rpm: declare darwin as badPlatform

[katexochen]

Description of changes

Support by rpm for darwin got removed in #196350.

This will be fixed once apple_sdk > 13.0 is available and we can check for that version.

Not sure if I got every package that (transitively) depends on rpm. Is there any good way to check this?

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/)
• 23.11 Release Notes (or backporting 23.05 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.

[reckenrode]

Looks good from a Darwin perspective.

[vcunat]

Can we explain why this is even needed? AFAIK it's acting transitively

$ nix eval -f. --argstr system x86_64-darwin apkg.outPath
error: Package ‘rpm-4.18.1’ in pkgs/tools/package-management/rpm/default.nix:72 is not available on the requested hostPlatform:

[katexochen]

@vcunat Maybe I understood things wrong. I'm trying to fix these evaluation errors thrown by hydra:
https://hydra.nixos.org/jobset/nixpkgs/trunk#tabs-errors

in job ‘libmodulemd.x86_64-darwin’:
error:
       … while calling the 'deepSeq' builtin

         at /nix/store/0555ml133awcl3gyl5v0lh93hbp3nngh-source/lib/customisation.nix:288:7:

          287|     in if drv == null then null else
          288|       lib.deepSeq drv' drv';
             |       ^
          289|

       … while evaluating the attribute 'outPath'

         at /nix/store/0555ml133awcl3gyl5v0lh93hbp3nngh-source/lib/customisation.nix:277:13:

          276|           value = commonAttrs // {
          277|             outPath = output.outPath;
             |             ^
          278|             drvPath = output.drvPath;

       (stack trace truncated; use '--show-trace' to show the full trace)

       error: Failed to evaluate rpm-4.18.1: «unsupported»: is not available on the requested hostPlatform:
         hostPlatform.config = "x86_64-apple-darwin"
         package.meta.platforms = [
           "aarch64-linux"
           "armv5tel-linux"
           "armv6l-linux"
           "armv7a-linux"
           "armv7l-linux"
           "i686-linux"
           "loongarch64-linux"
           "m68k-linux"
           "microblaze-linux"
           "microblazeel-linux"
           "mips-linux"
           "mips64-linux"
           "mips64el-linux"
           "mipsel-linux"
           "powerpc64-linux"
           "powerpc64le-linux"
           "riscv32-linux"
           "riscv64-linux"
           "s390-linux"
           "s390x-linux"
           "x86_64-linux"
         ]
         package.meta.badPlatforms = [ ]

[vcunat]

Right, the amount of eval warnings is a long-standing problem. I don't know off the top of my head what's the best approach here. Maybe this PR really is the best one.

[Artturin]

This is incorrect, the proper way seems to be setting badPlatforms in rpm to make it and it's dependants not show up on hydra https://ryantm.github.io/nixpkgs/stdenv/meta/#var-meta-badPlatforms

[katexochen]

This is incorrect, the proper way seems to be setting badPlatforms in rpm to make it and it's dependants not show up on hydra https://ryantm.github.io/nixpkgs/stdenv/meta/#var-meta-badPlatforms

Thanks @Artturin, this looks like the thing I want to do. Do I also need to add darwin back under platforms when setting badPlatforms = platforms.darwin; ?

[Artturin]

There should be a comment about when darwin can be readded

Explained

[github-actions]

Successfully created backport PR for release-23.11:

• [Backport release-23.11] rpm: declare darwin as badPlatform #269711

[Artturin]

Let's see if it works but we should not do this for each individual package which isn't supported on some platform. Instead, the root issue should be fixed.

[katexochen]

Let's see if it works but we should not do this for each individual package which isn't supported on some platform. Instead, the root issue should be fixed.

@Artturin So what is/was the root cause in your opinion? And how could a more general solution look like?

[katexochen]

Seems like the fix didn't work, the evaluation errors are still there: https://hydra.nixos.org/eval/1802300#tabs-errors

[rb2k]

It's been a while and I'm vaguely sorry to necromance this PR, but I am just doing some digging into what it would take to get microdnf to compile again on darwin and this is probably one of the best places for people to find this information :)

Upstream RPM can be built on darwin again starting with 4.19.1.1:
rpm-software-management/rpm#2807

So if this gets upgraded, this might be build-able again. Maybe I can convince some of my colleagues to take a look.

[reckenrode]

It’s possible, but it may also still require a newer SDK than is available currently in nixpkgs. I’m going to be working on a refactor and update for 24.11, but that’s not ready (or started) yet.