My previous message also replied to your other message inline, just in case you missed that.
On Mon, Mar 17, 2025 at 1:31 PM Michael Thomas <m...@mtcc.com> wrote: > > On 3/17/25 8:10 AM, Allen Robinson wrote: > > re: numbering (to keep this all in one email) > > > > We absolutely need ordering. Maybe we don't need an index number to > > have guarantees about that ordering surviving transit, which is > > potentially reasonable since we have the same requirement for any > > signed repeated headers. Upgrading all of those ordering SHOULDs into > > MUSTs may be a good idea. > Absent any evidence to the contrary, why should I or anybody else > believe that this is an actual problem in the field? I don't remember > why it's a SHOULD vs a MUST re: trace headers, but an update to DKIM > could change it to a MUST if need be. > > > > Is the index number hsrmful is some way that I'm not seeing? > > The point is why do something that isn't needed? But there is harm in > that i= is already a valid DKIM tag and since it's not been decided > whether it's a new protocol or an update it shouldn't make breaking > changes gratuitously. > Sure. If we collectively decide that an index tag should be included in whatever signature header a DKIM2 signer produces, and we also decide to implement DKIM2 as a DKIM-Signature header with other stuff around it, then the index tag would not be allowed to use the already-assigned i= tag, like maybe n= or hop=. I don't see this as harmful on its own, and this collision also wouldn't be a valid argument for why to implement DKIM2 as an entirely new header. > > Mike >
_______________________________________________ Ietf-dkim mailing list -- ietf-dkim@ietf.org To unsubscribe send an email to ietf-dkim-le...@ietf.org
