Mercurial > pidgin3 / changeset

• changeset
• raw

changeset

This should fix another crash I've seen. We weren't correctly

Tue, 18 Nov 2008 07:32:23 +0000

author

Mark Doliner <markdoliner@pidgin.im>

date

Tue, 18 Nov 2008 07:32:23 +0000

changeset 24708

a5f6cef9797e

parent 24707

2a6fd0a61c41

child 24709

4e9caef6c0fc

child 24710

08d2dde8722f

child 25934

7f05d5b0c921

child 25954

3bfb792dc380

child 26007

71077851aec8

This should fix another crash I've seen. We weren't correctly
cancelling the proxy connect when we destroyed the servconn
because we were only calling msn_servconn_disconnect() if we
had actually connected.

So always call msn_servconn_disconnect() from msn_servconn_destroy()
and let the disconnect function deal with only closing stuff that
is actually open.

Here's the backtrace:
#0 0x0000000000000000 in ?? ()
#1 0x00000000005f95e0 in connect_cb (data=0x2ab92ef6e7b0, source=7, error_message=0x0)
at servconn.c:185
#2 0x000000000056e7da in purple_proxy_connect_data_connected (connect_data=0x2ab92ef70260)
at proxy.c:598
#3 0x000000000056e90d in socket_ready_cb (data=0x2ab92ef70260, source=7, cond=PURPLE_INPUT_WRITE)
at proxy.c:657
#4 0x00000000004db70c in pidgin_io_invoke (source=0x2ab92ee4be50, condition=G_IO_OUT,
data=0x2ab92dfd3130) at gtkeventloop.cc:79

In frame 1, servconn is not valid (it has already been freed)

--- a/libpurple/protocols/msn/servconn.c	Tue Nov 18 07:16:49 2008 +0000
+++ b/libpurple/protocols/msn/servconn.c	Tue Nov 18 07:32:23 2008 +0000
@@ -69,8 +69,7 @@
 		return;
 	}
 
-	if (servconn->connected)
-		msn_servconn_disconnect(servconn);
+	msn_servconn_disconnect(servconn);
 
 	if (servconn->destroy_cb)
 		servconn->destroy_cb(servconn);