Resolve two FIXME notes in Array::sort routines. FIXME note in sort routines wondered whether there was a performance impact from testing for NaN values on integer types (which cannot have IEEE NaN values). Performance testing reveals there is no slowdown to using the function, presumably because library routines have a very quick shortcut to eject integers from consideration. * Array.cc (sort (dim, mode)): Add comment explaining special case for "stride == 1". Add note for future programmers that no special case is needed for integers to avoid sort_isnan() function call. Remove FIXME notes about impact on integers. * Array.cc (sort (sidx, dim, mode)): Add comment for "stride == 1" pointing to further explanation in Array::sort (dim,mode). Remove FIXME notes about impact on integers.

# line endings in repository match line endings on disc
* -text