mul_dir

mul_dir(src: V, m: M)

A faster alternative to multiply a direction by a 4x4 matrix, ignoring the homogeneous normalization.