Re: LYNX-DEV problem page

[Klaus Weide]

Laura,

 First, PLEASE state what version of the code exactly you are using
when you share your observations with us.  Apparently it is not always
the latest number of the development code, sometimes you seem to talk
about fotemods, it gets confusing.

On Thu, 8 May 1997, Laura Eaves wrote:

> Yesterday there were some problems with http://www.msnbc.com/
> 
> Since this page changes every day, the same problems aren't there today,
> but i had downloaded yesterday's page and put it in
> http://www.intac.com/~leaves/msnbc.html
> 
> First, link 4 is null due to the parser closing an anchor early.
> (bad html).  Both parsers seem to null out link 4.

That anchor has a MARQUEE as content.  Lynx treats that specially,
see the Users Guide, see the code.  That interferes with it being in an
anchor.  The "new parsing" closes the A anyway when it sees the MARQUEE,
so it becomes a "hidden link".  The "old parsing" (whose handling of A is
actually newer..) would need an additional check in HTML.c for me->inA
under case HTML_MARQUEE: to recover better from this.

Btw. the same kind of problem seems to exist for all kinds of tags with an
ID attribute if they appear within Anchor text... (not dependent on "which
parser" etc.)

> Also, links 2 and 3 are USEMAPS.  With the old parser, they come out
> [LINK]-[USEMAP], with the new parser they come out [USEMAP] as the
> preceeding anchor is closed.  (In this case, it looks like the new parser
> is right.)

There is an open HREF-less A (which shouldn't lead to a [LINK] string
anyway...)

Note that the "old parser" behavior here is not the same as recent
fotemods behavior.  The latter doesn't show those [LINK] strings either -
as long as BOLD_NAME_ANCHORS:TRUE is not in effect (see lynx.cfg).
For BOLD_NAME_ANCHORS:FALSE fotemods closes HREF-less anchors immediately,
in this case this prevents the inappropriate generation of [LINK].
I didn't include that in the devel code's HTML.c, so you get different
behavior there.

> Finally, this page, and other pages at this site, are sprinkled with
> ‘ and ’ strings.  Shouldn't these be translated into something else?

Yes, of course.  By the authors or authoring tools.  
Only they can know for sure what they mean.  Unless we want to pretend
that all authoring is done with one and the same code page, Microsoft's
CP 1252, and ignore not only that there are things beyond Microsoft
platforms, but also that Windows uses different code pages in different
parts of the worlds.  And then the next versions of those authoring
tools come out which do things correctly anyway (and I believe there is
hope for this). 

  Klaus

;
; To UNSUBSCRIBE:  Send a mail message to address@hidden
;                  with "unsubscribe lynx-dev" (without the
;                  quotation marks) on a line by itself.
;