#black-formatter
flat krakenBOT
flat krakenBOT
muted haven
@dense jungle I plan to merge this since there are code in the wild (as the user reported) that use this syntax, and I don't think being too permissive for black is a problem. WDYT?
dense jungle
<@!783088578363523104> I plan to merge this since there are code in the wild (as...
I was actually going to suggest holding off until CPython made a decision on whether to revert the syntax
muted haven
Ah, that is also fine. I am not too sure it will be reverted, but who knows
dense jungle
yeah I'd lean towards not reverting, but Pablo seems to think it's a possibility π
flat krakenBOT
bright glacier
I've been debugging why my power fix PR is changing formatting behavior when it shouldn't for over 40 minutes and I'm no where closer π¦
I know what function call in trans.py causes it but I have no idea what it is doing under the hood causing the weird behaviour
dense jungle
I know what function call in trans.py causes it but I have no idea what it is do...
I noticed the changed cases all had a **kwargs call in them, do you know if that's part of the story?
ruby trench
Shouldn't according to PEP 8 here https://www.python.org/dev/peps/pep-0008/#other-recommendations something like:
hypot2 = x * x + y * y
c = (a + b) * (a - b)
should be formatted to
hypot2 = x*x + y*y
c = (a+b) * (a-b)
But black does it the other way round.
Is this intentional?
lament crow
lament roost
Pre-commit definitely feels faster after bumping black's version to latest
bright glacier
hmm that's weird since doesn't pre-commit clone black instead of using PyPI? and on top of that 22.1.0 has a stability hotfix which reduces formatting throughput by 50%
I suppose the 3.10 parsing improvements might be what is responsible for the speedup you feel since the above is only applicable for changed files thankfully.
lament roost
I'm running the local black through pre-commit instead of the normal one
humble tulip
How is everyone doing? My name is James. This is my favorite channel. God bless, James!
unique mulch
So this is a setting in VS Code, just something to take note of
Maybe something to add to the Editor Integration part of the docs
lament crow
This is a Python server, not a Dart one.
Technically, Black's goal is to let you forget about formatting so the ruler shouldn't really matter as Black will wrap the code appropriately on its own.
flat krakenBOT
unique mulch
~~This is a Python server, not a Dart one.~~
Technically, Black's goal is to let...
I just couldn't be arsed to apply it on Python at the time
flat krakenBOT
bright glacier
looks like we procrastinated for too long on that blackd PR :p /lh
golden swallow
how do i connect black with pycharm
teal goblet
do you have black[d] installed?
unborn lion
How can I avoid deadlock(mysql) in python multithreaded script?
short forge
Not the channel for this, try #databases
teal goblet
Black somehow turned this. ```py
if sys.version_info >= (3, 10):
TypeAlias added in python 3.10
from typing import TypeAlias
else:
from typing_extensions import TypeAlias
Into this. ```py
if sys.version_info >= (3, 10):
peAlias added in python 3.10
from typing import TypeAlias
else:
from typing_extensions import TypeAlias
π What version?
teal goblet
22.1.0
bright glacier
any custom configuration? since I can't reproduce on the playground https://black.vercel.app/?version=stable&state=_Td6WFoAAATm1rRGAgAhARYAAAB0L-Wj4ADkAJhdAD2IimZxl1N_Wlws4TilGXALX_Sht8xBkY7n5b7nLGv2KllUexZZphFWaDAHV8JqyI4IWwi73FWyLMJ6vMmioF2ixolWXt38aB1QOEmLJu_HHR1QnVdQ-dsIdxcaMTMYKgm_weX60fcPC-D9PDyeOV2ufX2FhwAWvsW1acLa7AtQL0zEBH72Awez6fJd7ZVlr-aOAN53JMIAANEeUy5esIo0AAG0AeUBAAAzdw2fscRn-wIAAAAABFla
teal goblet
just changing the line-length to 79 for flake8
bright glacier
still not reproducible hmmm
elder tusk
consider ufmt / usort https://usort.readthedocs.io/en/stable/
it's much less likely to have these kinds of bugs: https://usort.readthedocs.io/en/stable/why.html#comparison-to-isort
dense jungle
Black itself is also unlikely to have this sort of bug for similar reasons
we do an AST safety check that makes sure the output is valid Python equivalent to the source
little parrot
isnt there an atomic mode for isort?
elder tusk
looks like it, seems to just be a check for syntax errors at the end. could've helped OP, but wouldn't have helped for things like https://github.com/PyCQA/isort/issues/1507
flat krakenBOT
errant barn
:D
damn.
gotta close my personal one too
seems that at least not a bot: https://github.com/felix-hilden/tekore/pull/269#issuecomment-1042921282
muted haven
I got the same exact reply on my repo, so who knows maybe it is a bot after all
bright glacier
so they ignore the fact you need permission from all contributors to relicense (which can be covered by the CLA) but sign CLAs...
lament crow
I can't explain it any other way than them being a troll
I know there are some die-hard fans of GPL but this doesn't even contain any reasoning why you would want to change license (not that it would change anything of course lol)
errant barn
they responded with reasons so I'm calling "no troll" :D
teal goblet
Why would you switch to a more restrictive license than MIT?
lament crow
so they ignore the fact you need permission from all contributors to relicense (...
You don't need permission from all contributors if you're going from more permissive license
Since you're just creating derivative work under stricter license which MIT doesn't disallow
Hence why it's permissive :D
tired shard
If anyone got time, would appreciate some reviews on my PRs 
dense jungle
^ um probably not
supple birch
print ('pp')
seems like a little unnecessary, no?
ivory pumice
seems necessary to me cheif
eternal egret
Does the psf/black vim plugin run a version of black sourced from the repo?
the vim plugin is running out of date black despite both my venv and system black being up to date
does something in the repo need to be updated separately for the vim plugin that got missed?
eternal egret
ahah, you need to separately run BlackUpgrade
mint barn
(I know very few things of Python) I just had an idea: can somebody write a programm that makes student notes more eco-friendly? For instance, I have notes on a course but all the pictures and word use a lot of space and inkt. I put it in that programm and is so adjusted that nothing will be deleted, but the best usage of space and as little inkt as possible is used. Still the notes are comprehensive etc. Can somebody make that??
glacial drum
(I know very few things of Python) I just had an idea: can somebody write a prog...
This isn't the channel for general Python questions. Take a read of #βο½how-to-get-help to see how you can get help with your question.
flat krakenBOT
topaz rapids
Hi
flat krakenBOT
flat krakenBOT
flat krakenBOT
bright glacier
I think I'm beginning to understand why the power op comment PR is changing formatting although it's still unclear what the root cause is
β― black test2.py --check --diff --color -l 79 --fast
transforming line
running transformer: <black.linegen.rhs object at 0x7ff5f41136d0>
called trailers
calling rhs via LOOP 0 set()
rhs called
rhs called
aaaaaaaaaaaaaa = 111111111111111111111111111111111111111111111111111111111111 + (
-- a ** a
-- )
calling rhs via BEST EFFORT
rhs called
rhs called
aaaaaaaaaaaaaa = 111111111111111111111111111111111111111111111111111111111111 + (
-- a ** a
-- )
transforming line
running transformer: <black.linegen.rhs object at 0x7ff5f4dc0310>
called trailers
calling rhs via LOOP 0 set()
rhs called
<black.linegen.rhs object at 0x7ff5f4dc0310> No brackets found
running transformer: <function hug_power_op at 0x7ff5f3f1fdc0>
<function hug_power_op at 0x7ff5f3f1fdc0> No doublestar token was found in the line.
transforming line
always fun debugging like this :P
thick hazel
```
β― black test2.py --check --diff --color -l 79 --fast
transforming line
runni...
imma be honest i dont understand a thing here
bright glacier
Well yeah I just added a handful print calls to figure out what's going wrong in a patch I'm currently working on :)
flat krakenBOT
celest sorrel
um guys
as a joke i made my black format to line length 5
and when i try and change it back to 120 its all kind of...fucked
is there a way to force 120 for everything? o.O
like
im scarred
from __future__ import (
annotations,
)
dull olive
checkout your previous commit?
celest sorrel
im literally gonna cry
celest sorrel
checkout your previous commit?
i cantttt, ive been working on this soduku for the past 4 hours
is there a formatter that forces a line length of 120?
bright glacier
let this be a lesson in don't run any code rewriter unless you're OK with the changes being not undoable ^^
celest sorrel
π
bright glacier
unless you're using a SVC tool like git
flat krakenBOT
bright glacier
@dense jungle I do have a work in progress patch to update diff-shades to force --preview, I can probably get it PRed soon
(I've just been having fun writing an "optimization" patch for mypyc and it was so engaging I ignored basically all notifications 
)
dense jungle
<@!783088578363523104> I do have a work in progress patch to update diff-shades ...
sounds great, that would be good for the docstrings PR
and have fun optimizing mypyc π
bright glacier
the C compiler is still mad at me so I got some more work left but that doesn't sound too bad.
bright glacier
Yup, that's why I said it was "in-progress", I still had to write the other workflow asserting stability
dense jungle
ah right π
I guess in that workflow we should skip projects like pyanalyze that turn on preview style
bright glacier
I was a little discouraged from writing it since a) it's probably a fair bit of work, and b) it involves changes to diff-shades itself since there's no way to force preview=False easily
... or yeah we could just skip them I guess
dense jungle
I think it's probably only pyanalyze, at least for now
bright glacier
hmmm, the diff-shades baseline run shouldn't be taking 8+ minutes. I wondered if I messed something up.
oh wait, I forgot I'm passing --preview ... anyway there's a crash
dense jungle
nice catch, I guess it gets confused about the comment in the middle of an implicit concatenation
bright glacier
No blockchain is involved here, we instead like to play with tree structures π
bright glacier
there is blockchain code here
If you're wondering whether this server has a blockchain / crypto / web3 channel, the answer is no. It's been asked many times in #community-meta but we haven't seen enough interest to warrant a whole new channel.
mint barn
What is this project about?
silent apex
What is this project about?
It's an auto-formatter for Python code.
mint barn
Wow thatβs awesome. Is it similar to HTML one where you press control shift f (or something) and it formats it to the right indexes.
silent apex
erm, it's often run from the command line, but a lot of editors have extensions to run it on save
mint barn
Sweet, great work!
silent apex
I've done nothing for it haha, thank all these wonderful people in purple
foggy trout
Oh, well never mind lol
silent apex
make it look like a label
If you just put the link https://github.com/psf/black/labels/F%3A%20linetoolong, github is smart and autoformats it. There's probably another way, but that's one I know of.
safe fog
civic plume
versed trout
Hi everyone, i want to ask u if anyone build befor a recommender system using ML ?
lament crow
@civic plume, @versed trout this is not a proper channel for this. Read the channel topic.
loud nebula
does anyone wanna play minecraft with me
bright glacier
!ot - A little off-topic for this channel :) perhaps ask in a OT channel?
late dewBOT
lusty crag
Sup, I'm working on a legacy python project and naturally it uses every python code style under the sun. I'm trying to introduce black in order to standardize it but there is quite a bit of resistance from the old devs. Chiefly they don't like:
""" Foo """becomes"""Foo"""foo(very_long_line)becomesfoo(\n\tvery_long_line\n)'this string'becomes"this string"- no alignment on "=" for variable assignments which are right next to each other
can I change these settings in black? From the docs it seems that fiddlin with small stuff like this is not really supported, am I out of luck/need to look to another automatic code formatter?
jikes, sorry for the formatting
lusty crag
Ok, thanks for the response π
ancient depot
Good morning my name is Raj Mukherjee and I am new to this community. I am pursuing my B.Tech degree from Srm University in artificial intelligence and I am working on python for quite a few years and I have also started working on Data Science, Artificial Intelligence and Machine Learning so as per my interest I would like help and contribute something to this community
south ridge
Good morning my name is Raj Mukherjee and I am new to this community. I am pursu...
You should checkout the github repo for this project
ancient depot
Ok i will check it out
bright glacier
wat is CI doing ...
I wonder if this is just that edge case where the fact the PR isn't tested with main merged in
flat krakenBOT
bright glacier
hmmm I don't really like this furo update
concretely I feel like the headings are now too bold, I could see how they could've seemed too dim previously but this feels like the other side of the spectrum π
Also, it's not that clear IMO that 22.1.0 is a higher heading than Highlights
https://black.readthedocs.io/en/latest/faq.html feels very shout-y now
fossil estuary
lament crow
Also, it's not that clear IMO that 22.1.0 is a higher heading than Highlights
honestly, I always found difference between different levels of headings in furo to be hard to notice sometimes
flat krakenBOT
flat krakenBOT
lusty oasis
I want to customize black by cmd
plain atlas
honestly, I always found difference between different levels of headings in furo...
I don't know why they switched to furo
I've found furo beautiful but also not beautiful
flat krakenBOT
dense jungle
the windows docs build is real slow
radiant iris
HiImManh#2550
everybody add this account to the dead toxinloght
beacause he make a lot of thing that make problem
flat krakenBOT
bright glacier
Given the initial negative feedback I'm inclined to reject the PR from a birds-eye view.
I haven't actually looked at the PR in detail yet.
dense jungle
what negative feedback? the downarrow emoji?
I think I mostly like it now, it really isn't much worse than the docstring processing we already do
I'm mostly scared of the backlash π
bright glacier
Yeah.. and I'm not particularly fond adding another option for this. I suspect there isn't enough consensus to make this feasible longterm.
dense jungle
what we choose creates the consensus for how Python is formatted
I'm mostly not joking
bright glacier
Yeah although I'm acutely aware of the issues that will get opened when there's any room for disagreement in style. I guess that's unavoidable though regardless if we change the style or not.
dense jungle
just means we get to close more
little parrot
what negative feedback? the downarrow emoji?
The reason they downvoted it was they use a a doc string format like """|coro
something something
"""" and this expand that a lot more
errant barn
Do reckon adding an option for docstring processing or more broadly AST safety would be better?
dense jungle
Do reckon adding an option for docstring processing or more broadly AST safety w...
isn't the del parens thing the only other AST-unsafe thing we do? I don't really want an option for that
errant barn
psf/black#212 would be another one, but I'm totally ok with just the docstring thing
toxic stormBOT
errant barn
but that is also another one where an option for it would be odd
at least with docstrings it makes some sense
dense jungle
yes, I don't think an option for that would be necessary, unless we can really find a case where it makes a difference in practice
Docstrings are actually visible at runtime and there are lots of tools that parse them, so it's more plausible that reformatting them will break something
topaz rapids
What are you all talking about
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
dense jungle
we should probably do a new release soon
bright glacier
Yeah, I would've brought it up sooner but .. I'm still the blocker. The mypyc build workflow is still not up. Given my terrible mental state it would probably be better to just do it manually again, maybe even asynchronously for this next release.
Thanks for holding down the fort while I've been checked out by the way, it's nice not having to stare at unreplied-to emails.
dense jungle
no worries! take care of yourself
neon loom
Yeah, personal life first mate. Can we/I help with any of the wheel build flow?
flint cedar
why does black turn
SomeClass(
param=(
"Lorem ipsum dolor sit amet"
"I forgot the other parts"
)
)
into
SomeClass(
param=(
"Lorem ipsum dolor sit amet" "I forgot the other parts"
)
)
because we haven't turned --experimental-string-processing on by default yet π
flint cedar
so i can just turn that on and it will leave the string alone?
it doesn't, it joins the two together into one string
dense jungle
yes, if it fits in one line
dense jungle
default line length is 88
crude blade
is black first or third party?
errant barn
first to us π
neon loom
What does that even mean?
bright glacier
we're not affiliated with the PSF or CPython project other than the former giving us some administrative help
We aren't the official formatter for Python if that's what you were asking
There's no such thing
lunar nova
anyone using black in neovim here?
I have the plugin working, but formatting using = still formats differently
bright glacier
Today's new furo release should fix the heading typography issues -- I'mma test it locally and then merge
Yeah it's a lot better -- still not perfect IMO, but it's actually readable
dense jungle
@bright glacier if you have time, do you know how I could get the anchor links in https://github.com/psf/black/pull/2839 to work?
I've been procrastinating that for a while now π
bright glacier
I've been procrastinating on documentation changes for even longer π
long frigate
(I'm sorry, I open the worst issues...)
dense jungle
(I'm sorry, I open the _worst_ issues...)
can confirm, https://github.com/psf/black/issues/1557
and thanks for reporting this!
long frigate
can confirm, https://github.com/psf/black/issues/1557
Looking at https://github.com/psf/black/issues?q=is%3Aissue+author%3AZac-HD+-label%3A"R%3A+duplicate"+-label%3A"T%3A+documentation", I reckon most of these are pretty gnarly...
dense jungle
https://github.com/psf/black/issues/1012 is pretty nasty π
bright glacier
man, I remember reading this great source on sphinx refs, just gotta find it
bright glacier
I feel like https://github.com/psf/black/runs/4999187973?check_suite_focus=true is a bug the more I look into it
Yes. _Black_ is strictly about formatting, nothing else. Black strives to ensure that
after formatting the AST is
[checked](the_black_code_style/current_style.md#ast-before-and-after-formatting) with
limited special cases where the code is allowed to differ. If issues are found, an error
is raised and the file is left untouched. Magical comments that influence linters and
other tools, such as `# noqa`, may be moved by _Black_. See below for more details.
it works here?
I wonder if we could just sidestep the issue by folding usage and configuration into a single document
Anyway I'mma see if I can create a reproducer because I have no idea what's wrong here
!remind 3H documentation is hard
late dewBOT
late dewBOT
!remind 3H documentation is hard
It has arrived!
mint barn
!remind 100Y 
late dewBOT
late dewBOT
mint barn
!remind delete 4187
late dewBOT
There does not seem to be anything matching your query.
mint barn
!remind 1s Oo
late dewBOT
mint barn
errant barn
please don't spam this channel (or any channel for that matter)
mint barn
oh
bright glacier
#bot-commands exists for a reason @mint barn ^^
proper rampart
did someone face this issue before?
proper rampart
??
little parrot
Do python3.10 -m black .
teal goblet
you may need to activate a virtualenv if you have it installed in one.
proper rampart
first pip doesn't work for
mac@MBP-de-mac site-packages % pip --version
zsh: command not found: pip
I did pip3 and it worked for me as you can see
mac@MBP-de-mac site-packages % pip3 --version
pip 22.0.4 from /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages/pip (python 3.10)
when i do pip install black doesn't work for
mac@MBP-de-mac site-packages % pip install black
zsh: command not found: pip
With this one i did pip3 also as you can see
mac@MBP-de-mac site-packages % pip3 install black
Requirement already satisfied: black in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (22.1.0)
Requirement already satisfied: mypy-extensions>=0.4.3 in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (from black) (0.4.3)
Requirement already satisfied: pathspec>=0.9.0 in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (from black) (0.9.0)
Requirement already satisfied: click>=8.0.0 in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (from black) (8.0.4)
Requirement already satisfied: tomli>=1.1.0 in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (from black) (2.0.1)
Requirement already satisfied: platformdirs>=2 in /Library/Frameworks/Python.framework/Versions/3.10/lib/python3.10/site-packages (from black) (2.5.1)
But I can't work with black i don't know why
mac@MBP-de-mac web3_py_simple_storage % black .
zsh: command not found: black
Please can someone help me
little parrot
Do python3.10 -m black .
@proper rampart
proper rampart
i did it but the black option doesn't work
little parrot
Wdym?
proper rampart
Wdym?
let me show a example
as you can see here
this is the problem that i face
little parrot
Oh idk what you'd need to do to fix it in vsc
flat krakenBOT
flat krakenBOT
bright glacier
Nope, it's all unofficial still. We want to define one eventually, but it's been low priority for time being.
glacial swan
ah alr
I am still doing my formatter function thing
Can't wait to do it officailly π
tawdry sorrel
HEY
flat krakenBOT
main flume
tired shard
Ok my exams are almost over, one remaining, will resume work on my PRs from Friday 
And maybe open some more ;)
bright glacier
!remind push diff-shades changes..?
late dewBOT
bright glacier
!remind new 2H push diff-shades changes..?
late dewBOT
late dewBOT
!remind new 2H push diff-shades changes..?
It has arrived!
mint barn
@main flume This channel is for discussions about the Black formatter - please try out commands in #bot-commands next time
hexed tangle
@hallow bane Do not spam the eval command in random channels. Please see #bot-commands for that
errant barn
Please observe the channel topic
hallow bane
K
bright glacier
!remind 50d can we test CPython 3.11-dev on GHA yet?
late dewBOT
bright glacier
that's roughly when the first beta should be available, currently yarl and aiohttp fail to compile due to a header file being deleted / moved upstream
topaz rapids
Hi
tired shard
@dense jungle psf/black#2525 should be ready for review now, sorry for the delay
toxic stormBOT
tired shard
So is psf/black#2770
tired shard
(all are ready)
dense jungle
ok will take a look in the next few days
bright glacier
Currently working on rewriting the diff-shades integration to add a job that forces the stable style and fails on changes.
Unfortunately it's a bit of a mess since it will require changes on diff-shades' end ... and the CI is still failing due to that ineffective monkeypatch.
Also this adds parallelization AND a pre-configuration job to the workflow so I think I'm going to be the only one capable of maintaining this for the forseeable future...
I guess more documentation that details how it works couldn't hurt.
I realize we haven't used the workflow dispatch feature of the workflow and it continues to be a thorn when adding features since often it doesn't need the same level of automation as pushes and PRs. It might be worth removing since I can't really think of a situation we'd use it anyway.
And even if it was useful, it's 100% feasible to do the same kind of thing locally, it's just more effort which is probably fine for such a rare usecase. It was easy to implement within the original requirements, but it's not worth it anymore.
Phew, I wasn't too indecisive this time.
dense jungle
good call!
bright glacier
the only minor issue is that the PRs we do make that stabilizes any preview formatting changes will be red thanks to diff-shades but that seems fine
at most it'll be a yearly occurrence done by a maintainer who should know what's up
tired shard
have a label to decide whether we want to run preview diff-shdaes or non preview?
bright glacier
the planned design runs both, one to preview changes and one to fail on stable changes
and I don't think it's worth keeping a label for these one off events
tired shard
yeah, I guessed that, but you said it would be red for preview formatting changes, so we can decide on which of the diff to follow from that? if you get what i mean
tired shard
oh ok, i read that wrong. my bad
bright glacier
the only minor issue is that the PRs we do make that stabilizes any preview formatting changes will be red thanks to diff-shades but that seems fine
bright glacier
oh ok, i read that wrong. my bad
no worries
tired shard
ok, are we running preview formatting or just the stable formatting on black?
i see the tests are failing when i include preview formatting changes in the source
dense jungle
I changed it recently to format Black itself with preview style
tired shard
even the tests?
like the test to check if source is formatted
ok nope
should i include that in 2770 or make a new one, i belive thats something we want as we are formatting with preview style
otherwise the tests would be fialing
dunno why that doesn't happen now with main branch 
bright glacier
oh hmm I wonder the test isolation recently added is overriding the preview=True in pyproject.toml
tired shard
uhhh which test isolation? could you point to the commit
bright glacier
like adding --config=tests/empty.toml everywhere
tired shard
ah, yeah that's it
bright glacier
Hmm actually I doubt it since the test checks per file and that would skip the configuration file anyway
I'm AFK tho so I can't check
tired shard
ok, i don't think we are adding empty config for the test_source_is_formated
can't see anything in the related code
late dewBOT
tests/test_format.py line 154
check_file(str(path), DEFAULT_MODE, data=False)```there ^
bright glacier
ah default mode, that'd do it
tired shard
Right, one more test is failing of test_piping in BlackTestCase
If the ``.ipynb`` originally had a trailing newline, it'll be preserved.
"""
- trailing_newline = (src_contents[-1] == "\n")
+ trailing_newline = src_contents[-1] == "\n"
modified = False
nb = json.loads(src_contents)
none of the tests are adding --preview
should i just add them by default to BlackRunner or add them spearately to each?
dense jungle
Right, one more test is failing of `test_piping` in `BlackTestCase`
```diff
...
only the tests that test black's own formatting. normal unit tests should run in stable mode
tired shard
even expressions.diff test?
dense jungle
why not?
tired shard
In the feature enum, if want to add anew feature for the detection purpose, what number will i give? And what are these numbers? Can't really understand them from just looking at them
dense jungle
In the feature enum, if want to add anew feature for the detection purpose, what...
I think they're just random
tired shard
ok, so i will just choose the next number for F_STRING_EQEQUAL_DEBUGGING feature, couldn't find a better name for f"{x=}"
dense jungle
ok, so i will just choose the next number for `F_STRING_EQEQUAL_DEBUGGING` featu...
I think "debug f-strings" is the usual term, so I'd suggest DEBUG_F_STRING
tired shard
ok, that's what i had earlier, found the feature names a bit more explicit, so included the operator name
Also, for the detection, I am just using "=" in code_inside_braces_of_string", or there are edge cases to this which i am not aware of?
dense jungle
Also, for the detection, I am just using `"=" in code_inside_braces_of_string"`,...
that seems wrong it would catch, f"{x == y}"
from https://docs.python.org/3/reference/lexical_analysis.html#formatted-string-literals seems like it should be at the very end or before ! or :
tired shard
oh oi forgot about those, yep i think regex would be a good fit then
tired shard
but wouldn't the parsing fail if we have a hanging =?
we can only parse x= in f"{x=]" since we don't have f-string support yet
or we check if the first char isn't ' or "? since if it is not and they are included after a var, then it is invalid syntax
dense jungle
not sure what you mean, f"{x+'=:'}" is also valid
lament crow
:=
dense jungle
you can in fact use walrus inside an f-string ```>>> f"{(x:=3)}"
'3'
x
3
(please don't)
tired shard
oh wow i didn't know tht
bright glacier
!file 18H file some visual output suggestions to rich-click to add more newlines
late dewBOT
bright glacier
ah fantastic
!remind new 18H file some visual output suggestions to rich-click to add more newlines
late dewBOT
bright glacier
It's very nice and I want to add it to diff-shades but it's a young project and has a lot of rough edges imo
also, yeah the lack of typing support is a hard no (fixed on main apparently tho)
flat krakenBOT
lament crow
aww, I looked at this diff, noticed Preview.hex_codes_in_unicode_sequences in self.mode and thought I somehow missed that you can actually enable a single preview feature rather than go all in
errant barn
Well not from the command line though
it's intended just for converting specific features to stable
late dewBOT
src/black/mode.py line 157
def __contains__(self, feature: Preview) -> bool:```I think the reason it's done like this is due to experimental string processing flag but anyhow, I know it's not possible from cli that's why I was sad lol
"aww" was for sad
errant barn
yeah ESP is a special case
tired shard
once esp is merged in? would this still remain or preview would be handled as a single config?
errant barn
will still remain, although the 2 lines specifically for ESP will be deleted of course
lament crow
would be nice if the preview flag was granular, not gonna lie
experimental string processing is a bit much for me while it's still experimental but a feature like the one above would be nice to enable sooner
errant barn
Sorry, it won't be π
lament crow
too many different possibilities of people setting these flags in I guess
bright glacier
bright glacier
I can't reproduce it in a windows terminal -- I bet rich changed something on windows
could also be a windows-specific rich bug, it seems it only occurs when a non-interactive terminal is detected
lament crow
rich 12.0.0 dropped colorama
though it should support legacy windows console on its own now
bright glacier
somehow printing is fully broken on the new terminal too, this makes no sense
I just love windows, it's totally not the source of many bugs over the years :) /s
(notepad++ may or may not be frozen right now thanks to windows update)
calm shuttle
e
bright glacier
yeaa according to the test output it's straight up not printing anything to stdout in a subprocess :/
lament crow
does it only happen in CI or can you repro it locally too
bright glacier
locally too if I'm not using a terminal directly
lament crow
strange
what about when you change command prompt's settings to use legacy console mode
bright glacier
I feel like it must be my fault somewhere since this would be a very bad bug but I have no idea where to start debugging lol
what even ...
i-
@lament crow have you seen something like this?
bright glacier
yup π
lament crow
you got my curiosity up so I'm gonna clone diff shades lol
bright glacier
This makes no sense though since running a rich application through a subprocess with stdout attached to a pipe should be tested on windows
and even if this somehow wasn't tested in their CI, it should've been reported by now!!! but I can't find an issue
lament crow
I can't repro
bright glacier
fun fun fun
lament crow
hmm, try uninstalling colorama
bright glacier
still broken
lament crow
bright glacier
(venv) PS R:\Programming\diff-shades> pip freeze
atomicwrites==1.4.0
attrs==21.4.0
black==22.1.0
click==8.0.4
commonmark==0.9.1
coverage==6.3.2
-e git+https://github.com/ichard26/diff-shades.git@d33bf5ae691f9e702e99a148ba521d207b6e8fb9#egg=diff_shades
iniconfig==1.1.1
mypy-extensions==0.4.3
packaging==21.3
pathspec==0.9.0
platformdirs==2.5.1
pluggy==1.0.0
py==1.11.0
Pygments==2.11.2
pyparsing==3.0.7
pytest==7.1.0
pytest-cov==3.0.0
rich==12.0.0
tomli==2.0.1
typing_extensions==4.1.1
I guess isolated virtualenv time?
bright glacier
hmm?
bright glacier
I'm in a virtualenv, it's diff-shades' tho
lament crow
I don't have 12 there lol
I'm in venv too
but subprocess.run(["python", ... runs the global python
apparently...
but how
lament crow
okay, now I can repro
lament crow
I have no clue why it's using wrong python though,
like, shouldn't subprocess use PATH to determine this...
well, whatever >.<
bright glacier
okay, now I can repro
wait how?
lament crow
by actually using right python by passing sys.executable
cerulean agate
!code
lament crow
my global env's rich is not 12
bright glacier
ah lol
lament crow
I just didn't realise it's using that
bright glacier
I don't have rich in my global env AFAIK
bright glacier
I stop using windows day to day by the time I learned to not install everything in the global env
lament crow
I mean, the thing I have in global env work fine :)
plain atlas
then apt would not work
lament crow
totally have them all installed as admin so that pip can write to C:\Python38\lib\site-packages
plain atlas
so I went to another machine running Linux and copied some of the files
apt works again
bright glacier
I mean on my ubuntu install I gave myself ownership of some files under /opt to make my python installations work lol
bright glacier
Probably worth a bug report because I'm running out of ideas (or I guess we could ping will here, but that sounds rude :) )
unless you have more ideas on how to debug these weird windows issues :p
lament crow
it assumes that a non-terminal is legacy console
and therefore uses Windows APIs to write to console rather than use stdout
lament crow
ye, it only has special handling for first stream
I assume the Windows APIs work that way
bright glacier
I've reached the limit of my rich and windows knowledge so I'm just listening :)
lament crow
I mean, you should report it to rich
bright glacier
I was just waiting on you whether you'd all of sudden discover why :)
I will do so though.
lament crow
I'm pretty sure it's because it can't use WriteConsole API (the one that's used for legacy console printing) when it's piped because output from WriteConsole is not capturable
but I haven't used those APIs so I can't say that for certain
WriteConsole fails if it is used with a standard handle that is redirected to a file
well
bright glacier
(I have never written C code on windows)
lament crow
I guess I could just read documentation then
bright glacier
(I have never written C code on windows)
well that's a slight lie, I have like three years ago? but I've forgotten all of it and anyway I have never worked with windll or the windows APIs
lament crow
I wrote a bit in university but like, not much...
well, not much in terms of the amount of different API functions that I used
because winapi is pretty extensive π
certainly never worked with console apis
bright glacier
import subprocess
import sys
code = """\
import sys
from rich.console import Console
console = Console(file=sys.stdout)
console.print("spam")
"""
err_code = code.replace("sys.stdout", "sys.stderr")
cmd = [sys.executable, "-"]
print("[capture_output=True via sys.stdout]")
proc = subprocess.run(cmd, input=code, capture_output=True, encoding="utf-8")
print(f"{proc.stdout=} {proc.stderr=}")
print("[capture_output=True via sys.stderr]")
proc = subprocess.run(cmd, input=err_code, capture_output=True, encoding="utf-8")
print(f"{proc.stdout=} {proc.stderr=}")
print("[capture_output=False via sys.stdout]")
proc = subprocess.run(cmd, input=code, capture_output=False, encoding="utf-8")
reproducer made
lament crow
seems fine
I would include this excerpt from WriteConsole's documentation:
WriteConsole fails if it is used with a standard handle that is redirected to a file. If an application processes multilingual output that can be redirected, determine whether the output handle is a console handle (one method is to call the GetConsoleMode function and check whether it succeeds). If the handle is a console handle, call WriteConsole. If the handle is not a console handle, the output is redirected and you should call WriteFile to perform the I/O.
https://docs.microsoft.com/en-us/windows/console/writeconsole
bright glacier
Anyway I guess I'll just ban rich==12.0.0 on windows. I need to get diff-shades functional in the time I wait for a response.
late dewBOT
!remind new 18H file some visual output suggestions to rich-click to add more ne...
It has arrived!
thick coyote
!e
bright glacier
This is probably the wrong channel, no?
long frigate
I got annoyed at the way list[int] tempts people to write tuple[int] for their type annotations, and decided to write a linter which forces tuple[int,] if you meant the latter (instead of tuple[int, ...] for variable-length).
Except that Black makes that too ugly to expect much adoption, so...
flat krakenBOT
flat krakenBOT
flat krakenBOT
bright glacier
I promise to get diff-shades ready for post-stable use soon! I'm just following up on the rich bug and then I need to test out the integration patch ^^
OK so bug.py is working correctly on master but diff-shades' test suite is still failing
I guess it's something specific to Console(record=True) now .... 
@lament crow what the hell
import subprocess
import sys
code = """\
import sys
from rich.console import Console
console = Console(file=sys.stdout, record=True)
console.print("spam")
print(f'recorded: {console.export_text()}')
"""
err_code = code.replace("sys.stdout", "sys.stderr")
cmd = [sys.executable, "-"]
print("[capture_output=True via sys.stdout]")
proc = subprocess.run(cmd, input=code, capture_output=True, encoding="utf-8")
print(f"{proc.stdout=} {proc.stderr=}")
print("[capture_output=True via sys.stderr]")
proc = subprocess.run(cmd, input=err_code, capture_output=True, encoding="utf-8")
print(f"{proc.stdout=} {proc.stderr=}")
print("[capture_output=False via sys.stdout]")
proc = subprocess.run(cmd, input=code, capture_output=False, encoding="utf-8")
(venv) PS R:\Programming\diff-shades> python .\bug_record.py
[capture_output=True via sys.stdout]
proc.stdout='spam\n' proc.stderr="Exception ignored in: <_io.TextIOWrapper name='<stdout>' mode='w' encoding='cp1252'>\nOSError: [Errno 9] Bad file descriptor\n"
[capture_output=True via sys.stderr]
proc.stdout='recorded: spam\n\n' proc.stderr='spam\n'
[capture_output=False via sys.stdout]
spam
recorded: spam
and now export_text is just plain broken
(venv) PS R:\Programming\diff-shades> python .\bug_record.py
[capture_output=True via sys.stdout]
proc.stdout='spam\nrecorded: spam\n\n' proc.stderr=''
[capture_output=True via sys.stderr]
proc.stdout='recorded: spam\n\n' proc.stderr='spam\n'
[capture_output=False via sys.stdout]
spam
recorded: spam
running with rich==11.2.0 works as expected
lament crow
report that on the issue
bright glacier
windows. just wonderful
bright glacier
minor correction, the recording stuff works but printing is still acting up
If I switch to using console.print instead of the built in print with 12.0.0 I get this weirdness
(venv) PS R:\Programming\diff-shades> python .\bug_record.py
[capture_output=True via sys.stdout]
proc.stdout='spam\n' proc.stderr='Traceback (most recent call last):\n File "<stdin>", line 6, in <module>\n File "R:\\Programming\\diff-shades\\venv\\lib\\site-packages\\rich\\console.py", line 1642, in print\n self._buffer.extend(new_segments)\n File "R:\\Programming\\diff-shades\\venv\\lib\\site-packages\\rich\\console.py", line 841, in __exit__\n self._exit_buffer()\n File "R:\\Programming\\diff-shades\\venv\\lib\\site-packages\\rich\\console.py", line 799, in _exit_buffer\n self._check_buffer()\n File "R:\\Programming\\diff-shades\\venv\\lib\\site-packages\\rich\\console.py", line 1935, in _check_buffer\n with open(file_no, "w") as output_file:\nOSError: [WinError 6] The handle is invalid\n'
[capture_output=True via sys.stderr]
proc.stdout='' proc.stderr='spam\nrecorded: spam\n\n'
[capture_output=False via sys.stdout]
spam
recorded: spam
although that doesn't explain why the test suite is failing, I feel like I'm missing something
OK actually stuff is still pretty broken
recording, using the built print, and even printing again with console's print are all failing in different ways with STDOUT
OK they've been reported.
bright glacier
hmm, sqlalchemy is still failing, unfortunately diff-shades doesn't let you just ignore errors for a project or file.
the best interim option would be not fail the workflow on file failures
plucky dock
Hey, I'm looking for volunteers who could help with the script benchmarks.
Just download this repository and inside from the terminal position type
python3 main.py "test" "test" 130 90
I need this data to make a table comparing password cracking times for a specific processor.
bright glacier
This is probably the wrong channel, no? Not sure how this is related to the psf/black oss project ^^
bright glacier
so close! but I need to rewrite the log file name or else this confusing diff difference will occur >.<
It's kinda scary how much time this one task has taken me, but this integration is probably one of more helpful improvements to developer experience so it's worth it π
bright glacier
hmm, sqlalchemy is still failing, unfortunately diff-shades doesn't let you just...
so I stopped failing the workflow if the baseline revision has failing files and also added --check-allow so we can specifically ignore certain already known issues (for the target revision)
flat krakenBOT
glacial swan
INFO:blib2to3.pgen2.driver:Generating grammar tables from /home/jdjg/JDBot/JDBot/lib/python3.8/site-packages/blib2to3/Grammar.txt
INFO:blib2to3.pgen2.driver:Writing grammar tables to /home/jdjg/.cache/black/21.12b0/Grammar3.8.10.final.0.pickle
INFO:blib2to3.pgen2.driver:Writing failed: [Errno 2] No such file or directory: '/home/jdjg/.cache/black/21.12b0/tmpzys_y126'
INFO:blib2to3.pgen2.driver:Generating grammar tables from /home/jdjg/JDBot/JDBot/lib/python3.8/site-packages/blib2to3/PatternGrammar.txt
INFO:blib2to3.pgen2.driver:Writing grammar tables to /home/jdjg/.cache/black/21.12b0/PatternGrammar3.8.10.final.0.pickle
so I installed black into venv
I am not sure why this is an issue now
didn't have this issue until yesterday
Version: 21.12b0
bright glacier
You can honestly ignore those, it may say error but black can recover from that one
@neon loom I copied black-primer to https://github.com/ichard26/black-primer. Do you want me to transfer it to you?
neon loom
Ready to make it an optional install in case someone uses it?
Well to point people there?
neon loom
Iβm cool to just be added as a moderator if you want a second person.
bright glacier
Iβll take this and move it to cooperlees/black-primer after the release and if people want it they can install it from there β¦
I didn't misread this, did I?
neon loom
Letβs share on the Readme that itβs unsupported and only there for legacy use cases - fork away.
Yeah we can do that or im happy if itβs easier to just use yours - I donβt care
What ever is easier
bright glacier
well my plan was to
- copy it into its own repo [done]
- delete black-primer from psf/black
- profit $
I was just trying to save you some effort by making the repo myself and then transferring it to you as you originally planned / stated
neon loom
Cheers
I just wanted it available in the low chance someone is like what I use primer β¦ etc etc
bright glacier
Cool, I had some free time and decided to do part of the work for ya :)
glad to hear I didn't misread and wasted an hour
nimble bough
the committed solution was a bit too naive
can u help me?
nimble bough
Cool, I had some free time and decided to do part of the work for ya :)
can u help me?
bright glacier
can u help?
Seriously, please stop pinging a bunch of random people for help. Or at least provide more context to make the ping worth it.
neon loom
glad to hear I didn't misread and wasted an hour <:lemon_sweat:75444188171862028...
Sorry - the transfer expires very quick. I thought it would wait till the am, seems not
bright glacier
I'm actually a bit sad π₯²
dense jungle
thanks, I was too lazy to find the file myself π
flat krakenBOT
Description
Closes #2921. Aims to remove redundant parentheses from with statements:
I currently have one failing test tests/data/parenthesized_context_managers.py since this code now removes unnecessary parentheses around one-line with statements, so...
bright glacier
If someone else could walk through this person through the tests and preview that'd be great ^^
dense jungle
If someone else could walk through this person through the tests and preview tha...
on it
topaz rapids
Hi
topaz rapids
If someone else could walk through this person through the tests and preview tha...
Walk through what
errant barn
this
I got stuck with this error can someone please help me out
flat krakenBOT
tired shard
https://github.com/psf/black/runs/5563343510?check_suite_focus=true#step:9:35 what would be the best way to go around this?
dense jungle
https://github.com/psf/black/runs/5563343510?check_suite_focus=true#step:9:35 wh...
wow that's a weird error
tired shard
yeah :(, mypy can be troubelsome sometimes
dense jungle
commented on the PR
flat krakenBOT
dense jungle
you successfully tricked me into reviewing the rest of the PR too π
tired shard
ty :D
spare cave
just wanted to say we use black in our class and it does make the code much easier to read
thank you
flat krakenBOT
tired shard
Felix, finally opened the issue :D
errant barn
mint ππΏ
flat krakenBOT
left grotto
hi . intermediate python coder here. i am looking to upskill my coding level, introduce type hinting into my code, use linters etc. i wanted to know what is the difference between black and the linters suggested in my vs code e.g. mypy, pylint, flake, etc.. and what is autopep8... do these all work to conform the code style with PEPs? so what is the difference between them?
dense jungle
hi . intermediate python coder here. i am looking to upskill my coding level, in...
they basically work at different levels. black and autopep8 are autoformatters, they do things like split up long lines and add whitespace
mypy is a type checker. it looks at type annotations in your code and tells you when you do something like 1 + "x" that breaks the type system
left grotto
why are there so many linters? do they all do the same job? and which one is the best?
dense jungle
flake8 and pylint are linters, which are kind of in between the two. they find (and sometimes fix) various mistakes or risky bits in your code
left grotto
this might seem like a dumb question but why not combine the functionality of autoformatters and linters into one package?
black + flake8+pylint all into one?
dense jungle
they do very different things internally. you could make a tool that wraps a bunch of them but not sure how useful that would be
in particular, autoformatters format code in place and linters and type checkers produce lists of errors, so the output format is very different
flat krakenBOT
vocal fjord
dense jungle
nice to see in https://github.com/PyCQA/pylint/pull/3577 that pylint is moving to get rid of checks that duplicate what Black does
(and thanks @tired shard and the PR author for digging up those links π )
tired shard
Btw ichard, would love to see a review on psf/black#2525, seems super close to merge :D
toxic stormBOT
flat krakenBOT
dense jungle
yes GH is having some issues today
tired shard
ah yep webhook performance looks degraded
new pfp π is that you?
a bit different looking than the previous one
dense jungle
new pfp π is that you?
what is pfp?
tired shard
profile pic
tired shard
nice beard and hat π
little parrot
"isort, unlike black is a fantastic ... tool" :^)
molten kernel
how do i configure black in Pipfile, like i can in pyproject.toml
i tried [black] and [tool.black]
dense jungle
how do i configure black in Pipfile, like i can in pyproject.toml
that's not supported
molten kernel
:( why not
dense jungle
pyproject.toml is the standard configuration file. we're not interested in adding more config files
molten kernel
pipenv is recommended by pypa
it may not be standard but really all that has to be done is a check for the Pipfile in the same directory, and look for tool.black in the same way
bright glacier
Oh man, it's weird not seeing a cute cat where ever on the internet you're present Jelle.
I'm sure I'll get used to this one too, just needs time :)
bright glacier
it may not be standard but really all that has to be done is a check for the Pip...
other that pyproject.toml being the one file the most python projects will have, there's a cost to supporting formats especially with code complexity and maintenance
dense jungle
yes. we have to figure out what to do if both config files are present
and whatever we do will probably confuse someone
bright glacier
adding Pipfile support might not be that expensive from a maintenance perspective, but there's this good analogy Brett Cannon said: you're presenting us a kitten (it's low cost on its own) but when they grow up it's not so low maintenance / and arguably cute
and we have to draw the line somewhere and we prefer to draw the line defensively erring on the side of keeping maintenance sustainable
software almost always marches towards gaining limitless complexity from "just one more feature" and we'd like to slow that down at least :)
flat krakenBOT
tired shard
Since this hasn't been a common request and it would complicate our internals, I'm now inclined to reject this enhancement request. We can revisit that decision if we get more use cases.
errant barn
<@!783088578363523104> <@!429445826251587616> what about PR #2527 ?
If I understand you correctly, no since that only deals with configuration values, not the location
molten kernel
at the time of that PR black even used Pipenv lol
errant barn
Yeah, we've since stopped using it π
tired shard
If I understand you correctly, no since that only deals with configuration value...
Oh, I didn't know it was a continuation of an ongoing discussion, my bad.
lot has changed since that PR
errant barn
None of the labelled issues? There are some unstables, but no obvious duplicates, at least at a quick glance
tired shard
i remember seeing an issue talking about fmt:off/skip on function definition lines
maybe wrong
errant barn
would be good to find though!
tired shard
ok nah can't find
atleast the labeled issues
ok found it
yeah it isn't labelled
errant barn
oh, nice detective work
tired shard
lol ty
bright glacier
oh how great
flat krakenBOT
tired shard
Btw ichard, would love to see a review on psf/black#2525, seems super close to m...
ig you missed this @bright glacier
And for adding isort to our dev dependencies, should I just rename test_requirements to dev_requirements or I have to add it somewhere else?
(or usort)
errant barn
they are our dev requirements essentially
so renaming? why not, but anyways that's where I think they go
bright glacier
euh, I disagree with renaming it
We don't specify which mypy version we use other than in .pre-commit-config.yaml
tired shard
i don't see any better place than π€·ββοΈ
bright glacier
same with flake8
tired shard
hmmm
i willl just do that then
to be consistent with others
And if you could put your thoughts on usort/isort, so we can choose one
bright glacier
it could be worth it but I'd keep it in a separate file (maybe a larger dev-dependences.txt file) to avoid slowing down the test session setup
tired shard
i never had the need of it, so its fine IMO
errant barn
that could be nice, since the wider file can include from the test deps, so the dev setup wouldn't be any less convenient
bright glacier
@neon loom damn, we're really raising the standard of linter and autoformatter CI, even pytest has their own primer: https://github.com/PyCQA/pylint/issues/4413
neon loom
Imitation is the best for of flattery
flat krakenBOT
flat krakenBOT
tired shard
I see some false formatting changes here, https://github.com/psf/black/compare/main...Shivansh-007:experiment-isort?expand=1
not sure how i fix them, haven't messed up with isort yet
the config looks correct though
atleast if the docs aren't hiding anything
bright glacier
you should be using known_first_party not known_third_party for black and friends
third_party is for dependencies and first_party is for in tree code
tired shard
oh shit yes, i didn't see that. ty π
bright glacier
also "adds an autoformatter to an autoformatter project" -> proceeds to use inconsistent formatting in the configuration π₯΄ lol
plz add spaces around the equal signs, kthxbye
tired shard
lol will fix
flat krakenBOT
bright glacier
no.
tired shard
ok π
bright glacier
Isort: Fix typo third_part -> first_party
the typos are taking no prisoners
I'm surprised at the diff size
tired shard
the only thing i am not 100% liking
i would have liked that to be broken down into 2-3 imports
also i still yet to try usort, it seems more promising
bright glacier
In general I prefer a single long from import over multiple imports but yeah it could be argued it's a bit much here
tired shard
yeah even i do, but this a bit too big
i think there is a config option for that
will see
another one
bright glacier
It might be attaching semantic meaning to the comment
Isort has some interesting logic keeping sections relatively intact these days
tired shard
yeah it does
i will give usort a try once
and then we could decide
but according to the features, i am more on usort
that pretty much sums it
Β΅sort operates on a strictly-parsed syntax tree of each file as a whole, rather than reading and parsing individual lines of text at a time. This guarantees that Β΅sort is modifying the actual Python syntax elements, along with any associated comments, and generating grammatically correct results after sorting.
one down side tho
Β΅sort uses case-insensitive, lexicographical sorting for both module names and imported items within each statement. This means uppercase or titlecase names are mixed with lowercase names, but this also provides a more consistent ordering of names, and ensures that frog, Frog, and FROG will always sort next to each other:
and there is no config oprtion for that
bright glacier
yeah I'd prefer constants in their own little group -> i.e. taking into account casing
tired shard
yeah
tired shard
> Β΅sort operates on a strictly-parsed syntax tree of each file as a whole, rathe...
couldn't see any issue re this in the current PR btw
bright glacier
It might be better to separate the normal vs preview test data files instead of repurposing the stable ones as preview
dense jungle
yes, we haven't really thought about how testing works with stable vs. preview
flat krakenBOT
dark kiln
Damn, this is some advanced sh*t.
mint barn
Who is it?
I wonder for already a really long time
It even does not have a profile
Mysterious @github#0000
Oh, a webhook
Than i got it
neon loom
yes, we haven't really thought about how testing works with stable vs. preview
This could be good PyCon sprint work for me maybe.
tired shard
Any updates on psf/black#2863
toxic stormBOT
dense jungle
Any updates on psf/black#2863
My feeling is that that's intended behavior
teal goblet
workaround: black --config ../../pyproject.toml
neon loom
Any updates on psf/black#2863
flake8 stops at first flake8 found too ... I feel this determinism is something we want ... but I get the other side of there being no black config we could also keep going
This could be a breaking change to people who already exercise this behavior tho in practice
flat krakenBOT
shrewd zinc
Is black supposed to fail if it encounters stuff like
f"some text {'\n'.join((f' β’ {i}' for i in list_of_strings))}"
?
errant barn
Does that work in Python 3.10? It fails on 3.9
With the same error that Black gives: "f-string expression part cannot include a backslash"
shrewd zinc
How do I see black errors in vscode?
At least the editor doesn't complain about anything
shrewd zinc
no, 3.8
bright glacier
yeah that's illegal syntax
errant barn
Yep, not a Black problem then I'm afraid
bright glacier
technically you could work around the issue by passing --fast but if the code isn't even valid Python, I don't see the point
errant barn
Can't always trust editors :D
bright glacier
!e print(f"{'\n'.join([1, 2])}")
shrewd zinc
What specifically is the issue?
late dewBOT
@bright glacier :x: Your eval job has completed with return code 1.
001 | File "<string>", line 1
002 | print(f"{'\n'.join([1, 2])}")
003 | ^
004 | SyntaxError: f-string expression part cannot include a backslash
The backslash
late dewBOT
@shrewd zinc :white_check_mark: Your eval job has completed with return code 0.
1-2
red sedge
I totally missed that backslash lol
They're looking to fix this in 3.12
Integrating the f-string parsing with the peg parser makes it possible to parse it
shrewd zinc
peg parser?
red sedge
flat krakenBOT
lament crow
Twitter
flat krakenBOT
flat krakenBOT
bright glacier
https://github.com/psf/black/pull/2948 ... uhhh where's the beta commit github?
Do I just manually merge the PR or something..?
errant barn
at least GHA was a bit slow earlier
bright glacier
guess who forgot to change the commit author before committing this manually
!remind 1M is the github .git-blame-ignore-rev blame support out of beta? don't forget to attribute https://github.com/henrikhorluck
late dewBOT
bright glacier
!remind edit duration 4276 1m
late dewBOT
dense jungle
!remind edit duration 4276 1m
impressed you managed to edit the duration within the 1-minute window you had
bright glacier
I'm not that slow I swear π
flat krakenBOT
bright glacier
gosh, why is this diff-shades integration so complex haha
anyway I think I should be able to update the PR in 30ish minutes assuming testing passes
neon loom
What makes it so complex? Anything I could help on to try not make it so, sorry I've been at a distance with it all
bright glacier
The additional features we need π -- originally the integration would only show the impact of black using the projects' own configuration (or the defaults) but now we want to move away from that and instead force the preview / stable style
neon loom
ahhh right
How you going to ensure preview does what it should? Just toggle if we expect changes or not?
bright glacier
had a real fun time trying to fix the latest bugbear error code too btw lol
bright glacier
B019, the diff-shades code is a bit hairy so I ended removing the caching altogether
neon loom
I just released another one today with more fixes for some false positives on new checks
ambv talked me out of removing 3.6 support - Will do it when flake8 itself does it
bright glacier
How you going to ensure preview does what it should? Just toggle if we expect ch...
Not sure what you mean, but now the integration has two jobs (well one more for run configuration, but we can ignore that) for PRs:
- preview-changes: basically what we had originally but now all projects use the preview style (it's more useful this way)
- assert-no-changes: this is new and will fail on any differences between the baseline and target revisions (it forces the stable style for all projects)
For pushes to main, only preview-changes runs.
neon loom
Yup cool. Sounds great and answers what I was trying to ask.
bright glacier
The more I hack on this the more I think it'd make for a great blog post.. something something about real world testing with downstream :)
but I got other things to write about first (and other tasks to finish first before that)
https://github.com/ichard26/black/runs/5668459799?check_suite_focus=true it's always so satisfying to see this complex piece of CI work well π
hopefully it stays that way and doesn't fail in a way I don't expect
neon loom
Cut issues for things / easy wishlist stuff you want and I'll come do some: https://github.com/ichard26/diff-shades/issues
So I can learn it
And @bright glacier - What power do you have now being on events team? Congrats
I was like, I like to party, I want a party hat
bright glacier
No more power actually, I just now get pings for all of the events python discord are planning and running 
bright glacier
I've been python discord staff for quite a while now
bright glacier
Run diff-shades compare --check assert-no-changes-main-450253afaf.json assert-no-changes-pr-15-8b2f244f34.json || echo "Please verify you didn't change the stable code style unintentionally!"
guess who messed up their bash conditional logic :p
I forgot to add the exit 1 at the end
neon loom
bash is evil
neon loom
I converted a script at work today to simple python cause I find it easy. from subprocess import run and have at it
bright glacier
it is configured but that bash command runs the second command if the former fails and I forgot that basically ignores the return code from GHA's perspective
late dewBOT
.github/workflows/changelog.yml lines 18 to 21
run: |
grep -Pz "\((\n\s*)?#${{ github.event.pull_request.number }}(\n\s*)?\)" CHANGES.md || \
(echo "Please add '(#${{ github.event.pull_request.number }})' change line to CHANGES.md (or if appropriate, ask a maintainer to add the 'skip news' label)" && \
exit 1)```see that && exit 1 ? that's what I need
Seems off topic for this channel, no? this is for the psf/black autoformatter project
lethal cairn
oh, sorry I tought it was for open sorce projects, Im new here
bright glacier
this channel is indeed part of the open source project category but each channel is dedicated to a specific one python discord admins has vetted
lethal cairn
ooh, I see, thanks man!
bright glacier
There's no channel for sharing your project, but the off topic channels usually work well enough assuming you'll doing it in good faith and are following the rest of the rules here.
!off-topic
late dewBOT
lethal cairn
thanks man!
bright glacier
Cut issues for things / easy wishlist stuff you want and I'll come do some: http...
haha, I'd be nervous having you look at the code given you're a professional dev and well ... I'm not one :p
I can do that though, I think I have a wishlist somewhere in notion π€
neon loom
haha, I'd be nervous having you look at the code given you're a professional dev...
I'm a network engineer ...
Who likes to code. My code is horrible too.
bright glacier
Coding is fun, although I've been doing way more webdev than Python stuff lately
neon loom
@bright glacier I'm submitting a talk to a conference in your poutine town