[1096] in bugtraq
Re: sendmail testing
daemon@ATHENA.MIT.EDU (Michael R. Widner)
Sat Feb 25 15:14:06 1995
From: "Michael R. Widner" <mmwidner@gsbux1.uchicago.edu>
To: hobbit@avian.org (*Hobbit*)
Date: Sat, 25 Feb 1995 13:20:24 -0600 (CST)
Cc: bugtraq@fc.net
Reply-To: widner@UChicago.EDU
In-Reply-To: <199502240857.DAA03998@narq.avian.org> from "*Hobbit*" at Feb 24, 95 03:57:32 am
Previously, *Hobbit* wrote:
> Anyone who wants to watch the bug in action, and whose sendmail is doing
> the identd lookups, can send a message to himself via the mailserver here.
> For instance, if you're foo@bar.com, send mail to foo%bar.com@avian.org
> and get your sniffer handy. If /tmp/fuckme shows up on your system, most
> likely owned by bin, you've got the problem.
I've been playing around with this bug a little bit. I understand what's
going on, but there is something I'm a bit confused about. I get all the
right stuff in the qf* file and if I have sendmail process it by doing a
sendmail -q I see the expected behavior. But when sendmail processes
the incomming messages immediately it ignores the extra R<whatever> line.
The stuff all shows up in the delivered message, but sendmail doesn't act
on it.
So what I'm confused about is this: What is different about the way
sendmail processes the queue files when it is called with the -q option
and when it is just processing off smtp connections? I see the exact same
behavior on 8.6.9 running on 4.1.3 and Linux.
--
Michael R. Widner
widner@uchicago.edu
