Commit | Line | Data |
---|---|---|
a0d0e21e LW |
1 | =head1 NAME |
2 | ||
184e9718 | 3 | perlipc - Perl interprocess communication (signals, fifos, pipes, safe subprocesses, sockets, and semaphores) |
a0d0e21e LW |
4 | |
5 | =head1 DESCRIPTION | |
6 | ||
4633a7c4 LW |
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. | |
10 | ||
11 | =head1 Signals | |
12 | ||
13 | Perl uses a simple signal handling model: the %SIG hash contains names or | |
14 | references of user-installed signal handlers. These handlers will be called | |
15 | with an argument which is the name of the signal that triggered it. A | |
16 | signal may be generated intentionally from a particular keyboard sequence like | |
a2eb9003 | 17 | control-C or control-Z, sent to you from another process, or |
4633a7c4 | 18 | triggered automatically by the kernel when special events transpire, like |
54310121 | 19 | a child process exiting, your process running out of stack space, or |
4633a7c4 LW |
20 | hitting file size limit. |
21 | ||
22 | For example, to trap an interrupt signal, set up a handler like this. | |
7b05b7e3 TC |
23 | Do as little as you possibly can in your handler; notice how all we do is |
24 | set a global variable and then raise an exception. That's because on most | |
25 | systems, libraries are not re-entrant; particularly, memory allocation and | |
26 | I/O routines are not. That means that doing nearly I<anything> in your | |
27 | handler could in theory trigger a memory fault and subsequent core dump. | |
4633a7c4 LW |
28 | |
29 | sub catch_zap { | |
30 | my $signame = shift; | |
31 | $shucks++; | |
32 | die "Somebody sent me a SIG$signame"; | |
54310121 | 33 | } |
4633a7c4 LW |
34 | $SIG{INT} = 'catch_zap'; # could fail in modules |
35 | $SIG{INT} = \&catch_zap; # best strategy | |
36 | ||
37 | The names of the signals are the ones listed out by C<kill -l> on your | |
38 | system, or you can retrieve them from the Config module. Set up an | |
39 | @signame list indexed by number to get the name and a %signo table | |
40 | indexed by name to get the number: | |
41 | ||
42 | use Config; | |
43 | defined $Config{sig_name} || die "No sigs?"; | |
44 | foreach $name (split(' ', $Config{sig_name})) { | |
45 | $signo{$name} = $i; | |
46 | $signame[$i] = $name; | |
47 | $i++; | |
54310121 | 48 | } |
4633a7c4 | 49 | |
6a3992aa | 50 | So to check whether signal 17 and SIGALRM were the same, do just this: |
4633a7c4 LW |
51 | |
52 | print "signal #17 = $signame[17]\n"; | |
54310121 | 53 | if ($signo{ALRM}) { |
4633a7c4 | 54 | print "SIGALRM is $signo{ALRM}\n"; |
54310121 | 55 | } |
4633a7c4 LW |
56 | |
57 | You may also choose to assign the strings C<'IGNORE'> or C<'DEFAULT'> as | |
58 | the handler, in which case Perl will try to discard the signal or do the | |
f648820c GS |
59 | default thing. |
60 | ||
19799a22 | 61 | On most Unix platforms, the C<CHLD> (sometimes also known as C<CLD>) signal |
f648820c GS |
62 | has special behavior with respect to a value of C<'IGNORE'>. |
63 | Setting C<$SIG{CHLD}> to C<'IGNORE'> on such a platform has the effect of | |
64 | not creating zombie processes when the parent process fails to C<wait()> | |
65 | on its child processes (i.e. child processes are automatically reaped). | |
66 | Calling C<wait()> with C<$SIG{CHLD}> set to C<'IGNORE'> usually returns | |
67 | C<-1> on such platforms. | |
68 | ||
69 | Some signals can be neither trapped nor ignored, such as | |
4633a7c4 LW |
70 | the KILL and STOP (but not the TSTP) signals. One strategy for |
71 | temporarily ignoring signals is to use a local() statement, which will be | |
72 | automatically restored once your block is exited. (Remember that local() | |
73 | values are "inherited" by functions called from within that block.) | |
74 | ||
75 | sub precious { | |
76 | local $SIG{INT} = 'IGNORE'; | |
77 | &more_functions; | |
54310121 | 78 | } |
4633a7c4 LW |
79 | sub more_functions { |
80 | # interrupts still ignored, for now... | |
54310121 | 81 | } |
4633a7c4 LW |
82 | |
83 | Sending a signal to a negative process ID means that you send the signal | |
fb73857a | 84 | to the entire Unix process-group. This code sends a hang-up signal to all |
85 | processes in the current process group (and sets $SIG{HUP} to IGNORE so | |
86 | it doesn't kill itself): | |
4633a7c4 LW |
87 | |
88 | { | |
89 | local $SIG{HUP} = 'IGNORE'; | |
90 | kill HUP => -$$; | |
91 | # snazzy writing of: kill('HUP', -$$) | |
92 | } | |
a0d0e21e | 93 | |
4633a7c4 LW |
94 | Another interesting signal to send is signal number zero. This doesn't |
95 | actually affect another process, but instead checks whether it's alive | |
54310121 | 96 | or has changed its UID. |
a0d0e21e | 97 | |
4633a7c4 LW |
98 | unless (kill 0 => $kid_pid) { |
99 | warn "something wicked happened to $kid_pid"; | |
54310121 | 100 | } |
a0d0e21e | 101 | |
4633a7c4 LW |
102 | You might also want to employ anonymous functions for simple signal |
103 | handlers: | |
a0d0e21e | 104 | |
4633a7c4 | 105 | $SIG{INT} = sub { die "\nOutta here!\n" }; |
a0d0e21e | 106 | |
4633a7c4 | 107 | But that will be problematic for the more complicated handlers that need |
54310121 | 108 | to reinstall themselves. Because Perl's signal mechanism is currently |
184e9718 | 109 | based on the signal(3) function from the C library, you may sometimes be so |
4633a7c4 LW |
110 | misfortunate as to run on systems where that function is "broken", that |
111 | is, it behaves in the old unreliable SysV way rather than the newer, more | |
112 | reasonable BSD and POSIX fashion. So you'll see defensive people writing | |
113 | signal handlers like this: | |
a0d0e21e | 114 | |
54310121 | 115 | sub REAPER { |
4633a7c4 | 116 | $waitedpid = wait; |
6a3992aa DL |
117 | # loathe sysV: it makes us not only reinstate |
118 | # the handler, but place it after the wait | |
54310121 | 119 | $SIG{CHLD} = \&REAPER; |
4633a7c4 LW |
120 | } |
121 | $SIG{CHLD} = \&REAPER; | |
122 | # now do something that forks... | |
123 | ||
816229cf | 124 | or better still: |
4633a7c4 | 125 | |
6a3992aa | 126 | use POSIX ":sys_wait_h"; |
54310121 | 127 | sub REAPER { |
4633a7c4 | 128 | my $child; |
816229cf NC |
129 | # If a second child dies while in the signal handler caused by the |
130 | # first death, we won't get another signal. So must loop here else | |
131 | # we will leave the unreaped child as a zombie. And the next time | |
132 | # two children die we get another zombie. And so on. | |
1450d070 | 133 | while (($child = waitpid(-1,WNOHANG)) > 0) { |
4633a7c4 | 134 | $Kid_Status{$child} = $?; |
54310121 | 135 | } |
6a3992aa | 136 | $SIG{CHLD} = \&REAPER; # still loathe sysV |
4633a7c4 LW |
137 | } |
138 | $SIG{CHLD} = \&REAPER; | |
139 | # do something that forks... | |
140 | ||
141 | Signal handling is also used for timeouts in Unix, While safely | |
142 | protected within an C<eval{}> block, you set a signal handler to trap | |
143 | alarm signals and then schedule to have one delivered to you in some | |
144 | number of seconds. Then try your blocking operation, clearing the alarm | |
145 | when it's done but not before you've exited your C<eval{}> block. If it | |
146 | goes off, you'll use die() to jump out of the block, much as you might | |
147 | using longjmp() or throw() in other languages. | |
148 | ||
149 | Here's an example: | |
150 | ||
54310121 | 151 | eval { |
4633a7c4 | 152 | local $SIG{ALRM} = sub { die "alarm clock restart" }; |
54310121 | 153 | alarm 10; |
4633a7c4 | 154 | flock(FH, 2); # blocking write lock |
54310121 | 155 | alarm 0; |
4633a7c4 LW |
156 | }; |
157 | if ($@ and $@ !~ /alarm clock restart/) { die } | |
158 | ||
8a4f6ac2 GS |
159 | If the operation being timed out is system() or qx(), this technique |
160 | is liable to generate zombies. If this matters to you, you'll | |
161 | need to do your own fork() and exec(), and kill the errant child process. | |
162 | ||
4633a7c4 LW |
163 | For more complex signal handling, you might see the standard POSIX |
164 | module. Lamentably, this is almost entirely undocumented, but | |
165 | the F<t/lib/posix.t> file from the Perl source distribution has some | |
166 | examples in it. | |
167 | ||
168 | =head1 Named Pipes | |
169 | ||
170 | A named pipe (often referred to as a FIFO) is an old Unix IPC | |
171 | mechanism for processes communicating on the same machine. It works | |
54310121 | 172 | just like a regular, connected anonymous pipes, except that the |
4633a7c4 LW |
173 | processes rendezvous using a filename and don't have to be related. |
174 | ||
175 | To create a named pipe, use the Unix command mknod(1) or on some | |
176 | systems, mkfifo(1). These may not be in your normal path. | |
177 | ||
178 | # system return val is backwards, so && not || | |
179 | # | |
180 | $ENV{PATH} .= ":/etc:/usr/etc"; | |
54310121 | 181 | if ( system('mknod', $path, 'p') |
4633a7c4 LW |
182 | && system('mkfifo', $path) ) |
183 | { | |
5a964f20 | 184 | die "mk{nod,fifo} $path failed"; |
54310121 | 185 | } |
4633a7c4 LW |
186 | |
187 | ||
188 | A fifo is convenient when you want to connect a process to an unrelated | |
189 | one. When you open a fifo, the program will block until there's something | |
54310121 | 190 | on the other end. |
4633a7c4 LW |
191 | |
192 | For example, let's say you'd like to have your F<.signature> file be a | |
193 | named pipe that has a Perl program on the other end. Now every time any | |
6a3992aa | 194 | program (like a mailer, news reader, finger program, etc.) tries to read |
4633a7c4 | 195 | from that file, the reading program will block and your program will |
6a3992aa | 196 | supply the new signature. We'll use the pipe-checking file test B<-p> |
4633a7c4 LW |
197 | to find out whether anyone (or anything) has accidentally removed our fifo. |
198 | ||
199 | chdir; # go home | |
200 | $FIFO = '.signature'; | |
201 | $ENV{PATH} .= ":/etc:/usr/games"; | |
202 | ||
203 | while (1) { | |
204 | unless (-p $FIFO) { | |
205 | unlink $FIFO; | |
54310121 | 206 | system('mknod', $FIFO, 'p') |
4633a7c4 | 207 | && die "can't mknod $FIFO: $!"; |
54310121 | 208 | } |
4633a7c4 LW |
209 | |
210 | # next line blocks until there's a reader | |
211 | open (FIFO, "> $FIFO") || die "can't write $FIFO: $!"; | |
212 | print FIFO "John Smith (smith\@host.org)\n", `fortune -s`; | |
213 | close FIFO; | |
6a3992aa | 214 | sleep 2; # to avoid dup signals |
4633a7c4 | 215 | } |
a0d0e21e | 216 | |
5a964f20 TC |
217 | =head2 WARNING |
218 | ||
219 | By installing Perl code to deal with signals, you're exposing yourself | |
220 | to danger from two things. First, few system library functions are | |
221 | re-entrant. If the signal interrupts while Perl is executing one function | |
222 | (like malloc(3) or printf(3)), and your signal handler then calls the | |
223 | same function again, you could get unpredictable behavior--often, a | |
224 | core dump. Second, Perl isn't itself re-entrant at the lowest levels. | |
225 | If the signal interrupts Perl while Perl is changing its own internal | |
226 | data structures, similarly unpredictable behaviour may result. | |
227 | ||
228 | There are two things you can do, knowing this: be paranoid or be | |
229 | pragmatic. The paranoid approach is to do as little as possible in your | |
230 | signal handler. Set an existing integer variable that already has a | |
231 | value, and return. This doesn't help you if you're in a slow system call, | |
232 | which will just restart. That means you have to C<die> to longjump(3) out | |
233 | of the handler. Even this is a little cavalier for the true paranoiac, | |
234 | who avoids C<die> in a handler because the system I<is> out to get you. | |
235 | The pragmatic approach is to say ``I know the risks, but prefer the | |
236 | convenience'', and to do anything you want in your signal handler, | |
237 | prepared to clean up core dumps now and again. | |
238 | ||
239 | To forbid signal handlers altogether would bars you from | |
240 | many interesting programs, including virtually everything in this manpage, | |
f783e543 | 241 | since you could no longer even write SIGCHLD handlers. |
5a964f20 | 242 | |
a0d0e21e | 243 | |
4633a7c4 LW |
244 | =head1 Using open() for IPC |
245 | ||
246 | Perl's basic open() statement can also be used for unidirectional interprocess | |
247 | communication by either appending or prepending a pipe symbol to the second | |
a2eb9003 | 248 | argument to open(). Here's how to start something up in a child process you |
4633a7c4 LW |
249 | intend to write to: |
250 | ||
54310121 | 251 | open(SPOOLER, "| cat -v | lpr -h 2>/dev/null") |
4633a7c4 LW |
252 | || die "can't fork: $!"; |
253 | local $SIG{PIPE} = sub { die "spooler pipe broke" }; | |
254 | print SPOOLER "stuff\n"; | |
255 | close SPOOLER || die "bad spool: $! $?"; | |
256 | ||
257 | And here's how to start up a child process you intend to read from: | |
258 | ||
259 | open(STATUS, "netstat -an 2>&1 |") | |
260 | || die "can't fork: $!"; | |
261 | while (<STATUS>) { | |
262 | next if /^(tcp|udp)/; | |
263 | print; | |
54310121 | 264 | } |
a2eb9003 | 265 | close STATUS || die "bad netstat: $! $?"; |
4633a7c4 LW |
266 | |
267 | If one can be sure that a particular program is a Perl script that is | |
268 | expecting filenames in @ARGV, the clever programmer can write something | |
269 | like this: | |
270 | ||
5a964f20 | 271 | % program f1 "cmd1|" - f2 "cmd2|" f3 < tmpfile |
4633a7c4 LW |
272 | |
273 | and irrespective of which shell it's called from, the Perl program will | |
274 | read from the file F<f1>, the process F<cmd1>, standard input (F<tmpfile> | |
275 | in this case), the F<f2> file, the F<cmd2> command, and finally the F<f3> | |
276 | file. Pretty nifty, eh? | |
277 | ||
54310121 | 278 | You might notice that you could use backticks for much the |
4633a7c4 LW |
279 | same effect as opening a pipe for reading: |
280 | ||
281 | print grep { !/^(tcp|udp)/ } `netstat -an 2>&1`; | |
282 | die "bad netstat" if $?; | |
283 | ||
284 | While this is true on the surface, it's much more efficient to process the | |
285 | file one line or record at a time because then you don't have to read the | |
19799a22 | 286 | whole thing into memory at once. It also gives you finer control of the |
4633a7c4 LW |
287 | whole process, letting you to kill off the child process early if you'd |
288 | like. | |
289 | ||
290 | Be careful to check both the open() and the close() return values. If | |
291 | you're I<writing> to a pipe, you should also trap SIGPIPE. Otherwise, | |
292 | think of what happens when you start up a pipe to a command that doesn't | |
293 | exist: the open() will in all likelihood succeed (it only reflects the | |
294 | fork()'s success), but then your output will fail--spectacularly. Perl | |
295 | can't know whether the command worked because your command is actually | |
296 | running in a separate process whose exec() might have failed. Therefore, | |
6a3992aa | 297 | while readers of bogus commands return just a quick end of file, writers |
4633a7c4 LW |
298 | to bogus command will trigger a signal they'd better be prepared to |
299 | handle. Consider: | |
300 | ||
5a964f20 TC |
301 | open(FH, "|bogus") or die "can't fork: $!"; |
302 | print FH "bang\n" or die "can't write: $!"; | |
303 | close FH or die "can't close: $!"; | |
304 | ||
305 | That won't blow up until the close, and it will blow up with a SIGPIPE. | |
306 | To catch it, you could use this: | |
307 | ||
308 | $SIG{PIPE} = 'IGNORE'; | |
309 | open(FH, "|bogus") or die "can't fork: $!"; | |
310 | print FH "bang\n" or die "can't write: $!"; | |
311 | close FH or die "can't close: status=$?"; | |
4633a7c4 | 312 | |
68dc0745 | 313 | =head2 Filehandles |
314 | ||
5a964f20 TC |
315 | Both the main process and any child processes it forks share the same |
316 | STDIN, STDOUT, and STDERR filehandles. If both processes try to access | |
45bc9206 | 317 | them at once, strange things can happen. You may also want to close |
5a964f20 TC |
318 | or reopen the filehandles for the child. You can get around this by |
319 | opening your pipe with open(), but on some systems this means that the | |
320 | child process cannot outlive the parent. | |
68dc0745 | 321 | |
322 | =head2 Background Processes | |
323 | ||
324 | You can run a command in the background with: | |
325 | ||
7b05b7e3 | 326 | system("cmd &"); |
68dc0745 | 327 | |
328 | The command's STDOUT and STDERR (and possibly STDIN, depending on your | |
329 | shell) will be the same as the parent's. You won't need to catch | |
330 | SIGCHLD because of the double-fork taking place (see below for more | |
331 | details). | |
332 | ||
333 | =head2 Complete Dissociation of Child from Parent | |
334 | ||
335 | In some cases (starting server processes, for instance) you'll want to | |
893af57a RS |
336 | completely dissociate the child process from the parent. This is |
337 | often called daemonization. A well behaved daemon will also chdir() | |
338 | to the root directory (so it doesn't prevent unmounting the filesystem | |
339 | containing the directory from which it was launched) and redirect its | |
340 | standard file descriptors from and to F</dev/null> (so that random | |
341 | output doesn't wind up on the user's terminal). | |
342 | ||
343 | use POSIX 'setsid'; | |
344 | ||
345 | sub daemonize { | |
346 | chdir '/' or die "Can't chdir to /: $!"; | |
347 | open STDIN, '/dev/null' or die "Can't read /dev/null: $!"; | |
348 | open STDOUT, '>/dev/null' | |
349 | or die "Can't write to /dev/null: $!"; | |
350 | defined(my $pid = fork) or die "Can't fork: $!"; | |
351 | exit if $pid; | |
352 | setsid or die "Can't start a new session: $!"; | |
353 | open STDERR, '>&STDOUT' or die "Can't dup stdout: $!"; | |
354 | } | |
5a964f20 | 355 | |
893af57a RS |
356 | The fork() has to come before the setsid() to ensure that you aren't a |
357 | process group leader (the setsid() will fail if you are). If your | |
358 | system doesn't have the setsid() function, open F</dev/tty> and use the | |
359 | C<TIOCNOTTY> ioctl() on it instead. See L<tty(4)> for details. | |
5a964f20 | 360 | |
893af57a RS |
361 | Non-Unix users should check their Your_OS::Process module for other |
362 | solutions. | |
68dc0745 | 363 | |
4633a7c4 LW |
364 | =head2 Safe Pipe Opens |
365 | ||
366 | Another interesting approach to IPC is making your single program go | |
367 | multiprocess and communicate between (or even amongst) yourselves. The | |
368 | open() function will accept a file argument of either C<"-|"> or C<"|-"> | |
369 | to do a very interesting thing: it forks a child connected to the | |
370 | filehandle you've opened. The child is running the same program as the | |
371 | parent. This is useful for safely opening a file when running under an | |
372 | assumed UID or GID, for example. If you open a pipe I<to> minus, you can | |
373 | write to the filehandle you opened and your kid will find it in his | |
374 | STDIN. If you open a pipe I<from> minus, you can read from the filehandle | |
375 | you opened whatever your kid writes to his STDOUT. | |
376 | ||
377 | use English; | |
378 | my $sleep_count = 0; | |
379 | ||
54310121 | 380 | do { |
c07a80fd | 381 | $pid = open(KID_TO_WRITE, "|-"); |
4633a7c4 LW |
382 | unless (defined $pid) { |
383 | warn "cannot fork: $!"; | |
384 | die "bailing out" if $sleep_count++ > 6; | |
385 | sleep 10; | |
54310121 | 386 | } |
4633a7c4 LW |
387 | } until defined $pid; |
388 | ||
389 | if ($pid) { # parent | |
c07a80fd | 390 | print KID_TO_WRITE @some_data; |
391 | close(KID_TO_WRITE) || warn "kid exited $?"; | |
4633a7c4 LW |
392 | } else { # child |
393 | ($EUID, $EGID) = ($UID, $GID); # suid progs only | |
54310121 | 394 | open (FILE, "> /safe/file") |
4633a7c4 LW |
395 | || die "can't open /safe/file: $!"; |
396 | while (<STDIN>) { | |
397 | print FILE; # child's STDIN is parent's KID | |
54310121 | 398 | } |
4633a7c4 | 399 | exit; # don't forget this |
54310121 | 400 | } |
4633a7c4 LW |
401 | |
402 | Another common use for this construct is when you need to execute | |
403 | something without the shell's interference. With system(), it's | |
54310121 | 404 | straightforward, but you can't use a pipe open or backticks safely. |
4633a7c4 LW |
405 | That's because there's no way to stop the shell from getting its hands on |
406 | your arguments. Instead, use lower-level control to call exec() directly. | |
407 | ||
54310121 | 408 | Here's a safe backtick or pipe open for read: |
4633a7c4 LW |
409 | |
410 | # add error processing as above | |
c07a80fd | 411 | $pid = open(KID_TO_READ, "-|"); |
4633a7c4 LW |
412 | |
413 | if ($pid) { # parent | |
c07a80fd | 414 | while (<KID_TO_READ>) { |
4633a7c4 | 415 | # do something interesting |
54310121 | 416 | } |
c07a80fd | 417 | close(KID_TO_READ) || warn "kid exited $?"; |
4633a7c4 LW |
418 | |
419 | } else { # child | |
420 | ($EUID, $EGID) = ($UID, $GID); # suid only | |
421 | exec($program, @options, @args) | |
422 | || die "can't exec program: $!"; | |
423 | # NOTREACHED | |
54310121 | 424 | } |
4633a7c4 LW |
425 | |
426 | ||
427 | And here's a safe pipe open for writing: | |
428 | ||
429 | # add error processing as above | |
c07a80fd | 430 | $pid = open(KID_TO_WRITE, "|-"); |
4633a7c4 LW |
431 | $SIG{ALRM} = sub { die "whoops, $program pipe broke" }; |
432 | ||
433 | if ($pid) { # parent | |
434 | for (@data) { | |
c07a80fd | 435 | print KID_TO_WRITE; |
54310121 | 436 | } |
c07a80fd | 437 | close(KID_TO_WRITE) || warn "kid exited $?"; |
4633a7c4 LW |
438 | |
439 | } else { # child | |
440 | ($EUID, $EGID) = ($UID, $GID); | |
441 | exec($program, @options, @args) | |
442 | || die "can't exec program: $!"; | |
443 | # NOTREACHED | |
54310121 | 444 | } |
4633a7c4 LW |
445 | |
446 | Note that these operations are full Unix forks, which means they may not be | |
447 | correctly implemented on alien systems. Additionally, these are not true | |
54310121 | 448 | multithreading. If you'd like to learn more about threading, see the |
184e9718 | 449 | F<modules> file mentioned below in the SEE ALSO section. |
4633a7c4 | 450 | |
7b05b7e3 | 451 | =head2 Bidirectional Communication with Another Process |
4633a7c4 LW |
452 | |
453 | While this works reasonably well for unidirectional communication, what | |
454 | about bidirectional communication? The obvious thing you'd like to do | |
455 | doesn't actually work: | |
456 | ||
c07a80fd | 457 | open(PROG_FOR_READING_AND_WRITING, "| some program |") |
4633a7c4 | 458 | |
9f1b1f2d GS |
459 | and if you forget to use the C<use warnings> pragma or the B<-w> flag, |
460 | then you'll miss out entirely on the diagnostic message: | |
4633a7c4 LW |
461 | |
462 | Can't do bidirectional pipe at -e line 1. | |
463 | ||
464 | If you really want to, you can use the standard open2() library function | |
7b05b7e3 | 465 | to catch both ends. There's also an open3() for tridirectional I/O so you |
4633a7c4 LW |
466 | can also catch your child's STDERR, but doing so would then require an |
467 | awkward select() loop and wouldn't allow you to use normal Perl input | |
468 | operations. | |
469 | ||
470 | If you look at its source, you'll see that open2() uses low-level | |
5a964f20 | 471 | primitives like Unix pipe() and exec() calls to create all the connections. |
4633a7c4 LW |
472 | While it might have been slightly more efficient by using socketpair(), it |
473 | would have then been even less portable than it already is. The open2() | |
474 | and open3() functions are unlikely to work anywhere except on a Unix | |
475 | system or some other one purporting to be POSIX compliant. | |
476 | ||
477 | Here's an example of using open2(): | |
478 | ||
479 | use FileHandle; | |
480 | use IPC::Open2; | |
5a964f20 | 481 | $pid = open2(*Reader, *Writer, "cat -u -n" ); |
4633a7c4 LW |
482 | print Writer "stuff\n"; |
483 | $got = <Reader>; | |
484 | ||
6a3992aa DL |
485 | The problem with this is that Unix buffering is really going to |
486 | ruin your day. Even though your C<Writer> filehandle is auto-flushed, | |
4633a7c4 | 487 | and the process on the other end will get your data in a timely manner, |
6a3992aa | 488 | you can't usually do anything to force it to give it back to you |
54310121 | 489 | in a similarly quick fashion. In this case, we could, because we |
4633a7c4 LW |
490 | gave I<cat> a B<-u> flag to make it unbuffered. But very few Unix |
491 | commands are designed to operate over pipes, so this seldom works | |
54310121 | 492 | unless you yourself wrote the program on the other end of the |
4633a7c4 LW |
493 | double-ended pipe. |
494 | ||
54310121 | 495 | A solution to this is the nonstandard F<Comm.pl> library. It uses |
4633a7c4 LW |
496 | pseudo-ttys to make your program behave more reasonably: |
497 | ||
498 | require 'Comm.pl'; | |
499 | $ph = open_proc('cat -n'); | |
500 | for (1..10) { | |
501 | print $ph "a line\n"; | |
502 | print "got back ", scalar <$ph>; | |
503 | } | |
a0d0e21e | 504 | |
4633a7c4 | 505 | This way you don't have to have control over the source code of the |
54310121 | 506 | program you're using. The F<Comm> library also has expect() |
507 | and interact() functions. Find the library (and we hope its | |
4633a7c4 | 508 | successor F<IPC::Chat>) at your nearest CPAN archive as detailed |
184e9718 | 509 | in the SEE ALSO section below. |
a0d0e21e | 510 | |
c8db1d39 TC |
511 | The newer Expect.pm module from CPAN also addresses this kind of thing. |
512 | This module requires two other modules from CPAN: IO::Pty and IO::Stty. | |
513 | It sets up a pseudo-terminal to interact with programs that insist on | |
514 | using talking to the terminal device driver. If your system is | |
515 | amongst those supported, this may be your best bet. | |
516 | ||
5a964f20 TC |
517 | =head2 Bidirectional Communication with Yourself |
518 | ||
519 | If you want, you may make low-level pipe() and fork() | |
520 | to stitch this together by hand. This example only | |
521 | talks to itself, but you could reopen the appropriate | |
522 | handles to STDIN and STDOUT and call other processes. | |
523 | ||
524 | #!/usr/bin/perl -w | |
525 | # pipe1 - bidirectional communication using two pipe pairs | |
526 | # designed for the socketpair-challenged | |
527 | use IO::Handle; # thousands of lines just for autoflush :-( | |
528 | pipe(PARENT_RDR, CHILD_WTR); # XXX: failure? | |
529 | pipe(CHILD_RDR, PARENT_WTR); # XXX: failure? | |
530 | CHILD_WTR->autoflush(1); | |
531 | PARENT_WTR->autoflush(1); | |
532 | ||
533 | if ($pid = fork) { | |
534 | close PARENT_RDR; close PARENT_WTR; | |
535 | print CHILD_WTR "Parent Pid $$ is sending this\n"; | |
536 | chomp($line = <CHILD_RDR>); | |
537 | print "Parent Pid $$ just read this: `$line'\n"; | |
538 | close CHILD_RDR; close CHILD_WTR; | |
539 | waitpid($pid,0); | |
540 | } else { | |
541 | die "cannot fork: $!" unless defined $pid; | |
542 | close CHILD_RDR; close CHILD_WTR; | |
543 | chomp($line = <PARENT_RDR>); | |
544 | print "Child Pid $$ just read this: `$line'\n"; | |
545 | print PARENT_WTR "Child Pid $$ is sending this\n"; | |
546 | close PARENT_RDR; close PARENT_WTR; | |
547 | exit; | |
548 | } | |
549 | ||
550 | But you don't actually have to make two pipe calls. If you | |
551 | have the socketpair() system call, it will do this all for you. | |
552 | ||
553 | #!/usr/bin/perl -w | |
554 | # pipe2 - bidirectional communication using socketpair | |
555 | # "the best ones always go both ways" | |
556 | ||
557 | use Socket; | |
558 | use IO::Handle; # thousands of lines just for autoflush :-( | |
559 | # We say AF_UNIX because although *_LOCAL is the | |
560 | # POSIX 1003.1g form of the constant, many machines | |
561 | # still don't have it. | |
562 | socketpair(CHILD, PARENT, AF_UNIX, SOCK_STREAM, PF_UNSPEC) | |
563 | or die "socketpair: $!"; | |
564 | ||
565 | CHILD->autoflush(1); | |
566 | PARENT->autoflush(1); | |
567 | ||
568 | if ($pid = fork) { | |
569 | close PARENT; | |
570 | print CHILD "Parent Pid $$ is sending this\n"; | |
571 | chomp($line = <CHILD>); | |
572 | print "Parent Pid $$ just read this: `$line'\n"; | |
573 | close CHILD; | |
574 | waitpid($pid,0); | |
575 | } else { | |
576 | die "cannot fork: $!" unless defined $pid; | |
577 | close CHILD; | |
578 | chomp($line = <PARENT>); | |
579 | print "Child Pid $$ just read this: `$line'\n"; | |
580 | print PARENT "Child Pid $$ is sending this\n"; | |
581 | close PARENT; | |
582 | exit; | |
583 | } | |
584 | ||
4633a7c4 | 585 | =head1 Sockets: Client/Server Communication |
a0d0e21e | 586 | |
6a3992aa | 587 | While not limited to Unix-derived operating systems (e.g., WinSock on PCs |
4633a7c4 | 588 | provides socket support, as do some VMS libraries), you may not have |
184e9718 | 589 | sockets on your system, in which case this section probably isn't going to do |
6a3992aa DL |
590 | you much good. With sockets, you can do both virtual circuits (i.e., TCP |
591 | streams) and datagrams (i.e., UDP packets). You may be able to do even more | |
4633a7c4 LW |
592 | depending on your system. |
593 | ||
594 | The Perl function calls for dealing with sockets have the same names as | |
595 | the corresponding system calls in C, but their arguments tend to differ | |
596 | for two reasons: first, Perl filehandles work differently than C file | |
597 | descriptors. Second, Perl already knows the length of its strings, so you | |
598 | don't need to pass that information. | |
a0d0e21e | 599 | |
4633a7c4 LW |
600 | One of the major problems with old socket code in Perl was that it used |
601 | hard-coded values for some of the constants, which severely hurt | |
602 | portability. If you ever see code that does anything like explicitly | |
603 | setting C<$AF_INET = 2>, you know you're in for big trouble: An | |
604 | immeasurably superior approach is to use the C<Socket> module, which more | |
605 | reliably grants access to various constants and functions you'll need. | |
a0d0e21e | 606 | |
68dc0745 | 607 | If you're not writing a server/client for an existing protocol like |
608 | NNTP or SMTP, you should give some thought to how your server will | |
609 | know when the client has finished talking, and vice-versa. Most | |
610 | protocols are based on one-line messages and responses (so one party | |
4a6725af | 611 | knows the other has finished when a "\n" is received) or multi-line |
68dc0745 | 612 | messages and responses that end with a period on an empty line |
613 | ("\n.\n" terminates a message/response). | |
614 | ||
5a964f20 TC |
615 | =head2 Internet Line Terminators |
616 | ||
617 | The Internet line terminator is "\015\012". Under ASCII variants of | |
618 | Unix, that could usually be written as "\r\n", but under other systems, | |
619 | "\r\n" might at times be "\015\015\012", "\012\012\015", or something | |
620 | completely different. The standards specify writing "\015\012" to be | |
621 | conformant (be strict in what you provide), but they also recommend | |
622 | accepting a lone "\012" on input (but be lenient in what you require). | |
623 | We haven't always been very good about that in the code in this manpage, | |
624 | but unless you're on a Mac, you'll probably be ok. | |
625 | ||
4633a7c4 | 626 | =head2 Internet TCP Clients and Servers |
a0d0e21e | 627 | |
4633a7c4 LW |
628 | Use Internet-domain sockets when you want to do client-server |
629 | communication that might extend to machines outside of your own system. | |
630 | ||
631 | Here's a sample TCP client using Internet-domain sockets: | |
632 | ||
633 | #!/usr/bin/perl -w | |
4633a7c4 LW |
634 | use strict; |
635 | use Socket; | |
636 | my ($remote,$port, $iaddr, $paddr, $proto, $line); | |
637 | ||
638 | $remote = shift || 'localhost'; | |
639 | $port = shift || 2345; # random port | |
640 | if ($port =~ /\D/) { $port = getservbyname($port, 'tcp') } | |
641 | die "No port" unless $port; | |
642 | $iaddr = inet_aton($remote) || die "no host: $remote"; | |
643 | $paddr = sockaddr_in($port, $iaddr); | |
644 | ||
645 | $proto = getprotobyname('tcp'); | |
646 | socket(SOCK, PF_INET, SOCK_STREAM, $proto) || die "socket: $!"; | |
647 | connect(SOCK, $paddr) || die "connect: $!"; | |
54310121 | 648 | while (defined($line = <SOCK>)) { |
4633a7c4 | 649 | print $line; |
54310121 | 650 | } |
4633a7c4 LW |
651 | |
652 | close (SOCK) || die "close: $!"; | |
653 | exit; | |
654 | ||
655 | And here's a corresponding server to go along with it. We'll | |
656 | leave the address as INADDR_ANY so that the kernel can choose | |
54310121 | 657 | the appropriate interface on multihomed hosts. If you want sit |
c07a80fd | 658 | on a particular interface (like the external side of a gateway |
659 | or firewall machine), you should fill this in with your real address | |
660 | instead. | |
661 | ||
662 | #!/usr/bin/perl -Tw | |
c07a80fd | 663 | use strict; |
664 | BEGIN { $ENV{PATH} = '/usr/ucb:/bin' } | |
665 | use Socket; | |
666 | use Carp; | |
5865a7df | 667 | my $EOL = "\015\012"; |
c07a80fd | 668 | |
54310121 | 669 | sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" } |
c07a80fd | 670 | |
671 | my $port = shift || 2345; | |
672 | my $proto = getprotobyname('tcp'); | |
51ee6500 | 673 | |
5865a7df | 674 | ($port) = $port =~ /^(\d+)$/ or die "invalid port"; |
6a3992aa | 675 | |
c07a80fd | 676 | socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!"; |
54310121 | 677 | setsockopt(Server, SOL_SOCKET, SO_REUSEADDR, |
c07a80fd | 678 | pack("l", 1)) || die "setsockopt: $!"; |
679 | bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!"; | |
680 | listen(Server,SOMAXCONN) || die "listen: $!"; | |
681 | ||
682 | logmsg "server started on port $port"; | |
683 | ||
684 | my $paddr; | |
685 | ||
686 | $SIG{CHLD} = \&REAPER; | |
687 | ||
688 | for ( ; $paddr = accept(Client,Server); close Client) { | |
689 | my($port,$iaddr) = sockaddr_in($paddr); | |
690 | my $name = gethostbyaddr($iaddr,AF_INET); | |
691 | ||
54310121 | 692 | logmsg "connection from $name [", |
693 | inet_ntoa($iaddr), "] | |
c07a80fd | 694 | at port $port"; |
695 | ||
54310121 | 696 | print Client "Hello there, $name, it's now ", |
5a964f20 | 697 | scalar localtime, $EOL; |
54310121 | 698 | } |
c07a80fd | 699 | |
54310121 | 700 | And here's a multithreaded version. It's multithreaded in that |
701 | like most typical servers, it spawns (forks) a slave server to | |
c07a80fd | 702 | handle the client request so that the master server can quickly |
703 | go back to service a new client. | |
4633a7c4 LW |
704 | |
705 | #!/usr/bin/perl -Tw | |
4633a7c4 LW |
706 | use strict; |
707 | BEGIN { $ENV{PATH} = '/usr/ucb:/bin' } | |
a0d0e21e | 708 | use Socket; |
4633a7c4 | 709 | use Carp; |
5865a7df | 710 | my $EOL = "\015\012"; |
a0d0e21e | 711 | |
4633a7c4 | 712 | sub spawn; # forward declaration |
54310121 | 713 | sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" } |
a0d0e21e | 714 | |
4633a7c4 LW |
715 | my $port = shift || 2345; |
716 | my $proto = getprotobyname('tcp'); | |
51ee6500 | 717 | |
5865a7df | 718 | ($port) = $port =~ /^(\d+)$/ or die "invalid port"; |
54310121 | 719 | |
c07a80fd | 720 | socket(Server, PF_INET, SOCK_STREAM, $proto) || die "socket: $!"; |
54310121 | 721 | setsockopt(Server, SOL_SOCKET, SO_REUSEADDR, |
c07a80fd | 722 | pack("l", 1)) || die "setsockopt: $!"; |
723 | bind(Server, sockaddr_in($port, INADDR_ANY)) || die "bind: $!"; | |
724 | listen(Server,SOMAXCONN) || die "listen: $!"; | |
a0d0e21e | 725 | |
4633a7c4 | 726 | logmsg "server started on port $port"; |
a0d0e21e | 727 | |
4633a7c4 LW |
728 | my $waitedpid = 0; |
729 | my $paddr; | |
a0d0e21e | 730 | |
816229cf | 731 | use POSIX ":sys_wait_h"; |
54310121 | 732 | sub REAPER { |
816229cf NC |
733 | my $child; |
734 | while (($waitedpid = waitpid(-1,WNOHANG)) > 0) { | |
735 | logmsg "reaped $waitedpid" . ($? ? " with exit $?" : ''); | |
736 | } | |
6a3992aa | 737 | $SIG{CHLD} = \&REAPER; # loathe sysV |
4633a7c4 LW |
738 | } |
739 | ||
740 | $SIG{CHLD} = \&REAPER; | |
741 | ||
54310121 | 742 | for ( $waitedpid = 0; |
743 | ($paddr = accept(Client,Server)) || $waitedpid; | |
744 | $waitedpid = 0, close Client) | |
4633a7c4 | 745 | { |
6a3992aa | 746 | next if $waitedpid and not $paddr; |
4633a7c4 LW |
747 | my($port,$iaddr) = sockaddr_in($paddr); |
748 | my $name = gethostbyaddr($iaddr,AF_INET); | |
749 | ||
54310121 | 750 | logmsg "connection from $name [", |
751 | inet_ntoa($iaddr), "] | |
4633a7c4 | 752 | at port $port"; |
a0d0e21e | 753 | |
54310121 | 754 | spawn sub { |
b921b357 | 755 | $|=1; |
5a964f20 TC |
756 | print "Hello there, $name, it's now ", scalar localtime, $EOL; |
757 | exec '/usr/games/fortune' # XXX: `wrong' line terminators | |
4633a7c4 LW |
758 | or confess "can't exec fortune: $!"; |
759 | }; | |
a0d0e21e | 760 | |
54310121 | 761 | } |
a0d0e21e | 762 | |
4633a7c4 LW |
763 | sub spawn { |
764 | my $coderef = shift; | |
a0d0e21e | 765 | |
54310121 | 766 | unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') { |
4633a7c4 | 767 | confess "usage: spawn CODEREF"; |
a0d0e21e | 768 | } |
4633a7c4 LW |
769 | |
770 | my $pid; | |
771 | if (!defined($pid = fork)) { | |
772 | logmsg "cannot fork: $!"; | |
773 | return; | |
774 | } elsif ($pid) { | |
775 | logmsg "begat $pid"; | |
6a3992aa | 776 | return; # I'm the parent |
4633a7c4 | 777 | } |
6a3992aa | 778 | # else I'm the child -- go spawn |
4633a7c4 | 779 | |
c07a80fd | 780 | open(STDIN, "<&Client") || die "can't dup client to stdin"; |
781 | open(STDOUT, ">&Client") || die "can't dup client to stdout"; | |
4633a7c4 LW |
782 | ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr"; |
783 | exit &$coderef(); | |
54310121 | 784 | } |
4633a7c4 LW |
785 | |
786 | This server takes the trouble to clone off a child version via fork() for | |
787 | each incoming request. That way it can handle many requests at once, | |
788 | which you might not always want. Even if you don't fork(), the listen() | |
789 | will allow that many pending connections. Forking servers have to be | |
790 | particularly careful about cleaning up their dead children (called | |
791 | "zombies" in Unix parlance), because otherwise you'll quickly fill up your | |
792 | process table. | |
793 | ||
794 | We suggest that you use the B<-T> flag to use taint checking (see L<perlsec>) | |
795 | even if we aren't running setuid or setgid. This is always a good idea | |
796 | for servers and other programs run on behalf of someone else (like CGI | |
797 | scripts), because it lessens the chances that people from the outside will | |
798 | be able to compromise your system. | |
799 | ||
800 | Let's look at another TCP client. This one connects to the TCP "time" | |
801 | service on a number of different machines and shows how far their clocks | |
802 | differ from the system on which it's being run: | |
803 | ||
804 | #!/usr/bin/perl -w | |
4633a7c4 LW |
805 | use strict; |
806 | use Socket; | |
807 | ||
808 | my $SECS_of_70_YEARS = 2208988800; | |
54310121 | 809 | sub ctime { scalar localtime(shift) } |
4633a7c4 | 810 | |
54310121 | 811 | my $iaddr = gethostbyname('localhost'); |
812 | my $proto = getprotobyname('tcp'); | |
813 | my $port = getservbyname('time', 'tcp'); | |
4633a7c4 LW |
814 | my $paddr = sockaddr_in(0, $iaddr); |
815 | my($host); | |
816 | ||
817 | $| = 1; | |
818 | printf "%-24s %8s %s\n", "localhost", 0, ctime(time()); | |
819 | ||
820 | foreach $host (@ARGV) { | |
821 | printf "%-24s ", $host; | |
822 | my $hisiaddr = inet_aton($host) || die "unknown host"; | |
823 | my $hispaddr = sockaddr_in($port, $hisiaddr); | |
824 | socket(SOCKET, PF_INET, SOCK_STREAM, $proto) || die "socket: $!"; | |
825 | connect(SOCKET, $hispaddr) || die "bind: $!"; | |
826 | my $rtime = ' '; | |
827 | read(SOCKET, $rtime, 4); | |
828 | close(SOCKET); | |
829 | my $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ; | |
830 | printf "%8d %s\n", $histime - time, ctime($histime); | |
a0d0e21e LW |
831 | } |
832 | ||
4633a7c4 LW |
833 | =head2 Unix-Domain TCP Clients and Servers |
834 | ||
a2eb9003 | 835 | That's fine for Internet-domain clients and servers, but what about local |
4633a7c4 LW |
836 | communications? While you can use the same setup, sometimes you don't |
837 | want to. Unix-domain sockets are local to the current host, and are often | |
54310121 | 838 | used internally to implement pipes. Unlike Internet domain sockets, Unix |
4633a7c4 LW |
839 | domain sockets can show up in the file system with an ls(1) listing. |
840 | ||
5a964f20 | 841 | % ls -l /dev/log |
4633a7c4 | 842 | srw-rw-rw- 1 root 0 Oct 31 07:23 /dev/log |
a0d0e21e | 843 | |
4633a7c4 LW |
844 | You can test for these with Perl's B<-S> file test: |
845 | ||
846 | unless ( -S '/dev/log' ) { | |
3ba19564 | 847 | die "something's wicked with the log system"; |
54310121 | 848 | } |
4633a7c4 LW |
849 | |
850 | Here's a sample Unix-domain client: | |
851 | ||
852 | #!/usr/bin/perl -w | |
4633a7c4 LW |
853 | use Socket; |
854 | use strict; | |
855 | my ($rendezvous, $line); | |
856 | ||
857 | $rendezvous = shift || '/tmp/catsock'; | |
858 | socket(SOCK, PF_UNIX, SOCK_STREAM, 0) || die "socket: $!"; | |
9607fc9c | 859 | connect(SOCK, sockaddr_un($rendezvous)) || die "connect: $!"; |
54310121 | 860 | while (defined($line = <SOCK>)) { |
4633a7c4 | 861 | print $line; |
54310121 | 862 | } |
4633a7c4 LW |
863 | exit; |
864 | ||
5a964f20 TC |
865 | And here's a corresponding server. You don't have to worry about silly |
866 | network terminators here because Unix domain sockets are guaranteed | |
867 | to be on the localhost, and thus everything works right. | |
4633a7c4 LW |
868 | |
869 | #!/usr/bin/perl -Tw | |
4633a7c4 LW |
870 | use strict; |
871 | use Socket; | |
872 | use Carp; | |
873 | ||
874 | BEGIN { $ENV{PATH} = '/usr/ucb:/bin' } | |
5865a7df | 875 | sub spawn; # forward declaration |
5a964f20 | 876 | sub logmsg { print "$0 $$: @_ at ", scalar localtime, "\n" } |
4633a7c4 LW |
877 | |
878 | my $NAME = '/tmp/catsock'; | |
879 | my $uaddr = sockaddr_un($NAME); | |
880 | my $proto = getprotobyname('tcp'); | |
881 | ||
c07a80fd | 882 | socket(Server,PF_UNIX,SOCK_STREAM,0) || die "socket: $!"; |
4633a7c4 | 883 | unlink($NAME); |
c07a80fd | 884 | bind (Server, $uaddr) || die "bind: $!"; |
885 | listen(Server,SOMAXCONN) || die "listen: $!"; | |
4633a7c4 LW |
886 | |
887 | logmsg "server started on $NAME"; | |
888 | ||
5a964f20 TC |
889 | my $waitedpid; |
890 | ||
816229cf | 891 | use POSIX ":sys_wait_h"; |
5a964f20 | 892 | sub REAPER { |
816229cf NC |
893 | my $child; |
894 | while (($waitedpid = waitpid(-1,WNOHANG)) > 0) { | |
895 | logmsg "reaped $waitedpid" . ($? ? " with exit $?" : ''); | |
896 | } | |
5a964f20 | 897 | $SIG{CHLD} = \&REAPER; # loathe sysV |
5a964f20 TC |
898 | } |
899 | ||
4633a7c4 LW |
900 | $SIG{CHLD} = \&REAPER; |
901 | ||
5a964f20 | 902 | |
54310121 | 903 | for ( $waitedpid = 0; |
904 | accept(Client,Server) || $waitedpid; | |
905 | $waitedpid = 0, close Client) | |
4633a7c4 LW |
906 | { |
907 | next if $waitedpid; | |
908 | logmsg "connection on $NAME"; | |
54310121 | 909 | spawn sub { |
4633a7c4 LW |
910 | print "Hello there, it's now ", scalar localtime, "\n"; |
911 | exec '/usr/games/fortune' or die "can't exec fortune: $!"; | |
912 | }; | |
54310121 | 913 | } |
4633a7c4 | 914 | |
5865a7df NC |
915 | sub spawn { |
916 | my $coderef = shift; | |
917 | ||
918 | unless (@_ == 0 && $coderef && ref($coderef) eq 'CODE') { | |
919 | confess "usage: spawn CODEREF"; | |
920 | } | |
921 | ||
922 | my $pid; | |
923 | if (!defined($pid = fork)) { | |
924 | logmsg "cannot fork: $!"; | |
925 | return; | |
926 | } elsif ($pid) { | |
927 | logmsg "begat $pid"; | |
928 | return; # I'm the parent | |
929 | } | |
930 | # else I'm the child -- go spawn | |
931 | ||
932 | open(STDIN, "<&Client") || die "can't dup client to stdin"; | |
933 | open(STDOUT, ">&Client") || die "can't dup client to stdout"; | |
934 | ## open(STDERR, ">&STDOUT") || die "can't dup stdout to stderr"; | |
935 | exit &$coderef(); | |
936 | } | |
937 | ||
4633a7c4 LW |
938 | As you see, it's remarkably similar to the Internet domain TCP server, so |
939 | much so, in fact, that we've omitted several duplicate functions--spawn(), | |
940 | logmsg(), ctime(), and REAPER()--which are exactly the same as in the | |
941 | other server. | |
942 | ||
943 | So why would you ever want to use a Unix domain socket instead of a | |
944 | simpler named pipe? Because a named pipe doesn't give you sessions. You | |
945 | can't tell one process's data from another's. With socket programming, | |
946 | you get a separate session for each client: that's why accept() takes two | |
947 | arguments. | |
948 | ||
949 | For example, let's say that you have a long running database server daemon | |
950 | that you want folks from the World Wide Web to be able to access, but only | |
951 | if they go through a CGI interface. You'd have a small, simple CGI | |
952 | program that does whatever checks and logging you feel like, and then acts | |
953 | as a Unix-domain client and connects to your private server. | |
954 | ||
7b05b7e3 TC |
955 | =head1 TCP Clients with IO::Socket |
956 | ||
957 | For those preferring a higher-level interface to socket programming, the | |
958 | IO::Socket module provides an object-oriented approach. IO::Socket is | |
959 | included as part of the standard Perl distribution as of the 5.004 | |
960 | release. If you're running an earlier version of Perl, just fetch | |
106325ad | 961 | IO::Socket from CPAN, where you'll also find modules providing easy |
7b05b7e3 TC |
962 | interfaces to the following systems: DNS, FTP, Ident (RFC 931), NIS and |
963 | NISPlus, NNTP, Ping, POP3, SMTP, SNMP, SSLeay, Telnet, and Time--just | |
964 | to name a few. | |
965 | ||
966 | =head2 A Simple Client | |
967 | ||
968 | Here's a client that creates a TCP connection to the "daytime" | |
969 | service at port 13 of the host name "localhost" and prints out everything | |
970 | that the server there cares to provide. | |
971 | ||
972 | #!/usr/bin/perl -w | |
973 | use IO::Socket; | |
974 | $remote = IO::Socket::INET->new( | |
975 | Proto => "tcp", | |
976 | PeerAddr => "localhost", | |
977 | PeerPort => "daytime(13)", | |
978 | ) | |
979 | or die "cannot connect to daytime port at localhost"; | |
980 | while ( <$remote> ) { print } | |
981 | ||
982 | When you run this program, you should get something back that | |
983 | looks like this: | |
984 | ||
985 | Wed May 14 08:40:46 MDT 1997 | |
986 | ||
987 | Here are what those parameters to the C<new> constructor mean: | |
988 | ||
13a2d996 | 989 | =over 4 |
7b05b7e3 TC |
990 | |
991 | =item C<Proto> | |
992 | ||
993 | This is which protocol to use. In this case, the socket handle returned | |
994 | will be connected to a TCP socket, because we want a stream-oriented | |
995 | connection, that is, one that acts pretty much like a plain old file. | |
996 | Not all sockets are this of this type. For example, the UDP protocol | |
997 | can be used to make a datagram socket, used for message-passing. | |
998 | ||
999 | =item C<PeerAddr> | |
1000 | ||
1001 | This is the name or Internet address of the remote host the server is | |
1002 | running on. We could have specified a longer name like C<"www.perl.com">, | |
1003 | or an address like C<"204.148.40.9">. For demonstration purposes, we've | |
1004 | used the special hostname C<"localhost">, which should always mean the | |
1005 | current machine you're running on. The corresponding Internet address | |
1006 | for localhost is C<"127.1">, if you'd rather use that. | |
1007 | ||
1008 | =item C<PeerPort> | |
1009 | ||
1010 | This is the service name or port number we'd like to connect to. | |
1011 | We could have gotten away with using just C<"daytime"> on systems with a | |
1012 | well-configured system services file,[FOOTNOTE: The system services file | |
1013 | is in I</etc/services> under Unix] but just in case, we've specified the | |
1014 | port number (13) in parentheses. Using just the number would also have | |
1015 | worked, but constant numbers make careful programmers nervous. | |
1016 | ||
1017 | =back | |
1018 | ||
1019 | Notice how the return value from the C<new> constructor is used as | |
1020 | a filehandle in the C<while> loop? That's what's called an indirect | |
1021 | filehandle, a scalar variable containing a filehandle. You can use | |
1022 | it the same way you would a normal filehandle. For example, you | |
1023 | can read one line from it this way: | |
1024 | ||
1025 | $line = <$handle>; | |
1026 | ||
1027 | all remaining lines from is this way: | |
1028 | ||
1029 | @lines = <$handle>; | |
1030 | ||
1031 | and send a line of data to it this way: | |
1032 | ||
1033 | print $handle "some data\n"; | |
1034 | ||
1035 | =head2 A Webget Client | |
1036 | ||
1037 | Here's a simple client that takes a remote host to fetch a document | |
1038 | from, and then a list of documents to get from that host. This is a | |
1039 | more interesting client than the previous one because it first sends | |
1040 | something to the server before fetching the server's response. | |
1041 | ||
1042 | #!/usr/bin/perl -w | |
1043 | use IO::Socket; | |
1044 | unless (@ARGV > 1) { die "usage: $0 host document ..." } | |
1045 | $host = shift(@ARGV); | |
5a964f20 TC |
1046 | $EOL = "\015\012"; |
1047 | $BLANK = $EOL x 2; | |
7b05b7e3 TC |
1048 | foreach $document ( @ARGV ) { |
1049 | $remote = IO::Socket::INET->new( Proto => "tcp", | |
1050 | PeerAddr => $host, | |
1051 | PeerPort => "http(80)", | |
1052 | ); | |
1053 | unless ($remote) { die "cannot connect to http daemon on $host" } | |
1054 | $remote->autoflush(1); | |
5a964f20 | 1055 | print $remote "GET $document HTTP/1.0" . $BLANK; |
7b05b7e3 TC |
1056 | while ( <$remote> ) { print } |
1057 | close $remote; | |
1058 | } | |
1059 | ||
1060 | The web server handing the "http" service, which is assumed to be at | |
4375e838 | 1061 | its standard port, number 80. If the web server you're trying to |
7b05b7e3 | 1062 | connect to is at a different port (like 1080 or 8080), you should specify |
c47ff5f1 | 1063 | as the named-parameter pair, C<< PeerPort => 8080 >>. The C<autoflush> |
7b05b7e3 TC |
1064 | method is used on the socket because otherwise the system would buffer |
1065 | up the output we sent it. (If you're on a Mac, you'll also need to | |
1066 | change every C<"\n"> in your code that sends data over the network to | |
1067 | be a C<"\015\012"> instead.) | |
1068 | ||
1069 | Connecting to the server is only the first part of the process: once you | |
1070 | have the connection, you have to use the server's language. Each server | |
1071 | on the network has its own little command language that it expects as | |
1072 | input. The string that we send to the server starting with "GET" is in | |
1073 | HTTP syntax. In this case, we simply request each specified document. | |
1074 | Yes, we really are making a new connection for each document, even though | |
1075 | it's the same host. That's the way you always used to have to speak HTTP. | |
1076 | Recent versions of web browsers may request that the remote server leave | |
1077 | the connection open a little while, but the server doesn't have to honor | |
1078 | such a request. | |
1079 | ||
1080 | Here's an example of running that program, which we'll call I<webget>: | |
1081 | ||
5a964f20 | 1082 | % webget www.perl.com /guanaco.html |
7b05b7e3 TC |
1083 | HTTP/1.1 404 File Not Found |
1084 | Date: Thu, 08 May 1997 18:02:32 GMT | |
1085 | Server: Apache/1.2b6 | |
1086 | Connection: close | |
1087 | Content-type: text/html | |
1088 | ||
1089 | <HEAD><TITLE>404 File Not Found</TITLE></HEAD> | |
1090 | <BODY><H1>File Not Found</H1> | |
1091 | The requested URL /guanaco.html was not found on this server.<P> | |
1092 | </BODY> | |
1093 | ||
1094 | Ok, so that's not very interesting, because it didn't find that | |
1095 | particular document. But a long response wouldn't have fit on this page. | |
1096 | ||
1097 | For a more fully-featured version of this program, you should look to | |
1098 | the I<lwp-request> program included with the LWP modules from CPAN. | |
1099 | ||
1100 | =head2 Interactive Client with IO::Socket | |
1101 | ||
1102 | Well, that's all fine if you want to send one command and get one answer, | |
1103 | but what about setting up something fully interactive, somewhat like | |
1104 | the way I<telnet> works? That way you can type a line, get the answer, | |
1105 | type a line, get the answer, etc. | |
1106 | ||
1107 | This client is more complicated than the two we've done so far, but if | |
1108 | you're on a system that supports the powerful C<fork> call, the solution | |
1109 | isn't that rough. Once you've made the connection to whatever service | |
1110 | you'd like to chat with, call C<fork> to clone your process. Each of | |
1111 | these two identical process has a very simple job to do: the parent | |
1112 | copies everything from the socket to standard output, while the child | |
1113 | simultaneously copies everything from standard input to the socket. | |
1114 | To accomplish the same thing using just one process would be I<much> | |
1115 | harder, because it's easier to code two processes to do one thing than it | |
1116 | is to code one process to do two things. (This keep-it-simple principle | |
5a964f20 TC |
1117 | a cornerstones of the Unix philosophy, and good software engineering as |
1118 | well, which is probably why it's spread to other systems.) | |
7b05b7e3 TC |
1119 | |
1120 | Here's the code: | |
1121 | ||
1122 | #!/usr/bin/perl -w | |
1123 | use strict; | |
1124 | use IO::Socket; | |
1125 | my ($host, $port, $kidpid, $handle, $line); | |
1126 | ||
1127 | unless (@ARGV == 2) { die "usage: $0 host port" } | |
1128 | ($host, $port) = @ARGV; | |
1129 | ||
1130 | # create a tcp connection to the specified host and port | |
1131 | $handle = IO::Socket::INET->new(Proto => "tcp", | |
1132 | PeerAddr => $host, | |
1133 | PeerPort => $port) | |
1134 | or die "can't connect to port $port on $host: $!"; | |
1135 | ||
1136 | $handle->autoflush(1); # so output gets there right away | |
1137 | print STDERR "[Connected to $host:$port]\n"; | |
1138 | ||
1139 | # split the program into two processes, identical twins | |
1140 | die "can't fork: $!" unless defined($kidpid = fork()); | |
1141 | ||
1142 | # the if{} block runs only in the parent process | |
1143 | if ($kidpid) { | |
1144 | # copy the socket to standard output | |
1145 | while (defined ($line = <$handle>)) { | |
1146 | print STDOUT $line; | |
1147 | } | |
1148 | kill("TERM", $kidpid); # send SIGTERM to child | |
1149 | } | |
1150 | # the else{} block runs only in the child process | |
1151 | else { | |
1152 | # copy standard input to the socket | |
1153 | while (defined ($line = <STDIN>)) { | |
1154 | print $handle $line; | |
1155 | } | |
1156 | } | |
1157 | ||
1158 | The C<kill> function in the parent's C<if> block is there to send a | |
1159 | signal to our child process (current running in the C<else> block) | |
1160 | as soon as the remote server has closed its end of the connection. | |
1161 | ||
7b05b7e3 TC |
1162 | If the remote server sends data a byte at time, and you need that |
1163 | data immediately without waiting for a newline (which might not happen), | |
1164 | you may wish to replace the C<while> loop in the parent with the | |
1165 | following: | |
1166 | ||
1167 | my $byte; | |
1168 | while (sysread($handle, $byte, 1) == 1) { | |
1169 | print STDOUT $byte; | |
1170 | } | |
1171 | ||
1172 | Making a system call for each byte you want to read is not very efficient | |
1173 | (to put it mildly) but is the simplest to explain and works reasonably | |
1174 | well. | |
1175 | ||
1176 | =head1 TCP Servers with IO::Socket | |
1177 | ||
5a964f20 | 1178 | As always, setting up a server is little bit more involved than running a client. |
7b05b7e3 TC |
1179 | The model is that the server creates a special kind of socket that |
1180 | does nothing but listen on a particular port for incoming connections. | |
c47ff5f1 | 1181 | It does this by calling the C<< IO::Socket::INET->new() >> method with |
7b05b7e3 TC |
1182 | slightly different arguments than the client did. |
1183 | ||
13a2d996 | 1184 | =over 4 |
7b05b7e3 TC |
1185 | |
1186 | =item Proto | |
1187 | ||
1188 | This is which protocol to use. Like our clients, we'll | |
1189 | still specify C<"tcp"> here. | |
1190 | ||
1191 | =item LocalPort | |
1192 | ||
1193 | We specify a local | |
1194 | port in the C<LocalPort> argument, which we didn't do for the client. | |
1195 | This is service name or port number for which you want to be the | |
1196 | server. (Under Unix, ports under 1024 are restricted to the | |
1197 | superuser.) In our sample, we'll use port 9000, but you can use | |
1198 | any port that's not currently in use on your system. If you try | |
1199 | to use one already in used, you'll get an "Address already in use" | |
19799a22 | 1200 | message. Under Unix, the C<netstat -a> command will show |
7b05b7e3 TC |
1201 | which services current have servers. |
1202 | ||
1203 | =item Listen | |
1204 | ||
1205 | The C<Listen> parameter is set to the maximum number of | |
1206 | pending connections we can accept until we turn away incoming clients. | |
1207 | Think of it as a call-waiting queue for your telephone. | |
1208 | The low-level Socket module has a special symbol for the system maximum, which | |
1209 | is SOMAXCONN. | |
1210 | ||
1211 | =item Reuse | |
1212 | ||
1213 | The C<Reuse> parameter is needed so that we restart our server | |
1214 | manually without waiting a few minutes to allow system buffers to | |
1215 | clear out. | |
1216 | ||
1217 | =back | |
1218 | ||
1219 | Once the generic server socket has been created using the parameters | |
1220 | listed above, the server then waits for a new client to connect | |
d1be9408 JF |
1221 | to it. The server blocks in the C<accept> method, which eventually accepts a |
1222 | bidirectional connection from the remote client. (Make sure to autoflush | |
7b05b7e3 TC |
1223 | this handle to circumvent buffering.) |
1224 | ||
1225 | To add to user-friendliness, our server prompts the user for commands. | |
1226 | Most servers don't do this. Because of the prompt without a newline, | |
1227 | you'll have to use the C<sysread> variant of the interactive client above. | |
1228 | ||
1229 | This server accepts one of five different commands, sending output | |
1230 | back to the client. Note that unlike most network servers, this one | |
1231 | only handles one incoming client at a time. Multithreaded servers are | |
f83494b9 | 1232 | covered in Chapter 6 of the Camel. |
7b05b7e3 TC |
1233 | |
1234 | Here's the code. We'll | |
1235 | ||
1236 | #!/usr/bin/perl -w | |
1237 | use IO::Socket; | |
1238 | use Net::hostent; # for OO version of gethostbyaddr | |
1239 | ||
1240 | $PORT = 9000; # pick something not in use | |
1241 | ||
1242 | $server = IO::Socket::INET->new( Proto => 'tcp', | |
1243 | LocalPort => $PORT, | |
1244 | Listen => SOMAXCONN, | |
1245 | Reuse => 1); | |
1246 | ||
1247 | die "can't setup server" unless $server; | |
1248 | print "[Server $0 accepting clients]\n"; | |
1249 | ||
1250 | while ($client = $server->accept()) { | |
1251 | $client->autoflush(1); | |
1252 | print $client "Welcome to $0; type help for command list.\n"; | |
1253 | $hostinfo = gethostbyaddr($client->peeraddr); | |
1254 | printf "[Connect from %s]\n", $hostinfo->name || $client->peerhost; | |
1255 | print $client "Command? "; | |
1256 | while ( <$client>) { | |
1257 | next unless /\S/; # blank line | |
1258 | if (/quit|exit/i) { last; } | |
1259 | elsif (/date|time/i) { printf $client "%s\n", scalar localtime; } | |
1260 | elsif (/who/i ) { print $client `who 2>&1`; } | |
1261 | elsif (/cookie/i ) { print $client `/usr/games/fortune 2>&1`; } | |
1262 | elsif (/motd/i ) { print $client `cat /etc/motd 2>&1`; } | |
1263 | else { | |
1264 | print $client "Commands: quit date who cookie motd\n"; | |
1265 | } | |
1266 | } continue { | |
1267 | print $client "Command? "; | |
1268 | } | |
1269 | close $client; | |
1270 | } | |
1271 | ||
1272 | =head1 UDP: Message Passing | |
4633a7c4 LW |
1273 | |
1274 | Another kind of client-server setup is one that uses not connections, but | |
1275 | messages. UDP communications involve much lower overhead but also provide | |
1276 | less reliability, as there are no promises that messages will arrive at | |
1277 | all, let alone in order and unmangled. Still, UDP offers some advantages | |
1278 | over TCP, including being able to "broadcast" or "multicast" to a whole | |
1279 | bunch of destination hosts at once (usually on your local subnet). If you | |
1280 | find yourself overly concerned about reliability and start building checks | |
6a3992aa | 1281 | into your message system, then you probably should use just TCP to start |
4633a7c4 LW |
1282 | with. |
1283 | ||
90034919 LC |
1284 | Note that UDP datagrams are I<not> a bytestream and should not be treated |
1285 | as such. This makes using I/O mechanisms with internal buffering | |
1286 | like stdio (i.e. print() and friends) especially cumbersome. Use syswrite(), | |
1287 | or better send(), like in the example below. | |
1288 | ||
4633a7c4 | 1289 | Here's a UDP program similar to the sample Internet TCP client given |
7b05b7e3 | 1290 | earlier. However, instead of checking one host at a time, the UDP version |
4633a7c4 LW |
1291 | will check many of them asynchronously by simulating a multicast and then |
1292 | using select() to do a timed-out wait for I/O. To do something similar | |
1293 | with TCP, you'd have to use a different socket handle for each host. | |
1294 | ||
1295 | #!/usr/bin/perl -w | |
1296 | use strict; | |
4633a7c4 LW |
1297 | use Socket; |
1298 | use Sys::Hostname; | |
1299 | ||
54310121 | 1300 | my ( $count, $hisiaddr, $hispaddr, $histime, |
1301 | $host, $iaddr, $paddr, $port, $proto, | |
4633a7c4 LW |
1302 | $rin, $rout, $rtime, $SECS_of_70_YEARS); |
1303 | ||
1304 | $SECS_of_70_YEARS = 2208988800; | |
1305 | ||
1306 | $iaddr = gethostbyname(hostname()); | |
1307 | $proto = getprotobyname('udp'); | |
1308 | $port = getservbyname('time', 'udp'); | |
1309 | $paddr = sockaddr_in(0, $iaddr); # 0 means let kernel pick | |
1310 | ||
1311 | socket(SOCKET, PF_INET, SOCK_DGRAM, $proto) || die "socket: $!"; | |
1312 | bind(SOCKET, $paddr) || die "bind: $!"; | |
1313 | ||
1314 | $| = 1; | |
1315 | printf "%-12s %8s %s\n", "localhost", 0, scalar localtime time; | |
1316 | $count = 0; | |
1317 | for $host (@ARGV) { | |
1318 | $count++; | |
1319 | $hisiaddr = inet_aton($host) || die "unknown host"; | |
1320 | $hispaddr = sockaddr_in($port, $hisiaddr); | |
1321 | defined(send(SOCKET, 0, 0, $hispaddr)) || die "send $host: $!"; | |
1322 | } | |
1323 | ||
1324 | $rin = ''; | |
1325 | vec($rin, fileno(SOCKET), 1) = 1; | |
1326 | ||
1327 | # timeout after 10.0 seconds | |
1328 | while ($count && select($rout = $rin, undef, undef, 10.0)) { | |
1329 | $rtime = ''; | |
1330 | ($hispaddr = recv(SOCKET, $rtime, 4, 0)) || die "recv: $!"; | |
1331 | ($port, $hisiaddr) = sockaddr_in($hispaddr); | |
1332 | $host = gethostbyaddr($hisiaddr, AF_INET); | |
1333 | $histime = unpack("N", $rtime) - $SECS_of_70_YEARS ; | |
1334 | printf "%-12s ", $host; | |
1335 | printf "%8d %s\n", $histime - time, scalar localtime($histime); | |
1336 | $count--; | |
1337 | } | |
1338 | ||
90034919 LC |
1339 | Note that this example does not include any retries and may consequently |
1340 | fail to contact a reachable host. The most prominent reason for this | |
1341 | is congestion of the queues on the sending host if the number of | |
a31a806a | 1342 | list of hosts to contact is sufficiently large. |
90034919 | 1343 | |
4633a7c4 LW |
1344 | =head1 SysV IPC |
1345 | ||
1346 | While System V IPC isn't so widely used as sockets, it still has some | |
1347 | interesting uses. You can't, however, effectively use SysV IPC or | |
1348 | Berkeley mmap() to have shared memory so as to share a variable amongst | |
1349 | several processes. That's because Perl would reallocate your string when | |
1350 | you weren't wanting it to. | |
1351 | ||
54310121 | 1352 | Here's a small example showing shared memory usage. |
a0d0e21e | 1353 | |
41d6edb2 | 1354 | use IPC::SysV qw(IPC_PRIVATE IPC_RMID S_IRWXU); |
0ade1984 | 1355 | |
a0d0e21e | 1356 | $size = 2000; |
41d6edb2 JH |
1357 | $id = shmget(IPC_PRIVATE, $size, S_IRWXU) || die "$!"; |
1358 | print "shm key $id\n"; | |
a0d0e21e LW |
1359 | |
1360 | $message = "Message #1"; | |
41d6edb2 | 1361 | shmwrite($id, $message, 0, 60) || die "$!"; |
0ade1984 | 1362 | print "wrote: '$message'\n"; |
41d6edb2 | 1363 | shmread($id, $buff, 0, 60) || die "$!"; |
0ade1984 | 1364 | print "read : '$buff'\n"; |
a0d0e21e | 1365 | |
0ade1984 JH |
1366 | # the buffer of shmread is zero-character end-padded. |
1367 | substr($buff, index($buff, "\0")) = ''; | |
1368 | print "un" unless $buff eq $message; | |
1369 | print "swell\n"; | |
a0d0e21e | 1370 | |
41d6edb2 JH |
1371 | print "deleting shm $id\n"; |
1372 | shmctl($id, IPC_RMID, 0) || die "$!"; | |
a0d0e21e LW |
1373 | |
1374 | Here's an example of a semaphore: | |
1375 | ||
0ade1984 JH |
1376 | use IPC::SysV qw(IPC_CREAT); |
1377 | ||
a0d0e21e | 1378 | $IPC_KEY = 1234; |
41d6edb2 JH |
1379 | $id = semget($IPC_KEY, 10, 0666 | IPC_CREAT ) || die "$!"; |
1380 | print "shm key $id\n"; | |
a0d0e21e | 1381 | |
a2eb9003 | 1382 | Put this code in a separate file to be run in more than one process. |
a0d0e21e LW |
1383 | Call the file F<take>: |
1384 | ||
1385 | # create a semaphore | |
1386 | ||
1387 | $IPC_KEY = 1234; | |
41d6edb2 JH |
1388 | $id = semget($IPC_KEY, 0 , 0 ); |
1389 | die if !defined($id); | |
a0d0e21e LW |
1390 | |
1391 | $semnum = 0; | |
1392 | $semflag = 0; | |
1393 | ||
1394 | # 'take' semaphore | |
1395 | # wait for semaphore to be zero | |
1396 | $semop = 0; | |
41d6edb2 | 1397 | $opstring1 = pack("s!s!s!", $semnum, $semop, $semflag); |
a0d0e21e LW |
1398 | |
1399 | # Increment the semaphore count | |
1400 | $semop = 1; | |
41d6edb2 | 1401 | $opstring2 = pack("s!s!s!", $semnum, $semop, $semflag); |
a0d0e21e LW |
1402 | $opstring = $opstring1 . $opstring2; |
1403 | ||
41d6edb2 | 1404 | semop($id,$opstring) || die "$!"; |
a0d0e21e | 1405 | |
a2eb9003 | 1406 | Put this code in a separate file to be run in more than one process. |
a0d0e21e LW |
1407 | Call this file F<give>: |
1408 | ||
4633a7c4 | 1409 | # 'give' the semaphore |
a0d0e21e LW |
1410 | # run this in the original process and you will see |
1411 | # that the second process continues | |
1412 | ||
1413 | $IPC_KEY = 1234; | |
41d6edb2 JH |
1414 | $id = semget($IPC_KEY, 0, 0); |
1415 | die if !defined($id); | |
a0d0e21e LW |
1416 | |
1417 | $semnum = 0; | |
1418 | $semflag = 0; | |
1419 | ||
1420 | # Decrement the semaphore count | |
1421 | $semop = -1; | |
41d6edb2 | 1422 | $opstring = pack("s!s!s!", $semnum, $semop, $semflag); |
a0d0e21e | 1423 | |
41d6edb2 | 1424 | semop($id,$opstring) || die "$!"; |
a0d0e21e | 1425 | |
7b05b7e3 | 1426 | The SysV IPC code above was written long ago, and it's definitely |
0ade1984 JH |
1427 | clunky looking. For a more modern look, see the IPC::SysV module |
1428 | which is included with Perl starting from Perl 5.005. | |
4633a7c4 | 1429 | |
41d6edb2 JH |
1430 | A small example demonstrating SysV message queues: |
1431 | ||
1432 | use IPC::SysV qw(IPC_PRIVATE IPC_RMID IPC_CREAT S_IRWXU); | |
1433 | ||
1434 | my $id = msgget(IPC_PRIVATE, IPC_CREAT | S_IRWXU); | |
1435 | ||
1436 | my $sent = "message"; | |
1437 | my $type = 1234; | |
1438 | my $rcvd; | |
1439 | my $type_rcvd; | |
1440 | ||
1441 | if (defined $id) { | |
1442 | if (msgsnd($id, pack("l! a*", $type_sent, $sent), 0)) { | |
1443 | if (msgrcv($id, $rcvd, 60, 0, 0)) { | |
1444 | ($type_rcvd, $rcvd) = unpack("l! a*", $rcvd); | |
1445 | if ($rcvd eq $sent) { | |
1446 | print "okay\n"; | |
1447 | } else { | |
1448 | print "not okay\n"; | |
1449 | } | |
1450 | } else { | |
1451 | die "# msgrcv failed\n"; | |
1452 | } | |
1453 | } else { | |
1454 | die "# msgsnd failed\n"; | |
1455 | } | |
1456 | msgctl($id, IPC_RMID, 0) || die "# msgctl failed: $!\n"; | |
1457 | } else { | |
1458 | die "# msgget failed\n"; | |
1459 | } | |
1460 | ||
4633a7c4 LW |
1461 | =head1 NOTES |
1462 | ||
5a964f20 TC |
1463 | Most of these routines quietly but politely return C<undef> when they |
1464 | fail instead of causing your program to die right then and there due to | |
1465 | an uncaught exception. (Actually, some of the new I<Socket> conversion | |
1466 | functions croak() on bad arguments.) It is therefore essential to | |
1467 | check return values from these functions. Always begin your socket | |
1468 | programs this way for optimal success, and don't forget to add B<-T> | |
1469 | taint checking flag to the #! line for servers: | |
4633a7c4 | 1470 | |
5a964f20 | 1471 | #!/usr/bin/perl -Tw |
4633a7c4 LW |
1472 | use strict; |
1473 | use sigtrap; | |
1474 | use Socket; | |
1475 | ||
1476 | =head1 BUGS | |
1477 | ||
1478 | All these routines create system-specific portability problems. As noted | |
1479 | elsewhere, Perl is at the mercy of your C libraries for much of its system | |
1480 | behaviour. It's probably safest to assume broken SysV semantics for | |
6a3992aa | 1481 | signals and to stick with simple TCP and UDP socket operations; e.g., don't |
a2eb9003 | 1482 | try to pass open file descriptors over a local UDP datagram socket if you |
4633a7c4 LW |
1483 | want your code to stand a chance of being portable. |
1484 | ||
5a964f20 TC |
1485 | As mentioned in the signals section, because few vendors provide C |
1486 | libraries that are safely re-entrant, the prudent programmer will do | |
1487 | little else within a handler beyond setting a numeric variable that | |
1488 | already exists; or, if locked into a slow (restarting) system call, | |
1489 | using die() to raise an exception and longjmp(3) out. In fact, even | |
1490 | these may in some cases cause a core dump. It's probably best to avoid | |
1491 | signals except where they are absolutely inevitable. This | |
1492 | will be addressed in a future release of Perl. | |
4633a7c4 LW |
1493 | |
1494 | =head1 AUTHOR | |
1495 | ||
1496 | Tom Christiansen, with occasional vestiges of Larry Wall's original | |
7b05b7e3 | 1497 | version and suggestions from the Perl Porters. |
4633a7c4 LW |
1498 | |
1499 | =head1 SEE ALSO | |
1500 | ||
7b05b7e3 TC |
1501 | There's a lot more to networking than this, but this should get you |
1502 | started. | |
1503 | ||
5a964f20 TC |
1504 | For intrepid programmers, the indispensable textbook is I<Unix Network |
1505 | Programming> by W. Richard Stevens (published by Addison-Wesley). Note | |
1506 | that most books on networking address networking from the perspective of | |
1507 | a C programmer; translation to Perl is left as an exercise for the reader. | |
7b05b7e3 TC |
1508 | |
1509 | The IO::Socket(3) manpage describes the object library, and the Socket(3) | |
1510 | manpage describes the low-level interface to sockets. Besides the obvious | |
1511 | functions in L<perlfunc>, you should also check out the F<modules> file | |
1512 | at your nearest CPAN site. (See L<perlmodlib> or best yet, the F<Perl | |
1513 | FAQ> for a description of what CPAN is and where to get it.) | |
1514 | ||
4633a7c4 | 1515 | Section 5 of the F<modules> file is devoted to "Networking, Device Control |
6a3992aa | 1516 | (modems), and Interprocess Communication", and contains numerous unbundled |
4633a7c4 LW |
1517 | modules numerous networking modules, Chat and Expect operations, CGI |
1518 | programming, DCE, FTP, IPC, NNTP, Proxy, Ptty, RPC, SNMP, SMTP, Telnet, | |
1519 | Threads, and ToolTalk--just to name a few. |