Mailing List Archive

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Stuart D. Gathman wrote:
> To handle checking that exp= is not used, I propose specifying the
> magic string "DEFAULT" as the expected explanation. This will match
> the default explanation, whatever it is. If an implementation supports
> setting the default explanation, then setting the default to "DEFAULT"
> in the test driver will do the trick. Otherwise, the test driver can
> check for the magic string and do something special.

I assume you are referring to 6.2/13 (the last paragraph of 6.2)?

I still think this corner of the SPF spec doesn't warrant the additional
complexity in the test suite required to test it. A mistake in a test
driver implementation with regard to this is about as likely as a mistake
of equivalent effect in the implementation itself, so there's really no
point.

But if you insist on adding this concept to the test suite spec, then
please give an example of what you proposed above. I don't understand
your proposal well enough to give any comments besides what I wrote above.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE83OGwL7PKlBZWjsRAmkBAJwMx9HJy8U3cgaPQuBs4EhAuFFsigCeKEO3
uV8+povvI63crsEhUyZr4os=
=2mcG
-----END PGP SIGNATURE-----

-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname=spf-devel@v2.listbox.com

Stuart D. Gathman wrote:

> setting the default to "DEFAULT" in the test driver will do
> the trick.

> Any comments?

+1. Julian apparently says that the exp= stuff isn't too
interesting, but SPF implementations wishing to support exp=
should get it right. A misleading exp= (from the wrong
record) could be bad.

Besides exp= is the only "real" modifier we have. And there
are macros only allowed in conjunction with exp= (otherwise
they'd be a serious privacy risk, allowing to track how far
spam gets).

And you want to test this stuff, IIRC the "DEFAULT" is your
third proposal. Make sure that implementations always using
the DEFAULT don't get a test-FAIL, that's a legal behaviour.

Conformance test cases can offer a third outcome: PASS, FAIL,
and INCONCLUSIVE (e.g. for an unsupported option).

Frank


-------
To unsubscribe, change your address, or temporarily deactivate your subscription,
please go to http://v2.listbox.com/member/?listname=spf-devel@v2.listbox.com