Sometimes blacklists are bad

I blog this quite late as it happened to me yesterday, but that was a long long day and I forgot to do a lot of things.

One of the things that pissed me off yesterday morning when I woke up was to find mails from OpenOffice’s bug tracker, and from a contact for a new job put in the Junk folder.

A quick look at it told me that the problem were the URIBL_*_SURBL checks from SpamAssassin, giving high rates to almost every address possible, present in the mails, included openoffice.org and firebirdsql.org.

The first thing I did was to set this in my local.cf for spamassassin:

score   URIBL_AB_SURBL  0.0
score   URIBL_PH_SURBL  0.0

but then I got the mails garbled by spamassassin’s safe report.. took me a while to find in spamassassin -d the correct command to kill that extra data written during the check.

So, why am I blogging all this? Well, to give an hit to the KMail users: if you set up SpamAssassin integration in KMail you got a “Filter Classify as NOT Spam” action to run on mails, that by default only uses sa-learn to bias the bayesian filter; if you go in “Configure Filters” from the “Settings” menu, you’ll find the configuration for that action; add a new entry on the filter actions “Pipe Through” spamassassin -d; now, every time you mark something as NOT spam even when it has been filtered through spamassassin already it will come back to its natural shape.

And this is it for this week’s issue of Flameeyes’s microscopic tips & tricks.