gnulib/doc/gnulib-intro.texi

@node Benefits
@section Benefits of using Gnulib

Gnulib is useful to enhance various aspects of a package:

@itemize @bullet
@item
Portability: With Gnulib, a package maintainer can program against the
POSIX and GNU libc APIs and nevertheless expect good portability to
platforms that don't implement POSIX.

@item
Maintainability: When a package uses modules from Gnulib instead of code
written specifically for that package, the maintainer has less code to
maintain.

@item
Security: Gnulib provides functions that are immune against vulnerabilities
that plague the uses of the corresponding commonplace functions. For
example, @code{asprintf}, @code{canonicalize_file_name} are not affected
by buffer sizing problems that affect @code{sprintf}, @code{realpath}.
@code{openat} does not have the race conditions that @code{open} has. Etc.

@item
Reliability: Gnulib provides functions that combine a call to a system
function with a check of the result. Examples are @code{xalloc},
@code{xprintf}, @code{xstrtod}, @code{xgetcwd}.

@item
Structure: Gnulib offers a way to structure code into modules, typically
one include file, one source code file, and one autoconf macro for each
functionality. Modularity helps maintainability.
@end itemize

@node Library vs Reusable Code
@section Library vs. Reusable Code

Classical libraries are installed as binary object code.  Gnulib is
different: It is used as a source code library.  Each package that uses
Gnulib thus ships with part of the Gnulib source code.  The used portion
of Gnulib is tailored to the package: A build tool, called
@code{gnulib-tool}, is provided that copies a tailored subset of Gnulib
into the package.

@menu
* Other source code libraries::
@end menu

@node Other source code libraries
@subsection Comparison with other source code libraries

Many programming languages nowadays have
@itemize
@item
an ecosystem of reusable source code packages, available through a central site,
@item
together with a tool that downloads dependencies from this central site.
@end itemize

Here is a table of such source code libraries.

@multitable @columnfractions .3 .35 .35
@headitem Language @tab Central site @tab Download tool
@item Python @tab @url{https://pypi.org,,pypi.org} @tab @samp{pip}
@item Java @tab @url{https://mvnrepository.com,,mvnrepository.com} @tab @samp{maven}, @samp{gradle}
@item C# @tab @url{https://nuget.org,,nuget.org} @tab @samp{nuget}
@item JavaScript, TypeScript @tab @url{https://npmjs.com,,npmjs.com} @tab @samp{npm}, @samp{yarn}
@item Common Lisp @tab @url{https://quicklisp.org,,quicklisp.org} @tab @samp{ql:quickload}
@item Emacs Lisp @tab @url{https://elpa.gnu.org,,elpa.gnu.org} @tab @samp{M-x package-install}
@item Rust @tab @url{https://crates.io,,crates.io} @tab @samp{cargo}
@item Go @tab @url{https://pkg.go.dev,,pkg.go.dev} @tab @samp{go mod}
@item Ruby @tab @url{https://rubygems.org,,rubygems.org} @tab @samp{gem}
@item Lua @tab @url{https://luarocks.org,,luarocks.org} @tab @samp{luarocks}
@item OCaml @tab @url{https://ocaml.org/packages,,ocaml.org/packages} @tab @samp{opam}
@item D @tab @url{https://code.dlang.org,,code.dlang.org} @tab @samp{dub}
@item Perl @tab @url{https://cpan.org,,cpan.org} @tab @samp{cpan}
@item PHP @tab @url{https://packagist.org,,packagist.org} @tab @samp{composer}
@item R @tab @url{https://cran.r-project.org,,cran.r-project.org} @tab @samp{R CMD INSTALL}
@end multitable

@cindex supply chain attacks
Most of them are vulnerable to @emph{supply chain attacks}.
These are cyber-attacks to the central site (library)
or to one of the packages in the library,
with the goal of attacking the packages that use that specific package
as a (direct or indirect) dependency.

Examples of such supply chain attacks:
@itemize
@item
A disenchanted package maintainer may replace the code of their package
with something that does not work any more.
@item
An attacker may plant a backdoor in a package.
@item
An attacker may register packages whose names only slightly differ
from well-known and heavily used packages (``typo-squatting'').
@end itemize

While some mitigations exist, they are often cumbersome to put in place.
For example, a developer could set up a partial mirror of the library,
where adding a package to the mirror is a manual operation,
together with a configuration or modification of the tool that
redirects download or installation requests to the mirror.

Gnulib is not vulnerable to such attacks, because
all of its code is managed in a single repository,
with a limited set of committers
and with established code review practices.

@node Portability and Application Code
@section Portability and Application Code

One of the goals of Gnulib is to make portable programming easy, on
the basis of the standards relevant for GNU (and Unix).  The objective
behind that is to avoid a fragmentation of the user community into
disjoint user communities according to the operating system, and
instead allow synergies between users on different operating systems.

Another goal of Gnulib is to provide application code that can be shared
between several applications.  Some people wonder: "What? glibc doesn't
have a function to copy a file?"  Indeed, the scope of a system's libc is
to implement the relevant standards (ISO C, POSIX) and to provide
access functions to the kernel's system calls, and little more.

There is no clear borderline between both areas.

For example, Gnulib has a facility for generating the name of backup
files.  While this task is entirely at the application level---no
standard specifies an API for it---the naïve code has
some portability problems because on some platforms the length of file
name components is limited to 30 characters or so.  Gnulib handles
that.

Similarly, Gnulib has a facility for executing a command in a
subprocess.  It is at the same time a portability enhancement (it
works on GNU, Unix, and Windows, compared to the classical
@code{fork}/@code{exec} idiom which is not portable to Windows), as well
as an application aid: it takes care of redirecting stdin and/or
stdout if desired, and emits an error message if the subprocess
failed.

@node Target Platforms
@section Target Platforms

Gnulib supports a number of platforms that we call the ``reasonable
portability targets''.  This class consists of widespread operating systems,
for three years after their last availability, or---for proprietary
operating systems---as long as the vendor provides commercial support for
it.  Already existing Gnulib code for older operating systems is usually
left in place for longer than these three years.  So it comes that programs
that use Gnulib run pretty well also on these older operating systems.

Some operating systems are not very widespread, but are Free Software and
are actively developed.  Such platforms are also supported by Gnulib, if
that OS's developers community keeps in touch with the Gnulib developers,
by providing bug reports, analyses, or patches.  For such platforms, Gnulib
supports only the versions of the last year or the last few months,
depending on the maturity of said OS project, the number of its users, and
how often these users upgrade.

Niche operating systems are generally unsupported by Gnulib, unless some
of their developers or users contribute support to Gnulib.

The degree of support Gnulib guarantees for a platform depends on the
amount of testing it gets from volunteers.  Platforms on which Gnulib
is frequently tested are the best supported.  Then come platforms with
occasional testing, then platforms which are rarely tested.  Usually,
we fix bugs when they are reported.  Except that some rarely tested
platforms are also low priority; bug fixes for these platforms can
take longer.

@menu
* Supported Platforms::
* Formerly Supported Platforms::
* Unsupported Platforms::
@end menu

@node Supported Platforms
@subsection Supported Platforms

As of 2024, the list of supported platforms is the following:

@itemize
@item
glibc systems.  With glibc 2.19 or newer, they are frequently tested.
@c [Not very relevant in the long term.]
@c The distributions Ubuntu, Debian, Fedora, CentOS are frequently tested.
@c Arch Linux is occasionally tested.
@c gNewSense, Trisquel, OpenSUSE are rarely tested.
About the kernels:
@itemize
@item
glibc on Linux is frequently tested.
@item
glibc on kFreeBSD is rarely tested.
@item
@c There is Alpine Linux 3.19, and also musl-gcc on Ubuntu.
musl libc on Linux is occasionally tested.
@item
Fil-C on Linux is occasionally tested.
@end itemize
@item
macOS@.  In versions 12.5, it's occasionally tested.  In version
10.5, it's rarely tested.
@item
FreeBSD 14.0 or newer is occasionally tested.
@item
OpenBSD 7.0 or newer is occasionally tested.
@item
NetBSD 10.0 or newer is occasionally tested.
@item
CheriBSD 25.03 or newer is occasionally tested.
@item
AIX 7.1 and 7.3 are occasionally tested.
@item
Solaris 10 and 11.4 are occasionally tested.  Solaris 9 is rarely
tested and low priority.
@item
Android is occasionally tested, through the Termux app on Android 11.
@item
Cygwin 2.9 and 3.4 are occasionally tested.  Cygwin 1.7.x no longer tested.
@item
Native Windows:
@itemize
@item
mingw is occasionally tested.  Only the latest version of mingw is
tested; older versions are not supported.
@item
MSVC 14 (Microsoft Visual Studio 2015 14.0) is occasionally tested.
Only ``release'' builds (compiler option @samp{-MD}) are supported,
not ``debug'' builds (compiler option @samp{-MDd}).
@end itemize
Note that some modules are currently unsupported on native Windows:
@code{mgetgroups}, @code{getugroups}, @code{idcache},
@code{userspec}, @code{openpty}, @code{login_tty}, @code{forkpty},
@code{pt_chown}, @code{grantpt}, @code{pty-h}, @code{savewd},
@code{mkancesdirs}, @code{mkdir-p}, @code{euidaccess}, @code{faccessat}.
The versions of Windows that are supported are Windows 10 and newer.
@item
GNU Hurd 0.9 is rarely tested.
@item
z/OS is not tested.
@item
Minix 3.3.0 is no longer tested.
@item
Haiku is no longer tested.
@item
uClibc on Linux is no longer tested.
@item
QNX is no longer tested.
@end itemize

@node Formerly Supported Platforms
@subsection Formerly Supported Platforms

The following platforms were supported in the past, but are no longer
supported:
@itemize
@item
glibc versions 2.1.x and older.
@item
Mac OS X 10.4 and older.
@item
AIX 6 and older.
@item
IBM XL C/C++ 13.1 and older.
@item
HP-UX 11.31.
@item
@c IRIX 6.5 cc has no option for C99 support. You would need to use gcc instead.
IRIX 6.5 and older.
@item
OSF/1 5.1.
@item
Solaris 8 and older.
@item
Interix.
@item
BeOS.
@end itemize

Gnulib supports these operating systems only in an unvirtualized environment.
When you run an OS inside a virtual machine, you have to be aware that the
virtual machine can bring in bugs of its own.  For example, floating-point
operations on Solaris can behave slightly differently in QEMU than on real
hardware.  And Haiku's @command{bash} program misbehaves in VirtualBox 3,
whereas it behaves fine in VirtualBox 4.

Similarly, running native Windows binaries on GNU/Linux under WINE is
rarely tested and low priority: WINE has a set of behaviours and bugs that
is slightly different from native Windows.

@node Unsupported Platforms
@subsection Unsupported Platforms

@cindex integer arithmetic portability
@cindex portability, integer arithmetic

Some platforms with C compilers are not supported by Gnulib because
the platforms violate Gnulib's C portability assumptions.  @xref{Other
portability assumptions}.

These assumptions are not required by the C or POSIX standards but
hold on almost all practical porting targets.  If you need to port
Gnulib code to a platform where these assumptions are not true, we
would appreciate hearing of any fixes.  We need fixes that do not
increase runtime overhead on standard hosts and that are relatively
easy to maintain.

These platforms are listed below to illustrate problems that Gnulib
and Gnulib-using code would have if it were intended to be portable to
all practical POSIX or C platforms.

@itemize @bullet
@item
In Clang 19 and earlier, the @option{-fsanitize=undefined} option
causes the program to crash if it adds zero to a null pointer --
behavior that is undefined in strict C17 and earlier, but which yields
a null pointer in C23 and later and on all practical porting targets,
and which the Gnulib portability guidelines allow.

If you use Clang 19 or earlier with @option{-fsanitize=undefined}, you can work
around the problem by also using @samp{-fno-sanitize=pointer-overflow},
although this may also disable some unrelated and useful pointer checks.
Perhaps someday the Clang developers will fix the infelicity.

@item
The IBM i's pointers are 128 bits wide and it lacks the two types
@code{intptr_t} and @code{uintptr_t}, which are optional in the C and
POSIX standards.  However, these two types are required for the XSI
extension to POSIX, and many Gnulib modules use them.  To work around
this compatibility problem, Gnulib-using applications can be run on
the IBM i's PASE emulation environment.  The IBM i's architecture
descends from the System/38 (1978).

@item
The Unisys ClearPath Dorado's machine word is 36 bits.  Its signed
integers use a ones'-complement representation that conforms to
C17 and earlier but not to C23 and later.  On these machines,
@code{CHAR_BIT == 9} and @code{INT_MIN == -INT_MAX}.  By default
@code{UINT_MAX} is @math{2^{36} - 2}, which does not conform to the C
requirement that it be one less than a power of two.  Although
compiler options can raise @code{UINT_MAX} to be @math{2^{36} - 1},
this can break system code that uses @math{-0} as a flag value.
This platform's architecture descends from the UNIVAC 1103 (1953).

@item
The Unisys ClearPath Libra's machine word is 48 bits
with a 4-bit tag and a 4-bit data extension.  Its
@code{unsigned int} uses the low-order 39 bits of the word, and
@code{int} uses the low-order 40 bits of the word with a
signed-magnitude representation that conforms to C17 and earlier
but not to C23 and later.  On these machines, @code{INT_MAX ==
UINT_MAX}, @code{INT_MIN == -INT_MAX}, and @code{sizeof (int) == 6}.
This platform's architecture descends from the Burroughs B5000 (1961).
@end itemize

The following platforms are not supported by Gnulib.  The cost of
supporting them would exceed the benefit because they are rarely used, or
poorly documented, or have been supplanted by other platforms, or diverge
too much from POSIX, or some combination of these and other factors.
Please don't bother sending us patches for them.

@itemize
@item
Windows 95/98/ME.
@item
DJGPP and EMX (the 32-bit operating systems running in DOS).
@item
MSDOS (the 16-bit operating system).
@item
Windows Mobile, Symbian OS, iOS.
@end itemize

@node Modules
@section Modules

Gnulib is divided into modules.  Every module implements a single
facility.  Modules can depend on other modules.

A module consists of a number of files and a module description.  The
files are copied by @code{gnulib-tool} into the package that will use it,
usually verbatim, without changes.  Source code files (.h, .c files)
reside in the @file{lib/} subdirectory.  Autoconf macro files reside in
the @file{m4/} subdirectory.  Build scripts reside in the
@file{build-aux/} subdirectory.

The module description contains the list of files; @code{gnulib-tool}
copies these files.  It contains the module's
dependencies; @code{gnulib-tool} installs them as well.  It also
contains the autoconf macro invocation (usually a single line or
nothing at all); @code{gnulib-tool} ensures this is invoked from the
package's @file{configure.ac} file.  And also a @file{Makefile.am}
snippet; @code{gnulib-tool} collects these into a @file{Makefile.am}
for the tailored Gnulib part.  The module description and include file
specification are for documentation purposes; they are combined into
@file{MODULES.html}.

The module system serves two purposes:

@enumerate
@item
It ensures consistency of the used autoconf macros and @file{Makefile.am}
rules with the source code.  For example, source code which uses the
@code{getopt_long} function---this is a common way to implement parsing
of command line options in a way that complies with the GNU standards---needs
the source code (@file{lib/getopt.c} and others), the autoconf macro
which detects whether the system's libc already has this function (in
@file{m4/getopt.m4}), and a few @file{Makefile.am} lines that create the
substitute @file{getopt.h} if not.  These three pieces belong together.
They cannot be used without each other.  The module description and
@code{gnulib-tool} ensure that they are copied altogether into the
destination package.

@item
It allows for scalability.  It is well-known since the inception of the
MODULA-2 language around 1978 that dissection into modules with
dependencies allows for building large sets of code in a maintainable way.
The maintainability comes from the facts that:

@itemize @bullet
@item
Every module has a single purpose; you don't worry about other parts of
the program while creating, reading or modifying the code of a module.

@item
The code you have to read in order to understand a module is limited to
the source of the module and the .h files of the modules listed as
dependencies.  It is for this reason also that we recommend to put the
comments describing the functions exported by a module into its .h file.
@end itemize

In other words, the module is the elementary unit of code in Gnulib,
comparable to a class in object-oriented languages like Java or C#.
@end enumerate

The module system is the basis of @code{gnulib-tool}.  When
@code{gnulib-tool} copies a part of Gnulib into a package, it first
compiles a module list, starting with the requested modules and adding all
the dependencies, and then collects the files, @file{configure.ac}
snippets and @file{Makefile.am} snippets.

@node Various Kinds of Modules
@section Various Kinds of Modules

There are modules of various kinds in Gnulib.  For a complete list of the
modules, see in @file{MODULES.html}.

@subsection Support for ISO C or POSIX functions.

When a function is not implemented by a system, the Gnulib module provides
an implementation under the same name.  Examples are the @samp{snprintf}
and @samp{readlink} modules.

Similarly, when a function is not correctly implemented by a system,
Gnulib provides a replacement.  For functions, we use the pattern

@smallexample
#if !HAVE_WORKING_FOO
# define foo rpl_foo
#endif
@end smallexample

@noindent
and implement the @code{foo} function under the name @code{rpl_foo}.  This
renaming is needed to avoid conflicts at compile time (in case the system
header files declare @code{foo}) and at link/run time (because the code
making use of @code{foo} could end up residing in a shared library, and
the executable program using this library could be defining @code{foo}
itself).

For header files, such as @code{stdint.h}, we provide
the substitute only if the system doesn't provide a correct one.  The
template of this replacement is distributed in a slightly different name,
with @samp{.in} inserted before the @samp{.h} extension, so that on
systems which do provide a correct
header file the system's one is used.

The modules in this category are supported in C++ mode as well.  This
means, while the autoconfiguration uses the C compiler, the resulting
header files and function substitutes can be used with a matching C++
compiler as well.

@subsection Enhancements of ISO C or POSIX functions

These are sometimes POSIX functions with GNU extensions also found in
glibc---examples: @samp{getopt}, @samp{fnmatch}---and often new
APIs---for example, for all functions that allocate memory in one way
or the other, we have variants which also include the error checking
against the out-of-memory condition.

@subsection Portable general use facilities

Examples are a module for copying a file---the portability problems
relate to the copying of the file's modification time, access rights,
and extended attributes---or a module for extracting the tail
component of a file name---here the portability to native Windows
requires a different API than the classical POSIX @code{basename} function.

@subsection Reusable application code

Examples are an error reporting function, a module that allows output of
numbers with K/M/G suffixes, or cryptographic facilities.

@subsection Object oriented classes

Examples are data structures like @samp{list}, or abstract output stream
classes that work around the fact that an application cannot implement an
stdio @code{FILE} with its logic.  Here, while staying in C, we use
implementation techniques like tables of function pointers, known from the
C++ language or from the Linux kernel.

@subsection Interfaces to external libraries

@mindex iconv
Examples are the @samp{iconv} module, which interfaces to the
@code{iconv} facility, regardless whether it is contained in libc or in
an external @code{libiconv}.
@mindex readline
Or the @samp{readline} module, which interfaces to the GNU readline library.

@subsection Build / maintenance infrastructure

@mindex maintainer-makefile
An example is the @samp{maintainer-makefile} module, which provides extra
Makefile tags for maintaining a package.

@node Collaborative Development
@section Collaborative Development

Gnulib is maintained collaboratively.  The mailing list is
@code{<bug-gnulib at gnu dot org>}.  Be warned that some people on the
list may be very active at some times and unresponsive at other times.

Every module has one or more maintainers.  While issues are discussed
collaboratively on the list, the maintainer of a module nevertheless has
a veto right regarding changes in his module.

All patches should be posted to the list, regardless whether they are
proposed patches or whether they are committed immediately by the
maintainer of the particular module.  The purpose is not only to inform
the other users of the module, but mainly to allow peer review.  It is not
uncommon that several people contribute comments or spot bugs after a
patch was proposed.

Conversely, if you are using Gnulib, and a patch is posted that affects
one of the modules that your package uses, you have an interest in
proofreading the patch.

@node Copyright
@section Copyright

Most modules are under the GPL@.  Some, mostly modules which can
reasonably be used in libraries, are under LGPL@.  Few modules are
under other licenses, such as LGPLv2+, unlimited, or public domain.

If the module description file says "GPL", it means "GPLv3+" (GPLv3
or newer, at the licensee's choice); if it says "LGPL", it means
"LGPLv3+" (LGPLv3 or newer, at the licensee's choice).

The source files, more precisely the files in @file{lib/} and
@file{build-aux/}, are under a license compatible with the module's
license.  Most often, they are under the same license.  But files can be
shared among several modules, and in these cases it can happen that a
source file is under a weaker license than noted in the module
description -- namely under the weakest license among the licenses of
the modules that contain the file.

Different licenses apply to files in special directories:

@table @file
@item modules/
Module description files are under this copyright:

@quotation
Copyright @copyright{} 20XX--20YY Free Software Foundation, Inc.@*
Copying and distribution of this file, with or without modification,
in any medium, are permitted without royalty provided the copyright
notice and this notice are preserved.  This file is offered as-is,
without any warranty.
@end quotation

@item m4/
Autoconf macro files are under this copyright:

@quotation
Copyright @copyright{} 20XX--20YY Free Software Foundation, Inc.@*
This file is free software; the Free Software Foundation
gives unlimited permission to copy and/or distribute it,
with or without modifications, as long as this notice is preserved.
This file is offered as-is, without any warranty.
@end quotation

or under this copyright:

@quotation
Copyright @copyright{} 20XX--20YY Free Software Foundation, Inc.@*
This file is free software, distributed under the terms of the GNU
General Public License as published by the Free Software Foundation;
either version 2 of the License, or (at your option) any later version.
As a special exception to the GNU General Public License, this file
may be distributed as part of a program that contains a configuration
script generated by Autoconf, under the same distribution terms as
the rest of that program.
@end quotation

We denote either of these license statements as ``unlimited''.

@item tests/
If a license statement is not present in a test module, the test files are
under GPL@.  Even if the corresponding source module is under LGPL, this is
not a problem, since compiled tests are not installed by ``make install''.

@item doc/
Documentation files are under this copyright:

@quotation
Copyright @copyright{} 2004--20YY Free Software Foundation, Inc.@*
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with no
Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts.  A
copy of the license is at @url{https://www.gnu.org/licenses/fdl-1.3.en.html}.
@end quotation
@end table

If you want to use some Gnulib modules under LGPL, you can do so by
passing the option @samp{--lgpl} to @code{gnulib-tool}.  This will
ensure that all imported modules can be used under the LGPL license.
Similarly, if you want some Gnulib modules
under LGPLv2+ (Lesser GPL version 2.1 or newer), you can do so by
passing the option @samp{--lgpl=2} to @code{gnulib-tool}.

Keep in mind that when you submit patches to files in Gnulib, you should
license them under a compatible license.  This means that sometimes the
contribution will have to be LGPL, if the original file is available
under LGPL@.  You can find out about it by looking at the license header
of the file.

@node Steady Development
@section Steady Development

Gnulib modules are continually adapted, to match new practices, to be
consistent with newly added modules, or simply as a response to build
failure reports.

If you are willing to report an occasional regression, we recommend to
use the newest version from git always, except in periods of major
changes.  Most Gnulib users do this.

@node Openness
@section Openness

Gnulib is open in the sense that we gladly accept contributions if they
are generally useful, well engineered, and if the contributors have signed
the obligatory papers with the FSF.

The module system is open in the sense that a package using Gnulib can
@enumerate
@item
locally patch or override files in Gnulib,
@item
locally add modules that are treated like Gnulib modules by
@code{gnulib-tool}.
@end enumerate

This is achieved by the @samp{--local-dir} option of @code{gnulib-tool}
(@pxref{Extending Gnulib}).