Message-Id: <199707080433.AAA14982@postman.opengroup.org>
To: new-httpd@apache.org
Subject: Re: connection handles 
In-reply-to: Your message of "Mon, 07 Jul 1997 20:48:43 EDT."
             <Pine.LNX.3.95dg3.970707204502.23100V-100000@twinlark.arctic.org>
Date: Tue, 08 Jul 1997 00:33:52 -0400
From: Doug MacEachern <dougm@opengroup.org>
Sender: new-httpd-owner@apache.org
Precedence: bulk
Reply-To: new-httpd@apache.org

Dean Gaudet <dgaudet@arctic.org> wrote:

> Well the proxy uses BUFF * as well.  The only thing I'm really debating is
> whether this belongs in BUFF or in conn_rec.  I'd lean towards
> putting it in conn_rec... but I'm not the one using it :)

conn_rec would be fine too, in terms of time-n-place requirements.
 
> What would happen if you tried both SSL and RPC (or whatever combo) in the
> same server?  How would the layering affect the handle?  Normally this sort
> of thing is dealt with by a vector with an entry for each module.

That's kinda why I called it "transport handle", it might happen that
ssl and rpc are configured into the same server, but I can't see using
more than one transport protocol for a given request (talking to
client (browser) anyhow).  The creation of module per-request configs
currently happens too late to be used for this.  per-server configs
are no good for storing per-request data in a thread model.

-Doug