[PATCH 2/4] mf/tests: Check H264 decoder first NV12 decoded frame.

[Rémi Bernon]

On 2/24/22 11:22, Rémi Bernon wrote:
> On 2/24/22 11:18, Nikolay Sivov wrote:
>>
>>
>> On 2/24/22 12:47, Rémi Bernon wrote:
>>> On 2/24/22 10:39, Marvin wrote:
>>>> Hi,
>>>>
>>>> While running your changed tests, I think I found new failures.
>>>> Being a bot and all I'm not very good at pattern recognition, so I 
>>>> might be
>>>> wrong, but could you please double-check?
>>>>
>>>> Full results can be found at:
>>>> https://testbot.winehq.org/JobDetails.pl?Key=109200
>>>>
>>>> Your paranoid android.
>>>>
>>>>
>>>> === w1064_tsign (64 bit report) ===
>>>>
>>>> mf:
>>>> mf.c:6939: Test failed: ProcessOutput returned 0xc00d6d72
>>>
>>> I'll have a look at this failure, sadly the H264 transform seems to 
>>> be asynchronous and the results aren't easily predictable.
>> In general if async mode is supported, it should be explicitly 
>> enabled. Does it really help if you sleep for a bit?
>>
> 
> Yes it helps, I didn't want to do that as it's always a bit dubious, and 
> hoped that a robust loop should be enough, but maybe there's no other 
> option.
> 
> It may not be truly async in the MF sense, but it definitely internally 
> queues input samples and processes them asynchronously.

I think there's also MF_LOW_LATENCY mode attributes which I believe 
helps too, I'll find a way.
-- 
Rémi Bernon <rbernon at codeweavers.com>