2025-07-01 06:34:22 for any alpine developers around: is version 3 of apk-tools considered stable enough to use? 2025-07-01 06:36:47 It's going to be merged to edge today, (reminder to ncopa), so I guess so 🤷 2025-07-01 06:44:00 good morning! 2025-07-01 06:45:44 good morning :) 2025-07-01 06:46:45 PureTryOut: wasn't edge renamed to stable? i recall reading a commit for that 2025-07-01 07:08:25 Lol no, just linux-edge to linux-stable, which is only the name of a kernel package 😉 2025-07-01 07:10:43 got it 2025-07-01 08:34:48 yes today is the day we merge apk-tools 3 2025-07-01 08:35:29 linux-edge linux-stable. zfs-stable. those are doomed to be misunderstood 2025-07-01 08:36:14 i think linux-mainline might have been a better name 2025-07-01 08:36:23 naming is hard 2025-07-01 08:42:15 Well linux-edge is removed. linux-stable is the same stable as kernel.org names it, just as linux-lts 2025-07-01 08:42:41 linux-edge would basically be linux-mainline but it had a different kernel config than the other kernels which was confusing and ultimately it got removed 2025-07-01 08:43:36 ncopa: re the breeze-icons build failures on x86 and armv7, it seems the problem is the memory usage that 32-bit allows. Any clue how to solve it or do I have to disable the package on that arch? 😢 2025-07-01 08:53:18 is there a way to put a regular service with an init file under some sort of supervisor/daemontools kind of thing? If yes, what is it called? 2025-07-01 08:53:48 service packagename status does report the correct state so it shouldn't be too complicated i'm thinking 2025-07-01 08:54:37 PureTryOut: so it runs out of addressable memory space? what is running out of memory? the compiler? 2025-07-01 08:55:00 might be possible to disable debugging, reduce symbols includeded, or disable LTO or similar 2025-07-01 08:56:32 > virtual memory exhausted: Out of memory 2025-07-01 08:56:42 It's barely any code on that package in the firstp lace 🤔 2025-07-01 08:56:51 *first place 2025-07-01 08:57:58 Actually, only seems to be x86 now 2025-07-01 08:58:06 report upstream 2025-07-01 08:58:40 Should I look at s6? runit? supervise-daemon? 2025-07-01 08:59:27 i think openrc supports supervise-daemon and s6 2025-07-01 09:00:51 ncopa: thank you 2025-07-01 09:11:27 PureTryOut: looks like they generate a cpp file with the image data embedded 2025-07-01 09:19:15 https://forums.gentoo.org/viewtopic-p-8855770.html?sid=1e08136b3f4b8ab1de7d20abd181c743 2025-07-01 09:27:16 PureTryOut: report it here: https://invent.kde.org/frameworks/breeze-icons/-/issues/ 2025-07-01 09:29:17 PureTryOut: this looks relevant https://invent.kde.org/frameworks/breeze-icons/-/commit/ddac172cba33db77143b9df4e5ad32be4e19cff6 2025-07-01 09:36:20 PureTryOut: this works https://tpaste.us/5Qej but you should probably rename the patches or download them 2025-07-01 09:36:39 It has been fixed upstream already 2025-07-01 09:50:03 ncopa: linux-mainline seems to be a clearer name for it 2025-07-01 09:50:11 ncopa: given that the MR for linux-stable was open for months there was enough discussion time. also i think i wrote it pretty clear in the release notes: https://wiki.alpinelinux.org/wiki/Draft_Release_Notes_for_Alpine_3.23.0#linux-stable_replaces_linux-edge 2025-07-01 09:50:40 linux-mainline is torvalds' tree, e.g. the rc kernels. linux-stable follows the latest release from greg's stable kernels 2025-07-01 09:51:04 there is a difference see https://kernel.org/ or https://www.kernel.org/category/faq.html 2025-07-01 09:59:15 yeah i know. and understand the reasoning. it still ends up a bit confusing in alpine context. 2025-07-01 09:59:34 and its not guaranteed that linux-mainline would be less confusing 2025-07-01 09:59:46 so i dont know whats best 2025-07-01 10:00:46 how would it be confusing in the context of alpine? i'm not connecting the dots here 2025-07-01 10:02:29 hi, i'm hitting a bug with docker-engine 28.2.2 (alpine 3.22) fixed in 28.3, I did see an MR to upgrade the package but I was wondering if it would be ported to alpine 3.22 (because it's a bug fix but it's not a patch release) 2025-07-01 10:02:31 what would you think "zfs-stable" mean if you saw it first time? 2025-07-01 10:02:58 good point 2025-07-01 10:03:22 but as i said. i'm not sure zfs-mainline woudl be any better 2025-07-01 10:03:22 my first thought would be the latest release 2025-07-01 10:03:25 teapot9: likely yes, but have to ask tomalok first 2025-07-01 10:04:11 teapot9: you can always create an issue if you havent already, and we can find a solution for it 2025-07-01 10:04:13 mebious_: but zfs-lts is just as confusing i guess, given that also zfs exists 2025-07-01 10:04:20 exactly 2025-07-01 10:04:20 though i guess it would be vague if the project doesn't publish releases in mulitple branches 2025-07-01 10:04:54 so im fine with keeping it as it is. changing it now will likely not make things less confusing 2025-07-01 10:05:14 i also though that linux-edge was a confusing name 2025-07-01 10:05:26 as i said: naming is hard 2025-07-01 10:05:41 yeah 2025-07-01 10:06:17 i mean now we have at least a compiled zfs for the stable kernel :p 2025-07-01 10:06:56 i only just registered that in my brain when you mentioned it 2025-07-01 10:07:09 achill ncopa: thanks, i'll create an issue and link it to the MR 2025-07-01 10:42:24 what about this crazy idea of calling the package... 'linux' :-) 2025-07-01 10:42:42 simply dropping the extra '-' 2025-07-01 10:43:03 ACTION _thinks_ that's what arch does, but he might be wrong 2025-07-01 10:43:57 yep, arch seems to be using 'linux' and 'linux-lts'. but I never used arch so I may be wrong 2025-07-01 10:45:09 that wouldve been even more consuing since the default e.g. in the images is linux-lts 2025-07-01 15:15:17 does librewolf need dbus? getting "lw already running but doesnt respond" when trying to open a url with `librewolf my.url` from a terminal (and its indeed already open and working) 2025-07-01 16:17:53 neat, apk tools 3 default in edge now! 2025-07-01 16:43:45 😱 apk3 tells me what packages are not available in a repo anymore. That's such a much needed feature! 2025-07-01 16:47:08 Oh, sick! 2025-07-01 16:47:35 Hello, where can i contact maintainers of apk-tools ? I'm encountering difficulties at understanding the .apk v3 format. 2025-07-01 16:48:00 #alpine-devel or https://gitlab.alpinelinux.org/alpine/apk-tools 2025-07-01 16:55:29 "#alpine-devel or https://gitlab..." <- I headed towards the channel thanks. I doubt the gitlab is an appropriate to ask questions. 2025-07-01 16:56:45 or even #apk-tools on oftc :p 2025-07-01 16:58:39 achill: that's even better thanks 2025-07-01 17:03:59 i don't understand something about apk. why does info have have a sub commands like apk info --print-arch but you can just do apk --print-arch, v3 does the same thing. 2025-07-01 17:04:15 what does info do? 2025-07-01 17:34:32 userdocs: --print-arch is a global option 2025-07-01 17:34:38 userdocs: that means it ignores the subcommand 2025-07-01 17:34:57 It's an option that acts as a command 2025-07-01 17:39:28 ok, i understand now. but let's say you are just looking for a global, you have to do apk somethign -h? 2025-07-01 17:45:37 apk -h 2025-07-01 17:45:56 oh, that just shows the commands 2025-07-01 17:46:05 but yes, every subcommand also has a section for global options 2025-07-01 17:46:25 maybe it's me i feel like global should 1: have it's own thing / be shown by apk -h and/or be under the info/list options. 2025-07-01 17:48:36 coz i can do this and it make no sense. apk upgrade --print-arch 2025-07-01 20:32:02 Plasma 6.4 in Alpine seems to not be able to theme the Aurorae anymore, it seems like a similar problem to https://forum.artixlinux.org/index.php/topic,8316.0.html 2025-07-01 20:32:30 https://lounge.swee.codes/uploads/8c07c60fc7fc75f4/image.png 2025-07-01 20:34:33 On that forum topic it says they fixed it with reinstalling Aurorae, but I have no idea what is the equivalent alpine package for that 2025-07-01 20:37:09 https://gitlab.alpinelinux.org/alpine/aports/-/issues/17324 2025-07-01 20:37:34 achill: oh cool, it's already reported 2025-07-01 20:37:41 yeah 2025-07-01 20:43:01 i can try to package it, but have no way to verify if it actually works 2025-07-01 21:33:53 go ahead, would love to test if that works :3 2025-07-01 21:35:48 feel free: https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/86406, you could use mrtest or download the CI artifacts 2025-07-01 21:58:13 frag, LW seems to use dbus to check if there is a running instance of itself. Im not running dbus and I cant seem to be able to spawn more LW processes without using a different profile. So you might need dbus to be able to call "librewolf example.com" and have it open in your already open browser. 2025-07-01 22:02:30 as a workaround there is --no-remote cli option for ignoring already running instances 2025-07-01 22:05:07 Is OpenPAM available through Aports by some name? 2025-07-01 22:23:28 achill: mrtest seems to use PostmarketOS gitlab instead of AlpineLinux gitlab 2025-07-01 22:23:42 you can add -a 2025-07-01 22:23:42 $ mrtest add 86406 2025-07-01 22:23:43 urllib.error.HTTPError: HTTP Error 404: Not Found 2025-07-01 22:23:43 Download https://gitlab.postmarketos.org/api/v4/projects/postmarketOS%2Fpmaports/merge_requests/86406 2025-07-01 22:23:48 achill: ah ok 2025-07-01 22:23:52 so mrtest add -a 86406 2025-07-01 22:24:57 done, restarting Plasma 2025-07-01 22:28:45 achill: https://lounge.swee.codes/uploads/861ee9c86e7cf921/image.png can confirm this fixes the issue! 2025-07-01 22:29:06 nice! thanks for testing 2025-07-01 22:34:20 I should also post that on the MR 2025-07-01 22:52:43 What's the package for the APKBUILD linter that the GitLab Pipeline uses? 2025-07-01 23:01:46 I think you mean atools-go 2025-07-02 09:02:02 tried to upgrade my server today (uses edge), and i got this nasty conflict around postgresql-bdr, a package i dont even have installed (apk del postgreql-bdr just returns OK): https://tpaste.us/Kxkg 2025-07-02 09:08:34 manually selecting the postgrelql version in world fixed it, but having this meaningless conflict is very strange 2025-07-02 11:26:25 Might be related to the fact that postgresql-bdr has a versioned postgresql provider but other postgresql providers are virtual (without version). This has technically never been supported but it is still used a lot in aports. 2025-07-02 17:37:11 I am trying to pxe boot a alpine kernel and initrd on a uefi system. 2025-07-02 17:40:51 Apologies, hit enter instead of tab, new "glove80 split keyboard, retraining my homerow magic" am getting "BdsDxe: failed to start Boot0001 "UEFI PXEv4 (MAC:5254006FF160)" from PciRoot(0x0)/Pci(0x2,0x0)/Pci(0x0,0x0)/MAC(5254006FF160,0x1)/IPv4(0.0.0.0,0x0,DHCP,0.0.0.0,0.0.0.0,0.0.0.0): Load Error" which makes me think the alpine image I built doesnt have support to boot from EFI. It works on a system using bios 2025-07-02 17:41:40 I'm testing against a qemu / kvm / virt-manager machine using /usr/share/edk2/x64/OVMF_VARS.4m.fd boot rom 2025-07-02 17:42:35 if you put that same kernel on a disk attached to the VM and boot it, does it work? 2025-07-02 17:45:13 hello, i have a little request, if there is someone running ZNC, could you compile this module for me https://paste.parinux.org/paste/996Tofn1#qN9qfr5LcYyzZ8UFnUGbCEsKR16Xz5InqV9jFbrKURB as I am unable to do so due to hw limitations and znc on Alpine doesnt come with znc-buildmod 2025-07-02 17:46:08 bon_: apk add znc-dev 2025-07-02 17:46:54 https://pkgs.alpinelinux.org/contents?file=znc-buildmod&path=&name=&branch=edge&repo=&arch= 2025-07-02 17:47:19 ^ exactly 2025-07-02 17:47:30 I am building the image using linuxkit. Using the Linuxkit run command they both boot. "/usr/bin/qemu-system-x86_64 -smp 1 -m 1024 -uuid ab281c0a-d135-42bc-830b-b720d1c76ca8 -pidfile linuxkit-state/qemu.pid -machine q35,accel=kvm:tcg -object rng-random,id=rng0,filename=/dev/urandom -device virtio-rng-pci,rng=rng0 -drive if=pflash,format=raw,file=./OVMF_CODE.fd -kernel linuxkit-kernel -initrd linuxkit-initrd.img -append 2025-07-02 17:47:30 console=tty0 console=ttyS0 console=ttyAMA0 -device virtio-net-pci,netdev=t0,mac=de:f9:8a:89:de:c7 -netdev user,id=t0 -nographic" 2025-07-02 17:48:55 dwfreed ikke thank you :D i havent noticed it needs a wildcard on search :) 2025-07-02 17:49:12 there are no wildcards in that search 2025-07-02 17:49:44 Why is there only QRCode *encoding* tools and no decoding ones (besides zbarimg)? 2025-07-02 17:50:16 because there aren't a lot of decoding tools in the first place, and one generally doesn't use linux to decode them 2025-07-02 17:50:20 i mean i just looked for znc that only shows znc package 2025-07-02 17:50:53 while znc* also shows znc-dev 2025-07-02 17:51:45 dwfreed, one needs a specific operating system in order to decode a QRcode? 2025-07-02 17:52:04 Windows maybe? 2025-07-02 17:52:04 feel free to write a decoder for linux 2025-07-02 17:52:16 zxing makes a good lib for it, I hear 2025-07-02 17:52:19 quinq: no, it's more about the formfactor of the device you are using generally 2025-07-02 17:52:36 wat 2025-07-02 17:53:53 except for special usecases, you probably won't use a server, desktop or even most laptops to scan QR codes 2025-07-02 17:54:28 i saw someone at congress manually decoding a QR code by hand 2025-07-02 17:54:34 or was it drawing? idk 2025-07-02 17:55:01 Ah, zxing works, thanks dwfreed 2025-07-02 17:55:27 Ah, indeed 2025-07-02 17:55:34 Scanning is a different matteer 2025-07-02 17:55:36 matter 2025-07-02 17:55:52 I'm talking about decoding, the QRcodes are already available as regular images 2025-07-02 17:57:41 zxing-2.3.0-r1 description: 2025-07-02 17:57:41 C++ port of ZXing 2025-07-02 17:58:03 Maybe not the best description when searching for code processing :D 2025-07-02 17:58:23 yeah... 2025-07-02 17:59:00 Anyway, thanks again for unblocking me! 2025-07-02 18:14:31 characters that can be encoded as ASCII are, even if there is UTF-8 chars in the text file? 2025-07-02 18:15:08 frag: utf-8 is the same as ASCII for characters less than 128 2025-07-02 18:15:33 oh 2025-07-02 18:17:21 UTF-8 was designed on purpose that way 2025-07-02 18:17:25 yes 2025-07-02 18:18:02 very nice, i guess there is not much purpose in keeping text ascii vs utf-8 2025-07-02 18:18:15 except some old tools cant handle utf8 2025-07-02 18:18:42 if it's all characters below 128, there is no difference (also technically ascii doesn't define characters greater than 127) 2025-07-02 19:00:29 The `llvm20` package contains `llvm-ar`, but previous versions such as `llvm19` do not contain their respective versions of llvm-ar (such as llvm-ar-19). Is this intentional? 2025-07-02 19:00:52 We build our game using Clang 19 and I would prefer not to mix Clang 19 with llvm-ar 20 2025-07-02 19:03:41 /usr/bin/llvm19-ar 2025-07-02 19:03:45 is provided by llvm19 2025-07-02 19:03:54 https://pkgs.alpinelinux.org/contents?name=llvm19&repo=main&branch=edge&arch=x86_64 2025-07-02 19:05:45 Ah, I see, apologies. I assumed that the binary name would match Fedora. 2025-07-02 19:07:50 That is slightly unfortunate, because it means I can't use the same environment configuration to build on my host machine (fedora) and inside an alpine distrobox. I will have to solve that with a symlink. 2025-07-02 20:53:43 im using defaul crond on alpine, how can i specify the user to which a cron job applies? 2025-07-02 21:03:46 i mean using the /etc/periodic/ .. but i might just skip it 2025-07-02 21:14:00 Use su within to change user 2025-07-02 21:14:05 Otherwise, there's crontab 2025-07-03 03:45:24 Anyone else using apkv3 feeling like it is slow compared to the prior version? 2025-07-03 03:45:38 Intel N3450, 4GB RAM, 4GB zswap, 64GB eMMC 2025-07-03 03:45:50 Not apt slow, but certainly noticeably slower 2025-07-03 03:46:18 4044 packages 2025-07-03 03:47:25 feels faster to me 2025-07-03 03:47:45 Does it nail a single thread while solving? 2025-07-03 03:48:03 N3450 performance in single-thread is pretty abysmal, but when all four cores kick in, it is tolerable 2025-07-03 04:40:57 apkv3 executes commands faster on my end 2025-07-03 05:01:55 Yeah, nah, many times slower here 2025-07-03 05:01:58 That isn't fun 2025-07-03 05:02:11 I wonder how to profile what is making it take so much longer 2025-07-03 05:40:15 did you get that harware from the garbage bin? 2025-07-03 05:40:25 heh 2025-07-03 06:19:20 A lot of people throw away perfectly working hardware, to replace it with the newest ultimate gamer mining alien riced consumable 2025-07-03 07:44:17 i thought the whole point of alpine was to be able to run in it on your smart toaster 2025-07-03 07:53:24 "i thought the whole point of..." <- i run it on my smart toothbrush, does that count ? 2025-07-03 07:53:54 "i thought the whole point of..." <- I want AI to cook my toast, not make deliriously assembled code or art. 2025-07-03 07:57:02 smart toothbrush does count and ai toaster sounds like a good idea. 2025-07-03 07:58:36 https://www.digikey.com/en/maker/projects/how-to-build-an-ai-powered-toaster/2268be5548e74ceca6830bf35f0f0f9e 2025-07-03 10:35:37 hi all 2025-07-03 10:44:35 @swfrd_ welcome to ignore 2025-07-03 10:47:10 @userdocs Alpine the OS runs flawlessly as ever. apkv3, the package manager, I am finding significantly slower to solve vs prior apk tools, since v3 became default in edge recently 2025-07-03 10:55:20 Saijin_Naib[m]: i understand, not really sure how what you can do to narrow down the cause though. 2025-07-03 11:15:32 Yeah... I want something to bring before I raise an Issue on the tracker, you know? 2025-07-03 11:16:16 Asking if others have slow solve and looking for commonalities seems like a decent start 2025-07-03 11:24:38 Any Cashapp? Chime? Zelle? Btc? Usdt?Skrill? Apple Pay? Pay pal? Venmo? BOA? Wells Fargo? Join the channel and got paid 💯📌 2025-07-03 11:24:38 https://t.me/+Cnk97ig86vUzNWZk 2025-07-03 11:40:50 I only have pen pineapple apple pay 2025-07-03 11:43:33 sell the pen and eat freuits salad? 2025-07-03 11:51:33 hi all back ... . 2025-07-03 12:38:35 alpine-standard-3.22.0-x86_64.iso doesn't boot using ventoy 2025-07-03 14:23:58 Anyone using niri compositor? 2025-07-03 14:50:50 @fission thanks, that'll be on repeat all day now 2025-07-03 17:28:12 hello, how can i get python3-certbot-dns-duckdns on alpine? is it part of some package? 2025-07-03 17:34:17 doesn't look like it 2025-07-03 17:42:25 ok, installed it with pipx 2025-07-03 21:30:14 hello :) 2025-07-03 21:31:24 After an update, I have this weird issue where `uname -r` says "6.12.33-0-lts" but the installed linux-lts is 6.12.34-0-lts. So e.g. all the modules in /lib/modules are for 6.12.34 and my running kernel doesn't want to load them 2025-07-03 21:31:42 did you restart? 2025-07-03 21:32:06 The thing is, `file /boot/vmlinuz-lts` says "version 6.12.34-0-lts 2025-07-03 21:32:13 pj: yes, I restarted many times now 2025-07-03 21:32:47 Of course I had not enabled the package cache so I can't easily reinstall linux-lts (no network interfaces, can't mount a vfat USB stick, etc) 2025-07-03 21:33:42 I ran `mkinitfs -c /etc/mkinitfs/mkinitfs.conf -b / 6.12.34-0-lts` and the command succeeded, but a reboot didn't change anything 2025-07-03 21:34:52 I checked my /boot/grub/grub.cfg and there the linux command refers to what I expect: `linux /alpinelinux/root@/boot/vmlinuz-lts root=ZFS=rpool/alpinelinux/root ro modules=[...]` 2025-07-03 21:36:09 findmnt -T /boot shows "/ rpool/alpinelinux/root zfs rw,relatime,xattr,posixacl,casesensitive", which seems fine 2025-07-03 21:37:00 and `apk info linux-lts` shows linux-lts-6.12.34-r0 2025-07-03 21:37:19 So really, it seems like 6.12.34 was installed, and I don't get how my system still loads 6.12.33 2025-07-03 21:38:22 could I try to re-run dracut or something? Would it make sense? I'm always confused with dracut and what exactly it does 2025-07-03 21:41:50 Oh: in /var/log/dmesg I only see references to 6.12.33-0-lts, nothing after. 2025-07-03 21:46:07 is your /boot mounted correctly? 2025-07-03 21:48:18 longnoserob[m]: I think it is, but I'm not sure how to be perfectly sure 2025-07-03 21:50:16 longnoserob[m]: or maybe it is not because I cannot mount any vfat partitions because I'm lacking the module 2025-07-03 21:52:00 I was hoping to force modprobe modules (it's only a patch update: 6.12.33 to 6.12.34, but modprobe --force gives me "Key was rejected by service" 2025-07-03 21:57:29 hmm 2025-07-03 21:59:28 yeah... 2025-07-03 22:07:05 what I don't get is that in /boot I have the new kernel. So how does it load the old kernel in the first place? 2025-07-03 22:07:14 I can't find it on my system 2025-07-03 22:18:20 naib: I think I noticed some slowdowns too. It would probably be nice to write some tool which checks performance (and size) across the git history to track such issues 2025-07-03 22:28:02 hi there 2025-07-03 22:29:17 Would you be able to share some working socks5 proxy in the US? 2025-07-03 22:29:51 wrong channel i think 2025-07-03 22:29:57 I'd like to watch a movie on YI but it is allowed only for US 2025-07-03 22:30:34 You are clever guys - you know many tips and tricks 2025-07-03 22:31:11 offtopic anyway 2025-07-03 22:32:28 maybe some of you from US can share your VPN - just asking... 2025-07-03 22:32:30 never mind 2025-07-03 22:33:12 have a great day! bye 2025-07-04 00:42:26 Saijin_Naib[m]: I'm here to help :) 2025-07-04 02:05:11 The gnome-keyring service file uses $KEYRING_COMPONENTS - where is this specified / can I change it? I would like to use it for ssh. 2025-07-04 02:23:20 @sertonix can you give me a basic summary of your system stats and package count? 2025-07-04 02:30:21 hey. I'm trying to use rustup-installed rustc on e.g. Alpine 3.22/amd64, and am consistently getting segfaults compiling stuff. Here's a gist with the build failure, and gdb bt: https://gist.github.com/thresheek/c7a5bb3852e197e16304aa2ef585d933 2025-07-04 02:31:24 It works fine with Alpine-3.22-provided rust/cargo, though. The problem is I build on older Alpines too (3.19 and 3.20), and the rustc shipped in those versions is too old for the code I need to build - hence I would have liked to use the rustup-provided rust toolchain. 2025-07-04 02:32:06 any idea if the issue I have could be fixed by e.g. providing some build flags? I'm totally new to rust world. Thanks 2025-07-04 02:44:15 thresh: I don't know much about rust (or anything, really :) but it looks like the APKBUILD file is a bit involved and several patches are applied, even for the edge builds 2025-07-04 02:47:43 fission, Yeah, that's what I've checked as well :) I guess my question is "why rust-lang ships broken toolchains for musl"... 2025-07-04 02:48:18 but maybe there's fix end-user like me can apply.. 2025-07-04 03:06:38 yeah good question 2025-07-04 03:10:28 I can't tell really, just eyeballing it, though alpine-target.patch seems like a possible candidate 2025-07-04 03:10:40 (for making it compile in a more alpine-compatible way) 2025-07-04 09:08:45 Hello, How much storage is needed to build qt6-qtbase package ? 2025-07-04 09:27:19 ! LaTeX Error: File `troll.cls' not found. 2025-07-04 09:35:20 maduncle94[m]: including os and systems packages? 2025-07-04 12:31:24 Hello, I am new around here, let me know if this is not appropriate place to ask :-) 2025-07-04 12:32:01 I have a problem with kernel drivers ath12k from latest upstream (vanilla?) kernel. Where should I report the issue ? 2025-07-04 12:32:16 1^ 2025-07-04 12:34:20 2. What system container do you recommend that support host file system ala bwrap and systemd-nspawn? 2025-07-04 12:34:22 I need to run a initrc within the container and that was poorly supported by docker (and podman?), and is not supported by bubblewrap 2025-07-04 12:34:50 by the way thanks for the awesome distro :happy: 2025-07-04 12:35:34 1. late search-fu gave me this: https://wireless.docs.kernel.org/en/latest/en/users/drivers/ath12k.html 2025-07-04 12:39:56 What init system do you ahve in mind within the container? I.e. systemd, openrc or something else? 2025-07-04 12:43:55 systemd 2025-07-04 12:44:22 openrc if glibc is available 2025-07-04 12:44:59 re bwrap as system container is 'not planned', ref: https://github.com/containers/bubblewrap/issues/668 2025-07-04 12:49:25 lxc/incus? 2025-07-04 12:50:18 naib: I found the commit and and created apk-tools#11126 2025-07-04 12:50:18 with lxc (and probably with incus) I can't boot from bare host fs (i do not know how to do it) 2025-07-04 12:54:33 I will try the documentationn https://linuxcontainers.org/incus/docs/main/reference/storage_dir/ 2025-07-04 12:54:39 that is what i nneed 2025-07-04 13:05:02 Is openrc for running sshd? or is there other services you expecting to run? 2025-07-04 13:06:51 I want to avoid using docker, and docker-compose 2025-07-04 13:07:16 at least try to avoid 2025-07-04 13:08:33 Are you trying to avoid containers but trying to use the features directly that they are built upon? 2025-07-04 13:09:47 I want to have all the setup locally in one container or vm. I prefer a system container because they are said to more lightweight. The reason having everything in a single container is that it does not require new tooling to track errors across services 2025-07-04 13:10:54 I read that is a lost cause, and that big tech use borg et al. and that I should embrace containers for security reasons. 2025-07-04 13:11:19 ease of use of docker, docker-compose, setups is far from ideal. 2025-07-04 13:11:22 ime 2025-07-04 13:12:29 Indeed, to me it sounds like you aren't interested in the benefits of containers and you would almost be better off with chroot for 'having a second system'. 2025-07-04 13:13:49 Years ago I would use Gentoo from a chroot inside Ubuntu, as Gentoo provided better tooling at the time for cross compiling to ARM. 2025-07-04 13:14:18 another reason to have that setup, is that I can spin an unrelated instance of the same project without again more tooling or configuratino because the system container isolate every instannce of the project 2025-07-04 13:14:45 chroot does not work because ports are shared with the host ? 2025-07-04 13:16:17 Indeed, it doesn't cover network namespaces which is what allows you to have processes on the same machine that both think they are talking on port 8080. 2025-07-04 13:17:20 So is your idea that you have one big container per project? or you have essentially one big base image that has all your favourite tools etc for running with different projects? 2025-07-04 13:20:52 sort of yes 2025-07-04 13:21:08 it is not an image, it is directory in the host file system 2025-07-04 13:23:31 let's try something else donno, how do you setup local dev environment ? 2025-07-04 13:25:30 I develop for Windows :) so perhaps not as helpful. 2025-07-04 13:26:16 you develop product relying on microservices? do you spin several services / webapps on your developmennt machine or vm ? 2025-07-04 13:29:14 I did work on a app that had a web server which I would spin up on the development machine. 2025-07-04 13:30:36 But sure, I will humour you and say, yes I've since moved to doing "microservices" (in scare quotes because it may be multiple processes but they not really match what microservice are 2025-07-04 13:32:09 For that the parts, run in the containers however some small parts can run on outside (and some parts don't), not that there is much point to that. 2025-07-04 13:34:09 As for your situation, I would suggest you take a look at runc and crun to see if they could help build your dream system. Essenitally with runc you essenitally have a directory (which is the rootfile system) and you can run a container (i.e isolated namespaces, cgroups etc) from that without the the conviences provided by Docker/Podman/nerdctl 2025-07-04 13:38:42 good call, I will again at runc 2025-07-04 13:39:33 If you are curious, here is a minimal example: https://tpaste.us/gbvy 2025-07-04 13:47:48 @sertonix thank you, I'll put further info there! 2025-07-04 15:29:49 I'm trying to figure out why py3-lsp-server doesn't format my files in neovim with :lua vim.lsp.buf.format(); using a pip-installed python-lsp-server in a venv it does. I assume it's missing hooking into the plugins for formatting like `yapf`? How would I get that to work? 2025-07-04 15:47:53 Donno, based on your instructions a wiki page has been created. https://wiki.alpinelinux.org/wiki/Runc. please feel free to update it. Thanks 2025-07-04 16:59:54 I upgraded from 3.19 to 3.20 and somehow zfs kernel module got missing since that. I am stuck trying to rectify that problem with my 100M boot partition. 2025-07-04 17:03:04 mkinitfs says I don't have enough space on the partition and I do not know what files I must keep or what are the alternative routes to a booting system that would mount my encrypted zfs volumes 2025-07-04 17:03:57 any ideas on how to proceed since I don't think deleting files nor trying to resize the boot partition helps at all without worsening the current situation? 2025-07-04 18:43:42 what's alpine's .bashrc equivalent for ash? i think it used to be .profile but it doesn't seem to be automatically sourced right after login anymore 2025-07-04 18:50:52 el[m]1, ash will use what's set by ENV 2025-07-04 18:53:11 I made this a while ago automatically load my aliases: 2025-07-04 18:53:30 ACTION sent a doas code block: https://matrix.org/oftc/media/v1/media/download/AZOhuzWlfWcfmlF9ZL8-5Vq_c67Bt6Vf16HS7Aa_BI8jQ7MlUhfuhxvUVTRAsoZ02nFR5qtZ457m9QJtmeyu_YBCeYHuJm2gAG1hdHJpeC5vcmcvVkdMQWxpdHhYY3VqYWpXVGlZSmJwS1Fo 2025-07-04 18:53:42 and just added all my aliases to ashrc 2025-07-04 18:54:03 I don't think that pasted properly 2025-07-04 18:54:50 ACTION sent a code block: https://matrix.org/oftc/media/v1/media/download/AaArm5pyNcWNk3buufdUiNys9HkGDu0PPvD-hXt9jz8u7mnvCCDfgp_nKtfTvNpKU-pCtsQVR1uh_iImwwxaCB1CeYHuOclQAG1hdHJpeC5vcmcvQ1lwWXZBSnRVSEpIU2Z5QWt3ZWptaUF4 2025-07-04 18:56:30 looking at it now, I probably don't need doas in there 2025-07-04 19:00:04 taterspie[m]: .profile is sourced only by login shells 2025-07-04 19:14:16 Ah okay, I did a bit more reading and I think understand the original question better now 2025-07-04 19:23:39 I managed to figure out chrooting into my encrypted zfs and ext boot partition with extended installation image on USB and then perform update and system now boots normally 2025-07-04 19:25:06 gg 2025-07-04 20:00:04 https://github.com/nomacs/nomacs/issues/1017 2025-07-04 20:00:14 can i fix it in alpine package? 2025-07-04 20:36:46 sure, why not ? 2025-07-04 21:18:59 jones: you could try running `upgrade-grub` and restarting 2025-07-04 21:19:20 i think i had an issue like that at some point which i fixed by doing the aforementioned 2025-07-04 21:46:00 Donno I was secretly dreaming of https://github.com/nestybox/sysbox?tab=readme-ov-file#using-sysbox a rootless system container 2025-07-05 03:47:46 I'm trying out zram-init for the first time, but when starting the service, I get 2025-07-05 03:47:47 zramctl: /dev/zram0: failed to set number of streams: No such file or directory 2025-07-05 18:13:54 is alpine sporting systemd? 2025-07-05 18:14:54 swfrd_: no 2025-07-05 18:28:12 swfrd_: one related project did that https://wiki.postmarketos.org/wiki/Systemd :) 2025-07-05 18:31:08 Hey there 2025-07-05 22:21:48 I am trying to compile llama-cpp for Alpine with Vulkan (it compiles fine without vulkan).... (full message at ) 2025-07-06 02:38:29 hi all cartel here ... . how are you all ... . 2025-07-06 09:01:57 Hi. Morning. 2025-07-06 09:02:13 hi 2025-07-06 09:02:18 Is there some changes in luks/cryptsetup from 3.21 to 3.22? 2025-07-06 09:03:52 I updated but now my luks key is not accepted, so I can't boot :( 2025-07-06 09:05:20 `git log v3.21.0..v3.22.0 cryptsetup/` returns nothing 2025-07-06 09:09:48 Maybe some luks key format/algo changed? 2025-07-06 09:11:47 what happens if you boot a live iso or attach the disk to another system and try to mount it? 2025-07-06 09:12:43 I will have to try that. I do have a snapshot (this is a vps) so I could probably/hopefully go back to that one. Trying a power off and power on now - but the vps is a bit slow to react 2025-07-06 09:14:25 just to be sure, maybe you also had just the uppercase case key enabled? 2025-07-06 09:15:05 i know that it did intrupt me a couple of times 2025-07-06 09:16:48 Yea. It worked after a full power cycle 2025-07-06 09:17:05 So probably something with keyboard setting not being correct 2025-07-06 09:23:56 Now I have a different problem :D 2025-07-06 09:29:57 a typo in fstab.. sigh. 2025-07-06 11:15:27 is autoconf not part of build-base? 2025-07-06 11:15:59 no. most projects don't need it. 2025-07-06 11:16:42 thanks, need to add it then maually 2025-07-06 11:47:57 Hey 2025-07-06 11:49:58 After upgrade my LXD container are not starting udev 2025-07-06 12:31:57 https://tpaste.us/9XxZ <-- That's what the log say's 2025-07-06 12:34:35 https://tpaste.us/nNb9 <-- That's better ;) 2025-07-06 14:45:38 alpine gitlab does not show git history but it shows "500: We're sorry, something went wrong on our end". Is that expected? 2025-07-06 15:11:22 on which page? creating of a MR? you need to rebsae your branch 2025-07-06 15:15:36 here are some random example, https://gitlab.alpinelinux.org/alpine/aports/-/commits/master/community/alembic?ref_type=heads and https://gitlab.alpinelinux.org/alpine/aports/-/commits/master/community/ali?ref_type=heads 2025-07-06 15:16:09 I thought it was due to reduce server loader 2025-07-06 15:16:13 *load 2025-07-06 15:51:50 oh yeah git log of a certian subdir also seems to be hard on the server 2025-07-06 15:53:25 I used github mirror to see the logs :( 2025-07-06 18:25:05 I'm having a bit of trouble getting xdg-desktop-portals to behave on Sway. If I run sway via `exec dbus-start-session sway`, then neither the gtk nor the wlr portals function. If I start outside of dbus with just `exec sway`, then the gtk portal works, but wlr still does not. What am I doing wrong? 2025-07-06 18:25:25 (in hindsight this might be a more generic question that would belong in sway's channel) 2025-07-06 18:26:59 (that said, my portals.conf is identical between fedora and alpine, and it works fine on fedora) 2025-07-06 18:51:43 Ok, if I DON'T run it in dbus then the steam flatpak doesn't function. So I have to run it in dbus. which breaks the portals... 2025-07-06 18:52:11 I will ask in the sway channnel 2025-07-06 19:28:57 does alpine have something to ask for packages? 2025-07-06 19:30:29 Biswa96[m]: alpine runs a cgit at https://git.alpinelinux.org/aports/ 2025-07-06 19:30:52 realroot[m]: what are you looking for, specifically? 2025-07-06 19:31:55 adding rocm for ollama https://gitlab.alpinelinux.org/milesalan/aports/-/blob/add-ollama/testing/ollama/APKBUILD 2025-07-06 19:37:52 realroot[m], i have the impression that by far the best way is to actually contribute it - if you can 2025-07-06 19:42:03 maybe i am scared of alpine being political 2025-07-06 19:46:12 realroot[m]: I would say that in general that Alpine does not really care about politics unless it's likely to cause problems for users, as was the case with e.g. Xlibre. 2025-07-06 19:46:27 what would be `depends="gcc-libs glibc"` in Alpine? `musl`? 2025-07-06 19:46:50 you would not need to specify. 2025-07-06 19:46:52 what problems can cause xlibre? 2025-07-06 19:48:11 first one is https://gitlab.archlinux.org/archlinux/packaging/packages/rocm-core/-/blob/main/PKGBUILD I think 2025-07-06 19:48:59 does `install -Dm644 rocm-ld.conf "$pkgdir/etc/ld.so.conf.d/rocm.conf"` make sense? 2025-07-06 19:49:31 realroot[m]: Xlibre, being a fork of Xorg, would impose an additional burden on the Alpine security team in order to ensure that all security patches by Xorg are present in Xlibre, that any security issues found in Xlibre are patched in a reasonable period of time, etc., with the added problem that Xlibre would necessitate dealing with behaviour which, were it to occur in Alpine 2025-07-06 19:49:33 community spaces, would be against Alpine's CoC. 2025-07-06 19:50:42 ld.so.conf is a glibc thing. 2025-07-06 19:51:18 From the Alpine Linux leadership this is about politics, not just software 2025-07-06 19:51:24 ` Open source is politics` 2025-07-06 19:52:44 okay, please go take a course on social psychology and report back about what 'apolitical' _actually_ means. :) 2025-07-06 19:53:17 (protip: any social interaction between two people, where one person has any kind of power over the other, is inherently political.) 2025-07-06 19:55:31 the decision to package a particular program is political, even without issues like Xorg vs Xlibre. the vast majority of the time no considerations need to be made about how upstream is likely to behave toward users, etc., and so the politics of the decision are invisible. 2025-07-06 19:57:34 anyway, the long & short of it is that you don't need to worry about it unless you have a reason to believe that upstream is likely to be problematic wrt Alpine or yourself. 2025-07-06 20:00:10 "ld.so.conf is a glibc thing." <- what would that be on alpine? 2025-07-06 20:00:48 there isn't a musl equivalent. 2025-07-06 20:02:11 content is /opt/rocm/lib where do I put it? 2025-07-06 20:02:41 ldconfig alternative was mentioned here https://wiki.musl-libc.org/faq 2025-07-06 20:05:38 so /lib/ld-musl-x86_64.path ? iiuc it should append it to the file how does apk do that? 2025-07-06 20:06:00 I do not have it here 2025-07-06 20:09:55 i think hook 2025-07-06 20:10:40 https://gitlab.alpinelinux.org/search?search=*hook*&nav_source=navbar&project_id=1&group_id=2&search_code=true&repository_ref=master 2025-07-06 20:23:55 does musl have a room? 2025-07-06 20:27:52 there is mailing list and irc for musl support https://musl.libc.org/support.html 2025-07-06 20:31:01 realroot[m]: yes, you should not need to poke it. 2025-07-06 20:46:20 realroot[m], if you're worried about rocm being politically troublesome, the difference between "asking for a package" and "contributing a package" is not the important distinction 2025-07-06 21:29:15 Hi 2025-07-06 21:31:02 Is it possible to cache apk packages with squid caching proxy? 2025-07-06 21:32:46 or a docker registry kind of alpine package repository registry? 2025-07-06 22:09:01 https://wiki.alpinelinux.org/wiki/How_to_setup_a_Alpine_Linux_mirror#Partial_mirror_using_nginx 2025-07-06 22:09:12 idk how up to date the article still is 2025-07-06 22:12:58 realroot[m], I was just stated a fact, and I'm not in the “Alpine Linux leadership” 2025-07-06 22:13:02 stating 2025-07-06 23:20:56 @realroot:nadeko.net I think the binaries that use libraries from /opt/rocm/lib should just have that path in their rpath. Putting it somewhere else means all binaries will try to load libraries from that directory. Depending on the contents of /opt/rocm/lib they could also rust be moved to /usr/lib 2025-07-06 23:21:24