[guix shell](https://guix.gnu.org/manual/devel/en/html_node/Invoking-guix-shell.html) allows users to create a shell and control what resources are available. How can I use the guix.scm file to specify an environment variable in this shell environment? In this specific example, I'm just looking for a TEST environment variable with the value 2. ChatGPT gave me two suggestions below which did not work, nor are these options defined in the [package reference](https://guix.gnu.org/manual/en/html_node/package-Reference.html) . Both give me the same error: *extraneous field initializers*. I also saw [this thread](https://lists.gnu.org/archive/html/help-guix/2022-05/msg00025.html) about the inability to define environment variables in a manifest file and one user said that it should be possible to do with guix.scm files (or, the type of files passed to guix shell with --file argument), but I'm not sure if that's correct. I tried the guile [setenv](https://www.gnu.org/software/guile/manual/html_node/Runtime-Environment.html#index-setenv) Scheme procedure but I get the error: *invalid field specifier*. Here's an example guix.scm with the three options I tried at the bottom:

(use-modules
	(guix)
	(guix build-system gnu)
	((guix licenses) #:prefix license:)
	(gnu packages bison)
	(gnu packages compression)
	(gnu packages flex)
	(gnu packages m4)
	(gnu packages tls)
)

(package
	(name "coreboot")
	(version "dan")
	(source (local-file "." #:recursive? #t))
	(build-system gnu-build-system)
	(native-inputs
		(list
			bison
			flex
			m4
			openssl
			zlib
		)
	)
	(synopsis "coreboot")
	(description "coreboot")
	(home-page "https://www.coreboot.org/ ")
	(license license:expat)
	(shell '(#:variables (,"TEST=2"))) ; Does not work
	(environment-variables '("TEST" "2")) ; Does not work
	(setenv "TEST" "2") ; Does not work
)

I've made my Guix system entirely declarative, always adding packages through guix package -m mymanifest.scm. Sometimes this fails because of conflicting dependencies. E.g. right now I have gcc-toolchain in my manifest, amongst dozens of other packages, and this message shows up:

guix package: error: profile contains conflicting entries for gcc-toolchain
guix package: error:   first entry: gcc-toolchain@12.3.0 /gnu/store/rfm800pq3q2midj29a4xlikdzjp1ps2l-gcc-toolchain-12.3.0
guix package: error:   second entry: gcc-toolchain@11.3.0 /gnu/store/ks87cpc36kh8hqwr569pks4yrzfl7mnv-gcc-toolchain-11.3.0
hint: You cannot have two different versions or variants of `gcc-toolchain' in the same profile.

I suppose gcc-toolchain that is included in the manifest is 12.3.0, and some other package has gcc-toolchain@11.3.0 as an (indirect) dependency. But I don't know which package that is. In this particular case, the problem was easily resolved, because I also included gcc in the manifest, and that resulted in the warning that package 'gcc' has been superseded by 'gcc-toolchain', and removing gcc seems to have fixed the issue. However, I'd like a more general applicable method to find the 'offending' package. That is, given a manifest file, say mymanifest.scm, how can I figure out through which dependency chain a given package, say gcc-toolchain@11.3.0, ends up in the profile? I usually do a binary search, by simply commenting out (or back in) half of the packages and see when the problem disappears (or reappears), but this is a manual and tedious process. The ugly naive way I can automate this myself would be to loop through all the packages, call guix graph on each package, and then grep for the problematic dependency. Better would be to figure out what guix graph is doing, and do that for a whole manifest. Or perhaps convert a manifest into a package (empty, except for dependencies), and then call guix graph on that. But there is probably a better way.

Guix is based on Nix. I know that Guix uses Scheme and has only free software in repos. What are the main differences between them?

I am moving GNOME Boxes virtual machines from Fedora to Guix following [the guide on Fedora wiki](https://fedoraproject.org/wiki/Move_boxes_virtual_machines) . On the Fedora workstation:

$ virsh list --all
 Id   Name           State
-------------------------------
 -    win10          shut off
$ cd ~
$ virsh dumpxml win10 >win10.xml
$ mv .local/share/gnome-boxes/images/win10 win10

Then, I moved win10 and win10.xml to the home directory at the new machine. On that Guix system:

$ guix install gnome-boxes
$ cd ~
$ mv win10 .local/share/gnome-boxes/images/win10
$ guix install libvirt
$ virsh create win10.xml
error: Failed to create domain from win10.xml
error: Cannot check QEMU binary /usr/bin/qemu-system-x86_64: No such file or directory

qemu-system-x84_64 is located in a different place:

$ type qemu-system-x86_64
qemu-system-x86_64 is /run/current-system/profile/bin/qemu-system-x86_64

So, how do I complete importing the virtual machine?

After booting Guix for the first time, I noticed that tapping on touchpad doesn't render any action. Apparently, Guix comes with a default configuration for touchpad allowing me to click on touchpad, which is mostly sufficient, but having tap disabled, which may be tiring with extensive touchpad use. This is similar to the default configuration one could encounter on Debian . For just in case, I used xinput list to find out the name of the touchpad, and xinput list-props followed by the touchpad name, to check its capabilities. Then I found the location of synclient binary with find, and, consulting the synaptics(4) man page online, invoked synclient with TapButton1=1 to enable one-finger tap, then with TapButton2=3 to set two-finger tap to right button event (which brings the right-click menu). While on it, I also enabled horizontal two-finger scrolling with HorizTwoFingerScroll=1, as well as natural scrolling by setting the VertScrollDelta and HorizScrollDelta to the same values which I learned from xinput list-props but negative (by adding - before each value). Now, one would typically add this settings to /etc/X11/xorg.conf.d/70-synaptics.conf. But since this is Guix, the file is not in /etc but in the GNU store and the touchpad driver is supposed to be configured in /etc/config.scm. The GNU Guix Reference Manual for version 1.3.0 doesn't mention how to configure synaptics. What can I do on a Guix system, so that I do not need configuring the touchpad myself after each login?

How can I activate the screen if it doesn't turn on after resuming from the S3 deep sleep state (suspend to RAM)? The test for suspending to RAM appears to be successful.

# echo core > /sys/power/pm_test
# echo mem > /sys/power/state

The device seems suspending.

# echo none > /sys/power/pm_test
# echo mem > /sys/power/state

The power LED turns off. I can resume it by pressing a key. The device powers on, but the screen remains unresponsive. I am able to type commands in the terminal and also hear system sounds.

# dmesg > log.txt
# reboot

Here are the dmesg messages recorded from the moment I initiated the suspension to RAM until right before I rebooted.

[  986.764160] PM: suspend entry (deep)
[  986.775044] Filesystems sync: 0.010 seconds
[  986.790717] rfkill: input handler enabled
[  986.791783] Freezing user space processes
[  986.793695] Freezing user space processes completed (elapsed 0.001 seconds)
[  986.793700] OOM killer disabled.
[  986.793702] Freezing remaining freezable tasks
[  986.794847] Freezing remaining freezable tasks completed (elapsed 0.001 seconds)
[  986.794880] printk: Suspending console(s) (use no_console_suspend to debug)
[  986.909707] ACPI: EC: interrupt blocked
[  986.946292] ACPI: PM: Preparing to enter system sleep state S3
[  986.946831] ACPI: EC: event blocked
[  986.946833] ACPI: EC: EC stopped
[  986.946834] ACPI: PM: Saving platform NVS memory
[  986.947201] Disabling non-boot CPUs ...
[  986.949337] smpboot: CPU 1 is now offline
[  986.952353] smpboot: CPU 2 is now offline
[  986.955071] smpboot: CPU 3 is now offline
[  986.957565] smpboot: CPU 4 is now offline
[  986.960071] smpboot: CPU 5 is now offline
[  986.960279] ACPI: PM: Low-level resume complete
[  986.960279] ACPI: EC: EC started
[  986.960279] ACPI: PM: Restoring platform NVS memory
[  986.960279] AMD-Vi: Virtual APIC enabled
[  986.960279] AMD-Vi: Virtual APIC enabled
[  986.960279] LVT offset 0 assigned for vector 0x400
[  986.960279] Enabling non-boot CPUs ...
[  986.960279] x86: Booting SMP configuration:
[  986.960279] smpboot: Booting Node 0 Processor 1 APIC 0x1
[  986.961196] ACPI: \_SB_.PLTF.C001: Found 3 idle states
[  986.961203] ACPI: FW issue: working around C-state latencies out of order
[  986.961407] CPU1 is up
[  986.961423] smpboot: Booting Node 0 Processor 2 APIC 0x2
[  986.961697] ACPI: \_SB_.PLTF.C002: Found 3 idle states
[  986.961703] ACPI: FW issue: working around C-state latencies out of order
[  986.961897] CPU2 is up
[  986.961919] smpboot: Booting Node 0 Processor 3 APIC 0x4
[  986.962448] ACPI: \_SB_.PLTF.C003: Found 3 idle states
[  986.962460] ACPI: FW issue: working around C-state latencies out of order
[  986.962924] CPU3 is up
[  986.962943] smpboot: Booting Node 0 Processor 4 APIC 0x5
[  986.963563] ACPI: \_SB_.PLTF.C004: Found 3 idle states
[  986.963580] ACPI: FW issue: working around C-state latencies out of order
[  986.964123] CPU4 is up
[  986.964137] smpboot: Booting Node 0 Processor 5 APIC 0x6
[  986.964691] ACPI: \_SB_.PLTF.C005: Found 3 idle states
[  986.964708] ACPI: FW issue: working around C-state latencies out of order
[  986.965199] CPU5 is up
[  986.966014] ACPI: PM: Waking up from system sleep state S3
[  986.966512] ACPI: EC: interrupt unblocked
[  986.968633] ACPI: EC: event unblocked
[  986.968812] pci 0000:00:00.2: can't derive routing for PCI INT A
[  986.968818] pci 0000:00:00.2: PCI INT A: no GSI
[  986.992288] nvme nvme0: Shutdown timeout set to 8 seconds
[  987.112414] nvme nvme0: 12/0/0 default/read/poll queues
[  987.214103] OOM killer enabled.
[  987.214107] Restarting tasks ... done.
[  987.216294] random: crng reseeded on system resumption
[  987.216901] PM: suspend exit
[  987.280708] ata1: SATA link down (SStatus 0 SControl 300)
[  987.281085] ata2: SATA link down (SStatus 0 SControl 300)
[  987.402492] rfkill: input handler disabled
[  991.340829] acpi AMDI0010:03: Already enumerated
[  996.400591] acpi AMDI0010:03: Already enumerated

Additional information about the system:

$ uname -rvm
6.3.10-gnu #1 SMP PREEMPT_DYNAMIC 1 x86_64
$ lspci -s 0000:00:00.2
00:00.2 IOMMU: Advanced Micro Devices, Inc. [AMD] Renoir/Cezanne IOMMU
$ lspci -k -s 02:00.0
02:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Renoir (rev c3)
	Subsystem: Acer Incorporated [ALI] Device 142b
	Kernel modules: amdgpu
$ shepherd --version | grep shepherd
shepherd (GNU Shepherd) 0.10.1

--- There is no such issue on Ubuntu or Fedora. But I can reproduce it there too by adding nomodeset to the kernel boot parameters. The issue persists on Trisquel and Guix systems both running blob-free kernels. So, I conclude that the issue is a consequence of a fallback mechanism when the kernel does not load firmware for AMD GPU with Ryzen 5 4500U. I suspect that it depends on the installed UEFI firmware capabilities but I am not sure how to investigate it. Considering that I have access to the terminal after the system wakes from sleep, could I simply reset the display from the command line somehow?

I installed the Guix OS in a partition of my HDD for testing, alongside my daily OS partition. Now the previous grub is overwritten and my daily distro is not accessible anymore. I didn't found "grub-mkconfig" in the packages search. How do you update grub list with Guix please ?

GNOME Files (Nautilus) fails to generate thumbnails on an up-to-date Guix system, which is a known outstanding issue . Small empty files just go to ~/.cache/thumbnails/fail/gnome-thumbnail-factory. While this is being debugged and fixed, is there another way to browse files with thumbnails being displayed?

How do I install NodeJS on GuixSD?

Running sudo guix gc --list-roots on my Guix System I see a lot of system generations lying around, going all the way back to when I first installed the system.

$ sudo guix gc --list-roots
...
/var/guix/profiles/system-22-link
/var/guix/profiles/system-1-link
/var/guix/profiles/system-10-link
/var/guix/profiles/system-11-link
/var/guix/profiles/system-12-link
/var/guix/profiles/system-13-link
/var/guix/profiles/system-14-link
/var/guix/profiles/system-15-link
...

These are not garbage collected when I run sudo guix gc --delete-generations=30d as I would expect. Why is this and how do I GC the system roots?

I want to install Guix System Distribution (or GuixSD, or just Guix) alongside some other distros. The problem is, Guix doesn't recognise Realtek drivers, which are nonfree, so I can't use WiFi. I can install Guix with Ethernet, but it is very unstable, so I want to install it with my existing Debian (buster) installation so I can then add the Nonguix channel and install the drivers. Problem is, I don't really know how. I'm pretty sure it includes installing Guix onto Debian, installing some packages, and strapping GuixSD onto the root (atleast, that's how NixOS does it, which Guix is inspired by) or it includes chrooting into the ISO. I would appreciate any help possible. ## System Info

# neofetch # some info was filtered

root@deblugo 
--------------- 
OS: Debian GNU/Linux 10 (buster) x86_64 
Host: HP Notebook 
Kernel: 4.19.0-16-amd64 
Packages: 1624 (dpkg) 
Shell: bash 5.0.3  
CPU: Intel Celeron N3060 (2) @ 2.480GHz 
GPU: Intel Atom/Celeron/Pentium Processor x5-E8000/J3xxx/N3xxx 
Memory: 1931MiB / 3803MiB 

# lsb_release -a
No LSB modules are available.
Distributor ID:	Debian
Description:	Debian GNU/Linux 10 (buster)
Release:	10
Codename:	buster

# uname -a
Linux deblugo 4.19.0-16-amd64 #1 SMP Debian 4.19.181-1 (2021-03-19) x86_64 GNU/Linux

# cat /etc/os-release
PRETTY_NAME="Debian GNU/Linux 10 (buster)"
NAME="Debian GNU/Linux"
VERSION_ID="10"
VERSION="10 (buster)"
VERSION_CODENAME=buster
ID=debian
HOME_URL="https://www.debian.org/ "
SUPPORT_URL="https://www.debian.org/support "
BUG_REPORT_URL="https://bugs.debian.org/ "

In Guix user can verify reproducibility of a package with guix challenge https://guix.gnu.org/manual/en/html_node/Invoking-guix-challenge.html . What is the Nix command for that?

When I search the Guix repository there are only versions of OpenJDK from 9 and up, but I want to install OpenJDK 8. Why is there no version 8?

$ guix search openjdk | recsel -p name,version
name: openjdk
version: 9.181

name: openjdk
version: 14.0

name: openjdk
version: 13.0

name: openjdk
version: 12.33

name: openjdk
version: 11.28

name: openjdk
version: 10.46

name: icedtea
version: 1.13.13

name: java-marlin-renderer
version: 0.9.4.2

name: icedtea
version: 3.7.0

name: icedtea
version: 2.6.13

On my system I have installed both thunar and thunar-volman:

$ guix package --list-installed | grep thunar
thunar-volman	0.9.5	out	/gnu/store/yp5xyyvvc59ivghszgyfwy3izd8jvqz0-thunar-volman-0.9.5
thunar	1.8.16	out	/gnu/store/4hy62y05g34hwrhlyk8vjhka47d5bmhq-thunar-1.8.16

But when volume management is disabled the advanced settings looks like follows: But gvfs is installed, at least according to the package manager.

$ guix package --list-installed | grep gvfs
gvfs	1.40.1	out	/gnu/store/qg1827ai9j5raypmdajw3k7cibsqdykz-gvfs-1.40.1

Is this a packaging issue or, as I think I've read before, is it that Guix hasn't solved how to deal with dynamic dependencies?

Given that Guix is a GNU project, it is no surprise that guix search spotify returns nothing. Spotify provide both Debian and Snap packages; is there a way to automatically convert one of these packages, or even the the package stream itself, to a Guix package definition? Even better, has someone alread done this?

On a Guix System the PID 1 is called Shepherd and is controlled using the command herd. I often find myself wanting to tab-complete commands beginning with herd, like:

$ sudo herd restart some

But lacking that, I can't even find a way to list all services known to (the) Shepherd. The help printout reads:

└$ sudo herd --help
herd [OPTIONS...] ACTION SERVICE [ARG...]
Apply ACTION (start, stop, status, etc.) on \
SERVICE with the ARGs.
  -s, --socket=FILE           send commands to FILE
      --help                  display this help and exit
      --usage                 display short usage message and exit
      --version               display version information and exit

and only seems to mention actions that should be taken on services. The info and man pages are both terse, merely listing the available actions.

herd [OPTION...] ACTION [SERVICE [ARG...]]

   It causes the ACTION of the SERVICE to be invoked.  When SERVICE is
omitted and ACTION is ‘status’ or ‘detailed-status’, the ‘root’ service
is used(1) (*note The root and unknown services::, for more information
on the ‘root’ service.)

   For each action, you should pass the appropriate ARGs.  Actions that
are available for every service are ‘start’, ‘stop’, ‘restart’,
‘status’, ‘enable’, ‘disable’, and ‘doc’.

I have this problem to install Guix System: After this, the screen turns black and nothing is displayed. I first downloaded the Guix System, extracted the iso with this command: "unxz" and then burned it to a 4.7GB DVD-R with "Brasero". graphics card: NVIDIA GT720

Say that you have a dynamically compiled binary for 32-bit Intel x86 that you want to run on your 64-bit x86-64 Guix System. The binary has been compiled on a different system. How would you go about running it? On Ubuntu the instructions to do the same thing can be found here: How to run 32-bit app in Ubuntu 64-bit?

Both seems to run just fine? Are there different systems for my normal user and if I run the command as root?

The [documentation](https://guix.gnu.org/manual/en/html_node/Invoking-guix-system.html) for guix system reconfigure says: > Build the operating system described in *file*, activate it, and switch to it. A cursory search didn't bring up any information on what or where that *file* is or how to generate it. As @muru commented, that page also says: > Once you have written an operating system declaration as seen in the previous section ...but that section is [Bootloader Configuration](https://guix.gnu.org/manual/en/html_node/Bootloader-Configuration.html) and likewise does not tell me where to find that file on my system. Since this step is part of a normal system upgrade (after runnnig guix pull as described in multiple places), I would've expected something a bit more obvious... but since I am a complete noob in GuixSD, I could easily be wrong.