nixpkgs: backports from upstream #1318

emilazy · 2025-02-03T19:46:22Z

Should fix the module after recent Nixpkgs changes.

cixel · 2025-02-03T19:59:42Z

This might be redundant with the logs in the failing checks, but when I update my flake input I now get the following:

building the system configuration...
error:
       … while evaluating the attribute 'config.system.build.toplevel'
         at /nix/store/32dh0g41x5w9zkqxqka3rj7h18blqhyh-source/lib/modules.nix:334:9:
          333|         options = checked options;
          334|         config = checked (removeAttrs config [ "_module" ]);
             |         ^
          335|         _module = checked (config._module);

       … while calling the 'seq' builtin
         at /nix/store/32dh0g41x5w9zkqxqka3rj7h18blqhyh-source/lib/modules.nix:334:18:
          333|         options = checked options;
          334|         config = checked (removeAttrs config [ "_module" ]);
             |                  ^
          335|         _module = checked (config._module);

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

       error: undefined variable 'darwinExpectedSystem'
       at /nix/store/f8vkfbnn50yf02ijxcf9rapw70bziv39-source/modules/nix/nixpkgs.nix:288:177:
          287|           assertion = cfg.constructedByUs -> !hasPlatform -> cfg.system == pkgsSystem;
          288|           message = "The nix-darwin nixpkgs.pkgs option was set to a Nixpkgs invocation that compiles to target system ${pkgsSystem} but nix-darwin was configured for system ${darwinExpectedSystem} via nix-darwin option nixpkgs.system. The nix-darwin system settings must match the Nixpkgs target system.";
             |                                                                                                                                                                                 ^
          289|         }

This got mangled in the backport a year and a half ago. Fixes: e25eeff

… externally This is a common footgun people hit often. Remove it. Backport of Nixpkgs commit ce87196a00214a0062ece1c3e03a9a97f563580f. Co-authored-by: K900 <[email protected]>

henrik-ch was also here :) Backport of Nixpkgs commit 11406bdc0e5af9b3c8a8d597da23349238c65277. Co-authored-by: Silvan Mosberger <[email protected]> Co-Authored-By: Valentin Gagarin <[email protected]>

system and config shouldn't both be specified — each will be filled in based on the other when the system is elaborated. Backport of Nixpkgs commit a3ba0495452cd8e72735ebd4472838e96902a259. Co-authored-by: Alyssa Ross <[email protected]>

jwiegley · 2025-02-03T20:34:41Z

Is it necessary for this PR to also reformat the entire file? It makes it very hard to see what the actual change is...

emilazy · 2025-02-03T20:42:44Z

Yes, since Nixpkgs upstream formatted it and it’s necessary to easily backport the changes after that point. We should do a treewide reformat but I’m holding out until Nixpkgs does the full treewide. I will add the reformatting commit to .git-blame-ignore-revs once I’ve gotten this actually working and no longer need to rebase it, though.

All the changes here are just backported from NixOS. If you want to review the backports I’d suggest going commit‐by‐commit, since otherwise you’re reading like ten upstream PRs at once.

Since the output of `lib.systems.elaborate` contains functions, an equality check with `==` does not suffice, `lib.systems.equals` should be used instead. Backport of Nixpkgs commit 3794246066409d7baac72e3fdfb0e4f66ef4a013. Co-authored-by: Jared Baur <[email protected]>

Backport of Nixpkgs commit e6057cfd59f278db3aeb058a4e1e0bcc24696267. Co-authored-by: Valentin Gagarin <[email protected]> Co-authored-by: Dominic Mills <[email protected]>

Backport of Nixpkgs commit 609e57485d1fa111e3a689498d9d338dc03a7bc5. Co-authored-by: Felix Buehler <[email protected]>

The assertion message should include the `nixpkgs.config` value, however it currently includes the entire `nixpkgs.config` _option_. This means the type, declarations, definitions, etc were all printed. Backport of Nixpkgs commit 1bd4da1848cb7b68858ebb2ca1f8b0e5fed46c58. Co-authored-by: Matt Sturgeon <[email protected]>

Backport of Nixpkgs commit 6d9dfef94ffd59a327573eea7bc709a84c44b3d2. Co-authored-by: Matt Sturgeon <[email protected]>

The description for options.nixpkgs.system already hints at this: Neither ${opt.system} nor any other option in nixpkgs.* is meant to be read by modules and configurations. Use pkgs.stdenv.hostPlatform instead. We can support this goal by not elaborating the systems anymore, forcing users to go via pkgs.stdenv. This will prevent problems when making the top-level package sets composable in the next commit. For this to work, you should pass a fully elaborated system to nixpkgs' localSystem or crossSystem options. Backport of Nixpkgs commit 0a19371146130c0e2a402fd0c35f8283b0e81910. Co-authored-by: Wolfgang Walther <[email protected]>

ossareh · 2025-02-03T20:54:14Z

Can confirm that as of 537056b I can now rebuild successfully. Thanks @emilazy!

cixel · 2025-02-03T21:02:02Z

Can confirm that as of 537056b I can now rebuild successfully. Thanks @emilazy!

Ditto, but only if I have nixpkgs.linux-builder.enable = false.

If I set it true, I get the following:

error:
       … while evaluating the attribute 'value'
         at /nix/store/32dh0g41x5w9zkqxqka3rj7h18blqhyh-source/lib/modules.nix:846:9:
          845|     in warnDeprecation opt //
          846|       { value = addErrorContext "while evaluating the option `${showOption loc}':" value;
             |         ^
          847|         inherit (res.defsFinal') highestPrio;

       … while evaluating the option `system.build':

       … while evaluating the attribute 'mergedValue'
         at /nix/store/32dh0g41x5w9zkqxqka3rj7h18blqhyh-source/lib/modules.nix:881:5:
          880|     # Type-check the remaining definitions, and merge them. Or throw if no definitions.
          881|     mergedValue =
             |     ^
          882|       if isDefined then

       … while evaluating definitions from `/nix/store/nl6r1g7n6jcn0pwjb3hikszf2lhn0163-source/modules/system':

       … while evaluating the option `nix.buildMachines."[definition 1-entry 1]".systems':

       … while evaluating definitions from `/nix/store/nl6r1g7n6jcn0pwjb3hikszf2lhn0163-source/modules/nix/linux-builder.nix':

       … while evaluating the option `nix.linux-builder.systems':

       … while evaluating definitions from `/nix/store/nl6r1g7n6jcn0pwjb3hikszf2lhn0163-source/modules/nix/linux-builder.nix':

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

       error: expected a set but found a string: "aarch64-linux"

This goes away if I manually set:

  nix.linux-builder.systems = [ pkgs.hostPlatform.system ];

It isn't obvious to me why the default isn't working.

emilazy · 2025-02-03T21:11:09Z

@wolfgangwalther Looks like darwin.linux-builder’s NixOS configuration is broken by your PR? I don’t think that’s something we can work around downstream since we need Hydra to build it (and I’m not sure how we’d stop reading nixpkgs.hostPlatform.system since it doesn’t seem like the package exposes its package set).

emilazy · 2025-02-03T21:33:10Z

Actually I guess we can do darwin.linux-builder.nixosConfig._module.args.pkgs.stdenv.hostPlatform.system. Though the darwin.linux-builder’s way of setting nixpkgs.hostPlatform still seems a bit questionable.

emilazy · 2025-02-03T21:35:19Z

@cixel #1319 should hopefully fix that; I’ve submitted it separately as it can be applied independently, but you’ll probably need to test with both.

cixel · 2025-02-03T21:57:37Z

Thanks for the quick turnaround. I did the following:

locally checked out emilazy:push-mmmkkprkuzzs (branch for linux-builder: don’t read nixpkgs.hostPlatform #1319)
rebased it onto emilazy:push-tmtstvpkwkow
set inputs.darwin.url = /path/to/local/repo
darwin-rebuild switch --flake /path/to/flake

and I get the following:

building the system configuration...
these 17 derivations will be built:
  /nix/store/jdmim32yd0pbha4pf6frac3sqdil3dfa-options.json.drv
  /nix/store/y2q70ndn9l28y9qif5wb7sc40jjmsmww-options.json.drv
  /nix/store/6jhi1k3jfamf4l9j638lqrj25x5f4dpi-darwin-manual-html.drv
  /nix/store/m3446bhiqznii8idnz48g0kcy45wg40c-darwin-help.drv
  /nix/store/wqif2s887hhxrnra8knxpfbad7iyi2jm-darwin-manpages.drv
  /nix/store/91m9i060wm2jg0mf2d4jlqisq1k8xh65-system-applications.drv
  /nix/store/mrhgqvrv90qgs2zry28l5dpcfvnlr9zf-system-path.drv
  /nix/store/md8dp72hxw75251dicgih6jkm9xhlpnr-darwin-system-25.05.drv
  /nix/store/71k581g5bqfyx2icfqny84dgb4gnr5ln-darwin-uninstaller.drv
  /nix/store/bbny1dkiq4sn8was9l4wy286yqzf1gnq-options.json.drv
  /nix/store/iry0n6vaifbqp1lnp6hw385x8c677lm1-options.json.drv
  /nix/store/w2fam0w31ysk34dj7ggshk7g2m1xhgxz-darwin-manual-html.drv
  /nix/store/9cwji11dwgr2qqvn5qlny3g8f867jbis-darwin-help.drv
  /nix/store/bw5ha2gflszpazzh6jc3ccffbjh17acd-darwin-manpages.drv
  /nix/store/6lqrlh9r4rywhmil4m3i5xfdqjyklli1-system-applications.drv
  /nix/store/fkj92jsx127s4vzib7lqq1ngahl5qyyl-system-path.drv
  /nix/store/lgsbvkr7azb999f74wdwa06lqglpkpxs-darwin-system-25.05.c5afeba.drv
building '/nix/store/bbny1dkiq4sn8was9l4wy286yqzf1gnq-options.json.drv' on 'ssh-ng://builder@linux-builder'...
building '/nix/store/jdmim32yd0pbha4pf6frac3sqdil3dfa-options.json.drv' on 'ssh-ng://builder@linux-builder'...
copying 0 paths...
error: build of '/nix/store/bbny1dkiq4sn8was9l4wy286yqzf1gnq-options.json.drv' on 'ssh-ng://builder@linux-builder' failed: error: a 'aarch64-darwin' with features {} is required to build '/nix/store/bbny1dkiq4sn8was9l4wy286yqzf1gnq-options.json.drv', but I am a 'aarch64-linux' with features {benchmark, big-parallel, gccarch-armv8-a, kvm, nixos-test}
error: builder for '/nix/store/bbny1dkiq4sn8was9l4wy286yqzf1gnq-options.json.drv' failed with exit code 1
error: 1 dependencies of derivation '/nix/store/iry0n6vaifbqp1lnp6hw385x8c677lm1-options.json.drv' failed to build
error: 1 dependencies of derivation '/nix/store/bw5ha2gflszpazzh6jc3ccffbjh17acd-darwin-manpages.drv' failed to build
error: 1 dependencies of derivation '/nix/store/w2fam0w31ysk34dj7ggshk7g2m1xhgxz-darwin-manual-html.drv' failed to build
error: 1 dependencies of derivation '/nix/store/6lqrlh9r4rywhmil4m3i5xfdqjyklli1-system-applications.drv' failed to build
error: 1 dependencies of derivation '/nix/store/lgsbvkr7azb999f74wdwa06lqglpkpxs-darwin-system-25.05.c5afeba.drv' failed to build

am I doing something wrong?

emilazy · 2025-02-04T03:21:19Z

Not sure. Seems like the Linux builder is trying to build part of your nix-darwin configuration (I guess?) but I don’t know why.

wolfgangwalther · 2025-02-04T17:52:04Z

Looks like darwin.linux-builder’s NixOS configuration is broken by your PR? I don’t think that’s something we can work around downstream since we need Hydra to build it (and I’m not sure how we’d stop reading nixpkgs.hostPlatform.system since it doesn’t seem like the package exposes its package set).

No, I don't think so. This piece of darwin-packages.nix, should still work, right?

              nixpkgs.hostPlatform = lib.mkDefault (toGuest stdenv.hostPlatform.system);

cixel · 2025-02-04T20:56:28Z

Not sure. Seems like the Linux builder is trying to build part of your nix-darwin configuration (I guess?) but I don’t know why.

Yeah, I think that in messing around with different branches and configs I worked my machine into a state where linux-builder was being told to build stuff it's not supposed to. Could've been my "workaround" above broke things. Unsure.

I reset things to an older state and then tried again. Things seem to work now.

Relevant entries from my flake.lock, in case these are useful:

    "darwin": {
      "inputs": {
        "nixpkgs": [
          "nixpkgs"
        ]
      },
      "locked": {
        "lastModified": 1738619295,
        "narHash": "sha256-E/zV6EsIuVrzCr07077srLAPpSvgEjtX/qJ4Nu4hPps=",
        "ref": "refs/heads/push-mmmkkprkuzzs",
        "rev": "c5afebab4fc5a86c170da96779233c32724ccc10",
        "revCount": 2042,
        "type": "git",
        "url": "file:///Users/ehdens/workspace/github.com/emilazy/nix-darwin"
      },
      "original": {
        "type": "git",
        "url": "file:///Users/ehdens/workspace/github.com/emilazy/nix-darwin"
      }
    },

    "nixpkgs": {
      "locked": {
        "lastModified": 1738678663,
        "narHash": "sha256-uPdqVr8gN3oYYcueL/Rh7wM4Y8/D6p9IvRmoBf+uQa8=",
        "owner": "NixOS",
        "repo": "nixpkgs",
        "rev": "00769b0532199db4e1bda59865f00f3a86232c75",
        "type": "github"
      },
      "original": {
        "owner": "NixOS",
        "ref": "nixpkgs-unstable",
        "repo": "nixpkgs",
        "type": "github"
      }
    },

Enzime

Not the biggest fan of the last commit as it's unclear of it's going to get reverted (see NixOS/nixpkgs#376988 (comment))

However, according to @wolfgangwalther it should still be backwards compatible (NixOS/nixpkgs#376988 (comment)) so I'm going to merge this now

to bring in LnL7/nix-darwin#1318

wolfgangwalther · 2025-02-05T18:26:56Z

However, according to @wolfgangwalther it should still be backwards compatible (NixOS/nixpkgs#376988 (comment)) so I'm going to merge this now

Let me clarify my comment: I was mostly thinking about two scenarios:

passing elaborated systems to the nixos/nixpkgs module
accessing config.nixpkgs.hostPlatform etc.

The fixes for those will be unproblematic, now that we had a revert.

I am not sure that this applies to the change in this PR, though: I still can't wrap my head around the "copying of the whole module", but I can imagine that you will get similar breakage now with this version in nix-darwin and the revert upstream.

So... better be prepared to apply the reverted changes here as well.

emilazy requested a review from Enzime February 3, 2025 19:46

emilazy mentioned this pull request Feb 3, 2025

pkgs/top-level: error: assertion '(((builtins).isAttrs localSystem) -> (! ((localSystem) ? parsed)))' failed #1317

Closed

emilazy and others added 4 commits February 3, 2025 20:25

nixpkgs: fix undefined variable in assertion

5b0cffe

This got mangled in the backport a year and a half ago. Fixes: e25eeff

nixpkgs: assert that nixpkgs.config is not set when pkgs is passed in…

962eb3f

… externally This is a common footgun people hit often. Remove it. Backport of Nixpkgs commit ce87196a00214a0062ece1c3e03a9a97f563580f. Co-authored-by: K900 <[email protected]>

nixpkgs: Rewrite overlays option docs

3cd3a79

henrik-ch was also here :) Backport of Nixpkgs commit 11406bdc0e5af9b3c8a8d597da23349238c65277. Co-authored-by: Silvan Mosberger <[email protected]> Co-Authored-By: Valentin Gagarin <[email protected]>

nixpkgs: use less confusing example systems

2df9e48

system and config shouldn't both be specified — each will be filled in based on the other when the system is elaborated. Backport of Nixpkgs commit a3ba0495452cd8e72735ebd4472838e96902a259. Co-authored-by: Alyssa Ross <[email protected]>

emilazy force-pushed the push-tmtstvpkwkow branch 2 times, most recently from 921dcde to fb70457 Compare February 3, 2025 20:29

emilazy force-pushed the push-tmtstvpkwkow branch from fb70457 to 537056b Compare February 3, 2025 20:42

emilazy and others added 8 commits February 3, 2025 20:44

nixpkgs: link to Nixpkgs manual for global configuration options

320bf02

Backport of Nixpkgs commit e6057cfd59f278db3aeb058a4e1e0bcc24696267. Co-authored-by: Valentin Gagarin <[email protected]> Co-authored-by: Dominic Mills <[email protected]>

nixpkgs: remove with lib;

bd1d467

Backport of Nixpkgs commit 609e57485d1fa111e3a689498d9d338dc03a7bc5. Co-authored-by: Felix Buehler <[email protected]>

nixpkgs: show definition files in config assertion

80eddf2

Backport of Nixpkgs commit 6d9dfef94ffd59a327573eea7bc709a84c44b3d2. Co-authored-by: Matt Sturgeon <[email protected]>

nixpkgs: format with nixfmt

dc1c716

git-blame-ignore-revs: add nixpkgs module formatting commit

5084b33

emilazy force-pushed the push-tmtstvpkwkow branch from 537056b to 7c72c01 Compare February 3, 2025 21:11

MattSturgeon mentioned this pull request Feb 4, 2025

[main] Update flake.lock & generated files nix-community/nixvim#2964

Merged

Enzime approved these changes Feb 5, 2025

View reviewed changes

Enzime merged commit ae406c0 into LnL7:master Feb 5, 2025
3 checks passed

EricCrosson added a commit to EricCrosson/dotfiles that referenced this pull request Feb 5, 2025

build: update flake inputs

5ffc891

to bring in LnL7/nix-darwin#1318

emilazy deleted the push-tmtstvpkwkow branch February 5, 2025 21:06

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

nixpkgs: backports from upstream #1318

nixpkgs: backports from upstream #1318

emilazy commented Feb 3, 2025

cixel commented Feb 3, 2025

jwiegley commented Feb 3, 2025

emilazy commented Feb 3, 2025

ossareh commented Feb 3, 2025 •

edited

Loading

cixel commented Feb 3, 2025 •

edited

Loading

emilazy commented Feb 3, 2025

emilazy commented Feb 3, 2025

emilazy commented Feb 3, 2025

cixel commented Feb 3, 2025 •

edited

Loading

emilazy commented Feb 4, 2025

wolfgangwalther commented Feb 4, 2025

cixel commented Feb 4, 2025 •

edited

Loading

Enzime left a comment

wolfgangwalther commented Feb 5, 2025

nixpkgs: backports from upstream #1318

nixpkgs: backports from upstream #1318

Conversation

emilazy commented Feb 3, 2025

cixel commented Feb 3, 2025

jwiegley commented Feb 3, 2025

emilazy commented Feb 3, 2025

ossareh commented Feb 3, 2025 • edited Loading

cixel commented Feb 3, 2025 • edited Loading

emilazy commented Feb 3, 2025

emilazy commented Feb 3, 2025

emilazy commented Feb 3, 2025

cixel commented Feb 3, 2025 • edited Loading

emilazy commented Feb 4, 2025

wolfgangwalther commented Feb 4, 2025

cixel commented Feb 4, 2025 • edited Loading

Enzime left a comment

Choose a reason for hiding this comment

wolfgangwalther commented Feb 5, 2025

ossareh commented Feb 3, 2025 •

edited

Loading

cixel commented Feb 3, 2025 •

edited

Loading

cixel commented Feb 3, 2025 •

edited

Loading

cixel commented Feb 4, 2025 •

edited

Loading