softfloat 1.0.0

Pure software floating-point implementation, with `const` and `no_std` support
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21

use super::F32;

/// Sign of Y, magnitude of X (F32)
///
/// Constructs a number with the magnitude (absolute value) of its
/// first argument, `x`, and the sign of its second argument, `y`.
pub(crate) const fn copysign(x: F32, y: F32) -> F32 {
    let mut ux = x.to_bits();
    let uy = y.to_bits();
    ux &= 0x7fffffff;
    ux |= uy & 0x80000000;
    F32::from_bits(ux)
}

#[cfg(test)]
mod test {
    #[test]
    fn sanity_check() {
        assert_eq!(f32!(1.0).copysign(f32!(-0.0)), f32!(-1.0))
    }
}