- From: Edward O'Connor <hober0@gmail.com>
- Date: Wed, 6 Oct 2010 12:20:29 -0700
- To: Marco Rogers <marco.rogers@gmail.com>
- Cc: www-archive@w3.org
Hi, > Interesting. I wouldn't put money on how long it'll be before you can > actually rely on this algorithm. FF4 ships with it. (FF3.something shipped with it too, actually, but it was turned off by default. Set option "html5.enable" to true to use it in FF36.) WebKit nightlies also ship with it. I can't remember if the current releases of Safari and Chrome do too, but they will soon. Dunno about Opera offhand. IE9's parser, while in some respects better than what came before, isn't quite there. All that said, the algorithm was (is) developed with backwards compatibility in mind--the DOM it produces is usually close to what a legacy browser would produce. Closer than some other random tag soup parser, anyway. > But I get your point. Is there a reference implementation of it? Not an official one, but there are some mature ones out there. html5lib implements it in Python and Ruby. Henri Sivonen's Java implementation powers validator.nu, and (in mechanically-translated-into-C++-form) is the FF parser too. As I mentioned on-list, node's own Aria Stewart has implemented it in JavaScript. And (apparently because I hate myself) I'm working on an implementation in Emacs Lisp. HTH. Ted
Received on Wednesday, 6 October 2010 19:21:17 UTC