Conversation
| conformance_automated = "Fail" | ||
| conformant = "Pass" |
There was a problem hiding this comment.
I know I said we could do this if it made sense, but @JelleZijlstra 's new invariant-checker won't like this. We could either put up a PR to make this assertion less pedantic (is this really required by the spec?), or I intend to put up a PR that will remove this Unknown | on the ty side shortly, so we could also just mark this Partial until then.
There was a problem hiding this comment.
Another option is to use ignore_errors to ignore certain errors.
| @@ -0,0 +1,19 @@ | |||
| conformance_automated = "Fail" | |||
| conformant = "Unsupported" | |||
There was a problem hiding this comment.
This "Unsupported" doesn't feel right for the name generics_variance, because we support a lot of variance checking, but I guess what this file is really testing is all about enforcing not using the wrong explicit-variance legacy TypeVar in the wrong place, and we don't yet do that...
| conformance_automated = "Fail" | ||
| conformant = "Partial" | ||
| notes = """ | ||
| Deliberately does not allow `str` to be narrowed to literal string types through equality or containment checks due to the possibility of `str` subclasses that could have unexpected equality semantics. |
There was a problem hiding this comment.
I think we should make a PR to remove these assertions from the conformance suite. Unsound narrowing behavior is not required by the spec and shouldn't be asserted by the conformance suite.
AlexWaygood
added
the
topic: conformance tests
3 participants
No description provided.