gh-105834: Add tests for calling issubclass() between two protocols

[AlexWaygood]

• Issue: Untested code in typing.py #105834

[carljm]

Your reasoning on the issue looks right to me.

[JelleZijlstra]

We need to improve our test coverage instead. This block covers a case like this (add this to test_basic_protocol):

diff --git a/Lib/test/test_typing.py b/Lib/test/test_typing.py index 3eb0fcad69..84e9f2650f 100644 --- a/Lib/test/test_typing.py +++ b/Lib/test/test_typing.py @@ -2474,6 +2474,11 @@ def f(): self.assertNotIsSubclass(types.FunctionType, P) self.assertNotIsInstance(f, P) + class HasAnno(Protocol): + meth: Callable[[Any], Any] + + self.assertIsSubclass(HasAnno, P) + def test_runtime_checkable_generic_non_protocol(self): # Make sure this doesn't raise AttributeError with self.assertRaisesRegex( 

On current main, this test passes, but with your patch applied, it no longer passes.

[bedevere-bot]

When you're done making the requested changes, leave the comment: I have made the requested changes; please review again.

[AlexWaygood]

Thanks @JelleZijlstra! I knew I had to be missing something 😖

Here's a complete repro:

>>> from typing import * >>> @runtime_checkable ... class CallableMembersProto(Protocol): ... def meth(self): ... ... >>> class NonCallableMembersProto(Protocol): ... meth: Callable[[], None] ... >>> issubclass(NonCallableMembersProto, CallableMembersProto) True

The last line would be False if we went ahead with this PR as it stands currently. I think you're right that this is definitely intended behaviour, and it seems untested currently.

I can see arguments both ways in terms of whether it's desirable behaviour, and it seems really unlikely to come up in practice. (PEP-544 doesn't seem to discuss this point at all.) But given that this has been established behaviour since Python 3.8, I agree that we should probably keep it. (I also can't find any significant performance improvement from removing this block of code.)

[AlexWaygood]

I have made the requested changes; please review again

[bedevere-bot]

Thanks for making the requested changes!

@carljm, @JelleZijlstra: please review the changes made to this pull request.

[miss-islington]

Thanks @AlexWaygood for the PR 🌮🎉.. I'm working now to backport this PR to: 3.11, 3.12.
🐍🍒⛏🤖

[miss-islington]

Sorry, @AlexWaygood, I could not cleanly backport this to 3.11 due to a conflict.
Please backport using cherry_picker on command line.
cherry_picker 70c075c194d3739ae10ce76265f05fa82ed46487 3.11

[bedevere-bot]

GH-105859 is a backport of this pull request to the 3.12 branch.

[bedevere-bot]

GH-105860 is a backport of this pull request to the 3.11 branch.