Frank Ellermann <nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> wrote:
David F. Skoll wrote:
<snip language about "proper failure report">
SHOULD, should, recommended, or RECOMMENDED, what's going to happen
with those huge log files, apart from deleting them by a cron-job ?
1) disk space is cheap;
2) they presumably will be rotated, and only deleted after a week or
more;
3) parsing them by a script looking for an incident will be simple.
If we avoid defining what a "proper failure report" is and leave it
up to SMTP implementors, I fear that e-mail will become completely
unreliable (instead of somewhat unreliable as it is today.)
Yes, but it's your proposal to say "SHOULD bounce" instead of the
"MUST bounce" in 2821. I fear that this is another a loophole to
avoid real (and painful) solutions, based on "reject at the border".
"Reject at the border" is already becoming a normal practice: be
patient...
I wouldn't be happy if an SPF PASS ends up in a log file deleted by
a cron job. Maybe your SHOULD needs a qualifier: "If the receiver
cannot validate the return-path then" etc.
We cannot presume universal use of SPF.
OTOH, I could live with your wording...
For the "normal" case (in my parallel universe that's an SPF PASS
or a similar case :-) I'd like to keep the "MUST bounce" as it was.
Don't hang on too tight to that horse -- it's pining for the
fjords... :^(
--
John Leslie <john(_at_)jlc(_dot_)net>