From: Eric Mahurin <eric.mahurin@...>
Date: 2006-08-14T15:33:53+09:00
Subject: Re: converting some autogenerated ruby code to C

------=_Part_41756_6619067.1155537231134
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Yep.  I haven't done a checkin to CVS in a while though.  I've been fairly
actively working on in the last few months.  One of the latest interesting
things I've figured out and put in is left recursion.  Normally this is
unheard of with LL parsers.  This opens up a many of the LR grammars (i.e.
from YACC) to be used with my parser (but I still need a (E)BNF converter to
make it easier).

Hopefully once Dominik implements some kind of *eval replacement, I'll be
able to use ruby2cext and get the performance of Regexp.  With a hand coded
character-by-character lexer (the kind Grammar generates), I see a 4-5X
performance improvement (using his CVS ruby2cext) which beats the fastest
equivalent Regexp/StringScanner lexer (only a small improvement using
ruby2cext).  This could make Grammar as a Regexp alternative - equivalent
performance and much, much more flexible.

I'll try to make a checkin soon so that you can play around with it.

On 8/13/06, Phil Tomson <philtomson@gmail.com> wrote:
>
>
> Eric Mahurin wrote:
> > One of the next things I want to do in my grammar package is to give
> > it a performance boost by generating C code instead of ruby code.
>
> Does this mean that new versions of Grammar will be released in the
> future and that works continues on it?
>
> Phil
>
>
>

------=_Part_41756_6619067.1155537231134--