Commit | Line | Data |
a0d0e21e |
1 | =head1 NAME |
2 | |
3 | perlref - Perl references and nested data structures |
4 | |
5 | =head1 DESCRIPTION |
6 | |
cb1a09d0 |
7 | Before release 5 of Perl it was difficult to represent complex data |
5a964f20 |
8 | structures, because all references had to be symbolic--and even then |
9 | it was difficult to refer to a variable instead of a symbol table entry. |
10 | Perl now not only makes it easier to use symbolic references to variables, |
11 | but also lets you have "hard" references to any piece of data or code. |
12 | Any scalar may hold a hard reference. Because arrays and hashes contain |
13 | scalars, you can now easily build arrays of arrays, arrays of hashes, |
14 | hashes of arrays, arrays of hashes of functions, and so on. |
a0d0e21e |
15 | |
16 | Hard references are smart--they keep track of reference counts for you, |
2d24ed35 |
17 | automatically freeing the thing referred to when its reference count goes |
7b8d334a |
18 | to zero. (Note: the reference counts for values in self-referential or |
2d24ed35 |
19 | cyclic data structures may not go to zero without a little help; see |
7b8d334a |
20 | L<perlobj/"Two-Phased Garbage Collection"> for a detailed explanation.) |
2d24ed35 |
21 | If that thing happens to be an object, the object is destructed. See |
22 | L<perlobj> for more about objects. (In a sense, everything in Perl is an |
23 | object, but we usually reserve the word for references to objects that |
24 | have been officially "blessed" into a class package.) |
25 | |
26 | Symbolic references are names of variables or other objects, just as a |
54310121 |
27 | symbolic link in a Unix filesystem contains merely the name of a file. |
2d24ed35 |
28 | The C<*glob> notation is a kind of symbolic reference. (Symbolic |
29 | references are sometimes called "soft references", but please don't call |
30 | them that; references are confusing enough without useless synonyms.) |
31 | |
54310121 |
32 | In contrast, hard references are more like hard links in a Unix file |
2d24ed35 |
33 | system: They are used to access an underlying object without concern for |
34 | what its (other) name is. When the word "reference" is used without an |
5a964f20 |
35 | adjective, as in the following paragraph, it is usually talking about a |
2d24ed35 |
36 | hard reference. |
37 | |
38 | References are easy to use in Perl. There is just one overriding |
39 | principle: Perl does no implicit referencing or dereferencing. When a |
40 | scalar is holding a reference, it always behaves as a simple scalar. It |
41 | doesn't magically start being an array or hash or subroutine; you have to |
42 | tell it explicitly to do so, by dereferencing it. |
a0d0e21e |
43 | |
5a964f20 |
44 | =head2 Making References |
45 | |
46 | References can be created in several ways. |
a0d0e21e |
47 | |
48 | =over 4 |
49 | |
50 | =item 1. |
51 | |
52 | By using the backslash operator on a variable, subroutine, or value. |
54310121 |
53 | (This works much like the & (address-of) operator in C.) Note |
5f05dabc |
54 | that this typically creates I<ANOTHER> reference to a variable, because |
a0d0e21e |
55 | there's already a reference to the variable in the symbol table. But |
56 | the symbol table reference might go away, and you'll still have the |
57 | reference that the backslash returned. Here are some examples: |
58 | |
59 | $scalarref = \$foo; |
60 | $arrayref = \@ARGV; |
61 | $hashref = \%ENV; |
62 | $coderef = \&handler; |
55497cff |
63 | $globref = \*foo; |
cb1a09d0 |
64 | |
5a964f20 |
65 | It isn't possible to create a true reference to an IO handle (filehandle |
66 | or dirhandle) using the backslash operator. The most you can get is a |
67 | reference to a typeglob, which is actually a complete symbol table entry. |
68 | But see the explanation of the C<*foo{THING}> syntax below. However, |
69 | you can still use type globs and globrefs as though they were IO handles. |
a0d0e21e |
70 | |
71 | =item 2. |
72 | |
5a964f20 |
73 | A reference to an anonymous array can be created using square |
a0d0e21e |
74 | brackets: |
75 | |
76 | $arrayref = [1, 2, ['a', 'b', 'c']]; |
77 | |
5a964f20 |
78 | Here we've created a reference to an anonymous array of three elements |
54310121 |
79 | whose final element is itself a reference to another anonymous array of three |
a0d0e21e |
80 | elements. (The multidimensional syntax described later can be used to |
184e9718 |
81 | access this. For example, after the above, C<$arrayref-E<gt>[2][1]> would have |
a0d0e21e |
82 | the value "b".) |
83 | |
cb1a09d0 |
84 | Note that taking a reference to an enumerated list is not the same |
85 | as using square brackets--instead it's the same as creating |
86 | a list of references! |
87 | |
54310121 |
88 | @list = (\$a, \@b, \%c); |
58e0a6ae |
89 | @list = \($a, @b, %c); # same thing! |
90 | |
54310121 |
91 | As a special case, C<\(@foo)> returns a list of references to the contents |
58e0a6ae |
92 | of C<@foo>, not a reference to C<@foo> itself. Likewise for C<%foo>. |
cb1a09d0 |
93 | |
a0d0e21e |
94 | =item 3. |
95 | |
5a964f20 |
96 | A reference to an anonymous hash can be created using curly |
a0d0e21e |
97 | brackets: |
98 | |
99 | $hashref = { |
100 | 'Adam' => 'Eve', |
101 | 'Clyde' => 'Bonnie', |
102 | }; |
103 | |
5a964f20 |
104 | Anonymous hash and array composers like these can be intermixed freely to |
a0d0e21e |
105 | produce as complicated a structure as you want. The multidimensional |
106 | syntax described below works for these too. The values above are |
107 | literals, but variables and expressions would work just as well, because |
108 | assignment operators in Perl (even within local() or my()) are executable |
109 | statements, not compile-time declarations. |
110 | |
111 | Because curly brackets (braces) are used for several other things |
112 | including BLOCKs, you may occasionally have to disambiguate braces at the |
113 | beginning of a statement by putting a C<+> or a C<return> in front so |
114 | that Perl realizes the opening brace isn't starting a BLOCK. The economy and |
115 | mnemonic value of using curlies is deemed worth this occasional extra |
116 | hassle. |
117 | |
118 | For example, if you wanted a function to make a new hash and return a |
119 | reference to it, you have these options: |
120 | |
121 | sub hashem { { @_ } } # silently wrong |
122 | sub hashem { +{ @_ } } # ok |
123 | sub hashem { return { @_ } } # ok |
124 | |
ebc58f1a |
125 | On the other hand, if you want the other meaning, you can do this: |
126 | |
127 | sub showem { { @_ } } # ambiguous (currently ok, but may change) |
128 | sub showem { {; @_ } } # ok |
129 | sub showem { { return @_ } } # ok |
130 | |
131 | Note how the leading C<+{> and C<{;> always serve to disambiguate |
132 | the expression to mean either the HASH reference, or the BLOCK. |
133 | |
a0d0e21e |
134 | =item 4. |
135 | |
5a964f20 |
136 | A reference to an anonymous subroutine can be created by using |
a0d0e21e |
137 | C<sub> without a subname: |
138 | |
139 | $coderef = sub { print "Boink!\n" }; |
140 | |
141 | Note the presence of the semicolon. Except for the fact that the code |
142 | inside isn't executed immediately, a C<sub {}> is not so much a |
143 | declaration as it is an operator, like C<do{}> or C<eval{}>. (However, no |
5a964f20 |
144 | matter how many times you execute that particular line (unless you're in an |
a0d0e21e |
145 | C<eval("...")>), C<$coderef> will still have a reference to the I<SAME> |
146 | anonymous subroutine.) |
147 | |
748a9306 |
148 | Anonymous subroutines act as closures with respect to my() variables, |
149 | that is, variables visible lexically within the current scope. Closure |
150 | is a notion out of the Lisp world that says if you define an anonymous |
151 | function in a particular lexical context, it pretends to run in that |
152 | context even when it's called outside of the context. |
153 | |
154 | In human terms, it's a funny way of passing arguments to a subroutine when |
155 | you define it as well as when you call it. It's useful for setting up |
156 | little bits of code to run later, such as callbacks. You can even |
54310121 |
157 | do object-oriented stuff with it, though Perl already provides a different |
158 | mechanism to do that--see L<perlobj>. |
748a9306 |
159 | |
160 | You can also think of closure as a way to write a subroutine template without |
161 | using eval. (In fact, in version 5.000, eval was the I<only> way to get |
162 | closures. You may wish to use "require 5.001" if you use closures.) |
163 | |
164 | Here's a small example of how closures works: |
165 | |
166 | sub newprint { |
167 | my $x = shift; |
168 | return sub { my $y = shift; print "$x, $y!\n"; }; |
a0d0e21e |
169 | } |
748a9306 |
170 | $h = newprint("Howdy"); |
171 | $g = newprint("Greetings"); |
172 | |
173 | # Time passes... |
174 | |
175 | &$h("world"); |
176 | &$g("earthlings"); |
a0d0e21e |
177 | |
748a9306 |
178 | This prints |
179 | |
180 | Howdy, world! |
181 | Greetings, earthlings! |
182 | |
183 | Note particularly that $x continues to refer to the value passed into |
cb1a09d0 |
184 | newprint() I<despite> the fact that the "my $x" has seemingly gone out of |
748a9306 |
185 | scope by the time the anonymous subroutine runs. That's what closure |
186 | is all about. |
187 | |
5a964f20 |
188 | This applies only to lexical variables, by the way. Dynamic variables |
748a9306 |
189 | continue to work as they have always worked. Closure is not something |
190 | that most Perl programmers need trouble themselves about to begin with. |
a0d0e21e |
191 | |
192 | =item 5. |
193 | |
194 | References are often returned by special subroutines called constructors. |
748a9306 |
195 | Perl objects are just references to a special kind of object that happens to know |
a0d0e21e |
196 | which package it's associated with. Constructors are just special |
197 | subroutines that know how to create that association. They do so by |
198 | starting with an ordinary reference, and it remains an ordinary reference |
5a964f20 |
199 | even while it's also being an object. Constructors are often |
200 | named new() and called indirectly: |
a0d0e21e |
201 | |
202 | $objref = new Doggie (Tail => 'short', Ears => 'long'); |
203 | |
5a964f20 |
204 | But don't have to be: |
205 | |
206 | $objref = Doggie->new(Tail => 'short', Ears => 'long'); |
207 | |
208 | use Term::Cap; |
209 | $terminal = Term::Cap->Tgetent( { OSPEED => 9600 }); |
210 | |
211 | use Tk; |
212 | $main = MainWindow->new(); |
213 | $menubar = $main->Frame(-relief => "raised", |
214 | -borderwidth => 2) |
215 | |
a0d0e21e |
216 | =item 6. |
217 | |
218 | References of the appropriate type can spring into existence if you |
5f05dabc |
219 | dereference them in a context that assumes they exist. Because we haven't |
a0d0e21e |
220 | talked about dereferencing yet, we can't show you any examples yet. |
221 | |
cb1a09d0 |
222 | =item 7. |
223 | |
55497cff |
224 | A reference can be created by using a special syntax, lovingly known as |
225 | the *foo{THING} syntax. *foo{THING} returns a reference to the THING |
226 | slot in *foo (which is the symbol table entry which holds everything |
227 | known as foo). |
cb1a09d0 |
228 | |
55497cff |
229 | $scalarref = *foo{SCALAR}; |
230 | $arrayref = *ARGV{ARRAY}; |
231 | $hashref = *ENV{HASH}; |
232 | $coderef = *handler{CODE}; |
36477c24 |
233 | $ioref = *STDIN{IO}; |
55497cff |
234 | $globref = *foo{GLOB}; |
235 | |
36477c24 |
236 | All of these are self-explanatory except for *foo{IO}. It returns the |
237 | IO handle, used for file handles (L<perlfunc/open>), sockets |
238 | (L<perlfunc/socket> and L<perlfunc/socketpair>), and directory handles |
239 | (L<perlfunc/opendir>). For compatibility with previous versions of |
240 | Perl, *foo{FILEHANDLE} is a synonym for *foo{IO}. |
55497cff |
241 | |
5f05dabc |
242 | *foo{THING} returns undef if that particular THING hasn't been used yet, |
243 | except in the case of scalars. *foo{SCALAR} returns a reference to an |
244 | anonymous scalar if $foo hasn't been used yet. This might change in a |
245 | future release. |
246 | |
5a964f20 |
247 | *foo{IO} is an alternative to the \*HANDLE mechanism given in |
248 | L<perldata/"Typeglobs and Filehandles"> for passing filehandles |
249 | into or out of subroutines, or storing into larger data structures. |
250 | Its disadvantage is that it won't create a new filehandle for you. |
251 | Its advantage is that you have no risk of clobbering more than you want |
252 | to with a typeglob assignment, although if you assign to a scalar instead |
253 | of a typeglob, you're ok. |
36477c24 |
254 | |
5a964f20 |
255 | splutter(*STDOUT); |
36477c24 |
256 | splutter(*STDOUT{IO}); |
5a964f20 |
257 | |
cb1a09d0 |
258 | sub splutter { |
259 | my $fh = shift; |
260 | print $fh "her um well a hmmm\n"; |
261 | } |
262 | |
5a964f20 |
263 | $rec = get_rec(*STDIN); |
36477c24 |
264 | $rec = get_rec(*STDIN{IO}); |
5a964f20 |
265 | |
cb1a09d0 |
266 | sub get_rec { |
267 | my $fh = shift; |
268 | return scalar <$fh>; |
269 | } |
270 | |
a0d0e21e |
271 | =back |
272 | |
5a964f20 |
273 | =head2 Using References |
274 | |
a0d0e21e |
275 | That's it for creating references. By now you're probably dying to |
276 | know how to use references to get back to your long-lost data. There |
277 | are several basic methods. |
278 | |
279 | =over 4 |
280 | |
281 | =item 1. |
282 | |
6309d9d9 |
283 | Anywhere you'd put an identifier (or chain of identifiers) as part |
284 | of a variable or subroutine name, you can replace the identifier with |
285 | a simple scalar variable containing a reference of the correct type: |
a0d0e21e |
286 | |
287 | $bar = $$scalarref; |
288 | push(@$arrayref, $filename); |
289 | $$arrayref[0] = "January"; |
290 | $$hashref{"KEY"} = "VALUE"; |
291 | &$coderef(1,2,3); |
cb1a09d0 |
292 | print $globref "output\n"; |
a0d0e21e |
293 | |
294 | It's important to understand that we are specifically I<NOT> dereferencing |
295 | C<$arrayref[0]> or C<$hashref{"KEY"}> there. The dereference of the |
296 | scalar variable happens I<BEFORE> it does any key lookups. Anything more |
297 | complicated than a simple scalar variable must use methods 2 or 3 below. |
298 | However, a "simple scalar" includes an identifier that itself uses method |
299 | 1 recursively. Therefore, the following prints "howdy". |
300 | |
301 | $refrefref = \\\"howdy"; |
302 | print $$$$refrefref; |
303 | |
304 | =item 2. |
305 | |
6309d9d9 |
306 | Anywhere you'd put an identifier (or chain of identifiers) as part of a |
307 | variable or subroutine name, you can replace the identifier with a |
308 | BLOCK returning a reference of the correct type. In other words, the |
309 | previous examples could be written like this: |
a0d0e21e |
310 | |
311 | $bar = ${$scalarref}; |
312 | push(@{$arrayref}, $filename); |
313 | ${$arrayref}[0] = "January"; |
314 | ${$hashref}{"KEY"} = "VALUE"; |
315 | &{$coderef}(1,2,3); |
36477c24 |
316 | $globref->print("output\n"); # iff IO::Handle is loaded |
a0d0e21e |
317 | |
318 | Admittedly, it's a little silly to use the curlies in this case, but |
319 | the BLOCK can contain any arbitrary expression, in particular, |
320 | subscripted expressions: |
321 | |
54310121 |
322 | &{ $dispatch{$index} }(1,2,3); # call correct routine |
a0d0e21e |
323 | |
324 | Because of being able to omit the curlies for the simple case of C<$$x>, |
325 | people often make the mistake of viewing the dereferencing symbols as |
326 | proper operators, and wonder about their precedence. If they were, |
5f05dabc |
327 | though, you could use parentheses instead of braces. That's not the case. |
a0d0e21e |
328 | Consider the difference below; case 0 is a short-hand version of case 1, |
329 | I<NOT> case 2: |
330 | |
331 | $$hashref{"KEY"} = "VALUE"; # CASE 0 |
332 | ${$hashref}{"KEY"} = "VALUE"; # CASE 1 |
333 | ${$hashref{"KEY"}} = "VALUE"; # CASE 2 |
334 | ${$hashref->{"KEY"}} = "VALUE"; # CASE 3 |
335 | |
336 | Case 2 is also deceptive in that you're accessing a variable |
337 | called %hashref, not dereferencing through $hashref to the hash |
338 | it's presumably referencing. That would be case 3. |
339 | |
340 | =item 3. |
341 | |
6da72b64 |
342 | Subroutine calls and lookups of individual array elements arise often |
343 | enough that it gets cumbersome to use method 2. As a form of |
344 | syntactic sugar, the examples for method 2 may be written: |
a0d0e21e |
345 | |
6da72b64 |
346 | $arrayref->[0] = "January"; # Array element |
347 | $hashref->{"KEY"} = "VALUE"; # Hash element |
348 | $coderef->(1,2,3); # Subroutine call |
a0d0e21e |
349 | |
6da72b64 |
350 | The left side of the arrow can be any expression returning a reference, |
a0d0e21e |
351 | including a previous dereference. Note that C<$array[$x]> is I<NOT> the |
352 | same thing as C<$array-E<gt>[$x]> here: |
353 | |
354 | $array[$x]->{"foo"}->[0] = "January"; |
355 | |
356 | This is one of the cases we mentioned earlier in which references could |
357 | spring into existence when in an lvalue context. Before this |
358 | statement, C<$array[$x]> may have been undefined. If so, it's |
359 | automatically defined with a hash reference so that we can look up |
360 | C<{"foo"}> in it. Likewise C<$array[$x]-E<gt>{"foo"}> will automatically get |
361 | defined with an array reference so that we can look up C<[0]> in it. |
5a964f20 |
362 | This process is called I<autovivification>. |
a0d0e21e |
363 | |
364 | One more thing here. The arrow is optional I<BETWEEN> brackets |
365 | subscripts, so you can shrink the above down to |
366 | |
367 | $array[$x]{"foo"}[0] = "January"; |
368 | |
369 | Which, in the degenerate case of using only ordinary arrays, gives you |
370 | multidimensional arrays just like C's: |
371 | |
372 | $score[$x][$y][$z] += 42; |
373 | |
374 | Well, okay, not entirely like C's arrays, actually. C doesn't know how |
375 | to grow its arrays on demand. Perl does. |
376 | |
377 | =item 4. |
378 | |
379 | If a reference happens to be a reference to an object, then there are |
380 | probably methods to access the things referred to, and you should probably |
381 | stick to those methods unless you're in the class package that defines the |
382 | object's methods. In other words, be nice, and don't violate the object's |
383 | encapsulation without a very good reason. Perl does not enforce |
384 | encapsulation. We are not totalitarians here. We do expect some basic |
385 | civility though. |
386 | |
387 | =back |
388 | |
389 | The ref() operator may be used to determine what type of thing the |
390 | reference is pointing to. See L<perlfunc>. |
391 | |
5a964f20 |
392 | The bless() operator may be used to associate the object a reference |
393 | points to with a package functioning as an object class. See L<perlobj>. |
a0d0e21e |
394 | |
5f05dabc |
395 | A typeglob may be dereferenced the same way a reference can, because |
a0d0e21e |
396 | the dereference syntax always indicates the kind of reference desired. |
397 | So C<${*foo}> and C<${\$foo}> both indicate the same scalar variable. |
398 | |
399 | Here's a trick for interpolating a subroutine call into a string: |
400 | |
cb1a09d0 |
401 | print "My sub returned @{[mysub(1,2,3)]} that time.\n"; |
402 | |
403 | The way it works is that when the C<@{...}> is seen in the double-quoted |
404 | string, it's evaluated as a block. The block creates a reference to an |
405 | anonymous array containing the results of the call to C<mysub(1,2,3)>. So |
406 | the whole block returns a reference to an array, which is then |
407 | dereferenced by C<@{...}> and stuck into the double-quoted string. This |
408 | chicanery is also useful for arbitrary expressions: |
a0d0e21e |
409 | |
184e9718 |
410 | print "That yields @{[$n + 5]} widgets\n"; |
a0d0e21e |
411 | |
412 | =head2 Symbolic references |
413 | |
414 | We said that references spring into existence as necessary if they are |
415 | undefined, but we didn't say what happens if a value used as a |
416 | reference is already defined, but I<ISN'T> a hard reference. If you |
417 | use it as a reference in this case, it'll be treated as a symbolic |
418 | reference. That is, the value of the scalar is taken to be the I<NAME> |
419 | of a variable, rather than a direct link to a (possibly) anonymous |
420 | value. |
421 | |
422 | People frequently expect it to work like this. So it does. |
423 | |
424 | $name = "foo"; |
425 | $$name = 1; # Sets $foo |
426 | ${$name} = 2; # Sets $foo |
427 | ${$name x 2} = 3; # Sets $foofoo |
428 | $name->[0] = 4; # Sets $foo[0] |
429 | @$name = (); # Clears @foo |
430 | &$name(); # Calls &foo() (as in Perl 4) |
431 | $pack = "THAT"; |
432 | ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval |
433 | |
434 | This is very powerful, and slightly dangerous, in that it's possible |
435 | to intend (with the utmost sincerity) to use a hard reference, and |
436 | accidentally use a symbolic reference instead. To protect against |
437 | that, you can say |
438 | |
439 | use strict 'refs'; |
440 | |
441 | and then only hard references will be allowed for the rest of the enclosing |
54310121 |
442 | block. An inner block may countermand that with |
a0d0e21e |
443 | |
444 | no strict 'refs'; |
445 | |
5a964f20 |
446 | Only package variables (globals, even if localized) are visible to |
447 | symbolic references. Lexical variables (declared with my()) aren't in |
448 | a symbol table, and thus are invisible to this mechanism. For example: |
a0d0e21e |
449 | |
5a964f20 |
450 | local $value = 10; |
a0d0e21e |
451 | $ref = \$value; |
452 | { |
453 | my $value = 20; |
454 | print $$ref; |
54310121 |
455 | } |
a0d0e21e |
456 | |
457 | This will still print 10, not 20. Remember that local() affects package |
458 | variables, which are all "global" to the package. |
459 | |
748a9306 |
460 | =head2 Not-so-symbolic references |
461 | |
a6006777 |
462 | A new feature contributing to readability in perl version 5.001 is that the |
463 | brackets around a symbolic reference behave more like quotes, just as they |
748a9306 |
464 | always have within a string. That is, |
465 | |
466 | $push = "pop on "; |
467 | print "${push}over"; |
468 | |
469 | has always meant to print "pop on over", despite the fact that push is |
470 | a reserved word. This has been generalized to work the same outside |
471 | of quotes, so that |
472 | |
473 | print ${push} . "over"; |
474 | |
475 | and even |
476 | |
477 | print ${ push } . "over"; |
478 | |
479 | will have the same effect. (This would have been a syntax error in |
a6006777 |
480 | Perl 5.000, though Perl 4 allowed it in the spaceless form.) Note that this |
748a9306 |
481 | construct is I<not> considered to be a symbolic reference when you're |
482 | using strict refs: |
483 | |
484 | use strict 'refs'; |
485 | ${ bareword }; # Okay, means $bareword. |
486 | ${ "bareword" }; # Error, symbolic reference. |
487 | |
488 | Similarly, because of all the subscripting that is done using single |
489 | words, we've applied the same rule to any bareword that is used for |
490 | subscripting a hash. So now, instead of writing |
491 | |
492 | $array{ "aaa" }{ "bbb" }{ "ccc" } |
493 | |
5f05dabc |
494 | you can write just |
748a9306 |
495 | |
496 | $array{ aaa }{ bbb }{ ccc } |
497 | |
498 | and not worry about whether the subscripts are reserved words. In the |
499 | rare event that you do wish to do something like |
500 | |
501 | $array{ shift } |
502 | |
503 | you can force interpretation as a reserved word by adding anything that |
504 | makes it more than a bareword: |
505 | |
506 | $array{ shift() } |
507 | $array{ +shift } |
508 | $array{ shift @_ } |
509 | |
510 | The B<-w> switch will warn you if it interprets a reserved word as a string. |
5f05dabc |
511 | But it will no longer warn you about using lowercase words, because the |
748a9306 |
512 | string is effectively quoted. |
513 | |
49399b3f |
514 | =head2 Pseudo-hashes: Using an array as a hash |
515 | |
516 | WARNING: This section describes an experimental feature. Details may |
517 | change without notice in future versions. |
518 | |
519 | Beginning with release 5.005 of Perl you can use an array reference |
520 | in some contexts that would normally require a hash reference. This |
521 | allows you to access array elements using symbolic names, as if they |
522 | were fields in a structure. |
523 | |
524 | For this to work, the array must contain extra information. The first |
525 | element of the array has to be a hash reference that maps field names |
526 | to array indices. Here is an example: |
527 | |
528 | $struct = [{foo => 1, bar => 2}, "FOO", "BAR"]; |
529 | |
530 | $struct->{foo}; # same as $struct->[1], i.e. "FOO" |
531 | $struct->{bar}; # same as $struct->[2], i.e. "BAR" |
532 | |
533 | keys %$struct; # will return ("foo", "bar") in some order |
534 | values %$struct; # will return ("FOO", "BAR") in same some order |
535 | |
536 | while (my($k,$v) = each %$struct) { |
537 | print "$k => $v\n"; |
538 | } |
539 | |
540 | Perl will raise an exception if you try to delete keys from a pseudo-hash |
541 | or try to access nonexistent fields. For better performance, Perl can also |
542 | do the translation from field names to array indices at compile time for |
543 | typed object references. See L<fields>. |
544 | |
545 | |
5a964f20 |
546 | =head2 Function Templates |
547 | |
548 | As explained above, a closure is an anonymous function with access to the |
549 | lexical variables visible when that function was compiled. It retains |
550 | access to those variables even though it doesn't get run until later, |
551 | such as in a signal handler or a Tk callback. |
552 | |
553 | Using a closure as a function template allows us to generate many functions |
554 | that act similarly. Suppopose you wanted functions named after the colors |
555 | that generated HTML font changes for the various colors: |
556 | |
557 | print "Be ", red("careful"), "with that ", green("light"); |
558 | |
559 | The red() and green() functions would be very similar. To create these, |
560 | we'll assign a closure to a typeglob of the name of the function we're |
561 | trying to build. |
562 | |
563 | @colors = qw(red blue green yellow orange purple violet); |
564 | for my $name (@colors) { |
565 | no strict 'refs'; # allow symbol table manipulation |
566 | *$name = *{uc $name} = sub { "<FONT COLOR='$name'>@_</FONT>" }; |
567 | } |
568 | |
569 | Now all those different functions appear to exist independently. You can |
570 | call red(), RED(), blue(), BLUE(), green(), etc. This technique saves on |
571 | both compile time and memory use, and is less error-prone as well, since |
572 | syntax checks happen at compile time. It's critical that any variables in |
573 | the anonymous subroutine be lexicals in order to create a proper closure. |
574 | That's the reasons for the C<my> on the loop iteration variable. |
575 | |
576 | This is one of the only places where giving a prototype to a closure makes |
577 | much sense. If you wanted to impose scalar context on the arguments of |
578 | these functions (probably not a wise idea for this particular example), |
579 | you could have written it this way instead: |
580 | |
581 | *$name = sub ($) { "<FONT COLOR='$name'>$_[0]</FONT>" }; |
582 | |
583 | However, since prototype checking happens at compile time, the assignment |
584 | above happens too late to be of much use. You could address this by |
585 | putting the whole loop of assignments within a BEGIN block, forcing it |
586 | to occur during compilation. |
587 | |
588 | Access to lexicals that change over type--like those in the C<for> loop |
589 | above--only works with closures, not general subroutines. In the general |
590 | case, then, named subroutines do not nest properly, although anonymous |
591 | ones do. If you are accustomed to using nested subroutines in other |
592 | programming languages with their own private variables, you'll have to |
593 | work at it a bit in Perl. The intuitive coding of this kind of thing |
594 | incurs mysterious warnings about ``will not stay shared''. For example, |
595 | this won't work: |
596 | |
597 | sub outer { |
598 | my $x = $_[0] + 35; |
599 | sub inner { return $x * 19 } # WRONG |
600 | return $x + inner(); |
601 | } |
602 | |
603 | A work-around is the following: |
604 | |
605 | sub outer { |
606 | my $x = $_[0] + 35; |
607 | local *inner = sub { return $x * 19 }; |
608 | return $x + inner(); |
609 | } |
610 | |
611 | Now inner() can only be called from within outer(), because of the |
612 | temporary assignments of the closure (anonymous subroutine). But when |
613 | it does, it has normal access to the lexical variable $x from the scope |
614 | of outer(). |
615 | |
616 | This has the interesting effect of creating a function local to another |
617 | function, something not normally supported in Perl. |
618 | |
cb1a09d0 |
619 | =head1 WARNING |
748a9306 |
620 | |
621 | You may not (usefully) use a reference as the key to a hash. It will be |
622 | converted into a string: |
623 | |
624 | $x{ \$a } = $a; |
625 | |
54310121 |
626 | If you try to dereference the key, it won't do a hard dereference, and |
184e9718 |
627 | you won't accomplish what you're attempting. You might want to do something |
cb1a09d0 |
628 | more like |
748a9306 |
629 | |
cb1a09d0 |
630 | $r = \@a; |
631 | $x{ $r } = $r; |
632 | |
633 | And then at least you can use the values(), which will be |
634 | real refs, instead of the keys(), which won't. |
635 | |
5a964f20 |
636 | The standard Tie::RefHash module provides a convenient workaround to this. |
637 | |
cb1a09d0 |
638 | =head1 SEE ALSO |
a0d0e21e |
639 | |
640 | Besides the obvious documents, source code can be instructive. |
641 | Some rather pathological examples of the use of references can be found |
642 | in the F<t/op/ref.t> regression test in the Perl source directory. |
cb1a09d0 |
643 | |
644 | See also L<perldsc> and L<perllol> for how to use references to create |
5a964f20 |
645 | complex data structures, and L<perltoot>, L<perlobj>, and L<perlbot> |
646 | for how to use them to create objects. |