One final response from me to this, because it is relevant to the IETF
last call:
On Fri, Jun 24, 2011 at 1:33 PM, Douglas Otis <dotis(_at_)mail-abuse(_dot_)org>
wrote:
Complaints from John, Dave, and Barry and others is likely and
understandably out of fatigue. They just want the process to be over.
No. We want to get the spec right, and to make sure it specifies the
right normative things.
Although DKIM will be securely hashing the headers fields which MUST include
the From header, developers are being told they must ignore multiple
singleton header fields discovered in the process.
No one is saying that anyone "must ignore" anything. If you think we
are, cite the words in the spec that say that.
What we are saying, and have said repeatedly in many places, is that
it is not the job of the DKIM specification to define the handling of
this normatively. Developers may certainly handle the presence of
multiple instances of header fields that are not allowed to have
multiple instances... in any manner they like, and we even suggest
(non-normatively) what they might do -- and ignoring the situation
isn't among our suggestions.
But it's not a normative part of the DKIM protocol, and it shouldn't be.
Now I'm done responding to this topic. The points have all been made,
on both sides.
Barry, DKIM chair
_______________________________________________
Ietf mailing list
Ietf(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf