Re: [xml] Feature request: callbacks

[Paolo Casarini <casarini CS UniBO IT>]

On Thu, 15 Mar 2001, Daniel Veillard wrote:

> > Isn't this feature part of DOM2 Mutation Events ? What's the reason why it
> > should be part of the node tree api ?
>
>   Dohhh ... right
> http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-eventgroupings-mutationevents
>
>   What freaks me is the following:
> http://www.w3.org/TR/DOM-Level-2-Events/events.html#Events-Registration-interfaces
>
> "The EventTarget interface is implemented by all Nodes in an implementation
>  which supports the DOM Event Model."
>
> Which is clearly something I try to avoid (changing all node data structures
> to add more stuff in them). I'm not sure I want to go there, I think the
> DOM events were implemented on top of libxml (Raph Levien used them for
> gill) without adding something specific.
>
> The initial need for those callbacks cames from the use of _private, i.e. 
> the application reserved pointer in the Node structure. Seems to me that
> if there is really a need to stick to the DOM interface then this can (and is
> ) done with _private and in this case the DOM event registration can be
> implemented at this layer with minimal penatly for all the other users of
> libxml.
I agree with you, mutation events belong to DOM implementations. The only
usefull callback is to avoid memory leak caused by the use of _private.
In my opinion the problem is memory leak not events.
Dealing with events is a DOM problem.

Paolo
--
Paolo Casarini - casarini cs unibo it