Il 7 Mar 2004 alle 12:15 Alan J. Flavell immise in rete
> On Sun, 7 Mar 2004, Steffen Heil wrote:
> > Is there a way to drop messages instead to freeze them, if (and only
> > if) they have null envelope senders?
> With respect, are you sure that's the right question? ...
> > I have lots of these mails, as everyone will, because of virii using
> > forged addresses.
> Are you talking about forged valid addresses, or forged invalid
> addresses?
I have a similar problem althought only slightly different:
I can assume that such frozable message lay in five groups:
1. Normal messages (with a legitimate from header)
2. Errors sent to a forged invalid address
3. Errors sent to a forged valid address
4. Errors sent to a valid address, for valid reasons.
5. valid errors senty to an invalid adrress (since someone misspelled
bot to and from address ! It occours often !)
Ideally one should drop category 2 and 3 and insist on the others.
category 1 is easily recognized ( valid from ), category 5 can be dropped
ince it would never go into the right mailbox.
The remaining are 2,3 and 4: here there is an help:
I can accept that if there is a delivery error and the return address
does not accept within 5 minutes i can drop the error. This way i could
just throw away all error message within such time.
I would not throw away immediately.
My other probles is another:
I could fire a second exim , to get all undelivered NON error messages
and try to delivery with different parameters [namely, through a smarthost
with a different connection] .
What is the command to "Delivery all NON error messages, staying in the
queue for MORE than 9 minutes, NOT having as a destination a domain
as *.abcd.ef " (ideally it should not try to reroute messages whose MX is
in localnet, if i cannot do via an IPmask a could also assume on trailing
part of the domain). ?
--
Leonardo Boselli
Nucleo Informatico e Telematico del Dipartimento Ingegneria Civile
Universita` di Firenze , V. S. Marta 3 - I-50139 Firenze
tel +39 0554796431 cell +39 3488605348 fax +39 055495333
http://www.dicea.unifi.it/~leo
