| 21 Oct 2024 |
 vcunat | This merge: 3fc30386251852085c2eab0731a6439f13140036 | 13:11:18 |
| vcunat | Some change in it didn't cause stdenv rebuild on master but it does here. | 13:11:42 |
 K900 | Yaaaay | 13:16:06 |
| vcunat | bf2137762146 bmake: 20240808 -> 20240921
| 13:16:34 |
| vcunat | So I guess I'll revert it on staging-next. | 13:16:51 |
 emily | yes we now depend on bmake to build locale data | 13:37:08 |
 Alyssa Ross | maybe would be better to revert on master, so there's less chance of people starting to rely on the update? | 13:37:12 |
| K900 | Agreed | 13:37:23 |
| emily | I guess we can reapply on staging | 13:37:48 |
| Alyssa Ross | Yeah | 13:37:51 |
| vcunat | Probably. Though I thought we might avoid stdenv rebuilds in some upcoming iteration. | 13:50:24 |
| vcunat | * Probably. Though I thought we might avoid stdenv rebuilds in some upcoming staging-next iteration. | 13:50:34 |
| Alyssa Ross | not sure I understand | 13:51:24 |
| vcunat | I thought we might have a period where staging wouldn't get any stdenv update. But maybe I was naive. | 13:54:47 |
| emily | I'm pretty sure we already have a Darwin rebuild in the pipeline for the next staging | 13:54:56 |
| emily | https://github.com/NixOS/nixpkgs/pull/349555 would be another. | 13:55:24 |
| emily | (which I'd really like to get in because otherwise every use of Rust bindgen for macOS stuff is going to need workarounds.) | 13:55:49 |
| vcunat | I see. OK. | 13:56:28 |
| emily | sorry to ruin your dreams 😅 | 13:57:03 |
| vcunat | Oh, that one's unmerged. But before you posted I saw that the darwin stdenv changed hash already in staging-next..staging anyway. So I reapplied the bmake update on staging. | 14:04:54 |
|  svrana joined the room. | 17:22:05 |
|  veleiro joined the room. | 20:01:13 |
 Randy Eckenrode | rcodesign failed to build on Darwin on staging-next. I finally have the fix. | 23:21:20 |
| Randy Eckenrode | emily: https://github.com/NixOS/nixpkgs/blob/master/pkgs/by-name/ld/ld64/0008-Disable-searching-in-standard-library-locations.patch | 23:46:05 |
| Randy Eckenrode | Try also disabling that patch along with the clang change. | 23:46:15 |
| Randy Eckenrode | The Swift stuff may in fact be my fault. 🫠| 23:46:29 |
| Randy Eckenrode | https://github.com/apple-oss-distributions/ld64/blob/47f477cb721755419018f7530038b272e9d0cdea/src/ld/Options.cpp#L4313-L4319 | 23:47:21 |
| Randy Eckenrode | You can overlay sysroots with ld64. Do other linkers work that way? | 23:47:49 |
| Randy Eckenrode | In reply to @emilazy:matrix.org
(which I'd really like to get in because otherwise every use of Rust bindgen for macOS stuff is going to need workarounds.) See also my comment in #macos:nixos.org. I may have broken this on the ld64 side. I wonder if it works using lld. | 23:48:37 |
| Randy Eckenrode | https://github.com/NixOS/nixpkgs/pull/350374 for rcodesign. | 23:55:17 |
