Re: [Spirit-general] Parse tree, and ast tree formats...

[Joel de G. <jo...@bo...>]

Hartmut Kaiser wrote:

> Giovanni Bajo wrote:
> 
> 
>>>Does it still makes sense to apply your patch to the CVS code base?
>>
>>No, because it breaks backward compatibility badly, and there 
>>are people who find the current approach better.
> 
> 
> I see.
> 
> 
>>>Or do you expect to get your new template policy based ready in a 
>>>reasonable amount of time?
>>
>>Not for 1.8.0 in any case. I think it will be too intrusive a 
>>change for it.
> 
> 
> But we urgently need a solution for this! We have this discussion already
> for over a year. Let's add at least a 'special' AST policy, which contains
> your patch, say non_collapsing_ast and non_collapsing_ast_parse et.al. (yes,
> the names are very ugly, but suggestions welcome). This may be changed into
> a simple typedef later on, if/when you've finished your template policy
> based implementation.

I agree with Hartmut. We need new solutions and we need them ASAP.
We can't wait any longer for another few months. I do not agree however
that we incorporate this into the 1.8.0 branch. The best we can do, IMO,
is to provide a solution for the HEAD branch which will then become
the basis of 1.9.0.

<< WRT 1.9.0, I'd also like to add the sub_grammar proposal (see my
other post re: parsing binary data. The class has been around for
quite some time now and I am very fond of it. I'm quite impressed
with its code size reduction. >>

-- 
Joel de Guzman
http://www.boost-consulting.com
http://spirit.sf.net