IRC channel logs
2024-02-26.log
back to list of logs
<efraim>berlin offload machine .161 isn't offloading to the childhurd correctly :/ <efraim>huh, the symlink from /etc/guix/machines.scm points to a file that no longer exists <nckx>sneek: later tell dodoyada You'd check with 'lsmod | grep librem'. Also, define 'installing', since I can think of at least two ways, of which I'd expect only one to have any effect (kernel-loadable-modules). <efraim>civodul: I'm looking at the guile-git test failure on powerpc-linux and I'm wondering if it's related to one of the set! functions which have been giving me problems on powerpc-linux for a while <civodul>at first sight i don’t see how this could be related <efraim>I used display and tested against x86_64 first, (blob-content blob) doesn't return anything on powerpc-linux <efraim>#vu8(72 195 169 108 108 111 32 87 111 114 108 100 10) <ayatsfer>is gnu/packages/crates-io.scm manually maintained or is there any automation to include new crates? <efraim>there is automation for importing new packages and some for updating them, but its generally manually maintained <ayatsfer>ah, and which is the automation for importing? <efraim>although we're about to merge the rust-team branch, so we'll have a bunch of new and updated pacakges soon <ayatsfer>I need to find some example for a crate with a workspace with multiple members... <civodul>BTW, ‘guix import’ has a new ‘--insert’ option that may be helpful in managing things like ‘crates-io.scm’ <mrvdb>What do I need to do in emacs to get 'C-c C-d d' (doc symbol at point) to work? Just set geiser-load-path is not sufficient apparently. <ayatsfer>$ make && ./pre-inst-env guix import --insert crate spinning <ayatsfer>guix import: error: spinning: invalid importer <rekado>ayatsfer: no, you're using "guix import" wrong :) <rekado>it expects the importer's name as its first argument <rekado>e.g."guix import cran" for the CRAN package importer <rekado>this is then followed by importer-specific arguments <hako>ayatsfer: `--insert FILE` <hako>guix import --insert example.scm crate ... <efraim>I'm going to assume that git_blob_rawsize in libgit2 itself works fine on powerpc, but I might have to write something to check that. If so that would leave the FFI for that function <efraim>I probably should check it then, since the other functions seem to work fine. I would've assumed the testsuite for libgit2 would've found something <rekado>ayatsfer: oops, sorry for confusing you with my response. <jpoiret>mrvdb: guile variables don't have source location attached to them unfotunately, only procedures <Kabouik>Do we have an alternative to Barrier/Synergy for Wayland in our packages? <reyman>I finally found some information about nfs-service-type, but i don't know where i need to create the mountpoint (ex "/srv/myserver") needed by mount.nfs in my config. <reyman>when i restart nfs i have : [exportfs] exportfs: Failed to stat /srv/myserver: No such file or directory <jpoiret>reyman: /srv/myserver/ doesn't exist on your system I presume? <reyman>jpoiret: i need to create it at startup as a bind mount ? <jpoiret>then you'll need to create it. It's the same for all filesystems on guix, we don't create the mountpoints automatically iirc <reyman>that's not clear in the doc because "file-system" need a "device" argument. My mountpoint is not linked to any source, this is only a folder. <efraim>is the parent directory there? would it be enough to have /srv? <reyman>perhaps i need to mount /srv as bindmount for ext4 , because my server disk accessible on nfs are ext4 <reyman>hum finally i will simply mount remote nfs on a folder in my home/, this is more secure, i don't want to crash system using bad options in "(file-systems ... " part <nutcase>Hi guix, I created a personal channel. It contains a .guix-authorizations file with the fingerprint of my gpg key I used to sign all commits and which is contained as .key file in the keyring branch of the channels repository. For channel introduction, I use the commit/gpg-fingerprint I used to commit the .guix-authorizations and gpg-sign that commit. If I run guix git authentication commit/signer combination on a later commit, guix <nutcase>complains that the younger commits have a sha1 signature, which is not permitted. <nutcase>what should I check to find out, what's going wrong? <nutcase>when I introduce my channel with the latest commit ID, guix considers the repo / commit as authenticated. It seems that the authentication chain is not traversed down to the .guix-authoriations introducing commit. <reyman>ok i understand my problem, i'm silly, nfs-service is the server part, to export a folder as NFS ... what i need is more automount fs or something like that <reedm>Can someone explain how diskk uuids work for the guix filesystem configuration? currently i have my boot drive set up with uuids (as configured by the installer). I also have a second drive that I've just identified as /dev/sdb. It mounts fine and works as expected, but nothing shows up for it when I run blkid. <reedm>is it possible if i add a third drive, that the assignments of /dev/sdb, /dev/sdc, etc... could get mixed up? <efraim>I use `lsblk -fs` to find the UUID/label <efraim>and yes, it's possible they might come up in a different order <efraim>it's the whole drive? No partition table for the filesystem to sit on? I know that's possible but I'm not sure what to suggest for that <reedm>ahh, is that not a recommended configuration? I'm quite new to anything filesystem-related <reedm>I dont have any data on the drive yet, so if theres a better way, I am happy to partition/format now <ekaitz>it's like crt1.o is added to the compiler as is, not with the absolute path <efraim>reedm: sorry, I wander off regularly. I normally use gparted (GUI) or cfdisk (terminal UI) to add partitions to a disk and then use mkfs to add the filesystem <umanwizard>Hi, I sent two mails to guix-patches today and I have not received a confirmation. Can someone please check if there are issues? I am using `git send-email` for the first time, so maybe the issue is on my end and it is somehow configured wrong. But I _think_ it should have worked. <umanwizard>The patches are against the rust-team branch. The first mail should have the subject: [PATCH 1/2] gnu: Add rust-1.76. <umanwizard>I am not an expert guix contributor so please let me know if I'm doing something wrong. <apteryx>umanwizard: I also sometimes don't receive confirmation, but they appear on issues.guix.gnu.org usually just fine <umanwizard>The messages appear now, but for some reason they got assigned to different bugs (#69407 and #69408) despite being in the same patch series and being sent with one `git send-email` command ... how can I prevent this in the future? <euouae>jpoiret: FYI kexec works fine and microcode is not an issue AFAIK unless guix does something special that I'm not aware of <umanwizard>podiki: thank you. Not sure how I missed that before. <podiki>umanwizard: welcome. make sure you are always looking at the latest/"devel" manual, could be it wasn't in the older version (1.4.0; confusing but consider that just the installer version) <lilyp>We should make debbugs recognize kthxbye <podiki>lilyp: how is gnome-team looking? <lilyp>you ought to see the mail soon <podiki>i was updating a package locally and saw that i updated libhandy which was done on gnome-team already, so i figured i'd wait <podiki>(gtk-layer-shell and swaynotificationcenter) <reyman>i have a strange behavior with guix system compilation, at the install bootloader step, the process enter a wait loop and nothing happen <reyman>i try with more verbose compilation <reedm>is there a guix-y way to set owner/group permissions for a directory in the config.scm file? <sham1>Maybe a service to set the permissions, but I honestly couldn't tell you