[raymath] Fix: QuaternionMultiply

[0stamina]

the multiplication algorithm was slightly off
previously when combining quaternion rotations objects would shrink now they rotate as expected

trying to get around the issue by using QuaternionToMatrix then QuaternionTransform doesn't work, and I'm not sure why

[raysan5]

@0stamina This is a sensible change, used by raylib animation system, did you verify it does not break animations rotations?

[0stamina]

turns out the multiplications for animations were swapped to account for the mistake in the formula, I can't find any other instance of using quaternion multiplication to check

[raysan5]

@0stamina I'm afraid I need some time to review this sensible change. Sending it without checking the implications in raylib from your side is very concerning and makes me doubt.

[casperbear]

AI doesn't recommend to merge the fix and suggests to check for problems on user side, in particular 1) whether a Left-Handed coordinate system is used in a Right-Handed library (RayLib/OpenGL). 2) whether arguments are passed in incorrect order to QuaternionMultiply - new/local rotation should be passed first, base/world rotation second