On 5-Dec-08, at 10:57 AM, daveloeffler wrote:
>
> I think you've uncovered a bug that had been masked by the is_one bug.
> If I'm not mistaken, I^-1 calls MultiplicativeGroupElement.__invert__,
> which tests whether self is one and if not returns 1/self. Obligingly,
> the coercion model (being a
That sounds right (and I'm sure it is).
What is worrying me slightly is that I have a vague memory that I once
edited that generic __invert__() code (possibly it was me who added
the is_one() test.
I would prbably be best to have __invert__() NotImplmented in the base
class, forcing all derived
I think you've uncovered a bug that had been masked by the is_one bug.
If I'm not mistaken, I^-1 calls MultiplicativeGroupElement.__invert__,
which tests whether self is one and if not returns 1/self. Obligingly,
the coercion model (being asked to calculate 1/self) calls __invert__
again!
(I conv
