IRC channel logs

2026-01-31.log

back to list of logs

<Guest47>I'm trying to run the `make check` tests from a git checkout, but they keep trying to download substitutes that fail with multiple 404 errors from mirrors
<Guest47>Here's a paste of one test in particular failing after attempting to download xd2mb8mkay5wlhz9gbq1s0w1pnsww9zv-bash
<Guest47> https://paste.debian.net/hidden/f0681418
<Guest10>Hi guys, the version of rust packaged in guix is a little behind what i need. I saw that you can use guix shell fhs containers to install rustup and the latest binary. My question is, how are you guys integrating that with emacs? Is there some easy way to make emacs automatically do everything in that container when i'm in some project dir? direnv
<Guest10>seems to only be about exporting variables
<Guest47>I'm having this issue, is there really no way to make my tests work until this is fixed? Can anyone else reproduce it?
<Guest47> https://codeberg.org/guix/guix/issues/4034
<untrusem->rust-1.90 substitute is waiting for build slot for 3 days
<untrusem-> https://ci.guix.gnu.org/build/17440292/details
<untrusem->Guest10: upto rust-1.90 is packaged, but its is rust-team branch
<Guest10>untrusem-: thanks. I'd still be interested in hearing about guix shell container workflows if anyone has one
<untrusem->I just copy the rust.scm file from rust-team branch in my channel and use them from there
<untrusem->guix shell's `-e` flag would help in this
<untrusem-> https://bpa.st/7AN4Y <- a little example
<untrusem->you might need to use `-L` too
<untrusem->(info "(guix)Invoking guix shell")
<untrusem->Guest10
<Guest10>Thanks, I know how to do that and I can now make a shell with rust-1.90 and work on my project. The thing I'm still interested in though, out of curiosity and in case i need something that *really* isn't packaged in guix is how people are using guix --container --emulate-fhs. I guess it would probably be something with tramp
<untrusem->I wish people could shared substitutes with each other
<untrusem->I need rust-1.90 but don't want to build it for 2 hours
<untrusem->Guest10, you mean specifically for rust binaries?
<untrusem->have you read this https://zie87.github.io/posts/guix-foreign-binaries/ ?
<Guest10>Yeah, i can install rust binaries in it. The blogpost announcing --emulate-fhs also has an example of using rustup in it. The piece i'm missing is how to integrate it nicely into emacs. You'd want to run lsps in the container, run your compile commands in it, etc.
<Guest10>With normal guix shells to get that integration all i have to do is make an .envrc and use emacs-direnv, but i don't think that can work with containers
<untrusem->I mean you could use a shellscript for that
<untrusem->for example https://codeberg.org/Adroit/run-appimage-on-guix
<untrusem->but I don't how to do that exactly
<untrusem->I mean the workflow of shell with containers with .envrc
<untrusem->wishlist - a way to share substitutes with each other
<untrusem->I don't think I am the first one to want that
<untrusem->Guest10, could you share your rust-1.90 nar
<untrusem->(info "(guix)Invoking guix archive")
<Guest10>i haven't built it yet either
<untrusem->ohh I see
<parra>respect to channel-with-substitutes-available I did an analogous mechanism for docker in my system:
<parra> https://github.com/metacall/guix/blob/df459d438a8aed815de0fef89d532e83c2b5725c/release.js#L217
<parra>it would be great to achieve something like this but for any ci farm 🗿
<futurile>Guest47: you can try ``guix build --no-substitutes --no-grafts --source <package>` to see if the source (origin) is correct for the package
<ngz>Hello. Is it me or guix refresh actually downgrades packages ? I tried "guix refresh -s module:emacs-xyz -u", and I get a lot of reports such as "updating from version 1.6.2 to version 1.5.1". Some packages do get updated, tho.
<ngz>I would say that around 80% of refreshes are bogus. Hmmmm.
<csantosb>I confirm, it happens
<futurile>systemd-oci talk at #FOSDEM is pretty interesting
<theesm1>futurile: what's it about? orchestrating oci image-based services with systemd?
<futurile>theesm1: yes exactly, and providing some tooling to make it easy to set-up. They have a good vision for the tooling which will make it pretty accessible if you're a systemd user from what I could see
<theesm1>seems like sth worth looking into, maybe there's a thing or two to take inspiration from so oci-service-type in guix could benefit from those ideas
<futurile>theesm1: yeah. In general, I think 'services' are quite hard to get into in Guix. It requires you to be quite an expert I guess.
<kestrelwx>Hello! Thanks everyone for the release work.
<futurile>kestrelwx: down to Rutherther, noe and efraim - they did all the heavy lifting!
<reyman>Hi !
<reyman>After a guix update today, my yubikey seems not recognized (again), someone have the same problem ?
<reyman>I use guix home
<Rutherther>reyman: https://codeberg.org/guix/guix/issues/5940
<Rutherther>if you're on foreign distro, likely the pcsc in your system has to be updated
<Rutherther>or the one you use from guix downgraded
<reyman>Rutherther: Thanks ! I try now
<reyman>hum, i'm using guix, i have this :
<reyman>gpg: erreur d'exécution de « /gnu/store/6cqp72xly1aclylz21jz1lnkfg9rwpj7-gnupg-2.4.8/bin/gpg-agent » : code de retour 2 gpg: failed to start gpg-agent '/gnu/store/6cqp72xly1aclylz21jz1lnkfg9rwpj7-gnupg-2.4.8/bin/gpg-agent': Erreur générale
<reyman>when i try gpg --card-status
<reyman>strangely my key works in firefox ..
<reyman>and the gpg-agent seems running when i run ps -aux | grep gpg-agent
<reyman>why gpg is so weird :(
<futurile>reyman: to be clear are you running Guix directly as a Linux distro, or are you running on top (hosted) of another distribution?
<reyman>i run guix as my linux distro futurile
<reyman>scdaemon[15338]: pcsc_establish_context failed: inserted card (0x8010006a) gpg-agent[15320]: DBG: chan_10 <- ERR 100663614 Service is not running <SCD> gpg-agent[15320]: DBG: chan_9 -> ERR 100663614 Service is not running <SCD>
<reyman>this is the debug when i run pkill gpg-agent; gpg-agent --daemon --no-detach -v -v --debug-level advanced
<reyman>so yes, seems there is a problem with pcslite
<reyman>i don't have a scdaemon.conf is it needed now ?
<reyman>Rutherther: the polkit rules is builded during system reconfigure , but i don't see the service policy-kit with hurd status
<Rutherther>policy-kit service? Where would that be coming from? I am not aware of a policy-kit shepherd service
<reyman>ah
<reyman>i add this : (simple-service 'polkit-pcsc polkit-service-type (list pcsc-lite))
<reyman>just after that : (service pcscd-service-type)
<Rutherther>that adds a polkit rule, yes. No changes to shepherd services expected
<reyman>perhaps i need a real reboot Rutherther?
<reyman>ok
<Rutherther>yes, you need to reboot for new polkit rules to be loaded
<reyman>ok !
<reyman>so see you later :)
<futurile>reyman: hope it works!
<reyman>YES
<reyman>thanks Rutherther and futurile
<reyman>it works now
<sunless>hi
<futurile>heya sunless
<sunless>just testing stuffs, my first time using irc ... just trying to get comfortable, good to know it works
<sunless>oh, i've remembered, i had an issue with emacs packages, i installed =emacs-docker= and it installed =emacs-elfeed= despite not specified in my home config, there is also another packaging also "sneaking in" magit too, i just don't know yet
<sunless>i'm now seeking ways to go back to package.el, probably editing $EMACSLOADPATH
<sunless>futurile: is it normal to have "big" emacs packages installed as dependencies?
<futurile>sunless: sorry I'm the worse person to ask, I don't use emacs
<quassel-guy>Hello o/
<sunless>futurile: oh, that's ... weird lmao, i'll be patient for an emacs user, i appreciate your consideration
<sunless>quassel-guy: hello
<quassel-guy>Hello sunless, I have a question on Rust developement, how is it normally done? like adding libraries and stuff like that?
<sunless>quassel-guy: i'm also struggling with my rust fhs container, but my guess to your problem is trying "export LD_LIBRARY_PATH="$LIBRARY_PATH""
<quassel-guy>sunless: No I have no idea how things are done, like do you write a manifest.scm file and guix shell into the environment every time you do rust developement?
<sunless>quassel-guy: okay, i actually can't explain much because i'm also a beginner who hasn't successfully set it up
<quassel-guy>sunless: OK, let's set this thing up togethre my friend
<sunless>quassel-guy: has written some blogs about it, i'm sure he can explain
<quassel-guy>sunless: who?
<sunless>quassel-guy: i'm planning to set it up tomorrow, we'll track each others progress buddy
<sunless>futurile
<sunless>futurile: you there?
<quassel-guy>sunless: OK bro Hope you get this working properly
<sunless>quassel-guy: yeah, i don't want to go back to working in a docker container
<nbguixuser0>I have a question about emacs-mini: I notice that emacs-mini's parallel-build option is turned off but I couldn't find the reason on codeberg issues, anyone has experience prolonged build time due to this?