unpacking dict in function call with optional arguments fails

[terencehonles]

Bug Report

Unpacking a dict when calling a function with **kwargs should not fail. This seems to be related to #4771 and #9676, but slightly different since the function definition is wider than those reports.

To Reproduce

from __future__ import annotations

from collections.abc import Mapping
from typing import Any, cast, TYPE_CHECKING


def test(one: int | None = None, **kwargs: Any) -> int | None:
    print(repr(kwargs))
    return one


if TYPE_CHECKING:
    reveal_type(test)
    reveal_type(test())
    reveal_type(test(test='one'))
    reveal_type(test(**{'test': 'one'}))

which outputs:

test.py:13: note: Revealed type is 'def (one: Union[builtins.int, None] =, **kwargs: Any) -> Union[builtins.int, None]'
test.py:14: note: Revealed type is 'Union[builtins.int, None]'
test.py:15: note: Revealed type is 'Union[builtins.int, None]'
test.py:16: error: Argument 1 to "test" has incompatible type "**Dict[str, str]"; expected "Optional[int]"  [arg-type]

Expected Behavior

There is no issue passing kwargs via a dict and you shouldn't need to cast to a Mapping type.

Actual Behavior

There is an error message, but reveal_type(test(**cast(Mapping[str, str], {'test': 'one'}))) (or saving as a temporary variable with Mapping as its type) will allow the code to succeed.

Your Environment

• Mypy version used: 0.800
• Mypy command-line flags: --show-error-codes
• Mypy configuration options from mypy.ini (and other config files):
• Python version used: 3.8