lynx-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

lynx-dev Mis-numbering of links...


From: David L. Potter
Subject: lynx-dev Mis-numbering of links...
Date: Tue, 24 Nov 1998 09:47:56 -0400 (AST)

Our version of Lynx... (which we call)

        Lynx Version 2.7.1ac-0.102+intl+csuite

produces a rather curious result in some rather narrow circumstances. The 
behaviour was first detected on a Chinese Language page... and has 
subsequently been duplicated on an English page... a screen print of 
which is below... as a start you'll notice there are two links 
numbered '[3]'...

======================================================================

                                   Bug Demo.
     _________________________________________________________________

   [1] xxxxx_this_is_length_critical_xxxxxxxxxxxxxxxxxxxxxxxxxx???? |
   BOFH [3] [3] [4] [=||=] 
     _________________________________________________________________
     
   Above, the link number for link # 2 is missing. Any of the "?"s may be
   deleted from the "length-critical" string and the bug persists. Add
   one more character to the string or delete more than four characters
   and the bug disappears. The problem seems to be that lynx gets
   confused if:
    1. a link number is to be at the end of one line with the label 
       wrapped to the line below that and
    2. a link following the link to be wrapped has no label specified and
    3. the link with no label is on the same line as the wrapped label.
                
   Also, some link numbers get duplicated and the first duplicate in a
   pair seems to have its URL listed as "hidden" with the others listed
   as "visible".

======================================================================

I'm curious if this rendering is consistent with the latest version of 
Lynx...

Here is the url of our test page:

        http://www.chebucto.ns.ca/~af380/bug-demo.html

Here is the original Chinese Page:

        http://www.cnnic.cn/cgi-bin/domainqc

----

Thanks to Norman DeForest who did all the work figuring out what was 
happening...


david potter



reply via email to

[Prev in Thread] Current Thread [Next in Thread]