IRC channel logs


back to list of logs

<the_tubular>I don't understand that... Why when user A guix puill && guix upgrade gets version B and user B guix pull && guix upgrade gets version C ?
***Xenguy_ is now known as Xenguy
<trevdev[m]>If you use guix home, and you have made an update to your own personal channel and you would like to benefit from that update without updating everything would you pull and reconfigure for it?
<char[m]>Thanks podiki and examors
<trevdev[m]>Could I potentially put my personal channel in a different file and pull it separately from my primary channels.scm, or will that just cause me to loose packages from one channel or the other?
<podiki[m]>you could pin channels to commits, so you only update one channel's commit maybe?
<sektor[m]>#+BEGIN_QUOTe... (full message at
<boud>In case anyone is interested, we now have a Libera Chat channel #maneage - with the aim of bridging it to our main matrix channel. For the moment the only person on #maneage on liberachat is me, though. :P
<boud>There has been some #maneage chat on #tukaani, related to the role of xz in maneage and a library symbol bug introduced by RHEL into CentOS 7 for xz.
<Kabouik>What are the use-modules required to enable virtualization with libvirt? I already added (gnu packages virtualization) but libvirt-service-type is still considered an unbound variable
<Kabouik>That was simply (use-service-modules virtualization), I just didn't find it in the manual so I did not try it immediately
<lilyp>service modules are for services and package modules for packages :)
<zamfofex>Some more updates: I decided to try to run the QEMU command to start the Hurd VM manually, and it seems to not have proper networking — when I try to connect to it with SSH, the QEMU command complains: ‘Slirp: Failed to send packet, ret: -1’ I’m not fully sure about what that means, but investigating it further, it seems my host does not allow using KVM within the VPS.
<zamfofex>Even if that’s not the cause of the issue, it means the VM would likely be prohibitively slow, so I’m investigating using DigitalOcean instead.
<unmatched-paren>Cairn: I've finally replied :)
<Luk6655>mystery solved!... it was /tmp that was taking most of space on my root fs. Now that I cleared it, /gnu/store is about 11GB on disc
<pinoaffe>Hi guix! I'm trying to package davmail, and I'm running into an issue during the install phase (using `install-from-pom`)
<pinoaffe>I suspect it has something to do with file encoding of the pom, the relevant section of the log may be found at
<pinoaffe>does anyone have a clue how I can solve this? I have tried finding potentially helpful snippets/phases in other packages, but have yet to find anything that does the trick
<pinoaffe>oh and the package definition is at
<unmatched-paren>pinoaffe: that home-page seems to 404, btw
<unmatched-paren>pinoaffe: I wonder if the file is using \r\n lines instead of \n?
<unmatched-paren>file says pom.xml: XML 1.0 document, Unicode text, UTF-8 text
***Dynom_ is now known as Guest1009
<strangepr0gram>hi can somebody please help me with my manifest issue
<unmatched-paren>strangepr0gram: sure
<vivien>Matrix bridge is going down
<unmatched-paren>"...going down, going down, Matrix bridge is going down, my fair lady" :P
<strangepr0gram>i cannot ugprade or install packages anymore, seems manifest is in an "unsupported format" ; for current profile manifest is ; i can't really remember anything significant i've done besides removing a channel which was provenant for 1/2 packages
<unmatched-paren>strangepr0gram: bug#57306 <- sounds like this one
<unmatched-paren>or, rather,
<strangepr0gram>thank you I'm trying time-machine now
<ulfvonbelow>to follow-up on my earlier issues with gdm, it turns out that the actual problem was that gdm apparently panics if it doesn't detect hardware acceleration, and it had trouble finding mesa because of some tinkering I did to get mesa use to go through libglvnd. In short, the issue was some changes I made in my local repository that gdm was less than clear about (and the occasional permissions issue when switching between gdm and sddm
<ulfvonbelow>obfuscating things further).
<ulfvonbelow>by the way, is there a way to select the keyboard layout from within gdm? This system is shared, and some users use different keyboard layouts
<strangepr0gram>hi i'm still not sure what to do regarding wrong manifest
<lilyp>ulfvonbelow: you'd have to meddle with gdm's dconf settings, but I suppose it's doable
<lilyp>not declaratively tho
<strangepr0gram>unmatched-paren: sorry for pestering you this is the `actual` profile manifest and not system manifest, i wonder if this one's the issue
<strangepr0gram>i understand correctly that /var/guix/profiles/per-user/.../guix-profile/manifest is the ~user~ manifest and ~/.config/guix/current is ~system~ manifest
<strangepr0gram>(im running guix on top of debian)
<unmatched-paren>~/.config/guix/current is the pull manifest
<unmatched-paren>/run/current-system/profile/manifest is the system manifest
<Kabouik>I see that that patch from 2020 never got an answer, was there something wrong with it?
<pkill9>it probably just got forgotten
<unmatched-paren>that's sadly quite common, there aren't enough people to keep track of everything
<Kabouik>matterhorn deserves some love, bud admittedly that's quite a big patch there
<unmatched-paren>Kabouik: i'll review it
<unmatched-paren>but i can't merge it, no commit access
<Kabouik>Awesome unmatched-paren, thanks a lot! It's a really nice client, especially as the GUI one is very CPU intensive.
<Kabouik>Speaking of, I need to finish my nmail patch (it works, I'm using nmail currently), but would prefer to wait for the dev to make a new release with a couple fixes first
<unmatched-paren>review sent
<unmatched-paren>And aerc 0.12.0 includes the new :import-mbox command, so I can finally view messages from before I subscribed :D
<Kabouik>Thanks a lot unmatched-paren, now crossing fingers in hope that the patcher is still around and using Guix
<unmatched-paren>If not, maybe you could adopt the patchset?
<Kabouik>I wanted to before I saw how big it is, it's a bit overwhelming for me'but surely I'll try if they are not around amymore
<unmatched-paren>Eh, it's not too many.
<unmatched-paren>The aerc patchset has 42 packages in the latest version :S
<Luk6655>did anyone manage to get gnome 42 work on guix?
<lilyp>parts of gnome 42 require stuff happening on core-updates
<Luk6655>I'm not sure what does this mean, is core-updates some package
<unmatched-paren>core-updates is a branch
<unmatched-paren>for updating/modifying packages that have lots of dependents
<unmatched-paren>so that we don't rebuild the world every time we merge a modification for one of those packages
<Luk6655>I see, so does it mean there exist package definitions for gnome 42, they are just not merged because it would trigger a lot of rebuilds?
<unmatched-paren>no, said package definition would require changes on core-updates
<Luk6655>if one uses the main channel yes, I assume people doing development on guix have their own local branches
<Luk6655>like I do
<Luk6655>so I'm really asking if anyone managed to get it working using their local repo for example
<unmatched-paren>probably not
<Luk6655>When did gnome 41 appear in guix? It was released in September 2021
*unmatched-paren git log --grep
<Luk6655>indeed, I was hoping someone may remember...
<unmatched-paren>November 2019
<unmatched-paren>it can't have been September 2021
<unmatched-paren>Oh, wait
<unmatched-paren>Oh, yeah, 19 was the date
<unmatched-paren>so November 2021
<Luk6655>I just looked at this article from 23rd Sept 2021
<Luk6655>the exact time doesn't really matter, I'm just interested how much time it took last time
<oriansj>unmatched-paren: you can do git log --date=iso-strict --grep
<unmatched-paren>oriansj: Oh, thanks!
<strangepr0gram>hi, i tried running "guix time-machine -C ~/.config/guix/channels.scm --commit=c9fbd40785a99e13a59d8e530830ce85220a9871 -- guix pull" as "bug was fixed in" c9fbd40785a99e13a59d8e530830ce85220a9871. i got a build error which is: "unbound variable ~S"
<AwesomeAdam54321>stranepro0gram: Are you guix time traveling to the future?
<Luk6655>yes, there is a log from Nov 2021 as you said
<Luk6655>gnu: gnome-shell: Update to 41.0.
<unmatched-paren>yeah, I did `git log --grep "gnome-shell: Update to 41"'
<Luk6655>the only reason why I'm interested in 42 is that the current version appears broken with an 'unmentionable' graphics driver
<strangepr0gram>AwesomeAdam54321: errrrrrrr d--dont think so, that commit is since july, and i started having the manifest issue only about a week or two ago
<Luk6655>even version 42 is partially broken still, but at least it is fixable, while gnome in general seems the only dm that can be easily forced to look good on a high dpi screen
<strangepr0gram>oh interesting
<Luk6655>mate is not that bad, with exception of the window-list applet that uses 16x16 icons and there is no way to change it without patching libwnck (or mate-panel of course)
<strangepr0gram> i've tried listing generations and at generation 35 it stops and says unsupported manifest etc... it doesn't event exist
<unmatched-paren>strangepr0gram: `guix describe`?
<unmatched-paren>strangepr0gram: the fix is after the commit you're on
<strangepr0gram>guix-time-machine -C ... --commit=ac8e987ce2
<strangepr0gram>-- guix pull * sorry i'm very new
<unmatched-paren>sorry, i don't know how to help you further
<AwesomeAdam54321>strangepr0gram: Can you try switch to a previous generation before the commit, and try to run the command again?
<AwesomeAdam54321>*before commit 72abf9b9b82d2f5f34b7d22b7aae6f2398336a93
<AwesomeAdam54321>Maybe you can run the command using /usr/local/bin/guix
<strangepr0gram>still same error
<AwesomeAdam54321>Sorry, I don't know how to solve the problem
<strangepr0gram>ok thank you
<strangepr0gram>oh weird it's solved now ?
<strangepr0gram>pulled from 36 to 34 (because 35 doesn't exist)
<strangepr0gram>er rolled-back*
<strangepr0gram>wait nope still not working
<strangepr0gram>works now
<strangepr0gram>:shrug: o_O
<strangepr0gram>seems guix doesn't allow one to do anything if there's atleast one missing user-generation
<strangepr0gram>wait no only guix package -u works, guix pull still doesn't
<Kabouik>Awesome examors, I didn't see your message until now about the GRUB options, but tried them and they seem to work: the screen is properly rotated after the GRUB screen (but not before; somehow it was properly rotated out of the box when the machine still had its Windows install)
<Kabouik>Touch coordinates are still wrong though, I'll check with other GPD Pocket 3 owners
<boji2>I am having trouble installing fresh guix, getting this error: /mnt/boot/efi doesnt look like EFI partition
<boji2>I am new so if its something obvious please excuse me
<examors>Kabouik: nice, glad to hear :)
<ulfvonbelow>is there any way to pass --allow-collisions or equivalent option to "guix home reconfigure"?
<unwox>boji2: are you using graphical installer?
<unmatched-paren>ulfvonbelow: you can't possibly allow collisions
<boji2>unwox: Yes I am :D
<ulfvonbelow>I mean it worked okay when I was managing my ~/.guix-profile with "guix package -u" and "guix package -i", etc
<unwox>boji2: did you create uefi-boot partition? it should be formatted as vfat and have mount point "/boot/efi"
<ulfvonbelow>it looks like the installer is expecting it to be mounted on /mnt/boot/efi for installing the bootloader, though
<boji2>unwox: oh I see, I did it as ext4 haha I was just going through install :) Using guix atm on my laptop and decided to put it on pc as daily driver, but not really sure what are benefits of making it EFI instead of regular since laptop is on "normal" install?
<unwox>by "normal" you mean bios?
<boji2>I suppose so, really fresh into computer world and linux :)
<unwox>boji2: well, for uefi setup you'll need at least 2 partitions: etx4-formatted root "/" and vfat-formatted /boot/efi (1-2 GB should be enough)
<boji2>unwox: Thanks a bunch :) Also would you consider EFI an upgrade or different option? Trying to wrap my mind around the concepts
<ulfvonbelow>the advantage of efi is that I have no idea how to tell grub to install to a particular UUID or partition name, but I do know how to tell guix to mount a particular UUID or partition name in a particular place in the filesystem, and I do know how to tell grub-efi to install to a particular place in the filesystem
<boji2>Thanks it clarifies things up a bit :D
<ulfvonbelow>there's probably a way to tell grub to install to a particular uuid, and in that case I can't think of any particular difference that really affects me
<unwox>there is not much difference between uefi and bios unless you are going to add more than 2TB of storage to your system i'd say
<unwox>ulfvonbelow: you can specify uuid for root partition (at least using mapped-devices)
<boji2>Is vfat-formated partiton fat32 or ntfs? In type selection cant find vfat option :)
<boji2>Thanks :D
***jesopo is now known as jess
<ulfvonbelow>unwox: right, the trouble though is specifying the 'target' field of bootloader-configuration as a uuid
<ulfvonbelow>don't wanna switch some wires around, reconfigure, and find out that one of my filesystems got wiped out because I specified /dev/sda1 or something like that
<boji2>Now getting EFI manager failed to register boot entry: Operation not permitted
<boji2>thats when choosing fat32 format for /boot/efi and ext4 for mount at /
<lilyp>did you actually mount the efi partition?
<unwox>ulfvonbelow: got you, you are right
<boji2>lilyp: not really sure as to what you mean? :)
<unwox>boji2: if i remember correctly graphical installer should give you a config.scm before actual installation. could you share it with us somehow?
<unwox>you may also need to disable "secure boot" in bios settings (f2/del on boot)
<boji2>Quickly checked bios to see if safe boot is off before posting log, only thing that could possibly impact in bios settings is sata mode: ahcii or intel rst premium with intel optane system acceleration, change that maybe, right now ahci is on and safe boot is off
<jpoiret>ahci should be on for linux to work properly iirc
<jpoiret>the pseudo-RAID mode is brokne
<boji2>Thanks, will try one more time see if works
<pkill9>does anyone ahve an example of extending the guix commadnline by using the guix/extensions directory?
<muradm>hello guix
<muradm>i had a question few days ago, what is the effective linux kernel configuration for default kernel packages? as far as i can tell from (gnu packages linux) it seems to be "make defconfig" + %default-extra-linux-options basically for x64 least
<muradm>i'm curious which kernel features would be mandatory to run guix if there are such?
***lukedashjr is now known as luke-jr
<jpoiret>i'm curious how people use emacs on guix. I often want to use modes corresponding to what i'm working on using guix shell, and so would need to launch an instance per project so that the lisp files are available to them
<jpoiret>this is quite the opposite to the common knowledge of "run one system-wide daemon", which is why I was wondering how other people manage
<unmatched-paren>muradm: Hi, did you get my previous message about wlgreet? I didn't see a response ping, apologies if there was one
<unmatched-paren>pkill9: See gwl and guix-modules
<lilyp>jpoiret: I have my precious packages managed user-wide with per-project settings per project – people do use direnv stuff, but I'm in the spawn another emacs camp as well
<jpoiret>maybe i'll consider even running it in a terminal, although I'd miss too many goodies
<jpoiret>it doesn't feel right to `guix shell` then `emacs &`
<lilyp>you don't need the ampersand tho?
<jpoiret>if I need to work in the same dir at the same time, since in any case i'll still have a terminal lying around
<lilyp>M-x term, shell, eshell ?
<jpoiret>i'm too scared of ncurses stuff breaking
<examors>muradm: I believe the kernel configs are in gnu/packages/aux-files/linux-libre
<lilyp>also, you can spawn a graphical emacs from a guix shell too
<muradm>unmatched-paren: hi, which ones and when?.. if we are talking this, mostlikely i missed somethings
<unmatched-paren>muradm: I won't be able to find it, since the logs search doesn't work, so I'll just ask again
<unmatched-paren>those patches fixing sway crashing when used with a greeter were merged, right?
<unmatched-paren>for some reason it's still crashing for me
<unmatched-paren> <- if i have to change something: here's the service source
<muradm>unmatched-paren: yes, they are merged, now it works fine, just make sure that you are providing correct groups to greeter-supplimentary-groups like '("seat" "video" "input")
<muradm>unmatched-paren: here is mine working
<muradm>with gtkgreet
<unmatched-paren>Oh, thanks
<muradm>examors: those should be additions depending on specific version i suppose
<unmatched-paren>do i need to add "seat" to my user too?
<muradm>i rememember there was an issue with hardlinks or alike which caused problems, so i wander if there is a reference list of features expected from kernel by guix
<muradm>unmatched-paren: yes of course
<unmatched-paren>okay, thanks
*muradm reboot.. brb
*unmatched-paren reconfigure
<unmatched-paren>WOOT IT WORKS
<unmatched-paren>tysm muradm
<unmatched-paren>i'm gonna send a patch for wlgreet now
*muradm custom linux attempt 1 fail 1 success 0
***wielaard is now known as mjw
<pinoaffe>unmatched-paren: I think the file does indeed use dos-style newlines, I didn't think that would be an issue
<pinoaffe>unmatched-paren: oh lol, that url should should be sourceforge rather than sorgeforge
<unmatched-paren>pinoaffe: I wonder whether the sxml parser doesn't accept \r\n, is all
<pinoaffe>that may indeed be the case
<unmatched-paren>try running it through dos2unix in a build phase
<pinoaffe>yeah I was wondering whether there's a canonical way to do that in a buildphase or whether I should just aadd dos2unix to the native inputs and use that
<unmatched-paren>probably the latter
<pinoaffe>I guess I could just use substitute*, that way I don't add another input
<unmatched-paren>Oh, yeah, fair enough
<pinoaffe>doesn't matter much either way
<pinoaffe>thanks for the help, I'll go and try that!
<singpolyma>inputs don't cost at install or run time, so 🤷‍♂️
<pinoaffe>okay so trying to run substitute* on the file results in very similar errors originating in the substitute* call, and running dos2unix does not fix it
<unmatched-paren>woot, ported greetd-wlgreet-sway-session to guix
***mark__ is now known as mjw
<unmatched-paren>sneek: later tell muradm: <- wlgreet patch sent! Thanks for your work on greetd; since you're the expert on it around here, would you mind doing a cursory review of it to see -if- what I got wrong? :P
<unmatched-paren>sneek: good bot
<unmatched-paren>Ah, only this works:
<unmatched-paren>sneek: goodbot
<unmatched-paren>wait, what? some
<unmatched-paren>someone did this in #guile, i swear
<unmatched-paren>Ah, it's just
<unmatched-paren>sneek: botsnack (I give up :))
<boji2>Still having issues with fresh install on pc, this time getting error: grub-install /mnt/boot/efi doesnt look like EFI partition