> I'm not 100% certain I'm "get"ting your idea here ... we do currently
> run inbound/outbound mail on different IPs, but the problem isn't with
> the connections themselves.
>
> From the example session transcript with spamd that I posted earlier:
>
> 250 Hello, spam sender. Pleased to be wasting your time.
> MAIL FROM: [EMAIL PROTECTED]
> 250 Ok to start over.
> RCPT TO: [EMAIL PROTECTED]
> 250 You are about to try to deliver spam. Your time will be spent, for
> nothing.
>
> For an actual MTA, the 250 code here indicates an open relay, because
> we are not the MX for checkor.com.
Nope, because you can reject it later on. Their test is flawed.
All our open relay checking we do here actually checks the full
dialogue to see if the message is accepted *and* checks for an actual
relay, if they aren't doing that, their test is amatuer hour crap.
> spamd of course doesn't know this
> (and I'm aware that fixing it might not be easy), but it is still
> triggering a false positive as an open relay.
Spamd won't change in this regard. It would completely break
spamd to accomodate some false idea of a "test".
