pub fn sincosdx(x: f64, sinx: &mut f64, cosx: &mut f64)
In order to minimize round-off errors, this function exactly reduces the argument to the range [-45, 45] before converting it to radians.