Re: Who should set the Sender: header field?

On Fri 17/Dec/2021 04:10:43 +0100 John C Klensin wrote:
> --On Thursday, December 16, 2021 21:30 -0500 John R Levine wrote:
>> On Thu, 16 Dec 2021, John C Klensin wrote:
>>> When that new way of doing things will break the multiple From
>>> option, mess up mailing lists (although people are welcome to
>>> use their various group and social network services instead),
>>> ...
>> 
>> I'm with you on mailing lists, but the support for multiple
>> From is spotty even outside of gorilla land.
>> 
>> I'm not sure what to say about it but we are doing people no
>> favors by suggesting that mail with multiple from address is
>> likely to work.
> 
> I would favor a statement in the proposed A/S about this (or any
> similar feature) warning people that the feature is not widely
> supported, that it cannot be expected to work universally in an
> interoperable way, may result in mail being bounced and/or lost,
> and hence should probably be avoided unless the sender knows it
> will be handled properly at the recipient end.  I'd be hesitant
> about removing the construction from 5322bis entirely, but could
> be convinced.
> 
> The _only_ thing I would object to is arguing that the standard
> must be changed because one or several perceived large Gorillas
> have decided to not support it.


Since G say they banned multi-mailbox From: after DMARC (with no statement 
about how/if it worked before), perhaps if we rewrite DMARC so as to not 
preclude the authentication of that kind of messages, they might eventually 
restore the functionality.  Authenticating the sender of a multi-mailbox From: 
message does not suggest that its end-to-end flow is likely to work, but 
neither denies it.


Best
Ale
--