build: allow windows-core 0.59, windows-targets 0.53

[ErichDonGubler]

Things appear to compile and test fine, so 🫡 woot!

[ErichDonGubler]

Mozilla is hoping to eventually consume this as part of a transition to windows 0.59, as a transitive dependency of neqo and wgpu. I'm hoping a release will follow soon after this gets merged.

Please don't read this as a demand, though; the timeline on a need to consume these changes isn't urgent. Thank you so much for the work you've put into this crate! 🙏🏻 It's amazing how we've been able to use it to abstract over static and dynamic linking.

[mxinden]

Wouldn't this be a better PR title?

- build: allow windows-core 0.59, windows-targets 0.53
+ build: allow windows-sys 0.60, windows-targets 0.54

[ErichDonGubler]

@mxinden: No, the title is correct, but the misunderstanding is (kek) understandable, because those are the numbers in the new version range. Note, however, that the range is exclusive; pasting from the diff., observe that <0.x is always the upper bound:

[target.'cfg(windows)'.dependencies.windows-targets]
version = ">=0.48, <0.54"

[target.'cfg(windows)'.dev-dependencies.windows-sys]
version = ">=0.52,<0.60"

[nagisa]

Thank you, this looks good to me, will make a release in a couple minutes.

[nagisa]

Oh well, um, so this PR has revealed that the CI here was broken for a couple past commits and then further that there is something very wrong with the windows-gnu target which will require more investigation before I can fix it. I'll merge the GHA fix now, so note that the failure that will show up for this PR afterwards is unrelated to your fix itself.

[ErichDonGubler]

@nagisa: I'm happy to volunteer more time to see if I can debug it. 😀 Did you want the help?

[nagisa]

If you have time/motivation, go for it. I don't anticipate setting up a Windows machine/VM anytime very soon. At the same time I don't think landing this PR is blocked on those checks passing, only the release of a new version of the crate is.