TODO

 - limit the length of markup elements that never end.   Perhaps by
   configurable limits on the length that markup can have and still
   be recongnized.  Report stuff as 'text' when this happens?
 - unicode support (when parsing Unicode strings the strings reported
   in callbacks should also be Unicode strings).
 - remove 255 char limit on literal argspec strings
 - implement backslash escapes in literal argspec string
 - <![%app1;[...]]> (parameter entities)
 - pic attribute (">" or "?>" are defaults)  useful where you want to
   have "?>" as pic, but don't want the rest of xml_mode. For instance
   to parse PHP files.
 - make literal tags configurable.  The current list is hardcoded
   to be ""script", "style", "textarea", "xmp", and "plaintext".

SGML FEATURES WE WILL PROBABLY IGNORE FOREVER
 - Empty tags: <> </>  (repeat previous start tag)
 - <foo<bar>  (same as <foo><bar>)
 - NET tags <name/.../
 

POSSIBLE OPTIMIZATIONS
 - none that I can think of right now
 - run the profiler


MINOR "BUGS" (alias FEATURES)
 - no way to clear "boolean_attribute_value".
 - <style> and <script> do not end with the first "</".


MSIE bug compatibility
 - <script><!-- ... '</script>' ... --></script>
                     - similar to the behaviour for end
                       tags, MSIE apparently has some kind
                       of quoted text skipping behaviour.
	               Unclear if it is really JavaScript
                       aware.