IRC channel logs

<yaircul>In Gentoo it is possible to have a USE glag equal to "-rust" to prevent my computer from rusting and breaking. Is there a similar thing for GNU Guix?

<ieure>yaircul, Package variants are the rough Guix analogue to USE flags. I am not sure what the "-rust" flag does, if you explain more, I might be able to direct you.

<YARs2>nope, gux istaller doesn't function on pi 5

<YARs2>hmmmm

<futurile>morning morning all

<efraim>o/

<sham1>Morning /o

<adanska>morning!

<user_oreloznog>\o

<aloysb`>Yesterday, I asked for some help with defining services. I made good progress, a few things clicked after digging into `match`. Thank you gabber (and a few others). I am now hitting a roadblock on the notino of "file-like". I do not understand how `(file-like pkg)` in `define-configuration` behave. I can't find the source code for it. I understand it has to do with g-exps. Would someone be able to give me a hint as to where to look? My

<aloysb`>understanding of host code vs. build code is not solid. Should I attempt to build a toy package to understand g-exp better? Thank you.

<aloysb`>My goal is to create a 'caddy' services - but the true end goal is to understand guix in (relative) depth.

<untrusem>aloysb` if its possible, do make a writeup about your workflow, as I would be interested in knowing that as I have to write my first service as well

<civodul>Hello Guix!

<civodul>‘guix system reconfigure’ fails while building ‘switch-to-system.drv’

<civodul>anyone see this?

<civodul>looks like a recent regression

<gabber>`guix lint foo` crashes with a `bad-response'. does that mean the CVE library is down or is it a hint towards a bigger issue?

<civodul>‘guix lint hello -c cve’ seems to work here; which checker is failing?

<gabber>it says "fetching CVE database for 2025... [cve]..." just before the relatively ugly backtrace is printed

<neonio>hi guix!

<gabber>\o

<untrusem>gabber: yeah gabber I am getting it too

<untrusem>the lint error

<csantosb>Same here with `guix lint hello -c cve`, "`("EOF while reading response body: ~a bytes of ~a" (10776695 23324652))'"

<aloysb`>untrusem: for sure, will do.

<yelninei>are there any plans on updating the guix package sometime soon?

<gabber>so this is probably a service issue? i had it work for 2025, 2024 but then fail for 2023

<untrusem>yelninei, A mail in mailing list would probably get you an answer :P

<untrusem>what process starts gpg-agent?

<gabber>untrusem: depends. do you have `gpg-agent` in your guix home services?

<untrusem>nope

<gabber>then i'm not sure (:

<untrusem>okay figured it out :P

<gabber>what was it?

<abbe__> https://www.irccloud.com/pastebin/VlcrvfpV/ switch-to-system.scm is broken for me with https://codeberg.org/guix/guix/commit/c226fb3b70c20fe2418aad2df2b4f2ec83472dfc (current guix revision)

<untrusem>yep this know abbe__

<untrusem>known*

<untrusem> https://codeberg.org/guix/guix/pulls/9516

<abbe__>oh, is it being tracked somewhere ?

<abbe__>i see, thanks!

<untrusem>ludo message earlier here about this issue

<abbe__>I see. loose dependency on 'json'

<abbe__>maybe

<cnx>lavandula, misc-world-rebuild is first in line for merging but the qa build servers have been struggling to evaluate it for some reasons

<cnx>see https://debbugs.gnu.org/80782

<cnx>it's been that way for a months, failing during builds of packages we've been able to build locally

<zaskyv>yagamaI

<zaskyv>detal mick

<zaskyv>sigeon pex

<zaskyv>i luv sigeon pex

<cwebber>I've never been involved in anything involving grafting, but the ffmpeg vulnerability feels like something we should get out a high priority fix to users https://social.coop/@cwebber/116810673204863384 / https://www.securityweek.com/ffmpeg-pixelsmash-flaw-allows-rce-on-video-players-media-servers-nas-appliances/

<untrusem>cwebber: there is a pr for this

<untrusem>#9442

<cwebber>oh good

<cwebber>whew

<untrusem>we need to get teams attention to that one

<csantosb>Why exactly `guix refresh --list-dependent --recursive emacs-tintin-mode` gives so many dependents ? 🤔

<csantosb>This is a leaf package

<futurile>could we please have the next GCD _just_ about an engineering concern. I'm even willing to argue about the commit log format now!

<futurile>(the second part was a joke btw - I'm not having a good day for jokes in my electronic comms ... though my snappy one-liners at the coffee shop today rocked)

<csantosb>Babeltrace depends on lttng-tools, which is a leaf; still, the --recursive flag sends tones of packages

<csantosb>futurile: next GCD is a more consensual topic, editors war 😯

<futurile>ACTION laughs ... then cries

<futurile>I'm all for it though - ed for the win!

<sham1>feat: Make sure that people follow the changelog style

<cwebber>yay the fix is merged, thx untrusem / SilverlightningY

<jlicht>futurile: I have one, we should rewrite it all in Rust! :D

<jlicht>although the polemics are getting kind of hard to follow, agreed

<untrusem> https://red.chatoyer.de/r/rust/comments/1ufaef8/comment/otqav8x/

<untrusem>Rust space problem

<identity>s/red.chatoyer.de/old.reddit.com/ if you do not want to hit the Anubis wall

<yelninei>janneke: Did you see guix/guix!9113

<ieure>ACTION is confused by all the Gitlab-style ticket references, when we don't use Gitlab.

<yelninei>ACTION just copy pasted what codeberg buttton put in their clipboard

<untrusem>cwebber: I don't the fix in master branch..

<cwebber>oh whoops I misread the "manually merged" button

<janneke>yelninei: nope, missed that

<sham1>Okay, there's something extremely odd going on with the `emacs-guix` package. I can't even list the package names because `unbound-variable` errors keep getting raised. But on the other hand, when I try to do the same thing on the guix repl, it works just fine

<ieure>sham1, What variable is unbound?

<untrusem>cwebber: Its just that i don't want to build ffmpeg on my t480 but i guess will need to do that

<untrusem>emacs-guix has its fair share of errors

<sham1>ieure: it might honestly be easier if I just post the error message in a pastebin

<untrusem>check the issues in emacs guix repo too

<untrusem>sham1:

<sham1>Probably should

<untrusem> https://codeberg.org/guix/emacs-guix/issues/15

<sham1>Ah, yes. That's the ticket. Well, I'm getting that error message, and with `shared-mime-info` as well. But the REPL seems to launch just fine

<sham1>I might have to try with guix/emacs-guix!18 then

<yelninei>janneke: It fixes XOpenDisplay (via XForwarding, assuming openssh is compiled with a functioning _PATH_BSHELL and can find xauth). So you can start running your X application

<janneke>yelninei: neat!

<guixer-without-d>hello everyone, i just did pull and reconfigure on my machine and now dns resolution does not work even though networkmanager (nmcli dev show) shows a valid DNS entry. using host/dig/nslookup with explicit dns server also works. Am I missing a service like dnsmasq? /etc/resolv.conf has only `#placeholder`

<ieure>guixer-without-d, It's a known issue, there's a bug about it.

<untrusem>its also in news file so you would have gotten a msg if you guix pulled

<identity>guix pull --news

<guixer-without-d>is the news message still somewhere on the system?

<identity>it should be, run that command

<guixer-without-d>i found the message in news and followed instructions, now it's fixed, thanks!

<identity>guixer-without-d: guix pull should tell you to run it with --news when there are any news, make sure to do that next time

<civodul>guixer-without-d: also, ‘guix pull --list-generations’ will show you previous news entries

<bjc>ieure: is that a permanent solution or just a stop-gap to be run on every reboot?

<guixer-without-d>i'm not updating guix very often so these are good tips. I usually, run `guix system reconfigure /etc/config.scm` and then `guix pull` for each user

<bjc>civodul: ‘guix pull -l’ doesn't list the entry for 86776ccd30bb44eb856fcf3a14a8338f38fe1367

<bjc>it has the one before and after it, but not that one

<bjc>interestingly, i should have seen that message in generation 49, but in generation 50 i see the message about postgresql, which comes before the network manager message in etc/news.scm

<bjc>i'd guess i did a pull, got the bad commit, pulled again later, got the news item for postgres, pulled again later, after news.scm was updated with network manager, but since that commit predates the commit for a news item i've already seen, it gets skipped

<ieure>untrusem, I don't believe that's the whole story, I observed this on systems I reconfigured and then rebooted.

<umanwizard>Anyone seen this?

<umanwizard>/tmp/guix-build-clang-runtime-22.1.8.drv-0/source/build/lib/fuzzer/libcxx_fuzzer_aarch64/build/include/c++/v1/__config:900:75: warning: requested ‘init_priority’ 100 is reserved for internal use [-Wprio-ctor-dtor]

IRC channel logs

2026-06-25.log