In one of our automated tests we had the following situation: a pipline which fed (H264) buffers into an interpipesink was started. Right after that we started a second pipeline which consumed these buffers via an interpipesrc. When doing this the GStreamer leak tracer reported some leaked tag events. Introducing a short wait between starting the two pipelines would remove the leak.
It seems that the leaked events arrive at the second pipeline while the (second) pipeline is not able to accept them (because it is still starting up?). The intention of this PR is to unref the events in such a situation.
In one of our automated tests we had the following situation: a pipline which fed (H264) buffers into an interpipesink was started. Right after that we started a second pipeline which consumed these buffers via an interpipesrc. When doing this the GStreamer leak tracer reported some leaked tag events. Introducing a short wait between starting the two pipelines would remove the leak.
It seems that the leaked events arrive at the second pipeline while the (second) pipeline is not able to accept them (because it is still starting up?). The intention of this PR is to unref the events in such a situation.