From ross at biostat.ucsf.edu  Thu Sep  8 22:04:00 2005
From: ross at biostat.ucsf.edu (Ross Boylan)
Date: Thu, 8 Sep 2005 22:04:00 -0700
Subject: [SA-exim] distinguishing fetchmail from others
Message-ID: <20050909050400.GE3502@wheat.boylan.org>

I'd like help using sa-exim with fetchmail.  Recent network
alterations have forced me to use fetchmail, which I know is not a
natural combination with acceptance-time scanning.  I'm hoping it's
temporary.

My old rules didn't scan any locally originating mail, but everything
coming through fetchmail appears local.  Something's got to change if
I'm to have any spam checking.  I need some way to distinguish
fetchmail from other sources of email.  Further, within the mails
retrieved by fetchmail some will be from me (as I cc myself), and I'd
like to avoid scanning them.

Does anyone have any suggestions about how to accomplish this?

My current thinking is that I need to tell exim4 to listen on a second
port (umh, is there some natural choice?) and have fetchmail deliver
to that port.  I think the port will be available for SAEximRunCond.

BACKGROUND

I'm running sa-exim on a Debian testing (aka etch, basically same as
just released sarge) system. sa-exim 4.2-2; exim4 4.52-2.

FALSE LEADS
Here are some approaches I don't think will work.

fetchmail provides no easy way to add a header that would tag the
email (I seem to recall it used to have this feature, but don't see it
now, in 6.2.5).

I think sa-exim runs before exim routers, so having routers add
headers also won't work.

I think that the application of SAEximRunCond can not itself be made
conditional on anything (e.g., port of incoming message).  That is,
the condition can test the port and expand to false, but I can't
control whether or not the condition is run or the sa-exim scanner is
used at all.

I'm assuming sa-exim operates using exim's local scann function hook.
Right?

Thanks for any help.
Ross Boylan


From jon at recipexperience.com  Tue Sep 27 02:21:23 2005
From: jon at recipexperience.com (Jon Pentland)
Date: Tue, 27 Sep 2005 10:21:23 +0100
Subject: [SA-exim] sending the -u option to spamc
Message-ID: <001301c5c344$d47bc200$ee00a8c0@JONPC>

After extensive searching I haven't been able to find a solution to a
particular *problem* i'm having.

I wish to run a multi-domain webserver and have separate preferences for
each domain (AFAIK spamassassin requires local user accounts as opposed to
the forwarding I require). As such it would be perfect if I could keep a
list of domains and have sa-exim send the -u option to spamc dependent on
these domains.

My knowledge of c is minimal, but I feel something like this would be 
possible by keeping a list of domains and users which are checked before 
sending the appropriate -u option. This would allow me to keep a set of 
options for each domain - which would be more than enough for my purposes.

My apologies if this has been answered before anywhere or if there is a
painfully simple way of doing this, but any light shed on the
subject would be fantastic.

Regards,

Jon Pentland