https://bz.apache.org/SpamAssassin/show_bug.cgi?id=7735
--- Comment #50 from Wolfgang Breyha <wbre...@gmx.net> --- All that I can say is that meta rules did/do not work as expected with SA4. This might be the case because of the one issue which is not in bugzilla yet which I sent you personally and I've patched myself where finished meta rules do not trigger a run of dependent meta rules. But still, even patched, I see meta rules which have no result at all if they depend on an undefined (meta) rule, because they stay unrun till the end. And I ask myself what happens exactly if you run the same rule twice with different parameters for undefined dependencies? Which result is taken at last? If it is a list of "||" the one run with "1" will always result in a hit. This is most likely not intended. And what happens if more then one binary or arithmetic clause is "unrun" and you have to run it even more often with all combinations? The 3.4.x style to handle "unhit" rules as 0 was easy to understand and clear. And undefined was unhit. >From my current experience with SA4 this change kills 3rd party rulesets like KAM. I will ask Kevin McGrail to comment on this topic as well. -- You are receiving this mail because: You are the assignee for the bug.