This came up in ManimCommunity/manim#4398 (comment).

I believe the class definition class B(A): ... should not be allowed by pyright (especially in strict mode), because it can lead to unsoundness, as demostrated in the example below. In fact, mypy currently rejects this (mypy playground)

Describe the bug
A clear and concise description of the behavior you are seeing and the expected behavior along with steps to reproduce it.

Code or Screenshots
Code sample in pyright playground

class A:
    def foo(self, /, **kwargs: object) -> None: ...

class B(A):
    def foo(self, /, *, x: int = 3, **kwargs: object) -> None:
        assert isinstance(x, int)   # of course, right?

def use_a(a: A):
    a.foo(x="hello")  # ok because "hello": str <: object

use_a(A())  # all fine
use_a(B())  # woops, AssertionError!

VS Code extension or command-line
Observed in the Pyright playground (online) and verified its run-time behaviour with Python 3.12.11 (macOS arm64)