regexp-to-DFA algorithm

I'm writing a regexp-based lexical analyzer generator based on a regexp-to-DFA direct translation algorithm described in the Dragon Book.

It calculates the nullable, firstpos, lastpos, and followpos functions for concatenation, alternation, and kleene-star nodes. I want to add quantifiers for + (one or more times) and ? (zero or one time).

nullable开发者_JAVA技巧, firstpos, and lastpos are easy to calculate, but I'm not sure about followpos. Is it better to not implement those quantifiers as part of the parsing stage, and instead rewrite it in the lexical analysis stage (I think this is only "syntactic sugar" to kleene-star and alternation)?