[evolution-patches] Free busy refresh

[JP Rosevear <jpr ximian com>]

Fixes a problem Dan found.

-JP
-- 
JP Rosevear <jpr ximian com>
Ximian, Inc.

? leak.patch
? ref-count.patch
? gui/ref-count.patch
Index: ChangeLog
===================================================================
RCS file: /cvs/gnome/evolution/calendar/ChangeLog,v
retrieving revision 1.1764
diff -u -r1.1764 ChangeLog
--- ChangeLog	14 May 2003 18:45:53 -0000	1.1764
+++ ChangeLog	15 May 2003 16:05:07 -0000
@@ -1,3 +1,9 @@
+2003-05-15  JP Rosevear  <jpr ximian com>
+
+	* gui/e-meeting-time-sel.c
+	(e_meeting_time_selector_refresh_free_busy): ref ourselves the
+	number of times we'll get called back
+
 2003-05-14  JP Rosevear  <jpr ximian com>
 
 	* gui/e-tasks.c (e_tasks_destroy): guard against multiple destroys
Index: gui/e-meeting-time-sel.c
===================================================================
RCS file: /cvs/gnome/evolution/calendar/gui/e-meeting-time-sel.c,v
retrieving revision 1.47
diff -u -r1.47 e-meeting-time-sel.c
--- gui/e-meeting-time-sel.c	29 Apr 2003 18:51:50 -0000	1.47
+++ gui/e-meeting-time-sel.c	15 May 2003 16:05:08 -0000
@@ -1252,7 +1252,16 @@
 	/* Ref ourselves in case we are called back after destruction,
 	 * we can do this because we will get a call back even after
 	 * an error */
-	gtk_object_ref (GTK_OBJECT (mts));
+	/* FIXME We should really have a mechanism to unqueue the
+	 * notification */
+	if (all) {
+		int i;
+		
+		for (i = 0; i < e_meeting_model_count_actual_attendees (mts->model); i++)
+			gtk_object_ref (GTK_OBJECT (mts));
+	} else {
+		gtk_object_ref (GTK_OBJECT (mts));
+	}
 	
 	if (all)
 		e_meeting_model_refresh_all_busy_periods (mts->model, &start, &end,