IRC channel logs


back to list of logs

***rekado_ is now known as rekado
<KE0VVT>`guix upgrade` takes a while. Two and a half hours.
<vagrantc>it depends.
<sneek>Welcome back atka :)
***wonko-th1-sane is now known as wonko-the-sane
***iyzsong-www is now known as iyzsong-w
<atka>sneek: botsnack
<the_tubular>atka: botsnack
<the_tubular>Dang, that bot has lattency issues ...
<atka>how's the_tubular today
<the_tubular>Pretty good, I'm playing with guix home a bit
<the_tubular>How about you ? Still learning ?
<atka>did you use the guix home import feature?
<the_tubular>No, what can you import with it ?
<atka>it backs up your current home and generates a skeleton home-configuration.scm for you
<atka>but I found some issues with it so I was just curious
<the_tubular>Umm, what issue did you find ?
<the_tubular>Also, did you play with the new shepherd ?
<atka>not yet, is it a guix pull away now?
<jfred>oof... my laptop's battery died unexpectedly during a `guix home reconfigure` and now there seem to be some empty derivation files in the store. `guix gc --verify=repair,contents` doesn't seem to rebuild those store items, and now subsequent runs of `guix home reconfigure` bail out with "error parsing derivation `/gnu/store/nnp[...]-node-v10.24.1.tar.xz.drv': expected string `Derive(['"
<jfred>any way I can fix this without a complete reinstall?
<atka>can you roll back your home and reconfigure?
<jfred>I don't actually have a new home generation since `guix home reconfigure` never actually finished
<jfred>so nothing to roll back to, I'm still on the previous generation
<jfred>I assume the store items that are empty are for things that would have been part of the new generation had that process completed
<atka>can you switch to a previous previous generation?
<jfred>I mean... I could, but my current generation isn't broken
<jfred>it's just that I can't build a new one
<atka>hmm, I was thinking maybe if you went back further it would possibly reconfigure
<atka>or maybe if you're confident in your current generation delete the others and garbage collect them
<atka>I was able to recover from guix system not being able to list-generations or reconfigure by deleting all generations and garbage collecting them, maybe home would work similarly
<jfred>oh, yeah garbage collecting might work... I was just trying `--repair` but yeah I shouldn't need any of the new/empty store items anyway as they aren't in use
<jfred>that did the trick \o/
<jfred>thanks for the tip :)
<atka>np, glad it worked!
<jfred>as a bonus, that freed up about 26 GB :P
<atka>yes at the expense of previous generations
<atka>if I'm not mistaken
<jfred>yeah, but better that than a busted guix store haha
<atka>very true
<xelxebar>ARGH. The ncurses package definition hard-codes version string deep into one of its phase procedures.
<xelxebar>Trying to get a def for ncurses@5, but that just foiled my efforts.
<xelxebar>Having a make-ncurses-package procedure or something would be nice for this particular thing. Not sure if that's a holistically reasonable thing to have, though.
<xelxebar>Hrm. And docker-compose is failing to build (commit 28b3763e33cd71fe281cd71a568c54fea7f4cf4f):
<mroh>yeah. since 66177c2b87 I believe.
<mroh>looks like, we need to keep an python-pyyaml-5 package (or hack requirements, maybe).
<AIM[m]>Idky but I can't seem to transfer files from my android to my guix machine
<AIM[m]>I've installed libmtp, gvfs, fuse, udisks
<civodul>Hello Guix!
<AIM[m]>How do I start udisks service without desktop-services
<civodul>AIM[m]: you can use the udisks service individually:
<AIM[m]>It's part of desktop-services aye?
<atka>AIM[m]: use service-modules-desktop and define your services
<atka>the udisk-service-type is exported by desktop.scm
<AIM[m]>(service-modules-desktop #:udisks udisks)
<AIM[m]>Idk I didn't get ya
<civodul>the manual doesn't mention service-modules-desktop
<AIM[m]>Still finding my way in guix and guile
<civodul>to use udisks, you would add (udisks-service) to the 'services' field of your config
***iyzsong-www is now known as iyzsong-w
<AIM[m]>civodul: Is this something to import
<civodul>so the value of 'services' field would look like, say: (append (list (udisks-service)) %base-services)
<atka>sorry I mean (use-service-modules dbus desktop) etc at the top of your config.scm, otherwise it would be unbound no?
<AIM[m]>atka: Yep Yep I have it
<civodul>atka: oh right, indeed
<AIM[m]>civodul: Thank you so much
<AIM[m]><atka> "sorry I mean (use-service-..." <- Thank you as well atka
<atka>AIM[m]: glad to help
<AIM[m]>Idky but I still get nothing
<AIM[m]>I installed fuse, libmtp, gvfs, udisks and enabled udisks service
<AIM[m]>Thunar doesn't show my Android device
<civodul>hmm i vaguely remember another service for Android devices but i can't find it
***iyzsong-www is now known as iyzsong-w
<AIM[m]>fuse ain't a service right?
<civodul>ah, found it: (simple-service 'mtp udev-service-type (list libmtp))
<civodul>this lets you access "MTP devices" (it's part of %desktop-services)
<civodul>where MTP is "Media Transfer Protocol"
<Guest6>I get the error:
<Guest6>bash: /path/to/binary: No such file or directory
<Guest6>for some binary files in guix. But I've no idea what is causing this. I looked at one of them with `ldd`, and it was missing a bunch of library deps. I adjusted LD_LIBRARY_PATH from the paths of the libs as I found them in /gnu/store (which doesn't seem like the right approach). This fixed the ldd error, but I still get the "No such file or
<Guest6>directory" error.
<civodul>Guest6: hi! in general, i'd recommend leaving LD_LIBRARY_PATH unset; fiddling with it is likely to lead to problems
<civodul>however, the "No such file" error is likely because the binary was built for another distro
<civodul>and expects /lib64/ld-linux*.so to exist
<civodul>it does not exist on Guix System
<civodul>so instead, you should rebuild it from source
<civodul>or, as a last resort, patch the binary with patchelf
<AIM[m]><civodul> "where MTP is "Media Transfer..." <- Ik Ik
<AIM[m]><civodul> "ah, found it: (simple-service '..." <- I'll try this now
<AIM[m]><civodul> "ah, found it: (simple-service '..." <- There's a space after 'mtp ???
<Guest6>Thanks civodul! So, if LD_LIBRARY_PATH is unset, but ldd is showing as missing libraries (as in this case) what should I have done? Or are the libraries somehow missing because of /lib64/ld-linux*.so problem? Naively, it seems to me like two problems...?
<xelxebar>mroh: Okay. Thanks for the ACK. Would it be safe to assume the issue is known and being handled?
<Guest6>Here is the output of ldd, it seems to have found ld-linux to me?
<Guest6> (0x00007fff808e0000)
<Guest6> => /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33/lib/ (0x00007f422c550000)
<Guest6> => /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33/lib/ (0x00007f422c54b000)
<Guest6> => /gnu/store/8qv5kb2fgm4c3bf70zcg9l6hkf3qzpw9-zlib-1.2.11/lib/ (0x00007f422c52e000)
<Guest6> => /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33/lib/ (0x00007f422c524000)
<Guest6> => /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33/lib/ (0x00007f422c362000)
<Guest6>    /lib64/ => /gnu/store/5h2w4qi9hk1qzzgi1w83220ydslinr4s-glibc-2.33/lib/ (0x00007f422e2d9000)
<Guest6> => /gnu/store/094bbaq6glba86h1d4cj16xhdi6fk2jl-gcc-10.3.0-lib/lib/ (0x00007f422c346000)
<civodul>Guest6: yes, it looks good
<civodul>(please use or similar service next time)
<Guest6>Yeah, sorry. I should have. Thanks for pointing that out.
<civodul>there aren't any missing libraries here, AFAICS
<AIM[m]>(simple-service 'mtp udev-service-type (list libmtp))
<AIM[m]>Is it mtp-udev-service-type ?
<Guest6>I'm sure why doesn't point to anything? But I'm not that familiar.
<Guest6>civodul I might be confusing the issue, that output was after setting LD_LIBRARY_PATH. Before setting it, the entry said "not found".
<AIM[m]>civodul: device shows up but it says: "Message recipient disconnected from message bus without replying"
*AIM[m] uploaded an image: (105KiB) < >
<spiral88>Hi there, I'm really excited to install Emacs 28.1. Anyone have an idea of how long it might take for the guix package to be updated? I know I could install emacs-next and will just do that if it'll be a while.
<phodina[m]>could somebody please tell me why guix evaulates this as error? I'm just attempting to use the gexp and the error message is not helpful since I'm attempting to modify phases.
<civodul>phodina[m]: hi! do you have #:use-module (guix gexp) in the define-module form?
<civodul>also, "-L foo.scm" is not what you want; should be "-L some/directory"
<AIM[m]>Am I supposed to run polkit-service and dbus-service ?
<AIM[m]>Doesn't work when put into services list
<phodina[m]><civodul> "phodina: hi! do you have #:use-..." <- yes, there is `-L.` or the `-L<dir>`, just how I build it ... but the question is about line 27. If I don't use gexps it works but I want to learn them and the message `'replace' may only be used within 'modify-inputs'` is so confusing ... I looked at `ppsspp` package as an example and can't figure out what's wrong
<phodina[m]>> <> phodina: hi! do you have #:use-module (guix gexp) in the define-module form?
<phodina[m]> * <del>yes, there is `-L.` or the `-L<dir>`, just how I build it ... but the question is about line 27. If I don't use gexps it works but I want to learn them and the message `'replace' may only be used within 'modify-inputs'` is so confusing ... I looked at `ppsspp` package as an example and can't figure out what's wrong</del>
<phodina[m]>Thanks, the issue was the `gexp` module was not listed in the modules to be used.
<ZhuZihao[m]>What do you think about whether gst-plugins-good should propagates gst-plugins-base or not?
<civodul>phodina[m]: yes, that's a confusing error
<AIM[m]>civodul: Any clue on how to fix the error when trying to mount mtp device on thunar: "Message recipient disconnected from message bus without replying"
<spiral88>In answer to my question, spinning up emacs 28 this way works nicely :) `guix shell emacs-next -- emacs`
<sneek>wb atka
<civodul>AIM[m]: i'm afraid i can't help here, i don't use such devices
<civodul>if people on this channel can't help, you can try the mailing list
<civodul>(you don't need to subscribe)
<civodul>spiral88: neat!
<civodul>speaking of which, shouldn't we set up an emacs-28 branch with the upgrade?
<civodul>i'm surprised this hasn't been done yet :-)
<AIM[m]>civodul: Okay
<AIM[m]>Thunar isn't even showing my usb tho
<AIM[m]>Seems to be not related to mtp?
<AIM[m]>But some gvfs issue?
<civodul>AIM[m]: could be, not sure
<civodul>BTW, if you use Xfce, i'd recommend using plain %desktop-services + xfce-desktop-service-type
<AIM[m]>I'm actually not on xfce
<civodul>it's likely to have things set up correctly
<civodul>ah ok
<AIM[m]>Kinda on dwm
<AIM[m]>gvfs just needs the gvfs package installed right?
<civodul>i think so
<AIM[m]>Do I need yo be part of disk group
<bost>Hi. I want to update openjdk 17.0.1 -> 17.0.2. It's a simple change of the version number and the checksum... however I'm curious how the checksum gets computed? `guix download` returns a checksum which is incorrect.
<rekado>bost: depends on the source files
<rekado>we’re using a git repository for the sources of openjdk 17
<jab>morning guix!
<bost>rekado: ???
<rekado>bost: you’re looking at a zip file and compute its hash. We aren’t using that zip file.
<ZhuZihao[m]>bost: You need to unzip and calculate the hash of the folder
<rekado>it’s a git repository
<rekado>get the hash of the git repository
<rekado>we are not fetching a zip file and we don’t use the hash of its unpacked contents
<ZhuZihao[m]>bost is using the Github archive, it just the content of specific commit of git repo
<ZhuZihao[m]>Though Github will recreate Github archive zipball, but the hash of inner content will be the same
<rekado>our package definition is using git-fetch
<bost>rekado: So what's command for calculating hash of a git repo?
<rekado>get the git repo and use guix hash -rx
<rekado>the manual has an example for this exact situation
<ZhuZihao[m]>rekado: I do understand this, but in practice you can use the hash of the unzipped result of Github archive. If you can git repo, your repository checkout should be fresh, you have to clone it from scratch or you may need extra cleanup
<ZhuZihao[m]>If you use git repo
<ZhuZihao[m]>Github Archive is undeterminism because it will re-create the zipball regularly and introduce undeterminism, But the inner content of it is actually idempotent
<ZhuZihao[m]>IIUC, -r is deprecated
<ZhuZihao[m]>guix hash -S nar -x <some_directory>
<bost>ZhuZihao[m]: deprecated? Hmm, I just executed it and it did not complain.
<ZhuZihao[m]>Aha, It's not deprecated now
<bost>ZhuZihao[m]: `info "(guix)Invoking guix hash"` says "... deprecated [but] ‘-r’ remains accepted as a convenient shorthand"
<tribals>Hi, folks!
<tribals>I still didn't understand how to use "upstream" guix from Git repo. I followed all instructions from "Building from Git" manual section, but still didn't get the "big picture".
<tribals>My questions are: do I need to `make` everything after any source code changes?
<tribals>Do I need to always run `./pre-inst-env` in "clean" environment produced by `guix environment --pure` (or `guix shell --pure`)?
<MohammadSalehKam>Hi... (full message at
***wielaard is now known as mjw
<rekado>tribals: yes, run “make” after changes. It shouldn’t take log at all when only a few files have changed.
<rekado>if you want to use Guix from git you need to use “./pre-inst-env”
<rekado>I dnon’t think you need to be in a pure environment to use it after you’ve built it.
<tribals>rekado: Thanks
<tribals>The point about make isn't clear to me. Guix is a program made of Guile sources. Guile can be compiled at runtime transparently. So, i made a few changes to `guix/home/services/shells.scm`. Then I want to immediately use those changes. I run guix with pre-inst-env: `./pre-inst-env guix home ...` pre-inst-env "sees" just introduced changes, see it is never than compiled version of this file, so it uses uncompiled source. Isn't make does same thing:
<tribals>compiles Guile sources to `.go` objects?
<bjc>i believe guile compares the modification dates on the files, and will use whichever is newer
<bjc>(recompiling the scheme file, if its the newer one)
<tribals>So, if guile recompiles files as needed, then do I still need to run make?
*attila_lendvai has rebased his stuff on master, and reconfigured his server, and doesn't see any issues with the new shepherd
<civodul>attila_lendvai: yay, thanks for reporting back!
<civodul>tribals: yes; auto compilation is turned off by the "guix" command
*civodul fearlessly upgrades Guix machines they have access to
<attila_lendvai>civodul, my pleasure! thanks for the fibers stuff and the new release!
<civodul>twas long overdue!
<civodul>a week of focus turns the idea into a thing
<tribals>civodul: that't missing part I was looking for, thanks ))
<ZhuZihao[m]>tribals: pass -j argument to make can help you compile faster
<ZhuZihao[m]>guile's auto compilation will not utilize the parallel compilation I think
<ZhuZihao[m]>But you don't need `make` for every minor changes.
<sneek>bae, you have 1 message!
<sneek>bae, nckx says: messages left in your absence:
<bae>I'm buying a (used) laptop in a couple days to be my "daily driver" since my HP died. I'm trying to decide if I want to try installing Guix using a video by System Crafters. I'll probably end up with an i5 something with 4GB ram and 256GB of storage. I've been learning Scheme for a couple months, so Guix sounds like an interesting choice. My biggest concern is I don't really understand how the system configuration works in Guix.
<civodul>bae: hi! i would hope that you can quickly understand system configuration better than with any other distro
<bae>Does Guix add an additional layer of complexity to the system? For instance, instead of modifying my .bashrc file, will I be modifying a Guile program that in turn modifies my .bashrc?
<civodul>because it's uniform
***the-porcupirate is now known as porcupirate
<civodul>for system services, you'd modify, say, the openssh-configuration record instead of /etc/sshd_config
<civodul>and there's an "extra layer" that "translates" from one to the other
<gnoo>bae: no, with guix you edit .bashrc manually unless you like guix so much that you use it for that as well (with guix-home)
<ekaitz>hi guix! fast question! is it possible to make the programs keep the debug symbols in the gnu-build-system?
<gnoo>bae: however, for things outside your home directory, you generally will need to use guix system to make changes to it.
<bae>civodul: OK, so in a sense it's like a graphical program that modifies a config file, it's a guile program, and I modify the source code.
<civodul>ekaitz: egun on! yes: #:strip-binaries? #f
<ekaitz>arratsalde on civodul ! eskerrik asko
<civodul>(that's beyond my understand of Euskal :-))
<ekaitz>translation: good afternoon! thank you
<civodul>heheh, alright
<bae>civodul: Will I have any trouble installing any programs with only 4GB ram? I heard that there are some programs that have to be compiled from source, and they might hang if I don't have enough memory.
<civodul>bae: yes; so in a sense, Guix adds a uniform config layer above the services that you use
<civodul>bae: compiling ungoogled-chromium or libreoffice is terrible, but normally you don't need to do this
<bjc>depending on what you install, 4gb will not be enough
<singpolyma>You should never have to compile from source
*attila_lendvai has sent his changes to define-configuration:
<civodul>the build farm at ci.guix does a rather good job at providing binaries in a timely fashion
<civodul>(it used to be different!)
<bae>bjc: civodul I'll be installing emacs, and graphical browser which I need for work. That's about it.
<bae>and EXWM
<bjc>if you don't have to compile it, it should be fine
<bae>bjc: I've actually been torn between trying two different distros, and they are like polar opposites: Guix, and KISS.
<bjc>as for configs: if guix doesn't have a wrapper around the config format, or if you'd prefer not to use it, you can also use plain files, either defined as strings in scheme, or as files on the filesystem
<attila_lendvai>the donate page at points to a broken link:
<attila_lendvai>also discusses "GuixSD"
<apteryx>I'm still a bit confused about package/inherit; should I use it when defining for example a 'java-antlr4-cpp-runtime', which shares the same source with java-antlr4-runtime but builds from a different sub-directory?
<gnoo>bae: which gen i5 is that? if it's an old one like 2nd/3rd gen, you might want to use something else since even system reconfigure and guix pull takes quite a lot of time in these
<bjc>guix is not fast, even on my 3950x
*attila_lendvai drops a mail to guix-devel about it
<gnoo>i refrain from doing guix pull a lot of the time since it can take upto 30 mins sometimes
<bae>gnoo: Yeah, that's part of the reason I've also been considering KISS.
<apteryx>attila_lendvai: perhaps you should contact the GNU webmasters to report it
<bjc>the way i see it, with how often i update my system, i don't need it to be all that fast
<gnoo>bae: i'd suggest void more than kiss, i had good experience with it.
<attila_lendvai>apteryx, judging from the url, it seems to be a user's own page... no?
<ZhuZihao[m]>ekaitz: Add the "debug" output and strip will move debug symbols to that place.
<ekaitz>ZhuZihao[m]: oh that's also cool, what civodul said didn't work :D
<bae>I tried void live usb. I couldn't get it to work right. Maybe Tiny Core would be a good choice?
<ekaitz>(probably because I did it wrong)
<attila_lendvai>apteryx, scratch that! it may be true, but it's not guix specific. i'll contact fsf.
<apteryx>attila_lendvai: seems like it probably was setup by the FSF
<gnoo>bae: tiny core it too tiny
<bae>gnoo: I'm using a live Slax usb right now. It is slow AF. So bad, I'ts almost unusable.
<bae>I'm on a borrowed computer.
<gnoo>a live distro will always be slow, especially on old hardware
<ekaitz>also ZhuZihao[m] do you know how to load them to gdb alongside the executable?
<bae>gnoo: Oh...OK. Yeah, this is an old Lenovo I'm on right now. I'll take another look at void. I think actually I was thinking of that anti-fascist distro that I couldn't get to work properly. Not it's fault, it just had so many different install options for "persistence" that I just didn't feel confident that I was doing it right.
<bae>gnoo: Thanks for the tip.
<ZhuZihao[m]>ekaitz: Have you try "symbol-file" command in GDB?
<ekaitz>ZhuZihao[m]: that worked! thank you very much!
<apteryx>attila_lendvai: thank you!
<attila_lendvai>apteryx, my pleasure! i may end up convincing some people with $$ to use/rely on Guix, too... we'll see.
***the-porcupirate is now known as porcupirate
<apteryx>neat :-)
<nouun>Has anyone packaged gh CLI?
<ZhuZihao[m]>Go packaging is quite annoying
<jackhill>speaking of ungoogled-chromium, it looks like the build is currently failing: I guess the package could use an update as well, but, ooof, what a package to dive into.
<garuda>Hi all, when Guix will have GNOME 42? I am waiting for it.
<sneek>Welcome back garuda, you have 1 message!
<sneek>garuda, podiki[m] says: I saw you asked about haskell-gi, actually I packaged it (working!) but isn't in guix just yet
***garuda is now known as Sughosha
<nouun>ZhuZihao[m]: Indeed, hate dealing with nested packages.
<nouun>Cannot, for the life of me, get to package.
<jeko>Hey Guix !
<sneek>Welcome back jeko, you have 1 message!
<sneek>jeko, gnucode says: that sway works really well on guix.
<jeko>anyone having trouble with emacs in a guix shell environment ?
<jeko>I can't fire geiser because of an issue with a root-project function
<jeko>oh thats not because of the guix shell environment…
<jeko>it s something else…
<jeko>nevermind then, not guix related
<ZhuZihao[m]>What do you think about the kinit of KDE?
<ZhuZihao[m]>The core concept of kinit is: It fork first and load the shared library for a KDE application to start it.
<ZhuZihao[m]>But many KDE package in Guix requires a environment variable wrapper to make it work well. kinit would break our wrappers
<ZhuZihao[m]>typo: the shared library for -> of
<civodul>found my first Shepherd-related bug:
<apteryx>ooh. I guess I do not use non-root LUKS devices in a Shepherd service
<apteryx>so this would be triggered by having a file system definition that lives on a LUKS-encrypted device?
<civodul>apteryx: yes, like an encrypted /home
<civodul>discovered on the laptop next to mine
<civodul>(fortunately the affected user was glad to be able to reboot into the old generation, phew! :-))
***the-porcupirate is now known as porcupirate
<civodul>hmm looks like there are other potential related system test failures:
<jab>how would I go about deleting the gpg sig for guix? effectively disabling substitues for regular it ill adviseable to delete the gpg file?
<jab>maybe guix --archive has a --delete option....
<PotentialUser-59>Anyone else having an issue where guix system reconfigure is stuck on receiving objects? It progresses really slowly and then stops progressing at some point. guix pull works fine
***iyzsong- is now known as iyzsong
<maximed>Confusing terminology of the day (Guix / Rust): ‘TARGET — the target triple that is being compiled for. Native code should be compiled for this triple.’ (from
<maximed>TARGET: simple enough, corresponds to "guix build rust-foo --target=TARGET" at first sight.
<maximed>But then: ‘native code should be compiled for this triple’. native = native-inputs = --system?
<maximed>jeko: Hi
<jeko>Anyone having trouble with Geiser recently ?
<maximed>jeko: FWIW, I'm not, but I don't use Geiser. Also, in the past, people have had problems with Geiser, maybe there are still some unresolved problems or sharp edges.
<jeko>For couple of days it does not want to run when I'm working in a git repo…
<maximed>jeko: I don't expect Geiser to care about whether you're working in a git repo, a mercurial repo or an unpacked tarball
<maximed>is it git-related?
<jeko>when there is no .git folder in the repo it seems to work
<jeko>in the directory*
<maximed>jeko: So, if you move the .git outside the repo, then it works, but if you put it back, it doesn't?
<maximed>(fwiw, geiser isn't Guix, maybe it has its own IRC channel, maybe #guile?)
<maximed>(#guile needs registration)
<apteryx>jeko: the version of git was updated; not sure if that could have anything to do with your problem
<jeko>maximed: Oh if I move the .git repo outside I still get the issue… Maybe it's related to Projectile extension because I get the error when I open a file inside my ~/Workspace (which is my projects root search path). When I open /tmp/test.scm it's OK I can run Geiser.
<jeko>apteryx: I can see a trace in Emacs'messages :
<jeko>Unable to activate package ‘magit’.
<jeko>Required package ‘git-commit-3.3.0’ is unavailable
<jeko>Unable to activate package ‘hydra’.
<jeko>Required package ‘lv-0’ is unavailable
<jeko>Unable to activate package ‘use-package’.
<jeko>Required package ‘bind-key-2.4’ is unavailable
<jeko>I'm gonna try to remove extension one by one and see the one causing troubles
<apteryx>perhaps magit / libgit is not compatible with the latest git bump
<jeko>I will start by this one haha
<jeko>with this one*
<jeko>i mean without
<maximed>antioxidant-build-system update: rust-pico-sys@0.1 now build with antioxidant
<kitzman>is it expected that gnu/system/hurd.scm:122 is `(initrd-modules (lambda _ '())`? guix system image works fine with it, but guix deploy complains (maybe i shouldn't attempt to overwrite it?)
<maximed>kitzman: FWIW, I don't think that the Hurd even has initrd-modules
<maximed>maybe "guix deploy" uses it even when it shouldn't? (wild guessing)
<kitzman>yes; guix deploy works fine if it's an empty list.
<maximed>kitzan: Do you mean that the (lambda _) should be dropped, just keeping (initrd-modules '())
<kitzman>yes. from other systems it seems like they're using lists.
<civodul>maximed: re antioxidant, yay!
<civodul>kitzman, maximed: initrd-modules should definitely be a list (empty list here)
<civodul>i can fix as you suggest
<Guest28>Ths should have said GuixSD:
<maximed>civodul: Now getting rust-tokio to build ...
<maximed>error occurred: Failed to find tool. Is `cc` installed?
<civodul>Guest28: "Guix System", even :-)
<anadon>The Agda package is broken and looks like there are three ways to get sources and three yet different ways to get the  stdlib which for some reason isn't bundled.  They also require new environmental variables be set.  Where do I even start?  The apparent package to add something resembling a stdlib, `agda-ial`, just straight up doesn't build.
<anadon>I can't learn and fix all the tools all at once.  I need to work on meson right now.
<kitzman>civodul: if possible; if easier/less time consuming i can open a bug
<maximed>antioxidant-build-system update: rust-generator@0.6.20 and dependencies now build with antioxidant
<civodul>kitzman: i'll do it, it's easier in this case :-)
<maximed>antioxidant-build-system update: rust-tokio@1.8.0 and dependencies now build with antioxidant
***caleb__ is now known as KE0VVT
<maximed>antioxidant-build-system update: rust-cloudflare-zlib-sys@0.3.0 and dependencies now build with antioxidant
<maximed>antioxidant-build-system update: rust-miniz-sys and dependencies now build with antioxidant
<civodul>now there are 3K Rust packages so... :-)
<maximed>civodul: Most packages don't need individual changes.
<maximed>E.g., at a time, antioxidant did not support enabling features, and no features were disabled by default. Now they are on by default, which makes a lot more packages build
<maximed>oops, mistyped a feature name, now I've to rebuild 50 dependents ...
<maximed>antioxidant-build-system update: rust-nix and dependencies now build with antioxidant
<maximed>(that's not nix the package manager or daemon)
<maximed>What I'm doing, is trying to make the first thing from (gnu packages rust-apps) build, then the second, and so on
<maximed>wait no, I'm doing them out-of-order actually
<maximed>The app 'hexyl' can be build already (unless it's broken again), now trying 'sniffglue'
<anadon>"sniffglue"?  Well that's a package name...
<maximed>... I cannot recommend talking about that package in front of health professionals or the police
<maximed>there's some room for confusion there
***mark_ is now known as mjw
<maximed>antioxidant-build-system update: rust-http-body and dependencies now build with antioxidant
<rekado>apteryx: on the 12th we will have the SAN extension installed. This may lead to a short downtime of the SAN if things go wrong — and since we aren’t using it via multipath I expect things to go wrong.
<rekado>between 10am and 12pm Berlin time it may be safer to shut down the head node of the build farm.
<rekado>is this feasible?
<mroh>Perhaps, we need the old python-pyyaml-5 for docker-compose (which doesn't build currently).
<maximed>antioxidant-build-system update: rust-h2 and dependencies (including a more recent tokio) now build with antioxidant
<kitzman>are derivations architecture-dependent?
<maximed>kitzman: Yes, they depend on --system and --target
<kitzman>hm. but for example, substitutable derivations are only target dependent, right?
<civodul>you can run "cat /gnu/store/...-foo.drv" to see those pieces of info