3 perlipc - Perl interprocess communication (signals, fifos, pipes, safe subprocesses, sockets, and semaphores)
7 The basic IPC facilities of Perl are built out of the good old Unix
8 signals, named pipes, pipe opens, the Berkeley socket routines, and SysV
9 IPC calls. Each is used in slightly different situations.
13 Perl uses a simple signal handling model: the %SIG hash contains names
14 or references of user-installed signal handlers. These handlers will
15 be called with an argument which is the name of the signal that
16 triggered it. A signal may be generated intentionally from a
17 particular keyboard sequence like control-C or control-Z, sent to you
18 from another process, or triggered automatically by the kernel when
19 special events transpire, like a child process exiting, your process
20 running out of stack space, or hitting file size limit.
22 For example, to trap an interrupt signal, set up a handler like this:
27 die "Somebody sent me a SIG$signame";
29 $SIG{INT} = 'catch_zap'; # could fail in modules
30 $SIG{INT} = \&catch_zap; # best strategy
32 Prior to Perl 5.7.3 it was necessary to do as little as you possibly
33 could in your handler; notice how all we do is set a global variable
34 and then raise an exception. That's because on most systems,
35 libraries are not re-entrant; particularly, memory allocation and I/O
36 routines are not. That meant that doing nearly I<anything> in your
37 handler could in theory trigger a memory fault and subsequent core
38 dump - see L<Deferred Signals> below.
40 The names of the signals are the ones listed out by C<kill -l> on your
41 system, or you can retrieve them from the Config module. Set up an
42 @signame list indexed by number to get the name and a %signo table
43 indexed by name to get the number:
46 defined $Config{sig_name} || die "No sigs?";
47 foreach $name (split(' ', $Config{sig_name})) {
53 So to check whether signal 17 and SIGALRM were the same, do just this:
55 print "signal #17 = $signame[17]\n";
57 print "SIGALRM is $signo{ALRM}\n";
60 You may also choose to assign the strings C<'IGNORE'> or C<'DEFAULT'> as
61 the handler, in which case Perl will try to discard the signal or do the
64 On most Unix platforms, the C<CHLD> (sometimes also known as C<CLD>) signal
65 has special behavior with respect to a value of C<'IGNORE'>.
66 Setting C<$SIG{CHLD}> to C<'IGNORE'> on such a platform has the effect of
67 not creating zombie processes when the parent process fails to C<wait()>
68 on its child processes (i.e. child processes are automatically reaped).
69 Calling C<wait()> with C<$SIG{CHLD}> set to C<'IGNORE'> usually returns
70 C<-1> on such platforms.
72 Some signals can be neither trapped nor ignored, such as
73 the KILL and STOP (but not the TSTP) signals. One strategy for
74 temporarily ignoring signals is to use a local() statement, which will be
75 automatically restored once your block is exited. (Remember that local()
76 values are "inherited" by functions called from within that block.)
79 local $SIG{INT} = 'IGNORE';
83 # interrupts still ignored, for now...
86 Sending a signal to a negative process ID means that you send the signal
87 to the entire Unix process-group. This code sends a hang-up signal to all
88 processes in the current process group (and sets $SIG{HUP} to IGNORE so
89 it doesn't kill itself):
92 local $SIG{HUP} = 'IGNORE';
94 # snazzy writing of: kill('HUP', -$$)
97 Another interesting signal to send is signal number zero. This doesn't
98 actually affect another process, but instead checks whether it's alive
99 or has changed its UID.
101 unless (kill 0 => $kid_pid) {
102 warn "something wicked happened to $kid_pid";
105 You might also want to employ anonymous functions for simple signal
108 $SIG{INT} = sub { die "\nOutta here!\n" };
110 But that will be problematic for the more complicated handlers that need
111 to reinstall themselves. Because Perl's signal mechanism is currently
112 based on the signal(3) function from the C library, you may sometimes be so
113 misfortunate as to run on systems where that function is "broken", that
114 is, it behaves in the old unreliable SysV way rather than the newer, more
115 reasonable BSD and POSIX fashion. So you'll see defensive people writing
116 signal handlers like this:
120 # loathe sysV: it makes us not only reinstate
121 # the handler, but place it after the wait
122 $SIG{CHLD} = \&REAPER;
124 $SIG{CHLD} = \&REAPER;
125 # now do something that forks...
129 use POSIX ":sys_wait_h";
132 # If a second child dies while in the signal handler caused by the
133 # first death, we won't get another signal. So must loop here else
134 # we will leave the unreaped child as a zombie. And the next time
135 # two children die we get another zombie. And so on.
136 while (($child = waitpid(-1,WNOHANG)) > 0) {
137 $Kid_Status{$child} = $?;
139 $SIG{CHLD} = \&REAPER; # still loathe sysV
141 $SIG{CHLD} = \&REAPER;
142 # do something that forks...
144 Signal handling is also used for timeouts in Unix, While safely
145 protected within an C<eval{}> block, you set a signal handler to trap
146 alarm signals and then schedule to have one delivered to you in some
147 number of seconds. Then try your blocking operation, clearing the alarm
148 when it's done but not before you've exited your C<eval{}> block. If it
149 goes off, you'll use die() to jump out of the block, much as you might
150 using longjmp() or throw() in other languages.
155 local $SIG{ALRM} = sub { die "alarm clock restart" };
157 flock(FH, 2); # blocking write lock
160 if ($@ and $@ !~ /alarm clock restart/) { die }
162 If the operation being timed out is system() or qx(), this technique
163 is liable to generate zombies. If this matters to you, you'll
164 need to do your own fork() and exec(), and kill the errant child process.
166 For more complex signal handling, you might see the standard POSIX
167 module. Lamentably, this is almost entirely undocumented, but
168 the F<t/lib/posix.t> file from the Perl source distribution has some
171 =head2 Handling the SIGHUP Signal in Daemons
173 A process that usually starts when the system boots and shuts down
174 when the system is shut down is called a daemon (Disk And Execution
175 MONitor). If a daemon process has a configuration file which is
176 modified after the process has been started, there should be a way to
177 tell that process to re-read its configuration file, without stopping
178 the process. Many daemons provide this mechanism using the C<SIGHUP>
179 signal handler. When you want to tell the daemon to re-read the file
180 you simply send it the C<SIGHUP> signal.
182 The implementation of such a mechanism in Perl using a normal signal
183 handler works only the first time the signal is sent. The solution to
184 this problem is to use C<POSIX> signal handlers if available.
186 The following example implements a simple daemon, which restarts
187 itself every time the C<SIGHUP> signal is received. The actual code is
188 located in the subroutine C<code()>, which simply prints some debug
189 info to show that it works and should be replaced with the real code.
195 use File::Basename ();
196 use File::Spec::Functions;
200 # make the daemon cross-platform, so exec always calls the script
201 # itself with the right path, no matter how the script was invoked.
202 my $script = File::Basename::basename($0);
203 my $SELF = catfile $FindBin::Bin, $script;
205 # POSIX unmasks the sigprocmask properly
206 my $sigset = POSIX::SigSet->new();
207 my $action = POSIX::SigAction->new('sigHUP_handler',
210 POSIX::sigaction(&POSIX::SIGHUP, $action);
213 print "got SIGHUP\n";
214 exec($SELF, @ARGV) or die "Couldn't restart: $!\n";
221 print "ARGV: @ARGV\n";
233 A named pipe (often referred to as a FIFO) is an old Unix IPC
234 mechanism for processes communicating on the same machine. It works
235 just like a regular, connected anonymous pipes, except that the
236 processes rendezvous using a filename and don't have to be related.
238 To create a named pipe, use the Unix command mknod(1) or on some
239 systems, mkfifo(1). These may not be in your normal path.
241 # system return val is backwards, so && not ||
243 $ENV{PATH} .= ":/etc:/usr/etc";
244 if ( system('mknod', $path, 'p')
245 && system('mkfifo', $path) )
247 die "mk{nod,fifo} $path failed";
251 A fifo is convenient when you want to connect a process to an unrelated
252 one. When you open a fifo, the program will block until there's something
255 For example, let's say you'd like to have your F<.signature> file be a
256 named pipe that has a Perl program on the other end. Now every time any
257 program (like a mailer, news reader, finger program, etc.) tries to read
258 from that file, the reading program will block and your program will
259 supply the new signature. We'll use the pipe-checking file test B<-p>
260 to find out whether anyone (or anything) has accidentally removed our fifo.
263 $FIFO = '.signature';
264 $ENV{PATH} .= ":/etc:/usr/games";
269 system('mknod', $FIFO, 'p')
270 && die "can't mknod $FIFO: $!";
273 # next line blocks until there's a reader
274 open (FIFO, "> $FIFO") || die "can't write $FIFO: $!";
275 print FIFO "John Smith (smith\@host.org)\n", `fortune -s`;
277 sleep 2; # to avoid dup signals
280 =head2 Deferred Signals
282 In Perls before Perl 5.7.3 by installing Perl code to deal with
283 signals, you were exposing yourself to danger from two things. First,
284 few system library functions are re-entrant. If the signal interrupts
285 while Perl is executing one function (like malloc(3) or printf(3)),
286 and your signal handler then calls the same function again, you could
287 get unpredictable behavior--often, a core dump. Second, Perl isn't
288 itself re-entrant at the lowest levels. If the signal interrupts Perl
289 while Perl is changing its own internal data structures, similarly
290 unpredictable behaviour may result.
292 There were two things you could do, knowing this: be paranoid or be
293 pragmatic. The paranoid approach was to do as little as possible in your
294 signal handler. Set an existing integer variable that already has a
295 value, and return. This doesn't help you if you're in a slow system call,
296 which will just restart. That means you have to C<die> to longjump(3) out
297 of the handler. Even this is a little cavalier for the true paranoiac,
298 who avoids C<die> in a handler because the system I<is> out to get you.
299 The pragmatic approach was to say ``I know the risks, but prefer the
300 convenience'', and to do anything you wanted in your signal handler,
301 and be prepared to clean up core dumps now and again.
303 In Perl 5.7.3 and later to avoid these problems signals are
304 "deferred"-- that is when the signal is delivered to the process by
305 the system (to the C code that implements Perl) a flag is set, and the
306 handler returns immediately. Then at strategic "safe" points in the
307 Perl interpreter (e.g. when it is about to execute a new opcode) the
308 flags are checked and the Perl level handler from %SIG is
309 executed. The "deferred" scheme allows much more flexibility in the
310 coding of signal handler as we know Perl interpreter is in a safe
311 state, and that we are not in a system library function when the
312 handler is called. However the implementation does differ from
313 previous Perls in the following ways:
317 =item Long running opcodes
319 As Perl interpreter only looks at the signal flags when it about to
320 execute a new opcode if a signal arrives during a long running opcode
321 (e.g. a regular expression operation on a very large string) then
322 signal will not be seen until operation completes.
324 =item Interrupting IO
326 When a signal is delivered (e.g. INT control-C) the operating system
327 breaks into IO operations like C<read> (used to implement Perls
328 E<lt>E<gt> operator). On older Perls the handler was called
329 immediately (and as C<read> is not "unsafe" this worked well). With
330 the "deferred" scheme the handler is not called immediately, and if
331 Perl is using system's C<stdio> library that library may re-start the
332 C<read> without returning to Perl and giving it a chance to call the
333 %SIG handler. If this happens on your system the solution is to use
334 C<:perlio> layer to do IO - at least on those handles which you want
335 to be able to break into with signals. (The C<:perlio> layer checks
336 the signal flags and calls %SIG handlers before resuming IO operation.)
338 Note that the default in Perl 5.7.3 and later is to automatically use
339 the C<:perlio> layer.
341 =item Signals as "faults"
343 Certain signals e.g. SEGV, ILL, BUS are generated as a result of
344 virtual memory or other "faults". These are normally fatal and there
345 is little a Perl-level handler can do with them. (In particular the
346 old signal scheme was particularly unsafe in such cases.) However if
347 a %SIG handler is set the new scheme simply sets a flag and returns as
348 described above. This may cause the operating system to try the
349 offending machine instruction again and - as nothing has changed - it
350 will generate the signal again. The result of this is a rather odd
351 "loop". In future Perl's signal mechanism may be changed to avoid this
352 - perhaps by simply disallowing %SIG handlers on signals of that
353 type. Until then the work-round is not to set a %SIG handler on those
354 signals. (Which signals they are is operating system dependant.)
356 =item Signals triggered by operating system state
358 On some operating systems certain signal handlers are supposed to "do
359 something" before returning. One example can be CHLD or CLD which
360 indicates a child process has completed. On some operating systems the
361 signal handler is expected to C<wait> for the completed child
362 process. On such systems the deferred signal scheme will not work for
363 those signals (it does not do the C<wait>). Again the failure will
364 look like a loop as the operating system will re-issue the signal as
365 there are un-waited-for completed child processes.
369 =head1 Using open() for IPC
371 Perl's basic open() statement can also be used for unidirectional
372 interprocess communication by either appending or prepending a pipe
373 symbol to the second argument to open(). Here's how to start
374 something up in a child process you intend to write to:
376 open(SPOOLER, "| cat -v | lpr -h 2>/dev/null")
377 || die "can't fork: $!";
378 local $SIG{PIPE} = sub { die "spooler pipe broke" };
379 print SPOOLER "stuff\n";
380 close SPOOLER || die "bad spool: $! $?";
382 And here's how to start up a child process you intend to read from:
384 open(STATUS, "netstat -an 2>&1 |")
385 || die "can't fork: $!";
387 next if /^(tcp|udp)/;
390 close STATUS || die "bad netstat: $! $?";
392 If one can be sure that a particular program is a Perl script that is
393 expecting filenames in @ARGV, the clever programmer can write something
396 % program f1 "cmd1|" - f2 "cmd2|" f3 < tmpfile
398 and irrespective of which shell it's called from, the Perl program will
399 read from the file F<f1>, the process F<cmd1>, standard input (F<tmpfile>
400 in this case), the F<f2> file, the F<cmd2> command, and finally the F<f3>
401 file. Pretty nifty, eh?
403 You might notice that you could use backticks for much the
404 same effect as opening a pipe for reading:
406 print grep { !/^(tcp|udp)/ } `netstat -an 2>&1`;
407 die "bad netstat" if $?;
409 While this is true on the surface, it's much more efficient to process the
410 file one line or record at a time because then you don't have to read the
411 whole thing into memory at once. It also gives you finer control of the
412 whole process, letting you to kill off the child process early if you'd
415 Be careful to check both the open() and the close() return values. If
416 you're I<writing> to a pipe, you should also trap SIGPIPE. Otherwise,
417 think of what happens when you start up a pipe to a command that doesn't
418 exist: the open() will in all likelihood succeed (it only reflects the
419 fork()'s success), but then your output will fail--spectacularly. Perl
420 can't know whether the command worked because your command is actually
421 running in a separate process whose exec() might have failed. Therefore,
422 while readers of bogus commands return just a quick end of file, writers
423 to bogus command will trigger a signal they'd better be prepared to
426 open(FH, "|bogus") or die "can't fork: $!";
427 print FH "bang\n" or die "can't write: $!";
428 close FH or die "can't close: $!";
430 That won't blow up until the close, and it will blow up with a SIGPIPE.
431 To catch it, you could use this:
433 $SIG{PIPE} = 'IGNORE';
434 open(FH, "|bogus") or die "can't fork: $!";
435 print FH "bang\n" or die "can't write: $!";
436 close FH or die "can't close: status=$?";
440 Both the main process and any child processes it forks share the same
441 STDIN, STDOUT, and STDERR filehandles. If both processes try to access
442 them at once, strange things can happen. You may also want to close
443 or reopen the filehandles for the child. You can get around this by
444 opening your pipe with open(), but on some systems this means that the
445 child process cannot outlive the parent.
447 =head2 Background Processes
449 You can run a command in the background with:
453 The command's STDOUT and STDERR (and possibly STDIN, depending on your
454 shell) will be the same as the parent's. You won't need to catch
455 SIGCHLD because of the double-fork taking place (see below for more
458 =head2 Complete Dissociation of Child from Parent
460 In some cases (starting server processes, for instance) you'll want to
461 completely dissociate the child process from the parent. This is
462 often called daemonization. A well behaved daemon will also chdir()
463 to the root directory (so it doesn't prevent unmounting the filesystem
464 containing the directory from which it was launched) and redirect its
465 standard file descriptors from and to F</dev/null> (so that random
466 output doesn't wind up on the user's terminal).
471 chdir '/' or die "Can't chdir to /: $!";
472 open STDIN, '/dev/null' or die "Can't read /dev/null: $!";
473 open STDOUT, '>/dev/null'
474 or die "Can't write to /dev/null: $!";
475 defined(my $pid = fork) or die "Can't fork: $!";
477 setsid or die "Can't start a new session: $!";
478 open STDERR, '>&STDOUT' or die "Can't dup stdout: $!";
481 The fork() has to come before the setsid() to ensure that you aren't a
482 process group leader (the setsid() will fail if you are). If your
483 system doesn't have the setsid() function, open F</dev/tty> and use the
484 C<TIOCNOTTY> ioctl() on it instead. See L<tty(4)> for details.
486 Non-Unix users should check their Your_OS::Process module for other
489 =head2 Safe Pipe Opens
491 Another interesting approach to IPC is making your single program go
492 multiprocess and communicate between (or even amongst) yourselves. The
493 open() function will accept a file argument of either C<"-|"> or C<"|-">
494 to do a very interesting thing: it forks a child connected to the
495 filehandle you've opened. The child is running the same program as the
496 parent. This is useful for safely opening a file when running under an
497 assumed UID or GID, for example. If you open a pipe I<to> minus, you can
498 write to the filehandle you opened and your kid will find it in his
499 STDIN. If you open a pipe I<from> minus, you can read from the filehandle
500 you opened whatever your kid writes to his STDOUT.
502 use English '-no_match_vars';
506 $pid = open(KID_TO_WRITE, "|-");
507 unless (defined $pid) {
508 warn "cannot fork: $!";
509 die "bailing out" if $sleep_count++ > 6;
512 } until defined $pid;
515 print KID_TO_WRITE @some_data;
516 close(KID_TO_WRITE) || warn "kid exited $?";
518 ($EUID, $EGID) = ($UID, $GID); # suid progs only
519 open (FILE, "> /safe/file")
520 || die "can't open /safe/file: $!";
522 print FILE; # child's STDIN is parent's KID
524 exit; # don't forget this
527 Another common use for this construct is when you need to execute
528 something without the shell's interference. With system(), it's
529 straightforward, but you can't use a pipe open or backticks safely.
530 That's because there's no way to stop the shell from getting its hands on
531 your arguments. Instead, use lower-level control to call exec() directly.
533 Here's a safe backtick or pipe open for read:
535 # add error processing as above
536 $pid = open(KID_TO_READ, "-|");
539 while (<KID_TO_READ>) {
540 # do something interesting
542 close(KID_TO_READ) || warn "kid exited $?";
545 ($EUID, $EGID) = ($UID, $GID); # suid only
546 exec($program, @options, @args)
547 || die "can't exec program: $!";
552 And here's a safe pipe open for writing:
554 # add error processing as above
555 $pid = open(KID_TO_WRITE, "|-");
556 $SIG{ALRM} = sub { die "whoops, $program pipe broke" };
562 close(KID_TO_WRITE) || warn "kid exited $?";
565 ($EUID, $EGID) = ($UID, $GID);
566 exec($program, @options, @args)
567 || die "can't exec program: $!";
571 Note that these operations are full Unix forks, which means they may not be
572 correctly implemented on alien systems. Additionally, these are not true
573 multithreading. If you'd like to learn more about threading, see the
574 F<modules> file mentioned below in the SEE ALSO section.
576 =head2 Bidirectional Communication with Another Process
578 While this works reasonably well for unidirectional communication, what
579 about bidirectional communication? The obvious thing you'd like to do
580 doesn't actually work:
582 open(PROG_FOR_READING_AND_WRITING, "| some program |")
584 and if you forget to use the C<use warnings> pragma or the B<-w> flag,
585 then you'll miss out entirely on the diagnostic message:
587 Can't do bidirectional pipe at -e line 1.
589 If you really want to, you can use the standard open2() library function
590 to catch both ends. There's also an open3() for tridirectional I/O so you
591 can also catch your child's STDERR, but doing so would then require an
592 awkward select() loop and wouldn't allow you to use normal Perl input
595 If you look at its source, you'll see that open2() uses low-level
596 primitives like Unix pipe() and exec() calls to create all the connections.
597 While it might have been slightly more efficient by using socketpair(), it
598 would have then been even less portable than it already is. The open2()
599 and open3() functions are unlikely to work anywhere except on a Unix
600 system or some other one purporting to be POSIX compliant.
602 Here's an example of using open2():
606 $pid = open2(*Reader, *Writer, "cat -u -n" );
607 print Writer "stuff\n";
610 The problem with this is that Unix buffering is really going to
611 ruin your day. Even though your C<Writer> filehandle is auto-flushed,
612 and the process on the other end will get your data in a timely manner,
613 you can't usually do anything to force it to give it back to you
614 in a similarly quick fashion. In this case, we could, because we
615 gave I<cat> a B<-u> flag to make it unbuffered. But very few Unix
616 commands are designed to operate over pipes, so this seldom works
617 unless you yourself wrote the program on the other end of the
620 A solution to this is the nonstandard F<Comm.pl> library. It uses
621 pseudo-ttys to make your program behave more reasonably:
624 $ph = open_proc('cat -n');
626 print $ph "a line\n";
627 print "got back ", scalar <$ph>;
630 This way you don't have to have control over the source code of the
631 program you're using. The F<Comm> library also has expect()
632 and interact() functions. Find the library (and we hope its
633 successor F<IPC::Chat>) at your nearest CPAN archive as detailed
634 in the SEE ALSO section below.
636 The newer Expect.pm module from CPAN also addresses this kind of thing.
637 This module requires two other modules from CPAN: IO::Pty and IO::Stty.
638 It sets up a pseudo-terminal to interact with programs that insist on
639 using talking to the terminal device driver. If your system is
640 amongst those supported, this may be your best bet.
642 =head2 Bidirectional Communication with Yourself
644 If you want, you may make low-level pipe() and fork()
645 to stitch this together by hand. This example only
646 talks to itself, but you could reopen the appropriate
647 handles to STDIN and STDOUT and call other processes.
650 # pipe1 - bidirectional communication using two pipe pairs
651 # designed for the socketpair-challenged
652 use IO::Handle; # thousands of lines just for autoflush :-(
653 pipe(PARENT_RDR, CHILD_WTR); # XXX: failure?
654 pipe(CHILD_RDR, PARENT_WTR); # XXX: failure?
655 CHILD_WTR->autoflush(1);
656 PARENT_WTR->autoflush(1);
659 close PARENT_RDR; close PARENT_WTR;
660 print CHILD_WTR "Parent Pid $$ is sending this\n";
661 chomp($line = <CHILD_RDR>);
662 print "Parent Pid $$ just read this: `$line'\n";
663 close CHILD_RDR; close CHILD_WTR;
666 die "cannot fork: $!" unless defined $pid;
667 close CHILD_RDR; close CHILD_WTR;
668 chomp($line = <PARENT_RDR>);
669 print "Child Pid $$ just read this: `$line'\n";
670 print PARENT_WTR "Child Pid $$ is sending this\n";
671 close PARENT_RDR; close PARENT_WTR;
675 But you don't actually have to make two pipe calls. If you
676 have the socketpair() system call, it will do this all for you.
679 # pipe2 - bidirectional communication using socketpair
680 # "the best ones always go both ways"
683 use IO::Handle; # thousands of lines just for autoflush :-(
684 # We say AF_UNIX because although *_LOCAL is the
685 # POSIX 1003.1g form of the constant, many machines
686 # still don't have it.
687 socketpair(CHILD, PARENT, AF_UNIX, SOCK_STREAM, PF_UNSPEC)
688 or die "socketpair: $!";
691 PARENT->autoflush(1);
695 print CHILD "Parent Pid $$ is sending this\n";
696 chomp($line = <CHILD>);
697 print "Parent Pid $$ just read this: `$line'\n";
701 die "cannot fork: $!" unless defined $pid;
703 chomp($line = <PARENT>);
704 print "Child Pid $$ just read this: `$line'\n";
705 print PARENT "Child Pid $$ is sending this\n";
710 =head1 Sockets: Client/Server Communication
712 While not limited to Unix-derived operating systems (e.g., WinSock on PCs
713 provides socket support, as do some VMS libraries), you may not have
714 sockets on your system, in which case this section probably isn't going to do
715 you much good. With sockets, you can do both virtual circuits (i.e., TCP
716 streams) and datagrams (i.e., UDP packets). You may be able to do even more
717 depending on your system.
719 The Perl function calls for dealing with sockets have the same names as
720 the corresponding system calls in C, but their arguments tend to differ
721 for two reasons: first, Perl filehandles work differently than C file
722 descriptors. Second, Perl already knows the length of its strings, so you
723 don't need to pass that information.
725 One of the major problems with old socket code in Perl was that it used
726 hard-coded values for some of the constants, which severely hurt
727 portability. If you ever see code that does anything like explicitly
728 setting C<$AF_INET = 2>, you know you're in for big trouble: An
729 immeasurably superior approach is to use the C<Socket> module, which more
730 reliably grants access to various constants and functions you'll need.
732 If you're not writing a server/client for an existing protocol like
733 NNTP or SMTP, you should give some thought to how your server will
734 know when the client has finished talking, and vice-versa. Most
735 protocols are based on one-line messages and responses (so one party
736 knows the other has finished when a "\n" is received) or multi-line
737 messages and responses that end with a period on an empty line
738 ("\n.\n" terminates a message/response).
740 =head2 Internet Line Terminators
742 The Internet line terminator is "\015\012". Under ASCII variants of
743 Unix, that could usually be written as "\r\n", but under other systems,
744 "\r\n" might at times be "\015\015\012", "\012\012\015", or something
745 completely different. The standards specify writing "\015\012" to be
746 conformant (be strict in what you provide), but they also recommend
747 accepting a lone "\012" on input (but be lenient in what you require).
748 We haven't always been very good about that in the code in this manpage,
749 but unless you're on a Mac, you'll probably be ok.
751 =head2 Internet TCP Clients and Servers
753 Use Internet-domain sockets when you want to do client-server
754 communication that might extend to machines outside of your own system.
756 Here's a sample TCP client using Internet-domain sockets:
761 my ($remote,$port, $iaddr, $paddr, $proto, $line);
763 $remote = shift || 'localhost';
764 $port = shift || 2345; # random port
765 if ($port =~ /\D/) { $port = getservbyname($port, 'tcp') }
766 die "No port" unless $port;
767 $iaddr = inet_aton($remote) || die "no host: $remote";
768 $paddr = sockaddr_in($port, $iaddr);
770 $proto = getprotobyname('tcp');
771 socket(SOCK, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
772 connect(SOCK, $paddr) || die "connect: $!";
773 while (defined($line = <SOCK>)) {
777 close (SOCK) || die "close: $!";
780 And here's a corresponding server to go along with it. We'll
781 leave the address as INADDR_ANY so that the kernel can choose
782 the appropriate interface on multihomed hosts. If you want sit
783 on a particular interface (like the external side of a gateway
784 or firewall machine), you should fill this in with your real address
789 BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
792 my $EOL = "\015\012";
794 sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }
796 my $port = shift || 2345;
797 my $proto = getprotobyname('tcp');
799 ($port) = $port =~ /^(\d+)$/ or die "invalid port";
801 socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
802 setsockopt(Server, SOL_SOCKET, SO_REUSEADDR,
803 pack("l", 1)) || die "setsockopt: $!";
804 bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!";
805 listen(Server,SOMAXCONN) || die "listen: $!";
807 logmsg "server started on port $port";
811 $SIG{CHLD} = \&REAPER;
813 for ( ; $paddr = accept(Client,Server); close Client) {
814 my($port,$iaddr) = sockaddr_in($paddr);
815 my $name = gethostbyaddr($iaddr,AF_INET);
817 logmsg "connection from $name [",
818 inet_ntoa($iaddr), "]
821 print Client "Hello there, $name, it's now ",
822 scalar localtime, $EOL;
825 And here's a multithreaded version. It's multithreaded in that
826 like most typical servers, it spawns (forks) a slave server to
827 handle the client request so that the master server can quickly
828 go back to service a new client.
832 BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
835 my $EOL = "\015\012";
837 sub spawn; # forward declaration
838 sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }
840 my $port = shift || 2345;
841 my $proto = getprotobyname('tcp');
843 ($port) = $port =~ /^(\d+)$/ or die "invalid port";
845 socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
846 setsockopt(Server, SOL_SOCKET, SO_REUSEADDR,
847 pack("l", 1)) || die "setsockopt: $!";
848 bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!";
849 listen(Server,SOMAXCONN) || die "listen: $!";
851 logmsg "server started on port $port";
856 use POSIX ":sys_wait_h";
859 while (($waitedpid = waitpid(-1,WNOHANG)) > 0) {
860 logmsg "reaped $waitedpid" . ($? ? " with exit $?" : '');
862 $SIG{CHLD} = \&REAPER; # loathe sysV
865 $SIG{CHLD} = \&REAPER;
867 for ( $waitedpid = 0;
868 ($paddr = accept(Client,Server)) || $waitedpid;
869 $waitedpid = 0, close Client)
871 next if $waitedpid and not $paddr;
872 my($port,$iaddr) = sockaddr_in($paddr);
873 my $name = gethostbyaddr($iaddr,AF_INET);
875 logmsg "connection from $name [",
876 inet_ntoa($iaddr), "]
881 print "Hello there, $name, it's now ", scalar localtime, $EOL;
882 exec '/usr/games/fortune' # XXX: `wrong' line terminators
883 or confess "can't exec fortune: $!";
891 unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') {
892 confess "usage: spawn CODEREF";
896 if (!defined($pid = fork)) {
897 logmsg "cannot fork: $!";
901 return; # I'm the parent
903 # else I'm the child -- go spawn
905 open(STDIN, "<&Client") || die "can't dup client to stdin";
906 open(STDOUT, ">&Client") || die "can't dup client to stdout";
907 ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
911 This server takes the trouble to clone off a child version via fork() for
912 each incoming request. That way it can handle many requests at once,
913 which you might not always want. Even if you don't fork(), the listen()
914 will allow that many pending connections. Forking servers have to be
915 particularly careful about cleaning up their dead children (called
916 "zombies" in Unix parlance), because otherwise you'll quickly fill up your
919 We suggest that you use the B<-T> flag to use taint checking (see L<perlsec>)
920 even if we aren't running setuid or setgid. This is always a good idea
921 for servers and other programs run on behalf of someone else (like CGI
922 scripts), because it lessens the chances that people from the outside will
923 be able to compromise your system.
925 Let's look at another TCP client. This one connects to the TCP "time"
926 service on a number of different machines and shows how far their clocks
927 differ from the system on which it's being run:
933 my $SECS_of_70_YEARS = 2208988800;
934 sub ctime { scalar localtime(shift) }
936 my $iaddr = gethostbyname('localhost');
937 my $proto = getprotobyname('tcp');
938 my $port = getservbyname('time', 'tcp');
939 my $paddr = sockaddr_in(0, $iaddr);
943 printf "%-24s %8s %s\n", "localhost", 0, ctime(time());
945 foreach $host (@ARGV) {
946 printf "%-24s ", $host;
947 my $hisiaddr = inet_aton($host) || die "unknown host";
948 my $hispaddr = sockaddr_in($port, $hisiaddr);
949 socket(SOCKET, PF_INET, SOCK_STREAM, $proto) || die "socket: $!";
950 connect(SOCKET, $hispaddr) || die "bind: $!";
952 read(SOCKET, $rtime, 4);
954 my $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ;
955 printf "%8d %s\n", $histime - time, ctime($histime);
958 =head2 Unix-Domain TCP Clients and Servers
960 That's fine for Internet-domain clients and servers, but what about local
961 communications? While you can use the same setup, sometimes you don't
962 want to. Unix-domain sockets are local to the current host, and are often
963 used internally to implement pipes. Unlike Internet domain sockets, Unix
964 domain sockets can show up in the file system with an ls(1) listing.
967 srw-rw-rw- 1 root 0 Oct 31 07:23 /dev/log
969 You can test for these with Perl's B<-S> file test:
971 unless ( -S '/dev/log' ) {
972 die "something's wicked with the log system";
975 Here's a sample Unix-domain client:
980 my ($rendezvous, $line);
982 $rendezvous = shift || '/tmp/catsock';
983 socket(SOCK, PF_UNIX, SOCK_STREAM, 0) || die "socket: $!";
984 connect(SOCK, sockaddr_un($rendezvous)) || die "connect: $!";
985 while (defined($line = <SOCK>)) {
990 And here's a corresponding server. You don't have to worry about silly
991 network terminators here because Unix domain sockets are guaranteed
992 to be on the localhost, and thus everything works right.
999 BEGIN { $ENV{PATH} = '/usr/ucb:/bin' }
1000 sub spawn; # forward declaration
1001 sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" }
1003 my $NAME = '/tmp/catsock';
1004 my $uaddr = sockaddr_un($NAME);
1005 my $proto = getprotobyname('tcp');
1007 socket(Server,PF_UNIX,SOCK_STREAM,0) || die "socket: $!";
1009 bind (Server, $uaddr) || die "bind: $!";
1010 listen(Server,SOMAXCONN) || die "listen: $!";
1012 logmsg "server started on $NAME";
1016 use POSIX ":sys_wait_h";
1019 while (($waitedpid = waitpid(-1,WNOHANG)) > 0) {
1020 logmsg "reaped $waitedpid" . ($? ? " with exit $?" : '');
1022 $SIG{CHLD} = \&REAPER; # loathe sysV
1025 $SIG{CHLD} = \&REAPER;
1028 for ( $waitedpid = 0;
1029 accept(Client,Server) || $waitedpid;
1030 $waitedpid = 0, close Client)
1033 logmsg "connection on $NAME";
1035 print "Hello there, it's now ", scalar localtime, "\n";
1036 exec '/usr/games/fortune' or die "can't exec fortune: $!";
1041 my $coderef = shift;
1043 unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') {
1044 confess "usage: spawn CODEREF";
1048 if (!defined($pid = fork)) {
1049 logmsg "cannot fork: $!";
1052 logmsg "begat $pid";
1053 return; # I'm the parent
1055 # else I'm the child -- go spawn
1057 open(STDIN, "<&Client") || die "can't dup client to stdin";
1058 open(STDOUT, ">&Client") || die "can't dup client to stdout";
1059 ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr";
1063 As you see, it's remarkably similar to the Internet domain TCP server, so
1064 much so, in fact, that we've omitted several duplicate functions--spawn(),
1065 logmsg(), ctime(), and REAPER()--which are exactly the same as in the
1068 So why would you ever want to use a Unix domain socket instead of a
1069 simpler named pipe? Because a named pipe doesn't give you sessions. You
1070 can't tell one process's data from another's. With socket programming,
1071 you get a separate session for each client: that's why accept() takes two
1074 For example, let's say that you have a long running database server daemon
1075 that you want folks from the World Wide Web to be able to access, but only
1076 if they go through a CGI interface. You'd have a small, simple CGI
1077 program that does whatever checks and logging you feel like, and then acts
1078 as a Unix-domain client and connects to your private server.
1080 =head1 TCP Clients with IO::Socket
1082 For those preferring a higher-level interface to socket programming, the
1083 IO::Socket module provides an object-oriented approach. IO::Socket is
1084 included as part of the standard Perl distribution as of the 5.004
1085 release. If you're running an earlier version of Perl, just fetch
1086 IO::Socket from CPAN, where you'll also find modules providing easy
1087 interfaces to the following systems: DNS, FTP, Ident (RFC 931), NIS and
1088 NISPlus, NNTP, Ping, POP3, SMTP, SNMP, SSLeay, Telnet, and Time--just
1091 =head2 A Simple Client
1093 Here's a client that creates a TCP connection to the "daytime"
1094 service at port 13 of the host name "localhost" and prints out everything
1095 that the server there cares to provide.
1099 $remote = IO::Socket::INET->new(
1101 PeerAddr => "localhost",
1102 PeerPort => "daytime(13)",
1104 or die "cannot connect to daytime port at localhost";
1105 while ( <$remote> ) { print }
1107 When you run this program, you should get something back that
1110 Wed May 14 08:40:46 MDT 1997
1112 Here are what those parameters to the C<new> constructor mean:
1118 This is which protocol to use. In this case, the socket handle returned
1119 will be connected to a TCP socket, because we want a stream-oriented
1120 connection, that is, one that acts pretty much like a plain old file.
1121 Not all sockets are this of this type. For example, the UDP protocol
1122 can be used to make a datagram socket, used for message-passing.
1126 This is the name or Internet address of the remote host the server is
1127 running on. We could have specified a longer name like C<"www.perl.com">,
1128 or an address like C<"204.148.40.9">. For demonstration purposes, we've
1129 used the special hostname C<"localhost">, which should always mean the
1130 current machine you're running on. The corresponding Internet address
1131 for localhost is C<"127.1">, if you'd rather use that.
1135 This is the service name or port number we'd like to connect to.
1136 We could have gotten away with using just C<"daytime"> on systems with a
1137 well-configured system services file,[FOOTNOTE: The system services file
1138 is in I</etc/services> under Unix] but just in case, we've specified the
1139 port number (13) in parentheses. Using just the number would also have
1140 worked, but constant numbers make careful programmers nervous.
1144 Notice how the return value from the C<new> constructor is used as
1145 a filehandle in the C<while> loop? That's what's called an indirect
1146 filehandle, a scalar variable containing a filehandle. You can use
1147 it the same way you would a normal filehandle. For example, you
1148 can read one line from it this way:
1152 all remaining lines from is this way:
1156 and send a line of data to it this way:
1158 print $handle "some data\n";
1160 =head2 A Webget Client
1162 Here's a simple client that takes a remote host to fetch a document
1163 from, and then a list of documents to get from that host. This is a
1164 more interesting client than the previous one because it first sends
1165 something to the server before fetching the server's response.
1169 unless (@ARGV > 1) { die "usage: $0 host document ..." }
1170 $host = shift(@ARGV);
1173 foreach $document ( @ARGV ) {
1174 $remote = IO::Socket::INET->new( Proto => "tcp",
1176 PeerPort => "http(80)",
1178 unless ($remote) { die "cannot connect to http daemon on $host" }
1179 $remote->autoflush(1);
1180 print $remote "GET $document HTTP/1.0" . $BLANK;
1181 while ( <$remote> ) { print }
1185 The web server handing the "http" service, which is assumed to be at
1186 its standard port, number 80. If the web server you're trying to
1187 connect to is at a different port (like 1080 or 8080), you should specify
1188 as the named-parameter pair, C<< PeerPort => 8080 >>. The C<autoflush>
1189 method is used on the socket because otherwise the system would buffer
1190 up the output we sent it. (If you're on a Mac, you'll also need to
1191 change every C<"\n"> in your code that sends data over the network to
1192 be a C<"\015\012"> instead.)
1194 Connecting to the server is only the first part of the process: once you
1195 have the connection, you have to use the server's language. Each server
1196 on the network has its own little command language that it expects as
1197 input. The string that we send to the server starting with "GET" is in
1198 HTTP syntax. In this case, we simply request each specified document.
1199 Yes, we really are making a new connection for each document, even though
1200 it's the same host. That's the way you always used to have to speak HTTP.
1201 Recent versions of web browsers may request that the remote server leave
1202 the connection open a little while, but the server doesn't have to honor
1205 Here's an example of running that program, which we'll call I<webget>:
1207 % webget www.perl.com /guanaco.html
1208 HTTP/1.1 404 File Not Found
1209 Date: Thu, 08 May 1997 18:02:32 GMT
1210 Server: Apache/1.2b6
1212 Content-type: text/html
1214 <HEAD><TITLE>404 File Not Found</TITLE></HEAD>
1215 <BODY><H1>File Not Found</H1>
1216 The requested URL /guanaco.html was not found on this server.<P>
1219 Ok, so that's not very interesting, because it didn't find that
1220 particular document. But a long response wouldn't have fit on this page.
1222 For a more fully-featured version of this program, you should look to
1223 the I<lwp-request> program included with the LWP modules from CPAN.
1225 =head2 Interactive Client with IO::Socket
1227 Well, that's all fine if you want to send one command and get one answer,
1228 but what about setting up something fully interactive, somewhat like
1229 the way I<telnet> works? That way you can type a line, get the answer,
1230 type a line, get the answer, etc.
1232 This client is more complicated than the two we've done so far, but if
1233 you're on a system that supports the powerful C<fork> call, the solution
1234 isn't that rough. Once you've made the connection to whatever service
1235 you'd like to chat with, call C<fork> to clone your process. Each of
1236 these two identical process has a very simple job to do: the parent
1237 copies everything from the socket to standard output, while the child
1238 simultaneously copies everything from standard input to the socket.
1239 To accomplish the same thing using just one process would be I<much>
1240 harder, because it's easier to code two processes to do one thing than it
1241 is to code one process to do two things. (This keep-it-simple principle
1242 a cornerstones of the Unix philosophy, and good software engineering as
1243 well, which is probably why it's spread to other systems.)
1250 my ($host, $port, $kidpid, $handle, $line);
1252 unless (@ARGV == 2) { die "usage: $0 host port" }
1253 ($host, $port) = @ARGV;
1255 # create a tcp connection to the specified host and port
1256 $handle = IO::Socket::INET->new(Proto => "tcp",
1259 or die "can't connect to port $port on $host: $!";
1261 $handle->autoflush(1); # so output gets there right away
1262 print STDERR "[Connected to $host:$port]\n";
1264 # split the program into two processes, identical twins
1265 die "can't fork: $!" unless defined($kidpid = fork());
1267 # the if{} block runs only in the parent process
1269 # copy the socket to standard output
1270 while (defined ($line = <$handle>)) {
1273 kill("TERM", $kidpid); # send SIGTERM to child
1275 # the else{} block runs only in the child process
1277 # copy standard input to the socket
1278 while (defined ($line = <STDIN>)) {
1279 print $handle $line;
1283 The C<kill> function in the parent's C<if> block is there to send a
1284 signal to our child process (current running in the C<else> block)
1285 as soon as the remote server has closed its end of the connection.
1287 If the remote server sends data a byte at time, and you need that
1288 data immediately without waiting for a newline (which might not happen),
1289 you may wish to replace the C<while> loop in the parent with the
1293 while (sysread($handle, $byte, 1) == 1) {
1297 Making a system call for each byte you want to read is not very efficient
1298 (to put it mildly) but is the simplest to explain and works reasonably
1301 =head1 TCP Servers with IO::Socket
1303 As always, setting up a server is little bit more involved than running a client.
1304 The model is that the server creates a special kind of socket that
1305 does nothing but listen on a particular port for incoming connections.
1306 It does this by calling the C<< IO::Socket::INET->new() >> method with
1307 slightly different arguments than the client did.
1313 This is which protocol to use. Like our clients, we'll
1314 still specify C<"tcp"> here.
1319 port in the C<LocalPort> argument, which we didn't do for the client.
1320 This is service name or port number for which you want to be the
1321 server. (Under Unix, ports under 1024 are restricted to the
1322 superuser.) In our sample, we'll use port 9000, but you can use
1323 any port that's not currently in use on your system. If you try
1324 to use one already in used, you'll get an "Address already in use"
1325 message. Under Unix, the C<netstat -a> command will show
1326 which services current have servers.
1330 The C<Listen> parameter is set to the maximum number of
1331 pending connections we can accept until we turn away incoming clients.
1332 Think of it as a call-waiting queue for your telephone.
1333 The low-level Socket module has a special symbol for the system maximum, which
1338 The C<Reuse> parameter is needed so that we restart our server
1339 manually without waiting a few minutes to allow system buffers to
1344 Once the generic server socket has been created using the parameters
1345 listed above, the server then waits for a new client to connect
1346 to it. The server blocks in the C<accept> method, which eventually accepts a
1347 bidirectional connection from the remote client. (Make sure to autoflush
1348 this handle to circumvent buffering.)
1350 To add to user-friendliness, our server prompts the user for commands.
1351 Most servers don't do this. Because of the prompt without a newline,
1352 you'll have to use the C<sysread> variant of the interactive client above.
1354 This server accepts one of five different commands, sending output
1355 back to the client. Note that unlike most network servers, this one
1356 only handles one incoming client at a time. Multithreaded servers are
1357 covered in Chapter 6 of the Camel.
1359 Here's the code. We'll
1363 use Net::hostent; # for OO version of gethostbyaddr
1365 $PORT = 9000; # pick something not in use
1367 $server = IO::Socket::INET->new( Proto => 'tcp',
1369 Listen => SOMAXCONN,
1372 die "can't setup server" unless $server;
1373 print "[Server $0 accepting clients]\n";
1375 while ($client = $server->accept()) {
1376 $client->autoflush(1);
1377 print $client "Welcome to $0; type help for command list.\n";
1378 $hostinfo = gethostbyaddr($client->peeraddr);
1379 printf "[Connect from %s]\n", $hostinfo->name || $client->peerhost;
1380 print $client "Command? ";
1381 while ( <$client>) {
1382 next unless /\S/; # blank line
1383 if (/quit|exit/i) { last; }
1384 elsif (/date|time/i) { printf $client "%s\n", scalar localtime; }
1385 elsif (/who/i ) { print $client `who 2>&1`; }
1386 elsif (/cookie/i ) { print $client `/usr/games/fortune 2>&1`; }
1387 elsif (/motd/i ) { print $client `cat /etc/motd 2>&1`; }
1389 print $client "Commands: quit date who cookie motd\n";
1392 print $client "Command? ";
1397 =head1 UDP: Message Passing
1399 Another kind of client-server setup is one that uses not connections, but
1400 messages. UDP communications involve much lower overhead but also provide
1401 less reliability, as there are no promises that messages will arrive at
1402 all, let alone in order and unmangled. Still, UDP offers some advantages
1403 over TCP, including being able to "broadcast" or "multicast" to a whole
1404 bunch of destination hosts at once (usually on your local subnet). If you
1405 find yourself overly concerned about reliability and start building checks
1406 into your message system, then you probably should use just TCP to start
1409 Note that UDP datagrams are I<not> a bytestream and should not be treated
1410 as such. This makes using I/O mechanisms with internal buffering
1411 like stdio (i.e. print() and friends) especially cumbersome. Use syswrite(),
1412 or better send(), like in the example below.
1414 Here's a UDP program similar to the sample Internet TCP client given
1415 earlier. However, instead of checking one host at a time, the UDP version
1416 will check many of them asynchronously by simulating a multicast and then
1417 using select() to do a timed-out wait for I/O. To do something similar
1418 with TCP, you'd have to use a different socket handle for each host.
1425 my ( $count, $hisiaddr, $hispaddr, $histime,
1426 $host, $iaddr, $paddr, $port, $proto,
1427 $rin, $rout, $rtime, $SECS_of_70_YEARS);
1429 $SECS_of_70_YEARS = 2208988800;
1431 $iaddr = gethostbyname(hostname());
1432 $proto = getprotobyname('udp');
1433 $port = getservbyname('time', 'udp');
1434 $paddr = sockaddr_in(0, $iaddr); # 0 means let kernel pick
1436 socket(SOCKET, PF_INET, SOCK_DGRAM, $proto) || die "socket: $!";
1437 bind(SOCKET, $paddr) || die "bind: $!";
1440 printf "%-12s %8s %s\n", "localhost", 0, scalar localtime time;
1444 $hisiaddr = inet_aton($host) || die "unknown host";
1445 $hispaddr = sockaddr_in($port, $hisiaddr);
1446 defined(send(SOCKET, 0, 0, $hispaddr)) || die "send $host: $!";
1450 vec($rin, fileno(SOCKET), 1) = 1;
1452 # timeout after 10.0 seconds
1453 while ($count && select($rout = $rin, undef, undef, 10.0)) {
1455 ($hispaddr = recv(SOCKET, $rtime, 4, 0)) || die "recv: $!";
1456 ($port, $hisiaddr) = sockaddr_in($hispaddr);
1457 $host = gethostbyaddr($hisiaddr, AF_INET);
1458 $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ;
1459 printf "%-12s ", $host;
1460 printf "%8d %s\n", $histime - time, scalar localtime($histime);
1464 Note that this example does not include any retries and may consequently
1465 fail to contact a reachable host. The most prominent reason for this
1466 is congestion of the queues on the sending host if the number of
1467 list of hosts to contact is sufficiently large.
1471 While System V IPC isn't so widely used as sockets, it still has some
1472 interesting uses. You can't, however, effectively use SysV IPC or
1473 Berkeley mmap() to have shared memory so as to share a variable amongst
1474 several processes. That's because Perl would reallocate your string when
1475 you weren't wanting it to.
1477 Here's a small example showing shared memory usage.
1479 use IPC::SysV qw(IPC_PRIVATE IPC_RMID S_IRWXU);
1482 $id = shmget(IPC_PRIVATE, $size, S_IRWXU) || die "$!";
1483 print "shm key $id\n";
1485 $message = "Message #1";
1486 shmwrite($id, $message, 0, 60) || die "$!";
1487 print "wrote: '$message'\n";
1488 shmread($id, $buff, 0, 60) || die "$!";
1489 print "read : '$buff'\n";
1491 # the buffer of shmread is zero-character end-padded.
1492 substr($buff, index($buff, "\0")) = '';
1493 print "un" unless $buff eq $message;
1496 print "deleting shm $id\n";
1497 shmctl($id, IPC_RMID, 0) || die "$!";
1499 Here's an example of a semaphore:
1501 use IPC::SysV qw(IPC_CREAT);
1504 $id = semget($IPC_KEY, 10, 0666 | IPC_CREAT ) || die "$!";
1505 print "shm key $id\n";
1507 Put this code in a separate file to be run in more than one process.
1508 Call the file F<take>:
1510 # create a semaphore
1513 $id = semget($IPC_KEY, 0 , 0 );
1514 die if !defined($id);
1520 # wait for semaphore to be zero
1522 $opstring1 = pack("s!s!s!", $semnum, $semop, $semflag);
1524 # Increment the semaphore count
1526 $opstring2 = pack("s!s!s!", $semnum, $semop, $semflag);
1527 $opstring = $opstring1 . $opstring2;
1529 semop($id,$opstring) || die "$!";
1531 Put this code in a separate file to be run in more than one process.
1532 Call this file F<give>:
1534 # 'give' the semaphore
1535 # run this in the original process and you will see
1536 # that the second process continues
1539 $id = semget($IPC_KEY, 0, 0);
1540 die if !defined($id);
1545 # Decrement the semaphore count
1547 $opstring = pack("s!s!s!", $semnum, $semop, $semflag);
1549 semop($id,$opstring) || die "$!";
1551 The SysV IPC code above was written long ago, and it's definitely
1552 clunky looking. For a more modern look, see the IPC::SysV module
1553 which is included with Perl starting from Perl 5.005.
1555 A small example demonstrating SysV message queues:
1557 use IPC::SysV qw(IPC_PRIVATE IPC_RMID IPC_CREAT S_IRWXU);
1559 my $id = msgget(IPC_PRIVATE, IPC_CREAT | S_IRWXU);
1561 my $sent = "message";
1567 if (msgsnd($id, pack("l! a*", $type_sent, $sent), 0)) {
1568 if (msgrcv($id, $rcvd, 60, 0, 0)) {
1569 ($type_rcvd, $rcvd) = unpack("l! a*", $rcvd);
1570 if ($rcvd eq $sent) {
1576 die "# msgrcv failed\n";
1579 die "# msgsnd failed\n";
1581 msgctl($id, IPC_RMID, 0) || die "# msgctl failed: $!\n";
1583 die "# msgget failed\n";
1588 Most of these routines quietly but politely return C<undef> when they
1589 fail instead of causing your program to die right then and there due to
1590 an uncaught exception. (Actually, some of the new I<Socket> conversion
1591 functions croak() on bad arguments.) It is therefore essential to
1592 check return values from these functions. Always begin your socket
1593 programs this way for optimal success, and don't forget to add B<-T>
1594 taint checking flag to the #! line for servers:
1603 All these routines create system-specific portability problems. As noted
1604 elsewhere, Perl is at the mercy of your C libraries for much of its system
1605 behaviour. It's probably safest to assume broken SysV semantics for
1606 signals and to stick with simple TCP and UDP socket operations; e.g., don't
1607 try to pass open file descriptors over a local UDP datagram socket if you
1608 want your code to stand a chance of being portable.
1610 As mentioned in the signals section, because few vendors provide C
1611 libraries that are safely re-entrant, the prudent programmer will do
1612 little else within a handler beyond setting a numeric variable that
1613 already exists; or, if locked into a slow (restarting) system call,
1614 using die() to raise an exception and longjmp(3) out. In fact, even
1615 these may in some cases cause a core dump. It's probably best to avoid
1616 signals except where they are absolutely inevitable. This
1617 will be addressed in a future release of Perl.
1621 Tom Christiansen, with occasional vestiges of Larry Wall's original
1622 version and suggestions from the Perl Porters.
1626 There's a lot more to networking than this, but this should get you
1629 For intrepid programmers, the indispensable textbook is I<Unix Network
1630 Programming> by W. Richard Stevens (published by Addison-Wesley). Note
1631 that most books on networking address networking from the perspective of
1632 a C programmer; translation to Perl is left as an exercise for the reader.
1634 The IO::Socket(3) manpage describes the object library, and the Socket(3)
1635 manpage describes the low-level interface to sockets. Besides the obvious
1636 functions in L<perlfunc>, you should also check out the F<modules> file
1637 at your nearest CPAN site. (See L<perlmodlib> or best yet, the F<Perl
1638 FAQ> for a description of what CPAN is and where to get it.)
1640 Section 5 of the F<modules> file is devoted to "Networking, Device Control
1641 (modems), and Interprocess Communication", and contains numerous unbundled
1642 modules numerous networking modules, Chat and Expect operations, CGI
1643 programming, DCE, FTP, IPC, NNTP, Proxy, Ptty, RPC, SNMP, SMTP, Telnet,
1644 Threads, and ToolTalk--just to name a few.