Looks like the package usrmerge is not installed. It is mandatory to install before switching to excalibur or ceres.

# apt-get install usrmerge

"unreadable" in case of an optical drive usually means "defective medium" - or more seldon "drive defective".
To find out: Try other CD/DVD's - as Altiod mentioned.

No!
Iso file systems are readable on any OS. At least should be.

 $ grep 'Arch' /boot/grub/grub.cfg

returns nothing.

I'm out of ideas by now.

Thanks for the replies!

That is excatly the point. Arch should be there, but it's not!
Silly me ... should have looked at /var/log/syslog before:

(...) 40grub2: debug: parsing: ### BEGIN /etc/grub.d/30_os-prober ###
(...) 40grub2: debug: parsing: menuentry 'Arch Linux (rolling) (on /dev/nvme0n1p3)' --class arch --class gnu-linux --class gnu --class os $menuentry_id_option 'osprober-gnulinux-simple-4b419f03-1a7f-4fa8-836c-4018cd10c5ca' {
(...) 40grub2: Skipping entry 'Arch Linux (rolling) (on /dev/nvme0n1p3)':
(...) 40grub2: appears to be an automatic reference taken from another menu.lst

So 40grub2 skiped the detected boot entry! But why?
For now it's "/usr/lib/linux-boot-probes/mounted/40grub2". To be continued ... .
Edit: There is no "menu.lst" below "/boot" or "/etc".

@g4sra: There is one bootable SSD, with one booting grub2 from "Devuan 6". There is no shared "/boot" partition, only "/boot/efi" is mounted on all three.

This sounds strange but reasonable  ... so checking for 'smapi'.

Result: Package 'tp-smapi-dkms' is installed. There is no reason why, but it is. This computer is a common ATX-PC, not a notebook.

And the liquorix kernel just compiles / works!
Thanks a lot.

EDIT: Knowing this makes your first reply even more understandable.

Thanks for the replies.
@JayneCobb: Welcome to the forum!
@steve_v: Nice explanation and thanks for the insight.

Still not sure how to proceed. Since I don't have a thinkpad, is the module needed at all? Maybe it's possible to disable the build of 'tp_smapi'.
Or more likely, just waiting and hoping for the next liquorix kernel; linux-6.16.1 is already released, with or without an updated driver.
I don't see me kernel-hacking these days.

I remember having QasMixer in my alsa-only-time. It was recommended from someone in this forum and worked good.

It works - shows alsa channels and also pipewire and the inactive hdmi-sound-devices. There are more Qas* sound tools.

Nice work and install tutorial!

Here is a short collection of incedences on my excalibur:
Compiling: pkg-config --cflags --libs gtk+-3.0 throwed an error, solved by installing libgtk-3-dev .
Scraping: There were two or three instances of the script running, beside free cpu cores.
You mentioned "/var/lib/dpkg/status" and I thought this text file would be processed to get package related information.
dpkg was querried a lot (multiple ps ax in the terminal to kill time, e.g.: dpkg-query --search -- galculator.desktop )

#-#-#-#-#-#

Sorting:
A doubled "ParaView" entry cought my attention - see code section.

Guess is:
The list is sorted by the name of the *.desktop file. The shown "Application Name" is from the 'Name="Application Name"'-field.
*Highly accurate* entries like "Document Viewer" are existing too.
I'm missing a reference to the executed binary itself.
Or maybe executable name for sorting?

**imaginary screen shot of the program list**

(... "O" section here?!)
ParaView
QjackCtl
qpwgraph
samplev1
Wireshark
Mousepad
ParaView Cilent
(... more "P" entries)

**end of imaginary screen shot**

$ find | grep desktop$ | grep -i paraview
/usr/share/applications/org.paraview.ParaView.desktop
/usr/share/applications/paraview.desktop

$ diff  /usr/share/applications/paraview.desktop /usr/share/applications/org.paraview.ParaView.desktop
3,5d2
< Name=ParaView Client
< GenericName=Data Viewer
< Comment=ParaView allows viewing of large data sets
7c4,7
< Terminal=false
---
> Name=ParaView
> Comment=Parallel visualization application
> Exec=paraview %f
> TryExec=paraview
9,11c9,10
< MimeType=application/x-paraview;
< Categories=Education;Math;Science;
< Exec=paraview
---
> StartupWMClass=paraview
> Categories=Qt;Science;DataVisualization;

@Prowler_Gr  Thanks for the head up.

First reading this made me clone my debian-trixie-VM-with-sysvinit-enabled before trying apt-get udate; apt-get upgrade.
grub still boots /sbin/init as PID 1, so no harm done.

I'm not exactly sure what the removed part of code does - assumtion is the code enables a way to have sysvinit, systemd and upstart installed in parallel and select init via grub.

Moore likely the other way around: Ask an AI, then get your GPU toasted.

@Glen: Thanks for sharing. Appreciated.