Mail and Resent-fields (was: Re: HELP!)

John C Klensin <KLENSIN@infoods.mit.edu> Mon, 09 August 1993 12:23 UTC

Received: from ietf.nri.reston.va.us by IETF.CNRI.Reston.VA.US id aa01018; 9 Aug 93 8:23 EDT
Received: from CNRI.RESTON.VA.US by IETF.CNRI.Reston.VA.US id aq00886; 9 Aug 93 8:23 EDT
Received: from venera.isi.edu by CNRI.Reston.VA.US id aa03217; 7 Aug 93 19:50 EDT
Received: from INFOODS.MIT.EDU by venera.isi.edu (5.65c/5.61+local-12) id <AA21284>; Sat, 7 Aug 1993 16:41:16 -0700
Received: from INFOODS.UNU.EDU by INFOODS.UNU.EDU (PMDF V4.2-13 #2603) id <01H1GY4V1BTS0000XU@INFOODS.UNU.EDU>; Sat, 7 Aug 1993 19:40:52 EDT
Date: Sat, 07 Aug 1993 19:40:52 -0400
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: John C Klensin <KLENSIN@infoods.mit.edu>
Subject: Mail and Resent-fields (was: Re: HELP!)
In-Reply-To: <199308072054.AA28328@can.isi.edu>
To: braden@isi.edu
Cc: ietf-hosts@isi.edu
Message-Id: <744766852.196433.KLENSIN@INFOODS.UNU.EDU>
X-Envelope-To: braden@ISI.EDU, ietf-hosts@ISI.EDU
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; CHARSET="US-ASCII"
Content-Transfer-Encoding: 7bit
Mail-System-Version: <MultiNet-MM(330)+TOPSLIB(156)+PMDF(4.2)@INFOODS.UNU.EDU>

>Note that he DID NOT just forward ("resend") my message, he wrote his
>own message and appended the text of my message.

Well...

There are a number of ways to get the RiceMail system to decide it is
resending, rather than replying.  I don't recall the details, although I
would be happy to hunt them up after INET.

The real problem is with the definitions of the Resent- fields, and
there are many such problems.  RFC-822 more or less says "here is the
syntax", but attaches few semantics and fewer rules.  My recollection of
the ancient history is that Dave didn't like the whole idea, but some of
"the committee" did, but Dave did the writeup, and...

The net result is that it is impossible to look at 822 and make a claim
that a Resent- field is being used improperly.  So we have:

  (i) Should RiceMail be using Resent- in this case (whatever this case
      is)?
  (ii) Should mail exploders that are acting at UA level (rather than
      being MTA/envelope-only exploders) be adding Resent-fields?
  (iii) If a user sets a personal forwarding address on the system and
      mailbox associated with an externally-known address, should a
      Resent-to field be applied in the forwarding process?
  (iv) In cases (ii) and (iii) above, should Resent-From be supplied,
      and, if so, how should it be set?
  (v) Is a mail system that rejects or trashes mail when Resent-To
      appears without Resent-From (or Resent-Date) behaving correctly?
  (vi) If a message arrives with From, To, Reply-to, and Resent-From,
       should automated replies go to Reply-to or Resent-From?
  (vii) If a message arrives with From and Resent-From (but not Reply-to
       or Resent-Reply-To), should automated replies go to From or
       Resent-From?
  (viii) If a UA receives a message containing a Resent-From line, what
       should it show in its table of contents or equivalent listing?
  
There are deployed implementations corresponding to all possible answers
above, and people willing to defend their choices to the death (and cite
822 text as the reason for them).

Proposed answers coming soon (if I ever dig out of the IESG swamp) to an
I-D near you.

>Isn't this wierd?  If so, how can it have gone unnoticed for so long?

  Yes.  But some of us have noticed, more's the pity.

    john