path: root/manual/startup.texi

@node Process Startup
@chapter Process Startup and Termination

@cindex process
@dfn{Processes} are the primitive units for allocation of system
resources.  Each process has its own address space and (usually) one
thread of control.  A process executes a program; you can have multiple
processes executing the same program, but each process has its own copy
of the program within its own address space and executes it
independently of the other copies.

This chapter explains what your program should do to handle the startup
of a process, to terminate its process, and to receive information
(arguments and the environment) from the parent process.

@menu
* Program Arguments::           Parsing your program's command-line arguments.
* Environment Variables::       How to access parameters inherited from
				 a parent process.
* Program Termination::         How to cause a process to terminate and
				 return status information to its parent.
@end menu

@node Program Arguments
@section Program Arguments
@cindex program arguments
@cindex command line arguments
@cindex arguments, to program

@cindex program startup
@cindex startup of program
@cindex invocation of program
@cindex @code{main} function
@findex main
The system starts a C program by calling the function @code{main}.  It
is up to you to write a function named @code{main}---otherwise, you
won't even be able to link your program without errors.

In ANSI C you can define @code{main} either to take no arguments, or to
take two arguments that represent the command line arguments to the
program, like this:

@smallexample
int main (int @var{argc}, char *@var{argv}[])
@end smallexample

@cindex argc (program argument count)
@cindex argv (program argument vector)
The command line arguments are the whitespace-separated tokens given in
the shell command used to invoke the program; thus, in @samp{cat foo
bar}, the arguments are @samp{foo} and @samp{bar}.  The only way a
program can look at its command line arguments is via the arguments of
@code{main}.  If @code{main} doesn't take arguments, then you cannot get
at the command line.

The value of the @var{argc} argument is the number of command line
arguments.  The @var{argv} argument is a vector of C strings; its
elements are the individual command line argument strings.  The file
name of the program being run is also included in the vector as the
first element; the value of @var{argc} counts this element.  A null
pointer always follows the last element: @code{@var{argv}[@var{argc}]}
is this null pointer.

For the command @samp{cat foo bar}, @var{argc} is 3 and @var{argv} has
three elements, @code{"cat"}, @code{"foo"} and @code{"bar"}.

If the syntax for the command line arguments to your program is simple
enough, you can simply pick the arguments off from @var{argv} by hand.
But unless your program takes a fixed number of arguments, or all of the
arguments are interpreted in the same way (as file names, for example),
you are usually better off using @code{getopt} to do the parsing.

In Unix systems you can define @code{main} a third way, using three arguments:

@smallexample
int main (int @var{argc}, char *@var{argv}[], char *@var{envp})
@end smallexample

The first two arguments are just the same.  The third argument
@var{envp} gives the process's environment; it is the same as the value
of @code{environ}.  @xref{Environment Variables}.  POSIX.1 does not
allow this three-argument form, so to be portable it is best to write
@code{main} to take two arguments, and use the value of @code{environ}.

@menu
* Argument Syntax::       By convention, options start with a hyphen.
* Parsing Options::       The @code{getopt} function.
* Example of Getopt:: 	  An example of parsing options with @code{getopt}.
* Long Options::	  GNU suggests utilities accept long-named options.
			   Here is how to do that.
* Long Option Example::   An example of using @code{getopt_long}.
@end menu

@node Argument Syntax
@subsection Program Argument Syntax Conventions
@cindex program argument syntax
@cindex syntax, for program arguments
@cindex command argument syntax

POSIX recommends these conventions for command line arguments.
@code{getopt} (@pxref{Parsing Options}) makes it easy to implement them.

@itemize @bullet
@item
Arguments are options if they begin with a hyphen delimiter (@samp{-}).

@item
Multiple options may follow a hyphen delimiter in a single token if
the options do not take arguments.  Thus, @samp{-abc} is equivalent to
@samp{-a -b -c}.

@item
Option names are single alphanumeric characters (as for @code{isalnum};
see @ref{Classification of Characters}).

@item
Certain options require an argument.  For example, the @samp{-o} command
of the @code{ld} command requires an argument---an output file name.

@item
An option and its argument may or may not appear as separate tokens.  (In
other words, the whitespace separating them is optional.)  Thus,
@w{@samp{-o foo}} and @samp{-ofoo} are equivalent.

@item
Options typically precede other non-option arguments.

The implementation of @code{getopt} in the GNU C library normally makes
it appear as if all the option arguments were specified before all the
non-option arguments for the purposes of parsing, even if the user of
your program intermixed option and non-option arguments.  It does this
by reordering the elements of the @var{argv} array.  This behavior is
nonstandard; if you want to suppress it, define the
@code{_POSIX_OPTION_ORDER} environment variable.  @xref{Standard
Environment}.

@item
The argument @samp{--} terminates all options; any following arguments
are treated as non-option arguments, even if they begin with a hyphen.

@item
A token consisting of a single hyphen character is interpreted as an
ordinary non-option argument.  By convention, it is used to specify
input from or output to the standard input and output streams.

@item
Options may be supplied in any order, or appear multiple times.  The
interpretation is left up to the particular application program.
@end itemize

@cindex long-named options
GNU adds @dfn{long options} to these conventions.  Long options consist
of @samp{--} followed by a name made of alphanumeric characters and
dashes.  Option names are typically one to three words long, with
hyphens to separate words.  Users can abbreviate the option names as
long as the abbreviations are unique.

To specify an argument for a long option, write
@samp{--@var{name}=@var{value}}.  This syntax enables a long option to
accept an argument that is itself optional.

Eventually, the GNU system will provide completion for long option names
in the shell.

@node Parsing Options
@subsection Parsing Program Options
@cindex program arguments, parsing
@cindex command arguments, parsing
@cindex parsing program arguments

Here are the details about how to call the @code{getopt} function.  To
use this facility, your program must include the header file
@file{unistd.h}.
@pindex unistd.h

@comment unistd.h
@comment POSIX.2
@deftypevar int opterr
If the value of this variable is nonzero, then @code{getopt} prints an
error message to the standard error stream if it encounters an unknown
option character or an option with a missing required argument.  This is
the default behavior.  If you set this variable to zero, @code{getopt}
does not print any messages, but it still returns the character @code{?}
to indicate an error.
@end deftypevar

@comment unistd.h
@comment POSIX.2
@deftypevar int optopt
When @code{getopt} encounters an unknown option character or an option
with a missing required argument, it stores that option character in
this variable.  You can use this for providing your own diagnostic
messages.
@end deftypevar

@comment unistd.h
@comment POSIX.2
@deftypevar int optind
This variable is set by @code{getopt} to the index of the next element
of the @var{argv} array to be processed.  Once @code{getopt} has found
all of the option arguments, you can use this variable to determine
where the remaining non-option arguments begin.  The initial value of
this variable is @code{1}.
@end deftypevar

@comment unistd.h
@comment POSIX.2
@deftypevar {char *} optarg
This variable is set by @code{getopt} to point at the value of the
option argument, for those options that accept arguments.
@end deftypevar

@comment unistd.h
@comment POSIX.2
@deftypefun int getopt (int @var{argc}, char **@var{argv}, const char *@var{options})
The @code{getopt} function gets the next option argument from the
argument list specified by the @var{argv} and @var{argc} arguments.
Normally these values come directly from the arguments received by
@code{main}.

The @var{options} argument is a string that specifies the option
characters that are valid for this program.  An option character in this
string can be followed by a colon (@samp{:}) to indicate that it takes a
required argument.

If the @var{options} argument string begins with a hyphen (@samp{-}), this
is treated specially.  It permits arguments that are not options to be
returned as if they were associated with option character @samp{\0}.

The @code{getopt} function returns the option character for the next
command line option.  When no more option arguments are available, it
returns @code{-1}.  There may still be more non-option arguments; you
must compare the external variable @code{optind} against the @var{argc}
parameter to check this.

If the option has an argument, @code{getopt} returns the argument by
storing it in the varables @var{optarg}.  You don't ordinarily need to
copy the @code{optarg} string, since it is a pointer into the original
@var{argv} array, not into a static area that might be overwritten.

If @code{getopt} finds an option character in @var{argv} that was not
included in @var{options}, or a missing option argument, it returns
@samp{?} and sets the external variable @code{optopt} to the actual
option character.  If the first character of @var{options} is a colon
(@samp{:}), then @code{getopt} returns @samp{:} instead of @samp{?} to
indicate a missing option argument.  In addition, if the external
variable @code{opterr} is nonzero (which is the default), @code{getopt}
prints an error message.
@end deftypefun

@node Example of Getopt
@subsection Example of Parsing Arguments with @code{getopt}

Here is an example showing how @code{getopt} is typically used.  The
key points to notice are:

@itemize @bullet
@item
Normally, @code{getopt} is called in a loop.  When @code{getopt} returns
@code{-1}, indicating no more options are present, the loop terminates.

@item
A @code{switch} statement is used to dispatch on the return value from
@code{getopt}.  In typical use, each case just sets a variable that
is used later in the program.

@item
A second loop is used to process the remaining non-option arguments.
@end itemize

@smallexample
@include testopt.c.texi
@end smallexample

Here are some examples showing what this program prints with different
combinations of arguments:

@smallexample
% testopt
aflag = 0, bflag = 0, cvalue = (null)

% testopt -a -b
aflag = 1, bflag = 1, cvalue = (null)

% testopt -ab
aflag = 1, bflag = 1, cvalue = (null)

% testopt -c foo
aflag = 0, bflag = 0, cvalue = foo

% testopt -cfoo
aflag = 0, bflag = 0, cvalue = foo

% testopt arg1
aflag = 0, bflag = 0, cvalue = (null)
Non-option argument arg1

% testopt -a arg1
aflag = 1, bflag = 0, cvalue = (null)
Non-option argument arg1

% testopt -c foo arg1
aflag = 0, bflag = 0, cvalue = foo
Non-option argument arg1

% testopt -a -- -b
aflag = 1, bflag = 0, cvalue = (null)
Non-option argument -b

% testopt -a -
aflag = 1, bflag = 0, cvalue = (null)
Non-option argument -
@end smallexample

@node Long Options
@subsection Parsing Long Options

To accept GNU-style long options as well as single-character options,
use @code{getopt_long} instead of @code{getopt}.  This function is
declared in @file{getopt.h}, not @file{unistd.h}.  You should make every
program accept long options if it uses any options, for this takes
little extra work and helps beginners remember how to use the program.

@comment getopt.h
@comment GNU
@deftp {Data Type} {struct option}
This structure describes a single long option name for the sake of
@code{getopt_long}.  The argument @var{longopts} must be an array of
these structures, one for each long option.  Terminate the array with an
element containing all zeros.

The @code{struct option} structure has these fields:

@table @code
@item const char *name
This field is the name of the option.  It is a string.

@item int has_arg
This field says whether the option takes an argument.  It is an integer,
and there are three legitimate values: @w{@code{no_argument}},
@code{required_argument} and @code{optional_argument}.

@item int *flag
@itemx int val
These fields control how to report or act on the option when it occurs.

If @code{flag} is a null pointer, then the @code{val} is a value which
identifies this option.  Often these values are chosen to uniquely
identify particular long options.

If @code{flag} is not a null pointer, it should be the address of an
@code{int} variable which is the flag for this option.  The value in
@code{val} is the value to store in the flag to indicate that the option
was seen.
@end table
@end deftp

@comment getopt.h
@comment GNU
@deftypefun int getopt_long (int @var{argc}, char **@var{argv}, const char *@var{shortopts}, struct option *@var{longopts}, int *@var{indexptr})
Decode options from the vector @var{argv} (whose length is @var{argc}).
The argument @var{shortopts} describes the short options to accept, just as
it does in @code{getopt}.  The argument @var{longopts} describes the long
options to accept (see above).

When @code{getopt_long} encounters a short option, it does the same
thing that @code{getopt} would do: it returns the character code for the
option, and stores the options argument (if it has one) in @code{optarg}.

When @code{getopt_long} encounters a long option, it takes actions based
on the @code{flag} and @code{val} fields of the definition of that
option.

If @code{flag} is a null pointer, then @code{getopt_long} returns the
contents of @code{val} to indicate which option it found.  You should
arrange distinct values in the @code{val} field for options with
different meanings, so you can decode these values after
@code{getopt_long} returns.  If the long option is equivalent to a short
option, you can use the short option's character code in @code{val}.

If @code{flag} is not a null pointer, that means this option should just
set a flag in the program.  The flag is a variable of type @code{int}
that you define.  Put the address of the flag in the @code{flag} field.
Put in the @code{val} field the value you would like this option to
store in the flag.  In this case, @code{getopt_long} returns @code{0}.

For any long option, @code{getopt_long} tells you the index in the array
@var{longopts} of the options definition, by storing it into
@code{*@var{indexptr}}.  You can get the name of the option with
@code{@var{longopts}[*@var{indexptr}].name}.  So you can distinguish among
long options either by the values in their @code{val} fields or by their
indices.  You can also distinguish in this way among long options that
set flags.

When a long option has an argument, @code{getopt_long} puts the argument
value in the variable @code{optarg} before returning.  When the option
has no argument, the value in @code{optarg} is a null pointer.  This is
how you can tell whether an optional argument was supplied.

When @code{getopt_long} has no more options to handle, it returns
@code{-1}, and leaves in the variable @code{optind} the index in
@var{argv} of the next remaining argument.
@end deftypefun

@node Long Option Example
@subsection Example of Parsing Long Options

@smallexample
@include longopt.c.texi
@end smallexample

@node Environment Variables
@section Environment Variables

@cindex environment variable
When a program is executed, it receives information about the context in
which it was invoked in two ways.  The first mechanism uses the
@var{argv} and @var{argc} arguments to its @code{main} function, and is
discussed in @ref{Program Arguments}.  The second mechanism uses
@dfn{environment variables} and is discussed in this section.

The @var{argv} mechanism is typically used to pass command-line
arguments specific to the particular program being invoked.  The
environment, on the other hand, keeps track of information that is
shared by many programs, changes infrequently, and that is less
frequently used.

The environment variables discussed in this section are the same
environment variables that you set using assignments and the
@code{export} command in the shell.  Programs executed from the shell
inherit all of the environment variables from the shell.
@c !!! xref to right part of bash manual when it exists

@cindex environment
Standard environment variables are used for information about the user's
home directory, terminal type, current locale, and so on; you can define
additional variables for other purposes.  The set of all environment
variables that have values is collectively known as the
@dfn{environment}.

Names of environment variables are case-sensitive and must not contain
the character @samp{=}.  System-defined environment variables are
invariably uppercase.

The values of environment variables can be anything that can be
represented as a string.  A value must not contain an embedded null
character, since this is assumed to terminate the string.


@menu
* Environment Access::    How to get and set the values of
			   environment variables.
* Standard Environment::  These environment variables have
			   standard interpretations.
@end menu

@node Environment Access
@subsection Environment Access
@cindex environment access
@cindex environment representation

The value of an environment variable can be accessed with the
@code{getenv} function.  This is declared in the header file
@file{stdlib.h}.
@pindex stdlib.h

@comment stdlib.h
@comment ANSI
@deftypefun {char *} getenv (const char *@var{name})
This function returns a string that is the value of the environment
variable @var{name}.  You must not modify this string.  In some non-Unix
systems not using the GNU library, it might be overwritten by subsequent
calls to @code{getenv} (but not by any other library function).  If the
environment variable @var{name} is not defined, the value is a null
pointer.
@end deftypefun


@comment stdlib.h
@comment SVID
@deftypefun int putenv (const char *@var{string})
The @code{putenv} function adds or removes definitions from the environment.
If the @var{string} is of the form @samp{@var{name}=@var{value}}, the
definition is added to the environment.  Otherwise, the @var{string} is
interpreted as the name of an environment variable, and any definition
for this variable in the environment is removed.

The GNU library provides this function for compatibility with SVID; it
may not be available in other systems.
@end deftypefun

@c !!! BSD function setenv

You can deal directly with the underlying representation of environment
objects to add more variables to the environment (for example, to
communicate with another program you are about to execute; see
@ref{Executing a File}).

@comment unistd.h
@comment POSIX.1
@deftypevar {char **} environ
The environment is represented as an array of strings.  Each string is
of the format @samp{@var{name}=@var{value}}.  The order in which
strings appear in the environment is not significant, but the same
@var{name} must not appear more than once.  The last element of the
array is a null pointer.

This variable is declared in the header file @file{unistd.h}.

If you just want to get the value of an environment variable, use
@code{getenv}.
@end deftypevar

Unix systems, and the GNU system, pass the initial value of
@code{environ} as the third argument to @code{main}.
@xref{Program Arguments}.

@node Standard Environment
@subsection Standard Environment Variables
@cindex standard environment variables

These environment variables have standard meanings.  This doesn't mean
that they are always present in the environment; but if these variables
@emph{are} present, they have these meanings.  You shouldn't try to use
these environment variable names for some other purpose.

@comment Extra blank lines make it look better.
@table @code
@item HOME
@cindex HOME environment variable
@cindex home directory

This is a string representing the user's @dfn{home directory}, or
initial default working directory.

The user can set @code{HOME} to any value.
If you need to make sure to obtain the proper home directory
for a particular user, you should not use @code{HOME}; instead,
look up the user's name in the user database (@pxref{User Database}).

For most purposes, it is better to use @code{HOME}, precisely because
this lets the user specify the value.

@c !!! also USER
@item LOGNAME
@cindex LOGNAME environment variable

This is the name that the user used to log in.  Since the value in the
environment can be tweaked arbitrarily, this is not a reliable way to
identify the user who is running a process; a function like
@code{getlogin} (@pxref{Who Logged In}) is better for that purpose.

For most purposes, it is better to use @code{LOGNAME}, precisely because
this lets the user specify the value.

@item PATH
@cindex PATH environment variable

A @dfn{path} is a sequence of directory names which is used for
searching for a file.  The variable @code{PATH} holds a path used
for searching for programs to be run.

The @code{execlp} and @code{execvp} functions (@pxref{Executing a File})
use this environment variable, as do many shells and other utilities
which are implemented in terms of those functions.

The syntax of a path is a sequence of directory names separated by
colons.  An empty string instead of a directory name stands for the
current directory (@pxref{Working Directory}).

A typical value for this environment variable might be a string like:

@smallexample
:/bin:/etc:/usr/bin:/usr/new/X11:/usr/new:/usr/local/bin
@end smallexample

This means that if the user tries to execute a program named @code{foo},
the system will look for files named @file{foo}, @file{/bin/foo},
@file{/etc/foo}, and so on.  The first of these files that exists is
the one that is executed.

@c !!! also TERMCAP
@item TERM
@cindex TERM environment variable

This specifies the kind of terminal that is receiving program output.
Some programs can make use of this information to take advantage of
special escape sequences or terminal modes supported by particular kinds
of terminals.  Many programs which use the termcap library
(@pxref{Finding a Terminal Description,Find,,termcap,The Termcap Library
Manual}) use the @code{TERM} environment variable, for example.

@item TZ
@cindex TZ environment variable

This specifies the time zone.  @xref{TZ Variable}, for information about
the format of this string and how it is used.

@item LANG
@cindex LANG environment variable

This specifies the default locale to use for attribute categories where
neither @code{LC_ALL} nor the specific environment variable for that
category is set.  @xref{Locales}, for more information about
locales.

@ignore
@c I doubt this really exists
@item LC_ALL
@cindex LC_ALL environment variable

This is similar to the @code{LANG} environment variable.  However, its
value takes precedence over any values provided for the individual
attribute category environment variables, or for the @code{LANG}
environment variable.
@end ignore

@item LC_COLLATE
@cindex LC_COLLATE environment variable

This specifies what locale to use for string sorting.

@item LC_CTYPE
@cindex LC_CTYPE environment variable

This specifies what locale to use for character sets and character
classification.

@item LC_MONETARY
@cindex LC_MONETARY environment variable

This specifies what locale to use for formatting monetary values.

@item LC_NUMERIC
@cindex LC_NUMERIC environment variable

This specifies what locale to use for formatting numbers.

@item LC_TIME
@cindex LC_TIME environment variable

This specifies what locale to use for formatting date/time values.

@item _POSIX_OPTION_ORDER
@cindex _POSIX_OPTION_ORDER environment variable.

If this environment variable is defined, it suppresses the usual
reordering of command line arguments by @code{getopt}.  @xref{Argument Syntax}.

@c !!! GNU also has COREFILE, CORESERVER, EXECSERVERS
@end table

@node Program Termination
@section Program Termination
@cindex program termination
@cindex process termination

@cindex exit status value
The usual way for a program to terminate is simply for its @code{main}
function to return.  The @dfn{exit status value} returned from the
@code{main} function is used to report information back to the process's
parent process or shell.

A program can also terminate normally by calling the @code{exit}
function.

In addition, programs can be terminated by signals; this is discussed in
more detail in @ref{Signal Handling}.  The @code{abort} function causes
a signal that kills the program.

@menu
* Normal Termination::          If a program calls @code{exit}, a
                                 process terminates normally.
* Exit Status::                 The @code{exit status} provides information
                                 about why the process terminated.
* Cleanups on Exit::            A process can run its own cleanup
                                 functions upon normal termination.
* Aborting a Program::          The @code{abort} function causes
                                 abnormal program termination.
* Termination Internals::       What happens when a process terminates.
@end menu

@node Normal Termination
@subsection Normal Termination

A process terminates normally when the program calls @code{exit}.
Returning from @code{main} is equivalent to calling @code{exit}, and
the value that @code{main} returns is used as the argument to @code{exit}.

@comment stdlib.h
@comment ANSI
@deftypefun void exit (int @var{status})
The @code{exit} function terminates the process with status
@var{status}.  This function does not return.
@end deftypefun

Normal termination causes the following actions:

@enumerate
@item
Functions that were registered with the @code{atexit} or @code{on_exit}
functions are called in the reverse order of their registration.  This
mechanism allows your application to specify its own ``cleanup'' actions
to be performed at program termination.  Typically, this is used to do
things like saving program state information in a file, or unlocking
locks in shared data bases.

@item
All open streams are closed, writing out any buffered output data.  See
@ref{Closing Streams}.  In addition, temporary files opened
with the @code{tmpfile} function are removed; see @ref{Temporary Files}.

@item
@code{_exit} is called, terminating the program.  @xref{Termination Internals}.
@end enumerate

@node Exit Status
@subsection Exit Status
@cindex exit status

When a program exits, it can return to the parent process a small
amount of information about the cause of termination, using the
@dfn{exit status}.  This is a value between 0 and 255 that the exiting
process passes as an argument to @code{exit}.

Normally you should use the exit status to report very broad information
about success or failure.  You can't provide a lot of detail about the
reasons for the failure, and most parent processes would not want much
detail anyway.

There are conventions for what sorts of status values certain programs
should return.  The most common convention is simply 0 for success and 1
for failure.  Programs that perform comparison use a different
convention: they use status 1 to indicate a mismatch, and status 2 to
indicate an inability to compare.  Your program should follow an
existing convention if an existing convention makes sense for it.

A general convention reserves status values 128 and up for special
purposes.  In particular, the value 128 is used to indicate failure to
execute another program in a subprocess.  This convention is not
universally obeyed, but it is a good idea to follow it in your programs.

@strong{Warning:} Don't try to use the number of errors as the exit
status.  This is actually not very useful; a parent process would
generally not care how many errors occurred.  Worse than that, it does
not work, because the status value is truncated to eight bits.
Thus, if the program tried to report 256 errors, the parent would
receive a report of 0 errors---that is, success.

For the same reason, it does not work to use the value of @code{errno}
as the exit status---these can exceed 255.

@strong{Portability note:} Some non-POSIX systems use different
conventions for exit status values.  For greater portability, you can
use the macros @code{EXIT_SUCCESS} and @code{EXIT_FAILURE} for the
conventional status value for success and failure, respectively.  They
are declared in the file @file{stdlib.h}.
@pindex stdlib.h

@comment stdlib.h
@comment ANSI
@deftypevr Macro int EXIT_SUCCESS
This macro can be used with the @code{exit} function to indicate
successful program completion.

On POSIX systems, the value of this macro is @code{0}.  On other
systems, the value might be some other (possibly non-constant) integer
expression.
@end deftypevr

@comment stdlib.h
@comment ANSI
@deftypevr Macro int EXIT_FAILURE
This macro can be used with the @code{exit} function to indicate
unsuccessful program completion in a general sense.

On POSIX systems, the value of this macro is @code{1}.  On other
systems, the value might be some other (possibly non-constant) integer
expression.  Other nonzero status values also indicate future.  Certain
programs use different nonzero status values to indicate particular
kinds of "non-success".  For example, @code{diff} uses status value
@code{1} to mean that the files are different, and @code{2} or more to
mean that there was difficulty in opening the files.
@end deftypevr

@node Cleanups on Exit
@subsection Cleanups on Exit

Your program can arrange to run its own cleanup functions if normal
termination happens.  If you are writing a library for use in various
application programs, then it is unreliable to insist that all
applications call the library's cleanup functions explicitly before
exiting.  It is much more robust to make the cleanup invisible to the
application, by setting up a cleanup function in the library itself
using @code{atexit} or @code{on_exit}.

@comment stdlib.h
@comment ANSI
@deftypefun int atexit (void (*@var{function}) (void))
The @code{atexit} function registers the function @var{function} to be
called at normal program termination.  The @var{function} is called with
no arguments.

The return value from @code{atexit} is zero on success and nonzero if
the function cannot be registered.
@end deftypefun

@comment stdlib.h
@comment SunOS
@deftypefun int on_exit (void (*@var{function})(int @var{status}, void *@var{arg}), void *@var{arg})
This function is a somewhat more powerful variant of @code{atexit}.  It
accepts two arguments, a function @var{function} and an arbitrary
pointer @var{arg}.  At normal program termination, the @var{function} is
called with two arguments:  the @var{status} value passed to @code{exit},
and the @var{arg}.

This function is included in the GNU C library only for compatibility
for SunOS, and may not be supported by other implementations.
@end deftypefun

Here's a trivial program that illustrates the use of @code{exit} and
@code{atexit}:

@smallexample
@include atexit.c.texi
@end smallexample

@noindent
When this program is executed, it just prints the message and exits.

@node Aborting a Program
@subsection Aborting a Program
@cindex aborting a program

You can abort your program using the @code{abort} function.  The prototype
for this function is in @file{stdlib.h}.
@pindex stdlib.h

@comment stdlib.h
@comment ANSI
@deftypefun void abort (void)
The @code{abort} function causes abnormal program termination.  This
does not execute cleanup functions registered with @code{atexit} or
@code{on_exit}.

This function actually terminates the process by raising a
@code{SIGABRT} signal, and your program can include a handler to
intercept this signal; see @ref{Signal Handling}.
@end deftypefun

@c Put in by rms.  Don't remove.
@cartouche
@strong{Future Change Warning:} Proposed Federal censorship regulations
may prohibit us from giving you information about the possibility of
calling this function.  We would be required to say that this is not an
acceptable way of terminating a program.
@end cartouche

@node Termination Internals
@subsection Termination Internals

The @code{_exit} function is the primitive used for process termination
by @code{exit}.  It is declared in the header file @file{unistd.h}.
@pindex unistd.h

@comment unistd.h
@comment POSIX.1
@deftypefun void _exit (int @var{status})
The @code{_exit} function is the primitive for causing a process to
terminate with status @var{status}.  Calling this function does not
execute cleanup functions registered with @code{atexit} or
@code{on_exit}.
@end deftypefun

When a process terminates for any reason---either by an explicit
termination call, or termination as a result of a signal---the
following things happen:

@itemize @bullet
@item
All open file descriptors in the process are closed.  @xref{Low-Level I/O}.
Note that streams are not flushed automatically when the process
terminates; @xref{I/O on Streams}.

@item
The low-order 8 bits of the return status code are saved to be reported
back to the parent process via @code{wait} or @code{waitpid}; see
@ref{Process Completion}.

@item
Any child processes of the process being terminated are assigned a new
parent process.  (On most systems, including GNU, this is the @code{init}
process, with process ID 1.)

@item
A @code{SIGCHLD} signal is sent to the parent process.

@item
If the process is a session leader that has a controlling terminal, then
a @code{SIGHUP} signal is sent to each process in the foreground job,
and the controlling terminal is disassociated from that session.
@xref{Job Control}.

@item
If termination of a process causes a process group to become orphaned,
and any member of that process group is stopped, then a @code{SIGHUP}
signal and a @code{SIGCONT} signal are sent to each process in the
group.  @xref{Job Control}.
@end itemize