Kang, Joseph S. said: > [snip] > >> From what I've seen the raw headers are not changed with the >> exception of the MailScanner spamassassin score which may be >> different because of RBL-SURBL's that have picked up on it. >> This doesn't appear to be a problem because sa-learn ignores >> the spam markup anyway. When I first implemented this I tried >> a Public folder but saw that the headers were hosed which is >> why I went to a regular user folder. This seems to work very well. > > I had meant to send a message to the list to ask for clarification on part > of this conversation. I'm curious, for the people that are seeing this TNEF > header format, what version of Exchange are you running? > > For better or worse, we're still using Exchange 5.5 and I'm set up with two > public folders for users to move/copy messages into. I use one of the > previously mentioned Perl IMAP scripts from my Linux sendmail gateway box to > perform additional training to my Bayes DB. > > -Joe K.
Hi Joe, I have seen this problem on Ex2K and Ex2K3, I don't remember if it was an issue under 5.5 or not. I have a copy of a test-imap-header.pl script written by Nick Burch that can be used to see what the headers look like after you pull a message from an IMAP folder... let me know if you would like a copy of the script. --matt
