On Monday, February 13, 2012 08:58:32 AM Murray S. Kucherawy wrote: > > -----Original Message----- > > From: [email protected] [mailto:[email protected]] On Behalf Of > > John Levine Sent: Monday, February 13, 2012 7:42 AM > > To: [email protected] > > Subject: Re: [marf] Message bodies in ARF reports > > > > Interesting question. > > > > ARF is a version of multipart/report as defined in RFC 3462. In that > > document, the message/rfc822 or message/headers is optional. In RFC > > 5965, we made it mandatory in section 2.d. We did that because the > > usage scenario we were thinking of was spam button reports where you > > always have a message in hand, not for any deeper reason. > > > > You are of course correct that SPF failures sometimes don't accept the > > message. I'd be willing to file an erratum that 2.d. should have said > > that the report MUST include the message if the reporting entity has > > received it. > > I'd support that.
That's great. Is MUST include if you have ... or SHOULD include (unless you do not have ...) a better construction? I thought the trend in IETF documents was away from MUSTs if they weren't essential. > I'd also support the idea that an SPF report can synthesize the > text/rfc822-headers for a message whose content has not actually arrived. > All you legally need is a From: field as I recall, and that can be > generated based on the rejected RFC5321.MailFrom. If by 'synthesize' you mean 'assume the RFC5321.MailFrom and the FRC5322.From are the same', then yes, it could do that. That's not the same as saying I think it's a good idea. In cases where this heuristic is wrong (I've seen numbers between 5% and 20% of mail), it will probably be deeply confusing to receivers of such reports. I don't think we should. Scott K _______________________________________________ marf mailing list [email protected] https://www.ietf.org/mailman/listinfo/marf
