Re: Interactive parsing with Bison

[Gary V. Vaughan]

A Tout le Monde:

Akim Demaille wrote:
> 
> Le 10 juil. 06 à 09:27, Richard Stallman a écrit :
> 
>>     Compiled C is faster that interpreted (and scanned many times) M4
>>     code.  Some constructs, particularly list processing and call stack
>>     manipulation, are implemented in M4, but extremely slow.  If they
>>     were implemented as extension in C, the speedup would be tremendous.
>>
>> Would it make sense to add a feature to M4 which is good for that job?
> 
> Yes, definitely IMHO).  The problem here is not the lack of
> understanding of
> what could be done (writing the C code for these primitives is actually
> quite simple), the problem is more like there is not enough resources
> to finish the modular infrastructure Gary started.  Once the infrastructure
> finished and released, it should be quite straightforward to use it.

Agreed 100%.  Happily, there has been a flurry of activity around M4
this last 6 weeks or so (mainly Eric Blake and Ralf Wildenhues) which
I'm encouraging in the hope it can be sustained.  A new stable M4 1.4.5
is right around the corner, after which work on the modular M4 in CVS
trunk will resume.

Despite the fact that the modular framework API on the trunk has not
yet fully stabilised, I would be delighted to accept utility modules
from other projects to improve our understanding of what needs to be
done to make the API as useful and flexible as possible.  Not to mention
that once a module (and associated tests) are in the M4 trunk, they
will be maintained to keep track of changes to the module API prior
to the release of M4-2.0.

Cheers,
        Gary.
-- 
Gary V. Vaughan      ())_.  address@hidden,gnu.org}
Research Scientist   ( '/   http://blog.azazil.net
GNU Hacker           / )=   http://trac.azazil.net/projects/libtool
Technical Author   `(_~)_   http://sources.redhat.com/autobook

signature.asc
Description: OpenPGP digital signature