Question

This is just a simple question. I’m currently using Mozilla’s Rhino to develop a little webapp. As one step, I need to get a webpage and filter all of it’s nodes. For doing that, I use E4X. I thought I could do this like that:

var pnodes = doc..*(p);

But that produces an error. How is it done right?

(BTW: this is just a step for increasing performance. The code already does well, it’s just a bit slow.)

Answer 1

You should be able to use the following:

doc..*.(name() == "p")

Note that this there is a bug in the Rhino and SpiderMonkey implementations where the filter expression name() == "p" is not correctly scoped to the current node, so none of the XML or XMLList methods are defined.

---

Another workable solution is to lookup all p nodes in the document and accumulate the parent of each in an array.

var elements = [];

for each (var p in doc..p) {
    var parent = p.parent();
    if(elements.indexOf(parent) === -1)
        elements.push(parent);
}