hi This problem was reported by many other users
Situation arises when globaluser is trained using dspam_train globaluser path_to_spam path_to_ham dspamhome/group file contains globalgroup:classification:*globaluser As per dspam readme this is used to provide out-of-the-box filtering to all new users problems immaterial of whether the mail is spam or ham the dspam header result would always be innocent we also noted that the header contained exactly the same values as below for all emails spam and ham X-DSPAM-Result: Innocent X-DSPAM-Confidence: 0.6000 X-DSPAM-Probability: 0.0000 Resolution this was resolved when we upgraded to dspam version 3.8 you can download it here http://www.zdziarski.com/projects/dspam/sources/dspam-3.8.0.tar.gz NOTE : this is still not released as a stable version so please take care. we will be using this on for the next 7 days on a test basis and updating this thread regds rajesh mahadevan
