This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
ASCIZ is not a typo
[perl5.git] / pod / perlfunc.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
3perlfunc - Perl builtin functions
4
5=head1 DESCRIPTION
6
7The functions in this section can serve as terms in an expression.
8They fall into two major categories: list operators and named unary
9operators. These differ in their precedence relationship with a
10following comma. (See the precedence table in L<perlop>.) List
11operators take more than one argument, while unary operators can never
12take more than one argument. Thus, a comma terminates the argument of
13a unary operator, but merely separates the arguments of a list
14operator. A unary operator generally provides a scalar context to its
2b5ab1e7 15argument, while a list operator may provide either scalar or list
a0d0e21e 16contexts for its arguments. If it does both, the scalar arguments will
5f05dabc 17be first, and the list argument will follow. (Note that there can ever
0f31cffe 18be only one such list argument.) For instance, splice() has three scalar
2b5ab1e7
TC
19arguments followed by a list, whereas gethostbyname() has four scalar
20arguments.
a0d0e21e
LW
21
22In the syntax descriptions that follow, list operators that expect a
23list (and provide list context for the elements of the list) are shown
24with LIST as an argument. Such a list may consist of any combination
25of scalar arguments or list values; the list values will be included
26in the list as if each individual element were interpolated at that
27point in the list, forming a longer single-dimensional list value.
cf264981 28Commas should separate elements of the LIST.
a0d0e21e
LW
29
30Any function in the list below may be used either with or without
31parentheses around its arguments. (The syntax descriptions omit the
5f05dabc 32parentheses.) If you use the parentheses, the simple (but occasionally
19799a22 33surprising) rule is this: It I<looks> like a function, therefore it I<is> a
a0d0e21e
LW
34function, and precedence doesn't matter. Otherwise it's a list
35operator or unary operator, and precedence does matter. And whitespace
36between the function and left parenthesis doesn't count--so you need to
37be careful sometimes:
38
68dc0745
PP
39 print 1+2+4; # Prints 7.
40 print(1+2) + 4; # Prints 3.
41 print (1+2)+4; # Also prints 3!
42 print +(1+2)+4; # Prints 7.
43 print ((1+2)+4); # Prints 7.
a0d0e21e
LW
44
45If you run Perl with the B<-w> switch it can warn you about this. For
46example, the third line above produces:
47
48 print (...) interpreted as function at - line 1.
49 Useless use of integer addition in void context at - line 1.
50
2b5ab1e7
TC
51A few functions take no arguments at all, and therefore work as neither
52unary nor list operators. These include such functions as C<time>
53and C<endpwent>. For example, C<time+86_400> always means
54C<time() + 86_400>.
55
a0d0e21e 56For functions that can be used in either a scalar or list context,
54310121 57nonabortive failure is generally indicated in a scalar context by
a0d0e21e
LW
58returning the undefined value, and in a list context by returning the
59null list.
60
5a964f20
TC
61Remember the following important rule: There is B<no rule> that relates
62the behavior of an expression in list context to its behavior in scalar
63context, or vice versa. It might do two totally different things.
a0d0e21e 64Each operator and function decides which sort of value it would be most
2b5ab1e7 65appropriate to return in scalar context. Some operators return the
5a964f20 66length of the list that would have been returned in list context. Some
a0d0e21e
LW
67operators return the first value in the list. Some operators return the
68last value in the list. Some operators return a count of successful
69operations. In general, they do what you want, unless you want
70consistency.
71
d1be9408 72A named array in scalar context is quite different from what would at
5a964f20
TC
73first glance appear to be a list in scalar context. You can't get a list
74like C<(1,2,3)> into being in scalar context, because the compiler knows
75the context at compile time. It would generate the scalar comma operator
76there, not the list construction version of the comma. That means it
77was never a list to start with.
78
79In general, functions in Perl that serve as wrappers for system calls
f86cebdf 80of the same name (like chown(2), fork(2), closedir(2), etc.) all return
5a964f20
TC
81true when they succeed and C<undef> otherwise, as is usually mentioned
82in the descriptions below. This is different from the C interfaces,
19799a22
GS
83which return C<-1> on failure. Exceptions to this rule are C<wait>,
84C<waitpid>, and C<syscall>. System calls also set the special C<$!>
5a964f20
TC
85variable on failure. Other functions do not, except accidentally.
86
cb1a09d0
AD
87=head2 Perl Functions by Category
88
89Here are Perl's functions (including things that look like
5a964f20 90functions, like some keywords and named operators)
cb1a09d0
AD
91arranged by category. Some functions appear in more
92than one place.
93
13a2d996 94=over 4
cb1a09d0
AD
95
96=item Functions for SCALARs or strings
97
22fae026 98C<chomp>, C<chop>, C<chr>, C<crypt>, C<hex>, C<index>, C<lc>, C<lcfirst>,
945c54fd
JH
99C<length>, C<oct>, C<ord>, C<pack>, C<q/STRING/>, C<qq/STRING/>, C<reverse>,
100C<rindex>, C<sprintf>, C<substr>, C<tr///>, C<uc>, C<ucfirst>, C<y///>
cb1a09d0
AD
101
102=item Regular expressions and pattern matching
103
ab4f32c2 104C<m//>, C<pos>, C<quotemeta>, C<s///>, C<split>, C<study>, C<qr//>
cb1a09d0
AD
105
106=item Numeric functions
107
22fae026
TM
108C<abs>, C<atan2>, C<cos>, C<exp>, C<hex>, C<int>, C<log>, C<oct>, C<rand>,
109C<sin>, C<sqrt>, C<srand>
cb1a09d0
AD
110
111=item Functions for real @ARRAYs
112
22fae026 113C<pop>, C<push>, C<shift>, C<splice>, C<unshift>
cb1a09d0
AD
114
115=item Functions for list data
116
ab4f32c2 117C<grep>, C<join>, C<map>, C<qw/STRING/>, C<reverse>, C<sort>, C<unpack>
cb1a09d0
AD
118
119=item Functions for real %HASHes
120
22fae026 121C<delete>, C<each>, C<exists>, C<keys>, C<values>
cb1a09d0
AD
122
123=item Input and output functions
124
22fae026
TM
125C<binmode>, C<close>, C<closedir>, C<dbmclose>, C<dbmopen>, C<die>, C<eof>,
126C<fileno>, C<flock>, C<format>, C<getc>, C<print>, C<printf>, C<read>,
127C<readdir>, C<rewinddir>, C<seek>, C<seekdir>, C<select>, C<syscall>,
128C<sysread>, C<sysseek>, C<syswrite>, C<tell>, C<telldir>, C<truncate>,
129C<warn>, C<write>
cb1a09d0
AD
130
131=item Functions for fixed length data or records
132
22fae026 133C<pack>, C<read>, C<syscall>, C<sysread>, C<syswrite>, C<unpack>, C<vec>
cb1a09d0
AD
134
135=item Functions for filehandles, files, or directories
136
22fae026 137C<-I<X>>, C<chdir>, C<chmod>, C<chown>, C<chroot>, C<fcntl>, C<glob>,
5ff3f7a4 138C<ioctl>, C<link>, C<lstat>, C<mkdir>, C<open>, C<opendir>,
1e278fd9
JH
139C<readlink>, C<rename>, C<rmdir>, C<stat>, C<symlink>, C<sysopen>,
140C<umask>, C<unlink>, C<utime>
cb1a09d0 141
cf264981 142=item Keywords related to the control flow of your Perl program
cb1a09d0 143
98293880
JH
144C<caller>, C<continue>, C<die>, C<do>, C<dump>, C<eval>, C<exit>,
145C<goto>, C<last>, C<next>, C<redo>, C<return>, C<sub>, C<wantarray>
cb1a09d0 146
54310121 147=item Keywords related to scoping
cb1a09d0 148
4375e838 149C<caller>, C<import>, C<local>, C<my>, C<our>, C<package>, C<use>
cb1a09d0
AD
150
151=item Miscellaneous functions
152
4375e838 153C<defined>, C<dump>, C<eval>, C<formline>, C<local>, C<my>, C<our>, C<reset>,
22fae026 154C<scalar>, C<undef>, C<wantarray>
cb1a09d0
AD
155
156=item Functions for processes and process groups
157
22fae026 158C<alarm>, C<exec>, C<fork>, C<getpgrp>, C<getppid>, C<getpriority>, C<kill>,
ab4f32c2 159C<pipe>, C<qx/STRING/>, C<setpgrp>, C<setpriority>, C<sleep>, C<system>,
22fae026 160C<times>, C<wait>, C<waitpid>
cb1a09d0
AD
161
162=item Keywords related to perl modules
163
22fae026 164C<do>, C<import>, C<no>, C<package>, C<require>, C<use>
cb1a09d0
AD
165
166=item Keywords related to classes and object-orientedness
167
22fae026
TM
168C<bless>, C<dbmclose>, C<dbmopen>, C<package>, C<ref>, C<tie>, C<tied>,
169C<untie>, C<use>
cb1a09d0
AD
170
171=item Low-level socket functions
172
22fae026
TM
173C<accept>, C<bind>, C<connect>, C<getpeername>, C<getsockname>,
174C<getsockopt>, C<listen>, C<recv>, C<send>, C<setsockopt>, C<shutdown>,
737dd4b4 175C<socket>, C<socketpair>
cb1a09d0
AD
176
177=item System V interprocess communication functions
178
22fae026
TM
179C<msgctl>, C<msgget>, C<msgrcv>, C<msgsnd>, C<semctl>, C<semget>, C<semop>,
180C<shmctl>, C<shmget>, C<shmread>, C<shmwrite>
cb1a09d0
AD
181
182=item Fetching user and group info
183
22fae026
TM
184C<endgrent>, C<endhostent>, C<endnetent>, C<endpwent>, C<getgrent>,
185C<getgrgid>, C<getgrnam>, C<getlogin>, C<getpwent>, C<getpwnam>,
186C<getpwuid>, C<setgrent>, C<setpwent>
cb1a09d0
AD
187
188=item Fetching network info
189
22fae026
TM
190C<endprotoent>, C<endservent>, C<gethostbyaddr>, C<gethostbyname>,
191C<gethostent>, C<getnetbyaddr>, C<getnetbyname>, C<getnetent>,
192C<getprotobyname>, C<getprotobynumber>, C<getprotoent>,
193C<getservbyname>, C<getservbyport>, C<getservent>, C<sethostent>,
194C<setnetent>, C<setprotoent>, C<setservent>
cb1a09d0
AD
195
196=item Time-related functions
197
22fae026 198C<gmtime>, C<localtime>, C<time>, C<times>
cb1a09d0 199
37798a01
PP
200=item Functions new in perl5
201
22fae026 202C<abs>, C<bless>, C<chomp>, C<chr>, C<exists>, C<formline>, C<glob>,
b76cc8ba 203C<import>, C<lc>, C<lcfirst>, C<map>, C<my>, C<no>, C<our>, C<prototype>,
4375e838 204C<qx>, C<qw>, C<readline>, C<readpipe>, C<ref>, C<sub*>, C<sysopen>, C<tie>,
22fae026 205C<tied>, C<uc>, C<ucfirst>, C<untie>, C<use>
37798a01
PP
206
207* - C<sub> was a keyword in perl4, but in perl5 it is an
5a964f20 208operator, which can be used in expressions.
37798a01
PP
209
210=item Functions obsoleted in perl5
211
22fae026 212C<dbmclose>, C<dbmopen>
37798a01 213
cb1a09d0
AD
214=back
215
60f9f73c
JH
216=head2 Portability
217
2b5ab1e7
TC
218Perl was born in Unix and can therefore access all common Unix
219system calls. In non-Unix environments, the functionality of some
220Unix system calls may not be available, or details of the available
221functionality may differ slightly. The Perl functions affected
60f9f73c
JH
222by this are:
223
224C<-X>, C<binmode>, C<chmod>, C<chown>, C<chroot>, C<crypt>,
225C<dbmclose>, C<dbmopen>, C<dump>, C<endgrent>, C<endhostent>,
226C<endnetent>, C<endprotoent>, C<endpwent>, C<endservent>, C<exec>,
ef5a6dd7
JH
227C<fcntl>, C<flock>, C<fork>, C<getgrent>, C<getgrgid>, C<gethostbyname>,
228C<gethostent>, C<getlogin>, C<getnetbyaddr>, C<getnetbyname>, C<getnetent>,
54d7b083 229C<getppid>, C<getpgrp>, C<getpriority>, C<getprotobynumber>,
60f9f73c
JH
230C<getprotoent>, C<getpwent>, C<getpwnam>, C<getpwuid>,
231C<getservbyport>, C<getservent>, C<getsockopt>, C<glob>, C<ioctl>,
232C<kill>, C<link>, C<lstat>, C<msgctl>, C<msgget>, C<msgrcv>,
2b5ab1e7 233C<msgsnd>, C<open>, C<pipe>, C<readlink>, C<rename>, C<select>, C<semctl>,
60f9f73c
JH
234C<semget>, C<semop>, C<setgrent>, C<sethostent>, C<setnetent>,
235C<setpgrp>, C<setpriority>, C<setprotoent>, C<setpwent>,
236C<setservent>, C<setsockopt>, C<shmctl>, C<shmget>, C<shmread>,
737dd4b4 237C<shmwrite>, C<socket>, C<socketpair>,
80cbd5ad
JH
238C<stat>, C<symlink>, C<syscall>, C<sysopen>, C<system>,
239C<times>, C<truncate>, C<umask>, C<unlink>,
2b5ab1e7 240C<utime>, C<wait>, C<waitpid>
60f9f73c
JH
241
242For more information about the portability of these functions, see
243L<perlport> and other available platform-specific documentation.
244
cb1a09d0
AD
245=head2 Alphabetical Listing of Perl Functions
246
a0d0e21e
LW
247=over 8
248
5b3c99c0 249=item -X FILEHANDLE
a0d0e21e 250
5b3c99c0 251=item -X EXPR
a0d0e21e 252
5b3c99c0 253=item -X
a0d0e21e
LW
254
255A file test, where X is one of the letters listed below. This unary
256operator takes one argument, either a filename or a filehandle, and
257tests the associated file to see if something is true about it. If the
7660c0ab 258argument is omitted, tests C<$_>, except for C<-t>, which tests STDIN.
19799a22 259Unless otherwise documented, it returns C<1> for true and C<''> for false, or
a0d0e21e
LW
260the undefined value if the file doesn't exist. Despite the funny
261names, precedence is the same as any other named unary operator, and
262the argument may be parenthesized like any other unary operator. The
263operator may be any of:
7e778d91
IZ
264X<-r>X<-w>X<-x>X<-o>X<-R>X<-W>X<-X>X<-O>X<-e>X<-z>X<-s>X<-f>X<-d>X<-l>X<-p>
265X<-S>X<-b>X<-c>X<-t>X<-u>X<-g>X<-k>X<-T>X<-B>X<-M>X<-A>X<-C>
a0d0e21e
LW
266
267 -r File is readable by effective uid/gid.
268 -w File is writable by effective uid/gid.
269 -x File is executable by effective uid/gid.
270 -o File is owned by effective uid.
271
272 -R File is readable by real uid/gid.
273 -W File is writable by real uid/gid.
274 -X File is executable by real uid/gid.
275 -O File is owned by real uid.
276
277 -e File exists.
8e7e0aa8
MJD
278 -z File has zero size (is empty).
279 -s File has nonzero size (returns size in bytes).
a0d0e21e
LW
280
281 -f File is a plain file.
282 -d File is a directory.
283 -l File is a symbolic link.
9c4d0f16 284 -p File is a named pipe (FIFO), or Filehandle is a pipe.
a0d0e21e
LW
285 -S File is a socket.
286 -b File is a block special file.
287 -c File is a character special file.
288 -t Filehandle is opened to a tty.
289
290 -u File has setuid bit set.
291 -g File has setgid bit set.
292 -k File has sticky bit set.
293
121910a4 294 -T File is an ASCII text file (heuristic guess).
2cdbc966 295 -B File is a "binary" file (opposite of -T).
a0d0e21e 296
95a3fe12 297 -M Script start time minus file modification time, in days.
a0d0e21e 298 -A Same for access time.
95a3fe12 299 -C Same for inode change time (Unix, may differ for other platforms)
a0d0e21e 300
a0d0e21e
LW
301Example:
302
303 while (<>) {
5b3eff12 304 chomp;
a0d0e21e 305 next unless -f $_; # ignore specials
5a964f20 306 #...
a0d0e21e
LW
307 }
308
5ff3f7a4
GS
309The interpretation of the file permission operators C<-r>, C<-R>,
310C<-w>, C<-W>, C<-x>, and C<-X> is by default based solely on the mode
311of the file and the uids and gids of the user. There may be other
312reasons you can't actually read, write, or execute the file. Such
313reasons may be for example network filesystem access controls, ACLs
314(access control lists), read-only filesystems, and unrecognized
315executable formats.
316
2b5ab1e7
TC
317Also note that, for the superuser on the local filesystems, the C<-r>,
318C<-R>, C<-w>, and C<-W> tests always return 1, and C<-x> and C<-X> return 1
5ff3f7a4
GS
319if any execute bit is set in the mode. Scripts run by the superuser
320may thus need to do a stat() to determine the actual mode of the file,
2b5ab1e7 321or temporarily set their effective uid to something else.
5ff3f7a4
GS
322
323If you are using ACLs, there is a pragma called C<filetest> that may
324produce more accurate results than the bare stat() mode bits.
5ff3f7a4
GS
325When under the C<use filetest 'access'> the above-mentioned filetests
326will test whether the permission can (not) be granted using the
468541a8 327access() family of system calls. Also note that the C<-x> and C<-X> may
5ff3f7a4
GS
328under this pragma return true even if there are no execute permission
329bits set (nor any extra execute permission ACLs). This strangeness is
330due to the underlying system calls' definitions. Read the
331documentation for the C<filetest> pragma for more information.
332
a0d0e21e
LW
333Note that C<-s/a/b/> does not do a negated substitution. Saying
334C<-exp($foo)> still works as expected, however--only single letters
335following a minus are interpreted as file tests.
336
337The C<-T> and C<-B> switches work as follows. The first block or so of the
338file is examined for odd characters such as strange control codes or
61eff3bc 339characters with the high bit set. If too many strange characters (>30%)
cf264981 340are found, it's a C<-B> file; otherwise it's a C<-T> file. Also, any file
a0d0e21e 341containing null in the first block is considered a binary file. If C<-T>
9124316e 342or C<-B> is used on a filehandle, the current IO buffer is examined
19799a22 343rather than the first block. Both C<-T> and C<-B> return true on a null
54310121 344file, or a file at EOF when testing a filehandle. Because you have to
4633a7c4
LW
345read a file to do the C<-T> test, on most occasions you want to use a C<-f>
346against the file first, as in C<next unless -f $file && -T $file>.
a0d0e21e 347
19799a22 348If any of the file tests (or either the C<stat> or C<lstat> operators) are given
28757baa 349the special filehandle consisting of a solitary underline, then the stat
a0d0e21e
LW
350structure of the previous file test (or stat operator) is used, saving
351a system call. (This doesn't work with C<-t>, and you need to remember
352that lstat() and C<-l> will leave values in the stat structure for the
5c9aa243 353symbolic link, not the real file.) (Also, if the stat buffer was filled by
cf264981 354an C<lstat> call, C<-T> and C<-B> will reset it with the results of C<stat _>).
5c9aa243 355Example:
a0d0e21e
LW
356
357 print "Can do.\n" if -r $a || -w _ || -x _;
358
359 stat($filename);
360 print "Readable\n" if -r _;
361 print "Writable\n" if -w _;
362 print "Executable\n" if -x _;
363 print "Setuid\n" if -u _;
364 print "Setgid\n" if -g _;
365 print "Sticky\n" if -k _;
366 print "Text\n" if -T _;
367 print "Binary\n" if -B _;
368
fbb0b3b3
RGS
369As of Perl 5.9.1, as a form of purely syntactic sugar, you can stack file
370test operators, in a way that C<-f -w -x $file> is equivalent to
cf264981 371C<-x $file && -w _ && -f _>. (This is only syntax fancy: if you use
fbb0b3b3
RGS
372the return value of C<-f $file> as an argument to another filetest
373operator, no special magic will happen.)
374
a0d0e21e
LW
375=item abs VALUE
376
54310121 377=item abs
bbce6d69 378
a0d0e21e 379Returns the absolute value of its argument.
7660c0ab 380If VALUE is omitted, uses C<$_>.
a0d0e21e
LW
381
382=item accept NEWSOCKET,GENERICSOCKET
383
f86cebdf 384Accepts an incoming socket connect, just as the accept(2) system call
19799a22 385does. Returns the packed address if it succeeded, false otherwise.
2b5ab1e7 386See the example in L<perlipc/"Sockets: Client/Server Communication">.
a0d0e21e 387
8d2a6795
GS
388On systems that support a close-on-exec flag on files, the flag will
389be set for the newly opened file descriptor, as determined by the
390value of $^F. See L<perlvar/$^F>.
391
a0d0e21e
LW
392=item alarm SECONDS
393
54310121 394=item alarm
bbce6d69 395
a0d0e21e 396Arranges to have a SIGALRM delivered to this process after the
cf264981 397specified number of wallclock seconds has elapsed. If SECONDS is not
d400eac8
JH
398specified, the value stored in C<$_> is used. (On some machines,
399unfortunately, the elapsed time may be up to one second less or more
400than you specified because of how seconds are counted, and process
401scheduling may delay the delivery of the signal even further.)
402
403Only one timer may be counting at once. Each call disables the
404previous timer, and an argument of C<0> may be supplied to cancel the
405previous timer without starting a new one. The returned value is the
406amount of time remaining on the previous timer.
a0d0e21e 407
4633a7c4 408For delays of finer granularity than one second, you may use Perl's
19799a22
GS
409four-argument version of select() leaving the first three arguments
410undefined, or you might be able to use the C<syscall> interface to
83df6a1d
JH
411access setitimer(2) if your system supports it. The Time::HiRes
412module (from CPAN, and starting from Perl 5.8 part of the standard
413distribution) may also prove useful.
2b5ab1e7 414
68f8bed4
JH
415It is usually a mistake to intermix C<alarm> and C<sleep> calls.
416(C<sleep> may be internally implemented in your system with C<alarm>)
a0d0e21e 417
19799a22
GS
418If you want to use C<alarm> to time out a system call you need to use an
419C<eval>/C<die> pair. You can't rely on the alarm causing the system call to
f86cebdf 420fail with C<$!> set to C<EINTR> because Perl sets up signal handlers to
19799a22 421restart system calls on some systems. Using C<eval>/C<die> always works,
5a964f20 422modulo the caveats given in L<perlipc/"Signals">.
ff68c719
PP
423
424 eval {
f86cebdf 425 local $SIG{ALRM} = sub { die "alarm\n" }; # NB: \n required
36477c24 426 alarm $timeout;
ff68c719 427 $nread = sysread SOCKET, $buffer, $size;
36477c24 428 alarm 0;
ff68c719 429 };
ff68c719 430 if ($@) {
f86cebdf 431 die unless $@ eq "alarm\n"; # propagate unexpected errors
ff68c719
PP
432 # timed out
433 }
434 else {
435 # didn't
436 }
437
91d81acc
JH
438For more information see L<perlipc>.
439
a0d0e21e
LW
440=item atan2 Y,X
441
442Returns the arctangent of Y/X in the range -PI to PI.
443
ca6e1c26 444For the tangent operation, you may use the C<Math::Trig::tan>
28757baa
PP
445function, or use the familiar relation:
446
447 sub tan { sin($_[0]) / cos($_[0]) }
448
a0d0e21e
LW
449=item bind SOCKET,NAME
450
451Binds a network address to a socket, just as the bind system call
19799a22 452does. Returns true if it succeeded, false otherwise. NAME should be a
4633a7c4
LW
453packed address of the appropriate type for the socket. See the examples in
454L<perlipc/"Sockets: Client/Server Communication">.
a0d0e21e 455
fae2c0fb 456=item binmode FILEHANDLE, LAYER
1c1fc3ea 457
a0d0e21e
LW
458=item binmode FILEHANDLE
459
1cbfc93d
NIS
460Arranges for FILEHANDLE to be read or written in "binary" or "text"
461mode on systems where the run-time libraries distinguish between
462binary and text files. If FILEHANDLE is an expression, the value is
463taken as the name of the filehandle. Returns true on success,
b5fe5ca2 464otherwise it returns C<undef> and sets C<$!> (errno).
1cbfc93d 465
d807c6f4
JH
466On some systems (in general, DOS and Windows-based systems) binmode()
467is necessary when you're not working with a text file. For the sake
468of portability it is a good idea to always use it when appropriate,
469and to never use it when it isn't appropriate. Also, people can
470set their I/O to be by default UTF-8 encoded Unicode, not bytes.
471
472In other words: regardless of platform, use binmode() on binary data,
473like for example images.
474
475If LAYER is present it is a single string, but may contain multiple
476directives. The directives alter the behaviour of the file handle.
477When LAYER is present using binmode on text file makes sense.
478
fae2c0fb 479If LAYER is omitted or specified as C<:raw> the filehandle is made
0226bbdb
NIS
480suitable for passing binary data. This includes turning off possible CRLF
481translation and marking it as bytes (as opposed to Unicode characters).
749683d2
YST
482Note that, despite what may be implied in I<"Programming Perl"> (the
483Camel) or elsewhere, C<:raw> is I<not> the simply inverse of C<:crlf>
fae2c0fb 484-- other layers which would affect binary nature of the stream are
0226bbdb
NIS
485I<also> disabled. See L<PerlIO>, L<perlrun> and the discussion about the
486PERLIO environment variable.
01e6739c 487
d807c6f4
JH
488The C<:bytes>, C<:crlf>, and C<:utf8>, and any other directives of the
489form C<:...>, are called I/O I<layers>. The C<open> pragma can be used to
490establish default I/O layers. See L<open>.
491
fae2c0fb
RGS
492I<The LAYER parameter of the binmode() function is described as "DISCIPLINE"
493in "Programming Perl, 3rd Edition". However, since the publishing of this
494book, by many known as "Camel III", the consensus of the naming of this
495functionality has moved from "discipline" to "layer". All documentation
496of this version of Perl therefore refers to "layers" rather than to
497"disciplines". Now back to the regularly scheduled documentation...>
498
01e6739c 499To mark FILEHANDLE as UTF-8, use C<:utf8>.
1cbfc93d 500
ed53a2bb 501In general, binmode() should be called after open() but before any I/O
01e6739c
NIS
502is done on the filehandle. Calling binmode() will normally flush any
503pending buffered output data (and perhaps pending input data) on the
fae2c0fb 504handle. An exception to this is the C<:encoding> layer that
01e6739c 505changes the default character encoding of the handle, see L<open>.
fae2c0fb 506The C<:encoding> layer sometimes needs to be called in
3874323d
JH
507mid-stream, and it doesn't flush the stream. The C<:encoding>
508also implicitly pushes on top of itself the C<:utf8> layer because
509internally Perl will operate on UTF-8 encoded Unicode characters.
16fe6d59 510
19799a22 511The operating system, device drivers, C libraries, and Perl run-time
30168b04
GS
512system all work together to let the programmer treat a single
513character (C<\n>) as the line terminator, irrespective of the external
514representation. On many operating systems, the native text file
515representation matches the internal representation, but on some
516platforms the external representation of C<\n> is made up of more than
517one character.
518
68bd7414
NIS
519Mac OS, all variants of Unix, and Stream_LF files on VMS use a single
520character to end each line in the external representation of text (even
5e12dbfa 521though that single character is CARRIAGE RETURN on Mac OS and LINE FEED
01e6739c
NIS
522on Unix and most VMS files). In other systems like OS/2, DOS and the
523various flavors of MS-Windows your program sees a C<\n> as a simple C<\cJ>,
524but what's stored in text files are the two characters C<\cM\cJ>. That
525means that, if you don't use binmode() on these systems, C<\cM\cJ>
526sequences on disk will be converted to C<\n> on input, and any C<\n> in
527your program will be converted back to C<\cM\cJ> on output. This is what
528you want for text files, but it can be disastrous for binary files.
30168b04
GS
529
530Another consequence of using binmode() (on some systems) is that
531special end-of-file markers will be seen as part of the data stream.
532For systems from the Microsoft family this means that if your binary
4375e838 533data contains C<\cZ>, the I/O subsystem will regard it as the end of
30168b04
GS
534the file, unless you use binmode().
535
536binmode() is not only important for readline() and print() operations,
537but also when using read(), seek(), sysread(), syswrite() and tell()
538(see L<perlport> for more details). See the C<$/> and C<$\> variables
539in L<perlvar> for how to manually set your input and output
540line-termination sequences.
a0d0e21e 541
4633a7c4 542=item bless REF,CLASSNAME
a0d0e21e
LW
543
544=item bless REF
545
2b5ab1e7
TC
546This function tells the thingy referenced by REF that it is now an object
547in the CLASSNAME package. If CLASSNAME is omitted, the current package
19799a22 548is used. Because a C<bless> is often the last thing in a constructor,
2b5ab1e7 549it returns the reference for convenience. Always use the two-argument
cf264981
SP
550version if a derived class might inherit the function doing the blessing.
551See L<perltoot> and L<perlobj> for more about the blessing (and blessings)
552of objects.
a0d0e21e 553
57668c4d 554Consider always blessing objects in CLASSNAMEs that are mixed case.
2b5ab1e7 555Namespaces with all lowercase names are considered reserved for
cf264981 556Perl pragmata. Builtin types have all uppercase names. To prevent
2b5ab1e7
TC
557confusion, you may wish to avoid such package names as well. Make sure
558that CLASSNAME is a true value.
60ad88b8
GS
559
560See L<perlmod/"Perl Modules">.
561
a0d0e21e
LW
562=item caller EXPR
563
564=item caller
565
5a964f20 566Returns the context of the current subroutine call. In scalar context,
28757baa 567returns the caller's package name if there is a caller, that is, if
19799a22 568we're in a subroutine or C<eval> or C<require>, and the undefined value
5a964f20 569otherwise. In list context, returns
a0d0e21e 570
748a9306 571 ($package, $filename, $line) = caller;
a0d0e21e
LW
572
573With EXPR, it returns some extra information that the debugger uses to
574print a stack trace. The value of EXPR indicates how many call frames
575to go back before the current one.
576
f3aa04c2 577 ($package, $filename, $line, $subroutine, $hasargs,
e476b1b5 578 $wantarray, $evaltext, $is_require, $hints, $bitmask) = caller($i);
e7ea3e70 579
951ba7fe 580Here $subroutine may be C<(eval)> if the frame is not a subroutine
19799a22 581call, but an C<eval>. In such a case additional elements $evaltext and
7660c0ab 582C<$is_require> are set: C<$is_require> is true if the frame is created by a
19799a22 583C<require> or C<use> statement, $evaltext contains the text of the
277ddfaf 584C<eval EXPR> statement. In particular, for an C<eval BLOCK> statement,
951ba7fe 585$filename is C<(eval)>, but $evaltext is undefined. (Note also that
0fc9dec4
RGS
586each C<use> statement creates a C<require> frame inside an C<eval EXPR>
587frame.) $subroutine may also be C<(unknown)> if this particular
588subroutine happens to have been deleted from the symbol table.
589C<$hasargs> is true if a new instance of C<@_> was set up for the frame.
590C<$hints> and C<$bitmask> contain pragmatic hints that the caller was
591compiled with. The C<$hints> and C<$bitmask> values are subject to change
592between versions of Perl, and are not meant for external use.
748a9306
LW
593
594Furthermore, when called from within the DB package, caller returns more
7660c0ab 595detailed information: it sets the list variable C<@DB::args> to be the
54310121 596arguments with which the subroutine was invoked.
748a9306 597
7660c0ab 598Be aware that the optimizer might have optimized call frames away before
19799a22 599C<caller> had a chance to get the information. That means that C<caller(N)>
7660c0ab 600might not return information about the call frame you expect it do, for
b76cc8ba 601C<< N > 1 >>. In particular, C<@DB::args> might have information from the
19799a22 602previous time C<caller> was called.
7660c0ab 603
a0d0e21e
LW
604=item chdir EXPR
605
c4aca7d0
GA
606=item chdir FILEHANDLE
607
608=item chdir DIRHANDLE
609
ce2984c3
PF
610=item chdir
611
ffce7b87 612Changes the working directory to EXPR, if possible. If EXPR is omitted,
0bfc1ec4 613changes to the directory specified by C<$ENV{HOME}>, if set; if not,
ffce7b87 614changes to the directory specified by C<$ENV{LOGDIR}>. (Under VMS, the
b4ad75f0
AMS
615variable C<$ENV{SYS$LOGIN}> is also checked, and used if it is set.) If
616neither is set, C<chdir> does nothing. It returns true upon success,
617false otherwise. See the example under C<die>.
a0d0e21e 618
c4aca7d0
GA
619On systems that support fchdir, you might pass a file handle or
620directory handle as argument. On systems that don't support fchdir,
621passing handles produces a fatal error at run time.
622
a0d0e21e
LW
623=item chmod LIST
624
625Changes the permissions of a list of files. The first element of the
4633a7c4 626list must be the numerical mode, which should probably be an octal
4ad40acf 627number, and which definitely should I<not> be a string of octal digits:
2f9daede 628C<0644> is okay, C<'0644'> is not. Returns the number of files
dc848c6f 629successfully changed. See also L</oct>, if all you have is a string.
a0d0e21e
LW
630
631 $cnt = chmod 0755, 'foo', 'bar';
632 chmod 0755, @executables;
f86cebdf
GS
633 $mode = '0644'; chmod $mode, 'foo'; # !!! sets mode to
634 # --w----r-T
2f9daede
TPG
635 $mode = '0644'; chmod oct($mode), 'foo'; # this is better
636 $mode = 0644; chmod $mode, 'foo'; # this is best
a0d0e21e 637
c4aca7d0
GA
638On systems that support fchmod, you might pass file handles among the
639files. On systems that don't support fchmod, passing file handles
640produces a fatal error at run time.
641
642 open(my $fh, "<", "foo");
643 my $perm = (stat $fh)[2] & 07777;
644 chmod($perm | 0600, $fh);
645
ca6e1c26
JH
646You can also import the symbolic C<S_I*> constants from the Fcntl
647module:
648
649 use Fcntl ':mode';
650
651 chmod S_IRWXU|S_IRGRP|S_IXGRP|S_IROTH|S_IXOTH, @executables;
652 # This is identical to the chmod 0755 of the above example.
653
a0d0e21e
LW
654=item chomp VARIABLE
655
313c9f5c 656=item chomp( LIST )
a0d0e21e
LW
657
658=item chomp
659
2b5ab1e7
TC
660This safer version of L</chop> removes any trailing string
661that corresponds to the current value of C<$/> (also known as
28757baa
PP
662$INPUT_RECORD_SEPARATOR in the C<English> module). It returns the total
663number of characters removed from all its arguments. It's often used to
664remove the newline from the end of an input record when you're worried
2b5ab1e7
TC
665that the final record may be missing its newline. When in paragraph
666mode (C<$/ = "">), it removes all trailing newlines from the string.
4c5a6083
GS
667When in slurp mode (C<$/ = undef>) or fixed-length record mode (C<$/> is
668a reference to an integer or the like, see L<perlvar>) chomp() won't
b76cc8ba 669remove anything.
19799a22 670If VARIABLE is omitted, it chomps C<$_>. Example:
a0d0e21e
LW
671
672 while (<>) {
673 chomp; # avoid \n on last field
674 @array = split(/:/);
5a964f20 675 # ...
a0d0e21e
LW
676 }
677
4bf21a6d
RD
678If VARIABLE is a hash, it chomps the hash's values, but not its keys.
679
a0d0e21e
LW
680You can actually chomp anything that's an lvalue, including an assignment:
681
682 chomp($cwd = `pwd`);
683 chomp($answer = <STDIN>);
684
685If you chomp a list, each element is chomped, and the total number of
686characters removed is returned.
687
442a8c12
NC
688If the C<encoding> pragma is in scope then the lengths returned are
689calculated from the length of C<$/> in Unicode characters, which is not
690always the same as the length of C<$/> in the native encoding.
691
15e44fd8
RGS
692Note that parentheses are necessary when you're chomping anything
693that is not a simple variable. This is because C<chomp $cwd = `pwd`;>
694is interpreted as C<(chomp $cwd) = `pwd`;>, rather than as
695C<chomp( $cwd = `pwd` )> which you might expect. Similarly,
696C<chomp $a, $b> is interpreted as C<chomp($a), $b> rather than
697as C<chomp($a, $b)>.
698
a0d0e21e
LW
699=item chop VARIABLE
700
313c9f5c 701=item chop( LIST )
a0d0e21e
LW
702
703=item chop
704
705Chops off the last character of a string and returns the character
5b3eff12 706chopped. It is much more efficient than C<s/.$//s> because it neither
7660c0ab 707scans nor copies the string. If VARIABLE is omitted, chops C<$_>.
4bf21a6d
RD
708If VARIABLE is a hash, it chops the hash's values, but not its keys.
709
5b3eff12 710You can actually chop anything that's an lvalue, including an assignment.
a0d0e21e
LW
711
712If you chop a list, each element is chopped. Only the value of the
19799a22 713last C<chop> is returned.
a0d0e21e 714
19799a22 715Note that C<chop> returns the last character. To return all but the last
748a9306
LW
716character, use C<substr($string, 0, -1)>.
717
15e44fd8
RGS
718See also L</chomp>.
719
a0d0e21e
LW
720=item chown LIST
721
722Changes the owner (and group) of a list of files. The first two
19799a22
GS
723elements of the list must be the I<numeric> uid and gid, in that
724order. A value of -1 in either position is interpreted by most
725systems to leave that value unchanged. Returns the number of files
726successfully changed.
a0d0e21e
LW
727
728 $cnt = chown $uid, $gid, 'foo', 'bar';
729 chown $uid, $gid, @filenames;
730
c4aca7d0
GA
731On systems that support fchown, you might pass file handles among the
732files. On systems that don't support fchown, passing file handles
733produces a fatal error at run time.
734
54310121 735Here's an example that looks up nonnumeric uids in the passwd file:
a0d0e21e
LW
736
737 print "User: ";
19799a22 738 chomp($user = <STDIN>);
5a964f20 739 print "Files: ";
19799a22 740 chomp($pattern = <STDIN>);
a0d0e21e
LW
741
742 ($login,$pass,$uid,$gid) = getpwnam($user)
743 or die "$user not in passwd file";
744
5a964f20 745 @ary = glob($pattern); # expand filenames
a0d0e21e
LW
746 chown $uid, $gid, @ary;
747
54310121 748On most systems, you are not allowed to change the ownership of the
4633a7c4
LW
749file unless you're the superuser, although you should be able to change
750the group to any of your secondary groups. On insecure systems, these
751restrictions may be relaxed, but this is not a portable assumption.
19799a22
GS
752On POSIX systems, you can detect this condition this way:
753
754 use POSIX qw(sysconf _PC_CHOWN_RESTRICTED);
755 $can_chown_giveaway = not sysconf(_PC_CHOWN_RESTRICTED);
4633a7c4 756
a0d0e21e
LW
757=item chr NUMBER
758
54310121 759=item chr
bbce6d69 760
a0d0e21e 761Returns the character represented by that NUMBER in the character set.
a0ed51b3 762For example, C<chr(65)> is C<"A"> in either ASCII or Unicode, and
1e54db1a
JH
763chr(0x263a) is a Unicode smiley face. Note that characters from 128
764to 255 (inclusive) are by default not encoded in UTF-8 Unicode for
765backward compatibility reasons (but see L<encoding>).
aaa68c4a 766
8a064bd6
JH
767Negative values give the Unicode replacement character (chr(0xfffd)),
768except under the L</bytes> pragma, where low eight bits of the value
769(truncated to an integer) are used.
770
974da8e5
JH
771If NUMBER is omitted, uses C<$_>.
772
b76cc8ba 773For the reverse, use L</ord>.
a0d0e21e 774
974da8e5
JH
775Note that under the C<bytes> pragma the NUMBER is masked to
776the low eight bits.
777
778See L<perlunicode> and L<encoding> for more about Unicode.
bbce6d69 779
a0d0e21e
LW
780=item chroot FILENAME
781
54310121 782=item chroot
bbce6d69 783
5a964f20 784This function works like the system call by the same name: it makes the
4633a7c4 785named directory the new root directory for all further pathnames that
951ba7fe 786begin with a C</> by your process and all its children. (It doesn't
28757baa 787change your current working directory, which is unaffected.) For security
4633a7c4 788reasons, this call is restricted to the superuser. If FILENAME is
19799a22 789omitted, does a C<chroot> to C<$_>.
a0d0e21e
LW
790
791=item close FILEHANDLE
792
6a518fbc
TP
793=item close
794
9124316e
JH
795Closes the file or pipe associated with the file handle, returning
796true only if IO buffers are successfully flushed and closes the system
797file descriptor. Closes the currently selected filehandle if the
798argument is omitted.
fb73857a
PP
799
800You don't have to close FILEHANDLE if you are immediately going to do
19799a22
GS
801another C<open> on it, because C<open> will close it for you. (See
802C<open>.) However, an explicit C<close> on an input file resets the line
803counter (C<$.>), while the implicit close done by C<open> does not.
fb73857a 804
dede8123
RGS
805If the file handle came from a piped open, C<close> will additionally
806return false if one of the other system calls involved fails, or if the
fb73857a 807program exits with non-zero status. (If the only problem was that the
dede8123 808program exited non-zero, C<$!> will be set to C<0>.) Closing a pipe
2b5ab1e7 809also waits for the process executing on the pipe to complete, in case you
b76cc8ba 810want to look at the output of the pipe afterwards, and
e5218da5
GA
811implicitly puts the exit status value of that command into C<$?> and
812C<${^CHILD_ERROR_NATIVE}>.
5a964f20 813
73689b13
GS
814Prematurely closing the read end of a pipe (i.e. before the process
815writing to it at the other end has closed it) will result in a
816SIGPIPE being delivered to the writer. If the other end can't
817handle that, be sure to read all the data before closing the pipe.
818
fb73857a 819Example:
a0d0e21e 820
fb73857a
PP
821 open(OUTPUT, '|sort >foo') # pipe to sort
822 or die "Can't start sort: $!";
5a964f20 823 #... # print stuff to output
fb73857a
PP
824 close OUTPUT # wait for sort to finish
825 or warn $! ? "Error closing sort pipe: $!"
826 : "Exit status $? from sort";
827 open(INPUT, 'foo') # get sort's results
828 or die "Can't open 'foo' for input: $!";
a0d0e21e 829
5a964f20
TC
830FILEHANDLE may be an expression whose value can be used as an indirect
831filehandle, usually the real filehandle name.
a0d0e21e
LW
832
833=item closedir DIRHANDLE
834
19799a22 835Closes a directory opened by C<opendir> and returns the success of that
5a964f20
TC
836system call.
837
a0d0e21e
LW
838=item connect SOCKET,NAME
839
840Attempts to connect to a remote socket, just as the connect system call
19799a22 841does. Returns true if it succeeded, false otherwise. NAME should be a
4633a7c4
LW
842packed address of the appropriate type for the socket. See the examples in
843L<perlipc/"Sockets: Client/Server Communication">.
a0d0e21e 844
cb1a09d0
AD
845=item continue BLOCK
846
cf264981
SP
847C<continue> is actually a flow control statement rather than a function. If
848there is a C<continue> BLOCK attached to a BLOCK (typically in a C<while> or
98293880
JH
849C<foreach>), it is always executed just before the conditional is about to
850be evaluated again, just like the third part of a C<for> loop in C. Thus
cb1a09d0
AD
851it can be used to increment a loop variable, even when the loop has been
852continued via the C<next> statement (which is similar to the C C<continue>
853statement).
854
98293880 855C<last>, C<next>, or C<redo> may appear within a C<continue>
19799a22
GS
856block. C<last> and C<redo> will behave as if they had been executed within
857the main block. So will C<next>, but since it will execute a C<continue>
1d2dff63
GS
858block, it may be more entertaining.
859
860 while (EXPR) {
861 ### redo always comes here
862 do_something;
863 } continue {
864 ### next always comes here
865 do_something_else;
866 # then back the top to re-check EXPR
867 }
868 ### last always comes here
869
870Omitting the C<continue> section is semantically equivalent to using an
19799a22 871empty one, logically enough. In that case, C<next> goes directly back
1d2dff63
GS
872to check the condition at the top of the loop.
873
a0d0e21e
LW
874=item cos EXPR
875
d6217f1e
GS
876=item cos
877
5a964f20 878Returns the cosine of EXPR (expressed in radians). If EXPR is omitted,
7660c0ab 879takes cosine of C<$_>.
a0d0e21e 880
ca6e1c26 881For the inverse cosine operation, you may use the C<Math::Trig::acos()>
28757baa
PP
882function, or use this relation:
883
884 sub acos { atan2( sqrt(1 - $_[0] * $_[0]), $_[0] ) }
885
a0d0e21e
LW
886=item crypt PLAINTEXT,SALT
887
ef2e6798
MS
888Creates a digest string exactly like the crypt(3) function in the C
889library (assuming that you actually have a version there that has not
cf264981 890been extirpated as a potential munitions).
ef2e6798
MS
891
892crypt() is a one-way hash function. The PLAINTEXT and SALT is turned
893into a short string, called a digest, which is returned. The same
894PLAINTEXT and SALT will always return the same string, but there is no
895(known) way to get the original PLAINTEXT from the hash. Small
896changes in the PLAINTEXT or SALT will result in large changes in the
897digest.
898
899There is no decrypt function. This function isn't all that useful for
900cryptography (for that, look for F<Crypt> modules on your nearby CPAN
901mirror) and the name "crypt" is a bit of a misnomer. Instead it is
902primarily used to check if two pieces of text are the same without
903having to transmit or store the text itself. An example is checking
904if a correct password is given. The digest of the password is stored,
cf264981 905not the password itself. The user types in a password that is
ef2e6798
MS
906crypt()'d with the same salt as the stored digest. If the two digests
907match the password is correct.
908
909When verifying an existing digest string you should use the digest as
910the salt (like C<crypt($plain, $digest) eq $digest>). The SALT used
cf264981 911to create the digest is visible as part of the digest. This ensures
ef2e6798
MS
912crypt() will hash the new string with the same salt as the digest.
913This allows your code to work with the standard L<crypt|/crypt> and
914with more exotic implementations. In other words, do not assume
915anything about the returned string itself, or how many bytes in the
916digest matter.
85c16d83
JH
917
918Traditionally the result is a string of 13 bytes: two first bytes of
919the salt, followed by 11 bytes from the set C<[./0-9A-Za-z]>, and only
ef2e6798
MS
920the first eight bytes of the digest string mattered, but alternative
921hashing schemes (like MD5), higher level security schemes (like C2),
922and implementations on non-UNIX platforms may produce different
923strings.
85c16d83
JH
924
925When choosing a new salt create a random two character string whose
926characters come from the set C<[./0-9A-Za-z]> (like C<join '', ('.',
d3989d75
CW
927'/', 0..9, 'A'..'Z', 'a'..'z')[rand 64, rand 64]>). This set of
928characters is just a recommendation; the characters allowed in
929the salt depend solely on your system's crypt library, and Perl can't
930restrict what salts C<crypt()> accepts.
e71965be 931
a0d0e21e 932Here's an example that makes sure that whoever runs this program knows
cf264981 933their password:
a0d0e21e
LW
934
935 $pwd = (getpwuid($<))[1];
a0d0e21e
LW
936
937 system "stty -echo";
938 print "Password: ";
e71965be 939 chomp($word = <STDIN>);
a0d0e21e
LW
940 print "\n";
941 system "stty echo";
942
e71965be 943 if (crypt($word, $pwd) ne $pwd) {
a0d0e21e
LW
944 die "Sorry...\n";
945 } else {
946 print "ok\n";
54310121 947 }
a0d0e21e 948
9f8f0c9d 949Of course, typing in your own password to whoever asks you
748a9306 950for it is unwise.
a0d0e21e 951
ef2e6798 952The L<crypt|/crypt> function is unsuitable for hashing large quantities
19799a22 953of data, not least of all because you can't get the information
ef2e6798 954back. Look at the L<Digest> module for more robust algorithms.
19799a22 955
f2791508
JH
956If using crypt() on a Unicode string (which I<potentially> has
957characters with codepoints above 255), Perl tries to make sense
958of the situation by trying to downgrade (a copy of the string)
959the string back to an eight-bit byte string before calling crypt()
960(on that copy). If that works, good. If not, crypt() dies with
961C<Wide character in crypt>.
85c16d83 962
aa689395 963=item dbmclose HASH
a0d0e21e 964
19799a22 965[This function has been largely superseded by the C<untie> function.]
a0d0e21e 966
aa689395 967Breaks the binding between a DBM file and a hash.
a0d0e21e 968
19799a22 969=item dbmopen HASH,DBNAME,MASK
a0d0e21e 970
19799a22 971[This function has been largely superseded by the C<tie> function.]
a0d0e21e 972
7b8d334a 973This binds a dbm(3), ndbm(3), sdbm(3), gdbm(3), or Berkeley DB file to a
19799a22
GS
974hash. HASH is the name of the hash. (Unlike normal C<open>, the first
975argument is I<not> a filehandle, even though it looks like one). DBNAME
aa689395
PP
976is the name of the database (without the F<.dir> or F<.pag> extension if
977any). If the database does not exist, it is created with protection
19799a22
GS
978specified by MASK (as modified by the C<umask>). If your system supports
979only the older DBM functions, you may perform only one C<dbmopen> in your
aa689395 980program. In older versions of Perl, if your system had neither DBM nor
19799a22 981ndbm, calling C<dbmopen> produced a fatal error; it now falls back to
aa689395
PP
982sdbm(3).
983
984If you don't have write access to the DBM file, you can only read hash
985variables, not set them. If you want to test whether you can write,
19799a22 986either use file tests or try setting a dummy hash entry inside an C<eval>,
aa689395 987which will trap the error.
a0d0e21e 988
19799a22
GS
989Note that functions such as C<keys> and C<values> may return huge lists
990when used on large DBM files. You may prefer to use the C<each>
a0d0e21e
LW
991function to iterate over large DBM files. Example:
992
993 # print out history file offsets
994 dbmopen(%HIST,'/usr/lib/news/history',0666);
995 while (($key,$val) = each %HIST) {
996 print $key, ' = ', unpack('L',$val), "\n";
997 }
998 dbmclose(%HIST);
999
cb1a09d0 1000See also L<AnyDBM_File> for a more general description of the pros and
184e9718 1001cons of the various dbm approaches, as well as L<DB_File> for a particularly
cb1a09d0 1002rich implementation.
4633a7c4 1003
2b5ab1e7
TC
1004You can control which DBM library you use by loading that library
1005before you call dbmopen():
1006
1007 use DB_File;
1008 dbmopen(%NS_Hist, "$ENV{HOME}/.netscape/history.db")
1009 or die "Can't open netscape history file: $!";
1010
a0d0e21e
LW
1011=item defined EXPR
1012
54310121 1013=item defined
bbce6d69 1014
2f9daede
TPG
1015Returns a Boolean value telling whether EXPR has a value other than
1016the undefined value C<undef>. If EXPR is not present, C<$_> will be
1017checked.
1018
1019Many operations return C<undef> to indicate failure, end of file,
1020system error, uninitialized variable, and other exceptional
1021conditions. This function allows you to distinguish C<undef> from
1022other values. (A simple Boolean test will not distinguish among
7660c0ab 1023C<undef>, zero, the empty string, and C<"0">, which are all equally
2f9daede 1024false.) Note that since C<undef> is a valid scalar, its presence
19799a22 1025doesn't I<necessarily> indicate an exceptional condition: C<pop>
2f9daede
TPG
1026returns C<undef> when its argument is an empty array, I<or> when the
1027element to return happens to be C<undef>.
1028
f10b0346
GS
1029You may also use C<defined(&func)> to check whether subroutine C<&func>
1030has ever been defined. The return value is unaffected by any forward
04891299 1031declarations of C<&func>. Note that a subroutine which is not defined
847c7ebe
DD
1032may still be callable: its package may have an C<AUTOLOAD> method that
1033makes it spring into existence the first time that it is called -- see
1034L<perlsub>.
f10b0346
GS
1035
1036Use of C<defined> on aggregates (hashes and arrays) is deprecated. It
1037used to report whether memory for that aggregate has ever been
1038allocated. This behavior may disappear in future versions of Perl.
1039You should instead use a simple test for size:
1040
1041 if (@an_array) { print "has array elements\n" }
1042 if (%a_hash) { print "has hash members\n" }
2f9daede
TPG
1043
1044When used on a hash element, it tells you whether the value is defined,
dc848c6f 1045not whether the key exists in the hash. Use L</exists> for the latter
2f9daede 1046purpose.
a0d0e21e
LW
1047
1048Examples:
1049
1050 print if defined $switch{'D'};
1051 print "$val\n" while defined($val = pop(@ary));
1052 die "Can't readlink $sym: $!"
1053 unless defined($value = readlink $sym);
a0d0e21e 1054 sub foo { defined &$bar ? &$bar(@_) : die "No bar"; }
2f9daede 1055 $debugging = 0 unless defined $debugging;
a0d0e21e 1056
19799a22 1057Note: Many folks tend to overuse C<defined>, and then are surprised to
7660c0ab 1058discover that the number C<0> and C<""> (the zero-length string) are, in fact,
2f9daede 1059defined values. For example, if you say
a5f75d66
AD
1060
1061 "ab" =~ /a(.*)b/;
1062
7660c0ab 1063The pattern match succeeds, and C<$1> is defined, despite the fact that it
cf264981 1064matched "nothing". It didn't really fail to match anything. Rather, it
2b5ab1e7 1065matched something that happened to be zero characters long. This is all
a5f75d66 1066very above-board and honest. When a function returns an undefined value,
2f9daede 1067it's an admission that it couldn't give you an honest answer. So you
19799a22 1068should use C<defined> only when you're questioning the integrity of what
7660c0ab 1069you're trying to do. At other times, a simple comparison to C<0> or C<""> is
2f9daede
TPG
1070what you want.
1071
dc848c6f 1072See also L</undef>, L</exists>, L</ref>.
2f9daede 1073
a0d0e21e
LW
1074=item delete EXPR
1075
01020589
GS
1076Given an expression that specifies a hash element, array element, hash slice,
1077or array slice, deletes the specified element(s) from the hash or array.
8216c1fd 1078In the case of an array, if the array elements happen to be at the end,
b76cc8ba 1079the size of the array will shrink to the highest element that tests
8216c1fd 1080true for exists() (or 0 if no such element exists).
a0d0e21e 1081
eba0920a
EM
1082Returns a list with the same number of elements as the number of elements
1083for which deletion was attempted. Each element of that list consists of
1084either the value of the element deleted, or the undefined value. In scalar
1085context, this means that you get the value of the last element deleted (or
1086the undefined value if that element did not exist).
1087
1088 %hash = (foo => 11, bar => 22, baz => 33);
1089 $scalar = delete $hash{foo}; # $scalar is 11
1090 $scalar = delete @hash{qw(foo bar)}; # $scalar is 22
1091 @array = delete @hash{qw(foo bar baz)}; # @array is (undef,undef,33)
1092
1093Deleting from C<%ENV> modifies the environment. Deleting from
01020589
GS
1094a hash tied to a DBM file deletes the entry from the DBM file. Deleting
1095from a C<tie>d hash or array may not necessarily return anything.
1096
8ea97a1e
GS
1097Deleting an array element effectively returns that position of the array
1098to its initial, uninitialized state. Subsequently testing for the same
cf264981
SP
1099element with exists() will return false. Also, deleting array elements
1100in the middle of an array will not shift the index of the elements
1101after them down. Use splice() for that. See L</exists>.
8ea97a1e 1102
01020589 1103The following (inefficiently) deletes all the values of %HASH and @ARRAY:
a0d0e21e 1104
5f05dabc
PP
1105 foreach $key (keys %HASH) {
1106 delete $HASH{$key};
a0d0e21e
LW
1107 }
1108
01020589
GS
1109 foreach $index (0 .. $#ARRAY) {
1110 delete $ARRAY[$index];
1111 }
1112
1113And so do these:
5f05dabc 1114
01020589
GS
1115 delete @HASH{keys %HASH};
1116
9740c838 1117 delete @ARRAY[0 .. $#ARRAY];
5f05dabc 1118
2b5ab1e7 1119But both of these are slower than just assigning the empty list
01020589
GS
1120or undefining %HASH or @ARRAY:
1121
1122 %HASH = (); # completely empty %HASH
1123 undef %HASH; # forget %HASH ever existed
2b5ab1e7 1124
01020589
GS
1125 @ARRAY = (); # completely empty @ARRAY
1126 undef @ARRAY; # forget @ARRAY ever existed
2b5ab1e7
TC
1127
1128Note that the EXPR can be arbitrarily complicated as long as the final
01020589
GS
1129operation is a hash element, array element, hash slice, or array slice
1130lookup:
a0d0e21e
LW
1131
1132 delete $ref->[$x][$y]{$key};
5f05dabc 1133 delete @{$ref->[$x][$y]}{$key1, $key2, @morekeys};
a0d0e21e 1134
01020589
GS
1135 delete $ref->[$x][$y][$index];
1136 delete @{$ref->[$x][$y]}[$index1, $index2, @moreindices];
1137
a0d0e21e
LW
1138=item die LIST
1139
19799a22
GS
1140Outside an C<eval>, prints the value of LIST to C<STDERR> and
1141exits with the current value of C<$!> (errno). If C<$!> is C<0>,
61eff3bc
JH
1142exits with the value of C<<< ($? >> 8) >>> (backtick `command`
1143status). If C<<< ($? >> 8) >>> is C<0>, exits with C<255>. Inside
19799a22
GS
1144an C<eval(),> the error message is stuffed into C<$@> and the
1145C<eval> is terminated with the undefined value. This makes
1146C<die> the way to raise an exception.
a0d0e21e
LW
1147
1148Equivalent examples:
1149
1150 die "Can't cd to spool: $!\n" unless chdir '/usr/spool/news';
54310121 1151 chdir '/usr/spool/news' or die "Can't cd to spool: $!\n"
a0d0e21e 1152
ccac6780 1153If the last element of LIST does not end in a newline, the current
df37ec69
WW
1154script line number and input line number (if any) are also printed,
1155and a newline is supplied. Note that the "input line number" (also
1156known as "chunk") is subject to whatever notion of "line" happens to
1157be currently in effect, and is also available as the special variable
1158C<$.>. See L<perlvar/"$/"> and L<perlvar/"$.">.
1159
1160Hint: sometimes appending C<", stopped"> to your message will cause it
1161to make better sense when the string C<"at foo line 123"> is appended.
1162Suppose you are running script "canasta".
a0d0e21e
LW
1163
1164 die "/etc/games is no good";
1165 die "/etc/games is no good, stopped";
1166
1167produce, respectively
1168
1169 /etc/games is no good at canasta line 123.
1170 /etc/games is no good, stopped at canasta line 123.
1171
2b5ab1e7 1172See also exit(), warn(), and the Carp module.
a0d0e21e 1173
7660c0ab
A
1174If LIST is empty and C<$@> already contains a value (typically from a
1175previous eval) that value is reused after appending C<"\t...propagated">.
fb73857a
PP
1176This is useful for propagating exceptions:
1177
1178 eval { ... };
1179 die unless $@ =~ /Expected exception/;
1180
ad216e65
JH
1181If LIST is empty and C<$@> contains an object reference that has a
1182C<PROPAGATE> method, that method will be called with additional file
1183and line number parameters. The return value replaces the value in
28a5cf3b 1184C<$@>. i.e. as if C<< $@ = eval { $@->PROPAGATE(__FILE__, __LINE__) }; >>
ad216e65
JH
1185were called.
1186
7660c0ab 1187If C<$@> is empty then the string C<"Died"> is used.
fb73857a 1188
52531d10
GS
1189die() can also be called with a reference argument. If this happens to be
1190trapped within an eval(), $@ contains the reference. This behavior permits
1191a more elaborate exception handling implementation using objects that
4375e838 1192maintain arbitrary state about the nature of the exception. Such a scheme
52531d10
GS
1193is sometimes preferable to matching particular string values of $@ using
1194regular expressions. Here's an example:
1195
da279afe 1196 use Scalar::Util 'blessed';
1197
52531d10
GS
1198 eval { ... ; die Some::Module::Exception->new( FOO => "bar" ) };
1199 if ($@) {
da279afe 1200 if (blessed($@) && $@->isa("Some::Module::Exception")) {
52531d10
GS
1201 # handle Some::Module::Exception
1202 }
1203 else {
1204 # handle all other possible exceptions
1205 }
1206 }
1207
19799a22 1208Because perl will stringify uncaught exception messages before displaying
52531d10
GS
1209them, you may want to overload stringification operations on such custom
1210exception objects. See L<overload> for details about that.
1211
19799a22
GS
1212You can arrange for a callback to be run just before the C<die>
1213does its deed, by setting the C<$SIG{__DIE__}> hook. The associated
1214handler will be called with the error text and can change the error
1215message, if it sees fit, by calling C<die> again. See
1216L<perlvar/$SIG{expr}> for details on setting C<%SIG> entries, and
cf264981 1217L<"eval BLOCK"> for some examples. Although this feature was
19799a22
GS
1218to be run only right before your program was to exit, this is not
1219currently the case--the C<$SIG{__DIE__}> hook is currently called
1220even inside eval()ed blocks/strings! If one wants the hook to do
1221nothing in such situations, put
fb73857a
PP
1222
1223 die @_ if $^S;
1224
19799a22
GS
1225as the first line of the handler (see L<perlvar/$^S>). Because
1226this promotes strange action at a distance, this counterintuitive
b76cc8ba 1227behavior may be fixed in a future release.
774d564b 1228
a0d0e21e
LW
1229=item do BLOCK
1230
1231Not really a function. Returns the value of the last command in the
6b275a1f
RGS
1232sequence of commands indicated by BLOCK. When modified by the C<while> or
1233C<until> loop modifier, executes the BLOCK once before testing the loop
1234condition. (On other statements the loop modifiers test the conditional
1235first.)
a0d0e21e 1236
4968c1e4 1237C<do BLOCK> does I<not> count as a loop, so the loop control statements
2b5ab1e7
TC
1238C<next>, C<last>, or C<redo> cannot be used to leave or restart the block.
1239See L<perlsyn> for alternative strategies.
4968c1e4 1240
a0d0e21e
LW
1241=item do SUBROUTINE(LIST)
1242
cf264981 1243This form of subroutine call is deprecated. See L<perlsub>.
a0d0e21e
LW
1244
1245=item do EXPR
1246
1247Uses the value of EXPR as a filename and executes the contents of the
ea63ef19 1248file as a Perl script.
a0d0e21e
LW
1249
1250 do 'stat.pl';
1251
1252is just like
1253
986b19de 1254 eval `cat stat.pl`;
a0d0e21e 1255
2b5ab1e7 1256except that it's more efficient and concise, keeps track of the current
ea63ef19 1257filename for error messages, searches the @INC directories, and updates
2b5ab1e7
TC
1258C<%INC> if the file is found. See L<perlvar/Predefined Names> for these
1259variables. It also differs in that code evaluated with C<do FILENAME>
1260cannot see lexicals in the enclosing scope; C<eval STRING> does. It's the
1261same, however, in that it does reparse the file every time you call it,
1262so you probably don't want to do this inside a loop.
a0d0e21e 1263
8e30cc93 1264If C<do> cannot read the file, it returns undef and sets C<$!> to the
2b5ab1e7 1265error. If C<do> can read the file but cannot compile it, it
8e30cc93
G
1266returns undef and sets an error message in C<$@>. If the file is
1267successfully compiled, C<do> returns the value of the last expression
1268evaluated.
1269
a0d0e21e 1270Note that inclusion of library modules is better done with the
19799a22 1271C<use> and C<require> operators, which also do automatic error checking
4633a7c4 1272and raise an exception if there's a problem.
a0d0e21e 1273
5a964f20
TC
1274You might like to use C<do> to read in a program configuration
1275file. Manual error checking can be done this way:
1276
b76cc8ba 1277 # read in config files: system first, then user
f86cebdf 1278 for $file ("/share/prog/defaults.rc",
b76cc8ba 1279 "$ENV{HOME}/.someprogrc")
2b5ab1e7 1280 {
5a964f20 1281 unless ($return = do $file) {
f86cebdf
GS
1282 warn "couldn't parse $file: $@" if $@;
1283 warn "couldn't do $file: $!" unless defined $return;
1284 warn "couldn't run $file" unless $return;
5a964f20
TC
1285 }
1286 }
1287
a0d0e21e
LW
1288=item dump LABEL
1289
1614b0e3
JD
1290=item dump
1291
19799a22
GS
1292This function causes an immediate core dump. See also the B<-u>
1293command-line switch in L<perlrun>, which does the same thing.
1294Primarily this is so that you can use the B<undump> program (not
1295supplied) to turn your core dump into an executable binary after
1296having initialized all your variables at the beginning of the
1297program. When the new binary is executed it will begin by executing
1298a C<goto LABEL> (with all the restrictions that C<goto> suffers).
1299Think of it as a goto with an intervening core dump and reincarnation.
1300If C<LABEL> is omitted, restarts the program from the top.
1301
1302B<WARNING>: Any files opened at the time of the dump will I<not>
1303be open any more when the program is reincarnated, with possible
b76cc8ba 1304resulting confusion on the part of Perl.
19799a22
GS
1305
1306This function is now largely obsolete, partly because it's very
1307hard to convert a core file into an executable, and because the
1308real compiler backends for generating portable bytecode and compilable
ac206dc8
RGS
1309C code have superseded it. That's why you should now invoke it as
1310C<CORE::dump()>, if you don't want to be warned against a possible
1311typo.
19799a22
GS
1312
1313If you're looking to use L<dump> to speed up your program, consider
1314generating bytecode or native C code as described in L<perlcc>. If
1315you're just trying to accelerate a CGI script, consider using the
210b36aa 1316C<mod_perl> extension to B<Apache>, or the CPAN module, CGI::Fast.
19799a22 1317You might also consider autoloading or selfloading, which at least
b76cc8ba 1318make your program I<appear> to run faster.
5a964f20 1319
aa689395
PP
1320=item each HASH
1321
5a964f20 1322When called in list context, returns a 2-element list consisting of the
aa689395 1323key and value for the next element of a hash, so that you can iterate over
74fc8b5f 1324it. When called in scalar context, returns only the key for the next
e902a979 1325element in the hash.
2f9daede 1326
ab192400 1327Entries are returned in an apparently random order. The actual random
504f80c1
JH
1328order is subject to change in future versions of perl, but it is
1329guaranteed to be in the same order as either the C<keys> or C<values>
4546b9e6
JH
1330function would produce on the same (unmodified) hash. Since Perl
13315.8.1 the ordering is different even between different runs of Perl
1332for security reasons (see L<perlsec/"Algorithmic Complexity Attacks">).
ab192400
GS
1333
1334When the hash is entirely read, a null array is returned in list context
19799a22
GS
1335(which when assigned produces a false (C<0>) value), and C<undef> in
1336scalar context. The next call to C<each> after that will start iterating
1337again. There is a single iterator for each hash, shared by all C<each>,
1338C<keys>, and C<values> function calls in the program; it can be reset by
2f9daede
TPG
1339reading all the elements from the hash, or by evaluating C<keys HASH> or
1340C<values HASH>. If you add or delete elements of a hash while you're
74fc8b5f
MJD
1341iterating over it, you may get entries skipped or duplicated, so
1342don't. Exception: It is always safe to delete the item most recently
1343returned by C<each()>, which means that the following code will work:
1344
1345 while (($key, $value) = each %hash) {
1346 print $key, "\n";
1347 delete $hash{$key}; # This is safe
1348 }
aa689395 1349
f86cebdf 1350The following prints out your environment like the printenv(1) program,
aa689395 1351only in a different order:
a0d0e21e
LW
1352
1353 while (($key,$value) = each %ENV) {
1354 print "$key=$value\n";
1355 }
1356
19799a22 1357See also C<keys>, C<values> and C<sort>.
a0d0e21e
LW
1358
1359=item eof FILEHANDLE
1360
4633a7c4
LW
1361=item eof ()
1362
a0d0e21e
LW
1363=item eof
1364
1365Returns 1 if the next read on FILEHANDLE will return end of file, or if
1366FILEHANDLE is not open. FILEHANDLE may be an expression whose value
5a964f20 1367gives the real filehandle. (Note that this function actually
19799a22 1368reads a character and then C<ungetc>s it, so isn't very useful in an
748a9306 1369interactive context.) Do not read from a terminal file (or call
19799a22 1370C<eof(FILEHANDLE)> on it) after end-of-file is reached. File types such
748a9306
LW
1371as terminals may lose the end-of-file condition if you do.
1372
820475bd
GS
1373An C<eof> without an argument uses the last file read. Using C<eof()>
1374with empty parentheses is very different. It refers to the pseudo file
1375formed from the files listed on the command line and accessed via the
61eff3bc
JH
1376C<< <> >> operator. Since C<< <> >> isn't explicitly opened,
1377as a normal filehandle is, an C<eof()> before C<< <> >> has been
820475bd 1378used will cause C<@ARGV> to be examined to determine if input is
67408cae 1379available. Similarly, an C<eof()> after C<< <> >> has returned
efdd0218
RB
1380end-of-file will assume you are processing another C<@ARGV> list,
1381and if you haven't set C<@ARGV>, will read input from C<STDIN>;
1382see L<perlop/"I/O Operators">.
820475bd 1383
61eff3bc 1384In a C<< while (<>) >> loop, C<eof> or C<eof(ARGV)> can be used to
820475bd
GS
1385detect the end of each file, C<eof()> will only detect the end of the
1386last file. Examples:
a0d0e21e 1387
748a9306
LW
1388 # reset line numbering on each input file
1389 while (<>) {
b76cc8ba 1390 next if /^\s*#/; # skip comments
748a9306 1391 print "$.\t$_";
5a964f20
TC
1392 } continue {
1393 close ARGV if eof; # Not eof()!
748a9306
LW
1394 }
1395
a0d0e21e
LW
1396 # insert dashes just before last line of last file
1397 while (<>) {
6ac88b13 1398 if (eof()) { # check for end of last file
a0d0e21e
LW
1399 print "--------------\n";
1400 }
1401 print;
6ac88b13 1402 last if eof(); # needed if we're reading from a terminal
a0d0e21e
LW
1403 }
1404
a0d0e21e 1405Practical hint: you almost never need to use C<eof> in Perl, because the
3ce0d271
GS
1406input operators typically return C<undef> when they run out of data, or if
1407there was an error.
a0d0e21e
LW
1408
1409=item eval EXPR
1410
1411=item eval BLOCK
1412
ce2984c3
PF
1413=item eval
1414
c7cc6f1c
GS
1415In the first form, the return value of EXPR is parsed and executed as if it
1416were a little Perl program. The value of the expression (which is itself
5a964f20 1417determined within scalar context) is first parsed, and if there weren't any
be3174d2
GS
1418errors, executed in the lexical context of the current Perl program, so
1419that any variable settings or subroutine and format definitions remain
cf264981 1420afterwards. Note that the value is parsed every time the C<eval> executes.
be3174d2
GS
1421If EXPR is omitted, evaluates C<$_>. This form is typically used to
1422delay parsing and subsequent execution of the text of EXPR until run time.
c7cc6f1c
GS
1423
1424In the second form, the code within the BLOCK is parsed only once--at the
cf264981 1425same time the code surrounding the C<eval> itself was parsed--and executed
c7cc6f1c
GS
1426within the context of the current Perl program. This form is typically
1427used to trap exceptions more efficiently than the first (see below), while
1428also providing the benefit of checking the code within BLOCK at compile
1429time.
1430
1431The final semicolon, if any, may be omitted from the value of EXPR or within
1432the BLOCK.
1433
1434In both forms, the value returned is the value of the last expression
5a964f20 1435evaluated inside the mini-program; a return statement may be also used, just
c7cc6f1c 1436as with subroutines. The expression providing the return value is evaluated
cf264981
SP
1437in void, scalar, or list context, depending on the context of the C<eval>
1438itself. See L</wantarray> for more on how the evaluation context can be
1439determined.
a0d0e21e 1440
19799a22
GS
1441If there is a syntax error or runtime error, or a C<die> statement is
1442executed, an undefined value is returned by C<eval>, and C<$@> is set to the
a0d0e21e 1443error message. If there was no error, C<$@> is guaranteed to be a null
19799a22 1444string. Beware that using C<eval> neither silences perl from printing
c7cc6f1c 1445warnings to STDERR, nor does it stuff the text of warning messages into C<$@>.
d9984052
A
1446To do either of those, you have to use the C<$SIG{__WARN__}> facility, or
1447turn off warnings inside the BLOCK or EXPR using S<C<no warnings 'all'>>.
1448See L</warn>, L<perlvar>, L<warnings> and L<perllexwarn>.
a0d0e21e 1449
19799a22
GS
1450Note that, because C<eval> traps otherwise-fatal errors, it is useful for
1451determining whether a particular feature (such as C<socket> or C<symlink>)
a0d0e21e
LW
1452is implemented. It is also Perl's exception trapping mechanism, where
1453the die operator is used to raise exceptions.
1454
1455If the code to be executed doesn't vary, you may use the eval-BLOCK
1456form to trap run-time errors without incurring the penalty of
1457recompiling each time. The error, if any, is still returned in C<$@>.
1458Examples:
1459
54310121 1460 # make divide-by-zero nonfatal
a0d0e21e
LW
1461 eval { $answer = $a / $b; }; warn $@ if $@;
1462
1463 # same thing, but less efficient
1464 eval '$answer = $a / $b'; warn $@ if $@;
1465
1466 # a compile-time error
5a964f20 1467 eval { $answer = }; # WRONG
a0d0e21e
LW
1468
1469 # a run-time error
1470 eval '$answer ='; # sets $@
1471
cf264981
SP
1472Using the C<eval{}> form as an exception trap in libraries does have some
1473issues. Due to the current arguably broken state of C<__DIE__> hooks, you
1474may wish not to trigger any C<__DIE__> hooks that user code may have installed.
2b5ab1e7
TC
1475You can use the C<local $SIG{__DIE__}> construct for this purpose,
1476as shown in this example:
774d564b
PP
1477
1478 # a very private exception trap for divide-by-zero
f86cebdf
GS
1479 eval { local $SIG{'__DIE__'}; $answer = $a / $b; };
1480 warn $@ if $@;
774d564b
PP
1481
1482This is especially significant, given that C<__DIE__> hooks can call
19799a22 1483C<die> again, which has the effect of changing their error messages:
774d564b
PP
1484
1485 # __DIE__ hooks may modify error messages
1486 {
f86cebdf
GS
1487 local $SIG{'__DIE__'} =
1488 sub { (my $x = $_[0]) =~ s/foo/bar/g; die $x };
c7cc6f1c
GS
1489 eval { die "foo lives here" };
1490 print $@ if $@; # prints "bar lives here"
774d564b
PP
1491 }
1492
19799a22 1493Because this promotes action at a distance, this counterintuitive behavior
2b5ab1e7
TC
1494may be fixed in a future release.
1495
19799a22 1496With an C<eval>, you should be especially careful to remember what's
a0d0e21e
LW
1497being looked at when:
1498
1499 eval $x; # CASE 1
1500 eval "$x"; # CASE 2
1501
1502 eval '$x'; # CASE 3
1503 eval { $x }; # CASE 4
1504
5a964f20 1505 eval "\$$x++"; # CASE 5
a0d0e21e
LW
1506 $$x++; # CASE 6
1507
2f9daede 1508Cases 1 and 2 above behave identically: they run the code contained in
19799a22 1509the variable $x. (Although case 2 has misleading double quotes making
2f9daede 1510the reader wonder what else might be happening (nothing is).) Cases 3
7660c0ab 1511and 4 likewise behave in the same way: they run the code C<'$x'>, which
19799a22 1512does nothing but return the value of $x. (Case 4 is preferred for
2f9daede
TPG
1513purely visual reasons, but it also has the advantage of compiling at
1514compile-time instead of at run-time.) Case 5 is a place where
19799a22 1515normally you I<would> like to use double quotes, except that in this
2f9daede
TPG
1516particular situation, you can just use symbolic references instead, as
1517in case 6.
a0d0e21e 1518
4968c1e4 1519C<eval BLOCK> does I<not> count as a loop, so the loop control statements
2b5ab1e7 1520C<next>, C<last>, or C<redo> cannot be used to leave or restart the block.
4968c1e4 1521
d819b83a
DM
1522Note that as a very special case, an C<eval ''> executed within the C<DB>
1523package doesn't see the usual surrounding lexical scope, but rather the
1524scope of the first non-DB piece of code that called it. You don't normally
1525need to worry about this unless you are writing a Perl debugger.
1526
a0d0e21e
LW
1527=item exec LIST
1528
8bf3b016
GS
1529=item exec PROGRAM LIST
1530
19799a22
GS
1531The C<exec> function executes a system command I<and never returns>--
1532use C<system> instead of C<exec> if you want it to return. It fails and
1533returns false only if the command does not exist I<and> it is executed
fb73857a 1534directly instead of via your system's command shell (see below).
a0d0e21e 1535
19799a22
GS
1536Since it's a common mistake to use C<exec> instead of C<system>, Perl
1537warns you if there is a following statement which isn't C<die>, C<warn>,
1538or C<exit> (if C<-w> is set - but you always do that). If you
1539I<really> want to follow an C<exec> with some other statement, you
55d729e4
GS
1540can use one of these styles to avoid the warning:
1541
5a964f20
TC
1542 exec ('foo') or print STDERR "couldn't exec foo: $!";
1543 { exec ('foo') }; print STDERR "couldn't exec foo: $!";
55d729e4 1544
5a964f20 1545If there is more than one argument in LIST, or if LIST is an array
f86cebdf 1546with more than one value, calls execvp(3) with the arguments in LIST.
5a964f20
TC
1547If there is only one scalar argument or an array with one element in it,
1548the argument is checked for shell metacharacters, and if there are any,
1549the entire argument is passed to the system's command shell for parsing
1550(this is C</bin/sh -c> on Unix platforms, but varies on other platforms).
1551If there are no shell metacharacters in the argument, it is split into
b76cc8ba 1552words and passed directly to C<execvp>, which is more efficient.
19799a22 1553Examples:
a0d0e21e 1554
19799a22
GS
1555 exec '/bin/echo', 'Your arguments are: ', @ARGV;
1556 exec "sort $outfile | uniq";
a0d0e21e
LW
1557
1558If you don't really want to execute the first argument, but want to lie
1559to the program you are executing about its own name, you can specify
1560the program you actually want to run as an "indirect object" (without a
1561comma) in front of the LIST. (This always forces interpretation of the
54310121 1562LIST as a multivalued list, even if there is only a single scalar in
a0d0e21e
LW
1563the list.) Example:
1564
1565 $shell = '/bin/csh';
1566 exec $shell '-sh'; # pretend it's a login shell
1567
1568or, more directly,
1569
1570 exec {'/bin/csh'} '-sh'; # pretend it's a login shell
1571
bb32b41a
GS
1572When the arguments get executed via the system shell, results will
1573be subject to its quirks and capabilities. See L<perlop/"`STRING`">
1574for details.
1575
19799a22
GS
1576Using an indirect object with C<exec> or C<system> is also more
1577secure. This usage (which also works fine with system()) forces
1578interpretation of the arguments as a multivalued list, even if the
1579list had just one argument. That way you're safe from the shell
1580expanding wildcards or splitting up words with whitespace in them.
5a964f20
TC
1581
1582 @args = ( "echo surprise" );
1583
2b5ab1e7 1584 exec @args; # subject to shell escapes
f86cebdf 1585 # if @args == 1
2b5ab1e7 1586 exec { $args[0] } @args; # safe even with one-arg list
5a964f20
TC
1587
1588The first version, the one without the indirect object, ran the I<echo>
1589program, passing it C<"surprise"> an argument. The second version
1590didn't--it tried to run a program literally called I<"echo surprise">,
1591didn't find it, and set C<$?> to a non-zero value indicating failure.
1592
0f897271
GS
1593Beginning with v5.6.0, Perl will attempt to flush all files opened for
1594output before the exec, but this may not be supported on some platforms
1595(see L<perlport>). To be safe, you may need to set C<$|> ($AUTOFLUSH
1596in English) or call the C<autoflush()> method of C<IO::Handle> on any
1597open handles in order to avoid lost output.
1598
19799a22 1599Note that C<exec> will not call your C<END> blocks, nor will it call
7660c0ab
A
1600any C<DESTROY> methods in your objects.
1601
a0d0e21e
LW
1602=item exists EXPR
1603
01020589 1604Given an expression that specifies a hash element or array element,
8ea97a1e
GS
1605returns true if the specified element in the hash or array has ever
1606been initialized, even if the corresponding value is undefined. The
1607element is not autovivified if it doesn't exist.
a0d0e21e 1608
01020589
GS
1609 print "Exists\n" if exists $hash{$key};
1610 print "Defined\n" if defined $hash{$key};
1611 print "True\n" if $hash{$key};
1612
1613 print "Exists\n" if exists $array[$index];
1614 print "Defined\n" if defined $array[$index];
1615 print "True\n" if $array[$index];
a0d0e21e 1616
8ea97a1e 1617A hash or array element can be true only if it's defined, and defined if
a0d0e21e
LW
1618it exists, but the reverse doesn't necessarily hold true.
1619
afebc493
GS
1620Given an expression that specifies the name of a subroutine,
1621returns true if the specified subroutine has ever been declared, even
1622if it is undefined. Mentioning a subroutine name for exists or defined
847c7ebe
DD
1623does not count as declaring it. Note that a subroutine which does not
1624exist may still be callable: its package may have an C<AUTOLOAD>
1625method that makes it spring into existence the first time that it is
1626called -- see L<perlsub>.
afebc493
GS
1627
1628 print "Exists\n" if exists &subroutine;
1629 print "Defined\n" if defined &subroutine;
1630
a0d0e21e 1631Note that the EXPR can be arbitrarily complicated as long as the final
afebc493 1632operation is a hash or array key lookup or subroutine name:
a0d0e21e 1633
2b5ab1e7
TC
1634 if (exists $ref->{A}->{B}->{$key}) { }
1635 if (exists $hash{A}{B}{$key}) { }
1636
01020589
GS
1637 if (exists $ref->{A}->{B}->[$ix]) { }
1638 if (exists $hash{A}{B}[$ix]) { }
1639
afebc493
GS
1640 if (exists &{$ref->{A}{B}{$key}}) { }
1641
01020589
GS
1642Although the deepest nested array or hash will not spring into existence
1643just because its existence was tested, any intervening ones will.
61eff3bc 1644Thus C<< $ref->{"A"} >> and C<< $ref->{"A"}->{"B"} >> will spring
01020589
GS
1645into existence due to the existence test for the $key element above.
1646This happens anywhere the arrow operator is used, including even:
5a964f20 1647
2b5ab1e7
TC
1648 undef $ref;
1649 if (exists $ref->{"Some key"}) { }
1650 print $ref; # prints HASH(0x80d3d5c)
1651
1652This surprising autovivification in what does not at first--or even
1653second--glance appear to be an lvalue context may be fixed in a future
5a964f20 1654release.
a0d0e21e 1655
afebc493
GS
1656Use of a subroutine call, rather than a subroutine name, as an argument
1657to exists() is an error.
1658
1659 exists &sub; # OK
1660 exists &sub(); # Error
1661
a0d0e21e
LW
1662=item exit EXPR
1663
ce2984c3
PF
1664=item exit
1665
2b5ab1e7 1666Evaluates EXPR and exits immediately with that value. Example:
a0d0e21e
LW
1667
1668 $ans = <STDIN>;
1669 exit 0 if $ans =~ /^[Xx]/;
1670
19799a22 1671See also C<die>. If EXPR is omitted, exits with C<0> status. The only
2b5ab1e7
TC
1672universally recognized values for EXPR are C<0> for success and C<1>
1673for error; other values are subject to interpretation depending on the
1674environment in which the Perl program is running. For example, exiting
167569 (EX_UNAVAILABLE) from a I<sendmail> incoming-mail filter will cause
1676the mailer to return the item undelivered, but that's not true everywhere.
a0d0e21e 1677
19799a22
GS
1678Don't use C<exit> to abort a subroutine if there's any chance that
1679someone might want to trap whatever error happened. Use C<die> instead,
1680which can be trapped by an C<eval>.
28757baa 1681
19799a22 1682The exit() function does not always exit immediately. It calls any
2b5ab1e7 1683defined C<END> routines first, but these C<END> routines may not
19799a22 1684themselves abort the exit. Likewise any object destructors that need to
2b5ab1e7
TC
1685be called are called before the real exit. If this is a problem, you
1686can call C<POSIX:_exit($status)> to avoid END and destructor processing.
87275199 1687See L<perlmod> for details.
5a964f20 1688
a0d0e21e
LW
1689=item exp EXPR
1690
54310121 1691=item exp
bbce6d69 1692
b76cc8ba 1693Returns I<e> (the natural logarithm base) to the power of EXPR.
a0d0e21e
LW
1694If EXPR is omitted, gives C<exp($_)>.
1695
1696=item fcntl FILEHANDLE,FUNCTION,SCALAR
1697
f86cebdf 1698Implements the fcntl(2) function. You'll probably have to say
a0d0e21e
LW
1699
1700 use Fcntl;
1701
0ade1984 1702first to get the correct constant definitions. Argument processing and
b76cc8ba 1703value return works just like C<ioctl> below.
a0d0e21e
LW
1704For example:
1705
1706 use Fcntl;
5a964f20
TC
1707 fcntl($filehandle, F_GETFL, $packed_return_buffer)
1708 or die "can't fcntl F_GETFL: $!";
1709
554ad1fc 1710You don't have to check for C<defined> on the return from C<fcntl>.
951ba7fe
GS
1711Like C<ioctl>, it maps a C<0> return from the system call into
1712C<"0 but true"> in Perl. This string is true in boolean context and C<0>
2b5ab1e7
TC
1713in numeric context. It is also exempt from the normal B<-w> warnings
1714on improper numeric conversions.
5a964f20 1715
19799a22 1716Note that C<fcntl> will produce a fatal error if used on a machine that
2b5ab1e7
TC
1717doesn't implement fcntl(2). See the Fcntl module or your fcntl(2)
1718manpage to learn what functions are available on your system.
a0d0e21e 1719
be2f7487 1720Here's an example of setting a filehandle named C<REMOTE> to be
1721non-blocking at the system level. You'll have to negotiate C<$|>
1722on your own, though.
1723
1724 use Fcntl qw(F_GETFL F_SETFL O_NONBLOCK);
1725
1726 $flags = fcntl(REMOTE, F_GETFL, 0)
1727 or die "Can't get flags for the socket: $!\n";
1728
1729 $flags = fcntl(REMOTE, F_SETFL, $flags | O_NONBLOCK)
1730 or die "Can't set flags for the socket: $!\n";
1731
a0d0e21e
LW
1732=item fileno FILEHANDLE
1733
2b5ab1e7
TC
1734Returns the file descriptor for a filehandle, or undefined if the
1735filehandle is not open. This is mainly useful for constructing
19799a22 1736bitmaps for C<select> and low-level POSIX tty-handling operations.
2b5ab1e7
TC
1737If FILEHANDLE is an expression, the value is taken as an indirect
1738filehandle, generally its name.
5a964f20 1739
b76cc8ba 1740You can use this to find out whether two handles refer to the
5a964f20
TC
1741same underlying descriptor:
1742
1743 if (fileno(THIS) == fileno(THAT)) {
1744 print "THIS and THAT are dups\n";
b76cc8ba
NIS
1745 }
1746
1747(Filehandles connected to memory objects via new features of C<open> may
1748return undefined even though they are open.)
1749
a0d0e21e
LW
1750
1751=item flock FILEHANDLE,OPERATION
1752
19799a22
GS
1753Calls flock(2), or an emulation of it, on FILEHANDLE. Returns true
1754for success, false on failure. Produces a fatal error if used on a
2b5ab1e7 1755machine that doesn't implement flock(2), fcntl(2) locking, or lockf(3).
19799a22 1756C<flock> is Perl's portable file locking interface, although it locks
2b5ab1e7
TC
1757only entire files, not records.
1758
1759Two potentially non-obvious but traditional C<flock> semantics are
1760that it waits indefinitely until the lock is granted, and that its locks
1761B<merely advisory>. Such discretionary locks are more flexible, but offer
cf264981
SP
1762fewer guarantees. This means that programs that do not also use C<flock>
1763may modify files locked with C<flock>. See L<perlport>,
2b5ab1e7
TC
1764your port's specific documentation, or your system-specific local manpages
1765for details. It's best to assume traditional behavior if you're writing
1766portable programs. (But if you're not, you should as always feel perfectly
1767free to write for your own system's idiosyncrasies (sometimes called
1768"features"). Slavish adherence to portability concerns shouldn't get
1769in the way of your getting your job done.)
a3cb178b 1770
8ebc5c01
PP
1771OPERATION is one of LOCK_SH, LOCK_EX, or LOCK_UN, possibly combined with
1772LOCK_NB. These constants are traditionally valued 1, 2, 8 and 4, but
ea3105be 1773you can use the symbolic names if you import them from the Fcntl module,
68dc0745
PP
1774either individually, or as a group using the ':flock' tag. LOCK_SH
1775requests a shared lock, LOCK_EX requests an exclusive lock, and LOCK_UN
ea3105be
GS
1776releases a previously requested lock. If LOCK_NB is bitwise-or'ed with
1777LOCK_SH or LOCK_EX then C<flock> will return immediately rather than blocking
68dc0745
PP
1778waiting for the lock (check the return status to see if you got it).
1779
2b5ab1e7
TC
1780To avoid the possibility of miscoordination, Perl now flushes FILEHANDLE
1781before locking or unlocking it.
8ebc5c01 1782
f86cebdf 1783Note that the emulation built with lockf(3) doesn't provide shared
8ebc5c01 1784locks, and it requires that FILEHANDLE be open with write intent. These
2b5ab1e7 1785are the semantics that lockf(3) implements. Most if not all systems
f86cebdf 1786implement lockf(3) in terms of fcntl(2) locking, though, so the
8ebc5c01
PP
1787differing semantics shouldn't bite too many people.
1788
becacb53
TM
1789Note that the fcntl(2) emulation of flock(3) requires that FILEHANDLE
1790be open with read intent to use LOCK_SH and requires that it be open
1791with write intent to use LOCK_EX.
1792
19799a22
GS
1793Note also that some versions of C<flock> cannot lock things over the
1794network; you would need to use the more system-specific C<fcntl> for
f86cebdf
GS
1795that. If you like you can force Perl to ignore your system's flock(2)
1796function, and so provide its own fcntl(2)-based emulation, by passing
8ebc5c01
PP
1797the switch C<-Ud_flock> to the F<Configure> program when you configure
1798perl.
4633a7c4
LW
1799
1800Here's a mailbox appender for BSD systems.
a0d0e21e 1801
7e1af8bc 1802 use Fcntl ':flock'; # import LOCK_* constants
a0d0e21e
LW
1803
1804 sub lock {
7e1af8bc 1805 flock(MBOX,LOCK_EX);
a0d0e21e
LW
1806 # and, in case someone appended
1807 # while we were waiting...
1808 seek(MBOX, 0, 2);
1809 }
1810
1811 sub unlock {
7e1af8bc 1812 flock(MBOX,LOCK_UN);
a0d0e21e
LW
1813 }
1814
1815 open(MBOX, ">>/usr/spool/mail/$ENV{'USER'}")
1816 or die "Can't open mailbox: $!";
1817
1818 lock();
1819 print MBOX $msg,"\n\n";
1820 unlock();
1821
2b5ab1e7
TC
1822On systems that support a real flock(), locks are inherited across fork()
1823calls, whereas those that must resort to the more capricious fcntl()
1824function lose the locks, making it harder to write servers.
1825
cb1a09d0 1826See also L<DB_File> for other flock() examples.
a0d0e21e
LW
1827
1828=item fork
1829
2b5ab1e7
TC
1830Does a fork(2) system call to create a new process running the
1831same program at the same point. It returns the child pid to the
1832parent process, C<0> to the child process, or C<undef> if the fork is
1833unsuccessful. File descriptors (and sometimes locks on those descriptors)
1834are shared, while everything else is copied. On most systems supporting
1835fork(), great care has gone into making it extremely efficient (for
1836example, using copy-on-write technology on data pages), making it the
1837dominant paradigm for multitasking over the last few decades.
5a964f20 1838
0f897271
GS
1839Beginning with v5.6.0, Perl will attempt to flush all files opened for
1840output before forking the child process, but this may not be supported
1841on some platforms (see L<perlport>). To be safe, you may need to set
1842C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method of
1843C<IO::Handle> on any open handles in order to avoid duplicate output.
a0d0e21e 1844
19799a22 1845If you C<fork> without ever waiting on your children, you will
2b5ab1e7
TC
1846accumulate zombies. On some systems, you can avoid this by setting
1847C<$SIG{CHLD}> to C<"IGNORE">. See also L<perlipc> for more examples of
1848forking and reaping moribund children.
cb1a09d0 1849
28757baa
PP
1850Note that if your forked child inherits system file descriptors like
1851STDIN and STDOUT that are actually connected by a pipe or socket, even
2b5ab1e7 1852if you exit, then the remote server (such as, say, a CGI script or a
19799a22 1853backgrounded job launched from a remote shell) won't think you're done.
2b5ab1e7 1854You should reopen those to F</dev/null> if it's any issue.
28757baa 1855
cb1a09d0
AD
1856=item format
1857
19799a22 1858Declare a picture format for use by the C<write> function. For
cb1a09d0
AD
1859example:
1860
54310121 1861 format Something =
cb1a09d0
AD
1862 Test: @<<<<<<<< @||||| @>>>>>
1863 $str, $%, '$' . int($num)
1864 .
1865
1866 $str = "widget";
184e9718 1867 $num = $cost/$quantity;
cb1a09d0
AD
1868 $~ = 'Something';
1869 write;
1870
1871See L<perlform> for many details and examples.
1872
8903cb82 1873=item formline PICTURE,LIST
a0d0e21e 1874
5a964f20 1875This is an internal function used by C<format>s, though you may call it,
a0d0e21e
LW
1876too. It formats (see L<perlform>) a list of values according to the
1877contents of PICTURE, placing the output into the format output
7660c0ab 1878accumulator, C<$^A> (or C<$ACCUMULATOR> in English).
19799a22 1879Eventually, when a C<write> is done, the contents of
cf264981
SP
1880C<$^A> are written to some filehandle. You could also read C<$^A>
1881and then set C<$^A> back to C<"">. Note that a format typically
19799a22 1882does one C<formline> per line of form, but the C<formline> function itself
748a9306 1883doesn't care how many newlines are embedded in the PICTURE. This means
4633a7c4 1884that the C<~> and C<~~> tokens will treat the entire PICTURE as a single line.
748a9306
LW
1885You may therefore need to use multiple formlines to implement a single
1886record format, just like the format compiler.
1887
19799a22 1888Be careful if you put double quotes around the picture, because an C<@>
748a9306 1889character may be taken to mean the beginning of an array name.
19799a22 1890C<formline> always returns true. See L<perlform> for other examples.
a0d0e21e
LW
1891
1892=item getc FILEHANDLE
1893
1894=item getc
1895
1896Returns the next character from the input file attached to FILEHANDLE,
b5fe5ca2
SR
1897or the undefined value at end of file, or if there was an error (in
1898the latter case C<$!> is set). If FILEHANDLE is omitted, reads from
1899STDIN. This is not particularly efficient. However, it cannot be
1900used by itself to fetch single characters without waiting for the user
1901to hit enter. For that, try something more like:
4633a7c4
LW
1902
1903 if ($BSD_STYLE) {
1904 system "stty cbreak </dev/tty >/dev/tty 2>&1";
1905 }
1906 else {
54310121 1907 system "stty", '-icanon', 'eol', "\001";
4633a7c4
LW
1908 }
1909
1910 $key = getc(STDIN);
1911
1912 if ($BSD_STYLE) {
1913 system "stty -cbreak </dev/tty >/dev/tty 2>&1";
1914 }
1915 else {
5f05dabc 1916 system "stty", 'icanon', 'eol', '^@'; # ASCII null
4633a7c4
LW
1917 }
1918 print "\n";
1919
54310121
PP
1920Determination of whether $BSD_STYLE should be set
1921is left as an exercise to the reader.
cb1a09d0 1922
19799a22 1923The C<POSIX::getattr> function can do this more portably on
2b5ab1e7
TC
1924systems purporting POSIX compliance. See also the C<Term::ReadKey>
1925module from your nearest CPAN site; details on CPAN can be found on
1926L<perlmodlib/CPAN>.
a0d0e21e
LW
1927
1928=item getlogin
1929
cf264981 1930This implements the C library function of the same name, which on most
5a964f20 1931systems returns the current login from F</etc/utmp>, if any. If null,
19799a22 1932use C<getpwuid>.
a0d0e21e 1933
f86702cc 1934 $login = getlogin || getpwuid($<) || "Kilroy";
a0d0e21e 1935
19799a22
GS
1936Do not consider C<getlogin> for authentication: it is not as
1937secure as C<getpwuid>.
4633a7c4 1938
a0d0e21e
LW
1939=item getpeername SOCKET
1940
1941Returns the packed sockaddr address of other end of the SOCKET connection.
1942
4633a7c4
LW
1943 use Socket;
1944 $hersockaddr = getpeername(SOCK);
19799a22 1945 ($port, $iaddr) = sockaddr_in($hersockaddr);
4633a7c4
LW
1946 $herhostname = gethostbyaddr($iaddr, AF_INET);
1947 $herstraddr = inet_ntoa($iaddr);
a0d0e21e
LW
1948
1949=item getpgrp PID
1950
47e29363 1951Returns the current process group for the specified PID. Use
7660c0ab 1952a PID of C<0> to get the current process group for the
4633a7c4 1953current process. Will raise an exception if used on a machine that
f86cebdf 1954doesn't implement getpgrp(2). If PID is omitted, returns process
19799a22 1955group of current process. Note that the POSIX version of C<getpgrp>
7660c0ab 1956does not accept a PID argument, so only C<PID==0> is truly portable.
a0d0e21e
LW
1957
1958=item getppid
1959
1960Returns the process id of the parent process.
1961
4d76a344
RGS
1962Note for Linux users: on Linux, the C functions C<getpid()> and
1963C<getppid()> return different values from different threads. In order to
1964be portable, this behavior is not reflected by the perl-level function
1965C<getppid()>, that returns a consistent value across threads. If you want
e3256f86
RGS
1966to call the underlying C<getppid()>, you may use the CPAN module
1967C<Linux::Pid>.
4d76a344 1968
a0d0e21e
LW
1969=item getpriority WHICH,WHO
1970
4633a7c4
LW
1971Returns the current priority for a process, a process group, or a user.
1972(See L<getpriority(2)>.) Will raise a fatal exception if used on a
f86cebdf 1973machine that doesn't implement getpriority(2).
a0d0e21e
LW
1974
1975=item getpwnam NAME
1976
1977=item getgrnam NAME
1978
1979=item gethostbyname NAME
1980
1981=item getnetbyname NAME
1982
1983=item getprotobyname NAME
1984
1985=item getpwuid UID
1986
1987=item getgrgid GID
1988
1989=item getservbyname NAME,PROTO
1990
1991=item gethostbyaddr ADDR,ADDRTYPE
1992
1993=item getnetbyaddr ADDR,ADDRTYPE
1994
1995=item getprotobynumber NUMBER
1996
1997=item getservbyport PORT,PROTO
1998
1999=item getpwent
2000
2001=item getgrent
2002
2003=item gethostent
2004
2005=item getnetent
2006
2007=item getprotoent
2008
2009=item getservent
2010
2011=item setpwent
2012
2013=item setgrent
2014
2015=item sethostent STAYOPEN
2016
2017=item setnetent STAYOPEN
2018
2019=item setprotoent STAYOPEN
2020
2021=item setservent STAYOPEN
2022
2023=item endpwent
2024
2025=item endgrent
2026
2027=item endhostent
2028
2029=item endnetent
2030
2031=item endprotoent
2032
2033=item endservent
2034
2035These routines perform the same functions as their counterparts in the
5a964f20 2036system library. In list context, the return values from the
a0d0e21e
LW
2037various get routines are as follows:
2038
2039 ($name,$passwd,$uid,$gid,
6ee623d5 2040 $quota,$comment,$gcos,$dir,$shell,$expire) = getpw*
a0d0e21e
LW
2041 ($name,$passwd,$gid,$members) = getgr*
2042 ($name,$aliases,$addrtype,$length,@addrs) = gethost*
2043 ($name,$aliases,$addrtype,$net) = getnet*
2044 ($name,$aliases,$proto) = getproto*
2045 ($name,$aliases,$port,$proto) = getserv*
2046
2047(If the entry doesn't exist you get a null list.)
2048
4602f195
JH
2049The exact meaning of the $gcos field varies but it usually contains
2050the real name of the user (as opposed to the login name) and other
2051information pertaining to the user. Beware, however, that in many
2052system users are able to change this information and therefore it
106325ad 2053cannot be trusted and therefore the $gcos is tainted (see
2959b6e3
JH
2054L<perlsec>). The $passwd and $shell, user's encrypted password and
2055login shell, are also tainted, because of the same reason.
4602f195 2056
5a964f20 2057In scalar context, you get the name, unless the function was a
a0d0e21e
LW
2058lookup by name, in which case you get the other thing, whatever it is.
2059(If the entry doesn't exist you get the undefined value.) For example:
2060
5a964f20
TC
2061 $uid = getpwnam($name);
2062 $name = getpwuid($num);
2063 $name = getpwent();
2064 $gid = getgrnam($name);
08a33e13 2065 $name = getgrgid($num);
5a964f20
TC
2066 $name = getgrent();
2067 #etc.
a0d0e21e 2068
4602f195
JH
2069In I<getpw*()> the fields $quota, $comment, and $expire are special
2070cases in the sense that in many systems they are unsupported. If the
2071$quota is unsupported, it is an empty scalar. If it is supported, it
2072usually encodes the disk quota. If the $comment field is unsupported,
2073it is an empty scalar. If it is supported it usually encodes some
2074administrative comment about the user. In some systems the $quota
2075field may be $change or $age, fields that have to do with password
2076aging. In some systems the $comment field may be $class. The $expire
2077field, if present, encodes the expiration period of the account or the
2078password. For the availability and the exact meaning of these fields
2079in your system, please consult your getpwnam(3) documentation and your
2080F<pwd.h> file. You can also find out from within Perl what your
2081$quota and $comment fields mean and whether you have the $expire field
2082by using the C<Config> module and the values C<d_pwquota>, C<d_pwage>,
2083C<d_pwchange>, C<d_pwcomment>, and C<d_pwexpire>. Shadow password
2084files are only supported if your vendor has implemented them in the
2085intuitive fashion that calling the regular C library routines gets the
5d3a0a3b 2086shadow versions if you're running under privilege or if there exists
cf264981
SP
2087the shadow(3) functions as found in System V (this includes Solaris
2088and Linux.) Those systems that implement a proprietary shadow password
5d3a0a3b 2089facility are unlikely to be supported.
6ee623d5 2090
19799a22 2091The $members value returned by I<getgr*()> is a space separated list of
a0d0e21e
LW
2092the login names of the members of the group.
2093
2094For the I<gethost*()> functions, if the C<h_errno> variable is supported in
2095C, it will be returned to you via C<$?> if the function call fails. The
7660c0ab 2096C<@addrs> value returned by a successful call is a list of the raw
a0d0e21e
LW
2097addresses returned by the corresponding system library call. In the
2098Internet domain, each address is four bytes long and you can unpack it
2099by saying something like:
2100
f337b084 2101 ($a,$b,$c,$d) = unpack('W4',$addr[0]);
a0d0e21e 2102
2b5ab1e7
TC
2103The Socket library makes this slightly easier:
2104
2105 use Socket;
2106 $iaddr = inet_aton("127.1"); # or whatever address
2107 $name = gethostbyaddr($iaddr, AF_INET);
2108
2109 # or going the other way
19799a22 2110 $straddr = inet_ntoa($iaddr);
2b5ab1e7 2111
19799a22
GS
2112If you get tired of remembering which element of the return list
2113contains which return value, by-name interfaces are provided
2114in standard modules: C<File::stat>, C<Net::hostent>, C<Net::netent>,
2115C<Net::protoent>, C<Net::servent>, C<Time::gmtime>, C<Time::localtime>,
2116and C<User::grent>. These override the normal built-ins, supplying
2117versions that return objects with the appropriate names
2118for each field. For example:
5a964f20
TC
2119
2120 use File::stat;
2121 use User::pwent;
2122 $is_his = (stat($filename)->uid == pwent($whoever)->uid);
2123
b76cc8ba
NIS
2124Even though it looks like they're the same method calls (uid),
2125they aren't, because a C<File::stat> object is different from
19799a22 2126a C<User::pwent> object.
5a964f20 2127
a0d0e21e
LW
2128=item getsockname SOCKET
2129
19799a22
GS
2130Returns the packed sockaddr address of this end of the SOCKET connection,
2131in case you don't know the address because you have several different
2132IPs that the connection might have come in on.
a0d0e21e 2133
4633a7c4
LW
2134 use Socket;
2135 $mysockaddr = getsockname(SOCK);
19799a22 2136 ($port, $myaddr) = sockaddr_in($mysockaddr);
b76cc8ba 2137 printf "Connect to %s [%s]\n",
19799a22
GS
2138 scalar gethostbyaddr($myaddr, AF_INET),
2139 inet_ntoa($myaddr);
a0d0e21e
LW
2140
2141=item getsockopt SOCKET,LEVEL,OPTNAME
2142
636e6b1f
TH
2143Queries the option named OPTNAME associated with SOCKET at a given LEVEL.
2144Options may exist at multiple protocol levels depending on the socket
2145type, but at least the uppermost socket level SOL_SOCKET (defined in the
2146C<Socket> module) will exist. To query options at another level the
2147protocol number of the appropriate protocol controlling the option
2148should be supplied. For example, to indicate that an option is to be
2149interpreted by the TCP protocol, LEVEL should be set to the protocol
2150number of TCP, which you can get using getprotobyname.
2151
2152The call returns a packed string representing the requested socket option,
2153or C<undef> if there is an error (the error reason will be in $!). What
2154exactly is in the packed string depends in the LEVEL and OPTNAME, consult
2155your system documentation for details. A very common case however is that
cf264981 2156the option is an integer, in which case the result will be a packed
636e6b1f
TH
2157integer which you can decode using unpack with the C<i> (or C<I>) format.
2158
2159An example testing if Nagle's algorithm is turned on on a socket:
2160
4852725b 2161 use Socket qw(:all);
636e6b1f
TH
2162
2163 defined(my $tcp = getprotobyname("tcp"))
2164 or die "Could not determine the protocol number for tcp";
4852725b
DD
2165 # my $tcp = IPPROTO_TCP; # Alternative
2166 my $packed = getsockopt($socket, $tcp, TCP_NODELAY)
2167 or die "Could not query TCP_NODELAY socket option: $!";
636e6b1f
TH
2168 my $nodelay = unpack("I", $packed);
2169 print "Nagle's algorithm is turned ", $nodelay ? "off\n" : "on\n";
2170
a0d0e21e
LW
2171
2172=item glob EXPR
2173
0a753a76
PP
2174=item glob
2175
d9a9d457
JL
2176In list context, returns a (possibly empty) list of filename expansions on
2177the value of EXPR such as the standard Unix shell F</bin/csh> would do. In
2178scalar context, glob iterates through such filename expansions, returning
2179undef when the list is exhausted. This is the internal function
2180implementing the C<< <*.c> >> operator, but you can use it directly. If
2181EXPR is omitted, C<$_> is used. The C<< <*.c> >> operator is discussed in
2182more detail in L<perlop/"I/O Operators">.
a0d0e21e 2183
3a4b19e4
GS
2184Beginning with v5.6.0, this operator is implemented using the standard
2185C<File::Glob> extension. See L<File::Glob> for details.
2186
a0d0e21e
LW
2187=item gmtime EXPR
2188
ce2984c3
PF
2189=item gmtime
2190
d1be9408 2191Converts a time as returned by the time function to an 8-element list
54310121 2192with the time localized for the standard Greenwich time zone.
4633a7c4 2193Typically used as follows:
a0d0e21e 2194
b76cc8ba 2195 # 0 1 2 3 4 5 6 7
48a26b3a 2196 ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday) =
a0d0e21e
LW
2197 gmtime(time);
2198
48a26b3a
GS
2199All list elements are numeric, and come straight out of the C `struct
2200tm'. $sec, $min, and $hour are the seconds, minutes, and hours of the
2201specified time. $mday is the day of the month, and $mon is the month
2202itself, in the range C<0..11> with 0 indicating January and 11
2203indicating December. $year is the number of years since 1900. That
2204is, $year is C<123> in year 2023. $wday is the day of the week, with
22050 indicating Sunday and 3 indicating Wednesday. $yday is the day of
b76cc8ba 2206the year, in the range C<0..364> (or C<0..365> in leap years.)
48a26b3a
GS
2207
2208Note that the $year element is I<not> simply the last two digits of
cf264981 2209the year. If you assume it is then you create non-Y2K-compliant
48a26b3a 2210programs--and you wouldn't want to do that, would you?
2f9daede 2211
abd75f24
GS
2212The proper way to get a complete 4-digit year is simply:
2213
2214 $year += 1900;
2215
2216And to get the last two digits of the year (e.g., '01' in 2001) do:
2217
2218 $year = sprintf("%02d", $year % 100);
2219
48a26b3a 2220If EXPR is omitted, C<gmtime()> uses the current time (C<gmtime(time)>).
a0d0e21e 2221
48a26b3a 2222In scalar context, C<gmtime()> returns the ctime(3) value:
0a753a76
PP
2223
2224 $now_string = gmtime; # e.g., "Thu Oct 13 04:54:34 1994"
2225
fe86afc2
NC
2226If you need local time instead of GMT use the L</localtime> builtin.
2227See also the C<timegm> function provided by the C<Time::Local> module,
2228and the strftime(3) and mktime(3) functions available via the L<POSIX> module.
7660c0ab 2229
fe86afc2
NC
2230This scalar value is B<not> locale dependent (see L<perllocale>), but is
2231instead a Perl builtin. To get somewhat similar but locale dependent date
2232strings, see the example in L</localtime>.
0a753a76 2233
62aa5637
MS
2234See L<perlport/gmtime> for portability concerns.
2235
a0d0e21e
LW
2236=item goto LABEL
2237
748a9306
LW
2238=item goto EXPR
2239
a0d0e21e
LW
2240=item goto &NAME
2241
7660c0ab 2242The C<goto-LABEL> form finds the statement labeled with LABEL and resumes
a0d0e21e 2243execution there. It may not be used to go into any construct that
7660c0ab 2244requires initialization, such as a subroutine or a C<foreach> loop. It
0a753a76 2245also can't be used to go into a construct that is optimized away,
19799a22 2246or to get out of a block or subroutine given to C<sort>.
0a753a76 2247It can be used to go almost anywhere else within the dynamic scope,
a0d0e21e 2248including out of subroutines, but it's usually better to use some other
19799a22 2249construct such as C<last> or C<die>. The author of Perl has never felt the
7660c0ab 2250need to use this form of C<goto> (in Perl, that is--C is another matter).
1b6921cb
BT
2251(The difference being that C does not offer named loops combined with
2252loop control. Perl does, and this replaces most structured uses of C<goto>
2253in other languages.)
a0d0e21e 2254
7660c0ab
A
2255The C<goto-EXPR> form expects a label name, whose scope will be resolved
2256dynamically. This allows for computed C<goto>s per FORTRAN, but isn't
748a9306
LW
2257necessarily recommended if you're optimizing for maintainability:
2258
2259 goto ("FOO", "BAR", "GLARCH")[$i];
2260
1b6921cb
BT
2261The C<goto-&NAME> form is quite different from the other forms of
2262C<goto>. In fact, it isn't a goto in the normal sense at all, and
2263doesn't have the stigma associated with other gotos. Instead, it
2264exits the current subroutine (losing any changes set by local()) and
2265immediately calls in its place the named subroutine using the current
2266value of @_. This is used by C<AUTOLOAD> subroutines that wish to
2267load another subroutine and then pretend that the other subroutine had
2268been called in the first place (except that any modifications to C<@_>
6cb9131c
GS
2269in the current subroutine are propagated to the other subroutine.)
2270After the C<goto>, not even C<caller> will be able to tell that this
2271routine was called first.
2272
2273NAME needn't be the name of a subroutine; it can be a scalar variable
cf264981 2274containing a code reference, or a block that evaluates to a code
6cb9131c 2275reference.
a0d0e21e
LW
2276
2277=item grep BLOCK LIST
2278
2279=item grep EXPR,LIST
2280
2b5ab1e7
TC
2281This is similar in spirit to, but not the same as, grep(1) and its
2282relatives. In particular, it is not limited to using regular expressions.
2f9daede 2283
a0d0e21e 2284Evaluates the BLOCK or EXPR for each element of LIST (locally setting
7660c0ab 2285C<$_> to each element) and returns the list value consisting of those
19799a22
GS
2286elements for which the expression evaluated to true. In scalar
2287context, returns the number of times the expression was true.
a0d0e21e
LW
2288
2289 @foo = grep(!/^#/, @bar); # weed out comments
2290
2291or equivalently,
2292
2293 @foo = grep {!/^#/} @bar; # weed out comments
2294
be3174d2
GS
2295Note that C<$_> is an alias to the list value, so it can be used to
2296modify the elements of the LIST. While this is useful and supported,
2297it can cause bizarre results if the elements of LIST are not variables.
2b5ab1e7
TC
2298Similarly, grep returns aliases into the original list, much as a for
2299loop's index variable aliases the list elements. That is, modifying an
19799a22
GS
2300element of a list returned by grep (for example, in a C<foreach>, C<map>
2301or another C<grep>) actually modifies the element in the original list.
2b5ab1e7 2302This is usually something to be avoided when writing clear code.
a0d0e21e 2303
a4fb8298 2304If C<$_> is lexical in the scope where the C<grep> appears (because it has
cf264981 2305been declared with C<my $_>) then, in addition to being locally aliased to
a4fb8298
RGS
2306the list elements, C<$_> keeps being lexical inside the block; i.e. it
2307can't be seen from the outside, avoiding any potential side-effects.
2308
19799a22 2309See also L</map> for a list composed of the results of the BLOCK or EXPR.
38325410 2310
a0d0e21e
LW
2311=item hex EXPR
2312
54310121 2313=item hex
bbce6d69 2314
2b5ab1e7 2315Interprets EXPR as a hex string and returns the corresponding value.
38366c11 2316(To convert strings that might start with either C<0>, C<0x>, or C<0b>, see
2b5ab1e7 2317L</oct>.) If EXPR is omitted, uses C<$_>.
2f9daede
TPG
2318
2319 print hex '0xAf'; # prints '175'
2320 print hex 'aF'; # same
a0d0e21e 2321
19799a22 2322Hex strings may only represent integers. Strings that would cause
53305cf1 2323integer overflow trigger a warning. Leading whitespace is not stripped,
38366c11
DN
2324unlike oct(). To present something as hex, look into L</printf>,
2325L</sprintf>, or L</unpack>.
19799a22 2326
ce2984c3 2327=item import LIST
a0d0e21e 2328
19799a22 2329There is no builtin C<import> function. It is just an ordinary
4633a7c4 2330method (subroutine) defined (or inherited) by modules that wish to export
19799a22 2331names to another module. The C<use> function calls the C<import> method
cea6626f 2332for the package used. See also L</use>, L<perlmod>, and L<Exporter>.
a0d0e21e
LW
2333
2334=item index STR,SUBSTR,POSITION
2335
2336=item index STR,SUBSTR
2337
2b5ab1e7
TC
2338The index function searches for one string within another, but without
2339the wildcard-like behavior of a full regular-expression pattern match.
2340It returns the position of the first occurrence of SUBSTR in STR at
2341or after POSITION. If POSITION is omitted, starts searching from the
2342beginning of the string. The return value is based at C<0> (or whatever
2343you've set the C<$[> variable to--but don't do that). If the substring
cf264981 2344is not found, C<index> returns one less than the base, ordinarily C<-1>.
a0d0e21e
LW
2345
2346=item int EXPR
2347
54310121 2348=item int
bbce6d69 2349
7660c0ab 2350Returns the integer portion of EXPR. If EXPR is omitted, uses C<$_>.
2b5ab1e7
TC
2351You should not use this function for rounding: one because it truncates
2352towards C<0>, and two because machine representations of floating point
2353numbers can sometimes produce counterintuitive results. For example,
2354C<int(-6.725/0.025)> produces -268 rather than the correct -269; that's
2355because it's really more like -268.99999999999994315658 instead. Usually,
19799a22 2356the C<sprintf>, C<printf>, or the C<POSIX::floor> and C<POSIX::ceil>
2b5ab1e7 2357functions will serve you better than will int().
a0d0e21e
LW
2358
2359=item ioctl FILEHANDLE,FUNCTION,SCALAR
2360
2b5ab1e7 2361Implements the ioctl(2) function. You'll probably first have to say
a0d0e21e 2362
a11c483f 2363 require "sys/ioctl.ph"; # probably in $Config{archlib}/ioctl.ph
a0d0e21e 2364
a11c483f 2365to get the correct function definitions. If F<sys/ioctl.ph> doesn't
a0d0e21e 2366exist or doesn't have the correct definitions you'll have to roll your
61eff3bc 2367own, based on your C header files such as F<< <sys/ioctl.h> >>.
5a964f20 2368(There is a Perl script called B<h2ph> that comes with the Perl kit that
54310121 2369may help you in this, but it's nontrivial.) SCALAR will be read and/or
4633a7c4 2370written depending on the FUNCTION--a pointer to the string value of SCALAR
19799a22 2371will be passed as the third argument of the actual C<ioctl> call. (If SCALAR
4633a7c4
LW
2372has no string value but does have a numeric value, that value will be
2373passed rather than a pointer to the string value. To guarantee this to be
19799a22
GS
2374true, add a C<0> to the scalar before using it.) The C<pack> and C<unpack>
2375functions may be needed to manipulate the values of structures used by
b76cc8ba 2376C<ioctl>.
a0d0e21e 2377
19799a22 2378The return value of C<ioctl> (and C<fcntl>) is as follows:
a0d0e21e
LW
2379
2380 if OS returns: then Perl returns:
2381 -1 undefined value
2382 0 string "0 but true"
2383 anything else that number
2384
19799a22 2385Thus Perl returns true on success and false on failure, yet you can
a0d0e21e
LW
2386still easily determine the actual value returned by the operating
2387system:
2388
2b5ab1e7 2389 $retval = ioctl(...) || -1;
a0d0e21e
LW
2390 printf "System returned %d\n", $retval;
2391
be2f7487 2392The special string C<"0 but true"> is exempt from B<-w> complaints
5a964f20
TC
2393about improper numeric conversions.
2394
a0d0e21e
LW
2395=item join EXPR,LIST
2396
2b5ab1e7
TC
2397Joins the separate strings of LIST into a single string with fields
2398separated by the value of EXPR, and returns that new string. Example:
a0d0e21e 2399
2b5ab1e7 2400 $rec = join(':', $login,$passwd,$uid,$gid,$gcos,$home,$shell);
a0d0e21e 2401
eb6e2d6f
GS
2402Beware that unlike C<split>, C<join> doesn't take a pattern as its
2403first argument. Compare L</split>.
a0d0e21e 2404
aa689395
PP
2405=item keys HASH
2406
504f80c1
JH
2407Returns a list consisting of all the keys of the named hash.
2408(In scalar context, returns the number of keys.)
2409
2410The keys are returned in an apparently random order. The actual
2411random order is subject to change in future versions of perl, but it
2412is guaranteed to be the same order as either the C<values> or C<each>
4546b9e6
JH
2413function produces (given that the hash has not been modified). Since
2414Perl 5.8.1 the ordering is different even between different runs of
2415Perl for security reasons (see L<perlsec/"Algorithmic Complexity
d6df3700 2416Attacks">).
504f80c1 2417
cf264981
SP
2418As a side effect, calling keys() resets the HASH's internal iterator
2419(see L</each>). In particular, calling keys() in void context resets
2420the iterator with no other overhead.
a0d0e21e 2421
aa689395 2422Here is yet another way to print your environment:
a0d0e21e
LW
2423
2424 @keys = keys %ENV;
2425 @values = values %ENV;
b76cc8ba 2426 while (@keys) {
a0d0e21e
LW
2427 print pop(@keys), '=', pop(@values), "\n";
2428 }
2429
2430or how about sorted by key:
2431
2432 foreach $key (sort(keys %ENV)) {
2433 print $key, '=', $ENV{$key}, "\n";
2434 }
2435
8ea1e5d4
GS
2436The returned values are copies of the original keys in the hash, so
2437modifying them will not affect the original hash. Compare L</values>.
2438
19799a22 2439To sort a hash by value, you'll need to use a C<sort> function.
aa689395 2440Here's a descending numeric sort of a hash by its values:
4633a7c4 2441
5a964f20 2442 foreach $key (sort { $hash{$b} <=> $hash{$a} } keys %hash) {
4633a7c4
LW
2443 printf "%4d %s\n", $hash{$key}, $key;
2444 }
2445
19799a22 2446As an lvalue C<keys> allows you to increase the number of hash buckets
aa689395
PP
2447allocated for the given hash. This can gain you a measure of efficiency if
2448you know the hash is going to get big. (This is similar to pre-extending
2449an array by assigning a larger number to $#array.) If you say
55497cff
PP
2450
2451 keys %hash = 200;
2452
ab192400
GS
2453then C<%hash> will have at least 200 buckets allocated for it--256 of them,
2454in fact, since it rounds up to the next power of two. These
55497cff
PP
2455buckets will be retained even if you do C<%hash = ()>, use C<undef
2456%hash> if you want to free the storage while C<%hash> is still in scope.
2457You can't shrink the number of buckets allocated for the hash using
19799a22 2458C<keys> in this way (but you needn't worry about doing this by accident,
55497cff
PP
2459as trying has no effect).
2460
19799a22 2461See also C<each>, C<values> and C<sort>.
ab192400 2462
b350dd2f 2463=item kill SIGNAL, LIST
a0d0e21e 2464
b350dd2f 2465Sends a signal to a list of processes. Returns the number of
517db077
GS
2466processes successfully signaled (which is not necessarily the
2467same as the number actually killed).
a0d0e21e
LW
2468
2469 $cnt = kill 1, $child1, $child2;
2470 kill 9, @goners;
2471
b350dd2f 2472If SIGNAL is zero, no signal is sent to the process. This is a
1e9c1022 2473useful way to check that a child process is alive and hasn't changed
b350dd2f
GS
2474its UID. See L<perlport> for notes on the portability of this
2475construct.
2476
2477Unlike in the shell, if SIGNAL is negative, it kills
4633a7c4
LW
2478process groups instead of processes. (On System V, a negative I<PROCESS>
2479number will also kill process groups, but that's not portable.) That
2480means you usually want to use positive not negative signals. You may also
1e9c1022
JL
2481use a signal name in quotes.
2482
2483See L<perlipc/"Signals"> for more details.
a0d0e21e
LW
2484
2485=item last LABEL
2486
2487=item last
2488
2489The C<last> command is like the C<break> statement in C (as used in
2490loops); it immediately exits the loop in question. If the LABEL is
2491omitted, the command refers to the innermost enclosing loop. The
2492C<continue> block, if any, is not executed:
2493
4633a7c4
LW
2494 LINE: while (<STDIN>) {
2495 last LINE if /^$/; # exit when done with header
5a964f20 2496 #...
a0d0e21e
LW
2497 }
2498
4968c1e4 2499C<last> cannot be used to exit a block which returns a value such as
2b5ab1e7
TC
2500C<eval {}>, C<sub {}> or C<do {}>, and should not be used to exit
2501a grep() or map() operation.
4968c1e4 2502
6c1372ed
GS
2503Note that a block by itself is semantically identical to a loop
2504that executes once. Thus C<last> can be used to effect an early
2505exit out of such a block.
2506
98293880
JH
2507See also L</continue> for an illustration of how C<last>, C<next>, and
2508C<redo> work.
1d2dff63 2509
a0d0e21e
LW
2510=item lc EXPR
2511
54310121 2512=item lc
bbce6d69 2513
d1be9408 2514Returns a lowercased version of EXPR. This is the internal function
ad0029c4
JH
2515implementing the C<\L> escape in double-quoted strings. Respects
2516current LC_CTYPE locale if C<use locale> in force. See L<perllocale>
983ffd37 2517and L<perlunicode> for more details about locale and Unicode support.
a0d0e21e 2518
7660c0ab 2519If EXPR is omitted, uses C<$_>.
bbce6d69 2520
a0d0e21e
LW
2521=item lcfirst EXPR
2522
54310121 2523=item lcfirst
bbce6d69 2524
ad0029c4
JH
2525Returns the value of EXPR with the first character lowercased. This
2526is the internal function implementing the C<\l> escape in
2527double-quoted strings. Respects current LC_CTYPE locale if C<use
983ffd37
JH
2528locale> in force. See L<perllocale> and L<perlunicode> for more
2529details about locale and Unicode support.
a0d0e21e 2530
7660c0ab 2531If EXPR is omitted, uses C<$_>.
bbce6d69 2532
a0d0e21e
LW
2533=item length EXPR
2534
54310121 2535=item length
bbce6d69 2536
974da8e5 2537Returns the length in I<characters> of the value of EXPR. If EXPR is
b76cc8ba 2538omitted, returns length of C<$_>. Note that this cannot be used on
2b5ab1e7
TC
2539an entire array or hash to find out how many elements these have.
2540For that, use C<scalar @array> and C<scalar keys %hash> respectively.
a0d0e21e 2541
974da8e5
JH
2542Note the I<characters>: if the EXPR is in Unicode, you will get the
2543number of characters, not the number of bytes. To get the length
2544in bytes, use C<do { use bytes; length(EXPR) }>, see L<bytes>.
2545
a0d0e21e
LW
2546=item link OLDFILE,NEWFILE
2547
19799a22 2548Creates a new filename linked to the old filename. Returns true for
b76cc8ba 2549success, false otherwise.
a0d0e21e
LW
2550
2551=item listen SOCKET,QUEUESIZE
2552
19799a22 2553Does the same thing that the listen system call does. Returns true if
b76cc8ba 2554it succeeded, false otherwise. See the example in
cea6626f 2555L<perlipc/"Sockets: Client/Server Communication">.
a0d0e21e
LW
2556
2557=item local EXPR
2558
19799a22 2559You really probably want to be using C<my> instead, because C<local> isn't
b76cc8ba 2560what most people think of as "local". See
13a2d996 2561L<perlsub/"Private Variables via my()"> for details.
2b5ab1e7 2562
5a964f20
TC
2563A local modifies the listed variables to be local to the enclosing
2564block, file, or eval. If more than one value is listed, the list must
2565be placed in parentheses. See L<perlsub/"Temporary Values via local()">
2566for details, including issues with tied arrays and hashes.
a0d0e21e 2567
a0d0e21e
LW
2568=item localtime EXPR
2569
ba053783
AL
2570=item localtime
2571
19799a22 2572Converts a time as returned by the time function to a 9-element list
5f05dabc 2573with the time analyzed for the local time zone. Typically used as
a0d0e21e
LW
2574follows:
2575
54310121 2576 # 0 1 2 3 4 5 6 7 8
a0d0e21e 2577 ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) =
ba053783 2578 localtime(time);
a0d0e21e 2579
48a26b3a 2580All list elements are numeric, and come straight out of the C `struct
ba053783
AL
2581tm'. C<$sec>, C<$min>, and C<$hour> are the seconds, minutes, and hours
2582of the specified time.
48a26b3a 2583
ba053783
AL
2584C<$mday> is the day of the month, and C<$mon> is the month itself, in
2585the range C<0..11> with 0 indicating January and 11 indicating December.
2586This makes it easy to get a month name from a list:
54310121 2587
ba053783
AL
2588 my @abbr = qw( Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec );
2589 print "$abbr[$mon] $mday";
2590 # $mon=9, $mday=18 gives "Oct 18"
abd75f24 2591
ba053783
AL
2592C<$year> is the number of years since 1900, not just the last two digits
2593of the year. That is, C<$year> is C<123> in year 2023. The proper way
2594to get a complete 4-digit year is simply:
abd75f24 2595
ba053783 2596 $year += 1900;
abd75f24 2597
ba053783
AL
2598To get the last two digits of the year (e.g., '01' in 2001) do:
2599
2600 $year = sprintf("%02d", $year % 100);
2601
2602C<$wday> is the day of the week, with 0 indicating Sunday and 3 indicating
2603Wednesday. C<$yday> is the day of the year, in the range C<0..364>
2604(or C<0..365> in leap years.)
2605
2606C<$isdst> is true if the specified time occurs during Daylight Saving
2607Time, false otherwise.
abd75f24 2608
48a26b3a 2609If EXPR is omitted, C<localtime()> uses the current time (C<localtime(time)>).
a0d0e21e 2610
48a26b3a 2611In scalar context, C<localtime()> returns the ctime(3) value:
a0d0e21e 2612
5f05dabc 2613 $now_string = localtime; # e.g., "Thu Oct 13 04:54:34 1994"
a0d0e21e 2614
fe86afc2
NC
2615This scalar value is B<not> locale dependent but is a Perl builtin. For GMT
2616instead of local time use the L</gmtime> builtin. See also the
2617C<Time::Local> module (to convert the second, minutes, hours, ... back to
2618the integer value returned by time()), and the L<POSIX> module's strftime(3)
2619and mktime(3) functions.
2620
2621To get somewhat similar but locale dependent date strings, set up your
2622locale environment variables appropriately (please see L<perllocale>) and
2623try for example:
a3cb178b 2624
5a964f20 2625 use POSIX qw(strftime);
2b5ab1e7 2626 $now_string = strftime "%a %b %e %H:%M:%S %Y", localtime;
fe86afc2
NC
2627 # or for GMT formatted appropriately for your locale:
2628 $now_string = strftime "%a %b %e %H:%M:%S %Y", gmtime;
a3cb178b
GS
2629
2630Note that the C<%a> and C<%b>, the short forms of the day of the week
2631and the month of the year, may not necessarily be three characters wide.
a0d0e21e 2632
62aa5637
MS
2633See L<perlport/localtime> for portability concerns.
2634
07698885 2635=item lock THING
19799a22 2636
01e6739c 2637This function places an advisory lock on a shared variable, or referenced
03730085 2638object contained in I<THING> until the lock goes out of scope.
a6d5524e 2639
f3a23afb 2640lock() is a "weak keyword" : this means that if you've defined a function
67408cae 2641by this name (before any calls to it), that function will be called
03730085
AB
2642instead. (However, if you've said C<use threads>, lock() is always a
2643keyword.) See L<threads>.
19799a22 2644
a0d0e21e
LW
2645=item log EXPR
2646
54310121 2647=item log
bbce6d69 2648
2b5ab1e7
TC
2649Returns the natural logarithm (base I<e>) of EXPR. If EXPR is omitted,
2650returns log of C<$_>. To get the log of another base, use basic algebra:
19799a22 2651The base-N log of a number is equal to the natural log of that number
2b5ab1e7
TC
2652divided by the natural log of N. For example:
2653
2654 sub log10 {
2655 my $n = shift;
2656 return log($n)/log(10);
b76cc8ba 2657 }
2b5ab1e7
TC
2658
2659See also L</exp> for the inverse operation.
a0d0e21e 2660
a0d0e21e
LW
2661=item lstat EXPR
2662
54310121 2663=item lstat
bbce6d69 2664
19799a22 2665Does the same thing as the C<stat> function (including setting the
5a964f20
TC
2666special C<_> filehandle) but stats a symbolic link instead of the file
2667the symbolic link points to. If symbolic links are unimplemented on
c837d5b4
DP
2668your system, a normal C<stat> is done. For much more detailed
2669information, please see the documentation for C<stat>.
a0d0e21e 2670
7660c0ab 2671If EXPR is omitted, stats C<$_>.
bbce6d69 2672
a0d0e21e
LW
2673=item m//
2674
2675The match operator. See L<perlop>.
2676
2677=item map BLOCK LIST
2678
2679=item map EXPR,LIST
2680
19799a22
GS
2681Evaluates the BLOCK or EXPR for each element of LIST (locally setting
2682C<$_> to each element) and returns the list value composed of the
2683results of each such evaluation. In scalar context, returns the
2684total number of elements so generated. Evaluates BLOCK or EXPR in
2685list context, so each element of LIST may produce zero, one, or
2686more elements in the returned value.
dd99ebda 2687
a0d0e21e
LW
2688 @chars = map(chr, @nums);
2689
2690translates a list of numbers to the corresponding characters. And
2691
4633a7c4 2692 %hash = map { getkey($_) => $_ } @array;
a0d0e21e
LW
2693
2694is just a funny way to write
2695
2696 %hash = ();
2697 foreach $_ (@array) {
4633a7c4 2698 $hash{getkey($_)} = $_;
a0d0e21e
LW
2699 }
2700
be3174d2
GS
2701Note that C<$_> is an alias to the list value, so it can be used to
2702modify the elements of the LIST. While this is useful and supported,
2703it can cause bizarre results if the elements of LIST are not variables.
2b5ab1e7
TC
2704Using a regular C<foreach> loop for this purpose would be clearer in
2705most cases. See also L</grep> for an array composed of those items of
2706the original list for which the BLOCK or EXPR evaluates to true.
fb73857a 2707
a4fb8298 2708If C<$_> is lexical in the scope where the C<map> appears (because it has
cf264981 2709been declared with C<my $_>) then, in addition to being locally aliased to
a4fb8298
RGS
2710the list elements, C<$_> keeps being lexical inside the block; i.e. it
2711can't be seen from the outside, avoiding any potential side-effects.
2712
205fdb4d
NC
2713C<{> starts both hash references and blocks, so C<map { ...> could be either
2714the start of map BLOCK LIST or map EXPR, LIST. Because perl doesn't look
2715ahead for the closing C<}> it has to take a guess at which its dealing with
2716based what it finds just after the C<{>. Usually it gets it right, but if it
2717doesn't it won't realize something is wrong until it gets to the C<}> and
2718encounters the missing (or unexpected) comma. The syntax error will be
2719reported close to the C<}> but you'll need to change something near the C<{>
2720such as using a unary C<+> to give perl some help:
2721
2722 %hash = map { "\L$_", 1 } @array # perl guesses EXPR. wrong
2723 %hash = map { +"\L$_", 1 } @array # perl guesses BLOCK. right
2724 %hash = map { ("\L$_", 1) } @array # this also works
2725 %hash = map { lc($_), 1 } @array # as does this.
2726 %hash = map +( lc($_), 1 ), @array # this is EXPR and works!
cea6626f 2727
205fdb4d
NC
2728 %hash = map ( lc($_), 1 ), @array # evaluates to (1, @array)
2729
2730or to force an anon hash constructor use C<+{>
2731
2732 @hashes = map +{ lc($_), 1 }, @array # EXPR, so needs , at end
2733
2734and you get list of anonymous hashes each with only 1 entry.
2735
19799a22 2736=item mkdir FILENAME,MASK
a0d0e21e 2737
5a211162
GS
2738=item mkdir FILENAME
2739
491873e5
RGS
2740=item mkdir
2741
0591cd52 2742Creates the directory specified by FILENAME, with permissions
19799a22
GS
2743specified by MASK (as modified by C<umask>). If it succeeds it
2744returns true, otherwise it returns false and sets C<$!> (errno).
491873e5
RGS
2745If omitted, MASK defaults to 0777. If omitted, FILENAME defaults
2746to C<$_>.
0591cd52 2747
19799a22 2748In general, it is better to create directories with permissive MASK,
0591cd52 2749and let the user modify that with their C<umask>, than it is to supply
19799a22 2750a restrictive MASK and give the user no way to be more permissive.
0591cd52
NT
2751The exceptions to this rule are when the file or directory should be
2752kept private (mail files, for instance). The perlfunc(1) entry on
19799a22 2753C<umask> discusses the choice of MASK in more detail.
a0d0e21e 2754
cc1852e8
JH
2755Note that according to the POSIX 1003.1-1996 the FILENAME may have any
2756number of trailing slashes. Some operating and filesystems do not get
2757this right, so Perl automatically removes all trailing slashes to keep
2758everyone happy.
2759
a0d0e21e
LW
2760=item msgctl ID,CMD,ARG
2761
f86cebdf 2762Calls the System V IPC function msgctl(2). You'll probably have to say
0ade1984
JH
2763
2764 use IPC::SysV;
2765
7660c0ab 2766first to get the correct constant definitions. If CMD is C<IPC_STAT>,
cf264981 2767then ARG must be a variable that will hold the returned C<msqid_ds>
951ba7fe
GS
2768structure. Returns like C<ioctl>: the undefined value for error,
2769C<"0 but true"> for zero, or the actual return value otherwise. See also
4755096e 2770L<perlipc/"SysV IPC">, C<IPC::SysV>, and C<IPC::Semaphore> documentation.
a0d0e21e
LW
2771
2772=item msgget KEY,FLAGS
2773
f86cebdf 2774Calls the System V IPC function msgget(2). Returns the message queue
4755096e
GS
2775id, or the undefined value if there is an error. See also
2776L<perlipc/"SysV IPC"> and C<IPC::SysV> and C<IPC::Msg> documentation.
a0d0e21e 2777
a0d0e21e
LW
2778=item msgrcv ID,VAR,SIZE,TYPE,FLAGS
2779
2780Calls the System V IPC function msgrcv to receive a message from
2781message queue ID into variable VAR with a maximum message size of
41d6edb2
JH
2782SIZE. Note that when a message is received, the message type as a
2783native long integer will be the first thing in VAR, followed by the
2784actual message. This packing may be opened with C<unpack("l! a*")>.
2785Taints the variable. Returns true if successful, or false if there is
4755096e
GS
2786an error. See also L<perlipc/"SysV IPC">, C<IPC::SysV>, and
2787C<IPC::SysV::Msg> documentation.
41d6edb2
JH
2788
2789=item msgsnd ID,MSG,FLAGS
2790
2791Calls the System V IPC function msgsnd to send the message MSG to the
2792message queue ID. MSG must begin with the native long integer message
2793type, and be followed by the length of the actual message, and finally
2794the message itself. This kind of packing can be achieved with
2795C<pack("l! a*", $type, $message)>. Returns true if successful,
2796or false if there is an error. See also C<IPC::SysV>
2797and C<IPC::SysV::Msg> documentation.
a0d0e21e
LW
2798
2799=item my EXPR
2800
307ea6df
JH
2801=item my TYPE EXPR
2802
1d2de774 2803=item my EXPR : ATTRS
09bef843 2804
1d2de774 2805=item my TYPE EXPR : ATTRS
307ea6df 2806
19799a22 2807A C<my> declares the listed variables to be local (lexically) to the
1d2de774
JH
2808enclosing block, file, or C<eval>. If more than one value is listed,
2809the list must be placed in parentheses.
307ea6df 2810
1d2de774
JH
2811The exact semantics and interface of TYPE and ATTRS are still
2812evolving. TYPE is currently bound to the use of C<fields> pragma,
307ea6df
JH
2813and attributes are handled using the C<attributes> pragma, or starting
2814from Perl 5.8.0 also via the C<Attribute::Handlers> module. See
2815L<perlsub/"Private Variables via my()"> for details, and L<fields>,
2816L<attributes>, and L<Attribute::Handlers>.
4633a7c4 2817
a0d0e21e
LW
2818=item next LABEL
2819
2820=item next
2821
2822The C<next> command is like the C<continue> statement in C; it starts
2823the next iteration of the loop:
2824
4633a7c4
LW
2825 LINE: while (<STDIN>) {
2826 next LINE if /^#/; # discard comments
5a964f20 2827 #...
a0d0e21e
LW
2828 }
2829
2830Note that if there were a C<continue> block on the above, it would get
2831executed even on discarded lines. If the LABEL is omitted, the command
2832refers to the innermost enclosing loop.
2833
4968c1e4 2834C<next> cannot be used to exit a block which returns a value such as
2b5ab1e7
TC
2835C<eval {}>, C<sub {}> or C<do {}>, and should not be used to exit
2836a grep() or map() operation.
4968c1e4 2837
6c1372ed
GS
2838Note that a block by itself is semantically identical to a loop
2839that executes once. Thus C<next> will exit such a block early.
2840
98293880
JH
2841See also L</continue> for an illustration of how C<last>, C<next>, and
2842C<redo> work.
1d2dff63 2843
4a66ea5a
RGS
2844=item no Module VERSION LIST
2845
2846=item no Module VERSION
2847
a0d0e21e
LW
2848=item no Module LIST
2849
4a66ea5a
RGS
2850=item no Module
2851
593b9c14 2852See the C<use> function, of which C<no> is the opposite.
a0d0e21e
LW
2853
2854=item oct EXPR
2855
54310121 2856=item oct
bbce6d69 2857
4633a7c4 2858Interprets EXPR as an octal string and returns the corresponding
4f19785b
WSI
2859value. (If EXPR happens to start off with C<0x>, interprets it as a
2860hex string. If EXPR starts off with C<0b>, it is interpreted as a
53305cf1
NC
2861binary string. Leading whitespace is ignored in all three cases.)
2862The following will handle decimal, binary, octal, and hex in the standard
2863Perl or C notation:
a0d0e21e
LW
2864
2865 $val = oct($val) if $val =~ /^0/;
2866
19799a22
GS
2867If EXPR is omitted, uses C<$_>. To go the other way (produce a number
2868in octal), use sprintf() or printf():
2869
2870 $perms = (stat("filename"))[2] & 07777;
2871 $oct_perms = sprintf "%lo", $perms;
2872
2873The oct() function is commonly used when a string such as C<644> needs
2874to be converted into a file mode, for example. (Although perl will
2875automatically convert strings into numbers as needed, this automatic
2876conversion assumes base 10.)
a0d0e21e
LW
2877
2878=item open FILEHANDLE,EXPR
2879
68bd7414
NIS
2880=item open FILEHANDLE,MODE,EXPR
2881
2882=item open FILEHANDLE,MODE,EXPR,LIST
2883
ba964c95
T
2884=item open FILEHANDLE,MODE,REFERENCE
2885
a0d0e21e
LW
2886=item open FILEHANDLE
2887
2888Opens the file whose filename is given by EXPR, and associates it with
ed53a2bb
JH
2889FILEHANDLE.
2890
2891(The following is a comprehensive reference to open(): for a gentler
2892introduction you may consider L<perlopentut>.)
2893
a28cd5c9
NT
2894If FILEHANDLE is an undefined scalar variable (or array or hash element)
2895the variable is assigned a reference to a new anonymous filehandle,
2896otherwise if FILEHANDLE is an expression, its value is used as the name of
2897the real filehandle wanted. (This is considered a symbolic reference, so
2898C<use strict 'refs'> should I<not> be in effect.)
ed53a2bb
JH
2899
2900If EXPR is omitted, the scalar variable of the same name as the
2901FILEHANDLE contains the filename. (Note that lexical variables--those
2902declared with C<my>--will not work for this purpose; so if you're
67408cae 2903using C<my>, specify EXPR in your call to open.)
ed53a2bb
JH
2904
2905If three or more arguments are specified then the mode of opening and
2906the file name are separate. If MODE is C<< '<' >> or nothing, the file
2907is opened for input. If MODE is C<< '>' >>, the file is truncated and
2908opened for output, being created if necessary. If MODE is C<<< '>>' >>>,
b76cc8ba 2909the file is opened for appending, again being created if necessary.
5a964f20 2910
ed53a2bb
JH
2911You can put a C<'+'> in front of the C<< '>' >> or C<< '<' >> to
2912indicate that you want both read and write access to the file; thus
2913C<< '+<' >> is almost always preferred for read/write updates--the C<<
2914'+>' >> mode would clobber the file first. You can't usually use
2915either read-write mode for updating textfiles, since they have
2916variable length records. See the B<-i> switch in L<perlrun> for a
2917better approach. The file is created with permissions of C<0666>
2918modified by the process' C<umask> value.
2919
2920These various prefixes correspond to the fopen(3) modes of C<'r'>,
2921C<'r+'>, C<'w'>, C<'w+'>, C<'a'>, and C<'a+'>.
5f05dabc 2922
6170680b
IZ
2923In the 2-arguments (and 1-argument) form of the call the mode and
2924filename should be concatenated (in this order), possibly separated by
68bd7414
NIS
2925spaces. It is possible to omit the mode in these forms if the mode is
2926C<< '<' >>.
6170680b 2927
7660c0ab 2928If the filename begins with C<'|'>, the filename is interpreted as a
5a964f20 2929command to which output is to be piped, and if the filename ends with a
f244e06d
GS
2930C<'|'>, the filename is interpreted as a command which pipes output to
2931us. See L<perlipc/"Using open() for IPC">
19799a22 2932for more examples of this. (You are not allowed to C<open> to a command
5a964f20 2933that pipes both in I<and> out, but see L<IPC::Open2>, L<IPC::Open3>,
4a4eefd0
GS
2934and L<perlipc/"Bidirectional Communication with Another Process">
2935for alternatives.)
cb1a09d0 2936
ed53a2bb
JH
2937For three or more arguments if MODE is C<'|-'>, the filename is
2938interpreted as a command to which output is to be piped, and if MODE
2939is C<'-|'>, the filename is interpreted as a command which pipes
2940output to us. In the 2-arguments (and 1-argument) form one should
2941replace dash (C<'-'>) with the command.
2942See L<perlipc/"Using open() for IPC"> for more examples of this.
2943(You are not allowed to C<open> to a command that pipes both in I<and>
2944out, but see L<IPC::Open2>, L<IPC::Open3>, and
2945L<perlipc/"Bidirectional Communication"> for alternatives.)
2946
2947In the three-or-more argument form of pipe opens, if LIST is specified
2948(extra arguments after the command name) then LIST becomes arguments
2949to the command invoked if the platform supports it. The meaning of
2950C<open> with more than three arguments for non-pipe modes is not yet
2951specified. Experimental "layers" may give extra LIST arguments
2952meaning.
6170680b
IZ
2953
2954In the 2-arguments (and 1-argument) form opening C<'-'> opens STDIN
b76cc8ba 2955and opening C<< '>-' >> opens STDOUT.
6170680b 2956
fae2c0fb
RGS
2957You may use the three-argument form of open to specify IO "layers"
2958(sometimes also referred to as "disciplines") to be applied to the handle
2959that affect how the input and output are processed (see L<open> and
2960L<PerlIO> for more details). For example
7207e29d 2961
9124316e
JH
2962 open(FH, "<:utf8", "file")
2963
2964will open the UTF-8 encoded file containing Unicode characters,
fae2c0fb
RGS
2965see L<perluniintro>. (Note that if layers are specified in the
2966three-arg form then default layers set by the C<open> pragma are
01e6739c 2967ignored.)
ed53a2bb
JH
2968
2969Open returns nonzero upon success, the undefined value otherwise. If
2970the C<open> involved a pipe, the return value happens to be the pid of
2971the subprocess.
cb1a09d0 2972
ed53a2bb
JH
2973If you're running Perl on a system that distinguishes between text
2974files and binary files, then you should check out L</binmode> for tips
2975for dealing with this. The key distinction between systems that need
2976C<binmode> and those that don't is their text file formats. Systems
8939ba94 2977like Unix, Mac OS, and Plan 9, which delimit lines with a single
ed53a2bb
JH
2978character, and which encode that character in C as C<"\n">, do not
2979need C<binmode>. The rest need it.
cb1a09d0 2980
fb73857a 2981When opening a file, it's usually a bad idea to continue normal execution
19799a22
GS
2982if the request failed, so C<open> is frequently used in connection with
2983C<die>. Even if C<die> won't do what you want (say, in a CGI script,
fb73857a 2984where you want to make a nicely formatted error message (but there are
5a964f20 2985modules that can help with that problem)) you should always check
19799a22 2986the return value from opening a file. The infrequent exception is when
fb73857a
PP
2987working with an unopened filehandle is actually what you want to do.
2988
cf264981 2989As a special case the 3-arg form with a read/write mode and the third
ed53a2bb 2990argument being C<undef>:
b76cc8ba
NIS
2991
2992 open(TMP, "+>", undef) or die ...
2993
f253e835
JH
2994opens a filehandle to an anonymous temporary file. Also using "+<"
2995works for symmetry, but you really should consider writing something
2996to the temporary file first. You will need to seek() to do the
2997reading.
b76cc8ba 2998
2ce64696 2999Since v5.8.0, perl has built using PerlIO by default. Unless you've
28a5cf3b 3000changed this (i.e. Configure -Uuseperlio), you can open file handles to
2ce64696 3001"in memory" files held in Perl scalars via:
ba964c95 3002
b996200f
SB
3003 open($fh, '>', \$variable) || ..
3004
3005Though if you try to re-open C<STDOUT> or C<STDERR> as an "in memory"
3006file, you have to close it first:
3007
3008 close STDOUT;
3009 open STDOUT, '>', \$variable or die "Can't open STDOUT: $!";
ba964c95 3010
cb1a09d0 3011Examples:
a0d0e21e
LW
3012
3013 $ARTICLE = 100;
3014 open ARTICLE or die "Can't find article $ARTICLE: $!\n";
3015 while (<ARTICLE>) {...
3016
6170680b 3017 open(LOG, '>>/usr/spool/news/twitlog'); # (log is reserved)
fb73857a 3018 # if the open fails, output is discarded
a0d0e21e 3019
6170680b 3020 open(DBASE, '+<', 'dbase.mine') # open for update
fb73857a 3021 or die "Can't open 'dbase.mine' for update: $!";
cb1a09d0 3022
6170680b
IZ
3023 open(DBASE, '+<dbase.mine') # ditto
3024 or die "Can't open 'dbase.mine' for update: $!";
3025
3026 open(ARTICLE, '-|', "caesar <$article") # decrypt article
fb73857a 3027 or die "Can't start caesar: $!";
a0d0e21e 3028
6170680b
IZ
3029 open(ARTICLE, "caesar <$article |") # ditto
3030 or die "Can't start caesar: $!";
3031
2359510d 3032 open(EXTRACT, "|sort >Tmp$$") # $$ is our process id
fb73857a 3033 or die "Can't start sort: $!";
a0d0e21e 3034
ba964c95
T
3035 # in memory files
3036 open(MEMORY,'>', \$var)
3037 or die "Can't open memory file: $!";
3038 print MEMORY "foo!\n"; # output will end up in $var
3039
a0d0e21e
LW
3040 # process argument list of files along with any includes
3041
3042 foreach $file (@ARGV) {
3043 process($file, 'fh00');
3044 }
3045
3046 sub process {
5a964f20 3047 my($filename, $input) = @_;
a0d0e21e
LW
3048 $input++; # this is a string increment
3049 unless (open($input, $filename)) {
3050 print STDERR "Can't open $filename: $!\n";
3051 return;
3052 }
3053
5a964f20 3054 local $_;
a0d0e21e
LW
3055 while (<$input>) { # note use of indirection
3056 if (/^#include "(.*)"/) {
3057 process($1, $input);
3058 next;
3059 }
5a964f20 3060 #... # whatever
a0d0e21e
LW
3061 }
3062 }
3063
ae4c5402 3064See L<perliol> for detailed info on PerlIO.
2ce64696 3065
a0d0e21e 3066You may also, in the Bourne shell tradition, specify an EXPR beginning
00cafafa
JH
3067with C<< '>&' >>, in which case the rest of the string is interpreted
3068as the name of a filehandle (or file descriptor, if numeric) to be
3069duped (as L<dup(2)>) and opened. You may use C<&> after C<< > >>,
3070C<<< >> >>>, C<< < >>, C<< +> >>, C<<< +>> >>>, and C<< +< >>.
3071The mode you specify should match the mode of the original filehandle.
3072(Duping a filehandle does not take into account any existing contents
cf264981 3073of IO buffers.) If you use the 3-arg form then you can pass either a
00cafafa 3074number, the name of a filehandle or the normal "reference to a glob".
6170680b 3075
eae1b76b
SB
3076Here is a script that saves, redirects, and restores C<STDOUT> and
3077C<STDERR> using various methods:
a0d0e21e
LW
3078
3079 #!/usr/bin/perl
eae1b76b
SB
3080 open my $oldout, ">&STDOUT" or die "Can't dup STDOUT: $!";
3081 open OLDERR, ">&", \*STDERR or die "Can't dup STDERR: $!";
818c4caa 3082
eae1b76b
SB
3083 open STDOUT, '>', "foo.out" or die "Can't redirect STDOUT: $!";
3084 open STDERR, ">&STDOUT" or die "Can't dup STDOUT: $!";
a0d0e21e 3085
eae1b76b
SB
3086 select STDERR; $| = 1; # make unbuffered
3087 select STDOUT; $| = 1; # make unbuffered
a0d0e21e
LW
3088
3089 print STDOUT "stdout 1\n"; # this works for
3090 print STDERR "stderr 1\n"; # subprocesses too
3091
eae1b76b
SB
3092 open STDOUT, ">&", $oldout or die "Can't dup \$oldout: $!";
3093 open STDERR, ">&OLDERR" or die "Can't dup OLDERR: $!";
a0d0e21e
LW
3094
3095 print STDOUT "stdout 2\n";
3096 print STDERR "stderr 2\n";
3097
ef8b303f
JH
3098If you specify C<< '<&=X' >>, where C<X> is a file descriptor number
3099or a filehandle, then Perl will do an equivalent of C's C<fdopen> of
3100that file descriptor (and not call L<dup(2)>); this is more
3101parsimonious of file descriptors. For example:
a0d0e21e 3102
00cafafa 3103 # open for input, reusing the fileno of $fd
a0d0e21e 3104 open(FILEHANDLE, "<&=$fd")
df632fdf 3105
b76cc8ba 3106or
df632fdf 3107
b76cc8ba 3108 open(FILEHANDLE, "<&=", $fd)
a0d0e21e 3109
00cafafa
JH
3110or
3111
3112 # open for append, using the fileno of OLDFH
3113 open(FH, ">>&=", OLDFH)
3114
3115or
3116
3117 open(FH, ">>&=OLDFH")
3118
ef8b303f
JH
3119Being parsimonious on filehandles is also useful (besides being
3120parsimonious) for example when something is dependent on file
3121descriptors, like for example locking using flock(). If you do just
3122C<< open(A, '>>&B') >>, the filehandle A will not have the same file
3123descriptor as B, and therefore flock(A) will not flock(B), and vice
3124versa. But with C<< open(A, '>>&=B') >> the filehandles will share
3125the same file descriptor.
3126
3127Note that if you are using Perls older than 5.8.0, Perl will be using
3128the standard C libraries' fdopen() to implement the "=" functionality.
3129On many UNIX systems fdopen() fails when file descriptors exceed a
3130certain value, typically 255. For Perls 5.8.0 and later, PerlIO is
3131most often the default.
4af147f6 3132
df632fdf
JH
3133You can see whether Perl has been compiled with PerlIO or not by
3134running C<perl -V> and looking for C<useperlio=> line. If C<useperlio>
3135is C<define>, you have PerlIO, otherwise you don't.
3136
6170680b
IZ
3137If you open a pipe on the command C<'-'>, i.e., either C<'|-'> or C<'-|'>
3138with 2-arguments (or 1-argument) form of open(), then
a0d0e21e 3139there is an implicit fork done, and the return value of open is the pid
7660c0ab 3140of the child within the parent process, and C<0> within the child
184e9718 3141process. (Use C<defined($pid)> to determine whether the open was successful.)
a0d0e21e
LW
3142The filehandle behaves normally for the parent, but i/o to that
3143filehandle is piped from/to the STDOUT/STDIN of the child process.
3144In the child process the filehandle isn't opened--i/o happens from/to
3145the new STDOUT or STDIN. Typically this is used like the normal
3146piped open when you want to exercise more control over just how the
3147pipe command gets executed, such as when you are running setuid, and
54310121 3148don't want to have to scan shell commands for metacharacters.
6170680b 3149The following triples are more or less equivalent:
a0d0e21e
LW
3150
3151 open(FOO, "|tr '[a-z]' '[A-Z]'");
6170680b
IZ
3152 open(FOO, '|-', "tr '[a-z]' '[A-Z]'");
3153 open(FOO, '|-') || exec 'tr', '[a-z]', '[A-Z]';
b76cc8ba 3154 open(FOO, '|-', "tr", '[a-z]', '[A-Z]');
a0d0e21e
LW
3155
3156 open(FOO, "cat -n '$file'|");
6170680b
IZ
3157 open(FOO, '-|', "cat -n '$file'");
3158 open(FOO, '-|') || exec 'cat', '-n', $file;
b76cc8ba
NIS
3159 open(FOO, '-|', "cat", '-n', $file);
3160
3161The last example in each block shows the pipe as "list form", which is
64da03b2
JH
3162not yet supported on all platforms. A good rule of thumb is that if
3163your platform has true C<fork()> (in other words, if your platform is
3164UNIX) you can use the list form.
a0d0e21e 3165
4633a7c4
LW
3166See L<perlipc/"Safe Pipe Opens"> for more examples of this.
3167
0f897271
GS
3168Beginning with v5.6.0, Perl will attempt to flush all files opened for
3169output before any operation that may do a fork, but this may not be
3170supported on some platforms (see L<perlport>). To be safe, you may need
3171to set C<$|> ($AUTOFLUSH in English) or call the C<autoflush()> method
3172of C<IO::Handle> on any open handles.
3173
ed53a2bb
JH
3174On systems that support a close-on-exec flag on files, the flag will
3175be set for the newly opened file descriptor as determined by the value
3176of $^F. See L<perlvar/$^F>.
a0d0e21e 3177
0dccf244 3178Closing any piped filehandle causes the parent process to wait for the
e5218da5
GA
3179child to finish, and returns the status value in C<$?> and
3180C<${^CHILD_ERROR_NATIVE}>.
0dccf244 3181
ed53a2bb
JH
3182The filename passed to 2-argument (or 1-argument) form of open() will
3183have leading and trailing whitespace deleted, and the normal
3184redirection characters honored. This property, known as "magic open",
5a964f20 3185can often be used to good effect. A user could specify a filename of
7660c0ab 3186F<"rsh cat file |">, or you could change certain filenames as needed:
5a964f20
TC
3187
3188 $filename =~ s/(.*\.gz)\s*$/gzip -dc < $1|/;
3189 open(FH, $filename) or die "Can't open $filename: $!";
3190
6170680b
IZ
3191Use 3-argument form to open a file with arbitrary weird characters in it,
3192
3193 open(FOO, '<', $file);
3194
3195otherwise it's necessary to protect any leading and trailing whitespace:
5a964f20
TC
3196
3197 $file =~ s#^(\s)#./$1#;
3198 open(FOO, "< $file\0");
3199
a31a806a 3200(this may not work on some bizarre filesystems). One should
106325ad 3201conscientiously choose between the I<magic> and 3-arguments form
6170680b
IZ
3202of open():
3203
3204 open IN, $ARGV[0];
3205
3206will allow the user to specify an argument of the form C<"rsh cat file |">,
3207but will not work on a filename which happens to have a trailing space, while
3208
3209 open IN, '<', $ARGV[0];
3210
3211will have exactly the opposite restrictions.
3212
19799a22 3213If you want a "real" C C<open> (see L<open(2)> on your system), then you
6170680b
IZ
3214should use the C<sysopen> function, which involves no such magic (but
3215may use subtly different filemodes than Perl open(), which is mapped
3216to C fopen()). This is
5a964f20
TC
3217another way to protect your filenames from interpretation. For example:
3218
3219 use IO::Handle;
3220 sysopen(HANDLE, $path, O_RDWR|O_CREAT|O_EXCL)
3221 or die "sysopen $path: $!";
3222 $oldfh = select(HANDLE); $| = 1; select($oldfh);
38762f02 3223 print HANDLE "stuff $$\n";
5a964f20
TC
3224 seek(HANDLE, 0, 0);
3225 print "File contains: ", <HANDLE>;
3226
7660c0ab
A
3227Using the constructor from the C<IO::Handle> package (or one of its
3228subclasses, such as C<IO::File> or C<IO::Socket>), you can generate anonymous
5a964f20
TC
3229filehandles that have the scope of whatever variables hold references to
3230them, and automatically close whenever and however you leave that scope:
c07a80fd 3231
5f05dabc<