Table of Contents
Table of Contents
nvf is a highly modular, configurable, extensible and easy to use Neovim configuration in Nix. Designed for flexibility and ease of use, nvf allows you to easily configure your fully featured Neovim instance with a few lines of Nix.
If you notice any issues with nvf, or this documentation, then please consider reporting them over at the issue tracker. Issues tab, in addition to the discussions tab is a good place as any to request new features.
You may also consider submitting bugfixes, feature additions and upstreamed changes that you think are critical over at the pull requests tab.
Table of Contents
Thanks to the portability of Nix, you can try out nvf without actually installing it to your machine. Below are the commands you may run to try out different configurations provided by this flake. As of v0.5, two specialized configurations are provided:
Nix - Nix language server + simple utility plugins
Maximal - Variable language servers + utility and decorative plugins
You may try out any of the provided configurations using the nix run
command
on a system where Nix is installed.
$ cachix use nvf # Optional: it'll save you CPU resources and time
$ nix run github:notashelf/nvf#nix # will run the default minimal configuration
Do keep in mind that this is susceptible to garbage collection meaning it will be removed from your Nix store once you garbage collect.
$ nix run github:notashelf/nvf#nix
$ nix run github:notashelf/nvf#maximal
Nix
configuration by default provides LSP/diagnostic support for Nix alongside
a set of visual and functional plugins. By running nix run .#
, which is the
default package, you will build Neovim with this config.
Maximal
is the ultimate configuration that will enable support for more
commonly used language as well as additional complementary plugins. Keep in
mind, however, that this will pull a lot of dependencies.
You are strongly recommended to use the binary cache if you would like to try the Maximal configuration.
While you can configure nvf yourself using the builder, you can also use the pre-built configs that are available. Here are a few default configurations you can use.
$ nix shell github:notashelf/nvf#maximal test.nix
It is the same fully configured Neovim as with the Nix configuration, but with every supported language enabled.
::: {.note} Running the maximal config will download a lot of packages as it is downloading language servers, formatters, and more. :::
$ nix run github:notashelf/nvf#nix test.nix
Enables all the of Neovim plugins, with language support for specifically Nix. This lets you see what a fully configured neovim setup looks like without downloading a whole bunch of language servers and associated tools.
There are multiple ways of installing nvf on your system. You may either choose the standalone installation method, which does not depend on a module system and may be done on any system that has the Nix package manager or the appropriate modules for NixOS and home-manager as described in the module installation section.
Table of Contents
It is possible to install nvf without depending on NixOS or Home-Manager as the
parent module system, using the neovimConfiguration
function exposed in the
extended library. This function will take modules
and extraSpecialArgs
as
arguments, and return the following schema as a result.
{
options = "The options that were available to configure";
config = "The outputted configuration";
pkgs = "The package set used to evaluate the module";
neovim = "The built neovim package";
}
An example flake that exposes your custom Neovim configuration might look like
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
nvf.url = "github:notashelf/nvf";
};
outputs = {
self,
nixpkgs,
...
} @ inputs: {
packages."x86_64-linux" = let
neovimConfigured = (inputs.nvf.lib.neovimConfiguration {
inherit (nixpkgs.legacyPackages."x86_64-linux") pkgs;
modules = [{
config.vim = {
# Enable custom theming options
theme.enable = true;
# Enable Treesitter
tree-sitter.enable = true;
# Other options will go here. Refer to the config
# reference in Appendix B of the nvf manual.
# ...
};
}];
});
in {
# Set the default package to the wrapped instance of Neovim.
# This will allow running your Neovim configuration with
# `nix run` and in addition, sharing your configuration with
# other users in case your repository is public.
default = neovimConfigured.neovim;
};
};
}
The above setup will allow to set up nvf as a standalone flake, which you can build independently from your system configuration while also possibly sharing it with others. The next two chapters will detail specific usage of such a setup for a package output in the context of NixOS or Home-Manager installation.
Your built Neovim configuration can be exposed as a flake output to make it easier to share across machines, repositories and so on. Or it can be added to your system packages to make it available across your system.
The following is an example installation of nvf
as a standalone package with
the default theme enabled. You may use other options inside config.vim
in
configModule
, but this example will not cover that extensively.
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
home-manager.url = "github:nix-community/home-manager";
nvf.url = "github:notashelf/nvf";
};
outputs = {nixpkgs, nvf, ...}: let
system = "x86_64-linux";
pkgs = nixpkgs.legacyPackages.${system};
configModule = {
# Add any custom options (and do feel free to upstream them!)
# options = { ... };
config.vim = {
theme.enable = true;
# and more options as you see fit...
};
};
customNeovim = nvf.lib.neovimConfiguration {
inherit pkgs;
modules = [configModule];
};
in {
# This will make the package available as a flake output under 'packages'
packages.${system}.my-neovim = customNeovim.neovim;
# Example nixosConfiguration using the configured Neovim package
nixosConfigurations = {
yourHostName = nixpkgs.lib.nixosSystem {
# ...
modules = [
# This will make wrapped neovim available in your system packages
{environment.systemPackages = [customNeovim.neovim];}
];
# ...
};
};
};
}
Your built Neovim configuration can be exposed as a flake output to make it easier to share across machines, repositories and so on. Or it can be added to your system packages to make it available across your system.
The following is an example installation of nvf
as a standalone package with
the default theme enabled. You may use other options inside config.vim
in
configModule
, but this example will not cover that extensively.
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
home-manager.url = "github:nix-community/home-manager";
nvf.url = "github:notashelf/nvf";
};
outputs = {nixpkgs, home-manager, nvf, ...}: let
system = "x86_64-linux";
pkgs = nixpkgs.legacyPackages.${system};
configModule = {
# Add any custom options (and do feel free to upstream them!)
# options = { ... };
config.vim = {
theme.enable = true;
# and more options as you see fit...
};
};
customNeovim = nvf.lib.neovimConfiguration {
inherit pkgs;
modules = [configModule];
};
in {
# This will make the package available as a flake output under 'packages'
packages.${system}.my-neovim = customNeovim.neovim;
# Example Home-Manager configuration using the configured Neovim package
homeConfigurations = {
"your-username@your-hostname" = home-manager.lib.homeManagerConfiguration {
# ...
modules = [
# This will make Neovim available to users using the Home-Manager
# configuration. To make the package available to all users, prefer
# environment.systemPackages in your NixOS configuration.
{home.packages = [customNeovim.neovim];}
];
# ...
};
};
};
}
Table of Contents
The below chapters will describe installing nvf as NixOS and Home-Manager modules. Note that those methods are mutually exclusive, and will likely cause path collisions if used simultaneously.
Table of Contents
The NixOS module allows us to customize the different vim
options from inside
the NixOS configuration without having to call for the wrapper yourself. It is
the recommended way to use nvf alongside the home-manager module depending
on your needs.
To use it, we first add the input flake.
{
inputs = {
obsidian-nvim.url = "github:epwalsh/obsidian.nvim";
nvf = {
url = "github:notashelf/nvf";
# you can override input nixpkgs
inputs.nixpkgs.follows = "nixpkgs";
# you can also override individual plugins
# for example:
inputs.obsidian-nvim.follows = "obsidian-nvim"; # <- this will use the obsidian-nvim from your inputs
};
};
}
Followed by importing the NixOS module somewhere in your configuration.
{
# assuming nvf is in your inputs and inputs is in the argset
# see example below
imports = [ inputs.nvf.nixosModules.default ];
}
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
nvf.url = "github:notashelf/nvf";
};
outputs = { nixpkgs, nvf, ... }: let
system = "x86_64-linux"; in {
# ↓ this is your host output in the flake schema
nixosConfigurations."yourUsername»" = nixpkgs.lib.nixosSystem {
modules = [
nvf.nixosModules.default # <- this imports the NixOS module that provides the options
./configuration.nix # <- your host entrypoint
];
};
};
}
Once the module is properly imported by your host, you will be able to use the
programs.nvf
module option anywhere in your configuration in order to
configure nvf.
programs.nvf = {
enable = true;
# your settings need to go into the settings attribute set
# most settings are documented in the appendix
settings = {
vim.viAlias = false;
vim.vimAlias = true;
vim.lsp = {
enable = true;
};
};
};
}
nvf exposes a lot of options, most of which are not referenced in the installation sections of the manual. You may find all available options in the appendix
Table of Contents
The home-manager module allows us to customize the different vim
options from
inside the home-manager configuration without having to call for the wrapper
yourself. It is the recommended way to use nvf alongside the NixOS module
depending on your needs.
To use it, we first add the input flake.
{
inputs = {
obsidian-nvim.url = "github:epwalsh/obsidian.nvim";
nvf = {
url = "github:notashelf/nvf";
# you can override input nixpkgs
inputs.nixpkgs.follows = "nixpkgs";
# you can also override individual plugins
# for example:
inputs.obsidian-nvim.follows = "obsidian-nvim"; # <- this will use the obsidian-nvim from your inputs
};
};
}
Followed by importing the home-manager module somewhere in your configuration.
{
# assuming nvf is in your inputs and inputs is in the argset
# see example below
imports = [ inputs.nvf.homeManagerModules.default ];
}
{
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
home-manager.url = "github:nix-community/home-manager";
nvf.url = "github:notashelf/nvf";
};
outputs = { nixpkgs, home-manager, nvf, ... }: let
system = "x86_64-linux"; in {
# ↓ this is your home output in the flake schema, expected by home-manager
"your-username@your-hostname" = home-manager.lib.homeManagerConfiguration
modules = [
nvf.homeManagerModules.default # <- this imports the home-manager module that provides the options
./home.nix # <- your home entrypoint
];
};
};
}
Once the module is properly imported by your host, you will be able to use the
programs.nvf
module option anywhere in your configuration in order to
configure nvf.
programs.nvf = {
enable = true;
# your settings need to go into the settings attribute set
# most settings are documented in the appendix
settings = {
vim.viAlias = false;
vim.vimAlias = true;
vim.lsp = {
enable = true;
};
};
};
}
nvf exposes a lot of options, most of which are not referenced in the installation sections of the manual. You may find all available options in the appendix
As of v0.5, you may now specify the Neovim package that will be wrapped with
your configuration. This is done with the vim.package
option.
{inputs, pkgs, ...}: {
# using the neovim-nightly overlay
vim.package = inputs.neovim-overlay.packages.${pkgs.system}.neovim;
}
The neovim-nightly-overlay always exposes an unwrapped package. If using a
different source, you are highly recommended to get an “unwrapped” version of
the neovim package, similar to neovim-unwrapped
in nixpkgs.
{ pkgs, ...}: {
# using the neovim-nightly overlay
vim.package = pkgs.neovim-unwrapped;
}
Table of Contents
nvf, by default, exposes a wide variety of plugins as module options for your convenience and bundles necessary dependencies into nvf’s runtime. In case a plugin is not available in nvf, you may consider making a pull request to nvf to include it as a module or you may add it to your configuration locally.
There are multiple ways of adding custom plugins to your nvf configuration.
You can use custom plugins, before they are implemented in the flake. To add a
plugin to the runtime, you need to add it to the vim.startPlugins
list
in your configuration.
Adding a plugin to startPlugins
will not allow you to configure the plugin
that you have added, but nvf provides multiple way of configuring any custom
plugins that you might have added to your configuration.
Just making the plugin to your Neovim configuration available might not always
be enough. In that case, you can write custom lua config using either
config.vim.lazy.plugins.*.setupOpts
config.vim.extraPlugins.*.setup
or
config.vim.luaConfigRC
.
The first option uses an extended version of lz.n
’s PluginSpec. setupModule
and setupOpt
can be used if the plugin uses a require('module').setup(...)
pattern. Otherwise, the before
and after
hooks should do what you need.
{
config.vim.lazy.plugins = {
aerial.nvim = {
# ^^^^^^^^^ this name should match the package.pname or package.name
package = aerial-nvim;
setupModule = "aerial";
setupOpts = {option_name = false;};
after = "print('aerial loaded')";
};
};
}
The second option uses an attribute set, which maps DAG section names to a
custom type, which has the fields package
, after
, setup
. They allow you to
set the package of the plugin, the sections its setup code should be after (note
that the extraPlugins
option has its own DAG scope), and the its setup code
respectively. For example:
config.vim.extraPlugins = with pkgs.vimPlugins; {
aerial = {
package = aerial-nvim;
setup = "require('aerial').setup {}";
};
harpoon = {
package = harpoon;
setup = "require('harpoon').setup {}";
after = ["aerial"]; # place harpoon configuration after aerial
};
}
The third option also uses an attribute set, but this one is resolved as a DAG directly. The attribute names denote the section names, and the values lua code. For example:
{
# this will create an "aquarium" section in your init.lua with the contents of your custom config
# which will be *appended* to the rest of your configuration, inside your init.vim
config.vim.luaConfigRC.aquarium = "vim.cmd('colorscheme aquiarum')";
}
One of the greatest strengths of nvf is the ability to order
snippets of configuration via the DAG system. It will allow specifying positions
of individual sections of configuration as needed. nvf provides helper functions
in the extended library, usually under inputs.nvf.lib.nvim.dag
that you may
use.
Please refer to the DAG section in the nvf manual to find out more about the DAG system.
As of version 0.7, we exposed an API for configuring lazy-loaded plugins via
lz.n
and lzn-auto-require
.
{
config.vim.lazy.plugins = {
"aerial.nvim" = {
package = pkgs.vimPlugins.aerial-nvim;
setupModule = "aerial";
setupOpts = {
option_name = true;
};
after = ''
-- custom lua code to run after plugin is loaded
print('aerial loaded')
'';
# Explicitly mark plugin as lazy. You don't need this if you define one of
# the trigger "events" below
lazy = true;
# load on command
cmd = ["AerialOpen"];
# load on event
event = ["BufEnter"];
# load on keymap
keys = [
{
key = "<leader>a";
action = ":AerialToggle<CR>";
}
];
};
};
}
As of version 0.5, we have a more extensive API for configuring plugins,
under vim.extraPlugins
. Instead of using DAGs exposed by the library, you may
use the extra plugin module as follows:
{
config.vim.extraPlugins = with pkgs.vimPlugins; {
aerial = {
package = aerial-nvim;
setup = ''
require('aerial').setup {
-- some lua configuration here
}
'';
};
harpoon = {
package = harpoon;
setup = "require('harpoon').setup {}";
after = ["aerial"];
};
};
}
Prior to version v0.5, the method of adding new plugins was adding the plugin
package to vim.startPlugins
and add its configuration as a DAG under one of
vim.configRC
or vim.luaConfigRC
. Users who have not yet updated to 0.5, or
prefer a more hands-on approach may use the old method where the load order of
the plugins is determined by DAGs.
To add a plugin not available in nvf as a module to your configuration, you may
add it to vim.startPlugins
in order to make it available to Neovim at
runtime.
{pkgs, ...}: {
# Add a Neovim plugin from Nixpkgs to the runtime.
vim.startPlugins = [pkgs.vimPlugins.aerial-nvim];
}
And to configure the added plugin, you can use the luaConfigRC
option to
provide configuration as a DAG using the nvf extended library.
{inputs, ...}: let
# This assumes you have an input called 'nvf' in your flake inputs
# and 'inputs' in your specialArgs. In the case you have passed 'nvf'
# to specialArgs, the 'inputs' prefix may be omitted.
inherit (inputs.nvf.lib.nvim.dag) entryAnywhere;
in {
vim.luaConfigRC.aerial-nvim= entryAnywhere ''
require('aerial').setup {
-- your configuration here
}
'';
}
One of the greatest strengths of nvf is its ability to get plugins from flake inputs and build them locally from any given source. For plugins that do not require any kind of additional building step, this is a powerful method of adding plugins to your configuration that are not packaged in nixpkgs, or those you want to track from source without relying on nixpkgs.
The additional plugins section details the addition
of new plugins to nvf under regular circumstances, i.e. while making a pull
request to the project. You may override those plugin inputs in your own
flake.nix
to change source versions, e.g., to use newer versions of plugins
that are not yet updated in nvf.
{
inputs = {
# ...
# The name here is arbitrary, you can name it whatever.
# This will add a plugin input called "your-neodev-input"
# that you can reference in a `follows` line.
your-neodev-input = {
url = "github:folke/neodev.nvim";
flake = false;
};
nvf = {
url = "github:notashelf/nvf";
# The name of the input must match for the follows line
# plugin-neodev-nvim is what the input is called inside nvf
# so you must match the exact name here.
inputs.plugin-neodev-nvim.follows = "your-neodev-input";
};
# ...
};
}
This will override the source for the neodev.nvim
plugin that is used in nvf
with your own input. You can update your new input via nix flake update
or
more specifically nix flake update <name of your input>
to keep it up to date.
While updating plugin inputs, make sure that any configuration that has been
deprecated in newer versions is changed in the plugin’s setupOpts
. If you
depend on a new version, requesting a version bump in the issues section is a
more reliable option.
Table of Contents
Language specific support means there is a combination of language specific
plugins, treesitter
support, nvim-lspconfig
language servers, and null-ls
integration. This gets you capabilities ranging from autocompletion to
formatting to diagnostics. The following languages have sections under the
vim.languages
attribute.
C/C++: vim.languages.clang.enable
Typescript/Javascript: vim.languages.ts.enable
Python: vim.languages.python.enable:
Markdown: vim.languages.markdown.enable
Adding support for more languages, and improving support for existing ones are great places where you can contribute with a PR.
In any of the opt.languages.<language>.lsp.package
options you can provide
your own LSP package, or provide the command to launch the language server, as a
list of strings. You can use this to skip automatic installation of a language
server, and instead use the one found in your $PATH
during runtime, for
example:
vim.languages.java = {
lsp = {
enable = true;
# this expects jdt-language-server to be in your PATH
# or in `vim.extraPackages`
package = ["jdt-language-server" "-data" "~/.cache/jdtls/workspace"];
};
}
Table of Contents
We conform to the NixOS options types for the most part, however, a noteworthy
addition for certain options is the
DAG (Directed acyclic graph)
type which is borrowed from home-manager’s extended library. This type is most
used for topologically sorting strings. The DAG type allows the attribute set
entries to express dependency relations among themselves. This can, for example,
be used to control the order of configuration sections in your luaConfigRC
.
The below section, mostly taken from the home-manager manual explains in more detail the overall usage logic of the DAG type.
lib.dag.entryAnywhere (value: T) : DagEntry<T>
Indicates that value
can be placed anywhere within the DAG. This is also the
default for plain attribute set entries, that is
foo.bar = {
a = lib.dag.entryAnywhere 0;
}
and
foo.bar = {
a = 0;
}
are equivalent.
lib.dag.entryAfter (afters: list string) (value: T) : DagEntry<T>
Indicates that value
must be placed after each of the attribute names in the
given list. For example
foo.bar = {
a = 0;
b = lib.dag.entryAfter [ "a" ] 1;
}
would place b
after a
in the graph.
lib.dag.entryBefore (befores: list string) (value: T) : DagEntry<T>
Indicates that value
must be placed before each of the attribute names in
the given list. For example
foo.bar = {
b = lib.dag.entryBefore [ "a" ] 1;
a = 0;
}
would place b
before a
in the graph.
lib.dag.entryBetween (befores: list string) (afters: list string) (value: T) : DagEntry<T>
Indicates that value
must be placed before the attribute names in the first
list and after the attribute names in the second list. For example
foo.bar = {
a = 0;
c = lib.dag.entryBetween [ "b" ] [ "a" ] 2;
b = 1;
}
would place c
before b
and after a
in the graph.
There are also a set of functions that generate a DAG from a list. These are
convenient when you just want to have a linear list of DAG entries, without
having to manually enter the relationship between each entry. Each of these
functions take a tag
as argument and the DAG entries will be named
${tag}-${index}
.
lib.dag.entriesAnywhere (tag: string) (values: [T]) : Dag<T>
Creates a DAG with the given values with each entry labeled using the given tag. For example
foo.bar = lib.dag.entriesAnywhere "a" [ 0 1 ];
is equivalent to
foo.bar = {
a-0 = 0;
a-1 = lib.dag.entryAfter [ "a-0" ] 1;
}
lib.dag.entriesAfter (tag: string) (afters: list string) (values: [T]) : Dag<T>
Creates a DAG with the given values with each entry labeled using the given tag.
The list of values are placed are placed after each of the attribute names in
afters
. For example
foo.bar =
{ b = 0; } // lib.dag.entriesAfter "a" [ "b" ] [ 1 2 ];
is equivalent to
foo.bar = {
b = 0;
a-0 = lib.dag.entryAfter [ "b" ] 1;
a-1 = lib.dag.entryAfter [ "a-0" ] 2;
}
lib.dag.entriesBefore (tag: string) (befores: list string) (values: [T]) : Dag<T>
Creates a DAG with the given values with each entry labeled using the given tag.
The list of values are placed before each of the attribute names in befores
.
For example
foo.bar =
{ b = 0; } // lib.dag.entriesBefore "a" [ "b" ] [ 1 2 ];
is equivalent to
foo.bar = {
b = 0;
a-0 = 1;
a-1 = lib.dag.entryBetween [ "b" ] [ "a-0" ] 2;
}
lib.dag.entriesBetween (tag: string) (befores: list string) (afters: list string) (values: [T]) : Dag<T>
Creates a DAG with the given values with each entry labeled using the given tag.
The list of values are placed before each of the attribute names in befores
and after each of the attribute names in afters
. For example
foo.bar =
{ b = 0; c = 3; } // lib.dag.entriesBetween "a" [ "b" ] [ "c" ] [ 1 2 ];
is equivalent to
foo.bar = {
b = 0;
c = 3;
a-0 = lib.dag.entryAfter [ "c" ] 1;
a-1 = lib.dag.entryBetween [ "b" ] [ "a-0" ] 2;
}
Table of Contents
From the previous chapter, it should be clear that DAGs are useful, because you can add code that relies on other code. However, if you don’t know what the entries are called, it’s hard to do that, so here is a list of the internal entries in nvf:
vim.luaConfigRC
(top-level DAG) (luaConfigPre
) - not a part of the actual DAG, instead, it’s simply
inserted before the rest of the DAG
globalsScript
- used to set globals defined in vim.globals
basic
- used to set basic configuration options
optionsScript
- used to set options defined in vim.o
theme
(this is simply placed before pluginConfigs
and lazyConfigs
,
meaning that surrounding entries don’t depend on it) - used to set up the
theme, which has to be done before other plugins
lazyConfigs
- lz.n
and lzn-auto-require
configs. If vim.lazy.enable
is false, this will contain each plugin’s config instead.
pluginConfigs
- the result of the nested vim.pluginRC
(internal option,
see the Custom Plugins page for adding your
own plugins) DAG, used to set up internal plugins
extraPluginConfigs
- the result of vim.extraPlugins
, which is not a
direct DAG, but is converted to, and resolved as one internally
mappings
- the result of vim.maps
Table of Contents
nvf is designed for the developer as much as it is designed for the end-user. We would like for any contributor to be able to propagate their changes, or add new features to the project with minimum possible friction. As such, below are the guides and guidelines written to streamline the contribution process and to ensure that your valuable input integrates into nvf’s development as seamlessly as possible without leaving any question marks in your head.
This section is directed mainly towards those who wish to contribute code into the project. If you instead wish to report a bug, or discuss a potential new feature implementation (which you do not wish to implement yourself) first look among the already open issues and if no matching issue exists you may open a new issue and describe your problem/request.
While creating an issue, please try to include as much information as you can, ideally also include relevant context in which an issue occurs or a feature should be implemented. If you wish to make a contribution, but feel stuck - please do not be afraid to submit a pull request, we will help you get it in.
You, naturally, would like to start by forking the repository to get started. If
you are new to Git and GitHub, do have a look at GitHub’s
Fork a repo guide for
instructions on how you can do this. Once you have a fork of nvf, you should
create a separate branch based on the most recent main
branch. Give your
branch a reasonably descriptive name (e.g. feature/debugger
or
fix/pesky-bug
) and you are ready to work on your changes
Implement your changes and commit them to the newly created branch and when you are happy with the result, and positive that it fulfills our Contributing Guidelines, push the branch to GitHub and create a pull request. The default pull request template available on the nvf repository will guide you through the rest of the process, and we’ll gently nudge you in the correct direction if there are any mistakes.
If your contribution tightly follows the guidelines, then there is a good chance it will be merged without too much trouble. Some of the guidelines will be strictly enforced, others will remain as gentle nudges towards the correct direction. As we have no automated system enforcing those guidelines, please try to double check your changes before making your pull request in order to avoid “faulty” code slipping by.
If you are uncertain how these rules affect the change you would like to make then feel free to start a discussion in the discussions tab ideally (but not necessarily) before you start developing.
Most, if not all, changes warrant changes to the documentation. Module options should be documented with Nixpkgs-flavoured Markdown, albeit with exceptions.
::: {.note} As of v0.5, nvf is itself documented using full markdown in both module options and the manual. With v0.6, this manual has also been converted to markdown in full. :::
The HTML version of this manual containing both the module option descriptions and the documentation of nvf (such as this page) can be generated and opened by typing the following in a shell within a clone of the nvf Git repository:
$ nix build .#docs-html
$ xdg-open $PWD/result/share/doc/nvf/index.html
Make sure your code is formatted as described in code-style section. To maintain consistency throughout the project you are encouraged to browse through existing code and adopt its style also in new code.
Similar to code style guidelines we encourage a consistent commit message format as described in commit style guidelines.
The commits in your pull request should be reasonably self-contained. Which means each and every commit in a pull request should make sense both on its own and in general context. That is, a second commit should not resolve an issue that is introduced in an earlier commit. In particular, you will be asked to amend any commit that introduces syntax errors or similar problems even if they are fixed in a later commit.
The commit messages should follow the seven rules, except for “Capitalize the subject line”. We also ask you to include the affected code component or module in the first line. A commit message ideally, but not necessarily, follow the given template from home-manager’s own documentation
{component}: {description}
{long description}
where {component}
refers to the code component (or module) your change
affects, {description}
is a very brief description of your change, and
{long description}
is an optional clarifying description. As a rare exception,
if there is no clear component, or your change affects many components, then the
{component}
part is optional. See
example commit message for a commit message
that fulfills these requirements.
The commit 69f8e47e9e74c8d3d060ca22e18246b7f7d988ef in home-manager contains the following commit message.
starship: allow running in Emacs if vterm is used
The vterm buffer is backed by libvterm and can handle Starship prompts
without issues.
Similarly, if you are contributing to nvf, you would include the scope of the commit followed by the description:
languages/ruby: init module
Adds a language module for Ruby, adds appropriate formatters and Treesitter grammars
Long description can be omitted if the change is too simple to warrant it. A minor fix in spelling or a formatting change does not warrant long description, however, a module addition or removal does as you would like to provide the relevant context, i.e. the reasoning behind it, for your commit.
Finally, when adding a new module, say modules/foo.nix
, we use the fixed
commit format foo: add module
. You can, of course, still include a long
description if you wish.
In case of nested modules, i.e modules/languages/java.nix
you are recommended
to contain the parent as well - for example languages/java: some major change
.
Keep lines at a reasonable width, ideally 80 characters or less. This also applies to string literals and module descriptions and documentation.
nvf is formatted by the
alejandra tool and the formatting is
checked in the pull request and push workflows. Run the nix fmt
command inside
the project repository before submitting your pull request.
While Alejandra is mostly opinionated on how code looks after formatting, certain changes are done at the user’s discretion based on how the original code was structured.
Please use one line code for attribute sets that contain only one subset. For example:
# parent modules should always be unfolded
# which means module = { value = ... } instead of module.value = { ... }
module = {
value = mkEnableOption "some description" // { default = true; }; # merges can be done inline where possible
# same as parent modules, unfold submodules
subModule = {
# this is an option that contains more than one nested value
someOtherValue = mkOption {
type = lib.types.bool;
description = "Some other description";
default = true;
};
};
}
If you move a line down after the merge operator, Alejandra will automatically unfold the whole merged attrset for you, which we do not want.
module = {
key = mkEnableOption "some description" // {
default = true; # we want this to be inline
}; # ...
}
For lists, it is mostly up to your own discretion how you want to format them, but please try to unfold lists if they contain multiple items and especially if they are to include comments.
# this is ok
acceptableList = [
item1 # comment
item2
item3 # some other comment
item4
];
# this is not ok
listToBeAvoided = [item1 item2 /* comment */ item3 item4];
# this is ok
acceptableList = [item1 item2];
# this is also ok if the list is expected to contain more elements
acceptableList= [
item1
item2
# more items if needed...
];
Once you have made your changes, you will need to test them thoroughly. If it is
a module, add your module option to configuration.nix
(located in the root of
this project) inside neovimConfiguration
. Enable it, and then run the maximal
configuration with nix run .#maximal -Lv
to check for build errors. If neovim
opens in the current directory without any error messages (you can check the
output of :messages
inside neovim to see if there are any errors), then your
changes are good to go. Open your pull request, and it will be reviewed as soon
as possible.
If it is not a new module, but a change to an existing one, then make sure the
module you have changed is enabled in the maximal configuration by editing
configuration.nix
, and then run it with nix run .#maximal -Lv
. Same
procedure as adding a new module will apply here.
As of 0.4, there exists an API for writing your own keybinds and a couple of useful utility functions are available in the extended standard library. The following section contains a general overview to how you may utilize said functions.
To set a mapping, you should define it in vim.keymaps
.
An example, simple keybinding, can look like this:
{
vim.keymaps = [
{
key = "<leader>wq";
mode = ["n"];
action = ":wq<CR>";
silent = true;
desc = "Save file and quit";
}
];
}
There are many settings available in the options. Please refer to the documentation to see a list of them.
nvf provides a helper function, so that you don’t have to write the mapping attribute sets every time:
mkKeymap
, which mimics neovim’s vim.keymap.set
function
You can read the source code of some modules to see them in action, but the usage should look something like this:
# plugindefinition.nix
{lib, ...}: let
inherit (lib.options) mkEnableOption;
inherit (lib.nvim.binds) mkMappingOption;
in {
options.vim.plugin = {
enable = mkEnableOption "Enable plugin";
# Mappings should always be inside an attrset called mappings
mappings = {
workspaceDiagnostics = mkMappingOption "Workspace diagnostics [trouble]" "<leader>lwd";
documentDiagnostics = mkMappingOption "Document diagnostics [trouble]" "<leader>ld";
lspReferences = mkMappingOption "LSP References [trouble]" "<leader>lr";
quickfix = mkMappingOption "QuickFix [trouble]" "<leader>xq";
locList = mkMappingOption "LOCList [trouble]" "<leader>xl";
symbols = mkMappingOption "Symbols [trouble]" "<leader>xs";
};
}
# config.nix
{
config,
lib,
options,
...
}: let
inherit (lib.modules) mkIf;
inherit (lib.nvim.binds) mkKeymap;
cfg = config.vim.plugin;
keys = cfg.mappings;
inherit (options.vim.lsp.trouble) mappings;
in {
config = mkIf cfg.enable {
vim.keymaps = [
(mkKeymap "n" keys.workspaceDiagnostics "<cmd>Trouble toggle diagnostics<CR>" {desc = mappings.workspaceDiagnostics.description;})
(mkKeymap "n" keys.documentDiagnostics "<cmd>Trouble toggle diagnostics filter.buf=0<CR>" {desc = mappings.documentDiagnostics.description;})
(mkKeymap "n" keys.lspReferences "<cmd>Trouble toggle lsp_references<CR>" {desc = mappings.lspReferences.description;})
(mkKeymap "n" keys.quickfix "<cmd>Trouble toggle quickfix<CR>" {desc = mappings.quickfix.description;})
(mkKeymap "n" keys.locList "<cmd>Trouble toggle loclist<CR>" {desc = mappings.locList.description;})
(mkKeymap "n" keys.symbols "<cmd>Trouble toggle symbols<CR>" {desc = mappings.symbols.description;})
];
};
}
If you have come across a plugin that has an API that doesn’t seem to easily allow custom keybindings, don’t be scared to implement a draft PR. We’ll help you get it done.
To add a new Neovim plugin, first add the source url in the inputs section of
flake.nix
with the prefix plugin-
{
inputs = {
# ...
plugin-neodev-nvim = {
url = "github:folke/neodev.nvim";
flake = false;
};
# ...
};
}
Prepending plugin-
to the name of the input will allow nvf to automatically
discover inputs that are marked as plugins, and make them available in
vim.startPlugins
or other areas that require a very specific plugin type as it
is defined in https://github.com/notashelf/nvf/blob/v0.8/lib/types/plugins.nix
The addition of the plugin-
prefix will allow nvf to autodiscover the
input from the flake inputs automatically, allowing you to refer to it in areas
that require a very specific plugin type as defined in lib/types/plugins.nix
You can now reference this plugin using its string name, the plugin will be built with the name and source URL from the flake input, allowing you to refer to it as a string.
config.vim.startPlugins = ["neodev-nvim"];
Most plugins is initialized with a call to require('plugin').setup({...})
.
We use a special function that lets you easily add support for such setup
options in a modular way: mkPluginSetupOption
.
Once you have added the source of the plugin as shown above, you can define the setup options like this:
# in modules/.../your-plugin/your-plugin.nix
{lib, ...}:
let
inherit (lib.types) bool int;
inherit (lib.nvim.types) mkPluginSetupOption;
in {
options.vim.your-plugin = {
setupOpts = mkPluginSetupOption "plugin name" {
enable_feature_a = mkOption {
type = bool;
default = false;
# ...
};
number_option = mkOption {
type = int;
default = 3;
# ...
};
};
};
}
# in modules/.../your-plugin/config.nix
{lib, config, ...}:
let
cfg = config.vim.your-plugin;
in {
vim.luaConfigRC = lib.nvim.dag.entryAnywhere ''
require('plugin-name').setup(${lib.nvim.lua.toLuaObject cfg.setupOpts})
'';
}
This above config will result in this lua script:
require('plugin-name').setup({
enable_feature_a = false,
number_option = 3,
})
Now users can set any of the pre-defined option field, and can also add their own fields!
# in user's config
{
vim.your-plugin.setupOpts = {
enable_feature_a = true;
number_option = 4;
another_field = "hello";
size = { # nested fields work as well
top = 10;
};
};
}
As you’ve seen above, toLuaObject
is used to convert our nix attrSet
cfg.setupOpts
, into a lua table. Here are some rules of the conversion:
nix null
converts to lua nil
number and strings convert to their lua counterparts
nix attrSet/list convert into lua tables
you can write raw lua code using lib.generators.mkLuaInline
. This function
is part of nixpkgs.
Example:
vim.your-plugin.setupOpts = {
on_init = lib.generators.mkLuaInline ''
function()
print('we can write lua!')
end
'';
}
If the plugin can be lazy-loaded, vim.lazy.plugins
should be used to add it.
Lazy plugins are managed by lz.n
.
# in modules/.../your-plugin/config.nix
{lib, config, ...}:
let
cfg = config.vim.your-plugin;
in {
vim.lazy.plugins.your-plugin = {
# instead of vim.startPlugins, use this:
package = "your-plugin";
# if your plugin uses the `require('your-plugin').setup{...}` pattern
setupModule = "your-plugin";
inherit (cfg) setupOpts;
# events that trigger this plugin to be loaded
event = ["DirChanged"];
cmd = ["YourPluginCommand"];
# keymaps
keys = [
# we'll cover this in detail in the keymaps section
{
key = "<leader>d";
mode = "n";
action = ":YourPluginCommand";
}
];
};
;
}
This results in the following lua code:
require('lz.n').load({
{
"name-of-your-plugin",
after = function()
require('your-plugin').setup({--[[ your setupOpts ]]})
end,
event = {"DirChanged"},
cmd = {"YourPluginCommand"},
keys = {
{"<leader>d", ":YourPluginCommand", mode = {"n"}},
},
}
})
A full list of options can be found [here](https://notashelf.github.io/nvf/options.html#opt-vim.lazy.plugins