Bug#805414: gdm3: disable pulseaudio to prevent capturing A2DP sink on session start

[Antoine Beaupré]

On 2017-07-02 02:41:26, Michael Biebl wrote:
> Am 02.07.2017 um 02:38 schrieb Michael Biebl:
>> Am 02.07.2017 um 02:30 schrieb Antoine Beaupré:
>>> Or do we treat this as a PA bug ("should release BT")?
>> 
>> I'm not sure, actually.
>
>
> https://wiki.archlinux.org/index.php/Talk:Bluetooth_headset#GDMs_pulseaudio_instance_captures_bluetooth_headset
>
> suggests that a custom PA config shipped by gdm could workaround this by
> unloading the bluetooth module.
>
> That feels a bit like a hack though.
> The real issue is probably that PA doesn't release the BT device.

So I have found this is already reported in the PA bugtracker in #845938
(merged with #848130) - let's treat this separately.

There's a workaround for GDM: disable bluetooth support as documented in
the Arch wiki. We could patch the gdm3 pacakge to do exactly that while
PA fixes its stuff.

The rationale behind disabling BT support there is that there's zero UI
in GDM3 to pair devices. Until those exists, it's meaningless to have BT
support at the login prompt: it will never work, short of the user
switching to a VT and hacking commands by hand, at which point all bets
are off and we can assume the user will figure out a way.

Therefore I would argue we can have a workaround here, in the gdm3
package, for now.

I'll trickle up the PA bugs on the side as well.

Sounds good?

A.

-- 
We must learn to live together as brothers or perish together as fools.
                        - Martin Luther King, Jr.