<unmatched-paren>but you should be aware that anything that runs in a graphical environment won't work, there's a simple patch to fix it, but it hasn't been merged yet <jab>unmatched-paren: I had used automatic login to virtual console before...but it's not working now. I've filed a bug for it. <Tirifto>Hello! Should I ever be getting a derivation built for a package for which there’s a substitute available? <oriansj>unmatched-paren: interesting they don't show up when I did guix package -s greet <oriansj>unmatched-paren: but I do appreciate that pointer *unmatched-paren had better send wlgreet <oriansj>unmatched-paren: I'm guessing one could just replace (service slim-service-type) with (service agreety-service-type) ?? <unmatched-paren>obviously you'd just replace that greetd-wlgreet stuff with another greetd-terminal-configuration for vt-1 <pkill9>does anyoen ahve examples of guix home configuration files? <jab>Hey raghavgururajan! <jab>I just thought of something pretty crazy.... <jab>OpenBSD supports an autoinstaller...Apparently you can install OpenBSD and specify via autoinstall, what services to enable... <jab>then you could use libguestfs, to change files and such.... <jab>guixRS channel could potentially support guix system vm openbsd ..... ? <jab>I found a blog post talking up it. Seems pretty cool. <haruno>Hey guys how is it going? I am new to guix and am at fresh gui installation. Got to partitioning part then got errors, from the log I would assume it has to do with my ethernet connection. So how do I acess network gui during setup, because at info iI couldnt find keybind :D <jts>So I'm working on a package that needs stddef.h at runtime, but it can't find it. since the header in question is in linux-libre-headers, I tried adding that as an input, but no dice. I'm guessing the fact that stddef.h is imported indirectly through stdio.h may be part of the issue. anyone know how I might work around this? <polyex>ty guix for making best OS and declarative system cm that exists <nckx>haruno: That doesn't sound right: a networking error breaking the partitioner? Hm. Could you share the log (a photo is OK if you can't share the text)? It's probably better to report this to bug-guix at gnu dot org, as you can tell the IRC channel is very quiet this time of (European) night. <pkill9>oh it's probably the xonotic-data <nckx>pkill9: Because that's the version of the game you're running, and you didn't update that. I wonder why --with-source=xonotic-data=http://dl.xonotic.org/xonotic-0.8.5.zip doesn't work (I get the dreaded ‘had no effect’ warning). <haruno>nckx: Thanks for reply, I am new to coding and linux in general so that was my earlier assumption but now I think error starts in ice-9/eval.scm guile-user misc error the usb from which I am installing is apperently busy. Any ideas? :D <nckx>You beat me to it: -data is the game, xonotic the engine. <nckx>haruno: Ideas to do what, exactly? <haruno>what could be causing the error since its fresh install and all :D <pkill9>nckx: maybe because xonotic-data isn't a public package? <nckx>I don't see what freshness would have to do with it. It might really help if you could share the error. The only thing we can say for sure is that without the error, we can't say. <nckx>pkill9: Nope, tried making it public with no effect. <nckx>It's probably something silly like that, though. *nckx heads off to bed, because xonotic-data won't finish download before I smash the keyboard with my sleeping face. Good luck to both o' ya. <yuu[m]>`use-package-modules` is just a macro for `(use-modules (gnu packages package-name))`? same for use-service-modules? <yuu[m]>/mnt/etc/guixs/yuuguix.scm:56:2: error: (bootloader (bootloader-configuration (bootloader grub-efi-bootloader) (targets (quote ("/efi")))) (keyboard-layout keyboard-layout)): invalid field specifier <yuu[m]>hm it was just keyboard-layout which should be in bootloader-configuration. <apteryx>guix install emacs emacs-paredit :-) <yuu[m]>apteryx: oh yeah gonna install thanks for the tip 😊 <yuu[m]>but also guix error message could be a bit more helpful <Lumine>Hi, I have a question for guix home savants, regarding managing a specific shell, fish. It isn't documented like zsh and doesn't have an rc file, so how would I go about configuring it? <vivien>I didn’t know that fish was something you could use as a shell <unmatched-paren>instead configure your terminal emulator to start it instead of ${SHELL} <unmatched-paren>and you need to configure it with guix home to make sure guix env vars are set up <Lumine>unmatched-paren: thank you! I think I have a better idea now <Lumine>I already know how to set up fish in my main config <paul_j>Quick question - quite often I receive two copies of emails in the different guix mailing lists. Is this something others are seeing, or do I need to dig a bit deeper into my setup? <paul_j>It's obviously not really an issue, but I am intrigued! <paul_j>Ok, so I will consider it a feature! :) <unmatched-paren>Probably some client-side issue, though, because it's almost always tobias whose emails are duplicated :) <Lumine>unmatched-paren: follow-up question, where can I found out what my environment variables are like you have them defined in guix home? <Lumine>The %env-vars are defined higher up <Lumine>I don't know what my env vars are <Lumine>So it's need to use basis then, gotcha. I just thought the fish service requires specific env vars <Lumine>Thanks, I'll tinker with fish after I get food delivery and report back :) <nckx>unmatched-paren: What's different about my mails? <unmatched-paren>I think it has happened for others before, but it's almost always your mails. <nckx>I just use mu4e like most people. <nckx>Could you give an example Message-Id? <unmatched-paren>Actually, looks like your most recent mails don't have that problem. <unmatched-paren>nckx: Here's the 'Resent-Message-Id': <handler.57127.B57127.166021279618473@debbugs.gnu.org> <unmatched-paren>nckx: Ah, yes, this is one of the messages with the "Sent on the go" signature. <nckx>Hmm, searching for that doesn't show anything. Is there no Message-Id? I thought that was an RFC violation, but apparently not. <nckx>I'm not sure what to do about that, I'll check K9's settings later. <nckx>(Odd that it does have a Message-ID there.) <nckx>I don't know what to do. It's not like I CC'd both you and the bug, which is what I thought might be happening, or something similar like that… <nckx>I meant: that was my hypothesis for why you were receiving dupes, because I do occasionally reply to such threads, and do CC you. <nckx>This is not such a mail so me confused. <nckx>One copy has X-Spam headers and one doesn't. <pkill9>nckx: did you find out why xonotic-data package doesnt get transformed? <pkill9>it bothers me that it isnt being transformed <nckx>Surely this can't just be because I ‘To: NNN@debbugs, …, bug-guix@debbugs’. Loads of people must do that. Also, debbugs *explicitly handles* that when checking whether to open a new bug number, so it knows it's harmless… <nckx>pkill9: No, but it's not like I've been investigating since we last spoke. I just logged on. <Not_Leader>i'm in the xonotic irc so i got thrown off for a second <Not_Leader>is this something involving xonotic–data not updating? <nckx>No, it's that --with-source=xonotic-data=URL does nothing. <nckx>Explicitly so, Guix will merrily point it out. <faust45>i am trying to setup guix code navigation in emacs, but have no success <faust45>do any knows how to navigate guile code in emacs? <nckx>pkill9: You can ignore my grumbling above. I think this is ‘by design’ & package transformations just aren't made for this. Just writing your own trivially inherited xonotic-data + xonotic packages probably is. <nckx>I don't think I've ever had a use case where package transformations *did* apply, so I've not used them. I think ever. <nckx>They address a very narrow use case, probably well. <nckx>If --with-latest= worked here, it would probably work here. <nckx>Is there a way to {white,black}list version numbers in properties? (With a regexp/lambda/…) <mroh>Love, Peace & Guix! Good morning! <mroh>faust45: I think, geiser is best for guile. But, if "only" navigation is required, perhaps something like dumb-jump is easier to setup. <faust45>@mroh: thanks! with geiser i am able jump to module definition, but coldnt jump to func def <faust45>mroh: jump to definition start working only after i run geiser-compile-current-buffer <faust45>mroh: but coldnt compile config.scm got error no code for module (my services kmonad) <unmatched-paren>faust45: You'll need to add the parent directory of the `my` directory to the load path <alethkit>How does one connect the Guix REPL from Emacs-Guix to a given scheme module? I am trying to set up Guix package autocompletion when editing service modules. <lilyp>to be fair, all variable autocompletion is like that tho <lilyp>don't forget to add a json rpc intermediate layer <alethkit>Let me try and rephrase it: I want geiser to connect to the Guix REPL such that "guix-repl-buffer" appears in "geiser-repl--repls" <alethkit>That way, Geiser's company integration works with the existing Guix repl without having to recompiler everything <alethkit>Do I really have to set up a remote Guix REPL and have geiser connect to that? It seems a bit roundabout. <lilyp>alethkit: wouldn't a regular advice around the spawning function work? <alethkit>lilyp: There must be some sort of hook to modify the load path, right? <lilyp>I'm not sure, I haven't dug into those paths <alethkit>I think the issue is that the Geiser REPL doesn't precompile them <sneek>Welcome back gnucode, you have 1 message! <sneek>gnucode, ArneBab says: guile-emacs worked, but it was too slow, because it was still missing support for byte-compiled files. <gnucode>ArneBab: thanks for giving me a tip about guile-emacs. <orneb>Never mind, I added exec dbus-daemon --session --address=unix:path=$XDG_RUNTIME_DIR/bus to my sway config file and now waybar shows up. <orneb>Does anyone use Sway with seatd and greetd? <orneb>meanwhile waybar is not showing up again. <nckx>haruno1: Did you manually mount anything during the installation? In fact, did you enter a shell (command line) at all? <haruno1>nckx: Hey :D Pretty much folllowed all instructions from yt video where it worked like this (normal gui install from .iso download from website) and the drive specified is the same as the one where installation media is <nckx>You mean that sda is the USB (or whatever) drive? <acrow>vagrantc: I've got fixes for NEWS and handling of those en-dash chars! <vagrantc>acrow: once this is "good enough" i'll probably take a git snapshot and upload guix to debian ... since a release seems to keep getting delayed for one reason or another <acrow>vagrantc: Aside from dependency on some regularity there is no longer anything guix specific in here. I'm happy to now have made a contribution to debian. Whoo hoo. <acrow>vagrantc: My contribs here appear to be more difficult. <acrow>vagrantc: Christine is going down to Tacoma to practise with Ea this AM. I'll be headed out to play with John Peng. What are you up to this morning? <vagrantc>acrow: nix/COPYING is not licensed under the GPL ... it is a copy of the LGPL :) <vagrantc>acrow: apparently, i'm reviewing your tool :) <haruno1>nckx: Sure thing, will do and report results :D <acrow>vagrantc: thanks -- though I'm sure it's better than it sounds. ;) <vagrantc>acrow: would it be crazy to get the UNKNOWN files listed with their copyrights individually, even though you worked so hard to consilidate everything? :) <acrow>vagrantc: Is there other debian tooling that is written in guile? <acrow>vagrantc: All files get listed individually if you add the -A flag. <haruno1>nckx: Oh its the latest one, I was using it :) <nckx>haruno1: Ah, pity, so it's still a relevant bug. Are you still at that error screen? I'd like to know why sda is ‘in use’, for example by checking the output of ‘mount’ at a shell prompt (Control+Alt+F2 or so should switch to one). <nckx>Frustrating that it doesn't report ‘why’, but it doesn't really get that information itself, so little it can do. <nckx>Could at least log/return the errno though. <vldn>mh my loginctl seems to hang if i try to shutdown, someone knows this problem? <haruno1>nckx: I did a create log dump in install and sended it, the installer-dump-28d087d1. Where can I locate it to send to you? :D <nckx>Please send it to bug-guix at gnu dot org, this will create a public bug report at issues.guix.gnu.org that everyone can read & reply to. <nckx>You found the dump, you just need to host/send it, right? <nckx>Just attach it to a mail (possibly compressed using ‘gzip’ if it's more than a few hundred kilobytes). <nckx>I'm not sure the dump contains what I asked for above, though, so please also include the contents of /proc/mountinfo. <nckx>* /proc/mounts I mean. Sorry, distracted :) ***user3456_ is now known as user3456
<Ox151>when guix weather fails to find substitutes, what is the alternative? Take the package scm declaration and run it locally? <podiki[m]>it just means if you do a guix build or guix install it will build it locally <podiki[m]>longest ones being things like llvm, icecat, libreoffice, the rust chain, ... <Ox151>ok, make sense. then i would have to specifiy --fallback? im am just asking because i package i am trying to install failed after the dependicies were installed and i did a weather and it showed no available substitutes <Ox151>podiki[m]: yikes, its rust-cargo specifically so i guess it will take a while <podiki[m]>you shouldn't have to specify --fallback if there are no subs, it will just build (but maybe if it is a network issue?) <podiki[m]>huh, it just failed on most recent try on the CI <podiki[m]>so that indicates likely it wil fail for you as well, which it did sounds like <podiki[m]>in the past the CI built it in ~400s, not bad ***LispyLights is now known as Aurora_v_kosmose
<nckx>There are some CI errors where ‘just try building it again’ is a reasonable option. This doesn't look like one. If a file's missing when building on CI, it will be missing when you build it at home. <nckx>* terms & conditions apply. <nckx>I've restarted the CI build out of curiosity but assume the error is legitimate. ***mark_ is now known as mjw
<acrow>vagrantc: There are three files with the #f license. They are the debian/changelog (which appears to contain no license0, and 2 po/packages/*gmo files (which appear to be build intermediates, again wo license). I think these can be, in general, ignored. So, I will have the program ignore all *.gmo and changelog files. WDYT? <acrow>vagrantc: Well, anyway, in my test file set. <vagrantc>acrow: yeah, i'll need to add an entry for debian/* but ironically nothing in debian/ has much in the way of license headers :) <acrow>No, it will still enter the files in the debian dir, just not changelog or files that glob *.gmo. If I understand you? <vagrantc>acrow: maybe allow for exclude patterns instead of singling out specific file types? *vagrantc sends a boat full of features creeping north <vagrantc>wish there was a "guix import pypi --pretend-it-is-not-a-whell SOMEPACKAGE" to at least spit out a bare package template <Cairn>Reading the manual. Are substitutes disabled by default when using the package manager? <Cairn>Like, when not using Guix System. <Cairn>Just curious if that line's up to day <nckx>sneek: later ask antipode: What's blocking #50238 anyway? <vivien>I’m trying to write a program, with the intent to make it work with guix. It is a custom web client that wraps curl to add some more intelligent stuff. How should I write my tests? For now, I am considering sampling a random port and a self-signed certificate authority, running httpd (unprivileged) on localhost:random-port and testing against that. How are my chances that it will work in the guix builder container? Have you eve <vivien>r found a package that has tests like this? <vagrantc>Cairn: should be enabled by default ... but certainly could depend on how you install it <nckx>Chances are extremely good. <nckx>I don't know which httpd you prefer but at least nginx can listen on a socket (file), which further improves those odds. <vagrantc>Cairn: guix-daemon both needs the relevent keys in /etc/guix/acl and ... either explicitly pass both substitute servers via --substitute-urls or a guix-daemon that defaults to both substitute servers ... <dirtcastle>where should I add the channel.scm for the root user? <nckx>/etc/guix is the default for all users. <nckx>So if /root/.config/guix/channels.scm exists, /etc/guix/channels.scm will be ignored. <nckx>Please remembering root != system-wide, or you is will having bad times. <nckx>Another day, another guix build: error: executing SQLite statement: database disk image is malformed. <pkill9>out of curiosity, does anyone use guix in qubes os? <nckx>klm: That looks like a problem on ‘your’ side. Does ‘ls -d /gnu/store/254la887jsdrpym642d289zkcaahqpp0-c*’ actually show anything? <klm>I'm sure it is :-) Yes, I've got /gnu/store/254la887jsdrpym642d289zkcaahqpp0-cups-pk-helper-0.2.7 <klm>Where did the question mark inside the store filename come from? <nckx>Just an encoding issue in the error message, I think (perhaps a ‘…’ character, distinct from ‘...’, 3 characters, not being properly rendered). <antipode>sneek: later tell nckx: (on #50238) AFAIK, nothing. <sneek>antipode, you have 1 message! <sneek>antipode, nckx says: What's blocking #50238 anyway? <sneek>nckx, antipode says: (on #50238) AFAIK, nothing. <nckx>klm: So the ? has nothing to do with the failure. Know who does have something to do with the failure? *nckx gets the fixing hammer. <Cairn>vagrantc, sorry, I mean when not using Guix System <dirtcastle>just now installed guix using guix system init . I can't login. I didn't give a password but it still asks me for a password . id there a default password ? <Cairn>But I guess that means it wouldn't be enabled by default during an install, since you'd need to modify etc <antipode>... the with-error-handling from the Guix that is being pulled is used, so no problem with travelling back in time to before we had that macro. <nckx>dirtcastle: root shouldn't have a password. Nobody else should be able to log in. <klm>nckx: Are you saying there's a bug and you're on it? <antipode>(on #50238) Besides, with-error-handling precedes "guix pull". <nckx>klm: I'm saying there's a bug in and I'm it. But yes. There's nothing you can do but wait :) <vivien>nckx, I’d love to use file sockets, but a lot of stuff is marked "https-only", which I won’t risk interpreting as "no unsafe-http". <klm>nckx: ok great, thanks! Looking forward to take a peek at your upcoming commit. <nckx>Uh, OK, it's going to be ‘gnu: cups-pk-helper: Fix stupid thing I missed. | This is a follow-up to my commit half an hour ago.’. <nckx>dirtcastle: If that's not documented, it probably should be. <vagrantc>Cairn: i also was referring to when not using Guix System <vagrantc>Cairn: foreign distro installs still have guix-daemon and guix-daemon still uses /etc/guix/acl and such <nckx>klm: By the way, if you don't need cups-pk-helper right now, you could comment it out in the mean time. ***antipode1 is now known as antipode
<nckx>The error is local to that. <dirtcastle>what is the command to make herd list all shepherd services both enabled and disabled. I installed both connman and network manager but I can't start it. <dirtcastle>is there a command to make shepherd list all services both enabled & disabled. I couldn't find it in the manual. <klm>nckx: What do I need to comment out for it to go away? I'm completely new to Guix <nckx>Never mind then, not worth it. Won't take long, I'm just dealing with an unrelated issue & then I'll push a fix. <klm>Ok, thanks. But I'd love to learn how to find out what's causing cups-pk-helper to get pulled in <klm>I'm running `guix graph --type=reverse-package cups-pk-helper` but it only lists itself <nckx>Pretty sure it's it %desktop-services. <nckx>Don't think there's a turnkey CLI way to check that. <mubarak_>I have a small WIFI router that I use to connect my laptop to through wifi. I found also that if I connect it to my laptop with "usb data cable", network-manager connect to it as 'Ethernet Network'. So I search online for how to setup usb0 interface in Linux. <mubarak_>After reading many pages I was able to setup usb0 from the command line and ping successfully after adding some dns servers ip to /etc/resolv.conf. <mubarak_>my questions is how can I add this file '/etc/udev/rules.d/70-persistent-net.rules' to guix? the live system that running from usb doesn't allow me to create that file <nckx>Nor will the final system. I'm sure this is documented in the manual. <nckx>Look for a system service regarding udev rules. <mubarak_>I can't connect through wifi because my wifi card doesn't have free firmware(intel). Thats why I was search for other ways to connect to the internet so that I can install a 100% free distro on my laptop <haruno1>nckx: hey sorry tried to find a way to export the required log but really not sure how to do it since the all options in the installer were included in the uploaded file <nckx>I'm not sure what ‘the required log’ and ‘the uploaded file’ refer to here. <nckx>Or, to be quite honest, ‘export’ 😛 <haruno1>nckx: from earlier: <nckx> You found the dump, you just need to host/send it, right? <haruno1><nckx> Just attach it to a mail (possibly compressed using ‘gzip’ if it's more than a few hundred kilobytes). <haruno1>* atka (~atka@static-198-54-131-90.cust.tzulo.com) has joined <haruno1><nckx> I'm not sure the dump contains what I asked for above, though, so please also include the contents of /proc/mountinfo. <haruno1><nckx> * /proc/mounts I mean. Sorry, distracted :) <nckx>I'm aware of the context. I just have trouble parsing your last message. <haruno1>oh sorry, where can I find /proc/mountinfo since its fresh install from latest build on website :) <nckx>Could you switch to a command line using Ctrl+Alt+F1? <nckx>Then just cat /proc/mounts (mountinfo was a typo). <nckx>Does it mention sda at all? <haruno1>thanks, yeah there isnt sda mentioned <nckx>No? Then why the hell is it ‘busy’ :-/ <nckx>Or whatever the phrase was. <haruno1>yeah it was being used somewhere else haha I am installing using usb flash stick and ventoy, maybe ventoy? <nckx>haruno1: I think I know what this is. <nckx>Etc. (various noises follow) <nckx>unmatched-paren: It does! But it's complicated. <nckx>It works *perfectly* if you perform the manual installation. <haruno1>haha yeah we found it, thanks a bunch guys, will run one without ventoy n test :D <nckx>Yeah, try that, I bet the problem will go away. <nckx>Yeah, no, so I dove into Guix+Ventoy a bit between those two different links, and a bit more afterwards, hence the ‘it's actually more complicated’ conclusion. <nckx>Ventoy is really nice, and quite clever. It doesn't do the old tired GRUB ‘loopback’ thing. It actually presents the ISO as a proper Linux device mapper, er, device. Really cool. <nckx>And this works very well, and avoids most compatibility weirdness with fake drives, because as far as the OS is concerned the drive is remarkably real. <nckx>But in this case the fact that it's still ‘hosted’ on a different device is, or that's my guess, still leaking through and causing phantom business. <nckx>All this is just a hypothesis I just came up with, mind you. But it sounds plausible. <nckx>unmatched-paren: There are more fun things like Guix's ventoy support being purely coincidental, because we pull in dm-crypt for LUKS support, and this happens to pull in dm-mod which is what Ventoy needs. It's why I previously said Guix didn't support ventoy: because I was booting an image without LUKS support, because that was ‘obviously’ irrelevant. Oops. <nckx>Or: whilst I redeployed this laptop from scratch using a Ventoy-booted ISO and it went swimmingly from start to finish… I kept the existing partition layout, because why would I repartition. Oops. <dirtcastle>oh I had to run dhclient -v interface to get internet working. anyways connman and network manager still doesn't work. <pkill9>whu does guix home rexonfiigure keep redownloading everything, it makes no sense <nckx>I think the right fix here would be to hide busy devices, possibly with a warning (‘Some devices are not available, blah blah’), but allowing the installation to continue. This wouldn't even be an artificial limitation because even mounting the Ventoy host partition with the ISO is explicitly unsupported already.