On Mon 18/Jan/2021 18:54:15 +0100 John R Levine wrote:
On Mon, 18 Jan 2021, Alessandro Vesely wrote:
At least, we could specify in the General Record Format that unrecognized
policies should be treated as p=none.
Absolutely not. Any DMARC record with an invalid p= is just broken. We write
standards that say how to interoperate, not how to guess what other people had
in mind when they implemented it wrong.
Right, but IF one day we introduce a new policy, we'll have to be compatible
with installed base. Please don't reply that we'll NEVER introduce new policies...
BTW, the current spec does not mean that an invalid p= implies the DMARC record
is broken. If it did, it wouldn't say to check rua= in that case.
Best
Ale
--
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
