Xalan-c, bdbxml and the document() function

[Ivan Bogouchev / Иван Богушев <iv...@gmail.com>]

Hi all,

I am considering the integration of Xalan-c and sleepycat's berkley dbxml.

It is no problem to extract a document from the db, and then 
reparse/transform it with Xalan (there is also the possibility to get 
directly a Xerces-c DOM tree from the db).
But I need to be able to load new documents coming from the db from within 
my stylesheets. 

So I guess one solution would be to write an extension function for the 
XalanTransformer that loads a document from the db (or even executes a 
XQuery), but I was wandering if it is possible to extend the existing 
document() function, in order to be able to do document('
bdbxml://container/document').

And now the question: is it possible to extend the document() function in 
such a way and if yes can someone give me some hints on where to start?


-- 
--Ivan--

Re: Xalan-c, bdbxml and the document() function

[da...@us.ibm.com]

> I am considering the integration of Xalan-c and sleepycat's berkley
> dbxml.
> 
> It is no problem to extract a document from the db, and then
> reparse/transform it with Xalan (there is also the possibility to get
> directly a Xerces-c DOM tree from the db). But I need to be able to
> load new documents coming from the db from within my stylesheets. 
> 
> So I guess one solution would be to write an extension function for
> the XalanTransformer that loads a document from the db (or even executes
> a XQuery), but I was wandering if it is possible to extend the existing
> document() function, in order to be able to do
> document('bdbxml://container/document').
> 
> And now the question: is it possible to extend the document() function 
in
> such a way and if yes can someone give me some hints on where to start?

What we really need in Xalan-C is a concept like the JAXP URIResolver, but 
designed with C++ in mind.  We had several proposals on the list a few 
years back, but didn't really get anywhere with them.  My suggestion is 
you take a look at the implementatation of the existing document() 
function, because you will see calls to parse URLs go through 
XSLTProcessorEnvSupportDefault::parseXML().  You can modify that class to 
intercept calls to specific URL schemes and supply the appropriate 
document.

Dave