Admins will be upgrading ORNL GitLab Servers on Saturday, 16 May 2026, from 7 AM until 11 AM EST. Repositories will experience intermittent outages during this time.
The propagated equivalent of `nativeBuildInputs`. This would be called `depsBuildHostPropagated` but for historical continuity. For example, if package `Y` has `propagatedNativeBuildInputs = [X]`, and package `Z` has `buildInputs = [Y]`, then package `Z` will be built as if it included package `X` in its `nativeBuildInputs`. If instead, package `Z` has `nativeBuildInputs = [Y]`, then `Z` will be built as if it included `X` in the `depsBuildBuild` of package `Z`, because of the sum of the two `-1` host offsets.
The propagated equivalent of `nativeBuildInputs`. This would be called `depsBuildHostPropagated` but for historical continuity. For example, if package `Y` has `propagatedNativeBuildInputs = [X]`, and package `Z` has `buildInputs = [Y]`, then package `Z` will be built as if it included package `X` in its `nativeBuildInputs`. Note that if instead, package `Z` has `nativeBuildInputs = [Y]`, then `X` will not be included at all.
-`services.dex` now restarts upon changes to the `.environmentFile` or entries in `.settings.staticClients[].secretFile` when the entry is a `path` type.
-`nixos-rebuild-ng`, a full rewrite of `nixos-rebuild` in Python, is available for testing. You can enable it by setting [system.rebuild.enableNg](options.html#opt-system.rebuild.enableNg) in your configuration (this will replace the old `nixos-rebuild`), or by adding `nixos-rebuild-ng` to your `environment.systemPackages` (in this case, it will live side-by-side with `nixos-rebuild` as `nixos-rebuild-ng`). It is expected that the next major version of NixOS (25.11) will enable `system.rebuild.enableNg` by default.
- A `nixos-rebuild build-image` sub-command has been added.
It allows users to build platform-specific (disk) images from their NixOS configurations. `nixos-rebuild build-image` works similar to the popular [nix-community/nixos-generators](https://github.com/nix-community/nixos-generators) project. See new [section on image building in the nixpkgs manual](https://nixos.org/manual/nixpkgs/unstable/#sec-image-nixos-rebuild-build-image). It is also available for `nixos-rebuild-ng`.
-`nixos-option` has been rewritten to a Nix expression called by a simple bash script. This lowers our maintenance threshold, makes eval errors less verbose, adds support for flake-based configurations, descending into `attrsOf` and `listOf` submodule options, and `--show-trace`.
It allows users to build platform-specific (disk) images from their NixOS configurations. `nixos-rebuild build-image` works similar to the popular [nix-community/nixos-generators](https://github.com/nix-community/nixos-generators) project. See new [section on image building in the nixpkgs manual](https://nixos.org/manual/nixpkgs/unstable/#sec-image-nixos-rebuild-build-image).
<!-- To avoid merge conflicts, consider adding your item at an arbitrary place in the list instead. -->
List of rules to be added to `/etc/xdg/pay-respects/rules`.
`pay-respects` will read the contents of these generated rules to recommend command corrections.
Each rule module should start with the `command` attribute that specifies the command name. See the [upstream documentation](https://codeberg.org/iff/pay-respects/src/branch/main/rules.md) for more information.
'';
};
aiIntegration=mkOption{
default=false;
example={
url="http://127.0.0.1:11434/v1/chat/completions";
model="llama3";
locale="nl-be";
};
description=''
Whether to enable `pay-respects`' LLM integration. When there is no rule for a given error, `pay-respects` can query an OpenAI-compatible API endpoint for command corrections.
- If this is set to `false`, all LLM-related features are disabled.
- If this is set to `true`, the default OpenAI endpoint will be used, using upstream's API key. This default API key may be rate-limited.
- You can also set a custom API endpoint, large language model and locale for command corrections. Simply access the `aiIntegration.url`, `aiIntegration.model` and `aiIntegration.locale` options, as described in the example.
- Take a look at the [services.ollama](#opt-services.ollama.enable) NixOS module if you wish to host a local large language model for `pay-respects`.
For all of these methods, you can set a custom secret API key by using the `_PR_AI_API_KEY` environment variable.
The accepted format is a lowercase [`ISO 639-1` language code](https://en.wikipedia.org/wiki/List_of_ISO_639_language_codes), followed by a dash '-', followed by a lowercase [`ISO 3166-1 alpha-2` country code](https://en.wikipedia.org/wiki/ISO_3166-1_alpha-2).