James Thornton <[EMAIL PROTECTED]> writes:
> The bounce would come from the upstream server, not the TMDA
> server. If the upstream server is an open-relay and the bounce goes
> to a forged address, the open-relay's postmaster would be
> responsible for sending it to the forged address, not the administer
> of the TMDA server.
So, you'd just be pushing the delivery of the bounce to the forged
address onto someone else's shoulders. This doesn't solve the forged
sender problem. It doesn't solve anything. The qmail-localfilter
page you quoted also says:
``The bottom line is that YOU will never be responsible for sending
bounces to random forged return paths.''
Bullocks. You are still responsible for the bounce because you
initiated it by rejecting the transmission. Someone else is just
handling the delivery of that bounce. This doesn't mean you are off
the hook.
> From http://www.jfitz.com/qmail-localfilter/index.html:
>
> "The good thing about having your server reject the message (with an
> SMTP 550 code, not a bounce) is twofold:
>
> - Spammers may eventually take you off their lists.
This is pure speculation.
> - If you block a message from a legitimate sender, their own
> mailserver will send them a failure notification
So what? They still get an unwanted bounce message.
> Challenging messages at the SMTP levels also avoids additional
> delays in sending the challenge resulting from potentially large
> backlogged queues on TMDA servers.
In practice, if the machine is that swamped, it won't be able to
accept/reject new SMTP connections in a timely manner either.
_____________________________________________
tmda-users mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-users
