On Wed, Jul 06, 2005 at 03:23:42PM +0100, Tony Finch wrote:
The "forwarding problem" is _only_ an issue for entities that support
SPF at either end. It is only a problem for the forwarder if they care
about it.
My systems don't publish or check SPF, but SPF causes problems for me
because it breaks my users' email, and I have to deal with the complaints.
I take it you are a forwarder. If not, you shouldn't experience
any problems and/or your statement in this paragraph isn't correct.
The rest of my reply is written with above in mind.
It isn't SPF that's causing you problems. It is you using my name
that's causing you problems. In short: you are causing you problems.
2. Have your forwarder reinject.
This doesn't work. (a) It requires changes to the forwarding site, which
is what you are trying to avoid. (b) Where do bounces from forwarding
failures go?
To you, not to me. You send a message to a destination, you deal
with the problems this causes. Why should I have to deal with your
users sending mail to your users at a different location ?
3. Whitelist your forwarders.
How do you tell which incoming email is forwarded and which is not? What
happens when the forwarding IP addresses change?
Well, why would you need to know?
if "example.net" is me,
if "example.org" is the forwarder,
if "example.com" is the final destination:
Original mail:
"mail from: <me(_at_)example(_dot_)net>"
"rcpt to: <your_user(_at_)example(_dot_)org>"
New mail, generated by you:
"mail from: <your_user(_at_)example(_dot_)org>"
"rcpt to: <your_user(_at_)example(_dot_)com>"
You say you don't do SPF. This means "example.com" does not find
a record and therefore doesn't check SPF. Problem solved.
Yes, you have to deal with bounces from "example.com" to "example.org".
You shouldn't blindly forward those back to "example.com".
Now a counter argument: suppose "your_user(_at_)example(_dot_)org" doesn't
want me to know about his "your_user(_at_)example(_dot_)com" address,
sending a bounce to "me(_at_)example(_dot_)net" (as opposed to
your_user(_at_)example(_dot_)org)
would be undesired behaviour, with or without SPF.
Alex