2024-02-01 00:51:50 Generally, are packages supposed to create necessary service users? On a fresh Alpine 3.19, I `apk add loki && service loki start` and get "start-stop-daemon: user `loki' not found" 2024-02-01 01:01:42 strange, that package has no pre-install/post-install to create a loki user 2024-02-01 13:00:08 hm, are these massive number of join/quit messages considered as spam? 2024-02-01 13:01:08 I have them hidden, so I don't even notice it 2024-02-01 13:01:18 But is generally considered spam 2024-02-01 13:56:53 The libcrypto3 package seems to include some perl scripts? Anybody an idea where they belong? 2024-02-01 13:56:53 https://pkgs.alpinelinux.org/contents?name=libcrypto3&arch=x86_64 2024-02-01 13:58:05 tsget? 2024-02-01 13:58:48 Everything in the /etc/ssl/misc/ directory 2024-02-01 14:02:15 It seems to be part of so called 'apps': https://github.com/openssl/openssl/blob/master/apps/build.info#L58 2024-02-01 16:47:02 Hi, 2024-02-01 16:48:51 I dont know if this is the right place to ask, but I would like to ask a question about partitions and the alpine linux on m1 hardware. Anyway here goes the question. I installed alpine linux with the asahi installer from mps. Everything worked well. I have also run the setup-alpine script. 2024-02-01 16:51:03 My current problem is that setup-disk does not recognize the big data partion. Any ideas how I can start using the "big parition" if it does not show up in disk-setup? 2024-02-01 16:57:38 If I run df -h it tell me that there is no space available. If I run lsblk nvme0n1p5 has 180GB availbale 2024-02-01 16:59:00 check df -i 2024-02-01 17:00:25 df -i tells me, that 8GB (11%) are used, and that 78793 are available on the root partition 2024-02-01 17:55:33 I managed to expand the size shown by df -i to 34M but, cfdisk still shows, that the parition size if 158GB. 2024-02-01 18:15:24 if there's a /dev entry for the partition in question, most likely adding an appropriate entry to /etc/fstab will do the trick. disclaimer: I have not had a chance to try Apple's aarch64 hardware yet, so I might be completely wrong. 2024-02-01 18:16:34 bepeop-no-robot partition size and fs size arent hard linked, try to grow your fs 2024-02-01 18:22:12 Thanks will try after dinner :) 2024-02-01 20:16:22 lehm: please fix your client and poke somebody on /msg ChanServ access #alpine-linux list (or me) once you have 2024-02-01 20:22:34 I have tried the following to enlarge the fs size with resize2fs and cfdisk and fdisk. 2024-02-01 20:22:40 Here are the logs. 2024-02-01 20:22:43 https://privatebin.io/?2702443184d1141c#CpF2UEsKyf4a2aEQsjahbk8gcm3ApEvoB2LNzm5Gs4vG 2024-02-01 20:23:15 Sadly I messed up the ext4 partition, so I will start tomorrow from scratch. 2024-02-01 20:23:56 Is there any more documentaiton than this blog post: https://arvanta.net/alpine/install-alpine-m1/ 2024-02-01 20:26:30 and the https://arvanta.net/alpine/alpine-installer.mp4 video and script https://arvanta.net/asahi/aai.sh from the video. 2024-02-01 20:31:21 bepeop-no-robot: the -i flag to df shows *inodes* not filesystem space 2024-02-01 20:31:44 note the usage percentage is 0 2024-02-01 20:32:23 when coworkers want to resize lvms or partitions, I always point them at fsadm 2024-02-01 20:32:42 Ok, lemme check if I missread and actually already succeded 2024-02-01 20:33:38 if this actuallly works, I will be very happy and contribute some documentation. 2024-02-01 20:33:55 I have been running fedora last week and it is so bloated 2024-02-01 21:11:57 @dwfreed you where correct. resize2fs solved my problem. Thanks for helping everybody 2024-02-01 21:43:06 Hi! What's the best way to keep an old kernel along with a new version when there's an upgrade? 2024-02-01 21:44:02 Sure I can keep a copy of it (with modules and initramfs), but I was hoping to have an easier way for doing that 2024-02-01 21:49:24 hmm.. maybe kernel-hooks can help 2024-02-01 22:05:58 I have installed the mesa-asahi drivers, but even the foot terminal has input lag. should I use different packages than the ones here: https://privatebin.net/?db86c7942384bd38#6vf2d6SqmDJMncCQbLf2XcMjjjg9dS8qZyNhMKrxnxGv 2024-02-01 22:11:11 Also how can I install julia or juliaup. I tried install juliaup from rust, but this error occurs. Please note that I added the print statement, so you know which folder is missing. https://privatebin.io/?440c3e19537eed71#6nUFpjJcbuiqEPWbuZQ5qAPdhdv75SU7SaJsjeLuVseJ 2024-02-01 22:12:30 "No such file or directory" it's probably downloading glibc binaries 2024-02-01 22:13:38 bepeop-no-robot: sorry, I didn't had time to update docs 2024-02-01 22:15:31 I grepped the file file and it contains 2024-02-01 22:15:32 cat build-script-build | grep libc 2024-02-01 22:15:34 __libc_start_main 2024-02-01 22:15:35 libc.musl-aarch64.so.1 2024-02-01 22:16:39 @mps I can help, just tell me what to write. Based on my experience today I, what is missing is: Resize after first boot and which graphic drivers are missing. 2024-02-01 22:19:29 bepeop-no-robot: graphic drivers don't missing if you install mesa-asahi 2024-02-01 22:21:07 main problem is in that resize root FS is too risky, better give the expected size need + 20-30% 2024-02-01 22:22:11 @mps ok, good thing it worked for me. If you need help with anything ping me 2024-02-01 22:22:59 bepeop-no-robot: would be nice if someone take all this burden from me ;-) 2024-02-01 22:23:46 ok, I'm kidding, now it is not so hard as it was two years ago at beginning 2024-02-01 22:24:43 pigz decompression is not multithreaded? 2024-02-01 22:25:39 nope it isnt 2024-02-01 22:29:38 bepeop-no-robot: if you plan to start again to install it could make video of last steps after first reboot, this could be useful for new users 2024-02-01 22:34:14 bepeop-no-robot: about slow foot, latest mesa asahi works though yes, sometimes it is not as on 'normal' mesa 2024-02-01 22:34:32 but it works fine all in all 2024-02-01 22:36:34 @mps Sure. I can install again and make the video. 2024-02-01 22:38:43 bepeop-no-robot: on what model you install? 2024-02-01 23:00:58 m1 mac mini 2024-02-01 23:09:50 ah, J274AP code 2024-02-01 23:11:22 it should work fine with alpine 2024-02-01 23:11:31 What a lucky day, building julia from scratch worked just fine. Maybe I can ever contribute my first APKBUILD 2024-02-01 23:15:06 good night now 2024-02-01 23:19:09 good night and thanks for all the great work and support <3 2024-02-02 08:49:15 Hello, what is the default passwd for alpine standard? I tried test123. It is incorrect. 2024-02-02 08:50:52 just enter key, i.e. no passwd 2024-02-02 08:51:26 Thanks 2024-02-02 08:57:37 still incorrect. I tried that for alpine and foo 2024-02-02 08:58:31 user is root 2024-02-02 09:00:06 that works 2024-02-02 09:04:21 wiki has this msg but not be "marked" 2024-02-02 10:52:33 based on https://wiki.insteps.net/AlpineLinux/Create-what-is-only-needed-repo , i have curated pkgs for xfce4 2024-02-02 10:52:43 its still initial stage, will add docs to it soon -> https://git.insteps.net/mess/apklist/ 2024-02-02 10:53:56 planning to add some pkgs to make it a basic "rescue disk" 2024-02-02 15:31:39 Hi, I just tried to use iwd on a pretty empty alpine linux install and it complains about a ton of missing ciphers and cryptographic functions 2024-02-02 15:31:47 Could there be a missing dependency here? 2024-02-02 15:32:37 can you pastebin the errors you are getting, e.g. using https://tpaste.us/ ? 2024-02-02 15:33:14 will do! 2024-02-02 15:36:17 https://paste.debian.net/hidden/692339ae 2024-02-02 15:37:49 perhaps a missing kernel module? 2024-02-02 15:43:43 What kernel are you running? 2024-02-02 15:59:42 missing drivers in kernel 2024-02-02 16:04:35 For some time I seem to have a bug in tmux that when I attach, some weird characters are being sent to it 2024-02-02 16:04:51 One example is 00/0000eee/d1d1 2024-02-02 16:04:59 Has anyone else seen that? 2024-02-02 16:05:30 which terminal emultor and which TERM environmental variable do you have set before attaching? 2024-02-02 16:05:40 s/emultor/emulator 2024-02-02 16:06:33 I'm on urxvt, TERM=rxvt 2024-02-02 16:06:59 interesting, that seems to be reset via ssh 2024-02-02 16:07:11 So originally, locally I have: rxvt-unicode-256color 2024-02-02 16:07:32 Then I login to a server and I have only rxvt left 2024-02-02 16:10:53 not seeing that, and i do everything in tmux local and remote, but i use foot (waynesworld). 2024-02-02 16:14:24 I did have something like that some decades ago 2024-02-02 16:14:43 That I needed to reduce the length of the terminal, but that was a long long long ago hack 2024-02-02 16:19:54 kill me 2024-02-02 16:19:56 [17:19] sun:~% which ssh 2024-02-02 16:19:56 ssh: aliased to TERM=rxvt ssh 2024-02-02 16:20:08 That is a bugfix from 10-15y ago 2024-02-02 16:20:50 oops 2024-02-02 16:21:01 Interesting, the problem persists though with TERM=rxvt-unicode-256color 2024-02-02 16:22:05 are you setting default-terminal in tmux.conf, and if so, to what value? 2024-02-02 16:22:36 I did not modify tmux.conf anywhere so far 2024-02-02 16:23:21 what does $TERM say inside tmux? 2024-02-02 16:23:31 i actually don't know what it defaults to. 2024-02-02 16:23:36 tmux-256color 2024-02-02 16:23:41 mps: ikke: All I needed was a reboot, I figured that it failed to load the modules because I performed an apk upgrade to a newer kernel version before starting iwd :) 2024-02-02 16:23:45 okay, so, not that then 2024-02-02 16:23:59 the problem is it happens every time I attach, but only once for each attach 2024-02-02 16:24:16 And it also happens when I spawn a new session 2024-02-02 16:24:27 oh now that's interesting. 2024-02-02 16:24:30 And I just realise it happens locally as well as remote 2024-02-02 16:24:31 so that rules out ssh 2024-02-02 16:24:53 locally I just got the following pasted: 10;rgb:0000/0000/0000]11;rgb:eeee/eeee/d1d1 2024-02-02 16:25:55 are you sourcing anything scripts or anything for your prompt 2024-02-02 16:26:23 my hunch is it's your env not tmux 2024-02-02 16:26:40 I do have the following on most machines for setting PS1: PS1='[\t] [\033[1m]\h.ge.nau.so[\033[0m]:\w$ ' 2024-02-02 16:26:48 I can remove that for testing 2024-02-02 16:26:55 no that's fine 2024-02-02 16:27:34 Using export PS1="stupidps1%% " does not fix it 2024-02-02 16:28:27 And it hapens on "all hosts" (I've to find a non-alpine linux host for testing) 2024-02-02 16:28:52 ok, it does not happen on Debian 2024-02-02 16:29:41 tmux defaults on debian may not be upstream 2024-02-02 16:30:03 i set some stuff in tmux.conf but i didn't keep notes on specifically why. 2024-02-02 16:30:39 set -g default-terminal 'tmux-256color' 2024-02-02 16:30:41 set -sa terminal-overrides ',tmux-256color:RGB' 2024-02-02 16:30:43 set -ga terminal-overrides ',t 2024-02-02 16:32:26 ^ that last line pasted wrong 2024-02-02 16:32:48 should be ',tmux-256color:Tc' 2024-02-02 16:33:10 hmm, those three don't change anything for me 2024-02-02 16:34:36 they're probably default anyhow. 2024-02-02 16:35:17 i do this in many of my configs, explicitly set things so that i know when the behavior changes, rather than relying on defaults 2024-02-02 17:52:01 Hi, is there a difference between starting X with startx in tty and starting X from .profile ? Or it does exactly the same thing ? 2024-02-02 18:01:13 should essentially be the same 2024-02-02 18:01:58 polkit or something like that may not be happy about it, though 2024-02-02 18:02:49 but i think that it more about how it is started rather than from whence 2024-02-02 18:03:37 ok, thank you 2024-02-02 19:30:04 hi, I'm installing alpine on a new laptop. I tried to follow this guide for zfs 2024-02-02 19:30:38 (ups) with encrypted datasheets https://gmelikov.github.io/openzfs-docs/Getting%20Started/Alpine%20Linux/Root%20on%20ZFS.html and after reboot I get an error on grub 2024-02-02 19:31:05 It doesn't found /alpinelinux/root@/vmlinuz-lts (which seems a pretty weird path) 2024-02-02 19:35:44 this is full grub menuentry https://tpaste.us/lgeN 2024-02-02 19:41:45 donoban: you're trying to install Alpine in an unsupported fashion 2024-02-02 19:42:35 well, I have it running on other server 2024-02-02 19:42:51 I think that I followed the same guide 2024-02-02 19:43:07 and perhaps it doesn't' work now precisely because it's doing something that isn't supported? 2024-02-02 19:44:12 it's probably some minor thing :) 2024-02-02 19:44:35 just because it may have worked with a previous version of Alpine in the past does not mean it will necessarily work with newer Alpine versions 2024-02-02 19:45:04 just because it works to install Alpine does not mean it won't necessarily break in the future with Alpine package updates 2024-02-02 19:45:33 that procedure is doing several "non standard" things 2024-02-02 19:46:13 the "grub workarounds"? 2024-02-02 19:46:40 including modifying a template file packaged by Grub 2024-02-02 19:47:17 so then updated versions of that file won't be installed as part of Grub package updates 2024-02-02 19:47:23 I'm reading on archlinux wiki "Using EFISTUB 2024-02-02 19:47:42 could I use that to boot directly with a zfs boot partition? 2024-02-02 19:47:44 basically if you go down this route "you're on you own" 2024-02-02 19:47:58 use what? 2024-02-02 19:48:05 something like 2024-02-02 19:48:09 # efibootmgr --create --disk your_esp_disk --part your_esp_partition_number --label "Arch Linux (ZFS)" --loader /vmlinuz-linux --unicode 'zfs=zroot/ROOT/default rw initrd=\initramfs-linux.img' 2024-02-02 19:48:35 I'm gonna try to adapt it 2024-02-02 19:48:46 Alpine doesn't "officially" support EFISTUB booting, i.e. setup-disk doesn't support it 2024-02-02 19:48:58 so again, you'd be on your own 2024-02-02 19:49:20 is there some supported way for boot encrytped zfs? 2024-02-02 19:49:31 did you look at setup-disk? 2024-02-02 19:49:34 maybe using LUKS? 2024-02-02 19:49:40 uhm 2024-02-02 19:49:51 I know it supports ZFS and I thought it supported encrypted ZFS 2024-02-02 19:50:21 oh nice 2024-02-02 19:55:41 there's been enough problems recently with Grub on Alpine Edge with "official" install configs without trying to handle 1001 weird/random non-standard installs/configs 2024-02-02 19:56:09 yeah I was thinking that it will be instantly broken after uprading to edge 2024-02-02 19:57:21 that's my point 2024-02-02 19:59:42 uhm this looks interesting https://github.com/zbm-dev/zfsbootmenu?tab=readme-ov-file 2024-02-02 20:01:16 again not supports by setup-disk and I don't think it is even packaged for Alpine 2024-02-02 20:02:41 yes yes 2024-02-02 20:03:40 looking at setup-disk source, the process for install with zfs doesn't look straighforward 2024-02-02 20:04:08 compared to that "random" document you linked to earlier? ;-) 2024-02-02 20:05:47 I don't feel that it handles the zfs partitioning, pools/datahset creations etc... 2024-02-02 20:08:51 nope, once you move away from ext4 then setup-disk doesn't necessarily do everything for other "supported" filesystems 2024-02-02 20:09:10 well my currennt problem is that it doesn't find --set=root c8ca0d36b8e712fc , I will try to find the right id 2024-02-02 20:09:36 "lsblk" 2024-02-02 20:09:39 or "blkid" 2024-02-02 20:09:39 I have to go now, thanks for trying to help 2024-02-02 20:10:00 will check, it's powered off now 2024-02-02 20:10:06 see you 2024-02-02 21:27:28 I'm gonna try this https://docs.zfsbootmenu.org/en/v2.3.x/guides/alpine/uefi.html it loosk pretty clean and support to boot a encrypted dataset 2024-02-02 22:08:22 yaeahyy it boots :D 2024-02-02 22:15:30 I updated guide (better tell short notes) about installing alpine on apple silicon here https://arvanta.net/alpine/install-alpine-m1/ if bepeop-no-robot appears here with questions 2024-02-02 22:16:00 or anyone else trying to install alpine on apple silicon 2024-02-02 22:16:31 note: asahi kernel still doesn't works flawlessly on M3 machines 2024-02-02 22:31:18 I need something better than zola to make such html guides 2024-02-03 02:05:48 ACTION network lost? 2024-02-03 02:06:45 anyone meet this could still use 1.85.2 2024-02-03 03:36:15 Hello, how to write on APKBUILD when we want to install our package on custom directory ? for example "/usr/lib/local_lib". Thanks. 2024-02-03 03:37:17 The library is compiled using meson. 2024-02-03 10:40:04 Solved... thanks 2024-02-03 15:13:22 Is it possinle to prevent gnome-apps-core from depending on fwupd, or st least prevent it from starting fwupd automatically? 2024-02-03 15:13:30 ACTION hi 2024-02-03 15:14:10 What is starting fwupd automatically? 2024-02-03 15:20:53 Gnome desktop, something 2024-02-03 15:21:25 When i try apk del fwupd it says gnome depends on it. 2024-02-03 15:21:45 When i reboot it gets started when i have logged in 2024-02-03 15:22:40 The init script for fwupd is disabled 2024-02-03 15:23:23 the init script isn't needed for KDE or Gnome desktop users. 2024-02-03 15:23:52 You could do `apk add -t fwupd`, which installs fwupd as a virtual package. But if something tries to run fwupd, it will fail 2024-02-03 15:25:31 Hm that's an ideq 2024-02-03 15:25:58 I did chmod -x. But I don't know if that cause issues 2024-02-03 15:26:28 This is a vm so it is rather not necessary to have fwupd :) 2024-02-03 15:27:44 ah 2024-02-03 15:31:48 Hm https://gist.tnonline.net/YU 2024-02-03 15:32:40 Oh, a soname dependency 2024-02-03 15:32:43 then you cannot get rid of it 2024-02-03 15:35:44 :/ isn't it a little weird that there's a hard dependency on it? 2024-02-03 15:38:01 What's even more weird is if you go to the app center and disable all the repos in gnome, it still launches fwupd. 2024-02-03 15:38:12 (I mean all the fwupd repos) 2024-02-03 15:43:35 why do you care if it's installed in the first place 2024-02-03 15:45:38 Uses ram. Keeps connecting to remote server 2024-02-03 15:46:40 Remove the fwupd repos. Latter problem solved. 2024-02-03 15:48:46 # rm -rf /etc/fwupd/remotes.d/*.conf 2024-02-03 15:56:36 Hm. Not sure. It still starts unless I remove exec bit on it 2024-02-03 16:12:04 And I guess future updates will overwrite / add back those files 2024-02-03 16:12:48 yes 2024-02-03 16:13:34 maybe it would be better to blank them. 2024-02-03 16:13:43 I saw a thread on debian forums about it. They have libfwupd as separate package 2024-02-03 16:14:31 Or at least at that time 2024-02-03 16:17:01 https://packages.debian.org/search?searchon=contents&keywords=libfwupd.so.2&mode=exactfilename&suite=bookworm&arch=any 2024-02-03 16:18:24 Sounds like an xy problem though. The real solution would be an option in gnome-apps-core to disable it 2024-02-03 16:18:38 Not by sabotaging the fwupd binary 2024-02-03 16:22:48 they don't even have a gsetting to do so. insane. 2024-02-03 16:24:21 Forza: what I'm wondering is whether it is even calling the binary if it directly links against the library 2024-02-03 16:24:23 Yes probably. But at the moment we can't. From the thread I read gnome only really depends on the lib and not the daemon, that's why it got split in two packages 2024-02-03 16:24:38 right, but doesn't that mean the solution is moot? 2024-02-03 16:25:21 I don't know. If we can split the package we don't need the daemon and still keep all deps that gnome needs 2024-02-03 16:25:46 But what are you solving then? 2024-02-03 16:25:56 I mean, it's not a bad idea to split them 2024-02-03 16:26:15 Not having fwupd running automatically when gnome runs 2024-02-03 16:26:48 Why does it start the daemon if it does not even need it¿ 2024-02-03 16:27:12 To check for updates? 2024-02-03 16:28:00 But maybe there's another way to disable this in gnome, but as zcrayfish said, there's no gsetting for that? 2024-02-03 16:30:24 It's not an important issue for me personally though. Just was surprised that it was running on my vm when I had disabled the init script. So I googled and found the debian solution. People there were against "call home" apps without asking the user first. 2024-02-03 18:40:21 is there any tool available that I can use to extract an alpine package file? 2024-02-03 18:40:45 tar xzf 2024-02-03 18:44:40 huh, I expected something else since I can't open it with an archiver 2024-02-03 18:49:21 maybe the archiver goes by extension 2024-02-03 19:37:45 hey, any clue about getting spice to work properly? im unable to use the clipboard 2024-02-03 19:37:56 alpine as a guest 2024-02-03 19:39:17 you haven't indicated which hypervisor you're referring to 2024-02-03 19:41:00 kvm/qemu 2024-02-03 19:41:08 running on a different distro 2024-02-03 19:41:11 ubuntu iirc 2024-02-03 19:41:51 Juest: did you install spice-vdagent on the guest? 2024-02-03 19:42:16 yes, it does nothing, it fails to open /dev/uinput and exits 2024-02-03 19:42:25 Does ring a bell 2024-02-03 19:42:35 vm configuration issue? 2024-02-03 19:42:54 i read a bit about it but wanting to make sure here 2024-02-03 19:43:05 unfortunately i dont have access to the settings 2024-02-03 19:44:21 starting/stopping qemu guest helper crashes remote-viewer lol 2024-02-03 19:44:30 Juest: Did you specify `-chardev spicevmc,name=vdagent,id=vdagent`? 2024-02-03 19:44:37 where? 2024-02-03 19:44:46 to qemu 2024-02-03 19:44:49 #14314 2024-02-03 19:45:42 probably they didnt specify that 2024-02-03 19:45:48 its a simple vm 2024-02-03 19:46:01 changing any configuration seems too complex/difficult with the virt-manager gui 2024-02-03 19:46:18 I don't know virt-manager, so no idea how to do it there 2024-02-03 19:46:35 virsh edit didnt seem to accept changes for some reason 2024-02-03 19:46:40 Oh, this one as well: 2024-02-03 19:46:42 -device virtserialport,nr=1,bus=virtio-serial0.0,chardev=vdagent,name=com.redhat.spice.0 2024-02-03 19:46:50 ah 2024-02-03 19:49:43 also discussed here: https://www.spinics.net/lists/spice-devel/msg43812.html 2024-02-03 19:55:55 okay 2024-02-03 21:54:20 For the academics among us, I've ported Zotero to Alpine Linux: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/60178. It's already available on my personal repo (https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/60178) for Alpine 3.18, 3.19 and edge. 2024-02-03 21:55:16 It builds from source by building Firefox 102, and feeds the build artifacts into Zotero's build system where it would fetch Firefox prebuilt tar.gz. 2024-02-03 21:55:49 Personal repo: https://lab.ilot.io/ayakael/repo-apk 2024-02-03 21:57:45 Though it's latest beta release, so version 7.0.0_beta57. Version 6.x.x won't be covered by this aport as it is based on Firefox 60. As of now, better-bibtex seems to work, but zotfile hasn't been ported to Zotero 7 yet (and may never be). 2024-02-04 07:51:30 hi 2024-02-04 07:51:38 hey 2024-02-04 08:00:26 why is doas so silly 2024-02-04 08:00:33 there is no way to do 'nopass'? 2024-02-04 08:02:10 still? 2024-02-04 08:05:06 sewn: what are you want to achieve? 2024-02-04 08:05:12 nopass 2024-02-04 08:05:37 there is 2024-02-04 08:05:47 I mean, exactly 2024-02-04 08:06:01 permit nopass :wheel? 2024-02-04 08:07:00 That was a bit silly :) 2024-02-04 08:11:04 i am doing that 2024-02-04 08:11:07 It just dont doas 2024-02-04 08:11:09 It's not doasing 2024-02-04 08:11:26 it don't? 2024-02-04 08:11:39 it don't 2024-02-04 08:14:12 i'm so confused 2024-02-04 08:14:16 im in the wheel group and it just don't 2024-02-04 08:14:37 when did you add yourself to the wheel group 2024-02-04 08:14:40 in the same shell? 2024-02-04 08:14:41 installation 2024-02-04 08:15:34 Check that groups returns wheel 2024-02-04 08:15:44 it does 2024-02-04 08:16:10 check that the doas.conf file is the one actually being used 2024-02-04 08:16:16 it probably isn't 2024-02-04 08:16:39 /sbin/setup-user does 'permit persist' in /etc/doas.d 2024-02-04 08:16:40 alpine has borked doas.conf multiple times 2024-02-04 08:16:49 so its preferring that over the /etc/doas.conf 2024-02-04 08:16:50 awesome 2024-02-04 08:17:55 :) 2024-02-04 08:18:07 at this point I don't even know how doas works in alpine 2024-02-04 08:18:28 and that's after I've been following the exact patch upstream 2024-02-04 08:19:02 i would have switched to ssu if i didn't realize /etc/doas.d 2024-02-04 08:19:05 ssu just works 2024-02-04 08:19:21 doas also works, just... not in alpine 2024-02-04 08:19:28 sudo 2024-02-04 08:19:30 ACTION ducks 2024-02-04 08:19:36 hmm 2024-02-04 08:19:58 What's ssu? 2024-02-04 08:20:04 https://github.com/illiliti/ssu 2024-02-04 08:20:27 (I use doas daily in Alpine, no issue) 2024-02-04 08:35:06 btw, I use this: http://www.repo.hu/projects/sud0/ 2024-02-04 08:35:22 "sud0 - minimalistic sudo for automation" 2024-02-04 08:35:48 the 0 is way too far to type 2024-02-04 08:36:13 hahaha 2024-02-04 08:37:09 *for automation* ? 2024-02-04 08:37:14 hm 2024-02-04 09:47:11 running xfce on laptop (v3.18.6), keep getting at boot, udev: unknown user/group "root:input" on line 103 2024-02-04 10:09:22 sounds like you need to make an input group 2024-02-04 10:28:20 i thought that setup-desktop xfce would take care of it 2024-02-04 10:30:24 looks like it may not be easy to install xfce in ram-based setup, https://tpaste.us/a1Kb 2024-02-04 10:32:18 hmm, input group is in baselayout 2024-02-04 10:33:44 ditching networkmanager for now, and using rc-service wpa_supplicant to start wifi 2024-02-04 10:53:16 anyone is using zvols on alpine? it seems that it doesn't populate /dev/zvol.. 2024-02-04 11:01:30 that presumably needs dev manager support; maybe nobody wrote mdev rules for it? 2024-02-04 11:21:49 uhm 2024-02-04 11:42:58 and no one mentioned to install network-manager-applet \o/, finally network vi applet 2024-02-04 11:56:44 where I can find the instructions to fix an unbootable installation after grub update? 2024-02-04 11:57:19 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.20.0 2024-02-04 11:57:19 https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.20.0 2024-02-04 11:57:22 lol 2024-02-04 11:57:27 ouch 2024-02-04 11:57:29 :D 2024-02-04 11:58:31 thanks but I can't boot anymore so I need to boot from a USB stick and then? 2024-02-04 11:59:01 yeah 2024-02-04 12:00:08 and desktop freeze, reverting to previous apkovl 2024-02-04 12:01:09 shinobi57474858: indeed, boot from usb, and then chroot into your system 2024-02-04 12:02:31 you need to manually mount /proc /dev /sys before chroot and find the efi partition and mount that too to /boot/efi 2024-02-04 12:03:10 the grub cfg revert in the wiki didnt work for me 2024-02-04 12:05:25 thanks ikke and nsz I'll give it a try 2024-02-04 12:07:40 has somebody write access to this release notes? "Shot-term work-around" is kinda funny 2024-02-04 12:09:23 Someone fixed it already 2024-02-04 12:10:38 wiki shows the old for me 2024-02-04 12:10:51 Oh, the title still has the typo 2024-02-04 12:11:12 yep, now it's good :) 2024-02-04 12:13:41 (I checked the history, and now it is clear to me the text had the same typo) 2024-02-04 13:57:12 Heh i had a grub boot failure on gentoo after updating grub. 2024-02-04 13:57:35 Switched over to refind instead. Works rather nicely. 2024-02-04 14:01:23 ok, installing networkmanager, networkmanger-wifi, network-manager-applet - after desktop setup work most of the time 2024-02-04 14:02:41 but if i lbu ci after that and reboot, desktop freeze 2024-02-04 14:16:09 cannot re-produce it again, but something add cron->/etc/init.d/cron in runlevels/default, should have been crond 2024-02-04 14:16:22 added^ 2024-02-04 14:20:10 Forza: https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.20.0 2024-02-04 14:21:16 dok: forza was not asking for help with this issue 2024-02-04 14:21:18 Dok. Thanks, i saw it. my issue with grub was a couple of weeks ago. 2024-02-04 14:21:25 :) 2024-02-04 14:21:53 And, refind is not a solution for systems booting in bios mode 2024-02-04 14:34:20 Forza: alright, i had this issue last week and i though you as well 2024-02-04 14:36:59 :) I have a server at work with alpine and grub, so I'll be mindful to check this before updating and rebooting 2024-02-04 14:58:27 anyone know how to solve this? not sure what extension it is, but probably excel spreadsheet or similar : "The resource cannot be displayed because the file extension is not being accepted by your browser." 2024-02-04 15:00:50 well what are you doing to trigger it? 2024-02-04 15:03:59 pressing a button on a website "open" expecting to get "historical cash transactions" 2024-02-04 15:04:32 its saxo bank, which is used for stock trading 2024-02-04 15:04:58 no save as? 2024-02-04 15:06:28 no, just get a pop-up with this message 2024-02-04 15:06:46 what browser? 2024-02-04 15:06:51 never seen something like this before.. 2024-02-04 15:07:07 qutebrowser, but also tried ff with same result 2024-02-04 15:07:34 its probably expecting links! jk... 2024-02-04 15:07:39 do you have a program installed that can open spreadsheets? 2024-02-04 15:08:47 i have gnumeric 2024-02-04 15:10:00 firefox's developer tool should show you exactly what it's trying to do 2024-02-04 15:19:56 huh, actually worked on ff, thought i had tried it .. 2024-02-04 16:52:40 hi all 2024-02-04 16:53:10 does anyone use MAILFROM in busybox crontab? Does not work. 2024-02-04 16:54:08 MAILTO respects. 2024-02-04 16:54:43 is it supposed to? 2024-02-04 16:55:14 how to check? :/ 2024-02-04 16:55:50 i checked crond, cronie - all failed. 2024-02-04 16:56:11 crond is another package 2024-02-04 16:56:38 "how to check?" - look at source code? 2024-02-04 16:56:45 so three cron software does not respect MAILFROM. 2024-02-04 16:56:52 minimal: ok 2024-02-04 16:56:52 i have not heard of MAILFROM in any cron variant 2024-02-04 16:57:01 what is it supposed to do? 2024-02-04 16:57:23 format From address and name in mail of cron output 2024-02-04 16:57:58 does busybox cron even do mail? how do you configure it ? 2024-02-04 16:58:24 yes Busybox crond supports "MAILTO=" 2024-02-04 16:58:25 https://git.busybox.net/busybox/tree/miscutils/crond.c 2024-02-04 16:58:25 dalias: MAILTO="...@..." 2024-02-04 16:58:44 as long as ENABLE_FEATURE_CROND_CALL_SENDMAIL is enabled 2024-02-04 16:59:27 ok. never mind then. 2024-02-04 17:00:49 ahh it looks for sendmail in path rather than /usr/lib/sendmail 2024-02-04 17:01:04 that would explain it never working :-p 2024-02-04 17:01:44 tankf33der: it also doesn't support "INVENTEDOPTION1=" or "INVENTEDOPTION2=" 2024-02-04 17:17:51 tankf33der: MAILFROM is supported by the cronie package 2024-02-04 17:55:13 minimal:sigh 2024-02-04 18:00:21 well you've got your options now... 2024-02-04 18:03:04 tried hours with cronie, no success 2024-02-04 18:03:14 MAILFROM does not respect 2024-02-04 18:03:27 switched back to busybox 2024-02-04 18:03:59 both the cronie manpages and the cronie code indicate support for MAILFROM 2024-02-04 18:05:05 sure, uses login@domain name without respect. 2024-02-04 18:05:20 root@my.domain or so 2024-02-04 18:05:21 :) 2024-02-04 18:05:25 how does code give respect? 2024-02-04 18:05:47 should be mike@gmail.com 2024-02-04 18:06:03 MAILFROM=mike@gmail.com 2024-02-04 18:06:07 MAILFROM="mike@gmail.com" 2024-02-04 18:06:10 MAILFROM="" 2024-02-04 18:06:16 all combinations checked. 2024-02-04 18:06:42 root@my.domain and bounce by remote MTA :) 2024-02-04 18:08:10 you haven't exactly defined the problem - have you looked at the emails being generated when you set MAILFROM? 2024-02-04 18:11:02 yeap, i see root@my.domain from string without gmail 2024-02-04 18:13:43 you see this on the machine itself? (not relying on looking at bounces) 2024-02-04 18:14:12 havre you checked the logfiles to see if cronie is giving warnings/errors? 2024-02-04 18:15:51 never mind, thank you for you time! 2024-02-04 18:17:26 tankf33der: you don't want to try and fix your problem? 2024-02-04 18:19:46 if i used cronie, i see mail in the log and it bounced and rejected because came from root@my.domain 2024-02-04 18:20:00 i already spent hours in this. 2024-02-04 18:20:08 i would like take a break. 2024-02-04 18:20:36 well from a brief look at the code I would expect when you tried MAILFROM="" that an error would appear in the logfiles 2024-02-04 18:20:57 as cronie does not appear to allow "<" or ">" as characters in an email address 2024-02-04 18:21:19 which is why I asked if you checked the logs 2024-02-04 18:21:22 no syntax error from cronie 2024-02-04 18:21:41 maybe i was on busybox all this time? :) 2024-02-04 18:24:20 now my tor relay is broken. what a day. 2024-02-04 19:59:21 I noticed in crontab that there are directories listed in there. Does that mean I can create the "daily" directory and add my executable script to it to run? 2024-02-04 20:00:14 jerkwad_robot_, depending which cron you have, /etc/cron.daily and the like may already exist 2024-02-04 20:00:53 check the man/info etc for your cron, some of them are slightly different but do the same thing 2024-02-04 20:02:04 tdr, thanks! 2024-02-04 20:03:14 the default crontab for root uses run-parts 2024-02-04 20:03:25 (in alpine linux) 2024-02-04 20:03:27 run-parts /etc/periodic/daily 2024-02-04 20:03:34 that's it 2024-02-04 20:04:12 so if I just create the folder and add the script it'll do its thing? 2024-02-04 20:04:21 The folder should already exist 2024-02-04 20:04:36 Oh, got it... yep they do 2024-02-04 20:04:45 The scripts need to be executable 2024-02-04 20:04:53 thumbsup 2024-02-04 20:05:24 There were some other restrictions, but not sure if they are still there 2024-02-04 23:41:16 except for expanding commands, i guess ` is used extremely little in (plain) text? thinking of using it for marking/highlighting text 2024-02-04 23:44:55 some people use `` in ASCII text due to the lack of “. idk. 2024-02-04 23:56:04 ah yes, quite rare tho 2024-02-05 00:33:38 `Don't forget GNU Info single quote marks.' 03blegh Hated those things. 2024-02-05 09:58:20 "The character was designed for typewriters to add a grave accent to a (lower-case[a]) base letter, by overtyping it atop that letter." -- https://en.wikipedia.org/wiki/Backtick 2024-02-05 11:21:59 some computer fonts, like SunOS console apparently, used to render ` as ‘ 2024-02-05 11:22:42 that explains why sometimes things use `foo' for quoting things like paths for printing 2024-02-05 14:44:45 Hi, I've been using Raspberry Pi 3s to run some network tests, with one generating packets and one receiving packets (pktgen & tcpdump). I recently decided to upgrade from 3.5 to 3.18 on said Raspberrys and ran into issues with the packets being occasionally delayed. Attempting to generate a packet every 1 ms. Realise that it's a big jump in versions but anyone have any idea of ways to solve or 2024-02-05 14:44:51 encountered something similar? 2024-02-05 17:30:24 I am really happy with how good alpine runs with asahi. Kudos to @mps and everybody who contributed 2024-02-05 19:16:39 I have a physical server running Alpine 3.19 from a little SATA DOM SSD (like these https://duckduckgo.com/?q=sata+dom+ssd&t=ffab&iax=images&ia=images ) -- I set that up with setup-alpine and it all boots fine, but now I've added larger data SSDs... 2024-02-05 19:20:16 ...and have set up 4 partitions on each (3 identical gpt partition tables on the 3 drives) and now am trying to set up mdadm RAID1 across these. A partition for swap, another I'd like to use for /var or /var/log, and a large data partition. All formatted ext4, no LUKS nor LVM nor ZFS nor BTRFS, trying to stay simple. 2024-02-05 19:21:49 The RAID1'd swap partition starts up fine on boot but the 2 ext4 partitions that I'd manually mounted at /mnt/md0 and /mnt/md1 don't mount on boot, Alpine says: 2024-02-05 19:22:33 mounting UUID=my-long-uuid on /mnt/md0 failed: No such file or directory 2024-02-05 19:23:07 so does /mnt/md0 exist? 2024-02-05 19:23:48 The directories exist and are read-writable for all 777 both before and after I (manually) mount them. The UUIDs in /etc/fstab match the /proc/mdstat info... 2024-02-05 19:24:40 does it match the filesystem UUID as shown by blkid? 2024-02-05 19:27:20 Yes there are 3 partitions for each of those UUIDs as raid members 2024-02-05 19:27:51 I mean blkid shows those UUIDs 3 times each with different PARTUUIDs and in fstab I have them only once 2024-02-05 19:28:16 I didn't ask about PARTUUIDs, I asked about (filesystem UUIDs) 2024-02-05 19:28:46 you don't mount a partition, you mount a filesystem so you use the FS UUID 2024-02-05 19:29:06 I was just trying to explain that blkid shows these exact UUIDs 3x as I think it's supposed to 2024-02-05 19:29:27 ok, I'm just trying to be completely clear on the situation 2024-02-05 19:29:29 ah k checking... 2024-02-05 19:30:03 blkid (1 of 3) shows 2024-02-05 19:30:44 maybe paste your - mdadm status ; blkid ; cat /etc/fstab ; not sure either myself 2024-02-05 19:30:57 SLASHdev/sda3: UUID="df0c5a14-9c25-d9c0-e730-9980b3d36b90" UUID_SUB="83db2e5e-5d54-43c1-9ffd-24514958a51f" LABEL="e3a:0" TYPE="linux_raid_member" PARTUUID="0d8da683-6f67-4ddd-a844-5b320b8d7f9e" 2024-02-05 19:31:52 that's not an entry shows a ext4 fs.........if it was then there would be TYPE="ext4" as one of the fields 2024-02-05 19:32:00 fstab has 2024-02-05 19:32:04 UUID=df0c5a14-9c25-d9c0-e730-9980b3d36b90 /mnt/md0 ext4 rw,relatime 0 0 2024-02-05 19:32:37 where is the "blkid" output for /dev/md0 ? 2024-02-05 19:32:44 Should fstab have auto in the options? 2024-02-05 19:33:28 SLASHdev/md0: UUID="29f91fd4-bed1-44f3-8a5b-337f9c693e73" BLOCK_SIZE="4096" TYPE="ext4" 2024-02-05 19:33:44 So I'm using the wrong UUID! 2024-02-05 19:33:49 so then you should be mounting UUID=29f91fd4-bed1-44f3-8a5b-337f9c693e73 2024-02-05 19:34:19 df0c5a14-9c25-d9c0-e730-9980b3d36b90 is not a filesystem 2024-02-05 19:36:47 Thank you @minimal rebooting that machine now 2024-02-05 19:39:45 ...and it has drives! Success, thank you! 2024-02-05 19:44:02 These names, /mnt/md0 and /mnt/md1 were just for my sanity as I tried to set this up and debug. My use cases for the 2 large partitions are 1 for log files or other heavy OS writing, and 1 for a bunch of .qcow VM files (I haven't figured out passthru stuff yet). 2024-02-05 19:45:23 How can I get Alpine to mount a system path like /var to my /dev/md0 and a-place-for-massive-virtual-disks to my /dev/md1? 2024-02-05 19:46:21 I mean I know mount commands and fstab now but these feel trickier given that there's stuff in /var already, I'd already run setup-alpine etc when it was just the boot drive. 2024-02-05 19:46:40 (when there was only the boot drive installed, I mean) 2024-02-05 19:46:54 easiest would be to boot some different media, and move the files from /var to your disk 2024-02-05 19:49:10 ...ah so dust off the Alpine install USB stick, boot from that again, move folders from SATA DOM boot drive to my more write-friendly, bigger, RAID'd SSD partitions... then... just edit my mount lines in /etc/fstab? 2024-02-05 19:50:02 yes 2024-02-05 19:55:30 Is there some way to open a man page directly in a given section? Something like `man sway --title CONFIGURATION`? 2024-02-05 19:59:13 Not aware of any option and cannot find anything in the man page 2024-02-05 20:03:55 LESS=$(echo -n + ; man ls | grep -n DESCRIPTION | cut -f1 -d:) man ls 2024-02-05 20:04:11 is my best idea 2024-02-05 20:04:18 (tested on debian) 2024-02-05 20:35:18 erm... had to take a call and laptop went to sleep, got booted, can't see what I missed from this web client -- how often do the IRC logs get updated, they seem to be a couple of days behind? 2024-02-05 20:35:51 the only thing you missed was responses to WhyNotHugo's man question 2024-02-05 20:36:13 ah k thx 2024-02-05 20:36:45 or perhaps not even that, based on timestamps :) 2024-02-05 20:38:54 I can't even see my old conversation but fortunately I will half-remember it forever 2024-02-05 20:39:00 Habbie: `man man | grep DESCRIPTION` yields no results, although `man man | grep D`. 2024-02-05 20:39:05 I wonder if there are control codes in the middle. 2024-02-05 20:39:12 possible! i was surprised it just worked 2024-02-05 20:39:37 i'm sure with a bit of tweaking my trick can work for you 2024-02-05 20:39:43 What paths in Linux are the most write-intensive and best routed to my write-friendly enterprise SSDs? My main candidates are the data drive where I'd put VM .qcows, and swap, and /var/log but are there other common ones? Should I do all of /var ? 2024-02-05 20:40:08 al2024, /tmp? unless that is tmpfs already 2024-02-05 20:40:46 Right yes /tmp is tmpfs already thank you though 2024-02-05 20:41:28 as for the rest of /var, i don't know much about alpine, but some distros have database servers store their data there too 2024-02-05 20:41:41 also, depending on what you have, /var might not be much bigger than /var/log anyway, so might as well do it? 2024-02-05 20:43:46 good point 2024-02-05 20:44:02 My plan for this machine is to just have libvirt/qemu/kvm VMs running and niginx proxying traffic to them. Some Alpine, some Ubuntu and maybe a PostgreSQL in one but that'll be my next mind-eff, how to have the drive Pg uses be as fast and passtthrough-y and native as possible. 2024-02-05 20:45:20 Is it insanely bad to have say U22.04 and U23.10 and Alpine VMs with different kernel versions runing on an Alpine host with yet another kernel version? I always wondered at what point I'm just asking for trouble. 2024-02-05 20:45:44 with kvm, none of them should care much, i think 2024-02-05 21:48:22 Hi, I'm using Alpine as a virtualbox guest. with SDDM and OpenBox.I'm having an issue where the properties '/VirtualBox/GuestInfo/OS/LoggedInUsers' and '/VirtualBox/GuestInfo/OS/NoLoggedInUsers' don't reflect the fact that I'm logged in on the guest (they stay 0 and true respectively). I installed the guest additions following Alpine's doc, and other features work well (clipboard sharing, display resizing, etc). Any help appreciated! 2024-02-05 22:11:43 frazooka: I have no idea about virtualbox but is 'w' or 'who' working inside the guest? 2024-02-05 22:18:44 donoban: 'who' shows an empty output as well, that's wierd isnt it !? 2024-02-05 22:22:22 How do I instruct cloud-init to use udhcpcd? 2024-02-05 22:23:54 donoban: you put me on the right track -> https://wiki.alpinelinux.org/wiki/GNU_core_utilities : "Some programs like wall(1) depend on a libc feature named "utmp", that cannot be implemented in libc securely. [...] That is why programs that use utmp - such as wall, who or w - output nothing but still exit successfully. This is expected, a feature, and cannot be fixed easily." 2024-02-05 22:24:11 yeah, i was going to paste it 2024-02-05 22:24:25 I suppose that if you add it, virtualbox should report it properly too 2024-02-05 22:25:11 It's a little bit outdated, since utmps is available 2024-02-05 22:25:51 seems like vbox guest additions need to be patched for that though. Will take a look later 2024-02-05 22:38:51 Ermine, donoban: so I did install utmps and setup the services, but 'who' still outputs nothing 2024-02-05 22:42:34 because virtualbox guest additions do not use it 2024-02-05 22:42:49 please wait for some time 2024-02-05 22:43:56 Ermine: sure, sorry + thanks! 2024-02-05 22:45:34 Ermine: what do you mean "How do I instruct cloud-init to use udhcpcd?" ? 2024-02-05 22:46:12 frazooka: do you have the utmpx daemons running? 2024-02-05 22:46:47 minimal: nevermind 2024-02-05 22:47:16 Ermine: ? well you asked the question so I expect you wanted it answered 2024-02-05 22:47:40 I think I found the answer myself 2024-02-05 22:47:53 you think? 2024-02-05 22:49:00 yes 2024-02-05 22:49:17 rather than you know 2024-02-05 22:49:38 true 2024-02-05 22:49:49 so in others you don't know the answer 2024-02-05 22:49:56 s/others/other words/ 2024-02-05 22:51:16 But it's almost 2am here and my server works, so I'm not going to investigate possibly spurious error messages in my logs right now 2024-02-05 22:51:28 minimal: I have utmp-prepare, utmpd, wtmpd and utmp-init services running. Are those the daemons or should I look for something else? 2024-02-05 22:52:00 Ermine: well I was offering to help but you don't seem to want help, even thought you asked for it 2024-02-05 22:53:19 frazooka: in general those are required for utmpx to function. So "who" should produce output 2024-02-05 22:54:17 For VirtualBoxGuest Additions however I expect that has not been compiled with utmps support 2024-02-05 22:54:25 minimal: I'm afraid it doesn't output anything 2024-02-05 22:54:32 minimal: thank you for the offering, but I figured out that it's not urgent (for me) for now, though I've believed it is when I've asked it 2024-02-05 22:55:09 minimal: scrap that, who -a does 2024-02-05 22:55:41 that's already something!! 2024-02-05 22:56:54 Ermine: I suspect udhcpcd is not relevant to whatever you're doing, udhcpc may possibly be relevant to what you're doing with cloud-init but I suspect not 2024-02-05 22:57:28 Ermine, minimal: and now the vbox property is correctly set! "/VirtualBox/GuestInfo/Net/Count" = '1' (woohoo) 2024-02-05 22:58:08 This unblocks me, thank you very much Ermine, minimal, donoban 2024-02-05 22:58:33 aah no, sorry wrong property, it's too late i should sleep 2024-02-05 22:59:17 I'll let Ermine look into it when they have the time :)) Good night! 2024-02-06 00:28:32 is there a way to install the gawk apk, but have it leave the /usr/bin/awk symlink alone? 🤔 2024-02-06 00:31:06 I don't see any /usr/bin/awk referenced in the gawk APKBUILD and it has no pre/postinstall script to do so 2024-02-06 00:32:16 the package oracle shows it at least 2024-02-06 00:32:18 https://pkgs.alpinelinux.org/contents?branch=edge&name=gawk&arch=x86_64&repo=main 2024-02-06 00:32:30 ah, its in the "make install" part 2024-02-06 00:33:25 I guess the only way would be to create a new subpackage for the softlink and to make it optional when gawk is installed 2024-02-06 00:34:03 that's a lot of work to satisfy just one user who wants their system to be weird :) :) :) 2024-02-06 00:35:16 well gawk is only depending on by 20 packages so its "weird" to have gawk installed on an Alpine system anyway lol 2024-02-06 00:35:55 indeed. 2024-02-06 00:39:30 well this is the dumb thing I did for the time being, unless someone comes up with better (or more users desire the behavior that I want)... I removed the awk symlink, made it a hardlink to busybox, and then set the immutable bit on the hardlink. 2024-02-06 00:39:48 sure apk will perpetually have one error, but it's worth it ;) 2024-02-06 00:41:33 ACTION flashbacks of people setting immutable bit on files to prevent puppet tasks from changing them rather than bothering to update puppet tasks to cater for specific scenario 2024-02-06 00:43:20 haha 2024-02-06 00:43:46 I'm not a fan of my own solution 2024-02-06 00:43:58 hopefully it is temporary 2024-02-06 00:44:44 you mean until someone else implements an optional sub-package for the softlink? 2024-02-06 00:47:06 or until I realize that the awk script I was trying to run, which the author insisted will only run on the GNU version, actually runs exactly the same in busybox... Thus I completely wasted my time. 😁 2024-02-06 08:49:43 /list 2024-02-06 14:33:20 i think, xfce live + basic rescue tools iso is quite doable, anyone with resource to host .iso can a have a look at, https://git.insteps.net/mess/apklist/ 2024-02-06 14:34:22 i have not tried, but script setup-bootable can be used 2024-02-06 17:18:49 has there been anything new regarding s6 and alpine? 2024-02-06 17:18:56 sewn: no 2024-02-06 17:19:05 why not :( 2024-02-06 17:19:37 It's a lot of work and more complex than people assume 2024-02-06 17:19:43 aw ok 2024-02-06 17:19:46 is there a status 2024-02-06 17:20:14 https://gitlab.alpinelinux.org/alpine/tsc/-/issues/78#note_374794 2024-02-06 17:23:46 good posts by skarnet there. 2024-02-06 17:30:43 aww why hasnt alpine switched to s6 2024-02-06 17:30:48 s6 is really well designed 2024-02-06 17:31:38 sewn: because s6 is not yet what we need / expect from it 2024-02-06 17:31:45 Laurent explains it 2024-02-06 18:41:17 why doesn't alpine have hooks for efibootmgr to make an efistub like void does? 2024-02-06 18:48:02 sewn: efibootmgr doesn't create a EFISTUB 2024-02-06 18:48:27 I assume you're referring to creating a boot variable to *use* a EFISTUB 2024-02-06 18:54:13 sewn: to which the answer would be, you can write your own hook using the kernel-hooks package 2024-02-06 18:56:19 that's borijg 2024-02-06 18:59:31 sewn: well EFISTUB booting is not currently "officially" supported by Alpine, i.e. not implemented by setup-disk 2024-02-06 19:00:30 indeed Alpine doesn't "officially" use EFI boot variables at all 2024-02-06 19:00:44 it use the UEFI fallback file 2024-02-06 19:05:15 lame 2024-02-06 19:05:22 I tried to set it up myself but it just dont 2024-02-06 19:07:17 "it just dont" what? 2024-02-06 19:09:14 it doesn't boot 2024-02-06 19:09:37 efibootmgr also segfaults 2024-02-06 19:09:40 this is awesome 2024-02-06 19:10:55 well if efibootmgr segfaults whilst trying to create a boot variable then it's possible/likely the boot variable wasn't created and so of course then the system won't boot using the non-created variable lol 2024-02-06 19:11:24 well no it created just fine, it's segfaulting in chroot 2024-02-06 19:11:27 you haven't indicated what, if any, boot variable was created 2024-02-06 19:11:36 so perhaps the boot variable is incorrect 2024-02-06 19:12:32 so saying "it doesn't work" without specifying exactly what "it" (i.e. the boot variable) actually is doesn't really enable anyone to help you 2024-02-06 19:13:01 so what exactly efibootmgr command is being run to create a boot variable? 2024-02-06 19:13:19 s/exactly/exact/ 2024-02-06 19:14:19 the one in the docs but without using a variable for the params 2024-02-06 19:14:32 the one in which docs? 2024-02-06 19:14:40 efi stub bootloader page 2024-02-06 19:14:43 we're not mindreaders 2024-02-06 19:14:49 https://wiki.alpinelinux.org/wiki/Bootloaders 2024-02-06 19:15:19 I should have mentioned it's in the wiki, mb 2024-02-06 19:15:37 I was using uki before but it was taking up too much space 2024-02-06 19:20:09 so that shows example commands, can you pastebin (or equivalent) the actual commands you ran and their output and also the output of "efibootmgr" afterwards (to show all the EFI vars that exist) plus blkid/lsblk output to show how your disk(s) is laid out 2024-02-06 19:23:14 https://0x0.st/HkEl.txt 2024-02-06 19:23:17 thanks for your time 2024-02-06 19:26:00 hmm it's complaining about not finding my btrfs root 2024-02-06 19:26:23 I suppose I'll need the modules= kernel parameter which is usually inserted by grub or kernel hooks 2024-02-06 19:26:29 right, so perhaps the nvme driver is not being loaded 2024-02-06 19:26:40 yep 2024-02-06 19:26:55 you shouldn't need "modules=btrfs" as the initram should automatically load the module for the relevant fs type 2024-02-06 19:27:37 well basically you'd need (almost) the same cmdline options as when you'd use Grub instead for booting 2024-02-06 19:28:24 so what would I need to add to cmdline 2024-02-06 19:28:37 modules=nvme? 2024-02-06 19:28:48 that's something you could try 2024-02-06 19:29:06 what's the cmdline of a regular nvme grub boot tho 2024-02-06 19:29:33 assuming nvme is actually included in the initramfs file (i.e does /etc/mkinitfs/mkinitfs.conf list nvme in the features?) 2024-02-06 19:30:21 did you not previously install using Grub as bootloader? 2024-02-06 19:30:43 I did, it's forced by setup-disk 2024-02-06 19:30:51 and it did work 2024-02-06 19:31:07 so then didn't you look at the Grub config before you decided to replace it with EFISTUB booting? 2024-02-06 19:31:44 I did, I just had a major skill issue 2024-02-06 19:31:56 I was in the tty at the time so I couldn't copy paste easily 2024-02-06 19:32:05 I could have sourced the .cfg file though 2024-02-06 19:33:04 setup-disk is a script so you'd need to figure out what cmdline it would you in your scenario, there's no single hardcoded set of cmdline options, it depends on circumstances 2024-02-06 19:33:27 o ok 2024-02-06 19:34:42 which goes back to my earlier point that EFISTUB booting is not "offically" supported by Alpine, if you decide to do something non-Official then you're basically expected to know what you're doing... 2024-02-06 19:35:53 yeah it's not my first time dealing with efibootmgr 2024-02-06 19:36:05 I am used to just copying the kernel by moving ot 2024-02-06 19:36:13 to bootx64.efi to avoid this mess 2024-02-06 19:36:47 copying kernel to bootx64.efi wouldn't work as then you'd have no cmdline defined... 2024-02-06 19:37:28 I used to have a custom kernel 2024-02-06 19:37:32 lol I made a typo 2024-02-06 19:37:42 I wrote mnitrd 2024-02-06 19:37:58 so a custom kernel with a hardcoded cmdline? 2024-02-06 19:38:10 yep 2024-02-06 19:38:27 so then you must have known the required cmdline values? ;-) 2024-02-06 19:38:36 no I just set root and went on 2024-02-06 19:38:40 anyway it's working 2024-02-06 19:38:57 so it was the missing "modules=nvme"? 2024-02-06 19:39:10 1. don't do --disk /dev/nvme0n1p2 2. don't make typos in parameters 3. know your modules 2024-02-06 19:39:27 i believe it could be from just not making it load initramfs 2024-02-06 19:39:50 because of a.. typo 2024-02-06 19:46:39 definitely do not make typos when configuring your kernel or bootloader, I've done that also :P 2024-02-06 19:56:32 moving to the hardcore cmdline bootx64.efi kernel "just works" to this is too much 2024-02-06 20:01:47 compiling your own custom kernel and embedding a initramfs inside it and hardcoding a cmdline inside it "just works" whereas adding a kernel-trigger script is "too much"? wow 2024-02-06 20:09:38 anyhoo 2024-02-06 20:09:52 is it not possible to replace udhcpc and wpa supplicant with just iwd? 2024-02-06 20:10:06 have you tried? 2024-02-06 20:10:13 yeah 2024-02-06 20:10:22 so, is it possible? 2024-02-06 20:10:41 probably 2024-02-06 20:10:55 so why ask the question then? ;-) 2024-02-06 20:11:03 I think I just need to comment wlan out of /etc/network/interfaces 2024-02-06 20:12:02 yep that's all 2024-02-06 20:12:15 that and the iwd EnableNetworkConfiguration option 2024-02-06 20:13:15 so you've asked a question and answered it yourself all in the space of 3 minutes ;-) 2024-02-06 20:14:24 sometimes all you need is a hear yourself talk 2024-02-06 20:14:30 it can be helpful at times 2024-02-07 10:16:01 Hi guys. I've installed openscap from official package, but to use it I also need scap-security-guide https://github.com/ComplianceAsCode/content/releases/tag/v0.1.71 Now I would like to build it and create a package for it. I took a look at Arc's PKGBUILD https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=scap-security-guide and this looks similar to falpine's APKBUILD. But 2024-02-07 10:16:03 I think first I have to try and build it manually. Ofcourse I tried it but I got: CMake Error at /usr/share/cmake/Modules/FindPackageHandleStandardArgs.cmake:230 (message): Could NOT find PY_yaml (missing: PY_YAML) which means I'm also missing python-yaml package :S And probably some more sine there are quite a few required packages in that Arch's PKGBUILD. I would also try to 2024-02-07 10:16:05 create other packages if I knew they would not depened on a ton of other dependencies. 2024-02-07 10:16:07 ACTION sighs 2024-02-07 10:43:29 this one? https://pkgs.alpinelinux.org/package/edge/main/x86_64/py3-yaml 2024-02-07 10:44:18 add it to depends or makedepends 2024-02-07 10:46:57 you could search first, packages' name maybe different 2024-02-07 11:20:44 Aaah 2024-02-07 11:20:52 Thank you qaqland 2024-02-07 11:54:31 I'm still missing a few py3 packages: PY_pcre2 PY_cmakelint PY_sphinxcontrib.autojinja PY_sphinx_rtd_theme .... cmake install log: https://termbin.com/nj87 2024-02-07 11:54:41 4 to be exact 2024-02-07 12:23:59 Hi there. I installed Alpine linux 2 years ago on my homelab, but when having a look at my storage configuration there are a few things that look wrong with my boot partition. I restarted countless time with no apparent issue. 2024-02-07 12:24:34 frojnd: libxml2-utils has xmllint 2024-02-07 12:26:34 The boot partition is a mdadm raid1 array, but the output of fdisk -l /dev/md0 makes no sense. Can someone assist me in understanding what's happening ? 2024-02-07 12:30:57 frojnd: that's a package that I was keenly interested in getting into Alpine as well. When I looked into it, it required packaging a bunch of libraries before even getting to touch scap itself. Things like py3-sigma and py3-vt. I was working through that list added deps as I could. 2024-02-07 12:31:36 I'd be happy to help, if you'd like. 2024-02-07 12:33:18 https://gitlab.com/Durrendal/WS-Alpine-Packages/-/blob/master/testing/scap-security-guide/APKBUILD?ref_type=heads 2024-02-07 12:34:38 This is what I had for the scape-security-guide package. It's a little old at this point, but I can see if it builds against the latest tag, and make an MR for it if you'd like. (Or if you want the maintainer title that's totally fine as well!) 2024-02-07 13:15:27 why is startx/xinit a dependency of xorg-server? 2024-02-07 13:16:36 same goes to wayland 2024-02-07 13:16:44 i suppose its due to mesa? 2024-02-07 13:18:03 durrendal: I'm not as keen as you (making pkgs for Alpine) but am willing to learn and why not by helping creating scap-securith-guide Question is if you are willing to help me :) 2024-02-07 13:33:09 That being said.. how can I continue your work? I see that you already created APKBUILD file. Can I just download APKBUILD and run: abuild -r ? 2024-02-07 13:33:31 qaqland: right again. Thank you 2024-02-07 13:33:51 frojnd: I don't mind helping at all. Yes that's correct, pull that apkbuild file and then run abuild checksum && abuild -r 2024-02-07 13:34:07 you'll want to adjust the pkgver to 1.71 before running abuild checksum 2024-02-07 13:34:31 I just checked on my end, and it looks like the only missing package now is linkchecker 2024-02-07 13:34:35 linkchecker (no such package) 2024-02-07 13:34:43 https://pypi.org/project/LinkChecker/ 2024-02-07 13:35:59 hi all! i'm trying to upgrade an old alpine 3.17 box which was a little abandoned, and it won't let me anymore, now, because `wget https://dl-cdn.alpinelinux.org/alpine/v3.17/main/x86_64/APKINDEX.tar.gz` ==> "SSL routines:tls_post_process_server_certificate:certificate verify failed". is there a trick to make this work gain? 2024-02-07 13:36:02 again* 2024-02-07 13:38:38 you could change the mirror from https to http 2024-02-07 13:38:56 frojnd: I maintain a whole bunch of things, so I don't mind adding one or two more things to the list. Do you want to try and maintain packages for Alpine long term, or do you just need scap? 2024-02-07 13:39:23 or check if 'apk --no-check-certificate' works but i think that came later 2024-02-07 13:41:21 that already did it (http), thank you knolle! 2024-02-07 13:41:34 durrendal: I want to try and maintain for Alpine long term. I also need scap :) 2024-02-07 13:41:35 (not a security-sensitive system, just my weechat VM) 2024-02-07 13:43:58 packages are signed anyway so you don't really lose a lot security wise 2024-02-07 13:44:59 frojnd: awesome ^^ I'm looking forward to your scap contribution! Let me know if you get stuck anywhere and I am more than happy to help out. 2024-02-07 13:45:23 Great. Will let you know :) what do you prefer irc/mail ? 2024-02-07 13:45:37 in fact, I have a py3-linkchecker apkbuild together you can use as a jumping off point to get past that missing dep issue 2024-02-07 13:46:42 Spledid 2024-02-07 13:48:09 Both work for me, I'm on irc more or less constantly (weechat bouncer). But if longer form communication is better you can reach me at wpsinatra@gmail.com 2024-02-07 13:50:11 frojnd: https://gitlab.com/Durrendal/WS-Alpine-Packages/-/blob/master/testing/py3-linkchecker/APKBUILD?ref_type=heads 2024-02-07 13:50:42 not 100% positive if that works, I threw it together real quick using toAPK (converted the aur pkgbuild to an apkbuild) 2024-02-07 13:50:51 but it should get you going in the right direction 2024-02-07 13:52:42 One general question, when running abuild -r and dependencies are being installed... will they be automatically removed if build fails? 2024-02-07 13:53:01 yes that's correct, they get installed as a meta package and then bulk removed at the end 2024-02-07 13:53:09 in the event of success or failure actualy 2024-02-07 13:53:20 s/actualy/actually/ 2024-02-07 13:53:49 Than that's even more practical than trying to first install dependencies by hand and then running cmake .... 2024-02-07 13:53:56 I mean using APKBUILD 2024-02-07 13:54:05 And abuild 2024-02-07 13:54:19 Yes 2024-02-07 13:54:55 agreed entirely. I would suggest some alpine system (vm, lxc/d or incus container, chroot, dedicated box) running edge repos just for building packages 2024-02-07 13:54:55 Even more practical to run individual steps if you are iterating 2024-02-07 13:55:04 just to hedge against accidentally breaking your main system 2024-02-07 13:55:21 abuild unpack deps prepare build 2024-02-07 13:55:43 If you then add a dependency, you can just run abuild deps again 2024-02-07 14:03:39 I didn't actually know that, thanks for mentioning it ikke! 2024-02-07 14:41:57 good to know knolle :) 2024-02-07 14:45:15 oh, dang, after upgrade to 3.19, still no https for me, e.g. for things like ruby's bundler :( 2024-02-07 15:19:40 oooh, /etc/ssl/certs/ca-certificates.crt was empty lol 2024-02-07 15:23:02 can anyone here visit youtube on luakit? 2024-02-07 15:23:12 bye all! 2024-02-07 22:45:12 Could I pretty please with a cherry on top get --no-chown option on apk upgrade & fix? 2024-02-07 22:57:29 hey 2024-02-07 22:57:59 how do I search for what package provides a certain command? having trouble surfacing the command with google 2024-02-07 22:58:50 xgpt: apk search cmd:foo 2024-02-07 22:59:21 CcxWrk: patches welcome, i think 2024-02-07 23:01:14 cmd:foo? 2024-02-07 23:01:46 where foo = whatever command you are looking for 2024-02-07 23:02:15 ohhh 2024-02-07 23:02:22 I see what's going on, it's only in edge lol 2024-02-07 23:02:29 thanks for the help y'all 2024-02-07 23:02:41 np 2024-02-07 23:31:47 CcxWrk: strace -e inject=chown:retval=0 2024-02-07 23:49:00 Ughh, that might actually be a good idea in the meantime 2024-02-08 10:28:49 is there a kernel boot option to log every loaded modules? 2024-02-08 10:29:55 1 or 2 modules are missing from my initramfs, but I couldn't tell what 2024-02-08 10:30:49 it boots fine from the supplied aarch64 initramfs, so definitely a module 2024-02-08 10:42:24 ls /sys/modules 2024-02-08 10:42:50 if you use dracut, you could add "rd.break" to get a shell 2024-02-08 10:43:15 and I remember "lsinitrd" 2024-02-08 10:43:41 and besides modules, don't forget about firmware files 2024-02-08 10:46:00 I'm interested in the load order 2024-02-08 10:46:17 I don't want to put all of them into the initramfs 2024-02-08 10:46:57 it consistently stops at a specific point 2024-02-08 10:47:54 so I'm hoping for 1 or 2 modules can fix it 2024-02-08 10:49:06 maybe you can increase loglevel=... for the kernel itself 2024-02-08 10:50:21 already on 7 :( 2024-02-08 10:52:14 it just don't want to spit module loads 2024-02-08 11:23:11 I'm affected by ryzen cpu bugs and it bricked the update while building the kernel image. There are no tools included eg lsblk or iwctl. 2024-02-08 11:23:31 How do you repair the filesystem with alpin linux iso image? 2024-02-08 11:24:38 boot from a live image? 2024-02-08 11:24:44 totallyrandom: the extended iso has more tools present by default 2024-02-08 11:27:58 alright ill try the extended version 2024-02-08 11:41:04 Can anybody tell me when /etc/grub.d/20-linux_xen has an effect? 2024-02-08 11:44:51 ikke: i logged in with root but the tools are not available on the extended version (-ash lsblk: not found) 2024-02-08 11:45:22 totallyrandom: you still need to install them 2024-02-08 11:45:57 apk add cmd:lsblk 2024-02-08 11:46:04 or apk add lsblk 2024-02-08 12:21:48 ikke: what about iwctl? its require by world 2024-02-08 12:24:58 i cant even add apk linux for rebuilding 2024-02-08 12:47:46 ok I managed to get a connection and I downloaded and reinstalled the latest lts kernel image. 2024-02-08 12:48:10 I'm still stuck at loading the initial ramdisk. Have I missed a step? 2024-02-08 12:48:48 initram-fs? 2024-02-08 12:49:02 And stuck in what sense? 2024-02-08 12:54:34 who is managing the aports pull requests? whoever they are they are doing an amazing job 2024-02-08 12:54:45 60369 PRs total but only 431 opened 2024-02-08 12:54:58 ikke: when the machine boots it cannot get pass the initial loading screen where it says loading linux-lts and loading initial ramdisk 2024-02-08 12:56:52 sewn: A couple of us 2024-02-08 12:56:59 i love u guys 2024-02-08 12:57:33 totallyrandom: Did you have the boot partition mounted when you reinstalled the kernel? 2024-02-08 12:57:55 ikke: yes I mounted it to /mnt/boot 2024-02-08 12:58:15 i have the root part and one single boot part 2024-02-08 12:58:32 totallyrandom: did you then chroot into /mnt? 2024-02-08 12:59:10 ikke: yes with the provided start chroot script from the wiki 2024-02-08 13:06:47 totallyrandom: can you try to add nomodeset to the kernel arguments when booting? (and perhaps remove quiet) 2024-02-08 14:02:23 ikke: its not working either 2024-02-08 14:03:19 fyi i had to interrupt power from the machine in order to reboot 2024-02-08 14:04:00 it's a common issue with 2nd gen ryzen when the system hangs for no reason 2024-02-08 16:49:34 does anyone has expereince with usage of mkimage script from aports repository to build alpine-aarch64.iso on x86 host platform? I tried to do the following: aports/scripts/mkimage.sh --tag edge --outdir ~/iso --arch aarch64 --repository https://dl-cdn.alpinelinux.org/alpine/edge/main --profile base but ~/iso is empty and script finished in 1 seconds what definetely isn't enough to build all the packages and put them to iso. 2024-02-08 16:58:47 olku: mkimage.sh does NOT build packages, it assembles an ISO using existing APK packages 2024-02-08 16:59:41 that's why you're telling it which repo(s) to use 2024-02-08 17:00:33 minimal: it makes sense, but then it should be an iso image present in outdir 2024-02-08 17:02:07 ok, can then abuild build a package for aarch64 but on x86 host? 2024-02-08 17:03:22 olku: ikke already gave you an answer a while ago in alpine-linux channel 2024-02-08 17:03:49 qemu-user + binfmt 2024-02-08 17:07:25 oh, missed that 2024-02-08 19:18:29 GNOME installation on ram-based setup went nice, just sticking to rc-service wpa_supplicant for now 2024-02-08 19:20:06 battery power status also seems better than xfce4 2024-02-08 19:32:23 no one grins, coz i think this is first time i am using wayland 2024-02-08 19:53:13 didn't have to install log/logout screen saver, work fine 2024-02-08 19:56:49 hmmm, needed to close/open lid twice to send it to screen off mode, something is delaying it on first close 2024-02-08 20:01:18 quick lid open/close , keeps the screen lit 2024-02-08 20:05:40 vkrishn, which wayland compositor? 2024-02-08 20:05:42 haven't a gnome experience for yrs, so i can say it is nice 2024-02-08 20:06:35 there comes first grin, all i did is setup-desktop gnome (v3.18.6) :-) 2024-02-08 20:07:00 this is 1st gen (intel) laptop, i5 2024-02-08 20:07:11 ah, same as mine. 2024-02-08 20:07:42 i am kinda liking the fonts 2024-02-08 20:09:24 i don't even know what's/how its running 2024-02-08 20:17:10 would be nice if someone wrote a wiki on "how to run xfce on wayland" 2024-02-08 20:23:21 fn keys response is better than xfce4 experience 2024-02-08 20:37:48 any hint on howto generate what pkgs+version are installed via apk ? 2024-02-08 20:38:34 found it , apk list -I 2024-02-08 20:41:39 this is what i am thinking, take this list, diff it with previous xfce4 list, then voilà i have a modular gnome-desktop-install method list 2024-02-08 20:43:03 which means, anyone can download it on separate folder and install/remove, update/upgrade it (in theory) 2024-02-08 20:44:07 aha, "modular desktop pkgs overlay" - modpov 2024-02-08 20:45:58 so also possible to build modular squashfs pre-installed files, though that needs some tweaking in initramfs 2024-02-08 21:02:55 oh my dell, it went to sleep, while i blah-blahed + made tea, now it does not want to wake 2024-02-08 21:03:08 found my first gnome bug on al 2024-02-08 21:05:46 so i closed the lid, took few tea sips, and re-opend it, and its back awake 2024-02-08 21:07:53 gnome has a sleep disorder 2024-02-08 21:12:01 mpv play nice, only 4 degree rise in temperature 2024-02-08 21:12:21 ok its 6 2024-02-08 21:18:25 Can anybody tell me if the returned events of a poll syscall can have both POLLIN and an error set? 2024-02-08 21:19:04 Or to state differently busybox acpid is stuck in a loop cause both POLLIN and an error is set. I would like to know if that is an busybox or musl issue 2024-02-08 21:20:26 probably this is what is causing sleep disorder, "[vo/gpu/wayland] GNOME's wayland compositor lacks support for the idle inhibit protocol. This means the screen can blank during playback." 2024-02-08 21:34:45 "Thanks to the heroic effort of omni and celeste, rust 1.76 landed today in #alpine after being stuck on 1.72 for quite some time." 2024-02-08 21:34:56 Very nicely done 2024-02-08 22:04:27 somercet: i think its "mutter" 2024-02-08 22:08:24 ah 2024-02-08 22:19:19 sertonix: saw your bug report on aerc. did you see that on alpine? just curious. 2024-02-08 22:20:19 Yes, why? 2024-02-08 22:20:46 haven't seen it, but i'm not using notmuch. 2024-02-08 22:28:18 sertonix: scratch that, it repros 2024-02-08 22:39:05 is WTFPL compatible with MIT,GPL in same installation ? 2024-02-08 23:45:35 turns out, pressing power-button after sleep, turn it back-on (this might be not same on other laptops) 2024-02-08 23:48:06 MIT and GPL aren't all that similar. and throwing WTFPL in there to boot. It would depend on the context I guess. 2024-02-08 23:49:13 You shouldn't include GPL code inside an MIT, or WTFPL project. 2024-02-08 23:58:04 but these 3pkgs(all types) can exists in same install ? 2024-02-08 23:58:41 eg. GNOME install pulls all 3 2024-02-08 23:59:30 thanks for answer 2024-02-09 11:35:25 hey, how much space would a minimal desktop install take? 2024-02-09 11:40:28 Juest: typically less than 1G, but if you tune the linux firmware installed, but can be quite small 2024-02-09 11:40:36 hmmm 2024-02-09 11:40:46 sure, i mean this is going to run in a experimental box i have 2024-02-09 11:40:55 the expectation will be that the system will not be updated frequently 2024-02-09 11:41:07 s/will be/is/ 2024-02-09 11:41:32 i should be good with running alpine as a desktop distro right? 2024-02-09 11:44:10 Desktops should be regularly updated 2024-02-09 11:44:23 it will be regularly updated 2024-02-09 11:44:30 but it will not be booted frequently 2024-02-09 11:44:58 is alpine suitable for desktop use? 2024-02-09 11:46:01 ikke: as far as i've seen, a complete desktop install on most distros can take up to 20gb 2024-02-09 11:46:25 and thats with package cache 2024-02-09 13:13:01 I have used alpine as my desktop OS for a year or two, it's certainly possible but perhaps not for everyone 2024-02-09 13:13:19 (or three actually?) 2024-02-09 13:15:23 I recently re-installed everything from scratch and my root filesystem (excluding /home) is 2.35G, using sway 2024-02-09 14:18:19 bananicorn: what command di you use to get that statistic 2024-02-09 14:38:44 hey, how would i install alpine to a specific partition? 2024-02-09 14:39:45 setup-disks says it will erase the disk im trying to install on 2024-02-09 14:40:42 sewn: That's what 'zfs list' reports 2024-02-09 14:40:58 'du -h' reports a little less 2024-02-09 14:41:19 I have my /home partition separate 2024-02-09 14:41:46 of course zfs list only works if you use zfs for your root filesystem ;) 2024-02-09 14:59:57 Juest: You can install to a mountpoint, so you can just mount your partitions under that mountpoint appropriately and then run setup-disk "/path/to/mountpoint“ 2024-02-09 15:00:21 ah interesting 2024-02-09 15:00:52 Perhaps you can also specify a partition directly, not sure 2024-02-09 15:01:40 setup-disk wants to partition it seems 2024-02-09 15:02:42 You can also set BOOTLOADER=none if you don't want it to install a bootloader 2024-02-09 15:07:25 having issues with the setup 2024-02-09 15:07:28 ERROR: alpine-baselayout-3.4.3-r2: failed to rename var/spool/.apk.1330e212b7b42701c34c05275c3aba9bcde6d872c51306ee to var/spool/mail. 2024-02-09 15:07:28 ERROR: 2 errors updating directory permissions 2024-02-09 15:07:28 ==> initramfs: creating /boot/initramfs-lts for 6.6.16-0-lts 2024-02-09 15:07:28 awk: Relink `/usr/lib/libncursesw.so.6' with `/lib/libc.so.6' for IFUNC symbol `strcpy' 2024-02-09 15:07:29 Segmentation fault 2024-02-09 15:29:34 im having issues 2024-02-09 15:29:44 i see a warning icon in gparted in alpine 2024-02-09 15:29:55 when im looking at the filesystem 2024-02-09 15:34:58 well im used to gparted 2024-02-09 15:35:03 but its not super usable here 2024-02-09 15:35:08 im lacking something else 2024-02-09 17:12:20 hey, any idea why installing grub wipes all other grub entries and otherwise i am not able to get a boot entry for alpine? 2024-02-09 17:26:06 Juest: i have been using alpine as my primary workstation with desktop/sway for about 2 years. in my config I am using about 3GB excluding home and other things not related to alpine. I have my home on zfs zpool. can't help you with the dual boot partition thing, just dedicate a full disk to it 2024-02-09 17:27:37 aquamo4k: im doing multiple boot setup 2024-02-09 17:27:38 i have it working 2024-02-09 17:27:51 the problem is that somehow grub-install wipes all other entries 2024-02-09 17:28:05 i dont know where you can customize grub-install 2024-02-09 17:28:15 yeah, that's why i never do multi-boot, just get another computer is my motto :-) 2024-02-09 17:28:35 everything is possible but not necessary is how I roll 2024-02-09 17:29:19 for some reason older distros such as debian based do not see newer grub created images 2024-02-09 17:30:37 its ok 2024-02-09 17:30:44 i can wait for someone else to assist 2024-02-09 17:35:46 It seems like if one of the distros you have in multiboot update the grub, then you might have to go through each other os and fix them to align with the changes to grub - especially if there is a bug or error during the update. 2024-02-09 17:36:59 some OSes do a scan for other OSes when setting up grub 2024-02-09 17:37:02 you can also chain grubs 2024-02-09 17:40:42 yeah, good point, chaining might be a safer approach. 2024-02-09 17:42:37 Juest: grub-install wipes "all other entries"? which other entries? where are they defined? 2024-02-09 17:43:05 minimal: talking about other grub efi entries in the firmware 2024-02-09 17:43:07 are you referring to another Linux distros Grub boot options? 2024-02-09 17:43:11 eg, ubuntu, artix, 2024-02-09 17:43:28 "in the firmware" - you mean EFI boot variables? 2024-02-09 17:43:31 yes 2024-02-09 17:43:36 as in, in the boot list 2024-02-09 17:43:58 the grub menu, or the EFI boot variables? they are different things 2024-02-09 17:44:09 grub-install can *set* EFI boot variables, via efibootmgr in installed, but I don't believe it ever deletes any existing ones 2024-02-09 17:44:21 would surprise me too 2024-02-09 17:44:24 the option to boot directly to the grub menu from the uefi firmware 2024-02-09 17:44:44 well, i ran `grub-install` and it removed the other grubx64.efi entries 2024-02-09 17:44:48 Juest: have you actually checked the EFI boot variables? 2024-02-09 17:44:52 yes i have 2024-02-09 17:44:58 i checked after running grub-install 2024-02-09 17:45:11 artix and ubuntu are gone 2024-02-09 17:45:17 ah, that's probably because you didn't specify "--bootloader-id" when you ran grub-install... 2024-02-09 17:45:30 ah 2024-02-09 17:45:33 really? 2024-02-09 17:45:49 though that would only "delete" (actually replace) a single entry called "grub" 2024-02-09 17:45:53 well, is it possible to customize grub-install without passing parameters? 2024-02-09 17:46:14 it deleted other grub-install boot entries that it didnt make 2024-02-09 17:46:21 how do you customise something without telling it the customisations? lol 2024-02-09 17:46:38 Juest: are you sure about that? 2024-02-09 17:46:41 i guess it was done by modifying grub-install 2024-02-09 17:46:43 yes i am sure 2024-02-09 17:46:54 1st, bo you have efibootmgr installed? 2024-02-09 17:47:00 of course 2024-02-09 17:47:12 of course? it is not installed by default 2024-02-09 17:47:19 AFAIK 2024-02-09 17:47:33 i know its not installed by default but its required for installing into a uefi system for the boot entry 2024-02-09 17:47:38 so i already added it 2024-02-09 17:47:41 ok 2024-02-09 17:47:43 beforehand 2024-02-09 17:47:50 so run it to look at your current boot entries 2024-02-09 17:48:05 what entries do you currently have? 2024-02-09 17:48:07 i already done a before and after of efibootmgr 2024-02-09 17:48:23 ok, so pastebin/or equivalent the before and after output... 2024-02-09 17:50:18 so you want me to reboot and reproduce the issue to show you? 2024-02-09 17:51:07 alpine's grub is incapable of booting the artix entry because it chains initramfs and ucodes 2024-02-09 17:51:15 well I asked to see your before and after efibootmgr output 2024-02-09 17:51:37 you're asking for me to reproduce it just to show you a before/after 2024-02-09 17:51:54 because i do not have that output handy since i rebooted/lost the terminal 2024-02-09 17:52:08 no, you said you'd already "done a before and after of efibootmgr" so I was asking to see that 2024-02-09 17:52:15 yeah i did not save that 2024-02-09 17:52:40 "alpine's grub is incapable of booting the artix entry" - being incapable of boot that is a different/unrealated issue to the one of EFI boot variables being deleted 2024-02-09 17:52:49 i know 2024-02-09 17:53:05 so why not try and get one problem resolved before raising another one? 2024-02-09 17:53:18 fine 2024-02-09 17:54:00 is it not clear to you that the boot entries generated by different versions of grub-install are getting deleted by alpine's grub-install? 2024-02-09 17:54:54 as for what boot entries, they basically point to the grubx64.efi binary on the respective path, eg efi/artix/grubx64.efi, efi/ubuntu/shimx64.efi 2024-02-09 17:55:45 Juest: you are saying that is what is happening,, I'm asking for some evidence of that happening 2024-02-09 17:56:17 i just said i didnt save the evidence you want 2024-02-09 17:56:29 so then how do you expect someone to help diagnose the problem? 2024-02-09 17:56:45 explaining the problem 2024-02-09 17:58:20 explaining the problem doesn't provide actual evidence of the issue though 2024-02-09 17:58:27 also, setup-disk did not install a efi boot entry in first place for whatever reason despite having installed efibootmgr 2024-02-09 17:58:44 setup-disk does NOT manipulate EFI boot variables at all 2024-02-09 17:58:50 it is designed that way 2024-02-09 17:58:58 it uses fallback EFI booting 2024-02-09 17:59:33 yeah my firmware does not support boot entries via fallback efi booting for hard drives, only usb devices 2024-02-09 17:59:34 setup-disk is not designed for multi-OS/distro booting, it assumes it is the only OS installed on a machine 2024-02-09 18:00:08 using a mount point enables you to install in a multi-os environment 2024-02-09 18:00:16 anyways 2024-02-09 18:00:28 Juest: are you *sure* about that? Is there an official doc from the motherboard manufacturer stating that? I'd need to check the UEFI spec but AFAIK fallback is standard behaviour for UEFI implementations 2024-02-09 18:00:53 while it is standard, not every manufacturer follows that 2024-02-09 18:01:09 Juest: a mount point does NOT change how setup-disk calls grub-install and so it still sets up fallback booting only 2024-02-09 18:01:14 this firmware's built in efi shell does not boot vmlinuz directly for instance 2024-02-09 18:01:49 Juest: perhaps - where did you determine that your motherboard doesn't support fallback for non-removable storage? 2024-02-09 18:02:14 which MB vendor is it? which UEFI vendor is it? 2024-02-09 18:02:30 by trial and error, there's no automatic entry that appears on the presence of EFI/BOOT/BOOTX64.EFI 2024-02-09 18:02:41 with usbs it does work 2024-02-09 18:02:46 gigabyte 2024-02-09 18:02:49 uefi 2 2024-02-09 18:03:14 uefi vendor.... 2024-02-09 18:03:26 do you know how to query all this from linux? 2024-02-09 18:04:01 dmidecode 2024-02-09 18:04:11 in the "BIOS Information" section 2024-02-09 18:06:22 BIOS information, System information and base board information 2024-02-09 18:06:31 System Information 2024-02-09 18:06:31 Product Name: GB-BSi3-6100 2024-02-09 18:06:31 Manufacturer: GIGABYTE 2024-02-09 18:06:46 Vendor: American Megatrends Inc. 2024-02-09 18:06:46 Version: F4 2024-02-09 18:06:46 Release Date: 12/08/2015 2024-02-09 18:06:59 Base Board Information 2024-02-09 18:07:00 Manufacturer: GIGABYTE 2024-02-09 18:07:00 Product Name: MKLP3AP-00 2024-02-09 18:07:02 Please use pastebin next time 2024-02-09 18:07:18 they're not that many lines 2024-02-09 18:07:27 sorry 2024-02-09 18:07:27 Just in case 2024-02-09 18:07:36 i have flood control on my client 2024-02-09 18:07:37 Juest: what about the "BIOS Information" section? 2024-02-09 18:07:59 minimal: vendor, version and release date is from bios information 2024-02-09 18:10:00 im not willing to upgrade the bios 2024-02-09 18:10:11 Warning: 2024-02-09 18:10:11 Because BIOS flashing is potentially risky, if you do not encounter problems using the current version of BIOS, it is recommended that you not flash the BIOS. To flash the BIOS, do it with caution. Inadequate BIOS flashing may result in system malfunction. 2024-02-09 18:10:21 so you're not a old UEFI - Gigabyte shows latest is F13 18/07/2023 2024-02-09 18:10:32 i guess 2024-02-09 18:10:44 so perhaps they've fixed some bugs 2024-02-09 18:10:50 but i have no information whatever the new bios updates support newer uefi features 2024-02-09 18:11:36 there's no change log 2024-02-09 18:12:01 probably becuase they don't want to admit to how many bugs they've fixed lol 2024-02-09 18:12:21 ACTION shrugs 2024-02-09 18:13:33 anyway without some output to look at I don't see how anyone can really help you regarding the "boot variable deletion" issue 2024-02-09 18:13:48 the issue is in grub-install 2024-02-09 18:14:17 so you said before, I have been asking for actual evidence to look at regarding that 2024-02-09 18:14:39 my assumption is because after grub-install is ran, the uefi entries disappear 2024-02-09 18:14:50 i'll try again with verbosity 2024-02-09 18:15:39 grub-install DOES have code to delete a EFI boot variable for a single "distributor" 2024-02-09 18:16:09 why is it deleting all distributors entries then? 2024-02-09 18:16:36 we're back to "you have provided no evidence of that occurring" ;-) 2024-02-09 18:16:54 you say it is occurring, I'm asking for solid evidence of that 2024-02-09 18:21:28 can i ask for someone else to help who is less demanding? 2024-02-09 18:21:34 as you did not pass grub-install the "--bootloader-id" option then it will default to "grub" and there delete any "grub" bootloader entries, not "ubuntu", "artix" etc 2024-02-09 18:21:53 less demanding? you mean you want someone who works on "feelings", not "facts"? 2024-02-09 18:22:15 it deleted artix because the path ends in grubx64.efi maybe? 2024-02-09 18:22:17 I *am* trying to help you but you're not providing any hard facts 2024-02-09 18:22:32 I doubt that 2024-02-09 18:22:54 you could of course run "grub-install" with "-v" option to see what iyt is actually doing... 2024-02-09 18:22:59 im doing something more elaborated just for you 2024-02-09 18:24:23 well if you'd rather I didn't help you then I can leave you to figure things out yourself - I'm not demanding, I just want to investigate based on facts and evidence 2024-02-09 18:29:23 it'd be nice if there was some independent resource tracking uefi implementations by board. 2024-02-09 18:29:34 what works and what doesn't. 2024-02-09 18:30:47 invoked: that would be quite complicated/resource intensive to do 2024-02-09 18:31:01 and would you retest for each UEFI revision on each MB? 2024-02-09 18:31:40 it would not be simple, and it would require community input, sure. 2024-02-09 18:31:57 but currently nobody tracks any of it 2024-02-09 18:32:28 having something would be better than nothing 2024-02-09 18:32:49 not everything in the UEFI spec is mandatory to implement 2024-02-09 18:33:02 most UEFIs use EDK II as a base 2024-02-09 18:33:05 that's not the question such a resource would be trying to answer 2024-02-09 18:33:12 it would just be saying what works and what doesn't. 2024-02-09 18:33:28 you've seen the recent various UEFI security issues? image loading bugs? network boot bugs? etc 2024-02-09 18:33:52 you know, sunlight is the best disinfectant 2024-02-09 18:34:08 the various vendors affects were because the bugs were (from memory) present in the EDK II code that they used 2024-02-09 18:36:39 part of the issue is when people buy shitty boards without having any way to know how specifically shitty they are in advance 2024-02-09 18:36:54 so there's not as much incentive for board makers to do things right 2024-02-09 18:38:44 agreed 2024-02-09 18:39:05 "if it manages to boot Windows then ship it" 2024-02-09 18:39:28 yep. 2024-02-09 18:48:24 https://paste.ee/p/9XFpr#Uw4IolBRALNl02rPHfnO3zm_GsJk0h7J https://paste.ee/p/4AVqt#0PsQ9ji5NFhU0shvFdCorp6aQpazLpxg https://paste.ee/p/FaMeg#GUDPJRs82Q7fqS_wBrSN2mljFPRqA3Bg 2024-02-09 18:49:21 Well, sometimes the kernel says "Your BIOS is broken" 2024-02-09 18:49:21 here you go minimal, attempting it on three installs with before/after 2024-02-09 18:49:41 i have to use pci=nomsi for this computer or hdmi audio wont work 2024-02-09 18:50:16 well sorry, i thought the ubuntu entry would get deleted too but it seems its not touched 2024-02-09 18:50:24 but artix is getting removed 2024-02-09 18:50:39 which is already customized when i run `grub-install` 2024-02-09 18:50:51 and it does not delete existing boot entries 2024-02-09 18:54:46 ok, so only the artix entry is getting removed for some reason 2024-02-09 18:55:32 grub-install will remove any entry named with the appropriate "distribution" and as you didn't specify any then "grub" is used by default 2024-02-09 18:57:23 so grub-install runs "efibootmgr" and looks for lines of the form "Boot" and then does a *substring* compare of "grub" with each of those - only "Boot0003" would have met that criteria 2024-02-09 18:57:49 and so only grub-install would have used efibootmgr to only delete "Boot0003" 2024-02-09 18:58:22 however as I said earlier, it you ran "grub-install -v" then perhaps you'd get more useful information into what grub-install is actually doing 2024-02-09 19:08:25 minimal: running with -v fills my terminal buffer so i cant copy everything, wait i'll just install netcat and do termbin 2024-02-09 19:08:37 but things go to stderr rather than stdout i believe 2024-02-09 19:08:56 2>&1 2024-02-09 19:09:58 ah thanks 2024-02-09 19:10:05 i'll reboot and dump grub-install -v 2024-02-09 19:15:02 different question 2024-02-09 19:15:19 how do i get graphical privilege escalation to work? 2024-02-09 19:15:21 polkit? 2024-02-09 19:16:20 I can see one possible situation where grub-install can make the wrong decision, if the distribution is set to "boot"...but that would have to be explicitly set as the default is "grub" (which your Alpine boot variable is using for its name) 2024-02-09 19:17:24 i notice its the exact same arguments on the partition specification 2024-02-09 19:17:50 what do you mean? the verbose output? 2024-02-09 19:18:45 in efibootmgr 2024-02-09 19:19:05 I don't understand what you mean 2024-02-09 19:20:46 Windows Boot Manager HD(11,GPT,7f6249f4-0b3b-48c2-8834-70e9ba25d211,0x37cff800,0x8c800) grub HD(11,GPT,7f6249f4-0b3b-48c2-8834-70e9ba25d211,0x37c41000,0x202000) 2024-02-09 19:21:55 anyways 2024-02-09 19:22:10 2>&1|nc ? 2024-02-09 19:23:08 https://termbin.com/eqgm 2024-02-09 19:23:19 well the partition referring to is the ESP partition as that is where *all* the various OSes EFIs files are... 2024-02-09 19:23:35 notice the last things 2024-02-09 19:23:38 is it different 2024-02-09 19:23:51 the only shows 2 lines of output 2024-02-09 19:23:54 whoops 2024-02-09 19:23:55 s/the/that/ 2024-02-09 19:23:57 i forgot to run with -v 2024-02-09 19:25:42 just out of curiosity, and I've only read 4/5 of this, but: is boot/efi a partition? If it ended up dismounted, that would account for the other files being "deleted" 2024-02-09 19:26:27 yes its a always mounted partition 2024-02-09 19:26:35 can you confirm it's mounted now? 2024-02-09 19:26:44 its in fstab 2024-02-09 19:27:08 UUID=728a6afa-d593-4f87-bc4e-af9593e6138b / ext4 rw,relatime 0 1 2024-02-09 19:27:09 UUID=339C-7C85 /boot/efi vfat rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=utf8,shortname=mixed,errors=remount-ro 0 2 2024-02-09 19:27:22 so far 2024-02-09 19:27:30 and in `df -T`? 2024-02-09 19:27:32 the relevant lines in grub-install -v output is 2024-02-09 19:27:33 grub-install: info: copying `/boot/grub/x86_64-efi/core.efi' -> `/boot/efi/EFI/grub/grubx64.efi'. 2024-02-09 19:27:34 grub-install: info: Registering with EFI: distributor = `grub', path = `\EFI\grub\grubx64.efi', ESP at hostdisk//dev/sda,gpt11. 2024-02-09 19:27:34 grub-install: info: executing efibootmgr --version /dev/null. 2024-02-09 19:27:34 grub-install: info: executing modprobe -q efivars. 2024-02-09 19:27:34 grub-install: info: executing efibootmgr -b 0003 -B. 2024-02-09 19:27:34 can you post the verbose grub-install output rather than get sidetracked about something else? 2024-02-09 19:28:03 sorry 2024-02-09 19:28:06 /dev/sda11 vfat 1.1G 407M 620M 40% /boot/efi 2024-02-09 19:28:13 cool 2024-02-09 19:28:20 "executing efibootmgr -b 0003 -B" - so that is grub-install deleting the Boot0003 variable.......which is Alpine 2024-02-09 19:28:44 i'll look at the output on artix 2024-02-09 19:28:49 and back in here 2024-02-09 19:29:01 so if there's no sign of any other "efibootmgr -b 000x -B" entries then it is not deleting other boot variables 2024-02-09 19:29:14 i accidentally ran grub-install without v 2024-02-09 19:31:04 https://termbin.com/rwbd 2024-02-09 19:31:04 artix grub-install 2024-02-09 19:31:54 Hmm? What about artix? :) 2024-02-09 19:32:57 alpine grub-install is deleting artix boot entry 2024-02-09 19:33:26 the artix grub-install output isn't relevant, what is relevant is verbose grub-install on Alpine when the existing Artix EFI boot variable is apparently being deleted 2024-02-09 19:34:15 https://termbin.com/ly5n 2024-02-09 19:34:20 solus is in the mix here too 2024-02-09 19:34:31 invoked: that is fine and its systemd-boot not grub 2024-02-09 19:34:59 i have a different problem in that distro where anytime it does the boot update procedure it causes all entries to duplicate 2024-02-09 19:35:51 minimal: the output does not tell anything about why it is deleting it 2024-02-09 19:35:58 i believe its because it finds grubx64.efi in it 2024-02-09 19:37:59 so the output DOES show both Boot0003 (Alpine) and Boot0005 (Artix) being deleted by grub-install 2024-02-09 19:38:14 it does not explain anywhere why 2024-02-09 19:38:30 the only remote occurrence is that both contain 'grubx64.efi' in the path 2024-02-09 19:38:59 as I said before grub-install runs "efibootmgr" to see the boot entries - that ONLY show the Boot000x value and the bootloader-id/distribution, it does NOT show the EFI filename 2024-02-09 19:40:02 i'll try booting alpine from efi shell instead 2024-02-09 19:43:40 https://termbin.com/dyd9 - here you go 2024-02-09 19:43:59 it still deletes artix's 2024-02-09 19:46:38 what does "efibootmgr" on its own show? 2024-02-09 19:52:00 minimal: you know what? i just decided to set GRUB_DISTRIBUTOR="alpine", this causes it to no longer delete other distributors grub 2024-02-09 19:52:38 you could also have run grub-install with "--bootloader-id=alpine" instead 2024-02-09 19:53:28 the point is whilst grub-install is indeed using efibootmgr to delete Boot0005 I can see no reason WHY from looking at the output and the grub code 2024-02-09 19:53:45 which version of Alpine are you using? which version of Grub? 2024-02-09 19:53:56 3.19.1 2024-02-09 19:54:06 so Grub 2.06? 2024-02-09 19:54:09 grub 2.06-r17 2024-02-09 19:56:24 can you please run "efibootmgr" (no options) ? 2024-02-09 20:04:57 minimal: i've ran efibootmgr without options and it had that verbosity 2024-02-09 20:05:05 what verbosity? 2024-02-09 20:05:07 only in ubuntu it appears to not have verbosity and only show the entries names 2024-02-09 20:05:16 ah, then that is the problem then 2024-02-09 20:05:25 ??? 2024-02-09 20:05:32 meh 2024-02-09 20:05:42 I don't have an Alpine UEFI VM handle to check so I was looking at "efibootmgr" output on Debian UEFI 2024-02-09 20:06:09 run it with -v, the output is very similar, with dots 2024-02-09 20:06:15 you mean that "efibootmgr" with NO options gives more output per boot entry than: 2024-02-09 20:06:31 "Boot0002* debian" for example 2024-02-09 20:06:38 yes 2024-02-09 20:06:45 then yes that is the problem 2024-02-09 20:06:53 on your end 2024-02-09 20:06:58 run efibootmgr -v 2024-02-09 20:07:02 I wonder why Alpine efibootmgr is outputting differently on Alpine 2024-02-09 20:07:15 because the debian one is modified 2024-02-09 20:07:19 **NO** grub-install runs "efibootmgr", not "efibootmgr -v" 2024-02-09 20:07:23 and alpine and artix do not modify it 2024-02-09 20:07:33 grub-install runs a efibootmgr action 2024-02-09 20:07:53 efibootmgr displays the final output after executing an action 2024-02-09 20:07:56 so to be 100% clear, when you run "efibootmgr" on Alpine with *no* options it shows more output per boot entry? 2024-02-09 20:08:05 yes 2024-02-09 20:08:16 ok, so it is a Grub bug then 2024-02-09 20:08:23 when you run eg, efibootmgr -B -b 003 it prints the output 2024-02-09 20:08:26 afterwards 2024-02-09 20:08:33 the resulting efi boot list 2024-02-09 20:08:50 grub-install does a *substring* match on the "efibootmgr" output looking for the distribution name 2024-02-09 20:09:03 ah i see the problem 2024-02-09 20:09:10 its matching 'grub' in grubx64.efi 2024-02-09 20:09:13 obviously on Debian etc efibootmgr does NOT show the EFI filename 2024-02-09 20:09:35 the substring search is designed to match on the boot entry name 2024-02-09 20:10:03 minimal: you assume wrongly by not understanding the evidence properly 2024-02-09 20:10:11 so any boot variable with the distribution name appeared anywhere in the 1st 80 characters will be matched and deleted 2024-02-09 20:10:21 I assumed what exactly? 2024-02-09 20:10:25 nowhere it shows that grub-install runs efibootmgr itself 2024-02-09 20:10:38 you think that the debian efibootmgr affects things 2024-02-09 20:10:47 debian-modified* 2024-02-09 20:10:58 efibootmgr shows the result after running a command 2024-02-09 20:11:12 grub-install DOES run the efibootmgr program, you've just provided verbose output showing it does 2024-02-09 20:11:44 what the verbose output does NOT log is that grub-install ALSO runs "efibootmgr" with no options via a pipe but does not verbosely log that it does so 2024-02-09 20:12:01 okay then 2024-02-09 20:12:26 it executes efibootmgr --version 2024-02-09 20:12:29 not efibootmgr alone 2024-02-09 20:12:37 the question is which different distros have either behaviour for efibootmgr 2024-02-09 20:12:38 i guess it doesnt log that for some reason 2024-02-09 20:12:57 Juest: no, I am saying that grub-install runs efibootmgr with NO options 2024-02-09 20:13:11 I am staring at the code where it does exactly that 2024-02-09 20:13:14 i can tell you for certain that debian based distros and maybe red hat distros do modify the provided software 2024-02-09 20:13:33 arch, alpine, do NOT modify the provided software and package as is 2024-02-09 20:13:49 right, and the Grub devs likely are mainly testing with "popular" distros and so based their code on those distros behaviour 2024-02-09 20:13:59 not really? 2024-02-09 20:14:07 you dont know that, do you? 2024-02-09 20:14:35 I did say "likely", which means I was not stating a fact 2024-02-09 20:14:36 alpine is using upstream grub 2024-02-09 20:14:53 I know, I'm the one who recently updated Alpine grub in Edge to 2.12... 2024-02-09 20:15:17 popular distros modify things themselves and use old versions with backports and support it 2024-02-09 20:15:47 but I'm talking about different behaviour between efibootmgr in distros, not grub-install 2024-02-09 20:16:00 the same goes for efibootmgr 2024-02-09 20:16:01 ... 2024-02-09 20:17:01 should i confirm this is happening by using a distributor named shim? 2024-02-09 20:17:23 if you want to risk breaking another installed distro? lol 2024-02-09 20:17:33 i can boot back to that distro 2024-02-09 20:17:43 also the name would need to appear in the first 80 characters of the line 2024-02-09 20:18:03 for some reason os-prober wont pick up artix nor alpine because those are newer grub versions 2024-02-09 20:18:16 ubuntu os-prober 2024-02-09 20:18:38 well then that would be an Ubuntu issue then? ;-) 2024-02-09 20:18:55 no, because its likely to also happen in debian 2024-02-09 20:19:01 it has to do with the features 2024-02-09 20:19:09 they backport security patches not features 2024-02-09 20:19:13 well not always 2024-02-09 20:19:14 and 2024-02-09 20:19:37 im using ubuntu 22.04 lts and it did not pick up neither distro 2024-02-09 20:19:48 i guess its because its too new? 2024-02-09 20:20:07 i have efi shell to the rescue really 2024-02-09 20:20:23 should i really bother to update the firmware? 2024-02-09 20:22:13 I don't see how the Grub version has anything to do with how os-prober detects EFI boot files 2024-02-09 20:23:50 ok 2024-02-09 20:27:33 so the default (no options) efibootmgr output was changed in March 2019 and shipped in efibootmgr 18 2024-02-09 20:28:13 so version before 18 would only show boot entry number and name and so grub-install would not have this issue then 2024-02-09 20:28:26 but with efibootmgr 18+ this issue can occur 2024-02-09 20:29:01 but it is really a grub bug as grub-install should be more careful in parsing the efibootmgr output 2024-02-09 20:41:13 Juest: I suspect your os-prober issue is that it doesn't know how to detect Alpine kernels and initramfs files 2024-02-09 20:44:05 same with artix? 2024-02-09 20:44:09 which runs 6.6 2024-02-09 20:44:52 possibly, I know nothing about artix 2024-02-09 20:45:17 its basically very similar to arch 2024-02-09 20:45:20 rolling release distro 2024-02-09 20:45:31 basically Alpine names its kernel and initramfs files in a "non typically" way so os-probe won't find them 2024-02-09 20:46:24 as Alpine doesn't support more than 1 kernel version to be installed its filename does NOT include the version unlike most distros 2024-02-09 20:46:40 os-prober has patterns for filenames to search for 2024-02-09 20:51:28 minimal: what version of os-prober do you have? 2024-02-09 20:51:37 also are you on sid or stable? 2024-02-09 20:54:12 I have various versions of Debian 2024-02-09 20:54:51 whichever you were testing on 2024-02-09 20:55:19 well I wasn't "testing", I just happened to efibootmgr to see typical output 2024-02-09 20:55:37 anyway 18 is packaged for some Debian versions 2024-02-09 20:56:54 the point is the problem only happens with efibootmgr 18+ and where someone runs grub-install and where the "distribution" value also happens to be a substring of any "efibootmgr" boot entries output 2024-02-09 20:57:23 yeah, open a issue 2024-02-09 20:57:31 sir yes sir 2024-02-09 20:57:49 or perhaps that would be too demanding? ;-) 2024-02-09 20:58:59 oops, wrong conversation, was still thinknig about efibootmgr one 2024-02-09 20:59:34 the os-prober one isn't an issue for me as I don't dualboot Alpine and Debian 2024-02-09 21:13:15 okay 2024-02-09 21:37:44 frazooka: re virtualbox guest additions: how did you try to get the list of logged in users? 2024-02-10 00:32:52 Hello, it's been a month I'm playing with APKBUILD. If I use other people's APKBUILD and modify it (e.g. add a patch), could I put my name on "maintainer" ? Thanks. 2024-02-10 00:40:54 would you be talking over package maintainership from them? 2024-02-10 00:42:22 I don't take the maintainership, so no. 2024-02-10 00:43:08 so then why would you update the package maintainer field? 2024-02-10 00:44:57 maintainer is the person who has maintainership of the package 2024-02-10 00:45:00 so no, don't update it 2024-02-10 00:47:46 I have a plan to publish my customized APKBUILD (Including some patchs I made), I thought would be make sense if I put my name. If that is wrong, I will leave the maintainer field. 2024-02-10 00:48:23 *won't update the maintainer field. 2024-02-10 00:51:23 yep, that is wrong 2024-02-10 00:53:31 Thanks, everything is clear now. 2024-02-10 00:53:42 you mean publish it somewhere else? not adding the changes to the Alpine repo? 2024-02-10 01:03:18 Yeah, I place it on my repo. 2024-02-10 01:04:32 why not submit the patches to Alpine? 2024-02-10 01:08:41 It's customized patch, for example, I add gaps patch for dwl APKBUILD. I don't think that will appropriate for vanilla package. 2024-02-10 01:12:24 Later. 2024-02-10 06:33:17 If anyone has a moment, I've cleaned up #60506 2024-02-10 06:34:56 I hope to be more regular with my maintenance 2024-02-10 06:35:04 The past two years have been an absolute shitshow haha 2024-02-10 06:39:16 !60506 2024-02-10 07:25:47 Thanks, Ikke 2024-02-10 07:25:49 Forgot the syntax 2024-02-10 07:26:55 ACTION uploaded an image: (87KiB) < https://matrix.org/_matrix/media/v3/download/matrix.org/mbmMvduwDAUOwUOLjgxxvdNe/image.png > 2024-02-10 07:27:03 Some of these I can put away, but... This list depresses me haha 2024-02-10 07:27:16 Aspirations > Skill 2024-02-10 11:26:55 Ermine: re virtualbox guest additions: I use vboxmanage guestproperty get "${VM_NAME}" "/VirtualBox/GuestInfo/OS/LoggedInUsers" (number of logged in users) 2024-02-10 11:35:12 frazooka: ok, thank you. Feature will work on Alpine Edge once my patch gets merged. Do you use any display managers by the chance? 2024-02-10 11:41:35 Ermine: Yes I use SDMM 2024-02-10 11:41:58 Ermine: thanks a lot for the work btw, thats awesome! 2024-02-10 12:01:29 Is wayfire broken an alpine? 2024-02-10 12:01:33 *on 2024-02-10 14:33:27 i dunno, is it? 2024-02-10 15:25:32 Hi, can you possibly help me understand the git's APKBUILD? I can see that git depends on e.g. libcurl (apk info -R), but it's not in the depends variable (APKBUILD). The variable doesn't seem to be set for git. https://git.alpinelinux.org/aports/tree/main/git/APKBUILD?h=3.19-stable 2024-02-10 15:29:49 x-yuri: depends= doesn't list all dependencies of a package. Abuild traces dependencies too by looking at list of libraries which package binaries use 2024-02-10 15:30:26 apk info probably says that git depends on so:libcurl.so or something like that 2024-02-10 15:31:14 That means "git depends on a package that provides libcurl.so" 2024-02-10 15:31:22 so:libcurl.so.4 to be precise. I see, thanks. 2024-02-10 16:19:07 x-yuri: you do see curl-dev in makedepends 2024-02-10 16:24:20 ikke: sure 2024-02-10 16:50:41 hey, which mta is suggested in this distro? 2024-02-10 16:52:58 suggested? none 2024-02-10 16:59:25 for simple stuff I've used dma in the past 2024-02-10 23:40:36 is anyone able to get "*y or "+y to work in vim/wayland 2024-02-10 23:40:51 or is there some secret to get it to work 2024-02-10 23:41:32 hey, i lost colors when i switched to bash, whats up with that? 2024-02-10 23:41:48 i guess i need to create the color stuff in .bashrc to make it work 2024-02-10 23:47:15 Vis resets the terminal on exit 2024-02-11 00:19:55 well? 2024-02-11 00:26:07 well what? 2024-02-11 02:12:46 orchardstreet22: what are you trying to do? 2024-02-11 06:00:04 xgpt: copy/paste from vim in a safe way, ideally using commands like "*y, "+y, "*p, "+p etc 2024-02-11 06:00:33 alpine's vim is compiled with clipboard support but those commands just aren't working for me in Alpine in wayland for some reason 2024-02-11 13:43:31 orchardstreet22: did you install vim explicitly with apk install vim? The "vi" that ships with alpine isn't full vim 2024-02-11 15:23:59 do you have a clipboard setup? 2024-02-11 15:30:26 aquamo4k xulfer: I did apk add vim wl-clipboard 2024-02-11 15:30:57 copypaste works fine with shift+ctrl+V, but that's not as safe as "+p etc 2024-02-11 15:31:25 also vi is aliased to vim in .profile so i'm never running vi 2024-02-11 15:32:47 sorry in .rc, regardless it's always vim 2024-02-11 17:36:21 hey, how do you upgrade from one point release to another? 2024-02-11 17:36:26 eg from 3.19 to 3.19.1 2024-02-11 17:37:01 you upgrade packages 2024-02-11 17:37:25 that's it? and the os release information? 2024-02-11 17:37:27 Juest: You don't have to explicitly upgrade to patch releases 2024-02-11 17:37:31 it still shows 3.19? 2024-02-11 17:37:46 there are not 3.19 and 3.19.1 repos, there is only a "3.19" repo which is continually updated 2024-02-11 17:38:21 there's no 3.19.1 on the getty text right? 2024-02-11 17:38:24 login text* 2024-02-11 17:39:03 i see, os-release says im on 3.19.1 but pretty name is 3.19 2024-02-11 17:39:18 Yes 2024-02-11 17:39:30 Because you are tracking the 3.19-stable branch 2024-02-11 17:39:41 You will always get the latest packages on that branch 2024-02-11 17:39:53 i see, thanks 2024-02-11 17:40:07 The only reason these patch releases are done is to build new docker images and isos 2024-02-11 17:40:20 is it easy to "convert" a install to different "spins"? 2024-02-11 17:40:58 as in, from standard to vm/virt, or from virt to extended 2024-02-11 17:41:20 extended only changes what packages are present on the iso 2024-02-12 03:52:01 'apk update' says its using version v20231219-3633-gfbaec483ee1, but my /var/cache/apk lists the latest APKINDEX (11-Feb-2024) as having been downloaded. I can't, for example, update ffmpeg, as 'apk list' shows "ffmpeg-6.1.1-r3 x86_64 {ffmpeg} [upgradable from: ffmpeg-6.0.1-r1]" but 'apk fix ffmpeg' shows "[APK unavailable, skipped] Reinstalling ffmpeg (6.0.1-r1)" 2024-02-12 03:52:15 https://bpa.st/RGUQ 2024-02-12 03:52:27 Why is there a version mismatch, and how can I fix it? 2024-02-12 04:58:43 bovis: what does `apk upgrade ffmpeg' do? 2024-02-12 09:21:06 hmm i'm packaging pizauth seems that i am getting this error 2024-02-12 09:21:08 https://github.com/ltratt/pizauth/blob/master/Makefile#L28 2024-02-12 09:21:16 install: can't create directory '/usr/local/share/examples/': Permission denied 2024-02-12 09:21:45 https://github.com/dngray/containers/blob/main/pizauth/pizauth/APKBUILD 2024-02-12 09:21:55 should i make the directory first i guess? or should i put it somewhere else 2024-02-12 09:28:08 dngray: you should install to $pkgdir 2024-02-12 09:28:38 Usuallt with DESTDIR="$pkgdir" 2024-02-12 09:30:10 oh yeah, mm 2024-02-12 09:30:24 btw this package is relevant, to anyone who uses mutt aerc or whatever 2024-02-12 09:30:32 tldr is that mailctl is not really maintained 2024-02-12 09:30:49 and author is re-writing it, but i found pizauth which looks to be better designed anyway 2024-02-12 09:30:58 and not in obscure haskell, but rather rust. 2024-02-12 09:31:38 and without hardcoding localhost https://github.com/pdobsan/mailctl/issues/37 2024-02-12 09:32:31 ikke: yeah so i won't do make install ill do install i n the APKBUILD 2024-02-12 09:32:32 instead. 2024-02-12 09:33:13 The Makefile does not support DESTDIR? 2024-02-12 09:33:24 oh wait it does. 2024-02-12 09:35:26 ah here we go https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#pkgdir 2024-02-12 09:38:48 if this works i might submit to aports 2024-02-12 09:38:53 because a lot of people will want this 2024-02-12 09:39:21 and seems to be most security orientated approach i've seen 2024-02-12 09:41:30 https://termbin.com/glm8 hmm what does this check mean 2024-02-12 09:41:55 > > > ERROR: pizauth*: Packages must not put anything under /srv, /usr/local or /opt 2024-02-12 09:41:56 is usr local not used on alpine? 2024-02-12 09:56:45 Not for packages 2024-02-12 09:59:19 so what do i want to change PREFIX to then? 2024-02-12 09:59:26 just /usr? 2024-02-12 10:00:48 Yes 2024-02-12 10:01:40 sorry for the newb questions been a while since i used alpine/packaged for it 2024-02-12 10:02:15 Don't worry 2024-02-12 11:07:11 ikke: that makefile installs some man files https://github.com/ltratt/pizauth/blob/master/Makefile#L24-L27 resulting in an error like https://termbin.com/rv1j 2024-02-12 11:07:21 i get this should be in a pizauth-doc subpackage 2024-02-12 11:07:35 in a case like that i guess i have to manually patch those lines out of the makefile 2024-02-12 11:08:24 oh or i guess i could set MANDIR 2024-02-12 11:08:38 though i think that's still not going to put it into the subpackage 2024-02-12 11:09:14 https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#doc() 2024-02-12 11:10:40 https://wiki.alpinelinux.org/wiki/APKBUILD_examples:Special_Subpackages_Doc maybe is what i want 2024-02-12 11:18:44 dngray: yes, just adding a $pkgname-doc subpackage is enough, abuild takes care of the rest 2024-02-12 11:19:03 oh so i don't need a doc function? 2024-02-12 11:19:08 nope 2024-02-12 11:19:26 There is a default implementation 2024-02-12 11:23:19 seems you're right, it worked 2024-02-12 12:02:23 How does alpine handle packages that aren't maintained anymore? For example dansguardian. Even debian doesn't package it anymore 2024-02-12 12:30:43 I guess you mean not maintained upstream anymore 2024-02-12 12:33:22 wrt dansguardian I've looked at it too, a while ago, perhaps it could be suggested for removal or at least move to community 2024-02-12 12:33:42 there's also e2guardian, the dansguardian fork 2024-02-12 12:39:33 sertonix: btw, your "main/*: add missing !check option" commit didn't say why it was missing, perhaps the MR did, but I assume it was for aports that didn't have a check() 2024-02-12 12:40:39 the way dansguardian has been handled the past ten yers, since its last upgrade, is making sure it builds 2024-02-12 13:02:18 omni: The e2guardian web page seems to be down though 2024-02-12 13:03:37 With the commit it is as you assumed. Probably should have explained it in the commit message 2024-02-12 13:18:46 I'm trying to build scap-security-guide with the help of durrendal: https://gitlab.com/Durrendal/WS-Alpine-Packages/-/blob/master/testing/scap-security-guide/APKBUILD?ref_type=heads I get an error with abuild I think? https://termbin.com/smgy /usr/bin/abuild: cd: line 14: can't cd to scap-security-guide-0.1.71: No such file or directory 2024-02-12 13:21:57 In a directory where I'm running abuild -r I see: https://termbin.com/ri7i 2024-02-12 13:22:26 so there is src and inside scap-security-guide-0.1.71 directory 2024-02-12 13:33:00 frojnd: could you try adding builddir="$srcdir/scap-security-guide-v$pkgver" ? (Only a guess though) 2024-02-12 13:34:13 Oh, the issue is the firts cd in the build() function. Try removing it 2024-02-12 14:17:40 sertonix: I think you are right, I had to put pwd, inside and I'm already inside src/scap-security-guide-v$pkgver 2024-02-12 14:21:49 Ok package is done. One question though. I've built it on edge (inside qemu) can I now install it on stable? 2024-02-12 14:23:45 Let just try it 2024-02-12 14:24:21 Was celebrating too quickly: -- Configuring incomplete, errors occurred! 2024-02-12 14:29:56 Allright, waiting for build to be complete 2024-02-12 14:30:31 How can I squeeze a bit more from host computer? Currently running guest alpine like this: qemu-system-x86_64 -m 8G -enable-kvm -machine q35 -device intel-iommu -cpu host -nic user,hostfwd=tcp::60022-:22 -hda 2024-02-12 14:30:39 More CPU power... 2024-02-12 14:52:20 https://wiki.alpinelinux.org/wiki/How_to_make_a_cross_architecture_chroot 2024-02-12 14:53:03 frojnd: i use this 2024-02-12 15:19:08 zcrayfish: apk upgrade ffmpeg gives me this: https://bpa.st/D7BQ 2024-02-12 15:23:08 apk add -s ffmpeg=6.1.1-r3 2024-02-12 15:26:47 ikke: I get conflicts https://bpa.st/LG6Q 2024-02-12 15:27:22 qaqland: interesting 2024-02-12 16:01:22 bovis: Which alpine version are you using? It seems like gdal needs to be upgraded/rebuild on that one 2024-02-12 16:02:56 ikke: Is there any way to force the use of a specific APKINDEX? https://bpa.st/RGUQ makes me think I'm being held back in some way. As far as I know, my system is up to date concerning 2023-12-19 information, but I did the full apk upgrade yesterday. 2024-02-12 16:03:27 sertonix: Version is edge 2024-02-12 16:03:45 bovis: it has nothing to do with what apk index is used 2024-02-12 16:04:06 You have packages installed that require older versions 2024-02-12 16:04:55 Is there a way to bring up the full list of those that require older versions? 2024-02-12 16:06:03 I might be running into issues with orphaned packages. 2024-02-12 16:09:13 maybe try apk -s upgrade -a -l --prune 2024-02-12 16:09:39 -s is to only see what would happen 2024-02-12 16:09:47 apk list --orphan 2024-02-12 16:09:55 bovis: what arch are you using? 2024-02-12 16:10:12 ikke: x86_64 2024-02-12 16:11:12 Is orphan everything I've installed that requires older dependencies, or everything locally I've installed that isn't from the edge repository? 2024-02-12 16:11:33 any package that does not exist anymore on any repository 2024-02-12 16:12:37 bovis: try apk add -s gdal=3.8.3-r3 2024-02-12 16:13:30 sertonix: Simulation is https://bpa.st/LT4A 2024-02-12 16:16:01 ikke: https://bpa.st/WMKQ gdal-3.8.3-r3 2024-02-12 16:17:20 bovis: and apk list --orphan? 2024-02-12 16:17:45 It seems that gmt is no longer provided 2024-02-12 16:19:23 I don't see that we ever had a package called gmt 2024-02-12 16:20:11 bovis: did you build this yourself? 2024-02-12 16:20:16 ikke: Yes. That's a local orphan. 2024-02-12 16:22:02 So either rebuild it yourself, or remove it 2024-02-12 16:22:29 I'll do that. 2024-02-12 16:24:00 Furthering the issue, how difficult would it be for me to become the maintainer of that package and put it into edge? I'm unsure of whether or not it could be built for other architectures, but I only have access to my own laptop. Am I equippped to maintain something? 2024-02-12 16:25:01 bovis: sure, we have at least our CI to catch build issues on other arches 2024-02-12 16:25:54 CI? 2024-02-12 16:26:23 Continuous integration 2024-02-12 16:26:40 jobs that build the package on various arches when you create an MR on gitlab 2024-02-12 16:37:50 Removing gmt worked, in that I was able to upgrade ffmpeg and a few others. I still see the December timestamp here though. https://bpa.st/RGUQ Why is that? 2024-02-12 16:38:37 Yet more orphans holding back something that I'm just unaware of? 2024-02-12 16:38:39 That's the latest snapshot 2024-02-12 16:38:54 It's not the date of the index or something 2024-02-12 16:39:22 (the last git tag that was created on edge) 2024-02-12 16:39:24 Oh. I assumed it should match http://dl-cdn.alpinelinux.org/alpine/edge/main/ which is 2024 2024-02-12 16:40:23 What you see is the git describe output, which is 3633 commits since the tag v20231219 was created 2024-02-12 16:40:43 It has nothing to do with when the index itself was created 2024-02-12 17:12:15 hmmm, plasma install first install did not go well, will retry(after reading docs if any) see what i missed 2024-02-12 17:13:40 i use kde for main desktop, so i would take a chunk of dark-chocolate and retry so that my experience with al remains ok 2024-02-12 17:15:17 sddm give 2 options to use i.e X11 and wayland. wayland hangs, X11 atleast starts but something not ok 2024-02-12 17:16:37 1. i think screen refresh issue, 2. cannot minimize/close apps 2024-02-12 17:18:20 testing on 1st gen, i5 laptop with 8gb ram 2024-02-12 17:18:49 so far xfce4, gnome installation went nice 2024-02-12 17:36:45 xinit wont't start 2024-02-12 17:48:11 vkrishn: "startx" disappeared for me the other day after an update. try apk add xinit ? 2024-02-12 17:49:59 ok 2024-02-12 17:50:17 already installed 2024-02-12 17:55:34 installed xterm 2024-02-12 17:57:09 Pastebin the error. I don't know if I can help. 2024-02-12 18:03:56 oh my xterm, now switching session back to Wayland in sddm works 2024-02-12 18:04:23 let me trace what i did simingly superb 2024-02-12 18:05:22 cannot see X(close) icon in apps like firefox, rest seems ok 2024-02-12 18:06:44 i don't like dark-chocolate somuch anyways, so keeping it aside 2024-02-12 18:10:18 after setup-desktop plama, apk add plasma xterm 2024-02-12 18:13:53 setup-devd udev 2024-02-12 18:14:21 apk add dbus dbus-x11 2024-02-12 18:14:41 /etc/init.d/dbus restart 2024-02-12 18:15:09 /etc/init.d/sddm restart 2024-02-12 18:15:37 then select Wayland, then login 2024-02-12 18:17:30 and all is running from ram, du -csh /usr - 1.7Gb 2024-02-12 18:19:51 now rebooting to see if it can be re-done 2024-02-12 18:22:57 So close: https://termbin.com/wv1u >>> ERROR: scap-security-guide*: Found uncompressed man pages: I think warning about missing -doc and error are connected? I'm just doing it by abuild -r 2024-02-12 18:34:44 frojnd: Does your APKBUILD have a line for subpackages="$pkgname-doc" ? 2024-02-12 18:52:18 kinda funny, i need to first login using session X11, logout then login session Wayland 2024-02-12 18:55:12 don't need to add dbus, dbux-x11, xinit as they are installed by setup-desktop 2024-02-12 19:00:05 mpv plays nice, temprature rise 4+, cpu usage reasonable 2024-02-12 19:26:35 lid close, lid open, login ok 2024-02-12 19:28:12 generating, apk list -I, will add to https://git.insteps.net/mess/apklist/ soonish 2024-02-12 19:28:48 looking forward to use kde/plasma 2024-02-12 19:54:49 apk add yakuake (nice) 2024-02-12 20:13:00 how do i change console fonts? 2024-02-12 20:16:53 setfont command or consolefont openrc service 2024-02-12 20:17:34 clicking on "virtual keyboard" on sddm login screen does not do anything 2024-02-12 20:19:15 Ermine: but i dont have the consolefont files, do i? 2024-02-12 20:21:08 Either you do, and they are in /usr/share/consolefonts, or you should install some 2024-02-12 20:21:43 For example, font-terminus package 2024-02-12 20:31:16 how do i remove dracut without removing grub-efi and linux-lts? 2024-02-12 20:32:28 you install some other initramfs creation alternative? 2024-02-12 20:33:12 you need some tool to create initramfs... 2024-02-12 20:33:32 minimal: installing dracut removes mkinitfs? 2024-02-12 20:33:53 install mkinitfs then 2024-02-12 20:34:43 I don't remember if it actually removes mkinitfs, I think so 2024-02-12 20:34:52 yeah it seems to have done that 2024-02-12 20:35:12 anyway both dracut and mkinitfs "provide" the same virtual package required by grub, kernel etc 2024-02-12 20:35:18 i see 2024-02-12 20:35:30 i dont have any man pages for some reason, i need dracut-docs? 2024-02-12 20:35:46 i see its all separated 2024-02-12 20:35:58 I thought you wanted to *remove* dracut 2024-02-12 20:36:50 now i have both packages installed 2024-02-12 20:37:15 dracut and dracut-docs? 2024-02-12 20:37:22 dracut and mkinitfs 2024-02-12 20:37:45 so then they don't remove each other 2024-02-12 20:37:47 how do i query what dependencies a package has installed? 2024-02-12 20:37:54 you must have earlier removed mkinitfs manually 2024-02-12 20:37:57 dracut removed mkinitfs 2024-02-12 20:38:01 i didnt remove mkinitfs 2024-02-12 20:38:06 but mkinitfs didnt remove dracut 2024-02-12 20:41:05 I just installed dracut on a Edge VM and it didn't remove mkinitfs 2024-02-12 20:45:30 weird 2024-02-12 20:45:46 idk how or why it did get removed, i didnt remove it 2024-02-12 20:48:39 i tried to remove dracut and it might have removed mkinitfs 2024-02-12 20:48:41 idk, very weird 2024-02-12 20:48:48 maybe mkinitfs wasnt installed explicitly 2024-02-12 20:49:18 how do i remove a explicity package without removing it because its a dependency? 2024-02-12 20:49:50 package that i installed explicitly 2024-02-12 20:51:59 apk wouldn't allow you to remove dependency of another package 2024-02-12 21:12:17 will alpine switch to dbus-broker 2024-02-12 21:16:44 sewn: I asked a while back, because it is claimed to be more performant, but I can not find good literature on that 2024-02-12 21:17:17 sewn: dbus-broker is in packages, just install it. Alpine does not install dbus by default. 2024-02-12 21:17:26 I have also not tried building/packaging it, so it may be up to us to prove it out first before asking for such a change 2024-02-12 21:17:47 Ermine: Whaa, since when?! 2024-02-12 21:19:28 Well, it's in testing since last August 2024-02-12 21:20:12 You can try filing a MR to move it to community 2024-02-12 21:25:36 Mmm... i'm not doing so hot with the packages I have, Community or Main doesn't sit well with me yet 2024-02-12 21:26:17 I wouldnt have been able to upgrade libopenraw without much hand-holding and a patch from Celeste 2024-02-12 23:58:53 i noticed the other day that alpine openrc scripts are a little more on the basic side compared to gentoo 2024-02-13 04:05:15 Ermine: dbus-daemon is provided by dbus which packages depend on 2024-02-13 09:31:11 durrendal: when you are here I've queried you... 2024-02-13 12:56:06 does apk info -W check for local installed packages or all repositories? 2024-02-13 12:56:45 Only installed packages 2024-02-13 12:56:59 You cannot use it to find what package to install to get a specific file 2024-02-13 13:02:33 Anyone have a good link for zfs tuning a harddrive array? 2024-02-13 16:25:34 hi there. is it possible to install gettext 0.18.3. In Alpine Linux the version is 0.22.3 but a project requrires 0.18.3 2024-02-13 16:38:05 olku: only if you build it yourself and pin it 2024-02-13 17:01:15 thanks 2024-02-13 20:24:42 i was trying to limit access to su by adding "auth required pam_wheel.so trust group=wheel" to /etc/pam.d/su so only users who are in the wheel group can escalate. i can still become root via su as a user without wheel though. do i need to do more? 2024-02-13 20:29:05 su from which Alpine package? Busybox su is not PAM-enabled 2024-02-13 20:29:53 well i guess it's busybox 2024-02-13 20:29:58 su can be provided by either busybox, shadow-login, or util-linux-login packages 2024-02-13 20:30:10 cits: guess? you can check 2024-02-13 20:30:29 e.g. what does "su --help" indicate? 2024-02-13 20:30:40 yeah it is, busybox. which one of the alternative should i pick? 2024-02-13 20:30:54 whichever one you want 2024-02-13 20:31:02 no downsides/pitfalls? 2024-02-13 20:31:21 that depends on what you're doing/looking for/etc 2024-02-13 20:32:16 i just wanted to harden a system and thought i'd be nice for my "service account" to not be able to become root, so nothing serious 2024-02-13 20:32:18 installing either shadow-login or util-linux-login will also replace some other binaries such as login 2024-02-13 20:33:57 alright, that did the job, thanks :) 2024-02-13 21:32:49 Hi all, I just finshed building a new box, with a data-install alpine nas server and it appear in.tftpd loads but not listening. netstat -an any suggetions? iptables is off 2024-02-13 21:35:12 txnintn: ps shows it is running? 2024-02-13 21:37:59 rc-status -a shows it run checking ps now 2024-02-13 21:39:15 yes 3 from the bottme. 2024-02-13 21:44:12 and for netstat you checked for 69/udp ? 2024-02-13 21:44:51 e.g. "netstat -an | grep udp |grep 69" 2024-02-13 21:44:58 I did a netstat -an 2024-02-13 21:46:00 the listening column at udp 69 is blank 2024-02-13 21:46:23 sorry the state column 2024-02-13 21:46:51 and the Local Address column? 2024-02-13 21:47:07 :::69 2024-02-13 21:47:39 that's a IPv6 listen 2024-02-13 21:47:49 sounds like it is listening for IPv6 only 2024-02-13 21:48:25 e.g. looking on a VM here for sshd there are 2 TCP entries: 0.0.0.0:22 (IPv4) and :::22 (IPv6) 2024-02-13 21:48:33 That's not going to work. not running ipv6 2024-02-13 21:49:09 well you much have IPv6 (at least on lo0) otherwise it couldn't listen 2024-02-13 21:50:36 I've been on a learning curve for 3 month or so, too old for ipv6 haha hah 2024-02-13 21:52:04 was does "ps" show for the tftpd process? 2024-02-13 21:52:58 s/was/what/ 2024-02-13 21:55:38 15502 root 0:00 /usr/sbin/in.tftpd -l -R 4096:32767 -s /bluehole/src/tftp/ 2024-02-13 21:56:06 and there are no messages for tftp in /var/log/ files? 2024-02-13 21:57:24 I install alpine acf and it's healthchecking me to death in messages 2024-02-13 21:58:35 at the bottom deamon.warn tftp: client does not accept options 2024-02-13 21:59:16 so that shows a TFTP client actually talked to tftpd, doesn't it? 2024-02-13 21:59:51 before that user.debug : will stop /usr/sbin/in.tftpd 2024-02-13 22:00:14 re: "the listening column at udp 69 is blank" - yes, it is UDP, not TCP-based, so state column is blank 2024-02-13 22:01:04 then deamon.warn sending signal 15 to pid 6096 2024-02-13 22:01:13 you can add a "-v" option to the /etc/conf.d/tftpd setting to try and get more info when you (re)start it 2024-02-13 22:09:50 backup, i could throw in a --ipv4 in there as well 2024-02-13 22:11:45 ok on the netstat, we got 0.0.0.0:* now 2024-02-13 22:13:04 did the ~"-v" show anything useful? (like it logging the address(es) it listens on?) 2024-02-13 22:16:26 yes, getting RRQ from 192.168.10.57 (rpi4) and listing a bunch of file names 2024-02-13 22:18:55 I have to fix the rpi, the last bootloader change killed the boot at rainbow and 7 flash on led.. 2024-02-13 22:20:30 you got it working using "--ipv4"? or it worked without that? 2024-02-13 22:22:13 the address are after the changes but the pi is not advancing, I'll try it again, 2024-02-13 22:24:18 it's trying, I jumped out of message and went back it, fresh rrq from 192.168.10.57 two minutes ago. 2024-02-13 22:28:58 It's pulling from the serial number directory in tftp, the file names are repeating, so is the client asking and files not sending? 2024-02-13 22:30:58 "that would be a TFTP matter" 2024-02-13 22:33:20 Let call it good for now, I'm going to we do the bootloader and try again. I would love to get the pi's all running diskless with a sd backup. 2024-02-13 22:33:57 Thanks for the help. 2024-02-13 22:36:47 bye have a good night. 2024-02-14 04:02:56 In the busybox syslog, where do you tell it to sent logs to a remote server? 2024-02-14 05:50:00 raub: all you need is written on https://wiki.alpinelinux.org/wiki/Syslog 2024-02-14 05:50:02 Configuration 2024-02-14 05:50:05 Edit /etc/conf.d/syslog to change the options used when running syslogd. All available options can be looked up with syslogd --help. 2024-02-14 05:50:08 -R HOST[:PORT]Log to HOST:PORT (default PORT:514) 2024-02-14 07:29:34 Has anyone built an alpine-netboot tarball with linux-edge from community as the kernel? How difficult would that be? I want to put Alpine on a spare machine and play around with bcachefs, which isn't in linux-lts. 2024-02-14 10:38:16 aron: thanks for the help, You made me find out the real issue: look at line 5 in https://paste.centos.org/view/b169b55c 2024-02-14 15:35:00 hey all! How do I simply restart networking without doing a reboot? 2024-02-14 15:38:06 service networking restart 2024-02-14 15:38:14 ugh, so I'm trying to *enable* ipv6 2024-02-14 15:38:46 I've added face eth0 ine6 auto to my confi 2024-02-14 15:38:48 I've added face eth0 ine6 auto to my config 2024-02-14 15:38:56 but it keeps the local address it already had 2024-02-14 15:39:20 what address you get on dhcp is up to your router 2024-02-14 15:39:24 oooh 2024-02-14 15:39:29 that's stateless config 2024-02-14 15:41:10 apparently I need to install dhcpcd? 2024-02-14 15:43:25 xgpt: what do you want to achieve? 2024-02-14 15:43:35 You dont need a separate stanza for ipv6 2024-02-14 15:45:30 lemme share my interfaces 2024-02-14 15:47:32 ikke: https://0x0.st/Hddk.txt 2024-02-14 15:47:48 I want to simply enable ipv6, it's running on a VPS which should be given an ipv6 from my host 2024-02-14 15:49:26 You do see a link local address? 2024-02-14 15:49:40 Starting with fe80 2024-02-14 15:49:45 IPv6 autoconf is enabled by default in the kernel 2024-02-14 15:50:02 Question is if you have somebody else giving router advertisements at the other end 2024-02-14 15:50:04 yes I see an address with fe80 2024-02-14 15:50:13 Then ipv6 is enabled 2024-02-14 15:50:14 it ought to , it does with my other VPS 2024-02-14 15:50:20 but is fe80 public? 2024-02-14 15:50:28 So you have docker installed 2024-02-14 15:50:29 No, it's link-local 2024-02-14 15:50:38 Do* 2024-02-14 15:50:40 I don't think I've got docker installed on that vps 2024-02-14 15:50:43 no, I don't' 2024-02-14 15:52:29 What does cat /proc/sys/net/ipv6/conf/eth0/accept_ra return? 2024-02-14 15:52:50 ikke: 1 2024-02-14 15:53:05 And forwarding (in the same dir) 2024-02-14 15:53:32 ikke: 0 2024-02-14 15:53:49 So far so good 2024-02-14 15:53:56 :) 2024-02-14 15:54:32 do I need dhcpcd? or is the udhcpc good enough? 2024-02-14 15:55:07 No, not if your router does support SLAAC 2024-02-14 15:55:08 xgpt: for slaac you don't need either 2024-02-14 15:55:26 But if your router actually only does DHCPv6, then yes you need a DHCPv6 client 2024-02-14 15:55:32 depends on what the vps provider allows 2024-02-14 15:57:05 oh woooow 2024-02-14 15:57:11 rack nerd doesn't enable ipv6 by default 2024-02-14 15:57:36 i myself explicitly set ipv6.disable=1 :) 2024-02-14 15:58:14 🤨 2024-02-14 15:58:29 Must be some miscomprehension 2024-02-14 15:58:44 duly noted 2024-02-14 15:58:46 this makes sense 2024-02-14 15:59:05 my server isn't going to pick up an ipv6 if the router isn't giving them out 2024-02-14 15:59:26 :) 2024-02-14 18:18:52 is it normal for dnsmasq to listen to port 67, https://tpaste.us/6MwB ? 2024-02-14 18:19:20 That's dhcp 2024-02-14 18:19:31 and dnsmasq is a dhcp server as well 2024-02-14 18:19:44 ok, thanks, i have a client calling it 2024-02-14 18:21:13 via - udhcpc -i wlan0 -B -qn 2024-02-14 18:21:22 thanks 2024-02-15 00:07:34 I am trying to build a custom package with `abuild rootbld`. It can download the sources and install the dependencies correctly, but then there is some resolve error when running cargo (it's a Rust project). Could it be that `abuild rootbld` doesn't give it access to the internet? If yes, can I enable it for the build? 2024-02-15 00:12:20 jonesv: https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#options the "net" option 2024-02-15 00:12:36 but yeah, by default there is no network access during build 2024-02-15 00:14:14 elly: ach and I had just read that section about the !check option, but I missed the net one. Thanks! 2024-02-15 00:17:38 I am making a package for the Rust plugins of gstreamer, which is an interesting exercise because it is in Rust. The `Cargo.lock` has 628 (!) dependencies. So it raises the question: am I expected to make a package for each dependency (just like for Python there are `py3-*` packages), or am I expected to link those 628 dependencies statically (the default in Rust)? 2024-02-15 00:18:31 Because, among others, the issue if I link them statically is: what do I do with the licenses? Should I add all the licenses of all the dependencies to that package? And for those that require to copy the license, should I copy all of them in a `-doc` package? 2024-02-15 00:28:54 That is an interesting problem 2024-02-15 00:29:32 Is it customary to ship licenses separate from the binaries in a -doc package? Does that violate the license? 2024-02-15 00:57:10 jonesv: hmmm, that's a good question... I went and looked at community/chromium which also has many dependencies under different licenses, a lot of which are vendored 2024-02-15 00:57:34 and it looks like that APKBUILD just uses the license of the chromium package itself (so BSD-3-Clause) and doesn't list the dependency licenses 2024-02-15 00:57:53 although for chromium, all the dependency licenses are available at runtime anyway 2024-02-15 00:58:41 I think that that approach makes sense... if you try to explain all 628 dependency licenses in the APKBUILD it will be illegible 2024-02-15 01:01:43 I think you will realistically have to link them statically, because if the dep graph is so large, there's no way you will avoid version conflicts between deps 2024-02-15 01:01:50 against other packages on the system 2024-02-15 01:05:43 (also, the idea of 628 dependencies is somewhat... overwhelming) 2024-02-15 06:25:23 How does alpine want udhcpc to integrate with wpa_supplicant? As I see it now there is a wpa_cli.sh but its never invoked? I keep manually haveing to run udhcpc -iwlan0 on boot for the network to function, its getting old, should I just install dhcpd from apk instead or is there a proper way to get the 2 to function properly together 2024-02-15 06:28:04 MaddieKalan: that's taken care of by ifupdown-ng (in /etc/network/interfaces) 2024-02-15 06:29:07 iface wlan0 2024-02-15 06:29:09 use dhcp 2024-02-15 06:48:29 Thank you, ikke. But I wonder, is this documented in the install process, I mean I've been switching my laptops from void to alpine so I've just installed it on another one and I didn't notice anything about editing /etc/network/interfaces, anyway I apreciate the help :) 2024-02-15 06:53:44 MaddieKalan: if you use setup-interfaces (which setup-alpine executes), then it would set it up for you as it should 2024-02-15 06:57:16 Weird, I did install the system with setup-alpine but it hasn't set it up for me? My best guess is it might have something to do with buggy (not sure if its a bug or on purpose) behaviour I found where it doesn't ask me for a WiFI password so I have to switch TTY and configure wpa_supplicant manually and restart it for networking to work. 2024-02-15 07:24:28 Is there a pre-built ARM image available for aarch64? 2024-02-15 07:24:53 Trying to install Alpine in a VM on an Apple M3 machine (not supported by Asahi yet) and I'd like to use UEFI if possible 2024-02-15 08:33:00 PureTryOut: I hope this is useful https://dl-cdn.alpinelinux.org/alpine/v3.19/releases/aarch64/ 2024-02-15 08:33:38 You probably want to grap the one tagged with virt. 2024-02-15 08:50:23 I already tried that one obviously 😉 Doesn't seem to do UEFI, or at least VirtualBox doesn't pick it up 2024-02-15 08:55:44 Oh nvm it seems to be a VirtualBox issue. I just tried virt-manager (qemu, libvirt) and it actually boots there 2024-02-15 09:46:55 SOrry I feel asleep at my desk, great to hear everything worked out :) 2024-02-15 10:08:06 Hi people Please help! since yesterday I'm trying to create a new VM with alpine-linux 3.19.1 (virtual, x64). I go through all the questions of setup-alpine as usual, point to dl-cdn.alpinelinux.org/alpine/v3.19/main ... but when I get to the final step I select "sda" and then I get a message about missing syslinux dependency. I thought it was a problem with 3.19.1, but I just retried 2024-02-15 10:08:06 with 3.19.0 and the problem is intermitent... Is the problem dl-cdn? Should I be using another mirror for installation? 2024-02-15 10:09:14 xgpt: I think it is fine to ship the licenses in a `-doc` package. The problem was more that if you have 628 of them, then the package doesn't make much sense anymore (and is difficult to maintain) 2024-02-15 10:11:08 elly: haha yeah, that many dependencies is a problem. But I think it is a consequence of the popularity of "language package managers" (in this case cargo) that make it super easy to just not think about dependencies at all... 2024-02-15 10:17:12 elly: The chromium case is interesting. I think it's easier for executables to statically link dependencies. Actually it probably make sense for an executable to have a mix between shared and static dependencies (e.g. use openssl from the system, but statically link this small dependency that nobody knows). But for a library it feels wrong... 2024-02-15 10:30:37 never mind, I tried http://dl-cdn.alpinelinux.org/alpine , dropping the v3.19/main part and it worked 2024-02-15 10:30:40 see ya! 2024-02-15 10:31:10 This is sad, is there really no aalib on alpine? Guess I can become a contributer! 2024-02-15 11:23:21 there used to be 2024-02-15 11:24:00 MaddieKalan: it was removed because it was unmaintained 2024-02-15 11:25:21 MaddieKalan: last version: https://gitlab.alpinelinux.org/alpine/aports/-/blob/b4a13e7314bfe25e63b2f1ae5014d5429c17b539/testing/aalib/APKBUILD 2024-02-15 11:28:03 I need an advice. I have a 65GB source file project and am wondering what's fastest method to search for a function. It's java/c++/c/bash project and currently I'm recursively grepping for a function through thousands and thousands of files... 2024-02-15 11:28:31 I could limit search to *.java that's for sure 2024-02-15 11:28:34 build an index 2024-02-15 11:28:58 as long as you're searching individual files each time, it's not going to get faster 2024-02-15 11:29:27 grawity: how can I build an index? (don't know the prcedure/meaning of that) 2024-02-15 11:29:28 but something like https://github.com/google/codesearch would be a one-time background task 2024-02-15 11:29:39 Thanks 2024-02-15 11:29:40 there's no single procedure other than "you use a program that does so" 2024-02-15 11:30:02 ctags as well 2024-02-15 11:30:03 cindex (the one I linked) is no longer maintained, but it used to work well for me 2024-02-15 11:39:03 ikke: What do you meany by unmaintained? The last update to aalib was in ~2001, the software is a finished product. 2024-02-15 11:40:01 MaddieKalan: the alpine package 2024-02-15 11:40:33 I will check the link you posted to see what you mean. 2024-02-15 11:42:22 Ah I see, alpine was on rc4, would the package be accepted if I updated it to rc5? 2024-02-15 11:51:38 MaddieKalan: I think so, yes 2024-02-15 11:52:08 The problem was that it remained in testing and was not touched for >6 months 2024-02-15 11:52:20 the expectation is that packages will not live forever in testing 2024-02-15 11:59:24 I mean it seems like the package works fine in other distros, so I suppose updateing should be fine? I'll look into becomeing a contributer. 2024-02-15 12:05:32 ifupdown-ng can into wpa_supplicant, but this feature is unused for some reason... 2024-02-15 12:06:39 How do I configure nftables so that a port where no process is listening DROPs instead of REJECTs? 2024-02-15 12:07:01 E.g.: I don't want to unconditionally REJECT; if a process is listening I want a normal SYN/ACK, but otherwise DROP. 2024-02-15 12:21:20 Is there a specific reason that not all keys are in /etc/apk/keys? 2024-02-15 12:24:30 WhyNotHugo: You can set the chain policy to drop 2024-02-15 12:25:34 I like to have a policy of drop for my input and forward chain with specific accept/reject where necessary 2024-02-15 12:26:34 Perhaps that does not answer your question actually 2024-02-15 12:28:51 MaddieKalan: something not updated since 2001 probably will have basic issues building (new cpu arch, stuff like that) so might save yourself some time by just looking for patches elsewhere 2024-02-15 12:30:16 (i didn't look at the old package, just saying) 2024-02-15 12:31:44 invoked: I am no alpine expert, but I belive alpine has a arch taging system, so that the package can be available only for x86_64 and i386, other archs can be gradually added as testing has been completed. 2024-02-15 12:32:13 (That way its only build for supported cpu arches) 2024-02-15 12:32:57 i'm saying something different. source from 2001 doesn't know about amd64 probably 2024-02-15 12:33:11 as one example 2024-02-15 12:33:20 bananicorn: I can't figure out how that work, do you have any pointers for me? 2024-02-15 12:33:44 WhyNotHugo: You mean setting a chain policy? 2024-02-15 12:34:04 A chain policy that replaces the outgoing REJECT with DROP. 2024-02-15 12:34:32 TBH, I don't fully understand chains, only rules 2024-02-15 12:35:26 you change the policy part of the chain 2024-02-15 12:35:29 At the start of the chain, there is a line or two that specify which hook the chain is attached to, eg. input/output/forward with which priority, followed by the policy 2024-02-15 12:36:02 I wrote a little article about nftables if you want a basic overview of how its rules work 2024-02-15 12:36:08 https://regrow.earth/diy-server/nftables-01-basics/ 2024-02-15 12:36:15 Perhaps that can help 2024-02-15 12:36:44 Basically you want "policy drop" instead of "policy reject" 2024-02-15 12:48:06 bananicorn: Thanks, that clarified a lot of concepts for me. 2024-02-15 12:48:22 Glad to hear, np :) 2024-02-15 12:48:24 For my specific issue, I'd need to create a rule in 'output' chain, right? 2024-02-15 12:48:56 (the `output` chain defined in Alpine's default /etc/nftables.nft) 2024-02-15 12:53:25 If you want to apply it to outgoing packets then yes 2024-02-15 13:20:21 bananicorn: Thanks for all the guidance :) 2024-02-15 13:22:16 np! 2024-02-15 16:22:51 I'm creating a package and during the build process I am seeing this: WARNING: Datastream component 'scap_org.open-scap_cref_pub-projects-security-oval-suse.linux.enterprise.12.xml' points out to the remote 'https://ftp.suse.com/pub/projects/security/oval/suse.linux.enterprise.12.xml'. Use '--fetch-remote-resources' option to download it. and then it skips the file... how can I 2024-02-15 16:22:53 dynamically fetch those files? This is APKBUILD so far: https://termbin.com/c3cs 2024-02-15 16:23:19 Note that package builds just fine but when using the program I get messages realted to that worning 2024-02-15 16:25:36 I see there is fetch for APKBUILD: This function is called to download the remote files in source. 2024-02-15 17:07:26 `service XXXX reload` shows coloured output when I ssh into my laptop... but not on interactive terminals (foot). Any ideas why this is? 2024-02-15 17:07:41 frojnd: Add the URL to sources= 2024-02-15 17:07:45 *source= 2024-02-15 17:11:34 i've noticed the colours sometimes, too. haven't bothered to look into it 2024-02-15 19:01:18 WhyNotHugo: i run foot + tmux and it works inside tmux. dunno. that's funny, though 2024-02-15 19:01:57 WhyNotHugo: there are multiple urls during the build process... 2024-02-15 19:02:16 WhyNotHugo: if you referred to `'https://ftp.suse.com/pub/projects/security/oval/suse.linux.enterprise.12.xml` 2024-02-15 20:52:33 frojnd: yeah, you add every url into source=, separated by space 2024-02-15 20:52:41 invoked: but not _outside_ tmux? 2024-02-15 21:10:54 nope. 2024-02-15 21:11:26 didn't notice before because colors don't matter too much, to me 2024-02-15 21:47:49 I just found out a weird thing 2024-02-16 02:09:04 same with the colors, i think when i switched from ash to bash i lost colors? 2024-02-16 02:09:17 well service colors work 2024-02-16 02:09:31 thats openrc sending colors and the terminal has to support it 2024-02-16 02:24:15 It just surprised me to ssh into my laptop and see coloured output for 'service'. 2024-02-16 02:24:25 The ssh session was running zsh, same as my usual shells. 2024-02-16 02:25:28 Huh, apparently it's `export TERM=xterm-256color` that makes openrc use colour. 2024-02-16 02:25:32 TERM=foot won't. 2024-02-16 02:34:07 there's also foot-extra in terminfo but same thing. 2024-02-16 02:35:02 i know foot (as per the faq on codeberg) supports a bunch of nonstandard stuff in its terminfo 2024-02-16 02:36:57 also it's not just the colors missing, it doesn't print the status on the right side (rc-service) 2024-02-16 09:13:42 Hm I've tried to add extra sources: https://termbin.com/9xax and then ran: apkbuild checksum and it downloads all sources except: >>> scap-security-guide: Fetching https://people.canonical.com/~ubuntu-security/oval/com.ubuntu.xenial.cve.oval.xml wget: server returned error: HTTP/1.1 404 Not Found and then checksum didn't match... because that url is actually 404 :s 2024-02-16 09:15:22 Will try and build without that ubuntu url and see if oscap will actually give anything else t than notaplicable for result 2024-02-16 10:07:45 I guess adding those sources to source isn't enough (APKBUILD) https://termbin.com/hpzx I still get those warning when building: https://termbin.com/2703 and when trying to run a test I still get notapplicable: doas oscap xccdf eval --profile xccdf_org.ssgproject.content_profile_rht-ccp --results-arf arf.xml --report report.html 2024-02-16 10:07:47 /usr/share/xml/scap/ssg/content/ssg-rhel7-ds.xml gives me: https://termbin.com/f3y3 2024-02-16 12:39:13 frojnd: You probably need to `mv` the file into the place where the build script expects it. 2024-02-16 13:11:54 WhyNotHugo: good thinking, now I have to find out where is that place :) 2024-02-16 14:04:01 I've tried to put it inside src/scap-security-guidee-0.1.71/build/ and inside src/scap-security-guidee-0.1.71/ but it's not recognized 2024-02-16 14:04:04 Still same warning 2024-02-16 17:52:38 hi. Is there a way to pass discard option to cryptsetup/luksin alpine via kernel command line? 2024-02-16 17:57:31 there is no /etc/crypttab 2024-02-16 17:59:04 cryptdiscards=yes 2024-02-16 17:59:12 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L579 2024-02-16 18:00:18 Do I put that in /etc/mkinitfs/mkinitfs.conf ? 2024-02-16 18:01:20 no 2024-02-16 18:01:26 your kernel cmdline 2024-02-16 18:02:36 Ah. thanks! 2024-02-16 18:05:54 works great :) 2024-02-16 18:07:25 How do you verify? 2024-02-16 18:07:40 fstrim -v / 2024-02-16 18:07:44 ah ok 2024-02-16 18:07:51 without it I get a not supported error 2024-02-16 18:08:21 also "lsblk -D" 2024-02-16 18:08:21 I was surprised how easy it was to setup alpine on luks+lvm 2024-02-16 18:08:55 now, I have to manually put in the passphrase on server boot via remote console though. But at least it should be "more secure" :o 2024-02-16 18:08:56 if the DISc-GRAN column is zero then discard not supported 2024-02-16 18:10:38 Forza: I did submit a MR to add SSH-based remote unlock for LUKS some time ago but it was never merged 2024-02-16 18:11:32 oh. sounds cool. why didn't it? 2024-02-16 18:11:46 a dropbear thing? 2024-02-16 18:14:13 yeah it ran dropbear in the initramfs 2024-02-16 18:14:39 at the time there was no testing framework for mkinitfs and so it wasn't accepted due to "risk" of any major changes breaking mkinitfs 2024-02-16 18:15:06 I haven't (yet) gotten around to updating it and resubmitting it since 2024-02-16 18:16:22 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/86 for referenfce 2024-02-16 18:17:04 there are other ways to unlock LUKS - Grub 2.12 I believe now supports automatic unlocking via TPM (need to verify that), dracut as alternative initramfs supports SSH-based unlock AFAIK, booster as alternative initramfs supports clevis-based unlocking 2024-02-16 18:46:38 Just how do you trust that on a remote cloud machine 2024-02-16 18:47:26 And 2 how do you move such an image and access it on another machine for rescue or various reasons 2024-02-16 18:49:45 Forza: "move such an image"? you're referring to a disk image of a Alpine install? 2024-02-16 18:50:22 also "trust that" - what do you mean by "that" specifically? 2024-02-16 19:12:33 minimal: yes 2024-02-16 19:14:27 Forza: well you trust an Alpine disk image on a cloud as much as you trust the Cloud Provider you're using I guess ;-) 2024-02-16 19:15:11 on some Clouds disk images can be encrypted and the keys kept in a Provider's "secure" keystore 2024-02-16 19:17:30 This is why I use luks with password on boot 2024-02-16 19:17:47 I guess the cloud provider can still do stuff if they want 2024-02-16 19:18:17 yes, your Vm is running on the Cloud Provider's hypervisor so then have, in theory, complete access to it ;-) 2024-02-16 19:18:44 so they could log you typing your LUKS passphrase 2024-02-16 19:19:20 Yup 2024-02-16 19:19:27 or just pull the encryption key from RAM 2024-02-16 19:19:47 Maybe encrypted ram 2024-02-16 19:19:57 some providers do offer VMs using the recent AMD and Intel "VM memory encryption etc" functionality 2024-02-16 19:20:22 but again if there are issues/security holes in that... 2024-02-16 20:37:59 how problematic would it be to have, bugs : sysret_ss_attrs null_seg spectre_v1 spectre_v2 spec_store_bypass retbleed , in /cat/cpuinfo ? 2024-02-16 20:45:28 add "smt_rsb srso div0" to it 2024-02-16 20:45:54 now showing in alpine-v3.18.6 2024-02-16 20:47:12 this is "AMD Ryzen 5 3500U with Radeon Vega" , i am fear shaking with bit now ;) 2024-02-16 21:02:48 is there a chart somewhere that says cpu and related bugs whether fixed or not 2024-02-16 23:38:47 vkrishn: /proc/cpuinfo bugs just lists what has been detected based on hardware, *not* whether they have been mitigated or not 2024-02-16 23:39:53 you can grep through dmesg (and other places) to see if they've been mitigated (source https://unix.stackexchange.com/questions/456425/what-does-the-bugs-section-of-proc-cpuinfo-actually-show and local testing) 2024-02-16 23:55:03 fluix: thanks, i have add the process to a local docs, i think someone pointed the same a month or so back, but i forgot. here what it says in vuln - https://tpaste.us/z5VJ 2024-02-16 23:55:44 i think retbleed is still an issue, SMT vulnerable 2024-02-16 23:57:08 would try to make a table as i go along 2024-02-16 23:57:51 a table of CPU and vulnerabilities? that depends on if you have mitigations enabled, which ones you have enabled, microcode updates, sysctl options, and probably more 2024-02-16 23:58:39 but that also need a doc page or atleast knowing ? 2024-02-16 23:59:58 would be good to have somewhere public, yeah, just keep in mind this isn't very specific to Alpine 2024-02-17 00:00:08 so if i say booting al-v3.18.6 iso, then a chart to that extent can be made ? 2024-02-17 00:00:30 i am not trying to do on other distro for now 2024-02-17 00:01:11 I don't think I understand what you're trying to do 2024-02-17 00:02:05 ok, when you say 1. mitigations enabled ? it means enabled during boot or in kernel or modules ? 2024-02-17 00:03:43 kernel, but there may be mitigations enabled elsewhere as you mention 2024-02-17 00:04:10 2. microcode updated - like in syslinux calling amd/intel microcodes (ucode) ? 2024-02-17 00:05:56 like in microcode. I don't know what syslinux is 2024-02-17 00:06:02 for alpine-standard-3.18.6-x86_64.iso it should be same for same hardware/user ? 2024-02-17 00:06:10 wdym? 2024-02-17 00:11:19 oh, i thought syslinux was famous - by syslinux i mean passing to cmdline variable 2024-02-17 00:12:46 you can load microcode in different ways 2024-02-17 00:13:06 ok 2024-02-17 00:15:42 so i shouldn't bother to put it somewhere ? 2024-02-17 00:50:14 I think having information on what CPU bugs exist and what mitigations there are is useful, I just was (and still somewhat am) confused about exactly how you want to publish this. you mention a "local page" and I think you are focusing to specifically on Alpine (this information would be useful for users of most distros, and solutions are usually not Alpine-specific). of course if you want to specficially point out Alpine-specific things that's fine 2024-02-17 00:56:32 i assume y'all have looked at /sys/devices/system/cpu/vulnerabilities 2024-02-17 01:02:34 invoked: https://tpaste.us/z5VJ 2024-02-17 01:04:18 fluix: i don't know how to proceed with it, but willing to put info of hardwares' what i can, suggestions welcomed 2024-02-17 01:07:58 at quick glance you mainly have SMT to think about, which you can probably disable in the bios 2024-02-17 01:08:15 if it's not fixed with amd's microcode by now it probably will not be, so you have to disable smt 2024-02-17 01:15:18 invoked: thanks, will it affect anything normal in OS functioning ? 2024-02-17 01:15:39 loss of performance, depending. 2024-02-17 01:16:51 will try to read about them more 2024-02-17 01:47:18 btw, i am still unable to get sound via alsa (on normal setup-desktop xfce), anyone any hints, would be appreciated 2024-02-17 03:38:00 got alsa working :) 2024-02-17 03:41:09 this helped, https://forum.artixlinux.org/index.php/topic,3837.0.html 2024-02-17 10:31:22 Hi. I am using syslog-ng and am getting this error: Error opening plugin module; module='cloud_auth', error='Error relocating /usr/lib/syslog-ng/libcloud_auth.so: _ZTVN10__cxxabiv117__class_type_infoE: symbol not [ ok ] 2024-02-17 10:33:39 https://gist.tnonline.net/LM looks like something is missing when I do ldd lbcloud_auth.so 2024-02-17 10:36:47 maybe this is relevant to this issue? Maybe the build for exim missed some library or did linking wrong ? https://stackoverflow.com/questions/51689973/libcabi-undefined-symbol-ztvn10-cxxabiv120-si-class-type-infoe 2024-02-17 11:14:27 Forza, is this a syslog-ng error, or Exim? 2024-02-17 11:14:47 oh dear..... syslog-ng oups! 2024-02-17 11:16:13 I am too into my email setup at the moment :D 2024-02-17 11:16:20 https://gitlab.alpinelinux.org/alpine/aports/-/issues/15797 I made an issue 2024-02-17 15:38:30 Hi, does anyone know which package provide the `man` page for /etc/network/interfaces ? 2024-02-17 16:13:01 ifupdown-ng-doc, if any 2024-02-17 16:45:00 thanks ikke 2024-02-18 00:05:50 p 2024-02-18 00:15:28 I working on the finishing touches on my raspberry router. I really enjoyed using ACF to do most of the configuring. I installed iptables acf-iptables and copied my default rules-save, and it's being overwritten. # Generated by iptables-save v1.8.10 (nf_tables) on Sat Feb 17 17:35:41 2024 2024-02-18 00:31:49 hello, what's the best way to handle multiple patch on APKBUILD ? I want the user to choose their prefered patch and the apply it. 2024-02-18 00:32:15 I make this (https://gist.githubusercontent.com/ahmadraniri/6345f0b9013031d5f1a603f96fbd2a26/raw/7057147d62be099986c82badbc64926413bf302e/APKBUILD) but somehow abuild apply the patch before user choose. 2024-02-18 00:32:42 *and then. 2024-02-18 01:04:21 bye, thanks. 2024-02-18 09:59:16 how does one get asan on alpine 2024-02-18 10:00:35 compiler-rt ok got it 2024-02-18 14:03:33 so far, tested wifis ath9k(atheros), broadcom(brcmsmac) , they install/work fine in v3.18.6 2024-02-18 14:05:12 i am not going for networkmanager for now, but just for knowing, for those wanting it, try not activating it during desktop startup 2024-02-18 14:06:07 thanks for devs making it happen, i am trying to setup one for regular use 2024-02-18 15:13:05 add realtek-8822CE to working wifis, will try to add to a webpage, thinking on it 2024-02-19 16:26:28 what other dep is needed to install and run android-tools ? 2024-02-19 16:26:55 adb devices - exits with error 2024-02-19 16:27:53 i wonder what the error could be… 2024-02-19 16:29:41 https://tpaste.us/d1vx 2024-02-19 17:08:50 something already listening at that address? 2024-02-19 17:09:00 permissions? 2024-02-19 17:10:02 strace? 2024-02-19 17:10:46 Other distros have a package with a set of udev rules for android 2024-02-19 17:14:20 lopid: fresh install-to-ram, should not be 2024-02-19 17:14:48 Ermine: yes, its in v3.19(community) 2024-02-19 17:15:14 my install is 3.18.6 2024-02-19 17:18:31 if that is needed pls backport it, if possible 2024-02-19 17:25:23 Ermine: obviously that would only apply if (e)udev is being used, rather than mdev 2024-02-19 17:26:02 the android-tools package doesn't declare any dep on eudev 2024-02-19 17:30:08 strace adb devices > dump.txt 2>&1, should be useful ? 2024-02-19 17:33:06 Hello ! I just recently downloaded Alpine on an old computer to use it as a "home server" (in a way), so everyone on my network can access a ssd drive. This part has been going on great, and works. But, I then wanted my screen to be shut down when the lid is down (so it doesn't consume electricity nor burns or anything), but I've been stumbling on that... I've tried installing xorg (xorg-setup-base) and use xset with a 2024-02-19 17:33:06 script but... it doesn't detect my display... Do i need to install xfce for this to work ? 2024-02-19 17:33:31 Ermine: ^ 2024-02-19 17:34:53 did you install a gui just for this? does the screen not turn off at a virtual console? 2024-02-19 17:38:25 I didn't install anything really, i'm on the basic terminal... couldn't get neither openbox or fluxbox to turn on... 2024-02-19 17:39:34 vkrishn: yes 2024-02-19 17:39:52 https://d.insteps.net/adb.strace.txt 2024-02-19 17:40:34 minimal: it would be nice to have set of mdev(d) rules for android 2024-02-19 17:41:35 lopid: of course, when I finally reach for help, it works... I managed to get setterm to work, after it just didn't work for the whole day... sorry for the bothering :/ 2024-02-19 17:42:07 Ermine: the general problem is mdev doesn't support "drop in" conf files to add to its configuration unlike udev, so mdev.conf would need to be modified for any/every rule required by every package 2024-02-19 17:45:34 minimal: didn't skarnet rant about this is -devel ? 2024-02-19 17:47:34 vkrishn: there's also log at /tmp/adb.0.log, can you send it as well? 2024-02-19 17:48:52 both myself and skarnet talked about this in the past, of trying to kludge in some mechanism to "merge" files 2024-02-19 17:53:29 ... to no avail 2024-02-19 17:55:09 https://tpaste.us/MR77 2024-02-19 17:58:54 currently there's approx 20 various udev rule packages, it would make sense if it was possible to also have corresponding mdev rule packages for those also 2024-02-19 18:03:09 hey sorry for the stupid question but how would i disable waiting for network during boot? 2024-02-19 18:10:53 vkrishn: can you do the strace again, but with -f flag? strace -f -o strace.log adb devices 2024-02-19 18:11:05 I'm always forgetting -f 2024-02-19 18:12:40 Juest: waiting for network? or waiting for DHCP? 2024-02-19 18:13:16 minimal: waiting for network to be up, dhcp takes a bit to gather but it should be quick 2024-02-19 18:13:35 during boot 2024-02-19 18:19:14 https://d.insteps.net/adb.strace2.txt 2024-02-19 18:19:20 Juest~: dhcp takes as long as it takes... 2024-02-19 18:19:49 minimal: usually dhcp is quick to get an ip address.... 2024-02-19 18:19:51 should 2024-02-19 18:20:00 not sure why it takes longer 2024-02-19 18:20:08 or i guess i dont perceive it taking longer for some reason 2024-02-19 18:21:04 you've just made conflicting statements: "it takes longer" and "i dont perceive it taking longer" 2024-02-19 18:22:18 still unclear exactly what you're asking - is it that you don't want the boot to wait until networking is up? If so then what should happen to any services that may depend on networking? 2024-02-19 18:23:05 s/networking is up/ethernet interface has an address/ 2024-02-19 18:28:32 vkrishn: take a look at output of ip addr. Probably there's no interface with 127.0.0.1 address. 2024-02-19 18:29:25 lo should have this address 2024-02-19 18:33:15 minimal: correct, services should still start fine after networking comes up, i just dont want boot to be held off by waiting for network, in artix it boots fast and got network just fine, those services can wait until network is up, boot doesnt require network services 2024-02-19 18:39:47 maybe its because i need to enable parallel processing? 2024-02-19 18:46:52 Ermine: thanks, rc-service loopback restart; rc-update add loopback 2024-02-19 18:47:02 works now :) 2024-02-19 18:50:47 don't know why that stopped, or never started 2024-02-19 18:55:00 ok, figured it out, i did manual interface install (wpa ... etc) and bypassed setup-interfaces 2024-02-19 18:59:54 setup-interfaces is about networking service afaik 2024-02-19 19:00:20 and /etc/network/interfaces has config for lo 2024-02-19 19:00:45 Juest: do you mean artix with openrc? 2024-02-19 19:00:50 yes Ermine 2024-02-19 19:01:33 probably should be asking this in the openrc channel but its a ghost town 2024-02-19 19:10:31 and "networking service" was not been added in boot or default yet, ahh! 2024-02-19 19:28:05 hardware fun fact, did you know that usb 3.0+ female port can be connected via a standard mobile micro-sd cable using one of its larger port 2024-02-19 19:28:21 though the speed would be of usb-2.0x 2024-02-19 19:31:36 micro-usb cable ^ 2024-02-19 21:06:25 so... 2024-02-19 22:05:22 -p 2024-02-20 17:13:40 mksquash (v3.18.6) gives lots of xattr messages while compressing a ntfs3 partition, but was ok while using knoppix(debian) 2024-02-20 17:22:25 I'm trying to install Steam using the wiki guide, but getting PolicyKit1 errors: https://wiki.alpinelinux.org/wiki/Steam 2024-02-20 17:22:38 `error: Failed to install org.freedesktop.Platform: Authorization error: The name org.freedesktop.PolicyKit1 was not provided by any .service files` 2024-02-20 17:22:45 am I missing an obvious package? 2024-02-20 17:23:11 Which arch and version of alpine? and which step on the wiki guide are you running into the issue? 2024-02-20 17:23:31 3.19.1, x86_64 2024-02-20 17:23:46 The error occurs on `flatpak install com.valvesoftware.Steam` 2024-02-20 17:25:18 not sure offhand, what happens if you change the command to `flatpak --user install com.valvesoftware.Steam' ? 2024-02-20 17:26:03 also the wiki article could use more clarity, you are running the install command as a regular user and not root? 2024-02-20 17:26:18 interesting, yes that seems to fix it 2024-02-20 17:26:31 I reran `flatpak --user install com.valvesoftware.Steam` as non-root 2024-02-20 17:26:44 preceded by `flatpak --user remote-add --if-not-exists flathub https://flathub.org/repo/flathub.flatpakrepo` as non-root 2024-02-20 17:26:56 cool 2024-02-20 17:27:04 I'll update the wiki, you're not the first to run into that \o/ 2024-02-20 17:27:24 thanks! I'll report back in a few mins when the install finishes 2024-02-20 17:34:55 seems like the flatpak installed correctly, but it fails to create the initial window 2024-02-20 17:35:04 `src/steamexe/updateui_xwin.cpp (339) : Could not open connection to X` 2024-02-20 17:35:16 I'm running sway, xwayland is installed 2024-02-20 17:36:47 ideas appreciated :) 2024-02-20 17:37:51 huh, so running it without the --user flag seems to work 2024-02-20 17:38:11 `flatpak run com.valvesoftware.Steam` not `flatpak --user run com.valvesoftware.Steam` 2024-02-21 00:10:26 In the spirit of Gentoo's user patches, how should I proceed if I wanted to compile an Alpine package with different options? I can totally get the APKBUILD, modify it and rebuild the package myself. But then what happens when the package is modified upstream? Do you have a strategy for that? 2024-02-21 00:12:33 I guess I would like for abuild to tell me that it will override my user-package, such that I know that I may have to rebuild/reinstall it myself... 2024-02-21 00:15:31 so the equiv of knowing "i used a different repo (local setup to get this" to block version bumps from removing your custom copy? 2024-02-21 00:27:47 tdr: that, or just letting me know that it is removing my custom copy I guess. If I overwrite the system package with my own, keeping the same pkgrel, then the next update will erase my custom package, right? I guess that's fine, but it means that I may need to take action, so it would be cool if abuild told me "I will erase your custom package" 2024-02-21 00:29:13 Or maybe I'm overthinking it and I can just remember which packages I customized (right now I just want to rebuild gstreamer to have intel media SDK support such that it can actually encode/decode h264/h265 with my iGPU) 2024-02-21 00:30:14 And that was my "ooooh, that's the point of the Gentoo user patches" revelation, I guess :D. Before that I always thought that Gentoo was about "compiling with the optimizations you want", but actually sometimes you really need to compile with a specific option. 2024-02-21 00:31:15 Or, actually, the solution is maybe to make my own `gstreamer-plugin-msdk` that just installs this one plugin next to the system gstreamer! 2024-02-21 00:47:59 creating your own local repo would be ideal but it's not trivial (but not hard either) 2024-02-21 00:49:16 but that's only worth it if you have several ormore packages you want to build & maintain special 2024-02-21 02:34:39 it there any tool could tell me files in an xxx.apk 2024-02-21 02:35:31 install it and apk info -L is a little complex 2024-02-21 03:23:17 qaqland: pkgs.alpinelinux.org ? 2024-02-21 03:52:02 The apk files are actually gzip-compressed tar files, and gzip can be streaming decompressed. You can download the header part of the apk file to get the file list, I guess. 2024-02-21 04:00:37 lindsay: pkgs.alpinelinux.org is easier if it's an existing APK file though yes "zcat file.apk | tar vtf -" does work 2024-02-21 08:34:15 minimal: this command really works! thankyou 2024-02-21 10:27:45 Hi, I've installed the `emacs` and `emacs-gtk3-nativecomp` packages, and everything's working so far nicely. Is there any other package I need to install/something I need to configure to get emacs' daemon functionality? OpenRC complains there that it can't find the `emacs` service 2024-02-21 14:38:38 invoked: how does it work if I have my own repo? Can my repo have only a few packages, and be search before the upstream repos? And then does that mean that my custom packages won't be updated by upstream (and therefore I need to check myself if they have updates)? 2024-02-21 15:05:58 jonesv: no, it won't work like that. apk iirc will always pull the highest version number 2024-02-21 15:42:45 invoked: but then it won't tell me that it "switched back" to the upstream package (because it has a highest pkgrel), right? 2024-02-21 15:44:45 Id name the package differently 2024-02-21 15:54:39 ikke: hmm but what if another package depends on it? I will end up with two conflicting packages, right? 2024-02-21 16:00:03 jonesv: you can make it provide the other package 2024-02-21 16:00:15 (make sure you provide a version as well) 2024-02-21 16:05:46 provides="otherpackage=$pkgver-r$pkgrel" 2024-02-21 16:21:58 remind me, what's the name of the package which has the ash shell manpage in it? 2024-02-21 16:22:13 Is there an ash shell man page? 2024-02-21 16:23:41 so i looked for the ash home page, and it pointed me to the dash git server 2024-02-21 16:23:57 looks like there are manpages in the source for dash but probably not what alpine is using 2024-02-21 16:24:19 I'll probably just yield and figure it out by reading them from the repo with nroff 2024-02-21 16:24:39 alpine is using busybox ash 2024-02-21 16:25:22 It does have some bashisms enabled though, so reading a manpage for dash would not give a complete picture 2024-02-21 16:40:07 good to know, just wanted to RTFM and *sigh* requires USL (use-the-source-luke) :-) 2024-02-21 16:57:41 ikke: oh right, I see! Thanks! 2024-02-21 22:14:57 All, I am trying to put the finishing touches on my raspberry pi router. I'm using ACF to create content in iptable, dhcp, cron. I have installed iptable-legacy, but iptables-save keeps over writing rules-save. 2024-02-21 22:46:36 hm, I'm wanting to copy my edited files and custom-added (e.g. compiled from source) files to a secondary disk, reinstall alpine, and restore the files, does apk audit catch all such files? I think it does have a few false positives, at least, such as all the busybox-extras symlinks (I guess the symlinks aren't part of package, but setup programmatically?), wondering if it also 2024-02-21 22:46:37 misses anything (other than /home and /tmp of course) 2024-02-21 22:47:09 (using apk audit --full -r --ignore-busybox-symlinks) 2024-02-21 23:10:31 txnintn: if I recall correctly, the files mention how to save the configuration 2024-02-21 23:11:19 You mean the iptables.conf file? 2024-02-21 23:17:21 I was just running a command line test on this issue, if I edit rules-save and start and restart, the file is overwritten. 2024-02-21 23:19:21 I believe there's a command to save it 2024-02-21 23:19:55 looks like you need to run iptables-save 2024-02-21 23:20:01 err, rules-save 2024-02-21 23:20:21 because the service runs rules-restore: https://git.alpinelinux.org/aports/tree/main/iptables/iptables.initd#n84 2024-02-21 23:22:23 So the app is inserting a backup over my hand edited update to rules-save 2024-02-21 23:27:56 In the acf on the expert page if I hit the save button below the rules-save file, it saves it but again when I hit restart goes back to default file. 2024-02-21 23:30:18 sorry I know nothing about ACF or what it's doing. good luck debugging 2024-02-21 23:32:55 I'm building my own house and I like to get the home automation and the reset of the network done and tested. Thanks for your input. 2024-02-22 09:29:05 tips on installing Alpine on RPi5+NVMe ? 2024-02-22 11:48:35 requesting opinion/experience: is it worth to run x86 VMs on x86_64 machine? pros/cons? 2024-02-22 11:50:55 (host and guests are all alpine, if it matters) 2024-02-22 15:31:55 aron: i'd say it depends on the applications you intend to run in the VMs 2024-02-22 15:32:31 ncopa: thanks for your reply 2024-02-22 15:32:51 for instance A, only light-weight application 2024-02-22 15:33:20 i.e. a commit hook would send a message to it 2024-02-22 15:33:21 the pro is that it uses slightly less memory. I'd only do it if the VM gets 2G ram or less 2024-02-22 15:33:35 instance B, a small CGI with lighttpd 2024-02-22 15:34:09 way less than 2G, I thought about 100-200M 2024-02-22 15:34:13 the con is that it as only 4GB addressable memory 2024-02-22 15:34:42 I'm a bit worried about the reduced instruction set, less registers, etc 2024-02-22 15:35:05 yeah, might run slightly slower 2024-02-22 15:35:20 but the less memory sounds good 2024-02-22 15:36:21 I'll try it how it goes 2024-02-22 15:37:36 👍 2024-02-22 15:39:29 ncopa: while we are on it, everything is -Os correct? 2024-02-22 15:39:50 Well, not _everything_ 2024-02-22 15:40:01 well, most things 2024-02-22 15:40:02 some packages override it to -O2 or -O3 even 2024-02-22 15:40:06 But it's the default 2024-02-22 15:40:14 yeah. that's what i thought. thx 2024-02-22 15:41:04 invoked: https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/default.conf#L1 2024-02-22 15:42:33 thx. i should have looked there myself. 2024-02-22 15:43:19 You'd first have to know where to look 2024-02-22 15:45:03 Do you have specific instructions how to configure iptables-legacy? 2024-02-22 15:46:03 Is this related to your earlier acf question? 2024-02-22 15:46:15 ikke: well, i was adding context around aron's question, otherwise i would have grepped the tree for confs :) 2024-02-22 15:47:52 I was hoping to use acf as my primary and ssh as secondary, as the router will be in a network cabinet. 2024-02-22 15:50:12 txnintn: I think you should just disable the iptables service and let acf do the saving 2024-02-22 15:50:38 there is not a lot to configure 2024-02-22 15:50:54 the only thing those services to is make sure the config is saved on shutdown and restored on boot 2024-02-22 15:54:47 That the issue, it's not saving in the cli or in acf. I did notice try to figure this out, libipq is not loading when adding iptables-legacy. So don't load iptables, ie rc-update iptables boot and start it with acf? 2024-02-22 15:56:55 fyi the cabinet is 12 feet above the floor. 2024-02-22 15:57:50 I'm not too familiar with how acf is supposed to work 2024-02-22 16:01:31 I thought it was like webmin application, I did configure dhcpd with it and a few other things. 2024-02-22 16:02:51 Yes, it is 2024-02-22 16:03:17 But not sure if the iptables module uses those services or supersedes them 2024-02-22 16:08:56 yeah, could be a 3 ring circus between acf, iptables and iptable-legacy. 2024-02-22 16:10:31 there is no distinction between iptables and iptables-legacy 2024-02-22 16:10:56 iptables-legacy provides the iptables interface for nftables 2024-02-22 16:13:31 and it appears to be mostly applies to save and restore function. 2024-02-22 16:24:48 From what I have found iptables-legacy-save and restore is to be used in lew of the iptables-save and restore. 2024-02-22 21:50:51 did xrdb disappear? 2024-02-22 21:51:00 I swear I had an xrdb binary a little while ago 2024-02-22 21:51:45 elly: someone removed startx as a dependency, so it probably got uninstalled along with some other things 2024-02-22 21:51:55 The package is still there 2024-02-22 21:52:11 oh, I see, yep 2024-02-22 21:52:19 it used to be pulled in implicitly for me 2024-02-22 21:52:21 thanks! 2024-02-22 22:39:13 My suggest for acf-iptables, maybe add a restore button for rules-save, it may make it a little more user friendly. 2024-02-22 22:41:29 txnintn: You'd have to open an issue on the relevant repository 2024-02-22 22:42:05 https://gitlab.alpinelinux.org/acf/acf-iptables 2024-02-22 22:51:54 Working on thanks for the link 2024-02-23 05:42:38 ahh sorry my cat disconnected my laptop (I was using a very old MacBook Air with a battery that won't charge and when he jumped on my lap he hit the MagSafe connector) -- I was spicy_carlton if someone responds 2024-02-23 05:44:18 l0rd_hex: I don't see any messages from spicy_carlton 2024-02-23 05:44:44 ikke: I just realized it didn't post from the webclient, thanks 2024-02-23 05:47:59 so my question is, I've been booting the alpine extended iso on the odd machine, creating a custom parition layout, mkfs.ext(2|4)ing them and mounting to /mnt, after which I run setup-disk /mnt -- I noticed that when I ran setup-alpine after the settings didn't stick after reboot but I was able to re-run setup-alpine and it works just fine. I was wondering if it's possible to use setup-disk with a 2024-02-23 05:48:05 custom partition table and run setup-alpine in one go 2024-02-23 06:50:09 Ohh, I think I figured it out, I run setup-alpine _first_ and then setup-disk /mnt 2024-02-23 08:49:07 Hello... Is it linux-edge not recommended for running waydroid ? Linux-lts works fine for waydroid. 2024-02-23 09:04:17 bye, thanks. 2024-02-23 15:44:32 any suggestion to improve on how to check if wifi is connected ? https://tpaste.us/QRgO 2024-02-23 16:03:32 iw dev wlan0 link 2024-02-23 16:08:27 i usually do that from cmdline, was wanting to script in cron 2024-02-23 16:09:29 pretty much anything you do at the command line you can put in cron. 2024-02-23 19:02:30 How do I remove a package form APKINDEX.tar.gz ? 2024-02-23 19:05:43 If i remberright extract apkindex to a folder, edit the file, delete the file from the index and recompress it. 2024-02-23 19:06:37 I was adding a file to apkindex. 2024-02-23 19:07:25 the first file is the list of files... 2024-02-23 19:10:28 So, I had to apk add linux-firmware-rtlwifi-20231111-r1.apk manually before I could install anything else because I didn't have wifi 2024-02-23 19:10:59 Now, though, I can apk add it from the repo 2024-02-23 19:11:16 But when I do, I get "1 error; 1448 MiB in 230 packages" 2024-02-23 19:11:17 that's it 2024-02-23 19:11:33 apk tells me there is an error, but doesn't say what the error was. 2024-02-23 19:12:41 So you manually added a driver to the apkcache? 2024-02-23 19:14:02 I don't know. I did apk add /path/to/linux-firmware-rtlwifi-20231111-r1.apk 2024-02-23 19:14:06 Is that what that does? 2024-02-23 19:17:43 well, I fixed that problem. All I had to do to switch to the repo version was apk del it first. 2024-02-23 19:17:58 but I still have that "1 error" 2024-02-23 19:18:01 If you manually add files to the apkcache, the APKINDEX.tar.gs (list) has to be updated 2024-02-23 19:18:28 I didn't do that. I told apk to install the file directly. I didn't touch the cache 2024-02-23 19:18:56 so then number of files in the cache and the list match 2024-02-23 19:19:39 I just figured out that I get that "1 error" no matter what apk command I do. 2024-02-23 19:21:45 Did you remove the # from the local repostitory while adding the file? 2024-02-23 19:23:11 I have no idea what you're talking about. All I did was download the .apk from a package mirror, save it to an SD card, plug the SD card into the computer with a fresh install of alpine linux but no internet, and run apk add /mnt/sdcard/linux-firmware-rtlwifi-20231111-r1.apk . 2024-02-23 19:25:09 I don't even know whether this "1 error" I keep seeing has anything to do with the rtlwifi package. 2024-02-23 19:25:14 What does the /etc/lbu/repositories list without a #? 2024-02-23 19:25:17 When I install any package, I still see it 2024-02-23 19:25:38 no such file 2024-02-23 19:26:31 /etc/apk/repositories lists http://mirror.csclub.uwaterloo.ca/alpine/v3.19/main and http://mirror.csclub.uwaterloo.ca/alpine/v3.19/community 2024-02-23 19:29:34 Ah, found an answer: https://dev.alpinelinux.org/~clandmeter/other/forum.alpinelinux.org/forum/general-discussion/apk-always-has-1-error-solved.html 2024-02-23 19:29:47 my bad... did you setup an apkcache local location (setup-apkcache) 2024-02-23 19:30:06 no, I *only* did what I said I did 2024-02-23 19:30:29 The error is gone now, though. I was from a totally different package 2024-02-23 19:30:39 apk fix fixed it 2024-02-23 19:30:46 I'm running serveral diskless setups 2024-02-23 19:31:09 Yeah I can't do diskless. I only have a gig of ram 2024-02-23 19:31:25 sounds like you got 2024-02-23 19:31:25 isn't that plenty? 2024-02-23 19:31:56 diskless loads the packages to ram. There's more than a gig of packages on this thing 2024-02-23 19:32:04 ok 2024-02-23 19:32:32 retroarch pulled in like 150 dependencies lol 2024-02-23 19:32:42 I guess it depends on what you want to do. 2024-02-23 19:32:54 retroarch and only retroarch 2024-02-23 19:33:17 This thing can't run a web browser with only a gig of ram 2024-02-23 19:33:30 Stupid, and it didn't always used to be like this, but true. 2024-02-23 19:33:43 My router is at 349mb in 137 packages... 2024-02-23 19:33:45 The web has grown out of control 2024-02-23 19:34:37 I used to run OS/2 it was written in machine code, 2024-02-23 19:36:55 How do I pacman -U --asdeps a package? 2024-02-23 19:37:06 *pacman -D --asdeps 2024-02-23 19:37:06 first 32-bit os multi threaded in 1992 with gui on a 230 megabyte hardrive. 2024-02-23 19:40:04 < Don't know pacman, played with arch a weebit about a year ago, ended up using alpine. 2024-02-23 19:41:19 I did apk add mesa without realising it was already installed as a dependenct. 2024-02-23 19:41:36 Now, it's explicitly installed and it has an entry in /etc/apk/world. 2024-02-23 19:42:08 I'd like to remove the entry. Is that safe to do with just vi, or is there come apk incantation I need to do instead? 2024-02-23 19:43:08 just do apk del; which removes it from world 2024-02-23 19:43:24 if it's still a dependency of something it will stay installed though 2024-02-23 19:45:37 oh, thanks, zcrayfish 2024-02-23 19:56:30 \o/ 2024-02-23 20:44:45 well, I've got retroarch working 2024-02-23 20:44:54 I'm playing donkey kong country right now lol 2024-02-23 20:45:31 now, I've gotta set up automatic login 2024-02-23 20:46:19 how many games for it? 2024-02-23 20:46:27 way too many games 2024-02-23 20:46:31 literally all the games 2024-02-23 20:46:39 every snes game ever made 2024-02-23 20:46:40 lol 2024-02-23 20:48:01 <- building my own house, if I ever get done, I'll look into it. 2024-02-24 01:41:34 Hi all, is there a .iso that I can download that has both non free firmware and wireless so I can install via the network? 2024-02-24 01:46:37 d3fragg3d: did you try the Extended ISO? 2024-02-24 01:47:32 oh I didnt. cool 2024-02-24 03:59:16 Edge has a conflict on libcmark between gnome-builder and gnome evolution right now, with builder being built against the older version 2024-02-24 03:59:40 What is the proper way to address this if builder doesn't have a new release, just submit an MR to bump the rel so it gets rebuilt against the newer libcmark? 2024-02-24 06:50:15 Saijin_Naib: yes 2024-02-24 06:51:43 Contrib.md covers the needed MR note style? 2024-02-24 06:54:46 : rebuild against 2024-02-24 07:23:50 Thanks as ever, ikke 2024-02-24 12:28:21 Setting up logbook was trivial. I left some notes for anyone curious: https://wiki.alpinelinux.org/wiki/Logbookd 2024-02-24 18:34:54 WhyNotHugo: what drove you to be interested in logbooks vs. conventional syslog daemons? 2024-02-24 22:24:16 xgpt: it stores logs from all applications into a single DB. So I can view logs of all services intermixed, or view just one servie, depending on my needs. 2024-02-24 22:26:41 It also does very sporadic writes to disk, so I'm going to use it on rpi (with an sd as storage). 2024-02-24 23:42:43 gotcha 2024-02-24 23:50:52 hello, I'm on alpine sway and have a question about flatpaks. Sometimes (like when I try and run Discord or play certain games on Steam) flatpak fails and my terminal says "error: while opening repository /var/lib/flatpak/repo: opening repo repo: opendir(/var/lib/flatpak/repo): No such file or directory". I'm having trouble finding what I'm doing wrong ,any advice? 2024-02-24 23:52:32 I pretty recently reflashed my computer so I haven't messed with it too much, I'm not sure if this is a permissions issue or what's going on. Most flatpaks are fine, but like I said, "flatpak --user run com.discordapp.Discord" always gets me that error 2024-02-25 00:03:42 Does it make any difference if you drop the --user flag? 2024-02-25 00:05:04 Trying to run any flatpak without the --user flag gives me that error unfortunately 2024-02-25 00:06:09 When I run discord with that flag it gives me that error too, but there's at least a few other lines (like Disabling Updates Already Done and 'failed to let stub process know of exit: bad file descriptor') 2024-02-25 00:06:53 honestly I was fine when it was just discord, since I can pull that up in a browser, but now a game in steam won't launch and terminal gives me that same error, which is the more annoying bit 2024-02-25 00:07:45 well at least we've ruled out an accidental system install. =) 2024-02-25 00:11:00 true! The game that won't launch runs through proton, while the one that does launch has a native steam release, so I'm downloading another windows-only steam game to try and narrow things down a bit 2024-02-25 00:15:29 that game failed too, so when Steam tries to run a game through proton it fails with that error, and when I try and run discord it fails with that error too 2024-02-25 00:26:34 I feel like this started when proton-ge was archived and became unavailable as a flatpak. Before that I'd always used the flatpak proton-ge to run proton games through steam 2024-02-25 22:21:03 Bonsoir 2024-02-25 22:21:12 I'm trying to use dhcpcd with bonding 2024-02-25 22:21:37 I've been using a bond interface for quite some time, with udhcpc 2024-02-25 22:22:14 According to the wiki, if I understand correctly, installing dhcpcd would be sufficient, but I get a “timeout” for some unknown reason 2024-02-25 22:22:21 (not sure what's actually timeouting too) 2024-02-25 22:23:46 Do you know if using dhcpcd actually requires some special configuration with ifupdown? 2024-02-25 22:27:10 quinq: what's your /etc/dhcpcd.conf and /etc/network/interfaces contents? do the logs in /var/logs/ give any info? 2024-02-25 22:28:44 dhcpcd conf is default, with added "nohook resolv.conf" and "noipv4ll" 2024-02-25 22:29:29 interfaces is loopback and "auto bond0\niface bond0 inet dhcp\nbunch of bond-* options" 2024-02-25 22:30:19 grep -F dhcp /var/log/dhcp returns nothing 2024-02-25 22:30:29 sorry, /var/log/messages 2024-02-25 22:32:28 strange, I'd expect to see dhcpcd log to /var/log/messages 2024-02-25 22:33:48 facepalm, I made a typo in the grep command 2024-02-25 22:33:51 do you have ifupdown-ng installed? 2024-02-25 22:34:20 That comes with base system, doesn't it? 2024-02-25 22:34:24 ifupdown-ng should use dhcpcd as the DHCP client in preference to dhclient/udhcpc 2024-02-25 22:34:44 it is the default, unless you install busybox-ifupdown 2024-02-25 22:34:54 Basically, waiting for carrier / carrier acquired / soliciting an IPv6 router / soliciting a DHCP lease / no IPv6 Routers available / timeout / dhcpcd exited 2024-02-25 22:35:01 (I'm assuming you're on Edge) 2024-02-25 22:35:20 Indeed, edge here, and dhcpcd is being used 2024-02-25 22:35:28 so you're using IPv6 only? 2024-02-25 22:35:33 no 2024-02-25 22:35:46 Though I have a working IPv6 network, in addition to the IPv4 one 2024-02-25 22:36:15 I can also see that the laptop is associated with the wireless AP during dhcpcd runtime 2024-02-25 22:36:53 the dhcpcd log entries mention bond0 as expected? 2024-02-25 22:36:59 and it works directly when I uninstall dhcpcd 2024-02-25 22:37:10 (using udhcpc) 2024-02-25 22:37:24 For testing, I do: doas rc-service networking restart 2024-02-25 22:37:36 Yes sorry, it's trying to up bond0 2024-02-25 22:39:13 I can see that udhcpc is still running, even after doas rc-service networking stop, which is strange 2024-02-25 22:39:27 Though I'm killing it before restarting with dhcpcd 2024-02-25 22:39:41 This looks like an openrc bug, doesn't it? 2024-02-25 22:40:55 /etc/init.networking runs ifup 2024-02-25 22:41:05 which for you is provided by ifupdown-ng package 2024-02-25 22:41:15 which will select dhcpcd 2024-02-25 22:41:23 so I don't see how udhcpc is being run 2024-02-25 22:41:44 It's being run from when I remode dhcpcd 2024-02-25 22:41:52 remode? 2024-02-25 22:41:54 And restart the network to check that it works without it 2024-02-25 22:42:01 s/remode/remove/ 2024-02-25 22:42:29 ok, that's because /etc/network/interfaces states DHCP and ifup (from ifupdown-ng) will look for a DHCP client 2024-02-25 22:42:42 as you uninstalled dhcpcd then it will use udhcpc instead 2024-02-25 22:42:53 Yes, I know that :) 2024-02-25 22:43:12 But I'm saying that it's strange it's still runnin when I'm stopping the networking service 2024-02-25 22:43:17 (or service group) 2024-02-25 22:43:44 Anyway, that was a side remark, not the first actual issue 2024-02-25 22:43:46 try running "ifdown --verbose bond0" and see what it reports 2024-02-25 22:43:53 trying 2024-02-25 22:45:09 interface is down, no mention of dhcp though 2024-02-25 22:45:17 oh 2024-02-25 22:45:23 udhcpc is gone though 2024-02-25 22:45:32 So the problem is that the rc/networking doesn't call ifdown? 2024-02-25 22:45:55 no, it does call ifdown, check the stop function in the script yourself 2024-02-25 22:47:05 unless ifquery isn't showing the interface 2024-02-25 22:49:59 I think I know what's the “bug” 2024-02-25 22:50:06 I need udhcpc for installing dhcpcd 2024-02-25 22:50:19 ? 2024-02-25 22:50:19 Then shutting down doesn't look at what's running, but what's installed 2024-02-25 22:50:27 And then tries to kill some dhcpcd instead of the running udhcpc 2024-02-25 22:53:46 well that's likely a issue due to missing pre-deinstall scripts (i.e. dhcpcd.pre-deinstall) to shutdown the client before the package is removed 2024-02-25 22:54:05 it's something in general I've been considering recently 2024-02-25 22:54:33 I'm not removing udhcpc, but installing dhcpcd (as it's supposed to take priority over udhcpc anyway) 2024-02-25 22:55:03 I was referring to when you earlier mentioned about removing dhcpcd 2024-02-25 22:55:47 That part causes no issue as far as I can see 2024-02-25 22:56:17 but yes there are some logic "tweaks" that should be made to the ifupdown-ng dhcp script 2024-02-25 22:57:22 "I'm not removing udhcpc" - well you can't as it is part of Busybox ;-) But "logically" you be able to - that's something that I've been working on recently 2024-02-25 22:57:35 s/you be/you should be/ 2024-02-25 23:00:03 Can I tell apk do install the packages I fetched locally? 2024-02-25 23:00:30 I'm trying to add dhcpcd while “offline” 2024-02-25 23:00:44 just "apk add " 2024-02-25 23:01:14 oh ^^ 2024-02-25 23:01:16 Thank you 2024-02-25 23:04:45 I get more output with the "debug" option in dhcpcd.conf 2024-02-25 23:04:58 Though I can just see its DISCOVER packets apparently timeout 2024-02-25 23:15:23 lel 2024-02-25 23:15:33 It's sending DHCP requests with the wrong MAC 2024-02-25 23:16:54 So most likely it's sending its requests on the wrong slave interface 2024-02-25 23:17:03 Not the actually selected one by the bond 2024-02-25 23:17:24 I guess nobody tested bonding with dhcpcd 2024-02-25 23:23:44 works now 2024-02-25 23:23:49 I had to change the bond parameters 2024-02-25 23:24:04 Maybe I'll spend the energy making a bug report to dhcpcd 2024-02-25 23:30:14 Thanks for the help, minimal :) 2024-02-26 00:11:38 quinq: which bond parameters did you need to change? 2024-02-26 00:17:18 minimal, main interface 2024-02-26 00:17:48 Was eth0, but wlan0 was the active one 2024-02-26 00:17:57 Changing it to wlan0 made it use wlan0 2024-02-26 00:19:25 So there's a bug in determining the currently active slave 2024-02-26 00:20:24 Maybe what happens is that when the bond gets up, it starts with default interface, and dhcpcd catches that at once, then the bond switches to the actually up slave, but dhcpcd doesn't follow 2024-02-26 00:21:03 Just an idea, I don't know how dhcpcd handles bonding internally 2024-02-26 00:21:06 Anyway, good night! 2024-02-26 15:18:32 f_: hi. also coming here, at your suggestion 2024-02-26 17:37:54 i've created an installer that makes it easier to install Alpine linux with ZFS and ZFSBootMenu, on any VPS provider: https://github.com/psy0rz/alpinebox 2024-02-26 17:38:10 is there a place i can add a reference? In the docs or wiki or something? 2024-02-26 17:46:34 I'm not sure we have any unofficial installer recommendation page 2024-02-26 21:54:23 Hi, when I try to run some, but not all, flatpaks I get "error: While opening repository /var/lib/flatpak/repo: opening repo: opendir(/var/lib/flatpak/repo): No such file or directory" and it fails to run. What does this mean? 2024-02-26 22:13:09 Seems to be a flatpak issue, found a workaround that works written (for alpine no less) on flatpak's github #5488. I'll add this to the wiki troubleshoot in case anyone else runs into this 2024-02-26 23:50:57 hey, did alpine remove drivers recently? I can't launch steam games that use proton anymore, and when I check the log it says "application tried to create a window, but no driver could be loaded". Linux steam games are fine. Did some drivers get removed or something? 2024-02-27 00:12:15 working fine here... 2024-02-27 00:12:45 Actually it seems the issue is because the flatpak for proton-ge was archived. Without it I can't launch proton games (the ones native to steam don't work). Forcing flatpak to install the old version works, but obviously isn't ideal. Will update wiki if I find solution 2024-02-27 00:29:27 zcrayfish what proton are you using? A flatpak, manual install, or one of the ones that steam automatically has? I'm having trouble with everything but the old proton-ge flatpak 2024-02-27 00:39:16 The ones that steam automatically installs. 2024-02-27 00:40:18 and proton experimental (also installed via steam) 2024-02-27 01:05:21 drat. I'm not sure what's wrong with my setup then. Ah well, it isn't a good solution but I guess I'll continue with the archived flatpak proton-ge. Appreciate it though 2024-02-27 14:09:54 hello 2024-02-27 14:18:08 how is everyone doing 2024-02-27 17:48:40 anyone live 2024-02-27 17:49:10 they live 2024-02-27 17:53:51 hello my freind 2024-02-27 17:54:00 I bough a linode 4 gb ram vps 2024-02-27 17:54:06 with alpine ... 2024-02-27 17:54:20 I have created a normal user after root offcourse 2024-02-27 17:55:17 what is going on ? I create a fork in my regular user and it it is not in /home/a51... it is in .config/zsh/.zprezto 2024-02-27 17:56:39 dear lord 2024-02-27 17:56:56 I just found out alpine-zsh-config 2024-02-27 17:58:52 this is not right 2024-02-27 18:03:55 dear lord 2024-02-27 18:04:13 ls is not regular ls, busybox ls falva ... 2024-02-27 18:05:54 order restored :) 2024-02-27 19:25:20 a little mistake locked my vps ... with ssh 2024-02-27 19:25:31 is it possible to log in ? 2024-02-27 19:27:44 a51: if you've completely locked yourself out of ssh, see: https://www.linode.com/docs/products/compute/compute-instances/guides/lish/ 2024-02-27 19:27:49 this is not alpine-specific tho. 2024-02-27 20:06:39 a51: is your ups linode? 2024-02-27 20:40:19 xgpt: they did say that previously 2024-02-27 20:40:33 "I bought a linode 4 gb ram vps" 2024-02-27 20:42:52 dwfreed: look at my join time to this channel? 2024-02-27 20:46:09 most people have logs :P 2024-02-27 20:46:20 and you were in the channel previously when they said it 2024-02-27 20:47:18 ugh 2024-02-27 20:47:20 I was 2024-02-27 20:47:25 I dislike my current IRC setup 2024-02-27 20:47:28 sorry dwfreed 2024-02-27 20:47:54 weechat probably has something similar to irssi's logresume.pl 2024-02-27 21:02:34 dwfreed: 1) how did you know I'm using weechat? 2) I'm using weechat w/ weechat relay + glowing bear. It does have a fair amount of scroll back, but on reconnection it displays the backlog as this greyed out text that blends with joins/parts/quits/etc. 2024-02-27 21:02:56 it's in your quit message 2024-02-27 21:03:03 2024-02-27 18:49:13 -!- xgpt [~xgpt@000277d8.user.oftc.net] has quit [Quit: WeeChat 4.2.1] 2024-02-27 21:03:20 and oof at the greyed out part 2024-02-27 21:03:31 maybe see if that can be changed 2024-02-27 21:04:04 dwfreed: yeah, I'm actually very interested in transitioning to ZNC + graphical clients 2024-02-27 21:07:39 to keep up with missed msgs, i go to weechat's logs dir, and do, tail -n 4 -f irc.*#* 2024-02-27 21:15:06 vkrishn: that's way too much for me 2024-02-27 21:16:45 vkrishn: I take it you're just in a few channels? 2024-02-27 21:17:38 yes, :) 2024-02-27 21:22:57 ACTION looks at his 170+ windows >.> <.< >.> 2024-02-27 21:23:34 I have anxiety thinking about that idea 2024-02-27 21:27:45 170, wow!, and nobody help you reading them ;) 2024-02-27 21:30:32 makes me tired trying to imagine it 2024-02-27 22:06:28 what packages gives cat ? 2024-02-27 22:06:44 sudo apk search cat yields nothing 2024-02-27 22:06:50 dwfreed: that many windows doesn't crash my computer. It does crash my productivity and sometimes my brain 2024-02-27 22:06:57 the hardware can handle it, but the wetware can't 2024-02-27 22:07:10 a51 cat? 2024-02-27 22:07:16 busybox? 2024-02-27 22:07:37 or do you need gnu coreutils a51 ? 2024-02-27 22:07:48 a51: https://pkgs.alpinelinux.org/contents?branch=edge&name=coreutils&arch=x86&repo=main 2024-02-27 22:13:08 a51: apk search cmd:cat 2024-02-27 22:13:32 coreutils is probably the right answer 2024-02-27 22:21:27 fluix: sorry I had to put /bin/cat 2024-02-27 22:21:31 it was installed 2024-02-27 22:21:54 but cat .ssh/ide25519.pub did not work I had to /bin/cat 2024-02-27 22:23:37 sounds like your PATH is missing some entries 2024-02-27 22:28:54 it is a linode-vps 2024-02-27 22:34:58 a51: that's super weird 2024-02-27 22:35:20 a51: I always try and install my VPS from iso via the virtual consoles. 2024-02-27 22:35:47 idk if linode does, I've been using vultr instead. back when I needed quick spin VPS 2024-02-27 22:35:56 a51: I always try and install my VPS from iso via the virtual consoles. 2024-02-27 22:38:33 xgpt: how ? 2024-02-27 22:38:40 xgpt: you are with linode ? 2024-02-27 22:49:49 a51: No? I don't even work in IT. 2024-02-27 22:49:55 I'm just an alpine user 2024-02-27 22:50:15 ok 2024-02-27 22:50:53 But when I install a VPS system, I generally install from iso instead of using the provider's existing images. 2024-02-28 01:22:15 in .zshrc I have export EDITOR=vim but visudo is open in nano 2024-02-28 02:29:09 a51: export VISUAL=$EDITOR immediately after EDITOR=vim 2024-02-28 02:32:16 a51: don't forget to "source ~/.zshrc" after editing ~/.zshrc. Either that or log out and log back in. Or type "zsh" and end up with a bunch of nested sessions (yuck) 2024-02-28 02:34:43 exec zsh --login? 2024-02-28 09:58:23 Does anyone know what alpine package contains i965_dri.so? xorg seems not to start because its missing. 2024-02-28 09:59:32 that looks like mesa 2024-02-28 10:00:01 maybe mesa-dri-gallium 2024-02-28 10:00:28 $ apk info -L mesa-dri-gallium|grep -F i915 2024-02-28 10:00:28 usr/lib/xorg/modules/dri/i915_dri.so 2024-02-28 10:00:31 indeed grawity :) 2024-02-28 10:02:32 perhaps not: https://pkgs.alpinelinux.org/contents?file=i965_dri.so&path=*&name=&branch=v3.19&arch=x86_64 2024-02-28 10:03:59 oh, right, i965_dri 2024-02-28 10:04:09 didn't that become iris_dri / crocus_dri in recent versions 2024-02-28 10:04:20 still mesa either way 2024-02-28 10:06:14 the only i965 is in libva-intel-driver 2024-02-28 10:06:18 https://pkgs.alpinelinux.org/contents?file=*i965*&path=*&name=&branch=v3.19&arch=x86_64 2024-02-28 10:07:31 MaddieKalan: do you have a manually customized xorg.conf or something? 2024-02-28 10:08:36 grawity: Pretty much a clean install. 2024-02-28 10:08:48 I belive mesa-dri-gallium is installed, I'll check. 2024-02-28 10:12:09 Weird I have libva-intel-driver aswell. Wonder what is makeing xorg want i965 though... 2024-02-28 10:12:59 do you have any more logs from that "won't start because it's missing"? 2024-02-28 10:13:22 iirc some of the paths are hardcoded based on GPU manufacturer 2024-02-28 10:14:33 I'll try uploading it. 2024-02-28 10:15:47 current mesa should have iris_dri.so for "modern" intel GPUs, crocus_dri for older ones (~haswell), and i915_dri for the really old stuff (GMA) 2024-02-28 10:16:55 grawity: https://0x0.st/HRL_.0.log 2024-02-28 10:17:04 duh, wrong grep, sorry ^^ 2024-02-28 10:22:59 I'm tempted to suggest uninstalling xf86-video-intel 2024-02-28 10:23:40 (and maybe -fbdev and -vesa as well...) 2024-02-28 10:24:08 looks like the driver name is really hardcoded in it 2024-02-28 10:25:48 xf86-video-intel-fbdev isn't available? 2024-02-28 10:42:18 I mean xf86-video-intel, video-fbdev, video-vesa 2024-02-28 10:48:35 x786-video-{intel,fbdev,vesa} was already installed. 2024-02-28 10:48:45 xf86-video-{intel,fbdev,vesa} was already installed.* 2024-02-28 11:59:47 yes, I was suggesting to *uninstall* them 2024-02-28 12:14:26 I mount an sshfs filesystem from a remove server. When the system goes to sleep, this gets unmounted. Is there some way to auto-remount on wakeup? 2024-02-28 12:14:37 I highly suspect that it is not possible without providing the key again. 2024-02-28 12:14:59 -o reconnect, but yes, it needs to always have access to the credentials 2024-02-28 12:15:43 Access to credentials might be annoying since my agent prompts me each for access (which can be untimely). 2024-02-28 12:16:07 Can I somehow use a shim mount on a directory so that when I try to read it, another fuse mount is overlaid on top? Kinda like lazy-mounting? 2024-02-28 12:16:16 autofs 2024-02-28 12:16:37 it'll be a bit annoying with SSH though, as all autofs mounts necessarily happen in the context of the autofs daemon (which runs as root) 2024-02-28 12:16:49 unless you find an autofs-ish FUSE variant 2024-02-28 12:16:57 haven't really searched for any myself 2024-02-28 12:18:11 A fuse-based autofs implementation would be perfect actually. 2024-02-28 12:18:18 It's a shame that the only impl is a kernel module. 2024-02-28 12:19:18 An annoying aspect of a monolithic kernel is that filesystem drivers can just be pop'ed out and used in a userspace context :( 2024-02-28 12:20:14 it's a shame linux doesn't have any mechanism called "file systems in user space" 2024-02-28 12:22:09 (tbh I don't like sshfs in particular, it feels like... you know that Rob Pike quote about "nowadays everyone thinks 'networking' is cobbling things together with SSH and shellscripts") 2024-02-28 12:22:40 grawity: what would you use to mount a filesystem over the Internet? 2024-02-28 12:23:16 I know of fuse, but my complaint was that I can't re-use the same kernel filesystem driver in userspace. 2024-02-28 12:23:17 sshfs when I must, nfsv4 with kerberos when I can 2024-02-28 12:23:57 I don't control the remote host, and I can't expose nfsv4 as non-root :( 2024-02-28 12:25:13 (I don't put *too* much trust in the kernel's nfs server implementation though, so it's mostly over wireguard or other VPNs, but other than that it's pretty smooth over WAN, I can just `cd /n/myserver/Music` and autofs does what's needed) 2024-02-28 14:11:05 Does any of the official alpine .iso allow to boot and install using the `linux-edge` kernel? Or would I need to build a custom .iso? 2024-02-28 14:18:34 KERNEL_FLAVOR=edge setup-alpine 2024-02-28 14:18:53 But you need to make sure the community repo is enabled 2024-02-28 14:43:24 @ikke, thanks! I can then use setup-bootable to overwrite the original installation media, and then reboot to linux-edge. Or is there an easier way to reload the kernel? 2024-02-28 15:26:40 The nas I'm building is hanging shutting down. * start-stop-daemon: 8 process(es) refused to stop [ !! ]. I belive is hanging on ttyS0 VT100. 2024-02-28 15:41:00 that KERNEL_FLAVOR is some neat trivia i didn't know 2024-02-28 17:07:29 how is everyone doing ? 2024-02-28 17:08:21 I choose alpine as my linode-vps because voidlinux was not there 2024-02-28 17:09:30 for people not accustomed ... it felt alien because I was confused ... ls was not like before ... thank god googling alpine-linux normal ls gave me the answer 2024-02-28 17:12:07 coreutils? 2024-02-28 17:12:33 https://wiki.alpinelinux.org/wiki/How_to_get_regular_stuff_working 2024-02-28 17:15:21 i read that as "how to make alpine linux more like other linux" which gets into philosophy 2024-02-28 17:18:38 i read it with a different word for "stuff" 2024-02-28 17:20:38 historical vs de facto 'norms' 2024-02-28 17:22:24 invoked: since I will be using alpine more, I will be accustomed to the way it works 2024-02-28 17:23:32 but for first time users ... I think there should be an asterisk * If ytou want "regular" commands install ... 2024-02-28 17:33:23 > KREYREN_oftc: Check latest linux-edge on A64 (TERES-I) please. I think everything is there though I didn't had time to test < -- Noted, we are doing kernel development atm so i will submit adjustments to the lts and edge to make teres work on all distros and document the modules 2024-02-28 19:31:05 a51: I think that the fact Alpine uses BusyBox is very clearly explained. I knew that before I even installed my first instance. This and musl are probably the main characteristics of Alpine. 2024-02-28 19:31:47 the gnu commands are not more "regular" imo, it depends where you come from 2024-02-28 19:32:02 alpine feels a lot more bsd-like than most other linux distros do so from a bsd perspective alpine is the regular one :P 2024-02-28 22:09:08 hey alpine I've been suffering with dbus for a while now with notify-send and pipewire. 2024-02-28 22:10:18 The service is running, I launch sway with dbus-run-session, and DBUS_SESSION_BUS_ADDRESS is defined, but each time I run notify-send, I get this error: 2024-02-28 22:10:23 Error calling StartServiceByName for org.freedesktop.Notifications: Process org.freedesktop.Notifications received signal 5 2024-02-28 22:11:06 Buttt, if I run 'dbus-run-session -- sh' inside sway and run the same command, it works! 2024-02-28 22:36:49 what do you use as your notification daemon? 2024-02-28 22:37:11 ang1e: ^ 2024-02-28 22:37:21 dunst. 2024-02-28 22:39:12 It seems to catch SIGTRAP for some reason 2024-02-28 22:41:32 which shouldn't appear 2024-02-28 22:42:24 how about trying mako? I'll need to try to reproduce this issue with dunst 2024-02-28 22:46:09 what's the quick way to switch daemons? 2024-02-28 22:46:44 ang1e, do you have any dbus zombies running? 2024-02-28 22:48:13 ang1e: you can replace 'exec dunst' or what do you have in your sway config with 'exec mako' and reload sway 2024-02-28 22:52:16 no zombies I could see in htop. 2024-02-28 22:52:51 ohhh, I'll be so mad if it's just a config error in sway. 2024-02-28 23:00:40 Ermine: that's definitely what it was. 2024-02-28 23:01:11 thanks for the help. I'll probably switch to mako too. 2024-02-29 17:36:03 Hopefully they figure out the steamwebhelper thing soon, the steam-for-linux issues page is popping up with posts. I'm just glad it's not a 'damn I broke my system again' moment for once lol 2024-02-29 21:20:39 i keep getting "Couldn't connect to accessibility bus: Failed to connect to socket /var/run/user/102/at-spi/bus_0: No such file or directory", whenever i launce any gui app like gedit from cmdline 2024-02-29 21:23:32 btw, installing kate/kwrite and working went well 2024-02-29 21:23:42 in xfce4 :)