On Sat, 04 Oct 2014 11:32:17 +0200 Axb wrote: > Unless somebody thinks this a terribly bad idea, I'll be adding a > 20_bayes_ignore_header.cf to the SA default rules to replace the few > (unmantained) bayes_ignore_header entries in local.cf > > comments?
I'd like to see some evidence that these headers are harmful before it's set unconditionally like that. I had a quick look at my own mail and found that barracuda headers made it through to X-Spam-tokens in only about 0.2% of spam and ham; and when they did they were mostly in the spam token lists for spam and equally spread between the two lists for ham.
