On Fri, Jan 01, 2010 at 07:45:55PM +0100, Vincent Lefevre wrote: > I reported 5 bugs against a package, so that I received a reply with: > > Cc: 563...@bugs.debian.org, 563...@bugs.debian.org, 563...@bugs.debian.org, > 563...@bugs.debian.org, 563...@bugs.debian.org > > which triggered the rule SUSPICIOUS_RECIPS with 3.2 points. For > a suspicious-something rule, the number of points is quite high. > Actually spam with similar recipients generally has much more > recipients. This rule should be fixed.
So, this bug is nearly 7 years old, but is still present and legitimate today. The SUSPICIOUS_RECIPS test looks for similar recipients for a given message, which is a common spam trait, and, at least based on the regular masschecks that the Spamassassin project runs, not a common trait of legitimate mail. One thing we could consider is whether to provide some negative score rules to ensure that mail from the BTS isn't flagged as spam, similar to other rules we currently provide in /etc/spamassassin/65_debian.cf. That'd help with this specific case, but not globally. I think that this is enough of a special case that that may be sufficient. In the immediate term, if this is still a concern at all, I recomment local configuration to whitelist mail from the BTS. noah
signature.asc
Description: PGP signature
