🚀 go-pugleaf

In article <1992Jul27.024211.20761@ryptyde.cts.com> scott@ryptyde.cts.com (Scott McClure) writes:
>
>Which is the correct field for a mailer to use when getting the address
>from a message to reply to it:
>
>Return-Path:   *or*   From:
>
>Seems to me that the Return-Path is often unreliable and should be
>avoided, but there still seem to be sites on the net that use this.

  You are supposed to use "Reply-To:" if there is one, otherwise "From:".
Don't ever use "Return-Path:".  It is highly unreliable.

Neil Rickert <rickert@mp.cs.niu.edu> writes:
|
|   In article <1992Jul27.024211.20761@ryptyde.cts.com> scott@ryptyde.cts.com (Scott McClure) writes:
|   >
|   >Which is the correct field for a mailer to use when getting the address
|   >from a message to reply to it:
|   >
|   >Return-Path:   *or*   From:
|   >
|   >Seems to me that the Return-Path is often unreliable and should be
|   >avoided, but there still seem to be sites on the net that use this.
|
|     You are supposed to use "Reply-To:" if there is one, otherwise "From:".
|   Don't ever use "Return-Path:".  It is highly unreliable.

Perhaps a little explanation of why it's unreliable would do.  The cause
is, as almost always when we discuss malfunctioning mail software, a
virus known as "sendmail" which has infected a number of systems using
BSD Unix, and has yet to be given the proper attention, i.e. be removed.

The Return-Path header is supposed to be inserted upon final delivery of
the message, as far as SMTP is concerned.  This may not always be final
delivery to the user, though, but on the Internet, it is.  Now, this
bogus SMTP daemon and mail rapist, sendmail, inserts Return-Path headers
at will and does not obey the standard.  And, which is more, it believes
that when a message has a Return-Path header, it should not replace it
with a good one.

The Return-Path is thus meant to be reflect the envelope sender, as SMTP
sees it, and it should, theoretically, be as reliable as bouncing mail
at the SMTP level to bounce it at the application level.  Not so in
practice.  Thanks to sendmail.

Help stomp out sendmail in our lifetime.

Best regards,
</Erik>

--
Erik Naggum             |  ISO  8879 SGML     |
                        |  ISO 10744 HyTime   |
+47-295-0313            |  ISO 10646 UCS      |      Memento, terrigena.
<erik@naggum.no>        |  ISO  9899 C        |      Memento, vita brevis.

In article <1992Jul27.024211.20761@ryptyde.cts.com>, Scott McClure writes:

> Which is the correct field for a mailer to use when getting the address
> from a message to reply to it:
> 
> Return-Path:   *or*   From:
> 

First choice (if present): Reply-To:
Second Choice: From:

Return-Path: is for sending error mails by the MTA, not the MUA!

Gruss, Christian

-- 
Christian Riede, Lange Strasse 88, D-W7500 Karlsruhe 51, Germany 
Tel. (+49) 721 {886091 | 608-4487 | 608-2749}  |   iso-8859-1:   
EMail: chr@senga.ka.sub.org, riede@ira.uka.de  |  äöüß==aeoeuess 

In article <1992Jul27.024211.20761@ryptyde.cts.com> scott@ryptyde.cts.com (Scott McClure) writes:

: I have heard conflicting reports on this.  Hopefully someone out there
: knows the whole story.
: 
: Which is the correct field for a mailer to use when getting the address
: from a message to reply to it:
: 
: Return-Path:   *or*   From:

The Return-Path field is to be used by the message transport, in the case
errors are detected in routing.  The From:, and Reply-To: fields are to be
used by user agents in constructing default return addresses.   The Return-Path
field usually reflects the SMTP MAIL FROM value, whereas the From: field is
what the original UA constructs.

Regards,

Tim Kehres

There was a period of time long ago during which I found that one or
two intermediate UUCP sites were mangling From: headers quite badly.
During that period of time, I found it convenient to tell MH to send
replies to the return-path field.

Even today, there are a few sites that will change a fully-qualified
domain-form From: address into bang-path form.  Other sites will
happily prepend their name to this bang-path.  Yet others *won't*.  The
final result is what looks like a complete bang-path, but it has some
names missing in the middle.  At that point the mail bounces with
"unknown UUCP host".  (And we started with a nice FQDN, which is
supposed to be Immutable and Always Replyable.)

The reply-path seems to be usually much better behaved.

Fortunately I haven't found it necessary lately to use the reply-path
for replies, but I can see why some people might.
-- 
Rahul Dhesi <dhesi@cirrus.com>
also: dhesi@rahul.net