We started with a very simple real expression. A result involving %i
or logs of negative numbers is unreasonable, especially when the
principal value is incorrect, and more especially when there is no
mathematical subtlety involved.
> the ambiguous result might be due to a wrong turn somewhere, but I am
> inclined to think that we can't tell from just looking at the result.
You are apparently limiting the term 'bug' to implementation bugs.
That is, if the implementation performs as intended, it is not a bug.
If we had a complete specification, I suppose I would agree with you
that this might possibly be a 'specification error' rather than an
'implementation error' (= bug). But we do not have a specification,
and anyway I would consider a specification error as much a bug as an
implementation error.
Even if it is not a 'bug' by your definition, it should be logged in
the bug list and at some point get, um, *improved* (we could only say
it was fixed if it was broken...).
-s