[PATCH 05/10][try 2] quartz/tests: Fix interface leaks in aggregation test.

[Alexander Dorofeyev]

Michael Karcher wrote:
> Am Donnerstag, den 17.07.2008, 20:24 +0300 schrieb Alexander Dorofeyev:
>> +    do
>> +    {
>> +        hr = IUnknown_Release(pUnkInner);
>> +    } while (hr);
> Release returns a ULONG, not a HRESULT. Do we ignore the difference in
> Wine?

I usually do not, but that's how it happens in surrounding code there 
(ADDREF_EXPECT macro etc). IMO this is no big deal, but if this is a causing 
questions I guess I'll change it to use an ULONG variable and and fix the 
existing code to do the same for the sake of consistency.