[Top][All Lists]

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

Re: lynx-dev yahoo cookies not working the same in 2.8.3.dev15

From: Klaus Weide
Subject: Re: lynx-dev yahoo cookies not working the same in 2.8.3.dev15
Date: Tue, 23 Nov 1999 21:18:04 -0600 (CST)

On Tue, 23 Nov 1999, Larry W. Virden wrote:

> Now I have the trace files.  What _appears_ to be happening is either the
> trace output is not outputing enough information, or, the new code is
> not getting all the cookie info - 

Or, a third possibility, the server is not sending the same info.
Discriminating based on some headers (.dev vs .rel? although that
seems unlikely).

> in the trace, I see:
> 2153c726
> < User-Agent: Lynx/2.8.2rel.1 libwww-FM/2.14
> ---
> > User-Agent: Lynx/2.8.3dev.15 libwww-FM/2.14
> 2156c729
> < Cookie: B=0rguq6g8mhrtu; 
> Y=v=1&n=1scejighn0oed&l=bl8h34d/o&p=m1n1a1s491030g&r=
> 01&lg=us; 
> T=z=6FxO4A6LGP4AtKtjB06V4KxNgY2MTc1NjE0MTY-&a=AAE&sk=DAADmo24uDGX1p&d=
> c2wBTVFFeE5qQXlNVFl6TmpFLQFhAUFBRQF6egE2RnhPNEFnV0E-; YM.State=YM%3d; 
> YM.Pref=fa
> 6fwd%3dattach%26fontsz%3dnormal%26msgwidth%3d82%26order%3ddown%26inc%3d200%26got
> o%3dmsg
> ---
> > Cookie: B=cknhieitsv5hk

That's what lynx is sending (as I'm sure you understand).  There should be
differences *before* that, in what lynx is *receiving*.  Well I assume
the above isn't the first request, in a chain of redirections.

The trace output isn't good at showing the raw response headers received,
in the cae of redirection responses.  But you *should* find lines like

   LYSetCookie called with [...]
   LYProcessSetCookies: [...]
   LYmktime: Parsing [...]
   LYProcessSetCookie: [...]
   store_cookie: Rejecting because [...] 

Look for differences in those.

If the request headers you quote (above) *are* for the very first request,
then there must be a difference in either the cookie file's contents, or
in the way lynx reads in and stores that contents initially.  (You should
be able to locate that in the trace log.)

> There are so many differences in the two traces that it becomes a struggle
> for me to figure out what's important.  The old trace is over 4000 lines
> long, while the new one is less than 2000 ....

Well, you have to find the *first* important difference (as you probably

Are you sure those two lynx invocation were starting out in a
comparable 'state'?  (as far as config settings; as far as contents of
your cookie files; and as far as previous interaction with the server

Actually, you should be able to find the first divergence of behavior
without wading through trace logs.  Start two lynx sesssions in parallel
(presumably using screen, or maybe switching with ^Z/fg), one of each
version.  Use -noredir.  [Well, first you should probably verify that
the whole thing works with -noredir at all.]  There will be no automatic
following of redirections, but you should be able to follow them as
links.  [If they provide those in the redirection responses.]  At each
step, i.e. after each redirection response, check the Cookie Jar pages
before proceeding.  When does the first difference show up?

Variant of the above (and you should also try this just 'normally'):
Disable all 'automatic' accepting of {cookies,invalid cookies}, but
also automatic rejecting, try to get prompted for everything. (And don't
answer with 'A'lways.)  You should notice where things start getting

Then start to concentrate on the now-narrowed-down area with trace logs,
packet sniffing, ...


reply via email to

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