On 04/19/2013 12:42 PM, Patrik Flykt wrote:
On Fri, 2013-04-19 at 10:09 +0200, Daniel Wagner wrote:
The 'policy->session != NULL' is basically testing if the refcount is
1
or 2. Okay, I agree that is rather stupid.
Just unref() as the last step, no data can be trusted after and unref.
What about
static void remove_policy(struct policy_data *policy)
{
connman_session_set_default_config(policy->config);
update_session(policy);
g_hash_table_remove(file_hash, policy->filename);
policy_unref(policy);
}
? If there is no session attached to the policy we will do some
unnecessary operation in connman_session_set_default_config(). But it
improves the readability of the code.
cheer,
daniel
_______________________________________________
connman mailing list
connman@connman.net
http://lists.connman.net/listinfo/connman