| 18 Jan 2026 |
 hexa | missing a cycle is expensive indeed | 22:30:16 |
| hexa | there are options to delay the start of a cycle or merge things into an active cycle | 22:32:53 |
 Mic92 | I wonder if we can convince upstream to restructure systemd into meson submodules. It's also not nice for normal package managers to have dependency cycles | 22:33:53 |
 Arian | hot take. maybe we should only support darwin on unstable and not on the release branches. Those release branches are for NixOS after all. Having NixOS updates delayed because of darwin feels backwards | 23:03:40 |
 Winter | didn’t we have release branch jobsets for darwin | 23:04:52 |
| Arian | yeh we have https://hydra.nixos.org/jobset/nixpkgs/nixpkgs-25.11-darwin but we don't have a separate staging-25.11 flow for just darwin iirc? | 23:06:13 |
 Fabián Heredia | https://hydra.nixos.org/build/318935477
Can someone cancel this build? Seems to be stuck printing OOM in the tail, been running for 15 hrs :C | 23:13:27 |
 emily | huh? nixpkgs-YY.MM-darwin exists | 23:18:50 |
| emily | there are lots of users of stable branches on Darwin | 23:19:03 |
| hexa | Redacted or Malformed Event | 23:27:20 |
| 19 Jan 2026 |
 Randy Eckenrode | I didn’t think that would work for some reason. | 00:19:11 |
| * Randy Eckenrode raises his hand. | 00:20:41 |
 Sergei Zimmerman (xokdvium) | This applies a fix to nix 2.33 with glibc 2.42 https://github.com/NixOS/nixpkgs/pull/481478 and also disables flaky aarch64-darwin tests since we haven’t come to a good resolution. I’ll try go get around ~tomorrow to sensing those to 25.11 too | 01:23:50 |
| Sergei Zimmerman (xokdvium) | * This applies a fix to nix 2.33 with glibc 2.42 https://github.com/NixOS/nixpkgs/pull/481478 and also disables flaky aarch64-darwin tests since we haven’t come to a good resolution. I’ll try go get around ~tomorrow to sending those to 25.11 too | 01:24:25 |
| Fabián Heredia | This should fix some of the breakage on staging-next related to glibc 2.42: https://github.com/NixOS/nixpkgs/pull/481503 | 02:29:11 |
| Fabián Heredia | Currently building | 02:29:18 |
 Vladimír Čunát |
there are lots of users of stable branches on Darwin
So far I don't see anyone complaining that it lags on a 10 days old commit. (because issues in Nix itself)
| 07:06:49 |
| Vladimír Čunát | It's not really possible to separate staging* for the platforms. We only have one master branch and one release-25.11 branch. | 07:07:55 |
| Vladimír Čunát | And changing that approach would seem really complicated. | 07:08:20 |
| Fabián Heredia | 
Download image.png | 16:11:59 |
| Fabián Heredia | Can someone help me with a new eval on staging-next? https://github.com/NixOS/nixpkgs/pull/481503
This should fix the biggest breakage on x86_64-linux (compiler-rt for rocm) | 16:12:06 |
 leona | done | 16:17:12 |
| Fabián Heredia | ty <3 | 16:19:54 |
| Randy Eckenrode | I would have caught the setproctitle issue sooner if it rebuilt my system configs more frequently. I got out of the habit while running a staging branch for a while (because full rebuilds on staging take forever). It wouldn’t be the first Darwin issue I’ve fixed. | 17:32:59 |
| Randy Eckenrode | What’s wrong with Swift on Darwin this time? ☹️ | 17:33:20 |
| Randy Eckenrode | Oh, swiftc crash during build. Not touching that one with a 10m pole. I’m super burnt out on the Swift 6.2 update. Those crashes are not fun to debug. | 17:37:27 |
 whispers (it/fae) | it's not darwin specific, swift is failing on master on linux too (iirc gcc 15 fallout, though idk if the cause is different on darwin) | 17:37:30 |
| Randy Eckenrode | Swift shouldn’t even be using GCC to build. | 17:37:45 |
| whispers (it/fae) | oh yeah right, darwin's issue are separate | 17:38:06 |
| Randy Eckenrode | AFAIK it’s technically possible but not really supported upstream. | 17:38:21 |
