| 2 Jul 2026 |
 K900 | Nice | 13:42:47 |
 hexa | I only tested on x86_64-linux fwiw | 13:43:06 |
| K900 | So did I, at least for now | 13:43:33 |
| K900 | I need to get all of my machines on latest staging-next but that requires big computer time and that's currently building KDE Gear 26.04.3 | 13:44:14 |
| hexa | yeah, what was left for me were browsers and browser-adjacent things | 13:45:46 |
| hexa | and I wasn't motivated enough to accept those yet | 13:46:05 |
 emily | btw I still have post-x86_64-darwin fixes that I'll try to put up PRs for in the next few days, but not anything that should cause meaningful numbers of rebuilds (or maybe any rebuilds at all in most cases) | 13:46:52 |
| K900 | webkitgtk and qtwebengine are cached and I'm on firefox nightly binary | 13:46:55 |
| K900 | So uh | 13:46:57 |
| K900 | No browsers | 13:47:02 |
| emily | mostly update script fixes, don't need to block the cycle really, but would suck to have them all break on merge | 13:47:11 |
 Alyssa Ross | I comment out firefox when I build staging-next. I don't have the RAM. | 13:47:29 |
| emily | some of it is an ast-grep treewide that I don't want to do before we're about to merge because the conflicts would be hell | 13:47:29 |
| emily | https://hydra.nixos.org/jobset/nixpkgs/staging-next/evals hm, what did we lose 20k builds to? | 13:48:28 |
| emily | oh, I guess they were failures | 13:48:34 |
| hexa | no, mostly changes pushed onto staging-next since the previous eval | 13:53:38 |
| hexa | like the pillow security update | 13:53:42 |
| hexa | probably the heaviest package among them | 13:53:56 |
| hexa | aarch64-darwin still looks pretty bleak | 13:54:15 |
| hexa | so I wasn't in a hurry | 13:54:20 |
| hexa | we are also basically ouf of work for the darwin machines with ~5k left queued | 13:54:45 |
| hexa | whereas linux has 33k and 43k | 13:55:03 |
| hexa | e.g. ffmpeg-headless and MoltenVK | 13:56:16 |
| K900 | There's a PR with bulk lld fixes somewhere | 13:56:47 |
 whispers [& it/fae] | https://github.com/NixOS/nixpkgs/pull/537712 has moltenvk | 13:57:46 |
| emily | oh didn't see that there was another one | 13:58:05 |
| emily | urgh I am so sorry about how much this is hitting | 13:58:13 |
| emily | the temporary in-tree workarounds aren't too bad, but I have a fear that tons of people's downstream builds are going to break | 13:58:31 |
| emily | Myria: re the wrapped Swift, does that mean that Swift itself fails to link programs, not just Swift itself fails to link? | 13:59:46 |
| emily | maybe we should just be introducing an ld64-fixed and using that where necessary :/ | 14:00:04 |
