On Fri, 10 Feb 2006 15:42:14 -0800, "John W. Baxter" <jwblist@???> wrote:
> On 2/10/06 3:47 AM, "C?dric MARCOUX" <cmarcx@???> wrote:
>
>> I would like to make a "spamtrap" account on exim using SpamAssassin.
>> Here is what I wan to do
>>
>> * user@??? receive a spam that has not been tagged as spam by
>> SpamAssassin
>> => the user@??? forward this mail to spam@???
>> => the .forward stored in the account spam@??? should add the
>> forwarded spammer email to the SpamAssassin blacklist in
>> /etc/mail/spamassassin/local.cf
>
> What to do plan to do when user1 of your system likes a particular
> message,
> and user2 thinks the same message is spam?
>
> Or, as is very common, attempts to unsubscribe from a mailing list he DID
> subscribe to, by marking it as spam?
>
> And what about the sender whose address was forged in a reported message
> who
> routinely sends mail to a friend who is one of your users, and suddenly
> can't?
>
> Even AOL gives some slack for those cases before blacklisting a sender.
>
>
>
> As to how to do it, one approach would be based on the logfile and
> logwrite
> commands found in Section 3.15 of the Exim filter specification (number
> may
> differ, this is from the Exim 4.50 docs (the version I have handy on my
> desktop machine).
>
> After suitable lines are written, either a periodic process, or something
> tailing the "log" file being written if you want instant gratification,
> would adjust a database which Exim consults to find banned senders.
swatch (perl script,
http://swatch.sourceforge.net/) can be used to read log files and execute a command/script
when a given pattern is recognized. Maybe that could help ?
Mathieu CATTIN
