Re: [ietf-smtp] Fwd: New Version Notification for draft-crocker-email-deliveredto-00.txt

[Dave Crocker <[e-mail]>]

On 2/6/2021 3:19 AM, Alessandro Vesely wrote:
> On Fri 05/Feb/2021 22:53:34 +0100 Dave Crocker wrote:
>> On 2/5/2021 1:44 PM, Valdis Klētnieks wrote:
>>>
>>> It's a plausible design that the agent doing final delivery accepts
>>> responsibility for delivery to multiple local mailboxes (I've often 
>>> had things
>>> that ended up delivering to root@localhost and valdis@localhost with one
>>> invocation of the local delivery agent, for example).
> 
> 
> Even then, actual delivery usually happens one mailbox at a time, in a 
> loop.

A variation on John's response:  A protocol specification is not an 
implementation specification.  Your comment is about an implementation 
choice.  The only requirement for a protocol spec is that it not prevent 
reasonable implementation.


>> The question for the specification is whether it makes clear what the 
>> specification calls for.  I think it does, but if readers aren't clear 
>> about it, it would be good to know what about it isn't clear enough.
> 
> 
> Perhaps, the case of delivery to multiple mailboxes should be 
> anticipated in Section 3.

I've added a pointer to Security Considerations.


>> A different question is how to implement the specification, so as to 
>> conform to the specification.
>>
>> For storing a separate copy for each recipient, that seems 
>> straightforward.
>>
>> For sharing a single copy of the message among multiple recipients, 
>> the answer probably is not straightforward...
> 
> 
> The case of shared IMAP folders is also straightforward.  The 
> Delivered-To: contains the mailbox of that folder if it has one, or the 
> mailbox of the recipient whose filter directed the message to that folder.

Shared folders are unrelated.  A Shared folder has a single address.

A different issue is implementations that store a single copy of a 
message sent to multiple addressees.  The text is meant merely to 
highlight a concern for divulging other addresses, not to suggest solutions.



d/

-- 
Dave Crocker
Brandenburg InternetWorking
bbiw.net