From: Colin Paul Adams <colin@co...> - 2005-08-05 06:07:30
Yesterday, I committed a change in which I scrapped namespace nodes
from the XPath data model.
This was done for performance reasons, but it also means the tasking
of creating a virtual data model over an XM_DOCUMENT should be easier.
This could be useful where a program has a large XM_DOCUMENT in
memory, and cannot afford to build a copy.
In most cases, where memory is no a limiting facter, it will perform
better to serialize the tree to an event stream, and build a normal
Neither of these options is supported at present (the only way to
transform such a tree at the moment is to serialize it to an xml text
representation). Each will require a URI scheme to identify the source
tree, as the XSLT standard requires a URI for the source document, and
the Gobo XSLT library accordingly requires this as an argument to the
Get latest updates about Open Source Projects, Conferences and News.