Joe Kocsis wrote:
[snip]
> 1) Figuring that if the to and cc are all part of the same send that I
> wont need the fail, because since I require sender verify I dont need
> for responses that are all to the same message.
> However I dont understand why auth'd users can send to non-existsent
> useres.. I have the line
> require verify = sender
This verifies the SENDER, not the recipient.
> before the
> accept authenticated = *
> so shouldnt that deny the message once it see's a bad email??
>
> 2) The line appears to work if the sender is sending multiple messages
> via the same connection.. as soon as you enter 3
> rctp to: commands the connection is dropped..
This is exactly what the previously quoted ACL snippet does.
> So perhaps its working after all.... but Im unable to test its
> functionality from my email clients, since Im sending the to's as part
> of one message sent?
Should work exactly the same as telnet. The server has to be told about all
the addresses it is meant to send to on RCPT TO lines, the TO: and CC: headers
are included in the email but aren't used by the server for delivery (until
you start messing with programs like fetchmail to retrieve mail - evil thing).
It does sound like the ACL is working perfectly, but if it's still not
seemingly working, perhaps you could furnish us with a listing of your
acl_smtp_rcpt - it's impossible to tell what is going on without knowing what
the ACL is set up to do. The following section of the default ACL could be
causing issues if your client has been added to the relay_from_hosts hostlist,
but it's just a random guess.
<quote>
# If control reaches this point, the domain is neither in +local_domains
# nor in +relay_to_domains.
accept hosts = +relay_from_hosts
</quote>
You could try running "exim -bdf -d-all+acl" and see exactly what is happening
when you try to send it an email. Or if you really want to see every single
decision it's making "-d+all".
Ted.
