Bug report

Bug description:

A few tests for typing.get_type_hints() test behavior that I feel is incorrect. Fixing it may be more trouble than it's worth, but I thought I'd record these.

1. test_get_type_hints_generic

This tests that the a key for the TypedDict is T | None, where T is the TypeVar of that name defined in test_typing.py. However, the relevant TypedDict field is defined in a separate module _typed_dict_helper that has its own T TypeVar. get_type_hints() should pick up the T from the helper module, not the one from test_typing.

This also affects another assertion in the same function.

2. test_get_type_hints_classes

This calls get_type_hints() on a class defined like this:

class B(Generic[T]):
    class A(Generic[T]):
        pass

    my_inner_a1: 'B.A'
    my_inner_a2: A
    my_outer_a: 'A'  # unless somebody calls get_type_hints with localns=B.__dict__

It checks that the unquoted A refers to the nested class and the quoted A refers to a different A class in the outer scope. But the fact that the name was quoted shouldn't influence scoping logic; it should resolve to the same name whether or not it's quoted.

I have a change that fixes both of these tests, but I'm not convinced we can afford to change this.

CPython versions tested on:

CPython main branch

Operating systems tested on:

No response

Linked PRs

• gh-137228: Improve get_type_hints() #137230