IRC channel logs


back to list of logs

<oriansj>just a stupid question but shouldn't login managers and the Window managers/Desktop environments they display be by definition global? For what would happen if 2 people installed say i3 How would any of the users know which version was which
<oriansj>(especially if one of the users built a custom version)
<joshuaBPMan>oriansj: If I understand you properly...good question.
<oriansj>joshuaBPMan: there are also security implications of users installing particular pieces of software (say screen savers, login interfaces and sudo) but I don't believe guix even has the concept of security impacting packages
<oriansj>one could say we should perhaps look closer at critical views of our work and probably take some time to talk to the people most critical of guix to find the areas where we really should be improving
<oriansj>notice for example, why don't we steal an idea from apt (but in reverse) where one could append --install-recommends which would add additional packages that are generally needed by the named package (like how you need to install nss-certs when you install icecat or bash-completions when you install bash)
<oriansj>why don't we have a guix unpack for installing pack files created by guix pack ?
<joshuaBPMan>oriansj: I honestly don't understand what your original question was.
<oriansj>joshuaBPMan: it was about the issues and security implications of installing login managers/Window managers by non-administrative users
<happy_gnu[m]>ok so I installed guile-wm as user
<happy_gnu[m]>did not get recognized
<happy_gnu[m]>then as root
<happy_gnu[m]>(i did a guix pull to make sure everything is ok)
<happy_gnu[m]>then I add
<happy_gnu[m]>guile-wm in (packages (cons* ratpoison i3-wm i3status dmenu ;window managers
<happy_gnu[m]>nss-certs ;for HTTPS access
<happy_gnu[m]>and I get this error
<happy_gnu[m]>guix system: erro: failed to loat '/etc/config.scm'" ... unbound variable: guile-wm
<happy_gnu[m]>what to do now?
<thomassgn>happy_gnu[m]: check what the location of guile-wm is. you ca do this by looking at location in the search results for it with `guix package -s guile-wm'
<thomassgn>the location (without .scm) should be added to your '(use-package-modules ...)' at the top of your configuration
<thomassgn>happy_gnu[m]: also, you might have to add it in the wm services bit, but look it up, or ask me again later. (to lazy now, omw to bed)
<happy_gnu[m]>thomassgn: OK thanks :)
<joshuaBPMan>crazy thought, how hard would it be to package an "emacs distribution" into a guix package? You could install additional helper programs for Emacs. Syntax checking, searching programs, a service to start it, etc.
<happy_gnu[m]>joshuaBPMan: like all Elpa and Melpa merged in one distribution?
<happy_gnu[m]>So I was able to install guile-wm
<happy_gnu[m]>and now it is on slim
<happy_gnu[m]>but when I enter password and it enters guile-wm it shows and error and logs out
<happy_gnu[m]>I will try to fix it in the morning though
<happy_gnu[m]>maybe I need a config file or something
<happy_gnu[m]>anyway GuixSD is awesome :)
<happy_gnu[m]>tomorrow I will start making packages
<happy_gnu[m]>I hope is not too hard :/
<sadiq[m]>And at last installation is complete
<sadiq[m]>How long does it take to boot the system?
<happy_gnu[m]>sadiq[m]: depends what is your system specs? what Desktop Enviroment are you using?
<happy_gnu[m]>for me on a VM with i3wm takes like 10 seconds
<sadiq[m]>Gnome. But it is now stuck for about 5 minutes
<happy_gnu[m]>sadiq[m]: thats weird did you see Grub?
<sadiq[m]>I see "Entering a new prompt" text. But I don't see any prompt
<sadiq[m]>I mean no any $PS
<sadiq[m]>The boot began. Seems like some error happened.
<happy_gnu[m]>so this is after Grub? or before?
<sadiq[m]>The text ERROR: LUKS partition not found is shown
<happy_gnu[m]>what did you use a USB?
<happy_gnu[m]>are you installing on VM or computer?
<sadiq[m]>Bare metal. Now I'm in guile prompt
<sadiq[m]>Bare metal. Now I'm in guile prompt
<happy_gnu[m]>well in the morning I cancelled my installation before time sadiq[m]
<happy_gnu[m]>what I did was boot again for usb image
<happy_gnu[m]>mount partitions (do not make them again, just mount them they should have everything)
<happy_gnu[m]>also herd start cow-store /mnt
<happy_gnu[m]>and guix system init /mnt/etc/config.scm /mnt
<happy_gnu[m]>everything that was done was already there
<happy_gnu[m]>It finished in like 5 minutes the last things
<sadiq[m]>That seems to start over
<sadiq[m]>Hm.. So should I remove (dependencies mapped-devices) defun call if cryptsetup isn't used? (In config.scm)
<rekado>Hi Guix!
<civodul>Hello Guix!
<pksadiq>why is installation always failing at certain package (adwaita-icon-theme)? seems like the file is downloaded partially
<civodul>pksadiq: could you give the URL that's causing problems?
<civodul>people reported something similar in the past, but i thought this was fixed
<ArneBab>when trying guix pull --url=...v0.13.0.tar.gz, I get a hash mismatch: Downloading (54.1MiB installed)... → python-2.7.12 14.7MiB 7.4MiB/s 00:02 [####################] 100.0% → hash mismatch in downloaded path `/gnu/store/q6rbp7s542jkhrhz04hsp2i60gw0h4as-python-2.7.12': expected aea4335a5e6d65a36ed74561b15f8242773c49110be30d8ab7b43590f0568e60
<ArneBab>, got 49b3fc5b436309b4d097ed3c84946d5cab742db6159d76f5ad7a1d7896a2760f
<pksadiq>civodul: I resumed with fallback. Now working fine. Sorry, I couldn't capture the url
<janneke>sorry for the cuirass-bug-spam, i'm working on a documentation patch
<civodul>ArneBab: you're upgrading from what version?
<civodul>i think you'll have to --fallback
<ArneBab>civodul: from more than a year ago. --fallback wasn’t recognized as option
<ArneBab>this worked: guix pull --url=
<civodul>ok but it's not a great idea :-)
<ng0>happy_gnu[m]: I came to a conclusion for a good name for my set of scripts which will generate a distributable New Moon / Pale Moon variant at some point: meissa.
<ng0>named after
<efraim>Can i put in my os-config (dependencirs /dev/sda2 /dev/sdb2 /dev/sdc2)
<ng0>is "libindicator" present in some branch I don't know about? It's not in master.
<ng0> <- I'd put it into (gnu packages indicator). Any other namespace ideas?
<ng0>"A set of symbols and convience functions that all indicators would like to use. Not of real use outside of the Ayatana indicators project."
<ng0>or "libindicator.scm"
<ng0>^ picking this one for now, it's easier to rebase the patches and rename it if someone doesn't agree.
<efraim>I'm having a hard time with this error, from the installer. (string-append "-root=" (uuid…
<efraim>Would that be passed to grub?
<efraim>I'm not sure where to debug it
<efraim>Swap-devices is a list or cons?
<ng0>I have this problem with atril. I provided "openjpeg" but in the last phase of install it can not find -lopenjpeg. Anything obvious I could try?
<ng0>looks like atm
<efraim>Try the older openjpeg or manually add -lopenjepg
<efraim>Also only fix introspection once :)
<ng0>oops :D
<ng0>yeah it's not final
<thomasd>ng0: also, you could try something like (string-append (assoc-ref %build-inputs "openjpeg") "/lib"), or "/include". In my experience, most configure-scripts seem to expect the lib or include directory in such an option (./configure --help might give a hint, or config.log)
<ng0>at least the error changed ;)
<ng0>ld: cannot find -lgs
<ng0>we stumble forwards, one failure at a time
<nee``>Hello, I have the same problem as nextos had on 2017-09-15 here when running reconfigure: I also updated to the new bootloader syntax in but it still fails with the same trace.
<ng0>lib gs... ghostscript right?
<thomasd>ng0: gs = ghostscript? it's not listed in the inputs
<nee``>Is there some way to set the default width for guile stack traces higher for the whole system, so it prints the full variables? I mean this setting:
<ng0>success \\o/ (minus the test suite)
<efraim>nee``: I'm also getting errors related to read-boot-parameters-file
<nee``>efraim: have you updated to the new syntax for bootloader?
<pksadiq>why does guixSD install 3.22 version of many packages while the list in guix websites lists 3.24 versions of those?
<pksadiq>gnome packages, I mean
<nee``>Also, the error says "/var/guix/profiles/system-1...". Is it reading all profiles there to generate the grub entries? Could it be failing to read an older profiles, because of the new syntax?
<efraim>nee``: it shouldn't be an issue, this is a new install
<nee``>pksadiq: I think the website gets regenerated with commits to master, maybe you need to `guix pull` and `guix system reconfigure /etc/config.scm` to update the gnome packages.
<nee``>pksadiq: `guix package -s ^gnome$` lists gnome 3.24.3 for me and I just did a pull a few minutes ago.
<thomasd>yay > 6000 packages ;-)
<nee``>`guix system list-generations` also fails with a similiar trace now:
<efraim>nee``: I also tried to build the config examples
<thomasd>nee : which version are you on?
<nee``>guix --version says 851cb6dd3996ef72ad7c23bd588f8ffb64a2c21e
<thomasd>I see that a bug in that code was fixed (or introduced? :) ) last week
<ng0>alright. 1 thing left on the road to an almost complete Mate Desktop: fixing caja extensions integration.
<thomasd>since you are on the latest version, might be that the fix introduced another issue
<ng0>fwiw, I have a system where since a recent'ish commit the luks disk is no longer found, backtrace on boot. But it's no loss to boot into older generation and set it up again from 0.
<ng0>the other systems, with not so many disk like this one, still work
<thomasd>nee: can you paste the content of your /var/guix/profiles/system-{xyz}-link/parameters files somewhere?
<thomasd>(or specifically, the “root-device” entry)
<nee``>here it is: note that I had the root device wrongly set to sda1 on my inital install, which I made by running `guix system init config.scm /` on a debian server
<pksadiq``>will --fallback build all packages from source?
<thomasd>nee: I think the problem is that the “store device” in your parameter files is “(store (device #f) (...))”, which is not handled by the code trying to parse that parameters file
<pksadiq``>ACTION sighs...
<thomasd>fwiw, on my machine it's “(device "my-root")”, i.e. the label of my root partition. Don't know what (device #f) means.
<pksadiq``>this is too much. I shall try this again some other day. Every single package is being built from source, and not just the failing package :(
<nee``>thomasd: Maybe it is related to the deprecated (device) field in the bootloader settings that was replaced by (target). I used device until today.
<thomasd>nee: hmm, that's possible. I don't have time to fully debug it now.
<thomasd>as a workaround, you could remove the “system-xyz-link” symlinks before running reconfigure or similar commands
<thomasd>maybe keep them somewhere else, so you can restore them afterwards
<thomasd>(and don't run “guix gc” in the mean time :-) )
<nee``>thomasd: okay thanks for the help
<nee``>Also here is the first config I used, it has sda1 corrected to sda3, but appart from that it's what should have created the system-1 from guix master on 2017-09-05:
<efraim>Deleting the offending system generations seems to work for me
<efraim>Thanks thomasd and nee``
<thomasd>efraim: good! I still think it could still be a bug, though.
<thomasd>from my limited understanding of gnu/system.scm,
<efraim>looks like the code for 'old-versions' could be improved
<thomasd>(store (device #f)) could be legal, meaning that the code parsing that file should be able to handle it (which it is not). What is the (store (device ...)) in the newly configured generations?
<nee``>thomasd: for me the latest one is still #f (generation 93 in my paste is the newest). Maybe because I never did a reconfigure with the (target) field in the old bootloader syntax and always sticked to the deprecated (device) field.
<thomasd>nee: I meant after removing the offending generations and re-running “guix system reconfigure”, as efraim seems to have done.
<thomasd>I suppose you should submit a bug if the issue persists.
<sadiq[m]>I have 200 MiB bandwidth more for today. Would it be a good idea to install the bare minimal and add the desktop + services later?
<thomasd>sadiq: I've done it like that twice. First install the minimal system, then upgrade from there.
<jonsger>sadiq[m]: if you are spare of bandwith, I'm not sure if guix is a good idea at all :P
<efraim>apteryx[m]: with your snippet for using bayfront and berlin as substitute urls, what did you have to add to teh top? (guix store)?
<efraim>Seems to be the case
<lfam>I can't handle it at the moment, but CVE-2017-9798 (Apache httpd memory leak) needs to be patched:
<sadiq[m]>jonsger: Have a 2GiB/day limited. Wasted the rest trying to install guixSD with desktop
<civodul>ACTION looks at the httpd CVE-2017-9798
<sadiq[m]>What is the approx size of master.tar.gz guix snapshot?
<apteryx[m]>efraim: yes!
<efraim>sadiq[m]: 50 MB?
<efraim>Not including the actual bootstrap binaries
<efraim>I still have yet to reboot, but it looks like I have GuixSD installed on a btrfs raid 1
<civodul>efraim: "guix lint -c cve binutils" is... interesting :-)
<efraim>Yeah, that was an incredibly long list
<efraim>I think the ones I actually looked at were found with AFL
<civodul>"guix lint -c cve" shows a lot of 2017 entries :-/
<efraim>there's a grub config option for 'search --set=root --label my-root'
<efraim>Interesting, but doesn't answer my question of installing to multiple drives
<civodul>efraim: installing to multiple drives?
<efraim>Btrfs handles the raid between 3 disks, but if I install grub to sda and that disk dies I still have my data but no grub
<efraim>Standard solution is to run 'sudo grub-install /dev/sd[abc]' when necessary
<sadiq[m]>Hm... Installation went well. Now how can I do a system wide 'guix pull' (that would effect every user, and not just the one that invoked)
<civodul>you mean run 'grub-install' 3 times, or pass it three targets?
<civodul>sadiq[m]: 'guix pull' is always per-user, never system-wide
<efraim>Apparently it wont take three taegets
<efraim>Gentoo handbook says it only has to be run for other devices after a grub update or other large changes, like dual booting
<sadiq[m]>civodul: what out deleting a user. And creating again after guix pull as root?
<sadiq[m]>Sorry for the typos, my phone's autocorrect is too much aggressive
<efraim>ln -s everyone's $HOME/.config/guix/latest to one person's?
<civodul>sadiq[m]: sorry, i don't understand what you have in mind
<sadiq[m]>With a copy of*
<sadiq[m]>civodul: create a new user with current user's profile
<civodul>good question, there's no simple way to do that currently
<civodul>but it'd be useful and not too hard to implement
<sadiq[m]>Hm.. Well, then I'll redownload all those again. Not a big deal. :)
<happy_gnu[m]>sadiq[m]: are you still here?
<happy_gnu[m]>I was sleeping
<happy_gnu[m]>Did you remove the part of encryption from the config.scm?
<happy_gnu[m]>ng0: that's pretty cool!
<happy_gnu[m]>(λ Orionis)
<apteryx[m]>efraim: I think our bootloader configuration should accept a list of devices; it would then install the bootloader on each of those.
<efraim>It would be convenient, even if it doesn't come up often
<efraim>Or a list of bootloaders might be better
<efraim>For flashing uboot and installing grub for example
<sadiq[m]>happy_gnu: reinstalled and fixed the issue :)
<happy_gnu[m]>sadiq[m]: oh great! :) is it working now?
<sneek>davexunit, you have 1 message.
<sneek>davexunit, alezost says: hi, I sent a tiny patch for Haunt to your "davet at" email a week ago, but now I have doubts if you still use this email :-) Should I re-send it to, or did you receive that message? thanks
<davexunit>sneek: later tell alezost I got the email but I haven't been able to look at the issue yet, though I can say that where compiled guile modules are supposed to go has never been clear and I copied from other devs.
<davexunit>that thinkpad x62 article above is very interesting.
<davexunit>it includes a script for improving battery life on laptops that I might try out on GuixSD
<efraim>I was thinking of combining some of the lines with {foo,bar}
<sadiq[m]>guix package --list-generations says error: profile '/var/guix/profiles/per-user/sadiq/guix-profile' does not exist
<sadiq[m]>happy_gnu: yup. Going good
<happy_gnu[m]>sadiq[m]: what does your config.scm says
<happy_gnu[m]>Ok I want to start packaging
<happy_gnu[m]>where do I start reading to learn how to do it?
<bavier`>happy_gnu[m]: there is a section in the manual
<happy_gnu[m]>I looked for Syncthing seems like it is not in
<happy_gnu[m]>It says it requires Git and Go
<happy_gnu[m]>Both are on Guix
<bavier`>happy_gnu[m]: check the email archives. istr there was maybe a start at a package
<efraim>lfam has a mostly working go-build-system and syncthing package
<happy_gnu[m]>mailing list just mention synthing on ""
<happy_gnu[m]>sorry I copied the link
<happy_gnu[m]>I meant "bug#27261: guix package -u should warn about non-existent packages "
<happy_gnu[m]>ok so if lfam already has it maybe I should look for something else then
<rekado>I’m having problems with my wee i686 server.
<rekado>Our substitute availability for i686 is pretty poor :(
<reed_>Hi all. When I install guix and run 'guix system init /mnt/etc/config.scm /mnt' it says "installing for i386-pc platform" but I downloaded the x86_64 image. Is this the expected output?
<efraim>yes, grub calls x86 and x86_64 machines with bios "i386-pc" and
<reed_>Thank you!
<happy_gnu[m]>Ok so I think I3blocks is good to start
<happy_gnu[m]>the only dependency it has that is not in guix is playerctl
<happy_gnu[m]>so I should start by packaging playerctl right?
<happy_gnu[m]>it only requires Glib
<happy_gnu[m]>there are no mentions of playerctl or i3blocks on mailing list
<stefanX2ovic>Hello, I have a few questions. Are you familiar with campaign? And can the GNU Guix support this campaign as an organisation?
<happy_gnu[m]>stefanX2ovic: I am pretty sure Guix Does
<happy_gnu[m]>After all is a GNU project
<happy_gnu[m]>And that campaign is by the Free Software Foundation Europe
<stefanX2ovic>happy_gnu[m]: It can be good for the project if we can get GuixSD logo there.
<happy_gnu[m]>stefanX2ovic: But isn't the GNU project under the Free Software Foundation which is a USA organization
<happy_gnu[m]>If anything the Free Software Foundation should support but FSFE is already there..
<stefanX2ovic>happy_gnu[m]: sure why not, I agree...
<stefanX2ovic>happy_gnu[m]: and I think it is a world wide campaign.
<stefanX2ovic>happy_gnu[m]: It can be good for publicity if we can somehow get GuixSD logo there.
<happy_gnu[m]>stefanX2ovic: is a EU campaign
<happy_gnu[m]>For EU
<happy_gnu[m]>Link literally ends in .eu
<stefanX2ovic>happy_gnu[m]: Correct, but people from all over the wold can sing this open letter.
<stefanX2ovic>happy_gnu[m]: Isn't GNOME project part of GNU and it has its logo on the site. Is there a reason why GuixSD logo can not be there as well?
<stefanX2ovic>Let's continue our discussion of campaign at .
<bavier`>stefanX2ovic: it's the GNOME Foundation, rather than the project itself, that is pledging its support.
<bavier`>stefanX2ovic: Guix has a "Guix Europe" foundation that would be in a better place to state such support
<stefanX2ovic>Isn't this "Guix Europe" foundation channel as well?
<bavier`>stefanX2ovic: yes
<stefanX2ovic>bavier`: Thank you. This is question for its members then.
<dustyweb>ispell-init-process: Error: The encoding "nil" is not known. This could also mean that the file "/gnu/store/af5mr9wz9d7rcsa0bwi82r9a98dmfx8p-aspell-" could not be opened for reading or does not exist.
<dustyweb>that's a new one
<dustyweb>when I tried ispell-buffer
<dustyweb>anyone else hit this?
<happy_gnu[m]>dustyweb: I had a similar bug moments ago
<happy_gnu[m]>nothing worked
<happy_gnu[m]>I rebooted
<happy_gnu[m]>and everything went back to normal
<happy_gnu[m]>I know is not a regular GNU/Linux approach to reboot, but it was because I mistakenly closed the virtual machine window
<ng0>well, there's openSuSe on the list of supporting organizations… someone representing Guix Europe would have to contact, that's all.
***LnL7 is now known as LnL
<stefanX2ovic>ng0: Great. Can we make this happen? Are the Guix Europe foundation members here?
<ng0>it's usually better to use the mailinglist for anything which requires people to be present
<ng0>some might be here.
<stefanX2ovic>ng0: Does the Guix Europe foundation have an official mail address or web page?
<ng0>there's a git for it
<stefanX2ovic>ng0: is down.
<happy_gnu[m]>stefanX2ovic: you might as well take a look at website
<ng0> yw
<ng0>ACTION -> noodles
<ng0>+ the news on the website.
<joshuaBPMan>hello, can someone point me to a config.scm that uses nginx? You can view my config here if you're really curious:
<ng0>same repo
<joshuaBPMan>thanks ng0. I haven't got a clue how that works, but thanks. :)
<ng0>I have some modification and old configs, but they are in a terrible state
<joshuaBPMan>So is the link you sent me your nginx service configuration?
<joshuaBPMan>Or is that how you build nginx?
<ng0>I think not even half of the corrections are pushed... proceed with extra reading skills:
<ng0>no, the link I sent is the one
<ng0>and there's a directory "nginx"
<ng0>which is used by the system config
<joshuaBPMan>ok. You're using a :#config-file ...
<joshuaBPMan>I guess I'll have to try that.
<joshuaBPMan>that's probably the easiest way to get set up and running.
<joshuaBPMan>ok... (nginx-service #:config-file "/etc/nginx/nginx.conf") is not working
<joshuaBPMan>i've got "web" in my use-service-modules and use-package-modules sections...
<joshuaBPMan>the error I get when I try to reconfigure is:
<joshuaBPMan>/etc/gnome-awesome.scm:163:18: /etc/gnome-awesome.scm:163:18: Unbound variable: nginx-service
<ng0>I'm not the best person to ask for advice. I have to leave soon. But others might help.
<joshuaBPMan>ok. Thanks for sending the links anyway!
<cbaines>joshuaBPMan, the procedure nginx-service was recently removed
<cbaines>you may need to switch to doing something like (service nginx-service-type ...)
<joshuaBPMan>cbaines: ok. Perhaps I'm reading outdated documentation.
<joshuaBPMan>cbaines: also do you know why it was removed? It seems a bit odd to remove that procedure. There's still a dovecot-service, dico-service, mysql-service procedures...
<cbaines>joshuaBPMan, I was the one who removed it, because I think that the new style is preferable, now that a <service-type> can have a default value
<cbaines>at some point it would be good if the services were a bit more consistent
<cbaines>as you can see, that hasn't happened yet
<cbaines>apologies if its caused some confusion
<civodul`>so, what freedom-respecting single-board computer would you recommend these days?
<civodul`>i just need a headless thing
<apteryx[m]>Eoma68 or something like it maybe?
<ng0>maybe in the next generation or the one after that, depending on what you need the eoma68 for
<rekado>eoma68 still isn’t ready for shipping. November maybe.
<apteryx[m]>There's also a nice power9 board in pre-orders from Raptor Engineering (Talos II), although this is quite an investment (5000 USD for a complete desktop)
<rekado>(but it looks like luke is busy arguing on mailing lists, so maybe it won’t happen this year)
<civodul`>ok, but among the cheap ARM things available today? :-)
<civodul`>rekado: re luke, that was also my impression :-)
<ng0>apteryx[m]: that's not a single-board computer, the talos ii ;)
<rekado>davexunit: have you tried the me_cleaner scripts with your x220?
<jonsger>civodul`: maybe risc-v ?
<rekado>davexunit: I’m looking for a replacement for my shattered x200s :-/
<efraim> the rock64 is the next board I would get, but it has mali graphics
<joshuaBPMan>cbaines: I'm sorry if I came across as rude. I was just confused. So the new style is (service <service-name> ...) ?
<ng0>why "but" efraim ?
<apteryx[m]>ng0: Oh, i failed to read single board as a SOC based board, sorry ;)
<joshuaBPMan>new defacto style
<cbaines>joshuaBPMan, no offence taken, its a good question
<efraim>lima drivers have stalled, and mali shows no interest in FOSS drivers
<joshuaBPMan>good :)
<cbaines>I prefer that style, but it's not defacto
<civodul`>efraim: what about Beagle Bone, how is it?
<efraim>i don't have one, but I hear they only have 512MiB of ram
<efraim>the beagle bone black at least
<ng0>I was able to compile coreboot on it until recently (system got too old)
<cbaines>I started using it when writing services, as it's simpler. The implementation of procedures like the now removed nginx-service are often very verbose, but all they do is wrap the service configuration with defaults
<ng0>efraim: do you think the rock64 would make a decent video streaming device (for kodi and webbrowser streams etc)?
<joshuaBPMan>cbaines: ok. I guess that makes sense.
<efraim>right now I stream from a 1st gen RPi with 256MB of ram, so I get the impression that if it doesn't need to transcode anything then just about any device should be ok
<jonsger>civodul`: is that hifive thing to slow? risc-v is open hardware, I don't know how it stands with firmware/drivers
<jonsger>but for me it's a really interesting thing to watch
<civodul`>jonsger: it does look nice, but i was looking for something already mature and that can run Guix
<jonsger>okay. 16kB RAM seems to be little to less for Guile... :)
<civodul`>efraim: the RPi3 has problems with the video drivers and boot firmware, is that right?
<Apteryx>civodul`: might be a silly thing to say, but maybe some older phone that is compatible with Replicant could be interesting?
<civodul`>efraim: i mean freedom problems
<efraim>civodul`: from what I've heard, yes
<Apteryx>One with at least 2 GB of RAM if possible (for guix pull ;)
<civodul`>bah, don't tell me about that one
<civodul>i kinda made progress on it:
<civodul>it keeps me awake at night...
<ng0>good (the progress) :)
<jonsger>Members of riscv foundation: AMD, Nvidia, IBM, Qualcomm, Samsung, Mediatek :P Do they not trusting in {PowerPC, x86, ARM}...
<efraim>civodul: the cubietruck might be an interesting option, a little older but it has mainline linux support
<efraim>freedom box has a wiki with some possible hardware options they're looking at
<Apteryx>civodul: thanks for your efforts in cornering the problem :)
<efraim>seems arbian has a couple of boards with mostly complete mainline support, the odroid xu4, tinkerboard
<efraim>gdk-pixbuf replacements are define-public on purpose?
<efraim>ACTION is about to head off to bed
<civodul>hmm maybe not
<nee``>sneek: later tell thomasd I moved the system links like you said and reconfigure worked now. Thank you!
<nee``>sneek: later tell thomasd in the new parameters file (store (device ...)) is still #f
<sneek>Will do.
<happy_gnu[m]>If a package is on github
<happy_gnu[m]>What should I put on
<happy_gnu[m]>source (origin (method url-fetch) (uri (string-append
<happy_gnu[m]>The .git link or the zip?
<bavier`>happy_gnu[m]: the release archive url is fine
<mray_web>anybody using riot here? I though should make me join here....
<happy_gnu[m]>mray_web: I do
<happy_gnu[m]>Are you on browser or mobile
<mray_web>happy_gnu[m]: i'm running the native desktop app, but use my own server
<mray_web>happy_gnu[m]: what is your matrix room id for this room?
<mray_web>to ask my actual question: when will there be an easy option to try guix sd? (like live USB)?!
<bavier`>mray_web: there is an x86_64 vm that you can try out in qemu