| 13 May 2026 |
 emily | I would really prefer to land it once it won't break everything | 16:55:19 |
 Vladimír Čunát | Hmm, it's only easy for the small jobset which only pulls stdenvs, yeah. | 16:55:55 |
 hexa | and I would prefer wasting less compute cycles for when we have to support three releases | 16:55:58 |
| emily | I can get those changes ready on a few days' notice but I am just trying to manage my time. I do not want to land something that will break the world before I have the fixes pushed out | 16:56:03 |
| emily | like I said, I can have it ready for the upcoming -next | 16:56:18 |
| hexa | ok | 16:56:27 |
| emily | overriding it on Hydra for the staging jobset specifically for now seems like the right fix until then? | 16:56:32 |
| emily | nothing else will be wasting cycles on x86_64-darwin | 16:56:48 |
| hexa | https://hydra.nixos.org/jobset/nixpkgs/staging#tabs-configuration done | 16:57:40 |
| Vladimír Čunát | I suppose. But you know... on nixpkgs side we deduplicated that list and then we override it on a remote place per-jobset... | 16:58:17 |
| Vladimír Čunát | but that's of little practical significance anyway. | 16:58:48 |
| hexa | the per jobset thing is a stopgap for now | 16:58:51 |
| hexa | I think that's a fair compromise | 16:59:16 |
| Vladimír Čunát | Yes, just easily forgotten. | 16:59:38 |
| Vladimír Čunát | But the list is very static anyway. | 16:59:53 |
| hexa | doesn't matter for staging, does it? | 16:59:56 |
| emily | I can remind about the override if I'm reminded about the -next :) | 17:00:11 |
| emily | but yeah, this stuff changes very little | 17:00:21 |
| emily | as I discovered by how much needs fixing if you remove anything | 17:00:26 |
| emily | it will require a bit of a dance with nixpkgs-ci | 17:00:37 |
| emily | as you need to merge changes to make CI not block on the platform and then merge changes dropping it etc. | 17:00:50 |
| emily | there's like a 3-PR sequence that has to land in order | 17:01:02 |
 K900 | Is it too late for a security-adjacent Pipewire update | 20:39:00 |
|  Adam joined the room. | 22:44:10 |
| 14 May 2026 |
|  ryand56 joined the room. | 05:08:56 |
| K900 | OK I'm just going to send it | 10:39:53 |
| K900 | And libpipewire treewide is on my list for after branchoff | 10:40:33 |
|  tuhana joined the room. | 14:34:37 |
 André Lima | So is the fix here to just remove the chromium patch from https://github.com/dtolnay/rustversion/blob/a1dfebcd35f6082886548156f02f3832f52c0853/src/bound.rs#L51-L55 and change the selectors that bytemuck uses from rustversion::before(2026-01-27) to rustversion::any(before(2026-01-27), before(1.95)) and rustversion::since(2026-01-27) to rustversion::any(since(2026-01-27), since(1.95))? | 15:08:33 |
| André Lima | * So is the fix here to just remove the chromium patch from the chromium derivation and change the selectors that bytemuck uses from rustversion::before(2026-01-27) to rustversion::any(before(2026-01-27), before(1.95)) and rustversion::since(2026-01-27) to rustversion::any(since(2026-01-27), since(1.95))? | 15:09:49 |
