Re: [libxml++] Proposition: nodes manipulation

[Christophe de V. <cde...@al...>]

Stefan Seefeld a écrit :

> Christophe de Vienne wrote:
>
>> Currently, the lifetime of nodes is controlled by libxml2, via 
>> callbacks on node ceation/deletion. The restrictions are that we 
>> cannot create a standalone node, nor detach and then delete a node 
>> like a normal class instance.
>
>
> Why do you believe this is a restriction ? Being able to unlink a node
> from its document will open up a big can of worms. Not only will you
> have to deal with resource allocation issues (until now nodes are owned
> and thus taken care of by a document), but there are other issues such
> as context data that are required for the node to make sense. Just think
> about namespaces.


I see what you mean. In my mind the restriction is compared to what 
libxml2 proposes.
Any issues with context datas are already adressed by libxml2, and 
ressource allocation issues could eventually be solved with smart 
pointers (cf my other post).

>
> Sometimes it's better for an API to make it hard (if not impossible)
> for users to do certain things, just because there are better ways
> to achieve what they want.


That why I want to discuss this API change. I'm not 100% sure it worth 
it (let's say I'm around 75% :-), but I'm sure it worth a try.
I think being able to construct a subtree independently from the final 
tree is interesting.

Regards,

Christophe