-----Original Message-----
From: owner-spf-discuss(_at_)v2(_dot_)listbox(_dot_)com
[mailto:owner-spf-discuss(_at_)v2(_dot_)listbox(_dot_)com]On Behalf Of Frank
Ellermann
Sent: Friday, August 05, 2005 7:56 AM
To: spf-discuss(_at_)v2(_dot_)listbox(_dot_)com
Subject: [spf-discuss] Re: SPF test suite
Marc Chametzky wrote:
test4.spam.co.nz "v=spf1 redirect=test4.spam.co.nz -all"
[...]
With the "-all" in there, the redirect modifier won't even
be used, so this should simply return "Fail", right?
Argh, yes, sorry, I missed the "-all". Scott has it right,
the "redirect=" has no effect, it's "allowed" (probably not
inteded - but validators trying to guess what a user means
are in trouble ;-)
Sorry also to Lennon, that policy should work. The spec.
says only "RECOMMENDED" to appear as the very last term,
and "all" vs. "redirect=" has its own rule in chapter 5.1:
Mechanisms after "all" will never be tested. Any "redirect"
modifier (Section 6.1) has no effect when there is an "all"
mechanism.
Tricky for validators, it's almost an error, but not in
practice. They could display two warnings:
1 - redirect=any with "all" is useless
2 - redirect=self would be an error without "all"
Bye, Frank
OK. Some more items for the To Do list.
My validator is already set up to raise warnings on some things that are not
formally errors, but almost certainly not what the publisher wanted.
Catching the existence of both a redirect= and an all in the same record is
a good candidate for that.
Also still on the To Do list is continuing to process to the end and collect
up ALL the errors. Currently it bails after the first error. I have yet to
negotiate with Stuart to see if pySPF can carry the impact of additional
code for this, because it isn't something that's operationally useful. It's
only for the validator.
Scott K