This document describes GNU Guix version 0.6, a functional package management tool written for the GNU system.
|• Introduction:||What is Guix about?|
|• Installation:||Installing Guix.|
|• Package Management:||Package installation, upgrade, etc.|
|• Programming Interface:||Using Guix in Scheme.|
|• Utilities:||Package management commands.|
|• GNU Distribution:||Software for your friendly GNU system.|
|• Contributing:||Your help needed!|
|• GNU Free Documentation License:||The license of this manual.|
|• Concept Index:||Concepts.|
|• Function Index:||Functions.|
GNU Guix1 is a functional package management tool for the GNU system. Package management consists of all activities that relate to building packages from sources, honoring their build-time and run-time dependencies, installing packages in user environments, upgrading installed packages to new versions or rolling back to a previous set, removing unused software packages, etc.
The term functional refers to a specific package management discipline. In Guix, the package build and installation process is seen as a function, in the mathematical sense. That function takes inputs, such as build scripts, a compiler, and libraries, and returns an installed package. As a pure function, its result depends solely on its inputs—for instance, it cannot refer to software or scripts that were not explicitly passed as inputs. A build function always produces the same result when passed a given set of inputs. It cannot alter the system’s environment in any way; for instance, it cannot create, modify, or delete files outside of its build and installation directories. This is achieved by running build processes in isolated environments (or containers), where only their explicit inputs are visible.
The result of package build functions is cached in the file system, in a special directory called the store (see The Store). Each package is installed in a directory of its own, in the store—by default under /nix/store. The directory name contains a hash of all the inputs used to build that package; thus, changing an input yields a different directory name.
This approach is the foundation of Guix’s salient features: support for transactional package upgrade and rollback, per-user installation, and garbage collection of packages (see Features).
Guix has a command-line interface, which allows users to build, install, upgrade, and remove packages, as well as a Scheme programming interface.
Last but not least, Guix is used to build a distribution of the GNU system, with many GNU and non-GNU free software packages. See GNU Distribution.
GNU Guix is available for download from its website at http://www.gnu.org/software/guix/. This section describes the software requirements of Guix, as well as how to install it and get ready to use it.
The build procedure for Guix is the same as for other GNU software, and is not covered here. Please see the files README and INSTALL in the Guix source tree for additional details.
|• Requirements:||Software needed to build and run Guix.|
|• Setting Up the Daemon:||Preparing the build daemon’s environment.|
|• Invoking guix-daemon:||Running the build daemon.|
GNU Guix depends on the following packages:
--disable-daemon was passed to
following packages are also needed:
When a working installation of the Nix package
manager is available, you
can instead configure Guix with
--disable-daemon. In that case,
Nix replaces the three dependencies above.
Guix is compatible with Nix, so it is possible to share the same store
between both. To do so, you must pass
configure not only the
--with-store-dir value, but also the same
--localstatedir value. The latter is essential because it
specifies where the database that stores metadata about the store is
located, among other things. The default values are
--disable-daemon is not required if
your goal is to share the store with Nix.
Operations such as building a package or running the garbage collector
are all performed by a specialized process, the Guix daemon, on
behalf of clients. Only the daemon may access the store and its
associated database. Thus, any operation that manipulates the store
goes through the daemon. For instance, command-line tools such as
guix package and
guix build communicate with the
daemon (via remote procedure calls) to instruct it what to do.
In a standard multi-user setup, Guix and its daemon—the
guix-daemon program—are installed by the system
administrator; /nix/store is owned by
guix-daemon runs as
root. Unprivileged users may use
Guix tools to build packages or otherwise access the store, and the
daemon will do it on their behalf, ensuring that the store is kept in a
consistent state, and allowing built packages to be shared among users.
guix-daemon runs as
root, you may not want package
build processes themselves to run as
root too, for obvious
security reasons. To avoid that, a special pool of build users
should be created for use by build processes started by the daemon.
These build users need not have a shell and a home directory: they will
just be used when the daemon drops
root privileges in build
processes. Having several such users allows the daemon to launch
distinct build processes under separate UIDs, which guarantees that they
do not interfere with each other—an essential feature since builds are
regarded as pure functions (see Introduction).
On a GNU/Linux system, a build user pool may be created like this (using
Bash syntax and the
# groupadd guix-builder # for i in `seq 1 10`; do useradd -g guix-builder -G guix-builder \ -d /var/empty -s `which nologin` \ -c "Guix build user $i" guix-builder$i; done
guix-daemon program may then be run as
# guix-daemon --build-users-group=guix-builder
This way, the daemon starts build processes in a chroot, under one of
guix-builder users. On GNU/Linux, by default, the chroot
environment contains nothing but the
Guix may also be used in a single-user setup, with
running as an unprivileged user. However, to maximize non-interference
of build processes, the daemon still needs to perform certain operations
that are restricted to
root on GNU/Linux: it should be able to
run build processes in a chroot, and to run them under different UIDs.
To that end, the
nix-setuid-helper program is provided; it is
a small C program (less than 300 lines) that, if it is made setuid
root, can be executed by the daemon to perform these operations
on its behalf. The
root-owned /etc/nix-setuid.conf file
is read by
nix-setuid-helper; it should contain exactly two
words: the user name under which the authorized
runs, and the name of the build users group.
If you are installing Guix as an unprivileged user and do not have the
ability to make nix-setuid-helper setuid-
root, it is still
possible to run
guix-daemon. However, build processes will
not be isolated from one another, and not from the rest of the system.
Thus, build processes may interfere with each other, and may access
programs, libraries, and other files available on the system—making it
much harder to view them as pure functions.
guix-daemon program implements all the functionality to
access the store. This includes launching build processes, running the
garbage collector, querying the availability of a build result, etc. It
is normally run as
root like this:
# guix-daemon --build-users-group=guix-builder
For details on how to set it up, Setting Up the Daemon.
guix-daemon launches build processes under
different UIDs, taken from the build group specified with
--build-users-group. In addition, each build process is run in a
chroot environment that only contains the subset of the store that the
build process depends on, as specified by its derivation
(see derivation), plus a set of specific
system directories. By default, the latter contains /dev and
/dev/pts. Furthermore, on GNU/Linux, the build environment is a
container: in addition to having its own file system tree, it has
a separate mount name space, its own PID name space, network name space,
etc. This helps achieve reproducible builds (see Features).
The following command-line options are supported:
Take users from group to run build processes (see build users).
Do not use substitutes for build products. That is, always build things locally instead of allowing downloads of pre-built binaries.
By default substitutes are used, unless the client—such as the
guix package command—is explicitly invoked with
When the daemon runs with
--no-substitutes, clients can still
explicitly enable substitution via the
remote procedure call (see The Store).
Cache build failures. By default, only successful builds are cached.
Use n CPU cores to build each derivation;
0 means as many
The default value is
1, but it may be overridden by clients, such
--cores option of
guix build (see Invoking guix build).
The effect is to define the
NIX_BUILD_CORES environment variable
in the build process, which can then use it to exploit internal
parallelism—for instance, by running
Allow at most n build jobs in parallel. The default value is
Produce debugging output.
This is useful to debug daemon start-up issues, but then it may be
overridden by clients, for example the
--verbosity option of
guix build (see Invoking guix build).
Add dir to the build chroot.
Doing this may change the result of build processes—for instance if they use optional dependencies found in dir when it is available, and not otherwise. For that reason, it is not recommended to do so. Instead, make sure that each derivation declares all the inputs that it needs.
Disable chroot builds.
Using this option is not recommended since, again, it would allow build processes to gain access to undeclared dependencies.
Disable compression of the build logs.
--lose-logs is used, all the build logs are kept in the
localstatedir. To save space, the daemon automatically compresses
them with bzip2 by default. This option disables that.
Disable automatic file “deduplication” in the store.
By default, files added to the store are automatically “deduplicated”: if a newly added file is identical as another one found in the store, the daemon makes the new file a hard link to the other file. This slightly increases the input/output load at the end of a build process. This option disables this.
On Linux-based systems, impersonate Linux 2.6. This means that the
uname system call will report 2.6 as the release number.
This might be helpful to build programs that (usually wrongfully) depend on the kernel version number.
Do not keep build logs. By default they are kept under
Assume system as the current system type. By default it is the
architecture/kernel pair found at configure time, such as
Listen for connections on socket, the file name of a Unix-domain socket. The default socket is localstatedir/daemon-socket/socket. This option is only useful in exceptional circumstances, such as if you need to run several daemons on the same machine.
The purpose of GNU Guix is to allow users to easily install, upgrade, and remove software packages, without having to know about their build procedure or dependencies. Guix also goes beyond this obvious set of features.
This chapter describes the main features of Guix, as well as the package management tools it provides.
|• Features:||How Guix will make your life brighter.|
|• Invoking guix package:||Package installation, removal, etc.|
|• Packages with Multiple Outputs:||Single source package, multiple outputs.|
|• Invoking guix gc:||Running the garbage collector.|
|• Invoking guix pull:||Fetching the latest Guix and distribution.|
When using Guix, each package ends up in the package store, in its
own directory—something that resembles
xxx is a base32 string.
Instead of referring to these directories, users have their own
profile, which points to the packages that they actually want to
use. These profiles are stored within each user’s home directory, at
alice installs GCC 4.7.2. As a result,
/home/alice/.guix-profile/bin/gcc points to
/nix/store/…-gcc-4.7.2/bin/gcc. Now, on the same machine,
bob had already installed GCC 4.8.0. The profile of
simply continues to point to
/nix/store/…-gcc-4.8.0/bin/gcc—i.e., both versions of GCC
coexist on the same system without any interference.
guix package command is the central tool to manage
packages (see Invoking guix package). It operates on those per-user
profiles, and can be used with normal user privileges.
The command provides the obvious install, remove, and upgrade
operations. Each invocation is actually a transaction: either
the specified operation succeeds, or nothing happens. Thus, if the
guix package process is terminated during the transaction,
or if a power outage occurs during the transaction, then the user’s
profile remains in its previous state, and remains usable.
In addition, any package transaction may be rolled back. So, if, for example, an upgrade installs a new version of a package that turns out to have a serious bug, users may roll back to the previous instance of their profile, which was known to work well. Similarly, the global system configuration is subject to transactional upgrades and roll-back (see Using the Configuration System).
All those packages in the package store may be garbage-collected. Guix can determine which packages are still referenced by the user profiles, and remove those that are provably no longer referenced (see Invoking guix gc). Users may also explicitly remove old generations of their profile so that the packages they refer to can be collected.
Finally, Guix takes a purely functional approach to package management, as described in the introduction (see Introduction). Each /nix/store package directory name contains a hash of all the inputs that were used to build that package—compiler, libraries, build scripts, etc. This direct correspondence allows users to make sure a given package installation matches the current state of their distribution. It also helps maximize build reproducibility: thanks to the isolated build environments that are used, a given build is likely to yield bit-identical files when performed on different machines (see container).
This foundation allows Guix to support transparent binary/source deployment. When a pre-built binary for a /nix/store path is available from an external source—a substitute, Guix just downloads it3; otherwise, it builds the package from source, locally.
guix package command is the tool that allows users to
install, upgrade, and remove packages, as well as rolling back to
previous configurations. It operates only on the user’s own profile,
and works with normal user privileges (see Features). Its syntax
guix package options
Primarily, options specifies the operations to be performed during the transaction. Upon completion, a new profile is created, but previous generations of the profile remain available, should the user want to roll back.
For each user, a symlink to the user’s default profile is automatically
created in $HOME/.guix-profile. This symlink always points to the
current generation of the user’s default profile. Thus, users can add
$HOME/.guix-profile/bin to their
variable, and so on.
In a multi-user setup, user profiles must be stored in a place
registered as a garbage-collector root, which
$HOME/.guix-profile points to (see Invoking guix gc). That
directory is normally
localstatedir is the value passed to
--localstatedir, and user is the user name. It must be
root, with user as the owner. When it does not
exist, or is not owned by user,
guix package emits an
error about it.
The options can be among the following:
package may specify either a simple package name, such as
guile, or a package name followed by a hyphen and version number,
guile-1.8.8. If no version number is specified, the
newest available version will be selected. In addition, package
may contain a colon, followed by the name of one of the outputs of the
package, as in
(see Packages with Multiple Outputs).
Sometimes packages have propagated inputs: these are dependencies that automatically get installed along with the required package.
An example is the GNU MPC library: its C header files refer to those of the GNU MPFR library, which in turn refer to those of the GMP library. Thus, when installing MPC, the MPFR and GMP libraries also get installed in the profile; removing MPC also removes MPFR and GMP—unless they had also been explicitly installed independently.
Besides, packages sometimes rely on the definition of environment
variables for their search paths (see explanation of
--search-paths below). Any missing or possibly incorrect
environment variable definitions are reported here.
Finally, when installing a GNU package, the tool reports the availability of a newer upstream version. In the future, it may provide the option of installing directly from the upstream version, even if that version is not yet in the distribution.
Install the package exp evaluates to.
exp must be a Scheme expression that evaluates to a
<package> object. This option is notably useful to disambiguate
between same-named variants of a package, with expressions such as
(@ (gnu packages base) guile-final).
Note that this option installs the first output of the specified package, which may be insufficient when needing a specific output of a multiple-output package.
--install, package may specify a version number
and/or output name in addition to the package name. For instance,
-r glibc:debug would remove the
debug output of
Upgrade all the installed packages. When regexp is specified, upgrade only installed packages whose name matches regexp.
Note that this upgrades package to the latest version of packages found
in the distribution currently installed. To update your distribution,
you should regularly run
guix pull (see Invoking guix pull).
Roll back to the previous generation of the profile—i.e., undo the last transaction.
When combined with options such as
--install, roll back occurs
before any other actions.
When rolling back from the first generation that actually contains installed packages, the profile is made to point to the zeroth generation, which contains no files apart from its own meta-data.
Installing, removing, or upgrading packages from a generation that has been rolled back to overwrites previous future generations. Thus, the history of a profile’s generations is always linear.
Report environment variable definitions, in Bash syntax, that may be needed in order to use the set of installed packages. These environment variables are used to specify search paths for files used by some of the installed packages.
For example, GCC needs the
environment variables to be defined so it can look for headers and
libraries in the user’s profile (see Environment Variables in Using the GNU Compiler Collection (GCC)). If GCC and, say, the C
library are installed in the profile, then
suggest setting these variables to
Use profile instead of the user’s default profile.
Show what would be done without actually doing it.
When substituting a pre-built binary fails, fall back to building packages locally.
Do not use substitutes for build products. That is, always build things locally instead of allowing downloads of pre-built binaries.
Same as for
guix build (see Invoking guix build).
Produce verbose output. In particular, emit the environment’s build log on the standard error port.
Use the bootstrap Guile to build the profile. This option is only useful to distribution developers.
In addition to these actions
guix package supports the
following options to query the current state of a profile, or the
availability of packages:
List the available packages whose synopsis or description matches
regexp. Print all the meta-data of matching packages in
recutils format (see GNU recutils databases in GNU recutils manual).
This allows specific fields to be extracted using the
command, for instance:
$ guix package -s malloc | recsel -p name,version name: glibc version: 2.17 name: libgc version: 7.2alpha6
List the currently installed packages in the specified profile, with the most recently installed packages shown last. When regexp is specified, list only installed packages whose name matches regexp.
For each installed package, print the following items, separated by
tabs: the package name, its version string, the part of the package that
is installed (for instance,
out for the default output,
include for its headers, etc.), and the path of this package in
List packages currently available in the software distribution (see GNU Distribution). When regexp is specified, list only installed packages whose name matches regexp.
For each package, print the following items separated by tabs: its name, its version string, the parts of the package (see Packages with Multiple Outputs), and the source location of its definition.
Return a list of generations along with their creation dates; for each generation, show the installed packages, with the most recently installed packages shown last. Note that the zeroth generation is never shown.
For each installed package, print the following items, separated by tabs: the name of a package, its version string, the part of the package that is installed (see Packages with Multiple Outputs), and the location of this package in the store.
When pattern is used, the command returns only matching generations. Valid patterns include:
--list-generations=1returns the first one.
--list-generations=1,8,2 outputs three generations in the
specified order. Neither spaces nor trailing commas are allowed.
--list-generations=2..9prints the specified generations and everything in between. Note that the start of a range must be lesser than its end.
It is also possible to omit the endpoint. For example,
--list-generations=2.., returns all generations starting from the
--list-generations=20dlists generations that are up to 20 days old.
When pattern is omitted, delete all generations except the current one.
This command accepts the same patterns as --list-generations.
When pattern is specified, delete the matching generations. When
pattern specifies a duration, generations older than the
specified duration match. For instance,
deletes generations that are more than one month old.
If the current generation matches, it is deleted atomically—i.e., by switching to the previous available generation. Note that the zeroth generation is never deleted.
Note that deleting generations prevents roll-back to them. Consequently, this command must be used with care.
Often, packages defined in Guix have a single output—i.e., the
source package leads exactly one directory in the store. When running
guix package -i glibc, one installs the default output of the
GNU libc package; the default output is called
out, but its name
can be omitted as shown in this command. In this particular case, the
default output of
glibc contains all the C header files, shared
libraries, static libraries, Info documentation, and other supporting
Sometimes it is more appropriate to separate the various types of files
produced from a single source package into separate outputs. For
instance, the GLib C library (used by GTK+ and related packages)
installs more than 20 MiB of reference documentation as HTML pages.
To save space for users who do not need it, the documentation goes to a
separate output, called
doc. To install the main GLib output,
which contains everything but the documentation, one would run:
guix package -i glib
The command to install its documentation is:
guix package -i glib:doc
Some packages install programs with different “dependency footprints”. For instance, the WordNet package install both command-line tools and graphical user interfaces (GUIs). The former depend solely on the C library, whereas the latter depend on Tcl/Tk and the underlying X libraries. In this case, we leave the command-line tools in the default output, whereas the GUIs are in a separate output. This allows users who do not need the GUIs to save space.
There are several such multiple-output packages in the GNU distribution.
Other conventional output names include
lib for libraries and
possibly header files,
bin for stand-alone programs, and
debug for debugging information (see Installing Debugging Files). The outputs of a packages are listed in the third column of
the output of
guix package --list-available (see Invoking guix package).
Packages that are installed but not used may be garbage-collected.
guix gc command allows users to explicitly run the garbage
collector to reclaim space from the /nix/store directory.
The garbage collector has a set of known roots: any file under
/nix/store reachable from a root is considered live and
cannot be deleted; any other file is considered dead and may be
deleted. The set of garbage collector roots includes default user
profiles, and may be augmented with
guix build --root, for
example (see Invoking guix build).
Prior to running
guix gc --collect-garbage to make space, it is
often useful to remove old generations from user profiles; that way, old
package builds referenced by those generations can be reclaimed. This
is achieved by running
guix package --delete-generations
(see Invoking guix package).
guix gc command has three modes of operation: it can be
used to garbage-collect any dead files (the default), to delete specific
--delete option), or to print garbage-collector
information. The available options are listed below:
Collect garbage—i.e., unreachable /nix/store files and sub-directories. This is the default operation when no option is specified.
When min is given, stop once min bytes have been collected.
min may be a number of bytes, or it may include a unit as a
suffix, such as
MiB for mebibytes and
GB for gigabytes.
When min is omitted, collect all the garbage.
Attempt to delete all the store files and directories specified as arguments. This fails if some of the files are not in the store, or if they are still live.
Show the list of dead files and directories still present in the store—i.e., files and directories no longer reachable from any root.
Show the list of live store files and directories.
In addition, the references among existing store files can be queried:
List the references (respectively, the referrers) of store files given as arguments.
List the requisites of the store files passed as arguments. Requisites include the store files themselves, their references, and the references of these, recursively. In other words, the returned list is the transitive closure of the store files.
Packages are installed or upgraded to the latest version available in
the distribution currently available on your local machine. To update
that distribution, along with the Guix tools, you must run
pull: the command downloads the latest Guix source code and package
descriptions, and deploys it.
guix package will use packages and package
versions from this just-retrieved copy of Guix. Not only that, but all
the Guix commands and Scheme modules will also be taken from that latest
guix sub-commands added by the update also
guix pull command is usually invoked with no arguments,
but it supports the following options:
Produce verbose output, writing build logs to the standard error output.
Download the source tarball of Guix from url.
By default, the tarball is taken from its canonical address at
gnu.org, for the stable branch of Guix.
Use the bootstrap Guile to build the latest Guix. This option is only useful to Guix developers.
GNU Guix provides several Scheme programming interfaces (APIs) to define, build, and query packages. The first interface allows users to write high-level package definitions. These definitions refer to familiar packaging concepts, such as the name and version of a package, its build system, and its dependencies. These definitions can then be turned into concrete build actions.
Build actions are performed by the Guix daemon, on behalf of users. In a standard setup, the daemon has write access to the store—the /nix/store directory—whereas users do not. The recommended setup also has the daemon perform builds in chroots, under a specific build users, to minimize interference with the rest of the system.
Lower-level APIs are available to interact with the daemon and the store. To instruct the daemon to perform a build action, users actually provide it with a derivation. A derivation is a low-level representation of the build actions to be taken, and the environment in which they should occur—derivations are to package definitions what assembly is to C programs.
This chapter describes all these APIs in turn, starting from high-level package definitions.
|• Defining Packages:||Defining new packages.|
|• The Store:||Manipulating the package store.|
|• Derivations:||Low-level interface to package derivations.|
|• The Store Monad:||Purely functional interface to the store.|
The high-level interface to package definitions is implemented in the
(guix packages) and
(guix build-system) modules. As an
example, the package definition, or recipe, for the GNU Hello
package looks like this:
(use-modules (guix packages) (guix download) (guix build-system gnu) (guix licenses)) (define hello (package (name "hello") (version "2.8") (source (origin (method url-fetch) (uri (string-append "mirror://gnu/hello/hello-" version ".tar.gz")) (sha256 (base32 "0wqd8sjmxfskrflaxywc7gqw7sfawrfvdxd9skxawzfgyy0pzdz6")))) (build-system gnu-build-system) (inputs `(("gawk" ,gawk))) (synopsis "GNU Hello") (description "Yeah...") (home-page "http://www.gnu.org/software/hello/") (license gpl3+)))
Without being a Scheme expert, the reader may have guessed the meaning
of the various fields here. This expression binds variable hello
<package> object, which is essentially a record
(see Scheme records in GNU Guile Reference Manual).
This package object can be inspected using procedures found in the
(guix packages) module; for instance,
There are a few points worth noting in the above package definition:
sourcefield of the package is an
<origin>object. Here, the
(guix download)is used, meaning that the source is a file to be downloaded over FTP or HTTP.
mirror://gnu prefix instructs
url-fetch to use one of
the GNU mirrors defined in
sha256 field specifies the expected SHA256 hash of the file
being downloaded. It is mandatory, and allows Guix to check the
integrity of the file. The
(base32 …) form introduces the
base32 representation of the hash. You can obtain this information with
guix download (see Invoking guix download) and
hash (see Invoking guix hash).
When needed, the
origin form can also have a
listing patches to be applied, and a
snippet field giving a
Scheme expression to modify the source code.
build-systemfield is set to gnu-build-system. The gnu-build-system variable is defined in the
(guix build-system gnu)module, and is bound to a
Naturally, gnu-build-system represents the familiar GNU Build
System, and variants thereof (see configuration and
makefile conventions in GNU Coding Standards). In a
nutshell, packages using the GNU Build System may be configured, built,
and installed with the usual
./configure && make && make check &&
make install command sequence. This is what gnu-build-system
In addition, gnu-build-system ensures that the “standard” environment for GNU packages is available. This includes tools such as GCC, Coreutils, Bash, Make, Diffutils, and Patch.
inputsfield specifies inputs to the build process—i.e., build-time or run-time dependencies of the package. Here, we define an input called
"gawk"whose value is that of the gawk variable; gawk is itself bound to a
Note that GCC, Coreutils, Bash, and other essential tools do not need to be specified as inputs here. Instead, gnu-build-system takes care of ensuring that they are present.
However, any other dependencies need to be specified in the
inputs field. Any dependency not specified here will simply be
unavailable to the build process, possibly leading to a build failure.
There are other fields that package definitions may provide. Of
particular interest is the
arguments field. When specified, it
must be bound to a list of additional arguments to be passed to the
build system. For instance, the above definition could be augmented
with the following field initializer:
(arguments `(#:tests? #f #:configure-flags '("--enable-silent-rules")))
These are keyword arguments (see keyword
arguments in Guile in GNU Guile Reference Manual). They are
passed to gnu-build-system, which interprets them as meaning “do
make check”, and “run configure with the
--enable-silent-rules flag”. The value of these keyword
parameters is actually evaluated in the build stratum—i.e., by a
Guile process launched by the daemon (see Derivations).
Once a package definition is in place4, the
package may actually be built using the
guix build command-line
tool (see Invoking guix build). Eventually, updating the package
definition to a new upstream version can be partly automated by the
guix refresh command (see Invoking guix refresh).
Behind the scenes, a derivation corresponding to the
object is first computed by the
That derivation is stored in a
.drv file under /nix/store.
The build actions it prescribes may then be realized by using the
build-derivations procedure (see The Store).
<derivation> object of package for system
package must be a valid
<package> object, and system
must be a string denoting the target system type—e.g.,
"x86_64-linux" for an x86_64 Linux-based GNU system. store
must be a connection to the daemon, which operates on the store
(see The Store).
Similarly, it is possible to compute a derivation that cross-builds a package for some other system:
<derivation> object of package cross-built from
system to target.
target must be a valid GNU triplet denoting the target hardware
and operating system, such as
(see GNU configuration triplets in GNU
Configure and Build System).
Conceptually, the store is where derivations that have been successfully built are stored—by default, under /nix/store. Sub-directories in the store are referred to as store paths. The store has an associated database that contains information such has the store paths referred to by each store path, and the list of valid store paths—paths that result from a successful build.
The store is always accessed by the daemon on behalf of its clients (see Invoking guix-daemon). To manipulate the store, clients connect to the daemon over a Unix-domain socket, send it requests, and read the result—these are remote procedure calls, or RPCs.
(guix store) module provides procedures to connect to the
daemon, and to perform RPCs. These are described below.
Connect to the daemon over the Unix-domain socket at file. When reserve-space? is true, instruct it to reserve a little bit of extra space on the file system so that the garbage collector can still operate, should the disk become full. Return a server object.
file defaults to %default-socket-path, which is the normal
location given the options that were passed to
Close the connection to server.
This variable is bound to a SRFI-39 parameter, which refers to the port where build and error logs sent by the daemon should be written.
Procedures that make RPCs all take a server object as their first argument.
#t when path is a valid store path.
Add text under file name in the store, and return its store path. references is the list of store paths referred to by the resulting store path.
Build derivations (a list of
<derivation> objects or
derivation paths), and return when the worker is done building them.
#t on success.
Note that the
(guix monads) module provides a monad as well as
monadic versions of the above procedures, with the goal of making it
more convenient to work with code that accesses the store (see The Store Monad).
This section is currently incomplete.
Low-level build actions and the environment in which they are performed are represented by derivations. A derivation contain the following pieces of information:
Derivations allow clients of the daemon to communicate build actions to
the store. They exist in two forms: as an in-memory representation,
both on the client- and daemon-side, and as files in the store whose
name end in
.drv—these files are referred to as derivation
paths. Derivations paths can be passed to the
procedure to perform the build actions they prescribe (see The Store).
(guix derivations) module provides a representation of
derivations as Scheme objects, along with procedures to create and
otherwise manipulate derivations. The lowest-level primitive to create
a derivation is the
Build a derivation with the given arguments, and return the resulting
When hash, hash-algo, and hash-mode are given, a fixed-output derivation is created—i.e., one whose result is known in advance, such as a file download.
When references-graphs is true, it must be a list of file name/store path pairs. In that case, the reference graph of each store path is exported in the build environment in the corresponding file, in a simple text format.
Here’s an example with a shell script as its builder, assuming store is an open connection to the daemon, and bash points to a Bash executable in the store:
(use-modules (guix utils) (guix store) (guix derivations)) (let ((builder ; add the Bash script to the store (add-text-to-store store "my-builder.sh" "echo hello world > $out\n" '()))) (derivation store "foo" bash `("-e" ,builder) #:env-vars '(("HOME" . "/homeless")))) ⇒ #<derivation /nix/store/…-foo.drv => /nix/store/…-foo>
As can be guessed, this primitive is cumbersome to use directly. An
improved variant is
build-expression->derivation, which allows
the caller to directly pass a Guile expression as the build script:
Return a derivation that executes Scheme expression exp as a
builder for derivation name. inputs must be a list of
(name drv-path sub-drv) tuples; when sub-drv is omitted,
"out" is assumed. modules is a list of names of Guile
modules from the current search path to be copied in the store,
compiled, and made available in the load path during the execution of
((guix build utils) (guix build
exp is evaluated in an environment where
%outputs is bound
to a list of output/path pairs, and where
%build-inputs is bound
to a list of string/output-path pairs made from inputs.
Optionally, env-vars is a list of string pairs specifying the name
and value of environment variables visible to the builder. The builder
terminates by passing the result of exp to
exit; thus, when
#f, the build is considered to have failed.
exp is built using guile-for-build (a derivation). When
guile-for-build is omitted or is
#f, the value of the
%guile-for-build fluid is used instead.
derivation procedure for the meaning of references-graphs.
Here’s an example of a single-output derivation that creates a directory containing one file:
(let ((builder '(let ((out (assoc-ref %outputs "out"))) (mkdir out) ; create /nix/store/…-goo (call-with-output-file (string-append out "/test") (lambda (p) (display '(hello guix) p)))))) (build-expression->derivation store "goo" builder)) ⇒ #<derivation /nix/store/…-goo.drv => …>
Remember that the build expression passed to
build-expression->derivation is run by a separate Guile process
than the one that calls
build-expression->derivation: it is run
by a Guile process launched by the daemon, typically in a chroot. So,
while there is a single language for both the host and the build
side, there are really two strata of code: the host-side, and the
This distinction is important to keep in mind, notably when using
higher-level constructs such as gnu-build-system (see Defining Packages). For this reason, Guix modules that are meant to be used in
the build stratum are kept in the
(guix build …) name
The procedures that operate on the store described in the previous sections all take an open connection to the build daemon as their first argument. Although the underlying model is functional, they either have side effects or depend on the current state of the store.
The former is inconvenient: the connection to the build daemon has to be carried around in all those functions, making it impossible to compose functions that do not take that parameter with functions that do. The latter can be problematic: since store operations have side effects and/or depend on external state, they have to be properly sequenced.
This is where the
(guix monads) module comes in. This module
provides a framework for working with monads, and a particularly
useful monad for our uses, the store monad. Monads are a
construct that allows two things: associating “context” with values
(in our case, the context is the store), and building sequences of
computations (here computations includes accesses to the store.) Values
in a monad—values that carry this additional context—are called
monadic values; procedures that return such values are called
Consider this “normal” procedure:
(define (profile.sh store) ;; Return the name of a shell script in the store that ;; initializes the 'PATH' environment variable. (let* ((drv (package-derivation store coreutils)) (out (derivation->output-path drv))) (add-text-to-store store "profile.sh" (format #f "export PATH=~a/bin" out))))
(guix monads), it may be rewritten as a monadic function:
(define (profile.sh) ;; Same, but return a monadic value. (mlet %store-monad ((bin (package-file coreutils "bin"))) (text-file "profile.sh" (string-append "export PATH=" bin))))
There are two things to note in the second version: the
parameter is now implicit, and the monadic value returned by
package-file—a wrapper around
derivation->output-path—is bound using
instead of plain
Calling the monadic
profile.sh has no effect. To get the desired
effect, one must use
(run-with-store (open-connection) (profile.sh)) ⇒ /nix/store/...-profile.sh
The main syntactic forms to deal with monads in general are described below.
return forms in body as being
Return a monadic value that encapsulates val.
Bind monadic value mval, passing its “contents” to monadic procedure mproc6.
Bind the variables var to the monadic values mval in
body. The form (var -> val) binds var to the
“normal” value val, as per
mlet* is to
let* is to
(see Local Bindings in GNU Guile Reference Manual).
The interface to the store monad provided by
(guix monads) is as
The store monad. Values in the store monad encapsulate accesses to the
store. When its effect is needed, a value of the store monad must be
“evaluated” by passing it to the
run-with-store procedure (see
Run mval, a monadic value in the store monad, in store, an open store connection.
Return as a monadic value the absolute file name in the store of the file containing text.
value in the absolute file name of file within the output directory of package. When file is omitted, return the name of the output directory of package.
Monadic version of
Monadic version of
package-derivation (see Defining Packages).
This section describes tools primarily targeted at developers and users who write new package definitions. They complement the Scheme programming interface of Guix in a convenient way.
|• Invoking guix build:||Building packages from the command line.|
|• Invoking guix download:||Downloading a file and printing its hash.|
|• Invoking guix hash:||Computing the cryptographic hash of a file.|
|• Invoking guix refresh:||Updating package definitions.|
guix build command builds packages or derivations and
their dependencies, and prints the resulting store paths. Note that it
does not modify the user’s profile—this is the job of the
guix package command (see Invoking guix package). Thus,
it is mainly useful for distribution developers.
The general syntax is:
guix build options package-or-derivation…
package-or-derivation may be either the name of a package found in
the software distribution such as
coreutils-8.20, or a derivation such as
/nix/store/…-coreutils-8.19.drv. Alternatively, the
--expression option may be used to specify a Scheme expression
that evaluates to a package; this is useful when disambiguation among
several same-named packages or package variants is needed.
The options may be zero or more of the following:
Build the package or derivation expr evaluates to.
For example, expr may be
(@ (gnu packages guile)
guile-1.8), which unambiguously designates this specific variant of
version 1.8 of Guile.
Alternately, expr may refer to a zero-argument monadic procedure
(see The Store Monad). The procedure must return a derivation as a
monadic value, which is then passed through
Build the packages’ source derivations, rather than the packages themselves.
guix build -S gcc returns something like
/nix/store/…-gcc-4.7.2.tar.bz2, which is GCC’s source tarball.
The returned source tarball is the result of applying any patches and
code snippets specified in the package’s
origin (see Defining Packages).
Attempt to build for system—e.g.,
the host’s system type.
An example use of this is on Linux-based systems, which can emulate
different personalities. For instance, passing
--system=i686-linux on an
x86_64-linux system allows users
to build packages in a complete 32-bit environment.
Cross-build for triplet, which must be a valid GNU triplet, such
"mips64el-linux-gnu" (see GNU
configuration triplets in GNU Configure and Build System).
Return the derivation paths, not the output paths, of the given packages.
Keep the build tree of failed builds. Thus, if a build fail, its build tree is kept under /tmp, in a directory whose name is shown at the end of the build log. This is useful when debugging build issues.
Do not build the derivations.
When substituting a pre-built binary fails, fall back to building packages locally.
Do not use substitutes for build products. That is, always build things locally instead of allowing downloads of pre-built binaries.
When the build or substitution process remains silent for more than seconds, terminate it and report a build failure.
Allow the use of up to n CPU cores for the build. The special
0 means to use as many CPU cores as available.
Make file a symlink to the result, and register it as a garbage collector root.
Use the given verbosity level. level must be an integer between 0 and 5; higher means more verbose output. Setting a level of 4 or more may be helpful when debugging setup issues with the build daemon.
Return the build log file names for the given package-or-derivations, or raise an error if build logs are missing.
This works regardless of how packages or derivations are specified. For instance, the following invocations are equivalent:
guix build --log-file `guix build -d guile` guix build --log-file `guix build guile` guix build --log-file guile guix build --log-file -e '(@ (gnu packages guile) guile-2.0)'
Behind the scenes,
guix build is essentially an interface to
package-derivation procedure of the
module, and to the
build-derivations procedure of the
When writing a package definition, developers typically need to download
the package’s source tarball, compute its SHA256 hash, and write that
hash in the package definition (see Defining Packages). The
guix download tool helps with this task: it downloads a file
from the given URI, adds it to the store, and prints both its file name
in the store and its SHA256 hash.
The fact that the downloaded file is added to the store saves bandwidth:
when the developer eventually tries to build the newly defined package
guix build, the source tarball will not have to be
downloaded again because it is already in the store. It is also a
convenient way to temporarily stash files, which may be deleted
eventually (see Invoking guix gc).
guix download command supports the same URIs as used in
package definitions. In particular, it supports
https URIs (HTTP over TLS) are supported provided the
Guile bindings for GnuTLS are available in the user’s environment; when
they are not available, an error is raised.
The following option is available:
Write the hash in the format specified by fmt. For more information on the valid values for fmt, Invoking guix hash.
guix hash command computes the SHA256 hash of a file.
It is primarily a convenience tool for anyone contributing to the
distribution: it computes the cryptographic hash of a file, which can be
used in the definition of a package (see Defining Packages).
The general syntax is:
guix hash option file
guix hash has the following option:
Write the hash in the format specified by fmt.
hexadecimal can be used as well).
If the --format option is not specified,
will output the hash in
nix-base32. This representation is used
in the definitions of packages.
The primary audience of the
guix refresh command is developers
of the GNU software distribution. By default, it reports any packages
provided by the distribution that are outdated compared to the latest
upstream version, like this:
$ guix refresh gnu/packages/gettext.scm:29:13: gettext would be upgraded from 0.18.1.1 to 0.18.2.1 gnu/packages/glib.scm:77:12: glib would be upgraded from 2.34.3 to 2.37.0
It does so by browsing each package’s FTP directory and determining the highest version number of the source tarballs therein7.
--update, it modifies distribution source files to
update the version numbers and source tarball hashes of those packages’
recipes (see Defining Packages). This is achieved by downloading
each package’s latest source tarball and its associated OpenPGP
signature, authenticating the downloaded tarball against its signature
gpg, and finally computing its hash. When the public
key used to sign the tarball is missing from the user’s keyring, an
attempt is made to automatically retrieve it from a public key server;
when it’s successful, the key is added to the user’s keyring; otherwise,
guix refresh reports an error.
The following options are supported:
Update distribution source files (package recipes) in place. Defining Packages, for more information on package definitions.
Select all the packages in subset, one of
core subset refers to all the packages at the core of the
distribution—i.e., packages that are used to build “everything
else”. This includes GCC, libc, Binutils, Bash, etc. Usually,
changing one of these packages in the distribution entails a rebuild of
all the others. Thus, such updates are an inconvenience to users in
terms of build time or bandwidth used to achieve the upgrade.
non-core subset refers to the remaining packages. It is
typically useful in cases where an update of the core packages would be
guix refresh can be passed one or more package
names, as in this example:
guix refresh -u emacs idutils
The command above specifically updates the
idutils packages. The
--select option would have no
effect in this case.
The following options can be used to customize GnuPG operation:
Use host as the OpenPGP key server when importing a public key.
Use command as the GnuPG 2.x command. command is searched
Guix comes with a distribution of free software8 that forms the basis of the GNU system. This
includes core GNU packages such as GNU libc, GCC, and Binutils, as well
as many GNU and non-GNU applications. The complete list of available
packages can be browsed
on-line or by
guix package (see Invoking guix package):
guix package --list-available
Our goal is to build a practical 100% free software distribution of Linux-based and other variants of GNU, with a focus on the promotion and tight integration of GNU components, and an emphasis on programs and tools that help users exert that freedom.
The GNU distribution is currently available on the following platforms:
x86_64 architecture, Linux-Libre kernel;
Intel 32-bit architecture (IA32), Linux-Libre kernel;
little-endian 64-bit MIPS processors, specifically the Loongson series, n32 application binary interface (ABI), and Linux-Libre kernel.
For information on porting to other architectures or kernels, See Porting.
|• Installing Debugging Files:||Feeding the debugger.|
|• Package Modules:||Packages from the programmer’s viewpoint.|
|• Packaging Guidelines:||Growing the distribution.|
|• Bootstrapping:||GNU/Linux built from scratch.|
|• Porting:||Targeting another platform or kernel.|
|• System Configuration:||Configuring a GNU system.|
Building this distribution is a cooperative effort, and you are invited to join! Contributing, for information about how you can help.
Program binaries, as produced by the GCC compilers for instance, are typically written in the ELF format, with a section containing debugging information. Debugging information is what allows the debugger, GDB, to map binary code to source code; it is required to debug a compiled program in good conditions.
The problem with debugging information is that is takes up a fair amount of disk space. For example, debugging information for the GNU C Library weighs in at more than 60 MiB. Thus, as a user, keeping all the debugging info of all the installed programs is usually not an option. Yet, space savings should not come at the cost of an impediment to debugging—especially in the GNU system, which should make it easier for users to exert their computing freedom (see GNU Distribution).
Thankfully, the GNU Binary Utilities (Binutils) and GDB provide a mechanism that allows users to get the best of both worlds: debugging information can be stripped from the binaries and stored in separate files. GDB is then able to load debugging information from those files, when they are available (see Separate Debug Files in Debugging with GDB).
The GNU distribution takes advantage of this by storing debugging
information in the
lib/debug sub-directory of a separate package
output unimaginatively called
debug (see Packages with Multiple Outputs). Users can choose to install the
of a package when they need it. For instance, the following command
installs the debugging information for the GNU C Library and for GNU
guix package -i glibc:debug -i guile:debug
GDB must then be told to look for debug files in the user’s profile, by
debug-file-directory variable (consider setting it
from the ~/.gdbinit file, see Startup in Debugging with
(gdb) set debug-file-directory ~/.guix-profile/lib/debug
From there on, GDB will pick up debugging information from the
.debug files under ~/.guix-profile/lib/debug.
debug output mechanism in Guix is implemented by the
gnu-build-system (see Defining Packages). Currently, it is
opt-in—debugging information is available only for those packages
whose definition explicitly declares a
debug output. This may be
changed to opt-out in the future, if our build farm servers can handle
the load. To check whether a package has a
debug output, use
guix package --list-available (see Invoking guix package).
From a programming viewpoint, the package definitions of the
distribution are provided by Guile modules in the
...) name space (see Guile modules in GNU Guile
Reference Manual). For instance, the
(gnu packages emacs)
module exports a variable named
emacs, which is bound to a
<package> object (see Defining Packages). The
packages) module provides facilities for searching for packages.
The distribution is fully bootstrapped and self-contained:
each package is built based solely on other packages in the
distribution. The root of this dependency graph is a small set of
bootstrap binaries, provided by the
bootstrap) module. For more information on bootstrapping,
The GNU distribution is nascent and may well lack some of your favorite packages. This section describes how you can help make the distribution grow. See Contributing, for additional information on how you can help.
Free software packages are usually distributed in the form of source code tarballs—typically tar.gz files that contain all the source files. Adding a package to the distribution means essentially two things: adding a recipe that describes how to build the package, including a list of other packages required to build it, and adding package meta-data along with that recipe, such as a description and licensing information.
In Guix all this information is embodied in package definitions. Package definitions provide a high-level view of the package. They are written using the syntax of the Scheme programming language; in fact, for each package we define a variable bound to the package definition, and export that variable from a module (see Package Modules). However, in-depth Scheme knowledge is not a prerequisite for creating packages. For more information on package definitions, Defining Packages.
Once a package definition is in place, stored in a file in the Guix
source tree, it can be tested using the
guix build command
(see Invoking guix build). For example, assuming the new package is
gnew, you may run this command from the Guix build tree:
./pre-inst-env guix build gnew --keep-failed
--keep-failed makes it easier to debug build failures since
it provides access to the failed build tree.
Once your package builds correctly, please send us a patch (see Contributing). Well, if you need help, we will be happy to help you too. Once the patch is committed in the Guix repository, the new package automatically gets built on the supported platforms by our continuous integration system.
Users can obtain the new package definition simply by running
guix pull (see Invoking guix pull). When
hydra.gnu.org is done building the package, installing the
package automatically downloads binaries from there (except when using
--no-substitutes). The only place where human intervention is
needed is to review and apply the patch.
|• Software Freedom:||What may go into the distribution.|
|• Package Naming:||What’s in a name?|
|• Version Numbers:||When the name is not enough.|
|• Python Modules:||Taming the snake.|
The GNU operating system has been developed so that users can have freedom in their computing. GNU is free software, meaning that users have the four essential freedoms: to run the program, to study and change the program in source code form, to redistribute exact copies, and to distribute modified versions. Packages found in the GNU distribution provide only software that conveys these four freedoms.
In addition, the GNU distribution follow the free software distribution guidelines. Among other things, these guidelines reject non-free firmware, recommendations of non-free software, and discuss ways to deal with trademarks and patents.
Some packages contain a small and optional subset that violates the
above guidelines, for instance because this subset is itself non-free
code. When that happens, the offending items are removed with
appropriate patches or code snippets in the package definition’s
origin form (see Defining Packages). That way,
build --source returns the “freed” source rather than the unmodified
A package has actually two names associated with it:
First, there is the name of the Scheme variable, the one following
define-public. By this name, the package can be made known in the
Scheme code, for instance as input to another package. Second, there is
the string in the
name field of a package definition. This name
is used by package management commands such as
guix package and
Both are usually the same and correspond to the lowercase conversion of the
project name chosen upstream. For instance, the GNUnet project is packaged
gnunet. We do not add
lib prefixes for library packages,
unless these are already part of the official project name. But see
Python Modules for special rules concerning modules for
the Python language.
We usually package only the latest version of a given free software
project. But sometimes, for instance for incompatible library versions,
two (or more) versions of the same package are needed. These require
different Scheme variable names. We use the name as defined
in Package Naming
for the most recent version; previous versions use the same name, suffixed
- and the smallest prefix of the version number that may
distinguish the two versions.
The name inside the package definition is the same for all versions of a package and does not contain any version number.
For instance, the versions 2.24.20 and 3.9.12 of GTK+ may be packaged as follows:
(define-public gtk+ (package (name "gtk+") (version "3.9.12") ...)) (define-public gtk+-2 (package (name "gtk+") (version "2.24.20") ...))
If we also wanted GTK+ 3.8.2, this would be packaged as
(define-public gtk+-3.8 (package (name "gtk+") (version "3.8.2") ...))
We currently package Python 2 and Python 3, under the Scheme variable names
python as explained in Version Numbers.
To avoid confusion and naming clashes with other programming languages, it
seems desirable that the name of a package for a Python module contains
Some modules are compatible with only one version of Python, others with both.
If the package Foo compiles only with Python 3, we name it
python-foo; if it compiles only with Python 2, we name it
python2-foo. If it is compatible with both versions, we create two
packages with the corresponding names.
If a project already contains the word
python, we drop this;
for instance, the module python-dateutil is packaged under the names
Bootstrapping in our context refers to how the distribution gets built “from nothing”. Remember that the build environment of a derivation contains nothing but its declared inputs (see Introduction). So there’s an obvious chicken-and-egg problem: how does the first package get built? How does the first compiler get compiled? Note that this is a question of interest only to the curious hacker, not to the regular user, so you can shamelessly skip this section if you consider yourself a “regular user”.
The GNU system is primarily made of C code, with libc at its core. The
GNU build system itself assumes the availability of a Bourne shell and
command-line tools provided by GNU Coreutils, Awk, Findutils, ‘sed’, and
‘grep’. Furthermore, build programs—programs that run
make, etc.—are written in Guile Scheme
(see Derivations). Consequently, to be able to build anything at
all, from scratch, Guix relies on pre-built binaries of Guile, GCC,
Binutils, libc, and the other packages mentioned above—the
These bootstrap binaries are “taken for granted”, though we can also re-create them if needed (more on that later).
The figure above shows the very beginning of the dependency graph of the
distribution, corresponding to the package definitions of the
packages bootstrap) module. At this level of detail, things are
slightly complex. First, Guile itself consists of an ELF executable,
along with many source and compiled Scheme files that are dynamically
loaded when it runs. This gets stored in the guile-2.0.7.tar.xz
tarball shown in this graph. This tarball is part of Guix’s “source”
distribution, and gets inserted into the store with
(see The Store).
But how do we write a derivation that unpacks this tarball and adds it
to the store? To solve this problem, the
derivation—the first one that gets built—uses
bash as its
builder, which runs
build-bootstrap-guile.sh, which in turn calls
tar to unpack the tarball. Thus, bash, tar,
xz, and mkdir are statically-linked binaries, also part of
the Guix source distribution, whose sole purpose is to allow the Guile
tarball to be unpacked.
guile-bootstrap-2.0.drv is built, we have a functioning
Guile that can be used to run subsequent build programs. Its first task
is to download tarballs containing the other pre-built binaries—this
is what the
.tar.xz.drv derivations do. Guix modules such as
ftp-client.scm are used for this purpose. The
module-import.drv derivations import those modules in a directory
in the store, using the original layout. The
module-import-compiled.drv derivations compile those modules, and
write them in an output directory with the right layout. This
corresponds to the
#:modules argument of
build-expression->derivation (see Derivations).
Finally, the various tarballs are unpacked by the
etc., at which point we have a working C tool chain.
Bootstrapping is complete when we have a full tool chain that does not
depend on the pre-built bootstrap tools discussed above. This
no-dependency requirement is verified by checking whether the files of
the final tool chain contain references to the /nix/store
directories of the bootstrap inputs. The process that leads to this
“final” tool chain is described by the package definitions found in
(gnu packages base) module.
The first tool that gets built with the bootstrap binaries is GNU Make, which is a prerequisite for all the following packages. From there Findutils and Diffutils get built.
Then come the first-stage Binutils and GCC, built as pseudo cross
--target equal to
--host. They are
used to build libc. Thanks to this cross-build trick, this libc is
guaranteed not to hold any reference to the initial tool chain.
From there the final Binutils and GCC are built. GCC uses
from the final Binutils, and links programs against the just-built libc.
This tool chain is used to build the other packages used by Guix and by
the GNU Build System: Guile, Bash, Coreutils, etc.
And voilà! At this point we have the complete set of build tools that
the GNU Build System expects. These are in the
variables of the
(gnu packages base) module, and are implicitly
used by any package that uses
gnu-build-system (see Defining Packages).
Because the final tool chain does not depend on the bootstrap binaries,
those rarely need to be updated. Nevertheless, it is useful to have an
automated way to produce them, should an update occur, and this is what
(gnu packages make-bootstrap) module provides.
The following command builds the tarballs containing the bootstrap binaries (Guile, Binutils, GCC, libc, and a tarball containing a mixture of Coreutils and other basic command-line tools):
guix build bootstrap-tarballs
The generated tarballs are those that should be referred to in the
(gnu packages bootstrap) module mentioned at the beginning of
Still here? Then perhaps by now you’ve started to wonder: when do we reach a fixed point? That is an interesting question! The answer is unknown, but if you would like to investigate further (and have significant computational and storage resources to do so), then let us know.
As discussed above, the GNU distribution is self-contained, and
self-containment is achieved by relying on pre-built “bootstrap
binaries” (see Bootstrapping). These binaries are specific to an
operating system kernel, CPU architecture, and application binary
interface (ABI). Thus, to port the distribution to a platform that is
not yet supported, one must build those bootstrap binaries, and update
(gnu packages bootstrap) module to use them on that platform.
Fortunately, Guix can cross compile those bootstrap binaries. When everything goes well, and assuming the GNU tool chain supports the target platform, this can be as simple as running a command like this one:
guix build --target=armv5tel-linux-gnueabi bootstrap-tarballs
Once these are built, the
(gnu packages bootstrap) module needs
to be updated to refer to these binaries on the target platform. In
glibc-dynamic-linker procedure in that module must
be augmented to return the right file name for libc’s dynamic linker on
that platform; likewise,
packages linux) must be taught about the new platform.
In practice, there may be some complications. First, it may be that the
extended GNU triplet that specifies an ABI (like the
above) is not recognized by all the GNU tools. Typically, glibc
recognizes some of these, whereas GCC uses an extra
configure flag (see
gcc.scm for examples of how to handle this).
Second, some of the required packages could fail to build for that
platform. Lastly, the generated binaries could be broken for some
This section documents work-in-progress. As such it may be incomplete, outdated, or open to discussions. Please discuss it on firstname.lastname@example.org.
The GNU system supports a consistent whole-system configuration mechanism. By that we mean that all aspects of the global system configuration—such as the available system services, timezone and locale settings, user accounts—are declared in a single place. Such a system configuration can be instantiated—i.e., effected.
One of the advantages of putting all the system configuration under the control of Guix is that it supports transactional system upgrades, and makes it possible to roll-back to a previous system instantiation, should something go wrong with the new one (see Features). Another one is that it makes it easy to replicate the exact same configuration across different machines, or at different points in time, without having to resort to additional administration tools layered on top of the system’s own tools.
This section describes this mechanism. First we focus on the system administrator’s viewpoint—explaining how the system is configured and instantiated. Then we show how this mechanism can be extended, for instance to support new system services.
|• Using the Configuration System:||Customizing your GNU system.|
|• Defining Services:||Adding new service definitions.|
The operating system is configured by filling in an
operating-system structure, as defined by the
module. A simple setup, with the default system services, the default
Linux-Libre kernel, initial RAM disk, and boot loader looks like this:
(use-modules (gnu system) (gnu system shadow) ; for 'user-account' (gnu system service) ; for 'lsh-service' (gnu packages base) ; Coreutils, grep, etc. (gnu packages bash) ; Bash (gnu packages system) ; dmd, Inetutils (gnu packages zile) ; Zile (gnu packages less) ; less (gnu packages guile) ; Guile (gnu packages linux)) ; procps, psmisc (define komputilo (operating-system (host-name "komputilo") (timezone "Europe/Paris") (locale "fr_FR.UTF-8") (users (list (user-account (name "alice") (password "") (uid 1000) (gid 100) (comment "Bob's sister") (home-directory "/home/alice")))) (packages (list coreutils bash guile-2.0 guix dmd inetutils findutils grep sed procps psmisc zile less)) (services (cons (lsh-service #:port 2222 #:allow-root-login? #t) %standard-services))))
This example should be self-describing. The
packages field lists
packages provided by the various
(gnu packages ...) modules above
(see Package Modules). These are the packages that will be globally
visible on the system, for all user accounts—i.e., in every user’s
PATH environment variable—in addition to the per-user profiles
(see Invoking guix package).
services field lists system services to be made
available when the system starts. The %standard-services list,
(gnu system) module, provides the basic services one
would expect from a GNU system: a login service (mingetty) on each tty,
syslogd, libc’s name service cache daemon (nscd), etc.
operating-system declaration above specifies that, in
addition to those services, we want the
lshd secure shell
daemon listening on port 2222, and allowing remote
(see Invoking lshd in GNU lsh Manual). Under the hood,
lsh-service arranges so that
lshd is started with the
right command-line options, possibly with supporting configuration files
generated as needed (see Defining Services).
Assuming the above snippet is stored in the my-system-config.scm
file, the (yet unwritten!)
guix system --boot
my-system-config.scm command instantiates that configuration, and makes
it the default GRUB boot entry. The normal way to change the system’s
configuration is by updating this file and re-running the
At the Scheme level, the bulk of an
is instantiated with the following monadic procedure (see The Store Monad):
Return a derivation that builds os, an
object (see Derivations).
The output of the derivation is a single directory that refers to all the packages, configuration files, and other supporting files needed to instantiate os.
(gnu system dmd) module defines several procedures that allow
users to declare the operating system’s services (see Using the Configuration System). These procedures are monadic
procedures—i.e., procedures that return a monadic value in the store
monad (see The Store Monad). Examples of such procedures include:
return the definition of a service that runs
offer a login service on the given console tty;
return a definition for libc’s name service cache daemon (nscd);
return a definition for a service that runs
(see Invoking guix-daemon).
The monadic value returned by those procedures is a service
definition—a structure as returned by the
Service definitions specifies the inputs the service depends on, and an
expression to start and stop the service. Behind the scenes, service
definitions are “translated” into the form suitable for the
configuration file of dmd, the init system (see Services in GNU
As an example, here is what the
nscd-service procedure looks
(define (nscd-service) (mlet %store-monad ((nscd (package-file glibc "sbin/nscd"))) (return (service (documentation "Run libc's name service cache daemon.") (provision '(nscd)) (start `(make-forkexec-constructor ,nscd "-f" "/dev/null" "--foreground")) (stop `(make-kill-destructor)) (respawn? #f) (inputs `(("glibc" ,glibc)))))))
inputs field specifies that this service depends on the
glibc package—the package that contains the
stop fields are expressions that
make use of dmd’s facilities to start and stop processes (see Service
De- and Constructors in GNU dmd Manual). The
field specifies the name under which this service is known to dmd, and
documentation specifies on-line documentation. Thus, the
deco start ncsd,
deco stop nscd, and
deco doc nscd will do what you would expect (see Invoking
deco in GNU dmd Manual).
This project is a cooperative effort, and we need your help to make it grow! Please get in touch with us on email@example.com. We welcome ideas, bug reports, patches, and anything that may be helpful to the project. We particularly welcome help on packaging (see Packaging Guidelines).
Please see the HACKING file that comes with the Guix source code for practical details about contributions.
Guix is based on the Nix package manager, which was designed and implemented by Eelco Dolstra. Nix pioneered functional package management, and promoted unprecedented features, such as transactional package upgrades and rollbacks, per-user profiles, and referentially transparent build processes. Without this work, Guix would not exist.
The Nix-based software distributions, Nixpkgs and NixOS, have also been an inspiration for Guix.
Copyright © 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc. http://fsf.org/ Everyone is permitted to copy and distribute verbatim copies of this license document, but changing it is not allowed.
The purpose of this License is to make a manual, textbook, or other functional and useful document free in the sense of freedom: to assure everyone the effective freedom to copy and redistribute it, with or without modifying it, either commercially or noncommercially. Secondarily, this License preserves for the author and publisher a way to get credit for their work, while not being considered responsible for modifications made by others.
This License is a kind of “copyleft”, which means that derivative works of the document must themselves be free in the same sense. It complements the GNU General Public License, which is a copyleft license designed for free software.
We have designed this License in order to use it for manuals for free software, because free software needs free documentation: a free program should come with manuals providing the same freedoms that the software does. But this License is not limited to software manuals; it can be used for any textual work, regardless of subject matter or whether it is published as a printed book. We recommend this License principally for works whose purpose is instruction or reference.
This License applies to any manual or other work, in any medium, that contains a notice placed by the copyright holder saying it can be distributed under the terms of this License. Such a notice grants a world-wide, royalty-free license, unlimited in duration, to use that work under the conditions stated herein. The “Document”, below, refers to any such manual or work. Any member of the public is a licensee, and is addressed as “you”. You accept the license if you copy, modify or distribute the work in a way requiring permission under copyright law.
A “Modified Version” of the Document means any work containing the Document or a portion of it, either copied verbatim, or with modifications and/or translated into another language.
A “Secondary Section” is a named appendix or a front-matter section of the Document that deals exclusively with the relationship of the publishers or authors of the Document to the Document’s overall subject (or to related matters) and contains nothing that could fall directly within that overall subject. (Thus, if the Document is in part a textbook of mathematics, a Secondary Section may not explain any mathematics.) The relationship could be a matter of historical connection with the subject or with related matters, or of legal, commercial, philosophical, ethical or political position regarding them.
The “Invariant Sections” are certain Secondary Sections whose titles are designated, as being those of Invariant Sections, in the notice that says that the Document is released under this License. If a section does not fit the above definition of Secondary then it is not allowed to be designated as Invariant. The Document may contain zero Invariant Sections. If the Document does not identify any Invariant Sections then there are none.
The “Cover Texts” are certain short passages of text that are listed, as Front-Cover Texts or Back-Cover Texts, in the notice that says that the Document is released under this License. A Front-Cover Text may be at most 5 words, and a Back-Cover Text may be at most 25 words.
A “Transparent” copy of the Document means a machine-readable copy, represented in a format whose specification is available to the general public, that is suitable for revising the document straightforwardly with generic text editors or (for images composed of pixels) generic paint programs or (for drawings) some widely available drawing editor, and that is suitable for input to text formatters or for automatic translation to a variety of formats suitable for input to text formatters. A copy made in an otherwise Transparent file format whose markup, or absence of markup, has been arranged to thwart or discourage subsequent modification by readers is not Transparent. An image format is not Transparent if used for any substantial amount of text. A copy that is not “Transparent” is called “Opaque”.
Examples of suitable formats for Transparent copies include plain ASCII without markup, Texinfo input format, LaTeX input format, SGML or XML using a publicly available DTD, and standard-conforming simple HTML, PostScript or PDF designed for human modification. Examples of transparent image formats include PNG, XCF and JPG. Opaque formats include proprietary formats that can be read and edited only by proprietary word processors, SGML or XML for which the DTD and/or processing tools are not generally available, and the machine-generated HTML, PostScript or PDF produced by some word processors for output purposes only.
The “Title Page” means, for a printed book, the title page itself, plus such following pages as are needed to hold, legibly, the material this License requires to appear in the title page. For works in formats which do not have any title page as such, “Title Page” means the text near the most prominent appearance of the work’s title, preceding the beginning of the body of the text.
The “publisher” means any person or entity that distributes copies of the Document to the public.
A section “Entitled XYZ” means a named subunit of the Document whose title either is precisely XYZ or contains XYZ in parentheses following text that translates XYZ in another language. (Here XYZ stands for a specific section name mentioned below, such as “Acknowledgements”, “Dedications”, “Endorsements”, or “History”.) To “Preserve the Title” of such a section when you modify the Document means that it remains a section “Entitled XYZ” according to this definition.
The Document may include Warranty Disclaimers next to the notice which states that this License applies to the Document. These Warranty Disclaimers are considered to be included by reference in this License, but only as regards disclaiming warranties: any other implication that these Warranty Disclaimers may have is void and has no effect on the meaning of this License.
You may copy and distribute the Document in any medium, either commercially or noncommercially, provided that this License, the copyright notices, and the license notice saying this License applies to the Document are reproduced in all copies, and that you add no other conditions whatsoever to those of this License. You may not use technical measures to obstruct or control the reading or further copying of the copies you make or distribute. However, you may accept compensation in exchange for copies. If you distribute a large enough number of copies you must also follow the conditions in section 3.
You may also lend copies, under the same conditions stated above, and you may publicly display copies.
If you publish printed copies (or copies in media that commonly have printed covers) of the Document, numbering more than 100, and the Document’s license notice requires Cover Texts, you must enclose the copies in covers that carry, clearly and legibly, all these Cover Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on the back cover. Both covers must also clearly and legibly identify you as the publisher of these copies. The front cover must present the full title with all words of the title equally prominent and visible. You may add other material on the covers in addition. Copying with changes limited to the covers, as long as they preserve the title of the Document and satisfy these conditions, can be treated as verbatim copying in other respects.
If the required texts for either cover are too voluminous to fit legibly, you should put the first ones listed (as many as fit reasonably) on the actual cover, and continue the rest onto adjacent pages.
If you publish or distribute Opaque copies of the Document numbering more than 100, you must either include a machine-readable Transparent copy along with each Opaque copy, or state in or with each Opaque copy a computer-network location from which the general network-using public has access to download using public-standard network protocols a complete Transparent copy of the Document, free of added material. If you use the latter option, you must take reasonably prudent steps, when you begin distribution of Opaque copies in quantity, to ensure that this Transparent copy will remain thus accessible at the stated location until at least one year after the last time you distribute an Opaque copy (directly or through your agents or retailers) of that edition to the public.
It is requested, but not required, that you contact the authors of the Document well before redistributing any large number of copies, to give them a chance to provide you with an updated version of the Document.
You may copy and distribute a Modified Version of the Document under the conditions of sections 2 and 3 above, provided that you release the Modified Version under precisely this License, with the Modified Version filling the role of the Document, thus licensing distribution and modification of the Modified Version to whoever possesses a copy of it. In addition, you must do these things in the Modified Version:
If the Modified Version includes new front-matter sections or appendices that qualify as Secondary Sections and contain no material copied from the Document, you may at your option designate some or all of these sections as invariant. To do this, add their titles to the list of Invariant Sections in the Modified Version’s license notice. These titles must be distinct from any other section titles.
You may add a section Entitled “Endorsements”, provided it contains nothing but endorsements of your Modified Version by various parties—for example, statements of peer review or that the text has been approved by an organization as the authoritative definition of a standard.
You may add a passage of up to five words as a Front-Cover Text, and a passage of up to 25 words as a Back-Cover Text, to the end of the list of Cover Texts in the Modified Version. Only one passage of Front-Cover Text and one of Back-Cover Text may be added by (or through arrangements made by) any one entity. If the Document already includes a cover text for the same cover, previously added by you or by arrangement made by the same entity you are acting on behalf of, you may not add another; but you may replace the old one, on explicit permission from the previous publisher that added the old one.
The author(s) and publisher(s) of the Document do not by this License give permission to use their names for publicity for or to assert or imply endorsement of any Modified Version.
You may combine the Document with other documents released under this License, under the terms defined in section 4 above for modified versions, provided that you include in the combination all of the Invariant Sections of all of the original documents, unmodified, and list them all as Invariant Sections of your combined work in its license notice, and that you preserve all their Warranty Disclaimers.
The combined work need only contain one copy of this License, and multiple identical Invariant Sections may be replaced with a single copy. If there are multiple Invariant Sections with the same name but different contents, make the title of each such section unique by adding at the end of it, in parentheses, the name of the original author or publisher of that section if known, or else a unique number. Make the same adjustment to the section titles in the list of Invariant Sections in the license notice of the combined work.
In the combination, you must combine any sections Entitled “History” in the various original documents, forming one section Entitled “History”; likewise combine any sections Entitled “Acknowledgements”, and any sections Entitled “Dedications”. You must delete all sections Entitled “Endorsements.”
You may make a collection consisting of the Document and other documents released under this License, and replace the individual copies of this License in the various documents with a single copy that is included in the collection, provided that you follow the rules of this License for verbatim copying of each of the documents in all other respects.
You may extract a single document from such a collection, and distribute it individually under this License, provided you insert a copy of this License into the extracted document, and follow this License in all other respects regarding verbatim copying of that document.
A compilation of the Document or its derivatives with other separate and independent documents or works, in or on a volume of a storage or distribution medium, is called an “aggregate” if the copyright resulting from the compilation is not used to limit the legal rights of the compilation’s users beyond what the individual works permit. When the Document is included in an aggregate, this License does not apply to the other works in the aggregate which are not themselves derivative works of the Document.
If the Cover Text requirement of section 3 is applicable to these copies of the Document, then if the Document is less than one half of the entire aggregate, the Document’s Cover Texts may be placed on covers that bracket the Document within the aggregate, or the electronic equivalent of covers if the Document is in electronic form. Otherwise they must appear on printed covers that bracket the whole aggregate.
Translation is considered a kind of modification, so you may distribute translations of the Document under the terms of section 4. Replacing Invariant Sections with translations requires special permission from their copyright holders, but you may include translations of some or all Invariant Sections in addition to the original versions of these Invariant Sections. You may include a translation of this License, and all the license notices in the Document, and any Warranty Disclaimers, provided that you also include the original English version of this License and the original versions of those notices and disclaimers. In case of a disagreement between the translation and the original version of this License or a notice or disclaimer, the original version will prevail.
If a section in the Document is Entitled “Acknowledgements”, “Dedications”, or “History”, the requirement (section 4) to Preserve its Title (section 1) will typically require changing the actual title.
You may not copy, modify, sublicense, or distribute the Document except as expressly provided under this License. Any attempt otherwise to copy, modify, sublicense, or distribute it is void, and will automatically terminate your rights under this License.
However, if you cease all violation of this License, then your license from a particular copyright holder is reinstated (a) provisionally, unless and until the copyright holder explicitly and finally terminates your license, and (b) permanently, if the copyright holder fails to notify you of the violation by some reasonable means prior to 60 days after the cessation.
Moreover, your license from a particular copyright holder is reinstated permanently if the copyright holder notifies you of the violation by some reasonable means, this is the first time you have received notice of violation of this License (for any work) from that copyright holder, and you cure the violation prior to 30 days after your receipt of the notice.
Termination of your rights under this section does not terminate the licenses of parties who have received copies or rights from you under this License. If your rights have been terminated and not permanently reinstated, receipt of a copy of some or all of the same material does not give you any rights to use it.
The Free Software Foundation may publish new, revised versions of the GNU Free Documentation License from time to time. Such new versions will be similar in spirit to the present version, but may differ in detail to address new problems or concerns. See http://www.gnu.org/copyleft/.
Each version of the License is given a distinguishing version number. If the Document specifies that a particular numbered version of this License “or any later version” applies to it, you have the option of following the terms and conditions either of that specified version or of any later version that has been published (not as a draft) by the Free Software Foundation. If the Document does not specify a version number of this License, you may choose any version ever published (not as a draft) by the Free Software Foundation. If the Document specifies that a proxy can decide which future versions of this License can be used, that proxy’s public statement of acceptance of a version permanently authorizes you to choose that version for the Document.
“Massive Multiauthor Collaboration Site” (or “MMC Site”) means any World Wide Web server that publishes copyrightable works and also provides prominent facilities for anybody to edit those works. A public wiki that anybody can edit is an example of such a server. A “Massive Multiauthor Collaboration” (or “MMC”) contained in the site means any set of copyrightable works thus published on the MMC site.
“CC-BY-SA” means the Creative Commons Attribution-Share Alike 3.0 license published by Creative Commons Corporation, a not-for-profit corporation with a principal place of business in San Francisco, California, as well as future copyleft versions of that license published by that same organization.
“Incorporate” means to publish or republish a Document, in whole or in part, as part of another Document.
An MMC is “eligible for relicensing” if it is licensed under this License, and if all works that were first published under this License somewhere other than this MMC, and subsequently incorporated in whole or in part into the MMC, (1) had no cover texts or invariant sections, and (2) were thus incorporated prior to November 1, 2008.
The operator of an MMC Site may republish an MMC contained in the site under CC-BY-SA on the same site at any time before August 1, 2009, provided the MMC is eligible for relicensing.
To use this License in a document you have written, include a copy of the License in the document and put the following copyright and license notices just after the title page:
Copyright (C) year your name. Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the section entitled ``GNU Free Documentation License''.
If you have Invariant Sections, Front-Cover Texts and Back-Cover Texts, replace the “with…Texts.” line with this:
with the Invariant Sections being list their titles, with the Front-Cover Texts being list, and with the Back-Cover Texts being list.
If you have Invariant Sections without Cover Texts, or some other combination of the three, merge those two alternatives to suit the situation.
If your document contains nontrivial examples of program code, we recommend releasing these examples in parallel under your choice of free software license, such as the GNU General Public License, to permit their use in free software.
|Jump to:||B C D F G M P R S|
|build environment:||Invoking guix-daemon|
|build users:||Setting Up the Daemon|
|chroot:||Setting Up the Daemon|
|container, build environment:||Invoking guix-daemon|
|cross-compilation:||Invoking guix build|
|daemon:||Setting Up the Daemon|
|functional package management:||Introduction|
|garbage collector:||Invoking guix gc|
|GNU Build System:||Defining Packages|
|monad:||The Store Monad|
|monadic functions:||The Store Monad|
|monadic values:||The Store Monad|
|multiple-output packages:||Packages with Multiple Outputs|
|package outputs:||Packages with Multiple Outputs|
|propagated inputs:||Invoking guix package|
|reproducible builds:||Invoking guix-daemon|
|search paths:||Invoking guix package|
|service definition:||Defining Services|
|store paths:||The Store|
|strata of code:||Derivations|
|system configuration:||System Configuration|
|Jump to:||B C D F G M P R S|
A B C D M O P R T V W
|The Store Monad|
|The Store Monad|
|The Store Monad|
|The Store Monad|
|Using the Configuration System|
|Using the Configuration System|
|The Store Monad|
|The Store Monad|
|The Store Monad|
|The Store Monad|
|The Store Monad|
|The Store Monad|
A B C D M O P R T V W
“Guix” is pronounced like “geeks”, or “ɡiːks” using the international phonetic alphabet (IPA).
On some systems
/dev/shm, which supports
shared memory, is a symlink to another directory such as
/run/shm, that is not is the chroot. When that is the
case, shared memory support is unavailable in the chroot environment.
The workaround is to make sure that /dev/shm is directly a
tmpfs mount point.
As of version 0.6, substitutes are downloaded from
http://hydra.gnu.org/ but are not authenticated—i.e.,
Guix cannot tell whether binaries it downloaded have been tampered with,
nor whether they come from the genuine
gnu.org build farm. This
will be fixed in future versions. In the meantime, concerned users can
--no-substitutes (see Invoking guix-daemon).
definitions like the one above may be automatically converted from the
Nixpkgs distribution using the
guix import command.
The term stratum in this context was coined by Manuel Serrano et al. in the context of their work on Hop.
This operation is commonly referred to as “bind”, but that name denotes an unrelated procedure in Guile. Thus we use this somewhat cryptic symbol inherited from the Haskell language.
Currently, this only works for GNU packages.
The term “free” here refers to the freedom provided to users of that software.