[master] flake.lock: Update

[home-manager-ci]

Automated update by the update-flake-lock GitHub Action.

Flake lock file updates:

• Updated input 'nixpkgs':
    'github:NixOS/nixpkgs/c06b4ae3d6599a672a6210b7021d699c351eebda?narHash=sha256-wvfdLLWJ2I9oEpDd9PfMA8osfIZicoQ5MT1jIwNs9Tk%3D' (2026-03-13)
  → 'github:NixOS/nixpkgs/b40629efe5d6ec48dd1efba650c797ddbd39ace0?narHash=sha256-TJ3lSQtW0E2JrznGVm8hOQGVpXjJyXY2guAxku2O9A4%3D' (2026-03-18)

This PR was most recently updated by workflow run 23315011266.

[teto]

nix run .#tests -- neovim work for nixos-unstable but not make test. Turns out one is using nixos-unstable while the other is using nixpkgs-unstable and currently nixpkgs-unstable is older than nixos-unstable.
This is confusing and we should use a single source of truth. I would favor nixos-unstable since:

• if you are on nixos, you are more likely to use nixos-unstable than nixpkgs-unstable
• the allegedly advantage of nixpkgs-unstable being more up to date is not necessarily true as showcased here.

[khaneliman]

the allegedly advantage of nixpkgs-unstable being more up to date is not necessarily true as showcased here.

It's extremely rare for it to not be the case, but it does happen.

If you are on nixos, you are more likely to use nixos-unstable than nixpkgs-unstable

We don't make that assertion, we support anywhere nix can be run.

Much rather track nixpkgs-unstable for home level confguration but I think flake.nix is nixos-unstable atm and we should match.

[khaneliman]

@teto looks like neovim breaking change from upstream. Needs update for module injecting config, test assertion, or update to module to use legacyWrapper also affected nixvim nix-community/nixvim#4219

[teto]

yeah, that's what #8734 is for I dont merge in nixpkgs before testing my PR against HM. How shall we proceed ? can I push here or we merge this one than the fixes ?

[khaneliman]

yeah, that's what #8734 is for I dont merge in nixpkgs before testing my PR against HM. How shall we proceed ? can I push here or we merge this one than the fixes ?

Ideally, we don't introduce broken behavior in a flake update. I wait for green CI so a push here would be fine.

[teto]

I think the neovim commit fixes the neovim test. The remaining failure I am not familiar with. Seems like a keepass dependency not building on aarch64