[feedback-report] Repeated optional fields distribution in feedback-report, and typo in IETF draft

[Mark Martinec]

(sorry if this is a duplicate)

> Is this a valid report(with fields being repeated, but other fields
> being in between)?
> 
> Authentication-Results: [ authres-header-1 ]
> Original-Rcpt-To: [orig-rcpt-to-1]
> Authentication-Results: [ authres-header-2 ]
> Original-Rcpt-To: [orig-rcpt-to-2]
> 
> Or should the repeated fields be placed consecutively like below?
> 
> Authentication-Results: [ authres-header-1 ]
> Authentication-Results: [ authres-header-2 ]
> Original-Rcpt-To: [orig-rcpt-to-1]
> Original-Rcpt-To: [orig-rcpt-to-2]

Murray S. Kucherawy wrote:
> To some degree we're documenting current usage.  So what are current
> implementers doing around this issue?

It doesn't make semantic sense to interleave them, as the
Original-Rcpt-To go one-per-recipient, and the Authentication-Results
are one-per-signature. Amavisd-new does all the Original-Rcpt-To first,
then all Authentication-Results.

  Mark