> Confused... Do you think something is wrong with the current code? No, I was just being explicit about the nonobvious quirks of the semantics. (They merit some comments in the eventual code.)
> IOW, I assume this test-case [...] > is right, correct? Yes. Thanks, Roland
