Comment 32 for bug 1406458

True, but ONLY for those seeds that had problems. Given that the blanks were a pretty severe artifact, I would not consider this a huge problem. Also, although ideally we want to be as backwards compatible as possible, the overall appearance is determined more by the other parameters than by the seed. If you just vary the seed you basically get different instantiations of the same kind of noise. Still, it is fairly easy to modify the code so that this does not happen (basically the idea would be to repeatedly loop over the vectors to fill in any remaining zeros, instead of having the do-loop per vector, ideally this should be done AFTER "shuffling lattice selectors", to avoid influencing those as well).