* Vector2i and Vector3i mul/div by a float results in Vector2 and Vector3 respectively.
* Create specializations to allow proper bindings.
This fixes#44408 and supersedes #44441 and keeps the same rule of int <op> float returnig float, like with scalars.