Happy times with the tokenizer!

I've setup a tester that will actually test the tokenizer again the html5 testsuite. More and more tests found are passing, and most of the states in the tokenizer has been filled with code. I'm now working on fixing attributes in start tokens.

Even though it's probably not the best rust code anyone ever has written, it kind of starting to make sense and it becomes easier to code my thoughts into a functional rust functions so I hope things will speed up as well.

For the next days I will be spending my spare time filling in the tokenizer states, checking line by line where i've made mistakes, and making sure all tests will pass.

From that point on, I might think about converting the tokens into a dom tree, or at least something that looks like that. But it's a part i still have to read a lot more into. For instance, when a script is detected, the dom generation should be halted to let the script do its magic with the dom, before we start the parser again. Since we don't have javascript in the system this will not be happening very soon.. Or maybe there is a (simple?) javascript v8 library i can use?