Re: [Boost-users] lambda parse tree access

From: Markus Werle (numerical.simulation_at_[hidden])
Date: 2006-12-18 07:49:03

Eric Niebler <eric <at> boost-consulting.com> writes:

> I am actively working on a generic expression template framework called 
> proto. It is already checked into Boost CVS as part of xpressive (see 
> boost/xpressive/proto on CVS HEAD). It is also being used in the 
> development of Spirit-2. I think it would be a very good fit for what 
> you are trying to do.

So I am too late again. 
Actually I am working on a rework (better complete redesign) of Daixtrose
called etl that fits the way things work at boost and also makes
use of fusion and mpl.
If you are interested in some fragments of my design thoughts, please contact
me by mail. I am very interested in having compile time differentiation
supported right from the beginning. Also I have some thoughts about how to
implement flexible disambiguation mechanisms and how to achieve non-intrusive
ETs for all classes.
After a first glance at proto I feel like disambiguation is not included, 
but I may be wrong ...

> Unfortunately, the code is still in some flux, and as such is not 
> documented. I have some read-me material you might find helpful, but 
> nothing complete or comprehensive. Let me know, and I can send it along.

Yes, please.
I have some comments on your code, but feel a little lost with
all the preprocessor stuff. If you can provide some meta info like 
e.g. design decisions this makes a first step much easier (
especially for coders like me with lower skills than Eric N.).

Markus