Re: mailing lists are complicated, was Email explained from first principles
Alessandro Vesely <[email protected]>
| Newsgroups | gmane.ietf.smtp |
|---|---|
| Message-ID | <[email protected]> |
On Fri 28/May/2021 20:18:19 +0200 John Levine wrote: > It appears that Alessandro Vesely <[email protected]> said: >> Mailing lists often do keep DKIM signatures intact. They only alter the rest >> of the message, but the transformations they do are stylized enough to allow to >> revert them and verify the original signatures. ... > > We've been around this barn a few times before. What you say may seem > true if you have only seen a few lists handled by a single list > package. In reality, I have seen lists add headers and footers, add, > delete, and reorganize MIME parts while changing the MIME boundary > strings, rewrite URLs in message bodies, and flatten HTML bodies into > text. There is no way to list all of the common changes, much less > describe how to reverse them. There are also lists that gave up making any change in order to not break DMARC. Some other lists cannot afford that, and turn to From: rewriting. If they can manage to restrain from irreversible changes, the original From: can be safely restored. Reverting From: rewriting is a practical approach which works for cooperating lists and cooperating author domains. > In any event, whatever you might want to do by reviving DKIM signatures is already > handled by ARC, so it'd be a lot more productive to encourage ARC adoption than > to invent yet another half-hearted DMARC workaround. ARC is a completely different thing. It introduces a third party, the sender, which modifies the message header in such a way to induce receivers to accept a message even if it fails DMARC. Semantically, it is very similar to Dave's Use of the Sender Header Field. ARC is good for global providers who are able to rate intermediate senders. I see no conflict between these two methods. Best Ale --