Re: [sieve] New Version Notification for draft-george-sieve-vacation-time-00

[Aaron Stone <[e-mail]>]

2010/2/6 Ned Freed <[e-mail]>:
>> 2010/2/6 Arnt Gulbrandsen <[e-mail]>:
>> > Дилян Палаузов writes:
>> >>
>> >> Considering the above way of determining the From address, why cannot the
>> >> requirement
>> >>
>> >>   Unless explicitly overridden with a :from parameter, the From field
>> >>   SHOULD be set to the address of the owner of the Sieve script.
>> >>
>> >> be met?
>> >
>> > Not sure when I'll get around to it, but I do mean to someday implement
>> > group-wide and system-wide sieves, ie. sieves written by someone, processing
>> > someone else's mail. I don't understand what that SHOULD means in that
>> > context.
>
> This really isn't that hard. We've had user sieve sieves, channel sieves,
> system-wide sieves, domain sieves, head of household sieves, mailing list
> sieves, and various other sorts. But in every case there's a fairly obvious
> choice for an owner address, e.g,. for a domain sieve it's the domain
> postmaster and if there isn't one of those, the system postmaster.
>
> The tricky thing in a multiple sieve environment isn't figuring out who owns
> the sieve but rather how to combine the results of multiple sieves. Jutta
> Degener's long-expirted multisieve draft laid out the approach we follow, more
> or less.
>
>> Event in that context, I'd still infer address based on the recipient
>> -- the system level script should still know whose mailbox the message
>> will be delivered to.
>
> IMO that's a *really* bad idea. You're forging mail when you do this.

Hmm. Fair point. If a system sieve were to use the vacation mechanism
to report that a user has been suspended, or a mailbox is over quota,
etc., the reply ought to come from a system address. Arnt, did you
have another use case in mind? (And is this a reasonable issue to
discuss or just a rathole to discuss the depth of?)

Aaron