Seth Woolley's Man Viewer

perlipc(1) - perlipc - Perl interprocess communication (signals, fifos, pipes, safe subprocesses, sockets, and semaphores) - man 1 perlipc

([section] manual, -k keyword, -K [section] search, -f whatis)
man plain no title

PERLIPC(1)             Perl Programmers Reference Guide             PERLIPC(1)

       perlipc - Perl interprocess communication (signals, fifos, pipes, safe
       subprocesses, sockets, and semaphores)

       The basic IPC facilities of Perl are built out of the good old Unix
       signals, named(5,8) pipes, pipe(2,8) opens, the Berkeley socket(2,7,n) routines, and
       SysV IPC calls.  Each is used in(1,8) slightly different situations.

       Perl uses a simple signal(2,7) handling model: the %SIG hash contains names
       or references of user-installed signal(2,7) handlers.  These handlers will
       be called with an argument which is the name of the signal(2,7) that trig-
       gered it.  A signal(2,7) may be generated intentionally from a particular
       keyboard sequence like control-C or control-Z, sent to you from another
       process, or triggered automatically by the kernel when special events
       transpire, like a child process exiting, your process running out of
       stack space, or hitting file(1,n) size limit.

       For example, to trap an interrupt signal(2,7), set(7,n,1 builtins) up a handler like this:

           sub catch_zap {
               my $signame = shift;
               die "Somebody sent me a SIG$signame";
           $SIG{INT} = 'catch_zap';  # could fail in(1,8) modules
           $SIG{INT} = \&catch_zap;  # best strategy

       Prior to Perl 5.7.3 it was necessary to do as little as you possibly
       could in(1,8) your handler; notice how all we do is set(7,n,1 builtins) a global variable
       and then raise(3,n) an exception.  That's because on most systems, libraries
       are not re-entrant; particularly, memory allocation and I/O routines
       are not.  That meant that doing nearly anything in(1,8) your handler could
       in(1,8) theory trigger a memory fault and subsequent core dump - see
       "Deferred Signals (Safe Signals)" below.

       The names of the signals are the ones listed out by "kill(1,2,1 builtins) -l" on your
       system, or you can retrieve them from the Config module.  Set up an
       @signame list indexed by number to get the name and a %signo table
       indexed by name to get the number:

           use Config;
           defined $Config{sig_name} || die "No sigs?";
           foreach $name (split(1,n)(' ', $Config{sig_name})) {
               $signo{$name} = $i;
               $signame[$i] = $name;

       So to check whether signal(2,7) 17 and SIGALRM were the same, do just this:

           print "signal(2,7) #17 = $signame[17]\n";
           if(3,n) ($signo{ALRM}) {
               print "SIGALRM is $signo{ALRM}\n";

       You may also choose to assign the strings 'IGNORE' or 'DEFAULT' as the
       handler, in(1,8) which case Perl will try to discard the signal(2,7) or do the
       default thing.

       On most Unix platforms, the "CHLD" (sometimes also known as "CLD") sig-
       nal(2,7) has special behavior with respect to a value of 'IGNORE'.  Setting
       $SIG{CHLD} to 'IGNORE' on such a platform has the effect of not creat-
       ing zombie processes when the parent process fails to "wait()" on its
       child processes (i.e. child processes are automatically reaped).  Call-
       ing "wait()" with $SIG{CHLD} set(7,n,1 builtins) to 'IGNORE' usually returns "-1" on
       such platforms.

       Some signals can be neither trapped nor ignored, such as the KILL and
       STOP (but not the TSTP) signals.  One strategy for temporarily ignoring
       signals is to use a local() statement, which will be automatically
       restored once your block is exited.  (Remember that local() values are
       "inherited" by functions called from within that block.)

           sub precious {
               local $SIG{INT} = 'IGNORE';
           sub more_functions {
               # interrupts still ignored, for now...

       Sending a signal(2,7) to a negative process ID means that you send(2,n) the sig-
       nal(2,7) to the entire Unix process-group.  This code sends a hang-up signal(2,7)
       to all processes in(1,8) the current process group (and sets $SIG{HUP} to
       IGNORE so it doesn't kill(1,2,1 builtins) itself):

               local $SIG{HUP} = 'IGNORE';
               kill(1,2,1 builtins) HUP => -$$;
               # snazzy writing of: kill(1,2,1 builtins)('HUP', -$$)

       Another interesting signal(2,7) to send(2,n) is signal(2,7) number zero.  This doesn't
       actually affect a child process, but instead checks whether it's alive
       or has changed its UID.

           unless (kill(1,2,1 builtins) 0 => $kid_pid) {
               warn "something wicked happened to $kid_pid";

       When directed at a process whose UID is not identical to that of the
       sending process, signal(2,7) number zero may fail because you lack permis-
       sion to send(2,n) the signal(2,7), even though the process is alive.  You may be
       able to determine the cause of failure using "%!".

           unless (kill(1,2,1 builtins) 0 => $pid or $!{EPERM}) {
               warn "$pid looks dead";

       You might also want to employ anonymous functions for simple signal(2,7)

           $SIG{INT} = sub { die "\nOutta here!\n" };

       But that will be problematic for the more complicated handlers that
       need to reinstall themselves.  Because Perl's signal(2,7) mechanism is cur-
       rently based on the signal(2,7)(3) function from the C library, you may
       sometimes be so misfortunate as to run on systems where that function
       is "broken", that is, it behaves in(1,8) the old unreliable SysV way rather
       than the newer, more reasonable BSD and POSIX fashion.  So you'll see
       defensive people writing signal(2,7) handlers like this:

           sub REAPER {
               $waitedpid = wait;
               # loathe sysV: it makes us not only reinstate
               # the handler, but place it after the wait
               $SIG{CHLD} = \&REAPER;
           $SIG{CHLD} = \&REAPER;
           # now do something that forks...

       or better still:

           use POSIX ":sys_wait_h";
           sub REAPER {
               my $child;
               # If a second child dies while in(1,8) the signal(2,7) handler caused by the
               # first death, we won't get another signal. So must loop here else
               # we will leave the unreaped child as a zombie. And the next time(1,2,n)
               # two children die we get another zombie. And so on.
               while (($child = waitpid(-1,WNOHANG)) > 0) {
                   $Kid_Status{$child} = $?;
               $SIG{CHLD} = \&REAPER;  # still loathe sysV
           $SIG{CHLD} = \&REAPER;
           # do something that forks...

       Signal handling is also used for timeouts in(1,8) Unix,   While safely pro-
       tected within an "eval{}" block, you set(7,n,1 builtins) a signal(2,7) handler to trap alarm(1,2)
       signals and then schedule to have one delivered to you in(1,8) some number
       of seconds.  Then try your blocking operation, clearing the alarm(1,2) when
       it's done but not before you've exited your "eval{}" block.  If it goes
       off, you'll use die() to jump out of the block, much as you might using
       longjmp() or throw() in(1,8) other languages.

       Here's an example:

           eval {
               local $SIG{ALRM} = sub { die "alarm(1,2) clock(3,n) restart" };
               alarm(1,2) 10;
               flock(1,2)(FH, 2);   # blocking write(1,2) lock
               alarm(1,2) 0;
           if(3,n) ($@ and $@ !~ /alarm(1,2) clock(3,n) restart/) { die }

       If the operation being timed out is system() or qx(), this technique is
       liable to generate zombies.    If this matters to you, you'll need to
       do your own fork() and exec(3,n,1 builtins)(), and kill(1,2,1 builtins) the errant child process.

       For more complex signal(2,7) handling, you might see the standard POSIX mod-
       ule.  Lamentably, this is almost entirely undocumented(2,3), but the
       t/lib/posix.t file(1,n) from the Perl source distribution has some examples
       in(1,8) it.

       Handling the SIGHUP Signal in(1,8) Daemons

       A process that usually starts when the system boots and shuts down when
       the system is shut down is called a daemon (Disk And Execution MONi-
       tor). If a daemon process has a configuration file(1,n) which is modified
       after the process has been started, there should be a way to tell that
       process to re-read its configuration file(1,n), without stopping the
       process. Many daemons provide this mechanism using the "SIGHUP" signal(2,7)
       handler. When you want to tell the daemon to re-read the file(1,n) you sim-
       ply send(2,n) it the "SIGHUP" signal.

       Not all platforms automatically reinstall their (native) signal(2,7) han-
       dlers after a signal(2,7) delivery.  This means that the handler works only
       the first time(1,2,n) the signal(2,7) is sent. The solution to this problem is to
       use "POSIX" signal(2,7) handlers if(3,n) available, their behaviour is

       The following example implements a simple daemon, which restarts itself
       every time(1,2,n) the "SIGHUP" signal(2,7) is received. The actual code is located
       in(1,8) the subroutine "code()", which simply prints some debug info(1,5,n) to show
       that it works and should be replaced with the real code.

         #!/usr/bin/perl -w

         use POSIX ();
         use FindBin ();
         use File::Basename ();
         use File::Spec::Functions;


         # make the daemon cross-platform, so exec(3,n,1 builtins) always calls the script
         # itself with the right path, no matter how the script was invoked.
         my $script = File::Basename::basename($0);
         my $SELF = catfile $FindBin::Bin, $script;

         # POSIX unmasks the sigprocmask properly
         my $sigset = POSIX::SigSet->new();
         my $action = POSIX::SigAction->new('sigHUP_handler',
         POSIX::sigaction(&POSIX::SIGHUP, $action);

         sub sigHUP_handler {
             print "got SIGHUP\n";
             exec(3,n,1 builtins)($SELF, @ARGV) or die "Couldn't restart: $!\n";


         sub code {
             print "PID: $$\n";
             print "ARGV: @ARGV\n";
             my $c = 0;
             while (++$c) {
                 sleep(1,3) 2;
                 print "$c\n";

Named Pipes
       A named(5,8) pipe(2,8) (often referred to as a FIFO) is an old Unix IPC mechanism
       for processes communicating on the same machine.  It works just like a
       regular, connected anonymous pipes, except that the processes ren-
       dezvous using a filename and don't have to be related.

       To create a named(5,8) pipe(2,8), use the Unix command mknod(1,2)(1) or on some sys-
       tems, mkfifo(1,3)(1).  These may not be in(1,8) your normal path.

           # system return val is backwards, so && not ||
           $ENV{PATH} .= ":/etc:/usr/etc";
           if(3,n)  (      system('mknod(1,2)',  $path, 'p')
                   && system('mkfifo(1,3)', $path) )
               die "mk{nod,fifo} $path failed";

       A fifo is convenient when you want to connect a process to an unrelated
       one.  When you open(2,3,n) a fifo, the program will block until there's some-
       thing on the other end.

       For example, let's say you'd like to have your .signature file(1,n) be a
       named(5,8) pipe(2,8) that has a Perl program on the other end.  Now every time(1,2,n)
       any program (like a mailer, news reader, finger program, etc.) tries to
       read(2,n,1 builtins) from that file(1,n), the reading program will block and your program
       will supply the new signature.  We'll use the pipe-checking file(1,n) test
       -p to find out whether anyone (or anything) has accidentally removed
       our fifo.

           chdir; # go home
           $FIFO = '.signature';
           $ENV{PATH} .= ":/etc:/usr/games";

           while (1) {
               unless (-p $FIFO) {
                   unlink(1,2) $FIFO;
                   system('mknod(1,2)', $FIFO, 'p')
                       && die "can't mknod(1,2) $FIFO: $!";

               # next line blocks until there's a reader
               open(2,3,n) (FIFO, "> $FIFO") || die "can't write(1,2) $FIFO: $!";
               print FIFO "John Smith (smith\\n", `fortune -s`;
               close(2,7,n) FIFO;
               sleep(1,3) 2;    # to avoid dup signals

       Deferred Signals (Safe Signals)

       In Perls before Perl 5.7.3 by installing Perl code to deal with sig-
       nals, you were exposing yourself to danger from two things.  First, few
       system library functions are re-entrant.  If the signal(2,7) interrupts
       while Perl is executing one function (like malloc(3) or printf(1,3,1 builtins)(3)), and
       your signal(2,7) handler then calls the same function again, you could get
       unpredictable behavior--often, a core dump.  Second, Perl isn't itself
       re-entrant at the lowest levels.  If the signal(2,7) interrupts Perl while
       Perl is changing its own internal data structures, similarly unpre-
       dictable behaviour may result.

       There were two things you could do, knowing this: be paranoid or be
       pragmatic.  The paranoid approach was to do as little as possible in(1,8)
       your signal(2,7) handler.  Set an existing integer variable that already has
       a value, and return.  This doesn't help you if(3,n) you're in(1,8) a slow system
       call, which will just restart.  That means you have to "die" to
       longjump(3) out of the handler.  Even this is a little cavalier for the
       true paranoiac, who avoids "die" in(1,8) a handler because the system is out
       to get you.  The pragmatic approach was to say ``I know the risks, but
       prefer the convenience'', and to do anything you wanted in(1,8) your signal(2,7)
       handler, and be prepared to clean up core dumps now and again.

       In Perl 5.7.3 and later to avoid these problems signals are
       "deferred"-- that is when the signal(2,7) is delivered to the process by the
       system (to the C code that implements Perl) a flag is set(7,n,1 builtins), and the han-
       dler returns immediately. Then at strategic "safe" points in(1,8) the Perl
       interpreter (e.g. when it is about to execute a new opcode) the flags
       are checked and the Perl level handler from %SIG is executed. The
       "deferred" scheme allows much more flexibility in(1,8) the coding of signal(2,7)
       handler as we know Perl interpreter is in(1,8) a safe state, and that we are
       not in(1,8) a system library function when the handler is called.  However
       the implementation does differ from previous Perls in(1,8) the following

       Long running opcodes
           As Perl interpreter only looks at the signal(2,7) flags when it about to
           execute a new opcode if(3,n) a signal(2,7) arrives during a long running
           opcode (e.g. a regular expression operation on a very large string(3,n))
           then signal(2,7) will not be seen until operation completes.

       Interrupting IO
           When a signal(2,7) is delivered (e.g. INT control-C) the operating sys-
           tem breaks into IO operations like "read(2,n,1 builtins)" (used to implement Perls
           <> operator). On older Perls the handler was called immediately
           (and as "read(2,n,1 builtins)" is not "unsafe" this worked well). With the
           "deferred" scheme the handler is not called immediately, and if(3,n)
           Perl is using system's "stdio" library that library may re-start
           the "read(2,n,1 builtins)" without returning to Perl and giving it a chance to call
           the %SIG handler. If this happens on your system the solution is to
           use ":perlio" layer to do IO - at least on those handles which you
           want to be able to break into with signals. (The ":perlio" layer
           checks the signal(2,7) flags and calls %SIG handlers before resuming IO

           Note that the default in(1,8) Perl 5.7.3 and later is to automatically
           use the ":perlio" layer.

           Note that some networking library functions like gethostbyname()
           are known to have their own implementations of timeouts which may
           conflict with your timeouts.  If you are having problems with such
           functions, you can try using the POSIX sigaction() function, which
           bypasses the Perl safe signals (note that this means subjecting
           yourself to possible memory corruption, as described above).
           Instead of setting $SIG{ALRM} try something like the following:

               use POSIX;
               sigaction SIGALRM, new POSIX::SigAction sub { die "alarm(1,2)\n" }
                   or die "Error setting SIGALRM handler: $!\n";

       Restartable system calls
           On systems that supported it, older versions of Perl used the
           SA_RESTART flag when installing %SIG handlers.  This meant that
           restartable system calls would continue rather than returning when
           a signal(2,7) arrived.  In order to deliver deferred signals promptly,
           Perl 5.7.3 and later do not use SA_RESTART.  Consequently,
           restartable system calls can fail (with $! set(7,n,1 builtins) to "EINTR") in(1,8)
           places where they previously would have succeeded.

           Note that the default ":perlio" layer will retry "read(2,n,1 builtins)", "write(1,2)"
           and "close(2,7,n)" as described above and that interrupted "wait" and
           "waitpid" calls will always be retried.

       Signals as "faults"
           Certain signals e.g. SEGV, ILL, BUS are generated as a result of
           virtual(5,8) memory or other "faults". These are normally fatal and
           there is little a Perl-level handler can do with them. (In particu-
           lar the old signal(2,7) scheme was particularly unsafe in(1,8) such cases.)
           However if(3,n) a %SIG handler is set(7,n,1 builtins) the new scheme simply sets a flag
           and returns as described above. This may cause the operating system
           to try the offending machine instruction again and - as nothing has
           changed - it will generate the signal(2,7) again. The result of this is
           a rather odd "loop". In future Perl's signal(2,7) mechanism may be
           changed to avoid this - perhaps by simply disallowing %SIG handlers
           on signals of that type. Until then the work-round is not to set(7,n,1 builtins) a
           %SIG handler on those signals. (Which signals they are is operating
           system dependant.)

       Signals triggered by operating system state
           On some operating systems certain signal(2,7) handlers are supposed to
           "do something" before returning. One example can be CHLD or CLD
           which indicates a child process has completed. On some operating
           systems the signal(2,7) handler is expected to "wait" for the completed
           child process. On such systems the deferred signal(2,7) scheme will not
           work for those signals (it does not do the "wait"). Again the fail-
           ure will look(1,8,3 Search::Dict) like a loop as the operating system will re-issue the
           signal(2,7) as there are un-waited-for completed child processes.

       If you want the old signal(2,7) behaviour back regardless of possible memory
       corruption, set(7,n,1 builtins) the environment variable "PERL_SIGNALS" to "unsafe" (a
       new feature since Perl 5.8.1).

Using open(2,3,n)() for IPC
       Perl's basic open(2,3,n)() statement can also be used for unidirectional
       interprocess communication by either appending or prepending a pipe(2,8)
       symbol to the second argument to open(2,3,n)().  Here's how to start something
       up in(1,8) a child process you intend to write(1,2) to:

           open(2,3,n)(SPOOLER, "| cat -v | lpr -h 2>/dev/null")
                           || die "can't fork: $!";
           local $SIG{PIPE} = sub { die "spooler pipe(2,8) broke" };
           print SPOOLER "stuff\n";
           close(2,7,n) SPOOLER || die "bad spool: $! $?";

       And here's how to start up a child process you intend to read(2,n,1 builtins) from:

           open(2,3,n)(STATUS, "netstat -an 2>&1 |")
                           || die "can't fork: $!";
           while (<STATUS>) {
               next if(3,n) /^(tcp|udp)/;
           close(2,7,n) STATUS || die "bad netstat: $! $?";

       If one can be sure that a particular program is a Perl script that is
       expecting filenames in(1,8) @ARGV, the clever programmer can write(1,2) something
       like this:

           % program f1 "cmd1|" - f2 "cmd2|" f3 < tmpfile

       and irrespective of which shell it's called from, the Perl program will
       read(2,n,1 builtins) from the file(1,n) f1, the process cmd1, standard input (tmpfile in(1,8)
       this case), the f2 file(1,n), the cmd2 command, and finally the f3 file.
       Pretty nifty, eh?

       You might notice that you could use backticks for much the same effect
       as opening a pipe(2,8) for reading:

           print grep { !/^(tcp|udp)/ } `netstat -an 2>&1`;
           die "bad netstat" if(3,n) $?;

       While this is true on the surface, it's much more efficient to process
       the file(1,n) one line or record at a time(1,2,n) because then you don't have to
       read(2,n,1 builtins) the whole thing into memory at once.  It also gives you finer con-
       trol of the whole process, letting you to kill(1,2,1 builtins) off the child process
       early if(3,n) you'd like.

       Be careful to check both the open(2,3,n)() and the close(2,7,n)() return values.  If
       you're writing to a pipe(2,8), you should also trap SIGPIPE.  Otherwise,
       think of what happens when you start up a pipe(2,8) to a command that
       doesn't exist: the open(2,3,n)() will in(1,8) all likelihood succeed (it only
       reflects the fork()'s success), but then your output will fail--spec-
       tacularly.  Perl can't know whether the command worked because your
       command is actually running in(1,8) a separate process whose exec(3,n,1 builtins)() might
       have failed.  Therefore, while readers of bogus commands return just a
       quick end of file(1,n), writers to bogus command will trigger a signal(2,7)
       they'd better be prepared to handle.  Consider:

           open(2,3,n)(FH, "|bogus")  or die "can't fork: $!";
           print FH "bang\n"   or die "can't write: $!";
           close(2,7,n) FH            or die "can't close: $!";

       That won't blow up until the close(2,7,n), and it will blow up with a SIGPIPE.
       To catch it, you could use this:

           $SIG{PIPE} = 'IGNORE';
           open(2,3,n)(FH, "|bogus")  or die "can't fork: $!";
           print FH "bang\n"   or die "can't write: $!";
           close(2,7,n) FH            or die "can't close: status=$?";


       Both the main process and any child processes it forks share the same
       STDIN, STDOUT, and STDERR filehandles.  If both processes try to access(2,5)
       them at once, strange things can happen.  You may also want to close(2,7,n) or
       reopen the filehandles for the child.  You can get around this by open-
       ing your pipe(2,8) with open(2,3,n)(), but on some systems this means that the
       child process cannot outlive the parent.

       Background Processes

       You can run a command in(1,8) the background with:

           system("cmd &");

       The command's STDOUT and STDERR (and possibly STDIN, depending on your
       shell) will be the same as the parent's.  You won't need to catch
       SIGCHLD because of the double-fork taking place (see below for more

       Complete Dissociation of Child from Parent

       In some cases (starting server processes, for instance) you'll want to
       completely dissociate the child process from the parent.  This is often
       called daemonization.  A well behaved daemon will also chdir() to the
       root directory (so it doesn't prevent unmounting the filesystem con-
       taining the directory from which it was launched) and redirect its
       standard file(1,n) descriptors from and to /dev/null (so that random(3,4,6) output
       doesn't wind up on the user's terminal).

           use POSIX 'setsid(2,8)';

           sub daemonize {
               chdir '/'               or die "Can't chdir to /: $!";
               open(2,3,n) STDIN, '/dev/null' or die "Can't read(2,n,1 builtins) /dev/null: $!";
               open(2,3,n) STDOUT, '>/dev/null'
                                       or die "Can't write(1,2) to /dev/null: $!";
               defined(my $pid = fork) or die "Can't fork: $!";
               exit(3,n,1 builtins) if(3,n) $pid;
               setsid(2,8)                  or die "Can't start a new session: $!";
               open(2,3,n) STDERR, '>&STDOUT' or die "Can't dup stdout: $!";

       The fork() has to come before the setsid(2,8)() to ensure that you aren't a
       process group leader (the setsid(2,8)() will fail if(3,n) you are).  If your sys-
       tem doesn't have the setsid(2,8)() function, open(2,3,n) /dev/tty(1,4) and use the
       "TIOCNOTTY" ioctl() on it instead.  See tty(1,4)(4) for details.

       Non-Unix users(1,5) should check their Your_OS::Process module for other

       Safe Pipe Opens

       Another interesting approach to IPC is making your single program go
       multiprocess and communicate between (or even amongst) yourselves.  The
       open(2,3,n)() function will accept(2,8) a file(1,n) argument of either "-|" or "|-" to
       do a very interesting thing: it forks a child connected to the filehan-
       dle you've opened.  The child is running the same program as the par-
       ent.  This is useful for safely opening a file(1,n) when running under an
       assumed UID or GID, for example.  If you open(2,3,n) a pipe(2,8) to minus, you can
       write(1,2) to the filehandle you opened and your kid will find it in(1,8) his
       STDIN.  If you open(2,3,n) a pipe(2,8) from minus, you can read(2,n,1 builtins) from the filehandle
       you opened whatever your kid writes to his STDOUT.

           use English '-no_match_vars';
           my $sleep_count = 0;

           do {
               $pid = open(2,3,n)(KID_TO_WRITE, "|-");
               unless (defined $pid) {
                   warn "cannot fork: $!";
                   die "bailing out" if(3,n) $sleep_count++ > 6;
                   sleep(1,3) 10;
           } until defined $pid;

           if(3,n) ($pid) {  # parent
               print KID_TO_WRITE @some_data;
               close(2,7,n)(KID_TO_WRITE) || warn "kid exited $?";
           } else {     # child
               ($EUID, $EGID) = ($UID, $GID); # suid progs only
               open(2,3,n) (FILE, "> /safe/file(1,n)")
                   || die "can't open(2,3,n) /safe/file: $!";
               while (<STDIN>) {
                   print FILE; # child's STDIN is parent's KID
               exit(3,n,1 builtins);  # don't forget this

       Another common use for this construct is when you need to execute some-
       thing without the shell's interference.  With system(), it's straight-
       forward, but you can't use a pipe(2,8) open(2,3,n) or backticks safely.  That's
       because there's no way to stop the shell from getting its hands on your
       arguments.   Instead, use lower-level control to call exec(3,n,1 builtins)() directly.

       Here's a safe backtick or pipe(2,8) open(2,3,n) for read:

           # add error(8,n) processing as above
           $pid = open(2,3,n)(KID_TO_READ, "-|");

           if(3,n) ($pid) {   # parent
               while (<KID_TO_READ>) {
                   # do something interesting
               close(2,7,n)(KID_TO_READ) || warn "kid exited $?";

           } else {      # child
               ($EUID, $EGID) = ($UID, $GID); # suid only
               exec(3,n,1 builtins)($program, @options, @args)
                   || die "can't exec(3,n,1 builtins) program: $!";
               # NOTREACHED

       And here's a safe pipe(2,8) open(2,3,n) for writing:

           # add error(8,n) processing as above
           $pid = open(2,3,n)(KID_TO_WRITE, "|-");
           $SIG{PIPE} = sub { die "whoops, $program pipe(2,8) broke" };

           if(3,n) ($pid) {  # parent
               for (@data) {
                   print KID_TO_WRITE;
               close(2,7,n)(KID_TO_WRITE) || warn "kid exited $?";

           } else {     # child
               ($EUID, $EGID) = ($UID, $GID);
               exec(3,n,1 builtins)($program, @options, @args)
                   || die "can't exec(3,n,1 builtins) program: $!";
               # NOTREACHED

       Since Perl 5.8.0, you can also use the list form of "open(2,3,n)" for pipes :
       the syntax

           open(2,3,n) KID_PS, "-|", "ps", "aux" or die $!;

       forks the ps(1) command (without spawning a shell, as there are more
       than three arguments to open(2,3,n)()), and reads its standard output via the
       "KID_PS" filehandle.  The corresponding syntax to read(2,n,1 builtins) from command
       pipes (with "|-" in(1,8) place of "-|") is also implemented.

       Note that these operations are full Unix forks, which means they may
       not be correctly implemented on alien systems.  Additionally, these are
       not true multithreading.  If you'd like to learn more about threading,
       see the modules file(1,n) mentioned below in(1,8) the SEE ALSO section.

       Bidirectional Communication with Another Process

       While this works reasonably well for unidirectional communication, what
       about bidirectional communication?  The obvious thing you'd like to do
       doesn't actually work:

           open(2,3,n)(PROG_FOR_READING_AND_WRITING, "| some program |")

       and if(3,n) you forget to use the "use warnings" pragma or the -w flag, then
       you'll miss out entirely on the diagnostic message:

           Can't do bidirectional pipe(2,8) at -e line 1.

       If you really want to, you can use the standard open2() library func-
       tion to catch both ends.  There's also an open3() for tridirectional
       I/O so you can also catch your child's STDERR, but doing so would then
       require an awkward select(2,7,2 select_tut)() loop and wouldn't allow you to use normal
       Perl input operations.

       If you look(1,8,3 Search::Dict) at its source, you'll see that open2() uses low-level prim-
       itives like Unix pipe(2,8)() and exec(3,n,1 builtins)() calls to create all the connections.
       While it might have been slightly more efficient by using socketpair(),
       it would have then been even less(1,3) portable than it already is.  The
       open2() and open3() functions are  unlikely to work anywhere except on
       a Unix system or some other one purporting to be POSIX compliant.

       Here's an example of using open2():

           use FileHandle;
           use IPC::Open2;
           $pid = open2(*Reader, *Writer, "cat -u -n" );
           print Writer "stuff\n";
           $got = <Reader>;

       The problem with this is that Unix buffering is really going to ruin
       your day.  Even though your "Writer" filehandle is auto-flushed, and
       the process on the other end will get your data in(1,8) a timely manner, you
       can't usually do anything to force it to give it back to you in(1,8) a simi-
       larly quick fashion.  In this case, we could, because we gave cat a -u
       flag to make it unbuffered.  But very few Unix commands are designed to
       operate over pipes, so this seldom works unless you yourself wrote the
       program on the other end of the double-ended pipe.

       A solution to this is the nonstandard library.  It uses pseudo-
       ttys to make your program behave more reasonably:

           require '';
           $ph = open_proc('cat -n');
           for (1..10) {
               print $ph "a line\n";
               print "got back ", scalar <$ph>;

       This way you don't have to have control over the source code of the
       program you're using.  The Comm library also has expect() and inter-
       act() functions.  Find the library (and we hope its successor
       IPC::Chat) at your nearest CPAN archive as detailed in(1,8) the SEE ALSO
       section below.

       The newer module from CPAN also addresses this kind of thing.
       This module requires two other modules from CPAN: IO::Pty and IO::Stty.
       It sets up a pseudo-terminal to interact with programs that insist on
       using talking to the terminal device driver.  If your system is amongst
       those supported, this may be your best bet.

       Bidirectional Communication with Yourself

       If you want, you may make low-level pipe(2,8)() and fork() to stitch this
       together by hand.  This example only talks to itself, but you could
       reopen the appropriate handles to STDIN and STDOUT and call other pro-

           #!/usr/bin/perl -w
           # pipe1 - bidirectional communication using two pipe(2,8) pairs
           #         designed for the socketpair-challenged
           use IO::Handle;     # thousands of lines just for autoflush :-(
           pipe(2,8)(PARENT_RDR, CHILD_WTR);                # XXX: failure?
           pipe(2,8)(CHILD_RDR,  PARENT_WTR);               # XXX: failure?

           if(3,n) ($pid = fork) {
               close(2,7,n) PARENT_RDR; close(2,7,n) PARENT_WTR;
               print CHILD_WTR "Parent Pid $$ is sending this\n";
               chomp($line = <CHILD_RDR>);
               print "Parent Pid $$ just read(2,n,1 builtins) this: `$line'\n";
               close(2,7,n) CHILD_RDR; close(2,7,n) CHILD_WTR;
           } else {
               die "cannot fork: $!" unless defined $pid;
               close(2,7,n) CHILD_RDR; close(2,7,n) CHILD_WTR;
               chomp($line = <PARENT_RDR>);
               print "Child Pid $$ just read(2,n,1 builtins) this: `$line'\n";
               print PARENT_WTR "Child Pid $$ is sending this\n";
               close(2,7,n) PARENT_RDR; close(2,7,n) PARENT_WTR;
               exit(3,n,1 builtins);

       But you don't actually have to make two pipe(2,8) calls.  If you have the
       socketpair() system call, it will do this all for you.

           #!/usr/bin/perl -w
           # pipe2 - bidirectional communication using socketpair
           #   "the best ones always go both ways"

           use Socket;
           use IO::Handle;     # thousands of lines just for autoflush :-(
           # We say AF_UNIX because although *_LOCAL is the
           # POSIX 1003.1g form of the constant, many machines
           # still don't have it.
                                       or  die "socketpair: $!";


           if(3,n) ($pid = fork) {
               close(2,7,n) PARENT;
               print CHILD "Parent Pid $$ is sending this\n";
               chomp($line = <CHILD>);
               print "Parent Pid $$ just read(2,n,1 builtins) this: `$line'\n";
               close(2,7,n) CHILD;
           } else {
               die "cannot fork: $!" unless defined $pid;
               close(2,7,n) CHILD;
               chomp($line = <PARENT>);
               print "Child Pid $$ just read(2,n,1 builtins) this: `$line'\n";
               print PARENT "Child Pid $$ is sending this\n";
               close(2,7,n) PARENT;
               exit(3,n,1 builtins);

Sockets: Client/Server Communication
       While not limited to Unix-derived operating systems (e.g., WinSock on
       PCs provides socket(2,7,n) support, as do some VMS libraries), you may not
       have sockets on your system, in(1,8) which case this section probably isn't
       going to do you much good.  With sockets, you can do both virtual(5,8) cir-
       cuits (i.e., TCP streams) and datagrams (i.e., UDP packets).  You may
       be able to do even more depending on your system.

       The Perl function calls for dealing with sockets have the same names as
       the corresponding system calls in(1,8) C, but their arguments tend to differ
       for two reasons: first, Perl filehandles work differently than C file(1,n)
       descriptors.  Second, Perl already knows the length of its strings, so
       you don't need to pass that information.

       One of the major problems with old socket(2,7,n) code in(1,8) Perl was that it used
       hard-coded values for some of the constants, which severely hurt porta-
       bility.  If you ever see code that does anything like explicitly set-
       ting "$AF_INET = 2", you know you're in(1,8) for big trouble:  An immeasur-
       ably superior approach is to use the "Socket" module, which more reli-
       ably grants access(2,5) to various constants and functions you'll need.

       If you're not writing a server/client for an existing protocol like
       NNTP or SMTP, you should give some thought to how your server will know
       when the client has finished talking, and vice-versa.  Most protocols
       are based on one-line messages and responses (so one party knows the
       other has finished when a "\n" is received) or multi-line messages and
       responses that end with a period on an empty line ("\n.\n" terminates a

       Internet Line Terminators

       The Internet line terminator is "\015\012".  Under ASCII variants of
       Unix, that could usually be written as "\r\n", but under other systems,
       "\r\n" might at times be "\015\015\012", "\012\012\015", or something
       completely different.  The standards specify writing "\015\012" to be
       conformant (be strict in(1,8) what you provide), but they also recommend
       accepting a lone "\012" on input (but be lenient in(1,8) what you require).
       We haven't always been very good about that in(1,8) the code in(1,8) this man-
       page, but unless you're on a Mac, you'll probably be ok.

       Internet TCP Clients and Servers

       Use Internet-domain sockets when you want to do client-server communi-
       cation that might extend to machines outside of your own system.

       Here's a sample TCP client using Internet-domain sockets:

           #!/usr/bin/perl -w
           use strict;
           use Socket;
           my ($remote,$port, $iaddr, $paddr, $proto, $line);

           $remote  = shift || 'localhost';
           $port    = shift || 2345;  # random(3,4,6) port
           if(3,n) ($port =~ /\D/) { $port = getservbyname($port, 'tcp') }
           die "No port" unless $port;
           $iaddr   = inet_aton($remote)               || die "no host: $remote";
           $paddr   = sockaddr_in($port, $iaddr);

           $proto   = getprotobyname('tcp');
           socket(2,7,n)(SOCK, PF_INET, SOCK_STREAM, $proto)  || die "socket: $!";
           connect(SOCK, $paddr)    || die "connect: $!";
           while (defined($line = <SOCK>)) {
               print $line;

           close(2,7,n) (SOCK)            || die "close: $!";
           exit(3,n,1 builtins);

       And here's a corresponding server to go along with it.  We'll leave the
       address as INADDR_ANY so that the kernel can choose the appropriate
       interface on multihomed hosts.  If you want sit on a particular inter-
       face (like the external side of a gateway or firewall machine), you
       should fill this in(1,8) with your real address instead.

           #!/usr/bin/perl -Tw
           use strict;
           BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
           use Socket;
           use Carp;
           my $EOL = "\015\012";

           sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }

           my $port = shift || 2345;
           my $proto = getprotobyname('tcp');

           ($port) = $port =~ /^(\d+)$/                        or die "invalid port";

           socket(2,7,n)(Server, PF_INET, SOCK_STREAM, $proto)        || die "socket: $!";
           setsockopt(Server, SOL_SOCKET, SO_REUSEADDR,
                                               pack(3,n,n pack-old)("l", 1))   || die "setsockopt: $!";
           bind(2,n,1 builtins)(Server, sockaddr_in($port, INADDR_ANY))        || die "bind: $!";
           listen(1,2,7)(Server,SOMAXCONN)                            || die "listen: $!";

           logmsg "server started on port $port";

           my $paddr;

           $SIG{CHLD} = \&REAPER;

           for ( ; $paddr = accept(2,8)(Client,Server); close(2,7,n) Client) {
               my($port,$iaddr) = sockaddr_in($paddr);
               my $name = gethostbyaddr($iaddr,AF_INET);

               logmsg "connection from $name [",
                       inet_ntoa($iaddr), "]
                       at port $port";

               print Client "Hello there, $name, it's now ",
                               scalar localtime, $EOL;

       And here's a multithreaded version.  It's multithreaded in(1,8) that like
       most typical servers, it spawns (forks) a slave server to handle the
       client request so that the master(5,8) server can quickly go back to service
       a new client.

           #!/usr/bin/perl -Tw
           use strict;
           BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
           use Socket;
           use Carp;
           my $EOL = "\015\012";

           sub spawn;  # forward declaration
           sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }

           my $port = shift || 2345;
           my $proto = getprotobyname('tcp');

           ($port) = $port =~ /^(\d+)$/                        or die "invalid port";

           socket(2,7,n)(Server, PF_INET, SOCK_STREAM, $proto)        || die "socket: $!";
           setsockopt(Server, SOL_SOCKET, SO_REUSEADDR,
                                               pack(3,n,n pack-old)("l", 1))   || die "setsockopt: $!";
           bind(2,n,1 builtins)(Server, sockaddr_in($port, INADDR_ANY))        || die "bind: $!";
           listen(1,2,7)(Server,SOMAXCONN)                            || die "listen: $!";

           logmsg "server started on port $port";

           my $waitedpid = 0;
           my $paddr;

           use POSIX ":sys_wait_h";
           sub REAPER {
               my $child;
               while (($waitedpid = waitpid(-1,WNOHANG)) > 0) {
                   logmsg "reaped $waitedpid" . ($? ? " with exit(3,n,1 builtins) $?" : '');
               $SIG{CHLD} = \&REAPER;  # loathe sysV

           $SIG{CHLD} = \&REAPER;

           for ( $waitedpid = 0;
                 ($paddr = accept(2,8)(Client,Server)) || $waitedpid;
                 $waitedpid = 0, close(2,7,n) Client)
               next if(3,n) $waitedpid and not $paddr;
               my($port,$iaddr) = sockaddr_in($paddr);
               my $name = gethostbyaddr($iaddr,AF_INET);

               logmsg "connection from $name [",
                       inet_ntoa($iaddr), "]
                       at port $port";

               spawn sub {
                   print "Hello there, $name, it's now ", scalar localtime, $EOL;
                   exec(3,n,1 builtins) '/usr/games/fortune'           # XXX: `wrong' line terminators
                       or confess "can't exec(3,n,1 builtins) fortune: $!";


           sub spawn {
               my $coderef = shift;

               unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') {
                   confess "usage: spawn CODEREF";

               my $pid;
               if(3,n) (!defined($pid = fork)) {
                   logmsg "cannot fork: $!";
               } elsif ($pid) {
                   logmsg "begat $pid";
                   return; # I'm the parent
               # else I'm the child -- go spawn

               open(2,3,n)(STDIN,  "<&Client")   || die "can't dup client to stdin";
               open(2,3,n)(STDOUT, ">&Client")   || die "can't dup client to stdout";
               ## open(2,3,n)(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
               exit(3,n,1 builtins) &$coderef();

       This server takes the trouble to clone off a child version(1,3,5) via fork()
       for each incoming request.  That way it can handle many requests at
       once, which you might not always want.  Even if(3,n) you don't fork(), the
       listen(1,2,7)() will allow that many pending connections.  Forking servers
       have to be particularly careful about cleaning up their dead children
       (called "zombies" in(1,8) Unix parlance), because otherwise you'll quickly
       fill up your process table.

       We suggest that you use the -T flag to use taint checking (see perlsec)
       even if(3,n) we aren't running setuid or setgid.  This is always a good idea
       for servers and other programs run on behalf of someone else (like CGI
       scripts), because it lessens the chances that people from the outside
       will be able to compromise your system.

       Let's look(1,8,3 Search::Dict) at another TCP client.  This one connects to the TCP "time(1,2,n)"
       service on a number of different machines and shows how far their
       clocks differ from the system on which it's being run:

           #!/usr/bin/perl  -w
           use strict;
           use Socket;

           my $SECS_of_70_YEARS = 2208988800;
           sub ctime { scalar localtime(shift) }

           my $iaddr = gethostbyname('localhost');
           my $proto = getprotobyname('tcp');
           my $port = getservbyname('time(1,2,n)', 'tcp');
           my $paddr = sockaddr_in(0, $iaddr);

           $| = 1;
           printf(1,3,1 builtins) "%-24s %8s %s\n",  "localhost", 0, ctime(time(1,2,n)());

           foreach $host(1,5) (@ARGV) {
               printf(1,3,1 builtins) "%-24s ", $host(1,5);
               my $hisiaddr = inet_aton($host(1,5))     || die "unknown host(1,5)";
               my $hispaddr = sockaddr_in($port, $hisiaddr);
               socket(2,7,n)(SOCKET, PF_INET, SOCK_STREAM, $proto)   || die "socket: $!";
               connect(SOCKET, $hispaddr)          || die "bind: $!";
               my $rtime = '    ';
               read(2,n,1 builtins)(SOCKET, $rtime, 4);
               my $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ;
               printf(1,3,1 builtins) "%8d %s\n", $histime - time(1,2,n), ctime($histime);

       Unix-Domain TCP Clients and Servers

       That's fine for Internet-domain clients and servers, but what about
       local communications?  While you can use the same setup(2,8), sometimes you
       don't want to.  Unix-domain sockets are local to the current host(1,5), and
       are often used internally to implement pipes.  Unlike Internet domain
       sockets, Unix domain sockets can show up in(1,8) the file(1,n) system with an
       ls(1) listing.

           % ls -l /dev/log
           srw-rw-rw-  1 root            0 Oct 31 07:23 /dev/log

       You can test for these with Perl's -S file(1,n) test:

           unless ( -S '/dev/log' ) {
               die "something's wicked with the log system";

       Here's a sample Unix-domain client:

           #!/usr/bin/perl -w
           use Socket;
           use strict;
           my ($rendezvous, $line);

           $rendezvous = shift || 'catsock';
           socket(2,7,n)(SOCK, PF_UNIX, SOCK_STREAM, 0)       || die "socket: $!";
           connect(SOCK, sockaddr_un($rendezvous))     || die "connect: $!";
           while (defined($line = <SOCK>)) {
               print $line;
           exit(3,n,1 builtins);

       And here's a corresponding server.  You don't have to worry about silly
       network terminators here because Unix domain sockets are guaranteed to
       be on the localhost, and thus everything works right.

           #!/usr/bin/perl -Tw
           use strict;
           use Socket;
           use Carp;

           BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
           sub spawn;  # forward declaration
           sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }

           my $NAME = 'catsock';
           my $uaddr = sockaddr_un($NAME);
           my $proto = getprotobyname('tcp');

           socket(2,7,n)(Server,PF_UNIX,SOCK_STREAM,0)        || die "socket: $!";
           bind(2,n,1 builtins)  (Server, $uaddr)                      || die "bind: $!";
           listen(1,2,7)(Server,SOMAXCONN)                    || die "listen: $!";

           logmsg "server started on $NAME";

           my $waitedpid;

           use POSIX ":sys_wait_h";
           sub REAPER {
               my $child;
               while (($waitedpid = waitpid(-1,WNOHANG)) > 0) {
                   logmsg "reaped $waitedpid" . ($? ? " with exit(3,n,1 builtins) $?" : '');
               $SIG{CHLD} = \&REAPER;  # loathe sysV

           $SIG{CHLD} = \&REAPER;

           for ( $waitedpid = 0;
                 accept(2,8)(Client,Server) || $waitedpid;
                 $waitedpid = 0, close(2,7,n) Client)
               next if(3,n) $waitedpid;
               logmsg "connection on $NAME";
               spawn sub {
                   print "Hello there, it's now ", scalar localtime, "\n";
                   exec(3,n,1 builtins) '/usr/games/fortune' or die "can't exec(3,n,1 builtins) fortune: $!";

           sub spawn {
               my $coderef = shift;

               unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') {
                   confess "usage: spawn CODEREF";

               my $pid;
               if(3,n) (!defined($pid = fork)) {
                   logmsg "cannot fork: $!";
               } elsif ($pid) {
                   logmsg "begat $pid";
                   return; # I'm the parent
               # else I'm the child -- go spawn

               open(2,3,n)(STDIN,  "<&Client")   || die "can't dup client to stdin";
               open(2,3,n)(STDOUT, ">&Client")   || die "can't dup client to stdout";
               ## open(2,3,n)(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
               exit(3,n,1 builtins) &$coderef();

       As you see, it's remarkably similar to the Internet domain TCP server,
       so much so, in(1,8) fact, that we've omitted several duplicate func-
       tions--spawn(), logmsg(), ctime(), and REAPER()--which are exactly the
       same as in(1,8) the other server.

       So why would you ever want to use a Unix domain socket(2,7,n) instead of a
       simpler named(5,8) pipe(2,8)?  Because a named(5,8) pipe(2,8) doesn't give you sessions.
       You can't tell one process's data from another's.  With socket(2,7,n) program-
       ming, you get a separate session for each client: that's why accept(2,8)()
       takes two arguments.

       For example, let's say that you have a long running database server
       daemon that you want folks from the World Wide Web to be able to
       access(2,5), but only if(3,n) they go through a CGI interface.  You'd have a
       small, simple CGI program that does whatever checks and logging you
       feel like, and then acts as a Unix-domain client and connects to your
       private server.

TCP Clients with IO::Socket
       For those preferring a higher-level interface to socket(2,7,n) programming,
       the IO::Socket module provides an object-oriented approach.  IO::Socket
       is included as part of the standard Perl distribution as of the 5.004
       release.  If you're running an earlier version(1,3,5) of Perl, just fetch
       IO::Socket from CPAN, where you'll also find modules providing easy
       interfaces to the following systems: DNS, FTP, Ident (RFC 931), NIS and
       NISPlus, NNTP, Ping, POP3, SMTP, SNMP, SSLeay, Telnet, and Time--just
       to name a few.

       A Simple Client

       Here's a client that creates a TCP connection to the "daytime" service
       at port 13 of the host(1,5) name "localhost" and prints out everything that
       the server there cares to provide.

           #!/usr/bin/perl -w
           use IO::Socket;
           $remote = IO::Socket::INET->new(
                               Proto    => "tcp",
                               PeerAddr => "localhost",
                               PeerPort => "daytime(13)",
                         or die "cannot connect to daytime port at localhost";
           while ( <$remote> ) { print }

       When you run this program, you should get something back that looks
       like this:

           Wed May 14 08:40:46 MDT 1997

       Here are what those parameters to the "new" constructor mean:

           This is which protocol to use.  In this case, the socket(2,7,n) handle
           returned will be connected to a TCP socket(2,7,n), because we want a
           stream-oriented connection, that is, one that acts pretty much like
           a plain old file.  Not all sockets are this of this type.  For
           example, the UDP protocol can be used to make a datagram socket(2,7,n),
           used for message-passing.

           This is the name or Internet address of the remote host(1,5) the server
           is running on.  We could have specified a longer name like
           "", or an address like "".  For demonstra-
           tion purposes, we've used the special hostname "localhost", which
           should always mean the current machine you're running on.  The cor-
           responding Internet address for localhost is "127.1", if(3,n) you'd
           rather use that.

           This is the service name or port number we'd like to connect to.
           We could have gotten away with using just "daytime" on systems with
           a well-configured system services file(1,n),[FOOTNOTE: The system ser-
           vices file(1,n) is in(1,8) /etc/services under Unix] but just in(1,8) case, we've
           specified the port number (13) in(1,8) parentheses.  Using just the num-
           ber would also have worked, but constant numbers make careful pro-
           grammers nervous.

       Notice how the return value from the "new" constructor is used as a
       filehandle in(1,8) the "while" loop?  That's what's called an indirect file-
       handle, a scalar variable containing a filehandle.  You can use it the
       same way you would a normal filehandle.  For example, you can read(2,n,1 builtins) one
       line from it this way:

           $line = <$handle>;

       all remaining lines from is this way:

           @lines = <$handle>;

       and send(2,n) a line of data to it this way:

           print $handle "some data\n";

       A Webget Client

       Here's a simple client that takes a remote host(1,5) to fetch a document
       from, and then a list of documents to get from that host.  This is a
       more interesting client than the previous one because it first sends
       something to the server before fetching the server's response.

           #!/usr/bin/perl -w
           use IO::Socket;
           unless (@ARGV > 1) { die "usage: $0 host(1,5) document ..." }
           $host(1,5) = shift(@ARGV);
           $EOL = "\015\012";
           $BLANK = $EOL x 2;
           foreach $document ( @ARGV ) {
               $remote = IO::Socket::INET->new( Proto     => "tcp",
                                                PeerAddr  => $host(1,5),
                                                PeerPort  => "http(80)",
               unless ($remote) { die "cannot connect to http daemon on $host(1,5)" }
               print $remote "GET $document HTTP/1.0" . $BLANK;
               while ( <$remote> ) { print }
               close(2,7,n) $remote;

       The web server handing the "http" service, which is assumed to be at
       its standard port, number 80.  If the web server you're trying to con-
       nect to is at a different port (like 1080 or 8080), you should specify
       as the named-parameter pair, "PeerPort => 8080".  The "autoflush"
       method is used on the socket(2,7,n) because otherwise the system would buffer
       up the output we sent it.  (If you're on a Mac, you'll also need to
       change every "\n" in(1,8) your code that sends data over the network to be a
       "\015\012" instead.)

       Connecting to the server is only the first part of the process: once
       you have the connection, you have to use the server's language.  Each
       server on the network has its own little command language that it
       expects as input.  The string(3,n) that we send(2,n) to the server starting with
       "GET" is in(1,8) HTTP syntax.  In this case, we simply request each speci-
       fied document.  Yes, we really are making a new connection for each
       document, even though it's the same host.  That's the way you always
       used to have to speak HTTP.  Recent versions of web browsers may
       request that the remote server leave the connection open(2,3,n) a little
       while, but the server doesn't have to honor such a request.

       Here's an example of running that program, which we'll call webget:

           % webget /guanaco.html
           HTTP/1.1 404 File Not Found
           Date: Thu, 08 May 1997 18:02:32 GMT
           Server: Apache/1.2b6
           Connection: close(2,7,n)
           Content-type: text/html

           <HEAD><TITLE>404 File Not Found</TITLE></HEAD>
           <BODY><H1>File Not Found</H1>
           The requested URL /guanaco.html was not found on this server.<P>

       Ok, so that's not very interesting, because it didn't find that partic-
       ular document.  But a long response wouldn't have fit on this page.

       For a more fully-featured version(1,3,5) of this program, you should look(1,8,3 Search::Dict) to
       the lwp-request program included with the LWP modules from CPAN.

       Interactive Client with IO::Socket

       Well, that's all fine if(3,n) you want to send(2,n) one command and get one
       answer, but what about setting up something fully interactive, somewhat
       like the way telnet works?  That way you can type a line, get the
       answer, type a line, get the answer, etc.

       This client is more complicated than the two we've done so far, but if(3,n)
       you're on a system that supports the powerful "fork" call, the solution
       isn't that rough.  Once you've made the connection to whatever service
       you'd like to chat with, call "fork" to clone your process.  Each of
       these two identical process has a very simple job to do: the parent
       copies everything from the socket(2,7,n) to standard output, while the child
       simultaneously copies everything from standard input to the socket.  To
       accomplish the same thing using just one process would be much harder,
       because it's easier to code two processes to do one thing than it is to
       code one process to do two things.  (This keep-it-simple principle a
       cornerstones of the Unix philosophy, and good software engineering as
       well, which is probably why it's spread to other systems.)

       Here's the code:

           #!/usr/bin/perl -w
           use strict;
           use IO::Socket;
           my ($host(1,5), $port, $kidpid, $handle, $line);

           unless (@ARGV == 2) { die "usage: $0 host(1,5) port" }
           ($host(1,5), $port) = @ARGV;

           # create a tcp connection to the specified host(1,5) and port
           $handle = IO::Socket::INET->new(Proto     => "tcp",
                                           PeerAddr  => $host(1,5),
                                           PeerPort  => $port)
                  or die "can't connect to port $port on $host: $!";

           $handle->autoflush(1);              # so output gets(3,n) there right away
           print STDERR "[Connected to $host:$port]\n";

           # split(1,n) the program into two processes, identical twins
           die "can't fork: $!" unless defined($kidpid = fork());

           # the if(3,n){} block runs only in(1,8) the parent process
           if(3,n) ($kidpid) {
               # copy the socket(2,7,n) to standard output
               while (defined ($line = <$handle>)) {
                   print STDOUT $line;
               kill(1,2,1 builtins)("TERM", $kidpid);                  # send(2,n) SIGTERM to child
           # the else{} block runs only in(1,8) the child process
           else {
               # copy standard input to the socket(2,7,n)
               while (defined ($line = <STDIN>)) {
                   print $handle $line;

       The "kill(1,2,1 builtins)" function in(1,8) the parent's "if(3,n)" block is there to send(2,n) a sig-
       nal(2,7) to our child process (current running in(1,8) the "else" block) as soon
       as the remote server has closed its end of the connection.

       If the remote server sends data a byte at time(1,2,n), and you need that data
       immediately without waiting for a newline (which might not happen), you
       may wish to replace the "while" loop in(1,8) the parent with the following:

           my $byte;
           while (sysread($handle, $byte, 1) == 1) {
               print STDOUT $byte;

       Making a system call for each byte you want to read(2,n,1 builtins) is not very effi-
       cient (to put it mildly) but is the simplest to explain and works rea-
       sonably well.

TCP Servers with IO::Socket
       As always, setting up a server is little bit more involved than running
       a client.  The model is that the server creates a special kind of
       socket(2,7,n) that does nothing but listen(1,2,7) on a particular port for incoming
       connections.  It does this by calling the "IO::Socket::INET->new()"
       method with slightly different arguments than the client did.

           This is which protocol to use.  Like our clients, we'll still spec-
           ify "tcp" here.

           We specify a local port in(1,8) the "LocalPort" argument, which we
           didn't do for the client.  This is service name or port number for
           which you want to be the server. (Under Unix, ports under 1024 are
           restricted to the superuser.)  In our sample, we'll use port 9000,
           but you can use any port that's not currently in(1,8) use on your sys-
           tem.  If you try to use one already in(1,8) used, you'll get an "Address
           already in(1,8) use" message.  Under Unix, the "netstat -a" command will
           show which services current have servers.

           The "Listen" parameter is set(7,n,1 builtins) to the maximum number of pending con-
           nections we can accept(2,8) until we turn away incoming clients.  Think
           of it as a call-waiting queue(1,3) for your telephone.  The low-level
           Socket module has a special symbol for the system maximum, which is

           The "Reuse" parameter is needed so that we restart our server manu-
           ally without waiting a few minutes to allow system buffers to clear(1,3x,3x clrtobot)

       Once the generic server socket(2,7,n) has been created using the parameters
       listed above, the server then waits for a new client to connect to it.
       The server blocks in(1,8) the "accept(2,8)" method, which eventually accepts a
       bidirectional connection from the remote client.  (Make sure to aut-
       oflush this handle to circumvent buffering.)

       To add to user-friendliness, our server prompts the user for commands.
       Most servers don't do this.  Because of the prompt without a newline,
       you'll have to use the "sysread" variant of the interactive client

       This server accepts one of five different commands, sending output back
       to the client.  Note that unlike most network servers, this one only
       handles one incoming client at a time.  Multithreaded servers are cov-
       ered in(1,8) Chapter 6 of the Camel.

       Here's the code.  We'll

        #!/usr/bin/perl -w
        use IO::Socket;
        use Net::hostent;              # for OO version(1,3,5) of gethostbyaddr

        $PORT = 9000;                  # pick something not in(1,8) use

        $server = IO::Socket::INET->new( Proto     => 'tcp',
                                         LocalPort => $PORT,
                                         Listen    => SOMAXCONN,
                                         Reuse     => 1);

        die "can't setup(2,8) server" unless $server;
        print "[Server $0 accepting clients]\n";

        while ($client = $server->accept(2,8)()) {
          print $client "Welcome to $0; type help for command list.\n";
          $hostinfo = gethostbyaddr($client->peeraddr);
          printf(1,3,1 builtins) "[Connect from %s]\n", $hostinfo ? $hostinfo->name : $client->peerhost;
          print $client "Command? ";
          while ( <$client>) {
            next unless /\S/;       # blank line
            if(3,n)    (/quit|exit(3,n,1 builtins)/i)    { last;                                     }
            elsif (/date|time(1,2,n)/i)    { printf(1,3,1 builtins) $client "%s\n", scalar localtime;  }
            elsif (/who/i )         { print  $client `who 2>&1`;                }
            elsif (/cookie/i )      { print  $client `/usr/games/fortune 2>&1`; }
            elsif (/motd/i )        { print  $client `cat /etc/motd 2>&1`;      }
            else {
              print $client "Commands: quit date who cookie motd\n";
          } continue {
             print $client "Command? ";
          close(2,7,n) $client;

UDP: Message Passing
       Another kind of client-server setup(2,8) is one that uses not connections,
       but messages.  UDP communications involve much lower overhead but also
       provide less(1,3) reliability, as there are no promises that messages will
       arrive at all, let alone in(1,8) order and unmangled.  Still, UDP offers
       some advantages over TCP, including being able to "broadcast" or "mul-
       ticast" to a whole bunch of destination hosts at once (usually on your
       local subnet).  If you find yourself overly concerned about reliability
       and start building checks into your message system, then you probably
       should use just TCP to start with.

       Note that UDP datagrams are not a bytestream and should not be treated
       as such. This makes using I/O mechanisms with internal buffering like
       stdio (i.e. print() and friends) especially cumbersome. Use syswrite(),
       or better send(2,n)(), like in(1,8) the example below.

       Here's a UDP program similar to the sample Internet TCP client given
       earlier.  However, instead of checking one host(1,5) at a time(1,2,n), the UDP ver-
       sion(1,3,5) will check many of them asynchronously by simulating a multicast
       and then using select(2,7,2 select_tut)() to do a timed-out wait for I/O.  To do some-
       thing similar with TCP, you'd have to use a different socket(2,7,n) handle for
       each host.

           #!/usr/bin/perl -w
           use strict;
           use Socket;
           use Sys::Hostname;

           my ( $count, $hisiaddr, $hispaddr, $histime,
                $host(1,5), $iaddr, $paddr, $port, $proto,
                $rin, $rout, $rtime, $SECS_of_70_YEARS);

           $SECS_of_70_YEARS      = 2208988800;

           $iaddr = gethostbyname(hostname());
           $proto = getprotobyname('udp');
           $port = getservbyname('time(1,2,n)', 'udp');
           $paddr = sockaddr_in(0, $iaddr); # 0 means let kernel pick

           socket(2,7,n)(SOCKET, PF_INET, SOCK_DGRAM, $proto)   || die "socket: $!";
           bind(2,n,1 builtins)(SOCKET, $paddr)                          || die "bind: $!";

           $| = 1;
           printf(1,3,1 builtins) "%-12s %8s %s\n",  "localhost", 0, scalar localtime time(1,2,n);
           $count = 0;
           for $host(1,5) (@ARGV) {
               $hisiaddr = inet_aton($host(1,5))    || die "unknown host(1,5)";
               $hispaddr = sockaddr_in($port, $hisiaddr);
               defined(send(2,n)(SOCKET, 0, 0, $hispaddr))    || die "send(2,n) $host: $!";

           $rin = '';
           vec($rin, fileno(SOCKET), 1) = 1;

           # timeout(1,3x,3x cbreak) after 10.0 seconds
           while ($count && select(2,7,2 select_tut)($rout = $rin, undef, undef, 10.0)) {
               $rtime = '';
               ($hispaddr = recv(SOCKET, $rtime, 4, 0))        || die "recv: $!";
               ($port, $hisiaddr) = sockaddr_in($hispaddr);
               $host(1,5) = gethostbyaddr($hisiaddr, AF_INET);
               $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ;
               printf(1,3,1 builtins) "%-12s ", $host(1,5);
               printf(1,3,1 builtins) "%8d %s\n", $histime - time(1,2,n), scalar localtime($histime);

       Note that this example does not include any retries and may conse-
       quently fail to contact a reachable host. The most prominent reason for
       this is congestion of the queues on the sending host(1,5) if(3,n) the number of
       list of hosts to contact is sufficiently large.

       While System V IPC isn't so widely used as sockets, it still has some
       interesting uses.  You can't, however, effectively use SysV IPC or
       Berkeley mmap() to have shared memory so as to share a variable amongst
       several processes.  That's because Perl would reallocate your string(3,n)
       when you weren't wanting it to.

       Here's a small example showing shared memory usage.

           use IPC::SysV qw(IPC_PRIVATE IPC_RMID S_IRWXU);

           $size = 2000;
           $id = shmget(IPC_PRIVATE, $size, S_IRWXU) || die "$!";
           print "shm key $id\n";

           $message = "Message #1";
           shmwrite($id, $message, 0, 60) || die "$!";
           print "wrote: '$message'\n";
           shmread($id, $buff, 0, 60) || die "$!";
           print "read(2,n,1 builtins) : '$buff'\n";

           # the buffer of shmread is zero-character end-padded.
           substr($buff, index($buff, "\0")) = '';
           print "un" unless $buff eq $message;
           print "swell\n";

           print "deleting shm $id\n";
           shmctl($id, IPC_RMID, 0) || die "$!";

       Here's an example of a semaphore:

           use IPC::SysV qw(IPC_CREAT);

           $IPC_KEY = 1234;
           $id = semget($IPC_KEY, 10, 0666 | IPC_CREAT ) || die "$!";
           print "shm key $id\n";

       Put this code in(1,8) a separate file(1,n) to be run in(1,8) more than one process.
       Call the file(1,n) take:

           # create a semaphore

           $IPC_KEY = 1234;
           $id = semget($IPC_KEY,  0 , 0 );
           die if(3,n) !defined($id);

           $semnum = 0;
           $semflag = 0;

           # 'take' semaphore
           # wait for semaphore to be zero
           $semop = 0;
           $opstring1 = pack(3,n,n pack-old)("s!s!s!", $semnum, $semop, $semflag);

           # Increment the semaphore count
           $semop = 1;
           $opstring2 = pack(3,n,n pack-old)("s!s!s!", $semnum, $semop,  $semflag);
           $opstring = $opstring1 . $opstring2;

           semop($id,$opstring) || die "$!";

       Put this code in(1,8) a separate file(1,n) to be run in(1,8) more than one process.
       Call this file(1,n) give:

           # 'give' the semaphore
           # run this in(1,8) the original process and you will see
           # that the second process continues

           $IPC_KEY = 1234;
           $id = semget($IPC_KEY, 0, 0);
           die if(3,n) !defined($id);

           $semnum = 0;
           $semflag = 0;

           # Decrement the semaphore count
           $semop = -1;
           $opstring = pack(3,n,n pack-old)("s!s!s!", $semnum, $semop, $semflag);

           semop($id,$opstring) || die "$!";

       The SysV IPC code above was written long ago, and it's definitely
       clunky looking.  For a more modern look(1,8,3 Search::Dict), see the IPC::SysV module which
       is included with Perl starting from Perl 5.005.

       A small example demonstrating SysV message queues:


           my $id = msgget(IPC_PRIVATE, IPC_CREAT | S_IRWXU);

           my $sent = "message";
           my $type_sent = 1234;
           my $rcvd;
           my $type_rcvd;

           if(3,n) (defined $id) {
               if(3,n) (msgsnd($id, pack(3,n,n pack-old)("l! a*", $type_sent, $sent), 0)) {
                   if(3,n) (msgrcv($id, $rcvd, 60, 0, 0)) {
                       ($type_rcvd, $rcvd) = unpack("l! a*", $rcvd);
                       if(3,n) ($rcvd eq $sent) {
                           print "okay\n";
                       } else {
                           print "not okay\n";
                   } else {
                       die "# msgrcv failed\n";
               } else {
                   die "# msgsnd failed\n";
               msgctl($id, IPC_RMID, 0) || die "# msgctl failed: $!\n";
           } else {
               die "# msgget failed\n";

       Most of these routines quietly but politely return "undef" when they
       fail instead of causing your program to die right then and there due to
       an uncaught exception.  (Actually, some of the new Socket conversion
       functions  croak() on bad arguments.)  It is therefore essential to
       check return values from these functions.  Always begin your socket(2,7,n)
       programs this way for optimal success, and don't forget to add -T taint
       checking flag to the #! line for servers:

           #!/usr/bin/perl -Tw
           use strict;
           use sigtrap;
           use Socket;

       All these routines create system-specific portability problems.  As
       noted elsewhere, Perl is at the mercy of your C libraries for much of
       its system behaviour.  It's probably safest to assume broken SysV
       semantics for signals and to stick with simple TCP and UDP socket(2,7,n) oper-
       ations; e.g., don't try to pass open(2,3,n) file(1,n) descriptors over a local UDP
       datagram socket(2,7,n) if(3,n) you want your code to stand a chance of being porta-

       Tom Christiansen, with occasional vestiges of Larry Wall's original
       version(1,3,5) and suggestions from the Perl Porters.

       There's a lot more to networking than this, but this should get you

       For intrepid programmers, the indispensable textbook is Unix Network
       Programming, 2nd Edition, Volume 1 by W. Richard Stevens (published by
       Prentice-Hall).  Note that most books on networking address the subject
       from the perspective of a C programmer; translation to Perl is left as
       an exercise for the reader.

       The IO::Socket(3) manpage describes the object library, and the
       Socket(3) manpage describes the low-level interface to sockets.
       Besides the obvious functions in(1,8) perlfunc, you should also check out
       the modules file(1,n) at your nearest CPAN site.  (See perlmodlib or best
       yet, the Perl FAQ for a description of what CPAN is and where to get

       Section 5 of the modules file(1,n) is devoted to "Networking, Device Control
       (modems), and Interprocess Communication", and contains numerous unbun-
       dled modules numerous networking modules, Chat and Expect operations,
       CGI programming, DCE, FTP, IPC, NNTP, Proxy, Ptty, RPC, SNMP, SMTP,
       Telnet, Threads, and ToolTalk--just to name a few.

perl v5.8.5                       2004-04-23                        PERLIPC(1)

References for this manual (incoming links)