1PERLSUB(1) Perl Programmers Reference Guide PERLSUB(1)
2
6 perlsub - Perl subroutines
7
9 To declare subroutines:
10 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
15 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
20 To define an anonymous subroutine at runtime:
22 $subref = sub BLOCK; # no proto
24 $subref = sub (PROTO) BLOCK; # with proto
25 $subref = sub : ATTRS BLOCK; # with attributes
26 $subref = sub (PROTO) : ATTRS BLOCK; # with proto and attributes
27 To import subroutines:
29 use MODULE qw(NAME1 NAME2 NAME3);
31 To call subroutines:
33 NAME(LIST); # & is optional with parentheses.
35 NAME LIST; # Parentheses optional if predeclared/imported.
36 &NAME(LIST); # Circumvent prototypes.
37 &NAME; # Makes current @_ visible to called subroutine.
38
40 Like many languages, Perl provides for user-defined subroutines. These
41 may be located anywhere in the main program, loaded in from other files
42 via the "do", "require", or "use" keywords, or generated on the fly
43 using "eval" or anonymous subroutines. You can even call a function
44 indirectly using a variable containing its name or a CODE reference.
45 The Perl model for function call and return values is simple: all func‐
47 tions are passed as parameters one single flat list of scalars, and all
48 functions likewise return to their caller one single flat list of
49 scalars. Any arrays or hashes in these call and return lists will col‐
50 lapse, losing their identities--but you may always use pass-by-refer‐
51 ence instead to avoid this. Both call and return lists may contain as
52 many or as few scalar elements as you'd like. (Often a function with‐
53 out an explicit return statement is called a subroutine, but there's
54 really no difference from Perl's perspective.)
55 Any arguments passed in show up in the array @_. Therefore, if you
57 called a function with two arguments, those would be stored in $_[0]
58 and $_[1]. The array @_ is a local array, but its elements are aliases
59 for the actual scalar parameters. In particular, if an element $_[0]
60 is updated, the corresponding argument is updated (or an error occurs
61 if it is not updatable). If an argument is an array or hash element
62 which did not exist when the function was called, that element is cre‐
63 ated only when (and if) it is modified or a reference to it is taken.
64 (Some earlier versions of Perl created the element whether or not the
65 element was assigned to.) Assigning to the whole array @_ removes that
66 aliasing, and does not update any arguments.
67 A "return" statement may be used to exit a subroutine, optionally spec‐
69 ifying the returned value, which will be evaluated in the appropriate
70 context (list, scalar, or void) depending on the context of the subrou‐
71 tine call. If you specify no return value, the subroutine returns an
72 empty list in list context, the undefined value in scalar context, or
73 nothing in void context. If you return one or more aggregates (arrays
74 and hashes), these will be flattened together into one large indistin‐
75 guishable list.
76 If no "return" is found and if the last statement is an expression, its
78 value is returned. If the last statement is a loop control structure
79 like a "foreach" or a "while", the returned value is unspecified. The
80 empty sub returns the empty list.
81 Perl does not have named formal parameters. In practice all you do is
83 assign to a "my()" list of these. Variables that aren't declared to be
84 private are global variables. For gory details on creating private
85 variables, see "Private Variables via my()" and "Temporary Values via
86 local()". To create protected environments for a set of functions in a
87 separate package (and probably a separate file), see "Packages" in
88 perlmod.
89 Example:
91 sub max {
93 my $max = shift(@_);
94 foreach $foo (@_) {
95 $max = $foo if $max < $foo;
96 }
97 return $max;
98 }
99 $bestday = max($mon,$tue,$wed,$thu,$fri);
100 Example:
102 # get a line, combining continuation lines
104 # that start with whitespace
105 sub get_line {
107 $thisline = $lookahead; # global variables!
108 LINE: while (defined($lookahead = <STDIN>)) {
109 if ($lookahead =~ /^[ \t]/) {
110 $thisline .= $lookahead;
111 }
112 else {
113 last LINE;
114 }
115 }
116 return $thisline;
117 }
118 $lookahead = <STDIN>; # get first line
120 while (defined($line = get_line())) {
121 ...
122 }
123 Assigning to a list of private variables to name your arguments:
125 sub maybeset {
127 my($key, $value) = @_;
128 $Foo{$key} = $value unless $Foo{$key};
129 }
130 Because the assignment copies the values, this also has the effect of
132 turning call-by-reference into call-by-value. Otherwise a function is
133 free to do in-place modifications of @_ and change its caller's values.
134 upcase_in($v1, $v2); # this changes $v1 and $v2
136 sub upcase_in {
137 for (@_) { tr/a-z/A-Z/ }
138 }
139 You aren't allowed to modify constants in this way, of course. If an
141 argument were actually literal and you tried to change it, you'd take a
142 (presumably fatal) exception. For example, this won't work:
143 upcase_in("frederick");
145 It would be much safer if the "upcase_in()" function were written to
147 return a copy of its parameters instead of changing them in place:
148 ($v3, $v4) = upcase($v1, $v2); # this doesn't change $v1 and $v2
150 sub upcase {
151 return unless defined wantarray; # void context, do nothing
152 my @parms = @_;
153 for (@parms) { tr/a-z/A-Z/ }
154 return wantarray ? @parms : $parms[0];
155 }
156 Notice how this (unprototyped) function doesn't care whether it was
158 passed real scalars or arrays. Perl sees all arguments as one big,
159 long, flat parameter list in @_. This is one area where Perl's simple
160 argument-passing style shines. The "upcase()" function would work per‐
161 fectly well without changing the "upcase()" definition even if we fed
162 it things like this:
163 @newlist = upcase(@list1, @list2);
165 @newlist = upcase( split /:/, $var );
166 Do not, however, be tempted to do this:
168 (@a, @b) = upcase(@list1, @list2);
170 Like the flattened incoming parameter list, the return list is also
172 flattened on return. So all you have managed to do here is stored
173 everything in @a and made @b empty. See "Pass by Reference" for alter‐
174 natives.
175 A subroutine may be called using an explicit "&" prefix. The "&" is
177 optional in modern Perl, as are parentheses if the subroutine has been
178 predeclared. The "&" is not optional when just naming the subroutine,
179 such as when it's used as an argument to defined() or undef(). Nor is
180 it optional when you want to do an indirect subroutine call with a sub‐
181 routine name or reference using the "&$subref()" or "&{$subref}()" con‐
182 structs, although the "$subref->()" notation solves that problem. See
183 perlref for more about all that.
184 Subroutines may be called recursively. If a subroutine is called using
186 the "&" form, the argument list is optional, and if omitted, no @_
187 array is set up for the subroutine: the @_ array at the time of the
188 call is visible to subroutine instead. This is an efficiency mechanism
189 that new users may wish to avoid.
190 &foo(1,2,3); # pass three arguments
192 foo(1,2,3); # the same
193 foo(); # pass a null list
195 &foo(); # the same
196 &foo; # foo() get current args, like foo(@_) !!
198 foo; # like foo() IFF sub foo predeclared, else "foo"
199 Not only does the "&" form make the argument list optional, it also
201 disables any prototype checking on arguments you do provide. This is
202 partly for historical reasons, and partly for having a convenient way
203 to cheat if you know what you're doing. See Prototypes below.
204 Subroutines whose names are in all upper case are reserved to the Perl
206 core, as are modules whose names are in all lower case. A subroutine
207 in all capitals is a loosely-held convention meaning it will be called
208 indirectly by the run-time system itself, usually due to a triggered
209 event. Subroutines that do special, pre-defined things include
210 "AUTOLOAD", "CLONE", "DESTROY" plus all functions mentioned in perltie
211 and PerlIO::via.
212 The "BEGIN", "CHECK", "INIT" and "END" subroutines are not so much sub‐
214 routines as named special code blocks, of which you can have more than
215 one in a package, and which you can not call explicitly. See "BEGIN,
216 CHECK, INIT and END" in perlmod
217 Private Variables via my()
219 Synopsis:
221 my $foo; # declare $foo lexically local
223 my (@wid, %get); # declare list of variables local
224 my $foo = "flurp"; # declare $foo lexical, and init it
225 my @oof = @bar; # declare @oof lexical, and init it
226 my $x : Foo = $y; # similar, with an attribute applied
227 WARNING: The use of attribute lists on "my" declarations is still
229 evolving. The current semantics and interface are subject to change.
230 See attributes and Attribute::Handlers.
231 The "my" operator declares the listed variables to be lexically con‐
233 fined to the enclosing block, conditional ("if/unless/elsif/else"),
234 loop ("for/foreach/while/until/continue"), subroutine, "eval", or
235 "do/require/use"'d file. If more than one value is listed, the list
236 must be placed in parentheses. All listed elements must be legal lval‐
237 ues. Only alphanumeric identifiers may be lexically scoped--magical
238 built-ins like $/ must currently be "local"ized with "local" instead.
239 Unlike dynamic variables created by the "local" operator, lexical vari‐
241 ables declared with "my" are totally hidden from the outside world,
242 including any called subroutines. This is true if it's the same sub‐
243 routine called from itself or elsewhere--every call gets its own copy.
244 This doesn't mean that a "my" variable declared in a statically enclos‐
246 ing lexical scope would be invisible. Only dynamic scopes are cut off.
247 For example, the "bumpx()" function below has access to the lexical $x
248 variable because both the "my" and the "sub" occurred at the same
249 scope, presumably file scope.
250 my $x = 10;
252 sub bumpx { $x++ }
253 An "eval()", however, can see lexical variables of the scope it is
255 being evaluated in, so long as the names aren't hidden by declarations
256 within the "eval()" itself. See perlref.
257 The parameter list to my() may be assigned to if desired, which allows
259 you to initialize your variables. (If no initializer is given for a
260 particular variable, it is created with the undefined value.) Commonly
261 this is used to name input parameters to a subroutine. Examples:
262 $arg = "fred"; # "global" variable
264 $n = cube_root(27);
265 print "$arg thinks the root is $n\n";
266 fred thinks the root is 3
267 sub cube_root {
269 my $arg = shift; # name doesn't matter
270 $arg **= 1/3;
271 return $arg;
272 }
273 The "my" is simply a modifier on something you might assign to. So
275 when you do assign to variables in its argument list, "my" doesn't
276 change whether those variables are viewed as a scalar or an array. So
277 my ($foo) = <STDIN>; # WRONG?
279 my @FOO = <STDIN>;
280 both supply a list context to the right-hand side, while
282 my $foo = <STDIN>;
284 supplies a scalar context. But the following declares only one vari‐
286 able:
287 my $foo, $bar = 1; # WRONG
289 That has the same effect as
291 my $foo;
293 $bar = 1;
294 The declared variable is not introduced (is not visible) until after
296 the current statement. Thus,
297 my $x = $x;
299 can be used to initialize a new $x with the value of the old $x, and
301 the expression
302 my $x = 123 and $x == 123
304 is false unless the old $x happened to have the value 123.
306 Lexical scopes of control structures are not bounded precisely by the
308 braces that delimit their controlled blocks; control expressions are
309 part of that scope, too. Thus in the loop
310 while (my $line = <>) {
312 $line = lc $line;
313 } continue {
314 print $line;
315 }
316 the scope of $line extends from its declaration throughout the rest of
318 the loop construct (including the "continue" clause), but not beyond
319 it. Similarly, in the conditional
320 if ((my $answer = <STDIN>) =~ /^yes$/i) {
322 user_agrees();
323 } elsif ($answer =~ /^no$/i) {
324 user_disagrees();
325 } else {
326 chomp $answer;
327 die "'$answer' is neither 'yes' nor 'no'";
328 }
329 the scope of $answer extends from its declaration through the rest of
331 that conditional, including any "elsif" and "else" clauses, but not
332 beyond it. See "Simple statements" in perlsyn for information on the
333 scope of variables in statements with modifiers.
334 The "foreach" loop defaults to scoping its index variable dynamically
336 in the manner of "local". However, if the index variable is prefixed
337 with the keyword "my", or if there is already a lexical by that name in
338 scope, then a new lexical is created instead. Thus in the loop
339 for my $i (1, 2, 3) {
341 some_function();
342 }
343 the scope of $i extends to the end of the loop, but not beyond it, ren‐
345 dering the value of $i inaccessible within "some_function()".
346 Some users may wish to encourage the use of lexically scoped variables.
348 As an aid to catching implicit uses to package variables, which are
349 always global, if you say
350 use strict 'vars';
352 then any variable mentioned from there to the end of the enclosing
354 block must either refer to a lexical variable, be predeclared via "our"
355 or "use vars", or else must be fully qualified with the package name.
356 A compilation error results otherwise. An inner block may countermand
357 this with "no strict 'vars'".
358 A "my" has both a compile-time and a run-time effect. At compile time,
360 the compiler takes notice of it. The principal usefulness of this is
361 to quiet "use strict 'vars'", but it is also essential for generation
362 of closures as detailed in perlref. Actual initialization is delayed
363 until run time, though, so it gets executed at the appropriate time,
364 such as each time through a loop, for example.
365 Variables declared with "my" are not part of any package and are there‐
367 fore never fully qualified with the package name. In particular,
368 you're not allowed to try to make a package variable (or other global)
369 lexical:
370 my $pack::var; # ERROR! Illegal syntax
372 my $_; # also illegal (currently)
373 In fact, a dynamic variable (also known as package or global variables)
375 are still accessible using the fully qualified "::" notation even while
376 a lexical of the same name is also visible:
377 package main;
379 local $x = 10;
380 my $x = 20;
381 print "$x and $::x\n";
382 That will print out 20 and 10.
384 You may declare "my" variables at the outermost scope of a file to hide
386 any such identifiers from the world outside that file. This is similar
387 in spirit to C's static variables when they are used at the file level.
388 To do this with a subroutine requires the use of a closure (an anony‐
389 mous function that accesses enclosing lexicals). If you want to create
390 a private subroutine that cannot be called from outside that block, it
391 can declare a lexical variable containing an anonymous sub reference:
392 my $secret_version = '1.001-beta';
394 my $secret_sub = sub { print $secret_version };
395 &$secret_sub();
396 As long as the reference is never returned by any function within the
398 module, no outside module can see the subroutine, because its name is
399 not in any package's symbol table. Remember that it's not REALLY
400 called $some_pack::secret_version or anything; it's just $secret_ver‐
401 sion, unqualified and unqualifiable.
402 This does not work with object methods, however; all object methods
404 have to be in the symbol table of some package to be found. See "Func‐
405 tion Templates" in perlref for something of a work-around to this.
406 Persistent Private Variables
408 Just because a lexical variable is lexically (also called statically)
410 scoped to its enclosing block, "eval", or "do" FILE, this doesn't mean
411 that within a function it works like a C static. It normally works
412 more like a C auto, but with implicit garbage collection.
413 Unlike local variables in C or C++, Perl's lexical variables don't nec‐
415 essarily get recycled just because their scope has exited. If some‐
416 thing more permanent is still aware of the lexical, it will stick
417 around. So long as something else references a lexical, that lexical
418 won't be freed--which is as it should be. You wouldn't want memory
419 being free until you were done using it, or kept around once you were
420 done. Automatic garbage collection takes care of this for you.
421 This means that you can pass back or save away references to lexical
423 variables, whereas to return a pointer to a C auto is a grave error.
424 It also gives us a way to simulate C's function statics. Here's a
425 mechanism for giving a function private variables with both lexical
426 scoping and a static lifetime. If you do want to create something like
427 C's static variables, just enclose the whole function in an extra
428 block, and put the static variable outside the function but in the
429 block.
430 {
432 my $secret_val = 0;
433 sub gimme_another {
434 return ++$secret_val;
435 }
436 }
437 # $secret_val now becomes unreachable by the outside
438 # world, but retains its value between calls to gimme_another
439 If this function is being sourced in from a separate file via "require"
441 or "use", then this is probably just fine. If it's all in the main
442 program, you'll need to arrange for the "my" to be executed early,
443 either by putting the whole block above your main program, or more
444 likely, placing merely a "BEGIN" code block around it to make sure it
445 gets executed before your program starts to run:
446 BEGIN {
448 my $secret_val = 0;
449 sub gimme_another {
450 return ++$secret_val;
451 }
452 }
453 See "BEGIN, CHECK, INIT and END" in perlmod about the special triggered
455 code blocks, "BEGIN", "CHECK", "INIT" and "END".
456 If declared at the outermost scope (the file scope), then lexicals work
458 somewhat like C's file statics. They are available to all functions in
459 that same file declared below them, but are inaccessible from outside
460 that file. This strategy is sometimes used in modules to create pri‐
461 vate variables that the whole module can see.
462 Temporary Values via local()
464 WARNING: In general, you should be using "my" instead of "local",
466 because it's faster and safer. Exceptions to this include the global
467 punctuation variables, global filehandles and formats, and direct
468 manipulation of the Perl symbol table itself. "local" is mostly used
469 when the current value of a variable must be visible to called subrou‐
470 tines.
471 Synopsis:
473 # localization of values
475 local $foo; # make $foo dynamically local
477 local (@wid, %get); # make list of variables local
478 local $foo = "flurp"; # make $foo dynamic, and init it
479 local @oof = @bar; # make @oof dynamic, and init it
480 local $hash{key} = "val"; # sets a local value for this hash entry
482 local ($cond ? $v1 : $v2); # several types of lvalues support
483 # localization
484 # localization of symbols
486 local *FH; # localize $FH, @FH, %FH, &FH ...
488 local *merlyn = *randal; # now $merlyn is really $randal, plus
489 # @merlyn is really @randal, etc
490 local *merlyn = 'randal'; # SAME THING: promote 'randal' to *randal
491 local *merlyn = \$randal; # just alias $merlyn, not @merlyn etc
492 A "local" modifies its listed variables to be "local" to the enclosing
494 block, "eval", or "do FILE"--and to any subroutine called from within
495 that block. A "local" just gives temporary values to global (meaning
496 package) variables. It does not create a local variable. This is
497 known as dynamic scoping. Lexical scoping is done with "my", which
498 works more like C's auto declarations.
499 Some types of lvalues can be localized as well : hash and array ele‐
501 ments and slices, conditionals (provided that their result is always
502 localizable), and symbolic references. As for simple variables, this
503 creates new, dynamically scoped values.
504 If more than one variable or expression is given to "local", they must
506 be placed in parentheses. This operator works by saving the current
507 values of those variables in its argument list on a hidden stack and
508 restoring them upon exiting the block, subroutine, or eval. This means
509 that called subroutines can also reference the local variable, but not
510 the global one. The argument list may be assigned to if desired, which
511 allows you to initialize your local variables. (If no initializer is
512 given for a particular variable, it is created with an undefined
513 value.)
514 Because "local" is a run-time operator, it gets executed each time
516 through a loop. Consequently, it's more efficient to localize your
517 variables outside the loop.
518 Grammatical note on local()
520 A "local" is simply a modifier on an lvalue expression. When you
522 assign to a "local"ized variable, the "local" doesn't change whether
523 its list is viewed as a scalar or an array. So
524 local($foo) = <STDIN>;
526 local @FOO = <STDIN>;
527 both supply a list context to the right-hand side, while
529 local $foo = <STDIN>;
531 supplies a scalar context.
533 Localization of special variables
535 If you localize a special variable, you'll be giving a new value to it,
537 but its magic won't go away. That means that all side-effects related
538 to this magic still work with the localized value.
539 This feature allows code like this to work :
541 # Read the whole contents of FILE in $slurp
543 { local $/ = undef; $slurp = <FILE>; }
544 Note, however, that this restricts localization of some values ; for
546 example, the following statement dies, as of perl 5.9.0, with an error
547 Modification of a read-only value attempted, because the $1 variable is
548 magical and read-only :
549 local $1 = 2;
551 Similarly, but in a way more difficult to spot, the following snippet
553 will die in perl 5.9.0 :
554 sub f { local $_ = "foo"; print }
556 for ($1) {
557 # now $_ is aliased to $1, thus is magic and readonly
558 f();
559 }
560 See next section for an alternative to this situation.
562 WARNING: Localization of tied arrays and hashes does not currently work
564 as described. This will be fixed in a future release of Perl; in the
565 meantime, avoid code that relies on any particular behaviour of local‐
566 ising tied arrays or hashes (localising individual elements is still
567 okay). See "Localising Tied Arrays and Hashes Is Broken" in
568 perl58delta for more details.
569 Localization of globs
571 The construct
573 local *name;
575 creates a whole new symbol table entry for the glob "name" in the cur‐
577 rent package. That means that all variables in its glob slot ($name,
578 @name, %name, &name, and the "name" filehandle) are dynamically reset.
579 This implies, among other things, that any magic eventually carried by
581 those variables is locally lost. In other words, saying "local */"
582 will not have any effect on the internal value of the input record sep‐
583 arator.
584 Notably, if you want to work with a brand new value of the default
586 scalar $_, and avoid the potential problem listed above about $_ previ‐
587 ously carrying a magic value, you should use "local *_" instead of
588 "local $_".
589 Localization of elements of composite types
591 It's also worth taking a moment to explain what happens when you
593 "local"ize a member of a composite type (i.e. an array or hash ele‐
594 ment). In this case, the element is "local"ized by name. This means
595 that when the scope of the "local()" ends, the saved value will be
596 restored to the hash element whose key was named in the "local()", or
597 the array element whose index was named in the "local()". If that ele‐
598 ment was deleted while the "local()" was in effect (e.g. by a
599 "delete()" from a hash or a "shift()" of an array), it will spring back
600 into existence, possibly extending an array and filling in the skipped
601 elements with "undef". For instance, if you say
602 %hash = ( 'This' => 'is', 'a' => 'test' );
604 @ary = ( 0..5 );
605 {
606 local($ary[5]) = 6;
607 local($hash{'a'}) = 'drill';
608 while (my $e = pop(@ary)) {
609 print "$e . . .\n";
610 last unless $e > 3;
611 }
612 if (@ary) {
613 $hash{'only a'} = 'test';
614 delete $hash{'a'};
615 }
616 }
617 print join(' ', map { "$_ $hash{$_}" } sort keys %hash),".\n";
618 print "The array has ",scalar(@ary)," elements: ",
619 join(', ', map { defined $_ ? $_ : 'undef' } @ary),"\n";
620 Perl will print
622 6 . . .
624 4 . . .
625 3 . . .
626 This is a test only a test.
627 The array has 6 elements: 0, 1, 2, undef, undef, 5
628 The behavior of local() on non-existent members of composite types is
630 subject to change in future.
631 Lvalue subroutines
633 WARNING: Lvalue subroutines are still experimental and the implementa‐
635 tion may change in future versions of Perl.
636 It is possible to return a modifiable value from a subroutine. To do
638 this, you have to declare the subroutine to return an lvalue.
639 my $val;
641 sub canmod : lvalue {
642 # return $val; this doesn't work, don't say "return"
643 $val;
644 }
645 sub nomod {
646 $val;
647 }
648 canmod() = 5; # assigns to $val
650 nomod() = 5; # ERROR
651 The scalar/list context for the subroutine and for the right-hand side
653 of assignment is determined as if the subroutine call is replaced by a
654 scalar. For example, consider:
655 data(2,3) = get_data(3,4);
657 Both subroutines here are called in a scalar context, while in:
659 (data(2,3)) = get_data(3,4);
661 and in:
663 (data(2),data(3)) = get_data(3,4);
665 all the subroutines are called in a list context.
667 Lvalue subroutines are EXPERIMENTAL
669 They appear to be convenient, but there are several reasons to be
670 circumspect.
671 You can't use the return keyword, you must pass out the value
673 before falling out of subroutine scope. (see comment in example
674 above). This is usually not a problem, but it disallows an
675 explicit return out of a deeply nested loop, which is sometimes a
676 nice way out.
677 They violate encapsulation. A normal mutator can check the sup‐
679 plied argument before setting the attribute it is protecting, an
680 lvalue subroutine never gets that chance. Consider;
681 my $some_array_ref = []; # protected by mutators ??
683 sub set_arr { # normal mutator
685 my $val = shift;
686 die("expected array, you supplied ", ref $val)
687 unless ref $val eq 'ARRAY';
688 $some_array_ref = $val;
689 }
690 sub set_arr_lv : lvalue { # lvalue mutator
691 $some_array_ref;
692 }
693 # set_arr_lv cannot stop this !
695 set_arr_lv() = { a => 1 };
696 Passing Symbol Table Entries (typeglobs)
698 WARNING: The mechanism described in this section was originally the
700 only way to simulate pass-by-reference in older versions of Perl.
701 While it still works fine in modern versions, the new reference mecha‐
702 nism is generally easier to work with. See below.
703 Sometimes you don't want to pass the value of an array to a subroutine
705 but rather the name of it, so that the subroutine can modify the global
706 copy of it rather than working with a local copy. In perl you can
707 refer to all objects of a particular name by prefixing the name with a
708 star: *foo. This is often known as a "typeglob", because the star on
709 the front can be thought of as a wildcard match for all the funny pre‐
710 fix characters on variables and subroutines and such.
711 When evaluated, the typeglob produces a scalar value that represents
713 all the objects of that name, including any filehandle, format, or sub‐
714 routine. When assigned to, it causes the name mentioned to refer to
715 whatever "*" value was assigned to it. Example:
716 sub doubleary {
718 local(*someary) = @_;
719 foreach $elem (@someary) {
720 $elem *= 2;
721 }
722 }
723 doubleary(*foo);
724 doubleary(*bar);
725 Scalars are already passed by reference, so you can modify scalar argu‐
727 ments without using this mechanism by referring explicitly to $_[0]
728 etc. You can modify all the elements of an array by passing all the
729 elements as scalars, but you have to use the "*" mechanism (or the
730 equivalent reference mechanism) to "push", "pop", or change the size of
731 an array. It will certainly be faster to pass the typeglob (or refer‐
732 ence).
733 Even if you don't want to modify an array, this mechanism is useful for
735 passing multiple arrays in a single LIST, because normally the LIST
736 mechanism will merge all the array values so that you can't extract out
737 the individual arrays. For more on typeglobs, see "Typeglobs and File‐
738 handles" in perldata.
739 When to Still Use local()
741 Despite the existence of "my", there are still three places where the
743 "local" operator still shines. In fact, in these three places, you
744 must use "local" instead of "my".
745 1. You need to give a global variable a temporary value, especially
747 $_.
748 The global variables, like @ARGV or the punctuation variables, must
750 be "local"ized with "local()". This block reads in /etc/motd, and
751 splits it up into chunks separated by lines of equal signs, which
752 are placed in @Fields.
753 {
755 local @ARGV = ("/etc/motd");
756 local $/ = undef;
757 local $_ = <>;
758 @Fields = split /^\s*=+\s*$/;
759 }
760 It particular, it's important to "local"ize $_ in any routine that
762 assigns to it. Look out for implicit assignments in "while" condi‐
763 tionals.
764 2. You need to create a local file or directory handle or a local
766 function.
767 A function that needs a filehandle of its own must use "local()" on
769 a complete typeglob. This can be used to create new symbol table
770 entries:
771 sub ioqueue {
773 local (*READER, *WRITER); # not my!
774 pipe (READER, WRITER) or die "pipe: $!";
775 return (*READER, *WRITER);
776 }
777 ($head, $tail) = ioqueue();
778 See the Symbol module for a way to create anonymous symbol table
780 entries.
781 Because assignment of a reference to a typeglob creates an alias,
783 this can be used to create what is effectively a local function, or
784 at least, a local alias.
785 {
787 local *grow = \&shrink; # only until this block exists
788 grow(); # really calls shrink()
789 move(); # if move() grow()s, it shrink()s too
790 }
791 grow(); # get the real grow() again
792 See "Function Templates" in perlref for more about manipulating
794 functions by name in this way.
795 3. You want to temporarily change just one element of an array or
797 hash.
798 You can "local"ize just one element of an aggregate. Usually this
800 is done on dynamics:
801 {
803 local $SIG{INT} = 'IGNORE';
804 funct(); # uninterruptible
805 }
806 # interruptibility automatically restored here
807 But it also works on lexically declared aggregates. Prior to
809 5.005, this operation could on occasion misbehave.
810 Pass by Reference
812 If you want to pass more than one array or hash into a function--or
814 return them from it--and have them maintain their integrity, then
815 you're going to have to use an explicit pass-by-reference. Before you
816 do that, you need to understand references as detailed in perlref.
817 This section may not make much sense to you otherwise.
818 Here are a few simple examples. First, let's pass in several arrays to
820 a function and have it "pop" all of then, returning a new list of all
821 their former last elements:
822 @tailings = popmany ( \@a, \@b, \@c, \@d );
824 sub popmany {
826 my $aref;
827 my @retlist = ();
828 foreach $aref ( @_ ) {
829 push @retlist, pop @$aref;
830 }
831 return @retlist;
832 }
833 Here's how you might write a function that returns a list of keys
835 occurring in all the hashes passed to it:
836 @common = inter( \%foo, \%bar, \%joe );
838 sub inter {
839 my ($k, $href, %seen); # locals
840 foreach $href (@_) {
841 while ( $k = each %$href ) {
842 $seen{$k}++;
843 }
844 }
845 return grep { $seen{$_} == @_ } keys %seen;
846 }
847 So far, we're using just the normal list return mechanism. What hap‐
849 pens if you want to pass or return a hash? Well, if you're using only
850 one of them, or you don't mind them concatenating, then the normal
851 calling convention is ok, although a little expensive.
852 Where people get into trouble is here:
854 (@a, @b) = func(@c, @d);
856 or
857 (%a, %b) = func(%c, %d);
858 That syntax simply won't work. It sets just @a or %a and clears the @b
860 or %b. Plus the function didn't get passed into two separate arrays or
861 hashes: it got one long list in @_, as always.
862 If you can arrange for everyone to deal with this through references,
864 it's cleaner code, although not so nice to look at. Here's a function
865 that takes two array references as arguments, returning the two array
866 elements in order of how many elements they have in them:
867 ($aref, $bref) = func(\@c, \@d);
869 print "@$aref has more than @$bref\n";
870 sub func {
871 my ($cref, $dref) = @_;
872 if (@$cref > @$dref) {
873 return ($cref, $dref);
874 } else {
875 return ($dref, $cref);
876 }
877 }
878 It turns out that you can actually do this also:
880 (*a, *b) = func(\@c, \@d);
882 print "@a has more than @b\n";
883 sub func {
884 local (*c, *d) = @_;
885 if (@c > @d) {
886 return (\@c, \@d);
887 } else {
888 return (\@d, \@c);
889 }
890 }
891 Here we're using the typeglobs to do symbol table aliasing. It's a tad
893 subtle, though, and also won't work if you're using "my" variables,
894 because only globals (even in disguise as "local"s) are in the symbol
895 table.
896 If you're passing around filehandles, you could usually just use the
898 bare typeglob, like *STDOUT, but typeglobs references work, too. For
899 example:
900 splutter(\*STDOUT);
902 sub splutter {
903 my $fh = shift;
904 print $fh "her um well a hmmm\n";
905 }
906 $rec = get_rec(\*STDIN);
908 sub get_rec {
909 my $fh = shift;
910 return scalar <$fh>;
911 }
912 If you're planning on generating new filehandles, you could do this.
914 Notice to pass back just the bare *FH, not its reference.
915 sub openit {
917 my $path = shift;
918 local *FH;
919 return open (FH, $path) ? *FH : undef;
920 }
921 Prototypes
923 Perl supports a very limited kind of compile-time argument checking
925 using function prototyping. If you declare
926 sub mypush (\@@)
928 then "mypush()" takes arguments exactly like "push()" does. The func‐
930 tion declaration must be visible at compile time. The prototype
931 affects only interpretation of new-style calls to the function, where
932 new-style is defined as not using the "&" character. In other words,
933 if you call it like a built-in function, then it behaves like a built-
934 in function. If you call it like an old-fashioned subroutine, then it
935 behaves like an old-fashioned subroutine. It naturally falls out from
936 this rule that prototypes have no influence on subroutine references
937 like "\&foo" or on indirect subroutine calls like "&{$subref}" or
938 "$subref->()".
939 Method calls are not influenced by prototypes either, because the func‐
941 tion to be called is indeterminate at compile time, since the exact
942 code called depends on inheritance.
943 Because the intent of this feature is primarily to let you define sub‐
945 routines that work like built-in functions, here are prototypes for
946 some other functions that parse almost exactly like the corresponding
947 built-in.
948 Declared as Called as
950 sub mylink ($$) mylink $old, $new
952 sub myvec ($$$) myvec $var, $offset, 1
953 sub myindex ($$;$) myindex &getstring, "substr"
954 sub mysyswrite ($$$;$) mysyswrite $buf, 0, length($buf) - $off, $off
955 sub myreverse (@) myreverse $a, $b, $c
956 sub myjoin ($@) myjoin ":", $a, $b, $c
957 sub mypop (\@) mypop @array
958 sub mysplice (\@$$@) mysplice @array, @array, 0, @pushme
959 sub mykeys (\%) mykeys %{$hashref}
960 sub myopen (*;$) myopen HANDLE, $name
961 sub mypipe (**) mypipe READHANDLE, WRITEHANDLE
962 sub mygrep (&@) mygrep { /foo/ } $a, $b, $c
963 sub myrand ($) myrand 42
964 sub mytime () mytime
965 Any backslashed prototype character represents an actual argument that
967 absolutely must start with that character. The value passed as part of
968 @_ will be a reference to the actual argument given in the subroutine
969 call, obtained by applying "\" to that argument.
970 You can also backslash several argument types simultaneously by using
972 the "\[]" notation:
973 sub myref (\[$@%&*])
975 will allow calling myref() as
977 myref $var
979 myref @array
980 myref %hash
981 myref &sub
982 myref *glob
983 and the first argument of myref() will be a reference to a scalar, an
985 array, a hash, a code, or a glob.
986 Unbackslashed prototype characters have special meanings. Any unback‐
988 slashed "@" or "%" eats all remaining arguments, and forces list con‐
989 text. An argument represented by "$" forces scalar context. An "&"
990 requires an anonymous subroutine, which, if passed as the first argu‐
991 ment, does not require the "sub" keyword or a subsequent comma.
992 A "*" allows the subroutine to accept a bareword, constant, scalar
994 expression, typeglob, or a reference to a typeglob in that slot. The
995 value will be available to the subroutine either as a simple scalar, or
996 (in the latter two cases) as a reference to the typeglob. If you wish
997 to always convert such arguments to a typeglob reference, use Sym‐
998 bol::qualify_to_ref() as follows:
999 use Symbol 'qualify_to_ref';
1001 sub foo (*) {
1003 my $fh = qualify_to_ref(shift, caller);
1004 ...
1005 }
1006 A semicolon separates mandatory arguments from optional arguments. It
1008 is redundant before "@" or "%", which gobble up everything else.
1009 Note how the last three examples in the table above are treated spe‐
1011 cially by the parser. "mygrep()" is parsed as a true list operator,
1012 "myrand()" is parsed as a true unary operator with unary precedence the
1013 same as "rand()", and "mytime()" is truly without arguments, just like
1014 "time()". That is, if you say
1015 mytime +2;
1017 you'll get "mytime() + 2", not mytime(2), which is how it would be
1019 parsed without a prototype.
1020 The interesting thing about "&" is that you can generate new syntax
1022 with it, provided it's in the initial position:
1023 sub try (&@) {
1025 my($try,$catch) = @_;
1026 eval { &$try };
1027 if ($@) {
1028 local $_ = $@;
1029 &$catch;
1030 }
1031 }
1032 sub catch (&) { $_[0] }
1033 try {
1035 die "phooey";
1036 } catch {
1037 /phooey/ and print "unphooey\n";
1038 };
1039 That prints "unphooey". (Yes, there are still unresolved issues having
1041 to do with visibility of @_. I'm ignoring that question for the
1042 moment. (But note that if we make @_ lexically scoped, those anonymous
1043 subroutines can act like closures... (Gee, is this sounding a little
1044 Lispish? (Never mind.))))
1045 And here's a reimplementation of the Perl "grep" operator:
1047 sub mygrep (&@) {
1049 my $code = shift;
1050 my @result;
1051 foreach $_ (@_) {
1052 push(@result, $_) if &$code;
1053 }
1054 @result;
1055 }
1056 Some folks would prefer full alphanumeric prototypes. Alphanumerics
1058 have been intentionally left out of prototypes for the express purpose
1059 of someday in the future adding named, formal parameters. The current
1060 mechanism's main goal is to let module writers provide better diagnos‐
1061 tics for module users. Larry feels the notation quite understandable
1062 to Perl programmers, and that it will not intrude greatly upon the meat
1063 of the module, nor make it harder to read. The line noise is visually
1064 encapsulated into a small pill that's easy to swallow.
1065 If you try to use an alphanumeric sequence in a prototype you will gen‐
1067 erate an optional warning - "Illegal character in prototype...".
1068 Unfortunately earlier versions of Perl allowed the prototype to be used
1069 as long as its prefix was a valid prototype. The warning may be
1070 upgraded to a fatal error in a future version of Perl once the majority
1071 of offending code is fixed.
1072 It's probably best to prototype new functions, not retrofit prototyping
1074 into older ones. That's because you must be especially careful about
1075 silent impositions of differing list versus scalar contexts. For exam‐
1076 ple, if you decide that a function should take just one parameter, like
1077 this:
1078 sub func ($) {
1080 my $n = shift;
1081 print "you gave me $n\n";
1082 }
1083 and someone has been calling it with an array or expression returning a
1085 list:
1086 func(@foo);
1088 func( split /:/ );
1089 Then you've just supplied an automatic "scalar" in front of their argu‐
1091 ment, which can be more than a bit surprising. The old @foo which used
1092 to hold one thing doesn't get passed in. Instead, "func()" now gets
1093 passed in a 1; that is, the number of elements in @foo. And the
1094 "split" gets called in scalar context so it starts scribbling on your
1095 @_ parameter list. Ouch!
1096 This is all very powerful, of course, and should be used only in moder‐
1098 ation to make the world a better place.
1099 Constant Functions
1101 Functions with a prototype of "()" are potential candidates for inlin‐
1103 ing. If the result after optimization and constant folding is either a
1104 constant or a lexically-scoped scalar which has no other references,
1105 then it will be used in place of function calls made without "&".
1106 Calls made using "&" are never inlined. (See constant.pm for an easy
1107 way to declare most constants.)
1108 The following functions would all be inlined:
1110 sub pi () { 3.14159 } # Not exact, but close.
1112 sub PI () { 4 * atan2 1, 1 } # As good as it gets,
1113 # and it's inlined, too!
1114 sub ST_DEV () { 0 }
1115 sub ST_INO () { 1 }
1116 sub FLAG_FOO () { 1 << 8 }
1118 sub FLAG_BAR () { 1 << 9 }
1119 sub FLAG_MASK () { FLAG_FOO ⎪ FLAG_BAR }
1120 sub OPT_BAZ () { not (0x1B58 & FLAG_MASK) }
1122 sub N () { int(OPT_BAZ) / 3 }
1124 sub FOO_SET () { 1 if FLAG_MASK & FLAG_FOO }
1126 Be aware that these will not be inlined; as they contain inner scopes,
1128 the constant folding doesn't reduce them to a single constant:
1129 sub foo_set () { if (FLAG_MASK & FLAG_FOO) { 1 } }
1131 sub baz_val () {
1133 if (OPT_BAZ) {
1134 return 23;
1135 }
1136 else {
1137 return 42;
1138 }
1139 }
1140 If you redefine a subroutine that was eligible for inlining, you'll get
1142 a mandatory warning. (You can use this warning to tell whether or not
1143 a particular subroutine is considered constant.) The warning is con‐
1144 sidered severe enough not to be optional because previously compiled
1145 invocations of the function will still be using the old value of the
1146 function. If you need to be able to redefine the subroutine, you need
1147 to ensure that it isn't inlined, either by dropping the "()" prototype
1148 (which changes calling semantics, so beware) or by thwarting the inlin‐
1149 ing mechanism in some other way, such as
1150 sub not_inlined () {
1152 23 if $];
1153 }
1154 Overriding Built-in Functions
1156 Many built-in functions may be overridden, though this should be tried
1158 only occasionally and for good reason. Typically this might be done by
1159 a package attempting to emulate missing built-in functionality on a
1160 non-Unix system.
1161 Overriding may be done only by importing the name from a module at com‐
1163 pile time--ordinary predeclaration isn't good enough. However, the
1164 "use subs" pragma lets you, in effect, predeclare subs via the import
1165 syntax, and these names may then override built-in ones:
1166 use subs 'chdir', 'chroot', 'chmod', 'chown';
1168 chdir $somewhere;
1169 sub chdir { ... }
1170 To unambiguously refer to the built-in form, precede the built-in name
1172 with the special package qualifier "CORE::". For example, saying
1173 "CORE::open()" always refers to the built-in "open()", even if the cur‐
1174 rent package has imported some other subroutine called "&open()" from
1175 elsewhere. Even though it looks like a regular function call, it
1176 isn't: you can't take a reference to it, such as the incorrect
1177 "\&CORE::open" might appear to produce.
1178 Library modules should not in general export built-in names like "open"
1180 or "chdir" as part of their default @EXPORT list, because these may
1181 sneak into someone else's namespace and change the semantics unexpect‐
1182 edly. Instead, if the module adds that name to @EXPORT_OK, then it's
1183 possible for a user to import the name explicitly, but not implicitly.
1184 That is, they could say
1185 use Module 'open';
1187 and it would import the "open" override. But if they said
1189 use Module;
1191 they would get the default imports without overrides.
1193 The foregoing mechanism for overriding built-in is restricted, quite
1195 deliberately, to the package that requests the import. There is a sec‐
1196 ond method that is sometimes applicable when you wish to override a
1197 built-in everywhere, without regard to namespace boundaries. This is
1198 achieved by importing a sub into the special namespace
1199 "CORE::GLOBAL::". Here is an example that quite brazenly replaces the
1200 "glob" operator with something that understands regular expressions.
1201 package REGlob;
1203 require Exporter;
1204 @ISA = 'Exporter';
1205 @EXPORT_OK = 'glob';
1206 sub import {
1208 my $pkg = shift;
1209 return unless @_;
1210 my $sym = shift;
1211 my $where = ($sym =~ s/^GLOBAL_// ? 'CORE::GLOBAL' : caller(0));
1212 $pkg->export($where, $sym, @_);
1213 }
1214 sub glob {
1216 my $pat = shift;
1217 my @got;
1218 local *D;
1219 if (opendir D, '.') {
1220 @got = grep /$pat/, readdir D;
1221 closedir D;
1222 }
1223 return @got;
1224 }
1225 1;
1226 And here's how it could be (ab)used:
1228 #use REGlob 'GLOBAL_glob'; # override glob() in ALL namespaces
1230 package Foo;
1231 use REGlob 'glob'; # override glob() in Foo:: only
1232 print for <^[a-z_]+\.pm\$>; # show all pragmatic modules
1233 The initial comment shows a contrived, even dangerous example. By
1235 overriding "glob" globally, you would be forcing the new (and subver‐
1236 sive) behavior for the "glob" operator for every namespace, without the
1237 complete cognizance or cooperation of the modules that own those names‐
1238 paces. Naturally, this should be done with extreme caution--if it must
1239 be done at all.
1240 The "REGlob" example above does not implement all the support needed to
1242 cleanly override perl's "glob" operator. The built-in "glob" has dif‐
1243 ferent behaviors depending on whether it appears in a scalar or list
1244 context, but our "REGlob" doesn't. Indeed, many perl built-in have
1245 such context sensitive behaviors, and these must be adequately sup‐
1246 ported by a properly written override. For a fully functional example
1247 of overriding "glob", study the implementation of "File::DosGlob" in
1248 the standard library.
1249 When you override a built-in, your replacement should be consistent (if
1251 possible) with the built-in native syntax. You can achieve this by
1252 using a suitable prototype. To get the prototype of an overridable
1253 built-in, use the "prototype" function with an argument of
1254 "CORE::builtin_name" (see "prototype" in perlfunc).
1255 Note however that some built-ins can't have their syntax expressed by a
1257 prototype (such as "system" or "chomp"). If you override them you
1258 won't be able to fully mimic their original syntax.
1259 The built-ins "do", "require" and "glob" can also be overridden, but
1261 due to special magic, their original syntax is preserved, and you don't
1262 have to define a prototype for their replacements. (You can't override
1263 the "do BLOCK" syntax, though).
1264 "require" has special additional dark magic: if you invoke your
1266 "require" replacement as "require Foo::Bar", it will actually receive
1267 the argument "Foo/Bar.pm" in @_. See "require" in perlfunc.
1268 And, as you'll have noticed from the previous example, if you override
1270 "glob", the "<*>" glob operator is overridden as well.
1271 In a similar fashion, overriding the "readline" function also overrides
1273 the equivalent I/O operator "<FILEHANDLE>".
1274 Finally, some built-ins (e.g. "exists" or "grep") can't be overridden.
1276 Autoloading
1278 If you call a subroutine that is undefined, you would ordinarily get an
1280 immediate, fatal error complaining that the subroutine doesn't exist.
1281 (Likewise for subroutines being used as methods, when the method
1282 doesn't exist in any base class of the class's package.) However, if
1283 an "AUTOLOAD" subroutine is defined in the package or packages used to
1284 locate the original subroutine, then that "AUTOLOAD" subroutine is
1285 called with the arguments that would have been passed to the original
1286 subroutine. The fully qualified name of the original subroutine magi‐
1287 cally appears in the global $AUTOLOAD variable of the same package as
1288 the "AUTOLOAD" routine. The name is not passed as an ordinary argument
1289 because, er, well, just because, that's why...
1290 Many "AUTOLOAD" routines load in a definition for the requested subrou‐
1292 tine using eval(), then execute that subroutine using a special form of
1293 goto() that erases the stack frame of the "AUTOLOAD" routine without a
1294 trace. (See the source to the standard module documented in
1295 AutoLoader, for example.) But an "AUTOLOAD" routine can also just emu‐
1296 late the routine and never define it. For example, let's pretend that
1297 a function that wasn't defined should just invoke "system" with those
1298 arguments. All you'd do is:
1299 sub AUTOLOAD {
1301 my $program = $AUTOLOAD;
1302 $program =~ s/.*:://;
1303 system($program, @_);
1304 }
1305 date();
1306 who('am', 'i');
1307 ls('-l');
1308 In fact, if you predeclare functions you want to call that way, you
1310 don't even need parentheses:
1311 use subs qw(date who ls);
1313 date;
1314 who "am", "i";
1315 ls -l;
1316 A more complete example of this is the standard Shell module, which can
1318 treat undefined subroutine calls as calls to external programs.
1319 Mechanisms are available to help modules writers split their modules
1321 into autoloadable files. See the standard AutoLoader module described
1322 in AutoLoader and in AutoSplit, the standard SelfLoader modules in
1323 SelfLoader, and the document on adding C functions to Perl code in per‐
1324 lxs.
1325 Subroutine Attributes
1327 A subroutine declaration or definition may have a list of attributes
1329 associated with it. If such an attribute list is present, it is broken
1330 up at space or colon boundaries and treated as though a "use
1331 attributes" had been seen. See attributes for details about what
1332 attributes are currently supported. Unlike the limitation with the
1333 obsolescent "use attrs", the "sub : ATTRLIST" syntax works to associate
1334 the attributes with a pre-declaration, and not just with a subroutine
1335 definition.
1336 The attributes must be valid as simple identifier names (without any
1338 punctuation other than the '_' character). They may have a parameter
1339 list appended, which is only checked for whether its parentheses
1340 ('(',')') nest properly.
1341 Examples of valid syntax (even though the attributes are unknown):
1343 sub fnord (&\%) : switch(10,foo(7,3)) : expensive;
1345 sub plugh () : Ugly('\(") :Bad;
1346 sub xyzzy : _5x5 { ... }
1347 Examples of invalid syntax:
1349 sub fnord : switch(10,foo(); # ()-string not balanced
1351 sub snoid : Ugly('('); # ()-string not balanced
1352 sub xyzzy : 5x5; # "5x5" not a valid identifier
1353 sub plugh : Y2::north; # "Y2::north" not a simple identifier
1354 sub snurt : foo + bar; # "+" not a colon or space
1355 The attribute list is passed as a list of constant strings to the code
1357 which associates them with the subroutine. In particular, the second
1358 example of valid syntax above currently looks like this in terms of how
1359 it's parsed and invoked:
1360 use attributes __PACKAGE__, \&plugh, q[Ugly('\(")], 'Bad';
1362 For further details on attribute lists and their manipulation, see
1364 attributes and Attribute::Handlers.
1365
1367 See "Function Templates" in perlref for more about references and clo‐
1368 sures. See perlxs if you'd like to learn about calling C subroutines
1369 from Perl. See perlembed if you'd like to learn about calling Perl
1370 subroutines from C. See perlmod to learn about bundling up your func‐
1371 tions in separate files. See perlmodlib to learn what library modules
1372 come standard on your system. See perltoot to learn how to make object
1373 method calls.
1374perl v5.8.8 2006-01-07 PERLSUB(1)