Tracking Issue for `erf` and `erfc` for `f16, f32, f64, f128`

[tgross35]

Feature gate: #![feature(float_erf)]

This is a tracking issue for the error function (erf) and complementary error function (erfc) in std. These functions are provided by libm and are also exposed in C's math.h (reference).

Public API

// in std, for now

impl f16 {
    pub fn erf(self) -> Self;
    pub fn erfc(self)-> Self;
}

impl f32 {
    pub fn erf(self) -> Self;
    pub fn erfc(self)-> Self;
}

impl f64 {
    pub fn erf(self) -> Self;
    pub fn erfc(self)-> Self;
}

impl f128 {
    pub fn erf(self) -> Self;
    pub fn erfc(self)-> Self;
}   

Steps / History

• ACP: Add erf and erfc to f32 and f64 libs-team#89
• Implementation: Implement f{16,32,64,128}::{erf,erfc} (#![feature(float_erf)]) #136324
• Final comment period (FCP)¹
• Stabilization PR

Unresolved Questions

• None yet.

Footnotes

1. https://std-dev-guide.rust-lang.org/feature-lifecycle/stabilization.html ↩

[tgross35]

If anyone is interested in adding these, feel free to do so. The extern functions need to be added at https://github.com/rust-lang/rust/blob/6ac88786155774da429213a1362d4cb085047af3/library/std/src/sys/cmath.rs, then the API at https://github.com/rust-lang/rust/blob/6ac88786155774da429213a1362d4cb085047af3/library/std/src/f64.rs (and other float files) will call them, similar to gamma.

@rustbot label +E-easy

[GrigorenkoPV]

The signatures look very wrong. I guess the functions in impl f64 should take an f64 and not an f32, etc. Those just look like a typo.

A less trivial question is, shouldn't they all take self instead of x: Self? To be consistent with gamma and the rest.

Anyways, don't mind if I do,
@rustbot claim

[tgross35]

Oh woah yeah, I was looking at math.h and confused myself on the signatures. Updated.

[kennytm]

As I mentioned in the other ACP (rust-lang/libs-team#352), stabilization of {f32, f64}::{erf, erfc} should also consider the status of {f32, f64}::{gamma, ln_gamma} #99842. Both features are going to link to libm (or contains of copy of its implementation) and thus subject to #26350.

[foxtrotactive]

Hi, could I take a stab at this? Would be happy to help out where needed

[tgross35]

@GrigorenkoPV already has a PR up at #136324 (commenting @rustbot claim assigns the PR to you) so there won't be much left to do here. But thank you for the interest! If you are looking to get involved, looking at issues labeled E-easy and/or E-help-wanted that don't have anyone assigned is the best way to find something to do.

[foxtrotactive]

Great thanks a lot @tgross35