sfos: Workaround for the no audio in call issue.

See merge request !36

Needs proper fix later.
Contributes to JB#49447

Fix aarch64 builds by using systemd macros. Contributes JB#49681

See merge request !35

Signed-off-by: David Greaves <david.greaves@jolla.com>

Signed-off-by: David Greaves <david.greaves@jolla.com>

Signed-off-by: David Greaves <david.greaves@jolla.com>

Signed-off-by: David Greaves <david.greaves@jolla.com>

Avoid having system and user session daemon running simultaneously.

See merge request !34

Wait until system daemon pid file disappears before starting user
session PulseAudio.

See JB#49307

Don't use %config(noreplace).

See merge request !31

[packaging] Update config as per policy. Contributes JB#44152
Signed-off-by: David Greaves <david.greaves@jolla.com>

Have static uid and gid for pulse and use only audio as additional
group for user pulse. Fixes JB#46368

sfos: packaging: Fix dependencies of system-wide. JB#45842

See merge request !29

Add After=dbus.socket and remove DefaultDependencies=no.

System-wide mode Pulseaudio

See merge request !28

Add configuration for system-wide mode. These new unit is not meant to
be started by user but instead required when making emergency calls
outside user session. PartOf ensures that the unit is stopped when
emergency-call.target unit is stopped so that it doesn't collide with
user session pulseaudio.

There are no conflicts to multi-user.target or
systemd-user-sessions.service because otherwise they may not be started
at all if this unit happens to be still running.

Unifying summaries and use _smp_mflags instead of old jobs.

sfos: packaging: Do not use gdbm even if available ; JB#44440

See merge request !27

Add `--with-database=simple` to avoid using gdbm even if it were
available.  This is due to gdbm getting updated to a GPLv3 version,
so we must make sure it doesn't get linked in things that go in
the default images.
Signed-off-by: Aapi Hämäläinen <aapi.hamalainen@jolla.com>

Update to PulseAudio 12.2.

See merge request !24

No need to act too much on oFono state changes, as
if the daemon disappears it's either going to appear
again shortly, or we are rebooting, or something similar.

Only does anything when headset_head_unit profile
is active.

If the initial Connect call fails it is possible that the
headset in question is still starting up and doesn't want
connection from our end just yet. In that case say that we
are connecting and wait for possible NewConnection from
the starting device. In case there isn't going to be any
NewConnection set up a timer to update the backend state.

HFP/HSP headsets may initiate transport when they are connected
but the transport is down. When this happens switch to the
profile associated with the headset.

As we are not running the sink or source threads transports
wouldn't get switched to IDLE unless we do so here.

When running in droid setup we don't create sinks or sources,
as the reading and writing to headset is done through
droid.sink and droid.source with sco ports enabled.