RE: document('') and URIResolver

Ok all, 

Enough diatribe off of the main concern which is:

How in my transformer specific URIResolver do I handle
"" or null href args in the URIResolver.resolve(href,
base) method?

Jon

--- Michael Kay <mike@xxxxxxxxxxxx> wrote:
> > While in principle it would be possible for a
> system to special case
> > this argument to document() and instead of
> reparsing the XML just to
> > re-use some in-memory tree constructed when the
> stylesheet was parsed,
> > this might be difficult in practice for a system
> to achieve as for
> > example the default white space stripping rules
> are different, and a
> > system might have done other in-place changes to
> teh stylesheet tree,
> > for example related to processing xsl:include or
> code-re-writing in a
> > code optimiser...
> 
> Yes; in fact, the processor might well not have the
> original source
> stylesheet tree still available at run-time, and it
> almost certainly won't
> have all the original whitespace. There are also
> complications because
> document("") does NOT select the current stylesheet
> module; it selects the
> resource whose URI is the same as the base URI of
> the element containing the
> call on document() - which might not be the same,
> for example, if the
> stylesheet module was assembled from external XML
> entities.
> 
> Michael Kay
> http://www.saxonica.com/
> 
> 


		
Yahoo! Mail
Stay connected, organized, and protected. Take the tour:
http://tour.mail.yahoo.com/mailtour.html