What Discord bot do we use with GitHub here?
#black-formatter
neon loom
gusty grail
What Discord bot do we use with GitHub here?
it's a webhook
although, joe might be doing some filtering
neon loom
Legend. Thanks
gusty grail
with cloudlfare workers
(https://support.discord.com/hc/en-us/articles/228383668-Intro-to-Webhooks article by discord which uses github as an example)
bright glacier
I don't know whether cibuildwheel can produce a pure python wheel
neon loom
I don't want it to and don't get why it is
bright glacier
usually you would do the sdist and pure wheel outside of cibuildwheel
I've seen some discussion about sdists in the issue tracker, but I don't know what that meant
Oh, you meant that cibuildwheel was building a pure python wheel locally
neon loom
Yes
bright glacier
uhhhhh, probably something to do with the envvars, but I have never ran cibuildwheel locally
neon loom
My goal was to try and get a local way to test it to learn
So I thought try move as much as I could to the pyproject.toml since that's supported now
How did you test it all, trial and error with the aciton?
bright glacier
yup.
neon loom
I got the docker image and it was building in there, but for some reason was trying to build a native wheel
bright glacier
the only thing that comes to mind is that configuration is faulty since the BLACK_USE_MYPYC environment variable should've been set in the docker container
I would love to try it myself, but I'm busy today
neon loom
That's cool - I might play again later
Giving bandersnatch it's once a year overhaul / love today
I want to update pypi's install but I can never get movement / agreement there ... They still on 4.4.0
and old requests + deps with bugs ๐
mild stratus
Hi folks!
Is there any Pycharm/IntelliJ plugins for black?
I'm find out well plugin like scalafmt on intellij.
In official docs, recommend use external tools
mint barn
Hi folks!
Is there any Pycharm/IntelliJ plugins for black?
I'm find out well pl...
There's an integration with pycharm, and it's written in the docs, have you tried / seen that?
mild stratus
There's an integration with pycharm, and it's written in the docs, have you trie...
Yeah, I'm tried and apply it on my Pycharm. But I find out more something like as native plugin just install from market place.
flat krakenBOT
flat krakenBOT
flat krakenBOT
lament crow
re: this, might be worth changing the workflow default in repository settings to read-only once this is merged so that by default the token has the minimum permissions
dense jungle
re: this, might be worth changing the workflow default in repository settings to...
good idea, done
lament crow
wasn't entirely sure if you had access to repository settings, guess you do then ๐
dense jungle
I have great powers
lament crow
and great responsibilities
neon loom
And the right person for the job
brave epoch
what exactly is this?
flat krakenBOT
neon loom
what exactly is this?
Black? I'd suggest going an reading the documentation linked in the channel topic. https://black.readthedocs.io/en/stable/
neon loom
Interesting - As this is fixing a bug/regression - Does this go under --preview or not?
dense jungle
Interesting - As this is fixing a bug/regression - Does this go under --preview ...
Anything in 22.1.0 that doesn't crash or eat comments or anything similarly terrible goes in the stable style. This should be preview only.
neon loom
Ok, thanks. Was being lazy and didn't want to go re-read the policy
flat krakenBOT
neon loom
I haven't checked, but lets make sure development docs state about preview style + link to it - I'll try get to it today
flat krakenBOT
flat krakenBOT
flat krakenBOT
carmine parcel
!cban 550341177585434635 racism
late dewBOT
:incoming_envelope: :ok_hand: applied ban to @broken stratus permanently.
molten hare
!cban 550341177585434635 racism
carmine parcel
hah
flat krakenBOT
flat krakenBOT
flat krakenBOT
errant barn
'Tis not the channel for it mate, try the dedicated helps!
lethal valley
Where? If you dont mind me asking?
lethal valley
'Tis not the channel for it mate, try the dedicated helps!
Sorry
errant barn
Also just posting an exercise without showing what you've done to solve it yourself could be considered impolite ๐ The various help channels are in the "available help channels" topic ๐๐ฟ
lethal valley
Also just posting an exercise without showing what you've done to solve it yours...
Oh...
errant barn
Better to show some effort I mean
lethal valley
Yea lol i got to the function but that al.
mint barn
jI
misty solstice
hello, is there any easy way to get black and isort working nicely together in a github workflow?
lament crow
just run one after the other
misty solstice
got it
lament crow
since it's in CI, I suggest running both without any diff flag and running git diff after both complete so that one can see what changes need to be made to make both black and isort happy
dense jungle
isort has a black-compatible profile to make sure they don't fight over formatting
misty solstice
isort has a black-compatible profile to make sure they don't fight over formatti...
Yeah that's what I use in development
But I just didn't know if I could add it in a similar way in the workflow
lament crow
I personally would suggest running isort after black due to some magic trailing comma stuff
lament crow
since it's in CI, I suggest running both without any diff flag and running `git ...
if you use pre-commit and run it in CI, you get this for free btw
austere lava
hello, is there any easy way to get black and isort working nicely together in a...
There's also https://github.com/omnilib/ufmt if you want something off-the-shelf. (note: it uses usort instead of isort)
I wonder if we should mention this in our docs
misty solstice
thank you! ๐
rigid gorge
what is the recommended way to use black with pytest? pytest-black hasnt been updated since october 2020 and shows deprecated warnings with the latest pytest
bright glacier
ah looks I need to ignore the tracebacks while comparing analyses for that weird diff-shades issue
rigid gorge
what is the recommended way to use black with pytest? pytest-black hasnt been up...
i could fix it myself and i believe there's a pending pr that fixes it but i dont want to use the pending pr because that'd involve installing a dependency from git as well as fixing myself and publishing a release could take a while
bright glacier
I have never used black with pytest and I don't see any warnings from black's side so I'm of little help.
rigid gorge
ah
was just wondering if anyone knows of alternate ways to use it with pytest i guess
austere lava
im curious why do you need it in pytest?
rigid gorge
we've been enforcing formatting by making a pytest test for whether code is properly blacked
so tests fail if the code isn't formatted properly
little parrot
have you had a look at pre-commit?
austere lava
right, as opposed to a custom CI step that runs black check?
rigid gorge
i have but that requires everyone who contributes to use it (which is possible to enforce, but kinda hard)
rigid gorge
right, as opposed to a custom CI step that runs `black check`?
what do you mean?
errant barn
or a tox run
rigid gorge
we use tox to test multiple python versions, can it test for black formatting too?
errant barn
yup
errant barn
surely you can just make an environment and run black there, no?
as you would any other tool
austere lava
what do you mean?
late dewBOT
.github/workflows/lint.yml lines 27 to 28
- name: Lint
uses: pre-commit/action@v2.0.3```in our case this runs precommit as part of the CI suite, which in turns run black check
rigid gorge
oh github lets you automatically apply code before a merge?
austere lava
no, you just get a red signal saying your formatting is broken
same as a broken test
errant barn
so I guess it would be: PR red flag -> go for CI, local test run -> go for tox env
bright glacier
you can use pre-commit.ci to have fixes automatically applied, some folks might like it, but it can be nice
rigid gorge
you can use pre-commit.ci to have fixes automatically applied, some folks might ...
sounds great, i'll google that!
glacial swan
hey friendos
is there a way to run black on commits automatically?
so I can check with a pr if it's ok but run if black thinks it's sus?
errant barn
the aforementioned pre-commit!
glacial swan
Can you help me set it up?
errant barn
here's our config: https://github.com/psf/black/blob/main/.pre-commit-config.yaml#L7
late dewBOT
.pre-commit-config.yaml line 7
- id: black```I'm sure pre-commit has lots of good example resources
glacial swan
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v2
- name: Run linter
uses: psf/black@stable
with:
options: "--line-length 120 --check"
src: "."
this is what I use right now
errant barn
Oh do you mean running it online instead of before every commit locally?
rigid gorge
is there a way to run black on commits automatically?
so I can check with a pr i...
oh lmfao thats a concise way to phrase the question i just had
errant barn
the convergence
glacial swan
I want black to be ran online if possible
with only the stuff it needs
in the speacil .yml for it
glacial swan
```yaml
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-l...
or does it need to go into this one?
errant barn
then a similar job with black --check will work just fine!
glacial swan
what do you suggest I do in this?
black . 120?
I don't know what to do to autoformat on commits?
errant barn
oh sorry, I misunderstood
glacial swan
I am also a bit overwhelmed rn
I don't care if I make it autoformat prs and stuff, I just want it to check the workflow if it fails, to autoformat to 120
bright glacier
I'll note that you should define version right now since the action version is independent from the black version
errant barn
So to reiterate, you'd like for your CI run to automatically format the code with black and produce another commit to a PR?
glacial swan
Yes
errant barn
Or just check if the code is formatted correctly?
glacial swan
if it fails
bright glacier
I'll note that you should define version right now since the action version is i...
.... I should change this, cause it's dumb.
hot beacon
JD, that isn't what we are asking for
glacial swan
I would rather it autoformat when the pr fails and updates
errant barn
Who are "we" :D
hot beacon
@errant barn He misunderstood, he just want it to autoformat on commits.
hot beacon
Who are "we" :D
Me and some other ppl in his bot support
only one person to be specific with me
glacial swan
I think Autoformatting on prs and stuff if black fails is better tho
errant barn
I'm so confused
hot beacon
We just want autoformat when new commit comes.
Not PR commit or anything, just repo commit
bright glacier
so on the main / master branch only?
glacial swan
that works for me but I still want the workflow check
hot beacon
so on the main / master branch only?
yeah that
bright glacier
So in essence:
MAIN / MASTER:
- run black
- and push changes in a new commit
EVERYTHING ELSE:
- run black
- do nothing but fail the workflow
glacial swan
I would to keep my workflow though for new users
glacial swan
So in essence:
MAIN / MASTER:
- run black
- and push changes in a new commit
E...
yes
glacial swan
as I now no longer need to push a black fix
and it's done automatically
will it commit if black passes, I would rather it not commit if it passes with flying colors, i.e. black has no change?
bright glacier
Well given I don't think a prebuilt solution exists, it's going to be custom.
mint barn
What are you guys talking about
mint barn
Is anyone gonna answer my question
glacial swan
and you know not commit when there's no changes
bright glacier
well it is possible, but it requires some custom GHA logic
glacial swan
What are you guys talking about
workflow stuff
hot beacon
What are you guys talking about
committing formatted code when commit in main/master branch, failing in workflow if black sees its not formatted
glacial swan
I am a big newbie to workflow
bright glacier
if black pushes a commit to main, should that workflow fail still?
hot beacon
what do you mean ?
bright glacier
So in essence:
MAIN / MASTER:
- run black
- and push changes in a new commit
E...
We are talking about how to run black in CI according to these requirements @mint barn
flat krakenBOT
bright glacier
what do you mean ?
For main, if black changes a file (and then pushes a new commit as agreed upon) should the job be marked as failed still?
mint barn
We are talking about how to run black in CI according to these requirements <@45...
What is Cl
hot beacon
For main, if black changes a file (and then pushes a new commit as agreed upon) ...
eh, whatever is simple is fine
bright glacier
continuous integration, basically running automated checks against a changeset before deployment
bright glacier
this kind of workflow is not simple though
mint barn
I've worked with the lighter workflow of the year and a half years old when he was the first
errant barn
pre-commit could make it easier, if you're open to having the check and correction locally
hot beacon
well, how about,
MAIN/MASTER:
- format, then push change to new commit
only
mint barn
Guys how about both
mint barn
Find a way to make it with both
glacial swan
There's no way to get a workflow check?
you can't like
check commit
then recheck?
mint barn
Make 2 different ones
bright glacier
I am working onn it
bright glacier
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v2
- name: Run linter
uses: psf/black@stable
with:
options: "--line-length 120"
src: "."
- name: Commit & push changes (main only)
if: github.event_name == 'push' && github.ref_name == 'main'
continue-on-error: true
run: |
git config user.name "github-actions"
git config user.email "github-actions@example.com"
git add -A
git commit -m "[ci] Autoformat with black"
git push
- name: Fail job if changes were made (PR only)
if: github.event_name == 'pull_request'
run: git diff-index --quiet HEAD -- || (git diff && exit 1)
@glacial swan this should work
this is what the auto pushed commit looks like: https://github.com/ichard26/testblackaction/commit/581c3961a16908fa0052657871bbef25433bde76
I'm getting pretty decent at writing gha workflows ๐
not that I recommend what I shared as I provide no guarantees, but if it works I guess why not
glacial swan
```yaml
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-l...
so update ci.yml?
And this should work?
bright glacier
I just took the original yaml file you gave us earlier and added the bits to support what you wanted
bright glacier
@dense jungle this weird diff-shades behaviour where it's claiming changes exist when it shows none should be fixed now: https://github.com/ichard26/diff-shades/commit/bd7f033cda81175be2c46b381c02269dfa3bfe03
glacial swan
lament crow
you should use an email of a bot in order for it to not show on the Contributors tab in repo's Insights
bright glacier
what branch are you pushing to and is this part of a PR?
I'll note that since the job is run twice (once for the push event and another time for the pull event) on local PRs, the one triggered by the push event is kinda useless / ignored / ran without doing anything meaningful
GitHub unfortunately still doesn't provide any good way of cancelling / skipping jobs triggered by the push event when the pull_request event (and its copy of the jobs) is also going to trigger
You can block push events from firing on all branches except main, or skip the pull_request job if it's coming from the target repository but either of these are not ideal.
Anyway I got some work to do now, so I got to go.
bright glacier
what is the problem?
glacial swan
No
it looks to work
I mean
does this look fine to you?
oops
no it doesn't autoformat
glacial swan
what is the problem?
the problem here is it doesn't push the new version
rigid gorge
```yaml
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-l...
looks a bit complicated, i got this to work for my similar use case:
name: Lint
permissions:
contents: write
on:
push:
branches:
- main
jobs:
lint:
runs-on: ubuntu-latest
steps:
- name: Checkout code
uses: actions/checkout@v3
- name: Run black check
uses: psf/black@stable
with:
options: "--check --diff"
- name: Run black format
if: failure()
uses: psf/black@stable
with:
options: ""
- name: Commit black format
if: failure()
uses: stefanzweifel/git-auto-commit-action@v4
with:
commit_message: "Apply black to code"
this only runs the commit for fixing code with black if there was any code to fix in the first place
and it only runs for pushes to the main branch, but that's easy to change
glacial swan
what is the problem?
can you help have this actually push to master?
I can even dm the github if you want
glacial swan
you'll need to change the main string to master in the workflow
one final is that it pushes when on master anyway, can we prevent that?
bright glacier
What?
dense jungle
<@783088578363523104> this weird diff-shades behaviour where it's claiming chang...
yay, thanks!
bright glacier
Yea that's expected, that's why there is continue-on-error
so it doesn't mark the job as failed, only the step
glacial swan
Yea that's expected, that's why there is continue-on-error
This is annoying though
bright glacier
Look, I can fix it tomorrow, but I'm busy now.
glacial swan
ok
Can you please dm me a fix when you are available if possible ?
Thank you in advance
dense jungle
@glacial swan what's the problem you're having? I might be able to help
glacial swan
This is annoying though
this
the check passes
but I get a black
error like this
dense jungle
can you link to the workflow on GitHub?
glacial swan
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v2
- name: Run linter
uses: psf/black@stable
with:
options: "--line-length 120"
src: "."
- name: Commit & push changes (main only)
if: github.event_name == 'push' && github.ref_name == 'master'
continue-on-error: true
run: |
git config user.name "github-actions"
git config user.email "github-actions@example.com"
git add -A
git commit -m "[ci] Autoformat with black"
git push
- name: Fail job if changes were made (PR only)
if: github.event_name == 'pull_request'
run: git diff-index --quiet HEAD -- || (git diff && exit 1)
Here's the workflow file
glacial swan
can you link to the workflow on GitHub?
That's both
dense jungle
```yaml
name: CI
on: [push, pull_request]
jobs:
black:
runs-on: ubuntu-l...
hm maybe it fails the workflow only if the last command failed?
glacial swan
the workflow passes
I just don't want to see black complaining
if it can't commit as it's an empty commit
dense jungle
on the run you linked Black didn't seem to want anything changed
25 files left unchanged.
then you need to write some more complicated shell script than you have
could also consider using pre-commit.ci for this, it takes care of applying the changes for you
dense jungle
no, that runs in CI
despite the name it doesn't require people to use local pre-commit hooks
glacial swan
so it just autoformats on every commit unless it good?
dense jungle
yes
glacial swan
jovial coral
Should I expect black to move the key to the next line assuming the default line length of 80/88?
GGG = {
"12345678901234567890123456789012345678901234567890": "12345678901234567890123456789012345678901234567890"
}
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
peak wolf
do you know how I can fix Dubugger?
flat krakenBOT
wild haven
Image to ASCII (Pillow Required)
For it to work, there has to be an image with the python code in the same folder and you have to typeout its full name + extenion.
import PIL.Image
# ascii characters used to build the output text
ASCII_CHARS = ["@", "#", "S", "%", "?", "*", "+", ";", ":", ",", ".", "1", "2", "3", "4", "5", "6", "7", "9", "8", "A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M", "N", "O", "P", "Q", "R", "T", "U", "V", "W", "X", "W", "Y", "Z", "0", "`", "~", ",", ".", "/", ";", ":", ";", "[", "]", "|", "-", "_", "="]
# resize image according to a new width
def resize_image(image, new_width=100):
width, height = image.size
ratio = height/width
new_height = int(new_width * ratio)
resized_image = image.resize((new_width, new_height))
return(resized_image)
# convert each pixel to grayscale
def grayify(image):
grayscale_image = image.convert("L")
return(grayscale_image)
# convert pixels to a string of ascii characters
def pixels_to_ascii(image):
pixels = image.getdata()
characters = "".join([ASCII_CHARS[pixel//25] for pixel in pixels])
return(characters)
def main(new_width=100):
# attempt to open image from user-input
path = input("Enter a valid pathname to an image:\n")
try:
image = PIL.Image.open(path)
except:
print(path, " is not a valid pathname to an image.")
return
# convert image to ascii
new_image_data = pixels_to_ascii(grayify(resize_image(image)))
# format
pixel_count = len(new_image_data)
ascii_image = "\n".join([new_image_data[index:(index+new_width)] for index in range(0, pixel_count, new_width)])
# print result
print(ascii_image)
# save result to "ascii_image.txt"
with open("ascii_image.txt", "w") as f:
f.write(ascii_image)
# run program
main()
not the channel for it mate
flat krakenBOT
potent mango
hi everyone,
any idea why my black is not setting the trailing comma? It is doing the line break but not setting the trailing comma itself.
my pyproject.toml has only 2 lines:
[tool.black]
line-length = 120
exclude="file/path/"
is it only on a single line?
potent mango
I have a dict like this
{
"a": fun(msg, "ts"),
"id": ...,
"key": ..., "created": ...
}
and running black . in console will make the linebreak to
{
"a": fun(msg, "ts"),
"id": ...,
"key": ...,
"created": ...
}
but its not setting the last comma at the end
errant barn
potent mango
the problem was a comment on the next line
https://black.vercel.app/?version=main&state=_Td6WFoAAATm1rRGAgAhARYAAAB0L-Wj4ADmAHRdAD2IimZxl1N_Wm_vzTQ-P7e88-HPpthubCyLPup10IjvULL9Te6KZ1NodUgQMtz1NLUGAc5rjTKK_QogZJuLtXRH2sY5WOk6D72XAqkz1z1husCxzNbP5ioWHtGyI7H0HAVp5zntQmX8C62vRP-twtlIMnwtAB0bycSlKUlRAAGQAecBAACY-DmpscRn-wIAAAAABFla
errant barn
hmm, oh!
potent mango
(cool website btw with the black example formatting
errant barn
and putting it onto the same line does produce the trailing comma
the comment I mean
potent mango
true, so I guess I found a bug ๐
errant barn
yeah, at least I can't come up with a reason as to why that would be :D
feel free to submit if you'd be so kind!
potent mango
man, I dont envy you developer dealing with all these edge cases xD
errant barn
:D feel free to also contribute ๐ envy is a grave sin you know
potent mango
sure, I can open a ticket on github about it.
๐ feel free to also contribute ๐ envy is a grave sin you know
I might actually give it a try ๐
errant barn
or which way it goes, anyhows
potent mango
thank you for your help and have a nice day ๐
errant barn
> ๐ feel free to also contribute ๐ envy is a grave sin you know
I might actual...
cool, we'll be there to help!
you too
flat krakenBOT
flat krakenBOT
potent mango
cool, we'll be there to help!
sooo, I got the black dev environment locally up and running (great job with that btw, worked on the first try) and set up the test case that is failing, and went into black._format_str_once with a debugger to understand what is happening
I landed at black.linegen.delimiter_split() where I think the source of the issue lies, but I am not sure about that.
I got to this chunk of code, where the current_line is checked and always the last leave is selected and checked. If I understand correctly by some backwards looping through the line leaves
When I let the debugger iterate and always check, what current_line.leaves[-1] prints, it prints the last lcomment line # "longkey": ...
errant barn
sooo, I got the black dev environment locally up and running (great job with tha...
Yo! Lemme try, but small disclaimer: I'm not really an expert on the formatting code either
also, great to hear that the initial process was smooth :D
errant barn
When I let the debugger iterate and always check, what `current_line.leaves[-1]`...
It seems like the candidate location for a fix, but since it's spelled out in the code I think we ought to make sure that this is unintended first
i.e. submit an issue and discuss more there
potent mango
I will do that, ๐๐
austere lava
sooo i've been playing around with this: https://zsol.github.io/format/
it's client-side black ๐
little parrot
Webasm?
dense jungle
very creative use of mocking
lament crow
fyi, two single quote issues (#633, #1794) have been necroed :P
violet canyon
Can I share that link to gh repo?
austere lava
sooo i've been playing around with this: https://zsol.github.io/format/
I've updated this, sprinkled some delight on it and tweeted about it
that's how products are developed right?
bright glacier
lol you're not that wrong
I'm just doing regular web development with pure JS, HTML, and CSS right now ๐
austere lava
you could be writing python instead! ๐
bright glacier
I only have to clean up this messy JS so I think I'm good for now. Anyhow, how long did that website take you?
austere lava
you mean to write it?
bright glacier
yeah
austere lava
about 3 hours net total
bright glacier
oh wow that's not long at all
austere lava
yeah and I had no idea how to use py-script before ๐
not counting my wife's pixel-pushing that resulted in the wonderful css
bright glacier
haha, it's wonderful
how is the performance would you say?
it's obviously going to be slow as this is first-gen tech, but I have a lot of stuff open right now (discord's and vscode's electrons are not being kind to my CPU)
austere lava
Much better than I expected, that site reformats on every keypress and it's usable
bright glacier
oh that's great to hear
mint barn
who else loves tkinter
flat krakenBOT
bright glacier
hmmm the first thing that comes to mind would be mypyc
.. which would be concerning
can't reproduce the issue
os or python version dependent?
stuck vapor
no idea, couldn't reproduce this either
mint barn
!format
late dewBOT
mint barn
how do i format my code
bright glacier
!raw 975377672714731530
late dewBOT
== Raw message ==
\`โ`โ`โpython
print('test')
`โ`โ`โ
sick :D
buoyant grove
Guys, I have this code:
oldfile = open("FILE-%s.csv" % (state),'r')
newfile = open("FILE2-%s.csv" % (state),'w')
newfile2 = open("FILE3-%s.csv" % (state),'w')
for line in oldfile:
if re.search('\$1,5',line) or re.search('\$1,6',line) or re.search('\$1,7',line) or re.search('\$1,8',line) or re.search('\$1,9',line) or re.search('\$2,',line) or re.search('\$3,',line) or re.search('\$4,',line) or re.search('\$5,',line) or re.search('\$6,',line) or re.search('\$7,',line) or re.search('\$8,',line) or re.search('\$9,',line) or re.search('\$10,',line) or re.search('\$11,',line) or re.search('\$12,',line) or re.search('\$13,',line) or re.search('\$14,',line) or re.search('\$15,',line) or re.search('\$16,',line) or re.search('\$17,',line) or re.search('\$18,',line) or re.search('\$19,',line) or re.search('\$20,',line):
newfile.write(line)
else:
newfile2.write(line)```I was trying to filter out any lines which contains $1,5xx,xxx, $1,6xx,xxx,... all the way up to $20million, meaning filter out anything that have value of $1,5 million and up. Now, I realize this is not how to do it. Because it's also removing $1,500, $1,600... $20,000 (not the million 7-digit value).
Not sure how to improve /fix the above
lament crow
Guys, I have this code:
```py
oldfile = open("FILE-%s.csv" % (state),'r')
newfil...
this is a wrong channel to ask
It seems like you're asking a coding question, you should read #โ๏ฝhow-to-get-help
buoyant grove
this is a wrong channel to ask
dang, just realized that now sorry
mint barn
๐
bright glacier
is this on topic for this channel @mint barn ?
mint barn
what is
bright glacier
that eval command
mint barn
what eval command
bright glacier
Ugh whatever. Just don't do that again.
mint barn
i have no idea what ur talking about ๐ ๐
errant barn
Please refrain from advertising projects on this channel, it's for Black development
flat krakenBOT
flat krakenBOT
bright glacier
ugh, we still can't run CI on Python 3.11-dev on GHA for some reason
creating build/temp.linux-x86_64-cpython-311/aiohttp
gcc -pthread -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -I/home/runner/work/black/black/.tox/ci-py/include -I/opt/hostedtoolcache/Python/3.11.0-beta.1/x64/include/python3.11 -c aiohttp/_websocket.c -o build/temp.linux-x86_64-cpython-311/aiohttp/_websocket.o
aiohttp/_websocket.c:198:12: fatal error: longintrepr.h: No such file or directory
198 | #include "longintrepr.h"
| ^~~~~~~~~~~~~~~
compilation terminated.
error: command '/usr/bin/gcc' failed with exit code 1
[end of output]
I thought the new Cython release today was supposed to fix this ... >.<
gcc -pthread -Wsign-compare -DNDEBUG -g -fwrapv -O3 -Wall -fPIC -I/home/runner/work/black/black/.tox/ci-py/include -I/opt/hostedtoolcache/Python/3.11.0-beta.1/x64/include/python3.11 -c frozenlist/_frozenlist.c -o build/temp.linux-x86_64-cpython-311/frozenlist/_frozenlist.o
frozenlist/_frozenlist.c: In function โ__Pyx_AddTracebackโ:
frozenlist/_frozenlist.c:432:62: error: dereferencing pointer to incomplete type โPyFrameObjectโ {aka โstruct _frameโ}
432 | #define __Pyx_PyFrame_SetLineNumber(frame, lineno) (frame)->f_lineno = (lineno)
| ^~
frozenlist/_frozenlist.c:7127:5: note: in expansion of macro โ__Pyx_PyFrame_SetLineNumberโ
7127 | __Pyx_PyFrame_SetLineNumber(py_frame, py_line);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
error: command '/usr/bin/gcc' failed with exit code 1
[end of output]
looks like the interpreter core optimisation work still breaks frozenlist
dense jungle
```
creating build/temp.linux-x86_64-cpython-311/aiohttp
gcc -pthrea...
does aiohttp distribute the generated .c files?
if so they need to recompile and release
bright glacier
I'd guess not given the traceback file is build/temp.linux-x86_64-cpython-311/aiohttp/_websocket.o
although tbf I don't know how included C would work when being built from a sdist
dense jungle
I checked, there's .c files in the sdist
bright glacier
yeah me too
bright glacier
my only comment on https://github.com/psf/black/pull/3035 is that if we remove immediate newlines after blocks, that might effectively decide black's stance on newlines after class signatures nevermind, please see below!
I don't remember enough of the discussion in regards to classes and following newlines, but I do know they are somewhere in the issue tracker
I'm still +1 on the change though.
bright glacier
~~my only comment on https://github.com/psf/black/pull/3035 is that if we remove...
hmm, actually black already has an opinion about this :)
--- test.py 2022-05-18 16:45:22.675999 +0000
+++ test.py 2022-05-18 16:45:23.703591 +0000
@@ -1,10 +1,10 @@
class A:
-
def func():
pass
class B:
"""docstring"""
+
def func():
pass
!remind 5H audit recent code style changes to make sure our style documentation covers 'em
late dewBOT
late dewBOT
!remind 5H audit recent code style changes to make sure our style documentation ...
It has arrived!
bright glacier
that invite link is for the whole server, not this channel specifically as is currently
flat krakenBOT
loud whale
!raw 975377672714731530
late dewBOT
You are not allowed to use that command here. Please use the #bot-commands channel instead.
rigid inlet
Hola!
I really like black and we use it in our project at work.
We also recently added http://www.pydocstyle.org/ to our checks. A feature I wish black would have is to format D400 - First line should end with a period . If I create a PR that autoformats a period on docstrings, what is the probability for it to get accepted?๐ค
dense jungle
Hola!
I really like black and we use it in our project at work.
We also recently...
I'd be hesitant about that. What if it ends with a question mark, or someone uses a docstring parser that doesn't want the first line to end with a period?
rigid inlet
I'd be hesitant about that. What if it ends with a question mark, or someone use...
The pseudocode would be: If last_char not in {'.','?','!'':'}: add('.'). And pep257 clearly states The docstring is a phrase ending in a period. https://peps.python.org/pep-0257/#:~:text=The docstring is a phrase ending in a period.
lament crow
Black's aim is to make no changes to AST
If you want to automatically make changes like this, you should probably consider other tools
Instagram's Fixit perhaps?
rigid inlet
If you want to automatically make changes like this, you should probably conside...
Hmmm, good point! We also use pre-commit, so I think my other alternative would be to make a standalone D400-autofixer... ๐
lament crow
Yeah
molten hare
And a ping 1 2
dense jungle
@bright glacier thanks for diff-shades!
bright glacier
sigh, OK so these notification settings will not work. Thanks for the messages!
flat krakenBOT
surreal socket
Guys how to enable black formatter on new projects in vsc? I have it in User Settings but on new project its not working, i installed it on project venv.
when i use it from terminal on file it is working, but cant get it owrking on save in vsc on new project
rigid inlet
Guys how to enable black formatter on new projects in vsc? I have it in User Set...
- How does your workspace-settings file look like? 2. - What is the output of Python in vsc when you run autoformatting? Look for the output tab between problems and terminal and select Python?
rigid inlet
Guys how to enable black formatter on new projects in vsc? I have it in User Set...
Also, you probably just forgot to set python interpreter path, ctrl+shit+P -> select python interpreter and select your .venv
surreal socket
1. How does your workspace-settings file look like? 2. - What is the output of P...
My workspace settings are clean - though it will use my user settings?
I have venv python interpreter selected
still brook
Did you already pip install black in your .venv?
And setting VSCode to format on save
surreal socket
yes i installed but didnt set the format on save
though this settings will be imported from my user settings
surreal socket
And setting VSCode to `format on save`
yeah that solved thge problem but do i have to set this on every new project?
still brook
I think you need install black for every .venv, but for setting u don't if you have login with github in vscode the setting will sync automatically
rigid inlet
If you look at output->python you can see what vscode executes. eg: ~/code/py/**/.venv/bin/python -m black --diff --quiet ~/code/py/***/bulletin.py
bright glacier
@dense jungle was this you? ๐
dense jungle
<@783088578363523104> was this you? ๐
it's very useful ๐
bright glacier
(the other pageload is probably some bot)
dense jungle
for pyanalyze I have a similar workflow to black where the PR number goes in the changelog, so I bookmarked the page
bright glacier
Ah neat, Python Discord is starting to use the same changelog workflow too so now I have four main users ๐
I rewrote the codebase a week-ish ago; it was my first serious webdev project.
dense jungle
I think I only learned about the url parameters because you mentioned it here at some point, that was a big usability improvement for me
bright glacier
Yeah it's an open issue to advertise that feature better. Someone did come by and make Next PR Number update the URL to include the query parameters if you do a query which helps a little bit.
I should just add a little lightbulb admonition somewhere that mentions the URL query parameters
bright glacier
webdev is just โจ great and totally not soul-suckingโจ
this page has not helped: https://developer.mozilla.org/en-US/docs/Web/CSS/Containing_block
plain atlas
lunar gulch
west crown
hey everyone, kindly pinging again about https://github.com/psf/black/pull/3035
is there anything missing from this PR? can we merge it?
Would really like to start using it in Preview mode ๐
errant barn
As you mentioned in the last comment, we'd still need at least another review. We've agreed it's the better style, but an approving review is still needed. Also, still no comments about removing the class body newline. And to that end: did you consider the effect of previous code wrt. defs along with class bodies? Which is line 505 in the diff.
west crown
Hey @errant barn , I pinged it here so other reviewers could take a look at it ๐ Wasn't sure if this is the right place or I should have just add another comment in the PR (please let me know for next time!).
Surely this needs more review. No questions about that
errant barn
That's ok, we're all here ๐
If there's something to discuss then it's better to do it on the issue, so that we leave the paper trail
west crown
I still think that classes should be an exception to the rule and should still be formatted according to the existing rule, due to the example I presented there
errant barn
Fair, although personally I prefer to remove them ๐๐ฟ but what I meant was: defs trigger the previous formatting as well. So are there similar exceptions there?
west crown
I agree that further discussions should be in the PR. This was just a gentle ping ๐
I'm not sure what you mean. If you could please explain it better in the PR itself, I would really appriciated it
errant barn
Actually, going through the code once more addressed my concern already! So don't worry about it
flat krakenBOT
wary pollen
Hi,
Could anyone please tell, how to add exclude folder in the setup.cfg
Following does not work:
[black]
exclude = folder/example/some_folder/
But this one works:
black --check . --exclude=folder/example/some_folder/
flat krakenBOT
bright glacier
Yeah it's an open issue to advertise that feature better. Someone did come by an...
Resolved now โจ
It probably shouldn't have taken 79 insertions(+) and 10 deletions(-) but oh well lol
slow ibex
some my suggestions:
- format content of annotations in quotes:
x: "Union[X,Y]"->x: "Union[X, Y]" - replace
'X'->Xin annotations iffrom __future__ import annotationsenabled or python>=3.11 - remove
from __future__ import <smth>if this future import is pointless for target python version
bright glacier
All of those suggestions sound reasonable at first sight, but all of them are also AST-unsafe and I'm wary of adding more unsafe transformations.
We could add a --ast-unsafe flag or whatever, but that would hurt black's goal of minimizing configurability
slow ibex
just noticed that black can format match-case syntax, cool!
bright glacier
Given we maintain our own CST LL(1)ish parser, we can use whatever hacks we like to support PEG syntax lol
not exactly maintainable, but welcome to an aging codebase
errant barn
We could add a --ast-unsafe flag or whatever, but that would hurt black's goal o...
These requests keep coming, it could be useful to decide this one way or the other
cunning thunder
Is this formatter avilable for pycharm?
fast heart
flat krakenBOT
dense jungle
These requests keep coming, it could be useful to decide this one way or the oth...
^ here's one solution
bright glacier
I gotta say I like reading these GH embeds as a super duper condensed summary of what happened. Sometimes I'm too lazy to open 13 emails or whatever ๐
red sedge
@dense jungle seems like an ast bug still. Where does peephole optimization and constant folding fit?
dense jungle
yeah I assume it gets optimized away later
but IIRC ast.parse is before the peepholer
ast.parse("1 + 1") also doesn't get constant folded
red sedge
Black's promise should probably only apply after peephole and constant folding
f"ABC" r"ABC" and u"ABC" all make the same pyc?
dense jungle
0 RESUME 0
1 2 LOAD_CONST 0 ('abc')
4 RETURN_VALUE
>>> dis.dis('r"abc"')
0 RESUME 0
1 2 LOAD_CONST 0 ('abc')
4 RETURN_VALUE
>>> dis.dis('u"abc"')
0 RESUME 0
1 2 LOAD_CONST 0 ('abc')
4 RETURN_VALUE
I think black should promise the same pyc and not worry about the ast
bright glacier
The pyc doesn't sound very stable imo
dense jungle
I feel like that's a harder promise to implement since the pyc format is much less stable
If we can simply say "ast.parse is the same before and after, no ifs or buts", that's a much more convincing promise
red sedge
The same ast means the same pyc right? So it's a looser promise?
dense jungle
The same ast means the same pyc right? So it's a looser promise?
It's harder to check for
red sedge
Maybe there's some IR that cPython could commit to
dense jungle
That feels like a lot of work for very little gain
The cases where we could make ast-unsafe changes that are pyc-safe are pretty obscure
red sedge
f-string normalisation
dense jungle
And under my proposal, we'd still do minor AST-unsafe changes (like useless fs) by default, we could just offer an additional mode that guarantees zero AST changes
red sedge
I'd have expected this to be removed in the AST phase
dense jungle
I'd have expected this to be removed in the AST phase
report it to CPython ๐
red sedge
.>
f-strings being integrated into the peg parser seems like this could get fixed up
r/f/u prefix is a CST concern imho
dense jungle
Black already drops u prefixes for what it's worth
red sedge
Does the AST expose u prefixes?
dense jungle
no
red sedge
Hmm redundant f prefixes is definitely an ast bug then
Are redundant r prefixes in the AST?
dense jungle
Are redundant r prefixes in the AST?
no, the AST just gives you the string as it would appear in Python
no triple quoting either
red sedge
Points to redundant f strings being exposed to the AST being a bug
And related to the fact that f-strings get parsed in a second pass
dense jungle
Yes, it's definitely unexpected. When I tried it I expected it to be the other way around, because we removed fs from implicitly concatenated strings and that is AST-safe
"Module(body=[Expr(value=JoinedStr(values=[FormattedValue(value=Name(id='x', ctx=Load()), conversion=-1), Constant(value='a')]))], type_ignores=[])"
>>> ast.dump(ast.parse('f"{x}" f"a"'))
"Module(body=[Expr(value=JoinedStr(values=[FormattedValue(value=Name(id='x', ctx=Load()), conversion=-1), Constant(value='a')]))], type_ignores=[])"
Seems like a public python function to peephole optimise an ast and return a new one would be popular
flat krakenBOT
flat krakenBOT
bright glacier
Black is not a linter so nope
And don't worry you have found the right place!
Yes. You could run black manually with the --diff to see what would be changed but that's it
What would you want to see?
What kind of problems?
Yeah that's out of the picture since black doesn't operate by first detecting issues and then fixing them. It only knows how to turn some code into properly formatted code. Kinda like setting something in a silicon mold
The line (number) would be a lot easier but that's still unlikely
I have no idea, I don't use VSCode often and I have never really used its formatting integration
austere lava
I don't think format on type is supported by Black at the moment
(nor any other python formatter I'm aware of)
little parrot
I don't think format on type is supported by Black at the moment
Would this be possible if black switched to libcst?
austere lava
not really
at the moment
well, it depends. If you wanna format unfinished files while you're typing, that would need an error-recoverable parser
which neither libcst nor black has at the moment
there are plans for libcst to gain this feature though
there are many other considerations though. People often use Black to detect syntax errors; that wouldn't be as easy with such a solution
little parrot
People use black to detect syntax errors 
?
silent apex
there are many other considerations though. People often use Black to detect syn...
Why is this? Error resilient parsing still emits errors, it just requires a good recovery strategy, does it not?
austere lava
yeah and it'd require rewriting Black
silent apex
there are plans for libcst to gain this feature though
Ooh that's quite interesting. Is libcst generated from a grammar or is it entirely handwritten?
austere lava
the parser is generated using PEG at the moment
silent apex
got it, got it
silent apex
oh yeah, I remember looking through that. I haven't had a very good experience with autogenerated parser and good error reporting/resilience.
austere lava
well, CPython has a generated parser ๐
it's the only mainstream language implementation I know of that doesn't hand-roll its parser
silent apex
well, CPython has a generated parser ๐
doesn't that still only find a single error at a time?
austere lava
yep
I mean it technically can find multiple errors but just reports the furthest one
silent apex
yeah, for the python parser I've been writing, I've pretty much resigned myself to handwritten for error recovery >.<
silent apex
I honestly don't know, beyond "I felt like it"
austere lava
best reason 100%
silent apex
does black operate directly on the libcst CST?
austere lava
I wish
silent apex
sorry, not libcst, but whatever other lib (blib2to3?)
slow ibex
I guess it works, but when you type your file has invalid syntax and black cant do anything
flat krakenBOT
drifting trellis
How can I make black not go insane on lists like this for command line arguments
I'd rather it use word wrap in this example than add 40 lines into my program.
bright glacier
You will have to disable black via # fmt: off/on or # fmt: skip
drifting trellis
I'm still interested in using black, just specifically not on this one list.
bright glacier
Hence the formatting control comment ^^
drifting trellis
Oh, didn't know that was even a thing. Didn't understand what you meant ๐
Thanks!
drifting trellis
๐ ty
drifting trellis
```python
# fmt: off
a = 'a'
# fmt: on
a = 'a' # fmt: skip
```
Btw this example was also helpful, as I had originally tried to use # fmt: skip on the line prior like you would # fmt: off
Thanks again
hot egret
You will have to disable black via # fmt: off/on or # fmt: skip
heard of this before, but unclear on exact allowed syntax - how exactly are these detected? do fmt controls still work if the line is, e.g. a = 2 # fmt: skip (reason why), or # explanation, so we will use # fmt: off? or does skip have to be at the end and on/off the only contents on a line?
bright glacier
I honestly don't know. I personally will try to avoid # fmt: skip since I don't trust it
the implementation is pretty buggy and I've been meaning to improve it but obviously that hasn't happened yet
I'm pretty familiar with # fmt: off/on though
lament crow
I'm team black all the way 
https://github.com/Pierre-Sassoulas/black-disable-checker
bright glacier
๐ my goodness
dense jungle
I love it
quartz stag
small request, i need a pep 9001 black formatter
silent apex
small request, i need a pep 9001 black formatter
That's certainly not a small request, that would require changing the parser that black uses a fair bit, as well as the actual formatting code.
So you'd basically need to rewrite the whole project.
red sedge
small request, i need a pep 9001 black formatter
How about this instead? https://pypi.org/project/flake99/
red sedge
How can I make black not go insane on lists like this for command line arguments
I tend to automatically generate these, or use "...".split(",")
drifting trellis
I tend to automatically generate these, or use "...".split(",")
I do too, but Iโm trying to make the code as straightforward to understand as possible.
In this context, this is simpler to understand.
bright glacier
@dense jungle how would you like to be credited in my mypyc blog series? I can do full name and link to your github profile, or would you like something else?
I'm still in the process of editing it, but I'm working on it again. Hopefully I can finish editing today ๐ค
exotic stratus
Hey, I've been trying out black and find its results pretty nice for the most part, but this is something I don't quite understand; why would the compact format (single line) be considered better by black? Imho, it's less readable when passing multiple kwargs in. Wondering about a second opinion on this.
dense jungle
<@783088578363523104> how would you like to be credited in my mypyc blog series?...
that sounds good, thanks!
little parrot
Hey, I've been trying out `black` and find its results pretty nice for the most ...
you can add a trailing comma to force the parameter list to be exploded
fluid hearth
Hey, I've been trying out `black` and find its results pretty nice for the most ...
yup I prefer the latter, btw : Embed doesn't do anything
bright glacier
After who knows how many hours, my mypyc series detailing the journey integrating mypyc into black is published: https://ichard26.github.io/blog/2022/05/31/compiling-black-with-mypyc-part-1/ ๐
bright glacier
After who knows how many hours, my mypyc series detailing the journey integratin...
little parrot
This is an interesting read thanks for writing this up
solid adder
Friends, bump @bright glacier's blog posts on making Black faster on Hacker News:
https://news.ycombinator.com/newest
solid adder
Please don't link to the submission directly
Bumping from a direct link doesn't do anything
bright glacier
lol, I'm not enough of a nerd to know how to use HN. I don't even have an account there, thanks y'all
solid adder
It's annoying the link has to be dug out of /newest but it's how the game is played
dense jungle
do you know for sure that's true? I've heard it before but who knows whether they changed the rules
solid adder
I don't know anybody at Y Combinator if that's what you're asking
But it's a well known rumor at least
fast heart
@bright glacier Which tool produced this graph?
I saw something like it on the discourse
bright glacier
I thought I linked 'em?
fast heart
I skipped the important line and looked at scalene ๐
bright glacier
but either way I used cProfile and then used gprof2dot (the yelp fork, not the original one) piping the output into the dot linux utility
it is really useful
solid adder
The blog post will fall off the first page of https://news.ycombinator.com/newest soon. Upvote it you cowards
silent apex
oh no I can't find it anymore
quartz stag
it is really useful
heya, if i had some questions re: mypyc could i ask them here or is it strictly black? although i might have to do with it in the future re: async
actually i answered what i was gonna ask thru: https://github.com/mypyc/mypyc/issues/868
tight walrus
Hi
flat krakenBOT
flat krakenBOT
wicked salmon
After who knows how many hours, my mypyc series detailing the journey integratin...
meow
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
bright glacier
what were they even thinking?
austere lava
I'm glad at least they provided reasonable repro steps instead of just the error ๐
lament crow
was it always like this? I know that call chains are at least somewhat preserved if they're longer than line length but did black always join it into one line if it fit?
austere lava
IIRC it's been like this for a long while now
dense jungle
yes
flat krakenBOT
flat krakenBOT
bright glacier
they are the maintainer of that plugin given the repo is owned by them..?
either way, I'm not sure if they know how python packaging works
flat krakenBOT
flat krakenBOT
dense jungle
weirdly https://github.com/cudatext-addons/cuda_fmt_python_black seems to have a copy of Black in it
bright glacier
what is even going on
errant barn
honestly
dense jungle
can we politely tell them to stop opening issues
and we're not going to fix their project for them
flat krakenBOT
dense jungle
interestingly CudaText has nothing to do with CUDA the GPU programming thing, it's an actual text editor
written in Pascal of all things
bright glacier
Yeah, I was going to say the CUDA team should know Python better but nope, it's not even CUDA related
flat krakenBOT
frozen yacht
I'm running into a case where I pass a file to black e.g. black module.py but it just says "No Python files are present to be formatted. Nothing to do"
Yet when I paste the code into the playground it does format it.
What gives?
bright glacier
What is emitted when you pass --verbose to the command?
frozen yacht
The file is snekbox/nsjail.py
What I have is force-exclude = ["snekbox/config_pb2.py"]
How does that match?
bright glacier
I have no idea but force-exclude's type is a string
frozen yacht
oh
bright glacier
hence regular expression in the output
frozen yacht
I though it was a list of regexes
In hindsight I suppose that doesn't make much sense
bright glacier
Ah fair enough, no worries. I wonder if the help output could be any clearer. I think we state it's a regex though.
frozen yacht
I think i just didn't read carefully
Or maybe confirmation bias from excludes of the same file for other tools requiring a list
dense jungle
wonder if we could give an error if it's a list
frozen yacht
That would be nice
Or more generally reject config files if they don't conform to the schema
Instead of having some silent behaviour
bright glacier
OK I really need to sit down and review some PRs.
I've been kind of out of it lately, but I have some spare time
flat krakenBOT
flat krakenBOT
flat krakenBOT
red sedge
3.10 should also raise this warning
probably want to run pytest with filterwarnings=error
black.concurrency.shutdown is redundant on 3.9 https://docs.python.org/3/library/asyncio-eventloop.html#asyncio.loop.shutdown_default_executor
earnest gust
can anyone help me with how to use it? i dont really understand the doc
ive installed it from github using terminal
halcyon rain
can anyone help me with how to use it? i dont really understand the doc
try black [filename].py
flat krakenBOT
halcyon rain
no way to set it as default formatter?
earnest gust
thank you
fast heart
Is this expected?
def benchmark(function=None, **kwargs):
if kwargs:
def _inner_decorator(function):
_benchmark(function, **kwargs)
return _inner_decorator
if function is not None:
_benchmark(function)
else:
raise TypeError("benchmark requires a function to benchmark")
Referring to the blank line after the if statement
Black seems to be enforcing it
errant barn
Is this expected?
```python
def benchmark(function=None, **kwargs):
if kwarg...
expected, but subject to change: psf/black#3089
toxic stormBOT
fast heart
expected, but subject to change: psf/black#3089
Thanks!
flat krakenBOT
flat krakenBOT
flat krakenBOT
flat krakenBOT
errant barn
"Any color you like" is a reference to Henry Ford
bright glacier
Hmm I thought black had support for some kind of fluid call structure? re. https://github.com/psf/black/issues/3100
I can't remember what it looks like though.
flat krakenBOT
bright glacier
myst-parser likes to pin sphinx hence the combined PR
bright glacier
We have 31 open pull requests ๐
Either way, I'm probably done for the night. I still have a lot of things to pick up, but at least I have a start now ๐
Can't wait until this goes down again ๐
austere lava
Hmm I thought black had support for some kind of fluid call structure? re. https...
Once it doesn't fit on the line, yeah
bright glacier
Once it doesn't fit on the line, yeah
Ah that explains it.
Thanks for clarifying!
fast heart
some_dict = {
"something_something":
r"Lorem ipsum dolor sit amet, an sed convenire eloquentiam \t"
r"signiferumque, duo ea vocibus consetetur scriptorem. Facer \t",
}
Under preview, what should this be mapped to?
I think the issue is related to the fact that r-strings aren't merged with StringMerger
Then I think StringParenWrapper makes a mistake
neon loom
ichard action spamming my inbox ๐
flat krakenBOT
bright glacier
ichard action spamming my inbox ๐
Haha, I found a little bit of free time
austere lava
re #3119 above, I don't know if we should accept partial solutions like this. I fully get that a majority of fstrings only have trivial expressions (names or attributes) in them, but i'm not entirely sure the added complexity is worth the benefit
neon loom
@errant barn - Are we missing something on https://github.com/psf/black/pull/3114 - This is definitely causing redundancy right?
Or am I missing something
bright glacier
Yeah, it's definitely redundant. They have a point that it shouldn't probably be in the test suite though
errant barn
<@273104906083106816> - Are we missing something on https://github.com/psf/black...
Currently the redundancy is local, since we have a pre-commit hook and we also test it. Moving it to tox wouldn't help, so if we wanted to get rid of it, we could let pre-commit handle local development and have a CI check dedicated to black
Oh is his most recent point true, that even our dev pre-commit runs with the latest released version? In that case the redundancy would be justified
errant barn
Yeah, it's definitely redundant. They have a point that it shouldn't probably be...
And I'm somewhat ambivalent about this. It's convenient to have it in the test suite, but perhaps not the most elegant thing. Getting rid of the hard coded list makes it better by a mile either way though
austere lava
I'd like to close https://github.com/psf/black/issues/333
- I don't see Black ever gaining support
- There are other tools that already do a great job (isort and usort come to mind)
errant barn
big +1
neon loom
And I'm somewhat ambivalent about this. It's convenient to have it in the test s...
Yeah Iโm all for that with a pathlib iterdir + deny list if needed
neon loom
I'd like to close https://github.com/psf/black/issues/333
Ya. This isnโt for us. Letโs just link isort and usort
Both are way slower so I donโt want that slowness in black
lament crow
Since isort 5 got released with proper black support, there really isn't a good argument to support optimizing imports in Black
and now there's also usort
The state of things is a lot better than it was back when the issue was made.
Once isort 5.11 is released, the only discrepancy I'm aware of - magic trailing commas - will be properly supported as well.
austere lava
Usort needs some love there
๐
on a related note, fastero is pretty cool; gotta remember to use it instead of timeit
neon loom
This how you profiled?
austere lava
Nah I used py-spy for profiling
bright glacier
Oh is his most recent point true, that even our dev pre-commit runs with the lat...
It is false-ish
late dewBOT
.pre-commit-config.yaml lines 7 to 13
- id: black
name: black
language: system
entry: black
minimum_pre_commit_version: 2.9.2
require_serial: true
types_or: [python, pyi]```we run black over the project as a system hook, so effectively we're just running black as a shell command
now if the black command doesn't point to the dev version of black, then it'll probably use the pypi version that's available somewhere on the system
for CI, we install the dev / currently being tested revision of black before running pre-commit
late dewBOT
.github/workflows/lint.yml lines 22 to 28
- name: Install dependencies
run: |
python -m pip install --upgrade pip
python -m pip install -e '.[d]'
- name: Lint
uses: pre-commit/action@v3.0.0```not sure why we install with the [d] extra to be honest, but we are using the development version
bright glacier
I'd like to close https://github.com/psf/black/issues/333
+1
mint barn
sussy baka
cloud tapir
Hi Team! May I have a pointer to whatโs the difference between black and black[jupyter]? I searched in Black Repo and documentation. It seems that black[jupyter] adds extension based on Black to support formatting Jupyter notebooks. Iโm curious if there are any features in vanilla black but not in black[jupyter]? Thanks a lot!
errant barn
Hi Team! May I have a pointer to whatโs the difference between black and black[j...
just formatting notebooks! And [jupyter] isn't anything special, if you've installed ipython elsewhere that'll do!
cloud tapir
Thanks!
flat krakenBOT
worn zenith
I just want to share with the black team that all roughly 40 different python containers we run at my job are now entirely scrubbed by black and it is a pre-commit hook now too. We used to have little tiffs over standards and personal preferences, but having an outside decision maker helped us a ton by removing any agency on the matter. We just set the line length limit to 120 and forced to keep trailing commas in pyproject.toml files, which is also great, because now that we have the pyproject.toml standardized we've started to standardize our pytest env on that too.
So thanks for the work.
austere lava
I just want to share with the `black` team that all roughly 40 different python ...
thanks for the feedback, it's great to read success stories like this ๐
mint barn
Bonk on https://github.com/psf/black/pull/2525#issuecomment-1154716116 cc @bright glacier
austere lava
is that pico seconds?
nah just milli and microseconds
I used the fastero pypi package
neon loom
https://github.com/psf/black/runs/6884433369?check_suite_focus=true - Can someone help me workout what this diff-shades error means
Error: Invalid value for 'analysis': File 'preview-changes-main-6c1bd08f16.json' does not exist.
Also, when we targeting next release. I've been slack here. Looking at our changes - i think it's well past time?
dense jungle
yes, we should do it soon
neon loom
I'll take a look later today and get a PR up for the release and the someone can panic and say they want something merged first like always happens ๐
bright glacier
hmm I'm not sure if I agree with removing black from pre-commit
I guess from a purity standpoint, it's not great, but functionally it's mostly fine.
bright glacier
https://github.com/psf/black/runs/6884433369?check_suite_focus=true - Can someon...
It's been a recurring problem, the latest release doesn't type check cleanly under mypyc and I didn't make the workflow robust enough to handle that.
It compiles black before testing it because it actually saves time. I can make it fall back to running black under interpreted if compilation fails if you'd like.
bright glacier
I guess from a purity standpoint, it's not great, but functionally it's mostly f...
My main problem is that if contributors don't run tests (and even if they do so, they have to run a special test environment), then they won't know their code changes aren't black compliant until CI runs the lint job. That introduces a cycle which is annoying.
pre-commit for this situation is far from perfect as it doesn't always pick up the dev version of Black (usually when you don't activate a virtual environment) when checking the project, but IMO it's better than not having it.
I'm OK with moving the self-check outside of the test suite, but assuming we can work around pre-commit's flaws, then I'd suggest keeping black in pre-commit.
Pretty sure we could disable the black hook in lint CI if we so wanted.
neon loom
It compiles black before testing it because it actually saves time. I can make i...
Or how can we help with the compilation problem here? Should we open an issue?
neon loom
Pretty sure we could disable the black hook in lint CI if we so wanted.
My main goal is not wasting CI and local dev time running it twice. I already dislike we have to run pre-commit and tox seperately anyways - But that seems to be the python standard CI way these days ...
I also fine running the stable version will cause confusion and pain if we have formatting changes locally but the stable version does not - This will happen in next release we move preview changes onto main by default in late December
I also personally don't have the energy to debate, so if people want it back, feel free to restore it. I just find it wasteful, potentially causing confusion and we don't get pre-commit running on our differnet supported environments.
Ideally it would be nice to have pre-commit to install black from the local commit.
errant barn
Hi, sorry but this channel is for Black development, try the #help category!
wicked schooner
ok
grand hatch
?
errant barn
Hi, sorry but this channel is for Black development, try the #help category!
bright glacier
psf/black#3111 makes no sense to me.
toxic stormBOT
bright glacier
I'll check the sdist for a few of the reported versions, but according to the tags, ParameterSource should exist ยฏ_(ใ)_/ยฏ
I wonder if a redistributor of click messed up
austere lava
hm yeah click.core.ParameterSource has been a valid name since click 8.0
fast heart
I assumed it would be ubuntu or debian but I couldn't reproduce in a ubuntu VM or in a ubuntu docker image
And the debian package worked fine in a virtual environment
Perhaps it could be something like conda causing two different versions to be installed and whichever python install is running black is using the wrong version
Ubuntu seems to have 5 python-click packages, 3 of which are out of date so Ubuntu isn't entirely ruled out
Ubuntu's package management system scares me ๐
Oh interestingly, all of them are out of date. It's just that 3 of them have versions < 8.0 the other two are 8.0.3
flat krakenBOT
little parrot
What's the reason black splits strings that are too long at the end of a word and moving the white space to the start of the new line?
dense jungle
What's the reason black splits strings that are too long at the end of a word an...
Don't remember if it was an explicit decision, but I think this way it's easier to verify the spaces are there
mild abyss
i wish
flat krakenBOT
flat krakenBOT
exotic stratus
Two questions:
- is there a reason other than character length per line that black decides to put quotes in own lines? (see picture)
- if I have
max-line-lengthset in my.flake8, does black consider it? Or should the line adjustment be made to some other config file?
dense jungle
Two questions:
1. is there a reason other than character length per line that bl...
black doesn't look at .flake8, it has its own config file
exotic stratus
The docs on line length aren't exactly clear on this https://black.readthedocs.io/en/stable/the_black_code_style/current_style.html#line-length
dense jungle
not sure what you mean by (1), I don't see quotes on its own line
bright glacier
Yup, presumably when it's collapsed it violates the line length limit hence why it's split.
exotic stratus
The github readme mentions reading opts from pyproject.toml. Should I use that, or own config file that was mentioned above?
dense jungle
that's the file I meant
bright glacier
Jelle was unclear, Black only looks at pyproject.toml by default
dense jungle
I never misspeak ๐
bright glacier
Or make typos.
exotic stratus
Alright, seems like [tool.black] is the key
Thank you
Confirming that it was the line length
Excuse me for bothering further, is there a verbose option when running black via the CLI or otherwise, where each change is explained with the reasoning and which black rule it met for the adjustment?
dense jungle
Excuse me for bothering further, is there a verbose option when running black vi...
no, Black doesn't really separate out its changes
and it doesn't have a concept of "rules" even internally
rather most of the complexity is in a big algorithm for how to fit the code within the line length
exotic stratus
Ah, I intuitively thought it was some sort of a priority-based rule engine. My fault for not digging into the sources. Thank you regardless
dense jungle
for completeness, there is a verbose mode, but it mostly prints info about where it finds the config and what files it looks at
steep void
Has anyone run into this? - Currently running black v20.8b0 in my environment and running into issue when running pip install I hit the following error
Collecting black==20.8b0
Using cached https://nexus.jx.cicd/black-20.8b0.tar.gz (1.1 MB)
Installing build dependencies ... done
Getting requirements to build wheel ... done
Preparing metadata (pyproject.toml) ... error
error: subprocess-exited-with-error
ร Preparing metadata (pyproject.toml) did not run successfully.
โ exit code: 1
โฐโ> [51 lines of output]
Traceback (most recent call last):
File "/private/var/folders/0t/f4fb584j0d9d7bb4g6f05vy40000gp/T/pip-build-env-pbooiros/overlay/lib/python3.7/site-packages/setuptools_scm/_entrypoints.py", line 72, in <module>
from importlib.metadata import entry_points # type: ignore
ModuleNotFoundError: No module named 'importlib.metadata'
It just started happening today
If I upgrade the version to one of the latest versions of black I dont see the issue pop up anymore
bright glacier
I wonder if the latest setuptools-scm release broke things, again.
steep void
@bright glacier I think you maybe right, I see new versions of setuptools-scm
bright glacier
Preparing metadata (pyproject.toml) ... done
can't reproduce on Python 3.7 with black==20.8b0
steep void
if you uninstall black and importlib.metadata manually then install black, I was able to reproduce
bright glacier
If I upgrade the version to one of the latest versions of `black` I dont see the...
which ones? the ones from this and last year should work fine because wheels exist for them so there's no build step involved where setuptools-scm could crash while installing.
bright glacier
if you uninstall black and importlib.metadata manually then install black, I was...
That's odd since pip's build isolation should mean this shouldn't affect the build ...
steep void
black==20.8b0 upgrade ==> 21.12b0
I couldnt go to latest because of conflicts with click but this is two versions back
bright glacier
yeah we have wheels available for 21.12b0 so that does eliminate setuptool-scm entirely
steep void
@bright glacier thanks for the help on this, I'll test more on the setuptool-scm recent release but I think we'll just stick with the newest version (21.12b0)
bright glacier
I still can't reproduce the issue even if I uninstall importlib.metadata before install black==20.8b0
dense jungle
what is your pip version?
steep void
pip 22.1.2
@bright glacier could this possible be the issue that we were talking about
importlibs with python 3.8 or higher
bright glacier
<@429445826251587616> I think you maybe right, I see new versions of setuptools-...
It was an issue with setuptools-scm. I chatted with the setuptools-scm maintainer and they confirmed.
steep void
Thanks @bright glacier on the follow up, I found the following solution to work.
- black v21.8b0 - Pin setuptools-scm build-time dependency version (#2457) due to setuptools-scm breaking changes.
https://github.com/psf/black/pull/2457/files#diff-50c86b7ed8ac2cf95bd48334961bf0530cdc77b5a56f852c5c61b89d735fd711 - Upgrading to black 21.12b0 takes care of this issue completely
bright glacier
Ah that's good to know, thanks for investigating further!
If you encounter other packaging weirdness (well except for poetry or PDM, those aren't PyPA projects), just know discord.gg/pypa exists
If it's related to Black, obviously ask here first, but just in case it comes in helpful ^^
steep void
thanks for sending this, I'll join this, I like finding weirdness in code and tracing its source
bright glacier
I'm going to see if I can work on the release CI situation to hopefully unblock a new release
I'd appreciate it if someone could go through the commit log / changelog and make sure any code style changes have been (sufficiently) documented in the (future) Black Code Style documentation