| 8 Sep 2021 |
 Rick (Mindavi) | This one: https://github.com/NixOS/hydra/pull/906 | 20:02:14 |
 das_j | feels like someone could just replace the `$(document).ready by something that is triggered when the tree is loaded (→ when the spinner is replaced) | 20:02:51 |
| das_j | * feels like someone could just replace the $(document).ready( by something that is triggered when the tree is loaded (→ when the spinner is replaced) | 20:02:58 |
 tomberek | my approach tries to use HTML summary/details as the mechanism to have toggled build trees | 21:26:09 |
| 9 Sep 2021 |
|  trofi joined the room. | 08:37:54 |
| trofi | Basic hydra UI question: https://hydra.nixos.org/jobset/nixpkgs/binutils-2.36 shows 4 build jobs that are queued for more than 2 weeks. Are they really queued for that long or it's an artifact of some cancellation? | 08:39:42 |
| Rick (Mindavi) | I'd say that they're still queued | 08:52:32 |
| das_j | Yes, cancelling builds removes from the list them iirc | 08:54:01 |
| das_j | * Yes, cancelling builds removes them from the list iirc | 08:54:05 |
| Rick (Mindavi) | Yeah, they get the status 'aborted' | 08:55:28 |
| trofi | Aha. If I look at specific examples at https://hydra.nixos.org/eval/1698592#tabs-unfinished I see that one of the dependencies is timed out: https://hydra.nixos.org/build/150507932#tabs-buildsteps. I would naively expect it to fail the queued job. Or there is a retry somewhere? | 09:53:03 |
| trofi | Best I cloud find is https://discourse.nixos.org/t/finding-most-depended-upon-packages-that-fail-to-build-in-hydra/10090/2 which fetches build logs one by one and parses HTML. Not very light for a casual run. | 21:56:10 |
| trofi | Filed https://github.com/NixOS/hydra/issues/1018 | 22:25:19 |
 sterni | I'd expect that such queries would be much to expensive to be something you can actually expose to users | 22:39:55 |
| sterni | (although you can always rework the database if you really want to) | 22:40:24 |
| 10 Sep 2021 |
 @grahamc:nixos.org | notably none of that strictly needs to query hydra | 01:13:58 |
| @grahamc:nixos.org | almost none | 01:14:06 |
| @grahamc:nixos.org | like, the client itself could do the evaluation and check the cache to identify popular missing store paths from the cache | 01:14:27 |
| trofi | grahamc (he/him): should I close it and file against nix (or maybe nixpkgs)? | 06:26:59 |
| 11 Sep 2021 |
| trofi | Another hydra question: suppose build failed for transient reasons and hydra marks build as broken (running example is llvm_10 in nixpkgs:master that failed 12 days ago: https://hydra.nixos.org/build/151485978). When will hydra try to rebuild it again? Only when derivation changes? Or there is an expiration built in as well? | 17:05:46 |
| Rick (Mindavi) | Only when derivation or one of the inputs change, or someone restart manually (afaik) | 17:13:15 |
 cransom | it won't automatically retry. inputs for the derivation would need to change, or someone can hit the retry button if they have privileges. | 17:29:55 |
| cransom | (as rick said) | 17:30:10 |
| *  lukegb (he/him) presses the retry button | 18:34:14 |
| sterni | In reply to @trofi:matrix.org
Another hydra question: suppose build failed for transient reasons and hydra marks build as broken (running example is llvm_10 in nixpkgs:master that failed 12 days ago: https://hydra.nixos.org/build/151485978). When will hydra try to rebuild it again? Only when derivation changes? Or there is an expiration built in as well? restart should propagate to reverse dependencies (in basically all cases?) imo, would be a nice quality of life feature | 18:41:05 |
| 13 Sep 2021 |
|  Florian | W3F changed their display name from Florian | W3F - OoO to Florian | W3F - OoO Mon/Tue. | 11:56:28 |
| 14 Sep 2021 |
|  moritz.hedtke joined the room. | 17:15:02 |
|  Rev. CornWallace III (novus ordo seclorum) changed their display name from cw (novus ordo seclorum) to Fauci's Fine Wines. | 21:35:04 |
| Rev. CornWallace III (novus ordo seclorum) changed their display name from Fauci's Fine Wines to Corn Wallet. | 21:35:38 |
| 15 Sep 2021 |
|  jbedo joined the room. | 04:00:52 |
