> It disappeared becaue the "shadow" root document fragment disappeared, > which was a quick-and-hacked implementation and caused no end of trouble. > > Our ResultTreeFrag specialization of XalanDocumentFragment also > disappeared, as it was offering no value and just created complications. > > The real "root" is now the result tree fragment itself. Is there any > reason why that won't work? > I'm just not sure how I add this to a nodeset. Do I have to clone it?
theNodeList->addNode((*it).second->rtree().cloneNode( false )); Should the deep flag be set? Does the nodelist then own (and therefore ultimately cleanup) the cloned tree? Thanks, Mark > Dave > > > > |---------+---------------------------> > | | "Mark Weaver" | > | | <[EMAIL PROTECTED]| > | | > | > | | | > | | 04/07/2004 05:26| > | | AM | > | | Please respond | > | | to xalan-c-users| > |---------+---------------------------> > > >----------------------------------------------------------------- > ---------------------------------------------------------| > | > | > | To: "Xalan-C-Users" > <[email protected]> > | > | cc: (bcc: David N Bertoni/Cambridge/IBM) > | > | Subject: rtree()->getNodesetRoot() > | > > >----------------------------------------------------------------- > ---------------------------------------------------------| > > > > I've recently moved up to Xalan 1.7 and it seems that the getNodeSetRoot() > function has disappeared from XalanDocumentFragment. This was used in one > of my (nasty) extension functions that implements a dictionary. > > Basically, I'm trying to stick the RTF to a nodeset in the following code: > > dictionary_t::iterator it = > m_dic.find(arg1->str().data()); > > XPathExecutionContext::BorrowReturnMutableNodeRefList > theNodeList(executionContext); > if (it != m_dic.end()) { > > theNodeList->addNode((*it).second->rtree().getNodesetRoot()); > } > return > executionContext.getXObjectFactory().createNodeSet(theNodeList); > > the idea being that the nodeset is empty if the dictionary key > could not be > found or contains the node otherwise. (There's also a corresponding > function to return all of the nodes). > > Could anyone advise as to the correct way to do this these days? > > Thanks, > > Mark > > > > >
