Re: DMARC and ietf.org

[Brian E Carpenter <[e-mail]>]

On 22/07/2016 13:24, [e-mail] wrote:
> 
>> On Jul 21, 2016, at 12:43 PM, [e-mail] wrote:
> 
>>>> On 7/21/2016 7:35 AM, Brian E Carpenter wrote:
>>>>> Yes. So I repeat the question: Since the most pragmatic, non-purity-based
>>>>> solution is to rewrite the sender field for mail from p=reject (or p=quarantine)
>>>>> domains, when will we change the IETF and IRTF mailmen to do so?
>>>
>>>                                
>>>> I'm sure you really meant this, but just to be careful, what with this
>>>> being a technical point in a technical forum, it's worth clarifying that
>>>> the rewriting is for the rfc5322.from field and not the rfc5322.sender
>>>> field.
>>>
>>> I have an additional suggestion.
>>>
>>> If we're going to do this - and I'm not going to offer an opinion on whether or
>>> not it should be done - I'd like to see it done in a fashion that's both
>>> detectable and reversible. That way people using sieve or procmail or whatever
>>> will be able to undo the damage.
>>>
>>> The most straightforward way to accomplish this would be to make copies of the
>>> original fields with different names, but of course many other approaches  are
>>> possible.
> 
>> I do not see MailMan settings to make that happen.  Maybe I missed something...
> 
> That's most unfortunate, and I have to say moves my position from neutral
> to "don't do it".
> 
> Reversible damage is one thing, irreversible damage another.

That's the dilemma. An agent that obeys p=reject does irreversible
damage too. I can figure out how to live with p=reject being treated
as p=quarantine, but not with "reject means reject". And it isn't
just me, it's every IETFer using a certain mail operator that has
threatened to obey p=reject Real Soon Now.

   Brian

   Brian