IRC channel logs


back to list of logs

<civodul>dgr: that file exists on my laptop but it’s not empty
<civodul>could there be some data corruption happening here?
<dgr>civodul: how could it happen or how can I force a redownload?
<civodul>dgr: maybe “guix gc --repair /gnu/store/b1dc9qrjaq1gdbrjwjg4z8xxqd6c5kp2-wget-”
<civodul>sorry, “guix build --repair /gnu/store/b1dc9qrjaq1gdbrjwjg4z8xxqd6c5kp2-wget-”
<dgr>guix build thinks mrepair is a package
<dgr>guix build: error: mrepair: unknown package
<dgr>guix build --check and --repair did download the substitute but is still empty
<civodul>actually this is a “graft”, so it cannot be repaired (there are no substitutes for itself, only for the ungrafted variant)
<civodul>what about “guix build --repair /gnu/store/h7cq6s58l5prqbiqxz46qd098d04nxzv-wget-”?
<dgr>guix build thinks mrepair is a package
<dgr>currently it's building without substitute
<dgr>but it looks like the whole toolchain gets build an not just wget
<lykso>Hey, does anyone have any pointers for referencing a local file in a udev rule? Doesn't seem like "udev-rule" takes a g-exp and I don't see anything I can refer to in the Guix repository.
<sergiodj>ngz: thanks for the feedback. I wasn't sure if using G-exps is the "modern" way or not; I initially did wrote the package using it but then switched to non-G-exps. I'll revisit the package and update it
<lykso>It looks a little bit convoluted to me, but this >> << got me to a working solution using file->udev-rule, computed-file, and call-with-output-file (the use of which I learned from redshift-shepherd-service in gnu/home/services/desktop.scm)
<trevdev>I'm refactoring my home packages from package->specification to imported symbols. One of my specifications was "stumpwm:lib". I don't know how to get this as a symbol. Suggestions?
<trevdev>`(,package-symbol "lib") <-- How I got the build output for the
<trevdev> "lib" output without specification->package for "package-symbol:lib"
<efraim>sneek: later tell trevdev stumpwm:lib was just merged back into stumpwm so I wouldn't worry too much about trying to use stumpwm:lib at this point
<sneek>Got it.
<efraim>sneek: botsnack
<fnat>I'd like to aggregate some of my Guix Home services together. E.g. individual services for, say, mbsync, msmtp, notmuch, and then a global email-home-service that includes all the previous ones.
<fnat>However, I don't see a 'depends-on' kind-of field in the home service record - field which I was thinking of using for the aggregation.
<fnat>Hm... it must be 'extensions'...
<fnat>ACTION goes and gives a look at the code.
<simendsjo>I'm testing guix home, but I don't know how I'm able to use X for the container. Any pointers to the correct place in the documentation?
<civodul>simendsjo: hi! to use X in ‘guix home container’, you have to pass ‘--expose=/tmp/.X11-unix’ at least
<civodul>you may also need to set DISPLAY in the shell
<dgr>while trying to fix my info/wget problem I tried `uix build --no-substitutes --repair wget` but now finished glibc can't be moved to the final destination because the folder already exists - I guess, I'm holding it wrong
<Deltafire>argh.. trying to install guix on a laptop and it's builing everything.. feels like i'm running gentoo :/
<dariqq>Hi, trying to set up guix deploy, and i am a bit confused about the authorize? field. Previously i used a service to add the key. I commented the service and after deploying i get a key error but a new generation was instantiated. Am i misunderstanding something?
<dariqq>Ah i think the problem is that the key gets authorized and then /etc/guix/acl gets overwritten by the new generation which does not have the key.
<dariqq>Seems to be already reported in #46760
<peanuts>"guix deploy doesn't seem to be authorizing the machine that is deploying to the remote"
<pastor>Hello, please; could anyone help me to understand `map-derivation`? It feels like it does not behave as it's supposed to. Anyone has a simple usage example?
<pastor>Am I doing something wrong here?
<attila_lendvai>how can i find out which commit/package made the chromium binary invalid? i want to pull/reconfigure, but don't want to build chromium locally...
<attila_lendvai>...nor icedove.
<attila_lendvai>ACTION tries to pull prior to the llvm fixes
<attila_lendvai>no luck... going back 3 days
<ngz>attila_lendvai: Can't you git bisect the repo?
<attila_lendvai>ngz, which repo? i'm doing `guix pull --commit=[hash] --allow-downgrades` and try to `guix home reconfigure ...` and hoping that i won't need bisecting, just pick an old enough commit.
<ngz>cbaines: If we ever need to restart "core-updates", please let me know: I have a couple small fixes to add. However, those small fixes do not warrant restarting the branch on their own.
<ngz>attila_lendvai: I thought "which commit" referred to Guix repo.
<ngz>cbaines: BTW bulding "core-updates" is moving slowly, and accessing to information about, e.g., failing packages is not possible due to "#fvector-fold-right: expected vector, got ~S#f#f" error. Do you know why?
<attila_lendvai>FTR, i went back to commit c6e6001a10d96b7c8f023e3d9b7022d935d9f2f6 to avoid building chromium locally
<attila_lendvai>ACTION relogins
<peanuts>"guix.git - GNU Guix and GNU Guix System"
<ngz>attila_lendvai: Is chromium still updated on Guix, BTW? I haven't noticed an update for a while.
<attila_lendvai>civodul, the manual designates guix-devel for shepherd patches, and i've sent a few shepherd patches to guix-devel, but... now it feels a bit like spam.
<ngz>(which is sad, because there are not many options to cast a screen through a ChromeCast)
<attila_lendvai>ngz, yeah, good point. i also searched for updates in the git repo, and the last one is from last summer.
<cbaines>ngz, is there a patch series for those small core-updates fixes?
<cbaines>ngz, also, where are you seeing that error?
<ngz>cbaines: I can create one. I think the changes could be split into 3 patches.
<ngz>cbaines: Here when following link pointing to the number of failed packages, for examples.
<cbaines>great, obviously they might want squashing in to other commits, but it's helpful to just have them as a patch series
<cbaines>ngz, ah, and yeah, I see the QA error, I'll try to fix that...
<cbaines>due to the number of changes on master, core-updates is going to be slow to get building again
<ngz>So you suggest to send the small patches to bug-guix, with [core-updates] tag?
<cbaines>to guix-patches, but yes
<ngz>Yup. Mumi is sending them.
<lechner>Hi, can g-exps serialize lists, or do i really have to to (list #$@arguments) ?
<lechner>attila_lendvai / provided civodul approves, we could give Shepherd a separate "package" on Debbugs (which we also did for Mumi). It would get us or similar
<lechner> can be modified to still show Shepherd reports, as it does for Mumi
<lechner>also, reports can be reassigned between shepherd and guix
<Guixnewbie13>Hi all,
<Guixnewbie13>I just installed a guix VM under debian
<Guixnewbie13>qemu-system-x86_64 -m 4096 -smp 2 -enable-kvm   -nic user,model=virtio-net-pci -boot menu=on,order=d   -drive file=guix-system.img   -drive media=cdrom,file=guix-system-install-1.4.0.x86_64-linux.iso
<Guixnewbie13>Guix VM is now started. I would like to switch to plasma desktop at next reboot.
<Guixnewbie13>Could you please confirm that the steps to follow are :
<Guixnewbie13>1/ sudo vi /etc/config.csm
<Guixnewbie13>   to replace (service  gnome-desktop-service-type)
<lechner>Guixnewbie13 / you were silenced/ please use a paste site
<Guixnewbie13>it is a question about "how to use (service plasma-desktop-service-type)"
<grtcdr>Hi y'all, having switched to systemd-homed, Guix has started to bug me (and rightly so) with every command about my user ID not existing in the default lookup path, i.e. /etc/passwd, my nsswitch.conf lists two lookup methods for passwd: files and systemd, I realize Guix does not support the latter but is there any way I can silence that message.
<simendsjo>civodul: Thanks! Exposing /tmp/.X11-unix wasn't enough, but I see two of the files has user 65543 and group overflow, so I guess I have to link the users/groups together somehow?
<grtcdr>The exact message is ""user with UID 1000 not found". I've found in the mailing list suggestions to install and enable the host system's nscd, but I've been using Guix for at least 6 months now on Arch Linux without nscd and things have been working great without it.
<ngz>OOC, where do you put your home-configuration.scm file, and the other files Guix Home brings in, such as .bashrc, before symlinking them?
<ngz>I tried ".config/guix/", but I'm not sure this is a good idea, as the directory has initially other purposes.