build: update windows-sys to 0.59.0

[ErichDonGubler]

Carries on the effort from #1820, with some fixes that should make tests compile now.

[Darksonn]

Which field in Inner is making this necessary?

[Darksonn]

It may make more sense to implement these traits for Inner instead, or to wrap the offending field in a newtype and implement them on that newtype.

[ErichDonGubler]

Which field in Inner is making this necessary?

We need to implement Send and Sync manually because HANDLE is no longer considered thread-safe. I think your instinct is right, and we need to capture the overlapped I/O contract with types ourselves now.

Will take some time to dig into that contract, and report what I learn here.

[ErichDonGubler]

It seems much easier to reason about the specific APIs' usage of Handles, rather than a category of resource backing a Handle, for the purpose of justifying unsafe impl {Send,Sync}. I'll implement Send and Sync directly on iocp::CompletionPort and named_pipe::Inner first.

If there's desire to attempt categorizing types of resources behind Handles as thread-safe, maybe a follow-up PR would be best?

[Darksonn]

There's a build failure.

     Checking mio v1.0.3 (D:\a\mio\mio)
  error[E0425]: cannot find function `null_mut` in this scope
    --> src\sys\windows\afd.rs:69:13
     |
  69 |             null_mut(),
     |             ^^^^^^^^ not found in this scope
     |
  help: consider importing this function
     |
  1  + use std::ptr::null_mut;
     |

[ErichDonGubler]

@Darksonn:

There's a build failure.

     Checking mio v1.0.3 (D:\a\mio\mio)
  error[E0425]: cannot find function `null_mut` in this scope
    --> src\sys\windows\afd.rs:69:13
     |
  69 |             null_mut(),
     |             ^^^^^^^^ not found in this scope
     |
  help: consider importing this function
     |
  1  + use std::ptr::null_mut;
     |

Ah, that was an interesting feature combination that I had to use to reproduce locally. 😅 Glad CI caught it!

[Snowiiii]

Any progress here ?

[ErichDonGubler]

@Snowiiii: None, until you'd asked, and I'd refreshed my memory on this PR. 😀

I've updated the contents of the PR and its conversations with my latest thinking. I think this is ready for a re-review from @Darksonn.