Re: lynx-dev Cookie bug in lynx2.8.5dev.2

[Thomas Dickey]

On Mon, Oct 22, 2001 at 08:13:37PM -0700, Doug Kaufman wrote:
> On Mon, 22 Oct 2001, Thomas Dickey wrote:
> 
> > I think that was this change:
> > 
> > * cookie path= should be a prefix of the request-URI path, so do not 
> > truncate
> >   request-URI path in LYSetCookie() [we got mistaken "invalid cookie 
> > path=..."
> >   prompt in some cases previously] -LP
> 
> So, if I understand this correctly, the change was made to avoid
> "invalid cookie path" errors, but now some sites are completely
> inaccessible to lynx because of the change. It seems that some servers
> have incompatible behaviors. Do we know if one of these behaviors is
> "correct"? Does anyone see a workaround to allow lynx to function with
> both types of server?

not offhand (perhaps LP is still subscribed, and can confirm if I'm looking
at the right chunk, etc).  The only other changes were for LV's problems,
but that context looks like the one I modified for LP's patch.
 
> If the previous problem was a warning, but the site still functioned,
> I would advocate reverting to the old behavior. If the site didn't
> function with the "invalid cookie path" warning, then further
> discussion of the appropriate behavior is indicated. Can we tell which
> types of servers demand which cookie paths and then customize for
> them? The server that wouldn't work with the current lynx behavior is
> identified as "Oracle_Web_Listener/4.0.8.2.3EnterpriseEdition".
> 
>                               Doug
> __ 
> Doug Kaufman
> Internet: address@hidden
> 
> 
> ; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden

-- 
Thomas E. Dickey <address@hidden>
http://invisible-island.net
ftp://invisible-island.net

; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to address@hidden