Re: [Tracker] REVIEW: external-crawler branch, part1

[Martyn Russell <martyn lanedo com>]

On 08/08/14 08:58, Philip Van Hoof wrote:
> In finalize:
>
> +       if (priv->roots_to_notify) {
> +               g_hash_table_unref (priv->roots_to_notify);
> +       }
>
> vs.
>
> notify_roots_finished:
>
> +       if (check_queues && g_hash_table_size (fs->priv->roots_to_notify) < 2) {
>
>
> So if you check priv->roots_to_notify for not being NULL in finalize,
> then isn't it possible that in notify_roots_finished the same member
> can also be NULL? Or is notify_roots_finished always disconnected in a
> timely matter?

It's highly unlikely. All functions that call notify_roots_finished() 
are called BEFORE finalize() which cleans up the hash table.

I've added some NULL checks anyway.

--
Regards,
Martyn

Founder & Director @ Lanedo GmbH.
http://www.linkedin.com/in/martynrussell