Hi Travis,
On Thu, Jan 12, 2017 at 12:38 AM, Travis Griggs <travisgriggs@xxxxxxxxx> wrote:
> I’m trying to bring my linux SBC acting as a peripheral up to debian stretch. I had been running on jessie, with my own build of bluez 5.41. I had been modifying the bluetooth.service to execute:
>
> /usr/lib/bluetooth/bluetoothd -d -E --noplugin=*
>
> The latest stretch/testing bluez seems to be 5.43 though:
>
> dpkg -l | grep bluez
> ii bluez 5.43-1 armhf Bluetooth tools and daemons
>
> I thought the -E flag was supposed to no longer be necessary after 5.42. So I removed it. I updated my own python code to match the small changes I found in example-advertisement. But I’m getting the following error:
>
> ERROR:dbus.connection:Unable to set arguments (dbus.ObjectPath('/nelson_irrigation/twig_pilot'), {}) according to signature None: <class 'ValueError'>: Unable to guess signature from an empty dict
Well RegisterAdvertisement is still using
GDBUS_EXPERIMENTAL_ASYNC_METHOD, what has been marked as stable is the
GATT APIs so for advertisement you still need -E.
> Which after googling, was the same error I found a while back when I had forgotten to turn -E on with 5.40.
>
> So I added the -E back in and it started working as hoped.
>
> Can someone explain why the -E is still needed with 5.43, for dbus code like that found in the test/examples? Did it not really get promoted after all? Or did I misunderstand the semantics of what was changing?--
> 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
--
Luiz Augusto von Dentz
--
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
