On 2016-08-29 22:42, Axb wrote:
On 08/30/2016 07:32 AM, jdow wrote:
On 2016-08-29 17:51, li...@rhsoft.net wrote:
Am 30.08.2016 um 02:45 schrieb John Hardin:
On Mon, 29 Aug 2016, Anthony Hoppe wrote:
I just learned about the sought ruleset via
https://wiki.apache.org/spamassassin/ImproveAccuracy. Is this ruleset
still actively maintained? I'm considering implementing it in my
environment, but want to make sure just in case.
Sadly, no. I think it's been at least a couple of years since they were
regenerated
but they still hit junkmails and are even part of the fedora default
install
pulled with the first "sa-update"
rpm -q --file /etc/mail/spamassassin/channel.d/sought.conf
spamassassin-3.4.1-9.fc24.x86_64
cat /etc/mail/spamassassin/channel.d/sought.conf
# http://wiki.apache.org/spamassassin/SoughtRules
CHANNELURL=sought.rules.yerp.org
KEYID=6C6191E3
# Ignore everything below.
return 0
The late lamented SARE rule sets also work remarkably well if you
captured a suitable snapshot.
{^_-}
But some are a huge source of FPs...
After SARE shutdown I dumped the last release (for posterity)
https://sourceforge.net/p/sare/code/HEAD/tree/rules/DO-NOTUSE-rulesets-obsolete-/
be VERY carefull when playing with them - they are a huge 7 year old can of
worms.
They sure can be. I spent time winnowing the particularly bad rules out. I still
make custom rules that are quite helpful, especially for large mailinglists such
as the LKML. I use a collection of meta rules to expand Bayes scores away from
about 80. Otherwise I get too much ham getting punished for too little spam
getting punished. Now it errs somewhat too far towards ham getting through
meaning LKML leaks spam. But, what gets through is usually quite humorous
considering the LKML relay.
{^_^} Joanne says, "One woman's spam is another man's ham. How ya gonna score
it?"
