[jdom-interest] Using JDOM and LARGE XML Files ...

[Jon Baer]

This sounds exactly what I need to do, write a custom builder.  Just to
better clarify here is my project:

http://www.digitalanywhere.com/projects/dany/index.html

Its a chatbot client/server uses XML but some categories will load over
1000 elements or more and currently it uses a brute-force parser for a
non-XML input, which I am trying to change but it might not make sense to
load up a single document (one document can load other files) and keep this
thing in memory if Im not going to use it.  It needs to fill a class called
Graphmaster which handles all the nodes.

My other comment was if anyone has thought of writing a method in
XMLOutputter that would/could write directly to a server via POST.  I think
this would be a nice convienent addition, is there anything against it?  I
did it a while back where I simply passed it off to another threaded class
called PostThread but I was hoping something like this (considering all the
new XPs out there) would have been in XMLOutputter already.

- Jon



Jason Hunter wrote:

> Jools Enticknap wrote:
> >
> > Well, you can't extend SAXHandler because it's package protected, and
> > I'm not sure that making it public is a Good Thing (TM).
> >
> > Either delete the unrequired data after the Document is parsed, or
> > just live with the extra memory usage.
>
> Or you could write a custom builder modeled after SAXBuilder but with
> your discriminatory logic.  It just wouldn't be made standard.
>
> -jh-