Add flake8 problem matcher

[jackwilsdon]

Description

This PR implements the flake8 half of #3614 using a problem matcher, giving us nice inline errors like this on PRs:

I've had to disable PY_COLORS here, as the ANSI escape sequences were tripping up the regex used in the matcher.

To Do

• Documentation. (If you've add a new command-line flag, for example, find the appropriate page under docs/ to describe it.)
• Changelog. (Add an entry to docs/changelog.rst near the top of the document.)
• Tests. (Encouraged but not strictly required.)

[jackwilsdon]

Hmm, I've noticed that even though we're setting the code group, GitHub isn't actually showing the error code anywhere. Is it worth us compounding this back into the message and dropping the code group completely? We'd end up with the code in the message like this:

[jtpavlock]

I just want to double check this won't be effected by the changes in #3694. In particular, the output from pre-commit run -a would display flake8 errors as such:

flake8...................................................................Failed
- hook id: flake8
- exit code: 1

beetsplug/chroma.py:1:1: FI10 __future__ import "division" missing
beetsplug/chroma.py:1:1: FI11 __future__ import "absolute_import" missing
beetsplug/chroma.py:1:1: FI13 __future__ import "print_function" missing

[jackwilsdon]

Provided we output ::add-matcher::.github/flake8-problem-matcher.json in one of our steps, it should continue to work with flake8 output no matter where it is:

[jtpavlock]

Hmm, I've noticed that even though we're setting the code group, GitHub isn't actually showing the error code anywhere. Is it worth us compounding this back into the message and dropping the code group completely? We'd end up with the code in the message like this:

I like it with the error code, but probably not worth it if it's too much trouble

[jackwilsdon]

Changed in 5d24cb0 - was a nice and easy fix 👍

[jtpavlock]

will anything change if you do [^\s] rather than [^ ] ? (sorry for not asking this before you changed it 😄 )

[jackwilsdon]

I don't think so — the match is working correctly but GitHub just doesn't display the code anywhere in their UI.

I'll give it a shot in a bit though to see if anything interesting happens.

[jtpavlock]

the match is working correctly but GitHub just doesn't display the code anywhere in their UI.

Ah okay, I see what you're saying now. Well this looks cool! I'm excited to see it in action. I think you should feel free to merge when you're ready 👍

[jackwilsdon]

Awesome, thanks!

[sampsyo]

Wow, this is awesome! I didn't know this was possible. This will be such a time saver!!

[jef]

Wow... This was a lot easier than expected. Nice work guys.

[jackwilsdon]

There may be some beneifts to using custom actions but this seemed like the path of least resistance for the time being.

[jef]

That makes sense to me. Much easier just to put in this way rather than wrap it up another action.

We could've used local actions, but IMO that would be even worse (readability) I suppose.

Nice work!