Logged In: YES 
user_id=356893

Generally speaking, you are better off conditioning random
junk pulled off the web (with uTidylib or similar) before
feeding it to HTMLParser, which tends to report errors when
it finds them.

See: http://www.w3.org/TR/html4/appendix/notes.html#h-B.3.2.1
for an explanation of why the error message is strictly correct.

Someone may step in with a patch to make HTMLParser more
tolerant in this case; there will always be something else
though.