Perhaps the page is malformed -- I notice at least one
other problem with it -- but I'd like to parse it. 
Relevant excerpts appear to be:

<script language="JavaScript">
<!--
um.menuCode[i].replace(/<\/(li|ul)>/ig,'</$1>\n');
-->
</script>

goahead() identifies the next interesting part of the
page as the </$1> inside the javascript.  It's not
seeing the comment.  Should it?  I changed
interesting_cdata to lookup for <! as one interesting
possibility and it parsed the comment successfully.