Re: announcing synfl, the next generation of fontlock

Stephen J. Turnbull wrote:

>     dvl> SGLR module
>
> Uh, do you mean an ELL?  I don't think that's the way to go.  You
> could provide the relevant hooks in C, but I see no reason to believe
> that a Lisp parser can't be plenty fast.

        Scannerless Generalized LR. In short, it's capable of keeping a parse
forest (instead of a single parse tree) and cut branches once ambiguities are
resolved. And yes, I planned to make it a module in the Jerry sense. Actually,
I thought that was what you did in synfl (since you spoke of a "module"). As
for speed, I don't know but I have the opposite feeling. I'd rather make the
parser in C, and provide hooks in Lisp when needed.


>     dvl> 2/ Next step would be incremental parsing. I currently don't
>     dvl> have any knowledge on this matter.
>
> I'm not sure what you mean.  With a properly designed grammar, I'm
> pretty sure that it's possible to restart synfl at the beginning of
> any extent corresponding to a nonterminal of the grammar.

        Like I said, I'm not knowledgeable in incremental parsing, but some
friends of mine tell me it's a complete different story.

-- 
Didier Verna, [email protected], http://www.lrde.epita.fr/~didier

EPITA / LRDE, 14-16 rue Voltaire   Tel.+33 (1) 44 08 01 85
94276 Le Kremlin-Bicêtre, France   Fax.+33 (1) 53 14 59 22   [email protected]