HTML parser appends extra html node to root

[jdm]

By default, the parser creates an HTMLElement node as the root of the document, and then appends all nodes encountered to that. It currently doesn't discriminate when it does so, so when it encounters an actual element in the source, it merrily appends a second one to the root. We should find some way to avoid this.

[jjjjw]

I have been looking into this and it seems to be a consequence of the fact that hubbub (the current html parser) does not provide a nice way to get at the root node after the parsing, nor do we want to implicitly assign the root on the fly in a callback as we parse. So this spurious root is created up front, and I actually don't see a good way to get around that. I certainly could be failing to grasp something.

I assume that an html parser in Rust is somewhere down the line? Then we can just haz the tree. This is my naive solution.

[jdm]

Yes, code apparently exists to automatically translate Gecko's Java-originating HTML5 parser into native Rust code.

[jjjjw]

Wait just second. This is all a misunderstanding. Upon second glance it seems that there is a confusion between hubbub's document node and the html node. I should be able to straighten that out so that the current parser works in the meantime.

[jdm]

We should be able to fix this now by passing the Document node to the parser and using that as the root.

[jdm]

This is fixed now.