IRC channel logs


back to list of logs

<mwette>efraim: the `(Auth ...) => (list `(Auth ...))
<NieDzejkob>PotentialUser-26: try adding elogind to your services if you haven't already
<PotentialUser-26>oof, systemd deps. trying now
<NieDzejkob>elogind is specifically systemd-less
<NieDzejkob>like eudev
<PotentialUser-26>wasn't it extracted from systemd though? I mainly want to avoid the issues listed here:
<PotentialUser-26>would like screen to still work and such
<PotentialUser-26>but with all that said... IT WORKS! thank you so much for your help!!!! feels so good to have a guixsd system up and running. now to learn more about it (and eventually, contribute back)
<dftxbs3e>NieDzejkob, worked great. awesome
<dftxbs3e>how does one raise an exception in GNU Guile's Scheme?
<NieDzejkob>dftxbs3e: see `info guile condition`, grep for `(condition' in guix's source for examples
<dftxbs3e>NieDzejkob, thanks!
<dftxbs3e>NieDzejkob, I'm really sorry it's very confusing for me. Where can I find the documentation for the 'substitute*' function?
<newUser-andrew>Hi! I'm pretty new to Guix, and am curious about best practices. In particular, I wonder at what point Guix usage /stops/ being the right call. I would obviously not want to use it to manage random static files like pictures, but it seems appropriate to use it for installing fonts. What about keyboard layouts or retroarch cores? If there's a better
<newUser-andrew> place for this info just lmk, thaks :)
<Gooberpatrol66>I don't know what I did, but now guix pull returns guix pull: error: symlink: File exists: "/var/guix/profiles/per-user/nathan/current-guix"
<dftxbs3e>Gooberpatrol66, probably a bug.. report to the guix-devel mailing list
<guix-vits>Hi! Emacs rules. IRC w.o. mouse.
<drakonis>not the only TUI irc client, but yes it rules
***jamesrichardson is now known as jr0
***jr0 is now known as j-r`
***j-r` is now known as jr0
***jr0 is now known as Guest20845
<kbw1111>hi, i have a question about guix packaging. How can I inherit from a package and add (rather than totally replacing) a native-input or input? I'm not familiar with guile/lisp so i'm not sure how to work with the fields of the package
<guix-vits>If somebody know how to defeat the error: "/linux/errno.h: No such file or directory"?
<guix-vits>i'm just installed the toolchain, and tried to compile "hello" ``./gnu/store/ahqgl4h89xqj695lgqvsaf6zh2nhy4pj-glibc-2.29/include/bits/errno.h
<guix-vits>:26:11: fatal error: linux/errno.h: No such file or directory''
<guix-vits>it is g++
<guix-vits>i'll try to install gcc-9, as gcc-7 (was installed when i'd not specified version) isn't working for me.
<guix-vits>Strange, i can compile this "hello" from tty, but not from the Sway. I'll start new session and see...
<guix-vits>Yes it's works now. Will restart this session. Strange: why do i need to restart graphics session to make the newly installed GCC to work?
<dftxbs3e>hello, how can I get the target triplet it's currently building for in Guix?
<dftxbs3e>as a string *
<dftxbs3e>been trying to pass (%current-target-system) to string-prefix? and it doesnt want it..
<guix-vits>dftxbs3e: try gcc -v
<dftxbs3e>guix-vits, I mean inside the context of GNU Guix
<dftxbs3e>invoking a shell program is a hack
<guix-vits>The crowd shared it with me, it is a devel-version of Guix Manual
<dftxbs3e>guix-vits, I've been already grepping that... but without success.
<dftxbs3e>That's why I'm asking here, can't find in the manual or in GNU Guix source code
<dftxbs3e>gave up. seems like it's just too complicated to conditionally patch a package with additional patches depending on the target cross compiling triplet.
<guix-vits>dftxbs3e: i'm sorry, not got it: why Guix will ever need some platform-specific patchs? Either the package ported, or not. Correct? Not dev myself.
<guix-vits>dftxbs3e: (string-append libc "/lib"
<guix-vits> ":" libc "/" target "/lib")))
<guix-vits>variable named target, in:
<guix-vits>But i'm not understand the code at all, maybe useless.
<guix-vits>And just in case:
<str1ngs>dftxbs3e (%current-target-system) does not look right % usually means a global variable. have you tried %current-target-system instead?
<str1ngs>guix-vits: you need to restart when guix tells you that the profile environment has changed. since the session needs to source the new variables.
<guix-vits>str1ngs: Thanks. Is there any way to just `. SOME_FILE` instead of logoff, then login?
<str1ngs>guix-vits: closing a terminal an creating a new terminal will resolve it for bash. but for X11 sessions that does not work. because environment variables need to propagate to child processes
<str1ngs>same goes for wayland clients . unless they have a mechanism to load environment variables
<guix-vits>str1ngs: Fun is, that i got errors from g++. Inside of terminal, sir! ;)
<guix-vits>True story...
<str1ngs>right in which case you would have to restart the terminal
<str1ngs>guix is will tell you when you need to do this. it's just overlooked .when you first start using guix install :)
<joshuaBPMan>what permissions do my html files need to be for nginx-service to serve them?
<guix-vits>str1ngs: This had a good side as well: now i know how to (
<guix-vits>(packages (append (list ... package-MAJOR_VERSION_NUMBER ...)))
<guix-vits>joshuaBPMan: (I've no idea, never ran it):
<guix-vits>but this one from 2010 year...
<dftxbs3e>str1ngs, I think I did try yes.
<joshuaBP`>guix-vits: thanks
<clode>hi there I have guix installed alongside Voidlinux
<clode>I can't get to have guix-daemon run with runit
<clode>I have a run file that contains '/var/guix/profiles/per-user/root/current-guix/bin/guix-daemon --build-users-group=guixbuild'
<clode>but the logs spit out 'fatal error unable to start, exec format error'
<clode>however it runs fine when doing it manually in the terminal
<clode>anyone has seen anything like this before/
<nckx>Good morning, Guix!
<g_bor[m]>hello guix!
<clode>looks like I don't know how to use runit
<nckx>clode: What was the problem?
<nckx>…and they quit again.
<PotentialUser-63>Hey there. I tried to configure xorg-server with (extra-config ...) in the system config but the settings do not seem to apply. Where should those settings end up? In which file should I find them? Help appreciated :)
<nckx>PotentialUser-63: If you run ‘ps auxwww | grep [b]in/X’ it will display the ‘-config’ file that is used. Mine contains all my extra-config customisations.
<PotentialUser-63>nckx: Thanks. Didn't think of the ps output to find the config.. the config from -config looks quite default, so I think my settings weren't appended
<nckx>PotentialUser-63: Did you reconfigure and restart your log-in manager/machine?
<nckx>Here's my configuration (%xorg.conf is a string):
<PotentialUser-63>nckx: I did reconfigure and did a full reboot. Here is my config: . %machine-xorg-libinput-config is also a string like in your config.
<NieDzejkob>sneek: later tell kbw1111: You can do something like (package (inherit some-other-package) (native-inputs `(("my-new-input" ,hello) ,@(package-native-inputs some-otherr-package))))
<sneek>Will do.
<NieDzejkob>dftxbs3e: see the docstring, in this case in guix/build/utils.scm just below (define-syntax substitute*
<nckx>PotentialUser-63: Hm, I've not used set-xorg-configuration myself (why break what works?) but don't immediately see a mistake in yours, sorry.
<PotentialUser-63>nckx: I had trouble with setting my xorg configuration without set-xorg-configuration ... So that's why I used it :D. But I will definately try your config, it looks promising! Thank you
<bricewge>PotentialUser-63: It only works with GDM.
<sneek>bricewge, you have 1 message.
<sneek>bricewge, vits95 says: Thanks. The online version got a pictures, and there is even a characters. What a waste of time it was that i didn't managed to read this. I most certainly will.
<NieDzejkob>set-xorg-configuration only works for gdm
<PotentialUser-63>hmm.. the documentation should mention that explicitly.. It says: "Scheme Procedure: set-xorg-configuration config [login-manager-service-type]". [login-manager-service-type] seemed generic enough for me to use it with slim
<bricewge>Yes, you are not the only that stumbled upon this.
<NieDzejkob>or we could modify slim and sddm to accept xorg configs like that
<bricewge>PotentialUser-63: You could submit a patch for it until someone modify the other login-manager.
<NieDzejkob>no need, I'll send a patch to modify the login-managers like, now
<bricewge>NieDzejkob: Thanks :)
<bricewge>Are there any example of packages that uses multiple sources?
<PotentialUser-63>bricewge: I need to bootstrap my system first and understand it better in order to be able to submit patches :). But in the future I would like to help if I can.
<PotentialUser-63>NieDzejkob: awesome. Thanks
<bricewge>I want to add translation and documentation to Kicad, but they are in different repositories. And the binary looks in $PREFIX/share to find them.
<NieDzejkob>bricewge: I'd look into make-linux-libre-source
<NieDzejkob>Is there a way to use additional channels with ./pre-inst-env?
<nckx>bricewge: You could add the repositor{y,ies} as an input and copy them to $out/share manually. See git's git-manpages input for an example. If the docs are large, you can put them in :doc and patch the programme to look for them there.
<nckx>Am I the only one who finds set-xorg-configuration to be a strange idiom? It isn't, but it feels, non-functional.
<NieDzejkob>the fact that you need to specify the login manager type makes it, eh...
<NieDzejkob>maybe we should deprecate it?
<nckx>I don't know. I've always kind of ignored it until now 😛 No objection from me.
<nckx>I wonder what the motivation was for adding it (if it's ‘modify-services is spooky for new users’, eh, maybe, but this is no fix).
<abralek>How can I define my own functions in a scope of build phases? gexp i suppose?
<abralek>Is there some example?
<NieDzejkob>abralek: see gnu/packages/geo.scm:516
<abralek>well yes, but that lambda is not available in other phases, right?
<abralek>I need to define something like invoke, find-files, etc. Like those defined in (guix build utils)
<NieDzejkob>hmm, you should be able to do (let ((helper _)) (modify-phases ...)), I think
<NieDzejkob>(that's because modify-phases runs build-side
<abralek>NieDzejkob: Indeed, I does work!
<abralek>NieDzejkob: Thanks!
<NieDzejkob>Is there any reason core-updates isn't using GCC 9?
<efraim>It's a lot of effort to make sure everything builds correctly with gcc-9
<nckx>At a certain point you have to go into ‘ship it’ mode and stop bumping things, which is currently the case.
<efraim>looking at the mailing list I think we're about to lockdown core-updates and start hammering into shape
<janneke>efraim: i was waiting for a final OK before merging wip-bootstrap last friday after
<efraim>janneke: ah, right. Don't know how I forgot about that.
<efraim>that's how I figured it was ok to push my change to the gnu-build-system to pass NOCONFIGURE during 'bootstrap
<janneke>civodul said i should ask mbakke before merging and so i did, but wanted a final OK from him i guess -- scary to merge such a large patch set :-)
<janneke>efraim: i'm rebuilding world -- looking good :-)
<janneke>already at gcc-4.9.4 right now
<efraim>i tested it out on bayfront, hopefully none of the ones I removed needed "indeed" or "please" as an actual input
<janneke>yes that was a cleanup long overdue, i guess
<apteryx>bricewge: the patches for booting off a btrfs subvolumes are out there and they work alright, if you were interested in that!
<sneek>apteryx, you have 1 message.
<sneek>apteryx, NieDzejkob says: interesting. slim-service-type adds xterm to the list of packages
<apteryx>NieDzejkob: that explains it!
<apteryx>NieDzejkob: we should just patch ratpoison to refer to it directly
<apteryx>that'd be more to the point
<bricewge>nckx: git-manpages is definitely interesting, thanks. In my case I need to run cmake && make install, so instead of just calling origin in input I could define a package that use the cmake-build-system. Or is it too ugly?
<bricewge>apteryx: Nice! :D It's time to break my system.
<nckx>bricewge: A separate whatwasitagain-docs package certainly makes sense in some situations, if the docs need their own build recipe & inputs. Sounds like that might be the case. You can still keep the -docs package private (define vs. define-public) and just copy its output into whatwasitagain:{out or doc}. Depends on the size increase, and how tied into the UI these docs are.
<nckx>It depends™. 🙂
<bricewge>Thank you, I'll do it as you suggest then and will adapt the output from there.
<mbakke>NieDzejkob: patches that bump to GCC 9 would be most welcome (but also delayed until the next core-updates round) :-)
<NieDzejkob>apteryx: Hmm, I'll try removing xterm and see if there's really a problem with ratpoison's help
<mbakke>janneke: I've read through some of the commits and things LGTM overall :-)
<efraim>glibc-intermediate on core-updates on powerpc can't find nptl and fails to build
<janneke>mbakke: ty, it's building gcc-7.5.0 right now; when i get to `hello' i'll push it to core-updates
<janneke>it's annoying that just seems to say "No"
<efraim>don't worry about my error, powerpc-linux-gnu isn't a supported platform yet
<janneke>efraim: okay, yes i was hoping that any change/fixe there would be orthogonal
<janneke>well, you can always hope, right ;)
<efraim>it gets further on core-updates than on master, on master it does at gcc-cross-boot0
<rekado_>I’m trying to reconfigure and get this odd error:
<rekado_>ice-9/psyntax.scm:3254:24: attempt to include relative file name but could not determine base dir
<rekado_>the error location isn’t very helpful.
<rekado_>I get this after guix pull. And older guix works fine.
<civodul>mbakke, janneke: re wip-bootstrap, sounds like great news!
<mbakke>I just remembered this patch:
<civodul>rekado_: it probably comes from the user of local-file in one of our files
<civodul>could you try to determine which one?
<mbakke>can anyone see potential problems with it?
***jr0 is now known as Guest33885
<g_bor[m]>mbakke: it looks good to me. Maybe I would move script exists to build modules, so that it can be reused, and the name of it does not really reflect the functionality...
<rekado_>civodul: it’s the (include "nginx/berlin.scm") in berlin.scm
<g_bor[m]>Something like find-non-directory-file would be more exact.
<g_bor[m]>Sorry, so non-directory-file-exists.
<guixy>Hello guix!
<guixy>I'm working on a wrapper for clojure.
<guixy>Guix makes two different jars for clojure. Which one should I try to target?
<guixy>I'm guessing the more generic jar is preferable.
<guixy>bye guix
<janneke>civodul: ah yes, i rebased this morning on core-updates and started a rebuild
<rekado_>hmm, fibers fails to build
<cbaines>rekado_, if it's the same failure I encountered, it's non-deterministic
<cbaines>fewer cores may help, but the -c argument to guix build doesn't apply to the test suite (where the failure occurs)
<rekado_>cbaines: should we disable parallel builds for fibers then?
<cbaines>rekado_, I think the test suite does it's own thing, so patching the test suite may be necessary
<cbaines>I talked to Andy at the Guix days about the issue, and he said the underlying problem is a bug in Guile
<apteryx>NieDzejkob: perhaps it was using a different keyboard which made me fail at displaying C-t ?, but I think there was.
<apteryx>bricewge: \o/
<janneke>efraim: have you built core-updates after your changes? my rebased wip-bootstrap (not published yet) has been stuck for about 1h building gcc-7.5.0 here:
<janneke> /gnu/store/5m2cnfwk87n36wv19cgam9w0fdyx35rn-bash-static-5.0.16/bin/bash ../../gcc-7.5.0/gcc/../move-if-change tmp-fixinc_list fixinc_list
<janneke>echo timestamp > s-fixinc_list
*janneke goes to try building 'hello' on core-updates first
<janneke>your fixes seemed mostly harmless...
<janneke>maybe it's bad luck, i'll ^C and retry my rebased wip-bootstrap too
<joshuaBPMan>morning guix!
<NieDzejkob>bricewge: it took longer than I thought it'd take because testing requires a reconfigure & reboot, but
<janneke>hmm, no ... it's something else; the build still runs but it seems like the connection with the offload server has been interrupted or lost
*janneke just built hello on core-updates
<joshuaBPMan>Hey guix, does anyone know what the status of running web apps on guix?
<joshuaBPMan>I guess the javascript is probably holding up back quite a bit.
<OriansJ>joshuaBPMan: well I don't think javascript has anyone dedicated to getting the required packages into guix yet.
<apteryx>nckx: o/
<apteryx>joshuaBPMan: heya!
<joshuaBPMan>OriansJ: yeah....I hear that is a real head-ache trying to figure out all the cycling dependancies.
<OriansJ>joshuaBPMan: that is a bit of an understatement as I hear; basically no distro has javascript without binaries being used in the build process.
<OriansJ>even NixOS hasn't cracked that nut yet; despite years of hammering from a few dedicated people.
<mbakke>Python 3.8.2 is due tomorrow, that will be my last change for this core-updates round, promise :P
*mbakke takes the day off
<joshuaBPMan>OriansJ: dang.
<bricewge>NieDzejkob: Thank you again! TIL syntax-rules and I'll be able to get rid of the strange (slim-configuration (xorg-configuration (xorg-configuration (keyboard-layout ...)))).
<OriansJ>joshuaBPMan: even basics like jquery, which depends upons jquery to be built
<joshuaBPMan>OriansJ: geez. It would be super cool to be able just create a wordpress install on the fly in guix, but it sounds like that is some time away.
<OriansJ>joshuaBPMan: not really, just include binaries and the build chain is easy to make. Like NixOS has
<joshuaBPMan>OriansJ: I suppose if you're willing to bend on your principles a little.
<OriansJ>joshuaBPMan: some people care more about results than principals (The ends justify the means approach and all)
<NieDzejkob>I recently noticed that bison uses pre-compiled parsers when built in Guix. Were there any past attempts to avoid that?
<OriansJ>NieDzejkob: I think Nyacc is supposed to help solve that issue but I could be wrong.
<civodul>NieDzejkob: you mean that Bison tarball includes pre-built parsers?
<civodul>i guess nobody has investigated yet
*nckx .oO Is there no equivalent of with-directory-excursion for environment variables?
<rekado_>nckx: there is with-environment-variable in guix/tests.scm
<nckx>rekado_: Ah, thanks. But not in build utils or somesuch?
*nckx asks because of 1b2b776 in case that wasn't obvious.
<janneke>bah, i really cannot build my rebased wip-bootstrap :-(
<janneke>seems that i goofed and missed the window to push it onto core-updates; now i might be looking at another painful week of bisecting
<janneke>looking at binutils and glibc updates...crap
<civodul>janneke: where is it failing?
<NieDzejkob>hmm, does modifying guix/build trigger a world rebuild?
<civodul>rekado_: congrats on GWL 0.2.0! \o/
<civodul>NieDzejkob: in general yes, because (guix build utils) is used by everything, for instance
<civodul>changing (guix build syscalls) is fine, though
<civodul>when in doubt, try "guix build libreoffice -n" :-)
<janneke>civodul: it "hangs" while building gcc-7.5.0 (building /gnu/store/bq70a7hsn2jav9zixwc60lv23x2m92vs-gcc-7.5.0.drv...)
<janneke>my 3rd attempt now, it hangs for about an hour at
<janneke>/gnu/store/5m2cnfwk87n36wv19cgam9w0fdyx35rn-bash-static-5.0.16/bin/bash ../../gcc-7.5.0/gcc/../move-if-change tmp-fixinc_list fixinc_list
<janneke>echo timestamp > s-fixinc_list
<janneke>and then it is probably killed because of silence, dunno
<civodul>janneke: what does the process tree show at that point?
<janneke>civodul: i haven't looked, new attempt
<civodul>janneke: i can spawn the build too
<civodul>from ef7e1b3631e15c0cda14f96522fe59ce93ee407e
<janneke>ty; yes, that's what i'm using
<NieDzejkob>civodul: Thanks!
<efraim>janneke: just saw your message. I built past hello with my first patch but not with my second
<janneke>efraim: thank you, i just built core-updates with both your patches
<janneke>efraim: also, i realised that another merge of master into core-updates happened
<janneke>just before your patches, that's a much more probable source of the problem
<janneke>i'm now doing a non-offload build and civodul is also trying a build
<janneke>hello mehlon
<mehlon>anyone know what the status of Tor Browser on guix is?
<mehlon>at some point there was some attempt by ng0 I recall, but I'm not sure what came from that
<janneke>mehlon: i think the focus has shifted to Tor,
<NieDzejkob>ah, damn it
<NieDzejkob>I thought I could remove rust-winapi from inputs because the packages kept building. Turns out that's only because the crate was still being imported transitively
<mehlon>hmm.. I see
<mehlon>I guess it's a non-trivial job to package it
<efraim>It might work if you set custom build and test flags, but probably not worth it
<bricewge>po4a keeps telling me to install Unicode::GCString but I can't manage to do it.
<bricewge>I know nothing about perl, so does .../lib/perl5/site_perl/5.30.0/x86_64-linux-thread-multi/Unicode/ is the correct path?
<bricewge>x86_64-linux-thread-multi seems funny.
*janneke goes to steal nckx quit msg
*nckx doesn't mind, spread the ♥
<NieDzejkob>I normally have Join/Part/Quit events hidden, but it seems like there's reason not to
<nckx>bricewge: /gnu/store/snw2qx2jph1a78n25bgq37xx19kigl60-perl-unicode-linebreak-2019.001/lib/perl5/site_perl/5.30.0/x86_64-linux-thread-multi/Unicode/
<nckx>Read that as: are you actually installing the correct (perl-unicode-linebreak) package?
<bricewge>nckx: Yes it's the one I'm using.
<bricewge>Is it normal that x86_64-linux-thread-multi is part of the path?
<bricewge>I tried wrapping po4a in PERL5LIB and setting the save environment variable before the build phase of my package but to no avail.
<nckx>bricewge: It's not unusual (🎵) but I don't know what the exact meaning is. Many packages have $that/auto/PackageName, others put PackageName directly in 5.30.0…
<bricewge>So I was wondering if the issue came from the package.
<nckx>bricewge: And you set PERL5LIB to $pkg/lib/perl5/site_perl? That's definitely correct, I've never had to add the x86… (nor does any Guix package).
<bricewge>nckx: Exactly.
<janneke>civodul: successfully built /gnu/store/bq70a7hsn2jav9zixwc60lv23x2m92vs-gcc-7.5.0.drv
<janneke>civodul: so, it seems i have been experiencing a guix offload problem, any ideas?
<janneke>i don't think i had problems with offload for > 1y or so, i did have to restart guix-publish some times
<nckx>NieDzejkob: #:configure-flags is respected by meson-build-system.
***jr0 is now known as j0r
***j0r is now known as j-r`
<NieDzejkob>nckx: ah, when I did * in vim to find all uses of the variable, it didn't flag the one where it's used because it thought that ,@ is part of the word
<bricewge>NieDzejkob: Sorry for forgetting guix lint, do you have a git hook to do it?
<NieDzejkob>bricewge: I'm in the middle of writing one. I will submit a patch when I'm done ;)
<NieDzejkob>(Don't worry about it. I also forgot to run it on a few of my packages ;D)
<civodul>janneke: uh, i'm not aware of offloading issues these days
<janneke>hmm, i should look into it then -- find out what's hanging; possibly try a build on the offload machine
*janneke almost built guile again, locally; melting their laptop
<nnasunknown>hello friends
<nnasunknown>can you help me please
<nnasunknown>i follow the instructions in
<janneke>nnasunknown: good, you have a question?
<bricewge>NieDzejkob: Nice! That's the kind of thing I'm sure to forget; better to automate it.
<nnasunknown>when i install a package like guix package -i hello i got the follow message :
<nnasunknown>substitute: /gnu/store/4yyw7brvww6dk9cwwvdq724qdfz2zn0v-bash-minimal-4.4.23/bin/bash: warning: setlocale: LC_ALL: cannot change locale (en_US.UTF-8)
<nnasunknown>i already installed glibc-locales and glibc-utf8-locales
<nnasunknown>in /etc/systemd/system/guix-daemon.service i have : Environment='GUIX_LOCPATH=/var/guix/profiles/per-user/root/guix-profile/lib/locale' LC_ALL=en_US.UTF-8
<civodul>nnasunknown: that looks good; could you make sure glibc-utf8-locales is installed in root's account?
<nnasunknown>in /var/guix/profiles/per-user/root/current-guix/share/locale ?
<nnasunknown>or : # guix package --list-generations ? as root
<civodul>nnasunknown: as root you can run "guix install glibc-utf8-locales"
<nckx>nnasunknown: Try sudo guix package -I.
<civodul>then it'll show up in /var/guix/profiles/per-user/root/guix-profile/lib/locale
<NieDzejkob>sudo -i guix package -I.
<NieDzejkob>you need a login shell
<noobly>is /dev/sda1 size: 2M type: BIOS boot, /dev/sda2 size: 3.7G type: linux swao, /dev/sda3 size: 234.8G type: Linux file system a good scheme for bios/GPT?
<NieDzejkob>yeah, you should be good
<noobly>i tried the graphical install again but this morning when I booted my laptop up (no idea why it turned off, maybe the cat?) I was lead to an inaccessible "GRUB _" prompt
<noobly>so I rebooted with my bootable usb, mounted /dev/sda3, and looked over the docs and my config.scm and have rerun guix system init /mnt/etc/config.scm /mnt
<noobly>interestingly when I mounted /dev/sda3 /mnt, ls /mnt only returned 3 directories: etc, gnu, lost+found, tmp and var.. why was there no /home?
<nckx>noobly: That would mean that system never booted.
<nckx>Or at least not very far.
<noobly>nckx: what would? the grub prompt, or my lack of directories under /mnt?
<noobly>(or both, more likely)
<noobly>am I taking the right steps forward to complete the install?
<nckx>Sorry, missing /home.
<noobly>nckx, ok well that's good, i was worried it meant i had to redownload all those packages
<noobly>my wifi sucks
<nckx>Ehm, I'm afraid ‘guix system init’ will not re-use any part of what's already in /mnt/gnu.
<nckx>Especially if you started the installer afresh (new cow-store &c.).
<noobly>should I stop it? or just let it ride?
<noobly>it takes about 24+ hours for all my stuff to finish
<noobly>and I didn't really change much to the bootloader config, so I think I'm just going to run into the same issue
<noobly>I'm not sure where to look or how to troubleshoot, usually I can checkout dmesg or some var/log but I wasn't sure where to look or what even to look for
<nckx>noobly: What's unclear is what exactly happened before the cat incident: was the installer still running, was it interrupted, did you actually boot into your new system, …?
<nckx>noobly: Definitely keep a camera around when you boot the installed system, hopefully it will print more than just GRUB_ and we have something to help with.
<nckx>The fact that /mnt/boot doesn't even exist makes me suspect your installation never even finished but what do I know.
<nckx>Less than your cat.
<noobly>nckx: well, I went to bed with cow-store grabbing it's downloads (if my terminology is right), and when I woke up the machine was off. I've been casually battling this for a few days, running through various iterations on the graphical install and command line. every time I end with a bootloader error, the first couple times I was thinking it was
<noobly>EFI vs BIOS related
<nnasunknown>can i guix pull ? acts like transaction?
<noobly>but now that I've got a more accurate partition idk what the issue is, maybe a cat just turned it off so I'll just see how this guix system init command resolves I guess?
<noobly>can i echo startup text into a file somehow?
<nckx>nnasunknown: Yes, guix pull will either succeed or leave your previous Guix working.
<noobly>rather than camera upon boot
<nckx>noobly: Eh, not that I know of, our logging in general is a mess and (early) boot logging doubly so.
<nckx>noobly: OK, if it's at all possible that your installation never finished this makes more sense. No other choice but to try it again.
<noobly>ok, is there a built in text-bin utility like wgetpaste or dpaste I should know about?
<nckx>And for the love of god put a laundry basket over your laptop or something.
<noobly>ok, so to be clear I'm doing the right thing by letting the guix system init command ride?
<noobly>lol, good idea!
<nckx>noobly: We have wgetpaste.
<nckx>I think it even works.
<noobly>does it come with the install image?
<noobly>im getting command not found, but maybe that's just me
<noobly>thanks for the help
<nckx>I don't think it does, or it was added after the (relatively old) 1.0.1 release.
<nnasunknown>this is the only output for sudo -i guix package -I : /gnu/store/4yyw7brvww6dk9cwwvdq724qdfz2zn0v-bash-minimal-4.4.23/bin/bash: warning: setlocale: LC_ALL: cannot change locale (en_US.UTF-8)
<nckx>nnasunknown: Then you need to sudo [-i] guix install glibc-utf8-locales.
<nckx>I'm sceptical of the -i thing but if Jakub says it's needed I believe them.
<nckx>Foreign distros are a wild & lawless place.
<NieDzejkob>oh, you need them to use root's package definitions, profiles are unaffected (only tested right now) :'D
<nnasunknown>ohhh friend , yo make my day
<nckx>Yo estoy feliz 🙂
<nnasunknown>idem compadre
<NieDzejkob>what is this alien language and how come you realized you both speak it? :D
<drakonis> /whois nnasunknown
<drakonis>pretend you didnt see that
<drakonis>NieDzejkob: alas, it is an elegant language.
<nckx>Ouch, caught peeking.
<nckx>NieDzejkob: I don't even.
<nckx>That's about the extent of my Spanish.
<nckx>Thank you annoying b-day song.
<noobly>soy un perdedor
<NieDzejkob>drakonis: I can't tell whether you legitimately messed up that command or did it on purpose
<drakonis>i did mess it up, yes.
<drakonis>happens sometimes.
<noobly>how do you do it lol
<nckx>noobly: The actual command is /whois <nick>, what drakonis did was add a leading space which makes it a regular message.
<nckx>A common habit when pasting /unix/file/names.
<drakonis>a recurrent mistake when you have tall speakers in front of your screen
<drakonis>i leaned incorrectly and it covered the message ar
<NieDzejkob>I didn't yet manage to pick up the habit of prefixing my messages with spaces when talking about paths
<nckx>‘In front of your screen’ is a bad place for anything.
<drakonis>its tall enough to cover the task bar and message bar so its kinda weird.
<NieDzejkob>I prefer my keyboard in front of the screen rather than behind it
<nckx>Oh, fun with prepositions.
*nckx wonders if there's a language with unambiguous ‘in front of’ and ‘in front of I CAN'T SEE’. Probably.
<NieDzejkob>I kinda feel like Polish, my native tongue, would have something like that, since it's quite a bloated language feature-wise, but I can't for the life of me think of how I would say that
<KE0VVT>NieDzejkob: Polish is bloated?
<nckx>Does libyaml have some huge dependency count obscured through inheritance (or something else)? The fact that it's still on 0.1.7 when 0.2.1 is in Guix for a single dependent is weird.
<nckx>Surely not all 128 dependents of 0.1.7 are incompatible with 0.2.
<sirgazil>civodul: I sent a patch to fix the headings issue we talked about recently (
<alextee[m]>nckx: noticed that too, pretty weird that the 2 is versioned and "libyaml" is 0.1.7. shouldn't it be the other way around?
<nckx>alextee[m]: That's what makes me think ‘someone had good reason to do it this way’ and ask here.
<nckx>Of course that thought is often mistaken, but not always 😉
<mjw>civodul, congrats on the new Shepherd release. It is one of those things I would love to hack on. If I had time... *sigh*
<mehlon>"wake up sheeple!" - shepherd
<mehlon>good mornevening guix
<NieDzejkob>nckx: has git blame got any hints?
<nckx>NieDzejkob: I haven't looked yet, got distracted by openssh update.
<drakonis>ah, a new shepherd release?
<civodul>thanks, mjw!
*civodul -> zZz
<civodul>night all!
<jackhill>nckx: yay, openssh update, thanks! I'm excited about ed25519-sk and ecdsa-sk key types.
<nckx>jackhill: Currently blocking you excitement is a test failure (‘nixbld’ is suddenly no longer a valid user…)
<nckx>‘User nixbld not allowed because shell /noshell does not exist’ ah.