Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread Alessandro Vesely
On Fri 21/Jan/2022 18:59:16 +0100 John R Levine wrote: Yes, that is the plan.  Please go back and look at the discussion when we talked about the tree walk in the first place. I don't remember that we ever convene that subdomains cannot override the org domain record. The "org domain record"

Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread Dotzero
On Sat, Jan 22, 2022 at 6:52 AM Alessandro Vesely wrote: > > No, the concept of Organizational Domain is foundational to DMARC. We > cannot > overthrow it to spare an extra lookup. When we talked about tree walk we > knew > that additional lookups might well have come out. > > To specify that

Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread John R Levine
Ale said: No, the concept of Organizational Domain is foundational to DMARC. ... We went though this whole discussion quite a while ago when we agreed to switch to the tree walk. Unless you have a compelling argument about why our previous analysis and decision was wrong, I think we're done.

Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread Douglas Foster
If a.b.example.com is considered aligned with c.example.com under RFC7489, but will be considered unaligned under DMARCbis, then we have a pretty significant incompatibility and need to move to DMARCv2. In the current example of comparing a.b.example.com and c.example.com to each other: if there

Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread Dotzero
On Sat, Jan 22, 2022 at 11:02 AM Douglas Foster < dougfoster.emailstanda...@gmail.com> wrote: > If a.b.example.com is considered aligned with c.example.com under > RFC7489, but will be considered unaligned under DMARCbis, then we have a > pretty significant incompatibility and need to move to DMAR

Re: [dmarc-ietf] Tree walk is screwed up

2022-01-22 Thread Douglas Foster
To correct my last post, we are not reduced to using a list, as we have already anticipated this situation with the PSD flag. The solution to the sub-lease situation is for the parent domain to publish psd=y. We may need two PSD flag, one for "I am a real PSD and never send mail and should not be