This is, and I can't stress this enough, a lot better than all the manual bounds checking and indexing that was going on before. Also fixes a small bug where "\u{}" wouldn't get rejected as invalid unicode escape sequence.