Oops, second paragraph should read: The header does *not* ... > On Mar 30, 2019, at 3:05 PM, Viktor Dukhovni <ietf-d...@dukhovni.org> wrote: > > Essentially, all MTAs are intermediate MTAs. The header is added to the > message via the sender's *MUA*, and conveys the same sender preference to > every SMTP relay (MTA) en-route. For this header to be effective at > achieving its goal, there needs to be no prior difference between the MSA's > first relay hop and further downstream relays. TLS-related delivery issues > can occur at any hop along the path from sender to recipient. > > The header does obligate every en-route MTA to ignore its local policy. > Where such local policy mandates TLS, it will likely trump the header. > Where the MTA has no explicit policy and would otherwise follow the > hints (DANE or MTA-STS) from the next-hop destination, the header should > generally allow delivery when operational errors would otherwise break > such delivery by promising and not delivering on the requirements of DANE > or MTA-STS.
-- Viktor. _______________________________________________ Uta mailing list Uta@ietf.org https://www.ietf.org/mailman/listinfo/uta