[905] in SIPB_Linux_Development

home help back first fref pref prev next nref lref last post

Changes to sendmail.cf for local accounts

daemon@ATHENA.MIT.EDU (Sam Hartman)
Mon Jan 16 12:53:56 1995

To: linux-dev@MIT.EDU
Cc: hartmans@MIT.EDU, hoffmann@MIT.EDU
Date: Mon, 16 Jan 1995 12:53:36 EST
From: Sam Hartman <hartmans@MIT.EDU>

	
	Yesterday, I was messing around with sendmail trying to get it
to handle mail to local accounts on my system properly.  I know that
several solutions have been used by different people: modifying the
SIPB foo-server configuration file, having all addresses appear to be
local, or just doing the wrong thing and having the address appear to
be at mit.edu.  I got my system working and addressing mail the way I
like, but in the process, found out from Ron Hoffmann that several
others had tried to do the same, or were asking accounts for mailing
lists so that the local accounts would work right.  He suggested that
if I had something that would be useful to others, I send it here.

	The way I have things set up to work is that all accounts
default to appearing from/sending to user@mit.edu.  However, if a
local name is in /etc/sendmail.local, then the account appars
from/sends to user@machine.mit.edu.  Of course, root, postmaster,
mailer-daemon, and all your local users can go in /etc/sendmail.local.
The only misfeature is that local aliases kind of need to go there
too, if you want them to continue working *all* the time; if they
never get replied to, they don't have to be placed in
/etc/sendmail.local.

	Another potential problem with the way I did things is that my
sendmail.cf is not  a derivative of any of the MIT sendmail.cf files,
but rather, was done using the m4 package that comes with Sendmail 8,
combined with a few custom tweaks to get it working right.  I.E.  I
don't have UUCP or Bitnet mail set up quite right.  If I knew the
relay hosts, I could fix this, though.  I'm not an expert Sendmail
hacker by any means, but I do have a basic grasp of what I'm doing,
and what code the m4 package I used produces; I've looked at the final
product and tested it for most of my uses.

	If anyone would be interested, or if anyone wants to suggest a
cleaner solution, I would be happy to package up the file with some
simple instructions for using it.

--Sam

home help back first fref pref prev next nref lref last post