email: improve bytes handling #9032
Conversation
This comment has been minimized.
This comment has been minimized.
stdlib/email/base64mime.pyi
Outdated
| def decode(string: str | bytes) -> bytes: ... | ||
| from _typeshed import ReadableBuffer | ||
|
|
||
| def header_length(bytearray: bytes | bytearray) -> int: ... |
There was a problem hiding this comment.
str would work as well, and since header_encode() explicitly checks for str, I think it's safer to allow it here as well. In fact, the tests use str:
| def header_length(bytearray: bytes | bytearray) -> int: ... | |
| def header_length(bytearray: str | bytes | bytearray) -> int: ... |
There was a problem hiding this comment.
My thinking was that this function is about counting bytes using len(), and that doesn't make sense for str because a character can be multiple bytes. I guess it works for ASCII strings, though.
There was a problem hiding this comment.
At first I was not going to comment this for that reason, but seeing the test using str tipped me to comment on this.
| @@ -12,7 +12,7 @@ __all__ = ["Message", "EmailMessage"] | |||
|
|
|||
| _T = TypeVar("_T") | |||
|
|
|||
| _PayloadType: TypeAlias = list[Message] | str | bytes | |||
| _PayloadType: TypeAlias = list[Message] | str | bytes | bytearray | |||
There was a problem hiding this comment.
This change means that get_payload() can suddenly return bytearray, which contradicts the documentation. set_payload() is accepting a bytearray, though (and converting it to a str).
There was a problem hiding this comment.
get_payload() returns Any in our stub, _PayloadType is only mentioned in a comment. I think the comment is still close enough, we can revisit it if we ever change the Any.
There was a problem hiding this comment.
For some reason I totally missed the comment in front of the return type ...
|
According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉 |
No description provided.