guix-devel/doc/contributing.texi

990 lines
40 KiB
Plaintext
Raw Normal View History

@node Contributing
@chapter Contributing
This project is a cooperative effort, and we need your help to make it
grow! Please get in touch with us on @email{guix-devel@@gnu.org} and
@code{#guix} on the Freenode IRC network. We welcome ideas, bug
reports, patches, and anything that may be helpful to the project. We
particularly welcome help on packaging (@pxref{Packaging Guidelines}).
@cindex code of conduct, of contributors
@cindex contributor covenant
We want to provide a warm, friendly, and harassment-free environment, so
that anyone can contribute to the best of their abilities. To this end
our project uses a ``Contributor Covenant'', which was adapted from
@url{http://contributor-covenant.org/}. You can find a local version in
the @file{CODE-OF-CONDUCT} file in the source tree.
Contributors are not required to use their legal name in patches and
on-line communication; they can use any name or pseudonym of their
choice.
@menu
* Building from Git:: The latest and greatest.
* Running Guix Before It Is Installed:: Hacker tricks.
* The Perfect Setup:: The right tools.
* Packaging Guidelines:: Growing the distribution.
* Coding Style:: Hygiene of the contributor.
* Submitting Patches:: Share your work.
@end menu
@node Building from Git
@section Building from Git
If you want to hack Guix itself, it is recommended to use the latest
version from the Git repository:
@example
git clone https://git.savannah.gnu.org/git/guix.git
@end example
The easiest way to set up a development environment for Guix is, of
course, by using Guix! The following command starts a new shell where
all the dependencies and appropriate environment variables are set up to
hack on Guix:
@example
guix environment guix --pure
@end example
@xref{Invoking guix environment}, for more information on that command.
If you are unable to use Guix when building Guix from a checkout, the
following are the required packages in addition to those mentioned in the
installation instructions (@pxref{Requirements}).
@itemize
@item @url{http://gnu.org/software/autoconf/, GNU Autoconf};
@item @url{http://gnu.org/software/automake/, GNU Automake};
@item @url{http://gnu.org/software/gettext/, GNU Gettext};
@item @url{http://gnu.org/software/texinfo/, GNU Texinfo};
@item @url{http://www.graphviz.org/, Graphviz};
@item @url{http://www.gnu.org/software/help2man/, GNU Help2man (optional)}.
@end itemize
On Guix, extra dependencies can be added by instead running @command{guix
environment} with @option{--ad-hoc}:
@example
guix environment guix --pure --ad-hoc help2man git strace
@end example
Run @command{./bootstrap} to generate the build system infrastructure
using Autoconf and Automake. If you get an error like this one:
@example
configure.ac:46: error: possibly undefined macro: PKG_CHECK_MODULES
@end example
@noindent
it probably means that Autoconf couldnt find @file{pkg.m4}, which is
provided by pkg-config. Make sure that @file{pkg.m4} is available. The
same holds for the @file{guile.m4} set of macros provided by Guile. For
instance, if you installed Automake in @file{/usr/local}, it wouldnt
look for @file{.m4} files in @file{/usr/share}. In that case, you have
to invoke the following command:
@example
export ACLOCAL_PATH=/usr/share/aclocal
@end example
@xref{Macro Search Path,,, automake, The GNU Automake Manual}, for
more information.
Then, run @command{./configure} as usual. Make sure to pass
@code{--localstatedir=@var{directory}} where @var{directory} is the
@code{localstatedir} value used by your current installation (@pxref{The
Store}, for information about this).
Finally, you have to invoke @code{make check} to run tests
(@pxref{Running the Test Suite}). If anything
fails, take a look at installation instructions (@pxref{Installation})
or send a message to the @email{guix-devel@@gnu.org, mailing list}.
@node Running Guix Before It Is Installed
@section Running Guix Before It Is Installed
In order to keep a sane working environment, you will find it useful to
test the changes made in your local source tree checkout without
actually installing them. So that you can distinguish between your
``end-user'' hat and your ``motley'' costume.
To that end, all the command-line tools can be used even if you have not
run @code{make install}. To do that, you first need to have an environment
with all the dependencies available (@pxref{Building from Git}), and then
simply prefix each command with
@command{./pre-inst-env} (the @file{pre-inst-env} script lives in the
top build tree of Guix; it is generated by @command{./configure}).
An example@footnote{The @option{-E} flag to
@command{sudo} guarantees that @code{GUILE_LOAD_PATH} is correctly set
such that @command{guix-daemon} and the tools it uses can find the Guile
modules they need.}:
@example
$ sudo -E ./pre-inst-env guix-daemon --build-users-group=guixbuild
$ ./pre-inst-env guix build hello
@end example
@noindent
Similarly, an example for a Guile session using the Guix modules:
@example
$ ./pre-inst-env guile -c '(use-modules (guix utils)) (pk (%current-system))'
;;; ("x86_64-linux")
@end example
@noindent
@cindex REPL
@cindex read-eval-print loop
@dots{} and for a REPL (@pxref{Using Guile Interactively,,, guile, Guile
Reference Manual}):
@example
$ ./pre-inst-env guile
scheme@@(guile-user)> ,use(guix)
scheme@@(guile-user)> ,use(gnu)
scheme@@(guile-user)> (define snakes
(fold-packages
(lambda (package lst)
(if (string-prefix? "python"
(package-name package))
(cons package lst)
lst))
'()))
scheme@@(guile-user)> (length snakes)
$1 = 361
@end example
The @command{pre-inst-env} script sets up all the environment variables
necessary to support this, including @env{PATH} and @env{GUILE_LOAD_PATH}.
Note that @command{./pre-inst-env guix pull} does @emph{not} upgrade the
local source tree; it simply updates the @file{~/.config/guix/current}
symlink (@pxref{Invoking guix pull}). Run @command{git pull} instead if
you want to upgrade your local source tree.
@node The Perfect Setup
@section The Perfect Setup
The Perfect Setup to hack on Guix is basically the perfect setup used
for Guile hacking (@pxref{Using Guile in Emacs,,, guile, Guile Reference
Manual}). First, you need more than an editor, you need
@url{http://www.gnu.org/software/emacs, Emacs}, empowered by the
wonderful @url{http://nongnu.org/geiser/, Geiser}. To set that up, run:
@example
guix package -i emacs guile emacs-geiser
@end example
Geiser allows for interactive and incremental development from within
Emacs: code compilation and evaluation from within buffers, access to
on-line documentation (docstrings), context-sensitive completion,
@kbd{M-.} to jump to an object definition, a REPL to try out your code,
and more (@pxref{Introduction,,, geiser, Geiser User Manual}). For
convenient Guix development, make sure to augment Guiles load path so
that it finds source files from your checkout:
@lisp
;; @r{Assuming the Guix checkout is in ~/src/guix.}
(with-eval-after-load 'geiser-guile
(add-to-list 'geiser-guile-load-path "~/src/guix"))
@end lisp
To actually edit the code, Emacs already has a neat Scheme mode. But in
addition to that, you must not miss
@url{http://www.emacswiki.org/emacs/ParEdit, Paredit}. It provides
facilities to directly operate on the syntax tree, such as raising an
s-expression or wrapping it, swallowing or rejecting the following
s-expression, etc.
@cindex code snippets
@cindex templates
@cindex reducing boilerplate
We also provide templates for common git commit messages and package
definitions in the @file{etc/snippets} directory. These templates can
be used with @url{http://joaotavora.github.io/yasnippet/, YASnippet} to
expand short trigger strings to interactive text snippets. You may want
to add the snippets directory to the @var{yas-snippet-dirs} variable in
Emacs.
@lisp
;; @r{Assuming the Guix checkout is in ~/src/guix.}
(with-eval-after-load 'yasnippet
(add-to-list 'yas-snippet-dirs "~/src/guix/etc/snippets"))
@end lisp
The commit message snippets depend on @url{https://magit.vc/, Magit} to
display staged files. When editing a commit message type @code{add}
followed by @kbd{TAB} to insert a commit message template for adding a
package; type @code{update} followed by @kbd{TAB} to insert a template
for updating a package; type @code{https} followed by @kbd{TAB} to
insert a template for changing the home page URI of a package to HTTPS.
The main snippet for @code{scheme-mode} is triggered by typing
@code{package...} followed by @kbd{TAB}. This snippet also inserts the
trigger string @code{origin...}, which can be expanded further. The
@code{origin} snippet in turn may insert other trigger strings ending on
@code{...}, which also can be expanded further.
@node Packaging Guidelines
@section Packaging Guidelines
@cindex packages, creating
The GNU distribution is nascent and may well lack some of your favorite
packages. This section describes how you can help make the distribution
grow.
Free software packages are usually distributed in the form of
@dfn{source code tarballs}---typically @file{tar.gz} files that contain
all the source files. Adding a package to the distribution means
essentially two things: adding a @dfn{recipe} that describes how to
build the package, including a list of other packages required to build
it, and adding @dfn{package metadata} along with that recipe, such as a
description and licensing information.
In Guix all this information is embodied in @dfn{package definitions}.
Package definitions provide a high-level view of the package. They are
written using the syntax of the Scheme programming language; in fact,
for each package we define a variable bound to the package definition,
and export that variable from a module (@pxref{Package Modules}).
However, in-depth Scheme knowledge is @emph{not} a prerequisite for
creating packages. For more information on package definitions,
@pxref{Defining Packages}.
Once a package definition is in place, stored in a file in the Guix
source tree, it can be tested using the @command{guix build} command
(@pxref{Invoking guix build}). For example, assuming the new package is
called @code{gnew}, you may run this command from the Guix build tree
(@pxref{Running Guix Before It Is Installed}):
@example
./pre-inst-env guix build gnew --keep-failed
@end example
Using @code{--keep-failed} makes it easier to debug build failures since
it provides access to the failed build tree. Another useful
command-line option when debugging is @code{--log-file}, to access the
build log.
If the package is unknown to the @command{guix} command, it may be that
the source file contains a syntax error, or lacks a @code{define-public}
clause to export the package variable. To figure it out, you may load
the module from Guile to get more information about the actual error:
@example
./pre-inst-env guile -c '(use-modules (gnu packages gnew))'
@end example
Once your package builds correctly, please send us a patch
(@pxref{Submitting Patches}). Well, if you need help, we will be happy to
help you too. Once the patch is committed in the Guix repository, the
new package automatically gets built on the supported platforms by
@url{@value{SUBSTITUTE-SERVER}, our continuous integration system}.
@cindex substituter
Users can obtain the new package definition simply by running
@command{guix pull} (@pxref{Invoking guix pull}). When
@code{@value{SUBSTITUTE-SERVER}} is done building the package, installing the
package automatically downloads binaries from there
(@pxref{Substitutes}). The only place where human intervention is
needed is to review and apply the patch.
@menu
* Software Freedom:: What may go into the distribution.
* Package Naming:: What's in a name?
* Version Numbers:: When the name is not enough.
* Synopses and Descriptions:: Helping users find the right package.
* Python Modules:: A touch of British comedy.
* Perl Modules:: Little pearls.
* Java Packages:: Coffee break.
* Fonts:: Fond of fonts.
@end menu
@node Software Freedom
@subsection Software Freedom
@c Adapted from http://www.gnu.org/philosophy/philosophy.html.
@cindex free software
The GNU operating system has been developed so that users can have
freedom in their computing. GNU is @dfn{free software}, meaning that
users have the @url{http://www.gnu.org/philosophy/free-sw.html,four
essential freedoms}: to run the program, to study and change the program
in source code form, to redistribute exact copies, and to distribute
modified versions. Packages found in the GNU distribution provide only
software that conveys these four freedoms.
In addition, the GNU distribution follow the
@url{http://www.gnu.org/distros/free-system-distribution-guidelines.html,free
software distribution guidelines}. Among other things, these guidelines
reject non-free firmware, recommendations of non-free software, and
discuss ways to deal with trademarks and patents.
Some otherwise free upstream package sources contain a small and optional
subset that violates the above guidelines, for instance because this subset
is itself non-free code. When that happens, the offending items are removed
with appropriate patches or code snippets in the @code{origin} form of the
package (@pxref{Defining Packages}). This way, @code{guix
build --source} returns the ``freed'' source rather than the unmodified
upstream source.
@node Package Naming
@subsection Package Naming
@cindex package name
A package has actually two names associated with it:
First, there is the name of the @emph{Scheme variable}, the one following
@code{define-public}. By this name, the package can be made known in the
Scheme code, for instance as input to another package. Second, there is
the string in the @code{name} field of a package definition. This name
is used by package management commands such as
@command{guix package} and @command{guix build}.
Both are usually the same and correspond to the lowercase conversion of
the project name chosen upstream, with underscores replaced with
hyphens. For instance, GNUnet is available as @code{gnunet}, and
SDL_net as @code{sdl-net}.
We do not add @code{lib} prefixes for library packages, unless these are
already part of the official project name. But @pxref{Python
Modules} and @ref{Perl Modules} for special rules concerning modules for
the Python and Perl languages.
Font package names are handled differently, @pxref{Fonts}.
@node Version Numbers
@subsection Version Numbers
@cindex package version
We usually package only the latest version of a given free software
project. But sometimes, for instance for incompatible library versions,
two (or more) versions of the same package are needed. These require
different Scheme variable names. We use the name as defined
in @ref{Package Naming}
for the most recent version; previous versions use the same name, suffixed
by @code{-} and the smallest prefix of the version number that may
distinguish the two versions.
The name inside the package definition is the same for all versions of a
package and does not contain any version number.
For instance, the versions 2.24.20 and 3.9.12 of GTK+ may be packaged as follows:
@lisp
(define-public gtk+
(package
(name "gtk+")
(version "3.9.12")
...))
(define-public gtk+-2
(package
(name "gtk+")
(version "2.24.20")
...))
@end lisp
If we also wanted GTK+ 3.8.2, this would be packaged as
@lisp
(define-public gtk+-3.8
(package
(name "gtk+")
(version "3.8.2")
...))
@end lisp
@c See <https://lists.gnu.org/archive/html/guix-devel/2016-01/msg00425.html>,
@c for a discussion of what follows.
@cindex version number, for VCS snapshots
Occasionally, we package snapshots of upstream's version control system
(VCS) instead of formal releases. This should remain exceptional,
because it is up to upstream developers to clarify what the stable
release is. Yet, it is sometimes necessary. So, what should we put in
the @code{version} field?
Clearly, we need to make the commit identifier of the VCS snapshot
visible in the version string, but we also need to make sure that the
version string is monotonically increasing so that @command{guix package
--upgrade} can determine which version is newer. Since commit
identifiers, notably with Git, are not monotonically increasing, we add
a revision number that we increase each time we upgrade to a newer
snapshot. The resulting version string looks like this:
@example
2.0.11-3.cabba9e
^ ^ ^
| | `-- upstream commit ID
| |
| `--- Guix package revision
|
latest upstream version
@end example
It is a good idea to strip commit identifiers in the @code{version}
field to, say, 7 digits. It avoids an aesthetic annoyance (assuming
aesthetics have a role to play here) as well as problems related to OS
limits such as the maximum shebang length (127 bytes for the Linux
kernel.) It is best to use the full commit identifiers in
@code{origin}s, though, to avoid ambiguities. A typical package
definition may look like this:
@lisp
(define my-package
(let ((commit "c3f29bc928d5900971f65965feaae59e1272a3f7")
(revision "1")) ;Guix package revision
(package
(version (git-version "0.9" revision commit))
(source (origin
(method git-fetch)
(uri (git-reference
(url "git://example.org/my-package.git")
(commit commit)))
(sha256 (base32 "1mbikn@dots{}"))
(file-name (git-file-name name version))))
;; @dots{}
)))
@end lisp
@node Synopses and Descriptions
@subsection Synopses and Descriptions
@cindex package description
@cindex package synopsis
As we have seen before, each package in GNU@tie{}Guix includes a
synopsis and a description (@pxref{Defining Packages}). Synopses and
descriptions are important: They are what @command{guix package
--search} searches, and a crucial piece of information to help users
determine whether a given package suits their needs. Consequently,
packagers should pay attention to what goes into them.
Synopses must start with a capital letter and must not end with a
period. They must not start with ``a'' or ``the'', which usually does
not bring anything; for instance, prefer ``File-frobbing tool'' over ``A
tool that frobs files''. The synopsis should say what the package
is---e.g., ``Core GNU utilities (file, text, shell)''---or what it is
used for---e.g., the synopsis for GNU@tie{}grep is ``Print lines
matching a pattern''.
Keep in mind that the synopsis must be meaningful for a very wide
audience. For example, ``Manipulate alignments in the SAM format''
might make sense for a seasoned bioinformatics researcher, but might be
fairly unhelpful or even misleading to a non-specialized audience. It
is a good idea to come up with a synopsis that gives an idea of the
application domain of the package. In this example, this might give
something like ``Manipulate nucleotide sequence alignments'', which
hopefully gives the user a better idea of whether this is what they are
looking for.
Descriptions should take between five and ten lines. Use full
sentences, and avoid using acronyms without first introducing them.
Please avoid marketing phrases such as ``world-leading'',
``industrial-strength'', and ``next-generation'', and avoid superlatives
like ``the most advanced''---they are not helpful to users looking for a
package and may even sound suspicious. Instead, try to be factual,
mentioning use cases and features.
@cindex Texinfo markup, in package descriptions
Descriptions can include Texinfo markup, which is useful to introduce
ornaments such as @code{@@code} or @code{@@dfn}, bullet lists, or
hyperlinks (@pxref{Overview,,, texinfo, GNU Texinfo}). However you
should be careful when using some characters for example @samp{@@} and
curly braces which are the basic special characters in Texinfo
(@pxref{Special Characters,,, texinfo, GNU Texinfo}). User interfaces
such as @command{guix package --show} take care of rendering it
appropriately.
Synopses and descriptions are translated by volunteers
@uref{http://translationproject.org/domain/guix-packages.html, at the
Translation Project} so that as many users as possible can read them in
their native language. User interfaces search them and display them in
the language specified by the current locale.
To allow @command{xgettext} to extract them as translatable strings,
synopses and descriptions @emph{must be literal strings}. This means
that you cannot use @code{string-append} or @code{format} to construct
these strings:
@lisp
(package
;; @dots{}
(synopsis "This is translatable")
(description (string-append "This is " "*not*" " translatable.")))
@end lisp
Translation is a lot of work so, as a packager, please pay even more
attention to your synopses and descriptions as every change may entail
additional work for translators. In order to help them, it is possible
to make recommendations or instructions visible to them by inserting
special comments like this (@pxref{xgettext Invocation,,, gettext, GNU
Gettext}):
@example
;; TRANSLATORS: "X11 resize-and-rotate" should not be translated.
(description "ARandR is designed to provide a simple visual front end
for the X11 resize-and-rotate (RandR) extension. @dots{}")
@end example
@node Python Modules
@subsection Python Modules
@cindex python
We currently package Python 2 and Python 3, under the Scheme variable names
@code{python-2} and @code{python} as explained in @ref{Version Numbers}.
To avoid confusion and naming clashes with other programming languages, it
seems desirable that the name of a package for a Python module contains
the word @code{python}.
Some modules are compatible with only one version of Python, others with both.
If the package Foo compiles only with Python 3, we name it
@code{python-foo}; if it compiles only with Python 2, we name it
@code{python2-foo}. If it is compatible with both versions, we create two
packages with the corresponding names.
If a project already contains the word @code{python}, we drop this;
for instance, the module python-dateutil is packaged under the names
@code{python-dateutil} and @code{python2-dateutil}. If the project name
starts with @code{py} (e.g.@: @code{pytz}), we keep it and prefix it as
described above.
@subsubsection Specifying Dependencies
@cindex inputs, for Python packages
Dependency information for Python packages is usually available in the
package source tree, with varying degrees of accuracy: in the
@file{setup.py} file, in @file{requirements.txt}, or in @file{tox.ini}.
Your mission, when writing a recipe for a Python package, is to map
these dependencies to the appropriate type of ``input'' (@pxref{package
Reference, inputs}). Although the @code{pypi} importer normally does a
good job (@pxref{Invoking guix import}), you may want to check the
following check list to determine which dependency goes where.
@itemize
@item
We currently package Python 2 with @code{setuptools} and @code{pip}
installed like Python 3.4 has per default. Thus you don't need to
specify either of these as an input. @command{guix lint} will warn you
if you do.
@item
Python dependencies required at run time go into
@code{propagated-inputs}. They are typically defined with the
@code{install_requires} keyword in @file{setup.py}, or in the
@file{requirements.txt} file.
@item
Python packages required only at build time---e.g., those listed with
the @code{setup_requires} keyword in @file{setup.py}---or only for
testing---e.g., those in @code{tests_require}---go into
@code{native-inputs}. The rationale is that (1) they do not need to be
propagated because they are not needed at run time, and (2) in a
cross-compilation context, it's the ``native'' input that we'd want.
Examples are the @code{pytest}, @code{mock}, and @code{nose} test
frameworks. Of course if any of these packages is also required at
run-time, it needs to go to @code{propagated-inputs}.
@item
Anything that does not fall in the previous categories goes to
@code{inputs}, for example programs or C libraries required for building
Python packages containing C extensions.
@item
If a Python package has optional dependencies (@code{extras_require}),
it is up to you to decide whether to add them or not, based on their
usefulness/overhead ratio (@pxref{Submitting Patches, @command{guix
size}}).
@end itemize
@node Perl Modules
@subsection Perl Modules
@cindex perl
Perl programs standing for themselves are named as any other package,
using the lowercase upstream name.
For Perl packages containing a single class, we use the lowercase class name,
replace all occurrences of @code{::} by dashes and prepend the prefix
@code{perl-}.
So the class @code{XML::Parser} becomes @code{perl-xml-parser}.
Modules containing several classes keep their lowercase upstream name and
are also prepended by @code{perl-}. Such modules tend to have the word
@code{perl} somewhere in their name, which gets dropped in favor of the
prefix. For instance, @code{libwww-perl} becomes @code{perl-libwww}.
@node Java Packages
@subsection Java Packages
@cindex java
Java programs standing for themselves are named as any other package,
using the lowercase upstream name.
To avoid confusion and naming clashes with other programming languages,
it is desirable that the name of a package for a Java package is
prefixed with @code{java-}. If a project already contains the word
@code{java}, we drop this; for instance, the package @code{ngsjava} is
packaged under the name @code{java-ngs}.
For Java packages containing a single class or a small class hierarchy,
we use the lowercase class name, replace all occurrences of @code{.} by
dashes and prepend the prefix @code{java-}. So the class
@code{apache.commons.cli} becomes package
@code{java-apache-commons-cli}.
@node Fonts
@subsection Fonts
@cindex fonts
For fonts that are in general not installed by a user for typesetting
purposes, or that are distributed as part of a larger software package,
we rely on the general packaging rules for software; for instance, this
applies to the fonts delivered as part of the X.Org system or fonts that
are part of TeX Live.
To make it easier for a user to search for fonts, names for other packages
containing only fonts are constructed as follows, independently of the
upstream package name.
The name of a package containing only one font family starts with
@code{font-}; it is followed by the foundry name and a dash @code{-}
if the foundry is known, and the font family name, in which spaces are
replaced by dashes (and as usual, all upper case letters are transformed
to lower case).
For example, the Gentium font family by SIL is packaged under the name
@code{font-sil-gentium}.
For a package containing several font families, the name of the collection
is used in the place of the font family name.
For instance, the Liberation fonts consist of three families,
Liberation Sans, Liberation Serif and Liberation Mono.
These could be packaged separately under the names
@code{font-liberation-sans} and so on; but as they are distributed together
under a common name, we prefer to package them together as
@code{font-liberation}.
In the case where several formats of the same font family or font collection
are packaged separately, a short form of the format, prepended by a dash,
is added to the package name. We use @code{-ttf} for TrueType fonts,
@code{-otf} for OpenType fonts and @code{-type1} for PostScript Type 1
fonts.
@node Coding Style
@section Coding Style
In general our code follows the GNU Coding Standards (@pxref{Top,,,
standards, GNU Coding Standards}). However, they do not say much about
Scheme, so here are some additional rules.
@menu
* Programming Paradigm:: How to compose your elements.
* Modules:: Where to store your code?
* Data Types and Pattern Matching:: Implementing data structures.
* Formatting Code:: Writing conventions.
@end menu
@node Programming Paradigm
@subsection Programming Paradigm
Scheme code in Guix is written in a purely functional style. One
exception is code that involves input/output, and procedures that
implement low-level concepts, such as the @code{memoize} procedure.
@node Modules
@subsection Modules
Guile modules that are meant to be used on the builder side must live in
the @code{(guix build @dots{})} name space. They must not refer to
other Guix or GNU modules. However, it is OK for a ``host-side'' module
to use a build-side module.
Modules that deal with the broader GNU system should be in the
@code{(gnu @dots{})} name space rather than @code{(guix @dots{})}.
@node Data Types and Pattern Matching
@subsection Data Types and Pattern Matching
The tendency in classical Lisp is to use lists to represent everything,
and then to browse them ``by hand'' using @code{car}, @code{cdr},
@code{cadr}, and co. There are several problems with that style,
notably the fact that it is hard to read, error-prone, and a hindrance
to proper type error reports.
Guix code should define appropriate data types (for instance, using
@code{define-record-type*}) rather than abuse lists. In addition, it
should use pattern matching, via Guiles @code{(ice-9 match)} module,
especially when matching lists.
@node Formatting Code
@subsection Formatting Code
@cindex formatting code
@cindex coding style
When writing Scheme code, we follow common wisdom among Scheme
programmers. In general, we follow the
@url{http://mumble.net/~campbell/scheme/style.txt, Riastradh's Lisp
Style Rules}. This document happens to describe the conventions mostly
used in Guiles code too. It is very thoughtful and well written, so
please do read it.
Some special forms introduced in Guix, such as the @code{substitute*}
macro, have special indentation rules. These are defined in the
@file{.dir-locals.el} file, which Emacs automatically uses. Also note
that Emacs-Guix provides @code{guix-devel-mode} mode that indents and
highlights Guix code properly (@pxref{Development,,, emacs-guix, The
Emacs-Guix Reference Manual}).
@cindex indentation, of code
@cindex formatting, of code
If you do not use Emacs, please make sure to let your editor knows these
rules. To automatically indent a package definition, you can also run:
@example
./etc/indent-code.el gnu/packages/@var{file}.scm @var{package}
@end example
@noindent
This automatically indents the definition of @var{package} in
@file{gnu/packages/@var{file}.scm} by running Emacs in batch mode. To
indent a whole file, omit the second argument:
@example
./etc/indent-code.el gnu/services/@var{file}.scm
@end example
@cindex Vim, Scheme code editing
If you are editing code with Vim, we recommend that you run @code{:set
autoindent} so that your code is automatically indented as you type.
Additionally,
@uref{https://www.vim.org/scripts/script.php?script_id=3998,
@code{paredit.vim}} may help you deal with all these parentheses.
We require all top-level procedures to carry a docstring. This
requirement can be relaxed for simple private procedures in the
@code{(guix build @dots{})} name space, though.
Procedures should not have more than four positional parameters. Use
keyword parameters for procedures that take more than four parameters.
@node Submitting Patches
@section Submitting Patches
Development is done using the Git distributed version control system.
Thus, access to the repository is not strictly necessary. We welcome
contributions in the form of patches as produced by @code{git
format-patch} sent to the @email{guix-patches@@gnu.org} mailing list.
This mailing list is backed by a Debbugs instance accessible at
@uref{https://bugs.gnu.org/guix-patches}, which allows us to keep track
of submissions. Each message sent to that mailing list gets a new
tracking number assigned; people can then follow up on the submission by
sending email to @code{@var{NNN}@@debbugs.gnu.org}, where @var{NNN} is
the tracking number (@pxref{Sending a Patch Series}).
Please write commit logs in the ChangeLog format (@pxref{Change Logs,,,
standards, GNU Coding Standards}); you can check the commit history for
examples.
Before submitting a patch that adds or modifies a package definition,
please run through this check list:
@enumerate
@item
If the authors of the packaged software provide a cryptographic
signature for the release tarball, make an effort to verify the
authenticity of the archive. For a detached GPG signature file this
would be done with the @code{gpg --verify} command.
@item
Take some time to provide an adequate synopsis and description for the
package. @xref{Synopses and Descriptions}, for some guidelines.
@item
Run @code{guix lint @var{package}}, where @var{package} is the
name of the new or modified package, and fix any errors it reports
(@pxref{Invoking guix lint}).
@item
Make sure the package builds on your platform, using @code{guix build
@var{package}}.
@item
We recommend you also try building the package on other supported
platforms. As you may not have access to actual hardware platforms, we
recommend using the @code{qemu-binfmt-service-type} to emulate them. In
order to enable it, add the following service to the list of services in
your @code{operating-system} configuration:
@lisp
(service qemu-binfmt-service-type
(qemu-binfmt-configuration
(platforms (lookup-qemu-platforms "arm" "aarch64" "mips64el"))
(guix-support? #t)))
@end lisp
Then reconfigure your system.
You can then build packages for different platforms by specifying the
@code{--system} option. For example, to build the "hello" package for
the armhf, aarch64, or mips64 architectures, you would run the following
commands, respectively:
@example
guix build --system=armhf-linux --rounds=2 hello
guix build --system=aarch64-linux --rounds=2 hello
guix build --system=mips64el-linux --rounds=2 hello
@end example
@item
@cindex bundling
Make sure the package does not use bundled copies of software already
available as separate packages.
Sometimes, packages include copies of the source code of their
dependencies as a convenience for users. However, as a distribution, we
want to make sure that such packages end up using the copy we already
have in the distribution, if there is one. This improves resource usage
(the dependency is built and stored only once), and allows the
distribution to make transverse changes such as applying security
updates for a given software package in a single place and have them
affect the whole system---something that bundled copies prevent.
@item
Take a look at the profile reported by @command{guix size}
(@pxref{Invoking guix size}). This will allow you to notice references
to other packages unwillingly retained. It may also help determine
whether to split the package (@pxref{Packages with Multiple Outputs}),
and which optional dependencies should be used. In particular, avoid adding
@code{texlive} as a dependency: because of its extreme size, use
@code{texlive-tiny} or @code{texlive-union} instead.
@item
For important changes, check that dependent package (if applicable) are
not affected by the change; @code{guix refresh --list-dependent
@var{package}} will help you do that (@pxref{Invoking guix refresh}).
@c See <https://lists.gnu.org/archive/html/guix-devel/2016-10/msg00933.html>.
@cindex branching strategy
@cindex rebuild scheduling strategy
Depending on the number of dependent packages and thus the amount of
rebuilding induced, commits go to different branches, along these lines:
@table @asis
@item 300 dependent packages or less
@code{master} branch (non-disruptive changes).
@item between 300 and 1,200 dependent packages
@code{staging} branch (non-disruptive changes). This branch is intended
to be merged in @code{master} every 3 weeks or so. Topical changes
(e.g., an update of the GNOME stack) can instead go to a specific branch
(say, @code{gnome-updates}).
@item more than 1,200 dependent packages
@code{core-updates} branch (may include major and potentially disruptive
changes). This branch is intended to be merged in @code{master} every
2.5 months or so.
@end table
All these branches are @uref{@value{SUBSTITUTE-SERVER},
tracked by our build farm} and merged into @code{master} once
everything has been successfully built. This allows us to fix issues
before they hit users, and to reduce the window during which pre-built
binaries are not available.
Generally, branches other than @code{master} are considered
@emph{frozen} if there has been a recent evaluation, or there is a
corresponding @code{-next} branch. Please ask on the mailing list or
IRC if unsure where to place a patch.
@c TODO: It would be good with badges on the website that tracks these
@c branches. Or maybe even a status page.
@item
@cindex determinism, of build processes
@cindex reproducible builds, checking
Check whether the package's build process is deterministic. This
typically means checking whether an independent build of the package
yields the exact same result that you obtained, bit for bit.
A simple way to do that is by building the same package several times in
a row on your machine (@pxref{Invoking guix build}):
@example
guix build --rounds=2 my-package
@end example
This is enough to catch a class of common non-determinism issues, such
as timestamps or randomly-generated output in the build result.
Another option is to use @command{guix challenge} (@pxref{Invoking guix
challenge}). You may run it once the package has been committed and
built by @code{@value{SUBSTITUTE-SERVER}} to check whether it obtains the same
result as you did. Better yet: Find another machine that can build it
and run @command{guix publish}. Since the remote build machine is
likely different from yours, this can catch non-determinism issues
related to the hardware---e.g., use of different instruction set
extensions---or to the operating system kernel---e.g., reliance on
@code{uname} or @file{/proc} files.
@item
When writing documentation, please use gender-neutral wording when
referring to people, such as
@uref{https://en.wikipedia.org/wiki/Singular_they, singular
``they''@comma{} ``their''@comma{} ``them''}, and so forth.
@item
Verify that your patch contains only one set of related changes.
Bundling unrelated changes together makes reviewing harder and slower.
Examples of unrelated changes include the addition of several packages,
or a package update along with fixes to that package.
@item
Please follow our code formatting rules, possibly running the
@command{etc/indent-code.el} script to do that automatically for you
(@pxref{Formatting Code}).
@item
When possible, use mirrors in the source URL (@pxref{Invoking guix download}).
Use reliable URLs, not generated ones. For instance, GitHub archives are not
necessarily identical from one generation to the next, so in this case it's
often better to clone the repository. Don't use the @command{name} field in
the URL: it is not very useful and if the name changes, the URL will probably
be wrong.
@end enumerate
When posting a patch to the mailing list, use @samp{[PATCH] @dots{}} as
a subject. You may use your email client or the @command{git
send-email} command (@pxref{Sending a Patch Series}). We prefer to get
patches in plain text messages, either inline or as MIME attachments.
You are advised to pay attention if your email client changes anything
like line breaks or indentation which could potentially break the
patches.
When a bug is resolved, please close the thread by sending an email to
@email{@var{NNN}-done@@debbugs.gnu.org}.
@unnumberedsubsec Sending a Patch Series
@anchor{Sending a Patch Series}
@cindex patch series
@cindex @code{git send-email}
@cindex @code{git-send-email}
When sending a patch series (e.g., using @code{git send-email}), please
first send one message to @email{guix-patches@@gnu.org}, and then send
subsequent patches to @email{@var{NNN}@@debbugs.gnu.org} to make sure
they are kept together. See
@uref{https://debbugs.gnu.org/Advanced.html, the Debbugs documentation}
for more information. You can install @command{git send-email} with
@command{guix install git:send-email}.
@c Debbugs bug: https://debbugs.gnu.org/db/15/15361.html