Carl Worth wrote: > I'm not sure what you're arguing here. > > Are you saying that this bug is so common that we can assume that > platforms that don't give the desired behavior for >> with negative > numbers must not exist? Yes, this accurately describes what I am thinking.