Dick St.Peters wrote:
Exactly what in RFC 1123 "broke" 251-forwarding?
They removed the source routes. In theory the Return-Path
was a literal "path" with 821, each forwarder added its
own name to the return path.
So if something went wrong the mail took exactly the same
path backwards to the originator. Forwarders had a vital
interest to forward only mail _not_ resulting in later
bounces, they'd get it back and would then be obliged to
"backward" it.
251-forwarding is extremely rare anyway.
The more important issue is that SPF FAIL breaks the
much-more-common alias forwarding described in RFC 1123
section 5.3.6(a).
ACK, but that's what I mean when I say 251-forwarding,
for some time I wrote 5.3.6(a)-forwarding, 251 is shorter.
The "much more common" from my POV is still only one case
where my FAIL-policy resulted in a pseudo-551 (next hop
rejected it, 251-forwarder bounced it to me). Only one
case in now 16 months (okay, I don't write many mails).
Bye, Frank