Xalan context support... (Question bundle :-)

[Didier VILLEVALOIS <dv...@techmetrix.net>]

Hi,

I'm trying to compile XSL stylesheets to Java with some XSL stylesheets. I
don't yet want to manage XPath expressions. So i have LOTS of QUESTIONS.
So i'd like to plug my generated Java code to Xalan processor. I need:
	- a path context to manage current selected node set, environnement
variables and namespaces.
	- an output context to manage output documents.

Here is what i understood from Xalan against my need:

1) I must instanciate a XSLTEngineImpl. Let's call it 'engine'.

2) With engine.getExecContext() i get the XPathSupport object which manage
path context. (BUT engine is itself an XPathSupport)

3) With new XPathProcessorImpl(processor.getXMLProcessorLiaison()) i get the
an XPath parser that creates XPath objects from a string. I may then execute
it with the context retrieved before.

4) To output i call the SAX methods of the DocumentHandler retrieved with
engine.getFormatterListener().

Can you tell me what are the initialisation of engine i must take care in
order that it's in an appropriate state (root input node, ouput document...)
?

Can you tell me from where does come the engine.getVarStack() used in
ElemVariable class? I can't find it in the docs.

What is the equivalent of getVarStack() for namespace stack? And why the
need of a PrefixResolver in the execute methods if there is a such namespace
stack ?
How can i know what namespaces have allready been added to output ?

I'm not sure which push/pop methods to use in XPathSupport!!
How may i pop lots of variables in one time ? Do you manage named scopes so
that i may pop multipe scopes in one pop(name) call ?

Are all these stuff are stable ? Or do you have a better interface that i
should plug on ?

I hope you can help me. Thanks for those that read up to here.
Didier.