Stanislav Terliakov

This is funny! Yes, type vars are now bound in the function body if typevar is used in argument types or return types (or comes from parent scope - class...

mypy simply does not support aliased `dataclass` at all. ```python from dataclasses import dataclass dataclass_alias = dataclass @dataclass class Ok: a: int = 1 @dataclass_alias class Bad: a: int =...

Hm, I don't see any straightforward solution that could set constraints priorities reliably. But your Literal suggestion sounds interesting, I'll try to expand this PR tomorrow to cover this case

OK, `Literal` constraints are significantly different, nothing prevents their precise inference - we already keep track of `last_known_value`. That should be handled in `join.py` instead and does not belong to...

The primer change is bad but unrelated. Opened #19358 to address it.

TBH, I tend to agree with @ilevkivskyi - there's a balance between "we want to support this case correctly" and "we don't want to embed too many library-specific details into...

@ilevkivskyi tagging you since you gave some feedback on the first revision. I managed to remove the scary part - error messages are less specific now, but only `special_sig =...

> LG, I don't really understand the bound start args logic, but it will be easy to tweak if it will cause problems. I just invented a counterexample to that...

Primer diff shows only one false positive removed. That error was caused exactly by mistreatment fixed in this PR.

This is being discussed at https://discuss.python.org/t/overloads-of-async-generators-inconsistent-coroutine-wrapping/56665.