We need to take the token regexes, build a state machine out of them, and write the code for the lexer based on the state machine