-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Wayne Schlitt wrote:
Julian Mehnle writes:
The problem with this is that it suggests the use of localpart crypto
schemes (like SRS and SES), but many of these schemes will not honor
the 63 characters limit for domain labels imposed by performing a DNS
lookup using the %{l} (localpart) macro.
This section also doesn't mention the problems with using SES and
mailing lists that check the 2821.MAILFROM for subscriber
confirmation, rather than the 2822.From:. It also doesn't mention
problems with replay attacks and how to deal with them. It doesn't
mention the key distribution problem. It doesn't mention a lot of
things that many page SES spec.
I really don't see much reason to add the 63 character label problem
and omit the other ten pages of explanation.
Excuse me, but do you think I'm stupid??
All those things you mentioned are not (and must not be) dealt with in the
SPF spec because they are unrelated to SPF.
The 63 characters limit for _domain_labels_ I'm talking of is a limit that
stems from the application of SPF's %{l} macro. Thus it is NOT unrelated.
I am NOT talking about the 64 characters limit for e-mail address
localparts. If _that_ was the case, I'd agree right on that it doesn't
belong in the SPF spec.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
iD8DBQFClOeKwL7PKlBZWjsRAmfMAKCTFfbtqIwx/K4w9zZ02AjHEj21HwCfbf1f
5K0srbsi1ivXKU3nr+HqcpI=
=ig2n
-----END PGP SIGNATURE-----