https://i.imgur.com/ejli3Vm.png
Just had this pop up in my neovim when trying to run :Black, its been working great until recently. I never had this message before. Currently on python 3.12 and nightly neovim
#black-formatter
analog anvil
flat krakenBOT
neon loom
Am I blind - is there any docs on the easiest way to build a compiled black wheel? I'm trying to change the Dockerfile to compile by default. Should
hatch build --ext
Get me a wheel?
crl-m1:black cooper$ /tmp/tb/bin/hatch build --ext
[wheel]
I'm a hatch / mypyc n00b
Once i get a compiled wheel I plan to install it to /opt/venv which we copy to the published docker image
re-reads ichard's blog post
Damn, it's all using setup.py ... not gonna helps me I think
Ok - Got a non compile - Now to workout how to make it try to compile
crl-m1:black cooper$ /tmp/tb/bin/hatch -v build -t wheel
[wheel]
Building `wheel` version `standard`
dist/black-23.10.1.dev1+g882d879.d20231022-py3-none-any.whl
At least got us to use hatch: https://github.com/psf/black/pull/3965 - Will take any pointers on how to make it compile the wheel ...
zinc shell
Hey, mmm I just had a very bad experience asking for help in #ot0-fear-of-python ...
I am lookin for help as I am trying to learn..
recently i installed a program called "X-Osint" from github and it was unstalled using sudo
does anyone know how to uninstall it?
i installed it in kali linux on windows 11
already tried sudo apt-get remove xosint
but it says "E: Unable to locate package xosint"
can anyone help me here?
i just want to unsintall it as it is of no use
neon loom
I don't think is the right channel to ask for help on this friend.
flat krakenBOT
flat krakenBOT
flat krakenBOT
neon loom
Sorry was in meetings and missed you merged. Thanks π
neon loom
Do we want to get a hotfix out today for the github action failure?
dense jungle
yeah probably. We've accumulated a few other improvements too
neon loom
If you want I can drive it (have to read docs to remember how :D)
Happy to try start the diff to make changelog is up to date etc.
dense jungle
If you want I can drive it (have to read docs to remember how :D)
sounds good, yes the docs page should be good. I always follow it when making a release
neon loom
Ok - I'll cut one soon - Got side tracked on $job work
Anything else we want merged or you're happy with where we are?
dense jungle
Ok - I'll cut one soon - Got side tracked on $job work
just merged one. there's a couple other PRs that are near ready but they can wait
neon loom
We need https://github.com/psf/black/pull/3957 for GitHub action - So I'mma gonna merge that
I guess we should open an issue for Windows Doc builds failing
flat krakenBOT
neon loom
I feel we should add the new release block after release so we don't pollute pypi.org + tagged commit with it :\
I disagreed with it in the docs π
dense jungle
I feel we should add the new release block after release so we don't pollute pyp...
sure that's fine too, just means one more thing to do later
neon loom
Yeah, I'll do a diff and change the docs right after release
@dense jungle Would you accept a PR adding a shell script to run that sed -i s/$OLD_REL/$NEW_REL/g for the files we have version strings in? I dislike editing those strings π
dense jungle
<@783088578363523104> Would you accept a PR adding a shell script to run that `s...
sure, it's really annoying to edit them manually
neon loom
I might write in python so it works for our lovely Windows people π
flat krakenBOT
dense jungle
I might write in python so it works for our lovely Windows people π
for extra points, make it also clean up the changelog π
neon loom
Yeah, I'll do that too. python I can write unittests too cause I don't trust myself
https://github.com/psf/black/issues/3970 - Add any features I forgot. No wife this week so I might actually write it tonight π
The compile build takes a long long time in docker
Cause we compile x86 and then cross compile aarch64
I should see if I can get them to go in parallel like ichard does for our wheel building
flat krakenBOT
neon loom
^^ Would love anyone with killer GitHub actions CI skills here to PR plz π
I am stuck in the Meta world, so my GitHub skills are limited
docker CI == 10x runtime regression - 5 mins to 50 mins ... lol π¦
dense jungle
docker CI == 10x runtime regression - 5 mins to 50 mins ... lol π¦
but the resulting image is several times faster π
neon loom
Ok - all built - Cutting release
flat krakenBOT
neon loom
Our docs should auto build etc. right?
Seems CI will move the stable branch and then a docs build will kick off?
dense jungle
Seems CI will move the stable branch and then a docs build will kick off?
yes, all should be automated
thanks for doing the release!
neon loom
No worries, good to do it after a long break π
neon loom
Hmm, seems the release breaks black-pre-commit-mirror
It needs some cronjob?
Wonder if we can click run now or something and then think of a better way
dense jungle
Wonder if we can click run now or something and then think of a better way
I just pushed some button there
neon loom
Yeah seemed to work
flat krakenBOT
neon loom
Does psf org have access to larger runners? cc: @solid adder ... Would like to move the docker ci there if we can ...
solid adder
Does psf org have access to larger runners? cc: <@688057416062861319> ... Would ...
No idea if we have larger runners. I'll ask Ee
neon loom
Thanks - No biggy if we don't
I just regressed the docker build 10x moving it to compile black (5 mins to 50 mins) - I was surprised it was that much
We should be able to ~1/2 that doing the aarch64 + x86_64 builds in parallel ... Just gotta workout how to do that
flat krakenBOT
neon loom
Anyone know how to install git on a Windows Github Action runner? If not, I'm gonna remove it from the CI for release tool CI for now and open a issue for someone to help with it ... Cause I haz 0 windows skills
I'll try google it now
neon loom
Damn, sorry about the noise people subscribed. It looks like I'm gonna have to move back to python3.12 release_test.py and add print statements to see why we're failing with more detail - coverage seems to hide everything but the last exception ... but it seems I'm not getting git tag output, but I felt I had mocked git running out ...
clever peak
out of curiosity, do yall use mypy's stubgen at all?
dense jungle
out of curiosity, do yall use mypy's stubgen at all?
not in Black
neon loom
Damn, sorry about the noise people subscribed. It looks like I'm gonna have to m...
Ok. Finally with a non shallow checkout my tests work β¦
Some nice passer-byer informed me! Makes sense. I didnβt realize that limited tags.
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
neon loom
So I'm a little lost, what do we need to do to make our MacOSX cp11 wheels happy? Are we waiting for fiixes or we have to make some >= 11_0 version? I feel threads contradict ...
I'd be happy to fuck a 23.10.2 or at least have it ready for 23.11.0
Do we want macos-max-compat?
gusty grail
Maybe
But that change is kinda specific
Like you'd have to be actually interested in reading that doc
dense jungle
I guess good that people are reading the docs, in that case π
flat krakenBOT
flat krakenBOT
elder tusk
Do we want `macos-max-compat`?
yeah, I think setting to false will fix, but hadnβt yet gotten a chance to test
flat krakenBOT
flat krakenBOT
gusty grail
https://hacktoberfest.com/participation/#spam
just putting this out here
flat krakenBOT
flat krakenBOT
flat krakenBOT
half meteor
Hey Black maintainers, I will finally have some time for https://github.com/psf/black/issues/1352#issuecomment-1379257052
Are you still interested in this? Since this will be a big change, I hope to get buy-in before I start sending PRs.
It's currently implemented in Pyink via this interface: https://github.com/google/pyink/blob/5fd9550dc109ed26801b5999d356c56ad6bb4eb1/src/pyink/__init__.py#L311-L320
The majority of the logic is in this convert_unchanged_lines function: https://github.com/google/pyink/blob/5fd9550dc109ed26801b5999d356c56ad6bb4eb1/src/pyink/__init__.py#L1221, which is similar to how fmt: off/on is implemented.
late dewBOT
src/pyink/__init__.py lines 311 to 320
@click.option(
"--pyink-lines",
multiple=True,
metavar="START-END",
help=(
'Range of lines to format. Must be specified as "START-END", index is'
" 1-based and inclusive on both ends."
),
default=None,
)```
`src/pyink/__init__.py` line 1221
```py
ink.convert_unchanged_lines(src_node, lines)```yes, I think this would be useful. Thanks!
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
neon loom
Damn, black is active at the moment
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
young sand
Looks like that might have been closed accidentally due to the PR description in #black-formatter message
flat krakenBOT
flat krakenBOT
flat krakenBOT
normal solar
Hello
mint barn
self bot
tight hollow
"D:\Program Files (x86)\Anaconda\Scripts\black.exe" C:\Users\sasa\Desktop\Lesson_4\Lesson 4.ipynb
I want to format "Lesson 4.ipynb". Obviously, there's a blank space in the middle. So how can I format this file without rename. BTW, it can be opened as start "Lesson 4.ipynb" in cmd. But I can't "C:\Users\sasa\Desktop\Lesson_4\Lesson 4.ipynb" in black. I would appreciate it if anyone could help me!
dense jungle
"D:\Program Files (x86)\Anaconda\Scripts\black.exe" C:\Users\sasa\Desktop\Lesson...
possibly enclose the name in quotes ("C:\Users\sasa\Desktop\Lesson_4\Lesson 4.ipynb"). But I don't use Windows and I've heard shell quoting in Windows is weird
tight hollow
PS C:\Users\sasa\Desktop\Lesson 3> "D:\Program Files (x86)\Anaconda\Scripts\black.exe" "C:\Users\sasa\Desktop\Lesson_4\Lesson 4.ipynb"
- ... da\Scripts\black.exe" "C:\Users\sasa\Desktop\Lesson_4\Lesson 4.i
pynb" -
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
β"C:\Users\sasa\Desktop\Lesson_4\Lesson 4
.ipynb"βγ
+ CategoryInfo : ParserError: (:) [], ParentContainsErro
rRecordException
+ FullyQualifiedErrorId : UnexpectedToken
I already tried. But it didn't work.maybe single quotes instead? for what it's worth that error doesn't come from Black
tight hollow
PS C:\Users\sasa\Desktop\Lesson 3> "D:\Program Files (x86)\Anaconda\Scripts\black.exe" 'C:\Users\sasa\Desktop\Lesson_4\Lesson 4.ipynb'
- ... da\Scripts\black.exe" 'C:\Users\sasa\Desktop\Lesson_4\Lesson 4.i
pynb' -
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
β'C:\Users\sasa\Desktop\Lesson_4\Lesson 4
.ipynb'βγ
+ CategoryInfo : ParserError: (:) [], ParentContainsErro
rRecordException
+ FullyQualifiedErrorId : UnexpectedToken
it still can't workCannot run program "D:\Program Files (x86)\Anaconda\Scripts\black.exe" (in directory "C:\Users\sasa\Desktop\Lesson 3 "): CreateProcess error=193, %1 is not a valid Win32 application. what happens with this, at first time, it can be formatted
flat krakenBOT
elder tusk
.......well i fixed it https://github.com/psf/black/pull/4017
flat krakenBOT
flat krakenBOT
echo pebble
Hello
I am trying to use black in a gitlab ci
black:
stage: linting
image: registry.gitlab.com/mafda/python-linting
script:
- python -m black . --check --verbose --diff --color
But it says : Error: No such option: --color Did you mean --code?
Anyone has any idea, please?
dense jungle
Hello
I am trying to use black in a gitlab ci
```yml
black:
stage: linting
...
which version are you using? it looks like we added --color in 2020 (commit 8d6d92aa5b5248b5ff70ebf7977f8af5cbcb10b9)
echo pebble
I think it's because of gitlab or something related.
Removed --color and it works
This image: registry.gitlab.com/mafda/python-linting
Do you recommend another image for this?
dense jungle
I think it's because of gitlab or something related.
Removed --color and it work...
No idea, never heard of that and I don't know how Gitlab CI works
seems like it hardcodes black==19.10b0
which just had its fourth birthday
that suggests it's not a well-maintained project
echo pebble
i understand. Thanks
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
we should do a release in the next few days after the current crop of PRs is merged
flat krakenBOT
flat krakenBOT
half meteor
The diff-shades preview checker seems to be failing on the CI?
dense jungle
The diff-shades preview checker seems to be failing on the CI?
yes, that's been happening intermittently. I haven't been able to find useful logs to debug it
flat krakenBOT
flat krakenBOT
dense jungle
the docker build on main has been running for 47 minutes... π±
flat krakenBOT
flat krakenBOT
dense rune
should https://github.com/psf/black/pull/3940 be applied to black-pre-commit-mirror?
dense jungle
should <https://github.com/psf/black/pull/3940> be applied to black-pre-commit-m...
quite possibly, feel free to send a PR
dense rune
quite possibly, feel free to send a PR
flat krakenBOT
half meteor
This is indeed caused by the diffing algorithm in adjusted_lines.
I'm not sure this is really fixable, since you should always be able to construct an edge case to make the diffing algorithm produce an undesired output.
What's possible is 1) get rid of the two-pass format in Black; 2) skip assert_stable when --line-ranges is used (https://github.com/psf/black/blob/58f31a70efe6509ce8213afac998bc5d5bb7e34d/src/black/__init__.py#L1472)
late dewBOT
src/black/__init__.py line 1472
lines = adjusted_lines(lines, src, dst)```We close ...
flat krakenBOT
flat krakenBOT
half meteor
yes, that's been happening intermittently. I haven't been able to find useful lo...
I tried diff-shades with the preview style locally a couple of times and it has been fine.
Looking at a recent run: https://github.com/psf/black/actions/runs/6814056685?pr=4037
The hosted runner: GitHub Actions 7 lost communication with the server. Anything in your workflow that terminates the runner process, starves it for CPU/Memory, or blocks its network access can cause this error.
Maybe it's using too much CPU or memory?
dense jungle
I tried diff-shades with the preview style locally a couple of times and it has ...
yes, that might be it. If so, we can consider running it on fewer projects
half meteor
yes, that might be it. If so, we can consider running it on fewer projects
I wonder if there is a good way to cap it at like 80% of the CPU, it will take longer time it might not get killed this way.
elder tusk
maybe we have a perf regression?
half meteor
Possible, but https://github.com/psf/black/actions/runs/6807549150/job/18510596967 is a successful run last night based on current main with pending changes, Analyze target revision took 10m 54s
I checked a few runs from past week on the main branch they took 10 to 14 minutes.
earnest sequoia
Is it expected that the playground can't match statements? https://black.vercel.app/?version=main&state=_Td6WFoAAATm1rRGAgAhARYAAAB0L-Wj4AGdAHldAD2IimZxl1N_WmBasjhgIpJRm9DfswfTHVK7-U27RtCcvConAPxeBZhIzFLWwR4J5lDZ9f93h4U-y9h1zPSvJP4tcEhTz0ZrRS8f59LHYgd3OOKRaNIkYFVCbwTZMZCSkozoe599gj1SFizhchKv3zLAu4JnTpkRhoAAAAAAjaC-ekbT5nYAAZUBngMAAFNSR-yxxGf7AgAAAAAEWVo=
I couldn't find an open issue for that
austere lava
Try ticking "skip temporary sanity checks" in the settings (bottom left)
earnest sequoia
yep, but why does that fail on the playground when it works locally?
austere lava
Because the playground uses a runtime older than 3.10
upbeat escarp
hi, this is a general question. is black supposed to automatically fix your code. I have the settings shown and can see on save it triggers, but it only highlights the errors instead of auto fixing them. I am on ubuntu wsl2 with black installed as a vscode extension (and in my venv as a sanity)
"editor.formatOnSave": true,
"python.formatting.provider": "black",
"[python]": {
"editor.defaultFormatter": "ms-python.black-formatter",
"editor.formatOnSave": true
}
}
bright glacier
Black doesn't flag any errors. You must have a linter or type checker installed, too. Perhaps flake8 or ruff or mypy?
Black only handles code formatting so while it could probably address many code formatting lint errors that flake8 (for example) would raise, that's just a happy side-effect :)
upbeat escarp
i have flake8, i was recommended to add on googling, do i have to remove it to get the auto-formatting of my code to work. i just want for example re sult = x**2 to be auto corrected on save
bright glacier
black doesn't fix errors
and yes, it does need syntactically valid code or else it will probably fail to format
upbeat escarp
oh.. so a very dumb question. how come it shows in this video the code auto formatted? https://youtu.be/YapHfCRCpig?t=342
bright glacier
I'm confused. That's what black is supposed to do (assuming VSCode is configured to call it automatically on save)
Black is not a linter. It merely modifies your code stylistically, adjusting indentation, spacing, line breaks, etc.
What code are you trying to format?
upbeat escarp
ah i see. so in my example re sult = x**2, because the variable is not valid, black wont format it.
But for bad styled objects like x = {'a': 38, 'b': 42, 'c': 900 } it will fix
bright glacier
badly formatted
bright glacier
ah i see. so in my example `re sult = x**2`, because the variable is not valid, ...
It's because Black won't be able to parse the code.
It needs to be able to parse (understand the structure of your code) to be able to format.
upbeat escarp
ah i understand. is there some other addon that does what i want to achieve
parse and format
upbeat escarp
yes
bright glacier
I wouldn't be surprised if they exist, but I would strongly recommend learning to resolve them by hand instead. Such tools probably wouldn't be particularly reliable.
upbeat escarp
i see, thank you for clarifying the difference/expectation
bright glacier
The lines are becoming more and more blurred. Ruff, a hypermodern Python linter does have some autofixing functionality for its lint rules.
I doubt it can handle syntax errors, but it can fix certain lint violations at least.
young sand
isort occasionally fixes SyntaxErrors for you: https://github.com/python/typeshed/pull/10055/commits/9bf3296654f10069e2b7382ac994d442eb8738e6
But I personally think of that as a bug rather than a feature π
isort also occasionally introduces new SyntaxErrors for you, so, you know β swings and roundabouts https://github.com/PyCQA/flake8-pyi/pull/411/commits/a69c12d4ba436dca486eab8b6c1d6b2d3b04f500
flat krakenBOT
flat krakenBOT
flat krakenBOT
neon loom
I won't die on a hill, but this makes me sad - One of my most hated things about rust formatting.
lament crow
While I like this formatting visually, it's not at all what I'd expect Black style to look like based on how Black style looked in the past
I know that that's why the changes to the style are only made once a year but it's still quite significantly different from what existing users (including me) might expect
neon loom
Yeah, when we started it, since indentation is so important to python, we wanted to make it cleaer everywhere and never wanted to optimize for vertical white space etc.
I find this much harder to easily read and pickup the logic
Now I admit I'm 200% bias as I've been reading python code in black style before black existed (as ambv, myself and other facebookers decided the style on whiteboards @ work)
https://github.com/pypa/bandersnatch/pull/1606/files/979a8a6c8a8da08767fb63a92381ab09394fedc8..52cc530d50ec5c46f2a148ca3d135b26cdb4bb3f - If people want to see the examples ... as I said ... I can live with it. Just not a fan.
half meteor
You still have a choice to use the magic trailing comma case by case here to get the previous behavior
neon loom
O do I; so we now respect the last trailing comma more accurately?
Side note: I wish we never did the magic trailing comma ... Was meant to be a consistent style ... Now, even limited as much as we have, the black style is quite flexible
half meteor
For this particular case, yes.
For our story we have a large developer base, and having some flexibility like this played a key role to adopt the formatter while keeping most of them happy.
flat krakenBOT
flat krakenBOT
elder tusk
I won't die on a hill, but this makes me sad - One of my most hated things about...
want to comment on https://github.com/psf/black/issues/4042 ? i don't think it's a foregone conclusion, after all, hug_parens is probably the biggest change we've ever made to black's style
also for folks who dislike hug_parens feature, i'm curious how you feel about multiline_string_handling feature
gusty grail
!pban 1174251412641828878 30d Looks like you're only here to advertise. We're not an ad-board.
late dewBOT
:incoming_envelope: :ok_hand: applied ban to @tawny palm until <t:1702630387:f> (1 month).
balmy siren
i am facing an issue where I have to keep single quotes as practice instead of double quotes,
dependencies = [
('slack_automation', '0001_initial'),
("slack_automation", "0001_initial"),
but when i use "black ." it changes it to double quotes
dense jungle
i am facing an issue where I have to keep single quotes as practice instead of d...
you can use black --skip-string-normalization so it doesn't change quote style. But if your assignments require very specific style, maybe Black isn't the right tool
balmy siren
you can use `black --skip-string-normalization` so it doesn't change quote style...
no just the double quotes are an issue,
but if i have already have the code in double quotes can this command change it back to single quotes?
dense jungle
no
balmy siren
this cmd is giving an error, black --skip-string-normalization
black --skip-string-normalization
Traceback (most recent call last):
File "<frozen runpy>", line 198, in run_module_as_main
File "<frozen runpy>", line 88, in run_code
File "C:\Work\workx-backend\venv\scripts\black.exe_main.py", line 4, in <module>
File "C:\Work\workx-backend\venv\Lib\site-packages\click_init.py", line 7, in <module>
from .core import Argument as Argument
File "C:\Work\workx-backend\venv\Lib\site-packages\click\core.py", line 615
rv = f'{' '.join(parent_command_path)} {rv}'
flat krakenBOT
flat krakenBOT
flat krakenBOT
half meteor
I took another look at possible preview style regressions.
I ran diff-shades at each recent commit on my M1 Mac Mini, here are the results: https://paste.pythondiscord.com/FAVQ
It did take noticeably longer after https://github.com/psf/black/commit/1d4c31aa589dc0c8633af7531f8cc09192917b38 and https://github.com/psf/black/commit/b7717c3f1e73d6b847e2534a2cebbb657b96caf8
The reason it takes longer is not that the newly added code has a perf issue. It's because more files will be reformatted, so more files will make Black perform a second formatting pass, doubling the time for affected files.
neon loom
Damn, splitting the builds for docker did nothing sadly + broke it as they tag seperately so we don't get same arch builds with same tags ... https://hub.docker.com/r/pyfound/black/tags
Times were:
- amd64: ~4 mins
- arm64: ~35 mins
Old flow was ~30-40 mins in total ... so we really won nothing here sadly ...
I'll put up a PR reverting it ...
- Potentially broke latest + latest_non_release tag users ... Once we revert and merge to main it will build together again and tag both architecture versions together again ...
neon loom
Also see while I watch building:
#21 62.12 [wheel]
#21 62.76 /root/.local/share/hatch/env/virtual/black/VsnhxLU2/black-build/lib/python3.12/site-packages/setuptools_scm/git.py:163: UserWarning: "/src" is shallow and may cause errors
#21 62.76 warnings.warn(f'"{wd.path}" is shallow and may cause errors')
So if we ever see weirdness ... we need to change to full checkouts in our build workflows ...
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
late dewBOT
:incoming_envelope: :ok_hand: applied timeout to @warm burrow until <t:1700599475:f> (10 minutes) (reason: emoji spam - sent 50 emojis).
The <@&831776746206265384> have been alerted for review.
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
glacial harness
Hi!
I was wondering, will this error be fixed in the next black release?
d = { "foo": "test" }
text1 = f"this is just a {d["foo"]}" # Black fails parsing that despite it being correct in Python 3.12
text2 = f"this is just a {d['foo']}" # no problem```Follow https://github.com/psf/black/issues/3746 for updates
glacial harness
Thank you :D
flat krakenBOT
late dewBOT
:incoming_envelope: :ok_hand: applied timeout to @wheat island until <t:1701122145:f> (10 minutes) (reason: burst spam - sent 8 messages).
The <@&831776746206265384> have been alerted for review.
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
half meteor
Another finding on the diff-shades preview failures:
I tried to print the memory usage when it's running w/ these changes: https://github.com/ichard26/diff-shades/compare/stable...yilei:diff-shades:sleep
-
A successful run used ~11.5% of ~15.6GB ram in the end
>>>> psutil.virtual_memory()=svmem(total=16757780480, available=14827339776, percent=11.5, used=1543008256, free=11242786816, active=1505480704, inactive=3320844288, buffers=217481216, cached=3754504192, shared=24207360, slab=519602176): https://github.com/yilei/black/actions/runs/7067923208/job/19241889065?pr=1 -
A failed run used up all ram before it was killed:
>>>> psutil.virtual_memory()=svmem(total=16757780480, available=24997888, percent=99.9, used=16486572032, free=159674368, active=11400617984, inactive=4878725120, buffers=360448, cached=111173632, shared=9478144, slab=147853312)https://github.com/yilei/black/actions/runs/7068262208/job/19242647307?pr=1
So there appears to be a memory leak but only sometime
The failed run's memory usage keeps growing overtime when it formats each file (not strictly though as sometime it decreases just a little bit)
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
so far the outcome seems to be that we got more friends reporting useful feedback
plain atlas
I don't see a reason to stop using black any time soon
black is easier for users to contribute to (because the users using a python formatter are more likely to know python than rust)
ionic mica
so far the outcome seems to be that we got more friends reporting useful feedbac...
Kind of like Biome and Prettier I'm assuming
in the JS world
rugged sparrow
could i get someone to look at this?
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
could i get reviews on my pull requests please? [#4072](https://github.com/psf/b...
reviewed the first, but now there's a merge conflict in the second
rugged sparrow
reviewed the first, but now there's a merge conflict in the second
fixed
slow ibex
i am using construct library: ```py
Var = c.Struct(
'name' / WStr,
'type' / c.Enum(
u8,
unknown=0,
int=1,
dword=2,
float=3,
str=4,
array=9,
),
'value' / c.Switch(
c.this.type,
{
'unknown': c.Computed(None),
'int': i32,
'dword': u32,
'float': f64,
'str': WStr,
'array': List(u32),
},
),
)
this gets formatted to this: ```py
Var = c.Struct(
'name' / WStr,
'type'
/ c.Enum(
u8,
unknown=0,
int=1,
dword=2,
float=3,
str=4,
array=9,
),
'value'
/ c.Switch(
c.this.type,
{
'unknown': c.Computed(None),
'int': i32,
'dword': u32,
'float': f64,
'str': WStr,
'array': List(u32),
},
),
)
in particular, this: ```py
'value' / c.Switch(
...
),
becomes this: ```py
'value'
/ c.Switch(
...
),
that hurts readability
is it intended?
i understand that it makes sense to split it into several lines if expression after / is long
in this case it is indeed long, but it is already split into several lines and first line is not that long
slow ibex
can i gently ping someone here to get response? π
dense jungle
can i gently ping someone here to get response? π
some thoughts in no particular order:
have you tried this in preview style? I think we made some related changes
it's sort of an uncommon pattern (an operator with a function call on the RHS), so we probably haven't thought about it too much
I don't think it's an unreasonable style though, it follows the operator precedence
slow ibex
have you tried this in preview style? I think we made some related changes
just tried, nothing changes
slow ibex
I don't think it's an unreasonable style though, it follows the operator precede...
should i open an issue?
dense jungle
should i open an issue?
if you like. I looked through https://github.com/psf/black/issues?q=is%3Aissue+is%3Aopen+label%3A"F%3A+linebreak"+operator and didn't see any existing issues that look like this
slow ibex
i dont really know how to describe it
dense jungle
something about line breaks in binary operations with a function call on the right
slow ibex
psf/black#4092
toxic stormBOT
slow ibex
Not sure about the title. Feel free to edit it
dense jungle
(was just testing that the hook works)
dense jungle
I think I'll do a stable release 23.12.0 and an alpha release 24.1a1 with the proposed 2024 stable style today, cc @elder tusk
flat krakenBOT
plain atlas
Prettier failed
dense jungle
ahead of you π
flat krakenBOT
dense jungle
two releases for the price of one
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
Python syntax has changed? Or it hasn't? This line was not flagged in the past.
It would be nice to add a bit more info to that error message (e.g. a link or at least a code what exactly is violated). This way the message is more annoying than helpful.
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
late dewBOT
:incoming_envelope: :ok_hand: applied timeout to @elder bay until <t:1703057903:f> (10 minutes) (reason: duplicates spam - sent 4 duplicate messages).
The <@&831776746206265384> have been alerted for review.
valid halo
Hello, the --preview feature seems to not working for multi line """ strings. as a result the indentation gets messed up if there is a large piece of text (prompts) in the code. Is there anything I can do about it?
neon loom
Got a pastebin or a small repo with black playground to share?
flat krakenBOT
flat krakenBOT
silent knoll
Hey, could I get https://github.com/psf/black/pull/4109 looked at? Ideally this should probably be merged soon to get in the 2024 stable style
flat krakenBOT
flat krakenBOT
neon loom
Is there a quicker way than PR'ing this commit? I see I can no longer pushing directly to main
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
warm crow
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
elder tusk
what's next for 2024 style? should we make a2?
i think overall things are shaping up nicely.
we should leave out string_processing and hug_parens_with_braces_and_square_brackets. we should also probably leave out wrap_long_dict_values_in_parens (it has three open issues, including crash)
flat krakenBOT
flat krakenBOT
dense jungle
what's next for 2024 style? should we make a2?
i think overall things are shapin...
i've been procrastinating this a bit but another alpha sounds good, or maybe we can just finalize it
i'll look more next week
wispy creek
Hi! Why the last enum case is not formated by Black?
class ResultType(IntEnum):
OK = (0,),
NOT_FOUND = (1,),
ANOTHER_ERROR = 2
glacial drum
Hi! Why the last enum case is not formated by Black?
```
class ResultType(IntEn...
what would you expect it to be formatted as?
wispy creek
what would you expect it to be formatted as?
All items with the same formatting
class ResultType(IntEnum):
OK = (0,),
NOT_FOUND = (1,),
ANOTHER_ERROR = (2,)
2 and (2,) are not the same, so that would change the meaning
the first is an int, the second is a tuple with an int in it
I am guessing you actually mean the code to be this, right? ```py
class ResultType(IntEnum):
OK = 0
NOT_FOUND = 1
ANOTHER_ERROR = 2
At first I had your code, but Black converted it to this:
class ResultType(IntEnum):
OK = (0,),
NOT_FOUND = (1,),
ANOTHER_ERROR = 2
It sounds like maybe you thought you had this code:
class ResultType(IntEnum) :
OK = 0
NOT_FOUND = 1
ANOTHER_ERROR = 2
But actually you had this, which means something different (the trailing comma creates a single-element tuple when it occurs on the right-hand side of an assignment) :
class ResultType(IntEnum) :
OK = 0,
NOT_FOUND = 1,
ANOTHER_ERROR = 2
hmm it's possible, thanks!!
flat krakenBOT
flat krakenBOT
flat krakenBOT
Using Black Formatter extension on vs code which uses v23.10.1
for some reason I get no line formatting by default (of 79 characters) however things work when I declare line-length and any value (including the default of 79)
I feel like this should be an easy bug to fix (by default just declare it with 79) not. But hopefully it will help those that got stuck with this issue https://github.com/...
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
azure plank
Messing around with preview, when formatting this py for _ in []: for _ in []: for _ in []: ValueError(f"first part {'long long long long long long long long long long contents'}")
it becomes this playground link py for _ in []: for _ in []: for _ in []: ValueError( "first part" f" {'long long long long long long long long long long contents'}" )
Why is this? I would expect it to be formatted like this py "first part " f"{'long long long long long long long long long long contents'}"
instead.
elder tusk
my guess would be it's because it's a pretty common bug for people to forget the trailing space when using implicit string concatenation. having the space be leading makes it obvious. i think this is the issue where this was decided, if you want to read through the history: https://github.com/psf/black/issues/182#issuecomment-552132608
(but i agree it's not an open and shut case, when manually doing implicit string concatenation i do it the other way. if you split things differently, black won't resplit unless you're over line length)
(also note that this feature will not make it into the 2024 stable style)
flat krakenBOT
clever peak
encountering a bug.
import asyncio
from telebot import asyncio_filters
from telebot.async_telebot import AsyncTeleBot
from telebot.asyncio_handler_backends import State, StatesGroup
from telebot.asyncio_storage import StateMemoryStorage
from telebot.types import KeyboardButton, Message, ReplyKeyboardMarkup
import logging
state_storage = StateMemoryStorage()
bot = AsyncTeleBot("", state_storage=state_storage)
logger = logging.getLogger(__name__)
def contains(text: str):
def wrapper(message: Message) -> bool:
return text in message.text
return wrapper
class Database:
def get_balance(message: Message) -> int:
return 3
def user_exists_or_create(message: Message):
return 3
@bot.message_handler(func=contains("Balance"))
async def balance(message: Message):
await bot.send_message(message.chat.id, "Please wait while we process your request")
shop_database.user_exists_or_create()
await bot.send_message(
message.chat.id, f"Your balance is ${shop_database.get_balance()}", reply_markup=ReplyKeyboardMarkup(resize_keyboard=True).add(KeyboardButton("π back")
))
return
if __name__ == "__main__":
shop_database = Database()
bot.add_custom_filter(asyncio_filters.StateFilter(bot))
asyncio.run(bot.polling(non_stop=False), debug=True)```when i format this with black
await bot.send_message(
message.chat.id, f"Your balance is ${shop_database.get_balance()}", reply_markup=ReplyKeyboardMarkup(resize_keyboard=True).add(KeyboardButton("π back")
))``` get's converted to
```py
await bot.send_message(message.chat.id, f"Your balance is ${shop_database.get_balance()}", reply_markup=ReplyKeyboardMarkup(resize_keyboard=True).add(KeyboardButton("π back"))
which is missing an closing ), sorry for the noise, this was what i was able to distill it to so that i could repro reliable, let me know if something like this has been reported before
misty trench
hi all! New around, sorry if my questions is obvious. Is there an approximate schedule for stable 2024?
dense jungle
hi all! New around, sorry if my questions is obvious. Is there an approximate sc...
by the end of the month!
flat krakenBOT
dense jungle
encountering a bug.
```py
import asyncio
from telebot import asyncio_filters
fro...
That doesn't sound like a bug that Black would have. You may be getting confused with some other tool
clever peak
That doesn't sound like a bug that Black would have. You may be getting confused...
i'm using the black formatter on VSC
which thinking about it means i should be opening an issue on MS black extension repo maybe
lunar mulch
This is probably a better channel to ask in! Is there an ETA for when Black v24.0.0 will be released? I am in the middle of a repository cleanup and would LOVE to include the new preview style, but I need it to be stable before I merge π
Ah, I see above end of the month. Great work, and I am looking forward to it!
flat krakenBOT
azure plank
How did I miss dummy_implementations, I 100% agree with hauntsaninja that it's a massive qol improvement
flat krakenBOT
Describe the solution you'd like
json schema can let editor which support LSP to complete some configuration file. Such as:
There are many tools have supported json schema:
Can [tool.black](https:...
rare kernel
Black formatter format like this
Any way to remove the remove the red squiggly line?
bright glacier
Black doesn't really format like that, do you have line wrapping enabled in VSC?
rare kernel
I havent touched any of the settings here
bright glacier
well, black doesn't format like that so I doubt it's black's fault
what's the error indicated the squiggly line anyway?
rare kernel
bright glacier
Right. Black can't fix that.
rare kernel
Can I disable that type of formatting
bright glacier
Black is not causing that.
Black won't even be able to format your code as that's invalid syntax (and will fail to parse).
bright glacier
Can I disable that type of formatting
Do you have a sample of code to share that reproduces this issue? Ideally with https://black.now.sh
rare kernel
Thats the formatted code
If I put it into one line like this
st.set_page_config(page_title="Results", page_icon=":chart_with_upwards_trend:", initial_sidebar_state="collapsed")
black changes it into
st.set_page_config(
page_title="Results", page_icon="π", initial_sidebar_state="collapsed
")
In here it format like this
st.set_page_config(
page_title="Results",
page_icon=":chart_with_upwards_trend:",
initial_sidebar_state="collapsed",
)
Why isnt it happenong here
rare kernel
nvm this
st.set_page_config(
page_title="Results", page_icon="π", initial_sidebar_state="collapsed"
)
I have format on save
in vscode
bright glacier
well yeah, but we're trying to figure out why black.now.sh isn't reproducing it
removing vscode as a variable is part of that process
rare kernel
how do i format without vscode
bright glacier
pip install black and run python with -m black <path-to-your-file>
actual commands will vary on how you use pip and python on your system
rare kernel
st.set_page_config(
page_title="Results", page_icon="π", initial_sidebar_state="collapsed"
)
Format like this
IDK why vscode black put the " in a new line
It also dont go over the 80ch col limit
wiht the "
bright glacier
Are you using a different version of Black under VSCode? That's the last variable on black's end that could cause this (well other than any black configuration that may be present).
I'm still not convinced this is black's fault, although it seems strange for vscode as well.
bright glacier
maybe, but it's a pretty strange bug
rare kernel
This is the version I have in pip
Anyways vscode black is happy with the way pip black formatted it
for now ill use that
bright glacier
I'm sorry that you're dealing with this, but I'm at a loss on how this could happen. I would take a closer look if I had time, but I don't.
Perhaps you could reach out to the VSCode Python folks and see if they have any ideas?
rare kernel
nah its all good, not big deal
I'll create a issue in there github repo later
their*
flat krakenBOT
orchid crow
Somewhat related, does anyone know how to use the pylsp plugin for black? https://github.com/python-lsp/python-lsp-black
I know it is not maintained by Black's dev
flat krakenBOT
dense jungle
We're getting close to finalizing the stable style! Would appreciate reviews on https://github.com/psf/black/pull/4096 which also adds the new --unstable style and documentation
flat krakenBOT
flat krakenBOT
unborn heart
Hello, I have submitted a PR (https://github.com/psf/black/actions/runs/7635201817/job/20800232618?pr=4165) and have this failing action. Can anyone offer some guidance as to going about fixing this?
dense jungle
Hello, I have submitted a PR (https://github.com/psf/black/actions/runs/76352018...
unfortunately it's been frequently failing, it's not related to your PR
sometimes if I retry enough times it works
I haven't had a chance to investigate deeper
seems related to memory usage or something
i guess for your PR it's pretty clear it won't affect formatting anyway
unborn heart
Got it, thank you.
dense jungle
also just commented on the issue
flat krakenBOT
azure plank
π
wintry grove
is it OK to ask about the ruff formatter here?
silent apex
is it OK to ask about the ruff formatter here?
This channel is specifically for collaboration in the Black project, but you could ask in #tools-and-devops
dense jungle
is it OK to ask about the ruff formatter here?
if you want to, but it's not a great place. I think Ruff has its own discord server
wintry grove
ok thank you, appreciate the response and the direction
elder tusk
ruff formatter bases its style choices on black, so this is a reasonable place to discuss that. but e.g. if you want help getting started using ruff formatter or something, ruff discord is the way to go!
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
time for 24.1
flat krakenBOT
flat krakenBOT
bright glacier
congrats!
dense jungle
the wheels aren't built yet π
bright glacier
are you blaming me for not making mypyc fast enough π
dense jungle
it looks like the actions are queued, so I think I should blame GitHub for not giving us enough compute resources
(but of course I'm grateful they give us CI infrastructure for free)
bright glacier
imagine the productivity increase if we didn't have to wait for CI, would be truly revolutionary
but yes, GHA is great as-is, I'm just poking fun :P
dense jungle
https://pypi.org/project/black/24.1.0/#files we got our wheels
elder tusk
hooray!
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
welp seems like this new cache filename doesn't work on windows, I guess I can't use | in a path?
ok @ worked
fun fact: if you set your max line length to an 80-digit number, Black must always have been broken on eCryptfs. I wonder why nobody reported that
flat krakenBOT
flat krakenBOT
drowsy jolt
fun fact: if you set your max line length to an 80-digit number, Black must alwa...
some PyPI stats from henryiii in the [PyPA server](#1128441623974531072 message), here's the most popular tool.black.line-length values:
All the ones over 100:
120: 5755
88: 3578
79: 2876
100: 2714
80: 834
99: 811
119: 485
110: 359
90: 269
95: 169
140: 165
160: 112
(Less than that, then I probably should start including strings & the other way(s?) to specify it like line_length) Total is 19,076 pyproject.tomls that have tool.black
dense jungle
some PyPI stats from henryiii in the [PyPA server](https://discord.com/channels/...
Thanks, that's really interesting! Surprised 80 isn't more common
flat krakenBOT
flat krakenBOT
lunar mulch
Excellent! I just converted over a moderately complex project with positive results. I was, however, a bit surprised that the vast majority of changes were blank line additions and deletions π€£ (upgrading from v22)
Overall, I like the formatting changes quite a bit, nice work!
(also... 120 line-length gets my vote every time! π€)
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
placid oasis
Hi Everyone I am looking at documentation to try contribute to the project. https://black.readthedocs.io/en/latest/contributing/the_basics.html
I have experience with codding, pull requests, debugging, PyCharm and etc. but new to Open source projects
azure plank
I was reading through the black issues, and I wonder if it would be worth adding a C: unstable style issue label since things like string_processing have been moved to unstable, or if it would be too much work swapping between preview and unstable tags as features get moved.
dense jungle
I was reading through the black issues, and I wonder if it would be worth adding...
Good point, I do think it might be annoying to maintain as we move features between preview and unstable. Maybe instead we should set a guideline to put the responsible preview feature in the issue title so it's easy to track
azure plank
Good point, I do think it might be annoying to maintain as we move features betw...
Sounds reasonable to me. While from what I've seen it's not common, it's also not unprecedented to have the broken feature name in the title. I suppose the only other consideration would be on issues where the cause isn't initially known. From the perspective of a reader, shuffling tags on issues seems like a less substantial action than editing the title, but I'm not sure what other people's take on that would be/what the general culture around it is. I'm also not familiar enough with github from a maintainer's perspective to know if there are different perms for adding tags vs editing issue titles that might need to be changed.
dense jungle
Sounds reasonable to me. While from what I've seen it's not common, it's also no...
Both edits can only be done by triagers or maintainers, but putting the feature name in the title has the advantage that it shouldn't change over time
dense jungle
Sounds reasonable to me. While from what I've seen it's not common, it's also no...
I added the name of the relevant feature to all open preview style issues, and changed the label description to say the feature name should be added
I forgot there is an open crash with the hug_parens change (https://github.com/psf/black/issues/4036), that means we should move it to unstable
Also you're right that it might not be clear initially which preview feature is responsible, but we can always change the title later
azure plank
Looks good. My one complaint would be I vastly prefer the look of the feature name being in a code block, but it's mostly just a nitpick.
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
Describe the bug
After upgrading to the latest version, when I rub python -m black ., I got files reformatted:
Then, when I commit, it against reformat the files restoring them:
Irrational beha...
flat krakenBOT
flat krakenBOT