Exotic Nix Targets | 333 Members | |
| 103 Servers |
| Sender | Message | Time |
|---|---|---|
| 30 Jul 2023 | ||
| I ran that without anything added to my sandbox. im re-running with `qemu-riscv64` added to my sandbox to see if this is another one of those things of enabled binfmts leaking into the sandbox and causing things to think it has an elf interpreter for an executable when that interpreter isn't actually on PATH like the other issue we were seeing | 02:12:35 | |
| be back in a bit | 02:12:55 | |
| ok i get the failure all the same so it's probably not related to the binfmts at least | 02:15:55 | |
| implr: to confirm, is that error log i linked above showing a failure on the same test yours fails on? | 02:26:38 | |
| I left mine running overnight and have gotten a failure in
This is quite possibly a bug in psutil's test suite where it fails when at 100% swap usage. | 03:20:14 | |
| yeah, psutil failure looks related to RAM usage. | 06:42:49 | |
coreutils' mv-i test looks simple: https://github.com/coreutils/coreutils/blob/master/tests/mv/i-1.pl . I wonder why it printed help banner. Aaaand I reproduced it in gentoo chroot as well \o/: $ nix build nixpkgs#coreutils --rebuild. | 06:53:26 | |
In reply to @artemis:arty.chatyep, same | 10:43:20 | |
| and it actually fails on native (sufficiently new) nixpkgs for me as well, so the problem is just coreutils9.3, not riscv/cross compile related | 10:43:56 | |
In reply to @implr:hackerspace.plNot every coreutils 9.3 fails on native. I have a working RV64GC_Zba_Zbb build on 23.11.git.7c67f006ea0M without any overrides (besides the custom GCC 13 stdenv). | 10:46:31 | |
| 21:50:30 | ||
| 31 Jul 2023 | ||
| 18:15:45 | ||
| 22:02:53 | ||
| 22:40:26 | ||
| 1 Aug 2023 | ||
| 01:12:22 | ||
| Download usellvm-16.patch | 14:04:38 | |
I managed to get pkgsLLVM.nix working with an import <nixpkgs> { overlays = [(self: super: { llvmPackages = self.llvmPackages_16; })]; } overlay; however, I did have fix a few other minor things (libseccomp doCheck = false; lowdown hardeningDisable fortify; boost NIX_CFLAGS_LINK = prev.NIX_CFLAGS_LINK + " --target=${self.stdenv.hostPlatform.config}";, and disable 3 tests from nix) | 14:06:30 | |
| raitobezarius: any thoughts on the above patch? I'm on x86_64-unknown-linux-gnu, I have the impression you primarily were trying to make a darwin platform work | 14:07:10 | |
| I've commented the hacks so hopefully the motivation might be clear. Do you think this is suitable to send as a patch to make llvmPackages_16 work? | 14:07:38 | |
| [apols, I have to run away for a few hours shortly] | 14:07:55 | |
| I'd say open a PR, expose the problem solved by this and tag the LLVM maintainers | 14:08:00 | |
| it's hard to review like this stuff | 14:08:06 | |
| raitobezarius: which branch do I send it for -- is master OK given that little will depend directly on llvmPackages_16? | 14:12:27 | |
| master is fine | 14:12:56 | |
| 14:15:22 | ||
| Sent as https://github.com/NixOS/nixpkgs/pull/246577. | 14:15:59 | |
| 2 Aug 2023 | ||
| raitobezarius: I've been doing some testing of the above on aarch64-darwin. I found that plain pkgsLLVM.stdenv doesn't build on master, I get a silent failure in compiler-rt. NIX_DEBUG exposes the issue that preConfigure is trying to set Is pkgsLLVM known (meant?) to work on this platform? Is there a darwin platform it works on? | 09:24:55 | |
| part of why I haven't touched this recently is that I don't have a darwin anything... | 13:17:21 | |
| sorear: Yeah, it's a challenge. I can see from the git blame lens (and also through the diff-the-nix-llvm-recipies-through-the-versions) as I hack on things that people have gone through many iterations of fixing things for different environments. | 13:20:44 | |
| I see you're only fixing this for _16 so if I want pkgsLLVM _15 to work I | 13:21:38 | |