[xml] How does xmlReader prevent tree building

[Joachim Zobel <jz-2006 heute-morgen de>]

Hi.

I have been reading xmlreader.c to find out how the reader does parsing
without tree building. However I did not find out, how tree building is
prevented. I would do a

ctxt->node = NULL;

after a StartElement event and something similar for other events. Thats
not what the reader does. Any hints?

For context on what I am trying to do see
http://www.heute-morgen.de/site/03_Web_Tech/50_Building_an_Apache_XML_Rewriting_Stack.shtml

Sincerely,
Joachim