Paul DuBois wrote:
I did some investigation into Jeni's suggestion of using a SAX-based
transform
to resolve the XIncludes.
Having tried to implement an XInclude processor myself, I
can tell you it's not easy. In particular, trying to avoid
building a full in-memory representation means that some
valid XInclude instructions can't be processed in conformance
to the spec, and it may even be impossible to detect this.
It is propably interesting to see whether your problem can be
traced back to an avoidable or unavoidable inefficiency in
libxml's xinclude processing. Perhaps you should also ask
on their user list.
J.Pietschmann