IRC channel logs

<PotentialUser-84>ieure I tried following you suggestion to have emacs be my shell

<PotentialUser-84>and every time I login then it logs out immediately

<PotentialUser-84>(shell (file-append emacs "/home/myusername/.guix-profile/emacs"))

<PotentialUser-84>this is what I added to the oper system user account in my config.scm ieure

<ieure>PotentialUser-84, That is wrong.

<ieure>PotentialUser-84, Follow the example I pointed you at.

<ieure>(shell (file-append emacs "/bin/emacs"))

<PotentialUser-84>but why not use the emacs of my user profile, if after all I'm just setting the shell for myself (and not other users)

<PotentialUser-84>besides, I suppose I must install emacs in the systemm profile before doing it... right? ieure

<ieure>No.

<PotentialUser-84>ok so no problem that /bin/emacs doesnt exist...

<ieure>It will exist.

<ieure>It is not /bin/emacs, it's the bin/emacs path within the Emacs store item.

<PotentialUser-84>ok I can believe that

<PotentialUser-84>but anyway why wouldnt my profile emacs not work...

<PotentialUser-84>I mean its just my own shell

<PotentialUser-84>so makes sensse to use my own emacs..

<ieure>PotentialUser-84, I think you could do that, but the form you're using to express it is wrong.

<ieure>PotentialUser-84, This approach is not idiomatic, since it undermines its declarative/immutable nature and risks breakage.

<ieure>ex. if a user uninstalls Emacs from their profile, they can no longer log in, thus cannot fix the problem.

<ieure>PotentialUser-84, When you have (file-append emacs "/some/string"), that appends "/some/string" to the store item path for emacs, which will be something like /gnu/store/c6qvzq6zf3apwmdh5q9k8204yfxx0cr1-emacs-30.2

<ieure>And /gnu/store/c6qvzq6zf3apwmdh5q9k8204yfxx0cr1-emacs-30.2/home/myusername/.guix-profile/emacs does not exist. So it will not work.

<FuncProgLinux>Why does our Go module system doesn't support modules? (I'm not asking as a "please do for me") I mean if there are technical limitations or if Go modules are not the guix way?

<Googulator>Hi all! I'm getting a strange "Permission denied" error in guix-manual while trying to bootstrap Guix on a new system. Unfortunately, in the exception, the failing call is cut off. Is there a way to get the builder to run with an increased $COLUMNS, or better yet, have it drop to a REPL when it hits the exception?

<ieure>FuncProgLinux, Because Guix is a volunteer project and nobody has volunteered to build it.

<FuncProgLinux>ieure: Ah I see, so it's unimplemented. I saw the comment in the go build system but it only said that the current one doesn't support modules. I wasn't able to find if it was on purpose or a wishlisted feature :)

<ieure>FuncProgLinux, https://codeberg.org/guix/guix/src/branch/master/guix/build/go-build-system.scm#L175-L178

<ieure>I believe go-build-system predates Go module support, and it wasn't added.

<ieure>There are likely technical challenges, the module system defaults to downloading everything and using a Google proxy to do it.

<PotentialUser-84>thanks ieure . by the way while I was sudo system reconfiguring with the bad syntax of the emacs in my profile, and erroneously disconnected from the internet, I got an error presented weirdly: many lines of backtrace and it looked like an error in the program; then after I connected to the internet everything went throu.... but that error with the

<PotentialUser-84>backtrace was bizarre... I wish I saved a screenshot, and I wonder if there is anything to report to maintainers about that

<FuncProgLinux>ieure: oof I didn't know the Google Proxy part :/

<ieure>FuncProgLinux, https://drewdevault.com/2022/05/25/Google-has-been-DDoSing-sourcehut.html

<ieure>FuncProgLinux, Fortunately, they decided to make the telemetry *built into the toolchain* opt-in. I have an extremely low opinion of Go.

<graywolf>for large part, supporting modules should be "easy", just download the .zip files from proxy (as fixed output derivations) and unpack them into expected path

<ieure>(I am also a paid Go programmer. But I do not think much of the language, or the ecosystem, or the people who designed it.)

<graywolf>but someone has to write it :)

<FuncProgLinux>ieure: I kinda like it because it reminds me of pascal...and because it got me my first job years ago lol. But I do agree that some of those design choices are quite...crappy.

<ieure>FuncProgLinux, := does not a Pascal make.

<FuncProgLinux>I know it's not the same xD

<FuncProgLinux>though I would have liked proc() as well for procedures

<FuncProgLinux>and guessing which end keyword needs the "." at the end.

<graywolf>geex the DeVault's blog post is from 2022... I feel old

<ieure>untrusem, 149.0.2-1 is out, do you want to update, or should I?

<ieure>graywolf, I have dotfiles older than some coworkers.

<FuncProgLinux>My editor almost doubles my age

<nox>> FuncProgLinux, https://drewdevault.com/2022/05/25/Google-has-been-DDoSing-sourcehut.html

<nox>very interesting, in spite of the source. this kind of abuse would probably illegal if this wasn't via the internet

<graywolf>Hm, is there a reason not to canonicalize paths in %patch-path?

<xjjk>hello! guix newbie here. I notice that on https://guix.gnu.org/manual/1.5.0/en/html_node/Application-Setup.html, there's an example package, but without instructions about where to put that definition or how to install it.

<bavier>xjjk: do you mean the example locales package?

<adanska>woah, something mustve caused a world rebuild

<adanska>or at least a really significant rebuild

<untrusem>ieure: I can update

<graywolf>Is something wrong with debbugs? I sent a patch yesterday, and still do not have back the created issue number

<kestrelwx>o/

<f1refly>I'm trying to run a python program that requires tktooltip. running `pip install tktooltip` does not seem to put it into my module path and I don't think there's a package for it. what should I do?

<trev>f1refly: guix shell python python-pip python-virtualenv, go into the venv and install?

<f1refly>it won't, for some reason, find tkinter in that case. I added python:tk to my guix shell line

<trev>a few days ago we were talking about the niri service. i finally remembered why it wasn't working for me...it's cause i use zsh and the service doesn't take that into consideration so my .zprofile is never loaded. i guess i could try to fix this upstream now

<graywolf>Hm, how can I create users in the image generated by `guix pack -f docker'?

<PotentialUser-84>ieure I managed to have emacs be the shell of my user. thanks for your instructions. now comes the problem of shutting down. I never worried about shuttingdown from bash but in emacs case I think its different: the shutting down properly is essential for housekeeping operations by emacs (delete session save file, delete .desktop files if dekstop is

<PotentialUser-84>being used to save sessions...) and Im not sure if shutting down from within emcas would allow it to do all this. In addition, there is the issue that I just cant manage: shell-command "shutdown now" just doesnt do anythingm and I wonder if this is because I need sudo privileges. But, again, for some reason unknown to me, emacs seems not design to

<PotentialUser-84>handle sudo commands (I wonder why).... so Im at loss

<f1refly>what do I have to do to have python:tk available to the python in my venv?

<f1refly>I tried installing it to my user profile but that did nothing

<ieure>f1refly, What exactly are you doing? `guix shell ...' and then running virtualenv inside that?

<ieure>PotentialUser-84, You're stressing over nothing, I halt reboot from inside Emacs all the time. I have no idea what you mean by "emacs seems not designed to handle sudo commands."

<f1refly>I have installed some dependencies with pip install to my local python profile and included the others plus some extra dependencies that were in the repo with guix shell

<f1refly>Because when I tried using a venv python was not able to find tkinter

<gnufairy>(II) config/udev: Adding input device PC Speaker (/dev/input/event5)

<gnufairy>(II) No input driver specified, ignoring this device.

<gnufairy>(II) This device may have been added with another device file.

<gnufairy>(II) config/udev: Adding input device ThinkPad Extra Buttons (/dev/input/event4)

<gnufairy>(**) ThinkPad Extra Buttons: Applying InputClass "evdev keyboard catchall"

<gnufairy>(**) ThinkPad Extra Buttons: Applying InputClass "libinput keyboard catchall"

<gnufairy>(**) ThinkPad Extra Buttons: Applying InputClass "evdev keyboard catchall"

<gnufairy>(II) Using input driver 'evdev' for 'ThinkPad Extra Buttons'

<gnufairy>(II) systemd-logind: got fd for /dev/input/event4 13:68 fd 34 paused 0

<ieure>gnufairy, Stop.

<gnufairy>(**) ThinkPad Extra Buttons: always reports core events

<gnufairy>(**) evdev: ThinkPad Extra Buttons: Device: "/dev/input/event4"

<gnufairy>(--) evdev: ThinkPad Extra Buttons: Vendor 0x17aa Product 0x5054

<gnufairy>(--) evdev: ThinkPad Extra Buttons: Found keys

<gnufairy>(II) evdev: ThinkPad Extra Buttons: Configuring as keyboard

<gnufairy>(**) Option "config_info" "udev:/sys/devices/platform/thinkpad_acpi/input/input5/event4"

<gnufairy>(II) XINPUT: Adding extended input device "ThinkPad Extra Buttons" (type: KEYBOARD, id 14)

<gnufairy>(**) Option "xkb_rules" "evdev"

<gnufairy>(**) Option "xkb_layout" "us"

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Activating service name='org.a11y.Bus' requested by ':1.0' (uid=983 pid=5437 comm="gnome-session-check-accelerated" label="unconfined")

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Successfully activated service 'org.a11y.Bus'

<gnufairy>gnome-session-binary[5432]: WARNING: Could not check if unit gnome-session-x11@gnome-login.target is active: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.systemd1 was not provided by any .service files

<gnufairy>_IceTransmkdir: Owner of /tmp/.ICE-unix should be set to root

<gnufairy>gnome-session-binary[5432]: WARNING: Desktop file /gnu/store/7gfbmfs8zc5x8ksqxhr17jrf1qsss45l-gdm-46.2/share/gdm/greeter/autostart/orca-autostart.desktop for application orca-autostart.desktop could not be parsed or references a missing TryExec binary

<gnufairy>libmutter-Message: 07:18:17.509: Running GNOME Shell (using mutter 46.9) as a X11 window and compositing manager

<gnufairy>(II) modeset(0): EDID vendor "LEN", prod id 16400

<gnufairy>(II) modeset(0): Printing DDC gathered Modelines:

<gnufairy>(II) modeset(0): Modeline "1280x800"x0.0 75.30 1280 1296 1344 1512 800 802 804 830 -hsync -vsync (49.8 kHz eP)

<gnufairy>(II) modeset(0): Modeline "1280x800"x0.0 59.26 1280 1328 1360 1440 800 803 809 823 -hsync -vsync (41.2 kHz e)

<gnufairy>(II) modeset(0): EDID vendor "LEN", prod id 16400

<gnufairy>(II) modeset(0): Printing DDC gathered Modelines:

<gnufairy>(II) modeset(0): Modeline "1280x800"x0.0 75.30 1280 1296 1344 1512 800 802 804 830 -hsync -vsync (49.8 kHz eP)

<gnufairy>(II) modeset(0): Modeline "1280x800"x0.0 59.26 1280 1328 1360 1440 800 803 809 823 -hsync -vsync (41.2 kHz e)

<gnufairy>** Message: 07:18:18.413: Obtained a high priority EGL context

<graywolf>gnufairy: can you use some pastebin service next time?

<gnufairy>** Message: 07:18:18.419: Obtained a high priority EGL context

<gnufairy>dbus-daemon[5447]: Activating service name='org.a11y.atspi.Registry' requested by ':1.1' (uid=983 pid=5456 comm="gnome-shell" label="unconfined")

<gnufairy>dbus-daemon[5447]: Successfully activated service 'org.a11y.atspi.Registry'

<gnufairy>SpiRegistry daemon is running with well-known name - org.a11y.atspi.Registry

<gnufairy>GNOME Shell-Message: 07:18:19.397: Unset XDG_SESSION_ID, getCurrentSessionProxy() called outside a user session. Asking logind directly.

<gnufairy>GNOME Shell-Message: 07:18:19.398: Will monitor session c12

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Activating service name='org.gnome.Shell.Screencast' requested by ':1.5' (uid=983 pid=5456 comm="gnome-shell" label="unconfined")

<gnufairy>The XKEYBOARD keymap compiler (xkbcomp) reports:

<gnufairy>> Warning: Could not resolve keysym XF86RefreshRateToggle

<gnufairy>> Warning: Could not resolve keysym XF86Accessibility

<gnufairy>> Warning: Could not resolve keysym XF86DoNotDisturb

<gnufairy>Errors from xkbcomp are not fatal to the X server

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Activating service name='org.gnome.Shell.Notifications' requested by ':1.5' (uid=983 pid=5456 comm="gnome-shell" label="unconfined")

<gnufairy>Window manager warning: Failed to parse saved session file: Failed to open file “/var/lib/gdm/.config/mutter/sessions/10b7e9618374ed25e7177565789736481100000054320000.ms”: No such file or directory

<gnufairy>portal is not running: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.portal.IBus was not provided by any .service files

<f1refly>hehe

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Successfully activated service 'org.gnome.Shell.Notifications'

<gnufairy>(gsd-sharing:5504): sharing-plugin-WARNING **: 07:18:20.297: Failed to StopUnit service: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.systemd1 was not provided by any .service files

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Activating service name='org.gnome.ScreenSaver' requested by ':1.23' (uid=983 pid=5556 comm="gsd-usb-protection" label="unconfined")

<gnufairy>(gsd-xsettings:5511): xsettings-plugin-WARNING **: 07:18:20.402: Failed to get current UI scaling factor: GDBus.Error:org.freedesktop.DBus.Error.NameHasNoOwner: Name "org.gnome.Mutter.X11" does not exist

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Successfully activated service 'org.gnome.Shell.Screencast'

<gnufairy>gnome-session-binary[5432]: CODE_FILE=../gnome-session-46.0/gnome-session/gsm-manager.c:CODE_LINE=1404:start_phase:Entering running state

<gnufairy>dbus-daemon[5431]: [session uid=983 pid=5431 pidfd=5] Successfully activated service 'org.gnome.ScreenSaver'

<gnufairy>(gsd-power:5560): power-plugin-CRITICAL **: 07:18:21.098: update_mutter_backlight: assertion 'backlights != NULL' failed

<gnufairy>(gsd-usb-protection:5556): usb-protection-plugin-WARNING **: 07:18:21.128: Failed to fetch USBGuard parameters: GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name org.usbguard1 was not provided by any .service files

<gnufairy>(gsd-media-keys:5534): media-keys-plugin-WARNING **: 07:18:21.286: Failed to grab accelerator for keybinding settings:hibernate

<gnufairy>(gsd-media-keys:5534): media-keys-plugin-WARNING **: 07:18:21.286: Failed to grab accelerator for keybinding settings:playback-repeat

<gnufairy>(gsd-media-keys:5534): media-keys-plugin-WARNING **: 07:18:21.286: Failed to grab accelerator for keybinding settings:rotate-video-lock

<gnufairy>(gsd-media-keys:5534): media-keys-plugin-WARNING **: 07:18:21.286: Failed to grab accelerator for keybinding settings:playback-random

<gnufairy>(gsd-media-keys:5534): media-keys-plugin-WARNING **: 07:18:21.286: Failed to grab accelerator for keybinding settings:screensaver

<gnufairy>** Message: 07:18:21.521: Failed to start camera monitor

<gnufairy>GNOME Shell-Message: 07:18:21.597: Failed to get PolKit permission: Polkit.Error: GDBus.Error:org.freedesktop.PolicyKit1.Error.Failed: Action org.freedesktop.bolt.enroll is not registered

<gnufairy>GNOME Shell-Message: 07:18:21.654: Error calling StartServiceByName for org.freedesktop.GeoClue2: Launch helper exited with unknown return code 253

<gnufairy>GNOME Shell-Message: 07:18:21.822: Registering session with GDM

<gnufairy>** (gnome-shell:5456): WARNING **: 07:18:26.319: Error checking authorization for action id org.freedesktop.bolt.enroll: GDBus.Error:org.freedesktop.PolicyKit1.Error.Failed: Action org.freedesktop.bolt.enroll is not registered

<gnufairy>** (gnome-shell:5456): WARNING **: 07:18:26.342: Error checking authorization for action id org.freedesktop.bolt.enroll: GDBus.Error:org.freedesktop.PolicyKit1.Error.Failed: Action org.freedesktop.bolt.enroll is not registered

<gnufairy>(**) Option "fd" "29"

<gnufairy>(II) event12 - Contour Design Perfit Optical: device removed

<gnufairy>(**) Option "fd" "33"

<gnufairy>(II) event10 - TPPS/2 IBM TrackPoint: device removed

<gnufairy>(II) AIGLX: Suspending AIGLX clients for VT switch

<gnufairy>(EE) modeset(0): failed to set mode: Permission denied

<gnufairy>(II) systemd-logind: got pause for 13:66

<gnufairy>(II) systemd-logind: got pause for 13:76

<gnufairy>(II) systemd-logind: got pause for 13:77

<gnufairy>(II) systemd-logind: got pause for 13:78

<gnufairy>(II) systemd-logind: got pause for 13:64

<gnufairy>(II) systemd-logind: got pause for 13:75

<gnufairy>(II) systemd-logind: got pause for 13:74

<gnufairy>(II) systemd-logind: got pause for 13:67

<gnufairy>(II) systemd-logind: got pause for 13:68

<gnufairy>XIO: fatal IO error 4 (Interrupted system call) on X server ":0"

<gnufairy> after 2381 requests (2381 known processed) with 7 events remaining.

<gnufairy>(gsd-sharing:5504): GLib-GIO-WARNING **: 07:18:27.125: Error releasing name org.gnome.SettingsDaemon.Sharing: The connection is closed

<gnufairy>(gsd-rfkill:5516): GLib-GIO-WARNING **: 07:18:27.127: Error releasing name org.gnome.SettingsDaemon.Rfkill: The connection is closed

<gnufairy>(gsd-smartcard:5521): GLib-GIO-WARNING **: 07:18:27.128: Error releasing name org.gnome.SettingsDaemon.Smartcard: The connection is closed

<gnufairy>(gsd-housekeeping:5546): GLib-GIO-WARNING **: 07:18:27.136: Error releasing name org.gnome.SettingsDaemon.Housekeeping: The connection is closed

<gnufairy>(gsd-wwan:5531): GLib-GIO-WARNING **: 07:18:27.136: Error releasing name org.gnome.SettingsDaemon.Wwan: The connection is closed

<gnufairy>(gsd-a11y-settings:5543): GLib-GIO-WARNING **: 07:18:27.138: Error releasing name org.gnome.SettingsDaemon.A11ySettings: The connection is closed

<gnufairy>(gsd-usb-protection:5556): GLib-GIO-WARNING **: 07:18:27.138: Error releasing name org.gnome.SettingsDaemon.UsbProtection: The connection is closed

<gnufairy>(gsd-sound:5538): GLib-GIO-WARNING **: 07:18:27.138: Error releasing name org.gnome.SettingsDaemon.Sound: The connection is closed

<gnufairy>(II) evdev: ThinkPad Extra Buttons: Close

<ieure>Can someone please mute gnufairy?

<gnufairy>(II) UnloadModule: "evdev"

<gnufairy>(II) systemd-logind: releasing fd for 13:68

<ieure>this sucks

<gnufairy>(II) UnloadModule: "libinput"

<gnufairy>(II) systemd-logind: releasing fd for 13:74

<gnufairy>(II) evdev: AT Translated Set 2 keyboard: Close

<gnufairy>(II) UnloadModule: "evdev"

<gnufairy>(II) systemd-logind: releasing fd for 13:67

<gnufairy>(II) UnloadModule: "libinput"

<graywolf>there even wasn't any question...

<orahcio>Hello. Yesterday my SSD ran out of available space, so I removed some cache paths and the swapfile to make my Guix system usable again. However, now every guix command is very slow to show any output. What could be causing this?

<ieure>orahcio, No idea. Run it under strace?

<orahcio>Now, I am trying to run `guix gc --verify=contents`

<orahcio>ieure: How can I run it under strace?

<ieure>orahcio, `guix shell strace -- sudo strace -f -o guix.out guix ...' -- then examine `guix.out'

<orahcio>ieure: ok, I will try this

<orahcio>ieure: But I have a problem, guix comand is very slow now, so also guix shell

<orahcio>maybe I need to wait the resulto of `guix gc --verify=contents,repair`?

<ieure>orahcio, I really don't know, then. What command(s) are you running that are slow?

<ieure>If you just run `guix shell' with no package arguments, is that slow? Is `guix --help' slow? `guix repl'?

<orahcio>ieure: the first one, sudo guix system reconfigure system.scm; guix shell strace the second one

<orahcio>ieure: guix --help and guix repl work fine

<ieure>orahcio, Okay, I'd try restarting the Guix daemon. `sudo herd restart guix-daemon'

<orahcio>ieure: `doas herd restart guix-daemon` is also very slow to show something

<orahcio>ieure: I am not try to restart my system with a older generation, maybe I need to try this before anything

<orahcio>Even the reboot command is slow, I think something with herd is not good

<squid64>Is dotnet proprietary? Can't find it in guix. But I can find mono.

<Googulator>AFAIK the new dotnet (non-Framework) is free, but it's Microsoft

<Googulator>and that's probably a good enough reason for non-inclusion in Guix

<squid64>Does guix keep all software made by corporations out of the repos no matter if they're free or not?

<ieure>No.

<ieure>But corpo stuff is often difficult to bootstrap, because they don't care about bootstrapping.

<squid64>I see

<Googulator>MS actually has a surprisingly good track record in this regard - Azure Linux is currently the only bootstrappable distribution with a mainstream (binary-oriented) package manager

<squid64>The mono project page tells people to download dotnet core and I am guessing they won't make new mono versions.

<Googulator>It's not perfect, because a precompiled JDK is a bootstrap seed

<Googulator>But massively better than Fedora or Ubuntu, which assume that you're building RPM/DEB packages on an already installed distro of the same type

<squid64>But I guess I can just use another language. I been learning Rust and Haskell these days. C# was just the first language I learned.

<squid64>The future of rust looks kinda worrying though.

<Googulator>And let's not talk about Haskell's bootstrap.

<squid64>Well at least haskell is on guix

<Googulator>It is, with a bootstrap chain rooted in a massive binary package.

<Guest20>Hey, can somebody give or link me to a basic/kick-start repository? I'm completely lost, I'm coming from NixOS, but am unsure about the basics.

<Guest20>I've found some configs, but they are too complicated or don't fit my needs.

<Guest20>I'm looking for a good practice folder structure with system/home separated.

<Guest20>It should be one folder, not separated to /etc, .config etc..

<Guest20>And I'd like to know, how to pin my channels to a .lock file.

<Guest20>Thank you and have a great day!

<orahcio_>ieure: I am in a older generation now, everything seems okay now. The guix system reconfigure works fine now. Thanks for the help

<ieure>Guest20, I think your questions are a better fit for the help-guix mailing list.

<Guest20>Alright, thank you!

<Googulator>I did some work a while back to connect the GHC 4.08 bootstrap chain to the GHC 7.x one, but never compleded it - and bootstrapping 4.08 from anything other than a binary or transpiled C code is a completely unsolved problem.

<Googulator>Even worse, GHC has been resisting a public bootstrap ever since it was first released

<Googulator>The first public release, v0.26, was also the one that ripped out the ability to be bootstrapped using a different Haskell compiler, an ability that was (according to 0.26's readme) present in earlier, non-public versions.

<scholastic>Hi folks

<squid64>I don't know much about this kind of thing but it does sound like a pain

<squid64>So I should be thankful ghc is even there

<scholastic>Question about guix. On Debian: `sudo shutdown -h +5` = perfect 5min shutdown timer.What is the EXACT Guix System equivalent command?

<squid64>I like haskell because it's easy to read and it's not like the other functional languages.

<cdegroot>scholastic: `sleep 300; sudo halt` :-)

<cdegroot>I don't think Shepherd does delayed shutdowns.

<cdegroot>(checked the docs: nope)

<scholastic>It's possible to do that in another way?

<cdegroot>There's always more than one way to skin a cat, but if you're looking for a built-in delay in Shepherd, the answer is no.

<squid64>You could always write a simple bash script that does just that, should be pretty easy.

<cdegroot>Should be simple enough to wrap it in a shell script and call it `shutdown` and put it in the path before Shepherd's shutdown.

<cdegroot>And if the itch is strong enough, create a patch to Shepherd. It's just some simple Guile code :)

<squid64>Is Shepherd all written in scheme alone?

<cdegroot>I noticed one little C module, but by approximation, "yes"

<scholastic> Brilliant idea! I'll write that script

<cdegroot>(You probably can also create a service, "delayed-shutdown", in your Guix system config)

<squid64>I see. That's pretty cool. I don't know much about scheme but I know emacs lisp.

<cdegroot>Note though, that sleep + halt is, I think, what pretty much happens these days. The time that `wall` blasted warnings out every minute is gone (again, "I think", but that was the only defining difference between "sleep + halt" and "shutdown -h +5")

<squid64>I installed Guix System because I wanted a libre distro, NixOS can be like that but it does allow non-free software so I wanted something different.

<squid64>I think Guix home is pretty cool too and nixos doesn't have something like that out of the box.

<bjc>nix has home-environment

<squid64>I guess

<squid64>But personally I like the way guix does it better

<kestrelwx>PotentialUser-84: If you have elogind you can shutdown with `loginctl poweroff` as your user.

<squid64>I might learn scheme as well eventually, at least it has more use than just the configuration.

<ieure>kestrelwx, There was a big convo about this in #emacs earlier.

<trev>is the easiest way to test a service by just reconfiguring through ./pre-inst-env ?

<trev>honestly cannot remember how i was testing services before. memory like a blackhole

<bjc>i use ‘./pre-inst-env guix system vm …’

<bjc>maybe there's a better way

<gabber>trev: i think both `guix system vm` and `guix system container` should work, it depends on what service exactly you want to test.

<kestrelwx>This one boots into OBS with the niri service. https://codeberg.org/guix/guix/pulls/7411#issuecomment-12551115

<kestrelwx>You'd need to add `zsh` as the user's shell.

<kestrelwx>I guess you could also just use your normal config anyway.

<trev>kestrelwx: funny, that's the service i want to test. it doesn't work with zsh as is

<gnufairy> https://paste.centos.org/view/a0d04cfd

<gnufairy>Upon updating my system, my x200 was no longer able to open the display manager gdm. I have currently rolled back my system to its previous state.

<gnufairy>Up there is my log, here is my /etc/config.scm.

<gnufairy> https://paste.centos.org/view/22a93047

<gnufairy>Please help.

<ieure>gnufairy, Did you change anything in your config between generations?

<ieure>gnufairy, And is the issue that GDM doesn't start? Or does GDM start, but you're not able to log in?

<gnufairy>I did not change anything besides attempt to install xorg-server and cvt to my user profile while trying to set up my external monitor. However this was the only change. GDM flashes briefly but then the screen just keeps flashing in the tty. It loads to the point where I can see the login prompt in the tty. Then it tries to loads gdm and fails in the manner I described. Switching to a tty works fine.

<ieure>gnufairy, How did you install xorg-server? In your user profile?

<gnufairy>guix install xorg-server cvt

<ieure>Uninstall xorg-server.

<gnufairy>okay

<ieure>gnufairy, The direct cause for your issue is line 326: (EE) modeset(0): failed to set mode: Permission denied

<ieure>gnufairy, And this is probably because xorg needs setuid, and nothing installed in your profile can be suid.

<gnufairy>I see. But I can fully use this 1st generation before doing the update and reconfiguring my system. However, I will now reattempt logging into the updated generation with xorg removed from my profile.

<Rutherther>does it really need setuid? I don't think it does, you can use the home startx command service without any support from your system for xorg

<ieure>gnufairy, I don't see a cvt package in Guix, is that something you packaged?

<ieure>Rutherther, Yeah, hmm, changing system generations wouldn't affect this either. But I think xorg-server shouldn't be in a user profile regardless.

<ieure>Be interesting to diff the configurations for those two system generations.

<Rutherther>yeah, definitely not and I would also think it's causing this. But not for the reason it would need setuid, but rather something like wrong modules being picked up (you need to give Xorg proper arguments to find modues) or incompatible versions

<Rutherther>also I do not really see a reason to install it, what led you to it, gnufairy?

<ieure>They parted, presumably to reboot.

<Rutherther>if you're interested in starting xorg without a login manager it is definitely possible, but not by installing xorg to profile

<Rutherther>ieure: yeah, maybe they will read the logs

<trev>kestrelwx: can i ask you if you know if we can get rid of `bash -c` in the niri service and just exec with `(list #$(file-append niri "/bin/niri") "--session")` instead?

<gnufairy>Sadly removing xorg from my profile did not fix the issue and it persists as previously stated.

<bjc>you can compare the working config in /run/{current,booted}-system/configuration.scm with the one that doesn't work

<bjc>if rollback works, then whatever's wrong will be in the diff

<ieure>Yeah, if you want to diff those and put it in a pastebin, that'd be helpful.

<Rutherther>gnufairy: also can you try logging in as a different user? Sometimes the .config files in your home can cause this, ie. on updates of a DE

<ieure>Rutherther, Symptom happens before login.

<Rutherther>Oh! so it can't be something guix installed then either

<gnufairy>Where would I find config.scm for another generation?

<gnufairy>Could it be that the updated gdm is Wayland first which is playing havoc on my poor X200 GPU. Does anyone have a similar issue?

<kestrelwx>trev: I understand `bash -l` is there to source `~/.bash_profile` and `~/.bashrc`.

<Rutherther>gnufairy: use "guix system list-generations" to see where their configuration file is stored (if it is, but it is by default)

<trev>ah..yeah...i wonder if i can just call $SHELL

<Air4x>Hi, I was trying to compile guix from source to then try to contribute a bunch of packages, I followed the guide in the manual at "(guix) Contributing" section. But 'make check' fails with "make: *** [Makefile:7657: check] Error 2" (here the entire output: https://paste.debian.net/hidden/ed81f974). Is it a problem or I can go and hack around at my heart content?

<Rutherther>Air4x: the link is not working. It is a problem, but not your problem, the test suite is going to fail.

<Air4x>Rutherther: the link seems to work from my side, anyway if the test suite failing is not a problem I'm going to have fun playing with the code :)

<Rutherther>right, it works, I am not sure what happened before

<Air4x>cosmic rays, probably

<Rutherther>see https://codeberg.org/guix/guix/issues/4225

<Rutherther>feel free to fix it :)

<ham5urg>When I try to boot a guix.iso (build yesterday) on a new Dell XPS Ultra x7 with Intel Arc GPU, the display blanks out. I can not boot. Any hints?

<ham5urg>With nomodeset I got the output again: kernel panic Attempted to kill init, exitcode 0x00000100

<ham5urg>I see a "segfault, likely on cpu 6(core 10, socket 0)"

<ham5urg>the same error with nomodeset and vga=normal in grub.

<lullaby>Is there a straightforward way to dim the screen on laptop running Guix system? Running sway/wayland. Tried ddcutil but apparently that only works for external monitors.

<robgssp>try `light`

<ieure>lullaby, Modern backlight control on Linux uses the sysfs interface. /sys/class/backlight/*/brightness

<ieure>lullaby, Rutherer mentioned a CLI tool they liked yesterday; I live in Emacs, so I wrote emacs-blight.

<Rutherther>it's "brillo"

<ieure>I wish the sysfs interface supported non-laptop displays, many monitors support adjusting brightness via DDC, but it's much less accessible than using the sysfs interface.

<ieure>Maybe ddcci-driver-linux will do this.

<robgssp>Hey, question about fixed-output derivations. I've got a FOD test-case that just calls curl that is able to reach external IPs, but fails to resolve DNS. Does anyone know what's needed to get DNS working in the builder env?

<robgssp>here's the test case: https://pastebin.com/8E1QViz5

<ieure>robgssp, You cannot access the network from inside the build environment, so DNS will never work.

<Rutherther>robgssp: are you on a foreign distro? Is /etc/resolv.conf symlink on your system?

<Rutherther>ieure: sure you can from FOD's, otherwise there would be no sources fetched ever :)

<ieure>Hmmm.

<robgssp>Rutherther: ah yeah, sure is

<robgssp>it's a nixos box

<robgssp>I think the builtin fetchers still work though?

<Rutherther>robgssp: okay. There is a bug now, it should be fixed in a matter of days (just upgrade of the guix package now). For the time being ensure /etc/resolv.conf is not a symlink and I think it could work. If not, let's debug further then

<robgssp>gotcha, I'll give it a shot

<Rutherther> https://codeberg.org/guix/guix/pulls/6959 fyi

<robgssp>yep, that worked! thanks.

<Rutherther>robgssp: how do you update the guix daemon, is it through nixpkgs? Just wondering how you will get the bug fix. For imperative distros after this https://codeberg.org/guix/guix/pulls/7689 gets merged users will just do "sudo guix pull"

<robgssp>yeah, it's through nixpkgs

<lullaby>robgssp, ieure, both `bright` and sysfs work for me. Thank you!

<robgssp>I can point guix's nix build at HEAD for the time being

<dajole>Has anyone here successfully set up dropbear to unlock an encrypted system via ssh? There is an old mailing list thread about it (https://mail.gnu.org/archive/html/help-guix/2022-09/msg00135.html), but it's a few years old and I wonder if things might have changed. There doesn't seem to be anything in the docks, aside from the dropbear-service-type docs.

<ieure>dajole, dropbear will not work on Guix with an encrypted root FS, since it has to be unlocked by the bootloader before the initrd is loaded.

<dajole>hm...that is unfortunate

<dajole>Do people simply not use full disk encryption for servers, or is there another solution?

<ieure>dajole, I use Debian for my servers.

<ieure>Well... mostly. I have a Guix build server. It's not in the same building I live in, but it's close enough that I go to where it is if I need to reboot. I've pondered a network KVM or serial console to get around that.

<ieure>But my other stuff runs Debian, because it either needs better ZFS support than Guix currently has, or because it's a VPC and I don't want to run unsupported weirdware in that environment.

<dajole>yeah, I may have to use a regular distro, too. I'm not really a sysadmin/devops person and just really liked the idea of having all the config declarative and visible in a single file. I suppose ansible might get part of the way there.

<ieure>dajole, My Debian setup uses Ansible. It's pretty terrible compared to Guix.

<ieure>Ansible is the least worst of the tooling I've used, which is Ansible, Chef, and Puppet.

<ieure>But I do not like it at all.

<dajole>:D thanks for the encouragement

<ham5urg>does guix.iso boot from a 'ventoy' usb-stick?

<ieure>ham5urg, Yes.

<ieure>dajole, The problem with layering some declarative tooling on top of an imperative platform is that the imperative stuff still exists, and any changes you make don't get captured, so actual state vs. declared state almost inevitably drifts. Guix only offering the declarative route means this failure mode only exists in extreme corner cases.

<ieure>I've seen (or caused/contributed) to this failure mode in every single tool-on-top-of-OS management system.

<ham5urg>ieure, if I boot via ventoy->boot in grub2 mode; where I add vga=normal nomodeset, I run into a kernel panik: /ventoy/init:65:1 unterminated '#! ... !#' comment

<dajole>yeah, that's what I love about Guix. Even though my knowledge is still rather superficial it's 'easier' in some ways than other distros because I can actually trace what changes I've made.

<ieure>ham5urg, No idea about that. I put the .iso on my Ventoy stick, boot it in normal mode, it works.

<ham5urg>ieure, I tried normal boot mode with vga=normal nomodeset and I see some kernel messages, the last is from sof-audio-pci-intel-ptl : error: sof_probe_work failed err: -19

<ieure>ham5urg, You have pretty new hardware, is it fully supported by Linux?

<ieure>Do other distro installers boot on it?

<ham5urg>ieure, got it today, have not tried any other distro

<ham5urg>I ordered intel cpu/gpu in hope of full linux support

<nutcase>ham5urg: it may be the case that your audio device needs non-free firmware, which is not available in Guix itself, but available from another channel.

<ham5urg>nutcase, yes, I was hoping to install without a custom nonguix iso and pull all firmware later into the system via a nonguix channel.

<ham5urg>Idk why a missing firmware stop the whole boot process.

<ham5urg>I tried vga=normal nomodeset intel_ish_ipc.disable=1 snd_sof_pci.disable=1 but no success

<nutcase>ham5urg: maybe, there is another reason for not being able to boot? Additionally, maybe you can disable the device in your bios.

<gnucode>hello, does anybody know if guix can run on Debian GNU Hurd ?

<gnucode>I think that's how guix was first ported to the Hurd...so it should be possible...

<gnucode>janneke: I'm thinking about trying to run guix on Debian GNU/Hurd. The advantage of this, is that I could try to set up Guix to use unionfs...though I'm not much of a coder. any tips for me in trying to install guix on Debian GNU/Hurd ?

<ham5urg>ieure, nutcase, I bootet artix. It went through

IRC channel logs

2026-04-08.log