On Wed, Sep 13, 2017 at 1:03 PM, Murray S. Kucherawy <[email protected]>
wrote:

> At the risk of bringing up the whole "cv=invalid" debate again...
>
> When a chain is invalid (say, an AMS is missing), Section 9.3 says to add
> a seal that only covers itself but uses N+1 for its "i=" value.  Could
> someone propose some informational text for the draft that explains why
> that decision was made?
>

Yes, will add such information. In short, the reason for covering only the
last ARCset is because it is impossible to determine exactly what the
"implicit-h" list would be if the chain is corrupt. That makes it
indeterminate as to whether one should believe the "failing" report since
validating the signature would be ambiguous.

--Kurt
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to