[comp.emacs] GNU Emacs vs. sendmail - battle of the titans

montnaro@sprite.steinmetz (Skip Montanaro) (11/30/87)

Internet mail to our local network of Suns goes through a VMS VAX which adds
a little something extra to return addresses. For instance, zip@zap.domain is
transformed into zip%zap.domain.tcpip@csbvax.steinmetz. Undoing this in the
sendmail.cf file is pretty easy, but sendmail only generates a From field in
the correct format, it doesn't rewrite the From: field. On the other hand,
GNU Emacs discards the From field (in rmail-nuke-pinhead-header, I believe),
preferring to display the From: field, which defeats my attempts at making
mail headers more readable (and useful to the mail software running on our
Suns).

I can patch the rmail.el code to pay attention to From, however, the business
of which field (From or From:) is better, preferable, or more authoritative
has always been somewhat perplexing to me, and I'd like to get it straight
before butchering things. What's the straight dope? Can I make sendmail
rewrite the From: field? Can somebody from the Free Software Foundation
explain rmail's behavior? 

I imagine this is either old hat or uninteresting to most readers of these
groups, so reply directly to me and I will summarize if there is enough
interest. 

Thanks,

Skip (montanaro@ge-crd.arpa or uunet!steinmetz!sprite!montanaro)

davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr) (12/01/87)

In article <8013@steinmetz.steinmetz.UUCP> montanaro@ge-crd.arpa (Skip Montanaro) writes:
| Internet mail to our local network of Suns goes through a VMS VAX which adds
| a little something extra to return addresses. For instance, zip@zap.domain is

One of the other burning questions is, should the "Reply-to:" field
override all others and be used instead of any from field. That would
solve your problem, if only everybody's mailers did the same thing.
-- 
	bill davidsen		(wedu@ge-crd.arpa)
  {uunet | philabs | seismo}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me

markl@PTT.LCS.MIT.EDU (12/04/87)

In-Reply-To: davidsen@steinmetz.steinmetz.UUCP's message of 1 Dec 87 16:21:47 GMT

Repository: PTT

Originating-Client: thyme



   From: davidsen@steinmetz.steinmetz.UUCP (William E. Davidsen Jr)
   Keywords: GNU Emacs, sendmail
   Date: 1 Dec 87 16:21:47 GMT
   Reply-To: davidsen@crdos1.UUCP (bill davidsen)
   Distribution: na

   One of the other burning questions is, should the "Reply-to:" field
   override all others and be used instead of any from field. That would
   solve your problem, if only everybody's mailers did the same thing.
   -- 
	   bill davidsen		(wedu@ge-crd.arpa)
     {uunet | philabs | seismo}!steinmetz!crdos1!davidsen
   "Stupidity, like virtue, is its own reward" -me

Depends on whose brand of standard you like.  If you follow the
Internet RFC-822 standard, there is a specific hierarchy of fields to
reply to.  It is, in priority order, "reply-to:", followed by "from:",
followed by "sender:" (if present).  

markl

Internet: markl@ptt.lcs.mit.edu

Mark L. Lambert
MIT Laboratory for Computer Science
Distributed Systems Group

----------

cfe+@andrew.cmu.edu.UUCP (12/08/87)

RFC-822 does indicate a standard order of fields to reply to, but it isn't all 
three of those.  It's to use the Reply-To: field, if present; otherwise, the 
From: field.  You never reply to a Sender: address.

What I mean by ``reply to'' is that when a user asks to reply to a message 
(say, X), then the user's mail agent may assist the user by providing values 
for the fields in the header of a new message.  Typically, the To:, Subject:, 
and In-reply-to: (if not also References:) fields are derived from the fields 
of message X.  When the mail agent provides a value for the To: field, it 
should use the value that was in the Reply-to: field of message X, or the 
From: field of message X in case message X had no Reply-to: field.  (Whether 
the user agent makes use of the contents of the To: or CC: fields of message X 
is pretty much up to the user agent.)

This case is different from a system's automatically returning an error 
message to the sender of a piece of mail.  In that case, RFC821 and RFC822 
technically disagree, but what it's come down to is that you return the mail 
to the argument of SMTP's MAIL FROM: command, if you got the mail via SMTP.