John C Klensin wrote:
Now, if you and a few thousand other people get this
message, I'd claim a reasonable case has been made
that it works
s/it/something/
Likely you didn't test EXPN, even if you did you can't
tell if it didn't work because EXPN is default disabled
or not implemented.
The IETF list happily ignored 2821 3.10 and added its
List header fields to the header. It would also ignore
the same MUST in 2821bis 3.9.
For a good test you'd have to send your message from an
"IPv6 only" address and account, and ask folks to reply,
and report problems in public.
You used ns.jck.com as EHLO, that's no test of a domain
literal containing NO-WS-CTL or similar oddities.
You likely didn't test HELO and a non-trivial reverse
path in your message. "Source routing" was deprecated
18 years ago, it's not obvious what happens if somebody
still tries (in vain) to use it, is it ignored or does
it cause havoc.
Your article arrived as MAIL FROM <ietf-bounces(_at_)ietf> at
GMane, I can guess why this was accepted as "originator
(as indicated by the reverse-path)", but 2821bis doesn't
explain (1) why this is not more specified as in RFC 821,
(2) why it's a crucial decision where geting it wrong is
no option, and (3) why it's not more exactly the same
situation as 1989 when RFC 1123 was published.
"Better than 2821" and "good enough for DS" are unrelated
concepts. Just because nobody uses 821 anymore (because
1123 rendered it useless) doesn't prove that the design
of 821 was wrong, or that the 1123 design inherited by
2821bs was "better".
And the utter contempt for mail standards demonstrated by
the approach to promote them without a proper WG is sad.
2821bis only demonstrates one thing, *never ever again*,
all wannabe-improvements of the standards process based
on this experience are dead on arrival.
Frank
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/ietf