2026-02-01 04:40:08 <hans[m]1> hey guys
2026-02-01 04:40:08 <hans[m]1> however apk add linux-firmware gives read only filesystem error
2026-02-01 04:40:08 <hans[m]1> but linux-firmware IS included on the iso under /media/whatever
2026-02-01 04:40:08 <hans[m]1> linux-firmware is not installed on alpine-extended-3.23.3-x86_64.iso
2026-02-01 07:21:13 <frag> test
2026-02-01 07:23:53 <longnoserob[m]> maybe failed
2026-02-01 07:25:04 <ikke> Inconclusive 
2026-02-01 07:25:26 <jvvv> succeeded nondeterministically
2026-02-01 16:57:58 <jpp1> How do we expand the font size in the console?
2026-02-01 16:59:23 <courge_musque[m]> Maybe ctrl and +
2026-02-01 17:00:16 <lopid> move head closer to monitor
2026-02-01 17:00:26 <quinq> jaja
2026-02-01 17:01:07 <quinq> You can use setfont
2026-02-01 17:02:58 <quinq> https://wiki.alpinelinux.org/wiki/Fonts#Console_font
2026-02-01 17:41:11 <jpp1> That's it \o/
2026-02-01 17:41:29 <jpp1> thanks
2026-02-01 17:57:54 <Kammus> I can't tell you how many times per day I lean into my monitor
2026-02-01 17:58:22 <lopid> is it a secret?
2026-02-01 18:14:22 <jpp1> How to reduce the size of the font because it's a bit big now?
2026-02-01 18:14:46 <jpp1> consolefont="ter-122n.psf.gz" ?
2026-02-01 18:15:28 <Biswa96[m]> just select a smaller one. it's ok to test multiple fonts.
2026-02-01 18:18:30 <mid> jpp1: `ls /usr/share/consolefonts/` should give the list of possible font files
2026-02-01 18:40:08 <jpp1> It's good, perfect.
2026-02-01 18:40:35 <jpp1> thanks mid :)
2026-02-01 19:20:51 <markand> \o/
2026-02-01 19:21:02 <markand> algitbot, <3 
2026-02-02 09:03:50 <edd> is there a "proper" way to set zfs arc parameters at boot on alpine? besides using a local.d script to poke the /sys/... files for zfs?
2026-02-02 10:02:34 <quinq> edd, like any other module, you should be able to set the module parameters in /etc/modprobe.d/
2026-02-02 10:03:42 <quinq> Also depending in you need it from initrd, you might have to add that file to the initrd image
2026-02-02 10:04:36 <edd> right, i read about this and found that somewhat scary
2026-02-02 10:05:17 <edd> seems easier to just poke the file in a local.d
2026-02-02 10:05:31 <quinq> Humm, no it's already in the base mikinitfs config: /etc/mkinitfs/features.d/base.files:/etc/modprobe.d/*.conf
2026-02-02 10:05:44 <quinq> edd, what seems scary?
2026-02-02 10:05:59 <edd> fiddling with initrd
2026-02-02 10:06:12 <quinq> You don't have to
2026-02-02 10:06:31 <edd> ok, maybe i'll try again later
2026-02-02 10:06:49 <edd> my initial attempt to add a zfs.conf didn't seem to work, but i probably screwed up
2026-02-02 10:07:10 <edd> i didn't regenerate any initrd stuff or anything, so maybe that's what
2026-02-02 10:07:58 <quinq> It's “very” simple: options module param=value
2026-02-02 10:08:18 <quinq> Where module would be zfs, param, whatever param you want to set, and value, well the actual value you want to set :)
2026-02-02 10:08:25 <quinq> One line for each
2026-02-02 10:08:36 <edd> and then just reboot?
2026-02-02 10:08:48 <quinq> Yeah edd, if you already have zfs in your initrd that would be the issue
2026-02-02 10:09:09 <edd> ok, i'll have a play later
2026-02-02 10:09:09 <quinq> Just run mkinitfs, done
2026-02-02 10:09:15 <quinq> Good luck! :)
2026-02-02 10:09:22 <edd> oh, well that's the missing step `mkinitfs`
2026-02-02 10:09:35 <edd> thanks for this
2026-02-02 10:09:51 <edd> i'm coming from another unix and this initrd stuff is all very alien
2026-02-02 10:10:12 <quinq> Yeah, each system has its own configuration and commands, but in the end it's the same concepts
2026-02-02 10:10:27 <quinq> https://wiki.alpinelinux.org/wiki/Initramfs_init maybe that can help
2026-02-02 10:11:13 <quinq> Though the example might be a bit misleading, it's not necessary to pass all those options, that's how it works by default
2026-02-02 10:11:28 <quinq> Just check what it does on the output, it'll tell you what it has generated
2026-02-02 10:11:55 <quinq> Ah no sorry, it says in the fine print afterwards “Therefore normally just running mkinitfs should be enough if you need to rebuild it.”
2026-02-02 10:18:36 <marcc> Heya, I'm trying to add a package and struggling with the doc part. The man page gets installed when running make install, and it's just a static file. Should I still add a subpackage with the suffic -doc then?
2026-02-02 10:18:54 <marcc> and should the function be `doc() { default_doc }` ?
2026-02-02 10:19:24 <ikke> marcc: not sure what you mean with a static file, but just adding the -doc subpkg should suffice
2026-02-02 10:20:00 <ikke> There is already a `doc` function in abuild, you only need to define it if you need to do something not covered by default_doc
2026-02-02 10:21:44 <marcc> ikke As in the file is just `rc.1`, when you run make install that file is moved to MANPREFIX/man1. I'm getting an error rc-doc*: Missing subpkgdir for rc-doc
2026-02-02 10:22:39 <marcc> But I'm guessing then I shouldn't need to define a doc function
2026-02-02 10:22:42 <ikke> That means the file is not installed as the default_doc function expects
2026-02-02 10:22:50 <marcc> ah
2026-02-02 10:24:42 <marcc> So then running Where can I read up on the default_doc function?
2026-02-02 10:24:56 <marcc> Where can I read up on the default_doc function?**
2026-02-02 10:34:31 <ikke> https://gitlab.alpinelinux.org/alpine/abuild/-/blob/master/abuild.in#L1947
2026-02-02 16:16:31 <runxiyu> omni: could you review https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/96836 please? thanks!
2026-02-02 16:58:23 <nephele_> Hey there, I'm looking for some help with a zfs problem. "zpool status" reports "label is missing or invalid" and for one device "was /dev/sdb1"; But /dev/sdb1 still exists! and i can't use zfs replace because it is already part of the pool
2026-02-02 16:58:57 <nephele_> I don't see an indication that the drive is damaged
2026-02-02 16:59:00 <dwfreed> you would probably have better luck asking the zfs people directly
2026-02-02 19:42:00 <realroot[m]1> >It requires the user to have libdbus system liblary (should be available on any systemd distro)
2026-02-02 19:42:00 <realroot[m]1> is that for systemd only so?
2026-02-02 19:42:14 <achill> no
2026-02-02 19:43:20 <realroot[m]1> >It requires the user to also have either the GTK, GNOME, or KDE XDG Desktop Portal backend installed at runtime.
2026-02-02 19:43:20 <realroot[m]1> is that wayland thing ?
2026-02-02 19:44:44 <realroot[m]1> it(rfd) uses zenity fallback and background is transparent
2026-02-02 20:17:32 <frag> in shell, $_ refers to the last word in the last command, it there a var for the previous word (in the current line)?
2026-02-02 20:18:52 <quinq> s/shell/some &/
2026-02-02 20:19:52 <frag> so you can: mv myfile ${lastword}.txt, or $(date +%s)${lastword}
2026-02-02 20:20:42 <lopid> alt-. ?
2026-02-02 20:21:46 <quinq> Honestly I don't undertsand what you mean, but no it doesn't exist
2026-02-02 20:22:04 <lopid> me neither. i thought i might until those examples
2026-02-02 20:23:07 <quinq> :)
2026-02-02 20:23:43 <quinq> Maybe like reusing the result of the last command, even if it's on the same line?
2026-02-02 20:24:43 <frag> so `mv myfile ${lastword}.txt` would be `mv myfile myfile.txt`
2026-02-02 20:25:17 <lopid> oh i just delete myfile and ctrl-y to paste it twice
2026-02-02 20:25:33 <lopid> ctrl-w ctrl-y ctrl-y 
2026-02-02 20:25:58 <frag> must be nice to have ctrl-y :c
2026-02-02 20:26:04 <lopid> it is
2026-02-02 20:26:08 <frag> busybox ash doesnt :/
2026-02-02 20:26:26 <lopid> useful to know that at the start
2026-02-02 20:28:17 <frag> huh, dash has ^w
2026-02-02 20:29:38 <quinq> Yeah, do htat, ^w ^y ^y :D
2026-02-02 20:30:54 <quinq> frag, I usually do foo{,.new}
2026-02-02 20:31:06 <quinq> But that's not portable either (end scandaleous that it isn't!)
2026-02-02 20:31:17 <lopid> yeah, htat too
2026-02-02 20:31:23 <quinq> s/end/and/
2026-02-02 20:33:18 <quinq> Alternatively
2026-02-02 20:33:18 <frag> quinq is that mksh? whats a working example?
2026-02-02 20:33:21 <quinq> vi mode
2026-02-02 20:33:29 <quinq> There you cal del and paste
2026-02-02 20:33:50 <humm> Alternatively
2026-02-02 20:33:53 <humm> emacs mode
2026-02-02 20:34:05 <humm> There you may be cble to del and paste
2026-02-02 20:34:16 <quinq> But not to type “able”
2026-02-02 20:34:34 <humm> But not type “can”
2026-02-02 20:36:15 <quinq> Typing is herd
2026-02-02 20:42:32 <jvvv> a herd of penguins hit it like a gnu
2026-02-02 21:56:45 <omni> runxiyu: yes, sorry, I'd just like to try and enable tests
2026-02-02 21:59:27 <runxiyu> omni: its alr!
2026-02-02 22:09:52 <omni> I'll do that in a separate MR later
2026-02-02 22:11:15 <shadow255>  /12
2026-02-02 22:12:44 <omni> runxiyu: just an FYI, I've sometimes thought of splitting it into multiple subpackages
2026-02-03 02:08:08 <WhyNotHugo> SyntheticBird[m] ffoss: dbus-broker needs a "controller", which is glue code to tell a service manager to activate a service. There's one for systemd, but there's also another for dinit IIRC. Should be feasible to write one for OpenRC.
2026-02-03 07:39:12 <runxiyu> omni: makes sense
2026-02-03 08:39:49 <thirdwhl> Hey, so I was trying to do an in place upgrade and went a few too many versions ahead at which point it threw an error about unresolvable dependencies when I tried to update apk-tools. When I then tried to upgrade to the next version it still says there's problems but now it says "Huh? Error reporter did not find the broken constraints" - can I come back from this or will I have to burn it down and start over?
2026-02-03 08:42:01 <thirdwhl> "apk can't find the broken dependencies. there are now three versions of apk-tools installed. in the distance, sirens"
2026-02-03 08:44:26 <thirdwhl> When I did the googling I should have done in the first place, I found a bug report that suggests this is an issue with 3.20
2026-02-03 08:44:30 <thirdwhl> which I am running
2026-02-03 08:51:14 <thirdwhl> ok to be fair the person reporting the issue was also running a few edge/testing packages so not exactly an apples to apples comparison
2026-02-03 08:52:18 <thirdwhl> ok going back to the original repos clears things up
2026-02-03 08:53:13 <thirdwhl> Switching back and it's still not having it
2026-02-03 08:54:26 <thirdwhl> At least it's good to know I haven't borked this thing, maybe I'm just stuck on 3.20 until I decide to nuke this thing from orbit
2026-02-03 09:01:58 <thirdwhl> oops. accidentally did an 'apk upgrade' instead of an update. whelp. either I've done what I set out to do or I've got to nuke this from orbit anyway. have a day folks!
2026-02-03 09:16:20 <ikke> I've upgraded a server from 3.19 to  3.23 the other day without issues
2026-02-03 09:16:46 <ikke> One good step would be to make sure you don't have any packages installed (in /etc/apk/world), which are no longer providd
2026-02-03 09:16:59 <ikke> Also make sure you always use --available when upgrading
2026-02-03 09:17:28 <dwfreed> ikke: note they're gone
2026-02-03 09:17:39 <ikke> dwfreed: oh, thanks for mentioning
2026-02-03 09:17:54 <dwfreed> they quit 10 seconds after their last line
2026-02-03 09:18:14 <ikke> I see the notification now, but it was  too low contrast on this client for me to notice
2026-02-03 09:56:39 <SyntheticBird[m]> WhyNotHugo, interesting. I didn't know
2026-02-03 11:27:00 <frag> echo ';forøvrig'|busybox grep -iw ";for"
2026-02-03 11:27:16 <frag> busybox grep fail here ^
2026-02-03 11:27:44 <frag> sbase/gnu grep works (or at least what how i expect)
2026-02-03 11:30:08 <lopid> works here with BusyBox v1.37.0
2026-02-03 11:31:42 <frag> lopid i expect to not get output, you get output from that command?
2026-02-03 11:32:03 <frag> i have same version, and using C locale, dunno if that matters
2026-02-03 11:32:17 <lopid> why would you expect that?
2026-02-03 11:32:54 <ikke> I guess due to `-w`
2026-02-03 11:33:08 <lopid> oic
2026-02-03 12:51:32 <frag> maybe they even have improved it lately, im quite sure i found examples where it doesnt or didnt work
2026-02-03 12:51:42 <frag> damn, wrong window
2026-02-03 13:41:13 <frag> gnugrep goesnt support '\t' without -P so have to use foo'$'\t''bar :/
2026-02-03 13:53:57 <humm> frag: or you just use a literal tab
2026-02-03 13:54:35 <frag> hmm yes interesting, another thing busybox ash doesnt support tho ;p
2026-02-03 14:04:40 <frag> yash is interesting, bit flashy and you have to config it to remove intro text and to get ^L working, nice completion(?)
2026-02-03 14:05:25 <frag> and it doesnt show full PWD as default
2026-02-03 15:11:43 <Sertonix[m]> frag: you can write a literal tab in ash scripts without issues
2026-02-03 15:12:21 <frag> in an editor yeah
2026-02-03 15:15:47 <frag> might be a good solution
2026-02-03 19:59:04 <frag> how to kill a process that wount die with kill -9 PID :S
2026-02-03 20:06:03 <ikke> Is in Disk / io sleep?
2026-02-03 22:50:31 <Antiz1> 👋
2026-02-03 22:51:57 <Antiz1> Anyone has ever achieved running Arch Linux (or any other OS requiring cgroupv2) has an *unprivileged* LXC container on Alpine?
2026-02-03 22:53:44 <Antiz1> As far as I understand, OpenRC doesn't handle cgroup delegation. 
2026-02-03 22:54:18 <Antiz1> I've tried to remount /sys/fs/cgroup with nsdelegate, but no success either.
2026-02-03 22:56:01 <Antiz1> I'm also hitting the same error when running starting the container: "Failed to mount cgroup2 (type cgroup2) on /sys/fs/cgroup (MS_NOSUIDIMS_NODEVIMS_NOEXEC "nsdelegate,memory_recursiveprot,memory_hugetlb_accounting"): Permission denied"
2026-02-03 22:57:52 <Antiz1> I managed to start a Debian container (which still accepts cgroup v1) by mounting /sys/fs/cgroup/systemd with the correct option & permissions. But no luck with Arch Linux (which only accepts cgroup v2 due to its recent version of systemd).
2026-02-03 22:59:48 <Antiz1> I'm now wondering if it's even possible at all to start an *unprivileged* LXC container running any systemd-based OS that requires cgroup v2 on Alpine (which means, all systemd-based OS once they switched to systemd v258+)
2026-02-03 23:00:52 <Antiz1> Though, if anyone every achieved this somehow, I'd be curious (and happy) to know how :)
2026-02-03 23:00:58 <Antiz1> s/every/ever
2026-02-04 00:19:18 <CRCinAU> So, I'm wondering..... systemd-networkd is not a thing available on Alpine, right?
2026-02-04 00:19:59 <pj> correct
2026-02-04 00:19:59 <CRCinAU> and if not, how do people handle IPv6 + DHCPv6-PD properly if they use Alpine as a 'router' type install?
2026-02-04 00:20:25 <CRCinAU> I say that as a lot of the 'old' tools just don't quite work properly in this scenario :(
2026-02-04 00:21:41 <CRCinAU> I mean, the basic router parts could be dnsmasq + nftables, but I've always struggled with random scripts and addon crap to make it all work properly when not using systemd-networkd (which annoyingly works almost out of the box)
2026-02-04 00:23:24 <CRCinAU> I redid my 'router' to use OPNSense, and that works OOTB. but I'm getting the itch to go back to nftables etc ;)
2026-02-04 00:24:05 <CRCinAU> I think I've prefer an Alpine install over Fedora to make things a bit more lightweight...... but not sure it'd work as easily.
2026-02-04 01:59:07 <Renan[m]> how i can use system setting proxy in alpine?
2026-02-04 03:18:36 <u0_a617> Woah, good old irc chat on irssi, i'm happy! Good to be back! Hello Gents, german informatician, fascinated by linux since age 10 always went back to microsoft because i had to, now not anymore, want to setup an alpine xen dom0 with fedora and another alpine, struggling on how to set up german keyboard layout de-neo on lightdm, set up in /etc/lightdm/lightdm.conf, searched display-setup-script and 
2026-02-04 03:18:42 <u0_a617> put /usr/bin/bash -c "setxkbmap de neo >/dev/null 2>&1 || :" result after lightdm service restart: a blinking cursor on tty7, thanks already for your help!
2026-02-04 03:20:16 <nanosleep> u0_a617, do you know if your lightdm is running under wayland or xorg? If it's xorg, you can probably force the keyboard map in an xorg.conf
2026-02-04 03:20:54 <nanosleep> on wayland.... I would hope there's an option in the config for that, lol
2026-02-04 03:21:41 <u0_a617> nanosleep: you meant that regards my question?
2026-02-04 03:23:31 <nanosleep> yes, if your lightdm greeter is running under xorg, it should inherit from any settings set in xorg.conf -- so you should be able to set the XkbLayout in an xorg.conf snippet. If it's running under wayland (which is also technically possible), I think it would be responsible for implementing the keyboard map itself
2026-02-04 03:26:43 <u0_a617> nanosleep: great, i tried to follow some steps on forums, some chatbot steps, but tbh you can forget about ai's steps, it's complete nonsense. Wasted 14! days trying to install fedora before with ai on zfs and then i figured ill install alpine xen and on top fedora on zfs. Thanks buddy, ill try the xorg thing, but i have actually only found an xinit.org hmm
2026-02-04 03:29:03 <nanosleep> google or claude something like "xorg.conf.d/20-keyboard.conf" and "XkbLayout" and you'll get where you need to be probably
2026-02-04 03:30:14 <nanosleep> scratch that. here's an ancient stackoverflow post that probably will do the trick -- https://unix.stackexchange.com/questions/130858/change-the-keyboard-layout-in-lightdm
2026-02-04 03:31:28 <u0_a617> nanosleep: That's exactly the page i tried but the display rule
2026-02-04 03:32:09 <u0_a617> nanosleep: i also only have the path /etc/X11/xinit/
2026-02-04 03:33:19 <nanosleep> xorg will always look for an /etc/X11/xorg.conf -- you could potentially just create that file and stick it in there
2026-02-04 03:33:45 <nanosleep> https://wiki.alpinelinux.org/wiki/Xorg
2026-02-04 03:37:01 <u0_a617> nanosleep: Great! Thank you! I have yet another question: I have an usb dongle realtek rtw89_8852bu and i already figured how to switch its disk mode to wireless mode, also got the managed mode of networkmanager to run, I only still don't understand why at the beginning I don't get any ping.
2026-02-04 03:39:39 <nanosleep> sorry I'm not clear on what you're asking exactly. You can't ping anything?
2026-02-04 03:40:06 <u0_a617> nanosleep: Right. 
2026-02-04 03:41:43 <u0_a617> nanosleep: Well, i need to express that i have been able to ping my router now (since my system is running again a few minutes again), but no dns resolving.
2026-02-04 03:42:52 <nanosleep> the obvious questions are what does the networkmanager network config look like (is it getting dns, or is it manually set? disabled?).. and then also, what does your /etc/resolv.conf look like?
2026-02-04 03:46:21 <u0_a617> resolv.conf has an empty file only with a comment # generated by networkmanager, so i think it should be managed by it, I have no NetworkManager.conf anymore since I wrote one and then my eth0 and wlan0 interfaces vanished completely (i guess because they were rmmod'ed)
2026-02-04 03:46:43 <u0_a617> nanosleep: not by me though
2026-02-04 03:47:43 <nanosleep> you should probably be using nmtui or nm-applet (if you're in a gui). Ideally, don't try to configure it by hand
2026-02-04 03:48:34 <nanosleep> and your resolv.conf being empty is the problem. Your networkmanager connection isn't getting dns (either because DHCP isn't sending it, or because it's been told to do manual dns config/disabled). You can confirm this by manually sticking a line like 'nameserver 1.1.1.1' in your resolv.conf
2026-02-04 03:49:05 <nanosleep> networkmanager is managing the file though, so it'll be reset the next time you connect. You need to fix the underlying issue with the connection in networkmanager
2026-02-04 03:50:26 <u0_a617> nanosleep: oh wow i guess that did it, or the networkmanager restart
2026-02-04 03:54:26 <u0_a617> nanosleep: Any idea from someone more knowladgeable than me about linux how to solve the most linux problems actually with the good old way (without ai)? maybe you have some hints, i want to get really good finally because i have already loved linux always more than windows just the dependencies of proprietary solutions kept me at windows, you know?
2026-02-04 03:57:22 <nanosleep> I think just keep doing what you're doing. Find a tech community of linux desktop users and learn through osmosis. Learning linux as a desktop is a daunting task because there are so many little things you have to learn, and they're changing all the time. You could definitely make the argument that it isn't worth it for most people unless there's a particular reason they're willing to endure the suffering
2026-02-04 03:59:32 <u0_a617> nanosleep: I choose the suffering and hard learning, because many people already give up when its hard and i have already learned many lessons, i am also proud on my system that im building up myself. May i know which irc client you are using?
2026-02-04 04:01:05 <nanosleep> I'm going to lose geek cred for this, but currently I'm chatting from hexchat
2026-02-04 04:01:55 <u0_a617> Is this a linux program?
2026-02-04 04:03:00 <nanosleep> yeah, it's a graphical client
2026-02-04 04:03:09 <u0_a617> nanosleep: Come on, we're in the same situation, i'm not actually very proud that i have used microslop for decades, but now i can finally migrate to linux
2026-02-04 04:08:56 <u0_a617> nanosleep: Thank you for your help, i'll try that now. Have a great day or night depending on where you are based.
2026-02-04 04:09:15 <nanosleep> likewise, np, glad I could help!
2026-02-04 09:11:22 <maduncle94[m]> Would you guys confirm that gpm service works? Specifically for alpine edge version. Thanks.
2026-02-04 10:38:48 <mutable_reference[m]> Hi everybody!... (full message at <https://matrix.org/oftc/media/v1/media/download/AWInGe6hN2eoTkpz1iy1VfTq_yScbYdUWoHUX_ex891QsyvS_aYx6_9Hb2gYDUeEuYq7eAiOF2RcbH4IoDoZhMpCeccFTBmgAG1hdHJpeC5vcmcvZkt2V0RvYXJqdkNOa1lMVXRHQWhxU2RR>)
2026-02-04 10:58:29 <pj> mutable_reference: FYI, you most likely won't receive a response because your message is so big it got truncated on the IRC side to only "Hi everybody!..."
2026-02-04 10:58:32 <pj> it's best to keep messages short
2026-02-04 10:59:03 <mutable_reference[m]> So it's best to just send individual short messages
2026-02-04 10:59:05 <mutable_reference[m]> like this
2026-02-04 10:59:08 <pj> yes
2026-02-04 10:59:22 <mutable_reference[m]> Thansk! 
2026-02-04 10:59:31 <mutable_reference[m]> I'm trying to setup a cache for apk for my local network so that all my computers can update without having to hit the upstream repo.
2026-02-04 10:59:42 <mutable_reference[m]> I have setup an nginx reverse proxy for this. I cache all requests for .apk files but not other requests so that e.g. the package list stays up to date.
2026-02-04 10:59:59 <mutable_reference[m]> Does it really work tho? Like am i getting something wrong
2026-02-04 11:00:15 <pj> (IRC<>matrix bridge should handle a bit bigger messages but it's up to 10 lines only)
2026-02-04 11:01:15 <mutable_reference[m]> Also is there a way to configure alpine to fallback to another mirror if one cannot be reached? I would like this for the case of e.g. my laptop so that when it is in my home network it'll fetch the updates from alpinecache.localdomain but when i'm out and about that domain name obviously does not resolve and alpine fails to update and does not fallback to the other mirrors and i need to manually edit /etc/apk/repositories
2026-02-04 11:05:24 <pj> for fallback I think you can just list each mirror and it will gracefully use the available one
2026-02-04 11:05:44 <pj> given that apk is used with --available
2026-02-04 11:06:25 <pj> and --latest
2026-02-04 11:13:34 <mutable_reference[m]> Sorry i don't appear to be getting --available to work. When i run apk add libgcc --no-check-certificate --latest  --available i get an error ERROR: command line: unrecognized option 'available'
2026-02-04 11:18:00 <mutable_reference[m]> Oh maybe `apk upgrade --available`? that runs but i still get `ERROR: libapk-3.0.4-r0: DNS: name does not exist` and if i run it again it again tries to do the same update again e.g `apk-tools (3.0.3-r2 -> 3.0.4-r0)` so the update does to appear to work
2026-02-04 12:51:39 <skully> hi, question about libexpat packaging: newer versions are available on several arches, but not on x86_64. is there a known reason, or is it simply not rebuilt yet?
2026-02-04 13:43:03 <jpp1> skully: https://pkgs.alpinelinux.org/packages?name=libexpat
2026-02-04 13:51:30 <skully> @jppl sorry, I was not accurate, 3.22 version, 64 bit exclusively, everything else seems fine, every other version is 2.7.4, just 64 bit is 2.7.3
2026-02-04 13:52:36 <skully> jpp1: ^
2026-02-04 13:54:53 <Sertonix[m]> skully: see #17938
2026-02-04 13:57:22 <skully> awesome, thanks very much!
2026-02-04 16:18:19 <realroot[m]> zenity is transparent here
2026-02-04 16:18:27 <realroot[m]> zenity --calendar etc.
2026-02-04 19:39:59 <runxiyu_> 3.22.3 installation medium works perfectly on my machine
2026-02-04 19:40:10 <runxiyu_> 3.23.3's fails to find the iwlwifi firmware file
2026-02-04 20:34:44 <runxiyu> will debug in detail later 
2026-02-05 02:49:17 <Renan[m]> [bluetoothctl]> scan on
2026-02-05 02:49:17 <Renan[m]> SetDiscoveryFilter failed: org.bluez.Error.NotReady
2026-02-05 02:49:17 <Renan[m]> Failed to start discovery: org.bluez.Error.NotReady
2026-02-05 02:49:19 <Renan[m]> someone help me
2026-02-05 02:50:13 <Renan[m]>         Hard blocked: no
2026-02-05 02:50:13 <Renan[m]> hopeful$ rfkill list
2026-02-05 02:50:13 <Renan[m]> 1: phy0: wlan
2026-02-05 02:50:13 <Renan[m]>         Soft blocked: no
2026-02-05 02:50:13 <Renan[m]> 18: hci0: bluetooth
2026-02-05 02:50:15 <Renan[m]>         Soft blocked: no
2026-02-05 02:50:15 <Renan[m]>         Hard blocked: no
2026-02-05 03:16:40 <JustSouptheyhe[m]> <Renan[m]> "[bluetoothctl]> scan on..." <- This is going to sound basic, but is the service running? The `NotReady` makes it seem like it could be stopped / hanging somewhere.
2026-02-05 14:06:49 <WhyNotHugo> For a package with 'makedepends="vulkan-loader-dev', abuild is tracing dependency 'vulkan-loader', but the package does not depend on vulkan-loader and does not link to it; only a subpacakge requires vulkan. Why is this being traced automatically?
2026-02-05 14:08:49 <ikke> Are you sure there is no hidden dependency?
2026-02-05 14:09:01 <ikke> soname, symlink or anything?
2026-02-05 14:09:37 <WhyNotHugo> ldd doesn't show anything depending on it. https://paste.sr.ht/blob/ac7078151f2733cc9afff08885a7141f957dc5e2
2026-02-05 14:10:41 <WhyNotHugo> The only non ".so" file is LICENCE
2026-02-05 14:11:18 <WhyNotHugo> 🤦 I put it in depends myself
2026-02-05 14:11:25 <ikke> -)
2026-02-05 14:11:27 <ikke> :-)
2026-02-05 14:11:46 <ikke> The hidden dependency 😛
2026-02-05 14:12:08 <WhyNotHugo> Problem was on top of the chair again
2026-02-05 14:12:21 <longnoserob[m]> level 8 issue
2026-02-05 14:15:27 <amadaluzia> Hey :) does Alpine Linux work in non SSE2 systems?
2026-02-05 14:15:40 <ikke> No
2026-02-05 14:15:59 <amadaluzia> Ah, gotcha
2026-02-05 14:16:11 <amadaluzia> Just wanted to make sure
2026-02-05 14:16:18 <ikke> https://gitlab.alpinelinux.org/alpine/tsc/-/issues/35
2026-02-05 14:17:30 <amadaluzia> > Error: access denied: denied by administrative rule 64b27fa8811729dbafef17bfccb6299e/f9eca3ed223160842689
2026-02-05 14:18:37 <ikke> Do you have a request id?
2026-02-05 14:19:46 <ikke> Oh, I see it's combined, sorry
2026-02-05 14:22:22 <ikke> amadaluzia: Do you have cookies disabled?
2026-02-05 14:23:00 <amadaluzia> https://disroot.org/upload/019c2e2f-55e3-7d8b-80e4-992c69212235/zb2rhYWLkYxbfscDUn3vUcfvUueU1A78nQt9adh6vJMHoGuAo.png
2026-02-05 14:24:02 <ikke> Does it happen again on the 2nd try?
2026-02-05 14:24:32 <ikke> It does not require third-party cookies, but first-party cookies must be enabled
2026-02-05 14:25:37 <amadaluzia> Yea
2026-02-05 14:25:58 <amadaluzia> Happening under Vanadium on GrapheneOS
2026-02-05 15:21:43 <zooma> hello !
2026-02-05 15:23:06 <zooma> please i've got some issue on downlding alpine-virt-3.23.3-x86_64.iso on official site. the sha256sum check doesn t match ....
2026-02-05 15:23:28 <zooma> did s.o have the same ?
2026-02-05 15:33:38 <ikke> Oh, already gone
2026-02-05 15:33:41 <ikke> But I cannot reproduce
2026-02-05 15:33:45 <ikke> checksums check out
2026-02-05 16:42:35 <WhyNotHugo> Seeing error "provides multiple versions of same shared object: so:", but then no so is listed as duplicate: https://gitlab.alpinelinux.org/WhyNotHugo/aports/-/jobs/2207438#L1402
2026-02-05 16:42:36 <WhyNotHugo> Is this a bug?
2026-02-05 17:05:48 <realroot[m]> does /etc/sysctl.conf with vm.swappiness=10 work in alpine?
2026-02-05 17:07:14 <courge_musque[m]> Well, it's from the kernel, why wouldn't it?
2026-02-05 17:07:36 <realroot[m]> i remember some error how do i check it is ok?
2026-02-05 17:07:41 <realroot[m]> *if it is
2026-02-05 17:08:22 <ikke> sysctl vm.swappiness
2026-02-05 17:09:51 <realroot[m]> it seems fine thanks
2026-02-05 17:11:00 <realroot[m]> are mkinitfs hooks available now?
2026-02-05 17:12:11 <ikke> There's no plugin system yet
2026-02-05 18:12:39 <wozniak> hi, I am trying to use andrewfs and it looks like the modules for it aren't built in the default kernel. Would it be accepted if I submitted a pr to aports that just enabled building them
2026-02-05 18:14:21 <wozniak> https://docs.kernel.org/filesystems/afs.html
2026-02-05 18:25:50 <ikke> seems to be still under development
2026-02-05 18:27:40 <dwfreed> afs has existed for ages
2026-02-05 18:27:58 <ikke> https://docs.kernel.org/filesystems/afs.html
2026-02-05 18:28:17 <ikke> says it's still under development
2026-02-05 18:28:22 <dwfreed> more likely the docs haven't been touched other than to convert them to rst
2026-02-05 18:28:33 <dwfreed> you know how programmers are with docs
2026-02-05 18:31:07 <dwfreed> https://github.com/torvalds/linux/commits/ca6e9049a0934fe72ffea6990c889205aff0a2cf/Documentation/filesystems/afs.txt?browsing_rename_history=true&new_path=Documentation/filesystems/afs.rst&original_branch=master
2026-02-05 18:31:18 <dwfreed> not many changes
2026-02-05 18:32:05 <wozniak> yeah looks like the afs impl itself has been way more active than the docs lol
2026-02-05 21:04:24 <RoadRoller01[m]> how could i make abuild temp dirs be in /tmp ? to be faster/not over writing on my nvme
2026-02-05 21:07:42 <visone[m]> RoadRoller01[m]: Mount the directory where the build is done on /tmp
2026-02-05 21:07:42 <visone[m]> ```
2026-02-05 21:07:42 <visone[m]> tmpfs		/var/cache/scratchpkg/work	tmpfs	noatime,size=72G 0	0
2026-02-05 21:07:42 <visone[m]> I have this line in venom for exactly that
2026-02-05 21:07:42 <visone[m]> ```
2026-02-05 21:08:09 <visone[m]> ```
2026-02-05 21:08:09 <visone[m]> * Mount the directory where the build is done on /tmp
2026-02-05 21:08:09 <visone[m]> tmpfs		/var/cache/scratchpkg/work	tmpfs	noatime,size=72G 0	0
2026-02-05 21:08:09 <visone[m]> I have this fstab line in venom for exactly that
2026-02-05 21:08:09 <visone[m]> ```
2026-02-05 21:11:03 <RoadRoller01[m]> Ohh man isn't there a better way):
2026-02-05 21:20:09 <dwfreed> build chroot is hardcoded to be in /var/tmp, where a random directory is created with mktemp
2026-02-05 22:40:37 <Sertonix[m]> The current way that rootbld works requires this, unfortunatly
2026-02-06 03:09:49 <Saijin_Naib[m]> If I need easy hands-off full-disk encryption of an existing disk, what is my path? I am using UEFI XFS over LVM2 with grub
2026-02-06 03:17:23 <Saijin_Naib[m]> Looks like none 😑
2026-02-06 03:38:01 <invoked> Saijin_Naib[m]: you will get a range of opinions here
2026-02-06 03:40:59 <invoked> my opinion: i would use lvm1 with ext4 in most cases. it's as well tested as it gets
2026-02-06 03:47:50 <invoked> s/lvm1/luks1/
2026-02-06 04:41:00 <Saijin_Naib[m]> Haha, it would not be other people if there weren't variable opinions
2026-02-06 04:41:24 <Saijin_Naib[m]> Okay, that sounds reasonable, but a constraint I have is that it needs to be in-place, which for all the devices I manage means XFS over LVM2
2026-02-06 04:41:31 <Saijin_Naib[m]> And no data-loss/re-format
2026-02-06 04:41:43 <Saijin_Naib[m]> I'm gathering it is not possible in-place
2026-02-06 05:08:53 <invoked> there's not any way to do that which i know of.
2026-02-06 05:27:45 <Saijin_Naib[m]> Yeah, crap. Ah well. Was just to lower my insurance, haha
2026-02-06 07:40:47 <bananicorn> After upgrading my laptop and server and a bunch of lxc containers (all running alpine) I am here to say that 4 years in I still love alpine just as much, thanks for your work everyone :)
2026-02-06 07:41:36 <runxiyu_> yay
2026-02-06 07:42:14 <bananicorn> Whenever I am forced to use anything else, I miss apk
2026-02-06 07:45:26 <runxiyu_> there are other distros that use apk btw, see chimera (which also uses v3 apk formats unlike alpine which uses v2)
2026-02-06 10:02:53 <bananicorn> runxiyu_: Chimera looks interesting, I was very in love with OpenBSD before I found Alpine so the BSD userland sounds appealing - might give it a try at some point
2026-02-06 10:03:46 <bananicorn> On another note, I am playing with apk-audit(8) right now and would love to see what actually changed in a file (as a diff). Is there a command I am missing for this?
2026-02-06 10:20:07 <bananicorn> (I ended up doing an apk fix -r to reinstall and get the .apk-new to diff with)
2026-02-06 17:13:58 <fckai> HI folks, have a great day. I have a issue configuring The rc-update is unable to setup the service because the /var/run/supervise-docker.pid don't exists. Who is responsible to create these pid file ? 
2026-02-06 17:16:21 <fckai> the concrete output of the rc-update is supervise-daemon: fopen `/var/run/supervise-docker.pid': No such file or directory
2026-02-06 17:20:04 <JustSouptheyhe[m]> Does /var/run/ exist? The pid file should be automatically handled if the service is written right.
2026-02-06 17:23:48 <fckai> yeah the /var/run directory exists and the other s services pid files are ok. only the docker on fail.
2026-02-06 17:24:48 <penguinz_rule[m]> this might be punting the problem down the road but start_pre with a touch command might let it wor k
2026-02-06 17:29:21 <penguinz_rule[m]> disregard what i said
2026-02-06 17:49:03 <fckai> I solve it by re-installing the package but don't started like it say the wiki, just install the package, add the service with rc-update and reboot. running service start somehow brake the install
2026-02-06 19:02:07 <alimj[m]> fcolista: Hi. If I am not mistaken, you might be the package maintainer for `glances` on Alpine.
2026-02-06 19:04:43 <alimj[m]> In that case, it might be a good idea to add py3-jinja2 to the glaces requirements.
2026-02-06 21:06:39 <maria[m]> Hi! After updating Chromium on Edge, I'm experiencing the same issue as in https://gitlab.alpinelinux.org/alpine/aports/-/issues/13579 -- random crashes and "Aw snap" on big websites. `--no-sandbox` fixes it for me. I am on a thinkpad t480, mesa-dri-gallium and mesa-vulkan-intel are installed
2026-02-06 21:06:43 <maria[m]> What data should I give? Thanks for help
2026-02-06 23:07:38 <ikke> lnl: ^
2026-02-07 09:48:56 <frag> hmm repo `file` only returns exit status above 0 when file not found if you add the -E option, posix man file specify >0 if "An error occurred", is repo `file` not posix then?
2026-02-07 10:46:07 <quinq> frag, “If the file named by the file operand does not exist, cannot be read, or the type of the file named by the file operand cannot be determined, this shall not be considered an error that affects the exit status.”
2026-02-07 10:47:26 <quinq> And the file from repo doesn't have -E option
2026-02-07 10:47:48 <quinq> $ apk info -W /usr/bin/file
2026-02-07 10:47:48 <quinq> /usr/bin/file is owned by file-5.46-r2
2026-02-07 10:47:55 <frag> quinq aha!
2026-02-07 10:47:59 <quinq> $ file --help|grep -F -e -E
2026-02-07 10:48:01 <quinq> nothing
2026-02-07 10:48:40 <quinq> Where did you see it was supposed to?
2026-02-07 10:49:12 <frag> its not listed but it works :) try : file -E sdfsdfsdf||echo fail .. saw it in an online man page
2026-02-07 10:50:06 <quinq> Quality software
2026-02-07 10:54:36 <quinq> a7eee67e21ff1747ef2fd573cb0021cc7fab5a7b added in the code in 2014, never documented ever after
2026-02-07 10:55:19 <quinq> Fault of having a weird embedded --help message that is different from the (usage) synopsis
2026-02-07 10:56:10 <quinq> At least it's in the man-page
2026-02-07 10:56:34 <quinq> Which tells you that -E differs from POSIX ;p
2026-02-07 10:56:49 <quinq> “On filesystem errors (file not found etc), instead of handling the error as regular output as POSIX mandates and keep going, issue an error message and exit.”
2026-02-07 11:31:39 <frag> you have to use extra code to get a cronjob to run every 4th/13th etc week?
2026-02-07 11:33:22 <frag> cant be done with just the 5 specifier?
2026-02-07 11:44:51 <lopid> none of those are week
2026-02-07 11:48:50 <frag> ay
2026-02-07 12:31:26 <pu2> Can anybody tell me why `/usr/sbin/setup-keymap ch ch-fr` results in "ERROR: loadkmap failed to stop" ?
2026-02-07 12:32:26 <lopid> it's the normal message what loadkmap stops/restarts
2026-02-07 12:32:29 <lopid> when
2026-02-07 12:35:44 <lopid> specifically, there is no explicit return value in /etc/init.d/loadkmap's stop function
2026-02-07 12:37:51 <pu2> hmm I don't remember having had that message in previous versions of Alpine
2026-02-07 12:37:51 <pu2> and how is it normal, in the sense that it does not change the keyboard mapping when it's supposed to
2026-02-07 12:38:17 <lopid> it's normal because that's what openrc outputs in that situation
2026-02-07 12:40:28 <pu2> I mean, the point is changing the keyboard layout... it gets changed in "/etc/conf.d/loadkmap" but will go effective only at reboot. Hardly normal to me.
2026-02-07 12:40:55 <lopid> if the keymap isn't changed then that's a different issue than you seeing that message, which is what you asked
2026-02-07 12:41:24 <lopid> i can't talk about how exactly loadkmap itself operates
2026-02-07 12:41:55 <pu2> They seem to be related though, since it's a "new" phenomenon. But ok, let me rephrase... why does setup-keymap not change the keymap as it used to?
2026-02-07 12:42:17 <lopid> that, i don't know
2026-02-07 12:45:44 <edd> hey folks, i set up msmtp to replac sendmail, as per https://wiki.alpinelinux.org/wiki/Relay_email_to_gmail_(msmtp,_mailx,_sendmail
2026-02-07 12:46:11 <edd> confirmed working with `mail <localuser>`
2026-02-07 12:46:28 <edd> but tasks run with cron are not mailing
2026-02-07 12:46:41 <edd> i see the cron tasks running in /var/log/messages
2026-02-07 12:46:49 <edd> but their output is not being mailed out
2026-02-07 12:47:20 <edd> the tasks are in root's crontab and i've checked mailing the root user manually sends an email
2026-02-07 12:47:22 <edd> any ideas?
2026-02-07 12:52:01 <edd> i haven't set MAILTO, but I don't think i should have to
2026-02-07 12:52:51 <greguu> hey, just bulding a rootfs for armv5tel-musl on void linux using xbps-src. works great, but everytime I do this, I also update a few patches (musl, busybox etc) from alpine repo to my void-packages repo and I feel sad that alpine does not support armv5tel natively..mostly because I want to run pmOS on armv5tel
2026-02-07 12:56:49 <greguu> and I do whinge about this every year or so, for the last decade or so, I may keep doing until armv5tel is dropped in mainline
2026-02-07 13:01:16 <edd> it works with MAILTO... maybe busybox cron just defaults to discarding any output
2026-02-07 13:01:45 <edd> on my openbsd boxes cron mails the user whose crontab it is by default
2026-02-07 13:52:19 <Sertonix[m]> greguu: If you don't mind building packages yourself you could try to compiling alpine for armv5tel. Still missing proper documentation but there is https://codeberg.org/Alpine-Ports/
2026-02-07 14:01:23 <jvvv> edd: do you have a 'MAILTO=...' line at the top of your users crontab file?
2026-02-07 14:14:27 <jvvv> edd: oh, sorry, I read further back but not your more recent... I see you already know about that.
2026-02-07 16:41:50 <freerig[m]> Hi, doas anyone here know how to setup Obex over Bluetooth on Alpine?
2026-02-07 17:06:25 <freerig[m]> <freerig[m]> "Hi, doas anyone here know how to..." <- Nevermind, I just had to run `doas /usr/lib/bluetooth/obexd -nsar /path/to/shared/files` and make the interface discoverable.
2026-02-07 17:06:25 <freerig[m]> Maybe I could write this in the wiki...
2026-02-07 17:26:44 <realroot[m]> <freerig[m]> "Hi, doas anyone here know how to..." <- to transder files?
2026-02-07 17:26:50 <realroot[m]> *transfer
2026-02-07 17:59:09 <freerig[m]> Yes
2026-02-07 17:59:09 <freerig[m]> I first used the "bluez-obexd" package (the solution I mentioned earlier), but then I realized that `bluez-tools` provides a simple `bt-obex -s` command to receive files, so in case anyone is interested...
2026-02-07 19:36:05 <edd> jvvv: thanks anyway
2026-02-07 20:14:51 <realroot[m]> i use blueman-manager
2026-02-08 10:46:29 <chatterino[m]> Hello! I'm trying to get my audio setup running under sway, however using pipewire I can't get it running. I already did the troubleshooting from the wiki however I'm still getting the following error:... (full message at <https://matrix.org/oftc/media/v1/media/download/ATeZO3QCjlRAiiALj8Eaz9yiLs_OUzFgnSKrWeYe1tVGJpOHvPF7DC4Cgpb8UpluseLSON7HWhAoaoPw2I6-ov5CechPU57QAG1hdHJpeC5vcmcvZ3JCdXBESE12REN6ZlZWYU5GY2FCRVlq>)
2026-02-08 10:46:56 <chatterino[m]> s/however/but/
2026-02-08 10:53:27 <chatterino[m]> I also can't set it to the gui runlevel
2026-02-08 10:53:27 <chatterino[m]>  ``* rc-update: `gui' is not a valid runlevel``
2026-02-08 11:13:52 <quinq> chatterino[m], I suppose it's invalid because you didn't create it
2026-02-08 14:00:35 <invoked> looking at that wiki page, the walkthrough doesn't have them create it.
2026-02-08 14:01:08 <invoked> https://wiki.alpinelinux.org/wiki/OpenRC#For_Wayland
2026-02-08 14:01:50 <invoked> oh wait, it does. sorry, i just woke up.
2026-02-08 16:28:51 <invoked> on a lark, since i was still using the pipewire libexec script, i walked through these instructions. a problem i ran into on one machine is that i didn't have rtkit installed, and it will hard fail on that.
2026-02-08 16:29:16 <invoked> the page does mention the rtkit dependency, but imo it does so in the wrong spot
2026-02-08 16:30:29 <invoked> it should be mentioned under "User service prerequisites" not under "Realtime scheduling", because pipewire will try to configure realtime by default, and fail if rtkit isn't there.
2026-02-08 16:31:20 <invoked> though this is only for people not using PAM, i guess.
2026-02-08 16:49:14 <invoked> also, the /etc/user/init.d/pipewire* scripts apparently need a change: s/ready=/notify=/
2026-02-08 21:28:04 <chatterino[m]> Thanks for the hints, I finally made it crawling through the several linked wiki pages.  
2026-02-08 21:28:18 <chatterino[m]> * made it after crawling through
2026-02-08 22:22:58 <Guest1961> Hi, I’m trying to get Alpine Linux infra working without kernel modules. From what I can tell, modules/modloop seem pretty baked into the workflow. Are there any guides or docs on running Alpine without them?
2026-02-08 22:23:02 <Guest1961> Right now I’m using aports to build a custom kernel with everything built-in (no modules), then using the aports scripts to generate a netboot image. That path ends up calling mkimg.base.sh, which invokes update-kernel, and that assumes modules/modloop are present.
2026-02-08 22:23:09 <Guest1961> Am I missing a supported approach here, or is this just not a use case Alpine really targets?
2026-02-08 22:31:27 <Sertonix[m]> Compiling a custom kernel seems like the the only way I can think off. Other than using something which doesn't use alpine kernels at all (like docker).
2026-02-08 22:41:40 <Guest1961> I see, okay, was hoping that we could get away from doing that so we can stay within the alpine tooling ecosystem. Makes upgrading kernels annoying haha. Appreciate the note
2026-02-08 23:01:00 <failxontour[m]> may i ask out of curiosity Guest1961 (IRC) why you would need a Alpine Linux without additional kernel modules why not use smth. such as lfs or bedrock?
2026-02-08 23:01:36 <failxontour[m]> if you go as deep as into disabling kernel modules and rebuilding in the first place?
2026-02-08 23:06:38 <failxontour[m]> and if you invoke aports should you iirc not fork that repo in the first place?
2026-02-08 23:14:29 <failxontour[m]> Building a custom kernel also voids the update, upgrade base security of a system such as Alpine in the first place right? What happens if there are 0-days in kernel level. I guess your idea here is the compromise is for speed or instead of running on bleeding edge distros like as with arch or gentoo.
2026-02-08 23:20:04 <failxontour[m]> The whole pro-argument to using a feature such as modules seems to me be laying in the dynamic nature and interoperability of systems, parts in a system. Baking a custom catered kernel might seem like a good idea on a embedded system off the internet, but when there is a active connection this seems to me like a build once and then deal with the consequences solution. While it might compromise on boot times when actively within a
2026-02-08 23:20:04 <failxontour[m]> system this should in my current POV not have much of any effect on performance especially on more common AMD64 systems.
2026-02-08 23:25:17 <failxontour[m]> at least those are the conclusions i draw from what you wrote... sorry for the WoT
2026-02-08 23:37:41 <failxontour[m]> If i talk stupid smack please be so kind and enlighten me
2026-02-08 23:37:47 <failxontour[m]> * enlighten me thanks
2026-02-09 06:01:38 <prabu> invoked, for the pipewire realtime scheduling issue faced by chatterino, i've updated the wiki https://wiki.alpinelinux.org/wiki/PipeWire#PipeWire_user_service. if something is still not accurate, i'll appreciate a note on that...
2026-02-09 09:36:46 <frag> didnt know echo $'this is a test' works.. is this specified anywhere in the repo man-pages-posix?
2026-02-09 09:37:41 <grawity> probably not as it's a non-POSIX extension to the syntax
2026-02-09 09:38:17 <frag> afaiu its posix now with $'\t', im assuming its the same with my previous example
2026-02-09 09:39:16 <grawity> ah, I suppose it is https://pubs.opengroup.org/onlinepubs/9799919799/utilities/V3_chap02.html#tag_19_02_04
2026-02-09 09:39:42 <grawity> but that section is not part of any *command* manual
2026-02-09 09:40:01 <frag> moving target this posix :>
2026-02-09 09:41:32 <ikke> POSIX is descriptive, not prescriptivee
2026-02-09 09:41:49 <grawity> imagine programming to a specification that was stuck in the 1980s
2026-02-09 09:43:38 <frag> rly?!
2026-02-09 09:50:29 <frag> thats what happened to christendom eh, now everything is allowed, smh..
2026-02-09 10:31:16 <quinq> It's been added to SUSv5 yes
2026-02-09 10:32:29 <quinq> So some shell will need a bit of time to adapt
2026-02-09 10:32:34 <quinq> s/shell/&s/
2026-02-09 10:33:33 <quinq> ikke, what do you mean by POSIX not being prescriptive? (I'm just not sure exactly what it means in English)
2026-02-09 10:37:15 <ikke> It does not define anything not already in (common) use
2026-02-09 10:37:41 <quinq> Ah, well mostly yeah
2026-02-09 10:37:50 <quinq> It trys to find common grounds
2026-02-09 10:38:00 <quinq> It does though sometimes “invent” things
2026-02-09 10:38:20 <quinq> Like defining a new interface (not existing) to solve commons problems
2026-02-09 10:39:20 <quinq> Or just problems that nobody tried to solve yet
2026-02-09 13:20:43 <andar1an> Has anyone seen this before: i915 0000:00:02.0: [drm] ERROR GT0: GuC firmware i915/tgl_guc_70.bin: fetch failed -ENOENT? Getting a wedged uhd 770 igpu, have linux-firmware-i915, intel-media-driver, linux-firmware-intel,mesa-va-gallium, mesa-dri-gallium, and mesa-vulkan-intel. Not sure if setting mesa loader env will solve this problem, been tinkering to see if I can get igpu to work as secondary gpu 
2026-02-09 13:21:35 <andar1an> not seeing HuC messages in kernel ring buffer, only GuC
2026-02-09 13:40:13 <andar1an> not sure what package in alpine contains firmware related to GuC. Can grab from kernal git, just wondering if it is supposed to be in a package. With packages installed seem to be missing both 1915_guc(and huc)_load_status files, so pretty sure I am just missing firmware
2026-02-09 13:41:12 <andar1an> thought I may need to update commandline. Didn't try that yet
2026-02-09 13:50:09 <andar1an> sorry for spam, updating commandline seemed to remedy
2026-02-09 14:27:29 <runxiyu> hmmm, im not really familiar with supervise-daemon but this is what i'm seeing on htop: supervise-daemon spawns, and it spawns a child but it becomes a zombie and stalls forever?
2026-02-09 14:28:50 <aron> try to execute the child from the terminal, and see what it's problem
2026-02-09 14:29:06 <aron> (I don't think this is a supervise-daemon issue)
2026-02-09 14:30:37 <runxiyu> aron: works fine on the terminal
2026-02-09 14:30:42 <runxiyu> from a tmux of the same user, that is
2026-02-09 14:31:10 <runxiyu> the line is: supervise-daemon symigd --start --stdout /var/log/symigd.log --stderr /var/log/symigd.log --user symig --group symig --respawn-delay 5 --respawn-period 50 -- /usr/local/bin/symigd -config /etc/lindenii/symig/symigd.conf                 
2026-02-09 14:31:29 <runxiyu> and '/usr/local/bin/symigd -config /etc/lindenii/symig/symigd.conf' in a 'su - symig' works fine
2026-02-09 14:32:36 <dwfreed> check symigd.log?
2026-02-09 14:33:29 <runxiyu> it hasn't been created at all
2026-02-09 14:33:59 <runxiyu> ill replace the symigd binary with a shell script that puts something in /tmp to see if it executes at all, urgh
2026-02-09 14:35:11 <runxiyu> nope. as far as i can see, symigd isn't executed at all (i have a 'env > /tmp/stuff' in it, and the file is never created :/) 
2026-02-09 14:36:06 <aron> is anything in /var/log/messages ?
2026-02-09 14:36:35 <runxiyu> lindenii daemon.info supervise-daemon[2876]: Supervisor command line: supervise-daemon symigd --start --stdout /var/log/symigd.log --stderr /var/log/symigd.log --user symig --group symig --respawn-delay 5 --respawn-period 50 -- /usr/local/bin/symigd -config /etc/lindenii/symig/symigd.conf
2026-02-09 14:36:50 <runxiyu> nothing else about supervise-daemon or symig
2026-02-09 14:36:56 <runxiyu> hm
2026-02-09 14:41:53 <aron> are permissions okay on /usr/local/bin/symigd's every component?
2026-02-09 14:43:28 <aron> who is the owner of those?
2026-02-09 14:44:02 <runxiyu> the binary is owned by root and 755; /etc/lindenii/symig/* is owned by symig:symig
2026-02-09 14:44:32 <runxiyu> hmm
2026-02-09 14:44:40 <aron> and the other path component of /usr/local/bin/symigd ?
2026-02-09 14:45:08 <aron> ( /usr/local /usr/local/bin )
2026-02-09 14:45:39 <runxiyu> oh path components? yeah, those are all owned by root:root; and i could su into symig's user and execute symigd properly
2026-02-09 14:47:53 <invoked> prabu: thanks. i assume you have some way of monitoring irc without being here. :)  one note: rtkit doesn't get pulled in automatically by installing the others, so might want to be explicit about adding it.
2026-02-09 14:48:14 <aron> and what /usr/local/bin/symigd does? it stays in the foreground, or forks itself to bg?
2026-02-09 14:48:58 <runxiyu> stays in the foreground (and prints nothing when executing successfully)
2026-02-09 14:49:21 <runxiyu> previously i replaced it with a script that just puts files in /tmp, and the files never got created
2026-02-09 14:49:37 <runxiyu> ill gdb the supervise-daemon, since i got openrc-dbg now
2026-02-09 14:50:17 <aron> hmm
2026-02-09 14:51:41 <aron> yeah, gdb it, it might reveal its secret!
2026-02-09 14:51:46 <runxiyu> it's stuck on open...??????
2026-02-09 14:51:53 <runxiyu> for hte record i have no network drives or anything
2026-02-09 14:52:06 <aron> open of what?
2026-02-09 14:52:30 <runxiyu> #2  0x00007f7249c013c7 in open (filename=<optimized out>, flags=flags@entry=0) at src/fcntl/open.c:16
2026-02-09 14:52:34 <runxiyu> not sure at the moment
2026-02-09 14:52:41 <runxiyu> ill check source
2026-02-09 14:53:01 <aron> "<optimized out>", not much help :)
2026-02-09 14:53:23 <aron> try figure it out from frame #3
2026-02-09 14:54:10 <runxiyu> yeah
2026-02-09 14:54:12 <runxiyu> trying
2026-02-09 14:54:48 <runxiyu> fifo_fd = open(fifopath, O_RDONLY);
2026-02-09 14:55:46 <runxiyu> uh... supervise-daemon's internal fifos for... something?
2026-02-09 14:55:51 <aron> sounds bad
2026-02-09 14:56:31 <aron> I'd guess it is some ulimit
2026-02-09 14:56:57 <runxiyu> or... some weird fifo-blockin issue, urgh
2026-02-09 14:59:23 <aron> can you strace it? to see what it wants to create/open
2026-02-09 14:59:32 <aron> maybe the fifo create fails
2026-02-09 15:00:02 <runxiyu> ill try when i can (walking around an airport and cant really use laptop rn)
2026-02-09 15:02:57 <runxiyu> i dont have any weird ulimit/cgroups/etc settings... uhh
2026-02-09 15:04:38 <aron> okay, one less thing to worry about, it worth to check it
2026-02-09 15:09:58 <aron> opening a named pipe blocks if no writer end opened
2026-02-09 15:10:21 <aron> so, something is messed up, and strace could reveil it
2026-02-09 15:13:54 <runxiyu> well, supervise-daemon self-forks, so stracing- oh wait
2026-02-09 15:14:33 <runxiyu> im not sure how to strace it as soon as it forks itself
2026-02-09 15:14:50 <aron> strace can follow forks
2026-02-09 15:14:57 <aron> with the -f flag
2026-02-09 15:15:00 <runxiyu> ty
2026-02-09 15:19:06 <runxiyu> well, apparently strace -f makes it...  fail on clock_nanosleep because of log file permission issues
2026-02-09 15:19:13 <runxiyu> heisenbug + fix one small issue
2026-02-09 15:20:13 <runxiyu> well, uhhhhhh
2026-02-09 15:20:21 <runxiyu> i'd really like to get to the root of the issue
2026-02-09 15:21:17 <rodger> and all this while you're puttering around an airport .. that in uniquely impressive
2026-02-09 15:21:26 <rodger> is*
2026-02-09 15:22:19 <runxiyu> im sat down at a bench/chair/thing
2026-02-09 15:24:02 <runxiyu> .... and afterwards, i can just start it with openrc no problem??!!
2026-02-09 15:24:05 <runxiyu> that makes no sense!!
2026-02-09 15:28:07 <runxiyu> ill try to spin up some containers and see if i could reproduce this :/
2026-02-09 15:54:22 <runxiyu> can't reproduce :/
2026-02-09 18:00:18 <andar1an> Over the weekend I was tinkering with some efi stubs, does anyone use stubby boot efi stub? The only one that seemed to work for me was the systemd efi stub. Stubby seems to be a derivative of that, and I didn't know how to figure out what was going wrong 
2026-02-09 18:01:48 <andar1an> Also, does anyone have systemd efi stub working for a password prompt to unlock a luks disk? Only way I could get that to work was cryptkey, but was prbly something i was missing 
2026-02-09 18:04:42 <andar1an> my naive thinking is the password prompt should be possible to input, but I am not sure.
2026-02-09 18:20:52 <fabled> andar1an: I do efistub boot. I ended up forking/fixing gummiboot which is the earlier version of systemd efi stub. As long as you include proper initramfs in the efi image things are good. The LUKS password is promoted by the initramfs - not the efi stub
2026-02-09 18:41:08 <andar1an> cool. I saw in the wiki that gummiboot wasn't working. WHat is your fork valled @fabled? 
2026-02-09 18:41:35 <fabled> Unfortunately it is not publicly available currently 
2026-02-09 18:41:59 <andar1an> ah darn. Will it be in aports if it will be eventually?
2026-02-09 18:43:07 <fabled> Not sure. The systemd one is a superset. With additional features. I did mostly work as we didn't have systemd efistub available at the time, and I needed one custom efi variable exported as an extension.
2026-02-09 18:46:22 <fabled> I wonder if stubbyboot-efistub works
2026-02-09 18:48:30 <andar1an> systemd efistub is working for me, just enjoying the learning right now. stubbyboot-efistub didn't work for me unfortunately. Been wanting to move to a uki and complete secureboot forever. Finally did, only thing I am uncertain about is password prompt for luks unlock. Not sure if I need to include a splash screen or something. The stub loading seems like it will be possible to do input, but when not using crpytkey it just says mounting with no promp
2026-02-09 18:48:36 <andar1an> with cryptkey is succeeds
2026-02-09 18:49:49 <andar1an> will dig in soon hopefully 
2026-02-09 18:51:00 <fabled> Seems stubby does not support the newer style handover protocol 
2026-02-09 18:51:20 <fabled> Not secureboot
2026-02-09 18:51:43 <fabled> Yes, the stub needs additional security stuff if you have secureboot enabled.
2026-02-09 18:52:20 <fabled> I did do that for my fork. But I suspect you'll be better off with the systemd-efistub
2026-02-09 18:54:19 <andar1an> thank you @fabled, this has guided learning a bit. stubby didn't work for me without secureboot as well. May just be my hardware
2026-02-09 18:54:53 <fabled> Do you use alpine kernel or custom built?
2026-02-09 18:56:00 <andar1an> alpine stable kernel
2026-02-09 18:56:35 <andar1an> sorry, that machine is lts
2026-02-09 18:57:53 <fabled> I think both of them still have EFI_HABDOVER_PROTOCOL=y which is the deprecated old style handover used by gummi/stubby
2026-02-09 18:58:25 <fabled> Did you use grub or put your .efi as the default boot image?
2026-02-09 18:58:41 <achill> hmm linux-{lts,stable} dont have EFI_HANDOVER_PROTOCOL set
2026-02-09 18:58:50 <achill> or at least shouldnt afaik
2026-02-09 18:58:59 <fabled> I think it comes from default config
2026-02-09 18:59:06 <andar1an> using a uki, grub doesn't exist on system
2026-02-09 19:00:58 <andar1an> when I was looking at the alpine kernel configs, I wasn't sure if dome kernal defaults would exist beyond those options too. I noticed when building a kernal the default output was longer iirc
2026-02-09 19:02:06 <andar1an> can look at that parameter when I get back to that machine
2026-02-09 19:06:12 <fabled> seems stubby does cmdline handling from the efi program LoadOptions. I remember seeing BIOS implementations that dump garbage there when just installing .efi as the default image (vs. proper efi boot configuration)
2026-02-09 19:06:50 <fabled> I ran earlier into issue that my stub worked with grub, and on *some* of the machiens when being the default boot image EFI/boot/bootx64.efi
2026-02-09 19:07:05 <fabled> so had to add some heuristic to detect if the commandline is valid or not
2026-02-09 19:07:15 <andar1an> i tried setting it at EFI/BOOT/bootx64.efi too, no go
2026-02-09 19:08:10 <fabled> but anyway, multiple issues with stubby efistub. cannot recommend it.
2026-02-09 19:08:16 <andar1an> nice, I was wondering how one may check the commandline. I just grabbed the one from proc 
2026-02-09 19:08:40 <fabled> that is available only after boot up
2026-02-09 19:08:43 <fabled> and efistub constructs it
2026-02-09 19:08:49 <andar1an> machine was previously using grub
2026-02-09 19:09:27 <andar1an> so when converting I wasn able to use, just needed to drom BOOT_IMAGE portion and make sure to build initramfs with correct fstab
2026-02-09 19:09:31 <fabled> typically the efistub uses the one from PE section ".cmdline". but it allows also the BIOS to override it. and some BIOSes have implementation issues that may cause the efistub to not work correctly. I suspect this happens with stubby
2026-02-09 19:11:38 <andar1an> that machine is a gen 4 x1 yoga thinkpad. dk the last time i updated bios/firmware on it haha. One is unable to directly manage secureboot keys on it though, need to use sbctl (thank goodness for that)
2026-02-09 19:12:00 <andar1an> maybe if I update, can try again
2026-02-09 19:12:29 <andar1an> either way, will be good to dig in to understand how process and parameters work
2026-02-09 19:25:21 <andar1an> jummpin off now, thank you for chat and help
2026-02-10 19:33:56 <enigma9o7_irc> I keep checking edge/i386 for firefox but no version is there, it's only there for other arches.  (librewolf 147 and firefox-esr 140 are there tho)
2026-02-10 19:36:46 <ikke> enigma9o7_irc: firefox dropped support for x86
2026-02-10 19:38:03 <enigma9o7_irc> google dropped support in chromium for x86 several years ago, hasn't stopped distros from continuing to build it....
2026-02-10 19:55:51 <invoked> most distros have dropped x86 altogether
2026-02-10 19:57:23 <ikke> enigma9o7_irc: firefox dropping x86 means they are no longer testing it, resulting in build on other failures.
2026-02-10 19:57:38 <ikke> It would require someone dedicated to fix those issues to maintain support
2026-02-10 19:58:11 <frag> dropping x86 :S because everyone uses phones now?
2026-02-10 19:58:31 <ikke> frag: x86_64 is still supported
2026-02-10 19:58:40 <ikke> as is armv7
2026-02-10 19:59:36 <frag> oh, i dont know the distincttion, but sounds less dramatic then
2026-02-10 19:59:54 <ikke> x86 is i686 / 32-bits
2026-02-10 20:00:34 <frag> oh yeah.. old stuff i guess
2026-02-10 20:03:31 <enigma9o7_irc> understood.   i see now it was dropped because it was failing to build
2026-02-10 21:38:55 <f_> It seems like the Matrix bridge is one-way now.
2026-02-10 21:39:02 <f_> or zero-way
2026-02-10 21:39:23 <f_> Anyway RoadRoller01 asked on matrix "why isn't lilo bootloader available in aports?"
2026-02-10 21:39:43 <f_> And Sertonix: none of your messages landed
2026-02-10 21:39:48 <f_> pj: same.
2026-02-10 21:53:17 <invoked> f_: you're the bridge now
2026-02-10 21:54:39 <invoked> dawg
2026-02-10 21:56:05 <f_> xD
2026-02-10 21:56:25 <pj-> f_: well, it was meant for Sertonix so I think it's fine
2026-02-11 06:33:26 <iodomi[m]1> you also have flatpak left which has terrible sandbox and is hardly maintained and chroot
2026-02-11 07:27:38 <idkrnx[m]> <iodomi[m]1> "you also have flatpak left which..." <- Hardly maintained?
2026-02-11 07:31:39 <jpp1> Flatpak is crap
2026-02-11 07:35:03 <iodomi[m]1> for browsers it is on itself its ok i cut down on almost all permissions by default
2026-02-11 07:35:21 <iodomi[m]1> <idkrnx[m]> "Hardly maintained?" <- that's what brave says
2026-02-11 07:42:59 <puretryout[m]1> I have a hard time believing such a shady company. Flatpak is maintained. It's open-source, you can check yourself.
2026-02-11 07:54:41 <iodomi[m]1> flatpak brave was built 7 days ago compared to 5h ago on github releases
2026-02-11 07:54:59 <iodomi[m]1> that is horrible for a chromium browser
2026-02-11 08:02:23 <quinq> Yes, you need to update your web browser every other our, otherwise it's become obsolete and unable to render last amateur $BIGCORP website
2026-02-11 08:02:32 <quinq> s/our/hour/
2026-02-11 08:03:30 <jpp1> chromium browser is crap
2026-02-11 08:04:52 <jpp1> https://www.waterfox.com/
2026-02-11 08:51:46 <franks2> ncopa: I saw your presentation on NHN container days, was it k3s you set up there? or what was it?
2026-02-11 08:52:05 <ikke> k0s probably
2026-02-11 08:52:24 <franks2> ah, yeah! awesome, thx!
2026-02-11 14:54:04 <andar1an> would it make sense to add a kopt_cryptpass options so that one could prompt for a password instead of just a key on init, or is that silly? I feel i have tried this before but it was long ago and I don't remember  
2026-02-11 14:59:57 <andar1an> i guess there is init_args too. Maybe should look into that first
2026-02-11 15:01:20 <RoadRoller01[m]> is that only me or now for some reason abuild -r doesn't work?
2026-02-11 15:04:43 <f_> RoadRoller01[m]: can you be more specific
2026-02-11 15:04:49 <f_> What do you mean by doesn't work
2026-02-11 15:05:06 <f_> jpp1: FYI Waterfox was bought by an advertisement company so it's shady too IMO
2026-02-11 15:06:18 <f_>  08:54 <iodomi[m]1> flatpak brave was built 7 days ago compared to 5h ago on github releases
2026-02-11 15:06:21 <f_>  08:54 <iodomi[m]1> that is horrible for a chromium browser
2026-02-11 15:06:41 <f_> I'm not sure any distro or package repo can upgrade a web browser every 5 hours
2026-02-11 15:06:50 <andar1an> is librewolf still okay?
2026-02-11 15:07:08 <f_> librewolf is nice
2026-02-11 15:07:18 <andar1an> agreed, I wish it was on mobile too
2026-02-11 15:07:28 <jpp1> Ah, okay. f_
2026-02-11 15:07:30 <f_> it is thanks to pmOS :P
2026-02-11 15:07:45 <f_> (mobile-config-firefox works on librewolf too)
2026-02-11 15:07:51 <andar1an> I can't put pmos on my phone. Been dreaming about it
2026-02-11 15:08:13 <f_> yeah I was half-joking :p
2026-02-11 15:08:18 <invoked> well, someone's gotta do the work
2026-02-11 15:08:32 <andar1an> I did see an oss project building a phone with an rpi cm module, imagine that of a rockchip sbc may work 
2026-02-11 15:08:35 <f_> I wonder if anyone has ported librewolf to android perhaps
2026-02-11 15:09:27 <andar1an> mainly what I am wondering. Seeing a port to android would be nice for interim. I still have a lot of lerning before I can learn to port my phone to pmos, and need a backup phone first lol
2026-02-11 15:12:15 <andar1an> been also following servo and blitz, wondering when a browser may come out of those engines. Was curious about a browser where one could pass parameters to select rendering engine in app. Is that silly? Do any browsers exists which allows one to specify rendering engine though some html? 
2026-02-11 15:12:39 <andar1an> sorry, that should be offtopic. My bad
2026-02-11 15:14:38 <invoked> servo is a long way off, probably, before it can be considered ready
2026-02-11 15:14:58 <invoked> mozilla gave up on it for that reason
2026-02-11 15:15:38 <andar1an> ya, but blitz is used for native apps in dioxus world. Just haven't been focussed on client side research for a bit so out of touch with state
2026-02-11 15:17:29 <andar1an> seems like 2026 has been good year for servo so far though based on their blog
2026-02-11 15:36:05 <RoadRoller01[m]> f_: idk it just broke i didn't change anything in `/etc/apk`,... (full message at <https://matrix.org/oftc/media/v1/media/download/AVwM8wg96yHpgCLWWB95tWSpZzGH6CAW2fbN3PptZCYvZQv9kt8uY5tLtzNPXBtiupQF7547tUYk8hJ67rvxu_5CeclXFu9AAG1hdHJpeC5vcmcvbnVBWWZCRGFqTlJOZ2NVT3JaUXlXRHhQ>)
2026-02-11 15:43:05 <Sertonix[m]> Can you try running abuild -rv ?
2026-02-11 15:44:10 <RoadRoller01[m]> do you want me to send all of it?
2026-02-11 15:44:54 <RoadRoller01[m]> in the end :
2026-02-11 15:44:54 <RoadRoller01[m]> ```
2026-02-11 15:44:54 <RoadRoller01[m]> + local '_quiet='
2026-02-11 15:44:54 <RoadRoller01[m]> + undeps
2026-02-11 15:44:54 <RoadRoller01[m]> >>> libimobiledevice: Uninstalling dependencies...
2026-02-11 15:44:55 <RoadRoller01[m]> + abuild-apk --no-interactive del --wait 30 .makedepends-libimobiledevice
2026-02-11 15:44:55 <RoadRoller01[m]> ERROR: No such package: .makedepends-libimobiledevice
2026-02-11 15:44:57 <RoadRoller01[m]> ```
2026-02-11 15:45:23 <Sertonix[m]> To some paste service
2026-02-11 15:45:41 <RoadRoller01[m]> 
ACTION sent a code block: https://matrix.org/oftc/media/v1/media/download/AfUSW2lHYd4WCX-rFm4EPIhJyKCNN-deH0RRvMSeb0HjXAsvSEWEoHakictqAKWyP3gY8J5IR98di2kcJ3A06cZCeclXo6mwAG1hdHJpeC5vcmcvSXVuTkJLdEVpWVpRRWZhR1ZUcEtKcUpY

2026-02-11 15:45:48 <Sertonix[m]> The last few lines don't help
2026-02-11 15:45:48 <RoadRoller01[m]> this is i think the useful part 
2026-02-11 15:46:17 <RoadRoller01[m]> RoadRoller01[m]: idk why it didn't use my other repos 
2026-02-11 15:48:46 <RoadRoller01[m]> here also:
2026-02-11 15:48:46 <RoadRoller01[m]> ```
2026-02-11 15:48:46 <RoadRoller01[m]>  repo_args='--repository '"'"'/home/road/packages//community'"'"' --repository '"'"'/home/road/packages//main'"'"' --repository '"'"'/home/road/packages//community'"'"
2026-02-11 15:48:46 <RoadRoller01[m]> ```
2026-02-11 15:49:27 <RoadRoller01[m]>  cat /etc/apk/repositories
2026-02-11 15:49:27 <RoadRoller01[m]> ```
2026-02-11 15:49:27 <RoadRoller01[m]> ```
2026-02-11 15:49:49 <RoadRoller01[m]> cat /etc/apk/repositories... (full message at <https://matrix.org/oftc/media/v1/media/download/AZpFV9ySgQ4INmxvOuDi8u4QFxXr65qsuuWbtS1njkPpzyyCV8rGwQOAC7_b0KW78cBt8yjdIXkvoOIO9rETkQ5CeclX4COwAG1hdHJpeC5vcmcvSWhDTU5zSEl3bWVlZmhYTEZzdXdkRFJk>)
2026-02-11 15:50:35 <RoadRoller01[m]> * cat /etc/apk/repositories... (full message at <https://matrix.org/oftc/media/v1/media/download/ATgh5gThfZrrXtB9S0EeoichTrqsiKXQvxDmLv9gx-kdF_Ef5Ea2bKIKPHQHcGjnpV6HcbC2SJPE99yNFr7DAhJCeclX63iQAG1hdHJpeC5vcmcvVHNSQlp4Y1hmbkRXYlVORWVQSWVQTWZl>)
2026-02-11 15:51:01 <Sertonix[m]> Since there is no --from in the apk command these will be taken automatically.
2026-02-11 15:51:20 <Sertonix[m]> Could you edit /usr/bin/abuild and comment out the deps "--quiet --simulate" line
2026-02-11 15:51:28 <RoadRoller01[m]> apk update does show that it recognized the others repos:... (full message at <https://matrix.org/oftc/media/v1/media/download/Aboo4Im6OJtEvTqZ-e1MK9IRzbsSUd7Yf0FO9PjXqTnF3vy9qWRavWPgqcVp--wOSFCGwXegUcNM-QzGodIXKn9CeclX-HQQAG1hdHJpeC5vcmcvTkFBZndUV2JsZUVSRFBnR0NIRGdReURv>)
2026-02-11 15:51:44 <RoadRoller01[m]> Sertonix[m]: Ok
2026-02-11 15:53:32 <RoadRoller01[m]>  abuild -r... (full message at <https://matrix.org/oftc/media/v1/media/download/Ac6Oq_2JvZPWwZW8ZOWf16qmOMuxPnRTTpah4YrlAPYWeOTiayA49FNiwrQNGN2GuxLIWveymK2FGzoqSnH-4I5CeclYFrDAAG1hdHJpeC5vcmcvcnN2ak5keWtSRW5Za2FNWmx4b2NDSEhk>)
2026-02-11 15:54:15 <Sertonix[m]> Try running apk fix :)
2026-02-11 15:54:55 <RoadRoller01[m]> its an error with  postmarketos-mkinitfs
2026-02-11 15:55:07 <runxiyu> is gitlab down?
2026-02-11 15:56:19 <RoadRoller01[m]> runxiyu: its for me
2026-02-11 15:56:55 <runxiyu> hmm
2026-02-11 15:57:01 <runxiyu> seems up again
2026-02-11 15:57:19 <RoadRoller01[m]> runxiyu: yes 
2026-02-11 15:58:03 <RoadRoller01[m]> it was in a cat nap 
2026-02-11 15:59:02 <runxiyu> ?
2026-02-11 15:59:48 <RoadRoller01[m]> runxiyu: gitlab 
2026-02-11 16:01:02 <runxiyu> oh lol
2026-02-11 16:03:18 <runxiyu> .u ‐
2026-02-11 16:03:19 <runxiyu> errr
2026-02-11 16:04:01 <RoadRoller01[m]> runxiyu: ?
2026-02-11 16:09:59 <Sertonix[m]> <RoadRoller01[m]> "its an error with  postmarketos..." <- Added to the pile of issues with this line in abuild
2026-02-11 16:32:26 <andar1an> so I think where I got stuck last time in understanding if I can update nlplug-findfs to accept a prompt.
2026-02-11 16:32:42 <andar1an> if one adds a cryptpass options to initramfs-init
2026-02-11 16:33:29 <andar1an> when cryptsetup is calling in nlplugfindfs, could there be an argument to expect user input during this process
2026-02-11 16:40:24 <andar1an> i only see test-passphrase and verify-passphrase though for cryptsetup open, so not sure
2026-02-11 16:42:49 <andar1an> was thinking maybe it would be possible to capute input in initshell and pipe, but now I think that won't work
2026-02-11 16:52:05 <SrainUser> Hello, i have great issue with running flatpak without pulseaudio, the problem is that even when i use apulse, there is no sound for some reason, i dont see any errors, it is just that there is no sound
2026-02-11 16:56:11 <andar1an> you using pipewire and alsa? 
2026-02-11 16:56:56 <andar1an> this may help: https://wiki.alpinelinux.org/wiki/PipeWire
2026-02-11 16:56:57 <SrainUser> well just alsa
2026-02-11 16:58:09 <andar1an> oh cool, it is a user service now. Unreal
2026-02-11 17:01:21 <SrainUser> well it works for most stuff, like firefox, mpv, wine and such, i just dont know how to make it work for flatpak, like the flatpak just doesnt have sound
2026-02-11 17:02:14 <andar1an> with pipewire and also flatpaks have sound for me
2026-02-11 17:04:17 <andar1an> there is a pipewire-pulse package if you are using pipewire, but if not I am not sure
2026-02-11 17:11:50 <p_6f3Ik7Suw> works for me 03> pipewire & pipewire-pulse & wireplumber & dbus-run-session flatpak run --device=all
2026-02-11 17:12:35 <andar1an> have you tried the user service without dbus-run-session? or is it still necessary?
2026-02-11 17:13:18 <SrainUser> well i dont use pipewire so uh
2026-02-11 17:13:52 <andar1an> what are you using? You said you weren't using pulse audio right? 
2026-02-11 17:14:47 <SrainUser> yeah, im using just alsa, as satndalone thing
2026-02-11 17:14:55 <andar1an> could one of those have come in as a dependency on another package 
2026-02-11 17:15:13 <andar1an> if you check your world file do you have either pulse or pipewire?
2026-02-11 17:15:23 <andar1an> oh ok
2026-02-11 17:15:31 <andar1an> i didn't know you could do that
2026-02-11 17:16:33 <andar1an> not sure if some flatpaks require pulse audio
2026-02-11 17:18:15 <invoked> i think flatpak does rely on pipewire-pulse generally
2026-02-11 17:18:29 <invoked> i'm not 100% sure how the portals work
2026-02-11 17:18:39 <andar1an> so you will then need either pipewire or pulse audio for sound there right?
2026-02-11 17:18:48 <runxiyu> RoadRoller01[m]: nvm, im just trying to fix https://gitlab.alpinelinux.org/alpine/aports/-/issues/17936
2026-02-11 17:19:08 <runxiyu> and i have a bot in another channel that checks unicode codepoints given to it, and i pasted into the wrong channel
2026-02-11 17:19:10 <SrainUser> technically apulse should be the replacement for pulseaudio
2026-02-11 17:19:45 <SrainUser> it doesnt seem to work on flatpak
2026-02-11 17:20:17 <SrainUser> it doesnt seem to work on any application requiring pulseaudio
2026-02-11 17:20:17 <andar1an> pipewire has been pretty solid in my experience
2026-02-11 17:20:38 <invoked> i can understand not wanting to add stuff beyond alsa, but fwiw audio glitching went away 100% with pipewire, anecdotally
2026-02-11 17:20:46 <SrainUser> my problem with pipewire is that, im into minimalism ig, and pipewire is basically everything in one
2026-02-11 17:21:14 <SrainUser> and alsa is basically already in kernel
2026-02-11 17:22:29 <SrainUser> however i think i should just go to apulse developers about it becuz there is no way it doesnt work if it works for everyone
2026-02-11 17:22:33 <SrainUser> who used it
2026-02-11 17:23:12 <andar1an> I can understand that. I just dk how to help there, sorry. Just know pipewire works for me with the limited flatpaks i have
2026-02-11 17:25:50 <invoked> you're still adding software with apulse
2026-02-11 17:26:03 <invoked> six/half-dozen
2026-02-11 17:27:30 <SrainUser> what do you mean invoked?
2026-02-11 17:28:10 <invoked> if you're trying to avoid pipewire to get a pulseaudio socket, you're still adding apulse to get the same result
2026-02-11 17:28:39 <invoked> and pipewire isn't heavy.
2026-02-11 17:31:38 <dogg0_> hey guys some advice? if you have a vm, thats a fileserver, is it best practice to have them all inside the 1 image? or have the vm imagte, and the filestore on a nfs or what
2026-02-11 17:33:35 <invoked> this is not advice, but i would never use nfs for that purpose.
2026-02-11 17:34:47 <invoked> there are lots of ways you could solve that problem which are not specific to alpine in any way.
2026-02-11 17:35:27 <dogg0_> i get that.  i just thought id ask you guys
2026-02-11 17:35:58 <dogg0_> if best practice is to have them in the same vm, or seperate them
2026-02-11 17:36:21 <invoked> i'm not sure there is a one-size-fits-all answer for that
2026-02-11 17:41:50 <dogg0_> i said nfs because this is my lan, but its mostly just a 'is it usually best to keep the os and files seperate'
2026-02-11 17:45:22 <invoked> nfs is a pseudo filesystem, so it's not a block device that you can newfs. it's complex and for some edge cases, completely unsuitable
2026-02-11 17:46:40 <invoked> the answer depends on a bunch of factors, but generally having a second device that you can mount from inside the vm would be the way i'd probably go.
2026-02-11 17:47:19 <invoked> because you probably need to think about backups, et al
2026-02-11 17:50:06 <dogg0_> thx invoked
2026-02-11 17:54:38 <invoked> np, good luck
2026-02-11 18:03:51 <andar1an> my xdg runtime dir is set,  but when trying to get the user services added for pipewire, I keep getting an XDG_RUNTIME_DIR unset error. 
2026-02-11 18:04:02 <andar1an> has anyone seen this before?
2026-02-11 18:04:15 <invoked> yeah.
2026-02-11 18:04:43 <andar1an> am i missing something? I have not used user level services much in openrc yet
2026-02-11 18:05:00 <invoked> i use the openrc user service now, but i still set XDG_RUNTIME_DIR in .profile
2026-02-11 18:05:36 <andar1an> it is set in my environment on login, I am not sure why it isn't seeing it
2026-02-11 18:05:42 <andar1an> printenv shows it
2026-02-11 18:06:09 <invoked> andar1an: http://okturing.com/src/17805/body
2026-02-11 18:06:20 <invoked> that's what i have .profile
2026-02-11 18:06:27 <andar1an> danke
2026-02-11 18:06:46 <andar1an> that is the manual way to set, ya. Just shouldn't need to to that
2026-02-11 18:07:05 <invoked> that's what i think display managers do, but i don't use em
2026-02-11 18:07:10 <invoked> s/display/login/
2026-02-11 18:07:14 <andar1an> depends which one
2026-02-11 18:07:28 <andar1an> elogind will do, but for seatd you need to set urself iirc
2026-02-11 18:07:36 <invoked> yeah, i cut out the middlemen
2026-02-11 18:07:51 <andar1an> haha does it conflict for you ever?
2026-02-11 18:08:17 <invoked> i mean, i'm old. i don't care about having my waifu on login screen
2026-02-11 18:08:27 <invoked> just give me a login prompt, i'm good
2026-02-11 18:08:30 <andar1an> i use tuigreet
2026-02-11 18:08:39 <andar1an> quite appreciate it
2026-02-11 18:09:16 <andar1an> and lolz to waifu haha. amazing
2026-02-11 18:10:33 <invoked> i've always launched my display from a shell, but i wouldn't tell people what to do on their systems
2026-02-11 18:11:27 <andar1an> variety is spice of life haha. ricing a machine is best part :P
2026-02-11 18:11:32 <invoked> especially since i'd have some cybersec tryhard flying at me telling me it's less secure
2026-02-11 18:14:02 <andar1an> do you use a dbus user service instead of dbus run session? 
2026-02-11 18:14:12 <invoked> not yet.
2026-02-11 18:14:15 <andar1an> so intrigued by this now
2026-02-11 18:14:39 <andar1an> may give it a shot. Was in progress of migrating this machine to uki. Darn rabbit holes lol
2026-02-11 18:21:05 <invoked> i've planned to check out stuff like tuigreet but, i've got bigger fish to fry
2026-02-11 18:21:31 <andar1an> if you use greetd it is pretty easy few moment thing
2026-02-11 18:22:09 <invoked> the problem is, i have a lot of comfy informational things happening when i get a login shell
2026-02-11 18:22:29 <invoked> which i lose if i switch to a login manager
2026-02-11 18:22:36 <andar1an> if I need to see that I just switch tty
2026-02-11 18:23:31 <invoked> it's stuff i want to see every time
2026-02-11 18:25:22 <andar1an> true. I still have a few dmedg things to figure out, but my boot is generally okay. there is 1 shutdown and 1 startup warning I would eventually like to tackle, but they haven't been problematic yet
2026-02-11 19:01:41 <RoadRoller01[m]> runxiyu: ?
2026-02-11 21:07:16 <andar1an> amazing - i was just able to migrate to openrc user services for dbus and pipewire. This is pretty darn cool. The one thing I am finding challenging though is checking on user services via rc-status -U or rc-service -U servicename status
2026-02-11 21:07:35 <andar1an> is there a special way you are supposed to check user services? 
2026-02-11 21:07:57 <andar1an> nvm. ignore me
2026-02-11 21:08:09 <andar1an> now that they are working properly the status is working properly. 
2026-02-11 21:10:21 <andar1an> will it be reasonable to update the user init.d file to include code to set XDG_RUNTIME_DIR and DBUS_SESSION_BUS_ADDRESS? I needed to do both for my setup, but unsure if it is a common need
2026-02-11 21:13:50 <andar1an> this is pretty awesome-sauce, you can even user supervisor with them. Time to stick ssh-agent there haha
2026-02-11 21:23:00 <RoadRoller01[m]> you could use pam-rundir
2026-02-11 21:26:15 <runxiyu> RoadRoller01[m]: well, nevermind
2026-02-11 21:26:59 <RoadRoller01[m]> runxiyu: ??
2026-02-11 21:28:09 <runxiyu> RoadRoller01[m]: previously i sent '.u -' because i was checking unicode characters for a main/groff issue im assigned to; i did it in here rather than the channel where i have a unicode character detection bot; then you asked me ? and i explained it a little and you sent me ? and uhhhh
2026-02-11 21:28:13 <runxiyu> anyways it doesn't matter lol
2026-02-11 21:28:45 <RoadRoller01[m]> ?
2026-02-11 21:33:01 <WhyNotHugo> andar1an: IIRC openrc has a mechanism for dependant daemons to import variables for their dependencies.
2026-02-11 21:33:10 <WhyNotHugo> It's probably not documented tho.
2026-02-11 21:33:49 <WhyNotHugo> Tip: most applications default to using $XDG_RUNTIME_DIR/bus for dbus if $DBUS_SESSION_BUS_ADDRESS is unset. Just make dbus listen on that socket
2026-02-11 21:43:02 <runxiyu> x86)     export GOARCH="386" ;;
2026-02-11 21:43:04 <runxiyu> https://gitlab.alpinelinux.org/runxiyu/aports/-/jobs/2215974
2026-02-11 21:43:08 <runxiyu> > MC[AL28]:APKBUILD:238:17:literal integers must not be quoted
2026-02-11 21:43:10 <runxiyu> thats a little silly
2026-02-11 21:43:22 <runxiyu> achill: should i remove the quotes? or somehow surpress the lint
2026-02-11 21:43:22 <runxiyu> ty
2026-02-11 21:43:50 <achill> i think both is fine
2026-02-11 21:44:15 <runxiyu> how'd i ignore the lint? (since it makes sense to stay consistent with the rest of the switch/case
2026-02-11 21:44:20 <achill> if if you're at upgrading go, you could check if the tests also fail for you locally and maybe find out why
2026-02-11 21:44:37 <andar1an> thanks @WhyNotHugo! Will check that out soon. Was happy just to get this working for now. I wasn't seeing anything in searches yet for how to do dependencies. Will dive in soon
2026-02-11 21:45:45 <achill> something like # shellcheck disable=2046 a line before
2026-02-11 21:45:48 <andar1an> and ya, I just added some logic similar to what i saw but updated to export XDG_RUNTIME_DIR=/run/user/$(id -u $user) and 		export DBUS_SESSION_BUS_ADDRESS="unix:path=${XDG_RUNTIME_DIR:-"/run/user/$(id -u $user)"}/bus"  because of what you mentioned
2026-02-11 21:46:15 <achill> the number should of course match the lintcheck or something
2026-02-11 21:48:04 <runxiyu> its not shellcheck, it was in apkbuild-lint
2026-02-11 21:54:42 <runxiyu> https://gitlab.alpinelinux.org/runxiyu/aports/-/jobs/2215982 hhmmm, sigill....
2026-02-11 21:57:18 <WhyNotHugo> andar1an: I wouldn't export `XDG_RUNTIME_DIR=/run/user/$(id -u $user)`: whatever is creating the directory should give you the exact value.
2026-02-11 21:57:28 <WhyNotHugo> Then again, if it works it works.
2026-02-11 21:57:57 <andar1an> currently using elogind, and while it was creating the directory, openrc wasn't seeing it
2026-02-11 21:58:12 <andar1an> the services would fail to start until I did that. DK why
2026-02-11 21:58:25 <andar1an> I haven't tried with seatd yet
2026-02-11 21:58:52 <andar1an> but I don't think seatd even exports xdg, so probably even less confilct
2026-02-11 21:59:38 <andar1an> gotta test turnstile again eventually, but right now getting back to moving this to a uki and secureboot haha
2026-02-11 22:00:32 <andar1an> was mainly using elogind and eudev to make Steam life easier, but it is being problematic with non-steam games recently for remote play, so maybe tinkering is soon haha
2026-02-11 22:36:36 <andar1an> kernel mode setting video resolution seemed to mess with dbus user service haha. dk why
2026-02-11 22:36:51 <andar1an> fixed 1 problem, created another haha
2026-02-12 02:20:38 <andar1an> Is anyone else having issues with recent elogind versions? 
2026-02-12 02:21:22 <andar1an> I am finding that is the root of problemds in many scenarios. I can roll back and it remedies issues, and after rolling back I can roll forward and it usually works again
2026-02-12 02:22:16 <andar1an> no errors output, but it likes to fail to start as a service in these scenarios. x86_64
2026-02-12 02:24:04 <andar1an> even with the openrc-user services, it seems that is why XDG RUNTIME DIR wasn't being read
2026-02-12 02:27:09 <andar1an> dk why i asked now, bout to log off. later all
2026-02-12 10:45:14 <staceee> my server fails to apk update. It stucks at [fetch https://.../APKINDEX.tar.gz]
2026-02-12 10:45:36 <staceee> while I can curl the tar.gz
2026-02-12 10:52:26 <staceee> well it eventually fetched the resources o_O
2026-02-12 16:05:14 <runxiyu> intermittent networking issues™
2026-02-12 16:19:16 <msrd0> should community packages with new bugfix releases get updated for edge (or the master branch) only or also on the latest stable alpine branch?
2026-02-12 16:19:48 <ikke> msrd0: If there are no breaking changes, they can be backported to latest-stable
2026-02-12 16:51:36 <lopid> stable kind of implies bug fixes, does it not?
2026-02-12 16:53:43 <frag> talking of bugs, in mpv edge for a long time, at least on my rpi/dwm switching windows/workspaces makes the video unstable, so cant watch, i downgraded to stable again so dont have more info :> just was surprised it was there long ago and revently..
2026-02-12 16:55:42 <frag> *recently
2026-02-12 18:07:55 <O[m]> what up chat
2026-02-12 21:50:41 <Saijin_Naib[m]> Anyone using beets? It is broken for me, and wondering if I should try fixing it up
2026-02-13 00:36:05 <RoadRoller01[m]> hello for some reason my system randomly reboot (as what i have thought):... (full message at <https://matrix.org/oftc/media/v1/media/download/AVfz1o66A2jZlkWT8lo3hqeJHOLQgFz537jlmZuKQX6lKrlmGtkUoHSGDYSwyFoero8h1qgtuq-3Kiz3MqkyQtJCecnIYuswAG1hdHJpeC5vcmcvSXRScmhTbUNLc1pYdW9JYWZIVEFjTGFs>)
2026-02-13 06:26:47 <jpp1> Un tuto en français pour installer alpine linux
2026-02-13 06:27:03 <jpp1> https://peertube.stream/w/oCu6c5KuxDcT2JSRvQRbpZ
2026-02-13 06:48:17 <jpp1> See the section: Changing console font in this tutorial
2026-02-13 06:48:42 <jpp1> https://wiki.alpinelinux.org/wiki/Fonts
2026-02-13 07:14:55 <hazardaj_nombroj> ^ i think these days it's preferred that user-wide fonts are installed into ~/.local/share/fonts, not ~/.fonts, but both will work
2026-02-13 07:23:44 <jpp1> For what?
2026-02-13 07:40:10 <frag> what does 200% VSZ (2+ of these with librewolf and two semi-heavy sites) or even 800% (6 of these with surf with 2 semi-heavy sites) mean in busybox top?
2026-02-13 08:58:22 <hazardaj_nombroj> okay, i'm having a weird problem, installing alpine on a machine, setting up greetd, but it seems like the greetd user can't execute anything, not even simple commands
2026-02-13 09:06:02 <jpp1> greetd user ?
2026-02-13 09:06:58 <hazardaj_nombroj> yeah, greetd runs as its own user
2026-02-13 09:07:24 <jpp1> greetd: is A login manager daemon
2026-02-13 09:11:32 <hazardaj_nombroj> really, it seems like it's impossible to execute commands as any other system user
2026-02-13 09:11:51 <hazardaj_nombroj> like i can `doas -u nobody id` on my laptop, but not on this new system
2026-02-13 09:19:53 <jpp1> https://wiki.alpinelinux.org/wiki/Greetd
2026-02-13 09:21:18 <hazardaj_nombroj> thanks, i'm already aware of that
2026-02-13 09:21:57 <hazardaj_nombroj> but there does seem to be some weirdness, where i can't run *any* commands as greetd or any other system user
2026-02-13 09:22:21 <hazardaj_nombroj> even though my doas config has "permit persist :wheel" with no restrictions on what commands can be run
2026-02-13 09:25:18 <hazardaj_nombroj> found the error
2026-02-13 09:25:28 <hazardaj_nombroj> somehow / was set to 0700 and owned by my user
2026-02-13 11:42:44 <mkjmkj> I have 5 APKBUILD files for various game engines, all compiles from sources. Do you think I should submit MR's for them? I have a feeling the gaming community around alpine is not big :) ...
2026-02-13 11:51:55 <Sertonix[m]> As long as it's not huge nobody will stop you
2026-02-13 11:52:27 <mkjmkj> Then I'll give it a try 
2026-02-13 11:54:13 <aron> mkjmkj: what kind of game engines?
2026-02-13 11:57:44 <mkjmkj> aron: So far I have; devilutionx, uzdoom, eduke32, voidwd, fury, ioquake3, nblood, serioussamclassicvk, yquake2, vkquake
2026-02-13 11:58:26 <aron> I always wanted to package ioquake3, now I don't have to :)
2026-02-13 11:59:35 <mkjmkj> voidwd -> voidsw
2026-02-13 12:02:10 <mkjmkj> ioquake3 is a little tricky, but doable. They don't seem to version/tag their source code either, so one have to take a stab at it :-)
2026-02-13 12:02:58 <aron> meh, that's so unprofessional
2026-02-13 12:03:57 <mkjmkj> I know, but hey; latest master "works" ;-D
2026-02-13 12:04:08 <aron> lol
2026-02-13 12:20:49 <vv221> mkjmkj, I’m the main developer of ./play.it (https://git.dotslashplay.it/play.it/about/) and am currently moving half my systems to Alpine. So I for sure care about such game engines ;)
2026-02-13 12:27:20 <mkjmkj> vv221: So does the rest of us with enough nostalgia and game engine curiosity :-D
2026-02-13 12:29:06 <vv221> I packaged some game engines for Debian lately, including julius (Caesar 3) and jazz2 resurrection (Jazz Jackrabbit 2). Hopefully I’m going to learn how to do that for Alpine soon.
2026-02-13 12:29:14 <mkjmkj> vv221: BTW, the URL link seems to only resolve to IPv6. I'm in Denmark (a third-world-IPv6 country :-( ) so I cannot access it.
2026-02-13 12:29:58 <vv221> Ah, sorry for that, I temporarily shut down IPv4 due to attacks by LLM scrapping bots.
2026-02-13 12:31:39 <vv221> To make it short: it’s a native packages generator for several Linux distros, focusing on commercial DRM-free games. So you give it some game installer as a setup*.exe or a something*.sh, and it turns that into a .deb (Debian), a .pkg.tar.lxz (Arch Linux) or whatever is Gentoo new binary packages format.
2026-02-13 12:32:12 <vv221> With Alpine support being the next one we’re focusing on.
2026-02-13 12:40:54 <mkjmkj> vv211, aron: MR for ioquake3 has been made: 97462
2026-02-13 12:42:36 <aron> mkjmkj: thanks, I'll check it soon
2026-02-13 12:55:38 <Kammus> Good morning!  Anyone have a solution for uart yelling "mdev: unknown user/group 'root:uucp' on line 53
2026-02-13 12:55:38 <Kammus> " about 800 times during boot? 
2026-02-13 13:14:42 <mkjmkj> does /etc/group exists and contain group 'uucp' ?
2026-02-13 13:50:00 <Kammus> yes
2026-02-13 13:56:58 <ikke> mkjmkj: fyi, it's not necessary to open a ticket for every package MR you make
2026-02-13 13:59:40 <mkjmkj> ikke: I don't wish to. I just thought I have to create a ticket when introducing a new package.
2026-02-13 14:00:06 <mkjmkj> I'm obviously not familiar enough with Alpine policies :-)
2026-02-13 14:03:23 <mkjmkj> Kammus: then I can't think of much from what you have told so far.
2026-02-13 14:53:24 <deadcade[m]> Is there a way for me to download the source of one specific package?
2026-02-13 14:53:25 <deadcade[m]> I can `git clone` all of `aports`, but if I only need `community/somepackage`, that seems like a waste. Does `apk` have a package source download maybe?
2026-02-13 15:01:19 <ikke> deadcade[m]: does git clone --depth=1 suite you?
2026-02-13 15:01:36 <ikke> gitlab has source archives
2026-02-13 15:01:46 <vv221> Unless I missed something, aports only includes build rules, not the actual upstream sources. So cloning the whole of it (especially with --depth=1 as suggested) should be quite light.
2026-02-13 15:02:04 <runxiyu> deadcade[m]: https://git-scm.com/docs/partial-clone
2026-02-13 15:02:42 <runxiyu> also that gives you the apkbuild, not the "package source" which you'd need abuild feth for
2026-02-13 15:02:44 <runxiyu> fetch*
2026-02-13 15:03:32 <deadcade[m]> runxiyu: In this case I am interested in the APKBUILD
2026-02-13 15:03:50 <deadcade[m]> ikke: Looking into how light this is now
2026-02-13 15:03:53 <runxiyu> deadcade[m]: try a partial clone then
2026-02-13 15:04:31 <ikke> --depth=1 seems simpler
2026-02-13 15:04:39 <ikke> and is fast enough
2026-02-13 15:06:28 <deadcade[m]> --depth=1 seems fine?
2026-02-13 15:07:19 <vv221> I fetches only the current commit, not the whole history. Since that repo only includes a couple text files for each supported package, it should be painless.
2026-02-13 15:08:31 <deadcade[m]> To try to avoid X/Y problem let me real quick explain what I'm trying to do:
2026-02-13 15:08:31 <deadcade[m]> I'm converting my hacky sh scripts to build a custom Alpine repository (both fully custom packages and patches) to run in Forgejo actions as a "fun" excercise (and to not have to have a whole separate virtual machine specifically set up for building Alpine packages)
2026-02-13 15:09:47 <deadcade[m]> Previously I git pull'ed aports, which is light enough, and gives a summary of which packages updated. I used that to check whether I need to rebuild the patched packages. In Forgejo actions that seems like an even worse way of going about it, so I want to attempt checking whether a package needs to be updated per package (and thus only downloading the APKBUILDs that I actually need)
2026-02-13 15:10:55 <deadcade[m]> This hacky script runs every night, I don't want to put more stress on Alpine gitlab than needed.
2026-02-13 15:16:50 <deadcade[m]> Just got it to use sparse-checkout and that's exactly what I need. Thank you!
2026-02-13 15:19:29 <Biswa96[m]> the github mirror repository can be cloned instead if you want not to stress alpine gitlab https://github.com/alpinelinux/aports
2026-02-13 15:20:04 <deadcade[m]> Yeah that's fair. Idrc about hitting MS infrastructure too hard, but they might.
2026-02-13 15:21:49 <vv221> MS is not going to notice ;)
2026-02-13 15:22:14 <deadcade[m]> Fair
2026-02-13 15:22:27 <vv221> A daily fetch is ~nothing compared to what they serve.
2026-02-13 20:08:36 <Kammus> ok, sorry, had to do some work there. "mdev: unknown user/group 'root:uucp' on line 53 - I have commented line 53 out, I have deleted, added, changed permissions. I guess I don't know what mdev is doing at that point in the boot process? 
2026-02-13 20:16:31 <invoked> serial belongs to uucp iirc
2026-02-13 21:28:22 <WhyNotHugo> Kammus: you need to regenerate the initfs for the new mdev.conf to be copied into it.
2026-02-13 21:42:41 <Kammus> That's what I was thinking, it must have a copy inside.  OK! every time I do this it breaks. I'm going to try to change root:udev to root:root and see if that makes it happy. 
2026-02-13 22:18:38 <Kammus> Thanks! I knew there was something dumb I was forgetting
2026-02-14 13:59:27 <mrsvcd> Hi, if some one is interested I hacked a bit on Open SIMH to get it to compile on Alpine: https://github.com/MrSVCD/mrsvcd-simh 
2026-02-14 14:02:04 <mrsvcd> This is my first project like this so please be kind in your critisism.
2026-02-14 14:03:25 <invoked> mrsvcd: that's cool! i'm on tuhs and i've got a light interest in that stuff, but (i think) there are others too who would appreciate it
2026-02-14 14:04:19 <invoked> imo you should consider submitting it
2026-02-14 14:05:10 <invoked> (that's also where you will get your best feedback, on gitlab)
2026-02-14 14:07:26 <longnoserob[m]> nice
2026-02-14 14:11:20 <mrsvcd> invoked: yes, that could be fun. 
2026-02-14 15:22:52 <mrsvcd> invoked: how do i submit?
2026-02-14 15:25:03 <invoked> mrsvcd: this will probably cover everything https://wiki.alpinelinux.org/wiki/Creating_an_Alpine_package
2026-02-14 15:25:58 <invoked> then, for new aports, just be patient... someone will get to it eventually
2026-02-14 15:26:57 <vv221> I assume GitLab is the only way to submit support for new packages? (I love git, but grew into hating GitLab over time)
2026-02-14 15:28:27 <vv221> (I know about the cgit instance, but the documentation only talks about GitLab merge requests)
2026-02-14 15:28:32 <invoked> yes. you could try gitlab from the command line though, if you don't like the web interface (i feel you).  see glab
2026-02-14 15:28:43 <vv221> Nah, thank you ;)
2026-02-14 15:29:16 <vv221> I know about it, but my dislike is much deeper than that (I self-hosted GitLab for years, and even was part of the team maintaining GitLab in Debian for a couple years).
2026-02-14 15:29:32 <invoked> barring any of that, maybe you could bribe a developer with bags of cash
2026-02-14 15:29:38 <invoked> 
ACTION joking

2026-02-14 15:30:11 <vv221> I’m very short on cash sadly, been unemployed for two years and a half so I could focus on free/libre software instead.
2026-02-14 15:30:38 <vv221> And I would feel dirty by paying someone to use GitLab in my stead ;)
2026-02-14 15:31:03 <vv221> But it’s OK, I will find other ways to contribute.
2026-02-14 15:31:05 <invoked> yeah, gitlab.  pros and cons
2026-02-14 15:31:09 <ikke> Officially there still is a mailing list, but it would require someone else then to make an MR out of it
2026-02-14 15:32:11 <vv221> Merge requests are used only for the initial submission, or they are used for packaging updates too, after a packaged reached the repository a first time?
2026-02-14 15:32:22 <vv221> *after a package reached
2026-02-14 15:33:47 <invoked> mrsvcd: forgot to mention, if nobody takes up your MR after a month or something, it'd a good idea to ping here about it, in case it got buried
2026-02-14 15:39:26 <invoked> alpine releases roughly every 6 months, so from my pov there's at least a couple months every year where new aports drop in priority
2026-02-14 15:41:23 <vv221> The stable releases are why I’m considering Alpine as my replacement for Debian stable, thanks for having these instead of providing only a rolling-release.
2026-02-14 15:42:00 <invoked> i'm not sure it's equivalent to debian stable though
2026-02-14 15:42:10 <invoked> not much is :)
2026-02-14 15:42:24 <invoked> debian does a lot of backporting
2026-02-14 15:43:38 <mrsvcd> Debian is a trustworthy, slow, lumbering elefant and that is why we like it.
2026-02-14 15:43:48 <vv221> Backporting of patches? We (I’m a Debian Developer) only do that for critical fixes, so I would not call that "a lot".
2026-02-14 15:44:37 <vv221> Of course over the huge amount of packages, it probably actually is a lot ;)
2026-02-14 15:45:53 <vv221> May main use case for a stable system with fixed releases is to have something where I can automate security updates (`apk update && apk upgrade` or something like that in a crontab) and forget about. Then only handle a big upgrade once or twice every year or so.
2026-02-14 15:46:06 <vv221> *My main use case
2026-02-14 15:46:41 <vv221> I would not automate updates like this on a rolling-release, for obvious reasons.
2026-02-14 15:52:06 <invoked> i mean, debian does more backporting than i think anyone else does. (could be wrong though)
2026-02-14 15:53:28 <mrsvcd> I think that Ubuntu does their fair share of it as well since they have 10-15 years of support contracts now.
2026-02-14 15:58:09 <mrsvcd> There is already a simh port in testing for a old version of simh (simh classic), what should I call my version?
2026-02-14 15:59:04 <mrsvcd> mrsvcd-simh seems to pretentios. (I am a Swede so it comes with that problem)
2026-02-14 15:59:48 <invoked> just simh
2026-02-14 16:00:16 <invoked> maybe diff out that existing simh port against yours to see what's different
2026-02-14 16:00:35 <invoked> might be helpful to make yours better
2026-02-14 16:01:28 <invoked> then do your MR against the existing simh, and maybe change the maintainer line if simh hasn't been updated in a long time
2026-02-14 16:01:37 <invoked> take it over
2026-02-14 16:10:25 <ikke> vv221: merge requests are used for any update into aports
2026-02-14 16:12:02 <mrsvcd> K, I feel I should also make a simh "classic" version as well but we will see how hard it is.
2026-02-14 16:30:36 <vv221> ikke, that makes sense, thanks for the confirmation.
2026-02-14 17:35:59 <jpp1> https://www.lenovo.com/ca/fr/glossary/alpine/
2026-02-14 17:43:47 <quinq> wut
2026-02-14 17:45:12 <abby> it's slop
2026-02-14 17:46:09 <SyntheticBird[m]> it's not
2026-02-14 17:46:28 <SyntheticBird[m]> lenovo has added alpine linux to their glossary
2026-02-14 17:46:43 <SyntheticBird[m]> worthless info but ok
2026-02-14 17:47:13 <jpp1> What is porridge abby?
2026-02-14 17:48:27 <abby> SyntheticBird[m]: it is https://www.reddit.com/r/Lenovo/comments/1bxurhx/lenovocom_has_a_glossary_page_full_of_ai/
2026-02-14 17:48:56 <quinq> Reads like slop
2026-02-14 17:49:15 <quinq> (in translation from English)
2026-02-14 17:49:35 <SyntheticBird[m]> my bad
2026-02-14 17:49:39 <SyntheticBird[m]> fuck slop
2026-02-14 17:50:29 <abby> jpp1: contenu généré par l'IA
2026-02-14 17:50:29 <quinq> jaja abby 
2026-02-14 17:50:51 <jpp1> ah ok
2026-02-14 17:53:07 <quinq> https://www.lenovo.com/ca/fr/glossary/what-are-word-games/
2026-02-14 17:53:26 <vv221> Wow, thanks for the warning.
2026-02-14 17:54:06 <vv221> Avoiding slop exposure, especially by accident, is becoming a full-time job…
2026-02-14 17:56:45 <quinq> That's the main economy of LLMs
2026-02-14 17:57:06 <quinq> Get an LLM to do a human's job
2026-02-14 17:57:10 <quinq> Fire those humans
2026-02-14 17:57:20 <quinq> Hire humans to fix the job done by LLM
2026-02-14 17:57:41 <lopid> that's the hope i hold on to
2026-02-14 17:58:21 <andar1an> I do find the capital expenditure funny. If people stick to local first AI, there isn't a real mechanism to recover the infrastructure spend lol.
2026-02-14 17:59:02 <andar1an> wonder what happens then
2026-02-14 18:00:19 <andar1an> corallory is the capital expense jacks up consumer cost, which makes it harder to afford
2026-02-14 18:00:59 <jpp1> thanks abby 
2026-02-14 18:06:58 <jpp1> It's a shame that OVH doesn't offer Alpine Linux for its VPS.
2026-02-14 18:08:26 <lopid> true of other providers, too
2026-02-14 18:11:00 <andar1an> can you use custom? 
2026-02-14 18:15:09 <lopid> i had to request alpine specifically. they did add an image of, but i doubt they keep it up to date
2026-02-14 18:15:16 <lopid> it
2026-02-14 18:17:05 <andar1an> The alpine iso available from alpine release is a challenge with their servers? 
2026-02-14 18:17:42 <lopid> they give you it already set up
2026-02-14 18:18:05 <andar1an> ah, sorry. I misunderstood
2026-02-14 18:18:51 <lopid> select distro, ssh in to your new distro host
2026-02-14 18:19:50 <lopid> i did offer to install it myself if they'd just give me a blank vm with the iso connected as a cd drive
2026-02-14 18:20:40 <lopid> that's how i installed my previous vm (slackware), but they no longer offer that ability
2026-02-14 18:21:40 <andar1an> locked in to the provider?
2026-02-14 18:22:10 <lopid> i think they just want to provide the network already set up
2026-02-14 18:22:37 <lopid> rather, a host with the network (maybe other things) already set up
2026-02-14 18:24:46 <lopid> bit i hear that cloud providers generally use cloud-init to set up networking etc on 1st boot
2026-02-14 18:25:49 <andar1an> alpine has cloud images
2026-02-14 18:26:15 <andar1an> I haven't personally used, but I can't imagine they are not awesome haha
2026-02-14 18:27:21 <andar1an> they can bootstrap with tiny cloud or cloud init
2026-02-14 18:37:26 <andar1an> has anyone updated mount-ro to use pivot_root? Wondering if it is silly to pivot back to initramfs for shutdown and reboot
2026-02-14 18:48:04 <bb> ---
2026-02-14 18:51:54 <bb> I have a question about window managers and desktops. I have an used laptop with 8 Gb of Ram and I am not sure if I should install alpine linux with Xfce or install something very lean... like dwm or fluxbox...?
2026-02-14 18:53:23 <andar1an> you could always install both and launch to see what you prefer
2026-02-14 18:56:48 <bb> Yeah, but I would like to keep it very clean :)  - do you think there is a big difference in performance?
2026-02-14 18:57:19 <andar1an> you can remove what you don't like
2026-02-14 18:57:45 <andar1an> I dk, I haven't used either. Sorry :(
2026-02-14 18:59:24 <bb> thank you any way - I guess I will start with the good old Xfce because it has everything out of the box :)
2026-02-14 18:59:36 <andar1an> there is also plasma
2026-02-14 18:59:55 <andar1an> when I used to use plasma, it was good on ram, but that was some time ago
2026-02-14 19:00:20 <andar1an> I think sway, plasma, and xfce are in setup-desktop scripts, but I haven't checked in a while
2026-02-14 19:01:11 <bb> I am using Xfce right now on debian - I am used to it and it works.Plasma and KDE are too fancy for me :)
2026-02-14 19:01:42 <andar1an> I loved it haha. Dolphin is probably on of my favorite file managers. Think the UX was so well done
2026-02-14 19:02:01 <andar1an> been trying to keep an eye on plasma bigscreen
2026-02-14 19:02:17 <andar1an> but I like focussing on smithay derivatives because im a rustacean haha
2026-02-14 19:03:17 <andar1an> both cosmic DE and Niri are available in alpine
2026-02-14 19:04:00 <bb> I am a simple user - you sound like a real nerd :)
2026-02-14 19:04:40 <bb> I mean a software dev or so :)
2026-02-14 19:06:40 <andar1an> nice to have variety of choice haha!
2026-02-14 19:07:07 <andar1an> gotta jump off, happy weekend
2026-02-14 19:08:03 <bb> happy weekend!
2026-02-14 21:00:03 <maria[m]> After the latest update on Edge, Chromium 145.0.7632.45 doesn't launch at all. I get:
2026-02-14 21:00:03 <maria[m]> third_party/crashpad/crashpad/snapshot/linux/process_reader_linux.cc:95 "sched_getscheduler: Function not implemented (38)"
2026-02-14 21:00:03 <maria[m]> dbus/bus.c "Could not parse server address"
2026-02-14 21:00:03 <maria[m]> Does not launch with --no-sandbox either. Dbus is running. I'm using x11 (dwm) and Pipewire. Hope Chromium gets fixed before the stable release
2026-02-14 21:00:03 <maria[m]> "Illegal instruction"
2026-02-14 21:00:18 <maria[m]> Does not launch with --no-sandbox either. Dbus is running. I'm using x11 (dwm) and Pipewire. Hope Chromium gets fixed before the stable release
2026-02-14 21:00:18 <maria[m]> dbus/bus.c "Could not parse server address"
2026-02-14 21:00:18 <maria[m]> "Illegal instruction"
2026-02-14 21:00:19 <maria[m]> * After the latest update on Edge, Chromium 145.0.7632.45 doesn't launch at all. I get:
2026-02-14 21:00:19 <maria[m]> third_party/crashpad/crashpad/snapshot/linux/process_reader_linux.cc:95 "sched_getscheduler: Function not implemented (38)"
2026-02-14 21:03:47 <jpp1> Is this a troll? No, because Chromium…
2026-02-14 21:04:13 <maria[m]> No, i actually don't know what to do. Did i make some stupid mistake?
2026-02-14 21:05:51 <realroot[m]> did some update break it?
2026-02-14 21:05:51 <jpp1> By installing Chromium safely
2026-02-14 21:06:02 <quinq> maria[m], it's a known issue
2026-02-14 21:06:18 <quinq> 15:40:47  lnl$ #17954
2026-02-14 21:06:18 <quinq> 15:40:47  algitbot$ community/chromium: segfault on launch after upgrade: https://gitlab.alpinelinux.org/alpine/aports/issues/17954
2026-02-14 21:06:21 <quinq> 15:41:10  quinq$ Thank you lnl :)
2026-02-14 21:06:39 <maria[m]> Ok, sorry
2026-02-14 21:06:44 <quinq> No problem!
2026-02-14 21:06:47 <maria[m]> Thanks for responding
2026-02-14 21:06:54 <realroot[m]> you can try to install previosu version
2026-02-14 21:18:09 <jpp1> maria[m]: If I may, I would advise you not to use Chromium, and to use Waterfox instead.
2026-02-14 21:18:59 <quinq> Alpine doesn't package waterfox
2026-02-14 21:19:48 <vv221> I see both "regular" and ESR Firefox are packages though, that’s nice.
2026-02-14 21:19:54 <vv221> *are packaged
2026-02-14 21:20:37 <quinq> and librewolf
2026-02-14 21:21:15 <jpp1> quinq: Damage
2026-02-14 21:22:11 <maria[m]> jpp1:
2026-02-14 21:22:11 <maria[m]> I am also a web developer, and I need to make sure that my website looks the same, or at least acceptable, on Blink-based browsers.
2026-02-14 21:22:11 <maria[m]> I already use Librewolf for most things. Unfortunately, Firefox forks have performance issues for me. My hardware is quite old, so I can't even browse a certain marketplace on Firefox, but I can on Chromium. Figma doesn't work for me on Firefox, either.
2026-02-14 21:22:43 <quinq> Which 10% of web developer care about
2026-02-14 21:22:48 <quinq> (I mean that as a compliment)
2026-02-14 21:23:56 <quinq> maria[m], as an alternative you could try some qt web browser, they use blink too
2026-02-14 21:25:12 <vv221> As a Web developer I only cared about Firefox… with the reasoning that bugs on Blink would get reported anyway ;)
2026-02-14 21:25:24 <vv221> (that was more than a couple years ago, Firefox was still relevant at the time)
2026-02-14 21:26:36 <quinq> # apk add falkon
2026-02-14 21:26:37 <quinq>     required by: falkon-25.12.1-r0[so:libpyside6.abi3.so.6.9]
2026-02-14 21:26:37 <quinq>   so:libpyside6.abi3.so.6.9 (no such package):
2026-02-14 21:26:37 <quinq> ERROR: unable to select packages:
2026-02-14 21:26:37 <quinq>   so:libshiboken6.abi3.so.6.9 (no such package):
2026-02-14 21:26:39 <quinq>     required by: falkon-25.12.1-r0[so:libshiboken6.abi3.so.6.9]
2026-02-14 21:26:42 <quinq> Looks like falkon is broken
2026-02-14 21:28:00 <courge_musque[m]> Yeah, same error, on edge
2026-02-14 21:28:06 <vv221> TIL falkon is still a thing. It used to be my daily driver (it was still called qupzilla), before it sneakily switched to Blink and I ended up with Google Chrome engine on my system unknowingly.
2026-02-14 21:28:50 <quinq> Yeah
2026-02-14 21:29:21 <quinq> Well, they switched from qtwebkit to qtwebengine, it's more a qt switch than a falkon switch I suppose
2026-02-14 21:29:25 <vv221> I still have some muscle memory today from a bug qupzilla had, maybe 10 years ago.
2026-02-14 21:29:39 <vv221> Sure, but falkon is the one that I used, so it took the blame instead of Qt.
2026-02-14 21:30:04 <quinq> :)
2026-02-14 21:30:07 <vv221> Then I investigated further, and wrote an apt_preferences (still on Debian) to ensure no variant of Chrome can get installed.
2026-02-14 21:30:43 <vv221> I’m very happy by the way to see I will be able to replicate that trivially thanks to /etc/apk/world
2026-02-14 21:30:59 <mrsvcd> maria[m]: What hardware are you running? cpu & ram specificaly.
2026-02-14 21:31:25 <courge_musque[m]> One day with servo, no chrome will be hidden in software
2026-02-14 21:31:35 <courge_musque[m]> Hopefully
2026-02-14 21:31:56 <vv221> I would rather focus on options that I will see by myself, not only my grandchildren ;P
2026-02-14 21:32:02 <quinq> jaja qutebrowser segfaulted too
2026-02-14 21:32:25 <quinq> 2026 and web browser are still as crappy as they were 30 years ago
2026-02-14 21:32:34 <quinq> Just eating way more resources in their evolution
2026-02-14 21:33:07 <courge_musque[m]> There is still a gap between something like firefox and chrome
2026-02-14 21:33:07 <quinq> vv221 ^^
2026-02-14 21:33:08 <vv221> My real Web browser is uBlock Origin + uMatrix (no longer maintained, I know), Firefox is only the engine powering them.
2026-02-14 21:33:18 <vv221> With *very* harsh default blocking rule.
2026-02-14 21:33:28 <courge_musque[m]> I don't how people can put up with with ads, redirection and stuff like that
2026-02-14 21:33:35 <vv221> The kind that does not even allow first-party CSS or images by default.
2026-02-14 21:33:49 <quinq> Now they embed popups in the DOM
2026-02-14 21:33:55 <quinq> Like this ugly google thing
2026-02-14 21:33:57 <courge_musque[m]> vv221: webasm for ublock is insane, it gets so fast at blocking stuff
2026-02-14 21:34:04 <quinq> “DO YOU WANT TO IDENTIFY WITH GOOGLE”
2026-02-14 21:34:08 <quinq> no, go away
2026-02-14 21:34:42 <courge_musque[m]> But chrome won't allow extentions to use webasm, when they allow them to access the NPU, a real piece of hardware, smh
2026-02-14 21:34:45 <vv221> Well, when the Web is only pure HTML, everything seems blazingly fast… or a blank page.
2026-02-14 21:34:46 <frag> what if you get free cookies?
2026-02-14 21:35:21 <vv221> frag, my setup lets cookies in… but not out.
2026-02-14 21:35:26 <vv221> (a feature of uMatrix)
2026-02-14 21:35:46 <quinq> My body does too
2026-02-14 21:35:48 <vv221> So I get free cookies, only I do not share them back. Not even with the website that gave them in the first place.
2026-02-14 21:36:07 <courge_musque[m]> vv221: That's why my website is pure html with tables for formating, even if the css crowd will say "uh no, you shouldn't do it like this"
2026-02-14 21:36:28 <maria[m]> mrsvcd: 16 GB DDR4, Core i5-8350U, Intel UHD Graphics 620
2026-02-14 21:36:41 <vv221> Bah, HTML is overkill already, I use XHTML ;P — http://notes.vv221.fr/blackhole.xhtml
2026-02-14 21:36:53 <quinq> Imagine having memory problems with 16GB
2026-02-14 21:37:10 <vv221> Wow.
2026-02-14 21:37:21 <maria[m]> well, the people in LKML said that all modern desktops have 32 gb :-)
2026-02-14 21:37:29 <vv221> They’re silly.
2026-02-14 21:37:35 <vv221> Or clueless.
2026-02-14 21:37:37 <vv221> Maybe both.
2026-02-14 21:37:49 <courge_musque[m]> Yeah, 32GB of DDR3 on a thinkpad
2026-02-14 21:38:20 <quinq> vv221, self-centered
2026-02-14 21:38:22 <courge_musque[m]> Now 32GB of DDR5 can be $600 in DDR5
2026-02-14 21:38:40 <vv221> I *do* have 32GB on my desktop… but only since last November or December. And I am someone who run literal *thousands* (around 4,000) video games on that computer.
2026-02-14 21:38:58 <vv221> (DDR3 here, most of my setup is components from 2012~2015)
2026-02-14 21:39:18 <quinq> Have 6GB here, that's already huge
2026-02-14 21:39:46 <vv221> Right, my secondary *gaming* computer has 8GB. My travel laptop probably something like 4GB (I would have to double-check).
2026-02-14 21:40:03 <mrsvcd> I have 32GB i5-6500T but I probably only need 16GB since I have never seen it go over half usage exept whent i do something stupid with images.
2026-02-14 21:40:43 <vv221> A Web browser should *never*, no exception, require 16GB of memory. No matter what is done in it. Even running a full VM through some silly Web technology. Even mining some scam-coin.
2026-02-14 21:41:23 <vv221> Here my 32GB are for a somewhat unusal use case: big tmpfs temporary filesystems.
2026-02-14 21:41:49 <Kammus> I think I have 32GB, but I can't remember why. Definitely not needed, but its fine. 
2026-02-14 21:41:52 <vv221> ("tmpfs temporary filesystems" feels like useless repetition…)
2026-02-14 21:41:58 <jpp1> I have 8 GB of RAM and that's more than enough for what I do.
2026-02-14 21:42:46 <maria[m]> Well, i lied -- that was the kernel bugzilla, not lkml: https://bugzilla.kernel.org/show_bug.cgi?id=196729#c74
2026-02-14 21:42:47 <mrsvcd> I do like having my 16GB ramdrive for temporary storage but i don't need it.
2026-02-14 21:44:27 <vv221> Ugh, Anubis…
2026-02-14 21:44:32 <mrsvcd> Dev machines probably have/had 32GB of ram or more but end users don't.
2026-02-14 21:45:10 <vv221> Even as a dev, I still had 18GB until very late 2025.
2026-02-14 21:45:37 <maria[m]> I could probably do fine with less RAM, but who am I to say no to free RAM :-)
2026-02-14 21:45:45 <vv221> (I had to bump from 8GB because of self-hosting GitLab on a computer that is my daily driver in addition to being a server and a gaming machine)
2026-02-14 21:46:09 <vv221> Free RAM is the best RAM, closely followed by cheap RAM ;)
2026-02-14 21:46:56 <vv221> The 8GB → 18GB here was free RAM, and added a lot of breathing room to investigate memory leaks in a specific game (War for the Overworld).
2026-02-14 21:47:37 <Kammus> I think that's why I ended up with 32Gb, someone said it woudl game better... I noticed no difference but that was years ago.
2026-02-14 21:49:46 <vv221> Here it only helped with investigating bugs in game, not with actually playing them.
2026-02-14 21:50:04 <vv221> When it comes to actual play, I never saw anything even using the 18GB I had prior.
2026-02-14 21:51:00 <vv221> Hey, I do not even find games saturating the 8GB of my laptop, outside of a couple ones with abnormal memory consumption (meaning prior versions of the same game or engine would stay below 4~6GB, so something is most probably wrong with the current ones).
2026-02-14 21:55:42 <Case_Of> oh vv221 you’re using alpine? did not know this
2026-02-14 21:58:00 <vv221> Still my first steps: I replaced my least critical Debian stable LXC container with Alpine.
2026-02-14 21:58:22 <Case_Of> nice
2026-02-14 21:58:27 <vv221> Now the other Debian stable containers are going to follow, especially the one hosting everything related to the dotslashplay.it domain.
2026-02-14 21:59:23 <vv221> After that it’s going to be the laptop, going from Debian stable to Gentoo. And finally the desktop/server/host/everything going for Debian unstable to Gentoo. In the end I should have only Gentoo (actual hardware/host) and Alpine (stable LXC containers).
2026-02-14 21:59:50 <vv221> What I have seen so far is a pleasure to work with!
2026-02-14 22:02:27 <Case_Of> vv221: is it a strong laptop? some package need a lot of ram to build
2026-02-14 22:05:03 <maria[m]> By the way, is there a reason why hplip is edge-only? That package is why I switched to edge in the first place.
2026-02-14 22:05:03 <maria[m]> I'm sure that "why is X edge-only" is a common question; some common reasons could probably be added to the FAQ or the Repositories page as well. Thanks!
2026-02-14 22:08:41 <maria[m]> I mean, sure, Alpine isn't meant to be used as a desktop system, but maybe my server needs to print as well :-p
2026-02-14 22:09:29 <jpp1> It was vv221 who introduced me to Alpine Linux, so I installed it in a VM :)
2026-02-14 22:09:31 <vv221> Case_Of, Gentoo now has binary packages.
2026-02-14 22:09:48 <vv221> And by that I mean a public repository of binary packages.
2026-02-14 22:10:29 <vv221> maria[m], you know you can run stable while still cherry-picking packages from edge? I do that here for lgogdownloader.
2026-02-14 22:12:54 <Case_Of> vv221: true
2026-02-14 22:13:22 <Case_Of> vv221: but if it is to go to binary packages, why not directly alpine?
2026-02-14 22:13:33 <Case_Of> vv221: that’s what I did on my laptop
2026-02-14 22:16:50 <vv221> I’m going to run Gentoo on my main system, so I would rather have the same one on my laptop.
2026-02-14 22:17:23 <vv221> Plus, running native Linux games (built on top of glibc, of course) on Alpine is probably not going to be fun.
2026-02-14 22:19:24 <jpp1> What window manager do you recommend for Alpine?
2026-02-14 22:19:27 <vv221> I’m still going to build some stuff locally anyway, as there is common stuff like pipewire/pulseaudio that I do not use.
2026-02-14 22:19:38 <vv221> I don’t use a window manager in my containers ;(
2026-02-14 22:19:40 <vv221> *;)
2026-02-14 22:19:48 <courge_musque[m]> I have no issue running my games with steam or directly with wine on alpine
2026-02-14 22:20:20 <vv221> I don’t use Steam (never did, never will), and am not going to run Windows builds when I have access to native ones.
2026-02-14 22:20:37 <courge_musque[m]> Well, it will be even easier then
2026-02-14 22:21:22 <vv221> How is that easier? Native Linux games all expect the GNU libc.so.6, I don’t expect them to play nicely on top of musl.
2026-02-14 22:21:42 <vv221> Unless I’m missing some obvious compatibility trick?
2026-02-14 22:21:49 <courge_musque[m]> The glib compat package helps
2026-02-14 22:21:57 <courge_musque[m]> I don't remember the name
2026-02-14 22:22:05 <vv221> Oh, I am indeed missing a compat trick then, thanks!
2026-02-14 22:22:30 <quinq> WINE usually works fine though
2026-02-14 22:23:08 <vv221> Sure, but once again I’m not using WINE for native games. Like I would not run win32 Firefox through WINE, despite that technically being an option ;)
2026-02-14 22:23:09 <courge_musque[m]> It's gcompat on alpine
2026-02-14 22:23:16 <jpp1> I didn't quite understand what a container was.
2026-02-14 22:23:17 <courge_musque[m]> That packages helped me a lot
2026-02-14 22:24:08 <vv221> Nice. I’m not sure I’m going to use gcompat myself, but it’s something I’m going to remember for when I’m giving advice.
2026-02-14 22:24:22 <courge_musque[m]> https://wiki.alpinelinux.org/wiki/Software_management#Running_glibc_programs
2026-02-14 22:24:36 <courge_musque[m]> It's being used for firefox to see drm stuff
2026-02-14 22:24:55 <courge_musque[m]> Not using drm myself, so no clue of it still works for doing that
2026-02-14 22:25:33 <vv221> I don’t use that either, actually I’ve been actively working *against* DRM for more than a decade.
2026-02-14 22:25:48 <vv221> That’s the reason I don’t use Steam, despite working a lot with commercial video games ;)
2026-02-14 22:26:57 <courge_musque[m]> For me, if you can download the game from steam, then start it outside of steam, it's fine
2026-02-14 22:27:10 <courge_musque[m]> But that's something like 40% of games on the platform
2026-02-14 22:27:52 <vv221> I strongly reject mandatory software even for delivery, not only for actually starting games.
2026-02-14 22:29:16 <vv221> Not that it’s a big limitation: as I said earlier I still own more than 4,000 *bought* commercial games, all DRM-free.
2026-02-14 22:29:35 <vv221> (it’s not like I have even the slightest change to play through all that during my lifetime)
2026-02-14 22:38:28 <mrsvcd> Is anyone running loongarch64 here? I just bought the hardware but it is lunar new year so it will tame some time to get here.
2026-02-15 00:16:56 <Saijin_Naib[m]> @vv221 it is not. Even with gcompat, I have not had a single game run in Lutris or Minigalaxy
2026-02-15 00:17:10 <Saijin_Naib[m]> We should do a community funding drive to expand gcompat compat
2026-02-15 00:17:50 <courge_musque[m]> I must have the weirdest alpine ever, with the lightdm killing itself and the games working fine
2026-02-15 00:18:09 <Saijin_Naib[m]> You certainly do 🤣
2026-02-15 00:26:31 <Habbie> courge_musque[m], sounds better than the other way around ;)
2026-02-15 00:35:11 <vv221> Saijin_Naib[m], right, feedback I got elsewhere is that it is far from being ready to replace a glibc in a container/chroot.
2026-02-15 00:35:31 <jpp1> which desktop courge_musque[m]?
2026-02-15 08:39:22 <Case_Of> does anyone tried to install alpine on an odroid xu4?
2026-02-15 08:46:18 <longnoserob[m]> Case_Of:  indirect, as I installed postmarketOS 
2026-02-15 09:06:17 <Case_Of> is it complicated?
2026-02-15 09:06:28 <Case_Of> and does it work fine?
2026-02-15 09:20:46 <longnoserob[m]> works ok, you can installed from bre-built image if you want to try
2026-02-15 09:51:40 <courge_musque[m]> @jpp1 xfce, lightdm just doesn't want to work for it, but I think that I'm close to find the issue
2026-02-15 10:13:04 <jpp1> 01:17:49 courge_musque[m] | I must have the weirdest alpine ever, with the lightdm killing itself and the games working fine
2026-02-15 10:13:37 <jpp1> that's to say?
2026-02-15 10:17:03 <jpp1> Because refusing to work is vague, right?
2026-02-15 10:24:17 <jpp1> What exactly is happening with lightdm?
2026-02-15 10:26:39 <courge_musque[m]> https://caca.libre.tw/screenshots/IMG_20260211_220625.avif
2026-02-15 10:26:40 <courge_musque[m]> When I boot, it's showing this
2026-02-15 10:27:05 <courge_musque[m]> If I wait a bit, I can change the tty and return to tty1
2026-02-15 10:27:44 <courge_musque[m]> But lightdm doesn't start, I can only login in tty
2026-02-15 10:28:00 <courge_musque[m]> I can start XFCE with startx, but not much more
2026-02-15 10:28:02 <courge_musque[m]> https://caca.libre.tw/logs/
2026-02-15 10:28:06 <courge_musque[m]> Here are the logs
2026-02-15 10:32:58 <lopid> what are permissions on /var/lib/lightdm/.Xauthority and the dir?
2026-02-15 10:37:22 <courge_musque[m]> When I saw the logs, I tried to put it at 777, but it didn't change anything
2026-02-15 10:53:53 <ikke> courge_musque[m]: what about namei -om /var/lib/lightdm/.Xauthority
2026-02-15 10:53:59 <ikke> (you may need to install namei)
2026-02-15 11:14:45 <courge_musque[m]> I did a new install, and now it works
2026-02-15 11:14:47 <courge_musque[m]> WHY
2026-02-15 11:15:01 <courge_musque[m]> I didn't change anything, still on v3.23
2026-02-15 14:37:04 <mrsvcd> How do I install a testing package? aka cherry picking.
2026-02-15 14:38:07 <achill> you enable the testing repository, see https://docs.alpinelinux.org/user-handbook/0.1a/Working/apk.html#_repositories_releases_and_mirrors
2026-02-15 14:39:15 <vv221> mrsvcd, "cherry picking" is called "tagging" here.
2026-02-15 14:40:30 <vv221> cf. https://wiki.alpinelinux.org/wiki/Repositories#Tagged_repository
2026-02-15 14:41:26 <vv221> "Using testing repository" below on the same page is a specific example of a tagged repository enabling access to testing packages.
2026-02-15 14:50:41 <mrsvcd> Thank you
2026-02-15 18:03:02 <mrsvcd> I apriciate how fast the irc log updates, nicley done.
2026-02-15 18:06:26 <mrsvcd> Okej, fast might be the wrong word but I appriciate that they exist.
2026-02-15 20:54:06 <dsc_> hello alpine pplz
2026-02-15 20:54:24 <courge_musque[m]> hi
2026-02-15 20:54:26 <dsc_> question: there is abuild, but there is also apk-tools v3 'mkpkg' 
2026-02-15 20:54:35 <dsc_> they both do roughly the same, yeah?
2026-02-15 20:55:12 <dsc_> i.e: you dont need `abuild` to create some custom .apk
2026-02-15 20:55:28 <dsc_> instead you could use `apk mkpkg`
2026-02-15 20:58:47 <ikke> dsc_: abuild does more than what makepkg provides
2026-02-15 20:58:59 <ikke> mkpkg*
2026-02-15 20:59:27 <ikke> But if all you care about is creating an apk file from files you have locally, it may suffice
2026-02-15 20:59:59 <dsc_> ikke: thanks :)
2026-02-16 08:48:35 <alpineuser> hello
2026-02-16 08:49:34 <alpineuser> testing 123
2026-02-16 08:50:50 <alpineuser> ok
2026-02-16 09:03:41 <frag> test successful, welcome to alpine irc!!!!!!!!!!!!!
2026-02-16 09:03:51 <quinq> test
2026-02-16 09:05:02 <frag> test failed, reconfigure your irc client and try again
2026-02-16 09:05:20 <quinq> ;_;
2026-02-16 09:05:27 <quinq> testing 123
2026-02-16 09:06:00 <jpp1> :)
2026-02-16 09:06:01 <frag> close enough, welcom aboard!
2026-02-16 10:03:28 <quinq> yeaaah \o/
2026-02-16 10:46:32 <jpp1> :)
2026-02-16 18:55:44 <mrsvcd> how do I test my apkbuild script?
2026-02-16 18:58:56 <quinq> abuild rootbld
2026-02-16 19:53:27 <MRX1337> hello !
2026-02-16 23:13:01 <Saijin_Naib[m]> @mrsvcd apkbuild-lint and abuild -r?
2026-02-16 23:48:42 <shahid> this package now has "python3~3.12 unknown dependency" https://pkgs.alpinelinux.org/package/v3.23/main/x86_64/at-spi2-core
2026-02-16 23:49:03 <shahid> looks to be the reason https://gitlab.gnome.org/GNOME/at-spi2-core/-/commit/7c8a9365d5009b43afe6d16bebefc14a8df410a7
2026-02-17 00:59:27 <swee> how do I check what packages use up the most space?
2026-02-17 01:04:36 <vv221> You might be able to write something relying on `du -chs $files` + `apk query --fields contents $package`. But maybe there is already some more appropriate tool doing that already.
2026-02-17 01:15:54 <vv221> swee, this is a bit of a massive one-liner, I strongly suggest turning it into a real script, but I think it should give you the information you’re seeking:
2026-02-17 01:16:24 <vv221> apk list --installed --manifest | cut -d' ' -f1 | xargs -n1 -P$(nproc) sh -c 'pkg=$1; files=$(apk query --fields contents $pkg | tail +2 | sed "s#^\s*#/#" | xargs -n1 -P$(nproc) realpath | sort -u); size=$(du -chs $files | tail -n1 | cut -f1); printf "%s - %s\\n" "$size" "$pkg"' -- 2>/dev/null | sort -rh
2026-02-17 01:17:18 <vv221> Here it gives me this biggest offender: 25M - python3
2026-02-17 01:20:29 <vv221> (it’s failing on empty packages, but it should not be too hard to skip them when $files is empty)
2026-02-17 01:40:18 <swee> vv221: https://bpa.st/GGBQ ran this on another computer, idk if that's right
2026-02-17 02:00:42 <quinq> Gladly those packages are not 6.5GB-big :D
2026-02-17 02:19:33 <quinq> swee, vv221, apk list -q -I | xargs apk info -s zfs | sed 'N;s/\([^ ]*\).*\n\(.*\)/\2 - \1/;n;d' | sort -rh
2026-02-17 02:54:15 <quinq> apk list -q -I | xargs apk info --format json -s | jq -r '.[] | [."installed-size", ."name"] | @tsv' | sort -nr
2026-02-17 02:56:25 <efahl> Try: apk --installed query --format json --fields name,installed-size '*'
2026-02-17 02:59:06 <quinq> even better
2026-02-17 02:59:19 <efahl> apk query is my favorite command
2026-02-17 02:59:43 <efahl> use '--fields all' to see what all the field names are
2026-02-17 03:00:18 <quinq> Yeah, I didn't know about it -_-
2026-02-17 03:00:46 <efahl> It's new in 3.0, lots of work was done on it in the last 6 months.
2026-02-17 03:01:31 <quinq> I did try the --fields from apk-info, but that failed hard
2026-02-17 03:04:45 <efahl> I've pretty much stopped using list and info, only use query now.
2026-02-17 03:05:48 <quinq> :)
2026-02-17 03:12:10 <efahl> Show me all the packages that install a file with 'dnsmasq' in the name:  apk --installed query --fields name --match contents '*dnsmasq*'
2026-02-17 03:12:14 <efahl> Very versatile
2026-02-17 07:15:14 <prabu> thanks efahl,vv221,quinq,swee. added to wiki https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#List_packages_by_installed_size
2026-02-17 07:21:36 <quinq> prabu, that looks too complex
2026-02-17 07:24:30 <quinq> oops, gone
2026-02-17 07:25:45 <aron> yeah, not a typical one-liner
2026-02-17 11:27:41 <vv221> efahl, well done, I looked for that installed-size field but somehow missed it.
2026-02-17 15:48:56 <vv221> Got the following error trying to build some software, does it look like a glibc / musl compatibility problem?
2026-02-17 15:48:58 <vv221> /usr/lib/gcc/x86_64-alpine-linux-musl/15.2.0/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find Scrt1.o: No such file or directory
2026-02-17 15:49:12 <vv221> The software I’m trying to build is: https://dacav.org/projects/supercazzola/
2026-02-17 15:54:56 <aron> vv221: looks like invalid linker flags for me
2026-02-17 15:55:57 <vv221> It does complain about some flag a bit below that error:
2026-02-17 15:56:02 <vv221> /usr/lib/gcc/x86_64-alpine-linux-musl/15.2.0/../../../../x86_64-alpine-linux-musl/bin/ld: cannot find -lssp_nonshared: No such file or directory
2026-02-17 15:56:13 <aron> but this is just a guess, without the full build log I can't say anything else
2026-02-17 15:56:39 <aron> better not cherrypick some lines
2026-02-17 15:57:40 <aron> "nonshared" indicates it wants to produce a static linked executable
2026-02-17 15:57:46 <aron> this can cause trouble
2026-02-17 15:58:00 <vv221> aron, here is the full log for context: https://paste.debian.net/plainh/2352ec7b
2026-02-17 16:00:53 <aron> I think this still hides something important
2026-02-17 16:02:21 <aron> hmm, do you have musl-dev installed?
2026-02-17 16:03:28 <vv221> Not at all, that would be a problem indeed ;)
2026-02-17 16:03:58 <aron> maybe better to install build-base
2026-02-17 16:04:09 <vv221> That was it, thanks!
2026-02-17 16:04:27 <vv221> And thanks again for build-base, I should have guessed such a meta-package is available.
2026-02-17 16:05:52 <aron> I'm glad it works now, yw
2026-02-17 16:09:46 <vv221> With that I might be able to get rid of iocaine, as this tarpit looks like a much less bloated alternative.
2026-02-17 16:11:06 <aron> less bloat is always good!
2026-02-17 16:27:40 <dsc_> so, there is debian, but there is also ubuntu
2026-02-17 16:28:01 <dsc_> ubuntu builds/extends what debian provides
2026-02-17 16:28:24 <dsc_> is there a similar thing for Alpine? are there distros who took Alpine, and made it into 'something more' ?
2026-02-17 16:28:44 <ikke> dsc_: does postmarketos count?
2026-02-17 16:28:56 <dsc_> ah yes, of course. forgot about that for a second
2026-02-17 16:28:59 <courge_musque[m]> Postmarkos
2026-02-17 16:29:00 <courge_musque[m]> Same idea lol
2026-02-17 16:29:05 <courge_musque[m]> Maybe adelie too?
2026-02-17 16:29:06 <dsc_> any others?
2026-02-17 16:29:16 <ikke> Adelie does not build on top of Alpine
2026-02-17 16:29:46 <invoked> there seems to be someone who comes in here every month or two saying they're going to make an alpine derivative but then we never hear from them again
2026-02-17 16:29:47 <courge_musque[m]> True, but I don't have any other idea
2026-02-17 16:31:08 <dsc_> invoked: yeah I am doing this, so was just wondering if there are others
2026-02-17 16:31:29 <invoked> i'll try not to scare you then
2026-02-17 16:31:41 <dsc_> if anything, I wil scare you ;)
2026-02-17 16:31:54 <dsc_> for example, I ripped RSA out of apk and replaced it with minisign
2026-02-17 16:31:59 <dsc_> apk-tools*
2026-02-17 16:32:14 <dsc_> and did some more things
2026-02-17 16:32:16 <dsc_> but its still wip
2026-02-17 16:32:28 <invoked> possibly, but i'm old. i once shoehorned bsd userland around linux
2026-02-17 16:33:07 <dsc_> invoked: chimera is interesting yeah :P
2026-02-17 16:34:25 <invoked> i would just say, be careful what you wish for. most people who have started up new distributions couldn't or wouldn't say with them, for reasons unknown
2026-02-17 16:34:35 <invoked> but i know. it's a lot of work
2026-02-17 16:35:01 <invoked> if you're dogfooding 100% it is probably the only way.
2026-02-17 16:35:01 <dsc_> its a bit offtopic but I have strategies to deal with the eventual maintainance burden
2026-02-17 16:37:24 <invoked> i don't do this often but, suggestion: get at least one person who shares your vision & passion for whatever this is
2026-02-17 16:41:02 <vv221> aron, thanks to your help my new bots trap is ready: http://blackhole.vv222.fr/
2026-02-17 16:41:23 <vv221> I still need to teach it how to correctly display non-ASCII characters, but it’s not high priority.
2026-02-17 16:41:26 <mason> vv221: nice
2026-02-17 16:41:44 <vv221> It’s based on the stuff I already documented here: http://notes.vv221.fr/blackhole.xhtml
2026-02-17 16:41:53 <vv221> But it no longer relies on iocaine/rust/cargo.
2026-02-17 16:43:11 <courge_musque[m]> Will you update your tutorial? I'm tempted to do the same on one of my servers
2026-02-17 16:43:26 <aron> vv221: that's cool
2026-02-17 16:43:37 <mason> vv221: http://blackhole.vv222.fr/ isn't answering just now
2026-02-17 16:43:43 <mason> at least from here
2026-02-17 16:43:53 <vv221> It should, but very slowly.
2026-02-17 16:44:02 <vv221> Unless you’re still on the old DNS entry.
2026-02-17 16:44:08 <mason> That could be. Let's see.
2026-02-17 16:44:18 <invoked> not here either.
2026-02-17 16:44:33 <mason> Ah, I'm *only* seeing IPv6 come back and I don't have a bridge for that.
2026-02-17 16:44:35 <vv221> ::42 is the old one (Debian + iocaine, no longer running), ::142 is the new one (Alpine + whatever that new tool is named).
2026-02-17 16:44:57 <vv221> Oh, right, I should have said it’s currently IPv6 only.
2026-02-17 16:45:01 <mason> I see ::142 but I can't talk to it given lack of IPv6.
2026-02-17 16:45:07 <mason> kk, known then!
2026-02-17 16:45:15 <vv221> I shut down IPv4 for now, as bots were hammering it too much.
2026-02-17 16:46:53 <vv221> mason, what you are missing is a bit of nonsensical text based on French translation of Pratchett novels, displaying very slowly and including a bunch of links with other similare pages. So a crawler following links can keep following these links endlessly, always finding new stuff.
2026-02-17 16:47:36 <vv221> I’m serving *millions* of requests from that trap for the Claude bot only, every day. Well, I did with IPv4, it might be less with IPv6 only.
2026-02-17 16:47:52 <vv221> Then you add all the other pillaging bots from the generative AI industry…
2026-02-17 16:48:07 <mason> vv221: nice nice
2026-02-17 17:06:57 <vv221> Well, new goals now: learn some C basics, add UTF-8 support to supercazzola, package it for Alpine.
2026-02-17 17:07:01 <vv221> ETA: nope.
2026-02-17 17:12:03 <mason> vv221: good goals
2026-02-17 17:21:14 <jpp1> What is supercazzola?
2026-02-17 17:22:42 <vv221> A spam generator: https://dacav.org/projects/supercazzola/
2026-02-17 17:25:21 <jpp1> ok thanks
2026-02-17 19:10:47 <jpp1> Do you know it? --> https://www.virtua.cloud/fr/cloud-servers/systems/alpine-linux
2026-02-17 20:22:13 <mrsvcd> Hi, I just tried to send a merge request for my simh APKBUILD. 
2026-02-17 20:22:45 <lpz> jpp1, Nothing much to know, when server's certificate is not trusted ...
2026-02-17 20:24:24 <jpp1> ?
2026-02-17 20:27:07 <Habbie> jpp1, virtua.cloud has a broken https setup that only works for some visitors
2026-02-17 20:28:30 <jpp1> Ah ok
2026-02-17 20:36:54 <mrsvcd> What kind of machine does the risc-v builds?
2026-02-17 20:37:12 <ikke> milk-v pioneer
2026-02-17 20:45:51 <mrsvcd> good night people
2026-02-18 01:38:51 <madtypewriter[m]> Hello! I have an Alpine 3.20 box I'm trying to upgrade. Editing /etc/apk/repositories to 3.21 or 3.22 then running apk add --upgrade apk-tools gives "ERROR: unable to select packages:", "Huh? Error reporter did not find the broken constraints.". Trying 3.23 or edge gives an error about a conflict between qt6-qtwayland-6.6.3-r0 and qt6-qtwayland-6.10.1-r0, except that nothing I can see (per world and apk dot --installed
2026-02-18 01:38:51 <madtypewriter[m]> qt6-qtwayland |grep qt6-qtwayland) depends on any qt6-qtwayland and apk del qt6-qtwayland throws the exact same error. I would be grateful for any suggestions anyone might be able to give as to how I can resolve this.
2026-02-18 01:43:42 <madtypewriter[m]> Oh, and I tried throwing on --force-broken-world (I have filesystem snapshots, so I don't mind trying dangerous things) and all that seems to do is sit indefinitely spinning a core at 100%.
2026-02-18 03:34:34 <prabu> madtypewriter, sometimes orphaned packages can result in errors not easy to track. identify and remove orphaned packages and see if it helps... https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#Listing_installed_packages . i'll appreciate if you can share the error in tpaste.us, so this can be documented in the wiki page..
2026-02-18 04:21:09 <madtypewriter[m]> Well, that's good cleanup to do regardless:) Handful of orphans, all gone
2026-02-18 04:30:24 <madtypewriter[m]> Okay, sorta. I deleted orphans on 3.20, but then if I bump my repos to 3.21+, apk list --orphaned yields a whole list but apk del fails out because of the usual error
2026-02-18 04:31:12 <madtypewriter[m]> And I can't kill the orphans on 3.20 because apparently in 3.20 they were dependencies of some packages I really do wanna keep. I mean, I guess I can uninstall+upgrade+reinstall if it really comes to it, but...
2026-02-18 04:37:41 <madtypewriter[m]> I've created https://tpaste.us/Xvxl which I believe showcases the problem fairly well. Working system on 3.20, breaks on 3.21, breaks a little more verbosely on 3.23, and the breakage is from a package that apk says isn't needed:)
2026-02-18 04:42:39 <madtypewriter[m]> Oh, hang on, does apk info --rdepends qt6-qtwayland only list direct... that's odd.
2026-02-18 04:43:17 <madtypewriter[m]> Okay, that clearly does not mean what I thought it meant. Trying to remove qt6-qtwayland on 3.20 complains that a lot of things depend on it.
2026-02-18 04:43:50 <madtypewriter[m]> I have some questions about why qt6-qtwayland is in the dependency graph for, say, qemu-ui-gtk, but...
2026-02-18 04:57:34 <madtypewriter[m]> Okay, unless I'm misunderstanding what apk dot does, https://tpaste.us/LyPV  is another good angle: My system has no legitimate reason to want qt6-qtwayland-6.6.3-r0
2026-02-18 05:06:28 <prabu> you may want to review /etc/apk/world to see if there are packages pinned there to specific version..take a backup of that file and try to change it to see what might be causing the issue.
2026-02-18 05:11:11 <prabu> after changing version numbers in /etc/apk/repositories, you may want to issue, apk upgrade --available  --available switch is used to force all packages to be upgraded..
2026-02-18 05:11:32 <madtypewriter[m]> Don't think so; `grep -F -e '>' -e '<' -e '=' /etc/apk/world` gives 0 results. Looking at my world file did remind me that I have a handful of testing packages, but I don't think that's the problem
2026-02-18 05:12:28 <madtypewriter[m]> And I thought the point of running apk add --upgrade apk-tools was that it didn't worry about all that. I would run apk upgrade --available next, but I don't get that far
2026-02-18 05:16:29 <madtypewriter[m]> Actually, since I'm playing on a clone anyways, let me just drop all the testing packages from my world and see if it helps...
2026-02-18 05:16:41 <prabu> for apk, /etc/apk/world is the source of truth..so try to simplify it and see if it helps. once upgrade is successfully completed, you can add back from your backup..
2026-02-18 05:17:22 <madtypewriter[m]> Oh, now there's an idea
2026-02-18 05:19:03 <madtypewriter[m]> Alright, if I replace my world file with the smallest thing I could think of - the world file from the official alpine:3.23 docker image - then it works:)
2026-02-18 05:19:08 <madtypewriter[m]> Now I just have to binary-search it!
2026-02-18 05:23:49 <madtypewriter[m]> it was obs-studio
2026-02-18 05:24:11 <madtypewriter[m]> If I just remove that from my world file, apk stops erroring out
2026-02-18 05:25:17 <madtypewriter[m]> Thanks, prabu !
2026-02-18 05:25:25 <prabu> good
2026-02-18 05:26:36 <prabu> i'll appreciate if you can add that error message to wiki troubleshooting section of apk page.. if not i'll add later.. thanks for sharing the pastebin earlier..
2026-02-18 05:28:52 <prabu> if you're uncomfortable with editing wiki, just share a small note here on what you did to resolve the issue. i'll format it for wiki later..
2026-02-18 05:29:33 <madtypewriter[m]> I'm not sure what to put in the wiki? If I can dig up my wiki login I can add a section on circular version conflicts or whatever, but the fix is... take stuff out of world until it works? What I did doesn't feel very general
2026-02-18 05:32:03 <prabu> that above message will do.thanks. 
2026-02-18 05:32:40 <madtypewriter[m]> Well, maybe it generalizes; 1. cp /etc/apk/world /etc/apk/world.real 2. replace /etc/apk/world with a fresh copy from a docker image or whatever 3. Try apk add --upgrade apk-tools -i 4. Use nvim -d /etc/apk/world /etc/apk/world.real or such to add things until it breaks again, then slowly reverse until you find the problem
2026-02-18 05:32:54 <madtypewriter[m]> That's... roughly... what I did
2026-02-18 05:34:13 <madtypewriter[m]> (I actually kept a /etc/apk/world /etc/apk/world.working in the middle because I didn't want to remember what I added/removed every time, and this whole thing happened on a freshly cloned ZFS filesystem because I've had system updates go sideways one too many times to run it on my real root, but...)
2026-02-18 05:35:47 <madtypewriter[m]> Huh. Actually, maybe that should get upstreamed to the wiki instead of living in my personal notes. We've got a ZFS page, but it's pretty sparse.
2026-02-18 05:36:36 <madtypewriter[m]> Oh, but it does point at the ZoL docs which suggest zfsbootmenu, which is arguably better
2026-02-18 05:59:46 <prabu> an entry added here..https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#Huh?_Error_reporter_did_not_find_the_broken_constraints
2026-02-18 06:09:15 <godSend23> hey guys
2026-02-18 06:09:23 <godSend23> i can’t seem to join the asahi rooms for some reason
2026-02-18 06:09:32 <ikke> Not sure what we can do about that
2026-02-18 06:10:29 <godSend23> idk the reason
2026-02-18 06:10:49 <godSend23> do they require additional verification?
2026-02-18 06:10:54 <ikke> I don't know
2026-02-18 06:11:17 <ikke> Some channels require you to be registered to chanserv
2026-02-18 06:11:22 <ikke> nickserv*
2026-02-18 06:11:26 <ikke> to handle spam
2026-02-18 06:11:39 <godSend23> ohhh
2026-02-18 06:12:45 <godSend23> i think i did that already
2026-02-18 06:13:04 <ikke> Anyway, this is not a general irc support channel
2026-02-18 10:06:54 <Sertonix[m]> prabu: When apk-tools solver returns "Huh?" that is always a bug. I would prefer if these bugs would be fixed instead of documented.
2026-02-18 13:14:08 <prabu> Sertonix, yes, i agree with you. added a note to file a bug in the wiki entry..
2026-02-18 13:40:31 <quinq> prabu, add a note to contact Sertonix[m] directly
2026-02-18 13:54:07 <runxiyu> so, i can't shut down my computer right now, but:
2026-02-18 13:55:06 <runxiyu> the 3.23 installation medium has wifi broken on my computer with Intel(R) Wi-Fi 6 AX201 160MHz
2026-02-18 13:55:08 <runxiyu> missing firmware
2026-02-18 14:44:41 <vv221> On Debian, I have a binary git-http-backend provided with the git package, that I can then use to serve git repositories through http (that’s the so-called "smart" backend): https://manpages.debian.org/unstable/git-man/git-http-backend.1.en.html
2026-02-18 14:44:55 <vv221> On Alpine, I expected to find it under /usr/libexec/git-core/, but it seems to be missing.
2026-02-18 14:45:29 <vv221> is there a command that could tell me if the file name "git-http-backend" is provided by some package I might not have installed yet, similar to Debian apt-file?
2026-02-18 14:47:25 <efahl> apk query to the rescue, probably something like   apk query --fields name,contents --match contents '*git-http-backend*'
2026-02-18 14:48:49 <efahl> I added "contents" in the fields list so you can confirm, but maybe you just need "name" there.
2026-02-18 14:49:32 <vv221> Got it, that was git-daemon. Let me check the command you suggested to see if it would have found it…
2026-02-18 14:50:08 <ikke> Afaik, it will only report it for packages that you have installed
2026-02-18 14:50:25 <vv221> Oh, right, I just installed that package…
2026-02-18 14:50:32 <ikke> There is no remote index for what files a package provides
2026-02-18 14:50:44 <vv221> Right, it does not check remote packages.
2026-02-18 14:50:56 <efahl> Should be the full repo list, you usually need to do 'apk --installed query ...' to restrict it.
2026-02-18 14:51:28 <vv221> No, it does indeed only report a result if git-daemon is already installed, no matter if I add --installed or not.
2026-02-18 14:51:44 <vv221> Maybe --fields contents only works for local packages?
2026-02-18 14:51:51 <vv221> (local = installed)
2026-02-18 14:52:00 <efahl> Did you do 'apk update' recently?  It should look in all the indexes available.
2026-02-18 14:53:17 <vv221> I did, it does not help.
2026-02-18 14:53:23 <vv221> Let me illustrate: https://paste.debian.net/plainh/f3487a36
2026-02-18 14:53:36 <ikke> efahl: it does not contain the file listings
2026-02-18 14:53:55 <ikke> That would take a lot of space
2026-02-18 14:54:38 <vv221> That makes sense, Debian has that in a distinct dedicated index so it is not fetched unless one has explicitly installed apt-file.
2026-02-18 14:54:39 <efahl> Oho, got it.  Only stuff like name, depends and all that ar in the indexes...
2026-02-18 14:54:45 <ikke> Ye
2026-02-18 14:55:49 <vv221> I should probably take the habit to check the list of sub-packages for a given package, I guess it would answer "What does provide that command?" in most cases.
2026-02-18 14:56:01 <vv221> (this is how I found about the git-daemon sub-package)
2026-02-18 14:58:42 <vv221> All good now, this first cgit instance has now been migrated \o/
2026-02-18 14:59:10 <vv221> This should make the next ones much easier to handle, now that I have identified the main differences with my previous setup.
2026-02-18 14:59:49 <vv221> (thankfully my Debian was already running OpenRC, reducing the diff)
2026-02-18 14:59:57 <f_> vv221: apk has some neat ways to check what packages contain what. Try e.g. `apk search cmd:ls`
2026-02-18 15:00:22 <f_> you can even do `apk add cmd:ls` and it'll pick the first package providing the command `ls`
2026-02-18 15:00:50 <vv221> Nice, thanks for the hint!
2026-02-18 15:01:20 <vv221> It would not have helped here (git-http-backend is in /usr/libexec), but is going to be useful in my other migrations.
2026-02-18 15:02:00 <f_> Yeah it wouldn't, but just thought I'd mention it, has helped me so many times
2026-02-18 15:07:08 <vv221> I need all these hints, I have not learned a new packages manager in almost 20 years…
2026-02-18 15:11:36 <efahl> There's a man/wiki page somewhere that describes all the --match options for apk query, talks about pseudo-fields and such...
2026-02-18 15:13:40 <efahl> Just for fun, I'm trying to map all the list, info and search commands into queries, as that's teaching me a lot.
2026-02-18 15:15:18 <efahl> apk query --match provides '*git-http-backend*'  might be the global search for the commmand (longer form of 'git search ...')
2026-02-18 18:28:48 <jpp1> How do I install the man pages?
2026-02-18 18:29:02 <jpp1> hi
2026-02-18 18:29:06 <lopid> mandoc
2026-02-18 18:29:07 <ikke> jpp1: apk add docs
2026-02-18 18:29:24 <ikke> that will automatically install any -doc subpkg for packages you have installed
2026-02-18 18:37:32 <jpp1> How do we install ls?
2026-02-18 18:40:11 <Saijin_Naib[m]> @jpp1 apk search cmd:ls
2026-02-18 18:40:25 <Saijin_Naib[m]> I use lsd with an alias
2026-02-18 18:41:50 <jpp1> Yes, but I'm in a VM and I can't scroll the window.
2026-02-18 18:44:00 <jpp1> and there are many lines
2026-02-18 18:47:48 <courge_musque[m]> maybe pipe the output in less
2026-02-18 18:49:43 <jpp1> How do we do it?
2026-02-18 18:50:10 <lopid> ls | less
2026-02-18 18:52:26 <jpp1> ok thanks
2026-02-18 18:52:39 <vv221> This docs meta-package is a really nice idea, I would miss it if I were to go back to Debian.
2026-02-18 19:03:37 <jpp1> No entry for ls in the manual
2026-02-18 19:43:32 <penguinz_rule[m]> the busybox manpagee are only within man busybox
2026-02-18 19:43:47 <penguinz_rule[m]> so you need to /ls and hit n a whole bunch of times
2026-02-18 19:51:20 <jvvv> or use a more specific seach regex, like: /^[[:space:]]*ls
2026-02-18 20:10:06 <jpp1> It's not clear, what exactly am I supposed to do?
2026-02-18 20:16:32 <jvvv> jpp1: what are you trying to do exactly?
2026-02-18 20:19:32 <jpp1> Read the man page for the ls command
2026-02-18 20:20:06 <jvvv> and if you run 'man ls', it returns "No entry ..."?
2026-02-18 20:20:57 <jpp1> Yes!
2026-02-18 20:21:56 <penguinz_rule[m]> the manpages arent per command lik3 gnu coreutils though
2026-02-18 20:22:02 <jvvv> ok, what penguinz_rule[m] was suggesting was to read the manual page for busybox 'man busybox', but you will need to scroll down quite a ways for the 'ls' portion
2026-02-18 20:24:04 <jvvv> also, you can install the posix manual pages: 'apk add man-pages-posix'
2026-02-18 20:24:08 <jpp1> Ah, okay.
2026-02-18 20:24:57 <jvvv> if you install man-pages-posix, then 'man ls' should give some reading material.
2026-02-18 20:25:45 <jpp1> So I need to type: man busybox?
2026-02-18 20:27:03 <jvvv> that will present the manual page for busybox. it includes descriptions of busybox's commands (on of which is ls)
2026-02-18 20:27:32 <jvvv> (one of which is ls)
2026-02-18 20:27:44 <lpz> 
ACTION wonders why not just drop "linux man ls" to the search engine of ones choosing?

2026-02-18 20:28:00 <vv221> Because search engines are crap.
2026-02-18 20:28:24 <lpz> vv221, Ah. Forgot that.
2026-02-18 20:28:47 <vv221> It’s easy to forget if you stopped using them ;)
2026-02-18 20:32:57 <jpp1> thanks
2026-02-18 20:34:23 <jvvv> jpp1: glad to help
2026-02-19 05:05:34 <runxiyu> hi! just asking if anyone has thoughts on https://gitlab.alpinelinux.org/alpine/aports/-/issues/17936 before i continue investigating it. my approach is going to be the 'push upstream and occasionally add patches' option
2026-02-19 05:05:49 <runxiyu> s/push upstream/poke upstream maintainers and supply patches when appropriate/g
2026-02-19 05:42:01 <dimalloy> hello
2026-02-19 05:42:17 <dimalloy> can I dualboot alpine?
2026-02-19 05:42:44 <Biswa96[m]> yeah, sure. It's Linux :)
2026-02-19 05:47:06 <dimalloy> I know it's linux but the script installer doesn't seem to let me install to a partition
2026-02-19 05:53:33 <Biswa96[m]> the partitions needs to set manually before installing. this wiki may help https://wiki.alpinelinux.org/wiki/Dualbooting
2026-02-19 06:19:04 <dimalloy> can I dualboot alpine?
2026-02-19 06:19:37 <dsc_> yes, it is linux :)
2026-02-19 06:20:11 <dsc_> the wiki may help: https://wiki.alpinelinux.org/wiki/Dualbooting
2026-02-19 07:33:31 <quinq> Hi
2026-02-19 07:33:32 <quinq> cat login.defs.apk-new
2026-02-19 07:33:32 <quinq> USERGROUPS_ENAB yes
2026-02-19 07:33:36 <quinq> is that expected
2026-02-19 07:33:37 <quinq> ?
2026-02-19 07:33:42 <quinq> (that there's only one line)
2026-02-19 07:38:02 <jpp1> Hi
2026-02-19 07:40:34 <jpp1> Be patient, we'll surely answer you. :)
2026-02-19 07:41:13 <jpp1> Personally, I don't know.
2026-02-19 07:44:21 <quinq> Can't be patient in the era of A.I.!
2026-02-19 07:51:39 <jpp1> AI is crap
2026-02-19 07:53:43 <jpp1> https://notes.vv221.fr/ia-génératives.xhtml
2026-02-19 07:55:19 <jpp1> An ecological disaster
2026-02-19 07:55:21 <jpp1> The generation of databases and the configuration of generative AI consumes an insane amount of resources (especially water) and energy, the production of which is extremely polluting. Generative AI running locally is not exempt, as the generation of its database has already been carried out upstream through these destructive processes. Each use of this AI validates the
2026-02-19 07:55:23 <jpp1> destruction of our conditions for survival in favor of the profits of a handful of companies.
2026-02-19 07:59:51 <runxiyu> hi im getting this while accessing gitlab
2026-02-19 07:59:54 <runxiyu> Error: access denied: denied by administrative rule 185e3abfb5c9a3f7f984cd58fe83fd69/4a29df8e66a9867d99b8
2026-02-19 08:00:09 <jpp1> Hi
2026-02-19 08:01:47 <jpp1> Be patient, we'll surely answer you. :)
2026-02-19 14:39:23 <mrsvcd> My aport merge request failed because riscv64 compile took over 1 hour. Is this common? (simh is quite slow to compile.)
2026-02-19 14:40:41 <Biswa96[m]> yup, it's common. 
2026-02-19 14:40:58 <ikke> mrsvcd: you can increase the timeout on your forks setttings (CI/CD settings)
2026-02-19 14:52:03 <mrsvcd> Ok, how? switch interruptible to false?
2026-02-19 14:58:17 <ikke> No
2026-02-19 14:58:39 <ikke> There's setting called "
2026-02-19 14:58:41 <ikke> "Timeout"
2026-02-19 15:01:04 <mrsvcd> ah, I was in the wrong place. 'Settings' -> 'CI/CD' then 'Timeout' not 'CI/CD configuration'
2026-02-19 15:04:37 <cousin_luigi> If I wanted to create a package from a git snapshot, should I always provide a tarball or can abuild produce one itself?
2026-02-19 15:05:11 <ikke> cousin_luigi: you should provide one
2026-02-19 15:05:21 <cousin_luigi> Ok, thanks.
2026-02-19 15:05:35 <ikke> many git archives can generate one on the fly
2026-02-19 15:05:38 <ikke> git forges*
2026-02-19 15:08:34 <vv221> Alpine packaging is fully based on realease/snapshot tarballs? I like that!
2026-02-19 15:09:19 <vv221> There are some people pushing to move away from tarballs in Debian, it played a part in my departure for greener pastures (Alpine being the first one).
2026-02-19 15:10:16 <frag> as opposed to what? iso files?
2026-02-19 15:10:37 <vv221> git tags
2026-02-19 15:11:01 <frag> oh, beyond my understanding nvm :)
2026-02-19 15:11:35 <vv221> I love git as a version control system, I use it several hours daily and have been for more than a decade.
2026-02-19 15:11:49 <vv221> But I see it as a version control system, not as a delivery system.
2026-02-19 15:12:36 <vv221> I like that tarballs are a clean and minimal interface between upstream and distributions.
2026-02-19 15:17:47 <frag> very simple/nice indeed
2026-02-19 15:19:05 <frag> busybox wget alpine.tar;busybox tar xf alpine.tar
2026-02-19 15:32:08 <andar1an> Hello all. I am debating an attempt at creating a penpot package for aports as I want to see what it is about. The self-host deployment is based on containerized orchestration of several components such as postgres db, a webui. and some other binaries. I know I have seen some packages that allow one to specify various databases, but memory isn't great this morning. Does anyone have any suggestions of a package that may be a good reference?
2026-02-19 15:33:28 <ikke> andar1an: it may depend. Zabbix has multiple builds against different database libraries
2026-02-19 15:37:20 <andar1an> cool, thank you. Will take a look. If helpful to visualize, the overall the docker compose elements of interest seem to be a postgres db, frontend webui, backend binary, an exporter which I am assuming is for metrics (maybe otel) but not sure yst, something called valkey which I dk what it does yet, and a mailcatch server. I am assuming that any reverse proxy stuff can be excluded from any attempt in aports. 
2026-02-19 15:38:09 <andar1an> just worried I will be going down rabbit hole, but if this seems feasible I want to give it a go
2026-02-19 15:47:07 <andar1an> cool, thanks so much @ikke, this seems like great reference
2026-02-19 16:05:12 <andar1an> on a tangent, does anyone using openrc user services understand how environment variables are supposed to propagate to a graphical session?
2026-02-19 16:09:25 <andar1an> have said variabled set in rc_env_allow, export them in user service, but for life of me I don't know how to make them appear yet
2026-02-19 16:10:39 <andar1an> exporting in a start_pre(){} function, but have tried others
2026-02-19 16:16:37 <vv221> andar1an, any specific example? Here it seems I only use XDG_RUNTIME_DIR, but I set it both in the openrc script and in ~/.profile
2026-02-19 16:17:06 <vv221> Oh, wait, I use that for the SSH agent too, let me check that…
2026-02-19 16:18:07 <vv221> Right, it do it the same: the variable is set both in the openrc script and in ~/.profile, so I don’t even try propagating it.
2026-02-19 16:20:05 <_5f> hola
2026-02-19 16:35:05 <chuyinweilai[m]> h
2026-02-19 16:45:59 <andar1an> @vv221, yes. Been messing with SSH_AUTH_SOCK and SSH_ASKPASS. With compositor export in config never had issue. Tring to use dbus user service for dbus session now instead of dbus-run-session, and it seems okay so far, but I am unable to export variables like that in a user init script for ssh-agent for example
2026-02-19 16:46:30 <andar1an> it creates the socket I specify in the correct location, and I can pass the environment variable through cli manually, but it doesn't exist in environment
2026-02-19 16:47:00 <andar1an> I don't set it in profile. I was previously setting in in compositor config for niri
2026-02-19 16:47:13 <andar1an> then it would be available everywhere
2026-02-19 16:47:40 <andar1an> I use nushell a lot, not sure if that reads .profile either
2026-02-19 16:54:36 <vv221> When not relying on ~/.profile, for example with GUI-only variables like the SDL ones, I export them in the script starting my GUI session.
2026-02-19 16:54:57 <andar1an> I am starting my gui session with greetd
2026-02-19 16:55:36 <andar1an> just been exploring openrc user services and trying to make that work in a similar fashion
2026-02-19 16:55:37 <vv221> I don’t know that one, I start it from a shell script I wrote, ending with `exec sway`.
2026-02-19 16:56:14 <vv221> I used to have a GUI session relying on user services… but at the time I was still using X.org and systemd, so nothing really transferable here ;)
2026-02-19 16:56:59 <andar1an> no worries. It has mostly worked out so far, only beein exploring for a few days. The main issue I have understanding right now is Environment variables
2026-02-19 16:57:34 <andar1an> and based on docs I don't think it is necessary to have an rc conf for a user, I think it falls back to system rc conf. But I am not 100% certain
2026-02-19 16:58:36 <andar1an> few avenues to tinker with later
2026-02-19 20:18:51 <mrsvcd> I must have submitted my merge request at the right time, it took like 15 min to build on riscv compared to timing out aka >1h last few tries.
2026-02-19 20:19:48 <mrsvcd> any way, Good Night
2026-02-20 13:17:20 <Guest3271> test
2026-02-20 13:24:21 <longnoserob[m]> Guest3271: maybe passed
2026-02-20 13:27:14 <Guest3271> hehe 
2026-02-20 13:50:19 <fabricionaweb> When we have a new release, where is the best place to get updated about it? Like, a gitlab watch, a rss, or something else
2026-02-20 13:58:03 <n2j3> https://www.alpinelinux.org/atom.xml 
2026-02-20 14:01:21 <fabricionaweb> that is the one I thought, but, I think, it is also a "blog"
2026-02-20 14:28:56 <cousin_luigi> Perhaps there's a feed for tags on some git repo?
2026-02-20 14:37:30 <cousin_luigi> fabricionaweb: https://github.com/alpinelinux/aports/tags.atom try this one
2026-02-20 15:10:05 <fabricionaweb> than we are trusting in tags from aports - I think its fine - but I was looking to get to know more ways, if exists
2026-02-20 15:12:28 <snchmt[m]> Hello, since yesterday I had issues when connecting my roccat keyboard to unlock the luks partition, after a lot of research I finally got it working by creating a new mkinitfs feature with the specific linux driver (here usb/hid-rocat-isku.ko). I did not have this issue with the installation ISO, inside the firmware the setting or the grub interface. Is it the way to go or should I report an issue somewhere ? Inside the kernel or
2026-02-20 15:12:28 <snchmt[m]> mkinitfs I didn't find any relevant code
2026-02-20 17:24:33 <andar1an> patching a rust dependency in apkbuild file actually isn't so bad! a non-default unwrap() seemed to be secret sauce. This kinda made my day haha. My patch may be making a test fail, but was meaning to learn this hahaha
2026-02-20 17:25:13 <andar1an> *unpack() not unwrap. Different things lol
2026-02-20 17:46:04 <andar1an> is it possible to reference a variable in APKBUILD in a patch? 
2026-02-20 17:46:41 <andar1an> e.g. if you are setting a version number variable, can you use that in patch, or must it be hardcoded? 
2026-02-20 17:46:56 <ikke> A patch is static
2026-02-20 17:47:11 <andar1an> Should I put a version number in patch name then?
2026-02-20 17:48:34 <fabricionaweb> you can use a string and replace with sed in prepare step
2026-02-20 17:48:56 <andar1an> also, should one patch tests of a dependency too? or is it okay to skip
2026-02-20 17:56:30 <ikke> We prefer static patches over sed replacements
2026-02-20 17:56:43 <ikke> When something changes, it's clear
2026-02-20 17:56:59 <ikke> If the patch does not apply, you investigate why
2026-02-20 17:57:08 <ikke> if the patch changes, it's clear in the git history
2026-02-20 17:57:24 <ikke> With sed replacements, you have no idea what is replaced, and if it still applies after upstream changes
2026-02-20 17:58:14 <andar1an> okie, that is good to know. Right now it is static, I just didn't know if a version was desired for clarity. But I guess the file may not change across versions.
2026-02-20 17:58:22 <andar1an> thank you @ikke
2026-02-20 22:50:43 <thresh> hey.  I'm using Alpine on EC2, and recently the golden image VMs I make on top of the official ones started to have issues on boot.  Only 3.21 and 3.22, it seems (3.19, 3.20, 3.23 are fine).  When setting up the machines I'm basically installing a bunch of packages, and I'm not messing with kernel modules etc.
2026-02-20 22:51:08 <thresh> Now on 3.22 (3.21 at least, but let's focus on 3.22), the VM seems to failback to busybox shell
2026-02-20 22:51:32 <thresh> [    0.461819] Mounting root...
2026-02-20 22:51:40 <thresh> mount: mounting [    5.891744] Mounting root: failed. 
2026-02-20 22:51:41 <thresh> UUID=c06b4640-7bac-42da-b825-9c3cf0b78c46 on /sysroot failed: No such file or directory
2026-02-20 22:52:27 <thresh> I'm not able to see anything in blkid (and nvme kernel module isnt loaded at this point)
2026-02-20 23:01:36 <vv221> TIL https://alpinelinux.org/posts/2025-10-01-usr-merge.html
2026-02-20 23:02:00 <vv221> I notice that my 3.23 setups of Alpine are *not* using aliased directories.
2026-02-20 23:02:12 <vv221> /bin and /usr/bin are still distincts.
2026-02-20 23:03:04 <vv221> I installed them directly as 3.23 (not upgraded from an older version), from lxc-create --template alpine
2026-02-21 00:01:33 <vv221> I was looking for needrestart, but then realised it is already built into apk-autoupdate. Well done!
2026-02-21 01:14:45 <Sertonix[m]> Yes, the timeline on the blog post isn't up to date
2026-02-21 04:56:51 <f_> vv221: yes usr-merge began then it went canceled because it broke `apk audit`
2026-02-21 04:57:29 <f_> it will still happen but at a later date
2026-02-21 06:27:09 <rnkn> how do I change my shell, there is no chsh; do I need to install shadow?
2026-02-21 06:28:25 <longnoserob[m]> looks like it
2026-02-21 06:29:03 <longnoserob[m]> but login shell is also defined in /etc/passwd so usermod should also work
2026-02-21 06:29:05 <ikke> or edit /etc/passwd directly
2026-02-21 06:34:27 <rnkn> hmm usermod is part of shadow too, and there's no vipw for editing /etc/passwd
2026-02-21 09:15:39 <chatterino[m]> Hello! Does anyone have an idea on how to delay the networking service or fasten up the dhcp discovery to decrease boot times? I already have networking in my async runlevel. 
2026-02-21 09:16:32 <cousin_luigi> chatterino[m]: Is your dhcp server slow to reply?
2026-02-21 09:17:17 <chatterino[m]> I would assume no, is there a way to measure it?
2026-02-21 09:17:40 <chatterino[m]> It's very fast on my phone.
2026-02-21 09:18:31 <Jenkler> why is the time it takes a problem? 
2026-02-21 09:19:26 <Jenkler> Its quite fast as it is. and its only once a bootup 
2026-02-21 09:19:26 <cousin_luigi> chatterino[m]: Are you present when dhcp discovery occurs? Do you actually see it hang?
2026-02-21 09:19:39 <chatterino[m]> cousin_luigi: Yes
2026-02-21 09:20:00 <chatterino[m]> It prints 3x the line for discovery
2026-02-21 09:20:17 <chatterino[m]> and then does it again for the eth0
2026-02-21 09:20:19 <cousin_luigi> How long does it take? And does it need to do that for multiple interfaces?
2026-02-21 09:20:29 <cousin_luigi> I mean, is it required for them all?
2026-02-21 09:20:42 <cousin_luigi> And is ipv6 also in use?
2026-02-21 09:21:03 <chatterino[m]> Yes it is required otherwise I would've simply use a static ip.
2026-02-21 09:21:52 <Jenkler> have you tried diffrent dhcp clients ? 
2026-02-21 09:22:18 <chatterino[m]> Yes it is very fast on my phone. I also just cleared the lease for it and tried again.
2026-02-21 09:22:39 <cousin_luigi> Your phone uses one interface, I suppose. How many interfaces are we talking about here?
2026-02-21 09:22:53 <chatterino[m]> Two
2026-02-21 09:23:02 <cousin_luigi> chatterino[m]: Perhaps there are scripts in ifup?
2026-02-21 09:23:07 <quinq> chatterino[m], mate it parallel in rc
2026-02-21 09:23:15 <cousin_luigi> that too
2026-02-21 09:23:20 <chatterino[m]> it already is
2026-02-21 09:23:43 <chatterino[m]> it is listed in my /etc/runlevels/async
2026-02-21 09:24:05 <cousin_luigi> chatterino[m]: do you use ifupdown?
2026-02-21 09:24:48 <quinq> I think that Jenkler meant different DHCP clients, on your machine that you feel is slow
2026-02-21 09:25:03 <Jenkler> Yes 
2026-02-21 09:25:50 <Jenkler> sometimes a client can be broken. I have seen it in Cloud Stack 
2026-02-21 09:26:19 <Jenkler> or more the dhcp server that needs another client 
2026-02-21 09:26:43 <chatterino[m]> cousin_luigi: i suppose
2026-02-21 09:26:52 <chatterino[m]> i don't use networkmanager
2026-02-21 09:26:59 <cousin_luigi> chatterino[m]: I mean there's also ifupdown-ng
2026-02-21 09:27:10 <cousin_luigi> Not sure if that might give some performance advantage.
2026-02-21 09:27:15 <chatterino[m]> yes I use that
2026-02-21 09:27:18 <cousin_luigi> It looks like there's a bottleneck somewhere else.
2026-02-21 09:27:20 <chatterino[m]> i checked the version with a cmd
2026-02-21 09:27:26 <cousin_luigi> chatterino[m]: ifupdown or ifupdown-ng ?
2026-02-21 09:27:39 <chatterino[m]> ifupdown-ng 0.12.1
2026-02-21 09:27:49 <cousin_luigi> sorry, I got confused
2026-02-21 09:28:06 <cousin_luigi> there's only ifupdown-ng on alpine. I thought it was on #devuan
2026-02-21 09:28:14 <chatterino[m]> The ifupdown-ng package is installed as part of alpine-base package and provides necessary tools for managing network configuration in Alpine Linux. 
2026-02-21 09:29:13 <quinq> I don't think so, you get busybox-ifupdown by default
2026-02-21 09:29:27 <quinq> ifupdown-ng has to be installed explicitely
2026-02-21 09:30:04 <chatterino[m]> Maybe I get smarter with this:
2026-02-21 09:30:04 <chatterino[m]> The ifupdown-ng scripts prioritize dhcpcd over udhcpc as per /usr/libexec/ifupdown-ng/dhcp script. 
2026-02-21 09:31:40 <Jenkler> I do not know if udhcpc is faster, but i could be worth trying :) 
2026-02-21 09:32:34 <Jenkler> *it 
2026-02-21 09:33:09 <cousin_luigi> I thought udhcp was used for ipv6 acquisition. At least it is on openwrt.
2026-02-21 09:33:13 <chatterino[m]> Might be worth a shot. The script itself only checks for any of these three dhcp clients and then uses one of them.
2026-02-21 09:33:33 <cousin_luigi> But yes, worth a try indeed. Still, I would like to know if timeout is involved.
2026-02-21 09:35:22 <chatterino[m]> I'm looking on how to improve udhcpc first since I prefer to stick with the default services.
2026-02-21 09:43:08 <quinq> IPv6 usually works with RAD, not really DHCPv6 (though it exists of course)
2026-02-21 09:43:28 <quinq> And I doubt busybox even supports it (though I don't know)
2026-02-21 10:01:47 <cousin_luigi> dhcpv6 is not supported by Android, for some reason.
2026-02-21 10:59:35 <vv221> f_, feels like déjà-vu, with usrmerge in Debian breaking dpkg in subtle ways ;)
2026-02-21 11:00:38 <f_> yes :P
2026-02-21 12:05:34 <vv221> Did I get bitten by aliased paths less than one hour after writing about /usr-merge? Nobody can prove that.
2026-02-21 12:06:38 <vv221> (unrelated with /usr-merge on Alpine, it’s something in my software that uses symbolic links, but still a funny coincidence)
2026-02-21 14:42:17 <Renan[m]> Alpine dont have support for BTRFS?
2026-02-21 14:42:40 <infinitywisdom[m]> It does
2026-02-21 14:42:58 <Renan[m]> infinitywisdom[m]: why gparted dont have the option to use btrfs?
2026-02-21 14:42:59 <infinitywisdom[m]> https://wiki.alpinelinux.org/wiki/Btrfs
2026-02-21 14:43:27 <infinitywisdom[m]> Did you install the btrfs packages?
2026-02-21 14:43:50 <infinitywisdom[m]> If they are not installed then gparted does not have access to them. Like many other file system types.
2026-02-21 14:45:07 <infinitywisdom[m]> If you intend to use btrfs with alpine itself you will have to do so at the point of installation. If you just want gparted to use btrfs for messing around with hard drive partitions in that format. Install the btrfs packages, restart gparted (or) the computer, and then you should be good to go.
2026-02-21 14:45:17 <Renan[m]> infinitywisdom[m]: no
2026-02-21 14:45:19 <Renan[m]> infinitywisdom[m]: ok, thanks
2026-02-21 14:45:37 <infinitywisdom[m]> No problem. Glad to help. Hope it works out for you.
2026-02-21 15:02:39 <Renan[m]> <infinitywisdom[m]> "No problem. Glad to help. Hope..." <- now my pendrive with btrfs sees to be read only, what i need to do?
2026-02-21 15:04:00 <Renan[m]> Renan[m]: i think it dont is in read-only
2026-02-21 15:04:05 <Renan[m]> but i cant modify
2026-02-21 15:05:33 <Renan[m]> I cant see other hd too
2026-02-21 15:05:44 <infinitywisdom[m]> Unsure. Usually you have to set the right flags when you originally partition the hard drive. My suggestion would be to use search.brave.com, or another AI to try and troubleshoot the problem.
2026-02-21 15:06:05 <Renan[m]> 
ACTION uploaded an image: (136KiB) < https://matrix.org/oftc/media/v1/media/download/AV6_P928ZsmWdXJoNCgsUAuNhJ_o2csRIu0qZevfQBPsAYFLOmm6_yc6GpmJ8JpLuJCH7QNfROZFqbbqbnJ_6EFCecyNWR-wAHVucmVkYWN0ZWQub3JnL05hV1dpTExMWHRqVVFoWWRDR21RVXhwWQ >

2026-02-21 15:07:27 <penguinz_rule[m]> do you have udisks2 installed
2026-02-21 15:07:55 <penguinz_rule[m]> udisks2
2026-02-21 15:07:55 <penguinz_rule[m]> udisks2-openrc
2026-02-21 15:07:55 <penguinz_rule[m]> udisks2-dev
2026-02-21 15:07:58 <penguinz_rule[m]> i have these for use with k3b which relies on udisks2
2026-02-21 15:08:28 <Renan[m]> penguinz_rule[m]: i think not
2026-02-21 15:10:29 <penguinz_rule[m]> idek why i have \1-dev but, it shouldnt make a difference
2026-02-21 15:12:10 <Renan[m]> 
ACTION uploaded an image: (55KiB) < https://matrix.org/oftc/media/v1/media/download/AaowO6hC4Q72u6lbG-yU4eKyPIavqBuLYzQ6IYAT7a1Z-NRc1mbbC2Ev9o7To43jeTL-QYuMoUf85MEM7b_D5_NCecyNsipAAHVucmVkYWN0ZWQub3JnL1ZERlZhc29JZE1ueEp1d1R5YnFtT2tMRw >

2026-02-21 15:12:45 <Renan[m]> also, i dont have kdewallet but i am pronpeted every time to make one
2026-02-21 15:16:42 <Renan[m]> also, i cant use the pendrive i  Fedora
2026-02-21 15:16:50 <Renan[m]> i have the same problem in fedora
2026-02-21 15:42:32 <Renan[m]> 
ACTION uploaded an image: I can see my pendrive in gparted but not in kde parted. (81KiB) < https://matrix.org/oftc/media/v1/media/download/AVP8iBjDmwZ9rwMsuF-TKDvd4T5e5SjKhoAvK0Dd0_TjEqE5liyt-GCs8X7x3WCoZQhUqW9gvKl2jthup7YJdCZCecyPbxLQAHVucmVkYWN0ZWQub3JnL0RUaUNwWFVhdlNGVk1Pa3FmcHlETE9OTw >

2026-02-21 15:43:26 <lopid> was that url the uuencoded image itself
2026-02-21 15:51:05 <cousin_luigi> Renan[m]: A regular screenshot would be more intelligible.
2026-02-21 15:56:22 <Renan[m]> cousin_luigi: i cant
2026-02-21 15:56:41 <cousin_luigi> Renan[m]: Aren't you in front of that machine?
2026-02-21 15:57:33 <Renan[m]> cousin_luigi: I need pendrive to login matrix
2026-02-22 05:13:04 <rnkn> okay editing /etc/passwd by hand to change my shell locked me out of my account
2026-02-22 05:13:18 <rnkn> is there seriously no way to change a user's shell in the base system?
2026-02-22 05:17:19 <rnkn> nevermind, I can the path to the shell wrong
2026-02-22 05:17:25 <rnkn> s/can/got/
2026-02-22 05:34:49 <swee> when will a package update in edge reach a stable release? waiting for plasma 6.6 to reach v3.23
2026-02-22 05:52:08 <Saijin_Naib[m]> @swee should be at next stable, roughly every 6mo, unless a security update or other big fix
2026-02-22 05:52:20 <Saijin_Naib[m]> In general
2026-02-22 12:20:36 <witcher01> i'm trying to use my bluetooth headset but whenever i try to play audio wireplumber seems to be crashing with SIGSEGV. is this potentially a known issue, maybe even with an available fix?
2026-02-22 12:56:58 <witcher01> looks like the problem resolved itself after trying to debug it for a bit
2026-02-22 14:42:38 <f_> Is anyone else having strange issues with chromium right now? I'm trying to run it on armv7 but all it does is crash the whole thing, requiring a force reboot, and I'm wondering if it's a pmOS issue or alpine issue
2026-02-22 14:43:19 <f_> Not really able to get much logs sadly, even starting it in gdb makes the whole thing freeze
2026-02-22 14:44:57 <longnoserob[m]> let me fire up my tf201
2026-02-22 14:45:20 <f_> ty!
2026-02-22 14:45:54 <longnoserob[m]> package is chromium?
2026-02-22 14:46:02 <f_> it worked on a similar device but with chromium 144.0.7559.109-r0
2026-02-22 14:46:07 <f_> yes chromium, latest from repos
2026-02-22 14:46:15 <quinq> From pmos or from alpine?
2026-02-22 14:46:27 <f_> pmOS does not package chromium
2026-02-22 14:46:46 <quinq> So you're using Alpine
2026-02-22 14:46:48 <f_> when I say pmOS issue I more meant something terribly wrong with the kernel running on this tablet
2026-02-22 14:47:06 <f_> it's a samsung exynos4 tablet running pmOS armv7 openrc
2026-02-22 14:47:32 <f_> the other tablet runs systemd
2026-02-22 14:47:44 <longnoserob[m]> edge/systemd/tegra-armv7/ja-JP
2026-02-22 14:54:06 <longnoserob[m]> hmm installing and i get i/o error
2026-02-22 14:54:42 <f_> uhhhh wat
2026-02-22 14:54:56 <f_> that same chromium works on my other exynos4 tablet (w/systemd)
2026-02-22 14:55:04 <f_> this is screwed up
2026-02-22 14:55:14 <longnoserob[m]> could be just my network
2026-02-22 14:59:28 <f_> https://ircb.dersco.re/uploads/funderscore/9fe15405-f_-paste.txt oh boy
2026-02-22 14:59:59 <f_> okay so it's probably not an alpine issue - sorry for the noise
2026-02-22 17:16:33 <dogg0> love alpine its almost like art and it just works
2026-02-22 17:17:28 <courge_musque[m]> Yesterday, I was helping some guy with an asus desktop that couldn't boot to anything through grub, I installed alpine on his laptop and it just worked fine
2026-02-22 17:18:17 <courge_musque[m]> (he absolutely wanted mint, so I updated the bios/uefi so we could install it, and wiping alpine form the disk broke my heart)
2026-02-22 17:24:43 <dogg0> courge_musque[m], its certainly not for everyone
2026-02-22 17:25:32 <courge_musque[m]> well, once setup with kde/pipewire/networkmanager/drivers, it's smooth sailing
2026-02-22 17:42:25 <dogg0> i just ran setup-desktop for the first time ever and works ootb nicely
2026-02-23 00:34:42 <Saijin_Naib[m]> @dogg0 setup-desktop IMO is one of the most significant developments for Alpine to be used in more traditional usecases
2026-02-23 00:35:00 <Saijin_Naib[m]> I can easily step someone through that and have them successfully boot into a DE and be happu
2026-02-23 00:48:26 <dogg0> Saijin_Naib[m], i needed a couple of desktop vms, and alpine-desktop got me there in just a handful of minutes
2026-02-23 00:48:51 <dogg0> love it !
2026-02-23 09:46:26 <Saijin_Naib[m]> It is incredible
2026-02-23 14:22:54 <RoadRoller01[m]> i forgot how to search which pkg containsa file or own it with apk
2026-02-23 14:23:15 <RoadRoller01[m]> i want to know who own /usr/sbin/remove-shell
2026-02-23 14:24:35 <jvvv> apk info -W PATH
2026-02-23 14:27:07 <RoadRoller01[m]> bro i thought it was within apk-query
2026-02-23 14:27:09 <RoadRoller01[m]> thansk
2026-02-23 14:29:37 <jvvv> yes, you can also use: apk query --match owner PATH
2026-02-23 14:31:09 <RoadRoller01[m]> i tries this
2026-02-23 14:31:12 <RoadRoller01[m]> apk query --match contents /usr/sbin/remove-shel
2026-02-23 14:31:27 <RoadRoller01[m]> apk query --match contents /usr/sbin/remove-shell
2026-02-23 14:31:34 <RoadRoller01[m]> RoadRoller01[m]: tried*
2026-02-23 14:31:45 <RoadRoller01[m]> and it didn't fine anything
2026-02-23 14:31:54 <RoadRoller01[m]> jvvv: yeah this works
2026-02-23 14:32:32 <RoadRoller01[m]> what is the difference between owning and containing?
2026-02-23 14:32:36 <jvvv> the contents field should be followed by a package name rather than a file path
2026-02-23 14:33:12 <jvvv> owner -> what package owns this path
2026-02-23 14:33:42 <jvvv> contents -> what files are contained in this package
2026-02-23 14:33:50 <RoadRoller01[m]> oooh
2026-02-23 14:34:24 <RoadRoller01[m]> the website didn't find neither why is that?
2026-02-23 14:34:27 <RoadRoller01[m]> https://pkgs.alpinelinux.org/contents?file=remove-shell&path=&name=&branch=edge&repo=&arch=
2026-02-23 14:35:00 <RoadRoller01[m]> apk query --match owner /usr/sbin/remove-shell... (full message at <https://matrix.org/oftc/media/v1/media/download/AS8o7ivuBX4vZmBqh0vuylRt5CG0H2YGXrxswStL6VsE3VP_X7oL0IaaWY2KPnlm50KYB0rG3F3_MmfQ40loLhVCec0wXUxwAG1hdHJpeC5vcmcvcU1JUGhwWnFQTVdvQVRPVXR1aExvUmxF>)
2026-02-23 14:36:39 <penguinz_rule[m]> why does alpine distribute busybox without ed? it seems strange to forgo it
2026-02-23 14:47:45 <jvvv> git history for busybox describes busybox ed (at the time of the alpine commit, 20221219) as having critical bugs
2026-02-23 15:01:49 <Sertonix[m]> [@roadroller01:matrix.org](https://matrix.to/#/@roadroller01:matrix.org) apk is (unfortunatly, in my opinion) following symlinks when searching for the owner package.
2026-02-23 15:44:29 <jvvv> remove-shell is exactly that, a busybox symlink, so won't show up as part of a package. the symlink is installed by the post-install script, so it is not tracked.
2026-02-23 16:51:52 <realroot[m]> i trying guestmount and it cannot find guestfs commands
2026-02-23 16:51:52 <realroot[m]> command
2026-02-23 16:52:45 <realroot[m]> libguestfs: error: cannot find any suitable libguestfs supermin, fixed or old-style appliance on LIBGUESTFS_PATH (search path: /usr/local/lib/guestfs) 
2026-02-23 17:03:27 <Shiz>    RoadRoller01[m] : i forgot how to search which pkg containsa file or own it with apk
2026-02-23 17:03:38 <Shiz> if it's a binary, a quicker way is `apk info cmd:<binary>`
2026-02-23 17:04:07 <Shiz> e.g.
2026-02-23 17:04:09 <Shiz> # apk info cmd:update-extlinux
2026-02-23 17:04:11 <Shiz> syslinux-6.04_pre1-r19 description:
2026-02-23 17:09:51 <dwfreed> realroot[m]: alpine does not ship any appliances for libguestfs; you can download prebuilt ones from here: https://download.libguestfs.org/binaries/appliance/
2026-02-23 17:11:26 <realroot[m]> does that work in alpine? cause of musl
2026-02-23 17:12:08 <dwfreed> it's run in a VM
2026-02-23 17:12:31 <dwfreed> libguestfs literally starts its own VM to mount the filesystem you give it
2026-02-23 17:12:46 <dwfreed> the "appliance" is the VM image that is run
2026-02-23 17:13:55 <realroot[m]> i wanted to do it to chroot into VM is that possible?
2026-02-23 17:14:42 <dwfreed> I mean, it still appears to be a filesystem (via FUSE) to the system you're running the command on, so yes
2026-02-23 17:15:33 <dwfreed> just don't expect any kind of performance and you'll be fine
2026-02-23 17:22:03 <realroot[m]> guestmount -i -a /var/lib/libvirt/images/image.qcow2  mnt-point ?
2026-02-23 17:23:42 <realroot[m]> i extracted the archive from https://download.libguestfs.org/binaries/appliance/ : `ls appliance initrd  kernel  root`
2026-02-23 17:27:11 <broke> hello, I was looking to do the /etc/security/limits.conf on my alpine system, how do you do it on alpine? couldn't find any wiki for it.
2026-02-23 18:47:01 <dwfreed> realroot[m]: I don't think you're supposed to extract it, I think you're supposed to leave the tarball as is and put it in the directory referenced by that message
2026-02-23 18:47:35 <dwfreed> but you'll have to read the guestfs docs for the specifics
2026-02-23 22:22:01 <RoadRoller01[m]> Shiz: thanks! The example didn't work in my system
2026-02-23 22:22:19 <RoadRoller01[m]> i think because i am in aarch64
2026-02-23 22:22:32 <RoadRoller01[m]> <Sertonix[m]> "[@roadroller01:matrix.org](https..." <- So this is a bug?
2026-02-23 22:22:40 <Shiz> oh, that's odd -- was it a package you had installed?
2026-02-23 22:22:51 <Shiz> oh wait, aarch64
2026-02-23 22:22:54 <Shiz> yeah that doesn't have syslinux :D
2026-02-23 22:24:49 <Sertonix[m]> RoadRoller01[m]: There is an issue open for that
2026-02-23 22:24:49 <RoadRoller01[m]> what is syslinux 🤔
2026-02-23 22:25:04 <Sertonix[m]> A bootloader
2026-02-23 22:25:43 <RoadRoller01[m]> it should be renamed 
2026-02-23 22:25:57 <RoadRoller01[m]> btlinux is better
2026-02-23 22:34:21 <user0> isn't syslinux discontinued or something?
2026-02-23 22:34:47 <Shiz> RoadRoller01[m]: but you can also e.g. try with `apk info cmd:ls`
2026-02-23 22:35:21 <Shiz> user0: don't think it is
2026-02-23 22:36:44 <user0> hmm
2026-02-23 22:36:50 <user0> according to wikipedia, it is
2026-02-23 22:37:10 <user0> also nice to see you here :)
2026-02-23 22:39:22 <pj> "last release was 7 years ago, but I see recent commits in 2019-xx-xx, surely it's wrong"
2026-02-23 22:39:39 <pj> 
ACTION remembers it's 2026

2026-02-23 22:39:53 <user0> yes, its 2026
2026-02-23 22:40:05 <vv221> [citation needed]
2026-02-23 22:41:53 <Ermine> see syslinux repository
2026-02-23 22:44:24 <vv221> (that was a direct reaction to the last message of user0 about the current year, I have no reason to doubt the claims about syslinux)
2026-02-23 22:46:04 <pj> my $(date) says it's 2017 so I'm eagerly waiting for the next 2 years of worth of commits 😎
2026-02-23 22:46:27 <Shiz> dang, I guess it is
2026-02-23 22:47:22 <Shiz> user0: (also, have we met? apologies, I'm really terrible with names)
2026-02-23 22:47:48 <user0> Shiz: yes, #copyparty on libera
2026-02-23 22:47:57 <Shiz> ah there we go
2026-02-23 22:48:09 <Shiz> that makes sense why the oftc /whois didn't show up anything familiar
2026-02-23 22:48:19 <user0> yeah
2026-02-23 22:52:41 <Shiz> user0: you could say i've been part of the furniture for a while :p https://git.alpinelinux.org/apk-tools/commit/?id=5ba27c90007b2441f1fe35365c753a5365f3a2de  
2026-02-23 22:53:25 <user0> oh, cool
2026-02-24 03:44:20 <runxiyu> https://paste.sr.ht/~runxiyu/f32b1486c45863a52bbf5853f90a438447140127
2026-02-24 03:44:28 <runxiyu> installation medium firmware issues (regression from 3.22)
2026-02-24 03:44:32 <runxiyu> where to report this?
2026-02-24 03:44:42 <runxiyu> linux-firmware package from 3.23 seems to work fineish
2026-02-24 03:44:48 <runxiyu> at least it didn't break when i upgraded
2026-02-24 04:13:32 <jvvv> runxiyu: is that paste from the installer environment?
2026-02-24 04:17:05 <runxiyu> yea
2026-02-24 04:19:53 <jvvv> which image are you booting, standard or extended?
2026-02-24 04:20:45 <dwfreed> intel wifi firmware was split into a separate package
2026-02-24 04:20:51 <dwfreed> it's not installed during building the image
2026-02-24 04:21:01 <dwfreed> so it's not in the modloop
2026-02-24 04:21:14 <jvvv> it is not in linux-firmware-intel?
2026-02-24 04:21:18 <dwfreed> it is
2026-02-24 04:21:27 <dwfreed> but linux-firmware-intel is not installed
2026-02-24 04:22:00 <jvvv> right, ok. but the apk is on the extended image
2026-02-24 04:22:34 <jvvv> so realistally, one could install it into  the installer env then load the module
2026-02-24 04:23:52 <jvvv> that would be what I would try if I had no available ethernet connections to plug in
2026-02-24 04:25:05 <dwfreed> aye
2026-02-24 04:27:55 <jvvv> tbh, I keep a 70' ethernet cable around I made up for just this kind of situation
2026-02-24 04:28:50 <jvvv> while a 70' patch cable is not optimal for regular use, works in a pinch
2026-02-24 04:29:09 <dwfreed> assuming the device has an ethernet port, anyway
2026-02-24 04:29:28 <runxiyu> extended
2026-02-24 04:29:47 <dwfreed> so just install linux-firmware-intel and then reload iwlwifi
2026-02-24 04:30:03 <runxiyu> iirc i tried a few weeks back but i had readonly filesystem issues despite / being a tmpfs
2026-02-24 04:30:09 <runxiyu> ill reboot and investigate in a few moments
2026-02-24 04:30:46 <jvvv> mount -o remount,rw /
2026-02-24 04:31:15 <dwfreed> the modloop is probably mounted at /lib/firmware (or a symlink)
2026-02-24 04:31:24 <jvvv> good point
2026-02-24 04:31:31 <dwfreed> overlayfs to the rescue
2026-02-24 04:35:46 <runxiyu> ill try after fisni writing another bug report, ty
2026-02-24 05:05:20 <jvvv> dwfreed: i guess it has been a while since i did much with overlayfs. took me all this time to prove out doing it that way in a vm
2026-02-24 05:06:43 <jvvv> but it is probably the easiest solution
2026-02-24 07:25:03 <frag> been using chromium a bit, was surprised a webcast did work, finally tried firefox where it worked. Which browser generally works better when it comes to actually functioning with different sites like this?
2026-02-24 07:27:12 <quinq> none
2026-02-24 07:27:37 <quinq> You'd have to ask the persons responsible for the sites in question
2026-02-24 07:27:42 <quinq> Most of the time they'll have no idea
2026-02-24 07:30:28 <frag> *didnt work
2026-02-24 07:30:37 <lopid> oooohh
2026-02-24 07:30:42 <frag> hehe
2026-02-24 07:32:58 <quinq> Though you could have a look its parameter regarding hardware acceleration and/or software codec decoding
2026-02-24 07:34:17 <frag> hmm might be my settings, try to make chromium more privte respecting..
2026-02-24 07:34:24 <frag> yeah
2026-02-24 07:39:22 <rnkn> hello, my keyboard backlight is controlled by the value of /sys/class/leds/input0::scrolllock/brightness which I activate with a simple acpi script. since upgrading to 3.23 I've had a problem where the backlight will deactivate upon entering any GNOME text field. where might I report/troubleshoot this (i.e. which gitlab project)?
2026-02-24 07:42:43 <lopid> try another distro that uses the same version, if you can
2026-02-24 07:44:03 <rnkn> lopid: another linux distro?
2026-02-24 07:44:09 <lopid> yes
2026-02-24 07:44:16 <lopid> a live image
2026-02-24 07:45:18 <rnkn> ah okay cool, thanks
2026-02-24 07:45:32 <rnkn> i.e. narrow it down to either Alpine or GNOME?
2026-02-24 07:45:36 <lopid> yep
2026-02-24 09:20:00 <rnkn> okay just tried on the Debian live image and the issue is not there, so it seems to be isolated to Alpine
2026-02-24 09:20:18 <rnkn> next step I think is isolating it to a likely culprit?
2026-02-24 09:28:17 <rnkn> whatever it is it seems to be repeatedly and aggressively resetting the value of /sys/class/leds/input0::scrolllock/brightness
2026-02-24 10:14:05 <lopid> try older kernels if you have them
2026-02-24 10:41:48 <rnkn> I don’t really have kernels lying around
2026-02-24 12:39:24 <vv221> frag, if your focus is privacy, I strongly suggest *not* using a chromium basis.
2026-02-24 12:39:39 <vv221> (not that firefox is good, but it is slightly "less bad")
2026-02-24 12:40:52 <frag> rly? hmm, ppl seem to have different opinions about it, kinda worn down resisting the survilance tho ;( frigging police state infrastructure..
2026-02-24 12:42:18 <courge_musque[m]> I've tried ungoogled chromium recently, it was an unusable nightmare, I don't know how people can put up with it
2026-02-24 12:43:09 <frag> huh, never tried ungoogled, but figured it would be good ..
2026-02-24 12:43:15 <vv221> Well, chromium is under the full control of Google, who are known to be one of the worst (if not the worst) actor in surveillance.
2026-02-24 12:44:21 <vv221> Here I go one step further, and do not even allow Blink. So no QtWebEngine for example.
2026-02-24 12:44:46 <frag> you use ff vv221? maybe librewolf?
2026-02-24 12:46:03 <vv221> Firefox for now, as it’s the only one packaged in Debian. With uBlock Origin and uMatrix, both with very strict rules (blocking much more things than the default). And some Firefox-specific setup to disallow as much as its anti-features than I can (like LLM creep).
2026-02-24 12:46:12 <frag> after qutebrowser started crashing i been kinda switching browsers, trying out different ones
2026-02-24 12:46:36 <vv221> When switching my main computers to Gentoo, I might switch to a Firefox fork too, if one is packaged. Librewolf is indeed part of my list.
2026-02-24 12:47:06 <vv221> But my main move here is unrelated to the Web browser: I try to rely on Web stuff as little as possible.
2026-02-24 12:47:32 <frag> yes, i guess thats the real solution
2026-02-24 12:47:48 <vv221> The least I see "http(s)" in any given day, the best I feel about it.
2026-02-24 12:48:24 <vv221> This is one part of Debian that I am going to miss: I could contribute with packaging and stuff relying only on git and e-mail.
2026-02-24 12:48:51 <vv221> Their bugtracker for example is fully mail-driven, the Web interface can be avoided.
2026-02-24 12:50:41 <frag> why not use alpine?
2026-02-24 12:51:09 <frag> i like the diskless solution a lot
2026-02-24 12:54:22 <vv221> That’s not the question, it would be "Why Alpine?" ;)
2026-02-24 12:54:40 <vv221> (I am already using it by the way, it’s replacing Debian stable in my LXC containers)
2026-02-24 12:55:02 <vv221> My choice was between Debian, Arch Linux and Gentoo for my main system.
2026-02-24 12:55:48 <vv221> But I am leaving Debian, and find no appeal at all in Arch Linux. And with the addition of public repositories of binary packages, Gentoo lost what I saw as its major downside.
2026-02-24 12:56:22 <vv221> About "Why these 3 systems?", they are the ones supported by ./play.it: http://doc.dotslashplay.it/index.en.xhtml
2026-02-24 12:57:38 <vv221> (someone is working on adding Alpine support, and I am probably going to help them, but it will not be ready for a while)
2026-02-24 13:13:51 <frag> what is this play.it?
2026-02-24 13:14:58 <vv221> The best software in human history (you might guess that I am its main developer). It’s a native packages generator for video games, you can think of it as alien but for DRM-free commercial video games.
2026-02-24 13:15:20 <vv221> So some kind of .tar/.sh/.exe → .deb/.pkg.tar.zst/etc. converter.
2026-02-24 13:15:45 <vv221> It’s kind of the opposite of all Steam clones.
2026-02-24 13:16:24 <vv221> Instead of trying to replicate the all-in-one captive experience, it integrates video games in your existing environment, relying only on standard tools.
2026-02-24 13:16:58 <frag> hmm.. weird to call it play .. also installed google play on my graphene phone, kinda cringe name for an app store
2026-02-24 13:16:59 <vv221> So you install/update them with your packages manager, start them from your usual applications menu, etc.
2026-02-24 13:17:13 <frag> maybe like plug and play, shrug
2026-02-24 13:17:16 <vv221> The real name is "./play.it"
2026-02-24 13:18:09 <vv221> Because originally it was a collection of shell scripts all named play-(…).sh, so would run the packaging process with: ./play-(…)
2026-02-24 13:21:47 <vv221> (I never used the Google Play store, or anything similar, it was not even all that popular under that name when I started working on ./play.it)
2026-02-24 13:23:25 <frag> "press play on tape"
2026-02-24 13:24:24 <vv221> I had to have a quick look around me to make sure, but it seems I no longer have any tape player here ;)
2026-02-24 16:13:28 <RoadRoller01[m]> i want to know which processes my battery while in zzz/suspend>
2026-02-24 16:13:28 <RoadRoller01[m]> As for some reason my device warm up while its in suspend.
2026-02-24 16:13:28 <RoadRoller01[m]> What do you  recommend? 
2026-02-24 19:44:49 <frag> one thing i like about chromium is that it uses the primary selection when copying links and similar..
2026-02-24 19:58:18 <frag> .. oh, seems librewolf does too..
2026-02-24 20:03:43 <dwfreed> That's just the nature of X11
2026-02-24 20:04:27 <dwfreed> (or GTK is doing that automatically)
2026-02-24 20:09:19 <mason> Or you can have a clipboard manager that helps.
2026-02-24 20:14:26 <frag> dont think i ever used a clipboard manager, feels wrong/bloat
2026-02-24 20:19:11 <dwfreed> I have so much useful stuff that gets cycled in and out of my clipboard, it's kind of nice
2026-02-24 20:19:48 <dwfreed> I run xfce, so I use xfce-clipman
2026-02-24 20:20:12 <invoked> primary selection and clipboard are meant to be separate things, but x11 left a lot of behavior undefined and up to toolkits, so users are left to untangle what's supported and what actually happens
2026-02-24 20:20:42 <invoked> most people think they're the same thing
2026-02-24 20:21:07 <enigma9o7_irc> and wayland got rid of the middle-click-to-paste-selection feature that x11 had built in.  some apps (like browsers) have added it, but only in their app
2026-02-24 20:21:23 <invoked> it's a mess
2026-02-24 20:24:41 <invoked> honestly wayland really fucked this part up
2026-02-24 20:24:52 <frag> google ai convinced me librewolf is a lot better for privacy
2026-02-24 20:26:22 <invoked> a better way to look at it (imo) is that librewolf tries to patch the bad changes mozilla puts in over time
2026-02-24 20:26:52 <frag> ... than chromium
2026-02-24 20:28:14 <invoked> i prefer to use non-google stuff over ungoogled stuff, wherever possible.
2026-02-24 20:29:17 <frag> kinda liked chromium, buut value privacy
2026-02-24 20:30:03 <invoked> just on principle
2026-02-24 20:34:19 <vv221> frag, don’t trust chromium, but don’t trust Google (or any other) LLM either.
2026-02-24 20:34:25 <vv221> These things will rot your brain.
2026-02-24 20:35:28 <ssm> http has baked in fingerprinting (useragent, referrer), try to gophers/gemini :) :) :)
2026-02-24 20:35:35 <vv221> enigma9o7_irc, I’m surprised to learn about that, because I kept middle-click paste in all of my GUI software after moving to Wayland. I can’t think of any exception.
2026-02-24 20:35:40 <ssm> gophers not a typo, gopher + tls
2026-02-24 20:35:49 <frag> vv221 sounded right tho.. eg sends info about sites you visit back to google so it can check if they are safe.... and that its hard to turn off all such stuff..
2026-02-24 20:36:16 <vv221> LLM do "sound right", it’s literally what they’re for ;)
2026-02-24 20:37:05 <ssm> cp shakespear shakespear_ai WAOUW 
2026-02-24 20:37:13 <ssm> 100% accurate AI copy!
2026-02-24 20:37:42 <invoked> vv221: xwayland is very good at what it does
2026-02-24 20:38:17 <invoked> (assuming your apps are still x11)
2026-02-24 20:38:26 <vv221> Sure, but most of my stuff is native Wayland.
2026-02-24 20:38:33 <vv221> (maybe even all of it now)
2026-02-24 20:38:52 <ssm> hope xwayland doesn't cause the wine effect where no one actually ports native software because "xwayland works good enough"
2026-02-24 20:38:53 <vv221> I’m still keeping wxayland ~forever, because video games.
2026-02-24 20:39:27 <invoked> it depends, i think. i'm not sure in all cases it's "check for wayland, then fallback to x11"
2026-02-24 20:39:28 <vv221> ssm, the wine effect is a bit different; it is actually a Valve effect. They campaigned hard *against* native ports once they got control of Proton.
2026-02-24 20:40:16 <vv221> (and sadly, it worked, as they managed to brainwash players once again, so they are the one putting pressure on devs)
2026-02-24 20:41:48 <invoked> ssm: not until wayland can get high framerates, i would guess
2026-02-24 21:10:09 <courge_musque[m]> vv221: in steam, you always get the native version of the game when you install it, and the proton version is opt-in
2026-02-24 21:10:55 <courge_musque[m]> For a lot of games, the native version has issues, so players intentionnaly chose the proton version
2026-02-24 21:10:55 <vv221> Sure, but it did not prevent them from campaigning *against* native builds.
2026-02-24 21:11:18 <courge_musque[m]> More of a dev issues than a valve issue imo
2026-02-24 21:11:21 <vv221> (mosr probably for the usual vendor lock-in reasons)
2026-02-24 21:11:39 <vv221> Well, *Valve* campaigns sound like a Valve issue to me, not a dev one ;)
2026-02-24 21:11:56 <vv221> Of course it’s devs making the final decision.
2026-02-24 21:12:20 <vv221> But when you get pressure from the only significant vendor that your will have to work with, it can be a bit much.
2026-02-24 21:12:27 <courge_musque[m]> They are a company that wants to make money, and clearly waiting for individual studios to change isn't the way for 90%+ support
2026-02-24 21:13:14 <courge_musque[m]> But yeah, their campaigns is from them, so of course you can say that they're biased
2026-02-24 21:14:23 <vv221> It’s more than a bias: they are technically dictating what devs of commercial games can or can’t do.
2026-02-24 21:15:10 <courge_musque[m]> Uh, that's a bit much
2026-02-24 21:15:14 <vv221> So when in their documentation for devs they write "Do not provide a native build, rely on Proton instead", I see that as a strong move *against* us (= Linux video games players).
2026-02-24 21:15:29 <vv221> No, not really: no dev can be profitable without Steam.
2026-02-24 21:15:33 <courge_musque[m]> The native version is litteraly prioritized in their proprietary and unavailable client
2026-02-24 21:15:45 <vv221> So you either play by their rules, or find another job.
2026-02-24 21:16:15 <courge_musque[m]> It's documentation, not some kind of rule
2026-02-24 21:16:29 <courge_musque[m]> Even if it's propaganda, they don't force devs to do it
2026-02-24 21:16:42 <courge_musque[m]> Devs weren't doing it anyways
2026-02-24 21:16:48 <vv221> Wrong.
2026-02-24 21:16:56 <vv221> They were doing it already.
2026-02-24 21:17:06 <vv221> It only stopped following Valve new instructions.
2026-02-24 21:17:34 <invoked> valve does have a monopolistic advantage (imho) but also i'm not sure it makes that much of a difference rn. valve has bigger fish to fry, like for instance, making a linux client trustworthy enough for AAA multiplayer
2026-02-24 21:17:34 <courge_musque[m]> No, there were only a handful of games with native builds
2026-02-24 21:17:35 <vv221> We had a big spike of native builds for many years.
2026-02-24 21:18:42 <vv221> I give up, I’m not here to be lectured by someone I literally never saw in my last 15 years working with video games distribution.
2026-02-24 21:19:08 <courge_musque[m]> What, where do I lecture you?
2026-02-24 21:23:01 <invoked> cheating at multiplayer is so endemic. between that, and child protection laws, there's a real legislative push for attestation in the os stack now
2026-02-24 21:24:19 <courge_musque[m]> Is there really cheating outside of shooter games?
2026-02-24 21:24:37 <invoked> there's cheating in everything multiplayer
2026-02-24 21:24:56 <invoked> i'm not going to say i'm an expert, though
2026-02-24 21:25:56 <invoked> there is also an intersection with "voter id" type laws which is seeking to regulate operating systems too
2026-02-24 21:26:35 <ssm> "scan verification can" 2 diffs away
2026-02-24 21:26:35 <invoked> i doubt much of this stuff will get through, and even if some does, whether it will hold up
2026-02-24 21:26:40 <invoked> nonetheless, it's happening
2026-02-24 21:28:28 <invoked> whether they will consider "linux" generically, as a "operating system" remains to be seen
2026-02-24 21:28:46 <courge_musque[m]> Hope that it will be slowed down because lawmakers don't know anything about OS/kernel stuff
2026-02-24 21:28:54 <invoked> they're writing it for all the vertically integrated ones, eg microsoft, apple, google/android
2026-02-24 21:30:07 <invoked> this concept has a lot of support in some places like, australia for instance
2026-02-24 21:30:32 <invoked> anyway, way off topic
2026-02-24 23:07:52 <banansoppa> i have a strange networking issue when trying to run alpine in vmm (openbsd virtual machine). Essentially while running the iso i cannot even ping the vm host at 100.64.1.2, meanwhile if i use alpine 3.8.5 from 2020 i can, debian also can.
2026-02-24 23:08:08 <banansoppa> idk if this is the right place to ask this tho, but has anyone tried running alpine in vmm recently?
2026-02-24 23:11:17 <jvvv> banansoppa: does the happen before or after running setup-alpine or setup-interfaces?
2026-02-24 23:13:06 <banansoppa> well, udhcp fails to get a lease. I have tried adding the interface manually as the dhcp server should, does not work either. So before, during and after. I dont know much about this network stuff.
2026-02-24 23:13:25 <banansoppa> in version 3.8.5 it manages to get a lease 
2026-02-24 23:15:30 <jvvv> maybe dmesg might shine some light... maybe it is a driver issue
2026-02-24 23:17:21 <banansoppa> what should i look for in the dmesg? im not really fluent in dmesg
2026-02-24 23:19:10 <jvvv> dmesg wil give you the kernels' messages, so I would be looking for issues/problems where it messages about the network devices
2026-02-24 23:19:23 <dwfreed> just pastebin the whole output
2026-02-24 23:35:40 <banansoppa> https://pastebin.com/dcnXG58f i grepped for things to do with net, eth, driver or wlan
2026-02-24 23:40:46 <jvvv> like dwfreed mentioned, the whole output would be better
2026-02-25 00:01:09 <banansoppa> https://pastebin.com/PApEtuR8 alright heres the full dmesg
2026-02-25 00:02:00 <banansoppa> im not used to copying files over serial so i had to figure out how to do that
2026-02-25 00:17:28 <jvvv> I do not see any signs of a network device being initialized.
2026-02-25 00:26:10 <banansoppa> probably some sort of firmware issue then?
2026-02-25 00:26:22 <banansoppa> i think im just gonna use a different distro atleast for the time being 
2026-02-25 00:28:03 <banansoppa> but thanks a lot for the help
2026-02-25 00:30:49 <jvvv> banansoppa: i am doing some reading on vmm. too little, too late by the time i know the in's and out's of  it all.
2026-02-25 00:32:46 <banansoppa> its fine, ill stick to debian or something. you were very helpful, goodnight!
2026-02-25 00:32:56 <jvvv> gn
2026-02-25 09:31:15 <el[m]1> `Failed during configuration: Have not found any log file for sshd jail` I get this after installing fail2ban. is the wiki perhaps outdated? https://wiki.alpinelinux.org/wiki/Fail2ban#Install it seems like more steps might be required than it says
2026-02-25 15:08:54 <f_> longnoserob[m]: hi, you around to test something on armv7 again? (one of qutebrowser, angelfish or falkon) Trying to debug sigint on qt6webengine
2026-02-25 15:10:46 <longnoserob[m]> f_: sorry was out doing karaoke..  could test later 
2026-02-25 15:11:00 <f_> no worries
2026-02-25 16:35:09 <realroot[m]> I copied appliance folder  into  /usr/local/lib/guestfs and then content of it but still error https://github.com/virt-manager/virt-manager/issues/802
2026-02-25 16:35:25 <realroot[m]> ls /usr/local/lib/guestfs  initrd  kernel  root
2026-02-25 17:08:30 <frag> dont see mpstat mentioned much, kinda neat/simple
2026-02-25 17:36:39 <Quantum_3[m]> is webrtc plugin missing from gst-plugins-bad? i recently upgraded my system (i am on edge) and it seems to be missing
2026-02-25 17:38:55 <dwfreed> I see it in https://pkgs.alpinelinux.org/contents?file=libgstwebrtc*&path=&name=&branch=edge&repo=&arch=x86_64
2026-02-25 17:48:54 <Shiz> 2
2026-02-25 18:09:09 <frag> setup-apkrepos seems to still make http links, instead of https, bug or feauture?
2026-02-25 18:09:53 <ikke> frag: From the automatic mirror selection?
2026-02-25 18:10:00 <quinq> Feature, then you don't have to rely on CA mafia
2026-02-25 18:10:49 <ikke> https://mirrors.alpinelinux.org/mirrors.txt
2026-02-25 19:15:11 <staceee> anyone has recommendation for VPS or bare metal provider that support Alpine Linux?
2026-02-25 19:16:35 <DuClare> you wouldn't want a vps that doesn't let you install whatever os you want
2026-02-25 19:16:46 <normikoto> off the top of my head, linode as alpine images, vultr le you load a custom iso, hetzner has an alpine iso you could load (thouugh no alpine vm image itself)
2026-02-25 19:17:08 <normikoto> s/le/lets/
2026-02-25 19:18:49 <normikoto> worst case scenario, you can use funky kexec shennanigans, though it's not for the faint of heart
2026-02-25 19:44:24 <invoked> i'm not sure i'd trust images upstream from alpine (i'm not sure how you verify those in eg cloud manager), but linode will let you use a supplied img file
2026-02-25 19:44:46 <invoked> 
ACTION reflects on trusting trust

2026-02-25 19:46:32 <elagost> vultr has alpine images that seemed to work ok, last I used them. They are pretty cheap.
2026-02-25 21:17:19 <normikoto> if anything i'd trust upstream alpine images over cloud providers
2026-02-25 21:30:27 <p3> hi. trying install alpine standard 64 bits on a lenovo 500w 3gen. but i'm having problems with the identification of wlan0 card./ i was daownload manually differents drivers like linux-intel dhcp iw and many others but when i try install in the installation some have problems to install ... sorry my poor english
2026-02-25 21:31:57 <courge_musque[m]> What is the model of the wifi card?
2026-02-25 22:50:43 <courge_musque[m]> Hello, when a game is crashing the computer, and you can't even switch to a tty, is there a way to escape without restarting the whole computer?
2026-02-25 22:55:00 <vv221> courge_musque[m], that sounds like a loss of input due to an X11 freeze, I used to take back control through the "magic keys", but I don’t know if Alpine enables that by default.
2026-02-25 22:55:36 <dwfreed> they are enabled by default in Linux; I don't think Alpine disables them
2026-02-25 22:55:40 <courge_musque[m]> Yeah it's on x11, where can I find more about the magic keys?
2026-02-25 22:55:57 <dwfreed> "SysRq magic keys" is what you're looking for
2026-02-25 22:56:06 <dwfreed> hope your keyboard has a print screen key
2026-02-25 22:56:43 <vv221> The list is here: https://en.wikipedia.org/wiki/Magic_SysRq_key
2026-02-25 22:57:06 <vv221> I would probably go with "k" first, to kill the X server and claim back the input devices.
2026-02-25 22:57:49 <vv221> That would be, I think: Alt + PrntScrn + K
2026-02-25 22:58:05 <vv221> Beware that’s it’s going to kill your X11 session and everything running in it.
2026-02-25 22:58:27 <dwfreed> you may need to follow it with r
2026-02-25 22:58:42 <dwfreed> you can chain multiple keys as long as you don't release alt
2026-02-25 22:59:49 <dwfreed> note that if you are using a display manager, if it restarts X automatically, *don't* use r after the k
2026-02-25 23:01:00 <courge_musque[m]> alt+prtscrn+k does nothing
2026-02-25 23:01:13 <courge_musque[m]> I have a regular usb keyboard
2026-02-25 23:01:38 <vv221> Slightly related, but that is not going to help right now: I never got such freezes making the whole computer unusable since I switched to Wayland. I still get games so badly coded they can take a lot of things with them, but they can no longer take the input with them.
2026-02-25 23:02:12 <vv221> (that is Wayland + XWayland instead of native X11)
2026-02-25 23:02:31 <courge_musque[m]> Yeah, I would love too, but it's for cinnamon, they won't have good support before at least april/may
2026-02-25 23:03:51 <vv221> Oh, right, I think my flatmate has exactly the same problem here.
2026-02-25 23:04:59 <vv221> Here I run *a lot* of badly coded games, so when I realised the problems I could avoid by ditching native X11, I switch to another desktop environment in the process. I was at *that* level of enthusiasm ;)
2026-02-25 23:05:16 <vv221> (I used to get freezes like you describe almost every week)
2026-02-25 23:05:45 <dwfreed> courge_musque[m]: try 'r' and then 'e'; note that 'e' if successful will terminate basically every process with SIGTERM
2026-02-25 23:05:49 <courge_musque[m]> It's not really for me, it's for a friend on mint cinnamon
2026-02-25 23:06:17 <dwfreed> if 'e' doesn't work, proceed to 'i', which will *kill* every process
2026-02-25 23:06:58 <dwfreed> if 'i' doesn't help, you might try 's' 'u' 'b' and if the system doesn't reboot after 'b', you're going to have to pull the plug
2026-02-25 23:07:23 <courge_musque[m]> Wait, I use the prntscrn from an usb keyboard connected to a laptop, does it matter?
2026-02-25 23:07:29 <dwfreed> that's fine
2026-02-25 23:07:46 <dwfreed> as long as the usb keyboard was plugged in before things stopped working
2026-02-25 23:08:19 <dwfreed> if not, you may have to use the laptop's built-in keyboard
2026-02-25 23:08:44 <courge_musque[m]> I managed to make cinnamon crash by using the combination mentionned on wikipedia:
2026-02-25 23:08:44 <courge_musque[m]> "Alt+Fn+S simultaneously, then releasing Fn and S while still holding Alt"
2026-02-25 23:09:02 <dwfreed> nice
2026-02-25 23:09:19 <courge_musque[m]> But cinnamon managed to recover
2026-02-25 23:09:55 <courge_musque[m]> What I would like is to go back to a state where x11 is dead, to just do a clean restart without killing ext4 in the process
2026-02-25 23:10:16 <dwfreed> that's what 's' 'u' 'b' is intended for
2026-02-25 23:10:18 <courge_musque[m]> Because fsck kills some stuff in the process
2026-02-25 23:10:35 <dwfreed> 's'ync, 'u'nmount (technically it's remount readonly), re'b'oot
2026-02-25 23:11:00 <dwfreed> it's a bit nicer than the old-fashioned three finger salute
2026-02-25 23:11:14 <courge_musque[m]> I did b at first, I'm a little dumb
2026-02-25 23:11:39 <dwfreed> 'b' by itself is the old-fashioned three finger solute
2026-02-25 23:11:44 <dwfreed> s/solute/salute/
2026-02-25 23:11:49 <courge_musque[m]> So I need to do alt+sysreq+s, then the same with u, then the same with b?
2026-02-25 23:12:17 <dwfreed> you can do: hold alt, press sysrq, press s, press u, press b, release alt
2026-02-25 23:12:33 <dwfreed> this way you don't need 3 hands to press all the keys
2026-02-25 23:13:11 <dwfreed> you may want to insert some waiting time between each of s, u, and b
2026-02-25 23:18:15 <vv221> Quick tip: wait a couple seconds before each combination: s, wait 1~2 seconds, u, wait 1~2 seconds, etc.
2026-02-25 23:18:18 <courge_musque[m]> K was good, idk why it doesn't work, but B/S/U do
2026-02-25 23:18:54 <vv221> Oops, dwfreed gave the exact same advice already…
2026-02-25 23:21:08 <courge_musque[m]> Okay, I think that I understand it now, thanks!
2026-02-25 23:21:19 <courge_musque[m]> (But K still doesn't work, weird)
2026-02-25 23:24:17 <dwfreed> check the value of /proc/sys/kernel/sysrq
2026-02-25 23:24:46 <courge_musque[m]> The keys seem to be listened only if they are on the main keyboard, not a virtual or usb one
2026-02-25 23:25:44 <vv221> That rings a bell, I think I had computers (laptops) where I had to use the built-in keyboard. An USB one would not do the trick.
2026-02-25 23:26:22 <courge_musque[m]> sysrq gives a value of 176
2026-02-25 23:26:47 <dwfreed> that explains it
2026-02-25 23:27:00 <dwfreed> https://www.kernel.org/doc/html/latest/admin-guide/sysrq.html
2026-02-25 23:28:03 <courge_musque[m]> 176 isn't even in the table
2026-02-25 23:28:17 <dwfreed> 176 is 128+32+16
2026-02-25 23:28:33 <dwfreed> it is a bit field displayed as a decimal number
2026-02-25 23:28:36 <courge_musque[m]> Ohhh
2026-02-25 23:28:56 <dwfreed> so the only thing enabled is SUB, basically
2026-02-25 23:29:12 <courge_musque[m]> So I need 240
2026-02-25 23:30:04 <vv221> Both my systems here (not Alpine) have 438. If I were to set it up myself, I would simply use 1 for "allow everything".
2026-02-25 23:31:01 <dwfreed> what vv221 said; debian defaults to 438, I would probably want 502 at a minimum (438 + 64, which is the important one from what it's missing)
2026-02-25 23:31:46 <courge_musque[m]> Okay, at 510, k works
2026-02-25 23:31:47 <courge_musque[m]> I'll try at 438
2026-02-25 23:32:16 <dwfreed> k probably won't work with 438, because it's probably part of 64
2026-02-25 23:32:58 <courge_musque[m]> Uh, it worked
2026-02-25 23:33:01 <dwfreed> oh, it looks like it's part of 4
2026-02-25 23:33:10 <courge_musque[m]> Guess I'll keep 438
2026-02-25 23:33:27 <dwfreed> 64 is probably just e and i
2026-02-25 23:33:38 <courge_musque[m]> Idk why a debian derivative doesn't keep debian default
2026-02-25 23:34:21 <courge_musque[m]> Is there a way to change the value at boot, or forever?
2026-02-25 23:34:26 <courge_musque[m]> Maybe as a kernel parameter?
2026-02-25 23:34:30 <dwfreed> sysctl.conf
2026-02-25 23:34:48 <dwfreed> there's probably already a line in that file (it lives in /etc) or in a file in /etc/sysctl.d
2026-02-25 23:35:53 <dwfreed> or check for a line in /usr/lib/sysctl.d (but don't edit a file there; if it exists there, copy the file to /etc/sysctl.d and fix it there)
2026-02-25 23:37:19 <courge_musque[m]> 438 was the default value, but commented
2026-02-25 23:37:24 <courge_musque[m]> Removing the # was enough
2026-02-25 23:37:25 <courge_musque[m]> Thanks a lot!
2026-02-25 23:37:52 <dwfreed> so they probably patch the 176 into the kernel build
2026-02-25 23:52:51 <vv221> On the topic of video games and graphic servers, I don’t really expect an anwser here but it’s worth trying: has anyone ever managed to run the native Linux build of Botanicula on Wayland (+ XWayland)?
2026-02-25 23:53:19 <vv221> (it’s an Adobe Air game, shipping its own build of the Air runtime)
2026-02-26 00:08:55 <invoked> normikoto: i misspoke, i meant vps providers as, i guess, downstream from alpine...
2026-02-26 00:42:18 <kr4kj3n> join #linux
2026-02-26 01:07:03 <jvvv> #17976 should be assigned to me. Is there a way for me to do that is does it need to be done by someone with commit rights?
2026-02-26 02:02:53 <jvvv> mio: thanks!
2026-02-26 02:03:29 <mio> jvvv: you're welcome
2026-02-26 12:15:13 <f_> vv221: is it proprietary?
2026-02-26 12:15:35 <f_> if so you might have some luck with running it inside Distrobox, if it links against glibc
2026-02-26 12:16:30 <f_> Or gcompat
2026-02-26 12:20:52 <penguinz_rule[m]> gcompat sadly doesnt seem to work for a decent amount of programs
2026-02-26 12:23:54 <penguinz_rule[m]> speak of the devil pharo actually works lol
2026-02-26 12:31:12 <vv221> f_, Botanicula? It is proprietary, but this is a Wayland problem, not a musl/glibc one (I get these problems with glibc).
2026-02-26 12:36:00 <f_> vv221: oh ok, what issues are you having?
2026-02-26 12:36:45 <f_> At worst you can possibly just run rootful Xwayland (`xwayland-run` for convenience, or `Xwayland :3` and set `DISPLAY=:3`)
2026-02-26 12:41:48 <vv221> Xwayland :3 does work around it, thanks!
2026-02-26 12:42:10 <f_> Yw!
2026-02-26 12:42:16 <vv221> It does not make it playable yet (only partial display), but at least it gives me something to investigate.
2026-02-26 12:42:49 <vv221> What are the main differences between regular Xwayland and this rootful one I just tried?
2026-02-26 12:43:45 <f_> Xwayland rootless means your Wayland compositor kind of acts like an X11 WM and thus it is seamless
2026-02-26 12:43:58 <f_> Xwayland rootful is comparable to running Xephyr inside an Xorg session
2026-02-26 12:44:13 <vv221> About the initial problem, I get nothing but a black window with the cryptic message "TypeError: Error #1009" in console. I think I managed to get a more detailed error many years ago, but I don’t remember *how*.
2026-02-26 12:44:26 <f_> lol algitbot 
2026-02-26 12:45:00 <vv221> f_, to be fair, I would have difficulties getting the significant differences between Xephyr and a regular Xorg session ;)
2026-02-26 12:45:26 <vv221> (in my mind, it’s made so they are as alike as possible)
2026-02-26 14:24:08 <frag> an annoying thing with librewolf/ff is it seems to obsessively make the ~/Downloads dir o.o
2026-02-26 14:25:07 <frag> even when i configure it to use ~
2026-02-26 14:28:44 <vv221> I do not have this problem here, I don’t think I have ever seen a ~/Downloads directory in the last 10 years ;)
2026-02-26 14:29:07 <frag> wut? weird
2026-02-26 14:30:53 <vv221> Let me check my setup..
2026-02-26 14:31:23 <hoel> I’m interested too, ff keeps creating ~/Downloads here.
2026-02-26 14:31:25 <vv221> OK, first of all, in ~/.config/user-dirs.dirs I have:
2026-02-26 14:31:27 <vv221> XDG_DOWNLOAD_DIR="$HOME/incoming"
2026-02-26 14:31:43 <vv221> I can not promise Firefox follows that, but it might play a part.
2026-02-26 14:32:37 <vv221> In a similar way, browser.download.dir in Firefox about:config is set to /home/${USER}/incoming
2026-02-26 14:32:46 <vv221> I suspect it is what really does the trick here.
2026-02-26 14:33:20 <vv221> browser.download.lastDir is set to the same path.
2026-02-26 14:33:41 <vv221> In addition browser.download.useDownloadDir is set to false.
2026-02-26 14:34:16 <vv221> And that might be it, I see no other relevant setting.
2026-02-26 14:34:49 <hoel> I don’t have browser.download.lastDir.
2026-02-26 14:34:57 <frag> im not accustomed to librewolf, "download.dir" comes up with options boolean/number/string in addition to the download dir that says ~
2026-02-26 14:35:54 <vv221> Literal "~"? Or "/home/something"?
2026-02-26 14:36:08 <frag> home something
2026-02-26 14:36:34 <vv221> OK, so similar to what I have here.
2026-02-26 14:36:56 <vv221> Maybe $XDG_DOWNLOAD_DIR is the real culprit here, not Firefox?
2026-02-26 14:37:00 <invoked> if firefox does not respect the XDG override that's probably a bug
2026-02-26 14:37:08 <frag> i have browser.download.lastDir in librewolf, it has those 3 options again
2026-02-26 14:37:35 <vv221> I would check ~/.config/user-dirs.dirs first, it might not even be a firefox/librewolf thing in the first place.
2026-02-26 14:38:51 <frag> i dont have that file/dir
2026-02-26 14:39:12 <frag> also i dont think almost no XDG environmental vars are set
2026-02-26 14:39:45 <invoked> https://www.phoronix.com/news/Firefox-147-XDG-Base-Directory
2026-02-26 14:46:40 <vv221> Oh, I missed a maybe important detail: I run Firefox ESR, currently that’s version 140.
2026-02-26 14:49:51 <frag> should be a law against this
2026-02-26 15:09:09 <vv221> Against ~/Downloads/ ?
2026-02-26 15:10:38 <frag> yes!
2026-02-26 15:11:03 <penguinz_rule[m]> i like to use ~/media but i still have ~/Downloads
2026-02-26 15:13:52 <frag> hmm ppl actually use dirs like that in home..
2026-02-26 15:14:40 <frag> now i run diskless so use ~ more as a tmp dir, but i think i always used home like that.. a dump/tmp dir/inbox
2026-02-26 15:22:09 <vv221> I have explicit ~/incoming/ and ~/outgoing/ directories for that.
2026-02-26 15:22:52 <vv221> There is already enough sutff in ~ without using it as a buffer ;)
2026-02-26 15:31:40 <invoked> i'm not sure default paths need legislation but mozilla is definitely too slow on some things.
2026-02-26 15:33:11 <invoked> i'm sympathetic to misgivings about fdo things, but this wasn't something that needed 20 years of due diligence. literally inconsequential
2026-02-26 15:37:47 <frag> neither seems to work for me user-dirs.dirs XDG_DOWNLOAD_DIR, whenever i start librewolf it creates ~/Downloads
2026-02-26 15:41:17 <frag> this seems to fix it: while :;sleep 1; rm -r ~/Downloads;done
2026-02-26 15:42:24 <invoked> firefox may support XDG Base but it may not support xdg-user-dirs
2026-02-26 15:42:48 <invoked> if that's the case, it's maybe not a bug
2026-02-26 15:48:43 <penguinz_rule[m]> supporting xdg-base while not supporting xdg-user-dirs is so scuffed
2026-02-26 15:49:29 <frag> it doesnt even follow its own config it seems...
2026-02-26 15:49:48 <frag> how has this not driven everyone mad?!
2026-02-26 15:52:19 <frag> my install is smaller now though :)
2026-02-26 15:52:55 <frag> 874.7MB so tiny!
2026-02-26 15:53:08 <frag> with chromium it was 905MB
2026-02-26 15:57:44 <vv221> Firefox, or more specifically Mozilla, *does* drive everyone furious.
2026-02-26 15:57:55 <vv221> But the alternatives are, well, Google.
2026-02-26 15:58:58 <vv221> (or Epiphany alias Gnome Web, but I’m not sure I can run uBlock Origin + uMatrix on top of that one)
2026-02-26 15:59:25 <frag> how good is zerohosts.txt compared to ublock?
2026-02-26 16:00:05 <Sheila> anything that doesn't run filters on websites is going to be inferior to something that does, unfrotunately.
2026-02-26 16:00:11 <frag> or zerohosts.sh or what its called, blocks sites from /etc/hosts.. used it with surf, seemed to work well for me
2026-02-26 16:00:12 <vv221> I don’t know this one, but I’ve never seen anything even barely as good as uBlock Origin (not to be confused with "ublock" without "origin" in the name, that one is a scam).
2026-02-26 16:00:39 <Sheila> e.g. zerohosts is not going to block ads on YouTube.
2026-02-26 16:01:18 <vv221> Oh, /etc/hosts-based blocks are not going to cut it for me.
2026-02-26 16:01:18 <frag> oh i use mpv for youtube
2026-02-26 16:01:38 <vv221> I use Ublock Origin to block *everything* by default, with an explicit allow-list for a couple websites.
2026-02-26 16:01:39 <frag> basically never view video in the browser
2026-02-26 16:01:52 <vv221> (no audio/video in browser here either)
2026-02-26 16:02:47 <frag> maybe try it out vv221, kinda neat solution independent of browser
2026-02-26 16:03:23 <vv221> Well, I wrote just above why it can not work for my use case ;)
2026-02-26 16:04:14 <vv221> I’m not blocking only a targeted list of known bad actors. I lump all of the Web together as "bad actors until proven otherwise".
2026-02-26 16:04:39 <vv221> You can not do that with /etc/hosts.
2026-02-26 16:05:08 <frag> okok
2026-02-26 16:05:22 <vv221> That would be, in pseudo-/etc/hosts:
2026-02-26 16:05:26 <vv221> * 127.0.0.1
2026-02-26 16:05:40 <frag> installed epiphany from repo, didnt seem to actually load sites
2026-02-26 16:05:41 <vv221> I guess the problem is obvious when written like that ;)
2026-02-26 16:06:56 <vv221> I suggested Epiphany because it is the only mainstream WebKit browser I can think of, so the only non-Firefox and non-Chrome one.
2026-02-26 16:07:30 <vv221> We used to have a lot of Qt browsers for that, but nowadays they all turned into Google Chrome skins.
2026-02-26 16:08:02 <frag> oh its webkit, yeah used surf a good while, its nice but didnt work on some stuff
2026-02-26 16:08:26 <vv221> All browsers but Chrome are going to "not work on some stuff" ;)
2026-02-26 16:08:50 <vv221> Because what used to be standards, has been replaced with similarity to Chrome. With Chrome being the de facto Web standard.
2026-02-26 16:09:38 <vv221> That’s part of the reason why most Web browser developers gave up and are now only developing skins to apply on top of Google Chrome.
2026-02-26 16:10:30 <invoked> it's because most people don't care enough to select software on better fundamental criteria
2026-02-26 16:10:39 <invoked> the problem is always people.
2026-02-26 16:10:55 <frag> actually had to use firefox since chromium didnt work on a webcast the other day, which started me up with browsercomparisons again hehe..
2026-02-26 16:12:09 <invoked> feels like i've said it a million times but, a web browser is arguably the most complex program most people will use directly
2026-02-26 16:12:23 <invoked> if making a web browser was easy, we'd have better options.
2026-02-26 16:14:11 <vv221> invoked, "people" like me would love to rune an ethical Web browser. But the convergence of ethical + usable is exactly 0 when it comes to Web browsers ;)
2026-02-26 16:14:23 <invoked> some of this is by design, now... especially with these vertically integrated companies on all the standards committees
2026-02-26 16:14:38 <vv221> People wanting something sadly is not enough to get said thing to spawn into existence.
2026-02-26 16:15:24 <invoked> if enough people wanted it, it would happen. often things are enough for people to not care/want something different
2026-02-26 16:15:26 <vv221> Web browsers do not have to be that complexe, it’s only because a handful of companies pushed hard to move all of our activities to the Web, because it makes all that easier for them to control.
2026-02-26 16:15:55 <vv221> Nah, I don’t believe in magic, thinking a lot about something does not make it happen ;P
2026-02-26 16:16:14 <invoked> supply and demand
2026-02-26 16:16:18 <invoked> not magic.
2026-02-26 16:16:19 <vv221> It’s a myth.
2026-02-26 16:16:30 <vv221> Free market is a religion, not a science.
2026-02-26 16:16:39 <vv221> People can only buy what is sold to them.
2026-02-26 16:17:25 <vv221> (but I don’t think this is a discussion that can happen through IRC)
2026-02-26 16:17:42 <invoked> people, en masse, have all the power they need to change things, they just don't leverage it.
2026-02-26 16:17:43 <vv221> (so I’m going to leave that as "agreee to disagree")
2026-02-26 16:18:23 <invoked> it's always work for someone else to do, nobody thinks they should get involved
2026-02-26 16:19:01 <invoked> you can say that for just about anything, including web browsers
2026-02-26 16:23:08 <invoked> the only reason we're using firefox right now is because a small handful of ex-netscape people decided to make a new browser (phoenix/firebird->firefox) because aol took over netscape
2026-02-26 16:24:24 <invoked> i'm not sure if ladybird is another example of that but we'll see, i guess
2026-02-26 16:25:00 <invoked> i've not dug into it deeply but some people whose opinions i trust seem to like where it's going
2026-02-26 16:31:44 <frag> uuuuuhh, maybe the ~/Downloads was human error, as i have a cleanup scripts that delets ~/.librewolf... will see...
2026-02-26 16:35:04 <invoked> frag: why are you removing ~/.librewolf ?
2026-02-26 16:35:28 <invoked> because your XDG settings are for somewhere else?
2026-02-26 16:35:56 <frag> i dont remember, but i run diskless, and sometimes i need space
2026-02-26 16:36:27 <invoked> well if you blow away those paths you're killing the whole profile
2026-02-26 16:36:52 <frag> yeah i usually have no/minimal config
2026-02-26 16:37:12 <invoked> hm, alright
2026-02-26 16:38:30 <vv221> It sounds like removing ~/.librewolf would indeed restore the default directory to ~/Downloads, thus lead to the symptoms you described.
2026-02-26 16:39:03 <vv221> Here I have persistent system-wide settings under /etc/firefox-esr/, maybe librewolf has something similar.
2026-02-26 16:39:56 <vv221> (I don’t know what a "diskless" Alpine is, so feel free to discard that advice if /etc/ is not persistent either)
2026-02-26 16:45:08 <invoked> the profile shouldn't be much disk. you can clear the cache (if you have it enabled) without having to sacrifice your settings every time
2026-02-26 16:45:33 <invoked> or you can have it make a new profile and copy your prefs in
2026-02-26 16:46:48 <invoked> (your prefs are user.js)
2026-02-26 16:55:59 <frag> you can make things persistent in diskless, but kinda nice to have a default system, and just make some small changes, also my uptime is often quite long so i dont have to set it every day ;p
2026-02-26 16:56:56 <invoked> you can do that with user.js but it would be manual. anything you change within firefox itself is generated into prefs.js
2026-02-26 16:57:08 <invoked> user.js would be a file you add which sets overrides for anything in prefs.js
2026-02-26 16:57:57 <frag> 131 lines :S
2026-02-26 16:59:47 <invoked> so, in your scenario, first you run 'firefox --ProfileManager', create a profile, copy your user.js in, then run it. something like that
2026-02-26 17:02:49 <vv221> frag, I take it you run Alpine stable, ot edge? Here on Debian unstable (close to Apine edge) my uptime never grows above a couple weeks due to Linux kernel updates.
2026-02-26 17:02:59 <vv221> *Alpine stable, not edge?
2026-02-26 17:03:48 <frag> stable yes, mpv on edge was buggy for me
2026-02-26 17:04:57 <vv221> I’ve not tried edge yet (I have no real use case for it), I only cherry pick a couple packages from edge/testing.
2026-02-26 17:05:49 <frag> how do you do that? stable only has sxmo-surf, edge has surf
2026-02-26 17:06:06 <frag> i think ill stick with librewolf, but....
2026-02-26 17:06:59 <invoked> no reason not to keep using librewolf. it's basically just firefox with less crap
2026-02-26 17:07:21 <brickster> pretty much just firefox but you don't have to undo all the bad defaults
2026-02-26 17:07:47 <brickster> it even builds from latest
2026-02-26 17:07:48 <frag> it seems nice yeah, im a bit suckless fan so kinda cool to use surf too, but gotta be practical..
2026-02-26 17:08:43 <invoked> sertonix works on both alpine and librewolf so you could say it's well supported
2026-02-26 17:09:39 <invoked> (sorry for the ping)
2026-02-26 17:12:28 <vv221> frag, if you asked about cherry-picking from testing above, I rely on tagged repositories for that: https://wiki.alpinelinux.org/wiki/Repositories#Using_testing_repository
2026-02-26 17:14:43 <frag> ah yes
2026-02-26 17:24:03 <Quantum_3[m]> somehow between v3.23 and edge (1.26.9 vs 1.28.0) the gstreamer webrtc plugin disappeared (gst-inspect-1.0 cant find it)
2026-02-26 17:24:03 <Quantum_3[m]> it looks like the /usr/lib/gstreamer-1.0/libgstwebrtc.so file is missing in edge (though i dont really see anything that changed in the APKBUILD that would cause the plugin to stop working) 
2026-02-26 17:24:03 <Quantum_3[m]> any fixes though? currently i just downgraded the package
2026-02-26 17:40:40 <frag> hmm seem to still get the Downloads: if i start librewolf and quit right away, no Downloads, if i start it, type in "test" so it duckduckgos it, then quit, it creates it :S
2026-02-26 17:43:44 <frag> librewolf creates ~/Downloads when it quites.. 
2026-02-26 17:44:01 <brickster> probably won't if you change the default folder?
2026-02-26 17:44:05 <brickster> for downloads I mean
2026-02-26 17:44:17 <brickster> idk never tried doing that 
2026-02-26 17:46:53 <frag> it does t regardless, bit-terror
2026-02-26 17:47:14 <ikke> Quantum_3[m]: Did you check upstream changelogs?
2026-02-26 17:51:46 <ikke> oh, last commit is 4 years ago
2026-02-26 17:54:15 <jvvv> I think gstreamer upstream shifted to different repo. Still on gitlab.fdo.
2026-02-26 17:54:28 <ikke> Not this? https://gitlab.com/gstreamer/gst-plugins-bad
2026-02-26 17:54:53 <ikke> oh, that's not fdo
2026-02-26 17:55:17 <jvvv> I'm looking at this: https://gitlab.freedesktop.org/gstreamer/gstreamer
2026-02-26 17:55:19 <Saijin_Naib[m]> #17948 I had other plugins go missing in 1.28
2026-02-26 17:58:22 <jvvv> I am checking to see if this change is related: https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/8433/diffs
2026-02-26 17:59:22 <ikke> I was checking buildlog differences
2026-02-26 18:00:07 <jvvv> The change I reffed would need ssdpd-dev I think... running some test builds
2026-02-26 18:02:45 <jvvv> I need to dig deeper I think
2026-02-26 18:05:35 <jvvv> One thing I note is that there is still a versioned libgstwebrtc lib in /usr/lib, but the unversioned plugin lib went missing from /usr/lib/gstreammer-1.0
2026-02-26 19:31:11 <jvvv> I think I've isolated reason gst-plugins-bad is missing the webrtc plugins. libnice needs bump to 0.1.23 and gst-plugins-bad need to depend on webrtc-audio-processing-2 (-1 might work, not sure but will look at it)
2026-02-26 19:32:37 <jvvv> libnice has at least on issue in tests that needs fix, otherwise should be easy so long as bumping it does not break something else
2026-02-26 22:08:32 <user0> hi, how good/bad is nvidia support on alpine currently?
2026-02-26 22:08:47 <user0> (someone asked me to ask this, i don't use nvidia gpus myself)
2026-02-26 22:09:43 <Saijin_Naib[m]> If you have a newer GPU that works properly with nouveau, it should be decent. I think RTX 2xxx and above are the cutoff for that
2026-02-26 22:10:13 <Saijin_Naib[m]> I had GTX 1050TI and GTX 650TI and GT 240, but they didn't behave right (no clockrate scaling), so I switched to Intel Arc
2026-02-26 22:10:35 <user0> mmh
2026-02-26 22:10:38 <vv221> Alpine has no non-free nvidia driver?
2026-02-26 22:10:44 <Saijin_Naib[m]> Correct
2026-02-26 22:10:53 <penguinz_rule[m]> the nonfree depends on glibc methinks
2026-02-26 22:10:56 <Saijin_Naib[m]> Just nouveau, which on the newer cards allows for clockrate scaling
2026-02-26 22:11:21 <user0> great, thanks
2026-02-26 22:11:22 <vv221> Sure, but "newer nVIDIA GPU" is something very few people could afford ;)
2026-02-26 22:11:23 <Saijin_Naib[m]> And no cuda runtimes, either, I don't think
2026-02-26 22:11:37 <Saijin_Naib[m]> Right, hence why I changed to Intel after something like 20 years on NVIDIA
2026-02-26 22:11:56 <vv221> (my only nvidia is a GTX 1050, it is never going to get correct nouveau support)
2026-02-26 22:12:27 <vv221> (my main system had been AMD for decades anyway, the nvidia one is mostly so I can debug problem specific to their GPUs)
2026-02-26 22:12:37 <vv221> *has been, as in still is
2026-02-26 22:12:38 <Saijin_Naib[m]> penguinz_ruleYeah, and myself and a number of other folks asked NVIDIA on their github to build against MUSL years ago and still silence from them
2026-02-26 22:21:08 <courge_musque[m]> Even nouveau doesn't work well on Alpine
2026-02-26 22:21:33 <courge_musque[m]> I got a segfault for two different graphics card when trying to use vulkan
2026-02-26 22:24:40 <courge_musque[m]> localhost:~# vulkaninfo --summary
2026-02-26 22:24:40 <courge_musque[m]> Segmentation fault
2026-02-26 22:25:13 <courge_musque[m]> So yeah, I'll stick to intel/amd for now
2026-02-26 22:28:21 <Saijin_Naib[m]> Did you try to capture any crash logs/dumps and open an issue?
2026-02-26 22:44:43 <courge_musque[m]> I should
2026-02-26 22:44:59 <courge_musque[m]> I'll do it next week
2026-02-26 22:53:15 <Saijin_Naib[m]> 😍
2026-02-26 23:44:20 <neartheythemorhehim[m]> hi. how do i get a repository's signing keys after adding it to /etc/apk/repositories? do i just check the server for *.pub and download it manually?
2026-02-27 01:08:01 <Sertonix[m]> There is no generic way. It depends on the repo
2026-02-27 04:38:56 <RoadRoller01[m]> hello
2026-02-27 04:39:16 <RoadRoller01[m]> i am trying to build chromium 
2026-02-27 04:39:29 <RoadRoller01[m]> but i am getting a lot of:
2026-02-27 04:39:29 <RoadRoller01[m]> which already exists!  Assume -R? [n]
2026-02-27 04:39:29 <RoadRoller01[m]> The next patch would create the file chrome/browser/bromite_flag_choices.h,
2026-02-27 04:39:40 <RoadRoller01[m]> how to male it always yes
2026-02-27 04:39:53 <RoadRoller01[m]> i tried yes | abuild ...
2026-02-27 04:40:04 <RoadRoller01[m]> but for some reason it didn't work
2026-02-27 04:49:36 <Biswa96[m]> you have to rebase the patches on top of new versions. chromium is huge :)
2026-02-27 04:51:08 <RoadRoller01[m]> its a dumb thing 
2026-02-27 04:51:18 <RoadRoller01[m]> there was just left overs from last build 
2026-02-27 04:51:24 <RoadRoller01[m]> i need to just abuild clean
2026-02-27 04:51:27 <RoadRoller01[m]> (:
2026-02-27 06:21:49 <jpp1> Chromium is Google, a gafam, so it's not for me.
2026-02-27 06:25:21 <RoadRoller01[m]> sadly firefix is lagy in my system 
2026-02-27 06:28:05 <jpp1> Firefox ESR ?
2026-02-27 06:31:52 <RoadRoller01[m]> its because it uses opengl 
2026-02-27 06:32:04 <RoadRoller01[m]> my system doesn't like opengl
2026-02-27 06:33:02 <RoadRoller01[m]> while chromium uses vulkan 
2026-02-27 06:34:38 <RoadRoller01[m]> you might like this https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/98116 i wish it doesn't fails a lot 
2026-02-27 06:42:15 <jpp1> In any case, Chromium is crap
2026-02-27 06:44:54 <RoadRoller01[m]> 🤔why is that? is it because of gafam or you had a bad experience 
2026-02-27 06:46:24 <grawity> if you can do vulkan, could you run firefox on top of zink?
2026-02-27 06:50:34 <jpp1> I must have had a bad experience with the gafam that goes without saying.
2026-02-27 06:52:05 <jpp1> https://fr.wikipedia.org/wiki/GAFAM
2026-02-27 07:05:00 <RoadRoller01[m]> grawity: what is zink?
2026-02-27 07:06:11 <RoadRoller01[m]> ~> apk search zink
2026-02-27 07:06:11 <RoadRoller01[m]> ~>
2026-02-27 09:07:21 <Ermine> opengl is by no means laggy. And firefox is not an AAA game
2026-02-27 09:07:52 <Ermine> Check if it uses software rendering for some reason
2026-02-27 09:38:43 <f_> jpp1: as much as I don't like google, sometimes there's no other choice
2026-02-27 09:40:15 <f_> Chromium is my go-to for the time being on low-end devices (think old samsung tablet from 2011 with an Armv7 CPU and a GPU supporting at most OpenGL ES 2.0), simply because firefox performs extremely badly because it uses software rendering while Chromium still supports OpenGL ES 2.0
2026-02-27 09:40:43 <f_> firefox dropped support for OpenGL ES 2.0 back in 2021 which is why it uses software rendering on these devices
2026-02-27 09:43:06 <f_> This whole web browser story is just bad, apart from the big two, there isn't really any fully functional web browser from the last decade written from scratch... simply because the whole thing is so huge to implement it takes ages
2026-02-27 09:43:51 <f_> anyway IIRC Alpine's chromium package does remove some telemetry
2026-02-27 09:44:39 <f_> RoadRoller01[m]: oooooh ungoogled-chromium! Interesting!
2026-02-27 10:09:40 <courge_musque[m]> In Servo We Pray 🙏
2026-02-27 10:10:01 <courge_musque[m]> maybe they will achieve to overtake chromium in the browser engine land
2026-02-27 10:14:55 <hazardaj_nombroj> does anyone also have a problem, in which gnome shell shows up as "Not Responding" when trying to log out or power off, and you need to do a second "power off" for the operation to actually happen?
2026-02-27 10:17:08 <hazardaj_nombroj> seems like it's logged on the gnome bugtracker, but with little attention https://gitlab.gnome.org/GNOME/gnome-shell/-/issues/8885
2026-02-27 10:18:44 <Ermine> courge_musque[m]: from rendering perspective, Servo uses literally the same engine as Firefox, WebRender. So it can't use Vulkan or GLES
2026-02-27 10:21:04 <courge_musque[m]> Well designed opengl is better than poorly written vulkan
2026-02-27 10:53:09 <f_> courge_musque[m]: well not really if I can't even run it at a decent performance on my GLES2-only device.
2026-02-27 10:53:44 <f_> The renderer servo uses is exactly the same renderer Firefox uses
2026-02-27 10:54:09 <f_> so it will be as slow as Firefox on this kind of hardware for sure
2026-02-27 12:05:08 <courge_musque[m]> They may change the renderer in the future, as opengl won't be eternally supported by the hardware manufacturer
2026-02-27 12:08:59 <snchmt[m]> Hello, is there maintenance on the gitlab ?
2026-02-27 12:26:24 <vv221> Yet another reason why Firefox is a pile of crap: https://phabricator.services.mozilla.com/D279515
2026-02-27 12:26:32 <vv221> (sadly alternatives are even worse)
2026-02-27 13:12:38 <hazardaj_nombroj> huh, so gnome does not read ~/.profile when run in GDM on wayland
2026-02-27 13:12:56 <hazardaj_nombroj> apparently, one is supposed to instead set environment variables via ~/.config/environment.d/, but that's a systemd-specific setting
2026-02-27 13:14:47 <hazardaj_nombroj> it seems like it doesn't read /etc/profile.d/, either
2026-02-27 13:14:48 <vv221> I don’t think it’s specific to gnome / gdm / wayland, from what I remember most (all?) login managers ignore ~/.profile.
2026-02-27 13:17:22 <hazardaj_nombroj> which means the current config for gcr-ssh-agent is wrong
2026-02-27 13:17:33 <hazardaj_nombroj> it adds a file to /etc/profile.d/, which gnome then ignores
2026-02-27 13:29:01 <hazardaj_nombroj> so basically there's no way to set, say, a custom PATH right now that applies to the GNOME session
2026-02-27 13:31:03 <Biswa96[m]> can anyone open alpinelinux gitlab now?
2026-02-27 13:33:01 <vv221> git.alpinelinux.org is reactive, gitlab.alpinelinux.org is not.
2026-02-27 13:35:35 <fabricionaweb> it is down
2026-02-27 13:51:07 <ikke> Sorry, it's under load, and I don't have time right now 
2026-02-27 14:04:45 <hazardaj_nombroj> some suggest that ~/.pam_environment could work for GNOME, but it doesn't on my system, even though it clearly uses PAM
2026-02-27 14:04:50 <hazardaj_nombroj> is there a setting i'm missing?
2026-02-27 20:58:39 <RoadRoller01[m]> how to make abuild use sccache instead of ccache so i could compile chromium
2026-02-27 21:12:11 <Sertonix[m]> You can probably create a directory with symlinks to sccache and prepend it to PATH.
2026-02-27 22:13:01 <RoadRoller01[m]> idk if this would do a difference 
2026-02-27 22:13:04 <RoadRoller01[m]> but my device now is crying 
2026-02-28 06:56:34 <qaqland> how can I manually trigger mkinitfs.trigger? i want to use a custome kernel
2026-02-28 07:06:43 <ikke> simple way: `apk fix mkinitfs`
2026-02-28 09:26:36 <vv221> `apk fix $package` is a reinstall command?
2026-02-28 09:28:23 <ikke> yes, so running mkinitfs would be a side-effect of that
2026-02-28 09:28:38 <ikke> mkinitfs.trigger*
2026-02-28 09:33:27 <vv221> Thanks, I’m storing that into my mental rosetta stone.
2026-02-28 12:37:18 <sylvie> *taps mic* this thing on?
2026-02-28 12:38:40 <sylvie> yo uh i am trying to get wayland working, elogind is installed and running but when i run loginctl it says "No sessions."
2026-02-28 12:38:57 <sylvie> thats not a typo or anything, it does not say "No session available", it says "No sessions."
2026-02-28 12:40:06 <sylvie> i tried running `setup-devd udev` because i thought that might be the right thing to do but it didnt seem to do anything
2026-02-28 12:40:53 <sylvie> i did run setup-wayland prior to that and i am rebooting to be sure the changes applied
2026-02-28 12:40:59 <ikke> sylvie: what window manager?
2026-02-28 12:41:06 <sylvie> ikke: dwl
2026-02-28 12:41:51 <sylvie> oh yeah it also says XDG_RUNTIME_DIR needs to be set, seems like elogind should be responsible for setting that
2026-02-28 12:44:55 <sylvie> umm im not in group seat, does that matter?
2026-02-28 12:45:24 <ikke> sylvie: I start sway with `dbus-run-session swa`
2026-02-28 12:45:26 <ikke> sway*
2026-02-28 12:45:27 <sylvie> ahahahah `addgroup: unknown group seat` umm this doesnt seem good
2026-02-28 12:45:44 <sylvie> uhhhh am i not supposed to just ./dwl?
2026-02-28 12:46:57 <ikke> It's not my expertise, so I don't know what or what's not supposed to work
2026-02-28 12:47:49 <sylvie> seems like youre "supposed" to have dbus anyway
2026-02-28 12:51:23 <sylvie> installing dbus does not automagically set dxg_runtime_dir sob
2026-02-28 12:52:06 <sylvie> hold on progress: different error message
2026-02-28 12:55:34 <sylvie> different error message after installing util-linux-login and it actually looks like meaningful progress
2026-02-28 12:57:09 <sylvie> yo wtf i can run wayland through ssh?
2026-02-28 12:57:21 <sylvie> anyway seems like i just need to install mesa
2026-02-28 12:57:42 <sylvie> oh wait i was NOT sshed into the laptop
2026-02-28 13:03:07 <sylvie> ahah it uhhh worked i think
2026-02-28 13:03:16 <sylvie>  err: fcft.c:529: monospace:size=10:dpi=96.00: failed to match font
2026-02-28 13:03:16 <sylvie> Could not load font
2026-02-28 13:05:54 <sylvie> this is great
2026-02-28 13:07:44 <sylvie> ikke: what you said didnt end up being all that relevant but i think i would have given up without a rubber ducky so thanks for your involvement
2026-02-28 15:45:24 <vv221> OpenRC question: it seems services (system or user) are started/stopped in the lexicographic order, so I could tweak that order by using number prefixes in the links under /etc/runlevels and ~/.config/rc/runlevels). Would I break things by doing that?
2026-02-28 15:45:58 <vv221> Maybe there is another, recommended way to handle priority between services?
2026-02-28 15:46:24 <vv221> (between services that have no requirement relationship between each other, or that would be too easy)
2026-02-28 15:47:11 <lopid> https://github.com/OpenRC/openrc/blob/master/service-script-guide.md#the-depend-function
2026-02-28 15:47:42 <vv221> That’s what I mean by « no requirement relationship between each other » ;)
2026-02-28 15:48:23 <lopid> adding priority across them would be a relationship
2026-02-28 15:49:01 <lopid> "A before B" forms a relationship just like "B depends on A"
2026-02-28 15:49:11 <vv221> It’s not really a "priority" order, more like « I want the syslog to always start first and stop last » or « Start LXC containers after everyting else, but stop them first ».
2026-02-28 15:50:24 <vv221> Putting "after syslog" and "before lxc" in all my scripts does not seem like a good way to handle that.
2026-02-28 15:50:35 <lopid> nod
2026-02-28 15:50:47 <lopid> i think the local service might start last and stop first. could be wrong
2026-02-28 15:52:42 <vv221> Actually all of that would go away with parallel processing of start/stop actions. Can that be done with OpenRC?
2026-02-28 15:52:59 <lopid> i think not
2026-02-28 15:54:05 <vv221> It looks like I found an "interesting" problem once again ;)
2026-02-28 15:54:17 <vv221> Well, I’m going to experiment a bit.
2026-02-28 15:54:22 <lopid> rc_depend_strict=NO might go some way
2026-02-28 15:54:52 <vv221> Is that one documented in some man page?
2026-02-28 15:55:04 <lopid> /etc/rc.conf
2026-02-28 15:55:42 <vv221> Embedded documentation, my favourite, thanks!
2026-02-28 15:57:06 <vv221> Well, this one starts with rc_parallel="NO"
2026-02-28 15:57:25 <lopid> i have lead you to water
2026-02-28 15:57:29 <vv221> With a big "WARNING", but at least it’s there.
2026-02-28 18:17:58 <madtypewriter[m]> I installed Alpine on a Steam Deck, and it's mostly worked great, but I can't do HDMI CEC commands through the dock to my TV because /dev/cec0 because the the linux-lts config doesn't enable the driver. What's the most sensible way to get it? In theory I'd think there was some way to build just the 1 or 2 .ko file(s) I need from the same source as the official linux-lts package, and ideally I could even package that, but I can't
2026-02-28 18:17:58 <madtypewriter[m]> quite figure out how to actually do it.
2026-02-28 18:25:06 <madtypewriter[m]> Sorry, that should say "because /dev/cec0 doesn't exist because the the linux-lts config doesn't enable the driver"
2026-02-28 19:20:12 <Renan[m]> when Cosmic will get ported to alpine?
2026-02-28 19:20:58 <vv221> I assume this has nothing to do with Cosmic Cuttlefish?
2026-02-28 19:23:32 <abby> cosmic, the DE from system76 presumably
2026-02-28 19:24:25 <achill> Renan[m]: it is already in alpine
2026-02-28 19:24:56 <achill> https://wiki.alpinelinux.org/wiki/COSMIC
2026-02-28 19:25:38 <Renan[m]> achill: when? i dont see it
2026-02-28 19:26:09 <achill> i think it's currently just in alpine edge
2026-02-28 19:26:30 <Renan[m]> oh, i cant install with the alpine-interface
2026-02-28 19:26:46 <Renan[m]> achill: oh, i like the normal version
2026-02-28 19:26:47 <Renan[m]> i will wait to it
2026-02-28 19:26:52 <Renan[m]> now i am using popos
2026-02-28 19:26:57 <Renan[m]> cosmic is very good
2026-02-28 21:27:40 <Saijin_Naib[m]> Now we need cinnamon and pantheon