The officially official Devuan Forum!

You are not logged in.

#1 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » Yesterday 22:14:03

Elogind bug report #938 was fixed and closed. New version 255.17-5 now in Ceres repo and works without issue.

#2 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-26 20:50:47

Thanks for the update and good news, abower.

Did more tests with seatd and openbox with the 255.17-4 updates for elogind, libelogind0 and libpam-elogind... no desktop with startx or xinit. Sent a bug report and was combined with yours.

#3 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-24 21:20:44

@fsmithred  I did what's below to see if the openbox desktop worked for me as it does for you.

[REMOVE, NOT USED] menu:amd64 2.1.51
[INSTALL, DEPENDENCIES] libobrender32v5:amd64 3.6.1-12+b4
[INSTALL, DEPENDENCIES] libobt2v5:amd64 3.6.1-12+b4
[INSTALL, DEPENDENCIES] libstartup-notification0:amd64 0.12-8+b1
[INSTALL, DEPENDENCIES] linux-base-6.18.5+deb14-amd64:amd64 6.18.5-1
[INSTALL] linux-image-6.18.5+deb14-amd64:amd64 6.18.5-1
[INSTALL] obconf:amd64 1:2.0.4+git20150213-3+b1
[INSTALL] openbox:amd64 3.6.1-12+b4
[REMOVE (PURGE)] fluxbox:amd64 1.3.7-1+b1
[UPGRADE] elogind:amd64 255.17-2 -> 255.17-4
[UPGRADE] libelogind0:amd64 255.17-2 -> 255.17-4
[UPGRADE] libpam-elogind:amd64 255.17-2 -> 255.17-4

Although it slows the OS considerably, I installed and rebooted to the latest kernel and edited .xinitrc for openbox. Neither xinit nor startx started the openbox desktop.

I downgraded the elogind pkgs to 255.17-2 and openbox started the desktop with xinit and startx.

Curious that with the 255.17-4 elogind pkgs installed, seatd allows a window mgr desktop to start from the console, but not elogind.

#4 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-24 21:16:41

Thanks, abower, good to know. Until I understand something of the changes made in one or more of those elogind 255.17-4 pkgs, your bug report (thanks for the link) and how our issues may or may not relate, you're absolved of any blame. smile

#5 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-24 16:43:59

Appreciate what you posted, fsmithred. Neither xinit nor startx worked with the upgrade. I didn't list libelogind-compat because it was already installed and upgraded to 255.17-4. When I downgraded the three pkgs above there was no dependency issue, so libelogind-compat remains upgraded and the desktop works.

Been a long time since using openbox. If after month or so and an update doesn't resolve anything with fluxbox, openbox may well be a good replacement.

Seatd is not installed but libseat1 is a dependency of xserver-xorg-core. Tried seatd last year thinking it would use less memory. For whatever reason, it used more memory. Other than trying seatd for a a few days, been using elogind + dependencies + the two recommends (libpam-elogind and polkitd) for a long while in Ceres without issues.

I upgraded the three pkgs and tried users, staff and operator as owner/group, but all were changed back to root when starting the desktop. One of those three 255.17-4 pkgs will not allow any owner/group for /tmp/.X11-unix other than root, and immediately changes it to root when another owner/group is used. 

Not ideal but is there a way to prevent or override the change to root as owner/group for /tmp/.X11-unix?

#6 Re: Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-24 01:57:56

mknoop wrote:

with xfce4 desktop

Thanks for replying. My guess ATM is the 255-17.4 upgrades were built with display mgrs in mind and haven't gotten around to OSs without one. An Xfce DE uses slim or lightdm, so no problem.

It's unstable smile and not the first time holding back upgrades for while has been necessary.

Log example:
Before upgrade and after downgrade, when xinit brings up the desktop, the local Xorg log shows:

[ 76525.071] _XSERVTransmkdir: Owner of /tmp/.X11-unix should be set to root

The /tmp/.X11-unix owner/group are my user name. If owner/group changed to root, no display, which suggests a permissions issue.

No Xorg log msg like above with upgrade and xinit failed, because owner/group of /tmp/.X11-unix had been changed to root. As root, I changed owner/group to my user name. Went back to my user acct but xinit failed because /tmp/.X11-unix owner/group had been auto-changed in a few seconds back to root. Tried # chmod 1777 /tmp but that also failed.

Anyone with wiser eyes than mine see something in the logs that stands out?

#7 Hardware & System Configuration » [SOLVED] Ceres upgrade problem » 2026-01-23 21:51:19

fanderal
Replies: 10

I run Ceres with no DE or display mgr. It's a console login and xinit starts a Fluxbox desktop.

Did aptitude update/upgrade yesterday. For maybe 6/8 months, an elogind upgrade killed the desktop. Elogind was to be upgraded so I dropped to the console to install the upgrades. When complete, entered xinit and got no screen, no desktop.

Used the Excalibur repo, downgraded the upgrades and the desktop worked. A process of elimination found these interdependent pkgs were the problem:

[DOWNGRADE] elogind:amd64 255.17-4 -> 255.17-2
[DOWNGRADE] libelogind0:amd64 255.17-4 -> 255.17-2
[DOWNGRADE] libpam-elogind:amd64 255.17-4 -> 255.17-2

Upgraded them again this morning and booted with different kernel, no desktop. Downgraded again and now on hold.

Xorg logs in ~/.local and parts 1 and 2 in /var
https://secureshare.in/local-Xorg.0.log
https://secureshare.in/var-Xorg.0.log-1
https://secureshare.in/var-Xorg.0.log-2

Would appreciate any insight into what those upgrades changed to cause this problem.

Thanks!

#8 Re: Other Issues » Looking for advice on VMs. » 2026-01-11 17:23:25

EnglishMohican wrote:

but the start-up reset sets up 144Hz

Does the BIOS have an option to default to 60Hz? If not and not needed, is there an option to disable 144Hz?

#9 Re: Other Issues » troubleshooting help needed alsa, » 2026-01-07 17:44:48

zapper wrote:

just rolled back using a vm qcow2 transfer with a less updated version of the same OS...

Unlikely makes a difference but when using the linux-libre kernels and headers last year I downloaded the deb pkgs and installed them locally.

Did starting a browser from the terminal show anything useful?

As an earlier kernel version and another audio driver didn't resolve it, don't know what else to suggest.

#10 Re: Installation » No terminals after installation of Excalibur » 2026-01-07 01:00:45

switching2Devuan wrote:

When logging into the installed DE (LXDE) and then trying to get to a terminal with CTRL-ALT-F(1-6)

To clarify words... using CTRL-ALT-F(1-6) doesn't open a terminal. It opens a console (TTY) with a black screen and login prompt. If, say, the active desktop is on TTY2 and CTRL-ALT-F2 is clicked, it should do nothing. If CTRL-ALT-F3 is clicked instead, the desktop remains active but the screen switches to the TTY3 console.

LXDE's terminal is lxterminal and can be opened from the menu or possibly with Ctrl-Alt-T.

#11 Re: Other Issues » troubleshooting help needed alsa, » 2026-01-06 18:41:32

Local sound works but not the browsers... odd. Is there something invoked when starting a browser (firejail, etc) that you can test sound without using?

My mobo card/driver:

$ lspci -v | grep Audio
00:03.0 Audio device: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor HD Audio Controller (rev 06) (prog-if 00 [HDA compatible])
00:1b.0 Audio device: Intel Corporation 9 Series Chipset Family HD Audio Controller (prog-if 00 [HDA compatible])

$ lspci -k -s 00:1b.0
00:1b.0 Audio device: Intel Corporation 9 Series Chipset Family HD Audio Controller
        Subsystem: Gigabyte Technology Co., Ltd Device a182
        Kernel driver in use: snd_hda_intel
        Kernel modules: snd_hda_intel

Your info shows the snd_hda_intel module loaded but not used as the driver. Have you tried changing the audio driver?

Last year I used an old linux-libre kernel for this old mobo and sound wasn't an issue, local or browser. Running an LTS kernel (from the repo?) might fix the sound.

#12 Re: Other Issues » troubleshooting help needed alsa, » 2026-01-05 20:16:28

Is it just the browsers or no sound at all for local music/video?

The 'HDA Intel PCH' sound card is for the Kabylake Intel mobo, yet a 'k6.18.3-gnu' alsa driver suggests an AMD mobo and kernel.

Are you running the just released 6.18.3 kernel with built in components for both AMD and Intel?

#13 Re: Off-topic » Is AI going to become a Frankenstein monster? » 2025-12-30 18:36:49

pcalvert wrote:

I think it's time to bring this old science-fiction movie to the attention of many more people:

Colossus: The Forbin Project
https://www.imdb.com/title/tt0064177/

Appreciate the tip.* Like the Terminator series which later followed a similar concept, corporate devs who are in it for the science and rarely consider (or understand) real world consequences are blamed. I tend to look at who mandates the devs'/scientists' efforts and arrogantly/foolishly perceive those efforts as a means to more power and control. Something like the 1967 Bond movies spoof, The President's Analyst.

* Watch/download Colossus: The Forbin Project at archive-org.

#14 Re: Off-topic » Is AI going to become a Frankenstein monster? » 2025-12-28 22:07:01

Not exactly a Frankenstein... something pervasively worse.

While the news is beginning to forcast cascading financial doom for AI:

Incoming AI CRASH - The Truth Behind the Hype No One Wants to Admit
Nov 2025
9:39
https://rumble.com/v722s8g-incoming-ai- … admit.html

Yet the not-so-publicized development is alive and well:

greenjeans wrote:

LLM's are not the frankenstein monster, THIS is the frankenstein monster: https://www.sciencealert.com/computers- … e-prepared

And stargate-scm's 'Intergalactic Assimilation' posted text gives a plausible picture of how it would happen: in the end we will all be assimilated

Consider how systemd became a reality, followed by blacklisting Master/Slave and Whitelist/Blacklist labels, followed by user/dev forums and lists rewriting rules of conduct so we would learn the approved of and proper way to communicate with one another... these inanities became the new normal, as have many other new normals. Is there any reason to believe this gov't funded and big-tech developed biotechnology wouldn't follow the same path?

#15 Re: Installation » LXQt cant install at Ceres » 2025-12-19 19:12:13

2. lxqt-session:amd64 Depends lxqt-qtplugin (>= 2.3.0~)
but none of the choices are installable:
[no choices]

Lxqt-qtplugin is the problem. Version 2.3.0 is not available now but will be available soon. This occasionally happens in Ceres.

LXQt system integration plugin for Qt
https://tracker.debian.org/pkg/lxqt-qtplugin

#16 Re: DIY » ALSA-only purists: Question, new GUI app for the mixer and EQ? » 2025-12-16 18:16:34

Had alsamixer and alsamixer -D equal in the menu but replaced both with your app... polished, easy to use and presets. Thanks greenjeans!

#17 Re: Hardware & System Configuration » [SOLVED] File in ~/.config/pulse » 2025-11-27 06:39:30

Altoid wrote:

Do libavdevice62, libguvcview-2.2, libsdl2 get removed if you purge mpv?

Good thought, but no.
Note: libguvcview-2.2 is a library for guvcview, an app to test/set webcam functions and create local videos.

# aptitude purge mpv
<snip>
The following packages will be REMOVED:  
  libdisplay-info3{u} liblua5.2-0{u} libmujs3{u} libpipewire-0.3-0t64{u} 
  libsixel1{u} libspa-0.2-modules{u} libva-wayland2{u} 
  libwebrtc-audio-processing-1-3{u} libxpresent1{u} mpv{p} 
0 packages upgraded, 0 newly installed, 10 to remove and 1 not upgraded.
Need to get 0 B of archives. After unpacking 11.8 MB will be freed.

However:

# aptitude purge libpulse0
The following packages will be REMOVED:  
  libasyncns0{u} libpulse0{p} 
0 packages upgraded, 0 newly installed, 2 to remove and 1 not upgraded.
Need to get 0 B of archives. After unpacking 1,044 kB will be freed.
The following packages have unmet dependencies:
 mpv : Depends: libpulse0 (>= 0.99.4) but it is not going to be installed
 libsphinxbase3t64 : Depends: libpulse0 (>= 0.99.1) but it is not going to be installed
 libavdevice61 : Depends: libpulse0 (>= 0.99.1) but it is not going to be installed
 libguvcview-2.2 : Depends: libpulse0 (>= 0.99.1) but it is not going to be installed
 libsdl2-classic : Depends: libpulse0 (>= 5.0) but it is not going to be installed
The following actions will resolve these dependencies:

      Remove the following packages:                             
1)      ffmpeg [7:7.1.2-1+b1 (now, unstable)]                    
2)      guvcview [2.2.1-1 (now, unstable)]                       
3)      libavdevice-dev [7:7.1.2-1+b1 (now, unstable)]           
4)      libavdevice61 [7:7.1.2-1+b1 (now, unstable)]             
5)      libavfilter-dev [7:7.1.2-1+b1 (now, unstable)]           
6)      libavfilter10 [7:7.1.2-1+b1 (now, unstable)]             
7)      libguvcview-2.2 [2.2.1-1 (now, unstable)]                
8)      libpocketsphinx3 [0.8+5prealpha+1-15+b4 (now, unstable)] 
9)      libsdl2-2.0-0 [2.32.10+dfsg-4 (now, unstable)]           
10)     libsdl2-classic [2.32.10+dfsg-4 (now, unstable)]         
11)     libsdl2-image-2.0-0 [2.8.8+dfsg-1 (now, unstable)]       
12)     libsphinxbase3t64 [0.8+5prealpha+1-21+b1 (now, unstable)]
13)     mpv [0.40.0-5+b1 (now, unstable)]                        
14)     pipewalker [1.1-1 (now, unstable)]                       

Accept this solution? [Y/n/q/?]y 

The following packages will be REMOVED:
  ffmpeg{a} guvcview{a} jackd2{u} libass9{u} libasyncns0{u} libavc1394-0{u} 
  libavdevice-dev{a} libavdevice61{a} libavfilter-dev{a} libavfilter10{a} 
  libavformat-dev{u} libbs2b0{u} libcaca0{u} libcdio-cdda2t64{u} 
  libcdio-paranoia2t64{u} libdc1394-25{u} libdisplay-info3{u} libdovi3{u} 
  libfftw3-double3{u} libflite1{u} libgsl28{u} libgslcblas0{u} 
  libguvcview-2.2{a} libiec61883-0{u} libjack-jackd2-0{u} liblilv-0-0{u} 
  liblua5.2-0{u} libmujs3{u} libmysofa1{u} libopenal-data{u} libopenal1{u} 
  libpipewire-0.3-0t64{u} libplacebo351{u} libpocketsphinx3{a} 
  libportaudio2{u} libpostproc-dev{u} libpostproc58{u} >> libpulse0{p} << 
  libraw1394-11{u} librubberband2{u} libsdl2-2.0-0{a} libsdl2-classic{a} 
  libsdl2-image-2.0-0{a} libserd-0-0{u} libshaderc1{u} libsixel1{u} 
  libsndio7.0{u} libsord-0-0{u} libspa-0.2-modules{u} libsphinxbase3t64{a} 
  libsratom-0-0{u} libswscale-dev{u} libswscale8{u} libunibreak6{u} 
  libv4l-0t64{u} libv4lconvert0t64{u} libva-wayland2{u} libvidstab1.1{u} 
  libwebrtc-audio-processing-1-3{u} libxpresent1{u} libzimg2{u} 
  libzix-0-0{u} mpv{a} pipewalker{a} python3-dbus{u} 
0 packages upgraded, 0 newly installed, 65 to remove and 1 not upgraded.
Need to get 0 B of archives. After unpacking 110 MB will be freed.
Do you want to continue? [Y/n/?]n

I can remove it but finding daily use replacements for mpv, ffmpeg and guvcview without needing libpulse0 may prove tedious and not worth the effort.

Altoid wrote:

If so, you could use another media player and in the process rid yourself of libpulse0.

Perhaps there are a few. Thought of vlc but libpulse0 is a dependency. Any suggestions?

#18 Re: Hardware & System Configuration » [SOLVED] File in ~/.config/pulse » 2025-11-26 06:57:24

Altoid wrote:

The apulse package is (?) a compatibility layer to allow pulseaudio applications to work with ALSA.

That's my understanding. Can't recall why I installed it.

Altoid wrote:

Q: do you want pulseaudio applications running in your box?

Nope. Won't use applications with a PA dependency.

Altoid wrote:

If not, check with aptitude why for the reason the apulse package was installed.
If there is no valid dependency, you can probably remove it.

No valid dependency. I removed it.

Altoid wrote:

Apparently it (libpulse0) is a package of pulseaudio libraries.
But if you do not have pulseaudio installed, what would the use case be?

No idea, yet there's a bit more. I set aptitude to purge libpulse0. It would have broken libavdevice62, libguvcview-2.2, libsdl2 and mpv, with no solution.

#19 Re: Hardware & System Configuration » [SOLVED] File in ~/.config/pulse » 2025-11-25 17:22:07

Pulseaudio never installed.

apt list | grep installed | grep pulse*
apulse/unstable,now 0.1.14-1 amd64 [installed]
libpulse0/unstable,now 17.0+dfsg1-2+b1 amd64 [installed]

All lines commented in libpulse0's /etc/pulse/client.conf.

No pulse dir/file in ~/.config or /tmp

aptitude why libpulse0
i  mpv Depends libpulse0 (>= 0.99.4)

Conky-std installed, no libpulse0 dep.

#20 Re: Desktop and Multimedia » [SOLVED] Can't suspend, reboot, or shut down in a WM » 2025-11-09 18:26:27

Got curious about IceWM and installed it. As fredbird67 posted, shutdown and reboot didn't work. In /usr/share/icewm/README:

The directory /etc/X11/icewm used to contain copies of default configuration files shipped with the icewm package. The original files were and are located in /usr/share/icewm/ directory. This was considered service to the users but it has also led to confusion and made package handling more complicated.

Therefore, the default configuration files are now only located in
/usr/share/icewm/ directory. To customize something for all system users, the system administrator shall copy the particular file to /etc/X11/icewm when needed, and icewm will automatically make use of that copy. Regular users can do the same, installing their version into ~/.icewm/.

As root, I copied the menu and preferences files from /usr/share/icewm to /etc/X11/icewm and edited these two lines in preferences:

#  Command to reboot the system.
From
# RebootCommand="test -e /run/systemd/system && systemctl reboot || sudo -n /sbin/reboot"
To
RebootCommand="sudo shutdown -r now"

and 

#  Command to shutdown the system.
From
# ShutdownCommand="test -e /run/systemd/system && systemctl poweroff || sudo -n /sbin/poweroff"
To
ShutdownCommand="sudo shutdown -h now"

NOPASSWORD lines for reboot and shutdown were already in /etc/sudoers so those two cmds in IceWM's menu worked. Same results had ~/.icewm been created with the edited preferences file. My system isn't set up for hibernate but when the line for it in preferences is edited, it should work.

#21 Re: Other Issues » Apt .sources changes » 2025-11-02 22:25:51

https://daniel-lange.com/archives/192-M … rixie.html

@stargate-scm  Thanks for the link. Wanted to disable that warning but like the last poster reported, it didn't work with apt-3.x. In a terminal, the apt warning can be disabled (without modernizing) after editing the format in the new apt.conf.d file

from
APT::Get::Update::SourceListWarnings "false";

to
APT::Get::Update::SourceListWarnings "0";

#22 Re: Desktop and Multimedia » A tale of two laptops » 2025-10-28 18:04:30

@paramtn  Dunno why it didn't work as I've used what I posted. Can't recall the system default but you might check /etc/default/console-setup to see if the FONTSIZE line was changed from default to FONTSIZE="16x32". And that the GRUB_GFXMODE line is commented in /etc/default/grub before invoking # update-grub.

#23 Re: Desktop and Multimedia » A tale of two laptops » 2025-10-27 18:39:00

To edit the grub font and size so both laptops are same/similar:

1. Pick a font for your grub menu. DejaVu Sans Mono is in this example.

2. Convert the font to grub2's format using the grub-mkfont command, and set the font size:

# grub-mkfont --output=/boot/grub/DejaVuSansMono.pf2 \
--size=24 /usr/share/fonts/truetype/dejavu/DejaVuSansMono.ttf

3. In /etc/default/grub, add this GRUB_FONT line with the path to your freshly-converted font:

GRUB_FONT=/boot/grub/DejaVuSansMono.pf2

4. Enter # update-grub

This edit changes the font size in grub and the console to ~14. Adjust the '--size=' in the grub-mkfont cmd as needed. If the font is still pre-edit size after updating grub and you don't want to reboot, enter # setupcon -f in any console. This edit does not use the GRUB_GFXMODE line for resolution (commented).

#24 Re: Devuan » Update of Devuan causes VirtualBox shared folder corruption » 2025-09-04 21:43:42

Don't know if this resolves the posted issue. Just noticed this thread since my previous post.

Between May/August of 2025, Ceres upgraded virtualbox from 7.0.20-dfsg-1.2 to 7.1.12-dfsg-2+b1. I think it was after the 7.1.10-dfsg-1 upgrade in June when VBox failed to start. Was odd because it was just an update. Did lsmod | grep vbox and in addition to vboxdrv not loaded, got this:

libkmod: ERROR: conf_files_filter_out: Directories inside directories are not supported: /etc/modprobe.d/virtualbox-dkms.conf

The error? Inside /etc/modprobe.d, VBox created the directory virtualbox-dkms.conf and inside it was the file virtualbox-dkms.modprobe.conf. I moved the directory out of /etc/modprobe.d and started the VBox modules: no more lsmod errors and VBox worked.

A search found bug report 1105784 filed in May. The patch to fix it was in July and noted:

This patch does two things:

It moves the modprobe.d file to /usr/lib/modprobe.d, instead of /etc/modprobe.d.

It actually installs the file correctly. <snip>

Why? Blacklisting the kvm module was required because it conflicted with VBox's modules.

The full VBox suite, upgraded to 7.1.12-dfsg-2+b1 two weeks ago, is in Ceres contrib and works without issue.

#25 Re: DIY » A Survey of User-made Content » 2025-09-04 21:36:05

greenjeans wrote:

Both the Mate terminal and LXterminal support mouse-wheel scrolling.

Thanks, greenjeans. What you quoted was possible alternatives for others who want to use a mouse. Gave lxterminal a spin anyway and it reminded me of why I use the Up/Dn arrows and PgUp/PgDn to scroll. Using the mouse means having to move my right hand back and forth between the mouse and the keyboard (aptitude, man pages, etc). I find it easier to just keep both hands on the keyboard.

About lxterminal: on this OS, the mouse defaults to skipping two lines instead of scrolling one line at a time. Wasn't useful in aptitude. smile Didn't see a scroll setting in the GUI prefs and didn't look further, just deleted it. Appreciate the thought.

Board footer

Forum Software