On 6/15/2019 6:13 PM, Steve Atkins wrote:


On Jun 15, 2019, at 9:25 PM, <dilyan.palau...@aegee.org> wrote:

Hello,

p=reject; pct=0 is equivalent to p=quarantine; pct=0.

I've not been following this thread too closely so I might
be missing something, but under current DMARC spec I don't
think that's so - see section 6.6.4.

If I've missed the point ... never mind, carry on.


If I follow myself, I think it could be expressed as:

p=reject; pct=0; is effectively equivalent to p=quarantine; pct=100;

Given the order of mail "restriction" or "filtering" from high to low of reaching the user's eyeballs:

  p=reject       never accepted or accepted/discarded
  p=quarantine   accepted, imported into spam box, outside inbox
  p=none         accepted, imported into inbox

The "pct" effectively forces a fallback to the next lower applicable policy once the pct of failed mail has been processed:

  p=reject; pct=X;  fallback to p=quarantine
  p=quarantine; pct=X;  fallback to p=none
  p=none;  pct=X  fallback to UNDEFINED, N/A

where X can be 0 to 100.

When pct=100, which is the default, then the fallback would not apply since the explicit domain policy is applied to all DMARC failed messages. The receiver rejects mail with p=reject and quarantines mail with p=quarantine.

If there is an explicit pct=0, then effectively, the fallback is to be applied immediately, thus:

p=reject; pct=0; is effectively equivalent to p=quarantine; pct=100;

and

p=quarantine; pct=0; is effectively equivalent to p=none; pct=100;

Because of the fallback and quarantine implementation complexity and how failed messages can reach users, the OP is proposing to abolish the quarantine policy.


--
HLS


_______________________________________________
dmarc mailing list
dmarc@ietf.org
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to