Remove fragile dependency constraint on ordered-float.

[kpreid]

Connections
The problematic line was introduced by #6847 and last altered by #7618.

Description
Bounds like <=5.0 should never be used unless necessary to avoid bugs, because it will cause a build failure if another package has a requirement like ~5.1.0 or >=5.1.0. Upper bounds should always cut at a major version boundary, not before it.

This PR replaces the <=5.0 bound with a well-behaved <6.0.

Testing
xtask test

Squash or Rebase?
Rebase

Checklist

• Run cargo fmt.
• Run taplo format.
• Run cargo clippy --tests. If applicable, add:
  • --target wasm32-unknown-unknown
• Run cargo xtask test to run tests.
• If this contains user-facing changes, add a CHANGELOG.md entry.

[ErichDonGubler]

I'm greatly surprised by Cargo's behavior here. Does using 5.* instead of 5.0 not resolve the issue?

[kpreid]

Does using 5.* instead of 5.0 not resolve the issue?

It seems like <=5.* would in fact do so, but the meaning of that syntax is not documented, and I have never seen anyone else using it.

[kpreid]

I filed rust-lang/cargo#16128 to ask for clarification about what <=5.* means.

[ErichDonGubler]

@kpreid: I have observed it in the wild, but I don't remember which crate it's happened with. I know it was part of doing supporting ecosystem work necessary to support Mozilla's consumption of #6876, but 🤷🏻‍♂️ I'd have to go hunting with that knowledge in hand to find it again, I think.

[a1phyr]

I think that <=5 would have worked too.

Anyway, I think that this would be nice to backport to v27.

[MarijnS95]

Came here through the back-link of #6876, reminiscent of a previous PR by @kpreid at Traverse-Research/gpu-allocator#253 where we identified that the <=5 bound proposed by @a1phyr would have worked too, which is strictly different from <=5.0 and doesn't look nearly as confusing as <=5.*.