On 6/20/2022 9:05 AM, Paulo Pinto via mailop wrote:
 >ARC is motivated by the cases where DKIM/SPF/DMARC information about the
 >author/originator get broken.

I'm truly trying to find a justification to break DKIM/SPF on a message after it is sent.

SPF is designed to be extremely fragile. It breaks when even simple MTA relaying is done through an MTA that is not pre-registered in the SPF record. Such relaying has been an essential part of Internet mail since before there was an Internet. SPF was designed after this entirely reasonable behavior was well-established.

The word 'justification' is probably awkward in this context, but the technical and operational details here are pretty simple.

DKIM was designed with an expectation that the basic message -- the part used to formulate the DKIM signature -- will not change. That's a reasonable assumption for a single posting/delivery sequence.

Mailing lists create multiple such sequences before 'final' delivery. Mailing lists can and do do all sorts of things to messages that wind up breaking the DKIM signature. They always have. Mailing lists, too, were well-established before there was an Internet and long before DKIM was developed.

These technologies were designed to work properly for only a subset of entirely reasonable email handling activity that has always existed.


SPF -> You should be aware of all the servers that can be involved in the message transaction

No, actually you shouldn't.  It's a requirement that doesn't scale.


DKIM -> The message should only be signed after it is complete and leaving your controlled environment. Any modification to the message afterwards is tampering and should not happen.

See above. DKIM is for a single posting/delivery sequence. Mailing lists entail multiple. Mailing lists operate at user-level, not the transport level. User-level software can and does do whatever it wants, prior to (re-) posting and always has.

d/
_______________________________________________
mailop mailing list
mailop@mailop.org
https://list.mailop.org/listinfo/mailop

Reply via email to