On Thu 20/Oct/2022 20:16:02 +0200 Brotman, Alex wrote:
I started adding some text around the "Report-ID" format. I ran into a bit of a hurdle, and thought it best to get group feedback. We decided a while ago to add language that the "Report-ID", "msg-id", and "unique-id" were the same. In the thread a few weeks ago, it was suggested the format should allow for optional "<>", and then a mix of alphanum, periods, and dashes. The draft states that the filename should include the same report-id used in the subject, however, the "<>" I believe are illegal in Windows filenames. I left a "NOTE" in the text in the area where we suggest they should be the same. It's quite possible I've misinterpreted the intent of the "ridtxt", and there's no reason to worry. If we really want them to be the same, I'm on the edge of suggesting that we create a format that is friendly across the board, and make it a MUST. I say the edge because it's possible I'm missing something, and wanted to get feedback.
I'd stay with the original:
unique-id = 1*(ALPHA / DIGIT)
It is a bit more lax than RFC4122 which specifies just hex. Most UUID
utilities seem to keep that format, albeit some may represent it with dashes or
dots for readability.
If at all possible, senders SHOULD set Message-Id like so::
Message-Id = "<" unique-id "@" id-right ">"
where id-right is defined in RFC5322. Not doing so can be a necessity if the
available MSA overrides the Message-Id anyway.
In the old spec, unique-id is optional in the filename. I'd rather make it
optional in the subject...
Best
Ale
--
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
