On July 19, 2022 8:43:16 AM UTC, Alessandro Vesely <[email protected]> wrote:
>On Mon 18/Jul/2022 12:02:15 +0200 Scott Kitterman wrote:
>> On July 18, 2022 9:37:17 AM UTC, Alessandro Vesely <[email protected]> wrote:
>>> On Sun 17/Jul/2022 19:05:58 +0200 John R. Levine wrote:
>>>> I made a pull request from Scott's examples, lightly edited.
>>>
>>> Could we please add a convoluted PSD example?
>>>
>>> 5322..From domain
>>> : psd.example.com
>>>
>>> 5321.MailFrom domain
>>> : mail.psd.example.com
>>>
>>> DKIM-Signature d=
>>> : signing.psd.example.com
>>>
>>>
>>> This'd be similar to the first example, except that _dmarc.psd has psd=y.
>>> The aim is to show the possibly unexpected effect of psd=y on relaxed
>>> alignment.
>>
>> I think we should limit examples to cases that actually exist. As far as I
>> know, the PSD below a non-PSD in the tree is entirely theoretical.
>
>
>Examples don't have to be real. There is one where _dmarc.com exists and has
>psd=y...
>
>A PSD below a non-PSD is instructive for a couple of points not covered in the
>text:
>
>* The psd=y tag is disregarded or not depending on the context, not the domain,
>
>* a PSD must publish a DMARC record if it is below a non-PSD, even if they
>send no mail.
>
If there are no PSDs below a non-PSD (which AFAIK is true), I don't think we
care. My suggestion is that we do not add anything more for the moment.
The best way to understand what examples we might want to add would be for
someone who hasn't been deeply involved in the conversation to implement this
and see what questions they have and what issues there are in the resulting
implementation. Unfortunately none of us can do that.
Maybe there's a computer science professor out there that can assign it to a
class and report back?
Scott K
_______________________________________________
dmarc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmarc
