Introduction

Nix is a purely functional package manager. This means that it treats packages like values in purely functional programming languages such as Haskell — they are built by functions that don’t have side-effects, and they never change after they have been built. Nix stores packages in the Nix store, usually the directory /nix/store, where each package has its own unique subdirectory such as

/nix/store/b6gvzjyb2pg0kjfwrjmg1vfhh54ad73z-firefox-33.1/

where b6gvzjyb2pg0… is a unique identifier for the package that captures all its dependencies (it’s a cryptographic hash of the package’s build dependency graph). This enables many powerful features.

Multiple versions

You can have multiple versions or variants of a package installed at the same time. This is especially important when different applications have dependencies on different versions of the same package — it prevents the “DLL hell”. Because of the hashing scheme, different versions of a package end up in different paths in the Nix store, so they don’t interfere with each other.

An important consequence is that operations like upgrading or uninstalling an application cannot break other applications, since these operations never “destructively” update or delete files that are used by other packages.

Complete dependencies

Nix helps you make sure that package dependency specifications are complete. In general, when you’re making a package for a package management system like RPM, you have to specify for each package what its dependencies are, but there are no guarantees that this specification is complete. If you forget a dependency, then the package will build and work correctly on your machine if you have the dependency installed, but not on the end user's machine if it's not there.

Since Nix on the other hand doesn’t install packages in “global” locations like /usr/bin but in package-specific directories, the risk of incomplete dependencies is greatly reduced. This is because tools such as compilers don’t search in per-packages directories such as /nix/store/5lbfaxb722zp…-openssl-0.9.8d/include, so if a package builds correctly on your system, this is because you specified the dependency explicitly. This takes care of the build-time dependencies.

Once a package is built, runtime dependencies are found by scanning binaries for the hash parts of Nix store paths (such as r8vvq9kq…). This sounds risky, but it works extremely well.

Multi-user support

Nix has multi-user support. This means that non-privileged users can securely install software. Each user can have a different profile, a set of packages in the Nix store that appear in the user’s PATH. If a user installs a package that another user has already installed previously, the package won’t be built or downloaded a second time. At the same time, it is not possible for one user to inject a Trojan horse into a package that might be used by another user.

Atomic upgrades and rollbacks

Since package management operations never overwrite packages in the Nix store but just add new versions in different paths, they are atomic. So during a package upgrade, there is no time window in which the package has some files from the old version and some files from the new version — which would be bad because a program might well crash if it’s started during that period.

And since packages aren’t overwritten, the old versions are still there after an upgrade. This means that you can roll back to the old version:

$ nix-env --upgrade --attr nixpkgs.some-package
$ nix-env --rollback

Garbage collection

When you uninstall a package like this…

$ nix-env --uninstall firefox

the package isn’t deleted from the system right away (after all, you might want to do a rollback, or it might be in the profiles of other users). Instead, unused packages can be deleted safely by running the garbage collector:

$ nix-collect-garbage

This deletes all packages that aren’t in use by any user profile or by a currently running program.

Functional package language

Packages are built from Nix expressions, which is a simple functional language. A Nix expression describes everything that goes into a package build task (a “derivation”): other packages, sources, the build script, environment variables for the build script, etc. Nix tries very hard to ensure that Nix expressions are deterministic: building a Nix expression twice should yield the same result.

Because it’s a functional language, it’s easy to support building variants of a package: turn the Nix expression into a function and call it any number of times with the appropriate arguments. Due to the hashing scheme, variants don’t conflict with each other in the Nix store.

Transparent source/binary deployment

Nix expressions generally describe how to build a package from source, so an installation action like

$ nix-env --install --attr nixpkgs.firefox

could cause quite a bit of build activity, as not only Firefox but also all its dependencies (all the way up to the C library and the compiler) would have to be built, at least if they are not already in the Nix store. This is a source deployment model. For most users, building from source is not very pleasant as it takes far too long. However, Nix can automatically skip building from source and instead use a binary cache, a web server that provides pre-built binaries. For instance, when asked to build /nix/store/b6gvzjyb2pg0…-firefox-33.1 from source, Nix would first check if the file https://cache.nixos.org/b6gvzjyb2pg0….narinfo exists, and if so, fetch the pre-built binary referenced from there; otherwise, it would fall back to building from source.

Nix Packages collection

We provide a large set of Nix expressions containing hundreds of existing Unix packages, the Nix Packages collection (Nixpkgs).

Managing build environments

Nix is extremely useful for developers as it makes it easy to automatically set up the build environment for a package. Given a Nix expression that describes the dependencies of your package, the command nix-shell will build or download those dependencies if they’re not already in your Nix store, and then start a Bash shell in which all necessary environment variables (such as compiler search paths) are set.

For example, the following command gets all dependencies of the Pan newsreader, as described by its Nix expression:

$ nix-shell '<nixpkgs>' --attr pan

You’re then dropped into a shell where you can edit, build and test the package:

[nix-shell]$ unpackPhase
[nix-shell]$ cd pan-*
[nix-shell]$ configurePhase
[nix-shell]$ buildPhase
[nix-shell]$ ./pan/gui/pan

Portability

Nix runs on Linux and macOS.

NixOS

NixOS is a Linux distribution based on Nix. It uses Nix not just for package management but also to manage the system configuration (e.g., to build configuration files in /etc). This means, among other things, that it is easy to roll back the entire configuration of the system to an earlier state. Also, users can install software without root privileges. For more information and downloads, see the NixOS homepage.

License

Nix is released under the terms of the GNU LGPLv2.1 or (at your option) any later version.

Quick Start

This chapter is for impatient people who don't like reading documentation. For more in-depth information you are kindly referred to subsequent chapters.

1. Install Nix:

   $ curl -L https://nixos.org/nix/install | sh
   

   The install script will use sudo, so make sure you have sufficient rights.

   For other installation methods, see the detailed installation instructions.

2. Run software without installing it permanently:

   $ nix-shell --packages cowsay lolcat
   

   This downloads the specified packages with all their dependencies, and drops you into a Bash shell where the commands provided by those packages are present. This will not affect your normal environment:

   [nix-shell:~]$ cowsay Hello, Nix! | lolcat
   

   Exiting the shell will make the programs disappear again:

   [nix-shell:~]$ exit
   $ lolcat
   lolcat: command not found
   

3. Search for more packages on search.nixos.org to try them out.

4. Free up storage space:

   $ nix-collect-garbage
   

Installation

This section describes how to install and configure Nix for first-time use.

The current recommended option on Linux and MacOS is multi-user.

Multi-user

This installation offers better sharing, improved isolation, and more security over a single user installation.

This option requires either:

• Linux running systemd, with SELinux disabled
• MacOS

$ bash <(curl -L https://nixos.org/nix/install) --daemon

Single-user

Single-user is not supported on Mac.

This installation has less requirements than the multi-user install, however it cannot offer equivalent sharing, isolation, or security.

This option is suitable for systems without systemd.

$ bash <(curl -L https://nixos.org/nix/install) --no-daemon

Distributions

The Nix community maintains installers for several distributions.

They can be found in the nix-community/nix-installers repository.

Supported Platforms

Nix is currently supported on the following platforms:

• Linux (i686, x86_64, aarch64).

• macOS (x86_64, aarch64).

Installing a Binary Distribution

To install the latest version Nix, run the following command:

$ curl -L https://nixos.org/nix/install | sh

This performs the default type of installation for your platform:

• Multi-user:
  • Linux with systemd and without SELinux
  • macOS
• Single-user:
  • Linux without systemd
  • Linux with SELinux

We recommend the multi-user installation if it supports your platform and you can authenticate with sudo.

The installer can configured with various command line arguments and environment variables. To show available command line flags:

$ curl -L https://nixos.org/nix/install | sh -s -- --help

To check what it does and how it can be customised further, download and edit the second-stage installation script.

Installing a pinned Nix version from a URL

Version-specific installation URLs for all Nix versions since 1.11.16 can be found at releases.nixos.org. The directory for each version contains the corresponding SHA-256 hash.

All installation scripts are invoked the same way:

$ export VERSION=2.19.2 
$ curl -L https://releases.nixos.org/nix/nix-$VERSION/install | sh

Multi User Installation

The multi-user Nix installation creates system users and a system service for the Nix daemon.

Supported systems:

• Linux running systemd, with SELinux disabled
• macOS

To explicitly instruct the installer to perform a multi-user installation on your system:

$ bash <(curl -L https://nixos.org/nix/install) --daemon

You can run this under your usual user account or root. The script will invoke sudo as needed.

Single User Installation

To explicitly select a single-user installation on your system:

$ bash <(curl -L https://nixos.org/nix/install) --no-daemon

In a single-user installation, /nix is owned by the invoking user. The script will invoke sudo to create /nix if it doesn’t already exist. If you don’t have sudo, manually create /nix as root:

$ su root
# mkdir /nix
# chown alice /nix

Installing from a binary tarball

You can also download a binary tarball that contains Nix and all its dependencies:

• Choose a version and system type
• Download and unpack the tarball
• Run the installer

Example

$ pushd $(mktemp -d)
$ export VERSION=2.19.2
$ export SYSTEM=x86_64-linux
$ curl -LO https://releases.nixos.org/nix/nix-$VERSION/nix-$VERSION-$SYSTEM.tar.xz
$ tar xfj nix-$VERSION-$SYSTEM.tar.xz
$ cd nix-$VERSION-$SYSTEM
$ ./install
$ popd

The installer can be customised with the environment variables declared in the file named install-multi-user.

Native packages for Linux distributions

The Nix community maintains installers for some Linux distributions in their native packaging format(https://nix-community.github.io/nix-installers/).

macOS Installation

We believe we have ironed out how to cleanly support the read-only root file system on modern macOS. New installs will do this automatically.

This section previously detailed the situation, options, and trade-offs, but it now only outlines what the installer does. You don't need to know this to run the installer, but it may help if you run into trouble:

• create a new APFS volume for your Nix store
• update /etc/synthetic.conf to direct macOS to create a "synthetic" empty root directory to mount your volume
• specify mount options for the volume in /etc/fstab
  • rw: read-write
  • noauto: prevent the system from auto-mounting the volume (so the LaunchDaemon mentioned below can control mounting it, and to avoid masking problems with that mounting service).
  • nobrowse: prevent the Nix Store volume from showing up on your desktop; also keeps Spotlight from spending resources to index this volume
• if you have FileVault enabled
  • generate an encryption password
  • put it in your system Keychain
  • use it to encrypt the volume
• create a system LaunchDaemon to mount this volume early enough in the boot process to avoid problems loading or restoring any programs that need access to your Nix store

Installing Nix from Source

If no binary package is available or if you want to hack on Nix, you can build Nix from its Git repository.

Prerequisites

• GNU Autoconf (https://www.gnu.org/software/autoconf/) and the autoconf-archive macro collection (https://www.gnu.org/software/autoconf-archive/). These are needed to run the bootstrap script.

• GNU Make.

• Bash Shell. The ./configure script relies on bashisms, so Bash is required.

• A version of GCC or Clang that supports C++20.

• pkg-config to locate dependencies. If your distribution does not provide it, you can get it from http://www.freedesktop.org/wiki/Software/pkg-config.

• The OpenSSL library to calculate cryptographic hashes. If your distribution does not provide it, you can get it from https://www.openssl.org.

• The libbrotlienc and libbrotlidec libraries to provide implementation of the Brotli compression algorithm. They are available for download from the official repository https://github.com/google/brotli.

• cURL and its library. If your distribution does not provide it, you can get it from https://curl.haxx.se/.

• The SQLite embedded database library, version 3.6.19 or higher. If your distribution does not provide it, please install it from http://www.sqlite.org/.

• The Boehm garbage collector (bdw-gc) to reduce the evaluator’s memory consumption (optional).

  To enable it, install pkgconfig and the Boehm garbage collector, and pass the flag --enable-gc to configure.

  For bdw-gc <= 8.2.4 Nix needs a small patch to be applied.

• The boost library of version 1.66.0 or higher. It can be obtained from the official web site https://www.boost.org/.

• The editline library of version 1.14.0 or higher. It can be obtained from the its repository https://github.com/troglobit/editline.

• The libsodium library for verifying cryptographic signatures of contents fetched from binary caches. It can be obtained from the official web site https://libsodium.org.

• Recent versions of Bison and Flex to build the parser. (This is because Nix needs GLR support in Bison and reentrancy support in Flex.) For Bison, you need version 2.6, which can be obtained from the GNU FTP server. For Flex, you need version 2.5.35, which is available on SourceForge. Slightly older versions may also work, but ancient versions like the ubiquitous 2.5.4a won't.

• The libseccomp is used to provide syscall filtering on Linux. This is an optional dependency and can be disabled passing a --disable-seccomp-sandboxing option to the configure script (Not recommended unless your system doesn't support libseccomp). To get the library, visit https://github.com/seccomp/libseccomp.

• On 64-bit x86 machines only, libcpuid library is used to determine which microarchitecture levels are supported (e.g., as whether to have x86_64-v2-linux among additional system types). The library is available from its homepage http://libcpuid.sourceforge.net. This is an optional dependency and can be disabled by providing a --disable-cpuid to the configure script.

• Unless ./configure --disable-unit-tests is specified, GoogleTest (GTest) and RapidCheck are required, which are available at https://google.github.io/googletest/ and https://github.com/emil-e/rapidcheck respectively.

Obtaining the Source

The most recent sources of Nix can be obtained from its Git repository. For example, the following command will check out the latest revision into a directory called nix:

$ git clone https://github.com/NixOS/nix

Likewise, specific releases can be obtained from the tags of the repository.

Building Nix from Source

After cloning Nix's Git repository, issue the following commands:

$ autoreconf -vfi
$ ./configure options...
$ make
$ make install

Nix requires GNU Make so you may need to invoke gmake instead.

The installation path can be specified by passing the --prefix=prefix to configure. The default installation directory is /usr/local. You can change this to any location you like. You must have write permission to the prefix path.

Nix keeps its store (the place where packages are stored) in /nix/store by default. This can be changed using --with-store-dir=path.

Warning

It is best not to change the Nix store from its default, since doing so makes it impossible to use pre-built binaries from the standard Nixpkgs channels — that is, all packages will need to be built from source.

Nix keeps state (such as its database and log files) in /nix/var by default. This can be changed using --localstatedir=path.

Using Nix within Docker

To run the latest stable release of Nix with Docker run the following command:

$ docker run -ti ghcr.io/nixos/nix
Unable to find image 'ghcr.io/nixos/nix:latest' locally
latest: Pulling from ghcr.io/nixos/nix
5843afab3874: Pull complete
b52bf13f109c: Pull complete
1e2415612aa3: Pull complete
Digest: sha256:27f6e7f60227e959ee7ece361f75d4844a40e1cc6878b6868fe30140420031ff
Status: Downloaded newer image for ghcr.io/nixos/nix:latest
35ca4ada6e96:/# nix --version
nix (Nix) 2.3.12
35ca4ada6e96:/# exit

What is included in Nix's Docker image?

The official Docker image is created using pkgs.dockerTools.buildLayeredImage (and not with Dockerfile as it is usual with Docker images). You can still base your custom Docker image on it as you would do with any other Docker image.

The Docker image is also not based on any other image and includes minimal set of runtime dependencies that are required to use Nix:

• pkgs.nix
• pkgs.bashInteractive
• pkgs.coreutils-full
• pkgs.gnutar
• pkgs.gzip
• pkgs.gnugrep
• pkgs.which
• pkgs.curl
• pkgs.less
• pkgs.wget
• pkgs.man
• pkgs.cacert.out
• pkgs.findutils

Docker image with the latest development version of Nix

To get the latest image that was built by Hydra run the following command:

$ curl -L https://hydra.nixos.org/job/nix/master/dockerImage.x86_64-linux/latest/download/1 | docker load
$ docker run -ti nix:2.5pre20211105

You can also build a Docker image from source yourself:

$ nix build ./\#hydraJobs.dockerImage.x86_64-linux
$ docker load -i ./result/image.tar.gz
$ docker run -ti nix:2.5pre20211105

Security

Nix has two basic security models. First, it can be used in “single-user mode”, which is similar to what most other package management tools do: there is a single user (typically root) who performs all package management operations. All other users can then use the installed packages, but they cannot perform package management operations themselves.

Alternatively, you can configure Nix in “multi-user mode”. In this model, all users can perform package management operations — for instance, every user can install software without requiring root privileges. Nix ensures that this is secure. For instance, it’s not possible for one user to overwrite a package used by another user with a Trojan horse.

Single-User Mode

In single-user mode, all Nix operations that access the database in prefix/var/nix/db or modify the Nix store in prefix/store must be performed under the user ID that owns those directories. This is typically root. (If you install from RPM packages, that’s in fact the default ownership.) However, on single-user machines, it is often convenient to chown those directories to your normal user account so that you don’t have to su to root all the time.

Multi-User Mode

To allow a Nix store to be shared safely among multiple users, it is important that users are not able to run builders that modify the Nix store or database in arbitrary ways, or that interfere with builds started by other users. If they could do so, they could install a Trojan horse in some package and compromise the accounts of other users.

To prevent this, the Nix store and database are owned by some privileged user (usually root) and builders are executed under special user accounts (usually named nixbld1, nixbld2, etc.). When a unprivileged user runs a Nix command, actions that operate on the Nix store (such as builds) are forwarded to a Nix daemon running under the owner of the Nix store/database that performs the operation.

Note

Multi-user mode has one important limitation: only root and a set of trusted users specified in nix.conf can specify arbitrary binary caches. So while unprivileged users may install packages from arbitrary Nix expressions, they may not get pre-built binaries.

Setting up the build users

The build users are the special UIDs under which builds are performed. They should all be members of the build users group nixbld. This group should have no other members. The build users should not be members of any other group. On Linux, you can create the group and users as follows:

$ groupadd -r nixbld
$ for n in $(seq 1 10); do useradd -c "Nix build user $n" \
    -d /var/empty -g nixbld -G nixbld -M -N -r -s "$(which nologin)" \
    nixbld$n; done

This creates 10 build users. There can never be more concurrent builds than the number of build users, so you may want to increase this if you expect to do many builds at the same time.

Running the daemon

The Nix daemon should be started as follows (as root):

$ nix-daemon

You’ll want to put that line somewhere in your system’s boot scripts.

To let unprivileged users use the daemon, they should set the NIX_REMOTE environment variable to daemon. So you should put a line like

export NIX_REMOTE=daemon

into the users’ login scripts.

Restricting access

To limit which users can perform Nix operations, you can use the permissions on the directory /nix/var/nix/daemon-socket. For instance, if you want to restrict the use of Nix to the members of a group called nix-users, do

$ chgrp nix-users /nix/var/nix/daemon-socket
$ chmod ug=rwx,o= /nix/var/nix/daemon-socket

This way, users who are not in the nix-users group cannot connect to the Unix domain socket /nix/var/nix/daemon-socket/socket, so they cannot perform Nix operations.

Environment Variables

To use Nix, some environment variables should be set. In particular, PATH should contain the directories prefix/bin and ~/.nix-profile/bin. The first directory contains the Nix tools themselves, while ~/.nix-profile is a symbolic link to the current user environment (an automatically generated package consisting of symlinks to installed packages). The simplest way to set the required environment variables is to include the file prefix/etc/profile.d/nix.sh in your ~/.profile (or similar), like this:

source prefix/etc/profile.d/nix.sh

NIX_SSL_CERT_FILE

If you need to specify a custom certificate bundle to account for an HTTPS-intercepting man in the middle proxy, you must specify the path to the certificate bundle in the environment variable NIX_SSL_CERT_FILE.

If you don't specify a NIX_SSL_CERT_FILE manually, Nix will install and use its own certificate bundle.

Set the environment variable and install Nix

$ export NIX_SSL_CERT_FILE=/etc/ssl/my-certificate-bundle.crt
$ curl -L https://nixos.org/nix/install | sh

In the shell profile and rc files (for example, /etc/bashrc, /etc/zshrc), add the following line:

export NIX_SSL_CERT_FILE=/etc/ssl/my-certificate-bundle.crt

Note

You must not add the export and then do the install, as the Nix installer will detect the presence of Nix configuration, and abort.

If you use the Nix daemon, you should also add the following to /etc/nix/nix.conf:

ssl-cert-file = /etc/ssl/my-certificate-bundle.crt

Proxy Environment Variables

The Nix installer has special handling for these proxy-related environment variables: http_proxy, https_proxy, ftp_proxy, all_proxy, no_proxy, HTTP_PROXY, HTTPS_PROXY, FTP_PROXY, ALL_PROXY, NO_PROXY.

If any of these variables are set when running the Nix installer, then the installer will create an override file at /etc/systemd/system/nix-daemon.service.d/override.conf so nix-daemon will use them.

Upgrading Nix

Note

These upgrade instructions apply where Nix was installed following the installation instructions in this manual.

Check which Nix version will be installed, for example from one of the release channels such as nixpkgs-unstable:

$ nix-shell -p nix -I nixpkgs=channel:nixpkgs-unstable --run "nix --version"
nix (Nix) 2.18.1

Warning

Writing to the local store with a newer version of Nix, for example by building derivations with nix-build or nix-store --realise, may change the database schema! Reverting to an older version of Nix may therefore require purging the store database before it can be used.

Linux multi-user

$ sudo su
# nix-env --install --file '<nixpkgs>' --attr nix cacert -I nixpkgs=channel:nixpkgs-unstable
# systemctl daemon-reload
# systemctl restart nix-daemon

macOS multi-user

$ sudo nix-env --install --file '<nixpkgs>' --attr nix cacert -I nixpkgs=channel:nixpkgs-unstable
$ sudo launchctl remove org.nixos.nix-daemon
$ sudo launchctl load /Library/LaunchDaemons/org.nixos.nix-daemon.plist

Single-user all platforms

$ nix-env --install --file '<nixpkgs>' --attr nix cacert -I nixpkgs=channel:nixpkgs-unstable

Uninstalling Nix

Multi User

Removing a multi-user installation depends on the operating system.

Linux

If you are on Linux with systemd:

1. Remove the Nix daemon service:

   sudo systemctl stop nix-daemon.service
   sudo systemctl disable nix-daemon.socket nix-daemon.service
   sudo systemctl daemon-reload
   

Remove files created by Nix:

sudo rm -rf /etc/nix /etc/profile.d/nix.sh /etc/tmpfiles.d/nix-daemon.conf /nix ~root/.nix-channels ~root/.nix-defexpr ~root/.nix-profile

Remove build users and their group:

for i in $(seq 1 32); do
  sudo userdel nixbld$i
done
sudo groupdel nixbld

There may also be references to Nix in

• /etc/bash.bashrc
• /etc/bashrc
• /etc/profile
• /etc/zsh/zshrc
• /etc/zshrc

which you may remove.

macOS

1. If system-wide shell initialisation files haven't been altered since installing Nix, use the backups made by the installer:

   sudo mv /etc/zshrc.backup-before-nix /etc/zshrc
   sudo mv /etc/bashrc.backup-before-nix /etc/bashrc
   sudo mv /etc/bash.bashrc.backup-before-nix /etc/bash.bashrc
   

   Otherwise, edit /etc/zshrc, /etc/bashrc, and /etc/bash.bashrc to remove the lines sourcing nix-daemon.sh, which should look like this:

   # Nix
   if [ -e '/nix/var/nix/profiles/default/etc/profile.d/nix-daemon.sh' ]; then
     . '/nix/var/nix/profiles/default/etc/profile.d/nix-daemon.sh'
   fi
   # End Nix
   

2. Stop and remove the Nix daemon services:

   sudo launchctl unload /Library/LaunchDaemons/org.nixos.nix-daemon.plist
   sudo rm /Library/LaunchDaemons/org.nixos.nix-daemon.plist
   sudo launchctl unload /Library/LaunchDaemons/org.nixos.darwin-store.plist
   sudo rm /Library/LaunchDaemons/org.nixos.darwin-store.plist
   

   This stops the Nix daemon and prevents it from being started next time you boot the system.

3. Remove the nixbld group and the _nixbuildN users:

   sudo dscl . -delete /Groups/nixbld
   for u in $(sudo dscl . -list /Users | grep _nixbld); do sudo dscl . -delete /Users/$u; done
   

   This will remove all the build users that no longer serve a purpose.

4. Edit fstab using sudo vifs to remove the line mounting the Nix Store volume on /nix, which looks like

   UUID=<uuid> /nix apfs rw,noauto,nobrowse,suid,owners
   

   or

   LABEL=Nix\040Store /nix apfs rw,nobrowse
   

   by setting the cursor on the respective line using the error keys, and pressing dd, and then :wq to save the file.

   This will prevent automatic mounting of the Nix Store volume.

5. Edit /etc/synthetic.conf to remove the nix line. If this is the only line in the file you can remove it entirely:

   if [ -f /etc/synthetic.conf ]; then
     if [ "$(cat /etc/synthetic.conf)" = "nix" ]; then
       sudo rm /etc/synthetic.conf
     else
       sudo vi /etc/synthetic.conf
     fi
   fi
   

   This will prevent the creation of the empty /nix directory.

6. Remove the files Nix added to your system, except for the store:

   sudo rm -rf /etc/nix /var/root/.nix-profile /var/root/.nix-defexpr /var/root/.nix-channels ~/.nix-profile ~/.nix-defexpr ~/.nix-channels
   

7. Remove the Nix Store volume:

   sudo diskutil apfs deleteVolume /nix
   

   This will remove the Nix Store volume and everything that was added to the store.

   If the output indicates that the command couldn't remove the volume, you should make sure you don't have an unmounted Nix Store volume. Look for a "Nix Store" volume in the output of the following command:

   diskutil list
   

   If you do find a "Nix Store" volume, delete it by running diskutil deleteVolume with the store volume's diskXsY identifier.

Note

After you complete the steps here, you will still have an empty /nix directory. This is an expected sign of a successful uninstall. The empty /nix directory will disappear the next time you reboot.

You do not have to reboot to finish uninstalling Nix. The uninstall is complete. macOS (Catalina+) directly controls root directories, and its read-only root will prevent you from manually deleting the empty /nix mountpoint.

Single User

To remove a single-user installation of Nix, run:

$ rm -rf /nix ~/.nix-channels ~/.nix-defexpr ~/.nix-profile

You might also want to manually remove references to Nix from your ~/.profile.

Nix Store

The Nix store is an abstraction to store immutable file system data (such as software packages) that can have dependencies on other such data.

There are multiple types of Nix stores with different capabilities, such as the default one on the local filesystem (/nix/store) or binary caches.

File System Object

Nix uses a simplified model of the file system, which consists of file system objects. Every file system object is one of the following:

• File

  • A possibly empty sequence of bytes for contents
  • A single boolean representing the executable permission

• Directory

  Mapping of names to child file system objects

• Symbolic link

  An arbitrary string. Nix does not assign any semantics to symbolic links.

File system objects and their children form a tree. A bare file or symlink can be a root file system object.

Nix does not encode any other file system notions such as hard links, permissions, timestamps, or other metadata.

Examples of file system objects

A plain file:

50 B, executable: false

An executable file:

122 KB, executable: true

A symlink:

-> /usr/bin/sh

A directory with contents:

├── bin
│   └── hello: 35 KB, executable: true
└── share
    ├── info
    │   └── hello.info: 36 KB, executable: false
    └── man
        └── man1
            └── hello.1.gz: 790 B, executable: false

A directory that contains a symlink and other directories:

├── bin -> share/go/bin
├── nix-support/
└── share/

Content-Addressing File System Objects

For many operations, Nix needs to calculate a content addresses of a file system object. Usually this is needed as part of content addressing store objects, since store objects always have a root file system object. But some command-line utilities also just work on "raw" file system objects, not part of any store object.

Every content addressing scheme Nix uses ultimately involves feeding data into a hash function, and getting back an opaque fixed-size digest which is deemed a content address. The various methods of content addressing thus differ in how abstract data (in this case, a file system object and its descendents) are fed into the hash function.

Serialising File System Objects

The simplest method is to serialise the entire file system object tree into a single binary string, and then hash that binary string, yielding the content address. In this section we describe the currently-supported methods of serialising file system objects.

Flat

A single file object can just be hashed by its contents. This is not enough information to encode the fact that the file system object is a file, but if we already know that the FSO is a single non-executable file by other means, it is sufficient.

Because the hashed data is just the raw file, as is, this choice is good for compatibility with other systems. For example, Unix commands like sha256sum or sha1sum will produce hashes for single files that match this.

Nix Archive (NAR)

For the other cases of file system objects, especially directories with arbitrary descendents, we need a more complex serialisation format. Examples of such serialisations are the ZIP and TAR file formats. However, for our purposes these formats have two problems:

• They do not have a canonical serialisation, meaning that given an FSO, there can be many different serialisations. For instance, TAR files can have variable amounts of padding between archive members; and some archive formats leave the order of directory entries undefined. This would be bad because we use serialisation to compute cryptographic hashes over file system objects, and for those hashes to be useful as a content address or for integrity checking, uniqueness is crucial. Otherwise, correct hashes would report false mismatches, and the store would fail to find the content.

• They store more information than we have in our notion of FSOs, such as time stamps. This can cause FSOs that Nix should consider equal to hash to different values on different machines, just because the dates differ.

• As a practical consideration, the TAR format is the only truly universal format in the Unix environment. It has many problems, such as an inability to deal with long file names and files larger than 2^33 bytes. Current implementations such as GNU Tar work around these limitations in various ways.

For these reasons, Nix has its very own archive format—the Nix Archive (NAR) format, which is carefully designed to avoid the problems described above.

The exact specification of the Nix Archive format is in protocols/nix-archive.md

Content addressing File System Objects beyond a single serialisation pass

Serialising the entire tree and then hashing that binary string is not the only option for content addressing, however. Another technique is that of a Merkle graph, where previously computed hashes are included in subsequent byte strings to be hashed.

In particular, the Merkle graphs can match the original graph structure of file system objects: we can first hash (serialised) child file system objects, and then hash parent objects using the hashes of their children in the serialisation (to be hashed) of the parent file system objects.

Currently, there is one such Merkle DAG content addressing method supported.

Git (experimental)

Warning

This method is part of the git-hashing experimental feature.

Git's file system model is very close to Nix's, and so Git's content addressing method is a pretty good fit. Just as with regular Git, files and symlinks are hashed as git "blobs", and directories are hashed as git "trees".

However, one difference between Nix's and Git's file system model needs special treatment. Plain files, executable files, and symlinks are not differentiated as distinctly addressable objects, but by their context: by the directory entry that refers to them. That means so long as the root object is a directory, there is no problem: every non-directory object is owned by a parent directory, and the entry that refers to it provides the missing information. However, if the root object is not a directory, then we have no way of knowing which one of an executable file, non-executable file, or symlink it is supposed to be.

In response to this, we have decided to treat a bare file as non-executable file. This is similar to do what we do with flat serialisation, which also lacks this information. To avoid an address collision, attempts to hash a bare executable file or symlink will result in an error (just as would happen for flat serialisation also). Thus, Git can encode some, but not all of Nix's "File System Objects", and this sort of content-addressing is likewise partial.

In the future, we may support a Git-like hash for such file system objects, or we may adopt another Merkle DAG format which is capable of representing all Nix file system objects.

Store Object

A Nix store is a collection of store objects with references between them. A store object consists of

• A file system object as data
• A set of store paths as references to other store objects

Store objects are immutable: Once created, they do not change until they are deleted.

Content-Addressing Store Objects

Just like File System Objects, Store Objects can also be content-addressed, unless they are input-addressed.

For store objects, the content address we produce will take the form of a Store Path rather than regular hash. In particular, the content-addressing scheme will ensure that the digest of the store path is solely computed from the

• file system object graph (the root one and its children, if it has any)
• references
• store directory
• name

of the store object, and not any other information, which would not be an intrinsic property of that store object.

For the full specification of the algorithms involved, see the specification of store path digests.

Content addressing each part of a store object

File System Objects

With all currently supported store object content addressing methods, the file system object is always content-addressed first, and then that hash is incorporated into content address computation for the store object.

References

With all currently supported store object content addressing methods, other objects are referred to by their regular (string-encoded-) store paths.

Self-references however cannot be referred to by their path, because we are in the midst of describing how to compute that path!

The alternative would require finding as hash function fixed point, i.e. the solution to an equation in the form

digest = hash(..... || digest || ....)

which is computationally infeasible. As far as we know, this is equivalent to finding a hash collision.

Instead we just have a "has self reference" boolean, which will end up affecting the digest.

Name and Store Directory

These two items affect the digest in a way that is standard for store path digest computations and not specific to content-addressing. Consult the specification of store path digests for further details.

Content addressing Methods

For historical reasons, we don't support all features in all combinations. Each currently supported method of content addressing chooses a single method of file system object hashing, and may offer some restrictions on references. The names and store directories are unrestricted however.

Flat

This uses the corresponding Flat method of file system object content addressing.

References are not supported: store objects with flat hashing and references can not be created.

Text

This also uses the corresponding Flat method of file system object content addressing.

References to other store objects are supported, but self references are not.

This is the only store-object content-addressing method that is not named identically with a corresponding file system object method. It is somewhat obscure, mainly used for "drv files" (derivations serialized as store objects in their "ATerm" file format). Prefer another method if possible.

Nix Archive

This uses the corresponding Nix Archive method of file system object content addressing.

References (to other store objects and self references alike) are supported so long as the hash algorithm is SHA-256, but not (neither kind) otherwise.

Git

Warning

This method is part of the git-hashing experimental feature.

This uses the corresponding Git method of file system object content addressing.

References are not supported.

Only SHA-1 is supported at this time. If SHA-256-based Git becomes more widespread, this restriction will be revisited.

Store Path

Example

/nix/store/a040m110amc4h71lds2jmr8qrkj2jhxd-git-2.38.1

A rendered store path

Nix implements references to store objects as store paths.

Think of a store path as an opaque, unique identifier: The only way to obtain store path is by adding or building store objects. A store path will always reference exactly one store object.

Store paths are pairs of

• A 20-byte digest for identification
• A symbolic name for people to read

Example

• Digest: b6gvzjyb2pg0kjfwrjmg1vfhh54ad73z
• Name: firefox-33.1

To make store objects accessible to operating system processes, stores have to expose store objects through the file system.

A store path is rendered to a file system path as the concatenation of

• Store directory (typically /nix/store)
• Path separator (/)
• Digest rendered in a custom variant of Base32 (20 arbitrary bytes become 32 ASCII characters)
• Hyphen (-)
• Name

Example

  /nix/store/b6gvzjyb2pg0kjfwrjmg1vfhh54ad73z-firefox-33.1
  |--------| |------------------------------| |----------|
store directory            digest                 name

Exactly how the digest is calculated depends on the type of store path. Store path digests are supposed to be opaque, and so for most operations, it is not necessary to know the details. That said, the manual has a full specification of store path digests.

Store Directory

Every Nix store has a store directory.

Not every store can be accessed through the file system. But if the store has a file system representation, the store directory contains the store’s file system objects, which can be addressed by store paths.

This means a store path is not just derived from the referenced store object itself, but depends on the store that the store object is in.

Note

The store directory defaults to /nix/store, but is in principle arbitrary.

It is important which store a given store object belongs to: Files in the store object can contain store paths, and processes may read these paths. Nix can only guarantee referential integrity if store paths do not cross store boundaries.

Therefore one can only copy store objects to a different store if

• The source and target stores' directories match

  or

• The store object in question has no references, that is, contains no store paths

One cannot copy a store object to a store with a different store directory. Instead, it has to be rebuilt, together with all its dependencies. It is in general not enough to replace the store directory string in file contents, as this may render executables unusable by invalidating their internal offsets or checksums.

Nix supports different types of stores:

• Dummy Store
• Experimental Local Overlay Store
• Experimental SSH Store
• Experimental SSH Store with filesystem mounted
• HTTP Binary Cache Store
• Local Binary Cache Store
• Local Daemon Store
• Local Store
• S3 Binary Cache Store
• SSH Store

Store URL format

Stores are specified using a URL-like syntax. For example, the command

# nix path-info --store https://cache.nixos.org/ --json \
  /nix/store/a7gvj343m05j2s32xcnwr35v31ynlypr-coreutils-9.1

fetches information about a store path in the HTTP binary cache located at https://cache.nixos.org/, which is a type of store.

Store URLs can specify store settings using URL query strings, i.e. by appending ?name1=value1&name2=value2&... to the URL. For instance,

--store ssh://machine.example.org?ssh-key=/path/to/my/key

tells Nix to access the store on a remote machine via the SSH protocol, using /path/to/my/key as the SSH private key. The supported settings for each store type are documented below.

The special store URL auto causes Nix to automatically select a store as follows:

• Use the local store /nix/store if /nix/var/nix is writable by the current user.

• Otherwise, if /nix/var/nix/daemon-socket/socket exists, connect to the Nix daemon listening on that socket.

• Otherwise, on Linux only, use the local chroot store ~/.local/share/nix/root, which will be created automatically if it does not exist.

• Otherwise, use the local store /nix/store.

Dummy Store

Store URL format: dummy://

This store type represents a store that contains no store paths and cannot be written to. It's useful when you want to use the Nix evaluator when no actual Nix store exists, e.g.

# nix eval --store dummy:// --expr '1 + 2'

Settings

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

Experimental Local Overlay Store

Warning

This store is part of an experimental feature.

To use this store, make sure the local-overlay-store experimental feature is enabled. For example, include the following in nix.conf:

extra-experimental-features = local-overlay-store

Store URL format: local-overlay

This store type is a variation of the [local store] designed to leverage Linux's Overlay Filesystem (OverlayFS for short). Just as OverlayFS combines a lower and upper filesystem by treating the upper one as a patch against the lower, the local overlay store combines a lower store with an upper almost-[local store]. ("almost" because while the upper fileystems for OverlayFS is valid on its own, the upper almost-store is not a valid local store on its own because some references will dangle.) To use this store, you will first need to configure an OverlayFS mountpoint appropriately as Nix will not do this for you (though it will verify the mountpoint is configured correctly).

Conceptual parts of a local overlay store

This is a more abstract/conceptual description of the parts of a layered store, an authoritative reference. For more "practical" instructions, see the worked-out example in the next subsection.

The parts of a local overlay store are as follows:

• Lower store:

  Specified with the lower-store setting.

  This is any store implementation that includes a store directory as part of the native operating system filesystem. For example, this could be a [local store], [local daemon store], or even another local overlay store.

  The local overlay store never tries to modify the lower store in any way. Something else could modify the lower store, but there are restrictions on this Nix itself requires that this store only grow, and not change in other ways. For example, new store objects can be added, but deleting or modifying store objects is not allowed in general, because that will confuse and corrupt any local overlay store using those objects. (In addition, the underlying filesystem overlay mechanism may impose additional restrictions, see below.)

  The lower store must not change while it is mounted as part of an overlay store. To ensure it does not, you might want to mount the store directory read-only (which then requires the [read-only] parameter to be set to true).

  • Lower store directory:

    Specified with lower-store.real setting.

    This is the directory used/exposed by the lower store.

    As specified above, Nix requires the local store can only grow not change in other ways. Linux's OverlayFS in addition imposes the further requirement that this directory cannot change at all. That means that, while any local overlay store exists that is using this store as a lower store, this directory must not change.

  • Lower metadata source:

    Not directly specified. A consequence of the lower-store setting, depending on the type of lower store chosen.

    This is abstract, just some way to read the metadata of lower store store objects. For example it could be a SQLite database (for the [local store]), or a socket connection (for the [local daemon store]).

    This need not be writable. As stated above a local overlay store never tries to modify its lower store. The lower store's metadata is considered part of the lower store, just as the store's file system objects that appear in the store directory are.

• Upper almost-store:

  Not directly specified. Instead the constituent parts are independently specified as described below.

  This is almost but not quite just a [local store]. That is because taken in isolation, not as part of a local overlay store, by itself, it would appear corrupted. But combined with everything else as part of an overlay local store, it is valid.

  • Upper layer directory:

    Specified with upper-layer setting.

    This contains additional store objects (or, strictly speaking, their file system objects that the local overlay store will extend the lower store with).

  • Upper store directory:

    Specified with the real setting. This the same as the base local store setting, and can also be indirectly specified with the root setting.

    This contains all the store objects from each of the two directories.

    The lower store directory and upper layer directory are combined via OverlayFS to create this directory. Nix doesn't do this itself, because it typically wouldn't have the permissions to do so, so it is the responsibility of the user to set this up first. Nix can, however, optionally check that that the OverlayFS mount settings appear as expected, matching Nix's own settings.

  • Upper SQLite database:

    Not directly specified. The location of the database instead depends on the state setting. It is is always ${state}/db.

    This contains the metadata of all of the upper layer store objects (everything beyond their file system objects), and also duplicate copies of some lower layer store object's metadta. The duplication is so the metadata for the closure of upper layer store objects can be found entirely within the upper layer. (This allows us to use the same SQL Schema as the [local store]'s SQLite database, as foreign keys in that schema enforce closure metadata to be self-contained in this way.)

Example filesystem layout

Here is a worked out example of usage, following the concepts in the previous section.

Say we have the following paths:

• /mnt/example/merged-store/nix/store

• /mnt/example/store-a/nix/store

• /mnt/example/store-b

Then the following store URI can be used to access a local-overlay store at /mnt/example/merged-store:

local-overlay://?root=/mnt/example/merged-store&lower-store=/mnt/example/store-a&upper-layer=/mnt/example/store-b

The lower store directory is located at /mnt/example/store-a/nix/store, while the upper layer is at /mnt/example/store-b.

Before accessing the overlay store you will need to ensure the OverlayFS mount is set up correctly:

mount -t overlay overlay \
  -o lowerdir="/mnt/example/store-a/nix/store" \
  -o upperdir="/mnt/example/store-b" \
  -o workdir="/mnt/example/workdir" \
  "/mnt/example/merged-store/nix/store"

Note that OverlayFS requires /mnt/example/workdir to be on the same volume as the upperdir.

By default, Nix will check that the mountpoint as been set up correctly and fail with an error if it has not. You can override this behaviour by passing check-mount=false if you need to.

Settings

• check-mount

  Check that the overlay filesystem is correctly mounted.

  Nix does not manage the overlayfs mount point itself, but the correct functioning of the overlay store does depend on this mount point being set up correctly. Rather than just assume this is the case, check that the lowerdir and upperdir options are what we expect them to be. This check is on by default, but can be disabled if needed.

  Default: true

• log

  directory where Nix will store log files.

  Default: /nix/var/log/nix

• lower-store

  Store URL for the lower store. The default is auto (i.e. use the Nix daemon or /nix/store directly).

  Must be a store with a store dir on the file system. Must be used as OverlayFS lower layer for this store's store dir.

  Default: empty

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• read-only

  Allow this store to be opened when its database is on a read-only filesystem.

  Normally Nix will attempt to open the store database in read-write mode, even for querying (when write access is not needed), causing it to fail if the database is on a read-only filesystem.

  Enable read-only mode to disable locking and open the SQLite database with the immutable parameter set.

  Warning Do not use this unless the filesystem is read-only.

  Using it when the filesystem is writable can cause incorrect query results or corruption errors if the database is changed by another process. While the filesystem the database resides on might appear to be read-only, consider whether another user or system might have write access to it.

  Default: false

• real

  Physical path of the Nix store.

  Default: /nix/store

• remount-hook

  Script or other executable to run when overlay filesystem needs remounting.

  This is occasionally necessary when deleting a store path that exists in both upper and lower layers. In such a situation, bypassing OverlayFS and deleting the path in the upper layer directly is the only way to perform the deletion without creating a "whiteout". However this causes the OverlayFS kernel data structures to get out-of-sync, and can lead to 'stale file handle' errors; remounting solves the problem.

  The store directory is passed as an argument to the invoked executable.

  Default: empty

• require-sigs

  Whether store paths copied into this store should have a trusted signature.

  Default: true

• root

  Directory prefixed to all other paths.

  Default: ``

• state

  Directory where Nix will store state.

  Default: /dummy

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• upper-layer

  Directory containing the OverlayFS upper layer for this store's store dir.

  Default: empty

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

Experimental SSH Store

Store URL format: ssh-ng://[username@]hostname

Experimental store type that allows full access to a Nix store on a remote machine.

Settings

• base64-ssh-public-host-key

  The public host key of the remote machine.

  Default: empty

• compress

  Whether to enable SSH compression.

  Default: false

• max-connection-age

  Maximum age of a connection before it is closed.

  Default: 4294967295

• max-connections

  Maximum number of concurrent connections to the Nix daemon.

  Default: 1

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• remote-program

  Path to the nix-daemon executable on the remote machine.

  Default: nix-daemon

• remote-store

  Store URL to be used on the remote machine. The default is auto (i.e. use the Nix daemon or /nix/store directly).

  Default: empty

• ssh-key

  Path to the SSH private key used to authenticate to the remote machine.

  Default: empty

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

Experimental SSH Store with filesystem mounted

Warning

This store is part of an experimental feature.

To use this store, make sure the mounted-ssh-store experimental feature is enabled. For example, include the following in nix.conf:

extra-experimental-features = mounted-ssh-store

Store URL format: mounted-ssh-ng://[username@]hostname

Experimental store type that allows full access to a Nix store on a remote machine, and additionally requires that store be mounted in the local file system.

The mounting of that store is not managed by Nix, and must by managed manually. It could be accomplished with SSHFS or NFS, for example.

The local file system is used to optimize certain operations. For example, rather than serializing Nix archives and sending over the Nix channel, we can directly access the file system data via the mount-point.

The local file system is also used to make certain operations possible that wouldn't otherwise be. For example, persistent GC roots can be created if they reside on the same file system as the remote store: the remote side will create the symlinks necessary to avoid race conditions.

Settings

• base64-ssh-public-host-key

  The public host key of the remote machine.

  Default: empty

• compress

  Whether to enable SSH compression.

  Default: false

• log

  directory where Nix will store log files.

  Default: /nix/var/log/nix

• max-connection-age

  Maximum age of a connection before it is closed.

  Default: 4294967295

• max-connections

  Maximum number of concurrent connections to the Nix daemon.

  Default: 1

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• real

  Physical path of the Nix store.

  Default: /nix/store

• remote-program

  Path to the nix-daemon executable on the remote machine.

  Default: nix-daemon

• remote-store

  Store URL to be used on the remote machine. The default is auto (i.e. use the Nix daemon or /nix/store directly).

  Default: empty

• root

  Directory prefixed to all other paths.

  Default: ``

• ssh-key

  Path to the SSH private key used to authenticate to the remote machine.

  Default: empty

• state

  Directory where Nix will store state.

  Default: /dummy

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

HTTP Binary Cache Store

Store URL format: http://..., https://...

This store allows a binary cache to be accessed via the HTTP protocol.

Settings

• compression

  NAR compression method (xz, bzip2, gzip, zstd, or none).

  Default: xz

• compression-level

  The preset level to be used when compressing NARs. The meaning and accepted values depend on the compression method selected. -1 specifies that the default compression level should be used.

  Default: -1

• index-debug-info

  Whether to index DWARF debug info files by build ID. This allows dwarffs to fetch debug info on demand

  Default: false

• local-nar-cache

  Path to a local cache of NARs fetched from this binary cache, used by commands such as nix store cat.

  Default: empty

• parallel-compression

  Enable multi-threaded compression of NARs. This is currently only available for xz and zstd.

  Default: false

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• secret-key

  Path to the secret key used to sign the binary cache.

  Default: empty

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

• write-nar-listing

  Whether to write a JSON file that lists the files in each NAR.

  Default: false

Local Binary Cache Store

Store URL format: file://path

This store allows reading and writing a binary cache stored in path in the local filesystem. If path does not exist, it will be created.

For example, the following builds or downloads nixpkgs#hello into the local store and then copies it to the binary cache in /tmp/binary-cache:

# nix copy --to file:///tmp/binary-cache nixpkgs#hello

Settings

• compression

  NAR compression method (xz, bzip2, gzip, zstd, or none).

  Default: xz

• compression-level

  The preset level to be used when compressing NARs. The meaning and accepted values depend on the compression method selected. -1 specifies that the default compression level should be used.

  Default: -1

• index-debug-info

  Whether to index DWARF debug info files by build ID. This allows dwarffs to fetch debug info on demand

  Default: false

• local-nar-cache

  Path to a local cache of NARs fetched from this binary cache, used by commands such as nix store cat.

  Default: empty

• parallel-compression

  Enable multi-threaded compression of NARs. This is currently only available for xz and zstd.

  Default: false

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• secret-key

  Path to the secret key used to sign the binary cache.

  Default: empty

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

• write-nar-listing

  Whether to write a JSON file that lists the files in each NAR.

  Default: false

Local Daemon Store

Store URL format: daemon, unix://path

This store type accesses a Nix store by talking to a Nix daemon listening on the Unix domain socket path. The store pseudo-URL daemon is equivalent to unix:///nix/var/nix/daemon-socket/socket.

Settings

• log

  directory where Nix will store log files.

  Default: /nix/var/log/nix

• max-connection-age

  Maximum age of a connection before it is closed.

  Default: 4294967295

• max-connections

  Maximum number of concurrent connections to the Nix daemon.

  Default: 1

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• real

  Physical path of the Nix store.

  Default: /nix/store

• root

  Directory prefixed to all other paths.

  Default: ``

• state

  Directory where Nix will store state.

  Default: /dummy

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

Local Store

Store URL format: local, root

This store type accesses a Nix store in the local filesystem directly (i.e. not via the Nix daemon). root is an absolute path that is prefixed to other directories such as the Nix store directory. The store pseudo-URL local denotes a store that uses / as its root directory.

A store that uses a root other than / is called a chroot store. With such stores, the store directory is "logically" still /nix/store, so programs stored in them can only be built and executed by chroot-ing into root. Chroot stores only support building and running on Linux when mount namespaces and user namespaces are enabled.

For example, the following uses /tmp/root as the chroot environment to build or download nixpkgs#hello and then execute it:

# nix run --store /tmp/root nixpkgs#hello
Hello, world!

Here, the "physical" store location is /tmp/root/nix/store, and Nix's store metadata is in /tmp/root/nix/var/nix/db.

It is also possible, but not recommended, to change the "logical" location of the Nix store from its default of /nix/store. This makes it impossible to use default substituters such as https://cache.nixos.org/, and thus you may have to build everything locally. Here is an example:

# nix build --store 'local?store=/tmp/my-nix/store&state=/tmp/my-nix/state&log=/tmp/my-nix/log' nixpkgs#hello

Settings

• log

  directory where Nix will store log files.

  Default: /nix/var/log/nix

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• read-only

  Allow this store to be opened when its database is on a read-only filesystem.

  Normally Nix will attempt to open the store database in read-write mode, even for querying (when write access is not needed), causing it to fail if the database is on a read-only filesystem.

  Enable read-only mode to disable locking and open the SQLite database with the immutable parameter set.

  Warning Do not use this unless the filesystem is read-only.

  Using it when the filesystem is writable can cause incorrect query results or corruption errors if the database is changed by another process. While the filesystem the database resides on might appear to be read-only, consider whether another user or system might have write access to it.

  Default: false

• real

  Physical path of the Nix store.

  Default: /nix/store

• require-sigs

  Whether store paths copied into this store should have a trusted signature.

  Default: true

• root

  Directory prefixed to all other paths.

  Default: ``

• state

  Directory where Nix will store state.

  Default: /dummy

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

S3 Binary Cache Store

Store URL format: s3://bucket-name

This store allows reading and writing a binary cache stored in an AWS S3 (or S3-compatible service) bucket. This store shares many idioms with the HTTP Binary Cache Store.

For AWS S3, the binary cache URL for a bucket named example-nix-cache will be exactly s3://example-nix-cache. For S3 compatible binary caches, consult that cache's documentation.

Anonymous reads to your S3-compatible binary cache

If your binary cache is publicly accessible and does not require authentication, it is simplest to use the [HTTP Binary Cache Store] rather than S3 Binary Cache Store with https://example-nix-cache.s3.amazonaws.com instead of s3://example-nix-cache.

Your bucket will need a bucket policy like the following to be accessible:

{
    "Id": "DirectReads",
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "AllowDirectReads",
            "Action": [
                "s3:GetObject",
                "s3:GetBucketLocation"
            ],
            "Effect": "Allow",
            "Resource": [
                "arn:aws:s3:::example-nix-cache",
                "arn:aws:s3:::example-nix-cache/*"
            ],
            "Principal": "*"
        }
    ]
}

Authentication

Nix will use the default credential provider chain for authenticating requests to Amazon S3.

Note that this means Nix will read environment variables and files with different idioms than with Nix's own settings, as implemented by the AWS SDK. Consult the documentation linked above for further details.

Authenticated reads to your S3 binary cache

Your bucket will need a bucket policy allowing the desired users to perform the s3:GetObject and s3:GetBucketLocation action on all objects in the bucket. The anonymous policy given above can be updated to have a restricted Principal to support this.

Authenticated writes to your S3-compatible binary cache

Your account will need an IAM policy to support uploading to the bucket:

{
  "Version": "2012-10-17",
  "Statement": [
    {
      "Sid": "UploadToCache",
      "Effect": "Allow",
      "Action": [
        "s3:AbortMultipartUpload",
        "s3:GetBucketLocation",
        "s3:GetObject",
        "s3:ListBucket",
        "s3:ListBucketMultipartUploads",
        "s3:ListMultipartUploadParts",
        "s3:PutObject"
      ],
      "Resource": [
        "arn:aws:s3:::example-nix-cache",
        "arn:aws:s3:::example-nix-cache/*"
      ]
    }
  ]
}

Examples

With bucket policies and authentication set up as described above, uploading works via nix copy (experimental).

• To upload with a specific credential profile for Amazon S3:

  $ nix copy nixpkgs.hello \
    --to 's3://example-nix-cache?profile=cache-upload&region=eu-west-2'
  

• To upload to an S3-compatible binary cache:

  $ nix copy nixpkgs.hello --to \
    's3://example-nix-cache?profile=cache-upload&scheme=https&endpoint=minio.example.com'
  

Settings

• buffer-size

  Size (in bytes) of each part in multi-part uploads.

  Default: 5242880

• compression

  NAR compression method (xz, bzip2, gzip, zstd, or none).

  Default: xz

• compression-level

  The preset level to be used when compressing NARs. The meaning and accepted values depend on the compression method selected. -1 specifies that the default compression level should be used.

  Default: -1

• endpoint

  The URL of the endpoint of an S3-compatible service such as MinIO. Do not specify this setting if you're using Amazon S3.

  Note

  This endpoint must support HTTPS and will use path-based addressing instead of virtual host based addressing.

  Default: empty

• index-debug-info

  Whether to index DWARF debug info files by build ID. This allows dwarffs to fetch debug info on demand

  Default: false

• local-nar-cache

  Path to a local cache of NARs fetched from this binary cache, used by commands such as nix store cat.

  Default: empty

• log-compression

  Compression method for log/* files. It is recommended to use a compression method supported by most web browsers (e.g. brotli).

  Default: empty

• ls-compression

  Compression method for .ls files.

  Default: empty

• multipart-upload

  Whether to use multi-part uploads.

  Default: false

• narinfo-compression

  Compression method for .narinfo files.

  Default: empty

• parallel-compression

  Enable multi-threaded compression of NARs. This is currently only available for xz and zstd.

  Default: false

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• profile

  The name of the AWS configuration profile to use. By default Nix will use the default profile.

  Default: empty

• region

  The region of the S3 bucket. If your bucket is not in us–east-1, you should always explicitly specify the region parameter.

  Default: us-east-1

• scheme

  The scheme used for S3 requests, https (default) or http. This option allows you to disable HTTPS for binary caches which don't support it.

  Note

  HTTPS should be used if the cache might contain sensitive information.

  Default: empty

• secret-key

  Path to the secret key used to sign the binary cache.

  Default: empty

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

• write-nar-listing

  Whether to write a JSON file that lists the files in each NAR.

  Default: false

SSH Store

Store URL format: ssh://[username@]hostname

This store type allows limited access to a remote store on another machine via SSH.

Settings

• base64-ssh-public-host-key

  The public host key of the remote machine.

  Default: empty

• compress

  Whether to enable SSH compression.

  Default: false

• max-connections

  Maximum number of concurrent SSH connections.

  Default: 1

• path-info-cache-size

  Size of the in-memory store path metadata cache.

  Default: 65536

• priority

  Priority of this store when used as a substituter. A lower value means a higher priority.

  Default: 0

• remote-program

  Path to the nix-store executable on the remote machine.

  Default: nix-store

• remote-store

  Store URL to be used on the remote machine. The default is auto (i.e. use the Nix daemon or /nix/store directly).

  Default: empty

• ssh-key

  Path to the SSH private key used to authenticate to the remote machine.

  Default: empty

• store

  Logical location of the Nix store, usually /nix/store. Note that you can only copy store paths between stores if they have the same store setting.

  Default: /nix/store

• system-features

  Optional system features available on the system this store uses to build derivations.

  Example: "kvm"

  Default: machine-specific

• trusted

  Whether paths from this store can be used as substitutes even if they are not signed by a key listed in the trusted-public-keys setting.

  Default: false

• want-mass-query

  Whether this store can be queried efficiently for path validity when used as a substituter.

  Default: false

Nix Language

The Nix language is designed for conveniently creating and composing derivations – precise descriptions of how contents of existing files are used to derive new files.

Tip

These pages are written as a reference. If you are learning Nix, nix.dev has a good introduction to the Nix language.

The language is:

• domain-specific

  It comes with built-in functions to integrate with the Nix store, which manages files and performs the derivations declared in the Nix language.

• declarative

  There is no notion of executing sequential steps. Dependencies between operations are established only through data.

• pure

  Values cannot change during computation. Functions always produce the same output if their input does not change.

• functional

  Functions are like any other value. Functions can be assigned to names, taken as arguments, or returned by functions.

• lazy

  Values are only computed when they are needed.

• dynamically typed

  Type errors are only detected when expressions are evaluated.

Overview

This is an incomplete overview of language features, by example.

''
  multi
   line
    string
''

Data Types

Primitives

• String

  Strings can be written in three ways.

  The most common way is to enclose the string between double quotes, e.g., "foo bar". Strings can span multiple lines. The special characters " and \ and the character sequence ${ must be escaped by prefixing them with a backslash (\). Newlines, carriage returns and tabs can be written as \n, \r and \t, respectively.

  You can include the results of other expressions into a string by enclosing them in ${ }, a feature known as string interpolation.

  The second way to write string literals is as an indented string, which is enclosed between pairs of double single-quotes, like so:

  ''
    This is the first line.
    This is the second line.
      This is the third line.
  ''
  

  This kind of string literal intelligently strips indentation from the start of each line. To be precise, it strips from each line a number of spaces equal to the minimal indentation of the string as a whole (disregarding the indentation of empty lines). For instance, the first and second line are indented two spaces, while the third line is indented four spaces. Thus, two spaces are stripped from each line, so the resulting string is

  "This is the first line.\nThis is the second line.\n  This is the third line.\n"
  

  Note that the whitespace and newline following the opening '' is ignored if there is no non-whitespace text on the initial line.

  Indented strings support string interpolation.

  Since ${ and '' have special meaning in indented strings, you need a way to quote them. $ can be escaped by prefixing it with '' (that is, two single quotes), i.e., ''$. '' can be escaped by prefixing it with ', i.e., '''. $ removes any special meaning from the following $. Linefeed, carriage-return and tab characters can be written as ''\n, ''\r, ''\t, and ''\ escapes any other character.

  Indented strings are primarily useful in that they allow multi-line string literals to follow the indentation of the enclosing Nix expression, and that less escaping is typically necessary for strings representing languages such as shell scripts and configuration files because '' is much less common than ". Example:

  stdenv.mkDerivation {
    ...
    postInstall =
      ''
        mkdir $out/bin $out/etc
        cp foo $out/bin
        echo "Hello World" > $out/etc/foo.conf
        ${if enableBar then "cp bar $out/bin" else ""}
      '';
    ...
  }
  

  Finally, as a convenience, URIs as defined in appendix B of RFC 2396 can be written as is, without quotes. For instance, the string "http://example.org/foo.tar.bz2" can also be written as http://example.org/foo.tar.bz2.

• Number

  Numbers, which can be integers (like 123) or floating point (like 123.43 or .27e13).

  See arithmetic and comparison operators for semantics.

• Path

  Paths are distinct from strings and can be expressed by path literals such as ./builder.sh.

  Paths are suitable for referring to local files, and are often preferable over strings.

  • Path values do not contain trailing slashes, . and .., as they are resolved when evaluating a path literal.
  • Path literals are automatically resolved relative to their base directory.
  • The files referred to by path values are automatically copied into the Nix store when used in a string interpolation or concatenation.
  • Tooling can recognize path literals and provide additional features, such as autocompletion, refactoring automation and jump-to-file.

  A path literal must contain at least one slash to be recognised as such. For instance, builder.sh is not a path: it's parsed as an expression that selects the attribute sh from the variable builder.

  Path literals may also refer to absolute paths by starting with a slash.

  Note

  Absolute paths make expressions less portable. In the case where a function translates a path literal into an absolute path string for a configuration file, it is recommended to write a string literal instead. This avoids some confusion about whether files at that location will be used during evaluation. It also avoids unintentional situations where some function might try to copy everything at the location into the store.

  If the first component of a path is a ~, it is interpreted such that the rest of the path were relative to the user's home directory. For example, ~/foo would be equivalent to /home/edolstra/foo for a user whose home directory is /home/edolstra. Path literals that start with ~ are not allowed in pure evaluation.

  Paths can be used in string interpolation and string concatenation. For instance, evaluating "${./foo.txt}" will cause foo.txt from the same directory to be copied into the Nix store and result in the string "/nix/store/<hash>-foo.txt".

  Note that the Nix language assumes that all input files will remain unchanged while evaluating a Nix expression. For example, assume you used a file path in an interpolated string during a nix repl session. Later in the same session, after having changed the file contents, evaluating the interpolated string with the file path again might not return a new store path, since Nix might not re-read the file contents. Use :r to reset the repl as needed.

  Path literals can also include string interpolation, besides being interpolated into other expressions.

  At least one slash (/) must appear before any interpolated expression for the result to be recognized as a path.

  a.${foo}/b.${bar} is a syntactically valid number division operation. ./a.${foo}/b.${bar} is a path.

  Lookup path literals such as <nixpkgs> also resolve to path values.

• Boolean

  Booleans with values true and false.

• Null

  The null value, denoted as null.

List

Lists are formed by enclosing a whitespace-separated list of values between square brackets. For example,

[ 123 ./foo.nix "abc" (f { x = y; }) ]

defines a list of four elements, the last being the result of a call to the function f. Note that function calls have to be enclosed in parentheses. If they had been omitted, e.g.,

[ 123 ./foo.nix "abc" f { x = y; } ]

the result would be a list of five elements, the fourth one being a function and the fifth being a set.

Note that lists are only lazy in values, and they are strict in length.

Elements in a list can be accessed using builtins.elemAt.

Attribute Set

An attribute set is a collection of name-value-pairs (called attributes) enclosed in curly brackets ({ }).

An attribute name can be an identifier or a string. An identifier must start with a letter (a-z, A-Z) or underscore (_), and can otherwise contain letters (a-z, A-Z), numbers (0-9), underscores (_), apostrophes ('), or dashes (-).

Syntax

name = identifier | string
identifier ~ [a-zA-Z_][a-zA-Z0-9_'-]*

Names and values are separated by an equal sign (=). Each value is an arbitrary expression terminated by a semicolon (;).

Syntax

attrset = { [ name = expr ; ]... }

Attributes can appear in any order. An attribute name may only occur once.

Example:

{
  x = 123;
  text = "Hello";
  y = f { bla = 456; };
}

This defines a set with attributes named x, text, y.

Attributes can be accessed with the . operator.

Example:

{ a = "Foo"; b = "Bar"; }.a

This evaluates to "Foo".

It is possible to provide a default value in an attribute selection using the or keyword.

Example:

{ a = "Foo"; b = "Bar"; }.c or "Xyzzy"

{ a = "Foo"; b = "Bar"; }.c.d.e.f.g or "Xyzzy"

will both evaluate to "Xyzzy" because there is no c attribute in the set.

You can use arbitrary double-quoted strings as attribute names:

{ "$!@#?" = 123; }."$!@#?"

let bar = "bar"; in
{ "foo ${bar}" = 123; }."foo ${bar}"

Both will evaluate to 123.

Attribute names support string interpolation:

let bar = "foo"; in
{ foo = 123; }.${bar}

let bar = "foo"; in
{ ${bar} = 123; }.foo

Both will evaluate to 123.

In the special case where an attribute name inside of a set declaration evaluates to null (which is normally an error, as null cannot be coerced to a string), that attribute is simply not added to the set:

{ ${if foo then "bar" else null} = true; }

This will evaluate to {} if foo evaluates to false.

A set that has a __functor attribute whose value is callable (i.e. is itself a function or a set with a __functor attribute whose value is callable) can be applied as if it were a function, with the set itself passed in first , e.g.,

let add = { __functor = self: x: x + self.x; };
    inc = add // { x = 1; };
in inc 1

evaluates to 2. This can be used to attach metadata to a function without the caller needing to treat it specially, or to implement a form of object-oriented programming, for example.

Language Constructs

Recursive sets

Recursive sets are like normal attribute sets, but the attributes can refer to each other.

rec-attrset = rec { [ name = expr ; ]... }

Example:

rec {
  x = y;
  y = 123;
}.x

This evaluates to 123.

Note that without rec the binding x = y; would refer to the variable y in the surrounding scope, if one exists, and would be invalid if no such variable exists. That is, in a normal (non-recursive) set, attributes are not added to the lexical scope; in a recursive set, they are.

Recursive sets of course introduce the danger of infinite recursion. For example, the expression

rec {
  x = y;
  y = x;
}.x

will crash with an infinite recursion encountered error message.

Let-expressions

A let-expression allows you to define local variables for an expression.

let-in = let [ identifier = expr ]... in expr

Example:

let
  x = "foo";
  y = "bar";
in x + y

This evaluates to "foobar".

Inheriting attributes

When defining an attribute set or in a let-expression it is often convenient to copy variables from the surrounding lexical scope (e.g., when you want to propagate attributes). This can be shortened using the inherit keyword.

Example:

let x = 123; in
{
  inherit x;
  y = 456;
}

is equivalent to

let x = 123; in
{
  x = x;
  y = 456;
}

and both evaluate to { x = 123; y = 456; }.

Note

This works because x is added to the lexical scope by the let construct.

It is also possible to inherit attributes from another attribute set.

Example:

In this fragment from all-packages.nix,

graphviz = (import ../tools/graphics/graphviz) {
  inherit fetchurl stdenv libpng libjpeg expat x11 yacc;
  inherit (xorg) libXaw;
};

xorg = {
  libX11 = ...;
  libXaw = ...;
  ...
}

libpng = ...;
libjpg = ...;
...

the set used in the function call to the function defined in ../tools/graphics/graphviz inherits a number of variables from the surrounding scope (fetchurl ... yacc), but also inherits libXaw (the X Athena Widgets) from the xorg set.

Summarizing the fragment

...
inherit x y z;
inherit (src-set) a b c;
...

is equivalent to

...
x = x; y = y; z = z;
a = src-set.a; b = src-set.b; c = src-set.c;
...

when used while defining local variables in a let-expression or while defining a set.

In a let expression, inherit can be used to selectively bring specific attributes of a set into scope. For example

let
  x = { a = 1; b = 2; };
  inherit (builtins) attrNames;
in
{
  names = attrNames x;
}

is equivalent to

let
  x = { a = 1; b = 2; };
in
{
  names = builtins.attrNames x;
}

both evaluate to { names = [ "a" "b" ]; }.

Functions

Functions have the following form:

pattern: body

The pattern specifies what the argument of the function must look like, and binds variables in the body to (parts of) the argument. There are three kinds of patterns:

• If a pattern is a single identifier, then the function matches any argument. Example:

  let negate = x: !x;
      concat = x: y: x + y;
  in if negate true then concat "foo" "bar" else ""
  

  Note that concat is a function that takes one argument and returns a function that takes another argument. This allows partial parameterisation (i.e., only filling some of the arguments of a function); e.g.,

  map (concat "foo") [ "bar" "bla" "abc" ]
  

  evaluates to [ "foobar" "foobla" "fooabc" ].

• A set pattern of the form { name1, name2, …, nameN } matches a set containing the listed attributes, and binds the values of those attributes to variables in the function body. For example, the function

  { x, y, z }: z + y + x
  

  can only be called with a set containing exactly the attributes x, y and z. No other attributes are allowed. If you want to allow additional arguments, you can use an ellipsis (...):

  { x, y, z, ... }: z + y + x
  

  This works on any set that contains at least the three named attributes.

  It is possible to provide default values for attributes, in which case they are allowed to be missing. A default value is specified by writing name ? e, where e is an arbitrary expression. For example,

  { x, y ? "foo", z ? "bar" }: z + y + x
  

  specifies a function that only requires an attribute named x, but optionally accepts y and z.

• An @-pattern provides a means of referring to the whole value being matched:

  args@{ x, y, z, ... }: z + y + x + args.a
  

  but can also be written as:

  { x, y, z, ... } @ args: z + y + x + args.a
  

  Here args is bound to the argument as passed, which is further matched against the pattern { x, y, z, ... }. The @-pattern makes mainly sense with an ellipsis(...) as you can access attribute names as a, using args.a, which was given as an additional attribute to the function.

  Warning

  args@ binds the name args to the attribute set that is passed to the function. In particular, args does not include any default values specified with ? in the function's set pattern.

  For instance

  let
    f = args@{ a ? 23, ... }: [ a args ];
  in
    f {}
  

  is equivalent to

  let
    f = args @ { ... }: [ (args.a or 23) args ];
  in
    f {}
  

  and both expressions will evaluate to:

  [ 23 {} ]
  

Note that functions do not have names. If you want to give them a name, you can bind them to an attribute, e.g.,

let concat = { x, y }: x + y;
in concat { x = "foo"; y = "bar"; }

Conditionals

Conditionals look like this:

if e1 then e2 else e3

where e1 is an expression that should evaluate to a Boolean value (true or false).

Assertions

Assertions are generally used to check that certain requirements on or between features and dependencies hold. They look like this:

assert e1; e2

where e1 is an expression that should evaluate to a Boolean value. If it evaluates to true, e2 is returned; otherwise expression evaluation is aborted and a backtrace is printed.

Here is a Nix expression for the Subversion package that shows how assertions can be used:.

{ localServer ? false
, httpServer ? false
, sslSupport ? false
, pythonBindings ? false
, javaSwigBindings ? false
, javahlBindings ? false
, stdenv, fetchurl
, openssl ? null, httpd ? null, db4 ? null, expat, swig ? null, j2sdk ? null
}:

assert localServer -> db4 != null; ①
assert httpServer -> httpd != null && httpd.expat == expat; ②
assert sslSupport -> openssl != null && (httpServer -> httpd.openssl == openssl); ③
assert pythonBindings -> swig != null && swig.pythonSupport;
assert javaSwigBindings -> swig != null && swig.javaSupport;
assert javahlBindings -> j2sdk != null;

stdenv.mkDerivation {
  name = "subversion-1.1.1";
  ...
  openssl = if sslSupport then openssl else null; ④
  ...
}

The points of interest are:

1. This assertion states that if Subversion is to have support for local repositories, then Berkeley DB is needed. So if the Subversion function is called with the localServer argument set to true but the db4 argument set to null, then the evaluation fails.

   Note that -> is the logical implication Boolean operation.

2. This is a more subtle condition: if Subversion is built with Apache (httpServer) support, then the Expat library (an XML library) used by Subversion should be same as the one used by Apache. This is because in this configuration Subversion code ends up being linked with Apache code, and if the Expat libraries do not match, a build- or runtime link error or incompatibility might occur.

3. This assertion says that in order for Subversion to have SSL support (so that it can access https URLs), an OpenSSL library must be passed. Additionally, it says that if Apache support is enabled, then Apache's OpenSSL should match Subversion's. (Note that if Apache support is not enabled, we don't care about Apache's OpenSSL.)

4. The conditional here is not really related to assertions, but is worth pointing out: it ensures that if SSL support is disabled, then the Subversion derivation is not dependent on OpenSSL, even if a non-null value was passed. This prevents an unnecessary rebuild of Subversion if OpenSSL changes.

With-expressions

A with-expression,

with e1; e2

introduces the set e1 into the lexical scope of the expression e2. For instance,

let as = { x = "foo"; y = "bar"; };
in with as; x + y

evaluates to "foobar" since the with adds the x and y attributes of as to the lexical scope in the expression x + y. The most common use of with is in conjunction with the import function. E.g.,

with (import ./definitions.nix); ...

makes all attributes defined in the file definitions.nix available as if they were defined locally in a let-expression.

The bindings introduced by with do not shadow bindings introduced by other means, e.g.

let a = 3; in with { a = 1; }; let a = 4; in with { a = 2; }; ...

establishes the same scope as

let a = 1; in let a = 2; in let a = 3; in let a = 4; in ...

Variables coming from outer with expressions are shadowed:

with { a = "outer"; };
with { a = "inner"; };
a

Does evaluate to "inner".

Comments

• Inline comments start with # and run until the end of the line.

  Example

  # A number
  2 # Equals 1 + 1
  

  2
  

• Block comments start with /* and run until the next occurrence of */.

  Example

  /*
  Block comments
  can span multiple lines.
  */ "hello"
  

  "hello"
  

  This means that block comments cannot be nested.

  Example

  /* /* nope */ */ 1
  

  error: syntax error, unexpected '*'
  
         at «string»:1:15:
  
              1| /* /* nope */ *
               |               ^
  

  Consider escaping nested comments and unescaping them in post-processing.

  Example

  /* /* nested *\/ */ 1
  

  1
  

Scoping rules

Nix is statically scoped, but with multiple scopes and shadowing rules.

• primary scope --- explicitly-bound variables

  • let
  • inherit
  • function arguments

• secondary scope --- implicitly-bound variables

  • with Primary scope takes precedence over secondary scope. See with for a detailed example.

String interpolation

String interpolation is a language feature where a string, path, or attribute name can contain expressions enclosed in ${ } (dollar-sign with curly brackets).

Such a construct is called interpolated string, and the expression inside is an interpolated expression.

Examples

String

Rather than writing

"--with-freetype2-library=" + freetype + "/lib"

(where freetype is a derivation), you can instead write

"--with-freetype2-library=${freetype}/lib"

The latter is automatically translated to the former.

A more complicated example (from the Nix expression for Qt):

configureFlags = "
  -system-zlib -system-libpng -system-libjpeg
  ${if openglSupport then "-dlopen-opengl
    -L${mesa}/lib -I${mesa}/include
    -L${libXmu}/lib -I${libXmu}/include" else ""}
  ${if threadSupport then "-thread" else "-no-thread"}
";

Note that Nix expressions and strings can be arbitrarily nested; in this case the outer string contains various interpolated expressions that themselves contain strings (e.g., "-thread"), some of which in turn contain interpolated expressions (e.g., ${mesa}).

Path

Rather than writing

./. + "/" + foo + "-" + bar + ".nix"

or

./. + "/${foo}-${bar}.nix"

you can instead write

./${foo}-${bar}.nix

Attribute name

Attribute names can be interpolated strings.

Example

let name = "foo"; in
{ ${name} = 123; }

{ foo = 123; }

Attributes can be selected with interpolated strings.

Example

let name = "foo"; in
{ foo = 123; }.${name}

123

Interpolated expression

An expression that is interpolated must evaluate to one of the following:

• a string

• a path

• an attribute set that has a __toString attribute or an outPath attribute

  • __toString must be a function that takes the attribute set itself and returns a string
  • outPath must be a string

  This includes derivations or flake inputs (experimental).

A string interpolates to itself.

A path in an interpolated expression is first copied into the Nix store, and the resulting string is the store path of the newly created store object.

Example

$ mkdir foo

Reference the empty directory in an interpolated expression:

"${./foo}"

"/nix/store/2hhl2nz5v0khbn06ys82nrk99aa1xxdw-foo"

A derivation interpolates to the store path of its first output.

Example

let
  pkgs = import <nixpkgs> {};
in
"${pkgs.hello}"

"/nix/store/4xpfqf29z4m8vbhrqcz064wfmb46w5r7-hello-2.12.1"

An attribute set interpolates to the return value of the function in the __toString applied to the attribute set itself.

Example

let
  a = {
    value = 1;
    __toString = self: toString (self.value + 1);
  };
in
"${a}"

"2"

An attribute set also interpolates to the value of its outPath attribute.

Example

let
  a = { outPath = "foo"; };
in
"${a}"

"foo"

If both __toString and outPath are present in an attribute set, __toString takes precedence.

Example

let
  a = { __toString = _: "yes"; outPath = throw "no"; };
in
"${a}"

"yes"

If neither is present, an error is thrown.

Example

let
  a = {};
in
"${a}"

error: cannot coerce a set to a string: { }

       at «string»:4:2:

            3| in
            4| "${a}"
             |  ^

Lookup path

Syntax

lookup-path = < identifier [ / identifier ]... >

A lookup path is an identifier with an optional path suffix that resolves to a path value if the identifier matches a search path entry.

The value of a lookup path is determined by builtins.nixPath.

See builtins.findFile for details on lookup path resolution.

Example

<nixpkgs>

/nix/var/nix/profiles/per-user/root/channels/nixpkgs

Example

<nixpkgs/nixos>

/nix/var/nix/profiles/per-user/root/channels/nixpkgs/nixos

String context

Note

This is an advanced topic. The Nix language is designed to be used without the programmer consciously dealing with string contexts or even knowing what they are.

A string in the Nix language is not just a sequence of characters like strings in other languages. It is actually a pair of a sequence of characters and a string context. The string context is an (unordered) set of string context elements.

The purpose of string contexts is to collect non-string values attached to strings via string concatenation, string interpolation, and similar operations. The idea is that a user can combine together values to create a build instructions for derivations without manually keeping track of where they come from. Then the Nix language implicitly does that bookkeeping to efficiently obtain the closure of derivation inputs.

Note

String contexts are not explicitly manipulated in idiomatic Nix language code.

String context elements come in different forms:

• deriving path

  A string context element of this type is a deriving path. They can be either of type constant or output, which correspond to the types of deriving paths.

  • Constant string context elements

    Example

    builtins.storePath creates a string with a single constant string context element:

    builtins.getContext (builtins.storePath "/nix/store/wkhdf9jinag5750mqlax6z2zbwhqb76n-hello-2.10")
    

    evaluates to

    {
      "/nix/store/wkhdf9jinag5750mqlax6z2zbwhqb76n-hello-2.10" = {
        path = true;
      };
    }
    

  • Output string context elements

    Example

    The behavior of string contexts are best demonstrated with a built-in function that is still experimental: builtins.outputOf. This example will not work with stable Nix!

    builtins.getContext
      (builtins.outputOf
        (builtins.storePath "/nix/store/fvchh9cvcr7kdla6n860hshchsba305w-hello-2.12.drv")
        "out")
    

    evaluates to

    {
      "/nix/store/fvchh9cvcr7kdla6n860hshchsba305w-hello-2.12.drv" = {
        outputs = [ "out" ];
      };
    }
    

• derivation deep

  derivation deep is an advanced feature intended to be used with the exportReferencesGraph derivation attribute. A derivation deep string context element is a derivation path, and refers to both its outputs and the entire build closure of that derivation: all its outputs, all the other derivations the given derivation depends on, and all the outputs of those.

  Example

  The best way to illustrate derivation deep string contexts is with builtins.addDrvOutputDependencies. Take a regular constant string context element pointing to a derivation, and transform it into a "Derivation deep" string context element.

  builtins.getContext
    (builtins.addDrvOutputDependencies
      (builtins.storePath "/nix/store/fvchh9cvcr7kdla6n860hshchsba305w-hello-2.12.drv"))
  

  evaluates to

  {
    "/nix/store/fvchh9cvcr7kdla6n860hshchsba305w-hello-2.12.drv" = {
      allOutputs = true;
    };
  }
  

Inspecting string contexts

Most basically, builtins.hasContext will tell whether a string has a non-empty context.

When more granular information is needed, builtins.getContext can be used. It creates an attribute set representing the string context, which can be inspected as usual.

Clearing string contexts

buitins.unsafeDiscardStringContext will make a copy of a string, but with an empty string context. The returned string can be used in more ways, e.g. by operators that require the string context to be empty. The requirement to explicitly discard the string context in such use cases helps ensure that string context elements are not lost by mistake. The "unsafe" marker is only there to remind that Nix normally guarantees that dependencies are tracked, whereas the returned string has lost them.

Constructing string contexts

builtins.appendContext will create a copy of a string, but with additional string context elements. The context is specified explicitly by an attribute set in the format that builtins.hasContext produces. A string with arbitrary contexts can be made like this:

1. Create a string with the desired string context elements. (The contents of the string do not matter.)
2. Dump its context with builtins.getContext.
3. Combine it with a base string and repeated builtins.appendContext calls.

Operators

Attribute selection

Syntax

attrset . attrpath [ or expr ]

Select the attribute denoted by attribute path attrpath from attribute set attrset. If the attribute doesn’t exist, return the expr after or if provided, otherwise abort evaluation.

An attribute path is a dot-separated list of attribute names.

Syntax

attrpath = name [ . name ]...

Has attribute

Syntax

attrset ? attrpath

Test whether attribute set attrset contains the attribute denoted by attrpath. The result is a Boolean value.

See also: builtins.hasAttr

After evaluating attrset and attrpath, the computational complexity is O(log(n)) for n attributes in the attrset

Arithmetic

Numbers are type-compatible: Pure integer operations will always return integers, whereas any operation involving at least one floating point number return a floating point number.

See also Comparison and Equality.

The + operator is overloaded to also work on strings and paths.

String concatenation

Syntax

string + string

Concatenate two strings and merge their string contexts.

Path concatenation

Syntax

path + path

Concatenate two paths. The result is a path.

Path and string concatenation

Syntax

path + string

Concatenate path with string. The result is a path.

Note

The string must not have a string context that refers to a store path.

String and path concatenation

Syntax

string + path

Concatenate string with path. The result is a string.

Important

The file or directory at path must exist and is copied to the store. The path appears in the result as the corresponding store path.

Update

Syntax

attrset1 // attrset2

Update attribute set attrset1 with names and values from attrset2.

The returned attribute set will have all of the attributes in attrset1 and attrset2. If an attribute name is present in both, the attribute value from the latter is taken.

Comparison

Comparison is

• arithmetic for numbers
• lexicographic for strings and paths
• item-wise lexicographic for lists: elements at the same index in both lists are compared according to their type and skipped if they are equal.

All comparison operators are implemented in terms of <, and the following equivalencies hold:

Equality

• Attribute sets and lists are compared recursively, and therefore are fully evaluated.
• Comparison of functions always returns false.
• Numbers are type-compatible, see arithmetic operators.
• Floating point numbers only differ up to a limited precision.

Logical implication

Equivalent to !b1 || b2.

Derivations

The most important built-in function is derivation, which is used to describe a single derivation: a specification for running an executable on precisely defined input files to repeatably produce output files at uniquely determined file system paths.

It takes as input an attribute set, the attributes of which specify the inputs to the process. It outputs an attribute set, and produces a store derivation as a side effect of evaluation.

Input attributes

Required

• name (String)

  A symbolic name for the derivation. It is added to the store path of the corresponding store derivation as well as to its output paths.

  Example

  derivation {
    name = "hello";
    # ...
  }
  

  The store derivation's path will be /nix/store/<hash>-hello.drv. The output paths will be of the form /nix/store/<hash>-hello[-<output>]

• system (String)

  The system type on which the builder executable is meant to be run.

  A necessary condition for Nix to build derivations locally is that the system attribute matches the current system configuration option. It can automatically build on other platforms by forwarding build requests to other machines.

  Example

  Declare a derivation to be built on a specific system type:

  derivation {
    # ...
    system = "x86_64-linux";
    # ...
  }
  

  Example

  Declare a derivation to be built on the system type that evaluates the expression:

  derivation {
    # ...
    system = builtins.currentSystem;
    # ...
  }
  

  builtins.currentSystem has the value of the system configuration option, and defaults to the system type of the current Nix installation.

• builder (Path | String)

  Path to an executable that will perform the build.

  Example

  Use the file located at /bin/bash as the builder executable:

  derivation {
    # ...
    builder = "/bin/bash";
    # ...
  };
  

  Example

  Copy a local file to the Nix store for use as the builder executable:

  derivation {
    # ...
    builder = ./builder.sh;
    # ...
  };
  

  Example

  Use a file from another derivation as the builder executable:

  let pkgs = import <nixpkgs> {}; in
  derivation {
    # ...
    builder = "${pkgs.python}/bin/python";
    # ...
  };
  

Optional

• args (List of String)

  Default: [ ]

  Command-line arguments to be passed to the builder executable.

  Example

  Pass arguments to Bash to interpret a shell command:

  derivation {
    # ...
    builder = "/bin/bash";
    args = [ "-c" "echo hello world > $out" ];
    # ...
  };
  

• outputs (List of String)

  Default: [ "out" ]

  Symbolic outputs of the derivation. Each output name is passed to the builder executable as an environment variable with its value set to the corresponding store path.

  By default, a derivation produces a single output called out. However, derivations can produce multiple outputs. This allows the associated store objects and their closures to be copied or garbage-collected separately.

  Example

  Imagine a library package that provides a dynamic library, header files, and documentation. A program that links against such a library doesn’t need the header files and documentation at runtime, and it doesn’t need the documentation at build time. Thus, the library package could specify:

  derivation {
    # ...
    outputs = [ "lib" "dev" "doc" ];
    # ...
  }
  

  This will cause Nix to pass environment variables lib, dev, and doc to the builder containing the intended store paths of each output. The builder would typically do something like

  ./configure \
    --libdir=$lib/lib \
    --includedir=$dev/include \
    --docdir=$doc/share/doc
  

  for an Autoconf-style package.

  The name of an output is combined with the name of the derivation to create the name part of the output's store path, unless it is out, in which case just the name of the derivation is used.

  Example

  derivation {
    name = "example";
    outputs = [ "lib" "dev" "doc" "out" ];
    # ...
  }
  

  The store derivation path will be /nix/store/<hash>-example.drv. The output paths will be

  • /nix/store/<hash>-example-lib
  • /nix/store/<hash>-example-dev
  • /nix/store/<hash>-example-doc
  • /nix/store/<hash>-example

  You can refer to each output of a derivation by selecting it as an attribute. The first element of outputs determines the default output and ends up at the top-level.

  Example

  Select an output by attribute name:

  let
    myPackage = derivation {
      name = "example";
      outputs = [ "lib" "dev" "doc" "out" ];
      # ...
    };
  in myPackage.dev
  

  Since lib is the first output, myPackage is equivalent to myPackage.lib.

• See Advanced Attributes for more, infrequently used, optional attributes.

• Every other attribute is passed as an environment variable to the builder. Attribute values are translated to environment variables as follows:

  • Strings are passed unchanged.

  • Integral numbers are converted to decimal notation.

  • Floating point numbers are converted to simple decimal or scientific notation with a preset precision.

  • A path (e.g., ../foo/sources.tar) causes the referenced file to be copied to the store; its location in the store is put in the environment variable. The idea is that all sources should reside in the Nix store, since all inputs to a derivation should reside in the Nix store.

  • A derivation causes that derivation to be built prior to the present derivation. The environment variable is set to the store path of the derivation's default output.

  • Lists of the previous types are also allowed. They are simply concatenated, separated by spaces.

  • true is passed as the string 1, false and null are passed as an empty string.

Builder execution

The builder is executed as follows:

• A temporary directory is created under the directory specified by TMPDIR (default /tmp) where the build will take place. The current directory is changed to this directory.

• The environment is cleared and set to the derivation attributes, as specified above.

• In addition, the following variables are set:

  • NIX_BUILD_TOP contains the path of the temporary directory for this build.

  • Also, TMPDIR, TEMPDIR, TMP, TEMP are set to point to the temporary directory. This is to prevent the builder from accidentally writing temporary files anywhere else. Doing so might cause interference by other processes.

  • PATH is set to /path-not-set to prevent shells from initialising it to their built-in default value.

  • HOME is set to /homeless-shelter to prevent programs from using /etc/passwd or the like to find the user's home directory, which could cause impurity. Usually, when HOME is set, it is used as the location of the home directory, even if it points to a non-existent path.

  • NIX_STORE is set to the path of the top-level Nix store directory (typically, /nix/store).

  • NIX_ATTRS_JSON_FILE & NIX_ATTRS_SH_FILE if __structuredAttrs is set to true for the derivation. A detailed explanation of this behavior can be found in the section about structured attrs.

  • For each output declared in outputs, the corresponding environment variable is set to point to the intended path in the Nix store for that output. Each output path is a concatenation of the cryptographic hash of all build inputs, the name attribute and the output name. (The output name is omitted if it’s out.)

• If an output path already exists, it is removed. Also, locks are acquired to prevent multiple Nix instances from performing the same build at the same time.

• A log of the combined standard output and error is written to /nix/var/log/nix.

• The builder is executed with the arguments specified by the attribute args. If it exits with exit code 0, it is considered to have succeeded.

• The temporary directory is removed (unless the -K option was specified).

• If the build was successful, Nix scans each output path for references to input paths by looking for the hash parts of the input paths. Since these are potential runtime dependencies, Nix registers them as dependencies of the output paths.

• After the build, Nix sets the last-modified timestamp on all files in the build result to 1 (00:00:01 1/1/1970 UTC), sets the group to the default group, and sets the mode of the file to 0444 or 0555 (i.e., read-only, with execute permission enabled if the file was originally executable). Note that possible setuid and setgid bits are cleared. Setuid and setgid programs are not currently supported by Nix. This is because the Nix archives used in deployment have no concept of ownership information, and because it makes the build result dependent on the user performing the build.

Advanced Attributes

Derivations can declare some infrequently used optional attributes.

• allowedReferences
  The optional attribute allowedReferences specifies a list of legal references (dependencies) of the output of the builder. For example,

  allowedReferences = [];
  

  enforces that the output of a derivation cannot have any runtime dependencies on its inputs. To allow an output to have a runtime dependency on itself, use "out" as a list item. This is used in NixOS to check that generated files such as initial ramdisks for booting Linux don’t have accidental dependencies on other paths in the Nix store.

• allowedRequisites
  This attribute is similar to allowedReferences, but it specifies the legal requisites of the whole closure, so all the dependencies recursively. For example,

  allowedRequisites = [ foobar ];
  

  enforces that the output of a derivation cannot have any other runtime dependency than foobar, and in addition it enforces that foobar itself doesn't introduce any other dependency itself.

• disallowedReferences
  The optional attribute disallowedReferences specifies a list of illegal references (dependencies) of the output of the builder. For example,

  disallowedReferences = [ foo ];
  

  enforces that the output of a derivation cannot have a direct runtime dependencies on the derivation foo.

• disallowedRequisites
  This attribute is similar to disallowedReferences, but it specifies illegal requisites for the whole closure, so all the dependencies recursively. For example,

  disallowedRequisites = [ foobar ];
  

  enforces that the output of a derivation cannot have any runtime dependency on foobar or any other derivation depending recursively on foobar.

• exportReferencesGraph
  This attribute allows builders access to the references graph of their inputs. The attribute is a list of inputs in the Nix store whose references graph the builder needs to know. The value of this attribute should be a list of pairs [ name1 path1 name2 path2 ... ]. The references graph of each pathN will be stored in a text file nameN in the temporary build directory. The text files have the format used by nix-store --register-validity (with the deriver fields left empty). For example, when the following derivation is built:

  derivation {
    ...
    exportReferencesGraph = [ "libfoo-graph" libfoo ];
  };
  

  the references graph of libfoo is placed in the file libfoo-graph in the temporary build directory.

  exportReferencesGraph is useful for builders that want to do something with the closure of a store path. Examples include the builders in NixOS that generate the initial ramdisk for booting Linux (a cpio archive containing the closure of the boot script) and the ISO-9660 image for the installation CD (which is populated with a Nix store containing the closure of a bootable NixOS configuration).

• impureEnvVars
  This attribute allows you to specify a list of environment variables that should be passed from the environment of the calling user to the builder. Usually, the environment is cleared completely when the builder is executed, but with this attribute you can allow specific environment variables to be passed unmodified. For example, fetchurl in Nixpkgs has the line

  impureEnvVars = [ "http_proxy" "https_proxy" ... ];
  

  to make it use the proxy server configuration specified by the user in the environment variables http_proxy and friends.

  This attribute is only allowed in fixed-output derivations (see below), where impurities such as these are okay since (the hash of) the output is known in advance. It is ignored for all other derivations.

  Warning

  impureEnvVars implementation takes environment variables from the current builder process. When a daemon is building its environmental variables are used. Without the daemon, the environmental variables come from the environment of the nix-build.

  If the configurable-impure-env experimental feature is enabled, these environment variables can also be controlled through the impure-env configuration setting.

• outputHash; outputHashAlgo; outputHashMode
  These attributes declare that the derivation is a so-called fixed-output derivation, which means that a cryptographic hash of the output is already known in advance. When the build of a fixed-output derivation finishes, Nix computes the cryptographic hash of the output and compares it to the hash declared with these attributes. If there is a mismatch, the build fails.

  The rationale for fixed-output derivations is derivations such as those produced by the fetchurl function. This function downloads a file from a given URL. To ensure that the downloaded file has not been modified, the caller must also specify a cryptographic hash of the file. For example,

  fetchurl {
    url = "http://ftp.gnu.org/pub/gnu/hello/hello-2.1.1.tar.gz";
    sha256 = "1md7jsfd8pa45z73bz1kszpp01yw6x5ljkjk2hx7wl800any6465";
  }
  

  It sometimes happens that the URL of the file changes, e.g., because servers are reorganised or no longer available. We then must update the call to fetchurl, e.g.,

  fetchurl {
    url = "ftp://ftp.nluug.nl/pub/gnu/hello/hello-2.1.1.tar.gz";
    sha256 = "1md7jsfd8pa45z73bz1kszpp01yw6x5ljkjk2hx7wl800any6465";
  }
  

  If a fetchurl derivation was treated like a normal derivation, the output paths of the derivation and all derivations depending on it would change. For instance, if we were to change the URL of the Glibc source distribution in Nixpkgs (a package on which almost all other packages depend) massive rebuilds would be needed. This is unfortunate for a change which we know cannot have a real effect as it propagates upwards through the dependency graph.

  For fixed-output derivations, on the other hand, the name of the output path only depends on the outputHash* and name attributes, while all other attributes are ignored for the purpose of computing the output path. (The name attribute is included because it is part of the path.)

  As an example, here is the (simplified) Nix expression for fetchurl:

  { stdenv, curl }: # The curl program is used for downloading.
  
  { url, sha256 }:
  
  stdenv.mkDerivation {
    name = baseNameOf (toString url);
    builder = ./builder.sh;
    buildInputs = [ curl ];
  
    # This is a fixed-output derivation; the output must be a regular
    # file with SHA256 hash sha256.
    outputHashMode = "flat";
    outputHashAlgo = "sha256";
    outputHash = sha256;
  
    inherit url;
  }
  

  The outputHash attribute must be a string containing the hash in either hexadecimal or "nix32" encoding, or following the format for integrity metadata as defined by SRI. The "nix32" encoding is an adaptation of base-32 encoding. The convertHash function shows how to convert between different encodings, and the nix-hash command has information about obtaining the hash for some contents, as well as converting to and from encodings.

  The outputHashAlgo attribute specifies the hash algorithm used to compute the hash. It can currently be "sha1", "sha256", "sha512", or null. outputHashAlgo can only be null when outputHash follows the SRI format.

  The outputHashMode attribute determines how the hash is computed. It must be one of the following values:

  • "flat"

    This is the default.

  • "recursive" or "nar"

    Compatibility

    "recursive" is the traditional way of indicating this, and is supported since 2005 (virtually the entire history of Nix). "nar" is more clear, and consistent with other parts of Nix (such as the CLI), however support for it is only added in Nix version 2.21.

  • "text"

    Warning

    The use of this method for derivation outputs is part of the dynamic-derivations experimental feature.

  • "git"

    Warning

    This method is part of the git-hashing experimental feature.

• __contentAddressed

  Warning This attribute is part of an experimental feature.

  To use this attribute, you must enable the ca-derivations experimental feature. For example, in nix.conf you could add:

  extra-experimental-features = ca-derivations
  

  If this attribute is set to true, then the derivation outputs will be stored in a content-addressed location rather than the traditional input-addressed one.

  Setting this attribute also requires setting outputHashMode and outputHashAlgo like for fixed-output derivations (see above).

  It also implicitly requires that the machine to build the derivation must have the ca-derivations system feature.

• passAsFile
  A list of names of attributes that should be passed via files rather than environment variables. For example, if you have

  passAsFile = ["big"];
  big = "a very long string";
  

  then when the builder runs, the environment variable bigPath will contain the absolute path to a temporary file containing a very long string. That is, for any attribute x listed in passAsFile, Nix will pass an environment variable xPath holding the path of the file containing the value of attribute x. This is useful when you need to pass large strings to a builder, since most operating systems impose a limit on the size of the environment (typically, a few hundred kilobyte).

• preferLocalBuild
  If this attribute is set to true and distributed building is enabled, then, if possible, the derivation will be built locally instead of being forwarded to a remote machine. This is useful for derivations that are cheapest to build locally.

• allowSubstitutes
  If this attribute is set to false, then Nix will always build this derivation (locally or remotely); it will not try to substitute its outputs. This is useful for derivations that are cheaper to build than to substitute.

  This attribute can be ignored by setting always-allow-substitutes to true.

  Note

  If set to false, the builder should be able to run on the system type specified in the system attribute, since the derivation cannot be substituted.

• __structuredAttrs
  If the special attribute __structuredAttrs is set to true, the other derivation attributes are serialised into a file in JSON format. The environment variable NIX_ATTRS_JSON_FILE points to the exact location of that file both in a build and a nix-shell. This obviates the need for passAsFile since JSON files have no size restrictions, unlike process environments.

  It also makes it possible to tweak derivation settings in a structured way; see outputChecks for example.

  As a convenience to Bash builders, Nix writes a script that initialises shell variables corresponding to all attributes that are representable in Bash. The environment variable NIX_ATTRS_SH_FILE points to the exact location of the script, both in a build and a nix-shell. This includes non-nested (associative) arrays. For example, the attribute hardening.format = true ends up as the Bash associative array element ${hardening[format]}.

• outputChecks
  When using structured attributes, the outputChecks attribute allows defining checks per-output.

  In addition to allowedReferences, allowedRequisites, disallowedReferences and disallowedRequisites, the following attributes are available:

  • maxSize defines the maximum size of the resulting store object.
  • maxClosureSize defines the maximum size of the output's closure.
  • ignoreSelfRefs controls whether self-references should be considered when checking for allowed references/requisites.

  Example:

  __structuredAttrs = true;
  
  outputChecks.out = {
    # The closure of 'out' must not be larger than 256 MiB.
    maxClosureSize = 256 * 1024 * 1024;
  
    # It must not refer to the C compiler or to the 'dev' output.
    disallowedRequisites = [ stdenv.cc "dev" ];
  };
  
  outputChecks.dev = {
    # The 'dev' output must not be larger than 128 KiB.
    maxSize = 128 * 1024;
  };
  

• unsafeDiscardReferences

  When using structured attributes, the attribute unsafeDiscardReferences is an attribute set with a boolean value for each output name. If set to true, it disables scanning the output for runtime dependencies.

  Example:

  __structuredAttrs = true;
  unsafeDiscardReferences.out = true;
  

  This is useful, for example, when generating self-contained filesystem images with their own embedded Nix store: hashes found inside such an image refer to the embedded store and not to the host's Nix store.

• requiredSystemFeatures

  If a derivation has the requiredSystemFeatures attribute, then Nix will only build it on a machine that has the corresponding features set in its system-features configuration.

  For example, setting

  requiredSystemFeatures = [ "kvm" ];
  

  ensures that the derivation can only be built on a machine with the kvm feature.

Import From Derivation

The value of a Nix expression can depend on the contents of a store object.

Passing an expression expr that evaluates to a store path to any built-in function which reads from the filesystem constitutes Import From Derivation (IFD):

• import expr
• builtins.readFile expr
• builtins.readFileType expr
• builtins.readDir expr
• builtins.pathExists expr
• builtins.filterSource f expr
• builtins.path { path = expr; }
• builtins.hashFile t expr
• builtins.scopedImport x drv

When the store path needs to be accessed, evaluation will be paused, the corresponding store object realised, and then evaluation resumed.

This has performance implications: Evaluation can only finish when all required store objects are realised. Since the Nix language evaluator is sequential, it only finds store paths to read from one at a time. While realisation is always parallel, in this case it cannot be done for all required store paths at once, and is therefore much slower than otherwise.

Realising store objects during evaluation can be disabled by setting allow-import-from-derivation to false. Without IFD it is ensured that evaluation is complete and Nix can produce a build plan before starting any realisation.

Example

In the following Nix expression, the inner derivation drv produces a file with contents hello.

# IFD.nix
let
  drv = derivation {
    name = "hello";
    builder = "/bin/sh";
    args = [ "-c" "echo -n hello > $out" ];
    system = builtins.currentSystem;
  };
in "${builtins.readFile drv} world"

nix-instantiate IFD.nix --eval --read-write-mode

building '/nix/store/348q1cal6sdgfxs8zqi9v8llrsn4kqkq-hello.drv'...
"hello world"

The contents of the derivation's output have to be realised before they can be read with readFile. Only then evaluation can continue to produce the final result.

Illustration

As a first approximation, the following data flow graph shows how evaluation and building are interleaved, if the value of a Nix expression depends on realising a store object. Boxes are data structures, arrow labels are transformations.

+----------------------+             +------------------------+
| Nix evaluator        |             | Nix store              |
|  .----------------.  |             |                        |
|  | Nix expression |  |             |                        |
|  '----------------'  |             |                        |
|          |           |             |                        |
|       evaluate       |             |                        |
|          |           |             |                        |
|          V           |             |                        |
|    .------------.    |             |  .------------------.  |
|    | derivation |----|-instantiate-|->| store derivation |  |
|    '------------'    |             |  '------------------'  |
|                      |             |           |            |
|                      |             |        realise         |
|                      |             |           |            |
|                      |             |           V            |
|  .----------------.  |             |    .--------------.    |
|  | Nix expression |<-|----read-----|----| store object |    |
|  '----------------'  |             |    '--------------'    |
|          |           |             |                        |
|       evaluate       |             |                        |
|          |           |             |                        |
|          V           |             |                        |
|    .------------.    |             |                        |
|    |   value    |    |             |                        |
|    '------------'    |             |                        |
+----------------------+             +------------------------+

In more detail, the following sequence diagram shows how the expression is evaluated step by step, and where evaluation is blocked to wait for the build output to appear.

.-------.     .-------------.                        .---------.
|Nix CLI|     |Nix evaluator|                        |Nix store|
'-------'     '-------------'                        '---------'
    |                |                                    |
    |evaluate IFD.nix|                                    |
    |--------------->|                                    |
    |                |                                    |
    |  evaluate `"${readFile drv} world"`                 |
    |                |                                    |
    |    evaluate `readFile drv`                          |
    |                |                                    |
    |   evaluate `drv` as string                          |
    |                |                                    |
    |                |instantiate /nix/store/...-hello.drv|
    |                |----------------------------------->|
    |                :                                    |
    |                :  realise /nix/store/...-hello.drv  |
    |                :----------------------------------->|
    |                :                                    |
    |                                                     |--------.
    |                :                                    |        |
    |      (evaluation blocked)                           |  echo hello > $out
    |                :                                    |        |
    |                                                     |<-------'
    |                :        /nix/store/...-hello        |
    |                |<-----------------------------------|
    |                |                                    |
    |  resume `readFile /nix/store/...-hello`             |
    |                |                                    |
    |                |   readFile /nix/store/...-hello    |
    |                |----------------------------------->|
    |                |                                    |
    |                |               hello                |
    |                |<-----------------------------------|
    |                |                                    |
    |      resume `"${"hello"} world"`                    |
    |                |                                    |
    |        resume `"hello world"`                       |
    |                |                                    |
    | "hello world"  |                                    |
    |<---------------|                                    |
.-------.     .-------------.                        .---------.
|Nix CLI|     |Nix evaluator|                        |Nix store|
'-------'     '-------------'                        '---------'

Built-in Constants

These constants are built into the Nix language evaluator:

builtins (set)

Contains all the built-in functions and values.

Since built-in functions were added over time, testing for attributes in builtins can be used for graceful fallback on older Nix installations:

# if hasContext is not available, we assume `s` has a context
if builtins ? hasContext then builtins.hasContext s else true

currentSystem (string)

The value of the eval-system or else system configuration option.

It can be used to set the system attribute for builtins.derivation such that the resulting derivation can be built on the same system that evaluates the Nix expression:

 builtins.derivation {
   # ...
   system = builtins.currentSystem;
}

It can be overridden in order to create derivations for different system than the current one:

$ nix-instantiate --system "mips64-linux" --eval --expr 'builtins.currentSystem'
"mips64-linux"

Note

Not available in pure evaluation mode.

currentTime (integer)

Return the Unix time at first evaluation. Repeated references to that name will re-use the initially obtained value.

Example:

$ nix repl
Welcome to Nix 2.15.1 Type :? for help.

nix-repl> builtins.currentTime
1683705525

nix-repl> builtins.currentTime
1683705525

The store path of a derivation depending on currentTime will differ for each evaluation, unless both evaluate builtins.currentTime in the same second.

Note

Not available in pure evaluation mode.

false (Boolean)

Primitive value.

It can be returned by comparison operators and used in conditional expressions.

The name false is not special, and can be shadowed:

nix-repl> let false = 1; in false
1

langVersion (integer)

The current version of the Nix language.

nixPath (list)

The value of the nix-path configuration setting: a list of search path entries used to resolve lookup paths.

Lookup path expressions are desugared using this and builtins.findFile:

<nixpkgs>

is equivalent to:

builtins.findFile builtins.nixPath "nixpkgs"

nixVersion (string)

The version of Nix.

For example, where the command line returns the current Nix version,

$ nix --version
nix (Nix) 2.16.0

the Nix language evaluator returns the same value:

nix-repl> builtins.nixVersion
"2.16.0"

null (null)

Primitive value.

The name null is not special, and can be shadowed:

nix-repl> let null = 1; in null
1

storeDir (string)

Logical file system location of the Nix store currently in use.

This value is determined by the store parameter in Store URLs:

$ nix-instantiate --store 'dummy://?store=/blah' --eval --expr builtins.storeDir
"/blah"

true (Boolean)

Primitive value.

It can be returned by comparison operators and used in conditional expressions.

The name true is not special, and can be shadowed:

nix-repl> let true = 1; in true
1

Built-in Functions

This section lists the functions built into the Nix language evaluator. All built-in functions are available through the global builtins constant.

For convenience, some built-ins can be accessed directly:

• derivation
• import
• abort
• throw

derivation attrs

derivation is described in its own section.

abort s

Abort Nix expression evaluation and print the error message s.

add e1 e2

Return the sum of the numbers e1 and e2.

addDrvOutputDependencies s

Create a copy of the given string where a single constant string context element is turned into a derivation deep string context element.

The store path that is the constant string context element should point to a valid derivation, and end in .drv.

The original string context element must not be empty or have multiple elements, and it must not have any other type of element other than a constant or derivation deep element. The latter is supported so this function is idempotent.

This is the opposite of builtins.unsafeDiscardOutputDependency.

all pred list

Return true if the function pred returns true for all elements of list, and false otherwise.

any pred list

Return true if the function pred returns true for at least one element of list, and false otherwise.

attrNames set

Return the names of the attributes in the set set in an alphabetically sorted list. For instance, builtins.attrNames { y = 1; x = "foo"; } evaluates to [ "x" "y" ].

attrValues set

Return the values of the attributes in the set set in the order corresponding to the sorted attribute names.

baseNameOf x

Return the base name of either a path value x or a string x, depending on which type is passed, and according to the following rules.

For a path value, the base name is considered to be the part of the path after the last directory separator, including any file extensions. This is the simple case, as path values don't have trailing slashes.

When the argument is a string, a more involved logic applies. If the string ends with a /, only this one final slash is removed.

After this, the base name is returned as previously described, assuming / as the directory separator. (Note that evaluation must be platform independent.)

This is somewhat similar to the GNU basename command, but GNU basename will strip any number of trailing slashes.

bitAnd e1 e2

Return the bitwise AND of the integers e1 and e2.

bitOr e1 e2

Return the bitwise OR of the integers e1 and e2.

bitXor e1 e2

Return the bitwise XOR of the integers e1 and e2.

break v

In debug mode (enabled using --debugger), pause Nix expression evaluation and enter the REPL. Otherwise, return the argument v.

catAttrs attr list

Collect each attribute named attr from a list of attribute sets. Attrsets that don't contain the named attribute are ignored. For example,

builtins.catAttrs "a" [{a = 1;} {b = 0;} {a = 2;}]

evaluates to [1 2].

ceil double

Converts an IEEE-754 double-precision floating-point number (double) to the next higher integer.

If the datatype is neither an integer nor a "float", an evaluation error will be thrown.

compareVersions s1 s2

Compare two strings representing versions and return -1 if version s1 is older than version s2, 0 if they are the same, and 1 if s1 is newer than s2. The version comparison algorithm is the same as the one used by nix-env -u.

concatLists lists

Concatenate a list of lists into a single list.

concatMap f list

This function is equivalent to builtins.concatLists (map f list) but is more efficient.

concatStringsSep separator list

Concatenate a list of strings with a separator between each element, e.g. concatStringsSep "/" ["usr" "local" "bin"] == "usr/local/bin".

convertHash args

Return the specified representation of a hash string, based on the attributes presented in args:

• hash

  The hash to be converted. The hash format is detected automatically.

• hashAlgo

  The algorithm used to create the hash. Must be one of

  • "md5"
  • "sha1"
  • "sha256"
  • "sha512"

  The attribute may be omitted when hash is an SRI hash or when the hash is prefixed with the hash algorithm name followed by a colon. That <hashAlgo>:<hashBody> syntax is supported for backwards compatibility with existing tooling.

• toHashFormat

  The format of the resulting hash. Must be one of

  • "base16"
  • "nix32"
  • "base32" (deprecated alias for "nix32")
  • "base64"
  • "sri"

The result hash is the toHashFormat representation of the hash hash.

Example

Convert a SHA256 hash in Base16 to SRI:

builtins.convertHash {
  hash = "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855";
  toHashFormat = "sri";
  hashAlgo = "sha256";
}

"sha256-47DEQpj8HBSa+/TImW+5JCeuQeRkm5NMpJWZG3hSuFU="

Example

Convert a SHA256 hash in SRI to Base16:

builtins.convertHash {
  hash = "sha256-47DEQpj8HBSa+/TImW+5JCeuQeRkm5NMpJWZG3hSuFU=";
  toHashFormat = "base16";
}

"e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"

Example

Convert a hash in the form <hashAlgo>:<hashBody> in Base16 to SRI:

builtins.convertHash {
  hash = "sha256:e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855";
  toHashFormat = "sri";
}

"sha256-47DEQpj8HBSa+/TImW+5JCeuQeRkm5NMpJWZG3hSuFU="

deepSeq e1 e2

This is like seq e1 e2, except that e1 is evaluated deeply: if it’s a list or set, its elements or attributes are also evaluated recursively.

dirOf s

Return the directory part of the string s, that is, everything before the final slash in the string. This is similar to the GNU dirname command.

div e1 e2

Return the quotient of the numbers e1 and e2.

elem x xs

Return true if a value equal to x occurs in the list xs, and false otherwise.

elemAt xs n

Return element n from the list xs. Elements are counted starting from 0. A fatal error occurs if the index is out of bounds.

fetchClosure args

Note

This function is only available if the fetch-closure experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = fetch-closure

Fetch a store path closure from a binary cache, and return the store path as a string with context.

This function can be invoked in three ways, that we will discuss in order of preference.

Fetch a content-addressed store path

Example:

builtins.fetchClosure {
  fromStore = "https://cache.nixos.org";
  fromPath = /nix/store/ldbhlwhh39wha58rm61bkiiwm6j7211j-git-2.33.1;
}

This is the simplest invocation, and it does not require the user of the expression to configure trusted-public-keys to ensure their authenticity.

If your store path is input addressed instead of content addressed, consider the other two invocations.

Fetch any store path and rewrite it to a fully content-addressed store path

Example:

builtins.fetchClosure {
  fromStore = "https://cache.nixos.org";
  fromPath = /nix/store/r2jd6ygnmirm2g803mksqqjm4y39yi6i-git-2.33.1;
  toPath = /nix/store/ldbhlwhh39wha58rm61bkiiwm6j7211j-git-2.33.1;
}

This example fetches /nix/store/r2jd... from the specified binary cache, and rewrites it into the content-addressed store path /nix/store/ldbh....

Like the previous example, no extra configuration or privileges are required.

To find out the correct value for toPath given a fromPath, use nix store make-content-addressed:

# nix store make-content-addressed --from https://cache.nixos.org /nix/store/r2jd6ygnmirm2g803mksqqjm4y39yi6i-git-2.33.1
rewrote '/nix/store/r2jd6ygnmirm2g803mksqqjm4y39yi6i-git-2.33.1' to '/nix/store/ldbhlwhh39wha58rm61bkiiwm6j7211j-git-2.33.1'

Alternatively, set toPath = "" and find the correct toPath in the error message.

Fetch an input-addressed store path as is

Example:

builtins.fetchClosure {
  fromStore = "https://cache.nixos.org";
  fromPath = /nix/store/r2jd6ygnmirm2g803mksqqjm4y39yi6i-git-2.33.1;
  inputAddressed = true;
}

It is possible to fetch an input-addressed store path and return it as is. However, this is the least preferred way of invoking fetchClosure, because it requires that the input-addressed paths are trusted by the Nix configuration.

builtins.storePath

fetchClosure is similar to builtins.storePath in that it allows you to use a previously built store path in a Nix expression. However, fetchClosure is more reproducible because it specifies a binary cache from which the path can be fetched. Also, using content-addressed store paths does not require users to configure trusted-public-keys to ensure their authenticity.

fetchGit args

Fetch a path from git. args can be a URL, in which case the HEAD of the repo at that URL is fetched. Otherwise, it can be an attribute with the following attributes (all except url optional):

• url

  The URL of the repo.

• name (default: source)

  The name of the directory the repo should be exported to in the store.

• rev (default: the tip of ref)

  The Git revision to fetch. This is typically a commit hash.

• ref (default: HEAD)

  The Git reference under which to look for the requested revision. This is often a branch or tag name.

  This option has no effect once shallow cloning is enabled.

  By default, the ref value is prefixed with refs/heads/. As of 2.3.0, Nix will not prefix refs/heads/ if ref starts with refs/.

• submodules (default: false)

  A Boolean parameter that specifies whether submodules should be checked out.

• exportIgnore (default: true)

  A Boolean parameter that specifies whether export-ignore from .gitattributes should be applied. This approximates part of the git archive behavior.

  Enabling this option is not recommended because it is unknown whether the Git developers commit to the reproducibility of export-ignore in newer Git versions.

• shallow (default: false)

  Make a shallow clone when fetching the Git tree. When this is enabled, the options ref and allRefs have no effect anymore.

• allRefs

  Whether to fetch all references (eg. branches and tags) of the repository. With this argument being true, it's possible to load a rev from any ref. (by default only revs from the specified ref are supported).

  This option has no effect once shallow cloning is enabled.

• verifyCommit (default: true if publicKey or publicKeys are provided, otherwise false)

  Whether to check rev for a signature matching publicKey or publicKeys. If verifyCommit is enabled, then fetchGit cannot use a local repository with uncommitted changes. Requires the verified-fetches experimental feature.

• publicKey

  The public key against which rev is verified if verifyCommit is enabled. Requires the verified-fetches experimental feature.

• keytype (default: "ssh-ed25519")

  The key type of publicKey. Possible values:

  • "ssh-dsa"
  • "ssh-ecdsa"
  • "ssh-ecdsa-sk"
  • "ssh-ed25519"
  • "ssh-ed25519-sk"
  • "ssh-rsa" Requires the verified-fetches experimental feature.

• publicKeys

  The public keys against which rev is verified if verifyCommit is enabled. Must be given as a list of attribute sets with the following form:

  {
    key = "<public key>";
    type = "<key type>"; # optional, default: "ssh-ed25519"
  }
  

  Requires the verified-fetches experimental feature.

Here are some examples of how to use fetchGit.

• To fetch a private repository over SSH:

  builtins.fetchGit {
    url = "git@github.com:my-secret/repository.git";
    ref = "master";
    rev = "adab8b916a45068c044658c4158d81878f9ed1c3";
  }
  

• To fetch an arbitrary reference:

  builtins.fetchGit {
    url = "https://github.com/NixOS/nix.git";
    ref = "refs/heads/0.5-release";
  }
  

• If the revision you're looking for is in the default branch of the git repository you don't strictly need to specify the branch name in the ref attribute.

  However, if the revision you're looking for is in a future branch for the non-default branch you will need to specify the the ref attribute as well.

  builtins.fetchGit {
    url = "https://github.com/nixos/nix.git";
    rev = "841fcbd04755c7a2865c51c1e2d3b045976b7452";
    ref = "1.11-maintenance";
  }
  

  Note

  It is nice to always specify the branch which a revision belongs to. Without the branch being specified, the fetcher might fail if the default branch changes. Additionally, it can be confusing to try a commit from a non-default branch and see the fetch fail. If the branch is specified the fault is much more obvious.

• If the revision you're looking for is in the default branch of the git repository you may omit the ref attribute.

  builtins.fetchGit {
    url = "https://github.com/nixos/nix.git";
    rev = "841fcbd04755c7a2865c51c1e2d3b045976b7452";
  }
  

• To fetch a specific tag:

  builtins.fetchGit {
    url = "https://github.com/nixos/nix.git";
    ref = "refs/tags/1.9";
  }
  

• To fetch the latest version of a remote branch:

  builtins.fetchGit {
    url = "ssh://git@github.com/nixos/nix.git";
    ref = "master";
  }
  

• To verify the commit signature:

  builtins.fetchGit {
    url = "ssh://git@github.com/nixos/nix.git";
    verifyCommit = true;
    publicKeys = [
        {
          type = "ssh-ed25519";
          key = "AAAAC3NzaC1lZDI1NTE5AAAAIArPKULJOid8eS6XETwUjO48/HKBWl7FTCK0Z//fplDi";
        }
    ];
  }
  

  Nix will refetch the branch according to the tarball-ttl setting.

  This behavior is disabled in pure evaluation mode.

• To fetch the content of a checked-out work directory:

  builtins.fetchGit ./work-dir
  

If the URL points to a local directory, and no ref or rev is given, fetchGit will use the current content of the checked-out files, even if they are not committed or added to Git's index. It will only consider files added to the Git repository, as listed by git ls-files.

fetchTarball args

Download the specified URL, unpack it and return the path of the unpacked tree. The file must be a tape archive (.tar) compressed with gzip, bzip2 or xz. The top-level path component of the files in the tarball is removed, so it is best if the tarball contains a single directory at top level. The typical use of the function is to obtain external Nix expression dependencies, such as a particular version of Nixpkgs, e.g.

with import (fetchTarball https://github.com/NixOS/nixpkgs/archive/nixos-14.12.tar.gz) {};

stdenv.mkDerivation { … }

The fetched tarball is cached for a certain amount of time (1 hour by default) in ~/.cache/nix/tarballs/. You can change the cache timeout either on the command line with --tarball-ttl number-of-seconds or in the Nix configuration file by adding the line tarball-ttl = number-of-seconds.

Note that when obtaining the hash with nix-prefetch-url the option --unpack is required.

This function can also verify the contents against a hash. In that case, the function takes a set instead of a URL. The set requires the attribute url and the attribute sha256, e.g.

with import (fetchTarball {
  url = "https://github.com/NixOS/nixpkgs/archive/nixos-14.12.tar.gz";
  sha256 = "1jppksrfvbk5ypiqdz4cddxdl8z6zyzdb2srq8fcffr327ld5jj2";
}) {};

stdenv.mkDerivation { … }

Not available in restricted evaluation mode.

fetchTree input

Note

This function is only available if the fetch-tree experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = fetch-tree

Fetch a file system tree or a plain file using one of the supported backends and return an attribute set with:

• the resulting fixed-output store path
• the corresponding NAR hash
• backend-specific metadata (currently not documented).

input must be an attribute set with the following attributes:

• type (String, required)

  One of the supported source types. This determines other required and allowed input attributes.

• narHash (String, optional)

  The narHash parameter can be used to substitute the source of the tree. It also allows for verification of tree contents that may not be provided by the underlying transfer mechanism. If narHash is set, the source is first looked up is the Nix store and substituters, and only fetched if not available.

A subset of the output attributes of fetchTree can be re-used for subsequent calls to fetchTree to produce the same result again. That is, fetchTree is idempotent.

Downloads are cached in $XDG_CACHE_HOME/nix. The remote source will be fetched from the network if both are true:

• A NAR hash is supplied and the corresponding store path is not valid, that is, not available in the store

  Note

  Substituters are not used in fetching.

• There is no cache entry or the cache entry is older than tarball-ttl

Source types

The following source types and associated input attributes are supported.

• "file"

  Place a plain file into the Nix store. This is similar to builtins.fetchurl

  • url (String, required)

    Supported protocols:

    • https

      Example

      fetchTree {
        type = "file";
        url = "https://example.com/index.html";
      }
      

    • http

      Insecure HTTP transfer for legacy sources.

      Warning

      HTTP performs no encryption or authentication. Use a narHash known in advance to ensure the output has expected contents.

    • file

      A file on the local file system.

      Example

      fetchTree {
        type = "file";
        url = "file:///home/eelco/nix/README.md";
      }
      

• "tarball"

  Download a tar archive and extract it into the Nix store. This has the same underyling implementation as builtins.fetchTarball

  • url (String, required)

    Example

    fetchTree {
      type = "tarball";
      url = "https://github.com/NixOS/nixpkgs/tarball/nixpkgs-23.11";
    }
    

• "git"

  Fetch a Git tree and copy it to the Nix store. This is similar to builtins.fetchGit.

  • url (String, required)

    The URL formats supported are the same as for Git itself.

    Example

    fetchTree {
      type = "git";
      url = "git@github.com:NixOS/nixpkgs.git";
    }
    

    Note

    If the URL points to a local directory, and no ref or rev is given, Nix will only consider files added to the Git index, as listed by git ls-files but use the current file contents of the Git working directory.

  • ref (String, optional)

    By default, this has no effect. This becomes relevant only once shallow cloning is disabled.

    A Git reference, such as a branch or tag name.

    Default: "HEAD"

  • rev (String, optional)

    A Git revision; a commit hash.

    Default: the tip of ref

  • shallow (Bool, optional)

    Make a shallow clone when fetching the Git tree. When this is enabled, the options ref and allRefs have no effect anymore.

    Default: true

  • submodules (Bool, optional)

    Also fetch submodules if available.

    Default: false

  • allRefs (Bool, optional)

    By default, this has no effect. This becomes relevant only once shallow cloning is disabled.

    Whether to fetch all references (eg. branches and tags) of the repository. With this argument being true, it's possible to load a rev from any ref. (Without setting this option, only revs from the specified ref are supported).

    Default: false

  • lastModified (Integer, optional)

    Unix timestamp of the fetched commit.

    If set, pass through the value to the output attribute set. Otherwise, generated from the fetched Git tree.

  • revCount (Integer, optional)

    Number of revisions in the history of the Git repository before the fetched commit.

    If set, pass through the value to the output attribute set. Otherwise, generated from the fetched Git tree.

The following input types are still subject to change:

• "path"
• "github"
• "gitlab"
• "sourcehut"
• "mercurial"

input can also be a URL-like reference. The additional input types and the URL-like syntax requires the flakes experimental feature to be enabled.

Example

Fetch a GitHub repository using the attribute set representation:

builtins.fetchTree {
  type = "github";
  owner = "NixOS";
  repo = "nixpkgs";
  rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
}

This evaluates to the following attribute set:

{
  lastModified = 1686503798;
  lastModifiedDate = "20230611171638";
  narHash = "sha256-rA9RqKP9OlBrgGCPvfd5HVAXDOy8k2SmPtB/ijShNXc=";
  outPath = "/nix/store/l5m6qlvfs9sdw14ja3qbzpglcjlb6j1x-source";
  rev = "ae2e6b3958682513d28f7d633734571fb18285dd";
  shortRev = "ae2e6b3";
}

Example

Fetch the same GitHub repository using the URL-like syntax:

builtins.fetchTree "github:NixOS/nixpkgs/ae2e6b3958682513d28f7d633734571fb18285dd"

fetchurl url

Download the specified URL and return the path of the downloaded file.

Not available in restricted evaluation mode.

filter f list

Return a list consisting of the elements of list for which the function f returns true.

filterSource e1 e2

Warning

filterSource should not be used to filter store paths. Since filterSource uses the name of the input directory while naming the output directory, doing so will produce a directory name in the form of <hash2>-<hash>-<name>, where <hash>-<name> is the name of the input directory. Since <hash> depends on the unfiltered directory, the name of the output directory will indirectly depend on files that are filtered out by the function. This will trigger a rebuild even when a filtered out file is changed. Use builtins.path instead, which allows specifying the name of the output directory.

This function allows you to copy sources into the Nix store while filtering certain files. For instance, suppose that you want to use the directory source-dir as an input to a Nix expression, e.g.

stdenv.mkDerivation {
  ...
  src = ./source-dir;
}

However, if source-dir is a Subversion working copy, then all those annoying .svn subdirectories will also be copied to the store. Worse, the contents of those directories may change a lot, causing lots of spurious rebuilds. With filterSource you can filter out the .svn directories:

src = builtins.filterSource
  (path: type: type != "directory" || baseNameOf path != ".svn")
  ./source-dir;

Thus, the first argument e1 must be a predicate function that is called for each regular file, directory or symlink in the source tree e2. If the function returns true, the file is copied to the Nix store, otherwise it is omitted. The function is called with two arguments. The first is the full path of the file. The second is a string that identifies the type of the file, which is either "regular", "directory", "symlink" or "unknown" (for other kinds of files such as device nodes or fifos — but note that those cannot be copied to the Nix store, so if the predicate returns true for them, the copy will fail). If you exclude a directory, the entire corresponding subtree of e2 will be excluded.

findFile search-path lookup-path

Find lookup-path in search-path.

A search path is represented list of attribute sets with two attributes:

• prefix is a relative path.
• path denotes a file system location The exact syntax depends on the command line interface.

Examples of search path attribute sets:

• {
    prefix = "nixos-config";
    path = "/etc/nixos/configuration.nix";
  }
  

• {
    prefix = "";
    path = "/nix/var/nix/profiles/per-user/root/channels";
  }
  

The lookup algorithm checks each entry until a match is found, returning a path value of the match:

• If lookup-path matches prefix, then the remainder of lookup-path (the "suffix") is searched for within the directory denoted by path. Note that the path may need to be downloaded at this point to look inside.
• If the suffix is found inside that directory, then the entry is a match. The combined absolute path of the directory (now downloaded if need be) and the suffix is returned.

Lookup path expressions are desugared using this and builtins.nixPath:

<nixpkgs>

is equivalent to:

builtins.findFile builtins.nixPath "nixpkgs"

flakeRefToString attrs

Note

This function is only available if the flakes experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = flakes

Convert a flake reference from attribute set format to URL format.

For example:

builtins.flakeRefToString {
  dir = "lib"; owner = "NixOS"; ref = "23.05"; repo = "nixpkgs"; type = "github";
}

evaluates to

"github:NixOS/nixpkgs/23.05?dir=lib"

floor double

Converts an IEEE-754 double-precision floating-point number (double) to the next lower integer.

If the datatype is neither an integer nor a "float", an evaluation error will be thrown.

foldl' op nul list

Reduce a list by applying a binary operator, from left to right, e.g. foldl' op nul [x0 x1 x2 ...] = op (op (op nul x0) x1) x2) ....

For example, foldl' (acc: elem: acc + elem) 0 [1 2 3] evaluates to 6 and foldl' (acc: elem: { "${elem}" = elem; } // acc) {} ["a" "b"] evaluates to { a = "a"; b = "b"; }.

The first argument of op is the accumulator whereas the second argument is the current element being processed. The return value of each application of op is evaluated immediately, even for intermediate values.

fromJSON e

Convert a JSON string to a Nix value. For example,

builtins.fromJSON ''{"x": [1, 2, 3], "y": null}''

returns the value { x = [ 1 2 3 ]; y = null; }.

fromTOML e

Convert a TOML string to a Nix value. For example,

builtins.fromTOML ''
  x=1
  s="a"
  [table]
  y=2
''

returns the value { s = "a"; table = { y = 2; }; x = 1; }.

functionArgs f

Return a set containing the names of the formal arguments expected by the function f. The value of each attribute is a Boolean denoting whether the corresponding argument has a default value. For instance, functionArgs ({ x, y ? 123}: ...) = { x = false; y = true; }.

"Formal argument" here refers to the attributes pattern-matched by the function. Plain lambdas are not included, e.g. functionArgs (x: ...) = { }.

genList generator length

Generate list of size length, with each element i equal to the value returned by generator i. For example,

builtins.genList (x: x * x) 5

returns the list [ 0 1 4 9 16 ].

genericClosure attrset

builtins.genericClosure iteratively computes the transitive closure over an arbitrary relation defined by a function.

It takes attrset with two attributes named startSet and operator, and returns a list of attrbute sets:

• startSet: The initial list of attribute sets.

• operator: A function that takes an attribute set and returns a list of attribute sets. It defines how each item in the current set is processed and expanded into more items.

Each attribute set in the list startSet and the list returned by operator must have an attribute key, which must support equality comparison. The value of key can be one of the following types:

• Number
• Boolean
• String
• Path
• List

The result is produced by calling the operator on each item that has not been called yet, including newly added items, until no new items are added. Items are compared by their key attribute.

Common usages are:

• Generating unique collections of items, such as dependency graphs.
• Traversing through structures that may contain cycles or loops.
• Processing data structures with complex internal relationships.

Example

builtins.genericClosure {
  startSet = [ {key = 5;} ];
  operator = item: [{
    key = if (item.key / 2 ) * 2 == item.key
         then item.key / 2
         else 3 * item.key + 1;
  }];
}

evaluates to

[ { key = 5; } { key = 16; } { key = 8; } { key = 4; } { key = 2; } { key = 1; } ]

getAttr s set

getAttr returns the attribute named s from set. Evaluation aborts if the attribute doesn’t exist. This is a dynamic version of the . operator, since s is an expression rather than an identifier.

getContext s

Return the string context of s.

The string context tracks references to derivations within a string. It is represented as an attribute set of store derivation paths mapping to output names.

Using string interpolation on a derivation will add that derivation to the string context. For example,

builtins.getContext "${derivation { name = "a"; builder = "b"; system = "c"; }}"

evaluates to

{ "/nix/store/arhvjaf6zmlyn8vh8fgn55rpwnxq0n7l-a.drv" = { outputs = [ "out" ]; }; }

getEnv s

getEnv returns the value of the environment variable s, or an empty string if the variable doesn’t exist. This function should be used with care, as it can introduce all sorts of nasty environment dependencies in your Nix expression.

getEnv is used in Nix Packages to locate the file ~/.nixpkgs/config.nix, which contains user-local settings for Nix Packages. (That is, it does a getEnv "HOME" to locate the user’s home directory.)

getFlake args

Note

This function is only available if the flakes experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = flakes

Fetch a flake from a flake reference, and return its output attributes and some metadata. For example:

(builtins.getFlake "nix/55bc52401966fbffa525c574c14f67b00bc4fb3a").packages.x86_64-linux.nix

Unless impure evaluation is allowed (--impure), the flake reference must be "locked", e.g. contain a Git revision or content hash. An example of an unlocked usage is:

(builtins.getFlake "github:edolstra/dwarffs").rev

groupBy f list

Groups elements of list together by the string returned from the function f called on each element. It returns an attribute set where each attribute value contains the elements of list that are mapped to the same corresponding attribute name returned by f.

For example,

builtins.groupBy (builtins.substring 0 1) ["foo" "bar" "baz"]

evaluates to

{ b = [ "bar" "baz" ]; f = [ "foo" ]; }

hasAttr s set

hasAttr returns true if set has an attribute named s, and false otherwise. This is a dynamic version of the ? operator, since s is an expression rather than an identifier.

hasContext s

Return true if string s has a non-empty context. The context can be obtained with getContext.

Example

Many operations require a string context to be empty because they are intended only to work with "regular" strings, and also to help users avoid unintentionally loosing track of string context elements. builtins.hasContext can help create better domain-specific errors in those case.

name: meta:

if builtins.hasContext name
then throw "package name cannot contain string context"
else { ${name} = meta; }

hashFile type p

Return a base-16 representation of the cryptographic hash of the file at path p. The hash algorithm specified by type must be one of "md5", "sha1", "sha256" or "sha512".

hashString type s

Return a base-16 representation of the cryptographic hash of string s. The hash algorithm specified by type must be one of "md5", "sha1", "sha256" or "sha512".

head list

Return the first element of a list; abort evaluation if the argument isn’t a list or is an empty list. You can test whether a list is empty by comparing it with [].

import path

Load, parse, and return the Nix expression in the file path.

Note

Unlike some languages, import is a regular function in Nix.

The path argument must meet the same criteria as an interpolated expression.

If path is a directory, the file default.nix in that directory is used if it exists.

Example

$ echo 123 > default.nix

Import default.nix from the current directory.

import ./.

123

Evaluation aborts if the file doesn’t exist or contains an invalid Nix expression.

A Nix expression loaded by import must not contain any free variables, that is, identifiers that are not defined in the Nix expression itself and are not built-in. Therefore, it cannot refer to variables that are in scope at the call site.

Example

If you have a calling expression

rec {
  x = 123;
  y = import ./foo.nix;
}

then the following foo.nix will give an error:

# foo.nix
x + 456

since x is not in scope in foo.nix. If you want x to be available in foo.nix, pass it as a function argument:

rec {
  x = 123;
  y = import ./foo.nix x;
}

and

# foo.nix
x: x + 456

The function argument doesn’t have to be called x in foo.nix; any name would work.

intersectAttrs e1 e2

Return a set consisting of the attributes in the set e2 which have the same name as some attribute in e1.

Performs in O(n log m) where n is the size of the smaller set and m the larger set's size.

isAttrs e

Return true if e evaluates to a set, and false otherwise.

isBool e

Return true if e evaluates to a bool, and false otherwise.

isFloat e

Return true if e evaluates to a float, and false otherwise.

isFunction e

Return true if e evaluates to a function, and false otherwise.

isInt e

Return true if e evaluates to an integer, and false otherwise.

isList e

Return true if e evaluates to a list, and false otherwise.

isNull e

Return true if e evaluates to null, and false otherwise.

This is equivalent to e == null.

isPath e

Return true if e evaluates to a path, and false otherwise.

isString e

Return true if e evaluates to a string, and false otherwise.

length e

Return the length of the list e.

lessThan e1 e2

Return true if the number e1 is less than the number e2, and false otherwise. Evaluation aborts if either e1 or e2 does not evaluate to a number.

listToAttrs e

Construct a set from a list specifying the names and values of each attribute. Each element of the list should be a set consisting of a string-valued attribute name specifying the name of the attribute, and an attribute value specifying its value.

In case of duplicate occurrences of the same name, the first takes precedence.

Example:

builtins.listToAttrs
  [ { name = "foo"; value = 123; }
    { name = "bar"; value = 456; }
    { name = "bar"; value = 420; }
  ]

evaluates to

{ foo = 123; bar = 456; }

map f list

Apply the function f to each element in the list list. For example,

map (x: "foo" + x) [ "bar" "bla" "abc" ]

evaluates to [ "foobar" "foobla" "fooabc" ].

mapAttrs f attrset

Apply function f to every element of attrset. For example,

builtins.mapAttrs (name: value: value * 10) { a = 1; b = 2; }

evaluates to { a = 10; b = 20; }.

match regex str

Returns a list if the extended POSIX regular expression regex matches str precisely, otherwise returns null. Each item in the list is a regex group.

builtins.match "ab" "abc"

Evaluates to null.

builtins.match "abc" "abc"

Evaluates to [ ].

builtins.match "a(b)(c)" "abc"

Evaluates to [ "b" "c" ].

builtins.match "[[:space:]]+([[:upper:]]+)[[:space:]]+" "  FOO   "

Evaluates to [ "FOO" ].

mul e1 e2

Return the product of the numbers e1 and e2.

outputOf derivation-reference output-name

Note

This function is only available if the dynamic-derivations experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = dynamic-derivations

Return the output path of a derivation, literally or using a placeholder if needed.

If the derivation has a statically-known output path (i.e. the derivation output is input-addressed, or fixed content-addresed), the output path will just be returned. But if the derivation is content-addressed or if the derivation is itself not-statically produced (i.e. is the output of another derivation), a placeholder will be returned instead.

derivation reference must be a string that may contain a regular store path to a derivation, or may be a placeholder reference. If the derivation is produced by a derivation, you must explicitly select drv.outPath. This primop can be chained arbitrarily deeply. For instance,

builtins.outputOf
  (builtins.outputOf myDrv "out")
  "out"

will return a placeholder for the output of the output of myDrv.

This primop corresponds to the ^ sigil for derivable paths, e.g. as part of installable syntax on the command line.

parseDrvName s

Split the string s into a package name and version. The package name is everything up to but not including the first dash not followed by a letter, and the version is everything following that dash. The result is returned in a set { name, version }. Thus, builtins.parseDrvName "nix-0.12pre12876" returns { name = "nix"; version = "0.12pre12876"; }.

parseFlakeRef flake-ref

Note

This function is only available if the flakes experimental feature is enabled.

For example, include the following in nix.conf:

extra-experimental-features = flakes

Parse a flake reference, and return its exploded form.

For example:

builtins.parseFlakeRef "github:NixOS/nixpkgs/23.05?dir=lib"

evaluates to:

{ dir = "lib"; owner = "NixOS"; ref = "23.05"; repo = "nixpkgs"; type = "github"; }

partition pred list

Given a predicate function pred, this function returns an attrset containing a list named right, containing the elements in list for which pred returned true, and a list named wrong, containing the elements for which it returned false. For example,

builtins.partition (x: x > 10) [1 23 9 3 42]

evaluates to

{ right = [ 23 42 ]; wrong = [ 1 9 3 ]; }

path args

An enrichment of the built-in path type, based on the attributes present in args. All are optional except path:

• path
  The underlying path.

• name
  The name of the path when added to the store. This can used to reference paths that have nix-illegal characters in their names, like @.

• filter
  A function of the type expected by builtins.filterSource, with the same semantics.

• recursive
  When false, when path is added to the store it is with a flat hash, rather than a hash of the NAR serialization of the file. Thus, path must refer to a regular file, not a directory. This allows similar behavior to fetchurl. Defaults to true.

• sha256
  When provided, this is the expected hash of the file at the path. Evaluation will fail if the hash is incorrect, and providing a hash allows builtins.path to be used even when the pure-eval nix config option is on.

pathExists path

Return true if the path path exists at evaluation time, and false otherwise.

placeholder output

Return a placeholder string for the specified output that will be substituted by the corresponding output path at build time. Typical outputs would be "out", "bin" or "dev".

readDir path

Return the contents of the directory path as a set mapping directory entries to the corresponding file type. For instance, if directory A contains a regular file B and another directory C, then builtins.readDir ./A will return the set

{ B = "regular"; C = "directory"; }

The possible values for the file type are "regular", "directory", "symlink" and "unknown".

readFile path

Return the contents of the file path as a string.

readFileType p

Determine the directory entry type of a filesystem node, being one of "directory", "regular", "symlink", or "unknown".

removeAttrs set list

Remove the attributes listed in list from set. The attributes don’t have to exist in set. For instance,

removeAttrs { x = 1; y = 2; z = 3; } [ "a" "x" "z" ]

evaluates to { y = 2; }.

replaceStrings from to s

Given string s, replace every occurrence of the strings in from with the corresponding string in to.

The argument to is lazy, that is, it is only evaluated when its corresponding pattern in from is matched in the string s

Example:

builtins.replaceStrings ["oo" "a"] ["a" "i"] "foobar"

evaluates to "fabir".

seq e1 e2

Evaluate e1, then evaluate and return e2. This ensures that a computation is strict in the value of e1.

sort comparator list

Return list in sorted order. It repeatedly calls the function comparator with two elements. The comparator should return true if the first element is less than the second, and false otherwise. For example,

builtins.sort builtins.lessThan [ 483 249 526 147 42 77 ]

produces the list [ 42 77 147 249 483 526 ].

This is a stable sort: it preserves the relative order of elements deemed equal by the comparator.

split regex str

Returns a list composed of non matched strings interleaved with the lists of the extended POSIX regular expression regex matches of str. Each item in the lists of matched sequences is a regex group.

builtins.split "(a)b" "abc"

Evaluates to [ "" [ "a" ] "c" ].

builtins.split "([ac])" "abc"

Evaluates to [ "" [ "a" ] "b" [ "c" ] "" ].

builtins.split "(a)|(c)" "abc"

Evaluates to [ "" [ "a" null ] "b" [ null "c" ] "" ].

builtins.split "([[:upper:]]+)" " FOO "

Evaluates to [ " " [ "FOO" ] " " ].

splitVersion s

Split a string representing a version into its components, by the same version splitting logic underlying the version comparison in nix-env -u.

storePath path

This function allows you to define a dependency on an already existing store path. For example, the derivation attribute src = builtins.storePath /nix/store/f1d18v1y…-source causes the derivation to depend on the specified path, which must exist or be substitutable. Note that this differs from a plain path (e.g. src = /nix/store/f1d18v1y…-source) in that the latter causes the path to be copied again to the Nix store, resulting in a new path (e.g. /nix/store/ld01dnzc…-source-source).

Not available in pure evaluation mode.

See also builtins.fetchClosure.

stringLength e

Return the number of bytes of the string e. If e is not a string, evaluation is aborted.

sub e1 e2

Return the difference between the numbers e1 and e2.

substring start len s

Return the substring of s from character position start (zero-based) up to but not including start + len. If start is greater than the length of the string, an empty string is returned. If start + len lies beyond the end of the string or len is -1, only the substring up to the end of the string is returned. start must be non-negative. For example,

builtins.substring 0 3 "nixos"

evaluates to "nix".

tail list

Return the list without its first item; abort evaluation if the argument isn’t a list or is an empty list.

Warning

This function should generally be avoided since it's inefficient: unlike Haskell's tail, it takes O(n) time, so recursing over a list by repeatedly calling tail takes O(n^2) time.

throw s

Throw an error message s. This usually aborts Nix expression evaluation, but in nix-env -qa and other commands that try to evaluate a set of derivations to get information about those derivations, a derivation that throws an error is silently skipped (which is not the case for abort).

toFile name s

Store the string s in a file in the Nix store and return its path. The file has suffix name. This file can be used as an input to derivations. One application is to write builders “inline”. For instance, the following Nix expression combines the Nix expression for GNU Hello and its build script into one file:

{ stdenv, fetchurl, perl }:

stdenv.mkDerivation {
  name = "hello-2.1.1";

  builder = builtins.toFile "builder.sh" "
    source $stdenv/setup

    PATH=$perl/bin:$PATH

    tar xvfz $src
    cd hello-*
    ./configure --prefix=$out
    make
    make install
  ";

  src = fetchurl {
    url = "http://ftp.nluug.nl/pub/gnu/hello/hello-2.1.1.tar.gz";
    sha256 = "1md7jsfd8pa45z73bz1kszpp01yw6x5ljkjk2hx7wl800any6465";
  };
  inherit perl;
}

It is even possible for one file to refer to another, e.g.,

builder = let
  configFile = builtins.toFile "foo.conf" "
    # This is some dummy configuration file.
    ...
  ";
in builtins.toFile "builder.sh" "
  source $stdenv/setup
  ...
  cp ${configFile} $out/etc/foo.conf
";

Note that ${configFile} is a string interpolation, so the result of the expression configFile (i.e., a path like /nix/store/m7p7jfny445k...-foo.conf) will be spliced into the resulting string.

It is however not allowed to have files mutually referring to each other, like so:

let
  foo = builtins.toFile "foo" "...${bar}...";
  bar = builtins.toFile "bar" "...${foo}...";
in foo

This is not allowed because it would cause a cyclic dependency in the computation of the cryptographic hashes for foo and bar.

It is also not possible to reference the result of a derivation. If you are using Nixpkgs, the writeTextFile function is able to do that.

toJSON e

Return a string containing a JSON representation of e. Strings, integers, floats, booleans, nulls and lists are mapped to their JSON equivalents. Sets (except derivations) are represented as objects. Derivations are translated to a JSON string containing the derivation’s output path. Paths are copied to the store and represented as a JSON string of the resulting store path.

toPath s

DEPRECATED. Use /. + "/path" to convert a string into an absolute path. For relative paths, use ./. + "/path".

toString e

Convert the expression e to a string. e can be:

• A string (in which case the string is returned unmodified).

• A path (e.g., toString /foo/bar yields "/foo/bar".

• A set containing { __toString = self: ...; } or { outPath = ...; }.

• An integer.

• A list, in which case the string representations of its elements are joined with spaces.

• A Boolean (false yields "", true yields "1").

• null, which yields the empty string.

toXML e

Return a string containing an XML representation of e. The main application for toXML is to communicate information with the builder in a more structured format than plain environment variables.

Here is an example where this is the case:

{ stdenv, fetchurl, libxslt, jira, uberwiki }:

stdenv.mkDerivation (rec {
  name = "web-server";

  buildInputs = [ libxslt ];

  builder = builtins.toFile "builder.sh" "
    source $stdenv/setup
    mkdir $out
    echo "$servlets" | xsltproc ${stylesheet} - > $out/server-conf.xml ①
  ";

  stylesheet = builtins.toFile "stylesheet.xsl" ②
   "<?xml version='1.0' encoding='UTF-8'?>
    <xsl:stylesheet xmlns:xsl='http://www.w3.org/1999/XSL/Transform' version='1.0'>
      <xsl:template match='/'>
        <Configure>
          <xsl:for-each select='/expr/list/attrs'>
            <Call name='addWebApplication'>
              <Arg><xsl:value-of select=\"attr[@name = 'path']/string/@value\" /></Arg>
              <Arg><xsl:value-of select=\"attr[@name = 'war']/path/@value\" /></Arg>
            </Call>
          </xsl:for-each>
        </Configure>
      </xsl:template>
    </xsl:stylesheet>
  ";

  servlets = builtins.toXML [ ③
    { path = "/bugtracker"; war = jira + "/lib/atlassian-jira.war"; }
    { path = "/wiki"; war = uberwiki + "/uberwiki.war"; }
  ];
})

The builder is supposed to generate the configuration file for a Jetty servlet container. A servlet container contains a number of servlets (*.war files) each exported under a specific URI prefix. So the servlet configuration is a list of sets containing the path and war of the servlet (①). This kind of information is difficult to communicate with the normal method of passing information through an environment variable, which just concatenates everything together into a string (which might just work in this case, but wouldn’t work if fields are optional or contain lists themselves). Instead the Nix expression is converted to an XML representation with toXML, which is unambiguous and can easily be processed with the appropriate tools. For instance, in the example an XSLT stylesheet (at point ②) is applied to it (at point ①) to generate the XML configuration file for the Jetty server. The XML representation produced at point ③ by toXML is as follows:

<?xml version='1.0' encoding='utf-8'?>
<expr>
  <list>
    <attrs>
      <attr name="path">
        <string value="/bugtracker" />
      </attr>
      <attr name="war">
        <path value="/nix/store/d1jh9pasa7k2...-jira/lib/atlassian-jira.war" />
      </attr>
    </attrs>
    <attrs>
      <attr name="path">
        <string value="/wiki" />
      </attr>
      <attr name="war">
        <path value="/nix/store/y6423b1yi4sx...-uberwiki/uberwiki.war" />
      </attr>
    </attrs>
  </list>
</expr>

Note that we used the toFile built-in to write the builder and the stylesheet “inline” in the Nix expression. The path of the stylesheet is spliced into the builder using the syntax xsltproc ${stylesheet}.

trace e1 e2

Evaluate e1 and print its abstract syntax representation on standard error. Then return e2. This function is useful for debugging.

If the debugger-on-trace option is set to true and the --debugger flag is given, the interactive debugger will be started when trace is called (like break).

traceVerbose e1 e2

Evaluate e1 and print its abstract syntax representation on standard error if --trace-verbose is enabled. Then return e2. This function is useful for debugging.

tryEval e

Try to shallowly evaluate e. Return a set containing the attributes success (true if e evaluated successfully, false if an error was thrown) and value, equalling e if successful and false otherwise. tryEval will only prevent errors created by throw or assert from being thrown. Errors tryEval will not catch are for example those created by abort and type errors generated by builtins. Also note that this doesn't evaluate e deeply, so let e = { x = throw ""; }; in (builtins.tryEval e).success will be true. Using builtins.deepSeq one can get the expected result: let e = { x = throw ""; }; in (builtins.tryEval (builtins.deepSeq e e)).success will be false.

typeOf e

Return a string representing the type of the value e, namely "int", "bool", "string", "path", "null", "set", "list", "lambda" or "float".

unsafeDiscardOutputDependency s

Create a copy of the given string where every derivation deep string context element is turned into a constant string context element.

This is the opposite of builtins.addDrvOutputDependencies.

This is unsafe because it allows us to "forget" store objects we would have otherwise refered to with the string context, whereas Nix normally tracks all dependencies consistently. Safe operations "grow" but never "shrink" string contexts. builtins.addDrvOutputDependencies in contrast is safe because "derivation deep" string context element always refers to the underlying derivation (among many more things). Replacing a constant string context element with a "derivation deep" element is a safe operation that just enlargens the string context without forgetting anything.

unsafeDiscardStringContext s

Discard the string context from a value that can be coerced to a string.

warn e1 e2

Evaluate e1, which must be a string and print iton standard error as a warning. Then return e2. This function is useful for non-critical situations where attention is advisable.

If the debugger-on-trace or debugger-on-warn option is set to true and the --debugger flag is given, the interactive debugger will be started when warn is called (like break).

If the abort-on-warn option is set, the evaluation will be aborted after the warning is printed. This is useful to reveal the stack trace of the warning, when the context is non-interactive and a debugger can not be launched.

zipAttrsWith f list

Transpose a list of attribute sets into an attribute set of lists, then apply mapAttrs.

f receives two arguments: the attribute name and a non-empty list of all values encountered for that attribute name.

The result is an attribute set where the attribute names are the union of the attribute names in each element of list. The attribute values are the return values of f.

builtins.zipAttrsWith
  (name: values: { inherit name values; })
  [ { a = "x"; } { a = "y"; b = "z"; } ]

evaluates to

{
  a = { name = "a"; values = [ "x" "y" ]; };
  b = { name = "b"; values = [ "z" ]; };
}

This chapter discusses how to do package management with Nix, i.e., how to obtain, install, upgrade, and erase packages. This is the “user’s” perspective of the Nix system — people who want to create packages should consult the chapter on the Nix language.

Profiles

Profiles and user environments are Nix’s mechanism for implementing the ability to allow different users to have different configurations, and to do atomic upgrades and rollbacks. To understand how they work, it’s useful to know a bit about how Nix works. In Nix, packages are stored in unique locations in the Nix store (typically, /nix/store). For instance, a particular version of the Subversion package might be stored in a directory /nix/store/dpmvp969yhdqs7lm2r1a3gng7pyq6vy4-subversion-1.1.3/, while another version might be stored in /nix/store/5mq2jcn36ldlmh93yj1n8s9c95pj7c5s-subversion-1.1.2. The long strings prefixed to the directory names are cryptographic hashes (to be precise, 160-bit truncations of SHA-256 hashes encoded in a base-32 notation) of all inputs involved in building the package — sources, dependencies, compiler flags, and so on. So if two packages differ in any way, they end up in different locations in the file system, so they don’t interfere with each other. Here is what a part of a typical Nix store looks like:

Of course, you wouldn’t want to type

$ /nix/store/dpmvp969yhdq...-subversion-1.1.3/bin/svn

every time you want to run Subversion. Of course we could set up the PATH environment variable to include the bin directory of every package we want to use, but this is not very convenient since changing PATH doesn’t take effect for already existing processes. The solution Nix uses is to create directory trees of symlinks to activated packages. These are called user environments and they are packages themselves (though automatically generated by nix-env), so they too reside in the Nix store. For instance, in the figure above, the user environment /nix/store/0c1p5z4kda11...-user-env contains a symlink to just Subversion 1.1.2 (arrows in the figure indicate symlinks). This would be what we would obtain if we had done

$ nix-env --install --attr nixpkgs.subversion

on a set of Nix expressions that contained Subversion 1.1.2.

This doesn’t in itself solve the problem, of course; you wouldn’t want to type /nix/store/0c1p5z4kda11...-user-env/bin/svn either. That’s why there are symlinks outside of the store that point to the user environments in the store; for instance, the symlinks default-42-link and default-43-link in the example. These are called generations since every time you perform a nix-env operation, a new user environment is generated based on the current one. For instance, generation 43 was created from generation 42 when we did

$ nix-env --install --attr nixpkgs.subversion nixpkgs.firefox

on a set of Nix expressions that contained Firefox and a new version of Subversion.

Generations are grouped together into profiles so that different users don’t interfere with each other if they don’t want to. For example:

$ ls -l /nix/var/nix/profiles/
...
lrwxrwxrwx  1 eelco ... default-42-link -> /nix/store/0c1p5z4kda11...-user-env
lrwxrwxrwx  1 eelco ... default-43-link -> /nix/store/3aw2pdyx2jfc...-user-env
lrwxrwxrwx  1 eelco ... default -> default-43-link

This shows a profile called default. The file default itself is actually a symlink that points to the current generation. When we do a nix-env operation, a new user environment and generation link are created based on the current one, and finally the default symlink is made to point at the new generation. This last step is atomic on Unix, which explains how we can do atomic upgrades. (Note that the building/installing of new packages doesn’t interfere in any way with old packages, since they are stored in different locations in the Nix store.)

If you find that you want to undo a nix-env operation, you can just do

$ nix-env --rollback

which will just make the current generation link point at the previous link. E.g., default would be made to point at default-42-link. You can also switch to a specific generation:

$ nix-env --switch-generation 43

which in this example would roll forward to generation 43 again. You can also see all available generations:

$ nix-env --list-generations

You generally wouldn’t have /nix/var/nix/profiles/some-profile/bin in your PATH. Rather, there is a symlink ~/.nix-profile that points to your current profile. This means that you should put ~/.nix-profile/bin in your PATH (and indeed, that’s what the initialisation script /nix/etc/profile.d/nix.sh does). This makes it easier to switch to a different profile. You can do that using the command nix-env --switch-profile:

$ nix-env --switch-profile /nix/var/nix/profiles/my-profile

$ nix-env --switch-profile /nix/var/nix/profiles/default

These commands switch to the my-profile and default profile, respectively. If the profile doesn’t exist, it will be created automatically. You should be careful about storing a profile in another location than the profiles directory, since otherwise it might not be used as a root of the garbage collector.

All nix-env operations work on the profile pointed to by ~/.nix-profile, but you can override this using the --profile option (abbreviation -p):

$ nix-env --profile /nix/var/nix/profiles/other-profile --install --attr nixpkgs.subversion

This will not change the ~/.nix-profile symlink.

Garbage Collection

nix-env operations such as upgrades (-u) and uninstall (-e) never actually delete packages from the system. All they do (as shown above) is to create a new user environment that no longer contains symlinks to the “deleted” packages.

Of course, since disk space is not infinite, unused packages should be removed at some point. You can do this by running the Nix garbage collector. It will remove from the Nix store any package not used (directly or indirectly) by any generation of any profile.

Note however that as long as old generations reference a package, it will not be deleted. After all, we wouldn’t be able to do a rollback otherwise. So in order for garbage collection to be effective, you should also delete (some) old generations. Of course, this should only be done if you are certain that you will not need to roll back.

To delete all old (non-current) generations of your current profile:

$ nix-env --delete-generations old

Instead of old you can also specify a list of generations, e.g.,

$ nix-env --delete-generations 10 11 14

To delete all generations older than a specified number of days (except the current generation), use the d suffix. For example,

$ nix-env --delete-generations 14d

deletes all generations older than two weeks.

After removing appropriate old generations you can run the garbage collector as follows:

$ nix-store --gc

The behaviour of the garbage collector is affected by the keep-derivations (default: true) and keep-outputs (default: false) options in the Nix configuration file. The defaults will ensure that all derivations that are build-time dependencies of garbage collector roots will be kept and that all output paths that are runtime dependencies will be kept as well. All other derivations or paths will be collected. (This is usually what you want, but while you are developing it may make sense to keep outputs to ensure that rebuild times are quick.) If you are feeling uncertain, you can also first view what files would be deleted:

$ nix-store --gc --print-dead

Likewise, the option --print-live will show the paths that won’t be deleted.

There is also a convenient little utility nix-collect-garbage, which when invoked with the -d (--delete-old) switch deletes all old generations of all profiles in /nix/var/nix/profiles. So

$ nix-collect-garbage -d

is a quick and easy way to clean up your system.

Garbage Collector Roots

The roots of the garbage collector are all store paths to which there are symlinks in the directory prefix/nix/var/nix/gcroots. For instance, the following command makes the path /nix/store/d718ef...-foo a root of the collector:

$ ln -s /nix/store/d718ef...-foo /nix/var/nix/gcroots/bar

That is, after this command, the garbage collector will not remove /nix/store/d718ef...-foo or any of its dependencies.

Subdirectories of prefix/nix/var/nix/gcroots are also searched for symlinks. Symlinks to non-store paths are followed and searched for roots, but symlinks to non-store paths inside the paths reached in that way are not followed to prevent infinite recursion.

This section lists advanced topics related to builds and builds performance

Sharing Packages Between Machines

Sometimes you want to copy a package from one machine to another. Or, you want to install some packages and you know that another machine already has some or all of those packages or their dependencies. In that case there are mechanisms to quickly copy packages between machines.

Serving a Nix store via HTTP

You can easily share the Nix store of a machine via HTTP. This allows other machines to fetch store paths from that machine to speed up installations. It uses the same binary cache mechanism that Nix usually uses to fetch pre-built binaries from https://cache.nixos.org.

The daemon that handles binary cache requests via HTTP, nix-serve, is not part of the Nix distribution, but you can install it from Nixpkgs:

$ nix-env --install --attr nixpkgs.nix-serve

You can then start the server, listening for HTTP connections on whatever port you like:

$ nix-serve -p 8080

To check whether it works, try the following on the client:

$ curl http://avalon:8080/nix-cache-info

which should print something like:

StoreDir: /nix/store
WantMassQuery: 1
Priority: 30

On the client side, you can tell Nix to use your binary cache using --substituters, e.g.:

$ nix-env --install --attr nixpkgs.firefox --substituters http://avalon:8080/

The option substituters tells Nix to use this binary cache in addition to your default caches, such as https://cache.nixos.org. Thus, for any path in the closure of Firefox, Nix will first check if the path is available on the server avalon or another binary caches. If not, it will fall back to building from source.

You can also tell Nix to always use your binary cache by adding a line to the nix.conf configuration file like this:

substituters = http://avalon:8080/ https://cache.nixos.org/

Serving a Nix store via SSH

You can tell Nix to automatically fetch needed binaries from a remote Nix store via SSH. For example, the following installs Firefox, automatically fetching any store paths in Firefox’s closure if they are available on the server avalon:

$ nix-env --install --attr nixpkgs.firefox --substituters ssh://alice@avalon

This works similar to the binary cache substituter that Nix usually uses, only using SSH instead of HTTP: if a store path P is needed, Nix will first check if it’s available in the Nix store on avalon. If not, it will fall back to using the binary cache substituter, and then to building from source.

Note

The SSH substituter currently does not allow you to enter an SSH passphrase interactively. Therefore, you should use ssh-add to load the decrypted private key into ssh-agent.

You can also copy the closure of some store path, without installing it into your profile, e.g.

$ nix-store --realise /nix/store/m85bxg…-firefox-34.0.5 --substituters
ssh://alice@avalon

This is essentially equivalent to doing

$ nix-copy-closure --from alice@avalon
/nix/store/m85bxg…-firefox-34.0.5

You can use SSH’s forced command feature to set up a restricted user account for SSH substituter access, allowing read-only access to the local Nix store, but nothing more. For example, add the following lines to sshd_config to restrict the user nix-ssh:

Match User nix-ssh
  AllowAgentForwarding no
  AllowTcpForwarding no
  PermitTTY no
  PermitTunnel no
  X11Forwarding no
  ForceCommand nix-store --serve
Match All

On NixOS, you can accomplish the same by adding the following to your configuration.nix:

nix.sshServe.enable = true;
nix.sshServe.keys = [ "ssh-dss AAAAB3NzaC1k... bob@example.org" ];

where the latter line lists the public keys of users that are allowed to connect.

Remote Builds

Nix supports remote builds, where a local Nix installation can forward Nix builds to other machines. This allows multiple builds to be performed in parallel and allows Nix to perform multi-platform builds in a semi-transparent way. For instance, if you perform a build for a x86_64-darwin on an i686-linux machine, Nix can automatically forward the build to a x86_64-darwin machine, if available.

To forward a build to a remote machine, it’s required that the remote machine is accessible via SSH and that it has Nix installed. You can test whether connecting to the remote Nix instance works, e.g.

$ nix store ping --store ssh://mac

will try to connect to the machine named mac. It is possible to specify an SSH identity file as part of the remote store URI, e.g.

$ nix store ping --store ssh://mac?ssh-key=/home/alice/my-key

Since builds should be non-interactive, the key should not have a passphrase. Alternatively, you can load identities ahead of time into ssh-agent or gpg-agent.

If you get the error

bash: nix-store: command not found
error: cannot connect to 'mac'

then you need to ensure that the PATH of non-interactive login shells contains Nix.

The list of remote build machines can be specified on the command line or in the Nix configuration file. For example, the following command allows you to build a derivation for x86_64-darwin on a Linux machine:

$ uname
Linux

$ nix build --impure \
  --expr '(with import <nixpkgs> { system = "x86_64-darwin"; }; runCommand "foo" {} "uname > $out")' \
  --builders 'ssh://mac x86_64-darwin'
[1/0/1 built, 0.0 MiB DL] building foo on ssh://mac

$ cat ./result
Darwin

It is possible to specify multiple build machines separated by a semicolon or a newline, e.g.

  --builders 'ssh://mac x86_64-darwin ; ssh://beastie x86_64-freebsd'

Remote build machines can also be configured in nix.conf, e.g.

builders = ssh://mac x86_64-darwin ; ssh://beastie x86_64-freebsd

Finally, remote build machines can be configured in a separate configuration file included in builders via the syntax @/path/to/file. For example,

builders = @/etc/nix/machines

causes the list of machines in /etc/nix/machines to be included. (This is the default.)

Tuning Cores and Jobs

Nix has two relevant settings with regards to how your CPU cores will be utilized: cores and max-jobs. This chapter will talk about what they are, how they interact, and their configuration trade-offs.

• max-jobs
  Dictates how many separate derivations will be built at the same time. If you set this to zero, the local machine will do no builds. Nix will still substitute from binary caches, and build remotely if remote builders are configured.

• cores
  Suggests how many cores each derivation should use. Similar to make -j.

The cores setting determines the value of NIX_BUILD_CORES. NIX_BUILD_CORES is equal to cores, unless cores equals 0, in which case NIX_BUILD_CORES will be the total number of cores in the system.

The maximum number of consumed cores is a simple multiplication, max-jobs * NIX_BUILD_CORES.

The balance on how to set these two independent variables depends upon each builder's workload and hardware. Here are a few example scenarios on a machine with 24 cores:

It is up to the derivations' build script to respect host's requested cores-per-build by following the value of the NIX_BUILD_CORES environment variable.

Verifying Build Reproducibility

You can use Nix's diff-hook setting to compare build results. Note that this hook is only executed if the results differ; it is not used for determining if the results are the same.

For purposes of demonstration, we'll use the following Nix file, deterministic.nix for testing:

let
  inherit (import <nixpkgs> {}) runCommand;
in {
  stable = runCommand "stable" {} ''
    touch $out
  '';

  unstable = runCommand "unstable" {} ''
    echo $RANDOM > $out
  '';
}

Additionally, nix.conf contains:

diff-hook = /etc/nix/my-diff-hook
run-diff-hook = true

where /etc/nix/my-diff-hook is an executable file containing:

#!/bin/sh
exec >&2
echo "For derivation $3:"
/run/current-system/sw/bin/diff -r "$1" "$2"

The diff hook is executed by the same user and group who ran the build. However, the diff hook does not have write access to the store path just built.

Spot-Checking Build Determinism

Verify a path which already exists in the Nix store by passing --check to the build command.

If the build passes and is deterministic, Nix will exit with a status code of 0:

$ nix-build ./deterministic.nix --attr stable
this derivation will be built:
  /nix/store/z98fasz2jqy9gs0xbvdj939p27jwda38-stable.drv
building '/nix/store/z98fasz2jqy9gs0xbvdj939p27jwda38-stable.drv'...
/nix/store/yyxlzw3vqaas7wfp04g0b1xg51f2czgq-stable

$ nix-build ./deterministic.nix --attr stable --check
checking outputs of '/nix/store/z98fasz2jqy9gs0xbvdj939p27jwda38-stable.drv'...
/nix/store/yyxlzw3vqaas7wfp04g0b1xg51f2czgq-stable

If the build is not deterministic, Nix will exit with a status code of 1:

$ nix-build ./deterministic.nix --attr unstable
this derivation will be built:
  /nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv
building '/nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv'...
/nix/store/krpqk0l9ib0ibi1d2w52z293zw455cap-unstable

$ nix-build ./deterministic.nix --attr unstable --check
checking outputs of '/nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv'...
error: derivation '/nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv' may
not be deterministic: output '/nix/store/krpqk0l9ib0ibi1d2w52z293zw455cap-unstable' differs

In the Nix daemon's log, we will now see:

For derivation /nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv:
1c1
< 8108
---
> 30204

Using --check with --keep-failed will cause Nix to keep the second build's output in a special, .check path:

$ nix-build ./deterministic.nix --attr unstable --check --keep-failed
checking outputs of '/nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv'...
note: keeping build directory '/tmp/nix-build-unstable.drv-0'
error: derivation '/nix/store/cgl13lbj1w368r5z8gywipl1ifli7dhk-unstable.drv' may
not be deterministic: output '/nix/store/krpqk0l9ib0ibi1d2w52z293zw455cap-unstable' differs
from '/nix/store/krpqk0l9ib0ibi1d2w52z293zw455cap-unstable.check'

In particular, notice the /nix/store/krpqk0l9ib0ibi1d2w52z293zw455cap-unstable.check output. Nix has copied the build results to that directory where you can examine it.

Note

Check paths are not protected against garbage collection, and this path will be deleted on the next garbage collection.

The path is guaranteed to be alive for the duration of the diff-hook's execution, but may be deleted any time after.

If the comparison is performed as part of automated tooling, please use the diff-hook or author your tooling to handle the case where the build was not deterministic and also a check path does not exist.

--check is only usable if the derivation has been built on the system already. If the derivation has not been built Nix will fail with the error:

error: some outputs of '/nix/store/hzi1h60z2qf0nb85iwnpvrai3j2w7rr6-unstable.drv' 
are not valid, so checking is not possible

Run the build without --check, and then try with --check again.

Using the post-build-hook

Implementation Caveats

Here we use the post-build hook to upload to a binary cache. This is a simple and working example, but it is not suitable for all use cases.

The post build hook program runs after each executed build, and blocks the build loop. The build loop exits if the hook program fails.

Concretely, this implementation will make Nix slow or unusable when the internet is slow or unreliable.

A more advanced implementation might pass the store paths to a user-supplied daemon or queue for processing the store paths outside of the build loop.

Prerequisites

This tutorial assumes you have configured an S3-compatible binary cache as a substituter, and that the root user's default AWS profile can upload to the bucket.

Set up a Signing Key

Use nix-store --generate-binary-cache-key to create our public and private signing keys. We will sign paths with the private key, and distribute the public key for verifying the authenticity of the paths.

# nix-store --generate-binary-cache-key example-nix-cache-1 /etc/nix/key.private /etc/nix/key.public
# cat /etc/nix/key.public
example-nix-cache-1:1/cKDz3QCCOmwcztD2eV6Coggp6rqc9DGjWv7C0G+rM=

Then update nix.conf on any machine that will access the cache. Add the cache URL to substituters and the public key to trusted-public-keys:

substituters = https://cache.nixos.org/ s3://example-nix-cache
trusted-public-keys = cache.nixos.org-1:6NCHdD59X431o0gWypbMrAURkbJ16ZPMQFGspcDShjY= example-nix-cache-1:1/cKDz3QCCOmwcztD2eV6Coggp6rqc9DGjWv7C0G+rM=

Machines that build for the cache must sign derivations using the private key. On those machines, add the path to the key file to the secret-key-files field in their nix.conf:

secret-key-files = /etc/nix/key.private

We will restart the Nix daemon in a later step.

Implementing the build hook

Write the following script to /etc/nix/upload-to-cache.sh:

#!/bin/sh

set -eu
set -f # disable globbing
export IFS=' '

echo "Uploading paths" $OUT_PATHS
exec nix copy --to "s3://example-nix-cache" $OUT_PATHS

Note

The $OUT_PATHS variable is a space-separated list of Nix store paths. In this case, we expect and want the shell to perform word splitting to make each output path its own argument to nix store sign. Nix guarantees the paths will not contain any spaces, however a store path might contain glob characters. The set -f disables globbing in the shell. If you want to upload the .drv file too, the $DRV_PATH variable is also defined for the script and works just like $OUT_PATHS.

Then make sure the hook program is executable by the root user:

# chmod +x /etc/nix/upload-to-cache.sh

Updating Nix Configuration

Edit /etc/nix/nix.conf to run our hook, by adding the following configuration snippet at the end:

post-build-hook = /etc/nix/upload-to-cache.sh

Then, restart the nix-daemon.

Testing

Build any derivation, for example:

$ nix-build --expr '(import <nixpkgs> {}).writeText "example" (builtins.toString builtins.currentTime)'
this derivation will be built:
  /nix/store/s4pnfbkalzy5qz57qs6yybna8wylkig6-example.drv
building '/nix/store/s4pnfbkalzy5qz57qs6yybna8wylkig6-example.drv'...
running post-build-hook '/home/grahamc/projects/github.com/NixOS/nix/post-hook.sh'...
post-build-hook: Signing paths /nix/store/ibcyipq5gf91838ldx40mjsp0b8w9n18-example
post-build-hook: Uploading paths /nix/store/ibcyipq5gf91838ldx40mjsp0b8w9n18-example
/nix/store/ibcyipq5gf91838ldx40mjsp0b8w9n18-example

Then delete the path from the store, and try substituting it from the binary cache:

$ rm ./result
$ nix-store --delete /nix/store/ibcyipq5gf91838ldx40mjsp0b8w9n18-example

Now, copy the path back from the cache:

$ nix-store --realise /nix/store/ibcyipq5gf91838ldx40mjsp0b8w9n18-example
copying path '/nix/store/m8bmqwrch6l3h8s0k3d673xpmipcdpsa-example from 's3://example-nix-cache'...
warning: you did not specify '--add-root'; the result might be removed by the garbage collector
/nix/store/m8bmqwrch6l3h8s0k3d673xpmipcdpsa-example

Conclusion

We now have a Nix installation configured to automatically sign and upload every local build to a remote binary cache.

Before deploying this to production, be sure to consider the implementation caveats.

This section lists commands and options that you can use when you work with Nix.

Common Options

Most Nix commands accept the following command-line options:

• --help

  Prints out a summary of the command syntax and exits.

• --version

  Prints out the Nix version number on standard output and exits.

• --verbose / -v

  Increases the level of verbosity of diagnostic messages printed on standard error. For each Nix operation, the information printed on standard output is well-defined; any diagnostic information is printed on standard error, never on standard output.

  This option may be specified repeatedly. Currently, the following verbosity levels exist:

  • 0 “Errors only”

    Only print messages explaining why the Nix invocation failed.

  • 1 “Informational”

    Print useful messages about what Nix is doing. This is the default.

  • 2 “Talkative”

    Print more informational messages.

  • 3 “Chatty”

    Print even more informational messages.

  • 4 “Debug”

    Print debug information.

  • 5 “Vomit”

    Print vast amounts of debug information.

• --quiet

  Decreases the level of verbosity of diagnostic messages printed on standard error. This is the inverse option to -v / --verbose.

  This option may be specified repeatedly. See the previous verbosity levels list.

• --log-format format

  This option can be used to change the output of the log format, with format being one of:

  • raw

    This is the raw format, as outputted by nix-build.

  • internal-json

    Outputs the logs in a structured manner.

    Warning

    While the schema itself is relatively stable, the format of the error-messages (namely of the msg-field) can change between releases.

  • bar

    Only display a progress bar during the builds.

  • bar-with-logs

    Display the raw logs, with the progress bar at the bottom.

• --no-build-output / -Q

  By default, output written by builders to standard output and standard error is echoed to the Nix command's standard error. This option suppresses this behaviour. Note that the builder's standard output and error are always written to a log file in prefix/nix/var/log/nix.

• --max-jobs / -j number

  Sets the maximum number of build jobs that Nix will perform in parallel to the specified number. Specify auto to use the number of CPUs in the system. The default is specified by the max-jobs configuration setting, which itself defaults to 1. A higher value is useful on SMP systems or to exploit I/O latency.

  Setting it to 0 disallows building on the local machine, which is useful when you want builds to happen only on remote builders.

• --cores

  Sets the value of the NIX_BUILD_CORES environment variable in the invocation of builders. Builders can use this variable at their discretion to control the maximum amount of parallelism. For instance, in Nixpkgs, if the derivation attribute enableParallelBuilding is set to true, the builder passes the -jN flag to GNU Make. It defaults to the value of the cores configuration setting, if set, or 1 otherwise. The value 0 means that the builder should use all available CPU cores in the system.

• --max-silent-time

  Sets the maximum number of seconds that a builder can go without producing any data on standard output or standard error. The default is specified by the max-silent-time configuration setting. 0 means no time-out.

• --timeout

  Sets the maximum number of seconds that a builder can run. The default is specified by the timeout configuration setting. 0 means no timeout.

• --keep-going / -k

  Keep going in case of failed builds, to the greatest extent possible. That is, if building an input of some derivation fails, Nix will still build the other inputs, but not the derivation itself. Without this option, Nix stops if any build fails (except for builds of substitutes), possibly killing builds in progress (in case of parallel or distributed builds).

• --keep-failed / -K

  Specifies that in case of a build failure, the temporary directory (usually in /tmp) in which the build takes place should not be deleted. The path of the build directory is printed as an informational message.

• --fallback

  Whenever Nix attempts to build a derivation for which substitutes are known for each output path, but realising the output paths through the substitutes fails, fall back on building the derivation.

  The most common scenario in which this is useful is when we have registered substitutes in order to perform binary distribution from, say, a network repository. If the repository is down, the realisation of the derivation will fail. When this option is specified, Nix will build the derivation instead. Thus, installation from binaries falls back on installation from source. This option is not the default since it is generally not desirable for a transient failure in obtaining the substitutes to lead to a full build from source (with the related consumption of resources).

• --readonly-mode

  When this option is used, no attempt is made to open the Nix database. Most Nix operations do need database access, so those operations will fail.

• --arg name value

  This option is accepted by nix-env, nix-instantiate, nix-shell and nix-build. When evaluating Nix expressions, the expression evaluator will automatically try to call functions that it encounters. It can automatically call functions for which every argument has a default value (e.g., { argName ? defaultValue }: ...).

  With --arg, you can also call functions that have arguments without a default value (or override a default value). That is, if the evaluator encounters a function with an argument named name, it will call it with value value.

  For instance, the top-level default.nix in Nixpkgs is actually a function:

  { # The system (e.g., `i686-linux') for which to build the packages.
    system ? builtins.currentSystem
    ...
  }: ...
  

  So if you call this Nix expression (e.g., when you do nix-env --install --attr pkgname), the function will be called automatically using the value builtins.currentSystem for the system argument. You can override this using --arg, e.g., nix-env --install --attr pkgname --arg system \"i686-freebsd\". (Note that since the argument is a Nix string literal, you have to escape the quotes.)

• --argstr name value

  This option is like --arg, only the value is not a Nix expression but a string. So instead of --arg system \"i686-linux\" (the outer quotes are to keep the shell happy) you can say --argstr system i686-linux.

• --attr / -A attrPath

  Select an attribute from the top-level Nix expression being evaluated. (nix-env, nix-instantiate, nix-build and nix-shell only.) The attribute path attrPath is a sequence of attribute names separated by dots. For instance, given a top-level Nix expression e, the attribute path xorg.xorgserver would cause the expression e.xorg.xorgserver to be used. See nix-env --install for some concrete examples.

  In addition to attribute names, you can also specify array indices. For instance, the attribute path foo.3.bar selects the bar attribute of the fourth element of the array in the foo attribute of the top-level expression.

• --expr / -E

  Interpret the command line arguments as a list of Nix expressions to be parsed and evaluated, rather than as a list of file names of Nix expressions. (nix-instantiate, nix-build and nix-shell only.)

  For nix-shell, this option is commonly used to give you a shell in which you can build the packages returned by the expression. If you want to get a shell which contain the built packages ready for use, give your expression to the nix-shell --packages convenience flag instead.

• -I path

  Add an entry to the Nix expression search path. This option may be given multiple times. Paths added through -I take precedence over NIX_PATH.

• --option name value

  Set the Nix configuration option name to value. This overrides settings in the Nix configuration file (see nix.conf5).

• --repair

  Fix corrupted or missing store paths by redownloading or rebuilding them. Note that this is slow because it requires computing a cryptographic hash of the contents of every path in the closure of the build. Also note the warning under nix-store --repair-path.

Note

See man nix.conf for overriding configuration settings with command line flags.

Common Environment Variables

Most Nix commands interpret the following environment variables:

• IN_NIX_SHELL

  Indicator that tells if the current environment was set up by nix-shell. It can have the values pure or impure.

• NIX_PATH

  A colon-separated list of directories used to look up the location of Nix expressions using paths enclosed in angle brackets (i.e., <path>), e.g. /home/eelco/Dev:/etc/nixos. It can be extended using the -I option.

  If NIX_PATH is not set at all, Nix will fall back to the following list in impure and unrestricted evaluation mode:

  1. $HOME/.nix-defexpr/channels
  2. nixpkgs=/nix/var/nix/profiles/per-user/root/channels/nixpkgs
  3. /nix/var/nix/profiles/per-user/root/channels

  If NIX_PATH is set to an empty string, resolving search paths will always fail. For example, attempting to use <nixpkgs> will produce:

  error: file 'nixpkgs' was not found in the Nix search path
  

• NIX_IGNORE_SYMLINK_STORE

  Normally, the Nix store directory (typically /nix/store) is not allowed to contain any symlink components. This is to prevent “impure” builds. Builders sometimes “canonicalise” paths by resolving all symlink components. Thus, builds on different machines (with /nix/store resolving to different locations) could yield different results. This is generally not a problem, except when builds are deployed to machines where /nix/store resolves differently. If you are sure that you’re not going to do that, you can set NIX_IGNORE_SYMLINK_STORE to 1.

  Note that if you’re symlinking the Nix store so that you can put it on another file system than the root file system, on Linux you’re better off using bind mount points, e.g.,

  $ mkdir /nix
  $ mount -o bind /mnt/otherdisk/nix /nix
  

  Consult the mount 8 manual page for details.

• NIX_STORE_DIR

  Overrides the location of the Nix store (default prefix/store).

• NIX_DATA_DIR

  Overrides the location of the Nix static data directory (default prefix/share).

• NIX_LOG_DIR

  Overrides the location of the Nix log directory (default prefix/var/log/nix).

• NIX_STATE_DIR

  Overrides the location of the Nix state directory (default prefix/var/nix).

• NIX_CONF_DIR

  Overrides the location of the system Nix configuration directory (default prefix/etc/nix).

• NIX_CONFIG

  Applies settings from Nix configuration from the environment. The content is treated as if it was read from a Nix configuration file. Settings are separated by the newline character.

• NIX_USER_CONF_FILES

  Overrides the location of the Nix user configuration files to load from.

  The default are the locations according to the XDG Base Directory Specification. See the XDG Base Directories sub-section for details.

  The variable is treated as a list separated by the : token.

• TMPDIR

  Use the specified directory to store temporary files. In particular, this includes temporary build directories; these can take up substantial amounts of disk space. The default is /tmp.

• NIX_REMOTE

  This variable should be set to daemon if you want to use the Nix daemon to execute Nix operations. This is necessary in multi-user Nix installations. If the Nix daemon's Unix socket is at some non-standard path, this variable should be set to unix://path/to/socket. Otherwise, it should be left unset.

• NIX_SHOW_STATS

  If set to 1, Nix will print some evaluation statistics, such as the number of values allocated.

• NIX_COUNT_CALLS

  If set to 1, Nix will print how often functions were called during Nix expression evaluation. This is useful for profiling your Nix expressions.

• GC_INITIAL_HEAP_SIZE

  If Nix has been configured to use the Boehm garbage collector, this variable sets the initial size of the heap in bytes. It defaults to 384 MiB. Setting it to a low value reduces memory consumption, but will increase runtime due to the overhead of garbage collection.