Finish-up UTF-8 handling

[mkilgore]

There's a few FIXMEs still left in the code around UTF-8 handling that need to be addressed. \u support is a big one, and also skipping UTF-8 characters when doing string lexing (though what we're doing might be sufficient, since individual UTF-8 bytes aren't ever valid ASCII values so we'd never mistake them for things like a double quote).