bluealsa-aplay not releasing hw audio card unless bluealsa-aplay killed about bluez-alsa HOT 6 CLOSED

I've fixed nasty bug which has caused infinite read loop (100% CPU). And right now the bluealsa-aplay will terminate when the phone has been disconnected.

HOWEVER, this fix does not resolve your problem completely. Right now, I'm not able to detect whether the phone has been disconnected unless streaming has occurred.

connect phone -> start bluealsa-aplay -> disconnect phone (without playing audio) -> bluealsa-aplay is still running

This disconnection "feature" will be possible in the future, alongside with the "real bluetooth speaker" functionality (accepting any MAC address).

from bluez-alsa.

Looking forward to this "This disconnection "feature" will be possible in the future, alongside with the "real bluetooth speaker" functionality (accepting any MAC address)."
Is the other fix in the git now?
Been monitoring the DBus signals but they only give me "connected, active and idle", not much use for using them as triggers for other things. Strangely last night I could see a org.bluez.MediaPlayer1 bus, its signals gave me "playing, paused", everything. They changed immediately with the music off the iphone as well. Very useful, but that DBus interface has disappeared ( bluealsa-aplay still plays O.K) so not sure what happened.

from bluez-alsa.

Been monitoring the DBus signals but they only give me "connected, active and idle", not much use for using them as triggers for other things

Use PropertiesChanged signal on org.bluez.Device1 and look for the Connected property. On disconnect clean up, stop stream etc.

trangely last night I could see a org.bluez.MediaPlayer1 bus, its signals gave me "playing, paused", everything.

The org.bluez.MediaPlayer1 interface is created on successful device connection when all the prerequisites like bluealsa and device support (some phone don't support it) are in place.

edit: very simple python script below to handle connect and disconnect.

from gi.repository import Gio, GLib


def on_properties_changed(proxy, changed, invalidated):
    changed_properties = changed.unpack()
    if "Connected" in changed_properties:
        if changed_properties["Connected"]:
            print("Configure and start my program")
        else:
            print("Kill my program and cleanup")


proxy = Gio.DBusProxy.new_for_bus_sync(
    Gio.BusType.SYSTEM,
    Gio.DBusProxyFlags.NONE,
    None,
    "org.bluez",
    "/org/bluez/hci0/dev_CC_61_E5_1A_39_B5",
    "org.bluez.Device1",
    None)


proxy.connect("g-properties-changed", on_properties_changed)

loop = GLib.MainLoop()
loop.run()

from bluez-alsa.

OK, based on Arkq's comment "This disconnection "feature" will be possible in the future, alongside with the "real bluetooth speaker" functionality (accepting any MAC address).", ultimately the "real bluetooth speaker" is actually what I'm after and trying to do. Based on that I'll wait this #issue is closed for me.
Thanks for your help.
Cheers.

from bluez-alsa.

@George-ou812 Bare in mind, that the "future" means few months (not days). This project is not a "Bluetooth speaker", but more likely pulseaudio alternative for Bluetooth Audio on Linux. Hence, this feature has very low priority for me. After all, it is possible to accomplish it already with a little bit of scripting (e.g. Python code provide by @infirit).

from bluez-alsa.

@arkq, appreciate your timescale's, "pulseaudio alternative for Bluetooth Audio on Linux" is perfect for me, pa is taking up far too much resources on my board just for BT. Look forward to see this sometime, keep up the good work!

from bluez-alsa.