6 @documentencoding UTF-8
7 @settitle GNU Guix Reference Manual
12 @c Identifier of the OpenPGP key used to sign tarballs and such.
13 @set OPENPGP-SIGNING-KEY-ID 3CE464558A84FDC69DB40CFB090B11993D9AEBB5
16 Copyright @copyright{} 2012, 2013, 2014, 2015, 2016, 2017 Ludovic Courtès@*
17 Copyright @copyright{} 2013, 2014, 2016 Andreas Enge@*
18 Copyright @copyright{} 2013 Nikita Karetnikov@*
19 Copyright @copyright{} 2014, 2015, 2016 Alex Kost@*
20 Copyright @copyright{} 2015, 2016 Mathieu Lirzin@*
21 Copyright @copyright{} 2014 Pierre-Antoine Rault@*
22 Copyright @copyright{} 2015 Taylan Ulrich Bayırlı/Kammer@*
23 Copyright @copyright{} 2015, 2016, 2017 Leo Famulari@*
24 Copyright @copyright{} 2015, 2016, 2017 Ricardo Wurmus@*
25 Copyright @copyright{} 2016 Ben Woodcroft@*
26 Copyright @copyright{} 2016 Chris Marusich@*
27 Copyright @copyright{} 2016, 2017 Efraim Flashner@*
28 Copyright @copyright{} 2016 John Darrington@*
29 Copyright @copyright{} 2016 ng0@*
30 Copyright @copyright{} 2016 Jan Nieuwenhuizen@*
31 Copyright @copyright{} 2016 Julien Lepiller@*
32 Copyright @copyright{} 2016 Alex ter Weele@*
33 Copyright @copyright{} 2017 Clément Lassieur@*
34 Copyright @copyright{} 2017 Mathieu Othacehe@*
35 Copyright @copyright{} 2017 Federico Beffa@*
36 Copyright @copyright{} 2017 Carlo Zancanaro@*
37 Copyright @copyright{} 2017 Thomas Danckaert@*
38 Copyright @copyright{} 2017 humanitiesNerd@*
39 Copyright @copyright{} 2017 Christopher Allan Webber@*
40 Copyright @copyright{} 2017 Marius Bakke@*
41 Copyright @copyright{} 2017 Hartmut Goebel
43 Permission is granted to copy, distribute and/or modify this document
44 under the terms of the GNU Free Documentation License, Version 1.3 or
45 any later version published by the Free Software Foundation; with no
46 Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A
47 copy of the license is included in the section entitled ``GNU Free
48 Documentation License''.
51 @dircategory System administration
53 * Guix: (guix). Manage installed software and system configuration.
54 * guix package: (guix)Invoking guix package. Installing, removing, and upgrading packages.
55 * guix gc: (guix)Invoking guix gc. Reclaiming unused disk space.
56 * guix pull: (guix)Invoking guix pull. Update the list of available packages.
57 * guix system: (guix)Invoking guix system. Manage the operating system configuration.
60 @dircategory Software development
62 * guix environment: (guix)Invoking guix environment. Building development environments with Guix.
63 * guix build: (guix)Invoking guix build. Building packages.
64 * guix pack: (guix)Invoking guix pack. Creating binary bundles.
68 @title GNU Guix Reference Manual
69 @subtitle Using the GNU Guix Functional Package Manager
70 @author The GNU Guix Developers
73 @vskip 0pt plus 1filll
74 Edition @value{EDITION} @*
82 @c *********************************************************************
86 This document describes GNU Guix version @value{VERSION}, a functional
87 package management tool written for the GNU system.
90 * Introduction:: What is Guix about?
91 * Installation:: Installing Guix.
92 * Package Management:: Package installation, upgrade, etc.
93 * Programming Interface:: Using Guix in Scheme.
94 * Utilities:: Package management commands.
95 * GNU Distribution:: Software for your friendly GNU system.
96 * Contributing:: Your help needed!
98 * Acknowledgments:: Thanks!
99 * GNU Free Documentation License:: The license of this manual.
100 * Concept Index:: Concepts.
101 * Programming Index:: Data types, functions, and variables.
104 --- The Detailed Node Listing ---
108 * Binary Installation:: Getting Guix running in no time!
109 * Requirements:: Software needed to build and run Guix.
110 * Running the Test Suite:: Testing Guix.
111 * Setting Up the Daemon:: Preparing the build daemon's environment.
112 * Invoking guix-daemon:: Running the build daemon.
113 * Application Setup:: Application-specific setup.
115 Setting Up the Daemon
117 * Build Environment Setup:: Preparing the isolated build environment.
118 * Daemon Offload Setup:: Offloading builds to remote machines.
122 * Features:: How Guix will make your life brighter.
123 * Invoking guix package:: Package installation, removal, etc.
124 * Substitutes:: Downloading pre-built binaries.
125 * Packages with Multiple Outputs:: Single source package, multiple outputs.
126 * Invoking guix gc:: Running the garbage collector.
127 * Invoking guix pull:: Fetching the latest Guix and distribution.
128 * Invoking guix pack:: Creating software bundles.
129 * Invoking guix archive:: Exporting and importing store files.
131 Programming Interface
133 * Defining Packages:: Defining new packages.
134 * Build Systems:: Specifying how packages are built.
135 * The Store:: Manipulating the package store.
136 * Derivations:: Low-level interface to package derivations.
137 * The Store Monad:: Purely functional interface to the store.
138 * G-Expressions:: Manipulating build expressions.
142 * package Reference :: The package data type.
143 * origin Reference:: The origin data type.
147 * Invoking guix build:: Building packages from the command line.
148 * Invoking guix edit:: Editing package definitions.
149 * Invoking guix download:: Downloading a file and printing its hash.
150 * Invoking guix hash:: Computing the cryptographic hash of a file.
151 * Invoking guix import:: Importing package definitions.
152 * Invoking guix refresh:: Updating package definitions.
153 * Invoking guix lint:: Finding errors in package definitions.
154 * Invoking guix size:: Profiling disk usage.
155 * Invoking guix graph:: Visualizing the graph of packages.
156 * Invoking guix environment:: Setting up development environments.
157 * Invoking guix publish:: Sharing substitutes.
158 * Invoking guix challenge:: Challenging substitute servers.
159 * Invoking guix copy:: Copying to and from a remote store.
160 * Invoking guix container:: Process isolation.
162 Invoking @command{guix build}
164 * Common Build Options:: Build options for most commands.
165 * Package Transformation Options:: Creating variants of packages.
166 * Additional Build Options:: Options specific to 'guix build'.
167 * Debugging Build Failures:: Real life packaging experience.
171 * System Installation:: Installing the whole operating system.
172 * System Configuration:: Configuring the operating system.
173 * Documentation:: Browsing software user manuals.
174 * Installing Debugging Files:: Feeding the debugger.
175 * Security Updates:: Deploying security fixes quickly.
176 * Package Modules:: Packages from the programmer's viewpoint.
177 * Packaging Guidelines:: Growing the distribution.
178 * Bootstrapping:: GNU/Linux built from scratch.
179 * Porting:: Targeting another platform or kernel.
183 * Limitations:: What you can expect.
184 * Hardware Considerations:: Supported hardware.
185 * USB Stick Installation:: Preparing the installation medium.
186 * Preparing for Installation:: Networking, partitioning, etc.
187 * Proceeding with the Installation:: The real thing.
188 * Installing GuixSD in a VM:: GuixSD playground.
189 * Building the Installation Image:: How this comes to be.
193 * Using the Configuration System:: Customizing your GNU system.
194 * operating-system Reference:: Detail of operating-system declarations.
195 * File Systems:: Configuring file system mounts.
196 * Mapped Devices:: Block device extra processing.
197 * User Accounts:: Specifying user accounts.
198 * Locales:: Language and cultural convention settings.
199 * Services:: Specifying system services.
200 * Setuid Programs:: Programs running with root privileges.
201 * X.509 Certificates:: Authenticating HTTPS servers.
202 * Name Service Switch:: Configuring libc's name service switch.
203 * Initial RAM Disk:: Linux-Libre bootstrapping.
204 * Bootloader Configuration:: Configuring the boot loader.
205 * Invoking guix system:: Instantiating a system configuration.
206 * Running GuixSD in a VM:: How to run GuixSD in a virtual machine.
207 * Defining Services:: Adding new service definitions.
211 * Base Services:: Essential system services.
212 * Scheduled Job Execution:: The mcron service.
213 * Log Rotation:: The rottlog service.
214 * Networking Services:: Network setup, SSH daemon, etc.
215 * X Window:: Graphical display.
216 * Printing Services:: Local and remote printer support.
217 * Desktop Services:: D-Bus and desktop services.
218 * Database Services:: SQL databases.
219 * Mail Services:: IMAP, POP3, SMTP, and all that.
220 * Messaging Services:: Messaging services.
221 * Kerberos Services:: Kerberos services.
222 * Web Services:: Web servers.
223 * DNS Services:: DNS daemons.
224 * VPN Services:: VPN daemons.
225 * Network File System:: NFS related services.
226 * Continuous Integration:: The Cuirass service.
227 * Power management Services:: The TLP tool.
228 * Miscellaneous Services:: Other services.
232 * Service Composition:: The model for composing services.
233 * Service Types and Services:: Types and services.
234 * Service Reference:: API reference.
235 * Shepherd Services:: A particular type of service.
239 * Software Freedom:: What may go into the distribution.
240 * Package Naming:: What's in a name?
241 * Version Numbers:: When the name is not enough.
242 * Synopses and Descriptions:: Helping users find the right package.
243 * Python Modules:: A touch of British comedy.
244 * Perl Modules:: Little pearls.
245 * Java Packages:: Coffee break.
246 * Fonts:: Fond of fonts.
250 * Building from Git:: The latest and greatest.
251 * Running Guix Before It Is Installed:: Hacker tricks.
252 * The Perfect Setup:: The right tools.
253 * Coding Style:: Hygiene of the contributor.
254 * Submitting Patches:: Share your work.
258 * Programming Paradigm:: How to compose your elements.
259 * Modules:: Where to store your code?
260 * Data Types and Pattern Matching:: Implementing data structures.
261 * Formatting Code:: Writing conventions.
266 @c *********************************************************************
268 @chapter Introduction
271 GNU Guix@footnote{``Guix'' is pronounced like ``geeks'', or ``ɡiːks''
272 using the international phonetic alphabet (IPA).} is a package
273 management tool for the GNU system. Guix makes it easy for unprivileged
274 users to install, upgrade, or remove packages, to roll back to a
275 previous package set, to build packages from source, and generally
276 assists with the creation and maintenance of software environments.
278 @cindex user interfaces
279 Guix provides a command-line package management interface
280 (@pxref{Invoking guix package}), a set of command-line utilities
281 (@pxref{Utilities}), as well as Scheme programming interfaces
282 (@pxref{Programming Interface}).
284 Its @dfn{build daemon} is responsible for building packages on behalf of
285 users (@pxref{Setting Up the Daemon}) and for downloading pre-built
286 binaries from authorized sources (@pxref{Substitutes}).
288 @cindex extensibility of the distribution
289 @cindex customization, of packages
290 Guix includes package definitions for many GNU and non-GNU packages, all
291 of which @uref{https://www.gnu.org/philosophy/free-sw.html, respect the
292 user's computing freedom}. It is @emph{extensible}: users can write
293 their own package definitions (@pxref{Defining Packages}) and make them
294 available as independent package modules (@pxref{Package Modules}). It
295 is also @emph{customizable}: users can @emph{derive} specialized package
296 definitions from existing ones, including from the command line
297 (@pxref{Package Transformation Options}).
299 @cindex Guix System Distribution
301 You can install GNU@tie{}Guix on top of an existing GNU/Linux system
302 where it complements the available tools without interference
303 (@pxref{Installation}), or you can use it as part of the standalone
304 @dfn{Guix System Distribution} or GuixSD (@pxref{GNU Distribution}).
305 With GNU@tie{}GuixSD, you @emph{declare} all aspects of the operating
306 system configuration and Guix takes care of instantiating the
307 configuration in a transactional, reproducible, and stateless fashion
308 (@pxref{System Configuration}).
310 @cindex functional package management
311 Under the hood, Guix implements the @dfn{functional package management}
312 discipline pioneered by Nix (@pxref{Acknowledgments}).
313 In Guix, the package build and installation process is seen
314 as a @emph{function}, in the mathematical sense. That function takes inputs,
315 such as build scripts, a compiler, and libraries, and
316 returns an installed package. As a pure function, its result depends
317 solely on its inputs---for instance, it cannot refer to software or
318 scripts that were not explicitly passed as inputs. A build function
319 always produces the same result when passed a given set of inputs. It
320 cannot alter the environment of the running system in
321 any way; for instance, it cannot create, modify, or delete files outside
322 of its build and installation directories. This is achieved by running
323 build processes in isolated environments (or @dfn{containers}), where only their
324 explicit inputs are visible.
327 The result of package build functions is @dfn{cached} in the file
328 system, in a special directory called @dfn{the store} (@pxref{The
329 Store}). Each package is installed in a directory of its own in the
330 store---by default under @file{/gnu/store}. The directory name contains
331 a hash of all the inputs used to build that package; thus, changing an
332 input yields a different directory name.
334 This approach is the foundation for the salient features of Guix: support
335 for transactional package upgrade and rollback, per-user installation, and
336 garbage collection of packages (@pxref{Features}).
339 @c *********************************************************************
341 @chapter Installation
343 @cindex installing Guix
344 GNU Guix is available for download from its website at
345 @url{http://www.gnu.org/software/guix/}. This section describes the
346 software requirements of Guix, as well as how to install it and get
349 Note that this section is concerned with the installation of the package
350 manager, which can be done on top of a running GNU/Linux system. If,
351 instead, you want to install the complete GNU operating system,
352 @pxref{System Installation}.
354 @cindex foreign distro
355 When installed on a running GNU/Linux system---thereafter called a
356 @dfn{foreign distro}---GNU@tie{}Guix complements the available tools
357 without interference. Its data lives exclusively in two directories,
358 usually @file{/gnu/store} and @file{/var/guix}; other files on your
359 system, such as @file{/etc}, are left untouched.
361 Once installed, Guix can be updated by running @command{guix pull}
362 (@pxref{Invoking guix pull}).
365 * Binary Installation:: Getting Guix running in no time!
366 * Requirements:: Software needed to build and run Guix.
367 * Running the Test Suite:: Testing Guix.
368 * Setting Up the Daemon:: Preparing the build daemon's environment.
369 * Invoking guix-daemon:: Running the build daemon.
370 * Application Setup:: Application-specific setup.
373 @node Binary Installation
374 @section Binary Installation
376 @cindex installing Guix from binaries
377 This section describes how to install Guix on an arbitrary system from a
378 self-contained tarball providing binaries for Guix and for all its
379 dependencies. This is often quicker than installing from source, which
380 is described in the next sections. The only requirement is to have
383 Installing goes along these lines:
387 @cindex downloading Guix binary
388 Download the binary tarball from
389 @indicateurl{ftp://alpha.gnu.org/gnu/guix/guix-binary-@value{VERSION}.@var{system}.tar.xz},
390 where @var{system} is @code{x86_64-linux} for an @code{x86_64} machine
391 already running the kernel Linux, and so on.
393 @c The following is somewhat duplicated in ``System Installation''.
394 Make sure to download the associated @file{.sig} file and to verify the
395 authenticity of the tarball against it, along these lines:
398 $ wget ftp://alpha.gnu.org/gnu/guix/guix-binary-@value{VERSION}.@var{system}.tar.xz.sig
399 $ gpg --verify guix-binary-@value{VERSION}.@var{system}.tar.xz.sig
402 If that command fails because you do not have the required public key,
403 then run this command to import it:
406 $ gpg --keyserver pgp.mit.edu --recv-keys @value{OPENPGP-SIGNING-KEY-ID}
410 and rerun the @code{gpg --verify} command.
411 @c end authentication part
418 # tar --warning=no-timestamp -xf \
419 guix-binary-@value{VERSION}.@var{system}.tar.xz
420 # mv var/guix /var/ && mv gnu /
423 This creates @file{/gnu/store} (@pxref{The Store}) and @file{/var/guix}.
424 The latter contains a ready-to-use profile for @code{root} (see next
427 Do @emph{not} unpack the tarball on a working Guix system since that
428 would overwrite its own essential files.
430 The @code{--warning=no-timestamp} option makes sure GNU@tie{}tar does
431 not emit warnings about ``implausibly old time stamps'' (such
432 warnings were triggered by GNU@tie{}tar 1.26 and older; recent
434 They stem from the fact that all the
435 files in the archive have their modification time set to zero (which
436 means January 1st, 1970.) This is done on purpose to make sure the
437 archive content is independent of its creation time, thus making it
441 Make @code{root}'s profile available under @file{~/.guix-profile}:
444 # ln -sf /var/guix/profiles/per-user/root/guix-profile \
448 Source @file{etc/profile} to augment @code{PATH} and other relevant
449 environment variables:
452 # GUIX_PROFILE=$HOME/.guix-profile \
453 source $GUIX_PROFILE/etc/profile
457 Create the group and user accounts for build users as explained below
458 (@pxref{Build Environment Setup}).
461 Run the daemon, and set it to automatically start on boot.
463 If your host distro uses the systemd init system, this can be achieved
466 @c Versions of systemd that supported symlinked service files are not
467 @c yet widely deployed, so we should suggest that users copy the service
470 @c See this thread for more information:
471 @c http://lists.gnu.org/archive/html/guix-devel/2017-01/msg01199.html
474 # cp ~root/.guix-profile/lib/systemd/system/guix-daemon.service \
476 # systemctl start guix-daemon && systemctl enable guix-daemon
479 If your host distro uses the Upstart init system:
482 # initctl reload-configuration
483 # cp ~root/.guix-profile/lib/upstart/system/guix-daemon.conf /etc/init/
487 Otherwise, you can still start the daemon manually with:
490 # ~root/.guix-profile/bin/guix-daemon --build-users-group=guixbuild
494 Make the @command{guix} command available to other users on the machine,
498 # mkdir -p /usr/local/bin
500 # ln -s /var/guix/profiles/per-user/root/guix-profile/bin/guix
503 It is also a good idea to make the Info version of this manual available
507 # mkdir -p /usr/local/share/info
508 # cd /usr/local/share/info
509 # for i in /var/guix/profiles/per-user/root/guix-profile/share/info/* ;
513 That way, assuming @file{/usr/local/share/info} is in the search path,
514 running @command{info guix} will open this manual (@pxref{Other Info
515 Directories,,, texinfo, GNU Texinfo}, for more details on changing the
519 @cindex substitutes, authorization thereof
520 To use substitutes from @code{hydra.gnu.org} or one of its mirrors
521 (@pxref{Substitutes}), authorize them:
524 # guix archive --authorize < ~root/.guix-profile/share/guix/hydra.gnu.org.pub
528 Each user may need to perform a few additional steps to make their Guix
529 environment ready for use, @pxref{Application Setup}.
532 Voilà, the installation is complete!
534 You can confirm that Guix is working by installing a sample package into
538 # guix package -i hello
541 The @code{guix} package must remain available in @code{root}'s profile,
542 or it would become subject to garbage collection---in which case you
543 would find yourself badly handicapped by the lack of the @command{guix}
544 command. In other words, do not remove @code{guix} by running
545 @code{guix package -r guix}.
547 The binary installation tarball can be (re)produced and verified simply
548 by running the following command in the Guix source tree:
551 make guix-binary.@var{system}.tar.xz
555 ... which, in turn, runs:
558 guix pack -s @var{system} --localstatedir guix
561 @xref{Invoking guix pack}, for more info on this handy tool.
564 @section Requirements
566 This section lists requirements when building Guix from source. The
567 build procedure for Guix is the same as for other GNU software, and is
568 not covered here. Please see the files @file{README} and @file{INSTALL}
569 in the Guix source tree for additional details.
571 GNU Guix depends on the following packages:
574 @item @url{http://gnu.org/software/guile/, GNU Guile}, version 2.0.9 or
575 later, including 2.2.x;
576 @item @url{http://gnupg.org/, GNU libgcrypt};
578 @uref{http://gnutls.org/, GnuTLS}, specifically its Guile bindings
579 (@pxref{Guile Preparations, how to install the GnuTLS bindings for
580 Guile,, gnutls-guile, GnuTLS-Guile});
581 @item @url{http://www.gnu.org/software/make/, GNU Make}.
584 The following dependencies are optional:
589 @url{http://savannah.nongnu.org/projects/guile-json/, Guile-JSON} will
590 allow you to use the @command{guix import pypi} command (@pxref{Invoking
591 guix import}). It is of
592 interest primarily for developers and not for casual users.
595 @c Note: We need at least 0.10.2 for 'channel-send-eof'.
596 Support for build offloading (@pxref{Daemon Offload Setup}) and
597 @command{guix copy} (@pxref{Invoking guix copy}) depends on
598 @uref{https://github.com/artyom-poptsov/guile-ssh, Guile-SSH},
599 version 0.10.2 or later.
602 When @url{http://zlib.net, zlib} is available, @command{guix publish}
603 can compress build byproducts (@pxref{Invoking guix publish}).
606 Unless @code{--disable-daemon} was passed to @command{configure}, the
607 following packages are also needed:
610 @item @url{http://sqlite.org, SQLite 3};
611 @item @url{http://www.bzip.org, libbz2};
612 @item @url{http://gcc.gnu.org, GCC's g++}, with support for the
616 @cindex state directory
617 When configuring Guix on a system that already has a Guix installation,
618 be sure to specify the same state directory as the existing installation
619 using the @code{--localstatedir} option of the @command{configure}
620 script (@pxref{Directory Variables, @code{localstatedir},, standards,
621 GNU Coding Standards}). The @command{configure} script protects against
622 unintended misconfiguration of @var{localstatedir} so you do not
623 inadvertently corrupt your store (@pxref{The Store}).
625 @cindex Nix, compatibility
626 When a working installation of @url{http://nixos.org/nix/, the Nix package
627 manager} is available, you
628 can instead configure Guix with @code{--disable-daemon}. In that case,
629 Nix replaces the three dependencies above.
631 Guix is compatible with Nix, so it is possible to share the same store
632 between both. To do so, you must pass @command{configure} not only the
633 same @code{--with-store-dir} value, but also the same
634 @code{--localstatedir} value. The latter is essential because it
635 specifies where the database that stores metadata about the store is
636 located, among other things. The default values for Nix are
637 @code{--with-store-dir=/nix/store} and @code{--localstatedir=/nix/var}.
638 Note that @code{--disable-daemon} is not required if
639 your goal is to share the store with Nix.
641 @node Running the Test Suite
642 @section Running the Test Suite
645 After a successful @command{configure} and @code{make} run, it is a good
646 idea to run the test suite. It can help catch issues with the setup or
647 environment, or bugs in Guix itself---and really, reporting test
648 failures is a good way to help improve the software. To run the test
655 Test cases can run in parallel: you can use the @code{-j} option of
656 GNU@tie{}make to speed things up. The first run may take a few minutes
657 on a recent machine; subsequent runs will be faster because the store
658 that is created for test purposes will already have various things in
661 It is also possible to run a subset of the tests by defining the
662 @code{TESTS} makefile variable as in this example:
665 make check TESTS="tests/store.scm tests/cpio.scm"
668 By default, tests results are displayed at a file level. In order to
669 see the details of every individual test cases, it is possible to define
670 the @code{SCM_LOG_DRIVER_FLAGS} makefile variable as in this example:
673 make check TESTS="tests/base64.scm" SCM_LOG_DRIVER_FLAGS="--brief=no"
676 Upon failure, please email @email{bug-guix@@gnu.org} and attach the
677 @file{test-suite.log} file. Please specify the Guix version being used
678 as well as version numbers of the dependencies (@pxref{Requirements}) in
681 Guix also comes with a whole-system test suite that tests complete
682 GuixSD operating system instances. It can only run on systems where
683 Guix is already installed, using:
690 or, again, by defining @code{TESTS} to select a subset of tests to run:
693 make check-system TESTS="basic mcron"
696 These system tests are defined in the @code{(gnu tests @dots{})}
697 modules. They work by running the operating systems under test with
698 lightweight instrumentation in a virtual machine (VM). They can be
699 computationally intensive or rather cheap, depending on whether
700 substitutes are available for their dependencies (@pxref{Substitutes}).
701 Some of them require a lot of storage space to hold VM images.
703 Again in case of test failures, please send @email{bug-guix@@gnu.org}
706 @node Setting Up the Daemon
707 @section Setting Up the Daemon
710 Operations such as building a package or running the garbage collector
711 are all performed by a specialized process, the @dfn{build daemon}, on
712 behalf of clients. Only the daemon may access the store and its
713 associated database. Thus, any operation that manipulates the store
714 goes through the daemon. For instance, command-line tools such as
715 @command{guix package} and @command{guix build} communicate with the
716 daemon (@i{via} remote procedure calls) to instruct it what to do.
718 The following sections explain how to prepare the build daemon's
719 environment. See also @ref{Substitutes}, for information on how to allow
720 the daemon to download pre-built binaries.
723 * Build Environment Setup:: Preparing the isolated build environment.
724 * Daemon Offload Setup:: Offloading builds to remote machines.
727 @node Build Environment Setup
728 @subsection Build Environment Setup
730 @cindex build environment
731 In a standard multi-user setup, Guix and its daemon---the
732 @command{guix-daemon} program---are installed by the system
733 administrator; @file{/gnu/store} is owned by @code{root} and
734 @command{guix-daemon} runs as @code{root}. Unprivileged users may use
735 Guix tools to build packages or otherwise access the store, and the
736 daemon will do it on their behalf, ensuring that the store is kept in a
737 consistent state, and allowing built packages to be shared among users.
740 When @command{guix-daemon} runs as @code{root}, you may not want package
741 build processes themselves to run as @code{root} too, for obvious
742 security reasons. To avoid that, a special pool of @dfn{build users}
743 should be created for use by build processes started by the daemon.
744 These build users need not have a shell and a home directory: they will
745 just be used when the daemon drops @code{root} privileges in build
746 processes. Having several such users allows the daemon to launch
747 distinct build processes under separate UIDs, which guarantees that they
748 do not interfere with each other---an essential feature since builds are
749 regarded as pure functions (@pxref{Introduction}).
751 On a GNU/Linux system, a build user pool may be created like this (using
752 Bash syntax and the @code{shadow} commands):
754 @c See http://lists.gnu.org/archive/html/bug-guix/2013-01/msg00239.html
755 @c for why `-G' is needed.
757 # groupadd --system guixbuild
758 # for i in `seq -w 1 10`;
760 useradd -g guixbuild -G guixbuild \
761 -d /var/empty -s `which nologin` \
762 -c "Guix build user $i" --system \
768 The number of build users determines how many build jobs may run in
769 parallel, as specified by the @option{--max-jobs} option
770 (@pxref{Invoking guix-daemon, @option{--max-jobs}}). To use
771 @command{guix system vm} and related commands, you may need to add the
772 build users to the @code{kvm} group so they can access @file{/dev/kvm},
773 using @code{-G guixbuild,kvm} instead of @code{-G guixbuild}
774 (@pxref{Invoking guix system}).
776 The @code{guix-daemon} program may then be run as @code{root} with the
777 following command@footnote{If your machine uses the systemd init system,
778 dropping the @file{@var{prefix}/lib/systemd/system/guix-daemon.service}
779 file in @file{/etc/systemd/system} will ensure that
780 @command{guix-daemon} is automatically started. Similarly, if your
781 machine uses the Upstart init system, drop the
782 @file{@var{prefix}/lib/upstart/system/guix-daemon.conf}
783 file in @file{/etc/init}.}:
786 # guix-daemon --build-users-group=guixbuild
791 This way, the daemon starts build processes in a chroot, under one of
792 the @code{guixbuilder} users. On GNU/Linux, by default, the chroot
793 environment contains nothing but:
795 @c Keep this list in sync with libstore/build.cc! -----------------------
798 a minimal @code{/dev} directory, created mostly independently from the
799 host @code{/dev}@footnote{``Mostly'', because while the set of files
800 that appear in the chroot's @code{/dev} is fixed, most of these files
801 can only be created if the host has them.};
804 the @code{/proc} directory; it only shows the processes of the container
805 since a separate PID name space is used;
808 @file{/etc/passwd} with an entry for the current user and an entry for
812 @file{/etc/group} with an entry for the user's group;
815 @file{/etc/hosts} with an entry that maps @code{localhost} to
819 a writable @file{/tmp} directory.
822 You can influence the directory where the daemon stores build trees
823 @i{via} the @code{TMPDIR} environment variable. However, the build tree
824 within the chroot is always called @file{/tmp/guix-build-@var{name}.drv-0},
825 where @var{name} is the derivation name---e.g., @code{coreutils-8.24}.
826 This way, the value of @code{TMPDIR} does not leak inside build
827 environments, which avoids discrepancies in cases where build processes
828 capture the name of their build tree.
831 The daemon also honors the @code{http_proxy} environment variable for
832 HTTP downloads it performs, be it for fixed-output derivations
833 (@pxref{Derivations}) or for substitutes (@pxref{Substitutes}).
835 If you are installing Guix as an unprivileged user, it is still possible
836 to run @command{guix-daemon} provided you pass @code{--disable-chroot}.
837 However, build processes will not be isolated from one another, and not
838 from the rest of the system. Thus, build processes may interfere with
839 each other, and may access programs, libraries, and other files
840 available on the system---making it much harder to view them as
841 @emph{pure} functions.
844 @node Daemon Offload Setup
845 @subsection Using the Offload Facility
849 When desired, the build daemon can @dfn{offload} derivation builds to
850 other machines running Guix, using the @code{offload} @dfn{build
851 hook}@footnote{This feature is available only when
852 @uref{https://github.com/artyom-poptsov/guile-ssh, Guile-SSH} is
854 feature is enabled, a list of user-specified build machines is read from
855 @file{/etc/guix/machines.scm}; every time a build is requested, for
856 instance via @code{guix build}, the daemon attempts to offload it to one
857 of the machines that satisfy the constraints of the derivation, in
858 particular its system type---e.g., @file{x86_64-linux}. Missing
859 prerequisites for the build are copied over SSH to the target machine,
860 which then proceeds with the build; upon success the output(s) of the
861 build are copied back to the initial machine.
863 The @file{/etc/guix/machines.scm} file typically looks like this:
867 (name "eightysix.example.org")
868 (system "x86_64-linux")
869 (host-key "ssh-ed25519 AAAAC3Nza@dots{}")
871 (speed 2.)) ;incredibly fast!
874 (name "meeps.example.org")
875 (system "mips64el-linux")
876 (host-key "ssh-rsa AAAAB3Nza@dots{}")
879 (string-append (getenv "HOME")
880 "/.ssh/identity-for-guix"))))
884 In the example above we specify a list of two build machines, one for
885 the @code{x86_64} architecture and one for the @code{mips64el}
888 In fact, this file is---not surprisingly!---a Scheme file that is
889 evaluated when the @code{offload} hook is started. Its return value
890 must be a list of @code{build-machine} objects. While this example
891 shows a fixed list of build machines, one could imagine, say, using
892 DNS-SD to return a list of potential build machines discovered in the
893 local network (@pxref{Introduction, Guile-Avahi,, guile-avahi, Using
894 Avahi in Guile Scheme Programs}). The @code{build-machine} data type is
897 @deftp {Data Type} build-machine
898 This data type represents build machines to which the daemon may offload
899 builds. The important fields are:
904 The host name of the remote machine.
907 The system type of the remote machine---e.g., @code{"x86_64-linux"}.
910 The user account to use when connecting to the remote machine over SSH.
911 Note that the SSH key pair must @emph{not} be passphrase-protected, to
912 allow non-interactive logins.
915 This must be the machine's SSH @dfn{public host key} in OpenSSH format.
916 This is used to authenticate the machine when we connect to it. It is a
917 long string that looks like this:
920 ssh-ed25519 AAAAC3NzaC@dots{}mde+UhL hint@@example.org
923 If the machine is running the OpenSSH daemon, @command{sshd}, the host
924 key can be found in a file such as
925 @file{/etc/ssh/ssh_host_ed25519_key.pub}.
927 If the machine is running the SSH daemon of GNU@tie{}lsh,
928 @command{lshd}, the host key is in @file{/etc/lsh/host-key.pub} or a
929 similar file. It can be converted to the OpenSSH format using
930 @command{lsh-export-key} (@pxref{Converting keys,,, lsh, LSH Manual}):
933 $ lsh-export-key --openssh < /etc/lsh/host-key.pub
934 ssh-rsa AAAAB3NzaC1yc2EAAAAEOp8FoQAAAQEAs1eB46LV@dots{}
939 A number of optional fields may be specified:
943 @item @code{port} (default: @code{22})
944 Port number of SSH server on the machine.
946 @item @code{private-key} (default: @file{~/.ssh/id_rsa})
947 The SSH private key file to use when connecting to the machine, in
950 @item @code{compression} (default: @code{"zlib@@openssh.com,zlib"})
951 @itemx @code{compression-level} (default: @code{3})
952 The SSH-level compression methods and compression level requested.
954 Note that offloading relies on SSH compression to reduce bandwidth usage
955 when transferring files to and from build machines.
957 @item @code{daemon-socket} (default: @code{"/var/guix/daemon-socket/socket"})
958 File name of the Unix-domain socket @command{guix-daemon} is listening
961 @item @code{parallel-builds} (default: @code{1})
962 The number of builds that may run in parallel on the machine.
964 @item @code{speed} (default: @code{1.0})
965 A ``relative speed factor''. The offload scheduler will tend to prefer
966 machines with a higher speed factor.
968 @item @code{features} (default: @code{'()})
969 A list of strings denoting specific features supported by the machine.
970 An example is @code{"kvm"} for machines that have the KVM Linux modules
971 and corresponding hardware support. Derivations can request features by
972 name, and they will be scheduled on matching build machines.
977 The @code{guile} command must be in the search path on the build
978 machines. In addition, the Guix modules must be in
979 @code{$GUILE_LOAD_PATH} on the build machine---you can check whether
980 this is the case by running:
983 ssh build-machine guile -c "'(use-modules (guix config))'"
986 There is one last thing to do once @file{machines.scm} is in place. As
987 explained above, when offloading, files are transferred back and forth
988 between the machine stores. For this to work, you first need to
989 generate a key pair on each machine to allow the daemon to export signed
990 archives of files from the store (@pxref{Invoking guix archive}):
993 # guix archive --generate-key
997 Each build machine must authorize the key of the master machine so that
998 it accepts store items it receives from the master:
1001 # guix archive --authorize < master-public-key.txt
1005 Likewise, the master machine must authorize the key of each build machine.
1007 All the fuss with keys is here to express pairwise mutual trust
1008 relations between the master and the build machines. Concretely, when
1009 the master receives files from a build machine (and @i{vice versa}), its
1010 build daemon can make sure they are genuine, have not been tampered
1011 with, and that they are signed by an authorized key.
1013 @cindex offload test
1014 To test whether your setup is operational, run this command on the
1021 This will attempt to connect to each of the build machines specified in
1022 @file{/etc/guix/machines.scm}, make sure Guile and the Guix modules are
1023 available on each machine, attempt to export to the machine and import
1024 from it, and report any error in the process.
1026 If you want to test a different machine file, just specify it on the
1030 # guix offload test machines-qualif.scm
1033 Last, you can test the subset of the machines whose name matches a
1034 regular expression like this:
1037 # guix offload test machines.scm '\.gnu\.org$'
1040 @node Invoking guix-daemon
1041 @section Invoking @command{guix-daemon}
1043 The @command{guix-daemon} program implements all the functionality to
1044 access the store. This includes launching build processes, running the
1045 garbage collector, querying the availability of a build result, etc. It
1046 is normally run as @code{root} like this:
1049 # guix-daemon --build-users-group=guixbuild
1053 For details on how to set it up, @pxref{Setting Up the Daemon}.
1056 @cindex container, build environment
1057 @cindex build environment
1058 @cindex reproducible builds
1059 By default, @command{guix-daemon} launches build processes under
1060 different UIDs, taken from the build group specified with
1061 @code{--build-users-group}. In addition, each build process is run in a
1062 chroot environment that only contains the subset of the store that the
1063 build process depends on, as specified by its derivation
1064 (@pxref{Programming Interface, derivation}), plus a set of specific
1065 system directories. By default, the latter contains @file{/dev} and
1066 @file{/dev/pts}. Furthermore, on GNU/Linux, the build environment is a
1067 @dfn{container}: in addition to having its own file system tree, it has
1068 a separate mount name space, its own PID name space, network name space,
1069 etc. This helps achieve reproducible builds (@pxref{Features}).
1071 When the daemon performs a build on behalf of the user, it creates a
1072 build directory under @file{/tmp} or under the directory specified by
1073 its @code{TMPDIR} environment variable; this directory is shared with
1074 the container for the duration of the build. Be aware that using a
1075 directory other than @file{/tmp} can affect build results---for example,
1076 with a longer directory name, a build process that uses Unix-domain
1077 sockets might hit the name length limitation for @code{sun_path}, which
1078 it would otherwise not hit.
1080 The build directory is automatically deleted upon completion, unless the
1081 build failed and the client specified @option{--keep-failed}
1082 (@pxref{Invoking guix build, @option{--keep-failed}}).
1084 The following command-line options are supported:
1087 @item --build-users-group=@var{group}
1088 Take users from @var{group} to run build processes (@pxref{Setting Up
1089 the Daemon, build users}).
1091 @item --no-substitutes
1093 Do not use substitutes for build products. That is, always build things
1094 locally instead of allowing downloads of pre-built binaries
1095 (@pxref{Substitutes}).
1097 By default substitutes are used, unless the client---such as the
1098 @command{guix package} command---is explicitly invoked with
1099 @code{--no-substitutes}.
1101 When the daemon runs with @code{--no-substitutes}, clients can still
1102 explicitly enable substitution @i{via} the @code{set-build-options}
1103 remote procedure call (@pxref{The Store}).
1105 @item --substitute-urls=@var{urls}
1106 @anchor{daemon-substitute-urls}
1107 Consider @var{urls} the default whitespace-separated list of substitute
1108 source URLs. When this option is omitted,
1109 @indicateurl{https://mirror.hydra.gnu.org https://hydra.gnu.org} is used
1110 (@code{mirror.hydra.gnu.org} is a mirror of @code{hydra.gnu.org}).
1112 This means that substitutes may be downloaded from @var{urls}, as long
1113 as they are signed by a trusted signature (@pxref{Substitutes}).
1116 @item --no-build-hook
1117 Do not use the @dfn{build hook}.
1119 The build hook is a helper program that the daemon can start and to
1120 which it submits build requests. This mechanism is used to offload
1121 builds to other machines (@pxref{Daemon Offload Setup}).
1123 @item --cache-failures
1124 Cache build failures. By default, only successful builds are cached.
1126 When this option is used, @command{guix gc --list-failures} can be used
1127 to query the set of store items marked as failed; @command{guix gc
1128 --clear-failures} removes store items from the set of cached failures.
1129 @xref{Invoking guix gc}.
1131 @item --cores=@var{n}
1133 Use @var{n} CPU cores to build each derivation; @code{0} means as many
1136 The default value is @code{0}, but it may be overridden by clients, such
1137 as the @code{--cores} option of @command{guix build} (@pxref{Invoking
1140 The effect is to define the @code{NIX_BUILD_CORES} environment variable
1141 in the build process, which can then use it to exploit internal
1142 parallelism---for instance, by running @code{make -j$NIX_BUILD_CORES}.
1144 @item --max-jobs=@var{n}
1146 Allow at most @var{n} build jobs in parallel. The default value is
1147 @code{1}. Setting it to @code{0} means that no builds will be performed
1148 locally; instead, the daemon will offload builds (@pxref{Daemon Offload
1149 Setup}), or simply fail.
1151 @item --max-silent-time=@var{seconds}
1152 When the build or substitution process remains silent for more than
1153 @var{seconds}, terminate it and report a build failure.
1155 The default value is @code{0}, which disables the timeout.
1157 The value specified here can be overridden by clients (@pxref{Common
1158 Build Options, @code{--max-silent-time}}).
1160 @item --timeout=@var{seconds}
1161 Likewise, when the build or substitution process lasts for more than
1162 @var{seconds}, terminate it and report a build failure.
1164 The default value is @code{0}, which disables the timeout.
1166 The value specified here can be overridden by clients (@pxref{Common
1167 Build Options, @code{--timeout}}).
1169 @item --rounds=@var{N}
1170 Build each derivation @var{n} times in a row, and raise an error if
1171 consecutive build results are not bit-for-bit identical. Note that this
1172 setting can be overridden by clients such as @command{guix build}
1173 (@pxref{Invoking guix build}).
1175 When used in conjunction with @option{--keep-failed}, the differing
1176 output is kept in the store, under @file{/gnu/store/@dots{}-check}.
1177 This makes it easy to look for differences between the two results.
1180 Produce debugging output.
1182 This is useful to debug daemon start-up issues, but then it may be
1183 overridden by clients, for example the @code{--verbosity} option of
1184 @command{guix build} (@pxref{Invoking guix build}).
1186 @item --chroot-directory=@var{dir}
1187 Add @var{dir} to the build chroot.
1189 Doing this may change the result of build processes---for instance if
1190 they use optional dependencies found in @var{dir} when it is available,
1191 and not otherwise. For that reason, it is not recommended to do so.
1192 Instead, make sure that each derivation declares all the inputs that it
1195 @item --disable-chroot
1196 Disable chroot builds.
1198 Using this option is not recommended since, again, it would allow build
1199 processes to gain access to undeclared dependencies. It is necessary,
1200 though, when @command{guix-daemon} is running under an unprivileged user
1203 @item --disable-log-compression
1204 Disable compression of the build logs.
1206 Unless @code{--lose-logs} is used, all the build logs are kept in the
1207 @var{localstatedir}. To save space, the daemon automatically compresses
1208 them with bzip2 by default. This option disables that.
1210 @item --disable-deduplication
1211 @cindex deduplication
1212 Disable automatic file ``deduplication'' in the store.
1214 By default, files added to the store are automatically ``deduplicated'':
1215 if a newly added file is identical to another one found in the store,
1216 the daemon makes the new file a hard link to the other file. This can
1217 noticeably reduce disk usage, at the expense of slightly increased
1218 input/output load at the end of a build process. This option disables
1221 @item --gc-keep-outputs[=yes|no]
1222 Tell whether the garbage collector (GC) must keep outputs of live
1225 When set to ``yes'', the GC will keep the outputs of any live derivation
1226 available in the store---the @code{.drv} files. The default is ``no'',
1227 meaning that derivation outputs are kept only if they are GC roots.
1229 @item --gc-keep-derivations[=yes|no]
1230 Tell whether the garbage collector (GC) must keep derivations
1231 corresponding to live outputs.
1233 When set to ``yes'', as is the case by default, the GC keeps
1234 derivations---i.e., @code{.drv} files---as long as at least one of their
1235 outputs is live. This allows users to keep track of the origins of
1236 items in their store. Setting it to ``no'' saves a bit of disk space.
1238 Note that when both @code{--gc-keep-derivations} and
1239 @code{--gc-keep-outputs} are used, the effect is to keep all the build
1240 prerequisites (the sources, compiler, libraries, and other build-time
1241 tools) of live objects in the store, regardless of whether these
1242 prerequisites are live. This is convenient for developers since it
1243 saves rebuilds or downloads.
1245 @item --impersonate-linux-2.6
1246 On Linux-based systems, impersonate Linux 2.6. This means that the
1247 kernel's @code{uname} system call will report 2.6 as the release number.
1249 This might be helpful to build programs that (usually wrongfully) depend
1250 on the kernel version number.
1253 Do not keep build logs. By default they are kept under
1254 @code{@var{localstatedir}/guix/log}.
1256 @item --system=@var{system}
1257 Assume @var{system} as the current system type. By default it is the
1258 architecture/kernel pair found at configure time, such as
1259 @code{x86_64-linux}.
1261 @item --listen=@var{endpoint}
1262 Listen for connections on @var{endpoint}. @var{endpoint} is interpreted
1263 as the file name of a Unix-domain socket if it starts with
1264 @code{/} (slash sign). Otherwise, @var{endpoint} is interpreted as a
1265 host name or host name and port to listen to. Here are a few examples:
1268 @item --listen=/gnu/var/daemon
1269 Listen for connections on the @file{/gnu/var/daemon} Unix-domain socket,
1270 creating it if needed.
1272 @item --listen=localhost
1273 @cindex daemon, remote access
1274 @cindex remote access to the daemon
1275 @cindex daemon, cluster setup
1276 @cindex clusters, daemon setup
1277 Listen for TCP connections on the network interface corresponding to
1278 @code{localhost}, on port 44146.
1280 @item --listen=128.0.0.42:1234
1281 Listen for TCP connections on the network interface corresponding to
1282 @code{128.0.0.42}, on port 1234.
1285 This option can be repeated multiple times, in which case
1286 @command{guix-daemon} accepts connections on all the specified
1287 endpoints. Users can tell client commands what endpoint to connect to
1288 by setting the @code{GUIX_DAEMON_SOCKET} environment variable
1289 (@pxref{The Store, @code{GUIX_DAEMON_SOCKET}}).
1292 The daemon protocol is @emph{unauthenticated and unencrypted}. Using
1293 @code{--listen=@var{host}} is suitable on local networks, such as
1294 clusters, where only trusted nodes may connect to the build daemon. In
1295 other cases where remote access to the daemon is needed, we recommend
1296 using Unix-domain sockets along with SSH.
1299 When @code{--listen} is omitted, @command{guix-daemon} listens for
1300 connections on the Unix-domain socket located at
1301 @file{@var{localstatedir}/daemon-socket/socket}.
1305 @node Application Setup
1306 @section Application Setup
1308 @cindex foreign distro
1309 When using Guix on top of GNU/Linux distribution other than GuixSD---a
1310 so-called @dfn{foreign distro}---a few additional steps are needed to
1311 get everything in place. Here are some of them.
1315 @anchor{locales-and-locpath}
1316 @cindex locales, when not on GuixSD
1318 @vindex GUIX_LOCPATH
1319 Packages installed @i{via} Guix will not use the locale data of the
1320 host system. Instead, you must first install one of the locale packages
1321 available with Guix and then define the @code{GUIX_LOCPATH} environment
1325 $ guix package -i glibc-locales
1326 $ export GUIX_LOCPATH=$HOME/.guix-profile/lib/locale
1329 Note that the @code{glibc-locales} package contains data for all the
1330 locales supported by the GNU@tie{}libc and weighs in at around
1331 110@tie{}MiB. Alternatively, the @code{glibc-utf8-locales} is smaller but
1332 limited to a few UTF-8 locales.
1334 The @code{GUIX_LOCPATH} variable plays a role similar to @code{LOCPATH}
1335 (@pxref{Locale Names, @code{LOCPATH},, libc, The GNU C Library Reference
1336 Manual}). There are two important differences though:
1340 @code{GUIX_LOCPATH} is honored only by the libc in Guix, and not by the libc
1341 provided by foreign distros. Thus, using @code{GUIX_LOCPATH} allows you
1342 to make sure the programs of the foreign distro will not end up loading
1343 incompatible locale data.
1346 libc suffixes each entry of @code{GUIX_LOCPATH} with @code{/X.Y}, where
1347 @code{X.Y} is the libc version---e.g., @code{2.22}. This means that,
1348 should your Guix profile contain a mixture of programs linked against
1349 different libc version, each libc version will only try to load locale
1350 data in the right format.
1353 This is important because the locale data format used by different libc
1354 versions may be incompatible.
1356 @subsection Name Service Switch
1358 @cindex name service switch, glibc
1359 @cindex NSS (name service switch), glibc
1360 @cindex nscd (name service caching daemon)
1361 @cindex name service caching daemon (nscd)
1362 When using Guix on a foreign distro, we @emph{strongly recommend} that
1363 the system run the GNU C library's @dfn{name service cache daemon},
1364 @command{nscd}, which should be listening on the
1365 @file{/var/run/nscd/socket} socket. Failing to do that, applications
1366 installed with Guix may fail to look up host names or user accounts, or
1367 may even crash. The next paragraphs explain why.
1369 @cindex @file{nsswitch.conf}
1370 The GNU C library implements a @dfn{name service switch} (NSS), which is
1371 an extensible mechanism for ``name lookups'' in general: host name
1372 resolution, user accounts, and more (@pxref{Name Service Switch,,, libc,
1373 The GNU C Library Reference Manual}).
1375 @cindex Network information service (NIS)
1376 @cindex NIS (Network information service)
1377 Being extensible, the NSS supports @dfn{plugins}, which provide new name
1378 lookup implementations: for example, the @code{nss-mdns} plugin allow
1379 resolution of @code{.local} host names, the @code{nis} plugin allows
1380 user account lookup using the Network information service (NIS), and so
1381 on. These extra ``lookup services'' are configured system-wide in
1382 @file{/etc/nsswitch.conf}, and all the programs running on the system
1383 honor those settings (@pxref{NSS Configuration File,,, libc, The GNU C
1386 When they perform a name lookup---for instance by calling the
1387 @code{getaddrinfo} function in C---applications first try to connect to
1388 the nscd; on success, nscd performs name lookups on their behalf. If
1389 the nscd is not running, then they perform the name lookup by
1390 themselves, by loading the name lookup services into their own address
1391 space and running it. These name lookup services---the
1392 @file{libnss_*.so} files---are @code{dlopen}'d, but they may come from
1393 the host system's C library, rather than from the C library the
1394 application is linked against (the C library coming from Guix).
1396 And this is where the problem is: if your application is linked against
1397 Guix's C library (say, glibc 2.24) and tries to load NSS plugins from
1398 another C library (say, @code{libnss_mdns.so} for glibc 2.22), it will
1399 likely crash or have its name lookups fail unexpectedly.
1401 Running @command{nscd} on the system, among other advantages, eliminates
1402 this binary incompatibility problem because those @code{libnss_*.so}
1403 files are loaded in the @command{nscd} process, not in applications
1406 @subsection X11 Fonts
1409 The majority of graphical applications use Fontconfig to locate and
1410 load fonts and perform X11-client-side rendering. The @code{fontconfig}
1411 package in Guix looks for fonts in @file{$HOME/.guix-profile}
1412 by default. Thus, to allow graphical applications installed with Guix
1413 to display fonts, you have to install fonts with Guix as well.
1414 Essential font packages include @code{gs-fonts}, @code{font-dejavu}, and
1415 @code{font-gnu-freefont-ttf}.
1417 To display text written in Chinese languages, Japanese, or Korean in
1418 graphical applications, consider installing
1419 @code{font-adobe-source-han-sans} or @code{font-wqy-zenhei}. The former
1420 has multiple outputs, one per language family (@pxref{Packages with
1421 Multiple Outputs}). For instance, the following command installs fonts
1422 for Chinese languages:
1425 guix package -i font-adobe-source-han-sans:cn
1428 @cindex @code{xterm}
1429 Older programs such as @command{xterm} do not use Fontconfig and instead
1430 rely on server-side font rendering. Such programs require to specify a
1431 full name of a font using XLFD (X Logical Font Description), like this:
1434 -*-dejavu sans-medium-r-normal-*-*-100-*-*-*-*-*-1
1437 To be able to use such full names for the TrueType fonts installed in
1438 your Guix profile, you need to extend the font path of the X server:
1441 xset +fp ~/.guix-profile/share/fonts/truetype
1444 @cindex @code{xlsfonts}
1445 After that, you can run @code{xlsfonts} (from @code{xlsfonts} package)
1446 to make sure your TrueType fonts are listed there.
1448 @subsection X.509 Certificates
1450 @cindex @code{nss-certs}
1451 The @code{nss-certs} package provides X.509 certificates, which allow
1452 programs to authenticate Web servers accessed over HTTPS.
1454 When using Guix on a foreign distro, you can install this package and
1455 define the relevant environment variables so that packages know where to
1456 look for certificates. @xref{X.509 Certificates}, for detailed
1459 @subsection Emacs Packages
1461 @cindex @code{emacs}
1462 When you install Emacs packages with Guix, the elisp files may be placed
1463 either in @file{$HOME/.guix-profile/share/emacs/site-lisp/} or in
1465 @file{$HOME/.guix-profile/share/emacs/site-lisp/guix.d/}. The latter
1466 directory exists because potentially there may exist thousands of Emacs
1467 packages and storing all their files in a single directory may be not
1468 reliable (because of name conflicts). So we think using a separate
1469 directory for each package is a good idea. It is very similar to how
1470 the Emacs package system organizes the file structure (@pxref{Package
1471 Files,,, emacs, The GNU Emacs Manual}).
1473 By default, Emacs (installed with Guix) ``knows'' where these packages
1474 are placed, so you do not need to perform any configuration. If, for
1475 some reason, you want to avoid auto-loading Emacs packages installed
1476 with Guix, you can do so by running Emacs with @code{--no-site-file}
1477 option (@pxref{Init File,,, emacs, The GNU Emacs Manual}).
1479 @subsection The GCC toolchain
1484 Guix offers individual compiler packages such as @code{gcc} but if you
1485 are in need of a complete toolchain for compiling and linking source
1486 code what you really want is the @code{gcc-toolchain} package. This
1487 package provides a complete GCC toolchain for C/C++ development,
1488 including GCC itself, the GNU C Library (headers and binaries, plus
1489 debugging symbols in the @code{debug} output), Binutils, and a linker
1492 @cindex attempt to use impure library, error message
1494 The wrapper's purpose is to inspect the @code{-L} and @code{-l} switches
1495 passed to the linker, add corresponding @code{-rpath} arguments, and
1496 invoke the actual linker with this new set of arguments. By default,
1497 the linker wrapper refuses to link to libraries outside the store to
1498 ensure ``purity''. This can be annoying when using the toolchain to
1499 link with local libraries. To allow references to libraries outside the
1500 store you need to define the environment variable
1501 @code{GUIX_LD_WRAPPER_ALLOW_IMPURITIES}.
1505 @c *********************************************************************
1506 @node Package Management
1507 @chapter Package Management
1510 The purpose of GNU Guix is to allow users to easily install, upgrade, and
1511 remove software packages, without having to know about their build
1512 procedures or dependencies. Guix also goes beyond this obvious set of
1515 This chapter describes the main features of Guix, as well as the
1516 package management tools it provides. Along with the command-line
1517 interface described below (@pxref{Invoking guix package, @code{guix
1518 package}}), you may also use Emacs Interface (@pxref{Top,,,
1519 emacs-guix, The Emacs-Guix Reference Manual}), after installing
1520 @code{emacs-guix} package (run @kbd{M-x guix-help} command to start
1524 guix package -i emacs-guix
1528 * Features:: How Guix will make your life brighter.
1529 * Invoking guix package:: Package installation, removal, etc.
1530 * Substitutes:: Downloading pre-built binaries.
1531 * Packages with Multiple Outputs:: Single source package, multiple outputs.
1532 * Invoking guix gc:: Running the garbage collector.
1533 * Invoking guix pull:: Fetching the latest Guix and distribution.
1534 * Invoking guix pack:: Creating software bundles.
1535 * Invoking guix archive:: Exporting and importing store files.
1541 When using Guix, each package ends up in the @dfn{package store}, in its
1542 own directory---something that resembles
1543 @file{/gnu/store/xxx-package-1.2}, where @code{xxx} is a base32 string.
1545 Instead of referring to these directories, users have their own
1546 @dfn{profile}, which points to the packages that they actually want to
1547 use. These profiles are stored within each user's home directory, at
1548 @code{$HOME/.guix-profile}.
1550 For example, @code{alice} installs GCC 4.7.2. As a result,
1551 @file{/home/alice/.guix-profile/bin/gcc} points to
1552 @file{/gnu/store/@dots{}-gcc-4.7.2/bin/gcc}. Now, on the same machine,
1553 @code{bob} had already installed GCC 4.8.0. The profile of @code{bob}
1554 simply continues to point to
1555 @file{/gnu/store/@dots{}-gcc-4.8.0/bin/gcc}---i.e., both versions of GCC
1556 coexist on the same system without any interference.
1558 The @command{guix package} command is the central tool to manage
1559 packages (@pxref{Invoking guix package}). It operates on the per-user
1560 profiles, and can be used @emph{with normal user privileges}.
1562 @cindex transactions
1563 The command provides the obvious install, remove, and upgrade
1564 operations. Each invocation is actually a @emph{transaction}: either
1565 the specified operation succeeds, or nothing happens. Thus, if the
1566 @command{guix package} process is terminated during the transaction,
1567 or if a power outage occurs during the transaction, then the user's
1568 profile remains in its previous state, and remains usable.
1570 In addition, any package transaction may be @emph{rolled back}. So, if,
1571 for example, an upgrade installs a new version of a package that turns
1572 out to have a serious bug, users may roll back to the previous instance
1573 of their profile, which was known to work well. Similarly, the global
1574 system configuration on GuixSD is subject to
1575 transactional upgrades and roll-back
1576 (@pxref{Using the Configuration System}).
1578 All packages in the package store may be @emph{garbage-collected}.
1579 Guix can determine which packages are still referenced by user
1580 profiles, and remove those that are provably no longer referenced
1581 (@pxref{Invoking guix gc}). Users may also explicitly remove old
1582 generations of their profile so that the packages they refer to can be
1585 @cindex reproducibility
1586 @cindex reproducible builds
1587 Finally, Guix takes a @dfn{purely functional} approach to package
1588 management, as described in the introduction (@pxref{Introduction}).
1589 Each @file{/gnu/store} package directory name contains a hash of all the
1590 inputs that were used to build that package---compiler, libraries, build
1591 scripts, etc. This direct correspondence allows users to make sure a
1592 given package installation matches the current state of their
1593 distribution. It also helps maximize @dfn{build reproducibility}:
1594 thanks to the isolated build environments that are used, a given build
1595 is likely to yield bit-identical files when performed on different
1596 machines (@pxref{Invoking guix-daemon, container}).
1599 This foundation allows Guix to support @dfn{transparent binary/source
1600 deployment}. When a pre-built binary for a @file{/gnu/store} item is
1601 available from an external source---a @dfn{substitute}, Guix just
1602 downloads it and unpacks it;
1603 otherwise, it builds the package from source, locally
1604 (@pxref{Substitutes}). Because build results are usually bit-for-bit
1605 reproducible, users do not have to trust servers that provide
1606 substitutes: they can force a local build and @emph{challenge} providers
1607 (@pxref{Invoking guix challenge}).
1609 Control over the build environment is a feature that is also useful for
1610 developers. The @command{guix environment} command allows developers of
1611 a package to quickly set up the right development environment for their
1612 package, without having to manually install the dependencies of the
1613 package into their profile (@pxref{Invoking guix environment}).
1615 @node Invoking guix package
1616 @section Invoking @command{guix package}
1618 @cindex installing packages
1619 @cindex removing packages
1620 @cindex package installation
1621 @cindex package removal
1622 The @command{guix package} command is the tool that allows users to
1623 install, upgrade, and remove packages, as well as rolling back to
1624 previous configurations. It operates only on the user's own profile,
1625 and works with normal user privileges (@pxref{Features}). Its syntax
1629 guix package @var{options}
1631 @cindex transactions
1632 Primarily, @var{options} specifies the operations to be performed during
1633 the transaction. Upon completion, a new profile is created, but
1634 previous @dfn{generations} of the profile remain available, should the user
1637 For example, to remove @code{lua} and install @code{guile} and
1638 @code{guile-cairo} in a single transaction:
1641 guix package -r lua -i guile guile-cairo
1644 @command{guix package} also supports a @dfn{declarative approach}
1645 whereby the user specifies the exact set of packages to be available and
1646 passes it @i{via} the @option{--manifest} option
1647 (@pxref{profile-manifest, @option{--manifest}}).
1650 For each user, a symlink to the user's default profile is automatically
1651 created in @file{$HOME/.guix-profile}. This symlink always points to the
1652 current generation of the user's default profile. Thus, users can add
1653 @file{$HOME/.guix-profile/bin} to their @code{PATH} environment
1654 variable, and so on.
1655 @cindex search paths
1656 If you are not using the Guix System Distribution, consider adding the
1657 following lines to your @file{~/.bash_profile} (@pxref{Bash Startup
1658 Files,,, bash, The GNU Bash Reference Manual}) so that newly-spawned
1659 shells get all the right environment variable definitions:
1662 GUIX_PROFILE="$HOME/.guix-profile" \
1663 source "$HOME/.guix-profile/etc/profile"
1666 In a multi-user setup, user profiles are stored in a place registered as
1667 a @dfn{garbage-collector root}, which @file{$HOME/.guix-profile} points
1668 to (@pxref{Invoking guix gc}). That directory is normally
1669 @code{@var{localstatedir}/profiles/per-user/@var{user}}, where
1670 @var{localstatedir} is the value passed to @code{configure} as
1671 @code{--localstatedir}, and @var{user} is the user name. The
1672 @file{per-user} directory is created when @command{guix-daemon} is
1673 started, and the @var{user} sub-directory is created by @command{guix
1676 The @var{options} can be among the following:
1680 @item --install=@var{package} @dots{}
1681 @itemx -i @var{package} @dots{}
1682 Install the specified @var{package}s.
1684 Each @var{package} may specify either a simple package name, such as
1685 @code{guile}, or a package name followed by an at-sign and version number,
1686 such as @code{guile@@1.8.8} or simply @code{guile@@1.8} (in the latter
1687 case, the newest version prefixed by @code{1.8} is selected.)
1689 If no version number is specified, the
1690 newest available version will be selected. In addition, @var{package}
1691 may contain a colon, followed by the name of one of the outputs of the
1692 package, as in @code{gcc:doc} or @code{binutils@@2.22:lib}
1693 (@pxref{Packages with Multiple Outputs}). Packages with a corresponding
1694 name (and optionally version) are searched for among the GNU
1695 distribution modules (@pxref{Package Modules}).
1697 @cindex propagated inputs
1698 Sometimes packages have @dfn{propagated inputs}: these are dependencies
1699 that automatically get installed along with the required package
1700 (@pxref{package-propagated-inputs, @code{propagated-inputs} in
1701 @code{package} objects}, for information about propagated inputs in
1702 package definitions).
1704 @anchor{package-cmd-propagated-inputs}
1705 An example is the GNU MPC library: its C header files refer to those of
1706 the GNU MPFR library, which in turn refer to those of the GMP library.
1707 Thus, when installing MPC, the MPFR and GMP libraries also get installed
1708 in the profile; removing MPC also removes MPFR and GMP---unless they had
1709 also been explicitly installed by the user.
1711 Besides, packages sometimes rely on the definition of environment
1712 variables for their search paths (see explanation of
1713 @code{--search-paths} below). Any missing or possibly incorrect
1714 environment variable definitions are reported here.
1716 @item --install-from-expression=@var{exp}
1718 Install the package @var{exp} evaluates to.
1720 @var{exp} must be a Scheme expression that evaluates to a
1721 @code{<package>} object. This option is notably useful to disambiguate
1722 between same-named variants of a package, with expressions such as
1723 @code{(@@ (gnu packages base) guile-final)}.
1725 Note that this option installs the first output of the specified
1726 package, which may be insufficient when needing a specific output of a
1727 multiple-output package.
1729 @item --install-from-file=@var{file}
1730 @itemx -f @var{file}
1731 Install the package that the code within @var{file} evaluates to.
1733 As an example, @var{file} might contain a definition like this
1734 (@pxref{Defining Packages}):
1737 @verbatiminclude package-hello.scm
1740 Developers may find it useful to include such a @file{guix.scm} file
1741 in the root of their project source tree that can be used to test
1742 development snapshots and create reproducible development environments
1743 (@pxref{Invoking guix environment}).
1745 @item --remove=@var{package} @dots{}
1746 @itemx -r @var{package} @dots{}
1747 Remove the specified @var{package}s.
1749 As for @code{--install}, each @var{package} may specify a version number
1750 and/or output name in addition to the package name. For instance,
1751 @code{-r glibc:debug} would remove the @code{debug} output of
1754 @item --upgrade[=@var{regexp} @dots{}]
1755 @itemx -u [@var{regexp} @dots{}]
1756 @cindex upgrading packages
1757 Upgrade all the installed packages. If one or more @var{regexp}s are
1758 specified, upgrade only installed packages whose name matches a
1759 @var{regexp}. Also see the @code{--do-not-upgrade} option below.
1761 Note that this upgrades package to the latest version of packages found
1762 in the distribution currently installed. To update your distribution,
1763 you should regularly run @command{guix pull} (@pxref{Invoking guix
1766 @item --do-not-upgrade[=@var{regexp} @dots{}]
1767 When used together with the @code{--upgrade} option, do @emph{not}
1768 upgrade any packages whose name matches a @var{regexp}. For example, to
1769 upgrade all packages in the current profile except those containing the
1770 substring ``emacs'':
1773 $ guix package --upgrade . --do-not-upgrade emacs
1776 @item @anchor{profile-manifest}--manifest=@var{file}
1777 @itemx -m @var{file}
1778 @cindex profile declaration
1779 @cindex profile manifest
1780 Create a new generation of the profile from the manifest object
1781 returned by the Scheme code in @var{file}.
1783 This allows you to @emph{declare} the profile's contents rather than
1784 constructing it through a sequence of @code{--install} and similar
1785 commands. The advantage is that @var{file} can be put under version
1786 control, copied to different machines to reproduce the same profile, and
1789 @c FIXME: Add reference to (guix profile) documentation when available.
1790 @var{file} must return a @dfn{manifest} object, which is roughly a list
1793 @findex packages->manifest
1795 (use-package-modules guile emacs)
1800 ;; Use a specific package output.
1801 (list guile-2.0 "debug")))
1804 @findex specifications->manifest
1805 In this example we have to know which modules define the @code{emacs}
1806 and @code{guile-2.0} variables to provide the right
1807 @code{use-package-modules} line, which can be cumbersome. We can
1808 instead provide regular package specifications and let
1809 @code{specifications->manifest} look up the corresponding package
1813 (specifications->manifest
1814 '("emacs" "guile@@2.2" "guile@@2.2:debug"))
1818 @cindex rolling back
1819 @cindex undoing transactions
1820 @cindex transactions, undoing
1821 Roll back to the previous @dfn{generation} of the profile---i.e., undo
1822 the last transaction.
1824 When combined with options such as @code{--install}, roll back occurs
1825 before any other actions.
1827 When rolling back from the first generation that actually contains
1828 installed packages, the profile is made to point to the @dfn{zeroth
1829 generation}, which contains no files apart from its own metadata.
1831 After having rolled back, installing, removing, or upgrading packages
1832 overwrites previous future generations. Thus, the history of the
1833 generations in a profile is always linear.
1835 @item --switch-generation=@var{pattern}
1836 @itemx -S @var{pattern}
1838 Switch to a particular generation defined by @var{pattern}.
1840 @var{pattern} may be either a generation number or a number prefixed
1841 with ``+'' or ``-''. The latter means: move forward/backward by a
1842 specified number of generations. For example, if you want to return to
1843 the latest generation after @code{--roll-back}, use
1844 @code{--switch-generation=+1}.
1846 The difference between @code{--roll-back} and
1847 @code{--switch-generation=-1} is that @code{--switch-generation} will
1848 not make a zeroth generation, so if a specified generation does not
1849 exist, the current generation will not be changed.
1851 @item --search-paths[=@var{kind}]
1852 @cindex search paths
1853 Report environment variable definitions, in Bash syntax, that may be
1854 needed in order to use the set of installed packages. These environment
1855 variables are used to specify @dfn{search paths} for files used by some
1856 of the installed packages.
1858 For example, GCC needs the @code{CPATH} and @code{LIBRARY_PATH}
1859 environment variables to be defined so it can look for headers and
1860 libraries in the user's profile (@pxref{Environment Variables,,, gcc,
1861 Using the GNU Compiler Collection (GCC)}). If GCC and, say, the C
1862 library are installed in the profile, then @code{--search-paths} will
1863 suggest setting these variables to @code{@var{profile}/include} and
1864 @code{@var{profile}/lib}, respectively.
1866 The typical use case is to define these environment variables in the
1870 $ eval `guix package --search-paths`
1873 @var{kind} may be one of @code{exact}, @code{prefix}, or @code{suffix},
1874 meaning that the returned environment variable definitions will either
1875 be exact settings, or prefixes or suffixes of the current value of these
1876 variables. When omitted, @var{kind} defaults to @code{exact}.
1878 This option can also be used to compute the @emph{combined} search paths
1879 of several profiles. Consider this example:
1882 $ guix package -p foo -i guile
1883 $ guix package -p bar -i guile-json
1884 $ guix package -p foo -p bar --search-paths
1887 The last command above reports about the @code{GUILE_LOAD_PATH}
1888 variable, even though, taken individually, neither @file{foo} nor
1889 @file{bar} would lead to that recommendation.
1892 @item --profile=@var{profile}
1893 @itemx -p @var{profile}
1894 Use @var{profile} instead of the user's default profile.
1897 Produce verbose output. In particular, emit the build log of the
1898 environment on the standard error port.
1901 Use the bootstrap Guile to build the profile. This option is only
1902 useful to distribution developers.
1906 In addition to these actions, @command{guix package} supports the
1907 following options to query the current state of a profile, or the
1908 availability of packages:
1912 @item --search=@var{regexp}
1913 @itemx -s @var{regexp}
1914 @cindex searching for packages
1915 List the available packages whose name, synopsis, or description matches
1916 @var{regexp}, sorted by relevance. Print all the metadata of matching packages in
1917 @code{recutils} format (@pxref{Top, GNU recutils databases,, recutils,
1918 GNU recutils manual}).
1920 This allows specific fields to be extracted using the @command{recsel}
1921 command, for instance:
1924 $ guix package -s malloc | recsel -p name,version,relevance
1938 Similarly, to show the name of all the packages available under the
1939 terms of the GNU@tie{}LGPL version 3:
1942 $ guix package -s "" | recsel -p name -e 'license ~ "LGPL 3"'
1949 It is also possible to refine search results using several @code{-s}
1950 flags. For example, the following command returns a list of board
1954 $ guix package -s '\<board\>' -s game | recsel -p name
1959 If we were to omit @code{-s game}, we would also get software packages
1960 that deal with printed circuit boards; removing the angle brackets
1961 around @code{board} would further add packages that have to do with
1964 And now for a more elaborate example. The following command searches
1965 for cryptographic libraries, filters out Haskell, Perl, Python, and Ruby
1966 libraries, and prints the name and synopsis of the matching packages:
1969 $ guix package -s crypto -s library | \
1970 recsel -e '! (name ~ "^(ghc|perl|python|ruby)")' -p name,synopsis
1974 @xref{Selection Expressions,,, recutils, GNU recutils manual}, for more
1975 information on @dfn{selection expressions} for @code{recsel -e}.
1977 @item --show=@var{package}
1978 Show details about @var{package}, taken from the list of available packages, in
1979 @code{recutils} format (@pxref{Top, GNU recutils databases,, recutils, GNU
1983 $ guix package --show=python | recsel -p name,version
1991 You may also specify the full name of a package to only get details about a
1992 specific version of it:
1994 $ guix package --show=python@@3.4 | recsel -p name,version
2001 @item --list-installed[=@var{regexp}]
2002 @itemx -I [@var{regexp}]
2003 List the currently installed packages in the specified profile, with the
2004 most recently installed packages shown last. When @var{regexp} is
2005 specified, list only installed packages whose name matches @var{regexp}.
2007 For each installed package, print the following items, separated by
2008 tabs: the package name, its version string, the part of the package that
2009 is installed (for instance, @code{out} for the default output,
2010 @code{include} for its headers, etc.), and the path of this package in
2013 @item --list-available[=@var{regexp}]
2014 @itemx -A [@var{regexp}]
2015 List packages currently available in the distribution for this system
2016 (@pxref{GNU Distribution}). When @var{regexp} is specified, list only
2017 installed packages whose name matches @var{regexp}.
2019 For each package, print the following items separated by tabs: its name,
2020 its version string, the parts of the package (@pxref{Packages with
2021 Multiple Outputs}), and the source location of its definition.
2023 @item --list-generations[=@var{pattern}]
2024 @itemx -l [@var{pattern}]
2026 Return a list of generations along with their creation dates; for each
2027 generation, show the installed packages, with the most recently
2028 installed packages shown last. Note that the zeroth generation is never
2031 For each installed package, print the following items, separated by
2032 tabs: the name of a package, its version string, the part of the package
2033 that is installed (@pxref{Packages with Multiple Outputs}), and the
2034 location of this package in the store.
2036 When @var{pattern} is used, the command returns only matching
2037 generations. Valid patterns include:
2040 @item @emph{Integers and comma-separated integers}. Both patterns denote
2041 generation numbers. For instance, @code{--list-generations=1} returns
2044 And @code{--list-generations=1,8,2} outputs three generations in the
2045 specified order. Neither spaces nor trailing commas are allowed.
2047 @item @emph{Ranges}. @code{--list-generations=2..9} prints the
2048 specified generations and everything in between. Note that the start of
2049 a range must be smaller than its end.
2051 It is also possible to omit the endpoint. For example,
2052 @code{--list-generations=2..}, returns all generations starting from the
2055 @item @emph{Durations}. You can also get the last @emph{N}@tie{}days, weeks,
2056 or months by passing an integer along with the first letter of the
2057 duration. For example, @code{--list-generations=20d} lists generations
2058 that are up to 20 days old.
2061 @item --delete-generations[=@var{pattern}]
2062 @itemx -d [@var{pattern}]
2063 When @var{pattern} is omitted, delete all generations except the current
2066 This command accepts the same patterns as @option{--list-generations}.
2067 When @var{pattern} is specified, delete the matching generations. When
2068 @var{pattern} specifies a duration, generations @emph{older} than the
2069 specified duration match. For instance, @code{--delete-generations=1m}
2070 deletes generations that are more than one month old.
2072 If the current generation matches, it is @emph{not} deleted. Also, the
2073 zeroth generation is never deleted.
2075 Note that deleting generations prevents rolling back to them.
2076 Consequently, this command must be used with care.
2080 Finally, since @command{guix package} may actually start build
2081 processes, it supports all the common build options (@pxref{Common Build
2082 Options}). It also supports package transformation options, such as
2083 @option{--with-source} (@pxref{Package Transformation Options}).
2084 However, note that package transformations are lost when upgrading; to
2085 preserve transformations across upgrades, you should define your own
2086 package variant in a Guile module and add it to @code{GUIX_PACKAGE_PATH}
2087 (@pxref{Defining Packages}).
2091 @section Substitutes
2094 @cindex pre-built binaries
2095 Guix supports transparent source/binary deployment, which means that it
2096 can either build things locally, or download pre-built items from a
2097 server. We call these pre-built items @dfn{substitutes}---they are
2098 substitutes for local build results. In many cases, downloading a
2099 substitute is much faster than building things locally.
2101 Substitutes can be anything resulting from a derivation build
2102 (@pxref{Derivations}). Of course, in the common case, they are
2103 pre-built package binaries, but source tarballs, for instance, which
2104 also result from derivation builds, can be available as substitutes.
2106 The @code{hydra.gnu.org} server is a front-end to a build farm that
2107 builds packages from the GNU distribution continuously for some
2108 architectures, and makes them available as substitutes. This is the
2109 default source of substitutes; it can be overridden by passing the
2110 @option{--substitute-urls} option either to @command{guix-daemon}
2111 (@pxref{daemon-substitute-urls,, @code{guix-daemon --substitute-urls}})
2112 or to client tools such as @command{guix package}
2113 (@pxref{client-substitute-urls,, client @option{--substitute-urls}
2116 Substitute URLs can be either HTTP or HTTPS.
2117 HTTPS is recommended because communications are encrypted; conversely,
2118 using HTTP makes all communications visible to an eavesdropper, who
2119 could use the information gathered to determine, for instance, whether
2120 your system has unpatched security vulnerabilities.
2123 @cindex digital signatures
2124 @cindex substitutes, authorization thereof
2125 To allow Guix to download substitutes from @code{hydra.gnu.org} or a
2127 must add its public key to the access control list (ACL) of archive
2128 imports, using the @command{guix archive} command (@pxref{Invoking guix
2129 archive}). Doing so implies that you trust @code{hydra.gnu.org} to not
2130 be compromised and to serve genuine substitutes.
2132 This public key is installed along with Guix, in
2133 @code{@var{prefix}/share/guix/hydra.gnu.org.pub}, where @var{prefix} is
2134 the installation prefix of Guix. If you installed Guix from source,
2135 make sure you checked the GPG signature of
2136 @file{guix-@value{VERSION}.tar.gz}, which contains this public key file.
2137 Then, you can run something like this:
2140 # guix archive --authorize < hydra.gnu.org.pub
2143 Once this is in place, the output of a command like @code{guix build}
2144 should change from something like:
2147 $ guix build emacs --dry-run
2148 The following derivations would be built:
2149 /gnu/store/yr7bnx8xwcayd6j95r2clmkdl1qh688w-emacs-24.3.drv
2150 /gnu/store/x8qsh1hlhgjx6cwsjyvybnfv2i37z23w-dbus-1.6.4.tar.gz.drv
2151 /gnu/store/1ixwp12fl950d15h2cj11c73733jay0z-alsa-lib-1.0.27.1.tar.bz2.drv
2152 /gnu/store/nlma1pw0p603fpfiqy7kn4zm105r5dmw-util-linux-2.21.drv
2160 $ guix build emacs --dry-run
2161 The following files would be downloaded:
2162 /gnu/store/pk3n22lbq6ydamyymqkkz7i69wiwjiwi-emacs-24.3
2163 /gnu/store/2ygn4ncnhrpr61rssa6z0d9x22si0va3-libjpeg-8d
2164 /gnu/store/71yz6lgx4dazma9dwn2mcjxaah9w77jq-cairo-1.12.16
2165 /gnu/store/7zdhgp0n1518lvfn8mb96sxqfmvqrl7v-libxrender-0.9.7
2170 This indicates that substitutes from @code{hydra.gnu.org} are usable and
2171 will be downloaded, when possible, for future builds.
2173 Guix ignores substitutes that are not signed, or that are not signed by
2174 one of the keys listed in the ACL. It also detects and raises an error
2175 when attempting to use a substitute that has been tampered with.
2178 Substitutes are downloaded over HTTP or HTTPS.
2179 The @code{http_proxy} environment
2180 variable can be set in the environment of @command{guix-daemon} and is
2181 honored for downloads of substitutes. Note that the value of
2182 @code{http_proxy} in the environment where @command{guix build},
2183 @command{guix package}, and other client commands are run has
2184 @emph{absolutely no effect}.
2186 When using HTTPS, the server's X.509 certificate is @emph{not} validated
2187 (in other words, the server is not authenticated), contrary to what
2188 HTTPS clients such as Web browsers usually do. This is because Guix
2189 authenticates substitute information itself, as explained above, which
2190 is what we care about (whereas X.509 certificates are about
2191 authenticating bindings between domain names and public keys.)
2193 The substitute mechanism can be disabled globally by running
2194 @code{guix-daemon} with @code{--no-substitutes} (@pxref{Invoking
2195 guix-daemon}). It can also be disabled temporarily by passing the
2196 @code{--no-substitutes} option to @command{guix package}, @command{guix
2197 build}, and other command-line tools.
2200 @unnumberedsubsec On Trusting Binaries
2202 Today, each individual's control over their own computing is at the
2203 mercy of institutions, corporations, and groups with enough power and
2204 determination to subvert the computing infrastructure and exploit its
2205 weaknesses. While using @code{hydra.gnu.org} substitutes can be
2206 convenient, we encourage users to also build on their own, or even run
2207 their own build farm, such that @code{hydra.gnu.org} is less of an
2208 interesting target. One way to help is by publishing the software you
2209 build using @command{guix publish} so that others have one more choice
2210 of server to download substitutes from (@pxref{Invoking guix publish}).
2212 Guix has the foundations to maximize build reproducibility
2213 (@pxref{Features}). In most cases, independent builds of a given
2214 package or derivation should yield bit-identical results. Thus, through
2215 a diverse set of independent package builds, we can strengthen the
2216 integrity of our systems. The @command{guix challenge} command aims to
2217 help users assess substitute servers, and to assist developers in
2218 finding out about non-deterministic package builds (@pxref{Invoking guix
2219 challenge}). Similarly, the @option{--check} option of @command{guix
2220 build} allows users to check whether previously-installed substitutes
2221 are genuine by rebuilding them locally (@pxref{build-check,
2222 @command{guix build --check}}).
2224 In the future, we want Guix to have support to publish and retrieve
2225 binaries to/from other users, in a peer-to-peer fashion. If you would
2226 like to discuss this project, join us on @email{guix-devel@@gnu.org}.
2229 @node Packages with Multiple Outputs
2230 @section Packages with Multiple Outputs
2232 @cindex multiple-output packages
2233 @cindex package outputs
2236 Often, packages defined in Guix have a single @dfn{output}---i.e., the
2237 source package leads to exactly one directory in the store. When running
2238 @command{guix package -i glibc}, one installs the default output of the
2239 GNU libc package; the default output is called @code{out}, but its name
2240 can be omitted as shown in this command. In this particular case, the
2241 default output of @code{glibc} contains all the C header files, shared
2242 libraries, static libraries, Info documentation, and other supporting
2245 Sometimes it is more appropriate to separate the various types of files
2246 produced from a single source package into separate outputs. For
2247 instance, the GLib C library (used by GTK+ and related packages)
2248 installs more than 20 MiB of reference documentation as HTML pages.
2249 To save space for users who do not need it, the documentation goes to a
2250 separate output, called @code{doc}. To install the main GLib output,
2251 which contains everything but the documentation, one would run:
2254 guix package -i glib
2257 @cindex documentation
2258 The command to install its documentation is:
2261 guix package -i glib:doc
2264 Some packages install programs with different ``dependency footprints''.
2265 For instance, the WordNet package installs both command-line tools and
2266 graphical user interfaces (GUIs). The former depend solely on the C
2267 library, whereas the latter depend on Tcl/Tk and the underlying X
2268 libraries. In this case, we leave the command-line tools in the default
2269 output, whereas the GUIs are in a separate output. This allows users
2270 who do not need the GUIs to save space. The @command{guix size} command
2271 can help find out about such situations (@pxref{Invoking guix size}).
2272 @command{guix graph} can also be helpful (@pxref{Invoking guix graph}).
2274 There are several such multiple-output packages in the GNU distribution.
2275 Other conventional output names include @code{lib} for libraries and
2276 possibly header files, @code{bin} for stand-alone programs, and
2277 @code{debug} for debugging information (@pxref{Installing Debugging
2278 Files}). The outputs of a packages are listed in the third column of
2279 the output of @command{guix package --list-available} (@pxref{Invoking
2283 @node Invoking guix gc
2284 @section Invoking @command{guix gc}
2286 @cindex garbage collector
2288 Packages that are installed, but not used, may be @dfn{garbage-collected}.
2289 The @command{guix gc} command allows users to explicitly run the garbage
2290 collector to reclaim space from the @file{/gnu/store} directory. It is
2291 the @emph{only} way to remove files from @file{/gnu/store}---removing
2292 files or directories manually may break it beyond repair!
2294 The garbage collector has a set of known @dfn{roots}: any file under
2295 @file{/gnu/store} reachable from a root is considered @dfn{live} and
2296 cannot be deleted; any other file is considered @dfn{dead} and may be
2297 deleted. The set of garbage collector roots includes default user
2298 profiles, and may be augmented with @command{guix build --root}, for
2299 example (@pxref{Invoking guix build}).
2301 Prior to running @code{guix gc --collect-garbage} to make space, it is
2302 often useful to remove old generations from user profiles; that way, old
2303 package builds referenced by those generations can be reclaimed. This
2304 is achieved by running @code{guix package --delete-generations}
2305 (@pxref{Invoking guix package}).
2307 The @command{guix gc} command has three modes of operation: it can be
2308 used to garbage-collect any dead files (the default), to delete specific
2309 files (the @code{--delete} option), to print garbage-collector
2310 information, or for more advanced queries. The garbage collection
2311 options are as follows:
2314 @item --collect-garbage[=@var{min}]
2315 @itemx -C [@var{min}]
2316 Collect garbage---i.e., unreachable @file{/gnu/store} files and
2317 sub-directories. This is the default operation when no option is
2320 When @var{min} is given, stop once @var{min} bytes have been collected.
2321 @var{min} may be a number of bytes, or it may include a unit as a
2322 suffix, such as @code{MiB} for mebibytes and @code{GB} for gigabytes
2323 (@pxref{Block size, size specifications,, coreutils, GNU Coreutils}).
2325 When @var{min} is omitted, collect all the garbage.
2327 @item --free-space=@var{free}
2328 @itemx -F @var{free}
2329 Collect garbage until @var{free} space is available under
2330 @file{/gnu/store}, if possible; @var{free} denotes storage space, such
2331 as @code{500MiB}, as described above.
2333 When @var{free} or more is already available in @file{/gnu/store}, do
2334 nothing and exit immediately.
2338 Attempt to delete all the store files and directories specified as
2339 arguments. This fails if some of the files are not in the store, or if
2340 they are still live.
2342 @item --list-failures
2343 List store items corresponding to cached build failures.
2345 This prints nothing unless the daemon was started with
2346 @option{--cache-failures} (@pxref{Invoking guix-daemon,
2347 @option{--cache-failures}}).
2349 @item --clear-failures
2350 Remove the specified store items from the failed-build cache.
2352 Again, this option only makes sense when the daemon is started with
2353 @option{--cache-failures}. Otherwise, it does nothing.
2356 Show the list of dead files and directories still present in the
2357 store---i.e., files and directories no longer reachable from any root.
2360 Show the list of live store files and directories.
2364 In addition, the references among existing store files can be queried:
2370 @cindex package dependencies
2371 List the references (respectively, the referrers) of store files given
2377 List the requisites of the store files passed as arguments. Requisites
2378 include the store files themselves, their references, and the references
2379 of these, recursively. In other words, the returned list is the
2380 @dfn{transitive closure} of the store files.
2382 @xref{Invoking guix size}, for a tool to profile the size of the closure
2383 of an element. @xref{Invoking guix graph}, for a tool to visualize
2384 the graph of references.
2388 Lastly, the following options allow you to check the integrity of the
2389 store and to control disk usage.
2393 @item --verify[=@var{options}]
2394 @cindex integrity, of the store
2395 @cindex integrity checking
2396 Verify the integrity of the store.
2398 By default, make sure that all the store items marked as valid in the
2399 database of the daemon actually exist in @file{/gnu/store}.
2401 When provided, @var{options} must be a comma-separated list containing one
2402 or more of @code{contents} and @code{repair}.
2404 When passing @option{--verify=contents}, the daemon computes the
2405 content hash of each store item and compares it against its hash in the
2406 database. Hash mismatches are reported as data corruptions. Because it
2407 traverses @emph{all the files in the store}, this command can take a
2408 long time, especially on systems with a slow disk drive.
2410 @cindex repairing the store
2411 @cindex corruption, recovering from
2412 Using @option{--verify=repair} or @option{--verify=contents,repair}
2413 causes the daemon to try to repair corrupt store items by fetching
2414 substitutes for them (@pxref{Substitutes}). Because repairing is not
2415 atomic, and thus potentially dangerous, it is available only to the
2416 system administrator. A lightweight alternative, when you know exactly
2417 which items in the store are corrupt, is @command{guix build --repair}
2418 (@pxref{Invoking guix build}).
2421 @cindex deduplication
2422 Optimize the store by hard-linking identical files---this is
2423 @dfn{deduplication}.
2425 The daemon performs deduplication after each successful build or archive
2426 import, unless it was started with @code{--disable-deduplication}
2427 (@pxref{Invoking guix-daemon, @code{--disable-deduplication}}). Thus,
2428 this option is primarily useful when the daemon was running with
2429 @code{--disable-deduplication}.
2433 @node Invoking guix pull
2434 @section Invoking @command{guix pull}
2436 @cindex upgrading Guix
2437 @cindex updating Guix
2438 @cindex @command{guix pull}
2440 Packages are installed or upgraded to the latest version available in
2441 the distribution currently available on your local machine. To update
2442 that distribution, along with the Guix tools, you must run @command{guix
2443 pull}: the command downloads the latest Guix source code and package
2444 descriptions, and deploys it.
2446 On completion, @command{guix package} will use packages and package
2447 versions from this just-retrieved copy of Guix. Not only that, but all
2448 the Guix commands and Scheme modules will also be taken from that latest
2449 version. New @command{guix} sub-commands added by the update also
2452 Any user can update their Guix copy using @command{guix pull}, and the
2453 effect is limited to the user who run @command{guix pull}. For
2454 instance, when user @code{root} runs @command{guix pull}, this has no
2455 effect on the version of Guix that user @code{alice} sees, and vice
2456 versa@footnote{Under the hood, @command{guix pull} updates the
2457 @file{~/.config/guix/latest} symbolic link to point to the latest Guix,
2458 and the @command{guix} command loads code from there. Currently, the
2459 only way to roll back an invocation of @command{guix pull} is to
2460 manually update this symlink to point to the previous Guix.}.
2462 The @command{guix pull} command is usually invoked with no arguments,
2463 but it supports the following options:
2467 Produce verbose output, writing build logs to the standard error output.
2469 @item --url=@var{url}
2470 Download the source tarball of Guix from @var{url}.
2472 By default, the tarball is taken from its canonical address at
2473 @code{gnu.org}, for the stable branch of Guix.
2475 With some Git servers, this can be used to deploy any version of Guix.
2476 For example, to download and deploy version 0.12.0 of Guix from the
2480 guix pull --url=https://git.savannah.gnu.org/cgit/guix.git/snapshot/v0.12.0.tar.gz
2483 It can also be used to deploy arbitrary Git revisions:
2486 guix pull --url=https://git.savannah.gnu.org/cgit/guix.git/snapshot/74d862e8a.tar.gz
2490 Use the bootstrap Guile to build the latest Guix. This option is only
2491 useful to Guix developers.
2494 In addition, @command{guix pull} supports all the common build options
2495 (@pxref{Common Build Options}).
2497 @node Invoking guix pack
2498 @section Invoking @command{guix pack}
2500 Occasionally you want to pass software to people who are not (yet!)
2501 lucky enough to be using Guix. You'd tell them to run @command{guix
2502 package -i @var{something}}, but that's not possible in this case. This
2503 is where @command{guix pack} comes in.
2507 @cindex application bundle
2508 @cindex software bundle
2509 The @command{guix pack} command creates a shrink-wrapped @dfn{pack} or
2510 @dfn{software bundle}: it creates a tarball or some other archive
2511 containing the binaries of the software you're interested in, and all
2512 its dependencies. The resulting archive can be used on any machine that
2513 does not have Guix, and people can run the exact same binaries as those
2514 you have with Guix. The pack itself is created in a bit-reproducible
2515 fashion, so anyone can verify that it really contains the build results
2516 that you pretend to be shipping.
2518 For example, to create a bundle containing Guile, Emacs, Geiser, and all
2519 their dependencies, you can run:
2522 $ guix pack guile emacs geiser
2524 /gnu/store/@dots{}-pack.tar.gz
2527 The result here is a tarball containing a @file{/gnu/store} directory
2528 with all the relevant packages. The resulting tarball contains a
2529 @dfn{profile} with the three packages of interest; the profile is the
2530 same as would be created by @command{guix package -i}. It is this
2531 mechanism that is used to create Guix's own standalone binary tarball
2532 (@pxref{Binary Installation}).
2534 Users of this pack would have to run
2535 @file{/gnu/store/@dots{}-profile/bin/guile} to run Guile, which you may
2536 find inconvenient. To work around it, you can create, say, a
2537 @file{/opt/gnu/bin} symlink to the profile:
2540 guix pack -S /opt/gnu/bin=bin guile emacs geiser
2544 That way, users can happily type @file{/opt/gnu/bin/guile} and enjoy.
2546 Alternatively, you can produce a pack in the Docker image format using
2547 the following command:
2550 guix pack -f docker guile emacs geiser
2554 The result is a tarball that can be passed to the @command{docker load}
2556 @uref{https://docs.docker.com/engine/reference/commandline/load/, Docker
2557 documentation} for more information.
2559 Several command-line options allow you to customize your pack:
2562 @item --format=@var{format}
2563 @itemx -f @var{format}
2564 Produce a pack in the given @var{format}.
2566 The available formats are:
2570 This is the default format. It produces a tarball containing all the
2571 specifies binaries and symlinks.
2574 This produces a tarball that follows the
2575 @uref{https://github.com/docker/docker/blob/master/image/spec/v1.2.md,
2576 Docker Image Specification}.
2579 @item --expression=@var{expr}
2580 @itemx -e @var{expr}
2581 Consider the package @var{expr} evaluates to.
2583 This has the same purpose as the same-named option in @command{guix
2584 build} (@pxref{Additional Build Options, @code{--expression} in
2585 @command{guix build}}).
2587 @item --system=@var{system}
2588 @itemx -s @var{system}
2589 Attempt to build for @var{system}---e.g., @code{i686-linux}---instead of
2590 the system type of the build host.
2592 @item --target=@var{triplet}
2593 @cindex cross-compilation
2594 Cross-build for @var{triplet}, which must be a valid GNU triplet, such
2595 as @code{"mips64el-linux-gnu"} (@pxref{Specifying target triplets, GNU
2596 configuration triplets,, autoconf, Autoconf}).
2598 @item --compression=@var{tool}
2599 @itemx -C @var{tool}
2600 Compress the resulting tarball using @var{tool}---one of @code{gzip},
2601 @code{bzip2}, @code{xz}, or @code{lzip}.
2603 @item --symlink=@var{spec}
2604 @itemx -S @var{spec}
2605 Add the symlinks specified by @var{spec} to the pack. This option can
2606 appear several times.
2608 @var{spec} has the form @code{@var{source}=@var{target}}, where
2609 @var{source} is the symlink that will be created and @var{target} is the
2612 For instance, @code{-S /opt/gnu/bin=bin} creates a @file{/opt/gnu/bin}
2613 symlink pointing to the @file{bin} sub-directory of the profile.
2615 @item --localstatedir
2616 Include the ``local state directory'', @file{/var/guix}, in the
2619 @file{/var/guix} contains the store database (@pxref{The Store}) as well
2620 as garbage-collector roots (@pxref{Invoking guix gc}). Providing it in
2621 the pack means that the store is ``complete'' and manageable by Guix;
2622 not providing it pack means that the store is ``dead'': items cannot be
2623 added to it or removed from it after extraction of the pack.
2625 One use case for this is the Guix self-contained binary tarball
2626 (@pxref{Binary Installation}).
2629 In addition, @command{guix pack} supports all the common build options
2630 (@pxref{Common Build Options}) and all the package transformation
2631 options (@pxref{Package Transformation Options}).
2634 @node Invoking guix archive
2635 @section Invoking @command{guix archive}
2637 @cindex @command{guix archive}
2639 The @command{guix archive} command allows users to @dfn{export} files
2640 from the store into a single archive, and to later @dfn{import} them.
2641 In particular, it allows store files to be transferred from one machine
2642 to the store on another machine.
2644 @cindex exporting store items
2645 To export store files as an archive to standard output, run:
2648 guix archive --export @var{options} @var{specifications}...
2651 @var{specifications} may be either store file names or package
2652 specifications, as for @command{guix package} (@pxref{Invoking guix
2653 package}). For instance, the following command creates an archive
2654 containing the @code{gui} output of the @code{git} package and the main
2655 output of @code{emacs}:
2658 guix archive --export git:gui /gnu/store/...-emacs-24.3 > great.nar
2661 If the specified packages are not built yet, @command{guix archive}
2662 automatically builds them. The build process may be controlled with the
2663 common build options (@pxref{Common Build Options}).
2665 To transfer the @code{emacs} package to a machine connected over SSH,
2669 guix archive --export -r emacs | ssh the-machine guix archive --import
2673 Similarly, a complete user profile may be transferred from one machine
2674 to another like this:
2677 guix archive --export -r $(readlink -f ~/.guix-profile) | \
2678 ssh the-machine guix-archive --import
2682 However, note that, in both examples, all of @code{emacs} and the
2683 profile as well as all of their dependencies are transferred (due to
2684 @code{-r}), regardless of what is already available in the store on the
2685 target machine. The @code{--missing} option can help figure out which
2686 items are missing from the target store. The @command{guix copy}
2687 command simplifies and optimizes this whole process, so this is probably
2688 what you should use in this case (@pxref{Invoking guix copy}).
2690 @cindex nar, archive format
2691 @cindex normalized archive (nar)
2692 Archives are stored in the ``normalized archive'' or ``nar'' format, which is
2693 comparable in spirit to `tar', but with differences
2694 that make it more appropriate for our purposes. First, rather than
2695 recording all Unix metadata for each file, the nar format only mentions
2696 the file type (regular, directory, or symbolic link); Unix permissions
2697 and owner/group are dismissed. Second, the order in which directory
2698 entries are stored always follows the order of file names according to
2699 the C locale collation order. This makes archive production fully
2702 When exporting, the daemon digitally signs the contents of the archive,
2703 and that digital signature is appended. When importing, the daemon
2704 verifies the signature and rejects the import in case of an invalid
2705 signature or if the signing key is not authorized.
2706 @c FIXME: Add xref to daemon doc about signatures.
2708 The main options are:
2712 Export the specified store files or packages (see below.) Write the
2713 resulting archive to the standard output.
2715 Dependencies are @emph{not} included in the output, unless
2716 @code{--recursive} is passed.
2720 When combined with @code{--export}, this instructs @command{guix
2721 archive} to include dependencies of the given items in the archive.
2722 Thus, the resulting archive is self-contained: it contains the closure
2723 of the exported store items.
2726 Read an archive from the standard input, and import the files listed
2727 therein into the store. Abort if the archive has an invalid digital
2728 signature, or if it is signed by a public key not among the authorized
2729 keys (see @code{--authorize} below.)
2732 Read a list of store file names from the standard input, one per line,
2733 and write on the standard output the subset of these files missing from
2736 @item --generate-key[=@var{parameters}]
2737 @cindex signing, archives
2738 Generate a new key pair for the daemon. This is a prerequisite before
2739 archives can be exported with @code{--export}. Note that this operation
2740 usually takes time, because it needs to gather enough entropy to
2741 generate the key pair.
2743 The generated key pair is typically stored under @file{/etc/guix}, in
2744 @file{signing-key.pub} (public key) and @file{signing-key.sec} (private
2745 key, which must be kept secret.) When @var{parameters} is omitted,
2746 an ECDSA key using the Ed25519 curve is generated, or, for Libgcrypt
2747 versions before 1.6.0, it is a 4096-bit RSA key.
2748 Alternatively, @var{parameters} can specify
2749 @code{genkey} parameters suitable for Libgcrypt (@pxref{General
2750 public-key related Functions, @code{gcry_pk_genkey},, gcrypt, The
2751 Libgcrypt Reference Manual}).
2754 @cindex authorizing, archives
2755 Authorize imports signed by the public key passed on standard input.
2756 The public key must be in ``s-expression advanced format''---i.e., the
2757 same format as the @file{signing-key.pub} file.
2759 The list of authorized keys is kept in the human-editable file
2760 @file{/etc/guix/acl}. The file contains
2761 @url{http://people.csail.mit.edu/rivest/Sexp.txt, ``advanced-format
2762 s-expressions''} and is structured as an access-control list in the
2763 @url{http://theworld.com/~cme/spki.txt, Simple Public-Key Infrastructure
2766 @item --extract=@var{directory}
2767 @itemx -x @var{directory}
2768 Read a single-item archive as served by substitute servers
2769 (@pxref{Substitutes}) and extract it to @var{directory}. This is a
2770 low-level operation needed in only very narrow use cases; see below.
2772 For example, the following command extracts the substitute for Emacs
2773 served by @code{hydra.gnu.org} to @file{/tmp/emacs}:
2777 https://hydra.gnu.org/nar/@dots{}-emacs-24.5 \
2778 | bunzip2 | guix archive -x /tmp/emacs
2781 Single-item archives are different from multiple-item archives produced
2782 by @command{guix archive --export}; they contain a single store item,
2783 and they do @emph{not} embed a signature. Thus this operation does
2784 @emph{no} signature verification and its output should be considered
2787 The primary purpose of this operation is to facilitate inspection of
2788 archive contents coming from possibly untrusted substitute servers.
2792 @c *********************************************************************
2793 @node Programming Interface
2794 @chapter Programming Interface
2796 GNU Guix provides several Scheme programming interfaces (APIs) to
2797 define, build, and query packages. The first interface allows users to
2798 write high-level package definitions. These definitions refer to
2799 familiar packaging concepts, such as the name and version of a package,
2800 its build system, and its dependencies. These definitions can then be
2801 turned into concrete build actions.
2803 Build actions are performed by the Guix daemon, on behalf of users. In a
2804 standard setup, the daemon has write access to the store---the
2805 @file{/gnu/store} directory---whereas users do not. The recommended
2806 setup also has the daemon perform builds in chroots, under a specific
2807 build users, to minimize interference with the rest of the system.
2810 Lower-level APIs are available to interact with the daemon and the
2811 store. To instruct the daemon to perform a build action, users actually
2812 provide it with a @dfn{derivation}. A derivation is a low-level
2813 representation of the build actions to be taken, and the environment in
2814 which they should occur---derivations are to package definitions what
2815 assembly is to C programs. The term ``derivation'' comes from the fact
2816 that build results @emph{derive} from them.
2818 This chapter describes all these APIs in turn, starting from high-level
2819 package definitions.
2822 * Defining Packages:: Defining new packages.
2823 * Build Systems:: Specifying how packages are built.
2824 * The Store:: Manipulating the package store.
2825 * Derivations:: Low-level interface to package derivations.
2826 * The Store Monad:: Purely functional interface to the store.
2827 * G-Expressions:: Manipulating build expressions.
2830 @node Defining Packages
2831 @section Defining Packages
2833 The high-level interface to package definitions is implemented in the
2834 @code{(guix packages)} and @code{(guix build-system)} modules. As an
2835 example, the package definition, or @dfn{recipe}, for the GNU Hello
2836 package looks like this:
2839 (define-module (gnu packages hello)
2840 #:use-module (guix packages)
2841 #:use-module (guix download)
2842 #:use-module (guix build-system gnu)
2843 #:use-module (guix licenses)
2844 #:use-module (gnu packages gawk))
2846 (define-public hello
2852 (uri (string-append "mirror://gnu/hello/hello-" version
2856 "0ssi1wpaf7plaswqqjwigppsg5fyh99vdlb9kzl7c9lng89ndq1i"))))
2857 (build-system gnu-build-system)
2858 (arguments '(#:configure-flags '("--enable-silent-rules")))
2859 (inputs `(("gawk" ,gawk)))
2860 (synopsis "Hello, GNU world: An example GNU package")
2861 (description "Guess what GNU Hello prints!")
2862 (home-page "http://www.gnu.org/software/hello/")
2867 Without being a Scheme expert, the reader may have guessed the meaning
2868 of the various fields here. This expression binds the variable
2869 @code{hello} to a @code{<package>} object, which is essentially a record
2870 (@pxref{SRFI-9, Scheme records,, guile, GNU Guile Reference Manual}).
2871 This package object can be inspected using procedures found in the
2872 @code{(guix packages)} module; for instance, @code{(package-name hello)}
2873 returns---surprise!---@code{"hello"}.
2875 With luck, you may be able to import part or all of the definition of
2876 the package you are interested in from another repository, using the
2877 @code{guix import} command (@pxref{Invoking guix import}).
2879 In the example above, @var{hello} is defined in a module of its own,
2880 @code{(gnu packages hello)}. Technically, this is not strictly
2881 necessary, but it is convenient to do so: all the packages defined in
2882 modules under @code{(gnu packages @dots{})} are automatically known to
2883 the command-line tools (@pxref{Package Modules}).
2885 There are a few points worth noting in the above package definition:
2889 The @code{source} field of the package is an @code{<origin>} object
2890 (@pxref{origin Reference}, for the complete reference).
2891 Here, the @code{url-fetch} method from @code{(guix download)} is used,
2892 meaning that the source is a file to be downloaded over FTP or HTTP.
2894 The @code{mirror://gnu} prefix instructs @code{url-fetch} to use one of
2895 the GNU mirrors defined in @code{(guix download)}.
2897 The @code{sha256} field specifies the expected SHA256 hash of the file
2898 being downloaded. It is mandatory, and allows Guix to check the
2899 integrity of the file. The @code{(base32 @dots{})} form introduces the
2900 base32 representation of the hash. You can obtain this information with
2901 @code{guix download} (@pxref{Invoking guix download}) and @code{guix
2902 hash} (@pxref{Invoking guix hash}).
2905 When needed, the @code{origin} form can also have a @code{patches} field
2906 listing patches to be applied, and a @code{snippet} field giving a
2907 Scheme expression to modify the source code.
2910 @cindex GNU Build System
2911 The @code{build-system} field specifies the procedure to build the
2912 package (@pxref{Build Systems}). Here, @var{gnu-build-system}
2913 represents the familiar GNU Build System, where packages may be
2914 configured, built, and installed with the usual @code{./configure &&
2915 make && make check && make install} command sequence.
2918 The @code{arguments} field specifies options for the build system
2919 (@pxref{Build Systems}). Here it is interpreted by
2920 @var{gnu-build-system} as a request run @file{configure} with the
2921 @code{--enable-silent-rules} flag.
2927 What about these quote (@code{'}) characters? They are Scheme syntax to
2928 introduce a literal list; @code{'} is synonymous with @code{quote}.
2929 @xref{Expression Syntax, quoting,, guile, GNU Guile Reference Manual},
2930 for details. Here the value of the @code{arguments} field is a list of
2931 arguments passed to the build system down the road, as with @code{apply}
2932 (@pxref{Fly Evaluation, @code{apply},, guile, GNU Guile Reference
2935 The hash-colon (@code{#:}) sequence defines a Scheme @dfn{keyword}
2936 (@pxref{Keywords,,, guile, GNU Guile Reference Manual}), and
2937 @code{#:configure-flags} is a keyword used to pass a keyword argument
2938 to the build system (@pxref{Coding With Keywords,,, guile, GNU Guile
2942 The @code{inputs} field specifies inputs to the build process---i.e.,
2943 build-time or run-time dependencies of the package. Here, we define an
2944 input called @code{"gawk"} whose value is that of the @var{gawk}
2945 variable; @var{gawk} is itself bound to a @code{<package>} object.
2947 @cindex backquote (quasiquote)
2950 @cindex comma (unquote)
2954 @findex unquote-splicing
2955 Again, @code{`} (a backquote, synonymous with @code{quasiquote}) allows
2956 us to introduce a literal list in the @code{inputs} field, while
2957 @code{,} (a comma, synonymous with @code{unquote}) allows us to insert a
2958 value in that list (@pxref{Expression Syntax, unquote,, guile, GNU Guile
2961 Note that GCC, Coreutils, Bash, and other essential tools do not need to
2962 be specified as inputs here. Instead, @var{gnu-build-system} takes care
2963 of ensuring that they are present (@pxref{Build Systems}).
2965 However, any other dependencies need to be specified in the
2966 @code{inputs} field. Any dependency not specified here will simply be
2967 unavailable to the build process, possibly leading to a build failure.
2970 @xref{package Reference}, for a full description of possible fields.
2972 Once a package definition is in place, the
2973 package may actually be built using the @code{guix build} command-line
2974 tool (@pxref{Invoking guix build}), troubleshooting any build failures
2975 you encounter (@pxref{Debugging Build Failures}). You can easily jump back to the
2976 package definition using the @command{guix edit} command
2977 (@pxref{Invoking guix edit}).
2978 @xref{Packaging Guidelines}, for
2979 more information on how to test package definitions, and
2980 @ref{Invoking guix lint}, for information on how to check a definition
2981 for style conformance.
2982 @vindex GUIX_PACKAGE_PATH
2983 Lastly, @pxref{Package Modules}, for information
2984 on how to extend the distribution by adding your own package definitions
2985 to @code{GUIX_PACKAGE_PATH}.
2987 Finally, updating the package definition to a new upstream version
2988 can be partly automated by the @command{guix refresh} command
2989 (@pxref{Invoking guix refresh}).
2991 Behind the scenes, a derivation corresponding to the @code{<package>}
2992 object is first computed by the @code{package-derivation} procedure.
2993 That derivation is stored in a @code{.drv} file under @file{/gnu/store}.
2994 The build actions it prescribes may then be realized by using the
2995 @code{build-derivations} procedure (@pxref{The Store}).
2997 @deffn {Scheme Procedure} package-derivation @var{store} @var{package} [@var{system}]
2998 Return the @code{<derivation>} object of @var{package} for @var{system}
2999 (@pxref{Derivations}).
3001 @var{package} must be a valid @code{<package>} object, and @var{system}
3002 must be a string denoting the target system type---e.g.,
3003 @code{"x86_64-linux"} for an x86_64 Linux-based GNU system. @var{store}
3004 must be a connection to the daemon, which operates on the store
3005 (@pxref{The Store}).
3009 @cindex cross-compilation
3010 Similarly, it is possible to compute a derivation that cross-builds a
3011 package for some other system:
3013 @deffn {Scheme Procedure} package-cross-derivation @var{store} @
3014 @var{package} @var{target} [@var{system}]
3015 Return the @code{<derivation>} object of @var{package} cross-built from
3016 @var{system} to @var{target}.
3018 @var{target} must be a valid GNU triplet denoting the target hardware
3019 and operating system, such as @code{"mips64el-linux-gnu"}
3020 (@pxref{Configuration Names, GNU configuration triplets,, configure, GNU
3021 Configure and Build System}).
3024 @cindex package transformations
3025 @cindex input rewriting
3026 @cindex dependency tree rewriting
3027 Packages can be manipulated in arbitrary ways. An example of a useful
3028 transformation is @dfn{input rewriting}, whereby the dependency tree of
3029 a package is rewritten by replacing specific inputs by others:
3031 @deffn {Scheme Procedure} package-input-rewriting @var{replacements} @
3032 [@var{rewrite-name}]
3033 Return a procedure that, when passed a package, replaces its direct and
3034 indirect dependencies (but not its implicit inputs) according to
3035 @var{replacements}. @var{replacements} is a list of package pairs; the
3036 first element of each pair is the package to replace, and the second one
3039 Optionally, @var{rewrite-name} is a one-argument procedure that takes
3040 the name of a package and returns its new name after rewrite.
3044 Consider this example:
3047 (define libressl-instead-of-openssl
3048 ;; This is a procedure to replace OPENSSL by LIBRESSL,
3050 (package-input-rewriting `((,openssl . ,libressl))))
3052 (define git-with-libressl
3053 (libressl-instead-of-openssl git))
3057 Here we first define a rewriting procedure that replaces @var{openssl}
3058 with @var{libressl}. Then we use it to define a @dfn{variant} of the
3059 @var{git} package that uses @var{libressl} instead of @var{openssl}.
3060 This is exactly what the @option{--with-input} command-line option does
3061 (@pxref{Package Transformation Options, @option{--with-input}}).
3063 A more generic procedure to rewrite a package dependency graph is
3064 @code{package-mapping}: it supports arbitrary changes to nodes in the
3067 @deffn {Scheme Procedure} package-mapping @var{proc} [@var{cut?}]
3068 Return a procedure that, given a package, applies @var{proc} to all the packages
3069 depended on and returns the resulting package. The procedure stops recursion
3070 when @var{cut?} returns true for a given package.
3074 * package Reference :: The package data type.
3075 * origin Reference:: The origin data type.
3079 @node package Reference
3080 @subsection @code{package} Reference
3082 This section summarizes all the options available in @code{package}
3083 declarations (@pxref{Defining Packages}).
3085 @deftp {Data Type} package
3086 This is the data type representing a package recipe.
3090 The name of the package, as a string.
3092 @item @code{version}
3093 The version of the package, as a string.
3096 An object telling how the source code for the package should be
3097 acquired. Most of the time, this is an @code{origin} object, which
3098 denotes a file fetched from the Internet (@pxref{origin Reference}). It
3099 can also be any other ``file-like'' object such as a @code{local-file},
3100 which denotes a file from the local file system (@pxref{G-Expressions,
3101 @code{local-file}}).
3103 @item @code{build-system}
3104 The build system that should be used to build the package (@pxref{Build
3107 @item @code{arguments} (default: @code{'()})
3108 The arguments that should be passed to the build system. This is a
3109 list, typically containing sequential keyword-value pairs.
3111 @item @code{inputs} (default: @code{'()})
3112 @itemx @code{native-inputs} (default: @code{'()})
3113 @itemx @code{propagated-inputs} (default: @code{'()})
3114 @cindex inputs, of packages
3115 These fields list dependencies of the package. Each one is a list of
3116 tuples, where each tuple has a label for the input (a string) as its
3117 first element, a package, origin, or derivation as its second element,
3118 and optionally the name of the output thereof that should be used, which
3119 defaults to @code{"out"} (@pxref{Packages with Multiple Outputs}, for
3120 more on package outputs). For example, the list below specifies three
3124 `(("libffi" ,libffi)
3125 ("libunistring" ,libunistring)
3126 ("glib:bin" ,glib "bin")) ;the "bin" output of Glib
3129 @cindex cross compilation, package dependencies
3130 The distinction between @code{native-inputs} and @code{inputs} is
3131 necessary when considering cross-compilation. When cross-compiling,
3132 dependencies listed in @code{inputs} are built for the @emph{target}
3133 architecture; conversely, dependencies listed in @code{native-inputs}
3134 are built for the architecture of the @emph{build} machine.
3136 @code{native-inputs} is typically used to list tools needed at
3137 build time, but not at run time, such as Autoconf, Automake, pkg-config,
3138 Gettext, or Bison. @command{guix lint} can report likely mistakes in
3139 this area (@pxref{Invoking guix lint}).
3141 @anchor{package-propagated-inputs}
3142 Lastly, @code{propagated-inputs} is similar to @code{inputs}, but the
3143 specified packages will be automatically installed alongside the package
3144 they belong to (@pxref{package-cmd-propagated-inputs, @command{guix
3145 package}}, for information on how @command{guix package} deals with
3148 For example this is necessary when a C/C++ library needs headers of
3149 another library to compile, or when a pkg-config file refers to another
3150 one @i{via} its @code{Requires} field.
3152 Another example where @code{propagated-inputs} is useful is for languages
3153 that lack a facility to record the run-time search path akin to the
3154 @code{RUNPATH} of ELF files; this includes Guile, Python, Perl, and
3155 more. To ensure that libraries written in those languages can find
3156 library code they depend on at run time, run-time dependencies must be
3157 listed in @code{propagated-inputs} rather than @code{inputs}.
3159 @item @code{self-native-input?} (default: @code{#f})
3160 This is a Boolean field telling whether the package should use itself as
3161 a native input when cross-compiling.
3163 @item @code{outputs} (default: @code{'("out")})
3164 The list of output names of the package. @xref{Packages with Multiple
3165 Outputs}, for typical uses of additional outputs.
3167 @item @code{native-search-paths} (default: @code{'()})
3168 @itemx @code{search-paths} (default: @code{'()})
3169 A list of @code{search-path-specification} objects describing
3170 search-path environment variables honored by the package.
3172 @item @code{replacement} (default: @code{#f})
3173 This must be either @code{#f} or a package object that will be used as a
3174 @dfn{replacement} for this package. @xref{Security Updates, grafts},
3177 @item @code{synopsis}
3178 A one-line description of the package.
3180 @item @code{description}
3181 A more elaborate description of the package.
3183 @item @code{license}
3184 @cindex license, of packages
3185 The license of the package; a value from @code{(guix licenses)},
3186 or a list of such values.
3188 @item @code{home-page}
3189 The URL to the home-page of the package, as a string.
3191 @item @code{supported-systems} (default: @var{%supported-systems})
3192 The list of systems supported by the package, as strings of the form
3193 @code{architecture-kernel}, for example @code{"x86_64-linux"}.
3195 @item @code{maintainers} (default: @code{'()})
3196 The list of maintainers of the package, as @code{maintainer} objects.
3198 @item @code{location} (default: source location of the @code{package} form)
3199 The source location of the package. It is useful to override this when
3200 inheriting from another package, in which case this field is not
3201 automatically corrected.
3206 @node origin Reference
3207 @subsection @code{origin} Reference
3209 This section summarizes all the options available in @code{origin}
3210 declarations (@pxref{Defining Packages}).
3212 @deftp {Data Type} origin
3213 This is the data type representing a source code origin.
3217 An object containing the URI of the source. The object type depends on
3218 the @code{method} (see below). For example, when using the
3219 @var{url-fetch} method of @code{(guix download)}, the valid @code{uri}
3220 values are: a URL represented as a string, or a list thereof.
3223 A procedure that handles the URI.
3228 @item @var{url-fetch} from @code{(guix download)}
3229 download a file from the HTTP, HTTPS, or FTP URL specified in the
3233 @item @var{git-fetch} from @code{(guix git-download)}
3234 clone the Git version control repository, and check out the revision
3235 specified in the @code{uri} field as a @code{git-reference} object; a
3236 @code{git-reference} looks like this:
3240 (url "git://git.debian.org/git/pkg-shadow/shadow")
3241 (commit "v4.1.5.1"))
3246 A bytevector containing the SHA-256 hash of the source. Typically the
3247 @code{base32} form is used here to generate the bytevector from a
3250 You can obtain this information using @code{guix download}
3251 (@pxref{Invoking guix download}) or @code{guix hash} (@pxref{Invoking
3254 @item @code{file-name} (default: @code{#f})
3255 The file name under which the source code should be saved. When this is
3256 @code{#f}, a sensible default value will be used in most cases. In case
3257 the source is fetched from a URL, the file name from the URL will be
3258 used. For version control checkouts, it is recommended to provide the
3259 file name explicitly because the default is not very descriptive.
3261 @item @code{patches} (default: @code{'()})
3262 A list of file names containing patches to be applied to the source.
3264 This list of patches must be unconditional. In particular, it cannot
3265 depend on the value of @code{%current-system} or
3266 @code{%current-target-system}.
3268 @item @code{snippet} (default: @code{#f})
3269 A G-expression (@pxref{G-Expressions}) or S-expression that will be run
3270 in the source directory. This is a convenient way to modify the source,
3271 sometimes more convenient than a patch.
3273 @item @code{patch-flags} (default: @code{'("-p1")})
3274 A list of command-line flags that should be passed to the @code{patch}
3277 @item @code{patch-inputs} (default: @code{#f})
3278 Input packages or derivations to the patching process. When this is
3279 @code{#f}, the usual set of inputs necessary for patching are provided,
3280 such as GNU@tie{}Patch.
3282 @item @code{modules} (default: @code{'()})
3283 A list of Guile modules that should be loaded during the patching
3284 process and while running the code in the @code{snippet} field.
3286 @item @code{patch-guile} (default: @code{#f})
3287 The Guile package that should be used in the patching process. When
3288 this is @code{#f}, a sensible default is used.
3294 @section Build Systems
3296 @cindex build system
3297 Each package definition specifies a @dfn{build system} and arguments for
3298 that build system (@pxref{Defining Packages}). This @code{build-system}
3299 field represents the build procedure of the package, as well as implicit
3300 dependencies of that build procedure.
3302 Build systems are @code{<build-system>} objects. The interface to
3303 create and manipulate them is provided by the @code{(guix build-system)}
3304 module, and actual build systems are exported by specific modules.
3306 @cindex bag (low-level package representation)
3307 Under the hood, build systems first compile package objects to
3308 @dfn{bags}. A @dfn{bag} is like a package, but with less
3309 ornamentation---in other words, a bag is a lower-level representation of
3310 a package, which includes all the inputs of that package, including some
3311 that were implicitly added by the build system. This intermediate
3312 representation is then compiled to a derivation (@pxref{Derivations}).
3314 Build systems accept an optional list of @dfn{arguments}. In package
3315 definitions, these are passed @i{via} the @code{arguments} field
3316 (@pxref{Defining Packages}). They are typically keyword arguments
3317 (@pxref{Optional Arguments, keyword arguments in Guile,, guile, GNU
3318 Guile Reference Manual}). The value of these arguments is usually
3319 evaluated in the @dfn{build stratum}---i.e., by a Guile process launched
3320 by the daemon (@pxref{Derivations}).
3322 The main build system is @var{gnu-build-system}, which implements the
3323 standard build procedure for GNU and many other packages. It
3324 is provided by the @code{(guix build-system gnu)} module.
3326 @defvr {Scheme Variable} gnu-build-system
3327 @var{gnu-build-system} represents the GNU Build System, and variants
3328 thereof (@pxref{Configuration, configuration and makefile conventions,,
3329 standards, GNU Coding Standards}).
3331 @cindex build phases
3332 In a nutshell, packages using it are configured, built, and installed with
3333 the usual @code{./configure && make && make check && make install}
3334 command sequence. In practice, a few additional steps are often needed.
3335 All these steps are split up in separate @dfn{phases},
3336 notably@footnote{Please see the @code{(guix build gnu-build-system)}
3337 modules for more details about the build phases.}:
3341 Unpack the source tarball, and change the current directory to the
3342 extracted source tree. If the source is actually a directory, copy it
3343 to the build tree, and enter that directory.
3345 @item patch-source-shebangs
3346 Patch shebangs encountered in source files so they refer to the right
3347 store file names. For instance, this changes @code{#!/bin/sh} to
3348 @code{#!/gnu/store/@dots{}-bash-4.3/bin/sh}.
3351 Run the @file{configure} script with a number of default options, such
3352 as @code{--prefix=/gnu/store/@dots{}}, as well as the options specified
3353 by the @code{#:configure-flags} argument.
3356 Run @code{make} with the list of flags specified with
3357 @code{#:make-flags}. If the @code{#:parallel-build?} argument is true
3358 (the default), build with @code{make -j}.
3361 Run @code{make check}, or some other target specified with
3362 @code{#:test-target}, unless @code{#:tests? #f} is passed. If the
3363 @code{#:parallel-tests?} argument is true (the default), run @code{make
3367 Run @code{make install} with the flags listed in @code{#:make-flags}.
3369 @item patch-shebangs
3370 Patch shebangs on the installed executable files.
3373 Strip debugging symbols from ELF files (unless @code{#:strip-binaries?}
3374 is false), copying them to the @code{debug} output when available
3375 (@pxref{Installing Debugging Files}).
3378 @vindex %standard-phases
3379 The build-side module @code{(guix build gnu-build-system)} defines
3380 @var{%standard-phases} as the default list of build phases.
3381 @var{%standard-phases} is a list of symbol/procedure pairs, where the
3382 procedure implements the actual phase.
3384 The list of phases used for a particular package can be changed with the
3385 @code{#:phases} parameter. For instance, passing:
3388 #:phases (modify-phases %standard-phases (delete 'configure))
3391 means that all the phases described above will be used, except the
3392 @code{configure} phase.
3394 In addition, this build system ensures that the ``standard'' environment
3395 for GNU packages is available. This includes tools such as GCC, libc,
3396 Coreutils, Bash, Make, Diffutils, grep, and sed (see the @code{(guix
3397 build-system gnu)} module for a complete list). We call these the
3398 @dfn{implicit inputs} of a package, because package definitions do not
3399 have to mention them.
3402 Other @code{<build-system>} objects are defined to support other
3403 conventions and tools used by free software packages. They inherit most
3404 of @var{gnu-build-system}, and differ mainly in the set of inputs
3405 implicitly added to the build process, and in the list of phases
3406 executed. Some of these build systems are listed below.
3408 @defvr {Scheme Variable} ant-build-system
3409 This variable is exported by @code{(guix build-system ant)}. It
3410 implements the build procedure for Java packages that can be built with
3411 @url{http://ant.apache.org/, Ant build tool}.
3413 It adds both @code{ant} and the @dfn{Java Development Kit} (JDK) as
3414 provided by the @code{icedtea} package to the set of inputs. Different
3415 packages can be specified with the @code{#:ant} and @code{#:jdk}
3416 parameters, respectively.
3418 When the original package does not provide a suitable Ant build file,
3419 the parameter @code{#:jar-name} can be used to generate a minimal Ant
3420 build file @file{build.xml} with tasks to build the specified jar
3421 archive. In this case the parameter @code{#:source-dir} can be used to
3422 specify the source sub-directory, defaulting to ``src''.
3424 The parameter @code{#:build-target} can be used to specify the Ant task
3425 that should be run during the @code{build} phase. By default the
3426 ``jar'' task will be run.
3430 @defvr {Scheme Variable} asdf-build-system/source
3431 @defvrx {Scheme Variable} asdf-build-system/sbcl
3432 @defvrx {Scheme Variable} asdf-build-system/ecl
3434 These variables, exported by @code{(guix build-system asdf)}, implement
3435 build procedures for Common Lisp packages using
3436 @url{https://common-lisp.net/project/asdf/, ``ASDF''}. ASDF is a system
3437 definition facility for Common Lisp programs and libraries.
3439 The @code{asdf-build-system/source} system installs the packages in
3440 source form, and can be loaded using any common lisp implementation, via
3441 ASDF. The others, such as @code{asdf-build-system/sbcl}, install binary
3442 systems in the format which a particular implementation understands.
3443 These build systems can also be used to produce executable programs, or
3444 lisp images which contain a set of packages pre-loaded.
3446 The build system uses naming conventions. For binary packages, the
3447 package name should be prefixed with the lisp implementation, such as
3448 @code{sbcl-} for @code{asdf-build-system/sbcl}.
3450 Additionally, the corresponding source package should be labeled using
3451 the same convention as python packages (see @ref{Python Modules}), using
3452 the @code{cl-} prefix.
3454 For binary packages, each system should be defined as a Guix package.
3455 If one package @code{origin} contains several systems, package variants
3456 can be created in order to build all the systems. Source packages,
3457 which use @code{asdf-build-system/source}, may contain several systems.
3459 In order to create executable programs and images, the build-side
3460 procedures @code{build-program} and @code{build-image} can be used.
3461 They should be called in a build phase after the @code{create-symlinks}
3462 phase, so that the system which was just built can be used within the
3463 resulting image. @code{build-program} requires a list of Common Lisp
3464 expressions to be passed as the @code{#:entry-program} argument.
3466 If the system is not defined within its own @code{.asd} file of the same
3467 name, then the @code{#:asd-file} parameter should be used to specify
3468 which file the system is defined in. Furthermore, if the package
3469 defines a system for its tests in a separate file, it will be loaded
3470 before the tests are run if it is specified by the
3471 @code{#:test-asd-file} parameter. If it is not set, the files
3472 @code{<system>-tests.asd}, @code{<system>-test.asd}, @code{tests.asd},
3473 and @code{test.asd} will be tried if they exist.
3475 If for some reason the package must be named in a different way than the
3476 naming conventions suggest, the @code{#:asd-system-name} parameter can
3477 be used to specify the name of the system.
3481 @defvr {Scheme Variable} cargo-build-system
3482 @cindex Rust programming language
3483 @cindex Cargo (Rust build system)
3484 This variable is exported by @code{(guix build-system cargo)}. It
3485 supports builds of packages using Cargo, the build tool of the
3486 @uref{https://www.rust-lang.org, Rust programming language}.
3488 In its @code{configure} phase, this build system replaces dependencies
3489 specified in the @file{Carto.toml} file with inputs to the Guix package.
3490 The @code{install} phase installs the binaries, and it also installs the
3491 source code and @file{Cargo.toml} file.
3494 @defvr {Scheme Variable} cmake-build-system
3495 This variable is exported by @code{(guix build-system cmake)}. It
3496 implements the build procedure for packages using the
3497 @url{http://www.cmake.org, CMake build tool}.
3499 It automatically adds the @code{cmake} package to the set of inputs.
3500 Which package is used can be specified with the @code{#:cmake}
3503 The @code{#:configure-flags} parameter is taken as a list of flags
3504 passed to the @command{cmake} command. The @code{#:build-type}
3505 parameter specifies in abstract terms the flags passed to the compiler;
3506 it defaults to @code{"RelWithDebInfo"} (short for ``release mode with
3507 debugging information''), which roughly means that code is compiled with
3508 @code{-O2 -g}, as is the case for Autoconf-based packages by default.
3511 @defvr {Scheme Variable} glib-or-gtk-build-system
3512 This variable is exported by @code{(guix build-system glib-or-gtk)}. It
3513 is intended for use with packages making use of GLib or GTK+.
3515 This build system adds the following two phases to the ones defined by
3516 @var{gnu-build-system}:
3519 @item glib-or-gtk-wrap
3520 The phase @code{glib-or-gtk-wrap} ensures that programs in
3521 @file{bin/} are able to find GLib ``schemas'' and
3522 @uref{https://developer.gnome.org/gtk3/stable/gtk-running.html, GTK+
3523 modules}. This is achieved by wrapping the programs in launch scripts
3524 that appropriately set the @code{XDG_DATA_DIRS} and @code{GTK_PATH}
3525 environment variables.
3527 It is possible to exclude specific package outputs from that wrapping
3528 process by listing their names in the
3529 @code{#:glib-or-gtk-wrap-excluded-outputs} parameter. This is useful
3530 when an output is known not to contain any GLib or GTK+ binaries, and
3531 where wrapping would gratuitously add a dependency of that output on
3534 @item glib-or-gtk-compile-schemas
3535 The phase @code{glib-or-gtk-compile-schemas} makes sure that all
3536 @uref{https://developer.gnome.org/gio/stable/glib-compile-schemas.html,
3537 GSettings schemas} of GLib are compiled. Compilation is performed by the
3538 @command{glib-compile-schemas} program. It is provided by the package
3539 @code{glib:bin} which is automatically imported by the build system.
3540 The @code{glib} package providing @command{glib-compile-schemas} can be
3541 specified with the @code{#:glib} parameter.
3544 Both phases are executed after the @code{install} phase.
3547 @defvr {Scheme Variable} ocaml-build-system
3548 This variable is exported by @code{(guix build-system ocaml)}. It implements
3549 a build procedure for @uref{https://ocaml.org, OCaml} packages, which consists
3550 of choosing the correct set of commands to run for each package. OCaml
3551 packages can expect many different commands to be run. This build system will
3554 When the package has a @file{setup.ml} file present at the top-level, it will
3555 run @code{ocaml setup.ml -configure}, @code{ocaml setup.ml -build} and
3556 @code{ocaml setup.ml -install}. The build system will assume that this file
3557 was generated by @uref{http://oasis.forge.ocamlcore.org/, OASIS} and will take
3558 care of setting the prefix and enabling tests if they are not disabled. You
3559 can pass configure and build flags with the @code{#:configure-flags} and
3560 @code{#:build-flags}. The @code{#:test-flags} key can be passed to change the
3561 set of flags used to enable tests. The @code{#:use-make?} key can be used to
3562 bypass this system in the build and install phases.
3564 When the package has a @file{configure} file, it is assumed that it is a
3565 hand-made configure script that requires a different argument format than
3566 in the @code{gnu-build-system}. You can add more flags with the
3567 @code{#:configure-flags} key.
3569 When the package has a @file{Makefile} file (or @code{#:use-make?} is
3570 @code{#t}), it will be used and more flags can be passed to the build and
3571 install phases with the @code{#:make-flags} key.
3573 Finally, some packages do not have these files and use a somewhat standard
3574 location for its build system. In that case, the build system will run
3575 @code{ocaml pkg/pkg.ml} or @code{ocaml pkg/build.ml} and take care of
3576 providing the path to the required findlib module. Additional flags can
3577 be passed via the @code{#:build-flags} key. Install is taken care of by
3578 @command{opam-installer}. In this case, the @code{opam} package must
3579 be added to the @code{native-inputs} field of the package definition.
3581 Note that most OCaml packages assume they will be installed in the same
3582 directory as OCaml, which is not what we want in guix. In particular, they
3583 will install @file{.so} files in their module's directory, which is usually
3584 fine because it is in the OCaml compiler directory. In guix though, these
3585 libraries cannot be found and we use @code{CAML_LD_LIBRARY_PATH}. This
3586 variable points to @file{lib/ocaml/site-lib/stubslibs} and this is where
3587 @file{.so} libraries should be installed.
3590 @defvr {Scheme Variable} python-build-system
3591 This variable is exported by @code{(guix build-system python)}. It
3592 implements the more or less standard build procedure used by Python
3593 packages, which consists in running @code{python setup.py build} and
3594 then @code{python setup.py install --prefix=/gnu/store/@dots{}}.
3596 For packages that install stand-alone Python programs under @code{bin/},
3597 it takes care of wrapping these programs so that their @code{PYTHONPATH}
3598 environment variable points to all the Python libraries they depend on.
3600 Which Python package is used to perform the build can be specified with
3601 the @code{#:python} parameter. This is a useful way to force a package
3602 to be built for a specific version of the Python interpreter, which
3603 might be necessary if the package is only compatible with a single
3604 interpreter version.
3606 By default guix calls @code{setup.py} under control of
3607 @code{setuptools}, much like @command{pip} does. Some packages are not
3608 compatible with setuptools (and pip), thus you can disable this by
3609 setting the @code{#:use-setuptools} parameter to @code{#f}.
3612 @defvr {Scheme Variable} perl-build-system
3613 This variable is exported by @code{(guix build-system perl)}. It
3614 implements the standard build procedure for Perl packages, which either
3615 consists in running @code{perl Build.PL --prefix=/gnu/store/@dots{}},
3616 followed by @code{Build} and @code{Build install}; or in running
3617 @code{perl Makefile.PL PREFIX=/gnu/store/@dots{}}, followed by
3618 @code{make} and @code{make install}, depending on which of
3619 @code{Build.PL} or @code{Makefile.PL} is present in the package
3620 distribution. Preference is given to the former if both @code{Build.PL}
3621 and @code{Makefile.PL} exist in the package distribution. This
3622 preference can be reversed by specifying @code{#t} for the
3623 @code{#:make-maker?} parameter.
3625 The initial @code{perl Makefile.PL} or @code{perl Build.PL} invocation
3626 passes flags specified by the @code{#:make-maker-flags} or
3627 @code{#:module-build-flags} parameter, respectively.
3629 Which Perl package is used can be specified with @code{#:perl}.
3632 @defvr {Scheme Variable} r-build-system
3633 This variable is exported by @code{(guix build-system r)}. It
3634 implements the build procedure used by @uref{http://r-project.org, R}
3635 packages, which essentially is little more than running @code{R CMD
3636 INSTALL --library=/gnu/store/@dots{}} in an environment where
3637 @code{R_LIBS_SITE} contains the paths to all R package inputs. Tests
3638 are run after installation using the R function
3639 @code{tools::testInstalledPackage}.
3642 @defvr {Scheme Variable} texlive-build-system
3643 This variable is exported by @code{(guix build-system texlive)}. It is
3644 used to build TeX packages in batch mode with a specified engine. The
3645 build system sets the @code{TEXINPUTS} variable to find all TeX source
3646 files in the inputs.
3648 By default it runs @code{luatex} on all files ending on @code{ins}. A
3649 different engine and format can be specified with the
3650 @code{#:tex-format} argument. Different build targets can be specified
3651 with the @code{#:build-targets} argument, which expects a list of file
3652 names. The build system adds only @code{texlive-bin} and
3653 @code{texlive-latex-base} (both from @code{(gnu packages tex}) to the
3654 inputs. Both can be overridden with the arguments @code{#:texlive-bin}
3655 and @code{#:texlive-latex-base}, respectively.
3657 The @code{#:tex-directory} parameter tells the build system where to
3658 install the built files under the texmf tree.
3661 @defvr {Scheme Variable} ruby-build-system
3662 This variable is exported by @code{(guix build-system ruby)}. It
3663 implements the RubyGems build procedure used by Ruby packages, which
3664 involves running @code{gem build} followed by @code{gem install}.
3666 The @code{source} field of a package that uses this build system
3667 typically references a gem archive, since this is the format that Ruby
3668 developers use when releasing their software. The build system unpacks
3669 the gem archive, potentially patches the source, runs the test suite,
3670 repackages the gem, and installs it. Additionally, directories and
3671 tarballs may be referenced to allow building unreleased gems from Git or
3672 a traditional source release tarball.
3674 Which Ruby package is used can be specified with the @code{#:ruby}
3675 parameter. A list of additional flags to be passed to the @command{gem}
3676 command can be specified with the @code{#:gem-flags} parameter.
3679 @defvr {Scheme Variable} waf-build-system
3680 This variable is exported by @code{(guix build-system waf)}. It
3681 implements a build procedure around the @code{waf} script. The common
3682 phases---@code{configure}, @code{build}, and @code{install}---are
3683 implemented by passing their names as arguments to the @code{waf}
3686 The @code{waf} script is executed by the Python interpreter. Which
3687 Python package is used to run the script can be specified with the
3688 @code{#:python} parameter.
3691 @defvr {Scheme Variable} haskell-build-system
3692 This variable is exported by @code{(guix build-system haskell)}. It
3693 implements the Cabal build procedure used by Haskell packages, which
3694 involves running @code{runhaskell Setup.hs configure
3695 --prefix=/gnu/store/@dots{}} and @code{runhaskell Setup.hs build}.
3696 Instead of installing the package by running @code{runhaskell Setup.hs
3697 install}, to avoid trying to register libraries in the read-only
3698 compiler store directory, the build system uses @code{runhaskell
3699 Setup.hs copy}, followed by @code{runhaskell Setup.hs register}. In
3700 addition, the build system generates the package documentation by
3701 running @code{runhaskell Setup.hs haddock}, unless @code{#:haddock? #f}
3702 is passed. Optional Haddock parameters can be passed with the help of
3703 the @code{#:haddock-flags} parameter. If the file @code{Setup.hs} is
3704 not found, the build system looks for @code{Setup.lhs} instead.
3706 Which Haskell compiler is used can be specified with the @code{#:haskell}
3707 parameter which defaults to @code{ghc}.
3710 @defvr {Scheme Variable} dub-build-system
3711 This variable is exported by @code{(guix build-system dub)}. It
3712 implements the Dub build procedure used by D packages, which
3713 involves running @code{dub build} and @code{dub run}.
3714 Installation is done by copying the files manually.
3716 Which D compiler is used can be specified with the @code{#:ldc}
3717 parameter which defaults to @code{ldc}.
3720 @defvr {Scheme Variable} emacs-build-system
3721 This variable is exported by @code{(guix build-system emacs)}. It
3722 implements an installation procedure similar to the packaging system
3723 of Emacs itself (@pxref{Packages,,, emacs, The GNU Emacs Manual}).
3725 It first creates the @code{@var{package}-autoloads.el} file, then it
3726 byte compiles all Emacs Lisp files. Differently from the Emacs
3727 packaging system, the Info documentation files are moved to the standard
3728 documentation directory and the @file{dir} file is deleted. Each
3729 package is installed in its own directory under
3730 @file{share/emacs/site-lisp/guix.d}.
3733 @defvr {Scheme Variable} font-build-system
3734 This variable is exported by @code{(guix build-system font)}. It
3735 implements an installation procedure for font packages where upstream
3736 provides pre-compiled TrueType, OpenType, etc. font files that merely
3737 need to be copied into place. It copies font files to standard
3738 locations in the output directory.
3741 Lastly, for packages that do not need anything as sophisticated, a
3742 ``trivial'' build system is provided. It is trivial in the sense that
3743 it provides basically no support: it does not pull any implicit inputs,
3744 and does not have a notion of build phases.
3746 @defvr {Scheme Variable} trivial-build-system
3747 This variable is exported by @code{(guix build-system trivial)}.
3749 This build system requires a @code{#:builder} argument. This argument
3750 must be a Scheme expression that builds the package output(s)---as
3751 with @code{build-expression->derivation} (@pxref{Derivations,
3752 @code{build-expression->derivation}}).
3762 Conceptually, the @dfn{store} is the place where derivations that have
3763 been built successfully are stored---by default, @file{/gnu/store}.
3764 Sub-directories in the store are referred to as @dfn{store items} or
3765 sometimes @dfn{store paths}. The store has an associated database that
3766 contains information such as the store paths referred to by each store
3767 path, and the list of @emph{valid} store items---results of successful
3768 builds. This database resides in @file{@var{localstatedir}/guix/db},
3769 where @var{localstatedir} is the state directory specified @i{via}
3770 @option{--localstatedir} at configure time, usually @file{/var}.
3772 The store is @emph{always} accessed by the daemon on behalf of its clients
3773 (@pxref{Invoking guix-daemon}). To manipulate the store, clients
3774 connect to the daemon over a Unix-domain socket, send requests to it,
3775 and read the result---these are remote procedure calls, or RPCs.
3778 Users must @emph{never} modify files under @file{/gnu/store} directly.
3779 This would lead to inconsistencies and break the immutability
3780 assumptions of Guix's functional model (@pxref{Introduction}).
3782 @xref{Invoking guix gc, @command{guix gc --verify}}, for information on
3783 how to check the integrity of the store and attempt recovery from
3784 accidental modifications.
3787 The @code{(guix store)} module provides procedures to connect to the
3788 daemon, and to perform RPCs. These are described below. By default,
3789 @code{open-connection}, and thus all the @command{guix} commands,
3790 connect to the local daemon or to the URI specified by the
3791 @code{GUIX_DAEMON_SOCKET} environment variable.
3793 @defvr {Environment Variable} GUIX_DAEMON_SOCKET
3794 When set, the value of this variable should be a file name or a URI
3795 designating the daemon endpoint. When it is a file name, it denotes a
3796 Unix-domain socket to connect to. In addition to file names, the
3797 supported URI schemes are:
3802 These are for Unix-domain sockets.
3803 @code{file:///var/guix/daemon-socket/socket} is equivalent to
3804 @file{/var/guix/daemon-socket/socket}.
3807 @cindex daemon, remote access
3808 @cindex remote access to the daemon
3809 @cindex daemon, cluster setup
3810 @cindex clusters, daemon setup
3811 These URIs denote connections over TCP/IP, without encryption nor
3812 authentication of the remote host. The URI must specify the host name
3813 and optionally a port number (by default port 44146 is used):
3816 guix://master.guix.example.org:1234
3819 This setup is suitable on local networks, such as clusters, where only
3820 trusted nodes may connect to the build daemon at
3821 @code{master.guix.example.org}.
3823 The @code{--listen} option of @command{guix-daemon} can be used to
3824 instruct it to listen for TCP connections (@pxref{Invoking guix-daemon,
3828 @cindex SSH access to build daemons
3829 These URIs allow you to connect to a remote daemon over
3830 SSH@footnote{This feature requires Guile-SSH (@pxref{Requirements}).}.
3831 A typical URL might look like this:
3834 ssh://charlie@@guix.example.org:22
3837 As for @command{guix copy}, the usual OpenSSH client configuration files
3838 are honored (@pxref{Invoking guix copy}).
3841 Additional URI schemes may be supported in the future.
3843 @c XXX: Remove this note when the protocol incurs fewer round trips
3844 @c and when (guix derivations) no longer relies on file system access.
3846 The ability to connect to remote build daemons is considered
3847 experimental as of @value{VERSION}. Please get in touch with us to
3848 share any problems or suggestions you may have (@pxref{Contributing}).
3852 @deffn {Scheme Procedure} open-connection [@var{uri}] [#:reserve-space? #t]
3853 Connect to the daemon over the Unix-domain socket at @var{uri} (a string). When
3854 @var{reserve-space?} is true, instruct it to reserve a little bit of
3855 extra space on the file system so that the garbage collector can still
3856 operate should the disk become full. Return a server object.
3858 @var{file} defaults to @var{%default-socket-path}, which is the normal
3859 location given the options that were passed to @command{configure}.
3862 @deffn {Scheme Procedure} close-connection @var{server}
3863 Close the connection to @var{server}.
3866 @defvr {Scheme Variable} current-build-output-port
3867 This variable is bound to a SRFI-39 parameter, which refers to the port
3868 where build and error logs sent by the daemon should be written.
3871 Procedures that make RPCs all take a server object as their first
3874 @deffn {Scheme Procedure} valid-path? @var{server} @var{path}
3875 @cindex invalid store items
3876 Return @code{#t} when @var{path} designates a valid store item and
3877 @code{#f} otherwise (an invalid item may exist on disk but still be
3878 invalid, for instance because it is the result of an aborted or failed
3881 A @code{&nix-protocol-error} condition is raised if @var{path} is not
3882 prefixed by the store directory (@file{/gnu/store}).
3885 @deffn {Scheme Procedure} add-text-to-store @var{server} @var{name} @var{text} [@var{references}]
3886 Add @var{text} under file @var{name} in the store, and return its store
3887 path. @var{references} is the list of store paths referred to by the
3888 resulting store path.
3891 @deffn {Scheme Procedure} build-derivations @var{server} @var{derivations}
3892 Build @var{derivations} (a list of @code{<derivation>} objects or
3893 derivation paths), and return when the worker is done building them.
3894 Return @code{#t} on success.
3897 Note that the @code{(guix monads)} module provides a monad as well as
3898 monadic versions of the above procedures, with the goal of making it
3899 more convenient to work with code that accesses the store (@pxref{The
3903 @i{This section is currently incomplete.}
3906 @section Derivations
3909 Low-level build actions and the environment in which they are performed
3910 are represented by @dfn{derivations}. A derivation contains the
3911 following pieces of information:
3915 The outputs of the derivation---derivations produce at least one file or
3916 directory in the store, but may produce more.
3919 The inputs of the derivations, which may be other derivations or plain
3920 files in the store (patches, build scripts, etc.)
3923 The system type targeted by the derivation---e.g., @code{x86_64-linux}.
3926 The file name of a build script in the store, along with the arguments
3930 A list of environment variables to be defined.
3934 @cindex derivation path
3935 Derivations allow clients of the daemon to communicate build actions to
3936 the store. They exist in two forms: as an in-memory representation,
3937 both on the client- and daemon-side, and as files in the store whose
3938 name end in @code{.drv}---these files are referred to as @dfn{derivation
3939 paths}. Derivations paths can be passed to the @code{build-derivations}
3940 procedure to perform the build actions they prescribe (@pxref{The
3943 The @code{(guix derivations)} module provides a representation of
3944 derivations as Scheme objects, along with procedures to create and
3945 otherwise manipulate derivations. The lowest-level primitive to create
3946 a derivation is the @code{derivation} procedure:
3948 @deffn {Scheme Procedure} derivation @var{store} @var{name} @var{builder} @
3949 @var{args} [#:outputs '("out")] [#:hash #f] [#:hash-algo #f] @
3950 [#:recursive? #f] [#:inputs '()] [#:env-vars '()] @
3951 [#:system (%current-system)] [#:references-graphs #f] @
3952 [#:allowed-references #f] [#:disallowed-references #f] @
3953 [#:leaked-env-vars #f] [#:local-build? #f] @
3954 [#:substitutable? #t]
3955 Build a derivation with the given arguments, and return the resulting
3956 @code{<derivation>} object.
3958 When @var{hash} and @var{hash-algo} are given, a
3959 @dfn{fixed-output derivation} is created---i.e., one whose result is
3960 known in advance, such as a file download. If, in addition,
3961 @var{recursive?} is true, then that fixed output may be an executable
3962 file or a directory and @var{hash} must be the hash of an archive
3963 containing this output.
3965 When @var{references-graphs} is true, it must be a list of file
3966 name/store path pairs. In that case, the reference graph of each store
3967 path is exported in the build environment in the corresponding file, in
3968 a simple text format.
3970 When @var{allowed-references} is true, it must be a list of store items
3971 or outputs that the derivation's output may refer to. Likewise,
3972 @var{disallowed-references}, if true, must be a list of things the
3973 outputs may @emph{not} refer to.
3975 When @var{leaked-env-vars} is true, it must be a list of strings
3976 denoting environment variables that are allowed to ``leak'' from the
3977 daemon's environment to the build environment. This is only applicable
3978 to fixed-output derivations---i.e., when @var{hash} is true. The main
3979 use is to allow variables such as @code{http_proxy} to be passed to
3980 derivations that download files.
3982 When @var{local-build?} is true, declare that the derivation is not a
3983 good candidate for offloading and should rather be built locally
3984 (@pxref{Daemon Offload Setup}). This is the case for small derivations
3985 where the costs of data transfers would outweigh the benefits.
3987 When @var{substitutable?} is false, declare that substitutes of the
3988 derivation's output should not be used (@pxref{Substitutes}). This is
3989 useful, for instance, when building packages that capture details of the
3990 host CPU instruction set.
3994 Here's an example with a shell script as its builder, assuming
3995 @var{store} is an open connection to the daemon, and @var{bash} points
3996 to a Bash executable in the store:
3999 (use-modules (guix utils)
4003 (let ((builder ; add the Bash script to the store
4004 (add-text-to-store store "my-builder.sh"
4005 "echo hello world > $out\n" '())))
4006 (derivation store "foo"
4007 bash `("-e" ,builder)
4008 #:inputs `((,bash) (,builder))
4009 #:env-vars '(("HOME" . "/homeless"))))
4010 @result{} #<derivation /gnu/store/@dots{}-foo.drv => /gnu/store/@dots{}-foo>
4013 As can be guessed, this primitive is cumbersome to use directly. A
4014 better approach is to write build scripts in Scheme, of course! The
4015 best course of action for that is to write the build code as a
4016 ``G-expression'', and to pass it to @code{gexp->derivation}. For more
4017 information, @pxref{G-Expressions}.
4019 Once upon a time, @code{gexp->derivation} did not exist and constructing
4020 derivations with build code written in Scheme was achieved with
4021 @code{build-expression->derivation}, documented below. This procedure
4022 is now deprecated in favor of the much nicer @code{gexp->derivation}.
4024 @deffn {Scheme Procedure} build-expression->derivation @var{store} @
4025 @var{name} @var{exp} @
4026 [#:system (%current-system)] [#:inputs '()] @
4027 [#:outputs '("out")] [#:hash #f] [#:hash-algo #f] @
4028 [#:recursive? #f] [#:env-vars '()] [#:modules '()] @
4029 [#:references-graphs #f] [#:allowed-references #f] @
4030 [#:disallowed-references #f] @
4031 [#:local-build? #f] [#:substitutable? #t] [#:guile-for-build #f]
4032 Return a derivation that executes Scheme expression @var{exp} as a
4033 builder for derivation @var{name}. @var{inputs} must be a list of
4034 @code{(name drv-path sub-drv)} tuples; when @var{sub-drv} is omitted,
4035 @code{"out"} is assumed. @var{modules} is a list of names of Guile
4036 modules from the current search path to be copied in the store,
4037 compiled, and made available in the load path during the execution of
4038 @var{exp}---e.g., @code{((guix build utils) (guix build
4039 gnu-build-system))}.
4041 @var{exp} is evaluated in an environment where @code{%outputs} is bound
4042 to a list of output/path pairs, and where @code{%build-inputs} is bound
4043 to a list of string/output-path pairs made from @var{inputs}.
4044 Optionally, @var{env-vars} is a list of string pairs specifying the name
4045 and value of environment variables visible to the builder. The builder
4046 terminates by passing the result of @var{exp} to @code{exit}; thus, when
4047 @var{exp} returns @code{#f}, the build is considered to have failed.
4049 @var{exp} is built using @var{guile-for-build} (a derivation). When
4050 @var{guile-for-build} is omitted or is @code{#f}, the value of the
4051 @code{%guile-for-build} fluid is used instead.
4053 See the @code{derivation} procedure for the meaning of
4054 @var{references-graphs}, @var{allowed-references},
4055 @var{disallowed-references}, @var{local-build?}, and
4056 @var{substitutable?}.
4060 Here's an example of a single-output derivation that creates a directory
4061 containing one file:
4064 (let ((builder '(let ((out (assoc-ref %outputs "out")))
4065 (mkdir out) ; create /gnu/store/@dots{}-goo
4066 (call-with-output-file (string-append out "/test")
4068 (display '(hello guix) p))))))
4069 (build-expression->derivation store "goo" builder))
4071 @result{} #<derivation /gnu/store/@dots{}-goo.drv => @dots{}>
4075 @node The Store Monad
4076 @section The Store Monad
4080 The procedures that operate on the store described in the previous
4081 sections all take an open connection to the build daemon as their first
4082 argument. Although the underlying model is functional, they either have
4083 side effects or depend on the current state of the store.
4085 The former is inconvenient: the connection to the build daemon has to be
4086 carried around in all those functions, making it impossible to compose
4087 functions that do not take that parameter with functions that do. The
4088 latter can be problematic: since store operations have side effects
4089 and/or depend on external state, they have to be properly sequenced.
4091 @cindex monadic values
4092 @cindex monadic functions
4093 This is where the @code{(guix monads)} module comes in. This module
4094 provides a framework for working with @dfn{monads}, and a particularly
4095 useful monad for our uses, the @dfn{store monad}. Monads are a
4096 construct that allows two things: associating ``context'' with values
4097 (in our case, the context is the store), and building sequences of
4098 computations (here computations include accesses to the store). Values
4099 in a monad---values that carry this additional context---are called
4100 @dfn{monadic values}; procedures that return such values are called
4101 @dfn{monadic procedures}.
4103 Consider this ``normal'' procedure:
4106 (define (sh-symlink store)
4107 ;; Return a derivation that symlinks the 'bash' executable.
4108 (let* ((drv (package-derivation store bash))
4109 (out (derivation->output-path drv))
4110 (sh (string-append out "/bin/bash")))
4111 (build-expression->derivation store "sh"
4112 `(symlink ,sh %output))))
4115 Using @code{(guix monads)} and @code{(guix gexp)}, it may be rewritten
4116 as a monadic function:
4119 (define (sh-symlink)
4120 ;; Same, but return a monadic value.
4121 (mlet %store-monad ((drv (package->derivation bash)))
4122 (gexp->derivation "sh"
4123 #~(symlink (string-append #$drv "/bin/bash")
4127 There are several things to note in the second version: the @code{store}
4128 parameter is now implicit and is ``threaded'' in the calls to the
4129 @code{package->derivation} and @code{gexp->derivation} monadic
4130 procedures, and the monadic value returned by @code{package->derivation}
4131 is @dfn{bound} using @code{mlet} instead of plain @code{let}.
4133 As it turns out, the call to @code{package->derivation} can even be
4134 omitted since it will take place implicitly, as we will see later
4135 (@pxref{G-Expressions}):
4138 (define (sh-symlink)
4139 (gexp->derivation "sh"
4140 #~(symlink (string-append #$bash "/bin/bash")
4145 @c <https://syntaxexclamation.wordpress.com/2014/06/26/escaping-continuations/>
4146 @c for the funny quote.
4147 Calling the monadic @code{sh-symlink} has no effect. As someone once
4148 said, ``you exit a monad like you exit a building on fire: by running''.
4149 So, to exit the monad and get the desired effect, one must use
4150 @code{run-with-store}:
4153 (run-with-store (open-connection) (sh-symlink))
4154 @result{} /gnu/store/...-sh-symlink
4157 Note that the @code{(guix monad-repl)} module extends the Guile REPL with
4158 new ``meta-commands'' to make it easier to deal with monadic procedures:
4159 @code{run-in-store}, and @code{enter-store-monad}. The former is used
4160 to ``run'' a single monadic value through the store:
4163 scheme@@(guile-user)> ,run-in-store (package->derivation hello)
4164 $1 = #<derivation /gnu/store/@dots{}-hello-2.9.drv => @dots{}>
4167 The latter enters a recursive REPL, where all the return values are
4168 automatically run through the store:
4171 scheme@@(guile-user)> ,enter-store-monad
4172 store-monad@@(guile-user) [1]> (package->derivation hello)
4173 $2 = #<derivation /gnu/store/@dots{}-hello-2.9.drv => @dots{}>
4174 store-monad@@(guile-user) [1]> (text-file "foo" "Hello!")
4175 $3 = "/gnu/store/@dots{}-foo"
4176 store-monad@@(guile-user) [1]> ,q
4177 scheme@@(guile-user)>
4181 Note that non-monadic values cannot be returned in the
4182 @code{store-monad} REPL.
4184 The main syntactic forms to deal with monads in general are provided by
4185 the @code{(guix monads)} module and are described below.
4187 @deffn {Scheme Syntax} with-monad @var{monad} @var{body} ...
4188 Evaluate any @code{>>=} or @code{return} forms in @var{body} as being
4192 @deffn {Scheme Syntax} return @var{val}
4193 Return a monadic value that encapsulates @var{val}.
4196 @deffn {Scheme Syntax} >>= @var{mval} @var{mproc} ...
4197 @dfn{Bind} monadic value @var{mval}, passing its ``contents'' to monadic
4198 procedures @var{mproc}@dots{}@footnote{This operation is commonly
4199 referred to as ``bind'', but that name denotes an unrelated procedure in
4200 Guile. Thus we use this somewhat cryptic symbol inherited from the
4201 Haskell language.}. There can be one @var{mproc} or several of them, as
4206 (with-monad %state-monad
4208 (lambda (x) (return (+ 1 x)))
4209 (lambda (x) (return (* 2 x)))))
4213 @result{} some-state
4217 @deffn {Scheme Syntax} mlet @var{monad} ((@var{var} @var{mval}) ...) @
4219 @deffnx {Scheme Syntax} mlet* @var{monad} ((@var{var} @var{mval}) ...) @
4221 Bind the variables @var{var} to the monadic values @var{mval} in
4222 @var{body}, which is a sequence of expressions. As with the bind
4223 operator, this can be thought of as ``unpacking'' the raw, non-monadic
4224 value ``contained'' in @var{mval} and making @var{var} refer to that
4225 raw, non-monadic value within the scope of the @var{body}. The form
4226 (@var{var} -> @var{val}) binds @var{var} to the ``normal'' value
4227 @var{val}, as per @code{let}. The binding operations occur in sequence
4228 from left to right. The last expression of @var{body} must be a monadic
4229 expression, and its result will become the result of the @code{mlet} or
4230 @code{mlet*} when run in the @var{monad}.
4232 @code{mlet*} is to @code{mlet} what @code{let*} is to @code{let}
4233 (@pxref{Local Bindings,,, guile, GNU Guile Reference Manual}).
4236 @deffn {Scheme System} mbegin @var{monad} @var{mexp} ...
4237 Bind @var{mexp} and the following monadic expressions in sequence,
4238 returning the result of the last expression. Every expression in the
4239 sequence must be a monadic expression.
4241 This is akin to @code{mlet}, except that the return values of the
4242 monadic expressions are ignored. In that sense, it is analogous to
4243 @code{begin}, but applied to monadic expressions.
4246 @deffn {Scheme System} mwhen @var{condition} @var{mexp0} @var{mexp*} ...
4247 When @var{condition} is true, evaluate the sequence of monadic
4248 expressions @var{mexp0}..@var{mexp*} as in an @code{mbegin}. When
4249 @var{condition} is false, return @code{*unspecified*} in the current
4250 monad. Every expression in the sequence must be a monadic expression.
4253 @deffn {Scheme System} munless @var{condition} @var{mexp0} @var{mexp*} ...
4254 When @var{condition} is false, evaluate the sequence of monadic
4255 expressions @var{mexp0}..@var{mexp*} as in an @code{mbegin}. When
4256 @var{condition} is true, return @code{*unspecified*} in the current
4257 monad. Every expression in the sequence must be a monadic expression.
4261 The @code{(guix monads)} module provides the @dfn{state monad}, which
4262 allows an additional value---the state---to be @emph{threaded} through
4263 monadic procedure calls.
4265 @defvr {Scheme Variable} %state-monad
4266 The state monad. Procedures in the state monad can access and change
4267 the state that is threaded.
4269 Consider the example below. The @code{square} procedure returns a value
4270 in the state monad. It returns the square of its argument, but also
4271 increments the current state value:
4275 (mlet %state-monad ((count (current-state)))
4276 (mbegin %state-monad
4277 (set-current-state (+ 1 count))
4280 (run-with-state (sequence %state-monad (map square (iota 3))) 0)
4285 When ``run'' through @var{%state-monad}, we obtain that additional state
4286 value, which is the number of @code{square} calls.
4289 @deffn {Monadic Procedure} current-state
4290 Return the current state as a monadic value.
4293 @deffn {Monadic Procedure} set-current-state @var{value}
4294 Set the current state to @var{value} and return the previous state as a
4298 @deffn {Monadic Procedure} state-push @var{value}
4299 Push @var{value} to the current state, which is assumed to be a list,
4300 and return the previous state as a monadic value.
4303 @deffn {Monadic Procedure} state-pop
4304 Pop a value from the current state and return it as a monadic value.
4305 The state is assumed to be a list.
4308 @deffn {Scheme Procedure} run-with-state @var{mval} [@var{state}]
4309 Run monadic value @var{mval} starting with @var{state} as the initial
4310 state. Return two values: the resulting value, and the resulting state.
4313 The main interface to the store monad, provided by the @code{(guix
4314 store)} module, is as follows.
4316 @defvr {Scheme Variable} %store-monad
4317 The store monad---an alias for @var{%state-monad}.
4319 Values in the store monad encapsulate accesses to the store. When its
4320 effect is needed, a value of the store monad must be ``evaluated'' by
4321 passing it to the @code{run-with-store} procedure (see below.)
4324 @deffn {Scheme Procedure} run-with-store @var{store} @var{mval} [#:guile-for-build] [#:system (%current-system)]
4325 Run @var{mval}, a monadic value in the store monad, in @var{store}, an
4326 open store connection.
4329 @deffn {Monadic Procedure} text-file @var{name} @var{text} [@var{references}]
4330 Return as a monadic value the absolute file name in the store of the file
4331 containing @var{text}, a string. @var{references} is a list of store items that the
4332 resulting text file refers to; it defaults to the empty list.
4335 @deffn {Monadic Procedure} interned-file @var{file} [@var{name}] @
4336 [#:recursive? #t] [#:select? (const #t)]
4337 Return the name of @var{file} once interned in the store. Use
4338 @var{name} as its store name, or the basename of @var{file} if
4339 @var{name} is omitted.
4341 When @var{recursive?} is true, the contents of @var{file} are added
4342 recursively; if @var{file} designates a flat file and @var{recursive?}
4343 is true, its contents are added, and its permission bits are kept.
4345 When @var{recursive?} is true, call @code{(@var{select?} @var{file}
4346 @var{stat})} for each directory entry, where @var{file} is the entry's
4347 absolute file name and @var{stat} is the result of @code{lstat}; exclude
4348 entries for which @var{select?} does not return true.
4350 The example below adds a file to the store, under two different names:
4353 (run-with-store (open-connection)
4354 (mlet %store-monad ((a (interned-file "README"))
4355 (b (interned-file "README" "LEGU-MIN")))
4356 (return (list a b))))
4358 @result{} ("/gnu/store/rwm@dots{}-README" "/gnu/store/44i@dots{}-LEGU-MIN")
4363 The @code{(guix packages)} module exports the following package-related
4366 @deffn {Monadic Procedure} package-file @var{package} [@var{file}] @
4367 [#:system (%current-system)] [#:target #f] @
4370 value in the absolute file name of @var{file} within the @var{output}
4371 directory of @var{package}. When @var{file} is omitted, return the name
4372 of the @var{output} directory of @var{package}. When @var{target} is
4373 true, use it as a cross-compilation target triplet.
4376 @deffn {Monadic Procedure} package->derivation @var{package} [@var{system}]
4377 @deffnx {Monadic Procedure} package->cross-derivation @var{package} @
4378 @var{target} [@var{system}]
4379 Monadic version of @code{package-derivation} and
4380 @code{package-cross-derivation} (@pxref{Defining Packages}).
4385 @section G-Expressions
4387 @cindex G-expression
4388 @cindex build code quoting
4389 So we have ``derivations'', which represent a sequence of build actions
4390 to be performed to produce an item in the store (@pxref{Derivations}).
4391 These build actions are performed when asking the daemon to actually
4392 build the derivations; they are run by the daemon in a container
4393 (@pxref{Invoking guix-daemon}).
4395 @cindex strata of code
4396 It should come as no surprise that we like to write these build actions
4397 in Scheme. When we do that, we end up with two @dfn{strata} of Scheme
4398 code@footnote{The term @dfn{stratum} in this context was coined by
4399 Manuel Serrano et al.@: in the context of their work on Hop. Oleg
4400 Kiselyov, who has written insightful
4401 @url{http://okmij.org/ftp/meta-programming/#meta-scheme, essays and code
4402 on this topic}, refers to this kind of code generation as
4403 @dfn{staging}.}: the ``host code''---code that defines packages, talks
4404 to the daemon, etc.---and the ``build code''---code that actually
4405 performs build actions, such as making directories, invoking
4406 @command{make}, etc.
4408 To describe a derivation and its build actions, one typically needs to
4409 embed build code inside host code. It boils down to manipulating build
4410 code as data, and the homoiconicity of Scheme---code has a direct
4411 representation as data---comes in handy for that. But we need more than
4412 the normal @code{quasiquote} mechanism in Scheme to construct build
4415 The @code{(guix gexp)} module implements @dfn{G-expressions}, a form of
4416 S-expressions adapted to build expressions. G-expressions, or
4417 @dfn{gexps}, consist essentially of three syntactic forms: @code{gexp},
4418 @code{ungexp}, and @code{ungexp-splicing} (or simply: @code{#~},
4419 @code{#$}, and @code{#$@@}), which are comparable to
4420 @code{quasiquote}, @code{unquote}, and @code{unquote-splicing},
4421 respectively (@pxref{Expression Syntax, @code{quasiquote},, guile,
4422 GNU Guile Reference Manual}). However, there are major differences:
4426 Gexps are meant to be written to a file and run or manipulated by other
4430 When a high-level object such as a package or derivation is unquoted
4431 inside a gexp, the result is as if its output file name had been
4435 Gexps carry information about the packages or derivations they refer to,
4436 and these dependencies are automatically added as inputs to the build
4437 processes that use them.
4440 @cindex lowering, of high-level objects in gexps
4441 This mechanism is not limited to package and derivation
4442 objects: @dfn{compilers} able to ``lower'' other high-level objects to
4443 derivations or files in the store can be defined,
4444 such that these objects can also be inserted
4445 into gexps. For example, a useful type of high-level objects that can be
4446 inserted in a gexp is ``file-like objects'', which make it easy to
4447 add files to the store and to refer to them in
4448 derivations and such (see @code{local-file} and @code{plain-file}
4451 To illustrate the idea, here is an example of a gexp:
4458 (symlink (string-append #$coreutils "/bin/ls")
4462 This gexp can be passed to @code{gexp->derivation}; we obtain a
4463 derivation that builds a directory containing exactly one symlink to
4464 @file{/gnu/store/@dots{}-coreutils-8.22/bin/ls}:
4467 (gexp->derivation "the-thing" build-exp)
4470 As one would expect, the @code{"/gnu/store/@dots{}-coreutils-8.22"} string is
4471 substituted to the reference to the @var{coreutils} package in the
4472 actual build code, and @var{coreutils} is automatically made an input to
4473 the derivation. Likewise, @code{#$output} (equivalent to @code{(ungexp
4474 output)}) is replaced by a string containing the directory name of the
4475 output of the derivation.
4477 @cindex cross compilation
4478 In a cross-compilation context, it is useful to distinguish between
4479 references to the @emph{native} build of a package---that can run on the
4480 host---versus references to cross builds of a package. To that end, the
4481 @code{#+} plays the same role as @code{#$}, but is a reference to a
4482 native package build:
4485 (gexp->derivation "vi"
4488 (system* (string-append #+coreutils "/bin/ln")
4490 (string-append #$emacs "/bin/emacs")
4491 (string-append #$output "/bin/vi")))
4492 #:target "mips64el-linux-gnu")
4496 In the example above, the native build of @var{coreutils} is used, so
4497 that @command{ln} can actually run on the host; but then the
4498 cross-compiled build of @var{emacs} is referenced.
4500 @cindex imported modules, for gexps
4501 @findex with-imported-modules
4502 Another gexp feature is @dfn{imported modules}: sometimes you want to be
4503 able to use certain Guile modules from the ``host environment'' in the
4504 gexp, so those modules should be imported in the ``build environment''.
4505 The @code{with-imported-modules} form allows you to express that:
4508 (let ((build (with-imported-modules '((guix build utils))
4510 (use-modules (guix build utils))
4511 (mkdir-p (string-append #$output "/bin"))))))
4512 (gexp->derivation "empty-dir"
4515 (display "success!\n")
4520 In this example, the @code{(guix build utils)} module is automatically
4521 pulled into the isolated build environment of our gexp, such that
4522 @code{(use-modules (guix build utils))} works as expected.
4524 @cindex module closure
4525 @findex source-module-closure
4526 Usually you want the @emph{closure} of the module to be imported---i.e.,
4527 the module itself and all the modules it depends on---rather than just
4528 the module; failing to do that, attempts to use the module will fail
4529 because of missing dependent modules. The @code{source-module-closure}
4530 procedure computes the closure of a module by looking at its source file
4531 headers, which comes in handy in this case:
4534 (use-modules (guix modules)) ;for 'source-module-closure'
4536 (with-imported-modules (source-module-closure
4537 '((guix build utils)
4539 (gexp->derivation "something-with-vms"
4541 (use-modules (guix build utils)
4546 The syntactic form to construct gexps is summarized below.
4548 @deffn {Scheme Syntax} #~@var{exp}
4549 @deffnx {Scheme Syntax} (gexp @var{exp})
4550 Return a G-expression containing @var{exp}. @var{exp} may contain one
4551 or more of the following forms:
4555 @itemx (ungexp @var{obj})
4556 Introduce a reference to @var{obj}. @var{obj} may have one of the
4557 supported types, for example a package or a
4558 derivation, in which case the @code{ungexp} form is replaced by its
4559 output file name---e.g., @code{"/gnu/store/@dots{}-coreutils-8.22}.
4561 If @var{obj} is a list, it is traversed and references to supported
4562 objects are substituted similarly.
4564 If @var{obj} is another gexp, its contents are inserted and its
4565 dependencies are added to those of the containing gexp.
4567 If @var{obj} is another kind of object, it is inserted as is.
4569 @item #$@var{obj}:@var{output}
4570 @itemx (ungexp @var{obj} @var{output})
4571 This is like the form above, but referring explicitly to the
4572 @var{output} of @var{obj}---this is useful when @var{obj} produces
4573 multiple outputs (@pxref{Packages with Multiple Outputs}).
4576 @itemx #+@var{obj}:output
4577 @itemx (ungexp-native @var{obj})
4578 @itemx (ungexp-native @var{obj} @var{output})
4579 Same as @code{ungexp}, but produces a reference to the @emph{native}
4580 build of @var{obj} when used in a cross compilation context.
4582 @item #$output[:@var{output}]
4583 @itemx (ungexp output [@var{output}])
4584 Insert a reference to derivation output @var{output}, or to the main
4585 output when @var{output} is omitted.
4587 This only makes sense for gexps passed to @code{gexp->derivation}.
4590 @itemx (ungexp-splicing @var{lst})
4591 Like the above, but splices the contents of @var{lst} inside the
4595 @itemx (ungexp-native-splicing @var{lst})
4596 Like the above, but refers to native builds of the objects listed in
4601 G-expressions created by @code{gexp} or @code{#~} are run-time objects
4602 of the @code{gexp?} type (see below.)
4605 @deffn {Scheme Syntax} with-imported-modules @var{modules} @var{body}@dots{}
4606 Mark the gexps defined in @var{body}@dots{} as requiring @var{modules}
4607 in their execution environment.
4609 Each item in @var{modules} can be the name of a module, such as
4610 @code{(guix build utils)}, or it can be a module name, followed by an
4611 arrow, followed by a file-like object:
4614 `((guix build utils)
4616 ((guix config) => ,(scheme-file "config.scm"
4617 #~(define-module @dots{}))))
4621 In the example above, the first two modules are taken from the search
4622 path, and the last one is created from the given file-like object.
4624 This form has @emph{lexical} scope: it has an effect on the gexps
4625 directly defined in @var{body}@dots{}, but not on those defined, say, in
4626 procedures called from @var{body}@dots{}.
4629 @deffn {Scheme Procedure} gexp? @var{obj}
4630 Return @code{#t} if @var{obj} is a G-expression.
4633 G-expressions are meant to be written to disk, either as code building
4634 some derivation, or as plain files in the store. The monadic procedures
4635 below allow you to do that (@pxref{The Store Monad}, for more
4636 information about monads.)
4638 @deffn {Monadic Procedure} gexp->derivation @var{name} @var{exp} @
4639 [#:system (%current-system)] [#:target #f] [#:graft? #t] @
4640 [#:hash #f] [#:hash-algo #f] @
4641 [#:recursive? #f] [#:env-vars '()] [#:modules '()] @
4642 [#:module-path @var{%load-path}] @
4643 [#:references-graphs #f] [#:allowed-references #f] @
4644 [#:disallowed-references #f] @
4645 [#:leaked-env-vars #f] @
4646 [#:script-name (string-append @var{name} "-builder")] @
4647 [#:local-build? #f] [#:substitutable? #t] [#:guile-for-build #f]
4648 Return a derivation @var{name} that runs @var{exp} (a gexp) with
4649 @var{guile-for-build} (a derivation) on @var{system}; @var{exp} is
4650 stored in a file called @var{script-name}. When @var{target} is true,
4651 it is used as the cross-compilation target triplet for packages referred
4654 @var{modules} is deprecated in favor of @code{with-imported-modules}.
4656 make @var{modules} available in the evaluation context of @var{exp};
4657 @var{modules} is a list of names of Guile modules searched in
4658 @var{module-path} to be copied in the store, compiled, and made available in
4659 the load path during the execution of @var{exp}---e.g., @code{((guix
4660 build utils) (guix build gnu-build-system))}.
4662 @var{graft?} determines whether packages referred to by @var{exp} should be grafted when
4665 When @var{references-graphs} is true, it must be a list of tuples of one of the
4669 (@var{file-name} @var{package})
4670 (@var{file-name} @var{package} @var{output})
4671 (@var{file-name} @var{derivation})
4672 (@var{file-name} @var{derivation} @var{output})
4673 (@var{file-name} @var{store-item})
4676 The right-hand-side of each element of @var{references-graphs} is automatically made
4677 an input of the build process of @var{exp}. In the build environment, each
4678 @var{file-name} contains the reference graph of the corresponding item, in a simple
4681 @var{allowed-references} must be either @code{#f} or a list of output names and packages.
4682 In the latter case, the list denotes store items that the result is allowed to
4683 refer to. Any reference to another store item will lead to a build error.
4684 Similarly for @var{disallowed-references}, which can list items that must not be
4685 referenced by the outputs.
4687 The other arguments are as for @code{derivation} (@pxref{Derivations}).
4690 @cindex file-like objects
4691 The @code{local-file}, @code{plain-file}, @code{computed-file},
4692 @code{program-file}, and @code{scheme-file} procedures below return
4693 @dfn{file-like objects}. That is, when unquoted in a G-expression,
4694 these objects lead to a file in the store. Consider this G-expression:
4697 #~(system* #$(file-append glibc "/sbin/nscd") "-f"
4698 #$(local-file "/tmp/my-nscd.conf"))
4701 The effect here is to ``intern'' @file{/tmp/my-nscd.conf} by copying it
4702 to the store. Once expanded, for instance @i{via}
4703 @code{gexp->derivation}, the G-expression refers to that copy under
4704 @file{/gnu/store}; thus, modifying or removing the file in @file{/tmp}
4705 does not have any effect on what the G-expression does.
4706 @code{plain-file} can be used similarly; it differs in that the file
4707 content is directly passed as a string.
4709 @deffn {Scheme Procedure} local-file @var{file} [@var{name}] @
4710 [#:recursive? #f] [#:select? (const #t)]
4711 Return an object representing local file @var{file} to add to the store; this
4712 object can be used in a gexp. If @var{file} is a relative file name, it is looked
4713 up relative to the source file where this form appears. @var{file} will be added to
4714 the store under @var{name}--by default the base name of @var{file}.
4716 When @var{recursive?} is true, the contents of @var{file} are added recursively; if @var{file}
4717 designates a flat file and @var{recursive?} is true, its contents are added, and its
4718 permission bits are kept.
4720 When @var{recursive?} is true, call @code{(@var{select?} @var{file}
4721 @var{stat})} for each directory entry, where @var{file} is the entry's
4722 absolute file name and @var{stat} is the result of @code{lstat}; exclude
4723 entries for which @var{select?} does not return true.
4725 This is the declarative counterpart of the @code{interned-file} monadic
4726 procedure (@pxref{The Store Monad, @code{interned-file}}).
4729 @deffn {Scheme Procedure} plain-file @var{name} @var{content}
4730 Return an object representing a text file called @var{name} with the given
4731 @var{content} (a string) to be added to the store.
4733 This is the declarative counterpart of @code{text-file}.
4736 @deffn {Scheme Procedure} computed-file @var{name} @var{gexp} @
4737 [#:options '(#:local-build? #t)]
4738 Return an object representing the store item @var{name}, a file or
4739 directory computed by @var{gexp}. @var{options}
4740 is a list of additional arguments to pass to @code{gexp->derivation}.
4742 This is the declarative counterpart of @code{gexp->derivation}.
4745 @deffn {Monadic Procedure} gexp->script @var{name} @var{exp}
4746 Return an executable script @var{name} that runs @var{exp} using
4747 @var{guile}, with @var{exp}'s imported modules in its search path.
4749 The example below builds a script that simply invokes the @command{ls}
4753 (use-modules (guix gexp) (gnu packages base))
4755 (gexp->script "list-files"
4756 #~(execl #$(file-append coreutils "/bin/ls")
4760 When ``running'' it through the store (@pxref{The Store Monad,
4761 @code{run-with-store}}), we obtain a derivation that produces an
4762 executable file @file{/gnu/store/@dots{}-list-files} along these lines:
4765 #!/gnu/store/@dots{}-guile-2.0.11/bin/guile -ds
4767 (execl "/gnu/store/@dots{}-coreutils-8.22"/bin/ls" "ls")
4771 @deffn {Scheme Procedure} program-file @var{name} @var{exp} @
4773 Return an object representing the executable store item @var{name} that
4774 runs @var{gexp}. @var{guile} is the Guile package used to execute that
4777 This is the declarative counterpart of @code{gexp->script}.
4780 @deffn {Monadic Procedure} gexp->file @var{name} @var{exp} @
4781 [#:set-load-path? #t]
4782 Return a derivation that builds a file @var{name} containing @var{exp}.
4783 When @var{set-load-path?} is true, emit code in the resulting file to
4784 set @code{%load-path} and @code{%load-compiled-path} to honor
4785 @var{exp}'s imported modules.
4787 The resulting file holds references to all the dependencies of @var{exp}
4788 or a subset thereof.
4791 @deffn {Scheme Procedure} scheme-file @var{name} @var{exp}
4792 Return an object representing the Scheme file @var{name} that contains
4795 This is the declarative counterpart of @code{gexp->file}.
4798 @deffn {Monadic Procedure} text-file* @var{name} @var{text} @dots{}
4799 Return as a monadic value a derivation that builds a text file
4800 containing all of @var{text}. @var{text} may list, in addition to
4801 strings, objects of any type that can be used in a gexp: packages,
4802 derivations, local file objects, etc. The resulting store file holds
4803 references to all these.
4805 This variant should be preferred over @code{text-file} anytime the file
4806 to create will reference items from the store. This is typically the
4807 case when building a configuration file that embeds store file names,
4811 (define (profile.sh)
4812 ;; Return the name of a shell script in the store that
4813 ;; initializes the 'PATH' environment variable.
4814 (text-file* "profile.sh"
4815 "export PATH=" coreutils "/bin:"
4816 grep "/bin:" sed "/bin\n"))
4819 In this example, the resulting @file{/gnu/store/@dots{}-profile.sh} file
4820 will reference @var{coreutils}, @var{grep}, and @var{sed}, thereby
4821 preventing them from being garbage-collected during its lifetime.
4824 @deffn {Scheme Procedure} mixed-text-file @var{name} @var{text} @dots{}
4825 Return an object representing store file @var{name} containing
4826 @var{text}. @var{text} is a sequence of strings and file-like objects,
4830 (mixed-text-file "profile"
4831 "export PATH=" coreutils "/bin:" grep "/bin")
4834 This is the declarative counterpart of @code{text-file*}.
4837 @deffn {Scheme Procedure} file-append @var{obj} @var{suffix} @dots{}
4838 Return a file-like object that expands to the concatenation of @var{obj}
4839 and @var{suffix}, where @var{obj} is a lowerable object and each
4840 @var{suffix} is a string.
4842 As an example, consider this gexp:
4845 (gexp->script "run-uname"
4846 #~(system* #$(file-append coreutils
4850 The same effect could be achieved with:
4853 (gexp->script "run-uname"
4854 #~(system* (string-append #$coreutils
4858 There is one difference though: in the @code{file-append} case, the
4859 resulting script contains the absolute file name as a string, whereas in
4860 the second case, the resulting script contains a @code{(string-append
4861 @dots{})} expression to construct the file name @emph{at run time}.
4865 Of course, in addition to gexps embedded in ``host'' code, there are
4866 also modules containing build tools. To make it clear that they are
4867 meant to be used in the build stratum, these modules are kept in the
4868 @code{(guix build @dots{})} name space.
4870 @cindex lowering, of high-level objects in gexps
4871 Internally, high-level objects are @dfn{lowered}, using their compiler,
4872 to either derivations or store items. For instance, lowering a package
4873 yields a derivation, and lowering a @code{plain-file} yields a store
4874 item. This is achieved using the @code{lower-object} monadic procedure.
4876 @deffn {Monadic Procedure} lower-object @var{obj} [@var{system}] @
4878 Return as a value in @var{%store-monad} the derivation or store item
4879 corresponding to @var{obj} for @var{system}, cross-compiling for
4880 @var{target} if @var{target} is true. @var{obj} must be an object that
4881 has an associated gexp compiler, such as a @code{<package>}.
4885 @c *********************************************************************
4889 This section describes Guix command-line utilities. Some of them are
4890 primarily targeted at developers and users who write new package
4891 definitions, while others are more generally useful. They complement
4892 the Scheme programming interface of Guix in a convenient way.
4895 * Invoking guix build:: Building packages from the command line.
4896 * Invoking guix edit:: Editing package definitions.
4897 * Invoking guix download:: Downloading a file and printing its hash.
4898 * Invoking guix hash:: Computing the cryptographic hash of a file.
4899 * Invoking guix import:: Importing package definitions.
4900 * Invoking guix refresh:: Updating package definitions.
4901 * Invoking guix lint:: Finding errors in package definitions.
4902 * Invoking guix size:: Profiling disk usage.
4903 * Invoking guix graph:: Visualizing the graph of packages.
4904 * Invoking guix environment:: Setting up development environments.
4905 * Invoking guix publish:: Sharing substitutes.
4906 * Invoking guix challenge:: Challenging substitute servers.
4907 * Invoking guix copy:: Copying to and from a remote store.
4908 * Invoking guix container:: Process isolation.
4911 @node Invoking guix build
4912 @section Invoking @command{guix build}
4914 @cindex package building
4915 @cindex @command{guix build}
4916 The @command{guix build} command builds packages or derivations and
4917 their dependencies, and prints the resulting store paths. Note that it
4918 does not modify the user's profile---this is the job of the
4919 @command{guix package} command (@pxref{Invoking guix package}). Thus,
4920 it is mainly useful for distribution developers.
4922 The general syntax is:
4925 guix build @var{options} @var{package-or-derivation}@dots{}
4928 As an example, the following command builds the latest versions of Emacs
4929 and of Guile, displays their build logs, and finally displays the
4930 resulting directories:
4933 guix build emacs guile
4936 Similarly, the following command builds all the available packages:
4939 guix build --quiet --keep-going \
4940 `guix package -A | cut -f1,2 --output-delimiter=@@`
4943 @var{package-or-derivation} may be either the name of a package found in
4944 the software distribution such as @code{coreutils} or
4945 @code{coreutils@@8.20}, or a derivation such as
4946 @file{/gnu/store/@dots{}-coreutils-8.19.drv}. In the former case, a
4947 package with the corresponding name (and optionally version) is searched
4948 for among the GNU distribution modules (@pxref{Package Modules}).
4950 Alternatively, the @code{--expression} option may be used to specify a
4951 Scheme expression that evaluates to a package; this is useful when
4952 disambiguating among several same-named packages or package variants is
4955 There may be zero or more @var{options}. The available options are
4956 described in the subsections below.
4959 * Common Build Options:: Build options for most commands.
4960 * Package Transformation Options:: Creating variants of packages.
4961 * Additional Build Options:: Options specific to 'guix build'.
4962 * Debugging Build Failures:: Real life packaging experience.
4965 @node Common Build Options
4966 @subsection Common Build Options
4968 A number of options that control the build process are common to
4969 @command{guix build} and other commands that can spawn builds, such as
4970 @command{guix package} or @command{guix archive}. These are the
4975 @item --load-path=@var{directory}
4976 @itemx -L @var{directory}
4977 Add @var{directory} to the front of the package module search path
4978 (@pxref{Package Modules}).
4980 This allows users to define their own packages and make them visible to
4981 the command-line tools.
4985 Keep the build tree of failed builds. Thus, if a build fails, its build
4986 tree is kept under @file{/tmp}, in a directory whose name is shown at
4987 the end of the build log. This is useful when debugging build issues.
4988 @xref{Debugging Build Failures}, for tips and tricks on how to debug
4993 Keep going when some of the derivations fail to build; return only once
4994 all the builds have either completed or failed.
4996 The default behavior is to stop as soon as one of the specified
4997 derivations has failed.
5001 Do not build the derivations.
5004 When substituting a pre-built binary fails, fall back to building
5007 @item --substitute-urls=@var{urls}
5008 @anchor{client-substitute-urls}
5009 Consider @var{urls} the whitespace-separated list of substitute source
5010 URLs, overriding the default list of URLs of @command{guix-daemon}
5011 (@pxref{daemon-substitute-urls,, @command{guix-daemon} URLs}).
5013 This means that substitutes may be downloaded from @var{urls}, provided
5014 they are signed by a key authorized by the system administrator
5015 (@pxref{Substitutes}).
5017 When @var{urls} is the empty string, substitutes are effectively
5020 @item --no-substitutes
5021 Do not use substitutes for build products. That is, always build things
5022 locally instead of allowing downloads of pre-built binaries
5023 (@pxref{Substitutes}).
5026 Do not ``graft'' packages. In practice, this means that package updates
5027 available as grafts are not applied. @xref{Security Updates}, for more
5028 information on grafts.
5030 @item --rounds=@var{n}
5031 Build each derivation @var{n} times in a row, and raise an error if
5032 consecutive build results are not bit-for-bit identical.
5034 This is a useful way to detect non-deterministic builds processes.
5035 Non-deterministic build processes are a problem because they make it
5036 practically impossible for users to @emph{verify} whether third-party
5037 binaries are genuine. @xref{Invoking guix challenge}, for more.
5039 Note that, currently, the differing build results are not kept around,
5040 so you will have to manually investigate in case of an error---e.g., by
5041 stashing one of the build results with @code{guix archive --export}
5042 (@pxref{Invoking guix archive}), then rebuilding, and finally comparing
5045 @item --no-build-hook
5046 Do not attempt to offload builds @i{via} the ``build hook'' of the daemon
5047 (@pxref{Daemon Offload Setup}). That is, always build things locally
5048 instead of offloading builds to remote machines.
5050 @item --max-silent-time=@var{seconds}
5051 When the build or substitution process remains silent for more than
5052 @var{seconds}, terminate it and report a build failure.
5054 By default, the daemon's setting is honored (@pxref{Invoking
5055 guix-daemon, @code{--max-silent-time}}).
5057 @item --timeout=@var{seconds}
5058 Likewise, when the build or substitution process lasts for more than
5059 @var{seconds}, terminate it and report a build failure.
5061 By default, the daemon's setting is honored (@pxref{Invoking
5062 guix-daemon, @code{--timeout}}).
5064 @item --verbosity=@var{level}
5065 Use the given verbosity level. @var{level} must be an integer between 0
5066 and 5; higher means more verbose output. Setting a level of 4 or more
5067 may be helpful when debugging setup issues with the build daemon.
5069 @item --cores=@var{n}
5071 Allow the use of up to @var{n} CPU cores for the build. The special
5072 value @code{0} means to use as many CPU cores as available.
5074 @item --max-jobs=@var{n}
5076 Allow at most @var{n} build jobs in parallel. @xref{Invoking
5077 guix-daemon, @code{--max-jobs}}, for details about this option and the
5078 equivalent @command{guix-daemon} option.
5082 Behind the scenes, @command{guix build} is essentially an interface to
5083 the @code{package-derivation} procedure of the @code{(guix packages)}
5084 module, and to the @code{build-derivations} procedure of the @code{(guix
5085 derivations)} module.
5087 In addition to options explicitly passed on the command line,
5088 @command{guix build} and other @command{guix} commands that support
5089 building honor the @code{GUIX_BUILD_OPTIONS} environment variable.
5091 @defvr {Environment Variable} GUIX_BUILD_OPTIONS
5092 Users can define this variable to a list of command line options that
5093 will automatically be used by @command{guix build} and other
5094 @command{guix} commands that can perform builds, as in the example
5098 $ export GUIX_BUILD_OPTIONS="--no-substitutes -c 2 -L /foo/bar"
5101 These options are parsed independently, and the result is appended to
5102 the parsed command-line options.
5106 @node Package Transformation Options
5107 @subsection Package Transformation Options
5109 @cindex package variants
5110 Another set of command-line options supported by @command{guix build}
5111 and also @command{guix package} are @dfn{package transformation
5112 options}. These are options that make it possible to define @dfn{package
5113 variants}---for instance, packages built from different source code.
5114 This is a convenient way to create customized packages on the fly
5115 without having to type in the definitions of package variants
5116 (@pxref{Defining Packages}).
5120 @item --with-source=@var{source}
5121 Use @var{source} as the source of the corresponding package.
5122 @var{source} must be a file name or a URL, as for @command{guix
5123 download} (@pxref{Invoking guix download}).
5125 The ``corresponding package'' is taken to be the one specified on the
5126 command line the name of which matches the base of @var{source}---e.g.,
5127 if @var{source} is @code{/src/guile-2.0.10.tar.gz}, the corresponding
5128 package is @code{guile}. Likewise, the version string is inferred from
5129 @var{source}; in the previous example, it is @code{2.0.10}.
5131 This option allows users to try out versions of packages other than the
5132 one provided by the distribution. The example below downloads
5133 @file{ed-1.7.tar.gz} from a GNU mirror and uses that as the source for
5134 the @code{ed} package:
5137 guix build ed --with-source=mirror://gnu/ed/ed-1.7.tar.gz
5140 As a developer, @code{--with-source} makes it easy to test release
5144 guix build guile --with-source=../guile-2.0.9.219-e1bb7.tar.xz
5147 @dots{} or to build from a checkout in a pristine environment:
5150 $ git clone git://git.sv.gnu.org/guix.git
5151 $ guix build guix --with-source=./guix
5154 @item --with-input=@var{package}=@var{replacement}
5155 Replace dependency on @var{package} by a dependency on
5156 @var{replacement}. @var{package} must be a package name, and
5157 @var{replacement} must be a package specification such as @code{guile}
5158 or @code{guile@@1.8}.
5160 For instance, the following command builds Guix, but replaces its
5161 dependency on the current stable version of Guile with a dependency on
5162 the legacy version of Guile, @code{guile@@2.0}:
5165 guix build --with-input=guile=guile@@2.0 guix
5168 This is a recursive, deep replacement. So in this example, both
5169 @code{guix} and its dependency @code{guile-json} (which also depends on
5170 @code{guile}) get rebuilt against @code{guile@@2.0}.
5172 This is implemented using the @code{package-input-rewriting} Scheme
5173 procedure (@pxref{Defining Packages, @code{package-input-rewriting}}).
5175 @item --with-graft=@var{package}=@var{replacement}
5176 This is similar to @code{--with-input} but with an important difference:
5177 instead of rebuilding the whole dependency chain, @var{replacement} is
5178 built and then @dfn{grafted} onto the binaries that were initially
5179 referring to @var{package}. @xref{Security Updates}, for more
5180 information on grafts.
5182 For example, the command below grafts version 3.5.4 of GnuTLS onto Wget
5183 and all its dependencies, replacing references to the version of GnuTLS
5184 they currently refer to:
5187 guix build --with-graft=gnutls=gnutls@@3.5.4 wget
5190 This has the advantage of being much faster than rebuilding everything.
5191 But there is a caveat: it works if and only if @var{package} and
5192 @var{replacement} are strictly compatible---for example, if they provide
5193 a library, the application binary interface (ABI) of those libraries
5194 must be compatible. If @var{replacement} is somehow incompatible with
5195 @var{package}, then the resulting package may be unusable. Use with
5200 @node Additional Build Options
5201 @subsection Additional Build Options
5203 The command-line options presented below are specific to @command{guix
5210 Build quietly, without displaying the build log. Upon completion, the
5211 build log is kept in @file{/var} (or similar) and can always be
5212 retrieved using the @option{--log-file} option.
5214 @item --file=@var{file}
5215 @itemx -f @var{file}
5217 Build the package or derivation that the code within @var{file}
5220 As an example, @var{file} might contain a package definition like this
5221 (@pxref{Defining Packages}):
5224 @verbatiminclude package-hello.scm
5227 @item --expression=@var{expr}
5228 @itemx -e @var{expr}
5229 Build the package or derivation @var{expr} evaluates to.
5231 For example, @var{expr} may be @code{(@@ (gnu packages guile)
5232 guile-1.8)}, which unambiguously designates this specific variant of
5233 version 1.8 of Guile.
5235 Alternatively, @var{expr} may be a G-expression, in which case it is used
5236 as a build program passed to @code{gexp->derivation}
5237 (@pxref{G-Expressions}).
5239 Lastly, @var{expr} may refer to a zero-argument monadic procedure
5240 (@pxref{The Store Monad}). The procedure must return a derivation as a
5241 monadic value, which is then passed through @code{run-with-store}.
5245 Build the source derivations of the packages, rather than the packages
5248 For instance, @code{guix build -S gcc} returns something like
5249 @file{/gnu/store/@dots{}-gcc-4.7.2.tar.bz2}, which is the GCC
5252 The returned source tarball is the result of applying any patches and
5253 code snippets specified in the package @code{origin} (@pxref{Defining
5257 Fetch and return the source of @var{package-or-derivation} and all their
5258 dependencies, recursively. This is a handy way to obtain a local copy
5259 of all the source code needed to build @var{packages}, allowing you to
5260 eventually build them even without network access. It is an extension
5261 of the @code{--source} option and can accept one of the following
5262 optional argument values:
5266 This value causes the @code{--sources} option to behave in the same way
5267 as the @code{--source} option.
5270 Build the source derivations of all packages, including any source that
5271 might be listed as @code{inputs}. This is the default value.
5274 $ guix build --sources tzdata
5275 The following derivations will be built:
5276 /gnu/store/@dots{}-tzdata2015b.tar.gz.drv
5277 /gnu/store/@dots{}-tzcode2015b.tar.gz.drv
5281 Build the source derivations of all packages, as well of all transitive
5282 inputs to the packages. This can be used e.g. to
5283 prefetch package source for later offline building.
5286 $ guix build --sources=transitive tzdata
5287 The following derivations will be built:
5288 /gnu/store/@dots{}-tzcode2015b.tar.gz.drv
5289 /gnu/store/@dots{}-findutils-4.4.2.tar.xz.drv
5290 /gnu/store/@dots{}-grep-2.21.tar.xz.drv
5291 /gnu/store/@dots{}-coreutils-8.23.tar.xz.drv
5292 /gnu/store/@dots{}-make-4.1.tar.xz.drv
5293 /gnu/store/@dots{}-bash-4.3.tar.xz.drv
5299 @item --system=@var{system}
5300 @itemx -s @var{system}
5301 Attempt to build for @var{system}---e.g., @code{i686-linux}---instead of
5302 the system type of the build host.
5304 An example use of this is on Linux-based systems, which can emulate
5305 different personalities. For instance, passing
5306 @code{--system=i686-linux} on an @code{x86_64-linux} system allows users
5307 to build packages in a complete 32-bit environment.
5309 @item --target=@var{triplet}
5310 @cindex cross-compilation
5311 Cross-build for @var{triplet}, which must be a valid GNU triplet, such
5312 as @code{"mips64el-linux-gnu"} (@pxref{Specifying target triplets, GNU
5313 configuration triplets,, autoconf, Autoconf}).
5315 @anchor{build-check}
5317 @cindex determinism, checking
5318 @cindex reproducibility, checking
5319 Rebuild @var{package-or-derivation}, which are already available in the
5320 store, and raise an error if the build results are not bit-for-bit
5323 This mechanism allows you to check whether previously installed
5324 substitutes are genuine (@pxref{Substitutes}), or whether the build result
5325 of a package is deterministic. @xref{Invoking guix challenge}, for more
5326 background information and tools.
5328 When used in conjunction with @option{--keep-failed}, the differing
5329 output is kept in the store, under @file{/gnu/store/@dots{}-check}.
5330 This makes it easy to look for differences between the two results.
5333 @cindex repairing store items
5334 @cindex corruption, recovering from
5335 Attempt to repair the specified store items, if they are corrupt, by
5336 re-downloading or rebuilding them.
5338 This operation is not atomic and thus restricted to @code{root}.
5342 Return the derivation paths, not the output paths, of the given
5345 @item --root=@var{file}
5346 @itemx -r @var{file}
5347 Make @var{file} a symlink to the result, and register it as a garbage
5351 Return the build log file names or URLs for the given
5352 @var{package-or-derivation}, or raise an error if build logs are
5355 This works regardless of how packages or derivations are specified. For
5356 instance, the following invocations are equivalent:
5359 guix build --log-file `guix build -d guile`
5360 guix build --log-file `guix build guile`
5361 guix build --log-file guile
5362 guix build --log-file -e '(@@ (gnu packages guile) guile-2.0)'
5365 If a log is unavailable locally, and unless @code{--no-substitutes} is
5366 passed, the command looks for a corresponding log on one of the
5367 substitute servers (as specified with @code{--substitute-urls}.)
5369 So for instance, imagine you want to see the build log of GDB on MIPS,
5370 but you are actually on an @code{x86_64} machine:
5373 $ guix build --log-file gdb -s mips64el-linux
5374 https://hydra.gnu.org/log/@dots{}-gdb-7.10
5377 You can freely access a huge library of build logs!
5380 @node Debugging Build Failures
5381 @subsection Debugging Build Failures
5383 @cindex build failures, debugging
5384 When defining a new package (@pxref{Defining Packages}), you will
5385 probably find yourself spending some time debugging and tweaking the
5386 build until it succeeds. To do that, you need to operate the build
5387 commands yourself in an environment as close as possible to the one the
5390 To that end, the first thing to do is to use the @option{--keep-failed}
5391 or @option{-K} option of @command{guix build}, which will keep the
5392 failed build tree in @file{/tmp} or whatever directory you specified as
5393 @code{TMPDIR} (@pxref{Invoking guix build, @code{--keep-failed}}).
5395 From there on, you can @command{cd} to the failed build tree and source
5396 the @file{environment-variables} file, which contains all the
5397 environment variable definitions that were in place when the build
5398 failed. So let's say you're debugging a build failure in package
5399 @code{foo}; a typical session would look like this:
5403 @dots{} @i{build fails}
5404 $ cd /tmp/guix-build-foo.drv-0
5405 $ source ./environment-variables
5409 Now, you can invoke commands as if you were the daemon (almost) and
5410 troubleshoot your build process.
5412 Sometimes it happens that, for example, a package's tests pass when you
5413 run them manually but they fail when the daemon runs them. This can
5414 happen because the daemon runs builds in containers where, unlike in our
5415 environment above, network access is missing, @file{/bin/sh} does not
5416 exist, etc. (@pxref{Build Environment Setup}).
5418 In such cases, you may need to run inspect the build process from within
5419 a container similar to the one the build daemon creates:
5424 $ cd /tmp/guix-build-foo.drv-0
5425 $ guix environment --no-grafts -C foo --ad-hoc strace gdb
5426 [env]# source ./environment-variables
5430 Here, @command{guix environment -C} creates a container and spawns a new
5431 shell in it (@pxref{Invoking guix environment}). The @command{--ad-hoc
5432 strace gdb} part adds the @command{strace} and @command{gdb} commands to
5433 the container, which would may find handy while debugging. The
5434 @option{--no-grafts} option makes sure we get the exact same
5435 environment, with ungrafted packages (@pxref{Security Updates}, for more
5438 To get closer to a container like that used by the build daemon, we can
5439 remove @file{/bin/sh}:
5445 (Don't worry, this is harmless: this is all happening in the throw-away
5446 container created by @command{guix environment}.)
5448 The @command{strace} command is probably not in the search path, but we
5452 [env]# $GUIX_ENVIRONMENT/bin/strace -f -o log make check
5455 In this way, not only you will have reproduced the environment variables
5456 the daemon uses, you will also be running the build process in a container
5457 similar to the one the daemon uses.
5460 @node Invoking guix edit
5461 @section Invoking @command{guix edit}
5463 @cindex @command{guix edit}
5464 @cindex package definition, editing
5465 So many packages, so many source files! The @command{guix edit} command
5466 facilitates the life of users and packagers by pointing their editor at
5467 the source file containing the definition of the specified packages.
5471 guix edit gcc@@4.9 vim
5475 launches the program specified in the @code{VISUAL} or in the
5476 @code{EDITOR} environment variable to view the recipe of GCC@tie{}4.9.3
5479 If you are using a Guix Git checkout (@pxref{Building from Git}), or
5480 have created your own packages on @code{GUIX_PACKAGE_PATH}
5481 (@pxref{Defining Packages}), you will be able to edit the package
5482 recipes. Otherwise, you will be able to examine the read-only recipes
5483 for packages currently in the store.
5486 @node Invoking guix download
5487 @section Invoking @command{guix download}
5489 @cindex @command{guix download}
5490 @cindex downloading package sources
5491 When writing a package definition, developers typically need to download
5492 a source tarball, compute its SHA256 hash, and write that
5493 hash in the package definition (@pxref{Defining Packages}). The
5494 @command{guix download} tool helps with this task: it downloads a file
5495 from the given URI, adds it to the store, and prints both its file name
5496 in the store and its SHA256 hash.
5498 The fact that the downloaded file is added to the store saves bandwidth:
5499 when the developer eventually tries to build the newly defined package
5500 with @command{guix build}, the source tarball will not have to be
5501 downloaded again because it is already in the store. It is also a
5502 convenient way to temporarily stash files, which may be deleted
5503 eventually (@pxref{Invoking guix gc}).
5505 The @command{guix download} command supports the same URIs as used in
5506 package definitions. In particular, it supports @code{mirror://} URIs.
5507 @code{https} URIs (HTTP over TLS) are supported @emph{provided} the
5508 Guile bindings for GnuTLS are available in the user's environment; when
5509 they are not available, an error is raised. @xref{Guile Preparations,
5510 how to install the GnuTLS bindings for Guile,, gnutls-guile,
5511 GnuTLS-Guile}, for more information.
5513 @command{guix download} verifies HTTPS server certificates by loading
5514 the certificates of X.509 authorities from the directory pointed to by
5515 the @code{SSL_CERT_DIR} environment variable (@pxref{X.509
5516 Certificates}), unless @option{--no-check-certificate} is used.
5518 The following options are available:
5521 @item --format=@var{fmt}
5523 Write the hash in the format specified by @var{fmt}. For more
5524 information on the valid values for @var{fmt}, @pxref{Invoking guix hash}.
5526 @item --no-check-certificate
5527 Do not validate the X.509 certificates of HTTPS servers.
5529 When using this option, you have @emph{absolutely no guarantee} that you
5530 are communicating with the authentic server responsible for the given
5531 URL, which makes you vulnerable to ``man-in-the-middle'' attacks.
5533 @item --output=@var{file}
5534 @itemx -o @var{file}
5535 Save the downloaded file to @var{file} instead of adding it to the
5539 @node Invoking guix hash
5540 @section Invoking @command{guix hash}
5542 @cindex @command{guix hash}
5543 The @command{guix hash} command computes the SHA256 hash of a file.
5544 It is primarily a convenience tool for anyone contributing to the
5545 distribution: it computes the cryptographic hash of a file, which can be
5546 used in the definition of a package (@pxref{Defining Packages}).
5548 The general syntax is:
5551 guix hash @var{option} @var{file}
5554 When @var{file} is @code{-} (a hyphen), @command{guix hash} computes the
5555 hash of data read from standard input. @command{guix hash} has the
5560 @item --format=@var{fmt}
5562 Write the hash in the format specified by @var{fmt}.
5564 Supported formats: @code{nix-base32}, @code{base32}, @code{base16}
5565 (@code{hex} and @code{hexadecimal} can be used as well).
5567 If the @option{--format} option is not specified, @command{guix hash}
5568 will output the hash in @code{nix-base32}. This representation is used
5569 in the definitions of packages.
5573 Compute the hash on @var{file} recursively.
5575 In this case, the hash is computed on an archive containing @var{file},
5576 including its children if it is a directory. Some of the metadata of
5577 @var{file} is part of the archive; for instance, when @var{file} is a
5578 regular file, the hash is different depending on whether @var{file} is
5579 executable or not. Metadata such as time stamps has no impact on the
5580 hash (@pxref{Invoking guix archive}).
5581 @c FIXME: Replace xref above with xref to an ``Archive'' section when
5586 When combined with @option{--recursive}, exclude version control system
5587 directories (@file{.bzr}, @file{.git}, @file{.hg}, etc.)
5590 As an example, here is how you would compute the hash of a Git checkout,
5591 which is useful when using the @code{git-fetch} method (@pxref{origin
5595 $ git clone http://example.org/foo.git
5601 @node Invoking guix import
5602 @section Invoking @command{guix import}
5604 @cindex importing packages
5605 @cindex package import
5606 @cindex package conversion
5607 @cindex Invoking @command{guix import}
5608 The @command{guix import} command is useful for people who would like to
5609 add a package to the distribution with as little work as
5610 possible---a legitimate demand. The command knows of a few
5611 repositories from which it can ``import'' package metadata. The result
5612 is a package definition, or a template thereof, in the format we know
5613 (@pxref{Defining Packages}).
5615 The general syntax is:
5618 guix import @var{importer} @var{options}@dots{}
5621 @var{importer} specifies the source from which to import package
5622 metadata, and @var{options} specifies a package identifier and other
5623 options specific to @var{importer}. Currently, the available
5628 Import metadata for the given GNU package. This provides a template
5629 for the latest version of that GNU package, including the hash of its
5630 source tarball, and its canonical synopsis and description.
5632 Additional information such as the package dependencies and its
5633 license needs to be figured out manually.
5635 For example, the following command returns a package definition for
5639 guix import gnu hello
5642 Specific command-line options are:
5645 @item --key-download=@var{policy}
5646 As for @code{guix refresh}, specify the policy to handle missing OpenPGP
5647 keys when verifying the package signature. @xref{Invoking guix
5648 refresh, @code{--key-download}}.
5653 Import metadata from the @uref{https://pypi.python.org/, Python Package
5654 Index}@footnote{This functionality requires Guile-JSON to be installed.
5655 @xref{Requirements}.}. Information is taken from the JSON-formatted
5656 description available at @code{pypi.python.org} and usually includes all
5657 the relevant information, including package dependencies. For maximum
5658 efficiency, it is recommended to install the @command{unzip} utility, so
5659 that the importer can unzip Python wheels and gather data from them.
5661 The command below imports metadata for the @code{itsdangerous} Python
5665 guix import pypi itsdangerous
5670 Import metadata from @uref{https://rubygems.org/,
5671 RubyGems}@footnote{This functionality requires Guile-JSON to be
5672 installed. @xref{Requirements}.}. Information is taken from the
5673 JSON-formatted description available at @code{rubygems.org} and includes
5674 most relevant information, including runtime dependencies. There are
5675 some caveats, however. The metadata doesn't distinguish between
5676 synopses and descriptions, so the same string is used for both fields.
5677 Additionally, the details of non-Ruby dependencies required to build
5678 native extensions is unavailable and left as an exercise to the
5681 The command below imports metadata for the @code{rails} Ruby package:
5684 guix import gem rails
5689 Import metadata from @uref{https://www.metacpan.org/, MetaCPAN}@footnote{This
5690 functionality requires Guile-JSON to be installed.
5691 @xref{Requirements}.}.
5692 Information is taken from the JSON-formatted metadata provided through
5693 @uref{https://api.metacpan.org/, MetaCPAN's API} and includes most
5694 relevant information, such as module dependencies. License information
5695 should be checked closely. If Perl is available in the store, then the
5696 @code{corelist} utility will be used to filter core modules out of the
5697 list of dependencies.
5699 The command command below imports metadata for the @code{Acme::Boolean}
5703 guix import cpan Acme::Boolean
5708 @cindex Bioconductor
5709 Import metadata from @uref{http://cran.r-project.org/, CRAN}, the
5710 central repository for the @uref{http://r-project.org, GNU@tie{}R
5711 statistical and graphical environment}.
5713 Information is extracted from the @code{DESCRIPTION} file of the package.
5715 The command command below imports metadata for the @code{Cairo}
5719 guix import cran Cairo
5722 When @code{--recursive} is added, the importer will traverse the
5723 dependency graph of the given upstream package recursively and generate
5724 package expressions for all those packages that are not yet in Guix.
5726 When @code{--archive=bioconductor} is added, metadata is imported from
5727 @uref{http://www.bioconductor.org/, Bioconductor}, a repository of R
5728 packages for for the analysis and comprehension of high-throughput
5729 genomic data in bioinformatics.
5731 Information is extracted from the @code{DESCRIPTION} file of a package
5732 published on the web interface of the Bioconductor SVN repository.
5734 The command below imports metadata for the @code{GenomicRanges}
5738 guix import cran --archive=bioconductor GenomicRanges
5744 Import metadata from @uref{http://www.ctan.org/, CTAN}, the
5745 comprehensive TeX archive network for TeX packages that are part of the
5746 @uref{https://www.tug.org/texlive/, TeX Live distribution}.
5748 Information about the package is obtained through the XML API provided
5749 by CTAN, while the source code is downloaded from the SVN repository of
5750 the Tex Live project. This is done because the CTAN does not keep
5753 The command command below imports metadata for the @code{fontspec}
5757 guix import texlive fontspec
5760 When @code{--archive=DIRECTORY} is added, the source code is downloaded
5761 not from the @file{latex} sub-directory of the @file{texmf-dist/source}
5762 tree in the TeX Live SVN repository, but from the specified sibling
5763 directory under the same root.
5765 The command below imports metadata for the @code{ifxetex} package from
5766 CTAN while fetching the sources from the directory
5767 @file{texmf/source/generic}:
5770 guix import texlive --archive=generic ifxetex
5774 Import metadata from a local copy of the source of the
5775 @uref{http://nixos.org/nixpkgs/, Nixpkgs distribution}@footnote{This
5776 relies on the @command{nix-instantiate} command of
5777 @uref{http://nixos.org/nix/, Nix}.}. Package definitions in Nixpkgs are
5778 typically written in a mixture of Nix-language and Bash code. This
5779 command only imports the high-level package structure that is written in
5780 the Nix language. It normally includes all the basic fields of a
5783 When importing a GNU package, the synopsis and descriptions are replaced
5784 by their canonical upstream variant.
5786 Usually, you will first need to do:
5789 export NIX_REMOTE=daemon
5793 so that @command{nix-instantiate} does not try to open the Nix database.
5795 As an example, the command below imports the package definition of
5796 LibreOffice (more precisely, it imports the definition of the package
5797 bound to the @code{libreoffice} top-level attribute):
5800 guix import nix ~/path/to/nixpkgs libreoffice
5805 Import metadata from the Haskell community's central package archive
5806 @uref{https://hackage.haskell.org/, Hackage}. Information is taken from
5807 Cabal files and includes all the relevant information, including package
5810 Specific command-line options are:
5815 Read a Cabal file from standard input.
5816 @item --no-test-dependencies
5818 Do not include dependencies required only by the test suites.
5819 @item --cabal-environment=@var{alist}
5820 @itemx -e @var{alist}
5821 @var{alist} is a Scheme alist defining the environment in which the
5822 Cabal conditionals are evaluated. The accepted keys are: @code{os},
5823 @code{arch}, @code{impl} and a string representing the name of a flag.
5824 The value associated with a flag has to be either the symbol
5825 @code{true} or @code{false}. The value associated with other keys
5826 has to conform to the Cabal file format definition. The default value
5827 associated with the keys @code{os}, @code{arch} and @code{impl} is
5828 @samp{linux}, @samp{x86_64} and @samp{ghc}, respectively.
5831 The command below imports metadata for the latest version of the
5832 @code{HTTP} Haskell package without including test dependencies and
5833 specifying the value of the flag @samp{network-uri} as @code{false}:
5836 guix import hackage -t -e "'((\"network-uri\" . false))" HTTP
5839 A specific package version may optionally be specified by following the
5840 package name by an at-sign and a version number as in the following example:
5843 guix import hackage mtl@@2.1.3.1
5848 The @code{stackage} importer is a wrapper around the @code{hackage} one.
5849 It takes a package name, looks up the package version included in a
5850 long-term support (LTS) @uref{https://www.stackage.org, Stackage}
5851 release and uses the @code{hackage} importer to retrieve its metadata.
5852 Note that it is up to you to select an LTS release compatible with the
5853 GHC compiler used by Guix.
5855 Specific command-line options are:
5858 @item --no-test-dependencies
5860 Do not include dependencies required only by the test suites.
5861 @item --lts-version=@var{version}
5862 @itemx -r @var{version}
5863 @var{version} is the desired LTS release version. If omitted the latest
5867 The command below imports metadata for the @code{HTTP} Haskell package
5868 included in the LTS Stackage release version 7.18:
5871 guix import stackage --lts-version=7.18 HTTP
5876 Import metadata from an Emacs Lisp Package Archive (ELPA) package
5877 repository (@pxref{Packages,,, emacs, The GNU Emacs Manual}).
5879 Specific command-line options are:
5882 @item --archive=@var{repo}
5883 @itemx -a @var{repo}
5884 @var{repo} identifies the archive repository from which to retrieve the
5885 information. Currently the supported repositories and their identifiers
5889 @uref{http://elpa.gnu.org/packages, GNU}, selected by the @code{gnu}
5890 identifier. This is the default.
5892 Packages from @code{elpa.gnu.org} are signed with one of the keys
5893 contained in the GnuPG keyring at
5894 @file{share/emacs/25.1/etc/package-keyring.gpg} (or similar) in the
5895 @code{emacs} package (@pxref{Package Installation, ELPA package
5896 signatures,, emacs, The GNU Emacs Manual}).
5899 @uref{http://stable.melpa.org/packages, MELPA-Stable}, selected by the
5900 @code{melpa-stable} identifier.
5903 @uref{http://melpa.org/packages, MELPA}, selected by the @code{melpa}
5910 Import metadata from the crates.io Rust package repository
5911 @uref{https://crates.io, crates.io}.
5914 The structure of the @command{guix import} code is modular. It would be
5915 useful to have more importers for other package formats, and your help
5916 is welcome here (@pxref{Contributing}).
5918 @node Invoking guix refresh
5919 @section Invoking @command{guix refresh}
5921 @cindex @command {guix refresh}
5922 The primary audience of the @command{guix refresh} command is developers
5923 of the GNU software distribution. By default, it reports any packages
5924 provided by the distribution that are outdated compared to the latest
5925 upstream version, like this:
5929 gnu/packages/gettext.scm:29:13: gettext would be upgraded from 0.18.1.1 to 0.18.2.1
5930 gnu/packages/glib.scm:77:12: glib would be upgraded from 2.34.3 to 2.37.0
5933 Alternately, one can specify packages to consider, in which case a
5934 warning is emitted for packages that lack an updater:
5937 $ guix refresh coreutils guile guile-ssh
5938 gnu/packages/ssh.scm:205:2: warning: no updater for guile-ssh
5939 gnu/packages/guile.scm:136:12: guile would be upgraded from 2.0.12 to 2.0.13
5942 @command{guix refresh} browses the upstream repository of each package and determines
5943 the highest version number of the releases therein. The command
5944 knows how to update specific types of packages: GNU packages, ELPA
5945 packages, etc.---see the documentation for @option{--type} below. There
5946 are many packages, though, for which it lacks a method to determine
5947 whether a new upstream release is available. However, the mechanism is
5948 extensible, so feel free to get in touch with us to add a new method!
5950 When passed @code{--update}, it modifies distribution source files to
5951 update the version numbers and source tarball hashes of those package
5952 recipes (@pxref{Defining Packages}). This is achieved by downloading
5953 each package's latest source tarball and its associated OpenPGP
5954 signature, authenticating the downloaded tarball against its signature
5955 using @command{gpg}, and finally computing its hash. When the public
5956 key used to sign the tarball is missing from the user's keyring, an
5957 attempt is made to automatically retrieve it from a public key server;
5958 when this is successful, the key is added to the user's keyring; otherwise,
5959 @command{guix refresh} reports an error.
5961 The following options are supported:
5965 @item --expression=@var{expr}
5966 @itemx -e @var{expr}
5967 Consider the package @var{expr} evaluates to.
5969 This is useful to precisely refer to a package, as in this example:
5972 guix refresh -l -e '(@@@@ (gnu packages commencement) glibc-final)'
5975 This command lists the dependents of the ``final'' libc (essentially all
5980 Update distribution source files (package recipes) in place. This is
5981 usually run from a checkout of the Guix source tree (@pxref{Running
5982 Guix Before It Is Installed}):
5985 $ ./pre-inst-env guix refresh -s non-core -u
5988 @xref{Defining Packages}, for more information on package definitions.
5990 @item --select=[@var{subset}]
5991 @itemx -s @var{subset}
5992 Select all the packages in @var{subset}, one of @code{core} or
5995 The @code{core} subset refers to all the packages at the core of the
5996 distribution---i.e., packages that are used to build ``everything
5997 else''. This includes GCC, libc, Binutils, Bash, etc. Usually,
5998 changing one of these packages in the distribution entails a rebuild of
5999 all the others. Thus, such updates are an inconvenience to users in
6000 terms of build time or bandwidth used to achieve the upgrade.
6002 The @code{non-core} subset refers to the remaining packages. It is
6003 typically useful in cases where an update of the core packages would be
6006 @item --manifest=@var{file}
6007 @itemx -m @var{file}
6008 Select all the packages from the manifest in @var{file}. This is useful to
6009 check if any packages of the user manifest can be updated.
6011 @item --type=@var{updater}
6012 @itemx -t @var{updater}
6013 Select only packages handled by @var{updater} (may be a comma-separated
6014 list of updaters). Currently, @var{updater} may be one of:
6018 the updater for GNU packages;
6020 the updater for GNOME packages;
6022 the updater for KDE packages;
6024 the updater for X.org packages;
6026 the updater for packages hosted on kernel.org;
6028 the updater for @uref{http://elpa.gnu.org/, ELPA} packages;
6030 the updater for @uref{http://cran.r-project.org/, CRAN} packages;
6032 the updater for @uref{http://www.bioconductor.org/, Bioconductor} R packages;
6034 the updater for @uref{http://www.cpan.org/, CPAN} packages;
6036 the updater for @uref{https://pypi.python.org, PyPI} packages.
6038 the updater for @uref{https://rubygems.org, RubyGems} packages.
6040 the updater for @uref{https://github.com, GitHub} packages.
6042 the updater for @uref{https://hackage.haskell.org, Hackage} packages.
6044 the updater for @uref{https://www.stackage.org, Stackage} packages.
6046 the updater for @uref{https://crates.io, Crates} packages.
6049 For instance, the following command only checks for updates of Emacs
6050 packages hosted at @code{elpa.gnu.org} and for updates of CRAN packages:
6053 $ guix refresh --type=elpa,cran
6054 gnu/packages/statistics.scm:819:13: r-testthat would be upgraded from 0.10.0 to 0.11.0
6055 gnu/packages/emacs.scm:856:13: emacs-auctex would be upgraded from 11.88.6 to 11.88.9
6060 In addition, @command{guix refresh} can be passed one or more package
6061 names, as in this example:
6064 $ ./pre-inst-env guix refresh -u emacs idutils gcc@@4.8
6068 The command above specifically updates the @code{emacs} and
6069 @code{idutils} packages. The @code{--select} option would have no
6070 effect in this case.
6072 When considering whether to upgrade a package, it is sometimes
6073 convenient to know which packages would be affected by the upgrade and
6074 should be checked for compatibility. For this the following option may
6075 be used when passing @command{guix refresh} one or more package names:
6079 @item --list-updaters
6081 List available updaters and exit (see @option{--type} above.)
6083 For each updater, display the fraction of packages it covers; at the
6084 end, display the fraction of packages covered by all these updaters.
6086 @item --list-dependent
6088 List top-level dependent packages that would need to be rebuilt as a
6089 result of upgrading one or more packages.
6091 @xref{Invoking guix graph, the @code{reverse-package} type of
6092 @command{guix graph}}, for information on how to visualize the list of
6093 dependents of a package.
6097 Be aware that the @code{--list-dependent} option only
6098 @emph{approximates} the rebuilds that would be required as a result of
6099 an upgrade. More rebuilds might be required under some circumstances.
6102 $ guix refresh --list-dependent flex
6103 Building the following 120 packages would ensure 213 dependent packages are rebuilt:
6104 hop@@2.4.0 geiser@@0.4 notmuch@@0.18 mu@@0.9.9.5 cflow@@1.4 idutils@@4.6 @dots{}
6107 The command above lists a set of packages that could be built to check
6108 for compatibility with an upgraded @code{flex} package.
6110 The following options can be used to customize GnuPG operation:
6114 @item --gpg=@var{command}
6115 Use @var{command} as the GnuPG 2.x command. @var{command} is searched
6116 for in @code{$PATH}.
6118 @item --key-download=@var{policy}
6119 Handle missing OpenPGP keys according to @var{policy}, which may be one
6124 Always download missing OpenPGP keys from the key server, and add them
6125 to the user's GnuPG keyring.
6128 Never try to download missing OpenPGP keys. Instead just bail out.
6131 When a package signed with an unknown OpenPGP key is encountered, ask
6132 the user whether to download it or not. This is the default behavior.
6135 @item --key-server=@var{host}
6136 Use @var{host} as the OpenPGP key server when importing a public key.
6140 The @code{github} updater uses the
6141 @uref{https://developer.github.com/v3/, GitHub API} to query for new
6142 releases. When used repeatedly e.g. when refreshing all packages,
6143 GitHub will eventually refuse to answer any further API requests. By
6144 default 60 API requests per hour are allowed, and a full refresh on all
6145 GitHub packages in Guix requires more than this. Authentication with
6146 GitHub through the use of an API token alleviates these limits. To use
6147 an API token, set the environment variable @code{GUIX_GITHUB_TOKEN} to a
6148 token procured from @uref{https://github.com/settings/tokens} or
6152 @node Invoking guix lint
6153 @section Invoking @command{guix lint}
6155 @cindex @command{guix lint}
6156 @cindex package, checking for errors
6157 The @command{guix lint} command is meant to help package developers avoid
6158 common errors and use a consistent style. It runs a number of checks on
6159 a given set of packages in order to find common mistakes in their
6160 definitions. Available @dfn{checkers} include (see
6161 @code{--list-checkers} for a complete list):
6166 Validate certain typographical and stylistic rules about package
6167 descriptions and synopses.
6169 @item inputs-should-be-native
6170 Identify inputs that should most likely be native inputs.
6175 @itemx source-file-name
6176 Probe @code{home-page} and @code{source} URLs and report those that are
6177 invalid. Suggest a @code{mirror://} URL when applicable. Check that
6178 the source file name is meaningful, e.g. is not
6179 just a version number or ``git-checkout'', without a declared
6180 @code{file-name} (@pxref{origin Reference}).
6183 @cindex security vulnerabilities
6184 @cindex CVE, Common Vulnerabilities and Exposures
6185 Report known vulnerabilities found in the Common Vulnerabilities and
6186 Exposures (CVE) databases of the current and past year
6187 @uref{https://nvd.nist.gov/download.cfm#CVE_FEED, published by the US
6190 To view information about a particular vulnerability, visit pages such as:
6194 @indicateurl{https://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-YYYY-ABCD}
6196 @indicateurl{https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-YYYY-ABCD}
6200 where @code{CVE-YYYY-ABCD} is the CVE identifier---e.g.,
6201 @code{CVE-2015-7554}.
6203 Package developers can specify in package recipes the
6204 @uref{https://nvd.nist.gov/cpe.cfm,Common Platform Enumeration (CPE)}
6205 name and version of the package when they differ from the name that Guix
6206 uses, as in this example:
6212 ;; CPE calls this package "grub2".
6213 (properties '((cpe-name . "grub2"))))
6217 Warn about obvious source code formatting issues: trailing white space,
6218 use of tabulations, etc.
6221 The general syntax is:
6224 guix lint @var{options} @var{package}@dots{}
6227 If no package is given on the command line, then all packages are checked.
6228 The @var{options} may be zero or more of the following:
6231 @item --list-checkers
6233 List and describe all the available checkers that will be run on packages
6238 Only enable the checkers specified in a comma-separated list using the
6239 names returned by @code{--list-checkers}.
6243 @node Invoking guix size
6244 @section Invoking @command{guix size}
6247 @cindex package size
6249 @cindex @command{guix size}
6250 The @command{guix size} command helps package developers profile the
6251 disk usage of packages. It is easy to overlook the impact of an
6252 additional dependency added to a package, or the impact of using a
6253 single output for a package that could easily be split (@pxref{Packages
6254 with Multiple Outputs}). Such are the typical issues that
6255 @command{guix size} can highlight.
6257 The command can be passed a package specification such as @code{gcc@@4.8}
6258 or @code{guile:debug}, or a file name in the store. Consider this
6262 $ guix size coreutils
6263 store item total self
6264 /gnu/store/@dots{}-coreutils-8.23 70.0 13.9 19.8%
6265 /gnu/store/@dots{}-gmp-6.0.0a 55.3 2.5 3.6%
6266 /gnu/store/@dots{}-acl-2.2.52 53.7 0.5 0.7%
6267 /gnu/store/@dots{}-attr-2.4.46 53.2 0.3 0.5%
6268 /gnu/store/@dots{}-gcc-4.8.4-lib 52.9 15.7 22.4%
6269 /gnu/store/@dots{}-glibc-2.21 37.2 37.2 53.1%
6273 The store items listed here constitute the @dfn{transitive closure} of
6274 Coreutils---i.e., Coreutils and all its dependencies, recursively---as
6275 would be returned by:
6278 $ guix gc -R /gnu/store/@dots{}-coreutils-8.23
6281 Here the output shows three columns next to store items. The first column,
6282 labeled ``total'', shows the size in mebibytes (MiB) of the closure of
6283 the store item---that is, its own size plus the size of all its
6284 dependencies. The next column, labeled ``self'', shows the size of the
6285 item itself. The last column shows the ratio of the size of the item
6286 itself to the space occupied by all the items listed here.
6288 In this example, we see that the closure of Coreutils weighs in at
6289 70@tie{}MiB, half of which is taken by libc. (That libc represents a
6290 large fraction of the closure is not a problem @i{per se} because it is
6291 always available on the system anyway.)
6293 When the package passed to @command{guix size} is available in the
6294 store, @command{guix size} queries the daemon to determine its
6295 dependencies, and measures its size in the store, similar to @command{du
6296 -ms --apparent-size} (@pxref{du invocation,,, coreutils, GNU
6299 When the given package is @emph{not} in the store, @command{guix size}
6300 reports information based on the available substitutes
6301 (@pxref{Substitutes}). This makes it possible it to profile disk usage of
6302 store items that are not even on disk, only available remotely.
6304 You can also specify several package names:
6307 $ guix size coreutils grep sed bash
6308 store item total self
6309 /gnu/store/@dots{}-coreutils-8.24 77.8 13.8 13.4%
6310 /gnu/store/@dots{}-grep-2.22 73.1 0.8 0.8%
6311 /gnu/store/@dots{}-bash-4.3.42 72.3 4.7 4.6%
6312 /gnu/store/@dots{}-readline-6.3 67.6 1.2 1.2%
6318 In this example we see that the combination of the four packages takes
6319 102.3@tie{}MiB in total, which is much less than the sum of each closure
6320 since they have a lot of dependencies in common.
6322 The available options are:
6326 @item --substitute-urls=@var{urls}
6327 Use substitute information from @var{urls}.
6328 @xref{client-substitute-urls, the same option for @code{guix build}}.
6330 @item --map-file=@var{file}
6331 Write a graphical map of disk usage in PNG format to @var{file}.
6333 For the example above, the map looks like this:
6335 @image{images/coreutils-size-map,5in,, map of Coreutils disk usage
6336 produced by @command{guix size}}
6338 This option requires that
6339 @uref{http://wingolog.org/software/guile-charting/, Guile-Charting} be
6340 installed and visible in Guile's module search path. When that is not
6341 the case, @command{guix size} fails as it tries to load it.
6343 @item --system=@var{system}
6344 @itemx -s @var{system}
6345 Consider packages for @var{system}---e.g., @code{x86_64-linux}.
6349 @node Invoking guix graph
6350 @section Invoking @command{guix graph}
6353 @cindex @command{guix graph}
6354 @cindex package dependencies
6355 Packages and their dependencies form a @dfn{graph}, specifically a
6356 directed acyclic graph (DAG). It can quickly become difficult to have a
6357 mental model of the package DAG, so the @command{guix graph} command
6358 provides a visual representation of the DAG. By default,
6359 @command{guix graph} emits a DAG representation in the input format of
6360 @uref{http://www.graphviz.org/, Graphviz}, so its output can be passed
6361 directly to the @command{dot} command of Graphviz. It can also emit an
6362 HTML page with embedded JavaScript code to display a ``chord diagram''
6363 in a Web browser, using the @uref{https://d3js.org/, d3.js} library, or
6364 emit Cypher queries to construct a graph in a graph database supporting
6365 the @uref{http://www.opencypher.org/, openCypher} query language.
6366 The general syntax is:
6369 guix graph @var{options} @var{package}@dots{}
6372 For example, the following command generates a PDF file representing the
6373 package DAG for the GNU@tie{}Core Utilities, showing its build-time
6377 guix graph coreutils | dot -Tpdf > dag.pdf
6380 The output looks like this:
6382 @image{images/coreutils-graph,2in,,Dependency graph of the GNU Coreutils}
6384 Nice little graph, no?
6386 But there is more than one graph! The one above is concise: it is the
6387 graph of package objects, omitting implicit inputs such as GCC, libc,
6388 grep, etc. It is often useful to have such a concise graph, but
6389 sometimes one may want to see more details. @command{guix graph} supports
6390 several types of graphs, allowing you to choose the level of detail:
6394 This is the default type used in the example above. It shows the DAG of
6395 package objects, excluding implicit dependencies. It is concise, but
6396 filters out many details.
6398 @item reverse-package
6399 This shows the @emph{reverse} DAG of packages. For example:
6402 guix graph --type=reverse-package ocaml
6405 ... yields the graph of packages that depend on OCaml.
6407 Note that for core packages this can yield huge graphs. If all you want
6408 is to know the number of packages that depend on a given package, use
6409 @command{guix refresh --list-dependent} (@pxref{Invoking guix refresh,
6410 @option{--list-dependent}}).
6413 This is the package DAG, @emph{including} implicit inputs.
6415 For instance, the following command:
6418 guix graph --type=bag-emerged coreutils | dot -Tpdf > dag.pdf
6421 ... yields this bigger graph:
6423 @image{images/coreutils-bag-graph,,5in,Detailed dependency graph of the GNU Coreutils}
6425 At the bottom of the graph, we see all the implicit inputs of
6426 @var{gnu-build-system} (@pxref{Build Systems, @code{gnu-build-system}}).
6428 Now, note that the dependencies of these implicit inputs---that is, the
6429 @dfn{bootstrap dependencies} (@pxref{Bootstrapping})---are not shown
6430 here, for conciseness.
6433 Similar to @code{bag-emerged}, but this time including all the bootstrap
6436 @item bag-with-origins
6437 Similar to @code{bag}, but also showing origins and their dependencies.
6440 This is the most detailed representation: It shows the DAG of
6441 derivations (@pxref{Derivations}) and plain store items. Compared to
6442 the above representation, many additional nodes are visible, including
6443 build scripts, patches, Guile modules, etc.
6445 For this type of graph, it is also possible to pass a @file{.drv} file
6446 name instead of a package name, as in:
6449 guix graph -t derivation `guix system build -d my-config.scm`
6453 All the types above correspond to @emph{build-time dependencies}. The
6454 following graph type represents the @emph{run-time dependencies}:
6458 This is the graph of @dfn{references} of a package output, as returned
6459 by @command{guix gc --references} (@pxref{Invoking guix gc}).
6461 If the given package output is not available in the store, @command{guix
6462 graph} attempts to obtain dependency information from substitutes.
6464 Here you can also pass a store file name instead of a package name. For
6465 example, the command below produces the reference graph of your profile
6466 (which can be big!):
6469 guix graph -t references `readlink -f ~/.guix-profile`
6473 This is the graph of the @dfn{referrers} of a store item, as returned by
6474 @command{guix gc --referrers} (@pxref{Invoking guix gc}).
6476 This relies exclusively on local information from your store. For
6477 instance, let us suppose that the current Inkscape is available in 10
6478 profiles on your machine; @command{guix graph -t referrers inkscape}
6479 will show a graph rooted at Inkscape and with those 10 profiles linked
6482 It can help determine what is preventing a store item from being garbage
6487 The available options are the following:
6490 @item --type=@var{type}
6491 @itemx -t @var{type}
6492 Produce a graph output of @var{type}, where @var{type} must be one of
6493 the values listed above.
6496 List the supported graph types.
6498 @item --backend=@var{backend}
6499 @itemx -b @var{backend}
6500 Produce a graph using the selected @var{backend}.
6502 @item --list-backends
6503 List the supported graph backends.
6505 Currently, the available backends are Graphviz and d3.js.
6507 @item --expression=@var{expr}
6508 @itemx -e @var{expr}
6509 Consider the package @var{expr} evaluates to.
6511 This is useful to precisely refer to a package, as in this example:
6514 guix graph -e '(@@@@ (gnu packages commencement) gnu-make-final)'
6519 @node Invoking guix environment
6520 @section Invoking @command{guix environment}
6522 @cindex reproducible build environments
6523 @cindex development environments
6524 @cindex @command{guix environment}
6525 @cindex environment, package build environment
6526 The purpose of @command{guix environment} is to assist hackers in
6527 creating reproducible development environments without polluting their
6528 package profile. The @command{guix environment} tool takes one or more
6529 packages, builds all of their inputs, and creates a shell
6530 environment to use them.
6532 The general syntax is:
6535 guix environment @var{options} @var{package}@dots{}
6538 The following example spawns a new shell set up for the development of
6542 guix environment guile
6545 If the needed dependencies are not built yet, @command{guix environment}
6546 automatically builds them. The environment of the new shell is an augmented
6547 version of the environment that @command{guix environment} was run in.
6548 It contains the necessary search paths for building the given package
6549 added to the existing environment variables. To create a ``pure''
6550 environment, in which the original environment variables have been unset,
6551 use the @code{--pure} option@footnote{Users sometimes wrongfully augment
6552 environment variables such as @code{PATH} in their @file{~/.bashrc}
6553 file. As a consequence, when @code{guix environment} launches it, Bash
6554 may read @file{~/.bashrc}, thereby introducing ``impurities'' in these
6555 environment variables. It is an error to define such environment
6556 variables in @file{.bashrc}; instead, they should be defined in
6557 @file{.bash_profile}, which is sourced only by log-in shells.
6558 @xref{Bash Startup Files,,, bash, The GNU Bash Reference Manual}, for
6559 details on Bash start-up files.}.
6561 @vindex GUIX_ENVIRONMENT
6562 @command{guix environment} defines the @code{GUIX_ENVIRONMENT}
6563 variable in the shell it spawns; its value is the file name of the
6564 profile of this environment. This allows users to, say, define a
6565 specific prompt for development environments in their @file{.bashrc}
6566 (@pxref{Bash Startup Files,,, bash, The GNU Bash Reference Manual}):
6569 if [ -n "$GUIX_ENVIRONMENT" ]
6571 export PS1="\u@@\h \w [dev]\$ "
6576 ... or to browse the profile:
6579 $ ls "$GUIX_ENVIRONMENT/bin"
6582 Additionally, more than one package may be specified, in which case the
6583 union of the inputs for the given packages are used. For example, the
6584 command below spawns a shell where all of the dependencies of both Guile
6585 and Emacs are available:
6588 guix environment guile emacs
6591 Sometimes an interactive shell session is not desired. An arbitrary
6592 command may be invoked by placing the @code{--} token to separate the
6593 command from the rest of the arguments:
6596 guix environment guile -- make -j4
6599 In other situations, it is more convenient to specify the list of
6600 packages needed in the environment. For example, the following command
6601 runs @command{python} from an environment containing Python@tie{}2.7 and
6605 guix environment --ad-hoc python2-numpy python-2.7 -- python
6608 Furthermore, one might want the dependencies of a package and also some
6609 additional packages that are not build-time or runtime dependencies, but
6610 are useful when developing nonetheless. Because of this, the
6611 @code{--ad-hoc} flag is positional. Packages appearing before
6612 @code{--ad-hoc} are interpreted as packages whose dependencies will be
6613 added to the environment. Packages appearing after are interpreted as
6614 packages that will be added to the environment directly. For example,
6615 the following command creates a Guix development environment that
6616 additionally includes Git and strace:
6619 guix environment guix --ad-hoc git strace
6622 Sometimes it is desirable to isolate the environment as much as
6623 possible, for maximal purity and reproducibility. In particular, when
6624 using Guix on a host distro that is not GuixSD, it is desirable to
6625 prevent access to @file{/usr/bin} and other system-wide resources from
6626 the development environment. For example, the following command spawns
6627 a Guile REPL in a ``container'' where only the store and the current
6628 working directory are mounted:
6631 guix environment --ad-hoc --container guile -- guile
6635 The @code{--container} option requires Linux-libre 3.19 or newer.
6638 The available options are summarized below.
6641 @item --root=@var{file}
6642 @itemx -r @var{file}
6643 @cindex persistent environment
6644 @cindex garbage collector root, for environments
6645 Make @var{file} a symlink to the profile for this environment, and
6646 register it as a garbage collector root.
6648 This is useful if you want to protect your environment from garbage
6649 collection, to make it ``persistent''.
6651 When this option is omitted, the environment is protected from garbage
6652 collection only for the duration of the @command{guix environment}
6653 session. This means that next time you recreate the same environment,
6654 you could have to rebuild or re-download packages.
6656 @item --expression=@var{expr}
6657 @itemx -e @var{expr}
6658 Create an environment for the package or list of packages that
6659 @var{expr} evaluates to.
6661 For example, running:
6664 guix environment -e '(@@ (gnu packages maths) petsc-openmpi)'
6667 starts a shell with the environment for this specific variant of the
6673 guix environment --ad-hoc -e '(@@ (gnu) %base-packages)'
6676 starts a shell with all the GuixSD base packages available.
6678 The above commands only use the default output of the given packages.
6679 To select other outputs, two element tuples can be specified:
6682 guix environment --ad-hoc -e '(list (@ (gnu packages bash) bash) "include")'
6685 @item --load=@var{file}
6686 @itemx -l @var{file}
6687 Create an environment for the package or list of packages that the code
6688 within @var{file} evaluates to.
6690 As an example, @var{file} might contain a definition like this
6691 (@pxref{Defining Packages}):
6694 @verbatiminclude environment-gdb.scm
6698 Include all specified packages in the resulting environment, as if an
6699 @i{ad hoc} package were defined with them as inputs. This option is
6700 useful for quickly creating an environment without having to write a
6701 package expression to contain the desired inputs.
6703 For instance, the command:
6706 guix environment --ad-hoc guile guile-sdl -- guile
6709 runs @command{guile} in an environment where Guile and Guile-SDL are
6712 Note that this example implicitly asks for the default output of
6713 @code{guile} and @code{guile-sdl}, but it is possible to ask for a
6714 specific output---e.g., @code{glib:bin} asks for the @code{bin} output
6715 of @code{glib} (@pxref{Packages with Multiple Outputs}).
6717 This option may be composed with the default behavior of @command{guix
6718 environment}. Packages appearing before @code{--ad-hoc} are interpreted
6719 as packages whose dependencies will be added to the environment, the
6720 default behavior. Packages appearing after are interpreted as packages
6721 that will be added to the environment directly.
6724 Unset existing environment variables when building the new environment.
6725 This has the effect of creating an environment in which search paths
6726 only contain package inputs.
6728 @item --search-paths
6729 Display the environment variable definitions that make up the
6732 @item --system=@var{system}
6733 @itemx -s @var{system}
6734 Attempt to build for @var{system}---e.g., @code{i686-linux}.
6739 Run @var{command} within an isolated container. The current working
6740 directory outside the container is mapped inside the container.
6741 Additionally, a dummy home directory is created that matches the current
6742 user's home directory, and @file{/etc/passwd} is configured accordingly.
6743 The spawned process runs as the current user outside the container, but
6744 has root privileges in the context of the container.
6748 For containers, share the network namespace with the host system.
6749 Containers created without this flag only have access to the loopback
6752 @item --expose=@var{source}[=@var{target}]
6753 For containers, expose the file system @var{source} from the host system
6754 as the read-only file system @var{target} within the container. If
6755 @var{target} is not specified, @var{source} is used as the target mount
6756 point in the container.
6758 The example below spawns a Guile REPL in a container in which the user's
6759 home directory is accessible read-only via the @file{/exchange}
6763 guix environment --container --expose=$HOME=/exchange --ad-hoc guile -- guile
6766 @item --share=@var{source}[=@var{target}]
6767 For containers, share the file system @var{source} from the host system
6768 as the writable file system @var{target} within the container. If
6769 @var{target} is not specified, @var{source} is used as the target mount
6770 point in the container.
6772 The example below spawns a Guile REPL in a container in which the user's
6773 home directory is accessible for both reading and writing via the
6774 @file{/exchange} directory:
6777 guix environment --container --share=$HOME=/exchange --ad-hoc guile -- guile
6781 @command{guix environment}
6782 also supports all of the common build options that @command{guix
6783 build} supports (@pxref{Common Build Options}).
6786 @node Invoking guix publish
6787 @section Invoking @command{guix publish}
6789 @cindex @command{guix publish}
6790 The purpose of @command{guix publish} is to enable users to easily share
6791 their store with others, who can then use it as a substitute server
6792 (@pxref{Substitutes}).
6794 When @command{guix publish} runs, it spawns an HTTP server which allows
6795 anyone with network access to obtain substitutes from it. This means
6796 that any machine running Guix can also act as if it were a build farm,
6797 since the HTTP interface is compatible with Hydra, the software behind
6798 the @code{hydra.gnu.org} build farm.
6800 For security, each substitute is signed, allowing recipients to check
6801 their authenticity and integrity (@pxref{Substitutes}). Because
6802 @command{guix publish} uses the signing key of the system, which is only
6803 readable by the system administrator, it must be started as root; the
6804 @code{--user} option makes it drop root privileges early on.
6806 The signing key pair must be generated before @command{guix publish} is
6807 launched, using @command{guix archive --generate-key} (@pxref{Invoking
6810 The general syntax is:
6813 guix publish @var{options}@dots{}
6816 Running @command{guix publish} without any additional arguments will
6817 spawn an HTTP server on port 8080:
6823 Once a publishing server has been authorized (@pxref{Invoking guix
6824 archive}), the daemon may download substitutes from it:
6827 guix-daemon --substitute-urls=http://example.org:8080
6830 By default, @command{guix publish} compresses archives on the fly as it
6831 serves them. This ``on-the-fly'' mode is convenient in that it requires
6832 no setup and is immediately available. However, when serving lots of
6833 clients, we recommend using the @option{--cache} option, which enables
6834 caching of the archives before they are sent to clients---see below for
6837 As a bonus, @command{guix publish} also serves as a content-addressed
6838 mirror for source files referenced in @code{origin} records
6839 (@pxref{origin Reference}). For instance, assuming @command{guix
6840 publish} is running on @code{example.org}, the following URL returns the
6841 raw @file{hello-2.10.tar.gz} file with the given SHA256 hash
6842 (represented in @code{nix-base32} format, @pxref{Invoking guix hash}):
6845 http://example.org/file/hello-2.10.tar.gz/sha256/0ssi1@dots{}ndq1i
6848 Obviously, these URLs only work for files that are in the store; in
6849 other cases, they return 404 (``Not Found'').
6851 The following options are available:
6854 @item --port=@var{port}
6855 @itemx -p @var{port}
6856 Listen for HTTP requests on @var{port}.
6858 @item --listen=@var{host}
6859 Listen on the network interface for @var{host}. The default is to
6860 accept connections from any interface.
6862 @item --user=@var{user}
6863 @itemx -u @var{user}
6864 Change privileges to @var{user} as soon as possible---i.e., once the
6865 server socket is open and the signing key has been read.
6867 @item --compression[=@var{level}]
6868 @itemx -C [@var{level}]
6869 Compress data using the given @var{level}. When @var{level} is zero,
6870 disable compression. The range 1 to 9 corresponds to different gzip
6871 compression levels: 1 is the fastest, and 9 is the best (CPU-intensive).
6874 Unless @option{--cache} is used, compression occurs on the fly and
6875 the compressed streams are not
6876 cached. Thus, to reduce load on the machine that runs @command{guix
6877 publish}, it may be a good idea to choose a low compression level, to
6878 run @command{guix publish} behind a caching proxy, or to use
6879 @option{--cache}. Using @option{--cache} has the advantage that it
6880 allows @command{guix publish} to add @code{Content-Length} HTTP header
6883 @item --cache=@var{directory}
6884 @itemx -c @var{directory}
6885 Cache archives and meta-data (@code{.narinfo} URLs) to @var{directory}
6886 and only serve archives that are in cache.
6888 When this option is omitted, archives and meta-data are created
6889 on-the-fly. This can reduce the available bandwidth, especially when
6890 compression is enabled, since this may become CPU-bound. Another
6891 drawback of the default mode is that the length of archives is not known
6892 in advance, so @command{guix publish} does not add a
6893 @code{Content-Length} HTTP header to its responses, which in turn
6894 prevents clients from knowing the amount of data being downloaded.
6896 Conversely, when @option{--cache} is used, the first request for a store
6897 item (@i{via} a @code{.narinfo} URL) returns 404 and triggers a
6898 background process to @dfn{bake} the archive---computing its
6899 @code{.narinfo} and compressing the archive, if needed. Once the
6900 archive is cached in @var{directory}, subsequent requests succeed and
6901 are served directly from the cache, which guarantees that clients get
6902 the best possible bandwidth.
6904 The ``baking'' process is performed by worker threads. By default, one
6905 thread per CPU core is created, but this can be customized. See
6906 @option{--workers} below.
6908 When @option{--ttl} is used, cached entries are automatically deleted
6909 when they have expired.
6911 @item --workers=@var{N}
6912 When @option{--cache} is used, request the allocation of @var{N} worker
6913 threads to ``bake'' archives.
6915 @item --ttl=@var{ttl}
6916 Produce @code{Cache-Control} HTTP headers that advertise a time-to-live
6917 (TTL) of @var{ttl}. @var{ttl} must denote a duration: @code{5d} means 5
6918 days, @code{1m} means 1 month, and so on.
6920 This allows the user's Guix to keep substitute information in cache for
6921 @var{ttl}. However, note that @code{guix publish} does not itself
6922 guarantee that the store items it provides will indeed remain available
6923 for as long as @var{ttl}.
6925 Additionally, when @option{--cache} is used, cached entries that have
6926 not been accessed for @var{ttl} may be deleted.
6928 @item --nar-path=@var{path}
6929 Use @var{path} as the prefix for the URLs of ``nar'' files
6930 (@pxref{Invoking guix archive, normalized archives}).
6932 By default, nars are served at a URL such as
6933 @code{/nar/gzip/@dots{}-coreutils-8.25}. This option allows you to
6934 change the @code{/nar} part to @var{path}.
6936 @item --public-key=@var{file}
6937 @itemx --private-key=@var{file}
6938 Use the specific @var{file}s as the public/private key pair used to sign
6939 the store items being published.
6941 The files must correspond to the same key pair (the private key is used
6942 for signing and the public key is merely advertised in the signature
6943 metadata). They must contain keys in the canonical s-expression format
6944 as produced by @command{guix archive --generate-key} (@pxref{Invoking
6945 guix archive}). By default, @file{/etc/guix/signing-key.pub} and
6946 @file{/etc/guix/signing-key.sec} are used.
6948 @item --repl[=@var{port}]
6949 @itemx -r [@var{port}]
6950 Spawn a Guile REPL server (@pxref{REPL Servers,,, guile, GNU Guile
6951 Reference Manual}) on @var{port} (37146 by default). This is used
6952 primarily for debugging a running @command{guix publish} server.
6955 Enabling @command{guix publish} on a GuixSD system is a one-liner: just
6956 instantiate a @code{guix-publish-service-type} service in the @code{services} field
6957 of the @code{operating-system} declaration (@pxref{guix-publish-service-type,
6958 @code{guix-publish-service-type}}).
6960 If you are instead running Guix on a ``foreign distro'', follow these
6965 If your host distro uses the systemd init system:
6968 # ln -s ~root/.guix-profile/lib/systemd/system/guix-publish.service \
6969 /etc/systemd/system/
6970 # systemctl start guix-publish && systemctl enable guix-publish
6974 If your host distro uses the Upstart init system:
6977 # ln -s ~root/.guix-profile/lib/upstart/system/guix-publish.conf /etc/init/
6978 # start guix-publish
6982 Otherwise, proceed similarly with your distro's init system.
6985 @node Invoking guix challenge
6986 @section Invoking @command{guix challenge}
6988 @cindex reproducible builds
6989 @cindex verifiable builds
6990 @cindex @command{guix challenge}
6992 Do the binaries provided by this server really correspond to the source
6993 code it claims to build? Is a package build process deterministic?
6994 These are the questions the @command{guix challenge} command attempts to
6997 The former is obviously an important question: Before using a substitute
6998 server (@pxref{Substitutes}), one had better @emph{verify} that it
6999 provides the right binaries, and thus @emph{challenge} it. The latter
7000 is what enables the former: If package builds are deterministic, then
7001 independent builds of the package should yield the exact same result,
7002 bit for bit; if a server provides a binary different from the one
7003 obtained locally, it may be either corrupt or malicious.
7005 We know that the hash that shows up in @file{/gnu/store} file names is
7006 the hash of all the inputs of the process that built the file or
7007 directory---compilers, libraries, build scripts,
7008 etc. (@pxref{Introduction}). Assuming deterministic build processes,
7009 one store file name should map to exactly one build output.
7010 @command{guix challenge} checks whether there is, indeed, a single
7011 mapping by comparing the build outputs of several independent builds of
7012 any given store item.
7014 The command output looks like this:
7017 $ guix challenge --substitute-urls="https://hydra.gnu.org https://guix.example.org"
7018 updating list of substitutes from 'https://hydra.gnu.org'... 100.0%
7019 updating list of substitutes from 'https://guix.example.org'... 100.0%
7020 /gnu/store/@dots{}-openssl-1.0.2d contents differ:
7021 local hash: 0725l22r5jnzazaacncwsvp9kgf42266ayyp814v7djxs7nk963q
7022 https://hydra.gnu.org/nar/@dots{}-openssl-1.0.2d: 0725l22r5jnzazaacncwsvp9kgf42266ayyp814v7djxs7nk963q
7023 https://guix.example.org/nar/@dots{}-openssl-1.0.2d: 1zy4fmaaqcnjrzzajkdn3f5gmjk754b43qkq47llbyak9z0qjyim
7024 /gnu/store/@dots{}-git-2.5.0 contents differ:
7025 local hash: 00p3bmryhjxrhpn2gxs2fy0a15lnip05l97205pgbk5ra395hyha
7026 https://hydra.gnu.org/nar/@dots{}-git-2.5.0: 069nb85bv4d4a6slrwjdy8v1cn4cwspm3kdbmyb81d6zckj3nq9f
7027 https://guix.example.org/nar/@dots{}-git-2.5.0: 0mdqa9w1p6cmli6976v4wi0sw9r4p5prkj7lzfd1877wk11c9c73
7028 /gnu/store/@dots{}-pius-2.1.1 contents differ:
7029 local hash: 0k4v3m9z1zp8xzzizb7d8kjj72f9172xv078sq4wl73vnq9ig3ax
7030 https://hydra.gnu.org/nar/@dots{}-pius-2.1.1: 0k4v3m9z1zp8xzzizb7d8kjj72f9172xv078sq4wl73vnq9ig3ax
7031 https://guix.example.org/nar/@dots{}-pius-2.1.1: 1cy25x1a4fzq5rk0pmvc8xhwyffnqz95h2bpvqsz2mpvlbccy0gs
7035 In this example, @command{guix challenge} first scans the store to
7036 determine the set of locally-built derivations---as opposed to store
7037 items that were downloaded from a substitute server---and then queries
7038 all the substitute servers. It then reports those store items for which
7039 the servers obtained a result different from the local build.
7041 @cindex non-determinism, in package builds
7042 As an example, @code{guix.example.org} always gets a different answer.
7043 Conversely, @code{hydra.gnu.org} agrees with local builds, except in the
7044 case of Git. This might indicate that the build process of Git is
7045 non-deterministic, meaning that its output varies as a function of
7046 various things that Guix does not fully control, in spite of building
7047 packages in isolated environments (@pxref{Features}). Most common
7048 sources of non-determinism include the addition of timestamps in build
7049 results, the inclusion of random numbers, and directory listings sorted
7050 by inode number. See @uref{https://reproducible-builds.org/docs/}, for
7053 To find out what is wrong with this Git binary, we can do something along
7054 these lines (@pxref{Invoking guix archive}):
7057 $ wget -q -O - https://hydra.gnu.org/nar/@dots{}-git-2.5.0 \
7058 | guix archive -x /tmp/git
7059 $ diff -ur --no-dereference /gnu/store/@dots{}-git.2.5.0 /tmp/git
7062 This command shows the difference between the files resulting from the
7063 local build, and the files resulting from the build on
7064 @code{hydra.gnu.org} (@pxref{Overview, Comparing and Merging Files,,
7065 diffutils, Comparing and Merging Files}). The @command{diff} command
7066 works great for text files. When binary files differ, a better option
7067 is @uref{https://diffoscope.org/, Diffoscope}, a tool that helps
7068 visualize differences for all kinds of files.
7070 Once you have done that work, you can tell whether the differences are due
7071 to a non-deterministic build process or to a malicious server. We try
7072 hard to remove sources of non-determinism in packages to make it easier
7073 to verify substitutes, but of course, this is a process that
7074 involves not just Guix, but a large part of the free software community.
7075 In the meantime, @command{guix challenge} is one tool to help address
7078 If you are writing packages for Guix, you are encouraged to check
7079 whether @code{hydra.gnu.org} and other substitute servers obtain the
7080 same build result as you did with:
7083 $ guix challenge @var{package}
7087 where @var{package} is a package specification such as
7088 @code{guile@@2.0} or @code{glibc:debug}.
7090 The general syntax is:
7093 guix challenge @var{options} [@var{packages}@dots{}]
7096 When a difference is found between the hash of a locally-built item and
7097 that of a server-provided substitute, or among substitutes provided by
7098 different servers, the command displays it as in the example above and
7099 its exit code is 2 (other non-zero exit codes denote other kinds of
7102 The one option that matters is:
7106 @item --substitute-urls=@var{urls}
7107 Consider @var{urls} the whitespace-separated list of substitute source
7112 Show details about matches (identical contents) in addition to
7113 information about mismatches.
7117 @node Invoking guix copy
7118 @section Invoking @command{guix copy}
7120 @cindex copy, of store items, over SSH
7121 @cindex SSH, copy of store items
7122 @cindex sharing store items across machines
7123 @cindex transferring store items across machines
7124 The @command{guix copy} command copies items from the store of one
7125 machine to that of another machine over a secure shell (SSH)
7126 connection@footnote{This command is available only when Guile-SSH was
7127 found. @xref{Requirements}, for details.}. For example, the following
7128 command copies the @code{coreutils} package, the user's profile, and all
7129 their dependencies over to @var{host}, logged in as @var{user}:
7132 guix copy --to=@var{user}@@@var{host} \
7133 coreutils `readlink -f ~/.guix-profile`
7136 If some of the items to be copied are already present on @var{host},
7137 they are not actually sent.
7139 The command below retrieves @code{libreoffice} and @code{gimp} from
7140 @var{host}, assuming they are available there:
7143 guix copy --from=@var{host} libreoffice gimp
7146 The SSH connection is established using the Guile-SSH client, which is
7147 compatible with OpenSSH: it honors @file{~/.ssh/known_hosts} and
7148 @file{~/.ssh/config}, and uses the SSH agent for authentication.
7150 The key used to sign items that are sent must be accepted by the remote
7151 machine. Likewise, the key used by the remote machine to sign items you
7152 are retrieving must be in @file{/etc/guix/acl} so it is accepted by your
7153 own daemon. @xref{Invoking guix archive}, for more information about
7154 store item authentication.
7156 The general syntax is:
7159 guix copy [--to=@var{spec}|--from=@var{spec}] @var{items}@dots{}
7162 You must always specify one of the following options:
7165 @item --to=@var{spec}
7166 @itemx --from=@var{spec}
7167 Specify the host to send to or receive from. @var{spec} must be an SSH
7168 spec such as @code{example.org}, @code{charlie@@example.org}, or
7169 @code{charlie@@example.org:2222}.
7172 The @var{items} can be either package names, such as @code{gimp}, or
7173 store items, such as @file{/gnu/store/@dots{}-idutils-4.6}.
7175 When specifying the name of a package to send, it is first built if
7176 needed, unless @option{--dry-run} was specified. Common build options
7177 are supported (@pxref{Common Build Options}).
7180 @node Invoking guix container
7181 @section Invoking @command{guix container}
7183 @cindex @command{guix container}
7185 As of version @value{VERSION}, this tool is experimental. The interface
7186 is subject to radical change in the future.
7189 The purpose of @command{guix container} is to manipulate processes
7190 running within an isolated environment, commonly known as a
7191 ``container'', typically created by the @command{guix environment}
7192 (@pxref{Invoking guix environment}) and @command{guix system container}
7193 (@pxref{Invoking guix system}) commands.
7195 The general syntax is:
7198 guix container @var{action} @var{options}@dots{}
7201 @var{action} specifies the operation to perform with a container, and
7202 @var{options} specifies the context-specific arguments for the action.
7204 The following actions are available:
7208 Execute a command within the context of a running container.
7213 guix container exec @var{pid} @var{program} @var{arguments}@dots{}
7216 @var{pid} specifies the process ID of the running container.
7217 @var{program} specifies an executable file name within the root file
7218 system of the container. @var{arguments} are the additional options that
7219 will be passed to @var{program}.
7221 The following command launches an interactive login shell inside a
7222 GuixSD container, started by @command{guix system container}, and whose
7226 guix container exec 9001 /run/current-system/profile/bin/bash --login
7229 Note that the @var{pid} cannot be the parent process of a container. It
7230 must be PID 1 of the container or one of its child processes.
7234 @c *********************************************************************
7235 @node GNU Distribution
7236 @chapter GNU Distribution
7238 @cindex Guix System Distribution
7240 Guix comes with a distribution of the GNU system consisting entirely of
7241 free software@footnote{The term ``free'' here refers to the
7242 @url{http://www.gnu.org/philosophy/free-sw.html,freedom provided to
7243 users of that software}.}. The
7244 distribution can be installed on its own (@pxref{System Installation}),
7245 but it is also possible to install Guix as a package manager on top of
7246 an installed GNU/Linux system (@pxref{Installation}). To distinguish
7247 between the two, we refer to the standalone distribution as the Guix
7248 System Distribution, or GuixSD.
7250 The distribution provides core GNU packages such as GNU libc, GCC, and
7251 Binutils, as well as many GNU and non-GNU applications. The complete
7252 list of available packages can be browsed
7253 @url{http://www.gnu.org/software/guix/packages,on-line} or by
7254 running @command{guix package} (@pxref{Invoking guix package}):
7257 guix package --list-available
7260 Our goal is to provide a practical 100% free software distribution of
7261 Linux-based and other variants of GNU, with a focus on the promotion and
7262 tight integration of GNU components, and an emphasis on programs and
7263 tools that help users exert that freedom.
7265 Packages are currently available on the following platforms:
7270 Intel/AMD @code{x86_64} architecture, Linux-Libre kernel;
7273 Intel 32-bit architecture (IA32), Linux-Libre kernel;
7276 ARMv7-A architecture with hard float, Thumb-2 and NEON,
7277 using the EABI hard-float application binary interface (ABI),
7278 and Linux-Libre kernel.
7281 little-endian 64-bit ARMv8-A processors, Linux-Libre kernel. This is
7282 currently in an experimental stage, with limited support.
7283 @xref{Contributing}, for how to help!
7285 @item mips64el-linux
7286 little-endian 64-bit MIPS processors, specifically the Loongson series,
7287 n32 ABI, and Linux-Libre kernel.
7291 GuixSD itself is currently only available on @code{i686} and @code{x86_64}.
7294 For information on porting to other architectures or kernels,
7298 * System Installation:: Installing the whole operating system.
7299 * System Configuration:: Configuring the operating system.
7300 * Documentation:: Browsing software user manuals.
7301 * Installing Debugging Files:: Feeding the debugger.
7302 * Security Updates:: Deploying security fixes quickly.
7303 * Package Modules:: Packages from the programmer's viewpoint.
7304 * Packaging Guidelines:: Growing the distribution.
7305 * Bootstrapping:: GNU/Linux built from scratch.
7306 * Porting:: Targeting another platform or kernel.
7309 Building this distribution is a cooperative effort, and you are invited
7310 to join! @xref{Contributing}, for information about how you can help.
7312 @node System Installation
7313 @section System Installation
7315 @cindex installing GuixSD
7316 @cindex Guix System Distribution
7317 This section explains how to install the Guix System Distribution (GuixSD)
7318 on a machine. The Guix package manager can
7319 also be installed on top of a running GNU/Linux system,
7320 @pxref{Installation}.
7324 @c This paragraph is for people reading this from tty2 of the
7325 @c installation image.
7326 You are reading this documentation with an Info reader. For details on
7327 how to use it, hit the @key{RET} key (``return'' or ``enter'') on the
7328 link that follows: @pxref{Top, Info reader,, info-stnd, Stand-alone GNU
7329 Info}. Hit @kbd{l} afterwards to come back here.
7331 Alternately, run @command{info info} in another tty to keep the manual
7337 * Limitations:: What you can expect.
7338 * Hardware Considerations:: Supported hardware.
7339 * USB Stick Installation:: Preparing the installation medium.
7340 * Preparing for Installation:: Networking, partitioning, etc.
7341 * Proceeding with the Installation:: The real thing.
7342 * Installing GuixSD in a VM:: GuixSD playground.
7343 * Building the Installation Image:: How this comes to be.
7347 @subsection Limitations
7349 As of version @value{VERSION}, the Guix System Distribution (GuixSD) is
7350 not production-ready. It may contain bugs and lack important
7351 features. Thus, if you are looking for a stable production system that
7352 respects your freedom as a computer user, a good solution at this point
7353 is to consider @url{http://www.gnu.org/distros/free-distros.html, one of
7354 the more established GNU/Linux distributions}. We hope you can soon switch
7355 to the GuixSD without fear, of course. In the meantime, you can
7356 also keep using your distribution and try out the package manager on top
7357 of it (@pxref{Installation}).
7359 Before you proceed with the installation, be aware of the following
7360 noteworthy limitations applicable to version @value{VERSION}:
7364 The installation process does not include a graphical user interface and
7365 requires familiarity with GNU/Linux (see the following subsections to
7366 get a feel of what that means.)
7369 Support for the Logical Volume Manager (LVM) is missing.
7372 More and more system services are provided (@pxref{Services}), but some
7376 More than 5,300 packages are available, but you may
7377 occasionally find that a useful package is missing.
7380 GNOME, Xfce, LXDE, and Enlightenment are available (@pxref{Desktop Services}),
7381 as well as a number of X11 window managers. However, some graphical
7382 applications may be missing, as well as KDE.
7385 You have been warned! But more than a disclaimer, this is an invitation
7386 to report issues (and success stories!), and to join us in improving it.
7387 @xref{Contributing}, for more info.
7390 @node Hardware Considerations
7391 @subsection Hardware Considerations
7393 @cindex hardware support on GuixSD
7394 GNU@tie{}GuixSD focuses on respecting the user's computing freedom. It
7395 builds around the kernel Linux-libre, which means that only hardware for
7396 which free software drivers and firmware exist is supported. Nowadays,
7397 a wide range of off-the-shelf hardware is supported on
7398 GNU/Linux-libre---from keyboards to graphics cards to scanners and
7399 Ethernet controllers. Unfortunately, there are still areas where
7400 hardware vendors deny users control over their own computing, and such
7401 hardware is not supported on GuixSD.
7403 @cindex WiFi, hardware support
7404 One of the main areas where free drivers or firmware are lacking is WiFi
7405 devices. WiFi devices known to work include those using Atheros chips
7406 (AR9271 and AR7010), which corresponds to the @code{ath9k} Linux-libre
7407 driver, and those using Broadcom/AirForce chips (BCM43xx with
7408 Wireless-Core Revision 5), which corresponds to the @code{b43-open}
7409 Linux-libre driver. Free firmware exists for both and is available
7410 out-of-the-box on GuixSD, as part of @var{%base-firmware}
7411 (@pxref{operating-system Reference, @code{firmware}}).
7413 @cindex RYF, Respects Your Freedom
7414 The @uref{https://www.fsf.org/, Free Software Foundation} runs
7415 @uref{https://www.fsf.org/ryf, @dfn{Respects Your Freedom}} (RYF), a
7416 certification program for hardware products that respect your freedom
7417 and your privacy and ensure that you have control over your device. We
7418 encourage you to check the list of RYF-certified devices.
7420 Another useful resource is the @uref{https://www.h-node.org/, H-Node}
7421 web site. It contains a catalog of hardware devices with information
7422 about their support in GNU/Linux.
7425 @node USB Stick Installation
7426 @subsection USB Stick Installation
7428 An installation image for USB sticks can be downloaded from
7429 @indicateurl{ftp://alpha.gnu.org/gnu/guix/guixsd-usb-install-@value{VERSION}.@var{system}.xz},
7430 where @var{system} is one of:
7434 for a GNU/Linux system on Intel/AMD-compatible 64-bit CPUs;
7437 for a 32-bit GNU/Linux system on Intel-compatible CPUs.
7440 @c start duplication of authentication part from ``Binary Installation''
7441 Make sure to download the associated @file{.sig} file and to verify the
7442 authenticity of the image against it, along these lines:
7445 $ wget ftp://alpha.gnu.org/gnu/guix/guixsd-usb-install-@value{VERSION}.@var{system}.xz.sig
7446 $ gpg --verify guixsd-usb-install-@value{VERSION}.@var{system}.xz.sig
7449 If that command fails because you do not have the required public key,
7450 then run this command to import it:
7453 $ gpg --keyserver pgp.mit.edu --recv-keys @value{OPENPGP-SIGNING-KEY-ID}
7457 and rerun the @code{gpg --verify} command.
7460 This image contains a single partition with the tools necessary for an
7461 installation. It is meant to be copied @emph{as is} to a large-enough
7464 To copy the image to a USB stick, follow these steps:
7468 Decompress the image using the @command{xz} command:
7471 xz -d guixsd-usb-install-@value{VERSION}.@var{system}.xz
7475 Insert a USB stick of 1@tie{}GiB or more into your machine, and determine
7476 its device name. Assuming that the USB stick is known as @file{/dev/sdX},
7477 copy the image with:
7480 dd if=guixsd-usb-install-@value{VERSION}.x86_64 of=/dev/sdX
7484 Access to @file{/dev/sdX} usually requires root privileges.
7487 Once this is done, you should be able to reboot the system and boot from
7488 the USB stick. The latter usually requires you to get in the BIOS or
7489 UEFI boot menu, where you can choose to boot from the USB stick.
7491 @xref{Installing GuixSD in a VM}, if, instead, you would like to install
7492 GuixSD in a virtual machine (VM).
7494 @node Preparing for Installation
7495 @subsection Preparing for Installation
7497 Once you have successfully booted the image on the USB stick, you should
7498 end up with a root prompt. Several console TTYs are configured and can
7499 be used to run commands as root. TTY2 shows this documentation,
7500 browsable using the Info reader commands (@pxref{Top,,, info-stnd,
7501 Stand-alone GNU Info}). The installation system runs the GPM mouse
7502 daemon, which allows you to select text with the left mouse button and
7503 to paste it with the middle button.
7506 Installation requires access to the Internet so that any missing
7507 dependencies of your system configuration can be downloaded. See the
7508 ``Networking'' section below.
7511 The installation system includes many common tools needed for this task.
7512 But it is also a full-blown GuixSD system, which means that you can
7513 install additional packages, should you need it, using @command{guix
7514 package} (@pxref{Invoking guix package}).
7516 @subsubsection Keyboard Layout
7518 @cindex keyboard layout
7519 The installation image uses the US qwerty keyboard layout. If you want
7520 to change it, you can use the @command{loadkeys} command. For example,
7521 the following command selects the Dvorak keyboard layout:
7527 See the files under @file{/run/current-system/profile/share/keymaps} for
7528 a list of available keyboard layouts. Run @command{man loadkeys} for
7531 @subsubsection Networking
7533 Run the following command see what your network interfaces are called:
7540 @dots{} or, using the GNU/Linux-specific @command{ip} command:
7546 @c http://cgit.freedesktop.org/systemd/systemd/tree/src/udev/udev-builtin-net_id.c#n20
7547 Wired interfaces have a name starting with @samp{e}; for example, the
7548 interface corresponding to the first on-board Ethernet controller is
7549 called @samp{eno1}. Wireless interfaces have a name starting with
7550 @samp{w}, like @samp{w1p2s0}.
7553 @item Wired connection
7554 To configure a wired network run the following command, substituting
7555 @var{interface} with the name of the wired interface you want to use.
7558 ifconfig @var{interface} up
7561 @item Wireless connection
7564 To configure wireless networking, you can create a configuration file
7565 for the @command{wpa_supplicant} configuration tool (its location is not
7566 important) using one of the available text editors such as
7570 zile wpa_supplicant.conf
7573 As an example, the following stanza can go to this file and will work
7574 for many wireless networks, provided you give the actual SSID and
7575 passphrase for the network you are connecting to:
7579 ssid="@var{my-ssid}"
7581 psk="the network's secret passphrase"
7585 Start the wireless service and run it in the background with the
7586 following command (substitute @var{interface} with the name of the
7587 network interface you want to use):
7590 wpa_supplicant -c wpa_supplicant.conf -i @var{interface} -B
7593 Run @command{man wpa_supplicant} for more information.
7597 At this point, you need to acquire an IP address. On a network where IP
7598 addresses are automatically assigned @i{via} DHCP, you can run:
7601 dhclient -v @var{interface}
7604 Try to ping a server to see if networking is up and running:
7610 Setting up network access is almost always a requirement because the
7611 image does not contain all the software and tools that may be needed.
7613 @cindex installing over SSH
7614 If you want to, you can continue the installation remotely by starting
7618 herd start ssh-daemon
7621 Make sure to either set a password with @command{passwd}, or configure
7622 OpenSSH public key authentication before logging in.
7624 @subsubsection Disk Partitioning
7626 Unless this has already been done, the next step is to partition, and
7627 then format the target partition(s).
7629 The installation image includes several partitioning tools, including
7630 Parted (@pxref{Overview,,, parted, GNU Parted User Manual}),
7631 @command{fdisk}, and @command{cfdisk}. Run it and set up your disk with
7632 the partition layout you want:
7638 If your disk uses the GUID Partition Table (GPT) format and you plan to
7639 install BIOS-based GRUB (which is the default), make sure a BIOS Boot
7640 Partition is available (@pxref{BIOS installation,,, grub, GNU GRUB
7643 @cindex EFI, installation
7644 @cindex UEFI, installation
7645 @cindex ESP, EFI system partition
7646 If you instead wish to use EFI-based GRUB, a FAT32 @dfn{EFI System Partition}
7647 (ESP) is required. This partition should be mounted at @file{/boot/efi} and
7648 must have the @code{esp} flag set. E.g., for @command{parted}:
7651 parted /dev/sda set 1 esp on
7654 Once you are done partitioning the target hard disk drive, you have to
7655 create a file system on the relevant partition(s)@footnote{Currently
7656 GuixSD only supports ext4 and btrfs file systems. In particular, code
7657 that reads partition UUIDs and labels only works for these file system
7660 Preferably, assign partitions a label so that you can easily and
7661 reliably refer to them in @code{file-system} declarations (@pxref{File
7662 Systems}). This is typically done using the @code{-L} option of
7663 @command{mkfs.ext4} and related commands. So, assuming the target root
7664 partition lives at @file{/dev/sda1}, a file system with the label
7665 @code{my-root} can be created with:
7668 mkfs.ext4 -L my-root /dev/sda1
7671 @cindex encrypted disk
7672 If you are instead planning to encrypt the root partition, you can use
7673 the Cryptsetup/LUKS utilities to do that (see @inlinefmtifelse{html,
7674 @uref{https://linux.die.net/man/8/cryptsetup, @code{man cryptsetup}},
7675 @code{man cryptsetup}} for more information.) Assuming you want to
7676 store the root partition on @file{/dev/sda1}, the command sequence would
7677 be along these lines:
7680 cryptsetup luksFormat /dev/sda1
7681 cryptsetup open --type luks /dev/sda1 my-partition
7682 mkfs.ext4 -L my-root /dev/mapper/my-partition
7685 Once that is done, mount the target root partition under @file{/mnt}
7686 with a command like (again, assuming @code{my-root} is the label of the
7690 mount LABEL=my-root /mnt
7693 Also mount any other partitions you would like to use on the target
7694 system relative to this path. If you have @file{/boot} on a separate
7695 partition for example, mount it at @file{/mnt/boot} now so it is found
7696 by @code{guix system init} afterwards.
7698 Finally, if you plan to use one or more swap partitions (@pxref{Memory
7699 Concepts, swap space,, libc, The GNU C Library Reference Manual}), make
7700 sure to initialize them with @command{mkswap}. Assuming you have one
7701 swap partition on @file{/dev/sda2}, you would run:
7708 Alternatively, you may use a swap file. For example, assuming that in
7709 the new system you want to use the file @file{/swapfile} as a swap file,
7710 you would run@footnote{This example will work for many types of file
7711 systems (e.g., ext4). However, for copy-on-write file systems (e.g.,
7712 btrfs), the required steps may be different. For details, see the
7713 manual pages for @command{mkswap} and @command{swapon}.}:
7716 # This is 10 GiB of swap space. Adjust "count" to change the size.
7717 dd if=/dev/zero of=/mnt/swapfile bs=1MiB count=10240
7718 # For security, make the file readable and writable only by root.
7719 chmod 600 /mnt/swapfile
7720 mkswap /mnt/swapfile
7721 swapon /mnt/swapfile
7724 Note that if you have encrypted the root partition and created a swap
7725 file in its file system as described above, then the encryption also
7726 protects the swap file, just like any other file in that file system.
7728 @node Proceeding with the Installation
7729 @subsection Proceeding with the Installation
7731 With the target partitions ready and the target root mounted on
7732 @file{/mnt}, we're ready to go. First, run:
7735 herd start cow-store /mnt
7738 This makes @file{/gnu/store} copy-on-write, such that packages added to it
7739 during the installation phase are written to the target disk on @file{/mnt}
7740 rather than kept in memory. This is necessary because the first phase of
7741 the @command{guix system init} command (see below) entails downloads or
7742 builds to @file{/gnu/store} which, initially, is an in-memory file system.
7744 Next, you have to edit a file and
7745 provide the declaration of the operating system to be installed. To
7746 that end, the installation system comes with three text editors: GNU nano
7747 (@pxref{Top,,, nano, GNU nano Manual}), GNU Zile (an Emacs clone), and
7748 nvi (a clone of the original BSD @command{vi} editor).
7749 We strongly recommend storing that file on the target root file system, say,
7750 as @file{/mnt/etc/config.scm}. Failing to do that, you will have lost your
7751 configuration file once you have rebooted into the newly-installed system.
7753 @xref{Using the Configuration System}, for an overview of the
7754 configuration file. The example configurations discussed in that
7755 section are available under @file{/etc/configuration} in the
7756 installation image. Thus, to get started with a system configuration
7757 providing a graphical display server (a ``desktop'' system), you can run
7758 something along these lines:
7762 # cp /etc/configuration/desktop.scm /mnt/etc/config.scm
7763 # zile /mnt/etc/config.scm
7766 You should pay attention to what your configuration file contains, and
7771 Make sure the @code{grub-configuration} form refers to the device you
7772 want to install GRUB on. You also need to specify the @code{grub-efi}
7773 package if you wish to use native UEFI boot.
7776 Be sure that your partition labels match the value of their respective
7777 @code{device} fields in your @code{file-system} configuration, assuming
7778 your @code{file-system} configuration sets the value of @code{title} to
7782 If there are encrypted or RAID partitions, make sure to add a
7783 @code{mapped-devices} field to describe them (@pxref{Mapped Devices}).
7786 Once you are done preparing the configuration file, the new system must
7787 be initialized (remember that the target root file system is mounted
7791 guix system init /mnt/etc/config.scm /mnt
7795 This copies all the necessary files and installs GRUB on
7796 @file{/dev/sdX}, unless you pass the @option{--no-bootloader} option. For
7797 more information, @pxref{Invoking guix system}. This command may trigger
7798 downloads or builds of missing packages, which can take some time.
7800 Once that command has completed---and hopefully succeeded!---you can run
7801 @command{reboot} and boot into the new system. The @code{root} password
7802 in the new system is initially empty; other users' passwords need to be
7803 initialized by running the @command{passwd} command as @code{root},
7804 unless your configuration specifies otherwise
7805 (@pxref{user-account-password, user account passwords}).
7807 @cindex upgrading GuixSD
7808 From then on, you can update GuixSD whenever you want by running
7809 @command{guix pull} as @code{root} (@pxref{Invoking guix pull}), and
7810 then running @command{guix system reconfigure} to build a new system
7811 generation with the latest packages and services (@pxref{Invoking guix
7812 system}). We recommend doing that regularly so that your system
7813 includes the latest security updates (@pxref{Security Updates}).
7815 Join us on @code{#guix} on the Freenode IRC network or on
7816 @file{guix-devel@@gnu.org} to share your experience---good or not so
7819 @node Installing GuixSD in a VM
7820 @subsection Installing GuixSD in a Virtual Machine
7822 @cindex virtual machine, GuixSD installation
7823 @cindex virtual private server (VPS)
7824 @cindex VPS (virtual private server)
7825 If you'd like to install GuixSD in a virtual machine (VM) or on a
7826 virtual private server (VPS) rather than on your beloved machine, this
7829 To boot a @uref{http://qemu.org/,QEMU} VM for installing GuixSD in a
7830 disk image, follow these steps:
7834 First, retrieve and decompress the GuixSD installation image as
7835 described previously (@pxref{USB Stick Installation}).
7838 Create a disk image that will hold the installed system. To make a
7839 qcow2-formatted disk image, use the @command{qemu-img} command:
7842 qemu-img create -f qcow2 guixsd.img 50G
7845 The resulting file will be much smaller than 50 GB (typically less than
7846 1 MB), but it will grow as the virtualized storage device is filled up.
7849 Boot the USB installation image in an VM:
7852 qemu-system-x86_64 -m 1024 -smp 1 \
7853 -net user -net nic,model=virtio -boot menu=on \
7854 -drive file=guixsd-usb-install-@value{VERSION}.@var{system} \
7855 -drive file=guixsd.img
7858 The ordering of the drives matters.
7860 In the VM console, quickly press the @kbd{F12} key to enter the boot
7861 menu. Then press the @kbd{2} key and the @kbd{RET} key to validate your
7865 You're now root in the VM, proceed with the installation process.
7866 @xref{Preparing for Installation}, and follow the instructions.
7869 Once installation is complete, you can boot the system that's on your
7870 @file{guixsd.img} image. @xref{Running GuixSD in a VM}, for how to do
7873 @node Building the Installation Image
7874 @subsection Building the Installation Image
7876 @cindex installation image
7877 The installation image described above was built using the @command{guix
7878 system} command, specifically:
7881 guix system disk-image gnu/system/install.scm
7884 Have a look at @file{gnu/system/install.scm} in the source tree,
7885 and see also @ref{Invoking guix system} for more information
7886 about the installation image.
7888 @node System Configuration
7889 @section System Configuration
7891 @cindex system configuration
7892 The Guix System Distribution supports a consistent whole-system configuration
7893 mechanism. By that we mean that all aspects of the global system
7894 configuration---such as the available system services, timezone and
7895 locale settings, user accounts---are declared in a single place. Such
7896 a @dfn{system configuration} can be @dfn{instantiated}---i.e., effected.
7898 One of the advantages of putting all the system configuration under the
7899 control of Guix is that it supports transactional system upgrades, and
7900 makes it possible to roll back to a previous system instantiation,
7901 should something go wrong with the new one (@pxref{Features}). Another
7902 advantage is that it makes it easy to replicate the exact same configuration
7903 across different machines, or at different points in time, without
7904 having to resort to additional administration tools layered on top of
7905 the own tools of the system.
7906 @c Yes, we're talking of Puppet, Chef, & co. here. ↑
7908 This section describes this mechanism. First we focus on the system
7909 administrator's viewpoint---explaining how the system is configured and
7910 instantiated. Then we show how this mechanism can be extended, for
7911 instance to support new system services.
7914 * Using the Configuration System:: Customizing your GNU system.
7915 * operating-system Reference:: Detail of operating-system declarations.
7916 * File Systems:: Configuring file system mounts.
7917 * Mapped Devices:: Block device extra processing.
7918 * User Accounts:: Specifying user accounts.
7919 * Locales:: Language and cultural convention settings.
7920 * Services:: Specifying system services.
7921 * Setuid Programs:: Programs running with root privileges.
7922 * X.509 Certificates:: Authenticating HTTPS servers.
7923 * Name Service Switch:: Configuring libc's name service switch.
7924 * Initial RAM Disk:: Linux-Libre bootstrapping.
7925 * Bootloader Configuration:: Configuring the boot loader.
7926 * Invoking guix system:: Instantiating a system configuration.
7927 * Running GuixSD in a VM:: How to run GuixSD in a virtual machine.
7928 * Defining Services:: Adding new service definitions.
7931 @node Using the Configuration System
7932 @subsection Using the Configuration System
7934 The operating system is configured by providing an
7935 @code{operating-system} declaration in a file that can then be passed to
7936 the @command{guix system} command (@pxref{Invoking guix system}). A
7937 simple setup, with the default system services, the default Linux-Libre
7938 kernel, initial RAM disk, and boot loader looks like this:
7940 @findex operating-system
7942 @include os-config-bare-bones.texi
7945 This example should be self-describing. Some of the fields defined
7946 above, such as @code{host-name} and @code{bootloader}, are mandatory.
7947 Others, such as @code{packages} and @code{services}, can be omitted, in
7948 which case they get a default value.
7950 Below we discuss the effect of some of the most important fields
7951 (@pxref{operating-system Reference}, for details about all the available
7952 fields), and how to @dfn{instantiate} the operating system using
7953 @command{guix system}.
7955 @unnumberedsubsubsec Globally-Visible Packages
7957 @vindex %base-packages
7958 The @code{packages} field lists packages that will be globally visible
7959 on the system, for all user accounts---i.e., in every user's @code{PATH}
7960 environment variable---in addition to the per-user profiles
7961 (@pxref{Invoking guix package}). The @var{%base-packages} variable
7962 provides all the tools one would expect for basic user and administrator
7963 tasks---including the GNU Core Utilities, the GNU Networking Utilities,
7964 the GNU Zile lightweight text editor, @command{find}, @command{grep},
7965 etc. The example above adds tcpdump to those, taken from the @code{(gnu
7966 packages admin)} module (@pxref{Package Modules}). The
7967 @code{(list package output)} syntax can be used to add a specific output
7971 (use-modules (gnu packages))
7972 (use-modules (gnu packages dns))
7976 (packages (cons (list bind "utils")
7980 @findex specification->package
7981 Referring to packages by variable name, like @var{tcpdump} above, has
7982 the advantage of being unambiguous; it also allows typos and such to be
7983 diagnosed right away as ``unbound variables''. The downside is that one
7984 needs to know which module defines which package, and to augment the
7985 @code{use-package-modules} line accordingly. To avoid that, one can use
7986 the @code{specification->package} procedure of the @code{(gnu packages)}
7987 module, which returns the best package for a given name or name and
7991 (use-modules (gnu packages))
7995 (packages (append (map specification->package
7996 '("tcpdump" "htop" "gnupg@@2.0"))
8000 @unnumberedsubsubsec System Services
8003 @vindex %base-services
8004 The @code{services} field lists @dfn{system services} to be made
8005 available when the system starts (@pxref{Services}).
8006 The @code{operating-system} declaration above specifies that, in
8007 addition to the basic services, we want the @command{lshd} secure shell
8008 daemon listening on port 2222 (@pxref{Networking Services,
8009 @code{lsh-service}}). Under the hood,
8010 @code{lsh-service} arranges so that @code{lshd} is started with the
8011 right command-line options, possibly with supporting configuration files
8012 generated as needed (@pxref{Defining Services}).
8014 @cindex customization, of services
8015 @findex modify-services
8016 Occasionally, instead of using the base services as is, you will want to
8017 customize them. To do this, use @code{modify-services} (@pxref{Service
8018 Reference, @code{modify-services}}) to modify the list.
8020 For example, suppose you want to modify @code{guix-daemon} and Mingetty
8021 (the console log-in) in the @var{%base-services} list (@pxref{Base
8022 Services, @code{%base-services}}). To do that, you can write the
8023 following in your operating system declaration:
8026 (define %my-services
8027 ;; My very own list of services.
8028 (modify-services %base-services
8029 (guix-service-type config =>
8032 (use-substitutes? #f)
8033 (extra-options '("--gc-keep-derivations"))))
8034 (mingetty-service-type config =>
8035 (mingetty-configuration
8036 (inherit config)))))
8040 (services %my-services))
8043 This changes the configuration---i.e., the service parameters---of the
8044 @code{guix-service-type} instance, and that of all the
8045 @code{mingetty-service-type} instances in the @var{%base-services} list.
8046 Observe how this is accomplished: first, we arrange for the original
8047 configuration to be bound to the identifier @code{config} in the
8048 @var{body}, and then we write the @var{body} so that it evaluates to the
8049 desired configuration. In particular, notice how we use @code{inherit}
8050 to create a new configuration which has the same values as the old
8051 configuration, but with a few modifications.
8053 @cindex encrypted disk
8054 The configuration for a typical ``desktop'' usage, with an encrypted
8055 root partition, the X11 display
8056 server, GNOME and Xfce (users can choose which of these desktop
8057 environments to use at the log-in screen by pressing @kbd{F1}), network
8058 management, power management, and more, would look like this:
8061 @include os-config-desktop.texi
8064 A graphical UEFI system with a choice of lightweight window managers
8065 instead of full-blown desktop environments would look like this:
8068 @include os-config-lightweight-desktop.texi
8071 @xref{Desktop Services}, for the exact list of services provided by
8072 @var{%desktop-services}. @xref{X.509 Certificates}, for background
8073 information about the @code{nss-certs} package that is used here.
8075 Again, @var{%desktop-services} is just a list of service objects. If
8076 you want to remove services from there, you can do so using the
8077 procedures for list filtering (@pxref{SRFI-1 Filtering and
8078 Partitioning,,, guile, GNU Guile Reference Manual}). For instance, the
8079 following expression returns a list that contains all the services in
8080 @var{%desktop-services} minus the Avahi service:
8083 (remove (lambda (service)
8084 (eq? (service-kind service) avahi-service-type))
8088 @unnumberedsubsubsec Instantiating the System
8090 Assuming the @code{operating-system} declaration
8091 is stored in the @file{my-system-config.scm}
8092 file, the @command{guix system reconfigure my-system-config.scm} command
8093 instantiates that configuration, and makes it the default GRUB boot
8094 entry (@pxref{Invoking guix system}).
8096 The normal way to change the system configuration is by updating this
8097 file and re-running @command{guix system reconfigure}. One should never
8098 have to touch files in @file{/etc} or to run commands that modify the
8099 system state such as @command{useradd} or @command{grub-install}. In
8100 fact, you must avoid that since that would not only void your warranty
8101 but also prevent you from rolling back to previous versions of your
8102 system, should you ever need to.
8104 @cindex roll-back, of the operating system
8105 Speaking of roll-back, each time you run @command{guix system
8106 reconfigure}, a new @dfn{generation} of the system is created---without
8107 modifying or deleting previous generations. Old system generations get
8108 an entry in the bootloader boot menu, allowing you to boot them in case
8109 something went wrong with the latest generation. Reassuring, no? The
8110 @command{guix system list-generations} command lists the system
8111 generations available on disk. It is also possible to roll back the
8112 system via the commands @command{guix system roll-back} and
8113 @command{guix system switch-generation}.
8115 Although the command @command{guix system reconfigure} will not modify
8116 previous generations, must take care when the current generation is not
8117 the latest (e.g., after invoking @command{guix system roll-back}), since
8118 the operation might overwrite a later generation (@pxref{Invoking guix
8121 @unnumberedsubsubsec The Programming Interface
8123 At the Scheme level, the bulk of an @code{operating-system} declaration
8124 is instantiated with the following monadic procedure (@pxref{The Store
8127 @deffn {Monadic Procedure} operating-system-derivation os
8128 Return a derivation that builds @var{os}, an @code{operating-system}
8129 object (@pxref{Derivations}).
8131 The output of the derivation is a single directory that refers to all
8132 the packages, configuration files, and other supporting files needed to
8133 instantiate @var{os}.
8136 This procedure is provided by the @code{(gnu system)} module. Along
8137 with @code{(gnu services)} (@pxref{Services}), this module contains the
8138 guts of GuixSD. Make sure to visit it!
8141 @node operating-system Reference
8142 @subsection @code{operating-system} Reference
8144 This section summarizes all the options available in
8145 @code{operating-system} declarations (@pxref{Using the Configuration
8148 @deftp {Data Type} operating-system
8149 This is the data type representing an operating system configuration.
8150 By that, we mean all the global system configuration, not per-user
8151 configuration (@pxref{Using the Configuration System}).
8154 @item @code{kernel} (default: @var{linux-libre})
8155 The package object of the operating system kernel to use@footnote{Currently
8156 only the Linux-libre kernel is supported. In the future, it will be
8157 possible to use the GNU@tie{}Hurd.}.
8159 @item @code{kernel-arguments} (default: @code{'()})
8160 List of strings or gexps representing additional arguments to pass on
8161 the command-line of the kernel---e.g., @code{("console=ttyS0")}.
8163 @item @code{bootloader}
8164 The system bootloader configuration object. @xref{Bootloader Configuration}.
8166 @item @code{initrd} (default: @code{base-initrd})
8168 @cindex initial RAM disk
8169 A two-argument monadic procedure that returns an initial RAM disk for
8170 the Linux kernel. @xref{Initial RAM Disk}.
8172 @item @code{firmware} (default: @var{%base-firmware})
8174 List of firmware packages loadable by the operating system kernel.
8176 The default includes firmware needed for Atheros- and Broadcom-based
8177 WiFi devices (Linux-libre modules @code{ath9k} and @code{b43-open},
8178 respectively). @xref{Hardware Considerations}, for more info on
8181 @item @code{host-name}
8184 @item @code{hosts-file}
8186 A file-like object (@pxref{G-Expressions, file-like objects}) for use as
8187 @file{/etc/hosts} (@pxref{Host Names,,, libc, The GNU C Library
8188 Reference Manual}). The default is a file with entries for
8189 @code{localhost} and @var{host-name}.
8191 @item @code{mapped-devices} (default: @code{'()})
8192 A list of mapped devices. @xref{Mapped Devices}.
8194 @item @code{file-systems}
8195 A list of file systems. @xref{File Systems}.
8197 @item @code{swap-devices} (default: @code{'()})
8198 @cindex swap devices
8199 A list of strings identifying devices or files to be used for ``swap
8200 space'' (@pxref{Memory Concepts,,, libc, The GNU C Library Reference
8201 Manual}). For example, @code{'("/dev/sda3")} or @code{'("/swapfile")}.
8202 It is possible to specify a swap file in a file system on a mapped
8203 device, provided that the necessary device mapping and file system are
8204 also specified. @xref{Mapped Devices} and @ref{File Systems}.
8206 @item @code{users} (default: @code{%base-user-accounts})
8207 @itemx @code{groups} (default: @var{%base-groups})
8208 List of user accounts and groups. @xref{User Accounts}.
8210 @item @code{skeletons} (default: @code{(default-skeletons)})
8211 A list target file name/file-like object tuples (@pxref{G-Expressions,
8212 file-like objects}). These are the skeleton files that will be added to
8213 the home directory of newly-created user accounts.
8215 For instance, a valid value may look like this:
8218 `((".bashrc" ,(plain-file "bashrc" "echo Hello\n"))
8219 (".guile" ,(plain-file "guile"
8220 "(use-modules (ice-9 readline))
8221 (activate-readline)")))
8224 @item @code{issue} (default: @var{%default-issue})
8225 A string denoting the contents of the @file{/etc/issue} file, which is
8226 displayed when users log in on a text console.
8228 @item @code{packages} (default: @var{%base-packages})
8229 The set of packages installed in the global profile, which is accessible
8230 at @file{/run/current-system/profile}.
8232 The default set includes core utilities and it is good practice to
8233 install non-core utilities in user profiles (@pxref{Invoking guix
8236 @item @code{timezone}
8237 A timezone identifying string---e.g., @code{"Europe/Paris"}.
8239 You can run the @command{tzselect} command to find out which timezone
8240 string corresponds to your region. Choosing an invalid timezone name
8241 causes @command{guix system} to fail.
8243 @item @code{locale} (default: @code{"en_US.utf8"})
8244 The name of the default locale (@pxref{Locale Names,,, libc, The GNU C
8245 Library Reference Manual}). @xref{Locales}, for more information.
8247 @item @code{locale-definitions} (default: @var{%default-locale-definitions})
8248 The list of locale definitions to be compiled and that may be used at
8249 run time. @xref{Locales}.
8251 @item @code{locale-libcs} (default: @code{(list @var{glibc})})
8252 The list of GNU@tie{}libc packages whose locale data and tools are used
8253 to build the locale definitions. @xref{Locales}, for compatibility
8254 considerations that justify this option.
8256 @item @code{name-service-switch} (default: @var{%default-nss})
8257 Configuration of the libc name service switch (NSS)---a
8258 @code{<name-service-switch>} object. @xref{Name Service Switch}, for
8261 @item @code{services} (default: @var{%base-services})
8262 A list of service objects denoting system services. @xref{Services}.
8264 @item @code{pam-services} (default: @code{(base-pam-services)})
8266 @cindex pluggable authentication modules
8267 Linux @dfn{pluggable authentication module} (PAM) services.
8268 @c FIXME: Add xref to PAM services section.
8270 @item @code{setuid-programs} (default: @var{%setuid-programs})
8271 List of string-valued G-expressions denoting setuid programs.
8272 @xref{Setuid Programs}.
8274 @item @code{sudoers-file} (default: @var{%sudoers-specification})
8275 @cindex sudoers file
8276 The contents of the @file{/etc/sudoers} file as a file-like object
8277 (@pxref{G-Expressions, @code{local-file} and @code{plain-file}}).
8279 This file specifies which users can use the @command{sudo} command, what
8280 they are allowed to do, and what privileges they may gain. The default
8281 is that only @code{root} and members of the @code{wheel} group may use
8288 @subsection File Systems
8290 The list of file systems to be mounted is specified in the
8291 @code{file-systems} field of the operating system declaration
8292 (@pxref{Using the Configuration System}). Each file system is declared
8293 using the @code{file-system} form, like this:
8297 (mount-point "/home")
8298 (device "/dev/sda3")
8302 As usual, some of the fields are mandatory---those shown in the example
8303 above---while others can be omitted. These are described below.
8305 @deftp {Data Type} file-system
8306 Objects of this type represent file systems to be mounted. They
8307 contain the following members:
8311 This is a string specifying the type of the file system---e.g.,
8314 @item @code{mount-point}
8315 This designates the place where the file system is to be mounted.
8318 This names the ``source'' of the file system. By default it is the name
8319 of a node under @file{/dev}, but its meaning depends on the @code{title}
8320 field described below.
8322 @item @code{title} (default: @code{'device})
8323 This is a symbol that specifies how the @code{device} field is to be
8326 When it is the symbol @code{device}, then the @code{device} field is
8327 interpreted as a file name; when it is @code{label}, then @code{device}
8328 is interpreted as a partition label name; when it is @code{uuid},
8329 @code{device} is interpreted as a partition unique identifier (UUID).
8331 UUIDs may be converted from their string representation (as shown by the
8332 @command{tune2fs -l} command) using the @code{uuid} form@footnote{The
8333 @code{uuid} form expects 16-byte UUIDs as defined in
8334 @uref{https://tools.ietf.org/html/rfc4122, RFC@tie{}4122}. This is the
8335 form of UUID used by the ext2 family of file systems and others, but it
8336 is different from ``UUIDs'' found in FAT file systems, for instance.},
8341 (mount-point "/home")
8344 (device (uuid "4dab5feb-d176-45de-b287-9b0a6e4c01cb")))
8347 The @code{label} and @code{uuid} options offer a way to refer to disk
8348 partitions without having to hard-code their actual device
8349 name@footnote{Note that, while it is tempting to use
8350 @file{/dev/disk/by-uuid} and similar device names to achieve the same
8351 result, this is not recommended: These special device nodes are created
8352 by the udev daemon and may be unavailable at the time the device is
8355 However, when the source of a file system is a mapped device (@pxref{Mapped
8356 Devices}), its @code{device} field @emph{must} refer to the mapped
8357 device name---e.g., @file{/dev/mapper/root-partition}---and consequently
8358 @code{title} must be set to @code{'device}. This is required so that
8359 the system knows that mounting the file system depends on having the
8360 corresponding device mapping established.
8362 @item @code{flags} (default: @code{'()})
8363 This is a list of symbols denoting mount flags. Recognized flags
8364 include @code{read-only}, @code{bind-mount}, @code{no-dev} (disallow
8365 access to special files), @code{no-suid} (ignore setuid and setgid
8366 bits), and @code{no-exec} (disallow program execution.)
8368 @item @code{options} (default: @code{#f})
8369 This is either @code{#f}, or a string denoting mount options.
8371 @item @code{mount?} (default: @code{#t})
8372 This value indicates whether to automatically mount the file system when
8373 the system is brought up. When set to @code{#f}, the file system gets
8374 an entry in @file{/etc/fstab} (read by the @command{mount} command) but
8375 is not automatically mounted.
8377 @item @code{needed-for-boot?} (default: @code{#f})
8378 This Boolean value indicates whether the file system is needed when
8379 booting. If that is true, then the file system is mounted when the
8380 initial RAM disk (initrd) is loaded. This is always the case, for
8381 instance, for the root file system.
8383 @item @code{check?} (default: @code{#t})
8384 This Boolean indicates whether the file system needs to be checked for
8385 errors before being mounted.
8387 @item @code{create-mount-point?} (default: @code{#f})
8388 When true, the mount point is created if it does not exist yet.
8390 @item @code{dependencies} (default: @code{'()})
8391 This is a list of @code{<file-system>} or @code{<mapped-device>} objects
8392 representing file systems that must be mounted or mapped devices that
8393 must be opened before (and unmounted or closed after) this one.
8395 As an example, consider a hierarchy of mounts: @file{/sys/fs/cgroup} is
8396 a dependency of @file{/sys/fs/cgroup/cpu} and
8397 @file{/sys/fs/cgroup/memory}.
8399 Another example is a file system that depends on a mapped device, for
8400 example for an encrypted partition (@pxref{Mapped Devices}).
8404 The @code{(gnu system file-systems)} exports the following useful
8407 @defvr {Scheme Variable} %base-file-systems
8408 These are essential file systems that are required on normal systems,
8409 such as @var{%pseudo-terminal-file-system} and @var{%immutable-store} (see
8410 below.) Operating system declarations should always contain at least
8414 @defvr {Scheme Variable} %pseudo-terminal-file-system
8415 This is the file system to be mounted as @file{/dev/pts}. It supports
8416 @dfn{pseudo-terminals} created @i{via} @code{openpty} and similar
8417 functions (@pxref{Pseudo-Terminals,,, libc, The GNU C Library Reference
8418 Manual}). Pseudo-terminals are used by terminal emulators such as
8422 @defvr {Scheme Variable} %shared-memory-file-system
8423 This file system is mounted as @file{/dev/shm} and is used to support
8424 memory sharing across processes (@pxref{Memory-mapped I/O,
8425 @code{shm_open},, libc, The GNU C Library Reference Manual}).
8428 @defvr {Scheme Variable} %immutable-store
8429 This file system performs a read-only ``bind mount'' of
8430 @file{/gnu/store}, making it read-only for all the users including
8431 @code{root}. This prevents against accidental modification by software
8432 running as @code{root} or by system administrators.
8434 The daemon itself is still able to write to the store: it remounts it
8435 read-write in its own ``name space.''
8438 @defvr {Scheme Variable} %binary-format-file-system
8439 The @code{binfmt_misc} file system, which allows handling of arbitrary
8440 executable file types to be delegated to user space. This requires the
8441 @code{binfmt.ko} kernel module to be loaded.
8444 @defvr {Scheme Variable} %fuse-control-file-system
8445 The @code{fusectl} file system, which allows unprivileged users to mount
8446 and unmount user-space FUSE file systems. This requires the
8447 @code{fuse.ko} kernel module to be loaded.
8450 @node Mapped Devices
8451 @subsection Mapped Devices
8453 @cindex device mapping
8454 @cindex mapped devices
8455 The Linux kernel has a notion of @dfn{device mapping}: a block device,
8456 such as a hard disk partition, can be @dfn{mapped} into another device,
8457 usually in @code{/dev/mapper/},
8458 with additional processing over the data that flows through
8459 it@footnote{Note that the GNU@tie{}Hurd makes no difference between the
8460 concept of a ``mapped device'' and that of a file system: both boil down
8461 to @emph{translating} input/output operations made on a file to
8462 operations on its backing store. Thus, the Hurd implements mapped
8463 devices, like file systems, using the generic @dfn{translator} mechanism
8464 (@pxref{Translators,,, hurd, The GNU Hurd Reference Manual}).}. A
8465 typical example is encryption device mapping: all writes to the mapped
8466 device are encrypted, and all reads are deciphered, transparently.
8467 Guix extends this notion by considering any device or set of devices that
8468 are @dfn{transformed} in some way to create a new device; for instance,
8469 RAID devices are obtained by @dfn{assembling} several other devices, such
8470 as hard disks or partitions, into a new one that behaves as one partition.
8471 Other examples, not yet implemented, are LVM logical volumes.
8473 Mapped devices are declared using the @code{mapped-device} form,
8474 defined as follows; for examples, see below.
8476 @deftp {Data Type} mapped-device
8477 Objects of this type represent device mappings that will be made when
8478 the system boots up.
8482 This is either a string specifying the name of the block device to be mapped,
8483 such as @code{"/dev/sda3"}, or a list of such strings when several devices
8484 need to be assembled for creating a new one.
8487 This string specifies the name of the resulting mapped device. For
8488 kernel mappers such as encrypted devices of type @code{luks-device-mapping},
8489 specifying @code{"my-partition"} leads to the creation of
8490 the @code{"/dev/mapper/my-partition"} device.
8491 For RAID devices of type @code{raid-device-mapping}, the full device name
8492 such as @code{"/dev/md0"} needs to be given.
8495 This must be a @code{mapped-device-kind} object, which specifies how
8496 @var{source} is mapped to @var{target}.
8500 @defvr {Scheme Variable} luks-device-mapping
8501 This defines LUKS block device encryption using the @command{cryptsetup}
8502 command from the package with the same name. It relies on the
8503 @code{dm-crypt} Linux kernel module.
8506 @defvr {Scheme Variable} raid-device-mapping
8507 This defines a RAID device, which is assembled using the @code{mdadm}
8508 command from the package with the same name. It requires a Linux kernel
8509 module for the appropriate RAID level to be loaded, such as @code{raid456}
8510 for RAID-4, RAID-5 or RAID-6, or @code{raid10} for RAID-10.
8513 @cindex disk encryption
8515 The following example specifies a mapping from @file{/dev/sda3} to
8516 @file{/dev/mapper/home} using LUKS---the
8517 @url{https://gitlab.com/cryptsetup/cryptsetup,Linux Unified Key Setup}, a
8518 standard mechanism for disk encryption.
8519 The @file{/dev/mapper/home}
8520 device can then be used as the @code{device} of a @code{file-system}
8521 declaration (@pxref{File Systems}).
8525 (source "/dev/sda3")
8527 (type luks-device-mapping))
8530 Alternatively, to become independent of device numbering, one may obtain
8531 the LUKS UUID (@dfn{unique identifier}) of the source device by a
8535 cryptsetup luksUUID /dev/sda3
8538 and use it as follows:
8542 (source (uuid "cb67fc72-0d54-4c88-9d4b-b225f30b0f44"))
8544 (type luks-device-mapping))
8547 @cindex swap encryption
8548 It is also desirable to encrypt swap space, since swap space may contain
8549 sensitive data. One way to accomplish that is to use a swap file in a
8550 file system on a device mapped via LUKS encryption. In this way, the
8551 swap file is encrypted because the entire device is encrypted.
8552 @xref{Preparing for Installation,,Disk Partitioning}, for an example.
8554 A RAID device formed of the partitions @file{/dev/sda1} and @file{/dev/sdb1}
8555 may be declared as follows:
8559 (source (list "/dev/sda1" "/dev/sdb1"))
8561 (type raid-device-mapping))
8564 The @file{/dev/md0} device can then be used as the @code{device} of a
8565 @code{file-system} declaration (@pxref{File Systems}).
8566 Note that the RAID level need not be given; it is chosen during the
8567 initial creation and formatting of the RAID device and is determined
8568 automatically later.
8572 @subsection User Accounts
8576 @cindex user accounts
8577 User accounts and groups are entirely managed through the
8578 @code{operating-system} declaration. They are specified with the
8579 @code{user-account} and @code{user-group} forms:
8585 (supplementary-groups '("wheel" ;allow use of sudo, etc.
8587 "video" ;video devices such as webcams
8588 "cdrom")) ;the good ol' CD-ROM
8589 (comment "Bob's sister")
8590 (home-directory "/home/alice"))
8593 When booting or upon completion of @command{guix system reconfigure},
8594 the system ensures that only the user accounts and groups specified in
8595 the @code{operating-system} declaration exist, and with the specified
8596 properties. Thus, account or group creations or modifications made by
8597 directly invoking commands such as @command{useradd} are lost upon
8598 reconfiguration or reboot. This ensures that the system remains exactly
8601 @deftp {Data Type} user-account
8602 Objects of this type represent user accounts. The following members may
8607 The name of the user account.
8611 This is the name (a string) or identifier (a number) of the user group
8612 this account belongs to.
8614 @item @code{supplementary-groups} (default: @code{'()})
8615 Optionally, this can be defined as a list of group names that this
8618 @item @code{uid} (default: @code{#f})
8619 This is the user ID for this account (a number), or @code{#f}. In the
8620 latter case, a number is automatically chosen by the system when the
8623 @item @code{comment} (default: @code{""})
8624 A comment about the account, such as the account owner's full name.
8626 @item @code{home-directory}
8627 This is the name of the home directory for the account.
8629 @item @code{create-home-directory?} (default: @code{#t})
8630 Indicates whether the home directory of this account should be created
8631 if it does not exist yet.
8633 @item @code{shell} (default: Bash)
8634 This is a G-expression denoting the file name of a program to be used as
8635 the shell (@pxref{G-Expressions}).
8637 @item @code{system?} (default: @code{#f})
8638 This Boolean value indicates whether the account is a ``system''
8639 account. System accounts are sometimes treated specially; for instance,
8640 graphical login managers do not list them.
8642 @anchor{user-account-password}
8643 @item @code{password} (default: @code{#f})
8644 You would normally leave this field to @code{#f}, initialize user
8645 passwords as @code{root} with the @command{passwd} command, and then let
8646 users change it with @command{passwd}. Passwords set with
8647 @command{passwd} are of course preserved across reboot and
8650 If you @emph{do} want to have a preset password for an account, then
8651 this field must contain the encrypted password, as a string.
8652 @xref{crypt,,, libc, The GNU C Library Reference Manual}, for more information
8653 on password encryption, and @ref{Encryption,,, guile, GNU Guile Reference
8654 Manual}, for information on Guile's @code{crypt} procedure.
8660 User group declarations are even simpler:
8663 (user-group (name "students"))
8666 @deftp {Data Type} user-group
8667 This type is for, well, user groups. There are just a few fields:
8671 The name of the group.
8673 @item @code{id} (default: @code{#f})
8674 The group identifier (a number). If @code{#f}, a new number is
8675 automatically allocated when the group is created.
8677 @item @code{system?} (default: @code{#f})
8678 This Boolean value indicates whether the group is a ``system'' group.
8679 System groups have low numerical IDs.
8681 @item @code{password} (default: @code{#f})
8682 What, user groups can have a password? Well, apparently yes. Unless
8683 @code{#f}, this field specifies the password of the group.
8688 For convenience, a variable lists all the basic user groups one may
8691 @defvr {Scheme Variable} %base-groups
8692 This is the list of basic user groups that users and/or packages expect
8693 to be present on the system. This includes groups such as ``root'',
8694 ``wheel'', and ``users'', as well as groups used to control access to
8695 specific devices such as ``audio'', ``disk'', and ``cdrom''.
8698 @defvr {Scheme Variable} %base-user-accounts
8699 This is the list of basic system accounts that programs may expect to
8700 find on a GNU/Linux system, such as the ``nobody'' account.
8702 Note that the ``root'' account is not included here. It is a
8703 special-case and is automatically added whether or not it is specified.
8710 A @dfn{locale} defines cultural conventions for a particular language
8711 and region of the world (@pxref{Locales,,, libc, The GNU C Library
8712 Reference Manual}). Each locale has a name that typically has the form
8713 @code{@var{language}_@var{territory}.@var{codeset}}---e.g.,
8714 @code{fr_LU.utf8} designates the locale for the French language, with
8715 cultural conventions from Luxembourg, and using the UTF-8 encoding.
8717 @cindex locale definition
8718 Usually, you will want to specify the default locale for the machine
8719 using the @code{locale} field of the @code{operating-system} declaration
8720 (@pxref{operating-system Reference, @code{locale}}).
8722 The selected locale is automatically added to the @dfn{locale
8723 definitions} known to the system if needed, with its codeset inferred
8724 from its name---e.g., @code{bo_CN.utf8} will be assumed to use the
8725 @code{UTF-8} codeset. Additional locale definitions can be specified in
8726 the @code{locale-definitions} slot of @code{operating-system}---this is
8727 useful, for instance, if the codeset could not be inferred from the
8728 locale name. The default set of locale definitions includes some widely
8729 used locales, but not all the available locales, in order to save space.
8731 For instance, to add the North Frisian locale for Germany, the value of
8735 (cons (locale-definition
8736 (name "fy_DE.utf8") (source "fy_DE"))
8737 %default-locale-definitions)
8740 Likewise, to save space, one might want @code{locale-definitions} to
8741 list only the locales that are actually used, as in:
8744 (list (locale-definition
8745 (name "ja_JP.eucjp") (source "ja_JP")
8746 (charset "EUC-JP")))
8750 The compiled locale definitions are available at
8751 @file{/run/current-system/locale/X.Y}, where @code{X.Y} is the libc
8752 version, which is the default location where the GNU@tie{}libc provided
8753 by Guix looks for locale data. This can be overridden using the
8754 @code{LOCPATH} environment variable (@pxref{locales-and-locpath,
8755 @code{LOCPATH} and locale packages}).
8757 The @code{locale-definition} form is provided by the @code{(gnu system
8758 locale)} module. Details are given below.
8760 @deftp {Data Type} locale-definition
8761 This is the data type of a locale definition.
8766 The name of the locale. @xref{Locale Names,,, libc, The GNU C Library
8767 Reference Manual}, for more information on locale names.
8770 The name of the source for that locale. This is typically the
8771 @code{@var{language}_@var{territory}} part of the locale name.
8773 @item @code{charset} (default: @code{"UTF-8"})
8774 The ``character set'' or ``code set'' for that locale,
8775 @uref{http://www.iana.org/assignments/character-sets, as defined by
8781 @defvr {Scheme Variable} %default-locale-definitions
8782 A list of commonly used UTF-8 locales, used as the default
8783 value of the @code{locale-definitions} field of @code{operating-system}
8787 @cindex normalized codeset in locale names
8788 These locale definitions use the @dfn{normalized codeset} for the part
8789 that follows the dot in the name (@pxref{Using gettextized software,
8790 normalized codeset,, libc, The GNU C Library Reference Manual}). So for
8791 instance it has @code{uk_UA.utf8} but @emph{not}, say,
8795 @subsubsection Locale Data Compatibility Considerations
8797 @cindex incompatibility, of locale data
8798 @code{operating-system} declarations provide a @code{locale-libcs} field
8799 to specify the GNU@tie{}libc packages that are used to compile locale
8800 declarations (@pxref{operating-system Reference}). ``Why would I
8801 care?'', you may ask. Well, it turns out that the binary format of
8802 locale data is occasionally incompatible from one libc version to
8805 @c See <https://sourceware.org/ml/libc-alpha/2015-09/msg00575.html>
8806 @c and <https://lists.gnu.org/archive/html/guix-devel/2015-08/msg00737.html>.
8807 For instance, a program linked against libc version 2.21 is unable to
8808 read locale data produced with libc 2.22; worse, that program
8809 @emph{aborts} instead of simply ignoring the incompatible locale
8810 data@footnote{Versions 2.23 and later of GNU@tie{}libc will simply skip
8811 the incompatible locale data, which is already an improvement.}.
8812 Similarly, a program linked against libc 2.22 can read most, but not
8813 all, of the locale data from libc 2.21 (specifically, @code{LC_COLLATE}
8814 data is incompatible); thus calls to @code{setlocale} may fail, but
8815 programs will not abort.
8817 The ``problem'' in GuixSD is that users have a lot of freedom: They can
8818 choose whether and when to upgrade software in their profiles, and might
8819 be using a libc version different from the one the system administrator
8820 used to build the system-wide locale data.
8822 Fortunately, unprivileged users can also install their own locale data
8823 and define @var{GUIX_LOCPATH} accordingly (@pxref{locales-and-locpath,
8824 @code{GUIX_LOCPATH} and locale packages}).
8826 Still, it is best if the system-wide locale data at
8827 @file{/run/current-system/locale} is built for all the libc versions
8828 actually in use on the system, so that all the programs can access
8829 it---this is especially crucial on a multi-user system. To do that, the
8830 administrator can specify several libc packages in the
8831 @code{locale-libcs} field of @code{operating-system}:
8834 (use-package-modules base)
8838 (locale-libcs (list glibc-2.21 (canonical-package glibc))))
8841 This example would lead to a system containing locale definitions for
8842 both libc 2.21 and the current version of libc in
8843 @file{/run/current-system/locale}.
8847 @subsection Services
8849 @cindex system services
8850 An important part of preparing an @code{operating-system} declaration is
8851 listing @dfn{system services} and their configuration (@pxref{Using the
8852 Configuration System}). System services are typically daemons launched
8853 when the system boots, or other actions needed at that time---e.g.,
8854 configuring network access.
8856 GuixSD has a broad definition of ``service'' (@pxref{Service
8857 Composition}), but many services are managed by the GNU@tie{}Shepherd
8858 (@pxref{Shepherd Services}). On a running system, the @command{herd}
8859 command allows you to list the available services, show their status,
8860 start and stop them, or do other specific operations (@pxref{Jump
8861 Start,,, shepherd, The GNU Shepherd Manual}). For example:
8867 The above command, run as @code{root}, lists the currently defined
8868 services. The @command{herd doc} command shows a synopsis of the given
8873 Run libc's name service cache daemon (nscd).
8876 The @command{start}, @command{stop}, and @command{restart} sub-commands
8877 have the effect you would expect. For instance, the commands below stop
8878 the nscd service and restart the Xorg display server:
8882 Service nscd has been stopped.
8883 # herd restart xorg-server
8884 Service xorg-server has been stopped.
8885 Service xorg-server has been started.
8888 The following sections document the available services, starting with
8889 the core services, that may be used in an @code{operating-system}
8893 * Base Services:: Essential system services.
8894 * Scheduled Job Execution:: The mcron service.
8895 * Log Rotation:: The rottlog service.
8896 * Networking Services:: Network setup, SSH daemon, etc.
8897 * X Window:: Graphical display.
8898 * Printing Services:: Local and remote printer support.
8899 * Desktop Services:: D-Bus and desktop services.
8900 * Database Services:: SQL databases.
8901 * Mail Services:: IMAP, POP3, SMTP, and all that.
8902 * Messaging Services:: Messaging services.
8903 * Kerberos Services:: Kerberos services.
8904 * Web Services:: Web servers.
8905 * DNS Services:: DNS daemons.
8906 * VPN Services:: VPN daemons.
8907 * Network File System:: NFS related services.
8908 * Continuous Integration:: The Cuirass service.
8909 * Power management Services:: The TLP tool.
8910 * Miscellaneous Services:: Other services.
8914 @subsubsection Base Services
8916 The @code{(gnu services base)} module provides definitions for the basic
8917 services that one expects from the system. The services exported by
8918 this module are listed below.
8920 @defvr {Scheme Variable} %base-services
8921 This variable contains a list of basic services (@pxref{Service Types
8922 and Services}, for more information on service objects) one would
8923 expect from the system: a login service (mingetty) on each tty, syslogd,
8924 the libc name service cache daemon (nscd), the udev device manager, and
8927 This is the default value of the @code{services} field of
8928 @code{operating-system} declarations. Usually, when customizing a
8929 system, you will want to append services to @var{%base-services}, like
8933 (cons* (avahi-service) (lsh-service) %base-services)
8937 @defvr {Scheme Variable} special-files-service-type
8938 This is the service that sets up ``special files'' such as
8939 @file{/bin/sh}; an instance of it is part of @code{%base-services}.
8941 The value associated with @code{special-files-service-type} services
8942 must be a list of tuples where the first element is the ``special file''
8943 and the second element is its target. By default it is:
8945 @cindex @file{/bin/sh}
8946 @cindex @file{sh}, in @file{/bin}
8948 `(("/bin/sh" ,(file-append @var{bash} "/bin/sh")))
8951 @cindex @file{/usr/bin/env}
8952 @cindex @file{env}, in @file{/usr/bin}
8953 If you want to add, say, @code{/usr/bin/env} to your system, you can
8957 `(("/bin/sh" ,(file-append @var{bash} "/bin/sh"))
8958 ("/usr/bin/env" ,(file-append @var{coreutils} "/bin/env")))
8961 Since this is part of @code{%base-services}, you can use
8962 @code{modify-services} to customize the set of special files
8963 (@pxref{Service Reference, @code{modify-services}}). But the simple way
8964 to add a special file is @i{via} the @code{extra-special-file} procedure
8968 @deffn {Scheme Procedure} extra-special-file @var{file} @var{target}
8969 Use @var{target} as the ``special file'' @var{file}.
8971 For example, adding the following lines to the @code{services} field of
8972 your operating system declaration leads to a @file{/usr/bin/env}
8976 (extra-special-file "/usr/bin/env"
8977 (file-append coreutils "/bin/env"))
8981 @deffn {Scheme Procedure} host-name-service @var{name}
8982 Return a service that sets the host name to @var{name}.
8985 @deffn {Scheme Procedure} login-service @var{config}
8986 Return a service to run login according to @var{config}, a
8987 @code{<login-configuration>} object, which specifies the message of the day,
8991 @deftp {Data Type} login-configuration
8992 This is the data type representing the configuration of login.
8997 @cindex message of the day
8998 A file-like object containing the ``message of the day''.
9000 @item @code{allow-empty-passwords?} (default: @code{#t})
9001 Allow empty passwords by default so that first-time users can log in when
9002 the 'root' account has just been created.
9007 @deffn {Scheme Procedure} mingetty-service @var{config}
9008 Return a service to run mingetty according to @var{config}, a
9009 @code{<mingetty-configuration>} object, which specifies the tty to run, among
9013 @deftp {Data Type} mingetty-configuration
9014 This is the data type representing the configuration of Mingetty, which
9015 provides the default implementation of virtual console log-in.
9020 The name of the console this Mingetty runs on---e.g., @code{"tty1"}.
9022 @item @code{auto-login} (default: @code{#f})
9023 When true, this field must be a string denoting the user name under
9024 which the system automatically logs in. When it is @code{#f}, a
9025 user name and password must be entered to log in.
9027 @item @code{login-program} (default: @code{#f})
9028 This must be either @code{#f}, in which case the default log-in program
9029 is used (@command{login} from the Shadow tool suite), or a gexp denoting
9030 the name of the log-in program.
9032 @item @code{login-pause?} (default: @code{#f})
9033 When set to @code{#t} in conjunction with @var{auto-login}, the user
9034 will have to press a key before the log-in shell is launched.
9036 @item @code{mingetty} (default: @var{mingetty})
9037 The Mingetty package to use.
9042 @deffn {Scheme Procedure} agetty-service @var{config}
9043 Return a service to run agetty according to @var{config}, an
9044 @code{<agetty-configuration>} object, which specifies the tty to run,
9048 @deftp {Data Type} agetty-configuration
9049 This is the data type representing the configuration of agetty, which
9050 implements virtual and serial console log-in. See the @code{agetty(8)}
9051 man page for more information.
9056 The name of the console this agetty runs on, as a string---e.g.,
9057 @code{"ttyS0"}. This argument is mandatory.
9059 @item @code{baud-rate} (default: @code{#f})
9060 A string containing a comma-separated list of one or more baud rates, in
9063 @item @code{term} (default: @code{#f})
9064 A string containing the value used for the @code{TERM} environment
9067 @item @code{eight-bits?} (default: @code{#f})
9068 When @code{#t}, the tty is assumed to be 8-bit clean, and parity detection is
9071 @item @code{auto-login} (default: @code{#f})
9072 When passed a login name, as a string, the specified user will be logged
9073 in automatically without prompting for their login name or password.
9075 @item @code{no-reset?} (default: @code{#f})
9076 When @code{#t}, don't reset terminal cflags (control modes).
9078 @item @code{host} (default: @code{#f})
9079 This accepts a string containing the "login_host", which will be written
9080 into the @file{/var/run/utmpx} file.
9082 @item @code{remote?} (default: @code{#f})
9083 When set to @code{#t} in conjunction with @var{host}, this will add an
9084 @code{-r} fakehost option to the command line of the login program
9085 specified in @var{login-program}.
9087 @item @code{flow-control?} (default: @code{#f})
9088 When set to @code{#t}, enable hardware (RTS/CTS) flow control.
9090 @item @code{no-issue?} (default: @code{#f})
9091 When set to @code{#t}, the contents of the @file{/etc/issue} file will
9092 not be displayed before presenting the login prompt.
9094 @item @code{init-string} (default: @code{#f})
9095 This accepts a string that will be sent to the tty or modem before
9096 sending anything else. It can be used to initialize a modem.
9098 @item @code{no-clear?} (default: @code{#f})
9099 When set to @code{#t}, agetty will not clear the screen before showing
9102 @item @code{login-program} (default: (file-append shadow "/bin/login"))
9103 This must be either a gexp denoting the name of a log-in program, or
9104 unset, in which case the default value is the @command{login} from the
9107 @item @code{local-line} (default: @code{#f})
9108 Control the CLOCAL line flag. This accepts one of three symbols as
9109 arguments, @code{'auto}, @code{'always}, or @code{'never}. If @code{#f},
9110 the default value chosen by agetty is @code{'auto}.
9112 @item @code{extract-baud?} (default: @code{#f})
9113 When set to @code{#t}, instruct agetty to try to extract the baud rate
9114 from the status messages produced by certain types of modems.
9116 @item @code{skip-login?} (default: @code{#f})
9117 When set to @code{#t}, do not prompt the user for a login name. This
9118 can be used with @var{login-program} field to use non-standard login
9121 @item @code{no-newline?} (default: @code{#f})
9122 When set to @code{#t}, do not print a newline before printing the
9123 @file{/etc/issue} file.
9125 @c Is this dangerous only when used with login-program, or always?
9126 @item @code{login-options} (default: @code{#f})
9127 This option accepts a string containing options that are passed to the
9128 login program. When used with the @var{login-program}, be aware that a
9129 malicious user could try to enter a login name containing embedded
9130 options that could be parsed by the login program.
9132 @item @code{login-pause} (default: @code{#f})
9133 When set to @code{#t}, wait for any key before showing the login prompt.
9134 This can be used in conjunction with @var{auto-login} to save memory by
9135 lazily spawning shells.
9137 @item @code{chroot} (default: @code{#f})
9138 Change root to the specified directory. This option accepts a directory
9141 @item @code{hangup?} (default: @code{#f})
9142 Use the Linux system call @code{vhangup} to do a virtual hangup of the
9145 @item @code{keep-baud?} (default: @code{#f})
9146 When set to @code{#t}, try to keep the existing baud rate. The baud
9147 rates from @var{baud-rate} are used when agetty receives a @key{BREAK}
9150 @item @code{timeout} (default: @code{#f})
9151 When set to an integer value, terminate if no user name could be read
9152 within @var{timeout} seconds.
9154 @item @code{detect-case?} (default: @code{#f})
9155 When set to @code{#t}, turn on support for detecting an uppercase-only
9156 terminal. This setting will detect a login name containing only
9157 uppercase letters as indicating an uppercase-only terminal and turn on
9158 some upper-to-lower case conversions. Note that this will not support
9161 @item @code{wait-cr?} (default: @code{#f})
9162 When set to @code{#t}, wait for the user or modem to send a
9163 carriage-return or linefeed character before displaying
9164 @file{/etc/issue} or login prompt. This is typically used with the
9165 @var{init-string} option.
9167 @item @code{no-hints?} (default: @code{#f})
9168 When set to @code{#t}, do not print hints about Num, Caps, and Scroll
9171 @item @code{no-hostname?} (default: @code{#f})
9172 By default, the hostname is printed. When this option is set to
9173 @code{#t}, no hostname will be shown at all.
9175 @item @code{long-hostname?} (default: @code{#f})
9176 By default, the hostname is only printed until the first dot. When this
9177 option is set to @code{#t}, the fully qualified hostname by
9178 @code{gethostname} or @code{getaddrinfo} is shown.
9180 @item @code{erase-characters} (default: @code{#f})
9181 This option accepts a string of additional characters that should be
9182 interpreted as backspace when the user types their login name.
9184 @item @code{kill-characters} (default: @code{#f})
9185 This option accepts a string that should be interpreted to mean "ignore
9186 all previous characters" (also called a "kill" character) when the types
9189 @item @code{chdir} (default: @code{#f})
9190 This option accepts, as a string, a directory path that will be changed
9193 @item @code{delay} (default: @code{#f})
9194 This options accepts, as an integer, the number of seconds to sleep
9195 before opening the tty and displaying the login prompt.
9197 @item @code{nice} (default: @code{#f})
9198 This option accepts, as an integer, the nice value with which to run the
9199 @command{login} program.
9201 @item @code{extra-options} (default: @code{'()})
9202 This option provides an "escape hatch" for the user to provide arbitrary
9203 command-line arguments to @command{agetty} as a list of strings.
9208 @deffn {Scheme Procedure} kmscon-service-type @var{config}
9209 Return a service to run @uref{https://www.freedesktop.org/wiki/Software/kmscon,kmscon}
9210 according to @var{config}, a @code{<kmscon-configuration>} object, which
9211 specifies the tty to run, among other things.
9214 @deftp {Data Type} kmscon-configuration
9215 This is the data type representing the configuration of Kmscon, which
9216 implements virtual console log-in.
9220 @item @code{virtual-terminal}
9221 The name of the console this Kmscon runs on---e.g., @code{"tty1"}.
9223 @item @code{login-program} (default: @code{#~(string-append #$shadow "/bin/login")})
9224 A gexp denoting the name of the log-in program. The default log-in program is
9225 @command{login} from the Shadow tool suite.
9227 @item @code{login-arguments} (default: @code{'("-p")})
9228 A list of arguments to pass to @command{login}.
9230 @item @code{hardware-acceleration?} (default: #f)
9231 Whether to use hardware acceleration.
9233 @item @code{kmscon} (default: @var{kmscon})
9234 The Kmscon package to use.
9239 @cindex name service cache daemon
9241 @deffn {Scheme Procedure} nscd-service [@var{config}] [#:glibc glibc] @
9242 [#:name-services '()]
9243 Return a service that runs the libc name service cache daemon (nscd) with the
9244 given @var{config}---an @code{<nscd-configuration>} object. @xref{Name
9245 Service Switch}, for an example.
9248 @defvr {Scheme Variable} %nscd-default-configuration
9249 This is the default @code{<nscd-configuration>} value (see below) used
9250 by @code{nscd-service}. It uses the caches defined by
9251 @var{%nscd-default-caches}; see below.
9254 @deftp {Data Type} nscd-configuration
9255 This is the data type representing the name service cache daemon (nscd)
9260 @item @code{name-services} (default: @code{'()})
9261 List of packages denoting @dfn{name services} that must be visible to
9262 the nscd---e.g., @code{(list @var{nss-mdns})}.
9264 @item @code{glibc} (default: @var{glibc})
9265 Package object denoting the GNU C Library providing the @command{nscd}
9268 @item @code{log-file} (default: @code{"/var/log/nscd.log"})
9269 Name of the nscd log file. This is where debugging output goes when
9270 @code{debug-level} is strictly positive.
9272 @item @code{debug-level} (default: @code{0})
9273 Integer denoting the debugging levels. Higher numbers mean that more
9274 debugging output is logged.
9276 @item @code{caches} (default: @var{%nscd-default-caches})
9277 List of @code{<nscd-cache>} objects denoting things to be cached; see
9283 @deftp {Data Type} nscd-cache
9284 Data type representing a cache database of nscd and its parameters.
9288 @item @code{database}
9289 This is a symbol representing the name of the database to be cached.
9290 Valid values are @code{passwd}, @code{group}, @code{hosts}, and
9291 @code{services}, which designate the corresponding NSS database
9292 (@pxref{NSS Basics,,, libc, The GNU C Library Reference Manual}).
9294 @item @code{positive-time-to-live}
9295 @itemx @code{negative-time-to-live} (default: @code{20})
9296 A number representing the number of seconds during which a positive or
9297 negative lookup result remains in cache.
9299 @item @code{check-files?} (default: @code{#t})
9300 Whether to check for updates of the files corresponding to
9303 For instance, when @var{database} is @code{hosts}, setting this flag
9304 instructs nscd to check for updates in @file{/etc/hosts} and to take
9307 @item @code{persistent?} (default: @code{#t})
9308 Whether the cache should be stored persistently on disk.
9310 @item @code{shared?} (default: @code{#t})
9311 Whether the cache should be shared among users.
9313 @item @code{max-database-size} (default: 32@tie{}MiB)
9314 Maximum size in bytes of the database cache.
9316 @c XXX: 'suggested-size' and 'auto-propagate?' seem to be expert
9317 @c settings, so leave them out.
9322 @defvr {Scheme Variable} %nscd-default-caches
9323 List of @code{<nscd-cache>} objects used by default by
9324 @code{nscd-configuration} (see above).
9326 It enables persistent and aggressive caching of service and host name
9327 lookups. The latter provides better host name lookup performance,
9328 resilience in the face of unreliable name servers, and also better
9329 privacy---often the result of host name lookups is in local cache, so
9330 external name servers do not even need to be queried.
9333 @anchor{syslog-configuration-type}
9336 @deftp {Data Type} syslog-configuration
9337 This data type represents the configuration of the syslog daemon.
9340 @item @code{syslogd} (default: @code{#~(string-append #$inetutils "/libexec/syslogd")})
9341 The syslog daemon to use.
9343 @item @code{config-file} (default: @code{%default-syslog.conf})
9344 The syslog configuration file to use.
9349 @anchor{syslog-service}
9351 @deffn {Scheme Procedure} syslog-service @var{config}
9352 Return a service that runs a syslog daemon according to @var{config}.
9354 @xref{syslogd invocation,,, inetutils, GNU Inetutils}, for more
9355 information on the configuration file syntax.
9358 @anchor{guix-configuration-type}
9359 @deftp {Data Type} guix-configuration
9360 This data type represents the configuration of the Guix build daemon.
9361 @xref{Invoking guix-daemon}, for more information.
9364 @item @code{guix} (default: @var{guix})
9365 The Guix package to use.
9367 @item @code{build-group} (default: @code{"guixbuild"})
9368 Name of the group for build user accounts.
9370 @item @code{build-accounts} (default: @code{10})
9371 Number of build user accounts to create.
9373 @item @code{authorize-key?} (default: @code{#t})
9374 @cindex substitutes, authorization thereof
9375 Whether to authorize the substitute keys listed in
9376 @code{authorized-keys}---by default that of @code{hydra.gnu.org}
9377 (@pxref{Substitutes}).
9379 @vindex %default-authorized-guix-keys
9380 @item @code{authorized-keys} (default: @var{%default-authorized-guix-keys})
9381 The list of authorized key files for archive imports, as a list of
9382 string-valued gexps (@pxref{Invoking guix archive}). By default, it
9383 contains that of @code{hydra.gnu.org} (@pxref{Substitutes}).
9385 @item @code{use-substitutes?} (default: @code{#t})
9386 Whether to use substitutes.
9388 @item @code{substitute-urls} (default: @var{%default-substitute-urls})
9389 The list of URLs where to look for substitutes by default.
9391 @item @code{max-silent-time} (default: @code{0})
9392 @itemx @code{timeout} (default: @code{0})
9393 The number of seconds of silence and the number of seconds of activity,
9394 respectively, after which a build process times out. A value of zero
9395 disables the timeout.
9397 @item @code{extra-options} (default: @code{'()})
9398 List of extra command-line options for @command{guix-daemon}.
9400 @item @code{log-file} (default: @code{"/var/log/guix-daemon.log"})
9401 File where @command{guix-daemon}'s standard output and standard error
9404 @item @code{lsof} (default: @var{lsof})
9405 The lsof package to use.
9407 @item @code{http-proxy} (default: @code{#f})
9408 The HTTP proxy used for downloading fixed-output derivations and
9411 @item @code{tmpdir} (default: @code{#f})
9412 A directory path where the @command{guix-daemon} will perform builds.
9417 @deffn {Scheme Procedure} guix-service @var{config}
9418 Return a service that runs the Guix build daemon according to
9422 @deffn {Scheme Procedure} udev-service [#:udev udev]
9423 Run @var{udev}, which populates the @file{/dev} directory dynamically.
9426 @deffn {Scheme Procedure} urandom-seed-service @var{#f}
9427 Save some entropy in @var{%random-seed-file} to seed @file{/dev/urandom}
9431 @defvr {Scheme Variable} %random-seed-file
9432 This is the name of the file where some random bytes are saved by
9433 @var{urandom-seed-service} to seed @file{/dev/urandom} when rebooting.
9434 It defaults to @file{/var/lib/random-seed}.
9439 @deffn {Scheme Procedure} console-keymap-service @var{files} ...
9440 @cindex keyboard layout
9441 Return a service to load console keymaps from @var{files} using
9442 @command{loadkeys} command. Most likely, you want to load some default
9443 keymap, which can be done like this:
9446 (console-keymap-service "dvorak")
9449 Or, for example, for a Swedish keyboard, you may need to combine
9450 the following keymaps:
9452 (console-keymap-service "se-lat6" "se-fi-lat6")
9455 Also you can specify a full file name (or file names) of your keymap(s).
9456 See @code{man loadkeys} for details.
9462 @deffn {Scheme Procedure} gpm-service [#:gpm @var{gpm}] @
9464 Run @var{gpm}, the general-purpose mouse daemon, with the given
9465 command-line @var{options}. GPM allows users to use the mouse in the console,
9466 notably to select, copy, and paste text. The default value of @var{options}
9467 uses the @code{ps2} protocol, which works for both USB and PS/2 mice.
9469 This service is not part of @var{%base-services}.
9472 @anchor{guix-publish-service-type}
9473 @deffn {Scheme Variable} guix-publish-service-type
9474 This is the service type for @command{guix publish} (@pxref{Invoking
9475 guix publish}). Its value must be a @code{guix-configuration}
9476 object, as described below.
9478 This assumes that @file{/etc/guix} already contains a signing key pair as
9479 created by @command{guix archive --generate-key} (@pxref{Invoking guix
9480 archive}). If that is not the case, the service will fail to start.
9483 @deftp {Data Type} guix-publish-configuration
9484 Data type representing the configuration of the @code{guix publish}
9488 @item @code{guix} (default: @code{guix})
9489 The Guix package to use.
9491 @item @code{port} (default: @code{80})
9492 The TCP port to listen for connections.
9494 @item @code{host} (default: @code{"localhost"})
9495 The host (and thus, network interface) to listen to. Use
9496 @code{"0.0.0.0"} to listen on all the network interfaces.
9498 @item @code{compression-level} (default: @code{3})
9499 The gzip compression level at which substitutes are compressed. Use
9500 @code{0} to disable compression altogether, and @code{9} to get the best
9501 compression ratio at the expense of increased CPU usage.
9503 @item @code{nar-path} (default: @code{"nar"})
9504 The URL path at which ``nars'' can be fetched. @xref{Invoking guix
9505 publish, @code{--nar-path}}, for details.
9507 @item @code{cache} (default: @code{#f})
9508 When it is @code{#f}, disable caching and instead generate archives on
9509 demand. Otherwise, this should be the name of a directory---e.g.,
9510 @code{"/var/cache/guix/publish"}---where @command{guix publish} caches
9511 archives and meta-data ready to be sent. @xref{Invoking guix publish,
9512 @option{--cache}}, for more information on the tradeoffs involved.
9514 @item @code{workers} (default: @code{#f})
9515 When it is an integer, this is the number of worker threads used for
9516 caching; when @code{#f}, the number of processors is used.
9517 @xref{Invoking guix publish, @option{--workers}}, for more information.
9519 @item @code{ttl} (default: @code{#f})
9520 When it is an integer, this denotes the @dfn{time-to-live} of the
9521 published archives. @xref{Invoking guix publish, @option{--ttl}}, for
9526 @anchor{rngd-service}
9527 @deffn {Scheme Procedure} rngd-service [#:rng-tools @var{rng-tools}] @
9528 [#:device "/dev/hwrng"]
9529 Return a service that runs the @command{rngd} program from @var{rng-tools}
9530 to add @var{device} to the kernel's entropy pool. The service will fail if
9531 @var{device} does not exist.
9534 @anchor{pam-limits-service}
9535 @cindex session limits
9538 @deffn {Scheme Procedure} pam-limits-service [#:limits @var{limits}]
9540 Return a service that installs a configuration file for the
9541 @uref{http://linux-pam.org/Linux-PAM-html/sag-pam_limits.html,
9542 @code{pam_limits} module}. The procedure optionally takes a list of
9543 @code{pam-limits-entry} values, which can be used to specify
9544 @code{ulimit} limits and nice priority limits to user sessions.
9546 The following limits definition sets two hard and soft limits for all
9547 login sessions of users in the @code{realtime} group:
9552 (pam-limits-entry "@@realtime" 'both 'rtprio 99)
9553 (pam-limits-entry "@@realtime" 'both 'memlock 'unlimited)))
9556 The first entry increases the maximum realtime priority for
9557 non-privileged processes; the second entry lifts any restriction of the
9558 maximum address space that can be locked in memory. These settings are
9559 commonly used for real-time audio systems.
9562 @node Scheduled Job Execution
9563 @subsubsection Scheduled Job Execution
9567 @cindex scheduling jobs
9568 The @code{(gnu services mcron)} module provides an interface to
9569 GNU@tie{}mcron, a daemon to run jobs at scheduled times (@pxref{Top,,,
9570 mcron, GNU@tie{}mcron}). GNU@tie{}mcron is similar to the traditional
9571 Unix @command{cron} daemon; the main difference is that it is
9572 implemented in Guile Scheme, which provides a lot of flexibility when
9573 specifying the scheduling of jobs and their actions.
9575 The example below defines an operating system that runs the
9576 @command{updatedb} (@pxref{Invoking updatedb,,, find, Finding Files})
9577 and the @command{guix gc} commands (@pxref{Invoking guix gc}) daily, as
9578 well as the @command{mkid} command on behalf of an unprivileged user
9579 (@pxref{mkid invocation,,, idutils, ID Database Utilities}). It uses
9580 gexps to introduce job definitions that are passed to mcron
9581 (@pxref{G-Expressions}).
9584 (use-modules (guix) (gnu) (gnu services mcron))
9585 (use-package-modules base idutils)
9587 (define updatedb-job
9588 ;; Run 'updatedb' at 3AM every day. Here we write the
9589 ;; job's action as a Scheme procedure.
9590 #~(job '(next-hour '(3))
9592 (execl (string-append #$findutils "/bin/updatedb")
9594 "--prunepaths=/tmp /var/tmp /gnu/store"))))
9596 (define garbage-collector-job
9597 ;; Collect garbage 5 minutes after midnight every day.
9598 ;; The job's action is a shell command.
9599 #~(job "5 0 * * *" ;Vixie cron syntax
9603 ;; Update the index database as user "charlie" at 12:15PM
9604 ;; and 19:15PM. This runs from the user's home directory.
9605 #~(job '(next-minute-from (next-hour '(12 19)) '(15))
9606 (string-append #$idutils "/bin/mkid src")
9611 (services (cons (mcron-service (list garbage-collector-job
9617 @xref{Guile Syntax, mcron job specifications,, mcron, GNU@tie{}mcron},
9618 for more information on mcron job specifications. Below is the
9619 reference of the mcron service.
9621 @deffn {Scheme Procedure} mcron-service @var{jobs} [#:mcron @var{mcron2}]
9622 Return an mcron service running @var{mcron} that schedules @var{jobs}, a
9623 list of gexps denoting mcron job specifications.
9625 This is a shorthand for:
9627 (service mcron-service-type
9628 (mcron-configuration (mcron mcron) (jobs jobs)))
9632 @defvr {Scheme Variable} mcron-service-type
9633 This is the type of the @code{mcron} service, whose value is an
9634 @code{mcron-configuration} object.
9636 This service type can be the target of a service extension that provides
9637 it additional job specifications (@pxref{Service Composition}). In
9638 other words, it is possible to define services that provide additional
9642 @deftp {Data Type} mcron-configuration
9643 Data type representing the configuration of mcron.
9646 @item @code{mcron} (default: @var{mcron2})
9647 The mcron package to use.
9650 This is a list of gexps (@pxref{G-Expressions}), where each gexp
9651 corresponds to an mcron job specification (@pxref{Syntax, mcron job
9652 specifications,, mcron, GNU@tie{}mcron}).
9658 @subsubsection Log Rotation
9661 @cindex log rotation
9663 Log files such as those found in @file{/var/log} tend to grow endlessly,
9664 so it's a good idea to @dfn{rotate} them once in a while---i.e., archive
9665 their contents in separate files, possibly compressed. The @code{(gnu
9666 services admin)} module provides an interface to GNU@tie{}Rot[t]log, a
9667 log rotation tool (@pxref{Top,,, rottlog, GNU Rot[t]log Manual}).
9669 The example below defines an operating system that provides log rotation
9670 with the default settings, for commonly encountered log files.
9673 (use-modules (guix) (gnu))
9674 (use-service-modules admin mcron)
9675 (use-package-modules base idutils)
9679 (services (cons* (mcron-service)
9680 (service rottlog-service-type)
9684 @defvr {Scheme Variable} rottlog-service-type
9685 This is the type of the Rottlog service, whose value is a
9686 @code{rottlog-configuration} object.
9688 Other services can extend this one with new @code{log-rotation} objects
9689 (see below), thereby augmenting the set of files to be rotated.
9691 This service type can define mcron jobs (@pxref{Scheduled Job
9692 Execution}) to run the rottlog service.
9695 @deftp {Data Type} rottlog-configuration
9696 Data type representing the configuration of rottlog.
9699 @item @code{rottlog} (default: @code{rottlog})
9700 The Rottlog package to use.
9702 @item @code{rc-file} (default: @code{(file-append rottlog "/etc/rc")})
9703 The Rottlog configuration file to use (@pxref{Mandatory RC Variables,,,
9704 rottlog, GNU Rot[t]log Manual}).
9706 @item @code{rotations} (default: @code{%default-rotations})
9707 A list of @code{log-rotation} objects as defined below.
9710 This is a list of gexps where each gexp corresponds to an mcron job
9711 specification (@pxref{Scheduled Job Execution}).
9715 @deftp {Data Type} log-rotation
9716 Data type representing the rotation of a group of log files.
9718 Taking an example from the Rottlog manual (@pxref{Period Related File
9719 Examples,,, rottlog, GNU Rot[t]log Manual}), a log rotation might be
9725 (files '("/var/log/apache/*"))
9726 (options '("storedir apache-archives"
9732 The list of fields is as follows:
9735 @item @code{frequency} (default: @code{'weekly})
9736 The log rotation frequency, a symbol.
9739 The list of files or file glob patterns to rotate.
9741 @item @code{options} (default: @code{'()})
9742 The list of rottlog options for this rotation (@pxref{Configuration
9743 parameters,,, rottlog, GNU Rot[t]lg Manual}).
9745 @item @code{post-rotate} (default: @code{#f})
9746 Either @code{#f} or a gexp to execute once the rotation has completed.
9750 @defvr {Scheme Variable} %default-rotations
9751 Specifies weekly rotation of @var{%rotated-files} and
9752 a couple of other files.
9755 @defvr {Scheme Variable} %rotated-files
9756 The list of syslog-controlled files to be rotated. By default it is:
9757 @code{'("/var/log/messages" "/var/log/secure")}.
9760 @node Networking Services
9761 @subsubsection Networking Services
9763 The @code{(gnu services networking)} module provides services to configure
9764 the network interface.
9766 @cindex DHCP, networking service
9767 @deffn {Scheme Procedure} dhcp-client-service [#:dhcp @var{isc-dhcp}]
9768 Return a service that runs @var{dhcp}, a Dynamic Host Configuration
9769 Protocol (DHCP) client, on all the non-loopback network interfaces.
9772 @defvr {Scheme Variable} static-networking-service-type
9773 This is the type for statically-configured network interfaces.
9774 @c TODO Document <static-networking> data structures.
9777 @deffn {Scheme Procedure} static-networking-service @var{interface} @var{ip} @
9778 [#:netmask #f] [#:gateway #f] [#:name-servers @code{'()}]
9779 Return a service that starts @var{interface} with address @var{ip}. If
9780 @var{netmask} is true, use it as the network mask. If @var{gateway} is true,
9781 it must be a string specifying the default network gateway.
9783 This procedure can be called several times, one for each network
9784 interface of interest. Behind the scenes what it does is extend
9785 @code{static-networking-service-type} with additional network interfaces
9792 @cindex network management
9793 @deffn {Scheme Procedure} wicd-service [#:wicd @var{wicd}]
9794 Return a service that runs @url{https://launchpad.net/wicd,Wicd}, a network
9795 management daemon that aims to simplify wired and wireless networking.
9797 This service adds the @var{wicd} package to the global profile, providing
9798 several commands to interact with the daemon and configure networking:
9799 @command{wicd-client}, a graphical user interface, and the @command{wicd-cli}
9800 and @command{wicd-curses} user interfaces.
9803 @cindex NetworkManager
9805 @defvr {Scheme Variable} network-manager-service-type
9806 This is the service type for the
9807 @uref{https://wiki.gnome.org/Projects/NetworkManager, NetworkManager}
9808 service. The value for this service type is a
9809 @code{network-manager-configuration} record.
9812 @deftp {Data Type} network-manager-configuration
9813 Data type representing the configuration of NetworkManager.
9816 @item @code{network-manager} (default: @code{network-manager})
9817 The NetworkManager package to use.
9819 @item @code{dns} (default: @code{"default"})
9820 Processing mode for DNS, which affects how NetworkManager uses the
9821 @code{resolv.conf} configuration file.
9825 NetworkManager will update @code{resolv.conf} to reflect the nameservers
9826 provided by currently active connections.
9829 NetworkManager will run @code{dnsmasq} as a local caching nameserver,
9830 using a "split DNS" configuration if you are connected to a VPN, and
9831 then update @code{resolv.conf} to point to the local nameserver.
9834 NetworkManager will not modify @code{resolv.conf}.
9841 @deffn {Scheme Variable} connman-service-type
9842 This is the service type to run @url{https://01.org/connman,Connman},
9843 a network connection manager.
9845 Its value must be an
9846 @code{connman-configuration} record as in this example:
9849 (service connman-service-type
9850 (connman-configuration
9854 See below for details about @code{connman-configuration}.
9857 @deftp {Data Type} connman-configuration
9858 Data Type representing the configuration of connman.
9861 @item @code{connman} (default: @var{connman})
9862 The connman package to use.
9864 @item @code{disable-vpn?} (default: @code{#f})
9865 When true, enable connman's vpn plugin.
9869 @cindex WPA Supplicant
9870 @defvr {Scheme Variable} wpa-supplicant-service-type
9871 This is the service type to run @url{https://w1.fi/wpa_supplicant/,WPA
9872 supplicant}, an authentication daemon required to authenticate against
9873 encrypted WiFi or ethernet networks. It is configured to listen for
9876 The value of this service is the @code{wpa-supplicant} package to use.
9877 Thus, it can be instantiated like this:
9880 (use-modules (gnu services networking))
9882 (service wpa-supplicant-service-type)
9887 @cindex real time clock
9888 @deffn {Scheme Procedure} ntp-service [#:ntp @var{ntp}] @
9889 [#:servers @var{%ntp-servers}] @
9890 [#:allow-large-adjustment? #f]
9891 Return a service that runs the daemon from @var{ntp}, the
9892 @uref{http://www.ntp.org, Network Time Protocol package}. The daemon will
9893 keep the system clock synchronized with that of @var{servers}.
9894 @var{allow-large-adjustment?} determines whether @command{ntpd} is allowed to
9895 make an initial adjustment of more than 1,000 seconds.
9898 @defvr {Scheme Variable} %ntp-servers
9899 List of host names used as the default NTP servers.
9903 @deffn {Scheme variable} inetd-service-type
9904 This service runs the @command{inetd} (@pxref{inetd invocation,,,
9905 inetutils, GNU Inetutils}) daemon. @command{inetd} listens for
9906 connections on internet sockets, and lazily starts the specified server
9907 program when a connection is made on one of these sockets.
9909 The value of this service is an @code{inetd-configuration} object. The
9910 following example configures the @command{inetd} daemon to provide the
9911 built-in @command{echo} service, as well as an smtp service which
9912 forwards smtp traffic over ssh to a server @code{smtp-server} behind a
9913 gateway @code{hostname}:
9918 (inetd-configuration
9922 (socket-type 'stream)
9929 (socket-type 'stream)
9933 (program (file-append openssh "/bin/ssh"))
9935 '("ssh" "-qT" "-i" "/path/to/ssh_key"
9936 "-W" "smtp-server:25" "user@@hostname")))))
9939 See below for more details about @code{inetd-configuration}.
9942 @deftp {Data Type} inetd-configuration
9943 Data type representing the configuration of @command{inetd}.
9946 @item @code{program} (default: @code{(file-append inetutils "/libexec/inetd")})
9947 The @command{inetd} executable to use.
9949 @item @code{entries} (default: @code{'()})
9950 A list of @command{inetd} service entries. Each entry should be created
9951 by the @code{inetd-entry} constructor.
9955 @deftp {Data Type} inetd-entry
9956 Data type representing an entry in the @command{inetd} configuration.
9957 Each entry corresponds to a socket where @command{inetd} will listen for
9961 @item @code{node} (default: @code{#f})
9962 Optional string, a comma-separated list of local addresses
9963 @command{inetd} should use when listening for this service.
9964 @xref{Configuration file,,, inetutils, GNU Inetutils} for a complete
9965 description of all options.
9967 A string, the name must correspond to an entry in @code{/etc/services}.
9968 @item @code{socket-type}
9969 One of @code{'stream}, @code{'dgram}, @code{'raw}, @code{'rdm} or
9971 @item @code{protocol}
9972 A string, must correspond to an entry in @code{/etc/protocols}.
9973 @item @code{wait?} (default: @code{#t})
9974 Whether @command{inetd} should wait for the server to exit before
9975 listening to new service requests.
9977 A string containing the user (and, optionally, group) name of the user
9978 as whom the server should run. The group name can be specified in a
9979 suffix, separated by a colon or period, i.e. @code{"user"},
9980 @code{"user:group"} or @code{"user.group"}.
9981 @item @code{program} (default: @code{"internal"})
9982 The server program which will serve the requests, or @code{"internal"}
9983 if @command{inetd} should use a built-in service.
9984 @item @code{arguments} (default: @code{'()})
9985 A list strings or file-like objects, which are the server program's
9986 arguments, starting with the zeroth argument, i.e. the name of the
9987 program itself. For @command{inetd}'s internal services, this entry
9988 must be @code{'()} or @code{'("internal")}.
9991 @xref{Configuration file,,, inetutils, GNU Inetutils} for a more
9992 detailed discussion of each configuration field.
9996 @deffn {Scheme Procedure} tor-service [@var{config-file}] [#:tor @var{tor}]
9997 Return a service to run the @uref{https://torproject.org, Tor} anonymous
10000 The daemon runs as the @code{tor} unprivileged user. It is passed
10001 @var{config-file}, a file-like object, with an additional @code{User tor} line
10002 and lines for hidden services added via @code{tor-hidden-service}. Run
10003 @command{man tor} for information about the configuration file.
10006 @cindex hidden service
10007 @deffn {Scheme Procedure} tor-hidden-service @var{name} @var{mapping}
10008 Define a new Tor @dfn{hidden service} called @var{name} and implementing
10009 @var{mapping}. @var{mapping} is a list of port/host tuples, such as:
10012 '((22 "127.0.0.1:22")
10013 (80 "127.0.0.1:8080"))
10016 In this example, port 22 of the hidden service is mapped to local port 22, and
10017 port 80 is mapped to local port 8080.
10019 This creates a @file{/var/lib/tor/hidden-services/@var{name}} directory, where
10020 the @file{hostname} file contains the @code{.onion} host name for the hidden
10023 See @uref{https://www.torproject.org/docs/tor-hidden-service.html.en, the Tor
10024 project's documentation} for more information.
10027 @deffn {Scheme Procedure} bitlbee-service [#:bitlbee bitlbee] @
10028 [#:interface "127.0.0.1"] [#:port 6667] @
10029 [#:extra-settings ""]
10030 Return a service that runs @url{http://bitlbee.org,BitlBee}, a daemon that
10031 acts as a gateway between IRC and chat networks.
10033 The daemon will listen to the interface corresponding to the IP address
10034 specified in @var{interface}, on @var{port}. @code{127.0.0.1} means that only
10035 local clients can connect, whereas @code{0.0.0.0} means that connections can
10036 come from any networking interface.
10038 In addition, @var{extra-settings} specifies a string to append to the
10039 configuration file.
10042 Furthermore, @code{(gnu services ssh)} provides the following services.
10046 @deffn {Scheme Procedure} lsh-service [#:host-key "/etc/lsh/host-key"] @
10047 [#:daemonic? #t] [#:interfaces '()] [#:port-number 22] @
10048 [#:allow-empty-passwords? #f] [#:root-login? #f] @
10049 [#:syslog-output? #t] [#:x11-forwarding? #t] @
10050 [#:tcp/ip-forwarding? #t] [#:password-authentication? #t] @
10051 [#:public-key-authentication? #t] [#:initialize? #t]
10052 Run the @command{lshd} program from @var{lsh} to listen on port @var{port-number}.
10053 @var{host-key} must designate a file containing the host key, and readable
10056 When @var{daemonic?} is true, @command{lshd} will detach from the
10057 controlling terminal and log its output to syslogd, unless one sets
10058 @var{syslog-output?} to false. Obviously, it also makes lsh-service
10059 depend on existence of syslogd service. When @var{pid-file?} is true,
10060 @command{lshd} writes its PID to the file called @var{pid-file}.
10062 When @var{initialize?} is true, automatically create the seed and host key
10063 upon service activation if they do not exist yet. This may take long and
10064 require interaction.
10066 When @var{initialize?} is false, it is up to the user to initialize the
10067 randomness generator (@pxref{lsh-make-seed,,, lsh, LSH Manual}), and to create
10068 a key pair with the private key stored in file @var{host-key} (@pxref{lshd
10069 basics,,, lsh, LSH Manual}).
10071 When @var{interfaces} is empty, lshd listens for connections on all the
10072 network interfaces; otherwise, @var{interfaces} must be a list of host names
10075 @var{allow-empty-passwords?} specifies whether to accept log-ins with empty
10076 passwords, and @var{root-login?} specifies whether to accept log-ins as
10079 The other options should be self-descriptive.
10084 @deffn {Scheme Variable} openssh-service-type
10085 This is the type for the @uref{http://www.openssh.org, OpenSSH} secure
10086 shell daemon, @command{sshd}. Its value must be an
10087 @code{openssh-configuration} record as in this example:
10090 (service openssh-service-type
10091 (openssh-configuration
10092 (x11-forwarding? #t)
10093 (permit-root-login 'without-password)))
10096 See below for details about @code{openssh-configuration}.
10099 @deftp {Data Type} openssh-configuration
10100 This is the configuration record for OpenSSH's @command{sshd}.
10103 @item @code{pid-file} (default: @code{"/var/run/sshd.pid"})
10104 Name of the file where @command{sshd} writes its PID.
10106 @item @code{port-number} (default: @code{22})
10107 TCP port on which @command{sshd} listens for incoming connections.
10109 @item @code{permit-root-login} (default: @code{#f})
10110 This field determines whether and when to allow logins as root. If
10111 @code{#f}, root logins are disallowed; if @code{#t}, they are allowed.
10112 If it's the symbol @code{'without-password}, then root logins are
10113 permitted but not with password-based authentication.
10115 @item @code{allow-empty-passwords?} (default: @code{#f})
10116 When true, users with empty passwords may log in. When false, they may
10119 @item @code{password-authentication?} (default: @code{#t})
10120 When true, users may log in with their password. When false, they have
10121 other authentication methods.
10123 @item @code{public-key-authentication?} (default: @code{#t})
10124 When true, users may log in using public key authentication. When
10125 false, users have to use other authentication method.
10127 Authorized public keys are stored in @file{~/.ssh/authorized_keys}.
10128 This is used only by protocol version 2.
10130 @item @code{x11-forwarding?} (default: @code{#f})
10131 When true, forwarding of X11 graphical client connections is
10132 enabled---in other words, @command{ssh} options @option{-X} and
10133 @option{-Y} will work.
10135 @item @code{challenge-response-authentication?} (default: @code{#f})
10136 Specifies whether challenge response authentication is allowed (e.g. via
10139 @item @code{use-pam?} (default: @code{#t})
10140 Enables the Pluggable Authentication Module interface. If set to
10141 @code{#t}, this will enable PAM authentication using
10142 @code{challenge-response-authentication?} and
10143 @code{password-authentication?}, in addition to PAM account and session
10144 module processing for all authentication types.
10146 Because PAM challenge response authentication usually serves an
10147 equivalent role to password authentication, you should disable either
10148 @code{challenge-response-authentication?} or
10149 @code{password-authentication?}.
10151 @item @code{print-last-log?} (default: @code{#t})
10152 Specifies whether @command{sshd} should print the date and time of the
10153 last user login when a user logs in interactively.
10155 @item @code{subsystems} (default: @code{'(("sftp" "internal-sftp"))})
10156 Configures external subsystems (e.g. file transfer daemon).
10158 This is a list of two-element lists, each of which containing the
10159 subsystem name and a command (with optional arguments) to execute upon
10162 The command @command{internal-sftp} implements an in-process SFTP
10163 server. Alternately, one can specify the @command{sftp-server} command:
10165 (service openssh-service-type
10166 (openssh-configuration
10168 '(("sftp" ,(file-append openssh "/libexec/sftp-server"))))))
10173 @deffn {Scheme Procedure} dropbear-service [@var{config}]
10174 Run the @uref{https://matt.ucc.asn.au/dropbear/dropbear.html,Dropbear SSH
10175 daemon} with the given @var{config}, a @code{<dropbear-configuration>}
10178 For example, to specify a Dropbear service listening on port 1234, add
10179 this call to the operating system's @code{services} field:
10182 (dropbear-service (dropbear-configuration
10183 (port-number 1234)))
10187 @deftp {Data Type} dropbear-configuration
10188 This data type represents the configuration of a Dropbear SSH daemon.
10191 @item @code{dropbear} (default: @var{dropbear})
10192 The Dropbear package to use.
10194 @item @code{port-number} (default: 22)
10195 The TCP port where the daemon waits for incoming connections.
10197 @item @code{syslog-output?} (default: @code{#t})
10198 Whether to enable syslog output.
10200 @item @code{pid-file} (default: @code{"/var/run/dropbear.pid"})
10201 File name of the daemon's PID file.
10203 @item @code{root-login?} (default: @code{#f})
10204 Whether to allow @code{root} logins.
10206 @item @code{allow-empty-passwords?} (default: @code{#f})
10207 Whether to allow empty passwords.
10209 @item @code{password-authentication?} (default: @code{#t})
10210 Whether to enable password-based authentication.
10214 @defvr {Scheme Variable} %facebook-host-aliases
10215 This variable contains a string for use in @file{/etc/hosts}
10216 (@pxref{Host Names,,, libc, The GNU C Library Reference Manual}). Each
10217 line contains a entry that maps a known server name of the Facebook
10218 on-line service---e.g., @code{www.facebook.com}---to the local
10219 host---@code{127.0.0.1} or its IPv6 equivalent, @code{::1}.
10221 This variable is typically used in the @code{hosts-file} field of an
10222 @code{operating-system} declaration (@pxref{operating-system Reference,
10223 @file{/etc/hosts}}):
10226 (use-modules (gnu) (guix))
10229 (host-name "mymachine")
10232 ;; Create a /etc/hosts file with aliases for "localhost"
10233 ;; and "mymachine", as well as for Facebook servers.
10234 (plain-file "hosts"
10235 (string-append (local-host-aliases host-name)
10236 %facebook-host-aliases))))
10239 This mechanism can prevent programs running locally, such as Web
10240 browsers, from accessing Facebook.
10243 The @code{(gnu services avahi)} provides the following definition.
10245 @deffn {Scheme Procedure} avahi-service [#:avahi @var{avahi}] @
10246 [#:host-name #f] [#:publish? #t] [#:ipv4? #t] @
10247 [#:ipv6? #t] [#:wide-area? #f] @
10248 [#:domains-to-browse '()] [#:debug? #f]
10249 Return a service that runs @command{avahi-daemon}, a system-wide
10250 mDNS/DNS-SD responder that allows for service discovery and
10251 "zero-configuration" host name lookups (see @uref{http://avahi.org/}), and
10252 extends the name service cache daemon (nscd) so that it can resolve
10253 @code{.local} host names using
10254 @uref{http://0pointer.de/lennart/projects/nss-mdns/, nss-mdns}. Additionally,
10255 add the @var{avahi} package to the system profile so that commands such as
10256 @command{avahi-browse} are directly usable.
10258 If @var{host-name} is different from @code{#f}, use that as the host name to
10259 publish for this machine; otherwise, use the machine's actual host name.
10261 When @var{publish?} is true, publishing of host names and services is allowed;
10262 in particular, avahi-daemon will publish the machine's host name and IP
10263 address via mDNS on the local network.
10265 When @var{wide-area?} is true, DNS-SD over unicast DNS is enabled.
10267 Boolean values @var{ipv4?} and @var{ipv6?} determine whether to use IPv4/IPv6
10271 @deffn {Scheme Variable} openvswitch-service-type
10272 This is the type of the @uref{http://www.openvswitch.org, Open vSwitch}
10273 service, whose value should be an @code{openvswitch-configuration}
10277 @deftp {Data Type} openvswitch-configuration
10278 Data type representing the configuration of Open vSwitch, a multilayer
10279 virtual switch which is designed to enable massive network automation
10280 through programmatic extension.
10283 @item @code{package} (default: @var{openvswitch})
10284 Package object of the Open vSwitch.
10290 @subsubsection X Window
10293 @cindex X Window System
10294 Support for the X Window graphical display system---specifically
10295 Xorg---is provided by the @code{(gnu services xorg)} module. Note that
10296 there is no @code{xorg-service} procedure. Instead, the X server is
10297 started by the @dfn{login manager}, currently SLiM.
10299 @deftp {Data Type} sddm-configuration
10300 This is the data type representing the sddm service configuration.
10303 @item @code{display-server} (default: "x11")
10304 Select display server to use for the greeter. Valid values are "x11"
10307 @item @code{numlock} (default: "on")
10308 Valid values are "on", "off" or "none".
10310 @item @code{halt-command} (default @code{#~(string-apppend #$shepherd "/sbin/halt")})
10311 Command to run when halting.
10313 @item @code{reboot-command} (default @code{#~(string-append #$shepherd "/sbin/reboot")})
10314 Command to run when rebooting.
10316 @item @code{theme} (default "maldives")
10317 Theme to use. Default themes provided by SDDM are "elarun" or "maldives".
10319 @item @code{themes-directory} (default "/run/current-system/profile/share/sddm/themes")
10320 Directory to look for themes.
10322 @item @code{faces-directory} (default "/run/current-system/profile/share/sddm/faces")
10323 Directory to look for faces.
10325 @item @code{default-path} (default "/run/current-system/profile/bin")
10326 Default PATH to use.
10328 @item @code{minimum-uid} (default 1000)
10329 Minimum UID to display in SDDM.
10331 @item @code{maximum-uid} (default 2000)
10332 Maximum UID to display in SDDM
10334 @item @code{remember-last-user?} (default #t)
10335 Remember last user.
10337 @item @code{remember-last-session?} (default #t)
10338 Remember last session.
10340 @item @code{hide-users} (default "")
10341 Usernames to hide from SDDM greeter.
10343 @item @code{hide-shells} (default @code{#~(string-append #$shadow "/sbin/nologin")})
10344 Users with shells listed will be hidden from the SDDM greeter.
10346 @item @code{session-command} (default @code{#~(string-append #$sddm "/share/sddm/scripts/wayland-session")})
10347 Script to run before starting a wayland session.
10349 @item @code{sessions-directory} (default "/run/current-system/profile/share/wayland-sessions")
10350 Directory to look for desktop files starting wayland sessions.
10352 @item @code{xorg-server-path} (default @code{xorg-start-command})
10353 Path to xorg-server.
10355 @item @code{xauth-path} (default @code{#~(string-append #$xauth "/bin/xauth")})
10358 @item @code{xephyr-path} (default @code{#~(string-append #$xorg-server "/bin/Xephyr")})
10361 @item @code{xdisplay-start} (default @code{#~(string-append #$sddm "/share/sddm/scripts/Xsetup")})
10362 Script to run after starting xorg-server.
10364 @item @code{xdisplay-stop} (default @code{#~(string-append #$sddm "/share/sddm/scripts/Xstop")})
10365 Script to run before stopping xorg-server.
10367 @item @code{xsession-command} (default: @code{xinitr })
10368 Script to run before starting a X session.
10370 @item @code{xsessions-directory} (default: "/run/current-system/profile/share/xsessions")
10371 Directory to look for desktop files starting X sessions.
10373 @item @code{minimum-vt} (default: 7)
10376 @item @code{xserver-arguments} (default "-nolisten tcp")
10377 Arguments to pass to xorg-server.
10379 @item @code{auto-login-user} (default "")
10380 User to use for auto-login.
10382 @item @code{auto-login-session} (default "")
10383 Desktop file to use for auto-login.
10385 @item @code{relogin?} (default #f)
10386 Relogin after logout.
10391 @cindex login manager
10392 @deffn {Scheme Procedure} sddm-service config
10393 Return a service that spawns the SDDM graphical login manager for config of
10394 type @code{<sddm-configuration>}.
10397 (sddm-service (sddm-configuration
10398 (auto-login-user "Alice")
10399 (auto-login-session "xfce.desktop")))
10403 @deffn {Scheme Procedure} slim-service [#:allow-empty-passwords? #f] @
10404 [#:auto-login? #f] [#:default-user ""] [#:startx] @
10405 [#:theme @var{%default-slim-theme}] @
10406 [#:theme-name @var{%default-slim-theme-name}]
10407 Return a service that spawns the SLiM graphical login manager, which in
10408 turn starts the X display server with @var{startx}, a command as returned by
10409 @code{xorg-start-command}.
10413 SLiM automatically looks for session types described by the @file{.desktop}
10414 files in @file{/run/current-system/profile/share/xsessions} and allows users
10415 to choose a session from the log-in screen using @kbd{F1}. Packages such as
10416 @var{xfce}, @var{sawfish}, and @var{ratpoison} provide @file{.desktop} files;
10417 adding them to the system-wide set of packages automatically makes them
10418 available at the log-in screen.
10420 In addition, @file{~/.xsession} files are honored. When available,
10421 @file{~/.xsession} must be an executable that starts a window manager
10422 and/or other X clients.
10424 When @var{allow-empty-passwords?} is true, allow logins with an empty
10425 password. When @var{auto-login?} is true, log in automatically as
10426 @var{default-user}.
10428 If @var{theme} is @code{#f}, use the default log-in theme; otherwise
10429 @var{theme} must be a gexp denoting the name of a directory containing the
10430 theme to use. In that case, @var{theme-name} specifies the name of the
10434 @defvr {Scheme Variable} %default-theme
10435 @defvrx {Scheme Variable} %default-theme-name
10436 The G-Expression denoting the default SLiM theme and its name.
10439 @deffn {Scheme Procedure} xorg-start-command [#:guile] @
10440 [#:configuration-file #f] [#:xorg-server @var{xorg-server}]
10441 Return a derivation that builds a @var{guile} script to start the X server
10442 from @var{xorg-server}. @var{configuration-file} is the server configuration
10443 file or a derivation that builds it; when omitted, the result of
10444 @code{xorg-configuration-file} is used.
10446 Usually the X server is started by a login manager.
10449 @deffn {Scheme Procedure} xorg-configuration-file @
10450 [#:drivers '()] [#:resolutions '()] [#:extra-config '()]
10451 Return a configuration file for the Xorg server containing search paths for
10452 all the common drivers.
10454 @var{drivers} must be either the empty list, in which case Xorg chooses a
10455 graphics driver automatically, or a list of driver names that will be tried in
10456 this order---e.g., @code{(\"modesetting\" \"vesa\")}.
10458 Likewise, when @var{resolutions} is the empty list, Xorg chooses an
10459 appropriate screen resolution; otherwise, it must be a list of
10460 resolutions---e.g., @code{((1024 768) (640 480))}.
10462 Last, @var{extra-config} is a list of strings or objects appended to the
10463 @code{text-file*} argument list. It is used to pass extra text to be added
10464 verbatim to the configuration file.
10467 @deffn {Scheme Procedure} screen-locker-service @var{package} [@var{name}]
10468 Add @var{package}, a package for a screen-locker or screen-saver whose
10469 command is @var{program}, to the set of setuid programs and add a PAM entry
10470 for it. For example:
10473 (screen-locker-service xlockmore "xlock")
10476 makes the good ol' XlockMore usable.
10480 @node Printing Services
10481 @subsubsection Printing Services
10483 @cindex printer support with CUPS
10484 The @code{(gnu services cups)} module provides a Guix service definition
10485 for the CUPS printing service. To add printer support to a GuixSD
10486 system, add a @code{cups-service} to the operating system definition:
10488 @deffn {Scheme Variable} cups-service-type
10489 The service type for the CUPS print server. Its value should be a valid
10490 CUPS configuration (see below). To use the default settings, simply
10493 (service cups-service-type)
10497 The CUPS configuration controls the basic things about your CUPS
10498 installation: what interfaces it listens on, what to do if a print job
10499 fails, how much logging to do, and so on. To actually add a printer,
10500 you have to visit the @url{http://localhost:631} URL, or use a tool such
10501 as GNOME's printer configuration services. By default, configuring a
10502 CUPS service will generate a self-signed certificate if needed, for
10503 secure connections to the print server.
10505 Suppose you want to enable the Web interface of CUPS and also add
10506 support for HP printers @i{via} the @code{hplip} package. You can do
10507 that directly, like this (you need to use the @code{(gnu packages cups)}
10511 (service cups-service-type
10512 (cups-configuration
10513 (web-interface? #t)
10515 (list cups-filters hplip))))
10518 The available configuration parameters follow. Each parameter
10519 definition is preceded by its type; for example, @samp{string-list foo}
10520 indicates that the @code{foo} parameter should be specified as a list of
10521 strings. There is also a way to specify the configuration as a string,
10522 if you have an old @code{cupsd.conf} file that you want to port over
10523 from some other system; see the end for more details.
10525 @c The following documentation was initially generated by
10526 @c (generate-documentation) in (gnu services cups). Manually maintained
10527 @c documentation is better, so we shouldn't hesitate to edit below as
10528 @c needed. However if the change you want to make to this documentation
10529 @c can be done in an automated way, it's probably easier to change
10530 @c (generate-documentation) than to make it below and have to deal with
10531 @c the churn as CUPS updates.
10534 Available @code{cups-configuration} fields are:
10536 @deftypevr {@code{cups-configuration} parameter} package cups
10540 @deftypevr {@code{cups-configuration} parameter} package-list extensions
10541 Drivers and other extensions to the CUPS package.
10544 @deftypevr {@code{cups-configuration} parameter} files-configuration files-configuration
10545 Configuration of where to write logs, what directories to use for print
10546 spools, and related privileged configuration parameters.
10548 Available @code{files-configuration} fields are:
10550 @deftypevr {@code{files-configuration} parameter} log-location access-log
10551 Defines the access log filename. Specifying a blank filename disables
10552 access log generation. The value @code{stderr} causes log entries to be
10553 sent to the standard error file when the scheduler is running in the
10554 foreground, or to the system log daemon when run in the background. The
10555 value @code{syslog} causes log entries to be sent to the system log
10556 daemon. The server name may be included in filenames using the string
10557 @code{%s}, as in @code{/var/log/cups/%s-access_log}.
10559 Defaults to @samp{"/var/log/cups/access_log"}.
10562 @deftypevr {@code{files-configuration} parameter} file-name cache-dir
10563 Where CUPS should cache data.
10565 Defaults to @samp{"/var/cache/cups"}.
10568 @deftypevr {@code{files-configuration} parameter} string config-file-perm
10569 Specifies the permissions for all configuration files that the scheduler
10572 Note that the permissions for the printers.conf file are currently
10573 masked to only allow access from the scheduler user (typically root).
10574 This is done because printer device URIs sometimes contain sensitive
10575 authentication information that should not be generally known on the
10576 system. There is no way to disable this security feature.
10578 Defaults to @samp{"0640"}.
10581 @deftypevr {@code{files-configuration} parameter} log-location error-log
10582 Defines the error log filename. Specifying a blank filename disables
10583 access log generation. The value @code{stderr} causes log entries to be
10584 sent to the standard error file when the scheduler is running in the
10585 foreground, or to the system log daemon when run in the background. The
10586 value @code{syslog} causes log entries to be sent to the system log
10587 daemon. The server name may be included in filenames using the string
10588 @code{%s}, as in @code{/var/log/cups/%s-error_log}.
10590 Defaults to @samp{"/var/log/cups/error_log"}.
10593 @deftypevr {@code{files-configuration} parameter} string fatal-errors
10594 Specifies which errors are fatal, causing the scheduler to exit. The
10599 No errors are fatal.
10602 All of the errors below are fatal.
10605 Browsing initialization errors are fatal, for example failed connections
10606 to the DNS-SD daemon.
10609 Configuration file syntax errors are fatal.
10612 Listen or Port errors are fatal, except for IPv6 failures on the
10613 loopback or @code{any} addresses.
10616 Log file creation or write errors are fatal.
10619 Bad startup file permissions are fatal, for example shared TLS
10620 certificate and key files with world-read permissions.
10623 Defaults to @samp{"all -browse"}.
10626 @deftypevr {@code{files-configuration} parameter} boolean file-device?
10627 Specifies whether the file pseudo-device can be used for new printer
10628 queues. The URI @uref{file:///dev/null} is always allowed.
10630 Defaults to @samp{#f}.
10633 @deftypevr {@code{files-configuration} parameter} string group
10634 Specifies the group name or ID that will be used when executing external
10637 Defaults to @samp{"lp"}.
10640 @deftypevr {@code{files-configuration} parameter} string log-file-perm
10641 Specifies the permissions for all log files that the scheduler writes.
10643 Defaults to @samp{"0644"}.
10646 @deftypevr {@code{files-configuration} parameter} log-location page-log
10647 Defines the page log filename. Specifying a blank filename disables
10648 access log generation. The value @code{stderr} causes log entries to be
10649 sent to the standard error file when the scheduler is running in the
10650 foreground, or to the system log daemon when run in the background. The
10651 value @code{syslog} causes log entries to be sent to the system log
10652 daemon. The server name may be included in filenames using the string
10653 @code{%s}, as in @code{/var/log/cups/%s-page_log}.
10655 Defaults to @samp{"/var/log/cups/page_log"}.
10658 @deftypevr {@code{files-configuration} parameter} string remote-root
10659 Specifies the username that is associated with unauthenticated accesses
10660 by clients claiming to be the root user. The default is @code{remroot}.
10662 Defaults to @samp{"remroot"}.
10665 @deftypevr {@code{files-configuration} parameter} file-name request-root
10666 Specifies the directory that contains print jobs and other HTTP request
10669 Defaults to @samp{"/var/spool/cups"}.
10672 @deftypevr {@code{files-configuration} parameter} sandboxing sandboxing
10673 Specifies the level of security sandboxing that is applied to print
10674 filters, backends, and other child processes of the scheduler; either
10675 @code{relaxed} or @code{strict}. This directive is currently only
10676 used/supported on macOS.
10678 Defaults to @samp{strict}.
10681 @deftypevr {@code{files-configuration} parameter} file-name server-keychain
10682 Specifies the location of TLS certificates and private keys. CUPS will
10683 look for public and private keys in this directory: a @code{.crt} files
10684 for PEM-encoded certificates and corresponding @code{.key} files for
10685 PEM-encoded private keys.
10687 Defaults to @samp{"/etc/cups/ssl"}.
10690 @deftypevr {@code{files-configuration} parameter} file-name server-root
10691 Specifies the directory containing the server configuration files.
10693 Defaults to @samp{"/etc/cups"}.
10696 @deftypevr {@code{files-configuration} parameter} boolean sync-on-close?
10697 Specifies whether the scheduler calls fsync(2) after writing
10698 configuration or state files.
10700 Defaults to @samp{#f}.
10703 @deftypevr {@code{files-configuration} parameter} space-separated-string-list system-group
10704 Specifies the group(s) to use for @code{@@SYSTEM} group authentication.
10707 @deftypevr {@code{files-configuration} parameter} file-name temp-dir
10708 Specifies the directory where temporary files are stored.
10710 Defaults to @samp{"/var/spool/cups/tmp"}.
10713 @deftypevr {@code{files-configuration} parameter} string user
10714 Specifies the user name or ID that is used when running external
10717 Defaults to @samp{"lp"}.
10721 @deftypevr {@code{cups-configuration} parameter} access-log-level access-log-level
10722 Specifies the logging level for the AccessLog file. The @code{config}
10723 level logs when printers and classes are added, deleted, or modified and
10724 when configuration files are accessed or updated. The @code{actions}
10725 level logs when print jobs are submitted, held, released, modified, or
10726 canceled, and any of the conditions for @code{config}. The @code{all}
10727 level logs all requests.
10729 Defaults to @samp{actions}.
10732 @deftypevr {@code{cups-configuration} parameter} boolean auto-purge-jobs?
10733 Specifies whether to purge job history data automatically when it is no
10734 longer required for quotas.
10736 Defaults to @samp{#f}.
10739 @deftypevr {@code{cups-configuration} parameter} browse-local-protocols browse-local-protocols
10740 Specifies which protocols to use for local printer sharing.
10742 Defaults to @samp{dnssd}.
10745 @deftypevr {@code{cups-configuration} parameter} boolean browse-web-if?
10746 Specifies whether the CUPS web interface is advertised.
10748 Defaults to @samp{#f}.
10751 @deftypevr {@code{cups-configuration} parameter} boolean browsing?
10752 Specifies whether shared printers are advertised.
10754 Defaults to @samp{#f}.
10757 @deftypevr {@code{cups-configuration} parameter} string classification
10758 Specifies the security classification of the server. Any valid banner
10759 name can be used, including "classified", "confidential", "secret",
10760 "topsecret", and "unclassified", or the banner can be omitted to disable
10761 secure printing functions.
10763 Defaults to @samp{""}.
10766 @deftypevr {@code{cups-configuration} parameter} boolean classify-override?
10767 Specifies whether users may override the classification (cover page) of
10768 individual print jobs using the @code{job-sheets} option.
10770 Defaults to @samp{#f}.
10773 @deftypevr {@code{cups-configuration} parameter} default-auth-type default-auth-type
10774 Specifies the default type of authentication to use.
10776 Defaults to @samp{Basic}.
10779 @deftypevr {@code{cups-configuration} parameter} default-encryption default-encryption
10780 Specifies whether encryption will be used for authenticated requests.
10782 Defaults to @samp{Required}.
10785 @deftypevr {@code{cups-configuration} parameter} string default-language
10786 Specifies the default language to use for text and web content.
10788 Defaults to @samp{"en"}.
10791 @deftypevr {@code{cups-configuration} parameter} string default-paper-size
10792 Specifies the default paper size for new print queues. @samp{"Auto"}
10793 uses a locale-specific default, while @samp{"None"} specifies there is
10794 no default paper size. Specific size names are typically
10795 @samp{"Letter"} or @samp{"A4"}.
10797 Defaults to @samp{"Auto"}.
10800 @deftypevr {@code{cups-configuration} parameter} string default-policy
10801 Specifies the default access policy to use.
10803 Defaults to @samp{"default"}.
10806 @deftypevr {@code{cups-configuration} parameter} boolean default-shared?
10807 Specifies whether local printers are shared by default.
10809 Defaults to @samp{#t}.
10812 @deftypevr {@code{cups-configuration} parameter} non-negative-integer dirty-clean-interval
10813 Specifies the delay for updating of configuration and state files, in
10814 seconds. A value of 0 causes the update to happen as soon as possible,
10815 typically within a few milliseconds.
10817 Defaults to @samp{30}.
10820 @deftypevr {@code{cups-configuration} parameter} error-policy error-policy
10821 Specifies what to do when an error occurs. Possible values are
10822 @code{abort-job}, which will discard the failed print job;
10823 @code{retry-job}, which will retry the job at a later time;
10824 @code{retry-this-job}, which retries the failed job immediately; and
10825 @code{stop-printer}, which stops the printer.
10827 Defaults to @samp{stop-printer}.
10830 @deftypevr {@code{cups-configuration} parameter} non-negative-integer filter-limit
10831 Specifies the maximum cost of filters that are run concurrently, which
10832 can be used to minimize disk, memory, and CPU resource problems. A
10833 limit of 0 disables filter limiting. An average print to a
10834 non-PostScript printer needs a filter limit of about 200. A PostScript
10835 printer needs about half that (100). Setting the limit below these
10836 thresholds will effectively limit the scheduler to printing a single job
10839 Defaults to @samp{0}.
10842 @deftypevr {@code{cups-configuration} parameter} non-negative-integer filter-nice
10843 Specifies the scheduling priority of filters that are run to print a
10844 job. The nice value ranges from 0, the highest priority, to 19, the
10847 Defaults to @samp{0}.
10850 @deftypevr {@code{cups-configuration} parameter} host-name-lookups host-name-lookups
10851 Specifies whether to do reverse lookups on connecting clients. The
10852 @code{double} setting causes @code{cupsd} to verify that the hostname
10853 resolved from the address matches one of the addresses returned for that
10854 hostname. Double lookups also prevent clients with unregistered
10855 addresses from connecting to your server. Only set this option to
10856 @code{#t} or @code{double} if absolutely required.
10858 Defaults to @samp{#f}.
10861 @deftypevr {@code{cups-configuration} parameter} non-negative-integer job-kill-delay
10862 Specifies the number of seconds to wait before killing the filters and
10863 backend associated with a canceled or held job.
10865 Defaults to @samp{30}.
10868 @deftypevr {@code{cups-configuration} parameter} non-negative-integer job-retry-interval
10869 Specifies the interval between retries of jobs in seconds. This is
10870 typically used for fax queues but can also be used with normal print
10871 queues whose error policy is @code{retry-job} or
10872 @code{retry-current-job}.
10874 Defaults to @samp{30}.
10877 @deftypevr {@code{cups-configuration} parameter} non-negative-integer job-retry-limit
10878 Specifies the number of retries that are done for jobs. This is
10879 typically used for fax queues but can also be used with normal print
10880 queues whose error policy is @code{retry-job} or
10881 @code{retry-current-job}.
10883 Defaults to @samp{5}.
10886 @deftypevr {@code{cups-configuration} parameter} boolean keep-alive?
10887 Specifies whether to support HTTP keep-alive connections.
10889 Defaults to @samp{#t}.
10892 @deftypevr {@code{cups-configuration} parameter} non-negative-integer keep-alive-timeout
10893 Specifies how long an idle client connection remains open, in seconds.
10895 Defaults to @samp{30}.
10898 @deftypevr {@code{cups-configuration} parameter} non-negative-integer limit-request-body
10899 Specifies the maximum size of print files, IPP requests, and HTML form
10900 data. A limit of 0 disables the limit check.
10902 Defaults to @samp{0}.
10905 @deftypevr {@code{cups-configuration} parameter} multiline-string-list listen
10906 Listens on the specified interfaces for connections. Valid values are
10907 of the form @var{address}:@var{port}, where @var{address} is either an
10908 IPv6 address enclosed in brackets, an IPv4 address, or @code{*} to
10909 indicate all addresses. Values can also be file names of local UNIX
10910 domain sockets. The Listen directive is similar to the Port directive
10911 but allows you to restrict access to specific interfaces or networks.
10914 @deftypevr {@code{cups-configuration} parameter} non-negative-integer listen-back-log
10915 Specifies the number of pending connections that will be allowed. This
10916 normally only affects very busy servers that have reached the MaxClients
10917 limit, but can also be triggered by large numbers of simultaneous
10918 connections. When the limit is reached, the operating system will
10919 refuse additional connections until the scheduler can accept the pending
10922 Defaults to @samp{128}.
10925 @deftypevr {@code{cups-configuration} parameter} location-access-control-list location-access-controls
10926 Specifies a set of additional access controls.
10928 Available @code{location-access-controls} fields are:
10930 @deftypevr {@code{location-access-controls} parameter} file-name path
10931 Specifies the URI path to which the access control applies.
10934 @deftypevr {@code{location-access-controls} parameter} access-control-list access-controls
10935 Access controls for all access to this path, in the same format as the
10936 @code{access-controls} of @code{operation-access-control}.
10938 Defaults to @samp{()}.
10941 @deftypevr {@code{location-access-controls} parameter} method-access-control-list method-access-controls
10942 Access controls for method-specific access to this path.
10944 Defaults to @samp{()}.
10946 Available @code{method-access-controls} fields are:
10948 @deftypevr {@code{method-access-controls} parameter} boolean reverse?
10949 If @code{#t}, apply access controls to all methods except the listed
10950 methods. Otherwise apply to only the listed methods.
10952 Defaults to @samp{#f}.
10955 @deftypevr {@code{method-access-controls} parameter} method-list methods
10956 Methods to which this access control applies.
10958 Defaults to @samp{()}.
10961 @deftypevr {@code{method-access-controls} parameter} access-control-list access-controls
10962 Access control directives, as a list of strings. Each string should be
10963 one directive, such as "Order allow,deny".
10965 Defaults to @samp{()}.
10970 @deftypevr {@code{cups-configuration} parameter} non-negative-integer log-debug-history
10971 Specifies the number of debugging messages that are retained for logging
10972 if an error occurs in a print job. Debug messages are logged regardless
10973 of the LogLevel setting.
10975 Defaults to @samp{100}.
10978 @deftypevr {@code{cups-configuration} parameter} log-level log-level
10979 Specifies the level of logging for the ErrorLog file. The value
10980 @code{none} stops all logging while @code{debug2} logs everything.
10982 Defaults to @samp{info}.
10985 @deftypevr {@code{cups-configuration} parameter} log-time-format log-time-format
10986 Specifies the format of the date and time in the log files. The value
10987 @code{standard} logs whole seconds while @code{usecs} logs microseconds.
10989 Defaults to @samp{standard}.
10992 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-clients
10993 Specifies the maximum number of simultaneous clients that are allowed by
10996 Defaults to @samp{100}.
10999 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-clients-per-host
11000 Specifies the maximum number of simultaneous clients that are allowed
11001 from a single address.
11003 Defaults to @samp{100}.
11006 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-copies
11007 Specifies the maximum number of copies that a user can print of each
11010 Defaults to @samp{9999}.
11013 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-hold-time
11014 Specifies the maximum time a job may remain in the @code{indefinite}
11015 hold state before it is canceled. A value of 0 disables cancellation of
11018 Defaults to @samp{0}.
11021 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-jobs
11022 Specifies the maximum number of simultaneous jobs that are allowed. Set
11023 to 0 to allow an unlimited number of jobs.
11025 Defaults to @samp{500}.
11028 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-jobs-per-printer
11029 Specifies the maximum number of simultaneous jobs that are allowed per
11030 printer. A value of 0 allows up to MaxJobs jobs per printer.
11032 Defaults to @samp{0}.
11035 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-jobs-per-user
11036 Specifies the maximum number of simultaneous jobs that are allowed per
11037 user. A value of 0 allows up to MaxJobs jobs per user.
11039 Defaults to @samp{0}.
11042 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-job-time
11043 Specifies the maximum time a job may take to print before it is
11044 canceled, in seconds. Set to 0 to disable cancellation of "stuck" jobs.
11046 Defaults to @samp{10800}.
11049 @deftypevr {@code{cups-configuration} parameter} non-negative-integer max-log-size
11050 Specifies the maximum size of the log files before they are rotated, in
11051 bytes. The value 0 disables log rotation.
11053 Defaults to @samp{1048576}.
11056 @deftypevr {@code{cups-configuration} parameter} non-negative-integer multiple-operation-timeout
11057 Specifies the maximum amount of time to allow between files in a
11058 multiple file print job, in seconds.
11060 Defaults to @samp{300}.
11063 @deftypevr {@code{cups-configuration} parameter} string page-log-format
11064 Specifies the format of PageLog lines. Sequences beginning with percent
11065 (@samp{%}) characters are replaced with the corresponding information,
11066 while all other characters are copied literally. The following percent
11067 sequences are recognized:
11071 insert a single percent character
11074 insert the value of the specified IPP attribute
11077 insert the number of copies for the current page
11080 insert the current page number
11083 insert the current date and time in common log format
11089 insert the printer name
11092 insert the username
11095 A value of the empty string disables page logging. The string @code{%p
11096 %u %j %T %P %C %@{job-billing@} %@{job-originating-host-name@}
11097 %@{job-name@} %@{media@} %@{sides@}} creates a page log with the
11100 Defaults to @samp{""}.
11103 @deftypevr {@code{cups-configuration} parameter} environment-variables environment-variables
11104 Passes the specified environment variable(s) to child processes; a list
11107 Defaults to @samp{()}.
11110 @deftypevr {@code{cups-configuration} parameter} policy-configuration-list policies
11111 Specifies named access control policies.
11113 Available @code{policy-configuration} fields are:
11115 @deftypevr {@code{policy-configuration} parameter} string name
11116 Name of the policy.
11119 @deftypevr {@code{policy-configuration} parameter} string job-private-access
11120 Specifies an access list for a job's private values. @code{@@ACL} maps
11121 to the printer's requesting-user-name-allowed or
11122 requesting-user-name-denied values. @code{@@OWNER} maps to the job's
11123 owner. @code{@@SYSTEM} maps to the groups listed for the
11124 @code{system-group} field of the @code{files-config} configuration,
11125 which is reified into the @code{cups-files.conf(5)} file. Other
11126 possible elements of the access list include specific user names, and
11127 @code{@@@var{group}} to indicate members of a specific group. The
11128 access list may also be simply @code{all} or @code{default}.
11130 Defaults to @samp{"@@OWNER @@SYSTEM"}.
11133 @deftypevr {@code{policy-configuration} parameter} string job-private-values
11134 Specifies the list of job values to make private, or @code{all},
11135 @code{default}, or @code{none}.
11137 Defaults to @samp{"job-name job-originating-host-name
11138 job-originating-user-name phone"}.
11141 @deftypevr {@code{policy-configuration} parameter} string subscription-private-access
11142 Specifies an access list for a subscription's private values.
11143 @code{@@ACL} maps to the printer's requesting-user-name-allowed or
11144 requesting-user-name-denied values. @code{@@OWNER} maps to the job's
11145 owner. @code{@@SYSTEM} maps to the groups listed for the
11146 @code{system-group} field of the @code{files-config} configuration,
11147 which is reified into the @code{cups-files.conf(5)} file. Other
11148 possible elements of the access list include specific user names, and
11149 @code{@@@var{group}} to indicate members of a specific group. The
11150 access list may also be simply @code{all} or @code{default}.
11152 Defaults to @samp{"@@OWNER @@SYSTEM"}.
11155 @deftypevr {@code{policy-configuration} parameter} string subscription-private-values
11156 Specifies the list of job values to make private, or @code{all},
11157 @code{default}, or @code{none}.
11159 Defaults to @samp{"notify-events notify-pull-method notify-recipient-uri
11160 notify-subscriber-user-name notify-user-data"}.
11163 @deftypevr {@code{policy-configuration} parameter} operation-access-control-list access-controls
11164 Access control by IPP operation.
11166 Defaults to @samp{()}.
11170 @deftypevr {@code{cups-configuration} parameter} boolean-or-non-negative-integer preserve-job-files
11171 Specifies whether job files (documents) are preserved after a job is
11172 printed. If a numeric value is specified, job files are preserved for
11173 the indicated number of seconds after printing. Otherwise a boolean
11174 value applies indefinitely.
11176 Defaults to @samp{86400}.
11179 @deftypevr {@code{cups-configuration} parameter} boolean-or-non-negative-integer preserve-job-history
11180 Specifies whether the job history is preserved after a job is printed.
11181 If a numeric value is specified, the job history is preserved for the
11182 indicated number of seconds after printing. If @code{#t}, the job
11183 history is preserved until the MaxJobs limit is reached.
11185 Defaults to @samp{#t}.
11188 @deftypevr {@code{cups-configuration} parameter} non-negative-integer reload-timeout
11189 Specifies the amount of time to wait for job completion before
11190 restarting the scheduler.
11192 Defaults to @samp{30}.
11195 @deftypevr {@code{cups-configuration} parameter} string rip-cache
11196 Specifies the maximum amount of memory to use when converting documents
11197 into bitmaps for a printer.
11199 Defaults to @samp{"128m"}.
11202 @deftypevr {@code{cups-configuration} parameter} string server-admin
11203 Specifies the email address of the server administrator.
11205 Defaults to @samp{"root@@localhost.localdomain"}.
11208 @deftypevr {@code{cups-configuration} parameter} host-name-list-or-* server-alias
11209 The ServerAlias directive is used for HTTP Host header validation when
11210 clients connect to the scheduler from external interfaces. Using the
11211 special name @code{*} can expose your system to known browser-based DNS
11212 rebinding attacks, even when accessing sites through a firewall. If the
11213 auto-discovery of alternate names does not work, we recommend listing
11214 each alternate name with a ServerAlias directive instead of using
11217 Defaults to @samp{*}.
11220 @deftypevr {@code{cups-configuration} parameter} string server-name
11221 Specifies the fully-qualified host name of the server.
11223 Defaults to @samp{"localhost"}.
11226 @deftypevr {@code{cups-configuration} parameter} server-tokens server-tokens
11227 Specifies what information is included in the Server header of HTTP
11228 responses. @code{None} disables the Server header. @code{ProductOnly}
11229 reports @code{CUPS}. @code{Major} reports @code{CUPS 2}. @code{Minor}
11230 reports @code{CUPS 2.0}. @code{Minimal} reports @code{CUPS 2.0.0}.
11231 @code{OS} reports @code{CUPS 2.0.0 (@var{uname})} where @var{uname} is
11232 the output of the @code{uname} command. @code{Full} reports @code{CUPS
11233 2.0.0 (@var{uname}) IPP/2.0}.
11235 Defaults to @samp{Minimal}.
11238 @deftypevr {@code{cups-configuration} parameter} string set-env
11239 Set the specified environment variable to be passed to child processes.
11241 Defaults to @samp{"variable value"}.
11244 @deftypevr {@code{cups-configuration} parameter} multiline-string-list ssl-listen
11245 Listens on the specified interfaces for encrypted connections. Valid
11246 values are of the form @var{address}:@var{port}, where @var{address} is
11247 either an IPv6 address enclosed in brackets, an IPv4 address, or
11248 @code{*} to indicate all addresses.
11250 Defaults to @samp{()}.
11253 @deftypevr {@code{cups-configuration} parameter} ssl-options ssl-options
11254 Sets encryption options. By default, CUPS only supports encryption
11255 using TLS v1.0 or higher using known secure cipher suites. The
11256 @code{AllowRC4} option enables the 128-bit RC4 cipher suites, which are
11257 required for some older clients that do not implement newer ones. The
11258 @code{AllowSSL3} option enables SSL v3.0, which is required for some
11259 older clients that do not support TLS v1.0.
11261 Defaults to @samp{()}.
11264 @deftypevr {@code{cups-configuration} parameter} boolean strict-conformance?
11265 Specifies whether the scheduler requires clients to strictly adhere to
11266 the IPP specifications.
11268 Defaults to @samp{#f}.
11271 @deftypevr {@code{cups-configuration} parameter} non-negative-integer timeout
11272 Specifies the HTTP request timeout, in seconds.
11274 Defaults to @samp{300}.
11278 @deftypevr {@code{cups-configuration} parameter} boolean web-interface?
11279 Specifies whether the web interface is enabled.
11281 Defaults to @samp{#f}.
11284 At this point you're probably thinking ``oh dear, Guix manual, I like
11285 you but you can stop already with the configuration options''. Indeed.
11286 However, one more point: it could be that you have an existing
11287 @code{cupsd.conf} that you want to use. In that case, you can pass an
11288 @code{opaque-cups-configuration} as the configuration of a
11289 @code{cups-service-type}.
11291 Available @code{opaque-cups-configuration} fields are:
11293 @deftypevr {@code{opaque-cups-configuration} parameter} package cups
11297 @deftypevr {@code{opaque-cups-configuration} parameter} string cupsd.conf
11298 The contents of the @code{cupsd.conf}, as a string.
11301 @deftypevr {@code{opaque-cups-configuration} parameter} string cups-files.conf
11302 The contents of the @code{cups-files.conf} file, as a string.
11305 For example, if your @code{cupsd.conf} and @code{cups-files.conf} are in
11306 strings of the same name, you could instantiate a CUPS service like
11310 (service cups-service-type
11311 (opaque-cups-configuration
11312 (cupsd.conf cupsd.conf)
11313 (cups-files.conf cups-files.conf)))
11317 @node Desktop Services
11318 @subsubsection Desktop Services
11320 The @code{(gnu services desktop)} module provides services that are
11321 usually useful in the context of a ``desktop'' setup---that is, on a
11322 machine running a graphical display server, possibly with graphical user
11323 interfaces, etc. It also defines services that provide specific desktop
11324 environments like GNOME and XFCE.
11326 To simplify things, the module defines a variable containing the set of
11327 services that users typically expect on a machine with a graphical
11328 environment and networking:
11330 @defvr {Scheme Variable} %desktop-services
11331 This is a list of services that builds upon @var{%base-services} and
11332 adds or adjusts services for a typical ``desktop'' setup.
11334 In particular, it adds a graphical login manager (@pxref{X Window,
11335 @code{slim-service}}), screen lockers,
11336 a network management tool (@pxref{Networking
11337 Services, @code{wicd-service}}), energy and color management services,
11338 the @code{elogind} login and seat manager, the Polkit privilege service,
11339 the GeoClue location service, an NTP client (@pxref{Networking
11340 Services}), the Avahi daemon, and has the name service switch service
11341 configured to be able to use @code{nss-mdns} (@pxref{Name Service
11345 The @var{%desktop-services} variable can be used as the @code{services}
11346 field of an @code{operating-system} declaration (@pxref{operating-system
11347 Reference, @code{services}}).
11349 Additionally, the @code{gnome-desktop-service} and
11350 @code{xfce-desktop-service} procedures can add GNOME and/or XFCE to a
11351 system. To ``add GNOME'' means that system-level services like the
11352 backlight adjustment helpers and the power management utilities are
11353 added to the system, extending @code{polkit} and @code{dbus}
11354 appropriately, allowing GNOME to operate with elevated privileges on a
11355 limited number of special-purpose system interfaces. Additionally,
11356 adding a service made by @code{gnome-desktop-service} adds the GNOME
11357 metapackage to the system profile. Likewise, adding the XFCE service
11358 not only adds the @code{xfce} metapackage to the system profile, but it
11359 also gives the Thunar file manager the ability to open a ``root-mode''
11360 file management window, if the user authenticates using the
11361 administrator's password via the standard polkit graphical interface.
11363 @deffn {Scheme Procedure} gnome-desktop-service
11364 Return a service that adds the @code{gnome} package to the system
11365 profile, and extends polkit with the actions from
11366 @code{gnome-settings-daemon}.
11369 @deffn {Scheme Procedure} xfce-desktop-service
11370 Return a service that adds the @code{xfce} package to the system profile,
11371 and extends polkit with the ability for @code{thunar} to manipulate the
11372 file system as root from within a user session, after the user has
11373 authenticated with the administrator's password.
11376 Because the GNOME and XFCE desktop services pull in so many packages,
11377 the default @code{%desktop-services} variable doesn't include either of
11378 them by default. To add GNOME or XFCE, just @code{cons} them onto
11379 @code{%desktop-services} in the @code{services} field of your
11380 @code{operating-system}:
11383 (use-modules (gnu))
11384 (use-service-modules desktop)
11387 ;; cons* adds items to the list given as its last argument.
11388 (services (cons* (gnome-desktop-service)
11389 (xfce-desktop-service)
11390 %desktop-services))
11394 These desktop environments will then be available as options in the
11395 graphical login window.
11397 The actual service definitions included in @code{%desktop-services} and
11398 provided by @code{(gnu services dbus)} and @code{(gnu services desktop)}
11399 are described below.
11401 @deffn {Scheme Procedure} dbus-service [#:dbus @var{dbus}] [#:services '()]
11402 Return a service that runs the ``system bus'', using @var{dbus}, with
11403 support for @var{services}.
11405 @uref{http://dbus.freedesktop.org/, D-Bus} is an inter-process communication
11406 facility. Its system bus is used to allow system services to communicate
11407 and to be notified of system-wide events.
11409 @var{services} must be a list of packages that provide an
11410 @file{etc/dbus-1/system.d} directory containing additional D-Bus configuration
11411 and policy files. For example, to allow avahi-daemon to use the system bus,
11412 @var{services} must be equal to @code{(list avahi)}.
11415 @deffn {Scheme Procedure} elogind-service [#:config @var{config}]
11416 Return a service that runs the @code{elogind} login and
11417 seat management daemon. @uref{https://github.com/andywingo/elogind,
11418 Elogind} exposes a D-Bus interface that can be used to know which users
11419 are logged in, know what kind of sessions they have open, suspend the
11420 system, inhibit system suspend, reboot the system, and other tasks.
11422 Elogind handles most system-level power events for a computer, for
11423 example suspending the system when a lid is closed, or shutting it down
11424 when the power button is pressed.
11426 The @var{config} keyword argument specifies the configuration for
11427 elogind, and should be the result of an @code{(elogind-configuration
11428 (@var{parameter} @var{value})...)} invocation. Available parameters and
11429 their default values are:
11432 @item kill-user-processes?
11434 @item kill-only-users
11436 @item kill-exclude-users
11438 @item inhibit-delay-max-seconds
11440 @item handle-power-key
11442 @item handle-suspend-key
11444 @item handle-hibernate-key
11446 @item handle-lid-switch
11448 @item handle-lid-switch-docked
11450 @item power-key-ignore-inhibited?
11452 @item suspend-key-ignore-inhibited?
11454 @item hibernate-key-ignore-inhibited?
11456 @item lid-switch-ignore-inhibited?
11458 @item holdoff-timeout-seconds
11462 @item idle-action-seconds
11464 @item runtime-directory-size-percent
11466 @item runtime-directory-size
11470 @item suspend-state
11471 @code{("mem" "standby" "freeze")}
11474 @item hibernate-state
11476 @item hibernate-mode
11477 @code{("platform" "shutdown")}
11478 @item hybrid-sleep-state
11480 @item hybrid-sleep-mode
11481 @code{("suspend" "platform" "shutdown")}
11485 @deffn {Scheme Procedure} polkit-service @
11486 [#:polkit @var{polkit}]
11487 Return a service that runs the
11488 @uref{http://www.freedesktop.org/wiki/Software/polkit/, Polkit privilege
11489 management service}, which allows system administrators to grant access to
11490 privileged operations in a structured way. By querying the Polkit service, a
11491 privileged system component can know when it should grant additional
11492 capabilities to ordinary users. For example, an ordinary user can be granted
11493 the capability to suspend the system if the user is logged in locally.
11496 @deffn {Scheme Procedure} upower-service [#:upower @var{upower}] @
11497 [#:watts-up-pro? #f] @
11498 [#:poll-batteries? #t] @
11499 [#:ignore-lid? #f] @
11500 [#:use-percentage-for-policy? #f] @
11501 [#:percentage-low 10] @
11502 [#:percentage-critical 3] @
11503 [#:percentage-action 2] @
11504 [#:time-low 1200] @
11505 [#:time-critical 300] @
11506 [#:time-action 120] @
11507 [#:critical-power-action 'hybrid-sleep]
11508 Return a service that runs @uref{http://upower.freedesktop.org/,
11509 @command{upowerd}}, a system-wide monitor for power consumption and battery
11510 levels, with the given configuration settings. It implements the
11511 @code{org.freedesktop.UPower} D-Bus interface, and is notably used by
11515 @deffn {Scheme Procedure} udisks-service [#:udisks @var{udisks}]
11516 Return a service for @uref{http://udisks.freedesktop.org/docs/latest/,
11517 UDisks}, a @dfn{disk management} daemon that provides user interfaces with
11518 notifications and ways to mount/unmount disks. Programs that talk to UDisks
11519 include the @command{udisksctl} command, part of UDisks, and GNOME Disks.
11522 @deffn {Scheme Procedure} colord-service [#:colord @var{colord}]
11523 Return a service that runs @command{colord}, a system service with a D-Bus
11524 interface to manage the color profiles of input and output devices such as
11525 screens and scanners. It is notably used by the GNOME Color Manager graphical
11526 tool. See @uref{http://www.freedesktop.org/software/colord/, the colord web
11527 site} for more information.
11530 @deffn {Scheme Procedure} geoclue-application name [#:allowed? #t] [#:system? #f] [#:users '()]
11531 Return a configuration allowing an application to access GeoClue
11532 location data. @var{name} is the Desktop ID of the application, without
11533 the @code{.desktop} part. If @var{allowed?} is true, the application
11534 will have access to location information by default. The boolean
11535 @var{system?} value indicates whether an application is a system component
11536 or not. Finally @var{users} is a list of UIDs of all users for which
11537 this application is allowed location info access. An empty users list
11538 means that all users are allowed.
11541 @defvr {Scheme Variable} %standard-geoclue-applications
11542 The standard list of well-known GeoClue application configurations,
11543 granting authority to the GNOME date-and-time utility to ask for the
11544 current location in order to set the time zone, and allowing the
11545 IceCat and Epiphany web browsers to request location information.
11546 IceCat and Epiphany both query the user before allowing a web page to
11547 know the user's location.
11550 @deffn {Scheme Procedure} geoclue-service [#:colord @var{colord}] @
11551 [#:whitelist '()] @
11552 [#:wifi-geolocation-url "https://location.services.mozilla.com/v1/geolocate?key=geoclue"] @
11553 [#:submit-data? #f]
11554 [#:wifi-submission-url "https://location.services.mozilla.com/v1/submit?key=geoclue"] @
11555 [#:submission-nick "geoclue"] @
11556 [#:applications %standard-geoclue-applications]
11557 Return a service that runs the GeoClue location service. This service
11558 provides a D-Bus interface to allow applications to request access to a
11559 user's physical location, and optionally to add information to online
11560 location databases. See
11561 @uref{https://wiki.freedesktop.org/www/Software/GeoClue/, the GeoClue
11562 web site} for more information.
11565 @deffn {Scheme Procedure} bluetooth-service [#:bluez @var{bluez}]
11566 Return a service that runs the @command{bluetoothd} daemon, which manages
11567 all the Bluetooth devices and provides a number of D-Bus interfaces.
11569 Users need to be in the @code{lp} group to access the D-Bus service.
11572 @node Database Services
11573 @subsubsection Database Services
11577 The @code{(gnu services databases)} module provides the following services.
11579 @deffn {Scheme Procedure} postgresql-service [#:postgresql postgresql] @
11580 [#:config-file] [#:data-directory ``/var/lib/postgresql/data''] @
11581 [#:port 5432] [#:locale ``en_US.utf8'']
11582 Return a service that runs @var{postgresql}, the PostgreSQL database
11585 The PostgreSQL daemon loads its runtime configuration from @var{config-file},
11586 creates a database cluster with @var{locale} as the default
11587 locale, stored in @var{data-directory}. It then listens on @var{port}.
11590 @deffn {Scheme Procedure} mysql-service [#:config (mysql-configuration)]
11591 Return a service that runs @command{mysqld}, the MySQL or MariaDB
11594 The optional @var{config} argument specifies the configuration for
11595 @command{mysqld}, which should be a @code{<mysql-configuration>} object.
11598 @deftp {Data Type} mysql-configuration
11599 Data type representing the configuration of @var{mysql-service}.
11602 @item @code{mysql} (default: @var{mariadb})
11603 Package object of the MySQL database server, can be either @var{mariadb}
11606 For MySQL, a temporary root password will be displayed at activation time.
11607 For MariaDB, the root password is empty.
11609 @item @code{port} (default: @code{3306})
11610 TCP port on which the database server listens for incoming connections.
11614 @defvr {Scheme Variable} redis-service-type
11615 This is the service type for the @uref{https://redis.io/, Redis}
11616 key/value store, whose value is a @code{redis-configuration} object.
11619 @deftp {Data Type} redis-configuration
11620 Data type representing the configuration of redis.
11623 @item @code{redis} (default: @code{redis})
11624 The Redis package to use.
11626 @item @code{bind} (default: @code{"127.0.0.1"})
11627 Network interface on which to listen.
11629 @item @code{port} (default: @code{6379})
11630 Port on which to accept connections on, a value of 0 will disable
11631 listening on a TCP socket.
11633 @item @code{working-directory} (default: @code{"/var/lib/redis"})
11634 Directory in which to store the database and related files.
11638 @node Mail Services
11639 @subsubsection Mail Services
11643 The @code{(gnu services mail)} module provides Guix service definitions
11644 for email services: IMAP, POP3, and LMTP servers, as well as mail
11645 transport agents (MTAs). Lots of acronyms! These services are detailed
11646 in the subsections below.
11648 @subsubheading Dovecot Service
11650 @deffn {Scheme Procedure} dovecot-service [#:config (dovecot-configuration)]
11651 Return a service that runs the Dovecot IMAP/POP3/LMTP mail server.
11654 By default, Dovecot does not need much configuration; the default
11655 configuration object created by @code{(dovecot-configuration)} will
11656 suffice if your mail is delivered to @code{~/Maildir}. A self-signed
11657 certificate will be generated for TLS-protected connections, though
11658 Dovecot will also listen on cleartext ports by default. There are a
11659 number of options, though, which mail administrators might need to change,
11660 and as is the case with other services, Guix allows the system
11661 administrator to specify these parameters via a uniform Scheme interface.
11663 For example, to specify that mail is located at @code{maildir~/.mail},
11664 one would instantiate the Dovecot service like this:
11667 (dovecot-service #:config
11668 (dovecot-configuration
11669 (mail-location "maildir:~/.mail")))
11672 The available configuration parameters follow. Each parameter
11673 definition is preceded by its type; for example, @samp{string-list foo}
11674 indicates that the @code{foo} parameter should be specified as a list of
11675 strings. There is also a way to specify the configuration as a string,
11676 if you have an old @code{dovecot.conf} file that you want to port over
11677 from some other system; see the end for more details.
11679 @c The following documentation was initially generated by
11680 @c (generate-documentation) in (gnu services mail). Manually maintained
11681 @c documentation is better, so we shouldn't hesitate to edit below as
11682 @c needed. However if the change you want to make to this documentation
11683 @c can be done in an automated way, it's probably easier to change
11684 @c (generate-documentation) than to make it below and have to deal with
11685 @c the churn as dovecot updates.
11687 Available @code{dovecot-configuration} fields are:
11689 @deftypevr {@code{dovecot-configuration} parameter} package dovecot
11690 The dovecot package.
11693 @deftypevr {@code{dovecot-configuration} parameter} comma-separated-string-list listen
11694 A list of IPs or hosts where to listen for connections. @samp{*}
11695 listens on all IPv4 interfaces, @samp{::} listens on all IPv6
11696 interfaces. If you want to specify non-default ports or anything more
11697 complex, customize the address and port fields of the
11698 @samp{inet-listener} of the specific services you are interested in.
11701 @deftypevr {@code{dovecot-configuration} parameter} protocol-configuration-list protocols
11702 List of protocols we want to serve. Available protocols include
11703 @samp{imap}, @samp{pop3}, and @samp{lmtp}.
11705 Available @code{protocol-configuration} fields are:
11707 @deftypevr {@code{protocol-configuration} parameter} string name
11708 The name of the protocol.
11711 @deftypevr {@code{protocol-configuration} parameter} string auth-socket-path
11712 UNIX socket path to the master authentication server to find users.
11713 This is used by imap (for shared users) and lda.
11714 It defaults to @samp{"/var/run/dovecot/auth-userdb"}.
11717 @deftypevr {@code{protocol-configuration} parameter} space-separated-string-list mail-plugins
11718 Space separated list of plugins to load.
11721 @deftypevr {@code{protocol-configuration} parameter} non-negative-integer mail-max-userip-connections
11722 Maximum number of IMAP connections allowed for a user from each IP
11723 address. NOTE: The username is compared case-sensitively.
11724 Defaults to @samp{10}.
11729 @deftypevr {@code{dovecot-configuration} parameter} service-configuration-list services
11730 List of services to enable. Available services include @samp{imap},
11731 @samp{imap-login}, @samp{pop3}, @samp{pop3-login}, @samp{auth}, and
11734 Available @code{service-configuration} fields are:
11736 @deftypevr {@code{service-configuration} parameter} string kind
11737 The service kind. Valid values include @code{director},
11738 @code{imap-login}, @code{pop3-login}, @code{lmtp}, @code{imap},
11739 @code{pop3}, @code{auth}, @code{auth-worker}, @code{dict},
11740 @code{tcpwrap}, @code{quota-warning}, or anything else.
11743 @deftypevr {@code{service-configuration} parameter} listener-configuration-list listeners
11744 Listeners for the service. A listener is either a
11745 @code{unix-listener-configuration}, a @code{fifo-listener-configuration}, or
11746 an @code{inet-listener-configuration}.
11747 Defaults to @samp{()}.
11749 Available @code{unix-listener-configuration} fields are:
11751 @deftypevr {@code{unix-listener-configuration} parameter} string path
11752 Path to the file, relative to @code{base-dir} field. This is also used as
11756 @deftypevr {@code{unix-listener-configuration} parameter} string mode
11757 The access mode for the socket.
11758 Defaults to @samp{"0600"}.
11761 @deftypevr {@code{unix-listener-configuration} parameter} string user
11762 The user to own the socket.
11763 Defaults to @samp{""}.
11766 @deftypevr {@code{unix-listener-configuration} parameter} string group
11767 The group to own the socket.
11768 Defaults to @samp{""}.
11772 Available @code{fifo-listener-configuration} fields are:
11774 @deftypevr {@code{fifo-listener-configuration} parameter} string path
11775 Path to the file, relative to @code{base-dir} field. This is also used as
11779 @deftypevr {@code{fifo-listener-configuration} parameter} string mode
11780 The access mode for the socket.
11781 Defaults to @samp{"0600"}.
11784 @deftypevr {@code{fifo-listener-configuration} parameter} string user
11785 The user to own the socket.
11786 Defaults to @samp{""}.
11789 @deftypevr {@code{fifo-listener-configuration} parameter} string group
11790 The group to own the socket.
11791 Defaults to @samp{""}.
11795 Available @code{inet-listener-configuration} fields are:
11797 @deftypevr {@code{inet-listener-configuration} parameter} string protocol
11798 The protocol to listen for.
11801 @deftypevr {@code{inet-listener-configuration} parameter} string address
11802 The address on which to listen, or empty for all addresses.
11803 Defaults to @samp{""}.
11806 @deftypevr {@code{inet-listener-configuration} parameter} non-negative-integer port
11807 The port on which to listen.
11810 @deftypevr {@code{inet-listener-configuration} parameter} boolean ssl?
11811 Whether to use SSL for this service; @samp{yes}, @samp{no}, or
11813 Defaults to @samp{#t}.
11818 @deftypevr {@code{service-configuration} parameter} non-negative-integer service-count
11819 Number of connections to handle before starting a new process.
11820 Typically the only useful values are 0 (unlimited) or 1. 1 is more
11821 secure, but 0 is faster. <doc/wiki/LoginProcess.txt>.
11822 Defaults to @samp{1}.
11825 @deftypevr {@code{service-configuration} parameter} non-negative-integer process-min-avail
11826 Number of processes to always keep waiting for more connections.
11827 Defaults to @samp{0}.
11830 @deftypevr {@code{service-configuration} parameter} non-negative-integer vsz-limit
11831 If you set @samp{service-count 0}, you probably need to grow
11833 Defaults to @samp{256000000}.
11838 @deftypevr {@code{dovecot-configuration} parameter} dict-configuration dict
11839 Dict configuration, as created by the @code{dict-configuration}
11842 Available @code{dict-configuration} fields are:
11844 @deftypevr {@code{dict-configuration} parameter} free-form-fields entries
11845 A list of key-value pairs that this dict should hold.
11846 Defaults to @samp{()}.
11851 @deftypevr {@code{dovecot-configuration} parameter} passdb-configuration-list passdbs
11852 A list of passdb configurations, each one created by the
11853 @code{passdb-configuration} constructor.
11855 Available @code{passdb-configuration} fields are:
11857 @deftypevr {@code{passdb-configuration} parameter} string driver
11858 The driver that the passdb should use. Valid values include
11859 @samp{pam}, @samp{passwd}, @samp{shadow}, @samp{bsdauth}, and
11861 Defaults to @samp{"pam"}.
11864 @deftypevr {@code{passdb-configuration} parameter} space-separated-string-list args
11865 Space separated list of arguments to the passdb driver.
11866 Defaults to @samp{""}.
11871 @deftypevr {@code{dovecot-configuration} parameter} userdb-configuration-list userdbs
11872 List of userdb configurations, each one created by the
11873 @code{userdb-configuration} constructor.
11875 Available @code{userdb-configuration} fields are:
11877 @deftypevr {@code{userdb-configuration} parameter} string driver
11878 The driver that the userdb should use. Valid values include
11879 @samp{passwd} and @samp{static}.
11880 Defaults to @samp{"passwd"}.
11883 @deftypevr {@code{userdb-configuration} parameter} space-separated-string-list args
11884 Space separated list of arguments to the userdb driver.
11885 Defaults to @samp{""}.
11888 @deftypevr {@code{userdb-configuration} parameter} free-form-args override-fields
11889 Override fields from passwd.
11890 Defaults to @samp{()}.
11895 @deftypevr {@code{dovecot-configuration} parameter} plugin-configuration plugin-configuration
11896 Plug-in configuration, created by the @code{plugin-configuration}
11900 @deftypevr {@code{dovecot-configuration} parameter} list-of-namespace-configuration namespaces
11901 List of namespaces. Each item in the list is created by the
11902 @code{namespace-configuration} constructor.
11904 Available @code{namespace-configuration} fields are:
11906 @deftypevr {@code{namespace-configuration} parameter} string name
11907 Name for this namespace.
11910 @deftypevr {@code{namespace-configuration} parameter} string type
11911 Namespace type: @samp{private}, @samp{shared} or @samp{public}.
11912 Defaults to @samp{"private"}.
11915 @deftypevr {@code{namespace-configuration} parameter} string separator
11916 Hierarchy separator to use. You should use the same separator for
11917 all namespaces or some clients get confused. @samp{/} is usually a good
11918 one. The default however depends on the underlying mail storage
11920 Defaults to @samp{""}.
11923 @deftypevr {@code{namespace-configuration} parameter} string prefix
11924 Prefix required to access this namespace. This needs to be
11925 different for all namespaces. For example @samp{Public/}.
11926 Defaults to @samp{""}.
11929 @deftypevr {@code{namespace-configuration} parameter} string location
11930 Physical location of the mailbox. This is in the same format as
11931 mail_location, which is also the default for it.
11932 Defaults to @samp{""}.
11935 @deftypevr {@code{namespace-configuration} parameter} boolean inbox?
11936 There can be only one INBOX, and this setting defines which
11938 Defaults to @samp{#f}.
11941 @deftypevr {@code{namespace-configuration} parameter} boolean hidden?
11942 If namespace is hidden, it's not advertised to clients via NAMESPACE
11943 extension. You'll most likely also want to set @samp{list? #f}. This is mostly
11944 useful when converting from another server with different namespaces
11945 which you want to deprecate but still keep working. For example you can
11946 create hidden namespaces with prefixes @samp{~/mail/}, @samp{~%u/mail/}
11948 Defaults to @samp{#f}.
11951 @deftypevr {@code{namespace-configuration} parameter} boolean list?
11952 Show the mailboxes under this namespace with the LIST command. This
11953 makes the namespace visible for clients that do not support the NAMESPACE
11954 extension. The special @code{children} value lists child mailboxes, but
11955 hides the namespace prefix.
11956 Defaults to @samp{#t}.
11959 @deftypevr {@code{namespace-configuration} parameter} boolean subscriptions?
11960 Namespace handles its own subscriptions. If set to @code{#f}, the
11961 parent namespace handles them. The empty prefix should always have this
11963 Defaults to @samp{#t}.
11966 @deftypevr {@code{namespace-configuration} parameter} mailbox-configuration-list mailboxes
11967 List of predefined mailboxes in this namespace.
11968 Defaults to @samp{()}.
11970 Available @code{mailbox-configuration} fields are:
11972 @deftypevr {@code{mailbox-configuration} parameter} string name
11973 Name for this mailbox.
11976 @deftypevr {@code{mailbox-configuration} parameter} string auto
11977 @samp{create} will automatically create this mailbox.
11978 @samp{subscribe} will both create and subscribe to the mailbox.
11979 Defaults to @samp{"no"}.
11982 @deftypevr {@code{mailbox-configuration} parameter} space-separated-string-list special-use
11983 List of IMAP @code{SPECIAL-USE} attributes as specified by RFC 6154.
11984 Valid values are @code{\All}, @code{\Archive}, @code{\Drafts},
11985 @code{\Flagged}, @code{\Junk}, @code{\Sent}, and @code{\Trash}.
11986 Defaults to @samp{()}.
11993 @deftypevr {@code{dovecot-configuration} parameter} file-name base-dir
11994 Base directory where to store runtime data.
11995 Defaults to @samp{"/var/run/dovecot/"}.
11998 @deftypevr {@code{dovecot-configuration} parameter} string login-greeting
11999 Greeting message for clients.
12000 Defaults to @samp{"Dovecot ready."}.
12003 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list login-trusted-networks
12004 List of trusted network ranges. Connections from these IPs are
12005 allowed to override their IP addresses and ports (for logging and for
12006 authentication checks). @samp{disable-plaintext-auth} is also ignored
12007 for these networks. Typically you would specify your IMAP proxy servers
12009 Defaults to @samp{()}.
12012 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list login-access-sockets
12013 List of login access check sockets (e.g. tcpwrap).
12014 Defaults to @samp{()}.
12017 @deftypevr {@code{dovecot-configuration} parameter} boolean verbose-proctitle?
12018 Show more verbose process titles (in ps). Currently shows user name
12019 and IP address. Useful for seeing who is actually using the IMAP
12020 processes (e.g. shared mailboxes or if the same uid is used for multiple
12022 Defaults to @samp{#f}.
12025 @deftypevr {@code{dovecot-configuration} parameter} boolean shutdown-clients?
12026 Should all processes be killed when Dovecot master process shuts down.
12027 Setting this to @code{#f} means that Dovecot can be upgraded without
12028 forcing existing client connections to close (although that could also
12029 be a problem if the upgrade is e.g. due to a security fix).
12030 Defaults to @samp{#t}.
12033 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer doveadm-worker-count
12034 If non-zero, run mail commands via this many connections to doveadm
12035 server, instead of running them directly in the same process.
12036 Defaults to @samp{0}.
12039 @deftypevr {@code{dovecot-configuration} parameter} string doveadm-socket-path
12040 UNIX socket or host:port used for connecting to doveadm server.
12041 Defaults to @samp{"doveadm-server"}.
12044 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list import-environment
12045 List of environment variables that are preserved on Dovecot startup
12046 and passed down to all of its child processes. You can also give
12047 key=value pairs to always set specific settings.
12050 @deftypevr {@code{dovecot-configuration} parameter} boolean disable-plaintext-auth?
12051 Disable LOGIN command and all other plaintext authentications unless
12052 SSL/TLS is used (LOGINDISABLED capability). Note that if the remote IP
12053 matches the local IP (i.e. you're connecting from the same computer),
12054 the connection is considered secure and plaintext authentication is
12055 allowed. See also ssl=required setting.
12056 Defaults to @samp{#t}.
12059 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer auth-cache-size
12060 Authentication cache size (e.g. @samp{#e10e6}). 0 means it's disabled.
12061 Note that bsdauth, PAM and vpopmail require @samp{cache-key} to be set
12062 for caching to be used.
12063 Defaults to @samp{0}.
12066 @deftypevr {@code{dovecot-configuration} parameter} string auth-cache-ttl
12067 Time to live for cached data. After TTL expires the cached record
12068 is no longer used, *except* if the main database lookup returns internal
12069 failure. We also try to handle password changes automatically: If
12070 user's previous authentication was successful, but this one wasn't, the
12071 cache isn't used. For now this works only with plaintext
12073 Defaults to @samp{"1 hour"}.
12076 @deftypevr {@code{dovecot-configuration} parameter} string auth-cache-negative-ttl
12077 TTL for negative hits (user not found, password mismatch).
12078 0 disables caching them completely.
12079 Defaults to @samp{"1 hour"}.
12082 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list auth-realms
12083 List of realms for SASL authentication mechanisms that need them.
12084 You can leave it empty if you don't want to support multiple realms.
12085 Many clients simply use the first one listed here, so keep the default
12087 Defaults to @samp{()}.
12090 @deftypevr {@code{dovecot-configuration} parameter} string auth-default-realm
12091 Default realm/domain to use if none was specified. This is used for
12092 both SASL realms and appending @@domain to username in plaintext
12094 Defaults to @samp{""}.
12097 @deftypevr {@code{dovecot-configuration} parameter} string auth-username-chars
12098 List of allowed characters in username. If the user-given username
12099 contains a character not listed in here, the login automatically fails.
12100 This is just an extra check to make sure user can't exploit any
12101 potential quote escaping vulnerabilities with SQL/LDAP databases. If
12102 you want to allow all characters, set this value to empty.
12103 Defaults to @samp{"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ01234567890.-_@@"}.
12106 @deftypevr {@code{dovecot-configuration} parameter} string auth-username-translation
12107 Username character translations before it's looked up from
12108 databases. The value contains series of from -> to characters. For
12109 example @samp{#@@/@@} means that @samp{#} and @samp{/} characters are
12110 translated to @samp{@@}.
12111 Defaults to @samp{""}.
12114 @deftypevr {@code{dovecot-configuration} parameter} string auth-username-format
12115 Username formatting before it's looked up from databases. You can
12116 use the standard variables here, e.g. %Lu would lowercase the username,
12117 %n would drop away the domain if it was given, or @samp{%n-AT-%d} would
12118 change the @samp{@@} into @samp{-AT-}. This translation is done after
12119 @samp{auth-username-translation} changes.
12120 Defaults to @samp{"%Lu"}.
12123 @deftypevr {@code{dovecot-configuration} parameter} string auth-master-user-separator
12124 If you want to allow master users to log in by specifying the master
12125 username within the normal username string (i.e. not using SASL
12126 mechanism's support for it), you can specify the separator character
12127 here. The format is then <username><separator><master username>.
12128 UW-IMAP uses @samp{*} as the separator, so that could be a good
12130 Defaults to @samp{""}.
12133 @deftypevr {@code{dovecot-configuration} parameter} string auth-anonymous-username
12134 Username to use for users logging in with ANONYMOUS SASL
12136 Defaults to @samp{"anonymous"}.
12139 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer auth-worker-max-count
12140 Maximum number of dovecot-auth worker processes. They're used to
12141 execute blocking passdb and userdb queries (e.g. MySQL and PAM).
12142 They're automatically created and destroyed as needed.
12143 Defaults to @samp{30}.
12146 @deftypevr {@code{dovecot-configuration} parameter} string auth-gssapi-hostname
12147 Host name to use in GSSAPI principal names. The default is to use
12148 the name returned by gethostname(). Use @samp{$ALL} (with quotes) to
12149 allow all keytab entries.
12150 Defaults to @samp{""}.
12153 @deftypevr {@code{dovecot-configuration} parameter} string auth-krb5-keytab
12154 Kerberos keytab to use for the GSSAPI mechanism. Will use the
12155 system default (usually @file{/etc/krb5.keytab}) if not specified. You may
12156 need to change the auth service to run as root to be able to read this
12158 Defaults to @samp{""}.
12161 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-use-winbind?
12162 Do NTLM and GSS-SPNEGO authentication using Samba's winbind daemon
12163 and @samp{ntlm-auth} helper.
12164 <doc/wiki/Authentication/Mechanisms/Winbind.txt>.
12165 Defaults to @samp{#f}.
12168 @deftypevr {@code{dovecot-configuration} parameter} file-name auth-winbind-helper-path
12169 Path for Samba's @samp{ntlm-auth} helper binary.
12170 Defaults to @samp{"/usr/bin/ntlm_auth"}.
12173 @deftypevr {@code{dovecot-configuration} parameter} string auth-failure-delay
12174 Time to delay before replying to failed authentications.
12175 Defaults to @samp{"2 secs"}.
12178 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-ssl-require-client-cert?
12179 Require a valid SSL client certificate or the authentication
12181 Defaults to @samp{#f}.
12184 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-ssl-username-from-cert?
12185 Take the username from client's SSL certificate, using
12186 @code{X509_NAME_get_text_by_NID()} which returns the subject's DN's
12188 Defaults to @samp{#f}.
12191 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list auth-mechanisms
12192 List of wanted authentication mechanisms. Supported mechanisms are:
12193 @samp{plain}, @samp{login}, @samp{digest-md5}, @samp{cram-md5},
12194 @samp{ntlm}, @samp{rpa}, @samp{apop}, @samp{anonymous}, @samp{gssapi},
12195 @samp{otp}, @samp{skey}, and @samp{gss-spnego}. NOTE: See also
12196 @samp{disable-plaintext-auth} setting.
12199 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list director-servers
12200 List of IPs or hostnames to all director servers, including ourself.
12201 Ports can be specified as ip:port. The default port is the same as what
12202 director service's @samp{inet-listener} is using.
12203 Defaults to @samp{()}.
12206 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list director-mail-servers
12207 List of IPs or hostnames to all backend mail servers. Ranges are
12208 allowed too, like 10.0.0.10-10.0.0.30.
12209 Defaults to @samp{()}.
12212 @deftypevr {@code{dovecot-configuration} parameter} string director-user-expire
12213 How long to redirect users to a specific server after it no longer
12214 has any connections.
12215 Defaults to @samp{"15 min"}.
12218 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer director-doveadm-port
12219 TCP/IP port that accepts doveadm connections (instead of director
12220 connections) If you enable this, you'll also need to add
12221 @samp{inet-listener} for the port.
12222 Defaults to @samp{0}.
12225 @deftypevr {@code{dovecot-configuration} parameter} string director-username-hash
12226 How the username is translated before being hashed. Useful values
12227 include %Ln if user can log in with or without @@domain, %Ld if mailboxes
12228 are shared within domain.
12229 Defaults to @samp{"%Lu"}.
12232 @deftypevr {@code{dovecot-configuration} parameter} string log-path
12233 Log file to use for error messages. @samp{syslog} logs to syslog,
12234 @samp{/dev/stderr} logs to stderr.
12235 Defaults to @samp{"syslog"}.
12238 @deftypevr {@code{dovecot-configuration} parameter} string info-log-path
12239 Log file to use for informational messages. Defaults to
12241 Defaults to @samp{""}.
12244 @deftypevr {@code{dovecot-configuration} parameter} string debug-log-path
12245 Log file to use for debug messages. Defaults to
12246 @samp{info-log-path}.
12247 Defaults to @samp{""}.
12250 @deftypevr {@code{dovecot-configuration} parameter} string syslog-facility
12251 Syslog facility to use if you're logging to syslog. Usually if you
12252 don't want to use @samp{mail}, you'll use local0..local7. Also other
12253 standard facilities are supported.
12254 Defaults to @samp{"mail"}.
12257 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-verbose?
12258 Log unsuccessful authentication attempts and the reasons why they
12260 Defaults to @samp{#f}.
12263 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-verbose-passwords?
12264 In case of password mismatches, log the attempted password. Valid
12265 values are no, plain and sha1. sha1 can be useful for detecting brute
12266 force password attempts vs. user simply trying the same password over
12267 and over again. You can also truncate the value to n chars by appending
12268 ":n" (e.g. sha1:6).
12269 Defaults to @samp{#f}.
12272 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-debug?
12273 Even more verbose logging for debugging purposes. Shows for example
12275 Defaults to @samp{#f}.
12278 @deftypevr {@code{dovecot-configuration} parameter} boolean auth-debug-passwords?
12279 In case of password mismatches, log the passwords and used scheme so
12280 the problem can be debugged. Enabling this also enables
12282 Defaults to @samp{#f}.
12285 @deftypevr {@code{dovecot-configuration} parameter} boolean mail-debug?
12286 Enable mail process debugging. This can help you figure out why
12287 Dovecot isn't finding your mails.
12288 Defaults to @samp{#f}.
12291 @deftypevr {@code{dovecot-configuration} parameter} boolean verbose-ssl?
12292 Show protocol level SSL errors.
12293 Defaults to @samp{#f}.
12296 @deftypevr {@code{dovecot-configuration} parameter} string log-timestamp
12297 Prefix for each line written to log file. % codes are in
12298 strftime(3) format.
12299 Defaults to @samp{"\"%b %d %H:%M:%S \""}.
12302 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list login-log-format-elements
12303 List of elements we want to log. The elements which have a
12304 non-empty variable value are joined together to form a comma-separated
12308 @deftypevr {@code{dovecot-configuration} parameter} string login-log-format
12309 Login log format. %s contains @samp{login-log-format-elements}
12310 string, %$ contains the data we want to log.
12311 Defaults to @samp{"%$: %s"}.
12314 @deftypevr {@code{dovecot-configuration} parameter} string mail-log-prefix
12315 Log prefix for mail processes. See doc/wiki/Variables.txt for list
12316 of possible variables you can use.
12317 Defaults to @samp{"\"%s(%u): \""}.
12320 @deftypevr {@code{dovecot-configuration} parameter} string deliver-log-format
12321 Format to use for logging mail deliveries. You can use variables:
12324 Delivery status message (e.g. @samp{saved to INBOX})
12336 Defaults to @samp{"msgid=%m: %$"}.
12339 @deftypevr {@code{dovecot-configuration} parameter} string mail-location
12340 Location for users' mailboxes. The default is empty, which means
12341 that Dovecot tries to find the mailboxes automatically. This won't work
12342 if the user doesn't yet have any mail, so you should explicitly tell
12343 Dovecot the full location.
12345 If you're using mbox, giving a path to the INBOX
12346 file (e.g. /var/mail/%u) isn't enough. You'll also need to tell Dovecot
12347 where the other mailboxes are kept. This is called the "root mail
12348 directory", and it must be the first path given in the
12349 @samp{mail-location} setting.
12351 There are a few special variables you can use, eg.:
12357 user part in user@@domain, same as %u if there's no domain
12359 domain part in user@@domain, empty if there's no domain
12364 See doc/wiki/Variables.txt for full list. Some examples:
12366 @item maildir:~/Maildir
12367 @item mbox:~/mail:INBOX=/var/mail/%u
12368 @item mbox:/var/mail/%d/%1n/%n:INDEX=/var/indexes/%d/%1n/%
12370 Defaults to @samp{""}.
12373 @deftypevr {@code{dovecot-configuration} parameter} string mail-uid
12374 System user and group used to access mails. If you use multiple,
12375 userdb can override these by returning uid or gid fields. You can use
12376 either numbers or names. <doc/wiki/UserIds.txt>.
12377 Defaults to @samp{""}.
12380 @deftypevr {@code{dovecot-configuration} parameter} string mail-gid
12382 Defaults to @samp{""}.
12385 @deftypevr {@code{dovecot-configuration} parameter} string mail-privileged-group
12386 Group to enable temporarily for privileged operations. Currently
12387 this is used only with INBOX when either its initial creation or
12388 dotlocking fails. Typically this is set to "mail" to give access to
12390 Defaults to @samp{""}.
12393 @deftypevr {@code{dovecot-configuration} parameter} string mail-access-groups
12394 Grant access to these supplementary groups for mail processes.
12395 Typically these are used to set up access to shared mailboxes. Note
12396 that it may be dangerous to set these if users can create
12397 symlinks (e.g. if "mail" group is set here, ln -s /var/mail ~/mail/var
12398 could allow a user to delete others' mailboxes, or ln -s
12399 /secret/shared/box ~/mail/mybox would allow reading it).
12400 Defaults to @samp{""}.
12403 @deftypevr {@code{dovecot-configuration} parameter} boolean mail-full-filesystem-access?
12404 Allow full file system access to clients. There's no access checks
12405 other than what the operating system does for the active UID/GID. It
12406 works with both maildir and mboxes, allowing you to prefix mailboxes
12407 names with e.g. /path/ or ~user/.
12408 Defaults to @samp{#f}.
12411 @deftypevr {@code{dovecot-configuration} parameter} boolean mmap-disable?
12412 Don't use mmap() at all. This is required if you store indexes to
12413 shared file systems (NFS or clustered file system).
12414 Defaults to @samp{#f}.
12417 @deftypevr {@code{dovecot-configuration} parameter} boolean dotlock-use-excl?
12418 Rely on @samp{O_EXCL} to work when creating dotlock files. NFS
12419 supports @samp{O_EXCL} since version 3, so this should be safe to use
12420 nowadays by default.
12421 Defaults to @samp{#t}.
12424 @deftypevr {@code{dovecot-configuration} parameter} string mail-fsync
12425 When to use fsync() or fdatasync() calls:
12428 Whenever necessary to avoid losing important data
12430 Useful with e.g. NFS when write()s are delayed
12432 Never use it (best performance, but crashes can lose data).
12434 Defaults to @samp{"optimized"}.
12437 @deftypevr {@code{dovecot-configuration} parameter} boolean mail-nfs-storage?
12438 Mail storage exists in NFS. Set this to yes to make Dovecot flush
12439 NFS caches whenever needed. If you're using only a single mail server
12441 Defaults to @samp{#f}.
12444 @deftypevr {@code{dovecot-configuration} parameter} boolean mail-nfs-index?
12445 Mail index files also exist in NFS. Setting this to yes requires
12446 @samp{mmap-disable? #t} and @samp{fsync-disable? #f}.
12447 Defaults to @samp{#f}.
12450 @deftypevr {@code{dovecot-configuration} parameter} string lock-method
12451 Locking method for index files. Alternatives are fcntl, flock and
12452 dotlock. Dotlocking uses some tricks which may create more disk I/O
12453 than other locking methods. NFS users: flock doesn't work, remember to
12454 change @samp{mmap-disable}.
12455 Defaults to @samp{"fcntl"}.
12458 @deftypevr {@code{dovecot-configuration} parameter} file-name mail-temp-dir
12459 Directory in which LDA/LMTP temporarily stores incoming mails >128
12461 Defaults to @samp{"/tmp"}.
12464 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer first-valid-uid
12465 Valid UID range for users. This is mostly to make sure that users can't
12466 log in as daemons or other system users. Note that denying root logins is
12467 hardcoded to dovecot binary and can't be done even if @samp{first-valid-uid}
12469 Defaults to @samp{500}.
12472 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer last-valid-uid
12474 Defaults to @samp{0}.
12477 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer first-valid-gid
12478 Valid GID range for users. Users having non-valid GID as primary group ID
12479 aren't allowed to log in. If user belongs to supplementary groups with
12480 non-valid GIDs, those groups are not set.
12481 Defaults to @samp{1}.
12484 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer last-valid-gid
12486 Defaults to @samp{0}.
12489 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer mail-max-keyword-length
12490 Maximum allowed length for mail keyword name. It's only forced when
12491 trying to create new keywords.
12492 Defaults to @samp{50}.
12495 @deftypevr {@code{dovecot-configuration} parameter} colon-separated-file-name-list valid-chroot-dirs
12496 List of directories under which chrooting is allowed for mail
12497 processes (i.e. /var/mail will allow chrooting to /var/mail/foo/bar
12498 too). This setting doesn't affect @samp{login-chroot}
12499 @samp{mail-chroot} or auth chroot settings. If this setting is empty,
12500 "/./" in home dirs are ignored. WARNING: Never add directories here
12501 which local users can modify, that may lead to root exploit. Usually
12502 this should be done only if you don't allow shell access for users.
12503 <doc/wiki/Chrooting.txt>.
12504 Defaults to @samp{()}.
12507 @deftypevr {@code{dovecot-configuration} parameter} string mail-chroot
12508 Default chroot directory for mail processes. This can be overridden
12509 for specific users in user database by giving /./ in user's home
12510 directory (e.g. /home/./user chroots into /home). Note that usually
12511 there is no real need to do chrooting, Dovecot doesn't allow users to
12512 access files outside their mail directory anyway. If your home
12513 directories are prefixed with the chroot directory, append "/." to
12514 @samp{mail-chroot}. <doc/wiki/Chrooting.txt>.
12515 Defaults to @samp{""}.
12518 @deftypevr {@code{dovecot-configuration} parameter} file-name auth-socket-path
12519 UNIX socket path to master authentication server to find users.
12520 This is used by imap (for shared users) and lda.
12521 Defaults to @samp{"/var/run/dovecot/auth-userdb"}.
12524 @deftypevr {@code{dovecot-configuration} parameter} file-name mail-plugin-dir
12525 Directory where to look up mail plugins.
12526 Defaults to @samp{"/usr/lib/dovecot"}.
12529 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list mail-plugins
12530 List of plugins to load for all services. Plugins specific to IMAP,
12531 LDA, etc. are added to this list in their own .conf files.
12532 Defaults to @samp{()}.
12535 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer mail-cache-min-mail-count
12536 The minimum number of mails in a mailbox before updates are done to
12537 cache file. This allows optimizing Dovecot's behavior to do less disk
12538 writes at the cost of more disk reads.
12539 Defaults to @samp{0}.
12542 @deftypevr {@code{dovecot-configuration} parameter} string mailbox-idle-check-interval
12543 When IDLE command is running, mailbox is checked once in a while to
12544 see if there are any new mails or other changes. This setting defines
12545 the minimum time to wait between those checks. Dovecot can also use
12546 dnotify, inotify and kqueue to find out immediately when changes
12548 Defaults to @samp{"30 secs"}.
12551 @deftypevr {@code{dovecot-configuration} parameter} boolean mail-save-crlf?
12552 Save mails with CR+LF instead of plain LF. This makes sending those
12553 mails take less CPU, especially with sendfile() syscall with Linux and
12554 FreeBSD. But it also creates a bit more disk I/O which may just make it
12555 slower. Also note that if other software reads the mboxes/maildirs,
12556 they may handle the extra CRs wrong and cause problems.
12557 Defaults to @samp{#f}.
12560 @deftypevr {@code{dovecot-configuration} parameter} boolean maildir-stat-dirs?
12561 By default LIST command returns all entries in maildir beginning
12562 with a dot. Enabling this option makes Dovecot return only entries
12563 which are directories. This is done by stat()ing each entry, so it
12564 causes more disk I/O.
12565 (For systems setting struct @samp{dirent->d_type} this check is free
12566 and it's done always regardless of this setting).
12567 Defaults to @samp{#f}.
12570 @deftypevr {@code{dovecot-configuration} parameter} boolean maildir-copy-with-hardlinks?
12571 When copying a message, do it with hard links whenever possible.
12572 This makes the performance much better, and it's unlikely to have any
12574 Defaults to @samp{#t}.
12577 @deftypevr {@code{dovecot-configuration} parameter} boolean maildir-very-dirty-syncs?
12578 Assume Dovecot is the only MUA accessing Maildir: Scan cur/
12579 directory only when its mtime changes unexpectedly or when we can't find
12580 the mail otherwise.
12581 Defaults to @samp{#f}.
12584 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list mbox-read-locks
12585 Which locking methods to use for locking mbox. There are four
12590 Create <mailbox>.lock file. This is the oldest and most NFS-safe
12591 solution. If you want to use /var/mail/ like directory, the users will
12592 need write access to that directory.
12594 Same as dotlock, but if it fails because of permissions or because there
12595 isn't enough disk space, just skip it.
12597 Use this if possible. Works with NFS too if lockd is used.
12599 May not exist in all systems. Doesn't work with NFS.
12601 May not exist in all systems. Doesn't work with NFS.
12604 You can use multiple locking methods; if you do the order they're declared
12605 in is important to avoid deadlocks if other MTAs/MUAs are using multiple
12606 locking methods as well. Some operating systems don't allow using some of
12607 them simultaneously.
12610 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list mbox-write-locks
12614 @deftypevr {@code{dovecot-configuration} parameter} string mbox-lock-timeout
12615 Maximum time to wait for lock (all of them) before aborting.
12616 Defaults to @samp{"5 mins"}.
12619 @deftypevr {@code{dovecot-configuration} parameter} string mbox-dotlock-change-timeout
12620 If dotlock exists but the mailbox isn't modified in any way,
12621 override the lock file after this much time.
12622 Defaults to @samp{"2 mins"}.
12625 @deftypevr {@code{dovecot-configuration} parameter} boolean mbox-dirty-syncs?
12626 When mbox changes unexpectedly we have to fully read it to find out
12627 what changed. If the mbox is large this can take a long time. Since
12628 the change is usually just a newly appended mail, it'd be faster to
12629 simply read the new mails. If this setting is enabled, Dovecot does
12630 this but still safely fallbacks to re-reading the whole mbox file
12631 whenever something in mbox isn't how it's expected to be. The only real
12632 downside to this setting is that if some other MUA changes message
12633 flags, Dovecot doesn't notice it immediately. Note that a full sync is
12634 done with SELECT, EXAMINE, EXPUNGE and CHECK commands.
12635 Defaults to @samp{#t}.
12638 @deftypevr {@code{dovecot-configuration} parameter} boolean mbox-very-dirty-syncs?
12639 Like @samp{mbox-dirty-syncs}, but don't do full syncs even with SELECT,
12640 EXAMINE, EXPUNGE or CHECK commands. If this is set,
12641 @samp{mbox-dirty-syncs} is ignored.
12642 Defaults to @samp{#f}.
12645 @deftypevr {@code{dovecot-configuration} parameter} boolean mbox-lazy-writes?
12646 Delay writing mbox headers until doing a full write sync (EXPUNGE
12647 and CHECK commands and when closing the mailbox). This is especially
12648 useful for POP3 where clients often delete all mails. The downside is
12649 that our changes aren't immediately visible to other MUAs.
12650 Defaults to @samp{#t}.
12653 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer mbox-min-index-size
12654 If mbox size is smaller than this (e.g. 100k), don't write index
12655 files. If an index file already exists it's still read, just not
12657 Defaults to @samp{0}.
12660 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer mdbox-rotate-size
12661 Maximum dbox file size until it's rotated.
12662 Defaults to @samp{2000000}.
12665 @deftypevr {@code{dovecot-configuration} parameter} string mdbox-rotate-interval
12666 Maximum dbox file age until it's rotated. Typically in days. Day
12667 begins from midnight, so 1d = today, 2d = yesterday, etc. 0 = check
12669 Defaults to @samp{"1d"}.
12672 @deftypevr {@code{dovecot-configuration} parameter} boolean mdbox-preallocate-space?
12673 When creating new mdbox files, immediately preallocate their size to
12674 @samp{mdbox-rotate-size}. This setting currently works only in Linux
12675 with some file systems (ext4, xfs).
12676 Defaults to @samp{#f}.
12679 @deftypevr {@code{dovecot-configuration} parameter} string mail-attachment-dir
12680 sdbox and mdbox support saving mail attachments to external files,
12681 which also allows single instance storage for them. Other backends
12682 don't support this for now.
12684 WARNING: This feature hasn't been tested much yet. Use at your own risk.
12686 Directory root where to store mail attachments. Disabled, if empty.
12687 Defaults to @samp{""}.
12690 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer mail-attachment-min-size
12691 Attachments smaller than this aren't saved externally. It's also
12692 possible to write a plugin to disable saving specific attachments
12694 Defaults to @samp{128000}.
12697 @deftypevr {@code{dovecot-configuration} parameter} string mail-attachment-fs
12698 File system backend to use for saving attachments:
12701 No SiS done by Dovecot (but this might help FS's own deduplication)
12703 SiS with immediate byte-by-byte comparison during saving
12704 @item sis-queue posix
12705 SiS with delayed comparison and deduplication.
12707 Defaults to @samp{"sis posix"}.
12710 @deftypevr {@code{dovecot-configuration} parameter} string mail-attachment-hash
12711 Hash format to use in attachment filenames. You can add any text and
12712 variables: @code{%@{md4@}}, @code{%@{md5@}}, @code{%@{sha1@}},
12713 @code{%@{sha256@}}, @code{%@{sha512@}}, @code{%@{size@}}. Variables can be
12714 truncated, e.g. @code{%@{sha256:80@}} returns only first 80 bits.
12715 Defaults to @samp{"%@{sha1@}"}.
12718 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer default-process-limit
12720 Defaults to @samp{100}.
12723 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer default-client-limit
12725 Defaults to @samp{1000}.
12728 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer default-vsz-limit
12729 Default VSZ (virtual memory size) limit for service processes.
12730 This is mainly intended to catch and kill processes that leak memory
12731 before they eat up everything.
12732 Defaults to @samp{256000000}.
12735 @deftypevr {@code{dovecot-configuration} parameter} string default-login-user
12736 Login user is internally used by login processes. This is the most
12737 untrusted user in Dovecot system. It shouldn't have access to anything
12739 Defaults to @samp{"dovenull"}.
12742 @deftypevr {@code{dovecot-configuration} parameter} string default-internal-user
12743 Internal user is used by unprivileged processes. It should be
12744 separate from login user, so that login processes can't disturb other
12746 Defaults to @samp{"dovecot"}.
12749 @deftypevr {@code{dovecot-configuration} parameter} string ssl?
12750 SSL/TLS support: yes, no, required. <doc/wiki/SSL.txt>.
12751 Defaults to @samp{"required"}.
12754 @deftypevr {@code{dovecot-configuration} parameter} string ssl-cert
12755 PEM encoded X.509 SSL/TLS certificate (public key).
12756 Defaults to @samp{"</etc/dovecot/default.pem"}.
12759 @deftypevr {@code{dovecot-configuration} parameter} string ssl-key
12760 PEM encoded SSL/TLS private key. The key is opened before
12761 dropping root privileges, so keep the key file unreadable by anyone but
12763 Defaults to @samp{"</etc/dovecot/private/default.pem"}.
12766 @deftypevr {@code{dovecot-configuration} parameter} string ssl-key-password
12767 If key file is password protected, give the password here.
12768 Alternatively give it when starting dovecot with -p parameter. Since
12769 this file is often world-readable, you may want to place this setting
12770 instead to a different.
12771 Defaults to @samp{""}.
12774 @deftypevr {@code{dovecot-configuration} parameter} string ssl-ca
12775 PEM encoded trusted certificate authority. Set this only if you
12776 intend to use @samp{ssl-verify-client-cert? #t}. The file should
12777 contain the CA certificate(s) followed by the matching
12778 CRL(s). (e.g. @samp{ssl-ca </etc/ssl/certs/ca.pem}).
12779 Defaults to @samp{""}.
12782 @deftypevr {@code{dovecot-configuration} parameter} boolean ssl-require-crl?
12783 Require that CRL check succeeds for client certificates.
12784 Defaults to @samp{#t}.
12787 @deftypevr {@code{dovecot-configuration} parameter} boolean ssl-verify-client-cert?
12788 Request client to send a certificate. If you also want to require
12789 it, set @samp{auth-ssl-require-client-cert? #t} in auth section.
12790 Defaults to @samp{#f}.
12793 @deftypevr {@code{dovecot-configuration} parameter} string ssl-cert-username-field
12794 Which field from certificate to use for username. commonName and
12795 x500UniqueIdentifier are the usual choices. You'll also need to set
12796 @samp{auth-ssl-username-from-cert? #t}.
12797 Defaults to @samp{"commonName"}.
12800 @deftypevr {@code{dovecot-configuration} parameter} hours ssl-parameters-regenerate
12801 How often to regenerate the SSL parameters file. Generation is
12802 quite CPU intensive operation. The value is in hours, 0 disables
12803 regeneration entirely.
12804 Defaults to @samp{168}.
12807 @deftypevr {@code{dovecot-configuration} parameter} string ssl-protocols
12808 SSL protocols to use.
12809 Defaults to @samp{"!SSLv2"}.
12812 @deftypevr {@code{dovecot-configuration} parameter} string ssl-cipher-list
12813 SSL ciphers to use.
12814 Defaults to @samp{"ALL:!LOW:!SSLv2:!EXP:!aNULL"}.
12817 @deftypevr {@code{dovecot-configuration} parameter} string ssl-crypto-device
12818 SSL crypto device to use, for valid values run "openssl engine".
12819 Defaults to @samp{""}.
12822 @deftypevr {@code{dovecot-configuration} parameter} string postmaster-address
12823 Address to use when sending rejection mails.
12824 %d expands to recipient domain.
12825 Defaults to @samp{"postmaster@@%d"}.
12828 @deftypevr {@code{dovecot-configuration} parameter} string hostname
12829 Hostname to use in various parts of sent mails (e.g. in Message-Id)
12830 and in LMTP replies. Default is the system's real hostname@@domain.
12831 Defaults to @samp{""}.
12834 @deftypevr {@code{dovecot-configuration} parameter} boolean quota-full-tempfail?
12835 If user is over quota, return with temporary failure instead of
12837 Defaults to @samp{#f}.
12840 @deftypevr {@code{dovecot-configuration} parameter} file-name sendmail-path
12841 Binary to use for sending mails.
12842 Defaults to @samp{"/usr/sbin/sendmail"}.
12845 @deftypevr {@code{dovecot-configuration} parameter} string submission-host
12846 If non-empty, send mails via this SMTP host[:port] instead of
12848 Defaults to @samp{""}.
12851 @deftypevr {@code{dovecot-configuration} parameter} string rejection-subject
12852 Subject: header to use for rejection mails. You can use the same
12853 variables as for @samp{rejection-reason} below.
12854 Defaults to @samp{"Rejected: %s"}.
12857 @deftypevr {@code{dovecot-configuration} parameter} string rejection-reason
12858 Human readable error message for rejection mails. You can use
12871 Defaults to @samp{"Your message to <%t> was automatically rejected:%n%r"}.
12874 @deftypevr {@code{dovecot-configuration} parameter} string recipient-delimiter
12875 Delimiter character between local-part and detail in email
12877 Defaults to @samp{"+"}.
12880 @deftypevr {@code{dovecot-configuration} parameter} string lda-original-recipient-header
12881 Header where the original recipient address (SMTP's RCPT TO:
12882 address) is taken from if not available elsewhere. With dovecot-lda -a
12883 parameter overrides this. A commonly used header for this is
12885 Defaults to @samp{""}.
12888 @deftypevr {@code{dovecot-configuration} parameter} boolean lda-mailbox-autocreate?
12889 Should saving a mail to a nonexistent mailbox automatically create
12891 Defaults to @samp{#f}.
12894 @deftypevr {@code{dovecot-configuration} parameter} boolean lda-mailbox-autosubscribe?
12895 Should automatically created mailboxes be also automatically
12897 Defaults to @samp{#f}.
12900 @deftypevr {@code{dovecot-configuration} parameter} non-negative-integer imap-max-line-length
12901 Maximum IMAP command line length. Some clients generate very long
12902 command lines with huge mailboxes, so you may need to raise this if you
12903 get "Too long argument" or "IMAP command line too large" errors
12905 Defaults to @samp{64000}.
12908 @deftypevr {@code{dovecot-configuration} parameter} string imap-logout-format
12909 IMAP logout format string:
12912 total number of bytes read from client
12914 total number of bytes sent to client.
12916 Defaults to @samp{"in=%i out=%o"}.
12919 @deftypevr {@code{dovecot-configuration} parameter} string imap-capability
12920 Override the IMAP CAPABILITY response. If the value begins with '+',
12921 add the given capabilities on top of the defaults (e.g. +XFOO XBAR).
12922 Defaults to @samp{""}.
12925 @deftypevr {@code{dovecot-configuration} parameter} string imap-idle-notify-interval
12926 How long to wait between "OK Still here" notifications when client
12928 Defaults to @samp{"2 mins"}.
12931 @deftypevr {@code{dovecot-configuration} parameter} string imap-id-send
12932 ID field names and values to send to clients. Using * as the value
12933 makes Dovecot use the default value. The following fields have default
12934 values currently: name, version, os, os-version, support-url,
12936 Defaults to @samp{""}.
12939 @deftypevr {@code{dovecot-configuration} parameter} string imap-id-log
12940 ID fields sent by client to log. * means everything.
12941 Defaults to @samp{""}.
12944 @deftypevr {@code{dovecot-configuration} parameter} space-separated-string-list imap-client-workarounds
12945 Workarounds for various client bugs:
12948 @item delay-newmail
12949 Send EXISTS/RECENT new mail notifications only when replying to NOOP and
12950 CHECK commands. Some clients ignore them otherwise, for example OSX
12951 Mail (<v2.1). Outlook Express breaks more badly though, without this it
12952 may show user "Message no longer in server" errors. Note that OE6
12953 still breaks even with this workaround if synchronization is set to
12956 @item tb-extra-mailbox-sep
12957 Thunderbird gets somehow confused with LAYOUT=fs (mbox and dbox) and
12958 adds extra @samp{/} suffixes to mailbox names. This option causes Dovecot to
12959 ignore the extra @samp{/} instead of treating it as invalid mailbox name.
12961 @item tb-lsub-flags
12962 Show \Noselect flags for LSUB replies with LAYOUT=fs (e.g. mbox).
12963 This makes Thunderbird realize they aren't selectable and show them
12964 greyed out, instead of only later giving "not selectable" popup error.
12966 Defaults to @samp{()}.
12969 @deftypevr {@code{dovecot-configuration} parameter} string imap-urlauth-host
12970 Host allowed in URLAUTH URLs sent by client. "*" allows all.
12971 Defaults to @samp{""}.
12975 Whew! Lots of configuration options. The nice thing about it though is
12976 that GuixSD has a complete interface to Dovecot's configuration
12977 language. This allows not only a nice way to declare configurations,
12978 but also offers reflective capabilities as well: users can write code to
12979 inspect and transform configurations from within Scheme.
12981 However, it could be that you just want to get a @code{dovecot.conf} up
12982 and running. In that case, you can pass an
12983 @code{opaque-dovecot-configuration} as the @code{#:config} parameter to
12984 @code{dovecot-service}. As its name indicates, an opaque configuration
12985 does not have easy reflective capabilities.
12987 Available @code{opaque-dovecot-configuration} fields are:
12989 @deftypevr {@code{opaque-dovecot-configuration} parameter} package dovecot
12990 The dovecot package.
12993 @deftypevr {@code{opaque-dovecot-configuration} parameter} string string
12994 The contents of the @code{dovecot.conf}, as a string.
12997 For example, if your @code{dovecot.conf} is just the empty string, you
12998 could instantiate a dovecot service like this:
13001 (dovecot-service #:config
13002 (opaque-dovecot-configuration
13006 @subsubheading OpenSMTPD Service
13008 @deffn {Scheme Variable} opensmtpd-service-type
13009 This is the type of the @uref{https://www.opensmtpd.org, OpenSMTPD}
13010 service, whose value should be an @code{opensmtpd-configuration} object
13011 as in this example:
13014 (service opensmtpd-service-type
13015 (opensmtpd-configuration
13016 (config-file (local-file "./my-smtpd.conf"))))
13020 @deftp {Data Type} opensmtpd-configuration
13021 Data type representing the configuration of opensmtpd.
13024 @item @code{package} (default: @var{opensmtpd})
13025 Package object of the OpenSMTPD SMTP server.
13027 @item @code{config-file} (default: @var{%default-opensmtpd-file})
13028 File-like object of the OpenSMTPD configuration file to use. By default
13029 it listens on the loopback network interface, and allows for mail from
13030 users and daemons on the local machine, as well as permitting email to
13031 remote servers. Run @command{man smtpd.conf} for more information.
13036 @subsubheading Exim Service
13038 @cindex mail transfer agent (MTA)
13039 @cindex MTA (mail transfer agent)
13042 @deffn {Scheme Variable} exim-service-type
13043 This is the type of the @uref{https://exim.org, Exim} mail transfer
13044 agent (MTA), whose value should be an @code{exim-configuration} object
13045 as in this example:
13048 (service exim-service-type
13049 (exim-configuration
13050 (config-file (local-file "./my-exim.conf"))))
13054 In order to use an @code{exim-service-type} service you must also have a
13055 @code{mail-aliases-service-type} service present in your
13056 @code{operating-system} (even if it has no aliases).
13058 @deftp {Data Type} exim-configuration
13059 Data type representing the configuration of exim.
13062 @item @code{package} (default: @var{exim})
13063 Package object of the Exim server.
13065 @item @code{config-file} (default: @code{#f})
13066 File-like object of the Exim configuration file to use. If its value is
13067 @code{#f} then use the default configuration file from the package
13068 provided in @code{package}. The resulting configuration file is loaded
13069 after setting the @code{exim_user} and @code{exim_group} configuration
13075 @subsubheading Mail Aliases Service
13077 @cindex email aliases
13078 @cindex aliases, for email addresses
13080 @deffn {Scheme Variable} mail-aliases-service-type
13081 This is the type of the service which provides @code{/etc/aliases},
13082 specifying how to deliver mail to users on this system.
13085 (service mail-aliases-service-type
13086 '(("postmaster" "bob")
13087 ("bob" "bob@@example.com" "bob@@example2.com")))
13091 The configuration for a @code{mail-aliases-service-type} service is an
13092 association list denoting how to deliver mail that comes to this
13093 system. Each entry is of the form @code{(alias addresses ...)}, with
13094 @code{alias} specifying the local alias and @code{addresses} specifying
13095 where to deliver this user's mail.
13097 The aliases aren't required to exist as users on the local system. In
13098 the above example, there doesn't need to be a @code{postmaster} entry in
13099 the @code{operating-system}'s @code{user-accounts} in order to deliver
13100 the @code{postmaster} mail to @code{bob} (which subsequently would
13101 deliver mail to @code{bob@@example.com} and @code{bob@@example2.com}).
13103 @node Messaging Services
13104 @subsubsection Messaging Services
13109 The @code{(gnu services messaging)} module provides Guix service
13110 definitions for messaging services: currently only Prosody is supported.
13112 @subsubheading Prosody Service
13114 @deffn {Scheme Variable} prosody-service-type
13115 This is the type for the @uref{http://prosody.im, Prosody XMPP
13116 communication server}. Its value must be a @code{prosody-configuration}
13117 record as in this example:
13120 (service prosody-service-type
13121 (prosody-configuration
13122 (modules-enabled (cons "groups" %default-modules-enabled))
13125 (int-component-configuration
13126 (hostname "conference.example.net")
13128 (mod-muc (mod-muc-configuration)))))
13131 (virtualhost-configuration
13132 (domain "example.net"))))))
13135 See below for details about @code{prosody-configuration}.
13139 By default, Prosody does not need much configuration. Only one
13140 @code{virtualhosts} field is needed: it specifies the domain you wish
13143 Prosodyctl will help you generate X.509 certificates and keys:
13146 prosodyctl cert request example.net
13149 The available configuration parameters follow. Each parameter
13150 definition is preceded by its type; for example, @samp{string-list foo}
13151 indicates that the @code{foo} parameter should be specified as a list of
13152 strings. Types starting with @code{maybe-} denote parameters that won't
13153 show up in @code{prosody.cfg.lua} when their value is @code{'disabled}.
13155 There is also a way to specify the configuration as a string, if you
13156 have an old @code{prosody.cfg.lua} file that you want to port over from
13157 some other system; see the end for more details.
13159 @c The following documentation was initially generated by
13160 @c (generate-documentation) in (gnu services messaging). Manually maintained
13161 @c documentation is better, so we shouldn't hesitate to edit below as
13162 @c needed. However if the change you want to make to this documentation
13163 @c can be done in an automated way, it's probably easier to change
13164 @c (generate-documentation) than to make it below and have to deal with
13165 @c the churn as Prosody updates.
13167 Available @code{prosody-configuration} fields are:
13169 @deftypevr {@code{prosody-configuration} parameter} package prosody
13170 The Prosody package.
13173 @deftypevr {@code{prosody-configuration} parameter} file-name data-path
13174 Location of the Prosody data storage directory. See
13175 @url{http://prosody.im/doc/configure}.
13176 Defaults to @samp{"/var/lib/prosody"}.
13179 @deftypevr {@code{prosody-configuration} parameter} file-name-list plugin-paths
13180 Additional plugin directories. They are searched in all the specified
13181 paths in order. See @url{http://prosody.im/doc/plugins_directory}.
13182 Defaults to @samp{()}.
13185 @deftypevr {@code{prosody-configuration} parameter} string-list admins
13186 This is a list of accounts that are admins for the server. Note that you
13187 must create the accounts separately. See @url{http://prosody.im/doc/admins} and
13188 @url{http://prosody.im/doc/creating_accounts}.
13189 Example: @code{(admins '("user1@@example.com" "user2@@example.net"))}
13190 Defaults to @samp{()}.
13193 @deftypevr {@code{prosody-configuration} parameter} boolean use-libevent?
13194 Enable use of libevent for better performance under high load. See
13195 @url{http://prosody.im/doc/libevent}.
13196 Defaults to @samp{#f}.
13199 @deftypevr {@code{prosody-configuration} parameter} module-list modules-enabled
13200 This is the list of modules Prosody will load on startup. It looks for
13201 @code{mod_modulename.lua} in the plugins folder, so make sure that exists too.
13202 Documentation on modules can be found at: @url{http://prosody.im/doc/modules}.
13203 Defaults to @samp{%default-modules-enabled}.
13206 @deftypevr {@code{prosody-configuration} parameter} string-list modules-disabled
13207 @samp{"offline"}, @samp{"c2s"} and @samp{"s2s"} are auto-loaded, but
13208 should you want to disable them then add them to this list.
13209 Defaults to @samp{()}.
13212 @deftypevr {@code{prosody-configuration} parameter} file-name groups-file
13213 Path to a text file where the shared groups are defined. If this path is
13214 empty then @samp{mod_groups} does nothing. See
13215 @url{http://prosody.im/doc/modules/mod_groups}.
13216 Defaults to @samp{"/var/lib/prosody/sharedgroups.txt"}.
13219 @deftypevr {@code{prosody-configuration} parameter} boolean allow-registration?
13220 Disable account creation by default, for security. See
13221 @url{http://prosody.im/doc/creating_accounts}.
13222 Defaults to @samp{#f}.
13225 @deftypevr {@code{prosody-configuration} parameter} maybe-ssl-configuration ssl
13226 These are the SSL/TLS-related settings. Most of them are disabled so to
13227 use Prosody's defaults. If you do not completely understand these options, do
13228 not add them to your config, it is easy to lower the security of your server
13229 using them. See @url{http://prosody.im/doc/advanced_ssl_config}.
13231 Available @code{ssl-configuration} fields are:
13233 @deftypevr {@code{ssl-configuration} parameter} maybe-string protocol
13234 This determines what handshake to use.
13237 @deftypevr {@code{ssl-configuration} parameter} file-name key
13238 Path to your private key file, relative to @code{/etc/prosody}.
13239 Defaults to @samp{"/etc/prosody/certs/key.pem"}.
13242 @deftypevr {@code{ssl-configuration} parameter} file-name certificate
13243 Path to your certificate file, relative to @code{/etc/prosody}.
13244 Defaults to @samp{"/etc/prosody/certs/cert.pem"}.
13247 @deftypevr {@code{ssl-configuration} parameter} file-name capath
13248 Path to directory containing root certificates that you wish Prosody to
13249 trust when verifying the certificates of remote servers.
13250 Defaults to @samp{"/etc/ssl/certs"}.
13253 @deftypevr {@code{ssl-configuration} parameter} maybe-file-name cafile
13254 Path to a file containing root certificates that you wish Prosody to trust.
13255 Similar to @code{capath} but with all certificates concatenated together.
13258 @deftypevr {@code{ssl-configuration} parameter} maybe-string-list verify
13259 A list of verification options (these mostly map to OpenSSL's
13260 @code{set_verify()} flags).
13263 @deftypevr {@code{ssl-configuration} parameter} maybe-string-list options
13264 A list of general options relating to SSL/TLS. These map to OpenSSL's
13265 @code{set_options()}. For a full list of options available in LuaSec, see the
13269 @deftypevr {@code{ssl-configuration} parameter} maybe-non-negative-integer depth
13270 How long a chain of certificate authorities to check when looking for a
13271 trusted root certificate.
13274 @deftypevr {@code{ssl-configuration} parameter} maybe-string ciphers
13275 An OpenSSL cipher string. This selects what ciphers Prosody will offer to
13276 clients, and in what order.
13279 @deftypevr {@code{ssl-configuration} parameter} maybe-file-name dhparam
13280 A path to a file containing parameters for Diffie-Hellman key exchange. You
13281 can create such a file with:
13282 @code{openssl dhparam -out /etc/prosody/certs/dh-2048.pem 2048}
13285 @deftypevr {@code{ssl-configuration} parameter} maybe-string curve
13286 Curve for Elliptic curve Diffie-Hellman. Prosody's default is
13287 @samp{"secp384r1"}.
13290 @deftypevr {@code{ssl-configuration} parameter} maybe-string-list verifyext
13291 A list of "extra" verification options.
13294 @deftypevr {@code{ssl-configuration} parameter} maybe-string password
13295 Password for encrypted private keys.
13300 @deftypevr {@code{prosody-configuration} parameter} boolean c2s-require-encryption?
13301 Whether to force all client-to-server connections to be encrypted or not.
13302 See @url{http://prosody.im/doc/modules/mod_tls}.
13303 Defaults to @samp{#f}.
13306 @deftypevr {@code{prosody-configuration} parameter} boolean s2s-require-encryption?
13307 Whether to force all server-to-server connections to be encrypted or not.
13308 See @url{http://prosody.im/doc/modules/mod_tls}.
13309 Defaults to @samp{#f}.
13312 @deftypevr {@code{prosody-configuration} parameter} boolean s2s-secure-auth?
13313 Whether to require encryption and certificate authentication. This
13314 provides ideal security, but requires servers you communicate with to support
13315 encryption AND present valid, trusted certificates. See
13316 @url{http://prosody.im/doc/s2s#security}.
13317 Defaults to @samp{#f}.
13320 @deftypevr {@code{prosody-configuration} parameter} string-list s2s-insecure-domains
13321 Many servers don't support encryption or have invalid or self-signed
13322 certificates. You can list domains here that will not be required to
13323 authenticate using certificates. They will be authenticated using DNS. See
13324 @url{http://prosody.im/doc/s2s#security}.
13325 Defaults to @samp{()}.
13328 @deftypevr {@code{prosody-configuration} parameter} string-list s2s-secure-domains
13329 Even if you leave @code{s2s-secure-auth?} disabled, you can still require
13330 valid certificates for some domains by specifying a list here. See
13331 @url{http://prosody.im/doc/s2s#security}.
13332 Defaults to @samp{()}.
13335 @deftypevr {@code{prosody-configuration} parameter} string authentication
13336 Select the authentication backend to use. The default provider stores
13337 passwords in plaintext and uses Prosody's configured data storage to store the
13338 authentication data. If you do not trust your server please see
13339 @url{http://prosody.im/doc/modules/mod_auth_internal_hashed} for information
13340 about using the hashed backend. See also
13341 @url{http://prosody.im/doc/authentication}
13342 Defaults to @samp{"internal_plain"}.
13345 @deftypevr {@code{prosody-configuration} parameter} maybe-string log
13346 Set logging options. Advanced logging configuration is not yet supported
13347 by the GuixSD Prosody Service. See @url{http://prosody.im/doc/logging}.
13348 Defaults to @samp{"*syslog"}.
13351 @deftypevr {@code{prosody-configuration} parameter} file-name pidfile
13352 File to write pid in. See @url{http://prosody.im/doc/modules/mod_posix}.
13353 Defaults to @samp{"/var/run/prosody/prosody.pid"}.
13356 @deftypevr {@code{prosody-configuration} parameter} virtualhost-configuration-list virtualhosts
13357 A host in Prosody is a domain on which user accounts can be created. For
13358 example if you want your users to have addresses like
13359 @samp{"john.smith@@example.com"} then you need to add a host
13360 @samp{"example.com"}. All options in this list will apply only to this host.
13362 Note: the name "virtual" host is used in configuration to avoid confusion with
13363 the actual physical host that Prosody is installed on. A single Prosody
13364 instance can serve many domains, each one defined as a VirtualHost entry in
13365 Prosody's configuration. Conversely a server that hosts a single domain would
13366 have just one VirtualHost entry.
13368 See @url{http://prosody.im/doc/configure#virtual_host_settings}.
13370 Available @code{virtualhost-configuration} fields are:
13372 all these @code{prosody-configuration} fields: @code{admins}, @code{use-libevent?}, @code{modules-enabled}, @code{modules-disabled}, @code{groups-file}, @code{allow-registration?}, @code{ssl}, @code{c2s-require-encryption?}, @code{s2s-require-encryption?}, @code{s2s-secure-auth?}, @code{s2s-insecure-domains}, @code{s2s-secure-domains}, @code{authentication}, @code{log}, plus:
13373 @deftypevr {@code{virtualhost-configuration} parameter} string domain
13374 Domain you wish Prosody to serve.
13379 @deftypevr {@code{prosody-configuration} parameter} int-component-configuration-list int-components
13380 Components are extra services on a server which are available to clients,
13381 usually on a subdomain of the main server (such as
13382 @samp{"mycomponent.example.com"}). Example components might be chatroom
13383 servers, user directories, or gateways to other protocols.
13385 Internal components are implemented with Prosody-specific plugins. To add an
13386 internal component, you simply fill the hostname field, and the plugin you wish
13387 to use for the component.
13389 See @url{http://prosody.im/doc/components}.
13390 Defaults to @samp{()}.
13392 Available @code{int-component-configuration} fields are:
13394 all these @code{prosody-configuration} fields: @code{admins}, @code{use-libevent?}, @code{modules-enabled}, @code{modules-disabled}, @code{groups-file}, @code{allow-registration?}, @code{ssl}, @code{c2s-require-encryption?}, @code{s2s-require-encryption?}, @code{s2s-secure-auth?}, @code{s2s-insecure-domains}, @code{s2s-secure-domains}, @code{authentication}, @code{log}, plus:
13395 @deftypevr {@code{int-component-configuration} parameter} string hostname
13396 Hostname of the component.
13399 @deftypevr {@code{int-component-configuration} parameter} string plugin
13400 Plugin you wish to use for the component.
13403 @deftypevr {@code{int-component-configuration} parameter} maybe-mod-muc-configuration mod-muc
13404 Multi-user chat (MUC) is Prosody's module for allowing you to create
13405 hosted chatrooms/conferences for XMPP users.
13407 General information on setting up and using multi-user chatrooms can be found
13408 in the "Chatrooms" documentation (@url{http://prosody.im/doc/chatrooms}),
13409 which you should read if you are new to XMPP chatrooms.
13411 See also @url{http://prosody.im/doc/modules/mod_muc}.
13413 Available @code{mod-muc-configuration} fields are:
13415 @deftypevr {@code{mod-muc-configuration} parameter} string name
13416 The name to return in service discovery responses.
13417 Defaults to @samp{"Prosody Chatrooms"}.
13420 @deftypevr {@code{mod-muc-configuration} parameter} string-or-boolean restrict-room-creation
13421 If @samp{#t}, this will only allow admins to create new chatrooms.
13422 Otherwise anyone can create a room. The value @samp{"local"} restricts room
13423 creation to users on the service's parent domain. E.g. @samp{user@@example.com}
13424 can create rooms on @samp{rooms.example.com}. The value @samp{"admin"}
13425 restricts to service administrators only.
13426 Defaults to @samp{#f}.
13429 @deftypevr {@code{mod-muc-configuration} parameter} non-negative-integer max-history-messages
13430 Maximum number of history messages that will be sent to the member that has
13431 just joined the room.
13432 Defaults to @samp{20}.
13439 @deftypevr {@code{prosody-configuration} parameter} ext-component-configuration-list ext-components
13440 External components use XEP-0114, which most standalone components
13441 support. To add an external component, you simply fill the hostname field. See
13442 @url{http://prosody.im/doc/components}.
13443 Defaults to @samp{()}.
13445 Available @code{ext-component-configuration} fields are:
13447 all these @code{prosody-configuration} fields: @code{admins}, @code{use-libevent?}, @code{modules-enabled}, @code{modules-disabled}, @code{groups-file}, @code{allow-registration?}, @code{ssl}, @code{c2s-require-encryption?}, @code{s2s-require-encryption?}, @code{s2s-secure-auth?}, @code{s2s-insecure-domains}, @code{s2s-secure-domains}, @code{authentication}, @code{log}, plus:
13448 @deftypevr {@code{ext-component-configuration} parameter} string component-secret
13449 Password which the component will use to log in.
13452 @deftypevr {@code{ext-component-configuration} parameter} string hostname
13453 Hostname of the component.
13458 @deftypevr {@code{prosody-configuration} parameter} non-negative-integer-list component-ports
13459 Port(s) Prosody listens on for component connections.
13462 @deftypevr {@code{prosody-configuration} parameter} string component-interface
13463 Interface Prosody listens on for component connections.
13464 Defaults to @samp{"127.0.0.1"}.
13467 It could be that you just want to get a @code{prosody.cfg.lua}
13468 up and running. In that case, you can pass an
13469 @code{opaque-prosody-configuration} record as the value of
13470 @code{prosody-service-type}. As its name indicates, an opaque configuration
13471 does not have easy reflective capabilities.
13472 Available @code{opaque-prosody-configuration} fields are:
13474 @deftypevr {@code{opaque-prosody-configuration} parameter} package prosody
13475 The prosody package.
13478 @deftypevr {@code{opaque-prosody-configuration} parameter} string prosody.cfg.lua
13479 The contents of the @code{prosody.cfg.lua} to use.
13482 For example, if your @code{prosody.cfg.lua} is just the empty
13483 string, you could instantiate a prosody service like this:
13486 (service prosody-service-type
13487 (opaque-prosody-configuration
13488 (prosody.cfg.lua "")))
13491 @node Kerberos Services
13492 @subsubsection Kerberos Services
13495 The @code{(gnu services kerberos)} module provides services relating to
13496 the authentication protocol @dfn{Kerberos}.
13498 @subsubheading Krb5 Service
13500 Programs using a Kerberos client library normally
13501 expect a configuration file in @file{/etc/krb5.conf}.
13502 This service generates such a file from a definition provided in the
13503 operating system declaration.
13504 It does not cause any daemon to be started.
13506 No ``keytab'' files are provided by this service---you must explicitly create them.
13507 This service is known to work with the MIT client library, @code{mit-krb5}.
13508 Other implementations have not been tested.
13510 @defvr {Scheme Variable} krb5-service-type
13511 A service type for Kerberos 5 clients.
13515 Here is an example of its use:
13517 (service krb5-service-type
13518 (krb5-configuration
13519 (default-realm "EXAMPLE.COM")
13520 (allow-weak-crypto? #t)
13523 (name "EXAMPLE.COM")
13524 (admin-server "groucho.example.com")
13525 (kdc "karl.example.com"))
13528 (admin-server "kerb-admin.argrx.edu")
13529 (kdc "keys.argrx.edu"))))))
13533 This example provides a Kerberos@tie{}5 client configuration which:
13535 @item Recognizes two realms, @i{viz:} ``EXAMPLE.COM'' and ``ARGRX.EDU'', both
13536 of which have distinct administration servers and key distribution centers;
13537 @item Will default to the realm ``EXAMPLE.COM'' if the realm is not explicitly
13538 specified by clients;
13539 @item Accepts services which only support encryption types known to be weak.
13542 The @code{krb5-realm} and @code{krb5-configuration} types have many fields.
13543 Only the most commonly used ones are described here.
13544 For a full list, and more detailed explanation of each, see the MIT
13545 @uref{http://web.mit.edu/kerberos/krb5-devel/doc/admin/conf_files/krb5_conf.html,,krb5.conf}
13549 @deftp {Data Type} krb5-realm
13550 @cindex realm, kerberos
13553 This field is a string identifying the name of the realm.
13554 A common convention is to use the fully qualified DNS name of your organization,
13555 converted to upper case.
13557 @item @code{admin-server}
13558 This field is a string identifying the host where the administration server is
13562 This field is a string identifying the key distribution center
13567 @deftp {Data Type} krb5-configuration
13570 @item @code{allow-weak-crypto?} (default: @code{#f})
13571 If this flag is @code{#t} then services which only offer encryption algorithms
13572 known to be weak will be accepted.
13574 @item @code{default-realm} (default: @code{#f})
13575 This field should be a string identifying the default Kerberos
13576 realm for the client.
13577 You should set this field to the name of your Kerberos realm.
13578 If this value is @code{#f}
13579 then a realm must be specified with every Kerberos principal when invoking programs
13580 such as @command{kinit}.
13582 @item @code{realms}
13583 This should be a non-empty list of @code{krb5-realm} objects, which clients may
13585 Normally, one of them will have a @code{name} field matching the @code{default-realm}
13591 @subsubheading PAM krb5 Service
13594 The @code{pam-krb5} service allows for login authentication and password
13595 management via Kerberos.
13596 You will need this service if you want PAM enabled applications to authenticate
13597 users using Kerberos.
13599 @defvr {Scheme Variable} pam-krb5-service-type
13600 A service type for the Kerberos 5 PAM module.
13603 @deftp {Data Type} pam-krb5-configuration
13604 Data type representing the configuration of the Kerberos 5 PAM module
13605 This type has the following parameters:
13607 @item @code{pam-krb5} (default: @code{pam-krb5})
13608 The pam-krb5 package to use.
13610 @item @code{minimum-uid} (default: @code{1000})
13611 The smallest user ID for which Kerberos authentications should be attempted.
13612 Local accounts with lower values will silently fail to authenticate.
13618 @subsubsection Web Services
13623 The @code{(gnu services web)} module provides the following service:
13625 @deffn {Scheme Procedure} nginx-service [#:nginx nginx] @
13626 [#:log-directory ``/var/log/nginx''] @
13627 [#:run-directory ``/var/run/nginx''] @
13628 [#:server-list '()] @
13629 [#:upstream-list '()] @
13630 [#:config-file @code{#f}]
13632 Return a service that runs @var{nginx}, the nginx web server.
13634 The nginx daemon loads its runtime configuration from @var{config-file}.
13635 Log files are written to @var{log-directory} and temporary runtime data
13636 files are written to @var{run-directory}. For proper operation, these
13637 arguments should match what is in @var{config-file} to ensure that the
13638 directories are created when the service is activated.
13640 As an alternative to using a @var{config-file}, @var{server-list} can be
13641 used to specify the list of @dfn{server blocks} required on the host and
13642 @var{upstream-list} can be used to specify a list of @dfn{upstream
13643 blocks} to configure. For this to work, use the default value for
13646 At startup, @command{nginx} has not yet read its configuration file, so it
13647 uses a default file to log error messages. If it fails to load its
13648 configuration file, that is where error messages are logged. After the
13649 configuration file is loaded, the default error log file changes as per
13650 configuration. In our case, startup error messages can be found in
13651 @file{/var/run/nginx/logs/error.log}, and after configuration in
13652 @file{/var/log/nginx/error.log}. The second location can be changed with the
13653 @var{log-directory} configuration option.
13657 @deffn {Scheme Variable} nginx-service-type
13658 This is type for the nginx web server.
13660 This service can be extended to add server blocks in addition to the
13661 default one, as in this example:
13664 (simple-service 'my-extra-server nginx-service-type
13665 (list (nginx-server-configuration
13667 (root "/srv/http/extra-website"))))
13671 @deftp {Data Type} nginx-server-configuration
13672 Data type representing the configuration of an nginx server block.
13673 This type has the following parameters:
13676 @item @code{http-port} (default: @code{80})
13677 Nginx will listen for HTTP connection on this port. Set it at @code{#f} if
13678 nginx should not listen for HTTP (non secure) connection for this
13679 @dfn{server block}.
13681 @item @code{https-port} (default: @code{443})
13682 Nginx will listen for HTTPS connection on this port. Set it at @code{#f} if
13683 nginx should not listen for HTTPS (secure) connection for this @dfn{server block}.
13685 Note that nginx can listen for HTTP and HTTPS connections in the same
13686 @dfn{server block}.
13688 @item @code{server-name} (default: @code{(list 'default)})
13689 A list of server names this server represents. @code{'default} represents the
13690 default server for connections matching no other server.
13692 @item @code{root} (default: @code{"/srv/http"})
13693 Root of the website nginx will serve.
13695 @item @code{locations} (default: @code{'()})
13696 A list of @dfn{nginx-location-configuration} or
13697 @dfn{nginx-named-location-configuration} records to use within this
13700 @item @code{index} (default: @code{(list "index.html")})
13701 Index files to look for when clients ask for a directory. If it cannot be found,
13702 Nginx will send the list of files in the directory.
13704 @item @code{ssl-certificate} (default: @code{"/etc/nginx/cert.pem"})
13705 Where to find the certificate for secure connections. Set it to @code{#f} if
13706 you don't have a certificate or you don't want to use HTTPS.
13708 @item @code{ssl-certificate-key} (default: @code{"/etc/nginx/key.pem"})
13709 Where to find the private key for secure connections. Set it to @code{#f} if
13710 you don't have a key or you don't want to use HTTPS.
13712 @item @code{server-tokens?} (default: @code{#f})
13713 Whether the server should add its configuration to response.
13718 @deftp {Data Type} nginx-upstream-configuration
13719 Data type representing the configuration of an nginx @code{upstream}
13720 block. This type has the following parameters:
13724 Name for this group of servers.
13726 @item @code{servers}
13727 Specify the addresses of the servers in the group. The address can be
13728 specified as a IP address (e.g. @samp{127.0.0.1}), domain name
13729 (e.g. @samp{backend1.example.com}) or a path to a UNIX socket using the
13730 prefix @samp{unix:}. For addresses using an IP address or domain name,
13731 the default port is 80, and a different port can be specified
13737 @deftp {Data Type} nginx-location-configuration
13738 Data type representing the configuration of an nginx @code{location}
13739 block. This type has the following parameters:
13743 URI which this location block matches.
13745 @anchor{nginx-location-configuration body}
13747 Body of the location block, specified as a string. This can contain many
13748 configuration directives. For example, to pass requests to a upstream
13749 server group defined using an @code{nginx-upstream-configuration} block,
13750 the following directive would be specified in the body @samp{proxy_pass
13751 http://upstream-name;}.
13756 @deftp {Data Type} nginx-named-location-configuration
13757 Data type representing the configuration of an nginx named location
13758 block. Named location blocks are used for request redirection, and not
13759 used for regular request processing. This type has the following
13764 Name to identify this location block.
13767 @xref{nginx-location-configuration body}, as the body for named location
13768 blocks can be used in a similar way to the
13769 @code{nginx-location-configuration body}. One restriction is that the
13770 body of a named location block cannot contain location blocks.
13777 @subsubsection DNS Services
13778 @cindex DNS (domain name system)
13779 @cindex domain name system (DNS)
13781 The @code{(gnu services dns)} module provides services related to the
13782 @dfn{domain name system} (DNS). It provides a server service for hosting
13783 an @emph{authoritative} DNS server for multiple zones, slave or master.
13784 This service uses @uref{https://www.knot-dns.cz/, Knot DNS}.
13786 An example configuration of an authoritative server for two zones, one master
13790 (define-zone-entries example.org.zone
13791 ;; Name TTL Class Type Data
13792 ("@@" "" "IN" "A" "127.0.0.1")
13793 ("@@" "" "IN" "NS" "ns")
13794 ("ns" "" "IN" "A" "127.0.0.1"))
13796 (define master-zone
13797 (knot-zone-configuration
13798 (domain "example.org")
13800 (origin "example.org")
13801 (entries example.org.zone)))))
13804 (knot-zone-configuration
13805 (domain "plop.org")
13806 (dnssec-policy "default")
13807 (master (list "plop-master"))))
13809 (define plop-master
13810 (knot-remote-configuration
13812 (address (list "208.76.58.171"))))
13816 (services (cons* (service knot-service-type
13817 (knot-confifguration
13818 (remotes (list plop-master))
13819 (zones (list master-zone slave-zone))))
13824 @deffn {Scheme Variable} knot-service-type
13825 This is the type for the Knot DNS server.
13827 Knot DNS is an authoritative DNS server, meaning that it can serve multiple
13828 zones, that is to say domain names you would buy from a registrar. This server
13829 is not a resolver, meaning that it can only resolve names for which it is
13830 authoritative. This server can be configured to serve zones as a master server
13831 or a slave server as a per-zone basis. Slave zones will get their data from
13832 masters, and will serve it as an authoritative server. From the point of view
13833 of a resolver, there is no difference between master and slave.
13835 The following data types are used to configure the Knot DNS server:
13838 @deftp {Data Type} knot-key-configuration
13839 Data type representing a key.
13840 This type has the following parameters:
13843 @item @code{id} (default: @code{""})
13844 An identifier for other configuration fields to refer to this key. IDs must
13845 be unique and must not be empty.
13847 @item @code{algorithm} (default: @code{#f})
13848 The algorithm to use. Choose between @code{#f}, @code{'hmac-md5},
13849 @code{'hmac-sha1}, @code{'hmac-sha224}, @code{'hmac-sha256}, @code{'hmac-sha384}
13850 and @code{'hmac-sha512}.
13852 @item @code{secret} (default: @code{""})
13853 The secret key itself.
13858 @deftp {Data Type} knot-acl-configuration
13859 Data type representing an Access Control List (ACL) configuration.
13860 This type has the following parameters:
13863 @item @code{id} (default: @code{""})
13864 An identifier for ether configuration fields to refer to this key. IDs must be
13865 unique and must not be empty.
13867 @item @code{address} (default: @code{'()})
13868 An ordered list of IP addresses, network subnets, or network ranges represented
13869 with strings. The query must match one of them. Empty value means that
13870 address match is not required.
13872 @item @code{key} (default: @code{'()})
13873 An ordered list of references to keys represented with strings. The string
13874 must match a key ID defined in a @code{knot-key-configuration}. No key means
13875 that a key is not require to match that ACL.
13877 @item @code{action} (default: @code{'()})
13878 An ordered list of actions that are permitted or forbidden by this ACL. Possible
13879 values are lists of zero or more elements from @code{'transfer}, @code{'notify}
13880 and @code{'update}.
13882 @item @code{deny?} (default: @code{#f})
13883 When true, the ACL defines restrictions. Listed actions are forbidden. When
13884 false, listed actions are allowed.
13889 @deftp {Data Type} zone-entry
13890 Data type represnting a record entry in a zone file.
13891 This type has the following parameters:
13894 @item @code{name} (default: @code{"@@"})
13895 The name of the record. @code{"@@"} refers to the origin of the zone. Names
13896 are relative to the origin of the zone. For example, in the @code{example.org}
13897 zone, @code{"ns.example.org"} actually refers to @code{ns.example.org.example.org}.
13898 Names ending with a dot are absolute, which means that @code{"ns.example.org."}
13899 refers to @code{ns.example.org}.
13901 @item @code{ttl} (default: @code{""})
13902 The Time-To-Live (TTL) of this record. If not set, the default TTL is used.
13904 @item @code{class} (default: @code{"IN"})
13905 The class of the record. Knot currently supports only @code{"IN"} and
13906 partially @code{"CH"}.
13908 @item @code{type} (default: @code{"A"})
13909 The type of the record. Common types include A (IPv4 address), AAAA (IPv6
13910 address), NS (Name Server) and MX (Mail eXchange). Many other types are
13913 @item @code{data} (default: @code{""})
13914 The data contained in the record. For instance an IP address associated with
13915 an A record, or a domain name associated with an NS record. Remember that
13916 domain names are relative to the origin unless they end with a dot.
13921 @deftp {Data Type} zone-file
13922 Data type representing the content of a zone file.
13923 This type has the following parameters:
13926 @item @code{entries} (default: @code{'()})
13927 The list of entries. The SOA record is taken care of, so you don't need to
13928 put it in the list of entries. This list should probably contain an entry
13929 for your primary authoritative DNS server. Other than using a list of entries
13930 directly, you can use @code{define-zone-entries} to define a object containing
13931 the list of entries more easily, that you can later pass to the @code{entries}
13932 field of the @code{zone-file}.
13934 @item @code{origin} (default: @code{""})
13935 The name of your zone. This parameter cannot be empty.
13937 @item @code{ns} (default: @code{"ns"})
13938 The domain of your primary authoritative DNS server. The name is relative to
13939 the origin, unless it ends with a dot. It is mandatory that this primary
13940 DNS server corresponds to an NS record in the zone and that it is associated
13941 to an IP address in the list of entries.
13943 @item @code{mail} (default: @code{"hostmaster"})
13944 An email address people can contact you at, as the owner of the zone. This
13945 is translated as @code{<mail>@@<origin>}.
13947 @item @code{serial} (default: @code{1})
13948 The serial number of the zone. As this is used to keep track of changes by
13949 both slaves and resolvers, it is mandatory that it @emph{never} decreases.
13950 Always increment it when you make a change in your zone.
13952 @item @code{refresh} (default: @code{(* 2 24 3600)})
13953 The frequency at which slaves will do a zone transfer. This value is a number
13954 of seconds. It can be computed by multiplications or with
13955 @code{(string->duration)}.
13957 @item @code{retry} (default: @code{(* 15 60)})
13958 The period after which a slave will retry to contact its master when it fails
13959 to do so a first time.
13961 @item @code{expiry} (default: @code{(* 14 24 3600)})
13962 Default TTL of records. Existing records are considered correct for at most
13963 this amount of time. After this period, resolvers will invalidate their cache
13964 and check again that it still exists.
13966 @item @code{nx} (default: @code{3600})
13967 Default TTL of inexistant records. This delay is usually short because you want
13968 your new domains to reach everyone quickly.
13973 @deftp {Data Type} knot-remote-configuration
13974 Data type representing a remote configuration.
13975 This type has the following parameters:
13978 @item @code{id} (default: @code{""})
13979 An identifier for other configuration fields to refer to this remote. IDs must
13980 be unique and must not be empty.
13982 @item @code{address} (default: @code{'()})
13983 An ordered list of destination IP addresses. Addresses are tried in sequence.
13984 An optional port can be given with the @@ separator. For instance:
13985 @code{(list "1.2.3.4" "2.3.4.5@@53")}. Default port is 53.
13987 @item @code{via} (default: @code{'()})
13988 An ordered list of source IP addresses. An empty list will have Knot choose
13989 an appropriate source IP. An optional port can be given with the @@ separator.
13990 The default is to choose at random.
13992 @item @code{key} (default: @code{#f})
13993 A reference to a key, that is a string containing the identifier of a key
13994 defined in a @code{knot-key-configuration} field.
13999 @deftp {Data Type} knot-keystore-configuration
14000 Data type representing a keystore to hold dnssec keys.
14001 This type has the following parameters:
14004 @item @code{id} (default: @code{""})
14005 The id of the keystore. It must not be empty.
14007 @item @code{backend} (default: @code{'pem})
14008 The backend to store the keys in. Can be @code{'pem} or @code{'pkcs11}.
14010 @item @code{config} (default: @code{"/var/lib/knot/keys/keys"})
14011 The configuration string of the backend. An example for the PKCS#11 is:
14012 @code{"pkcs11:token=knot;pin-value=1234 /gnu/store/.../lib/pkcs11/libsofthsm2.so"}.
14013 For the pem backend, the string reprensents a path in the filesystem.
14018 @deftp {Data Type} knot-policy-configuration
14019 Data type representing a dnssec policy. Knot DNS is able to automatically
14020 sign your zones. It can either generate and manage your keys automatically or
14021 use keys that you generate.
14023 Dnssec is usually implemented using two keys: a Key Signing Key (KSK) that is
14024 used to sign the second, and a Zone Signing Key (ZSK) that is used to sign the
14025 zone. In order to be trusted, the KSK needs to be present in the parent zone
14026 (usually a top-level domain). If your registrar supports dnssec, you will
14027 have to send them your KSK's hash so they can add a DS record in their zone.
14028 This is not automated and need to be done each time you change your KSK.
14030 The policy also defines the lifetime of keys. Usually, ZSK can be changed
14031 easily and use weaker cryptographic functions (they use lower parameters) in
14032 order to sign records quickly, so they are changed often. The KSK however
14033 requires manual interaction with the registrar, so they are changed less often
14034 and use stronger parameters because they sign only one record.
14036 This type has the following parameters:
14039 @item @code{id} (default: @code{""})
14040 The id of the policy. It must not be empty.
14042 @item @code{keystore} (default: @code{"default"})
14043 A reference to a keystore, that is a string containing the identifier of a
14044 keystore defined in a @code{knot-keystore-configuration} field. The
14045 @code{"default"} identifier means the default keystore (a kasp database that
14046 was setup by this service).
14048 @item @code{manual?} (default: @code{#f})
14049 Whether the key management is manual or automatic.
14051 @item @code{single-type-signing?} (default: @code{#f})
14052 When @code{#t}, use the Single-Type Signing Scheme.
14054 @item @code{algorithm} (default: @code{"ecdsap256sha256"})
14055 An algorithm of signing keys and issued signatures.
14057 @item @code{ksk-size} (default: @code{256})
14058 The length of the KSK. Note that this value is correct for the default
14059 algorithm, but would be unsecure for other algorithms.
14061 @item @code{zsk-size} (default: @code{256})
14062 The length of the ZSK. Note that this value is correct for the default
14063 algorithm, but would be unsecure for other algorithms.
14065 @item @code{dnskey-ttl} (default: @code{'default})
14066 The TTL value for DNSKEY records added into zone apex. The special
14067 @code{'default} value means same as the zone SOA TTL.
14069 @item @code{zsk-lifetime} (default: @code{(* 30 24 3600)})
14070 The period between ZSK publication and the next rollover initiation.
14072 @item @code{propagation-delay} (default: @code{(* 24 3600)})
14073 An extra delay added for each key rollover step. This value should be high
14074 enough to cover propagation of data from the master server to all slaves.
14076 @item @code{rrsig-lifetime} (default: @code{(* 14 24 3600)})
14077 A validity period of newly issued signatures.
14079 @item @code{rrsig-refresh} (default: @code{(* 7 24 3600)})
14080 A period how long before a signature expiration the signature will be refreshed.
14082 @item @code{nsec3?} (default: @code{#f})
14083 When @code{#t}, NSEC3 will be used instead of NSEC.
14085 @item @code{nsec3-iterations} (default: @code{5})
14086 The number of additional times the hashing is performed.
14088 @item @code{nsec3-salt-length} (default: @code{8})
14089 The length of a salt field in octets, which is appended to the original owner
14090 name before hashing.
14092 @item @code{nsec3-salt-lifetime} (default: @code{(* 30 24 3600)})
14093 The validity period of newly issued salt field.
14098 @deftp {Data Type} knot-zone-configuration
14099 Data type representing a zone served by Knot.
14100 This type has the following parameters:
14103 @item @code{domain} (default: @code{""})
14104 The domain served by this configuration. It must not be empty.
14106 @item @code{file} (default: @code{""})
14107 The file where this zone is saved. This parameter is ignored by master zones.
14108 Empty means default location that depends on the domain name.
14110 @item @code{zone} (default: @code{(zone-file)})
14111 The content of the zone file. This parameter is ignored by slave zones. It
14112 must contain a zone-file record.
14114 @item @code{master} (default: @code{'()})
14115 A list of master remotes. When empty, this zone is a master. When set, this
14116 zone is a slave. This is a list of remotes identifiers.
14118 @item @code{ddns-master} (default: @code{#f})
14119 The main master. When empty, it defaults to the first master in the list of
14122 @item @code{notify} (default: @code{'()})
14123 A list of slave remote identifiers.
14125 @item @code{acl} (default: @code{'()})
14126 A list of acl identifiers.
14128 @item @code{semantic-checks?} (default: @code{#f})
14129 When set, this adds more semantic checks to the zone.
14131 @item @code{disable-any?} (default: @code{#f})
14132 When set, this forbids queries of the ANY type.
14134 @item @code{zonefile-sync} (default: @code{0})
14135 The delay between a modification in memory and on disk. 0 means immediate
14138 @item @code{serial-policy} (default: @code{'increment})
14139 A policy between @code{'increment} and @code{'unixtime}.
14144 @deftp {Data Type} knot-configuration
14145 Data type representing the Knot configuration.
14146 This type has the following parameters:
14149 @item @code{knot} (default: @code{knot})
14152 @item @code{run-directory} (default: @code{"/var/run/knot"})
14153 The run directory. This directory will be used for pid file and sockets.
14155 @item @code{listen-v4} (default: @code{"0.0.0.0"})
14156 An ip address on which to listen.
14158 @item @code{listen-v6} (default: @code{"::"})
14159 An ip address on which to listen.
14161 @item @code{listen-port} (default: @code{53})
14162 A port on which to listen.
14164 @item @code{keys} (default: @code{'()})
14165 The list of knot-key-configuration used by this configuration.
14167 @item @code{acls} (default: @code{'()})
14168 The list of knot-acl-configuration used by this configuration.
14170 @item @code{remotes} (default: @code{'()})
14171 The list of knot-remote-configuration used by this configuration.
14173 @item @code{zones} (default: @code{'()})
14174 The list of knot-zone-configuration used by this configuration.
14180 @subsubsection VPN Services
14181 @cindex VPN (virtual private network)
14182 @cindex virtual private network (VPN)
14184 The @code{(gnu services vpn)} module provides services related to
14185 @dfn{virtual private networks} (VPNs). It provides a @emph{client} service for
14186 your machine to connect to a VPN, and a @emph{servire} service for your machine
14187 to host a VPN. Both services use @uref{https://openvpn.net/, OpenVPN}.
14189 @deffn {Scheme Procedure} openvpn-client-service @
14190 [#:config (openvpn-client-configuration)]
14192 Return a service that runs @command{openvpn}, a VPN daemon, as a client.
14195 @deffn {Scheme Procedure} openvpn-server-service @
14196 [#:config (openvpn-server-configuration)]
14198 Return a service that runs @command{openvpn}, a VPN daemon, as a server.
14200 Both can be run simultaneously.
14203 @c %automatically generated documentation
14205 Available @code{openvpn-client-configuration} fields are:
14207 @deftypevr {@code{openvpn-client-configuration} parameter} package openvpn
14208 The OpenVPN package.
14212 @deftypevr {@code{openvpn-client-configuration} parameter} string pid-file
14213 The OpenVPN pid file.
14215 Defaults to @samp{"/var/run/openvpn/openvpn.pid"}.
14219 @deftypevr {@code{openvpn-client-configuration} parameter} proto proto
14220 The protocol (UDP or TCP) used to open a channel between clients and
14223 Defaults to @samp{udp}.
14227 @deftypevr {@code{openvpn-client-configuration} parameter} dev dev
14228 The device type used to represent the VPN connection.
14230 Defaults to @samp{tun}.
14234 @deftypevr {@code{openvpn-client-configuration} parameter} string ca
14235 The certificate authority to check connections against.
14237 Defaults to @samp{"/etc/openvpn/ca.crt"}.
14241 @deftypevr {@code{openvpn-client-configuration} parameter} string cert
14242 The certificate of the machine the daemon is running on. It should be
14243 signed by the authority given in @code{ca}.
14245 Defaults to @samp{"/etc/openvpn/client.crt"}.
14249 @deftypevr {@code{openvpn-client-configuration} parameter} string key
14250 The key of the machine the daemon is running on. It must be the key whose
14251 certificate is @code{cert}.
14253 Defaults to @samp{"/etc/openvpn/client.key"}.
14257 @deftypevr {@code{openvpn-client-configuration} parameter} boolean comp-lzo?
14258 Whether to use the lzo compression algorithm.
14260 Defaults to @samp{#t}.
14264 @deftypevr {@code{openvpn-client-configuration} parameter} boolean persist-key?
14265 Don't re-read key files across SIGUSR1 or --ping-restart.
14267 Defaults to @samp{#t}.
14271 @deftypevr {@code{openvpn-client-configuration} parameter} boolean persist-tun?
14272 Don't close and reopen TUN/TAP device or run up/down scripts across
14273 SIGUSR1 or --ping-restart restarts.
14275 Defaults to @samp{#t}.
14279 @deftypevr {@code{openvpn-client-configuration} parameter} number verbosity
14282 Defaults to @samp{3}.
14286 @deftypevr {@code{openvpn-client-configuration} parameter} tls-auth-client tls-auth
14287 Add an additional layer of HMAC authentication on top of the TLS control
14288 channel to protect against DoS attacks.
14290 Defaults to @samp{#f}.
14294 @deftypevr {@code{openvpn-client-configuration} parameter} key-usage verify-key-usage?
14295 Whether to check the server certificate has server usage extension.
14297 Defaults to @samp{#t}.
14301 @deftypevr {@code{openvpn-client-configuration} parameter} bind bind?
14302 Bind to a specific local port number.
14304 Defaults to @samp{#f}.
14308 @deftypevr {@code{openvpn-client-configuration} parameter} resolv-retry resolv-retry?
14309 Retry resolving server address.
14311 Defaults to @samp{#t}.
14315 @deftypevr {@code{openvpn-client-configuration} parameter} openvpn-remote-list remote
14316 A list of remote servers to connect to.
14318 Defaults to @samp{()}.
14320 Available @code{openvpn-remote-configuration} fields are:
14322 @deftypevr {@code{openvpn-remote-configuration} parameter} string name
14325 Defaults to @samp{"my-server"}.
14329 @deftypevr {@code{openvpn-remote-configuration} parameter} number port
14330 Port number the server listens to.
14332 Defaults to @samp{1194}.
14337 @c %end of automatic openvpn-client documentation
14339 @c %automatically generated documentation
14341 Available @code{openvpn-server-configuration} fields are:
14343 @deftypevr {@code{openvpn-server-configuration} parameter} package openvpn
14344 The OpenVPN package.
14348 @deftypevr {@code{openvpn-server-configuration} parameter} string pid-file
14349 The OpenVPN pid file.
14351 Defaults to @samp{"/var/run/openvpn/openvpn.pid"}.
14355 @deftypevr {@code{openvpn-server-configuration} parameter} proto proto
14356 The protocol (UDP or TCP) used to open a channel between clients and
14359 Defaults to @samp{udp}.
14363 @deftypevr {@code{openvpn-server-configuration} parameter} dev dev
14364 The device type used to represent the VPN connection.
14366 Defaults to @samp{tun}.
14370 @deftypevr {@code{openvpn-server-configuration} parameter} string ca
14371 The certificate authority to check connections against.
14373 Defaults to @samp{"/etc/openvpn/ca.crt"}.
14377 @deftypevr {@code{openvpn-server-configuration} parameter} string cert
14378 The certificate of the machine the daemon is running on. It should be
14379 signed by the authority given in @code{ca}.
14381 Defaults to @samp{"/etc/openvpn/client.crt"}.
14385 @deftypevr {@code{openvpn-server-configuration} parameter} string key
14386 The key of the machine the daemon is running on. It must be the key whose
14387 certificate is @code{cert}.
14389 Defaults to @samp{"/etc/openvpn/client.key"}.
14393 @deftypevr {@code{openvpn-server-configuration} parameter} boolean comp-lzo?
14394 Whether to use the lzo compression algorithm.
14396 Defaults to @samp{#t}.
14400 @deftypevr {@code{openvpn-server-configuration} parameter} boolean persist-key?
14401 Don't re-read key files across SIGUSR1 or --ping-restart.
14403 Defaults to @samp{#t}.
14407 @deftypevr {@code{openvpn-server-configuration} parameter} boolean persist-tun?
14408 Don't close and reopen TUN/TAP device or run up/down scripts across
14409 SIGUSR1 or --ping-restart restarts.
14411 Defaults to @samp{#t}.
14415 @deftypevr {@code{openvpn-server-configuration} parameter} number verbosity
14418 Defaults to @samp{3}.
14422 @deftypevr {@code{openvpn-server-configuration} parameter} tls-auth-server tls-auth
14423 Add an additional layer of HMAC authentication on top of the TLS control
14424 channel to protect against DoS attacks.
14426 Defaults to @samp{#f}.
14430 @deftypevr {@code{openvpn-server-configuration} parameter} number port
14431 Specifies the port number on which the server listens.
14433 Defaults to @samp{1194}.
14437 @deftypevr {@code{openvpn-server-configuration} parameter} ip-mask server
14438 An ip and mask specifying the subnet inside the virtual network.
14440 Defaults to @samp{"10.8.0.0 255.255.255.0"}.
14444 @deftypevr {@code{openvpn-server-configuration} parameter} cidr6 server-ipv6
14445 A CIDR notation specifying the IPv6 subnet inside the virtual network.
14447 Defaults to @samp{#f}.
14451 @deftypevr {@code{openvpn-server-configuration} parameter} string dh
14452 The Diffie-Hellman parameters file.
14454 Defaults to @samp{"/etc/openvpn/dh2048.pem"}.
14458 @deftypevr {@code{openvpn-server-configuration} parameter} string ifconfig-pool-persist
14459 The file that records client IPs.
14461 Defaults to @samp{"/etc/openvpn/ipp.txt"}.
14465 @deftypevr {@code{openvpn-server-configuration} parameter} gateway redirect-gateway?
14466 When true, the server will act as a gateway for its clients.
14468 Defaults to @samp{#f}.
14472 @deftypevr {@code{openvpn-server-configuration} parameter} boolean client-to-client?
14473 When true, clients are allowed to talk to each other inside the VPN.
14475 Defaults to @samp{#f}.
14479 @deftypevr {@code{openvpn-server-configuration} parameter} keepalive keepalive
14480 Causes ping-like messages to be sent back and forth over the link so
14481 that each side knows when the other side has gone down. @code{keepalive}
14482 requires a pair. The first element is the period of the ping sending,
14483 and the second element is the timeout before considering the other side
14488 @deftypevr {@code{openvpn-server-configuration} parameter} number max-clients
14489 The maximum number of clients.
14491 Defaults to @samp{100}.
14495 @deftypevr {@code{openvpn-server-configuration} parameter} string status
14496 The status file. This file shows a small report on current connection.
14497 It is truncated and rewritten every minute.
14499 Defaults to @samp{"/var/run/openvpn/status"}.
14503 @deftypevr {@code{openvpn-server-configuration} parameter} openvpn-ccd-list client-config-dir
14504 The list of configuration for some clients.
14506 Defaults to @samp{()}.
14508 Available @code{openvpn-ccd-configuration} fields are:
14510 @deftypevr {@code{openvpn-ccd-configuration} parameter} string name
14513 Defaults to @samp{"client"}.
14517 @deftypevr {@code{openvpn-ccd-configuration} parameter} ip-mask iroute
14520 Defaults to @samp{#f}.
14524 @deftypevr {@code{openvpn-ccd-configuration} parameter} ip-mask ifconfig-push
14527 Defaults to @samp{#f}.
14534 @c %end of automatic openvpn-server documentation
14537 @node Network File System
14538 @subsubsection Network File System
14541 The @code{(gnu services nfs)} module provides the following services,
14542 which are most commonly used in relation to mounting or exporting
14543 directory trees as @dfn{network file systems} (NFS).
14545 @subsubheading RPC Bind Service
14548 The RPC Bind service provides a facility to map program numbers into
14549 universal addresses.
14550 Many NFS related services use this facility. Hence it is automatically
14551 started when a dependent service starts.
14553 @defvr {Scheme Variable} rpcbind-service-type
14554 A service type for the RPC portmapper daemon.
14558 @deftp {Data Type} rpcbind-configuration
14559 Data type representing the configuration of the RPC Bind Service.
14560 This type has the following parameters:
14562 @item @code{rpcbind} (default: @code{rpcbind})
14563 The rpcbind package to use.
14565 @item @code{warm-start?} (default: @code{#t})
14566 If this parameter is @code{#t}, then the daemon will read a
14567 state file on startup thus reloading state information saved by a previous
14573 @subsubheading Pipefs Pseudo File System
14577 The pipefs file system is used to transfer NFS related data
14578 between the kernel and user space programs.
14580 @defvr {Scheme Variable} pipefs-service-type
14581 A service type for the pipefs pseudo file system.
14584 @deftp {Data Type} pipefs-configuration
14585 Data type representing the configuration of the pipefs pseudo file system service.
14586 This type has the following parameters:
14588 @item @code{mount-point} (default: @code{"/var/lib/nfs/rpc_pipefs"})
14589 The directory to which the file system is to be attached.
14594 @subsubheading GSS Daemon Service
14597 @cindex global security system
14599 The @dfn{global security system} (GSS) daemon provides strong security for RPC
14601 Before exchanging RPC requests an RPC client must establish a security
14602 context. Typically this is done using the Kerberos command @command{kinit}
14603 or automatically at login time using PAM services (@pxref{Kerberos Services}).
14605 @defvr {Scheme Variable} gss-service-type
14606 A service type for the Global Security System (GSS) daemon.
14609 @deftp {Data Type} gss-configuration
14610 Data type representing the configuration of the GSS daemon service.
14611 This type has the following parameters:
14613 @item @code{nfs-utils} (default: @code{nfs-utils})
14614 The package in which the @command{rpc.gssd} command is to be found.
14616 @item @code{pipefs-directory} (default: @code{"/var/lib/nfs/rpc_pipefs"})
14617 The directory where the pipefs file system is mounted.
14623 @subsubheading IDMAP Daemon Service
14625 @cindex name mapper
14627 The idmap daemon service provides mapping between user IDs and user names.
14628 Typically it is required in order to access file systems mounted via NFSv4.
14630 @defvr {Scheme Variable} idmap-service-type
14631 A service type for the Identity Mapper (IDMAP) daemon.
14634 @deftp {Data Type} idmap-configuration
14635 Data type representing the configuration of the IDMAP daemon service.
14636 This type has the following parameters:
14638 @item @code{nfs-utils} (default: @code{nfs-utils})
14639 The package in which the @command{rpc.idmapd} command is to be found.
14641 @item @code{pipefs-directory} (default: @code{"/var/lib/nfs/rpc_pipefs"})
14642 The directory where the pipefs file system is mounted.
14644 @item @code{domain} (default: @code{#f})
14645 The local NFSv4 domain name.
14646 This must be a string or @code{#f}.
14647 If it is @code{#f} then the daemon will use the host's fully qualified domain name.
14652 @node Continuous Integration
14653 @subsubsection Continuous Integration
14655 @cindex continuous integration
14656 @uref{https://notabug.org/mthl/cuirass, Cuirass} is a continuous
14657 integration tool for Guix. It can be used both for development and for
14658 providing substitutes to others (@pxref{Substitutes}).
14660 The @code{(gnu services cuirass)} module provides the following service.
14662 @defvr {Scheme Procedure} cuirass-service-type
14663 The type of the Cuirass service. Its value must be a
14664 @code{cuirass-configuration} object, as described below.
14667 To add build jobs, you have to set the @code{specifications} field of
14668 the configuration. Here is an example of a service defining a build job
14669 based on a specification that can be found in Cuirass source tree. This
14670 service polls the Guix repository and builds a subset of the Guix
14671 packages, as prescribed in the @file{gnu-system.scm} example spec:
14674 (let ((spec #~((#:name . "guix")
14675 (#:url . "git://git.savannah.gnu.org/guix.git")
14676 (#:load-path . ".")
14678 ;; Here we must provide an absolute file name.
14679 ;; We take jobs from one of the examples provided
14681 (#:file . #$(file-append
14683 "/tests/gnu-system.scm"))
14685 (#:proc . hydra-jobs)
14686 (#:arguments (subset . "hello"))
14687 (#:branch . "master"))))
14688 (service cuirass-service-type
14689 (cuirass-configuration
14690 (specifications #~(list #$spec)))))
14693 While information related to build jobs is located directly in the
14694 specifications, global settings for the @command{cuirass} process are
14695 accessible in other @code{cuirass-configuration} fields.
14697 @deftp {Data Type} cuirass-configuration
14698 Data type representing the configuration of Cuirass.
14701 @item @code{log-file} (default: @code{"/var/log/cuirass.log"})
14702 Location of the log file.
14704 @item @code{cache-directory} (default: @code{"/var/cache/cuirass"})
14705 Location of the repository cache.
14707 @item @code{user} (default: @code{"cuirass"})
14708 Owner of the @code{cuirass} process.
14710 @item @code{group} (default: @code{"cuirass"})
14711 Owner's group of the @code{cuirass} process.
14713 @item @code{interval} (default: @code{60})
14714 Number of seconds between the poll of the repositories followed by the
14717 @item @code{database} (default: @code{"/var/run/cuirass/cuirass.db"})
14718 Location of sqlite database which contains the build results and previously
14719 added specifications.
14721 @item @code{port} (default: @code{8080})
14722 Port number used by the HTTP server.
14724 @item @code{specifications} (default: @code{#~'()})
14725 A gexp (@pxref{G-Expressions}) that evaluates to a list of specifications,
14726 where a specification is an association list
14727 (@pxref{Associations Lists,,, guile, GNU Guile Reference Manual}) whose
14728 keys are keywords (@code{#:keyword-example}) as shown in the example
14731 @item @code{use-substitutes?} (default: @code{#f})
14732 This allows using substitutes to avoid building every dependencies of a job
14735 @item @code{one-shot?} (default: @code{#f})
14736 Only evaluate specifications and build derivations once.
14738 @item @code{load-path} (default: @code{'()})
14739 This allows users to define their own packages and make them visible to
14740 cuirass as in @command{guix build} command.
14742 @item @code{cuirass} (default: @code{cuirass})
14743 The Cuirass package to use.
14747 @node Power management Services
14748 @subsubsection Power management Services
14750 @cindex power management with TLP
14751 The @code{(gnu services pm)} module provides a Guix service definition
14752 for the Linux power management tool TLP.
14754 TLP enables various powersaving modes in userspace and kernel.
14755 Contrary to @code{upower-service}, it is not a passive,
14756 monitoring tool, as it will apply custom settings each time a new power
14757 source is detected. More information can be found at
14758 @uref{http://linrunner.de/en/tlp/tlp.html, TLP home page}.
14760 @deffn {Scheme Variable} tlp-service-type
14761 The service type for the TLP tool. Its value should be a valid
14762 TLP configuration (see below). To use the default settings, simply
14765 (service tlp-service-type)
14769 By default TLP does not need much configuration but most TLP parameters
14770 can be tweaked using @code{tlp-configuration}.
14772 Each parameter definition is preceded by its type; for example,
14773 @samp{boolean foo} indicates that the @code{foo} parameter
14774 should be specified as a boolean. Types starting with
14775 @code{maybe-} denote parameters that won't show up in TLP config file
14776 when their value is @code{'disabled}.
14778 @c The following documentation was initially generated by
14779 @c (generate-tlp-documentation) in (gnu services pm). Manually maintained
14780 @c documentation is better, so we shouldn't hesitate to edit below as
14781 @c needed. However if the change you want to make to this documentation
14782 @c can be done in an automated way, it's probably easier to change
14783 @c (generate-documentation) than to make it below and have to deal with
14784 @c the churn as TLP updates.
14786 Available @code{tlp-configuration} fields are:
14788 @deftypevr {@code{tlp-configuration} parameter} package tlp
14793 @deftypevr {@code{tlp-configuration} parameter} boolean tlp-enable?
14794 Set to true if you wish to enable TLP.
14796 Defaults to @samp{#t}.
14800 @deftypevr {@code{tlp-configuration} parameter} string tlp-default-mode
14801 Default mode when no power supply can be detected. Alternatives are AC
14804 Defaults to @samp{"AC"}.
14808 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer disk-idle-secs-on-ac
14809 Number of seconds Linux kernel has to wait after the disk goes idle,
14810 before syncing on AC.
14812 Defaults to @samp{0}.
14816 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer disk-idle-secs-on-bat
14817 Same as @code{disk-idle-ac} but on BAT mode.
14819 Defaults to @samp{2}.
14823 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer max-lost-work-secs-on-ac
14824 Dirty pages flushing periodicity, expressed in seconds.
14826 Defaults to @samp{15}.
14830 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer max-lost-work-secs-on-bat
14831 Same as @code{max-lost-work-secs-on-ac} but on BAT mode.
14833 Defaults to @samp{60}.
14837 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list cpu-scaling-governor-on-ac
14838 CPU frequency scaling governor on AC mode. With intel_pstate driver,
14839 alternatives are powersave and performance. With acpi-cpufreq driver,
14840 alternatives are ondemand, powersave, performance and conservative.
14842 Defaults to @samp{disabled}.
14846 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list cpu-scaling-governor-on-bat
14847 Same as @code{cpu-scaling-governor-on-ac} but on BAT mode.
14849 Defaults to @samp{disabled}.
14853 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-scaling-min-freq-on-ac
14854 Set the min available frequency for the scaling governor on AC.
14856 Defaults to @samp{disabled}.
14860 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-scaling-max-freq-on-ac
14861 Set the max available frequency for the scaling governor on AC.
14863 Defaults to @samp{disabled}.
14867 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-scaling-min-freq-on-bat
14868 Set the min available frequency for the scaling governor on BAT.
14870 Defaults to @samp{disabled}.
14874 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-scaling-max-freq-on-bat
14875 Set the max available frequency for the scaling governor on BAT.
14877 Defaults to @samp{disabled}.
14881 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-min-perf-on-ac
14882 Limit the min P-state to control the power dissipation of the CPU, in AC
14883 mode. Values are stated as a percentage of the available performance.
14885 Defaults to @samp{disabled}.
14889 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-max-perf-on-ac
14890 Limit the max P-state to control the power dissipation of the CPU, in AC
14891 mode. Values are stated as a percentage of the available performance.
14893 Defaults to @samp{disabled}.
14897 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-min-perf-on-bat
14898 Same as @code{cpu-min-perf-on-ac} on BAT mode.
14900 Defaults to @samp{disabled}.
14904 @deftypevr {@code{tlp-configuration} parameter} maybe-non-negative-integer cpu-max-perf-on-bat
14905 Same as @code{cpu-max-perf-on-ac} on BAT mode.
14907 Defaults to @samp{disabled}.
14911 @deftypevr {@code{tlp-configuration} parameter} maybe-boolean cpu-boost-on-ac?
14912 Enable CPU turbo boost feature on AC mode.
14914 Defaults to @samp{disabled}.
14918 @deftypevr {@code{tlp-configuration} parameter} maybe-boolean cpu-boost-on-bat?
14919 Same as @code{cpu-boost-on-ac?} on BAT mode.
14921 Defaults to @samp{disabled}.
14925 @deftypevr {@code{tlp-configuration} parameter} boolean sched-powersave-on-ac?
14926 Allow Linux kernel to minimize the number of CPU cores/hyper-threads
14927 used under light load conditions.
14929 Defaults to @samp{#f}.
14933 @deftypevr {@code{tlp-configuration} parameter} boolean sched-powersave-on-bat?
14934 Same as @code{sched-powersave-on-ac?} but on BAT mode.
14936 Defaults to @samp{#t}.
14940 @deftypevr {@code{tlp-configuration} parameter} boolean nmi-watchdog?
14941 Enable Linux kernel NMI watchdog.
14943 Defaults to @samp{#f}.
14947 @deftypevr {@code{tlp-configuration} parameter} maybe-string phc-controls
14948 For Linux kernels with PHC patch applied, change CPU voltages. An
14949 example value would be @samp{"F:V F:V F:V F:V"}.
14951 Defaults to @samp{disabled}.
14955 @deftypevr {@code{tlp-configuration} parameter} string energy-perf-policy-on-ac
14956 Set CPU performance versus energy saving policy on AC. Alternatives are
14957 performance, normal, powersave.
14959 Defaults to @samp{"performance"}.
14963 @deftypevr {@code{tlp-configuration} parameter} string energy-perf-policy-on-bat
14964 Same as @code{energy-perf-policy-ac} but on BAT mode.
14966 Defaults to @samp{"powersave"}.
14970 @deftypevr {@code{tlp-configuration} parameter} space-separated-string-list disks-devices
14975 @deftypevr {@code{tlp-configuration} parameter} space-separated-string-list disk-apm-level-on-ac
14976 Hard disk advanced power management level.
14980 @deftypevr {@code{tlp-configuration} parameter} space-separated-string-list disk-apm-level-on-bat
14981 Same as @code{disk-apm-bat} but on BAT mode.
14985 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list disk-spindown-timeout-on-ac
14986 Hard disk spin down timeout. One value has to be specified for each
14987 declared hard disk.
14989 Defaults to @samp{disabled}.
14993 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list disk-spindown-timeout-on-bat
14994 Same as @code{disk-spindown-timeout-on-ac} but on BAT mode.
14996 Defaults to @samp{disabled}.
15000 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list disk-iosched
15001 Select IO scheduler for disk devices. One value has to be specified for
15002 each declared hard disk. Example alternatives are cfq, deadline and
15005 Defaults to @samp{disabled}.
15009 @deftypevr {@code{tlp-configuration} parameter} string sata-linkpwr-on-ac
15010 SATA aggressive link power management (ALPM) level. Alternatives are
15011 min_power, medium_power, max_performance.
15013 Defaults to @samp{"max_performance"}.
15017 @deftypevr {@code{tlp-configuration} parameter} string sata-linkpwr-on-bat
15018 Same as @code{sata-linkpwr-ac} but on BAT mode.
15020 Defaults to @samp{"min_power"}.
15024 @deftypevr {@code{tlp-configuration} parameter} maybe-string sata-linkpwr-blacklist
15025 Exclude specified SATA host devices for link power management.
15027 Defaults to @samp{disabled}.
15031 @deftypevr {@code{tlp-configuration} parameter} maybe-on-off-boolean ahci-runtime-pm-on-ac?
15032 Enable Runtime Power Management for AHCI controller and disks on AC
15035 Defaults to @samp{disabled}.
15039 @deftypevr {@code{tlp-configuration} parameter} maybe-on-off-boolean ahci-runtime-pm-on-bat?
15040 Same as @code{ahci-runtime-pm-on-ac} on BAT mode.
15042 Defaults to @samp{disabled}.
15046 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer ahci-runtime-pm-timeout
15047 Seconds of inactivity before disk is suspended.
15049 Defaults to @samp{15}.
15053 @deftypevr {@code{tlp-configuration} parameter} string pcie-aspm-on-ac
15054 PCI Express Active State Power Management level. Alternatives are
15055 default, performance, powersave.
15057 Defaults to @samp{"performance"}.
15061 @deftypevr {@code{tlp-configuration} parameter} string pcie-aspm-on-bat
15062 Same as @code{pcie-aspm-ac} but on BAT mode.
15064 Defaults to @samp{"powersave"}.
15068 @deftypevr {@code{tlp-configuration} parameter} string radeon-power-profile-on-ac
15069 Radeon graphics clock speed level. Alternatives are low, mid, high,
15072 Defaults to @samp{"high"}.
15076 @deftypevr {@code{tlp-configuration} parameter} string radeon-power-profile-on-bat
15077 Same as @code{radeon-power-ac} but on BAT mode.
15079 Defaults to @samp{"low"}.
15083 @deftypevr {@code{tlp-configuration} parameter} string radeon-dpm-state-on-ac
15084 Radeon dynamic power management method (DPM). Alternatives are battery,
15087 Defaults to @samp{"performance"}.
15091 @deftypevr {@code{tlp-configuration} parameter} string radeon-dpm-state-on-bat
15092 Same as @code{radeon-dpm-state-ac} but on BAT mode.
15094 Defaults to @samp{"battery"}.
15098 @deftypevr {@code{tlp-configuration} parameter} string radeon-dpm-perf-level-on-ac
15099 Radeon DPM performance level. Alternatives are auto, low, high.
15101 Defaults to @samp{"auto"}.
15105 @deftypevr {@code{tlp-configuration} parameter} string radeon-dpm-perf-level-on-bat
15106 Same as @code{radeon-dpm-perf-ac} but on BAT mode.
15108 Defaults to @samp{"auto"}.
15112 @deftypevr {@code{tlp-configuration} parameter} on-off-boolean wifi-pwr-on-ac?
15113 Wifi power saving mode.
15115 Defaults to @samp{#f}.
15119 @deftypevr {@code{tlp-configuration} parameter} on-off-boolean wifi-pwr-on-bat?
15120 Same as @code{wifi-power-ac?} but on BAT mode.
15122 Defaults to @samp{#t}.
15126 @deftypevr {@code{tlp-configuration} parameter} y-n-boolean wol-disable?
15127 Disable wake on LAN.
15129 Defaults to @samp{#t}.
15133 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer sound-power-save-on-ac
15134 Timeout duration in seconds before activating audio power saving on
15135 Intel HDA and AC97 devices. A value of 0 disables power saving.
15137 Defaults to @samp{0}.
15141 @deftypevr {@code{tlp-configuration} parameter} non-negative-integer sound-power-save-on-bat
15142 Same as @code{sound-powersave-ac} but on BAT mode.
15144 Defaults to @samp{1}.
15148 @deftypevr {@code{tlp-configuration} parameter} y-n-boolean sound-power-save-controller?
15149 Disable controller in powersaving mode on Intel HDA devices.
15151 Defaults to @samp{#t}.
15155 @deftypevr {@code{tlp-configuration} parameter} boolean bay-poweroff-on-bat?
15156 Enable optical drive in UltraBay/MediaBay on BAT mode. Drive can be
15157 powered on again by releasing (and reinserting) the eject lever or by
15158 pressing the disc eject button on newer models.
15160 Defaults to @samp{#f}.
15164 @deftypevr {@code{tlp-configuration} parameter} string bay-device
15165 Name of the optical drive device to power off.
15167 Defaults to @samp{"sr0"}.
15171 @deftypevr {@code{tlp-configuration} parameter} string runtime-pm-on-ac
15172 Runtime Power Management for PCI(e) bus devices. Alternatives are on
15175 Defaults to @samp{"on"}.
15179 @deftypevr {@code{tlp-configuration} parameter} string runtime-pm-on-bat
15180 Same as @code{runtime-pm-ac} but on BAT mode.
15182 Defaults to @samp{"auto"}.
15186 @deftypevr {@code{tlp-configuration} parameter} boolean runtime-pm-all?
15187 Runtime Power Management for all PCI(e) bus devices, except blacklisted
15190 Defaults to @samp{#t}.
15194 @deftypevr {@code{tlp-configuration} parameter} maybe-space-separated-string-list runtime-pm-blacklist
15195 Exclude specified PCI(e) device addresses from Runtime Power Management.
15197 Defaults to @samp{disabled}.
15201 @deftypevr {@code{tlp-configuration} parameter} space-separated-string-list runtime-pm-driver-blacklist
15202 Exclude PCI(e) devices assigned to the specified drivers from Runtime
15207 @deftypevr {@code{tlp-configuration} parameter} boolean usb-autosuspend?
15208 Enable USB autosuspend feature.
15210 Defaults to @samp{#t}.
15214 @deftypevr {@code{tlp-configuration} parameter} maybe-string usb-blacklist
15215 Exclude specified devices from USB autosuspend.
15217 Defaults to @samp{disabled}.
15221 @deftypevr {@code{tlp-configuration} parameter} boolean usb-blacklist-wwan?
15222 Exclude WWAN devices from USB autosuspend.
15224 Defaults to @samp{#t}.
15228 @deftypevr {@code{tlp-configuration} parameter} maybe-string usb-whitelist
15229 Include specified devices into USB autosuspend, even if they are already
15230 excluded by the driver or via @code{usb-blacklist-wwan?}.
15232 Defaults to @samp{disabled}.
15236 @deftypevr {@code{tlp-configuration} parameter} maybe-boolean usb-autosuspend-disable-on-shutdown?
15237 Enable USB autosuspend before shutdown.
15239 Defaults to @samp{disabled}.
15243 @deftypevr {@code{tlp-configuration} parameter} boolean restore-device-state-on-startup?
15244 Restore radio device state (bluetooth, wifi, wwan) from previous
15245 shutdown on system startup.
15247 Defaults to @samp{#f}.
15252 The @code{(gnu services pm)} module provides an interface to
15253 thermald, a CPU frequency scaling service which helps prevent overheating.
15255 @defvr {Scheme Variable} thermald-service-type
15256 This is the service type for
15257 @uref{https://01.org/linux-thermal-daemon/, thermald}, the Linux
15258 Thermal Daemon, which is responsible for controlling the thermal state
15259 of processors and preventing overheating.
15262 @deftp {Data Type} thermald-configuration
15263 Data type representing the configuration of @code{thermald-service-type}.
15266 @item @code{ignore-cpuid-check?} (default: @code{#f})
15267 Ignore cpuid check for supported CPU models.
15269 @item @code{thermald} (default: @var{thermald})
15270 Package object of thermald.
15276 @node Miscellaneous Services
15277 @subsubsection Miscellaneous Services
15281 @subsubheading Lirc Service
15283 The @code{(gnu services lirc)} module provides the following service.
15285 @deffn {Scheme Procedure} lirc-service [#:lirc lirc] @
15286 [#:device #f] [#:driver #f] [#:config-file #f] @
15287 [#:extra-options '()]
15288 Return a service that runs @url{http://www.lirc.org,LIRC}, a daemon that
15289 decodes infrared signals from remote controls.
15291 Optionally, @var{device}, @var{driver} and @var{config-file}
15292 (configuration file name) may be specified. See @command{lircd} manual
15295 Finally, @var{extra-options} is a list of additional command-line options
15296 passed to @command{lircd}.
15300 @subsubheading Spice Service
15302 The @code{(gnu services spice)} module provides the following service.
15304 @deffn {Scheme Procedure} spice-vdagent-service [#:spice-vdagent]
15305 Returns a service that runs @url{http://www.spice-space.org,VDAGENT}, a daemon
15306 that enables sharing the clipboard with a vm and setting the guest display
15307 resolution when the graphical console window resizes.
15310 @subsubsection Dictionary Services
15312 The @code{(gnu services dict)} module provides the following service:
15314 @deffn {Scheme Procedure} dicod-service [#:config (dicod-configuration)]
15315 Return a service that runs the @command{dicod} daemon, an implementation
15316 of DICT server (@pxref{Dicod,,, dico, GNU Dico Manual}).
15318 The optional @var{config} argument specifies the configuration for
15319 @command{dicod}, which should be a @code{<dicod-configuration>} object, by
15320 default it serves the GNU Collaborative International Dictonary of English.
15322 You can add @command{open localhost} to your @file{~/.dico} file to make
15323 @code{localhost} the default server for @command{dico} client
15324 (@pxref{Initialization File,,, dico, GNU Dico Manual}).
15327 @deftp {Data Type} dicod-configuration
15328 Data type representing the configuration of dicod.
15331 @item @code{dico} (default: @var{dico})
15332 Package object of the GNU Dico dictionary server.
15334 @item @code{interfaces} (default: @var{'("localhost")})
15335 This is the list of IP addresses and ports and possibly socket file
15336 names to listen to (@pxref{Server Settings, @code{listen} directive,,
15337 dico, GNU Dico Manual}).
15339 @item @code{handlers} (default: @var{'()})
15340 List of @code{<dicod-handler>} objects denoting handlers (module instances).
15342 @item @code{databases} (default: @var{(list %dicod-database:gcide)})
15343 List of @code{<dicod-database>} objects denoting dictionaries to be served.
15347 @deftp {Data Type} dicod-handler
15348 Data type representing a dictionary handler (module instance).
15352 Name of the handler (module instance).
15354 @item @code{module} (default: @var{#f})
15355 Name of the dicod module of the handler (instance). If it is @code{#f},
15356 the module has the same name as the handler.
15357 (@pxref{Modules,,, dico, GNU Dico Manual}).
15359 @item @code{options}
15360 List of strings or gexps representing the arguments for the module handler
15364 @deftp {Data Type} dicod-database
15365 Data type representing a dictionary database.
15369 Name of the database, will be used in DICT commands.
15371 @item @code{handler}
15372 Name of the dicod handler (module instance) used by this database
15373 (@pxref{Handlers,,, dico, GNU Dico Manual}).
15375 @item @code{complex?} (default: @var{#f})
15376 Whether the database configuration complex. The complex configuration
15377 will need a corresponding @code{<dicod-handler>} object, otherwise not.
15379 @item @code{options}
15380 List of strings or gexps representing the arguments for the database
15381 (@pxref{Databases,,, dico, GNU Dico Manual}).
15385 @defvr {Scheme Variable} %dicod-database:gcide
15386 A @code{<dicod-database>} object serving the GNU Collaborative International
15387 Dictionary of English using the @code{gcide} package.
15390 The following is an example @code{dicod-service} configuration.
15393 (dicod-service #:config
15394 (dicod-configuration
15395 (handlers (list (dicod-handler
15399 (list #~(string-append "dbdir=" #$wordnet))))))
15400 (databases (list (dicod-database
15403 (handler "wordnet")
15404 (options '("database=wn")))
15405 %dicod-database:gcide))))
15408 @subsubsection Version Control
15410 The @code{(gnu services version-control)} module provides the following services:
15412 @subsubheading Git daemon service
15414 @deffn {Scheme Procedure} git-daemon-service [#:config (git-daemon-configuration)]
15416 Return a service that runs @command{git daemon}, a simple TCP server to
15417 expose repositories over the Git protocol for anonymous access.
15419 The optional @var{config} argument should be a
15420 @code{<git-daemon-configuration>} object, by default it allows read-only
15421 access to exported@footnote{By creating the magic file
15422 "git-daemon-export-ok" in the repository directory.} repositories under
15427 @deftp {Data Type} git-daemon-configuration
15428 Data type representing the configuration for @code{git-daemon-service}.
15431 @item @code{package} (default: @var{git})
15432 Package object of the Git distributed version control system.
15434 @item @code{export-all?} (default: @var{#f})
15435 Whether to allow access for all Git repositories, even if they do not
15436 have the @file{git-daemon-export-ok} file.
15438 @item @code{base-path} (default: @file{/srv/git})
15439 Whether to remap all the path requests as relative to the given path.
15440 If you run git daemon with @var{(base-path "/srv/git")} on example.com,
15441 then if you later try to pull @code{git://example.com/hello.git}, git
15442 daemon will interpret the path as @code{/srv/git/hello.git}.
15444 @item @code{user-path} (default: @var{#f})
15445 Whether to allow @code{~user} notation to be used in requests. When
15446 specified with empty string, requests to @code{git://host/~alice/foo} is
15447 taken as a request to access @code{foo} repository in the home directory
15448 of user @code{alice}. If @var{(user-path "path")} is specified, the
15449 same request is taken as a request to access @code{path/foo} repository
15450 in the home directory of user @code{alice}.
15452 @item @code{listen} (default: @var{'()})
15453 Whether to listen on specific IP addresses or hostnames, defaults to
15456 @item @code{port} (default: @var{#f})
15457 Whether to listen on an alternative port, which defaults to 9418.
15459 @item @code{whitelist} (default: @var{'()})
15460 If not empty, only allow access to this list of directories.
15462 @item @code{extra-options} (default: @var{'()})
15463 Extra options will be passed to @code{git daemon}, please run
15464 @command{man git-daemon} for more information.
15469 @node Setuid Programs
15470 @subsection Setuid Programs
15472 @cindex setuid programs
15473 Some programs need to run with ``root'' privileges, even when they are
15474 launched by unprivileged users. A notorious example is the
15475 @command{passwd} program, which users can run to change their
15476 password, and which needs to access the @file{/etc/passwd} and
15477 @file{/etc/shadow} files---something normally restricted to root, for
15478 obvious security reasons. To address that, these executables are
15479 @dfn{setuid-root}, meaning that they always run with root privileges
15480 (@pxref{How Change Persona,,, libc, The GNU C Library Reference Manual},
15481 for more info about the setuid mechanism.)
15483 The store itself @emph{cannot} contain setuid programs: that would be a
15484 security issue since any user on the system can write derivations that
15485 populate the store (@pxref{The Store}). Thus, a different mechanism is
15486 used: instead of changing the setuid bit directly on files that are in
15487 the store, we let the system administrator @emph{declare} which programs
15488 should be setuid root.
15490 The @code{setuid-programs} field of an @code{operating-system}
15491 declaration contains a list of G-expressions denoting the names of
15492 programs to be setuid-root (@pxref{Using the Configuration System}).
15493 For instance, the @command{passwd} program, which is part of the Shadow
15494 package, can be designated by this G-expression (@pxref{G-Expressions}):
15497 #~(string-append #$shadow "/bin/passwd")
15500 A default set of setuid programs is defined by the
15501 @code{%setuid-programs} variable of the @code{(gnu system)} module.
15503 @defvr {Scheme Variable} %setuid-programs
15504 A list of G-expressions denoting common programs that are setuid-root.
15506 The list includes commands such as @command{passwd}, @command{ping},
15507 @command{su}, and @command{sudo}.
15510 Under the hood, the actual setuid programs are created in the
15511 @file{/run/setuid-programs} directory at system activation time. The
15512 files in this directory refer to the ``real'' binaries, which are in the
15515 @node X.509 Certificates
15516 @subsection X.509 Certificates
15518 @cindex HTTPS, certificates
15519 @cindex X.509 certificates
15521 Web servers available over HTTPS (that is, HTTP over the transport-layer
15522 security mechanism, TLS) send client programs an @dfn{X.509 certificate}
15523 that the client can then use to @emph{authenticate} the server. To do
15524 that, clients verify that the server's certificate is signed by a
15525 so-called @dfn{certificate authority} (CA). But to verify the CA's
15526 signature, clients must have first acquired the CA's certificate.
15528 Web browsers such as GNU@tie{}IceCat include their own set of CA
15529 certificates, such that they are able to verify CA signatures
15532 However, most other programs that can talk HTTPS---@command{wget},
15533 @command{git}, @command{w3m}, etc.---need to be told where CA
15534 certificates can be found.
15536 @cindex @code{nss-certs}
15537 In GuixSD, this is done by adding a package that provides certificates
15538 to the @code{packages} field of the @code{operating-system} declaration
15539 (@pxref{operating-system Reference}). GuixSD includes one such package,
15540 @code{nss-certs}, which is a set of CA certificates provided as part of
15541 Mozilla's Network Security Services.
15543 Note that it is @emph{not} part of @var{%base-packages}, so you need to
15544 explicitly add it. The @file{/etc/ssl/certs} directory, which is where
15545 most applications and libraries look for certificates by default, points
15546 to the certificates installed globally.
15548 Unprivileged users, including users of Guix on a foreign distro,
15549 can also install their own certificate package in
15550 their profile. A number of environment variables need to be defined so
15551 that applications and libraries know where to find them. Namely, the
15552 OpenSSL library honors the @code{SSL_CERT_DIR} and @code{SSL_CERT_FILE}
15553 variables. Some applications add their own environment variables; for
15554 instance, the Git version control system honors the certificate bundle
15555 pointed to by the @code{GIT_SSL_CAINFO} environment variable. Thus, you
15556 would typically run something like:
15559 $ guix package -i nss-certs
15560 $ export SSL_CERT_DIR="$HOME/.guix-profile/etc/ssl/certs"
15561 $ export SSL_CERT_FILE="$HOME/.guix-profile/etc/ssl/certs/ca-certificates.crt"
15562 $ export GIT_SSL_CAINFO="$SSL_CERT_FILE"
15565 As another example, R requires the @code{CURL_CA_BUNDLE} environment
15566 variable to point to a certificate bundle, so you would have to run
15567 something like this:
15570 $ guix package -i nss-certs
15571 $ export CURL_CA_BUNDLE="$HOME/.guix-profile/etc/ssl/certs/ca-certificates.crt"
15574 For other applications you may want to look up the required environment
15575 variable in the relevant documentation.
15578 @node Name Service Switch
15579 @subsection Name Service Switch
15581 @cindex name service switch
15583 The @code{(gnu system nss)} module provides bindings to the
15584 configuration file of the libc @dfn{name service switch} or @dfn{NSS}
15585 (@pxref{NSS Configuration File,,, libc, The GNU C Library Reference
15586 Manual}). In a nutshell, the NSS is a mechanism that allows libc to be
15587 extended with new ``name'' lookup methods for system databases, which
15588 includes host names, service names, user accounts, and more (@pxref{Name
15589 Service Switch, System Databases and Name Service Switch,, libc, The GNU
15590 C Library Reference Manual}).
15592 The NSS configuration specifies, for each system database, which lookup
15593 method is to be used, and how the various methods are chained
15594 together---for instance, under which circumstances NSS should try the
15595 next method in the list. The NSS configuration is given in the
15596 @code{name-service-switch} field of @code{operating-system} declarations
15597 (@pxref{operating-system Reference, @code{name-service-switch}}).
15600 @cindex .local, host name lookup
15601 As an example, the declaration below configures the NSS to use the
15602 @uref{http://0pointer.de/lennart/projects/nss-mdns/, @code{nss-mdns}
15603 back-end}, which supports host name lookups over multicast DNS (mDNS)
15604 for host names ending in @code{.local}:
15607 (name-service-switch
15608 (hosts (list %files ;first, check /etc/hosts
15610 ;; If the above did not succeed, try
15611 ;; with 'mdns_minimal'.
15613 (name "mdns_minimal")
15615 ;; 'mdns_minimal' is authoritative for
15616 ;; '.local'. When it returns "not found",
15617 ;; no need to try the next methods.
15618 (reaction (lookup-specification
15619 (not-found => return))))
15621 ;; Then fall back to DNS.
15625 ;; Finally, try with the "full" 'mdns'.
15630 Do not worry: the @code{%mdns-host-lookup-nss} variable (see below)
15631 contains this configuration, so you will not have to type it if all you
15632 want is to have @code{.local} host lookup working.
15634 Note that, in this case, in addition to setting the
15635 @code{name-service-switch} of the @code{operating-system} declaration,
15636 you also need to use @code{avahi-service} (@pxref{Networking Services,
15637 @code{avahi-service}}), or @var{%desktop-services}, which includes it
15638 (@pxref{Desktop Services}). Doing this makes @code{nss-mdns} accessible
15639 to the name service cache daemon (@pxref{Base Services,
15640 @code{nscd-service}}).
15642 For convenience, the following variables provide typical NSS
15645 @defvr {Scheme Variable} %default-nss
15646 This is the default name service switch configuration, a
15647 @code{name-service-switch} object.
15650 @defvr {Scheme Variable} %mdns-host-lookup-nss
15651 This is the name service switch configuration with support for host name
15652 lookup over multicast DNS (mDNS) for host names ending in @code{.local}.
15655 The reference for name service switch configuration is given below. It
15656 is a direct mapping of the configuration file format of the C library , so
15657 please refer to the C library manual for more information (@pxref{NSS
15658 Configuration File,,, libc, The GNU C Library Reference Manual}).
15659 Compared to the configuration file format of libc NSS, it has the advantage
15660 not only of adding this warm parenthetic feel that we like, but also
15661 static checks: you will know about syntax errors and typos as soon as you
15662 run @command{guix system}.
15664 @deftp {Data Type} name-service-switch
15666 This is the data type representation the configuration of libc's name
15667 service switch (NSS). Each field below represents one of the supported
15684 The system databases handled by the NSS. Each of these fields must be a
15685 list of @code{<name-service>} objects (see below).
15689 @deftp {Data Type} name-service
15691 This is the data type representing an actual name service and the
15692 associated lookup action.
15696 A string denoting the name service (@pxref{Services in the NSS
15697 configuration,,, libc, The GNU C Library Reference Manual}).
15699 Note that name services listed here must be visible to nscd. This is
15700 achieved by passing the @code{#:name-services} argument to
15701 @code{nscd-service} the list of packages providing the needed name
15702 services (@pxref{Base Services, @code{nscd-service}}).
15705 An action specified using the @code{lookup-specification} macro
15706 (@pxref{Actions in the NSS configuration,,, libc, The GNU C Library
15707 Reference Manual}). For example:
15710 (lookup-specification (unavailable => continue)
15711 (success => return))
15716 @node Initial RAM Disk
15717 @subsection Initial RAM Disk
15720 @cindex initial RAM disk
15721 For bootstrapping purposes, the Linux-Libre kernel is passed an
15722 @dfn{initial RAM disk}, or @dfn{initrd}. An initrd contains a temporary
15723 root file system as well as an initialization script. The latter is
15724 responsible for mounting the real root file system, and for loading any
15725 kernel modules that may be needed to achieve that.
15727 The @code{initrd} field of an @code{operating-system} declaration allows
15728 you to specify which initrd you would like to use. The @code{(gnu
15729 system linux-initrd)} module provides three ways to build an initrd: the
15730 high-level @code{base-initrd} procedure and the low-level
15731 @code{raw-initrd} and @code{expression->initrd} procedures.
15733 The @code{base-initrd} procedure is intended to cover most common uses.
15734 For example, if you want to add a bunch of kernel modules to be loaded
15735 at boot time, you can define the @code{initrd} field of the operating
15736 system declaration like this:
15739 (initrd (lambda (file-systems . rest)
15740 ;; Create a standard initrd that has modules "foo.ko"
15741 ;; and "bar.ko", as well as their dependencies, in
15742 ;; addition to the modules available by default.
15743 (apply base-initrd file-systems
15744 #:extra-modules '("foo" "bar")
15748 The @code{base-initrd} procedure also handles common use cases that
15749 involves using the system as a QEMU guest, or as a ``live'' system with
15750 volatile root file system.
15752 The @code{base-initrd} procedure is built from @code{raw-initrd} procedure.
15753 Unlike @code{base-initrd}, @code{raw-initrd} doesn't do anything high-level,
15754 such as trying to guess which kernel modules and packages should be included
15755 to the initrd. An example use of @code{raw-initrd} is when a user has
15756 a custom Linux kernel configuration and default kernel modules included by
15757 @code{base-initrd} are not available.
15759 The initial RAM disk produced by @code{base-initrd} or @code{raw-initrd}
15760 honors several options passed on the Linux kernel command line
15761 (that is, arguments passed @i{via} the @code{linux} command of GRUB, or the
15762 @code{-append} option of QEMU), notably:
15765 @item --load=@var{boot}
15766 Tell the initial RAM disk to load @var{boot}, a file containing a Scheme
15767 program, once it has mounted the root file system.
15769 GuixSD uses this option to yield control to a boot program that runs the
15770 service activation programs and then spawns the GNU@tie{}Shepherd, the
15771 initialization system.
15773 @item --root=@var{root}
15774 Mount @var{root} as the root file system. @var{root} can be a
15775 device name like @code{/dev/sda1}, a partition label, or a partition
15778 @item --system=@var{system}
15779 Have @file{/run/booted-system} and @file{/run/current-system} point to
15782 @item modprobe.blacklist=@var{modules}@dots{}
15783 @cindex module, black-listing
15784 @cindex black list, of kernel modules
15785 Instruct the initial RAM disk as well as the @command{modprobe} command
15786 (from the kmod package) to refuse to load @var{modules}. @var{modules}
15787 must be a comma-separated list of module names---e.g.,
15788 @code{usbkbd,9pnet}.
15791 Start a read-eval-print loop (REPL) from the initial RAM disk before it
15792 tries to load kernel modules and to mount the root file system. Our
15793 marketing team calls it @dfn{boot-to-Guile}. The Schemer in you will
15794 love it. @xref{Using Guile Interactively,,, guile, GNU Guile Reference
15795 Manual}, for more information on Guile's REPL.
15799 Now that you know all the features that initial RAM disks produced by
15800 @code{base-initrd} and @code{raw-initrd} provide,
15801 here is how to use it and customize it further.
15804 @cindex initial RAM disk
15805 @deffn {Monadic Procedure} raw-initrd @var{file-systems} @
15806 [#:linux-modules '()] [#:mapped-devices '()] @
15807 [#:helper-packages '()] [#:qemu-networking? #f] [#:volatile-root? #f]
15808 Return a monadic derivation that builds a raw initrd. @var{file-systems} is
15809 a list of file systems to be mounted by the initrd, possibly in addition to
15810 the root file system specified on the kernel command line via @code{--root}.
15811 @var{linux-modules} is a list of kernel modules to be loaded at boot time.
15812 @var{mapped-devices} is a list of device mappings to realize before
15813 @var{file-systems} are mounted (@pxref{Mapped Devices}).
15814 @var{helper-packages} is a list of packages to be copied in the initrd. It may
15815 include @code{e2fsck/static} or other packages needed by the initrd to check
15818 When @var{qemu-networking?} is true, set up networking with the standard QEMU
15819 parameters. When @var{virtio?} is true, load additional modules so that the
15820 initrd can be used as a QEMU guest with para-virtualized I/O drivers.
15822 When @var{volatile-root?} is true, the root file system is writable but any changes
15826 @deffn {Monadic Procedure} base-initrd @var{file-systems} @
15827 [#:mapped-devices '()] [#:qemu-networking? #f] [#:volatile-root? #f]@
15828 [#:virtio? #t] [#:extra-modules '()]
15829 Return a monadic derivation that builds a generic initrd. @var{file-systems} is
15830 a list of file systems to be mounted by the initrd like for @code{raw-initrd}.
15831 @var{mapped-devices}, @var{qemu-networking?} and @var{volatile-root?}
15832 also behaves as in @code{raw-initrd}.
15834 When @var{virtio?} is true, load additional modules so that the
15835 initrd can be used as a QEMU guest with para-virtualized I/O drivers.
15837 The initrd is automatically populated with all the kernel modules necessary
15838 for @var{file-systems} and for the given options. However, additional kernel
15839 modules can be listed in @var{extra-modules}. They will be added to the initrd, and
15840 loaded at boot time in the order in which they appear.
15843 Needless to say, the initrds we produce and use embed a
15844 statically-linked Guile, and the initialization program is a Guile
15845 program. That gives a lot of flexibility. The
15846 @code{expression->initrd} procedure builds such an initrd, given the
15847 program to run in that initrd.
15849 @deffn {Monadic Procedure} expression->initrd @var{exp} @
15850 [#:guile %guile-static-stripped] [#:name "guile-initrd"]
15851 Return a derivation that builds a Linux initrd (a gzipped cpio archive)
15852 containing @var{guile} and that evaluates @var{exp}, a G-expression,
15853 upon booting. All the derivations referenced by @var{exp} are
15854 automatically copied to the initrd.
15857 @node Bootloader Configuration
15858 @subsection Bootloader Configuration
15861 @cindex boot loader
15863 The operating system supports multiple bootloaders. The bootloader is
15864 configured using @code{bootloader-configuration} declaration. All the
15865 fields of this structure are bootloader agnostic except for one field,
15866 @code{bootloader} that indicates the bootloader to be configured and
15869 Some of the bootloaders do not honor every field of
15870 @code{bootloader-configuration}. For instance, the extlinux
15871 bootloader does not support themes and thus ignores the @code{theme}
15874 @deftp {Data Type} bootloader-configuration
15875 The type of a bootloader configuration declaration.
15879 @item @code{bootloader}
15880 @cindex EFI, bootloader
15881 @cindex UEFI, bootloader
15882 @cindex BIOS, bootloader
15883 The bootloader to use, as a @code{bootloader} object. For now
15884 @code{grub-bootloader}, @code{grub-efi-bootloader} and
15885 @code{extlinux-bootloader} are supported. @code{grub-efi-bootloader},
15886 allows to boot on modern systems using the @dfn{Unified Extensible
15887 Firmware Interface} (UEFI).
15889 Available bootloaders are described in @code{(gnu bootloader @dots{})}
15892 @item @code{device}
15893 This is a string denoting the boot device. It must be a device name
15894 understood by the bootloader @command{installer} command, such as
15895 @code{/dev/sda} or @code{(hd0)} (for GRUB, @pxref{Invoking grub-install,,, grub,
15898 @item @code{menu-entries} (default: @code{()})
15899 A possibly empty list of @code{menu-entry} objects (see below), denoting
15900 entries to appear in the bootloader menu, in addition to the current
15901 system entry and the entry pointing to previous system generations.
15904 @item @code{default-entry} (default: @code{0})
15905 The index of the default boot menu entry. Index 0 is for the entry of the
15908 @item @code{timeout} (default: @code{5})
15909 The number of seconds to wait for keyboard input before booting. Set to
15910 0 to boot immediately, and to -1 to wait indefinitely.
15912 @item @code{theme} (default: @var{#f})
15913 The bootloader theme object describing the theme to use. If no theme
15914 is provided, some bootloaders might use a default theme, that's true
15917 @item @code{terminal-outputs} (default: @code{'gfxterm})
15918 The output terminals used for the bootloader boot menu, as a list of
15919 symbols. GRUB accepts the values: @code{console}, @code{serial},
15920 @code{serial_@{0-3@}}, @code{gfxterm}, @code{vga_text},
15921 @code{mda_text}, @code{morse}, and @code{pkmodem}. This field
15922 corresponds to the GRUB variable GRUB_TERMINAL_OUTPUT (@pxref{Simple
15923 configuration,,, grub,GNU GRUB manual}).
15925 @item @code{terminal-inputs} (default: @code{'()})
15926 The input terminals used for the bootloader boot menu, as a list of
15927 symbols. For GRUB, the default is the native platform terminal as
15928 determined at run-time. GRUB accepts the values: @code{console},
15929 @code{serial}, @code{serial_@{0-3@}}, @code{at_keyboard}, and
15930 @code{usb_keyboard}. This field corresponds to the GRUB variable
15931 GRUB_TERMINAL_INPUT (@pxref{Simple configuration,,, grub,GNU GRUB
15934 @item @code{serial-unit} (default: @code{#f})
15935 The serial unit used by the bootloader, as an integer from 0 to 3.
15936 For GRUB it is choosen at run-time; currently GRUB chooses 0, which
15937 corresponds to COM1 (@pxref{Serial terminal,,, grub,GNU GRUB manual}).
15939 @item @code{serial-speed} (default: @code{#f})
15940 The speed of the serial interface, as an integer. For GRUB, the
15941 default value is chosen at run-time; currently GRUB chooses
15942 9600@tie{}bps (@pxref{Serial terminal,,, grub,GNU GRUB manual}).
15949 Should you want to list additional boot menu entries @i{via} the
15950 @code{menu-entries} field above, you will need to create them with the
15951 @code{menu-entry} form. For example, imagine you want to be able to
15952 boot another distro (hard to imagine!), you can define a menu entry
15957 (label "The Other Distro")
15958 (linux "/boot/old/vmlinux-2.6.32")
15959 (linux-arguments '("root=/dev/sda2"))
15960 (initrd "/boot/old/initrd"))
15965 @deftp {Data Type} menu-entry
15966 The type of an entry in the bootloader menu.
15971 The label to show in the menu---e.g., @code{"GNU"}.
15974 The Linux kernel image to boot, for example:
15977 (file-append linux-libre "/bzImage")
15980 For GRUB, it is also possible to specify a device explicitly in the
15981 file path using GRUB's device naming convention (@pxref{Naming
15982 convention,,, grub, GNU GRUB manual}), for example:
15985 "(hd0,msdos1)/boot/vmlinuz"
15988 If the device is specified explicitly as above, then the @code{device}
15989 field is ignored entirely.
15991 @item @code{linux-arguments} (default: @code{()})
15992 The list of extra Linux kernel command-line arguments---e.g.,
15993 @code{("console=ttyS0")}.
15995 @item @code{initrd}
15996 A G-Expression or string denoting the file name of the initial RAM disk
15997 to use (@pxref{G-Expressions}).
15998 @item @code{device} (default: @code{#f})
15999 The device where the kernel and initrd are to be found---i.e., for GRUB,
16000 @dfn{root} for this menu entry (@pxref{root,,, grub, GNU GRUB manual}).
16002 This may be a file system label (a string), a file system UUID (a
16003 bytevector, @pxref{File Systems}), or @code{#f}, in which case
16004 the bootloader will search the device containing the file specified by
16005 the @code{linux} field (@pxref{search,,, grub, GNU GRUB manual}). It
16006 must @emph{not} be an OS device name such as @file{/dev/sda1}.
16011 @c FIXME: Write documentation once it's stable.
16012 Fow now only GRUB has theme support. GRUB themes are created using
16013 the @code{grub-theme} form, which is not documented yet.
16015 @defvr {Scheme Variable} %default-theme
16016 This is the default GRUB theme used by the operating system if no
16017 @code{theme} field is specified in @code{bootloader-configuration}
16020 It comes with a fancy background image displaying the GNU and Guix
16025 @node Invoking guix system
16026 @subsection Invoking @code{guix system}
16028 Once you have written an operating system declaration as seen in the
16029 previous section, it can be @dfn{instantiated} using the @command{guix
16030 system} command. The synopsis is:
16033 guix system @var{options}@dots{} @var{action} @var{file}
16036 @var{file} must be the name of a file containing an
16037 @code{operating-system} declaration. @var{action} specifies how the
16038 operating system is instantiated. Currently the following values are
16043 Build the operating system described in @var{file}, activate it, and
16044 switch to it@footnote{This action (and the related actions
16045 @code{switch-generation} and @code{roll-back}) are usable only on
16046 systems already running GuixSD.}.
16048 This effects all the configuration specified in @var{file}: user
16049 accounts, system services, global package list, setuid programs, etc.
16050 The command starts system services specified in @var{file} that are not
16051 currently running; if a service is currently running, it does not
16052 attempt to upgrade it since this would not be possible without stopping it
16055 This command creates a new generation whose number is one greater than
16056 the current generation (as reported by @command{guix system
16057 list-generations}). If that generation already exists, it will be
16058 overwritten. This behavior mirrors that of @command{guix package}
16059 (@pxref{Invoking guix package}).
16061 It also adds a bootloader menu entry for the new OS configuration,
16062 ---unless @option{--no-bootloader} is passed. For GRUB, it moves
16063 entries for older configurations to a submenu, allowing you to choose
16064 an older system generation at boot time should you need it.
16067 @c The paragraph below refers to the problem discussed at
16068 @c <http://lists.gnu.org/archive/html/guix-devel/2014-08/msg00057.html>.
16069 It is highly recommended to run @command{guix pull} once before you run
16070 @command{guix system reconfigure} for the first time (@pxref{Invoking
16071 guix pull}). Failing to do that you would see an older version of Guix
16072 once @command{reconfigure} has completed.
16075 @item switch-generation
16076 @cindex generations
16077 Switch to an existing system generation. This action atomically
16078 switches the system profile to the specified system generation. It
16079 also rearranges the system's existing bootloader menu entries. It
16080 makes the menu entry for the specified system generation the default,
16081 and it moves the entries for the other generatiors to a submenu, if
16082 supported by the bootloader being used. The next time the system
16083 boots, it will use the specified system generation.
16085 The bootloader itself is not being reinstalled when using this
16086 command. Thus, the installed bootloader is used with an updated
16087 configuration file.
16089 The target generation can be specified explicitly by its generation
16090 number. For example, the following invocation would switch to system
16094 guix system switch-generation 7
16097 The target generation can also be specified relative to the current
16098 generation with the form @code{+N} or @code{-N}, where @code{+3} means
16099 ``3 generations ahead of the current generation,'' and @code{-1} means
16100 ``1 generation prior to the current generation.'' When specifying a
16101 negative value such as @code{-1}, you must precede it with @code{--} to
16102 prevent it from being parsed as an option. For example:
16105 guix system switch-generation -- -1
16108 Currently, the effect of invoking this action is @emph{only} to switch
16109 the system profile to an existing generation and rearrange the
16110 bootloader menu entries. To actually start using the target system
16111 generation, you must reboot after running this action. In the future,
16112 it will be updated to do the same things as @command{reconfigure},
16113 like activating and deactivating services.
16115 This action will fail if the specified generation does not exist.
16118 @cindex rolling back
16119 Switch to the preceding system generation. The next time the system
16120 boots, it will use the preceding system generation. This is the inverse
16121 of @command{reconfigure}, and it is exactly the same as invoking
16122 @command{switch-generation} with an argument of @code{-1}.
16124 Currently, as with @command{switch-generation}, you must reboot after
16125 running this action to actually start using the preceding system
16129 Build the derivation of the operating system, which includes all the
16130 configuration files and programs needed to boot and run the system.
16131 This action does not actually install anything.
16134 Populate the given directory with all the files necessary to run the
16135 operating system specified in @var{file}. This is useful for first-time
16136 installations of GuixSD. For instance:
16139 guix system init my-os-config.scm /mnt
16142 copies to @file{/mnt} all the store items required by the configuration
16143 specified in @file{my-os-config.scm}. This includes configuration
16144 files, packages, and so on. It also creates other essential files
16145 needed for the system to operate correctly---e.g., the @file{/etc},
16146 @file{/var}, and @file{/run} directories, and the @file{/bin/sh} file.
16148 This command also installs bootloader on the device specified in
16149 @file{my-os-config}, unless the @option{--no-bootloader} option was
16153 @cindex virtual machine
16155 @anchor{guix system vm}
16156 Build a virtual machine that contains the operating system declared in
16157 @var{file}, and return a script to run that virtual machine (VM).
16158 Arguments given to the script are passed to QEMU.
16160 The VM shares its store with the host system.
16162 Additional file systems can be shared between the host and the VM using
16163 the @code{--share} and @code{--expose} command-line options: the former
16164 specifies a directory to be shared with write access, while the latter
16165 provides read-only access to the shared directory.
16167 The example below creates a VM in which the user's home directory is
16168 accessible read-only, and where the @file{/exchange} directory is a
16169 read-write mapping of @file{$HOME/tmp} on the host:
16172 guix system vm my-config.scm \
16173 --expose=$HOME --share=$HOME/tmp=/exchange
16176 On GNU/Linux, the default is to boot directly to the kernel; this has
16177 the advantage of requiring only a very tiny root disk image since the
16178 store of the host can then be mounted.
16180 The @code{--full-boot} option forces a complete boot sequence, starting
16181 with the bootloader. This requires more disk space since a root image
16182 containing at least the kernel, initrd, and bootloader data files must
16183 be created. The @code{--image-size} option can be used to specify the
16188 Return a virtual machine or disk image of the operating system declared
16189 in @var{file} that stands alone. By default, @command{guix system}
16190 estimates the size of the image needed to store the system, but you can
16191 use the @option{--image-size} option to specify a value.
16193 You can specify the root file system type by using the
16194 @option{--file-system-type} option. It defaults to @code{ext4}.
16196 When using @code{vm-image}, the returned image is in qcow2 format, which
16197 the QEMU emulator can efficiently use. @xref{Running GuixSD in a VM},
16198 for more information on how to run the image in a virtual machine.
16200 When using @code{disk-image}, a raw disk image is produced; it can be
16201 copied as is to a USB stick, for instance. Assuming @code{/dev/sdc} is
16202 the device corresponding to a USB stick, one can copy the image to it
16203 using the following command:
16206 # dd if=$(guix system disk-image my-os.scm) of=/dev/sdc
16210 Return a script to run the operating system declared in @var{file}
16211 within a container. Containers are a set of lightweight isolation
16212 mechanisms provided by the kernel Linux-libre. Containers are
16213 substantially less resource-demanding than full virtual machines since
16214 the kernel, shared objects, and other resources can be shared with the
16215 host system; this also means they provide thinner isolation.
16217 Currently, the script must be run as root in order to support more than
16218 a single user and group. The container shares its store with the host
16221 As with the @code{vm} action (@pxref{guix system vm}), additional file
16222 systems to be shared between the host and container can be specified
16223 using the @option{--share} and @option{--expose} options:
16226 guix system container my-config.scm \
16227 --expose=$HOME --share=$HOME/tmp=/exchange
16231 This option requires Linux-libre 3.19 or newer.
16236 @var{options} can contain any of the common build options (@pxref{Common
16237 Build Options}). In addition, @var{options} can contain one of the
16241 @item --system=@var{system}
16242 @itemx -s @var{system}
16243 Attempt to build for @var{system} instead of the host system type.
16244 This works as per @command{guix build} (@pxref{Invoking guix build}).
16248 Return the derivation file name of the given operating system without
16251 @item --file-system-type=@var{type}
16252 @itemx -t @var{type}
16253 For the @code{disk-image} action, create a file system of the given
16254 @var{type} on the image.
16256 When this option is omitted, @command{guix system} uses @code{ext4}.
16258 @cindex ISO-9660 format
16259 @cindex CD image format
16260 @cindex DVD image format
16261 @code{--file-system-type=iso9660} produces an ISO-9660 image, suitable
16262 for burning on CDs and DVDs.
16264 @item --image-size=@var{size}
16265 For the @code{vm-image} and @code{disk-image} actions, create an image
16266 of the given @var{size}. @var{size} may be a number of bytes, or it may
16267 include a unit as a suffix (@pxref{Block size, size specifications,,
16268 coreutils, GNU Coreutils}).
16270 When this option is omitted, @command{guix system} computes an estimate
16271 of the image size as a function of the size of the system declared in
16274 @item --root=@var{file}
16275 @itemx -r @var{file}
16276 Make @var{file} a symlink to the result, and register it as a garbage
16279 @item --on-error=@var{strategy}
16280 Apply @var{strategy} when an error occurs when reading @var{file}.
16281 @var{strategy} may be one of the following:
16284 @item nothing-special
16285 Report the error concisely and exit. This is the default strategy.
16288 Likewise, but also display a backtrace.
16291 Report the error and enter Guile's debugger. From there, you can run
16292 commands such as @code{,bt} to get a backtrace, @code{,locals} to
16293 display local variable values, and more generally inspect the state of the
16294 program. @xref{Debug Commands,,, guile, GNU Guile Reference Manual}, for
16295 a list of available debugging commands.
16300 All the actions above, except @code{build} and @code{init},
16301 can use KVM support in the Linux-libre kernel. Specifically, if the
16302 machine has hardware virtualization support, the corresponding
16303 KVM kernel module should be loaded, and the @file{/dev/kvm} device node
16304 must exist and be readable and writable by the user and by the
16305 build users of the daemon (@pxref{Build Environment Setup}).
16308 Once you have built, configured, re-configured, and re-re-configured
16309 your GuixSD installation, you may find it useful to list the operating
16310 system generations available on disk---and that you can choose from the
16311 bootloader boot menu:
16315 @item list-generations
16316 List a summary of each generation of the operating system available on
16317 disk, in a human-readable way. This is similar to the
16318 @option{--list-generations} option of @command{guix package}
16319 (@pxref{Invoking guix package}).
16321 Optionally, one can specify a pattern, with the same syntax that is used
16322 in @command{guix package --list-generations}, to restrict the list of
16323 generations displayed. For instance, the following command displays
16324 generations that are up to 10 days old:
16327 $ guix system list-generations 10d
16332 The @command{guix system} command has even more to offer! The following
16333 sub-commands allow you to visualize how your system services relate to
16336 @anchor{system-extension-graph}
16339 @item extension-graph
16340 Emit in Dot/Graphviz format to standard output the @dfn{service
16341 extension graph} of the operating system defined in @var{file}
16342 (@pxref{Service Composition}, for more information on service
16348 $ guix system extension-graph @var{file} | dot -Tpdf > services.pdf
16351 produces a PDF file showing the extension relations among services.
16353 @anchor{system-shepherd-graph}
16354 @item shepherd-graph
16355 Emit in Dot/Graphviz format to standard output the @dfn{dependency
16356 graph} of shepherd services of the operating system defined in
16357 @var{file}. @xref{Shepherd Services}, for more information and for an
16362 @node Running GuixSD in a VM
16363 @subsection Running GuixSD in a Virtual Machine
16365 @cindex virtual machine
16366 To run GuixSD in a virtual machine (VM), one can either use the
16367 pre-built GuixSD VM image distributed at
16368 @indicateurl{ftp://alpha.gnu.org/guix/guixsd-vm-image-@value{VERSION}.@var{system}.tar.xz}
16369 , or build their own virtual machine image using @command{guix system
16370 vm-image} (@pxref{Invoking guix system}). The returned image is in
16371 qcow2 format, which the @uref{http://qemu.org/, QEMU emulator} can
16375 If you built your own image, you must copy it out of the store
16376 (@pxref{The Store}) and give yourself permission to write to the copy
16377 before you can use it. When invoking QEMU, you must choose a system
16378 emulator that is suitable for your hardware platform. Here is a minimal
16379 QEMU invocation that will boot the result of @command{guix system
16380 vm-image} on x86_64 hardware:
16383 $ qemu-system-x86_64 \
16384 -net user -net nic,model=virtio \
16385 -enable-kvm -m 256 /tmp/qemu-image
16388 Here is what each of these options means:
16391 @item qemu-system-x86_64
16392 This specifies the hardware platform to emulate. This should match the
16396 Enable the unprivileged user-mode network stack. The guest OS can
16397 access the host but not vice versa. This is the simplest way to get the
16400 @item -net nic,model=virtio
16401 You must create a network interface of a given model. If you do not
16402 create a NIC, the boot will fail. Assuming your hardware platform is
16403 x86_64, you can get a list of available NIC models by running
16404 @command{qemu-system-x86_64 -net nic,model=help}.
16407 If your system has hardware virtualization extensions, enabling the
16408 virtual machine support (KVM) of the Linux kernel will make things run
16412 RAM available to the guest OS, in mebibytes. Defaults to 128@tie{}MiB,
16413 which may be insufficient for some operations.
16415 @item /tmp/qemu-image
16416 The file name of the qcow2 image.
16419 The default @command{run-vm.sh} script that is returned by an invocation of
16420 @command{guix system vm} does not add a @command{-net user} flag by default.
16421 To get network access from within the vm add the @code{(dhcp-client-service)}
16422 to your system definition and start the VM using
16423 @command{`guix system vm config.scm` -net user}. An important caveat of using
16424 @command{-net user} for networking is that @command{ping} will not work, because
16425 it uses the ICMP protocol. You'll have to use a different command to check for
16426 network connectivity, for example @command{guix download}.
16428 @subsubsection Connecting Through SSH
16432 To enable SSH inside a VM you need to add a SSH server like @code{(dropbear-service)}
16433 or @code{(lsh-service)} to your VM. The @code{(lsh-service}) doesn't currently
16434 boot unsupervised. It requires you to type some characters to initialize the
16435 randomness generator. In addition you need to forward the SSH port, 22 by
16436 default, to the host. You can do this with
16439 `guix system vm config.scm` -net user,hostfwd=tcp::10022-:22
16442 To connect to the VM you can run
16445 ssh -o UserKnownHostsFile=/dev/null -o StrictHostKeyChecking=no -p 10022
16448 The @command{-p} tells @command{ssh} the port you want to connect to.
16449 @command{-o UserKnownHostsFile=/dev/null} prevents @command{ssh} from complaining
16450 every time you modify your @command{config.scm} file and the
16451 @command{-o StrictHostKeyChecking=no} prevents you from having to allow a
16452 connection to an unknown host every time you connect.
16454 @subsubsection Using @command{virt-viewer} with Spice
16456 As an alternative to the default @command{qemu} graphical client you can
16457 use the @command{remote-viewer} from the @command{virt-viewer} package. To
16458 connect pass the @command{-spice port=5930,disable-ticketing} flag to
16459 @command{qemu}. See previous section for further information on how to do this.
16461 Spice also allows you to do some nice stuff like share your clipboard with your
16462 VM. To enable that you'll also have to pass the following flags to @command{qemu}:
16465 -device virtio-serial-pci,id=virtio-serial0,max_ports=16,bus=pci.0,addr=0x5
16466 -chardev spicevmc,name=vdagent,id=vdagent
16467 -device virtserialport,nr=1,bus=virtio-serial0.0,chardev=vdagent,
16468 name=com.redhat.spice.0
16471 You'll also need to add the @pxref{Miscellaneous Services, Spice service}.
16473 @node Defining Services
16474 @subsection Defining Services
16476 The previous sections show the available services and how one can combine
16477 them in an @code{operating-system} declaration. But how do we define
16478 them in the first place? And what is a service anyway?
16481 * Service Composition:: The model for composing services.
16482 * Service Types and Services:: Types and services.
16483 * Service Reference:: API reference.
16484 * Shepherd Services:: A particular type of service.
16487 @node Service Composition
16488 @subsubsection Service Composition
16492 Here we define a @dfn{service} as, broadly, something that extends the
16493 functionality of the operating system. Often a service is a process---a
16494 @dfn{daemon}---started when the system boots: a secure shell server, a
16495 Web server, the Guix build daemon, etc. Sometimes a service is a daemon
16496 whose execution can be triggered by another daemon---e.g., an FTP server
16497 started by @command{inetd} or a D-Bus service activated by
16498 @command{dbus-daemon}. Occasionally, a service does not map to a
16499 daemon. For instance, the ``account'' service collects user accounts
16500 and makes sure they exist when the system runs; the ``udev'' service
16501 collects device management rules and makes them available to the eudev
16502 daemon; the @file{/etc} service populates the @file{/etc} directory
16505 @cindex service extensions
16506 GuixSD services are connected by @dfn{extensions}. For instance, the
16507 secure shell service @emph{extends} the Shepherd---the GuixSD
16508 initialization system, running as PID@tie{}1---by giving it the command
16509 lines to start and stop the secure shell daemon (@pxref{Networking
16510 Services, @code{lsh-service}}); the UPower service extends the D-Bus
16511 service by passing it its @file{.service} specification, and extends the
16512 udev service by passing it device management rules (@pxref{Desktop
16513 Services, @code{upower-service}}); the Guix daemon service extends the
16514 Shepherd by passing it the command lines to start and stop the daemon,
16515 and extends the account service by passing it a list of required build
16516 user accounts (@pxref{Base Services}).
16518 All in all, services and their ``extends'' relations form a directed
16519 acyclic graph (DAG). If we represent services as boxes and extensions
16520 as arrows, a typical system might provide something like this:
16522 @image{images/service-graph,,5in,Typical service extension graph.}
16524 @cindex system service
16525 At the bottom, we see the @dfn{system service}, which produces the
16526 directory containing everything to run and boot the system, as returned
16527 by the @command{guix system build} command. @xref{Service Reference},
16528 to learn about the other service types shown here.
16529 @xref{system-extension-graph, the @command{guix system extension-graph}
16530 command}, for information on how to generate this representation for a
16531 particular operating system definition.
16533 @cindex service types
16534 Technically, developers can define @dfn{service types} to express these
16535 relations. There can be any number of services of a given type on the
16536 system---for instance, a system running two instances of the GNU secure
16537 shell server (lsh) has two instances of @var{lsh-service-type}, with
16538 different parameters.
16540 The following section describes the programming interface for service
16541 types and services.
16543 @node Service Types and Services
16544 @subsubsection Service Types and Services
16546 A @dfn{service type} is a node in the DAG described above. Let us start
16547 with a simple example, the service type for the Guix build daemon
16548 (@pxref{Invoking guix-daemon}):
16551 (define guix-service-type
16555 (list (service-extension shepherd-root-service-type guix-shepherd-service)
16556 (service-extension account-service-type guix-accounts)
16557 (service-extension activation-service-type guix-activation)))
16558 (default-value (guix-configuration))))
16562 It defines three things:
16566 A name, whose sole purpose is to make inspection and debugging easier.
16569 A list of @dfn{service extensions}, where each extension designates the
16570 target service type and a procedure that, given the parameters of the
16571 service, returns a list of objects to extend the service of that type.
16573 Every service type has at least one service extension. The only
16574 exception is the @dfn{boot service type}, which is the ultimate service.
16577 Optionally, a default value for instances of this type.
16580 In this example, @var{guix-service-type} extends three services:
16583 @item shepherd-root-service-type
16584 The @var{guix-shepherd-service} procedure defines how the Shepherd
16585 service is extended. Namely, it returns a @code{<shepherd-service>}
16586 object that defines how @command{guix-daemon} is started and stopped
16587 (@pxref{Shepherd Services}).
16589 @item account-service-type
16590 This extension for this service is computed by @var{guix-accounts},
16591 which returns a list of @code{user-group} and @code{user-account}
16592 objects representing the build user accounts (@pxref{Invoking
16595 @item activation-service-type
16596 Here @var{guix-activation} is a procedure that returns a gexp, which is
16597 a code snippet to run at ``activation time''---e.g., when the service is
16601 A service of this type is instantiated like this:
16604 (service guix-service-type
16605 (guix-configuration
16607 (use-substitutes? #f)))
16610 The second argument to the @code{service} form is a value representing
16611 the parameters of this specific service instance.
16612 @xref{guix-configuration-type, @code{guix-configuration}}, for
16613 information about the @code{guix-configuration} data type. When the
16614 value is omitted, the default value specified by
16615 @code{guix-service-type} is used:
16618 (service guix-service-type)
16621 @var{guix-service-type} is quite simple because it extends other
16622 services but is not extensible itself.
16624 @c @subsubsubsection Extensible Service Types
16626 The service type for an @emph{extensible} service looks like this:
16629 (define udev-service-type
16630 (service-type (name 'udev)
16632 (list (service-extension shepherd-root-service-type
16633 udev-shepherd-service)))
16635 (compose concatenate) ;concatenate the list of rules
16636 (extend (lambda (config rules)
16638 (($ <udev-configuration> udev initial-rules)
16639 (udev-configuration
16640 (udev udev) ;the udev package to use
16641 (rules (append initial-rules rules)))))))))
16644 This is the service type for the
16645 @uref{https://wiki.gentoo.org/wiki/Project:Eudev, eudev device
16646 management daemon}. Compared to the previous example, in addition to an
16647 extension of @var{shepherd-root-service-type}, we see two new fields:
16651 This is the procedure to @dfn{compose} the list of extensions to
16652 services of this type.
16654 Services can extend the udev service by passing it lists of rules; we
16655 compose those extensions simply by concatenating them.
16658 This procedure defines how the value of the service is @dfn{extended} with
16659 the composition of the extensions.
16661 Udev extensions are composed into a list of rules, but the udev service
16662 value is itself a @code{<udev-configuration>} record. So here, we
16663 extend that record by appending the list of rules it contains to the
16664 list of contributed rules.
16667 There can be only one instance of an extensible service type such as
16668 @var{udev-service-type}. If there were more, the
16669 @code{service-extension} specifications would be ambiguous.
16671 Still here? The next section provides a reference of the programming
16672 interface for services.
16674 @node Service Reference
16675 @subsubsection Service Reference
16677 We have seen an overview of service types (@pxref{Service Types and
16678 Services}). This section provides a reference on how to manipulate
16679 services and service types. This interface is provided by the
16680 @code{(gnu services)} module.
16682 @deffn {Scheme Procedure} service @var{type} [@var{value}]
16683 Return a new service of @var{type}, a @code{<service-type>} object (see
16684 below.) @var{value} can be any object; it represents the parameters of
16685 this particular service instance.
16687 When @var{value} is omitted, the default value specified by @var{type}
16688 is used; if @var{type} does not specify a default value, an error is
16691 For instance, this:
16694 (service openssh-service-type)
16698 is equivalent to this:
16701 (service openssh-service-type
16702 (openssh-configuration))
16705 In both cases the result is an instance of @code{openssh-service-type}
16706 with the default configuration.
16709 @deffn {Scheme Procedure} service? @var{obj}
16710 Return true if @var{obj} is a service.
16713 @deffn {Scheme Procedure} service-kind @var{service}
16714 Return the type of @var{service}---i.e., a @code{<service-type>} object.
16717 @deffn {Scheme Procedure} service-value @var{service}
16718 Return the value associated with @var{service}. It represents its
16722 Here is an example of how a service is created and manipulated:
16726 (service nginx-service-type
16727 (nginx-configuration
16729 (log-directory log-directory)
16730 (run-directory run-directory)
16731 (file config-file))))
16736 (eq? (service-kind s) nginx-service-type)
16740 The @code{modify-services} form provides a handy way to change the
16741 parameters of some of the services of a list such as
16742 @var{%base-services} (@pxref{Base Services, @code{%base-services}}). It
16743 evaluates to a list of services. Of course, you could always use
16744 standard list combinators such as @code{map} and @code{fold} to do that
16745 (@pxref{SRFI-1, List Library,, guile, GNU Guile Reference Manual});
16746 @code{modify-services} simply provides a more concise form for this
16749 @deffn {Scheme Syntax} modify-services @var{services} @
16750 (@var{type} @var{variable} => @var{body}) @dots{}
16752 Modify the services listed in @var{services} according to the given
16753 clauses. Each clause has the form:
16756 (@var{type} @var{variable} => @var{body})
16759 where @var{type} is a service type---e.g.,
16760 @code{guix-service-type}---and @var{variable} is an identifier that is
16761 bound within the @var{body} to the service parameters---e.g., a
16762 @code{guix-configuration} instance---of the original service of that
16765 The @var{body} should evaluate to the new service parameters, which will
16766 be used to configure the new service. This new service will replace the
16767 original in the resulting list. Because a service's service parameters
16768 are created using @code{define-record-type*}, you can write a succinct
16769 @var{body} that evaluates to the new service parameters by using the
16770 @code{inherit} feature that @code{define-record-type*} provides.
16772 @xref{Using the Configuration System}, for example usage.
16776 Next comes the programming interface for service types. This is
16777 something you want to know when writing new service definitions, but not
16778 necessarily when simply looking for ways to customize your
16779 @code{operating-system} declaration.
16781 @deftp {Data Type} service-type
16782 @cindex service type
16783 This is the representation of a @dfn{service type} (@pxref{Service Types
16788 This is a symbol, used only to simplify inspection and debugging.
16790 @item @code{extensions}
16791 A non-empty list of @code{<service-extension>} objects (see below).
16793 @item @code{compose} (default: @code{#f})
16794 If this is @code{#f}, then the service type denotes services that cannot
16795 be extended---i.e., services that do not receive ``values'' from other
16798 Otherwise, it must be a one-argument procedure. The procedure is called
16799 by @code{fold-services} and is passed a list of values collected from
16800 extensions. It must return a value that is a valid parameter value for
16801 the service instance.
16803 @item @code{extend} (default: @code{#f})
16804 If this is @code{#f}, services of this type cannot be extended.
16806 Otherwise, it must be a two-argument procedure: @code{fold-services}
16807 calls it, passing it the initial value of the service as the first argument
16808 and the result of applying @code{compose} to the extension values as the
16812 @xref{Service Types and Services}, for examples.
16815 @deffn {Scheme Procedure} service-extension @var{target-type} @
16817 Return a new extension for services of type @var{target-type}.
16818 @var{compute} must be a one-argument procedure: @code{fold-services}
16819 calls it, passing it the value associated with the service that provides
16820 the extension; it must return a valid value for the target service.
16823 @deffn {Scheme Procedure} service-extension? @var{obj}
16824 Return true if @var{obj} is a service extension.
16827 Occasionally, you might want to simply extend an existing service. This
16828 involves creating a new service type and specifying the extension of
16829 interest, which can be verbose; the @code{simple-service} procedure
16830 provides a shorthand for this.
16832 @deffn {Scheme Procedure} simple-service @var{name} @var{target} @var{value}
16833 Return a service that extends @var{target} with @var{value}. This works
16834 by creating a singleton service type @var{name}, of which the returned
16835 service is an instance.
16837 For example, this extends mcron (@pxref{Scheduled Job Execution}) with
16841 (simple-service 'my-mcron-job mcron-service-type
16842 #~(job '(next-hour (3)) "guix gc -F 2G"))
16846 At the core of the service abstraction lies the @code{fold-services}
16847 procedure, which is responsible for ``compiling'' a list of services
16848 down to a single directory that contains everything needed to boot and
16849 run the system---the directory shown by the @command{guix system build}
16850 command (@pxref{Invoking guix system}). In essence, it propagates
16851 service extensions down the service graph, updating each node parameters
16852 on the way, until it reaches the root node.
16854 @deffn {Scheme Procedure} fold-services @var{services} @
16855 [#:target-type @var{system-service-type}]
16856 Fold @var{services} by propagating their extensions down to the root of
16857 type @var{target-type}; return the root service adjusted accordingly.
16860 Lastly, the @code{(gnu services)} module also defines several essential
16861 service types, some of which are listed below.
16863 @defvr {Scheme Variable} system-service-type
16864 This is the root of the service graph. It produces the system directory
16865 as returned by the @command{guix system build} command.
16868 @defvr {Scheme Variable} boot-service-type
16869 The type of the ``boot service'', which produces the @dfn{boot script}.
16870 The boot script is what the initial RAM disk runs when booting.
16873 @defvr {Scheme Variable} etc-service-type
16874 The type of the @file{/etc} service. This service can be extended by
16875 passing it name/file tuples such as:
16878 (list `("issue" ,(plain-file "issue" "Welcome!\n")))
16881 In this example, the effect would be to add an @file{/etc/issue} file
16882 pointing to the given file.
16885 @defvr {Scheme Variable} setuid-program-service-type
16886 Type for the ``setuid-program service''. This service collects lists of
16887 executable file names, passed as gexps, and adds them to the set of
16888 setuid-root programs on the system (@pxref{Setuid Programs}).
16891 @defvr {Scheme Variable} profile-service-type
16892 Type of the service that populates the @dfn{system profile}---i.e., the
16893 programs under @file{/run/current-system/profile}. Other services can
16894 extend it by passing it lists of packages to add to the system profile.
16898 @node Shepherd Services
16899 @subsubsection Shepherd Services
16901 @cindex shepherd services
16903 @cindex init system
16904 The @code{(gnu services shepherd)} module provides a way to define
16905 services managed by the GNU@tie{}Shepherd, which is the GuixSD
16906 initialization system---the first process that is started when the
16907 system boots, also known as PID@tie{}1
16908 (@pxref{Introduction,,, shepherd, The GNU Shepherd Manual}).
16910 Services in the Shepherd can depend on each other. For instance, the
16911 SSH daemon may need to be started after the syslog daemon has been
16912 started, which in turn can only happen once all the file systems have
16913 been mounted. The simple operating system defined earlier (@pxref{Using
16914 the Configuration System}) results in a service graph like this:
16916 @image{images/shepherd-graph,,5in,Typical shepherd service graph.}
16918 You can actually generate such a graph for any operating system
16919 definition using the @command{guix system shepherd-graph} command
16920 (@pxref{system-shepherd-graph, @command{guix system shepherd-graph}}).
16922 The @var{%shepherd-root-service} is a service object representing
16923 PID@tie{}1, of type @var{shepherd-root-service-type}; it can be extended
16924 by passing it lists of @code{<shepherd-service>} objects.
16926 @deftp {Data Type} shepherd-service
16927 The data type representing a service managed by the Shepherd.
16930 @item @code{provision}
16931 This is a list of symbols denoting what the service provides.
16933 These are the names that may be passed to @command{herd start},
16934 @command{herd status}, and similar commands (@pxref{Invoking herd,,,
16935 shepherd, The GNU Shepherd Manual}). @xref{Slots of services, the
16936 @code{provides} slot,, shepherd, The GNU Shepherd Manual}, for details.
16938 @item @code{requirements} (default: @code{'()})
16939 List of symbols denoting the Shepherd services this one depends on.
16941 @item @code{respawn?} (default: @code{#t})
16942 Whether to restart the service when it stops, for instance when the
16943 underlying process dies.
16946 @itemx @code{stop} (default: @code{#~(const #f)})
16947 The @code{start} and @code{stop} fields refer to the Shepherd's
16948 facilities to start and stop processes (@pxref{Service De- and
16949 Constructors,,, shepherd, The GNU Shepherd Manual}). They are given as
16950 G-expressions that get expanded in the Shepherd configuration file
16951 (@pxref{G-Expressions}).
16953 @item @code{documentation}
16954 A documentation string, as shown when running:
16957 herd doc @var{service-name}
16960 where @var{service-name} is one of the symbols in @var{provision}
16961 (@pxref{Invoking herd,,, shepherd, The GNU Shepherd Manual}).
16963 @item @code{modules} (default: @var{%default-modules})
16964 This is the list of modules that must be in scope when @code{start} and
16965 @code{stop} are evaluated.
16970 @defvr {Scheme Variable} shepherd-root-service-type
16971 The service type for the Shepherd ``root service''---i.e., PID@tie{}1.
16973 This is the service type that extensions target when they want to create
16974 shepherd services (@pxref{Service Types and Services}, for an example).
16975 Each extension must pass a list of @code{<shepherd-service>}.
16978 @defvr {Scheme Variable} %shepherd-root-service
16979 This service represents PID@tie{}1.
16983 @node Documentation
16984 @section Documentation
16986 @cindex documentation, searching for
16987 @cindex searching for documentation
16988 @cindex Info, documentation format
16990 @cindex manual pages
16991 In most cases packages installed with Guix come with documentation.
16992 There are two main documentation formats: ``Info'', a browseable
16993 hypertext format used for GNU software, and ``manual pages'' (or ``man
16994 pages''), the linear documentation format traditionally found on Unix.
16995 Info manuals are accessed with the @command{info} command or with Emacs,
16996 and man pages are accessed using @command{man}.
16998 You can look for documentation of software installed on your system by
16999 keyword. For example, the following command searches for information
17000 about ``TLS'' in Info manuals:
17004 "(emacs)Network Security" -- STARTTLS
17005 "(emacs)Network Security" -- TLS
17006 "(gnutls)Core TLS API" -- gnutls_certificate_set_verify_flags
17007 "(gnutls)Core TLS API" -- gnutls_certificate_set_verify_function
17012 The command below searches for the same keyword in man pages:
17016 SSL (7) - OpenSSL SSL/TLS library
17017 certtool (1) - GnuTLS certificate tool
17021 These searches are purely local to your computer so you have the
17022 guarantee that documentation you find corresponds to what you have
17023 actually installed, you can access it off-line, and your privacy is
17026 Once you have these results, you can view the relevant documentation by
17030 $ info "(gnutls)Core TLS API"
17040 Info manuals contain sections and indices as well as hyperlinks like
17041 those found in Web pages. The @command{info} reader (@pxref{Top, Info
17042 reader,, info-stnd, Stand-alone GNU Info}) and its Emacs counterpart
17043 (@pxref{Misc Help,,, emacs, The GNU Emacs Manual}) provide intuitive key
17044 bindings to navigate manuals. @xref{Getting Started,,, info, Info: An
17045 Introduction}, for an introduction to Info navigation.
17047 @node Installing Debugging Files
17048 @section Installing Debugging Files
17050 @cindex debugging files
17051 Program binaries, as produced by the GCC compilers for instance, are
17052 typically written in the ELF format, with a section containing
17053 @dfn{debugging information}. Debugging information is what allows the
17054 debugger, GDB, to map binary code to source code; it is required to
17055 debug a compiled program in good conditions.
17057 The problem with debugging information is that is takes up a fair amount
17058 of disk space. For example, debugging information for the GNU C Library
17059 weighs in at more than 60 MiB. Thus, as a user, keeping all the
17060 debugging info of all the installed programs is usually not an option.
17061 Yet, space savings should not come at the cost of an impediment to
17062 debugging---especially in the GNU system, which should make it easier
17063 for users to exert their computing freedom (@pxref{GNU Distribution}).
17065 Thankfully, the GNU Binary Utilities (Binutils) and GDB provide a
17066 mechanism that allows users to get the best of both worlds: debugging
17067 information can be stripped from the binaries and stored in separate
17068 files. GDB is then able to load debugging information from those files,
17069 when they are available (@pxref{Separate Debug Files,,, gdb, Debugging
17072 The GNU distribution takes advantage of this by storing debugging
17073 information in the @code{lib/debug} sub-directory of a separate package
17074 output unimaginatively called @code{debug} (@pxref{Packages with
17075 Multiple Outputs}). Users can choose to install the @code{debug} output
17076 of a package when they need it. For instance, the following command
17077 installs the debugging information for the GNU C Library and for GNU
17081 guix package -i glibc:debug guile:debug
17084 GDB must then be told to look for debug files in the user's profile, by
17085 setting the @code{debug-file-directory} variable (consider setting it
17086 from the @file{~/.gdbinit} file, @pxref{Startup,,, gdb, Debugging with
17090 (gdb) set debug-file-directory ~/.guix-profile/lib/debug
17093 From there on, GDB will pick up debugging information from the
17094 @code{.debug} files under @file{~/.guix-profile/lib/debug}.
17096 In addition, you will most likely want GDB to be able to show the source
17097 code being debugged. To do that, you will have to unpack the source
17098 code of the package of interest (obtained with @code{guix build
17099 --source}, @pxref{Invoking guix build}), and to point GDB to that source
17100 directory using the @code{directory} command (@pxref{Source Path,
17101 @code{directory},, gdb, Debugging with GDB}).
17103 @c XXX: keep me up-to-date
17104 The @code{debug} output mechanism in Guix is implemented by the
17105 @code{gnu-build-system} (@pxref{Build Systems}). Currently, it is
17106 opt-in---debugging information is available only for the packages
17107 with definitions explicitly declaring a @code{debug} output. This may be
17108 changed to opt-out in the future if our build farm servers can handle
17109 the load. To check whether a package has a @code{debug} output, use
17110 @command{guix package --list-available} (@pxref{Invoking guix package}).
17113 @node Security Updates
17114 @section Security Updates
17116 @cindex security updates
17117 @cindex security vulnerabilities
17118 Occasionally, important security vulnerabilities are discovered in software
17119 packages and must be patched. Guix developers try hard to keep track of
17120 known vulnerabilities and to apply fixes as soon as possible in the
17121 @code{master} branch of Guix (we do not yet provide a ``stable'' branch
17122 containing only security updates.) The @command{guix lint} tool helps
17123 developers find out about vulnerable versions of software packages in the
17128 gnu/packages/base.scm:652:2: glibc@@2.21: probably vulnerable to CVE-2015-1781, CVE-2015-7547
17129 gnu/packages/gcc.scm:334:2: gcc@@4.9.3: probably vulnerable to CVE-2015-5276
17130 gnu/packages/image.scm:312:2: openjpeg@@2.1.0: probably vulnerable to CVE-2016-1923, CVE-2016-1924
17134 @xref{Invoking guix lint}, for more information.
17137 As of version @value{VERSION}, the feature described below is considered
17141 Guix follows a functional
17142 package management discipline (@pxref{Introduction}), which implies
17143 that, when a package is changed, @emph{every package that depends on it}
17144 must be rebuilt. This can significantly slow down the deployment of
17145 fixes in core packages such as libc or Bash, since basically the whole
17146 distribution would need to be rebuilt. Using pre-built binaries helps
17147 (@pxref{Substitutes}), but deployment may still take more time than
17151 To address this, Guix implements @dfn{grafts}, a mechanism that allows
17152 for fast deployment of critical updates without the costs associated
17153 with a whole-distribution rebuild. The idea is to rebuild only the
17154 package that needs to be patched, and then to ``graft'' it onto packages
17155 explicitly installed by the user and that were previously referring to
17156 the original package. The cost of grafting is typically very low, and
17157 order of magnitudes lower than a full rebuild of the dependency chain.
17159 @cindex replacements of packages, for grafts
17160 For instance, suppose a security update needs to be applied to Bash.
17161 Guix developers will provide a package definition for the ``fixed''
17162 Bash, say @var{bash-fixed}, in the usual way (@pxref{Defining
17163 Packages}). Then, the original package definition is augmented with a
17164 @code{replacement} field pointing to the package containing the bug fix:
17171 (replacement bash-fixed)))
17174 From there on, any package depending directly or indirectly on Bash---as
17175 reported by @command{guix gc --requisites} (@pxref{Invoking guix
17176 gc})---that is installed is automatically ``rewritten'' to refer to
17177 @var{bash-fixed} instead of @var{bash}. This grafting process takes
17178 time proportional to the size of the package, usually less than a
17179 minute for an ``average'' package on a recent machine. Grafting is
17180 recursive: when an indirect dependency requires grafting, then grafting
17181 ``propagates'' up to the package that the user is installing.
17183 Currently, the length of the name and version of the graft and that of
17184 the package it replaces (@var{bash-fixed} and @var{bash} in the example
17185 above) must be equal. This restriction mostly comes from the fact that
17186 grafting works by patching files, including binary files, directly.
17187 Other restrictions may apply: for instance, when adding a graft to a
17188 package providing a shared library, the original shared library and its
17189 replacement must have the same @code{SONAME} and be binary-compatible.
17191 The @option{--no-grafts} command-line option allows you to forcefully
17192 avoid grafting (@pxref{Common Build Options, @option{--no-grafts}}).
17196 guix build bash --no-grafts
17200 returns the store file name of the original Bash, whereas:
17207 returns the store file name of the ``fixed'', replacement Bash. This
17208 allows you to distinguish between the two variants of Bash.
17210 To verify which Bash your whole profile refers to, you can run
17211 (@pxref{Invoking guix gc}):
17214 guix gc -R `readlink -f ~/.guix-profile` | grep bash
17218 @dots{} and compare the store file names that you get with those above.
17219 Likewise for a complete GuixSD system generation:
17222 guix gc -R `guix system build my-config.scm` | grep bash
17225 Lastly, to check which Bash running processes are using, you can use the
17226 @command{lsof} command:
17229 lsof | grep /gnu/store/.*bash
17233 @node Package Modules
17234 @section Package Modules
17236 From a programming viewpoint, the package definitions of the
17237 GNU distribution are provided by Guile modules in the @code{(gnu packages
17238 @dots{})} name space@footnote{Note that packages under the @code{(gnu
17239 packages @dots{})} module name space are not necessarily ``GNU
17240 packages''. This module naming scheme follows the usual Guile module
17241 naming convention: @code{gnu} means that these modules are distributed
17242 as part of the GNU system, and @code{packages} identifies modules that
17243 define packages.} (@pxref{Modules, Guile modules,, guile, GNU Guile
17244 Reference Manual}). For instance, the @code{(gnu packages emacs)}
17245 module exports a variable named @code{emacs}, which is bound to a
17246 @code{<package>} object (@pxref{Defining Packages}).
17248 The @code{(gnu packages @dots{})} module name space is
17249 automatically scanned for packages by the command-line tools. For
17250 instance, when running @code{guix package -i emacs}, all the @code{(gnu
17251 packages @dots{})} modules are scanned until one that exports a package
17252 object whose name is @code{emacs} is found. This package search
17253 facility is implemented in the @code{(gnu packages)} module.
17255 @cindex customization, of packages
17256 @cindex package module search path
17257 Users can store package definitions in modules with different
17258 names---e.g., @code{(my-packages emacs)}@footnote{Note that the file
17259 name and module name must match. For instance, the @code{(my-packages
17260 emacs)} module must be stored in a @file{my-packages/emacs.scm} file
17261 relative to the load path specified with @option{--load-path} or
17262 @code{GUIX_PACKAGE_PATH}. @xref{Modules and the File System,,,
17263 guile, GNU Guile Reference Manual}, for details.}. These package definitions
17264 will not be visible by default. Users can invoke commands such as
17265 @command{guix package} and @command{guix build} with the
17266 @code{-e} option so that they know where to find the package. Better
17267 yet, they can use the
17268 @code{-L} option of these commands to make those modules visible
17269 (@pxref{Invoking guix build, @code{--load-path}}), or define the
17270 @code{GUIX_PACKAGE_PATH} environment variable. This environment
17271 variable makes it easy to extend or customize the distribution and is
17272 honored by all the user interfaces.
17274 @defvr {Environment Variable} GUIX_PACKAGE_PATH
17275 This is a colon-separated list of directories to search for additional
17276 package modules. Directories listed in this variable take precedence
17277 over the own modules of the distribution.
17280 The distribution is fully @dfn{bootstrapped} and @dfn{self-contained}:
17281 each package is built based solely on other packages in the
17282 distribution. The root of this dependency graph is a small set of
17283 @dfn{bootstrap binaries}, provided by the @code{(gnu packages
17284 bootstrap)} module. For more information on bootstrapping,
17285 @pxref{Bootstrapping}.
17287 @node Packaging Guidelines
17288 @section Packaging Guidelines
17290 @cindex packages, creating
17291 The GNU distribution is nascent and may well lack some of your favorite
17292 packages. This section describes how you can help make the distribution
17293 grow. @xref{Contributing}, for additional information on how you can
17296 Free software packages are usually distributed in the form of
17297 @dfn{source code tarballs}---typically @file{tar.gz} files that contain
17298 all the source files. Adding a package to the distribution means
17299 essentially two things: adding a @dfn{recipe} that describes how to
17300 build the package, including a list of other packages required to build
17301 it, and adding @dfn{package metadata} along with that recipe, such as a
17302 description and licensing information.
17304 In Guix all this information is embodied in @dfn{package definitions}.
17305 Package definitions provide a high-level view of the package. They are
17306 written using the syntax of the Scheme programming language; in fact,
17307 for each package we define a variable bound to the package definition,
17308 and export that variable from a module (@pxref{Package Modules}).
17309 However, in-depth Scheme knowledge is @emph{not} a prerequisite for
17310 creating packages. For more information on package definitions,
17311 @pxref{Defining Packages}.
17313 Once a package definition is in place, stored in a file in the Guix
17314 source tree, it can be tested using the @command{guix build} command
17315 (@pxref{Invoking guix build}). For example, assuming the new package is
17316 called @code{gnew}, you may run this command from the Guix build tree
17317 (@pxref{Running Guix Before It Is Installed}):
17320 ./pre-inst-env guix build gnew --keep-failed
17323 Using @code{--keep-failed} makes it easier to debug build failures since
17324 it provides access to the failed build tree. Another useful
17325 command-line option when debugging is @code{--log-file}, to access the
17328 If the package is unknown to the @command{guix} command, it may be that
17329 the source file contains a syntax error, or lacks a @code{define-public}
17330 clause to export the package variable. To figure it out, you may load
17331 the module from Guile to get more information about the actual error:
17334 ./pre-inst-env guile -c '(use-modules (gnu packages gnew))'
17337 Once your package builds correctly, please send us a patch
17338 (@pxref{Contributing}). Well, if you need help, we will be happy to
17339 help you too. Once the patch is committed in the Guix repository, the
17340 new package automatically gets built on the supported platforms by
17341 @url{http://hydra.gnu.org/jobset/gnu/master, our continuous integration
17344 @cindex substituter
17345 Users can obtain the new package definition simply by running
17346 @command{guix pull} (@pxref{Invoking guix pull}). When
17347 @code{hydra.gnu.org} is done building the package, installing the
17348 package automatically downloads binaries from there
17349 (@pxref{Substitutes}). The only place where human intervention is
17350 needed is to review and apply the patch.
17354 * Software Freedom:: What may go into the distribution.
17355 * Package Naming:: What's in a name?
17356 * Version Numbers:: When the name is not enough.
17357 * Synopses and Descriptions:: Helping users find the right package.
17358 * Python Modules:: A touch of British comedy.
17359 * Perl Modules:: Little pearls.
17360 * Java Packages:: Coffee break.
17361 * Fonts:: Fond of fonts.
17364 @node Software Freedom
17365 @subsection Software Freedom
17367 @c Adapted from http://www.gnu.org/philosophy/philosophy.html.
17368 @cindex free software
17369 The GNU operating system has been developed so that users can have
17370 freedom in their computing. GNU is @dfn{free software}, meaning that
17371 users have the @url{http://www.gnu.org/philosophy/free-sw.html,four
17372 essential freedoms}: to run the program, to study and change the program
17373 in source code form, to redistribute exact copies, and to distribute
17374 modified versions. Packages found in the GNU distribution provide only
17375 software that conveys these four freedoms.
17377 In addition, the GNU distribution follow the
17378 @url{http://www.gnu.org/distros/free-system-distribution-guidelines.html,free
17379 software distribution guidelines}. Among other things, these guidelines
17380 reject non-free firmware, recommendations of non-free software, and
17381 discuss ways to deal with trademarks and patents.
17383 Some otherwise free upstream package sources contain a small and optional
17384 subset that violates the above guidelines, for instance because this subset
17385 is itself non-free code. When that happens, the offending items are removed
17386 with appropriate patches or code snippets in the @code{origin} form of the
17387 package (@pxref{Defining Packages}). This way, @code{guix
17388 build --source} returns the ``freed'' source rather than the unmodified
17392 @node Package Naming
17393 @subsection Package Naming
17395 @cindex package name
17396 A package has actually two names associated with it:
17397 First, there is the name of the @emph{Scheme variable}, the one following
17398 @code{define-public}. By this name, the package can be made known in the
17399 Scheme code, for instance as input to another package. Second, there is
17400 the string in the @code{name} field of a package definition. This name
17401 is used by package management commands such as
17402 @command{guix package} and @command{guix build}.
17404 Both are usually the same and correspond to the lowercase conversion of
17405 the project name chosen upstream, with underscores replaced with
17406 hyphens. For instance, GNUnet is available as @code{gnunet}, and
17407 SDL_net as @code{sdl-net}.
17409 We do not add @code{lib} prefixes for library packages, unless these are
17410 already part of the official project name. But @pxref{Python
17411 Modules} and @ref{Perl Modules} for special rules concerning modules for
17412 the Python and Perl languages.
17414 Font package names are handled differently, @pxref{Fonts}.
17417 @node Version Numbers
17418 @subsection Version Numbers
17420 @cindex package version
17421 We usually package only the latest version of a given free software
17422 project. But sometimes, for instance for incompatible library versions,
17423 two (or more) versions of the same package are needed. These require
17424 different Scheme variable names. We use the name as defined
17425 in @ref{Package Naming}
17426 for the most recent version; previous versions use the same name, suffixed
17427 by @code{-} and the smallest prefix of the version number that may
17428 distinguish the two versions.
17430 The name inside the package definition is the same for all versions of a
17431 package and does not contain any version number.
17433 For instance, the versions 2.24.20 and 3.9.12 of GTK+ may be packaged as follows:
17436 (define-public gtk+
17441 (define-public gtk+-2
17444 (version "2.24.20")
17447 If we also wanted GTK+ 3.8.2, this would be packaged as
17449 (define-public gtk+-3.8
17456 @c See <https://lists.gnu.org/archive/html/guix-devel/2016-01/msg00425.html>,
17457 @c for a discussion of what follows.
17458 @cindex version number, for VCS snapshots
17459 Occasionally, we package snapshots of upstream's version control system
17460 (VCS) instead of formal releases. This should remain exceptional,
17461 because it is up to upstream developers to clarify what the stable
17462 release is. Yet, it is sometimes necessary. So, what should we put in
17463 the @code{version} field?
17465 Clearly, we need to make the commit identifier of the VCS snapshot
17466 visible in the version string, but we also need to make sure that the
17467 version string is monotonically increasing so that @command{guix package
17468 --upgrade} can determine which version is newer. Since commit
17469 identifiers, notably with Git, are not monotonically increasing, we add
17470 a revision number that we increase each time we upgrade to a newer
17471 snapshot. The resulting version string looks like this:
17476 | | `-- upstream commit ID
17478 | `--- Guix package revision
17480 latest upstream version
17483 It is a good idea to strip commit identifiers in the @code{version}
17484 field to, say, 7 digits. It avoids an aesthetic annoyance (assuming
17485 aesthetics have a role to play here) as well as problems related to OS
17486 limits such as the maximum shebang length (127 bytes for the Linux
17487 kernel.) It is best to use the full commit identifiers in
17488 @code{origin}s, though, to avoid ambiguities. A typical package
17489 definition may look like this:
17493 (let ((commit "c3f29bc928d5900971f65965feaae59e1272a3f7")
17494 (revision "1")) ;Guix package revision
17496 (version (string-append "0.9-" revision "."
17497 (string-take commit 7)))
17500 (uri (git-reference
17501 (url "git://example.org/my-package.git")
17503 (sha256 (base32 "1mbikn@dots{}"))
17504 (file-name (string-append "my-package-" version
17510 @node Synopses and Descriptions
17511 @subsection Synopses and Descriptions
17513 @cindex package description
17514 @cindex package synopsis
17515 As we have seen before, each package in GNU@tie{}Guix includes a
17516 synopsis and a description (@pxref{Defining Packages}). Synopses and
17517 descriptions are important: They are what @command{guix package
17518 --search} searches, and a crucial piece of information to help users
17519 determine whether a given package suits their needs. Consequently,
17520 packagers should pay attention to what goes into them.
17522 Synopses must start with a capital letter and must not end with a
17523 period. They must not start with ``a'' or ``the'', which usually does
17524 not bring anything; for instance, prefer ``File-frobbing tool'' over ``A
17525 tool that frobs files''. The synopsis should say what the package
17526 is---e.g., ``Core GNU utilities (file, text, shell)''---or what it is
17527 used for---e.g., the synopsis for GNU@tie{}grep is ``Print lines
17528 matching a pattern''.
17530 Keep in mind that the synopsis must be meaningful for a very wide
17531 audience. For example, ``Manipulate alignments in the SAM format''
17532 might make sense for a seasoned bioinformatics researcher, but might be
17533 fairly unhelpful or even misleading to a non-specialized audience. It
17534 is a good idea to come up with a synopsis that gives an idea of the
17535 application domain of the package. In this example, this might give
17536 something like ``Manipulate nucleotide sequence alignments'', which
17537 hopefully gives the user a better idea of whether this is what they are
17540 Descriptions should take between five and ten lines. Use full
17541 sentences, and avoid using acronyms without first introducing them.
17542 Please avoid marketing phrases such as ``world-leading'',
17543 ``industrial-strength'', and ``next-generation'', and avoid superlatives
17544 like ``the most advanced''---they are not helpful to users looking for a
17545 package and may even sound suspicious. Instead, try to be factual,
17546 mentioning use cases and features.
17548 @cindex Texinfo markup, in package descriptions
17549 Descriptions can include Texinfo markup, which is useful to introduce
17550 ornaments such as @code{@@code} or @code{@@dfn}, bullet lists, or
17551 hyperlinks (@pxref{Overview,,, texinfo, GNU Texinfo}). However you
17552 should be careful when using some characters for example @samp{@@} and
17553 curly braces which are the basic special characters in Texinfo
17554 (@pxref{Special Characters,,, texinfo, GNU Texinfo}). User interfaces
17555 such as @command{guix package --show} take care of rendering it
17558 Synopses and descriptions are translated by volunteers
17559 @uref{http://translationproject.org/domain/guix-packages.html, at the
17560 Translation Project} so that as many users as possible can read them in
17561 their native language. User interfaces search them and display them in
17562 the language specified by the current locale.
17564 Translation is a lot of work so, as a packager, please pay even more
17565 attention to your synopses and descriptions as every change may entail
17566 additional work for translators. In order to help them, it is possible
17567 to make recommendations or instructions visible to them by inserting
17568 special comments like this (@pxref{xgettext Invocation,,, gettext, GNU
17572 ;; TRANSLATORS: "X11 resize-and-rotate" should not be translated.
17573 (description "ARandR is designed to provide a simple visual front end
17574 for the X11 resize-and-rotate (RandR) extension. @dots{}")
17578 @node Python Modules
17579 @subsection Python Modules
17582 We currently package Python 2 and Python 3, under the Scheme variable names
17583 @code{python-2} and @code{python} as explained in @ref{Version Numbers}.
17584 To avoid confusion and naming clashes with other programming languages, it
17585 seems desirable that the name of a package for a Python module contains
17586 the word @code{python}.
17588 Some modules are compatible with only one version of Python, others with both.
17589 If the package Foo compiles only with Python 3, we name it
17590 @code{python-foo}; if it compiles only with Python 2, we name it
17591 @code{python2-foo}. If it is compatible with both versions, we create two
17592 packages with the corresponding names.
17594 If a project already contains the word @code{python}, we drop this;
17595 for instance, the module python-dateutil is packaged under the names
17596 @code{python-dateutil} and @code{python2-dateutil}. If the project name
17597 starts with @code{py} (e.g. @code{pytz}), we keep it and prefix it as
17600 @subsubsection Specifying Dependencies
17601 @cindex inputs, for Python packages
17603 Dependency information for Python packages is usually available in the
17604 package source tree, with varying degrees of accuracy: in the
17605 @file{setup.py} file, in @file{requirements.txt}, or in @file{tox.ini}.
17607 Your mission, when writing a recipe for a Python package, is to map
17608 these dependencies to the appropriate type of ``input'' (@pxref{package
17609 Reference, inputs}). Although the @code{pypi} importer normally does a
17610 good job (@pxref{Invoking guix import}), you may want to check the
17611 following check list to determine which dependency goes where.
17616 We currently package Python 2 with @code{setuptools} and @code{pip}
17617 installed like Python 3.4 has per default. Thus you don't need to
17618 specify either of these as an input. @command{guix lint} will warn you
17622 Python dependencies required at run time go into
17623 @code{propagated-inputs}. They are typically defined with the
17624 @code{install_requires} keyword in @file{setup.py}, or in the
17625 @file{requirements.txt} file.
17628 Python packages required only at build time---e.g., those listed with
17629 the @code{setup_requires} keyword in @file{setup.py}---or only for
17630 testing---e.g., those in @code{tests_require}---go into
17631 @code{native-inputs}. The rationale is that (1) they do not need to be
17632 propagated because they are not needed at run time, and (2) in a
17633 cross-compilation context, it's the ``native'' input that we'd want.
17635 Examples are the @code{pytest}, @code{mock}, and @code{nose} test
17636 frameworks. Of course if any of these packages is also required at
17637 run-time, it needs to go to @code{propagated-inputs}.
17640 Anything that does not fall in the previous categories goes to
17641 @code{inputs}, for example programs or C libraries required for building
17642 Python packages containing C extensions.
17645 If a Python package has optional dependencies (@code{extras_require}),
17646 it is up to you to decide whether to add them or not, based on their
17647 usefulness/overhead ratio (@pxref{Submitting Patches, @command{guix
17654 @subsection Perl Modules
17657 Perl programs standing for themselves are named as any other package,
17658 using the lowercase upstream name.
17659 For Perl packages containing a single class, we use the lowercase class name,
17660 replace all occurrences of @code{::} by dashes and prepend the prefix
17662 So the class @code{XML::Parser} becomes @code{perl-xml-parser}.
17663 Modules containing several classes keep their lowercase upstream name and
17664 are also prepended by @code{perl-}. Such modules tend to have the word
17665 @code{perl} somewhere in their name, which gets dropped in favor of the
17666 prefix. For instance, @code{libwww-perl} becomes @code{perl-libwww}.
17669 @node Java Packages
17670 @subsection Java Packages
17673 Java programs standing for themselves are named as any other package,
17674 using the lowercase upstream name.
17676 To avoid confusion and naming clashes with other programming languages,
17677 it is desirable that the name of a package for a Java package is
17678 prefixed with @code{java-}. If a project already contains the word
17679 @code{java}, we drop this; for instance, the package @code{ngsjava} is
17680 packaged under the name @code{java-ngs}.
17682 For Java packages containing a single class or a small class hierarchy,
17683 we use the lowercase class name, replace all occurrences of @code{.} by
17684 dashes and prepend the prefix @code{java-}. So the class
17685 @code{apache.commons.cli} becomes package
17686 @code{java-apache-commons-cli}.
17693 For fonts that are in general not installed by a user for typesetting
17694 purposes, or that are distributed as part of a larger software package,
17695 we rely on the general packaging rules for software; for instance, this
17696 applies to the fonts delivered as part of the X.Org system or fonts that
17697 are part of TeX Live.
17699 To make it easier for a user to search for fonts, names for other packages
17700 containing only fonts are constructed as follows, independently of the
17701 upstream package name.
17703 The name of a package containing only one font family starts with
17704 @code{font-}; it is followed by the foundry name and a dash @code{-}
17705 if the foundry is known, and the font family name, in which spaces are
17706 replaced by dashes (and as usual, all upper case letters are transformed
17708 For example, the Gentium font family by SIL is packaged under the name
17709 @code{font-sil-gentium}.
17711 For a package containing several font families, the name of the collection
17712 is used in the place of the font family name.
17713 For instance, the Liberation fonts consist of three families,
17714 Liberation Sans, Liberation Serif and Liberation Mono.
17715 These could be packaged separately under the names
17716 @code{font-liberation-sans} and so on; but as they are distributed together
17717 under a common name, we prefer to package them together as
17718 @code{font-liberation}.
17720 In the case where several formats of the same font family or font collection
17721 are packaged separately, a short form of the format, prepended by a dash,
17722 is added to the package name. We use @code{-ttf} for TrueType fonts,
17723 @code{-otf} for OpenType fonts and @code{-type1} for PostScript Type 1
17728 @node Bootstrapping
17729 @section Bootstrapping
17731 @c Adapted from the ELS 2013 paper.
17733 @cindex bootstrapping
17735 Bootstrapping in our context refers to how the distribution gets built
17736 ``from nothing''. Remember that the build environment of a derivation
17737 contains nothing but its declared inputs (@pxref{Introduction}). So
17738 there's an obvious chicken-and-egg problem: how does the first package
17739 get built? How does the first compiler get compiled? Note that this is
17740 a question of interest only to the curious hacker, not to the regular
17741 user, so you can shamelessly skip this section if you consider yourself
17742 a ``regular user''.
17744 @cindex bootstrap binaries
17745 The GNU system is primarily made of C code, with libc at its core. The
17746 GNU build system itself assumes the availability of a Bourne shell and
17747 command-line tools provided by GNU Coreutils, Awk, Findutils, `sed', and
17748 `grep'. Furthermore, build programs---programs that run
17749 @code{./configure}, @code{make}, etc.---are written in Guile Scheme
17750 (@pxref{Derivations}). Consequently, to be able to build anything at
17751 all, from scratch, Guix relies on pre-built binaries of Guile, GCC,
17752 Binutils, libc, and the other packages mentioned above---the
17753 @dfn{bootstrap binaries}.
17755 These bootstrap binaries are ``taken for granted'', though we can also
17756 re-create them if needed (more on that later).
17758 @unnumberedsubsec Preparing to Use the Bootstrap Binaries
17760 @c As of Emacs 24.3, Info-mode displays the image, but since it's a
17761 @c large image, it's hard to scroll. Oh well.
17762 @image{images/bootstrap-graph,6in,,Dependency graph of the early bootstrap derivations}
17764 The figure above shows the very beginning of the dependency graph of the
17765 distribution, corresponding to the package definitions of the @code{(gnu
17766 packages bootstrap)} module. A similar figure can be generated with
17767 @command{guix graph} (@pxref{Invoking guix graph}), along the lines of:
17770 guix graph -t derivation \
17771 -e '(@@@@ (gnu packages bootstrap) %bootstrap-gcc)' \
17775 At this level of detail, things are
17776 slightly complex. First, Guile itself consists of an ELF executable,
17777 along with many source and compiled Scheme files that are dynamically
17778 loaded when it runs. This gets stored in the @file{guile-2.0.7.tar.xz}
17779 tarball shown in this graph. This tarball is part of Guix's ``source''
17780 distribution, and gets inserted into the store with @code{add-to-store}
17781 (@pxref{The Store}).
17783 But how do we write a derivation that unpacks this tarball and adds it
17784 to the store? To solve this problem, the @code{guile-bootstrap-2.0.drv}
17785 derivation---the first one that gets built---uses @code{bash} as its
17786 builder, which runs @code{build-bootstrap-guile.sh}, which in turn calls
17787 @code{tar} to unpack the tarball. Thus, @file{bash}, @file{tar},
17788 @file{xz}, and @file{mkdir} are statically-linked binaries, also part of
17789 the Guix source distribution, whose sole purpose is to allow the Guile
17790 tarball to be unpacked.
17792 Once @code{guile-bootstrap-2.0.drv} is built, we have a functioning
17793 Guile that can be used to run subsequent build programs. Its first task
17794 is to download tarballs containing the other pre-built binaries---this
17795 is what the @code{.tar.xz.drv} derivations do. Guix modules such as
17796 @code{ftp-client.scm} are used for this purpose. The
17797 @code{module-import.drv} derivations import those modules in a directory
17798 in the store, using the original layout. The
17799 @code{module-import-compiled.drv} derivations compile those modules, and
17800 write them in an output directory with the right layout. This
17801 corresponds to the @code{#:modules} argument of
17802 @code{build-expression->derivation} (@pxref{Derivations}).
17804 Finally, the various tarballs are unpacked by the
17805 derivations @code{gcc-bootstrap-0.drv}, @code{glibc-bootstrap-0.drv},
17806 etc., at which point we have a working C tool chain.
17809 @unnumberedsubsec Building the Build Tools
17811 Bootstrapping is complete when we have a full tool chain that does not
17812 depend on the pre-built bootstrap tools discussed above. This
17813 no-dependency requirement is verified by checking whether the files of
17814 the final tool chain contain references to the @file{/gnu/store}
17815 directories of the bootstrap inputs. The process that leads to this
17816 ``final'' tool chain is described by the package definitions found in
17817 the @code{(gnu packages commencement)} module.
17819 The @command{guix graph} command allows us to ``zoom out'' compared to
17820 the graph above, by looking at the level of package objects instead of
17821 individual derivations---remember that a package may translate to
17822 several derivations, typically one derivation to download its source,
17823 one to build the Guile modules it needs, and one to actually build the
17824 package from source. The command:
17827 guix graph -t bag \
17828 -e '(@@@@ (gnu packages commencement)
17829 glibc-final-with-bootstrap-bash)' | dot -Tps > t.ps
17833 produces the dependency graph leading to the ``final'' C
17834 library@footnote{You may notice the @code{glibc-intermediate} label,
17835 suggesting that it is not @emph{quite} final, but as a good
17836 approximation, we will consider it final.}, depicted below.
17838 @image{images/bootstrap-packages,6in,,Dependency graph of the early packages}
17840 @c See <http://lists.gnu.org/archive/html/gnu-system-discuss/2012-10/msg00000.html>.
17841 The first tool that gets built with the bootstrap binaries is
17842 GNU@tie{}Make---noted @code{make-boot0} above---which is a prerequisite
17843 for all the following packages. From there Findutils and Diffutils get
17846 Then come the first-stage Binutils and GCC, built as pseudo cross
17847 tools---i.e., with @code{--target} equal to @code{--host}. They are
17848 used to build libc. Thanks to this cross-build trick, this libc is
17849 guaranteed not to hold any reference to the initial tool chain.
17851 From there the final Binutils and GCC (not shown above) are built.
17853 from the final Binutils, and links programs against the just-built libc.
17854 This tool chain is used to build the other packages used by Guix and by
17855 the GNU Build System: Guile, Bash, Coreutils, etc.
17857 And voilà! At this point we have the complete set of build tools that
17858 the GNU Build System expects. These are in the @code{%final-inputs}
17859 variable of the @code{(gnu packages commencement)} module, and are
17860 implicitly used by any package that uses @code{gnu-build-system}
17861 (@pxref{Build Systems, @code{gnu-build-system}}).
17864 @unnumberedsubsec Building the Bootstrap Binaries
17866 @cindex bootstrap binaries
17867 Because the final tool chain does not depend on the bootstrap binaries,
17868 those rarely need to be updated. Nevertheless, it is useful to have an
17869 automated way to produce them, should an update occur, and this is what
17870 the @code{(gnu packages make-bootstrap)} module provides.
17872 The following command builds the tarballs containing the bootstrap
17873 binaries (Guile, Binutils, GCC, libc, and a tarball containing a mixture
17874 of Coreutils and other basic command-line tools):
17877 guix build bootstrap-tarballs
17880 The generated tarballs are those that should be referred to in the
17881 @code{(gnu packages bootstrap)} module mentioned at the beginning of
17884 Still here? Then perhaps by now you've started to wonder: when do we
17885 reach a fixed point? That is an interesting question! The answer is
17886 unknown, but if you would like to investigate further (and have
17887 significant computational and storage resources to do so), then let us
17891 @section Porting to a New Platform
17893 As discussed above, the GNU distribution is self-contained, and
17894 self-containment is achieved by relying on pre-built ``bootstrap
17895 binaries'' (@pxref{Bootstrapping}). These binaries are specific to an
17896 operating system kernel, CPU architecture, and application binary
17897 interface (ABI). Thus, to port the distribution to a platform that is
17898 not yet supported, one must build those bootstrap binaries, and update
17899 the @code{(gnu packages bootstrap)} module to use them on that platform.
17901 Fortunately, Guix can @emph{cross compile} those bootstrap binaries.
17902 When everything goes well, and assuming the GNU tool chain supports the
17903 target platform, this can be as simple as running a command like this
17907 guix build --target=armv5tel-linux-gnueabi bootstrap-tarballs
17910 For this to work, the @code{glibc-dynamic-linker} procedure in
17911 @code{(gnu packages bootstrap)} must be augmented to return the right
17912 file name for libc's dynamic linker on that platform; likewise,
17913 @code{system->linux-architecture} in @code{(gnu packages linux)} must be
17914 taught about the new platform.
17916 Once these are built, the @code{(gnu packages bootstrap)} module needs
17917 to be updated to refer to these binaries on the target platform. That
17918 is, the hashes and URLs of the bootstrap tarballs for the new platform
17919 must be added alongside those of the currently supported platforms. The
17920 bootstrap Guile tarball is treated specially: it is expected to be
17921 available locally, and @file{gnu/local.mk} has rules do download it for
17922 the supported architectures; a rule for the new platform must be added
17925 In practice, there may be some complications. First, it may be that the
17926 extended GNU triplet that specifies an ABI (like the @code{eabi} suffix
17927 above) is not recognized by all the GNU tools. Typically, glibc
17928 recognizes some of these, whereas GCC uses an extra @code{--with-abi}
17929 configure flag (see @code{gcc.scm} for examples of how to handle this).
17930 Second, some of the required packages could fail to build for that
17931 platform. Lastly, the generated binaries could be broken for some
17934 @c *********************************************************************
17935 @include contributing.texi
17937 @c *********************************************************************
17938 @node Acknowledgments
17939 @chapter Acknowledgments
17941 Guix is based on the @uref{http://nixos.org/nix/, Nix package manager},
17942 which was designed and
17943 implemented by Eelco Dolstra, with contributions from other people (see
17944 the @file{nix/AUTHORS} file in Guix.) Nix pioneered functional package
17945 management, and promoted unprecedented features, such as transactional
17946 package upgrades and rollbacks, per-user profiles, and referentially
17947 transparent build processes. Without this work, Guix would not exist.
17949 The Nix-based software distributions, Nixpkgs and NixOS, have also been
17950 an inspiration for Guix.
17952 GNU@tie{}Guix itself is a collective work with contributions from a
17953 number of people. See the @file{AUTHORS} file in Guix for more
17954 information on these fine people. The @file{THANKS} file lists people
17955 who have helped by reporting bugs, taking care of the infrastructure,
17956 providing artwork and themes, making suggestions, and more---thank you!
17959 @c *********************************************************************
17960 @node GNU Free Documentation License
17961 @appendix GNU Free Documentation License
17962 @cindex license, GNU Free Documentation License
17963 @include fdl-1.3.texi
17965 @c *********************************************************************
17966 @node Concept Index
17967 @unnumbered Concept Index
17970 @node Programming Index
17971 @unnumbered Programming Index
17972 @syncodeindex tp fn
17973 @syncodeindex vr fn
17978 @c Local Variables:
17979 @c ispell-local-dictionary: "american";