Dang, I've been looking for some low-hanging doc or mypy-type fruit for a first PR and y'all have done a real good job keeping this project nice and clean. :''']
#black-formatter
swift fossil
bright glacier
I never feel like the project is well-oiled as a maintainer lol :p
There are so many infra and process improvements possible!
errant barn
ππΏ I'll try to have a look on friday if it's still TODO, unfortunately no time before that
swift fossil
Yeah! I gott'a get a feel for how all'y'all do your thing and work from there. I'll try to read through the code a bit more or something. I'll look for somethin' light in the issues list to hit at.
bright glacier
CI is complicated to write, ah. Also pretty annoying to test lol
Hmm is there any benefit to restricting permissions to only contents: read in the workflows themselves if in the repository actions settings the default permissions are exactly that?
because they are indeed set to contents: read only by default
flat krakenBOT
errant barn
lament crow
that's done when it doesn't fit the line length limit
so if this is inside deeply indented code then that would happen
flat krakenBOT
lament crow
lol
errant barn
yep :D
mint barn
π even i need to add myself to authors
lament crow
it does also say "Multiple contributions by:" so even if their PR was merged, that's still not multiple!
lament crow
who are you to be the judge of what's a contribution! /s
fast heart
π π
errant barn
that you? :D
errant barn
oh nevermind then π
dense jungle
another episode in "Black gets blamed for things it didn't do" π
flat krakenBOT
swift fossil
Is there any desire for Black to be compatible with the VSCode Dev Containers extension?
The tl;dr is that it makes a nice little docker env that lets VSCode devs work inside of that while dev-ing. I haven't seen it in too many Python packages besides Pandas (https://github.com/pandas-dev/pandas/blob/main/.devcontainer.json) but there's a number of smaller CLI-type packages I've used that have it.
The only con I can think of is that it's an additional file in the root, which, admittedly, is kind'a gross.
(Secretly, I'm just trying to figure out something to mess with, and my day job is a lot of ops-type stuff. Not a biggie if y'all aren't into it.)
bright glacier
I'm a little confused, is this file meant to preconfigure an GitHub Codespaces environment? ... because I'm not even sure if the psf organisation has access.
Or is this also applicable to local development in VSCode?
Ah it's the latter reading through the docs
swift fossil
Ah, apologies for not bein' more specific: yeah, it's the latter thing. Local dev in VSCode. I don't think I've ever done Codespaces stuff, so I don't know much about that.
flat krakenBOT
fallow pine
is there documentation on all of the possible config options?
aside from black --help I can't seem to find anything
flat krakenBOT
lament crow
You can change line length limit to your liking.
You can't change how it splits once you go over it
plain atlas
# fmt: off
nasty code
# fmt: on
red sedge
is there an issue for black magic comma in context managers? I couldn't see one with a search https://github.com/psf/black/issues?q=magic+comma+context+manager
is it ever correct for black to emit ,) when not in a tuple?
dense jungle
normalization is what Black does. Before you run Black code is not normalized, afterwards it is
--skip-string-normalization skips the things Black does to strings (change quote style and mess with the prefix)
bright glacier
s = 'single quotes will not be changed to double quotes'
s = u"this redundent u prefix also won't be removed"
oh and also the prefix won't be reordered (can't remember what the current order is though)
bright glacier
is there an issue for black magic comma in context managers? I couldn't see one...
oof yeah that is definitely a bug
context managers in general could do with a style overhaul since it sucks right now
bright glacier
Yeah, but I am not sure if black handles those parenthesized ctxmgrs well either
red sedge
is it ever correct for black to emit `,)` when not in a tuple?
I'm sure I raised an issue about this somewhere
I wish black.now.sh had a bug report button
red sedge
omg my wish has been granted
bright glacier
yup
civic shuttle
Hello i would like to share my project feel free to download it, but please credit
lament crow
Hello i would like to share my project feel free to download it, but please cred...
this isn't a channel for sharing your projects
civic shuttle
i did not know
bright glacier
Are there any (hopefully low effort) PRs people want to see merged before a release? If not I'll just get Felix's PR in and start the release checklist for v22.6
I looked at the changelog and I don't think there's anything that was left half-finished user-facing wise so I think there's nothing truly blocking the release. Just a matter of me clicking buttons :)
This release is rather light especially given it's been three months, but that's just the byproduct of my lack of activity on the project. I should have more free time now though so that'll change :)
oh hi @errant barn FYI I am editing your PR locally, I'll push my changes soon. You can lemme know if they look good!
errant barn
oh hi <@273104906083106816> FYI I am editing your PR locally, I'll push my chang...
got a flow idea? nice
bright glacier
yeah, turns out sleep helps :)
errant barn
often :D
bright glacier
Pushed, please take a look :)
I wonder if this will work as a hacky preview :p
dammit, you weren't supposed to yoink the code within the file and preview it discord.
ah all of the CSS is broken, that's unfortunate
oh, apparently saving a whole webpage also downloads a whole folder with the JS and CSS assets... that's annoying. I would've thought Chrome would put everything in a single archive
errant barn
Nice, that looks good
bright glacier
Kk thank you! Not sure if I should wait for Jelle's review?
errant barn
I think we're fine, but if you'd like!
bright glacier
lol, Rich is always going to be my unofficial nickname, everyone thinks of it :p
bright glacier
progress! (the release CI item is deferred for the release after this one)
if anyone wants to help this user, please do, currently busy with release prep. https://github.com/psf/black/issues/1985#issuecomment-1167837098
bright glacier
Not really sure who is active right now so I just guessed for my review requests :p
flat krakenBOT
bright glacier
I'mma go for a walk/run now. When I come back (in ~3 hours) I'll cut the release.
austere lava
look look I contribute
flat krakenBOT
bright glacier
I'm a little nervous the mypyc wheels might fail to build, but the diff-shades workflow does mean the linux wheels should build without issue so that makes me hopeful.
http://mirror.es.its.nyu.edu/epel/7/x86_64/repodata/repomd.xml: [Errno 12] Timeout on http://mirror.es.its.nyu.edu/epel/7/x86_64/repodata/repomd.xml: (28, 'Connection timed out after 30001 milliseconds')
Trying other mirror.
grrrr, now I am properly nervous
eh, seems to building fine, just a hiccup at the start
flat krakenBOT
bright glacier
.<
orchid lichen
whatβs black formatter?
bright glacier
An autoformatter for Python code
orchid lichen
oh cool
bright glacier
!remind 30M check in on the 3rd rerun of the mypyc wheel build workflow, sigh
late dewBOT
late dewBOT
!remind 30M check in on the 3rd rerun of the mypyc wheel build workflow, *sigh*
It has arrived!
bright glacier
Irrelevant, Jelle helped me sort out my zip troubles.
bright glacier
We should make up our minds on psf/black#751
bright glacier
If we want to support reading python-requires in pyproject.toml then making --target-version the minimum is only option that makes sense. And I'd like to support reading python-requires so I vote for that.
dense jungle
we should make it the minimum
bright glacier
@neon loom I think you misunderstood psf/black#3124, it's asking that Black should read [project.python-requires] to automatically set --target-version AFAICT (based off the comment they linked)
toxic stormBOT
neon loom
Ya, it did. I also want to use hatch + pyproject.toml as much as we can - If it's out of place feel free to delete
Good chance my quick 20 PR / Issue read caused me to misunderstand
little parrot
anyone ever seen this error?
error: cannot format docs/conf.py: Can't pickle <enum 'COLOR_DIFF'>: attribute lookup COLOR_DIFF on black failed
...
Oh no! π₯ π π₯
10 files failed to reformat.```this is black 22.6.0 and python 3.10.4 if it helps
i cant seem to get a more verbose traceback even with --verbose
dense jungle
anyone ever seen this error?
```
error: cannot format docs/conf.py: Can't pickle...
weird, is your environment clean?
could happen if you manually changed enum.py or something
bright glacier
!remind 18H test out setuptools' experimental PEP 660 support
late dewBOT
pallid dove
Black seems to be handling comments in a way that i don't quite want. So i got this:
# @app.route("/shutdown", methods=["POST"])
# def shutdown():
# shutdown_server()
# return "Server shutting down..."
After formatting, i get:
# @app.route("/shutdown", methods=["POST"])
# def shutdown():
# shutdown_server()
# return "Server shutting down..."
I get why that happens but I don't want it to do that. I could be tripping but this didn't work like this in previous versions. I checked configurations and there doesn't seem to be a way to change it.
errant barn
We don't really consider comments to have any special status wrt formatting, although it is annoying that commented code will break
As a workaround you could use # fmt: off/on or encapsulate the code in a multiline string
errant barn
Black seems to be handling comments in a way that i don't quite want. So i got t...
You would perhaps agree that
# any old comments
# unaligned
# should be put back
huh? i'm not sure i understand
errant barn
To preserve commented out code like you presented would be bad, is what I'm trying to say. It's probably hard to determine that a comment contains code (and it is intended as such), and we don't want to take previous formatting into account in any of our formatting.
IIRC this probably wasn't changed in the latests release π€
errant barn
As a workaround you could use `# fmt: off/on` or encapsulate the code in a multi...
@pallid dove So if you were just after "disabling" it, this is probably it 
pallid dove
i suppose that'll work. thanks!
no idea why but my discord is being weird and i can't see any new messages except for when going into the search bar
errant barn
i suppose that'll work. thanks!
no idea why but my discord is being weird and i ...
I must religiously share the teachings of Uncle Bob, though :D "don't check in commented code" https://youtu.be/2a_ytyt9sf8?t=2263
pallid dove
Interesting. I guess he does make a good point.
flat krakenBOT
mint barn
damnn felix u are fast
flat krakenBOT
errant barn
;)
flat krakenBOT
late dewBOT
!remind 18H test out setuptools' experimental PEP 660 support
It has arrived!
bright glacier
man, do I hate past me's decisions how to structure the documentation
it's really awful
I think the fact psf/black is the first major project I've ever managed or maintained (my personal ones don't count) shows 
learning on the job is totally fine, but it's unfortunate that some mistakes are hard to undo, gah
lament crow
do you guys even need stable branch anymore?
since now Black is no longer beta
and it should be picked up properly
bright glacier
I think some integrations reference it
lament crow
π
bright glacier
I know our old pre-commit sample configuration did (which was a very very very bad idea)
we're kinda stuck with it Β―_(γ)_/Β―
bright glacier
do you guys even need `stable` branch anymore?
it's a tag right now btw, which I dislike
lament crow
yeah, branch is more flexible
you have to force push tags, right
I mean, you probably delete it and readd
bright glacier
nah git tag -f stable 22.6.0 is enough
bright glacier
yeah, branch is more flexible
mhmm, and pulling an updated tag is really annoying since git assumes they won't change
I think you have to delete your local tag before being able to pull an updated tag from a remote
bright glacier
according to some folks in the core team, it's safer cause it's harder to accidentally change it (you need to pass --force)
I disagree that's important enough
lament crow
branch protection with deployment requirement perhaps?
bright glacier
yup
but I think there's some pushback on that front too since it's added complexity for not much gain
bright glacier
yeah, maybe I'll suggest it, but I don't like suggesting ideas that might get shot down (this is a personal thing, nothing to do with the core team or any of its members)
lament crow
the whole stable branch thing is an added complexity, time to drop in (20)23? :P
bright glacier
and break a whole wack of old pre-commit configurations 
silent apex
are you not on the core team of psf/black?
bright glacier
I am part of the core team, but I don't feel like I have much authority when it comes to policy.
silent apex
ah
lament crow
whether it's tag or branch, I don't see why it shouldn't be automated
I assume it hasn't be done because someone would have to actually do it since that's the standard problem of OSS
or I guess development in general
bright glacier
nah, it's hard to automate it right now since the policy is to only update it once we're confident the release isn't broken
lament crow
having to trigger a separate workflow would still be less work than doing it manually
and it standardizes the process
bright glacier
yeah I suppose, and if we can't fix the weird issue with RTD picking up the old commit of an updated tag, we could just fire a rebuild request in the workflow :p
lament crow
guess this needs a reproducer
bright glacier
https://readthedocs.org/projects/black/builds/17325177/ .. check the commit built here
and this is the rebuild I forced https://readthedocs.org/projects/black/builds/17325383/
lament crow
yeah, I'm just thinking of making a separate project where we can freely test that out without playing with an important repo lol
maybe I'm volunteering
just, tomorrow...
bright glacier
I mean if you don't get to it, I'll get to it eventually
although I have an insanely long to-do list so that might as well be effectively never
lament crow
I see you guys don't have PR builder enabled on RTD
bright glacier
I think the last time I tried enabling that it didn't work since I don't have admin permissions to the GH repo Β―_(γ)_/Β―
lament crow
you have RTD access?
bright glacier
yup, I tried enabling PR preview in the RTD admin panel
lament crow
I suppose if the webhooks are outdated it might require admin perms on GH repo
which... is possible since this project has existed for a while
dense jungle
I have admin access if you need anything
lament crow
There's a switch for PR builder here:
https://readthedocs.org/dashboard/black/advanced/
Which in itself shouldn't require admin access on GH I think but it's possible you run into one of these issues in which case that access would be needed:
https://docs.readthedocs.io/en/stable/pull-requests.html#troubleshooting
bright glacier
I can try again right now and see what error it gives me
lament crow
half of it works at least
lament crow
it might require a reconnect
which the UI seems to make it seems as if it's more disruptive than it actually is (I think, I know I did that once but I can't recall if that's the actual place lmao)
bright glacier
well I made my RTD account using my GH account so I'll probably have to set up a password first
bright glacier
lolwat
lament crow
you click disconnect, it disconnects and then you can connect again
I have no clue what would happen if you were to log out
bright glacier
*phones support*
oh they fixed it
you can longer shoot yourself in the foot π
lament crow
I guess my account does have password set up then
I didn't know
holy, I even have it saved in password manager, when did I do that...
bright glacier
haha :p
flat krakenBOT
flat krakenBOT
bright glacier
@mint barn hey, can you clarify whether you want me to take over the PR psf/black#2916 for ya? the way you worded your reply is a bit unclear, and while it sounds you are going to come back to the PR, you haven't pushed any updates (to resolve the outstanding review comments)
toxic stormBOT
mint barn
<@456226577798135808> hey, can you clarify whether you want me to take over the ...
yeah u can take over, the second part meant that when u are iterating over my code, if something is unclear let me know
bright glacier
Sounds good, thanks for clarifying!
bright glacier
I think I managed to fix mypyc/mypyc#917 π
toxic stormBOT
bright glacier
The root issue is that mypyc didn't know builtins.set was a built-in name, so it guessed it came from the module globals. This didn't blow up anything up somehow... until the dataclasses commit which made the __annotations__ logic for dataclasses try to better preserve the type annotations (previously they would be erased to type)
flat krakenBOT
Description
The Actions Status badge (the one that displays Test: passing / Test: failing) had a different left side color from all the other badges. With this commit, a badge from the shields.io website is used, standardising the badge styles.
Comparing the change:
Checklist - did you ...
- [x] Add a CHANGELOG entry if necessary?
- [x] Add / ...
flat krakenBOT
bright glacier
Good catch on frozenset's missing load_address operation @dense jungle , I added it but I can't seem to write a test for it..?
---------------------------------------------------------------- Captured stdout call ----------------------------------------------------------------
run-python37.test:6: error: Module "typing" has no attribute "FrozenSet"
run-python37.test:74: error: Name "frozenset" is not defined
other.py:49: error: Name "frozenset" is not defined
============================================================== short test summary info ===============================================================
FAILED mypyc/test/test_run.py::TestRun::run-python37.test::testRunDataclass
Does mypy[c]'s test suite run against a custom typing and builtins or something?
yup, mypyc does use its own typing.pyi and builtins.pyi >.<
dense jungle
Good catch on frozenset's missing load_address operation <@783088578363523104> ,...
every mysterious problem with mypy tests can be tracked back to the fixtures
bright glacier
I updated the PR. PTAL
dense jungle
including this one it sounds like π
bright glacier
I just stole a bunch of stubs from typeshed to make frozenset work
I think we'll be able to use the latest mypy[c] once this fix is merged, unless another regression is awaiting us
dense jungle
I updated the PR. PTAL
weird that the __annotations__ turn into set instead of set[str]. does mypyc do that to all generics?
late dewBOT
mypyc/irbuild/classdef.py lines 149 to 157
def create_non_ext_info(self) -> NonExtClassInfo:
non_ext_bases = populate_non_ext_bases(self.builder, self.cdef)
non_ext_metaclass = find_non_ext_metaclass(self.builder, self.cdef, non_ext_bases)
non_ext_dict = setup_non_ext_dict(self.builder, self.cdef, non_ext_metaclass,
non_ext_bases)
# We populate __annotations__ for non-extension classes
# because dataclasses uses it to determine which attributes to compute on.
# TODO: Maybe generate more precise types for annotations
non_ext_anns = self.builder.call_c(dict_new_op, [], self.cdef.line)```There's a todo about that in the code actually :)
dense jungle
ah ok. I asked because you added a testcase doing this. If it's a general mypyc behavior it's fine, but if set behaves worse than e.g. list in this regard we should fix it in your PR
sounds like it's not specific to set/frozenset, so we can save this problem for another day
bright glacier
wait a sec, this code is crashing at runtime even with the patch applied
from typing import Set, Type, TypeVar
T = TypeVar("T")
def wat(a: Type[T]) -> Type[T]:
return a
@wat
class NonExt:
s: Set[str] = {"a", "b", "c"}
does that work with a different annotation?
I imagine mypyc's support for class decorators is probably patchy in general
bright glacier
nah, I just forgot to activate my virtual environment
β― python -i -c "import test"
>>> test.NonExt.__annotations__
{'s': <class 'set'>}
seems like all non-extensions classes (which include dataclasses AFAIK?) have a less precise annotations
β― python -i -c "import test"
>>> test.Ext.__annotations__
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
AttributeError: type object 'Ext' has no attribute '__annotations__'
lol and apparently extension classes don't even have __annotations__
bright glacier
ah ok. I asked because you added a testcase doing this. If it's a general mypyc ...
To answer your question, yeah this is general mypyc behaviour, also applies to list for example
β― python -i -c "import test"
>>> test.NonExt.__annotations__
{'s': <class 'list'>}
Does mypy have a somewhat regular release schedule or are we going to be waiting for an indeterminate amount of time for the next release?
Oh and thank you for the review!
bright glacier
@dense jungle FYI I've been waiting for you to write documentation explaining an exception to the stability policy: https://github.com/psf/black/pull/2874#pullrequestreview-888107240
I'm alright with doing it myself, but you write better than me :)
oak forge
Hi, would someone here, preferably a PyCharm user, by any chance has a spare moment to review a PR with improvement of Black integration docs? Namely the use of BlackConnect plugin, which in my opinion is a production-ready and awesome π https://github.com/psf/black/pull/3150
flat krakenBOT
lament crow
re β
>>> import sys
>>> import subprocess
>>> sys.stderr.encoding
'utf-8'
>>> subprocess.run(["python", "-c", "print(__import__('sys').stderr.encoding)"], stderr=subprocess.PIPE)
cp1250
CompletedProcess(args=['python', '-c', "print(__import__('sys').stderr.encoding)"], returncode=0, stderr=b'')
For stdout to get properly encoded, sys.stdout/err.buffer.write(string.encode()) can be used
in case of click.echo, sys.stdout/err.buffer is used automatically if passing bytes to it rather than str
so doing click.echo(string.encode()) would do the trick
I'm not sure in how many places black uses unicode characters, it might be enough to alter the out() function to encode what it's passed
bright glacier
Pretty sure that'd break some environments where they patch STDOUT/ERR
lament crow
huh?
bright glacier
The buffer attribute isn't required, I suppose we could fallback to what we're doing right now if it's missing tho
lament crow
getting binary stream is abstracted by click
I don't know if it uses .buffer specifically
bright glacier
Ah yeah. I'll (maybe) look into it tomorrow. It's 12:08 over here π
flat krakenBOT
bright glacier
turns out mypyc was missing LoadAddress ops for quite a few built-ins, python/mypy#13078
toxic stormBOT
bright glacier
one IRbuild test is failing though, not sure why yet
ah
Expected:
...
return r0
def f(c):
c :: __main__.C
r0, r1 :: int
r2, r3, r4 :: object (diff)
r5 :: str (diff)
r6, r7 :: object (diff)
L0: (diff)
r0 = c.__neg__() (diff)
r1 = c.__invert__() (diff)
r2 = load_address PyLong_Type (diff)
r3 = PyObject_CallFunctionObjArgs(r2, c, 0) (diff)
r4 = builtins :: module (diff)
r5 = 'float' (diff)
r6 = CPyObject_GetAttr(r4, r5) (diff)
r7 = PyObject_CallFunctionObjArgs(r6, c, 0) (diff)
return 1 (diff)
Actual:
...
return r0
def f(c):
c :: __main__.C
r0, r1 :: int
r2, r3, r4, r5 :: object (diff)
L0: (diff)
r0 = c.__neg__() (diff)
r1 = c.__invert__() (diff)
r2 = load_address PyLong_Type (diff)
r3 = PyObject_CallFunctionObjArgs(r2, c, 0) (diff)
r4 = load_address PyFloat_Type (diff)
r5 = PyObject_CallFunctionObjArgs(r4, c, 0) (diff)
return 1 (diff)
it was expecting the old way of looking up builtins.float
bright glacier
ugh, diff-shades got cancelled on https://github.com/psf/black/pull/2874 because I pushed a changelog only commit
is it worth rerunning CI to run it?
the test workflow doesn't have the concurrency limit (yet!) and it seems to be passing on the merge commit
I'm wondering if it would be worth it in the end to rename T: design to T: code style or simply T: style
I get the feeling it's not obvious to issue openers their code style issue should use that label
bright glacier
the test workflow doesn't have the concurrency limit (yet!) and it seems to be p...
https://github.com/psf/black/runs/7219918686?check_suite_focus=true OK the xdist bug still exists apparently. I triggered a rerun... hopefully it passes!
foggy ivy
!e import random
prisoners = ([i for i in range(0,99)])
#creates list of prisoners numbered 0-99
shuffled_prisoners = random.sample(prisoners, len(prisoners))
#randomizes prisoner list
shuffled_list = random.sample(prisoners, len(prisoners))
#randomizes prisoner list to get the best possible random boxes
boxes = [(k, v) for k,v in enumerate(shuffled_list)]
#creates a linked list of tuples of random numbers
def main(prisoners, boxes):
"""Takes prisoner list and linked list representing boxes. Records and returns the final results"""
count = 0
final_freed_prisoners = 0
for prisoner in prisoners:
prisoner_num = prisoner
the_next_num = prisoner_num
print(f"Prisoner {prisoner_num}:\n--------Try")
outcome = prisoner_loop(prisoner_num, the_next_num)
count += 1
if outcome >= 1:
final_freed_prisoners += 1
else:
print(f"-------------XXX\n\n")
return print(f"\n*******{final_freed_prisoners} prisoners found their number********")
def prisoner_loop(prisoner_num, the_next_num):
"""Takes in prisoner number and the number 'inside' the box to map it to the next box, if prisoner finds box with his number returns 1 else 0"""
freed_prisoners = 0
for next in range(50):
next_num_tup = boxes[the_next_num]
the_next_num = next_num_tup[1]
print(next_num_tup, " -- ", next + 1)
if the_next_num == prisoner_num:
freed_prisoners += 1
print(f"------------!!!\n")
break
else:
continue
return freed_prisoners
if name != 'public static void main(String[] args)':
main(shuffled_prisoners, boxes)
late dewBOT
@foggy ivy :white_check_mark: Your eval job has completed with return code 0.
001 | Prisoner 26:
002 | --------Try
003 | (26, 49) -- 1
004 | (49, 46) -- 2
005 | (46, 1) -- 3
006 | (1, 65) -- 4
007 | (65, 23) -- 5
008 | (23, 97) -- 6
009 | (97, 37) -- 7
010 | (37, 39) -- 8
011 | (39, 0) -- 9
... (truncated - too many lines)
Full output: too long to upload
foggy ivy
oh crap ya there were a ton of lines i forgot...
any way to get the output to lineup better?
it's making me crazy!!!!!!!!!!!!!!!
oops sorry wrong room
solemn sequoia
anyone had trouble with black producing code that violates pycodestyle e125? I found this ticket: https://github.com/PyCQA/pycodestyle/issues/126
bright glacier
I don't remember seeing a ticket about E125 in the psf/black issue tracker, but perhaps I'm forgetting something.
red sedge
anyone had trouble with black producing code that violates pycodestyle e125? I f...
Do you have a demo input?
solemn sequoia
one sec
solemn sequoia
Do you have a demo input?
maybe it was a false alarm or I hadn't saved my file
slim olive
Black insists on the following transformation:
async def _get_session_id(
auth_header: str | None = Header(
None,
alias="Bearer",
description="Access token from the auth endpoint.",
),
) -> UUID:
...
... into:
async def _get_session_id(
auth_header: str
| None = Header(
None,
alias="Bearer",
description="Access token from the auth endpoint.",
),
) -> UUID:
...
Isn't that strange? I don't see a reason for the line break ~ the line's not too long
lament crow
that doesn't crash?
I wouldn't expect Python's syntax to be that forgiving to newlines in that place without parens
I guess it is, huh
either way, it is a strange wrapping behavior for sure
dense jungle
I wouldn't expect Python's syntax to be that forgiving to newlines in that place...
within parentheses you can put newlines wherever you want
lament crow
I just didn't think that would apply in the case of function def
dense jungle
agree this isn't great formatting though
dense jungle
I just didn't think that would apply in the case of function def
it's implemented in the tokenizer, so it's just wherever there are parentheses
dull agate
Are there any beginner-friendly open-source Python projects?
lyric ginkgo
i didnt want to open an issue for this cause i might be doing something wrong, but I can't get --extend-exclude to work correctly when using pre-commit
the command I ran on terminal was: black -l 99 --extend-exclude auvsi_suas ., which correctly excluded all python files in server/auvsi_suas
using a basic config on pre-commit doesn't work the same way, and reformats all files within server/auvsi_suas that need reformatting
repos:
- repo: https://github.com/psf/black
rev: 22.6.0
hooks:
- id: black
args: [-l, '99', --extend-exclude, auvsi_suas]
language_version: python3.10
am I specifying arguments incorrectly? I tried using strings on all arguments, joining them together, etc.
the -l line length parameter works as intended, but --extend-exclude does not
which leads me to think it's caused by pre-commit passing in file names differently...
any ideas?
i've even tried using args: [-l, '99', --extend-exclude, \s*auvsi_suas\s*] to match all characters in the filename, but it still doesn't seem to work in pre-commit
bright glacier
Are there any beginner-friendly open-source Python projects?
Ours for one! Although our contributing docs could be way better (matter of fact I'm rewriting them right now)
!contribute -- You could also try contributing to one of the Python Discord projects, they tend to be more beginner friendly than us.
late dewBOT
bright glacier
which leads me to think it's caused by `pre-commit` passing in file names differ...
Yea... the problem is that pre-commit passes the filepaths directly to hooks. Black ignores --include --exclude and --extend-exclude for filepaths given directly via the command line as they only really make sense for file discovery.
You can either use pre-commit's own exclude configuration, or if you just want to make Black work standalone and via pre-commit, you can use the --force-exclude option instead (I can't remember whether it clears the default excludes, but IIRC it doesn't, so it should be like --extend-exclude, but all of the time, no exceptions)
lyric ginkgo
And using that, I can just specify the folder name, and it processes that within the path?
e.g. specifying auvsi_suas for server/auvsi_suas/client/file.py
bright glacier
I think so
not entirely sure how pre-commit's exclude works, but that should work for Black's force-exclude
oh wait, that's a subdirectory, you may have to do server/auvsi_suas/ or .*/auvsi_suas/
dull agate
Ours for one! Although our contributing docs could be way better (matter of fact...
Where do i go to contribute
flat krakenBOT
flat krakenBOT
SGTM. Thank you for tackling this bug!
P.S. you need to add the closing keyword before each issue you'd like to close automagically
scarlet mortar
What is this channel
bright glacier
Hi @dense jungle, not sure what went wrong but can you close https://github.com/mypyc/mypyc/issues/924 and https://github.com/mypyc/mypyc/issues/935 please?
dense jungle
Hi <@783088578363523104>, not sure what went wrong but can you close https://git...
I lack that power
the mypyc repo is separate and I have no permissions there
I'm trying to get somebody to give me access
bright glacier
If anyone with async know how is looking for a (hopefully simple) issue to contribute a fix for, psf/black#3110 is it
toxic stormBOT
bright glacier
What is this channel
It's the real-time communication channel for black, the uncompromising code autoformatter for Python. See the channel description for more ^^
bright glacier
Where do i go to contribute
Sorry for taking so long to get back to you, but if you want to contribute to Black, I'd recommend 1) reading https://black.readthedocs.io/en/latest/contributing/the_basics.html and 2) looking at the issue tracker for any issue labelled with "good first issue"
... if that flies over your head, I suggest instead taking a look at the sir-lancebot Python Discord project which is a fair bit easier to contribute for beginners
flat krakenBOT
flat krakenBOT
late dewBOT
setup.py line 121
"Programming Language :: Python :: 3.6",```If you call new_event_loop rather than get_event_loop/asyncio.run then the event loop isn't set which means the ChildWatcher isn't configured
So you should use get_event_loop on 3.6 and asyncio.run on 3.7+
neon loom
All that effort why didnβt you just do the PR? LoL
I just hate all the ugly if version everywhere. With how we use asyncio is ChildWatcher even of benefit here
red sedge
All that effort why didnβt you just do the PR? LoL
Do the PR?
frigid scroll
Is it possible to configure black that way that it always places optional commas like here?
@app_commands.rename(
user="nutzer",
case="fall",
)
``` (The comma after `"fall"` is not needed)hm we can probably drop 3.6 now?
neon loom
I looked. We still have a lot of 3.6 downloads
I havenβt read what our stable policy states about dropping major version support.
neon loom
Is it possible to configure black that way that it always places optional commas...
The magic comma is my least liked part of black - But I get it
I just always add when developing
bright glacier
If you call new_event_loop rather than get_event_loop/asyncio.run then the event...
Good morning. What's even the benefit of ChildWatcher? I'm out of the loop (heh) on async stuff.
red sedge
It's if you're using asyncio.create_subprocess_exec etc
There's actually a bunch of stuff that asyncio.run does that new_event_loop doesn't, and it looks like you'll be able to get rid of a lot of the stuff in concurrency.cancel and concurrency.shutdown
bright glacier
I have written basically zero async code so I'll defer to Cooper (and @austere lava if he's around) for that
austere lava
I can take a look in a few
red sedge
Is there a plan to drop (running on) 3.6? I thought there was an issue about it, then you don't need the conditional for asyncio.run
lament crow
I know you don't follow semver in black but still I feel like these kind of changes should be deferred until 23.0 same as with style changes.
Not that I use 3.6 for anything myself though so I can't really advocate for keeping that alive in 2022 :P
mint barn
hi
bright glacier
Is there a plan to drop (running on) 3.6? I thought there was an issue about it,...
There's no official plans. I don't think there's even an issue about dropping 3.6 yet :p
austere lava
Well eventually our deps will stop supporting it π
bright glacier
Plenty of our deps already don't support it, I've been keeping track. Lemme pull up my list
late dewBOT
The snippet you tried to send was too long. Please see #bot-commands for the full snippet.
lament crow
As long as those deps properly declare minimum python version, pip will resolve it
Hopefully not in too long time
bright glacier
Yeah, as long as we don't bump the minimum requirements for these deps, we should be fine.
austere lava
If by fine you mean not immediately breaking, yeah. But bumping the minimum reqs is a good thing. Not being able to do it is bad
bright glacier
We don't have a reason to bump the minimums right now though, but you do have a point.
neon loom
I think 3.6 should go in 23.0 too. And we amend the stable policy to dictate our plans there if others agree.
Well would be 23.1
bright glacier
I'm not sure if the stability policy mentions Python support at all.
AFAIK the closest equivalent we have in the docs is this FAQ entry https://black.readthedocs.io/en/stable/faq.html#which-python-versions-does-black-support
dense jungle
the stability policy is explicitly only about formatting output, not anything else
Dropping 3.6 in the 23.0 release seems reasonable though
with Python's yearly cadence that means we can drop a version every year
red sedge
So a pedantic reading of that section says you could drop parsing 3.6 without a deprecation cycle, but can't drop running on 3.6 without a deprecation cycle
austere lava
I looked at the PR and seems fine to me. @red sedge's suggestions make sense, we can improve the code even more. IMO they can go in a separate PR. I'll comment
bright glacier
So re. psf/black#3141, there's two ways of fixing it. A) we don't move the closing quotes if a single quote character is used, or B) we normalize ' / " to ''' / """ respectively allowing the buggy logic to "just work"
bright glacier
late dewBOT
src/black/linegen.py lines 330 to 331
# We could enforce triple quotes at this point.
quote = quote_char * quote_len```I think this PR should call asyncio.set_event_loop(loop) and then asyncio.set_event_loop(None) in the finally
Ah I can't ```suggest on the finally block
https://github.com/psf/black/blob/main/pyproject.toml#L39 you could also enable filterwarnings=['error', ...] here so you don't get caught out by deprecation warnings in the future
late dewBOT
pyproject.toml line 39
"incompatible_with_mypyc: run when testing mypyc compiled black"```Oh yeah lol
bright glacier
of course you show up in the one issue I'm looking for :p @red sedge
bright glacier
it's almost two years old :o
bright glacier
Hmm, what's the status on removing the two-format-pass hack? I'd love to get rid of it if we can. Our error handling isn't built for it and it makes debugging Black more annoying >.<
I know we have a few test cases in the test suite, but I wonder if I still have my old script that tested Black against the many cases I extracted from the issue tracker.
Oh my, I found the issue data JSON, although I have no idea where the testing script is
I FOUND IT
This code is so awful, but it still works somehow.
Cases summary:
0 failed, 71 passed, and 0 were invalid
Failed cases breakdown:
0 failed due to instability and 0 failed due to invalid output
Remember this output lol?
bright glacier
Case summary:
4 failed, 78 passed, and 0 were invalid
Failed cases breakdown:
4 failed due to instability and 0 failed due to invalid output
hmm, removing the two-format-pass hack does show we still need it, sadly :(
--- first pass
+++ second pass
@@ -31,13 +31,11 @@
for s in (0.0001, 0.1, 1.0, 10.0, 1000.0, 100000.0)
for sign in (1.0, -1.0)
)
),
(),
-) + (
- (0.0, 0.0, 0.0),
-)
+) + ((0.0, 0.0, 0.0),)
interesting failure mode for psf/black#2074
toxic stormBOT
bright glacier
--- source
+++ first pass
@@ -1,7 +1,7 @@
def some_function(self):
- ''''<text here>
+ """ '<text here>
<text here, since without another non-empty line black is stable>
- '''
+ """
pass
--- first pass
+++ second pass
@@ -1,7 +1,7 @@
def some_function(self):
- """ '<text here>
+ """'<text here>
<text here, since without another non-empty line black is stable>
"""
pass
lol this should be simple enough to fix
OK failure #3 is of the same class as the last one. Failure #4 is another trailing comma instability case (I think?)
--- source
+++ first pass
@@ -1,4 +1,7 @@
class Visitor:
def check(self):
- if a == 's':
- assert type.type.type.names == [self.typemap["array"], self.typemap["char"]], line + f"{var} initialization type mismatch"
+ if a == "s":
+ assert type.type.type.names == [
+ self.typemap["array"],
+ self.typemap["char"],
+ ], (line + f"{var} initialization type mismatch")
--- first pass
+++ second pass
@@ -2,6 +2,8 @@
def check(self):
if a == "s":
assert type.type.type.names == [
self.typemap["array"],
self.typemap["char"],
- ], (line + f"{var} initialization type mismatch")
+ ], (
+ line + f"{var} initialization type mismatch"
+ )
hmm @dense jungle while fixing some docstring bugs, I noticed --skip-string-normalization doesn't stop black from normalizing docstring prefixes
β― black wat.py --diff --color -S
--- wat.py 2022-07-13 18:07:36.858247 +0000
+++ wat.py 2022-07-13 18:08:02.246896 +0000
@@ -1,3 +1,3 @@
def wat():
- F"""a"""
+ f"""a"""
pass
would reformat wat.py
should I fix that? technically it would be permissible under the exception we recently added to the stability policy since it wouldn't affect formatted codebases
Also of course it was me who introduced this bug ... https://github.com/psf/black/pull/2285/files#diff-98d222e4d1eed36d13bf2de2b3685fd3fbe65ff82450f30b77bfc11916cee76dR229
dense jungle
```diff
β― black wat.py --diff --color -S
--- wat.py 2022-07-13 18:07:36.85824...
part of me wants skip-string-normalization to only apply to quote style, not the prefix
since nobody really complains about the prefix
bright glacier
well we're inconsistent in that regard Β―_(γ)_/Β―
I agree though, no one really complains about the prefix.
dense jungle
I'd prefer to fix this only in preview mode. It's technically allowed under the stability policy, but I'd prefer to use the new exception only for really critical issues like dropping comments
bright glacier
.. except for MagicPython's capitial R, but I don't see us removing that exception anytime soon
The conditionals involved is going to be such a pain but that makes sense.
bright glacier
Hi @dense jungle , sorry to bother you again today, but I'm waiting for your approval on https://github.com/psf/black/pull/2874. You can just check the changes since your last review as I didn't force-push the branch and you've already approved it.
(i.e. make sure I didn't mess up my changelog entries)
bright glacier
finally!
flat krakenBOT
flat krakenBOT
red sedge
If I get a concept ack on this I'll try and whip up a PR
bright glacier
I'm inclined to reject psf/black#2527, thoughts?
toxic stormBOT
bright glacier
(either way, Shivansh is quite busy and the PR is very stale so I doubt it'll ever see the light of day)
austere lava
I'm planning on joining some mypy folks on the Europython sprint tomorrow
Apart from https://github.com/psf/black/issues/2845 is there anything else that would be good to resolve with them?
Cc @solid adder in case you had something else in mind
lament crow
There's also this which might be the same?
https://github.com/psf/black/issues/2867
bright glacier
uhhh, I could try compiling black with the latest mypyc and see what other issues are present
Jelle and I landed quite a few fixed for dataclasses and namedtuples so it shouldn't be crashing anymore (hopefully)
bright glacier
uhhh, I could try compiling black with the latest mypyc and see what other issue...
lol, I got sidetracked by another namedtuple/dataclasses __annotations__ bug shantanu raised
this time with the added fun of segfaults :D
while I've actually learned some C since writing my mypyc series, this is actually my first time seriously using gdb to debug an issue π
Program received signal SIGSEGV, Segmentation fault.
insert_to_emptydict (mp=0x7ffff752ed80, key=0x7ffff766c530, hash=-156007990909362949, value=0x0) at Objects/dictobject.c:1132
1132 Objects/dictobject.c: No such file or directory.
(gdb) bt 5
#0 insert_to_emptydict (mp=0x7ffff752ed80, key=0x7ffff766c530, hash=-156007990909362949, value=0x0) at Objects/dictobject.c:1132
#1 0x000055555569a2b1 in PyDict_SetItem (op=0x7ffff752ed80, key=0x7ffff766c530, value=0x0) at Objects/dictobject.c:1542
#2 0x00007ffff7523131 in CPyDef___top_level__ () at build/__native.c:412
#3 0x00007ffff75237f8 in PyInit_test () at build/__native.c:136
#4 0x00005555557b5a49 in _PyImport_LoadDynamicModuleWithSpec (spec=0x7ffff7544820, fp=<optimized out>) at ./Python/importdl.c:164
(More stack frames follow...)
(gdb) frame 2
#2 0x00007ffff7523131 in CPyDef___top_level__ () at build/__native.c:412
412 cpy_r_r55 = CPyDict_SetItem(cpy_r_r52, cpy_r_r54, cpy_r_r53);
seems like mypyc is trying to set VTableMethod.__annotations__["cls"] to ClassIR even though it hasn't been allocated or initialized yet
dense jungle
while I've actually learned some C since writing my mypyc series, this is actual...
right, value=0x0 is wrong
bright glacier
the whole way _annotations_ is handled for dataclasses/namedtuples feels a bit off honestly
https://github.com/python/mypy/blob/4479e0c2a75b7b802adf628034fa83beee9572c1/mypyc/irbuild/classdef.py#L494-L503 this is the problematic code
late dewBOT
mypyc/irbuild/classdef.py lines 494 to 503
if typ is None:
# FIXME: if get_type_info is not provided, don't fall back to stmt.type?
ann_type = get_proper_type(stmt.type)
if isinstance(ann_type, Instance):
typ = load_type(builder, ann_type.type, stmt.line)
else:
typ = builder.add(LoadAddress(type_object_op.type, type_object_op.src, stmt.line))
key = builder.load_str(lvalue.name)
builder.call_c(dict_set_item_op, [non_ext.anns, key, typ], stmt.line)```get_proper_type doesn't know what to do with a stringified type annotation ('test.ClassIR'), so it just returns it unchanged, and this function decides to just add a LoadAddress op for ClassIR even though it won't be allocated until later
The simple fix is to just load a constant string literal if get_proper_type returns a string, but I bet there's other issues in this logic worth fixing by reevaluating how _annotations_ is handled
Either way, I'll stop here for now and actually try to build black with latest mypyc as I said I would
bright glacier
right, `value=0x0` is wrong
value=0x0 means that it's an variable that has been zero-ed, right?
I'm totally guessing here, presumably key is also a PyObject pointer? ... I'm going to read the C-API docs now.
dense jungle
value=0x0 means that it's an variable that has been zero-ed, right?
it means it's NULL
bright glacier
ah so a invalid pointer
dense jungle
I'm totally guessing here, presumably key is also a PyObject pointer? ... I'm g...
yeah op should be the dict and key the key
bright glacier
int PyDict_SetItem(PyObject *p, PyObject *key, PyObject *val)
Oh look, I actually guessed correctly even though I'm not familiar with the CPython C-API haha
surprisingly black type checks OK with mypy from master without any intervention
build/__native_f4ee935d5338f3d318fb.c:75886:16: error: variable 'cpy_r_r7' is uninitialized when used here [-Werror,-Wuninitialized]
CPy_DecRef(cpy_r_r7);
^~~~~~~~
build/__native_f4ee935d5338f3d318fb.c:75509:23: note: initialize the variable 'cpy_r_r7' to silence this warning
PyObject *cpy_r_r7;
^
= NULL
1 error generated.
whelp.
bright glacier
from typing import Iterator
class Line:
def __init__(self) -> None:
self.depth: int = 0
class LineGenerator:
def __init__(self) -> None:
self.current_line: Line = Line()
def line(self) -> Iterator[Line]:
if not self.current_line:
self.current_line.depth += 1
return
yield Line()
I've minimized the reproducer to this so far.
bright glacier
https://github.com/python/mypy/commit/3c0409203a59eccb0e04caca7bef97c3e1e83ba7 is the first bad commit.
bright glacier
Apart from https://github.com/psf/black/issues/2845 is there anything else that ...
Sooo I tried compiling Black with the latest mypyc and I ran into this issue: https://github.com/mypyc/mypyc/issues/941
This issue has one heck of a title :)
austere lava
Wow thanks for digging
surreal valley
Hello! I'm Anika Roy and I'm a beginner in open source and I'm trying to make my first contribution.I found a issue i think i can work on,could someone help me out by telling the prerequisites for this,what i should read up and try before starting with it?
austere lava
Hey @surreal valley π have you read this? https://github.com/psf/black/issues/2560#issuecomment-1183920300
austere lava
With https://github.com/python/mypy/pull/13144 black now compiles with mypyc
bright glacier
Oh neat! I have no idea how keepalives or refcounts work in mypyc yet π
I've been looking a lot at irbuild.classdef for other bugs
austere lava
keepalives are a bit of a mess IMO
bright glacier
Thanks @austere lava!
austere lava
but I've only gotten acquainted 3 hours ago π
np π I chatted with Jukka and he said they might be interested in CI making sure black compiles with mypyc @ HEAD
so I'll look at throwing something together now
bright glacier
If you need help lemme know, but either way that sounds great
austere lava
do you know an easy way off the top of your head to run our tests against the binary wheel?
coz compiling is one thing, ideally we'd also not crash
bright glacier
so uhh for cibuildwheel I cheat and I don't run the full test suite because it requires multiple runs and I don't know how to make tox use a compiled wheel for that
austere lava
where's the cibuildwheel config?
bright glacier
austere lava
I assume we run this from some github action but can't find it in .github
ooooh your own repo
bright glacier
Yea, I still need to integrate it upstream π
austere lava
π need help? I have some time today and tomorrow
bright glacier
IIRC I've made the necessary changes locally, but I need to test them because I'm not deploying 100+ lines of CD configuration without testing
Which is mostly a matter of pressing some buttons and making a sacrifice to the "build correctly on first try" gods :p
red sedge
are these test cases intentionally triggering the syntax deprecation warnings? https://github.com/psf/black/pull/3175/files#r922679150
austere lava
As soon as https://github.com/hauntsaninja/mypy_primer/pull/41 is merged, mypy will get CI signal on if mypyc can compile Black (and run our test suite)
so hopefully no new mypyc PRs will break our ability to release
surreal valley
Hey <@911862685463883817> π have you read this? https://github.com/psf/black/is...
yes i did π i looked through the resources mentioned too.I was just wondering if there was something else i should look up too before i startπ
bright glacier
Gah, I didn't think before talking and made an incorrect assertion mainline CI is going to fail. I forgot that ichard26/black-deps-ci is using the master/main versions of deps π
red sedge
@bright glacier also it's a bug in aiohttp
bright glacier
wat
red sedge
They're not closing their transports correctly
bright glacier
Yeah I can see the ResourceWarnings, but AFAIK pytest isn't failing because of that. It's failing on
FAILED tests/test_blackd.py::BlackDTestCase::test_blackd_diff - DeprecationWarning: Middleware decorator is deprecated since 4.0 and its behaviour is default, you can simply remove this decorator.
Oh wait I was looking at the BaseEventLoop.__del__
And misdiagnosed it
The resource warnings will fail pytest on python 3.8+
bright glacier
are these test cases intentionally triggering the syntax deprecation warnings? h...
No clue, I'd have to look at the blame. My guess is that it was not intentional, but merely a byproduct of the commit author trying to stress test our parser and/or formatting logic
red sedge
that's not what I'm seeing, Β―\_(γ)_/Β―
It's line 682
surreal valley
Nah that should be enough to get you started
Got it,thank youππ»
flat krakenBOT
half meteor
hi ππΌ
first time chatting here
is this a good place to ask for PR review timeline expectations from maintainers, since i've sent a few PRs earlier?
lament crow
it seems that all of these PRs have merge conflicts?
half meteor
the conflicts are all in CHANGES.md which changes frequently, they are also trivial to resolve
so i hope to get a better idea of when they can be reviewed before updating
plain atlas
we just adopted towncrier for our own project, it was pretty trivial to set up
DisnakeDev/disnake#600
toxic stormBOT
plain atlas
don't let the massive diff confuse you 
and for the concerns of getting change notes in alpha documentation: sphinxcontrib-towncrier
dense jungle
(https://github.com/psf/black/pulls/yilei)
sorry I've been busy π¦ I'll try to take a look
half meteor
sorry I've been busy π¦ I'll try to take a look
understood, and no worries! as long as they will be reviewed. and i hope to send more contributions once they are done.
also, i'm more like to see a review on #3162, if you only have the time to take a look at one of them.
bright glacier
oh yeah sorry I've been dealing with some personal troubles so I've been out of it for code reviews lately
will try to take a look on one of them at least today
half meteor
will try to take a look on one of them at least today
again, no worries! take your time π
flat krakenBOT
bright glacier
@dense jungle both of these fmt: skip PRs are technically changing the stable code style, is that OK?
I'd hate to have to lock these improvements under --preview, would be clunky for both us developers and users
dense jungle
<@783088578363523104> both of these fmt: skip PRs are technically changing the s...
only if it fixes a crash
bright glacier
OK, both do fix crashes (technically psf/black#3158 fixes an infinite loop, but that seems major enough similiar to a crash)
toxic stormBOT
bright glacier
although psf/black#3148 also fixes some non-crashing format issues (mostly related to comments being moved or deleted) along side with some instability errors
toxic stormBOT
bright glacier
PTAL for both. I won't merge until you are fine with them from a stability standpoint.
I'd love if you could do code reviews too, but I'm confident enough in them that it shouldn't be necessary. Policy reviews are all I need to land them if we're getting close to the release deadline
dense jungle
although psf/black#3148 also fixes some non-crashing format issues (mostly relat...
wouldn't that fall under the recent change to the policy? (the changes only affect formatting unformatted code)
flat krakenBOT
bright glacier
Yes, they would but you wanted to keep our uses of the exception as minimal as possible hence why I asked for comment.
plain atlas
dense jungle
oh I see, I think fixing bugs where we eat comments is definitely a good use
comments should not be eaten
plain atlas
Yes, they would but you wanted to keep our uses of the exception as minimal as p...
imagine using an exception that was recently changed right away
dense jungle
we did, for another bug where we also ate comments π
bright glacier
oh I see, I think fixing bugs where we eat comments is definitely a good use
I didn't actually know that the PR fixed that behaviour until after asking you for comment. I did not check the linked issues π
plain atlas
i mean
eating comments is actually smart
because it enforces a coding style where comments can never be out of date
oki i'll see myself out now
bright glacier
I just assumed that they would change the formatting output of code using fmt: skip to something way less broken / janky
Also, thank you for the reviews Jelle!
Now this release is basically entirely held up on things I've assigned to myself haha. Might see if I can delegate some of the items.
flat krakenBOT
half meteor
thank you both for your reviews! i'll update the PRs per comments and resolve the conflicts tomorrow.
fast heart
I'm not able to reproduce this one
Oh. The file they're running are those three lines repeated several times
I'm not sure how specific it is to the test case given. A file containing 100k "1"s runs quite slow with black
dense jungle
i wonder if it's the parser or the formatter
fast heart
Ditto 100k "()"s, becoming slower the more nested the brackets are
late dewBOT
Hey @fast heart!
dense jungle
Top few entries while profiling a black run of 100k lines of `(())`
thanks mypyc π
fast heart
π
half meteor
what's the best way to trigger a rerun of the failed actions on a PR (due to flakiness)? i can git commit --amend and do a forced push, but i want to avoid it if possible
delicate marsh
what's the best way to trigger a rerun of the failed actions on a PR (due to fla...
you can push an empty commit
dense jungle
what's the best way to trigger a rerun of the failed actions on a PR (due to fla...
you can ask a maintainer nicely. closing and reopening also works
if you link me the PR I can push some buttons
half meteor
if you link me the PR I can push some buttons
could you please push the buttons on https://github.com/psf/black/pull/3162 and https://github.com/psf/black/pull/3148 ? π the failures are coveralls server errors
dense jungle
could you please push the buttons on https://github.com/psf/black/pull/3162 and ...
done for 3162. 3148 still has a few jobs running, I'll wait for those and then restart the failing ones
dense jungle
so many issues closed π
bright glacier
I might finally make some real ground on my issue closures π
bright glacier
Yay, this month's issue delta is negative! Although that will probably change by the end of the month π
https://ichard26.github.io/ghstats/ (it updates daily around 10AM EST, but I triggered a manual update)
red sedge
so ambv was talking to me about how many 3.6 downloads black is getting and why it has no many 'null' python version downloads and that he's not seen anything like it on other projects, however anyio seems to also have this large proportion of weekday 'null' python version downloads
https://pypistats.org/packages/black
https://pypistats.org/packages/anyio
glacial drum
Are those from non-standard pip downloads? EG conda, some archival copier etc.
green mulch
π€
tacit bane
Hello!
How do I in django to query a value in the bank greater than 100 and less than 1000
silent apex
Hello!
How do I in django to query a value in the bank greater than 100 and less...
Consider asking in #web-development
red sedge
Are those from non-standard pip downloads? EG conda, some archival copier etc.
Does conda download via archive?
red sedge
Are those from non-standard pip downloads? EG conda, some archival copier etc.
Anyway the trick to find out is to run the query and filter to include only the nulls and then look at the user agent
But we're both too scared of big query to do it
austere lava
are you too scared of the outcome or bigquery itself?
elder niche
red sedge
are you too scared of the outcome or bigquery itself?
big query itself
neon loom
Is this all for asyncio.run?
austere lava
Is this all for asyncio.run?
what do you mean? π
austere lava
oh I think it's unrelated to the mypyc fun
red sedge
The whole drop 3.6
The discussion started with it being about 3.6 stuff
But then got onto why black was getting so many more nulls than other projects
neon loom
Wonder if we can work on getting more info on what the nulls mean - I guess thatβs your plan digging into the bigdata dataset
red sedge
Yeah my working hypothesis is that the user agent does have the python version in but the regex in pypistats is wrong
fast heart
For automatically formatting code with black using github actions, is there a reference for getting this setup?
And comitting to the repository/PR
bright glacier
not really, we did get a PR for "autoblack" but we never landed it
my suggestion would be pre-commit.ci (and don't push to main directly)
dense jungle
we use pre-commit.ci for this for PRs on typeshed, works really well
fast heart
Pre-commit seems to be it!
It's doing at least half the job, but I don't think it's pushing: https://results.pre-commit.ci/run/github/515882591/1658437211.MUgD5ZoBTnKLR_m5VbfGmg
I just copy-pasted from the typeshed configuration
dense jungle
I just copy-pasted from the typeshed configuration
you may have to give it some special permission?
fast heart
you may have to give it some special permission?
Haven't been able to find it anywhere
flat krakenBOT
