Check *_pb2.pyi files again

[Avasam]

I was looking at Ruff vs Flake8 differences in typeshed and noticed this.

Start checking *_pb2.pyi files again as the issue seems to be resolved.
Fixes all NQA102 "# noqa: F821" has no matching violations that were ignored (we don't have a generation script for stubs/s2clientprotocol so that was done manually)

[AlexWaygood]

All these *_pb2.pyi files in our protobuf stubs are autogenerated with mypy-protobuf -- did you use the script here to regenerate them, or did you edit them manually? :)

If you used the latest version of mypy-protobuf to regenerate these files, we should probably update the pinned version here, so these changes aren't overridden the next time the files are regenerated:

typeshed/scripts/generate_proto_stubs.sh

Line 16 in 608eca5

MYPY_PROTOBUF_VERSION=v3.4.0

[Avasam]

The script ends at the followign in git bash:

+ python3 -m venv venv
+ source venv/bin/activate
./scripts/generate_proto_stubs.sh: line 47: venv/bin/activate: No such file or directory

And immediately fails with the following in WSL:

/usr/bin/env: ‘bash\r’: No such file or directory

So I can't really validate locally

Edit: \r because the script is in CRLF, I'll open a PR to fix that.

[AlexWaygood]

(Yeah, I've never run the script locally either)

[Avasam]

This should be better. I could auto-remove redundant F821 with Ruff once astral-sh/ruff#3011 is fixed
Completing #10914 resolved this issue.

[Avasam]

While we're at it:
Our Black configs "force-exclude" .*_pb2.pyi files with the following comment:

Exclude protobuf files because they have long line lengths
Ideally, we could configure Black to allow longer line lengths
for just these files, but doesn't seem possible yet.

But isn't that exactly what Black is for?
Did it use to make Black crash?
Is it because they're generated, so we don't want to bother formatting line-length? But then, shouldn't we not care about formatting them at all? Why is line-length specifically of concern here?

The protobuf stubs generation script even tries to run black (which currently ends up doing nothing), so I'm a bit confused.

[github-actions]

According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉

[AlexWaygood]

Thanks!