16 Oct 2024 |
Alyssa Ross | gotta do something while i wait for the other regressions to bisect | 18:18:24 |
K900 | Once again I am betrayed by my overly high level of faith in humanity | 18:18:52 |
emily | ok, time to look at wezterm | 18:27:33 |
emily | emily@yuyuko ~> NIXPKGS_ALLOW_UNSUPPORTED_SYSTEM=1 nix why-depends --impure --derivation nixpkgs/staging-next#legacyPackages.x86_64-linux.{wezterm,apple-sdk}
'flake:nixpkgs/staging-next#legacyPackages.x86_64-linux.wezterm' does not depend on 'flake:nixpkgs/staging-next#legacyPackages.x86_64-linux.apple-sdk'
| 18:43:28 |
emily | it's evaluating it but not actually using it | 18:43:33 |
K900 | Fun | 18:44:28 |
emily | and yet
… while evaluating attribute 'buildInputs' of derivation 'wezterm-20240203-110809-5046fc22'
at /nix/store/55vkbxyl6snv6m9wqnmx289xzjs7rw7z-source/pkgs/stdenv/generic/make-derivation.nix:383:7:
382| depsHostHost = elemAt (elemAt dependencies 1) 0;
383| buildInputs = elemAt (elemAt dependencies 1) 1;
| ^
384| depsTargetTarget = elemAt (elemAt dependencies 2) 0;
… in the condition of the assert statement
at /nix/store/55vkbxyl6snv6m9wqnmx289xzjs7rw7z-source/lib/customisation.nix:352:23:
351| drvPath = assert condition; drv.${outputName}.drvPath;
352| outPath = assert condition; drv.${outputName}.outPath;
| ^
353| } //
| 18:51:36 |
emily | buildInputs = buildInputs
++ lib.optionals stdenv.hostPlatform.isDarwin [ libiconv ]
++ lib.optionals stdenv.hostPlatform.isMinGW [ windows.pthreads ];
lol I didn't know we had an iconv hack in buildRustPackage
| 18:52:19 |
K900 | wat | 18:53:07 |
emily | obsolete now | 18:53:25 |
emily | I'm kind of lost, unless stdenv.hostPlatform.isDarwin is somehow evaluating apple-sdk , in which case we would have heard about it sooner | 18:53:42 |
K900 | We have time for a Mesa 24.2.5 right | 19:03:48 |
emily | surely | 19:06:21 |
emily | so I'm wondering if Rust's environment capture thing that messed up the previous SDK pattern might be messing this one up too | 19:06:49 |
emily | hm, no, the setEnv stuff looks fine (well, fine except for being bad) | 19:09:06 |
emily | I wish it wasn't impossible to trace eval issues in Nixpkgs | 19:09:19 |
K900 | https://github.com/NixOS/nixpkgs/pull/349114 going in | 19:18:14 |
emily | ok I'm ripping out parts of the wezterm derivation until it stops complaining | 19:26:22 |
emily | oh | 19:26:31 |
emily | you know what it's going to be? | 19:26:34 |
emily | I'm going to bet money it's darwin.apple_sdk_11_0.callPackage | 19:26:43 |
emily | yep | 19:26:55 |
emily | ok | 19:26:58 |
emily | time to dive into the compat layer | 19:27:05 |
emily | yep | 19:27:57 |
emily | ok, fixed it. | 19:31:09 |
linj | I just merged Emacs packages update into staging-next. https://github.com/NixOS/nixpkgs/pull/348670 | 19:31:22 |
emily | K900: here you go https://github.com/NixOS/nixpkgs/pull/349117 | 19:34:51 |
emily | so excited for darwin.apple_sdk_11_0 to die | 19:35:13 |
K900 | Oh fuck my life | 19:35:15 |