| 20 Jan 2026 |
Randy Eckenrode | There’s a chance it’s not compatible with the headers (especially on x86_64-darwin). | 21:25:36 |
Randy Eckenrode | * | 21:25:51 |
emily | when I checked staging:stdenv had failed like 7 jobs in a row | 21:25:57 |
emily | but perhaps there just weren't any rebuilds then? | 21:26:15 |
Randy Eckenrode | It builds for me if I use sudo to not go through the daemon or if I run it interactively. | 21:26:37 |
Randy Eckenrode | There was a Gentoo issue I linked with the same error that made it around like it doesn’t happen consistently. | 21:27:14 |
Randy Eckenrode | https://bugs.gentoo.org/950943 | 21:27:36 |
emily | https://hydra.nixos.org/jobset/nixpkgs/staging | 21:27:58 |
emily | doesn't look healthy | 21:28:03 |
Sarah Clark | I'll pass the comment upstream. Thanks. | 21:28:13 |
emily | that won't stop it blocking -next or testing staging PRs though, right? | 21:29:04 |
emily | Meson rebuilds stdenv so it's always staging. you should have seen LLVM rebuild most likely. but I just meant it's hard to test a PR that rebuilds bootstrap when bootstrap is broken | 21:30:06 |
Sarah Clark | Ah, this isn't the Meson build (that's done, successfully). I was just picking up a PR review | 21:30:55 |
emily | ah, whoops. I was talking about Meson, yeah. | 21:32:19 |
Sarah Clark | No worries hug | 21:39:22 |
Sarah Clark | (going offline, have been fighting an infection and I'm now exhausted) | 21:41:08 |
samasaur | feel better! | 21:43:18 |
| 21 Jan 2026 |
Ihar Hrachyshka | wireshark-qt segfaulted in hydra when wrapping qt app in fixup phase and I can't reproduce it on any of machines I have.
The log: https://cache.nixos.org/log/qvza680ibwq9qq6prqmqajmgywq2hiz9-wireshark-qt-4.6.3.drv
wrapping Qt applications in /nix/store/vmqmdn9lwpbdjvmnpyr728w0k44xnhvy-wireshark-qt-4.6.3/bin /nix/store/vmqmdn9lwpbdjvmnpyr728w0k44xnhvy-wireshark-qt-4.6.3/sbin /nix/store/vmqmdn9lwpbdjvmnpyr728w0k44xnhvy-wireshark-qt-4.6.3/libexec /nix/store/vmqmdn9lwpbdjvmnpyr728w0k44xnhvy-wireshark-qt-4.6.3/Applications
/nix/store/n4lgqnsbzr7m5i50csw6wa3z89q7s8zg-wrap-qt6-apps-hook/nix-support/setup-hook: line 80: 69662 Done find "$targetDir" ! -type d -executable -print0
69663 Segmentation fault: 11 | while IFS= read -r -d '' file; do
isELF "$file" || isMachO "$file" || continue; if [ -h "$file" ]; then
target="$(readlink -e "$file")"; echo "wrapping $file -> $target"; rm "$file"; makeQtWrapper "$target" "$file";
else
if [ -f "$file" ]; then
echo "wrapping $file"; wrapQtApp "$file";
fi;
fi;
done
Any ideas on how to debug that or a probable cause? the code is really simple. Is it bash itself that segfaulted?
| 02:22:27 |
| @vali0004:matrix.org left the room. | 02:35:46 |
Austin Horstman | anyone familiar with what changed with latest firefox-devedition on macos? lol... build went from 1h 45 min to 4 hr timeout in hydra and my local build is around 3 hours atm :S | 03:01:43 |
Austin Horstman | * anyone familiar with what changed with latest firefox{-devedition}-unwrapped on macos? lol... build went from 1h 45 min to 4 hr timeout in hydra and my local build is around 3 hours atm :S | 03:10:24 |
vcunat | https://github.com/NixOS/nixpkgs/pull/457369 | 10:29:51 |
vcunat | i.e. the resulting binary should be noticeably more efficient. | 10:30:16 |