| 21 Nov 2023 |
 raitobezarius | I don't see any missing attribute name | 12:37:03 |
 Linux Hackerman | that's not eval though | 12:37:18 |
 Lily Foster | In reply to @raitobezarius:matrix.org
Also the lib tests is fucking flakey almost broken? it was actually a test on the nix package: https://gist.github.com/GrahamcOfBorg/8c7705fdef8ee5b57863a2127be92285#file-ofborg-eval-lib-tests-L1611 | 12:37:45 |
| raitobezarius | but wouldn't "Complete, with errors" mean that the evaluation is OK but some of the "evaluation tests" failed? | 12:37:50 |
| Lily Foster | * it was actually a test on the nix package: https://gist.github.com/GrahamcOfBorg/8c7705fdef8ee5b57863a2127be92285#file-ofborg-eval-lib-tests-L1611 | 12:37:52 |
| raitobezarius | In reply to @lily:lily.flowers
it was actually a test on the nix package: https://gist.github.com/GrahamcOfBorg/8c7705fdef8ee5b57863a2127be92285#file-ofborg-eval-lib-tests-L1611 how do you know if it's nix 2.3, nix 2.13 or nix 2.18 ? | 12:38:17 |
| raitobezarius | given nix 2.18 is still building during failures | 12:38:38 |
| raitobezarius | it's either Nix 2.3 or Nix 2.13 | 12:38:41 |
| raitobezarius | error: 1 dependencies of derivation '/nix/store/q9sy9gm77sj1dlln90qjxr58s2pc098x-nixpkgs-lib-tests-nix-2.3.17.drv' failed to build
| 12:39:01 |
| raitobezarius | it's nix 2.3 yep | 12:39:03 |
| Lily Foster | yeah i knew i saw the drv name somewhere | 12:39:14 |
| Lily Foster | glad you found it. i'm on mobile and was trying to search | 12:39:24 |
| raitobezarius | but locally I reproduced it for Nix 2.18 too | 12:39:48 |
| raitobezarius | the failure | 12:39:50 |
| raitobezarius | which started to make me very confused now | 12:39:58 |
| Lily Foster | this line is yhe actual drv that failed: https://gist.github.com/GrahamcOfBorg/8c7705fdef8ee5b57863a2127be92285#file-ofborg-eval-lib-tests-L1699 | 12:40:03 |
| raitobezarius | yep | 12:40:19 |
| Lily Foster | In reply to @raitobezarius:matrix.org
which started to make me very confused now maybe the hydra runners are just built different and the test doesn't fail there, and it just gets substituted | 12:40:41 |
| Lily Foster | actually that would be cursed. is this just a flaky upstream test? | 12:41:05 |
| raitobezarius | it failed locally too | 12:41:55 |
| Lily Foster | https://github.com/NixOS/nix/issues/4683 | 12:44:38 |
| Lily Foster | ^ maybe related? | 12:44:47 |
| Lily Foster | or some of the other options here idk: https://github.com/NixOS/nix/issues?q=hashDerivationModulo | 12:45:46 |
| Lily Foster | In reply to @raitobezarius:matrix.org
it failed locally too why did the nix 2.3.17 builds pass on hydra then? https://hydra.nixos.org/build/239849673 | 12:47:48 |
| Lily Foster | * why did the nix 2.3.17 builds pass on hydra then? https://hydra.nixos.org/build/239849673 | 12:48:20 |
| raitobezarius | It's flakey as hell is my only hypothesis | 12:49:12 |
| Lily Foster | then why don't i see failures or restarts in hydra 🤔 | 12:50:12 |
| Lily Foster | i mean it probably is flaky, i guess the hydra runners just get lucky | 12:50:29 |
| Lily Foster | i don't see how a computation could be flaky though, so that's terrifying as hell | 12:50:41 |
| Lily Foster | maybe open a nix issue? especially if you can repro on 2.18? | 12:52:13 |
