gnulib/doc/gnulib-tool.texi

@node Invoking gnulib-tool
@chapter Invoking gnulib-tool

@c Copyright (C) 2005-2017 Free Software Foundation, Inc.

@c Permission is granted to copy, distribute and/or modify this document
@c under the terms of the GNU Free Documentation License, Version 1.3 or
@c any later version published by the Free Software Foundation; with no
@c Invariant Sections, no Front-Cover Texts, and no Back-Cover
@c Texts.  A copy of the license is included in the ``GNU Free
@c Documentation License'' file as part of this distribution.

@pindex gnulib-tool
@cindex invoking @command{gnulib-tool}

The @command{gnulib-tool} command is the recommended way to import
Gnulib modules.  It is possible to borrow Gnulib modules in a package
without using @command{gnulib-tool}, relying only on the
meta-information stored in the @file{modules/*} files, but with a
growing number of modules this becomes tedious.  @command{gnulib-tool}
simplifies the management of source files, @file{Makefile.am}s and
@file{configure.ac} in packages incorporating Gnulib modules.

@file{gnulib-tool} is not installed in a standard directory that is
contained in the @code{PATH} variable.  It needs to be run directly in
the directory that contains the Gnulib source code.  You can do this
either by specifying the absolute filename of @file{gnulib-tool}, or
you can also use a symbolic link from a place inside your @code{PATH}
to the @file{gnulib-tool} file of your preferred and most up-to-date
Gnulib checkout, like this:
@smallexample
$ ln -s $HOME/gnu/src/gnulib.git/gnulib-tool $HOME/bin/gnulib-tool
@end smallexample

Run @samp{gnulib-tool --help} for information.  To get familiar with
@command{gnulib-tool} without affecting your sources, you can also try
some commands with the option @samp{--dry-run}; then
@code{gnulib-tool} will only report which actions it would perform in
a real run without changing anything.

@menu
* Which modules?::              Determining the needed set of Gnulib modules
* Initial import::              First import of Gnulib modules.
* Modified imports::            Changing the import specification.
* Simple update::               Tracking Gnulib development.
* Source changes::              Impact of Gnulib on your source files.
* Multiple instances::          Using Gnulib for both a library and a binary
* gettextize and autopoint::    Caveat: @code{gettextize} and @code{autopoint} users!
* Localization::                Handling Gnulib's own message translations.
* VCS Issues::                  Integration with Version Control Systems.
* Unit tests::                  Bundling the unit tests of the Gnulib modules.
* Conditional dependencies::    Avoiding unnecessary checks and compilations.
@end menu


@node Which modules?
@section Finding modules
@cindex Finding modules

There are three ways of finding the names of Gnulib modules that you can use
in your package:

@itemize
@item
You have the complete module list, sorted according to categories, in
@url{http://www.gnu.org/software/gnulib/MODULES.html}.

@item
If you are looking for a particular POSIX header or function replacement,
look in the chapters @ref{Header File Substitutes} and
@ref{Function Substitutes}.  For headers and functions that are provided by
Glibc but not standardized by POSIX, look in the chapters
@ref{Glibc Header File Substitutes} and @ref{Glibc Function Substitutes}.

@item
If you have already found the source file in Gnulib and are looking for the
module that contains this source file, you can use the command
@samp{gnulib-tool --find @var{filename}}.
@end itemize


@node Initial import
@section Initial import
@cindex initial import

Gnulib assumes that your project uses Autoconf.  When using Gnulib, you
will need to have Autoconf among your build tools.

Unless you use @command{gnulib-tool}'s @option{--gnu-make} option,
Gnulib also assumes that your project uses Automake at least in a
subdirectory of your project.  While the use of Automake in your
project's top level directory is an easy way to fulfil the Makefile
conventions of the GNU coding standards, Gnulib does not require it.

Invoking @samp{gnulib-tool --import} will copy source files, create a
@file{Makefile.am} to build them, generate a file @file{gnulib-comp.m4} with
Autoconf M4 macro declarations used by @file{configure.ac}, and generate
a file @file{gnulib-cache.m4} containing the cached specification of how
Gnulib is used.

Our example will be a library that uses Autoconf, Automake and
Libtool.  It calls @code{strdup}, and you wish to use gnulib to make
the package portable to C99 and C11 (which don't have @code{strdup}).

@example
~/src/libfoo$ gnulib-tool --import strdup
Module list with included dependencies:
  absolute-header
  extensions
  strdup
  string
File list:
  lib/dummy.c
  lib/strdup.c
  lib/string.in.h
  m4/absolute-header.m4
  m4/extensions.m4
  m4/gnulib-common.m4
  m4/strdup.m4
  m4/string_h.m4
Creating directory ./lib
Creating directory ./m4
Copying file lib/dummy.c
Copying file lib/strdup.c
Copying file lib/string.in.h
Copying file m4/absolute-header.m4
Copying file m4/extensions.m4
Copying file m4/gnulib-common.m4
Copying file m4/gnulib-tool.m4
Copying file m4/strdup.m4
Copying file m4/string_h.m4
Creating lib/Makefile.am
Creating m4/gnulib-cache.m4
Creating m4/gnulib-comp.m4
Finished.

You may need to add #include directives for the following .h files.
  #include <string.h>

Don't forget to
  - add "lib/Makefile" to AC_CONFIG_FILES in ./configure.ac,
  - mention "lib" in SUBDIRS in Makefile.am,
  - mention "-I m4" in ACLOCAL_AMFLAGS in Makefile.am,
  - invoke gl_EARLY in ./configure.ac, right after AC_PROG_CC,
  - invoke gl_INIT in ./configure.ac.
~/src/libfoo$
@end example

By default, the source code is copied into @file{lib/} and the M4
macros in @file{m4/}.  You can override these paths by using
@code{--source-base=DIRECTORY} and @code{--m4-base=DIRECTORY}.  Some
modules also provide other files necessary for building.  These files
are copied into the directory specified by @samp{AC_CONFIG_AUX_DIR} in
@file{configure.ac} or by the @code{--aux-dir=DIRECTORY} option.  If
neither is specified, the current directory is assumed.

@code{gnulib-tool} can make symbolic links instead of copying the
source files.  The option to specify for this is @samp{--symlink}, or
@samp{-s} for short.  This can be useful to save a few kilobytes of disk
space.  But it is likely to introduce bugs when @code{gnulib} is updated;
it is more reliable to use @samp{gnulib-tool --update} (see below)
to update to newer versions of @code{gnulib}.  Furthermore it requires
extra effort to create self-contained tarballs, and it may disturb some
mechanism the maintainer applies to the sources.  For these reasons,
this option is generally discouraged.

@code{gnulib-tool} will overwrite any pre-existing files, in
particular @file{Makefile.am}.  It is also possible to separate the
generated @file{Makefile.am} content (for building the gnulib library)
into a separate file, say @file{gnulib.mk}, that can be included by your
handwritten @file{Makefile.am}, but this is a more advanced use of
@code{gnulib-tool}.

Consequently, it is a good idea to choose directories that are not
already used by your projects, to separate gnulib imported files from
your own files.  This approach is also useful if you want to avoid
conflicts between other tools (e.g., @code{gettextize} that also copy
M4 files into your package.  Simon Josefsson successfully uses a source
base of @file{gl/}, and a M4 base of @file{gl/m4/}, in several
packages.

After the @samp{--import} option on the command line comes the list of
Gnulib modules that you want to incorporate in your package.  The names
of the modules coincide with the filenames in Gnulib's @file{modules/}
directory.

Some Gnulib modules depend on other Gnulib modules.  @code{gnulib-tool}
will automatically add the needed modules as well; you need not list
them explicitly.  @code{gnulib-tool} will also memorize which dependent
modules it has added, so that when someday a dependency is dropped, the
implicitly added module is dropped as well (unless you have explicitly
requested that module).

If you want to cut a dependency, i.e., not add a module although one of
your requested modules depends on it, you may use the option
@samp{--avoid=@var{module}} to do so.  Multiple uses of this option are
possible.  Of course, you will then need to implement the same interface
as the removed module.

A few manual steps are required to finish the initial import.
@code{gnulib-tool} printed a summary of these steps.

First, you must ensure Autoconf can find the macro definitions in
@file{gnulib-comp.m4}.  Use the @code{ACLOCAL_AMFLAGS} specifier in
your top-level @file{Makefile.am} file, as in:

@example
ACLOCAL_AMFLAGS = -I m4
@end example

You are now ready to call the M4 macros in @code{gnulib-comp.m4} from
@file{configure.ac}.  The macro @code{gl_EARLY} must be called as soon
as possible after verifying that the C compiler is working.
Typically, this is immediately after @code{AC_PROG_CC}, as in:

@example
...
AC_PROG_CC
gl_EARLY
...
@end example

If you are using @code{AC_PROG_CC_STDC}, the macro @code{gl_EARLY} must
be called after it, like this:

@example
...
AC_PROG_CC
AC_PROG_CC_STDC
gl_EARLY
...
@end example

The core part of the gnulib checks are done by the macro
@code{gl_INIT}.  Place it further down in the file, typically where
you normally check for header files or functions.  It must come after
other checks which may affect the compiler invocation, such as
@code{AC_MINIX}.  For example:

@example
...
# For gnulib.
gl_INIT
...
@end example

@code{gl_INIT} will in turn call the macros related with the
gnulib functions, be it specific gnulib macros, like @code{gl_FUNC_ALLOCA}
or Autoconf or Automake macros like @code{AC_FUNC_ALLOCA} or
@code{AM_FUNC_GETLINE}.  So there is no need to call those macros yourself
when you use the corresponding gnulib modules.

You must also make sure that the gnulib library is built.  Add the
@code{Makefile} in the gnulib source base directory to
@code{AC_CONFIG_FILES}, as in:

@example
AC_CONFIG_FILES(... lib/Makefile ...)
@end example

You must also make sure that @code{make} will recurse into the gnulib
directory.  To achieve this, add the gnulib source base directory to a
@code{SUBDIRS} Makefile.am statement, as in:

@example
SUBDIRS = lib
@end example

or if you, more likely, already have a few entries in @code{SUBDIRS},
you can add something like:

@example
SUBDIRS += lib
@end example

Finally, you have to add compiler and linker flags in the appropriate
source directories, so that you can make use of the gnulib library.
Since some modules (@samp{getopt}, for example) may copy files into
the build directory, @file{top_builddir/lib} is needed as well
as @file{top_srcdir/lib}.  For example:

@example
...
AM_CPPFLAGS = -I$(top_builddir)/lib -I$(top_srcdir)/lib
...
LDADD = lib/libgnu.a
...
@end example

Don't forget to @code{#include} the various header files.  In this
example, you would need to make sure that @samp{#include <string.h>}
is evaluated when compiling all source code files, that want to make
use of @code{strdup}.

In the usual case where Autoconf is creating a @file{config.h} file,
you should include @file{config.h} first, before any other include
file.  That way, for example, if @file{config.h} defines
@samp{restrict} to be the empty string on a non-C99 host, or a macro
like @samp{_FILE_OFFSET_BITS} that affects the layout of data
structures, the definition is consistent for all include files.
Also, on some platforms macros like @samp{_FILE_OFFSET_BITS} and
@samp{_GNU_SOURCE} may be ineffective, or may have only a limited
effect, if defined after the first system header file is included.

Finally, note that you cannot use @code{AC_LIBOBJ} or
@code{AC_REPLACE_FUNCS} in your @file{configure.ac} and expect the
resulting object files to be automatically added to @file{lib/libgnu.a}.
This is because your @code{AC_LIBOBJ} and @code{AC_REPLACE_FUNCS} invocations
from @file{configure.ac} augment a variable @code{@@LIBOBJS@@} (and/or
@code{@@LTLIBOBJS@@} if using Libtool), whereas @file{lib/libgnu.a}
is built from the contents of a different variable, usually
@code{@@gl_LIBOBJS@@} (or @code{@@gl_LTLIBOBJS@@} if using Libtool).


@node Modified imports
@section Modified imports

You can at any moment decide to use Gnulib differently than the last time.

There are two ways to change how Gnulib is used.  Which one you'll use,
depends on where you keep track of options and module names that you pass
to @code{gnulib-tool}.

@itemize @bullet
@item
If you store the options and module names in a file under your own
control, such as @file{autogen.sh}, @file{bootstrap},
@file{bootstrap.conf}, or similar, simply invoke @command{gnulib-tool}
again, with modified options and more or fewer module names.

@item
@code{gnulib-tool} remembers which modules were used last time.  If you
want to rely on @code{gnulib-tool}'s own memory of the last used
options and module names, you can use the commands
@command{gnulib-tool --add-import} and
@command{gnulib-tool --remove-import}.

So, if you only want to use more Gnulib modules, simply invoke
@command{gnulib-tool --add-import @var{new-modules}}.  The list of
modules that you pass after @samp{--add-import} is @emph{added} to the
previous list of modules.

Similarly, if you want to use fewer Gnulib modules, simply invoke
@command{gnulib-tool --remove-import @var{unneeded-modules}}.  The list
of modules that you pass after @samp{--remove-import} is @emph{removed}
from the previous list of modules.  Note that if a module is then still
needed as dependency of other modules, it will be used nevertheless.
If you want to @emph{really} not use a module any more, regardless of
whether other modules may need it, you need to use the @samp{--avoid}
option.

For other changes, such as different choices of @samp{--lib},
@samp{--source-base} or @samp{--aux-dir}, the normal way is to
modify manually the file @file{gnulib-cache.m4} in the M4 macros
directory, then launch @samp{gnulib-tool --add-import}.

The only change for which this doesn't work is a change of the
@samp{--m4-base} directory.  Because, when you pass a different value of
@samp{--m4-base}, @code{gnulib-tool} will not find the previous
@file{gnulib-cache.m4} file any more.  A possible solution is to
manually copy the @file{gnulib-cache.m4} into the new M4 macro directory.

In the @file{gnulib-cache.m4} file, the macros have the following meaning:
@table @code
@item gl_MODULES
The argument is a space separated list of the requested modules, not including
dependencies.

@item gl_AVOID
The argument is a space separated list of modules that should not be used,
even if they occur as dependencies.  Corresponds to the @samp{--avoid}
command line argument.

@item gl_SOURCE_BASE
The argument is the relative file name of the directory containing the gnulib
source files (mostly *.c and *.h files).  Corresponds to the
@samp{--source-base} command line argument.

@item gl_M4_BASE
The argument is the relative file name of the directory containing the gnulib
M4 macros (*.m4 files).  Corresponds to the @samp{--m4-base} command line
argument.

@item gl_TESTS_BASE
The argument is the relative file name of the directory containing the gnulib
unit test files.  Corresponds to the @samp{--tests-base} command line argument.

@item gl_LIB
The argument is the name of the library to be created.  Corresponds to the
@samp{--lib} command line argument.

@item gl_LGPL
The presence of this macro without arguments corresponds to the @samp{--lgpl}
command line argument.  The presence of this macro with an argument (whose
value must be 2 or 3) corresponds to the @samp{--lgpl=@var{arg}} command line
argument.

@item gl_LIBTOOL
The presence of this macro corresponds to the @samp{--libtool} command line
argument and to the absence of the @samp{--no-libtool} command line argument.
It takes no arguments.

@item gl_MACRO_PREFIX
The argument is the prefix to use for macros in the @file{gnulib-comp.m4}
file.  Corresponds to the @samp{--macro-prefix} command line argument.
@end table

@end itemize

@node Simple update
@section Simple update

When you want to update to a more recent version of Gnulib, without
changing the list of modules or other parameters, a simple call
does it:

@smallexample
$ gnulib-tool --add-import
@end smallexample

@noindent
This will create, update or remove files, as needed.

Note: From time to time, changes are made in Gnulib that are not backward
compatible.  When updating to a more recent Gnulib, you should consult
Gnulib's @file{NEWS} file to check whether the incompatible changes affect
your project.


@node Source changes
@section Changing your sources for use with Gnulib

Gnulib contains some header file overrides.  This means that when building
on systems with deficient header files in @file{/usr/include/}, it may create
files named @file{string.h}, @file{stdlib.h}, @file{stdint.h} or similar in
the build directory.  In the other source directories of your package you
will usually pass @samp{-I} options to the compiler, so that these Gnulib
substitutes are visible and take precedence over the files in
@file{/usr/include/}.

These Gnulib substitute header files rely on @file{<config.h>} being
already included.  Furthermore @file{<config.h>} must be the first include
in every compilation unit.  This means that to @emph{all your source files}
and likely also to @emph{all your tests source files} you need to add an
@samp{#include <config.h>} at the top.  Which source files are affected?
Exactly those whose compilation includes a @samp{-I} option that refers to
the Gnulib library directory.

This is annoying, but inevitable: On many systems, @file{<config.h>} is
used to set system dependent flags (such as @code{_GNU_SOURCE} on GNU systems),
and these flags have no effect after any system header file has been included.


@node Multiple instances
@section Using Gnulib for both a library and a binary

Your project might build both a library and some accompanying binaries
in the same source tree. In that case you might want to use different
modules for the library than for the binaries. Typically the binaries
might want to make use of @code{getopt-posix} or @code{version-etc},
while the library wants to stay clear of these modules for technical
or licensing reasons.

Let's assume that your project contains a @file{lib} directory where
the source of the library resides and a @file{src} directory for the
sources of the binaries as follows.

@example
.
|-- configure.ac
|-- lib
|   |-- foo.c
|   `-- Makefile.am
|-- Makefile.am
`-- src
    |-- bar.c
    `-- Makefile.am
@end example

You can now add two instances of Gnulib to your project in separate
source trees:

@example
~/src/libfoo$ gnulib-tool --import --lib=libgnu --source-base=gnulib \
              --m4-base=gnulib/m4 --macro-prefix=gl strndup
~/src/libfoo$ gnulib-tool --import --lib=libgnutools \
              --source-base=src/gnulib --m4-base=src/gnulib/m4 \
              --macro-prefix=gl_tools getopt-gnu
@end example

The first one will import the module @code{strndup} in @file{gnulib}
and the second one will import @code{getopt-gnu} in @file{src/gnulib}
and you will end up with the following source tree (many files omitted
in the interest of brevity):

@example
.
|-- configure.ac
|-- gnulib
|   |-- m4
|   |-- strndup.c
|-- lib
|   |-- foo.c
|   `-- Makefile.am
|-- Makefile.am
`-- src
    |-- bar.c
    |-- gnulib
    |   |-- getopt.c
    |   |-- getopt.in.h
    |   |-- m4
    `-- Makefile.am
@end example

Integration with your code is basically the same as outlined in
@ref{Initial import} with the one exception that you have to add both
the macro @code{gl_EARLY} and the macro @code{gl_tools_EARLY} to your
@file{configure.ac} (and of course also both macros @code{gl_INIT} and
@code{gl_tools_INIT}). Obviously the name of the second macro is
dependent on the value of the @option{--macro-prefix} option in your
@command{gnulib-tool} invocation.

@example
...
AC_PROG_CC
gl_EARLY
gl_tools_EARLY
...
# For gnulib.
gl_INIT
gl_tools_INIT
...
@end example

Also as outlined in @ref{Initial import} you will have to add compiler
and linker flags. For the library you might have to add something
along the line of the following to your @file{Makefile.am}:

@example
...
AM_CPPFLAGS = -I$(top_srcdir)/gnulib -I$(top_builddir)/gnulib
...
libfoo_la_LIBADD = $(top_builddir)/gnulib/libgnu.la
...
@end example

Correspondingly for the binary you will have to add something along
the lines of to the following:

@example
...
AM_CPPFLAGS = -I$(top_srcdir)/src/gnulib -I$(top_builddir)/src/gnulib
...
LIBADD = $(top_builddir)/src/gnulib/libgnutools.la
...
@end example

The name of the library that you have pass in the linker option
depends on the @option{--lib} option in @command{gnulib-tool}
invocation.

@node gettextize and autopoint
@section Caveat: @code{gettextize} and @code{autopoint} users

@cindex gettextize, caveat
@cindex autopoint, caveat
The programs @code{gettextize} and @code{autopoint}, part of
GNU @code{gettext}, import or update the internationalization infrastructure.
Some of this infrastructure, namely ca.@: 20 Autoconf macro files and the
@file{config.rpath} file, is also contained in Gnulib and may be imported
by @code{gnulib-tool}.  The use of @code{gettextize} or @code{autopoint}
will therefore overwrite some of the files that @code{gnulib-tool} has
imported, and vice versa.

Avoiding to use @code{gettextize} (manually, as package maintainer) or
@code{autopoint} (as part of a script like @code{autoreconf} or
@code{autogen.sh}) is not the solution: These programs also import the
infrastructure in the @file{po/} and optionally in the @file{intl/} directory.

The copies of the conflicting files in Gnulib are more up-to-date than
the copies brought in by @code{gettextize} and @code{autopoint}.  When a
new @code{gettext} release is made, the copies of the files in Gnulib will
be updated immediately.

The choice of which version of gettext to require depends on the needs
of your package.  For a package that wants to comply to GNU Coding
Standards, the steps are:

@enumerate
@item
When you run @code{gettextize}, always use the @code{gettextize} from the
matching GNU gettext release.  For the most recent Gnulib checkout, this is
the newest release found on @url{http://ftp.gnu.org/gnu/gettext/}.  For an
older Gnulib snapshot, it is the release that was the most recent release
at the time the Gnulib snapshot was taken.

@item
After running @code{gettextize}, invoke @code{gnulib-tool} and import
the @code{gettext} module.  Also, copy the latest version of gnulib's
@file{build-aux/po/Makefile.in.in} to your @file{po/} directory (this
is done for you if you use gnulib's @file{bootstrap} script).

@item
If you get an error message like
@code{*** error: gettext infrastructure mismatch:
using a Makefile.in.in from gettext version ...
but the Autoconf macros are from gettext version ...},
it means that a new GNU gettext release was made, and its Autoconf macros
were integrated into Gnulib and now mismatch the @file{po/} infrastructure.
In this case, fetch and install the new GNU gettext release and run
@code{gettextize} followed by @code{gnulib-tool}.
@end enumerate

On the other hand, if your package is not as concerned with compliance
to the latest standards, but instead favors development on stable
environments, the steps are:

@enumerate
@item
Determine the oldest version of @code{gettext} that you intend to
support during development (at this time, gnulib recommends going no
older than version 0.17).  Run @code{autopoint} (not
@code{gettextize}) to copy infrastructure into place (newer versions
of gettext will install the older infrastructure that you requested).

@item
Invoke @code{gnulib-tool}, and import the @code{gettext-h} module.
@end enumerate

Regardless of which approach you used to get the infrastructure in
place, the following steps must then be used to preserve that
infrastructure (gnulib's @file{bootstrap} script follows these rules):

@enumerate
@item
When a script of yours run @code{autopoint}, invoke @code{gnulib-tool}
afterwards.

@item
When you invoke @code{autoreconf} after @code{gnulib-tool}, make sure to
not invoke @code{autopoint} a second time, by setting the @code{AUTOPOINT}
environment variable, like this:
@smallexample
$ env AUTOPOINT=true autoreconf --install
@end smallexample
@end enumerate


@node Localization
@section Handling Gnulib's own message translations

Gnulib provides some functions that emit translatable messages using GNU
@code{gettext}.  The @samp{gnulib} domain at the
@url{http://translationproject.org/, Translation Project} collects
translations of these messages, which you should incorporate into your
own programs.

There are two basic ways to achieve this.  The first, and older, method
is to list all the source files you use from Gnulib in your own
@file{po/POTFILES.in} file.  This will cause all the relevant
translatable strings to be included in your POT file.  When you send
this POT file to the Translation Project, translators will normally fill
in the translations of the Gnulib strings from their ``translation
memory'', and send you back updated PO files.

However, this process is error-prone: you might forget to list some
source files, or the translator might not be using a translation memory
and provide a different translation than another translator, or the
translation might not be kept in sync between Gnulib and your package.
It is also slow and causes substantial extra work, because a human
translator must be in the loop for each language and you will need to
incorporate their work on request.

For these reasons, a new method was designed and is now recommended.  If
you pass the @code{--po-base=@var{directory}} and @code{--po-domain=@var{domain}}
options to @code{gnulib-tool}, then @code{gnulib-tool} will create a
separate directory with its own @file{POTFILES.in}, and fetch current
translations directly from the Translation Project (using
@command{rsync} or @command{wget}, whichever is available).
The POT file in this directory will be called
@file{@var{domain}-gnulib.pot}, depending on the @var{domain} you gave to the
@code{--po-domain} option (typically the same as the package name).
This causes these translations to reside in a separate message domain,
so that they do not clash either with the translations for the main part
of your package nor with those of other packages on the system that use
possibly different versions of Gnulib.
When you use these options, the functions in Gnulib are built
in such a way that they will always use this domain regardless of the
default domain set by @code{textdomain}.

In order to use this method, you must---in each program that might use
Gnulib code---add an extra line to the part of the program that
initializes locale-dependent behavior.  Where you would normally write
something like:

@example
@group
  setlocale (LC_ALL, "");
  bindtextdomain (PACKAGE, LOCALEDIR);
  textdomain (PACKAGE);
@end group
@end example

@noindent
you should add an additional @code{bindtextdomain} call to inform
gettext of where the MO files for the extra message domain may be found:

@example
@group
  bindtextdomain (PACKAGE "-gnulib", LOCALEDIR);
@end group
@end example

(This example assumes that the @var{domain} that you specified
to @code{gnulib-tool} is the same as the value of the @code{PACKAGE}
preprocessor macro.)

Since you do not change the @code{textdomain} call, the default message
domain for your program remains the same and your own use of @code{gettext}
functions will not be affected.


@node VCS Issues
@section Issues with Version Control Systems

If a project stores its source files in a version control system (VCS),
such as CVS, Subversion, or Git, one needs to decide which files to commit.

In principle, all files created by @code{gnulib-tool}, except
@file{gnulib-cache.m4}, can be treated like generated source files,
like for example a @file{parser.c} file generated from
@file{parser.y}.  Alternatively, they can be considered source files
and updated manually.

Here are the three different approaches in common use.  Each has its
place, and you should use whichever best suits your particular project
and development methods.

@enumerate
@item
In projects which commit all source files, whether generated or not,
into their VCS, the @code{gnulib-tool} generated files should all be
committed.  In this case, you should pass the option
@samp{--no-vc-files} to @code{gnulib-tool}, which avoids alteration of
VCS-related files such as @file{.gitignore}.

Gnulib also contains files generated by @command{make} (and removed by
@code{make clean}), using information determined by
@command{configure}.  For a Gnulib source file of the form
@file{lib/foo.in.h}, the corresponding @file{lib/foo.h} is such a
@command{make}-generated file.  These should @emph{not} be checked
into the VCS, but instead added to @file{.gitignore} or equivalent.

@item
In projects which customarily omit from their VCS all files that are
generated from other source files, none of these files and directories
are added into the VCS.  As described in @ref{Modified imports}, there
are two ways to keep track of options and module names that are passed
to @code{gnulib-tool}.  The command for restoring the omitted files
depends on it:

@itemize @bullet
@item
If they are stored in a file other than @code{gnulib-cache.m4}, such as
@file{autogen.sh}, @file{bootstrap}, @file{bootstrap.conf}, or similar,
the restoration command is the entire @code{gnulib-tool ... --import ...}
invocation with all options and module names.

@item
If the project relies on @code{gnulib-tool}'s memory of the last used
options and module names, then the file @file{gnulib-cache.m4} in the M4
macros directory must be added to the VCS, and the restoration command
is:

@smallexample
$ gnulib-tool --update
@end smallexample

The @samp{--update} option operates much like the @samp{--add-import}
option, but it does not offer the possibility to change the way Gnulib is
used.  Also it does not report in the ChangeLogs the files that it had to
add because they were missing.

@end itemize

Gnulib includes the file @file{build-aux/bootstrap} to aid a developer
in using this setup.  Furthermore, in projects that use git for
version control, it is possible to use a git submodule containing the
precise commit of the gnulib repository, so that each developer
running @file{bootstrap} will get the same version of all
gnulib-provided files.  The location of the submodule can be chosen to
fit the package's needs; here's how to initially create the submodule
in the directory @file{.gnulib}:

@smallexample
$ dir=.gnulib
$ git submodule add -- git://git.sv.gnu.org/gnulib.git $dir
$ git config alias.syncsub "submodule foreach git pull origin master"
@end smallexample

@noindent
Thereafter, @file{bootstrap} can run this command to update the
submodule to the recorded checkout level:

@smallexample
git submodule update --init $dir
@end smallexample

@noindent
and a developer can use this sequence to update to a newer version of
gnulib:

@smallexample
$ git syncsub
$ git add $dir
$ ./bootstrap
@end smallexample

@item
Some projects take a ``middle road'': they do commit Gnulib source
files as in the first approach, but they do not commit other derived
files, such as a @code{Makefile.in} generated by Automake.  This
increases the size and complexity of the repository, but can help
occasional contributors by not requiring them to have a full Gnulib
checkout to do a build, and all developers by ensuring that all
developers are working with the same version of Gnulib in the
repository.  It also supports multiple Gnulib instances within a
project.  It remains important not to commit the
@command{make}-generated files, as described above.

@end enumerate


@node Unit tests
@section Bundling the unit tests of the Gnulib modules

You can bundle the unit tests of the Gnulib modules together with your
package, through the @samp{--with-tests} option.  Together with
@samp{--with-tests}, you also specify the directory for these tests
through the @samp{--tests-base} option.  Of course, you need to add this
directory to the @code{SUBDIRS} variable in the @code{Makefile.am} of
the parent directory.

The advantage of having the unit tests bundled is that when your program
has a problem on a particular platform, running the unit tests may help
determine quickly if the problem is on Gnulib's side or on your package's
side.  Also, it helps verifying Gnulib's portability, of course.

The unit tests will be compiled and run when the user runs @samp{make check}.
When the user runs only @samp{make}, the unit tests will not be compiled.

In the @code{SUBDIRS} variable, it is useful to put the Gnulib tests directory
after the directory containing the other tests, not before:

@smallexample
SUBDIRS = gnulib-lib src man tests gnulib-tests
@end smallexample

@noindent
This will ensure that on platforms where there are test failures in either
directory, users will see and report the failures from the tests of your
program.

Note: In packages which use more than one invocation of @code{gnulib-tool}
in the scope of the same @code{configure.ac}, you cannot use
@samp{--with-tests}.  You will have to use a separate @code{configure.ac}
in this case.


@node Conditional dependencies
@section Avoiding unnecessary checks and compilations

@cindex conditional dependencies
In some cases, a module is needed by another module only on specific
platforms.  But when a module is present, its Autoconf checks are always
executed, and its @code{Makefile.am} additions are always enabled.  So
it can happen that some Autoconf checks are executed and some source files
are compiled, although no other module needs them on this particular
platform, just @emph{in case} some other module would need them.

The option @samp{--conditional-dependencies} enables an optimization of
configure checks and @code{Makefile.am} snippets that avoids this.  With
this option, whether a module is considered ``present'' is no longer decided
when @code{gnulib-tool} is invoked, but later, when @code{configure} is run.
This applies to modules that were added as dependencies while
@code{gnulib-tool} was run; modules that were passed on the command line
explicitly are always ``present''.

For example, the @code{timegm} module needs, on platforms
where the system's @code{timegm} function is missing or buggy, a replacement
that is based on a function @code{mktime_internal}.  The module
@code{mktime-internal} that provides this function provides it on all
platforms.  So, by default, the file @file{mktime-internal.c} will be
compiled on all platforms, even on glibc and BSD systems which have a
working @code{timegm} function.  When the option
@samp{--conditional-dependencies} is given, on the other hand, and if
@code{mktime-internal} was not explicitly required on the command line,
the file @file{mktime-internal.c} will only be compiled on the platforms
where the @code{timegm} needs them.

Conditional dependencies are specified in the module description by putting
the condition on the same line as the dependent module, enclosed in brackets.
The condition is a boolean shell expression that can assume that the
@code{configure.ac} snippet from the module description has already been
executed.  In the example above, the dependency from @code{timegm} to
@code{mktime-internal} is written like this:

@smallexample
Depends-on:
...
mktime-internal [test $HAVE_TIMEGM = 0 || test $REPLACE_TIMEGM = 1]
...
@end smallexample

Note: The option @samp{--conditional-dependencies} cannot be used together
with the option @samp{--with-tests}.  It also cannot be used when a package
uses @code{gnulib-tool} for several subdirectories, with different values
of @samp{--source-base}, in the scope of a single @code{configure.ac} file.