The alternative is modifying the proxy code to insert its own Received
header. That would add a lot of overhead to what is otherwise a very
lightweight front end defence on our SMTP servers.
I don't get it. The proxy gathers the info that it is putting into
the logs anyway and stuffs a Received: line to the next stage ahead of
the body of the message. This doesn't seem like a lot of work. It
should be able to do that before reading the body of the message from
the incoming connection so it doesn't need any extra buffering.
What am I missing? Or are you saying the proxy passes a pointer to
the socket rather than the contents?
Or better still, use XCLIENT. This is it's main use-case.
Seems reasonable. Are there descriptions of it beyond what's in that
ietf-smtp mailing list