Repository navigation
Pep8 doesn't speak to the walrus #12147
Description
Description
Pep8speaks doesn't like walrus operators.
Expected behavior
Now that we are py3.8+ (#11934), this should not trip up the checker.
Actual behavior
Pep8speaks flags := as invalid syntax.
Steps to Reproduce
Push code with a := to astropy.
System Details
N/A.
Activity
nstarman commentedon Sep 5, 2021
@pllim I see this issue has been raised in pep8speaks-org/pep8speaks#173
A suggested fix is to use
pycodestyleover flake8.Or we could just wait until pep8speaks-org/pep8speaks#173 is fixed...
[-]Pep8 speaks vs. the walrus[/-][+]Pep8 doesn't speak to the walrus[/+]pllim commentedon Sep 7, 2021
If upstream can fix it within reasonable time frame, we can wait?
nstarman commentedon Sep 8, 2021
It doesn't make the CI fail; just gives a false error in the PR message. So long as maintainers know to ignore it, it's not a showstopper. The PR to fix the PEP8speaks error is a bit old, but now that many NumPy affiliated packages are dropping py3.7, I hope it's resolved soon...
pllim commentedon Sep 8, 2021
@nstarman , I don't see why we cannot use pycodestyle for pep8speaks, but I'd rather not switching everything from flake8 to pycodestyle, does that make sense?
@bsipocz , any advise here?
bsipocz commentedon Sep 8, 2021
flake8 does much more than pycodestyle so there shouldn't be a downgrade switch, the early adoption of walrus doesn't worth it imo. For cases like this, do patch upstream.
nstarman commentedon Sep 8, 2021
Just confirming, pep8speaks only runs on the diff. So a walrus in the codebase wouldn't make future PRs error, right?
pllim commentedon Sep 8, 2021
Theoretically, no.
nstarman commentedon Sep 30, 2021
Can confirm it only complains on new PRs. There are now a few walruses in the code and they are not stirring up trouble.
Side note. pep8speaks also complains about positional-only arguments
nstarman commentedon Jan 18, 2023
We don't have pep8speaks anymore!