Greg Connor wrote:
For most domains, I believe the policy for From: or HELO or
Sender: will be very similar to their policy for MAIL FROM
Dubious, many domain owners probably think that an "equivalent
header" is no problem, but some will find out that e-mail has
more obscure features than they were aware of. Like the case
of "moderated newsgroups" mentioned the day before yesterday.
I am sensing an active and vocal segment of the list who
don't want to mess with v=spf1 at all.
Adding new optional modifiers ("option" as in "opt-in") is a
part of the design.
I think trying to relate multiple headers together and make
one "control" the others leads to more complexity without
adding much value.
The Return-Path is a dominant header. And the relation between
Resent/From/Sender is defined in STD 11. You can't simplify
SMTP, it's already supposed to be simple. ;-) The concept of
a sender policy doesn't work well with 2822-headers, they show
up in unexpected places, e.g. behind news2mail gateways at the
other end of the world.
if two headers have different domains, they would be checked
one at a time against different SPF records. Does that make
sense?
Yes. But in a mail with Resent-From it makes no sense to check
the From. Dito Sender vs. From, dito Resent-stuff vs. Sender.
get yelled at by the hardcore classic people ;)
[...]
The main idea here is that I really do think that MOST
domains will have the same policy for multiple headers
YELL, you asked for it, <eg>
The mailfrom line also includes one way to deal with the
HARDPASS issue...
So you'd redefine + = HARDPASS, ? = SOFTPASS, ~ = SOFTFAIL,
and - = HARDFAIL, is that correct ? Nice, I like it.
I'd like it even more without SOFTFAIL, because I don't see
receivers actively helping to debug sender policies beyond
Wayne's idea of a "validating evaluation".
Bye, Frank