Author: snowcrash+exim-users Date: To: Heiko Schlittermann, Ted Cooper CC: exim-users Subject: Re: [exim] using aux acls -- why is control not passing correctly?
hi,
On 6/9/07, Ted Cooper <elc@???> wrote: > As soon as you accept, no further ACLs are looked at. It looks like if
> you don't find a virus, you then hit the accept target. Try changing
> your logic to DENY or DEFER. With the deny ACL, you will not have to use
> endpass.
Yup. You're right. Works nicely (well, so far ...)
I'm fairly certain I had a 'great' reason for choosing the
accept/endpass approach here. But, for the life o' me atm, I dunno
why :-}
On 6/10/07, Heiko Schlittermann <hs@???> wrote: > > i've Exim 'talking' to remote clamd/spamd daemons over TCP,
> >
> > av_scanner = clamd: 10.0.0.105 3310
> > spamd_address = 10.0.0.105 783
>
> Both do not resend the mail, they just evaluate, right?
> The message is not modified using av_scanner and/or spamd.
That is correct. Well, at least that's the _intention_. All header
mods are done in the calling ACL within Exim.
> May be I'm wrong, but if a (clean and nice) message comes in from let's
> say <pitti@???> it is accepted because the sender is not in the
> no_virus_scan list and the aux_acl_virus acl accepts it. I do not see
> any reason why the next acl rule should be asked.
>
> Only if the message comes from +no_virus_scan it hits the next acl.
It seems I got the operation of the endpass bass-ackwards. Swear I
read that silly passage in the book a bunch o' times ...
