On 12/6/24 18:25, John Levine wrote:
> DMARC has been in active use for a decade, so we don't have to speculate about
> how it works. We know that hardly anyone publishes size limits for reports, 
> and
> I at least have not heard any complaints about oversize reports. This suggests
> both that reporters have figured out how to send reports of reasonable sizes,
> and that domains can handle the reports they get.

That's my impression as well. While trying to figure out why the size
discussion was left in, my mind went into the long grass trying to come
up with a reason, however far fetched.


> With respect to that particular example, if you send enough mail to get big
> reports, and your MTA that receives the reports publishes a SIZE that only
> accepts small messages, well, that wasn't very clever, was it? I don't see why
> we need to spend any time on an entirely hypothetical problem.

Is the whole size discussion so hypothetical that we should even remove
the "MAY send a message saying you have a report that couldn't be sent"
part?

I guess your 'it's obsolete and reporters aren't expected to pay
attention to it' below is an answer.


> With respect to taking the limit syntax out of the ABNF, I'd be OK leaving the
> limit stuff in with a note that it's obsolete and reporters aren't expected to
> pay attention to it.

It sort of is half in, stating that exclamation mark needs to be percent
encoded.

We could go the obs-dmarc-uri route, imitating how RFC 5322 does it.

obs-dmarc-uri = dmarc-uri [ "!" 1*DIGIT [ "k" / "m" / "g" / "t" ] ]
              ; Obsolete syntax, reporters should ignore the
              ; size if it is found in a DMARC Policy Record.

and then update dmarc-urilist with the alternative

dmarc-urilist = ( dmarc-uri *(*WSP "," *WSP dmarc-uri) )
              / ( obs-dmarc-uri *(*WSP "," *WSP obs-dmarc-uri) )


Daniel K.

_______________________________________________
dmarc mailing list -- dmarc@ietf.org
To unsubscribe send an email to dmarc-le...@ietf.org

Reply via email to