2026-06-01 00:50:45 Can somebody provide some insight regarding update_config_sub on #18225 2026-06-01 01:00:55 el[m]1: not sure about the cause, sorry 2026-06-01 01:09:11 Ok so it is definitely the GNOME section of the setup-desktop script that is broken 2026-06-01 01:09:40 I just did a manual installation of gnome and it worked. 2026-06-01 01:10:07 Although I will say that there is a lot of gnome packages missing doing it manually and some of them are not even loading right. 2026-06-01 01:10:20 Like gnome terminal for instance. that won't even start. 2026-06-01 01:16:49 Does anyone know who is in charge of gnome for alpine linux, or the setup-desktop script? 2026-06-01 01:19:03 team/gnome 2026-06-01 01:19:35 best is probably if you document your findings in an aports issue 2026-06-01 01:21:15 What is the best way to get in touch with them in real time? 2026-06-01 01:23:26 maybe check the alpine-devel channel? 2026-06-01 01:23:39 Is thatr bridged to matrix too? 2026-06-01 01:23:54 s/thatr/that/ 2026-06-01 01:23:57 guess so 2026-06-01 01:25:53 longnoserob[m]: Found it 2026-06-01 01:25:54 thank you 2026-06-01 01:32:48 good 2026-06-01 06:09:14 Hey folks. I'm noticing that my window manager gets a dbus session bus under /tmp due the dbus session.conf file (and being launched via dbus-run-session) and another dbus session bus gets created under /run/user due to the openrc user config file. It's causing some issues due to programs getting different addresses depending on how they are launched. Have I set this up wrong? 2026-06-01 06:21:17 I do not use dbus-run-session but tinydm 2026-06-01 06:42:28 I'm wondering if dbus-run-session is redundant with the user service creating one. It is a well known path so maybe I can have greetd provide it to the wm. 2026-06-01 06:45:11 i think so 2026-06-01 06:49:48 with the user service is redundant 2026-06-01 08:17:19 ptrckd: that's basically how it went in systemd – the user bus was originally for background stuff, but eventually having two became an issue, and the user bus became used as the one and only session bus (with the "one GUI session per UID" limit arising from that) 2026-06-01 08:18:25 I don't know how far openrc implements user services, but either way, once you have a user-wide bus, it's *probably* better to use that for everything, and dbus-run-session then becomes redundant 2026-06-01 10:04:18 WhyNotHugo: Maybe a missing cd thing. Will take a look later 2026-06-01 10:07:34 Okay, that worked. I used a wrapper script similar to what greetd says to do for sway and sourced the address from the openrc dbus user config. 2026-06-01 10:17:47 One other thing I stumbled on when cleaning up my neovim config... The neovim pkg symlinks the tree-sitter parsers from /usr/lib/tree-sitter such that neovim can find it but it doesn't do something similar for the tree-sitter queries in /usr/share/tree-sitter. Is that intentional? If not I'd be interested in trying to submit a fix if someone could give me some idea of what I'd need to do. 2026-06-01 10:18:45 There is an open issue describing the issues for that 2026-06-01 10:29:47 Thanks! I see #18217. 2026-06-01 11:18:21 uhhh 2026-06-01 11:18:24 im outside right now 2026-06-01 11:18:31 but if it hasnt been merged yeti 2026-06-01 11:18:43 please dont merge the arahli version bump 2026-06-01 11:18:56 theres an issue i spotted that ill need to fix when i get back home 2026-06-01 11:18:59 thanks! 2026-06-01 11:19:20 if it has been merged its fine too, just might be a hassle for reviewers to review another ME 2026-06-01 11:19:22 mr* 2026-06-01 11:44:13 gday, question: when is 2.24 planned? 2026-06-01 11:45:39 kopernic: ASAP. We're blocked by the riscv64 builder, which hasn't finished building all packages yet 2026-06-01 11:46:52 2.24? 2026-06-01 11:47:19 Oh, Alpine 3.24? 2026-06-01 11:47:43 (typo, yes, 3.24) 2026-06-01 11:56:48 i may have asked this before but is .24 moving to kernel 7.0 2026-06-01 11:57:12 IIRC 7.0 is not LTS? 2026-06-01 11:59:35 i thoufht there might have been an 7.x LTS, but i am apprehensive due to AI code ingestion 2026-06-01 12:01:33 I mean, after all Alpine uses Linux as the kernel 2026-06-01 12:01:43 If you get my point 2026-06-01 12:05:27 hey 2026-06-01 12:05:43 has anyone experience in launching an alpine install iso/image from kexec? 2026-06-01 12:06:20 I want to install alpine on my vserver using this, as I cannot upload custom ISOs there, similar to how one can do this with OpenBSD's bsd.rd 2026-06-01 12:06:26 cve: I don't but I know zfsbootmenu does it 2026-06-01 12:06:36 well, not images but it uses kexec 2026-06-01 12:07:59 https://wiki.alpinelinux.org/wiki/Kexec 2026-06-01 12:08:02 might be of use 2026-06-01 12:12:03 fwiw, the vserver is currently on NixoS 2026-06-01 12:25:32 cve: hope someone else can help you, I'm not too clued up on it 2026-06-01 13:07:58 cve: not kexec, but I have used extlinux to overwrite the /boot partition on a remote Debian VPS then boot from the virt iso using that. Similar processish 2026-06-01 13:08:51 https://wiki.alpinelinux.org/wiki/Replacing_non-Alpine_Linux_with_Alpine_remotely <- essentially what's outlined here 2026-06-01 13:29:50 durrendal: is it possible to wipe the boot partition during the process? like does it load the image fully into memory first? 2026-06-01 13:30:51 note that it's a sample size of like 2, but at least in my experience I was able to unmount my /boot partition and wipe/reinstall it as part of the process 2026-06-01 13:31:21 I'd recommend practicing on a VM if you can recreate a facsimile of your target environment 2026-06-01 13:33:50 """11:18:31 runxiyu_ │ but if it hasnt been merged yeti""" <<< and I really do not expect to get merged. far too buggy! ;-P 2026-06-01 15:35:38 does busybox tar invcrementaly update, eg if i backup 250GB dir to fil.tar every month? 2026-06-01 16:04:37 i'm on a postmarketOS system (hence feel free to send me away if this is too off-topic) and it seems like there's either an install user error on my behalf or a packaging issue with the out of tree drivers, and "akms build all" just hangs with no CPU load. is there some akms expert here who could help me extract a better error message so i can figure out if it's my fault or what might be going on? sorry again if this is the wrong room or 2026-06-01 16:04:37 not a useful thing to try 2026-06-01 17:31:56 hii 2026-06-01 17:32:18 hii dude is anyone here 2026-06-01 17:33:38 hello 2026-06-01 17:41:44 🫣 2026-06-01 18:01:58 dude, wheres my chan? 2026-06-01 18:06:43 el[m]1: postmarketOS doesn't really support akms 2026-06-01 18:12:43 el[m]1: perhaps try`akms build -k all`? 2026-06-01 18:13:32 or maybe `akms build -k all all` 2026-06-01 18:13:45 been a while since I used akms 2026-06-01 18:15:10 f_: i think this device was set up to use it, it uses the raspberry pi 5 alpine kernel. so it's not a pmOS kernel 2026-06-01 18:17:42 jvvv: i tried this now: `# akms build --kernel all all` but sadly it just hangs as well. no output, no cpu usage, no internet usage (i checked with nethogs) 2026-06-01 18:19:02 el[m]1: just to be sure, you did that as root or with doas/sudo/su ? 2026-06-01 18:19:41 ok, i now see the '# ', so guessing yes 2026-06-01 18:21:05 yeah it's running as root 2026-06-01 18:21:45 when i check ps aux, it kind of looks like it's just a shell doing nothing but perhaps i grepped the list poorly 2026-06-01 18:22:05 I wonder if namespace support is built into the kernel 2026-06-01 18:27:47 okay so it seems like `apk fix akms` actually finally said something: https://github.com/jirutka/akms/issues/22 and it says the AKMBUILD file is somehow broken. so i suggested here that the error message is improved 2026-06-01 18:27:55 jvvv: is there some way i could check? 2026-06-01 18:31:16 el[m]1: you could `ls /proc/self/ns` 2026-06-01 18:31:44 # cat /boot/config-6.12.85-0-rpi | grep USER_NS 2026-06-01 18:31:44 CONFIG_USER_NS=y 2026-06-01 18:31:44 # ls /proc/self/ns 2026-06-01 18:31:44 cgroup net time uts 2026-06-01 18:31:44 ipc pid time_for_children 2026-06-01 18:31:45 mnt pid_for_children user 2026-06-01 18:32:36 yeah, they are all there 2026-06-01 18:33:05 another think to double check is that overlayfs is available 2026-06-01 18:34:18 ACTION sent a code block: https://matrix.org/oftc/media/v1/media/download/AWi9FltjOPsGpIxO6Lp52PTrX-W3v2bmmgSbM_Gi3Jb70T2py_VMJ9CveTCb9eI8XevlhdwQsIaZAGLkUXw-tjRCeezJAg7AAG1hdHJpeC5vcmcvQ3VsdFBTeUN0VUxGRERXQ1NjVlhCanpu 2026-06-01 18:34:21 looks like it is 2026-06-01 18:35:12 lsmod | grep overlay ; ls /lib/modules/$(uname -r)/kernel/fs/overlayfs 2026-06-01 18:35:46 sorry, links like that are painful for me to try to reconstruct from copy/paste in irc 2026-06-01 18:39:17 # ls /lib/modules/$(uname -r)/kernel/fs/overlayfs 2026-06-01 18:39:17 overlay.ko.xz 2026-06-01 18:39:33 it looks like it's loaded. i didn't need to load it so i assume it was loaded from the start 2026-06-01 18:40:11 yeah, that shouldn't be it. just trying to eliminate low hanging fruit 2026-06-01 18:40:34 thank you, i appreciate it! 2026-06-01 18:42:26 I am looking at the scripts: /usr/sbin/akms and /usr/libexec/akms/akms-runas and /usr/libexec/akms/akms-build to see if there is a way turn on more verbosity 2026-06-01 18:44:05 akms has the --verbose (or -v) option 2026-06-01 18:45:28 that would be a start to try to see where it is hanging 2026-06-01 18:47:35 ``` 2026-06-01 18:47:35 # akms build -v all 2026-06-01 18:47:35 apparently it hangs immediately at the start, no output 😅 2026-06-01 18:47:35 ``` 2026-06-01 18:56:16 el[m]1: I think this may be an issue with akms. I just installed a -src pkg and akms hangs like you describe. I actually need this working in the near future so will be looking at this. 2026-06-01 18:56:45 oh you mean it's just generally broken on alpine edge? fascinating 2026-06-01 18:56:55 well i'm glad i reported something useful then 2026-06-01 18:57:26 if you want me to test out some experimental variant of it, ping me any time. i have good backups 2026-06-01 18:57:37 paging my ps output, it looks like there is something funky going on. 2026-06-01 18:58:09 yeah, cool, that would be good. I'll let you know if/when I narrow down the issue. 2026-06-01 18:59:00 so if you want something tested where it may render the machine unbootable or break the entire install beyond repair, i'm the right person to ask 2026-06-01 19:00:02 I don't think we will have to worry about breaking anything, but backups and recovery methods are always good 2026-06-01 19:35:42 el[m]1: see https://github.com/jirutka/akms/pull/21 ... that should fix it 2026-06-01 19:38:00 el[m]1: I will put in an MR for this shortly 2026-06-01 19:42:46 jvvv: neat! i think pmOSes mrtest can directly test out alpine merge requests, so i should be able to easily try it out that way 2026-06-01 19:52:58 mrtest in general is available on alpine, it's not a pmOS-only thing :) 2026-06-01 19:53:20 f_: I haven't gotten it to work on alpine (probably due to the child pipelines) 2026-06-01 19:53:59 it seems working fine for me if I add `-a` 2026-06-01 19:54:25 Yeah, I did add -a 2026-06-01 19:55:52 f_: oh oops i didnt know, lol 2026-06-01 19:57:59 ikke: i guess we'll see in a few minutes if i can get it to work on pmOS with alpine packages or if it's just generally broken with that 😮 2026-06-01 20:14:41 el[m]1: see !103348 2026-06-01 20:52:37 jvvv: ok it seems like mrtest cant do it: 2026-06-01 20:52:37 Download https://gitlab.alpinelinux.org/api/v4/projects/alpine%2Faports/merge_requests/441837 2026-06-01 20:52:37 # mrtest add -a 441837 2026-06-01 20:53:00 perhaps the best way to test would be for me to apply the fix in the shell script source code manually 2026-06-01 20:53:48 el[m]1: That was what I did initially to test the fix. 2026-06-01 20:55:27 can confirm it builds now. fix works for me. it's no longer stuck 2026-06-01 20:57:07 thank you for confirming 2026-06-01 20:57:33 the build finishes successfully! all the modules built! thank you so much for finding this fix 2026-06-01 20:58:03 you are very welcome 2026-06-01 21:06:42 btw mrtest worked fine for me: mrtest add -a 103348 2026-06-01 21:07:46 the number you used was the pipeline number i think. i used the MR number, then i just had to answer a bunch of questions 2026-06-01 21:09:33 maybe the pipeline would have worked also, but the one that you used 441837 failed becuase i had forgotten to fix up the checksums after adding the Patch-Source to the patch 2026-06-01 21:13:21 no, using the pipeline number failed also 2026-06-01 22:14:45 I'm seeing that backreferences don't work (or maybe have different syntax?) in Alpine's Busybox vs. Busybox on something like Ubuntu or Debian (I tried Alpine 3.2x so far). I'm definitely familiar with various RE features being disabled to save space or limit dependencies on various systems, but does anyone know (a) where I can find alpine-specific docs for the utilities it ships with, and (b) is this a disabled feature or does it have a 2026-06-01 22:14:45 different syntax? 2026-06-01 22:15:01 An example command is: echo 'A A' | sed -E 's!^(\w+) (\1)!matched \1 \1!g' 2026-06-01 22:15:27 (I meant to say, "backreferences in sed using extended regular expressions", sorry.) 2026-06-01 22:19:47 Where is the busybox documentation that explains that this extensions is supposed to work, and how? 2026-06-01 22:20:26 Also busybox sed non-conformant option to indicate it has to parse an ERE is -r 2026-06-01 22:23:37 Ah sorry, they have updated their inline documentation, but not their online one 2026-06-01 22:23:49 “-r,-E Use extended regex syntax” 2026-06-01 22:30:37 They gitweb is down too https://git.busybox.net/busybox/log/?h=1_38_stable 2026-06-01 22:40:21 quinq: I'm using the -E flag there. But I honestly don't know where the docs would be. It's easy to find man pages for generic utils, but I don't know where Alpine's are without installing them on a running distro every time. 2026-06-01 23:02:19 quinq: I generally go off this comment in the Busybox manual: "The utilities in BusyBox generally have fewer options than their full-featured GNU cousins; however, the options that are included provide the expected functionality and behave very much like their GNU counterparts." So running that command with GNU sed does the replacement. 2026-06-02 00:39:51 (It actually differs between eg. the busybox:1.37.0 docker image, and alpine:3.23 which uses busybox 1.37.0. So even the busybox docs wouldn't really help me here, I'd need the documentation for alpine's build options to busybox, or what features it uses.) 2026-06-02 01:05:51 Turns out backreferences in sed work just fine *without* the -E (and adjustments for syntax). So there's no feature disabled, it's just... different. I can't find anything in the APKBUILD that would account for that, either. Wild. 2026-06-02 03:42:35 they' tripin' 2026-06-02 05:15:27 My first guess would be to check what behaviour is expected by POSIX 2026-06-02 06:41:42 Sertonix[m]: I will groan inwardly and check the standard then. But what caught me out was that there's no change in "behaviour" as in, backreferences still work. It's only a change in syntax. And it works in the same version of busybox in busybox's docker image. 2026-06-02 06:42:04 I can understand behavioural changes from compilation options eg. disable backreferences; but I don't get why the same syntax would be accepted as valid but do something different. 2026-06-02 06:46:51 POSIX says the syntax should be accepted as an ERE: https://pubs.opengroup.org/onlinepubs/9799919799/basedefs/V1_chap09.html#tag_09_05_03 2026-06-02 06:49:40 Er, okay, it also says that backreferences are a BRE-only thing: https://pubs.opengroup.org/onlinepubs/9799919799/ 2026-06-02 06:50:36 So... does Alpine's implementation of Busybox patch this out of the upstream version? Where is that introduced? 2026-06-02 06:57:16 Rationale is here, too: https://pubs.opengroup.org/onlinepubs/9699919799/xrat/V4_xbd_chap09.html#tag_21_09_04 ("rejected by the standard developers as likely to decrease consensus"). 2026-06-02 06:57:49 I really wish there was some kind of "really truly POSIX-only" reference implementation of the standard that I could test against, instead of having to memorise the entire thing. 2026-06-02 11:55:44 I wonder if this could be a difference in glibc. 2026-06-02 23:46:20 anyone using Firefox on 3.23, i'd appreciate if you could upgrade (151.0.3 should be available any moment now on the mirrors) and test whether everything still works correctly 2026-06-03 01:20:39 prabu do you know if switching to limine from grub is documented, or have you done that yourself yet? 2026-06-03 02:31:58 I realise there's no wiki entry for limine, should have made one when I set it up long ago :( 2026-06-03 04:56:30 Ah, leshit 2026-06-03 04:56:50 I do not wish to footgun/brick more family computers testing how to transition from grub to limine 2026-06-03 04:58:23 bricking implies broken without feasibly recovery 2026-06-03 04:58:34 feasible* 2026-06-03 04:59:08 Soft-brick? 2026-06-03 04:59:15 Trash? 2026-06-03 05:07:30 Break boot? 2026-06-03 05:18:14 !user 2026-06-03 05:19:19 user? 2026-06-03 05:49:39 Saijin_Naib, it was not me who referred to liminine. i use refind. i did test limine, but it doesn't support my configuration. i place kernels/initrams inside /boot folder of each of my btrfs subvolumes(@,@rescue etc) and my efi partition is mounted at /boot/efi 2026-06-03 05:54:34 ptrc, i'm using firefox on 3.23 and the upgrade to 151.0.3 (64-bit) went smoothly. I don't see any regressions. Thanks for maintaining this. 2026-06-03 05:58:13 prabu thanks for confirming, and WhyNotHugo, same for you. I guess I'll dig and try when I have some time and a bootable recovery drive haha 2026-06-03 07:46:03 good day: alpine 64bit 3.23.4 curl 8.19.0 (x86_64-alpine-linux-musl) libcurl/8.19.0 OpenSSL/3.5.6 zlib/1.3.2 brotli/1.2.0 zstd/1.5.7 c-ares/1.34.6 libidn2/2.3.8 libpsl/0.21.5 nghttp2/1.69.0 no firewall settings (everything open) trying to curl to sycn.afraid.org is not working for us here, is this a c-ares issue we read about please or something else ? 2026-06-03 07:50:19 sorry sync.afraid.org - which curl resolves to ip 69.42.215.252 and then just hangs: curl -v -4 "http://sync.afraid.org/u/mytokenhere/address?u=anon&p=anon&d=anon.anon" - works fine on slackware if that helps 2026-06-03 07:51:17 mytokenhere - is whre a real token would go 2026-06-03 08:54:06 is anyone home that might be able to help my earlier posted question ? thank you 2026-06-03 10:05:04 habs have patience, people sleeping and stuff 2026-06-03 10:42:19 mozangle won't compile when building servo-fetch 2026-06-03 10:42:47 Can i fix it? https://paste.trom.tf/ketikukoji.sql 2026-06-03 11:32:13 habs: all I can guess is network issues 2026-06-03 11:41:27 f: no not network issues ... all other comms is fine, just via curl as described .... gnu curl is fine. any way dont worry ... apparently im too impatient so i will go elsewhere. peace and best wishes to all (sincerely) 2026-06-03 12:09:13 "gnu curl"... wget? 2026-06-03 17:27:50 we may never know. 2026-06-03 22:38:44 Hello all! I have a thing I'm stuck on. I'm trying to make a daily-driveable text-only system using linux 2026-06-03 22:39:56 i think you'll miss utf8 2026-06-03 22:40:08 Everything is going great, but I can't for the life of me figure out how to change the cursor from the underline (ideally to a full block, which is much more visibile). The console_codes man page had lots of helpful info about palette stuff, but doesn't mention cursor appearance at all. I'm willing to recompile the whole OS over this btw 2026-06-03 22:40:29 oh ig you're in insert mode 2026-06-03 22:40:50 nvm you just wanna customize it 2026-06-03 22:41:42 noodle: yeah I always want block cursor I think. Also I haven't missed utf8 so far, what would that be like emojis ig? 2026-06-03 22:42:10 yeah and other langs if you have to deal with em 2026-06-03 22:43:27 huh. That's kind of surprising for some reason 2026-06-03 22:47:38 i mean ascii stands for american standard code for information exchange :b first 127 codepoints do english and the remaining 127 in a byte can be packed with non-standard symbols like accented characters for support for some european languages 2026-06-03 22:47:53 but that's all, typically you'd swap charsets if you want more iirc 2026-06-03 22:48:13 s/exchange/interchange/ 2026-06-03 22:50:22 yeah, that explains why some of the system fonts have greek or cyrillic 2026-06-03 22:50:49 but smth like japanese or arabic just Can't fit in that encoding ig 2026-06-03 22:51:21 yeah 2026-06-03 22:52:06 i mean arabic has 28 characters, i bet representing all the ligatures would take lotas codepoints tho. idk. never tried it 2026-06-03 22:52:12 Even in western languages, the first time you need to talk about Mötley Crüe you're in trouble :P 2026-06-03 22:52:42 hm, your umlauts are showing up fine for me! 2026-06-03 22:52:51 👍 2026-06-03 22:53:02 I think that sent as a tofu, what's it supposed to be? 2026-06-03 22:53:11 thumbs up emoji 2026-06-03 22:53:51 Hello! 2026-06-03 22:53:51 ah. Yeah that must be what tofus look like on here haha 2026-06-03 22:56:20 Anyone have Samba server set up on their Alpine server? I'm having some trouble connecting from my Gentoo box and I'm 99% sure it's a server side issue. 2026-06-03 22:56:57 What makes you say that? 2026-06-03 22:58:02 It's always an "NT_STATUS_BAD_NETWORK_NAME" error message 2026-06-03 22:58:25 y1nz: i looked up the sequence for getting a blinking bloc at https://www.invisible-island.net/xterm/ctlseqs/ctlseqs.html and it don't work in my alpine tty, i guess TERM=linux is not vt520-capable? 2026-06-03 22:58:47 should be echo -ne '\e[5 q' 2026-06-03 22:58:56 sorry, 1 2026-06-03 23:00:42 This is the result from "testparm -s"# Global parameters 2026-06-03 23:00:42 [global] 2026-06-03 23:00:42 dos charset = cp866 2026-06-03 23:00:42 allow insecure wide links = Yes 2026-06-03 23:00:42 idmap config * : backend = tdb 2026-06-03 23:00:42 force user = 2026-06-03 23:00:44 [shared] 2026-06-03 23:00:46 path = /media/storage 2026-06-03 23:00:48 read only = No 2026-06-03 23:00:50 wide links = Yes 2026-06-03 23:02:20 noodle: yeah, I had found the same thing. Didn't work, which makes sense given it's not on the console_codes man page 2026-06-03 23:02:50 i guess 2026-06-03 23:04:06 o found it 2026-06-03 23:04:29 y1nz: from https://www.kernel.org/doc/html/latest/admin-guide/vga-softcursor.html 2026-06-03 23:05:01 weird but seems kernel console has it's own escape sequences for setting cursor shape 2026-06-03 23:08:31 I had also found that, and it also didn't work :( 2026-06-03 23:09:36 weird. it worked on my vm 2026-06-03 23:09:53 echo -e '\033[?6c' 2026-06-03 23:10:11 yeah, nothing! 2026-06-03 23:10:23 wild. maybe reset it first? 2026-06-03 23:10:28 reset it? 2026-06-03 23:10:53 ya like, i remember there being a sequence/command to reset a terminal to a default good state 2026-06-03 23:10:57 I mean I'm currently on the machine in question, also running mtm (to have multiple vts open) 2026-06-03 23:11:17 oh. idk wats that but try it first on a plain tty 2026-06-03 23:11:46 like i have the one i boot into before i start wayland 2026-06-03 23:11:47 here I have another machine to try on, just to sanity-check myself haha 2026-06-03 23:11:56 nice 2026-06-03 23:11:57 I also have this one tricked out with lightmode colors 8) 2026-06-03 23:12:05 dam 2026-06-03 23:12:07 so testing without that is a good idea 2026-06-03 23:12:09 too 2026-06-03 23:12:22 plan9 already comes with light mode by default :b 2026-06-03 23:12:27 mhm remove all customizations ig 2026-06-03 23:12:41 WAIT WTF it worked on my other machine?? 2026-06-03 23:12:44 to match my state. i just got a default uncustomized terminal 2026-06-03 23:12:46 oh 2026-06-03 23:12:50 nice :) 2026-06-03 23:12:57 uhh.. Hm. 2026-06-03 23:13:08 gotta eliminate possible stuff interfering with it now 2026-06-03 23:13:13 Might leave and see if this works in a freshly logged-in version of this terminal 2026-06-03 23:13:20 sounds good ya 2026-06-03 23:13:25 I bet it's my multiplexer hosing something up >.> 2026-06-03 23:13:33 see yall on the flip 2026-06-03 23:13:36 gl 2026-06-03 23:17:02 but 2026-06-03 23:17:05 wrong buffer sor 2026-06-03 23:19:59 noodle: OK im back. so it DOES work on this system, although weirdly then vim resets it to the stinky underline cursor. As does mtm..... 2026-06-03 23:20:19 y1nz: sad 2026-06-03 23:20:48 i think it's just too stupid to restore that bit of state when exiting? idk 2026-06-03 23:21:14 buut ya ig vim ppl would know 2026-06-03 23:21:25 nah, on entry! vim is the main place I want to have a more visible cursor haha 2026-06-03 23:21:33 o ya, maybe it's dropping it? 2026-06-03 23:21:39 this is at least good progress though haha. <3 2026-06-03 23:21:53 yea :) 2026-06-03 23:22:04 it seems like vi respects it, but vim does not 2026-06-03 23:22:05 what about vi? (it's superior) 2026-06-03 23:22:07 see? 2026-06-03 23:22:11 that's why it's superior 2026-06-03 23:22:15 hmmmmmmmm you make a strong case.... 2026-06-03 23:22:19 :b 2026-06-03 23:22:24 I just really would miss the pretty colors :( 2026-06-03 23:22:44 and my nice colorcolumn at 80....... 2026-06-03 23:23:01 i don't believe in line limits anymore 2026-06-03 23:23:10 i just make code shor if it's too long 2026-06-03 23:23:16 and ya syntax hilighting is hard to give up on 2026-06-03 23:23:21 took me a while 2026-06-03 23:24:07 huh, I've heard similar from other people. Do you think it was worth the switch? Like did you get back to the same speed of comprehension as you had with colors? 2026-06-03 23:24:11 alsois mtm like tmux? 2026-06-03 23:24:12 wats it do 2026-06-03 23:24:19 yeah 2026-06-03 23:24:34 i mean i don't remember, been years, but i think i make more readable code now 2026-06-03 23:24:50 yes! mtm is Minimal Terminal Multiplexer, suckless-recommended, ~1000 lines of C. I'm actually sorta messing with editing it for my own preferences rn lol 2026-06-03 23:24:57 without color cues you gotta focus more on good formatting / style 2026-06-03 23:25:08 oh fair 2026-06-03 23:25:20 I already am extremely staunch about my formatting haha 2026-06-03 23:25:42 you know what, maybe this is the opportunity. Perhaps I will give vi a shake 2026-06-03 23:26:15 yea. if you're already strict about formatting i think you won't miss syntax hilighting much 2026-06-03 23:26:27 but, gotta admit, text does melt into eachother a bit, especially on irc 2026-06-03 23:27:22 yeah, I *really* like my colors in other places lol 2026-06-03 23:27:32 mhm 2026-06-03 23:27:57 I wonder if I could get *some* colors into vi without having to deal with The Entirety of Vim... 2026-06-03 23:28:16 not sure u can do that 2026-06-03 23:28:26 Like coloring paired brackets or keywords 2026-06-03 23:30:06 noodle: can u resend me that link you had before? to the documentation of the cursor mess 2026-06-03 23:30:25 y1nz: the xterm one? 2026-06-03 23:30:36 https://www.invisible-island.net/xterm/ctlseqs/ctlseqs.html 2026-06-03 23:30:36 the one with the 3 parameters, that ended up working 2026-06-03 23:30:39 ya that 2026-06-03 23:31:26 o 3 parameters? 2026-06-03 23:32:06 ah no, the ohter 2026-06-03 23:32:34 that talked about the linux-term-specific control sequence, like ESC[?6c 2026-06-03 23:32:47 o 2026-06-03 23:32:52 it's in the backlog silly :3 2026-06-03 23:33:18 https://www.kernel.org/doc/html/latest/admin-guide/vga-softcursor.html 2026-06-03 23:33:41 i dont have the backlog bc im stupie D: 2026-06-03 23:33:50 aw :C 2026-06-03 23:34:24 do u just connect directly, or do u have a server do it somehow? 2026-06-03 23:34:28 like to save the backlogs 2026-06-03 23:35:03 *hands them a "-s oftc" to add to their catgirl commandline* 2026-06-03 23:35:11 oh 2026-06-03 23:35:33 i thought u meant ur client doesn't save backlog when you quit/rejoin. i don't use a bouncer 2026-06-03 23:36:11 also -l flag saves chat to disk in a format good for grepping 2026-06-03 23:36:23 wow yeah. It does have that option huh 2026-06-03 23:37:05 It should be illegal for a program to have more than like 5 command-line options >.> 2026-06-03 23:37:19 lmao. catgirl can use a config file 2026-06-03 23:37:34 I also hate config files 2026-06-03 23:38:46 u hate everything lol 2026-06-03 23:38:48 twinz 2026-06-03 23:39:12 Yeah that's why I like alpine and suckless and stuff. Less things = less to hate 2026-06-03 23:39:19 mhm 2026-06-03 23:39:39 I have the utopian dream to someday touch all of the code I regularly interact with 2026-06-03 23:40:27 you hate commandline options? here's how i connect to just oftc "ircrc -t '#some,#channels' irc.oftc.net | tee -a /mnt/mq/irc /tmp/oftc" 2026-06-03 23:42:06 also u can do it 2026-06-03 23:42:54 i choose simple software for the same reason. i think it's good to run software you know you can study easily 2026-06-03 23:43:15 I had used ii and lchat originally, but it was too cumbersome. It's a tricky balance 2026-06-03 23:43:31 ii has really cool concepts 2026-06-03 23:43:49 I'll def give ircrc a try. I'm mostly just intimidated bc I haven't gotten around to figuring out what tee does yet 2026-06-03 23:44:00 separating the server connection from the UI is neat 2026-06-03 23:44:29 and with a file ui. it's very unix-philosophy i think 2026-06-03 23:44:35 oh 2026-06-03 23:44:39 tee is just a T pipe 2026-06-03 23:44:57 water goes in one end and comes out 2 ends instead of a regular pipe 2026-06-03 23:45:36 2 or more, but that's the basic idea, the extra ends are the filenames you specify, otherwise it writes to stdout 2026-06-03 23:46:02 huh, what does the -a do? 2026-06-03 23:46:10 o it appends instead of truncating the file 2026-06-03 23:46:13 ah 2026-06-03 23:47:18 when I used lchat, I found it pretty cumbersome to keep track of the state of things. That was *very* early on though, maybe I'll hack on it more later 2026-06-03 23:47:26 it's on my list lol 2026-06-03 23:47:36 o yeah. keeping track of mentions was painful on it for me 2026-06-03 23:47:58 you need like an extra window with mentions on it 2026-06-03 23:48:22 but to pipe all dms/server/etc stuff to that window idk how to 2026-06-03 23:48:48 yeah lol, cumbersome 2026-06-03 23:49:11 there should be like, both a master pipe for catching hilights, and readable files for reading individual buffers 2026-06-03 23:49:14 but ya gl on it 2026-06-03 23:49:35 i want to add ii's concepts to ircrc sometime 2026-06-03 23:49:37 would be nice 2026-06-03 23:49:39 Tbh even something like just flattening the dir structure might have helped. Like channel1.nick-serv.in channel2.nick-serv.out etc all in one dir.. 2026-06-03 23:50:02 idk. i like the hierarchy 2026-06-03 23:50:35 makes it easier to find stuff without doing lotsa string comparison 2026-06-03 23:51:02 if it was all one dir, ls shows everything. Then you just grep whatever you're looking for 2026-06-03 23:51:15 but youll have to ls | grep 2026-06-03 23:51:17 which is bleh 2026-06-03 23:51:19 maybe that plus a mechanism for tracking mentions/updates... 2026-06-03 23:51:26 no need i think 2026-06-03 23:51:34 master output can be grepped/filtered 2026-06-03 23:51:47 whys that bleh? I love my ls | grep :C 2026-06-03 23:52:01 i wouldn't run that in a loop 2026-06-03 23:52:02 or somth 2026-06-03 23:52:09 only wait 2026-06-03 23:52:12 wait* 2026-06-03 23:52:16 wait** 2026-06-03 23:52:21 (waiting) 2026-06-03 23:52:23 i forgor what we're try do 2026-06-03 23:52:23 hm 2026-06-03 23:52:59 ya ig it might be fine bc i don't think you'd have to search for available buffers, the user will specify them themselves 2026-06-03 23:53:14 but, still, representing a namespace via dirs/files is more natural imo 2026-06-03 23:54:25 I often find that replacing '/'s in directory structures with '.'s is a pretty one-to-one substitution 2026-06-03 23:54:50 like semantically i mean lol 2026-06-03 23:56:50 but 2026-06-03 23:57:13 time to have all your files under / with .'s :3 2026-06-03 23:58:25 ya 2026-06-04 00:00:08 y1nz: I think if you want to reliably change something like the (default) cursor format you would need a program that processes all output. Could probably be done with a custom patch for mtm. 2026-06-04 00:03:38 Sertonix: yeah, I think that's the move here 2026-06-04 00:06:34 Anyway, i must go, thanks all for the helps <3 2026-06-04 05:23:50 what is alpine's plan once gitlab is 100% enshittified? 2026-06-04 05:27:59 100% sounds like a very high bar 2026-06-04 05:30:51 self-hosted forgejo sounds fine? 2026-06-04 05:30:56 quinq: that sounds like a challenge 2026-06-04 05:31:15 or the opposite ^^ 2026-06-04 05:31:26 Though yeah I'd like to know what Alpine wants from a forge 2026-06-04 05:31:40 runxiyu: I quite like forgejo but it's not me who needs to deal with it every day 2026-06-04 05:31:45 I feel for the devs 2026-06-04 05:31:53 Makes sense 2026-06-04 05:31:54 I mean 2026-06-04 05:32:04 GitLab is also much heavier on hardware and other resources, IIRC 2026-06-04 05:32:28 Though, of course there are a lot of things that could be improved with Forgejo etc such as better delta reuse when serving packs... 2026-06-04 05:32:31 Good and flexible CI important 2026-06-04 05:32:32 well, I guess that's more of Git's job 2026-06-04 05:33:16 One thing I miss as well in forgejo is nested projects 2026-06-04 05:33:24 (namespaces) 2026-06-04 05:34:14 good point yeah 2026-06-04 05:34:25 i mean, among the large ones, basically only gitlab does it properly iirc? 2026-06-04 05:34:29 (cgit doesn't count) 2026-06-04 05:34:36 That I'm aware of 2026-06-04 05:34:43 yeah 2026-06-04 05:35:36 I'm not sure, and I'd prefer a dev to correct me or something, but from what I can see it appears that Alpine doesn't need flexible/complex code review workflows for most cases? 2026-06-04 05:36:21 ikke: are namespaces different to organisations? 2026-06-04 05:36:31 rnkn: gitlab can nest groups 2026-06-04 05:36:47 afaik, you only have flat organizations in forgejo 2026-06-04 05:36:54 ie, all repos live in level 2 2026-06-04 05:37:38 if anyone has okular installed, can you try to open a PDF, highlight something with the yellow highlighter, click in the highlight to open the popup, type in some text, and save the PDF? 2026-06-04 05:37:43 seems to reliably crash okular for me 2026-06-04 05:37:50 currently rebuilding with -dbg 2026-06-04 05:38:06 runxiyu: I recall similar issues 2026-06-04 05:38:15 was a while ago already 2026-06-04 05:39:01 https://u.runxiyu.org/ce107f241562b91137e531099890fe2175305b4d140261e7d797e3c862527a22-screenshot_2026-06-04_05-38-35.png 2026-06-04 05:39:13 500 on gitlab when accessing my aports fork 2026-06-04 05:39:34 Can you copy that request id? 2026-06-04 05:39:46 01KT8HZM6EZ8BCG0JJR5DJRQV7 2026-06-04 05:39:48 thanks! 2026-06-04 05:42:04 gitaly timeout 2026-06-04 05:43:02 hmm 2026-06-04 05:44:18 Seems to work (now) for me 2026-06-04 05:46:46 gitlab or okular? 2026-06-04 05:46:51 well gitlab works for me too 2026-06-04 05:46:56 gitlab 2026-06-04 05:48:38 runxiyu: I can type in the popup 2026-06-04 05:49:02 save it 2026-06-04 05:49:08 save the file* 2026-06-04 05:49:41 something wrong with saveAs and kio causing gcc's stack clash detection to trigger... 2026-06-04 05:49:48 right, segfault 2026-06-04 05:50:02 maybe related to https://gitlab.alpinelinux.org/alpine/aports/-/work_items/16690? 2026-06-04 05:50:32 likely so 2026-06-04 05:53:54 alright, i think i know the issue now, writing up a patch 2026-06-04 05:54:29 ikke: apparently kio puts a 512kib object on the stack, and musl has small stacks 2026-06-04 05:55:31 ah yes, that would explain 2026-06-04 05:58:06 4-line patch 2026-06-04 05:58:07 testing 2026-06-04 05:59:26 [131/925] Building CXX object src/core/CMakeFiles/KF6KIOCore.dir/askuseractioninterface.cpp.o 2026-06-04 05:59:28 poor laptop 2026-06-04 06:03:24 rip 2026-06-04 07:16:26 ikke: my fix appears to work. opening an MR soon 2026-06-04 07:17:52 Cool, thanks 2026-06-04 07:18:50 ikke: should i bump the rel for okular in the same commit? 2026-06-04 07:21:04 Yes 2026-06-04 07:21:36 commitstyle seems to want one commit to patch kio, and another commit to bump okular's pkgrel, if im reading it right 2026-06-04 07:21:36 ikke: not really sure how to write the commit message 2026-06-04 07:22:11 oh, I misunderstood 2026-06-04 07:22:20 one commit per package, yes 2026-06-04 07:22:27 ah 2026-06-04 07:23:12 Does okular need to be rebuilt for this? 2026-06-04 07:24:04 actually, good point. im not sure 2026-06-04 07:24:43 if ABI stays the same, I assume not? 2026-06-04 07:24:54 yeah, i dont think there'll be any ABI changes 2026-06-04 07:24:55 (But this is not my expertise) 2026-06-04 07:28:01 I should probably submit it to KDE 2026-06-04 07:28:09 Yes, that would be good 2026-06-04 07:28:31 registering an account now 2026-06-04 07:35:03 wait they have a mailing list... 2026-06-04 07:35:07 anyways 2026-06-04 07:39:28 https://invent.kde.org/frameworks/kio/-/merge_requests/2257 2026-06-04 08:41:39 ikke: merged upstream! \o/ 2026-06-04 08:42:31 Nice, that was fast 2026-06-04 15:29:46 WhyNotHugothanks for the help with the MR, much appreciated! 2026-06-04 15:30:07 now I need to find time to actually do something with my ch32v boards 2026-06-04 16:01:24 if i wannt to send a patch for some aports, do i have to create a gitlab account and click a merge request, or can i also just send a patch to the maintainer? :o 2026-06-04 16:02:33 kroovy: You _can_ try to send a patch to the maintainer, but most would prefer a merge request 2026-06-04 16:05:47 i understand. it makes sense to me to only have *one* way of handling it. 2026-06-04 16:06:50 i'm just not a fan of clicking in gitlab :c 2026-06-04 16:07:00 glab-cli may be of help 2026-06-04 16:07:06 oha! 2026-06-04 16:07:08 (but you would need to have an account first) 2026-06-04 16:07:17 ikke: https://codeberg.org/forgejo/forgejo/issues/9550 tldr nesting doesn't seem likely to happen anytime soon. but i've seen people do stuff around multiple instances. 2026-06-04 16:07:28 invoked: Yeah, i've been looking through issues 2026-06-04 16:07:29 didn't now about glab-cli. thx :) 2026-06-04 16:07:37 :] 2026-06-04 16:07:46 I create all my MRs to aports using glab cli 2026-06-04 16:08:14 i can't imagine why you would not use glab, the webui is punishing 2026-06-04 16:10:52 I find it a bit pity forgejo and co focused on duplicating github 2026-06-04 16:11:58 ya. it don't even load as much pages as github witthout js ☠ 2026-06-04 16:15:57 btw hi ikke :3 2026-06-04 16:16:30 noodle: hi, though I don't remember in what context we have met before 2026-06-04 16:20:16 i vaguely remember dming a bit 2026-06-04 16:20:29 might've been someone else. srry if not 2026-06-04 16:21:04 Don't have any history, but no worry 2026-06-04 16:21:25 rad 2026-06-04 16:24:35 i should backup my logs better. i keep losing them when i recover from backups. rip 2026-06-04 16:25:18 is lightdm working for anyone in x86_64? 2026-06-04 16:33:51 hi all! In the light of https://byteiota.com/github-cli-cve-2026-48501-token-leak/ any idea whom could I ping to take a look at https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/103094 ? Thank you :-) 2026-06-04 16:45:30 alexaandru: do you perhaps wish to take over maintainership? 2026-06-04 16:51:48 Sure, I could do that. Alpine is my home now, and I like to keep my tools up to date anyway... 2026-06-04 16:55:54 Author indicates in their profile that they're busy and open for someone else to take over 2026-06-04 16:56:00 Maintainer* 2026-06-04 16:59:21 much appreciated! So what's the etiquette on taking over maintainership? Do I check with current maintainer or do I just submit a MR... ? 2026-06-04 16:59:43 You can make an MR, then we can ask the maintainer if it's okay 2026-06-04 17:02:17 sounds good. I see in the git log lots of "upgrade to X.Y.Z and take over maintainership"... I guess I'll do the same and couple that request with the next upgrade, less noise. You guys already have enough MRs as is :-) 2026-06-04 17:09:57 hi, how can i add a startup script that runs as root? i tried it in /etc/local.d/xdg.start but it doesnt work. It is a simple bash script from https://man.sr.ht/~kennylevinsen/seatd/#using-simple-scripts 2026-06-04 17:15:38 that should do. Did you make it executable? 2026-06-04 17:16:26 also make sure the local service is enabled 2026-06-05 04:48:07 invoked: reading through that issue I still can't see what benefit groups/projects represent, just organise repos cleanly 2026-06-05 04:50:11 like, the idea of re-architecting forgejo just to avoid setting repo permissions an extra couple of times? 2026-06-05 04:50:59 Consistency 2026-06-05 04:52:30 what are the things need to be kept consistent? 2026-06-05 04:53:07 permissions 2026-06-05 04:53:17 in gitlabs case, CI/CD variables 2026-06-05 04:55:41 those things can't just be set per-repo? I know it's probably annoying but compared to those in that issue asking to re-architect forgejo... 2026-06-05 04:57:34 Having to set them in each repo and keep in sync is annoying and leads to inconsisntency 2026-06-05 05:00:20 fair enough 2026-06-05 05:44:04 is the dump utility not in aports? 2026-06-05 05:44:12 https://linux.die.net/man/8/dump 2026-06-05 05:50:43 e2fsprogs-extra 2026-06-05 05:52:18 except in there the binary is 'dumpe2fs' because there are other dumps. the dump(8) you linked to is for ext2/3 2026-06-05 05:59:21 sdf 2026-06-05 06:26:58 invoked: ty 2026-06-05 06:42:41 okay dumpe2fs displays file system information, it's not the backup tool 2026-06-05 06:57:37 apk search cmd:restore turns up nothing too 2026-06-05 06:57:41 I think I am outta luck 2026-06-05 07:23:17 rnkn: oops 2026-06-05 07:23:30 i thought that was it, sorry. 2026-06-05 07:23:45 you can poke around with 'apk search cmd:dump' 2026-06-05 07:24:04 ACTION herding cats 2026-06-05 07:30:05 I think I can see why it's missing; from debian: "This is a port of the 4.4BSD filesystem backup suite." 2026-06-05 07:30:37 wondering if I should just use tar 2026-06-05 07:59:15 busybox tar does have the one-filesystem flag, hrmm 2026-06-05 07:59:22 doesn't* 2026-06-05 12:10:15 aelin: do you use selinux on Alpine? 2026-06-05 13:51:41 i have the weird issue on armhf pi0w where it cannot connect to either ssid of the extended home wifi network, but will connect to the a hotspot off my phone strangely enough 2026-06-05 13:51:55 because i yoloed the edge upgrade i did not record what upgraded 2026-06-05 13:52:05 but it was ~3 months ago to now 2026-06-05 14:02:02 penguinz_rule[m]: if you already had apk-tools v3, there is /var/log/apk.log (or something like that) 2026-06-05 14:05:36 will send pastebin of soon, thanks 2026-06-05 15:47:11 rnkn: if busybox tar is a hard requirement you could try using busybox find's -xdev flag in combination with it, which doesn't descend into directories of other filesystems 2026-06-05 17:15:33 penguinz_rule[m]: pis rule 2026-06-05 17:15:52 ? 2026-06-05 18:03:59 raspberry pi, i assume, though i did read it as piss rule at first 2026-06-06 14:34:14 internet resolved itself lol 2026-06-06 21:41:55 does anyone have an up to date guide for how to get yt-dlp working on alpine? 2026-06-06 21:42:15 since the last time I used it, it has started to require a javascript runtime 2026-06-06 21:42:58 I did apk add deno 2026-06-06 21:43:03 then pipx install --force 'yt-dlp[default]' 2026-06-06 21:43:48 but yt-dlp still claims "No supported JavaScript runtime could be found." 2026-06-06 21:47:11 i mean, yt-dlp 2026.03.17 is packaged 2026-06-06 21:53:00 I see 2026-06-06 21:53:27 can I get it without upgrading to 3.23 2026-06-06 22:09:01 venv probably 2026-06-06 22:09:05 unless python is too old on that machine 2026-06-06 22:09:29 actually nvm 2026-06-06 22:28:11 Should be feasible to backport yt-dlp-2026.03.17 to 3.22 2026-06-06 22:28:36 Or clone aports and `abuild -r` just that package if you're willing. 2026-06-06 22:29:15 Er, actually, don't do that, it might rebuild dependencies and get you into an ugly state. Backporting is the way. 2026-06-07 03:06:14 How can I change my nickname any help please 2026-06-07 03:07:28 Dicess: try /nick yournewnickhere 2026-06-07 03:09:22 Can I message you directly am lost 2026-06-07 03:10:05 Dicess: sure, #oftc can provide help too 2026-06-07 03:10:47 Can you join me to other good irc networks 2026-06-07 03:11:01 Pm first I can't do it myself thank you 2026-06-07 03:12:04 sure, pming 2026-06-07 03:12:52 Where am I going to see it 2026-06-07 03:14:00 no worries, found you 2026-06-07 11:40:16 Hello, I would like to replace GRUB with an EFI stub. I've created a new bootentry using efibootmgr, but when I reboot I still get show the GRUB menu before linux boots. I would've expected that this would skip GRUB, and directly boot vmlinuz. Maybe I've gotten it wrong and this is not to be expected, truthfully I don't have such deep understanding of the linux boot process. 2026-06-07 11:40:24 https://paste.sr.ht/~ludv/2875e7c8edda79c2177b2d10b8479719b20949ec 2026-06-07 12:04:28 stureplan: BootCurrent: 0002 indicates that it has booted from entry 0002, which is labelled EFISTUB. 2026-06-07 12:18:14 I am wondering if the '--loader /vmlinux-lts' should maybe be '--loader /boot/vmlinuz-lts'? Also, not sure if there should be some backslashes before the foreward slashes. 2026-06-07 12:20:16 I don't know the disk layout involved here, but my gpt layout is: part1=root part2=efi part3=swap . This puts /boot in the main root partition so there wouldn't be a /vmlinux-lts for me. 2026-06-07 12:35:08 jvvv: those paths are relative to the ESP (which I presume to be /boot in this case). And it's backslashes since EFI uses win/dos conventions. 2026-06-07 12:50:48 WhyNotHugo: Yes, I renamed "Alpine Linux" -> "Alpine Linux EFISTUB" between the two attempts, they are both EFI stubs, sorry for the confusion. Yes, it has has booted the correct bootentry (0002), but I don't understand why GRUB is still being executed. 2026-06-07 12:54:45 I tried adding an entry similar to stureplan, but adjust paths for my /boot directory being located on my root partition (initrd entrys and kernel). 2026-06-07 13:04:39 That entry came up but failed with notice that secureboot was not enabled. I enabled it and it failed anyways. Will have look into it. 2026-06-07 13:26:29 stureplan: I don't see any obvious indication either, sorry. 2026-06-07 13:43:07 anything up with the riscv64 builder? 2026-06-07 14:00:46 stureplan: Ok, I figured out what is going wrong for efistub. 2026-06-07 14:06:38 First, the kernel and initramfs images need to be on the efi system partition. Some efi implementations require the loader (the kernel) to have an .efi suffix. The 'initrd=...' portion of the commandline must have unescaped backslashes instead of foreward slashes: 'initrd=\EFI\initrd.img'. I copied my to kernel, initramfs, and ucode images to my /boot/efi/EFI and made sure they 2026-06-07 14:06:40 all met the old dos 8+3 filename req. 2026-06-07 14:08:26 The backslashes for the initrd= paths, if inside double quotes, need to be doubled (\\EFI\\initrd.img). 2026-06-07 14:09:40 Not sure the dos 8+3 filenaming rules are necessary or not, could depend on efi implementation. 2026-06-07 14:14:31 ninjin: did u work out ur stagit ideas? 2026-06-07 14:15:27 For my laptop, the kernel doesn't need to end in .efi and the initrd files did not need to follow the dos 8+3 rule. 2026-06-07 14:17:45 stureplan: After copying kernel + initrd images to /boot/efi/EFI, this was my efibootmgr cmdline: efibootmgr -c -L "efistub" -d /dev/nvme1n1 -p 1 -l '\EFI/vmlinuz-lts' -u 'root=UUID=be739bea-cb2f-4dbf-bb48-40fecba918b2 ro modules=sd-mod,usb-storage,ext4,nvme quiet rootfstype=ext4 initrd=\EFI\intel-ucode.img initrd=\EFI\initramfs-lts' 2026-06-07 14:19:53 Getting a setup like this to correctly get updated on kernel upgrades and ucode upgrades is TODO. 2026-06-07 15:12:45 Hi there :) I use a ThinkPad X220 and have been successfully using it with Bluetooth headphones but it just doesn't see my new Bluetooth mouse or keyboard. is there something I might be missing software wise or is it just that the X220's Bluetooth module is too old? 2026-06-07 15:14:09 Basically with bluetoothctl and bluetuith they just don't show up when scanning, other devices do though 2026-06-07 15:14:27 And the keyboard and mouse work fine via Bluetooth with other devices 2026-06-07 15:23:26 are they in pairing mode? 2026-06-07 15:23:36 Yup, definitely 2026-06-07 15:23:51 Other devices recognise them at the same time that the X220 doesn't 2026-06-07 15:39:37 bananicorn, could be, checking for compat should be pretty easy 2026-06-07 16:06:36 I thought Bluetooth was backwards compatible but I guess it's not 😅 2026-06-07 16:07:34 So software wise there is nothing specific needed for Bluetooth mouse/keyboard devices? 2026-06-07 16:29:29 jvvv: Thank you so much, it works now! I changed the slashes. My ESP is mounted at /boot, and instead of moving vmlinuz and intramfs to /boot/EFI, i just edited the command. I also noticed that I didn't have any ucode installed, so I installed amd-ucode and edited "intel-ucode.img" -> "amd-ucode.img" in the command. 2026-06-07 16:31:18 Might (aswell as the slashes) also have been the missing ucode that was causing trouble? 2026-06-07 21:11:13 stureplan: Glad you got it working. Not sure about the ucode causing failure... maybe. But good thing is we both learned some stuff today. 2026-06-08 18:23:54 Are GoAway blocks on the wiki temporary or permanent? 2026-06-08 18:38:16 WhyNotHugo: sorry, didn't see your message until now 2026-06-08 18:38:37 we're playing around with SELinux in postmarketOS using the tools packaged in Alpine 2026-06-08 18:38:41 (and the policy) 2026-06-08 19:08:33 aelin: I'm curious how you deal with the fact that labels need to be applied to package-supplied files. Are you applying the labels via packages themselves? 2026-06-08 20:06:23 i have a horrible apk hook that relabels the entire filesystem, a proper implementation would require some more elaborate integration in apk 2026-06-08 20:07:01 i.e. I would love to be able to relabel only added files, but hooks don't get those as a parameter 2026-06-08 20:08:02 see https://gitlab.postmarketos.org/postmarketos-selinux/pmaports/-/blob/main/extra-repos/selinux/postmarketos-base-selinux/apk-commit_hook 2026-06-08 22:35:41 aelin: mostly, we need to ensure that apk retains extended attributes and then tweak every single package to set rules for files it installs, right? 2026-06-08 23:17:09 WhyNotHugo: not really, the contexts for files are specified in the policy, we just need to apply them when files are created by apk 2026-06-08 23:17:17 see e.g. how rpm does this: https://github.com/rpm-software-management/rpm/blob/master/plugins/selinux.c 2026-06-08 23:18:52 aelin: I was thinking of a simple approach where the package declares context for files it provides, and then the administrator defines rules based on those policies. 2026-06-08 23:19:00 I guess your approach yields maximal flexibility tho. 2026-06-08 23:19:26 you do NOT want to write your entirely own policy, that's a mammoth task 2026-06-08 23:21:17 aelin: you can always have an options package for those who want all the upstream/usual rules. 2026-06-08 23:21:24 anyways, I'm happy to bikeshed integration for labeling based on file_contexts in apk 2026-06-08 23:21:27 I personally wanted only a few very focused/narrow rules (at least for now) 2026-06-08 23:23:47 aelin: tangential: I'm not a fan of the Fedora appraoch of having generated (and automatically overwritten) file in /etc/selinux: those sound like stuff that goes /var/lib, whereas /etc should only have the _real_ user-written configuration. 2026-06-08 23:24:02 yes, that's a known issue 2026-06-08 23:24:19 see also https://bugzilla.redhat.com/show_bug.cgi?id=1290659 2026-06-08 23:25:15 and the more general issue is tracked in https://github.com/uapi-group/specifications/issues/76 2026-06-08 23:29:02 aelin: the main limitation of your commit hook is that it re-labels every single file on every package update, right? triggers run in a chroot? 2026-06-08 23:32:04 yeah, my hook is basically just inefficient 2026-06-08 23:32:34 relabling the entire fs is horribly slow but for my current testing it's... good enough to get something working 2026-06-08 23:33:35 It's not the first time I see a trigger having issues due to running in a chroot. It's not really the chroot part that bothers here, but the sandboxing preventing changes to file attributes, right? 2026-06-08 23:34:08 the trigger has no issues really 2026-06-08 23:34:48 it's just that you'd be inclined to run restorecon -Rv / in the trigger, but that doesn't work if you're running the chroot on a kernel without SELinux 2026-06-08 23:34:52 setfiles just always works 2026-06-08 23:35:40 hook != trigger. You're using a hook, but does a trigger work? 2026-06-08 23:35:47 ah sorry 2026-06-08 23:35:56 haven't tried it actually 2026-06-08 23:36:06 I forgot why I made this decision 2026-06-08 23:38:52 iirc hooks are called before changes are synced to disk so I think this should be slightly more efficient 2026-06-08 23:39:29 but I guess you can probably make it work with both 2026-06-08 23:40:02 OTOH, triggers only operate on the newly installed files. 2026-06-08 23:40:27 but do triggers get a list of files as input? 2026-06-08 23:40:32 I think so 2026-06-08 23:40:56 Yeah, a bunch of triggers have `for i in "$@"; do`. 2026-06-08 23:41:34 What I don't know if configuraing a trigger for `/*` works. 2026-06-08 23:41:34 triggers probably make more sense then, but I'm not sure how the order of execution would then look like 2026-06-08 23:41:44 yeah iirc it gets the path that matched, right 2026-06-08 23:41:49 which would just be /* 2026-06-08 23:41:58 ACTION vaguely remembers things from a few weeks ago 2026-06-08 23:42:34 like, we would want to make sure that the contexts are set correctly before running arbitrary other scripts that may require them to be correct at execution time 2026-06-08 23:42:55 which is why this is probably just best suited to live inside apk itself 2026-06-08 23:43:57 i.e. some post-install script may invoke an executable that needs some label 2026-06-08 23:46:06 Honestly, I don't get why SELinux isn't just path-based, instead of needing to modify all file metadata to work 2026-06-08 23:46:15 Especially if the metadata is beign programatically updated based on paths 2026-06-08 23:47:32 there's a lot of reasons for that 2026-06-08 23:49:53 various things that come to mind: there's more than just files, users can just change the attributes on a file without having to modify their policy, it makes for a rather nice way to reach about things (*what* is this? rather than *where* is this?) and a lot more 2026-06-08 23:50:04 s/reach/reason/ 2026-06-08 23:52:20 Interesting. First time I get a really good answer to that :) 2026-06-08 23:55:43 not sure how much time you've spent staring at SELinux, if you haven't already I'd recommend watching https://www.youtube.com/watch?v=_WOKRaM-HI4 2026-06-08 23:56:27 aelin: saw that already, as recommended by you 2026-06-08 23:58:29 I don't want to pollute this chat more than necessary, if you want to help with the pmOS integration or e.g. trying to come up with a good solution in apk-tools I'd recommend joining #selinux:postmarketos.org, that's where we're currently fooling around and discussing implementation 2026-06-08 23:58:58 could probably get that bridged to IRC if f_ is around 2026-06-09 00:19:21 Woudl be nice to have a bridge, yes. 2026-06-09 05:30:41 aelin: What is the issue with having packages contain labels? 2026-06-09 05:49:22 you'd be reinventing the wheel, the policy already contains all the required information 2026-06-09 05:49:41 also means that you can't ship two different policies 2026-06-09 06:28:40 upgrade to VERSION_ID=3.24.0_rc2 2026-06-09 06:29:08 worked perfectly from v3.23. thanks to the team 2026-06-09 12:03:15 Sertonix[m]: the local administrator may have a policy overriding labels on some files 2026-06-09 12:03:39 The final labels to be applied to files are system policy, which often is just the default shipped by the distro, but may diverge. 2026-06-09 12:04:30 aelin: perhaps it's best to start with an issue in apk-tools collecting all the constraints for this 2026-06-09 12:28:11 aelin: Is there a way to have labels automatically applied to new files? 2026-06-09 15:38:55 so i've kinda hosed my alpine install (i can't open chromium/mpv bc of missing symbols in an .so) --- i'm on edge -- do you think i can uninstall pretty much everything and reinstall, or would i be better off just reinstalling alpine? 2026-06-09 15:39:02 does that question even make sense 2026-06-09 15:46:06 acdw: first, did you provide --available to the upgrade command? 2026-06-09 15:55:59 no 2026-06-09 16:05:04 i am now 2026-06-09 16:05:21 oh right i can't upgrade either bc of python 2026-06-09 16:59:25 on a CI, apk fails to add qbe 1.3 : https://gitlab.alpinelinux.org/StacyHarper/aports/-/jobs/2390016#L215 2026-06-09 16:59:51 does the revert of qbe, then the recent adding back produced this? 2026-06-09 17:00:09 maybe we should have used a -r1? 2026-06-09 18:20:24 Yes 2026-06-09 18:20:42 Or invalidate the cache 2026-06-09 18:21:01 -r1 is a great way to invalidate cache :) 2026-06-09 19:08:53 ack 2026-06-09 19:11:20 `repo-tools fastly purge pkg --repo community --origin qbe` is another great way 2026-06-09 19:31:19 is there a way to downgrade to the older linux-lts? apk add linux-lts=6.12.85-0 doesn't seem to do it 2026-06-09 19:31:47 el[m]1: you would need to have a repo that has that version 2026-06-09 19:31:59 We do not have the space to keep older versions of each package 2026-06-09 19:32:26 postmarketOS has some akms devices that no longer build, so it might be worth keeping the older linux-lts version (specifically the version before the one shipped today) at least for a few weeks 2026-06-09 19:33:06 so it would be useful to have at least one older version of the kernel available, if at all possible 2026-06-09 19:37:54 \o/ 2026-06-09 19:38:12 \o/ 2026-06-09 19:41:47 turns out i still had old copies of the kernel in my cache. so at least for myself right now i worked around it 2026-06-09 19:42:17 el[m]1: would be best to create a dedicated repo with that version then 2026-06-09 19:46:18 \o/ 2026-06-09 20:56:38 }o/ 2026-06-09 22:36:18 looks like some of the docker images didn't make it onto dockerhub; ppc64le, riscv64, and s390x in :latest are still from 3.23.4 2026-06-09 22:36:42 the 3.24 tag has 386, amd64, arm-v6, arm-v7, arm64-v8 2026-06-10 03:51:12 acdw, generally the issue you're facing might be caused by orphaned packages. These packages will prevent upgrading other packages due to resolver. So identify orphaned packages in your system and remove them first before you run upgrade again 2026-06-10 04:00:36 oh htanks! 2026-06-10 04:00:46 how can i identify orfphaned packages 2026-06-10 04:01:25 this issue is not properly captured in the https://wiki.alpinelinux.org/wiki/Alpine_Package_Keeper#Troubleshooting because there will be no error message. You'll see upgrades available, but can't upgrade. apk list --orphaned 2026-06-10 04:01:34 apk list --orphan 2026-06-10 04:03:43 apk list --orph 2026-06-10 04:14:31 when is openssl being updated on alpine? 2026-06-10 04:14:41 debian already shipped with an updated version 2026-06-10 04:20:43 Yeah, it's already been 16h 2026-06-10 04:41:14 bloink: merge requests are already open 2026-06-10 04:41:45 The timing is a bit unfortunate 2026-06-10 05:33:17 congrats on the new release :) 2026-06-10 07:38:26 i think i got the `grub-install ` invocation wrong as rebooting gave me:... (full message at ) 2026-06-10 07:38:46 (sorry, that's the invocation mentioned at: https://www.alpinelinux.org/posts/Alpine-3.24.0-released.html) 2026-06-10 08:02:35 I'm guessing packages for 3.24 aren't built yet? 2026-06-10 08:02:50 there's no perl in 3.24 branch 2026-06-10 08:19:16 as is the case with openssl, when a merge request is open what is missing from it to being deployed to users via apk upgrade 2026-06-10 08:19:35 i am a little new at this stuff so bear that in mind 2026-06-10 11:09:07 I'm not sure where to report this, but https://pkgs.alpinelinux.org/packages?name=php85-common&branch=v3.24&repo=&arch=x86_64&origin=&flagged=&maintainer= is not showing any packages. But other architectures are 2026-06-10 11:14:16 alpinemike5: it may still be importing the packages 2026-06-10 11:16:09 ok, cool... thanks for the info 2026-06-10 13:04:27 re: grub-install issue above -- used efibootmgr to learn that no entry existed for alpine. 2026-06-10 13:04:27 is that supposed to get created at some point during setup? 2026-06-10 13:04:27 anyway, creating one and adjusting the boot order seemed to be another "fix". 2026-06-10 13:26:47 oh i see: https://wiki.alpinelinux.org/wiki/UEFI#EFI_bootloaders 2026-06-10 13:28:47 it seems like the upgrade instructions could be improved slightly to mention the bit about putting a fallback bootloader into place? i see the code for it in setup-disk -- should have read more carefully. 2026-06-10 19:33:19 hello 2026-06-10 19:33:40 why is the first non-root user assigned to "/home/yes" -- is this a bug? 2026-06-10 19:34:51 ykrenet: can you elaborate?> 2026-06-10 19:35:16 did you name the first non-root user 'yes' ? 2026-06-10 19:35:16 Sounds like you answered 'yes' when it asked for a username? 2026-06-10 19:35:20 lol 2026-06-10 19:35:24 lol I didn't 2026-06-10 19:35:36 show their whole /etc/passwd entry 2026-06-10 19:35:46 the display-name shows up correctly with lightdm 2026-06-10 19:36:08 ykrenet: how did you install Alpine? 2026-06-10 19:36:09 ,/etc/passwd shows everything correctly except for the home-path 2026-06-10 19:37:17 setup-alpine was used 2026-06-10 19:40:35 hi : i just upgraded alpine to 3.24 but rspamd seems to fail to start 2026-06-10 19:40:53 lots of messages like those : "Error relocating /usr/bin/rspamd: fuzzy_parse_keypair: symbol not found" 2026-06-10 19:41:13 aren't those messages related to musl stuff ? 2026-06-10 19:41:36 that's just dynamic linking in general 2026-06-10 19:43:44 Lord: what's the output of `apk info --orphaned` and `apk version` 2026-06-10 19:43:44 i followed quickly after a question saying "no", so the prompt wasn't very clear -- turned out I must of typed it in. so there's no bug here thanks 2026-06-10 19:45:03 http://vpaste.net/jG1YG 2026-06-10 19:45:37 does not seem to load for me 2026-06-10 19:46:01 Ok, you have a bunch of packages that have not been updated 2026-06-10 19:46:24 i tried apk fix (does nothing) apk upgrade -a (does nothing) 2026-06-10 19:46:26 Lord: What does `apk add -s rspamd=4.2.4-r0` return? 2026-06-10 19:46:41 It's probably one of the orphaned packages 2026-06-10 19:46:55 you also still have linux-vanilla, so chance is you are running an ancient kernel 2026-06-10 19:47:03 rspamd-4.0.1-r0: 2026-06-10 19:47:05 breaks: world[rspamd=4.2.4-r0] 2026-06-10 19:47:07 satisfies: world[rspamd-controller] world[rspamd-fuzzy] world[rspamd-proxy] 2026-06-10 19:47:23 yes i'm on a custom kernel not installed from alpine 2026-06-10 19:47:39 ok 2026-06-10 19:47:59 Lord: is that the only output? I'd expect more output 2026-06-10 19:48:12 ERROR: unable to select packages: 2026-06-10 19:48:14 rspamd-4.0.1-r0: 2026-06-10 19:48:16 breaks: world[rspamd=4.2.4-r0] 2026-06-10 19:48:18 satisfies: world[rspamd-controller] world[rspamd-fuzzy] world[rspamd-proxy] 2026-06-10 19:48:49 It should also mention rspamd-4.2.4-r0: 2026-06-10 19:48:53 should i apk del rspamd then add it ? 2026-06-10 19:49:23 can you try: apk del rspamd-controller rspamd-fuzzy rspamd-proxy? 2026-06-10 19:51:34 yep, still fail to start, deleted rspamd too, then added but still fails 2026-06-10 19:52:57 why do you want to install 4.2.4-r0 ? in 3.24 the latest rspamd is 4.0.1-r0 2026-06-10 19:53:57 Oh, sorry, mistook the version from simdjson 2026-06-10 19:54:14 So can you repeat with =4.0.1-r0? 2026-06-10 19:55:04 ok it works 2026-06-10 19:55:36 but now it ditched my conf and complains it can't read its config file ^^ 2026-06-10 20:07:33 there is something i don't understand. i deleted rspamd. tried apk add rspamd, it creates it /etc/rspamd/ with every default files. it can't start with all the "symbol not found" 2026-06-10 20:08:00 what is the installed version of rspamd-libs 2026-06-10 20:08:24 (bet it's not 4.0.1) 2026-06-10 20:10:16 hmm 3.14 ok that's probably the roots of my problems 2026-06-10 20:10:42 rspamd 4.0.0 had some breaking changes: https://github.com/rspamd/rspamd/releases/tag/4.0.0, so the config issue could probably be sorted by looking at that page 2026-06-10 20:10:47 You have many orphaned packages, and some packages that are not upgraded 2026-06-10 20:10:53 something is holding those packages back 2026-06-10 20:11:04 usually orphaned packages 2026-06-10 20:13:32 asking apk to update a package the the version that is available generally reveals what is blocking it 2026-06-10 20:15:03 it looks like icu could be the culprit 2026-06-10 20:15:57 http://vpaste.net/2hFZI 2026-06-10 20:19:43 postgres15 2026-06-10 20:19:56 That's keeping the old version of icu installed 2026-06-10 20:20:28 yep just removed it and upgrade -a and some packages get upgraded :-d 2026-06-10 20:21:15 ok icu got upgrade, and now rspamd-libs got upgraded too and now rspamd starts 2026-06-10 20:21:28 guess i have to say goodbye to postgres15 :-( 2026-06-10 20:22:09 3.22 is the latest version that still had postgres15 2026-06-10 20:22:59 no more orphaned pkg :-) 2026-06-10 20:23:02 thanks a lot 2026-06-10 20:24:55 yw 2026-06-10 20:26:56 i'm used to gentoo's portage which is a bit more explicit why a package is held back 2026-06-10 20:45:23 yeah, portage explodes in output whenever anything goes wrong 2026-06-10 20:46:02 but it's usually somewhere buried in that output exactly why it exploded, and if you stare long enough you'll find it 2026-06-10 21:40:01 i see the openssl3.5.7 patch is merged, i genuinly ask myself why only risc64 is not available yet. (i don't wanna stress, i just want to understand) 2026-06-10 21:40:47 riscv is slow is probably the explanation 2026-06-10 21:42:20 i see https://xkcd.com/303/ 2026-06-10 21:42:35 :D 2026-06-10 21:48:17 oh you're right. it says at the bottom of the build log... 2026-06-10 21:48:31 x86_64 took 16minutes 2026-06-10 21:48:43 riscv took 58minutes (last version) 2026-06-10 21:48:46 https://build.alpinelinux.org/ shows it's currently building llvm18, which takes a long time 2026-06-10 21:49:04 oh cool 2026-06-10 21:49:25 damn, how cool is that 2026-06-10 21:50:18 the last build of llvm18 on riscv took 5 hours 15 minutes 2026-06-10 21:50:47 compared to 1 hour 3 minutes on x86_64 2026-06-10 21:50:52 so 4-5x slower 2026-06-10 21:51:14 yea, interesting! 2026-06-10 21:51:44 i checked https://build.alpinelinux.org/buildlogs/build-edge-riscv64/main/openssl/openssl-3.5.6-r0.log 2026-06-10 21:51:53 the live-view is cool, tho :D 2026-06-10 21:57:08 okay so it says it finished py3-phonenumbers 11 hours ago; if it's doing builds in parallel, that might make sense 2026-06-10 22:02:04 i guess this is not cross compiling (crosscompiling was decided against for its complexity/error-proneness?) 2026-06-10 22:02:36 correct, this is running on actual riscv hardware 2026-06-10 22:02:49 nice :D 2026-06-10 22:03:15 cross compiling tends to fall over a lot because upstreams don't properly think about it in their code/build systems 2026-06-10 22:03:32 i can imagine 2026-06-11 06:58:44 Hi, is there a way to have `vi` be provided by neovim? Like the update-alternatives system from Debian or the eselect vi system from Gentoo? 2026-06-11 07:02:16 ACTION uses nano :) 2026-06-11 07:14:53 just install neovim, imoen? 2026-06-11 07:16:17 alias vi=nvim 2026-06-11 07:17:43 longnoserob[m]: it is installed, but vi is still the one from busybox. 2026-06-11 07:23:09 there are a number of ways to make it so that when you type vi it calls to nvim, but alpine does not have an alternatives system 2026-06-11 07:23:26 jvvv: okay, thank you. 2026-06-11 07:23:56 if your shell supports aliases, you do like i said above and assign an alias 2026-06-11 07:24:27 I’d want it to be system-wide, so an alias won’t do. 2026-06-11 07:24:42 I’ll just remove the vi binary and install a symlink in its place. 2026-06-11 07:25:03 Before doing that I just wanted to make sure there wasn’t a proper way to do that. :) 2026-06-11 07:25:12 vi defaults to symlink to busybox 2026-06-11 07:25:34 Perfect then, it’s just a matter of redirecting the symlink. :D 2026-06-11 07:25:47 yep 2026-06-11 07:33:40 Alpine doesn't have a generic alternatives system, but for busybox applets there is 2026-06-11 07:34:05 As long as neovim provides a vi executable at the same path as busybox does, it will override it 2026-06-11 07:34:48 So perhaps having a neovim-vi subpackage that provides /usr/bin/vi and you can opt-in to it 2026-06-11 07:40:14 actually thats a nice idea, i would certainly use that over my symlink /usr/local/bin/vi -> /usr/bin/nvim rolled out by ansible 2026-06-11 09:42:36 Hello! 2026-06-11 09:42:37 https://gitlab.alpinelinux.org/alpine/infra/infra/-/work_items/10869 2026-06-11 09:47:55 el-brujo: Hey, I have seen the issues. We are just a bit in a busy period right now during the release 2026-06-11 10:17:00 ah ok @ikke thank you for your response 2026-06-11 18:07:05 upgraded my servers to 3.24 and everything went great :) although the binfmt service doesn't seem to be an adequate replacment of qemu-binfmt yet. 2026-06-11 18:20:49 Hi, first post, go easy :-)....rpi zero with ethernet hat (running pihole) alpine 3.23 but with edge repository enabled. Has been working great, but at some point did an upgrade which seems to have broken the ethernet on the hat. I had dtoverlay=dwc2,dr_mode=host in usercfg.txt. Looking at the kernel config, it says # CONFIG_USB_DWC2_HOST is not set - could this be the reason it stopped working? Any idea how to fix? 2026-06-11 18:26:24 tbh, i dont know for sure that this kernel config has changed, but it seems possible... 2026-06-11 18:28:36 I was just looking at the linux-rpi git history and it does not appear to have changed locally (i.e in Alpine's build files) so I am looking into upstreams' changes 2026-06-11 18:29:18 thankyou @jvvv 2026-06-11 18:53:25 toby: looking at the kernel configs going back as far as 3.23.0 release images, the config has not changed. CONFIG_USB_DWC2_HOST is not set in any of them. even on the 3.24.0 release, the _USB_DWC2* config items have stayed the same. 2026-06-11 18:54:05 toby: can you give me a manufacturer/model for the hat? 2026-06-11 18:56:08 also, looking in your dmesg log would be advisable. perhaps that could help shed some light 2026-06-11 19:00:33 while not ethernet related, this upstream issue looks like it might be related: https://github.com/raspberrypi/linux/issues/7282 2026-06-11 19:05:47 another possible culprit might be: https://github.com/raspberrypi/linux/issues/7349 2026-06-11 19:09:32 toby: what is output of `uname -r` 2026-06-11 19:12:42 its an early waveshare hat: https://thepihut.com/products/ethernet-and-usb-hub-hat-for-raspberry-pi - I think its that one. Thank you all for your help. I will follow up on your suggestions 2026-06-11 19:13:54 uname -r , 6.12.85-0-rpi, but that is since I upgraded 2026-06-11 19:17:44 huh, I wonder if 6.18.35-rpi fix you issue. There are several fixes related to DWC2 host mode that may address you troubles. 2026-06-11 19:18:07 s/you/your/g 2026-06-11 19:25:02 from personal experience from indirectly dealing with the rpi kernel at work: their 6.18 branch is much better kept up-to-date and receives fixes 2026-06-11 19:25:11 iirc they might drop 6.12 soon? don't quote me on that tho 2026-06-11 19:26:24 actually scratch that, i don't know. sleepy brain is kicking in 2026-06-11 19:28:43 well, I am not sure either, but the 6.18.y branch is seeing much more attention than 6.12.y, especially about the DWC2 driver code 2026-06-11 19:29:08 6.18.y is generally just seeing much more attention than 6.12.y, yeah 2026-06-11 19:31:02 two options I see: try upgrading to the 6.18 rpi kernel or downgrade to an earlier 6.12 rpi kernel that is not affected by the dwc2 woes 2026-06-11 19:31:52 i would probably try the first and switch to the second if that did not help 2026-06-11 19:34:23 since enabling edge repos on a stable installation isn't really supported, it might be good to upgrade to 3.24 and disable edge repos if you then don't need them anymore. this is what i'd do 2026-06-11 19:35:59 yes, this makes sense. I didnt know how to set up the repositories file at the time. I know slightly more now... 2026-06-11 19:39:51 I will return and post my findings - in case you are interested. I am amazed at the support from you guys. than you so much 2026-06-11 19:59:58 witcher01: thanks for clarifying that about edge and stable. I had actually forgotten that toby said about tracking edge. I was looking at it from the 3.24 perspective and agree with your suggestion. 2026-06-11 20:39:46 yay, openssl for riscv64 was finally finished baking :) i'm back with a freshly compiled irc client \o/ 2026-06-12 00:56:46 If anyone is feeling adventurous and would like to test out Alpine Linux on bare metal, or in a virtual machine with Network Manager. I could use some help with testing out a new piece of software. https://codeberg.org/infinitywisdom/Alpine-Linux-NM-Setup.git 2026-06-12 01:51:46 I just use iwd, it's much simpler for the "i want to connect to wifi and use Internet" use case. 2026-06-12 02:28:13 hey I've been using a new email address for awhile and want to change some of the aports I maintain over to it. Should I change in a separate commit for every package or in one for everything? 2026-06-12 02:40:28 how many packagess are that? 2026-06-12 03:23:46 23 2026-06-12 04:03:34 on mr with 23 commits, sounds doable 2026-06-12 07:00:52 hi all , We are bunch of Researcher , Come find Us for source code , research and phd social science at https://www.machophd.org/2026/06/12/elitez-0day-xc-our-news-skraito-here-god-husband-come-find-me-at-chat-live-for-source-code-os-researcher-chat-and-many-more-even-game-all-paternize-by-us-including-conver/ 2026-06-12 07:44:13 was there a "race-condition" with setting the color for riscv64? https://pkgs.alpinelinux.org/packages?name=openssl&branch=edge&repo=&arch=&origin=&flagged=&maintainer= 2026-06-12 07:44:25 CVE-2026-45447 was reported while the CVE-2026-45445+-fixed version was still building? :D 2026-06-12 07:44:41 s/+// 2026-06-12 08:02:13 what does "copy image" do in ff/librewolf? where is it stored? 2026-06-12 08:03:37 In your clipboard 2026-06-12 08:05:41 xsel doesnt seem to be able to output it tho 2026-06-12 08:06:11 What do you expect it to output? 2026-06-12 08:06:46 like cat file.jpg or something? 2026-06-12 08:07:21 it only works for gimp or similar? 2026-06-12 08:08:26 Images are stored with a specific mimetype 2026-06-12 08:11:13 copyq read image/png would cat the raw output 2026-06-12 08:11:51 Other clipboard tools probably also have a similar way of selecting what mime stream to use 2026-06-12 08:13:42 xsel only operates on the X selection buffer, and no copy button in an application (including ^X/^C) is going to put images into the X selection buffer. 2026-06-12 08:14:28 (and IMO, the X selection buffer should not exist.) 2026-06-12 08:15:52 dont have overview of the clip board/selection universe, but i mostly use PRIMARY selection, and i like it 2026-06-12 09:14:35 Is there a known issue with 3.24.0 on raspberry pi's causing no video after boot? v3.23.4 works fine. 2026-06-12 09:56:05 Sheila: it has always had an option to operate on the clipboard buffer too 2026-06-12 09:56:20 but it doesn't have anything like a --target option to request specific content type 2026-06-12 09:56:56 xclip does though, so it should be possible to `xclip -out -sel clip -target TARGETS` to see what types are available, and then `-target image/png` or whatever to request that type 2026-06-12 09:57:12 I think primary/secondary came from SunView originally 2026-06-12 09:57:33 SunView docs were literally the only place I found that actually described how secondary is supposed to work 2026-06-12 19:59:36 hello! i'm trying to find the wiki page on hosting an alpine mirror but so far nothing. does anyone have a link on hand? 2026-06-12 19:59:54 https://wiki.alpinelinux.org/wiki/How_to_setup_a_Alpine_Linux_mirror? 2026-06-12 20:00:07 thank you! 2026-06-12 20:02:35 ah, one issue: i'd rather not use rsync. is there an alternative? 2026-06-12 20:12:12 Use old rsync 2026-06-12 20:12:16 otherwise, no 2026-06-12 20:19:42 swirly, tar 2026-06-12 20:23:50 tar is not an alternative to rsync 2026-06-12 20:25:53 trying tar for that rn... coincidentally 2026-06-12 20:26:42 quinq: tar? how so? 2026-06-12 20:26:46 Yeah it is 2026-06-12 20:27:14 dwfreed: may do, though i will hear out quinq's tar idea first 2026-06-12 20:27:15 There's openrsync also, which implements most of rsync. 2026-06-12 20:27:32 used 4h36m to back up 215G from microsd to an external usb, using my rpi 2026-06-12 20:27:37 Well… tar your files and extract them wher you want to 2026-06-12 20:28:01 quinq: ...i meant for hosting an alpine mirror 2026-06-12 20:28:05 oh well, openrsync it is 2026-06-12 20:28:23 If it's for remote, use an ssh pipe, like tar -c dir | ssh 'cd /ext && tar -x' 2026-06-12 20:29:12 this is for an alpine mirror 2026-06-12 20:29:21 can't exactly do that when you don't control the other end 2026-06-12 20:29:52 Ah well, I didn't know it was about Alpine mirroring 2026-06-12 20:30:06 Indeed if it has a hard dependency on rsync… Then you have to use rsync 2026-06-12 20:30:17 you would have if you read up 4 lines :P 2026-06-12 20:30:45 (and that's still not an alternative to rsync's ability to only transmit changed files, and if you're willing to deal with the CPU cost, only changed *parts* of files) 2026-06-12 20:34:03 That's an alternative 2026-06-12 20:34:08 It just doesn't support *that* feature 2026-06-12 20:34:29 Otherwise that'd be an equivalent 2026-06-12 20:35:55 I suppose that's like saying walking is an alternative to driving :P 2026-06-12 20:36:11 oh wow, edge is huge... 2026-06-12 20:36:11 also the wiki page on hosting alpine mirrors has to be updated to include v3.24 in the disk usage 2026-06-12 20:37:05 chances are it's somewhere between 3.23 and edge 2026-06-12 20:38:04 of course, but its still something that's missing 2026-06-12 20:38:04 i'll add it once i can get openrsync to cooperate with the script in the page 2026-06-12 20:38:40 https://fs.swirly.moe/screenshot_k8JEXnDaVTgr2Ljk.png amazing 2026-06-12 20:39:43 such a tiny screen, jk 2026-06-12 20:40:50 does openrsync just not have --stats? 2026-06-12 20:43:17 Yeah dwfreed, it depends on the context 2026-06-12 20:43:33 walking is definitely an alternative to driving, on short distances 2026-06-12 20:43:54 Not so much for 10's of kilometers 2026-06-12 20:44:43 i like trains 2026-06-12 20:45:44 typically rsync deals with 10s of gigabytes or 10s of thousands of files (or both) for which tar is not a good alternative :P 2026-06-12 20:45:55 https://fs.swirly.moe/screenshot_CEH0W0D2koxOVhuM.png did it. included command line as well just in case someone else wanted to do something like this 2026-06-12 20:46:25 frag, me too 2026-06-12 20:49:14 https://fs.swirly.moe/screenshot_AZvvemqWLKHD0cWj.png "&2>" was a mistake. ctrl-c doesn't work... 2026-06-12 20:52:38 kill -9 :> 2026-06-12 20:52:50 yep 2026-06-12 20:53:20 "2> /dev/null" doesnt work either, inside nor outside of the script... bash is fun. 2026-06-12 20:53:43 oh well, its just a nitpick. i should actually set up my mirror 2026-06-12 20:54:34 openrsync would be better if you can use it 2026-06-12 20:55:04 youll never guess what im using 2026-06-12 20:55:07 it has some bugs 2026-06-12 20:55:23 but i've seen some activity on openbsd ml's fixing it recently so hopefully it works better soon 2026-06-12 20:56:47 accidentally obliterated my nginx config... amazing 2026-06-12 20:56:55 yeah. that's why i said "if you can use it" ... it might not work as expected 2026-06-12 20:57:39 invoked: fairs. thought you meant like, if they can run it on their system or somth. sorry 2026-06-12 20:58:33 no, my bad. i wasn't specific enough about what i meant. 2026-06-12 20:58:44 :) 2026-06-12 21:00:44 tar took 4h36m, rsync took 14m overwriting a 5month old backup, might still use tar tho, will see.. 2026-06-12 21:00:46 brb forgot some config 2026-06-12 21:00:51 one thing i miss from rsync is -q 2026-06-12 21:07:39 i use busybox tar tho, i guess the other tar has options to update or something... 2026-06-12 21:07:57 so tared the full 215GB 2026-06-12 21:11:01 'update' in tar is 'put a new file at the end with the same filename' 2026-06-12 21:13:27 (because tar does not differentiate between tapes and files, and you can't easily rewrite one part of a tape) 2026-06-12 21:15:35 seems to have:incremental backup, also, i guess that could be fast(?) 2026-06-12 21:18:04 that's a gnu extension 2026-06-12 21:18:16 which goes outside the tar file itself 2026-06-12 21:18:47 usually find solutions tho, maybe dont need all this crap, or maybe some of it is static or something 2026-06-12 21:22:37 https://mirrors.swirly.moe/alpine/edge/community/ its going 2026-06-12 21:22:47 \o/ 2026-06-12 21:23:08 itll only be edge for now, as this will mostly be for personal use and i prefer rolling. i may submit it for approval however, because who knows 2026-06-12 21:23:26 "approval" as in being added to the mirrorlist 2026-06-12 21:24:39 im surprised we support things like armhf and loongaarch64 but not ppc 2026-06-12 21:25:00 i mean i know we have ppc64le, i meant ppc32be 2026-06-12 21:26:39 oh wow are the irc logs updated in realtime? 2026-06-12 21:26:54 ok it takes like 2 seconds, thats expected 2026-06-12 21:26:58 i think im just bored 2026-06-12 21:27:05 ill begone for now 2026-06-12 21:31:26 https://fs.swirly.moe/screenshot_ZbwUSiXwaXHAigRA.png oh thats fun 2026-06-12 21:32:41 and my wifi has decided that now would be a fantastic time to crap out 2026-06-12 22:29:08 swirly: I done a ppc port but currently it's not updated: https://codeberg.org/Alpine-Ports/aports/pulls/1 2026-06-13 02:26:47 Hello. I see that Alpine 3.24 is released, but it's not available on https://www.alpinelinux.org/cloud yet. When it'll be available here? 2026-06-13 02:56:35 void join #alpine-cloud as they announce the cloud updates in there. 2026-06-13 03:13:47 Sertonix: better than nothing. will try on my ppc machine when i get the chance 2026-06-13 10:09:21 is there some process that modifies /etc/mkinitfs/mkinitfs.conf? 2026-06-13 10:09:21 one notable difference is that my existing version has `ide` in it for example. 2026-06-13 10:09:21 upon upgrading to 3.24 i noticed that a `.apk-new` version showed up and its content differs -- looks like it is close to (if not the same as) what's here: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/Makefile?ref_type=heads#L97 2026-06-13 10:10:23 if the file is modified by user, apk will create .apk-new file if it differs 2026-06-13 10:10:35 this happens during package upgrade/install 2026-06-13 10:11:16 i don't recall manually editing the file. 2026-06-13 10:11:17 i also tested installing 3.23.4 again and upgrading and the same thing appears to occur 2026-06-13 10:16:53 The install process might have created a file that slightly differs from the packaged one 2026-06-13 10:18:37 thanks, i'll try to look through the various alpine-* scripts. 2026-06-13 10:19:15 err, rather the setup-* scripts 2026-06-13 10:28:00 ah, looks like setup-disk may be the culprit. thanks a lot! 2026-06-13 13:51:09 sogaiu[m]: running update-conf will help with managing .apk-new files. The manual page for update-conf gives a good overview. 2026-06-13 14:07:11 Yes, but you generally don't want to use it for mkinitfs 2026-06-13 14:19:28 I think that is probably why, when I run update-conf, I end up zapping much of the changes. Those I don't zap, I usually use vimdiff on and bring in only the changes I think are pertinant to my setups. 2026-06-13 14:20:00 Yeah, same for me 2026-06-13 14:20:28 Using update-conf unjudiceously could easily lead to some real problems pretty quickly. 2026-06-13 14:20:47 spelling is wrong I think 2026-06-13 14:21:14 unjudicious 2026-06-13 14:25:01 My mkinitfs.conf is definitely one I have been careful with, but I also have edited it to tailor it my hardware and use cases. That has helped me to keep my initramfs images smaller. 2026-06-13 14:56:27 Hi, I noticed many issues with packages since the rollover to v3.24 that happen because some applications were build agains libs from v3.23. Is there any way to contribute / fix these issues if I've never done so before? 2026-06-13 14:57:58 lballardini: first make sure you do not have any packages held back due to dependencies 2026-06-13 14:58:20 3.24 has been built from scratch, it cannot have any dependencies from 3.23 2026-06-13 14:58:40 lballardini: apk list --orphaned; apk version 2026-06-13 15:00:09 lballardini: Also make sure you have provided --available to the upgrade command 2026-06-13 15:02:48 What I mean is that some packages stopped working after upgrading their dependencies, I think only some from community repos, like moonlight-qt for example, that were not updated for a bit. 2026-06-13 15:02:48 I don't have any orphaned ones installed and it was a full upgrade, with --available. 2026-06-13 15:03:31 Would I be able to contribute fresh builds of these applications that work with the new dependecies from 3.24? 2026-06-13 15:04:20 The blueman packages is another one which is broken at the moment for me 2026-06-13 15:11:09 lballardini: broken how? 2026-06-13 15:11:41 Error relocating /usr/lib/libQt6QuickControls2.so.6: _ZN20QFileSelectorPrivate13removeStaticsERK5QListI7QStringE: symbol not found 2026-06-13 15:11:43 and so on 2026-06-13 15:13:01 for moonlight 2026-06-13 15:19:33 Likely a missed rebuild when a -private-dev package was updated 2026-06-13 15:21:21 lballardini: Could you try removing moonlight-qt and reinstalling it? 2026-06-13 15:22:00 yep already tried that, did not work 2026-06-13 15:24:41 I think blueman might be the more important package for most people, because it's quasi standard for bluetooth applet in xfce4. 2026-06-13 15:24:41 There it's some python error: AttributeError: 'gi.repository.GLib' object has no attribute 'unix_signal_source_new' 2026-06-13 15:33:12 lballardini: Would you pastebin your /etc/apk/repositories? 2026-06-13 15:33:26 sure thing 2026-06-13 15:34:57 https://pastebin.com/bEwYVVEC 2026-06-13 15:39:26 Thanks. Yeah, nothing looks wrong with that. I've never used the 'latest-stable' path, always either a hard release or edge, but that shouldn't affect this. 2026-06-13 15:40:44 Yes that makes no difference for me, just checked again 2026-06-13 15:43:10 I do notice that qt6-qtdeclarative build is newer than the moonlight-qt build, so I am betting moonlight-qt needs a rev bump (quite likely other qt6 stuff may also). 2026-06-13 15:43:51 that was what I was thinking. 2026-06-13 15:43:51 Can I assist with that or do the maintainers need to get active? 2026-06-13 15:46:30 It would be good to confirm, though. Weird thing is I installed moonlight-qt on my edge box I am on now and it ran fine, older moonlight-qt build with newer qt6-qtdeclarative build, so some more due diligence would be best. 2026-06-13 15:47:56 I have an old laptop I was meaning to put 3.24 on anyways, I can do a fresh install on that and try to confirm issue. 2026-06-13 15:48:13 that would be appreciated 2026-06-13 15:48:22 I just noticed chromium is broken for me as well 2026-06-13 15:48:32 Error relocating /usr/lib/libavformat.so.62: av_exif_parse_buffer: symbol not found 2026-06-13 15:48:41 But I have some stuff I need to do right now, so it will have to wait for an hour or two. 2026-06-13 15:49:04 it's not urgent or anything, I just wanted to contribute 2026-06-13 15:49:12 I'm using the flatpak meanwhile 2026-06-13 15:50:22 I would guess that it's qt6-qtdeclarative that needs to be rebuild. And there could be differences between arches. 2026-06-13 15:50:46 I'm on x86_64 2026-06-13 15:51:18 A good possibility, Sertonix[m] makes a good point. 2026-06-13 15:53:48 lballardini: Could you try "apk policy qt6-qtdeclarative", just to be sure? 2026-06-13 15:54:21 qt6-qtdeclarative policy: 2026-06-13 15:54:21 6.11.1-r0: 2026-06-13 15:54:21 http://192.168.2.2/alpine/latest-stable/community 2026-06-13 15:54:21 lib/apk/db/installed 2026-06-13 15:55:29 Looks correct 2026-06-13 15:59:50 I don't seem to be able to reproduce the issue in a quick test VM. Is there any package version pinned in /etc/apk/world? 2026-06-13 16:00:57 nope 2026-06-13 16:01:48 Maybe output of "apk version -I"? 2026-06-13 16:03:57 v3.24.0-73-ge41cbd2ae99 [http://192.168.2.2/alpine/latest-stable/main] 2026-06-13 16:03:57 v3.24.0-70-gda4a9246e8f [http://192.168.2.2/alpine/latest-stable/community] 2026-06-13 16:05:53 The community repo does not seem to be up-to-date. Can you try to update it and then upgrade -a the packages again? (should show the same version as the main repo then) 2026-06-13 16:07:42 v3.24.0-73-ge41cbd2ae99 [http://dl-cdn.alpinelinux.org/alpine/v3.24/main] 2026-06-13 16:07:42 Didn't install any new packages / didn't change the issue 2026-06-13 16:07:42 v3.24.0-73-ge41cbd2ae99 [http://dl-cdn.alpinelinux.org/alpine/v3.24/community] 2026-06-13 16:07:42 But I did a quick test on a VM that was also updated yesterday and couldn't reproduce the issue there either 2026-06-13 16:07:47 weird 2026-06-13 16:10:50 Yes wierd. Could you try "strings /usr/lib/Qt6Core.so.6 | grep _ZN20QFileSelectorPrivate13removeStaticsERK5QListI7QStringE"? 2026-06-13 16:13:14 I noticed that every app using qt6-qtdeclarative has the same issue. 2026-06-13 16:13:14 I then removed qt6-declarative and every app depending on it and while trying to reinstall them I get: https://pastebin.com/avVAgQBW 2026-06-13 16:14:38 strings: '/usr/lib/Qt6Core.so.6': No such file 2026-06-13 16:14:38 After reinstalling just moonlight 2026-06-13 16:16:10 Then it sounds like some package is holding back upgrades. IIRC "apk upgrade -sl" can show more helpfull errors 2026-06-13 16:18:41 py3-requests and py3-cairosvg were the culprit, I was able to reinstall moonlight now and it is working fine again 2026-06-13 16:18:51 but why didn't it alert about this issue during the upgrade? 2026-06-13 16:18:58 initially I mean 2026-06-13 16:20:23 regardless, thanks for the help so far 2026-06-13 16:21:20 It is generally acceptable for apk-tools that not all packages are upgraded. In some cases it is useful/necesarry and in some cases it is not intended. I don't see any easy way to improve that. 2026-06-13 16:33:48 JK 2026-06-13 16:33:51 jjjj 2026-06-13 18:49:13 hi! i just upgraded from 3.23 to 3.24.1 and used the merge-usr script. no i can't elevate to root anymore using "su - " + root password since /bin/bbsuid appears to have lost its setuid bit. is this intentional? 2026-06-13 18:50:51 How did you invoke merge-usr ? 2026-06-13 18:51:20 on a root shell, first with --dryrun than without as suggested in the release notes 2026-06-13 18:52:17 So just the bare command? Hmm... There's a bug where that happens when merging sbin and bin, but that's not done on Alpine Linux (and is opt-in if you really wanted to break your system). 2026-06-13 18:52:33 Do other suid and sgid binaries have their bits? 2026-06-13 18:52:41 same happens with /bin/su from util-linux-login when installed, also loses setuid 2026-06-13 18:54:00 That's really strange. There hasn't been an update to merge-usr for three months, so I don't think anything changed recently. 2026-06-13 18:54:47 can't find any on the upgraded 3.24 system 2026-06-13 18:54:49 Thinking of the other suid binaries.... does your system have /usr/bin/unix_chkpwd, and is it sgid? 2026-06-13 18:54:53 its now very secure. too secure. :D 2026-06-13 18:55:17 wait, i'll check the other system that has util-linux-login installed 2026-06-13 18:55:44 $ ls -la /usr/sbin/unix_chkpwd 2026-06-13 18:55:45 -rwxr-sr-x 1 root shadow 22416 Dec 30 10:59 /usr/sbin/unix_chkpwd 2026-06-13 18:56:23 $ ls -la /bin/su 2026-06-13 18:56:23 -rwxr-xr-x 1 root root 47152 May 15 16:20 /bin/su 2026-06-13 18:56:24 That seems correct 2026-06-13 18:56:33 Okay, what in the world. 2026-06-13 18:56:51 Something must have gone horribly wrong, but I'm trying to figure out what could have triggered that. 2026-06-13 18:57:35 realpath /usr/sbin ? 2026-06-13 18:57:59 $ realpath /usr/sbin 2026-06-13 18:58:00 /usr/sbin 2026-06-13 18:58:19 Okay, so the sbin merge logic wasn't accidentally triggered. 2026-06-13 18:59:44 apk info merge-usr ? 2026-06-13 19:00:12 $ apk info merge-usr 2026-06-13 19:00:13 merge-usr-1.0.3_git20260218-r0 webpage: 2026-06-13 19:00:13 Script to produce a merged-usr system 2026-06-13 19:00:13 merge-usr-1.0.3_git20260218-r0 description: 2026-06-13 19:00:13 https://gitlab.alpinelinux.org/postmarketos/merge-usr 2026-06-13 19:00:15 merge-usr-1.0.3_git20260218-r0 installed size: 2026-06-13 19:00:17 18 KiB 2026-06-13 19:01:30 That's correct as well. You might be hitting (sorry but I have no idea how the bot works) https://gitlab.alpinelinux.org/postmarketOS/merge-usr/-/work_items/6 2026-06-13 19:01:56 But I've only ever seen that when debugging sbin merging downstream in postmarketOS, so its strange that this is coming up now. 2026-06-13 19:03:40 basically what i did: take a 3.23 system with latest patches applied, modify /etc/apk/repositories to 3.24, apk update, apk upgrade, apk add merge-usr, merge-usr, apk del merge-usr, reboot 2026-06-13 19:04:47 everything appears to be fine except the setuid bits 2026-06-13 19:05:38 Best I can guess is that maybe a package didn't update cleanly and that confused merge-usr, but I am unsure. If you have the root account, you can manually fix the suid binaries or move them from another system. 2026-06-13 19:05:56 (suid / sgid / stickybit) 2026-06-13 19:06:18 yes, i guess i'll just do that 2026-06-13 19:06:55 but people may get locked out of stuff because of this 2026-06-13 19:06:58 Sorry that this happened, but thank you for reporting it. I documented it on the issue just in-case it comes up again. 2026-06-13 19:07:00 might be worth mentioning 2026-06-13 19:07:13 alright :) 2026-06-13 19:07:36 merge-usr probably should just be dis-recommended on the wiki because its kind-of disputed right now if its the right path (to merging usr, not whether to merge it or not) 2026-06-13 19:33:47 oh i just noticed i sent nonsense messages, sorry about that. was working out keybinds 2026-06-13 19:44:45 yeah that's what I tell people the morning after 2026-06-13 19:47:59 Hi, quick question: If I am running a diskless + data partition installation, should /bin etc be symlinks after upgrading to 3.24? And if so, how do I debug why that's not the case for me? 2026-06-13 19:50:45 /bin is not a symlink by default 2026-06-13 19:51:58 ok so `merge-usr` script must be used for both disk and diskless installations? I figured that https://alpinelinux.org/posts/2025-10-01-usr-merge.html was a bit outdated since it mentions 3.23 instead of 3.24, but it's linked within the release notes of 3.24, so I was hoping that at least the rest is still up to date 2026-06-13 19:54:19 I'm not entirely sure (I'm not an Alpine dev), but I believe that disk and diskless installations use the same basic filesystem, so merge-usr would probably be used for both. I believe that blog meant that there were no special migration steps for diskless. 2026-06-13 19:56:12 btw, i can reproduce the merge-usr issue on alpine-virt-3.24.1-x86_64.iso, without even installing 2026-06-13 19:56:32 localhost:~# ls -la /bin/bbsuid 2026-06-13 19:56:32 (1/1) Installing merge-usr (1.0.3_git20260218-r0) 2026-06-13 19:56:32 ---s--x--x 1 root root 14224 Jan 10 15:38 /bin/bbsuid 2026-06-13 19:56:32 Executing busybox-1.37.0-r31.trigger 2026-06-13 19:56:32 localhost:~# apk add merge-usr 2026-06-13 19:56:34 OK: 10.5 MiB in 29 packages 2026-06-13 19:56:36 localhost:~# merge-usr 2026-06-13 19:56:39 INFO: configuration: root '' 2026-06-13 19:56:41 INFO: checking '/bin' 2026-06-13 19:56:44 INFO: checking '/sbin' 2026-06-13 19:56:44 well, I've understood the _"you do not need to do anything [...] correctly [...] with the links in place"_ as an "it should have /bin linked to /usr/bin without me running any command". Essentially diskless means there is no disk that would have any `/bin` folder or symlink, it's all assembled from the packages at boot 2026-06-13 19:56:46 INFO: checking '/lib' 2026-06-13 19:56:49 INFO: migrating '/bin' 2026-06-13 19:56:52 INFO: migrating '/sbin' 2026-06-13 19:56:54 INFO: migrating '/lib' 2026-06-13 19:56:57 INFO: creating symlink '/bin' -> 'usr/bin' 2026-06-13 19:56:59 INFO: creating symlink '/sbin' -> 'usr/sbin' 2026-06-13 19:57:02 INFO: creating symlink '/lib' -> 'usr/lib' 2026-06-13 19:57:04 localhost:~# ls -la /bin/bbsuid 2026-06-13 19:57:07 ---x--x--x 1 root root 14224 Jan 10 15:38 /bin/bbsuid 2026-06-13 19:57:33 Wow, thanks for the thorough testing! I'll note that in the issue as well. 2026-06-13 20:00:29 I'll see if I can step through the commits to find where this happened. 2026-06-13 20:04:58 Guest11299: please use a pastebin next time fyi 2026-06-13 20:05:04 like https://bpa.st/ 2026-06-13 20:19:13 or tpaste.us 2026-06-13 20:58:59 [Dominic](https://matrix.to/#/@msrd0:msrd0.de) Some diskless setups (when not using apkovl IIRC) need a kernel parameter to opt-into /usr symlinks 2026-06-13 23:21:15 Sertonix I see a .apkovl.tar.gz file in /media/vda3, so I guess I am using apkovl? 2026-06-14 02:44:02 ikke: (okular from a week or so ago) upgraded to kio 6.27.0 as packaged; seems fine to me now. can you test it? 2026-06-14 16:05:33 Hi, I'm not able to upgrade one of my alpine installations. I allready did 4 of them without any Probs. But this one show's me this: 2026-06-14 16:05:37 https://bpa.st/CRDA 2026-06-14 16:08:26 feuerteufel: what's the contents of /etc/apk/world? 2026-06-14 16:10:11 https://bpa.st/NAEQ 2026-06-14 16:10:30 Hope you mean this? 2026-06-14 16:11:33 Yeah, that was the right one. Nothing seems horribly wrong there. 2026-06-14 16:11:58 Maybe try upgrade with -V 2026-06-14 16:12:54 Give me a min. 2026-06-14 16:14:37 https://bpa.st/F74A 2026-06-14 16:15:13 It just printed one line? 2026-06-14 16:15:22 Yes 2026-06-14 16:15:51 You need a lowercase v 2026-06-14 16:16:03 capital is same as --version, lowercase is same as --verbose 2026-06-14 16:16:56 Oh, shoot. I read help and got them mixed up. My bad. 2026-06-14 16:17:20 https://bpa.st/2NCA 2026-06-14 16:17:41 Same as without option 2026-06-14 16:18:35 No prob. 2026-06-14 16:21:28 Does apk add still work? If so, maybe you can try apk-tools-static to see if your apk is broken somehow. 2026-06-14 16:23:20 https://bpa.st/4QMQ 2026-06-14 16:24:18 Use a command "apk-tools-static"? 2026-06-14 16:24:54 No, do apk add apk-tools-static and then instead of running apk, try apk.static. 2026-06-14 16:25:25 Ah 2026-06-14 16:25:28 If that fails too, then something is wrong with your system, not apk. 2026-06-14 16:26:49 But when I go back to 3.23 in /etc/apk/repositories it works 2026-06-14 16:30:42 https://bpa.st/ALPA 2026-06-14 16:31:08 I did some tests with apk.statik 2026-06-14 16:33:10 You can remove apk-tools-static since it didn't do anything. You can try playing wack-a-mole by removing packages in world and retrying upgrade until it works, then reinstalling the packages. 2026-06-14 16:37:01 feuerteufel: Do all your installations mix stable with edge? 2026-06-14 16:38:13 No, just one more and that is already done 2026-06-14 16:39:15 Ok. I would expect something is bound to break eventually, but that does not mean this is cause in this case. 2026-06-14 16:42:38 https://bpa.st/PW5A 2026-06-14 16:43:41 So You think uninstall those 5 packs and try again? 2026-06-14 16:45:06 I might help... some of those are in community now and may be the dependencies are futzing things 2026-06-14 16:46:00 s/I might/It might/ 2026-06-14 16:48:12 Can I download the database behind https://pkgs.alpinelinux.org/contents ? 2026-06-14 16:48:22 I want to make queries which the web UI does't support 2026-06-14 16:50:53 jvvv: I checked in pkgs and it looks like they are all in community 2026-06-14 16:51:07 feuerteufel: Also, if it was me, I would comment out the testing repo line in /etc/apk/repositories just to see if that helps. 2026-06-14 16:51:31 So. yeah, if those are your only packages from testing, you currently have no need for it. 2026-06-14 16:52:33 I did that already, then he says that a upgrade is not possible 2026-06-14 16:53:27 Weird, must be some dependencies issues going on. 2026-06-14 16:53:42 But I'll try to delete them then comment out testing 2026-06-14 16:58:34 it'll take a while, have to some other thinks first. Hope I'm done in half an hour 2026-06-14 16:58:59 No worries 2026-06-14 17:25:12 oh my god im finally done doing the alpine mirror stuff 2026-06-14 17:25:33 https://mirrors.swirly.moe/alpine/ only edge for now but like. its there!!! 2026-06-14 17:26:16 and sure enough 404s bring you to the main alpine mirror... amazing 2026-06-14 17:27:26 i should ask, to any mirror maintainers here, how often do you update your mirrors? 2026-06-14 17:27:39 looking for a good baseline, atm ive got a cronjob doing it daily 2026-06-14 17:39:29 swirly: the mirrors are updated shortly after packages have been uploaded, so not a set period 2026-06-14 17:45:10 jvvv: I think, I've got it! 2026-06-14 17:46:09 https://bpa.st/YSTQ 2026-06-14 17:47:45 It looks like hplip makes the trouble 2026-06-14 17:50:22 Well, actually, I think it is python2-2.7.18-r0 2026-06-14 17:52:15 Still, you are getting closer to having it all sorted. Progress is good I think 2026-06-14 17:53:31 Thats all without testing 2026-06-14 17:54:11 Do you have anything that actually depends on python2? 2026-06-14 17:54:17 Do you think I've to enable testing? 2026-06-14 17:55:13 Well, you still have the python2 issues... python2 is holding back libffi, which holding back other stuff 2026-06-14 17:55:34 I'm not quite sure 2026-06-14 17:56:38 'apk info -r python2' for a start 2026-06-14 17:56:58 looking for reverse dependencies 2026-06-14 18:00:24 That's one of my oldest Alpine installations. I think it comes from 2.x 2026-06-14 18:00:33 https://bpa.st/ZUCA 2026-06-14 18:01:58 According to that nothing is actually using it. I would remove python2{,-dev} then try installing hplip. 2026-06-14 18:02:08 I think, I'll try to uninstall all of that 2026-06-14 18:02:51 If that fails for conflict with libffi, I would do another 'apk upgrade -a'. 2026-06-14 18:07:35 https://bpa.st/TNOQ 2026-06-14 18:08:06 I had to uninstall python-dev also 2026-06-14 18:08:49 And now everything runs without testing 2026-06-14 18:09:07 I think you have it sorted 2026-06-14 18:09:15 Good job 2026-06-14 18:09:20 jvvv: Thanks a lot! 2026-06-14 18:09:29 welcome 2026-06-14 18:10:12 I'll see if everything else is running 2026-06-14 18:10:57 I should have asked you to do a 'apk list --orphaned' at some point earlier. 2026-06-14 18:11:34 I do that for my self when I have upgrade conflicts, but often forget when trying to help. 2026-06-14 18:12:29 It does what? 2026-06-14 18:12:58 It will show you packages that are no longer in any of your configured repos 2026-06-14 18:13:53 Orphaned packages are a common depenceny gremlin 2026-06-14 18:15:40 Ah ... Thank's one more for my Howto ;) 2026-06-14 18:15:40 s/depenceny/dependency/ 2026-06-14 18:47:37 3 more done, 2 more to go thanks again jvvv 2026-06-14 18:53:20 feuerteufel: You are very welcome. Glad you got it sorted. 2026-06-15 02:04:20 what package offers the GTK2 module "appmenu-gtk-module" ? 2026-06-15 02:43:58 swee: gtk2 is no longer packaged 2026-06-15 02:44:06 on alpine 2026-06-15 02:45:16 dang, so no hexchat? 2026-06-15 02:45:41 nope 2026-06-15 02:48:05 also, you might take a look at https://hexchat.github.io/news/2.16.2.html 2026-06-15 02:50:07 ACTION is aware hexchat is unmaintained 2026-06-15 02:53:01 ah, well, I feel for you. I never used hexchat, but there are programs I rather not have to find an alternative for. 2026-06-15 02:53:34 s/I rather/I would rather/ 2026-06-15 11:50:40 can someone please update the librewolf package? the browser console seems to be broken in the current version in alpine and there have been two bugfix releases since then (I'd do it, but building it would probably take days and I'm not doing that) 2026-06-15 11:51:16 WhyNotHugo: ^ 2026-06-15 12:58:11 Noisytoot1: I don't see any issues in the console with 151.0.2_p1 2026-06-15 12:58:22 Just rebased !103976 anyway, if you see issues after the update, please try with a clean profile. 2026-06-15 13:43:21 On a new alpine linux installation, based on edge, xorg does not see-or-search for any input devices, leaving no mouse/keyboard 2026-06-15 13:43:23 Same for regular user as is for root 2026-06-15 13:43:47 /dev/input/ contains a lot of devices, udev is started 2026-06-15 13:44:07 no hwdb or similiar dstuff installed? 2026-06-15 13:45:23 longnoserob[m]: No that I know of, it's a brand new installation 2026-06-15 13:46:09 I got libinput + xf86-input-libinput 2026-06-15 13:46:37 Good day, folks. I'm experiencing a number of dbus related issues on a fresh installation, having followed the standard setup process (sys), then proceeding with the prerequisites for, and finally the desktop setup process. Init claims dbus is running, but no process can connect to it. Any ideas where I should look first? 2026-06-15 13:47:16 what are you using this old X stuff anyway , nico🇨🇭 😝 2026-06-15 13:47:40 right... 2026-06-15 13:47:49 just joking 2026-06-15 13:48:24 what do the Xorg logs tell us? 2026-06-15 13:49:29 what's the current preferred paste-service? 2026-06-15 13:50:06 is there a ci or a script on how the isos, that are available for download, are created? 2026-06-15 13:50:22 https://pastebin.com/Vq52FbJT 2026-06-15 13:52:21 ivanpollak: https://gitlab.alpinelinux.org/alpine/aports/-/tree/master/scripts 2026-06-15 13:52:25 They are located there 2026-06-15 13:56:56 On a follow-up to my previous message: additionally, polkit fails with exit code 1. Running interactively brings up permission related issues on /var/empty 2026-06-15 13:58:50 ikke: thanks! 2026-06-15 13:58:54 nico🇨🇭: I woudl try to `apk add pciutils` and see if that chanegs anything 2026-06-15 13:59:32 I just restarted udev and at leas the keyboard started working. If I see it correctly, mdev should be disabled if udev is enabled, correct? 2026-06-15 13:59:41 ikke: so the official builds are generated using mkimage? 2026-06-15 13:59:57 yes 2026-06-15 14:00:01 telmich: not sure 2026-06-15 14:00:57 ikke: how about cross-compilation? i have an x86_64 machine and would like to generate a aarch64 iso. Is this possible too? 2026-06-15 14:01:35 cross-configuration (i guess)* 2026-06-15 14:01:57 There is no official support as far as I know 2026-06-15 14:02:45 so those isos should rather be generated using the underlying tools instead of mkimmage? 2026-06-15 14:05:15 Oh, there is support 2026-06-15 14:05:18 --arch 2026-06-15 14:05:44 just saw it. Ill try. Thanks! 2026-06-15 14:06:21 funny notebook 2026-06-15 14:06:32 Does not reboot / hangs after requesting system reboot 2026-06-15 14:12:48 mouse works after removing the synaptics driver 2026-06-15 14:13:17 ..aside from clicking... 2026-06-15 14:17:16 WhyNotHugo: the ctrl-shift-j browser console? 2026-06-15 14:36:43 Noisytoot1: I didn't know of that one. Yeah, works fine here. 2026-06-15 18:06:10 swee: https://github.com/ZoiteChat/zoitechat is hexchat based, maintained and uses Gtk3 2026-06-15 19:11:34 update on my weird issue from earlier: / had 700... disregard all my head-bashing... 2026-06-15 19:40:46 WhyNotHugo: weird. for me it shows error logs but I can't type anything in there, and it's the same in 151.0.2-1 in guix and in a new empty profile 2026-06-15 19:43:17 NotFoundError: No such JSProcessActor ‘BrowserToolboxDevToolsProcess’ 2026-06-15 19:52:32 WhyNotHugo: Apparently the issue was that I had devtools.chrome.enabled set to false (it definitely used to be true, something must've unset it) 2026-06-15 22:57:55 Noisytoot: glad that it's fixed :) 2026-06-16 02:23:24 curious as to the state of ai in contribution guidelines/code of conduct and in alpine in general 2026-06-16 03:41:17 https://gitlab.alpinelinux.org/alpine/council/-/work_items/697 2026-06-16 03:41:17 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/103822 2026-06-16 03:41:17 so...still processing? 2026-06-16 07:14:03 sogaiu[m]: there are some disagreements, especially thoes unrelated to AI. e.g. the way to coordinate this kind of problem. 2026-06-16 07:30:18 thanks for the summary 2026-06-16 07:42:05 hello, is there any reasons why xlibre package was rejected from adding into alpine linux package? 2026-06-16 07:43:24 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/86092#note_520485 2026-06-16 07:44:15 wait a second. 2026-06-16 07:45:12 Ariadne Conill is a trans woman who rejected the inclusion of xlibre package? 2026-06-16 07:46:45 😭😭? 2026-06-16 07:47:19 You're on very thin ice there, watch out with what you say 2026-06-16 08:36:02 geez, it's even worse: https://thelibre.news/lets-talk-about-xlibre/ 2026-06-16 09:55:27 > His email was so senseless that even Linus Torvalds himself wrote an angry reply saying that 2026-06-16 09:55:27 Sorry, but the bar for _that_ isn't too high. I learned a lot of Finnish words from him. 2026-06-16 10:10:51 See the link above. I'm just gonna have some more fun with it. Like... 2026-06-16 10:10:51 Yeah, so? I can make a repo with over 3K commits per _day_! 2026-06-16 10:10:51 ` well redhat killed xorg development and >Btw, X11Libre's X server got 1.3K commits in 2 weeks.` 2026-06-16 10:10:51 `* * * * * cd ~/my-repo && git pull && date +%s > stamp.txt && git add stamp.txt && git commit -a -m'Bump the stamp to annoy everyone' && git push` 2026-06-16 10:10:51 Easy. 2026-06-16 10:11:09 s/over/about/, s/3K/1.4K/ 2026-06-16 10:11:43 Politics aside, number of commits is not a reason for or against. 2026-06-16 10:16:01 Let me put it this way: I also have a project that's quite... political. Yet if I asked for it to be added, it'd likely happen. XLibre is different for one other reason: it interacts with a lot of other OS components. And is also very much security relevant (more so than Wayland, actually, because of its networking features that Wayland lacks), and thus effective channels for security collaboration need to exist. Which a single 2026-06-16 10:16:01 project maintainer just can't fulfill. 2026-06-16 10:16:57 * Let me put it this way: I also have a project that's quite... political. Yet if I asked for it to be added, it'd likely happen (it'd just be pointless, as its main use is as a browser extension and packaging it as an APK won't help anyone really). XLibre is different for one other reason: it interacts with a lot of other OS components. And is also very much security relevant (more so than Wayland, actually, because of its 2026-06-16 10:16:57 networking features that Wayland lacks), and thus effective channels for security collaboration need to exist. Which a single project maintainer just can't fulfill. 2026-06-16 10:18:10 So I guess when he has a team of 10 active maintainers, and it objectively works better than Xorg and also needs less maintenance overhead for the distro - that might be the time to ask again. I have high doubts that's ever gonna happen though. 2026-06-16 10:18:46 it's had a year to become relevant. it has not. idk that there needs to be anything more said about it. 2026-06-16 10:18:55 * the distro (whether it's due to bugs, or feature changes needing updates to build scripts etc.) - that 2026-06-16 10:18:56 Yes, can we please drop this topic? 2026-06-16 10:19:43 Honestly I'd rather see X.org (the server) entirely go away. Xwayland and libX11 are here to stay for another few decades though, I bet. 2026-06-16 10:21:07 Fun fact, I stopped using X11's server not just due to intel chipset tearing, but also due to objectively worse performance than on Wayland on the same GPU, and more overhead to set it up. With X11 I have to mess with config files, whereas Sway just works. 2026-06-16 10:21:09 * mess with system specific config files, 2026-06-16 10:22:35 I do have one technical question though. Is there a way to use Xwayland with a Wayland "dummy" compositor, and running an actual X11 window manager inside Xwayland? That then won't be able to support Wayland applications, but I do wonder if it's possible. 2026-06-16 10:23:36 As that might be a more promising path to keep X11 desktop environments alive. 2026-06-16 10:28:04 Ah, that's what https://wayback.freedesktop.org/ is. 2026-06-16 10:44:38 huh, made a new user and started startx, works... no longer need to add to input group etc? 2026-06-16 10:48:46 divVerent1, there's some potential advantages to using a xwayland base when remoting applications. If I am not mistaken, this allows for GPU-accel via wayland without needing VirtualGL for GLX indirection like in a traditional Xorg stack 2026-06-16 10:49:31 ofc. this would only be true if you run it through waypipe not ssh -Y|-X 2026-06-16 10:52:10 we dont have dwarfs in repo? 2026-06-16 10:53:18 frag: I had so much fighting with Xorg configs on both Debian and Alpine. 2026-06-16 10:53:35 My usual result from starting an X server is getting one with no keyboard, no mouse, till I find out which daemon I forgot to enable. 2026-06-16 10:53:39 Then have to set TearFree. 2026-06-16 11:10:37 Based on the name of who just joined, I have trouble keeping my strategic popcorn reserves at a sustainable level. 2026-06-16 11:16:27 apkbrowser shows packages missing in Anitya with their version in grey, nice! 2026-06-16 11:16:45 I was going to implement the feature only to found out it was already there! 2026-06-16 11:17:14 Something like https://www.phoronix.com/forums/forum/phoronix/latest-phoronix-articles/1639750-yserver-modern-x11-server-written-in-rust-with-the-help-of-claude-code or https://git.dec05eba.com/phoenix/ would be feasible one day maybe? 2026-06-16 11:21:40 doubtful. one of the big issues with X11 is that the video drivers are a shitshow, and redoing the X server implementation *will not solve that problem* unless the new impl uses Wayland's video drivers instead. 2026-06-16 11:43:14 realroot[m]: My favourite "modern X11" is Wayland. 2026-06-16 11:58:21 I do use wayland and like it for my purposes, although a year back when I tested around a few screen readers, I was not very impressed 2026-06-16 11:59:19 screen reading is still a missing part afaik 2026-06-16 11:59:44 I personally contend that bolting on accessibility frameworks onto GUI applications is not The Right Way Of Doing Things, but it's practical and solves (even if imperfectly) the needs of many 2026-06-16 12:00:24 Yeah 2026-06-16 12:00:33 someone must tackle this, and write a procol so that programs can actually describe their content in another way 2026-06-16 12:01:09 staceee: Adoption and standardization will always be an issue, but yes, and I think people are definitely doing this 2026-06-16 12:01:38 Just I wish it could be more complete before such widespread adoption of Wayland, or sth idk 2026-06-16 12:02:37 runxiyu: I could't get screen readers working 2026-06-16 12:02:49 And when I first used input methods, I had to implement scaling support myself 2026-06-16 12:03:10 Input method stuff is so much better than two years back 2026-06-16 12:03:14 Per my experience 2026-06-16 12:03:26 fcitx5 just works for me, smoother than on x11 2026-06-16 12:03:41 staceee: there's one movement to do this, with the impelemntation being in Rust, based on Chromium's API, and the IPC done via side-channel in a protocol where the documentation is "read the rust code". 2026-06-16 12:03:43 screen readers i dont use on a day to day basis so i dont have recent experience 2026-06-16 12:04:02 WhyNotHugo: That's... great. /s 2026-06-16 12:04:25 runxiyu: a lot of historical baggage in documentation for things like fcitx. You don't need any environment variables nothing. Just run it. Done. 2026-06-16 12:04:42 Input method is pretty broken on COSMIC tho :( 2026-06-16 12:04:53 KDE has screen reader support. I didn't investigate it very thoroughly because the quality of the voices on offer damn near made my ears bleed, though. 2026-06-16 12:05:22 :/ 2026-06-16 12:05:26 Input method have warts tho. You can't use more than one, and they rely on hacks to relay keystrokes they _don't_ handle. 2026-06-16 12:05:47 WhyNotHugo: virtual-keyboard? iirc 2026-06-16 12:05:54 or are you referring to sth different 2026-06-16 12:06:03 "modern X11 server" 2026-06-16 12:06:05 Sheila: voices are typically optimised for speaking fast, so tyipcally sighted users (including myself) think they're terrible, but we're not judging by the right standards. 2026-06-16 12:06:27 I was recently trying to make a stenography thingy and bailed due to not hnderstanding the protocols in depth enough 2026-06-16 12:06:50 runxiyu: yeah, input methods send unhandled keystrokes via virtual-keyboard, so compositors don't route virtual-keyboard to input methods, and you can't use a virtual keyboard with an input method (i.e.: fcitx with a tablet's virutal keybaord) 2026-06-16 12:06:59 WhyNotHugo: i tried the screen reader on iOS and it was both fast and sohnds nice 2026-06-16 12:07:22 runxiyu: What didn't you understand? I can help out. Also: https://git.sr.ht/~whynothugo/emoji-im :D 2026-06-16 12:07:35 Thanks, I'll look into that 2026-06-16 12:07:41 iirc the input method v2 goal is not to type things, but to insert/append/change text inputs 2026-06-16 12:07:46 WhyNotHugo: please be aware that I am deafblind, not sighted; while I *do* use computers visually, I am endeavouring to change that because I have photosensitivity issues, but that does not make me sighted. :p 2026-06-16 12:07:55 Well a few things, I'll get to explain things when I chexk on that project again 2026-06-16 12:08:06 which would be after doing some furgit stuff 2026-06-16 12:08:21 staceee: you're right. But input methods grab all keyboard input, and for some keystrokes they want to say "I'm not handling this, just pass it on to the application". That's the missing bit. 2026-06-16 12:09:13 I'm surprise no one ever wrote an input method client, as a bottom layer shell bar, to fix last typed word with typo, working for all programs 2026-06-16 12:09:27 WhyNotHugo: also, I have auditory processing issues, so I do need any screen reader I use to have reasonable audio quality. 2026-06-16 12:09:27 ah I see 2026-06-16 12:10:44 anyway. NVDA, JAWS, and VoiceOver (macOS and iOS) have acceptable audio quality, so it's frustrating that Linux is behind. 2026-06-16 12:11:17 Sheila: I'm honestly not happy at all with the quality of accessibility features on Wayland. As an appliction developer, I find that there's little to nothing that can be done, and at a Wayland level, there seems to be lots of policits (with some groups wanting to move accessibiilty out-of-bad into their own stacks). 2026-06-16 12:12:15 until someone write a protocol for wayland itself. I think everyone would then have to support this, gtk/qt etc 2026-06-16 12:12:44 staceee: seems pretty doable. You can use levenshtein as an initial version, but a purpose-trained langauge model is a perfect fit for that (you mostly need to know if the typed word has a neighbouring words with higher probability in the squence) 2026-06-16 12:12:59 but that could also be external to the compositor, but that predate its prupose 2026-06-16 12:13:21 staceee: there is AccessKit, but I don't know where that's at right now. and frankly I kinda resent feeling like I ought to. 2026-06-16 12:14:38 but that's Linux accessibility for you: you *cannot* just be an end-user. :\ 2026-06-16 12:14:53 Sheila: last I checked, the protocol was undocumented, "go check the code in Rust and reverse it from there", and it seemed to be _really_ complex. 2026-06-16 12:15:34 Which honestly, just feels like them saying "use our impleentation or gtfo". 2026-06-16 12:16:28 yeah, I saw that just now. at the very least they could have pre-generated docs for mere mortals. 2026-06-16 12:18:56 There's no spec to document; the code was written and that's the protocol. 2026-06-16 12:23:37 deeply ironic that the protocol for an accessibility toolkit isn't accessible, tbh. 2026-06-16 12:33:37 acccessibln't 2026-06-16 12:33:54 this is one thing i like about windows/mac. their accessibilty seems more serious 2026-06-16 12:35:37 yeah 2026-06-16 12:35:39 yes, which is why macOS is my daily driver. 2026-06-16 12:40:26 noodle: meh, only for some kinds. Trying to use macOS with a keybaord is a nightmare 2026-06-16 12:42:42 WhyNotHugo: ya i never used mac before :b just going off what i read / feel 2026-06-16 12:47:40 A couple of years ago in their first-boot wizard, most buttons you could focus and had to press Space (Enter didn't work), EXCEPT in one step where Enter worked and Space didn't, like wtf? 2026-06-16 12:51:57 Sheila: do you use brltty? I ask because I took over maintainence of liblouis and was wondering about ways to ensure usability of things like brltty. 2026-06-16 12:52:59 jvvv: I don't have a Braille display yet, and the only bare-metal Linux I have right now is a ThinkPad T500 I'm reluctant to pull out just for testing purposes at the moment. 2026-06-16 12:53:26 (well, and a Dell Latitude, but same issue.) 2026-06-16 12:55:36 Sheila: Ok, just thought I'd ask. I've tried looking into possibly purchasing some equipment for testing purposes, but I got bogged down in all the different types and not knowing which ones would have a more ubiquitous status. 2026-06-16 12:57:08 jvvv: I am probably not going to be a great tester for you anyway because I am having to relearn Braille with an Orbit Reader *specifically* (my fingers aren't very sensitive and so I *need* poster-quality displays to actually be able to read it). 2026-06-16 12:58:07 this is not me saying 'no', just making sure you're aware of my limitations. :) 2026-06-16 12:59:46 Sheila: Ok, thanks for your input anyways. If you ever think of ways that I can improve alpine's assessibility, don't hessitate to throw me a line. I am interested in these kinds of things because I worry that someday I may need them also. 2026-06-16 13:24:28 Darn, those things are **really** expensive. 2026-06-16 13:28:49 That was my reaction when I was looking into purchasing. Pretty difficult on a fixed income. 2026-06-16 14:23:17 WhyNotHugo: is there a wya to indicate "because of various issues with my machines i might not use alpine that much in the upcoming weeks, feel free to bypass me for things on packages i maintain" and such? 2026-06-16 14:24:24 runxiyu_: not that I'm aware of, maybe just mention it on #alpine-devel so others are aware? 2026-06-16 14:24:32 You're leaving us for plan9, aren't you? 2026-06-16 14:25:35 nah, ill definitely be back 2026-06-16 14:25:51 its just that my thinkpad is sorta broken so ill be using my macbook air m1 with a broken screen 2026-06-16 14:25:59 so i definitely need external display 2026-06-16 14:26:08 and asahi's fairydust driver doesn't appear to work well on asahi-alpine 2026-06-16 17:21:26 WhyNotHugo are you making a gummiboot fork or alternative? 2026-06-16 17:29:28 realroot[m]: https://git.sr.ht/~whynothugo/candyboot See issues, several features are missing. No security audit has been done. 2026-06-16 17:31:52 nice, are you planning on completing it? 2026-06-16 21:26:58 realroot[m]: yup, but it's not my #1 priority right now. 2026-06-16 21:27:22 Some sites in librewolf/firefox request "wants to access other apps and services on this device". What the hell does this mean? 2026-06-16 21:59:25 gummiboot is such a good name 2026-06-16 22:53:17 omni: it is, can't do better than that 2026-06-17 05:30:50 Hi all, I want to flag a package out of date. I don’t get why we’re using a Fedora tool, but I oauth’d Anitya and… well the package is already there, and Anitya shows the newest version, but Alpine’s version is old. So how do I flag it out-of-date? 2026-06-17 05:32:40 Is Alpine listed as a mapping for that package? 2026-06-17 05:32:49 If not, please add it 2026-06-17 05:33:27 Yes, the mapping is already there. 2026-06-17 05:33:53 But there’s no button to flag it as out-of-date. 2026-06-17 05:35:42 It should be done automatically. Anitya polls the repos (and often gets rate-limited), and then Alpine polls Anitya 2026-06-17 05:35:49 How out of date are we talking? 2026-06-17 05:36:16 And what package? 2026-06-17 05:36:21 Compare and contrast: https://pkgs.alpinelinux.org/package/edge/community/x86/py3-cloudflare versus https://release-monitoring.org/project/17053/ 2026-06-17 05:36:50 It already is flagged out of datre 2026-06-17 05:37:53 I hope the maintainer is okay. 2026-06-17 05:39:19 Is there anything I can do? 2026-06-17 05:40:52 1. Create an issue requesting the update. 2. Create an MR to update the version 2026-06-17 09:11:43 preferrably (2) via abump or such 2026-06-17 09:12:09 (do people read source diffs when bumping packages?) 2026-06-17 09:12:46 i dont do that for well-known things like make/autotools but on occasion i skim source diffs of less well known ones 2026-06-17 10:17:48 I think that diffs should be reviewed more but current tooling is lacking a bit. 2026-06-17 11:14:16 for those that build with network enabled 2026-06-17 11:14:28 theres also reviewing network deps.. 2026-06-17 13:25:43 davycrockett: there's a bug right now in apkbrowser where all packages will show they're not in anitya. 2026-06-17 13:25:58 ACTION needs to scroll before replying to messages 🤦‍♂️ 2026-06-17 19:45:43 the iso for ppc64le seems to be broken. I can't boot it in qemu using qemu-system-ppc64 -cdrom alpine-standard-3.24.0-ppc64le.iso -m 1024 -nographic -net nic -net user 2026-06-17 19:46:01 https://x0.at/0tHU.txt 2026-06-17 20:14:00 Noisytoot: can you make an issue for that? 2026-06-17 22:03:15 hi! seems like halloy is out of date in the alpine repos. 2026-7-2 was released a week ago. 2026-06-17 22:11:22 feel free to submit an MR: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/community/halloy/APKBUILD 2026-06-17 22:18:45 there is an open MR for halloy 2026.7.2 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/103657 2026-06-18 01:44:12 ikke: which repo should I make the issue on? not sure if aports is right 2026-06-18 02:06:04 aports repo is fine when it's not known 2026-06-18 03:20:34 Yes 2026-06-18 06:52:15 incus-feature vs incus? 2026-06-18 06:52:17 a bit confused on this 2026-06-18 06:52:21 quick search on the wiki didn't show much 2026-06-18 06:56:49 runxiyu_: apparently incus remains on LTS branches, while incus-feature will always update to the latest version 2026-06-18 06:56:56 interesting 2026-06-18 06:57:06 The comment in the APKBUILD file reveals that 2026-06-18 06:57:06 but i only see things like -client and -vm under incus-feature... 2026-06-18 06:57:19 errr, -vm is in non-feature 2026-06-18 06:57:23 lemme see if it's an aarch64 thing 2026-06-18 06:57:24 alr 2026-06-18 06:57:39 lemme reclone aports onto this machine which i'd have to do anyway in a few days XD 2026-06-18 07:07:04 uhmmmm 2026-06-18 07:07:21 incus-feature-vm installed qemu-system-x86_64 2026-06-18 07:07:25 but im not on x86_64 2026-06-18 07:08:01 but im pretty sure incus supports x86_64, aarch64, 1ppc64le, and s390x if i remember correctly? 2026-06-18 11:42:40 runxiyu_: when incus lts was last built it seems like the builders failed to build and upload the full list of subpackages. At least on edge. 3.24 has the full list same with 3.23 2026-06-18 11:43:19 I have an MR open to bump it and hopefully fix that it just needs merged, and larena's approval. 2026-06-18 11:43:24 ah, okay 2026-06-18 11:43:25 thanks! 2026-06-18 11:46:22 Np :) was similarly confused when testing something yesterday and got bit by it too 2026-06-18 12:44:49 Hmm, buildrepo should notice the missing subpackages and try to build the package 2026-06-18 13:01:25 ikke: I thought it was odd too. It could just be that they didn't get uploaded? I only checked dl-cdn, and the issue is only with edge 2026-06-18 17:05:43 when i start "easy effects", package easyeffects, it says it wants the "Linux Studio Audio Plugins". does anybody know what the package for those are? 2026-06-18 17:08:23 s/Audio// 2026-06-18 17:12:19 ok i found it, seems like the required packages are lsp-plugins lsp-plugins-lv2 2026-06-18 23:17:25 I'd like to configure some hosts in a declarative way, so they can be quickly wiped and re-created. What are others using to do this kind of thing on Alpine? 2026-06-18 23:44:46 /etc/apk/world :) 2026-06-19 00:12:50 WhyNotHugo: saltstack is my go to, ansible works, shell scripts cause they're fun 2026-06-19 00:14:31 Sertonix[m]: there's a lot of little tweaks beyond installing packages 2026-06-19 00:14:46 I do love the concept of apk/world tho 🫶 2026-06-19 00:24:10 Might not meet your declarative requirment but use opininated packages like 0+wifi that do the little tweaking. 2026-06-19 00:59:30 I did that previously (on Arch). I recall it being a bit burdensome, but I have a different agility with apk. 2026-06-19 01:00:14 Some things (e.g.: editing sshd_config) would have to happen via install hooks I guess, but those will almost always use seomthing like sed, no matter how many abstractions we have on top. 2026-06-19 01:07:40 durrendal: salt seems really neet. And is in aports, unlike puppet 2026-06-19 01:08:40 Indeed, and if you couple it with cloud-init or tinycloud you can easily answer the ssh setup/bootstrapping problem easily :) 2026-06-19 01:09:40 Unfortunately the version in our repos is a little rough, I can't upgrade it to 3008.1 until my saltext MRs are merged as the 3007 -> 3008 bump purged a ton of modules 2026-06-19 01:09:54 (they got slip out as "community supported") 2026-06-19 01:12:11 WhyNotHugo: if you do end up messing with salt or cloud-init and run into issues or have questions, happy to help. 2026-06-19 01:56:09 thanks! 2026-06-19 08:06:23 uhh, ~/.asoundrc no longer respected? 2026-06-19 08:06:58 oh, nvm 2026-06-19 08:20:27 seems to be a game i played in librewolf that keeps the alsa busy, even thought ive closed the tab :S how do i stop this (without restarting librewolf preferably) 2026-06-19 08:25:20 frag, fuser -v /dev/snd/* 2026-06-19 08:28:03 oh, nice, but im kinda worried about librewolf running something in the background, probably mining btc or something 2026-06-19 08:28:24 i played a game online, but closed the tab 2026-06-19 08:32:40 Web technologies for you 2026-06-19 08:41:04 crazy... 2026-06-19 08:41:14 stay off of the internet i tell you! 2026-06-19 08:48:10 WhyNotHugo: I would love to make tiny-cloud usable for that use case, and I think it may be close already 2026-06-19 10:57:41 the vim-common apk is currently 8 MiB, and drops to 2.8 when removing /usr/share/vim/vim92/{spell,doc}, and installed size drops by 14.4 MiB. I realize these are 2026-06-19 10:58:19 generally expected by vim, but would there be any interest in splitting these into another package, to make it easier to remove them? 2026-06-19 11:04:46 on a related note, I suppose it is not possible to recompress APKs from gz to zst without also resigning it? 2026-06-19 11:05:50 The apk v3 file format packages can be recompressed without resigning. The v2 file format not. 2026-06-19 11:41:18 Merging abuild 3.18.0_rc1 into alpine edge. Most notable change is that rootbld does not call apk-tools as root anymore. I hope not too much stuff is broken 2026-06-19 11:41:35 🎉 2026-06-19 12:02:00 frag: Firefox does the same for me pretty often too. 2026-06-19 12:02:20 I reported it upstream; looks like when there's a single muted/paused stream on any tab, it keeps the stream to pulseaudio alive 2026-06-19 12:05:48 Sertonix[m]: `abuild rootbld` still works, even after I run `apk add !abuild-sudo`. Maybe the latter doesn't need to be in install_if any more? 2026-06-19 12:06:07 I tried building a package with pkgusers too 2026-06-19 12:13:01 I would keep the install_if for non-rootbld usage 2026-06-19 12:14:02 Oh, it's required for that? 2026-06-19 12:23:39 WhyNotHugo i assume so yes, and probably chromium, i think chromium has an option for turning it off tho(?) 2026-06-19 12:26:53 WhyNotHugo: you can configure abuild to use doas instead but by default it uses abuild-sudo: https://codeberg.org/sertonix/aports/src/commit/f9fb58be71da1072119ae14411eaf5f179abbabf/custom/0+abuild/abuild.conf.in#L12 2026-06-19 12:46:12 can anyone try to run thunderbird on an aarch64 machine 2026-06-19 12:53:57 I don't have an aarch64 with a graphical interface 2026-06-19 12:54:02 +machine 2026-06-19 13:01:47 alr, hm 2026-06-19 13:01:57 could try ssh -X or waypipe or sth if you have the time 2026-06-19 13:08:56 Not atm, but later today perhaps 2026-06-19 13:09:00 thanks! 2026-06-19 13:09:11 my aarch64 machine is not strong enough to build thunderbird 2026-06-19 13:09:41 so im trying to make an alpine minirootfs on a debian cfarm and unshare'ing 2026-06-19 13:31:19 hmmm 2026-06-19 13:35:04 might be 16k pages 2026-06-19 13:36:39 byt firefox works very well? 2026-06-19 13:43:47 ftr firefox/thunderbird really need compression (ssh -C) to run reasonably over -X/-Y 2026-06-19 13:50:09 ok, it's somewhere on the network path 2026-06-19 13:50:37 i do suspect a stack overflow still..... 2026-06-19 13:56:21 runxiyu_: thunderbird should work fine on aarch64, what's the matter? 2026-06-19 13:57:06 f_: segfaulting when it touches network 2026-06-19 13:57:18 I could give it a shot, I've plenty of aarch64 hw w/ ready to go graphical interface 2026-06-19 13:57:52 uhhhhh 2026-06-19 13:58:00 it also only happens when i'm not using an http proxy 2026-06-19 13:58:41 also 2026-06-19 13:58:45 only when QUIC is used 2026-06-19 14:03:46 okay 2026-06-19 14:03:52 for people on 16k pages an aarch64 2026-06-19 14:03:59 does your firefox work when visiting http/3 sites? 2026-06-19 14:04:03 ah 2026-06-19 14:04:09 then I wouldn't be affected by this 2026-06-19 14:05:47 hm.. 2026-06-19 14:05:59 and indeed thunderbird works fine on amlogic a311d running gnome 2026-06-19 14:06:15 which is a pretty "normal" aarch64 SoC, iirc doesn't use 16k pages anyway 2026-06-19 14:19:39 el 2026-06-19 14:19:45 oops 2026-06-19 14:55:35 Hello, I need some help. I want to install Alpine Linux + bcachefs, but I'm having a problem: in the Live LTS 6.18 image, bcachefs isn't built into the kernel by default. I tried searching the APK list but couldn’t find anything (I tried installing bcachefs-tools, but that didn’t work-running `mount -t bcachefs` still gave an error saying the kernel doesn’t support bcachefs). Can you tell me how to download bcachefs so 2026-06-19 14:55:35 that it can mount bcachefs? (I understand that DKMS is required, but I couldn’t find any mention of DKMS in the APK list at all) 2026-06-19 14:57:34 Alpine has AKMS, but no idea if there is a source package for that 2026-06-19 14:59:29 AKMS as I understand it is a tool for building DKMS? (on Alpine Linux) 2026-06-19 15:05:25 DinoMax0n: bcachefs is no longer available in the kernel anymore 2026-06-19 15:05:33 it got dropped a few years back 2026-06-19 15:07:09 f_: hence dkms/akms to build it out-of-tree 2026-06-19 15:07:16 right 2026-06-19 15:08:43 DinoMax0n: akms does a similar thing as dkms, but uses apk to get the packages 2026-06-19 15:09:26 Not entirely true, but it has it's own format to build the modules 2026-06-19 15:10:38 ok, can AKMS build bcachefs in the kernel ? If so, how to do it ? 2026-06-19 15:12:09 Read https://github.com/jirutka/akms/ (and the linked man-pages) to find out how to build an AKMBUILD 2026-06-19 15:12:19 The rest is following the bcachefs documentation on how to build it 2026-06-19 15:13:09 Ok, thank you 2026-06-19 15:14:29 DinoMax0n: You can also look at the APKBUILD files for some *-src packages 2026-06-19 16:21:46 this may be me being a bit silly but should there be like a document somewhere that says what grub commands where done during installation. 2026-06-19 16:23:13 s/\.$/? 2026-06-19 16:30:36 penguinz_rule[m]: it may be helpful to read through the setup_grub function in /usr/sbin/setup-disk 2026-06-19 16:31:20 i have a learning question: why is openssl marked red? does it indicate that there is a newer version? "flagged as 4.0.1". and second question, why is riscv64 not marked the same? https://pkgs.alpinelinux.org/packages?name=openssl&branch=edge&arch= 2026-06-19 16:38:45 kroovy: yes, it means a newer version is available. I don't know exactly why it does not show up for riscv64 2026-06-19 16:39:34 kroovy: could be that the package for riscv64 was updated after it was flagged 2026-06-19 16:39:54 (the builder has issues and is slow, so it's regularly behind) 2026-06-19 16:40:10 yes, that was my theory... kind of a "race condition" 2026-06-19 16:41:00 tdback: thanks 2026-06-19 17:00:30 who here thinks ikke is awesome for answering so many questions all the time? :D 2026-06-19 17:00:51 o/ 2026-06-19 17:49:45 kroovy: agreed. kind, polite, accurate. a real treasure here on irc, we need much more of this kind of people 2026-06-19 21:32:05 Hello world 2026-06-20 02:42:55 penguinz_rule: re: setup-disk - i found that these lines: 2026-06-20 02:42:55 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/c6002e5e150a9f17b2b24244114839b4b17bda51/setup-disk.in#L332-334 2026-06-20 02:42:55 turned out to be relevant for me after upgrading recently in addition to the grub-install lines above it. 2026-06-20 02:42:55 without that extra copying, when i rebooted, i was getting `grub_memcy not found`. 2026-06-20 06:23:58 s/memcy/memcpy/ 2026-06-20 09:26:35 sogaiu: thanks 2026-06-20 11:16:00 does anyone else has experienced the error message crontab: must be suid to work properly when running crontab -e as non-root? is this related to /usr merge? 2026-06-20 11:25:53 when trying to edit your own, non-root, crontab? 2026-06-20 12:08:26 `abuild deps` does requires abuild-sudo. I wonder if this (manual) usage could not use doas/sudo explicitly instead. 2026-06-20 12:52:53 I have a vps that has ubuntu on it, and the provider doesn't provide alpine images. Is there a way to "hijack" an existing os install and transform it into alpine? I've tried a few times to do this by resizing the filesystem, unpacking the rootfs tarball to a different partition, but it doesn't work (and I can't get a console, so it's hard to know how it fails) 2026-06-20 12:55:07 Without console it's will be very difficult 2026-06-20 13:19:15 oh, turns out I _can_ get a console. Nice. 2026-06-20 13:23:13 probably going to use https://netboot.xyz/docs/booting/uefi/ to add a bootloader entry, then boot to that. Looks interesting! 2026-06-20 13:29:37 [@_oftc_lopid:matrix.org](https://matrix.to/#/@_oftc_lopid:matrix.org) Yes, exactly 2026-06-20 13:33:25 … 2026-06-20 13:34:41 I mean I understand why it would need setuid bit, but I didn't remove it, and it used to work 2026-06-20 13:35:36 it shouldn't, especially for a user crontab. which version are you using? 2026-06-20 13:36:52 oh, crontab does (crond doesn't) 2026-06-20 13:40:09 /usr/bin/crontab ⇒ /bin/bbsuid (busybox is version 1.37.0, alpine is 3.24 manually usr merged, I think I installed at 3.22) 2026-06-20 13:41:31 and /bin/bbsuid isn't suid now? 2026-06-20 13:43:32 lopid: /bin/bbsuid is suid for me. Try apk fix cmd:bbsuid ? 2026-06-20 13:44:02 msrd0 2026-06-20 13:44:09 ooh, I successfully hijacked a free ubuntu vps with an alpine install :) 2026-06-20 13:48:51 WhyNotHugo: That worked, thanks! 2026-06-20 14:20:23 elagost: free? NICE! 2026-06-20 14:32:38 The bbsuid errors with merge-usr are caused by a bug that's been fixed upstream but hasn't been accepted into Alpine yet. See: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/104092 2026-06-20 14:34:26 Basically, merge-usr didn't copy over the suid and sgid permissions when moving files, so every binary that was moved during the merge loses them. That MR fixes this by copying the permissions on move. 2026-06-20 14:34:45 Sadly, if you have already moved, you will need to manually re-add the permissions. 2026-06-20 14:43:03 WhyNotHugo: sadly it's oracle cloud, but it is still free, and has been for years. 2026-06-20 15:00:05 JustSouptheyhe[m]: Reviewed upstream commit, merged. 2026-06-20 15:00:41 Thanks a bunch! 2026-06-20 15:09:00 z 2026-06-20 17:12:57 can i list all packages installed from a specific repository which is optionally tagged? 2026-06-20 17:39:00 ikke: I opened the issue: https://gitlab.alpinelinux.org/alpine/aports/-/work_items/18281 2026-06-20 17:53:25 witcher01: One idea for narrowing the repositories is to use a custom repositories file that only includes the repository or repositories you want to search. This can be done with the apk global option --repositories-file. For the optionally tagged part of your question, I am not sure about that, maybe someone else knows. 2026-06-20 18:46:00 helo 2026-06-20 21:31:25 elagost: what approach did you take in the end? Just boot from install media? 2026-06-21 10:34:10 WhyNotHugo: I used netboot.xyz - I downloaded an efi executable, added it to the bootloader with efibootmgr, then booted to that from the console and it loaded the install media. 2026-06-21 18:35:23 runxiyu_: do input methods work for you on gtk4 applications? 2026-06-21 18:36:08 lemme try 2026-06-21 18:38:58 https://u.runxiyu.org/23c48224aa063074a6f8abfd6f1195e2793562de044cecec5565d6ff73c4f7f0-screenshot_1782067104.png 2026-06-21 18:39:00 it appears to 2026-06-21 18:39:06 using fcitx5 with niri on nautilus 2026-06-21 18:39:28 Huh. They don't for me, #18284 2026-06-21 18:39:40 fcitx5 on sway: https://gitlab.alpinelinux.org/alpine/aports/-/work_items/18284 2026-06-21 18:40:34 i could try sway a bit later 2026-06-21 18:40:47 No special configuration, right? 2026-06-21 18:40:57 nope 2026-06-21 18:41:36 what the fuck, I ran nautilus for like 10 seconds and my CPU is pegged at 200% with some runaway process it left behind. 2026-06-21 18:42:20 probably a search indexer? 2026-06-21 18:42:45 Yeah, apparently it left some daemon as a surprise gift after I ran it. It's going to scan a few hundread gigabytes of data just in case. 2026-06-21 18:43:31 This is borderline malware. They really are trying to imitate Windows. 2026-06-21 18:44:01 damn 2026-06-21 18:44:11 somehow i can't get that to trigger on my system 2026-06-21 18:44:15 reminds me of polkit, dbus, and co not getting cleaned up after programs that runs them 2026-06-21 18:44:17 what's the process called? 2026-06-21 18:44:24 localsearch-3 2026-06-21 18:44:34 It kept running for a few minutes after I killed nautilus, until I killed it. 2026-06-21 21:15:26 yeah its on my system too i just didnt notice 2026-06-21 21:15:28 insane 2026-06-21 21:18:40 WhyNotHugo: does not work on my sway eithet 2026-06-21 21:18:46 works on my niri though 2026-06-21 21:19:16 runxiyu: I have very loud obnouxious fans. 2026-06-21 21:19:39 200% CPU is easy to spot. 2026-06-21 21:19:53 So it doesn't work on sway, but does on niri? 2026-06-21 21:20:05 Can you send me `wayland-info` from niri? 2026-06-21 21:20:12 I'll compare with globals it exposes. 2026-06-21 21:20:35 If you can also send me a short output of `WAYLAND_DEBUG=1 nautilus` on niri that would be super. Want to see which globals it binds. 2026-06-21 21:28:07 hang on 2026-06-21 21:28:08 uhh 2026-06-21 21:29:49 WhyNotHugo: https://okturing.com/src/20385/body 2026-06-21 21:31:44 https://okturing.com/src/20387/body 2026-06-21 21:34:28 So niri has a `mutter_x11_interop` global objects. Other than that, it exposes a strict subset of what sway exposes. 2026-06-21 21:35:05 And nautilus does not bind to it :/ 2026-06-21 21:35:11 hm 2026-06-21 21:58:57 runxiyu: Huh, if I set `GTK_IM_MODULE=wayland`, then it works on sway. Is Niri setting this variable for you? 2026-06-21 21:59:18 I wonder why GTK3 can auto-detect that it's runnign on wayland but gtk4 can't. 2026-06-21 22:03:17 Because every GTK release is an improvement over the previous version 2026-06-21 22:04:14 I'm deeply concearned about the day that Firefox finally upgrades to GTK4 :( 2026-06-21 22:04:24 :( 2026-06-21 22:05:48 If I understand gdk/wayland/gdksettings-wayland.c correctly, the default behaviour is to ignore the running input method unless explicitly configured or overriden by the user. 2026-06-21 22:06:38 ah, yes, because I totally wasn't running that input method for a reason 2026-06-21 22:06:57 Surely a user who has installed an input method on their system and is running it would never want to actually USE that, who'd want to type in non-latin anyway? 2026-06-21 22:07:20 right?! 2026-06-21 22:11:43 WhyNotHugo: afaik they don't have any plans to update to gtk4? 2026-06-21 22:11:49 like, it's at the bottom of their list 2026-06-21 22:12:18 f_: The issue on the topic has plenty of activity (I should unsubscribe from it tbh) 2026-06-21 22:14:06 hm 2026-06-22 02:15:18 I have avahi and openrc installed. Why is avahi-openrc not auto-installed? 2026-06-22 02:16:49 Yooooooooooooooooooooooooooooo 2026-06-22 02:16:52 hiiiiiiiiiiiiii 2026-06-22 02:16:57 i'm crwlr 2026-06-22 02:17:07 o/ 2026-06-22 02:17:13 wow 2026-06-22 02:17:15 hi 2026-06-22 02:21:11 byee 2026-06-22 02:49:56 https://gitlab.alpinelinux.org/alpine/aports/-/work_items/17936 2026-06-22 02:49:59 any thoughts on this btw 2026-06-22 04:18:29 Hi, did something change w/ chrony from 3.23 to 3.24? It looks as I can't connect any client anymore since I updated and I'd like to find out probable fields to search. 2026-06-22 04:53:26 holgersson: There have been several changes to the chrony packaging within the last three months. I use it mainly as a client and for that it still seems to work as expected. But for details about the changes, you could look at https://git.alpinelinux.org/aports/log/main/chrony?h=3.24-stable. 2026-06-22 04:55:10 Only change since 3.23 is a upgrade from 2.6.0 to 2.6.1 2026-06-22 04:55:45 sorry, wrong package 2026-06-22 04:58:26 https://git.alpinelinux.org/aports/log/main/chrony?qt=range&q=3.23-stable..3.24-stable 2026-06-22 05:00:26 Nice, I didn't know about the '..' for comparing changes between to branches. That is much clearer than the link I pasted. 2026-06-22 05:00:38 s/to/two/ 2026-06-22 05:00:58 Me neither, I searched for it 2026-06-22 05:01:08 It's not the regular h= parameter 2026-06-22 05:02:22 It's pretty slick. Hopefully I remember that one as it will be handy to me. 2026-06-22 05:03:02 ftr, the .. syntax is something git itself uses as well 2026-06-22 05:03:59 Yeah, that is what caught my eyes. Like for doing a diff between commits or format-patch. :) 2026-06-22 05:09:51 Ok, I guess I never bothered looking at that part of the cgit interface. Now that I play with some, I realize I have missed some interesting ways to utilize the cgit page. Thanks! 2026-06-22 16:28:06 there are some broken symlinks in the lts modloop -- cd /.modloop/modules/firmware/ath11k/WCN6855/hw2.1/nfa765/; readlink amss.bin.zst; readlink "$(readlink amss.bin.zst)"; # amss.bin.zst -> ../../hw2.0/nfa765/amss.bin.zst -> ../../../QCA6698AQ/hw2.1/amss.bin.zst (the folder QCA6698AQ does not exist) 2026-06-22 16:28:31 not sure whether this should be filed as an aports or alpine-conf issue, or something else 2026-06-22 16:29:53 to clarify, this is on the live-cd, not a hdd install, and perhaps not really an issue at all 2026-06-22 16:41:56 Hey. I cannot find it. Where is the reference for /etc/init.d/dmcrypt file? I want it to skip mapping devices if the PC didn't receive keyboard input after some time 2026-06-22 16:42:21 sorry I meant /etc/conf.d/dmcrypt 2026-06-22 17:37:23 Hi Guys, in particular @jvvv and @witcher01 who tried to help me with an issue (a few days ago) where an upgrade appeared to break my rpi zero w + waveshare ethernet hat. I'm not sure when it changed, as I did upgrades without rebooting, and messed about with different repositories. Anyway, the long and short of it is that I trawled though the kernel configs (using make xconfig on my main linux box, and reading the notes) Eventually I stumbled on a new 2026-06-22 17:37:25 option, not mentioned in the alpine wiki, which is the g_multi USB gadget module. It works! It works! I said I would come back to report on my findings, so here it is. 2026-06-22 17:42:26 if it helps anyone else...SKU:16595 Part No.:ETH/USB HUB HAT Brand:Waveshare 2026-06-22 17:51:15 ps. my current kernel is 6.18.35-0-rpi 2026-06-22 17:59:54 toby: So you were able to get it working with the stock alpine rpi kernel? 2026-06-22 18:02:10 jvvv: yes CONFIG_USB_G_MULTI=m - I just had to modprobe the module 2026-06-22 18:04:25 toby: Nice. Congrats on getting it all working. 2026-06-22 18:05:17 root=UUID=91bfd7cc-e7c1-416d-b9cd-57f663d85d75 modules=sd-mod,g_multi,ext4,quiet rootfstype=ext4 2026-06-22 18:06:05 thanks jvvv, I was think ing the hat was broken, but it wasn't :) 2026-06-22 18:07:20 Good deal. 2026-06-22 18:09:38 I think I will review on the waveshare site, as it will be broken for everyone I suspect, I couldnt get it working on rpi debian trixie either 2026-06-22 18:14:02 Good forward thinking 2026-06-22 19:00:22 meow 2026-06-22 19:42:48 ed: An issue on the aports repo would be good to have 2026-06-22 20:09:42 so for this package conflict with icu, https://bpa.st/FVYA, do i remove the satisfies packages and then install later ? 2026-06-22 20:14:31 deleting some of my self-built packages and using -al and getting new errors 2026-06-22 20:19:51 https://bpa.st/HL6A 2026-06-22 20:21:13 Yeah, you have to find what package(s) is holding back some upgrades 2026-06-22 20:26:06 Looks a lot like qt5-qtwebengine is blocked. 2026-06-22 20:26:15 yeah 2026-06-22 20:26:36 penguinz_rule[m]: can you try `apk add -s qt5-qtwebengine=5.15.17-r13 2026-06-22 20:26:38 ` 2026-06-22 20:26:50 (assuming this is edge) 2026-06-22 20:27:01 this is .23->.24 2026-06-22 20:27:08 i was going to delete every package from there but i will try that 2026-06-22 20:28:28 Oh, qt5-qtwebengine has been moved to testing 2026-06-22 20:28:34 so not available for 3.24 2026-06-22 20:28:47 apk list --orphan 2026-06-22 20:29:36 there seems to be more than that that is blocking, so i will move to phone since i am removing firefox rn 2026-06-22 20:29:43 Should't apk-tools be giving a warning about removed packages then? 2026-06-22 20:30:35 Not everything listed in the error is a problem. Firefox likely isn't 2026-06-22 20:30:48 indeed 2026-06-22 20:32:05 it was qt5-webengine, i think i had that to instsll notepadqq locally 2026-06-22 20:32:11 Maybe a general rule could be to check conflicts of the version that should be installed and satisfies of the version that should be removed. 2026-06-22 20:34:26 yes, that's what I do 2026-06-22 20:35:15 icu-libs-78.1-r0 conflicts with icu-libs-76.1-r1 (makes sense). Question then is, what keeps icu-libs-76.1-r1 2026-06-22 20:35:34 satisfies: qt5-qtwebengine-5.15.17-r11[so:libicui18n.so.76] 2026-06-22 20:35:42 ye i went scorched earth when i shouldn't have 2026-06-22 20:45:39 thanks you all!! i successfully upgraded 2026-06-23 09:52:58 discord seem to have the same problem of running in the background, despite closing the tab, inf librewolf(and assumingly other browsers) 2026-06-23 09:53:23 tried to kill the process, but it killed the whole browser 2026-06-23 10:00:44 frag: i'd recommend a 3rd party discord client but beware, they might block your account if you get found out lol 2026-06-23 10:00:59 there're some for gtk and some for terminal iirc 2026-06-23 10:01:15 yeah, that would be a problem ;p 2026-06-23 10:01:30 mhm sadly 2026-06-23 10:07:45 Don't use proprietary platforms like that 2026-06-23 10:11:55 yup that's even better 2026-06-23 10:12:11 stoat.chat then matrix then xmpp to get a similar experience 2026-06-23 10:12:39 but yeah didn't wanna offtrack 2026-06-23 10:16:39 yeah yeah, some investment forums there that are nice not to miss :c 2026-06-23 10:27:12 fair. idk anything about investment tbh. maybe you can be the change you want to see and start one on irc/xmpp? 2026-06-23 10:27:27 i bet it'll be hard and would take time 2026-06-23 10:27:44 non-tech communities tend to be uncommon here 2026-06-23 14:18:51 what are the options in aports that we have for an irc server? (Im asking before search) 2026-06-23 14:26:45 looks like ngircd, inspircd & unrealird are available 2026-06-23 14:27:14 https://pkgs.alpinelinux.org/packages?name=*ircd&branch=edge&repo=&arch=&origin=&flagged=&maintainer= 2026-06-23 14:27:45 hm is ircd the standard 2026-06-23 14:27:56 there is ergo as well (that one I knew) 2026-06-23 14:28:52 thank you 2026-06-23 14:41:16 fabricionaweb: might be good to explain your use case. a number of people are in this chan who have experience with ircds over the years who might be able to save you some time. 2026-06-23 14:44:13 fabricionaweb: For a very small network with one server and preference for modern features, Ergo is probably the most appropriate 2026-06-23 15:16:34 wont be a big server, private, couple of friends and bots, few channels, maybe I look to ergo but nice to know the alternatives as well 2026-06-23 15:16:43 I appreciate the help 2026-06-23 15:21:33 i'd use ngircd unless impersonation is a risk, because you gotta setup atheme with it 2026-06-23 15:22:29 it's simple and lightweight 2026-06-23 15:22:59 I like unrealircd, personally. 2026-06-23 15:23:31 fair. forgot about it. thx for mentioning. i gotta look at it sometime 2026-06-23 15:23:31 what is atheme noodle? 2026-06-23 15:23:59 fabricionaweb: it's like a collection of irc services like nickserv, chanserv, etc 2026-06-23 15:24:07 so ppl can have accounts and channel ownership 2026-06-23 15:24:21 ah nice, yes that makes sense, we are used to it 2026-06-23 15:25:03 mhm. i'd just use xmpp for a private friend group 2026-06-23 15:25:23 mostly bc my old friend group used to hangout on discord 2026-06-23 15:39:08 hi. i write a p2p chat app and now wants to test it on WWW. there is anyone to help me? 2026-06-23 15:40:50 I needs someone to connect my chat server then chat with me a bit to see it works or not. 2026-06-23 15:45:11 it is open source and libre software. it is rests on https://codeberg.org/postala/ 2026-06-23 15:49:36 ACTION reads the disclaimer 2026-06-23 15:52:05 yes that disclaimer is horror. But its about p2papp not related here. i want to test p2pchat here. 2026-06-23 15:52:32 two app use same repo. 2026-06-23 15:53:11 see p2pchat Readme.md for details. 2026-06-23 21:48:14 helloo 2026-06-23 21:51:18 ive been using puppy linux on an old computer but i had also heard of alpine linux and so i wanted to try that, and since ive figured out how to use irc (originaly just for nethack) i might as well also try to use it here, and also irc people are very cool 2026-06-23 23:12:15 Hallo 2026-06-23 23:14:02 helloo 2026-06-23 23:14:45 halo 2026-06-23 23:14:50 (2000) 2026-06-23 23:15:03 (2001)* heck 2026-06-23 23:15:12 wow 2026-06-23 23:15:21 how are you doing? 2026-06-23 23:15:40 am alr thank 2026-06-23 23:15:41 hbu 2026-06-23 23:16:19 doing OK 2026-06-23 23:16:35 peak 2026-06-23 23:16:44 yup 2026-06-23 23:18:34 I am wondering if anyone here has had experience installing on a SBC? 2026-06-23 23:18:43 (not a PI) 2026-06-23 23:21:24 which sbc? 2026-06-23 23:21:35 radxa dragon q6a 2026-06-23 23:23:10 have a look at pmOS: https://wiki.postmarketos.org/wiki/Radxa_Dragon_Q6A_(radxa-dragon-q6a) 2026-06-23 23:24:26 OK, thank you 2026-06-23 23:29:49 so good news, i now have alpine running, bad news, my 20 year old computer with known overheating issues is not happy i think 2026-06-23 23:31:16 aw 2026-06-23 23:31:28 there should be some power management stuff you can tweak 2026-06-23 23:31:35 in bios and alpine itself 2026-06-23 23:32:27 @ultro404 sorry had to reconnect 2026-06-23 23:33:05 How long since last thermal paste exchange - or how long since last dust removal? 2026-06-23 23:34:44 noodle: i might go check the settings in the bios again, but alpine isnt installed yet i think (im actually waiting on hostname to restart) 2026-06-23 23:35:19 o why was it heating if it aint installed yet :b 2026-06-23 23:35:43 but ya juzz is right. start there 2026-06-23 23:35:52 then with alpine then with bios i think 2026-06-23 23:35:54 Juzz: a few months ago (like maybe 5 or 6 actually), i used the same thermal pad i use on my other laptop 2026-06-23 23:37:04 OK, what kind of thermal pad? What is the item overheating and what kind of coller i placed on top of it? 2026-06-23 23:37:24 coller ? cooler 2026-06-23 23:37:46 for a while i think the fan was just unplugged, after plugging it back in it no longer immediatly goes up to 110 degrees, thats actually why i was able to use puppylinux for a while, but also i was just going off ram for that 2026-06-23 23:38:38 OK, is this a laptop or a tower computer? 2026-06-23 23:38:47 for the cooler i could try to find one of the pictures i had taken a long time ago but i dont know if ill be able to, ill check for the thermal paste but i think it was tp-[a bunch of numbers] 2026-06-23 23:39:57 its ptm7950 2026-06-23 23:40:20 that should work great, so that shouldn't be the problem 2026-06-23 23:40:21 i'd just put on fresh thermal paste tbh 2026-06-23 23:40:36 (i don't use pads) 2026-06-23 23:40:36 also the ram is now burning hot to the touch 2026-06-23 23:40:42 aw :C 2026-06-23 23:40:58 and turns out i just had to press enter to get back to the command line 2026-06-23 23:41:00 Juzz: I am working on creating an image for dragon q6a. I have created local build for kernel and wifi/bluetooth driver/firmware. I just need to work on the mkimage script for creating the image. I really should finish it so I can test it out. 2026-06-23 23:41:04 ddr4 or ? 2026-06-23 23:41:35 @jvvv: Awesome, I would be happy to help test it 2026-06-23 23:42:08 ddr2 2026-06-23 23:43:03 jvvv: oooo that does sound fancy 2026-06-23 23:43:32 Juzz: I will get back on it tonight. I don't why I have been procrastinating at this point. 2026-06-23 23:43:50 jvvv: we all do that :P 2026-06-23 23:45:08 ultro404: do you know the temps of the southbridge? As I recall back with ddr2 we had a southbridge that could become temperemental when hot 2026-06-23 23:46:50 i have absolutly no clue what that is and this has actually been alot of new stuff mentionned than what i haad tried on pine64 so actually let me search this up cause that sounds important 2026-06-23 23:47:38 ultro404 do you know the make and model of motherboard? 2026-06-23 23:49:35 would that be in the manual or on the motherboard physically 2026-06-23 23:51:59 both 2026-06-23 23:52:50 on the motherboard it would most likely be between the pci slots - and it would be featured somewhat prominently on the manual 2026-06-23 23:54:56 so i know the cpu is an amd athlon II, because ive looked at it when i disassembled it before and also theres a sticker on the front of the computer, but the manuals and specs online dont have an amd one 2026-06-23 23:55:16 OK, that is progress 2026-06-23 23:55:17 i remember why i stopped working on this computer... actually i dont the keyboard feels great 2026-06-23 23:55:56 athlons were hot running 2026-06-23 23:56:04 oh also actually if your trying to help me i should probably tell you what the computer is 2026-06-23 23:56:17 its an hp g61-408ca 2026-06-23 23:56:37 that helps 2026-06-23 23:56:45 it also had a seperate gpu chip i think 2026-06-23 23:57:46 It's a laptop, I can gather from searching the name 2026-06-23 23:59:40 OH, i found the old pictures. also yeah its a laptop (it also has the french canadian keyboard) 2026-06-24 00:00:27 WOW 2026-06-24 00:00:38 one of the chips says southbridge 2026-06-24 00:00:59 should i figure out how to send the pictures or just write whats on the chips 2026-06-24 00:02:42 Yup, the southbridge might get hot - also removing dust from it can help quite a lot 2026-06-24 00:03:38 im rereading the messages i sent to my friend (cause thats how i found the pictures) and its kinda funny, i was wondering if that was a second gpu 2026-06-24 00:04:18 a trip dopwn memory lane :) 2026-06-24 00:05:16 also does igp tell you anything (like radeon igp) cause im wondering if the other chip which i thought was the first gpu is actually just the northbridge 2026-06-24 00:05:31 as far as I recall ddr2 memory modules did not have serious heat issues 2026-06-24 00:05:54 yup both north and south bridge might be a source of trouble if they get too hot 2026-06-24 00:06:05 yeah i dont think its the ram itself, its just really close to the heatsink 2026-06-24 00:08:10 also, really surprising thing, the computer is still horrifically hot but somehow it hasent died, like its been on just doing nothing but its been on this whole time, so i guess i had solved the temperature issue but im really not sure, i am kinda terrified of continuing though because its so hot 2026-06-24 00:08:57 is it hot in the exhaust? 2026-06-24 00:10:14 when it is exhaustin hot air - more hot than it feels on the keyboard/top side - then the cooler is still working 2026-06-24 00:10:34 its hot the whole way along the heatsink pipes, the air being blown out the side isnt actually that hot (like i need to touch the grate to feel it) but the area under the fan and under the cpu are horrifically hot 2026-06-24 00:10:44 if it is hotter top side/keyboard than at the exhaust then I would be worried 2026-06-24 00:10:54 ah 2026-06-24 00:11:02 i would be worried too then 2026-06-24 00:11:30 because then the cooler is not transporting the heat away properly 2026-06-24 00:11:31 wait im gonna go check something on my other computer 2026-06-24 00:13:47 you know... it probably shouldve been obvious that a fan blows air 2026-06-24 00:14:38 should i take the computer apart again to try to figure this out or just install alpine 2026-06-24 00:14:56 if it is hot all the way along the heat pipes - but not so much just outside the exhaust, then that might be the actual heatsink not being properly connected to the heat pipes or the blower not doing a good enough job 2026-06-24 00:15:38 i think im gonna take it apart 2026-06-24 00:16:02 some laptops have the heatpipes and heat sink separate from eachother but with a compound in between to make sure heat transfer is good, so check that 2026-06-24 00:16:59 could be good to do a good take apart and put back together again session 2026-06-24 00:19:40 remember to have cotton swabs, isopropyl alcohol, paper towels and new thermal paste/pad 2026-06-24 00:20:34 ok so ive also found a picture of the heatsink (on the cpu side), its a copper pipe that goes up to like a grey metalic holding thing... ok this is really hard to describe is imgur a good place to put pictures 2026-06-24 00:20:53 should work 2026-06-24 00:23:52 https://imgur.com/a/ylDCmxM wabam 2026-06-24 00:24:24 that looks like the cpu 2026-06-24 00:24:52 yeah the first picture is the athlon 2026-06-24 00:26:04 i also have clear pictures of like all the other chips though i dont know if that would help, also i checked near the southbridge chip right now and its really not that hot at all 2026-06-24 00:26:31 OK, but it wouldn't hurt to see 2026-06-24 00:27:49 some south-/north bridge chips weren't that hot 2026-06-24 00:28:40 if HP didn't put a heat sink on it it should cope with whatever temps should arise 2026-06-24 00:29:16 *should* because sometimes they were wrong - but that is very few cases 2026-06-24 00:29:29 https://imgur.com/a/EL6vy4U boop, though i realized that discord probably compresses pictures or does something cause there not as clear 2026-06-24 00:30:09 looks alright 2026-06-24 00:30:41 is there thermal pad on the chip on the right? 2026-06-24 00:30:41 but yeah like the warmth could be mistaken for the heat coming from the harddrive (or maybe the hard drive is hot because its also next to the cpu) 2026-06-24 00:32:05 yeah on the igp one 2026-06-24 00:33:03 OH 2026-06-24 00:33:28 yeah, that needs connection to the heatpipe as well 2026-06-24 00:33:29 so i have a picture of the whole motherboard 2026-06-24 00:34:11 https://imgur.com/a/l7Ddhma 2026-06-24 00:34:22 Hej sukkerfri0039 2026-06-24 00:36:24 (i have a question (about irc) should i say hello too, i havent been saying hello to the other people, but you said hello, but i dont know them either) 2026-06-24 00:36:26 ultro404: up there where the copper goes into the cooler - make sure it connects good to the heatsink and that the blower is cleaned out from dust and other stuff that might inhibit airflow 2026-06-24 00:37:04 ultro404: ohh.. I just saw a name that is from my country, so I said hi :P 2026-06-24 00:37:55 in english it is sugarfree :P 2026-06-24 00:38:12 ooooohhh... could i actually fix an issue with the pipe not touching the fan properly, cause that kinda looks like its welded there 2026-06-24 00:38:22 ooh 2026-06-24 00:38:38 the fan doesn't need to touch the pipe 2026-06-24 00:39:32 wait what did you mean by connects good 2026-06-24 00:40:09 sry brain fails me right now 2026-06-24 00:40:19 can't remember word in english 2026-06-24 00:40:54 ah, yeah its fine when i take it apart ill come back here 2026-06-24 00:41:14 the copper thing a majic with fins - that needs to connect to the heatpipe 2026-06-24 00:41:49 the fan just needs to direct it's air towards the heatsink 2026-06-24 00:42:34 is disassembling the actual fan part a permanent thing, because i do have another fan from a broken laptop i could take apart to look at 2026-06-24 00:43:33 oh yeah that makes sense 2026-06-24 00:43:55 no, usually the thing to the left of the heatpipe (the blower) just needs to blow air over the fins of the heatsink to the right 2026-06-24 00:44:28 you can see to the right that there are fins to allow air to travel through it 2026-06-24 00:46:30 oh thats why no air blowing out the side was immediatly recognisable as a bad thing 2026-06-24 00:46:43 yup 2026-06-24 00:47:36 however the blower on the left - it needs to have fairly clean blades in order to blow air as it should 2026-06-24 00:48:36 you can clean that out with either compressed air - or by using wet wipes, but in general they should be without anything sticking to the blades 2026-06-24 00:49:30 from your picture it has the intake on the back and blows out hot air to the side 2026-06-24 00:52:12 i remember last time i had cleaned them with qtips and alcohol 2026-06-24 00:52:25 and that is good 2026-06-24 00:52:32 they did have alot of dust stuck on them 2026-06-24 00:52:52 do you have pets? 2026-06-24 00:53:42 if only you could see the state the other chair my cat sleeps on in my room 2026-06-24 00:54:22 well, they are family members after all... It only increases the amount of times you need to clean things :P 2026-06-24 00:55:10 though for most of its time this computers been at my grandmas house (it was my aunts) but also it hadent been turned on in like 10 yearsish and it was just sitting in a closet so i havent had to experience the pain of cat hair stuck in the fan 2026-06-24 00:55:10 something most of us are prepared to do to have them around ;) 2026-06-24 00:55:50 ...i... i should clean my own laptop i think, its been in this house for 4 years and my cat used to sleep on it in winter 2026-06-24 00:56:01 usually not being turned on also doesn't increase the dust/hair problems, so that shouldn't be it... 2026-06-24 00:57:33 I had not cleaned my own PC in quite a few years - when I finally did, OMG a lot of dust was released... and at that time I didn't have any pets... 2026-06-24 00:57:48 though also the thermal paste it had before i had changed it looked pretty agreegious 2026-06-24 00:58:03 I should probably clean my computer soon :P 2026-06-24 00:58:12 oh 2026-06-24 01:01:29 OK, thermal paste if it looks dry and cracked then that is probably the reason - if it is moist and clingy then it's probably not that part that's the issue (however once you take it apart you need to put something new on) 2026-06-24 01:02:41 it apparently looked burnt according to my friend who actually builds pcs (and not just take apart laptops for fun) 2026-06-24 01:06:51 mmmmmmm i think im gonna go to bed, good night and thank you alot 2026-06-24 01:07:31 anytime, and sleep well :) 2026-06-24 03:17:58 Juzz: I don't think I am going to make any progress tonight. Will try again tomorrow. 2026-06-24 15:44:48 how come the gajim package (in community) depends on packages in testing (py3-dbus-fast and py3-bleak)? 2026-06-24 15:48:48 hmmmm 2026-06-24 15:49:03 cpuld you log/blame it? 2026-06-24 15:49:09 i wonder what 3.24 does 2026-06-24 15:53:56 blaming files in aports takes 10000 years 2026-06-24 15:54:33 Noisytoot1: Use commit-graph path bloom filters 2026-06-24 15:54:41 I think aports mostly has a linear history 2026-06-24 15:54:50 yes, that helps a lot 2026-06-24 15:55:14 furgit-legacy supported them :P 2026-06-24 15:56:45 Noisytoot1: Why do you say it depends on those packages? 2026-06-24 15:57:38 ikke: https://x0.at/Nm0Z.txt 2026-06-24 15:57:49 it might be indirect 2026-06-24 15:58:16 Nothing says that they are dependencies 2026-06-24 15:58:24 masked in @testing 2026-06-24 16:04:15 I'm not sure what's causing the error then 2026-06-24 16:06:52 I had py3-bleak (without the @testing) in my /etc/apk/world, but why was adding gajim triggering the error and how did it work in the first place? 2026-06-24 16:08:08 Noisytoot1: apk always try to resolve world completely 2026-06-24 16:33:38 good morning (or well its noon but ehh) 2026-06-24 16:56:01 woo, ive disassembled the whole thing and im at the fan now (now i just have to clean it and check that its blowing air properly 2026-06-24 17:05:50 ok so the fan isnt dusty at all (cause i cleaned it a few months ago) i do wanna test what the airflow is like 2026-06-24 17:11:39 ikke: it wasn't happening when I tried to install other packages though 2026-06-24 17:28:36 py3-bleak has probably not been updated due to missing the tag and is now holding back a package that needs to be updated to install gajim. 2026-06-24 17:50:41 ok so i learned alot 2026-06-24 17:52:41 i was right yesterday the entire path the heatsink pipes take and the cpu are hot, but on the fan side the area near the heatpipe is really hot, but the air being blown out is not as hot and theres also not that much air (the grates 2026-06-24 17:53:23 buh, accedentally sent my message, [the grates are also not that hot but id guess thats just cause theres less metal there)] 2026-06-24 18:15:06 used 4h36m to back up 215G from microsd to an external usb, using my rpi 2026-06-24 18:15:20 ups... disregar 2026-06-24 18:15:23 d 2026-06-24 18:36:20 ultro404: my most recent learning on the topic is that you want to blow fresh air INTO the heatsink/radiator, not pull hot air out. Pushing air increases pressure so it flows out naturally. 2026-06-24 18:37:38 my most recent learning is there are apparently bubble in my thermal paste 2026-06-24 18:41:28 https://imgur.com/a/FBLilt2 heres a picture 2026-06-24 18:42:23 im guessing its because the screws are all around the cpu and that chip is just kinda besides it ( 2026-06-24 18:43:53 why do i keep accedentally pressing enter 2026-06-25 19:25:54 jvvv: Have you had time to look at the installer for the Dragon Q6? 2026-06-25 19:31:44 Juzz: I am working on putting together an image right now. I am very close, it should only need a few tweaks. 2026-06-25 19:32:59 The part I am trying to ensure is that the wifi and bluetooth drivers and firmware are available in the installer environment. 2026-06-25 19:33:53 Since these are out of tree, I am having to learn parts of the image building I had not previously been familiar with. 2026-06-25 19:34:06 Ahhh, OK 2026-06-25 19:35:24 I am getting close, but won't publish anything until I have tested on my own dragon. If I were to mess up my sbc that is one thing, but I don't want to be responsible for messing up someone elses. 2026-06-25 19:37:26 Show if the image I am currently building builds successfully, I will be busy testing it on my for a short while. 2026-06-25 19:38:39 It will be an iso image that can be burned to a thumb drive. I've included drivers for nvme and ufs, so hopefully all bases covered. 2026-06-25 19:44:58 sounds awesome :) 2026-06-25 19:48:48 Well, I am hoping... just as soon I can get the firmware into to modloop, I will then be testing. So far that is proving weird, but I will figure it out. 2026-06-25 21:37:57 where I can find the latest release notes for v3.24 2026-06-25 21:38:29 amin: https://wiki.alpinelinux.org/wiki/Release_Notes_for_Alpine_3.24.0 2026-06-26 10:33:14 hi, im trying to use mkimage.sh to generate an iso for alpine linux. i got tha iso and it boots, but it throws me into a rescue shell from the initramfs. My script looks like this: https://paste.sr.ht/~ivanpollak/dc8eae8f26fec5a1ad5f1e995a1dad79c771daef. I already messaged mps and he told me that i should ask here. 2026-06-26 10:35:21 he told me that ill probably have to do something with /etc/mkinitfs/mkinitfs.conf. is that correct? if so, what would need to be set? im not really familiar with doing something to the linux kernel. 2026-06-26 10:43:36 Did you try without the root= option? 2026-06-26 10:59:06 ill do that later, thanks 2026-06-26 15:15:51 jvvv, ikke: Just read your answers about chrony from …days ago. Thanks! 2026-06-26 15:49:52 Sertonix: still throws me into the recovery shell. It tells me: Mounting boot media: failed 2026-06-26 15:55:43 In that case I would try to figure out if I can manually mount the boot media through the recovery shell. If that is not possible chances are that some required kernel module are missing from the initramfs. 2026-06-26 15:59:07 i think thats the case. I cant find the bootstick‘s file system. Can i add the required modules also with mkimage.sh? 2026-06-26 15:59:10 But I can't tell how difficult that is in your case. 2026-06-26 16:04:50 initfs_features="..." can indirectly include kernel modules but not sure which ones you need. profile_base already includes most. 2026-06-26 16:09:54 Do you have a working asahi linux system that could be used to figure out required kernel modules? 2026-06-26 16:10:22 uh 2026-06-26 16:10:24 hang on 2026-06-26 16:10:38 Sertonix[m]: codeberg.org/runxiyu/dotfiles 2026-06-26 16:10:53 config customized after copying config from linux-asahi from alpine 2026-06-26 16:11:03 but my config is for the fairydust kernel though 2026-06-26 16:11:13 to enable external monitors on macbook air 2026-06-26 16:14:19 i have found the modules: https://github.com/nix-community/nixos-apple-silicon/blob/main/apple-silicon-support/modules/kernel/default.nix 2026-06-26 16:15:38 ill try it in a second 2026-06-26 16:16:53 ivanpollak: You could try to build a modified mkinitfs version that includes an "asahi" feature which includes these kernel modules. Then pass the repository with the modified mkinitfa build to mkimage.sh. 2026-06-26 16:18:15 ill first try to just add them to the initfs_features to see if it works okay? 2026-06-26 16:18:32 It won't but sure :) 2026-06-26 16:18:44 why do you think that it wo t? 2026-06-26 16:20:21 ahh i think i see the issue, mkinitfs doesnt know about those festures right? 2026-06-26 16:20:51 mkinitfs features are mapped to kernel modules through $ROOTFS/etc/mkinitfs/features.d/$feature.modules files, not implicitly. 2026-06-26 16:20:54 Yes 2026-06-26 16:21:41 thanks ill do that. also thanks that you‘re explaining me all this! 2026-06-26 16:22:10 Note: In the case of mkimage.sh these files are not read from /etc/mkinitfs/features.d, which is why I proposed the patched mkinitfs package 2026-06-26 16:22:52 (mkimage.sh or to be more percise update-kernel creates an extra rootfs for that) 2026-06-26 17:32:39 Sertonix: is there a way to not write the full path for the needed kernel modules? for example; i dont know where apple-mailbox is located. Can i just write something like **/apple-mailbox? 2026-06-26 17:39:01 nevermind, i would probably have to go through every module and find it either way 2026-06-26 17:59:01 Certainly not ** globbing. Explicit list (eg. based on file list in package) is the way to go 2026-06-27 14:48:10 Hello, I've recently updated my machine packages to latest edge but since I can't get any video to play, I check inside dmesg but didn't found anything that seems related, and in /var/log/messages the only weird thing I see is a warning from pipewire telling that it has been initiated from something else than openrc (while on openrc). 2026-06-27 14:48:10 Do you have any idea how I can fix/debug that ? 2026-06-27 15:11:25 what do you mean by video? monitor doesn't receive a signal? 2026-06-27 15:12:36 Any video I'm playing is stuck during like 30 seconds, and then it plays but without any sound 2026-06-27 15:12:48 you mean a movie? 2026-06-27 15:13:21 I tested on youtube and my own jellyfin, both the same issue, if I rollback to alpine LTS it plays correctly 2026-06-27 15:13:32 hm 2026-06-27 17:35:13 has anyone tried using alpine as a vm guest with gpu acceleration? ive tried both wayland and x11 wms and both fallbcak to llvmpipe with "virtio-gpu: driver missing" errors 2026-06-27 17:36:02 the module itself is loaded, i can only guess that the mesa build doesnt support it? i looked at the abuild and couldnt find where to add it though 2026-06-27 18:17:01 update, i managed to get the wayland side working with WLR_RENDERER=pixman. this workaround isn't necessary on other distros like arch though 2026-06-27 18:18:48 alfmo: Do you have installed libdrm? 2026-06-27 18:20:21 visone[m]: yes. i also tried mpv with --vo=drm and that works as well 2026-06-27 18:29:21 alfmo: do you have any cards on `/dev/dri/` 2026-06-27 18:31:36 visone[m]: visone[m]: yes, card0 and rendererD128 2026-06-27 18:32:35 my user is in the video group but it also fails the same way as root so that rules out permission issues 2026-06-27 18:33:25 alfmo: Try setting this env `WLR_DRM_DEVICES=/dev/dri/card0` 2026-06-27 18:37:54 visone[m]: same error with that too. its using libegl as the backend and just running `eglinfo` also gives that same error, so i dont think its related to wayland 2026-06-27 18:49:35 i think its due to -Dgallium-drivers not containing virtio. the apkbuild only adds it to -Dvulkan-drivers 2026-06-27 18:49:53 in mesa that is 2026-06-27 20:36:26 yeah that worked. would be nice if it could be upstreamed, i think virtio-gpu is a common enough usecase to justify it. patch: https://bpa.st/HZRA 2026-06-27 20:40:09 Now I wonder why I have sway running even with "virtio_gpu: driver missing" error message. 2026-06-27 22:38:30 Sertonix[m]: it's probably falling back to software rendering 2026-06-28 05:41:32 Heyy 2026-06-28 05:48:16 I need some help im running alpine Linux diskless version downloaded from pi image (yea I'm running this on my raspberry pi. I was to have some applications on it permanently but as I'm following the docs on https://wiki.alpinelinux.org/wiki/Diskless_Mode I'm running to run the command "mkdir /media/( my partition id) " the output was can't create file directory /my partition : File exist. I am really confused on this part and could use 2026-06-28 05:48:16 some guidance. Thanks❤ 2026-06-28 06:03:06 "I need some help im running..." <- Yk what nvm 2026-06-28 10:17:15 has anyone managed to get swaylock working with fprintd, i've followd the instructions at https://wiki.alpinelinux.org/wiki/Fingerprint_Authentication_with_swaylock, installed util-linux-login, enrolled my fingerprint, sudo-rs+login work fine, but swaylock refuses to accept my fingerprint 2026-06-28 13:17:02 anyone else getting kernel oops with a recent radeon card? seems to do a null deref in the amd driver 2026-06-28 13:17:15 somewhat easy to reproduce with mpv 2026-06-28 13:22:42 looks like arch-install-scripts has started to conflict with alpine-conf at some point recently 2026-06-28 13:23:41 ah, there's already an aports issue for it 2026-06-28 15:39:54 Hi, I had some issues on the 3.24 update with a mirror and switched to dl-cdn.alpinelinux.org. What is the preferred way from project's side: a) use CDN or b) specific mirror? Speed does not really matter for me. 2026-06-28 17:45:14 holgersson: for us it doesn't really matter. But we're interested to learn if a specific mirror has issues 2026-06-29 00:18:35 Hey, I have a question about awall. I'm trying to follow the SSH to another host section from the zero-to-awall wiki page, but activating the policy gives an error 2026-06-29 00:18:37 https://wiki.alpinelinux.org/wiki/Zero-To-Awall#SSH_to_another_Host this one 2026-06-29 00:18:59 Type mismatch: 'filter.1.service' should be string, is table Type mismatch: 'dnat.1.service' should be string, is table Invalid port range at dnat.1.to-port: 22 2026-06-29 00:19:17 Sorry multiple lines got put into a single message 2026-06-29 05:43:49 qemu-ui-spice-core appears broken on aarch64 with missing symbols 2026-06-29 05:49:15 nvm 2026-06-29 06:21:33 Hello everybody, I remember reading somewhere that not all desktop environments are available on aarch (specfically KDE), but I cannot find that information back. Any pointers= 2026-06-29 06:39:39 pu: I am pretty sure plasma is available on aarch 2026-06-29 06:40:20 These days we even have an alpine-based linux distro for mobile phones which are mostly aarch 2026-06-29 06:40:58 Plasma mobile and GNOME are both available there, and I am pretty sure KDE too 2026-06-29 06:51:55 For questions about what software is available and on what architectures, fasted and most accurate way to find out is by checking on https://pkgs.alpinelinux.org 2026-06-29 06:54:16 By using '*plasma*' in the search box and selecting aarch64 in the architecture selection box, you should see that plasma is supported on aarch64 and that the package you are interested in is probably 'plasma-desktop' 2026-06-29 07:59:57 jvvv: I did check the packages, and setup-desktop does its job, but then there's no login manager on boot. Same process with xfce works perfectly. I'm tyring to install this on a Mac, under UTM.app. Worked fine for Debian and Fedora, both with KDE. 2026-06-29 09:06:06 DaKnig: "KDE" is the community making software, "Plasma" is the software and platform. So yes Plasma Mobile is available on aarch64, and so is Plasma Desktop. But you can not say KDE is available 😉 2026-06-29 09:06:53 pu: make sure the sddm service is enabled. I'm currently typing this message on an Alpine Linux installation using Plasma Desktop within a VM running in UTM on a M3 mac, works just fine 2026-06-29 11:13:28 "DaKnig: "KDE" is the community..." <- They talked about desktop environment 2026-06-29 11:50:48 Yes. But you said "Plasma Mobile ... available there, and I am pretty sure KDE too". As if KDE is an alternative to Plasma Mobile 😉 2026-06-29 11:51:51 I guess they mean Plasma Destkop 2026-06-29 11:52:01 which previously was just called KDE ages ago ;) 2026-06-29 12:05:46 greetings, intermediate alpine user here.. trying to install using the netboot image.. created a syslinux menu entry for alpine, but fails to find /sbin/init.. the docs are quite sparse.. suggestions are welcome! 2026-06-29 12:07:19 btw the User Handbook has no PXE instructions, outside scope of the handbook :-( 2026-06-29 12:18:23 f_: I know what they meant, I'm just explaining what the actual terms are 2026-06-29 12:18:40 ¯\_(ツ)_/¯ 2026-06-29 13:06:30 "DaKnig: "KDE" is the community..." <- kde means "K Desktop Environment" 2026-06-29 13:42:39 PureTryOut: sddm is enabled, but the screen stays blank. I can login via a serial console and I see that all services are running. Did you configure anything special in terms of graphics driver or virtual hardware? 2026-06-29 13:44:49 enable logging in the sddm config? 2026-06-29 13:53:14 Check /var/log/sddm.log for any obvious errors 2026-06-29 13:54:12 Upslaught: it hasn't meant that since they started with using Plasma since KDE Plasma 4.0 2026-06-29 14:36:59 PureTryOut: https://tpaste.us/WJxv 2026-06-29 15:32:26 It meant Kool Desktop Environment back in early days :P 2026-06-29 17:16:52 I have a raspberry pi with alpine in diskless mode, which works great. now I have updated to 3.24 but my kernel is unchanged (6.12.67-0-rpi), where I think 3.24 uses 6.18.x. how do I update the kernel? 2026-06-29 17:16:52 the docs mentions installing the mkinitfs package but I am unsure if this will yield a working raspberry pi kernel if following the steps (https://wiki.alpinelinux.org/wiki/Diskless_Mode#Upgrading_a_diskless_system) 2026-06-29 17:17:17 upgrading == changing to 3.24 repo, updating and upgrading all packages 2026-06-29 17:18:12 dont know the real answer, but i just install a new tar-ball, i install my system again every (seldom) reboot tho 2026-06-29 17:18:40 the kernel tar? 2026-06-29 17:19:13 a new tarball for the whole system https://alpinelinux.org/downloads/ 2026-06-29 17:19:27 yes, either you have to rewrite the boot medium 2026-06-29 17:19:37 frag: oh lol, I didn't even think of that 2026-06-29 17:19:38 but there is also an update-kernel script, but that one may require quite some memory 2026-06-29 17:20:03 ah, I just remount / as RW and unpack it? 2026-06-29 17:21:24 that is not your boot medium 2026-06-29 17:21:44 boot medium is usually on /media/... 2026-06-29 17:22:44 basically install the system from scratch, with new tar-ball 2026-06-29 17:23:24 ah, yes 2026-06-29 17:23:50 I will give it a try, thanks! 2026-06-29 17:24:15 good luck! 2026-06-30 04:35:18 ivanpollak: wait, asahi mkinitfs issues? 2026-06-30 04:35:26 well 2026-06-30 04:35:29 i dealt with that a few days ago 2026-06-30 04:35:40 https://codeberg.org/runxiyu/kasahi/raw/branch/main/mkinitramfs 2026-06-30 17:21:29 helloo 2026-06-30 17:46:48 damn, setup-disk makes things really easy 2026-06-30 17:47:48 and also i was able to go through the whole installation and my computer didnt over heat 2026-06-30 17:47:51 woo 2026-06-30 17:48:51 not bad! 2026-06-30 19:51:16 Hello, i am on alpine 3.24.1 and i am affected by this bug: https://bbs.archlinux.org/viewtopic.php?id=313014 2026-06-30 19:52:28 Is this fixed on edge? If yes, is it planned to be backported to 3.24? If not, should i report the bug? 2026-06-30 19:53:17 The issue is that glycin reports: Image type "legacy-xpm" is not supported 2026-06-30 19:53:38 It doesn't say what version of gdk-pixbuf2 it's fixed in 2026-06-30 19:55:12 From https://gitlab.gnome.org/GNOME/glycin the XPM format is supported by the glycin-image-rs package, which i have installed. 2026-06-30 20:01:40 2.44.6 and 2.44.7 include the fix, but "-Dlegacy_xpm=enabled" also needs to be added 2026-06-30 20:04:09 Looking at glycin work_items shows a couple bugs related to xpm loading as well. Just a cursory look, but a fix might involve both packages 2026-06-30 20:11:38 vicencb: we have 2.44.7 but indeed miss the flag 2026-06-30 20:12:44 would be happy to enable it if someone wants to file a issue or a MR, but now i gotta sleep :p 2026-06-30 20:23:43 I am preparing an MR for adding the '-Dlegacy_xpm=enabled' to gdk-pixbuf 2026-06-30 20:30:13 !104723 2026-06-30 20:34:03 vicencb: Thanks for reporting and for sharing the research you have already done 2026-06-30 20:47:40 `doas update-conf` consistently overwrites my changes to /etc/init.d/greetd, with "Autoupdating unchanged" 2026-06-30 20:49:45 I know this occurs because `is_modified` returns 1, and this happens because 'lbu status -a' doesn't list it 2026-06-30 20:50:53 and that happens because `apk audit --backup` doesn't list it 2026-06-30 20:51:06 but why that is... I have no clue. 2026-06-30 20:51:14 roselandgoose: It sounds like you are using the '-i' switch with update-conf, which is the only thing that will autoupdate in that script that I am aware of. 2026-06-30 20:51:24 except.. I'm not 2026-06-30 20:51:48 and if I was, it would just say "Autoupdating" 2026-06-30 20:52:00 (the unchanged bit is specific to the else case) 2026-06-30 20:52:08 Ah, I see it now, at the end of the script. 2026-06-30 20:52:15 yeah :) 2026-06-30 20:53:16 so I know the file is owned by greetd-openrc 2026-06-30 20:53:32 and I can see the entry in /lib/apk/db/installed for it 2026-06-30 20:53:42 I think there needs to more logic to that else case. It is weird, because I have never run into that myself 2026-06-30 20:54:12 I think update-conf is fine, if is_modified works correctly 2026-06-30 20:54:29 but something weird is happening with the apk-audit that gets called 2026-06-30 20:54:58 Thanks @jvvv for !104723. Any ideas on when that will reach 3.24? 2026-06-30 20:55:30 do you use a diskless, data-disk, or system-disk install? jvvv 2026-06-30 20:56:41 vicencb: It has passed CI, so when ever someone who has commit rights gets to it. I bet achill will look at it when wakes up 2026-06-30 20:57:25 roselandgoose: I use sys, and I think that it looks like that is the difference, after looking over is_modified function 2026-06-30 20:57:27 ok, thanks 2026-06-30 20:57:34 vicencb: yw 2026-06-30 20:58:18 I'm on a system desk install as well 2026-06-30 20:58:30 so that won't explain why you've never seen it.. 2026-06-30 21:00:47 roselandgoose: Well, now that I think about it, I have never run into a case where it prompts me for initd changes, so I am thinking that I AM effected, I just never thought about that... I try not mess directly with the initd files but rather try to use confd or other configurations 2026-06-30 21:01:41 I try to avoid initd changes too, but in this case it was the simplest/most-appropriate thing to do 2026-06-30 21:02:37 roselandgoose: One suggestion would be to make a copy of the greetd initd file to another name ( etc/init.d/greetd_mine ?) and modify that instead 2026-06-30 21:03:13 that would certainly be a fine work around 2026-06-30 21:03:24 then 'rc-update del greetd' and 'rc-update add greetd_mine' or whatever you name it 2026-06-30 21:04:22 yeah 2026-06-30 21:04:37 are you usr-merged ? 2026-06-30 21:05:31 On this box, yes, though I have some regrets that I won't mention since they preference and not policy objections per se 2026-06-30 21:06:00 hm, interesting 2026-06-30 21:06:35 my suspiscion is that apk-audit not working properly here will have something to do with that 2026-06-30 21:07:07 Yes, it has been brought up that usr-merge screws up apk-audit 2026-06-30 21:07:28 Or it did last I heard anyone talk about it 2026-06-30 21:07:44 yeah, but even weirder... this machine isn't usr-merged 2026-06-30 21:07:57 lemme see if I can reproduce on a laptop that is... 2026-06-30 21:09:59 hm, okay so that theory isn't supported atm. my usr-merged laptop shows the same thing: `apk audit /etc/init.d` doesn't list modified files 2026-06-30 21:10:16 although it does list added files (agetty.ttyx links) 2026-06-30 21:10:59 I'm gonna hop over to the -dev channel