[PATCH 6/7] quartz/tests: Add tests for MPEG audio decoder sample processing.

[Anton Baskanov]

On 04.05.2022 04:18, Zebediah Figura (she/her) wrote:
> On 5/3/22 13:28, Anton Baskanov wrote:
>> Signed-off-by: Anton Baskanov <baskanov at gmail.com>
>> ---
>>    dlls/quartz/tests/mpegaudio.c | 83 +++++++++++++++++++++++++++++++++++
>>    1 file changed, 83 insertions(+)
>>
>> diff --git a/dlls/quartz/tests/mpegaudio.c b/dlls/quartz/tests/mpegaudio.c
>> index def5dc90238..0044d3bef6f 100644
>> --- a/dlls/quartz/tests/mpegaudio.c
>> +++ b/dlls/quartz/tests/mpegaudio.c
>> @@ -846,6 +846,7 @@ struct testfilter
>>        struct strmbase_source source;
>>        struct strmbase_sink sink;
>>        const AM_MEDIA_TYPE *mt;
>> +    unsigned int got_sample;
>>    };
>>    
>>    static inline struct testfilter *impl_from_strmbase_filter(struct strmbase_filter *iface)
>> @@ -919,11 +920,39 @@ static HRESULT testsink_connect(struct strmbase_sink *iface, IPin *peer, const A
>>        return S_OK;
>>    }
>>    
>> +static HRESULT WINAPI testsink_Receive(struct strmbase_sink *iface, IMediaSample *sample)
>> +{
>> +    struct testfilter *filter = impl_from_strmbase_filter(iface->pin.filter);
>> +    unsigned int i;
>> +    BYTE *data;
>> +    HRESULT hr;
>> +    LONG size;
>> +
>> +    ++filter->got_sample;
>> +
>> +    size = IMediaSample_GetSize(sample);
>> +    ok(size == 3072, "Got size %lu.\n", size);
>> +    size = IMediaSample_GetActualDataLength(sample);
>> +    ok(size == 768, "Got valid size %lu.\n", size);
>> +
>> +    hr = IMediaSample_GetPointer(sample, &data);
>> +    ok(hr == S_OK, "Got hr %#lx.\n", hr);
>> +    for (i = 0; i < size / 2; ++i)
>> +    {
>> +        short value;
>> +        memcpy(&value, &data[i * 2], 2);
>> +        ok(abs(value) <= 1, "Got value %d.\n", value);
> 
> I don't know if we really need to bother with validating the output data.
> 
> (If we do, though, it seems a bit more idiomatic to just alias "data"
> with a "const short *" pointer.)
> 

Oh, I didn't realize that we are compiling with -fno-strict-aliasing, so 
I thought that this would break the aliasing rules.

Anyway, I've removed the data check in v2.