Received: from localhost (daemon@localhost) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id HAA02282; Fri, 29 Mar 1996 07:25:50 -0500 Received: by CS.UTK.EDU (bulk_mailer v1.4); Fri, 29 Mar 1996 07:23:59 -0500 Received: from mailhost.pipex.net (pp@mailhost.pipex.net [158.43.128.3]) by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id HAA02131; Fri, 29 Mar 1996 07:23:57 -0500 Received: from pipe.pipex.net (actually mailhost.pipex.net) by pipe.pipex.net with SMTP (PP); Fri, 29 Mar 1996 12:23:19 +0000 To: drums@cs.utk.edu Subject: Re: Headers and agents In-Reply-To: <65knI4UEcsB@khms.westfalen.de> References: <199603272242.RAA23893@wilma.cs.utk.edu> <19960327225504.23010.qmail@koobera.math.uic.edu> Date: Fri, 29 Mar 1996 12:23:17 +0000 From: Tim Goodwin Message-ID: > > > I suppose it wouldn't hurt to say that mail with return-path <> should > > > not be forwarded to lists. > > > > That would be a different type of disaster---bounce messages > > disappearing without a trace. > > No no no. Nobody said anything about disappearing without a trace. Consider a message where the originator address is a mailing list. If such a message bounces, the proposed rule means that the bounce must either be rerouted to an administrator address, or dropped. Yet there is absolutely no reason why the bounce should not be forwarded to the list, provided the null return-path is preserved. It is a mistake at any protocol level to base routing decisions on the source address. Of course, messages from <> should not be forwarded to a discussion list like DRUMS, but never appears as an originator address, so such a message could only be created by broken software. Mailing lists have wider applications than discussion lists. > > The right solution is to say that an > > envelope sender of <> must not be rewritten. I concur. Tim.