On Mon 06/Jun/2022 22:05:08 +0200 Les Barstow wrote:
On a technical note, “0” and “1” generate DMARC failure reports, while “d”
produces a DKIM failure report and “s” produces an SPF failure report.
No, they are the same format specified in RFC 6591.
DKIM Failure Reporting (RFC 6651) and SPF Authentication Failure Reporting (RFC
6652) allow to request reports in the DKIM or SPF records, respectively.
Those specification allow to specify the percentage (rp=) and the reason of
failure (rr=). DKIM even provides a string (rs=) to be output in case of
reject, so that the sender will find it in the mail log.
They /are/ slightly different in content (and specification).
In DMARC, the percentage is totally at the mercy of the report generator. The
reason, for example tempfail, locl policy, syntax error, cannot be specified.
Anyway, all three of those specs refer to RFC 6591 for the format. Up to rs=,
if you request all of them to the same target address, you won't be able to
tell which one triggered sending a specific report.
Should dmarc-failure-reports attempt a comparison with those report requests?
Technically, I suppose “0:d:s” could produce one of each. That is, to put it
mildly, ugly.
I agree with Todd here. "0:d:s" is redundant. I have no opinion on whether it
should be syntactically valid or not.
Maybe this needs more than a simple ABNF discussion?
Sure. In fact, the whole DMARC failure reporting is redundant. Such
redundancy was supposedly intentionally introduced in an attempt to stimulate
an increase in the number of report generators...
Best
Ale
--
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
