On Sun, Apr 29, 2012 at 8:22 AM, Luiz Augusto von Dentz
<luiz.dentz@xxxxxxxxx> wrote:
> Hi Lucas,
>
> On Sat, Apr 28, 2012 at 12:25 AM, Lucas De Marchi
> <lucas.demarchi@xxxxxxxxxxxxxx> wrote:
>> + if (p->exists != NULL && !p->exists(p, iface->user_data))
>> + continue;
>> +
>> + dbus_message_iter_open_container(&dict, DBUS_TYPE_DICT_ENTRY,
>> + NULL, &entry);
>> + dbus_message_iter_append_basic(&entry, DBUS_TYPE_STRING,
>> + p->name);
>> + dbus_message_iter_open_container(&entry, DBUS_TYPE_VARIANT,
>> + p->type, &value);
>> +
>> + if (!p->get(connection, message, p, &value, iface->user_data)) {
>> + dbus_message_iter_abandon_container(&entry, &value);
>> + dbus_message_iter_abandon_container(&dict, &entry);
>> + dbus_message_unref(reply);
>> + return NULL;
>> + }
>
> I would do this a little different, make the callback return the value
> (void *) so you don't even need to create any container if it returns
> NULL, if the value is set then you create the container and use the
> registered type to add the value to the container, obviously this is
> limited to values that doesn't require another container but IMO that
> should not be the case of any of our properties.
Yeah, I thought about that. But as you said it's limited to the basic
types. See the example Henrique sent for "ao". And there's the more
difficult one: a property that is a dictionary.
To cover these more complicated cases, we would need to check if it's
a property with basic signature and act differently based on that.
Lucas De Marchi
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
