lynx-dev Lynx's tracelog: a useless debugging tool?

[Jacob Poon]

When I go to this page:

http://www.geocities.com/Tokyo/3741/text.html

Lynx said it has 'bad HTML'.  Then I debug it with the built-in trace log,
it gives out tons of useless infomation before reaching the 'bad HTML'
error.  Even so, the error message does not clearly describe the exact
problem.  In my case, it gives the following: 

Bad HTML: Something not closed before BODY end tag
SGML: End </HTML>
HTML:end_element: Popped style off stack - Normal
Bad HTML: Something not closed before HTML end tag

I tried to test it with W3C validator and doesn't help either.  Too many
times it says something like this:

Error at line 16:
   <img src=rib_bar_.gif></a>
                           ^  end tag for element "A"
                              which is not open (explanation...)       

Even the validator itself is buggy because the actual source is follows:


<A> ...... <img src=rib_bar_.gif></a>

The W3C validator itself is case sensitive on tag nesting checks, but HTML
tags are not.  Therefore I can't find what bugged Lynx. 

In other words, unless I can choose what Lynx should (not) put into the
trace log, and Lynx describe the errors more precisely, the trace log
feature is rather useless IMO.