Martin Gregorie-2 wrote > On Mon, 2014-08-04 at 12:28 -0700, RobertGrimes wrote: >> First, I am running on a Windows 2003 server platform if that makes any >> difference with spamd. The only parameter I use is -l. >> > If your SA configuration files contain any relative address they might > be affected by one of the main differences between Windows and > Linux/UNIX. The difference lies in what the operating system regards as > a program's current working directory (CWD). In windows a program's CWD > is the directory it was loaded from while in most Unices and Linux the > CWD is the directory that the user running the program was in when the > program was started. > > A second difference is that filenames that start with a slash are > unambiguous in Linux/UNIX but IIRC in windows they are relative to the > user's current drive letter. > > In both operating systems relative file names are relative to the > program's CWD, so if the configuration files you're using for SA, Bayes > and/or any databases, Bayes uses contain relative file names, e.g. > configuration/local.cf rather than absolute names, e.g. C:\Program Files > \Spamassassin\Configuration\local.cf and these are defaulted names in a > configuration which assumed SA was running on a Linux/UNIX box, then > obviously SA won't work correctly. > > IOW, look at all the filenames used in your SA configuration and double > check that they make sense in a Windows environment and correct any that > need it. There may be a case for making them all into absolute file > names. > >> I have tried running SA from a command line and I get significantly >> larger >> scores than what is being generated via hMailServer. I was thinking that >> the >> SA headers being added was confusing the command line version. If they >> are >> stripped, why such huge differences, like from about 2 to 10ish? >> > That could be down to the differences I described above. It could also > be due to assess permissions preventing SA from seeing files that don't > belong to the user its running under, but as I've never run anything on > Windows Server I don't know whether this is significant or not, but > under Linux/UNIX this is a common cause of SA problems. > >> Also, I need to do something with BayesDB. >> > Don't forget that Bayes won't do anything until its seen more than 200 > spam and 200 ham messages: you have to train it on the differences > between ham and spam before it can be useful. > > > Martin
I am not sure the reference to Linux as I am only using windows. I have not modified any pathnames so I don't see how the backslash vs forwardslash is relevant unless it is buried somewhere. I am running the spamc command line in the exact same directory as where the spamd runs. They both are running under the same user account so permissions should be equal. I have been running SA for several months and get at least 100 spams a day. Learning should be in force. -- View this message in context: http://spamassassin.1065346.n5.nabble.com/New-at-SpamAssassin-how-to-not-get-headers-tp110712p110722.html Sent from the SpamAssassin - Users mailing list archive at Nabble.com.
