Roger Moser wrote:
No problem: v=spf1 sc=s1 m1 sc=s1,s2 m2 m3 sc=s2 m4
Please don't say v=spf1, these ideas don't work with existing
implementations. The spf2.0 notation is shorter, and your
notation is better readable for humans.
OTOH the spf2.0 notation works for all kinds of positional
modifiers, not only scopes, e.g. v=spf2 m1 q=hardpass m2 m3
yery easy to program: when encountering the "sc" modifier
set the flags accordingly. No need of look-ahead.
That's no problem, future spf2.0 implementations could handle
all these variations, and in Wayne's "validating evaluation"
the look-ahead is necessary and a feature.
A bad idea would be _three_ kinds of modifiers (global, local,
and your kind of positional).
Why is m1 not in s2? "m1" is followed by "only=s2" and
therefore according to the Sender-ID definition it is in s2.
In draft-ietf-marid-protocol-03 it's THE mechanism (3.6.3):
| Positional modifiers apply only to the mechanism they follow.
Same text in Sender-ID. In "m1 only=s1 m2 m3 m4 only=s2" the
"only=s2" affects m4, but not m1..m3.
Bye, Frank