On Mon, Sep 9, 2024 at 9:00 PM Luiz Augusto von Dentz <luiz.dentz@xxxxxxxxx> wrote: > > Hi Vibhav, > > Don't really like where this is going, if we need bearer specific API > like this then we need a whole new interface, besides you didn't > really explain the reason why you needed this, is that to determine if > there are SDP records to be fetched? ConnectionType is needed only to differentiate between LE and BR/EDR devices. This is because the NT kernel creates device objects for every discovered Bluetooth device, and BR/EDR and LE devices exist under different paths/hierarchies (I suppose the equivalent would be BR/EDR devices existing under /dev/bluetooth/devices/bredr, and LE devices under /dev/bluetooth/devices/le). Essentially, BR/EDR and LE devices can be differentiated in several significant ways, and I could not find a way to reliably do that. The current code checks if AddressType is random, otherwise tries to see if the device object has org.bluez.Gatt* interfaces to determine if the connection bearer is LE. However, I understand that's not reliable, and doesn't determine if the device is connected using dual-mode. That being said, Bearer does sound like a better name for the property, and I can modify the patch to that effect. However, I don't foresee the need for a ConnectBearer property for Wine. Ultimately, my goal is to implement Bluetooth functionality with the minimum possible API surface added to BlueZ, hence the small (ish) patches. -- Vibhav Pant vibhavp@xxxxxxxxx GPG: 7ED1 D48C 513C A024 BE3A 785F E3FB 28CB 6AB5 9598
