[Bug 765249] New: udpsrc: external socket should not be unreffed upon disposal

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=765249

            Bug ID: 765249
           Summary: udpsrc: external socket should not be unreffed upon
                    disposal
    Classification: Platform
           Product: GStreamer
           Version: 1.8.0
                OS: All
            Status: NEW
          Severity: major
          Priority: Normal
         Component: gst-plugins-good
          Assignee: gstreamer-bugs at lists.freedesktop.org
          Reporter: marcin at saepia.net
        QA Contact: gstreamer-bugs at lists.freedesktop.org
     GNOME version: ---

If udpsrc is used with external socket (passed via socket property), this
socket is unreffed upon element disposal, despite ownership being not
transferred.

Sample code:


public static int main(string[] argv) {
    Gst.init(ref argv);

    var socket = new Socket(SocketFamily.IPV4, SocketType.DATAGRAM,
SocketProtocol.UDP);
    socket.bind(new InetSocketAddress(new InetAddress.any(SocketFamily.IPV4),
0), true);

    var pipeline = new Gst.Pipeline("pipeline");

    dynamic Gst.Element source = Gst.ElementFactory.make("udpsrc", "source");
    source.socket = socket;

    dynamic Gst.Element sink = Gst.ElementFactory.make("fakesink", "sink");

    pipeline.add(source);
    pipeline.add(sink);

    source.link(sink);

    pipeline.set_state(Gst.State.PLAYING);
    pipeline.set_state(Gst.State.NULL);
    pipeline.set_state(Gst.State.PLAYING);


    return 0;
}

Compile with valac --pkg gstreamer-1.0 --pkg gio-2.0 udpsrc.vala


Produces:

(udpsrc:42922): GLib-GIO-CRITICAL **: g_inet_socket_address_get_address:
assertion 'G_IS_INET_SOCKET_ADDRESS (address)' failed

(udpsrc:42922): GLib-GIO-CRITICAL **: g_inet_address_get_is_multicast:
assertion 'G_IS_INET_ADDRESS (address)' failed

(udpsrc:42922): GLib-GObject-CRITICAL **: g_object_unref: assertion
'G_IS_OBJECT (object)' failed


If the intention was to transfer ownership of socket into element, it is not
documented behaviour. 

But IMO it should not behave like this - you may want to ensure that socket is
constantly bound across state changes (e.g. so no other process will intercept
it).

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.