Re: lynx-dev Radio buttons - cosmetic bug?

[Klaus Weide]

On Tue, 6 Jul 1999 address@hidden wrote:

> > Using Lynx 2.8.1dev.4 on WinNT4, I have a minor oddity in the 
> > rendering - there is a line break after the first button in a group.

Is there an earlier version (after the 'new', aka SortaSGML parsing was
introduced somewhere in 2.7.n) that doesn't have the problem?  I assume
there isn't, but if there is I'd like to know.

> > Not all the groups of buttons on the page show the effect. I see no 
> > reason for it in the HTML. The page validates as HTML4.0 and appears 
> > OK in other browsers. 
> >  
> > The problem page is <http://www.wolverhampton.gov.uk/compform.htm>. 
> >  
> > Note - the page contains lots of seemingly redundant <DIV>s and  
> > <SPAN>s. These are workarounds for bugs in Netscape's handling of CSS  
> > in forms. I haven't checked whether they have anything to do with the  
> > problem, but I do need them there for the benefit of Netscape users. 
> >  
> > -- 
> > Peter Thomson 
> 
> This is what I see with 2.8.2dev.3 (I don't see what you are talking about)
> [ snip ]
>    Gender: Female  (_)
>    Male  (_) not specified  (_)

That's one example of what he's talking about.  If you toggle
SortaSGML off with ^V (or ^V^V when in one of the text fields...),
it becomes one line.

You can also start lynx with -tagsoup or put TAGSOUP:TRUE in lynx.cfg.

Yes, it's a bug in Lynx...  the DTD info is not only only 'sorta' a DTD,
in the case of LABEL at least it's also clearly wrong.  I'll send a patch
for HTMLDTD.c to lynx-dev after I figure out what should be there instead
of what is there currently for LABEL and FIELDSET.

You can start lynx with -preparsed, then look at the source with '\'
to see what lynx makes of the markup which it (erroneously) thinks is
invalid.  You can even use the ^V toggle in source view to note the
differences.

You can look at trace output for the page (-trace, or ^T toggle),
and look at lines with " <- " (or actually "<TAB><- " when using looking
at Lynx.trace with an editor).  Do they make sense?  No, in this
case pretty much everything indicated by these lines about LABEL
is wrong.  (FIELDSET too, it appears.)

   Klaus