#black-formatter

1 messages Β· Page 22 of 1

lament crow
#

warning aren't really meant to be raised when they're not set to raise through the warning filters (or the -W flag)

bright glacier
#

(the date isn't set in stone FYI, I just chose that on the fly)

#

oh I was holding down ctrl

tribal thistle
#

Would it look better at the bottom?

bright glacier
#

I'm raising it within format_str so that's not really possible

tribal thistle
#

Right, okay

bright glacier
#

emitting it against an explicit --target-version py27 is easy within the main function but it gets harder when source autodetection is the only way for us to tell

#

Although yeah, this would get messy with multiple files as multiple warnings would be emitted

#

I suppose we could set a global flag and then access it within main but I don't know how that'd work with multiprocessing as the format_str function doesn't have access to the manager object.

#

Well, format_file_in_place could get a manager parameter but that feels like overengineering at this point

flat krakenBOT
bright glacier
#

I wonder what's the chances I'll get angry comments saying we shouldn't do this ^^

#

The raw LOC currently in the codebase related to supporting Python 2 isn't actually that high (~200-400?) but it's the other assumptions and guarantees we won't have to hold anymore that is the major win. For one CI will get a lot less complicated as we won't have to run specific parts of the test suite twice, once with the python2 extra and once without.

#

Two, there's formatting decisions being impacted by Python 2 which is .. annoying to say the least.

#

I'm sure there's more but that's the main ones from my perspective.

flat krakenBOT
haughty vector
#

Hi,
Question: Do you know of some "legacy code" migration cases to use Black as a formatter?
In the company I work for, we want to use Black for some old services but we want to know "success stories" in this type of migration.

Thanks for your answers (Black is a great tool!)

dense jungle
haughty vector
dense jungle
bright glacier
#

very incomplete haha

flat krakenBOT
bright glacier
#

we all have the power to make the world burn by breaking all of the CI πŸ˜„

mint barn
#

black is so cool

#

fits all my needs

unique mulch
#

There also exists a Flake8 plug-in for use with black

flat krakenBOT
#
wacul-yamada
[psf/black] Issue opened: #2524 Does not work on Apple Silicon

Describe the bug

black panics on Apple Silicon

To Reproduce

Traceback (most recent call last):
  File "/opt/homebrew/bin/black", line 5, in 
    from black import patched_main
  File "/opt/homebrew/lib/python3.8/site-packages/black/__init__.py", line 13, in 
    import regex as re
  File "/opt/homebrew/lib/python3.8/site-packages/regex/__init__.py", line 1, in 
    from .regex import *
  File "/opt/homebrew/lib/python3.8/site-packages/regex/regex.py", line 4...
flat krakenBOT
flat krakenBOT
#
Shivansh-007
[psf/black] Pull request opened: #2525 feat\(config\): Support \`\-\-config\` in pyproject\.toml

Description

You can now specify config key in pyproject.toml which does the same job as the CLI option --config. The path of config should be relative to the CWD it is running from. If the file is not found or is invalid i.e. it is a folder rather a TOML file it would raise click.exceptions.FileError, and if the file can't be parsed it would raise tomli.TOMLDecodeError. All black config keys in the custom config should be registered under [black] and they would overwrite th...

flat krakenBOT
#
Shivansh-007
[psf/black] Pull request opened: #2527 feat\(config\): Add support for path specific config

Lol, I am sorry for spamming 3 PRs. Had completed them before vacation and didn't get time to open them until now :)

Description

Woohoo! Module specific configuration! and a bit of restructuring of code in the black/__init__.py file. Mainly shifting all the "reformatting" logic to a separate function to increase the readability of code in the click CLI function. You can go through the code in the commit, code should be self explanatory, though I still need to update the documenta...

flat krakenBOT
red sedge
#

December is an odd time to do the removal

#

Just push it like 32 days and you get v22

errant barn
#

would be cool to do it at the same time with removing the beta also

#

In that sense the new year would be fit for purpose πŸ˜„

red sedge
#

Eg I plan to remove 3.6 support in twisted 22 rather than 21.12

#

Even though python 3.6 is EOL in 21.12

#

Also you might want to consider a black21 pypi package so people can have both installed

bright glacier
#

How would that work? like rename the entry point?

#

Also I'm not too convinced it'll help much compared to them just pinning black that's bound to happen after we drop support

bright glacier
red sedge
bright glacier
#

But what's the benefits?

red sedge
#

And yeah it would need the entry point renamed

#

It would mean I could have both installed for when I'm working in say VSCode

#

But yeah I could just set two paths and have two virtual environments

bright glacier
#

It's going to conflict with a newer version as they would have the same package name and entry point

red sedge
#

No they'd need to be renamed package and entry point

#

Yeah maybe it's not so useful

#

I suspect you'll see one emerge anyway

bright glacier
#

Yeah but given we strongly recommend pinning black hopefully less?

red sedge
#

Right but it's about being able to have both in one virtual environment

bright glacier
#

I just don't like it as it feels like people are going to come up and claim we will have to support it indefinitely as people will rely on that compatibility trick

red sedge
#

You could even have black[python2] just require black21 and run black21 when you have a python2.7 target version

bright glacier
#

But yes we could do it

red sedge
#

Then you could remove the python2 parsing today

bright glacier
#

I suppose it's still less of a maintenance burden since we won't be doing any feature developed on such a black21 project but I will bet some packaging stuff will go wrong one day and we'll be asked to fix it

red sedge
#

I don't use Python 2, so I'm unbiased

#

But yeah I hate to be in the position to ask for an extension, it's just to make the version number pretty

bright glacier
#

I honestly don't really care for Python 2 as I've literally never used it.

red sedge
#

Wow

#

I've pretty much only done large scale 2 to 3 migrations, very little actual feature development

bright glacier
#

I don't have Python 2 code to even migrate ignoring some bug cases people have sent in for black

#

I haven't been in the Python scene for too long haha

errant barn
#

But yeah, dropping beta at the same time was just a quick suggestion, doesn't really matter when we do it

flat krakenBOT
#
peterjc
[psf/black] Issue opened: #2528 Missing documentation on magic comma

Is this related to a problem? Please describe.

I am aware from the issue tracker of the "magic comma" whereby a trailing comma can sometimes influence what black does, but cannot see where this is documented - e.g. https://black.readthedocs.io/en/stable/faq.html#

Describe the solution you'd like

Have the "magic comma" documented.

Describe alternatives you've considered

I have browsed the issues but haven't hit on a succinct summary.

**Additional conte...

flat krakenBOT
sour jay
red sedge
#

It would be 21.12b0 right?

errant barn
#

I get the major version increment though :D

red sedge
#

It's technically not a major version because calver

#

But it feels like it

flat krakenBOT
neon loom
#

I see no reason to wait.

#

Also, people and CPU architectures ... Thanks for inflicting this on us Apple.

errant barn
#

To wait for pulling off the beta?

#

Or Py2 support or both πŸ˜„

flat krakenBOT
#
uriva
[psf/black] Issue opened: #2530 commet after brackets has odd style

i often see that when writing this:

some_function(some_other_function(tuple), another_function),  # my long comment that the formatter doesn't like

the autoformatter does this:

some_function(
    some_other_function(tuple), another_function
),  # my long comment that the formatter doesn't like

i would prefer it to do this:

# my long comment that the formatter doesn't like
some_function(some_other_function(tuple), another_function),
somber bridge
#

is black safe to use on production code?

little parrot
#

yep black should error out if the ast is different from the original

#

(it shouldnt actually change the ast)

red sedge
#

Don't run it on your production server though

dense jungle
#

Unless your production server is running the black playground

flat krakenBOT
#
pradyunsg
[psf/black] Pull request review submitted: #2529 Define a stability policy

A few suggestions inline, and a note.

Define this stability policy as something that applies to the stable releases, not all releases. I would also recommend against making any formatting changes when this policy is adopted -- it would be a good idea to have a beta release prior to the stable release, that contains any potentially disruptive changes. IMO your last beta release and first stable release only have one difference -- the version number -- that's how you know it's good enough ...

bright glacier
#

Wow a metric ton of stuff happened while I was alseep and then busy. I'm still busy unfortunately but in 2.5 hours I'll try to get caught up and get reviews dealt with.

#

I suppose my action to deprecate Python 2 convinced you @dense jungle to stop us from sitting around and doing nothing about the stability problem ^^

#

And thank you for that by the way! I wouldn't have done that personally as it's not really my domain.

mint barn
#

Hyy

bright glacier
#

Hi

flat krakenBOT
red sedge
#

Well it looks like black is switching to semver that also happens to be calver

dense jungle
flat krakenBOT
#
sbalasa
[psf/black] Issue opened: #2531 Python 3\.10 keywords aren't recognized

Describe the bug

Python 3.10 keywords aren't recognized

To Reproduce

Run black when match keyword is present in the file

For example:

  1. Create a file sample.py having match keyword
  2. Run black -l 80 sample.py
  3. See error -->
error: cannot format sample.py: Cannot parse: 2:10:     match option:
Oh no! πŸ’₯ πŸ’” πŸ’₯
1 file failed to reformat.

Expected behavior

Black formatter should format 3.10 related syntax/keywords too.

**Environment (...

flat krakenBOT
flat krakenBOT
#
ichard26
[psf/black] Pull request review submitted: #2526 \`\-\-verbose\` mode prints location of project root

Thanks for submitting the PRs! This one looks the easiest to review out of the bunch so I'll be tackling this one first ^^

I know this is a minor feature but this should at least get a changelog entry as it is an user-facing change.

Also, why do we want to emit a message with the sources passed on the command line? I don't see how it's useful and it can be a bit confusing in some cases like this:

~/programming/oss/black/docs on branch feat/project-root-verbose [$?] via Pyth...
bright glacier
#

Yeah our configuration and options handling code needs some work but it's hard figuring out what's supposed to be the refactor and the feature addition.

#

There's some things of the refactor from a quick skim I don't like, but whatever those are fixable. What's not fixable is the commit history. I really would have hoped the refactor would've came in its own commit but nope.

#

Not sure what the pipfile.lock changes are about, but they're definitely broken

bright glacier
#

OK, I just went over most of the changes that happened over the 24 hours. Is this a correct summary of the current situation?

  • Formatting Python 2 will be deprecated in the next release
  • Formatting Python 2 support will be removed right on January 1st, 2022 (because people still treat CalVar as SemVar for the most part) as part of a normal beta release (22.1b0?)
  • On the same day or soon after we adopt the stability policy and publish a stable release containing only documentation changes and a version bump (22.1.0?)
bright glacier
dense jungle
#

Honestly I don't feel strongly about the date, but thought it'd be useful as an indication that Black 22 will be a clean break to some extent

#

Fewer things for users to remember if last beta version == last Python 2 version == last 21 version

bright glacier
#

Yeah I suppose. I do like Pradyunsg's idea that the stable release should be as simple as possible to really indicate we're going stable, but it would make the Python 2 removal more confusing with the current timeline.

dense jungle
#

Maybe we release 22.0rc0 first, then 22.0.0 hopefully unchanged

bright glacier
#

basically my rundown but with a rc marker instead of the beta marker haha

dense jungle
#

got to use all the available options in versioning πŸ™‚

#

maybe we throw an 22.0a0 in just for completeness

bright glacier
#

So wait we are in agreement and I just am misreading all of this? πŸ˜… there's nothing to debate

#

I thought there was haha. I still do think Pradyunsg's suggestions are overall a good idea but I don't think they're controversial.

bright glacier
dense jungle
#

and format everything using tabs and single quotes

bright glacier
#

speaking of that, what's the game plan with ESP? We'll need to get it landed before we adopt the stability policy or else it's going to be tabled for a year 😐

dense jungle
#

There's still the performance issue. I can try to find some time to figure it out

bright glacier
dense jungle
#

Cooper also recently ran into an issue where it doesn't format well, I think you were also on that thread

bright glacier
#

Yup, something about function calls not respecting user defined splits even though that's part of the ESP rules.

bright glacier
dense jungle
bright glacier
#

I tried for a little bit and I couldn't either 😦

bright glacier
#

@tired shard hello!

#

(it took like 5 minutes to figure out what your discord username was πŸ™ƒ)

#

ah they're busy and off discord for the time being

errant barn
#

Yeah CalSemVer for the win πŸ˜… But a yearly release cycle doesn't seem that bad so I'm all for it. We got to pick some number anyway if we want to promise stability.

drowsy sonnet
#

Hi. It's Friyay.

errant barn
#

Let's merge everything πŸ₯³

drowsy sonnet
#

There's no reason to stick with CalVer if you really want SemVer. You can just switch. Version numbers are cheap. 🀷

errant barn
#

I do like the idea of a yearly cycle still, switching to SemVer completely would just mean more major releases along the year I guess.

drowsy sonnet
#

The main benefit of anchoring to the calendar IMO is the ease of planning of changes.

#

That's why pip is anchored on the year.

drowsy sonnet
# bright glacier <@!740498628854153236> you should probably read this as a general rundown from w...

Idk where the Dec 1st date came into being (I aggressively unsubscribe from issues, because... uh... 1k unread notifications is bad enough). I wouldn't promise a date.

Just say Python 2 is deprecated and will be removed in the first stable release. And also note in the stability policy that you're adopting annual formatting changes; without making formatting changes as part of the first stable release.

If you cut the stable release in Feb 2022 or April 2022, nothing is going to break. Give yourselves the flexibility to do things at whatever pace, instead of promising deadlines to largely ambivalent users.

#

That's my 2 cents anyway.

red sedge
#

Don't get into a situation where you promise a removal and then don't do it

#

Eg what CPython do, it means you lose trust

errant barn
#

Fair enough πŸ‘πŸΏ

red sedge
#

But also people need to see black as a free gift not as something with an SLA that just went down

red sedge
errant barn
#

Hmm interesting!

red sedge
#

The foo he's talking about parsing is an X.509 certificate and the database is on the computer of every person in some reasonably large country

flat krakenBOT
#
felix-hilden
[psf/black] Pull request review submitted: #2525 feat\(config\): Support \`\-\-config\` in pyproject\.toml

Richard also raised a good point about chaining this value in the original issue. I tend towards chaining being supported, which would require a loop of some sort to follow the TOML files.

Also, I think it would be more consistent to just require all TOMLs to have the same structure, i.e. a tool.black section instead of some other magical section. That way we could probably reduce the implementation to recursion, which wo...

dense jungle
#

For what it's worth, I don't think my proposal really is "CalSemVer": we promise no changes in formatting, but that doesn't mean no other new features. For example, we'd still be able to land Jupyter Notebook support or Python 3.10 support in the middle of the year.

#

The guarantee is just that if Black formats your code right now, it will keep formatting it the same way for the rest of the year

errant barn
#

Yeah that was a bit tongue in cheek anyway πŸ˜„ but to be fair, semantic does allow to add features in minor versions. Now if we want to make breaking changes to other than formatting mid year that's a different discussion, and I think we should be able to as long as it's communicated, which the changelog will do!

bright glacier
#

It also gets messy when we introduce an official API for black - now SemVar could have two distinct meanings (one is more obvious but it's still not great)

errant barn
#

Yeah true, the guy advocating for semver on the policy issue had good points but I also feel like it isn't worth it to switch

visual trench
#

hi, I am amateur, i need a road map for begin python tutorial, do u have some good refrences to read ?

tribal thistle
#

Hey there! You’ll get an answer to that question over in #python-discussion. This channel is specifically for the black formatter tool

bright glacier
#

FYI this is very off-topic here as this is a topical channel about the Black autoformatter project. Not sure what's the server's position on advertising / promotion, but this isn't the right place for this.

void harbor
#

Sorry! I thought it's for all opensourse projects.

tired shard
flat krakenBOT
#
Zac-HD
[psf/black] Pull request opened: #2532 Fix feature detection for positional\-only arguments in lambdas

Description

While working on shed, I discovered that feature detection for positional-only arguments doesn't work in lambdas, though the docstring of get_features_used() suggests that it should. I've therefore added syms.varargslist to the check, and a test case which demonstrates that it works.

Checklist

  • [x] Added a CHANGELOG entry
  • [x] Added test cases
  • [x] This change does not need documentation
bright glacier
#

I should really get back to working on diff-shades because enforcing no formatting changes without something like mypy-primer is going to be a pain.

plain atlas
#

ooo you got a new flower

#

i mean hi

#

hello

bright glacier
#

!remind 1D Get back to work on diff-shades

late dewBOT
#
Sure.

Your reminder will arrive on <t:1634089012:F>!

ID: 3261
bright glacier
#

actually, not at 9 PM lol

#

!help remind edit expiration

late dewBOT
#
Query "remind edit expiration" not found.

Did you mean:
remind edit duration
remind edit mentions
remind edit content
remind edit

plain atlas
#

!remind delete iirc and make new

bright glacier
#

!help remind edit duration

plain atlas
#

wat

#

did not know that was a thing

bright glacier
#

!remind edit 18H 3261

#

!remind edit 3261 18H

plain atlas
#

!remind edit duration 3261 18H

bright glacier
#

!remind edit duration 3261 18H

late dewBOT
#
Sure thing!

That reminder has been edited successfully!

ID: 3261
plain atlas
#

:)

bright glacier
#

there, jeez lol

#

I'm bloody staff and I still can't get the reminders command right :p

plain atlas
#

could you please hit the trashcan on some of those help messages? πŸ₯Ί

#

since they're so big

#

seriously

#

it takes up my whole screen

bright glacier
#

happy now?

plain atlas
#

well, no, technically I'm never happy.

bright glacier
#

sigh, how are you then today?

plain atlas
#

monotone

bright glacier
#

relatable

plain atlas
#

I haven't written code for my project in days.

#

just.... fixing a bunch of workflow problems πŸ₯²

bright glacier
#

Haha yeah, the debugging experience with GHA is not great to say the least

plain atlas
#

thankfully I haven't started on those yet

bright glacier
plain atlas
#

more like, wrote some more scripts for pre-commit

bright glacier
#

yellow not an interesting colour for ya?

plain atlas
#

well....

#

i was going to say "ay, you have a new colour"

#

but forgot to

bright glacier
#

my short term memory sucks so that's understandable ^^

bright glacier
plain atlas
#

lol

plain atlas
bright glacier
#

I was invited a long time ago but I only recently accepted staff.

bright glacier
plain atlas
#

because poetry has like either 50 or 100 dependencies and my script has.... one dependency.

bright glacier
#

ah

#

pre-commit itself has quite a few dependences though

plain atlas
#

i already use pre-commit so that's not a problem

#

and even so, the script can be run alone

bright glacier
#

(I know this because I've checked black's CI setup and pre-commit is responsible for like 50% of the installed test dependencies)

#

we don't even use pre-commit in CI so yeah 🀷

plain atlas
#

it also takes like, 20 seconds less than poetry πŸ₯²

bright glacier
#

^-- that's my new favourite ducky fyi

bright glacier
plain atlas
#

tldr rather than submit a pull request to fix poetry I would rather use pip and a bunch of hacks to get around it

#

halloween

bright glacier
#

got any good nick suggestions for halloween?

plain atlas
plain atlas
#

...itchard?

#

yes i was gone just to think of that

#

wow i am smart

bright glacier
#

I don't like it when people call it "i-chard" for personal reasons, hate how it sounds

plain atlas
#

oh dang

#

TIL

#

i've uh

bright glacier
#

been living eh?

#

breathing that O2

plain atlas
#

never thought the other one

bright glacier
#

wait what are you talking about?

plain atlas
#

i-chard

#

actually wait

#

i'm not sure how i used to pronounce it ....

bright glacier
#

Jake said i-chard

#

in the owners Q&A :'(

silent apex
#

The one you dislike

bright glacier
#

the hard ch

#

I-card is fine actually

silent apex
#

ok, I've always mentally pronounced it i-card

tired shard
#

Wee helloe, richard could you see the msot recent comment on the --verbose PR

bright glacier
plain atlas
#

welp time to go back to dealing with merge conflicts

#

i mean fun times

#

seriously i just wanna code again 😒

bright glacier
#

everyone's favourite pasttime amirite /s

plain atlas
#

yes

#

i have so many

#

sooooo many workflows and stuff i want to do

#

but the end goal will be so worth it

bright glacier
plain atlas
#

just a few clicks and a new version is on its way on docker, read the docs, pypi, etc

bright glacier
#

!remind 18H also look at the --verbose latest comment from Jason

late dewBOT
#
Aye aye, cap'n!

Your reminder will arrive on <t:1634068451:F>!

ID: 3262
tired shard
#

Alright, that's fine

plain atlas
#

hello jason, could you take a look at the configuration pr? πŸ₯Ί

bright glacier
tired shard
#

I commented in workflow thread arl, using release feature of github to publish on pypi, docker, run tests, discord notification, etc

plain atlas
#

but like, black already has pypi and docker scripts

bright glacier
#

so you copied and pasted, heh πŸ˜„

tired shard
#

yeah, just catching up with all the development on my contribution projects

plain atlas
#

i will be, perhaps

#

haven't added them yet

#

but i did realise a few days ago

bright glacier
#

that

plain atlas
#

one sec

#

these are a thing

#

so I'll set up a workflow that makes a new branch, runs some things, commits it, and then opens a pr

#

probably*

bright glacier
#

Yup yup I've used workflow_dispatch before

bright glacier
plain atlas
#

yeah i might not do that too

bright glacier
plain atlas
#

the key reason i want to do it is in case i need to make an urgent bugfix release while not at my pc and unable to use git

bright glacier
bright glacier
plain atlas
#

yeah

#

i honestly

#

likely won't make that full workflow

#

and if i do, its much later in time

#

like, when there's not much else to do

#

hmmm

#

while i'm thinking about all of this, and er, which channel we're in

#

I have to realise, I've almost never thought about my formatting

#

since nearly the very beginning of the project

bright glacier
#

honestly the more I program the less I use black consistently

plain atlas
#

oh?

bright glacier
#

Yeah, I used to run black on all of my scripts and whatever, but nowadays I don't run it nearly as much because there's some elements of the style I hate a lot.

plain atlas
#

yeah, i had to turn its format off for a line of my code

bright glacier
#

Mostly with breaking on the LHS vs RHS and how nested brackets are treated

plain atlas
#

but that's because the line of code is super genius and black was messing it up

late dewBOT
#

scripts/export_requirements.py lines 26 to 33

# fmt: off
MESSAGE = textwrap.indent(textwrap.dedent(
    f"""
    NOTICE: This file is automatically generated by scripts/{__file__.rsplit('/',1)[-1]!s}
    This is also automatically regenerated when an edit to pyproject.toml or poetry.lock is commited.
    """
), '# ').strip()
# fmt: on```
bright glacier
#

Oh yeah and that's another wish item

#

putting more than one function on a single line

#

that's probably not a good idea to enforce consistently though

plain atlas
#

(i'm kidding of course, this line isn't actually super genius)

#

but i like what i did here πŸ˜›

plain atlas
flat krakenBOT
rose moth
#

Hi guys, i use black in vs code. I want to disable it putting trailing commas, how can I do the same?

#

I am not sure if this is correct, I just added the arg --skip-magic-trailing-comma into my black args in vsc

unknown.png
#

Nope, still not working

dense jungle
late dewBOT
plain atlas
#

get to work!

bright glacier
#

ugggh I am super busy with other stuff today

plain atlas
#

same!

#

lol i was opening diff shades

#

and ended up on next-pr-number

#

again

#

where is the source for diff-shades?!

late dewBOT
bright glacier
plain atlas
#

ah okay

granite hearth
#

How would use see what black formats exactly

tribal thistle
#

I’m not a 100% sure what you’re asking, but I think you’re looking for the β€”diff option

flat krakenBOT
#
seetchong
[psf/black] Issue opened: #2534 black 19\.10b0 crashes

Run black in pipenv shell command line.

% black kueh.py
error: cannot format kueh.py: INTERNAL ERROR: Black produced different code on the second pass of the formatter. Please report a bug on https://github.com/psf/black/issues. This diff might be helpful: /var/folders/qs/vt7pvr2d3mlctnlr3zs52q3w0000gp/T/blk_vt7wm5xo.log
Oh no! πŸ’₯ πŸ’” πŸ’₯
1 file failed to reformat.

To Reproduce

  1. Take this file 'kueh.py'. [kueh.py.zip](https://github.com/psf/black/files/7336314/kueh.py....
flat krakenBOT
golden steppe
#

Does this adhere to PEP 8?

errant barn
#

Black? Yes!

rose moth
dense jungle
rose moth
#

I could so it from source but I think that would be redudant just for trailing commas and I would like to get updates on black

dense jungle
rose moth
#

Oh well, thanks

flat krakenBOT
plain atlas
#

what even was that

dense jungle
#

good question

red sedge
feral turret
#

yo guys can you help me

#

i trying to lern how to work in pyton scipt...

lament crow
#

This is not the right channel, that I can tell you for sure

mint barn
#

!json

late dewBOT
#

When using JSON, you might run into the following error:

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

This error could have appeared because you just created the JSON file and there is nothing in it at the moment.

Whilst having empty data is no problem, the file itself may never be completely empty.

You most likely wanted to structure your JSON as a dictionary. To do this, edit your empty JSON file so that it instead contains {}.

Different data types are also supported. If you wish to read more on these, please refer to this article.

plain atlas
#

@bright glacier i've finally remembered I can make issues so other people can do things πŸ₯²

bright glacier
#

awesome

vestal birch
#

Can someone give me a project, one which use python+ dbms/file handling

flat krakenBOT
#
paris-ci
[psf/black] Issue opened: #2536 INTERNAL ERROR: Black produced code that is not equivalent to the source when a docstring contains a control character\.

Describe the bug
Black couldn't format a file in my project, generating code that does not equal the source.

To Reproduce

error: cannot format black_error.py: INTERNAL ERROR: Black produced code that is not equivalent to the source. Please report a bug on https://github.com/psf/black/issues. This diff might be helpful:

           value=
             Constant(
               kind=
                 None,  # NoneType
               value=
-                'I ...
lament crow
#

I hate when people disregard templates

flat krakenBOT
errant barn
#

Wait, do we have an option to not look for a configuration file?

rose moth
#

Hey guys, how can I disable black to do this? Its too short and just adds extra lines to the code and makes it confusing

unknown.png
errant barn
#

As Jelle said yesterday, Black generally isn't configurable πŸ˜…

#

As the line gets longer and there are a few arguments, breaking the line is much more readable

flat krakenBOT
#
felix-hilden
[psf/black] Issue opened: #2537 CLI flag to disable looking for a configuration file

Is your feature request related to a problem? Please describe.
Given that Black always looks for a configuration file, up to the file system root (!), it could be convenient to be able to explicitly not look for any configuration files. This would be advantageous when debugging different options.

Describe the solution you'd like

$ black --config -
$ black --no-config

Or maybe I've just missed it, didn't look at the code yet. In which case it should be document...

rose moth
#

I wish we could specify after how many parameters it would break the line

errant barn
#

It wouldn't really be consistent, because long lines would have to be broken anyway. Line length is the way to go in my opinion at least

rose moth
#

Okay, maybe we could have both

errant barn
#

We appreciate the ideas, but the whole ethos of Black is to limit the number of style choices a user has to make. We're quite opposed to adding configuration values like that. It's nothing personal, and I also sometimes feel like some other style would be better in certain situations. But overall it isn't worth the inconsistency. But it's nice to see you're interested in improving the style! If you think something should definitely be formatted differently in most cases, feel free to open a discussion on the issue tracker (after checking similar ones of course) πŸ‘πŸΏ I hope I'm not too dismissive also πŸ˜…

rose moth
#

πŸ‘

#

Thanks for the info

flat krakenBOT
#
felix-hilden
[psf/black] Pull request opened: #2538 Update bug template

I've modified the bug issue template slightly to:

  • Make it more explicit that we want the user to check if the bug is already fixed, rather than finding out that it isn't reproduced on main and submitting the issue anyway
  • Provide an example formatting of the reproducing example's components, and not hide it in the HTML comment

Checklist

  • [x] Add a CHANGELOG entry if necessary?
  • [x] Add / update tests if necessary?
  • [x] Add new / update outdated documentation?
errant barn
tired shard
#

why not yaml templates? They are more easier to work with

errant barn
#

Oh, I didn't know they allowed yaml too, I'm looking at the guide just now!

#

Or did you mean the issue forms?

tired shard
#

yeah

lament crow
#

You should use forms, yeah

errant barn
#

Hmm, maybe that'd be worth it

#

especially for a bug report

lament crow
#

Isn't there PR open for it by any chance though?

#

Or at least issue

#

I recall seeing one

errant barn
#

Can't find one at a quick glance

lament crow
#

Guess I misremembered it then

tired shard
errant barn
tired shard
#

Yeah, I would push this to a separate issue later, others would probably agree with you there, verbose enhancement changes soon

tiny rock
#

Hi! I'd like to work on issue https://github.com/psf/black/issues/2370, one thing is I'm not sure where to start. Any advice on where I should look to catch blank lines after docstrings and where to implement its deletion?

errant barn
#

But lemme give my 2 cents there now

lament crow
#

biiiig can of worms lol

#

especially with people already having issues with the little changes done by Black since it's AST unsafe

errant barn
#

yeah πŸ˜…

lament crow
#

but well, technically it's a bit of off-topic for that issue since it was just about whitespace outside the docstring lol

#

so no AST safety issues involved there at least

#

definitely in favor of no blank lines after myself πŸ˜„

#

when it comes to the multi-line doc vs single-line doc, I also agree with how you formatted it yourself: ```py
def f():
pass

def f():
"""Doc."""
return stuff

def f():
"""
Multiline doc.

With some explanation.
"""
return stuff

Rather than the:
```py
def f():
    pass

def f():
    """Doc."""
    return stuff

def f():
    """Multiline doc.

    With some explanation.
    """
    return stuff
#

but well, that's entirely separate thing

#

and there's probably some open issue about it already lol

errant barn
#

yeah, I checked numpy's style guide and there it was actually as in your second example

#

which I was quite surprised by

lament crow
#

eww

#

I know some do that so maybe not surprised but still eww

errant barn
#

yep

lament crow
#

I sometimes have to do ```py
def f():
"""
Very long doc that would not pass length limit in a single line if I did not add the new line.
"""
return stuff

rather than this because it didn't pass line length in this specific case:

def f():
"""Very long doc that would not pass length limit in a single line if I did not add the new line."""
return stuff

But it always seems off
#

definitely not my preference for things that can fit in single line

#

technically could just do a split just so that it doesn't look weird:

def f():
    """
    Very long doc that would not pass length limit in a single line
    if I did not add the new line.
    """
    return stuff

but if it's unnecessary to pass line length, it feels like I just am doing it so that I can look at it lol

errant barn
#

I guess it could be the time to shorten the initial explanation and just elaborate in a full paragraph

lament crow
#

Sometimes it might be hard to shorten the initial explanation

#

or at least that's my feeling

dense jungle
#

We have a lint rule enforcing that you have a one-line summary for your docstring

flat krakenBOT
#
myw
[psf/black] Issue opened: #2539 Black deletes nonempty block comments before decorators left unfromatted with \# fmt: skip

Bug description:

Black deletes nonempty block comments before decorators left unfromatted with #fmt: skip.

As a legitimate real-world example (unlike the minimal test case I present below), I found this when black unexpectedly removed an # endregion comment in between two unit tests, the second of which was decorated with @parameterize. That decorator was skipped with fmt: skip to allow me to group/align rows of arguments.

For me right now, this is mostly an annoyance, ...

bright glacier
#

wow the fmt: skip implementation seems buggy then πŸ™ƒ

flat krakenBOT
elder tusk
rose moth
#

Thanks

sour jay
sour jay
viral path
#

!d del

late dewBOT
#
del

7.5. The del statement


del_stmt ::=  "del" target_list
``` Deletion is recursively defined very similar to the way assignment is defined. Rather than spelling it out in full details, here are some hints.

Deletion of a target list recursively deletes each target, from left to right.

Deletion of a name removes the binding of that name from the local or global namespace, depending on whether the name occurs in a [`global`](https://docs.python.org/3/reference/simple_stmts.html#global) statement in the same code block. If the name is unbound, a [`NameError`](https://docs.python.org/3/library/exceptions.html#NameError "NameError") exception will be raised.
viral path
#

!e
lst=[1,3,4]
del lst[1:1]
print(lst)

late dewBOT
#

@viral path :white_check_mark: Your eval job has completed with return code 0.

[1, 3, 4]
flat krakenBOT
flat krakenBOT
tribal thistle
#

Hmm

#

Trying to repro 2540, but so far it doesn't seem to be crashing

#

It's insisting it's python 2 though, don't know if that's related

#

Oh wait I didn't have a clean branch πŸ€¦β€β™‚οΈ πŸ€¦β€β™‚οΈ 🀑

#

ignore all I said

#

(Still not crashing on a clean branch)

tribal thistle
#

Well, managed to finish off that bug, but found a more interesting one

#

jpadilla/black-playground#49

toxic stormBOT
flat krakenBOT
flat krakenBOT
#
AdamYoblick
[psf/black] Issue opened: #2542 Error in Visual Studio when formatting with no newline at end of file

Describe the bug

When trying to format a document in Visual Studio using the black formatter, an error is thrown when there is no newline at the end of the file. Please see https://github.com/microsoft/PTVS/issues/6752 for details.

To Reproduce

  1. Create a python application in Visual Studio
  2. Paste the following code (no newline at end!)

# this is a comment
import os
import sys
spam(ham[1], {eggs: 2})
def foo(): pass


x = 1
y = 2
y = 2
# comment ...
austere lava
#

TIL this is invalid syntax since the new PEG parser:

((*foo),)
bright glacier
#

huh.

bright glacier
#

As I rewrite diff-shades I'm replacing my custom DIY output format and tooling for rich. I regret nothing because it looks way better already πŸ˜„

unknown.png
tribal thistle
#

That's really pretty, keep it up

bright glacier
#

My original thinking was something along the lines of "oh diff-shades is going to be too slow to require fancy output" but I forgot that the other maintainers have 6-core CPUs with each core being much faster too (or something similarly fast) πŸ˜…

tribal thistle
#

lol

bright glacier
#

I'm just sitting here with a dual-core AMD APU from 2015

#

the github actions runner is actually faster than my machine AFAICT

lament crow
#

diff-shades is going to be too slow to require fancy output
I would think that this is a good reason to have the fancy output

#

since that way you can see progress

bright glacier
#

I mean, my original output already had a custom percentage indicator

bright glacier
#

/me needs to find a way to get higher quality screenshots and screen recordings ...

silent apex
bright glacier
silent apex
#

Is that open source?

bright glacier
#

diff-shades? Yes, it's a tooling project for black so we can more easily compare formatting differences between two versions of black!

#

hence diff (reference to the diff format used for comparisons) and shades (as in shades of black)

bright glacier
silent apex
#

I'd love to see it, although I'm in no rush haha

bright glacier
# silent apex I'd love to see it, although I'm in no rush haha

here ya go: https://github.com/ichard26/diff-shades

once you cloned and installed it (I recommend in a venv) you can try this example command: diff-shades analyze example.json -s "black" -s "blackbench" -s "bandersnatch" (it shouldn't leave any artifacts, although if it does lemme know!) (actually as I reimplement more functionality, example.json will be saved FYI)

GitHub
GitHub - ichard26/diff-shades: The Black shade analyser and compari...

The Black shade analyser and comparison tool. Contribute to ichard26/diff-shades development by creating an account on GitHub.

silent apex
#

tyty

bright glacier
silent apex
#

woah

#

All of the output is so beautiful

#

great job!

bright glacier
bright glacier
#

Nice, mypy caught a subtle programming error! I accidentally append-ed a dict to a list even though I should've been appending a tuple that contains such dictionary.

trim orchid
#

black-formater xDD

bright glacier
#

I found this and it seems like a great way to cut back on the "cancel workflow" when repeatedly force-pushing

tribal thistle
#

Hm, it probably won't work for what you described since the ref will be changing, but modify that part, and yeah it'll work

#

We actually have it on PRs for pydis projects

bright glacier
#

I'm not actually going to use it for the time being as I have write permissions to psf/black and I don't mind the clicking (after all I have to deal with failure notifications)

bright glacier
tribal thistle
#

failure notifs that is lol

bright glacier
#

Yup, only on my commits though, it'd be painful for all of them πŸ‘€

tribal thistle
#

Well there go my plans for sliding into your emails

bright glacier
#

I'm sure my email is somewhere out there in the interwebs

tribal thistle
#

I think you signed it over when you made your blood sacrifice contract to join staff

#

But where's the fun in that

bright glacier
#

But yeah lol, just imagine getting a GHA failure email in which the log output is "hey $name, wanna grab lunch later?"

tribal thistle
#

lmfao

bright glacier
tribal thistle
#

Anyways I'll stop bothering you before I get a warning for derailing this channel πŸ₯²

bright glacier
#

Yeah OK, I'm sure you could find it on notion

#

f.warn 386599446571384843 lemon_sentimental

#

(that's my python bot's prefix)

tribal thistle
#

I am on the test server 🀑

amber karma
#

Does black follow sub-poetry project's configs rather than taking the top level config?

#

atm I essentially have one config inside a folder and black seems to be defaulting to using the nested config not the top level config from where it's ran

#

the other offender is flake8 but its harder to tell whether it's the culpret or not

errant barn
#

Not currently, but there are issues on the tracker discussing something like that: psf/black#1826

toxic stormBOT
amber karma
#

ah, thanks shall have a read.

flat krakenBOT
#
bukzor
[psf/black] Issue opened: #2543 What's up with the extra newline after from/import?

Describe the style change

See also: #90

Black currently forces blank lines after all imports, including imports outside the "import section" at module scope. I find that this yields super inappropriate results. Black does not have enough context (outside the module-scope "import block") to know whether adding a blank line will correctly denote intent or enhance readability. Black should refuse to guess; as in #90,

Example input

I'm having trouble generating a good exa...

bright glacier
#

oh hi Marco, I see you're typing something πŸ‘€

stuck vapor
bright glacier
#

Hmm, my immediate reaction is that if they're going to patch sys.std* they should support as much of the io.TextIOWrapper API as possible, but I assume this issue has existed for a long time and probably won't be fixed?

#

totally off topic, on the IPython repository I got this unfortunate message:

A user you’ve blocked has previously contributed to this repository.
πŸ™ƒ

stuck vapor
bright glacier
#

The person I blocked was just annoying with their relentless IMO spammy issues but nothing too serious other than wasted time and effort

#

Is there an issue about OutStream not having the buffer attribute somewhere? I tried looking (hence why I learned the user I blocked contributed to IPython) but I couldn't find one 😦

stuck vapor
flat krakenBOT
#
fofoni
[psf/black] Issue opened: #2544 Consecutive string literals inside dict literals

The output black gives to me when the value of a dict item is made up of two consecutive string literals (on different lines because of the line length limit) is ugly enough to look like a bug to me:

--- STDIN	2021-10-18 20:25:48.108927 +0000
+++ STDOUT	2021-10-18 20:25:52.127191 +0000
@@ -1,5 +1,4 @@
 d = {
-    "key":
-        "this entry is a long string."
-        "lalala this is the next part of the long string lalala"
+    "key": "this entry is a long string."
+    "...
stuck vapor
flat krakenBOT
bright glacier
solar totem
#

Can I use black on an entire package?

silent apex
#

Absolutely! I run it on projects with at least a dozen files often.

solar totem
#

How do you do that?

#

I'm trying to have black auto format everything on github actions

silent apex
#

I typically run it locally, and then add black in CI just to make sure it was ran locally (so without modifying the code in CI).

#

Off the top of my head, I believe the args I'm thinking of are --diff --check

solar totem
#

What's CI?

#

I'd like to run it via GH Actions so anyone who makes a PR has it auto-formatted as well

silent apex
#

CI is continuous integration, think GH Actions

silent apex
solar totem
#

Ah

bright glacier
#

@solar totem ^^ (I forgot to mark my message as a reply)

solar totem
#

Huh

#

This whole automation with github actions is wildly confusing

#

I want to release a new PyPi release when I make a new GH release, I want to run this testing file, and all sorts of other things

bright glacier
#

You can totally do that, it's just annoying to setup a workflow that'll automatically push a commit

bright glacier
solar totem
#

Yeah, it's all just very confusing

bright glacier
# solar totem Yeah, it's all just very confusing

Fundamentally it's basically just running a pre-defined series of steps on some server (usually from the cloud). Actions are just pre-built steps in a way, they can be as simple as checking out a repository to as complex as handling a full build of a suite of wheels.

#

This is a little off-topic for this channel though, I'd suggest #tools-and-devops for more discussion or help.

#

although something that is on-topic is this example github actions workflow we have in the docs:

name: Lint

on: [push, pull_request]

jobs:
  lint:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v2
      - uses: psf/black@stable
        with:
          version: "21.5b2"
#

(and Jelle before you @ me for using an old version, it's the first one that came to mind as a valid one πŸ™‚ )

errant barn
heavy heron
#

You can make GitHub Action make a commit for you

#

There are even some pre-built actions for it:

      - name: Commit changes
        uses: stefanzweifel/git-auto-commit-action@v4
        with:
          file_pattern: pyproject.toml
          commit_message: Bump version to ${{ github.event.release.tag_name }}```
short forge
#

You should be using both precommit with the black hook and GH actions ideally

solar totem
#

haha

flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
#
scottwn
[psf/black] Issue opened: #2547 Plugin can't find black python module

Describe the bug
I have installed Black using vim-plug. Installation is successful and I can update the plugin no problem. However when I run black inside Vim I get the following error

Error detected while processing /Users/scott/.vim/plugged/black/autoload/black.vim:
line  195:
Traceback (most recent call last):
  File "", line 97, in 
ModuleNotFoundError: No module named 'black'
Error detected while processing function black#Black:
line   10:
Traceback (most recent call ...
plain atlas
#

wait when did this channel get moved

#

I THOUGHT I WAS GOING CRAZY

mint barn
#

hm what's this

plain atlas
#

black formatter

flat krakenBOT
#
DanInSpace104
[psf/black] Issue opened: #2548 Black doesn't wrap long line but wraps shorter line

Describe the bug

Black fixes line with length of 116 but doesn't fix line with length of 121. When i tried to fix it myself Black reverted my changes.

To Reproduce

  1. Take this code:
def temp():
    datetime_start_result = f'{datetime_start[6:10]}-{datetime_start[3:5]}-{datetime_start[0:2]} {datetime_start[11:]}'
    datetime_finish_result = f'{datetime_finish[6:10]}-{datetime_finish[3:5]}-{datetime_finish[0:2]} {datetime_finish[11:]}'
  1. Run Black on it with...
flat krakenBOT
#
nipunn1313
[psf/black] Pull request opened: #2549 Run second pass of formatter

Description

Run the formatter twice to ensure stability.
Fixes #2488 and #2518

Admittedly, it doesn't fix the underlying issue. There are three
expected-failure tests in test_black.py, which prove the issue
(test_trailing_comma_optional_parens_stability{1,2,3})

This is a (somewhat terrible) bandaid which runs the entire formatter
twice to work around this problem. This bandaid actually already existed
in the stability-checking-validation code, but resulted in a non-stable
...

flat krakenBOT
plain atlas
#

TBF I actually think a cli option to run black twice and alert of any changes would be nice loll

errant barn
#

But it shouldn't be like that πŸ˜…

#

The only reason you'd want that is the stability issues, so it'd be better to fix them. Granted, they aren't fixed

tired shard
#

Hi πŸ‘‹ Would love to get some more reviews on psf/black#2526 psf/black#2525

toxic stormBOT
barren iris
errant barn
#

We try to not touch docstrings too much. Although personally I'd be in favor of normalising their style as well πŸ˜„ But doesn't pydocstyle itself support many styles?

flat krakenBOT
tired shard
#

https://github.com/psf/black/pull/2525#discussion_r732652768 @errant barn I am not sure if you understood it right there, what I meant was when black would come across config it would overwrite all the config on the main pyproject.toml with that found in the config specified in config.

So that test would do the same thing, where it just keeps a few vars in the main config and all of the original ones in test_black.toml(current name) and then would see if the values are from test_black.toml and not from the main one

errant barn
#

Sorry, I find it hard to read this message as well, let's go over it :D

#

So the previous tests parse_pyproject and read_pyproject asserted that 1) the values are read correctly and 2) they are carried over to the top level function being click friendly

#

This change introduces the ability to specify config to link to another file. So in my opinion it should be a separate test.

#

Maybe it should be the other way around, that the new file uses this test.toml as the link target.

tired shard
#

Ok, so say the pyproject.toml contains:

[tool.black]
a = 10
b = 20
c = 30
config = "black.toml"

now we specified that the config for black would be present in black.toml and like discussed on github we would just overwrite the config with the one found in the linked file i.e. black.toml

[tool.black]
a = 20
b = 30

So, this is the content of black.toml, that test is testing that the values a which is 10 in the main config is overwritten to be 20 which is found in the linked config.

--

Yeah, the original test goals are conserved here, it just edits adds an additional before giving it to the top level function and reading the main config

errant barn
#

Maybe it's just preference, but I don't like the idea of deleting simple tests just because more complex tests cover those cases too

#

I get what it does though

errant barn
#

That way we wouldn't have to make two new files, just one with the link

tired shard
#

its not deleting them (if deleting means the thing below, i am adding them back, i removed them earlier as i added them as two separate tests but now kept them as one as the original functionality has changed)

Screenshot_from_2021-10-20_16-48-16.png
tired shard
errant barn
#

The two tests are different right: parse and read. You removed some attributes from the parsing test, which now covers less values. Or am I tripping? :D

tired shard
#

ah i get it, the attributes which are deleted aren't meant to be deleted and I am pushing a change to fix that.

They aren't there as they were required for a previous functionality which was removed and then i forgot to apply the change here,

errant barn
#

Yes, please leave this test as it is and create a new one with the linking

#

I think just a single test is ok though, which "reads", not "parses"

#

but I like the idea of checking that they are overwritten πŸ‘πŸΏ

tired shard
#

πŸ‘ yeah will mention that in a bit

#

btw is black participating in hactoberfest, as in can contributors get hacktoberfest-accepted label on their pr

errant barn
#

uhm.. I remember seeing some discussion

#

yeah we do have that label

bright glacier
#

I can't add the hacktoberfest topic FYI only ambv can IIRC

#

so we'll have to use the accepted level

errant barn
#

I think that's fine :D

tired shard
#

we have to use the accepted label anyways if the repo has the label or not IIRC (i haven't looked into this year's rules)

errant barn
#

it's not like our review cycle is faster than issues and contributions πŸ˜…

#

But what's up @bright glacier, howya doing?

heavy heron
fast warren
#

Quick question about formatting :
If you have a really big integer (for example, 0xAA87CA22BE8B05378EB1C71EF320AD746E1D3B628BA79B9859F741E082542A385502F25DBF55296C3A545E3872760AB7), how should you write it ?
Should you just write it and ignore the 80 characters line limit ? Or is there a special way of writing integers to respect this limit that I'm not aware of ?

tribal thistle
#

Not completely sure for numbers, but it’s certainly okay to go over the limit in black’s style if you have to

red sedge
#

With a bytes literal

fast warren
red sedge
#

But you can split it

fast warren
#

Well, yes. But it'd be a bit confusing I believe. I think it'd be more readable to just have one big integer rather than a bytes string

#

*But it's a good idea, thank you

lament crow
#

If no topic, only label works

flat krakenBOT
#
tabrezm
[psf/black] Issue opened: #2551 Error trying to format line with parameter assignment and generator

Describe the bug

Black throws an error trying to format files containing a parameter assignment with a generator. This is valid Python code.

error: cannot format util.py: Cannot parse: 200:43:     if any(match := pattern_error.match(s) for s in buffer):
Oh no! πŸ’₯ πŸ’” πŸ’₯
1 file failed to reformat.

To Reproduce

Code sample:

    if any(match := pattern_error.match(s) for s in buffer):
        if match.group(2) == data_not_available:
            # Error O...
tired shard
errant barn
#

we'll do it on merge!

tired shard
#

cool thanks

mint barn
#

Black does not work on 3.10 for mac?

dense jungle
mint barn
#

Running black on 3.10 throws this error

image0.jpg
dense jungle
mint barn
#

Thx for the help πŸ™‚

bright glacier
#

Oh yeah, since last time we interacted I've become PyDis staff ^^

errant barn
plain atlas
#

hey you're technically on PyDis too!

errant barn
#

oh snap

#

closes server

#
  • *speaks in asterisks
red sedge
#

Saw there was some libcst rust work this week

#

Fingers crossed eh

dense jungle
#

We'd still have to rewrite all of Black to work with libcst 😱

red sedge
#

What about an # black: off that literally snipped the lines until # black: on?

bright glacier
#

sigh we got a long story to tell here :(

Screenshot_20211020-174018_Gmail.jpg
red sedge
#

Then pasted them back after format

dense jungle
bright glacier
#

it wouldn't require black's parser to support the snipped out code

#

currently we still parse the whole file and then turn any marked code into standalone comments (internally) which aren't touched by the formatting logic

red sedge
#

Yep

#

Then you'd do the ast check after

#

And it would be the authors fault if they built something that can't work

#

It would be brittle but probably work in a lot of cases

bright glacier
bright glacier
red sedge
#

The format comment would have to match the indentation

bright glacier
red sedge
#

Wow magic

dense jungle
bright glacier
red sedge
#

Can you imagine JavaScript standardising something that wasn't supported in all their type checkers and formatters

uncut patio
#

Javascript standardises things that don't even have support in any of their implementations every once in a while

red sedge
#

Ah yeah that one thing

#

But they changed the standards process pretty much because of that

red sedge
plain atlas
#

hm

#

lol

#

if i want to do something cursed I probably shouldn't be doing

#

i want to set black to format one folder with one line length

#

and a different folder of a different length

#

in the same project

#

lol

bright glacier
#

there's an open issue exactly for that

tired shard
#

My 3rd PR πŸ™ƒ

flat krakenBOT
#
nipunn1313
[psf/black] Pull request opened: #2552 Print out line diff on test failure

Description

It currently prints both ASTs - this also
adds the line diff, making it much easier to visualize
the changes as well. Not too verbose since it's only a diff.
Especially useful if the test file is long (and the ASTs are big)

Checklist - did you ...

  • [x ] Add a CHANGELOG entry if necessary? (should not be necessary for this)
  • [x ] Add / update tests if necessary?
  • [x ] Add new / update outdated documentation?
errant barn
#

πŸ‘€

tribal thistle
#

Morning y'all

errant barn
#

Morning πŸ˜„

lament crow
cyan current
#

Good Afternoon

red sedge
errant barn
#

now these are the ideas we need

#

black is so opinionated that it becomes a lobbyist

lament crow
#

Where do I sign? πŸ˜„

flat krakenBOT
tired shard
#

πŸŽ‰

flat krakenBOT
flat krakenBOT
#
chasefinch
[psf/black] Issue opened: #2553 Use explicit string concatenation in \(experimental\) string processing

Describe the style change

I suggest to use explicit string concatenation, instead of implicit concatenation, for (currently experimental) string processing, with whitespace pushed to the end of each line.

Examples in the current Black style

(
    "Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor"
    " incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis"
    " nostrud exercitation ullamco laboris nisi ut ali...
lament crow
#

ugh

flat krakenBOT
#
nick863
[psf/black] Issue opened: #2554 The question mark in comment results in extra spaces on the second pass\.

Repro steps:

  1. Create the minimal jupyter notebook, which will contain one cell with the function ands a comment, containing question mark;
def example_fn(param):
    """
    param: str    what to print?
    """
    print(param)
  1. Close the file and run
black-nb  Repro.ipynb

The error will be shown:
error: cannot format /path/to/nb/Repro.ipynb: INTERNAL ERROR: Black produced code that is not equivalent to the source on pass 1. Please report a bug on https://gi...

tribal thistle
#

Using question marks for params? That's a new one

flat krakenBOT
bright glacier
#

yay for spammy reviews :(

tribal thistle
#

Umm is

#

/officially supported

tribal thistle
#

Could not reproduce with just black

bright glacier
#

No one from the team has ever been involved with black-nb

tribal thistle
#

Thanks

#

They probably should intercept that error message and put their bug tracker lol

tribal thistle
#

confirmed that issue was in fact with black-nb

#

Also did a whole bunch of debug for them for free 🀑

dense jungle
#

thanks for digging into this @tribal thistle !

tribal thistle
#

I was just messing around, it's actually fun

flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
heavy heron
flat krakenBOT
flat krakenBOT
#
olliemath
[psf/black] Pull request opened: #2559 Allow install under pypy

Description

This adds a minimal set of changes to use black under pypy. It's supposed to be a more acceptable version of ...
with the key point being there are zero changes for cpython users.

This is a draft - feedback appreciated - I plan to do all of the following (if needed):

Checklist - did you ...

  • [ ] Add a CHANGELOG entry if necessary?
  • [ ] Add / update tests if necessary?
  • [ ] Add new / update outdated documentation?
heavy heron
hoary plume
#

github is an issue indeed

flat krakenBOT
#
joshua-cannon-techlabs
[psf/black] Issue opened: #2560 Suggestion: Mention contextlib\.ExitStack in "Use backslashes for with statements"?

Is this related to a problem? Please describe.

On Black's future style doc it talks about black handling a with with multiple context managers in a few ways (Python-version-dependent).

Describe the solution you'd like

Black should suggest to the user (in documentation only) that if they don't like Black's current or future behavior, that constructing a `c...

flat krakenBOT
#
gsnedders
[psf/black] Issue opened: #2561 if\-expression line breaks in function arguments

Describe the style change

Currently, if an if-expression within a function call needs to be broken across lines to avoid being an overly long line, it gets put horizontally in line with arguments which makes it hard to see where each argument in the call begins.

Examples in the current Black style

            yield Test(
                test_path=trimmed_path,
                expected_text_path=expected_text_paths[-1]
                if expected_text_paths...
flat krakenBOT
#
hugovk
[psf/black] Pull request opened: #2562 Declare support for Python 3\.10

Description

Python 3.10 final is already tested, just needs the Trove classifier.

Checklist - did you ...

  • [ ] Add a CHANGELOG entry if necessary?
  • [n/a] Add / update tests if necessary?
  • [n/a] Add new / update outdated documentation?

Is a changelog entry needed?

By the way, Black was first released in March 2018 and took the forward-thinking view to only support Python 3.6, which was the newest release, and therefore Black hasn't needed to drop ...

flat krakenBOT
#
cooperlees
[psf/black] Pull request review submitted: #2555 Add \-\-projects cli flag to black\-primer

Thanks for the idea /addition. I feel this is already achievable tho. Via config files.

Whats the use case and advantage of this over just making a dedicated subset config file(s)?

  • I guess so you can have a large config file and only run some projects

But I feel if you're doing that a lot just having multiple config files would also work. But I get de-duplication is nice.

Interested to hear so I can understand more. Thanks!

bright glacier
#

supporting a program across three OSes and five Python versions is quite a bit. Even my tiny diff-shades project's CI looks complex in a way haha :)

#

One thing I don't like about using pre-commit with mypy is that unfortunately the mypyc compiled wheel isn't used. Slows down linting more than necessary but I doubt pre-commit would accept a Python specific feature to pull from PyPI.

flat krakenBOT
#
andgineer
[psf/black] Issue opened: #2563 different code on the second pass of the formatter\.

produce error

error: cannot format /Users/SOROKAN6/projects/idse-pqsar-api/tests/test_ellipsize.py: INTERNAL ERROR: Black produced different code on the second pass of the formatter.

run black for the file and it will show the error.
but there is impossible to attach py-file

Environment (please complete the following information):

black, version 21.8b0
Python 3.7.6
MacOS 11.6

Does this bug also happen on main?
No

Additional context

[blk_9gtrwz8w.log](htt...

late dewBOT
#

Hey @golden mica!

Uh-oh! It looks like your message got zapped by our spam filter. We currently don't allow .txt attachments, so here are some tips to help you travel safely:

β€’ If you attempted to send a message longer than 2000 characters, try shortening your message to fit within the character limit or use a pasting service (see below)

β€’ If you tried to show someone your code, you can use codeblocks
(run !code-blocks in #bot-commands for more information) or use a pasting service like:

https://paste.pythondiscord.com

tribal thistle
#

Oooh I just discovered difflib can reverse a diff

#

Pretty handy when all you've got in an issue is the diff lol

bright glacier
tribal thistle
# 
something
- a
+ b
something

difflib can take ^ and convert it to just:

something
a
something

or just

something
b
something
#

basically apply the reverse of a diff and give you back the two inputs

bright glacier
#

oh wow

tribal thistle
#

!d difflib.restore

late dewBOT
#
difflib.restore 

difflib.restore(sequence, which)```
Return one of the two sequences that generated a delta.

Given a *sequence* produced by [`Differ.compare()`](https://docs.python.org/3/library/difflib.html#difflib.Differ.compare "difflib.Differ.compare") or [`ndiff()`](https://docs.python.org/3/library/difflib.html#difflib.ndiff "difflib.ndiff"), extract lines originating from file 1 or 2 (parameter *which*), stripping off line prefixes.

Example...
bright glacier
#

I do wish difflib could apply a diff though

muted haven
bright glacier
#

?

muted haven
#

It was pretty much a hacky solution, so not something I'd say useful

#

but can be revived for simpler use cases

bright glacier
#

Eh, Łukasz already said he'll be looking at your hacky solution πŸ˜„

muted haven
#

πŸ˜…

bright glacier
#

I wonder what would happen if we threw a fuzzer at this "solution"

muted haven
#

I'm pretty sure it can be implemented though did not try to work on it

#

Regarding fuzzing, I've used this script;

import ast
import sys
from pathlib import Path

with open(sys.argv[1]) as stream:
    tree = ast.parse(stream.read())
    
functions = {
    func.name: func
    for func in ast.walk(tree)
    if isinstance(func, ast.FunctionDef)
}

test_dir = Path("/tmp/patma_black_tests/")
test_dir.mkdir(exist_ok=True)

for name, node in functions.items():
    with open(test_dir / f"black_{name}.py", "w") as stream:
        stream.write(ast.unparse(node))

to extract all pattern matching cases from https://github.com/python/cpython/blob/main/Lib/test/test_patma.py

GitHub
cpython/test_patma.py at main Β· python/cpython

The Python programming language. Contribute to python/cpython development by creating an account on GitHub.

#

and then run this to estimate how much percentage of those my hack support;

from pathlib import Path
from collections import Counter
from black.parsing import lib2to3_parse

test_dir = Path("/tmp/patma_black_tests/")

counter = Counter()

for file in test_dir.glob("black_*.py"):
    try:
        lib2to3_parse(file.read_text())
    except Exception:
        stat = "fail"
    else:
        stat = "success"
    
    counter[stat] += 1
    print(f"{file.stem}: {stat}")

print(counter)
#

(just throwing, in case it would be useful for future effort)

bright glacier
#

Yup yup. This is way out of my wheelhouse haha, so I'll cc @solid adder who actually knows something about parsers ^^

#

Thanks for the work by the way, I'm rather impressed and this might end up being useful!

#

The only things I can sorta understand about parsing is tokenization and whatever grammar blib2to3 uses. I once tried writing a recursive-descent parser .. it didn't go well πŸ˜…

muted haven
bright glacier
#

Yeah, I wanted to learn how parsers worked by writing one by hand, turns out it was harder than I expected πŸ˜†

#

CPython's grammar is pretty complex now, ay? definitely wouldn't want to maintain its parser by hand

muted haven
bright glacier
#

Poll time! please vote on which one looks better (hint: it's all about the bold styling)

unknown.png
#

!poll "First or second one?" "first" "second"

late dewBOT
#
First or second one?

πŸ‡¦ - first
πŸ‡§ - second

bright glacier
#

Other suggestions welcomed too 🌺

tribal thistle
#

mmm if only difflib could do image collages

bright glacier
#

haha yeah, I'll figure something out when I implement a show commabd

#

haha yeah, I'll figure something out when I implement a show command

lament crow
#

the only difference is the bold font on the text between time and the square brackets/parentheses?

bright glacier
#

yup

heavy heron
#

..... looking for the differences

bright glacier
heavy heron
#

ooooh, I just saw that, wow

#

since this is so colorful, I'd go with bold

#

also, I am totally using this as an inspiration for my build system :P

heavy heron
bright glacier
heavy heron
#

haha, nice name

#

will keep that in mind

bright glacier
heavy heron
#

Rich is gud

#

I still don't know what I am using to make my build system

#

I could try some bash for the sake of it, but I already regret it

bright glacier
#

Bash scripting is something I'll continue to ignore πŸ˜„

heavy heron
#

haha

#

I just love that one, despite being so basic

bright glacier
#

the only "shell scripts" I have are my various aliases

# 
alias lexa "exa -lga"
alias actenv '. env/bin/activate.fish || begin; . venv/bin/activate.fish && echo "... but venv did work!"; return; end;'
alias sl "sl -elF"
alias blackenv ". /home/ichard26/.local/share/virtualenvs/black-Q9x7i-w-/bin/activate.fish"
alias wb "python -m webbrowser"
alias addpythons 'set -x PATH /opt/python3.7.8/bin/:/opt/python3.9.5/bin/:$PATH'
heavy heron
#

straight up aliases, not functions?

#

yeah, okay

#

that's cool

#

Fish user, I dig this

bright glacier
#

actenv makes me happy as I don't have to do . /venv/bin/activate.fish constantly 🌺

heavy heron
#

well, I don't have a very high opinion of fish, but it is the less worse one

#

I don't have aliases, just autocomplete haha

bright glacier
#

understandable

bright glacier
# 
diff_shades/cli.py:208: error: Incompatible types in assignment (expression has type "str", variable has type
"Union[Literal['nothing-changed'], Literal['reformatted'], Literal['failed']]")  [assignment]
        for type in ("nothing-changed", "reformatted", "failed"):
``` well that's annoying but understandable
bright glacier
#

hmmm right windows unicode support is still a pain

tired shard
bright glacier
#

nope, check the font

tired shard
#

Ah ok

slim chasm
bright glacier
#

not relevant to the channel ^^

hoary plume
#

yuh that's reason why i use fish

bold swift
#

@upper pollen Hiiii!!!

flat krakenBOT
flat krakenBOT
#
fabioz
[psf/black] Issue opened: #2564 Black slow startup to format a single file

I'm a bit annoyed that it takes too much time to format a single file when using black to auto-save contents in my IDE (in this case it's Eclipse/PyDev, but I guess it should be the same for any place using the command line to format a single file).

-- In Eclipse/PyDev, the integration (I've done) is done so that for each new file to be formatted, black is executed and contents are piped in/out.

For doing this execution (in my machine), black takes 0.3s, but out of that time, 0.255s...

flat krakenBOT
#
jeremyschulman
[psf/black] Issue opened: #2565 Black reforms Click docstring use "prevent rewrapping" controls

Describe the bug

As a Developer using the Click library, I often use the docstring formatting control character slash-b and slash-f as described here: https://click.palletsprojects.com/en/8.0.x/documentation/#preventing-rewrapping

When I use black to reform my code, black will "right shift" the content in an unexpected and undesired manner.

To Reproduce

Create a Click command using the control formatting:

def cli_design_report_interfaces(devices: Tuple[str]...
slow ibex
#

Black doesnt support python 3.10
It crashes on match-case code
When this support will be added?

bright glacier
# 
--- diff_shades/config.py    2021-10-26 01:33:59.410935 +0000
+++ diff_shades/config.py    2021-10-26 01:39:34.119995 +0000
@@ -42,15 +42,13 @@
         "django",
         "https://github.com/django/django.git",
         custom_arguments=[
             "--skip-string-normalization",
             "--extend-exclude",
-            (
-                "/((docs|scripts)/|django/forms/models.py"
-                "|tests/gis_tests/test_spatialrefsys.py"
-                "|tests/test_runner_apps/tagged/tests_syntax_error.py)"
-            ),
+            "/((docs|scripts)/|django/forms/models.py"
+            "|tests/gis_tests/test_spatialrefsys.py"
+            "|tests/test_runner_apps/tagged/tests_syntax_error.py)",
         ],
         python_requires=">=3.8",
     ),
     Project("diff-shades", "https://github.com/ichard26/diff-shades.git"),
     Project("flake8-bugbear", "https://github.com/PyCQA/flake8-bugbear.git"),

I .. do not like this formatting I got by turning on ESP lemon_angrysad

#

-the brackets were explicitly added to increase clarity and that's gone now

bright glacier
bright glacier
# 
.rw-rw-r--  109M ichard26 ichard26 25 Oct 22:18 full.json

... that is one big analysis data file from diff-shades πŸ‘€

plain atlas
bright glacier
#

yes.

#

i do not care

#

it's only 10000 serialized dataclass objects that make up the file πŸ˜„ /gen

bright glacier
#

I think I'm going to have implement some basic optimizations to help keep the file size a bit more manageable.

plain atlas
#

!d sql

late dewBOT
bright glacier
#

Although it's really compressible so it's the not worse in the world: that file is 17MB as a zipfile :)

bright glacier
plain atlas
#

actually that's the wrong object anyhow

#

!d sqlite3

late dewBOT
#
sqlite3

Source code: Lib/sqlite3/

SQLite is a C library that provides a lightweight disk-based database that doesn’t require a separate server process and allows accessing the database using a nonstandard variant of the SQL query language. Some applications can use SQLite for internal data storage. It’s also possible to prototype an application using SQLite and then port the code to a larger database such as PostgreSQL or Oracle.

The sqlite3 module was written by Gerhard HΓ€ring. It provides a SQL interface compliant with the DB-API 2.0 specification described by PEP 249, and requires SQLite 3.7.15 or newer.

bright glacier
#

I have never used a proper database before tbh

plain atlas
#

same tho

flat krakenBOT
flat krakenBOT
#
jofrev
[psf/black] Issue opened: #2567 Black messes up indentation of else block after for\-statement in fmt on/off block

Black https://github.com/psf/black/commit/62ed53

The MRE below shows the issue pretty clearly, I think, but there's also a [Playground][playground] available.
The docs state:

fmt: on/off have to be on the same level of indentation.

So this kind of single line block should be okay, shouldn't it? #fmt: skip didn't work for me, either.

For what it's worth, I came across this issue in a Docker...

bright glacier
dense jungle
bright glacier
#

Alright time to stretch my triaging muscles and see if I need to open a new issue πŸ˜„

#

... it's been a while :)

flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
bright glacier
#

huh, there's now a github command palette, neat!

flat krakenBOT
#
cooperlees
[psf/black] Issue opened: #2568 Docker upload CI is broken

Haven't debugged yet but seems a build level bug inside the container. Need to see if it can be reproduced locally.

error: failed to solve: process "/bin/sh -c pip install --no-cache-dir --upgrade pip setuptools wheel     && apt update && apt install -y git     && cd /src     && pip install --no-cache-dir .[colorama,d]     && rm -rf /src     && apt remove -y git     && apt autoremove -y     && rm -rf /var/lib/apt/lists/*" did not complete successfully: exit code: 1
Error: buildx fail...
tribal thistle
#

Future note for once the build is fixed: you can get the build running on PRs too, so you don't have to discover this when you merge into main.

#

With some caching, it shouldn't be too bad

tired shard
#

Could I get rereviews on my PR :D

bright glacier
#

I'm going to be busy for this entire week. Sorry, hopefully someone else can pick up your PRs.

#

also oh my goodness your cats Jelle πŸ₯°

flat krakenBOT
#
pmolodo
[psf/black] Issue opened: #2569 error on windows directory junctions

Describe the bug

If a windows directory junction is found while crawling a source root, and it points outside of the source root, then black will error. This happens even if the directory should be ignored!

To Reproduce

  1. Start a windows cmd prompt
  2. mkdir source_root
  3. cd source_root
  4. mkdir .git
  5. mklink /J dir_junction ..
  6. python -m black --exclude dir_junction .

Then you get this error:

Traceback (most recent call last):
  File "C:\pytho...
dense jungle
#

I guess I'll have to learn what a directory junction is

dense jungle
flat krakenBOT
bright glacier
#

And our dependencies are still on fire πŸ™ƒ

flat krakenBOT
tribal thistle
#

Which repo would that be targeting lmao

tired shard
# 
## Can I run black with PyPy?

Yes, there is _experimental_ support for PyPy 3.7 and higher. You cannot format Python 2
files under PyPy, because PyPy's inbuilt ast module does not support this. For the most
stable user-experience, we recommend using CPython if possible.
#

hmmm

bright glacier
#

a little farfetched I admit

tired shard
flat krakenBOT
#
nipunn1313
[psf/black] Pull request opened: #2572 Fix instability due to trailing comma logic

Description

Fixes #2488 and #2518

Consistently apply paren omission rules regardless of existence of magic trailing comma. Previously, the first run would expand + add a magic comma, and the second run would get a different paren omission rule, causing black to be unstable. We had some hax in black/init to work around this which this diff removes.

This changes the way black formats (adds more parens in some cases). Based on the black release stability policy, we'd want this t...

tribal thistle
#

Run black in a while loop till the output does not change

/s
||unless…||

plain atlas
#

Then run a different formatter, just to triple ensure no changes made.

flat krakenBOT
flat krakenBOT
bright glacier
# 
.rw-rw-r-- 109M ichard26 ichard26 28 Oct 16:54 full.json
.rw-rw-r--  12M ichard26 ichard26 28 Oct 16:58 full.tar.bz2
.rw-rw-r--  17M ichard26 ichard26 28 Oct 16:57 full.tar.gz
.rw-rw-r-- 9.8M ichard26 ichard26 28 Oct 16:59 full.tar.xz

eh with compression it's not too bad @plain atlas

#

although to be fair --xz is extremely slow and --bzip2 is a fair bit slower than gzip

plain atlas
#

Smh

bright glacier
#

?

bright glacier
#

Conflicting files

  • mypy.ini
  • src/black/init.py
  • src/black/linegen.py
  • src/blib2to3/README
  • tests/test_black.py
    well that's annoying >.<
plain atlas
#

git rebase main -i

bright glacier
#

oh gosh test_black.py has five conflicts 😦

#

i don't even know how it has this many

plain atlas
#

interesting

#

where is this

bright glacier
#

psf/black#2431

toxic stormBOT
plain atlas
#

(I own no MacOS running machine)
πŸ‘ good person

bright glacier
#

I completely forgot we added a --workers option

#

whyyy do we have so many options nowadays πŸ‘€

plain atlas
#

i want black to have a lil bit of plugin support someday

#

or i might hack it in to a script and set that to my path

tribal thistle
#

Doesn't that go against the general philosophy

plain atlas
#

i want to be able to run isort with black

#

;-;

#

is that too much to ask

bright glacier
#

honestly some days I just want to give up and merge fundamentally bad changes into the project

#

because people will constantly ask for them regardless of what you do

plain atlas
#

lmk when, i will 100% make a fundamentally bad pr

plain atlas
bright glacier
#

.<

bright glacier
plain atlas
#

yes

#

mostly because i can set vscode to format on save but i can't set vscode to sort imports on save and doing it manually is so much work ;-;

#

then again

#

one sec

stuck vapor
#

you can, for example, set vscode to run pre-commit on save. Then, if you have isort and black and everything there in pre-commit, you're laughing πŸŽ‰

bright glacier
#

I think I resolved the conflicts, hopefully a local test suite run passes and then the mypyc compiled one too 🀞

#

they were super sketchy to resolve

#

ugh, the new test_projects primer test is super slow

# 
21.51s call     tests/test_primer.py::PrimerCLITests::test_projects
0.43s call     tests/test_primer.py::PrimerLibTests::test_gen_check_output
0.01s call     tests/test_primer.py::PrimerCLITests::test_help_output

(2 durations < 0.005s hidden.  Use -vv to show these durations.)
``` that's not OK
bright glacier
#

what the hell

# 
      src/black/__init__.py:1311: error: unused "type: ignore" comment
      src/black/__init__.py:1313: error: unused "type: ignore" comment

oh for goodness sake

#

i'm done for today, nothing has gone well so far

tribal thistle
#

Taking a break is a good idea

bright glacier
#

**If anyone wants to be really nice to me by doing a favour then please read the following:
**
Right now the test suite is a bit fragile and some tests break when the CWD isn't the root of the project. And while usually this isn't a problem as tox and most people would run pytest from the root, it's not guaranteed. I'm working on getting Black compiled with mypyc and the test phase of automated wheel build is currently crashing as the CWD isn't set to the project root (by design of pypa/cibuildwheel).

It'd be awesome to not have to fix this myself as it's mostly just uncomplicated maintenance which sometimes I don't really feel like doing. All you should have to do is setup a development environment, cd into tests/ and run pytest and boom you should see quite a few failures πŸ‘€

Thank you!

silent apex
#

I shall attempt to debug this. This is on the main branch correct?

bright glacier
#

Mhmm.

#

If you don't feel like setting up a full development environment simply setting up pre-commit and a virtual environment + python -m pip install -r test_requirements.txt should get you all set up for debugging.

silent apex
#

Not using pipenv?

bright glacier
#

You can also use pipenv but I honestly don't trust it that much haha πŸ˜… ... and would totally understand if someone felt the same way

silent apex
#

Ah fair enough, I don't love it either

bright glacier
#

Note that you'll have to install black with the jupyter extra.

silent apex
#

Is it mostly ModuleNotFoundErrors or did I set up something really wrong?

bright glacier
#

uhhh that doesn't seem right

silent apex
#

All I did was activate a venv and run python -m pip install -r test_requirements.txt

bright glacier
#

You also did a local install (python -m pip install .[jupyter]) right?

silent apex
#

errr, no. woops

bright glacier
#

should've clarified that, whoops my bad

silent apex
#

woohoo, is this right?

unknown.png
bright glacier
#

oh fuck

#

Yeah you'll also need the d extra

silent apex
bright glacier
#

well some the CWD sensitive tests need that extra

silent apex
#

I now realize there are some contrib docs, maybe I should read them.

bright glacier
#

nope, nope, apparently I got everything misremembered.

#

python -m pip install .[jupyter,python2,d] is actually what should work (did not remember we had so many dependencies required for testing)

#

Sorry for the trouble dawn. I was expecting this to be easy and simple but turns out no :(

bright glacier
silent apex
#

tyty

bright glacier
#

if this doesn't work, I- don't even know

#

failures due to mismatched CWD should look similar to this

unknown.png
#

I assume this'll involve a lot of the THIS_DIR = Path(__file__).parent and ROOT = THIS_DIR.parent pattern

silent apex
#

erm, did I screw something up?

unknown.png
bright glacier
#

I guess zsh considers [] shell syntax?

#

you'll probably need to escape it

silent apex
#

ah yeah, I encased it in quotes and it worked

#

That's probably why my installations were failing

#

Thanks!

plain atlas
#

wait why is zsh doing a bunch of random things holup

#

But also

silent apex
#

So uh, for a couple of these tests, it seems like these errors are coming from Path().cwd(), so is it just a matter of trying if it works and then going up a directory and trying again if it fails?

plain atlas
#

Rather than fixing the problem, what about sidestepping it

#

And determining the project root directory, then changing the cwd to that dir

bright glacier
silent apex
#

That would be nice, actually

plain atlas
#

You do realise that's not possible, is it?

silent apex
#

So just a BASE_PATH and calling everything based on that?

plain atlas
#

The root would only be true on one machine

bright glacier
silent apex
#

Got it, makes sense

plain atlas
#

Ah yeah

#

That

silent apex
#

And then every path would be constructed from one of those two?

bright glacier
#

yup, pretty sure we do that in other parts of the test suite

#

lemme find an example

plain atlas
#

Yknow

silent apex
#

Ooh, yeah, you have it in tests.util, I shall look at that

plain atlas
#

import tests ; tests.__file__.parent

#

Or like that

bright glacier
#

that seems like a bad idea

plain atlas
#

Smh

bright glacier
late dewBOT
#

tests/test_black.py lines 186 to 196

def test_piping_diff_with_color(self) -> None:
    source, _ = read_data("expression.py")
    config = THIS_DIR / "data" / "empty_pyproject.toml"
    args = [
        "-",
        "--fast",
        f"--line-length={black.DEFAULT_LINE_LENGTH}",
        "--diff",
        "--color",
        f"--config={config}",
    ]```
bright glacier
#

there's an example, although the base is the tests directory instead of the project root

silent apex
#

There's a PROJECT_ROOT in test.util, I'll just import that when necessary

plain atlas
#

Hmmm

#

I need to write more tests for my project

bright glacier
#

And...?

silent apex
#

What's the different between these two tests if I'm using the project root as a base? Can I just change the pyproject_path and get rid of the second?

bright glacier
#

huh

silent apex
#

I just don't really understand the difference between those, besides the .parent in the second.

bright glacier
#

It seems like the former is all about finding the pyproject.toml in the CWD while the latter is focused on the finding one in the parent (hence the forced switch to tests via that context manager)

#

I need a moment to read the code tho

silent apex
#

ohh, wait, I see the difference

#

I can't believe I didn't notice that before

bright glacier
#

I have no idea why pyproject_path = Path(Path().cwd(), "pyproject.toml").resolve() is returning /home/ichard26/programming/oss/black/pyproject.toml even when the CWD is tests/ though

tired shard
#

what if we get the cwd and the tests path, and merge them to form the relative path?

#

IIRC most of the tests are failing due to that tests/ path right?

bright glacier
#

yeah well

#

because they're relying on the relative path root to be ./ aka project_root

silent apex
tired shard
#

Yeah, so we get the cwd and tests path, then we can merge them, i.e. if the path is a/b/c and you are in a it would get the difference b/c/ (merge is not the correct word 🀦 )

bright glacier
#

Honestly I don't even know what these tests are testing

tired shard
#

excessing data

silent apex
bright glacier
#

yes he did write the tests lol

silent apex
#

They're well documented tests, really.

bright glacier
#

anyway, the reason why test_code_option_config is failing is because there isn't a pyproject.toml in the tests directory

tired shard
# 
>>> from pathlib import Path
>>> Path('tests/data').relative_to('tests')
PosixPath('data')
#

something like this

silent apex
silent apex
bright glacier
#

so when CWD is set to tests/ the expected configuration path contains tests but black's automagical searching of parent directories meant it found the project's root configuration

#

(instead of just blowing up)

silent apex
#

ah, I see

tired shard
bright glacier
#

another way to fix the test would be use the change_directory context manager and force the CWD to root so the spirit of the test case is preserved but that's probably not super duper important

silent apex
#

I can do that, I don't mind

bright glacier
#

example:

______________________________________________________________ test_ipynb_and_pyi_flags ______________________________________________________________

    def test_ipynb_and_pyi_flags() -> None:
        nb = os.path.join("tests", "data", "notebook_trailing_newline.ipynb")
        result = runner.invoke(
            main,
            [
                nb,
                "--pyi",
                "--ipynb",
                "--diff",
            ],
        )
        assert isinstance(result.exception, SystemExit)
        expected = "Cannot pass both `pyi` and `ipynb` flags!\n"
>       assert result.output == expected
E       AssertionError: assert 'Usage: main ... not exist.\n' == 'Cannot pass ...ynb` flags!\n'
E         - Cannot pass both `pyi` and `ipynb` flags!
E         + Usage: main [OPTIONS] SRC ...
E         + Try 'main -h' for help.
E         + 
E         + Error: Invalid value for 'SRC ...': Path 'tests/data/notebook_trailing_newline.ipynb' does not exist.

test_ipynb.py:455: AssertionError
#

an easy fix would be to replace

nb = os.path.join("tests", "data", "notebook_trailing_newline.ipynb")
# w/
nb = DATA_DIR / "notebook_trailing_newline.ipynb"
tired shard
#

ah I didn't see that DATA_DIR existed in utils.py

silent apex
#

So once I get this worked out, I should just open a PR, correct?

bright glacier
#

Yup!

#

I'll try to review it soon after but I can't make promises as this week is a bit tight.

silent apex
#

How would I go about using those utils to replace os.path.join()?

#

I'm not completely sure on that.

bright glacier
silent apex
#

I tried that, it still failed

bright glacier
#

uhhhm

silent apex
#

Wait a minute... it succeeded right above, I probably screwed something up

#

ah, forgot to str() it

silent apex
late dewBOT
#

tests/test_ipynb.py line 346

r"File 'tests[/\\]data[/\\]notebook_which_cant_be_parsed\.ipynb' "```
flat krakenBOT
silent apex
flat krakenBOT
#
dawnofmidnight
[psf/black] Pull request opened: #2574 fix: allow tests to be run from the tests/ directory

Description

Fixes the issues described [on Discord](#black-formatter message).

At the moment, tests break when they're run from a directory that is not the project root. Specifically, tests/. This makes changes to use the PROJECT_ROOT, THIS_DIR, and DATA_DIR constants defined in tests.util so that tests will be runnable from multiple locations.

Checklist - did you ...

  • [x] Add a CHANGELO...
flat krakenBOT
#
MarcoGorelli
[psf/black] Pull request opened: #2576 fix deprecation warning from tomli

Description

Checklist - did you ...

  • [ ] Add a CHANGELOG entry if necessary?
  • [ ] Add / update tests if necessary?
  • [ ] Add new / update outdated documentation?

Currently, running tests shows

  /Users/marcogorelli/black-dev/venv/lib/python3.8/site-packages/tomli/_parser.py:71: DeprecationWarning: Text file object support is deprecated in favor of binary file objects. Use `open("foo.toml", "rb")` to open the file in binary mode.
    warnings.warn(
flat krakenBOT
lament crow
#

With black becoming stable, are there any plans to include Python APIs in the stability guarantees?

spare hamlet
#

Who is expert at fixing indent errors

#

I have a. Lot of. Those to fix