Well, I think the major problem is quality control on new rules. Not infrequently, the first pass at a new rule (particularly eval rules) can cause problems on particularly odd input messages -- normally having a rule in CVS for a while before release allows at least a few dozen people to have it installed for a while trying it out, so we generally know if there's a problem. If we had CVS rules being more widely distributed before they were stabilised it could easily cause some problems. We would need to have a separate release process for rules files, and some extensive testing system to make sure they work OK across a broad cross section of mail (say run each new rule against the corpus before each rules release). It might be possible to set up some automated release mechanism which did this on a weekly basis I suppose.
C Nathan Neulinger wrote: NN> If we did anything like this, I think it would be very important to NN> break the rule files up into two separate pieces to distinguish rules NN> that are very dependent on the particular .pm version - i.e. the eval() NN> stuff. NN> NN> Rules would almost need a major/minor version number like a library. NN> NN> -- Nathan NN> NN> On Tue, 2002-05-28 at 16:32, David T-G wrote: NN> > Hi, all -- NN> > NN> > ...and then Jim Howard said... NN> > % NN> > % Please forgive me and point me there if this has been addressed already... NN> > NN> > Same here. NN> > NN> > NN> > % NN> > % I'm using SA 2.11, and would love to update the rulebase on a regular NN> > % basis, but I don't want to reinstall and config SA again and again. NN> > NN> > I actually had a couple of off-list discussions about this as I was going NN> > through the throes of a non-system-wide installation, but I've kept mum NN> > as I learn how SA works, particularly as I watch for a new release and NN> > then see what's new or different. NN> > NN> > I think of SA as in the same class as a virus scanner; you have the NN> > engine, which changes rarely, and the signature files, which change NN> > often. One should be able to get the latter without having to get the NN> > baggage of the former (much less need an installation as well). NN> > NN> > I also haven't looked around the sourceforge site to see if the rules NN> > themselves are available, but it seems that a directory with the rules NN> > that is available by the web would be easy to keep up to date with wget NN> > to do the mirroring, and away we go. NN> > NN> > Anyway, I'm interested in how to solve the same problem and look forward NN> > to pointers. Maybe I've even completely misunderstood the structure of NN> > SA and need some reeducation :-) NN> > NN> > NN> > TIA & HAND NN> > NN> > :-D NN> > -- NN> > David T-G * It's easier to fight for one's principles NN> > (play) [EMAIL PROTECTED] * than to live up to them. -- fortune cookie NN> > (work) [EMAIL PROTECTED] NN> > http://www.justpickone.org/davidtg/ Shpx gur Pbzzhavpngvbaf Qrprapl Npg! NN> > NN> _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _______________________________________________ Spamassassin-talk mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
