<janneke>rain1: okay...possibly there's an easy solution for that <janneke>OriansJ``: have you seen the ELF_sym and ELF_str tables in my exit42 hex2? <janneke>shall we have the hex2 linker generate those? <dadinn_>janneke: i cannot see any /gnu/store/*-coreutils-*/ directory <Petter>dadinn_: not even with this? find /gnu/store -maxdepth 1 -type d -path '*-coreutils-*' <dadinn_>it is fresh download of 0.13.0, the tar does not have it <reepca>Question, which package could I get the mount(2) manpage and similar from? I've got glibc installed, but apparently those man pages aren't included with it. <janneke>reepca: l ~/.guix-profile/share/man/man2 <janneke>lrwxrwxrwx 8 root root 73 Jan 1 1970 /home/janneke/.guix-profile/share/man/man2 -> /gnu/store/l2wabl61mmwq70vjahm2cq92ggx7gkh3-man-pages-4.11/share/man/man2/ <lfam>reepca: mount(2) is the syscall. The executable mount(8) is from util-linux <lfam>Ah, and janneke answered too :) <OriansJ``>sneek: later tell janneke, I am unsure that the Hex2 is the right place for ELF symbol table generation but I am open to the idea <rekado>it requires users to preload the drmaa library provided by a cluster scheduler <rekado>usually preloading binaries from other systems does not work <rekado>I’d like to add an environment variable to disable the out-of-date warning. <rekado>our shared Guix installation is built from git and there is no ~/.config/guix/latest on the nodes <rekado>the warning actually confused people who then try to run “guix pull”, which fails in our environment. ***eiro is now known as odin
<rekado>Is there a website for “Guix Europe”? <rekado>I’m currently checking (again) with the institute if they could donate a couple of servers to either Guix Europe or the FSF. <koosha>After the installation , I get this : "grub-install: warning: Sector 33 is already in use by program 'FlexNet'; avoiding it. This software may cause boot or other problems in future. Please ask its authors not to store data in the boot track."" <rekado>koosha: it may not be a problem, but if you do have a problem booting this may be the cause. <koosha>rekado: I've just got Kernel panic <rekado>e.g. “failed to mount root device” or something like that? <koosha>rekado: It says : " [ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000000 " <civodul>rekado: there's no web site, just a blog post somewhere <civodul>rekado: re the out-of-date warning, there's already an env. var <rekado>there’s one, but it doesn’t help when there is no ~/.config/guix/latest <rekado>I only saw the variable to override the age <koosha>rekado: What do you think about the Kernel panic ? <rekado>koosha: I don’t know. That’s a very generic error message. <rekado>I just learned that our propeller-toolchain doesn’t produce usable binaries for the target :( <rekado>koosha: have there been any other errors during your installation? <koosha>rekado: No , It said that the installation finished without any error . <jlicht>the Xorg tearing fix allows me to use Gnome on GuixSD again without tearing my hair out :D <rekado>koosha: did you enable disk encryption? <rekado>or is this a barebones installation? <koosha>rekado: I used lightweight-desktop . <koosha>rekado: Maybe the warning caused the Kernel panic . <jlicht>Hmm, it seems our vpnc package is using the latest 'release' from Wed Nov 19 21:29:22 CET 2008 <jlicht>whereas to make it work with eduroam (afaik a european-wide network for students at most universities), we need a newer version <jlicht>or at least a more up-to-date version of the vpnc-script that is bundled with vpnc <jlicht>Should I just try to update the package definition to use an unreleased version of vpnc that bundles the right script? <rekado>the new NFS share is so much faster! <koosha>Any new idea about the my kernel panic problem ? <mekeor>rekado: do you use NFS to share data on multiple computers within your home-network? <mekeor>does NFS save the data on all computers, then? or how does it work? <rekado>it’s a central guix-daemon writing to a central /gnu, which sits on NFS <rekado>the NFS share is mounted read-only on all cluster nodes <mekeor>that sounds pretty nice. and where is the data actually stored? <mekeor>and where does NFS store the data? on one single computer? <mekeor>okay, anyway, i think, i understand it <jlicht>rekado: you wrote the guixr script to work with this setup as well, right? <rekado>but much of the script is no longer needed <rekado>we can now do GUIX_DAEMON_SOCKET=guix://somewhere-else:1234 <rekado>ACTION goes back to playing with NFS settings <civodul>rekado: FWIW i've reproduced the slowness locally simply by talking over localhost and adding a delay on the loopback device with tc(1) ***odin is now known as od1
<koosha>Why doesn't the OS have "sbin" directory ? <rekado>koosha: if a user installs a package that provides “sbin” it will be offered as part of their profile. <koosha>rekado: Could this kernel panic problem be a bug ? <rekado>koosha: in past case a kernel panic was usually caused by failure to mount the root file system. <rekado>koosha: this can happen when the operating system configuration is incorrect. <rekado>e.g. if the file system has a different label than the one used in the configuration <rekado>did you make any changes to the configuration? <efraim>For remote guix-daemon, can I point my x86_64 guix and an aarch64 daemon? <koosha>rekado: I don't think it could be because of the configuration file , I checked it a couple of times and it seems it's okay . <dadinn>rekado: following where I left yesterday, still trying to get to bootsrap guix no-strings attached, without relying too much on the live host environment I am in at the moment <rekado>dadinn: is there a reason why you cannot install GuixSD the usual way? <dadinn>and it seems to me that without coreutils I am kind of stuck <dadinn>because I won't be able to start up the daemon without the groups/users, to build it <rekado>I don’t understand. Can’t you just boot the USB image we provide and run guix system init? <dadinn>rekado: nope, it is not what I am doing. I am trying to do something similar how you would set up a system with debootsrap... kind of Archlinux-style <dadinn>I download and extract the guix binary tar <rekado>I recommend installing it the way it is meant to be installed, because that’s what we have most experience with. <dadinn>rekado: Ok, i might give a try. Does the SD image install method provide a way to install the kernel, bootloader (grub), etc, so at the end you have a bootable system? <civodul>efraim: clients can talk to any daemon, but the daemon may or may not be able to perform the requested builds <dadinn>civodul: I see. I think my problem is that I would need a more hands-on debootsrap/Arch like shell-based install method because my root is set up in a more exotic way, with home/var/swap residing on a ZFS pool and afaik GUIX won't support ZFS anytime soon <dadinn>civodul: and moving /var on a running system, after installation, is a PITA <dadinn>rekado: honestly, because of the above, I haven't even looked into GuixSD intall method, only experimented with a Debian hosted, and exploring from there if I could get it bootstraped without a Debian host <atw>I have not been following the mailing list. Is there progress on a maven build system/importer? I was exploring packaging leiningen before, but I think maven is a prerequisite. <civodul>but i don't think it anything has materialized yet <koosha>rekado: When I delete the "--load" swich in the grub editor be pressing the "e" button , it doesn't give me kernel panic anymore but it gets stuck at some second . <boskovits>I'm packaging a software that seems to belong to network.scm <boskovits>I'd like to know, where in the file should I put my definition? <rekado>boskovits: the end will be just fine <rekado>atw: you don’t need a maven-build-system <rekado>atw: it would be good to have one, but I’ve had great success with just the ant-build-system <rekado>it can also create build.xml files <sneek>From what I understand, the cake is a lie <koosha>God , this ditro doesn't want to be installed . <rekado>koosha: you haven’t yet posted your configuration file and the mention of a kernel panic alone is not enough to help you. <rekado>we need more information to help <koosha>rekado: Okay , I paste it in a couple of minutes ... <rekado>koosha: does your file system really have the label “my-root”? <jonsger>ACTION hopes guix pull will work so he can "work" with guix pack ^^ <rekado>koosha: in the GRUB shell, can you find a disk with that label? <rekado>koosha: do you use legacy BIOS or EFI? <civodul>jonsger: if it doesn't, that's a bug! <koosha>rekado: Have you read the message I wrote about the "--load" argument in grub ? <koosha>rekado: How can I get the disk's label in grub shell ? there are just : (hd0,msdos1) , (hd0,msdos2) and ... <jonsger>civodul: it didn't worked, but I installed guix new :) <civodul>jonsger: bummer; how did 'guix pull' fail? <sneek>janneke, you have 1 message. <sneek>janneke, OriansJ`` says: I am unsure that the Hex2 is the right place for ELF symbol table generation but I am open to the idea <janneke>OriansJ``: I'm not sure either...let's discuss it -- possibly later tonight/day <rain1>sorry i couldn't figure out what to do about the va list thing, thats a feature i never use in C & a bit more advanced than I know how to resolve <civodul>jonsger: could you send this log to bug-guix, along with the version of Guile being used and any other interesting piece of info you might find useful? <janneke>i didn't look yet myself, i'm preoccupied in fully merging hex2 with mes <efraim>Scaleway bumped their lower tier armv8 servers up and added some new top end ones <janneke>epronk tried (&asked me to help) to get GuixSD running on their 32 bit and later 64 offerings, which did not work -- segfaults because of missing/different math hardware <efraim>I tried 32 bit on scaleway and it didn't work, haven't tried 64 bit yet <efraim>The 64 bit might be worth looking into, I tried my best to make sure I didn't just add support for my boards :) <jonsger>guix pack: error: build failed: a `armhf-linux' is required to build `/gnu/store/8qyrjzvs5d3il5b8igwpcificxzmii3b-binutils-2.25.tar.xz.drv', but I am a `x86_64-linux' <jonsger>can't do guix pack cross-compilation? <quiliro>hello! good gnus! i could finally install in macbook air 2,1 from 2009 <quiliro>now i must install basic environment for hacking guixsd with emacs <quiliro>is there a model scm file for guix system reconfigure? <quiliro>for guix system install i used bare-bones.scm <quiliro>and my personal data (user, partition) <buenouanq>quiliro: I would love to talk about this more and get whatever we have to do easy and standardized and possibly automated. <buenouanq>Posting from a 2008 EFI Macbook running GuixSD. <buenouanq>Out of town right now though so it will have to wait a couple weeks. <quiliro>it just recognized the usb installer <quiliro>what i had problems was with was the config file <quiliro>buenouanq: where did you see tyhat message? <buenouanq>refind works great though I guess, I'd just rather be able to without <quiliro>buenouanq: what do you mean? you cannot boot with the new 0.13 usb installer? <quiliro>buenouanq: Posting from a 2008 EFI Macbook running GuixSD. <quiliro>another problem i had was installing directly to lightweight-desktop.scmqqq i had to install from bare-boes.scm <quiliro>buenouanq: have you tryed with 0.13? <quiliro>debian and parabola booted too in mine <buenouanq>it has nothing to do with the image, it's the laptop itself <buenouanq>if the firmware or whatever could be forcibly changed to allow this, I would be so happy <quiliro>probably the port or the usb are defective...i have had that problem <buenouanq>but as far as I know, I have to install refind or something similar <quiliro>how can you know which firmware you have <quiliro>i eliminated it...but i don't think that has anything to do with booting with the usb installer <buenouanq>idk man, I haven't used osx in many years, I only use it to install refind, then remove it when I have a working usable os installed <quiliro>buenouanq: can you sugest soething for me? <rekado>koosha: I think you can use “ls -l” to view additional details of the partitions <quiliro>will anyone paste some sample scm for using emacs in order to make new package descriptions? <quiliro>guix system reconfigure emacs-dev.scm <quiliro>or will that be better for the user only...not root? <koosha>rekado: Yes , it's label is "my-root" also in grub shell . <koosha>rekado: Do you want I give you the grub configuration file ? <ng0>did the change (if already applied) to fix profile collision introduce a new error? I just ran guix system reconfigure and have this: <ng0>guix system: error: profile contains conmflicting entries for gtk+:out <ng0>it's about gtk+2 and gtk+3 <ng0>and video-intel fails to be cloned from the upstream source. any messsage threads I missed about this? <jlicht>ng0: I had the same issues with my profile! <quiliro>where can i find a copy of lighweight-desktop.scm? <alezost>koosha: was your grub.cfg generated by guix or did you write it manually? <koosha>alezost: No , I didn't write it . <alezost>koosha: ah, so I think the problem is not in grub.cfg since it's the usual one <quiliro>ng0: i never had it...i only have bare-bones.scm <ng0>gnu/system/examples/ <reepca>alright, 3rd gpu arrived today... let's see how this works. <reepca>Well, the good news is that it seems this old nvidia card works. I can successfully use both of my 1600x1200 displays independently. The bad news is that once again, my ethernet interface was renamed to enp3s0 from enp2s0 and didn't automatically activate on startup. Also my third display, connected with a vga->hdmi adapter, isn't showing anything. <mekeor>quiliro: how do you mean your question "where is gnu/system/examples/ ?" <mekeor>quiliro: it's in the git-repository of guix <koosha>quiliro: After installing GuixSD (It finished without any error and just this warning : "Sector 32 is already in use by the program `FlexNet'; avoiding it. This software may cause boot or other problems in future. Please ask its authors not to store data in the boot track.") I got kernel panic : " Cannot boot because: Kernel panic - not syncing: Attempted to kill init! exitcode=0x00000000" . This is my system configuration file : htt <civodul>koosha: the mailing list may be more appropriate to investigate <civodul>perhaps you could send the details and config file to help-guix@gnu.org? <quiliro>koosha: i searched the net about that once it happened to me <quiliro>flexnet is some kind of backup software you can remove <quiliro>i would be thankful if you would report it to help-guix either if you solve it or not...i should have done that but i do not know now what my problem was or how to solve it..... i remember it was very easy to solve <quiliro>i remember it was something about zeroing that part of the disk by some method <quiliro>koosha: but you can easily find the solution by searching on duckduckgo.com <quiliro>loadkeys /gnu/store/..-kbd-2.0.4/share/keymaps/i386/dvorak/dvorak-es <quiliro>loadkeys /gnu/store/...-kbd-2.0.4/share/keymaps/i386/dvorak/dvorak-es.map.gz <reepca>agh, turns out the GeForce 8400 GS I got from newegg has a faulty HDMI port. Guess that's what I get for buying refurbished. Which means... sigh... on to the 4th graphics card. But hey, at least I can use my two larger screens in the meantime! <koosha>quasisane: So , you're saying the kernel panic is because of the flexnet ? <jonsger>reepca: it would be nice if some could reverse engineering the firmware of a more recent card :) <reepca>jonsger: I dunno how feasible that is. The only reason the older cards work is that the firmware is baked-in if I understand correctly. From what I can tell, firmware, especially without documentation, would be really, really rough to reverse-engineer. Throw on cryptography baked into the hardware and it seems pretty close to impossible. But I'd like to learn more about the process - perhaps next summer. <jonsger>reepca: Hawaii weights around 200kb :(