zero extend pointer converstions to unsigned long long

[Zack Weinberg]

Mike Stump <mrs@apple.com> writes:

>> Please figure out some way to disable the test (or, better,
>> conditionalize it so it does work) on those targets where it is known
>> not to work.
>
> I don't think it makes sense to run the test on non-zero extend 
> machines, as the test tests for zero-extend.

Think you can write a dg-require-zero-extend annotation?

zw