hi Kassen,
actually, you were right. this is a problem, since we have changed the vector math functions (vadd, vmul, etc) to work with various vector lengths. vnormalize should also work like that if it's not slower than the c version.accidentally mixed that up with the vector length sense of the word, which is of course a lot more cpu-heavy and a lot less useful beyond length 3 or 4.
g.
