TODO list
=========

  + Error checking
  + Documentation
  + Implement one or more tree builders
    - test/tree2.c would serve as a basic tree builder
    - NetSurf's libxml2 binding could do with being brought back here somehow
  + Parse error reporting (incl. acknowledging self-closing flags)
  + Implement extraneous chunk insertion/tokenisation
  + Statistical charset autodetection
  + Shared library, for those platforms that support such things
    - requires possibly prefixing more things with hubbub_
  + Optimise it
    - being clever with e.g. attribute allocation in tokeniser
    - tag name interning / replacing element_type_from_name()
    - Hixie's data (http://tinyurl.com/hixie-html5-data-2007)