[Mimedefang] problems with bayes (?) database access

[alan premselaar]

On 5/23/03 10:17 PM, "Dirk Mueller" <dmuell at gmx.net> wrote:

> On Fre, 23 Mai 2003, alan premselaar wrote:
> 
>> yesterday when this happened, i stopped sendmail/mimedefang, killed the lock
>> files, ran sa-learn --rebuild and then start sendmail/mimedefang again, it
>> seems to work for awhile, and do it again.
> 
> spamassassins autowhitelist or bayes database currently don't work with
> mimedefang, because they don't lock the database so concurrent access will
> destroy them. Don't use it, or if you do, then don't complain that it
> doesn't work ;-)

Interesting.  I've been using bayes w/o any problems until just a day or two
ago.  I don't use auto whitelist. but i'll keep that in mind.
> 
> Alternatively you can filter the messages through spamc (use
> action_external_filter()) and then you can enable auto-whitelisting and
> bayes again. however, its useless to enable bayes if you don't maintain it
> manually. It will even hurt your false positive rate badly.
> 
> 
>> I've just turned off the RCVD_IN_ORBS test as well as there's been
>> discussion that it no longer exists.
> 
> Use SpamAssassin 2.55
> 

according to the release notes of 2.55 it was only released to fix an issue
with an incompatibility with perl 5.005, which i'm not using.

apparently the ORBS service recently "disappeared off the face of the
earth", and many people on the SATalk list were having issues with timeouts
related to not being able to connect to it.

anyways, thanks for the input.


alan