does not work with mime type text/html

[TRPB]

a small adjustment to the sample code stops it working:

var xpath = require('xpath')
  , dom = require('xmldom').DOMParser

var xml = "<book><title>Harry Potter</title></book>"
var doc = new dom().parseFromString(xml, 'text/html');
var nodes = xpath.select("//title", doc)

console.log(nodes);

nodes is an empty array.

I don't know whether this is an issue with xmldom or xpath but this fixes it:

var xpath = require('xpath')
  , dom = require('xmldom').DOMParser
  , serializer = require('xmldom').XMLSerializer;

var xml = "<book><title>Harry Potter</title></book>"
var doc = new dom().parseFromString(new serializer().serializeToString(new dom().parseFromString(xml, 'text/html')));
var nodes = xpath.select("//title", doc)

console.log(nodes);

Obviously this is terribly inefficient but it does work.

[JLRishe]

There's already an open issue #27 on this. The library works with DOMs that are parsed that way, but parsing with that mime type implicitly puts the nodes in the http://www.w3.org/1999/xhtml namespace (this is the parser's doing, not the xpath library). That's why you're not finding any matches with //title.