This sounds a lot like the cutthough concept - see spec.txt chapter 44.22.
I have only heard of this being used when forwarding the message to
nearby friendly mail servers, and I see that you cannot use DKIM and
cutthough.
> I noticed in my log files that I have situations where some mail is received
> from externally which is supposed to be forwarded to externally (with the
> 'redirect' driver, or 'remote_smtp' transport?). Now if my spam filter didn't
> detect the message but it was rejected by the forward recipient, Exim will
> generate a bounce message and try to send it to an arbitrary sender which is
> mostly fake. This is causing backscatter, which may cause issues with my mail
> server reputation in general. (I know that the RFC requires that behaviour,
> but that RFC is broken and not practical, for well-known reasons.) The same
> probably happens when the remote mailbox is full.
>
> I tried to understand what the 'errors_to' directive does but it's hard to
> follow those descriptions. So I'm not sure whether it's useful in my case.
>
> So I had another idea. If queuing and bouncing is the problem, can I get rid
> of that? If a message comes in from an unauthenticated sender, that must be
> another mail server. Those have got time, there's nobody waiting in front of
> a screen for the message to be submitted. So could Exim just try to deliver
> the message immediately in this case? If that fails, it could directly reject
> the message in the waiting original connection. No need for bounces.
>
> If the remote error is permanent, so should be Exim's. Same for temporary
> errors. Passing back the remote error message is probably a privacy concern
> as it might disclose the forward address or provider.
>
> This must only apply to forwards for unauthenticated senders. Authenticated
> senders are my users, they provided a password and I know I can send them
> bounces safely. And they're waiting for completion of the submission.
>
> How would a configuration of this look like? Or where in the documentation
> can I find more information about that?
>
> -Yves
>
> --
> ## List details at https://lists.exim.org/mailman/listinfo/exim-users
> ## Exim details at http://www.exim.org/
> ## Please use the Wiki with this list - http://wiki.exim.org/
>
--
Andrew C. Aitchison Kendal, UK
andrew@???
