Conversation
| complex(np.bytes_("1")) # E: No overload variant | ||
|
|
There was a problem hiding this comment.
Minor regression due to python/typeshed#8473, i.e. any object that defines __float__ is now treated as valid by the complex constructor (even though this will raise during runtime as being a bytes subclass takes priority).
Nothing we can really do here I'm afraid.
| AR_void: npt.NDArray[np.void] | ||
|
|
||
| pointer: ctypes.pointer[Any] | ||
| pointer: ctypes._Pointer[Any] |
There was a problem hiding this comment.
Adapt to the ctypes.pointer changes in python/typeshed#8446.
| reveal_type(a.coords) # E: tuple[builtins.int, ...] | ||
| reveal_type(a.index) # E: int | ||
| reveal_type(iter(a)) # E: flatiter[ndarray[Any, dtype[str_]]] | ||
| reveal_type(iter(a)) # E: Any |
There was a problem hiding this comment.
A minor regression introduced by python/typeshed#7817.
Seems to be a mypy bug as pylance does produce the expected result here.
|
Mypy seems to use |
|
Thanks @BvB93 |
Bump mypy to version 0.981.
Marked as backport due as the test set required a few updates.