if (m_rgVideoBufferData)
{
av_freep(m_rgVideoBufferData);
}
is wrong, because it will only free the first pointer of the matrix. It should be av_freep(m_rgVideoBufferData[0])... according to docs. I will test and see if this approach fixes the leak
After a closer investigation, seems to be related to #172 Samples have wrong presentation time stamp. @khouzam maybe we should finally merge the time stamp PRs and start doing some nonregression tests?
The field bellow leaks memory
m_rgVideoBufferData
Also, in destructor
is wrong, because it will only free the first pointer of the matrix. It should be av_freep(m_rgVideoBufferData[0])... according to docs. I will test and see if this approach fixes the leak