min-max: max! and min! macros for Rust

Why?

Sometimes you want to find the maximum of a bunch of scalars. Usually you would write something like max(x1, max(x2, max(x3, x4))). The max! macro provided by this crate simplifies that to max!(x1, x2, x3, x4). (Note, for an iterable data structure, you would use xx.iter().max()).

Usage

Add this to your Cargo.toml:

min-max = "0.1"

Then, for example:

use min_max::*;

fn main() {
    let max = max!(1, 5, 7, 2, 4, 9, 3);
    assert_eq!(max, 9);
    let min = min!(1, 5, 7, 2, 4, 9, 3);
    assert_eq!(min, 1);
    let min_max = min_max!(1, 5, 7, 2, 4, 9, 3);
    assert_eq!(min_max, (min, max));
}

Does it work on floats?

Yep. But you need to use max_partial!/min_partial!

use min_max::*;

fn main() {
    let partial_max = max_partial!(1.8f64, 5.8, 7.8, 2.8, 4.8, 9.8, 3.8);
    assert!((9.8 - partial_max).abs() < 1e-5);
    let partial_min = min_partial!(1.8f64, 5.8, 7.8, 2.8, 4.8, 9.8, 3.8);
    assert!((1.8 - partial_min).abs() < 1e-5);
}

What about NaN?

Do not use when your data contains NaN. When NaN is at the end, NaN is returned. Otherwise, the min/max excluding NaN is returned.

use min_max::*;

fn main() {
    let partial_max = max_partial!(1.8, 5.8, f64::NAN, 2.8, 4.8, 9.8, 3.8);
    assert!((9.8 - partial_max).abs() < 1e-5);
    let partial_max = max_partial!(1.8, 5.8, 2.8, 4.8, 9.8, 3.8, f64::NAN);
    assert!(partial_max.is_nan());
    let partial_min = min_partial!(1.8, 5.8, f64::NAN, 2.8, 4.8, 9.8, 3.8);
    assert!((1.8 - partial_min).abs() < 1e-5);
    let partial_min = max_partial!(1.8, 5.8, 2.8, 4.8, 9.8, 3.8, f64::NAN);
    assert!(partial_min.is_nan());
}

Can I use custom types?

Sure, why not?

use min_max::*;

#[derive(Debug, Ord, PartialOrd, Eq, PartialEq, Clone, Copy)]
struct Point {
    x: u16,
    y: u16,
}

fn main() {
    let a = Point { x: 5, y: 8 };
    let b = Point { x: 10, y: 92 };
    let c = Point { x: 0, y: 3 };
    let max = max!(a, b, c);
    assert_eq!(max, b);
}

What's going on under the hood?

Well, max!(x1, x2, x3) expands to:

std::cmp::max(x1, std::cmp::max(x2, std::cmp::max(x3)))

and so on. min! works similarly, but with std::cmp::min.

min_partial! and max_partial uses the min and max functions from the partial-min-max crate.