apple-sdk: only rewrite old SDK paths

[reckenrode]

Rewriting SDK paths is only needed when a text-based stub is an umbrella framework that does not include the contents of the stubs for the libraries re-exported by the framework. This is only the case for older SDKs. Once macOS introduced the dyld cache (and removed system dylibs) in 11.0, all umbrella frameworks in the SDK include the contents of their re-exported stubs.

While rewriting newer SDKs is mostly harmless, it breaks Zig, which tries to interpret the paths relative to the SDKROOT.

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandboxing enabled in nix.conf? (See Nix manual)
  • sandbox = relaxed
  • sandbox = true
• Tested, as applicable:
  • NixOS test(s) (look inside nixos/tests)
  • and/or package tests
  • or, for functions and "core" functionality, tests in lib/tests or pkgs/test
  • made sure NixOS tests are linked to the relevant packages
• Tested compilation of all packages that depend on this change using nix-shell -p nixpkgs-review --run "nixpkgs-review rev HEAD". Note: all changes have to be committed, also see nixpkgs-review usage
• Tested basic functionality of all binary files (usually in ./result/bin/)
• 24.11 Release Notes (or backporting 23.11 and 24.05 Release notes)
  • (Package updates) Added a release notes entry if the change is major or breaking
  • (Module updates) Added a release notes entry if the change is significant
  • (Module addition) Added a release notes entry if adding a new NixOS module
• Fits CONTRIBUTING.md.

---

Add a 👍 reaction to pull requests you find important.

[nixos-discourse]

This pull request has been mentioned on NixOS Discourse. There might be relevant details there:

https://discourse.nixos.org/t/the-darwin-sdks-have-been-updated/55295/26

[adamcstephens]

Successfully built zig using this on aarch64-darwin.

[cpick]

It looks like zig.x86_64-darwin has been removed from the latest staging-next eval jobset. I'm still learning how jobs are built on the staging -> staging-next -> trunk pipeline so I may be misinterpreting, but I wanted to make sure this removal was intentional?

[reckenrode]

This change isn’t in staging-next yet. I assume Zig will be built in the last (before 24.11) cycle coming up.

[cpick]

Thank you for your reply @reckenrode (here and elsewhere), it's helping me learn how the system works!
I based my assumption off of:

$ git branch --remote --contains e166c9bb94431551b3ac4c0674261144171e5775
  origin/emily/push-nkyqpvqvxwzn
  origin/llvm-19
  origin/staging
  origin/staging-next

And the fact that #354201 was created a few days after this change was merged into staging?

I have a hard time navigating Hydra, but I think the latest evaluation of nixpkgs:staging-next is being built off of https://github.com/NixOS/nixpkgs.git revision 127a31c
which contained this change:

$ git merge-base --is-ancestor e166c9bb94431551b3ac4c0674261144171e5775 127a31c3cf6e9ed385392357d72a65bbd5afe8b0 ; echo $?
0

And I think that's the evaluation that's reporting that zig.x86_64-darwin has been removed from?

[reckenrode]

Oh, sorry. I hadn’t realized they started the last cycle. I thought they were still on the last 24.05 cycle. 😅

I don’t know why it wouldn’t be part of the evaluation. It seems like it should be. I asked in #staging:nixos.org to see if someone there knows.

[reckenrode]

Thanks to @lilyinstarlight for the pointer: it was removed because of a Python eval error on staging-next. Once that’s fixed, Zig and a bunch of other stuff that got dropped should be part of the eval after that on staging-next.

[emilazy]

Should be fixed now. Sorry about that.