From: Andy Heninger (andy.heninger@gmail.com)
Date: Tue Sep 25 2007 - 20:36:48 CDT
Just to add my perspective on this thread,
All of the discussions and possibilities for language aware Unicode regular
expression processing are fascinating, but they aren't implemented anywhere
(at least not publicly) yet, let alone being mature enough to push up as
some sort of a formal recommendation. POSIX regular expressions don't count
- they're not Unicode, and the locale sensitive features are pretty much a
failure (which we should take as a caution.)
What is very important to have right in UTS-18 are the things that
implementors of conventional, mainstream regular expressions really should
be doing for Unicode support. Recommendations for the Perls, PCREs and
Javas of the world, and they do pay at least some attention to what UTS-18
says.
The changes to the descriptions of new-line handling, or fixing the problems
that Mike pointed out with the description of multi-line mode, these
_really_ need to be perfect.
The more advanced stuff needs implementation experience to get a better
handle on what works in practice. And implementors should have a free hand
to try things, without feeling too constrained by a UTS.
-- Andy
This archive was generated by hypermail 2.1.5 : Tue Sep 25 2007 - 20:38:28 CDT