On Thu, Apr 28, 2022 at 9:39 PM Dave Crocker via mailop <[email protected]> wrote:
> > On 4/28/2022 1:52 PM, Dave Crocker via mailop wrote: > > If writing a formal specification, yes, one needs careful language. > > This isn't that exercise. > > This prompted me to consider language that might be suitable for an RFC. > Perhaps: > > An MTA that is relaying a message SHOULD NOT attempt to repair > problems it detects with the message. > > If the MTA continues relaying the message, it MAY send an informal > multipart/report note, back to the RCTP-TO address of the message, > indicating the nature of the problem with the message. > RCPT-TO being the relay destination or the original destination? I would think you'd want to go to the MAIL-FROM as the creator of the message? > The details of the report would have to be provided, since this doesn't > fit within DSN details. > I think automated reports of these types of failures would be a bad idea, slightly less bad if it was only done for authenticated senders, and if rate limited... and sent to the postmaster and not the original sender, who is unlikely to have any ability to change how the message was formatted. Brandon
_______________________________________________ mailop mailing list [email protected] https://list.mailop.org/listinfo/mailop
