| 5 Aug 2022 |
 K900 | Or any other hypervisor for that matter | 14:53:27 |
 grahamc (he/him) | qemu is pretty instrumentable, I'm sure it could be done, but I imagine the cost/benefit is pretty skewed | 14:53:58 |
| grahamc (he/him) | especially when the risk is extremely low ("oh dear, keyboard is busted" rollback) | 14:54:19 |
 Vladimír Čunát | As the tests are so many, the error rate gets amplified a lot, so cutting down the list might have the best improvement/effort ratio. | 14:59:08 |
| grahamc (he/him) | yeah | 14:59:22 |
 Winter (she/her) | hey folks, does anyone know why https://hydra.nixos.org/build/186505922 is scheduled to be built on Linux when this is in a Darwin jobset? | 18:37:29 |
| Winter (she/her) | it seems that it's always built on Linux for whatever reason... nothing stands out in release.nix | 18:48:41 |
| Winter (she/her) | nothing stands out in the jobset configuration 😕 | 18:49:56 |
| Winter (she/her) | or manual/default.nix for that matter | 18:51:18 |
| Vladimír Čunát | These attributes are shared for all platforms. | 18:54:37 |
| Winter (she/her) | where's that defined? nothing stands out that would make it different than any other package... | 18:55:07 |
| Vladimír Čunát | It's the normal packages that get defined in a different way. | 18:56:59 |
| Vladimír Čunát | And you couldn't completely fix it anyway. There are almost always multiple platforms in the single eval. How would you prefer to choose which is the "default"? | 18:57:48 |
| Winter (she/her) | Ideally the manual would be built across all platforms, because there is a package used there that isn't built otherwise | 18:58:26 |
| Vladimír Čunát | I think through importing this file, this is the line: https://github.com/NixOS/nixpkgs/blob/master/pkgs/top-level/release-lib.nix#L14 | 18:58:29 |
| Winter (she/her) | In reply to @winterqt:nixos.dev
Ideally the manual would be built across all platforms, because there is a package used there that isn't built otherwise So that would require some compilation that isn't done on x86_64-linux. | 18:58:54 |
| Winter (she/her) | s/isn't/aren't | 19:00:01 |
| Vladimír Čunát | That seems fishy in itself, but I think for manual and perhaps some others it makes sense to switch their build to every platform. | 19:04:19 |
| Vladimír Čunát | And if they were channel blockers, depend on all the platform variants. | 19:04:33 |
| Winter (she/her) | In reply to @vcunat:matrix.org
And if they were channel blockers, depend on all the platform variants. what do you mean by this? block if any of the doc builds fail? | 19:05:08 |
| Vladimír Čunát | Yes. | 19:05:27 |
| Vladimír Čunát | Probably the same with *lib-tests, too. | 19:05:53 |
| Vladimír Čunát | At a quick glance all of that sounds like we want to validate it on all platforms. | 19:06:44 |
| Vladimír Čunát | (even if the multiple results of e.g. the manual most likely won't be useful) | 19:07:37 |
| Winter (she/her) | yeah, i think that would be ideal | 19:10:48 |
| Winter (she/her) | Redacted or Malformed Event | 19:19:07 |
 Pol | Can we restart that build please? https://hydra.nixos.org/build/186499521 | 20:22:13 |
| Pol | (still the same) | 20:22:16 |
| 8 Aug 2022 |
| K900 | Can someone take a look at https://hydra.nixos.org/build/186672095/nixlog/30? | 04:12:15 |
| K900 | Looks very stuck | 04:12:22 |
