On Wed, Jun 29, 2005 at 05:28:22PM -0400, Greg A. Woods said:
> [ On Saturday, June 25, 2005 at 11:46:53 (-0400), Stephen Gran wrote: ]
> > Subject: Re: [exim] a large number of domains fronted by Exim are refusing bounces...
> >
> > What if I put instead
> > IF $sender IS $the_null_return_path THEN
> > file_message(/dev/null)
> >
> > The facility to ignore bounces still exists, and your campaign to hobble
> > ACL's has changed nothing.
>
> Such a rule does not cause bounces to be rejected at the protocol level.
I guess you missed my point. If you expose the $sender_is_null or
senders = : at any point, you have exposed it. It could also be written
as
IF $sender IS $the_null_return_path THEN
:fail:
to cause a reject at the protocol level.
Ignoring this (suppose you also managed to hobble .forward files to
not be able to use this combination), what have you achieved? You have
forced the receiving MTA to accept a message that will never be seen,
for what pupose exactly? Shortening _your_ queue because you have
accepted a message that you know you can't return? Who wins here?
Not the recipient, seemingly.
> > Give me one reason why I should accept a message to root@ with a null
> > sender, originating outside my network. root never sends a message to
> > anyone, and never directly receives a message (aliases and all that).
>
> If root doesn't receive messages then block _all_ messages to root
> regardless of what sender address they arrive with (e.g. if they're not
> delivered by a command-line agent running on the localhost).
I do. However, some aliases, like postmaster, can legitimately get
messages with a non-null sender, but are never used for outbound email.
What do you suggest doing with a null sender message to postmaster,
hostmaster, webmaster, and other roll accounts that never send outbound
email? You have said:
> However if those addresses do exist then they _MUST_ accept valid
> messages, from valid sources, especially when those messages are sent
> with a null return path.
>
> Delivery error notifications are far from the only kind of message which
> might (and in some cases "SHOULD") be sent with a null return path, as a
> quick grep through the current RFC corpus will reveal, never mind all
> the other non-standardized uses.
I still don't see why the addresses under discussion (which do, I assure
you, exist), need to accept a null sender email. Can you explain why
they need to? You have yet to provide a real reason, as far as I can
tell.
--
--------------------------------------------------------------------------
| Stephen Gran | "Call immediately. Time is running |
| steve@??? | out. We both need to do something |
| http://www.lobefin.net/~steve | monstrous before we die." -- Message |
| | from Ralph Steadman to Hunter Thompson |
--------------------------------------------------------------------------
