| 4 Jan 2026 |
 Janne | cc hexa | 22:54:04 |
 hexa | ohhhhhhhhh | 22:55:16 |
| hexa | streaming logs | 22:55:17 |
| Janne | tbh, for how shitty the implementation is, it works surprisingly well | 22:58:35 |
| Janne | unfortunately, I don't know if the builds is finished and won't disconnect the streamā¦ maybe something for the future | 22:59:58 |
| 5 Jan 2026 |
|  Kleha joined the room. | 06:10:10 |
| 6 Jan 2026 |
|  eyduh (she/they) changed their profile picture. | 18:20:47 |
| eyduh (she/they) changed their profile picture. | 18:25:30 |
| 11 Jan 2026 |
|  ghpzin changed their display name from ghpzin (moved to @ghpzin:envs.net) to ghpzin. | 15:04:43 |
|  @ghpzin:envs.net left the room. | 16:18:30 |
| 14 Jan 2026 |
 Mic92 | Janne: https://github.com/NixOS/hydra/pull/1558 looks like the error on hydra.nixos.org is because of a missing migration file | 08:44:26 |
| Janne | Thank you, I wasn't aware of that list. And an additional thank you because you made the list obsolete in the first place š„³ | 13:30:15 |
| 15 Jan 2026 |
| Mic92 | https://github.com/NixOS/org/issues/227 Move nix-eval-jobs from nix-community to NixOS | 17:16:50 |
| 17 Jan 2026 |
 Sergei Zimmerman (xokdvium) | S3 uploads seem messed up sometimes? https://hydra.nixos.org/build/318599546
The uploads succeeds apparently though. I wonder if it's a bug with aws-sdk-cpp or something. | 20:01:34 |
 vcunat | I only see this one lately. (and quite a lot) | 20:11:21 |
| vcunat | A few months maybe. | 20:11:47 |
| Sergei Zimmerman (xokdvium) | Mhm aws-sdk-cpp did start defaulting to the new checksum algo. I wonder if there's a way to help diagnose this on the infra side. Maybe there's something useful in the AWS logs (just guessing here). One datapoint is that CRC64NVME has been buggy https://github.com/aws/aws-sdk-cpp/issues/3496, but that's with multipart upload? | 20:18:28 |
| Sergei Zimmerman (xokdvium) | There's the option of switching back to crc32 | 20:28:14 |
| Sergei Zimmerman (xokdvium) | But that would have to be implemented in nix and backported. Certainly doable, but would be nice to first get to the root of the issue | 20:28:47 |
 Arian | I expectd nix-eval-jobs --constituents ./nixos/release-small.nix to works but it doesn't. am I holding it wrong? | 22:06:36 |
| Arian | I get:
aggregate job 'tested' references non-existent job 'nixos.channel'
aggregate job 'tested' references non-existent job 'nixpkgs.tarball'
aggregate job 'tested' references non-existent job 'nixpkgs.release-checks'
aggregate job 'tested' references non-existent job 'nixos.tests.installer.lvm.x86_64-linux'
aggregate job 'tested' references non-existent job 'nixos.tests.installer.separateBoot.x86_64-linux'
aggregate job 'tested' references non-existent job 'nixos.tests.installer.simple.x86_64-linux'
aggregate job 'tested' references non-existent job 'nixos.dummy.aarch64-linux'
aggregate job 'tested' references non-existent job 'nixos.dummy.x86_64-linux'
aggregate job 'tested' references non-existent job 'nixos.manual.aarch64-linux'
aggregate job 'tested' references non-existent job 'nixos.manual.x86_64-linux'
aggregate job 'tested' references non-existent job 'nixos.tests.acme.http01-builtin.aarch64-linux'
aggregate job 'tested' references non-existent job 'nixos.tests.acme.http01-builtin.x86_64-linux'
| 22:06:54 |
| Arian | Oh apparently when we moved hydra to nix-eval-jobs we just pass --force-recurse ? Is all this recurseForDerivations annotations that we carefully put in all the hydra jobs in release.nix completely superfluous these days? Can all that code die now? | 22:15:51 |
| vcunat | That's on nixpkgs side to distinguish where to recurse and where not to. | 22:18:17 |
| vcunat | (recursing everywhere would've been easy even in the past, I believe) | 22:18:30 |
| vcunat | For example right now on master we have a full python312Packages set, but we don't want Hydra to build (or eval) it all. | 22:20:15 |
| Arian | Ah I see! | 22:25:52 |
| 19 May 2021 |
|  @eelco:nixos.org changed the history visibility to "world_readable" from "shared". | 15:39:32 |
|  cw (28 cycles) joined the room. | 15:40:25 |
|  @grahamc:nixos.org | 16:00:29 |
| @eelco:nixos.orgchanged room power levels. | 16:01:22 |
