Problem with @overload and Iterator/Generator

See last line (Playground):

from collections.abc import Iterator, Generator
from typing import Any, Iterable, TypeVar, overload
from typing_extensions import assert_type

T = TypeVar("T")
IterableT = TypeVar("IterableT", bound=Iterable)


@overload
def read_iterator(iterable: Iterator[T]) -> list[T]: ...


@overload
def read_iterator(iterable: IterableT) -> IterableT: ...


def read_iterator(iterable: Iterable) -> Iterable:
    if isinstance(iterable, Iterator):
        return list(iterable)
    else:
        return iterable


def _create_generator() -> Generator[int, Any, None]:
    for i in range(3):
        yield i


def test_read_iterator() -> None:
    an_iterator = iter(range(3))
    assert_type(read_iterator(an_iterator), list[int])

    a_range, a_list, a_tuple, a_set, a_dict = range(3), [1, 2, 3], (1, 2, 3), {1, 2, 3}, {"a": 1}
    assert_type(read_iterator(a_range), range)
    assert_type(read_iterator(a_list), list[int])
    assert_type(read_iterator(a_tuple), tuple[int, int, int])
    assert_type(read_iterator(a_set), set[int])
    assert_type(read_iterator(a_dict), dict[str, int])
    
    a_generator = _create_generator()
    assert_type(a_generator, Generator[int, Any, None])
    assert_type(read_iterator(a_generator), list[int])  # error: Expression is of type "Any", not "list[int]"  [assert-type]

Mypy doesn't like this code:

error: Expression is of type "Any", not "list[int]"  [assert-type]

Please remember that Generator is a subtype of Iterator:

>>> isinstance(a_generator, Iterator)
True

I think this is a bug in Mypy. The Generator should match the signature of def read_iterator(iterable: Iterator[T]) -> list[T]

Versions:

> python -V             
Python 3.11.1

> mypy --version
mypy 1.11.2 (compiled: yes)

❯ system_profiler SPSoftwareDataType
Software:

    System Software Overview:

      System Version: macOS 13.6.7 (22G720)
      Kernel Version: Darwin 22.6.0

Related: https://github.com/python/typing/issues/253

Start by reproducing the overload-resolution case from the linked mypy-play Playground, focusing on how Generator and Iterator are matched and how the TypeVar return type is inferred. Done means the final assert_type(read_iterator(a_generator), list[int]) passes without the reported error, while the other assertions remain valid.