I wrote: > I don't see a way to indicate that a maxSet relationship need only > hold if the size is nonzero. Perhaps maxSet(NULL) == 0 should be > assumed?
Nope, silly me, I keep thinking "how many" when I should be thinking "max index", so 0 is no good, it means one location can be set. Perhaps -1? Changing the literal value constructed in my patch still produces a version that accepts my test case. Ken