No rtsp teardown with fakesink or appsink.

[Tim-Philipp Müller]

On Tue, 2015-09-15 at 17:15 +0200, Thibault Jochem wrote:

Hi,

> I tried that but the EOS message seems to never arrive.
> I will use the --gst-debug-level=4 for now and will check if it's
> fixed on newer versions.

Might be related to this bug:
https://bugzilla.gnome.org/show_bug.cgi?id=706387
> 
> cheers,
> 
> On Tue, Sep 15, 2015 at 3:20 PM, Tim Müller <tim at centricular.com>
> wrote:
>         On Tue, 2015-09-15 at 13:27 +0200, Thibault Jochem wrote:
>         
>         Hi Thibault,
>         
>         > I'm having an issue where I can't have gstreamer to send
>         rtsp teardown
>         > when I'm using an appsink to grab pictures from the stream.
>         >
>         >
>         > I wrote a small test program, trying several sinks, and
>         everythings
>         > seems Ok using d3dvideosink, but as soon as I plug a
>         fakesink or an
>         > appsink, the rtsp socket seems to be closed before the
>         session is
>         > cleanly closed. This is how I close the stream.
>         >
>         >
>         >     gst_element_send_event(pipeline, gst_event_new_eos());
>         >     gst_element_set_state (pipeline, GST_STATE_NULL);
>         >     gst_object_unref (pipeline);
>         >     gst_object_unref (bus);
>         >
>         >
>         > Setting the gst-debug-level to 4 seems to work around the
>         problem so
>         > it make me think of a race condition.
>         >
>         >
>         > Is it a bug, or am I missing something to cleanly close the
>         pipeline ?
>         
>         I don't know if it helps, but sending the EOS event and then
>         immediately
>         setting state to NULL may not be what you want. You'd have to
>         wait for
>         the EOS event to make it through the pipeline and arrive at
>         all the
>         sinks, which will happen asynchronously. So perhaps you should
>         wait for
>         an EOS message on the pipeline bus before shutting down the
>         pipeline.
>         (Note that there may be a bug in rtpbin or rtpjitterbuffer
>         that it
>         doesn't forward the eos event properly, I'm not sure if that
>         has been
>         fixed or not right now).
>         
>         Cheers
>          -Tim
>         
>         --
>         Tim Müller, Centricular Ltd - http://www.centricular.com
>         
>         Join us at the GStreamer Conference: 8-9 October 2015 in
>         Dublin, Ireland
>         
>         
>         _______________________________________________
>         gstreamer-devel mailing list
>         gstreamer-devel at lists.freedesktop.org
>         http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel
> 
> 
> 
> 
> -- 
> Thibault Jochem
> www.laboiteaselfie.com
> _______________________________________________
> gstreamer-devel mailing list
> gstreamer-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel