Module random 

🔬This is a nightly-only experimental API. (random #130703)

Random value generation.

This module provides two low-level interfaces for random number generation:

• The Rng trait abstracts over all random number generators (RNGs) and is intended to be used in all cases where the choice of RNG is left to the user. It provides the fill_bytes method that fills a buffer of u8s with freshly-generated random data.
• The SystemRng implements the Rng trait by asking the operating system for cryptographically-secure random data on every call to fill_bytes.

In the future, higher-level interfaces for features like sampling distributions may be added to this module. Until that time, users of fill_bytes should take care to avoid sampling bias when using the filled byte buffer to create an instance of another type. In particular, the modulo operation is not suitable for constraining the range of an unconstrained number:

let mut buf = [0; 2];
rng.fill_bytes(&mut buf);
// 💀 **DO NOT DO THIS** 💀
// Numbers below ca. 22000 will be twice as likely.
let very_bad_random_number = u16::from_ne_bytes(buf) % 45000;

Examples

Generating a version 4/variant 1 UUID represented as text:

#![feature(random)]

use std::random::{Rng, SystemRng};

fn uuid(rng: &mut impl Rng) -> String {
    let mut buf = [0; 16];
    rng.fill_bytes(&mut buf);
    // Use little-endian to make the result reproducible across architectures.
    let bits = u128::from_le_bytes(buf);
    let g1 = (bits >> 96) as u32;
    let g2 = (bits >> 80) as u16;
    let g3 = (0x4000 | (bits >> 64) & 0x0fff) as u16;
    let g4 = (0x8000 | (bits >> 48) & 0x3fff) as u16;
    let g5 = (bits & 0xffffffffffff) as u64;
    format!("{g1:08x}-{g2:04x}-{g3:04x}-{g4:04x}-{g5:012x}")
}

println!("{}", uuid(&mut SystemRng));

Structs

SystemRngExperimental

The system random number generator.

Traits

DistributionExperimental

A trait representing a distribution of random values for a type.

RngExperimental

A source of randomness.

Functions

randomExperimental

Generates a random value from a distribution, using the default random source.