I have played around with the regular expression support in NPP 2.9, and like it very much, especially the (recently added) backreference support.
I have, however, found a few missing features, and since I cannot find any documentation for the regular expression support, I'll put them here:
1. Support for shorthand character classes such as \w, \d, \s and the negated \W, \D, \S. E.g. \w being synonymous with [a-zA-Z0-9_]. Saves a LOT of typing ;-). See http://www.regular-expressions.info/charclass.html
2. Support for the optional "?" operator to denote 0 or 1 matches of the preceding regexp. Example "</?([a-zA-Z0-9]+)>". See http://www.regular-expressions.info/optional.html
3. Support for multi-line operation (the ability to have a regular expression that spans multiple text lines). Possibly implemented via a "multi-line" checkbox.See http://www.regular-expressions.info/modifiers.html
3.1. Support for anchors, the ^ and $ (beginning- and end-of-line) and \b (word boundary). See http://www.regular-expressions.info/wordboundaries.html
4. Support for greedy/lazy matching via the the *? and +? operator. See http://www.regular-expressions.info/repeat.html
In any case, it would be nice to have some indication of which regular expression syntaxes are supported in NPP, since trial-and-error is a difficult way to discover feature levels ;-)
A good regexp syntax reference, for the various language implementations may be found at http://www.greenend.org.uk/rjk/2002/06/regexp.html
I have copied the core of this posting as a formal feature request (where it really belongs) http://sourceforge.net/tracker/?atid=612385&group_id=95717&func=browse as Request ID = 1185698
By the way... I noticed that NPP 2.9 actually do support the ^ and $ anchors (beginning and end of line) http://www.regular-expressions.info/anchors.html
I also tried positive and negative lookahead, which doesn't seem to work, either.