On Fri, 15 Dec 2000 14:57:44 +0900, ts <decoux / moulon.inra.fr> wrote:
> * how do it compare to GNU regexp in term of rapidity ?

I have no idea, never having benchmarked GNU regexp.  PCRE was
certainly fast enough to be usable.  Though the previous Python regex
engine, from code written by Tatu Ylonen, was faster on really small
strings because of smaller startup costs, PCRE had better asymptotic
behaviour on really large strings. 

> * Is it easy to modify it to add support for japanese encoding ?

Probably not.  Philip Hazel, PCRE's author, has been adding UTF-8
support, but from what little I know of Japanese encodings, UTF-8
support won't really help there.  (I may certainly be wrong.)  Note
that Python 1.6 uses a different engine written by Fredrik Lundh and
called SRE, which handles 8-bit and 16-bit characters, but doesn't try
to handle any specific encoding.  SRE isn't available as a standalone
library, though I'd encourage someone to package it; my project list
is so long that I'll probably never get around to doing so.

--amk