| 28 Jun 2026 |
 K900 | OK I mad a slack account for this | 12:05:43 |
| K900 | I hope y'all respect the sacrifices I'm making | 12:05:55 |
 hexa | didn't we just move off of pnpm9 and now 10 is marked vulnerable? | 12:06:40 |
| hexa | they grow up so fast | 12:06:47 |
| hexa | https://github.com/NixOS/nixpkgs/issues/529285 | 12:06:55 |
| K900 | /shrug | 12:06:56 |
| K900 | Anyway here's what I have so far | 12:23:11 |
| K900 | https://github.com/NixOS/nixpkgs/pull/536281 | 12:23:11 |
| K900 | Still need to sort out pandas and lix | 12:24:13 |
| K900 | And probably bang my head against valkey until it builds | 12:24:24 |
| hexa | needs fmt | 12:34:20 |
| K900 | Will do once I'm out of fixes | 12:34:33 |
 Grimmauld (any/all) | huh the gst-vaapi thing is funny, it still built against gst 1.28.1 up to .3, and now it is broken? I guess i shouldn't be surprised, sorry that slipped through, that one is on me | 12:49:03 |
| K900 | Well upstream 1.28 changelog says "you can try but we don't promise anything lul" | 12:52:15 |
 Randy Eckenrode | I suspect https://github.com/NixOS/nixpkgs/commit/23c05205023464c322a0df9ac0834c00f4da9fa8 broke ld64 on aarch64-darwin. | 13:28:29 |
| Randy Eckenrode | It usually works, but it fails to link glib, and it’s failing to link something in stage 0 of the Swift bootstrap in my Swift work. | 13:29:38 |
| Randy Eckenrode | * | 13:29:45 |
| Randy Eckenrode | I’ve reverted locally to see if it fixes the issue, but it’s going to take a few hours to re-bootstrap and build through glib. | 13:34:42 |
| Randy Eckenrode | * | 13:34:48 |
| Randy Eckenrode | (Mostly Glib since Swift hasn’t landing. The Swift bootstrap is also a bit lengthy.) | 13:35:28 |
 emily | oh no :/ | 13:41:18 |
| emily | maybe we can hack around it by building a variant w/o hardening just for glib | 13:41:39 |
| emily | and fixing up on staging | 13:41:45 |
| Randy Eckenrode | Yeah. I probably wrote that comment too suggesting it was x86_64-darwin only. I don’t think I ever experienced the problem on aarch64-darwin, but there it is. | 13:44:09 |
| Randy Eckenrode | If it’s only Glib, using LLD also works. | 13:44:21 |
| Randy Eckenrode | * | 13:45:01 |
| emily | I violated the golden rule: never trust comments, even if you wrote them yourself | 13:46:06 |
| emily | I expect it's not the only thing, but maybe it's rare enough that we can hold on for a cycle. | 13:47:05 |
| emily | using LLD instead sounds good to me | 13:47:12 |
 PixelHamster | Is the best way to check the staging->staging-next merges via the git log or is there a website that lists them? | 13:50:14 |
