| 9 Jun 2026 |
 raitobezarius | i bisected the issue to the performance curl patch | 18:22:18 |
| raitobezarius | https://github.com/NixOS/nixpkgs/pull/528906#issuecomment-4662501242 | 18:22:19 |
| raitobezarius | cc Scrumplex | 18:22:59 |
 Sergei Zimmerman (xokdvium) | Looks like it's failing the tests on darwin? | 18:27:32 |
 leona | I can look tomorrow | 18:28:39 |
| 10 Jun 2026 |
 dramforever | aaaaaa | 00:05:50 |
| dramforever | re: lix i mean | 00:05:55 |
| dramforever | Download r1g8ou6zotub1.mp4 | 00:06:47 |
| Sergei Zimmerman (xokdvium) | Curl and lix are sworn enemies apparently | 00:32:09 |
 hexa | lix uses curl's async api and I think that's where the problems in curl are. | 00:38:55 |
 Myria | If this helps at all, had a similar issue in staging with 3.4.3: https://github.com/NixOS/nixpkgs/pull/525465 | 01:24:09 |
 vcunat | A set of kernel updates -> 33k NixOS jobs to build. (and we haven't evaluated 25.11 yet) | 08:03:27 |
 K900 | @emily wanted to remove non-blocking tests from the evalf | 08:04:07 |
| K900 | @emily wanted to remove non-blocking tests from the eval | 08:04:13 |
| vcunat | The weird part is that nixpkgs:unstable got like 15k as well. | 08:04:33 |
| vcunat | (and I failed to find any big PR except staging-nixos) | 08:05:00 |
| vcunat | But in terms of time spent by Hydra on it, the (non-blocking) NixOS tests will have a vast majority. | 08:05:52 |
 Grimmauld (any/all) | We could also try migrating more shit to nspawn container tests, my understanding is those are cheaper to run? | 08:07:13 |
 Emil Thorsøe | I somehow got the idea that this was already decided? | 08:07:22 |
| K900 | It wasn't decided that we should go through with it | 08:07:53 |
| vcunat | Running them isn't the expensive part for Hydra.nixos.org currently. | 08:08:26 |
| K900 | But we have rough consensus that we want to try it and see the results | 08:08:28 |
| Emil Thorsøe | https://discourse.nixos.org/t/security-fixes-should-bypass-staging-unbound-cve-batch-as-a-case-study/78147/11 oh yeah, too strong a read of this | 08:08:56 |
| leona | https://github.com/NixOS/nixpkgs/pull/530261 | 08:20:37 |
| leona | that's the guarded version for -next. The same also works on staging, so I'll post the unguarded version there and on staging-26.05 | 08:20:57 |
| Sergei Zimmerman (xokdvium) | In reply to @leona:leona.is
https://github.com/NixOS/nixpkgs/pull/530261 Ah, gotta love the slop comments:
https://github.com/tridge/rsync/blob/pr-chmod-test-fix/t_chmod_secure.c#L59C3-L61C64
O_RESOLVE_BENEATH is a bsd thing… | 10:44:16 |
 Alyssa Ross | O_BENEATH is a BSD thing | 11:07:57 |
| Sergei Zimmerman (xokdvium) | I think O_BENEATH now is deprecated because it had some SNAFUs | 11:10:10 |
| Sergei Zimmerman (xokdvium) | https://reviews.freebsd.org/D25886 | 11:10:21 |
| Alyssa Ross | oh TIL | 11:10:24 |
