PulseAudio implementation

[riagriag]

With the help of cursor I started working on implementing PulseAudio.

So far I managed to port the termux implementation to localdesktop, start a pulseaudio server and play a wav file on startup.
For some reason the audio plays only the first time on a fresh installation.

The number of dependencies needed has vastly increased:

I believe to output sound from the linux distro, die pulseaudio server needs to be started from the app and listening to a tcp stream. The guest system needs to run a pulseaudio client, supplying that tcp stream?

While the prototype works, I still have to go through the code to see how much slop there is.

https://filebin.net/glf2g09lstibudt5

[riagriag]

Making progress:

PXL_20260530_163114125.1.mp4

Pulseaudio server is run by the app and the guest system outputs a stream to the server and as you can hear, android play the sound then.

[Htypyth]

now this is good feature. 👍🏻

[mister-teddy]

@riagriag magician 🧙

[riagriag]

More progress:

PXL_20260531_103155106.TS.1.mp4

Added pulseaudio to default_install() and default_check().
On launch, the audio stream of the whole lxqt session is directed to the pulse server:

fn default_launch() -> String {
format!(
"export PULSE_SERVER={PULSE_GUEST_SERVER}; XDG_RUNTIME_DIR=/tmp Xwayland -hidpi :1 2>&1 & while [ ! -e /tmp/.X11-unix/X1 ]; do sleep 0.1; done; XDG_SESSION_TYPE=x11 DISPLAY=:1 dbus-run-session startlxqt 2>&1"
)
}

So basically it is installed and started and works out of the box, without the user having to do anything.

[riagriag]

Trying to decrease the dependencies, but so for a release built would about double in size to about 6.5 MB.
Basically we only need the tcp stream and could drop all the codec stuff. Which would require a custom termux pulseaudio built. This shouldn't be too much of a hassle and seeing that the last pulseaudio release was in 2024, it might be a one time thing.

[riagriag]

Just going through the AI generated code and preparing for a commit. Having trouble porting it to XFCE, as I've used LXQT previously. Trying to make it agnostic in terms of the DE used. Will leave the dependencies as they are for now.

[riagriag]

Allright, should work for all DEs now, by adding a PulseAudio config file on setup, so each program knows where to direct the audio stream.
I'll make a pull request once I find out how this works (haven't done one yet).

[riagriag]

Pull request is created
#283

Please check and double check as I've never done this before.

[riagriag]

TODO:

• pulseaudio termux build with adjusted DT_NEEDED to drop some *.so files
• maybe add termux packages build system for reproducibility (its quite huge)
• documentation
• feel free to suggest something to be added here

[mister-teddy]

v2.0.0 is building your PulseAudio implementation 🚀

[mgrjanuszchmiel]

Dear elite developers,
Please, is it possible to really use Pulseaudio after starting The installed system? I do not see at all so I would like to install Orca from The Xfce4 terminal. I know, that when Archlinux is being used with Proot, it is always necessary to start orca by typing
dbus-run-session orca &
BUt it have been necessary only when Run application dialog from some actively executed desktop environment have not been used.
So may be, that
pacman -S orca
and pacman -S
espeak-ng
would be enough.
I would only like to know, if Pulseaudio is being initialised automatically, or is it necessary to type some commands.
Thank you very much for your explanation.