* use warnings module to ... raise warnings!

Make sure the output dtype of uniform is the one given as dtype parameter, and that normal Python int/float do not upcast it

That way, we use the CudaNdarray, instead of transferring back and forth
to the GPU, and the Gpu ops have CudaNdarrays as inputs, not numpy
ndarrays or scalars.

This removes the need for a special 'ndarray' case, and should generalize to
Python and numpy numeric types, as well as CudaNdarray, and even other types.

Also test it.
This should fix the problem experienced by Justin Bayer.