| 2 Jun 2026 |
 Sarah Clark | Needs:
mkdir -p $out/${python.sitePackages}/objc_classes
mv objc_classes/test $out/${python.sitePackages}/objc_classes
| 18:47:45 |
| Sarah Clark | * Needs $out:
mkdir -p $out/${python.sitePackages}/objc_classes
mv objc_classes/test $out/${python.sitePackages}/objc_classes
| 18:47:56 |
| Sarah Clark | Yeah, I looked for the casing differences first. aarch64-linux feels like a smoking gun. | 18:54:39 |
 Pol | Pushed again, thanks ! | 19:08:09 |
| Pol | Argh, saw your comment too late, I just disabled the failing test | 19:44:59 |
| Pol | Everything's green now ! :) | 20:03:19 |
| Pol | Thank you !!!! | 20:03:22 |
| 3 Jun 2026 |
 Randy Eckenrode | Saw that and left a series of replies. I didn’t know it was someone at Apple. | 00:14:43 |
| Randy Eckenrode | Some of it’s Swift. Some of it (particularly on Linux) is Nixpkgs. Swift really wants a sysroot. Darwin is close enough to mostly work, but Linux isn’t even close. | 00:15:29 |
| Randy Eckenrode | Also, the changing bootstrap requirements have required me to start over a few times. | 00:16:28 |
| Randy Eckenrode | This may sound like a dumb question, but are x86_64-linux and aarch64-linux using the same filesystem? | 00:17:15 |
| Randy Eckenrode | I’ve been bitten by ZFS doing case normalization by default. I had to create a volume with it disabled and mount it at /nix/var/nix/b for builds. | 00:18:08 |
 hexa | zfs doesn't do normalization unless explicitly set up. | 00:19:52 |
| hexa | Redacted or Malformed Event | 00:19:58 |
| Randy Eckenrode | * I’ve been bitten by ZFS doing case Unicode normalization by default. I had to create a volume with it disabled and mount it at /nix/var/nix/b for builds. | 00:20:43 |
| Randy Eckenrode | Meant Unicode normalization. Is that what you mean? | 00:21:16 |
| Randy Eckenrode | When I created my pools, they have it set to formD even though I didn’t do that myself. | 00:21:26 |
| Randy Eckenrode | At least I don’t think I did. 🤔 | 00:23:11 |
| Randy Eckenrode | Oh, maybe I did. I recall following OpenZFS’s NixOS guide, which does have you create the pool with formD normalization. | 00:24:37 |
| Randy Eckenrode | https://github.com/openzfs/openzfs-docs/blob/06e21fd1e7f379ee8139c4624e311d82ac25cc40/docs/Getting%20Started/NixOS/Root%20on%20ZFS.rst | 00:24:45 |
 K900 | @Randy Eckenrode I think the Mesa horkage starts around https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/41359 | 11:09:11 |
| K900 | Which is interesting | 11:09:16 |
| K900 | Because apparently zink+kk stack works? | 11:09:24 |
| Randy Eckenrode | I knew there was a hacky attempt to use Zink with KosmicKrisp. They’ve implemented a bunch of extensions recently. It works now? 😮 | 11:18:53 |
| Randy Eckenrode | * | 11:19:19 |
| Randy Eckenrode | * | 11:19:31 |
| Randy Eckenrode | I’m still waiting on the tessellation MR before trying it again with FFXIV. | 11:24:35 |
| Sarah Clark | Excellent question. The aarch64-linux failures are happening on Hydra; I don't know which filesystem it's using. | 17:23:43 |
| Randy Eckenrode | If it’s Hydra, I’d assume the builders are set up similarly between architectures and sensibly. Otherwise, they would fail building Git. | 17:24:59 |
| Sarah Clark | Just so. My next step is to compile a list of affected packages to look for similarities. I may also run a bisection. | 17:27:59 |
