Cmplx division by 0 and NaN

[huonw]

Should 1/(0+0i) be NaN+NaN i or fail, or something else.

see also #1284

[catamorphism]

Nominating for milestone 2, backwards-compatible

[catamorphism]

Visiting for triage.

[graydon]

Worth implementing correctly, but if anyone depends on its present behavior they're depending on a bug.

[graydon]

(also in libextra, not slated for 1.0-timeline stability commitment)

[graydon]

just a bug, removing milestone/nomination.

[treeman]

No longer relevant.

Explicit test for failure message with #16475.

[treeman]

My mistake. Still relevant.

For Complex32 and Complex64 still outputs -NaN+-NaNi. Testcase:

extern crate num;
use num::complex::{Complex32, Complex64, Complex};

fn main() {
    let _0_0i: Complex64 = Complex { re: 0.0, im: 0.0 };
    let _1_0i: Complex64 = Complex { re: 1.0, im: 0.0 };

    println!("{}", _1_0i / _0_0i);
}

While Complex<int> fails.

Do we want NaN+NaN i or do we want it to fail?

[gifnksm]

libnum crate moved to rust-lang/num repository.
@nick29581 Could you clone this issue to num repository?

[rust-highfive]

This issue has been moved to the RFCs repo: rust-num/num#20