💡 [REQUEST] - Better random for fuzzing

[SteMak]

Summary

Trident uses fair random generator which is not a good solution for fuzzing toolkit

For example, the following assert did not reverted for over 100B executions

#[flow]
fn flow1(&mut self) {
    let a = self.trident.random_from_range(0..u64::MAX) as u128;
    let b = self.trident.random_from_range(0..u64::MAX) as u128;
    let c = self.trident.random_from_range(0..u64::MAX) as u128;

    assert!(a * b > c);
}

Basic Example

Unfair random based on various byte sequences can be used

// Funny sample of unfair random breaking the assert above
fn unfair_u64(&mut self) -> u64 {
    let pow = self.trident.random_from_range(7..71);
    let divisor = self.trident.random_from_range(131..247);
    return (2u128.pow(pow) / divisor) as u64;
}

Drawbacks

Good unfair random is a way more complex than the fair one, it is usually statefull which takes additional computation resources

Unresolved questions

No response

Implementation PR

No response

Reference Issues

No response