Worth a go here as my trusy old Vaio netbook has **** sound (and I don't want pulseaudio).

I got bluealsa working but it was not so simple. For a start packages bluez-alsa-utils libasound2-plugin-bluez (you need both) have only a systemd service file, no sysv script. I needed bluetooth daemon running first and a custom ~/.asoundrc

pcm.SoundCoreMini {
        type bluealsa
        device "XX:XX:XX:XX:XX:XX"
        profile "a2dp"
        hint { show on description "Sound Core Mini"}
}

Having already installed blueman I could use those tools to get the device ID. The device can then be manipulated using blueman-manager.

I repackaged it with a custom init script but havent got that quite right as yet. It won't detach from the shell and hangs on bootup. But I can start it manually after disabling the init daemon.

Running chimaera you *might* get away with just a few lib deps from ceres. The original packages will break beowulf. My crude "backport" rebuild from source can run on beowulf.

Here's the offending custom init script, if anyone wants to debug it. More later (if) I get time to revisit this. Or maybe it's better not to have this start automatically..

#!/bin/sh
### BEGIN INIT INFO
# Provides: bluez-alsa

# Default-Start:     2 3 4 5
# Default-Stop:      0 1 6

# Required-Start:	$syslog $local_fs $remote_fs bluetooth 
# Required-Stop:	$syslog $local_fs $remote_fs 
# Short-Description: Bluealsa daemon
### END INIT INFO

. /lib/lsb/init-functions
prog=bluez-alsa
if test -f -/etc/default/bluez-alsa; then
	. -/etc/default/bluez-alsa 
fi

PIDFILE=/var/run/$prog.pid
DESC="Bluealsa daemon"
start() {
	log_daemon_msg "Starting $DESC" "$prog"
	start_daemon -p $PIDFILE /usr/bin/bluealsa $OPTIONS
	if [ $? -ne 0 ]; then
		log_end_msg 1
		exit 1
	fi
	if [ $? -eq 0 ]; then
		log_end_msg 0
	fi
	exit 0
}

stop() {
	log_daemon_msg "Stopping $DESC" "$prog"
	killproc -p $PIDFILE /usr/bin/bluealsa
	if [ $? -ne 0 ]; then
		log_end_msg 1
		exit 1
	fi
	if [ $? -eq 0 ]; then
		log_end_msg 0
	fi
}

force_reload() {
	stop
	start

}

case "$1" in
	start)
		start
		;;
	stop)
		stop
		;;
	force-reload)
		force_reload
		;;
	restart)
		stop
		start
		;;

	*)
		echo "$Usage: $prog {start|stop|force-reload|restart}"
		exit 2
esac

My installed bluetooth-related packages:

root@vaio:~# dpkg -l|grep blue
ii  blueman                                  2.0.8-1+deb10u1                     amd64        Graphical bluetooth manager
ii  bluez                                    5.50-1.2~deb10u2                    amd64        Bluetooth tools and daemons
ii  bluez-alsa-utils                         3.0.0-2                             amd64        Bluetooth Audio ALSA Backend (utils)
ii  bluez-firmware                           1.2-4                               all          Firmware for Bluetooth devices
ii  bluez-obexd                              5.50-1.2~deb10u2                    amd64        bluez obex daemon
ii  libasound2-plugin-bluez:amd64            3.0.0-2                             amd64        Bluetooth Audio ALSA Backend (plugins)
ii  libbluetooth-dev:amd64                   5.50-1.2~deb10u2                    amd64        Development files for using the BlueZ Linux Bluetooth library
ii  libbluetooth3:amd64                      5.50-1.2~deb10u2                    amd64        Library to use the BlueZ Linux Bluetooth stack
ii  libpam-blue                              0.9.0-3                             amd64        PAM module for local authenticaction with bluetooth devices

EDIT: example use:

:~$ mpv --audio-device=alsa/SoundCoreMini 'Sweet Georgia Brown 1.m4a'

Or in VLC, pick  it from the audio device GUI.

Cheers all (and thanks for chimaera), D

I guess some have their use-case reasons to automount removables. I don't know why.

Mounting an mtp device should be as simple as:

jmtpfs /path/to/your/mountpoint/
fusermount -u /path/to/your/mountpoint/

Or you could script that with some checks and a fancy yad dialog.

(OT) thanks xinomilo for mentioning  "lineageOS" (android AOSP based without bloat/spyware), probably as clean as you will get on such devices. The only network connection here at the moment is 4g. I did quite a bit of research on android (lack of) security and run lineage (without gapps).

Recent Youtube changes broke gtk-youtube-viewer (maybe other clients). It was forked to "gtk straw viewer" and works again. Well here and for now at least.. https://github.com/trizen/straw-viewer

Using a youtube client avoids heavyweight browser loading, ads, tracking... This one is quite light on deps and system resources but has a very good descriptive gui. No qt deps. Probably needs latest youtube-dl to run properly https://github.com/ytdl-org/youtube-dl

I posted (probably the only) deb package: http://exegnulinux.net/apt/pool/main/s/straw-viewer/

Sources, dsc et al are at the same url if anyone wants to inspect, compile or modify yourself. If interested, please post back if you see anything wrong with the package.

EDIT: youtube-dl (and anything else that uses it) will puke unless you update it quite often.

Maybe or not your problem is related but:

My ceres live builds, with *default* 5x kernels (all of them), hang every time in seemingly random stages of bootup. Often while still in initramfs. This older HP (Intel) laptop is all I have around at the moment. They boot fine with some combinations of "failsafe graphics" parameters on the cmdline (maybe breaks some acpi-related functions):

noapic noapm nodma nomce nolapic nosmp

Seems a major kernel issue. A search for, e.g. "linux 5x intel hangs", returns much info but little good news, at least not soon. No such issues here with Beowulf and Ascii systems.

:~$ uname -r && inxi -CG
5.4.0-4-amd64
CPU:
  Topology: Single Core model: Intel Core2 T5500 bits: 64 type: MCP
  L2 cache: 2048 KiB
  Speed: 1662 MHz min/max: N/A Core speed (MHz): 1: 1662
Graphics:
  Device-1: Intel Mobile 945GM/GMS 943/940GML Express Integrated Graphics
  driver: i915 v: kernel
  Display: x11 server: X.Org 1.20.7 driver: intel
  unloaded: fbdev,modesetting,vesa resolution: 1024x768~60Hz
  Message: Unable to show advanced data. Required tool glxinfo missing.

My usual method is an out-of-the-box live image (Refracta is the prime candidate but there are others) with LUKS encrypted full persistence. The live image is tailored to suit as many machines as possible. The persistence is a custom overlayfs, can be a (decent size) file or a partition. On boot it will prompt for the LUKS key, if you don't have that it will continue to boot to only the basic, default live image. The security of the rest is down to the strength of your encryption key.

It's quite surprising how quickly and efficiently such a system can run..

And is already packaged, working, supported and documented by TDE for all Devuan systems including Beowulf and Ceres https://wiki.trinitydesktop.org/DevuanInstall

ExeGNU (which tracks current Devuan) includes a custom selected "minimal" TDE (and openbox) plus some custom scripts, packages and defaults.

Devuan mainstream didn't actually reject TDE but at this point probably has other priorities..

Hazy memory here too but the refracta-lang config script was meant to set the system locale and automatically sort the user keyboard layout for xfce, nothing to do with lightdm. It was meant to avoid long-winded typing at the boot menu. I saw it working with lxdm.

I do remember, newer lightdm was not very good with the language selector at the login screen. That would affect multiple users on the same box who might prefer different languages and was a different issue.

RE iso size.. might there be a way to avoid having two copies of initrd (one in /live and one in the squash, 25 M each as gzip)?

First I heard of it.. maybe they used some ideas or scripts from ExeGNU or Refracta I might have been involved with..

TDE doesn't need anything gtk at all. As well as the KDE3 desktop itself, they forked the original QT stuff (now TQT). TDE/TQT runs fine on Devuan and has very many good packages.. but maybe not everything you might want.

For example, synaptic and firefox-esr require gtk3 libs (not here, I use apt and palemoon).

I do use the refracta tools, which use yad. Yad's Debian version requires gtk3 but can easily be repackaged with gtk2.

It's going to get more difficult to escape gtk3 entirely..

As for the Falkon browser, here's a (not very impressive) review: https://www.dedoimedo.com/computers/falkon.html ..

Thanks fsmithred for testing this. Will sort these and any other glitches found and do a new version at some point. This was done from Jessie using a fresh devuan bootstrap (ascii) base, I don't yet know if the ascii debootstrap version fixes any of this.

Probably because /etc/default/grub needs the "GRUB_DISTRIBUTOR" line changed from Debian to Devuan before update-grub is run.

BTW, I see in /etc: debian_version and devuan_version. This probably means other Devuan systems get listed as Debian since, although upstream osprober now supports Devuan, it will find debian_version first. Delete all instances of it!

Cant reproduce that yet, still investigating. My only removables are FAT pen drives, in case that makes any difference

Needs a post-install script to sed edit /etc/trinity/tdm/tdmrc as the exegnu-installer does. Maybe it has but with errors like you said.

one tip on iso size.. xz compressed initrd saves ~20MB (edit /etc/initramfstools/initramfs.conf)

If anyone a live-media partition mounted RW, there is an updated initrd patch somewhere on the sourceforge page (/files?). Debian's live-boot doesn't support that. Maybe useful for refracta2usb.

I was missing "desktop-base". Also there was a small error in your code box (take out the "fi; "):

# for i in [0-9]*; do convert /usr/share/icons/hicolor/scalable/emblems/emblem-devuan.svg -resize ${i%/} $i/apps/kmenu.png; fi; done

All working now, thanks samhain.

Confirmed here, using vdev "sdb1" is referenced and symlinked in /dev/disk/by-label and /dev/disk/by-uuid but nothing for it in /dev/disk/by-id.