Re: New regex lib? (fwd)

[Rasmus Lerdorf <ra...@lerdorf.on.ca>]

Guys, for PHP4 (and Apache2) we should probably look into using Henry's
new regex library.

---------- Forwarded message ----------
Date: Mon, 18 Jan 1999 11:35:05 -0500 (EST)
From: Henry Spencer <he...@spsystems.net>
To: Rasmus Lerdorf <ra...@lerdorf.on.ca>
Subject: Re: New regex lib?

Sorry for slow response...  Life has been exciting, not to mention
unpleasant, for the last few days.

> Anyway, I am just wondering what the state of your regex library rewrite
> is.  I seem to recall you mentioning somewhere that you were doing a
> complete rewrite...

Indeed so, and it has about reached the point of being shippable; I just
have not done the packaging yet.  It's shipping with Tcl/Tk's 8.1 beta
release, but I need to spend a bit of time putting together a standalone
version -- currently it relies on the Tcl 8.1 environment.  I hope to do
that in the next few weeks. 

To sum up in a sentence or two...  It has substantial new functionality,
including most of the Perl extensions (although a few of them are not done
quite the same way).  It can handle large character sets, although as yet
only in fixed-width encodings.  Performance is not yet sparkling but is
now decent, unlike my old alpha-release code that shipped with 4.4.  A
simple test set puts Tcl 8.1 regex performance roughly on par -- some
things faster, some slower -- with 8.0 (essentially my original 1986
regexp package), which is noteworthy when you consider that 8.0 had 8-bit
characters and 8.1 is using Unicode. 

                                                          Henry Spencer
                                                       henry@spsystems.net
                                                     (henry@zoo.toronto.edu)