***dutchie_ is now known as dutchie
<sebboh>Hi, I have just finished manually mounting all the filesystems from my old PC in the right places with the right options. So I've got this /etc/mtab that lists all that. In inferior distros, I could copy that to /etc/fstab and reboot and everything would be fine. (Maybe I'd edit it a bit, to use labels.) Does any method exist to generate a (file-systems ...) form from the input I have available? (or several <atw>mpd service users - do you find that having other things running that use pulseaudio -- including pavucontrol -- cause mpd to be unable to open the audio output? <nallastri>hello guix! How do I activate nouveau drivers on the Guix System? I tried guix install xf86-video-nouveau , but after that glx-info still didn't show the discrete GPU. <buhman>nallastri: is this optimus hardware? <buhman>ah, well, do you know how to use DRI_PRIME ? <nallastri>is it related to bulmblebee? I used to use that back on Arch and Debian <nallastri>it looks quite doable i'll try right away, thanks again <Sharlatan>Hi, where I can read about how to write services for Guix? I'm interested to make one for Emacs server. <Gamayun>Sharlatan: I should get around to setting one up as well at some point. Do share it if you create one! ;-) <PotentialUser-95>I was looking for /root/.config/guix/current/lib/systemd and I could not find it. Turns out it was there when I installed guix under current-guix-1-link but not since second version on wards. Any idea hat could be the reason for this? <PotentialUser-95>I was looking for /root/.config/guix/current/lib/systemd and I could not find it. Turns out it was there when I installed guix under current-guix-1-link but not since second version on wards. Any idea what could be the reason for this? <nckx>Consecutive pastes, too. <nckx>PotentialUser-95: Er, no, not really. <ItsMarlin>if i try to feed it something, it says it couldn't find the graph file <ItsMarlin>i look at the gui and it can't see any files apart form the directories <nckx>On my Debian-derivative system, /root/.config/guix/current/lib/systemd existed, now I'm pulling as root for the first time, but that will take a while. <nckx>It didn't exist in my regular user's .config/guix/current, though, which is up-to-date. <PotentialUser-95>nckx: Yes same with me. My regular user has only guix and guile folder under lib. And so is under root. I was rereading installation steps and it refers to systemd folder for guix-daemon service and I could not find it. This is when I found out that in the latest version it is no longer there. Nothing seems broken due to this but I still I am tryin <nckx>Core 2 Duo with busted power supply, ‘current CPU frequency is 629 MHz’, truly a joy to guix pull on. <FennecCode>Does anyone know how to make Emacs show new packages from a pull in guix-search? 🙂🤔 <criticalcat>Does "guix package -i blah" download and install a substitute? <bavier`>criticalcat: 'guix install' is an alias for 'guix package -i' <criticalcat>What does this mean? guix pull: error: symlink: File exists: "/var/guix/profiles/per-user/root/current-guix" <nckx>It's a bug in Guix I've just hit myself, but I haven't found out what it means yet. <criticalcat>It was following a failed "sudo guix install ..." for me. <str1ngs>it means, that it the root exists. so somewhere guix lost track of the profiles somehow <nckx>Yes. It does this with disturbing regularity for at least the past year, it seems from a short search :-/ <str1ngs>I ideally, you can try and figure out which symlink it's trying to use and remove it. but that has effects on the GC <str1ngs>probably since its the guix profile, not too much of a an effect. but it's something to consider <criticalcat>User A and root have symlinks in /var/guix/profiles/per-user/ but not the user B <criticalcat>What does it mean to point the dead link current-guix-0-link to the target pointed to by one of the other profiles? <nckx>s/dead/broken/, or you'll confuse easily confused old people like me. <nckx>criticalcat: What do you mean by other profile? Other user, or other generation of the same? <nckx>(I thought you meant the HTTP link above and was trying fruitlessly to break it.) <nckx>OK, so I managed to fix it here but not in a way I'd recommend. I removed /var/guix/profiles/per-user/root/current-guix and ran `my user's guix` pull as root. <nckx>That seems to have worked but it's also probably the long way 'round. <criticalcat>Suppose /var/guix/profiles/per-user/A/ has "current-guix-1-link -> /gnu/store/..." and "current-guix -> current-guix-1-link" and similar for guix-profile, but for B, there is only "/var/guix/profiles/per-user/B/current-guix -> current-guix-0-link" and that is a broken link. What would it mean to make "current-guix-0-link" target the same thing that A's "current-guix-1-link"? <nckx>criticalcat: I don't think it would harm anything, but I don't know if it would fix the error. <nckx>Actually, what you're suggesting is what I did but without the intermediate re-symlinking. <nckx>I deleted root's guix (rm /var/guix/profiles/per-user/root/current-guix), then ran my user's guix as root (su, not sudo). That's what your method would do. I *think*. <nckx>But maybe your symlink would still be in the way, causing the old error. <criticalcat>I tried removing the symlink and running as self, perm issue, changed dir to be ... I think I am too tired to have it make sense. <criticalcat>I am processing "ran my user's guix as root" and I think I get it. <nckx>criticalcat: Well, I can confirm that ‘rm /var/guix/profiles/per-user/root/current-guix; su; /home/criticalcat/.config/guix/current/bin/guix pull’ would work. <nckx>But if you're tired, maybe don't try that today 🙂 <criticalcat>Giving it a try, although not going to interrupt the thing that I just did which is slightly different, sudo su -; guix pull <nckx>Guix is really remarkably hard to break. Barring corruption or serious fuckuppery, you can usually get back to a working system in, like, 3 commands. <nckx>criticalcat: No, don't interrupt, who knows. <nckx>criticalcat: Guix System's sudo defaults to ‘sudo -E’ behaviour, many other distro's doesn't, it's sometimes easier to avoid sudo altogether than to figure out which parts of your user's environment leak through when. <criticalcat>I probably should clarify I am running guix package manager on Trisquel 8 <nckx>That said, I always thought ‘sudo su -’ was equal to ‘su -’, just a different password 🤷 <nckx>Oh, see, I *think* that's one of them distroes that doesn't env_keep by default, but don't take my word on it. <nckx>Or ‘Defaults env_reset’. <criticalcat>Maybe I need a "it feels like you just sudo-ed me on a non-guix system, one advantage of guix is you don't need to sudo, just be yourself" message. <nckx>Actually, it's called Guix System now, so "non-Guix-System system". <nckx>So you can say you run a non-non-Guix-System system when you switch! <buhman>kkebreau: I just tried core-updates, and the boot process hangs <nckx>buhman: Could you send that to the list as well (guix-devel, I don't think bug-guix is appropriate for core-updates)? It risks getting lost here. <buhman>I am intimidated by mailing lists <nckx>Don't feel obligated. It's just that #guix is particularly empty this time of (Euro)night. <tune>the repo for wlstream was deleted <tune>wf-recorder may be a similar program but is not yet packaged <nckx>tune: Can the source still be downloaded from one of Guix's content-addressed mirrors? <tune>not sure what happened, and I haven't tried. I seem to have it already installed, although I didn't get around to using it yet <buhman>nckx: hmm, do I need to be subscribed to send? I don't see my message in the august archive <nckx>buhman: The list is moderated by humans, if this is your first mail to the list you'll have to wait for one of them to wake up 🙂 *nckx would answer their own question about wlstream if ‘guix build wlstream --source --no-substitutes’ didn't insist on BUILDING PYTHON. <buhman>better that the mods tell me I did something wrong in posting than the whole list recieving my fuck-up, so that's fine with me. <nckx>--no-grafts was the magic word. <nckx>tune: The ‘content-addressed mirror at berlin.guixsd.org’ has wlstream source, so it's basically safe forever unless it badly breaks or a CVE gets found. <tune>that makes me feel better about learning to use it then <nckx>(Or rekado forgets to back up berlin.) *nckx wonders how/if back-ups are actually handled. Most of berlin is ‘cache data’, but these sources can suddenly become valuable data. <buhman>tune: are you using sway? I'm interested in reading your config.scm if you're willing to share. <buhman>nice, I had no idea sddm could do this <tune>yeah sddm has been pretty nice <tune>I was using gdm before but it wasn't able to launch sway <tune>so I was just launching from a tty until someone here helped me remove gdm from %desktop-services so sddm could work <rekado>nckx: there are no backups for Berlin. <rekado>it’s just big, somewhat redundant storage. <sammich>guix environments are my favourite thing. It makes hacking on peoples projects so easy <roptat>the store is append-only (well, except for the gc) <roptat>usually, if there's something you want to change, you create a new package, or a service or something else, that appends to the store <roptat>tell us what you want to do, and we'll help you find the best way to do it <Lukas4454>i am sending mails with some sound driver developer, for tests i need to change "usr/share/alsa/ucm/chtnau8824/HiFi.conf" <Lukas4454>the original driver comtains 250 lines to configure the soundcard <Lukas4454>clearly the ALSA developers did something wrong <Lukas4454>and i need to fix it so i can tell the devs how to fix it <buenouanq>Icecat and Calibre won't build for me right now <buenouanq>I've never had this problem with guix before, what do? <roptat>Lukas4454, I'm not sure how to do that... you can boot from the installer and mount everything rw, but note that changing you store will break guix assumptions <roptat>basically, you're voiding your warranty :p <Lukas4454>i would need to submit the change to alsa right? <roptat>if the file comes from them, that's where it should be fixed <roptat>we can always try and apply a fix ourselves while waiting for an upstream release, too <Lukas4454>because i would like to change that file first and check if that fixes it, and after that write to the alsa-devel mailList <Lukas4454>in the worst case i reinstall Guix with the same config to regain warranty <mbakke>buhman: The boot hang is due to entropy starvation; moving your mouse or mashing the keyboard will allow it to proceed..... <nixo_>Hi guix! Does anybody know a nice way to run guix in a CI? I've a NixOS machine running drone-ci (it does support docker), I want to build my software using guix. Any suggestion? <pinoaffe>efraim: okidoki, I found some other similar packages and am mirroring them <rekado>nckx: just asked: the MDC firewall *can* track domain names after all. I’ll ask them to punch holes for your build nodes. <rekado>hmm, ibus libpinyin isn’t working any more :-/ <rekado>bmwiedemann1: do you need a big file containing *all* packages, or would you be satisfied with one file per package? <rekado>(we don’t have that either, but I suppose it would be easy to add) <bmwiedemann1>rekado: one file would be easier to pull with wget. I also do git clone for some projects, pulling many small files. <rekado>bmwiedemann1: we recently migrated off GNU infrastructure, so a few things have changed. I’ll take a look at this later. <bmwiedemann1>would be good if this does not invent something new that bitrots away because nobody else uses it <rekado>bmwiedemann1: the clients do something weirder due to the way Guix works. <rekado>bmwiedemann1: nah, much weirder. <rekado>we have this infrastructure already for generating web pages that show available packages for an arbitrary version of Guix; should be easy to extend it to generate that package.json. <rekado>I’d rather do this right the first time, so it may take a little longer. <rekado>I’ll look into this tonight though. <Lukas4454>i posted to the mailinglist of ALSA, noone answers :( <minall>How do you use guix environment?, and why? <dongcarl>minall: `guix environment --ad-hoc cowsay -- cowsay 'Hi, miniall!'` <atw>hello minall ! I do as dongcarl does a lot <dongcarl>but there's a lot more that `guix environment` can do! <dongcarl>for example, maybe you learn about this newfangled language called python <minall>Yes, guix environment seems to be amazing, but I don't see all of it <dongcarl>will modify your environment so that you have python in your path! <minall>For example, If i were to make a code on python, why would I use guix environment, instead of just creating the script, for example with emacs? <minall>When we use guix environment a path is created? <dongcarl>minall: It's more modular, as in, you don't modify your default environment, you can just "try it out" <minall>So, is it to try instead of installing first? <minall>If I want to try a package, but not to install it, can I use guix environment? <dongcarl>minall: Well, it depends on your interpretation of the word "install", but yes! <dongcarl>it won't install to your default `$PATH` <minall>It will install on another place, so if I like it I should install it fully? <dongcarl>It installs to `/gnu/store/blahblah-profile`, and then prepends that to your `$PATH` if that makes sense? <minall>For example, I want to try a desktop environment... can I use: guix environment mate, and I'll have mate installed and ready to use on something like a VM? maybe I'm thinking wrong lol <minall>Yes, but it is not install on my path with guix environment <dongcarl>minall: Are you familiar with how `$PATH` works? <minall>Yes, /gnu/store, if I install something, the package and their dependencies are install in /gnu/store <minall>But with guix environment, the package that I install using guix environment will have his directory on the /gnu/store, but separated? I mean, not installed on the system? <dongcarl>When you type something like `cowsay 'Hi, miniall!` on your command line, your shell will look for an executable named "cowsay" under your `$PATH`, which may be something like `/usr/local/sbin:/usr/local/bin:/usr/bin` <minall>Yes, but of course, with different paths on guix <dongcarl>Well yeah, so technically, guix always installs to your system, but whether or not it remains in your $PATH is your choosing <minall>Yes, totally, so we can assume that whats in our $PATH is installed on the system fully right? <dongcarl>minall: Sure, technically you can have an invalid directory in your $PATH, but for most purposes yes <minall>Yes, so guix environment is to install other packages that I don't want YET to be on my system? <dongcarl>you don't want YET to be in your _default_ $PATH <minall>Thanks, that was very aclaratory!! <minall>It can be used to to -develop code without polluting your environment-, have you used it like this ? <dongcarl>minall: In Bitcoin Core, we use `guix environment` to create an environment where all the tools needed to cross-compile is available <dongcarl>so that the tools don't pollute the developer/maintainer's base system, but are still available when they need it <minall>So for example, I'll do a python project, that uses several dependencies, but I don't want -for some reason- to have installed those dependencies, I'll use guix environment for that? <dongcarl>minall: right! or you can use virtualenv or something python-specific <minall>Yes! i readed that on the manual <minall>Python can have virtualenv, but guix can do it with any language! <dongcarl>Guix: virtualenv but like for all the things^TM <minall>Thanks dongcarl, you're very helpful <minall>btw is there a way to test a config.scm after installing it? <minall>Like booting the machine, or something like that? <dongcarl>minall: Oh... I don't use GuixSD, so someone else can answer that <dongcarl>I only use Guix as a package manager on top of some other distro <minall>You can't run xorg twice right? I'm thinking something a little insane like: guix environment mate, then a window pops out when I can try everything about it, and if I like it, install it! <minall>That's a little insane but, welp <minall>Well debian and some other distros enabled by default Wayland now, so that's something, seems that Wayland is not a simple project to compare and leave <dongcarl>minall: Yeah never tried, but good luck! <mbakke>dongcarl: that's one side of the issue <joshuaBPMan1>this is really weird. I just tried starting my booting my macbook 7,1 with the archlinux usb stick that I just created... <mbakke>the other part of the issue is that currently, on the core-updates branch, glibc is the *only* "system header" configured by default <joshuaBPMan1>the error message I got was "initramfs unpacking failed: XZ-compressed data is corrupt. <mbakke>dongcarl: so packages that use -Werror may get build failures due to warnings in included headers <joshuaBPMan1>perhaps I need to watch a youtube video explaining how to make a bootable usb stick image. Because I seem to be doing it wrong... <dongcarl>mbakke: Is there documentation of best practices regarding `-isystem` usage? Or just avoid like the plague? <dongcarl>for context, I'm reviewing something in Bitcoin that wants to add it to silence warnings from dependencies, while being careful not to trigger the `include_next` thing <mbakke>dongcarl: -isystem IIUC is mainly to ignore warnings from included headers. Well-behaved packages set it for each of their dependencies. <roptat>minall, you can test a config.scm by building a VM with guix system vm or vm-image <mbakke>dongcarl: As long as you don't add glibc with -isystem you should be fine. <mbakke>dongcarl: np! it's good to talk about these things <mbakke>It would be great to make C{,PLUS}_INCLUDE_PATH work again with GCC7 and later so that we don't have to manually set them in many packages. <mbakke>But I haven't found a good solution yet. <mbakke>perhaps we could patch GCC to ignore headers that look like glibc headers :P <dongcarl>mbakke: Oh have they stopped working? Could you point me to the ticket? *bavier slogging through test failures in their 'rpmlint' package <minall>roptat: Thank you! i didn't knew about that command! <minall>roptat: It seems that guix system vm or vm-image doesn't exist, maybe I'm doing something wrong? <nothingmuch>i was trying to access an lvm drive on guixsd and was surprised to discovered it's not available - is there some reason underlying this? i didn't find much online <jonsger>nothingmuch: Support for the Logical Volume Manager (LVM) is missing. <- section 3.1 of the manual <Lukas4454>i think i install windows on my PC and sell it :( <roptat>minall, it should exist, have you tried "guix system --help"? <jonsger>Lukas4454: you could also use guix on a foreigen distro like Debian or openSUSE <nothingmuch>jonsger: yep, i found that much, and also a few discussions of this over the years and someone's personal package definition, i just wondered if it's because nobody ever found the time or because there is a reason behind this <jonsger>nothingmuch: I would say that no one invest the time. I further guess that it's not trivial, because if it's trivial we had support already :P <minall>It does, maybe I was writing something wrong, I'm trying guix system vm now! <bavier>Lukas4454: for small, cheap notebooks I've had success with the asus eeepcs <mbakke>dongcarl: For GCC7, I had to go back to only using CPATH, because C_INCLUDE_PATH would pull in glibc as a system header and create all sorts of problems. <dongcarl>mbakke: This is supposedly an upstream gcc problem, yes? <dongcarl>Will need to take some time to look at this <dongcarl>Have cropped up in my experiments here and there <minall>Welp, I maded an -vm- of my config.scm, it finished, but how do I start the vm? <minall>I maded it using guix system vm /etc/config.scm <mbakke>minall: try running the returned /gnu/store...-run-vm.sh script <mbakke>you can pass it command-line arguments that will be passed onto the Qemu process, i.e. ...-run-vm.sh -m 1g -smp 2 <minall>I got a QEMU window, But I seem to have an error, in procedure mount: no such file or directory <minall>Is that an error of the VM? or when I restart I'll have that error? <enderby->hi, i'm using go from guix to build exercism, and I'm getting "cc1: error: /home/username/.guix-profile/etc/profile/include: Not a directory". I've never used go to build anything else, so i'm not sure if this is a guix thing or go <Lukas4454>i need to add this line load-module module-udev-detect tsched=0 <mbakke>Lukas4454: does the file exist from before? <mbakke>Lukas4454: I mean, does it not work to just create /etc/pulse/default.pa ? <nixo_>Lukas4454: does not ~/.config/pulse/default.pa work? <Lukas4454>forget it, i install wi dows on there and sell it <buenouanq>icecat is failing to build and it looks like a problem with ...inbdr6-rust-1.19.0 <atw>successfully built /gnu/store/9kq0nhppmnk8krihvk4wwh9ibwbqa0jc-icecat-60.8.0-guix1.drv on guix --version => 000faac0a23542285ab0f2999b84a7001eb3068e <buenouanq>atw: I'm on an old thinkpad, the i686 guixsd image <atw>hmm, I am on 64bit. It could matter but I am not sure <rekado>the dependency has been failing to build for a long time, it seems <atw>that link is broken for me, seems like I can't even ping the host <rekado>yeah, that’s just Cuirass saying nope once in a while. <rekado>I feel motivated to fix it, but I don’t actually have time to do that. <sebboh>When I install samba, am I supposed to get some herd service to go with that? For smbd? *pkill9 created a build system for automagically wrapping games created with the LOVE framework <OriansJ>pkill9: Creative Commons NonComercial <OriansJ>as You may not use the material for commercial purposes; it does not comply with Free Software Guidelines and thus can not be in Guix main but like blobbed Linux, there can be guix package definitions for it <alexanderbarbosa>hey, I know we can split up strings with s, but what about spliting up symbols? eg: ch01.xhtml#ch1 --> ch01.xhtml <desttinghim>I am trying to package a library, and I think I've almost got it <desttinghim>But I am getting this error: `error: %outputs: unbound variable %outputs` <desttinghim>I read through the packaging tutorial and am using the method shown to try and pass the prefix in <pkill9>desttinghim: changing the quote (') to a quasiquote (`) in the make-flags line fixed the error for me <pkill9>(though there's another error, a /usr path is hardcoded it looks like <desttinghim>Dang, that was the error I was hoping to resolve with the prefix