B022: No arguments passed to contextlib.suppress

[jpy-git]

Closes #222.

This covers the case where contextlib.suppress is used

import contextlib

with contextlib.suppress():  
    raise ValueError

However, it does not currently cover the case where suppress is imported from contextlib

from contextlib import suppress

with suppress():  
    raise ValueError

Do you know of any existing flake8 helper functions/methods to tie the suppress function node to it's import?
I don't want to just ignore the one word suppress without checking incase it's actually a user defined function.

[cooperlees]

You're to quick - Sorry - I just merged another PR. I can look at cleaning up here or if you beat me that would be awesome too :D

[jpy-git]

You're to quick - Sorry - I just merged another PR. I can look at cleaning up here or if you beat me that would be awesome too :D

hehe no worries, I've resolved them now 😄

Would be good to get your take on this PR currently only covering contextlib.suppress rather just suppress since we'd have to resolve scope to figure out that the suppress being called is definitely the one from contextlib?

[cooperlees]

All makes sense to me, and test show us working and handling * expansion which is good.

[cooperlees]

Would be good to get your take on this PR currently only covering contextlib.suppress rather just suppress since we'd have to resolve scope to figure out that the suppress being called is definitely the one from contextlib?

Ahhh, I didn't think of this. Yeah, I feel assuming with suppress would be to noisy and we shouldn't do that.

[jpy-git]

Ahhh, I didn't think of this. Yeah, I feel assuming with suppress would be to noisy and we shouldn't do that.

Yeah that was my thinking too, better to avoid false positives 😄

Maybe we initially keep the rule like this and then I can make a separate issue to find a way to track variable/function scope when traversing the AST as a future refinement (I think pyflakes has something similar for tracking if imports get used).