hal and bluetooth

[David Zeuthen]

On Wed, 2005-01-19 at 21:52 +0200, Paul Ionescu wrote:
> > You need complex software and UI to handle all the discovery and
> > authentication anyway - just like networking (e.g. SMB, IMAP etc.).
> 
> We should not mess with the complex software and UI to handle discovery
> and authentication, but only to signal the presence and the attributes of
> an already connected wireless device (paired and authenticated and active).
> I see your point and I don't want either that HAL to do discovery and
> authentication and other stuff like this, because this is not his job, but
> merely to see that a new device was added in /sysfs and report it to dbus,
> to whom it may be interested.
> 

I suppose we *could* do this; I don't know much about Bluetooth but I
guess what we would do would be to add the following device objects

 - Bluetooth Adapter (for instance a USB device)
  - Bluetooth device (mouse)
   - Input device    (logical device - input.physdev
                      pointing to the parent)
  - Bluetooth device (phone/pda)
   - Serial port device  (logical device - serial.physdev 
                          pointing to the parent)

The benefit here is that a program only needs to speak with HAL
in order to discover e.g. input devices or serial devices. 

Now, I'm not sure is this is a good idea, it seems like a slippery
slope, in particular I don't know how well this plays with existing
Bluetooth userspace software (I'm thinking all the hcid, GNOME
Bluetooth, KDE Bluetooth etc.) - do you know?

Whether we should export what Service Profiles each Bluetooth
device supports is also another question.

David


_______________________________________________
hal mailing list
hal at lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/hal