mark broken package broken

[Et7f3]

Found while testing #225044

They are broken on master. So I just add metadata to avoid wasting CPU for other users (PyTorch fail near the end of the build)

Description of changes

Things done

• Built on platform(s)
  • x86_64-linux
  • aarch64-linux
  • x86_64-darwin
  • aarch64-darwin
• For non-Linux: Is sandbox = true set in nix.conf? (See Nix manual)
• 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/)
• 23.05 Release Notes (or backporting 22.11 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.

[SomeoneSerge]

pytorch

I don't have access to Darwin to test, but could it be related to merging #222273?

EDIT: logs at https://hydra.nixos.org/build/215378730/nixlog/1

[Et7f3]

Will test if reverting this PR help. Thanks for the pointer.

[Et7f3]

$ git checkout origin/master
$ git revert -m 1 c5b01a1
$ nix build .#python3Packages.pytorch

and it build.

[SomeoneSerge]

I presume that the aligned_alloc in question is this (one of these) function(s) wrapped in a bunch of ifdefs: https://github.com/mreineck/pocketfft/blame/076cb3d2536b7c5d0629093ad886e10ac05f3623/pocketfft_hdronly.h#L156

Updating pytorch -> updating submodules -> maybe some ifdefs have changed? E.g. this is from 2 months ago: mreineck/pocketfft@5fb931a. Although it always had the mingw32 term, it seems. Also cf the related change from October: mreineck/pocketfft@764893f. Also doesn't explain why aarch64 builds fine

[SomeoneSerge]

I'm skeptical but maybe we can try

patches = [
  ...
  (fetchpatch { revert = true; url = "https://github.com/mreineck/pocketfft/commit/764893f3f232d778608025d1256a1c814ce0139e.patch"; hash = ""; })
  (fetchpatch { revert = true; url = "https://github.com/mreineck/pocketfft/commit/5fb931a048d4ed6e2be304bbfb608f655f3da6bf.patch"; hash = ""; })
];

[RaitoBezarius]

Please can we ping the maintainers of those packages so they can chime in and potentially fix the packages?

[Et7f3]

I presume that the aligned_alloc in question is this (one of these) function(s) wrapped in a bunch of ifdefs: https://github.com/mreineck/pocketfft/blame/076cb3d2536b7c5d0629093ad886e10ac05f3623/pocketfft_hdronly.h#L156

Updating pytorch -> updating submodules -> maybe some ifdefs have changed? E.g. this is from 2 months ago: mreineck/pocketfft@5fb931a. Although it always had the mingw32 term, it seems. Also cf the related change from October: mreineck/pocketfft@764893f. Also doesn't explain why aarch64 builds fine

This issue seems to be specific to nixpkgs

I see a fix

nixpkgs/pkgs/applications/networking/browsers/ladybird/default.nix

Lines 54 to 58 in a6c2a73

	] ++ lib.optionals (stdenv.isDarwin && lib.versionOlder stdenv.targetPlatform.darwinSdkVersion "11.0") [
	# error: use of undeclared identifier 'aligned_alloc'
	"-include mm_malloc.h"
	"-Daligned_alloc=_mm_malloc"
	]);

Or another is to force taking the false branch that add a fallback.

[Et7f3]

I put in draft because I think I will remove the one I opened PR to fix and will do another round with the help of hydra to spot more failure.

[SuperSandro2000]

missing comment

[wegank]

Fixed in #226657.