Module computation on float

[denis-migdal]

Hi,

Currently, float modulo is computed as:

if($B.$isinstance(other, _b_.int)){
        other = _b_.int.numerator(other)
        return fast_float((self.value % other + other) % other)
    }

    if($B.$isinstance(other, float)){
        // use truncated division
        // cf https://en.wikipedia.org/wiki/Modulo_operation
        var q = Math.floor(self.value / other.value),
            r = self.value - other.value * q
        if(r == 0 && other.value < 0){
            return fast_float(-0)
        }
        return fast_float(r)
    }

Is there a reason to not simply use the following formula ?

const mod = (a % b + b) % b; // the formula used everywhere in the code.
if( mod === 0 && b < 0 )         // handle -0.
    return -0;
return mod;