Re: [xml] redicting parts of trees

[Rob Richards <rrichards ctindustries net>]

I was more concerned about getting locked into a specified context 
structure at this point and unsure about how flexible changing it was 
going to be going forward. As I havent had a chance to do anything more 
than read over the code - hadnt added it to my tree as I needed to test 
the existing code for the 2.6.20 release - I couldnt say for sure at 
this point what I might need for context support.

As long as the context structure isnt going to be set in stone at this 
point - however it ends up in the release- then it shouldnt be an issue. 
The functions are fine argument wise as I agree anything outside the 
current argument lists should be handled via the context.

And context naming, yes I am content with that. It is specific to the 
DOM functions as I was just going to suggest xmlDOMCtxt so the one in 
CVS is pretty much the same.

Rob

Daniel Veillard wrote:

> On Mon, Jun 27, 2005 at 09:23:43AM -0400, Rob Richards wrote:
>  
>
>
>  I am the one who suggested to Kasimier to push to CVs precisely
> to get testing, but labelling it as experimental. I think we can grow the
> structure fairly reasilly in the future. Are you afraid of PHP-5.2 compiled
> against a vertion of libxml2 post 2.6.20 and running with 2.6.20 and hence
> a limited context ? If this is the case we can arbitrarily put a function
> pointer in the lot and a couple more int and data pointers for good measure
> and feel safe about it.
>
> Daniel
>
>