On Thu, 2010-08-12 at 14:58 +0200, Mark Martinec wrote: > Vernon, > > > I have a few emails with attachments that have been quarantined as spam. > > I would normally use altermime to strip the attachment and then sa-learn > > --ham. Is this still the recommended way of unlearning messsages with > > attachments? > > I see no good reason for stripping attachments before learning. > Just feed the quarantine file to sa-learn or to spamassassin --report > e.g.:
A while back, i set the $sa_mail_body_size_limit to undef to combat emails with large jpeg and pdf attachments, well not large but larger that the 200*1024 limit. These messages were detected under this setting. When attempting to unlearn (sa-learn -D --ham (or --forget) {message}), it returns: [16026] info: archive-iterator: skipping large message Forgot tokens from 0 message(s) (0 message(s) examined) I've been using Amavis (then amavisd-new) and spamassassin long enough to realize that i need to re-educate myself of the proper techniques, hence the original question. Been doing ALOT of reading - many different opinions and trying to get the most optimal configurations to handle anywhere from 15k to 60k messages per day - small but still important. Does anyone have suggestions on the optimal message size limit or will the default 200*1024 suffice. Vernon ------------------------------------------------------------------------------ This SF.net email is sponsored by Make an app they can't live without Enter the BlackBerry Developer Challenge http://p.sf.net/sfu/RIM-dev2dev _______________________________________________ AMaViS-user mailing list AMaViS-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amavis-user Please visit http://www.ijs.si/software/amavisd/ regularly For administrativa requests please send email to rainer at openantivirus dot org