IRC channel logs


back to list of logs

<rekado>Kolev: the lilypond package installs $out/share/emacs/site-lisp
<rekado>it contains several Emacs modes for Lilypond files
<Kolev>rekado: Shouldn't Emacs be able to find M-x lilypond-mode?
<civodul>anyone has a subscribe link to ?
<graywolf>civodul: here you go
<civodul>oh, thank you graywolf :-)
<civodul>ACTION discovers the existence of gittuf
<graywolf>ACTION discovers that guile module system is... hard
<graywolf>Why does this fail inside (guix channels) with "load*: unbound variable"?
<peanuts>"debian Pastezone"
<civodul>graywolf: ‘use-modules’ is meant to be used at the top-level only
<civodul>that it “works” in other context is accidental
<graywolf>omg (@ (guix ui) load*) works
<civodul>but check out for example ‘read-operating-system’ to implement that sort of thing
<graywolf>Will do, thanks for the pointer :)
<civodul>and ‘make-user-module’
<graywolf>Oh, make-user-module creates a *new* module, with those listed imported
<graywolf>Shame on me for missing that completely
<graywolf>I see, it uses (module-use!) under the hood instead of the (use-modules)
<graywolf>civodul: Thanks a lot, as usual ^_^
<civodul>yw :-)
<civodul>Guile modules are glorified hash tables
<KE0VVT>Hm. I added (add-to-list 'load-path "/gnu/store/vs02xknnxsj6cscxc0a1l5hk3nyl34dx-lilypond-2.24.3/share/emacs/site-lisp") but still cannot execute lilypond- anything.
<freakingpenguin>KE0VVT: Did you run the code in lilypond-init.el?
<KE0VVT>freakingpenguin: I just did (add-to-list 'load path ...) and (require 'lilypond-mode).
<ram_man>I ran "guix install glibc-locales" and I got "while setting up the build environment: executing `/gnu/store/mzfkrxd4w8vqrmyrx169wj8wyw7r8i37-bash': Exec format error". Does anyone know what I can do?
<KE0VVT>freakingpenguin: Do I actually need to open up lilypond-init.el and eval the buffer?
<KE0VVT>freakingpenguin: Added lilypond-init.el to my init.el. It works now! Thank you!
<jackhill>sneek: later tell lilyp I've been using epiphany from gnome-team for a few days now. No problems to report!
<sneek>Will do.
<lilyp>nice to hear
<sneek>lilyp, you have 1 message!
<sneek>lilyp, jackhill says: I've been using epiphany from gnome-team for a few days now. No problems to report!
<lilyp>has someone looked into why ibus doesn't produce useful emoji anymore?
<acidbong>moin, fellas. would it be possible to install a Guix System from a different distro with guix on it? I see that the installation requires `herd`, which isn't available outside Guix System. Or can I temporarily have a guix shell with herd in it, just for installation purposes?
<iyzsong>acidbong: yes it can, and 'herd' is only used for 'cow-store' which is useful in livecd environment where the host /gnu/store doesn't have enough space (backed by RAM). on a different distro, if its /gnu/store have enough space then all is okay, if it's also livecd environment then some steps to setup a overlayfs is needed to duplicate what 'cow-store' does.
<acidbong>but there's no analogy to nixos' `nixos-generate-config` tool, right? the ISO only seems to store templates
<jlicht>I'm running into an issue w.r.t. overriding packages that come bundled with emacs, specifically those using eieio's defclass: while the help system correctly points to my override, the used classes are still the built-in ones
<jlicht>reproducer: `guix shell --pure emacs emacs-eglot -- emacs -Q`, `M-x describe-symbol` -> jsonrpc-connection -> notice that the help buffer lists the details from the built-in jsonrpc class, while the link to "jsonrpc.el" links to the more recent version (as propagated by emacs-eglot)
<dariqq>when wrapping a python file is there a nice way to only reference the GUIX_PYTHONPATH of the inputs and not native-inputs? Otherwise the native inputs end up in the final package
<dariqq>(when simply using (getenv GUIX_PYTHONPATH)
<lilyp>Sadly no, because there is no clean separation when doing native builds (the default)
<lilyp>You can use disallowed-references
<dariqq>from what i see disallowed-references cause a build failure if the final package includes a disallowed-reference. for #70460 it is only 1 python package that is requried during runtime and I specify the pythonpath manually
<peanuts>"[PATCH 0/4] Update power-profiles-daemon and add a service-type for it."
<mrvdb>I'm affected by on my machine (powerpc64le). Is there anything I can do locally to make sure 'guix pull' succeeds?
<peanuts>"nss@3.99 is really hard to build"
<cbaines>mrvdb, not that I know of unfortunately
<cbaines>I'm waiting to see what QA thinks of the changes here which I think will help
<peanuts>"[PATCH] gnu: nss/fixed: Tweak performance test."
<mrvdb>cbaines, thanks
<ngz>Using Guix or Guile tooling, how can I write the output of a shell command, such as "ls" to a file? I tried (call-with-output-file file (lambda (port) (invoke "ls" ...))), but this doesn't seem to work.
<cbaines>ngz, I'd suggest using spawn and specifying #:output as the port to the file
<acidbong>is systemd-boot support planned in the future? it can be compiled separately from systemd if necessary (Gentoo does that for it, tmpfiles and udev)
<ngz>cbaines: hmmm, thanks. This looks like low level (need to close port manually, etc.)
<lilyp>spawn can do I/O redirection?
<lilyp>and here I am doing dirty fork/exec :(
<apteryx>is there so substitute for 'guix build -S linux-libre' ?
<lilyp>There ought to be one, but idk if one's currently built
<janneke>sneek: later tell acidbong: i haven't seen a system-config generation tool outside of the installer, that might be neat to have
<sneek>Got it.
<janneke>sneek: botsnack
<janneke>shouldn't be too difficult to adapt something like that from the installer, I guess?
<Kabouik>Any idea what could be wrong here? I just pulled and updated all my packages so I should have only nss-certs 3.99. I do have nss-certs in config.scm too, but not a specific version, so that should not conflict right?
<apteryx>Kabouik: did you pull recently? that should now be handled automatically with a warning
<apteryx>(the warning says you no longer need to provide nss-certs yourself, as its now part of %base-packages)
<Kabouik>I pullled just before the reconfigure apteryx, got no issues
<Kabouik>I'll try to just remove it from my config then
<apteryx>weird that itdoesn't work; it should warn that you have multiple nss-certs packages, and keep only the newest
<apteryx>per commit 9593750698 and c85a7e1b2b
<Kabouik>I also ran a `guix install nss-certs` afterwards just in case and it did install 3.99, so I'm not sure about 3.88.1
<apteryx>(2 weeks old)
<Kabouik>Removing nss-certs from my config.scm seems to have solved it apteryx, thank you.
<apteryx>np... though I wonder why the the safeguard in 9593750698 didn't help here
<apteryx>what did your packages definition look like before?
<Kabouik>I'm not sure I understand the question, you want me to list all my packages listed in config.scm?
<Kabouik>I had no private-channel package with nss-certs as input (let alone a specific version).
<apteryx>just the 'packages' field of your operating-system record, the way it was defined when you had the issue
<apteryx>woohoo, I managed to package a working 'sysdig'. I needed to enable BTF debug info in the kernel.
<Kabouik>I realize I have another issue since the package upgrade. Emacs now errors out with "Debugger entered--Lisp error: (void-function defhydra)". despite emacs-hydra being properly installed. Has anyone encountered that issue? I asked on #emacs but they think this may be a Guix thing.
<apteryx>Kabouik: your sourcehut pages gives me: 401 Unauthorized
<Kabouik>Right, sorry, it's private.
<dariqq>has anyone from gnome-team seen yet?
<peanuts>"CVE-2024-34397: GDBus signal subscriptions for well-known names are vulnerable to unicast spoofing (#3268) ? Issues ? GNOME / GLib ? GitLab"
<anthk_>btw, with guix, is it possible to generate an install image with all the packages being able to being installed offline?
<ieure>No :(
<jab>hey guix! I'm playing with guix on a talos II. "guix pull" failed on a nss-certs...
<jab>Is anyone aware of this bug? Shall I file a bug report?
<anthk_>jab: are you the one with the gaming channel under guix?
<jab>anthk_: nope. I'm on hardlimit.
<peanuts>"The GNU Guy - Hardlimit"
<cbaines>jab, it's a known issue
<peanuts>"nss@3.99 is really hard to build"
<cbaines>(assuming it's nss rather than nss-certs that failed to build)
<ieure>jab, Are you on Guix System?
<jab>ieure: I am on chimera linux. using the guix package manager on top.
<jab>cbaines: thanks!
<lilyp>has anyone else encountered/investigated the gnome oddity that drag and drop file icons appear to get stuck on release?
<jab>I am trying to build guix from source...again on a power 9. I did a fresh git pull.
<jab>guix shell -D guix ; make authenticate ; make
<jab>then it complains about "making all in po/guix"
<ngz>I think I found why modular TeX Live is so slow…
<jab>and it fails. it that the guix manual that is failing to build?
<jab>here is the make log:
<peanuts>"debian Pastezone"
<jab>let me try guix shell -D guix -CPW maybe that is how you are supposed to build it.
<jab>that's just sad....guix shell -D guix help2man git strace -CPW; make
<jab>seems to be working.
<freakingpenguin>rekado: How can I add to the mumi test data? The data fetched from tests/data doesn't seem to contain any patch series and it seems more complicated than "place .mbox files".
<podiki>your make log shows missing sed
<freakingpenguin>jab: guix shell ....; make should run make after you exit the shell container, I'm surprised that would work.
<podiki>i think you would still be in the shell no? (i never tried, but you only exit the shell if you give guix shell ... -- <command> and it finishes <command>)
<freakingpenguin>Pretty sure the host shell catches ; and interpret it as "run guix shell blah until it exits, then run make".
<podiki>anyway, the error pasted was that sed was missing
<podiki>could be, i don't know
<podiki>(on zsh it didn't seem to exit in my quick test but was still in the shell; and didn't run what was after ; until after I exited)
<podiki>which I guess is what you were saying
<cwebber>new blogpost about some exciting new nlnet grants, including Shepherd on Goblins!
<peanuts>"Distributed System Daemons: More Than a Twinkle in Goblins' Eye -- Spritely Institute"
<lilyp>okay, turns out it's an Xorg only bug
<lilyp>goodbye, xorg
<freakingpenguin>cwebber: Neato, thanks for sharing!
<podiki>cwebber: thanks for sharing! Typo in the forum link at the end (institute misspelled)
<cwebber>oops fixing, thx podiki
<anthk_>I'd love some sicp compat for guile at guile-studio
<bigbookofbug>hi all! i had a question about the proper method to supply commands during the initrd process. i have a btrfs partition and would like to recursively delete the "root" subvol during the boot process (attempting to create an "impermanence" setup similar to what i had used when using NixOs). I've read the documentation on initrd, and know what shell commands to apply, but am unsure where/how to apply
<bigbookofbug>them in the configuration
<bigbookofbug>would this be something i declare in the `initrd` field. the docs show an example for qemu netowrking, and wanted to know if something similar could be achieved via `(initrd (lamba (system* "args" "here")))`
<bigbookofbug>i notice there's a `gnu.load` option as well, and wonder if that would provide a better means of mounting and recusively deleting the root subvolume
<panosalevro>is there a way to prevent `guix upgrade` from ever building a package when there's no substitute? the builds take a long time and almost always fail
<civodul>panosalevro: hi! there’s the ‘--do-not-upgrade=REGEXP’ option
<civodul>can be useful in such cases
<civodul>bigbookofbug: ‘gnu.load’ is used internally to specify the “boot program”, the thing that will spawn shepherd (PID 1)
<panosalevro>civodul: i tried that but i wasn't sure if i was using it correctly. if i have 2 packages, is it `guix upgrade --do-not-upgrade=package1,package2`?
<taeltydes>any advice for getting more helpful error messaging (formatting, specificity?). I'm trying to write some services for matterbridge but on system reconfigure the invalid specifier itself does not specify which field.