[Merge] lp:~phablet-team/media-hub/gst-1.5 into lp:media-hub

[Jim Hodapp]

Review: Approve code



Diff comments:

> 
> === modified file 'src/core/media/gstreamer/playbin.cpp'
> --- src/core/media/gstreamer/playbin.cpp	2015-06-01 21:34:46 +0000
> +++ src/core/media/gstreamer/playbin.cpp	2015-07-21 16:26:24 +0000
> @@ -28,15 +28,22 @@
>  
>  namespace
>  {
> -void setup_video_sink_for_buffer_streaming(GstElement* video_sink)
> +void setup_video_sink_for_buffer_streaming(GstElement* pipeline)
>  {
>      // Get the service-side BufferQueue (IGraphicBufferProducer) and associate it with
>      // the SurfaceTextureClientHybris instance
>      IGBPWrapperHybris igbp = decoding_service_get_igraphicbufferproducer();
>      SurfaceTextureClientHybris stc = surface_texture_client_create_by_igbp(igbp);
> +
>      // Because mirsink is being loaded, we are definitely doing * hardware rendering.
> -    surface_texture_client_set_hardware_rendering (stc, TRUE);
> -    g_object_set (G_OBJECT (video_sink), "surface", static_cast<gpointer>(stc), static_cast<char*>(NULL));
> +    surface_texture_client_set_hardware_rendering(stc, TRUE);
> +
> +    GstContext *context = gst_context_new("gst.mir.MirContext", TRUE);

Ok yeah, I was wondering if that might be a limitation. It's not worth the effort currently. It would be something to do when we go in and change all of the surface_texture_client and stc references in this code and the hybris code.

> +    GstStructure *structure = gst_context_writable_structure(context);
> +    gst_structure_set(structure, "gst_mir_context", G_TYPE_POINTER, stc, NULL);
> +
> +    /* Propagate context in pipeline (needed by amchybris and mirsink) */
> +    gst_element_set_context(pipeline, context);
>  }
>  }
>  #else  // MEDIA_HUB_HAVE_HYBRIS_MEDIA_COMPAT_LAYER


-- 
https://code.launchpad.net/~phablet-team/media-hub/gst-1.5/+merge/265413
Your team Ubuntu Phablet Team is subscribed to branch lp:~phablet-team/media-hub/fix-black-video-bugs.