| 13 Jun 2023 |
|  Julien changed their display name from JulienMalka (stop speaking german when) to Julien (stop speaking german when). | 14:53:54 |
|  K900 changed their display name from K900 to K900 (Old). | 20:51:42 |
|  @federicodschonborn:matrix.org changed their profile picture. | 20:55:44 |
| Julien changed their display name from Julien (stop speaking german when) to Julien. | 21:05:28 |
 magic_rb |
how do you people deal with having packages, nixos systems, (nixng systems) and others in one flake? since the hydraJobs attribute is flat, its hard to logically group things with good names, unless i do overlay-X overlay-Y nixos-Z nixng-W | 21:20:26 |
|  K900 (deprecated) joined the room. | 21:46:20 |
| 14 Jun 2023 |
 jackdk | hydraJobs can be a nested attrset, you're just not allowed to have '.' in any attr name | 06:39:31 |
| jackdk | How can I chase down why hydra seems to be holding gcroots for basically everything? We've set each jobset to keep only 2 evaluations but I keep finding really old links in various gcroot dirs, which I would not expect | 06:46:20 |
|  raphi changed their display name from raphi (element unread channel fix when) to raphi. | 07:03:13 |
|  hexa left the room. | 16:17:15 |
| 15 Jun 2023 |
|  @vector:tchncs.de removed their profile picture. | 15:50:57 |
| @vector:tchncs.de removed their display name vector. | 15:52:33 |
| @vector:tchncs.de left the room. | 15:54:17 |
| 16 Jun 2023 |
| jackdk | In reply to @jackdk:matrix.org
How can I chase down why hydra seems to be holding gcroots for basically everything? We've set each jobset to keep only 2 evaluations but I keep finding really old links in various gcroot dirs, which I would not expect I am the architect of my own suffering: the cronjob which ran hydra-update-gc-roots didn't run in the same environment as all the other hydra stuff, so it dutifully maintained the wrong gc roots dir. | 05:08:19 |
 Rick (Mindavi) | I should've asked the question that I was thinking of: is gc running 😅 | 09:56:48 |
| 17 Jun 2023 |
|  @lab-cat:matrix.org left the room. | 12:18:13 |
| 18 Jun 2023 |
|  @autiboy:matrix.org joined the room. | 18:06:52 |
|  toonn changed their profile picture. | 20:48:35 |
| 19 Jun 2023 |
|  Ryan Yin joined the room. | 04:30:12 |
|  @antifuchs:asf.computer joined the room. | 18:12:14 |
| @antifuchs:asf.computer | am I holding the manual at https://hydra.nixos.org/build/219217323/download/1/hydra/ wrong or is it meant to look like a completely unstyled html document? | 18:12:59 |
| @antifuchs:asf.computer | (the styley buttons at the top seem to indicate it's meant to look like something else) | 18:13:19 |
| raphi | 
Download image.png | 18:27:56 |
| raphi | yeah it's supposed to be resonably pretty mdBook
likely culprit: | 18:27:57 |
| raphi | * yeah it's supposed to be resonably pretty mdBook
likely culprit (when inspecting the response headers of the .css files): | 18:28:18 |
| raphi | https://github.com/NixOS/hydra/issues/1278 | 18:44:31 |
| @antifuchs:asf.computer | ah, oh, neat! thanks for digging that up. seems like https://github.com/NixOS/hydra/pull/1268 is intended to fix it? | 18:47:24 |
| 20 Jun 2023 |
| @antifuchs:asf.computer | I think I'm hitting a bootstrapping loop: I have a private repo on github for which I'd like to define a jobset with hydra; I don't quite understand how to authorize it to see the repo on github. Is the github_authorization sufficient to give it access? ... and how does that work with the flake jobset URL type? | 00:14:59 |
| @antifuchs:asf.computer | (or do those require git+ssh & some ssh private key somewhere? | 00:15:10 |
| @antifuchs:asf.computer | * (or do those require git+ssh & some ssh private key somewhere?) | 00:15:15 |
