pub fn float_max(a: f64, b: f64) -> f64
The language’s max of two reals: a NaN gives way to the other operand, and of two equal values the one with the positive sign, so max(0, -0.0) is 0.0 (IEEE 754-2019 maximumNumber).
max
max(0, -0.0)
0.0
maximumNumber