#black-formatter

1 messages ยท Page 15 of 1

lament crow
#

they said it's a good thing that pre-commit doesn't track changes

plain atlas
#

Guess I should add --check to my precommit config....

#

Stop reformatting and then rollback lmao

lament crow
#

using the stable tag (which can change) is what's not good

frozen yacht
bright glacier
#

well the use of the rev: stable makes it look like pre-commit will track changes, but it won't (which IMO is good because different versions of black have different formatting so upgrading should an explicit decision), explicitly configuring a specific version of black is better

frozen yacht
#

Understood, thanks

bright glacier
#

No worries, this area could definitely do with some better docs

#

And the release process for 21.6b0 has been officially completed ๐ŸŽ‰

flat krakenBOT
flat krakenBOT
#
jack1142
[psf/black] Pull request opened: #2330 Add EOF and trailing whitespace fixer to pre\-commit config

Unix tools often rely upon the files ending with a newline character to show the output properly (e.g. when the file has no newline at the end, cat file will cause the prompt to show right after its output without a newline in between). This may help minimize the diffs and ensures consistency across all files.

I've also added auto-removal of trailing whitespace since I don't think that outside of test data (which is excluded from pre-commit) anyone would actually want to have trailing wh...

crisp patio
#

what is black?

#

oh its a prettier for python?

lament crow
#

yes, but it probably has even fewer options to configure than prettier

#

but that's why it's called opinionated code formatter :)

crisp patio
#

its also newer right..?

#

so maybe give it some time?

lament crow
#

not that much newer

#

and what do you mean by "give it some time"?

silent apex
#

It's supposed to take away the confusion of why to format what in a certain way, and they do that by not giving you a choice.

crisp patio
#

oh

#

atleast theres no 200 different ways to do it and everybody chooses something else

flat krakenBOT
bright glacier
#

well I just approved the CI run for that PR, it's probably going to fail due to missing typeshed packages unfortunately, I wish pipenv locking was faster

#

!remind 4H review the pre-commit PR and update the requirement files to include typeshed packages

late dewBOT
#
I'll allow it.

Your reminder will arrive in 4 hours!

ID: 2374, Due ยท
amber karma
#

How do i get started with black? I've never used it before so looking for some tips n tricks

dense jungle
heavy heron
bright glacier
#

While I'm not happy with it, I'm also honestly scared to change what works "fine enough". IIRC zsol was working on porting black to poetry but the work was never finished. Also IIRC ambv was supportive. I guess we can look into it again.

tribal thistle
#

I will say one thing, poetry works nowhere as nice with different python versions

#

I did a full write up back then

bright glacier
#

psf/black#1411

tribal thistle
#

I can dig it up

toxic stormBOT
heavy heron
#

Haha, "the" write up

tribal thistle
#

Did a lot of them back then. Was quite chatty

lament crow
#

poetry is a whole different build system, pipenv is just (or mostly?) for development on the other hand

flat krakenBOT
tribal thistle
#

sir-lancebot#623

toxic stormBOT
lament crow
#

oh, so you can use it without switching the build system as well?

bright glacier
tribal thistle
#

Though Iโ€™d think it would be quite nice for this?

#

The build system that is

bright glacier
#

I mean, I am sorta being a hypocritical because I just used flit for my newest project but shrug

#

with its own build system

lament crow
#

I mean, newest project is not the same as existing project

#

I'm not trying to say poetry (as build system) is bad though, it's just different which means that more stuff needs to be considered before pursuing it

#

I feel like install time could be longer since it's non-default build system (and so pip needs to install it before build)?

#

not by much I would assume though

bright glacier
bright glacier
lament crow
#

I guess that's true

#

did not think about it lol

#

I haven't used alternative build systems much yet so I guess I just don't know much yet

bright glacier
#

Although you aren't wrong in a way, if we start compiling black with mypyc (which is what I am working on right now) then our wheels won't be pure python anymore. In that case though, I'd expect us to still publish a pure python wheel with 100% python and 0% mypyc alongside with the mypyc compiled wheels

bright glacier
lament crow
#

oh ye, I did

#

only like once but well

#

was rather recent so I also could take advantage of PEP 621

bright glacier
#

There's too many packaging peps I can't remember lol

#

!pep 621

late dewBOT
#
**PEP 621 - Storing project metadata in pyproject.toml**

Link

Status

Final

Created

22-Jun-2020

Type

Standards Track

bright glacier
#

IIRC flit just landed experimental support for that pep right?

lament crow
#

it's almost 3 months since it landed but it is still the latest release so ye

bright glacier
#

unfortunately there's no support for standardised pyproject.toml based requirements

lament crow
#

wdym

#

There should be

#

I definitely used it

#

sec

bright glacier
#

takluyver/flit#373

toxic stormBOT
bright glacier
#

unless pep 621 also defines and standardises requirement metadata, which in that case, my bad!

lament crow
#

it did, ye

#

well

#

it might have been separated, not sure

#

but it definitely was part of it in the end

# 
[project]
# snip
dependencies = [
    "dearpygui~=0.6.415",
]

[project.optional-dependencies]
dev = [
    "black==21.5b2",
    "flake8==3.9.2",
    "isort==5.8.0",
    "mypy==0.812",
]
bright glacier
#

Oh, pep 621 defined the general structure of the pyproject.toml project metadata while pep 631 is like a spec for the requirement metadata?

#

!pep 631

late dewBOT
#
**PEP 631 - Dependency specification in pyproject.toml based on PEP 508**

Link

Status

Accepted

Created

20-Aug-2020

Type

Standards Track

lament crow
#

there were 2 competing PEPs for specifying dependencies

#

but the one that "won" got merged into 621

bright glacier
#

Hmm, I gotta try and migrate my project to the 621/631 standardized metadata format, TIL, thanks!

lament crow
#

Personally I used PEP 621 because I found flit's definition of entry points a bit confusing

#

but they're generally rather comparable

tribal thistle
#

Hmm why the double labeling

lament crow
#

one thing to keep in mind is that you need to define version and description as dynamic fields if you want flit to export it from your code

heavy heron
tribal thistle
#

Toml header & dict

bright glacier
#

wasn't there another pep for sdist metadata and the dynamic marker... sigh

heavy heron
#

Probably, they are often many PEPs on different aspect of a subject for some probably very good reason

bright glacier
#

I agree with keeping them separate but man is it annoying to keep up with the packaging changes sometimes

mint barn
#

Does the "rtfd" in black.rtfd.io stand for "read the fucking docs" lol?

lament crow
#

pretty much :3

#

it's short domain of ReadTheDocs's

bright glacier
#

it's the price we get to pay for a short url from readthedocs I guess lol /j

mint barn
#

Oh ok
Still pretty nice lol

flat krakenBOT
worn zenith
#

is there a reason why the quotes/strings argument is -S instead of -s?

#

I always get it wrong on the first try because the help example isn't next to other text, so I don't notice the capitalization the first time.

glad pilot
#

I don't know if this is why Black chose that or not, but there's a kind of loose convention that lowercase switches are for enabling some behavior and uppercase are for disabling some behavior. That's not always very consistent across tools, though

worn zenith
#

I've always noticed that lowercase switches were the default scenario, and that uppercase was for switches that would be abbreviated in the same way, but were implemented later.

flat krakenBOT
late dewBOT
#

@bright glacier

It has arrived!

Here's your reminder: review the pre-commit PR and update the requirement files to include typeshed packages.
[Jump back to when you created the reminder](#black-formatter message)

bright glacier
#

oh right

worn zenith
#

I'm sure there's copiuous discussion on this, but why is line comment length not affected by the length parameter?

errant barn
#

There's an ongoing issue about splitting comment lines as well! See psf/black#1713

toxic stormBOT
worn zenith
#

Oh I was referecning block comments, like this docstring, which presumably is an entirely separate formatting discussion:

unknown.png
#

IN this particular case, they're all bounded to 120 characters instead of black's default 88, so at least this specific instance is handled invisibly wiht -l 120

errant barn
#

I might be wrong on this one, but I don't think Black formats docstrings besides a few modifications to indentation and newlines.

worn zenith
#

RIght, it didn't. I was wondering if there was already an existing discussion with respect to that before I went asserting my opinion.

errant barn
#

There's psf/black#2289 at least!

toxic stormBOT
worn zenith
#

Very fresh issue indeed.

flat krakenBOT
flat krakenBOT
bright glacier
#

and I was about to approve that

dense jungle
#

got to be faster

bright glacier
#

Jelle and I end up handling the same issue / PR at the same time so frequently

#

it's bonkers

bright glacier
dense jungle
#

I already merged it, sorry ๐Ÿ˜›

bright glacier
#

I meant this one: psf/black#2331 ๐Ÿ˜„

toxic stormBOT
dense jungle
#

oh yeah I merged the trailing line one. that one needs a Pipfile update I believe

bright glacier
#

Yep.

#

I can't wait once the "expose static metadata in simple api" for indexes pep is accepted and implemented, should cut down on resolving times significantly ๐Ÿ™‚

#

That will probably take a while tho

dense jungle
#

might not be too bad? I bet the PyPI people really want it too so they can save on bandwidth

errant barn
#

Hmm, does RTD need to be triggered manually when the stable tag is moved? The docs aren't updated to 21.6

bright glacier
#

I'll look into it

#

also, I don't think I did the pipenv locking correctly

#

or the github dependency review tool is broken

#

unfortunately it appears prettier changed their JSON formatting style so it's impossible to tell what pipenv lock changed

lament crow
bright glacier
#

and I forgot to include types-click .......... whelp time to relock

lament crow
#

but hmm, indeed they seem out of date ๐Ÿค”

bright glacier
#

yeah can confirm they are out of date, wonder what happened, looking into it right now

#

oh, it looks like I somehow messed up the updating of the stable tag

#

wait no

errant barn
#

Oh? It was on the correct commit though

bright glacier
#

my local copy is wrong

#

points to the right commit, you're right felix

#

huh then

#

more detective work I guess

#

(by the way, this is why I like using a branch for stable, a tag isn't automatically updated)

errant barn
#

Fair enough, the situation will improve when the real stable is released since the latest tag will be built

bright glacier
#

let's try rebuilding stable

#

Good catch Felix!

errant barn
#

Well at least it's good now :D

#

I wonder if RTD can be configured to have stable on any tag, not just stable tags (assuming it has a distinction rather than it was assigned to only build the tag named stable)

bright glacier
#

Well, if there isn't a specific "stable" rev in the git repo, it automatically points the stable docs to the latest tag IIRC

#

but we need the "stable" rev (i.e. a "stable" tag) for other situations so yeah

errant barn
#

Right okay!

#

Oh, nevermind then ๐Ÿ˜…

bright glacier
#

Like the GitHub Actions action (not a typo) we provide

flat krakenBOT
#
bbugyi200
[psf/black] Issue opened: #2334 I'm not sure how often \`pathlib\.Path\` is used with the augmented assignment operator rather than the regular one \(\`/\`\) nor do I know how oft...

I'm not sure how often pathlib.Path is used with the augmented assignment operator rather than the regular one (/) nor do I know how often the string would be long enough that it would need to be split but all that aside, /= can definitely appear with strings on the right-hand of the assignment operator too:

from pathlib import Path
x = Path.cwd()
x /= "directory/subdirectory/file.txt"

_Originally posted by @jack1142 in https://github.com/psf/black/issues/2312#issuecom...

bright glacier
#

and once again my displeasure with the Pipfile file format is back again

#

although that's also because I don't know to write environment markers in any requirement files by heart

bright glacier
#

well, that's great ๐Ÿ˜…

Screenshot_from_2021-06-13_16-19-58.png
#

whew, a simple Alt + F2 and "r" fixed it

foggy sapphire
#

yo guys i just downloaded this black formatter bc i was curious

#

could you help me run it on a file

bright glacier
#

black [your-file]

foggy sapphire
#

in command line?

bright glacier
#

yes

foggy sapphire
#

also do i need to have a virtual environment active

bright glacier
#

it's a command line only tool (there are library features but we don't officially support them just yet)

foggy sapphire
#

o wow nice

#

thanks!

#

woahhhhh this is cool

bright glacier
#

but black doesn't need a virtual environment to run

foggy sapphire
#

thanks dude!

bright glacier
#

No worries, have fun!

tribal thistle
#

is there an "official" email contact for black? Would that just be PSF?

errant barn
late dewBOT
#

setup.py line 59

author_email="lukasz@langa.pl",```
tribal thistle
#

Thanks

errant barn
#

Also, any tips on speedrunning ล? :D Every time I care I end up searching his github and copying the character. I couldn't find an alt code (that works) for it either

bright glacier
#

For Ubuntu I use ctrl shift alt + 141 to get that L (sorry gboard doesn't have that L for some damn reason)

#

no idea whether that would work for you tho

errant barn
#

Yeah for some reason that's "รฌ" for me :D

bright glacier
#

Windows?

errant barn
#

Yuup

bright glacier
#

whelp IIRC it does alt codes differently

#

0321 is something I found via a Google search, maybe try that?

errant barn
#

:D that would be an "A"

bright glacier
#

I give up then :(

errant barn
#

Ima have to create a macro for this :D

lament crow
errant barn
#

Oh God..

lament crow
#

more seriously though, you could actually make a custom keyboard layout

#

And choose some key combination that isn't used for writing ล

#

could even go for regular (Shift+)AltGr+L if your keyboard layout doesn't use that

errant barn
#

Yeah that would be nice!

lament crow
#

the above is called Microsoft Keyboard Layout Creator btw

tribal thistle
#
  • mdit-py-plugins requires Python ~=3.6, so it will not be satisfied for Python >=4.0
bright glacier
#

Context please?

tribal thistle
#

poetry knows something I don't ๐Ÿ˜

#

I'm getting to it lol

#

Was giving setting up a poetry branch a shot

bright glacier
#

I'm the middle of debugging this CLI application I'm making

bright glacier
tribal thistle
#

Got roadblocked by dependencies. MyST-Parser requires mdit-py-plugins

#

and somewhere in there python 4 came up

tribal thistle
bright glacier
#

A benchmarking suite for black

tribal thistle
#

that sounds great for what I'm testing lol

bright glacier
#

it's for my mypyc work ๐Ÿ˜„

#

I'm on like rewrite #4 for my data modeling

#

I think this one should stick though

tribal thistle
#

lmao

#

Well if you need any help/willing victims for testing at some point

bright glacier
#

Yeah, I'll make an announcement once I have an usable dev version done since I'm going to need some actual files to benchmark black against

#

Right now my target files is a checkout of the black repo from rev 21.6b0

#

here's a sneakpeek ๐Ÿ˜„

Screenshot_from_2021-06-13_17-30-40.png
tribal thistle
#

oh wow

#

I was not expecting it to look good

#

well, that came out wrong lol

bright glacier
#

oh it's fine, the actual benchmarking hardwork is handled by pyperf

#

!pypi pyperf

late dewBOT
bright glacier
#

blackbench is basically a collection of benchmarks + black specific configuration support for those benchmarks + a fancy frontend to pyperf

#

it's like pyperformance but for black

#

!pypi pyperformance

late dewBOT
bright glacier
worn zenith
#

Its a profiling tool specific to Black, a tool that itself provides only offline delays?

bright glacier
#

Can you please explain what "offline delays" means here? I don't understand, sorry.

worn zenith
#

Like a slow compiler, it has no bearing on the runtime of the target code. That is, no matter how long it takes "Black" to be run against a library, say for the sake of argument, requests, it will only be run once, and then requests will continue to operate at whatever speed it can.

#

rust suffers from this, in that the rust compiler can be very slow compared to the same program written in c and compiled with gcc, but the ultimate output is comparable.

bright glacier
#

Okay I understand, thanks for the explanation. Yes, blackbench is intended a performance measurement tool for black (it doesn't collect deep enough data IMO to count as profiling).

#

It's because I have a project to improve the performance of black using mypyc and I needed someway to measure the real-world improvements.

worn zenith
#

Not to deminish your work, but since I've never run black against a large dataset, how slow is it really?

#

And, in truth, first-time high-latency may be OK, if subsequent runs are vastly improved.

#

IDK if subsequent runs are in fact faster though.

bright glacier
#

Subsequent runs should be faster due the cache black maintains but the first runs can be quite slow

#

although black is a multi-core application so it depends

#

the main benefit of increasing black formatting throughout IMO is when you configure your editor / IDE to run black automatically on save

#

decreasing that delay between save and formatting finished makes those users' lives easier and less annoying

worn zenith
#

INdeed it would, especially for tools that do background autosaves.

#

I'd like to see IDE notations before automatic reformatting; simply highlight the lines like PyCharm can do for pip-8/flake-8 line lengths and suggest the fix, offering the keyboard shortcut for formatting that segment, instead of just rearranging the code automatically.

#

Especially if I have autosaves enabled and I'm still typing.

#

Though thats out of your scope.

#

But low latency coverage would be very helpful for such a product.

bright glacier
#

That would be hard to implement for Black since 1) we would rather not provide first-party support for formatting a specific range of lines in a file, and 2) black doesn't really provide any fine-enough output to implement that cleanly in an IDE

worn zenith
#

Yea I'm sure its a very complex thing to accomplish.

bright glacier
#

Yep. It especially gets tricky when the range isn't valid Python.

worn zenith
#

I do wish there were a way to recognize patterns. It undoes some of my formatting to make them consistent:

unknown.png
tribal thistle
#

Hmm have y'all ever run into I/O and event loop and file permission issues when running primer? This is related to my poetry work, but I figure I may as well ask before I go mad digging into the code.

Full TB:
https://paste.pythondiscord.com/avabamasax.lua

It's one hell of an ugly traceback. I think it printed things intertwined.

worn zenith
#

though it does look like adding a trialing , to every method call makes them the same, even if that is a bit aggressive.

tribal thistle
#

the magical trailing comma lol

tribal thistle
#

Aaaaalright, and I'm throwing in the towel for tonight. The primer thing is still an issue, but I'll look more into that soon. I've gotten some nice work done.

I've got two branches to show for it:
https://github.com/HassanAbouelela/black/tree/poetry - Exactly what it sounds like. Moved from pipenv to poetry, no other changes.

https://github.com/HassanAbouelela/black/tree/poetry-build - This one is a bit more involved. It moved to poetry as the dep manager & build and publishing tool. It might not be desired, but I made it as a POC and for fun anyways. There are still things that need to be added (see missing.md). the published package is here: https://test.pypi.org/project/black-poetry/

https://github.com/HassanAbouelela/black/blob/poetry-build/missing.md actually applies to both branches for the most part.

bright glacier
worn zenith
#

I saw that. Though it appears that running black without - C, then with, then without, is not idem potent, since it looks like it can remove some trailing commas. Git diff shows deltas. I need more testing to narrow down the details.

worn zenith
#

Alright so -C is not just skip magic trailing commas, in some instances it will affirmatively remove trailing commas.

All done! โœจ ๐Ÿฐ โœจ
9 files left unchanged.

C:\Users\Xavier\PycharmProjects\SplitFileReaderProject>.\venv\Scripts\python.exe -m black -l120 -S -C .\split_file_reader\
reformatted split_file_reader\__init__.py
reformatted split_file_reader\split_file_reader\__main__.py
reformatted split_file_reader\http_file_reader\http_file_reader.py
reformatted split_file_reader\split_file_writer\split_file_writer.py
reformatted split_file_reader\split_file_reader\split_file_reader.py
All done! โœจ ๐Ÿฐ โœจ
5 files reformatted, 4 files left unchanged.

C:\Users\Xavier\PycharmProjects\SplitFileReaderProject>.\venv\Scripts\python.exe -m black -l120 -S .\split_file_reader\
All done! โœจ ๐Ÿฐ โœจ
9 files left unchanged.```
unknown.png
#

1st to show that there is no diff if run without -C.

#

Second makes the changes with -C

#

Third to show that no -C makes no further changes.

#

Diff has an omitted trailing comma on some arguments.

#

IDK if this is expected behavior, seems slight contrary to the documentation.

dense jungle
#

Looks like expected behavior to me

worn zenith
#

The problem I have is that the documentation for command arg -C is

                                 Don't use trailing commas as a reason to
                                 split lines.```
However, this means that executing the command with `-C` will cause some trailing commas to go away, meaning the command `python.exe -m black -l120 -S .\split_file_reader\` produces different output, depending on whether `-C` has _ever_ been used in the command,  not just on this particular run.
#

My assertion is that, if this is expected behavior, the documentation should say: -C, --skip-magic-trailing-comma Don't use trailing commas as a reason to split lines, and remove extraneous trailing commas in method signatures.

dense jungle
#

I do agree that that documentation can be improved. It's not just about method signatures though, but also affects lists and other collection displays

worn zenith
#

Because its not just a skip, its an affirmative change to the code.

#

Sure, method signatures is the only delta I noticed in my code base, so it was the only thing I could assert was changed.

#

I also argue that the long-form command should be --remove-magic-trailing-comma, as skip implies a passive selection.

#

or similar

worn zenith
# worn zenith I do wish there were a way to recognize patterns. It undoes some of my formatti...

OK, so, if I add a magic trailing comma to all of the method calls, then the call to python.exe -m black -l120 -S .\split_file_reader\ makes them all the same. But if I run it with -C, then it removes some of the magic trailing comma and returns the code to this mixed formatting state. Then, running python.exe -m black -l120 -S .\split_file_reader\ does not rstore the magic trailing comma

#

Sorry, I quoted a screenshot showing what I meant.

#

Is there an equivalent --insert-magic-trailing-comma option to aggressively insert them where possible?

dense jungle
#

No, there isn't. The point of the magic trailing comma is to preserve existing formatting to minimize diffs, not to create a different style

worn zenith
#

Fair enough

bright glacier
#

Someone is working on adapting the PEG parser generator to generate Python instead of C. I just accepted their additions in the google doc, it's still a WIP and would definitely need further work regarding whitespace, but I'm already surprised

austere lava
#

I've got indented blocks and simple statements parsing! ๐Ÿ™‚ I can now parse chained pass statements like pass; pass; pass; pass ๐Ÿ˜„

#

or even pass; ...; pass !

bright glacier
#

And continuing on the apparent train of progress in the past few days: I got the first development version of blackbench, my very own benchmarking suite for black, completed!

neon loom
#

All maintainers should be sent a black bench at once to sit on and celebrate.

bright glacier
#

lol, .. Anyway, blackbench 21.6.dev1 doesn't come with much benchmark data to run black against. It's basically a 21.6b0 checkout of black's source code + one string micro target (based off of Jelle's discovery of ESP causing slowdowns). I'd appreciate more code files to throw to black.

dense jungle
#

could be nice to have a mode that prints out how long every file took to format

#

that way I can run it on quora's codebase and see what's slowest

bright glacier
#

I'm looking for two kinds of targets: "normal" and "micro" . Normal targets should represent real-world code (so the benchmarking data actually represents real-world performance). Micro targets should be small and are focused on one area of black formatting (and mostly exist to measure performance in a specific area, like string processing).

bright glacier
dense jungle
bright glacier
#

I would be supportive of a such feature, although I think we would need to be clear that black won't be providing in-depth timing data ๐Ÿ™‚

#

I'm just scared of the slippery slope that could occur.

dense jungle
#

Agree, I might just hack it in locally and see the results

flat krakenBOT
flat krakenBOT
#
aviyadeveloper
[psf/black] Issue opened: #2335 Version 21\.6b0 breaks github action

Using the github action with the specific version or with the @stable tag directing to it results in the following error:

Run psf/black@stable
  with:
    options: --check --diff
    src: .
/usr/local/bin/python3: can't open file '/home/runner/work/_actions/psf/black/stable/action/main.py': [Errno 2] No such file or directory
Error: Process completed with exit code 2.

Running the action with:

psf/black@21.5b1

solves the issue.

bright glacier
#

Looks like I broke the GitHub action somehow

#

don't understand the failure though, will have to look into it later I guess

errant barn
#

I do find it odd that it would break that easily for the newest release version as well lemon_eyes could be worth considering moving to that branch setup you talked about

bright glacier
#

I have my doubts moving to a stable branch will fix it but I have no idea at this point.

#

Anyway, what a morning to wake up to

errant barn
#

I don't know a damn thing about it either ๐Ÿ˜… Yeah indeed.

flat krakenBOT
neon loom
#

I still donโ€™t get why the branch helps at all.

#

And no ones been able to explain it. Every time they try they go o yeah a tag will do the same โ€ฆ

errant barn
#

Yeah, someone said that pre-commit assumes the tag to stay in one place so there are problems with caching, but I can't claim to know anything about that ๐Ÿ˜„

#

Oh wait. The same goes for moving branches

lament crow
#

pre-commit is not meant to be used with branches, nor moving tags

#

but that doesn't mean you as maintainers can't use them

#

people just shouldn't use them with pre-commit

#

also, personally I don't think people should use stable tag/branch on their actions workflows either since that leads to unstable CI results whenever there's a new Black release

#

from what I understand, the main purpose of the stable tag is so that RTD generates the stable docs in addition to the latest docs that correspond to main branch

errant barn
#

That's a very fair point

flat krakenBOT
bright glacier
#

you have to configure a separate version for the actual version of black used by the action (well it defaults to whatever pypi returns as the latest release)

#

none of this still explains why it's breaking though

#

this workflow using psf/black@stable is working just fine ... ๐Ÿ˜‘

#

why is our integration story so damn frustrating ๐Ÿ˜ฆ

#

is caching the issue here?

#

btw, pre-commit is completely unrelated to all of this unless we're talking about branch vs tag

errant barn
#

The previous comments would suggest that, but I've no idea why 21.6 doesn't work

bright glacier
#

and even then, it's still IMO pointless to dicuss

#

until I can get a functional reproduction, I can't do anything to fix it

#

well, I think replacing all references to stable in our integration guides is probably a good idea

#

and maybe we could go as far to get rid of the stable tag since it seems to be an endless source of frustration

bright glacier
neon loom
#

Itโ€™s the price of moving forward. Stress less mate - people can roll back while we work it out !

#

There will always be issues โ€ฆ

#

Iโ€™d rather have issues then sit dormant โ€ฆ

bright glacier
#

I've not enjoyed working on the integrations side of this project tbh. But yeah, I do feel better taking a mental break from this mess.

tribal thistle
#

If you want to look at this in a more positive light, it just means a lot of people depend on your work

#

You and all the other maintainers who keep dealing with all this mess to provide a nice product

#

Thanks for your work

errant barn
#

Yeah, not so motivating when the work you did yesterday comes to bite you in the morning..

bright glacier
#

That's the nature of OSS for ya these days though, we'll get through it together, eventually ๐Ÿ˜€

plain atlas
bright glacier
#

People do open issues to say thanks and I really like that, but I haven't seen one for psf/black yet ๐Ÿ˜„

#

It's really rare tho

dense jungle
#

It feels a lot better to me than having an issue tracker where nothing ever happens ๐Ÿ™‚ if people open issues, that at least means the project is making a difference

frozen badge
bright glacier
#

ex. PyCQA/isort#1265

toxic stormBOT
errant barn
#

Oh wow ๐Ÿ˜„ never seen something like that. Awesome though!

bright glacier
#

atleast they opened an issue so we actually know there's something wrong

plain atlas
ivory vine
#

Can Black be easily integrated into PyCharm? Or do I have to go through the usual process of adding it as an External Tool

errant barn
ivory vine
#

Oh nice!

flat krakenBOT
#
andyphelps
[psf/black] Issue opened: #2337 Zero length stdin input causes black to fail \(regression\)

Describe the bug

Using latest version 21.6b0 piping an empty file into black as below causes "string index out of range". This is a regression from 21.5b2.

NOTE: referencing an empty file works as expected this is only when piping into - as example below.
i.e. this works:

touch __init__.py
python -m black __init__.py
All done! โœจ ๐Ÿฐ โœจ
1 file left unchanged.

To Reproduce

touch __init__.py
cat __init__.py | python -m black -
error: cannot format -: str...
bright glacier
#

Congrats on your first issue closures Felix!

errant barn
#

๐Ÿ˜„ Thanks!

flat krakenBOT
flat krakenBOT
bright glacier
#

๐Ÿ‘€ you're on a roll

errant barn
#

It seems like a ton of issues are essentially the same parens split happening only after two operators ๐Ÿ‘€

bright glacier
#

doesn't surprise me

errant barn
#

Yeah ๐Ÿ˜…

scenic shadow
#

photoblobWhy is your logo a parody of the Ford logo? I'm assuming that's what it is since they're so similar

glacial drum
#

"Any customer can have a car painted any colour that he wants, so long as it is black." - Henry Ford.

#

The Model T only came in black, so that efficiency and improved quality could be achieved. Spraying different colours would have required a break in the production line, higher increased costs, more staff, more equipment, a more complicated process, and the risk of the wrong colour being applied.

scenic shadow
#

Aha! That's clever

#

Though later wasn't ford also pioneers of the business model of releasing new designs and colours every couple of years so their new cars became more desireable, which is something companies like apple still do today.

vernal field
#

Can i make the black formatter stop formatting a certain piece of code

errant barn
vernal field
lament crow
# 
# fmt: off
subprocess.run([
    sys.executable, "-m", "pip",
    "install",
    "-upgrade",
    "-c", "constraints.txt",
    "foo", "bar", "baz",
])
# fmt: on
frozen badge
#

who owns black.vercel.app? would be curious whether the URLs cant be shortened by default

errant barn
#

Yeah that would be nice

lament crow
errant barn
#

anyways: here are more examples

frozen badge
lament crow
#

I like how you managed to find a bug while trying to provide example lol

errant barn
#

:D

flat krakenBOT
#
felix-hilden
[psf/black] Issue opened: #2339 Comment before a fmt: off line produces\.\. bugs

Describe the bug
Comments behave oddly before # fmt: skip lines. I'll try to see if this has been reported and narrow it down some more later, but for now here's a few cases.

To Reproduce

please_dont_do_this = heres_one(horrible, line(with_=lots, of=looooooooooooooooooooooong), names) # fmt: skip

This works as expected: [playground](https://black.vercel.app/?version=main&state=_Td6WFoAAATm1rRGAgAhARYAAAB0L-Wj4AC8AINdAD2IimZxl1N_WmOMtVsri9V-H1EXxMcIX_iOXr4PWWi...

bright glacier
#

๐Ÿ‘€

errant barn
#

Seems like it's eating up some characters from the earlier line

#

BTW: thank heavens (heavens being Josรฉ) for the Playground, it's so easy to provide examples there!

flat krakenBOT
#
davidhuynh1996
[psf/black] Issue opened: #2340 Issue: file "/urs/lib/python3\.7/ast\.py", line 35, in parse

#!/usr/bin/env python
import urllib
import json
import os
import RPi.GPIO as GPIO
import Adafruit_DHT

GPIO.setwarnings(False)
GPIO.setmode(GPIO.BCM)
GPIO.setup(16, GPIO.OUT)
GPIO.setup(23, GPIO.OUT)
GPIO.setup(24, GPIO.OUT)
GPIO.setup(25, GPIO.OUT)
GPIO.setup(18, GPIO.OUT)
pwm = GPIO.PWM(18, 500)
pwm.start(100)
GPIO.setup(13, GPIO.OUT)
pwm1 = GPIO.PWM(13, 50)
from flask import Flask
from flask import request
from flask import make_response

Flask app should start in g...

dense jungle
#

@errant barn you're much nicer than me, I would just have closed that pile of code ๐Ÿ˜„

flat krakenBOT
errant barn
errant barn
#

I swear, all the parentheses design issues and bugs and linebreak things are somehow connected, I'm going mad ๐Ÿ˜„ Gotta start diving into the splitting code I think.

bright glacier
#

And I'm completely checked out of black's development since I've been working on blackbench :D

errant barn
bright glacier
#

Working on version 21.6.dev2 :D

flat krakenBOT
bright glacier
#

oh damn the Duolingo bird is using our software

silent apex
#

Woah nice

bright glacier
#

Quick Q: for blackbench, would supporting Python 3.6 and 3.7 be helpful? Given that it's a dev tool for a dev tool (lol) I don't think it's super necessary to keep a large support range. Although Black itself does support 3.6.2+ so an argument can be made that some developers of Black might be running 3.6 or 3.7. Right now Blackbench officially supports 3.8+ although the test suite passes on 3.7. Some work will be required to achieve 3.6 compatibility.

#

(yes I do realize after posting that this isn't really a "quick" question per se ๐Ÿ˜… )

errant barn
#

Honestly, 3.6 EOL is in 6 months ๐Ÿ˜„

dense jungle
#

I think it's fine to require 3.8+ for running blackbench itself

errant barn
#

And I'm waiting eagerly for it

dense jungle
#

It may still be helpful to be able to run Black on 3.6 though

errant barn
#

Agreed

dense jungle
#

run Black inside of blackbench that is

#

running Black standalone on 3.6 is definitely useful because that's what I have to do ๐Ÿ˜„

bright glacier
errant barn
#

Fair enough ๐Ÿ˜… actually is that something that should be taken into account when we start chopping the supported versions to reach 3.6?

#

I'm all for cold-blooded action, but then again, I'm not the one stuck on those platforms

bright glacier
#

well the thing is that most people don't move their platform if they don't have to, so until python 3.6 usage goes down (which probably won't be happening anytime soon due Python 3.6 being shipped by default) we can't leave those users out in the cold

errant barn
#

I think that's reasonable

bright glacier
#

Thankfully 3.6 isn't that annoying to support (python 2 on the other hand...)

errant barn
#

It's not as big of a burden as Py2 I suppose

#

yeah :D

bright glacier
#

But OK, for blackbench, I'll keep the status quo, but I'd be down to officially support 3.7 too.

#

Jelle's idea is something I've thought about but I'm not sure how I would implement it, anyway it's definitely a feature for later

errant barn
#

What are you on btw? I remember it being Linux, but not Ubuntu then I guess.

bright glacier
#

Actually still Ubuntu, but Ubuntu LTS 20.04.01 ๐Ÿ˜„

errant barn
#

Ah :D

bright glacier
#

20.04.02 actually

errant barn
#

Oh right, 18 isn't LTS

bright glacier
#

18.04 is LTS

tender berry
#

It is

errant barn
#

Oh it's every two years, not four, nevermind

bright glacier
#

and the support window is 5 years ๐Ÿ‘€

errant barn
#

Yeaaah...

#

But fair enough for wanting to provide a stable system

bright glacier
#

Oh looks like GitHub moved the recent activity section on the dashboard

heavy heron
#

That's.. interesting

#

I'm not a fan of the recent activity section, it is often having everything but what I need

flat krakenBOT
#
davidhuynh1996
[psf/black] Issue opened: #2342 Python 3: Traceback \(most recent call last\):

Describe the bug

To Reproduce

Expected behavior

Environment (please complete the following information):

  • Version:
  • OS and Python version:

Does this bug also happen on main?

<!-- To answer this, you have two options:

  1. Use the online formatter at , which will use
    the latest main branch.
  2. Or run Black on your machine:
    • create a new virtualenv (make sure it's the same Python version);
    • clone this repository;
    • run...
twin schooner
#

does black follow pep8?

errant barn
twin schooner
#

Cool

foggy sapphire
#

yo just a question, does black have support for the match statement yet?

#

or the new type hinting features

dense jungle
#

no for match

#

not sure what new type hinting features you mean, but most recent typing changes don't involve new syntax, so Black should already support them

foggy sapphire
#

x: int | str

dense jungle
#

that's not new syntax, so it should just work

foggy sapphire
dense jungle
#

although I think there are some complaints about how we format it

little parrot
#

what are the complaints?

dense jungle
frozen yacht
#

I'm trying to figure out the overlap between flake8 and black. I notice the black project itself uses flake8. Very few of pycodestyle's errors are ignored. Is there a large overlap between pycodestyle and black? Do you not bother ignoring the overlpa cause it wouldn't make a difference?

dense jungle
#

But in general Black should solve pretty much all formatting-related things flake8 complains about, and Black output should mostly be flake8-clean

#

There are a few exceptions, I think the main one is spacing in slices

frozen yacht
#

flake8's styling just comes from pycodestyle I think. I want to figure out if pycodestyle does something extra. If not, I may be able to just use pyflakes alone rather than flake8, which would be fewer dependencies.

dense jungle
#

skimming that list I think all but the E4* and W6* errors should be handled by Black

frozen yacht
#

That's actually very helpful, thank you.

dense jungle
#

oh there's some more in E7* that aren't covered by Black

frozen yacht
#

Oh yeah, good catch. Those aren't really style issues so I wish those were in a separate library. I think bugbear covers a few of those. I'll try to look around for alternatives. I might end up using flake8 anyway just so I can customise pyflakes.

dense jungle
#

and I guess there's also line length. Black occasionally isn't able to make lines fit in the line length limit

frozen yacht
#

It has an optional feature that breaks up strings to address that, right?

dense jungle
#

yes, that fixes a lot of them but not all

frozen yacht
#

Is that considered stable now? I saw you talking about some bugs with it earlier.

dense jungle
#

we're getting close. I think the only remaining known issue is the performance one

frozen yacht
#

Oh that's quite the jump in time

#

Luckily for me I don't deal with 10's of thousands of lines of code

dense jungle
#

that's what quadratic performance will do for you ๐Ÿ™‚

heavy heron
lament crow
#

Stack Overflow seems like a terrible suggestion lol

errant barn
errant barn
lament crow
#

Asking on SO is doomed to fail :p

#

But searching SO for answers is a different deal!

errant barn
#

Yeah it can really be a trial by fire, but I find it very useful

lament crow
#

I'd say avoid it if you can lol

#

SO is a knowledge base so it's not a good place to ask most of the questions people struggle with

errant barn
#

Significantly more effort, and not for the most basic questions, that's true. But funnily enough I've found that even trying to formulate an acceptable question there helps a lot in solving the problem

bright glacier
#

Rubber duck debugging FTW ๐Ÿฆ†

plain atlas
#

Rubber duck debugging: explaining your code to a rubber duck to see what's wrong with the code

crisp copper
#

rubber ducky is not intimidating , and listens.

flat krakenBOT
heavy heron
#

Random curiosity, what do you use 2to3 for in Black?

little parrot
#

its for parsing

#

ast.unparse messes with comments and adds brackets and stuff

austere lava
#

to be precise it's a fork of 2to3 that's been heavily customized for Black

clear field
errant barn
clear field
frigid crown
#

h

bright glacier
#

I just made up a new term for this PR review I'm in the middle of doing lol: conditionalizing

#

i.e. adding if... elif... else blocks

flat krakenBOT
#
ichard26
[psf/black] Pull request review submitted: #2343 Adding TokenError to lib2to3 function

Hi there,

Thanks for the PR! It's great to see this getting closer to being fixed. I looked over your work and I have a few suggestions and requests to make, so let's get started:

It would be extra fantastic to catch and print out a prettier error message for IndentationError as well. The main difficulty would probably be getting the line no. and col. information out of the exception object since this exception is a built-in exception and has a different argument setup than TokenError (...

bright glacier
#

sigh, I don't like how commenting a thread in a PR within a review causes that comment to be duplicated, one version on the PR review thread and another one on the original comment thread :/

#

feels good to get back on working directly on black though, I needed a little break from blackbench

#

has anyone tried using python_version environment marker in an entry for additional_dependencies? I tried this and it ain't working :/

  - repo: https://github.com/pre-commit/mirrors-mypy
    rev: v0.902
    hooks:
      - id: mypy
        exclude: ^docs/conf.py
        additional_dependencies:
          [
            types-dataclasses >= 0.1.3; python_version<'3.7',
            types-toml >= 0.1.1,
            types-typed-ast >= 1.4.1,
            types-click >= 7.1.2,
          ]
#

oh wait, I'm being an idiot

#

I forgot my python environment is 3.8

#

although for some reason, mypy is still complaining about missing dataclasses stubs even on 3.8 when types-dataclasses is not installed

#

I guess this is due to python_version=3.6 in mypy.ini, so it looks like we will need to install types-dataclasses unconditionally? /cc @dense jungle

flat krakenBOT
dense jungle
bright glacier
#

Yeah... but I don't think bumping python_version is a suitable solution here, so unconditional install it is. Thanks for confirming my thinking!

opal shuttle
#

In Django, how does it know I'm creating a Choice (model)? I'm following the official tutorial, but I'm confused how it knows this is a Choice... ```python

Give the Question a couple of Choices. The create call constructs a new

Choice object, does the INSERT statement, adds the choice to the set

of available choices and returns the new Choice object. Django creates

a set to hold the "other side" of a ForeignKey relation

(e.g. a question's choice) which can be accessed via the API.

q = Question.objects.get(pk=1)

Display any choices from the related object set -- none so far.

q.choice_set.all()
<QuerySet []>

Create three choices.

q.choice_set.create(choice_text='Not much', votes=0)
<Choice: Not much>
q.choice_set.create(choice_text='The sky', votes=0)
<Choice: The sky>
c = q.choice_set.create(choice_text='Just hacking again', votes=0)

Choice objects have API access to their related Question objects.

c.question
<Question: What's up?>

And vice versa: Question objects get access to Choice objects.

q.choice_set.all()
<QuerySet [<Choice: Not much>, <Choice: The sky>, <Choice: Just hacking again>]>
q.choice_set.count()
3

#

it's about 3/4 of the way down the page.

#

sorry wrong channel!

bright glacier
#

no worries

dense jungle
#

yeah you're using single-quoted strings. what is this

delicate marsh
#

hahahaha

neon loom
#

So, I'm being lazy here - What's left for us to do to enable string processing on by default?

dense jungle
#

the performance issue only I think

neon loom
#

What do we think we need to do here? Seems your preformatted setting didn't help. Should we try get Lukasz to think about it or anyone else?

dense jungle
#

I haven't tried that hard to get this to work; I feel like some variant of preformatted should work but haven't had time to try

#

also have you tried ESP on facebook codebases? wouldn't be surprised if there's other lurking issues we'll find as we throw enough code at it

bright glacier
#

OK, I hope this is the last pipenv lock I will have to do for the mypy 0.9xx PR

bright glacier
#

oh fun, I never checked if our docs build fine with sphinx 4 now the upgrade isn't blocked by myst-parser

#

i'mma do that right now

#

and the answer on this linux box is that yes it does work phew

#

although I don't understand why typed-ast is removed by the pipenv lock run I just did, but I'm no pipenv wizard and have no idea whether it's fine or not (although I did do a pipenv install --dev afterwards and it worked fine on my 3.8 environment so ...?)

#

And because I got distracted in the mypyc issues again, I found this lovely gem: mypyc/mypyc#707. Seems like this could've broken a few things in our codebase in a more stripped environment. I wonder what other Python incompatibilities currently exist (that I don't know of) ... hmm

toxic stormBOT
neon loom
flat krakenBOT
bright glacier
#

Oh lovely ๐Ÿ˜…

flat krakenBOT
flat krakenBOT
flat krakenBOT
#
MarcoGorelli
[psf/black] Issue opened: #2345 "If someone would like to push this \[Jupyter support\] PR to completion, I would be happy to reopen\.

https://github.com/psf/black/pull/1814 was closed, with the message:

If someone would like to push this PR to completion, I would be happy to reopen.

I'd be quite interested in seeing black support Jupyter Notebooks natively. Indeed, I wrote nbQA for the purpose of running this and other tools on notebooks, but if notebooks could be handled by the tools themselves, that'd be even better.

A few questions / discussion points before I begin:

  • wh...
flat krakenBOT
bright glacier
#

FYI I want to review the stdin pr, I have a nitpicky suggestion :)

#

unfortunately I am really busy this morning tho, so that will have to wait

flat krakenBOT
flat krakenBOT
bright glacier
#

fyi again: I no longer want to do a review for the stdin pr, jelle's review made me realize that my nitpicky suggestion isn't really necessary

flat krakenBOT
tribal thistle
#

isn't this true of literally any python package tho

#

ah Jelle got to it

flat krakenBOT
dense jungle
#

we fixed something like this at work by having a shell script that first cds to an empty directory and then invokes the Python code. don't think that's something we can do for black though

tribal thistle
#

Could do what PIP does and bundle the dependencies with the package

#

But that's nightmare fuel

dense jungle
#

that wouldn't even work with the standard library

#

and I feel like this kind of thing tends to happen more with the standard library because it has more generic names

tribal thistle
#

I'm thinking along the lines of copying the content of the files into black_number. Would be really annoying to handle upgrades, and deal with different python versions.

dense jungle
#

including the standard library? yes that would be a nightmare

heavy heron
#

Well, in fact, you can alter the python path when black is starting up to make sure to prioritize the right imports. Although not sure how portable that would be.

dense jungle
#

That would still only solve part of the problem, because site.py still runs first

austere lava
#

This is only a problem if black is installed in the same virtualenv you're working on

bright glacier
#

interesting change

Screenshot_from_2021-06-23_14-53-32.png
#

not sure if I like it, first impressions: I feel like hovering over the issue number was good enough and now issue mentions will be too noisy

lament crow
#

it's task lists

bright glacier
#

oh

#

๐Ÿ˜…

lament crow
#

in public beta rn

#

they also are planning many changes to project boards

#

well, not project planning in general as it's not going to be just boards anymore

bright glacier
#

oh wow

#

damn those are some rather big changes

#

thanks for letting me know

flat krakenBOT
bright glacier
#

and lemme guess that Jelle will also merge that PR lol

#

incredibly quick

flat krakenBOT
bright glacier
#

or not lol :p

bright glacier
#

I may or may not have forgotten to install a version of black in my blackbench CI ... ๐Ÿ˜…

#

... yeahhh my test suite failed on GHA ๐Ÿ˜›

#

I should probably add a runtime check to make sure that black is actually installed in the environment *before* running any benchmarks

bright glacier
#

wait how is this green? pytest clearly failed ...

#

I mean the test failures are valid, but I would've expected the workflow to be red all over the place, not green ๐Ÿ˜•

plain atlas
bright glacier
#

I've never really understood the point of following - I mean, I don't care either way

dense jungle
late dewBOT
#

.github/workflows/ci.yaml line 51

run: |```
dense jungle
#

it's essentially running pytest; coverage and the exit code of the first command gets ignored

bright glacier
#

oh yeah that's a good guess lol, lemme look into it

#

and reading the github docs, this annoying behaviour only happens on windows because fast-fail isn't supported for cmd.exe ???

#

that's surprising

#

wait a sec, but gha windows uses pwsh which does have fail-fast enabled by default?

tribal thistle
#

That seems to be the case

#

I can't actually see the coverage run

#

I do see output, just not the command itself tho

plain atlas
bright glacier
tribal thistle
#

Hmm

#

It feeds all commands at once to the pwsh executable, which may have something to do with it

#

perhaps not

#

Hmm docs say it does use the last status code

bright glacier
#

but yeah, the tests are failing cause of slash vs backslashes, how lovely lol (this was a big source of pain for the action rewrite)

#

although it says this for pwsh:

Fail-fast behavior when possible. For pwsh and powershell built-in shell, we will prepend $ErrorActionPreference = 'stop' to script contents.

#

I don't have much experience with powershell, so it's definitely possible I'm misreading what $ErrorActionPreference actually configures

#

anyway it doesn't matter, I changed it so the workflow just uses bash everywhere now so now it's failing as expected ๐Ÿ˜„

#

Thanks for the help @dense jungle, @tribal thistle, and @plain atlas !

plain atlas
#

I'll take the credit for being here

#

๐Ÿ˜› scale and jelle did it all, i only followed you

bright glacier
#

I sorely underestimated how much initial documentation I would write for blackbench ๐Ÿ™ƒ

#

I am currently writing the "running benchmarks" section and still got to write the "analyzing results" bit. This all still ignores the fact I need to write contributing.md ๐Ÿ˜…

#

I originally thought I wouldn't need docs beyond a readme but it looks like I do

#

thankfully documentation is literally my specialty so it shouldn't be too much work once I get the contents written up first

frozen badge
#

thankfully documentation is literally my specialty
what are your favourite tools that make documentation easier?

bright glacier
#

For while iterating on docs, I find sphinx-autobuild to be very helpful

#

For simple (github) markdown documents, there's grip

#

!pypi sphinx-autobuild

late dewBOT
bright glacier
#

!pypi grip

late dewBOT
#
grip v4.5.2

Render local readme files before sending off to GitHub.

bright glacier
#

sphinx-programoutput is a simple but works well way of (excluding an encoding issue) including click based CLIs help output dynamically (although for argparse I'd recommend sphinx-argparse-cli which is specifically tailored for argparse)

#

sphinx-copybutton and sphinx-inline-tabs are also good tools for easily improving the reader experience

#

And this isn't really a tool per se, but unsurprisingly, learning some of the many directives and roles provided by docutils and sphinx is IMO worth it

#

Sphinx's extlinks, intersphinx, and autodoc features also deserve a honorable mention

#

So yeah, that's a bunch of what I've used over the past year of documentation (mostly for psf/black)

#

I hope something out of that dump was useful ๐Ÿ˜„

frozen badge
#

thanks, appreciate the info

flat krakenBOT
#
whitespace-negative
[psf/black] Issue opened: #2348 Failing to parse single\-quote f\-string with backslash inside it

Describe the bug

The formatter is failing to parse single quote f-string as it cannot include a backslash inside it.

To Reproduce

  1. Take this test.py containing:
f"\"{b}\"{' ' * (long-len(b)+1)}: \"{sts}\",\n"
  1. Run black without arguments.
  2. See error -->
  File "/var/task/black/__init__.py", line 986, in assert_equivalent
    dst_ast = parse_ast(dst)
  File "/var/task/black/parsing.py", line 125, in parse_ast
    raise SyntaxError(
f'"{b}"{\' \' * ...
dense jungle
#

^ I thought that was also a syntax error at runtime but apparently not

#

backslashes are disallowed in the {} part of f-strings though

bright glacier
#

Me too, I thought it was an invalid issue (or maybe about an unhandled exception) but it's not ๐Ÿ™‚

flat krakenBOT
#
bzoracler
[psf/black] Issue opened: #2349 Inline comments shift strangely in multi\-line function calls

Describe the bug

The last inline comment in multi-line function call shifts to underneath the last parameter in fmt: on...fmt: off block

To Reproduce

Running default black arguments on this code block

def f(**kwargs):
    pass


p = f(
    # fmt: off
    param_1=1,  # First comment
    param_2=2,  # Second comment shifts <---------
    # fmt: on
    param_3         =               5 # This line Needs formatting
)

Formats it to

def f(**k...
echo ridge
#

Hi, I'd like to use black for my package. But I'm still in the early phase of developing, and I'm using comment blocks to organise ideas. Is it possible to tell black to not re-format comment blocks? I've got something like this: https://paste.xinu.at/JWlPV/python, which black will turn into this: https://paste.xinu.at/PDSyPY/python

So the differences aren't huge, but I try to run tox (which calls black) before each push, and I keep forgetting to reset these comments. It would be awesome if I could just tell black to ignore the comments.

class.py
frozen badge
#

have you considered doing the indentation after the # instead of before?

echo ridge
#

hmm, I haven't. the indentation was pycharm's idea. let me try that

frozen badge
#

i tested in black.now.sh and it does preserve indention inside the comment, so that should work unless some other tool is interfereing.

#

otherwise you would need to put # fmt: off before and # fmt: on after to tell black to ignore everything in between

#

oh, literally the embed above has a bug about this tho

echo ridge
#

oh but it's good to know this formatting command exists. thanks!

#

the indent inside the comment works, that's cool for me then

#

thanks!

frozen badge
#

generally comments on the same level should also be easier for IDEs to detect and allow for mass-uncommenting. i know VSCode has issues with comments which dont start on the same level

echo ridge
#

mass-uncommenting isn't a deal here tho because it's not commented out code but just additional info

white peak
#

I can't seen to figure out ...how to write out file on python?

austere lava
late dewBOT
#

Grammar/python.gram line 622

| (tok='!=' { _PyPegen_check_barry_as_flufl(p, tok) ? NULL : tok}) a=bitwise_or {_PyPegen_cmpop_expr_pair(p, NotEq, a) }```
dense jungle
#

from __future__ import barry_as_flufl makes != illegal and uses <> instead I believe

austere lava
#

I've spent way too much time on this kekw (โ•ฏยฐโ–กยฐ๏ผ‰โ•ฏ๏ธต โ”ปโ”โ”ป

dense jungle
#

It's an absolutely crucial feature

austere lava
#

clearly

#

I could be laying the groundwork here for the only CST parser in the planet which supports this feature

dense jungle
# 
>>> from __future__ import barry_as_FLUFL
>>> 1 <> 2
True
>>> 2 != 1
  File "<stdin>", line 1
    2 != 1
       ^
SyntaxError: with Barry as BDFL, use '<>' instead of '!='
bright glacier
#

!pep 401

late dewBOT
#
**PEP 401 - BDFL Retirement**

Link

Status

April Fool!

Created

01-Apr-2009

Type

Process

bright glacier
#

damn the April fools joke is older than I would've initially thought

#

oh lol, lib2to3 actually has support for that syntax

dense jungle
#

but black doesn't, somehow

bright glacier
#

apparently the built-in AST module doesn't support it :(

dense jungle
#

cpython should be ashamed of itself

#

indeed, black --fast does work

muted haven
# bright glacier apparently the built-in AST module doesn't support it :(

Sadly there is no flags parameter for ast.parse() but you can always use compile

>>> import ast
>>> import __future__
>>> tree = compile('a <> b', '<test>', 'eval', flags=ast.PyCF_ONLY_AST | __future__.CO_FUTURE_BARRY_AS_BDFL)
>>> ast.dump(tree)
"Expression(body=Compare(left=Name(id='a', ctx=Load()), ops=[NotEq()], comparators=[Name(id='b', ctx=Load())]))"
>>>
flat krakenBOT
#
ImmoWetzel
[psf/black] Issue opened: #2350 Style question

Hi,

we use black in the company to format the python code for some small internal tools. But nobody could explain why black converts a readable code into non readable.. There is for sure an idea behind this.
Which ?

example:


    parser.add_argument("-o", "--output_dir",   default=None, help="fixed folder of the BaseConfig log output and modified XML file; or automatic /tmp/ ")
    parser.add_argument("-a", "--arg_file",     default=None, help="Robot Project argument file (when...
foggy sapphire
#

can black format multiple files at once?

#

or do i just have to type each path

dense jungle
#

yes, you can pass a directory for example

frozen badge
#

it can format entire folders at once if you give it the folder path. i always just do black . to format everything in the current directory

foggy sapphire
bright glacier
#

Hmm ok, so I just finished doing some proper benchmarks (like system tuned + blackbench) to check mypyc's performance increases (with my initial compatibility changes). With safety checks always running format is 1.74x faster. With safety checks disabled format is 1.90x faster. And blib2to3 parsing is 1.72x faster. That's rather disappointing TBH. I doubt my mypyc specific optimizations skills (ie. basically none) will really do anything meaningful to that (although I do have a string specific optimization that does work ready).

dense jungle
#

That's still a great speedup! I wonder if we can find places where it's not as efficient as possible with some profiling

#

Or find places where the generated C code is not as efficient

bright glacier
#

(although to be clear, no shade is intended at mypyc, there's probably some unideal issues with mypyc favourability in our codebase)

frozen badge
#

is blackbench at a state where anyone can easily use it?

#

i would love to throw viztracer and scalene at it to see where the holdups are

bright glacier
#

usable: sorta? easy or user friendly: no. I'm working on the second development version and I'm making progress but it's not ready in terms of features nor actual profiling data. The documentation will also need some work.

bright glacier
frozen badge
#

both of them handle subprocesses just fine, with scalene even handling gpu bound code

bright glacier
#

whelp, guess who has zero experience with viztracer nor scalene lol

frozen badge
#

if anything this can serve as a recommendation to you then, they are both great profilers

bright glacier
#

I've been just using cProfile + yelp-gprof2dot all of this time ๐Ÿ˜›

frozen badge
#

yeah cProfile is not really something i would use for any large project

#

never heard of the other

bright glacier
#

sorry I meant that yelp-gprof2dot is the data frontend for cProfile's pstats output

frozen badge
#

ah, i just use snakeviz for cprofile

bright glacier
#

it produces a nice visual call graph

#

I'll look into those profilers as I try again at optimizing the codebase for mypyc, thanks for the recommendation!

frozen badge
bright glacier
#

yep!

frozen badge
#

i was hoping for the bot

bright glacier
#

ah, I thought you were asking if that change was the hackiest bit, which is right

frozen badge
#

ye

#

thats quite

#

yeaaaa

bright glacier
#

but hmm I guess a feature enhancement is due for that bot?

#

it is pretty rare though so it would be reasonable to not support it

frozen badge
late dewBOT
#

src/black/linegen.py lines 368 to 371

# HACK: functions (like rhs) compiled by mypyc don't retain their __name__
# attribute which is needed in `โ€‹run_transformer`โ€‹ further down. Unfortunately
# a nested class breaks mypyc too. So a class must be created via type ...
rhs = type("rhs", (), {"__call__": _rhs})()```
dense jungle
#

I feel like the fact that we're using __name__ in the first place is the hacky bit ๐Ÿ˜›

#

Overall this doesn't look bad, several of the changes actually improve clarity or are neutral

bright glacier
#

I might go with your old suggestion from when ambv's pr was originally reviewed

bright glacier
bright glacier
#

hmm I found py-spy and it's pretty nice and supports C extensions!

#

I tried viztracer and scalene and while they look good, I think I'll need some more time to get used to them

frozen badge
#

yeah, personally scalene fills all need i have for profiling extensions, be they C or something else, but the output format needs to be learned

frozen badge
#

TFW black doubles the length of this script im supposed to refactor Sweat

bright glacier
#

I'm doing more benchmarking with a few micro optimizations this time and for some reason compiled black is way more instable during benchmarks, the performance data varies way more :/

#

it's still better than a completely untuned system though so yeah, at least my tuning wasn't useless lol

bright glacier
#

Another side effect of compiling black is that the import time is better. A quick dirty benchmark shows 1.17x increase running "import black" and compiling more of black might improve the times even more (although it's a tradeoff between import time and distribution size)

heavy heron
#

Random curiosity, what do you mean by compiling?

#

Is Black using C extensions?

little parrot
#

I think it's using mypyc

bright glacier
#

Whelp, my (probably) naive optimization attempts netted me a 3% (which could easily be noise) increase in performance for parsing only (translates to a 1% boost for format-fast) ... OK, time to finally test an optimization for strings that I feel rather confident about, I want at least one win here ๐Ÿ˜…

bright glacier
#

I'm going to call it a day in terms of squeezing out more performance out of mypyc compiled black for now. I might (read: probably) pick it up again later, but this has been too much benchmarking in a short timeframe ๐Ÿ˜„ . I think I'll start investigating automatically building wheels for all of the platforms black targets (or continue work on blackbench, I'm not sure, just something different lol).

dense jungle
#

I recently learned about cibuildwheel, it's great for building windows/mac wheels easily

bright glacier
#

Did push that string micro-optimization though, but then again, it was a general optimization, not specific to mypyc. It was just avoiding some unnecessary regex compiles which actually represented 10% IIRC of the total runtime for the strings micro target (i.e. that list of strings that Jelle discovered causes slowdowns with ESP).

bright glacier
dense jungle
bright glacier
#

oh hmm, well I'll see, all of this is just me guessing ๐Ÿ˜„

plain atlas
#

Ngl that's a very weird image for a repo

bright glacier
#

uhhh

dense jungle
#

Black for now is even simpler than asynq since there's no build step

#

It's my profile picture ๐Ÿ˜„

#

Might want to change it at some point

bright glacier
#

the cat is a nice touch

bright glacier
#

Status update: I managed to get black compiled using cibuildwheel as a manylinux2001 wheel for 3.6, but unfortunately the test suite breaks because it looks like I missed a few tests which use mocks (that aren't supported by mypyc compiled black) with the special incompatible_with_mypyc marker

#

Other than forgetting a lot of stupid and dumb things, it overall is going pretty smooth

#

well so far, Windows and MacOS might be annoying (I've been only testing locally on Linux). I hope they don't need any compiler toolchain changes unlike Linux which needs clang.

#

I wonder if I'll need to mark every single test that uses mocks as incompatible, locally I didn't have to but for 3.6 more tests are failing than I expected.

#

So it's far too early for any sort production grade testing yet, but we are getting closer to shipping compiled black, so it's probably time to increase the testing range from "one guy and their x86-64 Ubuntu 20.04.02 Python 3.8.5 environment" to something a big bigger ๐Ÿ˜‰

#

*once I get at least the Linux portion of this automated wheel build workflow working

#

(and yes this is me rambling as I wait for the workflow to do its thing while hoping it doesn't fail)

silent apex
bright glacier
#

I'll write you down as a potential tester, thanks!

#

(now I need to find somewhere to keep these notes lol_

#

YAY the workflow didn't crash!

#

although the performance increase is even lower now, ~55% ๐Ÿ˜ข

#

will need to run proper benchmarks though (just maybe it actually won't be as bad as it looks from rough testing) , but that will wait until later cause I'm still tired of running benchmarks lol

flat krakenBOT
#
adamklaff
[psf/black] Issue opened: #2351 "IndexError: list index out of range" on all files

I'm running black via pipx, and I am getting this on every file I run black on. Any ideas?

  File "/Users/user1/.local/bin/black", line 8, in 
    sys.exit(patched_main())
  File "/Users/user1/.local/pipx/venvs/black/lib/python3.9/site-packages/black/__init__.py", line 1128, in patched_main
    main()
  File "/Users/user1/.local/pipx/venvs/black/lib/python3.9/site-packages/click/core.py", line 1137, in __call__
    return self.main(*args, **kwarg...
flat krakenBOT
heavy heron
mint igloo
valid halo
#

Hi, I am just getting started with linting and stuff. Is there any way to make black compatible with flake8 instead of making flake8 compatible with black? (max-line-length and E203)

errant barn
bright glacier
# flat kraken

It's probably worth an enhancement to catch errors from pathspec and present a more understandable error message about your .gitignore being invalid

bright glacier
#

ImportError: DLL load failed: A dynamic link library (DLL) initialization routine failed., whelp this is beyond my knowledge of windows :/

frozen badge
bright glacier
frozen badge
#

Based on the above issue, plausibly these could be causing the issue:

The right MSVC redistributable is not available.
Mix of 32-bit Python and 64-bit compiled module, or vice versa, somehow. This will clearly not work.

bright glacier
#

was about going to copy and paste that issue ๐Ÿ˜„

frozen badge
#

are visual studio build tools installed in the environment?

bright glacier
#

uhh, the thing is that I don't really know what I'm supposed to be looking for

frozen badge
#

hm, yea, that seems good

bright glacier
#

Apparently a workaround is just using interpreted mypy[c]

frozen badge
#

actually, i think i might have found it

#

the windows env only has MSVC redistributable 2010

#

it has 2019 minimal and debug runtime but im not sure whether that works

bright glacier
#

Hmm, I did see some issues about the visual studio on gha windows (v2019) linking to a rather new DLL, that seems like a good guess

bright glacier
#

Oh lovely, I can't get installing a newer runtime via choco to fix the issue, I guess I'll just use that interpreted mypy workaround

solid adder
bright glacier
#

In other news: who wants to bet that backslashes vs slashes are once again a source of pain for me ...

#

sigh - you thought I would've learned by now to be extra viligant

solid adder
#

The fear of touching unrelated lines is backwards. Yes, with one giant commit you reformat the world and you need to ignore this particular commit from browsing history with git blame (easy to do). However, when you do this single commit, everything that comes after is semantic changes so your commit history is clean in the sense that it actually shows what changed in terms of meaning, not style.

With tools like darker, you forever expose yourself to incremental commits that are a mix of semantic changes with stylistic changes, making it much harder to see what changed. Additionally, this gives up some of Black's most important features around making diffs smaller. But the worst thing that since you never applied a single sweeping reformatting commit, there is no way for you to ever decouple formatting changes from semantic changes while browsing git history. It's simply impossible because the changes are mingled with partial reformattings.

bright glacier
#

And a technical concern too: I don't see any logic regarding the safety checks format_file_contents implements poking around at the source code.

solid adder
#

What do you mean?

bright glacier
#

Darker passes in the code via format_str which doesn't implement the safety checks. It's probably fine (especially since the instability issues are pretty much fixed by now) but it's not ideal.

#

It's minor compared to the issues you presented above though ๐Ÿ™‚

tribal thistle
#

You make a fair point, but to offer an alternative perspective, youโ€™re putting a lot of faith in the tool working, because to be quite frank, no one is going to review a change that reformats the entire code base (esp on larger projects).

This is slightly alleviated by having things like unit tests & making sure your program just runs, but in my experience (all my experience is open source and personal, I donโ€™t program professionally) unit test coverage is never high enough.

solid adder
#

@tribal thistle this is why Black guarantees AST equivalence. Yeah, now we have this one exception which is docstrings but whitespace in docstrings is so unlikely to break anything that it's nit-picking at this point.

tribal thistle
#

Thatโ€™s something I hadnโ€™t considered

lament crow
late dewBOT
#

src/darker/__main__.py line 115

verify_ast_unchanged(```
lament crow
#

not that I would recommend the tool nonetheless, just posting for correctness :P

flat krakenBOT
#
trolliama
[psf/black] Issue opened: #2352 Black doesn't format scrapy file on vscode

Black doesn't format scrapy file

I am doing a scrapy project and i am unable to format the spider file. I am using vs code and i tried to format the file with cntrl+shift+p and "Format Document with..." and didn't work. I also used the command $ black -l 79 file_name.py and nothing happened

import scrapy


class PokemonSpider(scrapy.Spider):
    name = "pokemon"
    allowed_domains = ["pokemondb.net"]
    start_urls = ["https://pokemondb.net/pokedex/all"]

    d...
bright glacier
#

Status update on the black-mypyc-wheels situation: Linux is good to go for expanded testing and Windows has been successfully compiled for. Unfortunately the generated wheel fails to pass the test suite, failing numerous times with "Reached allegedly unreachable code!" which doesn't sound particularly good :-)

#

I do have a windows machine, but it's been a while since I've last booted it so it will take a while to debug it.

lament crow
#

and not like darker isn't already messing the diffs anyway ๐Ÿ˜„

bright glacier
#

Yep, it's more of an annoyance than a risk

flat krakenBOT
red sedge
#

darker seems like a technical marvel though

#

I'm glad it didn't exist when I was considering black because I'd definitely have used it

flat krakenBOT
flat krakenBOT
#
cdce8p
[psf/black] Issue opened: #2355 No trailing comma for one argument function definitions

Describe the style change

Black should not insert a trailing comma for one argument function definitions. This would be consistent with the style for two and more arguments.

Examples in the current Black style

# trailing comma is added after 'some_argument'
def onnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnne(
    some_argument,
):
    pass


# for two+ args -> no comma is added
def twwwwwwwwwwwwwwwwwwwwwwwww...
flat krakenBOT
flat krakenBOT
bright glacier
#

Full of magic
I love that description for the trailing comma label ๐Ÿ˜„

flat krakenBOT
flat krakenBOT
flat krakenBOT
#
ichard26
[psf/black] Issue opened: #2359 Present a more user\-friendly error if \.gitgignore

Is your feature request related to a problem? Please describe.

Yes. When .gitignore is invalid, Black crashes with a confusing IndexError exception and it's not obvious that the .gitignore file is the issue. This is worsen by the fact, it's non-obvious Black uses .gitignore at all (maintainer note: it's been documented so it's fine, just describing how this is even worse of a problem than at first glance).

This is also a problem if say reading fails for some reason.

*...

heavy heron
flat krakenBOT
#
felix-hilden
[psf/black] Issue opened: #2360 Make SRC/CODE mandatory and mutually exclusive

Is your feature request related to a problem? Please describe.
From #2356, which made (assuming it got merged) SRC appear as a mandatory CLI parameter: Strictly speaking, SRC is optional both in the way that Black accepts no arguments (and then prints "nothing to do") and with --code "". I think accepting SRC and then skipping it when -c is passed is a bit confusing.

Describe the solution you'd like
I'll propose we make SRC and --code mutually exclusive and require one of...

flat krakenBOT
#
felix-hilden
[psf/black] Pull request review submitted: #2278 Parenthesize conditional expressions

Thanks for the submission! I'd be more inclined to accept this if the tests contained counterexamples and other cases as well. Like imploding a short enough line or exploding from a single line (not already multiline like in the test). How about test expressions elsewhere, like positional arguments, or in entirely different contexts? Anyways, I think this is a solid addition to the style!

Also, I'm not entirely familiar with the visiting code, should the new visit_test be called somewher...

plain atlas
#

<@&831776746206265384>, see above, user id 819219783576190976 if deleted before investigation

sonic tundra
#

!warn @subtle ravine Please don't promote here without permission.

late dewBOT
#

:incoming_envelope: :ok_hand: applied warning to @subtle ravine.

plain atlas
#

i mean Hi :D

sonic tundra
#

Thanks for reporting this!

plain atlas
#

np

plain atlas
#

this is one of them

plain atlas
#

the esoteric one liner that you wrote

heavy heron
#

Yup

heavy summit
#

...if you did, run an antivirus

plain atlas
#

oof. I only saw that a person was in the wrong channel and advertising. Didn't even go that far.

#

never clicked the links

#

thanks for the info tho

bright glacier
#

And just finished writing a module called clickhacks.py for blackbench ๐Ÿ˜…

#

The lack of spacing between options in the help output was bothering me too much so I ended up subclassing click.formatting.HelpFormatter ๐Ÿ™‚

tribal thistle
#

:O
Not super relevant, but that's reminded me I should go check on coup

bright glacier
#

Coup what?

tribal thistle
#

lol let's pretend my keyboard is broken and I meant to say cloup all along

frozen badge
#

!pypi cloup

late dewBOT
#
cloup v0.9.1

Adds features to Click: option groups, constraints, subcommand sections and help themes.

tribal thistle
#

(FWIW, they've merged the changes quite a while back, but if you're fine with subclassing the formatter, that's probably a more logical solution than adding coup)

frozen badge
#

coup

#

@tribal thistle fix your L key

tribal thistle
#

mao

deft wyvern
#

We payed

bright glacier
#

Damn y'all with the games ๐Ÿ˜„

#

Anyway, I got this patch that adds a MANIFEST.in which cuts the sdist size by roughly 2x, should I add a changelog entry? (seems IMO overkill but it is technically user-facing)

#

I'm going to go with "no" (also saves me an edit-commit-push cycle).

flat krakenBOT
bright glacier
#

yeah ... I couldn't think of any more "goodbye" expressions off the top of my head for English so I ended up using French ๐Ÿ˜›

#

Good thing the commit message probably isn't really that important in this instance ๐Ÿ˜…

plain atlas
#

||Noel||

frozen badge
bright glacier
#

cue yeethon's dramatic entrance

flat krakenBOT
#
cooperlees
[psf/black] Pull request review submitted: #2362 Add MANIFEST\.in to cut down sdist by ~2x

Why do we want to do this? I've had experiences where I've wanted to use projects Docker files, run CI + linting locally before, so I feel we'd want most of these files ... Will go with that others think, but I like having all the config Giles. the .github and action etc. make sense to prune.

I guess we can say to do this (and I guess the way I usually do it) is you clone the repo. dist is the minimum to just "build" the python project ...

flat krakenBOT
#
ichard26
[psf/black] Pull request review submitted: #2278 Parenthesize conditional expressions

First of all, the formatting changes are a strong +1 personally. I found this example in the primer output and it's so much more readable now:

--- pyanalyze/typeshed.py	2021-05-30 03:06:13.755715 +0000
+++ pyanalyze/typeshed.py	2021-05-30 03:06:32.534930 +0000
@@ -448,13 +448,13 @@
                 arg = arg.replace(kind=SigParameter.POSITIONAL_OR_KEYWORD)
             cleaned_arguments.append(arg)
         return Signature.make(
             cleaned_arguments,
            ...
flat krakenBOT
#
felix-hilden
[psf/black] Pull request review submitted: #2362 Add MANIFEST\.in to cut down sdist by ~2x

Yay! This seems pretty reasonable ๐Ÿ‘Œ

Maybe also:

  • global-exclude *.py[cod] __pycache__ *.so
  • prune docs/_build

although I'm not sure if they are needed because we release with a workflow. Perhaps the first should be added just to be sure.

Are all other files really included by default? I was under the impression that each data file needed to be included explicitly. That might have something to do with the setup configuration though.

flat krakenBOT
#
felix-hilden
[psf/black] Pull request opened: #2363 Use setuptools\.find\_packages in setup

Hi! Just a quick PR changing our setup a bit for the better. If we were to introduce new modules or refactor the source further down into submodules, maintaining the list in setup may become a pain to deal with. No more.

find_packages only includes proper packages, so I had to add an init file to primer. Shouldn't affect anything though, since it's already used as a package.

Feel free to shoot this down if it's not what we want! I didn't feel like this was big enough to go through an ...

bright glacier
#

It's been like a year of me using Vim and now I'm finally customizing it to include more than just a few basic tweaks ๐Ÿ˜„

mint barn
#

does black change all of your ' with "

silent apex
#

It should if I remember correctly

bright glacier
#

It does, unless you pass -S / --skip-string-normalization.

flat krakenBOT
#
faizanalhassan
[psf/black] Issue opened: #2364 INTERNAL ERROR: Black produced different code on the second pass of the for matter

Hi, today I found a strange error while running black. It also referred to a log file, which might be helpful to resolve the error. Please check the logs:

Mode(target_versions={, , }, line_length=120, string_normalization=True, experimental_string_processing=False, is_pyi=False)
--- source
+++ first pass
@@ -125,12 +125,12 @@
                 # Search on the base of sale no if no search key is found
                 qry = qry.filter(cls.no.ilike(f"%{no_search_key}%"))
            ...
sour verge
flat krakenBOT
errant barn
#

I feel a bit wrong tagging bugs that are resolved in new releases with invalid. How have you dealt with them? Should we even set up some dedicated tag, like R: old?

lament crow
# sour verge How do I set global config on Mac? From what I can see (https://black.readthedoc...

Per the documentation:

You can use a โ€œglobalโ€ configuration, stored in a specific location in your home directory. This will be used as a fallback configuration, that is, it will be used if and only if Black doesnโ€™t find any configuration as mentioned above. Depending on your operating system, this configuration file should be stored as:

  • Windows: ~\.black
  • Unix-like (Linux, MacOS, etc.): $XDG_CONFIG_HOME/black (~/.config/black if the XDG_CONFIG_HOME environment variable is not set)

Note that these are paths to the TOML file itself (meaning that they shouldnโ€™t be named as pyproject.toml), not directories where you store the configuration. Here, ~ refers to the path to your home directory. On Windows, this will be something like C:\\Users\UserName.

#

emphasis mine

#

that would mean you should create a ~/.config/black file, not ~/.config/black/pyproject.toml

sour verge
#

Aha! Awesome thanks ๐Ÿ™‚

silent apex
#

Does anyone know of a strict flake8 confg that matches with black's code style?

heavy heron
#

Isn't there a flake8 plugging that basically runs `black --ceheck

plain atlas
#

that'd be kinda useless

#

๐Ÿ˜›

#

if you mean black --check, that's useful. and iirc its a pre-commit thing

heavy heron
#

lol yeah

tribal thistle
silent apex
#

Thank you!

tribal thistle
#

Though I'm not sure that a flake compliant run would be black compliant

#

To check that something has the black style, the easiest method would probably be running --check like ARU mentioned

silent apex
#

Ty

#

I'm trying to add both into my CI

flat krakenBOT
#
hauntsaninja
[psf/black] Issue opened: #2365 Document "fmt: off" under Usage, docs site is hard to use

I found myself wanting to turn black off for a block of code, so I searched my brain and it came up with # format: off. This didn't work, so I looked up black's docs. I went to: https://black.readthedocs.io/en/stable/ from which I clicked on "Usage and Configuration" from which I clicked on "The basics" from which I expanded some drop downs (whose content isn't searchable until expanded), which led me nowhere. This is a lot of clicks and no result compared to the old days when everything wa...

bright glacier
#

Sigh, and I thought I was done with major documentation changes ... there's a totally valid oversight pointed out in the above issue

#

I got some ideas to fix it (that don't involve changing the layout again) but I wish I didn't have to deal with another structural issue ๐Ÿ˜…

errant barn
silent apex
bright glacier
#

Outdated could also work with PRs so sure I'm happy with that.

#

OT: that's one cute cat picture! @silent apex

silent apex
#

Ty, found it on the internet and loved it

dense jungle
#

I like "outdated"

bright glacier
#

@errant barn and R: outdated has been created!

flat krakenBOT
#
JelleZijlstra
[psf/black] Pull request review submitted: #2278 Parenthesize conditional expressions

Thanks, agree that this is an improvement in formatting. I might just go edit pyanalyze now to apply this formatting :)

I feel like this could cause stability problems if Black inserts newlines in the first run, then parens in the second. Could you add some tests where everything is in one line in the original input?

The str(node) is indeed a bit awkward. An alternative could be to search for something like newline tokens in the children of the node, but I feel that that may be more f...

flat krakenBOT
bright glacier
#

I can't tell whether I'm adding useful commentary or am just ranting on the documentation structure issue ๐Ÿ˜…

dense jungle
#

it is useful!

bright glacier
#

I guess if it's any consolation for my poor brain, at least one day I'll get the docs in a good place, and I think it's fair to say they are better than how they were before.

#

Just looks like that road is longer and more bumpy that I'd have hoped ๐Ÿ˜„

#

Good excuse to get back into adding to the documentation again tho! I kinda forgot about it (getting distracted by mypyc and then blackbench stuff) and there's still some docs - mostly contributing docs - that I want to write but haven't found the motivation to do so yet.

bright glacier
#

gosh I'm booting windows up since like 11 months ago and the repeated "switch to Microsoft products please" screen is such a surprise to me from Linux ๐Ÿ˜†

#

Me: "no I don't need to pin the free office app to my taskbar windows... "

#

and then they have the audacity to say "All done! Windows is all yours. Have fun!" ๐Ÿ˜‚

#

Ah now time for my favorite things of all: windows update!

#

Oh dear there's also the 20H2 feature update to do too ๐Ÿ˜”

#

at least that one is optional

#

but anyway, time to figure out how to build black with mypyc on windows ... Locally!

#

the one thing I will note is that Windows UI looks and in general feels better (FTR for me!)

#

and the mouse pad experience is better too (palm rejection actually works consistently)

#

But in all seriousness, I like both Ubuntu and windows (I dual boot them for a reason!)

plain atlas
#

๐Ÿ‘€

bright glacier
#

what, is liking Windows scandalous?

plain atlas
#

i use a user agent spoofer on *.google.com and microsoft websites so they think i'm using their browsers and don't nag me

plain atlas
#

nah i use windows

#

and use vsc

#

and windows terminal

#

and edge on occasion

#

and the store

#

and winget

#

etc, i use way too many microsoft products

bright glacier
#

and GitHub too lol

plain atlas
#

and onedrive

#

and outlook sometimes :shudders:

bright glacier
#

And by extension azure (I've heard GitHub actions is powered by azure)

plain atlas
#

and powershell but that's excusable

plain atlas
#

but windows and linux are on azure

bright glacier
#

I use Microsoft authenticator too

plain atlas
plain atlas
#

the onedrive app is complete bollocks on android tho

bright glacier
#

the backup feature is why

plain atlas
#

the entire android app integration with onedrive is completely and utterly the worst thing i have ever seen

bright glacier
#

I would use Google's 2fa but it doesn't backup and I don't trust myself to be careful enough to not lose my 2fa creds

plain atlas
#

there is an extremely useless feature in the onedrive app

#

extremely

bright glacier
plain atlas
#

its the pin lock

#

i use both

silent apex
plain atlas
#

.โ€ฆ.that's horrible.

silent apex
#

In plain text, I might add

silent apex
#

But I do it

plain atlas
# plain atlas i use both

my windows computer is signed in to one drive and i didn't realize that until recently but basically my documents and shit are in the onedrive folders

bright glacier
#

well unless there's like some rouge employee, you should be fine (tm)

#

aw crap I probably won't be able to do one thing I planned to do so on my windows installation

#

My visual C++ installs are quite modern

#

still can investigate the "reached apparently unreachable block" issue though

plain atlas
#

i also use like insider every windows software

#

oh and

#

i have customized my cmd shell lmao

bright glacier
#

I customized my fish shell and vim setup but not windows because I hate it for development

plain atlas
#

i need to work on the colors but

silent apex
#

So much nicer

plain atlas
#

they aren't really the same

#

i use powershell

#

but here's a funny thing

#

what's most cursed about this image?

cmd_1KTEgSJqbl.png
bright glacier
#

I don't like a single update has completed yet in terms of windows update

#

think*

#

the backslashes for me personally

#

The amount of times my code breaks because of them is uncountable

plain atlas
#

lol

silent apex
#

I'm so confused

plain atlas
#

i also use cmd

bright glacier
#

seems about right (that time is wrong btw, don't worry, I'm not up that late! lol)

Screenshot_192.png
plain atlas
#

rgh

#

is vscode insiders not insiders right now

#

@bright glacier oh btw you should disable print spooler for the time being

bright glacier
#

oh yeah i forgot about that issue

#

good call

plain atlas
#

also don't run chdsk

bright glacier
#

? is that vulnerable too (wouldn't be surprised if so)

#

I'm sad now.

unknown.png
errant barn
#

Also @bright glacier, try to not be too hard on yourself ๐Ÿ˜… I stand by my words on the issue that the documentation restructuring was essential. We went from a heap of issues in the documentation to significantly less! Which rather than having no issues, is not incompetence, just realistic ๐Ÿ–ค

flat krakenBOT
#
andrewv99
[psf/black] Issue opened: #2366 Black silently changes expression type when wrapping expressions in parenthesis

Describe the bug

Black sometimes wraps an expression that is already long (enough), or possibly silently becomes long (enough) as a result of refactoring across a codebase and then applying Black formatting to it. The result is that the expression becomes a tuple of length 1, which is not the same as a scalar, and that change in the code by Black then silently causes type sensitive bugs to manifest that were not there before.
[BlackTrailingCommaBug.zip](https://github.com/psf/black/fi...

errant barn
#

Also, also: I'm off for a week of disconnect and lack of electricity by the lake, so I'm not here to discuss. Feel free to do as you please for my PRs and such if you'd like to take them forward!

lament crow
#

the app used to completely freeze when I tried to open my Documents folder which has 601 items at current moment ๐Ÿ˜ฌ