Re: [Rhythmbox-devel] Bluetooth, AVRCP and Rhythmbox

[John Frankish <j-frankish slb com>]

> -----Original Message-----
> From: Bastien Nocera [mailto:hadess hadess net]
> Sent: Tuesday, 18 May, 2010 13:12
> To: John Frankish
> Cc: rhythmbox-devel gnome org
> Subject: Re: [Rhythmbox-devel] Bluetooth, AVRCP and Rhythmbox
> 
> On Tue, 2010-05-18 at 06:35 +0200, John Frankish wrote:
> <snip>
> > > Thanks for the feedback.
> > >
> > > Yes, the uinput module is loaded and I've tried with alsa alone and
> > > with alsa+pulseaudio
> > >
> > > At the first remote AVRCP key press, I see this in dmesg:
> > >
> > > input: 00:1B:66:00:31:88 as /devices/virtual/input/input11
> > >
> > > ..and if I run bluetoothd in the foreground, I see (for example):
> > >
> > > bluetoothd[4253]: AVRCP: uinput initialized for 00:1B:66:00:31:88
> > > bluetoothd[4253]: AVCTP Connected
> > > ...
> > > bluetoothd[4253]: AVRCP: BACKWARD pressed
> > > bluetoothd[4253]: Got 8 bytes of data for AVCTP session 0xb79d86e0
> > > bluetoothd[4253]: AVCTP transaction 8, packet type 0, C/R 0, IPID0,
> > > PID 0x110E
> > > bluetoothd[4253]: AVRCP command 0x0, subunit_type 0x09, subunit_id 0x0,
> > > opcode 0x7C, 2 operands
> > > bluetoothd[4253]: AVRCP: BACKWARD released
> > > <snip>
> > > ...
> > >
> > > But there is no corresponding action in rhythmbox
> > >
> > > John
> > ----------------
> > I understand the communications chain should be:
> >
> > AVRCP signal -> bluetoothd -> uinput -> X -> gnome-settings-daemon ->
> rhythmbox
> 
> Could you test "evtest" on the device created? There should be a device
> created in /dev/input/* for the fake input, check whether events are
> sent there.
I don't see a new /dev/input/* being created, but there is:

/sys/devices/virtual/input/input11/event11

$ cat /proc/bus/input/devices
...
I: Bus=0005 Vendor=0000 Product=0000 Version=0000
N: Name="00:1B:66:00:31:88"
P: Phys=
S: Sysfs=/devices/virtual/input/input11
U: Uniq=
H: Handlers=kbd event11 
B: EV=100007
B: KEY=10300 168 0 0 0 0 0
B: REL=0
> > Is gnome-settings-daemon the only way to do this or is there another
> way?
> 
> It's not the only way to do it, but it's the supported way for GNOME
> applications, and Rhythmbox in particular.
> 
> > The AVRCP signal is getting as far as uinput, but using "dbus-monitor --session", I cannot see any dbus
> > AVRCP events in X and gnome-settings-daemon seems to have a problem (debug output below)
> >
> > How does the AVRCP signal get from uinput to X?
> >
> > John
> > ----------------
> > DEBUG:         Property: 'bluez.name' = 'Sennheiser PXC310BT'
> > DEBUG:         Property: 'device.icon_name' = 'audio-headphones-bluetooth'
> > DEBUG: Starting keyboard manager
> > [1274118894,000,xklavier.c:xkl_engine_start_listen/]    The backend does not require manual layout
> > management - but it is provided by the application
> > DEBUG: Starting a11y_keyboard manager
> > DEBUG: Starting media_keys manager
> > DEBUG: Unable to parse: 'XF86AudioMute'
> > DEBUG: Unable to parse: 'XF86AudioLowerVolume'
> > DEBUG: Unable to parse: 'XF86AudioRaiseVolume'
> > <snip>
> 
> Huh. Did you compile gnome-settings-daemon yourself? That does look
> like
> a problem indeed...
> 
> Cheers

Yes, I compiled it myself, but the dependencies appeared to have all been met and there were no error messages. What would be needed to enable 'XF86AudioMute', etc to be parsed?

John