Re: [exim] trusted_senders question

Top Page
Delete this message
Reply to this message
Author: Stephen Gran
Date:  
To: exim-users
Subject: Re: [exim] trusted_senders question
On Fri, Nov 19, 2004 at 01:20:35PM -0500, Stephen Gran said:
> Hello all,
>
> I have a strangeness I'd like to resolve, but I'm not sure how to start
> debugging. We use a webmail app on one of our systems, and it can only
> inject email via the binary. It does use -f at least, and we have added
> the user that apache runs as to trusted_users.
>
> So, this what we have:
> /usr/sbin/exim4 -bP | grep trusted
> trusted_groups =
> trusted_users = uucp:www-data:list
> untrusted_set_sender =
>
> However, when a user makes a typo sending an email, and ends up with an
> invalid address in the to: we get the following behavior:
>
> 2004-11-19 12:46:34 1CVCqA-0007mJ-Jx <= www-data@??? U=www-data P=local S=8548 id=1100886394.419e317a6c896@??? from <www-data@???> for jusjoken@???
> 2004-11-19 12:46:37 1CVCqA-0007mJ-Jx => jusjoken@??? R=dnslookup T=remote_smtp S=8822 H=mail.lincsatmail.com [207.179.143.254] C="250 2.0.0 iAJHka6w015326 Message accepted for delivery" QT=3s DT=2s
> 2004-11-19 12:46:37 1CVCqA-0007mJ-Jx Completed
>
> 2004-11-19 12:46:34 1CVCqA-0007mK-Ld <= <> R=1CVCqA-0007mJ-Jx U=Debian-exim P=local S=9210 from <> for www-data@???
> 2004-11-19 12:46:34 1CVCqA-0007mK-Ld => linforce (root@???) <www-data@???> R=procmail T=procmail_pipe S=9356 QT=0s DT=0s
> 2004-11-19 12:46:34 1CVCqA-0007mK-Ld Completed


Just a follow up.

I have found that adding -odi to the daemon's command line results in
better behavior, but still not perfect. Now what we get for user typos
is one DSN to the user, and one to www-data. Now at least users are
getting their DSN's, but so is the admin. I can probably hack up the
ACL to stop that, but is this the desirable behavior? It seems not.

Thanks all,
--
--------------------------------------------------------------------------
|  Stephen Gran                  | Gumperson's Law:  The probability of a  |
|  steve@???             | given event occurring is inversely      |
|  http://www.lobefin.net/~steve | proportional to its desirability.       |

--------------------------------------------------------------------------