Re: Is it not recommanded to learn a message already flaged as spam?
On Thu, Dec 02, 2004 at 10:04:27AM -0500, Matt Kettler wrote: At 10:17 AM 12/2/2004 +0100, Nicolas wrote: With mutt, I'd like to define a macro which learn the mail as spam, report it to razor, and delete it. I'd like to know if it is not recommanded to learn a mail as spam, while it's already flaged as spam by SA? It IS recommended to learn mail that's already been flagged. Even if it's flagged BAYES_99 SA can still learn worthwhile tokens from a message. sa-learn recognizes SA's own spam tags, and will automatically strip those out before learning it. The only thing I'd avoid in training messages is I'd not intentionally train the same message twice. But even this is only because it's a minor waste of time.. SA will just ignore them, no harm done, but it's pointless to go out of your way to retrain the same message. Also, if you use spamassassin -r on the message, it will strip tags, learn as spam, and report it to razor, spamcop and any other hash systems you have installed (ie: dcc or pyzor.) So all your macro needs to do is call spamssassin -r message.txt and then delete the message. Thanks for the help, guys! Nicolas -- --- OxStOnE -- O - Z750 Linux --- ._ /\_ --- Powered -- (x) (x) ~~~
Re: Is it not recommanded to learn a message already flaged as spam?
At 10:17 AM 12/2/2004 +0100, Nicolas wrote: With mutt, I'd like to define a macro which learn the mail as spam, report it to razor, and delete it. I'd like to know if it is not recommanded to learn a mail as spam, while it's already flaged as spam by SA? It IS recommended to learn mail that's already been flagged. Even if it's flagged BAYES_99 SA can still learn worthwhile tokens from a message. sa-learn recognizes SA's own spam tags, and will automatically strip those out before learning it. The only thing I'd avoid in training messages is I'd not intentionally train the same message twice. But even this is only because it's a minor waste of time.. SA will just ignore them, no harm done, but it's pointless to go out of your way to retrain the same message. Also, if you use spamassassin -r on the message, it will strip tags, learn as spam, and report it to razor, spamcop and any other hash systems you have installed (ie: dcc or pyzor.) So all your macro needs to do is call spamssassin -r message.txt and then delete the message.
Re: Is it not recommanded to learn a message already flaged as spam?
On Thu, Dec 02, 2004 at 10:17:55AM +0100, Nicolas wrote: With mutt, I'd like to define a macro which learn the mail as spam, report it to razor, and delete it. I'd like to know if it is not recommanded to learn a mail as spam, while it's already flaged as spam by SA? That's because I see some of my mails which are flagged as spam which aren't listed in razor. I don't want to use several keys (one to learn spam, one to report to razor). Reporting spam via spamassassin -r will report it to Razor (also pyzor, DCC and SpamCop if they configured) and learn the mail as spam. No need for multiple macros. Yes, I would recommend it. Michael pgpuJEBJP2E1K.pgp Description: PGP signature
RE: Is it not recommanded to learn a message already flaged as spam?
Cool thanks, I was getting kind of confused. So I guess my next task will be to add razor. thanks -Original Message- From: Matt Kettler [mailto:[EMAIL PROTECTED] Sent: Thursday, December 02, 2004 7:04 AM To: Nicolas; spamassassin-users mailing-list Subject: Re: Is it not recommanded to learn a message already flaged as spam? At 10:17 AM 12/2/2004 +0100, Nicolas wrote: With mutt, I'd like to define a macro which learn the mail as spam, report it to razor, and delete it. I'd like to know if it is not recommanded to learn a mail as spam, while it's already flaged as spam by SA? It IS recommended to learn mail that's already been flagged. Even if it's flagged BAYES_99 SA can still learn worthwhile tokens from a message. sa-learn recognizes SA's own spam tags, and will automatically strip those out before learning it. The only thing I'd avoid in training messages is I'd not intentionally train the same message twice. But even this is only because it's a minor waste of time.. SA will just ignore them, no harm done, but it's pointless to go out of your way to retrain the same message. Also, if you use spamassassin -r on the message, it will strip tags, learn as spam, and report it to razor, spamcop and any other hash systems you have installed (ie: dcc or pyzor.) So all your macro needs to do is call spamssassin -r message.txt and then delete the message.