Re: [PATCH 0/1] MR8172: mfreadwrite: Fix stride in media type ouput.

[Nikolay Sivov (＠nsivov)]

Nikolay Sivov (@nsivov) commented about dlls/mfreadwrite/reader.c:
> - if (SUCCEEDED(hr))
> - {
> -
>    GUID major_type, subtype;
> -
>    UINT64 frame_size;
> -
>    LONG stride;
> -
> -
>    type = reader->streams[index].current;
> -
> -
>    if (SUCCEEDED(IMFMediaType_GetMajorType(type, &major_type))
> -
>            && IsEqualGUID(&major_type, &MFMediaType_Video)
> -
>            && SUCCEEDED(IMFMediaType_GetGUID(type, &MF_MT_SUBTYPE, &subtype))
> -
>            && SUCCEEDED(IMFMediaType_GetUINT64(type, &MF_MT_FRAME_SIZE, &frame_size))
> -
>            && SUCCEEDED(MFGetStrideForBitmapInfoHeader(subtype.Data1, frame_size >> 32, &stride)))
> -
>        IMFMediaType_SetUINT32(type, &MF_MT_DEFAULT_STRIDE, abs(stride));
> - }
> -

If we really need this it should go to helpers where type is set from the source or from a transform. Right now it looks like a workaround to me. How can we be sure that it's a reader responsibility to modify type like this, and not something missing in source/transform?