ENH: Allow ufuncs to request the iterator to provide contiguous arrays #30413

Proposed new feature or change:

In #29699, a ufunc loop for sin and cos is added that, if the input array is not contiguous, makes a copy to a contiguous buffer so that the result can be calculated using a vectorized SIMD loop. This makes a lot of sense, but as discussed there, it seems a bit strange to do this in the ufunc loop when the iterator is already able to do exactly this for non-aligned data or dtype mismatches. Indeed, it has the NPY_ITER_CONTIG flag for this purpose, so it would seem to be only a matter of allowing the ufunc/array method to specific that this should be used, either by a flag of its own or perhaps more simply, by only providing the contiguous loop in an array method.