Gtk Plug bug ...

[Michael Meeks <michael helixcode com>]

Hi,

	I managed to repeat this bug, and Federico and myself generated 
a patch which involves re-ordering the plug unrealize do this:

{
	GtkPlug *plug;

	g_return_if_fail (widget != NULL);
	g_return_if_fail (GTK_IS_PLUG (widget));

	plug = GTK_PLUG (widget);

	if (GTK_WIDGET_CLASS (parent_class)->unrealize)
		(* GTK_WIDGET_CLASS (parent_class)->unrealize) (widget);

	if (plug->socket_window != NULL) {
		gdk_window_set_user_data (plug->socket_window, NULL);
		gdk_window_unref (plug->socket_window);
		plug->socket_window = NULL;
	}
}

	Since elsewhere we do:

	widget->window = gdk_window_new (plug->socket_window, 
					 &attributes, attributes_mask);

	And the parent_class unrealize destroys widget->window.

	Anyhow; the purify log looks like this:

   Started   sample-control-factory (   2 errors, 0 leaked bytes)
      Purify instrumented sample-control-factory (pid 5687 at Thu Aug 10
17:39:35 2000)
      FMR: Free memory read
      This is occurring while in:
            gdk_window_internal_destroy [gdkwindow.c:591]
                         if (private->parent)
                           {
                             GdkWindowPrivate *parent_private =
(GdkWindowPrivate *)private->parent;
            =>               if (parent_private->children)
                               parent_private->children = g_list_remove
(parent_private->children, window);
                           }
                         
            gdk_window_destroy [gdkwindow.c:685]
            gtk_widget_real_unrealize [gtkwidget.c:4550]
            bonobo_plug_unrealize [bonobo-plug.c:78]
            gtk_marshal_NONE__NONE [gtkmarshal.c:344]
            gtk_signal_real_emit [gtksignal.c:1440]
            gtk_signal_emit [gtksignal.c:552]
            gtk_widget_unrealize [gtkwidget.c:1700]
            gtk_widget_shutdown [gtkwidget.c:4384]
            gtk_window_shutdown [gtkwindow.c:656]
      Reading 4 bytes from 0x1198b8 in the heap.
      Address 0x1198b8 is 48 bytes into a freed  block at 0x119888 of 52
bytes.
      This block was allocated from:
            malloc         [rtlib.o]
            g_malloc       [gmem.c:177]
            gdk_window_foreign_new [gdkwindow.c:519]
            bonobo_plug_construct [bonobo-plug.c:45]
            bonobo_plug_new [bonobo-plug.c:56]
            impl_Bonobo_Control_set_window [bonobo-control.c:283]
            _ORBIT_skel_Bonobo_Control_set_window [Bonobo-skels.c:3758]
            ORBit_POA_handle_request [orbit_poa.c:507]
            ORBit_handle_incoming_request [server.c:89]
            ORBit_handle_incoming_message [server.c:159]
      There have been 0 frees since this block was freed from:
            free           [rtlib.o]
            g_free         [gmem.c:411]
            gdk_window_unref [gdkwindow.c:740]
            bonobo_plug_unrealize [bonobo-plug.c:73]
                       if (plug->socket_window != NULL)
                       {
                               gdk_window_set_user_data
(plug->socket_window, NULL);
            =>                 gdk_window_unref (plug->socket_window);
                               plug->socket_window = NULL;
                       }
               
            gtk_marshal_NONE__NONE [gtkmarshal.c:344]
            gtk_signal_real_emit [gtksignal.c:1440]
            gtk_signal_emit [gtksignal.c:552]
            gtk_widget_unrealize [gtkwidget.c:1700]
            gtk_widget_shutdown [gtkwidget.c:4384]
            gtk_window_shutdown [gtkwindow.c:656]
      FMW: Free memory write

	HTH,

		Michael.

-- 
 mmeeks@gnu.org  <><, Pseudo Engineer, itinerant idiot