[Spambayes] SpamBayes 1.1a4 on Windows XP (correction)

[Kyle Allender]

skip at pobox.com wrote:
>     Kyle> I also discovered today that the proxied messages are being stored
>     Kyle> in the directory:
>
>     Kyle> C:\Documents and Settings\NetworkService\Application 
>     Kyle> Data\SpamBayes\Proxy\pop3proxy-spam-cache
>
>     Kyle> I found these when AVG performed its scan of the system and found
>     Kyle> quite a few trojans hiding in this directory.  So it would appear
>     Kyle> as though the training begins to occur with messages being cached,
>     Kyle> but the database isn't being written to perhaps?
>
>     Kyle> Can anyone shed light on this situation?
>
> Mail comes into the system and the pop3proxy caches them for your later
> consideration when training.  SpamBayes doesn't execute any of this code and
> no other programs in your system should access them.  Still, if this is an
> issue for you, you should be running some sort of virus scanner on the front
> end which scrubs viruses from email attachments before any other tool (like
> Outlook, Thunderbird or SpamBayes) sees the message.
>
>   
> ------------------------------------------------------------------------
>
>
> No virus found in this incoming message.
> Checked by AVG - www.avg.com 
> Version: 8.5.325 / Virus Database: 270.12.21/2103 - Release Date: 05/07/09 18:05:00
>
>   
That's pretty much how I figured it was supposed to work.  The caveat 
here is that I have not yet configured the proxy in the spambayes 
config.  I had hoped to train on previously collected spam messages 
first, _then_ set up the proxy so that the appropriate filtering would work.

Since I don't have anything executing attachments in Thunderbird, I'm OK 
with the messages hitting spambayes and then AVG AntiVirus.  The reverse 
would also be OK.  My main goal is to get rid of the spam in my inbox - 
I can deal with the antivirus as a separate issue.

If I've not configured the proxy why would there be cached spam 
messages?  Is the training partially succeeding even though the logs 
indicate otherwise?

K