#archlinux-ports | Logs for 2025-09-02

Back
[04:07:54] -!- drathir_tor has quit [Remote host closed the connection]
[05:20:19] <Solskogen> gromit: we just renamed the repo to [any]
[05:22:23] <Solskogen> we don't have sig files for packages in release (yet) - so you just have to trust us. SigLevel=Never is your friend ;-)
[05:27:33] <Solskogen> the any repo have sig files
[05:44:04] -!- drathir_tor has joined #archlinux-ports
[10:12:40] <Solskogen> gromit: you'll be awarded internet points if you're able to get rocm-llvm to compile and/or work on aarch64
[10:14:35] <jelle> https://src.fedoraproject.org
[10:14:36] <phrik> Title: Tree - rpms/rocm-runtime - src.fedoraproject.org (at src.fedoraproject.org)
[10:16:12] <gromit> Solskogen: I can of course look into that at some point, but is there any point in getting it to run? Are there ARM chipsets that would benefit from ROCm support?
[10:16:56] <jelle> debian seems to build it aarch64/ppc64 so :)
[10:16:58] <jelle> !next
[10:16:59] <phrik> Another satisfied customer! NEXT!
[10:17:31] <Solskogen> Some might want to use their AMD GPU on their ARM board :-)
[10:18:05] <Solskogen> oh, btw. Please reinstall pacman to get the correct mirrors and makepkg.conf.
[10:19:40] <jelle> gromit: more serious ARM boards have PCI
[10:22:13] <Solskogen> my Orion rig has a Geforce 2060
[10:29:41] <gromit> jelle: I only have an unserious VM right now hehe
[10:30:15] <jelle> well, it is not that readily available basically only for enterprise
[10:30:58] <jelle> which funnily enough also has UEFI :D
[10:32:37] <gromit> Solskogen: done :)
[10:44:56] <Solskogen> done?! how?!
[10:50:18] <gromit> Solskogen: I was referring to the pacman thingy, not rocm-llvm :p
[10:58:38] <Solskogen> oh, okay :-D
[11:00:00] <Solskogen> just so we're clear... just because a package is in the release-repo doesn't mean that I can guarantee that the package works - or even starts.
[11:00:14] <Solskogen> (but everything I've tried works!)
[11:09:07] <gromit> Solskogen: sure thats clear, it's kinda the same for the main arch repos ;)
[11:10:19] <gromit> I got kinda surprised by the amount of storage it needs to compile the linux kernel :o Directly filled up the root partition of my vserver x)
[11:13:10] <Solskogen> how much memory do you have?
[11:16:00] <gromit> Solskogen: 40 GB, but now I have attached a storage volume via CIFS :) Speed doesn't really matter in this setup I guess :D
[11:16:36] <Solskogen> 40G should be enough for tmpfs :-)
[14:44:21] <nl6720> what does `uname -m` return on AArch64, and is it consistent on all AArch64 systems?
[14:45:09] <gromit> nl6720: uname -m gives me "aarch64"
[14:46:48] <nl6720> gromit: thanks!
[14:47:03] <nl6720> I couldn't find a consistent answer online
[15:38:49] <gromit> noice, built ant booted mainline successfully
[16:25:47] <Solskogen> It's quite consistent
[16:26:04] <Solskogen> but some stuff thinks arm64 is better
[16:29:25] <Solskogen> heh, I got virtualbox compiled. I wonder if that works!
[16:31:00] <Solskogen> (for those who feel very adventurous can enable the staging repo)
[16:51:47] * gromit is feeling adventurous
[17:13:30] <Solskogen> virtualbox did not work (not that I thought it would)
[17:53:36] <bschnei> from what I've seen, most of the world outside of the kernel uses aarch64 to refer to ARMv8 and later instruction sets.
[17:53:46] <bschnei> kernel uses arm64
[18:57:06] <Solskogen> go likes arm64
[19:00:18] <Solskogen> I'm still a bit mad that amd64 didn't stick
[19:35:19] <jelle> debian uses amd64
[20:06:29] <Solskogen> how come some packages have two checksums? both b2sum and sha512 for instance
[20:08:17] -!- h|weechat has quit [Ping timeout: 248 seconds]
[20:10:28] -!- h|weechat has joined #archlinux-ports
[20:10:43] <anthraxx|M> Solskogen: dual hashing makes it resistant if one algo may be broken. Colliding both at the same time is unlikely.
[20:31:15] <Solskogen> colliding on sha512 is quite unlikely as well :-)
[20:47:58] <Solskogen> jelle: https://gitlab.archlinux.org :-)
[20:48:00] <phrik> Title: Move umockdev into depends. meson fails if !check is set: (!1) · Merge requests · Arch Linux / Packaging / Packages / feedbackd · GitLab (at gitlab.archlinux.org)
[20:48:20] <jelle> ooo in this case I can just bug my coworker :P
[20:48:54] <jelle> oh this is feedbackd, not umockdev
[20:56:11] <Solskogen> I'm afraid you have to utter one or two opinions here :-)
[20:59:15] <Solskogen> question: how do you cope with stuff like svt-hvec? upstream project is discontinued, and the github repo was closed in 2024.
[22:00:40] -!- drathir_tor has quit [Remote host closed the connection]
[22:01:02] -!- drathir_tor has joined #archlinux-ports
[22:10:13] -!- titus_livius has joined #archlinux-ports
[23:12:12] <bschnei> Solskogen: is it FTBFS? Gentoo can be a good source for patches. I have an MR open for potrace which has been unmaintained for 6 years. loqs found a patch at Gentoo that allowed it to build on aarch64 with LTO enabled 😎