ned+ietf-smtp(_at_)mrochek(_dot_)com wrote:
John Leslie writes:
> It may be worth noting that sender address rewrite should not be
> done when forwarding a message via TBR.
>
> (I can't offhand think of a reason you'd actually want to...)
Sieve processors executing redirect statement are permitted to change
the MAIL FROM, or leave it.
With one important exception: If the return address is empty you're not
supposed to fill it in. If you do you can end up creating a bounce loop.
I've been wondering if this restriction/concern needs to be mentioned in
2821bis.
I think it should.