The parser currently does extra loops 'n stuff to try to parse whitespace+comments. What if we could do a quick check before parsing and tell the parser if it could skip that (potentially expensive) step? Not sure if there's any positive impact, but it's worth exploring.