[Spambayes] SpamBayes 1.0a9 (0.9)

Mikhail Yakoubov qub at qub.com
Thu Feb 19 11:03:58 EST 2004


Tony Meyer wrote:
 
> > No, CVS sb_imapfilter.py 1.26 (and 1.27 too) won't work this way --
> [...]
> 
> You need message.py from CVS as well.  Sorry - I wrote the message to you
> before I had finished checking in the changes and testing them.  The FAQ
> about this problem does say that you need both.

Yes, it works this way, but has _very_ undesired side effect. Since I've
substituded both files with newer versions yesterday, ham get stuck in
my Inbox. That is, my client refuses to consider it newly arrived, as it
was before, and thus don't apply filtering rules to it. So, ham don't
get sorted, and I have to sort it manualy, as Messenger's filtering
mechanic works only on the messages in Inbox it considers newly arrived.

Not sure what flags are responsible for the message to be considered
fresh by client, but the fact is that before the change Spambayes
treated them client-friendy (was completely trasparent, avoided
interfering with future client processing), and has stopped doing so
now. At least, now the ham messages get marked as Read, and before they
were Unread. The same Read flags are on now on all the Spam and Unsure
messages sorted by Spambayes itself, not the way it was before.

FAQ at
http://spambayes.sourceforge.net/faq.html#sb-imapfilter-prints-out-skipping-unparseable-message-but-the-message-vanishes
states that one need message.py (revision 1.46) to fix the multipart
problem, which is obviously an error (1.46 is the one included in the
release_1_0_a9 itself). 1.48 seems to be the right choice, in view of
its timestamp.

I use
http://cvs.sourceforge.net/viewcvs.py/*checkout*/spambayes/spambayes/scripts/sb_imapfilter.py?rev=1.27
http://cvs.sourceforge.net/viewcvs.py/*checkout*/spambayes/spambayes/spambayes/message.py?rev=1.48
not the 1.26 and 1.46 as suggested by the FAQ.


Regards,
Mike Yakoubov.



More information about the Spambayes mailing list