You are not logged in.
Pages: 1
You need to install dbus-x11 instead
For me that is the worst option.
I want as little of dbus on my system as possible.
Why? See here
https://wiki.hyperbola.info/doku.php?id … us_failure
and my post on dbus in the antiX forum
https://www.antixforum.com/forums/topic … ost-118320
Thanks for the hint.
Yes, especially since there are no active dbus daemons on my system, dbus-turnstile-user-session probably is of no effect.
But if I remove it, audacious will go with it, unless I install some dbus dummy.
So I rather keep dbus-turnstile-user-session installed.
It seems that I have discovered the source of the mystery.
My system is based on excalibur and the dbus daemons are excludes by a dbus-bin pin. In excalibur there is no turnstile. So installed it from antiX and the related packages and runit service too. Among these there is the optional package 'dbus-turnstile-user-session'. It contains a script '/etc/user_session.d/01_dbus':
#!/bin/sh
DBUS_SESSION_BUS_ADDRESS="unix:path=$XDG_RUNTIME_DIR/bus"
dbus-daemon --session --nofork --address=$DBUS_SESSION_BUS_ADDRESS &if [ "$XDG_RUNTIME_DIR" = "/run/user/$(id -u)" ]; then
echo export DBUS_SESSION_BUS_ADDRESS="$DBUS_SESSION_BUS_ADDRESS" >> "$XDG_RUNTIME_DIR/profile"
fi
Somehow this script catches the dependencies to dbus-bin.
If I remove 'dbus-turnstile-user-session' (which depends on turnstile), all programs that need dbus-bin will be removed.
So now I understand, why I had got no dependency complains, when I installed audacious.
Yes, it is very weird. Sound works well and is pure alsa, nothing else installed.
Dbus daemons are excluded – only libdbus is installed.
Fortunately I see is no need to remove turnstile.
My guess: The cause is solver 3.
I had not checked sound. Indeed, removing turnstile deactivates sound, even if I set audacious to alsa output.
'refracta_14_test_xlibre_turnstile-20260507_1411.iso' has the same login packages installed as my system.
But I have turnstile as well. If I would purge it, audacious, mousepad, xfburn and gcr would be removed too.
I could not detect why. Reverse depends of turnstile only show turnstile-dbgsym.
To my delight building my system was very uncomplicated. But then I realized, that Refracta 13.3 has libsystemd0 onboard.
I have replaced libsystemd0 by libelogind-compat and libelogind0. Then I tried to remove libelogind0 or replace it by a dummy, but that did brake the system.
This would be possible by using packages from the antiX nosystemd repo. To do so I would have to rise the priority of the antiX nosystemd repo, but then I shall get conflicts with the runit-services package. Maybe it is possible with sysv-init, but not with runit.
This motivated me to check your 'refracta_14_test_xlibre_turnstile-20260507_1411.iso'
Maybe you have found a better solution.
You have installed elogind/libelogind0 installed and seatd, libseat1 and turnstile as well. Both fulfill the same purpose. So which one is active? To figure that out I have run two tests:
(1) # apt purge seatd turnstile
After reboot the system did not show any differences.
seatd and turnstile seem to have no effect.
(2) # apt purge elogind (and keep seatd + turnstile)
This removed lxqt-panel. So Lxqt became unusable, but openbox was still working fine. Then I removed turnstile as well. Again this did not make any difference.
So in both cases turnstile seems to be irrelevant.
Now I have integrated these insights in /etc/refractasnaoshot.conf
The compression sections of my refractasnaoshot.conf now looks like this:
-----------------------------
[...]
# Uncomment one of the lines below to use xz or zstd compression for smaller iso.
# small and slow # relative size of the iso-image
#mksq_opt="-comp xz" # (-4.1% to zstd -19 | -15,4% to zstd -1)
#mksq_opt="-comp xz -b 262144" # (-2.0% to xz)
#mksq_opt="-comp xz -b 524288" # (-3.8% to xz)
#mksq_opt="-comp xz -b 1048576" # (-5.5% to xz)
# smaller and slower:
#mksq_opt="-comp xz -Xbcj x86" # (-1.5% to xz)
#mksq_opt="-comp xz -Xbcj x86 -b 524288" # (-5.4% to xz)
#mksq_opt="-comp xz -Xbcj x86 -b 1048576" # (-6.9% to xz)
# not as small but faster
#mksq_opt="-comp zstd -Xcompression-level 22" # (= zstd -19 | --ultra ?)
#mksq_opt="-comp zstd -Xcompression-level 19" # (-11.8% to zstd -1)
mksq_opt="-comp zstd -Xcompression-level 15" # (-11.1% to zstd -1)
#mksq_opt="-comp zstd -Xcompression-level 8" # (-7.0% to zstd -1)
#mksq_opt="-comp zstd -Xcompression-level 3" # (-3.1% to zstd -1)
#mksq_opt="-comp zstd -Xcompression-level 1" # (-17.2% to lz4)
#mksq_opt="-comp lz4" # (+20,8% to zstd -1)
[...]
When I was creating an iso-image of my new Refracta based system, I have noticed, that zstd -22 and zstd -19 deliver the same results. Obviously the zstd ultra mode does not work properly within Refractasnapshot. This inspired me to do some testing of different compression options.
Here are my results:
Total size of the system on qemu hd:
/ 4,09 GB + /home/ 289 MB = 4,37 GB
xz: 1572 MB (-4.1% to zstd -19 | -15,4% to zstd -1)
xz -b 262144: 1540 MB (-2.0% to xz)
xz -b 524288: 1512 MB (-3.8% to xz)
xz -b 1048576: 1486 MB (-5.5% to xz)
xz -Xbcj x86 1548 MB (-1.5% to xz)
zstd -22: 1639 MB ('zstd --ultra -22' does not work)
zstd -19: 1639 MB (-11.8% to zstd -1)
zstd -15: 1652 MB (-11.1% to zstd -1)
zstd -8: 1727 MB (-7.0% to zstd -1)
zstd -3: 1800 MB (-3.1% to zstd -1)
zstd -1: 1858 MB (-17.2% to lz4)
lz4: 2245 MB (+20,8% to zstd -1)
Relative compression speed according to Allan Jude (2017):
[https://openzfs.org/w/images/b/b3/03-Op … in_ZFS.pdf]
zstd -19 = 1x | zstd --ultra -22 = 0.58X
zstd -15 = 2.3x | zstd -8 = 16.1x
zstd -3 = 59.1x | zstd -1 = 101.5x
lz4 = 172.3x
According to Northon Torga (2024) xz and zstd -19 have a similar compression speed:
[https://ntorga.com/gzip-bzip2-xz-zstd-7z-brotli-or-lz4/]
zstd -19 = 1x | xz = 0.96x
Decompression speed of xz is about 5 times slower than zstd:
xz = ~0,18x | zstd = 1x | lz4 = ~ 3.2x
Depending on hardware and kind of files the actual results may differ.
Based on 'refracta_13.3_nox_amd64-20260501_1521.iso' I have created a similar system in qemu.
Already two years ago I had build such a system based on antiX-23 (bookworm). It is running stable on my desktop pc as my work horse.
These have been my steps:
(1) install refracta_13.3_nox with runit and seatd
(2) exclude unwanted packages via pins in /etc/apt/preferences.d:
00dbus-bin 00java-common 00libelogind0 00libsystemd0 00mono-runtime-common 00nodejs 00polkitd
(3) add antiX-26 repos to refracta
Give Devuan a higher priority than antiX in /etc/apt/preferences.d
To stay as close as possible to original Devuan I have turned on the antiX repos only when needed.
(i.e.: Arandr from Devuan cannot be installed – the antiX version is fine. Plank needs gconf-gsettings-backend from antiX to function properly.)
Once the base system has been completed, I shall keep the antiX repos turned on for upgrades.
(4) install from antiX-26:
# apt install turnstile runit-service-turnstiled libturnstile0
# apt install runit-service-seatd
(5) install my proven Openbox desktop
no display manager – just xinitrc | no pipewire – pure alsa
This setup should work with all window managers.
Desktops do not work. They all need active dbus daemons, whereas this setup uses libdbus-1-3 only.
(6) Power Management
For power management I use acpid, xset and the script suspend_if_idle (combined with i3lock and kdocker).
[https://gitlab.com/antiX-Linux/antix-go … type=heads]
(7) use lxterminal as pseudo gui
i.e.; obsession is crippled, lxsession needs polkit
As workaround in the openbox menu I use commands like the following:
<execute>lxterminal --geometry=39x4 --title=PowerOff -e bash -c "echo -- ???_PowerOff_??? __ [ENTER] = yes --; read line; (openbox --exit && sudo poweroff); echo read line"
<execute>lxterminal --geometry=90x32 --title=clamscan_Downloads -e "clamscan -v -r ~/Downloads; echo '<Enter> to Close Terminal'; read line"
So far I have found a solution for all my software needs within this setup.
The system is light, fast and stable and maybe a little more secure than standard.
I had to create one single dummy for pkexec, because gparted as dependency of refractainstaller-gui demands it. Of course pkexec is not really needed. Gparted starts as well with 'lxsu' or 'su-to-root -X -c'.
An extra benefit: The limitations have inspired me to find better alternatives to standard apps.
Example: I use SQLiteStudio instead of the bloated LibreOffice-Base.
Linux-libre works well, but I prefer the Devuan kernel for the option of non-free-firmware.
So far xorg has not caused any problems. Therefore I have not yet tested xlibre. It is in the antiX-26 repo. '#apt install -s xlibre' does not show any problems, but all the xserver-xorg-input-* packages will be removed.

screenshot runit-service-manager https://postimg.cc/RqQnGVvg

screenshot #apt install xlibre https://postimg.cc/kDFtwX7f

screenshot openbox desktop https://postimg.cc/mhNFV2TS
There seem to be two more bugs in refracta_13.2_nox_amd64-20251205_0850
(1) The iso boots well with efi in Qemu. But if installed to a Qemu VM with efi and runit, it does not boot. I could fix it with bootrepair-antix.
(2) The language settings do survive the installation, but not the keyboard setting. This could be fixed by
# dpkg-reconfigure console-setup (all settings were ok - only had to confirm all the settings)
# update-initramfs -u (to make keyboard settings permanent)
Pages: 1