Function with generic class and generic value arguments are not inferred when generic is a TypedDict

Bug Report

When defining a function expecting a generic class and a generic value — tied to the same TypeVar —, mypy is not able to infer the type of the argument if the generic type is a TypedDict.

It works well for scalar types, if the variable is explicitly defined as the typed dict type, or even more surprising, if the first generic argument is binded via a functools.partial.

To Reproduce

https://mypy-play.net/?mypy=latest&python=3.12&gist=a6d36519fbd58e5c9b77999eedd16132

import typing
import functools

T = typing.TypeVar("T")


class A(typing.Generic[T]): ...


def f(a: A[T], t: T) -> None: ...


class D(typing.TypedDict):
    x: int


a_typed_dict = A[D]()
f(a_typed_dict, {"x": 1})  # error: Cannot infer type argument 1 of "f"  [misc]

Expected Behavior

The t argument is correctly validated as a D typed dict.

Actual Behavior

The following error message is returned:

error: Cannot infer type argument 1 of "f"  [misc]

Your Environment

• Mypy version used: 1.15
• Python version used: 3.12

まず、リンクされた mypy-play の再現コードを mypy 1.15 と Python 3.12 で実行します。TypedDict 引数の検証時における A[D] のジェネリック型推論を追跡します。完了条件は、呼び出しが "Cannot infer type argument 1" エラーなしで {"x": 1} を D として受け入れることです。