Tracking issue for the convert::FloatToInt trait

[SimonSapin]

PR #66841 adds this supporting trait for the {f32,f64}::approx_unchecked_to method (tracking issue: #67058) to the core::convert and std::convert modules:

mod private {
    pub trait Sealed {}
}
pub trait FloatToInt<Int>: private::Sealed + Sized {
    #[doc(hidden)]
    unsafe fn approx_unchecked(self) -> Int;
}

// Macro-generated for all combinations of primitive integer
// and primitive floating point types:
impl FloatToInt<$Int> for $Float {…}

The name and location (and usage of the sealed-trait pattern) are chosen to allow later adding additional methods with different conversion semantics between those types, as proposed in https://internals.rust-lang.org/t/pre-rfc-add-explicitly-named-numeric-conversion-apis/11395

[dodomorandi]

What's the current state of the art? Is it just a matter of low priority respect to other things?

And even if it is not the best place to ask: this will be the foundation for one (or more) RFC(s) to expose safe APIs like in the pre-RFC?

[SimonSapin]

The to_int_unchecked method of f32 and f64 were previously tracked at #67058 and are stable since 1.44. Technically only their supporting trait is tracked here.

For the wider idea of adding several conversion methods, what’s needed now is someone to champion it. Polish https://internals.rust-lang.org/t/pre-rfc-add-explicitly-named-numeric-conversion-apis/11395 (especially around prior art and alternatives) and formally propose and RFC, drive the discussion, etc. I’ve been meaning to since drafting it but you know, life happened. It doesn’t need to be me though ;)