# HG changeset patch
# User Mark Doliner <mark@kingant.net>
# Date 1048280371 0
# Node ID adf9ca39578f5f7a3f6631e9feb2296a84844441
# Parent  e6b5cb0f2bd20ea2978033d8ed9f486ae2f39511
[gaim-migrate @ 5174]
Christian, I got the following backtrace after opening and closing some
convos:
#4  <signal handler called>
#5  0x08066e47 in gaim_window_get_active_conversation (win=0x82b44d8)
    at conversation.c:775
#6  0x08085599 in update_send_as_selection (win=0x82b44d8) at gtkconv.c:1964
#7  0x4067fa4c in g_timeout_dispatch () from /usr/lib/libglib-2.0.so.0

So although I couldn't reproduce it, I guess it's still possible for
Gaim to crash there.  This fixes that by making sure that win still
exists before it tries to update the send as menu for it.  The other
way to fix this would be to keep track of the g_timer thingy and then
remove it when destroying win.  I don't know which was is better.
Probably the other way.

committer: Tailor Script <tailor@pidgin.im>

diff -r e6b5cb0f2bd2 -r adf9ca39578f src/gtkconv.c
--- a/src/gtkconv.c	Fri Mar 21 20:33:03 2003 +0000
+++ b/src/gtkconv.c	Fri Mar 21 20:59:31 2003 +0000
@@ -1961,6 +1961,9 @@
 	GtkWidget *menu;
 	GList *child;
 
+	if (g_list_find(gaim_get_windows(), win) == NULL)
+		return FALSE;
+
 	conv = gaim_window_get_active_conversation(win);