On Tue, Jul 18, 2017 at 10:02 AM, Alexey Melnikov <alexey.melni...@isode.com > wrote:
> I've started implementing ARC and have a few minor comments on the draft: > > 5.1.2.2. Computing the 'b' Tag Value for ARC-Message-Signature > > As with ARC-Seal and DKIM-Signature header fields, the order of > header fields signed MUST be done in bottom-up order. > > Upon rereading this I am not sure this is very clear. Maybe give an > example, especially if multiple instances of a particular header field are > included. > See section 5.1.1.3 of the document which specifies the way to build the header list for hashing. ARC-Authentication-Results needs ABNF (I know the document says in prose > that the syntax is effectively "i=<number>; ...", but also specifying this > with ABNF would be much better, especially in order to avoid people > inserting CFWS in places they shouldn't.) > See the -06 version which I posted today with some extra notes regarding open questions related to AAR. Once we settle those questions, I'll also add details (and ABNF) for the AAR and handling multiple A-R headers (another open question on the list). For now, if your environment creates multiple A-R headers with the same authserv-id, they need to have the (not "none") resinfo sections concatenated to create the AAR. I've taken a note to update this in the next draft after we resolve the open questions regarding content in Thursday's session. Examples need updating. > Thanks - I'm sure that there's a lot more problems in the examples than just the ones you listed :-/ I suspect that we should just entirely remove the examples right now. If you want a real life example, you can find it (for the next week or so) at https://pastebin.com/HyMriMKe or other citations on the #ietf99 channel in the slack group. --Kurt
_______________________________________________ dmarc mailing list dmarc@ietf.org https://www.ietf.org/mailman/listinfo/dmarc
