Update to Windows-sys v0.59 #579
Conversation
| @@ -938,9 +938,11 @@ pub(crate) fn unix_sockaddr(path: &Path) -> io::Result<SockAddr> { | |||
| } | |||
|
|
|||
| storage.sun_family = crate::sys::AF_UNIX as sa_family_t; | |||
| // SAFETY: casting `[u8]` to `[i8]` is safe. | |||
| let b = unsafe { &*(bytes as *const [u8] as *const [i8]) }; | |||
There was a problem hiding this comment.
Seems a bit fragile to rely on char being i8.
There was a problem hiding this comment.
Missed this comment, will send a pr to change it to char
There was a problem hiding this comment.
Oh, the type itself actually uses i8, not a type alias: https://docs.rs/windows-sys/latest/windows_sys/Win32/Networking/WinSock/struct.SOCKADDR_UN.html.
|
This PR is intended as a successor to #545, but I have concerns about not using a range version like I had authored there: #545 (comment) It seems unnecessarily hostile to hard-upgrade when not necessary for users who are already trying to tackle the big migrations the |
|
Cargo's support for range dependencies is not good. I do not recommend attempting to use it. |
|
@ErichDonGubler I agree with @Darksonn here. In addition I also don't want to support multiple window-sys versions with breaking changes in between, the instability of the latest version is enough of a maintenance burden as it is. |
No description provided.