black-formatter

[bright glacier]

Gee thanks! No worries, better done than never 🙂

[flat kraken]

JelleZijlstra

[psf/black] Issue opened: #2318 Switch to a new parser

Currently, Black uses a vendored version of lib2to3 for parsing. This works well for parsing Python 2 and early Python 3, but Python has now moved on to a PEG-based parser (PEP 617), and lib2to3 is no longer being maintained.

So we need a new parser. There are a few existing options that we could leverage (Parso, LibCST), but it's going to be a lot of work to do the migration. WE're d...

[dense jungle]

there you go

[flat kraken]

ichard26

[psf/black] Pull request review submitted: #2281 Possible fix for issue with indentation and fmt: skip

I'm in roughly the same boat as Jelle, but I did spent a few hours somewhat recently in this specific part of the codebase trying to fix a bug (I wasn't successful) and this looks good to me as well.

Thank you for the PR! It's awesome to see one bug fixed.

JelleZijlstra

[psf/black] Pull request closed: #2281 Possible fix for issue with indentation and fmt: skip

JelleZijlstra

[black:main] 1 new commit

40fae18 Possible fix for issue with indentation and fmt... - enzet

JelleZijlstra

[psf/black] Issue closed: #2254 fmt:skip fails with internal error

ichard26

[psf/black] Pull request review submitted: #2306 Mention comment non\-processing in documentation

Looking good!

I was going to suggest adding how Black tries extra hard for certain pragma comments to avoiding them, but I ended up deciding against it because it's not a guarantee so not exactly useful to readers :P (also it's mentioned in the linked AST safety section anyway)

Thanks a million!

[bright glacier]

@dense jungle do you want to do a quick lookover for the above docs PR or are happy with it (and I'll merge if so)?

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2306 Mention comment non\-processing in documentation

[bright glacier]

whelp that's my answer lol.

[dense jungle]

I think it's good to go

[plain atlas]

hm, no checks passed/failed webhook sending?

[bright glacier]

I think the service that's handling the communication between GH and this channel is filtering that sort of messages out

[plain atlas]

ah, it isn't being sent directly to a discord webhook then?

[bright glacier]

That would be way too noisy since there's pretty much no filtering configuration on their side

[plain atlas]

uh lemme view the source

workers#1

[toxic storm]

GitHub

Add short URL worker code

[bright glacier]

You get to choose "no webhook", "push events", or "or let me know of everything" IIRC

[delicate marsh]

yeah, that's not handled in the worker source @plain atlas, that's just disabled in the repo 🙂

[bright glacier]

Well it turns out I was wrong on the last option, but overall, you don't get much choice 🙂

[plain atlas]

ah lol

[plain atlas]

yeah, i agree.

[plain atlas]

could you tell me how the workers work in #ot2-the-original-pubsta?

[bright glacier]

Completely OT but I'm developing the ability to just know where to add line breaks in the little commit message box GitHub gives you when merging. I managed to nail the 72 character rule of thumb without checking somehow.

[flat kraken]

ichard26

[psf/black] Issue closed: #1017 Black's behavior with comments is not described in the style guide

ichard26

[psf/black] Pull request closed: #2306 Mention comment non\-processing in documentation

ichard26

[black:main] 1 new commit

a9eab85 Mention comment non-processing in documentation... - felix-hilden

[bright glacier]

OK, so before we release, here's my list of PRs we might consider landing first:

1. https://github.com/psf/black/pull/2301 (Switch toml TOML library for tomli)
2. https://github.com/psf/black/pull/2285 (Regression fix: leave R prefixes capitalization alone)
3. https://github.com/psf/black/pull/2297 (Normalise string prefix order)

I think 2 is safe to land (although it needs a review), 1 might be okay although I have no real experience in this area, and 3 seems like it's awaiting discussion on our support policy for Python 2.

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2285 Regression fix: leave R prefixes capitalization alone

JelleZijlstra

[psf/black] Issue closed: #1244 Preserve r\-string prefix capitalization

JelleZijlstra

[psf/black] Pull request closed: #2285 Regression fix: leave R prefixes capitalization alone

[bright glacier]

Although it definitely could be possible 3 is OK and I just managed to miss all of the relevant discussion. IIRC Cooper said they wanted to get it in.

[flat kraken]

JelleZijlstra

[black:main] 1 new commit

00e7e12 Regression fix: leave R prefixes capitalization... - ichard26

[bright glacier]

Thank you for the review @dense jungle !

[errant barn]

It is ok, but myself and Jelle would prefer r first, which imo could be reason to wait until Py2 is removed to avoid special cases

[flat kraken]

jack1142

[psf/black] Pull request opened: #2319 Support named escapes \(\`\\N\{\.\.\.\}\`\) in string processing

Fixes #1468 (or at least the issue that's in the comment there)

Not an awfully big amount of tests added here but I'm not sure if there's a need for more.

[lament crow]

Sooo after trying to contribute I've noticed some small issues with local testing:

• pre-commit run --all-files fails because src/black/output.py has one line that exceeds the line length
  • perhaps this command should be included in CI to stop a PR that would break this requirement? This was it wouldn't affect a contributor who hasn't actually done anything wrong.
• tox -e py is failing on the optional no_python2 test when I run it locally. I followed the contributing guide (the pipenv way) from Black's docs (https://black.readthedocs.io/en/stable/contributing/the_basics.html) before I was running any testing so the way I installed shouldn't be a problem? perhaps something causes the typed-ast dependency to be installed? Testing on Windows

I ignored both of these since I knew neither of these problems were caused by me but I just wanted to share some feedback as first-time contributor

[errant barn]

Yeah, Windows has had some issues 😅 If you'd be so kind, Richard has created a dedicated issue for the contributing experience: psf/black#2238 Would be great to add this in! (and don't worry, I'll add it later if you haven't yet)

[toxic storm]

GitHub

Improving the contributing experience

[lament crow]

I'll add later if I don't forget lol

Or I could, just copy-paste now I guess

Done, thanks for linking the right place to put it on

[flat kraken]

felix-hilden

[psf/black] Pull request review submitted: #2319 Support named escapes \(\`\\N\{\.\.\.\}\`\) in string processing

Hi, thanks for submitting! I'm not familiar with this part of the code base so take it with a grain of salt, but I left some thoughts below. I'd like for the more experienced maintainers to take a look too.

[tribal thistle]

Hmm which line is exceeding the limit? I just ran pre-commit on master and it passed

It should also be passing in master, there's already a CI action for it
https://github.com/psf/black/blob/main/.github/workflows/lint.yml

[lament crow]

λ pre-commit run --all-files
black....................................................................Passed
flake8...................................................................Failed
- hook id: flake8
- exit code: 1

src/black/output.py:48:92: B950 line too long (91 > 80 characters)

mypy.....................................................................Passed
prettier.................................................................Passed

[tribal thistle]

Could you try running pipenv run flake8 in the root directory?

[errant barn]

Ah is it another case of psf/black#2299

[toxic storm]

GitHub

Pre-commit Can't Find Black

[tribal thistle]

Sorta, my best guess is it isn't using the correct flake config

We've run into this a lot in #dev-contrib because our setup has a ton of plugins, and people would usually be running their own flakes without them

Hmm not sure that's it anymore. I couldn't repro it

I tried with 39

[errant barn]

It does say that the line length limit is 80 though

So something's fishy. Any updates @lament crow?

But now that I think about it, I don't think I've had problems running other than Black with precommit even though I have basically nothing installed on the system interpreter

[lament crow]

ah sorry

forgot

[errant barn]

And the linked issue was about not finding black..

[lament crow]

running pipenv run flake8 in root is wrong because that runs it on tests data as well

so I've got a big wall of errors :P

a wall that is apparently so big that I'm past console's history buffer

but anyway

line length limit is 88, right?

so 91 is still past 88

[tribal thistle]

88 + 10%

[lament crow]

80+10%

or 79

[tribal thistle]

I should better explain

[lament crow]

it's a bugbear thing

[tribal thistle]

the specified length is 80, bugbear adds 10% on it's own

ya

[errant barn]

Oh so it doesn't show in the error, the increased line length?

[lament crow]

well, 80+10% is 88 so however we look at it, 91 is higher

[errant barn]

yea

[lament crow]

        for lineno, line in enumerate(self.lines, start=1):
            length = len(line) - 1
            if length > 1.1 * self.max_line_length:
                yield B950(lineno, length, vars=(length, self.max_line_length))

[errant barn]

oh, it's this comment line: https://github.com/psf/black/blob/main/src/black/output.py#L48

[late dew]

src/black/output.py line 48

# See https://www.gnu.org/software/diffutils/manual/html_node/Incomplete-Lines.html```

[errant barn]

It's within line length for me though nvm, it's not

[tribal thistle]

Funny enough, I can't get any length to trigger it

What other flake8 rules are there to break

[lament crow]

but ye, this one is not going to work correctly because .flake8 does not have the excludes set

only pre-commit config does

[errant barn]

(does discord have strikethrough MD?)

[tribal thistle]

~~something~~

[lament crow]

not that this is the issue here

[errant barn]

yeah the line is longer than it should

[tribal thistle]

So flake definitely works, just not for this rule

Which is

interesting

[errant barn]

I'm proper confused now :D after removing .flake8 I can lint the file and it does complain

[lament crow]

do you run flake8 with cwd of the project root?

[errant barn]

Ah, E501 is the very line len error that is ignored in the config

[lament crow]

oh ye no

I'm getting the B950 one

[errant barn]

Yep, for some reason it doesn't trigger

[tribal thistle]

Actually, all B9 errors aren't triggering for me

I wonder if it's a bugbear bug

[errant barn]

the difference in this was the exclusion

[lament crow]

exclusion?

you removed it from .pre-commit-config.yaml?

[errant barn]

Exclusion of E501 in .flake8

[lament crow]

huh

[errant barn]

I think it's time to MWE this up :D

[tribal thistle]

Hmm

do you have a clean environment to test bugbear?

If so, do you get any errors on this:

class Test:
    def abc(xyz):
        pass

This should in theory be failing, but I can't get it to fail

[errant barn]

Yeah that fails

[tribal thistle]

Does it fail in black

[errant barn]

No!

Also I couldn't repro the line length issue alone

Even when straight up copying the file

[tribal thistle]

Seems bugbear isn't working at all

[bright glacier]

Completely lost at what is going on really but does flake8 even recognize bugbear?

Does flake8 --version include bugbear?

[errant barn]

yup

[tribal thistle]

3.9.2 (flake8-bugbear: 21.4.3, mccabe: 0.6.1, pycodestyle: 2.7.0, pyflakes: 2.3.1) CPython 3.9.4 on Windows

[bright glacier]

OK that's the end of me being useful since I'm on my phone 😅

[errant barn]

The interesting thing would be why did it work for Jackenmen then

[lament crow]

ye, I do find it interesting it only works for me lol

actually, I'm gonna check one thing in a sec

[errant barn]

Normally it's the bug that can't be reproduced, not the other way around :D

[lament crow]

ye, I found the reason it works differently for me

I do have a "global" flake8 config

since if a project doesn't have one, I like to have sensible defaults

[flake8]
max-line-length = 88
select = B,C,E,F,W
ignore = E203, W503

[errant barn]

Could you share that?

Perfect

[lament crow]

but project-specific one should override it

well

extend it

[errant barn]

Yeah..

[lament crow]

extend-ignore seems to be what's causing the issues here

like, if you replace it with ignore then it works fine (as in, fails) in CI

[errant barn]

But it doesn't ignore any bugbear errors 🤔

[lament crow]

Bugbear's opinionated warnings require explicit selection.

so perhaps it doesn't work well with extending smh

perhaps bugbear in some way adds into the default ignores

[errant barn]

would you try removing your select?

and maybe set the line length to 80

[lament crow]

I did, bugbear still fails for me

only thing that matters is the ignore

if I remove that from mine, then it works

[errant barn]

Black's ignore you mean?

[lament crow]

mine

[errant barn]

right

[lament crow]

if you replace Black's extend-ignore with ignore then it also works so it's really something about the extending working in unexpected way

[errant barn]

So for some reason the extend-ignore is handled in some odd ways

Yeah

YAY a reproduction!

Paste your config in a fresh environment, then call flake8 file.py --extend-ignore=E501

And actually, having an empty configuration named .flake8 and calling like above produces that same situation, whereas without it bugbear errors are still selected

[lament crow]

I'm confused, did you reproduce it yourself or do you want me to do something lol

[errant barn]

Which I'm not sure should happen either, since B9XX is opt-in if I remember correctly :D

Yeah I did

[lament crow]

btw, this is in bugbear's code

    @staticmethod
    def add_options(optmanager):
        """Informs flake8 to ignore B9xx by default."""
        optmanager.extend_default_ignore(disabled_by_default)

...

disabled_by_default = ["B901", "B902", "B903", "B950"]

[errant barn]

Yeah this is just getting more complicated by the minute :D

[lament crow]

I don't know why this default ignore doesn't actually do anything when I don't have a "global" configuration file with ignore in it but well

Do you plan on reporting that on flake's repo or should I look into it

maybe at least we will get a clearer answer lol

[errant barn]

I think we should really nail down a list of different situations and the behaviors

Because this is confusing as hell

I'll reset my venv and start from scratch

[lament crow]

1. Install flake8 and flake8-bugbear
2. Put this into a ~/.flake8 (or %userprofile\.flake8):

[flake8]
ignore = E203, W503

3. Create a simple project directory with .flake8 file:

[flake8]
max-line-length = 80
extend-ignore = E203, E266, E501
select = B,C,E,F,W,T4,B9

4. Make a test.py file:

# very long comments, which is certainly longer than 88 characters which is what bugbear should error on

5. Run flake8
6. See results (or rather, lack of any)

now that I wrote this, I need to actually see if I can reproduce it like that :D

oh wait, it was the other way around, right

so it's not a bug

because bugbear extend default ignores

which you guys extend further

so it's black's config fault after all, yay

[errant barn]

I have to go for some time but I'll get back to this

Or I'll have to go soon, but in the meantime I can't get flake8 alone to complain about line length

[lament crow]

flake8 alone, you mean without bugbear?

[errant barn]

yeah...

[lament crow]

are you ignoring E501

[errant barn]

nowhere, also explicitly selecting every error

[lament crow]

oh right

pycodestyle is disabled by default

from bugbear's readme:

Note that pycodestyle also has a bunch of warnings that are disabled by default. Those get enabled as soon as there is an ignore = line in your configuration.

:D

[errant barn]

Yeah, flake8 file.py --select=C,E,F,W does nothing

[errant barn]

what.

[lament crow]

what what

[errant barn]

flake8 file.py --select=C,E,F,W --ignore=E001 does work

[lament crow]

I should have quoted more:

I think this behavior is surprising so Bugbear's opinionated warnings require explicit selection.

YE, I think it's surprising too!

[errant barn]

Yeah, but I'm still on the flake8 only errors :D

I'd expect selecting the damn errors is enough

[lament crow]

ignore takes priority over select

[errant barn]

Maybe I'm just confused, if I'm ignoring nothing, shouldn't and explicit --select=E select all E-class errors :D

[lament crow]

are you ignoring nothing though?

because not saying you ignore something equals ignoring defaults

confusing, I know :D

[errant barn]

that's sooo dumb xD

--ignore= works too

[lament crow]

I'm not going to say it's dumb because I'm sure there must have been some reason to do it that way

but I sure can say it's confusing the hell out of me lol

[errant barn]

Yeah most probably

But kay well we'll move on from this

[lament crow]

so basically, you guys need to use regular ignore or else you run into surprising behaviors of flake8

[errant barn]

OOOH everything clicked!

I see now :D you're the best dude

[lament crow]

I'll look into making a PR considering I already spent time on this anyway lol

[errant barn]

Yeah that would be much appreciated

[bright glacier]

and I'm still thoroughly confused, looks like I will have to read over this chat a few times over at least

[errant barn]

Having the extends is convenient, but I guess it's better to enforce a list, people can then extend on the command line

[bright glacier]

anyway, great work y'all

[flat kraken]

jack1142

[psf/black] Pull request opened: #2320 Fix flake8 configuration by switching from extend\-ignore to ignore

Basically, bugbear's B9 checks (and possibly other checks that you possibly want to run) weren't actually running due to the use of extend-ignore.

[flat kraken]

pquentin

[psf/black] Issue opened: #2321 INTERNAL ERROR: Black produced different code on the second pass of the formatter after "pre\-existing trailing commas when not in the rightmost br...

Describe the bug

INTERNAL ERROR: Black produced different code on the second pass of the formatter.

The first bad commit is https://github.com/psf/black/commit/586d24236e6b57bc3b5da85fdbe2563835021076 according to git bisect.

To Reproduce

Take this file:

def test(self, othr):
    return (1 == 2 and
            (name, description, self.default, self.selected, self.auto_generated, self.parameters, self.meta_data, self.schedule) ==
            (name, description,...

[flat kraken]

felix-hilden

[psf/black] Pull request review submitted: #2320 Fix flake8 configuration by switching from extend\-ignore to ignore

Yeah not much to say, checks out!

[lament crow]

👀 I just noticed that your changelog entry check doesn't duplicate despite having label as an event

did GitHub finally fix it :o

[flat kraken]

felix-hilden

[psf/black] Pull request review submitted: #2319 Support named escapes \(\`\\N\{\.\.\.\}\`\) in string processing

In my opinion this looks much better now, thanks a ton for being so speedy to modify!

[tribal thistle]

I wonder why it was using extend in the first place

It was slipped in to this commit, but it isn't mentioned
https://github.com/psf/black/commit/573b8de54470dbad8bcaaebd5a28dad507c44666

[errant barn]

For CLI usage I would imagine

[tribal thistle]

Would that be something not worth keeping then?

[errant barn]

I mean, ignore in config and extend in CLI is functionally similar to extend in config and ignore in CLI

[tribal thistle]

What's been done seems like the best solution currently possible, but it seems the most ideal solution would be getting flake8(-bugbear) to support extend

Alright, sounds good

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2320 Fix flake8 configuration by switching from extend\-ignore to ignore

HassanAbouelela

[psf/black] Pull request review submitted: #2320 Fix flake8 configuration by switching from extend\-ignore to ignore

Everything looks good. Crazy this has gone unnoticed for so long.

JelleZijlstra

[psf/black] Pull request closed: #2320 Fix flake8 configuration by switching from extend\-ignore to ignore

JelleZijlstra

[black:main] 1 new commit

229498e Fix flake8 configuration by switching from exte... - jack1142

[tribal thistle]

On a slightly different note, does anyone review PSF projects locally? I wasn't able to submit my review due to oauth apps restrictions

[lament crow]

gh cli?

[dense jungle]

You mean using the Github app or something? I only use the website

[tribal thistle]

pycharm, but I imagine the cli would suffer from the same thing

Error occurred while submitting a review
Although you appear to have the correct authorization credentials, the psf organization has enabled OAuth App access restrictions, meaning that data access to third-parties is limited. For more information on these restrictions, including how to enable this app, visit https://docs.github.com/articles/restricting-access-to-your-organization-s-data

[lament crow]

you could use a PAT instead of oauth authentication as a workaround

[tribal thistle]

That would probably work, but pycharm signs out on restart, which would mean generating a new token every time I start it up

[lament crow]

I would tell you to just make an oauth app request, I would assume nobody's going to come to your home and murder you for it, maybe they just won't accept :)

[tribal thistle]

Hahaha

[lament crow]

but then again, I'm not part of PSF

or a PSF project

so maybe they do murder

[tribal thistle]

Anyways, it's just clicking the submit button that needs to be done on site, so it's probably not a big deal

Oh, I take that back

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2319 Support named escapes \(\`\\N\{\.\.\.\}\`\) in string processing

sharpblade4

[psf/black] Issue opened: #2322 Black fails due to non\-equivalent code generation

Describe the bug

Format failure due to non-equivalent code to the original, probably because of replacing tab with spaces inside a doc-string.

To Reproduce

1. create a file something.py
2. paste the content:

   def foo() -> None:
       """ Some documentatios:
                   'serialization.format' = '	',
       """
       pass
   
   

3. run black something.py
4. observe error like:

   error: cannot format something.p...

sharpblade4

[psf/black] Issue closed: #2322 Black fails due to non\-equivalent code generation

[lament crow]

looks like they did fix it indeed, but I ran into another problem which makes it unsuitable for my use case :(

[neon loom]

Wow - much activity

[bright glacier]

What are we waiting for @neon loom in terms of new release?

Or are you hoping I know so I can let you know so we do a release?... cause in that case, I don't exactly know. psf/black#2319 seems OK but I have no real skills in this area of the codebsae.

[toxic storm]

GitHub

Support named escapes (\N{...}) in string processing

[errant barn]

Yeah it seems to be okay, my last comment can be ignored for now

[neon loom]

I’ve just been waiting for a slow down in action to draw the lone in the same personally

[bright glacier]

lol

[neon loom]

*line in the sand

[bright glacier]

blame the apple

it's not "an apple keeps the doctor away", it's actually "an apple causes a breakout of typos"

[errant barn]

Named escapes could be good since it's related to the string processing

[bright glacier]

(for context, cooper has an iphone)

[flat kraken]

JelleZijlstra

[psf/black] Issue closed: #1468 Black produces code with invalid syntax by splitting \`\\n\` in half

JelleZijlstra

[psf/black] Pull request closed: #2319 Support named escapes \(\`\\N\{\.\.\.\}\`\) in string processing

JelleZijlstra

[black:main] 1 new commit

62402a3 Support named escapes (\N{...}) in string pro... - jack1142

[bright glacier]

Oh right, I think that's Jack's first contribution so it's time to bust out my saved reply... Well sorta, I mean, they saw the "improving the contributing experience" issue already.

I'll just say thank you as always 😄

[errant barn]

Also if you're still looking for easy wins, the AST errors just need a decisive mypy guru ;) but the impact is minimal though psf/black#2304

[toxic storm]

GitHub

Improve AST safety parsing error message

[bright glacier]

TBH, I've been hesitant since I may end rewriting the whole the module during my mypyc work.

[errant barn]

Oh!

[bright glacier]

But it's not 100% likely so I'll merge I guess.

Honestly I'm a fan of asserts for this sort of things.

[errant barn]

Yeah consider it done, gimme a minute

[bright glacier]

I was just gonna check it out and do the changes myself and push but that also works 🙂

My GitHub Notifications are always a mess, I can never keep my inbox empty 😑

[tribal thistle]

Someone once told me “select all -> mark read”, and get a fresh start

Ofc I didn’t but oh well

[bright glacier]

"View all 164 notifications" - oh how fun

[tribal thistle]

How do you even have that many lol

Are you subscribed to a bunch of things?

[bright glacier]

I watch two repositories and have a bad habit of not clicking "done".

[errant barn]

Oh actually, the assert doesn't work, so should we go with the test instead?

[bright glacier]

It's psf/black and pypa/bandersnatch by the way.

[bright glacier]

Wait what

[errant barn]

still produces the error

but moving to the if doesn't

[bright glacier]

............

[errant barn]

Yeah fun

So maybe we'll go with that since you're already working on it and type: ignore seems kinda bad

pushed

[lament crow]

happy to help, was honestly more fun than I thought it will be

[lament crow]

and you're no match to me 😬

I get email notifications and I kinda don't have time to mark those as done on GitHub

[bright glacier]

Ooo I wonder if typeshed's overloads for ast.parse don't have version based conditionals

[errant barn]

But how would that change assert vs if? 🤔

[bright glacier]

actually let's just move on, I don't know what I am saying 😅

yeah typeshed does have the version based conditionals anyway

https://github.com/python/typeshed/blob/d143a5d04b0bd279d19d0c85908f58a497784fbe/stdlib/ast.pyi#L158-L176

wonder why the assert doesn't work then

[errant barn]

Yeah seems weird since it's equivalent in the code flow

[neon loom]

Cooper has apple everything basically. Not home pods or HomeKit crap tho.

So we trying to get that 1 more PR in for the push? I think I'm gonna merge my own primer STDIN as it hurts nothing, is fast and might catch something one day ...

[plain atlas]

Are there two coopers?

or was that third person?

[bright glacier]

very much third person I believe

[plain atlas]

Ah, aru is sorry he didn't understand. 😄

[errant barn]

Yup checks passed! Feel free to mrrg

[bright glacier]

Richard hates your usage of third person /s

[plain atlas]

Felix just passed checks

[bright glacier]

That sounds like Felix was sus for something out of context 😅

[errant barn]

All the checks: health, IQ, not-a-pedo, you name it

[bright glacier]

Uhh, strange reference but OK

[errant barn]

Am I going too far again :D

Maybe it's time to dial it down again

[plain atlas]

Purple name, check
Too far, check

[errant barn]

:E

Yeah I might consider acting like someone responsible

[plain atlas]

Contributor of black, check

[plain atlas]

Aha same

[bright glacier]

It's like saying "I'm no criminal...", sounds suspicious doesn't it :p

[errant barn]

:D trust me, I'm trustworthy

[neon loom]

What is acting responsible?

[plain atlas]

Lol

who

[plain atlas]

My code definitely follows pep 8

And follows black too

Definitely 100%

I'd never write code in a different style ever

[errant barn]

Aaanyways, the thing before all this stuff: merge and then release?

[plain atlas]

Oops sorry for derailing the channel lol

[bright glacier]

I certainly without a shred of doubt am the most qualified person to lead the psf/black project. My experience goes far beyond the skyline into outer-space. My knowledge runs deep, as deep as the ocean. I have the will of 1000 developers... OK that's enough of me being ridiculous lol (so much /s)

[neon loom]

Yeah, I'm defering to @errant barn + @bright glacier to declare when release is wanted and I'll do the few clicks

[neon loom]

Yeah, I hear @solid adder has got nothing worth giving ...

[plain atlas]

Sounds like the beginning of a copy pasta lmao

[neon loom]

returns to $dayJob painful work

[plain atlas]

So black will be releasing as stable eventually?

[errant barn]

Naah, just the next beta release

Eventually? Yeah

[plain atlas]

Ah, whew

[neon loom]

That's not the discussion today. Today is cutting another beta release with all our fixes.

[plain atlas]

Whew :+1:

[neon loom]

Then get experimental string formatting on by default. Then we can discuss the stable

[plain atlas]

Good, good

[bright glacier]

Hmm, the stable tag only includes changes before and on May 10th even though the newest release is from May 31st

[neon loom]

O yeah, I never moved that.

I dislike that tag. We need to automate it somehow. lol

[flat kraken]

jack1142

[psf/black] Pull request opened: #2323 Add coverage files to gitignore

I've been getting a bunch of these when running tox -e py

[bright glacier]

The only thing I'm not sure about with the AST PR is this:

Also, target versions include Pythons all the way from 3.3. Should they be here (and in the version list below) too?

[errant barn]

Yeah that's fair

I initially removed them because they overlapped with the typed_ast versions 6 and 7

[solid adder]

Damn it. Before I read it until the end I was ready to give you the nuclear codes and run away

Open a bakery in Provence

Or buy a fishing boat in the Maldives

OK that's enough silliness from me, I get seasick real easy

[plain atlas]

Ah yes. I knew changing notifications to all messages for this channel would be a bad decision. Too much fun stuff happens here, and it's distracting in a good way 🙂

[bright glacier]

You can never have too much fun

[flat kraken]

felix-hilden

[psf/black] Pull request review submitted: #2323 Add coverage files to gitignore

Now that I run that exact command it happens to me as well, so seems like a no-brainer. Thanks!

[bright glacier]

I am not the biggest fan of pastries tbh

[dense jungle]

that's almost as bad as using tabs for formatting

[plain atlas]

!ot unless that's somehow related to black

[bright glacier]

It's a reference to ambv's (maybe dream) of running and starting a bakery in southern france

[flat kraken]

jack1142

[psf/black] Pull request opened: #2324 Don't run Docker workflow on push to forks

Docker workflow is used for publishing images to DockerHub which I'm pretty sure people that just try to contribute wouldn't want and the workflow fails because the forks don't have the credentials set up. While it is set to only run on main some people like to keep their fork up-to-date and GitHub even acknowledged that recently making it extremely easy to update from the web.

This is also a...

[errant barn]

@solid adder Still here? Could you chime in a bit to removing versions from AST parsing in psf/black#2304 I think you could be the person that knows the most about this.

[toxic storm]

GitHub

Improve AST safety parsing error message

[lament crow]

I call myself "the bug magnet"

:P

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2324 Don't run Docker workflow on push to forks

JelleZijlstra

[psf/black] Pull request review submitted: #2323 Add coverage files to gitignore

[solid adder]

Give me 15 minutes

I'm fetching a package

[flat kraken]

ichard26

[psf/black] Pull request review submitted: #2324 Don't run Docker workflow on push to forks

Whelp I'm copying and pasting for bandersnatch which quite literally is facing the exact same issue. Thanks!

[bright glacier]

actually I was wrong in that PR review

the bandersnatch issue is similar but not the same 😞

sometimes checking before doing something is better ay

[plain atlas]

uh--
should the docker workflow only run on pushes to main?

[lament crow]

yes

you don't want the branches created on psf repo to be pushed to dockerhub

[plain atlas]

ah, it already runs on main only

good

[lament crow]

also iirc, to keep the ambv/black->psf/black redirect, Łukasz needs to make branches on psf/black repo and can't use the fork I suppose

unless GitHub changed something

a bit annoying that you can't choose a fork name if you ask me :P

[bright glacier]

I'm pretty sure that's current behaviour ATM still

[plain atlas]

yeah it sucks

if ambv were to ever press the fork button it will break the redirect.

because it will just fork without asking if you really meant to fork ;-;

[lament crow]

unless you have repo creation access on a different org

[bright glacier]

Well, sorta

[lament crow]

in that case you have a popup

[plain atlas]

true

[bright glacier]

yep!

[lament crow]

where you have to choose where you want to fork

[plain atlas]

unless its already forked there

then i'm not sure

[lament crow]

it still asks

[bright glacier]

I considered adding a conditional to block the docker workflow from running if the test workflow fails, but I suck at conditionals so yeah no. Also, the chances of people pushing directly to main with breaking changes is low, right?

[lament crow]

do you guys use protected branches?

[errant barn]

Yeah seems reasonable

[bright glacier]

Nope IIRC

[lament crow]

ah well

[plain atlas]

😓

[lament crow]

could make sense to make a dep then if someone does push stuff directly to main

[errant barn]

Using them could be a good thing though, since we have a reasonable number of people available for review

[plain atlas]

not using protected branches is a bad idea imo

means that anyone with push access can just push to main

[bright glacier]

I agree in general. This is something Łukasz would have to configure.

[bright glacier]

Honestly our current set up means that I can make a release with only write perms

[errant barn]

That's convenient to be fair

[bright glacier]

protected branches would obviously help but not 100% block danger

[lament crow]

you can make github releases with just write perms anyway

[plain atlas]

e.e

that's freaky

[bright glacier]

Well I consider PyPI releases to be a lot more meaningful than GH releases 😛

I think it's worth bringing up with the rest of the core team later

[plain atlas]

this is a portion of the required status checks on my repo

[lament crow]

don't your gh release workflow push to pypi?

[bright glacier]

Yes

[plain atlas]

🤦

[lament crow]

then you can release to pypi as well, unless you use the environments
https://docs.github.com/en/actions/reference/environments

[plain atlas]

i legit just looked at my own repo's workflows to see what black had

[flat kraken]

JelleZijlstra

[psf/black] Pull request closed: #2324 Don't run Docker workflow on push to forks

JelleZijlstra

[black:main] 1 new commit

2c5150c Don't run Docker workflow on push to forks (#2324) - jack1142

JelleZijlstra

[psf/black] Pull request closed: #2323 Add coverage files to gitignore

JelleZijlstra

[black:main] 1 new commit

77021f0 Add coverage files to gitignore (#2323) - jack1142

[lament crow]

that would then actually require a review before push

[bright glacier]

that's real nice!

[plain atlas]

yo what

where is the guide for this

[lament crow]

^

[bright glacier]

I gave up on environments when it didn't work with github pages

[lament crow]

it's in beta rn

[bright glacier]

but I think that's just something specific with pages

[plain atlas]

to note, i'm pretty sure you don't need that to set up automatic pushes to the other stuff and yada yada

[lament crow]

environments are additional layer of security

[bright glacier]

@errant barn @dense jungle while I have your attention, would it be worth implementing (or discussing) stronger requirements / protections for pushes (direct or merges) to main?

[lament crow]

it depends on the project but imo it's not always that great, especially if you want to release a hotfix

and nobody's around

[dense jungle]

Don't feel a strong need for changes. @solid adder actually likes to push straight to main

as far as I can tell

[lament crow]

but generally, it's quite cool, just not for everyone's tastes

[dense jungle]

I'd like to be able to merge my own PRs without review if I feel it's urgent

[bright glacier]

I think protecting the PyPI upload might be a good call but in general I don't care about the rest

[dense jungle]

(if you disagree later we'd be able to revert)

[plain atlas]

this sounds worrysome

[lament crow]

repo admins can overcome the branch protection

[plain atlas]

^

[lament crow]

I don't know if you are one though

[dense jungle]

none of us are repo admins I believe

maybe Lukasz is

[bright glacier]

well he is one but even his powers are limited due to the repo being under the psf org

[lament crow]

they're not limited by much, he can't block people from the repo since that can only be org-wide but pretty much anything concerning the repo can be done by repo's admins

[errant barn]

Yeah I think it's a good idea, but I'm not against the current flow since it seems to work. The Encode org (responsible for HTTPX and Starlette and such) has used it to a great effect, since they give write access for essentially all contributors. But here I don't think it's essential

[lament crow]

oh and he can't remove the repo :3

[plain atlas]

🤔 You're part of encode?

[bright glacier]

That surprises me, I'd argue repo specific bans would be useful in enough situations

[errant barn]

Well I've put in like one commit but technically yeah :D

[bright glacier]

lol

[lament crow]

Yeah, it's a bit meh, but sadly GitHub only has global blocking

[bright glacier]

in that case, I would consider myself a part of pypa/pip

[plain atlas]

aa, i find httpx pretty nice but one thing is very wonky to me lol

[bright glacier]

cause I made one docs commit

[lament crow]

per-org/user

[plain atlas]

i'm also surprised that starlette still uses requests and not httpx for the test client

[dense jungle]

I contributed one space to PHP, so I'm part of PHP

[errant barn]

But my point was that the branch protection works for them since they have a massive amount of people with write access, but for us not as much

[lament crow]

well, there are different kinds of protections

you don't have to require reviewers

[plain atlas]

^

[lament crow]

you could just require passing status checks

[plain atlas]

the minimum that branch protection does is just prevent deleting the branch and pushing directly to main

[bright glacier]

gh blocks you from deleting the default branch right lol

[lament crow]

the minimum is preventing force-push

[bright glacier]

I'd hope so

[plain atlas]

but repo admins override that anyways

[plain atlas]

nah you can re-enable that

[solid adder]

Yeah I sometimes push to main but not when code changes are involved. But changing docs or tests esp. for some typo or inclusion of a new sentence (or testimonial by zzzeek 😂) I'd just do in main for convenience.

But if we find that we should protect main then I'm fine with that.

[lament crow]

well, then minimum is nothing lol

[plain atlas]

no, minimum prevents pushing directly to main

that's all

[lament crow]

it doesn't

I have branch protections that just prevent me from force-pushing

[bright glacier]

I mean, admins powers are supposed to be not given out like candy lol

[plain atlas]

you're a repo admin i assume

[errant barn]

Yeah not really dangerous admin powers though 😅

[lament crow]

yes, but I include administrators in the protection

[solid adder]

Fun fact, in cpython I have to push the merge to the release branch for the signed tags to remain immutable. So I would untick "include administrators" every release I'm making for a single push.

But our release workflow in Black is much less involved so this problem doesn't really exist here

[plain atlas]

ah, i see it now

[bright glacier]

I have never used code owners

[lament crow]

yeah, it's quite granular

[plain atlas]

turning that off and this off allows pushes to main

[lament crow]

I use code owners but not the requirements of the review from code owner

[bright glacier]

Also, y'all use dark mode for GH? I'm still using the light theme

[lament crow]

I use the extension for that

I don't like the one that GH has

[plain atlas]

i use the middle mode

[bright glacier]

that's the lower contrast one right?

[lament crow]

I still have it set to dark in case I decide to open it in mobile browser

[bright glacier]

and because I can't stop myself from being even more hated, I also use light mode for discord 😛

[lament crow]

I'm wondering how the new high contrast ones will look

https://github.com/github/roadmap/issues/202

[plain atlas]

/s ofc

[bright glacier]

now that's a threat 😉

[lament crow]

just don't blind us with screenshots of your discord and we might still become friends one day

:)

[lament crow]

wow

rude /s

[plain atlas]

believe it or not, I use VSC with the github dark theme

[bright glacier]

"why you gotta be so damn rude" or how ever that song goes just popped into my head lol

[lament crow]

I don't use VSC 😬

I wanted to

repeatedly

[plain atlas]

¯_(ツ)_/¯

all of my stuff is ssh so i use vsc

[errant barn]

Dontcha know I'm hooman too

[lament crow]

but there was always some small thing that annoyed me when I tried to switch and I went back to Sublime

[plain atlas]

this channel is no longer about black, this is the community within the community

[bright glacier]

it's the monkeys within the community

[errant barn]

Yeah we're going off road and fast :D

[lament crow]

since we're not talking about black, here's a cool thing that showed up on GH's roadmap today:
https://github.com/github/roadmap/issues/211

[plain atlas]

and a few minutes later, some other maintainer steers us back on track

[bright glacier]

Developers and users of Black are invited to use this channel for discussions on the development and usage of the tool.
Yeah screw that. 😄

[plain atlas]

Developers and users of Black are invited to use this channel for discussions.
FTFY

[lament crow]

hey, to smoothly develop the tool, we need tools like GitHub!

[bright glacier]

cooper is waiting on us (me and felix) to declare black ready for a new release and he's gonna come back and discover that we have been -100% productive

[lament crow]

or well, text editor :D

[plain atlas]

fwiw @lament crow, i find sublime better than n++

[lament crow]

aaand that's why you shouldn't do ETAs in OSS, takes the fun out of it :)

[errant barn]

I'm just waiting on Wu-cash to receive his package

[lament crow]

why does that sound like an insult to sublime lol

[plain atlas]

because i really hate one thing about sublime

this is n++, @bright glacier will love it

[bright glacier]

I actually used to use N++ when I still used Windows

[plain atlas]

this is sublime

[lament crow]

there is a light mode in sublime

[bright glacier]

I switch my VSC theme like every two weeks

[silent apex]

Saaaame

[plain atlas]

and why i hate sublime is visible in the screenshot above

you might not see it

[bright glacier]

I'm currently using "Solarized Light"

[plain atlas]

UNREGISTERED

i hate that text 😔

and I won't register it either

[lament crow]

then buy it

[plain atlas]

so we have a stalemate here

[lament crow]

the biggest change I've done is when I switched from ST3 to ST4 and there was a new color scheme

[bright glacier]

although I am a big fan of Notis themes too for VSC

[lament crow]

Mariana

[errant barn]

Kinda reminds me of my audio editor, has my mum's name in the license since she bought it at first 😄

[plain atlas]

lolno

[lament crow]

that's not that much for 3 years

[bright glacier]

"hey google, translate 99 freedom dollars into maple surup dollars"

[lament crow]

I'm actually disappointed google doesn't answer to that properly

[plain atlas]

its over 99 iirc

[plain atlas]

lol i ended up googling this https://www.dailymail.co.uk/news/article-2331770/New-Canadian-100-dollar-bills-reek-like-maple-syrup-national-bank-says-notes-fragrance-free.html

[bright glacier]

omg I found this awesome wiki article

https://en.wikipedia.org/wiki/Great_Canadian_Maple_Syrup_Heist

[lament crow]

honestly, Netflix subscription costs more than Sublime if you split Sublime's cost into months :P

[plain atlas]

anyways its $119.90

[plain atlas]

i don't have any subs 😛

[lament crow]

well, buy one coffee less per month or sth?

[bright glacier]

Coffee is bitter.

[plain atlas]

nope i buy very few things lol

[silent apex]

Yes, but why not just use an editor that's free? Unless it has some very nice features like PyCharm Pro.

[plain atlas]

i mean

[lament crow]

if you're happy with your editor then sure

[plain atlas]

pycharm and sublime are both free i think

just they remind you to pay them 😦

[bright glacier]

Nagware IIRC

[errant barn]

The OpenAI GPT-3 language model has this to say: "99 freedom dollars translated into maple surup dollars is roughly 340 million. That's 2 per second. I'm going to out there and say this
amount will keep increasing, as long as maple keeps getting more popular, and as
long as there's inflation"

[bright glacier]

Vive le Canada!

[lament crow]

pycharm community edition is just free

no nagging

pro is paid, probs with like 30 days trial

[plain atlas]

so this is my vsc theme

[bright glacier]

GPT is scary sometimes

[tribal thistle]

Pycharm pro is free if you have the right connections 🤡

[lament crow]

Sublime Text technically has this clause:

Sublime Text may be downloaded and evaluated for free, however a license must be purchased for continued use. There is currently no enforced time limit for the evaluation.

[plain atlas]

Name: GitHub Theme
Id: github.github-vscode-theme
Description: GitHub theme for VS Code
Version: 4.1.1
Publisher: GitHub
VS Marketplace Link: https://marketplace.visualstudio.com/items?itemName=GitHub.github-vscode-theme

[lament crow]

or you're maintainer of OSS project :3

[plain atlas]

ah, you know the same guy that joe knows?

[silent apex]

Or if you're a student

[bright glacier]

Tell that to the users of 7zip

[tribal thistle]

you can convert currency directly here

[lament crow]

did you mean winrar

[tribal thistle]

.wa short 100 freedom units to goose

[toxic storm]

Wolfram Alpha

Failed to get response.

[plain atlas]

LOL

[tribal thistle]

Close enough

[bright glacier]

and I am an idoit

[lament crow]

lol

[bright glacier]

can't spell lol

[bright glacier]

HEY, it's called "Canadian goose" not "goose" /s

[plain atlas]

here i ran it on a bot that returns the errors lol

[bright glacier]

it doesn't understand freedom

oh no

we're screwed

[errant barn]

turns out WolframAlpha servers contain oil - the foreign ministry

[bright glacier]

ok it's american enough then

.wa what is canada

[plain atlas]

@tribal thistle when can sirlancebot allow the command in this channel 🥺

[bright glacier]

it's not necessary honestly

[plain atlas]

personally lance and @late dew have the worst errors

[tribal thistle]

I honestly don’t even know what’s filtering it

[plain atlas]

like, one of them deletes the invoke, the other doesn't

[tribal thistle]

.wa what is canada

Wa does not know

[toxic storm]

Wolfram Alpha

View original for a bigger picture.

[plain atlas]

did you forget you have admin / helper / pydis core roles?!

[tribal thistle]

Wdym?

[plain atlas]

it doesn't allow the partners to use the cmd here but since you are staff it allows you to use them here

[tribal thistle]

I know that much, I don’t know if it’s a whitelist or a perm problem

[plain atlas]

ah

whitelist i think

[bright glacier]

"Canada is home to overilypolite maple syrup eating bearded axe wielding lumberers who say 'a-boat'"

may be true or not, I cannot confirm or deny

[tribal thistle]

Can you confirm?

Ah

[errant barn]

@bright glacier To bring it back on topic once more, I think I'll call it a night. IMO the AST error messages don't have to be in the release since the impact is so small, so feel free to ask Cooper to release.

I've had heaps of fun tonight though :D thanks everyone!

[bright glacier]

Alright, sounds good. Sleep tight and don't let the bed bugs bite 😆

[plain atlas]

yeah, imma head out too

typescript and rust await

or, await typescript and rust

[flat kraken]

ambv

[psf/black] Pull request closed: #1904 WIP: Added function for vertically aligning inline comments

[solid adder]

I agree with your find on the 3.3-3.5 parsing

Commented on the PR

[neon loom]

So much black love going around here

*on

[bright glacier]

I feel bad for Felix, his review circles are only dull greys, no pretty green or sharp reds :/

[plain atlas]

are you trying to speed run maintainer? lmao

[errant barn]

Any Vim plugin wizards here? We've been considering (re)moving our Black plugin for a while now. It's in poor condition and none of us really works on it. Would love to hear from you! psf/black#2027

[toxic storm]

GitHub

Reconsider the vim plugin provided in this repository

[tired shard]

echos godlygeek ||using vim for than 15 years iirc||

[bright glacier]

I guess at that point you would have used a ton of plugins. Personnel I use vim as my lightweight editor (VSCode for anything complex) and I have zero plugins (that might change once I finally upgrade my vim config). It's been like a year and a half of vim so far.

[indigo depot]

i just use the terminal to run black in vim ¯_(ツ)_/¯

[tired shard]

yeah same

[plain atlas]

@glad pilot 👀

[glad pilot]

I know vimscript well. If you're looking for help bringing the plugin into a supportable state, I might be able to help.

[errant barn]

Aye, that's great to hear!

For example I think help would be greatly appreciated on the stale vim pull requests 😅 https://github.com/psf/black/pulls?q=is%3Apr+is%3Aopen+label%3A"C%3A+vim"

[errant barn]

Sorry, I may have started a bit fast 😄 How familiar are you with Black? (also, I think I've seen your name pop up somewhere)

[bold sierra]

is this where the smart people talk?

[glad pilot]

You may know me from here as a mod, or from Black as the guy who complained loudly about Black changing docstrings in a way that changes the AST. Other than that, maybe a mailing list or something 🙂

I'm familiar with Black as a user, but I've never poked into its implementation at all.

I'm moderately shocked to learn that it has a vim plugin - never knew that, and it seems... odd

[errant barn]

Oh yeah there was a bunch of discussion about the empty docstring issue and such!

[glad pilot]

yep, that'd be me. I came on a bit strong, I think, heh

[errant barn]

no worries, strong personalities are needed too

[errant barn]

Right on, join and you'll become one too ;) let me try that again: So it must be the place for you too ;);)

[bold sierra]

Haha nahh bro

One day maybe

[errant barn]

How so? Should it no be a plugin or :D

[glad pilot]

the existence of a vim plugin is totally reasonable - I use a neovim plugin that has a mapping for running Black. Maintaining it first party as part of the Black repo seems like an odd choice, though - especially if none of the Black maintainers have much vimscript expertise

[errant barn]

Yeah 😅 hence the discussion to drop it

[dense jungle]

Yes, we're not too happy about it either. Theoretically it has benefits for discoverability but not sure it works well in practice

[glad pilot]

it seems to not be mentioned by the README.md, so I'm not sure how discoverable it is in practice.

if you decide you want to keep it, I probably can help you work through the open PRs. Nothing stands out to me as too complicated.

[errant barn]

Another option we discussed was pulling it out in a dedicated repo or combining with some existing plugins

[glad pilot]

I'm a bit unclear on how users would even install this plugin. I guess they could point a vim plugin manager at the Black repo - the plugin manager would clone the repo into one of vim's runtime customization directories, and vim would see the autoload/ and plugin/ directories and use them as a vim plugin - and there would just be a vestigial copy of Black that's also there and not used for anything

[dense jungle]

It's mentioned in https://black.readthedocs.io/en/stable/integrations/editors.html#vim

[glad pilot]

Plug 'psf/black', { 'branch': 'stable' } - so, yeah, that works the way I just said. They clone an entire copy of the Black repo, and only use 2 files out of that clone 🙂

or maybe 3, if there's a help file I haven't spotted

Well, I dunno - if you guys decide you want to maintain it, I can probably help you maintain it, but I think there's a plenty good argument for not maintaining it and making it the problem of someone with more expertise.

And on the other-other hand, if you do remove it, any vim users who ask their plugin manager to update their plugins will pull down a new version of Black that doesn't include the vim plugin, and it'll just break for them. Which is pretty annoying.

[errant barn]

That's a pickle for sure. For now I think that it's worth at least taking care of current contributions. Then we can perhaps start asking around if there would be people interested in taking this on in a longer run and evaluating the options.

[glad pilot]

I'll try to take a closer look at those PRs tonight

[dense jungle]

thank you!

[plain atlas]

nice, pickle

[flat kraken]

cooperlees

[psf/black] Pull request opened: #2325 Update CHANGES\.md for 21\.6b0 release

[bright glacier]

whelp, I'mma probably get a review request for that 😄

[neon loom]

https://tenor.com/view/yay-excited-gogogo-gif-11945751

[flat kraken]

JelleZijlstra

[psf/black] Pull request review submitted: #2325 Update CHANGES\.md for 21\.6b0 release

cooperlees

[psf/black] Pull request closed: #2325 Update CHANGES\.md for 21\.6b0 release

cooperlees

[black:main] 1 new commit

93c10bf Update CHANGES.md for 21.6b0 release (#2325) - cooperlees

[bright glacier]

somehow this is crashing mypyc (heavily simplified from src/black/parsing)

import ast
from typing import Any, Union

ast3: Any
from typed_ast import ast3


def stringify_ast(value: Any) -> None:
    # this line is the issue, removing it or removing the typed_ast AST types avoids the crash
    if isinstance(value, (ast.AST, ast3.AST)):
        pass

    else:
        stripped = [line for line in value.splitlines()]

I get:

❯ mypyc test.py
Traceback (most recent call last):
  File "mypyc/irbuild/builder.py", line 158, in accept
  File "mypy/nodes.py", line 1956, in accept
  File "mypyc/irbuild/visitor.py", line 242, in visit_list_comprehension
  File "mypyc/irbuild/expression.py", line 651, in transform_list_comprehension
  File "mypyc/irbuild/for_helpers.py", line 97, in translate_list_comprehension
  File "mypyc/irbuild/for_helpers.py", line 156, in comprehension_helper
  File "mypyc/irbuild/for_helpers.py", line 122, in handle_loop
  File "mypyc/irbuild/for_helpers.py", line 58, in for_loop_helper
  File "mypyc/irbuild/for_helpers.py", line 278, in make_for_loop_generator
  File "mypyc/irbuild/builder.py", line 702, in _analyze_iterable_item_type
test.py:14: KeyError: <mypy.nodes.CallExpr object at 0x7f2753e2af20>

If I get rid of the Any type annotation for ast3, mypyc stops crashing ...

[flat kraken]

cooperlees

[psf/black] New release published: 21\.6b0

[neon loom]

https://github.com/psf/black/actions - Off goes PyPI upload etc.

[dense jungle]

@bright glacier did you already update to the latest release of mypy? it just came out after a long time, so there's a lot of changes

[bright glacier]

oh yeah I forgot one came like what in this week?

[dense jungle]

I guess that error would happen if it didn't run type inference for some expressions?

yes

[neon loom]

I say now: https://github.com/psf/black/blob/main/.pre-commit-config.yaml#L22

[late dew]

.pre-commit-config.yaml line 22

rev: v0.812```

[neon loom]

*no

[dense jungle]

updating may give you new bugs or new features 😄

[neon loom]

We need precommit cit

*ci

[bright glacier]

I think that's a no-go for us unfortunately since we use a local hook

https://github.com/pre-commit-ci/issues/issues/13

[neon loom]

Hmm - I wonder if we can PR that. I might look one day.

O seems there is an issue at least

[bright glacier]

It's marked with a "pro" label so probably out of reach for us.

oh of course I already forgot third-party stubs need to be installed seperately

oops 😛

[lament crow]

you could just always make a workflow running pre-commit autoupdate right

[bright glacier]

I do have an abandoned script somewhere on my HDD that handles that lol

not really keen on actually using it tho

and yeah no, mypy 0.902 did not fix the KeyError sadly

[dense jungle]

Probably it thinks the else branch is unreachable and therefore doesn't analyze it

also Quora is now running black 21.6b0 😄

[bright glacier]

oh yeaaa now I'm a traitor again because my local black is outdated

gimme a moment

❯ python -m pip install black --upgrade
Requirement already satisfied: black in /opt/python3.8.5/lib/python3.8/site-packages (21.5b2)
Collecting black
  Downloading black-21.6b0-py3-none-any.whl (140 kB)
     |████████████████████████████████| 140 kB 206 kB/s 
Requirement already satisfied: mypy-extensions>=0.4.3 in /opt/python3.8.5/lib/python3.8/site-packages (from black) (0.4.3)
Requirement already satisfied: toml>=0.10.1 in /opt/python3.8.5/lib/python3.8/site-packages (from black) (0.10.1)
Requirement already satisfied: regex>=2020.1.8 in /opt/python3.8.5/lib/python3.8/site-packages (from black) (2020.7.14)
Requirement already satisfied: appdirs in /opt/python3.8.5/lib/python3.8/site-packages (from black) (1.4.4)
Requirement already satisfied: pathspec<1,>=0.8.1 in /opt/python3.8.5/lib/python3.8/site-packages (from black) (0.8.1)
Requirement already satisfied: click>=7.1.2 in /opt/python3.8.5/lib/python3.8/site-packages (from black) (7.1.2)
Installing collected packages: black
  Attempting uninstall: black
    Found existing installation: black 21.5b2
    Uninstalling black-21.5b2:
      Successfully uninstalled black-21.5b2
  WARNING: The scripts black, black-primer and blackd are installed in '/opt/python3.8.5/bin' which is not on PATH.
  Consider adding this directory to PATH or, if you prefer to suppress this warning, use --no-warn-script-location.
Successfully installed black-21.6b0
WARNING: You are using pip version 21.0.1; however, version 21.1.2 is available.
You should consider upgrading via the '/usr/bin/python -m pip install --upgrade pip' command.

totally readable

[neon loom]

Your pip much old

[bright glacier]

fixed as well 😄

[neon loom]

https://tenor.com/view/okay-im-caught-up-im-up-to-speed-up-to-date-up-to-par-okay-im-in-the-loop-now-gif-15083376

[dense jungle]

Python 3.8 though?

[neon loom]

Ok Quora 3.6 guy 😛

[bright glacier]

I do have Python 3.9.5 compiled on my machine I just haven't migrated my main env yet

[dense jungle]

I'm trying 😄

[bright glacier]

hmm, I might have found a simple workaround for the KeyError, I'm hoping that it still works in src/black/parsing.py

yeah I think you're right on the general issue @dense jungle , removing the Any annotation or using this discovered workaround makes mypy yell at me to put back a type ignore (there was originally one in that branch but then using Any, it wasn't needed)

YAY, I fixed it with an actually not too invasive workaround 😄

diff --git a/src/black/parsing.py b/src/black/parsing.py
index c912348..79808ca 100644
--- a/src/black/parsing.py
+++ b/src/black/parsing.py
@@ -5,6 +5,11 @@ import ast
 import sys
 from typing import Any, Iterable, Iterator, List, Set, Tuple, Type, Union
 
+if sys.version_info < (3, 8):
+    from typing_extensions import Final
+else:
+    from typing import Final
+
 # lib2to3 fork
 from blib2to3.pytree import Node, Leaf
 from blib2to3 import pygram, pytree
@@ -134,6 +139,10 @@ def parse_ast(src: str) -> Union[ast.AST, ast3.AST, ast27.AST]:
     return ast27.parse(src)
 
 
+ast3_AST: Final[Type] = ast3.AST
+ast27_AST: Final[Type] = ast27.AST
+
+
 def stringify_ast(
     node: Union[ast.AST, ast3.AST, ast27.AST], depth: int = 0
 ) -> Iterator[str]:
@@ -173,7 +182,7 @@ def stringify_ast(
                 elif isinstance(item, (ast.AST, ast3.AST, ast27.AST)):
                     yield from stringify_ast(item, depth + 2)
 
-        elif isinstance(value, (ast.AST, ast3.AST, ast27.AST)):
+        elif isinstance(value, (ast.AST, ast3_AST, ast27_AST)):
             yield from stringify_ast(value, depth + 2)
 
         else:

OK, next release we're turning on the improved string handling features, what bugs do we still have to squash first?

Glancing at the issue tracker, I see two:

• https://github.com/psf/black/issues/2313 (experimental string processing crash on very long variable name)
• https://github.com/psf/black/issues/2314 (Performance regression with experimental string processing)

[lament crow]

2313 is related to string splitter removing the parent of originally passed line here and never being set back by anything else (so in-place change of something passed by the caller and something else assuming that the parent will exist somewhere else):
https://github.com/psf/black/blob/main/src/black/trans.py#L1038
But that's as far as I went before deciding that it's out of my league for now lol

[late dew]

src/black/trans.py line 1038

insert_str_child = insert_str_child_factory(LL[string_idx])```

[neon loom]

GVR is all over the new parser doc …

[dense jungle]

nice!

[bright glacier]

I'm incredibly confused, could you explain please?

I'm bored while I wait for clang to compile black :p

[dense jungle]

he added some comments to https://docs.google.com/document/d/1QPqm01E7MIRi_l4jrgCGOFVThP_HINnkm4hmVZC_CSg/edit?fbclid=IwAR03VBxQLhXkrZiO7BDUQerkeQU7MOy51BHtJ9xqRMHonBF7_dF1Tef53fs#

and suggested we write our own 😄

[flat kraken]

Ryan-McPartlan

[psf/black] Pull request opened: #2326 Fix incorrect document referance

.pre-commit-config.yaml directs the user to README.md for information about "Version control integrations", but as of #2174 this information is now located at docs/integrations/source_version_control.md.

This commit fixes this broken reference.

[neon loom]

I did lol @ that

[bright glacier]

oh wow lol

I'm thoroughly surprised

It would be relatively easy to adapt the PEG parser generator used by CPython to generate Python instead of C

I mean, that's a valid suggestion

[plain atlas]

hi :D

[flat kraken]

ichard26

[psf/black] Pull request review submitted: #2326 Fix incorrect document referance

Ayy thank you! This is great!

[neon loom]

We have money to pay someone to try do the GVR suggestion too …

[plain atlas]

I just read that, nice

althoughveryminordetailitsTBDnotTBA

[dense jungle]

it stands for "to be ascertained"

[weak path]

no

[plain atlas]

oh

well

i read it as to be announced lol

[dense jungle]

yes, just kidding

[plain atlas]

i did add a con of one of them for you 😛

[dense jungle]

oh yeah, sounds like that's still open: https://github.com/davidhalter/parso/issues/138

[plain atlas]

lmao

this is very funny

[flat kraken]

memeplex

[psf/black] Issue opened: #2327 Inconsistent formatting of list comprehension concatenation

Describe the bug

If concatenating 2 comprehensions, content is indented and surrounded by a pair of parentheses.

To Reproduce

This is an example. The first variable (pre_grid) is assigned the concatenation of two comprehensions, while the second one (gr_grid) concatenates five. The difference in formatting is misleading in this context because both variables have similar semantics. I'm unable to find a good reason why they should be formatted different.

pr...

[plain atlas]

psf/black#1948 led to Instagram/LibCST#490 which depends on Instagram/LibCST#285

[toxic storm]

GitHub

Allow grouping parens around with...as for 3.9
Black fails to parse parenthesized (manager() as x) in with statements
3.10+ support

[dense jungle]

yes I just found that issue too, let me link it to the doc

[plain atlas]

👍

well this is a bit of a turn off Instagram/LibCST#341 👀

[toxic storm]

GitHub

[BUG] Generated Code not equivalent to AST

[flat kraken]

JelleZijlstra

[psf/black] Pull request closed: #2326 Fix incorrect document referance

JelleZijlstra

[black:main] 1 new commit

05b54b8 Fix incorrect document referance (#2326) - Ryan-McPartlan

[bright glacier]

the benefits of doing all of this mypyc related benchmarking is that I have reasonable decent benchmarking tools, which is great because I had planned to create some a long time ago but never did

FYI: I'm just going to say random stuff as I wait for this benchmarking to finish

[plain atlas]

@dense jungle you miiiight want to take a look at this open issue

[plain atlas]

favourite color?

[bright glacier]

Uhm

[dense jungle]

yes, that one could cause issues if we end up using libcst

@bright glacier there's only one right answer

[bright glacier]

Blue and Pink are both great colours

[dense jungle]

wrong

[bright glacier]

purple

[dense jungle]

wrong

[plain atlas]

black ofc

[bright glacier]

no!,

[plain atlas]

and favourite colour with at least one bit flipped on?

[bright glacier]

told ya already

[plain atlas]

this is the best pink

vscode colorized pink logo

[bright glacier]

that looks awesome!

[plain atlas]

i'm just gonna keep vsc as my discord pfp but colorized it for the seasons lol

anyways, cya

[bright glacier]

see ya later

[plain atlas]

oo

I just realized

I have now contributed to the same google document that the founder of python has contributed to

my life has reached its climax, there is nothing higher than this honour.

[bright glacier]

whelp, initial work on further optimizing black after getting mypyc working ain't so successful 😦

[plain atlas]

uh

why is this in a foreign language?

send help I can't read the docs of python

was trying to read the ast docs right now to help with the doc above

[bright glacier]

the universe is trying to convince you to not use Python by removing one of its strengths, its docs /j

[plain atlas]

imagine if the docs were python's strengths /s

[plain atlas]

good point, nice knowing ya. Imma go learn rust have fun!

[bright glacier]

oh fun

[plain atlas]

there's even a built in formatter

[flat kraken]

cooperlees

[psf/black] Issue closed: #2310 Add STDIN primer test

cooperlees

[psf/black] Pull request closed: #2315 Add STDIN test to primer

cooperlees

[black:main] 1 new commit

aa31a11 Add STDIN test to primer (#2315) - cooperlees

[glad pilot]

I took a stab at 2 of the 3 open C: vim PRs, and commented with patches in lieu of being able to push to the PR branch. The third, I don't much like - https://github.com/psf/black/pull/1319/files is proposing to change the public interface of the plugin, so that the :Black command takes arguments. It wants to take them in a strange format - comma-separated pairs of key=val pairs - which I don't think I've ever seen another vim command take. Vim commands that take arguments usually look more like how CLIs take flags and arguments, and I don't think I've ever seen a vim command take comma-separated arguments instead of space-separated. It's strange.

The approach here is basically a hack, I think - the format they chose for the arguments was chosen so that they could be passed directly to a Python function as kwargs, despite that being a non-idiomatic way to pass arguments to a vim command.

I'm not even sure that I buy the argument that it makes sense to be able to pass arguments to a single call to Black, instead of saving settings into a pyproject.toml file or at least setting the g:black_* variables in your vimrc. The whole point of Black should be that you're running it repeatedly to keep the style consistent over time - I'm not sure I understand why you'd want to make it easy to run with different settings than it has been configured to use.

[errant barn]

Thanks a ton! Now we can really make better decisions with them 👍🏿

[errant barn]

Also, the stable docs seem to be missing the 21.5b2 and the latest release. I suspect this is because of the stable git tag, so it should probably be moved!

[mint barn]

hELLO

[errant barn]

@keen oasis I'm not so sure this is appropriate here. The rules of PyDis are quite strict about advertising, and this is a channel about a specific project!

[keen oasis]

@errant barn deleted

[errant barn]

Thank you, much appreciated. Though I hope that you find what you were looking for elsewhere!

[flat kraken]

mironnn

[psf/black] Issue opened: #2328 No new line at file end

Add a new line at the end of python file or not to remove it

Currently black removes extra blank line after python code at the end of file.
Tested inside the docker build from the repository.

root@8c79a371a51b:/# black test.py --check --diff --verbose
--- test.py	2021-06-11 09:53:10.665099 +0000
+++ test.py	2021-06-11 09:53:15.809714 +0000
@@ -1,6 +1,5 @@
 def main():
     print(11)


 main()
-
would reformat test.py
Oh no! 💥 💔 💥
1 file would be reformatte...

[magic bison]

whats this?

a code formatter but what exactly

[errant barn]

For Python! Have a look at the pinned issues messages mate 😄

[magic bison]

how does it work?

like how does it know what to fix

[bright glacier]

pinned issues mate?

I think you meant pinned messages

[errant barn]

Already there mate ;)

[errant barn]

It uses and edits a syntax tree to produce new code

[bright glacier]

it has logic that takes whatever input and turns it into code that follows the "black code style"

[magic bison]

ok

[bright glacier]

it has no true sense of whether code is good or bad it just knows how to take some code and make changes that follow the black code style

[magic bison]

so its like pylint but it fixes stuff instead of telling you whats wrong

[errant barn]

eexactly!

[magic bison]

oh ok

[bright glacier]

in a way yes, but we don't have error messages and we can't, black is a formatter not linter

[lament crow]

black is just about style, pylint is about more

but black is very good at style ;)

basically let the computers worry for code style, not humans

[neon loom]

Yeah ichard noticed that. I'll move in tonight/tomorrow if no 21.6b0 issues come in

[bright glacier]

!remind 1D check on 21.6b0 release status and maybe move stable tag

[late dew]

Fat chance.

Sorry, you can't do that here!

[bright glacier]

oh lol

I guess I'll just use the bot commands channel 😦

[tribal thistle]

Uhh

We could probably enable that here

[bright glacier]

ah looking at the code, it seems like staff can make a reminder anywhere while everyone else (including me!) can make in these places:

https://github.com/python-discord/bot/blob/7a1996b204472ec58fbc6f4af5caa09300e72f95/config-default.yml#L244-L246

[late dew]

config-default.yml lines 244 to 246

reminder_whitelist:
    - *BOT_CMD
    - *DEV_CONTRIB```

[tribal thistle]

Should be a single line change

Would you like to PR this category into that?

I would do it, but not at PC

[bright glacier]

Would it need an issue for further discussion first?

[tribal thistle]

A core dev can sign off with no issue

You can link to my messages here

[bright glacier]

Cool then, I can PR it later once my time is freed up in an hour and a half

me wishes I could use reminders here so I don't forget to do the PR

[tribal thistle]

Hahahaha

Cool trick I got

[bright glacier]

i.e. privliiges

[tribal thistle]

!remind 90M 429445826251587616 dew it

[bright glacier]

lol

[dense jungle]

such power

[tribal thistle]

!remind 429445826251587616 90M dew it

[late dew]

ROGER THAT

Your reminder will arrive in 1 hour and 30 minutes and will mention 1 other(s)!

ID: 2366, Due · Jun 11, 2021, 6:44 PM

[tribal thistle]

Or I could just fuck it up

[bright glacier]

you could ping me in that reminder too can't you with your admin powers

OH

[tribal thistle]

Ye

[bright glacier]

that is my id

I didn't recognize it lol

[tribal thistle]

I wouldn’t at a glance either haha

[dense jungle]

just curious, how do you find that id?

[tribal thistle]

Press/right click on a user and click “copy ID”

You may need to enable developer mode first

[bright glacier]

can confirm this requirement

[dense jungle]

thanks, that did it

[late dew]

@tribal thistle @bright glacier

It has arrived!

Here's your reminder: dew it.
[Jump back to when you created the reminder](#black-formatter message)

[plain atlas]

very nice reminder mentions

now dew it 👀

oh since I'm making a bunch of vsc logos

i made one for black

nah this is the real one

[bright glacier]

the contrast is so bad it's beautiful /s

[bright glacier]

ugh, the setup required for testing changes for pydis-bot seems to be really involved

[frozen yacht]

We know and have plans to address it, but not enough time to work on that stuff lately.

[bright glacier]

It's acceptable, I've just gotten used to being really simple to get a dev env quickly for the python projects I work on

[frozen yacht]

If you're just whitelisting this channel, it should only be a config change and could forego testing.

[bright glacier]

We'll see, I don't trust myself enough to do an one line change and I might need a pydis-bot test env later so I'll try for now and see how much effort it will really take 🙂

[bright glacier]

Yeah no I do not feel like setting this up, got like half way done and I don't feel like doing this lol

Does show how big this server really is though

[tribal thistle]

Yeah it’s not fun to set up lol

[bright glacier]

everything except for the channels is acceptable, but my goodness there is a lot of channels

[tribal thistle]

Lol

The list is terrifying

[bright glacier]

it really is

[tribal thistle]

Anyways, you could add it in without testing, or I could have a look at it in the morning

[bright glacier]

The good news out of it is that it was my first time using poetry lol

[tribal thistle]

Yeah? How was it 👀

[bright glacier]

I had a not so great experience with pipenv (although I still use it for black) so I gave up on that class of tools forever until now

But poetry actually felt polished, I'm not sure if the "handle everything" nature of these tools is something I'll grow to like, but at least I feel like care was taken :p

[tribal thistle]

Poetry actually is less “handle everything” than pipenv

It doesn’t have scripts

Ofc we have a plug-in to do it…

[bright glacier]

So wait, I just make the oneliner change in /config-default.yml right?

[tribal thistle]

This category is probably not in the constants file, so it’ll have to be added in as well

[bright glacier]

oh fun

[tribal thistle]

That just involves adding the ID to the yaml, and the type (int) to the constants.py

[bright glacier]

So correct me if I'm wrong, but I need to do:

• get the ID of this category and put in the YAML file

• add the type to constants.py

[tribal thistle]

Yeah that’s basically it

[bright glacier]

• and finally add this channel to that other section

[tribal thistle]

Yeah that step is pretty important too lol

[bright glacier]

OK, not too bad although I'm the guy who would make 10 force pushes for a small change 😅

[tribal thistle]

I can’t remember if the filter handles categories. If it does, I think it would make sense to add the entire OSS category, if not, we could add just black. I should check lol

[tribal thistle]

Well, the policy allows you to do it 🤡

[bright glacier]

Is the server template up to date? It didn't exactly look like it

I mean, it sounds like a pain to keep up to date, but I'm curious

[tribal thistle]

So basically, yeah, but, it doesn’t have all the channels. It has all the channels you need to setup and run the bot, but fewer channels than the main server to keep it navigable

[plain atlas]

hey i like it!

[tribal thistle]

The only things you’d have to add on to the template are one emoji of your choice

Oh we probably drowned this channel

[bright glacier]

This should probably be documented somewhere lol

[tribal thistle]

Would you like to continue in #dev-contrib?

[plain atlas]

sometimes 10 commits are made for a small change and then just forcepush or squash merged

[bright glacier]

Yeah probably

although honestly

this channel gets derailed constantly by me

[tribal thistle]

Lol I’ll take this one

I can get excited inducting a new sacrificial goat to the academy of pydis contributing

[bright glacier]

!remind 20H check on 21.6b0 ~release~ bugginess status and maybe move stable tag

[late dew]

Sure.

Your reminder will arrive in 20 hours!

ID: 2368, Due · Jun 12, 2021, 6:33 PM

[bright glacier]

I make that reminder and yet I don't see any new bugs since the release yesterday

just psf/black#2328 which doesn't seem like it should exist, IIRC it was fixed in 20.8b1

[toxic storm]

GitHub

No new line at file end

[flat kraken]

fpdotmonkey

[psf/black] Issue opened: #2329 \`\# fmt: off\` fails to stop formatting on a comment block

Describe the bug

I have this file header in my source that I'd like to not mess with since it's got license information for code that isn't mine. So I wrap it in # fmt: off; ...; # fmt: on to preserve the block, but the block gets formatted anyway (a space gets added after the #).

Before black:

# fmt: off
#===- cindex-dump.py - cindex/Python Source Dump -------------*- python -*--===#
#
# Part of the LLVM Project, under the Apache License v2.0 with LLVM Excepti...

[bright glacier]

that issue is a duplicate fyi

I see we are both triaging it at the same time

[flat kraken]

ichard26

[psf/black] Issue closed: #2329 \`\# fmt: off\` fails to stop formatting on a comment block

[bright glacier]

our reaction time is sometimes nuts

[plain atlas]

this is an odd Testimonial

[flat kraken]

ichard26

[psf/black] Issue closed: #1446 Fix Formatting Errors on black\-primer CI

ichard26

[psf/black] Issue closed: #1696 Comments that are vertically aligned across 3 or more lines can remain vertically aligned\.

[bright glacier]

Alright, I got the 2 vim PRs godlygeek suggested patches for updated, they should be ready for review (not sure by who tho) and then merging

Trying to get the open PR count lower to reduce the burden on PR review resources

[plain atlas]

oof

[bright glacier]

one of mine I take it as?

[plain atlas]

its not 😛

the funny thing is, its old enough that it only had 26 checks

rather than the now 36

[bright glacier]

and we might get more once I hook up mypyc to black

[plain atlas]

😦

what's mypc?

do ping im gonna go back to the rust server for now~~

[bright glacier]

a compiler for statically typed python code, turns such code into a Python C extension for speed improvements @plain atlas

[plain atlas]

so black will be a c extension?

[bright glacier]

Perhaps obvious by the name, it's based off mypy

[bright glacier]

Yes, but that should be an implementation detail (well there's a few breakage concerns but they should be minor)

The main one is that monkeypatching will break (this does actually break some of our test suite) but expecting that to work (and stay supported) for an application that technically isn't officially a library is IMO a stretch

[flat kraken]

godlygeek

[psf/black] Pull request review submitted: #1869 Vim plugin fix string normalization option

[glad pilot]

I spotted a typo.... Or, merge-o? a minor mistake in one of them. The other looks fine, but I made a nitpicky suggestion about the changelog entry.

[tribal thistle]

Hey @gritty marsh please keep messages on topic

[muted haven]

Is the lib2to3 parser framework powerful enough to support features from the new PEG parser? CPython docs say lib2to3 is deprecated because it can’t support PEG syntax.
Chiming in late, it depends on the complexity though it is possible (at least as far as my experimentations go). Both parenthesized context managers and soft-keywords can be implemented, though in case of ambiguity it might require a bit more annotations (something like cut in the PEP 617, to commit to a rule). It is kinda harder when you try to do it on a parser with error handling (parso) but for black it should be easier to implement those 'hacks'.

[dense jungle]

Thank you!

[flat kraken]

godlygeek

[psf/black] Pull request review submitted: #1871 Find pyproject from vim relative to current file

LGTM

godlygeek

[psf/black] Pull request review submitted: #1869 Vim plugin fix string normalization option

LGTM

[glad pilot]

The third vim PR - https://github.com/psf/black/pull/1319 - is a feature request, not a bug report.

Two features, in fact:

1. the ability to set the target version for the plugin's call to Black (new vim variable for setting the default, plus ability to read a per project setting from pyproject.toml)
2. the ability to override every setting on the vim command line by making the :Black command take arguments.

If a maintainer decides that the features should be implemented, I can help make a more idiomatic implementation of them, but I'm probably not in a position to be making decisions about the direction of the plugin on my own.

[late dew]

@bright glacier

It has arrived!

Here's your reminder: check on 21.6b0 ~release~ bugginess status and maybe move stable tag.
[Jump back to when you created the reminder](#black-formatter message)

[bright glacier]

ah yes

[bright glacier]

@neon loom I suggest to move the stable tag, the newline being stripped issue is unreproducible for both me (using the docker steps the OP provided) and Felix (using a normal dev env since he doesn't know docker AFAIK)

[neon loom]

Got for it f you want

I can’t for an hour or so

It’s in the docs

[bright glacier]

Ah ok, sounds good!

Let's hope I don't break anything 😅

[frozen yacht]

Stable tag reminded me of something I wanted to ask. Does the pre-commit hook fail if it detects black needs to format something?

I was avoiding using stable because I figured I need to run it separately, and "stable" is obviously not on pypi

[bright glacier]

Well, honestly I wouldn't recommend using stable for pre-commit (even though IIRC that's what is in the pre-commit integrations docs)

pre-commit won't track changes to the stable tag

which is IMO a good thing (also git tag ]safety makes that annoying anyway)

But yes, the pre-commit hook should fail if it makes modifications to a file watched by pre-commit

although you can also use the --check flag to make black return a non zero return code if it would make changes (it doesn't actually touch the files tho)

[frozen yacht]

Ah okay I thought it was going to just fix it all and include it in the commit. Now that I think about it, I'm not sure if pre-commit even has the means to do that.

[bright glacier]

IIRC the pre-commit developer rejected an feature request for automatically adding the changes from a failing hook

[frozen yacht]

Wouldn't it be ideal to use --check in pre-commit always then? It's redundant to actually perform the fixes.

[lament crow]

how is it redundant?

[frozen yacht]

I mean in CI

[bright glacier]

in CI yes

[frozen yacht]

Sorry, important to mention that

[lament crow]

well

you can still benefit from diff in CI if you use --show-diff-on-failure

[bright glacier]

it doesn't really matter tho

[lament crow]

or rather contributors can

[bright glacier]

pre-commit and black both have diff options

[lament crow]

and speedwise should be mostly the same

[frozen yacht]

Would there be a nice way to integrate that output with GitHub annotations?

[bright glacier]

Not implemented although accepted: psf/black#1764

[toxic storm]

GitHub

add support for Github Annotations

[plain atlas]

Ugh so that's why

[frozen yacht]

Okay, cool. I'll look forward to that then.

[bright glacier]

I wouldn't hold your breath, our feature development leadtime ain't great, especially since I expect significant discussion for this implementation because that will be the first output format (other than diff) black will support

[frozen yacht]

Why not? You said it won't track changes but then go on to say that's a good thing.

[plain atlas]

This is why black reformats the code with a hook but doesn't actually save the code ;-;