Regex introduction?

Hi,

With strings not being lists of codes/chars, the idea to have meaningful
and efficient operations on them becomes more appealing.

For a long time, I thought that the solution was the approach started
(for SWI-Prolog) by Michael Hendrix in the pack regex
(http://www.swi-prolog.org/pack/list?p=regex). If (code) lists are the
prime data structure to work on, this makes sense.

On the other hand, regular expressions are a well established field and
redoing it all in Prolog seems a waste. Just like building your own
Prolog like inference engine in language X :-) As strings are low-level
data structures anyway, we can use standard regex implementations on the
low-level data.

I'm seriously considering to add regular expression support based on the
pcre (Perl Compatible Regular Expressions) library. This seems to be the
library of choice for most modern languages and provides support for the
two internal representations used in SWI-Prolog for Unicode text.

PCRE will (at least initially) be provided as a package. There will be a
low level interface that quite directly implements the full C api of
PCRE. The idea is to provide a re_compile that compiles a pattern with
options into a `blob', which makes it efficiently available and subject
to (atom-)GC and an re_exec that returns positions for matches and an
re_replace that does a regex replacement, returning a new string. On top
of that, we can define more high level libraries.

It seems only YAP and XSB provide regular expressions as a native
feature. YAP based on Henry Spencer's library and XSB on POSIX 2 regex
support.

Did I miss something?  Any precedence of well balanced interfaces for
Prolog?  The YAP interface is rather limited.  The XSB one is richer,
but a bit weird due to issues with atom-GC in XSB you should not be
bothered with as a user.

	Thanks --- Jan