I think I've made it faster. topic/faster-type-notation-parser-regex - it still passes the t/040_type_constraints/014_type_notation_parser.t test, so it functionally appears correct. But maybe someone is interested in doing some profiling and proving it?
Cheers, Sam
