2 X<subroutine> X<function>
4 perlsub - Perl subroutines
8 To declare subroutines:
9 X<subroutine, declaration> X<sub>
11 sub NAME; # A "forward" declaration.
12 sub NAME(PROTO); # ditto, but with prototypes
13 sub NAME : ATTRS; # with attributes
14 sub NAME(PROTO) : ATTRS; # with attributes and prototypes
16 sub NAME BLOCK # A declaration and a definition.
17 sub NAME(PROTO) BLOCK # ditto, but with prototypes
18 sub NAME : ATTRS BLOCK # with attributes
19 sub NAME(PROTO) : ATTRS BLOCK # with prototypes and attributes
21 use feature 'signatures';
22 sub NAME(SIG) BLOCK # with signature
23 sub NAME :ATTRS (SIG) BLOCK # with signature, attributes
24 sub NAME :prototype(PROTO) (SIG) BLOCK # with signature, prototype
26 To define an anonymous subroutine at runtime:
27 X<subroutine, anonymous>
29 $subref = sub BLOCK; # no proto
30 $subref = sub (PROTO) BLOCK; # with proto
31 $subref = sub : ATTRS BLOCK; # with attributes
32 $subref = sub (PROTO) : ATTRS BLOCK; # with proto and attributes
34 use feature 'signatures';
35 $subref = sub (SIG) BLOCK; # with signature
36 $subref = sub : ATTRS(SIG) BLOCK; # with signature, attributes
38 To import subroutines:
41 use MODULE qw(NAME1 NAME2 NAME3);
44 X<subroutine, call> X<call>
46 NAME(LIST); # & is optional with parentheses.
47 NAME LIST; # Parentheses optional if predeclared/imported.
48 &NAME(LIST); # Circumvent prototypes.
49 &NAME; # Makes current @_ visible to called subroutine.
53 Like many languages, Perl provides for user-defined subroutines.
54 These may be located anywhere in the main program, loaded in from
55 other files via the C<do>, C<require>, or C<use> keywords, or
56 generated on the fly using C<eval> or anonymous subroutines.
57 You can even call a function indirectly using a variable containing
58 its name or a CODE reference.
60 The Perl model for function call and return values is simple: all
61 functions are passed as parameters one single flat list of scalars, and
62 all functions likewise return to their caller one single flat list of
63 scalars. Any arrays or hashes in these call and return lists will
64 collapse, losing their identities--but you may always use
65 pass-by-reference instead to avoid this. Both call and return lists may
66 contain as many or as few scalar elements as you'd like. (Often a
67 function without an explicit return statement is called a subroutine, but
68 there's really no difference from Perl's perspective.)
69 X<subroutine, parameter> X<parameter>
71 In a subroutine that uses signatures (see L</Signatures> below),
72 arguments are assigned into lexical variables introduced by the
73 signature. In the current implementation of perl they are also
74 accessible in the C<@_> array in the same way as for non-signature
75 subroutines, but accessing them in this manner is now discouraged inside
76 such a signature-using subroutine.
78 In a subroutine that does not use signatures, any arguments passed in
79 show up in the array C<@_>. Therefore, if you called a function with
80 two arguments, those would be stored in C<$_[0]> and C<$_[1]>. The
81 array C<@_> is a local array, but its elements are aliases for the
82 actual scalar parameters. In particular, if an element C<$_[0]> is
83 updated, the corresponding argument is updated (or an error occurs if it
84 is not updatable). If an argument is an array or hash element which did
85 not exist when the function was called, that element is created only
86 when (and if) it is modified or a reference to it is taken. (Some
87 earlier versions of Perl created the element whether or not the element
88 was assigned to.) Assigning to the whole array C<@_> removes that
89 aliasing, and does not update any arguments.
90 X<subroutine, argument> X<argument> X<@_>
92 When not using signatures, Perl does not otherwise provide a means to
93 create named formal parameters. In practice all you do is assign to a
94 C<my()> list of these. Variables that aren't declared to be private are
95 global variables. For gory details on creating private variables, see
96 L</"Private Variables via my()"> and L</"Temporary Values via local()">.
97 To create protected environments for a set of functions in a separate
98 package (and probably a separate file), see L<perlmod/"Packages">.
100 A C<return> statement may be used to exit a subroutine, optionally
101 specifying the returned value, which will be evaluated in the
102 appropriate context (list, scalar, or void) depending on the context of
103 the subroutine call. If you specify no return value, the subroutine
104 returns an empty list in list context, the undefined value in scalar
105 context, or nothing in void context. If you return one or more
106 aggregates (arrays and hashes), these will be flattened together into
107 one large indistinguishable list.
109 If no C<return> is found and if the last statement is an expression, its
110 value is returned. If the last statement is a loop control structure
111 like a C<foreach> or a C<while>, the returned value is unspecified. The
112 empty sub returns the empty list.
113 X<subroutine, return value> X<return value> X<return>
120 $max = $foo if $max < $foo;
124 $bestday = max($mon,$tue,$wed,$thu,$fri);
128 # get a line, combining continuation lines
129 # that start with whitespace
132 $thisline = $lookahead; # global variables!
133 LINE: while (defined($lookahead = <STDIN>)) {
134 if ($lookahead =~ /^[ \t]/) {
135 $thisline .= $lookahead;
144 $lookahead = <STDIN>; # get first line
145 while (defined($line = get_line())) {
149 Assigning to a list of private variables to name your arguments:
152 my($key, $value) = @_;
153 $Foo{$key} = $value unless $Foo{$key};
156 Because the assignment copies the values, this also has the effect
157 of turning call-by-reference into call-by-value. Otherwise a
158 function is free to do in-place modifications of C<@_> and change
160 X<call-by-reference> X<call-by-value>
162 upcase_in($v1, $v2); # this changes $v1 and $v2
164 for (@_) { tr/a-z/A-Z/ }
167 You aren't allowed to modify constants in this way, of course. If an
168 argument were actually literal and you tried to change it, you'd take a
169 (presumably fatal) exception. For example, this won't work:
170 X<call-by-reference> X<call-by-value>
172 upcase_in("frederick");
174 It would be much safer if the C<upcase_in()> function
175 were written to return a copy of its parameters instead
176 of changing them in place:
178 ($v3, $v4) = upcase($v1, $v2); # this doesn't change $v1 and $v2
180 return unless defined wantarray; # void context, do nothing
182 for (@parms) { tr/a-z/A-Z/ }
183 return wantarray ? @parms : $parms[0];
186 Notice how this (unprototyped) function doesn't care whether it was
187 passed real scalars or arrays. Perl sees all arguments as one big,
188 long, flat parameter list in C<@_>. This is one area where
189 Perl's simple argument-passing style shines. The C<upcase()>
190 function would work perfectly well without changing the C<upcase()>
191 definition even if we fed it things like this:
193 @newlist = upcase(@list1, @list2);
194 @newlist = upcase( split /:/, $var );
196 Do not, however, be tempted to do this:
198 (@a, @b) = upcase(@list1, @list2);
200 Like the flattened incoming parameter list, the return list is also
201 flattened on return. So all you have managed to do here is stored
202 everything in C<@a> and made C<@b> empty. See
203 L</Pass by Reference> for alternatives.
205 A subroutine may be called using an explicit C<&> prefix. The
206 C<&> is optional in modern Perl, as are parentheses if the
207 subroutine has been predeclared. The C<&> is I<not> optional
208 when just naming the subroutine, such as when it's used as
209 an argument to defined() or undef(). Nor is it optional when you
210 want to do an indirect subroutine call with a subroutine name or
211 reference using the C<&$subref()> or C<&{$subref}()> constructs,
212 although the C<< $subref->() >> notation solves that problem.
213 See L<perlref> for more about all that.
216 Subroutines may be called recursively. If a subroutine is called
217 using the C<&> form, the argument list is optional, and if omitted,
218 no C<@_> array is set up for the subroutine: the C<@_> array at the
219 time of the call is visible to subroutine instead. This is an
220 efficiency mechanism that new users may wish to avoid.
223 &foo(1,2,3); # pass three arguments
224 foo(1,2,3); # the same
226 foo(); # pass a null list
229 &foo; # foo() get current args, like foo(@_) !!
231 foo; # like foo() iff sub foo predeclared, else
232 # a compile-time error
234 foo; # like foo() iff sub foo predeclared, else
235 # a literal string "foo"
237 Not only does the C<&> form make the argument list optional, it also
238 disables any prototype checking on arguments you do provide. This
239 is partly for historical reasons, and partly for having a convenient way
240 to cheat if you know what you're doing. See L</Prototypes> below.
243 Since Perl 5.16.0, the C<__SUB__> token is available under C<use feature
244 'current_sub'> and C<use 5.16.0>. It will evaluate to a reference to the
245 currently-running sub, which allows for recursive calls without knowing
246 your subroutine's name.
249 my $factorial = sub {
252 return($x * __SUB__->( $x - 1 ) );
255 The behavior of C<__SUB__> within a regex code block (such as C</(?{...})/>)
256 is subject to change.
258 Subroutines whose names are in all upper case are reserved to the Perl
259 core, as are modules whose names are in all lower case. A subroutine in
260 all capitals is a loosely-held convention meaning it will be called
261 indirectly by the run-time system itself, usually due to a triggered event.
262 Subroutines whose name start with a left parenthesis are also reserved the
263 same way. The following is a list of some subroutines that currently do
264 special, pre-defined things.
268 =item documented later in this document
272 =item documented in L<perlmod>
274 C<CLONE>, C<CLONE_SKIP>
276 =item documented in L<perlobj>
280 =item documented in L<perltie>
282 C<BINMODE>, C<CLEAR>, C<CLOSE>, C<DELETE>, C<DESTROY>, C<EOF>, C<EXISTS>,
283 C<EXTEND>, C<FETCH>, C<FETCHSIZE>, C<FILENO>, C<FIRSTKEY>, C<GETC>,
284 C<NEXTKEY>, C<OPEN>, C<POP>, C<PRINT>, C<PRINTF>, C<PUSH>, C<READ>,
285 C<READLINE>, C<SCALAR>, C<SEEK>, C<SHIFT>, C<SPLICE>, C<STORE>,
286 C<STORESIZE>, C<TELL>, C<TIEARRAY>, C<TIEHANDLE>, C<TIEHASH>,
287 C<TIESCALAR>, C<UNSHIFT>, C<UNTIE>, C<WRITE>
289 =item documented in L<PerlIO::via>
291 C<BINMODE>, C<CLEARERR>, C<CLOSE>, C<EOF>, C<ERROR>, C<FDOPEN>, C<FILENO>,
292 C<FILL>, C<FLUSH>, C<OPEN>, C<POPPED>, C<PUSHED>, C<READ>, C<SEEK>,
293 C<SETLINEBUF>, C<SYSOPEN>, C<TELL>, C<UNREAD>, C<UTF8>, C<WRITE>
295 =item documented in L<perlfunc>
297 L<< C<import> | perlfunc/use >>, L<< C<unimport> | perlfunc/use >>,
298 L<< C<INC> | perlfunc/require >>
300 =item documented in L<UNIVERSAL>
304 =item documented in L<perldebguts>
306 C<DB::DB>, C<DB::sub>, C<DB::lsub>, C<DB::goto>, C<DB::postponed>
308 =item undocumented, used internally by the L<overload> feature
310 any starting with C<(>
314 The C<BEGIN>, C<UNITCHECK>, C<CHECK>, C<INIT> and C<END> subroutines
315 are not so much subroutines as named special code blocks, of which you
316 can have more than one in a package, and which you can B<not> call
317 explicitly. See L<perlmod/"BEGIN, UNITCHECK, CHECK, INIT and END">
321 X<formal parameter> X<parameter, formal>
323 Perl has a facility to allow a subroutine's formal parameters to be
324 declared by special syntax, separate from the procedural code of the
325 subroutine body. The formal parameter list is known as a I<signature>.
327 This facility must be enabled before it can be used. It is enabled
328 automatically by a C<use v5.36> (or higher) declaration, or more
329 directly by C<use feature 'signatures'>, in the current scope.
331 The signature is part of a subroutine's body. Normally the body of a
332 subroutine is simply a braced block of code, but when using a signature,
333 the signature is a parenthesised list that goes immediately before the
334 block, after any name or attributes.
338 sub foo :lvalue ($a, $b = 1, @c) { .... }
340 The signature declares lexical variables that are
341 in scope for the block. When the subroutine is called, the signature
342 takes control first. It populates the signature variables from the
343 list of arguments that were passed. If the argument list doesn't meet
344 the requirements of the signature, then it will throw an exception.
345 When the signature processing is complete, control passes to the block.
347 Positional parameters are handled by simply naming scalar variables in
348 the signature. For example,
350 sub foo ($left, $right) {
351 return $left + $right;
354 takes two positional parameters, which must be filled at runtime by
355 two arguments. By default the parameters are mandatory, and it is
356 not permitted to pass more arguments than expected. So the above is
360 die "Too many arguments for subroutine" unless @_ <= 2;
361 die "Too few arguments for subroutine" unless @_ >= 2;
364 return $left + $right;
367 An argument can be ignored by omitting the main part of the name from
368 a parameter declaration, leaving just a bare C<$> sigil. For example,
370 sub foo ($first, $, $third) {
371 return "first=$first, third=$third";
374 Although the ignored argument doesn't go into a variable, it is still
375 mandatory for the caller to pass it.
377 A positional parameter is made optional by giving a default value,
378 separated from the parameter name by C<=>:
380 sub foo ($left, $right = 0) {
381 return $left + $right;
384 The above subroutine may be called with either one or two arguments.
385 The default value expression is evaluated when the subroutine is called,
386 so it may provide different default values for different calls. It is
387 only evaluated if the argument was actually omitted from the call.
391 sub foo ($thing, $id = $auto_id++) {
392 print "$thing has ID $id";
395 automatically assigns distinct sequential IDs to things for which no
396 ID was supplied by the caller. A default value expression may also
397 refer to parameters earlier in the signature, making the default for
398 one parameter vary according to the earlier parameters. For example,
400 sub foo ($first_name, $surname, $nickname = $first_name) {
401 print "$first_name $surname is known as \"$nickname\"";
404 An optional parameter can be nameless just like a mandatory parameter.
407 sub foo ($thing, $ = 1) {
411 The parameter's default value will still be evaluated if the corresponding
412 argument isn't supplied, even though the value won't be stored anywhere.
413 This is in case evaluating it has important side effects. However, it
414 will be evaluated in void context, so if it doesn't have side effects
415 and is not trivial it will generate a warning if the "void" warning
416 category is enabled. If a nameless optional parameter's default value
417 is not important, it may be omitted just as the parameter's name was:
419 sub foo ($thing, $=) {
423 Optional positional parameters must come after all mandatory positional
424 parameters. (If there are no mandatory positional parameters then an
425 optional positional parameters can be the first thing in the signature.)
426 If there are multiple optional positional parameters and not enough
427 arguments are supplied to fill them all, they will be filled from left
430 After positional parameters, additional arguments may be captured in a
431 slurpy parameter. The simplest form of this is just an array variable:
433 sub foo ($filter, @inputs) {
434 print $filter->($_) foreach @inputs;
437 With a slurpy parameter in the signature, there is no upper limit on how
438 many arguments may be passed. A slurpy array parameter may be nameless
439 just like a positional parameter, in which case its only effect is to
440 turn off the argument limit that would otherwise apply:
442 sub foo ($thing, @) {
446 A slurpy parameter may instead be a hash, in which case the arguments
447 available to it are interpreted as alternating keys and values.
448 There must be as many keys as values: if there is an odd argument then
449 an exception will be thrown. Keys will be stringified, and if there are
450 duplicates then the later instance takes precedence over the earlier,
451 as with standard hash construction.
453 sub foo ($filter, %inputs) {
454 print $filter->($_, $inputs{$_}) foreach sort keys %inputs;
457 A slurpy hash parameter may be nameless just like other kinds of
458 parameter. It still insists that the number of arguments available to
459 it be even, even though they're not being put into a variable.
461 sub foo ($thing, %) {
465 A slurpy parameter, either array or hash, must be the last thing in the
466 signature. It may follow mandatory and optional positional parameters;
467 it may also be the only thing in the signature. Slurpy parameters cannot
468 have default values: if no arguments are supplied for them then you get
469 an empty array or empty hash.
471 A signature may be entirely empty, in which case all it does is check
472 that the caller passed no arguments:
478 Prior to Perl 5.36 these were considered experimental, and emitted a
479 warning in the C<experimental::signatures> category. From Perl 5.36
480 onwards this no longer happens, though the warning category still exists
481 for back-compatibility with code that attempts to disable it with a
484 no warnings 'experimental::signatures';
486 In the current perl implementation, when using a signature the arguments
487 are still also available in the special array variable C<@_>. However,
488 accessing them via this array is now discouraged, and should not be
489 relied upon in newly-written code as this ability may change in a future
490 version. Code that attempts to access the C<@_> array will produce
491 warnings in the C<experimental::args_array_with_signatures> category when
495 # This line emits the warning seen below
496 print "Arguments are @_";
501 Use of @_ in join or string with signatured subroutine is
504 There is a difference between the two ways of accessing the arguments:
505 C<@_> I<aliases> the arguments, but the signature variables get
506 I<copies> of the arguments. So writing to a signature variable only
507 changes that variable, and has no effect on the caller's variables, but
508 writing to an element of C<@_> modifies whatever the caller used to
509 supply that argument.
511 There is a potential syntactic ambiguity between signatures and prototypes
512 (see L</Prototypes>), because both start with an opening parenthesis and
513 both can appear in some of the same places, such as just after the name
514 in a subroutine declaration. For historical reasons, when signatures
515 are not enabled, any opening parenthesis in such a context will trigger
516 very forgiving prototype parsing. Most signatures will be interpreted
517 as prototypes in those circumstances, but won't be valid prototypes.
518 (A valid prototype cannot contain any alphabetic character.) This will
519 lead to somewhat confusing error messages.
521 To avoid ambiguity, when signatures are enabled the special syntax
522 for prototypes is disabled. There is no attempt to guess whether a
523 parenthesised group was intended to be a prototype or a signature.
524 To give a subroutine a prototype under these circumstances, use a
525 L<prototype attribute|attributes/Built-in Attributes>. For example,
527 sub foo :prototype($) { $_[0] }
529 It is entirely possible for a subroutine to have both a prototype and
530 a signature. They do different jobs: the prototype affects compilation
531 of calls to the subroutine, and the signature puts argument values into
532 lexical variables at runtime. You can therefore write
534 sub foo :prototype($$) ($left, $right) {
535 return $left + $right;
538 The prototype attribute, and any other attributes, must come before
539 the signature. The signature always immediately precedes the block of
540 the subroutine's body.
542 =head2 Private Variables via my()
543 X<my> X<variable, lexical> X<lexical> X<lexical variable> X<scope, lexical>
544 X<lexical scope> X<attributes, my>
548 my $foo; # declare $foo lexically local
549 my (@wid, %get); # declare list of variables local
550 my $foo = "flurp"; # declare $foo lexical, and init it
551 my @oof = @bar; # declare @oof lexical, and init it
552 my $x : Foo = $y; # similar, with an attribute applied
554 B<WARNING>: The use of attribute lists on C<my> declarations is still
555 evolving. The current semantics and interface are subject to change.
556 See L<attributes> and L<Attribute::Handlers>.
558 The C<my> operator declares the listed variables to be lexically
559 confined to the enclosing block, conditional
560 (C<if>/C<unless>/C<elsif>/C<else>), loop
561 (C<for>/C<foreach>/C<while>/C<until>/C<continue>), subroutine, C<eval>,
562 or C<do>/C<require>/C<use>'d file. If more than one value is listed, the
563 list must be placed in parentheses. All listed elements must be
564 legal lvalues. Only alphanumeric identifiers may be lexically
565 scoped--magical built-ins like C<$/> must currently be C<local>ized
566 with C<local> instead.
568 Unlike dynamic variables created by the C<local> operator, lexical
569 variables declared with C<my> are totally hidden from the outside
570 world, including any called subroutines. This is true if it's the
571 same subroutine called from itself or elsewhere--every call gets
575 This doesn't mean that a C<my> variable declared in a statically
576 enclosing lexical scope would be invisible. Only dynamic scopes
577 are cut off. For example, the C<bumpx()> function below has access
578 to the lexical $x variable because both the C<my> and the C<sub>
579 occurred at the same scope, presumably file scope.
584 An C<eval()>, however, can see lexical variables of the scope it is
585 being evaluated in, so long as the names aren't hidden by declarations within
586 the C<eval()> itself. See L<perlref>.
589 The parameter list to my() may be assigned to if desired, which allows you
590 to initialize your variables. (If no initializer is given for a
591 particular variable, it is created with the undefined value.) Commonly
592 this is used to name input parameters to a subroutine. Examples:
594 $arg = "fred"; # "global" variable
596 print "$arg thinks the root is $n\n";
597 fred thinks the root is 3
600 my $arg = shift; # name doesn't matter
605 The C<my> is simply a modifier on something you might assign to. So when
606 you do assign to variables in its argument list, C<my> doesn't
607 change whether those variables are viewed as a scalar or an array. So
609 my ($foo) = <STDIN>; # WRONG?
612 both supply a list context to the right-hand side, while
616 supplies a scalar context. But the following declares only one variable:
618 my $foo, $bar = 1; # WRONG
620 That has the same effect as
625 The declared variable is not introduced (is not visible) until after
626 the current statement. Thus,
630 can be used to initialize a new $x with the value of the old $x, and
633 my $x = 123 and $x == 123
635 is false unless the old $x happened to have the value C<123>.
637 Lexical scopes of control structures are not bounded precisely by the
638 braces that delimit their controlled blocks; control expressions are
639 part of that scope, too. Thus in the loop
641 while (my $line = <>) {
647 the scope of $line extends from its declaration throughout the rest of
648 the loop construct (including the C<continue> clause), but not beyond
649 it. Similarly, in the conditional
651 if ((my $answer = <STDIN>) =~ /^yes$/i) {
653 } elsif ($answer =~ /^no$/i) {
657 die "'$answer' is neither 'yes' nor 'no'";
660 the scope of $answer extends from its declaration through the rest
661 of that conditional, including any C<elsif> and C<else> clauses,
662 but not beyond it. See L<perlsyn/"Simple Statements"> for information
663 on the scope of variables in statements with modifiers.
665 The C<foreach> loop defaults to scoping its index variable dynamically
666 in the manner of C<local>. However, if the index variable is
667 prefixed with the keyword C<my>, or if there is already a lexical
668 by that name in scope, then a new lexical is created instead. Thus
672 for my $i (1, 2, 3) {
676 the scope of $i extends to the end of the loop, but not beyond it,
677 rendering the value of $i inaccessible within C<some_function()>.
680 Some users may wish to encourage the use of lexically scoped variables.
681 As an aid to catching implicit uses to package variables,
682 which are always global, if you say
686 then any variable mentioned from there to the end of the enclosing
687 block must either refer to a lexical variable, be predeclared via
688 C<our> or C<use vars>, or else must be fully qualified with the package name.
689 A compilation error results otherwise. An inner block may countermand
690 this with C<no strict 'vars'>.
692 A C<my> has both a compile-time and a run-time effect. At compile
693 time, the compiler takes notice of it. The principal usefulness
694 of this is to quiet C<use strict 'vars'>, but it is also essential
695 for generation of closures as detailed in L<perlref>. Actual
696 initialization is delayed until run time, though, so it gets executed
697 at the appropriate time, such as each time through a loop, for
700 Variables declared with C<my> are not part of any package and are therefore
701 never fully qualified with the package name. In particular, you're not
702 allowed to try to make a package variable (or other global) lexical:
704 my $pack::var; # ERROR! Illegal syntax
706 In fact, a dynamic variable (also known as package or global variables)
707 are still accessible using the fully qualified C<::> notation even while a
708 lexical of the same name is also visible:
713 print "$x and $::x\n";
715 That will print out C<20> and C<10>.
717 You may declare C<my> variables at the outermost scope of a file
718 to hide any such identifiers from the world outside that file. This
719 is similar in spirit to C's static variables when they are used at
720 the file level. To do this with a subroutine requires the use of
721 a closure (an anonymous function that accesses enclosing lexicals).
722 If you want to create a private subroutine that cannot be called
723 from outside that block, it can declare a lexical variable containing
724 an anonymous sub reference:
726 my $secret_version = '1.001-beta';
727 my $secret_sub = sub { print $secret_version };
730 As long as the reference is never returned by any function within the
731 module, no outside module can see the subroutine, because its name is not in
732 any package's symbol table. Remember that it's not I<REALLY> called
733 C<$some_pack::secret_version> or anything; it's just $secret_version,
734 unqualified and unqualifiable.
736 This does not work with object methods, however; all object methods
737 have to be in the symbol table of some package to be found. See
738 L<perlref/"Function Templates"> for something of a work-around to
741 =head2 Persistent Private Variables
742 X<state> X<state variable> X<static> X<variable, persistent> X<variable, static> X<closure>
744 There are two ways to build persistent private variables in Perl 5.10.
745 First, you can simply use the C<state> feature. Or, you can use closures,
746 if you want to stay compatible with releases older than 5.10.
748 =head3 Persistent variables via state()
750 Beginning with Perl 5.10.0, you can declare variables with the C<state>
751 keyword in place of C<my>. For that to work, though, you must have
752 enabled that feature beforehand, either by using the C<feature> pragma, or
753 by using C<-E> on one-liners (see L<feature>). Beginning with Perl 5.16,
754 the C<CORE::state> form does not require the
757 The C<state> keyword creates a lexical variable (following the same scoping
758 rules as C<my>) that persists from one subroutine call to the next. If a
759 state variable resides inside an anonymous subroutine, then each copy of
760 the subroutine has its own copy of the state variable. However, the value
761 of the state variable will still persist between calls to the same copy of
762 the anonymous subroutine. (Don't forget that C<sub { ... }> creates a new
763 subroutine each time it is executed.)
765 For example, the following code maintains a private counter, incremented
766 each time the gimme_another() function is called:
769 sub gimme_another { state $x; return ++$x }
771 And this example uses anonymous subroutines to create separate counters:
775 return sub { state $x; return ++$x }
778 Also, since C<$x> is lexical, it can't be reached or modified by any Perl
781 When combined with variable declaration, simple assignment to C<state>
782 variables (as in C<state $x = 42>) is executed only the first time. When such
783 statements are evaluated subsequent times, the assignment is ignored. The
784 behavior of assignment to C<state> declarations where the left hand side
785 of the assignment involves any parentheses is currently undefined.
787 =head3 Persistent variables with closures
789 Just because a lexical variable is lexically (also called statically)
790 scoped to its enclosing block, C<eval>, or C<do> FILE, this doesn't mean that
791 within a function it works like a C static. It normally works more
792 like a C auto, but with implicit garbage collection.
794 Unlike local variables in C or C++, Perl's lexical variables don't
795 necessarily get recycled just because their scope has exited.
796 If something more permanent is still aware of the lexical, it will
797 stick around. So long as something else references a lexical, that
798 lexical won't be freed--which is as it should be. You wouldn't want
799 memory being free until you were done using it, or kept around once you
800 were done. Automatic garbage collection takes care of this for you.
802 This means that you can pass back or save away references to lexical
803 variables, whereas to return a pointer to a C auto is a grave error.
804 It also gives us a way to simulate C's function statics. Here's a
805 mechanism for giving a function private variables with both lexical
806 scoping and a static lifetime. If you do want to create something like
807 C's static variables, just enclose the whole function in an extra block,
808 and put the static variable outside the function but in the block.
813 return ++$secret_val;
816 # $secret_val now becomes unreachable by the outside
817 # world, but retains its value between calls to gimme_another
819 If this function is being sourced in from a separate file
820 via C<require> or C<use>, then this is probably just fine. If it's
821 all in the main program, you'll need to arrange for the C<my>
822 to be executed early, either by putting the whole block above
823 your main program, or more likely, placing merely a C<BEGIN>
824 code block around it to make sure it gets executed before your program
830 return ++$secret_val;
834 See L<perlmod/"BEGIN, UNITCHECK, CHECK, INIT and END"> about the
835 special triggered code blocks, C<BEGIN>, C<UNITCHECK>, C<CHECK>,
838 If declared at the outermost scope (the file scope), then lexicals
839 work somewhat like C's file statics. They are available to all
840 functions in that same file declared below them, but are inaccessible
841 from outside that file. This strategy is sometimes used in modules
842 to create private variables that the whole module can see.
844 =head2 Temporary Values via local()
845 X<local> X<scope, dynamic> X<dynamic scope> X<variable, local>
846 X<variable, temporary>
848 B<WARNING>: In general, you should be using C<my> instead of C<local>, because
849 it's faster and safer. Exceptions to this include the global punctuation
850 variables, global filehandles and formats, and direct manipulation of the
851 Perl symbol table itself. C<local> is mostly used when the current value
852 of a variable must be visible to called subroutines.
856 # localization of values
858 local $foo; # make $foo dynamically local
859 local (@wid, %get); # make list of variables local
860 local $foo = "flurp"; # make $foo dynamic, and init it
861 local @oof = @bar; # make @oof dynamic, and init it
863 local $hash{key} = "val"; # sets a local value for this hash entry
864 delete local $hash{key}; # delete this entry for the current block
865 local ($cond ? $v1 : $v2); # several types of lvalues support
868 # localization of symbols
870 local *FH; # localize $FH, @FH, %FH, &FH ...
871 local *merlyn = *randal; # now $merlyn is really $randal, plus
872 # @merlyn is really @randal, etc
873 local *merlyn = 'randal'; # SAME THING: promote 'randal' to *randal
874 local *merlyn = \$randal; # just alias $merlyn, not @merlyn etc
876 A C<local> modifies its listed variables to be "local" to the
877 enclosing block, C<eval>, or C<do FILE>--and to I<any subroutine
878 called from within that block>. A C<local> just gives temporary
879 values to global (meaning package) variables. It does I<not> create
880 a local variable. This is known as dynamic scoping. Lexical scoping
881 is done with C<my>, which works more like C's auto declarations.
883 Some types of lvalues can be localized as well: hash and array elements
884 and slices, conditionals (provided that their result is always
885 localizable), and symbolic references. As for simple variables, this
886 creates new, dynamically scoped values.
888 If more than one variable or expression is given to C<local>, they must be
889 placed in parentheses. This operator works
890 by saving the current values of those variables in its argument list on a
891 hidden stack and restoring them upon exiting the block, subroutine, or
892 eval. This means that called subroutines can also reference the local
893 variable, but not the global one. The argument list may be assigned to if
894 desired, which allows you to initialize your local variables. (If no
895 initializer is given for a particular variable, it is created with an
898 Because C<local> is a run-time operator, it gets executed each time
899 through a loop. Consequently, it's more efficient to localize your
900 variables outside the loop.
902 =head3 Grammatical note on local()
905 A C<local> is simply a modifier on an lvalue expression. When you assign to
906 a C<local>ized variable, the C<local> doesn't change whether its list is viewed
907 as a scalar or an array. So
909 local($foo) = <STDIN>;
910 local @FOO = <STDIN>;
912 both supply a list context to the right-hand side, while
914 local $foo = <STDIN>;
916 supplies a scalar context.
918 =head3 Localization of special variables
919 X<local, special variable>
921 If you localize a special variable, you'll be giving a new value to it,
922 but its magic won't go away. That means that all side-effects related
923 to this magic still work with the localized value.
925 This feature allows code like this to work :
927 # Read the whole contents of FILE in $slurp
928 { local $/ = undef; $slurp = <FILE>; }
930 Note, however, that this restricts localization of some values ; for
931 example, the following statement dies, as of perl 5.10.0, with an error
932 I<Modification of a read-only value attempted>, because the $1 variable is
933 magical and read-only :
937 One exception is the default scalar variable: starting with perl 5.14
938 C<local($_)> will always strip all magic from $_, to make it possible
939 to safely reuse $_ in a subroutine.
941 B<WARNING>: Localization of tied arrays and hashes does not currently
943 This will be fixed in a future release of Perl; in the meantime, avoid
944 code that relies on any particular behavior of localising tied arrays
945 or hashes (localising individual elements is still okay).
946 See L<perl58delta/"Localising Tied Arrays and Hashes Is Broken"> for more
950 =head3 Localization of globs
951 X<local, glob> X<glob>
957 creates a whole new symbol table entry for the glob C<name> in the
958 current package. That means that all variables in its glob slot ($name,
959 @name, %name, &name, and the C<name> filehandle) are dynamically reset.
961 This implies, among other things, that any magic eventually carried by
962 those variables is locally lost. In other words, saying C<local */>
963 will not have any effect on the internal value of the input record
966 =head3 Localization of elements of composite types
967 X<local, composite type element> X<local, array element> X<local, hash element>
969 It's also worth taking a moment to explain what happens when you
970 C<local>ize a member of a composite type (i.e. an array or hash element).
971 In this case, the element is C<local>ized I<by name>. This means that
972 when the scope of the C<local()> ends, the saved value will be
973 restored to the hash element whose key was named in the C<local()>, or
974 the array element whose index was named in the C<local()>. If that
975 element was deleted while the C<local()> was in effect (e.g. by a
976 C<delete()> from a hash or a C<shift()> of an array), it will spring
977 back into existence, possibly extending an array and filling in the
978 skipped elements with C<undef>. For instance, if you say
980 %hash = ( 'This' => 'is', 'a' => 'test' );
984 local($hash{'a'}) = 'drill';
985 while (my $e = pop(@ary)) {
990 $hash{'only a'} = 'test';
994 print join(' ', map { "$_ $hash{$_}" } sort keys %hash),".\n";
995 print "The array has ",scalar(@ary)," elements: ",
996 join(', ', map { defined $_ ? $_ : 'undef' } @ary),"\n";
1003 This is a test only a test.
1004 The array has 6 elements: 0, 1, 2, undef, undef, 5
1006 The behavior of local() on non-existent members of composite
1007 types is subject to change in future. The behavior of local()
1008 on array elements specified using negative indexes is particularly
1009 surprising, and is very likely to change.
1011 =head3 Localized deletion of elements of composite types
1012 X<delete> X<local, composite type element> X<local, array element> X<local, hash element>
1014 You can use the C<delete local $array[$idx]> and C<delete local $hash{key}>
1015 constructs to delete a composite type entry for the current block and restore
1016 it when it ends. They return the array/hash value before the localization,
1017 which means that they are respectively equivalent to
1020 my $val = $array[$idx];
1022 delete $array[$idx];
1029 my $val = $hash{key};
1035 except that for those the C<local> is
1036 scoped to the C<do> block. Slices are
1045 my $a = delete local $hash{a};
1050 my @nums = delete local @$a[0, 2]
1052 # $a is [ undef, 8 ]
1054 $a[0] = 999; # will be erased when the scope ends
1056 # $a is back to [ 7, 8, 9 ]
1059 # %hash is back to its original state
1061 This construct is supported since Perl v5.12.
1063 =head2 Lvalue subroutines
1064 X<lvalue> X<subroutine, lvalue>
1066 It is possible to return a modifiable value from a subroutine.
1067 To do this, you have to declare the subroutine to return an lvalue.
1070 sub canmod : lvalue {
1071 $val; # or: return $val;
1077 canmod() = 5; # assigns to $val
1078 nomod() = 5; # ERROR
1080 The scalar/list context for the subroutine and for the right-hand
1081 side of assignment is determined as if the subroutine call is replaced
1082 by a scalar. For example, consider:
1084 data(2,3) = get_data(3,4);
1086 Both subroutines here are called in a scalar context, while in:
1088 (data(2,3)) = get_data(3,4);
1092 (data(2),data(3)) = get_data(3,4);
1094 all the subroutines are called in a list context.
1096 Lvalue subroutines are convenient, but you have to keep in mind that,
1097 when used with objects, they may violate encapsulation. A normal
1098 mutator can check the supplied argument before setting the attribute
1099 it is protecting, an lvalue subroutine cannot. If you require any
1100 special processing when storing and retrieving the values, consider
1101 using the CPAN module Sentinel or something similar.
1103 =head2 Lexical Subroutines
1104 X<my sub> X<state sub> X<our sub> X<subroutine, lexical>
1106 Beginning with Perl 5.18, you can declare a private subroutine with C<my>
1107 or C<state>. As with state variables, the C<state> keyword is only
1108 available under C<use feature 'state'> or C<use 5.010> or higher.
1110 Prior to Perl 5.26, lexical subroutines were deemed experimental and were
1111 available only under the C<use feature 'lexical_subs'> pragma. They also
1112 produced a warning unless the "experimental::lexical_subs" warnings
1113 category was disabled.
1115 These subroutines are only visible within the block in which they are
1116 declared, and only after that declaration:
1118 # Include these two lines if your code is intended to run under Perl
1119 # versions earlier than 5.26.
1120 no warnings "experimental::lexical_subs";
1121 use feature 'lexical_subs';
1123 foo(); # calls the package/global subroutine
1125 foo(); # also calls the package subroutine
1127 foo(); # calls "state" sub
1128 my $ref = \&foo; # take a reference to "state" sub
1131 bar(); # calls "my" sub
1133 You can't (directly) write a recursive lexical subroutine:
1140 This example fails because C<baz()> refers to the package/global subroutine
1141 C<baz>, not the lexical subroutine currently being defined.
1143 The solution is to use L<C<__SUB__>|perlfunc/__SUB__>:
1146 __SUB__->(); # calls itself
1149 It is possible to predeclare a lexical subroutine. The C<sub foo {...}>
1150 subroutine definition syntax respects any previous C<my sub;> or C<state sub;>
1151 declaration. Using this to define recursive subroutines is a bad idea,
1154 my sub baz; # predeclaration
1155 sub baz { # define the "my" sub
1156 baz(); # WRONG: calls itself, but leaks memory
1159 Just like C<< my $f; $f = sub { $f->() } >>, this example leaks memory. The
1160 name C<baz> is a reference to the subroutine, and the subroutine uses the name
1161 C<baz>; they keep each other alive (see L<perlref/Circular References>).
1163 =head3 C<state sub> vs C<my sub>
1165 What is the difference between "state" subs and "my" subs? Each time that
1166 execution enters a block when "my" subs are declared, a new copy of each
1167 sub is created. "State" subroutines persist from one execution of the
1168 containing block to the next.
1170 So, in general, "state" subroutines are faster. But "my" subs are
1171 necessary if you want to create closures:
1176 ... do something with $x ...
1181 In this example, a new C<$x> is created when C<whatever> is called, and
1182 also a new C<inner>, which can see the new C<$x>. A "state" sub will only
1183 see the C<$x> from the first call to C<whatever>.
1185 =head3 C<our> subroutines
1187 Like C<our $variable>, C<our sub> creates a lexical alias to the package
1188 subroutine of the same name.
1190 The two main uses for this are to switch back to using the package sub
1191 inside an inner scope:
1198 # need to use the outer foo here
1204 and to make a subroutine visible to other packages in the same scope:
1206 package MySneakyModule;
1208 our sub do_something { ... }
1210 sub do_something_with_caller {
1212 () = caller 1; # sets @DB::args
1213 do_something(@args); # uses MySneakyModule::do_something
1216 =head2 Passing Symbol Table Entries (typeglobs)
1219 B<WARNING>: The mechanism described in this section was originally
1220 the only way to simulate pass-by-reference in older versions of
1221 Perl. While it still works fine in modern versions, the new reference
1222 mechanism is generally easier to work with. See below.
1224 Sometimes you don't want to pass the value of an array to a subroutine
1225 but rather the name of it, so that the subroutine can modify the global
1226 copy of it rather than working with a local copy. In perl you can
1227 refer to all objects of a particular name by prefixing the name
1228 with a star: C<*foo>. This is often known as a "typeglob", because the
1229 star on the front can be thought of as a wildcard match for all the
1230 funny prefix characters on variables and subroutines and such.
1232 When evaluated, the typeglob produces a scalar value that represents
1233 all the objects of that name, including any filehandle, format, or
1234 subroutine. When assigned to, it causes the name mentioned to refer to
1235 whatever C<*> value was assigned to it. Example:
1238 local(*someary) = @_;
1239 foreach $elem (@someary) {
1246 Scalars are already passed by reference, so you can modify
1247 scalar arguments without using this mechanism by referring explicitly
1248 to C<$_[0]> etc. You can modify all the elements of an array by passing
1249 all the elements as scalars, but you have to use the C<*> mechanism (or
1250 the equivalent reference mechanism) to C<push>, C<pop>, or change the size of
1251 an array. It will certainly be faster to pass the typeglob (or reference).
1253 Even if you don't want to modify an array, this mechanism is useful for
1254 passing multiple arrays in a single LIST, because normally the LIST
1255 mechanism will merge all the array values so that you can't extract out
1256 the individual arrays. For more on typeglobs, see
1257 L<perldata/"Typeglobs and Filehandles">.
1259 =head2 When to Still Use local()
1260 X<local> X<variable, local>
1262 Despite the existence of C<my>, there are still three places where the
1263 C<local> operator still shines. In fact, in these three places, you
1264 I<must> use C<local> instead of C<my>.
1270 You need to give a global variable a temporary value, especially $_.
1272 The global variables, like C<@ARGV> or the punctuation variables, must be
1273 C<local>ized with C<local()>. This block reads in F</etc/motd>, and splits
1274 it up into chunks separated by lines of equal signs, which are placed
1278 local @ARGV = ("/etc/motd");
1281 @Fields = split /^\s*=+\s*$/;
1284 It particular, it's important to C<local>ize $_ in any routine that assigns
1285 to it. Look out for implicit assignments in C<while> conditionals.
1289 You need to create a local file or directory handle or a local function.
1291 A function that needs a filehandle of its own must use
1292 C<local()> on a complete typeglob. This can be used to create new symbol
1296 local (*READER, *WRITER); # not my!
1297 pipe (READER, WRITER) or die "pipe: $!";
1298 return (*READER, *WRITER);
1300 ($head, $tail) = ioqueue();
1302 See the Symbol module for a way to create anonymous symbol table
1305 Because assignment of a reference to a typeglob creates an alias, this
1306 can be used to create what is effectively a local function, or at least,
1310 local *grow = \&shrink; # only until this block exits
1311 grow(); # really calls shrink()
1312 move(); # if move() grow()s, it shrink()s too
1314 grow(); # get the real grow() again
1316 See L<perlref/"Function Templates"> for more about manipulating
1317 functions by name in this way.
1321 You want to temporarily change just one element of an array or hash.
1323 You can C<local>ize just one element of an aggregate. Usually this
1324 is done on dynamics:
1327 local $SIG{INT} = 'IGNORE';
1328 funct(); # uninterruptible
1330 # interruptibility automatically restored here
1332 But it also works on lexically declared aggregates.
1336 =head2 Pass by Reference
1337 X<pass by reference> X<pass-by-reference> X<reference>
1339 If you want to pass more than one array or hash into a function--or
1340 return them from it--and have them maintain their integrity, then
1341 you're going to have to use an explicit pass-by-reference. Before you
1342 do that, you need to understand references as detailed in L<perlref>.
1343 This section may not make much sense to you otherwise.
1345 Here are a few simple examples. First, let's pass in several arrays
1346 to a function and have it C<pop> all of then, returning a new list
1347 of all their former last elements:
1349 @tailings = popmany ( \@a, \@b, \@c, \@d );
1354 foreach $aref ( @_ ) {
1355 push @retlist, pop @$aref;
1360 Here's how you might write a function that returns a
1361 list of keys occurring in all the hashes passed to it:
1363 @common = inter( \%foo, \%bar, \%joe );
1365 my ($k, $href, %seen); # locals
1366 foreach $href (@_) {
1367 while ( $k = each %$href ) {
1371 return grep { $seen{$_} == @_ } keys %seen;
1374 So far, we're using just the normal list return mechanism.
1375 What happens if you want to pass or return a hash? Well,
1376 if you're using only one of them, or you don't mind them
1377 concatenating, then the normal calling convention is ok, although
1380 Where people get into trouble is here:
1382 (@a, @b) = func(@c, @d);
1384 (%a, %b) = func(%c, %d);
1386 That syntax simply won't work. It sets just C<@a> or C<%a> and
1387 clears the C<@b> or C<%b>. Plus the function didn't get passed
1388 into two separate arrays or hashes: it got one long list in C<@_>,
1391 If you can arrange for everyone to deal with this through references, it's
1392 cleaner code, although not so nice to look at. Here's a function that
1393 takes two array references as arguments, returning the two array elements
1394 in order of how many elements they have in them:
1396 ($aref, $bref) = func(\@c, \@d);
1397 print "@$aref has more than @$bref\n";
1399 my ($cref, $dref) = @_;
1400 if (@$cref > @$dref) {
1401 return ($cref, $dref);
1403 return ($dref, $cref);
1407 It turns out that you can actually do this also:
1409 (*a, *b) = func(\@c, \@d);
1410 print "@a has more than @b\n";
1412 local (*c, *d) = @_;
1420 Here we're using the typeglobs to do symbol table aliasing. It's
1421 a tad subtle, though, and also won't work if you're using C<my>
1422 variables, because only globals (even in disguise as C<local>s)
1423 are in the symbol table.
1425 If you're passing around filehandles, you could usually just use the bare
1426 typeglob, like C<*STDOUT>, but typeglobs references work, too.
1432 print $fh "her um well a hmmm\n";
1435 $rec = get_rec(\*STDIN);
1438 return scalar <$fh>;
1441 If you're planning on generating new filehandles, you could do this.
1442 Notice to pass back just the bare *FH, not its reference.
1447 return open (FH, $path) ? *FH : undef;
1451 X<prototype> X<subroutine, prototype>
1453 Perl supports a very limited kind of compile-time argument checking
1454 using function prototyping. This can be declared in either the PROTO
1455 section or with a L<prototype attribute|attributes/Built-in Attributes>.
1456 If you declare either of
1459 sub mypush :prototype(\@@)
1461 then C<mypush()> takes arguments exactly like C<push()> does.
1463 If subroutine signatures are enabled (see L</Signatures>), then
1464 the shorter PROTO syntax is unavailable, because it would clash with
1465 signatures. In that case, a prototype can only be declared in the form
1469 function declaration must be visible at compile time. The prototype
1470 affects only interpretation of new-style calls to the function,
1471 where new-style is defined as not using the C<&> character. In
1472 other words, if you call it like a built-in function, then it behaves
1473 like a built-in function. If you call it like an old-fashioned
1474 subroutine, then it behaves like an old-fashioned subroutine. It
1475 naturally falls out from this rule that prototypes have no influence
1476 on subroutine references like C<\&foo> or on indirect subroutine
1477 calls like C<&{$subref}> or C<< $subref->() >>.
1479 Method calls are not influenced by prototypes either, because the
1480 function to be called is indeterminate at compile time, since
1481 the exact code called depends on inheritance.
1483 Because the intent of this feature is primarily to let you define
1484 subroutines that work like built-in functions, here are prototypes
1485 for some other functions that parse almost exactly like the
1486 corresponding built-in.
1488 Declared as Called as
1490 sub mylink ($$) mylink $old, $new
1491 sub myvec ($$$) myvec $var, $offset, 1
1492 sub myindex ($$;$) myindex &getstring, "substr"
1493 sub mysyswrite ($$$;$) mysyswrite $buf, 0, length($buf) - $off, $off
1494 sub myreverse (@) myreverse $a, $b, $c
1495 sub myjoin ($@) myjoin ":", $a, $b, $c
1496 sub mypop (\@) mypop @array
1497 sub mysplice (\@$$@) mysplice @array, 0, 2, @pushme
1498 sub mykeys (\[%@]) mykeys $hashref->%*
1499 sub myopen (*;$) myopen HANDLE, $name
1500 sub mypipe (**) mypipe READHANDLE, WRITEHANDLE
1501 sub mygrep (&@) mygrep { /foo/ } $a, $b, $c
1502 sub myrand (;$) myrand 42
1503 sub mytime () mytime
1505 Any backslashed prototype character represents an actual argument
1506 that must start with that character (optionally preceded by C<my>,
1507 C<our> or C<local>), with the exception of C<$>, which will
1508 accept any scalar lvalue expression, such as C<$foo = 7> or
1509 C<< my_function()->[0] >>. The value passed as part of C<@_> will be a
1510 reference to the actual argument given in the subroutine call,
1511 obtained by applying C<\> to that argument.
1513 You can use the C<\[]> backslash group notation to specify more than one
1514 allowed argument type. For example:
1516 sub myref (\[$@%&*])
1518 will allow calling myref() as
1526 and the first argument of myref() will be a reference to
1527 a scalar, an array, a hash, a code, or a glob.
1529 Unbackslashed prototype characters have special meanings. Any
1530 unbackslashed C<@> or C<%> eats all remaining arguments, and forces
1531 list context. An argument represented by C<$> forces scalar context. An
1532 C<&> requires an anonymous subroutine, which, if passed as the first
1533 argument, does not require the C<sub> keyword or a subsequent comma.
1535 A C<*> allows the subroutine to accept a bareword, constant, scalar expression,
1536 typeglob, or a reference to a typeglob in that slot. The value will be
1537 available to the subroutine either as a simple scalar, or (in the latter
1538 two cases) as a reference to the typeglob. If you wish to always convert
1539 such arguments to a typeglob reference, use Symbol::qualify_to_ref() as
1542 use Symbol 'qualify_to_ref';
1545 my $fh = qualify_to_ref(shift, caller);
1549 The C<+> prototype is a special alternative to C<$> that will act like
1550 C<\[@%]> when given a literal array or hash variable, but will otherwise
1551 force scalar context on the argument. This is useful for functions which
1552 should accept either a literal array or an array reference as the argument:
1556 die "Not an array or arrayref" unless ref $aref eq 'ARRAY';
1560 When using the C<+> prototype, your function must check that the argument
1561 is of an acceptable type.
1563 A semicolon (C<;>) separates mandatory arguments from optional arguments.
1564 It is redundant before C<@> or C<%>, which gobble up everything else.
1566 As the last character of a prototype, or just before a semicolon, a C<@>
1567 or a C<%>, you can use C<_> in place of C<$>: if this argument is not
1568 provided, C<$_> will be used instead.
1570 Note how the last three examples in the table above are treated
1571 specially by the parser. C<mygrep()> is parsed as a true list
1572 operator, C<myrand()> is parsed as a true unary operator with unary
1573 precedence the same as C<rand()>, and C<mytime()> is truly without
1574 arguments, just like C<time()>. That is, if you say
1578 you'll get C<mytime() + 2>, not C<mytime(2)>, which is how it would be parsed
1579 without a prototype. If you want to force a unary function to have the
1580 same precedence as a list operator, add C<;> to the end of the prototype:
1582 sub mygetprotobynumber($;);
1583 mygetprotobynumber $a > $b; # parsed as mygetprotobynumber($a > $b)
1585 The interesting thing about C<&> is that you can generate new syntax with it,
1586 provided it's in the initial position:
1590 my($try,$catch) = @_;
1597 sub catch (&) { $_[0] }
1602 /phooey/ and print "unphooey\n";
1605 That prints C<"unphooey">. (Yes, there are still unresolved
1606 issues having to do with visibility of C<@_>. I'm ignoring that
1607 question for the moment. (But note that if we make C<@_> lexically
1608 scoped, those anonymous subroutines can act like closures... (Gee,
1609 is this sounding a little Lispish? (Never mind.))))
1611 And here's a reimplementation of the Perl C<grep> operator:
1618 push(@result, $_) if &$code;
1623 Some folks would prefer full alphanumeric prototypes. Alphanumerics have
1624 been intentionally left out of prototypes for the express purpose of
1625 someday in the future adding named, formal parameters. The current
1626 mechanism's main goal is to let module writers provide better diagnostics
1627 for module users. Larry feels the notation quite understandable to Perl
1628 programmers, and that it will not intrude greatly upon the meat of the
1629 module, nor make it harder to read. The line noise is visually
1630 encapsulated into a small pill that's easy to swallow.
1632 If you try to use an alphanumeric sequence in a prototype you will
1633 generate an optional warning - "Illegal character in prototype...".
1634 Unfortunately earlier versions of Perl allowed the prototype to be
1635 used as long as its prefix was a valid prototype. The warning may be
1636 upgraded to a fatal error in a future version of Perl once the
1637 majority of offending code is fixed.
1639 It's probably best to prototype new functions, not retrofit prototyping
1640 into older ones. That's because you must be especially careful about
1641 silent impositions of differing list versus scalar contexts. For example,
1642 if you decide that a function should take just one parameter, like this:
1646 print "you gave me $n\n";
1649 and someone has been calling it with an array or expression
1653 func( $text =~ /\w+/g );
1655 Then you've just supplied an automatic C<scalar> in front of their
1656 argument, which can be more than a bit surprising. The old C<@foo>
1657 which used to hold one thing doesn't get passed in. Instead,
1658 C<func()> now gets passed in a C<1>; that is, the number of elements
1659 in C<@foo>. And the C<m//g> gets called in scalar context so instead of a
1660 list of words it returns a boolean result and advances C<pos($text)>. Ouch!
1662 If a sub has both a PROTO and a BLOCK, the prototype is not applied
1663 until after the BLOCK is completely defined. This means that a recursive
1664 function with a prototype has to be predeclared for the prototype to take
1672 This is all very powerful, of course, and should be used only in moderation
1673 to make the world a better place.
1675 =head2 Constant Functions
1678 Functions with a prototype of C<()> are potential candidates for
1679 inlining. If the result after optimization and constant folding
1680 is either a constant or a lexically-scoped scalar which has no other
1681 references, then it will be used in place of function calls made
1682 without C<&>. Calls made using C<&> are never inlined. (See
1683 L<constant> for an easy way to declare most constants.)
1685 The following functions would all be inlined:
1687 sub pi () { 3.14159 } # Not exact, but close.
1688 sub PI () { 4 * atan2 1, 1 } # As good as it gets,
1689 # and it's inlined, too!
1693 sub FLAG_FOO () { 1 << 8 }
1694 sub FLAG_BAR () { 1 << 9 }
1695 sub FLAG_MASK () { FLAG_FOO | FLAG_BAR }
1697 sub OPT_BAZ () { not (0x1B58 & FLAG_MASK) }
1699 sub N () { int(OPT_BAZ) / 3 }
1701 sub FOO_SET () { 1 if FLAG_MASK & FLAG_FOO }
1702 sub FOO_SET2 () { if (FLAG_MASK & FLAG_FOO) { 1 } }
1704 (Be aware that the last example was not always inlined in Perl 5.20 and
1705 earlier, which did not behave consistently with subroutines containing
1706 inner scopes.) You can countermand inlining by using an explicit
1717 sub bonk_val () { return 12345 }
1719 As alluded to earlier you can also declare inlined subs dynamically at
1720 BEGIN time if their body consists of a lexically-scoped scalar which
1721 has no other references. Only the first example here will be inlined:
1726 *INLINED = sub () { $var };
1733 *NOT_INLINED = sub () { $var };
1736 A not so obvious caveat with this (see [RT #79908]) is that the
1737 variable will be immediately inlined, and will stop behaving like a
1738 normal lexical variable, e.g. this will print C<79907>, not C<79908>:
1742 *RT_79908 = sub () { $x };
1745 print RT_79908(); # prints 79907
1747 As of Perl 5.22, this buggy behavior, while preserved for backward
1748 compatibility, is detected and emits a deprecation warning. If you want
1749 the subroutine to be inlined (with no warning), make sure the variable is
1750 not used in a context where it could be modified aside from where it is
1756 *INLINED = sub () { $x };
1758 # Warns. Future Perl versions will stop inlining it.
1762 *ALSO_INLINED = sub () { $x };
1765 Perl 5.22 also introduces the experimental "const" attribute as an
1766 alternative. (Disable the "experimental::const_attr" warnings if you want
1767 to use it.) When applied to an anonymous subroutine, it forces the sub to
1768 be called when the C<sub> expression is evaluated. The return value is
1769 captured and turned into a constant subroutine:
1772 *INLINED = sub : const { $x };
1775 The return value of C<INLINED> in this example will always be 54321,
1776 regardless of later modifications to $x. You can also put any arbitrary
1777 code inside the sub, at it will be executed immediately and its return
1778 value captured the same way.
1780 If you really want a subroutine with a C<()> prototype that returns a
1781 lexical variable you can easily force it to not be inlined by adding
1782 an explicit C<return>:
1786 *RT_79908 = sub () { return $x };
1789 print RT_79908(); # prints 79908
1791 The easiest way to tell if a subroutine was inlined is by using
1792 L<B::Deparse>. Consider this example of two subroutines returning
1793 C<1>, one with a C<()> prototype causing it to be inlined, and one
1794 without (with deparse output truncated for clarity):
1796 $ perl -MO=Deparse -le 'sub ONE { 1 } if (ONE) { print ONE if ONE }'
1801 print ONE() if ONE ;
1803 $ perl -MO=Deparse -le 'sub ONE () { 1 } if (ONE) { print ONE if ONE }'
1809 If you redefine a subroutine that was eligible for inlining, you'll
1810 get a warning by default. You can use this warning to tell whether or
1811 not a particular subroutine is considered inlinable, since it's
1812 different than the warning for overriding non-inlined subroutines:
1814 $ perl -e 'sub one () {1} sub one () {2}'
1815 Constant subroutine one redefined at -e line 1.
1816 $ perl -we 'sub one {1} sub one {2}'
1817 Subroutine one redefined at -e line 1.
1819 The warning is considered severe enough not to be affected by the
1820 B<-w> switch (or its absence) because previously compiled invocations
1821 of the function will still be using the old value of the function. If
1822 you need to be able to redefine the subroutine, you need to ensure
1823 that it isn't inlined, either by dropping the C<()> prototype (which
1824 changes calling semantics, so beware) or by thwarting the inlining
1825 mechanism in some other way, e.g. by adding an explicit C<return>, as
1828 sub not_inlined () { return 23 }
1830 =head2 Overriding Built-in Functions
1831 X<built-in> X<override> X<CORE> X<CORE::GLOBAL>
1833 Many built-in functions may be overridden, though this should be tried
1834 only occasionally and for good reason. Typically this might be
1835 done by a package attempting to emulate missing built-in functionality
1836 on a non-Unix system.
1838 Overriding may be done only by importing the name from a module at
1839 compile time--ordinary predeclaration isn't good enough. However, the
1840 C<use subs> pragma lets you, in effect, predeclare subs
1841 via the import syntax, and these names may then override built-in ones:
1843 use subs 'chdir', 'chroot', 'chmod', 'chown';
1847 To unambiguously refer to the built-in form, precede the
1848 built-in name with the special package qualifier C<CORE::>. For example,
1849 saying C<CORE::open()> always refers to the built-in C<open()>, even
1850 if the current package has imported some other subroutine called
1851 C<&open()> from elsewhere. Even though it looks like a regular
1852 function call, it isn't: the CORE:: prefix in that case is part of Perl's
1853 syntax, and works for any keyword, regardless of what is in the CORE
1854 package. Taking a reference to it, that is, C<\&CORE::open>, only works
1855 for some keywords. See L<CORE>.
1857 Library modules should not in general export built-in names like C<open>
1858 or C<chdir> as part of their default C<@EXPORT> list, because these may
1859 sneak into someone else's namespace and change the semantics unexpectedly.
1860 Instead, if the module adds that name to C<@EXPORT_OK>, then it's
1861 possible for a user to import the name explicitly, but not implicitly.
1862 That is, they could say
1866 and it would import the C<open> override. But if they said
1870 they would get the default imports without overrides.
1872 The foregoing mechanism for overriding built-in is restricted, quite
1873 deliberately, to the package that requests the import. There is a second
1874 method that is sometimes applicable when you wish to override a built-in
1875 everywhere, without regard to namespace boundaries. This is achieved by
1876 importing a sub into the special namespace C<CORE::GLOBAL::>. Here is an
1877 example that quite brazenly replaces the C<glob> operator with something
1878 that understands regular expressions.
1883 @EXPORT_OK = 'glob';
1889 my $where = ($sym =~ s/^GLOBAL_// ? 'CORE::GLOBAL' : caller(0));
1890 $pkg->export($where, $sym, @_);
1896 if (opendir my $d, '.') {
1897 @got = grep /$pat/, readdir $d;
1904 And here's how it could be (ab)used:
1906 #use REGlob 'GLOBAL_glob'; # override glob() in ALL namespaces
1908 use REGlob 'glob'; # override glob() in Foo:: only
1909 print for <^[a-z_]+\.pm\$>; # show all pragmatic modules
1911 The initial comment shows a contrived, even dangerous example.
1912 By overriding C<glob> globally, you would be forcing the new (and
1913 subversive) behavior for the C<glob> operator for I<every> namespace,
1914 without the complete cognizance or cooperation of the modules that own
1915 those namespaces. Naturally, this should be done with extreme caution--if
1916 it must be done at all.
1918 The C<REGlob> example above does not implement all the support needed to
1919 cleanly override perl's C<glob> operator. The built-in C<glob> has
1920 different behaviors depending on whether it appears in a scalar or list
1921 context, but our C<REGlob> doesn't. Indeed, many perl built-in have such
1922 context sensitive behaviors, and these must be adequately supported by
1923 a properly written override. For a fully functional example of overriding
1924 C<glob>, study the implementation of C<File::DosGlob> in the standard
1927 When you override a built-in, your replacement should be consistent (if
1928 possible) with the built-in native syntax. You can achieve this by using
1929 a suitable prototype. To get the prototype of an overridable built-in,
1930 use the C<prototype> function with an argument of C<"CORE::builtin_name">
1931 (see L<perlfunc/prototype>).
1933 Note however that some built-ins can't have their syntax expressed by a
1934 prototype (such as C<system> or C<chomp>). If you override them you won't
1935 be able to fully mimic their original syntax.
1937 The built-ins C<do>, C<require> and C<glob> can also be overridden, but due
1938 to special magic, their original syntax is preserved, and you don't have
1939 to define a prototype for their replacements. (You can't override the
1940 C<do BLOCK> syntax, though).
1942 C<require> has special additional dark magic: if you invoke your
1943 C<require> replacement as C<require Foo::Bar>, it will actually receive
1944 the argument C<"Foo/Bar.pm"> in @_. See L<perlfunc/require>.
1946 And, as you'll have noticed from the previous example, if you override
1947 C<glob>, the C<< <*> >> glob operator is overridden as well.
1949 In a similar fashion, overriding the C<readline> function also overrides
1950 the equivalent I/O operator C<< <FILEHANDLE> >>. Also, overriding
1951 C<readpipe> also overrides the operators C<``> and C<qx//>.
1953 Finally, some built-ins (e.g. C<exists> or C<grep>) can't be overridden.
1956 X<autoloading> X<AUTOLOAD>
1958 If you call a subroutine that is undefined, you would ordinarily
1959 get an immediate, fatal error complaining that the subroutine doesn't
1960 exist. (Likewise for subroutines being used as methods, when the
1961 method doesn't exist in any base class of the class's package.)
1962 However, if an C<AUTOLOAD> subroutine is defined in the package or
1963 packages used to locate the original subroutine, then that
1964 C<AUTOLOAD> subroutine is called with the arguments that would have
1965 been passed to the original subroutine. The fully qualified name
1966 of the original subroutine magically appears in the global $AUTOLOAD
1967 variable of the same package as the C<AUTOLOAD> routine. The name
1968 is not passed as an ordinary argument because, er, well, just
1969 because, that's why. (As an exception, a method call to a nonexistent
1970 C<import> or C<unimport> method is just skipped instead. Also, if
1971 the AUTOLOAD subroutine is an XSUB, there are other ways to retrieve the
1972 subroutine name. See L<perlguts/Autoloading with XSUBs> for details.)
1975 Many C<AUTOLOAD> routines load in a definition for the requested
1976 subroutine using eval(), then execute that subroutine using a special
1977 form of goto() that erases the stack frame of the C<AUTOLOAD> routine
1978 without a trace. (See the source to the standard module documented
1979 in L<AutoLoader>, for example.) But an C<AUTOLOAD> routine can
1980 also just emulate the routine and never define it. For example,
1981 let's pretend that a function that wasn't defined should just invoke
1982 C<system> with those arguments. All you'd do is:
1985 our $AUTOLOAD; # keep 'use strict' happy
1986 my $program = $AUTOLOAD;
1987 $program =~ s/.*:://;
1988 system($program, @_);
1994 In fact, if you predeclare functions you want to call that way, you don't
1995 even need parentheses:
1997 use subs qw(date who ls);
2002 A more complete example of this is the Shell module on CPAN, which
2003 can treat undefined subroutine calls as calls to external programs.
2005 Mechanisms are available to help modules writers split their modules
2006 into autoloadable files. See the standard AutoLoader module
2007 described in L<AutoLoader> and in L<AutoSplit>, the standard
2008 SelfLoader modules in L<SelfLoader>, and the document on adding C
2009 functions to Perl code in L<perlxs>.
2011 =head2 Subroutine Attributes
2012 X<attribute> X<subroutine, attribute> X<attrs>
2014 A subroutine declaration or definition may have a list of attributes
2015 associated with it. If such an attribute list is present, it is
2016 broken up at space or colon boundaries and treated as though a
2017 C<use attributes> had been seen. See L<attributes> for details
2018 about what attributes are currently supported.
2019 Unlike the limitation with the obsolescent C<use attrs>, the
2020 C<sub : ATTRLIST> syntax works to associate the attributes with
2021 a pre-declaration, and not just with a subroutine definition.
2023 The attributes must be valid as simple identifier names (without any
2024 punctuation other than the '_' character). They may have a parameter
2025 list appended, which is only checked for whether its parentheses ('(',')')
2028 Examples of valid syntax (even though the attributes are unknown):
2030 sub fnord (&\%) : switch(10,foo(7,3)) : expensive;
2031 sub plugh () : Ugly('\(") :Bad;
2032 sub xyzzy : _5x5 { ... }
2034 Examples of invalid syntax:
2036 sub fnord : switch(10,foo(); # ()-string not balanced
2037 sub snoid : Ugly('('); # ()-string not balanced
2038 sub xyzzy : 5x5; # "5x5" not a valid identifier
2039 sub plugh : Y2::north; # "Y2::north" not a simple identifier
2040 sub snurt : foo + bar; # "+" not a colon or space
2042 The attribute list is passed as a list of constant strings to the code
2043 which associates them with the subroutine. In particular, the second example
2044 of valid syntax above currently looks like this in terms of how it's
2047 use attributes __PACKAGE__, \&plugh, q[Ugly('\(")], 'Bad';
2049 For further details on attribute lists and their manipulation,
2050 see L<attributes> and L<Attribute::Handlers>.
2054 See L<perlref/"Function Templates"> for more about references and closures.
2055 See L<perlxs> if you'd like to learn about calling C subroutines from Perl.
2056 See L<perlembed> if you'd like to learn about calling Perl subroutines from C.
2057 See L<perlmod> to learn about bundling up your functions in separate files.
2058 See L<perlmodlib> to learn what library modules come standard on your system.
2059 See L<perlootut> to learn how to make object method calls.