[rfc][icedtea-web] renewed tagsoup

[Jiri Vanek]

Again, I'm trying to reincarnate old Omair's  tagsoup patch.
As this was developed by Omair, then rewritten by Adam, and now by Me, I think it is quite reviwed :)
tests are adapted, and few added

Ok for head? Some tests included.

Also do not forget  taht the results of this patch are visible only with tagsoup installed.  Without it, or with -xml it behaves as normlay.
(note, this will need to go to wiki)
By the way the PR1026 is fixed by this, and there can come more

I have already made quite deep testing, and some more testing will come, so please do not halt this for tests. Soem reproducers are in row.

J.