stylix/home-manager-integration: enable even when stylix.enable = false

[ThinkChaos]

Always enable this, even when !stylix.enable: if the HM module isn't imported, it prevents users from setting any stylix related options in their HM config.
That would go against the standard pattern for enable options: they should not affect what you're allowed to write in config, thus we need to import the module; nor affect the value of the module's options, thus we need to propagate the NixOS config.stylix values.

---

• I certify that I have the right to submit this contribution under the MIT license
• Commit messages adhere to Stylix commit conventions
• Theming changes adhere to the Stylix style guide
• Changes have been tested locally
• Changes have been tested in testbeds
• Each commit in this PR is suitable for backport to the current stable branch

[trueNAHO]

Thanks for the patch, but this seems to be a duplicate of the following pending patch review:

From e698150f68a2e443c9c01133ee072abaab3de47b Mon Sep 17 00:00:00 2001
From: NAHO <90870942+trueNAHO@users.noreply.github.com>
Date: Tue, 19 Aug 2025 23:36:32 +0200
Subject: [PATCH 2/3] stylix/home-manager-integration: do not guard on
 stylix.enable

Do not guard on stylix.enable because the original issue from commit
73b7d0f30039 ("stylix: guard home-manager-integration config (#1494)")
seems to have resolved itself in the meantime.

This allows testbeds to declare Stylix target options when Stylix is
disabled.

Reverts: 73b7d0f30039 ("stylix: guard home-manager-integration config (#1494)")
---
 stylix/home-manager-integration.nix | 32 ++++++++++++++---------------
 1 file changed, 15 insertions(+), 17 deletions(-)

diff --git a/stylix/home-manager-integration.nix b/stylix/home-manager-integration.nix
index a169e1a4f..51b1cbd59 100644
--- a/stylix/home-manager-integration.nix
+++ b/stylix/home-manager-integration.nix
@@ -208,22 +208,20 @@ in
     };
   };

-  config = lib.mkIf config.stylix.enable (
-    lib.optionalAttrs (options ? home-manager) (
-      lib.mkMerge [
-        (lib.mkIf config.stylix.homeManagerIntegration.autoImport {
-          home-manager.sharedModules =
-            [
-              config.stylix.homeManagerIntegration.module
-            ]
-            ++ lib.optionals config.stylix.homeManagerIntegration.followSystem copyModules;
-        })
-        (lib.mkIf config.home-manager.useGlobalPkgs {
-          home-manager.sharedModules = lib.singleton {
-            config.stylix.overlays.enable = false;
-          };
-        })
-      ]
-    )
+  config = lib.optionalAttrs (options ? home-manager) (
+    lib.mkMerge [
+      (lib.mkIf config.stylix.homeManagerIntegration.autoImport {
+        home-manager.sharedModules =
+          [
+            config.stylix.homeManagerIntegration.module
+          ]
+          ++ lib.optionals config.stylix.homeManagerIntegration.followSystem copyModules;
+      })
+      (lib.mkIf config.home-manager.useGlobalPkgs {
+        home-manager.sharedModules = lib.singleton {
+          config.stylix.overlays.enable = false;
+        };
+      })
+    ]
   );
 }

-- #1857

I will close this for now as duplicate.

[ThinkChaos]

Thanks for the reply!

I don't see that patch in the linked MR, is it pushed somewhere else?
Also if it's part of a big PR that's been in review for a while, maybe it makes sense to merge this before?

EDIT: found the PR #1857

[trueNAHO]

if it's part of a big PR that's been in review for a while

Reviewing this exact patch is pending in #1857 (comment), with the danger being that it reverts a previously known to work fix.

I am in favor of this change because it seems to work, although this assumption could be wrong. Since it changes the core, I would like approval from at least one of the other Stylix maintainers.

maybe it makes sense to merge this before?

Yes, it could be cherry-picked and merged earlier, but that does not change that its review has been pending. Please continue this discussion in #1857 (comment) to avoid further fragmentation.

I don't see that patch in the linked MR, is it pushed somewhere else?

[...]

EDIT: found the PR #1857

Sorry for providing the wrong GitHub URL. At least it was the correct patch. I updated my previous message.