Commit | Line | Data |
---|---|---|
a0d0e21e | 1 | =head1 NAME |
d74e8afc | 2 | X<reference> X<pointer> X<data structure> X<structure> X<struct> |
a0d0e21e LW |
3 | |
4 | perlref - Perl references and nested data structures | |
5 | ||
a1e2a320 GS |
6 | =head1 NOTE |
7 | ||
8 | This is complete documentation about all aspects of references. | |
9 | For a shorter, tutorial introduction to just the essential features, | |
10 | see L<perlreftut>. | |
11 | ||
a0d0e21e LW |
12 | =head1 DESCRIPTION |
13 | ||
cb1a09d0 | 14 | Before release 5 of Perl it was difficult to represent complex data |
5a964f20 TC |
15 | structures, because all references had to be symbolic--and even then |
16 | it was difficult to refer to a variable instead of a symbol table entry. | |
17 | Perl now not only makes it easier to use symbolic references to variables, | |
18 | but also lets you have "hard" references to any piece of data or code. | |
19 | Any scalar may hold a hard reference. Because arrays and hashes contain | |
20 | scalars, you can now easily build arrays of arrays, arrays of hashes, | |
21 | hashes of arrays, arrays of hashes of functions, and so on. | |
a0d0e21e LW |
22 | |
23 | Hard references are smart--they keep track of reference counts for you, | |
2d24ed35 | 24 | automatically freeing the thing referred to when its reference count goes |
7c2ea1c7 | 25 | to zero. (Reference counts for values in self-referential or |
2d24ed35 | 26 | cyclic data structures may not go to zero without a little help; see |
2b4f771d | 27 | L</"Circular References"> for a detailed explanation.) |
2d24ed35 CS |
28 | If that thing happens to be an object, the object is destructed. See |
29 | L<perlobj> for more about objects. (In a sense, everything in Perl is an | |
30 | object, but we usually reserve the word for references to objects that | |
31 | have been officially "blessed" into a class package.) | |
32 | ||
33 | Symbolic references are names of variables or other objects, just as a | |
54310121 | 34 | symbolic link in a Unix filesystem contains merely the name of a file. |
d1be9408 | 35 | The C<*glob> notation is something of a symbolic reference. (Symbolic |
2d24ed35 CS |
36 | references are sometimes called "soft references", but please don't call |
37 | them that; references are confusing enough without useless synonyms.) | |
d74e8afc ITB |
38 | X<reference, symbolic> X<reference, soft> |
39 | X<symbolic reference> X<soft reference> | |
2d24ed35 | 40 | |
54310121 | 41 | In contrast, hard references are more like hard links in a Unix file |
2d24ed35 CS |
42 | system: They are used to access an underlying object without concern for |
43 | what its (other) name is. When the word "reference" is used without an | |
5a964f20 | 44 | adjective, as in the following paragraph, it is usually talking about a |
2d24ed35 | 45 | hard reference. |
d74e8afc | 46 | X<reference, hard> X<hard reference> |
2d24ed35 CS |
47 | |
48 | References are easy to use in Perl. There is just one overriding | |
903c0e71 PM |
49 | principle: in general, Perl does no implicit referencing or dereferencing. |
50 | When a scalar is holding a reference, it always behaves as a simple scalar. | |
51 | It doesn't magically start being an array or hash or subroutine; you have to | |
52 | tell it explicitly to do so, by dereferencing it. | |
53 | ||
5a964f20 | 54 | =head2 Making References |
d74e8afc | 55 | X<reference, creation> X<referencing> |
5a964f20 TC |
56 | |
57 | References can be created in several ways. | |
a0d0e21e LW |
58 | |
59 | =over 4 | |
60 | ||
61 | =item 1. | |
d74e8afc | 62 | X<\> X<backslash> |
a0d0e21e LW |
63 | |
64 | By using the backslash operator on a variable, subroutine, or value. | |
d962e436 | 65 | (This works much like the & (address-of) operator in C.) |
7c2ea1c7 | 66 | This typically creates I<another> reference to a variable, because |
a0d0e21e LW |
67 | there's already a reference to the variable in the symbol table. But |
68 | the symbol table reference might go away, and you'll still have the | |
69 | reference 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 |
77 | It isn't possible to create a true reference to an IO handle (filehandle |
78 | or dirhandle) using the backslash operator. The most you can get is a | |
79 | reference to a typeglob, which is actually a complete symbol table entry. | |
80 | But see the explanation of the C<*foo{THING}> syntax below. However, | |
81 | you can still use type globs and globrefs as though they were IO handles. | |
a0d0e21e LW |
82 | |
83 | =item 2. | |
d74e8afc ITB |
84 | X<array, anonymous> X<[> X<[]> X<square bracket> |
85 | X<bracket, square> X<arrayref> X<array reference> X<reference, array> | |
a0d0e21e | 86 | |
5a964f20 | 87 | A reference to an anonymous array can be created using square |
a0d0e21e LW |
88 | brackets: |
89 | ||
90 | $arrayref = [1, 2, ['a', 'b', 'c']]; | |
91 | ||
5a964f20 | 92 | Here we've created a reference to an anonymous array of three elements |
54310121 | 93 | whose final element is itself a reference to another anonymous array of three |
a0d0e21e | 94 | elements. (The multidimensional syntax described later can be used to |
c47ff5f1 | 95 | access this. For example, after the above, C<< $arrayref->[2][1] >> would have |
a0d0e21e LW |
96 | the value "b".) |
97 | ||
7c2ea1c7 | 98 | Taking a reference to an enumerated list is not the same |
cb1a09d0 AD |
99 | as using square brackets--instead it's the same as creating |
100 | a list of references! | |
101 | ||
54310121 | 102 | @list = (\$a, \@b, \%c); |
5566fa15 | 103 | @list = \($a, @b, %c); # same thing! |
58e0a6ae | 104 | |
54310121 | 105 | As a special case, C<\(@foo)> returns a list of references to the contents |
b6429b1b GS |
106 | of C<@foo>, not a reference to C<@foo> itself. Likewise for C<%foo>, |
107 | except that the key references are to copies (since the keys are just | |
108 | strings rather than full-fledged scalars). | |
cb1a09d0 | 109 | |
a0d0e21e | 110 | =item 3. |
d74e8afc ITB |
111 | X<hash, anonymous> X<{> X<{}> X<curly bracket> |
112 | X<bracket, curly> X<brace> X<hashref> X<hash reference> X<reference, hash> | |
a0d0e21e | 113 | |
5a964f20 | 114 | A reference to an anonymous hash can be created using curly |
a0d0e21e LW |
115 | brackets: |
116 | ||
117 | $hashref = { | |
5566fa15 SF |
118 | 'Adam' => 'Eve', |
119 | 'Clyde' => 'Bonnie', | |
a0d0e21e LW |
120 | }; |
121 | ||
5a964f20 | 122 | Anonymous hash and array composers like these can be intermixed freely to |
a0d0e21e LW |
123 | produce as complicated a structure as you want. The multidimensional |
124 | syntax described below works for these too. The values above are | |
125 | literals, but variables and expressions would work just as well, because | |
126 | assignment operators in Perl (even within local() or my()) are executable | |
127 | statements, not compile-time declarations. | |
128 | ||
129 | Because curly brackets (braces) are used for several other things | |
130 | including BLOCKs, you may occasionally have to disambiguate braces at the | |
131 | beginning of a statement by putting a C<+> or a C<return> in front so | |
132 | that Perl realizes the opening brace isn't starting a BLOCK. The economy and | |
133 | mnemonic value of using curlies is deemed worth this occasional extra | |
134 | hassle. | |
135 | ||
136 | For example, if you wanted a function to make a new hash and return a | |
137 | reference to it, you have these options: | |
138 | ||
139 | sub hashem { { @_ } } # silently wrong | |
140 | sub hashem { +{ @_ } } # ok | |
141 | sub hashem { return { @_ } } # ok | |
142 | ||
ebc58f1a GS |
143 | On the other hand, if you want the other meaning, you can do this: |
144 | ||
555bd962 BG |
145 | sub showem { { @_ } } # ambiguous (currently ok, |
146 | # but may change) | |
ebc58f1a GS |
147 | sub showem { {; @_ } } # ok |
148 | sub showem { { return @_ } } # ok | |
149 | ||
7c2ea1c7 | 150 | The leading C<+{> and C<{;> always serve to disambiguate |
ebc58f1a GS |
151 | the expression to mean either the HASH reference, or the BLOCK. |
152 | ||
a0d0e21e | 153 | =item 4. |
d74e8afc ITB |
154 | X<subroutine, anonymous> X<subroutine, reference> X<reference, subroutine> |
155 | X<scope, lexical> X<closure> X<lexical> X<lexical scope> | |
a0d0e21e | 156 | |
5a964f20 | 157 | A reference to an anonymous subroutine can be created by using |
a0d0e21e LW |
158 | C<sub> without a subname: |
159 | ||
160 | $coderef = sub { print "Boink!\n" }; | |
161 | ||
7c2ea1c7 GS |
162 | Note the semicolon. Except for the code |
163 | inside not being immediately executed, a C<sub {}> is not so much a | |
a0d0e21e | 164 | declaration as it is an operator, like C<do{}> or C<eval{}>. (However, no |
5a964f20 | 165 | matter how many times you execute that particular line (unless you're in an |
19799a22 | 166 | C<eval("...")>), $coderef will still have a reference to the I<same> |
a0d0e21e LW |
167 | anonymous subroutine.) |
168 | ||
748a9306 | 169 | Anonymous subroutines act as closures with respect to my() variables, |
7c2ea1c7 | 170 | that is, variables lexically visible within the current scope. Closure |
748a9306 LW |
171 | is a notion out of the Lisp world that says if you define an anonymous |
172 | function in a particular lexical context, it pretends to run in that | |
7c2ea1c7 | 173 | context even when it's called outside the context. |
748a9306 LW |
174 | |
175 | In human terms, it's a funny way of passing arguments to a subroutine when | |
176 | you define it as well as when you call it. It's useful for setting up | |
177 | little bits of code to run later, such as callbacks. You can even | |
54310121 | 178 | do object-oriented stuff with it, though Perl already provides a different |
179 | mechanism to do that--see L<perlobj>. | |
748a9306 | 180 | |
7c2ea1c7 GS |
181 | You might also think of closure as a way to write a subroutine |
182 | template without using eval(). Here's a small example of how | |
183 | closures work: | |
748a9306 LW |
184 | |
185 | sub newprint { | |
5566fa15 SF |
186 | my $x = shift; |
187 | return sub { my $y = shift; print "$x, $y!\n"; }; | |
a0d0e21e | 188 | } |
748a9306 LW |
189 | $h = newprint("Howdy"); |
190 | $g = newprint("Greetings"); | |
191 | ||
192 | # Time passes... | |
193 | ||
194 | &$h("world"); | |
195 | &$g("earthlings"); | |
a0d0e21e | 196 | |
748a9306 LW |
197 | This prints |
198 | ||
199 | Howdy, world! | |
200 | Greetings, earthlings! | |
201 | ||
7c2ea1c7 GS |
202 | Note particularly that $x continues to refer to the value passed |
203 | into newprint() I<despite> "my $x" having gone out of scope by the | |
204 | time the anonymous subroutine runs. That's what a closure is all | |
205 | about. | |
748a9306 | 206 | |
5a964f20 | 207 | This applies only to lexical variables, by the way. Dynamic variables |
748a9306 LW |
208 | continue to work as they have always worked. Closure is not something |
209 | that most Perl programmers need trouble themselves about to begin with. | |
a0d0e21e LW |
210 | |
211 | =item 5. | |
d74e8afc | 212 | X<constructor> X<new> |
a0d0e21e | 213 | |
63acfd00 | 214 | References are often returned by special subroutines called constructors. Perl |
215 | objects are just references to a special type of object that happens to know | |
216 | which package it's associated with. Constructors are just special subroutines | |
217 | that know how to create that association. They do so by starting with an | |
218 | ordinary reference, and it remains an ordinary reference even while it's also | |
219 | being an object. Constructors are often named C<new()>. You I<can> call them | |
220 | indirectly: | |
221 | ||
222 | $objref = new Doggie( Tail => 'short', Ears => 'long' ); | |
223 | ||
224 | But that can produce ambiguous syntax in certain cases, so it's often | |
225 | better to use the direct method invocation approach: | |
5a964f20 TC |
226 | |
227 | $objref = Doggie->new(Tail => 'short', Ears => 'long'); | |
228 | ||
229 | use Term::Cap; | |
230 | $terminal = Term::Cap->Tgetent( { OSPEED => 9600 }); | |
231 | ||
232 | use Tk; | |
233 | $main = MainWindow->new(); | |
234 | $menubar = $main->Frame(-relief => "raised", | |
235 | -borderwidth => 2) | |
236 | ||
a0d0e21e | 237 | =item 6. |
d74e8afc | 238 | X<autovivification> |
a0d0e21e LW |
239 | |
240 | References of the appropriate type can spring into existence if you | |
5f05dabc | 241 | dereference them in a context that assumes they exist. Because we haven't |
a0d0e21e LW |
242 | talked about dereferencing yet, we can't show you any examples yet. |
243 | ||
cb1a09d0 | 244 | =item 7. |
d74e8afc | 245 | X<*foo{THING}> X<*> |
cb1a09d0 | 246 | |
55497cff | 247 | A reference can be created by using a special syntax, lovingly known as |
248 | the *foo{THING} syntax. *foo{THING} returns a reference to the THING | |
249 | slot in *foo (which is the symbol table entry which holds everything | |
250 | known as foo). | |
cb1a09d0 | 251 | |
55497cff | 252 | $scalarref = *foo{SCALAR}; |
253 | $arrayref = *ARGV{ARRAY}; | |
254 | $hashref = *ENV{HASH}; | |
255 | $coderef = *handler{CODE}; | |
36477c24 | 256 | $ioref = *STDIN{IO}; |
55497cff | 257 | $globref = *foo{GLOB}; |
c0bd1adc | 258 | $formatref = *foo{FORMAT}; |
171e2879 FC |
259 | $globname = *foo{NAME}; # "foo" |
260 | $pkgname = *foo{PACKAGE}; # "main" | |
55497cff | 261 | |
171e2879 FC |
262 | Most of these are self-explanatory, but C<*foo{IO}> |
263 | deserves special attention. It returns | |
7c2ea1c7 GS |
264 | the IO handle, used for file handles (L<perlfunc/open>), sockets |
265 | (L<perlfunc/socket> and L<perlfunc/socketpair>), and directory | |
266 | handles (L<perlfunc/opendir>). For compatibility with previous | |
39b99f21 | 267 | versions of Perl, C<*foo{FILEHANDLE}> is a synonym for C<*foo{IO}>, though it |
83677dc5 RS |
268 | is discouraged, to encourage a consistent use of one name: IO. On perls |
269 | between v5.8 and v5.22, it will issue a deprecation warning, but this | |
270 | deprecation has since been rescinded. | |
55497cff | 271 | |
7c2ea1c7 GS |
272 | C<*foo{THING}> returns undef if that particular THING hasn't been used yet, |
273 | except in the case of scalars. C<*foo{SCALAR}> returns a reference to an | |
5f05dabc | 274 | anonymous scalar if $foo hasn't been used yet. This might change in a |
275 | future release. | |
276 | ||
171e2879 FC |
277 | C<*foo{NAME}> and C<*foo{PACKAGE}> are the exception, in that they return |
278 | strings, rather than references. These return the package and name of the | |
279 | typeglob itself, rather than one that has been assigned to it. So, after | |
280 | C<*foo=*Foo::bar>, C<*foo> will become "*Foo::bar" when used as a string, | |
281 | but C<*foo{PACKAGE}> and C<*foo{NAME}> will continue to produce "main" and | |
282 | "foo", respectively. | |
283 | ||
7c2ea1c7 | 284 | C<*foo{IO}> is an alternative to the C<*HANDLE> mechanism given in |
5a964f20 TC |
285 | L<perldata/"Typeglobs and Filehandles"> for passing filehandles |
286 | into or out of subroutines, or storing into larger data structures. | |
287 | Its disadvantage is that it won't create a new filehandle for you. | |
7c2ea1c7 GS |
288 | Its advantage is that you have less risk of clobbering more than |
289 | you want to with a typeglob assignment. (It still conflates file | |
290 | and directory handles, though.) However, if you assign the incoming | |
291 | value to a scalar instead of a typeglob as we do in the examples | |
292 | below, there's no risk of that happening. | |
36477c24 | 293 | |
5566fa15 SF |
294 | splutter(*STDOUT); # pass the whole glob |
295 | splutter(*STDOUT{IO}); # pass both file and dir handles | |
5a964f20 | 296 | |
cb1a09d0 | 297 | sub splutter { |
5566fa15 SF |
298 | my $fh = shift; |
299 | print $fh "her um well a hmmm\n"; | |
cb1a09d0 AD |
300 | } |
301 | ||
5566fa15 | 302 | $rec = get_rec(*STDIN); # pass the whole glob |
7c2ea1c7 | 303 | $rec = get_rec(*STDIN{IO}); # pass both file and dir handles |
5a964f20 | 304 | |
cb1a09d0 | 305 | sub get_rec { |
5566fa15 SF |
306 | my $fh = shift; |
307 | return scalar <$fh>; | |
cb1a09d0 AD |
308 | } |
309 | ||
a0d0e21e LW |
310 | =back |
311 | ||
5a964f20 | 312 | =head2 Using References |
d74e8afc | 313 | X<reference, use> X<dereferencing> X<dereference> |
5a964f20 | 314 | |
a0d0e21e LW |
315 | That's it for creating references. By now you're probably dying to |
316 | know how to use references to get back to your long-lost data. There | |
317 | are several basic methods. | |
318 | ||
319 | =over 4 | |
320 | ||
321 | =item 1. | |
322 | ||
6309d9d9 | 323 | Anywhere you'd put an identifier (or chain of identifiers) as part |
324 | of a variable or subroutine name, you can replace the identifier with | |
325 | a simple scalar variable containing a reference of the correct type: | |
a0d0e21e LW |
326 | |
327 | $bar = $$scalarref; | |
328 | push(@$arrayref, $filename); | |
329 | $$arrayref[0] = "January"; | |
330 | $$hashref{"KEY"} = "VALUE"; | |
331 | &$coderef(1,2,3); | |
cb1a09d0 | 332 | print $globref "output\n"; |
a0d0e21e | 333 | |
19799a22 | 334 | It's important to understand that we are specifically I<not> dereferencing |
a0d0e21e | 335 | C<$arrayref[0]> or C<$hashref{"KEY"}> there. The dereference of the |
19799a22 | 336 | scalar variable happens I<before> it does any key lookups. Anything more |
a0d0e21e LW |
337 | complicated than a simple scalar variable must use methods 2 or 3 below. |
338 | However, a "simple scalar" includes an identifier that itself uses method | |
339 | 1 recursively. Therefore, the following prints "howdy". | |
340 | ||
341 | $refrefref = \\\"howdy"; | |
342 | print $$$$refrefref; | |
343 | ||
344 | =item 2. | |
345 | ||
6309d9d9 | 346 | Anywhere you'd put an identifier (or chain of identifiers) as part of a |
347 | variable or subroutine name, you can replace the identifier with a | |
348 | BLOCK returning a reference of the correct type. In other words, the | |
349 | previous examples could be written like this: | |
a0d0e21e LW |
350 | |
351 | $bar = ${$scalarref}; | |
352 | push(@{$arrayref}, $filename); | |
353 | ${$arrayref}[0] = "January"; | |
354 | ${$hashref}{"KEY"} = "VALUE"; | |
355 | &{$coderef}(1,2,3); | |
36477c24 | 356 | $globref->print("output\n"); # iff IO::Handle is loaded |
a0d0e21e LW |
357 | |
358 | Admittedly, it's a little silly to use the curlies in this case, but | |
359 | the BLOCK can contain any arbitrary expression, in particular, | |
360 | subscripted expressions: | |
361 | ||
5566fa15 | 362 | &{ $dispatch{$index} }(1,2,3); # call correct routine |
a0d0e21e LW |
363 | |
364 | Because of being able to omit the curlies for the simple case of C<$$x>, | |
365 | people often make the mistake of viewing the dereferencing symbols as | |
366 | proper operators, and wonder about their precedence. If they were, | |
5f05dabc | 367 | though, you could use parentheses instead of braces. That's not the case. |
a0d0e21e | 368 | Consider the difference below; case 0 is a short-hand version of case 1, |
19799a22 | 369 | I<not> case 2: |
a0d0e21e | 370 | |
5566fa15 SF |
371 | $$hashref{"KEY"} = "VALUE"; # CASE 0 |
372 | ${$hashref}{"KEY"} = "VALUE"; # CASE 1 | |
373 | ${$hashref{"KEY"}} = "VALUE"; # CASE 2 | |
374 | ${$hashref->{"KEY"}} = "VALUE"; # CASE 3 | |
a0d0e21e LW |
375 | |
376 | Case 2 is also deceptive in that you're accessing a variable | |
377 | called %hashref, not dereferencing through $hashref to the hash | |
378 | it's presumably referencing. That would be case 3. | |
379 | ||
380 | =item 3. | |
381 | ||
6da72b64 CS |
382 | Subroutine calls and lookups of individual array elements arise often |
383 | enough that it gets cumbersome to use method 2. As a form of | |
384 | syntactic sugar, the examples for method 2 may be written: | |
a0d0e21e | 385 | |
6da72b64 CS |
386 | $arrayref->[0] = "January"; # Array element |
387 | $hashref->{"KEY"} = "VALUE"; # Hash element | |
388 | $coderef->(1,2,3); # Subroutine call | |
a0d0e21e | 389 | |
6da72b64 | 390 | The left side of the arrow can be any expression returning a reference, |
19799a22 | 391 | including a previous dereference. Note that C<$array[$x]> is I<not> the |
c47ff5f1 | 392 | same thing as C<< $array->[$x] >> here: |
a0d0e21e LW |
393 | |
394 | $array[$x]->{"foo"}->[0] = "January"; | |
395 | ||
396 | This is one of the cases we mentioned earlier in which references could | |
397 | spring into existence when in an lvalue context. Before this | |
398 | statement, C<$array[$x]> may have been undefined. If so, it's | |
399 | automatically defined with a hash reference so that we can look up | |
c47ff5f1 | 400 | C<{"foo"}> in it. Likewise C<< $array[$x]->{"foo"} >> will automatically get |
a0d0e21e | 401 | defined with an array reference so that we can look up C<[0]> in it. |
5a964f20 | 402 | This process is called I<autovivification>. |
a0d0e21e | 403 | |
19799a22 | 404 | One more thing here. The arrow is optional I<between> brackets |
a0d0e21e LW |
405 | subscripts, so you can shrink the above down to |
406 | ||
407 | $array[$x]{"foo"}[0] = "January"; | |
408 | ||
409 | Which, in the degenerate case of using only ordinary arrays, gives you | |
410 | multidimensional arrays just like C's: | |
411 | ||
412 | $score[$x][$y][$z] += 42; | |
413 | ||
414 | Well, okay, not entirely like C's arrays, actually. C doesn't know how | |
415 | to grow its arrays on demand. Perl does. | |
416 | ||
417 | =item 4. | |
418 | ||
419 | If a reference happens to be a reference to an object, then there are | |
420 | probably methods to access the things referred to, and you should probably | |
421 | stick to those methods unless you're in the class package that defines the | |
422 | object's methods. In other words, be nice, and don't violate the object's | |
423 | encapsulation without a very good reason. Perl does not enforce | |
424 | encapsulation. We are not totalitarians here. We do expect some basic | |
425 | civility though. | |
426 | ||
427 | =back | |
428 | ||
7c2ea1c7 GS |
429 | Using a string or number as a reference produces a symbolic reference, |
430 | as explained above. Using a reference as a number produces an | |
431 | integer representing its storage location in memory. The only | |
432 | useful thing to be done with this is to compare two references | |
433 | numerically to see whether they refer to the same location. | |
d74e8afc | 434 | X<reference, numeric context> |
7c2ea1c7 GS |
435 | |
436 | if ($ref1 == $ref2) { # cheap numeric compare of references | |
5566fa15 | 437 | print "refs 1 and 2 refer to the same thing\n"; |
7c2ea1c7 GS |
438 | } |
439 | ||
440 | Using a reference as a string produces both its referent's type, | |
441 | including any package blessing as described in L<perlobj>, as well | |
442 | as the numeric address expressed in hex. The ref() operator returns | |
443 | just the type of thing the reference is pointing to, without the | |
444 | address. See L<perlfunc/ref> for details and examples of its use. | |
d74e8afc | 445 | X<reference, string context> |
a0d0e21e | 446 | |
5a964f20 TC |
447 | The bless() operator may be used to associate the object a reference |
448 | points to with a package functioning as an object class. See L<perlobj>. | |
a0d0e21e | 449 | |
5f05dabc | 450 | A typeglob may be dereferenced the same way a reference can, because |
7c2ea1c7 | 451 | the dereference syntax always indicates the type of reference desired. |
a0d0e21e LW |
452 | So C<${*foo}> and C<${\$foo}> both indicate the same scalar variable. |
453 | ||
454 | Here's a trick for interpolating a subroutine call into a string: | |
455 | ||
cb1a09d0 AD |
456 | print "My sub returned @{[mysub(1,2,3)]} that time.\n"; |
457 | ||
458 | The way it works is that when the C<@{...}> is seen in the double-quoted | |
459 | string, it's evaluated as a block. The block creates a reference to an | |
460 | anonymous array containing the results of the call to C<mysub(1,2,3)>. So | |
461 | the whole block returns a reference to an array, which is then | |
462 | dereferenced by C<@{...}> and stuck into the double-quoted string. This | |
463 | chicanery is also useful for arbitrary expressions: | |
a0d0e21e | 464 | |
184e9718 | 465 | print "That yields @{[$n + 5]} widgets\n"; |
a0d0e21e | 466 | |
35efdb20 DL |
467 | Similarly, an expression that returns a reference to a scalar can be |
468 | dereferenced via C<${...}>. Thus, the above expression may be written | |
469 | as: | |
470 | ||
471 | print "That yields ${\($n + 5)} widgets\n"; | |
472 | ||
0a044a7c DR |
473 | =head2 Circular References |
474 | X<circular reference> X<reference, circular> | |
475 | ||
476 | It is possible to create a "circular reference" in Perl, which can lead | |
477 | to memory leaks. A circular reference occurs when two references | |
478 | contain a reference to each other, like this: | |
479 | ||
480 | my $foo = {}; | |
481 | my $bar = { foo => $foo }; | |
482 | $foo->{bar} = $bar; | |
483 | ||
484 | You can also create a circular reference with a single variable: | |
485 | ||
486 | my $foo; | |
487 | $foo = \$foo; | |
488 | ||
489 | In this case, the reference count for the variables will never reach 0, | |
490 | and the references will never be garbage-collected. This can lead to | |
491 | memory leaks. | |
492 | ||
493 | Because objects in Perl are implemented as references, it's possible to | |
494 | have circular references with objects as well. Imagine a TreeNode class | |
495 | where each node references its parent and child nodes. Any node with a | |
496 | parent will be part of a circular reference. | |
497 | ||
498 | You can break circular references by creating a "weak reference". A | |
499 | weak reference does not increment the reference count for a variable, | |
500 | which means that the object can go out of scope and be destroyed. You | |
501 | can weaken a reference with the C<weaken> function exported by the | |
502 | L<Scalar::Util> module. | |
503 | ||
504 | Here's how we can make the first example safer: | |
505 | ||
506 | use Scalar::Util 'weaken'; | |
507 | ||
508 | my $foo = {}; | |
509 | my $bar = { foo => $foo }; | |
510 | $foo->{bar} = $bar; | |
511 | ||
512 | weaken $foo->{bar}; | |
513 | ||
514 | The reference from C<$foo> to C<$bar> has been weakened. When the | |
515 | C<$bar> variable goes out of scope, it will be garbage-collected. The | |
516 | next time you look at the value of the C<< $foo->{bar} >> key, it will | |
517 | be C<undef>. | |
518 | ||
519 | This action at a distance can be confusing, so you should be careful | |
520 | with your use of weaken. You should weaken the reference in the | |
521 | variable that will go out of scope I<first>. That way, the longer-lived | |
522 | variable will contain the expected reference until it goes out of | |
523 | scope. | |
524 | ||
a0d0e21e | 525 | =head2 Symbolic references |
d74e8afc ITB |
526 | X<reference, symbolic> X<reference, soft> |
527 | X<symbolic reference> X<soft reference> | |
a0d0e21e LW |
528 | |
529 | We said that references spring into existence as necessary if they are | |
530 | undefined, but we didn't say what happens if a value used as a | |
19799a22 | 531 | reference is already defined, but I<isn't> a hard reference. If you |
7c2ea1c7 | 532 | use it as a reference, it'll be treated as a symbolic |
19799a22 | 533 | reference. That is, the value of the scalar is taken to be the I<name> |
a0d0e21e LW |
534 | of a variable, rather than a direct link to a (possibly) anonymous |
535 | value. | |
536 | ||
537 | People frequently expect it to work like this. So it does. | |
538 | ||
539 | $name = "foo"; | |
5566fa15 SF |
540 | $$name = 1; # Sets $foo |
541 | ${$name} = 2; # Sets $foo | |
542 | ${$name x 2} = 3; # Sets $foofoo | |
543 | $name->[0] = 4; # Sets $foo[0] | |
544 | @$name = (); # Clears @foo | |
545 | &$name(); # Calls &foo() | |
a0d0e21e | 546 | $pack = "THAT"; |
5566fa15 | 547 | ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval |
a0d0e21e | 548 | |
7c2ea1c7 | 549 | This is powerful, and slightly dangerous, in that it's possible |
a0d0e21e LW |
550 | to intend (with the utmost sincerity) to use a hard reference, and |
551 | accidentally use a symbolic reference instead. To protect against | |
552 | that, you can say | |
553 | ||
554 | use strict 'refs'; | |
555 | ||
556 | and then only hard references will be allowed for the rest of the enclosing | |
54310121 | 557 | block. An inner block may countermand that with |
a0d0e21e LW |
558 | |
559 | no strict 'refs'; | |
560 | ||
5a964f20 TC |
561 | Only package variables (globals, even if localized) are visible to |
562 | symbolic references. Lexical variables (declared with my()) aren't in | |
563 | a symbol table, and thus are invisible to this mechanism. For example: | |
a0d0e21e | 564 | |
5a964f20 | 565 | local $value = 10; |
b0c35547 | 566 | $ref = "value"; |
a0d0e21e | 567 | { |
5566fa15 SF |
568 | my $value = 20; |
569 | print $$ref; | |
54310121 | 570 | } |
a0d0e21e LW |
571 | |
572 | This will still print 10, not 20. Remember that local() affects package | |
573 | variables, which are all "global" to the package. | |
574 | ||
748a9306 LW |
575 | =head2 Not-so-symbolic references |
576 | ||
0480bf32 | 577 | Brackets around a symbolic reference can simply |
903c0e71 PM |
578 | serve to isolate an identifier or variable name from the rest of an |
579 | expression, just as they always have within a string. For example, | |
748a9306 LW |
580 | |
581 | $push = "pop on "; | |
582 | print "${push}over"; | |
583 | ||
7c2ea1c7 | 584 | has always meant to print "pop on over", even though push is |
0480bf32 | 585 | a reserved word. This is generalized to work the same |
903c0e71 | 586 | without the enclosing double quotes, so that |
748a9306 LW |
587 | |
588 | print ${push} . "over"; | |
589 | ||
590 | and even | |
591 | ||
592 | print ${ push } . "over"; | |
593 | ||
0480bf32 | 594 | will have the same effect. This |
748a9306 LW |
595 | construct is I<not> considered to be a symbolic reference when you're |
596 | using strict refs: | |
597 | ||
598 | use strict 'refs'; | |
5566fa15 SF |
599 | ${ bareword }; # Okay, means $bareword. |
600 | ${ "bareword" }; # Error, symbolic reference. | |
748a9306 | 601 | |
903c0e71 PM |
602 | Similarly, because of all the subscripting that is done using single words, |
603 | the same rule applies to any bareword that is used for subscripting a hash. | |
604 | So now, instead of writing | |
748a9306 LW |
605 | |
606 | $array{ "aaa" }{ "bbb" }{ "ccc" } | |
607 | ||
5f05dabc | 608 | you can write just |
748a9306 LW |
609 | |
610 | $array{ aaa }{ bbb }{ ccc } | |
611 | ||
612 | and not worry about whether the subscripts are reserved words. In the | |
613 | rare event that you do wish to do something like | |
614 | ||
615 | $array{ shift } | |
616 | ||
617 | you can force interpretation as a reserved word by adding anything that | |
618 | makes it more than a bareword: | |
619 | ||
620 | $array{ shift() } | |
621 | $array{ +shift } | |
622 | $array{ shift @_ } | |
623 | ||
9f1b1f2d GS |
624 | The C<use warnings> pragma or the B<-w> switch will warn you if it |
625 | interprets a reserved word as a string. | |
5f05dabc | 626 | But it will no longer warn you about using lowercase words, because the |
748a9306 LW |
627 | string is effectively quoted. |
628 | ||
49399b3f | 629 | =head2 Pseudo-hashes: Using an array as a hash |
d74e8afc | 630 | X<pseudo-hash> X<pseudo hash> X<pseudohash> |
49399b3f | 631 | |
6d822dc4 MS |
632 | Pseudo-hashes have been removed from Perl. The 'fields' pragma |
633 | remains available. | |
e0478e5a | 634 | |
5a964f20 | 635 | =head2 Function Templates |
d74e8afc ITB |
636 | X<scope, lexical> X<closure> X<lexical> X<lexical scope> |
637 | X<subroutine, nested> X<sub, nested> X<subroutine, local> X<sub, local> | |
5a964f20 | 638 | |
b5c19bd7 DM |
639 | As explained above, an anonymous function with access to the lexical |
640 | variables visible when that function was compiled, creates a closure. It | |
641 | retains access to those variables even though it doesn't get run until | |
642 | later, such as in a signal handler or a Tk callback. | |
5a964f20 TC |
643 | |
644 | Using a closure as a function template allows us to generate many functions | |
c2611fb3 | 645 | that act similarly. Suppose you wanted functions named after the colors |
5a964f20 TC |
646 | that generated HTML font changes for the various colors: |
647 | ||
648 | print "Be ", red("careful"), "with that ", green("light"); | |
649 | ||
7c2ea1c7 | 650 | The red() and green() functions would be similar. To create these, |
5a964f20 | 651 | we'll assign a closure to a typeglob of the name of the function we're |
d962e436 | 652 | trying to build. |
5a964f20 TC |
653 | |
654 | @colors = qw(red blue green yellow orange purple violet); | |
655 | for my $name (@colors) { | |
5566fa15 | 656 | no strict 'refs'; # allow symbol table manipulation |
5a964f20 | 657 | *$name = *{uc $name} = sub { "<FONT COLOR='$name'>@_</FONT>" }; |
d962e436 | 658 | } |
5a964f20 TC |
659 | |
660 | Now all those different functions appear to exist independently. You can | |
661 | call red(), RED(), blue(), BLUE(), green(), etc. This technique saves on | |
662 | both compile time and memory use, and is less error-prone as well, since | |
663 | syntax checks happen at compile time. It's critical that any variables in | |
664 | the anonymous subroutine be lexicals in order to create a proper closure. | |
665 | That's the reasons for the C<my> on the loop iteration variable. | |
666 | ||
667 | This is one of the only places where giving a prototype to a closure makes | |
668 | much sense. If you wanted to impose scalar context on the arguments of | |
669 | these functions (probably not a wise idea for this particular example), | |
670 | you could have written it this way instead: | |
671 | ||
672 | *$name = sub ($) { "<FONT COLOR='$name'>$_[0]</FONT>" }; | |
673 | ||
674 | However, since prototype checking happens at compile time, the assignment | |
675 | above happens too late to be of much use. You could address this by | |
676 | putting the whole loop of assignments within a BEGIN block, forcing it | |
677 | to occur during compilation. | |
678 | ||
58e2a187 CW |
679 | Access to lexicals that change over time--like those in the C<for> loop |
680 | above, basically aliases to elements from the surrounding lexical scopes-- | |
681 | only works with anonymous subs, not with named subroutines. Generally | |
682 | said, named subroutines do not nest properly and should only be declared | |
683 | in the main package scope. | |
684 | ||
685 | This is because named subroutines are created at compile time so their | |
686 | lexical variables get assigned to the parent lexicals from the first | |
687 | execution of the parent block. If a parent scope is entered a second | |
688 | time, its lexicals are created again, while the nested subs still | |
689 | reference the old ones. | |
690 | ||
691 | Anonymous subroutines get to capture each time you execute the C<sub> | |
692 | operator, as they are created on the fly. If you are accustomed to using | |
693 | nested subroutines in other programming languages with their own private | |
694 | variables, you'll have to work at it a bit in Perl. The intuitive coding | |
695 | of this type of thing incurs mysterious warnings about "will not stay | |
d962e436 | 696 | shared" due to the reasons explained above. |
58e2a187 | 697 | For example, this won't work: |
5a964f20 TC |
698 | |
699 | sub outer { | |
700 | my $x = $_[0] + 35; | |
701 | sub inner { return $x * 19 } # WRONG | |
702 | return $x + inner(); | |
b432a672 | 703 | } |
5a964f20 TC |
704 | |
705 | A work-around is the following: | |
706 | ||
707 | sub outer { | |
708 | my $x = $_[0] + 35; | |
709 | local *inner = sub { return $x * 19 }; | |
710 | return $x + inner(); | |
b432a672 | 711 | } |
5a964f20 TC |
712 | |
713 | Now inner() can only be called from within outer(), because of the | |
58e2a187 CW |
714 | temporary assignments of the anonymous subroutine. But when it does, |
715 | it has normal access to the lexical variable $x from the scope of | |
716 | outer() at the time outer is invoked. | |
5a964f20 TC |
717 | |
718 | This has the interesting effect of creating a function local to another | |
719 | function, something not normally supported in Perl. | |
720 | ||
f0d99131 | 721 | =head1 WARNING: Don't use references as hash keys |
d74e8afc | 722 | X<reference, string context> X<reference, use as hash key> |
748a9306 LW |
723 | |
724 | You may not (usefully) use a reference as the key to a hash. It will be | |
725 | converted into a string: | |
726 | ||
727 | $x{ \$a } = $a; | |
728 | ||
54310121 | 729 | If you try to dereference the key, it won't do a hard dereference, and |
184e9718 | 730 | you won't accomplish what you're attempting. You might want to do something |
cb1a09d0 | 731 | more like |
748a9306 | 732 | |
cb1a09d0 AD |
733 | $r = \@a; |
734 | $x{ $r } = $r; | |
735 | ||
736 | And then at least you can use the values(), which will be | |
737 | real refs, instead of the keys(), which won't. | |
738 | ||
5a964f20 TC |
739 | The standard Tie::RefHash module provides a convenient workaround to this. |
740 | ||
f0d99131 | 741 | =head2 Postfix Dereference Syntax |
821361b6 RS |
742 | |
743 | Beginning in v5.20.0, a postfix syntax for using references is | |
744 | available. It behaves as described in L</Using References>, but instead | |
745 | of a prefixed sigil, a postfixed sigil-and-star is used. | |
746 | ||
747 | For example: | |
748 | ||
749 | $r = \@a; | |
750 | @b = $r->@*; # equivalent to @$r or @{ $r } | |
751 | ||
752 | $r = [ 1, [ 2, 3 ], 4 ]; | |
753 | $r->[1]->@*; # equivalent to @{ $r->[1] } | |
754 | ||
1c2511e0 AC |
755 | In Perl 5.20 and 5.22, this syntax must be enabled with C<use feature |
756 | 'postderef'>. As of Perl 5.24, no feature declarations are required to make | |
757 | it available. | |
821361b6 RS |
758 | |
759 | Postfix dereference should work in all circumstances where block | |
760 | (circumfix) dereference worked, and should be entirely equivalent. This | |
761 | syntax allows dereferencing to be written and read entirely | |
762 | left-to-right. The following equivalencies are defined: | |
763 | ||
2bcaee20 FC |
764 | $sref->$*; # same as ${ $sref } |
765 | $aref->@*; # same as @{ $aref } | |
766 | $aref->$#*; # same as $#{ $aref } | |
767 | $href->%*; # same as %{ $href } | |
768 | $cref->&*; # same as &{ $cref } | |
769 | $gref->**; # same as *{ $gref } | |
821361b6 RS |
770 | |
771 | Note especially that C<< $cref->&* >> is I<not> equivalent to C<< | |
864eb29a RS |
772 | $cref->() >>, and can serve different purposes. |
773 | ||
774 | Glob elements can be extracted through the postfix dereferencing feature: | |
775 | ||
776 | $gref->*{SCALAR}; # same as *{ $gref }{SCALAR} | |
821361b6 RS |
777 | |
778 | Postfix array and scalar dereferencing I<can> be used in interpolating | |
779 | strings (double quotes or the C<qq> operator), but only if the | |
1c2511e0 | 780 | C<postderef_qq> feature is enabled. |
821361b6 RS |
781 | |
782 | =head2 Postfix Reference Slicing | |
783 | ||
784 | Value slices of arrays and hashes may also be taken with postfix | |
785 | dereferencing notation, with the following equivalencies: | |
786 | ||
787 | $aref->@[ ... ]; # same as @$aref[ ... ] | |
788 | $href->@{ ... }; # same as @$href{ ... } | |
789 | ||
864eb29a RS |
790 | Postfix key/value pair slicing, added in 5.20.0 and documented in |
791 | L<the KeyE<sol>Value Hash Slices section of perldata|perldata/"Key/Value Hash | |
792 | Slices">, also behaves as expected: | |
821361b6 RS |
793 | |
794 | $aref->%[ ... ]; # same as %$aref[ ... ] | |
795 | $href->%{ ... }; # same as %$href{ ... } | |
796 | ||
797 | As with postfix array, postfix value slice dereferencing I<can> be used | |
798 | in interpolating strings (double quotes or the C<qq> operator), but only | |
1c2511e0 | 799 | if the C<postderef_qq> L<feature> is enabled. |
821361b6 | 800 | |
f0d99131 | 801 | =head2 Assigning to References |
82848c10 FC |
802 | |
803 | Beginning in v5.22.0, the referencing operator can be assigned to. It | |
804 | performs an aliasing operation, so that the variable name referenced on the | |
805 | left-hand side becomes an alias for the thing referenced on the right-hand | |
806 | side: | |
807 | ||
808 | \$a = \$b; # $a and $b now point to the same scalar | |
809 | \&foo = \&bar; # foo() now means bar() | |
810 | ||
baabe3fb | 811 | This syntax must be enabled with C<use feature 'refaliasing'>. It is |
82848c10 | 812 | experimental, and will warn by default unless C<no warnings |
baabe3fb | 813 | 'experimental::refaliasing'> is in effect. |
82848c10 FC |
814 | |
815 | These forms may be assigned to, and cause the right-hand side to be | |
816 | evaluated in scalar context: | |
817 | ||
818 | \$scalar | |
819 | \@array | |
820 | \%hash | |
821 | \&sub | |
822 | \my $scalar | |
823 | \my @array | |
824 | \my %hash | |
825 | \state $scalar # or @array, etc. | |
826 | \our $scalar # etc. | |
827 | \local $scalar # etc. | |
828 | \local our $scalar # etc. | |
829 | \$some_array[$index] | |
830 | \$some_hash{$key} | |
831 | \local $some_array[$index] | |
832 | \local $some_hash{$key} | |
833 | condition ? \$this : \$that[0] # etc. | |
834 | ||
df706e5b FC |
835 | Slicing operations and parentheses cause |
836 | the right-hand side to be evaluated in | |
e05542ee | 837 | list context: |
82848c10 | 838 | |
e05542ee FC |
839 | \@array[5..7] |
840 | (\@array[5..7]) | |
841 | \(@array[5..7]) | |
df706e5b FC |
842 | \@hash{'foo','bar'} |
843 | (\@hash{'foo','bar'}) | |
844 | \(@hash{'foo','bar'}) | |
82848c10 FC |
845 | (\$scalar) |
846 | \($scalar) | |
847 | \(my $scalar) | |
848 | \my($scalar) | |
849 | (\@array) | |
850 | (\%hash) | |
851 | (\&sub) | |
852 | \(&sub) | |
853 | \($foo, @bar, %baz) | |
854 | (\$foo, \@bar, \%baz) | |
855 | ||
856 | Each element on the right-hand side must be a reference to a datum of the | |
857 | right type. Parentheses immediately surrounding an array (and possibly | |
858 | also C<my>/C<state>/C<our>/C<local>) will make each element of the array an | |
859 | alias to the corresponding scalar referenced on the right-hand side: | |
860 | ||
861 | \(@a) = \(@b); # @a and @b now have the same elements | |
862 | \my(@a) = \(@b); # likewise | |
863 | \(my @a) = \(@b); # likewise | |
864 | push @a, 3; # but now @a has an extra element that @b lacks | |
865 | \(@a) = (\$a, \$b, \$c); # @a now contains $a, $b, and $c | |
866 | ||
867 | Combining that form with C<local> and putting parentheses immediately | |
868 | around a hash are forbidden (because it is not clear what they should do): | |
869 | ||
870 | \local(@array) = foo(); # WRONG | |
871 | \(%hash) = bar(); # wRONG | |
872 | ||
873 | Assignment to references and non-references may be combined in lists and | |
874 | conditional ternary expressions, as long as the values on the right-hand | |
875 | side are the right type for each element on the left, though this may make | |
876 | for obfuscated code: | |
877 | ||
878 | (my $tom, \my $dick, \my @harry) = (\1, \2, [1..3]); | |
879 | # $tom is now \1 | |
880 | # $dick is now 2 (read-only) | |
881 | # @harry is (1,2,3) | |
882 | ||
883 | my $type = ref $thingy; | |
74bfae27 | 884 | ($type ? $type eq 'ARRAY' ? \@foo : \$bar : $baz) = $thingy; |
82848c10 FC |
885 | |
886 | The C<foreach> loop can also take a reference constructor for its loop | |
887 | variable, though the syntax is limited to one of the following, with an | |
888 | optional C<my>, C<state>, or C<our> after the backslash: | |
889 | ||
890 | \$s | |
891 | \@a | |
892 | \%h | |
893 | \&c | |
894 | ||
895 | No parentheses are permitted. This feature is particularly useful for | |
896 | arrays-of-arrays, or arrays-of-hashes: | |
897 | ||
898 | foreach \my @a (@array_of_arrays) { | |
899 | frobnicate($a[0], $a[-1]); | |
900 | } | |
901 | ||
902 | foreach \my %h (@array_of_hashes) { | |
74bfae27 | 903 | $h{gelastic}++ if $h{type} eq 'funny'; |
82848c10 FC |
904 | } |
905 | ||
906 | B<CAVEAT:> Aliasing does not work correctly with closures. If you try to | |
907 | alias lexical variables from an inner subroutine or C<eval>, the aliasing | |
908 | will only be visible within that inner sub, and will not affect the outer | |
909 | subroutine where the variables are declared. This bizarre behavior is | |
910 | subject to change. | |
911 | ||
5c703779 FC |
912 | =head1 Declaring a Reference to a Variable |
913 | ||
d4062d50 FC |
914 | Beginning in v5.26.0, the referencing operator can come after C<my>, |
915 | C<state>, C<our>, or C<local>. This syntax must be enabled with C<use | |
916 | feature 'declared_refs'>. It is experimental, and will warn by default | |
917 | unless C<no warnings 'experimental::refaliasing'> is in effect. | |
5c703779 FC |
918 | |
919 | This feature makes these: | |
920 | ||
921 | my \$x; | |
922 | our \$y; | |
923 | ||
924 | equivalent to: | |
925 | ||
926 | \my $x; | |
927 | \our $x; | |
928 | ||
929 | It is intended mainly for use in assignments to references (see | |
930 | L</Assigning to References>, above). It also allows the backslash to be | |
931 | used on just some items in a list of declared variables: | |
932 | ||
933 | my ($foo, \@bar, \%baz); # equivalent to: my $foo, \my(@bar, %baz); | |
934 | ||
cb1a09d0 | 935 | =head1 SEE ALSO |
a0d0e21e LW |
936 | |
937 | Besides the obvious documents, source code can be instructive. | |
7c2ea1c7 | 938 | Some pathological examples of the use of references can be found |
a0d0e21e | 939 | in the F<t/op/ref.t> regression test in the Perl source directory. |
cb1a09d0 AD |
940 | |
941 | See also L<perldsc> and L<perllol> for how to use references to create | |
82e1c0d9 | 942 | complex data structures, and L<perlootut> and L<perlobj> |
5a964f20 | 943 | for how to use them to create objects. |