| 13 Apr 2026 |
 Mic92 | hexa (signing key rotation when): Simon Hauser I started integrating ofborg infra in the nixos infra repo, in the hope that we than have an easier time keeping everything up-to-date: https://github.com/NixOS/infra/pull/1009 | 07:37:13 |
| Mic92 | I also noticed my age keys were not enrolled for these machines | 07:37:40 |
| Mic92 | dgrig: did this happen? https://github.com/NixOS/infra/issues/968 | 08:40:02 |
 hexa | it did | 11:16:38 |
| hexa | https://github.com/FiloSottile/age/issues/537#issuecomment-1907361675
https://github.com/FiloSottile/age/issues/142#issuecomment-1001161195 | 11:27:33 |
| hexa | also https://github.com/str4d/rage/issues/272#issuecomment-2179600832 | 11:30:36 |
 uep | https://hydra.nixos.org/jobset/nixos/unstable shows 4 days since the last eval. I know there were disk space and other issues, as above, but I'm assuming they're now resolved. Is there still an issue? | 22:23:41 |
| uep | Or, as I suddenly realise, is this just awkward timing and rounding - an eval may kick off any minute on the every-2-days schedule, with one having been missed? | 22:24:30 |
| uep | I'll assume it's that for now, there's no real reason to expect an eval to have started immediately after the issue was resolved. | 22:27:13 |
| hexa | dovecot changes caused eval errors | 22:31:20 |
| hexa | merged https://github.com/NixOS/nixpkgs/pull/509691 just now to unblock | 22:32:01 |
| hexa | still not enough | 22:32:49 |
| hexa | error:
Failed assertions:
- The option definition `services.dovecot2.enableImap' in `/home/hexa/git/nixos/master/nixos/tests/gitlab/gitlab.nix' no longer has any effect; please remove it
Set 'services.dovecot2.settings.protocols.imap = true/false;' instead.
- services.dovecot2: Since Dovecot 2.4, the option 'services.dovecot2.settings.dovecot_config_version' must be explicitly set.
To retain compatibility with future updates, set the following and manually update as needed.
services.dovecot2.settings.dovecot_config_version = "2.4.3";
Alternatively, you can automatically update to newer versions of the configuration format, which might break compatibility with future updates.
services.dovecot2.settings.dovecot_config_version = config.services.dovecot2.package.version;
See <https://doc.dovecot.org/latest/installation/upgrade/2.3-to-2.4.html>.
- services.dovecot2: Since Dovecot 2.4, the option 'services.dovecot2.settings.dovecot_storage_version' must be explicitly set.
Set it to the oldest version the storage should stay compatible with, for example the following for the currently selected version.
services.dovecot2.settings.dovecot_storage_version = "2.4.3";
See <https://doc.dovecot.org/latest/installation/upgrade/2.3-to-2.4.html>.
note: trace involved the following derivations:
derivation 'vm-test-run-gitlab'
derivation 'nixos-test-driver-gitlab'
derivation 'nixos-vm'
| 22:33:00 |
| hexa | https://githuib.com/nixos/nixpkgs/pull/509688 | 22:33:21 |
| hexa | Redacted or Malformed Event | 22:33:47 |
| 14 Apr 2026 |
 leona | also we should merge staging-nixos bc. of https://github.com/NixOS/nixpkgs/pull/509378 | 07:53:01 |
 Vladimír Čunát | Right, we only have ~ 10% of nixos.tests.* built on the current eval. | 07:57:38 |
| Vladimír Čunát | So it makes sense for efficiency. | 07:58:47 |
| Vladimír Čunát | https://github.com/NixOS/nixpkgs/pull/509851 | 08:00:28 |
| Vladimír Čunát | * https://github.com/NixOS/nixpkgs/pull/509851 🚀 | 08:00:39 |
| Vladimír Čunát | It also switches the default Nix branch. | 08:01:33 |
| Vladimír Čunát | So it's not really minor 😅 | 08:01:51 |
| Vladimír Čunát | But best to get that fast, given the stabilization of 26.05. | 08:03:34 |
 Sergei Zimmerman (xokdvium) | I think that already got merged to master before | 08:07:28 |
| Vladimír Čunát | Does anyone understand what needs to be done about those redirects in NixOS manual? | 08:18:20 |
| Vladimír Čunát | I'm really confused. The NixOS test driver added some option, which needs some other action apparently? While I do see the error, I don't get what the redirect should be like. | 08:19:41 |
| Vladimír Čunát | (and I can't even browse the manual, as it refuses to build) | 08:21:06 |
| Vladimír Čunát | I feel like we've gotten to annoyingly long iteration times on some things. Between this check failing, CI running a few times, merge queue, and nixos-unstable eval, I need at least 3 hours to do this "trivial" operation - that's if I prioritize kicking it whenever it needs to - longer if I did not. | 08:36:13 |
| Vladimír Čunát | (and I'm even skipping some steps, e.g. pushing to staging-nixos directly, even though it's shouting at me) | 08:37:26 |
| Vladimír Čunát | I wouldn't mind so much if it didn't need so much attention. Because it's always wait 10-15 minutes then do something, then again wait and again it needs some button-pushing. | 08:39:05 |
