#14693 introduced the bevy_math::ops module, which allows bevy_math to avoid using the Rust standard floating point functions with unspecified precision. Based on comments there, it seems that making the module public to allow library authors downstream of bevy_math to do the same (for the sake of platform-agnostic determinism in floating point operations) would be desirable, so we should probably do that.

f64 versions of the operations could also be introduced in a PR that addresses this, but that part could also be left to a follow-up — on the other hand, the f64 versions are not useful internally in bevy_math, so they should probably be introduced once the module is already public.