-
Notifications
You must be signed in to change notification settings - Fork 122
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
infra: fix ruff configuration and add a few checks #573
Conversation
@henryiii - I've gotta think on this -- the confluence of all the different settings is making my head hurt 😅 I don't think we can blindly set |
Did you mean to replace isort with ruff? You haven't opted into it. |
Not yet. I hit this and wanted to fix it first. And it might not be configurable enough yet. |
But you removed isort's config. |
Oops, thought I reverted that. 😅 |
Some of our files are being skipped because they have "build" in the name. Noticed this in pypa/build#573, and turns out problems were being masked here too. Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
a7a9c3d
to
dd574e0
Compare
That's better. Since we have custom spacing set for isort (not sure why?), I'm pretty sure we can't switch isort to Ruff (without changing it, that is). I'm not sure what the correct solution to the exclude problem is, but it does seem that the current situation is not ideal. But I might just be more susceptible to it, due to working on a lot of building stuff. :) I do think it's better to err on the side of not silently ignoring directors, as a linter can't tell if something is missed, it just never reports anything for that directory. I wouldn't mind adding manually adding extend-excludes for directories that are pretty common but not hidden. |
I think you accidentally reverted the ruff config changes too 😅 |
Signed-off-by: Henry Schreiner <henryschreineriii@gmail.com>
dd574e0
to
8e0edf8
Compare
I was looking at adding a few checks (trying to check the validity of replacing isort), and discovered fixes/checks weren't being applied. It turns out
build
is being ignored by default, so the whole package insrc/build
is being ignored! For the pre-commit check, I don't think the default excludes really make much sense, and I've got other projects with a submodule named build, too, so while it especially affects us, it does affect others too (scikit-build-core is one with a build submodule). I see why it might be ignored by default if you aren't running in pre-commit.CC @charliermarsh, would it make sense to pass
--exclude=""
or something like that for the pre-commit hook? Maybe the respect gitignore flag? Pre-commit has it's own system for skipping files, it seems broken that it's not being respected. Well, at least if one of the excluded directories is something you actually want included. I wouldn't mind keeping the other directories there so this can be run outside of pre-commit more easily, but I can't have ".../build" excluded. Maybe there just needs to be an "include" or "dont-exclude"?