stubtest: ignore more dunder pos-only errors#12294
Merged
hauntsaninja merged 1 commit intopython:masterfrom Mar 5, 2022
Merged
stubtest: ignore more dunder pos-only errors#12294hauntsaninja merged 1 commit intopython:masterfrom
hauntsaninja merged 1 commit intopython:masterfrom
Conversation
This isn't actually a reversion. This logic was asymmetrical for reasons lost to time. Although I suspect it was this way because it introduced only a few errors on typeshed. What changed was that as a result of python#12203 we actually started checking a lot more dunder methods. Previously, we only checked dunders if either: a) it was a special dunder, like __init__ or __call, or b) the dunder was defined on the actual stub class itself In particular, we started checking every dunder redefined at runtime that the stub just inherited from object. A possible intermediate option would be to not check positional-only arguments in the case where a stub inherits the definition. I'll experiment with that once typeshed CI is green.
JelleZijlstra
approved these changes
Mar 5, 2022
Contributor
|
According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉 |
AlexWaygood
reviewed
Mar 5, 2022
| # These two are basically useless for type checkers | ||
| "__hash__", | ||
| "__getattr__", # resulting behaviour might be typed explicitly | ||
| "__setattr__", # defining this on a class can cause worse type checking |
Member
There was a problem hiding this comment.
Hmm, this one seems to be rare enough that I think I'd prefer to consider each case separately and allowlist them in typeshed. I can't think of an example where including it in the stub would improve type checking, but you never know?
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This isn't actually a reversion. This logic was asymmetrical for reasons
lost to time. Although I suspect it was this way because the delta
only a few errors on typeshed.
What changed was that as a result of #12203 we actually started checking
a lot more dunder methods. Previously, we only checked dunders if either:
a) it was a special dunder, like
__init__or__call__, orb) the dunder was defined on the actual stub class itself
In particular, we started checking every dunder redefined at runtime
that the stub just inherited from object.
A possible intermediate option would be to not check positional-only
arguments in the case where a stub inherits the definition. I'll
experiment with that once typeshed CI is green.