On Fri, Oct 11, 2002 at 10:25:27PM -0700, Marc MERLIN wrote:
> On Wed, Oct 09, 2002 at 09:40:21AM +0100, Matthew Byng-Maddick wrote:
> > On Tue, Oct 08, 2002 at 10:44:27AM -0700, Kevin P. Fleming wrote:
> > > This is fine. What this means to me is that once I have confirmed that
> > > your server accepts all local parts (or at least a sufficiently random
> > > one), I can cache that fact and stop sending sender verification
> > That sounds like an unimaginably bad idea to me, and defeats a lot of the
> > point of having callouts in the first place.
> Why would that be?
> They accept null envelopes, but they will also accept any rcpt to you
> throw at them.
Explain to me how you can know this? I want to know that I can bounce it
back to them if I have to (otherwise I wouldn't be doing callouts in the
first place), but I don't believe I can know the latter half of your
argument.
> Why waste your resources and theirs to do callbacks in this case?
I'm not against cacheing the callbacks, I just don't think you can infer
what you think is inferable.
MBM
--
Matthew Byng-Maddick <mbm@???> http://colondot.net/
