use sort_unstable if you're deduplicating anyways

[miguelraz]

Objective

• To make bevy_math faster 🚀

Solution

• Unstable sorts are potentially destructive, but if you're deduplicating anyways, there's no risk of that, so you might as well use a faster version.

It also helps that it makes less allocations than the normal .sort().

The changes to bevy_ecs's unit tests and examples/ui/text aren't on any critical path, so most of the gains (if any) are for bevy_math.

Testing

• Did you test these changes? If so, how?

I did cargo test -p bevy_math and got all tests passing:

test result: ok. 117 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 1.06s

Likewise for cargo test -p bevy_ecs:

test result: ok. 478 passed; 0 failed; 3 ignored; 0 measured; 0 filtered out; finished in 163.98s

• Are there any parts that need more testing?

No.

• How can other people (reviewers) test your changes? Is there anything specific they need to know?

Nope.

• If relevant, what platforms did you test these changes on, and are there any important ones you can't test?

Windows 11, not important.

---

Showcase

I'd be happy to add a local criterion benchmark if someone points me towards a small representative example to do the before/after comparison - this is my first Bevy PR so I don't know what would be a decent stress test for this code path.

[github-actions]

Welcome, new contributor!

Please make sure you've read our contributing guide, as well as our policy regarding AI usage, and we look forward to reviewing your pull request shortly ✨

[mardzie]

Code looks good to me and the reasoning holds up. I am not familiar with the deeper context though.

[SpecificProtagonist]

As none of these floats can be NaN, this is correct.