| 23 Oct 2025 |
 Randy Eckenrode | Yes. | 00:39:49 |
| Randy Eckenrode | https://github.com/apple-oss-distributions/distribution-macOS/tree/macos-260 | 00:40:00 |
 emily | feature freeze is in a couple days, maybe those should be for 26.05 😅 | 00:40:01 |
| emily | forgot to look at the bump today, sorry | 00:40:22 |
| emily | hopefully tomorrow | 00:40:24 |
| Randy Eckenrode | But text_cmds has an updated sed! 🤪 | 00:43:19 |
| Randy Eckenrode | More importantly. ld64 updates. | 00:43:46 |
| Randy Eckenrode | We might need the ld64 update for the 26.0 SDK. | 00:44:37 |
| Randy Eckenrode | https://github.com/apple-oss-distributions/ld64/blob/1a4389663d65d6630e4b3e31ace2a86b6183b452/src/ld/Options.cpp#L4888 | 00:45:01 |
| Randy Eckenrode | Or at least cherry-pick that patch until I can do it. | 00:45:41 |
 Ihar Hrachyshka | ok re hardening flag, if no one gets to fix it in a day or two, I may post a patch on weekend. I suspect I will make mistakes though. :) so feel free to beat me to it. | 00:45:57 |
 samasaur | if the 26.0 source releases are out I can try and include that in my apple-sdk_26 PR | 00:50:28 |
| samasaur | ...the one i still haven't opened | 00:50:34 |
| Randy Eckenrode | I was complaining about the source release updates in pkgs/os-specific/darwin/apple-source-releases. Those updates can be non-trivial. The SDK is just providing the sources. | 00:59:23 |
|  reno joined the room. | 01:00:51 |
| Randy Eckenrode | nix-repl> llvmPackages_21.clang
«derivation /nix/store/45l6x4prnm0fj9snbcaaknh5wbd5kxzj-clang-wrapper-21.1.2.drv»
nix-repl> (llvmPackages_21.override { }).clang
«derivation /nix/store/axdpznj47124az3zjf788fr3lsw7853m-clang-wrapper-21.1.2.drv»
| 01:06:52 |
| Randy Eckenrode | That doesn’t seem right. | 01:06:55 |
| samasaur | ah, gotcha | 01:07:50 |
| Randy Eckenrode | Ah, https://github.com/NixOS/nixpkgs/pull/445668#issuecomment-3326150125 | 01:09:51 |
| Randy Eckenrode | nix run github:NixOS/nixpkgs#nix-diff $(nix derivation show github:nixos/nixpkgs#legacyPackages.aarch64-darwin.php84 | jq -r 'keys[0]') $(nix derivation show .#legacyPackages.aarch64-darwin.php84 | jq -r 'keys[0]')
warning: Git tree '/Users/reckenrode/Developer/nixpkgs' is dirty
$ nix run github:NixOS/nixpkgs#nix-diff $(nix derivation show github:nixos/nixpkgs#legacyPackages.aarch64-darwin.casadi | jq -r 'keys[0]') $(nix derivation show .#legacyPackages.aarch64-darwin.casadi | jq -r 'keys[0]')
warning: Git tree '/Users/reckenrode/Developer/nixpkgs' is dirty
| 01:50:23 |
| Randy Eckenrode | No differences. I had to use an ugly hack in the last stage. | 01:50:37 |
| Randy Eckenrode | The “no-op” stage is doing things. | 01:50:49 |
| samasaur | that's always reassuring | 01:50:56 |
| Randy Eckenrode | https://github.com/NixOS/nixpkgs/pull/445668/files#r2453728239 | 01:54:07 |
| emily | thanks, I hate it | 02:04:35 |
| Randy Eckenrode | If it can get rebuilds down to zero, and that allows us to have a working llvmPackages scope, I’ll live with it until it gets corrected. | 02:06:19 |
| samasaur | https://github.com/NixOS/nixpkgs/pull/454747 | 02:12:33 |
| emily | I am enjoying the continuation of the tradition of rambling in the SDK PRs :D | 02:14:58 |
| emily | we should do it as a relay, every year someone has to add their first SDK version | 02:15:25 |
| samasaur | it looked like so much fun when you did it :) | 02:16:52 |
