| 1 Aug 2021 |
 lukegb (he/him) | I see the problem, GDM 40.1 changed the props that get checked | 22:20:51 |
| lukegb (he/him) | https://github.com/NixOS/nixpkgs/pull/132366 for that one | 22:27:37 |
| 2 Aug 2021 |
 Jamie | what's up with nixos.tests.shadow.x86_64-linux on unstable-small? https://hydra.nixos.org/build/149348475 | 06:31:42 |
| Jamie | * what's up with nixos.tests.shadow.x86_64-linux on unstable? https://hydra.nixos.org/build/149348475 | 06:31:47 |
| Jamie | it succeeds for me locally | 06:31:57 |
| Jamie | i guess it needs a retry? | 06:32:05 |
| Jamie | the gnome test also appears to have failed and succeeds for me, could that get a retry too? | 06:45:45 |
| Jamie | oh weird, the test as it succeeds for me has a different store path to what's showing on hydra | 06:49:39 |
| Jamie | maybe i'm reproducing wrong | 06:49:42 |
 Vladimír Čunát | I'm restarting the gnome test a few times (tested locally, too). | 06:49:42 |
| Vladimír Čunát | Oh, I meant the gnome-xorg one evaluation before. | 06:52:09 |
| Vladimír Čunát | * Oh, I meant the gnome-xorg test one evaluation before. | 06:52:21 |
| Jamie | yeah now the problem is wayland i guess haha | 06:52:25 |
| Jamie | I'm pretty confused about my repro though...
~/P/nixpkgs ❯❯❯ git status
HEAD detached at aaf9676fbb7
nothing to commit, working tree clean
~/P/nixpkgs ❯❯❯ nix-instantiate nixos/release-combined.nix -A nixos.tests.gnome.x86_64-linux
warning: you did not specify '--add-root'; the result might be removed by the garbage collector
/nix/store/xv9jiajw9040bqllrjj8v27ra1rphcr1-vm-test-run-gnome.drv
| 06:58:07 |
| Jamie | whereas hydra says the test is /nix/store/08qsx05zb7xis63qwxjdbjk092sgwn50-vm-test-run-gnome.drv | 06:58:23 |
| Jamie | on that same commit | 06:58:39 |
| Jamie | so while it succeeds for me, idk if i'm actually testing something different | 06:59:38 |
| Jamie | hm yeah there's some evaluation impurity i guess? it matches if i evaluate it from /nix/store/v316zxg1w7x8343xs9lmhkh5hvjmprsq-source | 07:02:26 |
| Vladimír Čunát | .git/ or something? Most likely no significant difference. I've been reproducing tests like this for years. | 07:03:39 |
| Jamie | cool repro with the matching hash succeeded | 07:03:56 |
| Jamie | yeah i suppose so | 07:03:59 |
| Jamie | yay and it succeeded on hydra | 07:04:31 |
| Vladimír Čunát | Yes, both tests succeeded now. | 07:04:42 |
| Jamie | now for chromuium :D | 07:04:54 |
| Vladimír Čunát | Yes, that will be several hours at least, I'm afraid. | 07:05:12 |
| Jamie | i wonder if my local chromium build on my desktop will beat it | 07:05:26 |
| Vladimír Čunát | Yes, that's not so hard. The machine is doing multiple large builds at once, and it's a rather old server. | 07:06:31 |
| Jamie | yeah:( | 07:06:49 |
| Jamie | and when you write your software with the intention of building it on the distributed compilation infrastructure of google... :P | 07:08:09 |
 nf | can someone walk me through why most of the builds of ungoogled-chromium have 1s build times, despite the hashes being different each time? | 11:28:58 |
