Hi,
I've recently installed Spamassassin 2.63 on a RedHat Linux 9.0 with
Sendmail. I've followed William Stearns's guide for integration of
Spamassassin with Sendmail and found this guide very useful. However,
I've got some questions to which I could find no answer and was
therefore hoping that someone could help me with them. Following are the
questions:

1. I've configured shared whitelist and bayes databases as described in
the guide. I've now realised that, whilst the mbox file for the
'spamtrap' account is always 0 in size (/var/spool/mail), the equivalent
'hamtrap' mbox file is not. In fact, it seems to increase as more
falsepositives are reported. Should I be concerned about this? Ideally,
I'd like to delete all mail reported to spamtrap and hamtrap, once the
necessary learning has been done...

2. Secondly, the guide mentions that the second difference between the
.procmailrc files of hamtrap and spamtrap are the lines: learned-spam
vs. #learned-ham. I've realised that this in effect create a file called
learned-spam in /home/spamtrap/mail that stores all the spam messages.
Is this necessary or can I safely delete the file and remove the
learned-spam line from the .procmailrc file?

3. Finally, I'm also trying to configure 'restricting who can report' as
mentioned in the guide. The problem is that I can't find an exact match
to the header fields as mentioned in the guide. Specifically, the guide
makes use of the ^Resent-* header fields, however, I could not find
these fields when I resent messages myself. Could it be that my mail
client (Microsoft Outlook) does not include them? The closest match I
found was a Sender field which I'm currently making use of. Is it ok to
just use this field (There was no equivalent field to the
resent-message-Id field as mentioned in the guide)?

Thanks for all your help.
Joe

Reply via email to