#archlinux-ports | Logs for 2025-10-24

Back
[00:53:45] <bschnei> The MR Solskogen is referring to is here: https://gitlab.archlinux.org Reports of the generic linux package working (or not) and really any feedback at all is appreciated!
[00:53:46] <phrik> Title: Sign in · GitLab (at gitlab.archlinux.org)
[01:00:21] -!- hrtk_ has quit [Ping timeout: 252 seconds]
[01:01:07] -!- hrtk_ has joined #archlinux-ports
[01:11:16] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[01:41:00] -!- h|weechat has quit [Ping timeout: 256 seconds]
[01:48:01] -!- h|weechat has joined #archlinux-ports
[01:51:29] -!- h|weechat has quit [Max SendQ exceeded]
[02:02:18] -!- h|weechat has joined #archlinux-ports
[02:20:49] -!- hrtk_ has joined #archlinux-ports
[02:27:46] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[02:30:25] -!- hrtk_ has joined #archlinux-ports
[02:38:25] -!- hrtk_ has quit [Ping timeout: 246 seconds]
[03:53:14] -!- hcmb_ has joined #archlinux-ports
[03:53:14] hcmb is now known as Guest5434
[03:53:14] -!- Guest5434 has quit [Killed (iridium.libera.chat (Nickname regained by services))]
[03:53:15] hcmb_ is now known as hcmb
[05:08:10] -!- hrtk_ has joined #archlinux-ports
[05:17:28] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[06:30:19] nl6720_ is now known as nl6720
[06:30:32] -!- nl6720 has quit []
[06:30:55] -!- nl6720 has joined #archlinux-ports
[07:06:13] -!- hrtk_ has joined #archlinux-ports
[07:12:22] -!- hrtk_ has quit [Remote host closed the connection]
[07:17:00] -!- hrtk_ has joined #archlinux-ports
[07:23:00] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[07:24:50] -!- hrtk_ has joined #archlinux-ports
[07:30:22] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[07:31:09] -!- hrtk_ has joined #archlinux-ports
[07:37:26] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[08:01:14] -!- hrtk_ has joined #archlinux-ports
[08:18:32] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[08:43:44] -!- hrtk_ has joined #archlinux-ports
[08:51:19] -!- hrtk_ has quit [Ping timeout: 240 seconds]
[09:30:19] -!- hrtk_ has joined #archlinux-ports
[09:39:34] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[10:18:06] -!- hrtk has quit [Ping timeout: 256 seconds]
[10:25:11] -!- hrtk has joined #archlinux-ports
[11:41:25] -!- nan25q has joined #archlinux-ports
[11:42:56] <nan25q> where can i read up on the aarch64 effort and help support it? i've been running alarm on multiple SBCs for over a decade and have plenty of practical experience
[11:43:41] <nan25q> goes without saying i'm deeply disappointed by the stagnation of alarm
[12:01:16] <strit|M> I think the best place is this chat, currently. There's a bootstrap tarball by drzee linked further up in the chat.
[12:03:25] <nan25q> can i find logs somewhere?
[12:03:53] <Antiz> nan25q: This is an early stage effort
[12:04:19] <nan25q> i'm a developer, not a user
[12:04:34] <Antiz> Currently the point is to find packages in the official repo that do not build as-is on aarch64 and submit fixing MR for those
[12:05:21] <Antiz> Within scope defined in this mail : https://lists.archlinux.org
[12:05:23] <phrik> Title: First step for non-x86_64 contributions - Arch-dev-public - lists.archlinux.org (at lists.archlinux.org)
[12:07:06] <Antiz> Solskogen, bschnei and DrZee have some further work done on their side (e.g. a repo with packages they were able to build so far and a tarball for bootstrapping)
[12:07:10] <Antiz> But this isn't official yet.
[12:07:55] <gromit> Antiz: to expand: This is not only about the packages but also about fostering the tooling and contribution ecosystem
[12:08:18] <Antiz> gromit: I was about to talk about it but you're too fast :P
[12:08:20] <nan25q> is that repo public? the bootstrapping aspect is what's most interesting to me
[12:08:31] <Antiz> nan25q: I don't think it is for now.
[12:08:41] <Antiz> But we have created https://gitlab.archlinux.org on our gitlab
[12:08:42] <phrik> Title: Arch Linux / Ports / AArch64 / project-management · GitLab (at gitlab.archlinux.org)
[12:08:57] <Antiz> See some issues opened there with tasks opened around tooling and so on
[12:08:57] <nan25q> (i currently have a frankenstein bootstrap method that utilizes armbian *shudder*)
[12:09:39] <gromit> nan25q: the packages in the repo from DrZee are bootstrapped via archlinuxarm, we'll need to figure out how to bootstrap from arch in the future
[12:09:45] <gromit> See https://gitlab.archlinux.org for a tracking issue
[12:09:46] <phrik> Title: Bootstrap the toolchain (#1) · Issues · Arch Linux / Ports / AArch64 / project-management · GitLab (at gitlab.archlinux.org)
[12:10:03] <gromit> Antiz: sorry, this time you were faster ;)
[12:10:11] <Antiz> gromit: :P
[12:11:28] <Antiz> nan25q: https://arch-linux-repo.drzee.net
[12:11:29] <phrik> Title: Index of arch/tarballs/os/aarch64/ (at arch-linux-repo.drzee.net)
[12:11:44] <Antiz> You can find the mentioned bootstrap tarball here
[12:12:35] <jelle> but too.. negate the trusting trust attack you will need two different compilers
[12:14:10] <DrZee> gromit: "the PKGs are bootstrap from archlinuxarm" .... not any more. We are now compiling running everything off itself. Yes for the first PKG compilation we use arch Linux arm as it allowed us to use familiar tooling (makepkg etc.) but we are no longer dependent in that.
[12:15:37] <gromit> DrZee: yes this is exactly why we need to figure it out, see the linked issue
[12:15:50] <gromit> This is about the initial bootstrap and the trusting trust issue
[12:41:50] -!- hrtk_ has joined #archlinux-ports
[12:45:24] <Antiz> gromit: Alternatively we could go the way Allan suggested
[12:45:53] <Antiz> i.e. bootstrap from multiple (at least 2 I guess) distros and compare the result
[12:46:28] <Antiz> It's not as trustworthy as bootstrapping from arch I guess, but it's a pretty common way to at least *mitigate* the trusting trust issue as far as I understand
[12:46:54] <Antiz> Basically what jelle suggested by using two different compilers IIUC
[12:46:56] <nan25q> if you can get bit for bit reproducible builds then this approach is good
[12:47:15] <nan25q> otherwise not sure how the comparison will help
[12:48:32] <jelle> Antiz: for mitigating tursting trust you need to compile one with llvm and the other with gcc, imo that is not what we should be doing
[12:49:18] <jelle> https://dwheeler.com
[12:49:19] <phrik> Title: Fully Countering Trusting Trust through Diverse Double-Compiling (DDC) - Countering Trojan Horse attacks on Compilers (at dwheeler.com)
[12:49:24] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[12:49:27] <Antiz> nan25q: Yes obviously, that's what I meant by "compare"
[12:49:38] <Antiz> Reproducible builds as in bit-by-bit identical
[12:50:26] <jelle> fedora has pacman packaged /o\
[12:51:12] <Antiz> jelle: Ah yeah right, so we could either build it twice with llvm and gcc or build it twice with same compiler but on different distros (and expect reproducible builds), correct?
[12:51:36] <Antiz> twice with llvm and gcc on ALARM that is.
[12:52:11] <Antiz> I guess the best scenario would still be able to bootstrap from (actual) Arch but the two above scenarios could be acceptable solutions too I guess?
[12:52:12] <jelle> wut?
[12:52:37] <Antiz> jelle: You meant two different compiler from the same distro?
[12:52:48] <nan25q> if we can boostrap a x86->aarch64 toolchain from existing arch, it seems that good enough right?
[12:53:08] <jelle> I am not sure where the trusting trust angle came in but I don't see it being too relevant
[12:53:23] <nan25q> either using something like crosstool-ng, or not
[12:53:32] <Antiz> nan25q: Sure, but we need to figure this out
[12:54:03] <Antiz> jelle: Well, as far as I understand, the fact that the package are initially bootstrapped from ALARM implies a trusting trust issue, right?
[12:54:27] <jelle> Antiz: its the same issue as you building a new ghc, openjdk or gcc
[12:54:32] <Antiz> As we rely on prebuilt compilers binaries from ALARM (in which we cannot technically establish trust)
[12:55:24] <Antiz> jelle: Sure.
[12:56:18] <Antiz> But, just to make sure I got things right, the idea would either be to be able to boostrap from Arch (given we trust our own binaries)
[12:56:48] <Antiz> Or alternatively bootstrap from different distro*s* and / or different compiler*s* and expect the same result
[12:57:17] <Antiz> Which would indicate that the compiler is not infected (or that they are both infected the same way)
[12:57:17] <jelle> documenting how to we can bootstrap another architecture is very valueable imo
[12:57:39] <jelle> Antiz: that all hinges on the toolchain being reproducible
[12:57:52] <Antiz> jelle: Yep!
[12:58:13] <jelle> and gcc-libs isn't
[12:59:45] <Antiz> Humm yeah https://reproducible.archlinux.org
[13:00:12] <jelle> you are signing yourself up for a lot of work and we can't even bootstrap our own packages atm :p
[13:01:15] <Antiz> jelle: I primarily wanted to make sure I got a correct *wide* picture of the issue, I did not volunteer :P
[13:01:58] <Antiz> Maybe I can bring the gcc-libs case at the reproducible builds summit next week :p
[13:29:00] -!- mkurz has joined #archlinux-ports
[13:30:30] -!- hrtk_ has joined #archlinux-ports
[13:36:59] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[13:39:07] -!- hrtk_ has joined #archlinux-ports
[13:45:28] -!- hrtk__ has joined #archlinux-ports
[13:47:03] -!- joske has joined #archlinux-ports
[13:47:11] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[13:52:09] -!- hrtk__ has quit [Ping timeout: 265 seconds]
[13:56:17] -!- nan25q has quit [Quit: Client closed]
[13:58:37] -!- mkurz has quit [Quit: Konversation terminated!]
[14:18:01] -!- joske has quit [Ping timeout: 264 seconds]
[14:50:02] -!- pl0rt has joined #archlinux-ports
[15:33:57] -!- hrtk_ has joined #archlinux-ports
[15:40:34] -!- hrtk_ has quit [Ping timeout: 244 seconds]
[16:07:16] -!- hrtk_ has joined #archlinux-ports
[16:13:24] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[16:14:30] -!- hrtk_ has joined #archlinux-ports
[16:21:02] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[16:28:52] -!- hrtk_ has joined #archlinux-ports
[16:34:07] -!- hrtk_ has quit [Ping timeout: 240 seconds]
[16:35:26] -!- hrtk_ has joined #archlinux-ports
[16:37:22] <Solskogen> the problem is that not everything can be cross compiled.
[16:40:23] <Solskogen> I'm packaging composable-kernel now - and the build has been running for almost 9 hours. Is that normal?! I see Feodra does it in 3 minutes, so either they're just faking it - or something is terribly wrong on my end.
[16:41:20] <Solskogen> is the build logs for x86_64 available somewhere?
[16:42:52] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[16:43:51] <Antiz> Solskogen: Sort of
[16:44:36] <Antiz> Ah nevermind they are, forgot we added that
[16:45:02] <Solskogen> nice. where? :-)
[16:45:23] <Antiz> Before, you needed to check build logs on our rebuilder instance we use for reproducible builds at https://reproducible.archlinux.org
[16:45:25] <phrik> Title: Arch Linux Reproducible Status (at reproducible.archlinux.org)
[16:45:43] <Solskogen> according to that one, it didn't build.
[16:45:43] <Antiz> But we now have a dedicated button on the archweb packages page :)
[16:46:22] <Antiz> Hummm nevermind, it seems build logs are only shown there for unreproducible packages
[16:46:29] <Antiz> So https://reproducible.archlinux.org is still the way to go
[16:46:31] <phrik> Title: Arch Linux Reproducible Status (at reproducible.archlinux.org)
[16:47:16] <Antiz> You can search for the package you need in the list (don't forget to expand the "good" list) and each packages have a dedicated "build log" button link there
[16:47:26] <Solskogen> https://reproducible.archlinux.org
[16:47:33] <Solskogen> but it comes without timestamps :/
[16:47:37] <Solskogen> (aaand it doesn't build)
[16:47:58] <Solskogen> probably because it EATS ram and cpu.
[16:48:12] <Antiz> Ah yup, it don't we expose build time anywhere though
[16:48:18] <Solskogen> I saw it used 64GB ram and 96GB of swap on our end
[16:48:21] <Antiz> I don't think we expose*
[16:48:40] <Antiz> Yeah it probably fails to build on the rebuilder because of lack of resources indeed
[16:49:00] <Antiz> FWIW, buildbtw should improve stuff on that front.
[16:49:38] <Antiz> Builds will be triggered in GitLab CI (so you'd be able to consult them, including timestamps)
[16:49:50] <Solskogen> I'm having trouble understanding how someone can create a software package like that - release it and be more or less happy with that amount of resources needed just to compile it.
[16:50:29] <Solskogen> I'm not gonna complain about onnxruntime anymore
[16:51:24] <Antiz> :/
[17:05:42] -!- hrtk_ has joined #archlinux-ports
[17:12:10] -!- hrtk__ has joined #archlinux-ports
[17:12:20] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[17:18:05] -!- hrtk__ has quit [Remote host closed the connection]
[17:40:13] -!- pl0rt has quit [Quit: Client closed]
[17:52:38] <tpkessler|M> <Solskogen> "I'm packaging composable-..." <- Yes, that's expected. It's heavily templated C++ code to facilitate loop unrolling for common tensor contractions.
[17:53:59] <tpkessler|M> Solskogen: one option is to skip all rocm packages for non x86_64 archs. AMD currently doesn't support ARM.
[18:04:37] -!- hrtk_ has joined #archlinux-ports
[18:11:50] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[18:24:53] <Solskogen> I got rocm-llvm 7.0.2 to compile at least.
[18:24:59] <Solskogen> and ollama was happy.
[18:29:02] <Solskogen> can anyone please send me a invite to matrix? I want to try to use that instead of irc.
[19:35:38] -!- hrtk_ has joined #archlinux-ports
[19:42:05] -!- hrtk_ has quit [Ping timeout: 265 seconds]
[19:42:51] -!- hrtk_ has joined #archlinux-ports
[19:52:08] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[19:53:52] -!- hrtk_ has joined #archlinux-ports
[19:59:25] -!- hrtk_ has quit [Ping timeout: 264 seconds]
[20:00:24] -!- hrtk_ has joined #archlinux-ports
[20:06:17] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[20:09:11] -!- hrtk_ has joined #archlinux-ports
[20:19:58] -!- hrtk_ has quit [Remote host closed the connection]
[20:42:34] -!- hrtk_ has joined #archlinux-ports
[20:50:08] -!- hrtk__ has joined #archlinux-ports
[20:50:46] -!- hrtk_ has quit [Ping timeout: 256 seconds]
[20:55:48] -!- hrtk__ has quit [Ping timeout: 260 seconds]
[21:20:15] -!- phrik has quit [Remote host closed the connection]
[21:23:24] -!- phrik has joined #archlinux-ports
[21:25:30] -!- hrtk_ has joined #archlinux-ports
[21:30:52] -!- hrtk_ has quit [Ping timeout: 244 seconds]
[21:33:03] -!- hrtk_ has joined #archlinux-ports
[21:40:00] -!- hrtk_ has quit [Remote host closed the connection]
[21:40:23] -!- hrtk_ has joined #archlinux-ports
[21:49:49] -!- hrtk_ has quit [Ping timeout: 264 seconds]
[21:52:39] -!- hrtk_ has joined #archlinux-ports
[22:00:37] -!- hrtk_ has quit [Ping timeout: 264 seconds]
[22:10:11] -!- titus_livius has joined #archlinux-ports
[22:23:26] -!- kxxt|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- qixiqaq|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- asuna-qwq|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- rabenda|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- strit|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- felixonmars|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- anthraxx|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- antiz|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- coelacanthus|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- orhun|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- C0rn3j|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- foxboron|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- archmatrix has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- xeonacid|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- heftig has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- wszqkzqk|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- dvzrv|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- jelle|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- tpkessler|M has quit [Quit: Bridge terminating on SIGTERM]
[22:23:26] -!- bgyorgy|M has quit [Quit: Bridge terminating on SIGTERM]
[22:36:05] -!- archmatrix has joined #archlinux-ports
[22:36:05] -!- wszqkzqk|M has joined #archlinux-ports
[22:37:43] -!- strit|M has joined #archlinux-ports
[22:37:44] -!- heftig has joined #archlinux-ports
[22:37:45] -!- anthraxx|M has joined #archlinux-ports
[22:37:46] -!- kxxt|M has joined #archlinux-ports
[22:37:47] -!- coelacanthus|M has joined #archlinux-ports
[22:37:48] -!- rabenda|M has joined #archlinux-ports
[22:37:48] -!- bgyorgy|M has joined #archlinux-ports
[22:37:49] -!- dvzrv|M has joined #archlinux-ports
[22:37:50] -!- jelle|M has joined #archlinux-ports
[22:37:51] -!- antiz|M has joined #archlinux-ports
[22:37:52] -!- asuna-qwq|M has joined #archlinux-ports
[22:37:52] -!- felixonmars|M has joined #archlinux-ports
[22:37:56] -!- tpkessler|M has joined #archlinux-ports
[22:37:57] -!- foxboron|M has joined #archlinux-ports
[22:38:01] -!- xeonacid|M has joined #archlinux-ports
[22:38:02] -!- qixiqaq|M has joined #archlinux-ports
[22:38:11] -!- orhun|M has joined #archlinux-ports
[22:38:12] -!- C0rn3j|M has joined #archlinux-ports
[23:18:01] -!- marmis has quit [Quit: Bye!]