On Fri, Feb 18, 2005 at 09:08:09AM -0500, Larry W. Virden wrote:
> I'm seeing a rather peculiar behavior and I was wondering
> if anyone else has seen it.
> 
> The most frequent site I visit via lynx is http://wiki.tcl.tk/ ,
> a wiki that deals with the Tcl programming language. 
> 
> In the past week I've begun to see a really weird behavior.
> If I access the site in the traditional manner, I get random
> but frequent http 400 error messages.
> 
> If I attempt to access the site via Mozilla, I don't see these
> errors.
> 
> But, even weirder, if I access the site with lynx -trace , I don't
> see the errors.

It's possible (haven't seen any recent examples) that the trace code is
initializing things as a side-effect.  More common is for the tracing to
slow things down so timing problems are not as apparent.  (Mozilla is
certainly slow, so that's consistent behavior).
 
> With the -trace flag, I do see some lynx output going to stderr 
> instead of the Lynx.trace file:
> /home/lwv28/i/src/Unix/lynx/t/lynx2-8-6/src/LYMain.c: 4074: 
> parse_arg(arg_name=http://wiki.tcl.tk/4, mask=2, count=2)
> /home/lwv28/i/src/Unix/lynx/t/lynx2-8-6/src/LYMain.c: 4121: parse_arg 
> startfile:http://wiki.tcl.tk/4

That's normal.  Lynx doesn't open the trace file right away, so some useful
information gets written to stderr, which is the default stream for this.

-- 
Thomas E. Dickey
http://invisible-island.net
ftp://invisible-island.net


_______________________________________________
Lynx-dev mailing list
[email protected]
http://lists.nongnu.org/mailman/listinfo/lynx-dev

Reply via email to