IRC channel logs


back to list of logs

<xavierm02>nckx: Why ****?
<nckx>xavierm02: ;; Chromiums build processes may consume up to 8GiB of memory per core. | ;; Disable parallel builds to avoid thrashing end user systems. | #:parallel-build? #f
<nckx>xavierm02: I won't bore the room with my opinions on that piece of s**tware ☺
<rubic88>I just have alsa-utils installed and it's working for me.
<nckx>rubic88: Maybe you use %desktop-packages or whatever it's called and it's included?
<nckx>sneek_: later tell rubic88: Maybe you use %desktop-packages or whatever it's called and it's included?
<bzp>thank you very much, I can increase and reduce the volume :D
<bzp>How do I solve the problem of themes gtk in guix?
<erudition>oh yeah I tried installing chromium the other day and it took me a while to realize why the screen was off when I came back... every time I tried it. Looks like I found a reliable way to completely hose a system with only 4G of ram
<erudition>gave up on that pretty fast
<Marlin[m]>Do you guys want me to make substitutes for chromium?
<Marlin[m]>it builds pretty fast on my machine
<erudition>If you want - I don't need it yet on that PC, but in general I think we should be using torrents for substitutes anyway so that would be the ideal
<erudition>oh my bad, I just looked up ungoogled-chromium, I didn't realize it's in guix. I had tried installing normal chromium from an unofficial channel, but I'll try this ungoogled one
<xavierm02>Is there some smart way of having things not show up as /gnu/store/<stuff>/thing.thing in top etc?
<nckx>xavierm02: ‘p’
<OriansJ>xavierm02: use htop
<nckx>* in htop.
<nckx>Which I added looong ago for exactly that reason in Nix ☺
<nckx>‘c’ in top.
<nckx>which has the worst man page evar.
<nckx>(Use htop though, it's better.)
<xavierm02>Perfect! thank you :)
<nckx>Ooh. It might already be installing files!
<xavierm02>How much can I trust packages in Guix? When you commit a package to guix, are you supposed to understand the code base of the project enough to ensure that there's nothing fishy, or just to ensure that it build, and check that there's nothing fishy is delegated to the user?
<xavierm02>and check -> and checking
<xavierm02>chromium build and working!
<xavierm02>so 7 hours and 20 min. But I installed something else that needed compiling, so it's probably a bit less
<bzp>someone knows how to automatically mount a pendrive from ranger?
<rubic88>xavierm02: Awesome!
<sneek_>rubic88, you have 1 message.
<sneek_>rubic88, nckx says: Maybe you use %desktop-packages or whatever it's called and it's included?
<nckx>bzp: That's not ranger's job but something like udisk's (which I think is part of %desktop-services I believe).
<nckx>rubic88: That was about pactl. Bad nckx for not including more context.
<rubic88>nckx: Understood. You're probably correct that it is included in %desktop.
*nckx → 😴
<nckx>xavierm02: We don't have the manpower to audit packages more than any other distribution.
*nckx → really 😴
<xavierm02>There's hope to run everything in a very contrained environment though
<rvgn>xavierm02, nckx: I added the user to libvirt group but still couldn't connect to libvirtd.
<rvgn>I am able to connect to libvirtd by running virt-manager as root. Do anyone know how to make any user eligible to connect to libvirtd??
<OriansJ>rvgn: simply add the user to libvirt
<rvgn>OriansJ I did. But still not connecting.
<rvgn>permission denied.
<OriansJ>rvgn: did you logout after being added to the group?
<rvgn>I rebooted after doing guix system reconfigure with user added to group libvirt.
<OriansJ>rvgn: and what do you get when you run: sudo ls -hal /var/lib/libvirt/
<OriansJ>and ls -hal /etc/libvirt/storage/
<rvgn>total 16K
<rvgn>drwxr-xr-x 4 root root 4.0K Jun 21 20:21 .
<rvgn>drwxr-xr-x 11 root root 4.0K Jun 21 16:45 ..
<rvgn>drwx--x--x 2 root root 4.0K Jun 21 20:21 images
<rvgn>drwxr-xr-x 8 root root 4.0K Jun 21 18:56 qemu
<rvgn>OriansJ that was the output
<OriansJ>rvgn: and your /etc/libvirt/libvirtd.conf has unix_sock_group = "libvirt" and unix_sock_rw_perms = "0770"
<rvgn>OriansJ Sorry, are you asking me to check or something else?
<OriansJ>rvgn: yes I am asking you to check the contents of your /etc/libvirt/libvirtd.conf
<OriansJ>(also make sure the group libvirt exists and the user account is a member of it)
<rvgn>That files doesn't exsist
<rvgn>OriansJ How can I check for a group and it's users?
<OriansJ>less /etc/group
<OriansJ>you can find the file by name alone with: find / -iname 'libvirtd.conf'
<pkill9>you might need to specify the group in the libvirt configuration
<PotentialUser-23>Hi, I have a question about guix store locking, how does gc know not to collect derivations that are being used for builds in progress?
<rvgn>pkill9 You mean in config.scm?
<rvgn>pkill9 Shouldn't that be default?
<pkill9>i would think so, but idk, you might need to specify it anyway, i have in mine and i seem to remember that it may be needed
<rvgn>pkill9 Cool, I'll try that.
<pkill9>a while back i was trying to get it working and virt-manager kept failing to connect, and that may have fixed it
<pkill9>ah yea, looking at the service definition, it seems the default group is set to "root"
<rvgn>pkill9 Did you declare "tls-port" as well?
<pkill9>no, just "unix-sock-group"
<rvgn>(service libvirt-service-type
<rvgn> (libvirt-configuration
<rvgn> (unix-sock-group "libvirt")))
<rvgn>pkill9 Correct??
<pkill9>yea that's what i have
<rvgn>Awesome. Thanks a lot.
*rvgn is reconfiguring the guix system and will be back after reboot.
<pkill9>rvgn: also consider using `guix system vm` to run a virtual machine built using your guix configuration and testing with that to avoid having to reboot :>
<rvgn>pkill9 It works now. thanks a lot.
<rvgn>pkill9 Oh wow! I didn't know that. So how it works? Does "guix system vm" creates and starts a vm?
<PotentialUser-23>The manual explains system vm pretty well
<pkill9>it creates a script that when you execute will start a vm with qemu
<pkill9>it returns the script path in stdout, so if you wrap the command with "$(guix system vm config.scm)" then it will build and start the vm
<rvgn>pkill9 Thanks. Btw, Now I am able to install an vm in virt-manager as it gives error that "virtlogd" not found.
<rvgn>pkill9 Should I enable service virtlog-service-type as well?
<pkill9>yea i think so
<pkill9>i have it in my config
<rvgn>Cool Thanks!
<pkill9>does that error prevent you from running a vm?
<rvgn>PotentialUser-23 Thanks! I will check it out.
<rvgn>pkill9 Do you use "bridge" or "passthrough" for networking in VM?
<rvgn>g_bor[m] nckx Just leaving you both a update message. My training modules' flash components works with gnash in trisquel. But the system is too slow inside virtual machine as my host is X200T. So if could make gnash work in guix, it would be awesome. Thanks!
<minall>Hello guix!
<minall>I'm having a little problem with xorg, I request help :D
<PotentialUser-23>what is the problem?
<str1ngs>rvgn: IIRC you need virtlog yes
<minall>I can't start xorg, and because of that, when I start the pc, It goes on a loop when it tries to start GDM, can't, and starts again, and my screen goes like on/off, until GDM does goes on an error of 'to many open files', and stops trying to load GDM after the process of GDM number like 503, and then i can use my pc normally
<minall>I have some errors that I think may be interesting on my /var/log/gdm/greeter.log,
<minall>The first one, that I think is of importance, "(EE) Unable to locate/open config directory: "/gnu/store/fjvhaqqi26kjgypr03bd3c0znha1dxkv-xorg.conf.d""
<minall>then followed by failed to load module fbdev, module doesn't exist, and vesa, module doesn't exist
<minall>And then, the one that i think links to the first error: "(EE) Screen(s) found, but none have a usable configuration."
<minall>Fatal server error:
<minall>(EE) no screens found(EE)
<minall>I'm at a loss, i don't know what to do
<ilyaigpetrov>hi. I want to crosscompile guix system image for raspberry pi zero wh, what exact command do I need?
<ilyaigpetrov>the docs has this: guix system disk-image --system=armhf-linux -e '((@ (gnu system install) os-with-u-boot) (@ (gnu system install) installation-os) "A20-OLinuXino-Lime2")'
<ilyaigpetrov>But I'm not sure about u-boot and the board name
<ilyaigpetrov>ah, I guess building system for arm is not yet possible
<emacsomancer>when I try to offload builds, I eventually end up with an error, despite `guix offload status` producing reasonable looking ouput
<emacsomancer>ERROR: In procedure put-bytevector: Throw to key `guile-ssh-error' with args `("write_to_channel_port" "Remote channel is closed" #<input-\output: channel (open) 2738b80> #f)'.
<emacsomancer>though it takes a while to get to that point, displaying a number of "process NNNN acquired build slot '/var/guix/offload/IP.ADD/0' \n "load on machine ...." messages
<g_bor[m]>ilyaigpetrov: hello, I am also interested im this, but it seems that we need an additional platform. The zero has a different cpu from the rest of the modern pi-s, wich are now supported.
<ilyaigpetrov>g_bor[m]: have you tried installing guix from on raspbian? I get segmentation fault.
<ilyaigpetrov>> Generally, your best bet is to use Raspbian - which is (mostly) Debian armhf rebuilt by members of Debian for the RPi's ARMv6+VFP2 ARM variant.
<ilyaigpetrov>so we need rebuilding
<g_bor[m]>ilyaigpetrov: no, our current armhf-linux is not compatible with raspberry zero. The correct triplet would be something like arm-linux-gnueabihf.
<g_bor[m]>Yes, that is the one.
<g_bor[m]>I believe that vagrantc might have something for you...
<ilyaigpetrov>will try pinging him once he is online or will leave him a memo
<g_bor[m]>ilyaigpetrov: ok. Do you know sneek?
<g_bor[m]>sneek: botsnack :)
<ilyaigpetrov>thanks, done
<nckx>rvgn: OK, so the libvirt service should depend on virtlog if it won't work at all without it. I'll make a note of that.
<nckx>Following the upstream default socket group of ‘root’ when *everyone* else uses ’libvirt’ doesn't sound very wise either.
<nckx>rvgn: Thank you for your report and I hope someone has time to [help you] get Gnash working on Guix this weekend. ☺
<mbakke>Howdy Guix!
<mbakke>I will be AFK until Wednesday or so.
<ArneBab_>packaging rust is awful :-( for i in $(for i in $(grep -o ',[^ ()][^ ()]*' more/packages/gnuzilla.scm | sed s.,.. | xargs); do grep -q "define-public $i" more/packages/gnuzilla.scm || echo $i; done | sort -u | grep rust | sed s/^rust-// | xargs) ; do guix import crate $i >> more/packages/gnuzilla.scm || (sleep 30 && guix import crate $i >> more/packages/gnuzilla.scm) ; done
<A0x1E>Yeah, who make a fix for version 1.30.1?
<ArneBab_>downloading missing rust packages iteratively: for i in $(for i in $(grep -o ',[^ ()][^ ()]*' more/packages/gnuzilla.scm | sed s.,.. | xargs); do grep -q "name \"$i\"" more/packages/gnuzilla.scm || echo $i; done | sort -u | grep rust | sed s/^rust-// | xargs) ; do guix import crate $i >> more/packages/gnuzilla.scm || (sleep 30 && guix import crate $i >> more/packages/gnuzilla.scm) ; done
<ItsMarlin>mornin guix
<ItsMarlin>i was talking with stallman about my amd gpu and the proprietary firmware
<ItsMarlin>he asked if i wanted to work in reverse engineering it
<ItsMarlin>hopefully i can do something useful
<Gamayun>ItsMarlin: Nite from here... ;) Sounds like a hefty task...
<OriansJ>ItsMarlin: you might find this useful
<OriansJ>The coreboot project is using it to reverse engineer firmware binaries themselves
<janneke>would it be OK for the emacsy package to put guile-2.2 in propagated-inputs?
<mfg>hi, did anyone try to compile sml/nj or mlton for guix?
<mfg>or maybe even cakeml :D
<OriansJ>mfg: well it doesn't ring a bell but if there is a guix package for it; then the answer is definitely a yes
<mfg>OriansJ: Hm than that's a no :D
<OriansJ>mfg: plus there are some languages we haven't bootstrapped yet; thus are either binary rooted (working on removing those [like haskell]) or not in the list yet
<mfg>Ah i see. The good thing is sml/nj can be bootstrapped with a C Compiler and mlton with sml/nj.
<mfg>But this seems to only work on 32-bit systems because sml/nj doesn't fully support 64-bit yet.
<OriansJ>mfg: one only needs to do it correctly once in a guix script and share it and everyone gets it
<OriansJ>mfg: most 64bit systems have backwards compatability modes for 32bit binaries
<mfg>yes i know, but the build script complains that it can't generate 32-bit executables. Maybe gcc is 64-bits only when i'm on x86_64 guix?
<mfg>(the distro not the pkg manager)
<OriansJ>mfg: well perhaps the question is how to use the 32bit version of gcc and binutils to produce the desired result?
<mfg>Also the build system is a single shell script so i'm not sure which build-system i should try in the package definition.
*mfg is away for a couple of hours
<tune>ohhh man I'm getting some font breakage in new terminals
<tune>using font-terminus
<tune>seeing all empty squares
<tune>ah I think this happened months ago. I gotta install font-terminus:pcf-8bit
<tune>the normal one is broken or something. I had both in my profile but when I applied a manifest I didn't have the 8bit version
<tune>okay, yeah that fixed it. thank goodness.
*kmicu puts on a tin foil hat: OriansJ works on bootstrapping, works for gov, and recommends NSA tools.
<OriansJ>kmicu: good, now don't trust any code that I write (I could have done anything)
<OriansJ>in fact the more people who don't trust and audit the F&ck out of the code I submit the better
<OriansJ>No one should have to trust me to have a trusted bootstrap; that is why one only has to audit the 2,874 bytes of hex code and generate their own 357byte hex0 assembler
<OriansJ>Then they need only audit the 4Kish lines of cc_x86.M1 (assembly) then everything else is straight forward C code to audit
<OriansJ>all of the pieces can easily be replaced by anyone interested in doing so; just like they are all easy to port to arbitrary hardware
<arshin>Are there any ideas floating around for "distributed verification" of FOSS projects? like a blockchain that says "these 43 nodes verified this project".
<OriansJ>arshin: yes guix challenge and public disclosure
<OriansJ>kmicu: all of the code is 100% deterministic and one needs only build their own hardware and bootstrap 5 pieces to check all platforms
*arshin reading about guix challenge. Is "public disclosure" some project or just a generic term for putting source code online?
<OriansJ>arshin: it is a generic term for making the entire world aware of a discrepency (I think gitian does something similar though)
<OriansJ>kmicu: and even if you trust me, assume my machine could have been compromised by someone else and audit the code and generate your own binaries
<arshin>OriansJ: thanks for the pointers. "making the world aware" - just by word of mouth (email lists, irc etc) or is there something blockchain-like to build public record of verifications? Guix challenge seems to be only about "local" verification
<OriansJ>arshin: well all validation chains must start locally but I am sure there are somethings out in the blockchain world that might extend this further (I just haven't had time to check yet)
<ArneBab_>I packaged around 400 rust packages, but now I get errors trying to install:
<ArneBab_>guix/build-system/cargo.scm:128:4: Invalid keyword: ("rust-lazy-static" #<package rust-lazy-static@1.3.0 more/packages/gnuzilla.scm:2148 3b59370>)
<ArneBab_>curiously I can start an install of rust-lazy-static directly.
<minall>hello guix!
<minall>Where can I find all the configuration files of the guixsd image ? for example, /etc/configuration/bare-bones.scm
<nckx>sneek_: later tell minall: They are renamed from <guix.git>/gnu/system/examples/*.tmpl .
<sneek_>Got it.
<nckx>sneek_: tasty botsnack.
<nckx>tune: Yeah, the upstream Terminus project stopped installing those 8-bit codepages entirely. My compromise was to put them in a separate output so Terminus is still Terminus as released by upstream, but keeps providing files that many need.
<nckx>o/ bzp.
<bavier`>hi guix
<xavierm02>Is there a good reason for the (gnu system) module not to export <operating-system>? Because of this, I can't match-record on an <operating-system> record.
<rvgn>Hello Guix!
<rvgn>Folks! Could anyone able to help me with this please?
<xavierm02>since the VM doesn't work, you could try booting on your other distro in a usb key
<rvgn> xavierm02 The issue is I do not have access to any USB drives atm. :(
<xavierm02>then you can try to rezise your partition to make a bit of space on your laptop
<xavierm02>or use a dvd or w/e
<rvgn>Good Idea!
<emacsomancer>I'm getting the sort of error msg described here: when I try to do 'guix copy' or 'guix offload'
<emacsomancer>what's not clear to me is how to resolve it
*nckx sees already uses that \o/
<bandali>nckx, hey, can you update pages?
<nckx>bandali: Probably.
<nckx>I haven't tried since the subdomain migration.
<bandali>nckx, cool. i see that the newest blog post on is missing on
<bandali>i think ludo forgot to regenerate/update
<bandali>hopefully i’ll get around to setting up redirects for to in the next couple of days, and from there onwards there’ll only be one website
<rvgn>nckx bandali Is it possible to use as main site and forward all requests for to
<rvgn>Oh wow! That's one hell of a co-incidence bandali xD
<bandali>rvgn, that’s the plan. i still haven’t gotten around to doing it
*nckx was confused there for a moment yes 😛
<bandali>i think it should be possible
<rvgn>bandali Cool!
<bandali>perhaps a bit tedious, since it might require individual redirect rules, instead of a nice regex
<nckx>bandali: Hm, I have no idea how the ‘new’ site is set up.
<nckx>bandali: Ew. Why's that?
<rekado>rvgn: yes, we will use eventually
<rekado>there’s still some work to be done there before the switch
<rekado>but I’m running low on free time to do it.
<bandali>nckx, i see. as for why, i think is a bunch of manual redirects
<rvgn>rekado Great
<rekado>I’m responsible for the lack of an update to
<rekado>will try to do it this weekend.
<rvgn>Ah I see
<bandali>though, maybe a simple htaccess rule to redirect to on the cvs repo might suffice
<nckx>rekado: How is the new site updated? I assume (s//hope/) that CVS is gone, but is some manual prodding still needed?
<rekado>bandali: yes, but someone needs to check what rules need to be set up
<rvgn>rekado Also, the SSL is missing at correct?
<nckx>I'd like to be able to help with such trivial fixes if needed.
<bandali>rekado, right. that someone is me :p i’ve been super busy, but hopefully in the next couple of days :)
<rekado>bandali: I don’t have time now to do this so I’ll just rebuild the site .
<rekado>rvgn: no, TLS works fine.
<nckx>Can both of y'all stop being so busy and get to work!
<bandali>rvgn, no i think https for was sorted out
<bandali>lol nckx
<rekado>nckx: no CVS. It’s really just pointing at, which is a little directory in my home directory on berlin…
<rekado>more duct tape
<bandali>one day, maybe one day, we could have proper automated deploys of the site via a ci service :)
<rvgn>rekado bandali Ah now I see https works at if I manually specify https. Possible to setup forced-https??
<bandali>as a gnu webmaster and someone who has to deal with various sites, i’m pretty happy with the overall state of things
*nckx <half-serious> has rms changed his mind about forcing HTTPS now?
*nckx is referring to a probably almost decade-old opinion by now, never mind.
<nckx>(Forcing was bad because freedom to type http:)
<bandali>one of my longer term goals is to allow people to use git instead of cvs for their projects’ web pages
<rekado>one of the reasons why we pushed for moving the project HTML pages out of CVS is because in the past we would regenerate them quite often, which wasn’t welcome.
*bandali sighs
<rekado>the first step then was to disable regeneration of the package list.
*rvgn confused with the difference between git and CVS.
*rvgn wonders if git is one type of CVS?
<bandali>rvgn, cvs is a specific software
<nckx>rvgn: CVS is a type of VCS, and git is another. Confusing!
<rekado>rvgn: CVS is a … mature version control system with a central server and slow operations. “git” is a little more recent, pretty fast, decentralized (no central server needed, but often used in practise).
<bandali>perhaps you’re think of vcs
<nckx>(VCS: version control system.)
<nckx>(CVS: a very old VCS)
<rvgn>bandali Oh yeah, that's VCS.
<rvgn>rekado Gotcha!
*rekado just updated
<rekado>sorry for the delay
<bandali>nckx, re https: it seems still doesn’t redirect to https
<bandali>so perhaps rms still hasn’t changed his opinion ?
<nckx>I wish I had a clue where to find that mail (I swear I'm not making it up) but good luck searching the Web for ‘https’ :-/
<rvgn>bandali For me it redirects to https
<bandali>rvgn, that may be your browser’s auto-complete history
<bandali>it doesn’t for me. i guess one could verify with curl perhaps
<rvgn>Ah I see.
<bandali>or maybe httpseverwhere if you use that addon
<nckx>curl -LI confirms.
<rvgn>bandali yes I use https-everyehere extension
<nckx>rvgn: It's something else still: HSTS.
<nckx>‘Strict-Transport-Security: max-age=63072000’
<nckx>So if you visit it using https:// once, your browser will force HTTPS for that many seconds afterwards.
<grubles>is it possible to easily enable non-free firmware loading (eg for a wifi nic)
<nckx>But http: works if you haven't (or if you clear settings or use curl).
<rvgn>nckx I see.
<nckx>grubles: We can't help you with that here (or any official Guix communication channel), sorry.
<grubles>that's fine. thank you.
<nckx>(And this isn't some subtle ploy to now tell you an ‘unofficial’ channel.)
<grubles>i think i have a wifi card that has free firmware. just need to swap out the current one.
<nckx>grubles: I hope you do ☺ I've lost my ath9k USB dongle somewhere in the house ☹
<bavier`>why do I keep having to build mariadb everytime I do a 'guix package -u' ?
<bavier`>rhetorical question ^
<rvgn>nckx I came across something shocking. Remember the patch you did for youtube-viewer? It stopped working after I did fresh reinstall of guix system couple of days ago. I double checked after doing guix package -u twice. Still did not work.
<nckx>Uhm. I don't *recall* it depending on .config/ or whatever… Weird?
<rvgn>On double-click, only SD videos opens and plays, not HD videos.
<nckx>rvgn: Could you check if ~/.config/youtube-viewer/<your-favourite>.conf doesn't contain anything fishy that could override the patching I applied in that commit?
<rvgn>nckx I could do that.
<nckx>And give me an example of a video that doesn't work for you?
<nckx>Thanks ☺
<rvgn>nckx Same one as last time. Search for "best pop songs". Title is New Pop Songs Playlist 2019 - Billboard Hot 100 Chart - Top Songs 2019 (Vevo Hot This Week). Thumbnail will be Ariana Grande.
*nckx installs yt-v again.
<nckx>(It's a handy tool & I use it once in a while but you<TAB> for youtube-dl is just too ingrained in my muscle memory.)
<nckx>rvgn: I'm ‘afraid’ that double-clicking Ariana's face pops up MPV just fine here.
<nckx>Let me retry.
<nckx>Still works after restarting yt-v.
<nckx>rvgn: Does it print anything useful when you start gtk-youtube-viewer from a terminal?
<nckx>Here it's only ‘2019.06.21’, the version of youtube-dl being used.
<rvgn>nckx Just a secx
<rvgn>Gtk-WARNING **: Unable to locate theme engine in module_path: "adwaita", at /gnu/store/xsd6a9yki1lc5w9x62yr7l7n3bslyncr-perl-gtk2-1.24992/lib/perl5/site_perl/5.28.0/x86_64-linux-thread-multi/ line 126.
<nckx>rvgn: Just to rule out silly bugs: I double-click the thumbnail, not the text.
<rvgn>Other than that. No other errors while opening gtk-youtube-viewer from terminal.
<nckx>Hm. Shouldn't be related but you could install adwaita-icon-theme just to make sure.
<nckx>Grr. Why u no error.
<rvgn>That warning i usual but worked fine after you patch.
<rvgn>How can I check which guix commit I am using?
<nckx>guix describe
<nckx>Will tell you which guix you've last pulled; having updated yt-v with that guix is your responsibility.
<rvgn>Oh wait, the actuall error shows up in terminal when I double click a video. Just a moment
<rvgn>nckx Please find the error at
<nckx>rvgn: Your youtube-dl is not the latest version, which means your youtube-viewer isn't the latest version; see my last message above ☺
<nckx>guix pull && guix upgrade youtube-viewer
<nckx>This will update youtube-dl too.
<lsl88>hi guix!
<rvgn>nckx Intersting I did guix pull and guix package -u thrice after fresh reinstall of guix system.
<rvgn>Anyway, I retry.
<nckx>rvgn: guix gc --references `which youtube-viewer` | grep youtube-dl
<rvgn>nckx Question: If I do "sudo guix system reconfigure" as "regular user", which guix version will be used? latest version in root user profile or recent version in regular user profile???
<nckx>should say 2019.06.21.
<nckx>rvgn: Is youtube-viewer in your system-packages?
<rvgn>nckx No I am just asking
<nckx>On Guix System, it should use rvgn's guix, not root's. You can run ‘sudo which guix’ to make sure.
<nckx>(Or whatever you call yourself ☺ )
<rvgn>Cool! Thanks
<rvgn>You are right. The version I have currently is 2019.05.20
<nckx>Hmm. Assuming you've pulled, ‘guix upgrade youtube-viewer’ (or ‘.’ for everything) should *really* fix that. o_O
<nckx>Hi lsl88! o/ I lost you between the other messages.
<lsl88>nckx: heyyy! :)
<lsl88>I'm talking to a folk that is really interested in Guix, and of course I am promoting, like usual, both Guix and the community ;)
<lsl88>(we need to publish the videos)
<lsl88>there was a paper about the differences with Nix, may anybody tell me where to find it?
<rvgn>nckx Worked like a charm. Thanks!
<nckx>rvgn: 👍
<nckx>lsl88: Actually I'm not sure which paper you mean, at all.
*nckx throws random Guix papers at lsl88.
*lsl88 thanks Tobias! I was talking about the second one
*lsl88 sends Tobias to bed early
<nckx>Oh ffs it's ten o'clock!
<lsl88>nckx: your mother says go to bed at most at 12
<rvgn>nckx Is there special meaning to your nick. if you don't mind me asking.
<nckx>rvgn: It's a somewhat arbitrary substring of my name (Geerinckx-Rice); it used to be unique on the Web until somebody decided to name a gene after me.
<nckx>That & 4 consonants in a row being unusual for anglophones, so I've had to spell those 4 letters all my life.
<nckx>lsl88: By sheer coincidence and totally not because you told me so, mum, I probably will.
<rvgn>nckx I see
<lsl88>nckx: hahahah ;) happy to hear that you are gonna get plenty of sleep :) and I liked the story of your nickname
<nckx>rvgn: …so it means absolutely nothing but is very tied to ‘me’ ☺ Hope that was a non-disappointing answer.
<leungbk>I've been struggling with this package definition for the past couple of days---would anyone be willing to offer some help?
<leungbk>The issue is that once I use the #modules argument on line 176, I get:
<leungbk>ERROR: In procedure symlink:
<leungbk>In procedure symlink: File exists
<leungbk>Without the #modules argument, the build succeeds, but I need the #modules there to install the Elisp file.
<nckx>leungbk: Is that supposed to be added to the end of haskell-something.scm?
<leungbk>Yes, for haskell.scm. `#:use-module (gnu packages emacs)` needs to be added up at the file also.
<leungbk>*the top of the file
*nckx stops trying to turn the file into something stand-alone.
<rvgn>nckx LoL nothing disappointing at all. I was just curious.
<nckx>leungbk: So, with the major caveat that I don't actually know anything about the Haskell build system: when I comment out #:modules and its modules, I get the same symlink error.
<leungbk>nckx: Sorry, I misremembered. Seems that commenting out both #:modules and #:imported-modules (so, empty arguments basically) causes the build to succeed.
<leungbk>I was sort of hoping the symlink error was caused by some obvious syntax issue with the pkg def that someone could spot in a flash, haha. No worries though, thanks for trying to help.
<leungbk>I've tried jamming in other modules in hopes that the error is simply caused by missing module imports, but no dice.
<nckx>leungbk: Sorry, I was yelling at sport men. The only thing that looks weird is (guix build utils) in #:w-i-m. Remove just that. If I'm rite, it will magically work.
<kmicu>nckx: Belgium in U-21? 😅
*kmicu ’s listening to Much better than U-21
<leungbk>nckx: The issue is that (guix build utils) is needed for the (modify-phases) invocation (currently commented out) that I want to be able to do eventually.
<nckx>leungbk: Actually, I don't think so…
*nckx tests
<nckx>kmicu: Good guess. That was earlier.
<leungbk>Hmm, I recall that being a problem the other day, I'll retest.
<nckx>Royals in U-performing.
<nckx>leungbk: !!!
<nckx>I don't blame you for not trusting randos on the 'net but I think that's it.
*nckx is pretty trustworthy as randos go, tho'.
<nckx>…and with that, my sport men have lost yet again and I'm off to bed.
<leungbk>nckx: Ohh, I thought (guix build utils) would be needed in both #:modules and #:imported-modules. Thanks for your help!
<nckx>leungbk: Happy to.
<nckx>Good night all o/
<leungbk>Have a good one.