, this definitely fixed it however, reproducing is a bitch however.

[ 1 ]

ridiculous that this super important feature is not even answered and I am paying for this

[ 1 ]

stupid question maybe but is there a reason this health check is sent over the message bus

[ 0 ]

it is an useless class

[ 0 ]

interesting does exit bypass atexit registered hooks i guess you do not get the dumps written unless you explicitly call llvm profile write file

[ 0 ]

i hoped to be able to output log messages which is not possible at the moment it seems a bit cleaner to kill off your goroutines before falling off the end of main i am not sure whether go does anything there anyway i have removed for now

[ 0 ]

this if term is active is now useless please remove

[ 0 ]

nit starting c this could be done directly in the declaration bool have token false

[ 0 ]

what the hell is this comment

[ 1 ]

test attach interface has a mock decorator calling it with like this will override the mocked argument call with key value arguments instead

[ 0 ]

tenant id is a kwarg let us use debtcollector removals removed kwarg here or oslo log versionutils report deprecated feature

[ 0 ]

this test failed on my saio because i already had a container in the test account is it possible to perform the check and be tolerant of existing containers

[ 0 ]

Ah. It's github fucking up. Sorry.

[ 1 ]

you can use get template function here

[ 0 ]

sometimes directories turn into byte files when there is xfs filesystem corruption mostly it is just paranoia you get enough clusters with enough different operators and someone s gonna do something dumb in here

[ 0 ]

Damn, you're right. Good catch. :(

[ 1 ]

added the mocks to the bug even though the mocks are only available on desktop and do not include managed prefs this change in particular will make the experience on android consistent with what we have today on desktop the page is accessible but controls are disabled the specific change has been recommended by rfeng our uxer

[ 0 ]

because no separator (space) should be no separator not an empty string. Everything in dm is nullable so I don't get the "typefucking" comment.

[ 1 ]

Could probably do it with a precompile step, but that starts getting into "clusterfuck of a buildsystem" territory.

[ 1 ]

hardcoded urls should be removed

[ 0 ]

I copy-pasta'd some code and now I feel bad. That said, I took off the default return ..() so help me god if I fucked up everything.

[ 1 ]

i was actually removing that while you wrote this, ain't that some shit

[ 1 ]

how? the old one is while attached_hand charge_counter = 0 and sle- oh oh oh wait fuck woops

[ 1 ]

enable security gruop should be enable security group

[ 0 ]

Supervolcano means evacuating more than just the impact area.... at least a quarter of the continental US is screwed by the Yellowstone caldera, and I can't imagine this would be different.

[ 1 ]

oh god why did that slip by me

[ 1 ]

i wish we could get rid of that automagic thing but i guess we have to live with this forever

[ 0 ]

use patcher for these mocks you have a plenty of methods with such mocks so it is better to initialize them in setup method

[ 0 ]

I didn't do this in my branch when I worked on it but holy shit this snowflakey shit for windows really should be done differently. This doesn't even work with railings atm.

[ 1 ]

ter/furn ids are weird like that. I think we'll need a header for this crap, it's messy. For now, it's no big deal.

[ 1 ]

missing field drive size initializer werror wmissing field initializers cgptaddparams params

[ 0 ]

Why a fix should be tied to a vote, this is ridiculous... There are lots of us with this issue, are any docs there describing how to fix it?

[ 1 ]

Sure, and if your code did a bunch of obfuscated or wacky things, I wouldn't have merged it (though I was probably bad to not allow PEP8 to finish - ahem).

[ 1 ]

good point how silly of me

[ 0 ]

@ninjanomnom adding all this brewery shit is gonna make the edible component incompatible. Shouldn't this brew shit be in a component?

[ 1 ]

ifdef config cgroup cpuacct else static inline int blah endif

[ 0 ]

hot damn lol. are you sure we can't do this more like `background: platformColors[platform].background`?

[ 1 ]

Who is newline, and why did the shit my pants????

[ 1 ]

i am not sure what you mean by reference error how about from line if lock file prefix if lock file prefix endswith lock file name s s lock file prefix safe name else lock file name s s lock file prefix safe name else lock file name safe name

[ 0 ]

add expect false using a path id that does not exist

[ 0 ]

Also, do me a solid and swap it to this: ``\t\tplaysound(A, 'sound/machines/twobeep.ogg', 50, 1) \t\tA.audible_message("[icon2html(A, hearers(A))] *[fuck]*", null, 1)``

[ 1 ]

english is not my native language but it looks like the message should be refined for example ipset utility does not exist on the host that you are running please

[ 0 ]

Well well, didn't find this issue when searching for it before. But now that I created a new one, there it "suddenly" :) is: #5788. Closing this one as duplicate of #5788. Solution: Increase precision. This sucks though, as it applies to all measurements, causing the database to eat up more disk space.

[ 1 ]

He's saying you're making people suicidal by wearing a mask. You've set the number too high. Also the message is fucking terrible.

[ 1 ]

then let us fix nailgun agent i really don like to see that data assumptions so deep in the code boolean update by agent must die or used at least somewhere really near the http handler not so deep in the code

[ 0 ]

lol this was dumb this exchange is hardcoded for rabbit t t i will add a config option like we have for other nodes

[ 0 ]

this code moved with some modification to test image util sh

[ 0 ]

damn, this `'...'` thing is magic

[ 1 ]

os brick is not usable in kilo so this comment should probably be removed

[ 0 ]

Blend is for fucking fags, stop using it

[ 1 ]

hi marios so far this patch is just a test to check whether this can mitigate lock wait timeouts the aim of this patch is to validate the hypothesis regarding these lock wait timeout failures honestly i have never been a fan of this approach and i would push for this patch to be merged only if we are unable to do anything better

[ 0 ]

The INL network is ridiculous. I'm not afraid; I'm merging

[ 1 ]

hmm, what precisely fails there, I have trouble reading the output...It appears to fail in the umount loop? Unfortunately the log data is all suppressed because it contains weird chars apparently. Do you get that data from journalctl? if so, could you add "--all" to its invocation, so that it prints even the rubbish it got?

[ 1 ]

There seems to be an awful lot of non-include path related activity in that commit..

[ 1 ]

prefer you use delete key as there is already a delete group delete host etc

[ 0 ]

Learn to interop with existing toolset folks! This is a ridiculous debate.

[ 1 ]

that sounds like it is blocking the user right my concern is it is pretty easy for users to type a dumb search e g search for the letter a accidentally and then be stuck in a frozen browser what do you think

[ 0 ]

this info seems useless

[ 0 ]

Same here. Its ridiculous! Is it so hard to filter scammers? I think People who have github accounts can protect their data...

[ 1 ]

name

[ 0 ]

if extract thread list is enabled it splits the raw data and the ps dump as well as the procfs dump previously it returned just the raw data and the mappings from pid to cmdline i went ahead and modified it so it returned the rest of the text dump as well which contains our proc dump and any future dumps that get added later any future text dumps will have to implement two cases either extraxt cmdline extracts the raw data or it doesn t then it gets passed to extract tgid which will either extract the raw data or it wont as you can see this is going to get cumbersome i will fix this now and have the raw data be extracted before any sort of text extraction happens not sure why this was not done in the first case

[ 0 ]

You need emag to make blood bro, so cost is upped, other way you get stupid clones that basically do nothing but impersonating some random dude they are clones of.

[ 1 ]

Still not quite right. It's now very easy to forget the `&` and pass random crap to `free`. Using `omf_free` should look _exactly_ like `free` to avoid this.

[ 1 ]

Jesus Christ. Monthly remember that we need outfit datums to fix this kind of shit.

[ 1 ]

move browser view into this block or inline getbrowserviewforbrowser call

[ 0 ]

Just to avoid multiple calls on the same frame, yeah. Can be removed if you think it's ugly.

[ 1 ]

maybe can we go for non admin examples or is that silly

[ 0 ]

And rest of request is nonsense mostly

[ 1 ]

need to add the secondary index vars otherwise puppet will complain so should add secondary index false secondary index type lucene

[ 0 ]

<*shakes fist*> darn you markdown!

[ 1 ]

useless empty line at end of file removal json dump does not create empty line at end of file json specification simply does not care

[ 0 ]

this is copy pasted verbatim from the neutron patch but it is silly just do for event in self pending events

[ 0 ]

please move it before putils execute invoke

[ 0 ]

It's been almost three years since opencv got java bindings. How the hell is the maven support still missing? This is ridiculous.

[ 1 ]

ideally it should be le agent cache inconsistency port id is not stored for s port mac

[ 0 ]

why admin creds are used for share creation

[ 0 ]

>Razor Component component model Please no! That just looks ridiculous \U0001f61d.

[ 1 ]

This is going to introduce a delay to using the status bar, god no. This is also going to be absurdly expensive, every time your mouse fucking MOVES it calls an extra proc for no reason.

[ 1 ]

Good point, I'll add some handling shit for that to end the fulton attempt there if it gets anchored before the fulton starts

[ 1 ]

this should be just using trans instead of blocktrans

[ 0 ]

seriously what the fuck are you doing

[ 1 ]