This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Update Pod-Simple to CPAN version 3.17
[perl5.git] / pod / perlref.pod
CommitLineData
a0d0e21e 1=head1 NAME
d74e8afc 2X<reference> X<pointer> X<data structure> X<structure> X<struct>
a0d0e21e
LW
3
4perlref - Perl references and nested data structures
5
a1e2a320
GS
6=head1 NOTE
7
8This is complete documentation about all aspects of references.
9For a shorter, tutorial introduction to just the essential features,
10see L<perlreftut>.
11
a0d0e21e
LW
12=head1 DESCRIPTION
13
cb1a09d0 14Before release 5 of Perl it was difficult to represent complex data
5a964f20
TC
15structures, because all references had to be symbolic--and even then
16it was difficult to refer to a variable instead of a symbol table entry.
17Perl now not only makes it easier to use symbolic references to variables,
18but also lets you have "hard" references to any piece of data or code.
19Any scalar may hold a hard reference. Because arrays and hashes contain
20scalars, you can now easily build arrays of arrays, arrays of hashes,
21hashes of arrays, arrays of hashes of functions, and so on.
a0d0e21e
LW
22
23Hard references are smart--they keep track of reference counts for you,
2d24ed35 24automatically freeing the thing referred to when its reference count goes
7c2ea1c7 25to zero. (Reference counts for values in self-referential or
2d24ed35 26cyclic data structures may not go to zero without a little help; see
7b8d334a 27L<perlobj/"Two-Phased Garbage Collection"> for a detailed explanation.)
2d24ed35
CS
28If that thing happens to be an object, the object is destructed. See
29L<perlobj> for more about objects. (In a sense, everything in Perl is an
30object, but we usually reserve the word for references to objects that
31have been officially "blessed" into a class package.)
32
33Symbolic references are names of variables or other objects, just as a
54310121 34symbolic link in a Unix filesystem contains merely the name of a file.
d1be9408 35The C<*glob> notation is something of a symbolic reference. (Symbolic
2d24ed35
CS
36references are sometimes called "soft references", but please don't call
37them that; references are confusing enough without useless synonyms.)
d74e8afc
ITB
38X<reference, symbolic> X<reference, soft>
39X<symbolic reference> X<soft reference>
2d24ed35 40
54310121 41In contrast, hard references are more like hard links in a Unix file
2d24ed35
CS
42system: They are used to access an underlying object without concern for
43what its (other) name is. When the word "reference" is used without an
5a964f20 44adjective, as in the following paragraph, it is usually talking about a
2d24ed35 45hard reference.
d74e8afc 46X<reference, hard> X<hard reference>
2d24ed35
CS
47
48References are easy to use in Perl. There is just one overriding
49principle: Perl does no implicit referencing or dereferencing. When a
50scalar is holding a reference, it always behaves as a simple scalar. It
51doesn't magically start being an array or hash or subroutine; you have to
52tell it explicitly to do so, by dereferencing it.
a0d0e21e 53
5a964f20 54=head2 Making References
d74e8afc 55X<reference, creation> X<referencing>
5a964f20
TC
56
57References can be created in several ways.
a0d0e21e
LW
58
59=over 4
60
61=item 1.
d74e8afc 62X<\> X<backslash>
a0d0e21e
LW
63
64By using the backslash operator on a variable, subroutine, or value.
7c2ea1c7
GS
65(This works much like the & (address-of) operator in C.)
66This typically creates I<another> reference to a variable, because
a0d0e21e
LW
67there's already a reference to the variable in the symbol table. But
68the symbol table reference might go away, and you'll still have the
69reference that the backslash returned. Here are some examples:
70
71 $scalarref = \$foo;
72 $arrayref = \@ARGV;
73 $hashref = \%ENV;
74 $coderef = \&handler;
55497cff 75 $globref = \*foo;
cb1a09d0 76
5a964f20
TC
77It isn't possible to create a true reference to an IO handle (filehandle
78or dirhandle) using the backslash operator. The most you can get is a
79reference to a typeglob, which is actually a complete symbol table entry.
80But see the explanation of the C<*foo{THING}> syntax below. However,
81you can still use type globs and globrefs as though they were IO handles.
a0d0e21e
LW
82
83=item 2.
d74e8afc
ITB
84X<array, anonymous> X<[> X<[]> X<square bracket>
85X<bracket, square> X<arrayref> X<array reference> X<reference, array>
a0d0e21e 86
5a964f20 87A reference to an anonymous array can be created using square
a0d0e21e
LW
88brackets:
89
90 $arrayref = [1, 2, ['a', 'b', 'c']];
91
5a964f20 92Here we've created a reference to an anonymous array of three elements
54310121 93whose final element is itself a reference to another anonymous array of three
a0d0e21e 94elements. (The multidimensional syntax described later can be used to
c47ff5f1 95access this. For example, after the above, C<< $arrayref->[2][1] >> would have
a0d0e21e
LW
96the value "b".)
97
7c2ea1c7 98Taking a reference to an enumerated list is not the same
cb1a09d0
AD
99as using square brackets--instead it's the same as creating
100a list of references!
101
54310121 102 @list = (\$a, \@b, \%c);
58e0a6ae
GS
103 @list = \($a, @b, %c); # same thing!
104
54310121 105As a special case, C<\(@foo)> returns a list of references to the contents
b6429b1b
GS
106of C<@foo>, not a reference to C<@foo> itself. Likewise for C<%foo>,
107except that the key references are to copies (since the keys are just
108strings rather than full-fledged scalars).
cb1a09d0 109
a0d0e21e 110=item 3.
d74e8afc
ITB
111X<hash, anonymous> X<{> X<{}> X<curly bracket>
112X<bracket, curly> X<brace> X<hashref> X<hash reference> X<reference, hash>
a0d0e21e 113
5a964f20 114A reference to an anonymous hash can be created using curly
a0d0e21e
LW
115brackets:
116
117 $hashref = {
118 'Adam' => 'Eve',
119 'Clyde' => 'Bonnie',
120 };
121
5a964f20 122Anonymous hash and array composers like these can be intermixed freely to
a0d0e21e
LW
123produce as complicated a structure as you want. The multidimensional
124syntax described below works for these too. The values above are
125literals, but variables and expressions would work just as well, because
126assignment operators in Perl (even within local() or my()) are executable
127statements, not compile-time declarations.
128
129Because curly brackets (braces) are used for several other things
130including BLOCKs, you may occasionally have to disambiguate braces at the
131beginning of a statement by putting a C<+> or a C<return> in front so
132that Perl realizes the opening brace isn't starting a BLOCK. The economy and
133mnemonic value of using curlies is deemed worth this occasional extra
134hassle.
135
136For example, if you wanted a function to make a new hash and return a
137reference to it, you have these options:
138
139 sub hashem { { @_ } } # silently wrong
140 sub hashem { +{ @_ } } # ok
141 sub hashem { return { @_ } } # ok
142
ebc58f1a
GS
143On the other hand, if you want the other meaning, you can do this:
144
145 sub showem { { @_ } } # ambiguous (currently ok, but may change)
146 sub showem { {; @_ } } # ok
147 sub showem { { return @_ } } # ok
148
7c2ea1c7 149The leading C<+{> and C<{;> always serve to disambiguate
ebc58f1a
GS
150the expression to mean either the HASH reference, or the BLOCK.
151
a0d0e21e 152=item 4.
d74e8afc
ITB
153X<subroutine, anonymous> X<subroutine, reference> X<reference, subroutine>
154X<scope, lexical> X<closure> X<lexical> X<lexical scope>
a0d0e21e 155
5a964f20 156A reference to an anonymous subroutine can be created by using
a0d0e21e
LW
157C<sub> without a subname:
158
159 $coderef = sub { print "Boink!\n" };
160
7c2ea1c7
GS
161Note the semicolon. Except for the code
162inside not being immediately executed, a C<sub {}> is not so much a
a0d0e21e 163declaration as it is an operator, like C<do{}> or C<eval{}>. (However, no
5a964f20 164matter how many times you execute that particular line (unless you're in an
19799a22 165C<eval("...")>), $coderef will still have a reference to the I<same>
a0d0e21e
LW
166anonymous subroutine.)
167
748a9306 168Anonymous subroutines act as closures with respect to my() variables,
7c2ea1c7 169that is, variables lexically visible within the current scope. Closure
748a9306
LW
170is a notion out of the Lisp world that says if you define an anonymous
171function in a particular lexical context, it pretends to run in that
7c2ea1c7 172context even when it's called outside the context.
748a9306
LW
173
174In human terms, it's a funny way of passing arguments to a subroutine when
175you define it as well as when you call it. It's useful for setting up
176little bits of code to run later, such as callbacks. You can even
54310121
PP
177do object-oriented stuff with it, though Perl already provides a different
178mechanism to do that--see L<perlobj>.
748a9306 179
7c2ea1c7
GS
180You might also think of closure as a way to write a subroutine
181template without using eval(). Here's a small example of how
182closures work:
748a9306
LW
183
184 sub newprint {
185 my $x = shift;
186 return sub { my $y = shift; print "$x, $y!\n"; };
a0d0e21e 187 }
748a9306
LW
188 $h = newprint("Howdy");
189 $g = newprint("Greetings");
190
191 # Time passes...
192
193 &$h("world");
194 &$g("earthlings");
a0d0e21e 195
748a9306
LW
196This prints
197
198 Howdy, world!
199 Greetings, earthlings!
200
7c2ea1c7
GS
201Note particularly that $x continues to refer to the value passed
202into newprint() I<despite> "my $x" having gone out of scope by the
203time the anonymous subroutine runs. That's what a closure is all
204about.
748a9306 205
5a964f20 206This applies only to lexical variables, by the way. Dynamic variables
748a9306
LW
207continue to work as they have always worked. Closure is not something
208that most Perl programmers need trouble themselves about to begin with.
a0d0e21e
LW
209
210=item 5.
d74e8afc 211X<constructor> X<new>
a0d0e21e 212
63acfd00 213References are often returned by special subroutines called constructors. Perl
214objects are just references to a special type of object that happens to know
215which package it's associated with. Constructors are just special subroutines
216that know how to create that association. They do so by starting with an
217ordinary reference, and it remains an ordinary reference even while it's also
218being an object. Constructors are often named C<new()>. You I<can> call them
219indirectly:
220
221 $objref = new Doggie( Tail => 'short', Ears => 'long' );
222
223But that can produce ambiguous syntax in certain cases, so it's often
224better to use the direct method invocation approach:
5a964f20
TC
225
226 $objref = Doggie->new(Tail => 'short', Ears => 'long');
227
228 use Term::Cap;
229 $terminal = Term::Cap->Tgetent( { OSPEED => 9600 });
230
231 use Tk;
232 $main = MainWindow->new();
233 $menubar = $main->Frame(-relief => "raised",
234 -borderwidth => 2)
235
a0d0e21e 236=item 6.
d74e8afc 237X<autovivification>
a0d0e21e
LW
238
239References of the appropriate type can spring into existence if you
5f05dabc 240dereference them in a context that assumes they exist. Because we haven't
a0d0e21e
LW
241talked about dereferencing yet, we can't show you any examples yet.
242
cb1a09d0 243=item 7.
d74e8afc 244X<*foo{THING}> X<*>
cb1a09d0 245
55497cff
PP
246A reference can be created by using a special syntax, lovingly known as
247the *foo{THING} syntax. *foo{THING} returns a reference to the THING
248slot in *foo (which is the symbol table entry which holds everything
249known as foo).
cb1a09d0 250
55497cff
PP
251 $scalarref = *foo{SCALAR};
252 $arrayref = *ARGV{ARRAY};
253 $hashref = *ENV{HASH};
254 $coderef = *handler{CODE};
36477c24 255 $ioref = *STDIN{IO};
55497cff 256 $globref = *foo{GLOB};
c0bd1adc 257 $formatref = *foo{FORMAT};
55497cff 258
7c2ea1c7
GS
259All of these are self-explanatory except for C<*foo{IO}>. It returns
260the IO handle, used for file handles (L<perlfunc/open>), sockets
261(L<perlfunc/socket> and L<perlfunc/socketpair>), and directory
262handles (L<perlfunc/opendir>). For compatibility with previous
39b99f21 263versions of Perl, C<*foo{FILEHANDLE}> is a synonym for C<*foo{IO}>, though it
264is deprecated as of 5.8.0. If deprecation warnings are in effect, it will warn
265of its use.
55497cff 266
7c2ea1c7
GS
267C<*foo{THING}> returns undef if that particular THING hasn't been used yet,
268except in the case of scalars. C<*foo{SCALAR}> returns a reference to an
5f05dabc
PP
269anonymous scalar if $foo hasn't been used yet. This might change in a
270future release.
271
7c2ea1c7 272C<*foo{IO}> is an alternative to the C<*HANDLE> mechanism given in
5a964f20
TC
273L<perldata/"Typeglobs and Filehandles"> for passing filehandles
274into or out of subroutines, or storing into larger data structures.
275Its disadvantage is that it won't create a new filehandle for you.
7c2ea1c7
GS
276Its advantage is that you have less risk of clobbering more than
277you want to with a typeglob assignment. (It still conflates file
278and directory handles, though.) However, if you assign the incoming
279value to a scalar instead of a typeglob as we do in the examples
280below, there's no risk of that happening.
36477c24 281
7c2ea1c7
GS
282 splutter(*STDOUT); # pass the whole glob
283 splutter(*STDOUT{IO}); # pass both file and dir handles
5a964f20 284
cb1a09d0
AD
285 sub splutter {
286 my $fh = shift;
287 print $fh "her um well a hmmm\n";
288 }
289
7c2ea1c7
GS
290 $rec = get_rec(*STDIN); # pass the whole glob
291 $rec = get_rec(*STDIN{IO}); # pass both file and dir handles
5a964f20 292
cb1a09d0
AD
293 sub get_rec {
294 my $fh = shift;
295 return scalar <$fh>;
296 }
297
a0d0e21e
LW
298=back
299
5a964f20 300=head2 Using References
d74e8afc 301X<reference, use> X<dereferencing> X<dereference>
5a964f20 302
a0d0e21e
LW
303That's it for creating references. By now you're probably dying to
304know how to use references to get back to your long-lost data. There
305are several basic methods.
306
307=over 4
308
309=item 1.
310
6309d9d9
PP
311Anywhere you'd put an identifier (or chain of identifiers) as part
312of a variable or subroutine name, you can replace the identifier with
313a simple scalar variable containing a reference of the correct type:
a0d0e21e
LW
314
315 $bar = $$scalarref;
316 push(@$arrayref, $filename);
317 $$arrayref[0] = "January";
318 $$hashref{"KEY"} = "VALUE";
319 &$coderef(1,2,3);
cb1a09d0 320 print $globref "output\n";
a0d0e21e 321
19799a22 322It's important to understand that we are specifically I<not> dereferencing
a0d0e21e 323C<$arrayref[0]> or C<$hashref{"KEY"}> there. The dereference of the
19799a22 324scalar variable happens I<before> it does any key lookups. Anything more
a0d0e21e
LW
325complicated than a simple scalar variable must use methods 2 or 3 below.
326However, a "simple scalar" includes an identifier that itself uses method
3271 recursively. Therefore, the following prints "howdy".
328
329 $refrefref = \\\"howdy";
330 print $$$$refrefref;
331
332=item 2.
333
6309d9d9
PP
334Anywhere you'd put an identifier (or chain of identifiers) as part of a
335variable or subroutine name, you can replace the identifier with a
336BLOCK returning a reference of the correct type. In other words, the
337previous examples could be written like this:
a0d0e21e
LW
338
339 $bar = ${$scalarref};
340 push(@{$arrayref}, $filename);
341 ${$arrayref}[0] = "January";
342 ${$hashref}{"KEY"} = "VALUE";
343 &{$coderef}(1,2,3);
36477c24 344 $globref->print("output\n"); # iff IO::Handle is loaded
a0d0e21e
LW
345
346Admittedly, it's a little silly to use the curlies in this case, but
347the BLOCK can contain any arbitrary expression, in particular,
348subscripted expressions:
349
54310121 350 &{ $dispatch{$index} }(1,2,3); # call correct routine
a0d0e21e
LW
351
352Because of being able to omit the curlies for the simple case of C<$$x>,
353people often make the mistake of viewing the dereferencing symbols as
354proper operators, and wonder about their precedence. If they were,
5f05dabc 355though, you could use parentheses instead of braces. That's not the case.
a0d0e21e 356Consider the difference below; case 0 is a short-hand version of case 1,
19799a22 357I<not> case 2:
a0d0e21e
LW
358
359 $$hashref{"KEY"} = "VALUE"; # CASE 0
360 ${$hashref}{"KEY"} = "VALUE"; # CASE 1
361 ${$hashref{"KEY"}} = "VALUE"; # CASE 2
362 ${$hashref->{"KEY"}} = "VALUE"; # CASE 3
363
364Case 2 is also deceptive in that you're accessing a variable
365called %hashref, not dereferencing through $hashref to the hash
366it's presumably referencing. That would be case 3.
367
368=item 3.
369
6da72b64
CS
370Subroutine calls and lookups of individual array elements arise often
371enough that it gets cumbersome to use method 2. As a form of
372syntactic sugar, the examples for method 2 may be written:
a0d0e21e 373
6da72b64
CS
374 $arrayref->[0] = "January"; # Array element
375 $hashref->{"KEY"} = "VALUE"; # Hash element
376 $coderef->(1,2,3); # Subroutine call
a0d0e21e 377
6da72b64 378The left side of the arrow can be any expression returning a reference,
19799a22 379including a previous dereference. Note that C<$array[$x]> is I<not> the
c47ff5f1 380same thing as C<< $array->[$x] >> here:
a0d0e21e
LW
381
382 $array[$x]->{"foo"}->[0] = "January";
383
384This is one of the cases we mentioned earlier in which references could
385spring into existence when in an lvalue context. Before this
386statement, C<$array[$x]> may have been undefined. If so, it's
387automatically defined with a hash reference so that we can look up
c47ff5f1 388C<{"foo"}> in it. Likewise C<< $array[$x]->{"foo"} >> will automatically get
a0d0e21e 389defined with an array reference so that we can look up C<[0]> in it.
5a964f20 390This process is called I<autovivification>.
a0d0e21e 391
19799a22 392One more thing here. The arrow is optional I<between> brackets
a0d0e21e
LW
393subscripts, so you can shrink the above down to
394
395 $array[$x]{"foo"}[0] = "January";
396
397Which, in the degenerate case of using only ordinary arrays, gives you
398multidimensional arrays just like C's:
399
400 $score[$x][$y][$z] += 42;
401
402Well, okay, not entirely like C's arrays, actually. C doesn't know how
403to grow its arrays on demand. Perl does.
404
405=item 4.
406
407If a reference happens to be a reference to an object, then there are
408probably methods to access the things referred to, and you should probably
409stick to those methods unless you're in the class package that defines the
410object's methods. In other words, be nice, and don't violate the object's
411encapsulation without a very good reason. Perl does not enforce
412encapsulation. We are not totalitarians here. We do expect some basic
413civility though.
414
415=back
416
7c2ea1c7
GS
417Using a string or number as a reference produces a symbolic reference,
418as explained above. Using a reference as a number produces an
419integer representing its storage location in memory. The only
420useful thing to be done with this is to compare two references
421numerically to see whether they refer to the same location.
d74e8afc 422X<reference, numeric context>
7c2ea1c7
GS
423
424 if ($ref1 == $ref2) { # cheap numeric compare of references
425 print "refs 1 and 2 refer to the same thing\n";
426 }
427
428Using a reference as a string produces both its referent's type,
429including any package blessing as described in L<perlobj>, as well
430as the numeric address expressed in hex. The ref() operator returns
431just the type of thing the reference is pointing to, without the
432address. See L<perlfunc/ref> for details and examples of its use.
d74e8afc 433X<reference, string context>
a0d0e21e 434
5a964f20
TC
435The bless() operator may be used to associate the object a reference
436points to with a package functioning as an object class. See L<perlobj>.
a0d0e21e 437
5f05dabc 438A typeglob may be dereferenced the same way a reference can, because
7c2ea1c7 439the dereference syntax always indicates the type of reference desired.
a0d0e21e
LW
440So C<${*foo}> and C<${\$foo}> both indicate the same scalar variable.
441
442Here's a trick for interpolating a subroutine call into a string:
443
cb1a09d0
AD
444 print "My sub returned @{[mysub(1,2,3)]} that time.\n";
445
446The way it works is that when the C<@{...}> is seen in the double-quoted
447string, it's evaluated as a block. The block creates a reference to an
448anonymous array containing the results of the call to C<mysub(1,2,3)>. So
449the whole block returns a reference to an array, which is then
450dereferenced by C<@{...}> and stuck into the double-quoted string. This
451chicanery is also useful for arbitrary expressions:
a0d0e21e 452
184e9718 453 print "That yields @{[$n + 5]} widgets\n";
a0d0e21e 454
35efdb20
DL
455Similarly, an expression that returns a reference to a scalar can be
456dereferenced via C<${...}>. Thus, the above expression may be written
457as:
458
459 print "That yields ${\($n + 5)} widgets\n";
460
0a044a7c
DR
461=head2 Circular References
462X<circular reference> X<reference, circular>
463
464It is possible to create a "circular reference" in Perl, which can lead
465to memory leaks. A circular reference occurs when two references
466contain a reference to each other, like this:
467
468 my $foo = {};
469 my $bar = { foo => $foo };
470 $foo->{bar} = $bar;
471
472You can also create a circular reference with a single variable:
473
474 my $foo;
475 $foo = \$foo;
476
477In this case, the reference count for the variables will never reach 0,
478and the references will never be garbage-collected. This can lead to
479memory leaks.
480
481Because objects in Perl are implemented as references, it's possible to
482have circular references with objects as well. Imagine a TreeNode class
483where each node references its parent and child nodes. Any node with a
484parent will be part of a circular reference.
485
486You can break circular references by creating a "weak reference". A
487weak reference does not increment the reference count for a variable,
488which means that the object can go out of scope and be destroyed. You
489can weaken a reference with the C<weaken> function exported by the
490L<Scalar::Util> module.
491
492Here's how we can make the first example safer:
493
494 use Scalar::Util 'weaken';
495
496 my $foo = {};
497 my $bar = { foo => $foo };
498 $foo->{bar} = $bar;
499
500 weaken $foo->{bar};
501
502The reference from C<$foo> to C<$bar> has been weakened. When the
503C<$bar> variable goes out of scope, it will be garbage-collected. The
504next time you look at the value of the C<< $foo->{bar} >> key, it will
505be C<undef>.
506
507This action at a distance can be confusing, so you should be careful
508with your use of weaken. You should weaken the reference in the
509variable that will go out of scope I<first>. That way, the longer-lived
510variable will contain the expected reference until it goes out of
511scope.
512
a0d0e21e 513=head2 Symbolic references
d74e8afc
ITB
514X<reference, symbolic> X<reference, soft>
515X<symbolic reference> X<soft reference>
a0d0e21e
LW
516
517We said that references spring into existence as necessary if they are
518undefined, but we didn't say what happens if a value used as a
19799a22 519reference is already defined, but I<isn't> a hard reference. If you
7c2ea1c7 520use it as a reference, it'll be treated as a symbolic
19799a22 521reference. That is, the value of the scalar is taken to be the I<name>
a0d0e21e
LW
522of a variable, rather than a direct link to a (possibly) anonymous
523value.
524
525People frequently expect it to work like this. So it does.
526
527 $name = "foo";
528 $$name = 1; # Sets $foo
529 ${$name} = 2; # Sets $foo
530 ${$name x 2} = 3; # Sets $foofoo
531 $name->[0] = 4; # Sets $foo[0]
532 @$name = (); # Clears @foo
533 &$name(); # Calls &foo() (as in Perl 4)
534 $pack = "THAT";
535 ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval
536
7c2ea1c7 537This is powerful, and slightly dangerous, in that it's possible
a0d0e21e
LW
538to intend (with the utmost sincerity) to use a hard reference, and
539accidentally use a symbolic reference instead. To protect against
540that, you can say
541
542 use strict 'refs';
543
544and then only hard references will be allowed for the rest of the enclosing
54310121 545block. An inner block may countermand that with
a0d0e21e
LW
546
547 no strict 'refs';
548
5a964f20
TC
549Only package variables (globals, even if localized) are visible to
550symbolic references. Lexical variables (declared with my()) aren't in
551a symbol table, and thus are invisible to this mechanism. For example:
a0d0e21e 552
5a964f20 553 local $value = 10;
b0c35547 554 $ref = "value";
a0d0e21e
LW
555 {
556 my $value = 20;
557 print $$ref;
54310121 558 }
a0d0e21e
LW
559
560This will still print 10, not 20. Remember that local() affects package
561variables, which are all "global" to the package.
562
748a9306
LW
563=head2 Not-so-symbolic references
564
a6006777
PP
565A new feature contributing to readability in perl version 5.001 is that the
566brackets around a symbolic reference behave more like quotes, just as they
748a9306
LW
567always have within a string. That is,
568
569 $push = "pop on ";
570 print "${push}over";
571
7c2ea1c7 572has always meant to print "pop on over", even though push is
748a9306
LW
573a reserved word. This has been generalized to work the same outside
574of quotes, so that
575
576 print ${push} . "over";
577
578and even
579
580 print ${ push } . "over";
581
582will have the same effect. (This would have been a syntax error in
7c2ea1c7 583Perl 5.000, though Perl 4 allowed it in the spaceless form.) This
748a9306
LW
584construct is I<not> considered to be a symbolic reference when you're
585using strict refs:
586
587 use strict 'refs';
588 ${ bareword }; # Okay, means $bareword.
589 ${ "bareword" }; # Error, symbolic reference.
590
591Similarly, because of all the subscripting that is done using single
592words, we've applied the same rule to any bareword that is used for
593subscripting a hash. So now, instead of writing
594
595 $array{ "aaa" }{ "bbb" }{ "ccc" }
596
5f05dabc 597you can write just
748a9306
LW
598
599 $array{ aaa }{ bbb }{ ccc }
600
601and not worry about whether the subscripts are reserved words. In the
602rare event that you do wish to do something like
603
604 $array{ shift }
605
606you can force interpretation as a reserved word by adding anything that
607makes it more than a bareword:
608
609 $array{ shift() }
610 $array{ +shift }
611 $array{ shift @_ }
612
9f1b1f2d
GS
613The C<use warnings> pragma or the B<-w> switch will warn you if it
614interprets a reserved word as a string.
5f05dabc 615But it will no longer warn you about using lowercase words, because the
748a9306
LW
616string is effectively quoted.
617
49399b3f 618=head2 Pseudo-hashes: Using an array as a hash
d74e8afc 619X<pseudo-hash> X<pseudo hash> X<pseudohash>
49399b3f 620
6d822dc4
MS
621Pseudo-hashes have been removed from Perl. The 'fields' pragma
622remains available.
e0478e5a 623
5a964f20 624=head2 Function Templates
d74e8afc
ITB
625X<scope, lexical> X<closure> X<lexical> X<lexical scope>
626X<subroutine, nested> X<sub, nested> X<subroutine, local> X<sub, local>
5a964f20 627
b5c19bd7
DM
628As explained above, an anonymous function with access to the lexical
629variables visible when that function was compiled, creates a closure. It
630retains access to those variables even though it doesn't get run until
631later, such as in a signal handler or a Tk callback.
5a964f20
TC
632
633Using a closure as a function template allows us to generate many functions
c2611fb3 634that act similarly. Suppose you wanted functions named after the colors
5a964f20
TC
635that generated HTML font changes for the various colors:
636
637 print "Be ", red("careful"), "with that ", green("light");
638
7c2ea1c7 639The red() and green() functions would be similar. To create these,
5a964f20
TC
640we'll assign a closure to a typeglob of the name of the function we're
641trying to build.
642
643 @colors = qw(red blue green yellow orange purple violet);
644 for my $name (@colors) {
645 no strict 'refs'; # allow symbol table manipulation
646 *$name = *{uc $name} = sub { "<FONT COLOR='$name'>@_</FONT>" };
647 }
648
649Now all those different functions appear to exist independently. You can
650call red(), RED(), blue(), BLUE(), green(), etc. This technique saves on
651both compile time and memory use, and is less error-prone as well, since
652syntax checks happen at compile time. It's critical that any variables in
653the anonymous subroutine be lexicals in order to create a proper closure.
654That's the reasons for the C<my> on the loop iteration variable.
655
656This is one of the only places where giving a prototype to a closure makes
657much sense. If you wanted to impose scalar context on the arguments of
658these functions (probably not a wise idea for this particular example),
659you could have written it this way instead:
660
661 *$name = sub ($) { "<FONT COLOR='$name'>$_[0]</FONT>" };
662
663However, since prototype checking happens at compile time, the assignment
664above happens too late to be of much use. You could address this by
665putting the whole loop of assignments within a BEGIN block, forcing it
666to occur during compilation.
667
58e2a187
CW
668Access to lexicals that change over time--like those in the C<for> loop
669above, basically aliases to elements from the surrounding lexical scopes--
670only works with anonymous subs, not with named subroutines. Generally
671said, named subroutines do not nest properly and should only be declared
672in the main package scope.
673
674This is because named subroutines are created at compile time so their
675lexical variables get assigned to the parent lexicals from the first
676execution of the parent block. If a parent scope is entered a second
677time, its lexicals are created again, while the nested subs still
678reference the old ones.
679
680Anonymous subroutines get to capture each time you execute the C<sub>
681operator, as they are created on the fly. If you are accustomed to using
682nested subroutines in other programming languages with their own private
683variables, you'll have to work at it a bit in Perl. The intuitive coding
684of this type of thing incurs mysterious warnings about "will not stay
685shared" due to the reasons explained above.
686For example, this won't work:
5a964f20
TC
687
688 sub outer {
689 my $x = $_[0] + 35;
690 sub inner { return $x * 19 } # WRONG
691 return $x + inner();
b432a672 692 }
5a964f20
TC
693
694A work-around is the following:
695
696 sub outer {
697 my $x = $_[0] + 35;
698 local *inner = sub { return $x * 19 };
699 return $x + inner();
b432a672 700 }
5a964f20
TC
701
702Now inner() can only be called from within outer(), because of the
58e2a187
CW
703temporary assignments of the anonymous subroutine. But when it does,
704it has normal access to the lexical variable $x from the scope of
705outer() at the time outer is invoked.
5a964f20
TC
706
707This has the interesting effect of creating a function local to another
708function, something not normally supported in Perl.
709
cb1a09d0 710=head1 WARNING
d74e8afc 711X<reference, string context> X<reference, use as hash key>
748a9306
LW
712
713You may not (usefully) use a reference as the key to a hash. It will be
714converted into a string:
715
716 $x{ \$a } = $a;
717
54310121 718If you try to dereference the key, it won't do a hard dereference, and
184e9718 719you won't accomplish what you're attempting. You might want to do something
cb1a09d0 720more like
748a9306 721
cb1a09d0
AD
722 $r = \@a;
723 $x{ $r } = $r;
724
725And then at least you can use the values(), which will be
726real refs, instead of the keys(), which won't.
727
5a964f20
TC
728The standard Tie::RefHash module provides a convenient workaround to this.
729
cb1a09d0 730=head1 SEE ALSO
a0d0e21e
LW
731
732Besides the obvious documents, source code can be instructive.
7c2ea1c7 733Some pathological examples of the use of references can be found
a0d0e21e 734in the F<t/op/ref.t> regression test in the Perl source directory.
cb1a09d0
AD
735
736See also L<perldsc> and L<perllol> for how to use references to create
5a964f20
TC
737complex data structures, and L<perltoot>, L<perlobj>, and L<perlbot>
738for how to use them to create objects.