On Thursday, February 24, 2022 12:46:13 PM EST John Levine wrote:
> It appears that Scott Kitterman  <[email protected]> said:
> >> 1.  Take your domain, chop it to the last five labels if it's longer than
> >> that.
> >> 
> >> 2.  Walk up the tree starting at the original domain, and at each level
> >> look for a DMARC record.
> >> 
> >> 3.  If you find one with a psd flag, stop.
> >> 
> >> 4.  If you find one without a psd flag, remember it and keep going.
> >> 
> >> 5.  If you reach the root, stop.
> >> 
> >> If you found a record with psd=n, that is the org domain.
> >> 
> >> If you found a record with psd=y, the label below it is the org domain.
> >> 
> >> Otherwise the org domain is the last DMARC record you found.
> >> 
> >> The rest doesn't change:
> >> 
> >> The policy domain is the original domain if it had a DMARC record,
> >> otherwise the org domain. The org domain might not have a DMARC
> >> record. Relaxed alignment still means that two names have the same org
> >> domain.
> >> 
> >> If you found no records at all, there is no org domain and no policy but
> >> so
> >> what, there's nothing to do.
> >
> >Yes, with the minor proviso that is it's longer than 5, you would start
> >with the exact match and then jump to 5, but that's a detail.
> 
> Right, that's what I meant.
> 
> There may be a few other corner cases, e.g., if the original domain
> had psd=y, I think you ignore the psd flag since there's no lower
> domain that could be the org domain, or maybe perversely pretend it
> was psd=n so it is its own org domain. That allows the PSDs that have
> mail servers to continue to work no worse than now.

Yes.  That's another one.

Scott K


_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to