I am using replit
#black-formatter
spare hamlet
tribal thistle
Ironically, this is both the wrong and right channel to ask in
If you'd like help with your python code, check out #โ๏ฝhow-to-get-help
Black may be able to fix some of your simpler problems though, its a tool that auto formats your code
bright glacier
@plain atlas can you ask your question again, it got deleted unintentionally
vivid pasture
Sorry about that
bright glacier
No worries, thanks for the swift action ๐
plain atlas
how are python warnings set to be emitted during running? Ik its something with the W flag....
plain atlas
<@717983911824588862> can you ask your question again, it got deleted unintentio...
yes ik it wasn't a black question but ๐
okay, got it
python -W always
or PYTHONWARNINGS=always
bright glacier
... now I think about it, I should probably check windows
tribal thistle
U call?
bright glacier
Uhm what?
tribal thistle
I usually like to break things on windows, anything I can help with?
bright glacier
Oh alright then, if you could setup a quick test environment for dawn's PR and cd into some random directory (\tests should do) that isn't the project root and verify no tests break that'd be awesome.
should be roughly these steps (yeah i'm not using pipenv)
although you'll have to checkout dawn's PR instead just cloning psf/black
tribal thistle
Tested on 3.9, no unexpected failures
bright glacier
alright so dawn's PR is good to ship :)
silent apex
woohoo!
bright glacier
thanks @tribal thistle and @silent apex 
silent apex
Ahh, I completely missed Marco's second comment, thanks for fixing it!
I swear I copied that code directly from his comment, I guess I removed it for some reason
bright glacier
@red sedge in what way?
bright glacier
yea not sure what happened to that link
red sedge
Sometimes two and sometimes 2
It will be harder to grep for them when removing the code
bright glacier
I mean, I explicitly split the PR into two commits so we could just revert the second one
But sure I can do that
red sedge
Black helps avoid a whole bunch of pedantry, it has to spill out somewhere
flat krakenBOT
bright glacier
Black helps avoid a whole bunch of pedantry, it has to spill out somewhere
fair enough ๐
red sedge
One has a semi and the other doesn't...
bright glacier
huh?
red sedge
A semi colon
bright glacier
how does a semi-colon get into my code, sigh, where is it?
red sedge
https://github.com/psf/black/pull/2523/files#diff-3b01c50563d8a5ba8893fc84303571a9b8a57bbcd2823ad65164008846d884acR77 missing a semi
red sedge
Yeah I wouldn't have mentioned it if it didn't go with the two/2
Deprecate; python two support
bright glacier
Oh wow not only did mypyc NOT subtly change black's formatting behaviour (woo!) main has no formatting differences compared to 21.9b0 as well ๐
plain atlas
!remind @bright glacier 1y that bug that just cropped up? its def related to this. /hj
late dewBOT
plain atlas
damn it
bright glacier
I can't either unfortunately, mod+ only
plain atlas
aha
bright glacier
building python from source with all of main optimizations disabled + interpreted mypy = you're gonna be having a long "grab a cup of coffee break"
dense jungle
Let's merge some PRs
bright glacier
I'm opening one right now :)
somehow we didn't notice that the loop parameter was removed in the various asyncio high-level APIs in 3.10
I'm surprised we haven't gotten a report about this since this code should crash if hit
bright glacier
Other than ลukasz, I'm pretty sure the Python 2 deprecation PR has enough reviews. Should we merge it @dense jungle ?
dense jungle
Other than ลukasz, I'm pretty sure the Python 2 deprecation PR has enough review...
I'll take a look soon, I'm like two PRs away from it
plain atlas
Considering how major it is, maybe not??
dense jungle
Looking at the --verbose one right now
plain atlas
-v is fun
bright glacier
ah yeah I ignored that one because I didn't want to think through their suggestion
bright glacier
Considering how major it is, maybe not??
Honestly the people on Python 2 should of knew this would be coming eventually.
dense jungle
I am fully expecting to merge it to be clear, just want to take another look at the code
plain atlas
Although
Should it be pending deprecation warning and then next month switch to deprecation warnings?
Because deprecation warnings can break stuff
bright glacier
Well black is still an application so I'm not sure how I'd feel using the warnings API
toxic stormBOT
bright glacier
I mean, black both doesn't have an official API and Python 2 has been sun-setted for a long time now
dense jungle
I don't think the warning API is a good fit for us
bright glacier
Not yet although we want to work on this eventually. Personally I plan to work on this when I finish up with the various projects I'm currently working on (it's a lot btw).
plain atlas
Okay yeah
Just checked
Looks fine, mostly
What does err() do?
I'm on mobile right now so couldn't find the definition
bright glacier
One potential problem I do want to bring up is that right now each call to format_str with Python 2 warns
... that may be too noisy
flat krakenBOT
plain atlas
Could save a variable once it's warned once and only warn if that variable isn't x
bright glacier
yeah that seems like a decent idea
plain atlas
I mean, black both doesn't have an official API and Python 2 has been sun-setted...
Is there a plan for an official api sometime?
bright glacier
....
bright glacier
Not yet although we want to work on this eventually. Personally I plan to work o...
^
plain atlas
Smh I missed that
Imagine finishing a project tho
I add a new one like every week or so
dense jungle
I'm trying to understand the print/exec stuff in the Python 2 PR now. Is the point of that to make sure we notice that the code is Python 2-only so we warn about it?
There's a bunch of other things we could look for, though I'm not sure it's worth the effort (except Exception, e:, 2L, def f((a, b)): come to mind)
oh and backticks though I'm not sure anyone ever used those
flat krakenBOT
plain atlas
There's a bunch of other things we could look for, though I'm not sure it's wort...
Hmmmmmmmmmmmmmmmmmmmmm
I wonder
Wouldn't the mode need to be set to python2 to format python 2 code?
Eg the list of modes to format for in pyproject.toml
dense jungle
nope
2L
By default we use a lib2to3 grammar that is compatible with both Python 2 and Python 3
plain atlas
How is 2L valid in py2??
silent apex
Why doesn't black have an official API? Is there a reason there's not one, besides just not making it?
plain atlas
Because Ichard hasn't made it
dense jungle
How is 2L valid in py2??
It means a long literal
In Python 2 int was limited to a machine int (32 or 64 bit) and long was the arbitrary-precision type
dense jungle
Why doesn't black have an official API? Is there a reason there's not one, besid...
We just haven't gotten around to it
plain atlas
Why doesn't black have an official API? Is there a reason there's not one, besid...
Because it wasn't a thing initially, and no one has done it as black has grown over adoption, it will also be tough to do it as its a large refactoring
I think
silent apex
We just haven't gotten around to it
Got it, makes sense
plain atlas
It means a long literal
Ahhhhh ok
bright glacier
I'm trying to understand the print/exec stuff in the Python 2 PR now. Is the poi...
Yup!
those were the only Python 2 only things I knew off the top of my head
dense jungle
They're probably also the easiest to detect ๐ . I'm a bit worried it won't be enough since exec is rare anywhere and print is also rare in many big codebases
Then again, I don't really want to go out of my way to help people who still write Python 2-only code in 2021
Reading the what's new brought up a few more: 0360 for octal literals; raise Exception, "message" syntax; [x for x in 1, 2] (needs parens in Python 3). Honestly those are all pretty marginal though
bright glacier
there's also the automatic unpacking in parameters (mind blown btw)
dense jungle
Yes, I listed that above. That was pretty useful actually ๐ฆ I had to make some code uglier when migrating to Python 3 to deal with that
dense jungle
I think we should merge the Python 2 PR as is but there are conflicts blocking a rebase
flat krakenBOT
bright glacier
I'll take a look tomorrow
plain atlas
lmao the test stuff that was causing a problem and not allowing the tests to be ran from a different directory....
I just made a simple script for my project so that isn't a problem lmao
although
i can't pass anything to it anymore, so uh, forget i said anything
fallow briar
e
tired shard
thanks jelle for the review, will take a look soon
gusty grail
@median garnet This channel is for discussion about usage and development of Black, a python code formatter. I've deleted your messages since it's off-topic for this channel. See #โ๏ฝhow-to-get-help for instructions on claiming a help channel.
sturdy charm
How do i compile python in windows 7 without actually having to download python (exe)
what are python developers trying to hide ?
bright glacier
Wrong channel, but also why?
bright glacier
the windows binaries are already compiled with PGO
bright glacier
that's OK, just please use a different channel then. I'm not sure which one would be ideal though.
sturdy charm
so no one knows and python developers are trying to hide something ?
bright glacier
Well I don't work with C on a daily basis on Windows, and I doubt the core developers of a pure python project (psf/black) would know either. #c-extensions or #internals-and-peps *may* know more but I honestly have no idea.
TIL black-21.9b1.dev51+gb0b3709-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_12_x86_64.manylinux2010_x86_64.whl is a valid wheel filename
that's alotta platform and ABI tags ๐
what's blocking the 0.920 mypy release btw @dense jungle ?
man, I hate you windows
oh lol I forgot mypyc doesn't support python 3.10 officially
well I guess I'll skip windows on python 3.10 then for the time being
bright glacier
yea... the # fmt: skip implementation is incredibly buggy ๐ฆ
bright glacier
because we don't support match case syntax
slow ibex
shouldn't it be skipped?
bright glacier
that's not how the fmt: off/on implementation works, the whole file is still parsed
slow ibex
ok, thanks
bright glacier
should probably include the python 2 deprecation today then
I've been working on the mypyc project so far today
neon loom
It ready for me to click merge?
bright glacier
which PR?
neon loom
Your python2 deprecation - Looks it to me
neon loom
Just did this to see if we still work apart from that known tricky unicode problem: https://github.com/psf/black/pull/new/remove_regex
@bright glacier - Want to see how much slower this is too?
bright glacier
As in you want me to benchmark it?
neon loom
If it's easy
bright glacier
I'm a little too busy for that for the time being
neon loom
No worries
If I get bored later I might try your tool
Just interested cause if we're not to much slower i might see if we can fix the unicode issues via monkey patching re possibly - Ideally we'd just fix it in cpython but that will take years to become the version everyone uses (but if I work it out I might submit the PR - but I need to remember all the context)
plain atlas
How do i compile python in windows 7 without actually having to download python ...
Idk how on windows, but ik how on Linux so it's definitely possible given its all open source and none of how to do it is hidden
plain atlas
Your python2 deprecation - Looks it to me
Depends.... If you're gonna do a Halloween release the version in that pr should be updated to 22.1b
blazing pike
does lack support match statements in 3.10 yet?
tribal thistle
Noot yet
blazing pike
sadness
dense jungle
@bright glacier about mypy 0.920, I don't know what's blocking it. mypy releases are done by the people at Dropbox and I don't have much insight into when they can do it
I heard they're no longer working on mypy full-time so it's probably just whenever they have time
I think the Python 2 PR needs to be rebased because we wanted to keep the two commits separate. I can do that later in the afternoon after I recover from climbing this hill ๐
dense jungle
there we go
flat krakenBOT
flat krakenBOT
neon loom
Ok, think I'll merge the docker and then cut a release unless anyone says not to - PR coming.
bright glacier
Ok, think I'll merge the docker and then cut a release unless anyone says not to...
awesome thanks!
bright glacier
oh it wasn't rebase onto main, oh well
neon loom
Why do we have more manual things to do @ release ๐ฆ Booo
dense jungle
oh it wasn't rebase onto main, oh well
It was actually
neon loom
We also have a lot of deprecation errors to fix
bright glacier
It was actually
oh is GitHub mobile's UI not great at demonstrating that then?
I'mma check on a laptop
neon loom
I think I'll put up a PR to remove this version manual step. I don't think the step it adds for each release now is worth it
bright glacier
neon loom
And TIL we have two files called the_basics.md
bright glacier
Do I just not understand what a rebase merge is like or does that look like a squash merge?
bright glacier
And TIL we have two files called the_basics.md
Yea.. I regret calling them that
neon loom
Damn, another file. We haz so much manual editing now. I might at least make a set_version.py or something that goes and does all the version replacing ...
bright glacier
we could probably automate it using a local sphinx extension
bright glacier
damn this will be our sixth release in a row in terms monthly releases, way to go on the consistent quicker deploy cycle
once again thank you @neon loom for getting the ball rolling on this front :)
flat krakenBOT
bright glacier
We should probably let people know via other channels that this release is important as we now have a formatting stability policy AND Python 2 is deprecated
neon loom
Where do you propose?
bright glacier
twitter and discuss.python.org are the only ones I know we have access to
we could also do a mailing list or two but do people even pay attention to those these days?
anyway, assuming https://github.com/psf/black/runs/4062923508?check_suite_focus=true passes (which it should) can we close psf/black#2568?
toxic stormBOT
silent apex
Where do you propose?
My suggestion is to stick a little banner at the top of https://black.rtfd.io, and maybe a note on the GitHub README
Maybe "banner" is the wrong word, but one of those 20px lines that are on the top of every page, or at least the first
bright glacier
also great ideas
bright glacier
Maybe "banner" is the wrong word, but one of those 20px lines that are on the to...
something like the banner on this page? https://pradyunsg.me/furo/quickstart/
silent apex
Yeah, the "if you like this theme" thing
neon loom
anyway, assuming https://github.com/psf/black/runs/4062923508?check_suite_focus=...
Yup, was just waiting to see it make it to docker hub
48MB. Nice
flat krakenBOT
bright glacier
We also have a lot of deprecation errors to fix
speaking of deprecation warnings ... we totally didn't realize that the loop parameter was removed in the various asyncio APIs in 3.10 ... I'm surprised we haven't had a report about this
neon loom
Banner etc. + twitter could be good. Maybe you start a post @bright glacier and I'll report and we nudge @ambv too
neon loom
speaking of deprecation warnings ... we totally didn't realize that the loop par...
yeah, we should open a issue for each of them to cleanup.
I'm heading for dinner now - But if you don't get to it I'll try get some issues opened so we can fix em
Let me know if you tweet - I don't really look at twitter much
bright glacier
thankfully the loop parameter was only being passed in some exception / cleanup code so it'd be rare for those (crashing) pieces of code to be hit but yea we should probably set up pytest to fail loudly on new warnings
neon loom
Yeah, lets get it clean and fail on warnings
bright glacier
I'm heading for dinner now - But if you don't get to it I'll try get some issues...
Have a great dinner :)
tired shard
Hey @errant barn ,
Re: https://github.com/psf/black/pull/2526#issuecomment-947563571
What do you think would be a good way to handle this, we are checking for configuration first so it is logged earlier, but if we don't check it earlier and black fails due to a config value (eg. pyi and ipynb- both were passed), it would never reach the project root log, so we need to log the configuration earlier.
errant barn
hmm true, didn't consider that
tired shard
Please
this isn't the right channel. feel free to have a look at our resources https://www.pythondiscord.com/resources/ or check out #โ๏ฝhow-to-get-help if you need help
lean path
this isn't the right channel. feel free to have a look at our resources https://...
Thanks for help
bright glacier
I'm a little nervous that @muted haven 's PR is going to break mypyc compatibility as it touches blib2to3 quite significantly but I can't imagine it would be horrible to fix ... unless there's dataclasses involved in which case CI gonna's be red :)
also the merge conflicts will be ... a bit of a pain to deal with :)
plain atlas
lol
bright glacier
I suspect this will undo some of my blib2to3 optimizations as well eek!
plain atlas
that was a good read
why are consumers of open source and free software the most entitled
dense jungle
I'm a little nervous that <@!686127960440438795> 's PR is going to break mypyc c...
I hope it won't be too bad, I'm excited about both of those things landing
dense jungle
why are consumers of open source and free software the most entitled
Maybe if they yell harder next time I'll change my mind
plain atlas
lol i'm reading all of the issues that they linked
there is even a fricking solution
If you want tabs, there's a tool for doing that: https://linux.die.net/man/1/expand
actually that does it in reverse
holy shit
plain atlas
Maybe if they yell harder next time I'll change my mind
you may already remember reading this, such a pedantic comment: https://github.com/psf/black/issues/47#issuecomment-439596440
dense jungle
oh is that the guy with the dictionary definitions?
plain atlas
yess
LOL
"Use tabs instead of spaces for indentation. "
"Tabs are always equal to 4 spaces."
which isn't true tho
tabs are just tabs
muted haven
I'm a little nervous that <@!686127960440438795> 's PR is going to break mypyc c...
uh, sorry about that. I can type the code, but not sure if there are any stuff that would be problematic in mypyc
plain atlas
lmfao
dense jungle
uh, sorry about that. I can type the code, but not sure if there are any stuff t...
for context the potentially conflicting changes are https://github.com/psf/black/pull/2431/files#diff-752cb00af77ac75adb6cb5b53d3c28691db40cee3e2f517d755bcea983e7ec99 vs. https://github.com/isidentical/black/commit/c22fca62bdeca503ef1fa56647b410cbd3d5af81. Looks like the addtoken method might get some gnarly conflicts and the match code needs some annotations.
silent apex
This is so rude my god
red sedge
I just don't get how this is such a thing
plain atlas
Because\tpeople\tneed\ttheir\ttabs
red sedge
Pretty sure there's a fork/wrapper that runs black and replaces spaces with tabs
tribal thistle
Many lol
red sedge
Tabs for indentation spaces for alignment
plain atlas
Wouldn't py3 choke on that?
uncut quiver
Hi, what is the function that reformats a file's content?
tribal thistle
Are you asking for the cli command, or are you looking for the function in the source code
uncut quiver
The function in the source code as I'm trying to reformat a string of code
tribal thistle
Call chain is:
main -> reformat_one/reformat_many -> format_file_in_place -> format_file_contents -> format_str (also check_stability)
If you're reformatting just a string, you may be able to skip right down to format_str, but option parsing is still a thing
uncut quiver
Alright thank you
tribal thistle
You could use reformat_code though, which does take a string, no file
uncut quiver
Yeah I will try both
sour jay
you may already remember reading this, such a pedantic comment: https://github.c...
picks an opinionated linter
Why can't I indent my code with tabs?
plain atlas
*picks an opinionated linter*
Why can't I indent my code with tabs?
uses python
Why can't I use brackets?
sour jay
;-;
plain atlas
Honestly i don't even know what that user was thinking
intellectually inconsistent
Like
WTF does that even mean
sour jay
Only intellectually consistent people would get that ๐ค
plain atlas
Smh
plain atlas
i don't get that, tbh
flat krakenBOT
dense jungle
more pipenv fun
bright glacier
more pipenv fun
You know that it's my favourite pastime :p /s
neon loom
https://twitter.com/brettsky/status/1453076484256505859?s=20
Brett is welcome to do a PR removing regex ...
https://github.com/psf/black/tree/remove_regex - Works on many files I tested with, but fails with those unicode tests ...
dense jungle
https://github.com/psf/black/tree/remove_regex - Works on many files I tested wi...
Yes we'll probably need to change something in blib2to3
neon loom
blib2to3 itself, too, uses regex ...
dense jungle
Yes, I think that's where we need it. But maybe we can do something manual in a few places so it still supports those exotic identifiers
flat krakenBOT
bright glacier
@muted haven I'll do some testing on this PR as I (finally) have free time. I'll check mypyc compat and behaviour equivalence with my own tool, diff-shades, with is basically black-primer / gallery but more integrated and full-featured (eventually, it's a WIP). If I feel like it I'll do some benchmarking but I make no promises on this one :)
I honestly have no idea why those two tests are failing either :(
dense jungle
Apparently it's not printing the Python 2 deprecation message? Not sure why that would be
muted haven
AFAIK it can't currently parse the python2 code, but don't know why it shouldn't affect much
I have not yet found time to research it
bright glacier
Yea... the grammar fallback code should handle it just fine ...
muted haven
hopefully on weekend I can get to some of these issues
bright glacier
Oh and just to be clear please don't feel like your work on gallery was pointless, at the bare minimum it spawned the creation of black-primer, which then inspired mypy-primer which then inspired me to create diff-shades haha :)
Also I'm fairly confident other people from the core team has used gallery before, I just haven't
Your creation is a parent lol
muted haven
Oh no worries on that. Gallery was an interesting experience for me, since it also inspired another project (reiz) that I ended up writing a paper about.
bright glacier
Of course, so gallery is the source of many great things ๐ /lh /gen
muted haven
Indeed
Also it was my first contribution to black, so there's that. Hopefully this would be the second one (and perhaps more to come)
bright glacier
Oh my, the merge conflict on src/blib2to3/pgen2/parser.py with the mypyc branch as the base is way smaller than I expected ๐ :)
recursive merge strategy for the win I suppose!
Traceback (most recent call last):
File "/home/ichard26/.local/share/virtualenvs/black-Q9x7i-w-/bin/black", line 5, in <module>
from black import patched_main
File "src/black/__init__.py", line 39, in <module>
from black.nodes import STARS, syms, is_simple_decorator_expression
File "src/black/nodes.py", line 32, in <module>
pygram.initialize(CACHE_DIR)
File "src/blib2to3/pygram.py", line 175, in initialize
python_symbols = _python_symbols(python_grammar)
File "src/blib2to3/pygram.py", line 32, in __init__
setattr(self, name, symbol)
AttributeError: '_python_symbols' object has no attribute 'asexpr_test'
first issue trying to run the mypyc compiled version of the match PR :)
looks like we're just missing an entry in this list: https://github.com/psf/black/pull/2586/files#diff-f226b3f8be705bbb2a754559f1cd759b495e58bae3a9f21746af5c5f10a80753R135-R142
bright glacier
1โฏ black test.py --diff -v --fast
Using configuration from /home/ichard26/programming/tools/diff-shades/pyproject.toml.
Traceback (most recent call last):
File "/home/ichard26/programming/tools/diff-shades/venv/lib/python3.8/site-packages/black/__init__.py", line 643, in reformat_one
if changed is not Changed.CACHED and format_file_in_place(
File "/home/ichard26/programming/tools/diff-shades/venv/lib/python3.8/site-packages/black/__init__.py", line 796, in format_file_in_place
dst_contents = format_file_contents(src_contents, fast=fast, mode=mode)
File "/home/ichard26/programming/tools/diff-shades/venv/lib/python3.8/site-packages/black/__init__.py", line 920, in format_file_contents
dst_contents = format_str(src_contents, mode=mode)
File "/home/ichard26/programming/tools/diff-shades/venv/lib/python3.8/site-packages/black/__init__.py", line 1057, in format_str
src_node = lib2to3_parse(src_contents.lstrip(), mode.target_versions)
File "/home/ichard26/programming/tools/diff-shades/venv/lib/python3.8/site-packages/black/parsing.py", line 96, in lib2to3_parse
raise exc from None
black.parsing.InvalidInput: Cannot parse: 3:6: match xt:
error: cannot format test.py: Cannot parse: 3:6: match xt:
Oh no! ๐ฅ ๐ ๐ฅ
1 file would fail to reformat.
~/programming/tools/diff-shades on main [$?โก] via v3.8.5 (venv) took 400ms
123โฏ cat test.py
x = hi
match xt:
case (a, n):
pass
what-
pip install-ing a local clone with the match PR did not grant black the magical ability to parse match ๐ฆ
plain atlas
are you running the correct black?
dense jungle
and you don't have local .so files lying around?
bright glacier
no compiled C
plain atlas
o
dense jungle
it stands for shared object
plain atlas
ah
bright glacier
and you don't have local .so files lying around?
none as far as I can tell
bright glacier
lol
12 failed, 251 passed, 3 skipped, 12 deselected, 3 xfailed, 12 warnings in 22.15s
``` OK this might take a while to fix ...hmm, looks like the token.* numbers got shifted around?
yup they did
I don't understand how the numerical IDs are calculated tho
flat krakenBOT
flat krakenBOT
flat krakenBOT
errant barn
Yes and no: no on the new syntax but yes on general runtime
onyx quarry
is there a way to just have the formatter skip the lines so i can still format the rest of the file
fmt: off doesnt seem to be doing anything
as far as stopping a crash
errant barn
We still need to parse the file, so unfortunately I think not
onyx quarry
what do you suggest, should I just turn formatting off for the file until its supported?
errant barn
you can follow psf/black#2242
toxic stormBOT
errant barn
In the meantime, probably yeah
onyx quarry
Thanks
flat krakenBOT
bright glacier
@silent apex do you mean all releases by "minor" or did you mean to say "major"?
silent apex
Err, I meant whatever 21.09 and 21.10 are
Idk what versioning scheme black uses
Apologies if that was unclear, the only versioning I speak is semver
dense jungle
we don't really have major vs. minor versions, it's just a matter of how many we get out in a month
which does make it more awkward to select which versions to show in the docs
plain atlas
calsemver ;-;
silent apex
Well then uh, whatever a new version is (21.09, 21.10, 21.11, etc.)
The terminology doesn't really matter
tribal thistle
Powerplay: 21.101
uncut ivy
!e
flat krakenBOT
dense jungle
Seems like we got some dependency problems again: https://github.com/psf/black/runs/4117987697?check_suite_focus=true. Something with coverage?
dense jungle
Don't think so, but we may not have pinned this dep
tribal thistle
Yeah, it's a known issue with the latest version of coveralls
Actually, this isn't the first time a change in coverage.py has broken the coveralls package
A lot of the issue history is fixing that
bright glacier
Since I'm getting closer to integrating my mypyc work into main, I was wondering on how we'd collect field reports from users. My original idea was to use Google Forms but a GitHub issue would also work well. Google Forms would be more structured but the lack of transparency can be bit of an issue for some people. Additionally an issue would be easier to ask follow-up questions within while a Google Form would require a "contact info" field.
... I guess my pros-cons list is leaning towards a gh issue :)
sour jay
Doesn't it come pretty naturally for people to just find the repository and open an issue?
dense jungle
Yes, I think issues are fine
silent apex
GitHub issues or discussions (a single thread maybe, since there are comments on replies) sounds fine to me personally
bright glacier
feels good to be making actual progress on diff-shades even if I'm still horribly slow :)
src/diff_shades/analysis.py: # TODO: record log files if available
src/diff_shades/analysis.py: # TODO: allow more control w/ black.Mode so we could use diff-shades to compare
src/diff_shades/analysis.py: # TODO: refactor this and related functions cuz it's a bit of a mess :)
src/diff_shades/cli.py: # TODO: implement a list view
src/diff_shades/cli.py: # TODO: implement a diff + failures view
src/diff_shades/cli.py: # TODO: allow filtering of projects and files checked
src/diff_shades/cli.py: # TODO: more informative output (in particular on the differences)
src/diff_shades/cli.py: # TODO: Gracefully warn but accept analyses that weren't set up the exact same way.
src/diff_shades/cli.py: "[italic]-> Eventually this will be just a warning, but that's a TODO"
src/diff_shades/cli.py:# TODO: implement this command :p
got a ton of TODOs tho haha
open crystal
black lives matter
bright glacier
Yes they do! .. although it's a little off-topic here fwiw
open crystal
richard
bright glacier
yeah?
open crystal
could you help me with something in #help-bread pls
slow ibex
this is terrible```py
# it is kw argument
partner_id_and_days=CustomCallable(
decode=lambda buf: 0
if buf.read_uint(pos=buf.pos) == 0
else (buf.read_uint(), buf.read_uint()),
encode=lambda buf, obj: buf.write_uint(0)
if obj == 0
else (buf.write_uint(obj[0]), buf.write_uint(obj[1])),
),
# ... other arguments
thanks ๐
bright glacier
psf/black#2278
toxic stormBOT
thin coyote
hello guys
bright glacier
OK we have less than three months to get black ready for a stable release ... there's a lot of work to be done still:
improve Python 2 detection- integrate and deploy mypyc
implement basic / experimental match supportland PyPy support- merge psf/black#2278 (maybe lock this under the --future flag?)
- merge psf/black#1879 (and lock the formatting under the --future flag)
- turn on ESP
toxic stormBOT
bright glacier
what's annoying about the first three is that they're connected and merging one of them will complicate the others
silent apex
OK we have less than three months to get black ready for a stable release ... th...
- turn on ESP
What's that?
bright glacier
our codename for --experimental-string-processing
silent apex
Ah, got it
bright glacier
sigh, looks like I'll have to unbork CI: https://github.com/psf/black/actions/runs/1429457406
bright glacier
am I being dumb or is the following valid code:
value = (1, 2)
match value:
case x, y:
pass
I can't get the match support PR to accept this fwiw ๐ฆ
silent apex
am I being dumb or is the following valid code:
```python
value = (1, 2)
match ...
The interpreter accepts it ๐คทโโ๏ธ
on 3.10.0, at least
bright glacier
I can't get the match support PR to accept this fwiw ๐ฆ
cc @muted haven
the following works:
value = (1, 2)
match value:
case (x, y):
pass
although I have no idea why, I'd need to take a closer look at the grammar to figure out why
silent apex
I'd guess it's like x, y = (1, 2) and (x, y) = (1, 2)
bright glacier
match_stmt: "match" subject_expr ':' NEWLINE INDENT case_block+ DEDENT
subject_expr: namedexpr_test
# cases
case_block: "case" patterns [guard] ':' suite
guard: 'if' namedexpr_test
patterns: or_test ['as' or_test]
this is the main bit of the blib2to3 grammar that's powering the PR but I don't understand patterns in particular or_test very well
bright glacier
OK we have less than three months to get black ready for a stable release ... th...
bright glacier
OK we have less than three months to get black ready for a stable release ... th...
for the time being I'll work on item 1 and 3
muted haven
If you could post it under the PR, I'll try to check if it works or not.
bright glacier
If you could post it under the PR, I'll try to check if it works or not.
fyi I figured out why the python 2 tests are failing under your pr
the TLDR is that I messed up when writing up the python 2 deprecation PR
in more details I was using the token.* constants to detect the python 2 features. token.* is only guaranteed to be constant below a node type value <256 and the Python 2 nodes were above that. I should've been using pygram.python_symbols instead
tired shard
For python2 detection can't we just use https://github.com/python/cpython/tree/main/Lib/lib2to3 and see if there is a diff?
bright glacier
I'm confused, could you elaborate?
plain atlas
!d lib2to3
late dewBOT
bright glacier
We use lib2to3 as our main parser, just a fork of it
plain atlas
fyi @tired shard black doesn't know that the code is there at all, black can format py2 without python2 installed.
tired shard
lib2to3 reads python2 code and applies some fixes to make it compatible with python3 right? So if we just use this tool to convert the source and then check the original code and to check if there is a diff, we can find if the original code is python2 or 3.
Example
-print "printing something"
+print("printing something")
This is the first think that popped in my mind.
bright glacier
That's probably a bad idea from a performance perspective
plain atlas
the problem in that case is insane performance ^^
bright glacier
anyway I'm almost done with my solution
tired shard
well yeah
bright glacier
cache
plain atlas
ah, where do i nuke that?
bright glacier
I actually have to clear the cache everytime I test something out
tired shard
~/.cache/black
bright glacier
tired shard
isn't there a clear cache option
bright glacier
nope.
tired shard
it would good to have one imo
bright glacier
I've actually made my own dropblackcache utility for this exact reason lol
muted haven
Thanks @bright glacier for the info, I'll try to get back to PR tomorrow and fix the case you sent!
plain atlas
oh well, ig it didn't make any file changes since they were already formatted 
welp, my git status is clean
bright glacier
Thanks <@429445826251587616> for the info, I'll try to get back to PR tomorrow a...
Yup, and I'll put up another PR to improve and solidify my python 2 detection code your PR shouldn't bork it ^^
I also have a local patch fixing some other issues for your branch fyi
plain atlas
welp, my git status is clean
ran different formatter to mess up my code so i can give black a real test
30 modified files now
late dewBOT
You are not allowed to use that command here. Please use the #bot-commands channel instead.
plain atlas
...
bright glacier
!snf 906600530388852747 906600635267424267
late dewBOT
Snowflakes
plain atlas
....that was 25 seconds
no, that was 5 seconds
wtf is wrong with time
41 files reformatted, 20 files left unchanged.
black . 23.61s user 0.75s system 613% cpu 3.974 total
bright glacier
are you reading the user time
cuz it won't be accurate due to multicore usage
IIRC
bright glacier
I honestly don't remember if time measures the wall clock or CPU time so I can't answer that quesiton
plain atlas
sure, it could work, but like
i was trying to get the cold evidence for it but i'm just gonna have to settle for my word...
50 files or in the repo and black will take around 10 seconds on pre-commit
even with no changes
yapf . -rip
bright glacier
oh yea ci is going to break most likely
muted haven
I also have a local patch fixing some other issues for your branch fyi
Amazing, if it is related to the parts I introduced feel free to push my branch (I think maintainers have that privilege by default on GitHub, not sure).
bright glacier
looks like hypothesis's devtooling is now 3.8+ https://github.com/psf/black/runs/4127052658?check_suite_focus=true#step:5:282
@tribal thistle so wait, how am I supposed to fix coveralls breaking the action?
apparently the most recent release of coveralls-python should've fixed this but no?
tribal thistle
<@!386599446571384843> so wait, how am I supposed to fix coveralls breaking the ...
Gah it pisses me off so much. The fix is upgrading, or downgrading coverage.py, either to before the change that broke this (6.0) or after the fix (not yet released, could ask nedbat for an ETA). More details in TheKevJames/coveralls-python#333
Either way, my professional recommendation is pin both coveralls and coverage
tribal thistle
Or maybe give your thoughts on TheKevJames/coveralls-python#334 ๐
toxic stormBOT
red sedge
It does worry me about the coverage py bus factor...
tribal thistle
Mmm, to get help, make sure to post your code as text instead of a screenshot
Isnโt pandas huge?
If it takes that long usually, itโs probably best to disable it
bright glacier
Yeah but it took over an hour when Django only took a quarter of that
it usually doesn't take an hour
Either diff-shades is broken or this PR I'm testing has seriously broken something... Or both!
tribal thistle
Is this diff shades or primer
Ah
Roll back, test again. Keep rolling back until things work
plain atlas
Either diff-shades is broken or this PR I'm testing has seriously broken somethi...
Or all three!
bright glacier
Roll back, test again. Keep rolling back until things work
Yeah I will but man will it take a while given how slow even a normal non-buggy run takes
30m on my laptop ๐ฆ
tribal thistle
Lol right, hadnโt thought about that
bright glacier
I looked into using a VM but I don't have access
tribal thistle
If it wasnโt so late rn, Iโd donate some CPU time
Tomorrow if youโre still at it
plain atlas
30m on my laptop ๐ฆ
Holy hell that's long
bright glacier
thankfully it's just one project acting up
plain atlas
If it wasnโt so late rn, Iโd donate some CPU time
Running black test suite on pydis cluster ๐คก
bright glacier
so I can just focus my testing on that which should cut down the testing significantly
plain atlas
Tbf you could probably formulate a workflow, remove concurrency options, and just push your changes and then come back to it once gha has finished running them
Iirc there's already a workflow for this
bright glacier
honestly it would be faster but I don't feel like making it
We don't have one for diff-shades
plain atlas
Ah.
bright glacier
I'm testing the experimental match statement pr fwiw
tribal thistle
Running black test suite on pydis cluster ๐คก
Was just gonna run it on my laptop. Itโs pretty good, could probably crunch it out quicker
plain atlas
Well, I'm going to go continue to shoot my future self in the foot while I procrastinate
bright glacier
Yeah #dualCoreCPULife
plain atlas
2Core2Life
tribal thistle
Is the code youโre testing pushed to a repo somewhere?
bright glacier
yea, ichard26/diff-shades although it's non functional on GitHub, I have a few hotfixes locally I haven't pushed since I don't know if they're contributing to the weird slowdown
bright glacier
also it's pretty involved to repeat the run I got set up right now and since I'm eating (what else am I going to do) so yeah
tribal thistle
It isnโt, we removed the command
tribal thistle
Idk then
bright glacier
.gh help
toxic stormBOT
Command Help
plain atlas
Its locked
bright glacier
.gh ichard26/diff-shades
plain atlas
;-;
bright glacier
.gh repo ichard26/diff-shades
tribal thistle
Uhh so where does analyze clone to lol
Also, on current main, 30 seconds in for a little above 50% on django
bright glacier
a temporary directory
tribal thistle
Holy mother of deprecation warnings
bright glacier
also this isn't that useful as I'm testing a pr of black
plain atlas
Holy mother of deprecation warnings
just set PYTHONWARNINGS to ignore:::
bright glacier
Holy mother of deprecation warnings
yea locally I slapped a redirect_stderr context manager but I don't know if that's causing the slow down
bright glacier
just set PYTHONWARNINGS to ignore:::
it's not using the warnings API
tribal thistle
Alright nice, total analysis time was 4:06, and a minute clone time (and it only crashed once! ๐ )
bright glacier
Also, on current main, 30 seconds in for a little above 50% on django
How fast is your laptop? ๐
plain atlas
Inb4 16
bright glacier
nah 64
tribal thistle
4, but is this actually utilizing more than one lol
bright glacier
yea it's multicored
plain atlas
Actually, the sun is weaker than scale's legs with that much heat
bright glacier
wth ๐
tribal thistle
Dude. Laptop connected to desk monitor. Ultimate setup
plain atlas
Smh I don't use my laptop on legs either smh mh smh
tribal thistle
Was this meant to reformat like 18 of the 23 projects lol
bright glacier
I don't remember but that sounds close enough to right
plain atlas
@tribal thistle how fast is your internet?
bright glacier
Lemme get back to my computer
tribal thistle
<@386599446571384843> how fast is your internet?
Uni internet goes brrr
bright glacier
25/2 for me ๐ฆ
plain atlas
Although I'd hope it's shallow cloning
tribal thistle
25/2 black/clone time?
bright glacier
It is
plain atlas
Also I would hope it clones to a cache dir instead of tempfile smh
bright glacier
25/2 black/clone time?
No, down/up Mbps
tribal thistle
Ah
Lol I got so lost
This is shaping up really nicely so far btw
Just gotta suppress all them pesky warnings
plain atlas
Oh yeah I was meaning to rethink my life choices and read more documentation
bright glacier
Also I would hope it clones to a cache dir instead of tempfile smh
That's what the --work-dir is for
plain atlas
I have written way too much extra code when apparently it's part of the stdlib, or even worse, part of builtins...
plain atlas
Bye
bright glacier
alright I'm back on my laptop
bright glacier
Itโs just a 100mb json
mostly just input and output of the formatting
tribal thistle
Work on my code instead
bright glacier
lol
plain atlas
alright I'm back on my laptop
I've been on mobile this whole time
late dewBOT
tribal thistle
Lol
plain atlas
I definitely do not have a few classes which exist to just make a generation methods iterators
I really don't have them, nope. None
bright glacier
If you want to rerun diff-shades like how I did it then read the following:
- clone https://github.com/psf/black/pull/2586 and install it
- set up diff-shades (I presume this is already done) in the same env
- download the analysis provided here: https://github.com/ichard26/black-mypyc-wheels/issues/1#issuecomment-955093775
- unpack it
- run
diff-shades analyze plz-gimme-a-reasonable-name.json --repeat-projects-from my-analysis-you-just-unzipped.json
the reason why you need that analysis file is so diff-shades clones the right revision of the projects
plain atlas
I really don't have them, nope. None
class GenerateID:
def __next__(self):
return generate_id()
so when comparing analyses only differences in the formatting output will be caught
bright glacier
This is shaping up really nicely so far btw
thanks for the compliment! hopefully the rest of the core team agrees since I think this will be quite helpful as we stabilize black :)
plain atlas
shit sorry for interrupting
plain atlas
As apology I'll go review some code 
bright glacier
a run against a mainline version of black should see zero failures
assuming that diff-shades isn't borked, this is bad news for isidentical's match PR
plain atlas
Hmmmm
bright glacier
wait what
bright glacier
uhhhhhh this isn't a normal failure mode for black
plain atlas
I haven't looked at the code at all
plain atlas
a run against a mainline version of black should see zero failures
This is only if the projects checked out arent changed between runs
bright glacier
two things
by failures I mean black erroring
and --repeat-projects-from exists for this very reason
plain atlas
uhhhhhh this isn't a normal failure mode for black
I think either there's too many open file pointers as once or file permissions got messed up, not sure tho
plain atlas
Damn
tribal thistle
Taking significantly longer, I wonder if my internet is just bad. Two minutes so far
plain atlas
My computer runs basically at 100% all the time
bright glacier
no, I think it's because I wrote this stupid code
import contextlib
import os
with open(os.devnull) as blackhole:
with contextlib.redirect_stderr(blackhole):
black.format_file_contents
I forgot the "w" mode argument :(
plain atlas
Ahhhhhh
bright glacier
so this entire analysis is junk anyway
plain atlas
Wait why stderr and not stdout?
bright glacier
because that's where the deprecation message is printed to
tribal thistle
Where does devnul go on windows
bright glacier
Taking significantly longer, I wonder if my internet is just bad. Two minutes so...
btw I can push my (fixed) output suppress code so you can rerun diff-shades without a bunch of deprecation msgs
bright glacier
Where does devnul go on windows
the NUL device
bright glacier
to be fair the analysis I gave you is rather old so the cloning is much more expensive as it can't just do a simple shallow clone
tribal thistle
Ah, I see
plain atlas
because that's where the deprecation message is printed to
Ah.
bright glacier
you should probably use the --work-dir option btw so reruns don't have to clone as much
plain atlas
I wonder....
I feel like the deprecation message could use the warning api
And then in the main file
bright glacier
you were literally here when we discussed this
plain atlas
Configure the warnings to be printed
Yes but I explored more with the warning api since then smh
!d warnings
late dewBOT
plain atlas
Tldr I learned this was a thing
tribal thistle
nice, I'll check it out once cloning is done
bright glacier
in the mean while I'm going to check pandas by itself
via the -s / --select option
I know, I really tried to make this a half-decent tool
plain atlas
Using click?
bright glacier
nice, I'll check it out once cloning is done
once you're done could you upload your analysis somewhere?
plain atlas
you were literally here when we discussed this
Well yeah, but if warnings can be configured to print them I don't see a problem with using it now lol
tribal thistle
Yeah sure
bright glacier
I didn't take the time to implement decent diffing capabilities across analyses yet
right now the compare command is literally a binary "no changes found!" or "differences found!" operation
fwiw diff-shades is almost 1000 lines of python code already
well OK it appears that diff-shades (in particular the monkeypatching of sys.stdout / sys.stderr) isn't the one to blame
it's progressing at a much more reasonable rate with a pypi version of black
tribal thistle
Hm, that's what I used my first time around, just pip install black
bright glacier
although to be fair I did change out the output suppression code to a non broken one
yeah that super slow run was with the match pr i LINKED ABOVE
WHOOPS
tribal thistle
Mmm gobbling up my Keyboard interrupts lol
bright glacier
kids, don't type with intense lag
i didn't realize I pressed caplocks until it was too late
sometimes using a compute heavy multi-core program can be annoying
tribal thistle
But yeah, significant speed loss on that PR from what little I've managed to get running
bright glacier
that's not good
inb4 a war erupts between people who want performance vs people who want match support /j
I wonder why it's so slow
tribal thistle
(btw suppression seems to still be broken on the latest push, as well as... most output)
It's just outputting it for everything now
tribal thistle
bright glacier
there's two main ways it could be slower, 1) parsing takes way longer as there's now a proxy in the heart of the parser to handle soft keywords, 2) the match support tweaked the formatting -> causing second passes -> slowdown
bright glacier
wtf
tribal thistle
It might not actually be slower, it seems to have actually crashed the first time around
just froze
plain atlas
it's progressing at a much more reasonable rate with a pypi version of black
Why would that even have a difference??
tribal thistle
i'm 80% through django in 1:50
bright glacier
Why would that even have a difference??
because the match PR basically rewrites the core of our parser
plain atlas
So main and pypi shouldn't have much of a difference, should they?
bright glacier
is this a windows thing?
tribal thistle
Maybe? It had the normal nice output previously, but a similar output when it hit the projects with deprecation warnings
Is the order different this time
Btw, it's back to normal now
After all the deprecation warnings have subsided
plain atlas
Did you resize the window?
bright glacier
@contextmanager
def suppress_output() -> Iterator:
with open(os.devnull, "w", encoding="utf-8") as blackhole:
with redirect_stdout(blackhole), redirect_stderr(blackhole):
yield
this is how I'm suppressing the output which works on linux
src = path.read_text("utf8")
try:
with suppress_output():
dst = black.format_file_contents(src, fast=False, mode=mode)
except black.NothingChanged:
return NothingChangedResult(src=src)
and at the callsite
plain atlas
That's a generator, not an iterator smh
tribal thistle
Did you resize the window?
It was broken before resizing. It seems redrawing the progress bar doesn't flush the line, and when the python warnings are going, that causes a
mess
bright glacier
are you sure the version of diff-shades you're using has the supress_output call in analysis.py:check_file ?
because this suppression worked on windows for a different funtion
tribal thistle
Yeah, I have the files open in the background ๐
I can test again once this run is finished
Which project throws it
tribal thistle
I'll check it out in a bit then
btw, we're up at 7m@40% now, so slower
But not too much
Wait no, it seems to have hung Just a really slow files
bright glacier
it probably hasn't (hopefully)
the chunksize I passed to multiprocessing.Pool isn't great but I haven't seen any major slowdowns from it yet
only chunk-y progress
tribal thistle
Does pandas just have files with like a million lines
bright glacier
yeah its files are stupidly large
tribal thistle
It feels like each file is taking well over a minute at this point
bright glacier
hmm maybe the slowdown I experienced wasn't just me
bright glacier
I'm really frustrated the suppression isn't working on windows fwiw
bright glacier
uhhhhh
bright glacier
what the hell is my code doing
tribal thistle
So it seems to have finished, got stuck during some transition, and locked the terminal
A few button presses ougtha convince it
Yeah that did it
plain atlas
@tribal thistle how is your cpu process so low smh ;-;
bright glacier
well if it happened the first time then we can't blame my suppression code as it didn't exist on github when you first ran diff-shades
tribal thistle
It was 100% for 10 minutes straight...
bright glacier
I doubt it's a rich bug
tribal thistle
Back @100 now
tribal thistle
if it's possible, do you think you could get the progress bar to display LOCs instead of files? It isn't a very accurate metric lol
60% after running pandas and django is uh...
bright glacier
yeah there's a huge difference between typeshed and django for instance
even tho by file count they're similar IIRC?
tribal thistle
hard to tell now, that output has long scrolled out of my terminal backlog
tribal thistle
I wonder if the package just wasn't installed properly
Tbf I didn't use editable, just ran a reinstall
gonna try to fix it, we're almost done
plain atlas
hard to tell now, that output has long scrolled out of my terminal backlog
that reminds me of when something was going so wrong.... i'm not sure why i didn't pipe the output LOL
bright glacier
forcing colour mode in rich on gha windows is not a good idea btw lol
plain atlas
i ended up setting my terminal to give me an extra 1.5k lines of scrollback buffer
should've just piped it
bright glacier
I have my terminal set up for 15000 lines of scrollback IIRC
plain atlas
2.5k here
but that's also an integrated terminal in vscode
idk how long windows terminal is configured for
bright glacier
I- was mistaken ๐
tribal thistle
So i was wrong
2380 (typeshed) vs 2860 for django
It hadn't scrolled away, it just hadn't started
bright glacier
yeah it's in alphabetical order
bright glacier
plain atlas
wait which 5 projects are already formatted with black?
bright glacier
this is a run with 21.9b0 IIRC
so wait the match PR introduces no failures and is only like 4 times slower?
tribal thistle
Uhh... I just did a very stupid thing
Accidentally overwrote the output when testing out the deprecation warnings ๐ฅฒ
bright glacier
oops
so wait is the suppression code borked for you?
because in that case I have no idea how I'm supposed to fix it
other than monkeypatching black.output.err but that won't always work IIRC right now
tribal thistle
It seems to be working now, just appears that running install without uninstalling didn't actually work
bright glacier
probably because I don't always change the version
actually wait no
isn't pip supposed to reinstall regardless of the version if you pass a local dir
plain atlas
i think uh force may be needed and it might not remove all files
tribal thistle
Running everything again now to regen the output, but we're back to clean output again
nope, neverind
broken again
bright glacier
...
tribal thistle
seems to have worked on attrs, but it's breaking on
seems to be django
Yeah that does it
bright glacier
i'm so annoyed
bright glacier
*sigh*
i can't even reproduce the python two deprecation warnings having removed the suppression code
why is everything not going to plan
tribal thistle
because we are programmers
bright glacier
Yeah.
tribal thistle
Is there a problem with the warnings themselves
bright glacier
oh huh interesting 21.10b0 doesn't warn on attrs but the match PR does
I have no idea why
uhhh that's even more confusing
oh right I know why
because my python 2 deprecation code was sketchy as heck
bright glacier
No, attrs is broken again now
let's just ignore which projects the deprecation messages are on because the logic in black is currently broken
is the output suppression code working?
tribal thistle
Which output does it supress
bright glacier
the deprecation warnings
tribal thistle
I'm seeing them right now, so probably not
bright glacier
great /s
bright glacier
I guess mixing click.secho, contextlib.redirect_std*, multiprocessing, and rich together is breaking (on) windows ๐
I guess I'll have to boot up windows tomorrow
I mean I guess I could run black in a subprocess but then I'd get rid of my fancy progress bar
bright glacier
It worked at first, then it stopped
Honestly knowing myself, all of my code is broken on windows somehow
bright glacier
I gave up on diff-shades
tribal thistle
Time to upload it before I destroy it
bright glacier
I don't need the analysis
tribal thistle
oh lol
bright glacier
because the PR is broken anyway
tribal thistle
What was the PR doing anyways
bright glacier
match support :)
tribal thistle
mmm
bright glacier
I don't understand how redirecting stderr doesn't work on windows and yet it does on Linux (probably due to multiprocessing I suppose)
But oh well I'm too frustrated with all of this diff-shades work to continue work on it for the time being
plain atlas
i probably should have timed this
i decided to format my project's venv for the fun of it
i wonder what will be broken when done lmfao
oh shit, aside from all of the packages, i also have to get through the stdlib
ooo
@bright glacier if you want a challenge get black to format pyrepl
!pypi pyrepl
late dewBOT
plain atlas
of my very long time to format my entire venv for no reason other than write cycles and boredom
1998 files reformatted, 584 files left unchanged, 4 files failed to reformat.
those four files are from pyrepl
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/historical_reader.py
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pymdownx/magiclink.py
error: cannot format /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/keymaps.py: cannot use --safe with this file; failed to parse source file.
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/module_lister.py
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/keymap.py
error: cannot format /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/pygame_console.py: cannot use --safe with this file; failed to parse source file.
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pymdownx/superfences.py
error: cannot format /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/pygame_keymap.py: cannot use --safe with this file; failed to parse source file.
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/simple_interact.py
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/trace.py
reformatted /home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/pyrepl/python_reader.py
@tribal thistle what the heck on my windows install it works totally fine
I hardcoded black to always print the deprecated message and yet it doesn't leak ๐
also man the new windows terminal is wayy sexier
can I now blame your system for being weird ๐ or is there another edge case where it breaks ๐
plain atlas
lol reformatting everything actually broke coverage
bright glacier
uhhhhh like the tool or the stat
plain atlas
the tool
bright glacier
should I be concened?
plain atlas
(i did also run isort)
ImportError: cannot import name '__version__' from partially initialized module 'coverage' (most likely due to a circular import) (/home/aru/.cache/pypoetry/virtualenvs/modmail-lmlBH83T-py3.8/lib/python3.8/site-packages/coverage/__init__.py)
bright glacier
I doubt black caused this
plain atlas
me either
bright glacier
but yeah this is why I'm strongly against black reordering imports
plain atlas
i'm reinstalling coverage and just running black on it
one minute
uhhhhhh
I'm going to create a new venv and just format coverage ๐
plain atlas
coverage seems like the only thing that broke with isort formatting
bright glacier
inb4 formatting black with isort breaks it /j
plain atlas
but doesn't seem like anything broke with black formatting
plain atlas
inb4 formatting black with isort breaks it /j
it didn't /srs
the black i used to format coverage was formatted with my isort config
bright glacier
is /srs an actual tone indictor, TIL
plain atlas
yes, its serious
bright glacier
oh of course lol
plain atlas
lmao
bright glacier
i'm used to using /gen as in genuine for the same purpose
plain atlas
what i want is a tone indicator for half serious
eg what if i used difflib to split these functions
half-joking doesn't make sense with that
half serious makes more sense
plain atlas
idfk lmao
bright glacier
uhhh whoops /g /lh
plain atlas
what is /g and /lh smh
bright glacier
/g = genuine | /lh = lighthearted
late dewBOT
You are not allowed to use that command here. Please use the #bot-commands channel instead.
bright glacier
!charinfo ih
late dewBOT
bright glacier
you should become staff for dem commands \hj
plain atlas
perhaps someday ๐
plain atlas
ooh i don't think i'll get another chance to use this tone indicator
someday i will break black /th
/th threat
bright glacier
I have a list of tone indicators with me /li
plain atlas
๐
smh
lol
honestly at some point there's too many tone indicators imo
like
some of them are great but at some point the ones that aren't used as much.... idk
I think the concept is great tho
and i would like to see.....
moved to #ot2-the-original-pubsta message
tribal thistle
@bright glacier The first run after the update did actually work without deprecation warnings
But then I ran a full run, and that broke it
It broke for both individual and full runs
gleaming grail
What is this thread for?
little parrot
black - The uncompromising Python code formatter
bright glacier
It broke for both individual and full runs
what-
bright glacier
<@!429445826251587616> The first run after the update did actually work without ...
do you have any ideas or am I going to have just try various things until it works on your machine TM
tribal thistle
Lol
I can do some debugging when I get home
Iโll try to get a solid repro method
flat krakenBOT
bright glacier
Various ideas I have include: moving where I patch sys.stderr, redirecting at the file descriptor level, using a in-memory blackhole, and monkey patching click and black.output utilities... one of these has to work ๐
tribal thistle
All 3 when
bright glacier
if you machine is a certified application stress tester which it seems to be already
errant barn
Well naturally we'll answer: Yes of course it is ๐ saves you a heap of trouble especially when working in a team
In terms of learning curve, it doesn't really have one. Once you've set it up, calling it is as easy as it gets
mint barn
oh , sounds promising
i'm just a bit confused on what it actually does
read about it before but what i got in a nutshell is that it makes ur code look better
errant barn
Have a look at the pinned messages!
yeah exactly that
doesn't change it in any other way, just visually
mint barn
sounds great , does it follow any specify guidelines like pep8 or any other ?
errant barn
it's a strict version of pep8
mint barn
fair
errant barn
opinionated though, so you won't be able to use your style
mint barn
is ambv the same person in this video https://www.youtube.com/watch?v=Xbl7XjFYsN4&list=PLhNSoGM2ik6SIkVGXWBwerucXjgP1rHmB
errant barn
same face
mint barn
opinionated though, so you won't be able to use *your* style
wouldn't matter if that style looks better than mine , its mostly an opinion but somethings are surely more readable than others
tribal thistle
A more reliable way to quit diff-shades would be uhh
great lol
@bright glacier can you replicate the fix working on the parser rewrite branch? Switching the black version to main seems to have solved the problems with the deprecation warnings
bright glacier
A more reliable way to quit diff-shades would be uhh
I guess this involves installing a signal handler?
bright glacier
<@!429445826251587616> can you replicate the fix working on the parser rewrite b...
that's because the deprecation warnings are misfiring on the parser branch because of my questionable code
bright glacier
<@!429445826251587616> can you replicate the fix working on the parser rewrite b...
I mean, I guess so? Instead of using the buggy branch I just hard coded black to always emit the deprecation warnings
tribal thistle
Ah yes, can confirm