IRC channel logs

2018-06-22.log

back to list of logs

<Copenhagen_Bram>hello
<nckx>o/
<pkill9>hi
<apteryx`>Copenhagen_Bram: I think I saw a bug fixed like today which was guix system reconfigure crashing with an error similar to what you shown us yesterday.
<Copenhagen_Bram>Hmm.
<apteryx`>you should update your guix and try again
<Copenhagen_Bram>I think I fixed that, by using a different guix version
<apteryx`>OK, so you're already passed that bump?
<Copenhagen_Bram># ls -l $(which guix)
<Copenhagen_Bram>lrwxrwxrwx 2 root root 56 Dec 31 1969 /root/.config/guix/current/bin/guix -> /gnu/store/gyanngcnr1vnx95c0xs0yi78r3ss1n2d-guix-command
<Copenhagen_Bram>yeh
<Copenhagen_Bram>this is the guix that doesn't work
<apteryx`>try guix --version to get the commit
<Copenhagen_Bram>a34801ca685df32e1df50597a96a578acb66cfb0
<apteryx`>nckx: ddcutil works :)
<apteryx`>I can change brightness using sudo ddcutil setvcp 10 100 (maximum) or sudo ddcutil setvcp 10 1 (minimum). Amazing! Thanks :)
<apteryx`>Now I'll try giving my user access to /dev/i2c-* without sudo.
<nckx>apteryx`: Great news! I was just heading to bed. I'll fix up the description and send it in, hopefully tomorrow...
<apteryx`>and will be on my way to having a functional keyboard shortcut to adjust brightness in ratpoison \\m/
<apteryx`>nckx: cool :) good night!
<nckx>\\o
<mbakke>janneke:
<mbakke>Oops.
<mbakke>Can you elaborate on the crt1.o error?
<janneke>mbakke: ld says: crt1.o: access beyond end of merged section (1728814200 + 0)
<janneke>i found that value is `addend', which is read from crt1.o
<janneke>but i'm going to compile binutils-2.14 with a full, proper gcc i686 toolchain and see what happens
***Gamayun_ is now known as Gamayun
<sahithi-ihtihas>Hi
<brendyn>hi
<sahithi-ihtihas>when I tried this (map (lambda (n) (match:substring m n)) (iota (- (match:count m) 1) 1))))
<sahithi-ihtihas>I am getting a whole string instead of list of substrings
<sahithi-ihtihas> https://paste.debian.net/1030281
<civodul>Hello Guix!
<sneek>civodul, you have 1 message.
<sneek>civodul, wingo says: yes, find_slot_map for subr_stub_code returns null
<roptat>hi civodul :)
<roptat>make dist failed with:
<roptat>make[2]: *** Aucune règle pour fabriquer la cible « nix/libstore/schema.sql », nécessaire pour « distdir-am ». Arrêt.
<roptat>I tried to run it to generate the tarball to send to the TP
<civodul>oh, that must be a mistake of mine
<civodul>i moved that file to guix/store recently
<civodul>lemme see
<civodul>efraim: i can connect to your overdrive from here, but from berlin 'guix offload test' times out
<civodul>ideas?
<civodul>roptat: fixed!
<roptat>civodul: thanks :)
<civodul>yw
<civodul>our GNOME stack is starting to be quite outdated
<efraim>civodul: the IP address changed, could that cause it?
<efraim>Or perhaps the ssh tunneling is causing problems
<rekado>I still cannot run alsamixer since the merge of core-updates.
<rekado>Do I really need to install some alsa plugins into my profile to fix the alsa service?
<roptat>rekado: what's the error message?
<rekado>roptat: the usual about libasound_module_ctl_pulse.so not being found.
<rekado>roptat: shouldn’t the *service* take care of this?
<rekado>including setting the environment variable to find the plugins?
<g_bor>hello guix!
<g_bor>I'm looking at #30889. gnu:Add pipewalker. I've seen the recommendation to use a sourceforge mirror. Do we have any policy which mirror to use?
<roptat>rekado: is there a way to set this variable though?
<g_bor>Never mind, I've found it.
<roptat>g_bor: mirror://sourceforge I think :)
<g_bor>Thanks :)
<roptat>rekado: also, the env var works only when alsa-lib and alsa-plugins are in the same profile
<rekado>roptat: I only have alsa-utils in my profile.
<rekado>I don’t have a custom ~/.asoundrc any more.
<rekado>do I need to install both alsa-lib and alsa-plugins in addition to alsa-utils now, just so I can use alsamixer?
<roptat>I think so
<rekado>that’s unfortunate :(
<roptat>I agree, the service should take care of this
<roptat>but I don't know how to do that
<g_bor>#31488 look good to me, but triggers 212 rebuilds. It adds an output to git. WDYT? I think this should go to staging?
<roptat>g_bor: that's less than 300
<g_bor>You are right :)
<roptat>if you fear it could cause some build failures though, it may be better to send it to staging
<roptat>you can also try and build some of these dependent packages
<g_bor>Actually I don't think there will be any failures. It adds an output... Does not touch anything else. Ok, I will push this then, and close.
<roptat>g_bor: did you start working on translating the manual?
<roptat>civodul: make dist creates guix-UNKNOWN.tar.gz, is that expected?
<roptat>also, DIST_DEPENDS_ON_UPDATE_PO is set to no, but po files are still updated by dist
<g_bor>roptat: Actually not yet. I'm focusing on java instead for now. I've also started working on a prometheus service.
<rekado>hmm, recent changes to the repository seem to cause Guix to freeze
<rekado>I’m trying to do ./pre-inst-env guix package -i alsa-plugins but it just sits there.
<g_bor>I've found that mirror://sourceforge generates and url ending with project. It seems, that pipewalker has a different uri scheme, not containing project. Is these an easy way to get rid of that, or is it acceptable to use downloads.sourceforge.net directly, as was in the original patch?
<rekado>it’s fine after a reboot :-/
<rekado>I wonder if that’s related to the bug civodul reported.
<g_bor>It seems that #30831 was reopened. Anyone with more insight? (Update rust to 1.24.1)
<g_bor>It seems it should be closed, we have 1.25.0 now. WDYT?
<jorgesumle>guix pull: error: symlink: Permission denied: "/home/user/.config/guix/latest.new"
<jorgesumle>this error appears when I execute git pull.
<jorgesumle>The file doesn't even exist
<jorgesumle>What should I do?
<rekado>jorgesumle: what version of Guix are you using?
<rekado>does /home/user/.config/guix/ exist as a directory?
<rekado>does the current user running “guix pull” have write access to that directory?
<g_bor>it seems that guix package -s gets relevance really wrong when searching for python...
<g_bor>All python packages are relevance 10...
<jorgesumle>guix (GNU Guix) d801cd0509c9c9523300c41582ec3d8e59e15539
<jorgesumle>Yes, it is a directory and has a file inside called latest
<jorgesumle>No, for some reaso it is root:root
<g_bor>rekado_: python 3.7 expected release date is 2018-06-27. Do we want this in 0.15.0?
<rekado>g_bor: it would cause a lot of rebuilds.
<rekado>jorgesumle: did you run any of this as root?
<rekado>or with sudo?
<rekado>jorgesumle: you need to fix the permissions of that directory
<jorgesumle>I did that
<jorgesumle>I'm waiting to see if it worked
<jorgesumle>So far so good
<rekado>roptat: I installed alsa-lib and alsa-plugins into my profile, but ALSA_PLUGIN_DIR is not set.
<jorgesumle>I still have the annoying "guile: warning: failed to install locale"
<rekado>roptat: I see that there is only a “native-search-path” specification.
<rekado>jorgesumle: make sure that you have “glibc-locales” installed, that the GUIX_LOCPATH environment variable points at the “lib/locale” directory, and that you’re using the same version of the libc.
<rekado>jorgesumle: this is described in the manual.
<g_bor>Ok, so it will only get to the next release. Fine. Just asking because we have an open bug for packaging 3.7 beta2, but I think it is not needed any more. What should be done with it? The patch is almost good for 3.7, but I'm sure that Hartmut has it at hand :)
<g_bor>I think I will leave this alone for now :)
<jorgesumle>rekado: export GUIX_LOCPATH=$HOME/.guix-profile/lib/locale in Bashrc and package installed
<jorgesumle>I'm going to wait to see if the guix pull fixes anything
<jorgesumle>It takes a lot of time
<rekado>roptat: I’ve set ALSA_PLUGIN_DIR manually to $HOME/.guix-profile/lib/alsa-lib and then ran alsamixer, but it crashes :(
<rekado>roptat: error is “munmap_chunk(): invalid pointer”
<rekado>ACTION runs guix refresh -t gnome -u
<jorgesumle>git pull worked
<roptat>rekado: whoops
<roptat>ok, native-search-path was a mistake
<roptat>the crash is weird though
<rekado>ACTION builds lots of GNOME updates on core-updates
<jlicht>hello guix
<roptat>rekado: somehow alsamixer doesn't understand ALSA_PLUGIN_DIR, but aplay does...
<roptat>no crash here
<roptat>oh I think I missed a place where ALSA_PLUGIN_DIR is not properly used
<roptat>that may be the reason for this crash
<rekado>the reason why I want to use alsamixer is that my headphone output seems to be muted, and I can’t seem to find a place in pavucontrol to unmute it.
<rekado>huh, seems fine now.
<roptat>I'll send a patch anyway
<rekado>thanks
<rekado>(I meant: my headphones work now)
<roptat>I don't if that will solve the crash though
<civodul>roptat: the "UNKNOWN" is not expected; could it be cause 'git' isn't in $PATH?
<roptat>oh maybe
<civodul>normally "make dist" creates a sensible tarball name based on 'git describe'
<roptat>no, git is in PATH
<civodul>is it a proper checkout, with the .git directory?
<civodul>does "git describe" work?
<roptat>yes it does
<civodul>alright
<civodul>could you inspect the beginning of the "make dist" output, where it runs build-aux/git-version-gen, etc.?
<roptat>it doesn't seem to run it
<roptat>make dist 2>&1 | grep git-version didn't show anything
<civodul>hmm does it help if you run "autoreconf" beforehand?
<roptat>yes it does, thanks :)
<roptat>I must have not had git at some point
<jlicht>sneek: later tell Rukako: where you able to get anywhere with shepherd + signalfd already :)?
<sneek>Got it.
<jlicht>sneek: botsnack
<sneek>:)
<Digit>hi. after figuring out why i couldnt install, was the daemon's not running, i try start it, but get: "error: cannot bind to socket `/var/guix/daemon-socket/socket': Address already in use" but searching for that gets no results.
<roptat>hi Digit, are you trying to install guix or guixSD?
<Digit>guix, on devuan
<roptat>ok, so you have installed it and it fails to install some package in your profile, or did I not understand the problem?
<Digit>i thought i had installed it, and used it in the past. it's there anyway, and my shell history is littered with guix commands
<roptat>iirc, devuan uses sysvinit, do you have a service for guix-daemon?
<roptat>is there any error message in its log?
<Digit>first i attempted to install some package, n got these two lines: http://dpaste.com/18RY5PR n then (after searching both those lines for a while), i gather the daemon's not running, then i find how n where, launch it, n get that line above. and then i'm stumped
<Digit>no idea how to check guix's log. not yet at all familiarised with its ways.
<Digit>oh, wait, you mean devuan's logs for guix, ... there, i'm just as lost. n_n
<roptat>ok, so running "ps aux | grep guix-daemon" doesn't show anything except grep, right?
<roptat>can you check that /var/guix and /var/guix/daemon-socket are owned by root?
<Digit>yep
<roptat>also, you ran guix-daemon as root, right?
<Digit>yep, both owned by root, and /gnu/store/5drb0ijbszvy8xmps89qcav1p4vy9wqr-guix-0.11.0/bin/guix-daemon is -r-xr-xr-x
<roptat>wow, that's an old guix
<Digit>been a while. hence the rusty unfamiliarity.
<roptat>the internet seems to think it can be a selinux issue
<Digit>well, i dont recall doing any selinux enhancing.
<kristianpaul>o/
<roptat>Digit: sorry, I don't really have other ideas
<Digit>well, thnx anyway, helped me understand the predicament a little better.
<Digit>besides making an attempt to reinstall guix entirely [(later, exhausted)], letting "error: cannot bind to socket `/var/guix/daemon-socket/socket': Address already in use" and the error upon guix package -i whatevers due to guix-daemon not running continuing to taunt me, is my plan.
<Digit>oops, i somehow misread "also, you ran guix-daemon as root, right?" as can run as non-root. o-O idk how my brain does these things. running as root now. this madness ends here. lol.
<roptat>Digit: so does it work now?
<roptat>guix can be run as non-root, but guix-daemon must be run as root
<roptat>civodul: as you can see, I've sent the new pot files to the TP :)
<Digit>yep. tho fresh bumps of oldness to get past, but yep. all's well, can run my guix commands again.
<roptat>cool
<roptat>you should really update guix though
<Digit>guix package -u guix i think was the first thing i tested it on. is that sufficient to update it?
<roptat>no, because this installs the version of guix guix knows of
<roptat>which is before 0.11.0
<roptat>the way you update guix is by running guix pull, but your version is too old to update to recent versions
<Digit>so... a full reinstall needed anyway?
<roptat>I think the only way is to reinstall from the guix binary
<pmikkelsen>hi guix
<apteryx>o/
<nckx>Hm. So I had working Unicode on GuixSD for... about a week, which was nice, I guess.
<mbakke>civodul: Hydra evaluations are failing with "no code for module (sqlite3)".
<g_bor>hello guix!
<davexunit>hello everyone, been awhile
<davexunit>I'm reinstalling guix from scratch on an ubuntu workstation at my office and ran into an issue I'm not sure how to deal with. I followed the binary installation instructions for 0.14 as normal, but get an error when I run 'guix pull' https://paste.fedoraproject.org/paste/kqQIFFXj4pPIEKvkDs8P4Q
<davexunit>is master broken or is an issue with transitioning to the new 'guix pull'?
<davexunit>maybe I need to run 'guix pull' with a snapshot from just before the new implementation got merged, then pull again?
<civodul>heya davexunit
<civodul>still the same old Guile bugs
<civodul>i'm working on it, but i feel very lonely :-)
<civodul>mbakke: oh!
<davexunit>civodul: :( so is this a thing where maybe if I try it again it will work?
<civodul>mbakke: no sqlite3 here: https://hydra.gnu.org/jobset/guix/master#tabs-errors
<civodul>davexunit: yes
<brendyn>civodul: You're not alone!
<civodul>mbakke: oh it's here: https://hydra.gnu.org/jobset/gnu/master#tabs-errors
<civodul>well, it's not new, is it?
<civodul>:-)
<davexunit>how's everyone doing? 0.15 seems to be around the corner?
<civodul>davexunit: doing good! though i'd really like these crashes to be gone :-)
<davexunit>civodul: what do you know about that problem?
<civodul>davexunit: very roughly it looks like a GC issue, where items on the VM stack are reclaimed too early
<civodul>actually the 'with-monad' thing might be different
<civodul>hard to tel
<civodul>+l
<civodul>see https://bugs.gnu.org/27476
<civodul>backdoor bundle: https://arstechnica.com/information-technology/2018/06/backdoored-images-downloaded-5-million-times-finally-removed-from-docker-hub/
<davexunit>civodul: ah, thanks for explaining
<davexunit>civodul: and wow, there's a serious real-world example of why Docker isn't so great
<civodul>yup
<davexunit>make sure to include that in every guix talk now
<davexunit>;)
<numerobis>Hi guix! Beginner question: How do I select the version of a package to install using "guix package -i"? I saw that both version 1.64 and version 1.66 of Boost are present, but can't find the way to specify the one to install....
<civodul>numerobis: you can run "guix package -i boost@1.64"
<civodul>see https://www.gnu.org/software/guix/manual/html_node/Invoking-guix-package.html
<numerobis>Thanks! I tried - and _, but not @
<civodul>heh :-)
<civodul>ACTION has to go
<numerobis>And sorry, I have read the manual but somehow I missed that...
<pkill9_>who is working on adding KDE again?
<Digit>think it will be safe enough keeping the same old /gnu/store from my 0.11.0?
<Digit>i'm currently following https://www.gnu.org/software/guix/manual/html_node/Binary-Installation.html and upon the mv of step2, "mv: cannot move 'gnu' to '/gnu': Directory not empty". would it manage to "just work" and pick up all the old guix's packages?
<pkill9_>yeah, not neccessarily profiles though, but guix will use substitutes safely if they're already in the store
<davexunit>Digit: I don't recommend doing that.
<davexunit>guix keeps a sqlite database for the entire contents of /gnu/store
<pkill9_>oh nvm ignore me
<davexunit>so if you dump a new binary installation on top of an existing /gnu/store, guix will no longer know about those old things
<pkill9_>i don't wanna give false info
<davexunit>any profiles that refer to those old items will still work just fine, but they are all orphaned, basically
<Digit>ACTION nods
<davexunit>I actually just did a fresh install on a system that already had guix today
<davexunit>I noted which packages I had installed, wiped out all the guix directories, and reinstalled
<Digit>so, i might as well rm -rf /gnu. presumably, no harm in that, as far as i see... besides losing all that, n starting with a fresh slate. ... idk if there's a simple world file (like gentoo has) that wont quibble over specific versions n deps needing listed too, that i could extract from /gnu or /var/guix (already gone, oops) or /home/digit/whereverthatconfigisat?
<davexunit>I ran 'guix package --list-installed' and saved the output
<Digit>yeah, my old guix was already broken enough i'm not sure even --list-installed would have workd. n_n too late to find out now.
<davexunit>yeah if /var/guix is gone you won't be able to even look at the old manifest file
<Digit>also, since i already removed /var/guix, i facepalm in the general direction of: ls -lha .guix-profile \\lrwxrwxrwx 1 digit digit 46 Oct 7 2016 .guix-profile -> /var/guix/profiles/per-user/digit/guix-profile
<Digit>yep
<davexunit>I typically keep a Scheme file that produces a manifest checked into a git repo so that I have a definitive list of what software I use
<davexunit>and run 'guix package --manifest=manifest.scm'
<janneke>huh, what's going on here? linked with -static:
<Digit>no inbuilt feature equivalent portage's world file in guix? what i've seen so far, guix likes to be fussy on specifics, rather than human fuzzy.
<janneke>/gnu/store/dc8sgwajvwglw81m4bflvkk9rkmhzfq0-binutils-boot0-sin-2.10.1/bin/ar
<janneke>bash: /gnu/store/dc8sgwajvwglw81m4bflvkk9rkmhzfq0-binutils-boot0-sin-2.10.1/bin/ar: No such file or directory
<janneke> ldd /gnu/store/dc8sgwajvwglw81m4bflvkk9rkmhzfq0-binutils-boot0-sin-2.10.1/bin/ld
<janneke>$ file /gnu/store/dc8sgwajvwglw81m4bflvkk9rkmhzfq0-binutils-boot0-sin-2.10.1/bin/ld
<janneke>/gnu/store/dc8sgwajvwglw81m4bflvkk9rkmhzfq0-binutils-boot0-sin-2.10.1/bin/ld: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), dynamically linked, interpreter /lib/ld-linux.so.2, for GNU/Linux 2.6.32, with debug_info, not stripped
<janneke>ow, bah! libtool removes the -static again
<janneke>sorry for the noise, thanks for listening
<nckx>Could someone briefly explain what ‘warning: rewriting hashes in `%1%'; cross fingers’ means? I've always dismissed it as graft-related but it happens during --source builds too.
<nckx>warning: rewriting hashes in `/gnu/store/gz6l0arf9yjylaqnxpz2cz2bzwz1v4gj-patchage-1.0.0.tar.bz2'; cross fingers
<janneke>OriansJ: congrats or your M2-Planet release!
<linuxhelphumanit>is it posible to install guixsd in a desktop computer ?
<roptat>linuxhelphumanit: yes, that's what I'm using now :)
<linuxhelphumanit>when you install the guixsd which size does the boot partition needs in mb
<janneke>linuxhelphumanit: i stopped using separate partitions
<janneke>depending on your needs, 30GB is a reasonable size for /
<janneke>but as a developer i often need more in the store, and i don't want to sacrifice that to a separate boot partition
<taylan>boot partition or root partition? :)
<sneek>taylan, you have 1 message.
<sneek>taylan, nckx says: running ‘sudo fc-cache -f’ fixed that for me. You can check fontconfig's worldview with ‘fc-list’.
<taylan>ooh
<taylan>sneek: botsnack
<sneek>:)
<janneke>hmm, oops -- can guix have a separate boot partition?
<linuxhelphumanit>i am just going to create a boot and a root partition based on that why size do you recomend me to use for boot ?
<taylan>this page contains an example with an EFI /boot partition: https://www.gnu.org/software/guix/manual/html_node/Using-the-Configuration-System.html
<janneke>f
<janneke>f
<taylan>I don't know either to be honest but I would assume that 1G should be more than enough for /boot...
<taylan>the Ubuntu Community Help Wiki says 250M to 1G :P
<janneke>yes, sorry linuxhelphumanit, i never used separate boot -- i assumed root
<taylan>linuxhelphumanit: I'd go with 1G unless you have a tiny disk and want to save space
<taylan>(tiny nowadays meaning anything below 100G I suppose?)
<roptat>according to du, my /boot takes 10MB
<taylan>:)
<roptat>but it's not separate
<roptat>so it doesn't have any kernel or initrd
<roptat>not sure if that would be the case with separate partitions
<taylan>oh, where does guix put the kernel and initrd if not on /boot?
<roptat>/gnu/store
<taylan>oooh :)
<g_bor>I've managed to fill a 300MB boot using Ubuntu, and not using autoremove, it took about 5 years :)
<roptat>maybe forever with guixsd then :)
<g_bor>I have seen a patch adding kdenlive on the queue. There is aslo another related patch, which seems to be not sorted out. I added a block. It would be nice to have this for the videos project.
<g_bor>The problematic bug is tracked as #30770. Should we contact the author again, on try to fix this?
<nckx>ACTION uses a 5 MiB /boot. Probably ‘don't try this at home, kids’ territory.
<nckx>s|/boot|/boot/efi| though.
<nckx>taylan: Did fc-cache help anything? I'm curious since I've been having my own font trouble since I left that message.
<pkill9_>what will it take for the linux desktop to have a solid user interface?
<g_bor>Is core-updates now open?
<nckx>g_bor: Yes, I think so.
<g_bor>ok, then I will send a reminder to push the doxygen update.
<g_bor>I cannot make sense of #30256: scripts: environment: Add --no-cwd.
<g_bor>Anyone with more insight?
<roptat>civodul: I need a good translation for "relocatable" :)
<linuxhelphumanit>but do i necesary need a boot partition or can this os be installed just with a root and a swap?
<roptat>you don't need a separate /boot (although you'll need a /boot folder)
<roptat>directory*
<linuxhelphumanit>i mean that when i am partittioning tmy dik to install this os i just need a root and a swap partition?
<nckx>linuxhelphumanit: Depends on whether you boot with BIOS or UEFI.
<roptat>exactly
<linuxhelphumanit>i boot with bios
<roptat>that's the case with other any other GNU/Linux distro btw
<nckx>OK. Then you don't need multiple partitions.
<nckx>(In fact, GuixSD doesn't really support a separate /boot in cases where you'd actually *need* it, yet. But that won't affect your simple setup.)
<linuxhelphumanit>i have a problem desscompressing the xz file when i try to execute the xz -d command there it says bash: xz: no se encontró la orden
<linuxhelphumanit>when i try to execute the xd -z command there its say bash: xz: order not found
<roptat>so that's a bash error?
<nckx>linuxhelphumanit: Which OS are you running now?
<civodul>roptat: how about "relocalisable" or "relogeable" ?
<linuxhelphumanit>i am using trisquel 7.0
<linuxhelphumanit>i am using Trisquel 7.0 sugar toast last word
<nckx>linuxhelphumanit: Then you need to install the xz-utils package, if my sources are correct.
<linuxhelphumanit>how do i do it?
<nckx>...I don't know.
<linuxhelphumanit>executing the sudo apt-get install xz-utils command in terminal?
<taylan>nckx: let me try if fc-cache helped just now. brb, restarting stuff.
<nckx>linuxhelphumanit: I guess so. I've never used Trisquel.
<taylan>works
<roptat>civodul: I went with "repositionnable"
<nckx>taylan: \\o/
<nckx>ACTION giggles ‘I helped someone!’
<civodul>roptat: even better, well done
<taylan>nckx: thank you! ^^
<pkill9_>i wish to add one or a few packages to guix at some opint
<pkill9_>point*