This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
various pod tweaks (from M.J.T. Guy <mjtg@cus.cam.ac.uk>)
[perl5.git] / pod / perlfaq8.pod
CommitLineData
68dc0745
PP
1=head1 NAME
2
d92eb7b0 3perlfaq8 - System Interaction ($Revision: 1.39 $, $Date: 1999/05/23 18:37:57 $)
68dc0745
PP
4
5=head1 DESCRIPTION
6
7This section of the Perl FAQ covers questions involving operating
8system interaction. This involves interprocess communication (IPC),
9control over the user-interface (keyboard, screen and pointing
10devices), and most anything else not related to data manipulation.
11
12Read the FAQs and documentation specific to the port of perl to your
46fc3d4c 13operating system (eg, L<perlvms>, L<perlplan9>, ...). These should
68dc0745
PP
14contain more detailed information on the vagaries of your perl.
15
16=head2 How do I find out which operating system I'm running under?
17
d92eb7b0
GS
18The $^O variable ($OSNAME if you use English) contains an indication of
19the name of the operating system (not its release number) that your perl
20binary was built for.
68dc0745
PP
21
22=head2 How come exec() doesn't return?
23
24Because that's what it does: it replaces your currently running
25program with a different one. If you want to keep going (as is
26probably the case if you're asking this question) use system()
27instead.
28
29=head2 How do I do fancy stuff with the keyboard/screen/mouse?
30
31How you access/control keyboards, screens, and pointing devices
32("mice") is system-dependent. Try the following modules:
33
34=over 4
35
36=item Keyboard
37
38 Term::Cap Standard perl distribution
39 Term::ReadKey CPAN
40 Term::ReadLine::Gnu CPAN
41 Term::ReadLine::Perl CPAN
42 Term::Screen CPAN
43
44=item Screen
45
46 Term::Cap Standard perl distribution
47 Curses CPAN
48 Term::ANSIColor CPAN
49
50=item Mouse
51
52 Tk CPAN
53
54=back
55
c8db1d39
TC
56Some of these specific cases are shown below.
57
58=head2 How do I print something out in color?
59
60In general, you don't, because you don't know whether
61the recipient has a color-aware display device. If you
62know that they have an ANSI terminal that understands
63color, you can use the Term::ANSIColor module from CPAN:
64
65 use Term::ANSIColor;
66 print color("red"), "Stop!\n", color("reset");
67 print color("green"), "Go!\n", color("reset");
68
69Or like this:
70
71 use Term::ANSIColor qw(:constants);
72 print RED, "Stop!\n", RESET;
73 print GREEN, "Go!\n", RESET;
74
75=head2 How do I read just one key without waiting for a return key?
76
77Controlling input buffering is a remarkably system-dependent matter.
d92eb7b0 78On many systems, you can just use the B<stty> command as shown in
c8db1d39
TC
79L<perlfunc/getc>, but as you see, that's already getting you into
80portability snags.
81
82 open(TTY, "+</dev/tty") or die "no tty: $!";
83 system "stty cbreak </dev/tty >/dev/tty 2>&1";
84 $key = getc(TTY); # perhaps this works
85 # OR ELSE
86 sysread(TTY, $key, 1); # probably this does
87 system "stty -cbreak </dev/tty >/dev/tty 2>&1";
88
89The Term::ReadKey module from CPAN offers an easy-to-use interface that
90should be more efficient than shelling out to B<stty> for each key.
91It even includes limited support for Windows.
92
93 use Term::ReadKey;
94 ReadMode('cbreak');
95 $key = ReadKey(0);
96 ReadMode('normal');
97
98However, that requires that you have a working C compiler and can use it
99to build and install a CPAN module. Here's a solution using
100the standard POSIX module, which is already on your systems (assuming
101your system supports POSIX).
102
103 use HotKey;
104 $key = readkey();
105
106And here's the HotKey module, which hides the somewhat mystifying calls
107to manipulate the POSIX termios structures.
108
109 # HotKey.pm
110 package HotKey;
111
112 @ISA = qw(Exporter);
113 @EXPORT = qw(cbreak cooked readkey);
114
115 use strict;
116 use POSIX qw(:termios_h);
117 my ($term, $oterm, $echo, $noecho, $fd_stdin);
118
119 $fd_stdin = fileno(STDIN);
120 $term = POSIX::Termios->new();
121 $term->getattr($fd_stdin);
122 $oterm = $term->getlflag();
123
124 $echo = ECHO | ECHOK | ICANON;
125 $noecho = $oterm & ~$echo;
126
127 sub cbreak {
128 $term->setlflag($noecho); # ok, so i don't want echo either
129 $term->setcc(VTIME, 1);
130 $term->setattr($fd_stdin, TCSANOW);
131 }
132
133 sub cooked {
134 $term->setlflag($oterm);
135 $term->setcc(VTIME, 0);
136 $term->setattr($fd_stdin, TCSANOW);
137 }
138
139 sub readkey {
140 my $key = '';
141 cbreak();
142 sysread(STDIN, $key, 1);
143 cooked();
144 return $key;
145 }
146
147 END { cooked() }
148
149 1;
150
151=head2 How do I check whether input is ready on the keyboard?
152
153The easiest way to do this is to read a key in nonblocking mode with the
154Term::ReadKey module from CPAN, passing it an argument of -1 to indicate
155not to block:
156
157 use Term::ReadKey;
158
159 ReadMode('cbreak');
160
161 if (defined ($char = ReadKey(-1)) ) {
162 # input was waiting and it was $char
163 } else {
164 # no input was waiting
165 }
166
167 ReadMode('normal'); # restore normal tty settings
168
169=head2 How do I clear the screen?
170
d92eb7b0 171If you only have do so infrequently, use C<system>:
c8db1d39
TC
172
173 system("clear");
174
175If you have to do this a lot, save the clear string
176so you can print it 100 times without calling a program
177100 times:
178
179 $clear_string = `clear`;
180 print $clear_string;
181
182If you're planning on doing other screen manipulations, like cursor
183positions, etc, you might wish to use Term::Cap module:
184
185 use Term::Cap;
186 $terminal = Term::Cap->Tgetent( {OSPEED => 9600} );
187 $clear_string = $terminal->Tputs('cl');
188
189=head2 How do I get the screen size?
190
191If you have Term::ReadKey module installed from CPAN,
192you can use it to fetch the width and height in characters
193and in pixels:
194
195 use Term::ReadKey;
196 ($wchar, $hchar, $wpixels, $hpixels) = GetTerminalSize();
197
198This is more portable than the raw C<ioctl>, but not as
199illustrative:
200
201 require 'sys/ioctl.ph';
202 die "no TIOCGWINSZ " unless defined &TIOCGWINSZ;
203 open(TTY, "+</dev/tty") or die "No tty: $!";
204 unless (ioctl(TTY, &TIOCGWINSZ, $winsize='')) {
205 die sprintf "$0: ioctl TIOCGWINSZ (%08x: $!)\n", &TIOCGWINSZ;
206 }
207 ($row, $col, $xpixel, $ypixel) = unpack('S4', $winsize);
208 print "(row,col) = ($row,$col)";
209 print " (xpixel,ypixel) = ($xpixel,$ypixel)" if $xpixel || $ypixel;
210 print "\n";
211
68dc0745
PP
212=head2 How do I ask the user for a password?
213
214(This question has nothing to do with the web. See a different
215FAQ for that.)
216
217There's an example of this in L<perlfunc/crypt>). First, you put
218the terminal into "no echo" mode, then just read the password
219normally. You may do this with an old-style ioctl() function, POSIX
220terminal control (see L<POSIX>, and Chapter 7 of the Camel), or a call
221to the B<stty> program, with varying degrees of portability.
222
223You can also do this for most systems using the Term::ReadKey module
224from CPAN, which is easier to use and in theory more portable.
225
c8db1d39
TC
226 use Term::ReadKey;
227
228 ReadMode('noecho');
229 $password = ReadLine(0);
230
68dc0745
PP
231=head2 How do I read and write the serial port?
232
233This depends on which operating system your program is running on. In
234the case of Unix, the serial ports will be accessible through files in
235/dev; on other systems, the devices names will doubtless differ.
236Several problem areas common to all device interaction are the
237following
238
239=over 4
240
241=item lockfiles
242
243Your system may use lockfiles to control multiple access. Make sure
244you follow the correct protocol. Unpredictable behaviour can result
245from multiple processes reading from one device.
246
247=item open mode
248
249If you expect to use both read and write operations on the device,
250you'll have to open it for update (see L<perlfunc/"open"> for
251details). You may wish to open it without running the risk of
252blocking by using sysopen() and C<O_RDWR|O_NDELAY|O_NOCTTY> from the
253Fcntl module (part of the standard perl distribution). See
254L<perlfunc/"sysopen"> for more on this approach.
255
256=item end of line
257
258Some devices will be expecting a "\r" at the end of each line rather
259than a "\n". In some ports of perl, "\r" and "\n" are different from
260their usual (Unix) ASCII values of "\012" and "\015". You may have to
261give the numeric values you want directly, using octal ("\015"), hex
262("0x0D"), or as a control-character specification ("\cM").
263
264 print DEV "atv1\012"; # wrong, for some devices
265 print DEV "atv1\015"; # right, for some devices
266
267Even though with normal text files, a "\n" will do the trick, there is
268still no unified scheme for terminating a line that is portable
46fc3d4c 269between Unix, DOS/Win, and Macintosh, except to terminate I<ALL> line
68dc0745
PP
270ends with "\015\012", and strip what you don't need from the output.
271This applies especially to socket I/O and autoflushing, discussed
272next.
273
274=item flushing output
275
276If you expect characters to get to your device when you print() them,
c8db1d39
TC
277you'll want to autoflush that filehandle. You can use select()
278and the C<$|> variable to control autoflushing (see L<perlvar/$|>
279and L<perlfunc/select>):
68dc0745
PP
280
281 $oldh = select(DEV);
282 $| = 1;
283 select($oldh);
284
285You'll also see code that does this without a temporary variable, as in
286
287 select((select(DEV), $| = 1)[0]);
288
c8db1d39
TC
289Or if you don't mind pulling in a few thousand lines
290of code just because you're afraid of a little $| variable:
291
292 use IO::Handle;
293 DEV->autoflush(1);
294
68dc0745
PP
295As mentioned in the previous item, this still doesn't work when using
296socket I/O between Unix and Macintosh. You'll need to hardcode your
297line terminators, in that case.
298
299=item non-blocking input
300
301If you are doing a blocking read() or sysread(), you'll have to
302arrange for an alarm handler to provide a timeout (see
303L<perlfunc/alarm>). If you have a non-blocking open, you'll likely
304have a non-blocking read, which means you may have to use a 4-arg
305select() to determine whether I/O is ready on that device (see
306L<perlfunc/"select">.
307
308=back
309
c8db1d39
TC
310While trying to read from his caller-id box, the notorious Jamie Zawinski
311<jwz@netscape.com>, after much gnashing of teeth and fighting with sysread,
312sysopen, POSIX's tcgetattr business, and various other functions that
313go bump in the night, finally came up with this:
314
315 sub open_modem {
316 use IPC::Open2;
317 my $stty = `/bin/stty -g`;
318 open2( \*MODEM_IN, \*MODEM_OUT, "cu -l$modem_device -s2400 2>&1");
319 # starting cu hoses /dev/tty's stty settings, even when it has
320 # been opened on a pipe...
321 system("/bin/stty $stty");
322 $_ = <MODEM_IN>;
323 chop;
324 if ( !m/^Connected/ ) {
325 print STDERR "$0: cu printed `$_' instead of `Connected'\n";
326 }
327 }
328
68dc0745
PP
329=head2 How do I decode encrypted password files?
330
331You spend lots and lots of money on dedicated hardware, but this is
332bound to get you talked about.
333
334Seriously, you can't if they are Unix password files - the Unix
c8db1d39
TC
335password system employs one-way encryption. It's more like hashing than
336encryption. The best you can check is whether something else hashes to
337the same string. You can't turn a hash back into the original string.
338Programs like Crack
339can forcibly (and intelligently) try to guess passwords, but don't
340(can't) guarantee quick success.
68dc0745
PP
341
342If you're worried about users selecting bad passwords, you should
343proactively check when they try to change their password (by modifying
344passwd(1), for example).
345
346=head2 How do I start a process in the background?
347
348You could use
349
350 system("cmd &")
351
352or you could use fork as documented in L<perlfunc/"fork">, with
353further examples in L<perlipc>. Some things to be aware of, if you're
354on a Unix-like system:
355
356=over 4
357
c8db1d39 358=item STDIN, STDOUT, and STDERR are shared
68dc0745
PP
359
360Both the main process and the backgrounded one (the "child" process)
361share the same STDIN, STDOUT and STDERR filehandles. If both try to
362access them at once, strange things can happen. You may want to close
363or reopen these for the child. You can get around this with
364C<open>ing a pipe (see L<perlfunc/"open">) but on some systems this
365means that the child process cannot outlive the parent.
366
367=item Signals
368
369You'll have to catch the SIGCHLD signal, and possibly SIGPIPE too.
370SIGCHLD is sent when the backgrounded process finishes. SIGPIPE is
371sent when you write to a filehandle whose child process has closed (an
372untrapped SIGPIPE can cause your program to silently die). This is
373not an issue with C<system("cmd&")>.
374
375=item Zombies
376
377You have to be prepared to "reap" the child process when it finishes
378
379 $SIG{CHLD} = sub { wait };
380
381See L<perlipc/"Signals"> for other examples of code to do this.
382Zombies are not an issue with C<system("prog &")>.
383
384=back
385
386=head2 How do I trap control characters/signals?
387
c8db1d39
TC
388You don't actually "trap" a control character. Instead, that character
389generates a signal which is sent to your terminal's currently
390foregrounded process group, which you then trap in your process.
391Signals are documented in L<perlipc/"Signals"> and chapter 6 of the Camel.
68dc0745 392
46fc3d4c 393Be warned that very few C libraries are re-entrant. Therefore, if you
68dc0745
PP
394attempt to print() in a handler that got invoked during another stdio
395operation your internal structures will likely be in an
396inconsistent state, and your program will dump core. You can
397sometimes avoid this by using syswrite() instead of print().
398
399Unless you're exceedingly careful, the only safe things to do inside a
400signal handler are: set a variable and exit. And in the first case,
401you should only set a variable in such a way that malloc() is not
402called (eg, by setting a variable that already has a value).
403
404For example:
405
406 $Interrupted = 0; # to ensure it has a value
407 $SIG{INT} = sub {
408 $Interrupted++;
409 syswrite(STDERR, "ouch\n", 5);
410 }
411
412However, because syscalls restart by default, you'll find that if
413you're in a "slow" call, such as E<lt>FHE<gt>, read(), connect(), or
414wait(), that the only way to terminate them is by "longjumping" out;
46fc3d4c 415that is, by raising an exception. See the time-out handler for a
68dc0745
PP
416blocking flock() in L<perlipc/"Signals"> or chapter 6 of the Camel.
417
418=head2 How do I modify the shadow password file on a Unix system?
419
c8db1d39
TC
420If perl was installed correctly, and your shadow library was written
421properly, the getpw*() functions described in L<perlfunc> should in
422theory provide (read-only) access to entries in the shadow password
423file. To change the file, make a new shadow password file (the format
424varies from system to system - see L<passwd(5)> for specifics) and use
d92eb7b0 425pwd_mkdb(8) to install it (see L<pwd_mkdb(8)> for more details).
68dc0745
PP
426
427=head2 How do I set the time and date?
428
429Assuming you're running under sufficient permissions, you should be
430able to set the system-wide date and time by running the date(1)
431program. (There is no way to set the time and date on a per-process
432basis.) This mechanism will work for Unix, MS-DOS, Windows, and NT;
433the VMS equivalent is C<set time>.
434
435However, if all you want to do is change your timezone, you can
436probably get away with setting an environment variable:
437
438 $ENV{TZ} = "MST7MDT"; # unixish
439 $ENV{'SYS$TIMEZONE_DIFFERENTIAL'}="-5" # vms
c8db1d39 440 system "trn comp.lang.perl.misc";
68dc0745
PP
441
442=head2 How can I sleep() or alarm() for under a second?
443
444If you want finer granularity than the 1 second that the sleep()
445function provides, the easiest way is to use the select() function as
446documented in L<perlfunc/"select">. If your system has itimers and
447syscall() support, you can check out the old example in
448http://www.perl.com/CPAN/doc/misc/ancient/tutorial/eg/itimers.pl .
449
450=head2 How can I measure time under a second?
451
65acb1b1
TC
452In general, you may not be able to. The Time::HiRes module (available
453from CPAN) provides this functionality for some systems.
68dc0745 454
65acb1b1
TC
455If your system supports both the syscall() function in Perl as well as
456a system call like gettimeofday(2), then you may be able to do
457something like this:
68dc0745
PP
458
459 require 'sys/syscall.ph';
460
461 $TIMEVAL_T = "LL";
462
463 $done = $start = pack($TIMEVAL_T, ());
464
d92eb7b0 465 syscall(&SYS_gettimeofday, $start, 0) != -1
68dc0745
PP
466 or die "gettimeofday: $!";
467
468 ##########################
469 # DO YOUR OPERATION HERE #
470 ##########################
471
472 syscall( &SYS_gettimeofday, $done, 0) != -1
473 or die "gettimeofday: $!";
474
475 @start = unpack($TIMEVAL_T, $start);
476 @done = unpack($TIMEVAL_T, $done);
477
478 # fix microseconds
479 for ($done[1], $start[1]) { $_ /= 1_000_000 }
480
481 $delta_time = sprintf "%.4f", ($done[0] + $done[1] )
482 -
483 ($start[0] + $start[1] );
484
485=head2 How can I do an atexit() or setjmp()/longjmp()? (Exception handling)
486
487Release 5 of Perl added the END block, which can be used to simulate
488atexit(). Each package's END block is called when the program or
c8db1d39
TC
489thread ends (see L<perlmod> manpage for more details).
490
491For example, you can use this to make sure your filter program
492managed to finish its output without filling up the disk:
493
494 END {
495 close(STDOUT) || die "stdout close failed: $!";
496 }
497
498The END block isn't called when untrapped signals kill the program, though, so if
499you use END blocks you should also use
68dc0745
PP
500
501 use sigtrap qw(die normal-signals);
502
503Perl's exception-handling mechanism is its eval() operator. You can
504use eval() as setjmp and die() as longjmp. For details of this, see
46fc3d4c 505the section on signals, especially the time-out handler for a blocking
68dc0745
PP
506flock() in L<perlipc/"Signals"> and chapter 6 of the Camel.
507
508If exception handling is all you're interested in, try the
509exceptions.pl library (part of the standard perl distribution).
510
511If you want the atexit() syntax (and an rmexit() as well), try the
512AtExit module available from CPAN.
513
514=head2 Why doesn't my sockets program work under System V (Solaris)? What does the error message "Protocol not supported" mean?
515
516Some Sys-V based systems, notably Solaris 2.X, redefined some of the
517standard socket constants. Since these were constant across all
518architectures, they were often hardwired into perl code. The proper
519way to deal with this is to "use Socket" to get the correct values.
520
521Note that even though SunOS and Solaris are binary compatible, these
522values are different. Go figure.
523
524=head2 How can I call my system's unique C functions from Perl?
525
526In most cases, you write an external module to do it - see the answer
527to "Where can I learn about linking C with Perl? [h2xs, xsubpp]".
528However, if the function is a system call, and your system supports
529syscall(), you can use the syscall function (documented in
530L<perlfunc>).
531
532Remember to check the modules that came with your distribution, and
533CPAN as well - someone may already have written a module to do it.
534
535=head2 Where do I get the include files to do ioctl() or syscall()?
536
537Historically, these would be generated by the h2ph tool, part of the
538standard perl distribution. This program converts cpp(1) directives
539in C header files to files containing subroutine definitions, like
540&SYS_getitimer, which you can use as arguments to your functions.
541It doesn't work perfectly, but it usually gets most of the job done.
542Simple files like F<errno.h>, F<syscall.h>, and F<socket.h> were fine,
543but the hard ones like F<ioctl.h> nearly always need to hand-edited.
544Here's how to install the *.ph files:
545
46fc3d4c 546 1. become super-user
68dc0745
PP
547 2. cd /usr/include
548 3. h2ph *.h */*.h
549
550If your system supports dynamic loading, for reasons of portability and
551sanity you probably ought to use h2xs (also part of the standard perl
552distribution). This tool converts C header files to Perl extensions.
553See L<perlxstut> for how to get started with h2xs.
554
555If your system doesn't support dynamic loading, you still probably
556ought to use h2xs. See L<perlxstut> and L<ExtUtils::MakeMaker> for
557more information (in brief, just use B<make perl> instead of a plain
558B<make> to rebuild perl with a new static extension).
559
560=head2 Why do setuid perl scripts complain about kernel problems?
561
562Some operating systems have bugs in the kernel that make setuid
563scripts inherently insecure. Perl gives you a number of options
564(described in L<perlsec>) to work around such systems.
565
566=head2 How can I open a pipe both to and from a command?
567
568The IPC::Open2 module (part of the standard perl distribution) is an
c8db1d39
TC
569easy-to-use approach that internally uses pipe(), fork(), and exec() to do
570the job. Make sure you read the deadlock warnings in its documentation,
571though (see L<IPC::Open2>). See L<perlipc/"Bidirectional Communication
572with Another Process"> and L<perlipc/"Bidirectional Communication with
573Yourself">
574
575You may also use the IPC::Open3 module (part of the standard perl
576distribution), but be warned that it has a different order of
577arguments from IPC::Open2 (see L<IPC::Open3>).
68dc0745 578
3fe9a6f1
PP
579=head2 Why can't I get the output of a command with system()?
580
46fc3d4c
PP
581You're confusing the purpose of system() and backticks (``). system()
582runs a command and returns exit status information (as a 16 bit value:
c8db1d39 583the low 7 bits are the signal the process died from, if any, and
46fc3d4c 584the high 8 bits are the actual exit value). Backticks (``) run a
3fe9a6f1
PP
585command and return what it sent to STDOUT.
586
46fc3d4c
PP
587 $exit_status = system("mail-users");
588 $output_string = `ls`;
3fe9a6f1 589
68dc0745
PP
590=head2 How can I capture STDERR from an external command?
591
592There are three basic ways of running external commands:
593
594 system $cmd; # using system()
595 $output = `$cmd`; # using backticks (``)
596 open (PIPE, "cmd |"); # using open()
597
598With system(), both STDOUT and STDERR will go the same place as the
599script's versions of these, unless the command redirects them.
600Backticks and open() read B<only> the STDOUT of your command.
601
602With any of these, you can change file descriptors before the call:
603
604 open(STDOUT, ">logfile");
605 system("ls");
606
607or you can use Bourne shell file-descriptor redirection:
608
609 $output = `$cmd 2>some_file`;
610 open (PIPE, "cmd 2>some_file |");
611
612You can also use file-descriptor redirection to make STDERR a
613duplicate of STDOUT:
614
615 $output = `$cmd 2>&1`;
616 open (PIPE, "cmd 2>&1 |");
617
618Note that you I<cannot> simply open STDERR to be a dup of STDOUT
619in your Perl program and avoid calling the shell to do the redirection.
620This doesn't work:
621
622 open(STDERR, ">&STDOUT");
623 $alloutput = `cmd args`; # stderr still escapes
624
625This fails because the open() makes STDERR go to where STDOUT was
626going at the time of the open(). The backticks then make STDOUT go to
627a string, but don't change STDERR (which still goes to the old
628STDOUT).
629
630Note that you I<must> use Bourne shell (sh(1)) redirection syntax in
631backticks, not csh(1)! Details on why Perl's system() and backtick
632and pipe opens all use the Bourne shell are in
633http://www.perl.com/CPAN/doc/FMTEYEWTK/versus/csh.whynot .
c8db1d39 634To capture a command's STDERR and STDOUT together:
68dc0745 635
c8db1d39
TC
636 $output = `cmd 2>&1`; # either with backticks
637 $pid = open(PH, "cmd 2>&1 |"); # or with an open pipe
638 while (<PH>) { } # plus a read
639
640To capture a command's STDOUT but discard its STDERR:
641
642 $output = `cmd 2>/dev/null`; # either with backticks
643 $pid = open(PH, "cmd 2>/dev/null |"); # or with an open pipe
644 while (<PH>) { } # plus a read
645
646To capture a command's STDERR but discard its STDOUT:
647
648 $output = `cmd 2>&1 1>/dev/null`; # either with backticks
649 $pid = open(PH, "cmd 2>&1 1>/dev/null |"); # or with an open pipe
650 while (<PH>) { } # plus a read
651
652To exchange a command's STDOUT and STDERR in order to capture the STDERR
653but leave its STDOUT to come out our old STDERR:
654
655 $output = `cmd 3>&1 1>&2 2>&3 3>&-`; # either with backticks
656 $pid = open(PH, "cmd 3>&1 1>&2 2>&3 3>&-|");# or with an open pipe
657 while (<PH>) { } # plus a read
658
659To read both a command's STDOUT and its STDERR separately, it's easiest
660and safest to redirect them separately to files, and then read from those
661files when the program is done:
662
663 system("program args 1>/tmp/program.stdout 2>/tmp/program.stderr");
664
665Ordering is important in all these examples. That's because the shell
666processes file descriptor redirections in strictly left to right order.
667
668 system("prog args 1>tmpfile 2>&1");
669 system("prog args 2>&1 1>tmpfile");
670
671The first command sends both standard out and standard error to the
672temporary file. The second command sends only the old standard output
673there, and the old standard error shows up on the old standard out.
68dc0745
PP
674
675=head2 Why doesn't open() return an error when a pipe open fails?
676
65acb1b1
TC
677Because the pipe open takes place in two steps: first Perl calls
678fork() to start a new process, then this new process calls exec() to
679run the program you really wanted to open. The first step reports
680success or failure to your process, so open() can only tell you
681whether the fork() succeeded or not.
682
683To find out if the exec() step succeeded, you have to catch SIGCHLD
684and wait() to get the exit status. You should also catch SIGPIPE if
685you're writing to the child--you may not have found out the exec()
3fe9a6f1 686failed by the time you write. This is documented in L<perlipc>.
68dc0745 687
65acb1b1
TC
688In some cases, even this won't work. If the second argument to a
689piped open() contains shell metacharacters, perl fork()s, then exec()s
690a shell to decode the metacharacters and eventually run the desired
691program. Now when you call wait(), you only learn whether or not the
692I<shell> could be successfully started. Best to avoid shell
693metacharacters.
694
68dc0745 695On systems that follow the spawn() paradigm, open() I<might> do what
65acb1b1 696you expect--unless perl uses a shell to start your command. In this
68dc0745
PP
697case the fork()/exec() description still applies.
698
699=head2 What's wrong with using backticks in a void context?
700
701Strictly speaking, nothing. Stylistically speaking, it's not a good
702way to write maintainable code because backticks have a (potentially
d92eb7b0 703humongous) return value, and you're ignoring it. It's may also not be very
68dc0745
PP
704efficient, because you have to read in all the lines of output, allocate
705memory for them, and then throw it away. Too often people are lulled
706to writing:
707
708 `cp file file.bak`;
709
710And now they think "Hey, I'll just always use backticks to run programs."
711Bad idea: backticks are for capturing a program's output; the system()
712function is for running programs.
713
714Consider this line:
715
716 `cat /etc/termcap`;
717
718You haven't assigned the output anywhere, so it just wastes memory
719(for a little while). Plus you forgot to check C<$?> to see whether
720the program even ran correctly. Even if you wrote
721
722 print `cat /etc/termcap`;
723
724In most cases, this could and probably should be written as
725
726 system("cat /etc/termcap") == 0
727 or die "cat program failed!";
728
d92eb7b0 729Which will get the output quickly (as it is generated, instead of only
c8db1d39 730at the end) and also check the return value.
68dc0745
PP
731
732system() also provides direct control over whether shell wildcard
733processing may take place, whereas backticks do not.
734
735=head2 How can I call backticks without shell processing?
736
737This is a bit tricky. Instead of writing
738
739 @ok = `grep @opts '$search_string' @filenames`;
740
741You have to do this:
742
743 my @ok = ();
744 if (open(GREP, "-|")) {
745 while (<GREP>) {
746 chomp;
747 push(@ok, $_);
748 }
749 close GREP;
750 } else {
751 exec 'grep', @opts, $search_string, @filenames;
752 }
753
754Just as with system(), no shell escapes happen when you exec() a list.
d92eb7b0 755Further examples of this can be found in L<perlipc/"Safe Pipe Opens">.
68dc0745 756
d92eb7b0
GS
757Note that if you're stuck on Microsoft, no solution to this vexing issue
758is even possible. Even if Perl were to emulate fork(), you'd still
759be hosed, because Microsoft gives no argc/argv-style API. Their API
760always reparses from a single string, which is fundamentally wrong,
761but you're not likely to get the Gods of Redmond to acknowledge this
762and fix it for you.
c8db1d39 763
54310121 764=head2 Why can't my script read from STDIN after I gave it EOF (^D on Unix, ^Z on MS-DOS)?
68dc0745
PP
765
766Because some stdio's set error and eof flags that need clearing. The
767POSIX module defines clearerr() that you can use. That is the
768technically correct way to do it. Here are some less reliable
769workarounds:
770
771=over 4
772
773=item 1
774
775Try keeping around the seekpointer and go there, like this:
776
777 $where = tell(LOG);
778 seek(LOG, $where, 0);
779
780=item 2
781
782If that doesn't work, try seeking to a different part of the file and
783then back.
784
785=item 3
786
787If that doesn't work, try seeking to a different part of
788the file, reading something, and then seeking back.
789
790=item 4
791
792If that doesn't work, give up on your stdio package and use sysread.
793
794=back
795
796=head2 How can I convert my shell script to perl?
797
798Learn Perl and rewrite it. Seriously, there's no simple converter.
799Things that are awkward to do in the shell are easy to do in Perl, and
800this very awkwardness is what would make a shell->perl converter
801nigh-on impossible to write. By rewriting it, you'll think about what
802you're really trying to do, and hopefully will escape the shell's
46fc3d4c 803pipeline datastream paradigm, which while convenient for some matters,
68dc0745
PP
804causes many inefficiencies.
805
806=head2 Can I use perl to run a telnet or ftp session?
807
46fc3d4c
PP
808Try the Net::FTP, TCP::Client, and Net::Telnet modules (available from
809CPAN). http://www.perl.com/CPAN/scripts/netstuff/telnet.emul.shar
810will also help for emulating the telnet protocol, but Net::Telnet is
811quite probably easier to use..
812
813If all you want to do is pretend to be telnet but don't need
814the initial telnet handshaking, then the standard dual-process
815approach will suffice:
816
817 use IO::Socket; # new in 5.004
818 $handle = IO::Socket::INET->new('www.perl.com:80')
819 || die "can't connect to port 80 on www.perl.com: $!";
820 $handle->autoflush(1);
821 if (fork()) { # XXX: undef means failure
822 select($handle);
823 print while <STDIN>; # everything from stdin to socket
824 } else {
825 print while <$handle>; # everything from socket to stdout
826 }
827 close $handle;
828 exit;
68dc0745
PP
829
830=head2 How can I write expect in Perl?
831
832Once upon a time, there was a library called chat2.pl (part of the
c8db1d39
TC
833standard perl distribution), which never really got finished. If you
834find it somewhere, I<don't use it>. These days, your best bet is to
835look at the Expect module available from CPAN, which also requires two
836other modules from CPAN, IO::Pty and IO::Stty.
68dc0745
PP
837
838=head2 Is there a way to hide perl's command line from programs such as "ps"?
839
840First of all note that if you're doing this for security reasons (to
841avoid people seeing passwords, for example) then you should rewrite
842your program so that critical information is never given as an
843argument. Hiding the arguments won't make your program completely
844secure.
845
846To actually alter the visible command line, you can assign to the
847variable $0 as documented in L<perlvar>. This won't work on all
848operating systems, though. Daemon programs like sendmail place their
849state there, as in:
850
851 $0 = "orcus [accepting connections]";
852
853=head2 I {changed directory, modified my environment} in a perl script. How come the change disappeared when I exited the script? How do I get my changes to be visible?
854
855=over 4
856
857=item Unix
858
859In the strictest sense, it can't be done -- the script executes as a
860different process from the shell it was started from. Changes to a
861process are not reflected in its parent, only in its own children
862created after the change. There is shell magic that may allow you to
863fake it by eval()ing the script's output in your shell; check out the
92c2ed05 864comp.unix.questions FAQ for details.
68dc0745 865
68dc0745
PP
866=back
867
868=head2 How do I close a process's filehandle without waiting for it to complete?
869
870Assuming your system supports such things, just send an appropriate signal
871to the process (see L<perlfunc/"kill">. It's common to first send a TERM
872signal, wait a little bit, and then send a KILL signal to finish it off.
873
874=head2 How do I fork a daemon process?
875
876If by daemon process you mean one that's detached (disassociated from
877its tty), then the following process is reported to work on most
878Unixish systems. Non-Unix users should check their Your_OS::Process
879module for other solutions.
880
881=over 4
882
883=item *
884
b5a41e52 885Open /dev/tty and use the TIOCNOTTY ioctl on it. See L<tty(4)>
c8db1d39
TC
886for details. Or better yet, you can just use the POSIX::setsid()
887function, so you don't have to worry about process groups.
68dc0745
PP
888
889=item *
890
891Change directory to /
892
893=item *
894
895Reopen STDIN, STDOUT, and STDERR so they're not connected to the old
896tty.
897
898=item *
899
900Background yourself like this:
901
902 fork && exit;
903
904=back
905
1a91aff4
GS
906The Proc::Daemon module, available from CPAN, provides a function to
907perform these actions for you.
908
68dc0745
PP
909=head2 How do I make my program run with sh and csh?
910
911See the F<eg/nih> script (part of the perl source distribution).
912
68dc0745
PP
913=head2 How do I find out if I'm running interactively or not?
914
915Good question. Sometimes C<-t STDIN> and C<-t STDOUT> can give clues,
916sometimes not.
917
918 if (-t STDIN && -t STDOUT) {
919 print "Now what? ";
920 }
921
922On POSIX systems, you can test whether your own process group matches
923the current process group of your controlling terminal as follows:
924
925 use POSIX qw/getpgrp tcgetpgrp/;
926 open(TTY, "/dev/tty") or die $!;
65acb1b1 927 $tpgrp = tcgetpgrp(fileno(*TTY));
68dc0745
PP
928 $pgrp = getpgrp();
929 if ($tpgrp == $pgrp) {
930 print "foreground\n";
931 } else {
932 print "background\n";
933 }
934
935=head2 How do I timeout a slow event?
936
937Use the alarm() function, probably in conjunction with a signal
b350dd2f 938handler, as documented in L<perlipc/"Signals"> and chapter 6 of the
68dc0745
PP
939Camel. You may instead use the more flexible Sys::AlarmCall module
940available from CPAN.
941
942=head2 How do I set CPU limits?
943
944Use the BSD::Resource module from CPAN.
945
946=head2 How do I avoid zombies on a Unix system?
947
948Use the reaper code from L<perlipc/"Signals"> to call wait() when a
949SIGCHLD is received, or else use the double-fork technique described
950in L<perlfunc/fork>.
951
952=head2 How do I use an SQL database?
953
954There are a number of excellent interfaces to SQL databases. See the
955DBD::* modules available from
956http://www.perl.com/CPAN/modules/dbperl/DBD .
c8db1d39
TC
957A lot of information on this can be found at
958http://www.hermetica.com/technologia/perl/DBI/index.html .
68dc0745
PP
959
960=head2 How do I make a system() exit on control-C?
961
962You can't. You need to imitate the system() call (see L<perlipc> for
963sample code) and then have a signal handler for the INT signal that
c8db1d39
TC
964passes the signal on to the subprocess. Or you can check for it:
965
966 $rc = system($cmd);
967 if ($rc & 127) { die "signal death" }
68dc0745
PP
968
969=head2 How do I open a file without blocking?
970
971If you're lucky enough to be using a system that supports
972non-blocking reads (most Unixish systems do), you need only to use the
973O_NDELAY or O_NONBLOCK flag from the Fcntl module in conjunction with
974sysopen():
975
976 use Fcntl;
977 sysopen(FH, "/tmp/somefile", O_WRONLY|O_NDELAY|O_CREAT, 0644)
978 or die "can't open /tmp/somefile: $!":
979
d92eb7b0 980
68dc0745 981
d92eb7b0
GS
982
983=head2 How do I install a module from CPAN?
984
985The easiest way is to have a module also named CPAN do it for you.
986This module comes with perl version 5.004 and later. To manually install
987the CPAN module, or any well-behaved CPAN module for that matter, follow
988these steps:
68dc0745
PP
989
990=over 4
991
992=item 1
993
994Unpack the source into a temporary area.
995
996=item 2
997
998 perl Makefile.PL
999
1000=item 3
1001
1002 make
1003
1004=item 4
1005
1006 make test
1007
1008=item 5
1009
1010 make install
1011
1012=back
1013
1014If your version of perl is compiled without dynamic loading, then you
1015just need to replace step 3 (B<make>) with B<make perl> and you will
1016get a new F<perl> binary with your extension linked in.
1017
c8db1d39
TC
1018See L<ExtUtils::MakeMaker> for more details on building extensions.
1019See also the next question.
1020
1021=head2 What's the difference between require and use?
1022
1023Perl offers several different ways to include code from one file into
1024another. Here are the deltas between the various inclusion constructs:
1025
1026 1) do $file is like eval `cat $file`, except the former:
5e3006a4 1027 1.1: searches @INC and updates %INC.
c8db1d39
TC
1028 1.2: bequeaths an *unrelated* lexical scope on the eval'ed code.
1029
1030 2) require $file is like do $file, except the former:
1031 2.1: checks for redundant loading, skipping already loaded files.
1032 2.2: raises an exception on failure to find, compile, or execute $file.
1033
1034 3) require Module is like require "Module.pm", except the former:
1035 3.1: translates each "::" into your system's directory separator.
1036 3.2: primes the parser to disambiguate class Module as an indirect object.
1037
1038 4) use Module is like require Module, except the former:
1039 4.1: loads the module at compile time, not run-time.
1040 4.2: imports symbols and semantics from that package to the current one.
1041
1042In general, you usually want C<use> and a proper Perl module.
46fc3d4c
PP
1043
1044=head2 How do I keep my own module/library directory?
1045
1046When you build modules, use the PREFIX option when generating
1047Makefiles:
1048
1049 perl Makefile.PL PREFIX=/u/mydir/perl
1050
1051then either set the PERL5LIB environment variable before you run
1052scripts that use the modules/libraries (see L<perlrun>) or say
1053
1054 use lib '/u/mydir/perl';
1055
65acb1b1
TC
1056This is almost the same as:
1057
1058 BEGIN {
1059 unshift(@INC, '/u/mydir/perl');
1060 }
1061
1062except that the lib module checks for machine-dependent subdirectories.
46fc3d4c
PP
1063See Perl's L<lib> for more information.
1064
1065=head2 How do I add the directory my program lives in to the module/library search path?
1066
1067 use FindBin;
7b8d334a 1068 use lib "$FindBin::Bin";
46fc3d4c
PP
1069 use your_own_modules;
1070
1071=head2 How do I add a directory to my include path at runtime?
1072
1073Here are the suggested ways of modifying your include path:
1074
1075 the PERLLIB environment variable
1076 the PERL5LIB environment variable
c2611fb3 1077 the perl -Idir command line flag
46fc3d4c
PP
1078 the use lib pragma, as in
1079 use lib "$ENV{HOME}/myown_perllib";
1080
1081The latter is particularly useful because it knows about machine
1082dependent architectures. The lib.pm pragmatic module was first
1083included with the 5.002 release of Perl.
68dc0745 1084
65acb1b1
TC
1085=head2 What is socket.ph and where do I get it?
1086
1087It's a perl4-style file defining values for system networking
1088constants. Sometimes it is built using h2ph when Perl is installed,
1089but other times it is not. Modern programs C<use Socket;> instead.
1090
fc36a67e
PP
1091=head1 AUTHOR AND COPYRIGHT
1092
65acb1b1 1093Copyright (c) 1997-1999 Tom Christiansen and Nathan Torkington.
5a964f20
TC
1094All rights reserved.
1095
1096When included as part of the Standard Version of Perl, or as part of
1097its complete documentation whether printed or otherwise, this work
d92eb7b0 1098may be distributed only under the terms of Perl's Artistic License.
5a964f20
TC
1099Any distribution of this file or derivatives thereof I<outside>
1100of that package require that special arrangements be made with
1101copyright holder.
1102
1103Irrespective of its distribution, all code examples in this file
1104are hereby placed into the public domain. You are permitted and
1105encouraged to use this code in your own programs for fun
1106or for profit as you see fit. A simple comment in the code giving
1107credit would be courteous but is not required.