1PERLVAR(1) Perl Programmers Reference Guide PERLVAR(1)
2
6 perlvar - Perl predefined variables
7
9 The Syntax of Variable Names
10 Variable names in Perl can have several formats. Usually, they must
11 begin with a letter or underscore, in which case they can be
12 arbitrarily long (up to an internal limit of 251 characters) and may
13 contain letters, digits, underscores, or the special sequence "::" or
14 "'". In this case, the part before the last "::" or "'" is taken to be
15 a package qualifier; see perlmod. A Unicode letter that is not ASCII
16 is not considered to be a letter unless "use utf8" is in effect, and
17 somewhat more complicated rules apply; see "Identifier parsing" in
18 perldata for details.
19 Perl variable names may also be a sequence of digits, a single
21 punctuation character, or the two-character sequence: "^" (caret or
22 CIRCUMFLEX ACCENT) followed by any one of the characters "[][A-Z^_?\]".
23 These names are all reserved for special uses by Perl; for example, the
24 all-digits names are used to hold data captured by backreferences after
25 a regular expression match.
26 Since Perl v5.6.0, Perl variable names may also be alphanumeric strings
28 preceded by a caret. These must all be written in the form "${^Foo}";
29 the braces are not optional. "${^Foo}" denotes the scalar variable
30 whose name is considered to be a control-"F" followed by two "o"'s.
31 These variables are reserved for future special uses by Perl, except
32 for the ones that begin with "^_" (caret-underscore). No name that
33 begins with "^_" will acquire a special meaning in any future version
34 of Perl; such names may therefore be used safely in programs. $^_
35 itself, however, is reserved.
36 Perl identifiers that begin with digits or punctuation characters are
38 exempt from the effects of the "package" declaration and are always
39 forced to be in package "main"; they are also exempt from "strict
40 'vars'" errors. A few other names are also exempt in these ways:
41 ENV STDIN
43 INC STDOUT
44 ARGV STDERR
45 ARGVOUT
46 SIG
47 In particular, the special "${^_XYZ}" variables are always taken to be
49 in package "main", regardless of any "package" declarations presently
50 in scope.
51
53 The following names have special meaning to Perl. Most punctuation
54 names have reasonable mnemonics, or analogs in the shells.
55 Nevertheless, if you wish to use long variable names, you need only
56 say:
57 use English;
59 at the top of your program. This aliases all the short names to the
61 long names in the current package. Some even have medium names,
62 generally borrowed from awk. For more info, please see English.
63 Before you continue, note the sort order for variables. In general, we
65 first list the variables in case-insensitive, almost-lexigraphical
66 order (ignoring the "{" or "^" preceding words, as in "${^UNICODE}" or
67 $^T), although $_ and @_ move up to the top of the pile. For variables
68 with the same identifier, we list it in order of scalar, array, hash,
69 and bareword.
70 General Variables
72 $ARG
73 $_ The default input and pattern-searching space. The following
74 pairs are equivalent:
75 while (<>) {...} # equivalent only in while!
77 while (defined($_ = <>)) {...}
78 /^Subject:/
80 $_ =~ /^Subject:/
81 tr/a-z/A-Z/
83 $_ =~ tr/a-z/A-Z/
84 chomp
86 chomp($_)
87 Here are the places where Perl will assume $_ even if you don't
89 use it:
90 · The following functions use $_ as a default argument:
92 abs, alarm, chomp, chop, chr, chroot, cos, defined, eval,
94 evalbytes, exp, fc, glob, hex, int, lc, lcfirst, length,
95 log, lstat, mkdir, oct, ord, pos, print, printf, quotemeta,
96 readlink, readpipe, ref, require, reverse (in scalar context
97 only), rmdir, say, sin, split (for its second argument),
98 sqrt, stat, study, uc, ucfirst, unlink, unpack.
99 · All file tests ("-f", "-d") except for "-t", which defaults
101 to STDIN. See "-X" in perlfunc
102 · The pattern matching operations "m//", "s///" and "tr///"
104 (aka "y///") when used without an "=~" operator.
105 · The default iterator variable in a "foreach" loop if no
107 other variable is supplied.
108 · The implicit iterator variable in the "grep()" and "map()"
110 functions.
111 · The implicit variable of "given()".
113 · The default place to put the next value or input record when
115 a "<FH>", "readline", "readdir" or "each" operation's result
116 is tested by itself as the sole criterion of a "while" test.
117 Outside a "while" test, this will not happen.
118 $_ is a global variable.
120 However, between perl v5.10.0 and v5.24.0, it could be used
122 lexically by writing "my $_". Making $_ refer to the global $_
123 in the same scope was then possible with "our $_". This
124 experimental feature was removed and is now a fatal error, but
125 you may encounter it in older code.
126 Mnemonic: underline is understood in certain operations.
128 @ARG
130 @_ Within a subroutine the array @_ contains the parameters passed
131 to that subroutine. Inside a subroutine, @_ is the default
132 array for the array operators "pop" and "shift".
133 See perlsub.
135 $LIST_SEPARATOR
137 $" When an array or an array slice is interpolated into a double-
138 quoted string or a similar context such as "/.../", its
139 elements are separated by this value. Default is a space. For
140 example, this:
141 print "The array is: @array\n";
143 is equivalent to this:
145 print "The array is: " . join($", @array) . "\n";
147 Mnemonic: works in double-quoted context.
149 $PROCESS_ID
151 $PID
152 $$ The process number of the Perl running this script. Though you
153 can set this variable, doing so is generally discouraged,
154 although it can be invaluable for some testing purposes. It
155 will be reset automatically across "fork()" calls.
156 Note for Linux and Debian GNU/kFreeBSD users: Before Perl
158 v5.16.0 perl would emulate POSIX semantics on Linux systems
159 using LinuxThreads, a partial implementation of POSIX Threads
160 that has since been superseded by the Native POSIX Thread
161 Library (NPTL).
162 LinuxThreads is now obsolete on Linux, and caching "getpid()"
164 like this made embedding perl unnecessarily complex (since
165 you'd have to manually update the value of $$), so now $$ and
166 "getppid()" will always return the same values as the
167 underlying C library.
168 Debian GNU/kFreeBSD systems also used LinuxThreads up until and
170 including the 6.0 release, but after that moved to FreeBSD
171 thread semantics, which are POSIX-like.
172 To see if your system is affected by this discrepancy check if
174 "getconf GNU_LIBPTHREAD_VERSION | grep -q NPTL" returns a false
175 value. NTPL threads preserve the POSIX semantics.
176 Mnemonic: same as shells.
178 $PROGRAM_NAME
180 $0 Contains the name of the program being executed.
181 On some (but not all) operating systems assigning to $0
183 modifies the argument area that the "ps" program sees. On some
184 platforms you may have to use special "ps" options or a
185 different "ps" to see the changes. Modifying the $0 is more
186 useful as a way of indicating the current program state than it
187 is for hiding the program you're running.
188 Note that there are platform-specific limitations on the
190 maximum length of $0. In the most extreme case it may be
191 limited to the space occupied by the original $0.
192 In some platforms there may be arbitrary amount of padding, for
194 example space characters, after the modified name as shown by
195 "ps". In some platforms this padding may extend all the way to
196 the original length of the argument area, no matter what you do
197 (this is the case for example with Linux 2.2).
198 Note for BSD users: setting $0 does not completely remove
200 "perl" from the ps(1) output. For example, setting $0 to
201 "foobar" may result in "perl: foobar (perl)" (whether both the
202 "perl: " prefix and the " (perl)" suffix are shown depends on
203 your exact BSD variant and version). This is an operating
204 system feature, Perl cannot help it.
205 In multithreaded scripts Perl coordinates the threads so that
207 any thread may modify its copy of the $0 and the change becomes
208 visible to ps(1) (assuming the operating system plays along).
209 Note that the view of $0 the other threads have will not change
210 since they have their own copies of it.
211 If the program has been given to perl via the switches "-e" or
213 "-E", $0 will contain the string "-e".
214 On Linux as of perl v5.14.0 the legacy process name will be set
216 with prctl(2), in addition to altering the POSIX name via
217 "argv[0]" as perl has done since version 4.000. Now system
218 utilities that read the legacy process name such as ps, top and
219 killall will recognize the name you set when assigning to $0.
220 The string you supply will be cut off at 16 bytes, this is a
221 limitation imposed by Linux.
222 Mnemonic: same as sh and ksh.
224 $REAL_GROUP_ID
226 $GID
227 $( The real gid of this process. If you are on a machine that
228 supports membership in multiple groups simultaneously, gives a
229 space separated list of groups you are in. The first number is
230 the one returned by "getgid()", and the subsequent ones by
231 "getgroups()", one of which may be the same as the first
232 number.
233 However, a value assigned to $( must be a single number used to
235 set the real gid. So the value given by $( should not be
236 assigned back to $( without being forced numeric, such as by
237 adding zero. Note that this is different to the effective gid
238 ($)) which does take a list.
239 You can change both the real gid and the effective gid at the
241 same time by using "POSIX::setgid()". Changes to $( require a
242 check to $! to detect any possible errors after an attempted
243 change.
244 Mnemonic: parentheses are used to group things. The real gid
246 is the group you left, if you're running setgid.
247 $EFFECTIVE_GROUP_ID
249 $EGID
250 $) The effective gid of this process. If you are on a machine
251 that supports membership in multiple groups simultaneously,
252 gives a space separated list of groups you are in. The first
253 number is the one returned by "getegid()", and the subsequent
254 ones by "getgroups()", one of which may be the same as the
255 first number.
256 Similarly, a value assigned to $) must also be a space-
258 separated list of numbers. The first number sets the effective
259 gid, and the rest (if any) are passed to "setgroups()". To get
260 the effect of an empty list for "setgroups()", just repeat the
261 new effective gid; that is, to force an effective gid of 5 and
262 an effectively empty "setgroups()" list, say " $) = "5 5" ".
263 You can change both the effective gid and the real gid at the
265 same time by using "POSIX::setgid()" (use only a single numeric
266 argument). Changes to $) require a check to $! to detect any
267 possible errors after an attempted change.
268 $<, $>, $( and $) can be set only on machines that support the
270 corresponding set[re][ug]id() routine. $( and $) can be
271 swapped only on machines supporting "setregid()".
272 Mnemonic: parentheses are used to group things. The effective
274 gid is the group that's right for you, if you're running
275 setgid.
276 $REAL_USER_ID
278 $UID
279 $< The real uid of this process. You can change both the real uid
280 and the effective uid at the same time by using
281 "POSIX::setuid()". Since changes to $< require a system call,
282 check $! after a change attempt to detect any possible errors.
283 Mnemonic: it's the uid you came from, if you're running setuid.
285 $EFFECTIVE_USER_ID
287 $EUID
288 $> The effective uid of this process. For example:
289 $< = $>; # set real to effective uid
291 ($<,$>) = ($>,$<); # swap real and effective uids
292 You can change both the effective uid and the real uid at the
294 same time by using "POSIX::setuid()". Changes to $> require a
295 check to $! to detect any possible errors after an attempted
296 change.
297 $< and $> can be swapped only on machines supporting
299 "setreuid()".
300 Mnemonic: it's the uid you went to, if you're running setuid.
302 $SUBSCRIPT_SEPARATOR
304 $SUBSEP
305 $; The subscript separator for multidimensional array emulation.
306 If you refer to a hash element as
307 $foo{$x,$y,$z}
309 it really means
311 $foo{join($;, $x, $y, $z)}
313 But don't put
315 @foo{$x,$y,$z} # a slice--note the @
317 which means
319 ($foo{$x},$foo{$y},$foo{$z})
321 Default is "\034", the same as SUBSEP in awk. If your keys
323 contain binary data there might not be any safe value for $;.
324 Consider using "real" multidimensional arrays as described in
326 perllol.
327 Mnemonic: comma (the syntactic subscript separator) is a semi-
329 semicolon.
330 $a
332 $b Special package variables when using "sort()", see "sort" in
333 perlfunc. Because of this specialness $a and $b don't need to
334 be declared (using "use vars", or "our()") even when using the
335 "strict 'vars'" pragma. Don't lexicalize them with "my $a" or
336 "my $b" if you want to be able to use them in the "sort()"
337 comparison block or function.
338 %ENV The hash %ENV contains your current environment. Setting a
340 value in "ENV" changes the environment for any child processes
341 you subsequently "fork()" off.
342 As of v5.18.0, both keys and values stored in %ENV are
344 stringified.
345 my $foo = 1;
347 $ENV{'bar'} = \$foo;
348 if( ref $ENV{'bar'} ) {
349 say "Pre 5.18.0 Behaviour";
350 } else {
351 say "Post 5.18.0 Behaviour";
352 }
353 Previously, only child processes received stringified values:
355 my $foo = 1;
357 $ENV{'bar'} = \$foo;
358 # Always printed 'non ref'
360 system($^X, '-e',
361 q/print ( ref $ENV{'bar'} ? 'ref' : 'non ref' ) /);
362 This happens because you can't really share arbitrary data
364 structures with foreign processes.
365 $OLD_PERL_VERSION
367 $] The revision, version, and subversion of the Perl interpreter,
368 represented as a decimal of the form 5.XXXYYY, where XXX is the
369 version / 1e3 and YYY is the subversion / 1e6. For example,
370 Perl v5.10.1 would be "5.010001".
371 This variable can be used to determine whether the Perl
373 interpreter executing a script is in the right range of
374 versions:
375 warn "No PerlIO!\n" if $] lt '5.008';
377 When comparing $], string comparison operators are highly
379 recommended. The inherent limitations of binary floating point
380 representation can sometimes lead to incorrect comparisons for
381 some numbers on some architectures.
382 See also the documentation of "use VERSION" and "require
384 VERSION" for a convenient way to fail if the running Perl
385 interpreter is too old.
386 See "$^V" for a representation of the Perl version as a version
388 object, which allows more flexible string comparisons.
389 The main advantage of $] over $^V is that it works the same on
391 any version of Perl. The disadvantages are that it can't
392 easily be compared to versions in other formats (e.g. literal
393 v-strings, "v1.2.3" or version objects) and numeric comparisons
394 can occasionally fail; it's good for string literal version
395 checks and bad for comparing to a variable that hasn't been
396 sanity-checked.
397 The $OLD_PERL_VERSION form was added in Perl v5.20.0 for
399 historical reasons but its use is discouraged. (If your reason
400 to use $] is to run code on old perls then referring to it as
401 $OLD_PERL_VERSION would be self-defeating.)
402 Mnemonic: Is this version of perl in the right bracket?
404 $SYSTEM_FD_MAX
406 $^F The maximum system file descriptor, ordinarily 2. System file
407 descriptors are passed to "exec()"ed processes, while higher
408 file descriptors are not. Also, during an "open()", system
409 file descriptors are preserved even if the "open()" fails
410 (ordinary file descriptors are closed before the "open()" is
411 attempted). The close-on-exec status of a file descriptor will
412 be decided according to the value of $^F when the corresponding
413 file, pipe, or socket was opened, not the time of the "exec()".
414 @F The array @F contains the fields of each line read in when
416 autosplit mode is turned on. See perlrun for the -a switch.
417 This array is package-specific, and must be declared or given a
418 full package name if not in package main when running under
419 "strict 'vars'".
420 @INC The array @INC contains the list of places that the "do EXPR",
422 "require", or "use" constructs look for their library files.
423 It initially consists of the arguments to any -I command-line
424 switches, followed by the default Perl library, probably
425 /usr/local/lib/perl, followed by ".", to represent the current
426 directory. ("." will not be appended if taint checks are
427 enabled, either by "-T" or by "-t", or if configured not to do
428 so by the "-Ddefault_inc_excludes_dot" compile time option.)
429 If you need to modify this at runtime, you should use the "use
430 lib" pragma to get the machine-dependent library properly
431 loaded also:
432 use lib '/mypath/libdir/';
434 use SomeMod;
435 You can also insert hooks into the file inclusion system by
437 putting Perl code directly into @INC. Those hooks may be
438 subroutine references, array references or blessed objects.
439 See "require" in perlfunc for details.
440 %INC The hash %INC contains entries for each filename included via
442 the "do", "require", or "use" operators. The key is the
443 filename you specified (with module names converted to
444 pathnames), and the value is the location of the file found.
445 The "require" operator uses this hash to determine whether a
446 particular file has already been included.
447 If the file was loaded via a hook (e.g. a subroutine reference,
449 see "require" in perlfunc for a description of these hooks),
450 this hook is by default inserted into %INC in place of a
451 filename. Note, however, that the hook may have set the %INC
452 entry by itself to provide some more specific info.
453 $INPLACE_EDIT
455 $^I The current value of the inplace-edit extension. Use "undef"
456 to disable inplace editing.
457 Mnemonic: value of -i switch.
459 @ISA Each package contains a special array called @ISA which
461 contains a list of that class's parent classes, if any. This
462 array is simply a list of scalars, each of which is a string
463 that corresponds to a package name. The array is examined when
464 Perl does method resolution, which is covered in perlobj.
465 To load packages while adding them to @ISA, see the parent
467 pragma. The discouraged base pragma does this as well, but
468 should not be used except when compatibility with the
469 discouraged fields pragma is required.
470 $^M By default, running out of memory is an untrappable, fatal
472 error. However, if suitably built, Perl can use the contents
473 of $^M as an emergency memory pool after "die()"ing. Suppose
474 that your Perl were compiled with "-DPERL_EMERGENCY_SBRK" and
475 used Perl's malloc. Then
476 $^M = 'a' x (1 << 16);
478 would allocate a 64K buffer for use in an emergency. See the
480 INSTALL file in the Perl distribution for information on how to
481 add custom C compilation flags when compiling perl. To
482 discourage casual use of this advanced feature, there is no
483 English long name for this variable.
484 This variable was added in Perl 5.004.
486 $OSNAME
488 $^O The name of the operating system under which this copy of Perl
489 was built, as determined during the configuration process. For
490 examples see "PLATFORMS" in perlport.
491 The value is identical to $Config{'osname'}. See also Config
493 and the -V command-line switch documented in perlrun.
494 In Windows platforms, $^O is not very helpful: since it is
496 always "MSWin32", it doesn't tell the difference between
497 95/98/ME/NT/2000/XP/CE/.NET. Use "Win32::GetOSName()" or
498 Win32::GetOSVersion() (see Win32 and perlport) to distinguish
499 between the variants.
500 This variable was added in Perl 5.003.
502 %SIG The hash %SIG contains signal handlers for signals. For
504 example:
505 sub handler { # 1st argument is signal name
507 my($sig) = @_;
508 print "Caught a SIG$sig--shutting down\n";
509 close(LOG);
510 exit(0);
511 }
512 $SIG{'INT'} = \&handler;
514 $SIG{'QUIT'} = \&handler;
515 ...
516 $SIG{'INT'} = 'DEFAULT'; # restore default action
517 $SIG{'QUIT'} = 'IGNORE'; # ignore SIGQUIT
518 Using a value of 'IGNORE' usually has the effect of ignoring
520 the signal, except for the "CHLD" signal. See perlipc for more
521 about this special case.
522 Here are some other examples:
524 $SIG{"PIPE"} = "Plumber"; # assumes main::Plumber (not
526 # recommended)
527 $SIG{"PIPE"} = \&Plumber; # just fine; assume current
528 # Plumber
529 $SIG{"PIPE"} = *Plumber; # somewhat esoteric
530 $SIG{"PIPE"} = Plumber(); # oops, what did Plumber()
531 # return??
532 Be sure not to use a bareword as the name of a signal handler,
534 lest you inadvertently call it.
535 If your system has the "sigaction()" function then signal
537 handlers are installed using it. This means you get reliable
538 signal handling.
539 The default delivery policy of signals changed in Perl v5.8.0
541 from immediate (also known as "unsafe") to deferred, also known
542 as "safe signals". See perlipc for more information.
543 Certain internal hooks can be also set using the %SIG hash.
545 The routine indicated by $SIG{__WARN__} is called when a
546 warning message is about to be printed. The warning message is
547 passed as the first argument. The presence of a "__WARN__"
548 hook causes the ordinary printing of warnings to "STDERR" to be
549 suppressed. You can use this to save warnings in a variable,
550 or turn warnings into fatal errors, like this:
551 local $SIG{__WARN__} = sub { die $_[0] };
553 eval $proggie;
554 As the 'IGNORE' hook is not supported by "__WARN__", you can
556 disable warnings using the empty subroutine:
557 local $SIG{__WARN__} = sub {};
559 The routine indicated by $SIG{__DIE__} is called when a fatal
561 exception is about to be thrown. The error message is passed
562 as the first argument. When a "__DIE__" hook routine returns,
563 the exception processing continues as it would have in the
564 absence of the hook, unless the hook routine itself exits via a
565 "goto &sub", a loop exit, or a "die()". The "__DIE__" handler
566 is explicitly disabled during the call, so that you can die
567 from a "__DIE__" handler. Similarly for "__WARN__".
568 The $SIG{__DIE__} hook is called even inside an "eval()". It
570 was never intended to happen this way, but an implementation
571 glitch made this possible. This used to be deprecated, as it
572 allowed strange action at a distance like rewriting a pending
573 exception in $@. Plans to rectify this have been scrapped, as
574 users found that rewriting a pending exception is actually a
575 useful feature, and not a bug.
576 "__DIE__"/"__WARN__" handlers are very special in one respect:
578 they may be called to report (probable) errors found by the
579 parser. In such a case the parser may be in inconsistent
580 state, so any attempt to evaluate Perl code from such a handler
581 will probably result in a segfault. This means that warnings
582 or errors that result from parsing Perl should be used with
583 extreme caution, like this:
584 require Carp if defined $^S;
586 Carp::confess("Something wrong") if defined &Carp::confess;
587 die "Something wrong, but could not load Carp to give "
588 . "backtrace...\n\t"
589 . "To see backtrace try starting Perl with -MCarp switch";
590 Here the first line will load "Carp" unless it is the parser
592 who called the handler. The second line will print backtrace
593 and die if "Carp" was available. The third line will be
594 executed only if "Carp" was not available.
595 Having to even think about the $^S variable in your exception
597 handlers is simply wrong. $SIG{__DIE__} as currently
598 implemented invites grievous and difficult to track down
599 errors. Avoid it and use an "END{}" or CORE::GLOBAL::die
600 override instead.
601 See "die" in perlfunc, "warn" in perlfunc, "eval" in perlfunc,
603 and warnings for additional information.
604 $BASETIME
606 $^T The time at which the program began running, in seconds since
607 the epoch (beginning of 1970). The values returned by the -M,
608 -A, and -C filetests are based on this value.
609 $PERL_VERSION
611 $^V The revision, version, and subversion of the Perl interpreter,
612 represented as a version object.
613 This variable first appeared in perl v5.6.0; earlier versions
615 of perl will see an undefined value. Before perl v5.10.0 $^V
616 was represented as a v-string rather than a version object.
617 $^V can be used to determine whether the Perl interpreter
619 executing a script is in the right range of versions. For
620 example:
621 warn "Hashes not randomized!\n" if !$^V or $^V lt v5.8.1
623 While version objects overload stringification, to portably
625 convert $^V into its string representation, use "sprintf()"'s
626 "%vd" conversion, which works for both v-strings or version
627 objects:
628 printf "version is v%vd\n", $^V; # Perl's version
630 See the documentation of "use VERSION" and "require VERSION"
632 for a convenient way to fail if the running Perl interpreter is
633 too old.
634 See also "$]" for a decimal representation of the Perl version.
636 The main advantage of $^V over $] is that, for Perl v5.10.0 or
638 later, it overloads operators, allowing easy comparison against
639 other version representations (e.g. decimal, literal v-string,
640 "v1.2.3", or objects). The disadvantage is that prior to
641 v5.10.0, it was only a literal v-string, which can't be easily
642 printed or compared, whereas the behavior of $] is unchanged on
643 all versions of Perl.
644 Mnemonic: use ^V for a version object.
646 ${^WIN32_SLOPPY_STAT}
648 If this variable is set to a true value, then "stat()" on
649 Windows will not try to open the file. This means that the
650 link count cannot be determined and file attributes may be out
651 of date if additional hardlinks to the file exist. On the
652 other hand, not opening the file is considerably faster,
653 especially for files on network drives.
654 This variable could be set in the sitecustomize.pl file to
656 configure the local Perl installation to use "sloppy" "stat()"
657 by default. See the documentation for -f in perlrun for more
658 information about site customization.
659 This variable was added in Perl v5.10.0.
661 $EXECUTABLE_NAME
663 $^X The name used to execute the current copy of Perl, from C's
664 "argv[0]" or (where supported) /proc/self/exe.
665 Depending on the host operating system, the value of $^X may be
667 a relative or absolute pathname of the perl program file, or
668 may be the string used to invoke perl but not the pathname of
669 the perl program file. Also, most operating systems permit
670 invoking programs that are not in the PATH environment
671 variable, so there is no guarantee that the value of $^X is in
672 PATH. For VMS, the value may or may not include a version
673 number.
674 You usually can use the value of $^X to re-invoke an
676 independent copy of the same perl that is currently running,
677 e.g.,
678 @first_run = `$^X -le "print int rand 100 for 1..100"`;
680 But recall that not all operating systems support forking or
682 capturing of the output of commands, so this complex statement
683 may not be portable.
684 It is not safe to use the value of $^X as a path name of a
686 file, as some operating systems that have a mandatory suffix on
687 executable files do not require use of the suffix when invoking
688 a command. To convert the value of $^X to a path name, use the
689 following statements:
690 # Build up a set of file names (not command names).
692 use Config;
693 my $this_perl = $^X;
694 if ($^O ne 'VMS') {
695 $this_perl .= $Config{_exe}
696 unless $this_perl =~ m/$Config{_exe}$/i;
697 }
698 Because many operating systems permit anyone with read access
700 to the Perl program file to make a copy of it, patch the copy,
701 and then execute the copy, the security-conscious Perl
702 programmer should take care to invoke the installed copy of
703 perl, not the copy referenced by $^X. The following statements
704 accomplish this goal, and produce a pathname that can be
705 invoked as a command or referenced as a file.
706 use Config;
708 my $secure_perl_path = $Config{perlpath};
709 if ($^O ne 'VMS') {
710 $secure_perl_path .= $Config{_exe}
711 unless $secure_perl_path =~ m/$Config{_exe}$/i;
712 }
713 Variables related to regular expressions
715 Most of the special variables related to regular expressions are side
716 effects. Perl sets these variables when it has a successful match, so
717 you should check the match result before using them. For instance:
718 if( /P(A)TT(ER)N/ ) {
720 print "I found $1 and $2\n";
721 }
722 These variables are read-only and dynamically-scoped, unless we note
724 otherwise.
725 The dynamic nature of the regular expression variables means that their
727 value is limited to the block that they are in, as demonstrated by this
728 bit of code:
729 my $outer = 'Wallace and Grommit';
731 my $inner = 'Mutt and Jeff';
732 my $pattern = qr/(\S+) and (\S+)/;
734 sub show_n { print "\$1 is $1; \$2 is $2\n" }
736 {
738 OUTER:
739 show_n() if $outer =~ m/$pattern/;
740 INNER: {
742 show_n() if $inner =~ m/$pattern/;
743 }
744 show_n();
746 }
747 The output shows that while in the "OUTER" block, the values of $1 and
749 $2 are from the match against $outer. Inside the "INNER" block, the
750 values of $1 and $2 are from the match against $inner, but only until
751 the end of the block (i.e. the dynamic scope). After the "INNER" block
752 completes, the values of $1 and $2 return to the values for the match
753 against $outer even though we have not made another match:
754 $1 is Wallace; $2 is Grommit
756 $1 is Mutt; $2 is Jeff
757 $1 is Wallace; $2 is Grommit
758 Performance issues
760 Traditionally in Perl, any use of any of the three variables "$`", $&
762 or "$'" (or their "use English" equivalents) anywhere in the code,
763 caused all subsequent successful pattern matches to make a copy of the
764 matched string, in case the code might subsequently access one of those
765 variables. This imposed a considerable performance penalty across the
766 whole program, so generally the use of these variables has been
767 discouraged.
768 In Perl 5.6.0 the "@-" and "@+" dynamic arrays were introduced that
770 supply the indices of successful matches. So you could for example do
771 this:
772 $str =~ /pattern/;
774 print $`, $&, $'; # bad: performance hit
776 print # good: no performance hit
778 substr($str, 0, $-[0]),
779 substr($str, $-[0], $+[0]-$-[0]),
780 substr($str, $+[0]);
781 In Perl 5.10.0 the "/p" match operator flag and the "${^PREMATCH}",
783 "${^MATCH}", and "${^POSTMATCH}" variables were introduced, that
784 allowed you to suffer the penalties only on patterns marked with "/p".
785 In Perl 5.18.0 onwards, perl started noting the presence of each of the
787 three variables separately, and only copied that part of the string
788 required; so in
789 $`; $&; "abcdefgh" =~ /d/
791 perl would only copy the "abcd" part of the string. That could make a
793 big difference in something like
794 $str = 'x' x 1_000_000;
796 $&; # whoops
797 $str =~ /x/g # one char copied a million times, not a million chars
798 In Perl 5.20.0 a new copy-on-write system was enabled by default, which
800 finally fixes all performance issues with these three variables, and
801 makes them safe to use anywhere.
802 The "Devel::NYTProf" and "Devel::FindAmpersand" modules can help you
804 find uses of these problematic match variables in your code.
805 $<digits> ($1, $2, ...)
807 Contains the subpattern from the corresponding set of capturing
808 parentheses from the last successful pattern match, not
809 counting patterns matched in nested blocks that have been
810 exited already.
811 Note there is a distinction between a capture buffer which
813 matches the empty string a capture buffer which is optional.
814 Eg, "(x?)" and "(x)?" The latter may be undef, the former not.
815 These variables are read-only and dynamically-scoped.
817 Mnemonic: like \digits.
819 @{^CAPTURE}
821 An array which exposes the contents of the capture buffers, if
822 any, of the last successful pattern match, not counting
823 patterns matched in nested blocks that have been exited
824 already.
825 Note that the 0 index of @{^CAPTURE} is equivalent to $1, the 1
827 index is equivalent to $2, etc.
828 if ("foal"=~/(.)(.)(.)(.)/) {
830 print join "-", @{^CAPTURE};
831 }
832 should output "f-o-a-l".
834 See also "$digits", "%{^CAPTURE}" and "%{^CAPTURE_ALL}".
836 Note that unlike most other regex magic variables there is no
838 single letter equivalent to "@{^CAPTURE}".
839 This variable was added in 5.25.7
841 $MATCH
843 $& The string matched by the last successful pattern match (not
844 counting any matches hidden within a BLOCK or "eval()" enclosed
845 by the current BLOCK).
846 See "Performance issues" above for the serious performance
848 implications of using this variable (even once) in your code.
849 This variable is read-only and dynamically-scoped.
851 Mnemonic: like "&" in some editors.
853 ${^MATCH}
855 This is similar to $& ($MATCH) except that it does not incur
856 the performance penalty associated with that variable.
857 See "Performance issues" above.
859 In Perl v5.18 and earlier, it is only guaranteed to return a
861 defined value when the pattern was compiled or executed with
862 the "/p" modifier. In Perl v5.20, the "/p" modifier does
863 nothing, so "${^MATCH}" does the same thing as $MATCH.
864 This variable was added in Perl v5.10.0.
866 This variable is read-only and dynamically-scoped.
868 $PREMATCH
870 $` The string preceding whatever was matched by the last
871 successful pattern match, not counting any matches hidden
872 within a BLOCK or "eval" enclosed by the current BLOCK.
873 See "Performance issues" above for the serious performance
875 implications of using this variable (even once) in your code.
876 This variable is read-only and dynamically-scoped.
878 Mnemonic: "`" often precedes a quoted string.
880 ${^PREMATCH}
882 This is similar to "$`" ($PREMATCH) except that it does not
883 incur the performance penalty associated with that variable.
884 See "Performance issues" above.
886 In Perl v5.18 and earlier, it is only guaranteed to return a
888 defined value when the pattern was compiled or executed with
889 the "/p" modifier. In Perl v5.20, the "/p" modifier does
890 nothing, so "${^PREMATCH}" does the same thing as $PREMATCH.
891 This variable was added in Perl v5.10.0.
893 This variable is read-only and dynamically-scoped.
895 $POSTMATCH
897 $' The string following whatever was matched by the last
898 successful pattern match (not counting any matches hidden
899 within a BLOCK or "eval()" enclosed by the current BLOCK).
900 Example:
901 local $_ = 'abcdefghi';
903 /def/;
904 print "$`:$&:$'\n"; # prints abc:def:ghi
905 See "Performance issues" above for the serious performance
907 implications of using this variable (even once) in your code.
908 This variable is read-only and dynamically-scoped.
910 Mnemonic: "'" often follows a quoted string.
912 ${^POSTMATCH}
914 This is similar to "$'" ($POSTMATCH) except that it does not
915 incur the performance penalty associated with that variable.
916 See "Performance issues" above.
918 In Perl v5.18 and earlier, it is only guaranteed to return a
920 defined value when the pattern was compiled or executed with
921 the "/p" modifier. In Perl v5.20, the "/p" modifier does
922 nothing, so "${^POSTMATCH}" does the same thing as $POSTMATCH.
923 This variable was added in Perl v5.10.0.
925 This variable is read-only and dynamically-scoped.
927 $LAST_PAREN_MATCH
929 $+ The text matched by the last bracket of the last successful
930 search pattern. This is useful if you don't know which one of
931 a set of alternative patterns matched. For example:
932 /Version: (.*)|Revision: (.*)/ && ($rev = $+);
934 This variable is read-only and dynamically-scoped.
936 Mnemonic: be positive and forward looking.
938 $LAST_SUBMATCH_RESULT
940 $^N The text matched by the used group most-recently closed (i.e.
941 the group with the rightmost closing parenthesis) of the last
942 successful search pattern.
943 This is primarily used inside "(?{...})" blocks for examining
945 text recently matched. For example, to effectively capture
946 text to a variable (in addition to $1, $2, etc.), replace
947 "(...)" with
948 (?:(...)(?{ $var = $^N }))
950 By setting and then using $var in this way relieves you from
952 having to worry about exactly which numbered set of parentheses
953 they are.
954 This variable was added in Perl v5.8.0.
956 Mnemonic: the (possibly) Nested parenthesis that most recently
958 closed.
959 @LAST_MATCH_END
961 @+ This array holds the offsets of the ends of the last successful
962 submatches in the currently active dynamic scope. $+[0] is the
963 offset into the string of the end of the entire match. This is
964 the same value as what the "pos" function returns when called
965 on the variable that was matched against. The nth element of
966 this array holds the offset of the nth submatch, so $+[1] is
967 the offset past where $1 ends, $+[2] the offset past where $2
968 ends, and so on. You can use $#+ to determine how many
969 subgroups were in the last successful match. See the examples
970 given for the "@-" variable.
971 This variable was added in Perl v5.6.0.
973 %{^CAPTURE}
975 %LAST_PAREN_MATCH
976 %+ Similar to "@+", the "%+" hash allows access to the named
977 capture buffers, should they exist, in the last successful
978 match in the currently active dynamic scope.
979 For example, $+{foo} is equivalent to $1 after the following
981 match:
982 'foo' =~ /(?<foo>foo)/;
984 The keys of the "%+" hash list only the names of buffers that
986 have captured (and that are thus associated to defined values).
987 If multiple distinct capture groups have the same name, then
989 $+{NAME} will refer to the leftmost defined group in the match.
990 The underlying behaviour of "%+" is provided by the
992 Tie::Hash::NamedCapture module.
993 Note: "%-" and "%+" are tied views into a common internal hash
995 associated with the last successful regular expression.
996 Therefore mixing iterative access to them via "each" may have
997 unpredictable results. Likewise, if the last successful match
998 changes, then the results may be surprising.
999 This variable was added in Perl v5.10.0. The "%{^CAPTURE}"
1001 alias was added in 5.25.7.
1002 This variable is read-only and dynamically-scoped.
1004 @LAST_MATCH_START
1006 @- "$-[0]" is the offset of the start of the last successful
1007 match. "$-["n"]" is the offset of the start of the substring
1008 matched by n-th subpattern, or undef if the subpattern did not
1009 match.
1010 Thus, after a match against $_, $& coincides with "substr $_,
1012 $-[0], $+[0] - $-[0]". Similarly, $n coincides with "substr
1013 $_, $-[n], $+[n] - $-[n]" if "$-[n]" is defined, and $+
1014 coincides with "substr $_, $-[$#-], $+[$#-] - $-[$#-]". One
1015 can use "$#-" to find the last matched subgroup in the last
1016 successful match. Contrast with $#+, the number of subgroups
1017 in the regular expression. Compare with "@+".
1018 This array holds the offsets of the beginnings of the last
1020 successful submatches in the currently active dynamic scope.
1021 "$-[0]" is the offset into the string of the beginning of the
1022 entire match. The nth element of this array holds the offset
1023 of the nth submatch, so "$-[1]" is the offset where $1 begins,
1024 "$-[2]" the offset where $2 begins, and so on.
1025 After a match against some variable $var:
1027 "$`" is the same as "substr($var, 0, $-[0])"
1029 $& is the same as "substr($var, $-[0], $+[0] - $-[0])"
1030 "$'" is the same as "substr($var, $+[0])"
1031 $1 is the same as "substr($var, $-[1], $+[1] - $-[1])"
1032 $2 is the same as "substr($var, $-[2], $+[2] - $-[2])"
1033 $3 is the same as "substr($var, $-[3], $+[3] - $-[3])"
1034 This variable was added in Perl v5.6.0.
1036 %{^CAPTURE_ALL}
1038 %- Similar to "%+", this variable allows access to the named
1039 capture groups in the last successful match in the currently
1040 active dynamic scope. To each capture group name found in the
1041 regular expression, it associates a reference to an array
1042 containing the list of values captured by all buffers with that
1043 name (should there be several of them), in the order where they
1044 appear.
1045 Here's an example:
1047 if ('1234' =~ /(?<A>1)(?<B>2)(?<A>3)(?<B>4)/) {
1049 foreach my $bufname (sort keys %-) {
1050 my $ary = $-{$bufname};
1051 foreach my $idx (0..$#$ary) {
1052 print "\$-{$bufname}[$idx] : ",
1053 (defined($ary->[$idx])
1054 ? "'$ary->[$idx]'"
1055 : "undef"),
1056 "\n";
1057 }
1058 }
1059 }
1060 would print out:
1062 $-{A}[0] : '1'
1064 $-{A}[1] : '3'
1065 $-{B}[0] : '2'
1066 $-{B}[1] : '4'
1067 The keys of the "%-" hash correspond to all buffer names found
1069 in the regular expression.
1070 The behaviour of "%-" is implemented via the
1072 Tie::Hash::NamedCapture module.
1073 Note: "%-" and "%+" are tied views into a common internal hash
1075 associated with the last successful regular expression.
1076 Therefore mixing iterative access to them via "each" may have
1077 unpredictable results. Likewise, if the last successful match
1078 changes, then the results may be surprising.
1079 This variable was added in Perl v5.10.0. The "%{^CAPTURE_ALL}"
1081 alias was added in 5.25.7.
1082 This variable is read-only and dynamically-scoped.
1084 $LAST_REGEXP_CODE_RESULT
1086 $^R The result of evaluation of the last successful "(?{ code })"
1087 regular expression assertion (see perlre). May be written to.
1088 This variable was added in Perl 5.005.
1090 ${^RE_DEBUG_FLAGS}
1092 The current value of the regex debugging flags. Set to 0 for
1093 no debug output even when the "re 'debug'" module is loaded.
1094 See re for details.
1095 This variable was added in Perl v5.10.0.
1097 ${^RE_TRIE_MAXBUF}
1099 Controls how certain regex optimisations are applied and how
1100 much memory they utilize. This value by default is 65536 which
1101 corresponds to a 512kB temporary cache. Set this to a higher
1102 value to trade memory for speed when matching large
1103 alternations. Set it to a lower value if you want the
1104 optimisations to be as conservative of memory as possible but
1105 still occur, and set it to a negative value to prevent the
1106 optimisation and conserve the most memory. Under normal
1107 situations this variable should be of no interest to you.
1108 This variable was added in Perl v5.10.0.
1110 Variables related to filehandles
1112 Variables that depend on the currently selected filehandle may be set
1113 by calling an appropriate object method on the "IO::Handle" object,
1114 although this is less efficient than using the regular built-in
1115 variables. (Summary lines below for this contain the word HANDLE.)
1116 First you must say
1117 use IO::Handle;
1119 after which you may use either
1121 method HANDLE EXPR
1123 or more safely,
1125 HANDLE->method(EXPR)
1127 Each method returns the old value of the "IO::Handle" attribute. The
1129 methods each take an optional EXPR, which, if supplied, specifies the
1130 new value for the "IO::Handle" attribute in question. If not supplied,
1131 most methods do nothing to the current value--except for "autoflush()",
1132 which will assume a 1 for you, just to be different.
1133 Because loading in the "IO::Handle" class is an expensive operation,
1135 you should learn how to use the regular built-in variables.
1136 A few of these variables are considered "read-only". This means that
1138 if you try to assign to this variable, either directly or indirectly
1139 through a reference, you'll raise a run-time exception.
1140 You should be very careful when modifying the default values of most
1142 special variables described in this document. In most cases you want
1143 to localize these variables before changing them, since if you don't,
1144 the change may affect other modules which rely on the default values of
1145 the special variables that you have changed. This is one of the
1146 correct ways to read the whole file at once:
1147 open my $fh, "<", "foo" or die $!;
1149 local $/; # enable localized slurp mode
1150 my $content = <$fh>;
1151 close $fh;
1152 But the following code is quite bad:
1154 open my $fh, "<", "foo" or die $!;
1156 undef $/; # enable slurp mode
1157 my $content = <$fh>;
1158 close $fh;
1159 since some other module, may want to read data from some file in the
1161 default "line mode", so if the code we have just presented has been
1162 executed, the global value of $/ is now changed for any other code
1163 running inside the same Perl interpreter.
1164 Usually when a variable is localized you want to make sure that this
1166 change affects the shortest scope possible. So unless you are already
1167 inside some short "{}" block, you should create one yourself. For
1168 example:
1169 my $content = '';
1171 open my $fh, "<", "foo" or die $!;
1172 {
1173 local $/;
1174 $content = <$fh>;
1175 }
1176 close $fh;
1177 Here is an example of how your own code can go broken:
1179 for ( 1..3 ){
1181 $\ = "\r\n";
1182 nasty_break();
1183 print "$_";
1184 }
1185 sub nasty_break {
1187 $\ = "\f";
1188 # do something with $_
1189 }
1190 You probably expect this code to print the equivalent of
1192 "1\r\n2\r\n3\r\n"
1194 but instead you get:
1196 "1\f2\f3\f"
1198 Why? Because "nasty_break()" modifies "$\" without localizing it first.
1200 The value you set in "nasty_break()" is still there when you return.
1201 The fix is to add "local()" so the value doesn't leak out of
1202 "nasty_break()":
1203 local $\ = "\f";
1205 It's easy to notice the problem in such a short example, but in more
1207 complicated code you are looking for trouble if you don't localize
1208 changes to the special variables.
1209 $ARGV Contains the name of the current file when reading from "<>".
1211 @ARGV The array @ARGV contains the command-line arguments intended
1213 for the script. $#ARGV is generally the number of arguments
1214 minus one, because $ARGV[0] is the first argument, not the
1215 program's command name itself. See "$0" for the command name.
1216 ARGV The special filehandle that iterates over command-line
1218 filenames in @ARGV. Usually written as the null filehandle in
1219 the angle operator "<>". Note that currently "ARGV" only has
1220 its magical effect within the "<>" operator; elsewhere it is
1221 just a plain filehandle corresponding to the last file opened
1222 by "<>". In particular, passing "\*ARGV" as a parameter to a
1223 function that expects a filehandle may not cause your function
1224 to automatically read the contents of all the files in @ARGV.
1225 ARGVOUT The special filehandle that points to the currently open output
1227 file when doing edit-in-place processing with -i. Useful when
1228 you have to do a lot of inserting and don't want to keep
1229 modifying $_. See perlrun for the -i switch.
1230 IO::Handle->output_field_separator( EXPR )
1232 $OUTPUT_FIELD_SEPARATOR
1233 $OFS
1234 $, The output field separator for the print operator. If defined,
1235 this value is printed between each of print's arguments.
1236 Default is "undef".
1237 You cannot call "output_field_separator()" on a handle, only as
1239 a static method. See IO::Handle.
1240 Mnemonic: what is printed when there is a "," in your print
1242 statement.
1243 HANDLE->input_line_number( EXPR )
1245 $INPUT_LINE_NUMBER
1246 $NR
1247 $. Current line number for the last filehandle accessed.
1248 Each filehandle in Perl counts the number of lines that have
1250 been read from it. (Depending on the value of $/, Perl's idea
1251 of what constitutes a line may not match yours.) When a line
1252 is read from a filehandle (via "readline()" or "<>"), or when
1253 "tell()" or "seek()" is called on it, $. becomes an alias to
1254 the line counter for that filehandle.
1255 You can adjust the counter by assigning to $., but this will
1257 not actually move the seek pointer. Localizing $. will not
1258 localize the filehandle's line count. Instead, it will
1259 localize perl's notion of which filehandle $. is currently
1260 aliased to.
1261 $. is reset when the filehandle is closed, but not when an open
1263 filehandle is reopened without an intervening "close()". For
1264 more details, see "I/O Operators" in perlop. Because "<>"
1265 never does an explicit close, line numbers increase across
1266 "ARGV" files (but see examples in "eof" in perlfunc).
1267 You can also use "HANDLE->input_line_number(EXPR)" to access
1269 the line counter for a given filehandle without having to worry
1270 about which handle you last accessed.
1271 Mnemonic: many programs use "." to mean the current line
1273 number.
1274 IO::Handle->input_record_separator( EXPR )
1276 $INPUT_RECORD_SEPARATOR
1277 $RS
1278 $/ The input record separator, newline by default. This
1279 influences Perl's idea of what a "line" is. Works like awk's
1280 RS variable, including treating empty lines as a terminator if
1281 set to the null string (an empty line cannot contain any spaces
1282 or tabs). You may set it to a multi-character string to match
1283 a multi-character terminator, or to "undef" to read through the
1284 end of file. Setting it to "\n\n" means something slightly
1285 different than setting to "", if the file contains consecutive
1286 empty lines. Setting to "" will treat two or more consecutive
1287 empty lines as a single empty line. Setting to "\n\n" will
1288 blindly assume that the next input character belongs to the
1289 next paragraph, even if it's a newline.
1290 local $/; # enable "slurp" mode
1292 local $_ = <FH>; # whole file now here
1293 s/\n[ \t]+/ /g;
1294 Remember: the value of $/ is a string, not a regex. awk has to
1296 be better for something. :-)
1297 Setting $/ to a reference to an integer, scalar containing an
1299 integer, or scalar that's convertible to an integer will
1300 attempt to read records instead of lines, with the maximum
1301 record size being the referenced integer number of characters.
1302 So this:
1303 local $/ = \32768; # or \"32768", or \$var_containing_32768
1305 open my $fh, "<", $myfile or die $!;
1306 local $_ = <$fh>;
1307 will read a record of no more than 32768 characters from $fh.
1309 If you're not reading from a record-oriented file (or your OS
1310 doesn't have record-oriented files), then you'll likely get a
1311 full chunk of data with every read. If a record is larger than
1312 the record size you've set, you'll get the record back in
1313 pieces. Trying to set the record size to zero or less is
1314 deprecated and will cause $/ to have the value of "undef",
1315 which will cause reading in the (rest of the) whole file.
1316 As of 5.19.9 setting $/ to any other form of reference will
1318 throw a fatal exception. This is in preparation for supporting
1319 new ways to set $/ in the future.
1320 On VMS only, record reads bypass PerlIO layers and any
1322 associated buffering, so you must not mix record and non-record
1323 reads on the same filehandle. Record mode mixes with line mode
1324 only when the same buffering layer is in use for both modes.
1325 You cannot call "input_record_separator()" on a handle, only as
1327 a static method. See IO::Handle.
1328 See also "Newlines" in perlport. Also see "$.".
1330 Mnemonic: / delimits line boundaries when quoting poetry.
1332 IO::Handle->output_record_separator( EXPR )
1334 $OUTPUT_RECORD_SEPARATOR
1335 $ORS
1336 $\ The output record separator for the print operator. If
1337 defined, this value is printed after the last of print's
1338 arguments. Default is "undef".
1339 You cannot call "output_record_separator()" on a handle, only
1341 as a static method. See IO::Handle.
1342 Mnemonic: you set "$\" instead of adding "\n" at the end of the
1344 print. Also, it's just like $/, but it's what you get "back"
1345 from Perl.
1346 HANDLE->autoflush( EXPR )
1348 $OUTPUT_AUTOFLUSH
1349 $| If set to nonzero, forces a flush right away and after every
1350 write or print on the currently selected output channel.
1351 Default is 0 (regardless of whether the channel is really
1352 buffered by the system or not; $| tells you only whether you've
1353 asked Perl explicitly to flush after each write). STDOUT will
1354 typically be line buffered if output is to the terminal and
1355 block buffered otherwise. Setting this variable is useful
1356 primarily when you are outputting to a pipe or socket, such as
1357 when you are running a Perl program under rsh and want to see
1358 the output as it's happening. This has no effect on input
1359 buffering. See "getc" in perlfunc for that. See "select" in
1360 perlfunc on how to select the output channel. See also
1361 IO::Handle.
1362 Mnemonic: when you want your pipes to be piping hot.
1364 ${^LAST_FH}
1366 This read-only variable contains a reference to the last-read
1367 filehandle. This is set by "<HANDLE>", "readline", "tell",
1368 "eof" and "seek". This is the same handle that $. and "tell"
1369 and "eof" without arguments use. It is also the handle used
1370 when Perl appends ", <STDIN> line 1" to an error or warning
1371 message.
1372 This variable was added in Perl v5.18.0.
1374 Variables related to formats
1376 The special variables for formats are a subset of those for
1378 filehandles. See perlform for more information about Perl's formats.
1379 $ACCUMULATOR
1381 $^A The current value of the "write()" accumulator for "format()"
1382 lines. A format contains "formline()" calls that put their
1383 result into $^A. After calling its format, "write()" prints
1384 out the contents of $^A and empties. So you never really see
1385 the contents of $^A unless you call "formline()" yourself and
1386 then look at it. See perlform and "formline PICTURE,LIST" in
1387 perlfunc.
1388 IO::Handle->format_formfeed(EXPR)
1390 $FORMAT_FORMFEED
1391 $^L What formats output as a form feed. The default is "\f".
1392 You cannot call "format_formfeed()" on a handle, only as a
1394 static method. See IO::Handle.
1395 HANDLE->format_page_number(EXPR)
1397 $FORMAT_PAGE_NUMBER
1398 $% The current page number of the currently selected output
1399 channel.
1400 Mnemonic: "%" is page number in nroff.
1402 HANDLE->format_lines_left(EXPR)
1404 $FORMAT_LINES_LEFT
1405 $- The number of lines left on the page of the currently selected
1406 output channel.
1407 Mnemonic: lines_on_page - lines_printed.
1409 IO::Handle->format_line_break_characters EXPR
1411 $FORMAT_LINE_BREAK_CHARACTERS
1412 $: The current set of characters after which a string may be
1413 broken to fill continuation fields (starting with "^") in a
1414 format. The default is " \n-", to break on a space, newline,
1415 or a hyphen.
1416 You cannot call "format_line_break_characters()" on a handle,
1418 only as a static method. See IO::Handle.
1419 Mnemonic: a "colon" in poetry is a part of a line.
1421 HANDLE->format_lines_per_page(EXPR)
1423 $FORMAT_LINES_PER_PAGE
1424 $= The current page length (printable lines) of the currently
1425 selected output channel. The default is 60.
1426 Mnemonic: = has horizontal lines.
1428 HANDLE->format_top_name(EXPR)
1430 $FORMAT_TOP_NAME
1431 $^ The name of the current top-of-page format for the currently
1432 selected output channel. The default is the name of the
1433 filehandle with "_TOP" appended. For example, the default
1434 format top name for the "STDOUT" filehandle is "STDOUT_TOP".
1435 Mnemonic: points to top of page.
1437 HANDLE->format_name(EXPR)
1439 $FORMAT_NAME
1440 $~ The name of the current report format for the currently
1441 selected output channel. The default format name is the same
1442 as the filehandle name. For example, the default format name
1443 for the "STDOUT" filehandle is just "STDOUT".
1444 Mnemonic: brother to $^.
1446 Error Variables
1448 The variables $@, $!, $^E, and $? contain information about different
1449 types of error conditions that may appear during execution of a Perl
1450 program. The variables are shown ordered by the "distance" between the
1451 subsystem which reported the error and the Perl process. They
1452 correspond to errors detected by the Perl interpreter, C library,
1453 operating system, or an external program, respectively.
1454 To illustrate the differences between these variables, consider the
1456 following Perl expression, which uses a single-quoted string. After
1457 execution of this statement, perl may have set all four special error
1458 variables:
1459 eval q{
1461 open my $pipe, "/cdrom/install |" or die $!;
1462 my @res = <$pipe>;
1463 close $pipe or die "bad pipe: $?, $!";
1464 };
1465 When perl executes the "eval()" expression, it translates the "open()",
1467 "<PIPE>", and "close" calls in the C run-time library and thence to the
1468 operating system kernel. perl sets $! to the C library's "errno" if
1469 one of these calls fails.
1470 $@ is set if the string to be "eval"-ed did not compile (this may
1472 happen if "open" or "close" were imported with bad prototypes), or if
1473 Perl code executed during evaluation "die()"d. In these cases the
1474 value of $@ is the compile error, or the argument to "die" (which will
1475 interpolate $! and $?). (See also Fatal, though.)
1476 Under a few operating systems, $^E may contain a more verbose error
1478 indicator, such as in this case, "CDROM tray not closed." Systems that
1479 do not support extended error messages leave $^E the same as $!.
1480 Finally, $? may be set to a non-0 value if the external program
1482 /cdrom/install fails. The upper eight bits reflect specific error
1483 conditions encountered by the program (the program's "exit()" value).
1484 The lower eight bits reflect mode of failure, like signal death and
1485 core dump information. See wait(2) for details. In contrast to $! and
1486 $^E, which are set only if an error condition is detected, the variable
1487 $? is set on each "wait" or pipe "close", overwriting the old value.
1488 This is more like $@, which on every "eval()" is always set on failure
1489 and cleared on success.
1490 For more details, see the individual descriptions at $@, $!, $^E, and
1492 $?.
1493 ${^CHILD_ERROR_NATIVE}
1495 The native status returned by the last pipe close, backtick
1496 ("``") command, successful call to "wait()" or "waitpid()", or
1497 from the "system()" operator. On POSIX-like systems this value
1498 can be decoded with the WIFEXITED, WEXITSTATUS, WIFSIGNALED,
1499 WTERMSIG, WIFSTOPPED, WSTOPSIG and WIFCONTINUED functions
1500 provided by the POSIX module.
1501 Under VMS this reflects the actual VMS exit status; i.e. it is
1503 the same as $? when the pragma "use vmsish 'status'" is in
1504 effect.
1505 This variable was added in Perl v5.10.0.
1507 $EXTENDED_OS_ERROR
1509 $^E Error information specific to the current operating system. At
1510 the moment, this differs from "$!" under only VMS, OS/2, and
1511 Win32 (and for MacPerl). On all other platforms, $^E is always
1512 just the same as $!.
1513 Under VMS, $^E provides the VMS status value from the last
1515 system error. This is more specific information about the last
1516 system error than that provided by $!. This is particularly
1517 important when $! is set to EVMSERR.
1518 Under OS/2, $^E is set to the error code of the last call to
1520 OS/2 API either via CRT, or directly from perl.
1521 Under Win32, $^E always returns the last error information
1523 reported by the Win32 call "GetLastError()" which describes the
1524 last error from within the Win32 API. Most Win32-specific code
1525 will report errors via $^E. ANSI C and Unix-like calls set
1526 "errno" and so most portable Perl code will report errors via
1527 $!.
1528 Caveats mentioned in the description of "$!" generally apply to
1530 $^E, also.
1531 This variable was added in Perl 5.003.
1533 Mnemonic: Extra error explanation.
1535 $EXCEPTIONS_BEING_CAUGHT
1537 $^S Current state of the interpreter.
1538 $^S State
1540 --------- -------------------------------------
1541 undef Parsing module, eval, or main program
1542 true (1) Executing an eval
1543 false (0) Otherwise
1544 The first state may happen in $SIG{__DIE__} and $SIG{__WARN__}
1546 handlers.
1547 The English name $EXCEPTIONS_BEING_CAUGHT is slightly
1549 misleading, because the "undef" value does not indicate whether
1550 exceptions are being caught, since compilation of the main
1551 program does not catch exceptions.
1552 This variable was added in Perl 5.004.
1554 $WARNING
1556 $^W The current value of the warning switch, initially true if -w
1557 was used, false otherwise, but directly modifiable.
1558 See also warnings.
1560 Mnemonic: related to the -w switch.
1562 ${^WARNING_BITS}
1564 The current set of warning checks enabled by the "use warnings"
1565 pragma. It has the same scoping as the $^H and "%^H"
1566 variables. The exact values are considered internal to the
1567 warnings pragma and may change between versions of Perl.
1568 This variable was added in Perl v5.6.0.
1570 $OS_ERROR
1572 $ERRNO
1573 $! When referenced, $! retrieves the current value of the C
1574 "errno" integer variable. If $! is assigned a numerical value,
1575 that value is stored in "errno". When referenced as a string,
1576 $! yields the system error string corresponding to "errno".
1577 Many system or library calls set "errno" if they fail, to
1579 indicate the cause of failure. They usually do not set "errno"
1580 to zero if they succeed. This means "errno", hence $!, is
1581 meaningful only immediately after a failure:
1582 if (open my $fh, "<", $filename) {
1584 # Here $! is meaningless.
1585 ...
1586 }
1587 else {
1588 # ONLY here is $! meaningful.
1589 ...
1590 # Already here $! might be meaningless.
1591 }
1592 # Since here we might have either success or failure,
1593 # $! is meaningless.
1594 Here, meaningless means that $! may be unrelated to the outcome
1596 of the "open()" operator. Assignment to $! is similarly
1597 ephemeral. It can be used immediately before invoking the
1598 "die()" operator, to set the exit value, or to inspect the
1599 system error string corresponding to error n, or to restore $!
1600 to a meaningful state.
1601 Mnemonic: What just went bang?
1603 %OS_ERROR
1605 %ERRNO
1606 %! Each element of "%!" has a true value only if $! is set to that
1607 value. For example, $!{ENOENT} is true if and only if the
1608 current value of $! is "ENOENT"; that is, if the most recent
1609 error was "No such file or directory" (or its moral equivalent:
1610 not all operating systems give that exact error, and certainly
1611 not all languages). The specific true value is not guaranteed,
1612 but in the past has generally been the numeric value of $!. To
1613 check if a particular key is meaningful on your system, use
1614 "exists $!{the_key}"; for a list of legal keys, use "keys %!".
1615 See Errno for more information, and also see "$!".
1616 This variable was added in Perl 5.005.
1618 $CHILD_ERROR
1620 $? The status returned by the last pipe close, backtick ("``")
1621 command, successful call to "wait()" or "waitpid()", or from
1622 the "system()" operator. This is just the 16-bit status word
1623 returned by the traditional Unix "wait()" system call (or else
1624 is made up to look like it). Thus, the exit value of the
1625 subprocess is really ("$? >> 8"), and "$? & 127" gives which
1626 signal, if any, the process died from, and "$? & 128" reports
1627 whether there was a core dump.
1628 Additionally, if the "h_errno" variable is supported in C, its
1630 value is returned via $? if any "gethost*()" function fails.
1631 If you have installed a signal handler for "SIGCHLD", the value
1633 of $? will usually be wrong outside that handler.
1634 Inside an "END" subroutine $? contains the value that is going
1636 to be given to "exit()". You can modify $? in an "END"
1637 subroutine to change the exit status of your program. For
1638 example:
1639 END {
1641 $? = 1 if $? == 255; # die would make it 255
1642 }
1643 Under VMS, the pragma "use vmsish 'status'" makes $? reflect
1645 the actual VMS exit status, instead of the default emulation of
1646 POSIX status; see "$?" in perlvms for details.
1647 Mnemonic: similar to sh and ksh.
1649 $EVAL_ERROR
1651 $@ The Perl error from the last "eval" operator, i.e. the last
1652 exception that was caught. For "eval BLOCK", this is either a
1653 runtime error message or the string or reference "die" was
1654 called with. The "eval STRING" form also catches syntax errors
1655 and other compile time exceptions.
1656 If no error occurs, "eval" sets $@ to the empty string.
1658 Warning messages are not collected in this variable. You can,
1660 however, set up a routine to process warnings by setting
1661 $SIG{__WARN__} as described in "%SIG".
1662 Mnemonic: Where was the error "at"?
1664 Variables related to the interpreter state
1666 These variables provide information about the current interpreter
1667 state.
1668 $COMPILING
1670 $^C The current value of the flag associated with the -c switch.
1671 Mainly of use with -MO=... to allow code to alter its behavior
1672 when being compiled, such as for example to "AUTOLOAD" at
1673 compile time rather than normal, deferred loading. Setting
1674 "$^C = 1" is similar to calling "B::minus_c".
1675 This variable was added in Perl v5.6.0.
1677 $DEBUGGING
1679 $^D The current value of the debugging flags. May be read or set.
1680 Like its command-line equivalent, you can use numeric or
1681 symbolic values, e.g. "$^D = 10" or "$^D = "st"". See
1682 "-Dnumber" in perlrun. The contents of this variable also
1683 affects the debugger operation. See "Debugger Internals" in
1684 perldebguts.
1685 Mnemonic: value of -D switch.
1687 ${^ENCODING}
1689 This variable is no longer supported.
1690 It used to hold the object reference to the "Encode" object
1692 that was used to convert the source code to Unicode.
1693 Its purpose was to allow your non-ASCII Perl scripts not to
1695 have to be written in UTF-8; this was useful before editors
1696 that worked on UTF-8 encoded text were common, but that was
1697 long ago. It caused problems, such as affecting the operation
1698 of other modules that weren't expecting it, causing general
1699 mayhem.
1700 If you need something like this functionality, it is
1702 recommended that use you a simple source filter, such as
1703 Filter::Encoding.
1704 If you are coming here because code of yours is being adversely
1706 affected by someone's use of this variable, you can usually
1707 work around it by doing this:
1708 local ${^ENCODING};
1710 near the beginning of the functions that are getting broken.
1712 This undefines the variable during the scope of execution of
1713 the including function.
1714 This variable was added in Perl 5.8.2 and removed in 5.26.0.
1716 Setting it to anything other than "undef" was made fatal in
1717 Perl 5.28.0.
1718 ${^GLOBAL_PHASE}
1720 The current phase of the perl interpreter.
1721 Possible values are:
1723 CONSTRUCT
1725 The "PerlInterpreter*" is being constructed via
1726 "perl_construct". This value is mostly there for
1727 completeness and for use via the underlying C variable
1728 "PL_phase". It's not really possible for Perl code to
1729 be executed unless construction of the interpreter is
1730 finished.
1731 START This is the global compile-time. That includes,
1733 basically, every "BEGIN" block executed directly or
1734 indirectly from during the compile-time of the top-
1735 level program.
1736 This phase is not called "BEGIN" to avoid confusion
1738 with "BEGIN"-blocks, as those are executed during
1739 compile-time of any compilation unit, not just the top-
1740 level program. A new, localised compile-time entered
1741 at run-time, for example by constructs as "eval "use
1742 SomeModule"" are not global interpreter phases, and
1743 therefore aren't reflected by "${^GLOBAL_PHASE}".
1744 CHECK Execution of any "CHECK" blocks.
1746 INIT Similar to "CHECK", but for "INIT"-blocks, not "CHECK"
1748 blocks.
1749 RUN The main run-time, i.e. the execution of
1751 "PL_main_root".
1752 END Execution of any "END" blocks.
1754 DESTRUCT
1756 Global destruction.
1757 Also note that there's no value for UNITCHECK-blocks. That's
1759 because those are run for each compilation unit individually,
1760 and therefore is not a global interpreter phase.
1761 Not every program has to go through each of the possible
1763 phases, but transition from one phase to another can only
1764 happen in the order described in the above list.
1765 An example of all of the phases Perl code can see:
1767 BEGIN { print "compile-time: ${^GLOBAL_PHASE}\n" }
1769 INIT { print "init-time: ${^GLOBAL_PHASE}\n" }
1771 CHECK { print "check-time: ${^GLOBAL_PHASE}\n" }
1773 {
1775 package Print::Phase;
1776 sub new {
1778 my ($class, $time) = @_;
1779 return bless \$time, $class;
1780 }
1781 sub DESTROY {
1783 my $self = shift;
1784 print "$$self: ${^GLOBAL_PHASE}\n";
1785 }
1786 }
1787 print "run-time: ${^GLOBAL_PHASE}\n";
1789 my $runtime = Print::Phase->new(
1791 "lexical variables are garbage collected before END"
1792 );
1793 END { print "end-time: ${^GLOBAL_PHASE}\n" }
1795 our $destruct = Print::Phase->new(
1797 "package variables are garbage collected after END"
1798 );
1799 This will print out
1801 compile-time: START
1803 check-time: CHECK
1804 init-time: INIT
1805 run-time: RUN
1806 lexical variables are garbage collected before END: RUN
1807 end-time: END
1808 package variables are garbage collected after END: DESTRUCT
1809 This variable was added in Perl 5.14.0.
1811 $^H WARNING: This variable is strictly for internal use only. Its
1813 availability, behavior, and contents are subject to change
1814 without notice.
1815 This variable contains compile-time hints for the Perl
1817 interpreter. At the end of compilation of a BLOCK the value of
1818 this variable is restored to the value when the interpreter
1819 started to compile the BLOCK.
1820 When perl begins to parse any block construct that provides a
1822 lexical scope (e.g., eval body, required file, subroutine body,
1823 loop body, or conditional block), the existing value of $^H is
1824 saved, but its value is left unchanged. When the compilation
1825 of the block is completed, it regains the saved value. Between
1826 the points where its value is saved and restored, code that
1827 executes within BEGIN blocks is free to change the value of
1828 $^H.
1829 This behavior provides the semantic of lexical scoping, and is
1831 used in, for instance, the "use strict" pragma.
1832 The contents should be an integer; different bits of it are
1834 used for different pragmatic flags. Here's an example:
1835 sub add_100 { $^H |= 0x100 }
1837 sub foo {
1839 BEGIN { add_100() }
1840 bar->baz($boon);
1841 }
1842 Consider what happens during execution of the BEGIN block. At
1844 this point the BEGIN block has already been compiled, but the
1845 body of "foo()" is still being compiled. The new value of $^H
1846 will therefore be visible only while the body of "foo()" is
1847 being compiled.
1848 Substitution of "BEGIN { add_100() }" block with:
1850 BEGIN { require strict; strict->import('vars') }
1852 demonstrates how "use strict 'vars'" is implemented. Here's a
1854 conditional version of the same lexical pragma:
1855 BEGIN {
1857 require strict; strict->import('vars') if $condition
1858 }
1859 This variable was added in Perl 5.003.
1861 %^H The "%^H" hash provides the same scoping semantic as $^H. This
1863 makes it useful for implementation of lexically scoped pragmas.
1864 See perlpragma. All the entries are stringified when accessed
1865 at runtime, so only simple values can be accommodated. This
1866 means no pointers to objects, for example.
1867 When putting items into "%^H", in order to avoid conflicting
1869 with other users of the hash there is a convention regarding
1870 which keys to use. A module should use only keys that begin
1871 with the module's name (the name of its main package) and a "/"
1872 character. For example, a module "Foo::Bar" should use keys
1873 such as "Foo::Bar/baz".
1874 This variable was added in Perl v5.6.0.
1876 ${^OPEN}
1878 An internal variable used by PerlIO. A string in two parts,
1879 separated by a "\0" byte, the first part describes the input
1880 layers, the second part describes the output layers.
1881 This variable was added in Perl v5.8.0.
1883 $PERLDB
1885 $^P The internal variable for debugging support. The meanings of
1886 the various bits are subject to change, but currently indicate:
1887 0x01 Debug subroutine enter/exit.
1889 0x02 Line-by-line debugging. Causes "DB::DB()" subroutine to
1891 be called for each statement executed. Also causes
1892 saving source code lines (like 0x400).
1893 0x04 Switch off optimizations.
1895 0x08 Preserve more data for future interactive inspections.
1897 0x10 Keep info about source lines on which a subroutine is
1899 defined.
1900 0x20 Start with single-step on.
1902 0x40 Use subroutine address instead of name when reporting.
1904 0x80 Report "goto &subroutine" as well.
1906 0x100 Provide informative "file" names for evals based on the
1908 place they were compiled.
1909 0x200 Provide informative names to anonymous subroutines based
1911 on the place they were compiled.
1912 0x400 Save source code lines into "@{"_<$filename"}".
1914 0x800 When saving source, include evals that generate no
1916 subroutines.
1917 0x1000
1919 When saving source, include source that did not compile.
1920 Some bits may be relevant at compile-time only, some at run-
1922 time only. This is a new mechanism and the details may change.
1923 See also perldebguts.
1924 ${^TAINT}
1926 Reflects if taint mode is on or off. 1 for on (the program was
1927 run with -T), 0 for off, -1 when only taint warnings are
1928 enabled (i.e. with -t or -TU).
1929 This variable is read-only.
1931 This variable was added in Perl v5.8.0.
1933 ${^SAFE_LOCALES}
1935 Reflects if safe locale operations are available to this perl
1936 (when the value is 1) or not (the value is 0). This variable
1937 is always 1 if the perl has been compiled without threads. It
1938 is also 1 if this perl is using thread-safe locale operations.
1939 Note that an individual thread may choose to use the global
1940 locale (generally unsafe) by calling "switch_to_global_locale"
1941 in perlapi. This variable currently is still set to 1 in such
1942 threads.
1943 This variable is read-only.
1945 This variable was added in Perl v5.28.0.
1947 ${^UNICODE}
1949 Reflects certain Unicode settings of Perl. See perlrun
1950 documentation for the "-C" switch for more information about
1951 the possible values.
1952 This variable is set during Perl startup and is thereafter
1954 read-only.
1955 This variable was added in Perl v5.8.2.
1957 ${^UTF8CACHE}
1959 This variable controls the state of the internal UTF-8 offset
1960 caching code. 1 for on (the default), 0 for off, -1 to debug
1961 the caching code by checking all its results against linear
1962 scans, and panicking on any discrepancy.
1963 This variable was added in Perl v5.8.9. It is subject to
1965 change or removal without notice, but is currently used to
1966 avoid recalculating the boundaries of multi-byte UTF-8-encoded
1967 characters.
1968 ${^UTF8LOCALE}
1970 This variable indicates whether a UTF-8 locale was detected by
1971 perl at startup. This information is used by perl when it's in
1972 adjust-utf8ness-to-locale mode (as when run with the "-CL"
1973 command-line switch); see perlrun for more info on this.
1974 This variable was added in Perl v5.8.8.
1976 Deprecated and removed variables
1978 Deprecating a variable announces the intent of the perl maintainers to
1979 eventually remove the variable from the language. It may still be
1980 available despite its status. Using a deprecated variable triggers a
1981 warning.
1982 Once a variable is removed, its use triggers an error telling you the
1984 variable is unsupported.
1985 See perldiag for details about error messages.
1987 $# $# was a variable that could be used to format printed numbers.
1989 After a deprecation cycle, its magic was removed in Perl
1990 v5.10.0 and using it now triggers a warning: "$# is no longer
1991 supported".
1992 This is not the sigil you use in front of an array name to get
1994 the last index, like $#array. That's still how you get the
1995 last index of an array in Perl. The two have nothing to do
1996 with each other.
1997 Deprecated in Perl 5.
1999 Removed in Perl v5.10.0.
2001 $* $* was a variable that you could use to enable multiline
2003 matching. After a deprecation cycle, its magic was removed in
2004 Perl v5.10.0. Using it now triggers a warning: "$* is no
2005 longer supported". You should use the "/s" and "/m" regexp
2006 modifiers instead.
2007 Deprecated in Perl 5.
2009 Removed in Perl v5.10.0.
2011 $[ This variable stores the index of the first element in an
2013 array, and of the first character in a substring. The default
2014 is 0, but you could theoretically set it to 1 to make Perl
2015 behave more like awk (or Fortran) when subscripting and when
2016 evaluating the index() and substr() functions.
2017 As of release 5 of Perl, assignment to $[ is treated as a
2019 compiler directive, and cannot influence the behavior of any
2020 other file. (That's why you can only assign compile-time
2021 constants to it.) Its use is highly discouraged.
2022 Prior to Perl v5.10.0, assignment to $[ could be seen from
2024 outer lexical scopes in the same file, unlike other compile-
2025 time directives (such as strict). Using local() on it would
2026 bind its value strictly to a lexical block. Now it is always
2027 lexically scoped.
2028 As of Perl v5.16.0, it is implemented by the arybase module.
2030 See arybase for more details on its behaviour.
2031 Under "use v5.16", or "no feature "array_base"", $[ no longer
2033 has any effect, and always contains 0. Assigning 0 to it is
2034 permitted, but any other value will produce an error.
2035 Mnemonic: [ begins subscripts.
2037 Deprecated in Perl v5.12.0.
2039 Assigning a non-zero value be fatal in Perl v5.30.0.
2041perl v5.28.2 2019-03-28 PERLVAR(1)