Implement TryFrom<f64> for f32

[clarfonthey]

This will only perform the conversion if it can be done without any loss of precision.

[rust-highfive]

r? @TimNN

(rust_highfive has picked a reviewer for you, use r? to override)

[scottmcm]

Previous conversations about TryFrom<float> for integers: #47857

[kennytm]

Please add test cases for these.

Also, this would make e.g. assert_eq!(f32::try_from(0.1_f64), None), which may be too strict? I'd expect it return None only if overflow occurs. Perhaps TryFrom isn't the proper way to introduce this operation, but rather separate methods allowing user to specify what to do with overflow, underflow, inexact exceptions and rounding direction (IEEE-754-2008 §4.3).

[clarfonthey]

I'll take a look at these in a bit. You're right on the inexactness, although it's the same if you try f32::try_from(f64::consts::pi). You are fundamentally losing accuracy there.

[hanna-kruppe]

See https://internals.rust-lang.org/t/how-should-we-provide-fallible-float-to-int-conversions/6708/1

[pietroalbini]

Ping from triage @clarcharr! It's been a while since we heard from you, will you have time to work on this again soon?

[clarfonthey]

@pietroalbini I haven't had time to look at this in the past few days, although I'm probably fine closing this for now. I'll open a new PR with tests once I get the time.

[pietroalbini]

Sure.