The NodeSet class can act as either a NodeVector, NodeList, or NodeIterator. However, in order for it to act as a NodeVector or NodeList, it's required that setShouldCacheNodes(true) be called before the first nextNode() is called, in order that nodes can be added as they are fetched. Derived classes that implement iterators must override runTo(int index), in order that they may run the iteration to the given index.
Note that we directly implement the DOM's NodeIterator interface. We do not emulate all the behavior of the standard NodeIterator. In particular, we do not guarantee to present a "live view" of the document ... but in XSLT, the source document should never be mutated, so this should never be an issue.
Thought: Should NodeSet really implement NodeList and NodeIterator, or should there be specific subclasses of it which do so? The advantage of doing it all here is that all NodeSets will respond to the same calls; the disadvantage is that some of them may return less-than-enlightening results when you do so.
@xsl.usage advanced | |