IRC channel logs

2022-09-07.log

back to list of logs

<pkill9>.sneek tell unwox no, in my earlier tries of runnign with pipewire I wasn't runing wireplumber
<sneek>unwox, pkill9 says: no, in my earlier tries of runnign with pipewire I wasn't runing wireplumber
<pkill9>.sneek later tell unwox no, in my earlier tries of runnign with pipewire I wasn't runing wireplumber
<sneek>Okay.
<pkill9>.sneek later tell iska does pulseaudio get run for you when you start an audio application if you aren't already running pipewire?
<sneek>Got it.
<pkill9>ty snek
<pkill9>ty sneek
<pkill9>ah I disabled autospawn in pulseaudio's client.conf
<pkill9>now I wonder if pipewire is supposed to autospawn, I don't think it is
<pkill9>especially as it seems like it needs 3 parts
<pkill9>for pulseaudio applications
<pkill9>so probably better handled as a service
<pkill9>why does pipewire not contain pipewire-jack?
<pkill9>ah nice there is no pipewire-jack binary
<pkill9>atleast not in the archlinux package
<Kirtai>Hi, just installed and the update is failing to build rav1e, Any thoughts?
<Kirtai>Seems to be missing nasm for some reason
<pkill9>does anyone run a wayland compositor as a home service?
<apteryx>sneek: later tell civodul re gnome importer: gedit: updating from version 40.1 to version 43.alpha -> perhaps too aggressive now? :-/
<sneek>Okay.
<Kirtai>Seems the solution was to throw more memory at the VM...
<florhizome[m]><pkill9> "does anyone run a wayland..." <- rde people do. Probably the best way to set up env vars etc
<abcdw>rekado: Thank you for the fix!
<jpoiret>pkill9: pipewire isn't supposed to autospawn, and neither pipewire-pulse
<abcdw>rekado: Thank you for mentioning it, yes, yas snippets no longer found. They were moved to subdirectory, fixed it in https://git.savannah.gnu.org/cgit/guix.git/commit/?id=39b8f85a6e
<ulfvonbelow>question: has anyone here ever tried to directly init an EFI system from a non-EFI system?
<ulfvonbelow>I think the process is a bit broken at the moment
<ulfvonbelow>it seems grub-install needs to be passed --no-nvram and e.g. --target x86_64-efi
***silver.libera.chat sets mode: +o ChanServ
***silver.libera.chat sets mode: +o ChanServ
***silver.libera.chat sets mode: +o ChanServ
***wielaard is now known as mjw
<efraim>how do I use this-package-input using not the 'out' output of the package
<abrenon>hi guix
<efraim>I don't believe GUIX_EXTENSIONS_PATH is documented in the manual
<abcdw>efraim: Unfortunately it's not and probably not many people aware of it.
<rekado>it’s not documented because we hadn’t decided if it was a good idea.
<rekado>now that it’s used by at least two extensions out there we might as well commit to it.
<efraim>I'm playing with my proof of concept one for 'guix send-email' -> 'git send-email' and I was looking at the patches again
<efraim>bug 57317
<efraim>I'm not sure about removing the search path from guix
***Dynom_ is now known as Guest9264
<efraim>with just the first one I don't have $GUIX_ENVIRONMENT/etc/profile sourced inside of 'guix shell'
<efraim>ok, with the first patch adding GUIX_EXTENSIONS_PATH to the generated profile and then installing my extension instead of just entering a guix shell, after 'bash -l' it shows up with 'guix help'
<pkill9>florhizome[m]: who are rde people?
<ulfvonbelow>TIL: libgit2 doesn't support the 'dumb http' protocol
<muradm>hello guix
<unwox>hello!
<sneek>unwox, you have 1 message!
<sneek>unwox, pkill9 says: no, in my earlier tries of runnign with pipewire I wasn't runing wireplumber
<muradm>gnu/packages/linux.scm:888: https://paste.rs/Ff0
<muradm>what is the meaning and purpose of build variable
<muradm>it is not used anywhere
<rekado>muradm: I bet it can be removed then.
<efraim>it's obvious after the fact, but running 'guix weather' regularly against my main machine made my guix-publish log blow up in size
<efraim>"blow up" compared to previous values
<muradm>rekado: not important but just as reminder 57642
<ulfvonbelow>is gdm broken for anyone else?
<ulfvonbelow>gdm-session-worker: GLib-GObject: g_object_unref: assertion 'G_IS_OBJECT (object)' failed
<efraim>ulfvonbelow: clear your mesa cache, in ~/.cache/mesa_shader_cashe, and the one fro GDM in /var/lib/gdm/.cache/mesa_shader_cache (i think)
<ulfvonbelow>but this is a fresh install
<apteryx>ulfvonbelow: is this really a critical error? I've seen it before.
<antipode>antioxidant now supports more (all?) forms of tests
<apteryx>antipode: neat!
<Andronikos>I want to compile GNU Artanis locally but I get "In procedure dlopen: file "libnss3.so", message "libnss3.so: cannot open shared object file: No such file or directory"". Using "guix shell autoconf make guile pkg-config nspr nss". What do I do wrong?
<apteryx>patch the dlopen call locally, or set LD_LIBRARY_PATH
<nij_>Anyone tried guix in a VM running over a macbook m2?
<nij_>(I mean the full guix OS.)
<apteryx>I doubt so
<nij_>Why not? I'd like to learn some reasons.. (m a noob in CS)
<apteryx>GNU stands for user freedom, while Apple stands for profits and lock-in is one.
<apteryx>oh wait, I missed the "in a VM" part
<apteryx>that should be doable
<apteryx>assuming QEMU runs on Mac (which I'm guessing it does), it should be very simple; just by downloading the QEMU image from the Guix website and invoking QEMU with it.
<Andronikos>apteryx: How do I set LD_LIBRARY_PATH? I mean how do I get the correct value.
<apteryx>actually did you use the 'gcc-toolchain' package to build the package?
<apteryx>OK, but looking at the above error it won't help because of dlopen
<apteryx>Andronikos: try 'LD_LIBRARY_PATH=$GUIX_ENVIRONMENT/lib/nss artanis ...'
<nij_>oh ok
<antipode>Downside of supporting more forms of tests is more test failures to resolve ...
<dominicm>Has anyone had issues with guix's libreoffice? I can't install extensions, use spellcheck, etc.
<rekado>dominicm: this sounds plausible. All packages in Guix are installed to /gnu/store, which is not writable. It’s possible that libreoffice attempts to install extensions to its prefix directory in /gnu/store.
<rekado>perhaps it can be patched to prefer a user-writable location (e.g. in ~/.local/share)
<dominicm>possibly yeah, that was my first thought but the error I got did not seem indicative of that. I'll try to look later.
<Andronikos>apteryx: Do you mean "LD_LIBRARY_PATH=$GUIX_ENVIRONMENT/lib/nss"? Don't understand what you mean with "artanis ..". Did try the first and it still does not find the lib.
***mark__ is now known as mjw
<PotentialUser-8>Hello. Are there any Schemers / Guixers in Strasbourg? I moved to that wonderful place recently and I look for more experienced people as mentor. I can offer beer (me) and Sushi / Kimchi (my wife)
<dlowe>I visited once. It's quite a beautiful city.
<PotentialUser-8>It really is.
<antipode>Could this build be restarted: https://ci.guix.gnu.org/build/1370400/details (network error)?
<panosalevro>PotentialUser-8: im studying remotely at the uni of strasbourg and might visit some time
<PotentialUser-8>panosalevro: Sounds good. I will stay here for around three years, maybe longer
<PotentialUser-8>I sent you a PM. 👍
<apteryx>Andronikos: I meant to use the LD_LIBRARY_PATH=$GUIX_ENVIRONMENT/lib/nss as a prefix to the artanis binary you built
<apteryx>or you could 'export LD_LIBRARY_PATH=$GUIX_ENVIRONMENT/lib/nss' to affect your shell session
<antipode>Nevermind, the latest commit rebuilds all the antioxidated crates anyway.
<shcv[m]>there seems to be a bug in pypy? `guix shell pypy` tries to build pypy-7.3.5, and fails with `Unbound variable: out` in the install phase
<apteryx>shcv[m]: it's a bug. I'll fix it (s/out/#$output/)
<apteryx>should be fixed with 9ec4233181, although it'll take some moments for the build farm to build it
***civodul` is now known as civodul
<phodina[m]>Does somebody know how to call `cargo` within from `meson`?... (full message at https://libera.ems.host/_matrix/media/r0/download/libera.chat/8f48fd88c7061084d39b585a725e7db35baba87c)
<apteryx>shcv[m]: hm, the pypy build failed with a different error now: bin/pypy3.debug: error: offset + size of segment 5 (type 1) exceeds total size
<davidl>efraim: reg. the bin output for #$this-package-input pcre (or some other package) - I think it's not possible. If you want the bin output you should have been able to use maybe: (inputs (list some-package `(,pcre "bin") some-other-package ))), but it doesn't work because I think the implementation is limited for the new format. I get issues with the hash when #$this-package-input pcre is used even though the inputs list appears to be accepted.
***hexology- is now known as hexology
<vldn>can i loop over a list and define functions with it?
<vldn>e.g run the define command in a recursive loop?
<antipode>sneek: later tell efraim: #$(gexp-input (this-package-...) "another-output")
<sneek>Got it.
<antipode>sneek: later tell efraim: Or #+, for native.
<sneek>Will do.
<antipode>vldn: You can use 'define' inside a loop (e.g. for-each), to define a local function inside the lambda of the for-each.
<antipode>That might not be the behaviour you want.
<antipode>For some loopy behaviour, a macro could be useful.
<antipode>But I do not know what goal you are trying to achieve, so I can't tell.
<antipode>To be more concrete, I'd like some concrete example (code, even if it doesn't compile)
<vldn>just ideas atm
<vldn>i want to write some kind of parser, that i have list with functions like "init" "player" "var" etc. with voids from another language and loop over it to generate somekind of language parser
<vldn>write guile code and after execution it compiles to this other language and outputs to a file
<vldn>maybe there is some easy way to do it
<vldn>functionwise the same as a html generator
<apteryx>seems libgit as used by guix doesn't honor GIT_ASKPASS, fails in CI when using a <git-checkout> object.
<apteryx>CI being Jenkins
<apteryx>the error is: error authenticating: no auth sock variable