IRC channel logs

<vntsuyo>is it possible to attach to a running guix shell?

<sneek>Welcome back untrusem!!

<untrusem>sneek: botsnack

<guixtroubles>Hey everyone, I recently installed guix on an XPS 13 and found out that guix doesn't have access to the full memory space. htop and free think there are only 4.3 GB

<guixtroubles>when there should be 16 GB there

<Rutherther>newguixbie: did you install i686?

<csantosb>Rutherther, I've been away for a while, non fixes pushes to master or next ?

<Rutherther>csantosb: updates/changes to next-master, unless they're fixes. I think additions of new packages/services don't really matter if they go to master.

<csantosb>Ok, thanks, and congrats for the rc1 !

<csantosb>By the way, regarding merges in next-master of pr targeted to master, I'm closing the issue with a hint on the commit, see #4974 for example. The usual "Manually Merged" blue button is useless in this case.

<Rutherther>it should be working as long as the PR has next-master as target. It can be retargeted, however it might bring some new commits to the UI and start becoming a bit confusing when the PR has been made to target master in the first place

<PotentialUser-93>hello! I sent an email to the devel mailing list yesterday but it doesn't show up on the archive: does anyone know why? :)

<PotentialUser-93>My email is about getting BcacheFS work as root FS in Guix. I also emailed separately the BCacheFS amiling list and it seems there is no strong obstruction to make this work... that's why then I messaged emacs-devel

<PotentialUser-93>sorry I meant guix devel

<identity>PotentialUser-93: slow, human filter

<identity>you probably need to wait for about a week for your first email to land

<PotentialUser-93>Aaah ok Cause on guix-help it went through immediately therefore I thought would be the same. It's because I sent on guix help the same message but then realised it's really for devel because it's for soething not yet implemented

<untrusem>we have codeberg too

<PotentialUser-93>yeah but codeberg requires making yet another account (and also seems to me less suitable than the devel mailing list )

<untrusem>I mean as te development is shifter there, it has more chance for getting reviewed there

<PotentialUser-93>thx for the suggestion, but for now the fact that I have to make an account puts me off. let's see if anyone replies to the mailinglist

<EnhA3>Is there an ISO image for installing GUIX on cloud virtual machiens?

<identity>EnhA3: not that i know of, but (info "(guix-cookbook) Running Guix on a Linode Server") and (info "(guix-cookbook) Running Guix on a Kimsufi Server") may be of interest

<gabber>EnhA3: you can craft your own minimal image that you want deployed on the target machine at ease, then deploy that

<gabber>i've described the process here: https://mit.teil.space/03-setup.html#org4597c2a

<untrusem>nice

<gabber>i move a big section in the cookbook to another place. now git interprets empty lines as being "the same code as before" leaving me with giant, ugly diffs. is there a workaround?

<untrusem>you can use difftastic

<untrusem>or delta

<untrusem>you use emacs gabber?

<gabber>yes

<gabber>i'm less concerned about what i do (it's a rather simple change) but that it remains reviewable

<untrusem> https://tsdh.org/posts/2022-08-01-difftastic-diffing-with-magit.html

<untrusem>ohh then I don't think difftastic will help, its basically a diff tool that acts of syntax change

<gabber>i guess i have to find a way to do that change that *git understands* then

<nixnixnix>strange, i can't install docker-compose

<nixnixnix> https://paste.debian.net/hidden/2298c140 << log here

<nixnixnix>my way to install is via config.scm

<gabber>what's the exact command you invoke, nixnixnix ?

<nixnixnix>guix system reconfigure /etc/config.scm

<nixnixnix>even guix install docker-compose produced the same error

<gabber>`guix shell docker-compose` spins up an environment with docker-compose ready

<gabber>have you pull'ed recently? has the same configuration worked before?

<nixnixnix>let me pull first

<nixnixnix>eh, let me try via shell first

<nixnixnix>ok, shell error too

<gabber>that is odd

<gabber>you pulled right now?

<nixnixnix>i done pull one hour ago.. let me try again.. it should be dont have different effect

<nixnixnix>the problem is python-dockerpty-0.4.1

<nixnixnix>i cannot install; i wonder why

<gabber>could you maybe share all the output?

<nixnixnix>gabber: thx! https://termbin.com/slqe

<gabber>ok, i get the same error with a recent checkout. would you mind opening an issue on codeberg, nixnixnix ?

<gabber>or, if you feel confident, try to fix the issue?

<nixnixnix>let me open an issue on codeberg

<nixnixnix>oh yeah, fix the issue? interesting.. g

<gabber>it seems that the `check' phase of python-dockerpty tries to do invoke command (or argument) "test" which seemingly does not exist?

<untrusem>ohh it is relevant to pyproject migration

<yelninei> https://codeberg.org/guix/guix/issues/5016

<nixnixnix>thank you

<nixnixnix>can i add https://codeberg.org/guix/guix/commit/8148ca84373ad3985bf522dfc892d512266fe0c8 on config.scm?

<nixnixnix>if cannot, where should i add?

<gabber>you can pass the commit hash to `guix time-machine --commit=<HASH> -- shell docker-compose`

<gabber>or similar command

<gabber>or—if you use multiple channels—craft your own channels.scm file (`guix describe --format=channels > channel-pin.scm`) and change the hash of the main guix repository

<nixnixnix>i'm new.. i still dont understand about it

<nixnixnix>:x

<gabber>then use this file (channel-pin.scm) for your `guix time-machine` invokation

<nixnixnix>i wanted to put config.scm, so i dont need another step for another user

<gabber>so, you want to use the commit you just posted to reconfigure your system?

<gabber>sorry, i don't know what your config.scm is. is it a system or home configuration declaration perhaps?

<nixnixnix>perhaps i said wrong. not all those on the commits, but the variables that has green background

<nixnixnix> https://codeberg.org/guix/guix/commit/8148ca84373ad3985bf522dfc892d512266fe0c8

<nixnixnix>configscm is my system configuration

<gabber>i am not sure the commit linked by you solves your issue

<gabber>it is already included in the main guix repository

<gabber>nixnixnix: have you opened the issue yet? it is possible people willing to help you are already lurking around and just waiting for the opportunity to fix some python package (:

<nixnixnix>yeah, i will do later if i can access my laptop

<nixnixnix>thx gabber

<gabber>no problem, i will open the issue for you then

<gabber>nixnixnix: https://codeberg.org/guix/guix/issues/5164

<nixnixnix> https://codeberg.org/guix/guix/src/commit/63b71a5152f8fc318ce598563fa7f605006f8074/gnu/packages/docker.scm#L278

<nixnixnix>i should edit this one

<nixnixnix>gabber: thanks mate :-)

<gabber>HTH

<nixnixnix>by the way

<nixnixnix>how to install docker.scm?

<gabber>you cannot "install" docker.scm, it's where all the different package definitions live. you can install the packages defined in there.

<gabber>it's where both docker-compose and python-dockerpty are definedd

<nixnixnix>thank you gabber

<luca>Hi, has anyone considered enabling landlock in the kernel? Ref https://docs.kernel.org/userspace-api/landlock.html#boot-time-configuration

<bdju> https://github.com/lxqt/qterminal/wiki#how-to-set-my-preferred-font-for-monospace I tried following these examples but I'm not seeing my monospace font used in nheko updating. I did fc-cache -rfv afterward and also restarted nheko.

<bdju>It says in LXQt 2.4 or newer that the graphical appearance config app has a setting for the monospace font. It looks like guix just has 2.2. I wonder how hard that would be to update.

<bdju>Using --with-latest fetches 2.3.0 of lxqt-config and fails to build.

<bdju>`fc-match monospace` shows my new settings but nheko hasn't changed at all. Maybe I'm hitting a bug.

<untrusem>> Does Shepherd allow for sandboxing of services the way systemd can?

<Rutherther>untrusem: shepherd is much simpler than systemd, so it doesn't have sandboxing built-in. Yes, it allows for you to do sandboxing yourself. Guix has least-authority-wrapper for that. Unfortunately it's not easy to wrap already made services, but custom ones are easy to wrap

<newguixbie>rutherther: yes I installed i686

<Rutherther>newguixbie: then that's your problem. 32bit cannot support more than 4 GB of RAM, it cannot address it, it has just 32 bits. You should've installed x86_64

<Rutherther>...if you wanted more than that

<newguixbie>I eventually figured I can launch hurd images by generating them hurd-qcow2 and passing -nographic … I’m _guessing_ that the networking errors are because the images are not getting past grub, they all hang at “Welcome to grub” even from recent commits (although I haven’t tried latest 1.5 stuff)

<newguixbie>I’m curious what configurations are known to run hurd successfully

<newguixbie>rutherther: I’ve mostly run into ram issues building newer larger packages, qt6 and guix pull. guix pull seems resolved if I run from checkout, the makefile compiles the files in batches

<newguixbie>rutherther: oh! I think you typed my name earlier instead of somebody else’s and I happen to have installed i686 but you were likely talking to somebody about ram issues

<Rutherther>guess so, you both have guix in the name and I didn't check the auto completion

IRC channel logs

2025-12-27.log