> -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of > Alessandro Vesely > Sent: Tuesday, October 11, 2011 12:21 AM > To: [email protected] > Subject: Re: [marf] New Version Notification - > draft-ietf-marf-authfailure-report-03.txt > > Good point. As an alternative, we could identify the > report-triggering results by reference, letting generators declare > their authserv-id. For example, adding the following line in the > second part of the current example: > > Authentication-Server-Id: mta1011.mail.tp2.someisp.com > > This alternative may allow a generator to piecewise write the A-R > fields directly to the relevant stream, rather than building them in > memory. At the recipient's, this alternative just prevents the > embarrassment of choice; e.g., what if the second part's A-R fields > seem to differ from their third part's correlatives?
There's nothing requiring a verifier to affix an A-R field in the first place, nor is there any guarantee that the verifier has properly stripped forgeries. Thus, what's in the second part is exactly what you're suggesting, and is possibly more reliable than what's in the third part. Moreover, the copy that's in the third part might be extracted before the local one is even added. It's right where it is, and doesn't need modification. _______________________________________________ marf mailing list [email protected] https://www.ietf.org/mailman/listinfo/marf
