Data.ByteString candidate 3

Donald Bruce Stewart dons at cse.unsw.edu.au
Thu Apr 27 03:16:00 EDT 2006


john:
> On Wed, Apr 26, 2006 at 08:19:34PM +1000, Donald Bruce Stewart wrote:
> > bulat.ziganshin:
> > > Hello Donald,
> > > btw, what will be really useful now, imho, is the interface to
> > > Text.Regex. how about working on it as next stage?
> > 
> > This is already done actually, here:
> >     http://www.cse.unsw.edu.au/~dons/code/lambdabot/Lib/Regex.hsc
> >     http://www.cse.unsw.edu.au/~dons/code/hmp3/Regex.hsc
> 
> I have a regex interface to PCRE and some neat typeclass tricks to give
> you perls (=~) operator but much more powerful here.
> 
> http://repetae.net/john/computer/haskell/JRegex/
> 
> It would be nice to get a PCRE binding in the libraries if it is
> available.
> 
> if there is interest in including this in the fptools libraries I can
> revisit and clean-up/modernize the code.

We really longed for a high performance regex lib in the standard
libraries while working on the shootout earlier this year. Text.Regex is
far too inefficient due to all the pack/unpackings. and even then C's
regexes aren't so great.  In fact, Chris K ended up writing
Tex.Regex.Lazy as a result of this effort.

Here's a nice benchmark for you code:
    http://shootout.alioth.debian.org/gp4/benchmark.php?test=regexdna&lang=all 

I wonder if JRegex would give us a faster entry?

After fast IO, regexes are the other thing we need to improve for ghc
6.6, I think. So at least the people who worked on the shootout would be
interested :)

-- Don



More information about the Libraries mailing list