- From: Dave Kristol <dmk@bell-labs.com>
- Date: Tue, 29 Jul 1997 18:18:42 -0400
- To: Foteos Macrides <MACRIDES@sci.wfbr.edu>
- Cc: http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com, http-state@lists.research.bell-labs.com
Foteos Macrides wrote: > [...] > > The problem is that you're counting on $Version=0; being treated > as a bad cookie by the old server or script, and don't know what error > handling it's using. It indeed would be desireable for the UA to > communicate that it supports version 1 or greater cookie handling in > those cases for which it is using cached version 0 cookies and doesn't > yet know if the server or script can handle new cookies. The simplest, > most efficient way might be to send: > > Cookie2: $Version="1" > Cookie: realoldnameA=realoldvalueA; realoldnameB=realoldvalueB[; ...] > > If it's an old server or script, the Cookie2 request header will be > ignored. Otherwise, the server or script will use Set-Cookie2, the > UA will not send the Cookie2 probe, and use Cookie: $Version="1"; ... > so the updating is mutually know by the State Management partners > with negligable excess network traffic having been expended, and with > both the server/script and UA thereafter enjoyed all the benefits > of the modern State Management protocol. I like this approach. Dave Kristol
Received on Tuesday, 29 July 1997 15:23:41 UTC