URL parsing regex

Url.js

(function () {

  // Regexp is based on Crockford's "Javascript, the Good Parts"

  // (?:([A-Za-z]+):)? - protocol (http, ftp, etc...) (optional)
  // (\/{0,3}) - slash(es)
  // (?:(\w+):(\w+)@)? - username:password@ (optional) -- does this need to support hyphens and dots?
  // ([\w\-.]+) - host name (shortened from [0-9.\-A-Za-z]+)
  // (?::(\d+))? - port number (optional)
  // (?:\/([^?#]*))? - match path (optional) (caution, matches all character except ? and #, so it will also match line-ending characters, control character, and lots of other characters)
  // (?:\?([^#]*))? - match querystring (optional)
  // (?:#(.*))? - match fragment (optional)

  var re = /^(?:([A-Za-z]+):)?(\/{0,3})(?:(\w+):(\w+)@)?([\w\-.]+)(?::(\d+))?(?:\/([^?#]*))?(?:\?([^#]*))?(?:#(.*))?$/,
      names = ["href", "protocol", "slash", "username", "password", "host", "port", "pathname", "search", "hash"],
      len = names.length,
      cache = {},
      url = {

      // Always return at least an object w/ empty values.

      parse: function (href) {

        if (cache[href])
          return cache[href];
  
        var values = href ? re.exec(href) : [],
            result = {},
            count = len;

        // Skip 'href' value in index 0

        while (--count)
          result[names[count]] = values[count] || "";
  
        cache[href] = result;
  
        return result;
      },

      // https://developer.mozilla.org/En/Same_origin_policy_for_JavaScript

      isSameOrigin: function (hrefA, hrefB) {
        var parse = url.parse,
            a = parse(hrefA),
            b = parse(hrefB);

        // FIXME: Handle empty responses from parse()

        return a.protocol == b.protocol &&
               a.host == b.host &&
               a.port == b.port;
      }
    };

  // Expose public interface

  window.url = url;
  
})();