What are the most recent developments in DFA/NFA-based regex parsing (that&#039;s about finite automata as such, not about so-called `NFA engines&#039;)? There is a classic solution for the base problem; a limited support for lookaheads in lex; I know of a submatch capturing method using FA. Any other features? Lookaheads, greediness, backrefs etc.?<BR><BR>BTW I&#039;m mostly interested in theory of supported features rather than implementations, though links to the latter are welcome, too.