IRC channel logs


back to list of logs

<eikcaz>Anyone using a guix machine as a wifi router? Seems like a pretty simple deal to plug in the ethernet dirrectly from my modem and turn on "hotspot" in gnome-control-center. Any reason I shouldn't do this?
<eikcaz>Any service I run on that machine gets port forwarded anyway, so it seems like it wouldn't increase my attack surface
<the_tubular>You mean an access point ?
<eikcaz>I mean both. Hotspots do take on the role of routers, as well as access points
<eikcaz>When you connect to a hotspot, it assigns a local IP address, and translates between local and IP addresses, so it is acting as a router, but since it allows wirless connections, it is also a wireless access point
<eikcaz>s/and IP/and global IP/
<brendn>Hi I noticed when using home-gpg-agent-service-type that when when stopping the shepherd service with herd stop gpg-agent, and then starting it again that the gpg-agent process would not respawn. Is that the case for anyone else? Is that intended?
<eikcaz>I'm not familiar with that particular service, but I'm guessing that the service does something other than launch gpg-agent. What happens when you try to use a GPG key?
<eikcaz>In my experience, gpg-agent starts automatically when I try to decrypt a password, which uses gpg-agent behind the scenes
<brendn>eikcaz: It looks like the service does do a few things other than run gpg-agent. It also seems to set some environment variables and the gpg-agent.conf file. But it does extend the home-shepherd-service-type to run gpg-agent under shepherd. When first logging in I can see the gpg-agent process running and it is listed under herd status. I think it might only be launching gpg-agent when ssh support is enabled, which I do have enabled.
<brendn>But after stopping the service and restarting it, the gpg-agent process doesn't launch again
<eikcaz>Hmm. Is it useable? Like, does it refuse to start even when you try to use a GPG key?
<freakingpenguin>Hey all. What would the mcron guile syntax for "run a job every monday" look like? I looked at (info "(mcron) Every second Sunday") but that example is a bit more complicated then what I need.
<jakef>freakingpenguin: not sure, but it might be worth checking out an alternative to mcron using shepherd timers
<peanuts>"Re: Shepherd timers"
<freakingpenguin>Looks interesting but I think I'll wait until that makes its way into shepherd stable haha.
<rebiw>Does home-dotfiles-service-type require subsequent reconfigure steps?.
<rebiw>After changing some dotfile I mean.
<freakingpenguin>rebiw: As in you update some dotfiles in your config and want to update the files in $HOME? I think so.
<rebiw>Yes, I'm currently using home-files-service-type  I don't know if I should move some to the dotfiles service.
<apteryx>yay, I've found why chat messages couldn't be displayed in linphone
<apteryx>debian investigated and resolved that in 2021
<PuercoPop>Question about, is there something guix specific about this bug or is it something to be reported upstream? I am experiencing the but in the same scenario as report 70211, jsonrpc and eglot.
<peanuts>"Emacs packages that override built-in features ignored when compiled"
<PuercoPop>Is there a reason why the built-in jsonrpc package is compiled to native could but not the newer jsonrpc installed as part of a separate derivation?
<elpogo>when i start a "guix repl" and set a breakpoint with ",break (@@ (guix ci) find-latest-commit-with-substitutes)" , the repl doesn't actually hit the breakpoint when i invoke the procedure with ((@@ (guix ci) find-latest-commit-with-substitutes) "") ; anyone know what i'm doing wrong?
<elpogo>the procedure does run and returns the commit without ever entering the interactive debugger
<metsomedog>Anyone else unable to download the iso from ?
<metsomedog>getting a json {"error": "Could not find the requested build product."} on the url
<metsomedog>or rather, anyone *not* getting such an error?
<metsomedog>I have an old iso from december locally, but would rather use latest for a new install
<raghavgururajan>apteryx: IIRC, it was something to do with SIMPLE support.
<civodul>Hello Guix!
<metsomedog>civodul: hello :)
<metsomedog>civodul: as I wrote about an hour ago, there seems to be a problem with the "latest" iso from, just for your info
<civodul>metsomedog: hi, thanks!
<civodul>i think this was reported before
<civodul>haven’t had a chance to look into it :-/
<metsomedog>civodul: no worries
<metsomedog>I managed to build an iso with the channel that shall not be named
<civodul>oh good
<jeko>Yoo Guixters!
<civodul>hey jeko!
<Kabouik>What would be the easiest way to provide a reproducible framework for R analyses along with a research manuscript. I'm using Guix system, but am not very familiar with how manifests work (I didn't use one for the analyses). I have used the guix-cran channel. May I ping you for that topic rekado?
<Kabouik>That was a question by the way, just forgot the question mark. :<
<mehrad>Does anyone know why on foreign distros, there is a file named `/etc/profile.d/` rather than `/etc/profile.d/`? I'm trying to make KDE Plasma to recognize the software I have installed via Guix
<mehrad>Apparently the file I
<mehrad>Apparently the file I'm at is the correct file:
<peanuts>"[PATCH] home: services: environment-variables: Fix XDG base directories."
<mehrad>peanuts: are you by any chance the peanutbutterandcrackers on github who wrote this
<peanuts>mehrad: Hi, for comments please contact my maintainers at
<peanuts>"Guix Desktop-Environment Integration Guide ? GitHub"
<mehrad>I see, this is a bot. Well...
<civodul>Kabouik: hi! i think rekado is unavailable these days, but you can definitely ask questions about manifests & co. here
<civodul>to get started with manifests, you can take a look at
<peanuts>"Writing Manifests (GNU Guix Reference Manual)"
<mehrad>civodul: Thanks, but how a manifest is related to my question?
<civodul>mehrad: i was replying to Kabouik
<mehrad>I have installed Guix on a foreign distro and have `guix install` a software. Now I'm looking for a way to integrate it with my DE (in my case KDE Plasma). I read on the Github gist which I liked to above, that the `/etc/profile.d/` is where the magic happens.
<mehrad>civodul: Aaah, sorry, I apparently joined after that and have not seen those. Sorry :)
<civodul>mehrad: the file is named so that it’s sourced last
<civodul>that’s the usual convention
<civodul>but then, if you ‘guix install’ things *and* have that file sourced, it should work
<civodul>maybe see also
<peanuts>"Application Setup (GNU Guix Reference Manual)"
<raghavgururajan>sneek: later tell apteryx: IIRC, it was something to do with SIMPLE support. Is debian using same version as us?
<civodul>metsomedog: images at are back (except the Hurd one, but soon)
<civodul>i believe i’ve fixed the underlying issue
<civodul>(just explained it on guix-devel)
<mehrad>civodul: Thanks for the explanation. I was thinking the same way since the zzz convention is also followed in R (another GNU project). The reality is that it doesn't work. To be fair, I have not restarted my machine, but I have opened new terminals to force running profile shells.
<civodul>mehrad: right, it’ll only work once this has been sourced
<civodul>this typically means you have to log out and re-log into your graphical environment after your first ‘guix install’
<mehrad>civodul: I thought that too, so I sourced it `source /etc/profile.d/` and then tried to run the software, and still got error that it is not found. Even `which` command says that it is not found
<mehrad>I was kinda hoping that Guix as a package manager integrates itself to the host at least on the most popular/major DEs like Gnome and KDE
<metsomedog>what do you mean by integrates btw? that you can find the applications installed by guix in the "launcher" of kde/gnome?
<mehrad>metsomedog: ideally, yes. But at this point they are not even in the path although these two are the first two entries in $PATH: /home/mehrad/.config/guix/current/bin:/home/mehrad/.config/guix/current/bin
<metsomedog>I know this is not helpful, but for me the guix applications show up in the (xfce) launcher on a foreign distro, likely due to the file you mentioned in /etc/profile.d/ ; have you tried to check the paths of those environement variables (in manually to see that the files exist?
<mehrad>metsomedog: Thanks for the suggestion. I checked and the paths are valid and corresponding directory exists
<metsomedog>mehrad: ok, and did you install some applications with guix install, or maybe with guix home?
<mehrad>metsomedog: guix install
<metsomedog>and do they show up in ~/.config/guix/current/bin ?
<mehrad>metsomedog: no
<mehrad>I only have guix and guix-daemon there
<mehrad>but as far as I understand (most probably I'm wrong), the `guix install foo` should install `foo` in the default profile
<metsomedog>hmm, I'm personally using guix home to install all packages as described in a scm file ( if interested), which stores the files in a lsightly different location, so I can't confirm whether there should be more files in the ~/.config/guix/current/bin directory.
<peanuts>"Home Configuration (GNU Guix Reference Manual)"
<futurile>Morning all
<mehrad>metsomedog: thanks anyways. I definitely don't use guix home (nothing against it, just haven't used it).
<mange>mehrad, metsomedog: The ~/.config/guix/current profile is for Guix itself. If you use "guix install" it installs to ~/.guix-profile, and if you use guix home it installs to ~/.guix-home/profile. I don't know about the foreign distribution setup, but I usually source $profile_dir/etc/profile to get things to show up.
<metsomedog>mehrad: maybe you can try to run 'guix package --list-installed=".*"' -- it should list all installed packages and their path
<mehrad>mange: You are absolutely right. the ~/.guix-profile/bin contains tons of symlinks to lots of things including the software I just installed. should I modify the and add this?
<mange>I don't know about that script. Does it come from the Guix install script, or something? I source this script to set mine up: Feel free to steal anything from there that's useful.
<peanuts>"guix ?"
<mehrad>mange: I'm almost certain that it has shipped with guix install script. It is discussed here as well:
<peanuts>"[PATCH] home: services: environment-variables: Fix XDG base directories."
<mehrad>mange: Thanks for the pastebin link. I see lots of similarities between your script and this The major difference is that yours have lots of checks in place and the other one is just exporting some variables
<mehrad>mange: this is what the /etc/profile.d/ looks like:
<peanuts>"Disroot Bin - Encrypted pastebin by PrivateBin"
<mange>Are the lines that source the profiles really commented out? The ## at the start of the lines that actually source the $GUIX_PROFILE/etc/profile bits look suspiciously like comments.
<mehrad>mange: Thanks for your input. that file is created by this script: but I might have modified it in the past. I don't remember, but it is not impossible. I now see that there are some new parts in that script. I will overwrite it and see how it goes
<mange>Yeah, you can see in sys_create_init_profile where it writes that script, but it doesn't have the ## before those lines sourcing the profile. I imagine that's causing your problems.
<mehrad>mange: Thanks. I'm in the middle of a project and cannot reboot the computer so i cannot report if it works at the moment, but by sourcing the file into bash, I can run the software. So I guess the issue is resolved and will come into affect after reboot (or logout+login)
<mehrad>thanks again
<freakingpenguin>Hmm, restarted a Guix vps and suddenly it can't reach outbound, although it can ping the gateway. Fun times.
<freakingpenguin>Seems to somehow be a routing problem. An IP is assigned but the default gateway isn't which is curious since both should be coming from dhcp to my understanding.
<apteryx>daviid: hello! I see you were the one who reported; I'm trying to integrate the fixes to Guix but something is still not working with the chat (it's as you described there: receiving works, and clicking on the notification shows the conversation with its history, but leaving this view, it vanishes and cannot be found again)
<sneek>apteryx, you have 1 message!
<sneek>apteryx, raghavgururajan says: IIRC, it was something to do with SIMPLE support. Is debian using same version as us?
<peanuts>"#983365 - linphone-desktop: chat messages - Debian Bug report logs"
<apteryx>daviid: hm, apparently the soci fix was enough for you, re-reading this:
<peanuts>"#983365 - linphone-desktop: chat messages - Debian Bug report logs"
<daviid>apteryx: i actually never used the debian package, i tried, didn't work, report to try to help ... didn't get it to work, the debian package i mean, and i kept using the upstream appimage (which i never removed, update regurlarly ...)
<ngz>Does anyone know what use have upstream-source-compilers? They do not seem involved in the package updating process. OTOH, they are related to G-exps, so anything could happen…
<raghavgururajan>daviid: Can you send me a message to ?
<freakingpenguin>Hmm, that DHCP issue seems to have been transient. Cleared up after another reboot with some troubleshooting in between, although I don't see any dhclient log messages mentioning the gateway IP address.
<freakingpenguin>My understanding is the route I added via ip route is volatile, so that shouldn't still be around.
<PotentialUser-47>hi all! Started playing with Guix with sway and waybar, so far it's great. Out of curiosity, here a Xwayland question: When I kill Xwayland (via killall), it kills waybar. I thought waybar is a wayland app, moreover it's not listed by xlsclients (xterm is for example), and killing Xwayland doesn't kill waybar under say Arch/Debian. Any ideas?
<PotentialUser-47>Also gdm is disabled, just using base-services with some added services. So I login into sway from tty
<PotentialUser-47>I meant to say I voluntarily disabled gdm. Ideally I don't want anything running in X.
<freakingpenguin>Hmm, my problem seems to come from the server having both internal and public eth device with ips set via DHCP. The internal somehow gets set as the default route.
<pkill9>PotentialUser-47: maybe one of the modules in waybar depends on xwayland, and maybe waybar crashes when thqt module crashes, idk
<raghavgururajan>PotentialUser-47: Also, I suspect dbus.
<raghavgururajan>PotentialUser-47: Probably what pkill9 said. Check for any build-flags to enable features that depends on Xwayland.
<lechner>Hi, why does this package definition provide only graft from 1.08.1 please?
<peanuts>"View paste EISQ"
<stephen-loom>In emacs and info I can't follow links to `(libc)Locales` or any other (libc) for that matter, emacs says the page isn't installed. How do I diagnose which package to install to get this page? I'm guessing I install the glibc package but is there a recommended way to figure this out without trial and error?