infundibulum

I was looking around the Mozilla XPath Documentation because I wanted to write a bookmarklet to lists headings in pages (I’m lazy like that). And low and behold, the first example I found did (almost) just that. Apparently you do something like:


var headings = document.evaluate("//h2", document, null, XPathResult.ANY_TYPE,null);


But that doesn’t really answer my question about how use XPath to get all the headings into that headings variable .

Digging around on the same site, I discovered a Firefox Extension which does everything I had in mind and then some: Document Map. It produces nifty outlines like this.

But that still doesn’t scratch the itch, of course — how do I get the XPATH to do what I want? So I asked my homie Jonas (er, can you have a homie in another country?), and he found this (in Java documentation, of all places):

Finding Elements by Absolute Location in a DOM Document Using XPath (Java Developers Almanac Example)

Which has some nice examples. Anyway, here’s the answer, apparently:

XPath 1.0 does not support regular expressions to match element names.
However, it is possible to perform some very simple matches on element names.

    // Get all elements whose name starts with el
    xpath = "//*[starts-with(name(), 'el')]";  // 2 3 5 7 8 9

    // Get all elements whose name contains with lem1
    xpath = "//*[contains(name(), 'lem1')]";   // 2 8

So I guess this is the answer:

xpath = "//*[starts-with(name(), 'h')]";

Assuming that there aren’t any other tags that start with h. Which is a dumb assumption. Er… are there any? XPath syntax is a little nutty-looking, if you ask me, but I guess it just takes some getting used to.

But whatever, for now problem solved.

But wait there’s more:

To really learn XPath:

Mark Pilgrim’s Dive into Greasemonkey also has a list of further reading links, one of which points to this XPath Tutorial by Example. It’s already been translated into seven languages, so I guess it must not suck. = )