Hi Lazarus, sorry for the late reply. I had some talks with upstream about your wish but apparently he's too busy at the moment to take care of it.
On Sun, Nov 02, 2008 at 09:09:21AM +0000, Lazarus Long wrote:
> The script doesn't deal well with SpamAssassin when this one is configured to
> store AWL and Bayesian rules in SQL containers.
I was trying to understand what your patch does but I still have some
issues with it. :) Can you send me a spamassassin configuration with SQL
based rules configured? I didn't find reliable information about it and
your suggestions do have a point.
E.g.:
> It may benefit with better error checking, especially if 'bayes_path' isn't
> defined but SQL isn't configured as the storage for SpamAssassin, and the dual
This one is definitly worth some effort.
> meaning of 'sa:user' should be documented in the configuration file or a new
> option should be defined to allow loading AWL and Bayesian rules to a single
> user ('amavis' should be the default) on the SQL database.
And I don't think it's good to have variables with double meaning. I'm
not ready with reviewing but it's better to be clear in choice of
variables.
> Also some pointers to the configuration of SpamAssassin for SQL would ease the
> usage of this package in those circunstances, for instance a reference to
> </usr/share/doc/spamassassin/sql/README*>.
Better documentation is always good. :) I'll have a look at it when I'm
more into your patch.
Thanks for your contribution, anyways. I'll try to release a patched
version ASAP but I'll probably need your help since I don't have a test
environment available here.
Hauke
signature.asc
Description: Digital signature
