On 4/29/06 8:00 AM, "Stephen J. Turnbull" <stephen@xemacs.org> wrote:
Sender doesn't instruct *conformant* MTAs at all, does it? AFAIK the only thing that a RFC 2821-conforming MTA looks at is the Return-Path header, and it's supposed to remove that.
There is no Return-Path: header during transmission of a message. The Return-Path header is added in the process of delivery. There is a return path, stated in the MAIL FROM:<example@example.com> SMTP command. (That command *can* have more stuff related to authentication.) The return path is what should be used as the address of a bounce if a mail system foolishly accepts a message and then creates a bounce.
Notice that if an MTA rejects a message (or one or more of the recipients of the message), it is not bouncing or creating a bounce. It is issuing an error response...the MTA (or MUA in the case of message submission) that was trying to send creates a bounce message if appropriate (for message submission, the MUA notifies the user--or pretends to: Microsoft by default hides the notification remarkably well).
While multi-line text associated with the rejection code is provided for, MUAs are very poor about showing it if a submission is rejected--some show only the first line; some only the last line. Even some MTAs "improve" the text of the rejection.
--John