Silent B028 from flake8-bugbear

[msmdev]

No description provided.

[msmdev]

@michalk8, flake8-bugbear started throwing B028 after the merge of #54 (but not before, meaning that all CI tests in the fix/update-precommits branch passed before merging!). So I played around a bit by adding # noqa B28 to all warnings in question, but those were removed as unnecessary when running pre-commit...
In effect, after running pre-commit run --all-files, all my changes were reversed and the flake8-bugbear again throws B028. I assume there is an inconsistency in the pre-commit config? Anyhow, I'm confused that this didn't show up in fix/update-precommits before the merge... what am I missing here?

[msmdev]

Also pre-commit run --all-files graciously passes both here and in main, but linting is failing in CI.

[michalk8]

I'd suggest removing this line: https://github.com/msmdev/pyGPCCA/blob/main/.github/workflows/ci.yml#L32 (the cache should be invalidated when pre-commits change).

[msmdev]

I'd suggest removing this line: https://github.com/msmdev/pyGPCCA/blob/main/.github/workflows/ci.yml#L32

What is the effect of this line, exactly?

[michalk8]

What is the effect of this line, exactly?

It's an alternative key, when it cannot be found here.

[msmdev]

What is the effect of this line, exactly?

It's an alternative key, when it cannot be found here.

So it's basically using a old cached version of the pre-commit-config.yaml?

[michalk8]

What is the effect of this line, exactly?

It's an alternative key, when it cannot be found here.

So it's basically using a old cached version of the pre-commit-config.yaml?

Yes, it shouldn't be there. But looking closely at the run, no cache-hit happened, so it's something else.

However, I think the B028 should be fixed in the code; tried locally (with different Python versions), can't also seem to reproduce the error on CI.

UPDATE: Can reproduce after I remove ~/.cache/pre-commit.

[msmdev]

What is the effect of this line, exactly?

It's an alternative key, when it cannot be found here.

So it's basically using a old cached version of the pre-commit-config.yaml?

Yes, it shouldn't be there. But looking closely at the run, no cache-hit happened, so it's something else.

However, I think the B028 should be fixed in the code; tried locally (with different Python versions), can't also seem to reproduce the error on CI.

UPDATE: Can reproduce after I remove ~/.cache/pre-commit.

What do you mean by "the B028 should be fixed in the code"?
If I add # noqa B028 in the respective lines, they are removed by yesqa during linting... so I can't prevent flake8 from complaining.

[msmdev]

Anyhow, flake8 is only complaining during the GitHub workflow - locally flake8 is graciously passing without # noqa B028...

[msmdev]

@michalk8, ooook, ignoring B028 in tox.ini solved the problem, but I'm still quite upset why flake8 is issuing B028 during GitHub workflows and not locally. -.-

[msmdev]

@michalk8, ooook, ignoring B028 in tox.ini solved the problem, but I'm still quite upset why flake8 is issuing B028 during GitHub workflows and not locally. -.-

@michalk8, would you please review? Thank you!

[michalk8]

LGTM, thanks @msmdev !