Looking over the PostScript parser I mentioned this morning, it appears be a handful of mostly self-contained textual parsers resembling little more than a lexer. There's several of these for different PostScript constructs, or loops around them.
It defers to a seperate file to parse literal values.
P.S. It appears I've overlooked a whole CF2 Fonts system, so that's what I'll be studying tomorrow! Looks to be interesting...