* support http://swish-e.org/docs/swish-config.html#dontbumppositiononendtags

* noindex comments handler

* tokenizer: regexp lib

* StringList parse_line_into_words should support UTF8 (widechar)

* html parser features for many 2.x config opts (img, a.href, etc)

* support full-content swishdescription regardless of root element name.

* switch to libunistring http://www.gnu.org/software/libunistring/ for
  all UTF-8 related string handling.