Re: [boost] Re: Formal Review: The Spirit Parser Framework.

From: Joel de Guzman (djowel_at_[hidden])
Date: 2002-10-15 02:23:10

From: "Carl Daniel" <cpdaniel_at_[hidden]>

> If Spirit doesn't depend on Phoenix, why is this here?  (It appears that
> this part of Spirit does in fact depend on Phoenix - is this part of what's
> being reviewed?).

The immediate plan is to write spirit::closure code that will optionally
use phoenix or boost::lambda. I still have to talk to the BLL authors
regarding this. There are still some issues in LL that need to be
addressed. For instance, many Spirit users use Borland. BLL is not yet
ported to Borland and some more compilers, Phoenix is. BLL has a
fixed arity limit of 10, Phonix has a parametizable max arity. Plus a 
few more...

Right now, BLL has incorporated Phoenix adaptable closures though. That is
most important. BLL and Phoenix now are interchangeable in this regard.

Another plan is to move the adaptable closures facility in a library of its
own and perhaps submit it for boost review. Adaptable closures are such
useful beasts in the realm of FPC++. It can be useful in other applications
besides parsing. For instance in "Techniques for Scientific C++" by
Todd Veldhuizen (http://osl.iu.edu/~tveldhui/papers/techniques/techniques01.html#l46), 
chapter 1.13 talks about "Dynamic scoping in C++" and presents a limited and
bare-bones version of Phoenix's adaptable closures.

Regards,
--Joel