Am 2018-12-05 15:59, schrieb Ralph Seichter: > From rspamd's perspective, such a high score means that the message is > almost certainly spam. What would be the point of "learning" something > that is already known? To train the bayes filter.