Stub for ossaudiodev #4944
Stub for ossaudiodev #4944
Conversation
stdlib/3/ossaudiodev.pyi
Outdated
| # TODO: oss_audio_device return type | ||
| @overload | ||
| def open(mode: str) -> Any: ... | ||
| @overload | ||
| def open(device: str, mode: str) -> Any: ... |
There was a problem hiding this comment.
mode is Literal["r", "w", "rw"], according to the documentation.
There was a problem hiding this comment.
I was wondering about this — wouldn't we need to have the broader str signature anyway, so that we fallback correctly? https://www.python.org/dev/peps/pep-0586/#interactions-with-overloads
There was a problem hiding this comment.
The situation described in the PEP is a bit different than here, I think. The PEP example has an overload, where other types depend on the literal value. Here, no other type depends on the literal. Also, the example in the PEP would be better written as mode: Literal["r", "rb"] = pick_file_mode(...), as this would be better than just returning str from pick_file_mode().
Off-topic aside: AnyOf could also help in the PEP example: Type checkers could treat a function foo with overloadsdef foo(x: Literal["b"]) -> bytes: ... and def foo(x: Literal["s"]) -> str: ... as also accepting def foo(x: Literal["s", "b"]) -> AnyOf[str, bytes]: ....
There was a problem hiding this comment.
Thanks, made the change!
One of the few remaining un-stubbed modules in the stdlib.
Note that this wasn't mentioned in #4545 since that list of
modules was determined on
sys.platform == 'darwin'