- From: David W. Morris <dwm@shell.portal.com>
- Date: Sun, 25 Feb 1996 18:36:31 -0800 (PST)
- To: http working group <http-wg%cuckoo.hpl.hp.com@hplb.hpl.hp.com>
On Thu, 22 Feb 1996, BearHeart / Bill Weinman wrote: > I would like to suggest a couple of things for the persistent > session spec: > > 1) A specification for a CGI variable that would indicate > to a CGI program that a session is persistent. Perhaps called > SERVER_CONNECTION it would have the content of the *server's* > Connection: header for the last transaction. Using this with the > HTTP_CONNECTION variable would allow a CGI program to determine > if it's dealing with a persistent connection. > > This information would be useful for CGI state management. > > 2) A method for a CGI program to originate request for a > persistent connection. Simply make the protocol bi-directional. > If the server originates a "Connection: persist" to the client, > the client may respond back with "Connection: persist" and the > exchanges could continue as already specified. These suggestions would only make sense if the persistant connection is directly between a end user client and server. They immediately break if a proxy is involved. I am assuming that the latest draft hasn't changed completely the earlier sense of persistent connections which makes the UA->proxy and proxy->server connection persistentcy independant. If it has, then I'll have more to say. Dave Morris
Received on Sunday, 25 February 1996 18:39:44 UTC