You don't seem to be paying attention. You're top-posting again, for a
start. It is early where you are, I suppose -- you probably didn't even
notice that the list broke the GPG signature, did you? I thought that
was a nice touch, but subtlety can be lost if you try it before the
recipient has had sufficient coffee in the morning :)
On Wed, 2004-09-15 at 08:49 -0500, Daniel Taylor wrote:
Well, it does fix Mr. Woodhouses objections, as this
mechanism would allow at least two of the forgeries he
presented to be caught, if not the lot.
I stated no objections, and it's not clear that any of the four examples
I presented actually _were_ forgeries. They _could_ have been, but they
just as well have really been authored by Joe.
Can you explain how the use of the %{l} macro would fix the problem in
any of the four cases I presented? Perhaps I'm being dim, but I just
can't see it. We assume that Joe is a real person and that
joeuser(_at_)example(_dot_)com or joeuser(_at_)hosteddomain(_dot_)com is a real
address which
accepts bounces in each case, of course.
If we have a macro which for which the implementation substitutes the
RFC1413 ident, perhaps that might help with #4 because we can make it
permit only mail sent from the mailer-daemon user, not from any user of
the box. But I don't see the relevance of the %{l} macro at all.
--
dwmw2