| 26 Jun 2026 |
 K900 | Oh | 18:45:01 |
| K900 | v4l-utils moved to Qt6 | 18:45:11 |
| K900 | That explains it | 18:45:13 |
 Grimmauld (any/all) | Thanks for the reminder, will take a look. I didn't finish reviewing upstream changes, but i don't see regressions in the tests i have built against it | 18:58:32 |
| Grimmauld (any/all) | I am quite confident we can just send it | 18:58:53 |
 hexa | then by all means | 18:59:16 |
| hexa | send it | 18:59:17 |
| hexa | emily https://hydra.nixos.org/build/333216997 | 19:32:34 |
 emily | uh | 19:33:19 |
| emily | well whatever that is was a recent merge | 19:33:23 |
| emily | libxml2… Grimmauld (any/all)? | 19:33:27 |
| emily | oh wait isn't this just the GC bug | 19:33:33 |
| emily | there's a GC bug right? | 19:33:38 |
| hexa | could be | 19:33:41 |
| hexa | we can retry | 19:33:50 |
| emily | no libxml2 merges in a while | 19:33:55 |
| Grimmauld (any/all) | github.com/NixOS/nixpkgs/pull/527525 was libxml2, and i did built to cmake and a bunch of python things on that | 19:39:08 |
| emily | whatever it was would have had to be in the last day | 19:39:24 |
| emily | so pretty sure it's just the bug | 19:39:28 |
| emily | oh I wonder if I know what the bug is | 19:39:51 |
| emily | I actually ran into a GC bug recently | 19:40:02 |
| emily | is it already diagnosed? | 19:40:12 |
| hexa | I'm building on the community builder in parallel | 19:40:29 |
| emily | I believe structured attributes + remote builders can result in deps being GC'd mid-build | 19:40:31 |
| emily | by it I mean the GC bug | 19:40:36 |
| hexa | https://github.com/NixOS/infra/pull/1104 hm | 19:41:01 |
| emily | ok, not the bug I was thinking of | 19:41:59 |
 Sergei Zimmerman (xokdvium) | Remote builds are generally ugh... very racy yeah | 19:43:35 |
| Sergei Zimmerman (xokdvium) | Some work to be done there certainly | 19:43:49 |
| Sergei Zimmerman (xokdvium) | Specifically https://github.com/NixOS/nix/pull/15719, which I'll return to | 19:44:16 |
