Russell Jones wrote: > Alrighty, now that I've had some time to look at it, I would like to > make sure I understand this correctly. > > In the man page for it, it says: bayes_path /path/filename > (default: ~/.spamassassin/bayes) > > So, if I change the ~/.spamassassin/bayes to, for example, > /var/spamassassin/bayes, then instead of each user having their own > bayes database, all information learned from all of the mail accounts > handled by this server will be stored in the /var/spamassassin/bayes > folder, and will be used accordingly, correct? No, it will be stored in the /var/spamassassin/ folder. For the above setting to work, there MUST NOT be a folder named /var/spamassassin/bayes/.
The bayes part is part of the filename, not a part of the path. That's the tricky part. that I warned you about. Basically, whatever you pass, SA will append "_toks" to it and use that as the resulting filename as your token database. so bayes_path /var/spamassassin/bayes causes SA to create the file /var/spamassassin/bayes_toks, NOT /var/spamassassin/bayes/bayes_toks. Matt Kettler wrote: > > WARNING 2: read the docs on bayes_path VERY carefully. Note that the > parameter is a path AND a partial filename, not just a path. > ie: bayes_path /var/spamassassin/bayes/bayes will cause SA to > create the file " /var/spamassassin/bayes/bayes_toks"
