On Thu, Dec 26, 2002 at 05:09:57PM -0500, Jeremy Nixon wrote: > I suspect my problem is related to the order in which meta rules are > evaluated, or the "legality" of nesting them. Is there some limit to > the depth to which meta rules can be nested inside other meta rules? > I had one like "meta __RULE (__FOO && __BAR)" which didn't work, but > when I replaced the __BAR with a copy and paste from the "meta __BAR" > definition, it suddenly worked.
There's a bug about this in bugzilla, #1187: http://bugzilla.spamassassin.org/show_bug.cgi?id=1187 In the bug, I wrote up a pseudo-code algorithm for the meta rules to deal with dependencies, but we haven't had time to get it actually integrated into the SA code yet. With a 2.50 release being actively worked on, this one is being held off until 2.6/3.0. > If I'm right, and the problem is the order in which the rules are > being evaluated, it should be easy enough to fix; best would be to > do them in the order they're defined, but alphabetical would probably > be easier to implement with a strategically placed sort() call. I think you're right. The bug mentions running meta tests alphabetically, but that may have been a "this is what we can do" rather than a "this is what currently happens". I'll look at the code and see what I can come up with. I'd think, as you say, a strategically placed sort() would do the trick for the time being. -- Randomly Generated Tagline: "For all those who say that Microsoft never creates anything original, just remember that they have now invented the cross-platform virus." - Phillip Karlsson
msg11464/pgp00000.pgp
Description: PGP signature
