2026-08-01 12:12:30 Could someone close this: v 2026-08-01 12:12:31 https://gitlab.alpinelinux.org/alpine/aports/-/work_items/16484 2026-08-01 12:13:16 Getting blocked from the wiki with Request Id 5a2d80dffd4d44964d650f6de883d4fb. 2026-08-01 12:13:31 Er, it worked a few seconds later. 2026-08-02 09:14:36 hi all! Could someone take a look at !106338 ? I tried to backport the github-cli security fixes from master to 3.24, it's my 1st non-master MR, hope I got things right... :-) Thank you!! 2026-08-02 13:33:31 Hi! In an APKBUILD there is one main package "M", and two (non mutually exclusive) subpackages "A" and "B". Is it possible to make apk to automatically install the subpackage "A", say as a default backend, if the user only requests to install "M" without requesting apk to add, along with it, neither "A" nor "B"? 2026-08-02 13:42:41 well, maybe this use case is bad, better always depend on "A", or even on none of them, and provide good documentation πŸ˜‰ 2026-08-02 14:05:32 I suppose put A in the package dependencies, and not B 2026-08-02 14:22:52 yes but I wanted to not necessarily install A if only M and B were requested, but I don't think it's worth it in my case, sorry for the noise 2026-08-02 15:19:16 A and B can provide C (as unversioned virtual provider) and M can depend on C. A can use provider_priority to be prefered 2026-08-02 15:22:23 But if A and B provide C, wouldn't they be mutually exclusive? 2026-08-02 16:58:05 Not if C is unversioned 2026-08-02 17:31:38 ok 2026-08-02 17:41:21 thanks! 2026-08-03 08:08:06 gitlab is unavailable, linode is performing emergency maintenance 2026-08-03 08:22:08 It's back 2026-08-03 19:25:21 can anyone help me package my first daemon? i have abuilds forked and i was wondering if i should create it under testing or under community 2026-08-03 19:27:09 nickbors: new aports start in testing/, with exception for direct dependency, e.g. of an existing aport in community/ 2026-08-03 19:27:52 i only depend on zig at build time 2026-08-03 19:31:51 sorry, meant if whatever you're packaging is needed by an existing aport to build/run 2026-08-03 19:32:09 if not, it can probably start in testing/ 2026-08-03 19:32:59 okay got it haha 2026-08-03 19:33:02 thanks for the help 2026-08-03 19:33:08 you're welcome 2026-08-03 19:33:24 mio: does the version start with a v? 2026-08-03 19:34:22 my naming convention is vX.X.X currently at https://git.nickbors.cc/-.tar.gz so i assume pkgver should be e.g v1.1.1 2026-08-03 19:34:46 or do i leave the v in the url but make the pkgver without it 2026-08-03 19:34:47 pkgver? no 2026-08-03 19:35:07 maybe have a look at https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#pkgver 2026-08-03 19:35:40 so in your case, 1.1.1 2026-08-03 19:36:20 and url can use v$pkgver if that is how you've tagged it 2026-08-03 19:36:29 okay got it 2026-08-03 20:02:24 mio: sorry to bug you but do i need a .initd if my package already has an openrc conf (i made it for alpine) 2026-08-03 20:03:12 i wont need to install that right? or i can just install it from $srcdir/contrib/openrc/$pkgname 2026-08-03 20:05:28 the latter 2026-08-03 20:09:04 (e.g. install to "$pkgdir"/etc/init.d/ and add an $pkgname-openrc to subpackages) 2026-08-03 20:09:41 thanks for the details 2026-08-03 20:10:43 you're welcome 2026-08-03 20:20:52 mio: my daemon requires a database. How would i go about migrations? just pre/post upgrade? 2026-08-03 20:21:41 my daemon isnt in charge of the schema because unix philosophy but i suppose id want migrations to be a thing haha 2026-08-03 20:29:58 not sure, sorry, maybe others can better advise there. some packages include migration scripts 2026-08-03 20:31:31 different from .{pre,post}-upgrade 2026-08-03 21:00:12 nickbors: what I typically do is write a migration function into the software itself, then when you load the DB you can check the schema version and walk it through the update process on first load. 2026-08-03 21:00:33 that doesn't work for everything though so take it with a grain of salt, it's just the pattern I've personally reached for. 2026-08-03 21:01:06 durrendal: yeah i settled on that haha 2026-08-03 21:01:09 thanks for the tip 2026-08-03 21:02:38 no problem, sometimes the path of least resistance is worth it. Plus in the future if another maintainer (or you) port it to a different distro the logic remains bound to the application rather than a specific packaging system 2026-08-03 21:03:47 durrendal: thats always a plus haha 2026-08-03 21:04:01 i was thinking of making a seperate util (like how postgres has pg_migrate) 2026-08-03 21:04:43 i have 3 scripts now its ALMOST time for a ctl haha 2026-08-03 21:04:54 but im avoiding doing real cli arg parsing lol 2026-08-03 21:06:03 nothing wrong with that method either, sometimes a handful of shell scripts with clear schema/1_$x.sh 2_$x.sh labelling is easier 2026-08-03 21:07:12 if you went that route you probably could wire it into post-install scripts. Or, maybe your openrc script could have a migrate function that invokes those scripts which you package alongside everything 2026-08-03 21:08:05 no i meant like one big migrate shell script that just read the schema version, find awk sed chained its way and piped to psql haha. The other scripts are for adding/deleting users 2026-08-04 13:19:08 a bit of self-promo (!): I have submitted !106361, a new aport that packages `dav-next`, a humble work of mine. I know that the approval (or not) or new aports can take long time, but I say this here because before Β± officially announcing my humble great work (🀣) to the πŸ‘€ of the World, and because a nginx module it is not easy to compile / install / setup, I'd like to have it at least packaged for Alpine and hopefull 2026-08-04 13:19:08 y have some feedback from interested people in its community. If there is a better way to proceed then of course feel free to tell me! 2026-08-04 18:10:54 mid, i have no feedback, but cool project :) 2026-08-04 18:18:09 thank you! 2026-08-05 03:26:28 hey guys! I just merged + mailed my first aport (a pastebin service i made in zig). If any of you have time, could you sanity check !106437? any help would be much appreciated :) 2026-08-05 05:41:03 nickbors: in the subpkg function, patze(), you are using the already defined builddir value ("$srcdir"/patze-v$pkgver") but the PWD for that function should already be $builddir 2026-08-05 05:43:38 actually you have done that in the package() function also... redundant 2026-08-05 05:44:46 it's a pretty minor nit, though I do think you should adjust it 2026-08-05 05:45:44 otherwise, looks good to me, especially for your first aport 2026-08-05 06:20:16 thanks jvvv, thats encouraging to hear! ill get around to that 2026-08-05 08:32:10 yw 2026-08-05 10:19:11 I'd like to tag an edge snapshot. would be great if we could get the edge builders idle 2026-08-05 10:19:37 so maybe hold big/heavy merges for now? 2026-08-05 15:37:01 riscv64 16/24 2026-08-05 18:14:06 edge snapshot tagged. feel free to push again 2026-08-05 18:16:00 will do docker image update later tonight 2026-08-05 20:05:05 does alpine (or apk) sandbox builds from the network? 2026-08-05 20:05:09 (either by default or optionally) 2026-08-05 20:05:18 my understanding was no (when summarising what various distros/pms do) 2026-08-05 20:05:39 sam_: default, no. `abuild rootbuild` would 2026-08-05 20:06:04 thank you! 2026-08-05 20:07:19 Some packages still opt-in to networking using the `net` option 2026-08-05 20:07:46 go/rust packages for example would need `net` 2026-08-06 11:58:44 git pull from https://gitlab.alpinelinux.org/alpine/aports.git appears to hang 2026-08-06 11:58:51 i wonder if Anubis is involved 2026-08-06 12:00:52 oh it went through 2026-08-06 12:25:54 hey, when did ~/packages change to ~/.local/share/abuild 2026-08-06 12:27:50 Oh did it? Nice! Although that shouldn't happen automatically and break workflows ofc. Still glad I don't need my ~ littered anymore though πŸ˜„ 2026-08-06 13:04:10 i might add a symlink, to be fair ;) 2026-08-06 13:22:04 Habbie, PureTryOut : pretty sure that's configurable with an env var. idk if the default changed recently or not 2026-08-06 13:22:13 the default feels changed to me 2026-08-06 13:22:27 in fact seeing what this 3.24 abuild has done in a few minutes will confirm that 2026-08-06 13:22:35 (unless it changed before the 3.24 release) 2026-08-06 13:24:42 i think its $REPODEST, per abuild(1) ENVIRONMENT section 2026-08-06 13:25:15 is that var set in your env? 2026-08-06 13:26:08 it's not set in my 3.24 env, where abuild just now created ~/packages 2026-08-06 13:26:28 REPODEST is also not set in my edge env 2026-08-06 13:27:35 but on the edge machine .local/share/abuild received the built packages? interesting 2026-08-06 13:27:46 yes 2026-08-06 13:28:09 checkout /etc/abuild.conf as well, where these variables are conventionally set 2026-08-06 13:28:32 otherwise I must be mistaken/missing something that has changed 2026-08-06 13:28:55 i'll do some more digging in a bit 2026-08-06 13:34:02 https://gitlab.alpinelinux.org/alpine/abuild/-/work_items/9997 2026-08-06 13:34:11 changed in abuild 3.18.0_rc3 2026-08-06 13:35:01 https://gitlab.alpinelinux.org/alpine/abuild/-/merge_requests/531 2026-08-06 13:35:17 in /etc/abuild.conf, the #REPODEST line indeed is different between 3.24 and edge 2026-08-06 13:43:04 I think the change is applied by /usr/share/abuild/default.conf:32 2026-08-06 13:43:43 That is on edge. I don't have a 3.24 box up rn 2026-08-06 13:47:35 After extracting the lastest abuild for 3.24, that conditional is not in the default.conf 2026-08-06 13:51:14 Looking at the various places in usr/bin/abuild and usr/share/abuild/functions.sh, I don't see how REPODEST would default to .local/share/abuild unless it is set in your /etc/abuild.conf or .abuild/abuild.conf 2026-08-06 13:52:19 That is, for 3.24 stable. On edge, it depends on if .local/share/abuild exists or not, and if REPODEST is already set 2026-08-06 13:59:07 on 3.24 i still get ~/packages, so that fits 2026-08-06 13:59:42 on edge it happens even without ~/.local present beforehand 2026-08-06 13:59:53 note that i'm not complaining and don't need help, i was just briefly surprised 2026-08-06 13:59:59 what worries me is how many other people are surprised ;) 2026-08-06 14:01:27 Sertonix[m]: shouldn't usr/share/default.conf first check for REPODEST being set before defaulting to anything? The way it is now, it looks like it will override REPODEST being set in etc/abuild.conf or .abuild/abuild.conf 2026-08-06 14:02:32 or is default.conf sourced before etc/abuild.conf and .abuild/abuild.conf? 2026-08-06 14:03:20 default.conf is sourced first 2026-08-06 14:07:24 Sorry, I see it now, in readconfig() 2026-08-06 17:02:21 any idea if anything can be done about tilix (and gtkd) conflicting with libunwind-dev by way of ldc-runtime->llvm-libunwind (not -dev) 2026-08-06 20:20:07 it would be nice if apkbuild-lint also marked things WARNING or ERROR, like abuild validate does 2026-08-07 02:18:52 Habbie: Just to make sure, was there a ~/packages directory before ~/.locale/share/abuild was used? 2026-08-07 07:24:13 Sertonix[m], no 2026-08-07 07:36:54 anyone have bandwidth to review !99156 !106234 !106236 !97494 2026-08-07 08:58:44 Hi, I have one open PR that needs review :D aports:add !106038 2026-08-07 08:58:58 good morning everyone, happy friday, happy coding 2026-08-08 16:17:23 Gcc16.2 is green !101711 but rv64 looks timed out 2026-08-09 13:08:14 Are Aports CI pipelines currently under heavy load? 2026-08-09 13:09:41 andar1an: I only see 1 job running 2026-08-09 13:09:56 How so? 2026-08-09 13:10:00 weird. CI warning I am getting is about load. 2026-08-09 13:10:06 Will try to rerun jobs 2026-08-09 13:10:12 andar1an: perhaps gitlab load? 2026-08-09 13:10:17 I've seen that message more often 2026-08-09 13:10:18 ah, maybe 2026-08-09 13:10:34 that is what the error said, I had assumed this was self-hosted on alpine infra 2026-08-09 13:10:54 thank you Ikke, will wait an dtry again 2026-08-09 13:10:59 yes, it is 2026-08-09 13:11:14 But I guess gitlab (the software) has some rate limiting 2026-08-09 13:12:54 im trying to spam less haha, I have been bad at that. I currently got better at using cache. But post-local build/testing not much i can do 2026-08-09 13:57:25 is there a way to skip a test in Go like with Rust, or do I need to patch test file with t.skip? 2026-08-09 13:57:37 -skip 2026-08-09 13:57:57 nice, I dk why that isn't coming up in help or searches. Thanks Ikke 2026-08-09 13:59:57 i lie. it was embedded in help paragraph 2026-08-09 14:00:47 Yw 2026-08-09 14:25:54 FYI, I have enabled the AutoLabeler service in aports-qa-bot again. It was disabled a long time ago due to a bug 2026-08-09 14:28:38 And also a new service that should provide feedback on soname bumps 2026-08-09 17:36:26 go CLI interface is amazingly terrible so don't beat yourself about it 2026-08-10 13:03:16 Can we merge !105805? The installed file is completely broken in its current state. 2026-08-10 15:39:26 Hi, are there any known issues with scripts/bootstrap.sh from aports? 2026-08-10 15:39:26 The path mentioned is a symlink to "/usr/lib/libgcc_s.so.1", which makes libtool resolve it to the HOST's libgcc, which is x86-64 so can't be linked with an armv7 binary. 2026-08-10 15:39:26 Running with "CBUILDROOT=... scripts/boostrap.sh armv7". Compiles fine at first, but fails at "gmp" with linking error: "/home/kubasz/sysroot-armv7/usr/lib/libgcc_s.so: file not recognized: file format not recognized" 2026-08-10 15:39:26 Manually deleting the symlink and making it relative fixes the issue, but it reappears again when building "binutils" with a similar error: "/usr/lib/libc.so: file not recognized: file format not recognized" 2026-08-10 15:39:26 I'm trying to bootstrap for armv7 on an Alpine 3.24.1 x86-64 host (clean install, tried real hardware and Docker). Asked in #alpine-ports first but got redirected here. This doesn't seem to be specific to ARM, other architectures fail the same way. 2026-08-10 15:50:40 it seems like the same issue is described in alpine/aports#16368, but that one is already two years old 2026-08-10 21:21:36 πŸ‘‹ 2026-08-10 21:22:23 I got few stale / pending MRs, allow me to link them here in case anyone has time to look at them πŸ€— 2026-08-10 21:22:28 !99837 2026-08-10 21:22:34 !105847 2026-08-10 21:22:39 !102477 2026-08-10 22:55:09 how do nushell completions work when it comes to packaging? is it similar to other shells? Are there any examples to reference? not sure what to grep for nu 2026-08-10 22:57:56 dont see something in /usr/share that may be related yet. Wondering if it may be built-in or maybe in home dir 2026-08-11 02:30:52 kuba2k29: sounds like we shuold make it a relative symlink 2026-08-11 08:20:37 hi all, i have weird requests - which package takes the longest time to build from aports? 2026-08-11 08:21:08 are anywhere stats about that? 2026-08-11 08:55:32 Get in touch with this platform for greatness you’ll definitely thank me laterβ€¨β„ΉοΈβ€οΈβ€¨πŸ‘‡πŸ»πŸ‘‡πŸ»πŸ‘‡πŸ» 2026-08-11 08:55:33 https://t.me/+pa-CiKYv9-5lNWM8 2026-08-11 10:19:44 WhyNotHugo: yeah, I fixed it this way too and it compiled (mostly) fine after that. Though the workaround for #16368 was still needed 2026-08-11 14:32:55 indy: could be chromium, but we may have aports that take longer to build than that 2026-08-11 15:06:38 can I get !106695 merged? it fixes 2 nasty RCEs that affect gitea and a slew of other security issues that have been fixed since the maintainer was last active. 2026-08-11 15:09:16 durrendal: `-buildmode=exe"` Is this due to CGO? 2026-08-11 15:09:31 Because it's better to enable CGO than to switch from pie to exe 2026-08-11 15:09:54 oh is that the preferred method? I thought it was the other way around, let me fix that 2026-08-11 15:11:33 Yeah, pie is a security feature 2026-08-11 15:11:43 So switching to exe weakens security 2026-08-11 15:29:21 ah yeah that makes sense, I've changed it for only the arches that need it. 2026-08-11 15:34:04 durrendal: it might be that you need to patch a Makefile or something. CGO is enabled by default, so generally it's the projects build system that disables it 2026-08-11 15:36:00 oh, the build passed 2026-08-11 15:39:05 that's what I'm used to having to do, woodpecker does it that way, but for some reason redeclaring the export lets it go through 2026-08-11 15:39:28 Yeah, I guess it checks if it has been explicitly exported 2026-08-11 15:39:46 I see other CVE fixes for 1.26 as well: https://blog.gitea.com/release-of-1.26.0/ 2026-08-11 15:39:58 And other minor versions 2026-08-11 15:50:10 there's about 60 between the last packaged version and this one, I honestly wasn't certain if it made sense to list every CVE in secfix since we're skipping between versions and never updated the package during those. 2026-08-11 15:51:00 Hmm, it's bit of a question. Officially we could only list CVEs that we patch ourselves (ie, where the fixed version is earlier than what upstream fixed) 2026-08-11 15:52:05 oh so in this case we'd list no CVEs under secfixes, since 1.27.1 is the official fix for the two listed 2026-08-11 15:52:52 that might be cleaner/higher value since apkbuilds aren't vulnerability databases 2026-08-11 15:53:45 Yup, but convention atm is to list the CVEs even if they are fixed in the same version as upstream 2026-08-11 16:00:40 if you want I can go back and add the CVEs between 1.25.4 and 1.27.1? They're not hard since Gitea on on top of things. 2026-08-12 01:56:35 I screwed up and merged an MR where the commits weren't properly squashed, !105569 2026-08-12 01:56:51 I suppose we'll live with that now rather than rewrite? πŸ˜” 2026-08-12 01:58:03 it happens. and yeah, no rewriting 2026-08-12 01:59:43 My shame shall live forever in commit history 2026-08-12 02:36:18 it's a badge of honor :) 2026-08-12 05:24:35 that xen upgrade broke something for me, I'll try and figure out what and why 2026-08-12 05:54:54 it's upstream changes to /etc/xen/xen-network-common.sh that removed a function I've relied on for years 2026-08-12 06:21:02 fix was simple 2026-08-12 07:05:53 omni: this change? https://xenbits.xen.org/gitweb/?p=xen.git;a=commitdiff;h=280e4ca4dffd1558d237804a8bf349b9c3874660;hp=6c9344ee3d48b8bb7758c72d9e4eeb8841b8f7ec 2026-08-12 07:39:39 dne: yes, so the fix was to just change to setup_bridge_port in my script 2026-08-12 07:55:01 ha "update APKBUILD" 2026-08-12 07:55:13 WhyNotHugo: no worries. it happens 2026-08-12 11:01:18 omni, thanks for tip, i saw some riscv64 taking 4h where x86_64 took 1/2h 2026-08-12 11:42:09 FYI I broke py3-openssl and a bit more with !106736 going to fix it or maybe better to revert? 2026-08-12 13:22:25 WhyNotHugo: it's tedious, but it can be good to have a habit of checking the Commits tab in the gitlab ui 2026-08-12 13:22:42 I typically do 2026-08-12 13:23:09 one of the merges I'm most embarrassed about is probably when I merged some large debug log file 2026-08-12 13:25:38 I intentionally only have the https remote to upstream and not the ssh remote πŸ˜… 2026-08-12 13:26:46 hmm, well, when I repackaged uutils with the uutils-coreutils as a drop-in replacement for coreutils and it broke builds soon before a release, _that_ was embarrassing 2026-08-12 14:31:19 Hi, anyone to merge this one? 2026-08-12 14:31:19 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/103453 2026-08-12 14:32:14 I thought it would be merged sooner but I don't know the timetable of aports/testing 2026-08-12 14:32:14 This is the C3 compiler (c3c) 2026-08-12 15:37:55 just let me know on the MR, thanks! 2026-08-12 17:56:46 i had a look at it. I think it only needs to be squashed into a single commit 2026-08-12 17:56:52 language looks interesting 2026-08-12 17:57:00 a lot to like there 2026-08-12 19:22:33 ncopa: I normally "Squash and merge" on github, I'd assume the same for gitlab 2026-08-12 19:23:01 c3 lang is indeed interesting 2026-08-12 19:36:00 squash and merge breaks commit authorship 2026-08-12 19:38:04 ncopa: thanks, I've squashed and rebased. https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/103453 2026-08-12 21:33:13 Is ther a guide for using mkimage.sh ? 2026-08-12 21:33:36 https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/scripts/mkimage.sh?ref_type=heads 2026-08-12 21:43:16 pretty short script, isn't the --help enough? 2026-08-12 22:56:57 mkimage.sh: line 17: /usr/share/abuild/functions.sh: No such file or directory 2026-08-12 23:29:16 apk add abuild 2026-08-13 08:06:54 Does anyone know if there are plans to bring the skiasharp package that's currently in the edge repository to the current stable repo? I think jellyfin needs it to be able to create library thumbnails because it packages libHarfBuzzSharp.so. I'm also wondering if missing that library is also causing jellyfin-ffmpeg to segfault when attempting to burn-in subtitles. 2026-08-13 09:08:40 any volunteer to take over perl maintenance? 2026-08-13 09:24:22 ncopa: who is "eligible" for this? I mean, I used to hack some perl stuff some (long) time ago, like porting Β± successfully to TomTom GPS systems or Android, but I'm not an Alpine maintainer so I don't know if I'm eligible for this… Is there a trial period? πŸ˜‰ 2026-08-13 11:12:15 you can have a trial period if you want. basically, it is just adding your name into the APKBUILD and bump pkgrel (so the package gets rebuilt). After that it is just to review any merge requests, comment and give a thumbs up or approve or something and ping someone with merge privileges to ensure it gets merged. 2026-08-13 11:13:36 The most complicated part is when there are new minor releases, eg when 5.45 arrives, and we need to rebuild lots of packages 2026-08-13 11:14:24 But I expect that you can get help with it when that happens 2026-08-13 11:16:48 looking at the git log, does not seem like it will too complicated 2026-08-13 12:32:47 ok, good, so I'm volunteer! about the trial period, I'm sorry it was a joke, I noticed too late it was very French people (and ofc possibly some other countries) oriented 2026-08-13 13:00:12 well, if you find out it does not work out, we just remove you as maintainer. 2026-08-13 13:45:41 "Required by (1483)": that's understandable, of course, for a package like perl. How are they managed usually? I guess there is a script somewhere, or I can write one, but are there pitfalls that I may not think of? except for build failures that must be managed manually? 2026-08-13 20:40:46 Gitaly fetching of a repo in ci for riscv seems to have been stuck for like 1hr+ now. Is this normal? Should i restart job? 2026-08-13 21:08:14 I restart a job once for those weird issue. If the same error message shown second time, I wait for reviews. 2026-08-13 21:08:29 * error message is shown second 2026-08-13 21:19:18 Ty 2026-08-14 10:29:44 ncopa: !106817 … 2 builds failed but since I only changed the maintainer comment they must not be related, although I'm interested to know why they failed, maybe I just have to restart them? 2026-08-14 13:01:03 ok, jobs restarted and passed, must have been transient errors, ok for this 2026-08-14 17:36:20 Anyone have time to help me understand the failure of linking with simde for !106744 2026-08-14 17:46:21 It's complaing about a function in x86 headers while building for non-x86, so I'm guessing there's some target arch detection in its build system that isn't accounting for those arches 2026-08-15 05:35:12 Arnavion ah, I think you nailed it. I do not believe those arches are fully supported by SIMDE yet, so it looks like I need to make simde-dev an arch-conditional makedepends, so supported arches get it and unsupported don't and just build normally 2026-08-15 12:50:22 kuba2k29: The libgcc_s.so looks a lot like #18307 but so far the follow-up MR wasn't created by JustSoup (they/them) 2026-08-15 12:52:09 For now one needs to manually revert 3ce8cecea0d7e2c819d65bcc8915f30a2de3d80c to have cross compilation be functional. 2026-08-16 03:19:23 Is it normal that `abuild rootbld` fails for rust packages? I'm getting a "Could not resolve host: index.crates.io" error 2026-08-16 03:20:50 vpzom: maybe it needs options=net ? 2026-08-16 03:21:35 ah yep, looks like I have two fixes to submit then :p 2026-08-16 07:32:45 hi all! I have this MR !106863 for github-cli for which I am a maintainer. Is there anything else for me to do other than comment "looks good" and thumbs up it? (after reviewing it, it goes without saying). I don't see an approve button or anything else similar... Thank you! :-) 2026-08-16 07:43:42 a comment is sufficient 2026-08-16 07:57:17 thank you! :-) 2026-08-16 10:31:27 Does anyone have time to have a look at !70266? 2026-08-16 10:44:57 hjaekel: i'm a bit confused why you're making changes other than pkgrel bumps in the dependents with the rebuild commits 2026-08-16 10:46:33 if those changes are really needed then they should get proper commits and not hide version bumps/other changes in a commit that only says "rebuild" in the message 2026-08-16 10:46:45 lotheac for example qgis did not build any more because of a SIP update, so I had to fix that to get the build running 2026-08-16 10:49:33 I mentioned that in the second line of the commit message, should I move that to the first line? 2026-08-16 10:50:17 it depends on what the changes are. so for qgis, you're not just rebuilding, you're upgrading the package 2026-08-16 10:50:24 so the message should reflect that 2026-08-16 10:50:42 as in, it would normally be "testing/gqis: upgrade to 3.44.13", right? 2026-08-16 10:51:53 if it's *only* a pkgrel bump then the rebuild commit message is fine, but if you're making other changes, recording why in the commit logs makes it easier to review 2026-08-16 10:52:15 as another example you're removing a patch from community/pdal, but there's no clue as to why 2026-08-16 10:56:22 Ok, I will fix that 2026-08-16 10:59:29 pushed 2026-08-16 11:03:20 thanks. i'll check the other commits too and leave comments 2026-08-16 12:23:47 lotheac I updated the MR, let me know if something else should be changed 2026-08-16 12:24:07 and thank you for the review 2026-08-16 12:24:54 seems reasonable enough. i will probably merge it tomorrow if ci passes 2026-08-16 12:30:20 ci will take some hours 2026-08-16 12:31:58 yeah, that's why "tomorrow" :) 2026-08-16 12:51:48 I got a package I'm packaging flagged that it is out of date. however I already have a MR !106738 for the previous version, if someone can take a look at it, since I had to do some changes to the build 2026-08-16 12:52:06 maintaining* 2026-08-16 13:50:43 if you think the package is ok to build this new way I can rebase with 0.3.2 instead of 0.3.1 which is in the MR now. 2026-08-16 14:12:36 Hi, Can you please re-review MR !101551? 2026-08-16 20:24:45 what archive format is used for apk? i'd like to compile my stuff but also just have it managed by the package manager 2026-08-16 20:24:53 nor do i really want to fork aports for that 2026-08-16 20:32:39 tar.gz - but you can just write a quick APKBUILD file out-of-aports-tree and have abuild -r build it. 2026-08-16 20:34:07 also note, if you go down this road, the deps of that package will "pin" and possibly block updates of other pkgs that depend on those deps. so best is to have no deps for such out-of-aports packages, which will keep your system healthy update-wise, but your handcompiled stuff might break after upgrades 2026-08-16 20:36:26 unvFK9I3vgc: got it, i can't really use aports as i'm currently running a chimera VM for testing this stuff. i assume they do something different because i am unable to extract cached apk files 2026-08-16 20:37:23 (even though they both use apk? i'm mainly asking here because alpine maintains apk so it's worth asking their dev channel) 2026-08-16 20:38:51 I believe chimera is using apk v3 while alpine is still on v2 2026-08-16 20:39:11 oh got it, i though alpine was using apk v3 since v3.23 2026-08-16 20:40:02 "The package manager has been upgraded to v3, but keeping the v2 index and package format for now." 2026-08-16 21:23:21 hrm, I'm upgrading a diskless 3.22 on an rpi to 3.24, and ran into a few things...but my main gripe is that the FAT of the boot partition starts at sector 0. Wouldn't be a problem if the boot partition were big enough to hold a new kernel, but now that I'm attempting to resize it, all tools (parted, *fdisk, ...) freak out... 2026-08-16 21:25:17 ...should we improve the rpi story a bit? give a wee bit bigger boot partition to begin with? or am I going completely wrong about this anyway and diskless installs should be upgraded differently to begin with, not just rpis 2026-08-17 00:42:54 I have an upgrade MR for tcpreplay. The upgrade includes 3 security fixes but there are no CVEs for them. There is some GHSA-* id's which I guess are github security advisories. Can the GHSA id be used instead of a CVE id? 2026-08-17 06:45:20 j_v: looks like you can e.g. see https://security.alpinelinux.org/vuln/GHSA-77vh-xpmg-72qh 2026-08-17 06:45:43 (but 'CVE-2026-47215 / GHSA-wqcr-7rf3-f64m' on a single line was not parsed correctly in community/singularity/APKBUILD ) 2026-08-17 06:46:10 (or rather only the CVE part was kept) 2026-08-17 06:46:41 The '/' should not be there 2026-08-17 06:46:44 just space delimited 2026-08-17 06:49:00 hmm, https://security.alpinelinux.org/srcpkg/py3-jinja2 has 'CVE-2025-27516 GHSA-cpwx-vrp4-4pq7' in the APKBUILD but only the CVEs are listed; I'd naively think it just keeps the first word (but honestly I don't think it's a problem... Just grepped for GHSA to check for j_v) 2026-08-17 06:49:21 Right 2026-08-17 08:28:11 morning, I have an open PR asking to ping review !105454 2026-08-17 11:28:20 oh btw, just to clarify, is 10.0_git20251207 > 10.0 ? 2026-08-17 11:28:58 er, it's in the docs, sorry for the noise 2026-08-17 11:29:01 (https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#pkgver) 2026-08-17 11:33:10 would be nice if someone could take a look at !97202, I feel like I'm probably being very impatient, but it has been sitting around for so long the versions became outdated 2026-08-17 12:28:58 Asmadeus: Thanks for the input. Because there are no CVEs for the ones I'm talking about, I just did it like this: '# - GHSA-p3f2-88rg-9mch'. I based it on what I found in other APKBUILDs 2026-08-17 22:04:12 Ariadne: Do you have a preferred method for submissions for macifrename? Is the email addr in the APKBUILD ok for submitting patchsets? 2026-08-17 22:04:30 github 2026-08-17 22:04:40 Ok, thanks 2026-08-17 22:13:07 Didn't find a mirror on github, only the repo on sr.ht. I emailed it to the email you list on your about page. Hope that is ok 2026-08-18 02:34:21 If an upstream lists vulnerabilities without CVEs but links to an oss-security post (https://www.openwall.com/lists/oss-security/2026/08/17/1) is there any convention for tracking the fixes? 2026-08-18 04:21:50 jvvv: no, we don't have anything for that. Sometimes someone will report back with assigned CVEs 2026-08-18 04:26:46 ikke: thanks for heads up. I linked to the post in the commit details but left it out of the APKBUILd itself, so I guess that should be ok. Seems like anyways 2026-08-18 04:27:05 yeah, it's fine 2026-08-18 04:27:18 We can always add it later 2026-08-18 04:27:34 True 2026-08-18 04:28:46 mio caught me on one of my recent MRs where I missed the security fixes so I am a bit hyper vigilant atm 2026-08-18 04:29:09 mio is good 2026-08-18 04:32:49 We keep eachother sharp 2026-08-18 20:32:18 Could someone review !103005? The requested changes are now ready. 2026-08-18 22:44:13 Hi team, I noticed wget on busybox doesn't support the no_proxy env variable. Can I submit a patch to the busybox apk to add support? 2026-08-18 23:16:06 Is anyone else ksmbd broken on 3.24.1? I'm getting a strange core dump 2026-08-18 23:21:00 https://gist.github.com/mterron/377eab6e46f30817b1a9bf9492e85c18 2026-08-19 05:54:03 terror: feel free to submit a patch to upstream busybox https://lists.busybox.net/mailman/listinfo/busybox but be aware that they get a lot of patches and it does not seem they have resources to review all of them. A tip is to add the feature you need but at the same time reduce size (use make bloatcheck to measure size). If you manage to do that 2026-08-19 05:54:03 you usually get their attention 2026-08-19 05:55:34 would be good if you could report the ksmb bug upstream as well 2026-08-19 05:59:58 what is no_proxy btw? There appears to be an option -Y on/off Use proxy 2026-08-19 06:15:38 I suspect it's yet another (mis)feature of GNU wget 2026-08-19 07:25:02 it's a "standard" environment variable; e.g. curl and py3-requests support it too ;) 2026-08-19 07:26:38 because curl supports it doesn't mean it's a good idea :/ configuration via environment variables is confusing and error-prone 2026-08-19 07:27:17 Why is it confusing? 2026-08-19 07:27:27 and more error prone relative to other options? 2026-08-19 07:29:19 apparently it originated in CERN's libwww in 1994… 2026-08-19 08:04:43 we are updating gitlab right now. its down for a few mins 2026-08-19 09:58:25 does anyone has an opinion on overriding firmware on diskless? https://gitlab.alpinelinux.org/alpine/aports/-/work_items/18239 2026-08-19 10:00:49 the original idea was that if /firmware exist on the boot media (eg USB disk or sd-card) it will be copied instead of using the firmware shipped with kernel. This way you could add extra firmware if needed, but you would need to have a copy of all firmware, not just a single one extra. this makes it possible to "delete" firmware by omitting copying 2026-08-19 10:00:49 it. if you want prevent any firmware to be loaded you simply have an empty dir there. 2026-08-19 10:01:29 but that idea is now questioned, and it suggested that the /firmware from boot media is appended to the one shipped with kernel 2026-08-19 10:02:04 i like the original idea because it is stupid and simple 2026-08-19 10:02:42 even if it probably is more convenient to append, as mentioned in the above work item 2026-08-19 11:07:23 ncopa: it could support both 2026-08-19 16:04:55 Should i just wait indefinitely for new aport mr's to be reviewed? Or should I @team/maintainers in PR. There is no rush, just looking at list and wondering 2026-08-19 16:07:07 Would one get feedback if it wouldn't be added so it could be closed? 2026-08-19 16:09:31 Also, it seems ci automatically creates anitya rule? 2026-08-19 16:18:43 andar1an[m]: @team/mentors, rather than maintainers :p 2026-08-19 16:19:13 and yeah, many developers don't actively check new aport MRs, so a ping would probably get some attention 2026-08-19 16:25:29 My bad, I knew it was an m lol 2026-08-19 16:25:46 Cool, thx 2026-08-19 16:27:11 About the Anitya rule, is that generally automatic? 2026-08-19 17:54:17 andar1an[m]: which Anitya rule? 2026-08-19 17:56:41 For all packages on pkgs.alpinelinux.org. is the anitya rule automatically created for new aports? For a recent package that was added, I saw that it is already listing version status and I didn't create rule. 2026-08-19 17:57:13 hm, i don't think we currently have any code that would do that? 2026-08-19 17:57:33 But then for others i see "rule not in anitya" 2026-08-19 17:58:17 Okay, guess one can always check after merge. Just pretty sure i didn't add a rule so i was unsure 2026-08-19 17:59:00 hm, id.fedoraproject.org is timing out for me 2026-08-19 17:59:12 so i can't even log into Anitya lol 2026-08-19 17:59:31 Ya, first 7 pages almost are showing anitya errors. Didn't try to login 2026-08-19 17:59:55 (On pkgs.alpinelinux) 2026-08-19 18:00:09 Well, warnings not errors 2026-08-19 18:03:15 I have issue getting into release-monitoring.org with fedora id also. 2026-08-19 18:03:20 Also gateway timeout 2026-08-19 18:03:33 Hope their infra is okay 2026-08-19 20:06:54 hey, a very simple (and embarrassing) change to package metadata reported here has a patch and backports ready to review, if any approvers are available: https://gitlab.alpinelinux.org/alpine/aports/-/work_items/18368 2026-08-19 20:07:07 (four linked MRs with the same contents in the comment) 2026-08-19 20:47:20 ptrc: thanks! 2026-08-19 20:47:35 to be fair, v3.22 is now failing because of some silly checksum issue :p 2026-08-19 20:47:48 but it's an easy fix, so no worries 2026-08-19 20:47:50 and you're welcome! 2026-08-19 20:51:50 ah, is it missing one of the other commits from 3.23+? let me check 2026-08-19 20:51:53 no, no 2026-08-19 20:52:04 the source URL + checksum was updated in 2025 2026-08-19 20:52:09 but it wasn't rebuilt since 2026-08-19 20:52:20 and currently the builders are a bit silly and cache sources by just filename 2026-08-19 20:52:38 so it would hit the cache, download the github.com sources, and try to compare the checksum against sr.ht 2026-08-19 20:52:51 ah! ok 2026-08-20 03:38:51 Anyone able to review !106236? It is a bump for an xfce4 panel plugin 2026-08-20 08:31:36 I see that "Merge" is enabled on an MR which is 81 commits behind the target branch. Does GitLab do the rebase implicitly now? 2026-08-20 08:31:41 Or would this create a merge commit? 2026-08-20 08:31:56 Example: !107020 2026-08-20 08:32:23 Spoiler: it rebases implicitly. 2026-08-20 08:32:28 SUPER nice ✨️ 2026-08-20 08:32:31 2026-08-20 08:32:33 2026-08-20 08:32:35 Yes 2026-08-20 08:32:40 huh, not sure where those empty lines came from 2026-08-20 08:36:02 Added a notification explaining it for developers/maintainers 2026-08-20 08:48:08 cool :) 2026-08-20 11:16:08 automatic merge ? 2026-08-20 14:10:42 quite nice to not have to rebase before merge 2026-08-20 14:51:49 Is the Alpine infrastructure having issues? 2026-08-20 14:52:07 gitlab apparently is 2026-08-20 16:24:09 Is it something to do with cloud infra? Having trouble with other services too, specific to EU 2026-08-20 16:26:48 Dk what gitlab runs on. Probably unrelated 2026-08-20 16:31:04 No, issue is gitaly load 2026-08-20 16:31:16 oh ok 2026-08-20 16:31:24 Either by webscraping or excessive fetching / cloning 2026-08-20 16:31:30 :( 2026-08-20 16:55:30 The gitlab instance has some proof of work service that sits infront? What is the goaway service? 2026-08-20 16:55:46 https://git.gammaspectra.live/git/go-away/ 2026-08-20 17:00:24 Cool, thanks for link. Looks like they use forgejo, nice haha. They had a recent security update. Will need a minor version bump. Should check for mr 2026-08-20 17:00:54 go-away had a version bump? 2026-08-20 17:01:00 No, forgejo 2026-08-20 17:01:02 ah ok 2026-08-20 17:01:11 Should have been more clear 2026-08-20 17:02:00 Forgejo taught me about Go templates, pretty cool haha. 2026-08-20 17:02:22 if go-away is out of date feel free to stalk me on IRC so I bump it 2026-08-20 17:02:48 I didn't check, i can compare while checking for forgejo mr 2026-08-20 17:02:52 right now it isn't 2026-08-20 17:02:59 0.7.0 is the latest 2026-08-20 17:03:12 Gitlab is bot usable for me atm 2026-08-20 17:03:21 Ok, good haha 2026-08-20 17:03:32 go-away did not have a release for quite some time 2026-08-20 17:04:38 I've been using anubis, it has been good for the minimal level of scrapers i get 2026-08-20 17:15:52 Lot's of scraping from GCP 2026-08-20 17:19:54 block google? 2026-08-20 17:21:17 Doing it, but we have CI servers hosted there, so cannot be a blanket ban 2026-08-20 17:21:23 oh 2026-08-20 17:26:03 ptrc: can you kill your auto retry? 2026-08-20 17:26:05 with anubis i was able to define bot policies in yaml config, the weights were helpful for that. 2026-08-20 17:26:08 oh, oof 2026-08-20 17:26:11 i had no idea that's still a thing 2026-08-20 17:26:15 it's a cronjob.. somewhere..... 2026-08-20 17:26:20 do you have any idea where it's coming from? 2026-08-20 17:26:25 No 2026-08-20 17:26:57 sorryyyyy 2026-08-20 17:26:59 give me a second 2026-08-20 17:27:01 i'll try and find it 2026-08-20 17:27:27 Maybe it's something else 2026-08-20 17:28:00 Are you able to rate limit spammy ips? 2026-08-20 17:28:23 Or do you already? 2026-08-20 17:28:43 There are too many IPs 2026-08-20 17:28:55 Good ol exponential backoff haha 2026-08-20 17:29:06 In router you can have a dynamic list 2026-08-20 17:29:24 So when repetitions occur you can go through stages of denial 2026-08-20 17:29:42 Doesn't need to be statically defined necessarily 2026-08-20 17:29:53 Just need to track initial connections 2026-08-20 17:30:45 oh yeah, it's not me, it's sending `{"id":"retry"}` 2026-08-20 17:31:01 my cronjobs were sending random bullshit in there 2026-08-20 17:32:52 😹 2026-08-20 17:40:46 netcup gmbh 2026-08-20 18:01:07 Not sure who / what that is, but I've blocked it for now 2026-08-20 18:06:11 Seems like things are recovering 2026-08-20 19:18:43 load is nice and low again 2026-08-20 20:15:05 netcup is a cheap vps 2026-08-20 20:16:10 Yeah, but could be managed by someone where 2026-08-20 20:16:13 here* 2026-08-20 20:36:40 i was getting hit with crawler ddos last couple of days then it stopped today 2026-08-20 20:55:31 I find crawlers came in waves 2026-08-20 20:55:45 Usually at similar times daily 2026-08-20 20:56:02 Sure does seem like ddos haha 2026-08-20 20:56:17 Just so darn many 2026-08-20 22:54:42 Anyone know of any go packages that embed a ui (preferably svelte). I am having trouble with appearance though there are no errors, and tests all pass 2026-08-20 23:12:10 I believe woodpecker and gitea both do this, not sure they're svelte though 2026-08-20 23:24:28 Thank you.Ya, forgejo is based on gitea, so I have seen that. Zulip-Desktop uses vite, which is what this project does. But for some reason the assets and pages all get created but it doesn't appear correct. Upstream bin looks fine 2026-08-20 23:25:10 Gonna check out woodpecker meow 2026-08-20 23:31:17 Even using the makefile is problematic. Future me problem. Thx for help 2026-08-21 06:04:42 hi all! When someone has a chance, I have a security upgrade for github-cli: !107142 Thank you! :-) 2026-08-21 15:37:44 heads up: the ceph packages have a severe CVEs, I created a work item for the ceph upgrade also impacting clients at https://gitlab.alpinelinux.org/alpine/aports/-/work_items/18412 2026-08-21 17:57:54 Why alpine-keys installs to usr/share/apk/keys and etc/apk/keys? 2026-08-21 18:42:37 All Alpine keys versus relevant keys for the current architecture 2026-08-21 20:09:58 but apk-tools reads /usr/share/apk/keys? 2026-08-21 20:12:05 also something is copying the keys into etc/apk/keys because I have everything from usr/share/keys 2026-08-21 20:24:27 apk-tools reads /etc/apk/keys and /lib/apk/keys 2026-08-21 20:26:01 Hi all, does anybody have time to review !105939? 2026-08-22 16:26:40 andar1an[m]: another couple I know of are adguardhome and gomodel (I think gomodel recently switched to svelte) 2026-08-22 16:28:33 Thanks iggy, I created an issue upstream, because even using make file same issue. There is a rename of ui directory in next release. Im assuming it may have to do with some dependency to start, but it looks like it is creating all assets, so it is weird. 2026-08-22 16:29:11 I haven't used the ui before, so didn't catch this until looking again recently 2026-08-22 16:32:35 telmich: I also suggested we make 20 the default since when the next release comes out 19 will be EOL ... see MR #104853 2026-08-22 16:32:47 the next Alpine release that is 2026-08-22 17:17:59 !107066 / !106817 ? 2026-08-22 17:19:30 mps is not even here... 2026-08-22 17:19:46 He is in other channels 2026-08-22 17:19:57 yeah one of #alpine-arm #alpine-loongarch #alpine-riscv64 2026-08-22 17:20:14 where perl maintenance is probably off-topic 2026-08-22 18:05:48 iggy: Makes a lot of sense 2026-08-22 19:21:59 ikke: thanks for your comment on his MR… I guess he could not see that he was making an MR that conflicted with mine…? I mean, without checking all other MRs 2026-08-22 19:22:20 Correct 2026-08-22 19:22:33 ah, he just answered 2026-08-22 22:03:35 For related packages (i.e. packages and needed dependencies not in aports) is it okay to do on 1 branch? 2026-08-22 22:03:52 Or keep them separate 2026-08-22 22:47:55 andar1an[m]: 1 branch, 1 commit/aport, if the dependencies/upgrades are required for the primary aport to build/function. also please mention it in the MR description 2026-08-22 22:48:19 Sounds good, that is what i have done and assumed 2026-08-22 22:48:37 When i am done this other one I will comment properly 2026-08-22 22:48:52 Is it possible to add a checkdepends for a secondary architecture 2026-08-22 22:49:13 I.e. if test is running on x86_64, but you also need i686 files for a lackage 2026-08-22 22:50:11 Gonna call it soon, but just want to commit this and need to do that first 2026-08-22 22:50:47 Apk add --arch in apkbuild won't work 2026-08-22 22:55:05 not sure it is supported ... the files may have to be supplied in an archive, or for now skip the tests that require the files 2026-08-22 22:57:41 Okie. Ty. I had assumed it may be possible with a tag like for repos 2026-08-22 22:57:53 Like pkgname@arch 2026-08-22 22:58:18 Will just skip for now 2026-08-23 10:12:05 noticed the following line in kdevelop's .PKGINFO: 2026-08-23 10:12:09 `provides = cmd:kdevelop!=26.04.3-r0` 2026-08-23 10:12:20 anyone knows where that comes from? doesn't seem to be abuild 2026-08-23 10:12:50 oh my dog nevermind 2026-08-23 10:12:54 it literally has 2026-08-23 10:13:00 an executable named `kdevelop!` 2026-08-23 10:15:38 do we not have literally any kind of sanitization against this 2026-08-23 10:15:57 cc: Sertonix[m] 2026-08-23 10:18:29 kdevelop!! 2026-08-23 10:22:19 https://ptrc.gay/JSoeFumv.txt 2026-08-23 10:22:24 that's such a funny failure case 2026-08-23 10:35:26 Although it rather looks like kind of an apk bug, putting '=' characters in a binary name doesn't sound very safe 2026-08-23 10:35:35 Why not abuild-10 2026-08-23 10:36:24 my point is that when you *do* put a random special character in a binary name, it gets put directly into the provide 2026-08-23 10:36:40 and then that influences apk parsing 2026-08-23 10:36:50 ptrc: The sanitization only excludes abuild at the moment 2026-08-23 10:36:54 fortunately `!=` is not a valid constraint, so apk parses it correctly 2026-08-23 10:37:05 *excludes @ 2026-08-23 10:37:16 ah, welp 2026-08-23 10:37:34 ah ok ptrc, I thought the provide was on purpose, thanks :) 2026-08-23 10:37:42 realistically it's the fault of the format that it puts the constraint in-band 2026-08-23 10:55:10 Oh my dog is my new favorite phrase haha. Need to use 😹 2026-08-23 10:55:45 Very wow 2026-08-23 11:00:21 Does anyone handle missing symbols that are from glibc libraries by building with gcompat? Last few days i have hit missing symbols a few times and dk how to handle properly. 2026-08-23 11:00:21 One of the symbols though comes from libdl and I don't think that is on gcompat. Does one need to like map to system calls somehow 2026-08-23 11:02:38 In aports it's not allowed to use gcompat like that. It's better to figuure out where these symbols come from. Common issues would be precompiled binaries or bundled copies of glibc header files. 2026-08-23 11:06:12 Thank you. One ks from libdl on a current package i am working on 2026-08-23 11:06:14 ptrc: Hm, I wonder if we should encode provider names instead of ignoring. 2026-08-23 11:06:17 *traced provider names 2026-08-23 11:07:08 Can you think of any examples that do that Sertonix? Not sure what to grep for 2026-08-23 11:07:10 andar1an[m]: happens every once in a while, i had to add some symbols to gcompat a while ago 2026-08-23 11:08:37 Should i be informing about symbols somewhere? I have gcompat installed. At least 3 i am missing across various build attempts for different packages 2026-08-23 11:09:28 I would start with finding the ELF file that uses the symbols and check where the file comes from 2026-08-23 11:09:59 Ldd isn't liking when i try to check 2026-08-23 11:10:19 But i may have looked at the wrong file, will try this morning 2026-08-23 11:11:12 I know 1 of 2 symbols comes from libdl 2026-08-23 11:13:54 I think readelf can list the usef symbols of a single ELF file 2026-08-23 11:16:25 Ya, or binsider. Shoulda thought of that but didnt. Thank you 2026-08-23 11:17:05 It's a coffee morning lol 2026-08-23 13:23:08 is it not possible to build mingw-w64-gcc for aarch64? im naiive here 2026-08-23 14:13:03 It is probably possible 2026-08-23 14:13:28 just wondering about people that may game on aarch64 devices 2026-08-23 14:13:41 currently it only builds for x86 and x86_64 2026-08-23 14:14:42 assuming any windows / wine use also, not just gaming 2026-08-23 14:16:51 Running windows binaries is mostly using precompiled binaries. What would you need mingw-w64-gcc for? 2026-08-23 14:19:24 specifically it relates to checks in icoextract package I am working on, but I could also just skip the checks 2026-08-23 14:20:02 would make my life way easier, because right now I have to go through and patch out tests for other archs for each arch 2026-08-23 14:20:55 doesn't actually pertain to build dependencies - so if that is acceptable, im happy to just skip haha 2026-08-23 14:22:38 do other windows compat packages not use wingw to generate tests? this is first time I am exposed to this 2026-08-23 14:23:41 it is interesting, but it is also challenging to make work in ci 2026-08-23 14:27:06 in a way, all test cases could be covered skipping alternate arches because all arches that are built hit test cases, but I don't know if that is apples to apples 2026-08-23 14:31:45 I also keep relying on ci to test alternate archs - can't find way to do something like CHOST=x86 abuild rootbld. Don't wanna keep spamming infra 2026-08-23 14:44:13 It's CBUILD 2026-08-23 14:49:07 thank you! maybe that can be added to abuild help. I only see CHOST and CTARGET related to cross-compiliation 2026-08-23 14:49:38 SO NICE! THANK YOU HAHA 2026-08-23 14:56:04 The issue is that people thinkbthey want cross-compilation when they don't. You want compilation in an emulator 2026-08-23 14:56:42 (And many places continue to call it cross compilation) 2026-08-23 14:56:49 compilation in an emulator is slower though 2026-08-23 14:57:19 that's why cross-compilers exist 2026-08-23 14:58:50 would one using rootbld with CBUILD flag be considered cross-compilation or compilation in an emulator? I want to say the latter, but I don't want to get language wrong 2026-08-23 14:59:00 ptrc: I think the abuild issue should be fixed now 2026-08-23 14:59:39 andar1an: compilation in emulator 2026-08-23 14:59:51 Sertonix[m]: thanks <3 2026-08-23 14:59:55 90% of apkbuilds in aports aren't setup for cross-compilation IIRC 2026-08-23 15:01:03 compilation in an emulator: CBUILD=x86 abuild rootbld 2026-08-23 15:01:03 in abuild terms there is also cross creation: CTARGET=x86 abuild rootbld 2026-08-23 15:01:03 cross compilation: CHOST=x86 abuild rootbld 2026-08-23 15:01:43 CHOST and CTARGET won't work either as base-build-x86 won't exist 2026-08-23 15:02:01 at least that is error i get 2026-08-23 15:04:42 Cause cross compilation requires setup 2026-08-23 15:05:09 ah, definitely didn't do that yet haha. 2026-08-23 15:05:21 I like compilation through an emulator haha -- seems easier 2026-08-23 15:05:48 Yes, the easy way is what most people want 2026-08-23 15:07:44 ci I imagine is where you want cross? 2026-08-23 15:08:03 not so bad there to set up 2026-08-23 21:18:18 Hi there, when registering for alpine's gitlab account should I expect confirmation mail to come immidiately or me using quite unusual address could've broken it? 2026-08-23 21:21:23 hi, what do you mean by "quite unusual"? 2026-08-23 21:23:54 I have a habit of using @mydomain.tld for all accounts I'm registering + tld of that alias email is .business so some websites don't accept it but most often if they don't they just reject it straight away 2026-08-23 21:33:37 yeah, no, that would definitely not interfere here 2026-08-23 21:33:45 what's your email host / MX? 2026-08-23 21:36:00 I'm using mxroute so mxrouting.net 2026-08-24 07:49:19 andar1an[m]: Sertonix[m] https://gitlab.alpinelinux.org/alpine/abuild/-/merge_requests/561 2026-08-24 07:51:21 Yea no, still no gitlab confirmation mail sadly, would it be possible to either contact somebody to approve it or have it deleted so I can re-create it with different email? 2026-08-24 09:57:32 Cool, thanks @WhyNotHugo 2026-08-24 14:37:11 hi, planning an MR to enable CONFIG_ACPI_APEI_GHES=y in main/linux-lts. looks like it was =m until 441cf9f (3.0.17 -> 3.2.1, jan 2012), where the tristate->bool change upstream silently turned it into n. anyone know if that was intentional? 2026-08-24 15:22:02 I doubt anyone still knows after 14 years :P 2026-08-24 16:26:36 Ok so I guess I'll submit that mr soon :) 2026-08-24 17:40:30 fyi, ncopa prefers to do those changes along with kernel updates 2026-08-24 17:45:05 So when is the best time to submit it? 2026-08-24 17:48:24 Best way is to open an issue, then ncop will make the change. You can make an MR, but ncopa will probably cherry-pick the commit 2026-08-24 17:51:46 I see. Well I opened the request anyway in draft mode, I'll probably open an issue repeating the MR body 2026-08-24 23:17:00 Should we support removing certain aports from main? For example, main/perl-inc-latest hasn't been updated upstream for over 10 years, and no other aport depends on it. Should we move it to community or testing, or simply remove it? 2026-08-25 09:38:40 when reverting an upgrade, should you also bump the pkgrel from what it was before the upgrade? 2026-08-25 09:38:46 or does that not matter 2026-08-25 09:49:29 yes 2026-08-25 09:49:47 It matters mostly for the cdn, which caches apk files by name 2026-08-25 09:51:03 ok, thanks 2026-08-25 14:08:15 this one should be powerful https://www.sifive.com/development-platforms/sifive-bigsky-sf-2u870-datacenter (for riscv problems) 2026-08-25 14:10:41 for the low low price of like, a house? 2026-08-25 14:50:20 tomalok: seems like your last fix to libpathrs is not sufficient 2026-08-25 14:50:26 Still getting relocation errors 2026-08-25 14:53:07 jvoisin: IDK I guess it's expensive 2026-08-25 14:53:12 jvoisin: The price of what a house cost when they were still attainable 2026-08-25 14:57:54 tomalok: seems like runc needs to be rebuilt 2026-08-25 14:58:15 It's now built against librspath.so, without the .0 suffix 2026-08-25 17:37:18 ikke got a report on it overnight -- libpathrs is rebuilt -- though there's no riscv64 APK in place yet, and that's what's blocking rebuild of runc. that edge buildier still out of service/degraded? 2026-08-25 17:42:24 achill: Not sure how you manage the -stable .config files, I think ncopa does the -lts ones manually. I'm looking to get something like https://gitlab.alpinelinux.org/alpine/aports/-/commit/b42805fa4cf6ac1b7b34feb2187fb37f610c816a in -stable. Would a ticket asking for that work or do you manage it with diffs? 2026-08-25 17:44:52 Does not look manual to me 2026-08-25 17:50:03 I'm pretty sure he spins up an actual `make menuconfig` or whatever to make the changes to make sure the deps and all that are all right 2026-08-25 17:52:26 ikke - https://pkgs.alpinelinux.org/packages?name=runc&branch=edge&arch= now showing runc-1.5.1-r3, except riscv64 2026-08-25 19:22:29 tomalok: the runner is back, but still building packagees 2026-08-25 19:23:29 s/runner/builder 2026-08-25 20:46:32 yeah i'd expect it to have a bit of a backlog 2026-08-26 05:08:17 tomalok: community has been uploaded 2026-08-26 18:10:37 Why does Alpine have tar and img downloads for Raspberry Pi? What are the differences? 2026-08-26 18:30:22 just two different format. see the 'Manual method' section in this page https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Preparing_the_installation_media 2026-08-26 21:08:12 Thank you for the reply. Most distros only provide one download file format, typically iso. What is the driving reason for different file formats, iso vs tar, for Alpine downloads? 2026-08-26 21:19:30 what distros provide an iso for raspberry pi? you can't boot a raspberry pi with a livecd/liveusb like an x86 computer. instead, you write the .img to an sd card or partition and copy the tar to the sd card 2026-08-26 21:24:50 You can write to usb i believe for lite models as a caveat i think 2026-08-26 21:25:39 Or maybe that is to use sd card. I may be backwards 2026-08-26 21:26:17 Was able to use usb on either a pi with or without mmc πŸ˜… 2026-08-26 21:27:42 but need to update overlay config likely for usb controller. Then you can install to another disk like nvmr 2026-08-26 21:27:50 s/nvme 2026-08-26 21:29:10 yes you can use usb or nvme for newer ones but it's still the same, no livecd/liveusb 2026-08-26 21:30:18 What do you mean by liveusb? Sorry, confused. If booting off the usb, would that be different? 2026-08-26 21:31:11 (I understand files are different) just confused about what to call it 2026-08-26 21:33:29 On mine i used an sd or usb to install to an nvme, which is why im confused 2026-08-26 21:37:02 like x86 2026-08-26 21:38:28 where you have an iso9660 filesystem image (usually distributed as a .iso) written to an external media (cd/dvd/usb) that is special in that it's usually booting a compressed userspace image where any writes aren't saved back to the disk after shutdown 2026-08-26 21:38:53 Ah, that's it, thank you! 2026-08-26 21:39:15 The writes. Perfect delineation now. Merci beacoup 2026-08-26 21:39:30 on raspberry pi, if you used an sd/usb to bootstrap, there's nothing different about it compared to a normal system on sd/usb, you're just using it to get started on nvme 2026-08-26 21:39:58 you could put the nvme in another computer or usb adapter and install it directly too 2026-08-26 21:40:16 Ya, im working towards pxe for these 2026-08-26 21:40:24 Just haven't gotten to it yet 2026-08-26 21:51:17 Sorry if not best way to share, but all the pi talk lately made me want to share haha. Finally started putting work into these: https://imgur.com/a/4LeoC1y 2026-08-26 21:51:17 Yay Alpine 2026-08-26 22:09:10 abby: Alpine-linux downloads provides the option of iso or tar for raspberry pi. I am trying to understand why and if the raspberry pi iso is like the x64 Alpine iso? Why does Alpine provide tar and iso for raspberry pi, but tar only for some and iso only for others? 2026-08-26 22:11:03 I think it is tar and img for pi. And i think the img is raw write to disk so it will include partition tables and sectors. I may be wrong though 2026-08-26 22:11:20 sk8board[m]: no, alpine has .img not .iso. these are different: img is just a disk image, iso is a special format 2026-08-26 22:13:22 Would the reasoning be because pi users who want to just write alpine to sd or disk for quick use through something like rpi imager don't need to worry about partition structure abby? 2026-08-26 22:21:24 no, it's because raspberry pi and most arm platforms don't support booting an external device like x86 bios/uefi does 2026-08-26 22:21:40 usually they'll have a special boot process or bootloader 2026-08-26 22:22:26 now things like ampere or snapdragon laptops have uefi support so they can use a live iso like x86 2026-08-26 22:26:34 Cool, thanks 2026-08-26 22:56:00 Mistake, thank you for the explanation. 2026-08-26 22:56:21 s/Mistake/My mistake/ 2026-08-27 08:10:11 Hmm.. I've got a noisy user asking about openssh updates for 3.24 because there are plenty of CVEs, and won't take "but if you look at the actual content the CVEs don't affect you" for an answer.. Is there a reason the main/openssh updates have only been pushed to master or would a backport MR to 3.24 be acceptable? 2026-08-27 08:42:09 security by checkbox 2026-08-27 08:57:36 Asmadeus: Either check whether backports for 10.3 are available, or assess what the impact would be of upgrading to 10.4 2026-08-27 09:23:22 https://metadata.ftp-master.debian.org/changelogs//main/o/openssh/openssh_10.4p1-5_changelog 2026-08-27 11:58:42 They upgraded to 10.4 in debian testing but not in stable; for stable they've stayed on 10.0 but backported fixes from up to 10.3 according to https://metadata.ftp-master.debian.org/changelogs//main/o/openssh/openssh_10.0p1-7+deb13u4_changelog 2026-08-27 12:03:26 honestly the "Potentially-incompatible changes" in both https://www.openssh.org/txt/release-10.4 and https://www.openssh.org/txt/release-10.5 look unlikely to break anything to me, but I very much agree with ikke's "security by checkbox" comment, so I'm not going to push very hard for this - was mostly curious about the reason not to backport as I've rarely had troubles with openssh minor upgrades 2026-08-27 12:06:58 (I say that, but the "must restart openssh" adventure a couple of years back was for 9.9, so that actually wasn't backed by much thought . . . :D) 2026-08-27 12:07:50 otoh openssh announced they're dealing with a large number of security reports lately (who hasn't) and will likely publish security releases more often, so it's going to be a hassle to keep up one way or another.. 2026-08-27 12:16:05 Well, it would be great if it were just security releases 2026-08-27 12:16:23 But projects more and more are just combining security releases with feature releases and even breaking changes 2026-08-27 12:22:36 Yeah openssh in particular has never been known for making a distinction, they actually just wrote "more frequent releases to get bugfixes into users' hands more quickly" (it's in the 10.5 release notes) 2026-08-27 12:40:48 !107433 2026-08-27 13:02:49 Thanks 2026-08-27 13:03:25 It only updates to 10.4, but 10.5 fixes some issues as well, not sure if we should directly go to 10.5 2026-08-27 16:41:41 Greetings all. I have a small doubt - does package() strip setgid perms from the directories created? If so, how can I bypass that? 2026-08-27 17:33:37 would someone be able to look at !103937? It's been sitting for a bit 2026-08-27 18:40:10 flamingradian[m]: I've been thinking about this some. One thing that comes to mind that last time I looked at a related topic, I think I found out that the files in question did not actually have the suid/sgid bits set. Not saying that is the case for you. Another issue that could be affecting this could be build time permissions, especially if building in a non-privileged 2026-08-27 18:40:12 container like rootbld does. Just some ideas 2026-08-27 18:51:06 Might also have something to do with fakeroot 2026-08-27 18:54:45 uh did you mean flamboyantpenguin 2026-08-27 18:57:53 Yeah, sorry, tab completion gone wrong, pebcak for not checking before posting 2026-08-27 19:01:50 Also, filtering out join/leave messages doesn't help either in this case 2026-08-28 11:53:51 jvvv: I did set the perms via install command on package(). No use. Did chmod too. Tried post-install, that did not work either. Do you mind looking at the APKBUILD? 2026-08-28 12:38:28 Here's it for reference 2026-08-28 12:38:28 https://code.dawn.org.in/dawn/linchpin 2026-08-28 12:55:10 The bug will be fixed in the next apk-tools release: https://gitlab.alpinelinux.org/alpine/apk-tools/-/work_items/11216 2026-08-28 13:44:19 Sertonix[m]: (The bug will be fixed in the next apk-tools release: https:/…) - I see, thank you. Will this be available in 3.24.1 itself or next release? 2026-08-28 13:57:55 I don't know yet 2026-08-28 15:57:18 I have noted in packaging that git-sub-modules can make things more tedious. 2026-08-28 15:57:18 In a mono-repo structure does anyone use sparse checkout, is that a reasonable option? 2026-08-28 15:57:18 For various hosts alpine setups scripts/image generators i have a monorepo with a subdir per host. Recently i am adding githooks for permissions. 2026-08-29 13:19:35 I think it may only make sense in scenarios where not an external dependency, but gonna try today 2026-08-29 16:24:09 could someone please merge !107500 ? thanks! 2026-08-29 19:00:11 thanks ikke! 2026-08-29 20:12:15 In the glab.a.o interface for an MR, I ticked the notification icon thinking that would just be a notication for myself. I didn't just turn on notifications for something more general, like other users, did I? 2026-08-29 20:13:32 I have guest status in aports, so I worry in a case that this that I may have done the wrong thing 2026-08-29 20:14:55 s/case that/case like/ 2026-08-30 05:07:13 jvvv: that's just for yourself 2026-08-30 05:12:31 good deal, thanks for heads up 2026-08-30 17:40:21 Is gitlab under heavy load? 2026-08-30 17:40:48 AI scrapers gonna scrape 2026-08-30 17:43:45 This be true, but that doesn't answer question haha. Dk if just me or others are seeing degraded performance. 2026-08-30 17:44:53 Oh, you meant at this moment. It was talked about in the main alpine channel. 2026-08-30 17:45:04 Looks like the person who mentioned it there is about to reply 2026-08-30 17:49:07 there was an alert in the channel with alerts, I think i-k-k-e (don't want to highlight him because he's probably busy with this) is on it 2026-08-30 17:49:33 Ikke is always fixing this stuff. Dog bless him 2026-08-30 17:49:56 So much for not mentioning him πŸ™ƒ 2026-08-30 17:50:09 You need an @ i think haha 2026-08-30 17:50:42 andar1an[m]: I get pinged when someone just mentions my name 2026-08-30 17:51:48 That seems undesireable 2026-08-30 17:51:52 andar1an[m]: this is irc not discord, just mentioning a name pings the person xD 2026-08-30 17:52:45 Good to know 2026-08-30 17:52:51 f_: But also Matrix (where I am) does that, too. 2026-08-30 17:53:11 Im on Matrix client atm. I need an @ 2026-08-30 17:53:25 yeah, if you have a common name like `hey` or something, probably a little annoying... for most people it's probably fine 2026-08-30 17:53:33 anyways doesn't matter 2026-08-30 17:53:38 Probably depends on the client 2026-08-30 17:53:40 Just good to know not just me 2026-08-30 17:53:53 but yeah, I get pinged when someone says i.e. let's get jiggy with it 2026-08-30 17:54:18 Lol exactly use case i waa wondering about 😹 2026-08-30 17:54:39 iggy: That's IRC-side, right? 2026-08-30 17:54:49 luckily, nobody says stuff like that anymore 2026-08-30 17:55:03 iggy: that's a pretty bad irc client 2026-08-30 17:55:08 mine doesn't do that 2026-08-30 17:55:11 Im a big fan of iggy azalea 2026-08-30 17:55:16 Did that do it? 2026-08-30 17:55:16 iggy: "Quiet piggy" 2026-08-30 17:55:18 I think it's the client, but most client's emulate each other 2026-08-30 17:55:40 yeah, I got highlights for all of those 2026-08-30 17:55:49 Oof 2026-08-30 17:56:24 Why i assumed there would be a trigger symbol even in irc 2026-08-30 17:56:58 my old client had a setting to only hilight on "whole words", but it doesn't misfire often enough for it to be annoying 2026-08-30 17:58:01 Imagine if f_ didnt have the underscore lol 2026-08-30 17:58:26 "For f sake" haha 2026-08-30 17:58:42 I believe I've seen him in other spaces as "funderscore" 2026-08-30 17:58:57 someone already took the 'f' nick on libera and on oftc 2 chars is the minimum 2026-08-30 17:59:02 yes I am funderscore too 2026-08-30 17:59:16 Awesome haha 2026-08-30 17:59:23 I use funderscore or f_ interchangeably, even here on irc 2026-08-30 18:00:10 I'm realizing this getting pretty far removed from alpine development. Would there be a better alpine space for digressions like this? 2026-08-30 18:00:37 Off topic, also why i tried to drop earlier haha. At least it's a weekend 2026-08-30 18:00:58 (I meant offtopic channel) 2026-08-30 18:01:22 #alpine-offtopic 2026-08-30 18:02:35 Let me steer things back to something devel-relevent. I've got jellyfin on my alpine server, and right now it's usability is degraded due to a lack of libHarfBuzzSharp. Looks like the jellyfin package in edge, as well as the skiasharp package in edge, provide for this library. Is there any plan to backport either of these into the current stable? 2026-08-30 18:57:51 grossaffe[m]: stable backports are typically only for security upgrades and such, not package adding 2026-08-30 18:58:07 If you want these, run edge 2026-08-30 18:58:21 Or wait until the next alpine major release 2026-08-30 18:58:50 Bug-fixes, too, no? 2026-08-30 18:59:35 yes 2026-08-30 19:06:45 I'm not exactly sure how all of this comes into play in the packaging, but I believe everything was fine in Alpine 3.23, but 3.24 broke it. 2026-08-30 19:07:05 Looks like there was some discussion on it in this issue: https://gitlab.alpinelinux.org/alpine/aports/-/work_items/17830 2026-08-30 19:09:31 you can build them in stable 2026-08-30 19:10:22 And I saw that the packages that supply libHarfBuzzSharp that are in edge aren't providing any newer minor revision; in the case of skiasharp, it's package 3.119.4-r0 vs 3.119.2-r0, and for Jellyfin, it's 10.11.11-r3 vs 10.11.11-r0 2026-08-30 19:16:02 "you can build them in stable" <- I'll look into that if it's decided there will be no back-porting. I haven't dealt with alpine packaging to this point. 2026-08-30 19:18:43 it's pretty easy, get the directory containing the APKBUILD file for the package you wanna build, and run abuild -r on it. then you can install the package from the local repo under ~/packages/ - but take note, that this will probably pin a bunch of deps to the version your custom pkg depends on, and makes system upgrades probably incomplete. if so, you might wanna uninstall the 2026-08-30 19:18:45 custom package, upgrade, and then rebuild it and reinstall it. 2026-08-30 19:20:19 [some details omitted] 2026-08-30 19:20:37 you can also try to just pull it from edge with edge as masked repo 2026-08-30 19:20:48 be careful 2026-08-30 19:25:12 Yeah, that definitely adds a complication to sysadmin. I'd have to do some calculus on the additional admin burden versus the current degraded functionality. 2026-08-30 19:25:12 And to be clear, I'm not looking for *new* functionality here, but rather this is something that was broken by the 3.24 release that worked in the 3.23 release. It just took me some time to recognize and diagnose the issue. 2026-08-30 19:25:12 In Jellyfin, I've confirmed that it manifests in a failure to create library thumbnails, as it uses libHarfBuzzSharp to draw text on top of the thumbnail. It may also be causing a much greater issue causing segfaulting when attempting to do subtitle burn-in. 2026-08-30 22:37:33 Hey all. Apologies if this ends up being contentious. After Debian's recent GR, there are a number of folks I respect jumping ship, and I know there'd be strong interest in Alpine if there were a policy published governing LLM contributions or LLM-aided contributions. I've not yet managed to find one, and I wonder if such a thing exists or is being worked on. 2026-08-30 22:40:40 here's the latest update I've seen about it: https://gitlab.alpinelinux.org/alpine/council/-/work_items/697#note_636853 2026-08-30 22:40:50 craftyguy[m]: Thank you! 2026-08-30 22:48:49 How often do council meetings happen? Didn't even know this was a thing lol 2026-08-30 22:51:51 Man, that would have been a good conversation. Too late now haha 2026-08-30 23:27:44 If a dev has time, could they take a look at !107602? I approved it just now 2026-08-31 00:28:50 omni: thank you 2026-08-31 00:38:24 jvvv: np, did it before I saw that you had mentioned it here =) 2026-08-31 00:39:07 all good, appreciated either way :) 2026-08-31 08:08:53 fossdd: I made the initial commit in !107324, could you help me take care the rebase and make the pipelines green? I want to take a break for the next two weekends. Many thanks 2026-08-31 08:48:25 hi! could someone take a look at !106569? 2026-08-31 08:53:15 sure~ 2026-08-31 08:54:16 ptrc: thanks :3 2026-08-31 12:19:23 Gitlab okay right now? 2026-08-31 12:20:17 Getting 500 2026-08-31 12:30:36 Should be fine now 2026-08-31 12:32:59 https://cdn.roaring.earth/wp-content/uploads/2017/09/Screen-Shot-2017-09-06-at-4.40.28-PM.png 2026-08-31 12:35:10 Thank you ikke, it is 2026-08-31 12:36:04 Looks like a live action scene from the lion king lol 2026-08-31 12:41:04 it's how I imagined ikke as he typed the above 2026-08-31 12:41:38 If anyone has time this week can I please get !87654, !106811, !98418, and !106683 looked at. Just trying to get off list. If not desired can close. 2026-08-31 12:41:43 Lmao yes omni 2026-08-31 12:42:06 Oh crap, im sorry, i didn't know bot would do that 2026-08-31 12:49:21 it is very convenient 2026-08-31 12:50:00 Also, do any apk packages support nushell completions? I only grep bash, zsh, and fish 2026-08-31 12:51:02 Yes, i just wouldn't have done 4 at once lol 2026-08-31 17:53:31 When it comes to command CBUILD=aarch64 abuild rootbld i am getting pre and post install errors for busybox, abuild alpine-baselayout, and openrc when environment is being set up. Has anyone encountered similar b4? 2026-08-31 17:54:42 Error code is 127 2026-08-31 17:57:40 have you tried restarting qemu-binfmt service? 2026-08-31 18:25:23 No i haven't, I have libvirtd service running, no qemu-binfmt service directly. But i can try to add that or see if therr is a qemu dependency in a bit. Thanks mio 2026-08-31 18:25:54 Silly not to think qemu was necessary somehow. Just assumed it would be bundled through command 2026-08-31 18:26:20 Much appreciated 2026-08-31 18:34:53 andar1an[m]: maybe try installing qemu-openrc (and qemu-aarch64 if not already added) and starting the qemu-binfmt service. hope you will eventually get it working 2026-08-31 18:37:30 Registering the binfmt automatically would be possible but the detection would have reliability issues 2026-08-31 18:50:10 I have it installed, just didn't know was necessary for rootbld 2026-08-31 18:53:18 Will be sure to add it and then i imagine can do for all archs, nice. TIL 2026-08-31 18:53:21 Ty 2026-08-31 18:55:41 you're welcome 2026-08-31 20:06:49 could someone please review/merge this one? !103779 (it's not stale) 2026-08-31 20:11:19 I got a few in earlier, but if you filter on "new aport" there are 330 open MRs... 2026-08-31 20:13:17 ya I'm not complaining, thank you and other maintainers for your work reviewing this stuff :) 2026-08-31 20:14:58 I didn't mean to imply that you were 2026-08-31 20:15:19 Makes me feel bad having open mrs haha. Why i have been curious about a triage or criteria system to know if one should keep open 2026-08-31 20:15:57 but it's like.. that number is the number of MRs I used to be comfortable with in total 2026-08-31 20:16:00 I have no issue having mr not merged, can do separately, so i worry about pollution when trying 2026-08-31 20:16:38 P.s. good to note you also don't seem to need system emulator for rootbld, just arch emulator for binary 2026-08-31 20:17:07 Thanks so much for help again. Hope i spam ci less now haha. Was missing openrc pkg and specific arch emulators 2026-08-31 20:18:42 you shouldn't feel bad, I too have a bunch of open MRs 2026-08-31 20:19:24 But it creates extra work to triage and sort, no? Could there be a way to make that easier? 2026-08-31 20:20:46 Closing things is also probably as valuable as merging 2026-08-31 20:21:34 omni: thank you :) 2026-08-31 23:58:43 There may be issue with linux-firmware update and bluetooth drivers for some systems. Can dog in more tomorrow