The tokenizer is currently passing 0 out of the 2932 tokenization tests from html5lib. :) Of course, my test runner is just stubbed.