Compute the hypotenuse avoiding overflow and underflow (single-precision). - View it on GitHub