R: R: Evaluating how good is a rule

2007-05-07 Thread Giampaolo Tomassoni
> -Messaggio originale- > Da: Luis Hernán Otegui [mailto:[EMAIL PROTECTED] > > Well, you could try asking the amavis list, tough. I think you could > actually tell amavis to copy the messages which hit a certain rule to > a certain directory, or to look for a specific tag by SA. But this i

Re: R: Evaluating how good is a rule

2007-05-07 Thread Luis Hernán Otegui
Well, you could try asking the amavis list, tough. I think you could actually tell amavis to copy the messages which hit a certain rule to a certain directory, or to look for a specific tag by SA. But this is getting kinda off-toopic, I think... Regards, Luix 2007/5/7, Giampaolo Tomassoni <[EM

R: Evaluating how good is a rule

2007-05-07 Thread Giampaolo Tomassoni
> -Messaggio originale- > Da: Luis Hernán Otegui [mailto:[EMAIL PROTECTED] > > Well, it all depends on how do you run SA (as a content filter, > through AMAVIS, via procmail, etc). Via AMAVIS you could use the log > parser which MrC wrote, which works like a charm here. It could work > as

Re: Evaluating how good is a rule

2007-05-07 Thread Luis Hernán Otegui
Well, it all depends on how do you run SA (as a content filter, through AMAVIS, via procmail, etc). Via AMAVIS you could use the log parser which MrC wrote, which works like a charm here. It could work as a control, since it tell you which rules hitted harder, which ones didn't, etc. SA cannot cop

Evaluating how good is a rule

2007-05-06 Thread Giampaolo Tomassoni
Dears, I'm something in need to write custom rules in order to penalize some kind of messages running through my SA. Now, of course I do apply a 0.001 score to my own test rules in order not to create FPs on received mails. Also, things are setup such that no SA header is added to messages yieldi