<mbakke>are there any interesting messages from sshd in /var/log/secure ? <quiliro>for ssh: authentication failure rhost:192.168.0.102 <mbakke>before that? anything about public keys or permissions? <quiliro>Permission denied, please try again. <quiliro>Permission denied, please try again. <mbakke>quiliro: can you try to "ssh -v 192.168.0.104" and verify that it tries to use the "id_rsa" key? <quiliro>debug1: identity file /home/quiliro/.ssh/id_rsa type 1 <quiliro>debug1: key_load_public: No such file or directory <quiliro>but /home/quiliro/.ssh/id_rsa exists <quiliro>debug1: Host '192.168.0.104' is known and matches the ECDSA host key. <quiliro>debug1: Found key in /home/quiliro/.ssh/known_hosts:1 <quiliro>debug1: Offering RSA public key: /home/quiliro/.ssh/id_rsa <quiliro>debug1: Authentications that can continue: publickey,password <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_dsa <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_ecdsa <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_ed25519 <quiliro>debug1: Next authentication method: password <wsddn>Quick question, should I file a bug report or something for the bug that jonsger and I have (make check for guix fails)? <mbakke>please also include the test-suite.log <mbakke>quiliro: it looks like it offers the rsa key, but it's not accepted by the server <mbakke>so the key itself should be fine <mbakke>strange that /var/log/secure does not mention the failed log-in attempt <mbakke>quiliro: oops, the sshd log is in /var/log/messages ! <quiliro>i was thinking that the key belongs to user quiliro and the user i am connecting to is root <mbakke>quiliro: no, they key should work on whatever user you add it to <quiliro>invalid user quiliro on other messages <quiliro>debug1: Next authentication method: publickey <quiliro>debug1: Offering RSA public key: /home/quiliro/.ssh/id_rsa <quiliro>debug1: Authentications that can continue: publickey,password <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_dsa <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_ecdsa <quiliro>debug1: Trying private key: /home/quiliro/.ssh/id_ed25519 <quiliro>debug1: Next authentication method: password <ng0>didn't follow the flood, but do you actually have the key in your .ssh/config in a block? <mbakke>quiliro: I can't reproduce it, but I see no messages when using a wrong key either <mbakke>quiliro: can you double check that /root/.ssh/authorized_keys looks correct? <ng0>see openssh pubkey configuration. But if you ask, I assume you already know this. I was just checking if this has been ignored <ng0>sneek: later tell catonano: I received your mail. I'll apply the patches tomorrow. <quiliro>ng0: i have no idea...just following instructions from mbakke <quiliro>but it _looks_ the same asauthorized_keys <ng0>then I hope marius instructed you right <mbakke>ng0: yes, we've verified that ssh tries to use ~/.ssh/id_rsa, but it's not accepted <mbakke>quiliro: did you copy id_rsa, or id_rsa.pub? <ng0>id_rsa.pub is what goes into auth_keys <ng0>id_rsa is the private key <ng0>ie the key of the keypair which needs to be in your local place <mbakke>quiliro: aaah, yes ng0 is right, sorry for the unclear instructions! <ng0>think of .pub as the key, and the id_rsa without .pub as the lock to the key <quiliro>oh...i was wondering why i didn't have to put my pub on th server and which was the private one!...i will copy pub then <ng0>no, other way around <ng0>correcting myself. you are right <quiliro>cool..np mbakke...better to learn more! <mbakke>quiliro: I learned something too, to read questions closely and to troubleshoot ssh access :D <quiliro>mbakke: great! thank you very much. it worked....you are a great troubleshooter along with ng0 <quiliro>i cannot use loadkeys to modify the keyboard to dvorak-es <quiliro>i have decided to write manuals from now on in Esperanto like scientists used to do in ancient times in latin so everyone could understand...that way...more people could learn a second language with more ease <quiliro>on the other hand, ssh-daemon is already up....the manual says to lift the daemon with herd start <mbakke>quiliro: is it because the loadkeys command is not available? <quiliro>mbakke: no...loadkeys is available locally on the target machine <quiliro>but it has no effect when running it remotely <quiliro>heck i am so happy because dvorak-es is included <mbakke>quiliro: does it not work to run that command before you ssh in? <quiliro>it is not possible for some reason in my debian <mbakke>and yeah, I noticed ssh-daemon was already loaded too, will try to fix it <quiliro>another thing is that on the previous usb install the hard disk was recognized as sda and the usdb as sdb so that eased installation because guix system init will use sdb for installation and that i think will conflict at boot because the hard disk will become sda <quiliro>besides, on the examples, the manual uses sda1 and label my-root for the / partition and sda2 for the swap partition <mbakke>quiliro: the manual recommends using labels to avoid that problem <mbakke>unfortunately, the guix boot code does not work with FAT labels yet <mbakke>yeah you have to use the drive paths that you'll have when the target system boots <mbakke>I've never seen a usb key get named before actual hard drives though <quiliro>what will happen when rebooting, it will not use the config.scm file to consider (bootloader (grub-configuration (device "/dev/sda"))) <quiliro>i notice bare-bones.scm changed from <mbakke>quiliro: ugh, you're right, for the first install you will need to use the current disk values in grub configuration <mbakke>quiliro: when you first run `guix system init`, the bootloader device must be the drive path that you have when the USB stick is booted <mbakke>but, once you boot the installed system, you may have to update the bootloader device <mbakke>...in the configuration... I can't sentences today.. <quiliro>will it boot?...if it will, do i have to guix system reconfigure after modifying bare-bone.scm? <mbakke>if you get no errors from `guix system init`, it will almost certainly boot <mbakke>but, once you have booted, you may have to update your configuration file before `guix system reconfigure` <mbakke>maybe we should use UUIDs for the bootloader configuration instead <mbakke>and definitely for the swap and FAT partitions <quiliro>on another issue...i have been installing with 0.12 for 3 days with installation giving errors....i checked and /tmp has the downloaded and compiled files...so now i want to continue. i assume i would have no problem. but i have decided to use another disk. how do i copy the files to the other disk? with cp? <quiliro>i assume i would have no problem using the same disk with 0.13 <quiliro>but how can i mirror the files with permissions and owners right <quiliro>the machine was a celeron d... i am changing to a core 2 duo <mbakke>quiliro: you can use `guix copy` to transfer store items from one machine to another <quiliro>both disks are in the same machine now <mbakke>e.g. `guix copy --from=192.168.3.100 /gnu/store/xyz...` <quiliro>one is in /mnt and the other in /mnt1 <mbakke>in that case, I'm not actually sure what to do <mbakke>quiliro: you can try, but I'm not sure if the guix-daemon database will find it <quiliro>something changes if the files are bit identical? <quiliro>because of the position on the drive? <mbakke>the daemon stores some information about the installed packages in /var/guix/db <mbakke>but I don't know much about it, it might work <mbakke>quiliro: the safest way is probably booting a virtual machine with one of the disks attached <mbakke>and use `guix copy --from=vm ...` <quiliro>it will take a lot of time for me because i have never used VM <quiliro>i will try cp...if it doesn't work i will try guix copy <mbakke>but, I think we have a lot more substitutes now than three days ago, so you shouldn't have to compile all that much <quiliro>root@gnu ~# cp --preserve=all --recursive /mnt1/etc /mnt <quiliro>root@gnu ~# cp --preserve=all --recursive /mnt1/tmp/ /mnt <mbakke>wait, I dont think that will work <mbakke>the daemon does not load anything from /tmp <mbakke>I thought you meant the /gnu/store items <quiliro>oh...but it did last time i shutdown and restarted installation <mbakke>you probably have "/tmp/guix-build-foo" folders? <mbakke>or what are you trying to copy, exactly? <mbakke>also, you should not modify /etc either <mbakke>that is generated from your system configuration <mbakke>what is '/tmp/guix-inst'? never seen that before <quiliro>yes but it downloads what is missing from there <quiliro>it is the /gnu/store directory before finishing installation <mbakke>aha, so you copied /gnu/store to /tmp/guix-inst <mbakke>the files in /etc/ will be overwritten by `guix system init/reconfigure` <mbakke>except things you've added manually <quiliro>i will 'herd start cow-store /mnt' now <quiliro>the files in /mnt/etc will be used by guix system init <mbakke>guix system init only takes a configuration file <mbakke>the configuration file can be anywhere <mbakke>so, /mnt is a new hard drive, correct? <mbakke>and you have the old guixsd installation at /oldmnt (say) <quiliro>yes...but it is not an old guix installation...it is an old incomplete installation <quiliro>because an incomplete installation has no /gnu/store and many other things <mbakke>if you transfer /gnu/store and /var/guix/db from the old system, that should be safe, I think <mbakke>anyway, I doubt the guix daemon will load anything from /tmp, that's a very unsafe thing to do! <mbakke>quiliro: I found the code that uses it <mbakke>there is a good chance it will work, but should be safe, anyway <mbakke>what are the contents of <mnt>/tmp/guix-inst, out of curiosity? <mbakke>is there only the ".rw-store" folder, and that only contains store items? <mbakke>it's only used in the init process of the installation image, the code that adds the`cow-store` herd service to be exact <quiliro>yes...i changed it in /mnt/etc/bare-bones.scm <quiliro>that was meant to answer what you mentioned long before <quiliro>mbakke: <mnt>/tmp/guix-inst contains the same files as /gnu/store on a running system <mbakke>quiliro: okay, thanks for confirming <mbakke>sorry for the confusion (again!) <mbakke>hope you have better luck with the installation this time :) <mbakke>I have many guixsd systems, so I can do the installation without a manual now :P <quiliro>i have 5 versions of glibc-utf8-locales-2.25.drv <quiliro>perhaps i will have problems with garbage collecting <mbakke>quiliro: that's normal, `guix gc` will clean up the entries that are not referenced from any profiles/gcroots <quiliro>how many running physical machines with guixsd do you have? <quiliro>mbakke: how can you stand the time for installation? you have another machine compile the and provide the binaries? <quiliro>i wish i could have something such as freedombone on guixsd <mbakke>one laptop with an oldish i5 CPU, one HTPC with old intel atom CPU, and one more powerful workstation <mbakke>I use offload or guix copy from the powerful systems <quiliro>perhaps i can run a virtual machine with parabola on top of guix to provide freedombone...check freedombone.net <quiliro>it is for generating binaries for other architectures <quiliro>maybe some raspberry pis.....oh...not possible with guixsd <quiliro>i wonder if beaglebone black is supported by guixsd <mbakke>not yet, but I think some are working on it <quiliro>i think i can ask a university to let me use one of its blades <mbakke>quiliro: that would be cool, but you need guix installed on it (doesn't have to be guixsd though) <mbakke>quiliro: AFAICT freedombone just generates a system image with some preinstalled software in it, correct? <mbakke>you can do the same with a guix configuration file as long as the system is supported <quiliro>mbakke: yes...it installs some software and uses one user to administer a la webmin <quiliro>how much time does the installation take on an older machine with offload using that heavyweight machine? <mbakke>quiliro: I have a 16-core server (virtual machine) from a previous job with guixsd :O <mbakke>and sometimes `guix copy` from it <mbakke>yeah, in exchange for answering questions over email when necessary :P <quiliro>mbakke: how much time will an installation take in average for barebones for you? <mbakke>quiliro: I always start with a minimal config, so no less than 30 minutes <mbakke>then set up desktop etc after I've booted into the guixsd installation <mbakke>lol, i meant "no more than".. :P <quiliro>how much time will it take to have atom up and running with desktop? <mbakke>quiliro: with substitutes, pretty fast.. but I also have a good internet connection <mbakke>no idea, lol.. just a few minutes with binary substitutes <mbakke>but it takes an hour to build the `guix pull` snapshot on it <mbakke>quiliro: IIRC you were unlucky and did not get any substitutes <mbakke>the situation should be much better now <mystified>Are there any dideo installation guides up anywhere ? <mystified>Are there any video installation guides up anywhere ? <quiliro>mystified: it is easier on text...video is to get a general idea <mystified>I tried a few times to install over the past 8 mnths or so.. <quiliro>you can even get better help on the mailing list help-guix@gnu.org <quiliro>mystified: that way more people can help even if it is not realtime <mystified>My opinion on Sysd, or dislike of it, is an understatement. <mystified>Thus experimenting with Distros without SysD <mystified>I'm no techie.. But I like getting my hands dirty :) <quiliro>the problem on installation is usually not paying attention to details in configuration file config.scm <quiliro>and using old machines makes installation very long <quiliro>your installation will be abreeze compared to mine <mystified>how ram is required running a window managr like xfce or openbox <quiliro>you can run any windo manager with 2 GB <mbakke>mystified: my HTPC running i3wm and kodi uses about 330M <quiliro>i have debian with xfce with a core 2 duo and it is fast <mystified>I'm running obrevenge at this moment. Like Trueos, Nutyx manjaro-openrc <mystified>If using a distro with SystemD then I prefer OB-Revenge] <quiliro>noobly: i am a newbie but would you please post the link to the mailing list inquiry? <noobly>quiliro: genode-main@lists.sourceforge.net. <quiliro>no...that is the address of the mail system...i am asking for the web page where i can read your report <quiliro>"Interesting. I wonder how much of a stretch it'd be to use a fork of NixOS, GuixSD (since it uses a Hurd microkernel as well), instead of NixOS. But thanks for the write up! I will certainly be following this soon" <quiliro>noobly: i don't understand why virtualize something on top of a virutalization <quiliro>phase `check' failed after 876.9 seconds <quiliro>builder for `/gnu/store/zprh9lkkxawavl72asmi2nn6ml95dbcc-guix-0.13.0.drv' failed with exit code 1 <quiliro>guix system: error: build failed: build of `/gnu/store/zprh9lkkxawavl72asmi2nn6ml95dbcc-guix-0.13.0.drv' failed <bms_>Hello. Does anyone know what the extent of 0.13.0's UEFI support is? For instance, will an old install automatically use UEFI on my iMac, or must I perform a fresh install? <bms_>Well, on a Mac, I think it is just an Apple-modified EFI, yes. <quiliro>i was told i could install guixsd with 0.13 <bms_>Yeah, I'm assuming fresh installs work, just wondering about updates for past installations. <bms_>I can currently boot into GuixSD, with a little GRUB2 scripting every time I boot. I know I could get around this if I really tried, but I'd rather make sure that I have to. <marusich>bms_, for additional information about it, you might want to check the version of the manual that is in the master branch. I think some commits which update the manual to document the new support were accidentally omitted from the release. <bms_>Alright, thanks, I'll check it. <noobly>quiliro: its for added security I believe, so the option is there <bms_>On Git, in doc/guix.texi. I browser-searched for UEFI. You need to import (gnu bootloaders) in your system .scm file. And add (grub grub-efi) to your Grub expression. <fusion809>Howdy folks has anyone had any success installing GuixSD on a VirtualBox VM? I saw the VM guide but it's for QEMU and I tried using VBoxManage to convert the USB image to a VDI. Booted it but there were failures that prevented it from booting properly. <j-r>fusion809: I did such many months ago. IIRC I just loaded the usb image onta a sata controller and booted from it. <j-r>I was unable to install the guest additions and then lost interest. <fusion809>Oh so you added it sort of like a disk image? <j-r>yes, like a normal disk image. <fusion809>Hey I'm presently trying to installing GuixSD to a VM the old fashion way by following the installation guide. It's going fairly well except one error I'm getting whenever I install a package and when I set the mirror, "guile: warning: failed to install locale <fusion809>warning: failed to install locale: invalid argument" <reepca>fusion809: are "glibc-utf8-locales" or "glibc-locales" installed for your current user (guix package -I will list them)? <fusion809>Hmm glibc-locales isn't installed the utf8 one is. <reepca>utf8 should work in most cases, what's the value of GUIX_LOCPATH environment variable? <fusion809>It's unset. Btw I've been wondering is GuixSD meant to be installed on a host OS or can be it installed as standalone OS? The install guide doesn't mention installing it to a blank root partition it seems to just be about installing to an existing Linux distro. <reepca>Ah, are you reading section 2 ("installation")? <reepca>Ah yeah, for GuixSD you want section 6, "GNU Distribution" <reepca>it's able to be installed both on a generic GNU/Linux system as a package manager and as its own distribution, so there are separate instructions <fusion809>Thanks. I must admit GuixSD is an interesting OS, 'tis unusual like NixOS. Kinda sad to hear that Ututo XS is practically dead as it's now Ubuntu-based and nothing like it used to be. <fusion809>There are several articles that GNU-Distribution article links to, must admit I'm having difficulty finding which one is for setting locales <fusion809>Or did ya mean that's the guide to install it to a new disk? <quiliro>fusion809: that is the guide for the distro <quiliro>fusion809: what you are installing is a pckage manager in another distro when you intall the binary <quiliro>that is why you have problem with locales <quiliro>fusion809: check the locales section on the guix manual if you still want just the package manager without the distro <quiliro>it should be available on the distro not on the package manager <reepca>Are you currently running the system image in your VM? <fusion809>Nope, I had some trouble with that so I thought I'd following the binary installation guide inside a VM for Arch Linux (I choose Arch Linux as it has systemd so following the service commands should be simple) <quiliro>guixsd (distro) should have herd....guix (package manager) shouldn't <reepca>fusion809: and you have a separate virtual disk thingy in your vm that you can install to from your arch linux image? <quiliro>so you use systemd instead of herd in a foreign distro <reepca>alright, then you could try installing from guix (package manager) on top of arch linux. The "guix system init" stuff is part of the package manager, you can run it from anything with guix (package manager) on it, the installation image is just nice because it's a predictable environment. <fusion809>Would it be good to ditch the Arch Linux environment and try with the installation image? I haven't tried using it as a SATA image like you suggested yet, so I might try that if you think it's a good idea. <brendyn>That's what I'm doing tonight. Desktop programs don't interoperate well when guix is on another distro <quiliro>brendyn: they should....what problems do you have? <brendyn>They don't find icons and have trouble mounting drives <brendyn>I've only got 3 patches in this last release. next time I want... 50! ***andreh1 is now known as andreh
<reepca>I don't know much about images and VMs and stuff - I know that I had success in putting GuixSD on a flash drive several months ago by just creating a file system on it, mounting it, and running "guix system init /mount/place/here myconfig.scm" <fusion809>The one good thing about VBox is it's easy to take screenshots. <fusion809>Is this a bug I should report at GuixSD's bug tracker? <fusion809>Are those errors deal-breakers, like will they prevent the installation? <reepca>¯\\_(ツ)_/¯ it looks like they're only problems with the other ttys. What happens if you press ctrl+alt+f2, for example? is that terminal dead? <fusion809>In VBox VMs it's right Ctrl + F2 (if I press Ctrl+Alt+F2 I'll change to tty2 on my host system) <reepca>Then I would guess it's not a deal breaker, but might be worth mentioning somewhere. Not sure where. <fusion809>Hey in the /mnt/etc/config.scm file the mapped-device section. I didn't encrypt my disk because it doesn't interest me. How do I deal with the mapped devices if my disk isn't encrypt. <reepca>I would guess that you just delete the mapped-devices expression. <reepca>It's not a required field (see 6.2.1 "Using the Configuration System", there's an example "barebones install" that doesn't use it) <reepca>Ah, it's used in listing the file-systems, in the (dependencies mapped-devices) expression. That's also optional, so delete that part of the file-system declaration <reepca>Just look at the first example, the "barebones" one, then add the stuff from the second example that you want <fusion809>I think I've followed it as I removed the mapped devices sec of the file-systems part of the config. Now guix init is giving: "guix system: error: build failed: getting info about the Nix store mount point: No such file or directory" <fusion809>Rofl I wanted the Vim editor in the USB image (just so it's there if I need to edit more files) so I ran guix package -i vim and it returned the same error as guix init... <reepca>guix package -i vim is giving a "guix system" error? Exactly the same as for the other one? <fusion809>Ya. Except without the guix system part of the error, replaced with guix package <fusion809>I'm gonna take a screenshot to show ya the output of both <fusion809>I was gonna suggest pastebining the config.scm in case you's can see an issue with it, but unfortunately there's no wgetpaste or pastebinit in this USB image and as guix isn't working for me it seems like that's not a possibility. <fusion809>I'm afraid that's not likely to work, running curl returns: "-bash: curl: command not found" <fusion809>nano works fine as does zile so it's odd curl and wget aren't available <reepca>Hm, unfortunate indeed. How about I pastebin what I'd try and you see if it's equivalent to what you've got? <fusion809>Seems identical to that, except, of course, updated for my own preferences (like sdX replaced with sda) <reepca>What's the value of NIX_STORE_DIR? <efraim>It can be blank, I believe its only used for 'guix import nix' <fusion809>Ya 'tis blank, sorry for the delay had to eat. <destt>I'm back with more errors making a package <destt>this time make is having trouble finding apr.h <destt>paste.lisp.org/display/347324 <destt>the paste has everything new and the error I'm getting <destt>Any clues on why apr.h is not being found? <quiliro>same problem this time "FAIL: tests/store.scm" <quiliro>in a few minutes i will report it to the mailing list <quiliro>i should look for the error log in the directory of the failed package <quiliro>is that directory /tmp/guix-build-guix-0.13.0.drv-0/source/ in that case? <quiliro>fusion809: i installed curl in the usb with guix package -i curl <fusion809>Seems you missed when I mentioned my errors when installing packages with guix package -i <PACKAGE>. It gives me "guix package: error: build failed: getting info about the Nix store mount point: No such file or directory" <reepca>Alright, so the file it's complaining doesn't exist is specified by processEnvironment() and directly by guix-register. processEnvironment() is only called by guix-register. It seems reasonable that the culprit is somewhere in guix-register. <quiliro>i wonder why you get that error and i don't <OCmylife>I have the same problem with my installation tests/store.scm fail <OCmylife>I'm trying to install it on bare metal (nvme) <reepca>the problem is that guix-register, despite being a separate program, is actually still just a node in a deep tree of abstraction. The fastest way to figure out which file it's complaining about would be to add a print statement and recompile or to break out a debugger... but for some reason I get the feeling if curl isn't installed, then neither is gdb. <destt>anybody have any idea why the package I linked earlier won't build? <OCmylife>test-name: verify-store + check-contents location: /tmp/guix-build-guix-0.13.0.drv-0/source/tests/store.scm:815 source: + (test-assert + "verify-store + check-contents" + (with-store + s + (let* ((text (random-text)) + (drv (build-expression->derivation + s + "corrupt" + `(let ((out (assoc-ref %outputs "out"))) + (call-with-output-file + <quiliro>will-never-be-found.drv what derivation is that? <reepca>apparently a poorly-named one... <Apteryx>It'd be nice if I could shadow a subset of packages in some gigantic module using GUIX_PACKAGE_PATH. But I think it completely replaces the module of the same reference. Right? <efraim>not ready for a bug report yet, but sablevm-classpath FTBFS on aarch64 <rekado_>quiliro GuixSD is NOT a fork of NixOS. <rekado_>(I know it’s not your comment, but it bears repeating anyway :)) <quiliro>i don't understand why you tell me that <quiliro>rekado_: what comment of mine implies guixsd is a fork of nixos? <quiliro>erased everything except bare-bones.scm <quiliro>b back after removing network card to insert usb memory <rekado_>quiliro: you pasted a comment above that included this statement. <quiliro>i forgot that they key was already copied in the usb memory <reepca>quiliro: it was a reply to noobly quoting a post he made <quiliro>rekado_: (23:13:10) fusion809: Thanks. I must admit GuixSD is an interesting OS, 'tis unusual like NixOS. Kinda sad to hear that Ututo XS is practically dead as it's now Ubuntu-based and nothing like it used to be. <rekado_><quiliro> [04:06:36] "Interesting. I wonder how much of a stretch it'd be to <rekado_> use a fork of NixOS, GuixSD (since it uses a Hurd microkernel as <rekado_> well), instead of NixOS. But thanks for the write up! I will <rekado_>I don’t want to get into a debate about attribution. I already wrote “(I know it’s not your comment, but it bears repeating anyway :))” <quiliro>rekado_: i was asking him if that was his post <Fox-_->I'd like to run "Commander S" to do scheme, using scsh <Fox-_->but there is no package and the install.scm is broken <Fox-_->cannot run guixSD 0.3 in qemu because video emulation FAIL. <quiliro>i have noticed a typo I have made in the installation <Fox-_->disk order of qemu start command ? <sneek>Welcome back civodul, you have 1 message. <sneek>civodul, rekado says: Should we start writing NEWS for the next release as we make commits (to avoid the painful process of writing it weeks later)? I just pushed the new Java bootstrap and think that could be mentioned there. <civodul>perhaps at the time of release we'll find that we have to trim it a bit, to keep the most important items, but that's fine <civodul>really happy about the new Java bootstrap! <civodul>do you think you could write about it? :-) <quiliro>i have a meeting tomorrow with the chancellor of a small public university....i would like to ask him to have his students volunteer to guix.....they do not know much about installing gnu....but they could learn....what do you recommend about this? <rekado_>civodul: yes, I’ll write about it soon. <quiliro>civodul: they could perhaps translate or test and report bugs <quiliro>i have the same failure: FAIL: tests/store.scm <jlicht>I am trying to add some custom udev rules to my GuixSD system, and thought I could do it by adding a simple `(simple-service 'some-rule udev-service-type (list <myrule-pkg>))` to my operating system declaration <jlicht>but it seems that my rules are not included at all :O <jlicht>is there any way in which I can debug this? <quiliro>actual-error: + (srfi-34 + #<condition &nix-protocol-error [message: "path `dtmp/guix-tests/store/462z3fnl7bs44vp9s97jyg1z74nsfvly-tar' is not in the Nix store" status: 1] efdae0>) <quiliro> test-name: verify-store + check-contents location: /tmp/guix-build-guix-0.13.0.drv-0/source/tests/store.scm:815 <quiliro>test-name: dead path can be explicitly collected location: /tmp/guix-build-guix-0.13.0.drv-0/source/tests/store.scm:178 <quiliro>exactly the same as the error I have already reported <user0984>I think this 0.13 I downloaded has issues. The test which checks 'store' is failing when Guix is build. <user0984>doing a second build right now to see ifi'm wrong <solene>user0984: I had a problem yesterday night when building guix to install 0.13 on amd64, a test failed but I wanted to try again to have logs <quiliro>check the updated bug report on the above link <user0984>yep, third run this time with what is included on the image, still failing <quiliro>could do it quicker with 0.13 because remotely installing makes it possible to chat, copy and paste errors and email them <user0984>quiliro,this update doesn't really help. i only have 1 harddisk and the installer and i know what i'm doing <user0984>impressive that this booksized bugreport made it through debbugs <quiliro>user0984: it should not be reported that way? <user0984>it should, it's just that i thought debbugs has a limit of size when textfiles get put on moderation approval (mailman i meant) <user0984>i will just watch the commit log until someone fixes this <fusion809>Hey folks I've noticed a flaw in the virtual machine documentation. For me running: qemu-img create -f qcow2 guixsd-usb-install-0.13.0.x86_64-linux.img 5G creates the qcow2 file and names it guixsd-usb-install-0.13.0.x86_64-linux.img thereby overwriting the USB image file. <fusion809>Plus on top of that the new guixsd-usb-install-0.13.0.x86_64-linux.img file is then only 196K in size, much smaller than the 1.1GB USB img file, making me wonder if it truly contains the required data to boot GuixSD <rekado_>fusion809: my copy of the manual lists this command: qemu-img create -f qcow2 guixsd.img 5G <rekado_>in step 3 you then add both this guixsd.img and the file you downloaded as drives when invoking qemu-system-x86_64 <fusion809>Ya but where do you get guixsd.img? Isn't the file you get from the xz-compressed USB image? If so I'm guessing the name of the file shouldn't make any diff <rekado_>you need a “virtual disk” and the image. <rekado_>qemu-img only creates the virtual disk. <fusion809>Must admit not used to using QEMU, I usually just use VBox but due to issues I mentioned earlier today on this channel I can't use VBox to run GuixSD effectively. <jlicht>Does anyone have any experience adding custom udev rules to GuixSD? <fusion809> -net user -net nic,model=virtio -boot menu=on \\ <fusion809> -drive file=guixsd-usb-install-0.13.0.x86_64-linux.img <fusion809>Ah I fixed it by reversing the order of drives. <fusion809>Might be worthwhile changing the docs this time unless I made an oversight (which is possible, as I said QEMU isn't a virtualization method I'm very used to) <fusion809>I suppose the docs do mention pressing F12 which does fix the issue, but still reversing the order might reduce the odds of someone making an error. <fusion809>But pressing Enter still seems to get me a console. <fusion809>Now internet is giving me some dramas. ipconfig -a reveals eth0 and lo as available connections. I ran `ifconfig eth0 up && dhclient -v eth0` to set the net up and then ran `ping -c 3 gnu.org` and it said 'ping: Unknown host`. <civodul>m-o, rekado_: you did mention that Cuirass has to be restarted, that it doesn't automatically poll repos, right? <civodul>at least that's what seems to happen on bayfront, which is ridiculous <m-o>well it polls the given git repo and store the last evaluated commit in database <m-o>each time a new commit is detected (!= from database sha1), a new evaluation is triggerexd <civodul>ok but it doesn't seem to notice new commits ***Acou_Bass is now known as Piece_Maker
<m-o>that's a new issue then :) <m-o>you have to remove database and restart if you want to *force* a new evaluation <m-o>(didn't found an easier way) <civodul>oh, so there *is* indeed a problem :-) <civodul>i think we've reached the point where it's less work to fix it than to keep restarting manually ;-) <m-o>for sure but I'm not sure how to notify cuirass that you want a new evaluation <civodul>it's supposed to poll and start a new evaluation when a new commit comes in <m-o>yes it does so (module bugs), the problem rekado and I were talking is when cuirass has seen new commit XXX, started and failed an evaluation and you want to reevaluate this commit XXX without removing database and restarting cuirass. Because, once cuirass has added XXX as the last evaluated commit in database, you have no way to tell it to restart evaluation on XXX. <solene>I can't get the usb image to boot on a Toshiba portégé laptop. The usb flash DEL shows that it tries do read, the screen lights up and display a "_" char and then goes black, and loop infinitely. I rewrited the usb disk with no better result. Did someone encountered this issue before ? <jlicht>rekado_: would you be so kind as to point me in the right direction (or the proper section of the manual) :-)? <rekado_>jlicht: I just modify the udev-service-type with a new udev-configuration that has a “rules” field listing all of my custom udev rules. <rekado_>jlicht: I don’t know the right section, I’m afraid <rekado_>jlicht: but I can show you the configuration I use <jlicht>rekado_: thanks a bunch, I will also pilfer some other parts of your config while I am at it <rekado_>civodul, m-o There are a couple of problems with Cuirass. What keeps me from relying on it here at the institute is that it stops doing things all the time. <rekado_>if anything fails it stops until the service is restarted <jlicht>rekado_: unless I am mistaken, looking at your config (and my observations), it seems that e.g. `(simple-service 'mtp udev-service-type (list libmtp))' in %desktop-services does not do anything currently. <rekado_>jlicht: oh, that could be a remnant of some work that never made it to master <jlicht>rekado_: it is in the current guix source tree, but your config and my own tests lead me to believe it is not effective at all <jlicht>rekado_: so your config is perfect no worries ;-) <roelj>Does anyone have any experience with CMake and missing "-pthreads" during its configure phase? <civodul>jlicht: it adds udev rules for mtp devices <civodul>i don't use that though so i don't know if it works as expected <civodul>roelj: no but hey, that's CMake! ;-) <j-r>Hello Guix. I've setup an offload build daemon. guix offload test works. However when I run guix package -i blah or guix package -u everything is built locally. How does the offload daemon get called during package builds? <civodul>j-r: the daemon looks for /etc/guix/machines.scm and calls the offload "hook" when it finds a matching machine <civodul>also, if you do not want to build anything on the local machine, you must run "guix-daemon --max-jobs=0" <jlicht>civodul: I would say so as well, but I cannot seem to find these udev rules in my system profile <civodul>jlicht: the udev rules are not in your system profile, but they're in a "udev-rules" item referred to by the udev process <davidl>Im trying to LUKS-encrypt my partition with the latest 0.13 image but I get an error: "CPU feature 'AVX registers' is not supported" and then when I check the partition with cryptsetup isLuks /dev/sda1 it returns false. <davidl>this wasn't the case with the 0.12 usb image. <j-r>civodul: I have /etc/guix/machines.scm configured properly (I think) guix offload test finds it and sends a test thing to it.... Anyway I add --max-jobs=0 to guix-daemon. Now I have a message that says either increase --max-jobs or enable distributed builds. I can find nothing in the manual about how to enable distributed builds. <jlicht>civodul: How would one find this udev-rules item? `ps' does not seem to help, and neither does herd status udev. <quiliro>it is great being mentioned in reddit <quiliro>it does not matter if someone speaks badly.....we are in the heads of people <quiliro>it is worse that no one has nothing to say about guix <rekado_>I resubmitted the release announcement to HN. <rekado_>hmm, in my attempt to import all of bioconductor I found that not all Bioconductor packages are actually hosted on Bioconductor infrastructure. <davidl>brendyn: I posted some links there to where to get free-software compatible hardware. <davexunit>rekado_: thanks for resubmitting the guix release announcement to HN. I got 9 upvotes yesterday but it never made it to the front page. <brendyn>Is it getting more now? It was on the 4th page pretty quickly the other day <rekado_>brendyn: it’s on the front page now (position 9) with 10 points. <brendyn>I guess I will post a little something <civodul>jlicht: guix gc -R $(readlink -f /run/current-system) | grep udev-rules <civodul>j-r: are you sure the 'system' field of the machine matches? <brendyn>What examples are there of Guix utilising featurse of scheme/lisp that aren't available in most other languages? <brendyn>ACTION wraps solene in parentheses and locks them in the cupboard <jlicht>brendyn: g-exps are quite magical <brendyn>I'm still only coming to understand those. I've never used them <ng___>qiliro: did you make any progress? <roelj>What's up with the "building of `/gnu/store/apkns3l6z8mgnjraci7pvk8zvj7hjspa-guix-0.13.0-1.a6d728b.drv' timed out after 3600 seconds of silence"? I can't build Guix with 1GB of memory? <ng___>I wonder what to do about 27032, looking at the uefi computer I wanted to set this up on <rekado_>brendyn: I once had a variant of the linux-libre package with the rt patches applied. <rekado_>it’s pretty easy to build this with Guix, but I found that I don’t need it, even for audio work. <roelj>ng___, civodul: Thanks for the pointers. <brendyn>rekado_: I'd like to use it as an example, do you still have it? <rekado_>brendyn: no, I don’t. But it wasn’t much more work than creating a variant of linux-libre (by inheriting), taking the patches, and adding them to the list of patches that ought to be applied to the sources. <ng___>did anyone have a succesful uefi install using the guixsd usb install medium yet? It can't just be 4 people with problems.. <ng___>just from what I experience here <ng___>report the report? or report in your report that your report is similar to 27032? <quiliro>report that my report is similar to 27032 <quiliro>and then post later reports on 27032 <ng___>do what you think you must do :) <ng___>you can simply write to the bug mentioned <ng___>To reply to this bug, email your comments to 27032 AT debbugs.gnu.org. <- in the webview <ng___>if you want the full traffic of all reported bugs, it is hooked to to bug-guix at gnu.org <quiliro>but i want all bugs to reach my email <ng___>then you can either use emacs debbugs or subscribe to the bug-guix@ list <ng___>it is relatively silent compared to guix-patches@ <fusion809>Is it just me or does Guix take forever to perform an update (with `guix pull && guix package -u`)? From the latest USB img I decided to do an update before installing Vim as running `guix package -i vim` suggested and it's still running after almost two hours. Is this usual or has it got something to do with the fact I'm running it on a USB img session? <ng___>now if only people would stop crossposting, I would be able to filter emails for once, but people can not be fixed. <quiliro>ng___: that is what i am asking: how to subscribe to the bug-guix@ list <fusion809>Ah thanks. Makes me feel better to know it wasn't my own screw up, granted what I had screwed up to cause it would escape me as all I've done is set up the net before running these guix commands. <rekado_>I wonder where to put all those bioconductor packages. <rekado_>the bioinformatics module is already ~ 10k lines long. <rekado_>with the packages I just imported it’s four times as large. <rekado_>maybe a new module called “gnu packages bioconductor”? <roelj>rekado_: That may be a good idea indeed. <roelj>How can I use a " in a substitute*? I escaped it like so: \\" but that doesn't seem to work.. <ng___>quiliro: the usual mailman way <quiliro>ng___: i do not find the mailman susbscription page <ng___>I find mlmmj easier to handle because I could just tell you message this address which is almost always canonical and do not open the webbrowser <rekado_>ng___ fusion809 “guix package -i” should not take long. <rekado_>that’s unrelated to “guix pull” taking up lots of memory <fusion809>Ya that's why I've decided to ditch `guix pull & guix package -u` <rekado_>roelj: and if that’s cmake you might be able to pass -DCMAKE_SAMTOOLS_DIR as a configure flag. <quiliro>rekado_: i thought you said he couldn't escape you. :-D <rekado_>roelj: one \\ for the (, another \\ for Guile. <roelj>Hmm. ERROR: Wrong type to apply: "set\\\\(SAMTOOLS_DIR \\"/usr/bin\\"\\\\)" <ng___>hm... fail again. I will just wait now. <ng___>thought deleting partitions and removing fragments could help. didn't <roelj>rekado_: I guess it really wants me to -DCMAKE_SAMTOOLS_DIR.. <rekado_>roelj: substitute* takes a file name first <roelj>Hmm.. Not escaping seems to work.. <rekado_>roelj: could you show us the actual code? <rekado_>you really do need to escape the parens <brendyn>Ok well I posted my little thing to HN <roelj>rekado_: Will my computer explode now? :O <roelj>rekado_: Ha! The substitute failed. :) <rekado_>this will match “setSAMTOOLS_DIR "/usr/bin"” <quiliro>"Paste number 347349: To escape, or not to escape. That is the question." :-D <roelj>rekado_: But.. double backslashes leads to "unexpected command: set\\(", and single backslashes lead to "illegal escape character #\\)". <rekado_>when do you get “unexpected command”? <rekado_>roelj: do not escape the ( in the replacement <efraim>I always check aria2 in bittorrent.scm to see how many backslashes I used before <roelj>rekado_: Yikes.. I hadn't thought of that. Thanks! <roelj>efraim: Good to know :) Thanks! <efraim>Does anyone know offhand how much RAM a librebooted macbook 2,1 can hold? <mbakke>efraim: 2,1! I used to have one of those with 4G, FWIW <quiliro>mbakke: but not the osne supported by libreboot <quiliro>you can check the model on the back if it is a macbook air <quiliro>i got mine (a1304 circa 2009) as a gift last month too <quiliro>i can't believe an 8 year old machine can run so well <quiliro>when everything is disposable nowadays <quiliro>only intel chipset works in libreboot <Apteryx>Can I use thunks in search-path-specification? Like... I'd like something to be formed dynamically there (EMACSLOADPATH="share/emacs/site-lisp:share/emacs/site-lisp/guix.d/ELPA-NAME-VERSION) --> The capitalized part depend on the package being looked at. <sirgazil>`guix pull` messages are looking tidy :) <Apteryx>If you ask me why: because the current discovery mechanism is limited to profiles only (guix-emacs.el looks for those folders in the current profile). This doesn't work to find other emacs dependencies at build time and would require me to fiddle with EMACSLoADPATH manually, which seems wrong. <sirgazil>Hmm, now I get failed: 504, "Gateway Time-out"... <lantw44>I am a maintainer of unofficial guix package on Arch User Repository and Fedora Copr. I found the systemd service files provided by guix 0.13.0 use /var/guix/profiles/per-user/root/guix-profile/bin/guix* in ExecStart=, but this path doesn't exist when guix package is installed by pacman or rpm. Should I patch it back to /usr/bin/guix*, or ask users to manually start guix-daemon and install guix with guix before <lantw44>they can use these systemd service files? <quiliro>i had some problem wit starting mariadb because of /var permissions too <roelj>lantw44: Is the /gnu/store included in those packages? In that case, the path to guix in the store would be the right way to substitute the faulty path with. <quiliro>i just created the directories as the guix user <ng___>catonano: I have some issues related to switching to uefi on my main computer, so I'll go through your patches as soon as I can do it again. <catonano>Ng0 ok, kein Problem. Thank you for your concern <sneek>Welcome back catonano, you have 1 message. <sneek>catonano, ng0 says: I received your mail. I'll apply the patches tomorrow. <catonano>Iantw44 i am interested in your Fedora copr <catonano>Gotta go. The phone battery is almost dry <lantw44>roelj: No, /gnu/store is empty, files in it are created by guix-daemon on the first start <civodul>lantw44: there were changes in the systemd files after the previous release; could you report the problem to bug-guix@gnu.org? <civodul>looks like we screwed one thing while fixing another one :-) <ng___>even when I'm not mentioned, it's cool that it found its way from overlay into upstream :) <roelj>lantw44: civodul: What do think about shipping the service file? Can it be fixed? <civodul>but i think the problem is that, depending on how you install Guix, the service file may contain the wrong file name <Apteryx>civodul: Do you think search-path-specification can be include some kind of lazy evaluated code for dynamically resolving a path? <Apteryx>I'd like to replace the guix-emacs.el mechanism by search-path-specifications, so that the elisp discovery mechanism works even at build time. Normally EMACSLOADPATH=share/emacs/site-lisp would have been enough (I think), but we are using these share/emacs/site-lisp/guix.d/ELPA-NAME-VERSION subdirs which complicate the task. <ng___>could a workaround for 27032 be to try with bios instead of uefi? not that I think that this test fails because of uefi, but you never know.. <civodul>Apteryx: no idea, that's a good question for guix-devel though :-) <civodul>it seems like i'm postponing all the discussions today, terrible :-) <Apteryx>civodul: no problem :) You must have a lot on your plate ;) <avoine>congratulation for the 0.13 release! <civodul>Apteryx: seriously though, alezost or Arun or Federico may have useful feedback <ng___>ACTION tries with bios just to exclude the impossible <civodul>ACTION caught the flu on top of that <lantw44>civodul: roelj: I sent a message to bug-guix@gnu.org ('systemd service files installed by guix 0.13.0 cannot used by other distributions <quiliro>ng___: it is not a uefi problem...i am using bios <quiliro>did you read the summary of my post? it is in one of the last messages <ng___>yes, but I have to agree with rekado it is very verbose and doesn't make the problem clear <ng___>ricardo's reply is after that <quiliro>message: "path `dtmp/guix-tests/store/462z3fnl7bs44vp9s97jyg1z74nsfvly-tar' is not in the Nix store" <rekado_>I was only commenting on message 17 (I always quote the relevant parts I’m replying to) <quiliro>rekado_: what do you think of message 20? <quiliro>it mentions two lines in /tmp/guix-build-guix-0.13.0.drv-0/source/tests/store.scm <quiliro>ng___: i can see line 815 and the next one say: <quiliro>(test-assert "verify-store + check-contents" <quiliro> ;; XXX: This test is I/O intensive. <quiliro>perhaps it is too intensive for a such an old machine? <catonano>and with the new confiiguraton it doesn't boot anymore <catonano>I could boot with thhe previous configuration but the onts in my terminal are mangled again <ng___>quiliro: I have ran this succesfully on an older GuixSD isntallation and the one where it is failing is an i7 ~4GHz 8+ GB Ram device <ng___>*older system with GuixSD, sorry <catonano>thank god it works wiith previious configurations ! <ng___>qiliro: i'll take a look when I have finished watching this utterly crap bad movie <quiliro>ng___: will you please try init instead of reconfigure on that powerful machine? <quiliro>and you still have the "FAIL: tests/store.scm" <ng___>I don't have anything ancient enough anymore to be i686 <quiliro>how about compiling for another arch....what is that called? <ng___>my last true 32 bit hardware went away in 2016 <ng___>doesn't work with init afaik <ng___>maybe I should just go from minimal Debian to GuixSD.. but 1) the error could still appear and 2) this doesn't make the make check failure disappear. <brendyn>Archlinux is beginning to phase out i686 support <brendyn>I didn't think it was all that old though <ng___>not that old, just not sold that much anymore where I am <quiliro>if you need something from debian you could run it in a vm <ng___>no. I meant to install GuixSD while this make check fails. <quiliro>;;; Failed to autoload make-page-map in (charting): <quiliro>;;; ERROR: missing interface for module (charting) <ng___>that's been around for a while <bavier`>quiliro: that message is harmless iirc <quiliro>;;; Failed to autoload read-pid-file in (shepherd service): <quiliro>;;; ERROR: missing interface for module (shepherd service) <quiliro>path `dtmp/guix-tests/store/462z3fnl7bs44vp9s97jyg1z74nsfvly-tar' is not in the Nix store <destt>having a problem using guile-sdl2 <destt>it's not appearing as an input <destt>e.g. `guix environment guile-sdl2` <destt>guile doesn't find it when you try and import it <destt>like so: `(use-modules (sdl2))` <mbakke>destt: try guix environment --ad-hoc guile-sdl2 <mbakke>without --ad-hoc, guix environment instead pulls in all dependencies of guile-sdl2 (e.g. for hacking on it) <destt>that doesn't seem to work either :/ <destt>but other packages (like guile-opengl) will work <davidl>building guix 0.13.0 fails the tests/store.scm test here when running guix system init. <destt>`guix environment --ad-hoc guile-opengl` then `guile` and finally `(use-modules (gl))` <davidl>quiliro: the tests/store.scm failed for me when I ran guix system init which started building guix. I think this is a bug. <SovereignBleak>Congratulations on the release everyone. BTRFS support is perfectly timed since I just switched to a BTRFS partitionless outlay with NixOS and Manjaro on their own subvolumes acting in place of partitions. Does GuixSD allow me to mount one of these subvolumes and install to it? Or will it balk at my zero partitions? <quiliro>trying 'guix system init /mnt/etc/bare-bones.scm /mnt --fallback' <SovereignBleak>By “these subvolumes” I mean creating a guixsd subvolume, mounting it then writing to it. <davidl>SovereignBleak: I tried this before and despite setting the btrfs-subvol as default mount-volume on my brtfs-partition that I ran the guix system init command on, Grub would mount the GuixSD root at /btrfs_sub/ <davidl>however, I think you can pass extra mount options in the system declaration, where you could possibly do something similar to mount -o subvolid=<my-subvol> <davidl>I mounted the subvol with mount -o subvolid at /mnt during install, ran btrfs set-default subvolid=257, remounted and installed to that but couldn't boot. <destt>I am getting the message `warning: failed to install locale: Invalid argument` with every guix command I run <reggggieee>but you have to install glibc-locales / glibc-utf8-locales as the user that's running guix commands <catonano>it's taking so long to compile the kernel :-/ <fusion809>Hey I've been running guix system init in a VM for the past several hours and I'm wondering if I close this VM and open it up later and run this same command will all the progress this command's made towards installing GuixSD to my virtual drive be lost? <buenouanq>I imagine so. Guix commands either complete successfully, or fail and do nothing. This is part of it"s functional nature. Being stateless or whatever, there is no `progress' as you are thinking about it. <buenouanq>I could totally be wrong though, I'm not really sure how any of this works. <davexunit>some sort of progress meter could be possible, but it is a bit tricky <davexunit>would be nice to know how far along you are, for sure. <fusion809>By progress I'm interested in knowing will all these hours be in vein? Like will the guix system init command take just as long to install the system as it would if I hadn't have spent these several hours running it? <davexunit>assuming that you're system is writing to disk <davexunit>and not a ramdisk or something that you would lose <buenouanq>I believe it will have to start over, but I could be wrong. <davexunit>any derivation that is successfully built will be cached on disk <davexunit>it's not all or nothing like that, that would be very bad. <buenouanq>This is just like any other distro vm install though - You haven't successfully installed it yet, what are you expecting to boot to? <fusion809>Yep I know I can't boot it, I'd still be booting the USB img <davexunit>in that guix it's not a guix problem because guix caches everything it builds, but a problem of using a volatile storage medium that will go away when the vm is shut down <davexunit>I recommend initializing the system with a minimal system config, booting into the new system, and then reconfiguring with a bigger config <fusion809>Well I'll play it safe. Just leave my PC on. Hopefully it'll be finished when I awaken. <jonsger>I'm now trying to install guixsd-0.12 with qemu to see if it works <davexunit>would highly recommend using that. 0.12 is 5 months old. <jonsger>davexunit: yes, but it fails several times during install, so I'll try 0.12 to see if it's a problem of my qemu setup <ng___>most strange. in an ad-hoc environment with guix the tests whic hare failing otherwise are succeeding now <destt>reggggieee, does this mean I should be installing nscd through my local package manager (apt)? <destt>nvm, just had to set GUIX_LOCPATH :P <zacts>congrats on the recent release <reggggieee>i just upgraded guix (on foreign distro) and now when I do guix commands, i get a huge number of "";;; ERROR: In procedure load-thunk-from-memory: Invalid or incomplete multibyte or wide character <reggggieee>;;; WARNING: loading compiled file <file> failed" <rekado_>reggggieee: that’s guile 2.0 trying to read guile 2.2 modules, I think <rekado_>reggggieee: are you using guix from git? <reggggieee>original installation was from the binary i think, but i just did a guix pull and upgraded a lot of programs <reggggieee>also, rekado_ ludo said you were probably the one to thank for a lot of the audio programs in guix - so thank you! <rekado_>there have been lots of contributors to the audio and music modules in Guix, but I certainly use a lot of them. <rekado_>reggggieee: “guix pull” should be enough. Maybe also upgrade the “guix” package itself. <reggggieee>ok, i'd just done both of those, which resulted in those error messages. i'll try again <rekado_>reggggieee: with the binary installation you set up guix for the root user. Don’t forget to update the root user’s guix as well. <catonano>I wrote on the help mailing list a ew minutes ago <reggggieee>catonano: ah, and updating with root user's guix doesn't help then? (i'm still in the process of pulling) <paroneayea>davexunit: I see you wrote the mkdtemp! procedure <paroneayea>I'm surprised this one hasn't been submitted upstream to Guile! <paroneayea>I'm trying to figure out how to get "lein repl" up and running in guix. I figured if I had clojure installed in my profile it would be there but it doesn't seem to e 
