IRC channel logs
2025-05-20.log
back to list of logs
<vagrantc>well, guix pull works again as of f2303dac31baac861a1f062dd9f7a84beba1b0a9 ... but there do seem to be a fair number of rebuilds <redacted>The AppImages I'm making with guix pack -f appimage -RR aren't running on Ubuntu. They exit immediately with no output unless I run them with GUIX_EXECUTION_ENGINE=proot. <redacted>Given that the AppImage tries and fails to mount() a tmpfs directory due to a permissions error, it looks like it's not falling back properly when user namespaces aren't supported. <apteryx>efraim: rust-1.82 breaks when updating curl to 8.11; would you happen to know how I can patch this? <efraim>apteryx: I'll take a look at it now, rebasing my core-packages tree <clone>Hi, when I try to run dev.zed.Zed in flatpak on guix system I get a portal error. I have flatpak, xdg-desktop-portal, xdg-desktop-portal-gnome, and xdg-dbus-proxy all installed in the system profile. If someone could try running it on their system to see if it's a configuration error on my part or has any ideas about what it could be that would be great. <clone>Portal call failed: Failed to start command: Failed to execute child process “/home/clone/.local/share/flatpak/app/dev.zed.Zed/x86_64/stable/{SNIP}/files/bin/zed” (No such file or directory) <clone>the file definitely DOES exist <abbe>most likely one of the library it's depending on is missing <ruther`>not a library. an interpreter. probably the dynamic linker <clone>so that would mean that the flatpak package is incorrectly made? <identity>clone: it is probably that Zed ships with something that needs dynamic system libraries, and Guix is not really all about it. you could probably fix that by using guix shell --container or poking at it with patchelf. or, you know, packaging Zed proper! <tae>When I try to pull but using Codeberg, I get: <tae>guix pull: error: Git error: could not find appropriate mechanism for credentials <tae>Does the --url flag require your Git installation to match with your Guix setup? <identity>tae: are you pulling directly from codeberg or from a local checkout? <tae>identity: I have an installation from Savannah, which times out, so I am trying to update it from Codeberg with --url=<codeberg-url> <identity>tae: the guix repo is not around yet, you want guix-mirror <apteryx>when is the etc/profile file of a profile sourced? e.g. if I work in 'guix shell' ? <apteryx>the shell launched in the new environment explicitly sources it? <ruther`>apteryx: I dont think the file is sourced when using guix shell. But I am not completely sure. I think guix shell reads the search paths directly and sets env accordingly <ruther`>(before actually launching the shell, and the env is then inherited by the process) <apteryx>ruther`: I think etc/profile is not even consulted when entering a 'guix shell'. It's the $profile/manifest which gets read, and then evaluate-search-paths is called with the search paths it contained onto profile. <nutcase>Hi guix. I have a user question in context of libvirtd/virt-manager. When I want to create a VM, I get "Unable to complete install: 'Failed to connect socket to '/var/run/libvirt/virtlogd-sock': No such file or directory'". What am I doing wrong / what could be missing? (I am not familiar with libvirtd and virt-manager) <nutcase>nevermind, I just rtfm and noticed, that I am missing the virtlog-service-type in my operating-system's configuration (I guess that's my problem) <decfed>redacted: could you please send over the exact invocation of guix pack and your Ubuntu version? I believe I tested -RR at least with package hello on an old Ubuntu in the past. <redacted>decfed: guix pack -f appimage -RR --entry-point=bin/luanti minetest <civodul>redacted: perhaps you’d be better off with a plain tarball, i.e., without ‘-f appimage’ <redacted>civodul: Yeah, there are workarounds. I could just use -f deb too if they're on Ubuntu. An AppImage would make it *somewhat* easier to do things with others: no need to explain what tar is, just have them run this file. <redacted>For context, someone else I'm trying to share programs with is on Ubuntu. After an AppImage I sent them didn't work, I've been testing in a VM. <civodul>redacted: oh i see; self-extraction is indeed an advantage of AppImage <civodul>if you manage to get more info, please consider emailing the details to bug-guix so we can keep track of it <krisbalintona>What wayland window managers do you all use, if any? I've been using hyprland but recently became curious whether there are any guile WMs around <identity>krisbalintona: there is an (abandoned?) X.org WM written in Guile, the package is in gnu/packages/guile-wm.scm; i personally used Sway until the papercuts started adding up and i switched to GNOME (though if GNOME was not around i would definitely be using Sway right now) <krisbalintona>identity: I see. Thanks for sharing. I'm curious though what about Gnome is appealing to you. Do you have some sort of tiling plugin for it? <identity>krisbalintona: i spend most of my time inside Emacs, so i do not really need tiling from my WM. in Sway, i have to set up multiple things manually (like night light tinting, for example), and even after that many things do not work (no idea how to auto-mount external media; fonts look Wrong for some reason, maybe scaling). with GNOME i get all of that, with (almost) none of the setup <krisbalintona>Yeah, that makes sense. Though I've always been bothered by how "in your face" the animations and certain other UI things are in Gnome. But I'm not sure if those are things you can change somehow <krisbalintona>identity: Maybe I'll give Gnome a fair shot in the future (especially since I mostly Emacs too lol) <identity>krisbalintona: you probably can change most of the things you want with extensions, iirc there is an extension called something like "Just Right" that lets you tweak/disable a bunch of things <identity>krisbalintona: it was "Just Perfection", gnome-shell-extension-just-perfection <hanker>How can I use `opensmtpd-extras` with the `opensmtpd-service-type`? It seems, by default, OpenSMTPD doesn't see the needed libexecs <cdegroot>krisbalintona: I use StumpWM (Common Lisp, not Guile, but Lisp so it counts ;-)) on X11 and I think someone "ported" it to Wayland, but you can pry my Xorg server from my cold dead hands so I won't know about the state of the project <identity>Mahogany is StumpWM but for Wayland, it is still in very early stage of development though <cdegroot>(I'm sorta converging on XFCE+StumpWM+some KDE bits, but I've been thinking I've settled on an X11 setup now for close to four decades now so not holding my breath. In any case, too much work to throw it away at some fancy almost-ready new thing like Wayland ;-)) <futurile>Q: does anyone other cbaines know how to restart the QA site/server/builder etc? <futurile>I assume they are off doing real life things - so someone else has to restart it <ieure>futurile, Host reverse-resolves to guix.aquilenet.fr, so guessing you'd need SSH/sudo access to that box. <ieure>futurile, Given that, you can probably figure it out, likely it's a system herd service. <ieure>I don't know if anyone else has that access, though. <civodul>futurile: hey! since a few days we cannot connect to bayfront (neither SSH nor serial access) <civodul>fellow hackers from Aquilenet are going to the data center in a few hours <poplin>hello, im a beginner with guix, but i'm exploring running services within containers. I've tried several configurations but in all cases I find the same problem: within the container, accessed by nsenter, herd does not seems to run (herd status hangs, and individual services statuses are shown as stopped). As (one of) a template of configuration file I used the file found in guix manual "Using-the-Configuration-System". Any help? <dariqq>poplin: I saw something recently with the shepherd-system-log when it cant open /proc/kmsg (e.g. in a container). Could you try removing that service from the %base-services (with something like modify-services) <futurile>civodul: ah got it, thanks for the update <bacs>is there somewhere I can grab an iso - getting a bit lost on mirrors atm <bacs>but struggling to find the equivalent on the mirrors <bacs>hmmm. None of the mirrors seem to mirror the website content. All have links pointing to the Bordeaux server which is down. <bacs>I'll wait for the server to come back up <bacs>JLambda: perfect - that's what I need <bacs>hey vagrantc - was just trying to spin up a VM to test on GUIX. Picked a bad time :-) <vagrantc>bacs: guix is a fun project with various reliability promises that are largely well met, but the infrastructure often has troubles keeping up with the ambitions of the developers <poplin>dariqq: thanks, I removed shepherd-system-log (and also nscd) but nothing changed <bacs>vagrantc: aye - it looks interesting. I just picked a bad moment ;-) Not every project has the resources of Debian. Perhaps we'll be able to help in some way with some of what we're working on. <bacs>vagrantc: very pleased with how clear the GUIX doco is <bacs>vagrantc: the Bordeaux server is hosted by Aquilenet, who are our ISP, so *normally* the GUIX server is as close to me as a DC server gets <rynn>ci.guix.gnu org is unreachable, and bourdeaux.guix.gnu.org is struggling for me. <dariqq>poplin: something interisting the shepherd log messages of the container (/var/log/messages)? I tried the " Using the Configuration System" example replacing the shepherd-syslog with the inetuils syslog-service and "only" got an exception in the loopback service <ieure>PotentialUser-47, Can you be more specific? <PotentialUser-47>does it run only in "compilation time" or does it keep running in the bg? <ieure>PotentialUser-47, The Guix daemon runs in the background. <ieure>I think this is what you're asking. <PotentialUser-47>can it move stuff and manage configuration files just like in "compilation time"? <ruther>PotentialUser-47: the daemon touches only the /gnu/store and /var/guix <ieure>PotentialUser-47, Your questions feel like an "XY problem." What's the thing you actually seek to understand? <ruther>PotentialUser-47: What is not understandable about that? <ieure>Which these questions are part of? <PotentialUser-47>im actually having problems with nix bcz it only "does stuff" when u swich systems, and im seeking solution :3 <ruther>PotentialUser-47: yes, Nix will build only when you ask it to. It is the same with Guix. It's a build system, nothing more, so why would it do anything when you don't want to build anything? <PotentialUser-47>i switch monitors a lot and it would be really nice if for example, i could do smth when a monitor changed, especially bcz my wm is very simple so i cant do logic on its config file <ruther>PotentialUser-47: that is not responsibility of Guix/Nix. You need a daemon that will watch for monitor changes and execute a script to switch when they do change <ruther>autorandr for x, kanshi for wayland <ruther>PotentialUser-47: what network server? <ruther>PotentialUser-47: guix doesn't have anything to do with that, so I don't understand the question <ruther>so you mean a bug in Guix System? I don't know of such a bug <ieure>PotentialUser-47, I daily drive Guix System and I haven't seen anything like that. <luca>Weirdly enough I've had a more stable experience with guix and network manager than on my previous distro. Even though for all intents and purposes it doesn't matter. A fresh start helps I suppose <identity>PotentialUser-47: you will find the answer to that in the #nonguix channel <poplin>dariqq: to my surprise, no /var/log/messages file (always using the "Using the Configuration System" example). I tried the same config in a different guix (virtual) machine (with a desktop-like configuration), and herd worked normally. That suggests that the first system (server like configuration) is missing something...