Re: [boost] scope of ptree-xml parser

From: Daniel Wesslén (daniel_at_[hidden])
Date: 2006-04-24 09:39:25

Thorsten Ottosen wrote:
> Daniel Wesslén wrote:
>> Thorsten Ottosen wrote:
>>>> When boost gets a full XML library then the property_tree XML parser 
>>>> could (should, IMO) be replaced with a wrapper over that one. What's the 
>>>> problem?
>>> There is no need for that. As long as the scope of the xml-parser for
>>> a property tree is well-defined wrt. what subset of xml it can handle, 
>>> the user is free to choose.
>>>
>>> It is not obvious how more advanced xml-features should map to a 
>>> property-tree, and it is quite thinkable that it does not make much 
>>> sense to use those together.
>>
>> Schema validation, XInclude and encodings could be transparently 
>> supported. Sure, some advanced features wouldn't map neatly to a 
>> property_tree, but that is no reason to rule out support for those that do.
> 
> I'm not qualified enough to reply to those claims.
> 
> But let's use this sub-thread to discuss the scope of the xml-parser.

Given that property_tree isn't meant to be a XML library as such, I 
don't think there is much need. I agree that is should be clearly 
documented which features of XML the parser can handle. After that is 
done, there is no reason not to add features that can fit in the tree.

I was simply suggesting that when we get a full-featured parser, then 
that one should be used for property_tree as well, to support as much of 
XML as possible within the constraints of what's feasible to store in 
the ptree.

-- 
Daniel Wesslén