On Wed, 2004-04-21 at 16:39, Philip Hazel wrote:
> On Wed, 21 Apr 2004, John Horne wrote:
>
> > > Did you try running a test with -bh? and with -d as well?
> > >
> > No. I'll have to rig something up because most of our checks ignore
> > anything within the University.
>
> Er, the whole point of -bh is that you specify an IP address from which
> Exim pretends it has received a connection. Just specify 1.2.3.4 or
> something, and you will be testing from "outside"...
>
Yes, sorry about that - not thinking.
I see the problem now too. In effect what I do is:
(MAIL ACL)
set acl_m7 = 0
check if the sending host or the sender is locally banned or rbl
listed - if it is then set acl_m3
verify the sender *IF* acl_m3 is not set
if the verification fails then set acl_m7
(RCPT ACL)
if acl_m7 = 0 then deny the recipient and say that the sender is
unrouteable.
check if m3 is set and if so deny the recipient saying that the
sending host is locally or rbl banned.
The problem is that if the MAIL ACL sets acl_m3 then the RCPT ACL will
see acl_m7 as 0 - which is wrong. The RCPT tests should be the other way
round (check acl_m3 then check acl_m7).
Many thanks for your help. An extra pair of eyes looking at the problem
usually helps :-)
John.
--
---------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: John.Horne@??? Fax: +44 (0)1752 233839
