Daniel said, "One minor note: "spamasassin" does not use spamd. To use spamd, you need to run "spamc"."
Right. Just being thorough. The point is that when I call spamasassin directly (rather than via spamc/spamd) I get a valid return result. Daniel said, ""get hosed"? I suggest you work on the descriptiveness there" I was just trying to save space. When I said "get hosed" I meant it behaved as I described in the first message in the thread. Specifically, it returns the "0/0" error result whenever it is given a message that would be marked as spam (as evidenced by passing the same message through spamassassin directly). Daniel said, "Also, the wiki I mentioned has an explicit section on how to test SpamAssassin." Yeah, I went through the wiki and google before hitting the listserv. I hate to ask questions for which the answers are already available. I tested my stuff as per the wiki. Among other tests (as described in the first message), that is how I came to understand the pattern of the error; which is to say that spamc returns an error when it checks spam, but not when it checks ham. Since it defaults to safely /not/ marking something as spam if it errors out, I get no spam hits. Daniel said, "[...] then it's 99% likely installed wrong." I used apt-get on two different machines and got the same error separately. Specifically, I used Synaptic on Fedora Core 2 to install the 2.64-2.1.fc2.dag spamassassin package from the Dag Wieers repository. I did the same on the other machine except that it is Fedora Core 1, and the package was 2.64-2.1.fc1.dag. Daniel said, "If your "spam" is not _really_ spam (READ THE WIKI), then well, argh!" The spam is really spam. Again, if I call spamassassin directly, it marks it as such. It's only when I call it via spamc/spamd that I get the problem. In the /var/log/maillog file, I see that when it checks a message that is not spam, I get this entry: Sep 14 17:54:36 localhost spamd[2034]: connection from localhost.localdomain [127.0.0.1] at port 33289 Sep 14 17:54:36 localhost spamd[4844]: info: setuid to tom succeeded Sep 14 17:54:36 localhost spamd[4844]: checking message <[EMAIL PROTECTED]> for tom:500. Sep 14 17:54:36 localhost spamd[4844]: clean message (0.0/5.0) for tom:500 in 0.7 seconds, 3032 bytes. Whereas, if the message is spam, (and testing it via a direct call to spamassassin at the command line confirms this!) then the log looks like this: Sep 14 18:09:51 localhost spamd[2034]: connection from localhost.localdomain [127.0.0.1] at port 33312 Sep 14 18:09:51 localhost spamd[4895]: info: setuid to tom succeeded Sep 14 18:09:51 localhost spamd[4895]: checking message <[EMAIL PROTECTED]> for tom:500. Note that in the second case, there is never an "indentified spam" return result, nor a "clean message" return result. It returns nothing according to the log. I am totally at a loss as to what could cause this behavior. Is there something else I can try to further troubleshoot it? -Tom Caudron
