On Thu, 8 Feb 2001, Marc Haber wrote:
> >What I don't understand is why this happens. Why doesn't the bounce go
> >back to the message's sender?
>
> It goes back to the message's sender, and is copied to postmaster.
Oh, you're using errors_copy. Well, if you ask for copies, you are going
to get them!
> >The new callback verify stuff can do this.
>
> This sounds nice. Are the results of the callback cached by exim in
> some way, or is a new callback being done for each message coming in
> for an address? If the address verifies, is the message delivered
> through the connection opened for address verification, or is a new
> connection used for that?
Sorry, I was talking nonsense. The callback stuff is for *sender*
verification. There's no equivalent for receiver verification. If there
were, I guess it should be called "call forward", and if it were
implemented, no, the message would not be delivered down that
connection. (Consider: a message has 300 recipients - you don't want to
hold open 300 connections while you are receiving the 20Mb of data
before you can deliver. Anyway, Exim just doesn't work like that.
Receiving and delivering are entirely separate actions.)
> hosts listed in sender_verify_hosts_callback are hosts where incoming
> mail is to be delivered to, or where mail is coming in from?
From.
--
Philip Hazel University of Cambridge Computing Service,
ph10@??? Cambridge, England. Phone: +44 1223 334714.
