Revert upstream patch that seems to be causing firefox and thunderbird to

crash, while mozilla dev's are looking where the actual bug is.

Submitted by:	cperciva@

x11-toolkits/gtk20/Makefile

@@ -5,7 +5,7 @@
 
 PORTNAME=	gtk
 PORTVERSION=	2.24.19
-PORTREVISION?=	0
+PORTREVISION?=	1
 CATEGORIES=	x11-toolkits
 MASTER_SITES=	GNOME
 MASTER_SITE_SUBDIR=	sources/gtk+/${PORTVERSION:C/^([0-9]+\.[0-9]+).*/\1/}

x11-toolkits/gtk20/files/patch-gdk_gdkwindow.c

@@ -0,0 +1,37 @@
+Revert the following Gtk+ commit [1], this commit is seems to cause
+firefox/thunderbird to crash [2]. Reverting this patch will make ff/tb stop 
+crashing while upstream is looking at the exact cause.
+
+[1] https://git.gnome.org/browse/gtk+/commit/?h=gtk-2-24&id=692a0e5906c5da7f85c16c9d6cbb0d3ed8b4a576
+[2] https://bugzilla.mozilla.org/show_bug.cgi?id=887587
+
+All information:
+http://lists.freebsd.org/pipermail/freebsd-gnome/2013-July/028852.html
+
+--- gdk/gdkwindow.c.orig	2013-07-25 11:01:16.000000000 +0200
++++ gdk/gdkwindow.c	2013-07-25 11:02:24.000000000 +0200
+@@ -5463,7 +5463,6 @@
+   GdkWindowImplIface *impl_iface;
+   gboolean save_region = FALSE;
+   GdkRectangle clip_box;
+-  int iteration;
+ 
+   /* Ensure the window lives while updating it */
+   g_object_ref (window);
+@@ -5471,15 +5470,8 @@
+   /* If an update got queued during update processing, we can get a
+    * window in the update queue that has an empty update_area.
+    * just ignore it.
+-   *
+-   * We run this multiple times if needed because on win32 the
+-   * first run can cause new (synchronous) updates from
+-   * gdk_window_flush_outstanding_moves(). However, we
+-   * limit it to two iterations to avoid any potential loops.
+    */
+-  iteration = 0;
+-  while (private->update_area &&
+-	 iteration++ < 2)
++  if (private->update_area)
+     {
+       GdkRegion *update_area = private->update_area;
+       private->update_area = NULL;