So I've had dspam 3.9.0 up and running but have never had any luck getting it to "train" using your script. Every time the script kicks off I get the following in my logs. It will delete the messages but there's no training going on. These to lines appear for every message in a .Junk folder (with different signature numbers of course). Got 3.9.1 rc1 up and going tonight and it's the same issue.
Richard have you run into this?
dspam[31337]: Signature retrieval for '1,4c24317f145301375414451' failed
dspam[31337]: Unable to find a valid signature. Aborting.
ya I've never been able to train dspam after snow leopard upgrade and dspam 3.9.0 If you are running it now how did you train it? I've seen at least 100k messages of spam and never has one been tagged as spam.
The dspam database has updates getting written to it but none of it's spam related - it's all spam_hits = 0 and the dspam_stats table shows a bit fat 0 for spam_learned. It's learned lots of innocent's somehow however.
I have to acknowlegde shamefully that my production server is running 3.6.8. Still need to perform the upgrade. There where issues with 3.8 and 3.9.0 therefore I could postpone it. But now there is no more excuse. All problems are solved ;-)
Will post more info when I have it, curently my setup is as described but I think it will change.
#
So I've had dspam 3.9.0 up and running but have never had any luck getting it to "train" using your script. Every time the script kicks off I get the following in my logs. It will delete the messages but there's no training going on. These to lines appear for every message in a .Junk folder (with different signature numbers of course). Got 3.9.1 rc1 up and going tonight and it's the same issue.
Richard have you run into this?
dspam[31337]: Signature retrieval for '1,4c24317f145301375414451' failed
dspam[31337]: Unable to find a valid signature. Aborting.
#
I've had 3.9.0 running but never tested training yet. I'm currently looking at 3.9.1 RC1.
However, please note that there was a small issue in 3.9.0 that interfered with the classification.
#
ya I've never been able to train dspam after snow leopard upgrade and dspam 3.9.0 If you are running it now how did you train it? I've seen at least 100k messages of spam and never has one been tagged as spam.
The dspam database has updates getting written to it but none of it's spam related - it's all spam_hits = 0 and the dspam_stats table shows a bit fat 0 for spam_learned. It's learned lots of innocent's somehow however.
#
I have to acknowlegde shamefully that my production server is running 3.6.8. Still need to perform the upgrade. There where issues with 3.8 and 3.9.0 therefore I could postpone it. But now there is no more excuse. All problems are solved ;-)
Will post more info when I have it, curently my setup is as described but I think it will change.