IRC channel logs


back to list of logs

<nckx>I just used the crate importer for the first time (yay!) and… inputs are keyword arguments now? Hm.
<vagrantc>openjdk doesn't seem to be available on i686-linux
<nckx>Do I want to know why this is.
<vagrantc>i'm trying to explore how to get diffoscope building on something other than x86_64-linux, since i've added all these native-inputs for tests
<vagrantc>some of them are expressly x86 only ... but a few just seem to be failing to build...
<vagrantc>(e.g. openjdk)
<vagrantc>is there a way to get guix weather to be more verbose about which things weren't available and which were?
<wdkrnls>Hi Guix, does anyone know a channel which might provide a package for rstudio?
<nckx>sneek: later tell wdkrnls: ?
<sneek>Got it.
<mbakke>apparently * fails to resolve on Arch:
<sneek>mbakke, you have 1 message.
<sneek>mbakke, raghavgururajan says: How are things with core-updates-->master merging for gnome 3.30. Just following up as it has been a while. :-)
<nckx>Hm. I blindly assumed that said channel only contains free software because it's rekado's, but I didn't check. If you see any non-free software in there please shut down your computer and accept my apologies.
<nckx>‘…on Arch’ wait that's not how what.
<nckx>mbakke: Arch (well, systemd) is doing a right thing. Who'dathunkit.
<nckx>(I'm guessing some poor GNU sysadmin is being pinged to death right about now, but have you talked to anyone about this? DNSSEC is basically saying doesn't exist.)
*nckx then notices the dates on that issue; groans.
<mbakke>nckx: the right thing would be to omit the `ad` flag in the response, as other resolvers do
*nckx takes back the stuff about the right thing.
<nckx>I jumped to dnsviz before reading (actually, loading) the whole thread.
<nckx>mbakke: I got confused by this:
<nckx>The → delegation.
<tune>how do I disable sleep on lid close while plugged in to power? I keep my laptop closed in a dock and it just went to sleep when I went to tty7
<vagrantc>i've used elogind-inhibit
<vagrantc>elogind-inhibit --what=handle-lid-switch --mode=block -- sleep 30d
<vagrantc>presuming 30 days is long enough for you
<vagrantc>or you can pass it whatever command you want; it'll keep it inhibited till the command finishes
<mbakke>tune: you can also add (handle-lid-switch 'inhibit) to the elogind configuration
<nckx>tune: Have you tried setting ‘handle-lid-switch’ to ‘ignore’ in ‘elogind-service’?
<mbakke>errh, ignore, indeed
*nckx was just typing ‘or perhaps inhibit’ 🙂
<mbakke>the elogind part of the manual could use an overhaul
<nckx>(If I seem exceptionally stupid tonight: I'm tunnelling through DNS at a whopping 1-2 KiB/s, with latency to match.)
<tune>ugh my mouse can't interact with icecat since the sleep issue
<tune>restarted it even. all keyboard stuff works. what a nightmare
<mbakke>nckx: yay for captive portals :-)
<nckx>No, true, still better than a modern 802.1x portal 🙂
<tune>I have never configured elogind. can this be done in my config.scm?
<vagrantc>yeah, it depends on if you want the default behavior to not suspend on lid close or just a one-off
<nckx>tune: Yes. vagrantc's trick is a hammer but a very good one.
<nckx>I've used it before, with something like ‘-- cat’ as the eternal guard.
<tune>killing xwayland fixed the mouse issue but also made all my windows disappear
<tune>I don't know what I'm doing with the service thing but I guess I'll look that up
<Blukunfando>Isn’t the whole point of IceCat to grab the mouse and leisurely kill it, revelling in its agony?
<tune>I see there is a handle-lid-switch-docked, but its default is already "ignore" and it clearly didn't help. what a horrible world we live in
<Guest96752>I am having issue on making a package installed by guix discoverable. Im am trying to use tuxboot to make a live clonezila system but it is complaining about libpng12. I already had installed it with guix, does someone knows what could be wrong? I am using guix manager on top of a gentoo system
<tune>does this look good? (elogind-service (handle-lid-switch "ignore"))
<tune>wait, no
<nckx>tune: So the issue really is ‘logind does what you ask as long as you're watching X, but as soon as you switch to a vt WHO ARE YOU BYE MUST SHUT DOWN’? Because that sounds credible nowadays.
<tune>(elogind-service (elogind-configuration (handle-lid-switch "ignore")))
<tune>hm I see
<nckx>tune: Manual says (elogind-service #:config (elogind-configuration (foo bar))
<tune>that looks odd... alright then
<nckx>It does, but it also still uses the procedure syntax (instead of -service-type) so it might just predate the current conventions. I use only the vanilla (elogind-service) so my system.scm is of no help.
<tune>something like this then? (elogind-service #:config (elogind-configuration (handle-lid-switch "ignore")))
<tune>or do I remove the parens from the change
<nckx>My guess, unfortunately, is that this is indeed one of those ‘seat’/‘session’ concepts that the systemd people have and that adding h-l-s ignore won't help.
<nckx>No, looks good, although the quotes might be too much. Manual isn't very clear but I think a symbol is expected, not a string.
<tune>yeah I had to remove the quotes, although now it fails on unbound variable
<tune>some sort of formatting issue now I guess, keeps blaming unrelated lines
<nckx>Or try 'ignore.
<nckx>Yes, we're just guessing at this point.
<tune>I don't actually know lisp, so this is all very painful
<nckx>It shouldn't (and normally isn't) this painful.
<tune>I just shoved it in the middle and clearly disrupted something
<nckx>tune: Yes, the (elogind-service) is outside of your (list …).
<tune>oh the tor section ended the list so I'm outside it I think
<tune>man, what a mess
<nckx>tune: Just move one ) from unix)))) to ignore)))
<nckx>I think.
<tune>I'm not gonna stack it all on the last item again, that doesn't make sense to my brain. gonna leave the last one on its own line even if it's less pretty
<tune>agh same error
<nckx>tune: Any chance you're willing to share your complete file? Or at least the full error?
<tune>wl-copy is broken so I can't do this as easily as I want
<tune>might paste them separately
<nckx>tune: ‘guix install wgetpaste’ might be of interest.
<nckx>OK, so it's not doing magic syntax stuff (probably for the best). Try 'ignore instead of ignore.
<nckx>Well that escawaifud quickly.
<tune>yeah sorry haha
<tune>I have a few pasting aliases and maybe that one wasn't the best. normally I could've cat'd a file to my clipboard and done it all on the debian paste
<tune>instead did a curl thing
<tune>new error: guix system: error: service 'file-system-/run/systemd' provided more than once
<nckx>Oh, I don't care, the sudden switch just made me laugh.
<nckx>OK, that's Guix/Guile's enigmatic way of hinting that you now probably have 2 elogind-configurations, and you need to remove the default one like you already remove gdm below.
<tune>does the order here matter? like should I do the removal before the services list?
<tune>this suddenly feels really complicated
<nckx>Heh. Yeah. So my problems are 1) I don't use any of this remove stuff to begin with, I just write my (services) from scratch and 2) I don't know off the top of my head whether the old-style (elogind-service) sets the expected ‘service-type’ field. So I'm the worst person to help you with this.
<nckx>That said, you can *try* replacing (eq? …) with (or (eq? (service-kind service) gdm-service-type) (eq? (service-kind service) elogind-service-type))
<nckx>and just being really lucky.
<nckx>That's always a plan.
<tune>this is the second time I've regretted using %desktop-services, but not sure I want to get caught up in changing that right now
<nckx>TBH you probably don't.
<nckx>I'm just a control freak.
<tune>so it's now actually updating, then I'll see what I end up with
<nckx>(Aside, re: <gonna leave the last ) on its own line even if it's less pretty>. One of the reasons that lispers [can afford to] write this way is because they *never* format their code by hand, so they read code by indentation, not by counting brackets. When you pasted your incorrect snippet, the first thing I did was paste it into emacs and auto-indented the whole thing. It made the error obvious. So write whatever you feel comfortable with, but consider
<nckx>configuring your editor to help you with the boring stuff. *Everyone* else does.)
<nckx>You're scheming in hard mode 🙂
*nckx AFK for a while.
<tune>whoa, a font issue from the other day is gone now. sweet.
<zacts>huh, I might go ahead and try out guix system on a second laptop.
<zacts>so, should I be installing most of my packages as a normal user or should it be systemwide? What is the common convention?
<tune>normal user
<tune>you can use a manifest if you still want a declarative list of packages for the user
<zacts>can I auto-generate a manifest?
<tune>I think so, but I haven't done it. I manage mine manually in a text editor
*zacts searches the doc on how to create manifests
<tune>you're welcome
<tune>so wl-clipboard 1.0 is broken on the current sway version. there's a 2.0.0_beta2 release now. how easily can the guix package be updated to this version?
<tune>nckx: I've got a semi-configured emacs and neovim but I don't totally get how my indentation works in either. I was editing in neovim earlier, but maybe I'll try to use emacs for my config.scm more
<tune>for now I've done C-x h TAB to change the indentation, but it's not immediately obvious to me what changed
<tune>at least emacs seems to do a better job of it automatically
<apteryx>hello Guix!
***tristanC_ is now known as tristanC
<chipb>ugh. I'm trying to use a guixsd vm, but I'm running into it suspending itself due to inactivity regardless of my being actively connected to it via SSH.
<chipb>any hints on how to turn that helpful feature off?
<chipb>libvirt's pm/suspend-to-{mem,disk} for the domain are set with enable=no, so by my reckoning, the VM shouldn't even see its hardware as capable of being suspended...
<chipb>(which I think may correspond to the "-global ICH9-LPC.disable_s3=1 -global ICH9-LPC.disable_s4=1" I see on my qemu's cmdline list...)
<chipb>hm. I *think* it's gdm doing it from its session. but that doesn't explain why the VM even thinks it can suspend to mem.
<apteryx>chipb: weird! perhaps you could try using a SlIM + ratpoison as the WM to make sure GDM/Gnome really is the culprit?
***slyfox_ is now known as slyfox
<Zaab1t>/gnu/store/q19l04vd2za80mk1845pz7r8cz29qk43-bash-minimal-4.4.23/bin/bash: warning: setlocale: LC_ALL: cannot change locale (en_US.utf8)
<Zaab1t>this warning is all over the place whenever I install anything
<efraim>do we have a package that uses the trivial-build-system and substitute*
<nckx>efraim: Random grep says freepats, certdata2pem, docbook-x[ms]l, epipe…
*nckx just did grep -C20 trivial-build-system gnu/packages/*scm | less, ‘/substitute’, scroll up; ugly but it gets the job done.
<efraim>nckx: thanks. Unfortunately it's still not liking part of my code, I'm sure I'm missing an input though
<apteryx>would anyone knows how to launch shepherd from a REPL?
<apteryx>((@ (shepherd) main) "help") works... but I'd expect just ((@ (shepherd) main)) to do the same as the shepherd script
<apteryx>yet it doesn't
<apteryx>what the shepherd script does is: (apply (@ (shepherd) main) (cdr (command-line)))
<apteryx>in the usual case (no arguments given), the cdr of (command-line) is '()
<apteryx>reading the doc of apply, it seems that this should be equivalent to calling shepherd directly with a list containing an empty list '(())
<apteryx>ah; just (main) *does* work. It's Geiser that was somehow masking the output
<quiliro>Saluton Giksujo!
<quiliro>Mi havas problemon kun loadkeys.
***jonsger1 is now known as jonsger
***fr33domlover1 is now known as fr33domlover
<olivuser>Hello everyone. Has anyone else experienced problems installing the game "freeorion"? Whenever I try to install it, my computer starts to hang - regardless of the cores-setting...
<olivuser>I did pull&upgrade before installing if that is of relevanve
<apteryx>hmm, let me try it
<zacts>why does `guix pull`, `guix package -u`, and then `guix install icecat` seem to build icecat from source?
<zacts>I'm hoping to not have to build anything from source on this computer
<Blukunfando>I’m building IceCat from source, too.
<janas>zacts: Maybe the package was updated recently and the build server doesn't have it built yet?
<janas>I just ran 'guix weather icecat', it says that there is no substitute available
<zacts>oh I see
<zacts>janas: I didn't know about that command
<zacts>thanks :-)
<quiliro>with that command line I get:
<quiliro> '' returned 500 ("Internal Server Error")
<quiliro>why is that?
<quiliro>if i make a guix upgrade, the current generation right after the completion is the one of the upgrade?
<zacts>janas: can I tell guix to only upgrade via binaries when I run `guix package -u`?
<zacts>oh cool
<zacts>I like the idea of a peer to peer system eventually
<janas>zacts: I know that there is a no-substitutes option, but I don't think there is a no-build one :(
<sherab>I am trying to install guix with qemu
<janas>I guess you could rerun 'guix weather icecat' every few hours until it reports a substitute
<zacts>maybe this could be a feature request to put in the issue tracker?
<sherab>gollowing the manuel i retrieved and decompressed the file and created the disk image that will hold the installed system
<sherab>my troubles begin with trying to boot the usb installation image in a vm
<zacts>I'll add it, if you think it might be cool
<zacts>janas: I'll do that for now
<sherab>here's the output from the command line qemu-system-x86_64 -m 1024 -smp 1 \ -net user -net nic,model=virtio -boot menu=on \ -drive file=guix-system-install-1.0.1.x86_64-linux.iso \ -drive file=guixsd.img
<sherab>qemu-system-x86_64: -smp 1: Could not open ' -net': No such file or directory
<quiliro>I get nothing displayed on ....Would someone test it please?
<quiliro>sherab: What happens if you put 'file=guix-system-install-1.0.1.x86_64-linux.iso' right after 'qemu-system-x86_64'?
<sherab>quiliro: without the -m 1024 -smp1 \ ??
<sherab>or do you mean like this: qemu-system-x86_64 -m 1024 -smp 1 \ file=guix-system-install-1.0.1.x86_64-linux.iso
<sherab>quiliro: here's the output:
<sherab>sherab@Sherab-MacBookAir:~/Downloads$ qemu-system-x86_64 -m 1024 -smp 1 \ file=guix-system-install-1.0.1.x86_64-linux.iso
<sherab>qemu-system-x86_64: -smp 1: Could not open ' file=guix-system-install-1.0.1.x86_64-linux.iso': No such file or directory
<sherab>but i do have a guix-system-install-1.0.1.x86_64-linux.iso in my folder
<quiliro> did not generate a virtual machine image
<sherab>quiliro: i think i did as i have the following all in the same folder I am working from : sherab@Sherab-MacBookAir:~/Downloads$ ls
<sherab> quixsd.img
<sherab>guix-system-install-1.0.1.x86_64-linux.iso reMarkable-installer-171.exe
<sherab>ihh maybe misspelling!!!
<quiliro>sherab: so you should use quixsd.img
<sherab>quiliro: yeah but i mv quixsd.img guixsd.img and it still not working
<quiliro>sherab: you are not using that are using guix-system-install-1.0.1.x86_64-linux.iso
<zacts>ok I added a feature request
<quiliro>sherab: added both!
<quiliro>sherab: why?
<sherab>quiliro: here's my latest : qemu-system-x86_64 -m 1024 -smp 1 \ -drive file=guix-system-install-1.0.1.x86_64-linux.iso \ -drive file=guixsd.img
<sherab>qemu-system-x86_64: -smp 1: Could not open ' -drive': No such file or directory
<quiliro>sherab: still adding both
<sherab>quiliro: i am sorry this might be too complicated for me - i am a newbee and i just followed the lines in the manual.
<sherab>the manual says qemu-system-x86_64 -m 1024 -smp 1 \
<sherab>then there are four lines which i just copied and pasted with alteration for the system
<quiliro>sherab: I am newbie too...but you specifying two files to boot from....
<sherab>quiliro: am i?
<sherab>as i said i am just following the manual
<quiliro>file=guix-system-install-1.0.1.x86_64-linux.iso and
<sherab>this is what the manual is saying to type to terminal. ?
<quiliro>i see...let me study
<sherab>yeah thanks for helping
<quiliro>sherab: please try all in one line and removing all \
<quiliro> she did you do this?: qemu-system-x86_64 -net nic,model=help
<quiliro>sherab: did you do this?: qemu-system-x86_64 -net nic,model=help
<quiliro>all in one line
<quiliro>qemu-system-x86_64 -net nic,model=help
<sherab>yeah still no luck so i am typing it all one word at a time maybe i made a misspel... i am working on it...
<quiliro>did you create your own image?
<sherab>i downloaded the and decompressed something then i think i created the image - following the instructions
<sherab>quiliro: give me few minutes i'll do everything from begining...
<quiliro>sherab: the best! do not change it literally
<quiliro>do it literally to avoid spelling errors
<sherab>yeah ok i'll start and let you know in few minutes where i am with it all.
<quiliro>i have not installed qemu, so i cannot test with you
<quiliro>i can just help with the logic
<sherab>ok i have it on a usb but it will not boot to macbook air
<quiliro>can you see the usb on your macos?
<quiliro>or you cannot boot before macos on boot menu?
<sherab>quiliro: i have linux mint on my macbookair and i want to try and install quix on it instead
<nckx>quiliro: issues.guix is simply overloaded. Nothing you can do.
<quiliro>nckx: thanks
*nckx o/ Guix.
<quiliro>sherab: you do not need a virtual machine image for that
<sherab>quiliro: i know
<quiliro>what you need a vm image is for testing or running one Operating System inside that what you want?
<nckx>quiliro: What's your loadkeys problem?
<nckx>(Too busy/tired for Esperanto, sorry.)
<quiliro>posted on the bug-guix mailing list
<sherab>quiliro: yes
<sherab>i am not using the usb at the moment
<quiliro>sherab: give a minute until i look for my bug report
<sherab>quiliro: cool
<quiliro>nckx: and the next one on the thread
<quiliro>nckx: bug#37479: International locale error
<nckx>quiliro: I saw those, but they don't mention loadkeys at all…
<quiliro>$ loadkeys
<quiliro>Couldn't get a file descriptor referring to the console
<quiliro>forgot to post that too
<quiliro>will do it now
<nckx>quiliro: Are you trying to run it from X? loadkeys only works on the linux VT.
<nckx>I use it regularly, hence my interest. I don't use Gnome/big desktops, so I'm less helpful there.
<quiliro>nckx: you are right....but the problem is on tty1 also....cannot type ^g on one character
<quiliro>as one character
<nckx>quiliro: Here's what I used to get ‘eo’ to work (which isn't part of the *completely* arbitrary %default-locale-definitions in (gnu system locale)). However, es_ES is already in that list, so I don't know why it still fails for you. I have no ‘*locale*’ packages installed anywhere, user nor system profile.
<sherab>quiliro: i think we're making progress.
<quiliro>nckx: i would like EO!!!!
<nckx>I can only promise that the above snippet worked for me. Nothing more was needed 🙂
<quiliro>sherab: nice...what happened?
<nckx>Hm, ^g (ĝ) doesn't work for me either on the console, only under X. I use a compose key, though.
<quiliro>nckx: dvorak-es keyboard is my only diff probably
<sherab>well i get a new terminal that looks like the VM terminal
<sherab>but something else comes in my linux terminal warning: TCG doesn't support requested feature: CPUID.01H:ECX.vmx [bit 5]
<quiliro>nckx: how do you see pictures on the console?
<quiliro>sherab: error and not working or error and working?
<sherab>quiliro: hang on tight something promising
<sherab>is happening
<quiliro>sherab: nice
<quiliro>25 degrees celsius inside 24 outside
<sherab>quiliro: i might not come back to you now but i think we're good. thank you v much !!
<sherab>[becasue i need to go]
<quiliro>sherab: :-)
<efraim>I wish it was that cool. We're still at 27 indoors at 1930
<quiliro>it is very hot now here
<quiliro>usually 20 average
<quiliro>27 would be very hot for the decade and 5 very cold for the decade
<quiliro>I am at/in Quito, Ecuador.
<quiliro>nckx: will your config be good as locale for grub and X1 as well?
<quiliro>nckx: how do you change locale? between en_GB" "eo" and "zh_CN?
<quiliro>nckx: could there be a "eo_EC"?
<quiliro>nckx: how can i see what "eo" has?
<quiliro>efraim: it never snows or gets above 25 degrees Celsius here
<quiliro>in all my life
<nckx>quiliro: 1a) almost never, I do run X you know 1b) with fbida 2a) I don't really 2b) with LANG=foo or LC_ALL=foo <command> 3) no 4) It's just eo, it doesn't have a country code by design (it used to be ‘eo_EO’ but that has been deprecated in favour of just ‘eo’).
<nckx>quiliro: 1c) The ‘fbida’ provides the ‘fbi’ command, so that can be confusing 2c) I only use zh_CN daily but tried eo for a week for fun.
<nckx>s/provides/package provides/
<efraim>quiliro: that sounds nice. here in Israel we have like 5 months of 30+ days
<nckx>quiliro: Good to know that the weather in Belgium today was as nice as in Ecuador 🙂
<quiliro>we have nice weather but we have a tyrant government
<quiliro>which attacks smart people like julian assange and ola bini
<nckx>We have shit weather and a shit government, so hey.
<quiliro>perhaps because we are easy going
<quiliro>as shitty as our government?
<quiliro>i have never gone to ski
<quiliro>i have seen the snow but on the mountains
<quiliro>the government attacks smart people directed by the usa
<nckx>Just incompetent. It's become something of a joke. We had elections in May(!). We still don't have a government.
<nckx>Our record is 541 days between election & new government. I think that was in 2010.
<quiliro>that is great! no goverment....
<nckx>Yeah. No.
<quiliro>like cataluña
<quiliro>in the 30's
<nckx>It's not actually great in real life 🙂 Being a wealthy EU member means we can afford to be without an acting government for a ridiculously long time without people setting fire to things or the US coming to Help, but it does affect things badly in the long run.
<nckx>It is not an anarchist paradise.
*nckx sad anarchist.
<quiliro>nckx: do you have nice fonts on the terminal?
<quiliro>it could be!
<quiliro>under the circumstances
<nckx>nckx: That's enough off-topic discussion, I think 🙂
<nckx>nckx: OK.
<nckx>quiliro: Yes, in that I run a custom kernel and have configured said kernel to use the 6x11 font instead of the standard VGA.
<quiliro>could be an option :-D
<nckx>Whether that's ‘nice’ is debatable but it makes more of a stack trace fit on the screen…
<quiliro>is there some picure?
<quiliro>nice fonts probably are easier on the eyes
<nckx>quiliro: Are you advocating the overthrow of our glorious maintainer?
<quiliro>i do not think our glorious maintainer is a master...unless you are talking about the git server :-D
<quiliro>it does have master
<quiliro>nckx: but master in that case, allows branches with equal rights
<quiliro>that would be an interesting question on help-guix: Is Guix an anarchy? If so, which type, individualist or collectivist?
<nckx>Sorry, it was the 6x10 font (CONFIG_FONT_6x10=y). Obviously not worth compiling a custom kernel just for that, but it's a nice touch.
<nckx>I'm sure it's not your definition of ‘nice’ but it's certainly ‘nice to have’ on a non-high-DPI screen.
<emacsomancer>on a foreign distro, trying to update guix I'm getting a collision error for guile-gcrypt
<quiliro>nckx: it does look nice
<quiliro>nckx: i cannot see that character
<nckx>Oh, :- )
<quiliro>just abox with 01F 642
<quiliro> is just a :-) ?
<nckx>As you can see (‘What█s new?’), the encoding isn't completely set up the way it should be, but I only use VTs for rescuing my X.
<quiliro>nckx: how do you deal with X? DE or just minimal? and how?
<nckx>Just realised that the black-on-white scheme is custom too (using ‘vt.color=0x70 vt.default_…’).
<nckx>quiliro: Just i3+i3bar+dmenu (with custom frecency wrapper script).
<quiliro>nckx: what do you think about exwm?
<nckx>My previous laptop needed every spare CPU cycle I could squeeze from it & now I'm used to it.
<nckx>quiliro: *Still* haven't tried it :-/
<quiliro>looking forward to your take on use emacs,richt?
<nckx>I do.
<nckx>I told Pierre (not that they care or remember) that I'd try it in February; I should keep my promise before next FOSDEM…
<quiliro>do you use Magit?
<nckx>quiliro: Lightly. I use it when emacs is invoked as $EDITOR (so for rebase -i &c.), but I run the git commands themselves from the command line.
<nckx>Purely out of habit.
<nckx>I'd probably be more efficient with more emacs in my life.
<nckx>emacsomancer: Could you paste the full error?
<nckx>(To or similar service.)
<quiliro>nckx: but in eshel or in shell you are still in emacs!
*nckx does not use eshell.
<quiliro>nckx: why?
<quiliro>nckx: i think i will set up a Guix academy
<nckx>¯\_(ツ)_/¯ If I do try out exwm and like it I probably will, but until then my wm is my wm and my terminal is my terminal.
<nckx>Plus my muscle memory is prone to typing ‘less’ a lot.
<quiliro>i have mentored minall for learning Guix...and i think i can replicate it with others
<nckx>quiliro: That would be great!
<quiliro>nckx: any suggestions? where to find students, what the curricula would be, methology, etc
<nckx>emacsomancer: Er, wow. That's gorier than expected. Not really a foreign Guix user/fan so can't help you much I'm afraid. Which command(-equivalent) would that be? ‘guix pull’? Does running it from the command line produce the same error?
<emacsomancer>nckx: that's the update command. guix pull is fine. I'll try it from the command line.
<nckx>emacsomancer: But updating what? You're not trying to ‘guix package -i guix’ or ‘guix package -u guix’, I hope? Because that is a big no-no.
<nckx>(Just checkin'.)
<emacsomancer>nckx: just `guix package -u`
<quiliro>emacsomancer: as regular user?
<emacsomancer>I mean, via guix-emacs
<nckx>Oh, OK. Your ‘trying to update guix’ had me worried for a sec.
<nckx>You don't really have any other choice with that nick.
<emacsomancer>that's true
<quiliro>emacsomancer: does 'guix upgrade' on eshell or shell work?
<nckx>quiliro: Not really. My own personal teaching style is a bit more theoretical/let's-understand-what-Scheme-is than is probably ideal at such events, so I tend to prefer tutoring more experienced users (just new to Scheme/Guix) one-on-one.
<quiliro>nckx: but command line is easy....that is just the beggining....the hard part is guix....
<quiliro>maybe i write that because i do not know how yet!
<quiliro>i still have trouble with packaging
<quiliro>perhaps i should read sicp
<quiliro>i want to use emacs with ledger-mode to make invoices and register them on ledger-cli ... is there already some documented way? (with automatic numbering and such)
<quiliro>i am very versatile with ledger-mode and magit already
<quiliro>so i have a journal file which is automatically recognized by ledger-mode because it ends in .ledger
<emacsomancer>commandline doesn't seem to try to update guix itself (and so doesn't fail)
<sherab>does it make sense that a new installation with qemu takes more than an hour to download many many many files??
<nckx>quiliro: I'm not sure where the ‘command line’ came from. I don't understand the distinction.
<nckx>Sure, ‘teaching’ people to type ‘guix install’ is easy. It's also not really teaching them anything, I think we agree on that.
<nckx>emacsomancer: If ‘guix pull’ does work, I'm starting to think this is an emacs-guix bug or at least gotcha. ‘guix package -u’ should *never* update guix itself.
<nckx>This all coated in a veneer of ‘I'm not really familiar with your use-case’, of course.
<nckx>But the *never* is a hard one.
<emacsomancer>nckx: `guix` is listed under user-installed packages. should I remove it?
<quiliro>sherab: hould depending on what you changed
<pkill9>yea, guix shouldn't be installed to a profile
<nckx>sherab: As long as it's downloading substitutes and not downloading/building sources, I guess so. It depends on your connection to
<pkill9>it gets "installed" by the `guix pull` command into ~/.config/guix/current
<quiliro>emacsomancer: perhapl you asked emacs-guix to install guix
<nckx>emacsomancer: If ‘guix package -I ^guix$’ returns anything other than nothing: hell yes.
<emacsomancer>I have some vague memory (maybe?) of installing guix directly to work around some issue with LOCALE - but I may be misremembering
<quiliro>emacsomancer: guix remove guix
<emacsomancer>I just removed it via emacs-guix
<quiliro>emacsomancer: please tell me your work process when you manage to get expertise on emacs-guix... i am a fan
<emacsomancer>quiliro: what do you mean? just how i use emacs-guix?
<quiliro>emacsomancer: yes...I use it but would like to know how others use it.
<quiliro>i usually enter via M-x guix
<emacsomancer>I tend to do `M-x guix-pull`, `M-x guix-installed-user-packages` and then `^` to mark packages needing updating; or `guix-all-packages` to get a list of packages if i want to install something new
<quiliro>then i p and a
<quiliro>emacsomancer: thanks
*nckx .oO Is ‘Cuirass’ just a punny pun on ‘CI’ and ‘Guix’? Or am I missing something obvious?
<zacts>I notice that `guix weather libreoffice` fails too
<c0c0>i'm trying to install guix. i booted from the installation usb stick, the grub menu with "GNU Guix installation..." shows up, and in the next step, i see some output such as "setting up setuid programs in '/run/setuid-programs'...". after that screen i arrive at a blank screen that stays forever. not sure what the problem is. before the blank screen, i see two errors, one of which is "Driver 'pcspkr'
<c0c0>is already registered, aborting...". i don't know what that means. any hints?
<zacts>I had an issue with the installer as well. You might want to try manually installing.
<zacts>c0c0: ^
<quiliro>c0c0: probably video card problem
<c0c0>quiliro: any idea how i could fix it? i'm running another linux-libre distribution on that machine, so i don't think it's a compatibility issue
<atw>nckx: if there's a pun in there it's lost on me
<nckx>c0c0: The pcspkr line is irrelevant & harmless.
<nckx>atw: Wel, ‘pun’… I can see ‘CI’ and Guix's ‘ui’ in there if I squint hard enough, 's all I meant. That and it is le French because honhon. That's all.
<atw>hehe had to wonder if I was missing some subtle multilingual pun
<c0c0>could "error in finalization thread: bad file descriptor" give any indication for why i end up with a blank screen?
<nckx>c0c0: Nope, that's another thing that shows up on everyone's screen while booting. Also harmless. We have too many of such lines, that is true.
<runner0101> │ │ contrun[m]
<c0c0>alright, does any of these give an indication for what might be going wrong?
<c0c0>that's all information i see
<quiliro>c0c0: you could try passing parameters to grub when you boot ....something like "vesa" I remember from old days
<quiliro>c0c0: maybe the image is not well copied to the USB even
<quiliro>c0c0: check the usb on another machine
<quiliro>c0c0: a machine which has different hardware, and a machine that has less dirt on the usb ports and a better state of hard disk and mainboard (hardware state and driver problems often cause problems booting)
<sherab>almost 3 hours of downloading and building is that normal??
<janneke>sherab: it depends on what you're doing.
<sherab>i rewally don't know what i am doing...;-)
<sherab>i donwloaded and decompressed a file
<sherab>then tried to instle in in qemu vm
<sherab>this was around 17:45 now its 21:32
*vagrantc waves :)
***Digitteknohippie is now known as Digit
<c0c0>quiliro: thanks for your hints. i tried a different usb stick, getting a blank screen again. tried on a different machine, and it works there
<c0c0>don't know if there's anything i can do to install it on that machine where i get a blank screen?
<vagrantc>so, i made some adjustments to the native-inputs of diffoscope that had some conditionally based on architecture ... but looking at: it didn't even attempt to build on aarch64 or armhf
<civodul>vagrantc: "guix show diffoscope" now says it's only supported on x86
<sneek>Welcome back civodul, you have 1 message.
<sneek>civodul, nckx says: Russians are back.
<civodul>heh :-)
<civodul>thanks nckx
<civodul>vagrantc: that means something among its dependencies is limited to those systems
<civodul>via 'supported-systems'
<civodul>that's why it's no longer being built on ARM
<vagrantc>civodul: huh... i successfully built it locally on armhf-linux, though ...
<civodul>yes, but the 'supported-systems' fields in the dependency graph say it's x86-only
<vagrantc>this is complicated :)
<civodul>'supported-systems' is mostly a hint for "guix show" and for Cuirass
<civodul>it can be wrong :-)
<civodul>so maybe that's what happens here: one package is wrongfully restricted to x86
<civodul>you can find out which one by commenting dependencies one by one and running "guix show diffoscope"
<vagrantc>on my armhf-linux system, i'm seeing systems: x86_64-linux i686-linux armhf-linux aarch64-linux mips64el-linux
<vagrantc>is that different from supported-systems?
<vagrantc>earlier, it did have some native-inputs that were only x86