BlueZ dbus binding is device dependent

[Iain Hibbert]

On Sat, 11 Apr 2009, Bruce Simpson wrote:

> So, just to confirm, I read the code examples in this BlueZ Wiki article;
> and it does indeed appear that the naming convention BlueZ uses in DBus is
> BlueZ dependent:-
>    http://wiki.bluez.org/wiki/HOWTO/DiscoveringDevices

first, that API may not be the same as the BlueZ 4.x series which I think
has been evolving. The wiki main page says the API is best defined by the
documents in the bluez releases.

second, I'm not sure if it is a total disaster. The device being
referenced is at least a string "/org/bluez/hci0" rather than the integer
dev_id that the bluez library uses. That can be worked on fairly easily,
plus I might have seen a comment on the bluez list (which I subscribed to
but don't really read) that hardcoded defs are to be frowned on, I think
there is a way to find local devices.

Then, the module name "org.bluez" is probably not needed to change - for
instance if a GNOME application uses the bluez dbus module, I think that
all calls end up passed to bluetoothd and 'all that is required' is having
a bluez compatible bluetoothd. Whether the bluez bluetoothd can be easily
patched or if a fork or a rewrite is required I don't know. I looked at
the bluez sources but I don't have any context for it so am just
confused..

iain