Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id LAA21965; Tue, 15 Aug 1995 11:30:04 -0400 X-Resent-To: drums@CS.UTK.EDU ; Tue, 15 Aug 1995 11:30:00 EDT Errors-to: owner-drums@CS.UTK.EDU Received: from mastodon.CS.Berkeley.EDU by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id LAA21946; Tue, 15 Aug 1995 11:29:58 -0400 Received: from mastodon.CS.Berkeley.EDU (localhost.Berkeley.EDU [127.0.0.1]) by mastodon.CS.Berkeley.EDU (8.6.11/8.6.9) with ESMTP id IAA07758; Tue, 15 Aug 1995 08:31:01 -0700 Message-Id: <199508151531.IAA07758@mastodon.CS.Berkeley.EDU> To: perry@piermont.com From: Eric Allman cc: ietf-drums Subject: Re: "Reply-To" In-reply-to: Mail from "Perry E. Metzger" dated Tue, 15 Aug 1995 11:19:23 EDT <199508151519.LAA23133@panix4.panix.com> Date: Tue, 15 Aug 1995 08:31:00 -0700 Sender: eric@mastodon.CS.Berkeley.EDU "My own" mailer (by which I assume you mean sendmail) makes no such restriction and never has: it restricts you from setting the _envelope_ from address, but not the _header_ from address, which is what we are talking about. Obviously, the Reply-To: header has no impact at the envelope level. Given that the standards already make it clear that the From: _header_ can be set by the user (and we are talking about standards here -- getting the implementations (typically of user agents, not MTAs) in line is another matter entirely. I am simply agreeing with Robert: the standards already permit what Jacob originally proposed, and a more useful semantic (to my mind, at least -- obviously not to yours) is to override, per the mailing list example I first used. eric ============= In Reply To: ============================================== : From: "Perry E. Metzger" : Subject: Re: "Reply-To" : Date: Tue, 15 Aug 1995 11:19:23 -0400 : : Eric Allman writes: : > I'm not arguing that there are many _possible_ semantics -- but as : > Robert pointed out, you CAN change your From: line to have the : > appropriate address (which would work even for UAs that don't grok : > Reply-To) -- just because you don't choose to do it this way doesn't : > mean you couldn't. : : It isn't entirely a matter of choice -- most modern MTAs like your own : only trust certain users to set the From: address. As such, most : ordinary users aren't in a position to do this in order to assure : appropriate handling. : : What *might* work nicely would be to have such MTAs switch to a policy : of inserting a Sender: line if the From: has been set and otherwise : permit arbitrary munging of the From: : : .pm