- From: Benjamin Franz <snowhare@netimages.com>
- Date: Wed, 11 Jun 1997 06:43:11 -0700 (PDT)
- To: www-talk@w3.org
On Tue, 10 Jun 1997 S.N.Brodie@ecs.soton.ac.uk wrote: > Benjamin Franz wrote: > > > > Accepting the cookie. The server seriously wants that cookie. It will not > > stop looping until it gets it. :( > > ARGH! > > Thanks for the information. Does NS/IE suffer the same problem if you > disable cookies in them too, then? Yep. Remember the stateless nature of the web. Without the cookie or URL state - the server ALWAYS believes you are a new person hitting it - and so it tries to get you to accept a cookie. In my test, MSIE prompted me for three cookies (which I declined ), and then prompted me for the first one again: Clearly it was looping. I haven't had a chance to test it with NS4.0 with silent cookie rejection, but I expect either a crash, a silent infinite loop or an error. (I can't test it right now because NS4.0Pro final is crashing every time I launch it - it doesn't seem to like my moving back and forth between network connections. Seems a bit...er...buggy for a final release.) Personally, I hope it goes into a silent infinite loop. On a high speed link that would result in an occasional punishing load on the offending 'accept my cookie or else' server as NS4.0 deploys more and you get people who turn off cookies altogether visiting. Its ok to use redirects to set cookies on the same server - but good design accounts for the possibility that the cookie might *not* be accepted by adding URL state as a fallback catch to prevent *exactly* this kind of loop. -- Benjamin Franz
Received on Wednesday, 11 June 1997 09:43:13 UTC