Re: [Bug-gnubg] bug: use of uninitialized value at set.c:CommandSetFullScreen

[Daisuke Takahashi]

Hi,

I confirmed the previous patch does not affect the fullscreen issue, but this 
one will fix it (at least on my OSX). It moves gtk_window_unfullscreen in order 
to make all widgets are showed before switching from fullscreen mode.
I wonder why fFullScreen is adjusted before ShowAllPanels(). Please check it if 
I do not misunderstand your code.
Thank you very much.


Regards,
Daisuke Takahashi

--- gtkgame.c   29 Jun 2014 14:53:45 -0000      1.886
+++ gtkgame.c   14 Jul 2014 13:32:51 -0000
@@ -1835,15 +1835,13 @@
         if (g_signal_handler_is_connected(G_OBJECT(ptl), id))
             g_signal_handler_disconnect(G_OBJECT(ptl), id);
 
-        gtk_window_unfullscreen(ptl);
-        gtk_window_set_decorated(ptl, TRUE);
-
         if (showingPanels) {
-            fFullScreen = TRUE; /* Avoid panel sizing code */
             ShowAllPanels(NULL, 0, NULL);
-            fFullScreen = FALSE;
         }
 
+        gtk_window_unfullscreen(ptl);
+        gtk_window_set_decorated(ptl, TRUE);
+        
         if (changedRP) {
             gtk_widget_set_sensitive(pmiRP, TRUE);
             changedRP = FALSE;



On 2014/07/14, at 9:14, Michael Petch wrote:

> 
> 
> On 2014-07-13 5:38 PM, Daisuke Takahashi wrote:
>> In addition to the OSX issue
> 
> So If I understand correctly you think this fix resolves the initial
> display problem were the main screen doesn't render on OS/X properly
> (the entire window appear white) include the analysis pane etc.
> 
> 
> -- 
> Michael Petch
> GNU Backgammon Maintainer / Developer
> OpenPGP FingerPrint=D81C 6A0D 987E 7DA5 3219 6715 466A 2ACE 5CAE 3304