<buenouanq>when might the next version release be coming up? <pkill9>what's the difference between /run/booted-system and /run/current-system? ***enick_345 is now known as sturm
<sturm>If I'm looking to run a `curl URL | bash` type installer in a container but getting `curl: (60) server certificate verification failed. CAfile: none CRLfile: none`, what package might I be missing? Container is run like this `guix environment --ad-hoc wget --container --network --share=$HOME/Downloads` <sturm>sorry, replace "wget" with "curl" in the above container command <sturm>Ah, adding `--expose=/etc/ssl/certs` does the trick to share the host's certs. I wonder still though if you can install these directly in the container. Adding "nss-certs" as an ad-hoc depedency doesn't seem to do it. <lfam>pkill9: The booted-system is the system that was booted. If you then reconfigure, I'd expect the current-system will have changed (have not read the code) <lfam>sneek: later tell pkill9: The booted-system is the system that was booted. If you then reconfigure, I'd expect the current-system will have changed (have not read the code) <lfam>The latest OpenSSH fails its tests, although I think the failure is basically a test bug (the test tries to create a file without first creating its parent directory) <lfam>sneek: later tell sohomb83: In Mutt, start a message in the normal way. After writing your message and leaving the editor, but before sending, press 'a' to attach files <lfam>sohomb83: In Mutt, start a message in the normal way. After writing your message and leaving the editor, but before sending, press 'a' to attach files <sohomb83>lfam: a friend suggested me to try out git send-email. I think i will try it once. :-) (thank you though!! ) <lfam>sohomb83: Sure, both methods work :) <lo_mlatu>Ask questions again...I now have compiled an extra kernel module, after defining a package for it, how can I configure to load that module(the .ko file) at boot time <lo_mlatu>the `initrd-modules` field seems can only detect the modules included in the kernel, `initrd` seems can do the job but...is there an easy way? rather than rewrite the whole gexp <grillon>yesturday guix system reconfigure stopped with gnome-desktop needed but not installed. So I installed it(as root) and restart guix system reconfigure. then because I'm not alone to use the computer I had to restart fully working <grillon>I think the best for me is starting in a vm. Is there a command to do p to v? I have read it's easy to reproduce a system with guix. <civodul>lo_mlatu: you can use --skip-checks if needed <civodul>that is, if you think 'guix system' makes a wrong diagnostic regarding modules, then use 'guix system reconfigure --skip-checks' <civodul>grillon: could you explain exactly what you typed and what you got? <civodul>it may be easier to do that on help-guix@gnu.org though <lo_mlatu>civodul: Sorry maybe I didn't explain it well...`initrd-modules` seems to accept a list of strings which represents the filename of the modules, during system reconfiguration it will try to find corresponding .ko files in the kernel directory( I guess, didn't fully check the source code) <lo_mlatu>Now I have a module file which is "out of kernel", how can I make a configure so the build process can find it? <civodul>the relevant code is in (gnu system linux-initrd) <civodul>as you can see, it assumes that modules come from the kernel package <lo_mlatu>yes, and didn't find a option to extend its search path <civodul>we would need to change that to, for instance, allow users to specify a list of packages to search modules in <lo_mlatu>I see the `kodir` defined in the raw-initrd, does it specify the "module search path"? <civodul>it's the 'linux' parameter passed to flat-linux-module-directory that determines where modules are taken frmo <civodul>it's not even a "search path" since everything is taken from the same place <civodul>yes, that's what we would need to do <zybell_>how about patching the kernel with search path symlinks. I mean symlinks which encode multiple filenames and refer to the first file for which they are not broken. <zybell_>that would be more generally usable. <civodul>lo_mlatu: to be clear, is it a module needed in the initrd, or can it be loaded later? <lo_mlatu>well..it can be loaded later, but without the module my keyboard and touchpad won't work, so I want to load it at boot time <lo_mlatu>initrd is the closest way I find to do that in guixsd....maybe there is a easier way? <civodul>lo_mlatu: that makes sense, yes (an out-of-tree module for a keyboard? weird) <civodul>so yes, i'd suggest changing flat-linux-module-directory so it can search for modules in several directories, not just one <lo_mlatu>zybell_: sorry I'm not familiar with guix, can `activation-service-type` achieve this? <civodul>that runs after the root file system has been mounted, so not exactly what you want <lo_mlatu>civodul: well, I'm using MacBook which has an spi keyboard... <zybell_>'mkdir -p /etc/sv/mymod;echo "modprobe mymodule;sleep 5">/etc/sv/mymod/run' <zybell_>and ln -s /etc/sv/mymod /var/service <zybell_>for systemd you need a run once unit. ***Piece_Maker is now known as Acou_Bass
<grillon>sure I'll contact help-guix@gnu.org to continue because my trying time window is too small for a chat. thank you for your help <civodul>lo_mlatu: i think you have networking issues :-) ***Piece_Maker is now known as Acou_Bass
***pareidolia_ is now known as pareidolia
***erdic_ is now known as erdic
***exio is now known as exio4
***MinceR_ is now known as MinceR
***ocharles_ is now known as ocharles
<pkill9>hmm, disk-image isn't persistent on with USB disk <sneek>pkill9, lfam says: The booted-system is the system that was booted. If you then reconfigure, I'd expect the current-system will have changed (have not read the code) <thomassgn>mbakke: bavier` Have just into the 4.14.32 kernel and are having the exact same problems with the atheros wifi card. I saw linux libre had a release message the other day with something about wireless cards(think it was ralink, but maybe...), doing a guix pull now and will be building the 4.16 kernel to see if it helps... <thorwil>civodul: hi! if (with-imported-modules '((guix build utils)) is necessary in x11-socket-directory-service, then what happened here that it seemed to work without? <sneek>Welcome back thorwil, you have 1 message. <lo_mlatu>Hi it's me again...I have installed `linux-libre-headers`, but I didn't find header files such as `modules.h` or `platform_device.h` which is needed for compiling a module..can any one help me? <pkill9_>does anyone have a GuixSD configuration for their desktop I could see? ***pkill9_ is now known as pkill9
<thorwil>pkill9: are you after something specific <pkill9>nah, I just want to check if I'm missing anything ***tilpner_ is now known as tilpner
<efraim>Yeah, thats how I do my manifest, makes it much,easier <thorwil>pkill9: it's odd to see that outside of a manifest, though <thorwil>pkill9: here, it's just (packages (cons* foo bar baz %base-packages)) <pkill9>thorwil: do you need to put (use-module) for all the packages you specify there? <thorwil>but yeah, i had to figure out some use-modules <pkill9>what does 'cons' and ';cons*' mean? <bavier`>pkill9: prepends something to the head of a list <bavier`>pkill9: cons* accepts multiple items, e.g. (cons* 1 2 3 (list 4 5 6)) -> '(1 2 3 4 5 6) <thomassgn>So linux-libre 4.16-rc7 has a few new dependencies, and fails to build here. Just stops building (during compilation (CC and AR going on). Trying 4.16 now <thomassgn>haha, I just realised also, rc7 is release candidate so linux-libre-4.16 is higher version than 4.16-rc7, isn't it? <lo_mlatu>Sorry for bothering..I have installed `linux-libre-headers`, but I didn't find header files such as `modules.h` or `platform_device.h` which is needed for compiling a kernel module, so how can I get these files? <zybell_>I had this problem years ago on another dist. A header was not in the headers-pkg but of course in the full source. <lo_mlatu>zybell_: that should works, I'll check that <thomassgn>And linux-libre-4.16 fails at what looks like the same place... wat? <thomassgn>civodul: Thanks for feedback on the libreoffice patch. I'm looking at it now. <thomassgn>and what was the trick to make the po files not be shown as untracked? checkout a branch from upstream? <pkill9>anyone using GuixSD in their server? <ng0>do I have an déjà vu or did we someone else make an april fools in the last 3 years which backfired and confused people in a similar way than the systemd one? <pkill9>can GRUB be configured to use MBR partition system? <pkill9>since i'd rather avoid wiping my harddrive to convert it to GPT <ng0>grub works with mbr and gpt <thomassgn>you don't really specify mbr. You more or less don't specify efi and your good. <thomassgn>(not sure if the efi grub requires gpt, I would assume so) <thorwil>pkill9: i thought it was written out somewhere, but can't find it ... i mean to recall that you need grub-bootloader and NOT grub-efi-bootloader <thomassgn>You don't specify mbr or gpt in relation to your bootloader <thomassgn>some of the schemes (like some efi (maybe all efi)) requires gpt. <thomassgn>I get "guix build: warning: ambiguous package specification `linux-libre@4.16'" from 'guix build -K linux-libre@4.16' and then it chooses linux-libre-4.16-rc7... Why? <thomassgn>And how can I make it choose what I write instead? <zybell_>efi needs gpt,because it finds the bootpartition by a special gptID. gpt doesnt exclude partitioning with mbr. A disk can be partitioned with both. Mostly its done by entering the whole gptdisk into the mbr as one strange/alien partition, so that BI/OS that don't understand gpt don't touch. But with care one can enter the bootpart into mbr too. Because its formatted as FAT syslinux should work. <jonsger[m]>is it intended that the qemu image doesn't have "mkfs.fat32"? <pkill9>does the GuixSD installation image come with an installation guide in a text file? <mbakke>pkill9: In the installation image, the manual is available on one of the VT's. Try Ctrl-Alt-F2 (and so on). <jonsger[m]>ACTION has after a long time running GuixSD again :) <bavier`>pkill9: cool, let us know how it goes <jonsger[m]>disk partitioning is still the main pain for me, the rest is pretty easy (<-- give in that command, then that...)