[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kfm-devel
Subject:    Re: Regular expression library choices
From:       Rob Gillen <rgillen () webmd ! net>
Date:       2001-11-06 16:23:51
[Download RAW message or body]

Okay, I got word back from John Maddock, the author of Regex++.  I 
learned something new, and something about POSIX regular expressions, 
something that perhaps wasn't too clear to me.  This is a snippet of 
what he has to say:

     The main difference is in the way things are matched: POSIX (and
     regex++) uses the "leftmost longest rule", while perl uses a
     "greedy/non-greedy" mechanism - for example ".*([a-zA-Z]*).*" $1
     will match the first word under POSIX rules and the last word under
     perl/Javascript rules.

     I keep meaning to investigate adding a "perl compatible matching
     mode" to regex++, but haven't found the time (and probably won't
     this side of Christmas).

So, I was wrong, and it appears, unfortunately, that Regex++ probably 
isn't going to work at this time.  It's a shame, because it really is a 
very nice library.

Rob


Rob Gillen wrote:

> Are you sure that Regex++ does not understand ECMAScript regular 
> expressions?  Perhaps I am misunderstanding something.  Regex++ 
> understands POSIX expressions for compatibility with the older regex 
> library.  As you no doubt know, ECMAScript regular expressions are based 
> on Perl5 regular expressions.  Regex++ does indeed seem to handle almost 
> everything that Perl regular expressions can handle.  There are probably 
> a few exceptions, but the only ones that I can immediately find in the 
> documentation is zero-width positive and negative lookaheads.  I've 
> written author of Regex++ for his thoughts on Regex++ vs. ECMAScript 
> regex capabilities.  Hopefully, soon, I'll hear back from him on the 
> topic and I'll let everyone know more.


-- 
Rob Gillen
WebMD, Midtown-Atlanta

[prev in list] [next in list] [prev in thread] [next in thread]