#archlinux-ports | Logs for 2025-09-22

Back
[00:35:24] -!- marmis has quit [Ping timeout: 258 seconds]
[00:35:53] -!- marmis has joined #archlinux-ports
[00:49:43] -!- marmis has quit [Quit: Bye!]
[01:06:57] -!- marmis has joined #archlinux-ports
[01:08:29] -!- marmis has quit [Client Quit]
[01:10:05] -!- marmis has joined #archlinux-ports
[01:25:47] -!- marmis has quit [Quit: Bye!]
[01:29:46] -!- marmis has joined #archlinux-ports
[03:18:53] -!- marmis has quit [Quit: Bye!]
[03:31:03] -!- marmis has joined #archlinux-ports
[04:23:37] -!- marmis has quit [Quit: Bye!]
[04:34:26] -!- marmis has joined #archlinux-ports
[06:06:29] -!- idealseal has quit [Ping timeout: 250 seconds]
[06:07:07] -!- idealseal has joined #archlinux-ports
[06:33:40] -!- marmis has quit [Quit: Bye!]
[06:37:33] -!- marmis has joined #archlinux-ports
[07:45:53] -!- rossy has joined #archlinux-ports
[08:58:34] -!- hcmb has joined #archlinux-ports
[09:06:00] -!- hcmb has quit [Changing host]
[09:06:00] -!- hcmb has joined #archlinux-ports
[09:17:43] -!- hcmb has quit [Remote host closed the connection]
[09:53:09] -!- marmis has quit [Quit: Bye!]
[09:54:54] -!- marmis has joined #archlinux-ports
[10:55:59] -!- hcmb has joined #archlinux-ports
[10:56:50] -!- hcmb has quit [Changing host]
[10:56:50] -!- hcmb has joined #archlinux-ports
[11:02:40] <hcmb> Hej, I am here because of https://bbs.archlinux.org
[11:02:42] <phrik> Title: Bring ARM into the fold? (Page 3) / Other Architectures / Arch Linux Forums (at bbs.archlinux.org)
[11:03:30] <hcmb> is there a way for me to read up on that (by accessing the room history it something) or to get a summary of the discussion?
[11:03:30] <Antiz> hcmb: 👋
[11:06:19] <Antiz> hcmb: Is there anything you'd like to have some input on in particular ?
[11:08:19] <Antiz> hcmb: The sum up is that some people here started some community effort around porting Arch to aarch64 and approached us about it to eventually start collaborating more closely (in the scope defined in this thread: https://lists.archlinux.org)
[11:08:35] <hcmb> mostly just what the progress looks like, and if there is any way I could help (I think mostly with testing stuff).
[11:09:57] <Antiz> This is still at an early stage but some milestones have already been reached. A working system can already be bootstrapped / booted for instance :)
[11:10:22] <hcmb> Antiz: ah, very nice, that is what I was hoping for :) then I'll just stay here (and set up some permanent IRC presence once I get home), and see how it develops
[11:10:32] <Antiz> hcmb: Sure :)
[11:10:55] <Antiz> hcmb: For now the scope for help is mostly check which packaging are not building on aarch6
[11:11:00] <Antiz> aarch64*
[11:11:04] -!- hcmb has quit [Read error: Connection reset by peer]
[11:11:23] -!- hcmb has joined #archlinux-ports
[11:11:45] <Antiz> And open some merge requests at https://gitlab.archlinux.org (in the scope defined in the above mail thread)
[11:11:46] <phrik> Title: Packages · GitLab (at gitlab.archlinux.org)
[11:11:58] <Antiz> open some merge requests to fix it that is.
[11:12:30] <Antiz> Some other areas will need some work/help at some point, but at the moment this is where we're at.
[11:12:42] -!- hcmb has quit [Changing host]
[11:12:42] -!- hcmb has joined #archlinux-ports
[11:12:49] -!- hcmb has quit [Read error: Connection reset by peer]
[11:13:06] -!- hcmb has joined #archlinux-ports
[11:13:27] -!- hcmb has quit [Changing host]
[11:13:27] -!- hcmb has joined #archlinux-ports
[12:50:33] <Solskogen> which reminds me that I need to document how to get it up and running
[12:51:54] <Solskogen> hcmb: one way of helping is just install and start using it and report back any issues you might see.
[14:06:56] <bschnei> hcmb: what ARM device do you have/would you be using to test?
[14:07:40] <hcmb> rpi3 and 4 for now, so nothing exciting for now
[14:12:40] <gromit> https://gokrazy.org I checked and the gokrazy project seems to use the RPI kernel for the pi5, but I think I heard people say in here that it is somewhat well-supported from mainline right?
[14:12:42] <phrik> Title: Supported platforms :: gokrazy (at gokrazy.org)
[14:12:58] <jelle> gromit: if you dont need internet
[14:13:13] <jelle> *ethernet
[14:13:27] <gromit> jelle: Ah so wifi works but ethernet does not? 🤔
[14:14:27] <jelle> gromit: https://www.phoronix.com
[14:15:02] <gromit> jelle: lul the beer bottle in the background :D
[14:15:09] <jelle> its suse
[14:15:27] <jelle> well michael likes beer!
[14:58:48] -!- nl6720_ has joined #archlinux-ports
[14:59:03] -!- nl6720 has quit [Ping timeout: 258 seconds]
[15:00:56] -!- nl6720_ has quit [Client Quit]
[15:01:47] -!- nl6720 has joined #archlinux-ports
[15:30:10] <Solskogen> I have the pi5 kernel in our repos
[16:51:52] <bschnei> hcmb: 1) we don't know if our kernel config will work on either of those devices yet (help welcome!) and 2) packages hosted at https://arch-linux-repo.drzee.net will not work for your devices as they are built for ARMv8.2. I have a separate repo with packages that should work on the Pi4. For more background see: https://gitlab.archlinux.org
[16:51:52] <bschnei> JpaWQiOiI2IiwiZnVsbF9wYXRoIjoiYXJjaGxpbnV4L3BvcnRzL2FhcmNoNjQvcHJvamVjdC1tYW5hZ2VtZW50IiwiaWQiOjIxMTk1MH0%3D
[16:51:53] <phrik> Title: Issues · Arch Linux / Ports / AArch64 / project-management · GitLab (at gitlab.archlinux.org)
[16:52:20] <bschnei> better URI: archlinux/ports/aarch64/project-management#6
[17:51:19] -!- hcmb_ has joined #archlinux-ports
[17:51:36] -!- hcmb_ has quit [Client Quit]
[17:51:49] -!- hcmb_ has joined #archlinux-ports
[17:52:50] -!- hcmb_ has quit [Client Quit]
[17:53:08] -!- hcmb_ has joined #archlinux-ports
[17:56:15] -!- hcmb_ has quit [Client Quit]
[17:56:24] -!- hcmb_ has joined #archlinux-ports
[17:56:24] -!- hcmb_ has quit [Client Quit]
[17:56:34] -!- hcmb_ has joined #archlinux-ports
[17:59:11] -!- hcmb_ has quit [Client Quit]
[17:59:21] -!- hcmb_ has joined #archlinux-ports
[18:06:45] -!- hcmb_ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
[18:09:59] -!- hcmb_ has joined #archlinux-ports
[18:11:34] -!- hcmb_ has quit [Client Quit]
[18:11:52] -!- hcmb_ has joined #archlinux-ports
[18:12:36] -!- hcmb has quit [Remote host closed the connection]
[18:12:45] -!- hcmb_ has quit [Client Quit]
[18:13:13] -!- hcmb has joined #archlinux-ports
[18:28:48] -!- hcmb has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
[18:37:15] <gromit> bschnei: you get the full URL if you open the issue instead of previewing
[18:38:08] <bschnei> ya sorry about that. https://gitlab.archlinux.org
[18:38:09] <phrik> Title: Packages fail to build without LSE support (#6) · Issues · Arch Linux / Ports / AArch64 / project-management · GitLab (at gitlab.archlinux.org)
[18:51:56] -!- hcmb has joined #archlinux-ports
[18:55:19] <hcmb> very nice with the issue tracker, thanks for the link. When I have some time I will try it out on one of my boards
[19:03:54] <bschnei> hcmb: message me when you are ready and I can give you URIs for ARMv8 package repos. They are very incomplete but should work on v8. The bigger challenge will be a working kernel. I don't own a Pi4 making it hard to check what is required.
[19:38:13] <idealseal> Are there any notes on how to get started?
[19:44:13] -!- Guest46 has joined #archlinux-ports
[19:46:04] -!- Guest46 has quit [Client Quit]
[19:57:59] <Solskogen> we need a place for the initial tarball.
[19:58:12] <Solskogen> DrZee?
[21:05:24] <DrZee> I'm working on operationalizing tar balls ... using the archiso script ... just trying to figure out why archlinux-docker rolls it's own rootfs instead of using a tarball and/or using pacstrap to make rootfs .... anyone know why?
[21:13:36] <Antiz> DrZee: pacstrap requires root privileges
[21:14:15] <DrZee> yeah ok ... but why build your own rootfs if it's technically just is the tarball?
[21:14:24] <DrZee> or do I miss something
[21:14:28] <Antiz> So relying on `fakeroot / fakechroot` allows for rootless builds
[21:15:20] <Antiz> DrZee: I'm not sure I follow
[21:15:34] <Antiz> https://gitlab.archlinux.org <-- This generate the tarball
[21:15:35] <phrik> Title: scripts/make-rootfs.sh · master · Arch Linux / archlinux-docker · GitLab (at gitlab.archlinux.org)
[21:16:39] -!- marmis has quit [Ping timeout: 258 seconds]
[21:17:06] -!- marmis has joined #archlinux-ports
[21:17:58] <Antiz> DrZee: So this script generates the tarball that is later used a rootfs for the image. Making your own rootfs tarballs also allows to inject specific stuff in there.
[21:18:22] <Antiz> For instance the docker image inject custom pacman config during rootfs generation: https://gitlab.archlinux.org
[21:18:23] <phrik> Title: scripts/make-rootfs.sh · master · Arch Linux / archlinux-docker · GitLab (at gitlab.archlinux.org)
[21:20:33] <Antiz> DrZee: "yeah ok ... but why build your own rootfs if it's technically just is the tarball?" <-- I'm not sure to understand what you mean here. The rootfs indeed is just a tarball but said tarball still has to be generated (which is what the script does). Am I missing / misunderstanding something maybe?
[21:23:29] <DrZee> if you regularly publish tarballs (from archiso) why not download that tarball and use it for the docker base instead of going through the process of making you own tarball in that script too? that is what I was thinking of ....
[21:24:27] <Antiz> Because the docker image has specific needs
[21:26:26] <Antiz> The tarball produced from archiso and uploaded to mirror each month is a "generic" one.
[21:28:47] <Antiz> The docker image has special needs / config that has to be set up in the rootfs: e.g. custom pacman config to not extract some specific stuff and save some space, not generate locales, comment some default option in pacman that do not work and/or makes no sense in a container, etc...
[21:31:07] <Antiz> With such specific needs, it's way easier / more straightforward to generate your own rootfs rather than downloading the generic tarball to then strip / remove / modify stuff from it
[21:31:18] <DrZee> yeah sure but those customisations can be made after the tarball is unpacked into the build folder the docker script sets up ... IMHO it would make it more logic to show the docker modification clearly
[21:38:03] <Antiz> Yeah well, IDK if there's an historical reason for that.
[21:38:49] <Antiz> The docker image used to rely on pacstrap to generate rootfs, but that was probably later dropped later to avoid requiring root privileges (which makes things easier / more secure)
[21:40:44] <Antiz> So I think it was just "let's use pacstrap" initially and then later "let's use fakeroot/fakechroot instead so we can have rootless build / CD". Relying on the "premade" archiso tarball probably never was considered as it implies more loops and hacky-ish stuff.
[21:44:10] <Antiz> More loops and hack-ish stuff as in, taking something premade and modify it according to your needs versus doing it from scratch directly including what you need. It's less "pragmactic / kiss" (Arch principles wise) I guess :P
[21:44:44] <Antiz> Although I get that the KISS vision / definition may be subject to debate :)
[21:54:45] <Antiz> DrZee: Is there any reason why you were looking at the Docker image by the way? Because we can agree that the Docker image is definitely a specific case here. To generate tarballs for the aarch64 port, it probably makes sense to the "classic" archiso way ^^
[21:55:16] <Antiz> to go the classic way*
[21:59:45] <DrZee> I'm building a pipeline (CI/CD) where I build the tarballs for aarch64 .... I need an aarch64 docker image for that so looked into archlinux-docker .. and build a working docker image for aarch64... but when looking under the hood I was wondering...
[22:00:00] <Antiz> Ah okay, makes sense
[22:04:10] <DrZee> I'll try to avoid using "full" virtual machines if not necessary ... the AUR packages I maintain are all build on a arch Linux x86 docker image ... works perfectly.
[22:05:50] <Antiz> To be fair, I don't know the actual / historical decisions that were made regarding the Docker image's rootfs general. What I said is based on my understanding & the work I've done on the WSL Arch image (which is heavily inspired by the way the Docker image is built), mixed with a bit of guessing / assumption 👼
[22:06:19] <Solskogen> isn't pacstrap supposed to be used?
[22:06:32] <Antiz> Some people in #archlinux-releng might have more input and be able to confirm (or not) what I said.
[22:07:44] <Solskogen> or is it so that pacstrap and fakeroot doesn't play ball?
[22:10:23] -!- titus_livius has joined #archlinux-ports
[22:10:26] <Antiz> Solskogen: No they don't, pacstrap needs actual/real root privileges
[22:17:21] -!- marmis has quit [Quit: Bye!]
[22:18:28] -!- marmis has joined #archlinux-ports
[22:24:39] -!- marmis has quit [Quit: Bye!]
[22:25:46] -!- marmis has joined #archlinux-ports
[23:10:29] -!- gromit has quit [Quit: ZNC 1.10.1+docker-git--rc11.10.x-znc-1.10.0-12-gee54fb12 - https://znc.in]
[23:14:35] -!- gromit has joined #archlinux-ports
[23:27:40] <bschnei> idealseal: welcome! Are you looking to help with testing? packaging? Both? :)