Debian Codegraph @

upstream/debian-unstable/man1/perlref.1

raw
1 .\" -*- mode: troff; coding: utf-8 -*-
2 .\" Automatically generated by Pod::Man 5.0102 (Pod::Simple 3.45)
3 .\"
4 .\" Standard preamble:
5 .\" ========================================================================
6 .de Sp \" Vertical space (when we can't use .PP)
7 .if t .sp .5v
8 .if n .sp
9 ..
10 .de Vb \" Begin verbatim text
11 .ft CW
12 .nf
13 .ne \\$1
14 ..
15 .de Ve \" End verbatim text
16 .ft R
17 .fi
18 ..
19 .\" \*(C` and \*(C' are quotes in nroff, nothing in troff, for use with C<>.
20 .ie n \{\
21 . ds C` ""
22 . ds C' ""
23 'br\}
24 .el\{\
25 . ds C`
26 . ds C'
27 'br\}
28 .\"
29 .\" Escape single quotes in literal strings from groff's Unicode transform.
30 .ie \n(.g .ds Aq \(aq
31 .el .ds Aq '
32 .\"
33 .\" If the F register is >0, we'll generate index entries on stderr for
34 .\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index
35 .\" entries marked with X<> in POD. Of course, you'll have to process the
36 .\" output yourself in some meaningful fashion.
37 .\"
38 .\" Avoid warning from groff about undefined register 'F'.
39 .de IX
40 ..
41 .nr rF 0
42 .if \n(.g .if rF .nr rF 1
43 .if (\n(rF:(\n(.g==0)) \{\
44 . if \nF \{\
45 . de IX
46 . tm Index:\\$1\t\\n%\t"\\$2"
47 ..
48 . if !\nF==2 \{\
49 . nr % 0
50 . nr F 2
51 . \}
52 . \}
53 .\}
54 .rr rF
55 .\" ========================================================================
56 .\"
57 .IX Title "PERLREF 1"
58 .TH PERLREF 1 2025-11-16 "perl v5.40.1" "Perl Programmers Reference Guide"
59 .\" For nroff, turn off justification. Always turn off hyphenation; it makes
60 .\" way too many mistakes in technical documents.
61 .if n .ad l
62 .nh
63 .SH NAME
64 perlref \- Perl references and nested data structures
65 .IX Xref "reference pointer data structure structure struct"
66 .SH NOTE
67 .IX Header "NOTE"
68 This is complete documentation about all aspects of references.
69 For a shorter, tutorial introduction to just the essential features,
70 see perlreftut.
71 .SH DESCRIPTION
72 .IX Header "DESCRIPTION"
73 Before release 5 of Perl it was difficult to represent complex data
74 structures, because all references had to be symbolic\-\-and even then
75 it was difficult to refer to a variable instead of a symbol table entry.
76 Perl now not only makes it easier to use symbolic references to variables,
77 but also lets you have "hard" references to any piece of data or code.
78 Any scalar may hold a hard reference. Because arrays and hashes contain
79 scalars, you can now easily build arrays of arrays, arrays of hashes,
80 hashes of arrays, arrays of hashes of functions, and so on.
81 .PP
82 Hard references are smart\-\-they keep track of reference counts for you,
83 automatically freeing the thing referred to when its reference count goes
84 to zero. (Reference counts for values in self-referential or
85 cyclic data structures may not go to zero without a little help; see
86 "Circular References" for a detailed explanation.)
87 If that thing happens to be an object, the object is destructed. See
88 perlobj for more about objects. (In a sense, everything in Perl is an
89 object, but we usually reserve the word for references to objects that
90 have been officially "blessed" into a class package.)
91 .PP
92 Symbolic references are names of variables or other objects, just as a
93 symbolic link in a Unix filesystem contains merely the name of a file.
94 The \f(CW*glob\fR notation is something of a symbolic reference. (Symbolic
95 references are sometimes called "soft references", but please don't call
96 them that; references are confusing enough without useless synonyms.)
97 .IX Xref "reference, symbolic reference, soft symbolic reference soft reference"
98 .PP
99 In contrast, hard references are more like hard links in a Unix file
100 system: They are used to access an underlying object without concern for
101 what its (other) name is. When the word "reference" is used without an
102 adjective, as in the following paragraph, it is usually talking about a
103 hard reference.
104 .IX Xref "reference, hard hard reference"
105 .PP
106 References are easy to use in Perl. There is just one overriding
107 principle: in general, Perl does no implicit referencing or dereferencing.
108 When a scalar is holding a reference, it always behaves as a simple scalar.
109 It doesn't magically start being an array or hash or subroutine; you have to
110 tell it explicitly to do so, by dereferencing it.
111 .SS "Making References"
112 .IX Xref "reference, creation referencing"
113 .IX Subsection "Making References"
114 References can be created in several ways.
115 .PP
116 \fIBackslash Operator\fR
117 .IX Xref "\\ backslash"
118 .IX Subsection "Backslash Operator"
119 .PP
120 By using the backslash operator on a variable, subroutine, or value.
121 (This works much like the & (address-of) operator in C.)
122 This typically creates \fIanother\fR reference to a variable, because
123 there's already a reference to the variable in the symbol table. But
124 the symbol table reference might go away, and you'll still have the
125 reference that the backslash returned. Here are some examples:
126 .PP
127 .Vb 5
128 \& $scalarref = \e$foo;
129 \& $arrayref = \e@ARGV;
130 \& $hashref = \e%ENV;
131 \& $coderef = \e&handler;
132 \& $globref = \e*foo;
133 .Ve
134 .PP
135 It isn't possible to create a true reference to an IO handle (filehandle
136 or dirhandle) using the backslash operator. The most you can get is a
137 reference to a typeglob, which is actually a complete symbol table entry.
138 But see the explanation of the \f(CW*foo{THING}\fR syntax below. However,
139 you can still use type globs and globrefs as though they were IO handles.
140 .PP
141 \fISquare Brackets\fR
142 .IX Xref "array, anonymous [ [] square bracket bracket, square arrayref array reference reference, array"
143 .IX Subsection "Square Brackets"
144 .PP
145 A reference to an anonymous array can be created using square
146 brackets:
147 .PP
148 .Vb 1
149 \& $arrayref = [1, 2, [\*(Aqa\*(Aq, \*(Aqb\*(Aq, \*(Aqc\*(Aq]];
150 .Ve
151 .PP
152 Here we've created a reference to an anonymous array of three elements
153 whose final element is itself a reference to another anonymous array of three
154 elements. (The multidimensional syntax described later can be used to
155 access this. For example, after the above, \f(CW\*(C`$arrayref\->[2][1]\*(C'\fR would have
156 the value "b".)
157 .PP
158 Taking a reference to an enumerated list is not the same
159 as using square brackets\-\-instead it's the same as creating
160 a list of references!
161 .PP
162 .Vb 2
163 \& @list = (\e$x, \e@y, \e%z);
164 \& @list = \e($x, @y, %z); # same thing!
165 .Ve
166 .PP
167 As a special case, \f(CW\*(C`\e(@foo)\*(C'\fR returns a list of references to the contents
168 of \f(CW@foo\fR, not a reference to \f(CW@foo\fR itself. Likewise for \f(CW%foo\fR,
169 except that the key references are to copies (since the keys are just
170 strings rather than full-fledged scalars).
171 .PP
172 \fICurly Brackets\fR
173 .IX Xref "hash, anonymous { {} curly bracket bracket, curly brace hashref hash reference reference, hash"
174 .IX Subsection "Curly Brackets"
175 .PP
176 A reference to an anonymous hash can be created using curly
177 brackets:
178 .PP
179 .Vb 4
180 \& $hashref = {
181 \& \*(AqAdam\*(Aq => \*(AqEve\*(Aq,
182 \& \*(AqClyde\*(Aq => \*(AqBonnie\*(Aq,
183 \& };
184 .Ve
185 .PP
186 Anonymous hash and array composers like these can be intermixed freely to
187 produce as complicated a structure as you want. The multidimensional
188 syntax described below works for these too. The values above are
189 literals, but variables and expressions would work just as well, because
190 assignment operators in Perl (even within \fBlocal()\fR or \fBmy()\fR) are executable
191 statements, not compile-time declarations.
192 .PP
193 Because curly brackets (braces) are used for several other things
194 including BLOCKs, you may occasionally have to disambiguate braces at the
195 beginning of a statement by putting a \f(CW\*(C`+\*(C'\fR or a \f(CW\*(C`return\*(C'\fR in front so
196 that Perl realizes the opening brace isn't starting a BLOCK. The economy and
197 mnemonic value of using curlies is deemed worth this occasional extra
198 hassle.
199 .PP
200 For example, if you wanted a function to make a new hash and return a
201 reference to it, you have these options:
202 .PP
203 .Vb 3
204 \& sub hashem { { @_ } } # silently wrong
205 \& sub hashem { +{ @_ } } # ok
206 \& sub hashem { return { @_ } } # ok
207 .Ve
208 .PP
209 On the other hand, if you want the other meaning, you can do this:
210 .PP
211 .Vb 4
212 \& sub showem { { @_ } } # ambiguous (currently ok,
213 \& # but may change)
214 \& sub showem { {; @_ } } # ok
215 \& sub showem { { return @_ } } # ok
216 .Ve
217 .PP
218 The leading \f(CW\*(C`+{\*(C'\fR and \f(CW\*(C`{;\*(C'\fR always serve to disambiguate
219 the expression to mean either the HASH reference, or the BLOCK.
220 .PP
221 \fIAnonymous Subroutines\fR
222 .IX Xref "subroutine, anonymous subroutine, reference reference, subroutine scope, lexical closure lexical lexical scope"
223 .IX Subsection "Anonymous Subroutines"
224 .PP
225 A reference to an anonymous subroutine can be created by using
226 \&\f(CW\*(C`sub\*(C'\fR without a subname:
227 .PP
228 .Vb 1
229 \& $coderef = sub { print "Boink!\en" };
230 .Ve
231 .PP
232 Note the semicolon. Except for the code
233 inside not being immediately executed, a \f(CW\*(C`sub {}\*(C'\fR is not so much a
234 declaration as it is an operator, like \f(CW\*(C`do{}\*(C'\fR or \f(CW\*(C`eval{}\*(C'\fR. (However, no
235 matter how many times you execute that particular line (unless you're in an
236 \&\f(CWeval("...")\fR), \f(CW$coderef\fR will still have a reference to the \fIsame\fR
237 anonymous subroutine.)
238 .PP
239 Anonymous subroutines act as closures with respect to \fBmy()\fR variables,
240 that is, variables lexically visible within the current scope. Closure
241 is a notion out of the Lisp world that says if you define an anonymous
242 function in a particular lexical context, it pretends to run in that
243 context even when it's called outside the context.
244 .PP
245 In human terms, it's a funny way of passing arguments to a subroutine when
246 you define it as well as when you call it. It's useful for setting up
247 little bits of code to run later, such as callbacks. You can even
248 do object-oriented stuff with it, though Perl already provides a different
249 mechanism to do that\-\-see perlobj.
250 .PP
251 You might also think of closure as a way to write a subroutine
252 template without using \fBeval()\fR. Here's a small example of how
253 closures work:
254 .PP
255 .Vb 6
256 \& sub newprint {
257 \& my $x = shift;
258 \& return sub { my $y = shift; print "$x, $y!\en"; };
259 \& }
260 \& $h = newprint("Howdy");
261 \& $g = newprint("Greetings");
262 \&
263 \& # Time passes...
264 \&
265 \& &$h("world");
266 \& &$g("earthlings");
267 .Ve
268 .PP
269 This prints
270 .PP
271 .Vb 2
272 \& Howdy, world!
273 \& Greetings, earthlings!
274 .Ve
275 .PP
276 Note particularly that \f(CW$x\fR continues to refer to the value passed
277 into \fBnewprint()\fR \fIdespite\fR "my \f(CW$x\fR" having gone out of scope by the
278 time the anonymous subroutine runs. That's what a closure is all
279 about.
280 .PP
281 This applies only to lexical variables, by the way. Dynamic variables
282 continue to work as they have always worked. Closure is not something
283 that most Perl programmers need trouble themselves about to begin with.
284 .PP
285 \fIConstructors\fR
286 .IX Xref "constructor new"
287 .IX Subsection "Constructors"
288 .PP
289 References are often returned by special subroutines called constructors. Perl
290 objects are just references to a special type of object that happens to know
291 which package it's associated with. Constructors are just special subroutines
292 that know how to create that association. They do so by starting with an
293 ordinary reference, and it remains an ordinary reference even while it's also
294 being an object. Constructors are often named \f(CWnew()\fR. You \fIcan\fR call them
295 indirectly:
296 .PP
297 .Vb 1
298 \& $objref = new Doggie( Tail => \*(Aqshort\*(Aq, Ears => \*(Aqlong\*(Aq );
299 .Ve
300 .PP
301 But that can produce ambiguous syntax in certain cases, so it's often
302 better to use the direct method invocation approach:
303 .PP
304 .Vb 1
305 \& $objref = Doggie\->new(Tail => \*(Aqshort\*(Aq, Ears => \*(Aqlong\*(Aq);
306 \&
307 \& use Term::Cap;
308 \& $terminal = Term::Cap\->Tgetent( { OSPEED => 9600 });
309 \&
310 \& use Tk;
311 \& $main = MainWindow\->new();
312 \& $menubar = $main\->Frame(\-relief => "raised",
313 \& \-borderwidth => 2)
314 .Ve
315 .PP
316 This indirect object syntax is only available when
317 \&\f(CW\*(C`use feature "indirect"\*(C'\fR is in effect,
318 and that is not the case when \f(CW\*(C`use v5.36\*(C'\fR (or
319 higher) is requested, it is best to avoid indirect object syntax entirely.
320 .PP
321 \fIAutovivification\fR
322 .IX Xref "autovivification"
323 .IX Subsection "Autovivification"
324 .PP
325 References of the appropriate type can spring into existence if you
326 dereference them in a context that assumes they exist. Because we haven't
327 talked about dereferencing yet, we can't show you any examples yet.
328 .PP
329 \fITypeglob Slots\fR
330 .IX Xref "*foo{THING} *"
331 .IX Subsection "Typeglob Slots"
332 .PP
333 A reference can be created by using a special syntax, lovingly known as
334 the *foo{THING} syntax. *foo{THING} returns a reference to the THING
335 slot in *foo (which is the symbol table entry which holds everything
336 known as foo).
337 .PP
338 .Vb 9
339 \& $scalarref = *foo{SCALAR};
340 \& $arrayref = *ARGV{ARRAY};
341 \& $hashref = *ENV{HASH};
342 \& $coderef = *handler{CODE};
343 \& $ioref = *STDIN{IO};
344 \& $globref = *foo{GLOB};
345 \& $formatref = *foo{FORMAT};
346 \& $globname = *foo{NAME}; # "foo"
347 \& $pkgname = *foo{PACKAGE}; # "main"
348 .Ve
349 .PP
350 Most of these are self-explanatory, but \f(CW*foo{IO}\fR
351 deserves special attention. It returns
352 the IO handle, used for file handles ("open" in perlfunc), sockets
353 ("socket" in perlfunc and "socketpair" in perlfunc), and directory
354 handles ("opendir" in perlfunc). For compatibility with previous
355 versions of Perl, \f(CW*foo{FILEHANDLE}\fR is a synonym for \f(CW*foo{IO}\fR, though it
356 is discouraged, to encourage a consistent use of one name: IO. On perls
357 between v5.8 and v5.22, it will issue a deprecation warning, but this
358 deprecation has since been rescinded.
359 .PP
360 \&\f(CW*foo{THING}\fR returns undef if that particular THING hasn't been used yet,
361 except in the case of scalars. \f(CW*foo{SCALAR}\fR returns a reference to an
362 anonymous scalar if \f(CW$foo\fR hasn't been used yet. This might change in a
363 future release.
364 .PP
365 \&\f(CW*foo{NAME}\fR and \f(CW*foo{PACKAGE}\fR are the exception, in that they return
366 strings, rather than references. These return the package and name of the
367 typeglob itself, rather than one that has been assigned to it. So, after
368 \&\f(CW\*(C`*foo=*Foo::bar\*(C'\fR, \f(CW*foo\fR will become "*Foo::bar" when used as a string,
369 but \f(CW*foo{PACKAGE}\fR and \f(CW*foo{NAME}\fR will continue to produce "main" and
370 "foo", respectively.
371 .PP
372 \&\f(CW*foo{IO}\fR is an alternative to the \f(CW*HANDLE\fR mechanism given in
373 "Typeglobs and Filehandles" in perldata for passing filehandles
374 into or out of subroutines, or storing into larger data structures.
375 Its disadvantage is that it won't create a new filehandle for you.
376 Its advantage is that you have less risk of clobbering more than
377 you want to with a typeglob assignment. (It still conflates file
378 and directory handles, though.) However, if you assign the incoming
379 value to a scalar instead of a typeglob as we do in the examples
380 below, there's no risk of that happening.
381 .PP
382 .Vb 2
383 \& splutter(*STDOUT); # pass the whole glob
384 \& splutter(*STDOUT{IO}); # pass both file and dir handles
385 \&
386 \& sub splutter {
387 \& my $fh = shift;
388 \& print $fh "her um well a hmmm\en";
389 \& }
390 \&
391 \& $rec = get_rec(*STDIN); # pass the whole glob
392 \& $rec = get_rec(*STDIN{IO}); # pass both file and dir handles
393 \&
394 \& sub get_rec {
395 \& my $fh = shift;
396 \& return scalar <$fh>;
397 \& }
398 .Ve
399 .SS "Using References"
400 .IX Xref "reference, use dereferencing dereference"
401 .IX Subsection "Using References"
402 That's it for creating references. By now you're probably dying to
403 know how to use references to get back to your long-lost data. There
404 are several basic methods.
405 .PP
406 \fISimple Scalar\fR
407 .IX Subsection "Simple Scalar"
408 .PP
409 Anywhere you'd put an identifier (or chain of identifiers) as part
410 of a variable or subroutine name, you can replace the identifier with
411 a simple scalar variable containing a reference of the correct type:
412 .PP
413 .Vb 6
414 \& $bar = $$scalarref;
415 \& push(@$arrayref, $filename);
416 \& $$arrayref[0] = "January";
417 \& $$hashref{"KEY"} = "VALUE";
418 \& &$coderef(1,2,3);
419 \& print $globref "output\en";
420 .Ve
421 .PP
422 It's important to understand that we are specifically \fInot\fR dereferencing
423 \&\f(CW$arrayref[0]\fR or \f(CW$hashref{"KEY"}\fR there. The dereference of the
424 scalar variable happens \fIbefore\fR it does any key lookups. Anything more
425 complicated than a simple scalar variable must use methods 2 or 3 below.
426 However, a "simple scalar" includes an identifier that itself uses method
427 1 recursively. Therefore, the following prints "howdy".
428 .PP
429 .Vb 2
430 \& $refrefref = \e\e\e"howdy";
431 \& print $$$$refrefref;
432 .Ve
433 .PP
434 \fIBlock\fR
435 .IX Subsection "Block"
436 .PP
437 Anywhere you'd put an identifier (or chain of identifiers) as part of a
438 variable or subroutine name, you can replace the identifier with a
439 BLOCK returning a reference of the correct type. In other words, the
440 previous examples could be written like this:
441 .PP
442 .Vb 6
443 \& $bar = ${$scalarref};
444 \& push(@{$arrayref}, $filename);
445 \& ${$arrayref}[0] = "January";
446 \& ${$hashref}{"KEY"} = "VALUE";
447 \& &{$coderef}(1,2,3);
448 \& $globref\->print("output\en"); # iff IO::Handle is loaded
449 .Ve
450 .PP
451 Admittedly, it's a little silly to use the curlies in this case, but
452 the BLOCK can contain any arbitrary expression, in particular,
453 subscripted expressions:
454 .PP
455 .Vb 1
456 \& &{ $dispatch{$index} }(1,2,3); # call correct routine
457 .Ve
458 .PP
459 Because of being able to omit the curlies for the simple case of \f(CW$$x\fR,
460 people often make the mistake of viewing the dereferencing symbols as
461 proper operators, and wonder about their precedence. If they were,
462 though, you could use parentheses instead of braces. That's not the case.
463 Consider the difference below; case 0 is a short-hand version of case 1,
464 \&\fInot\fR case 2:
465 .PP
466 .Vb 4
467 \& $$hashref{"KEY"} = "VALUE"; # CASE 0
468 \& ${$hashref}{"KEY"} = "VALUE"; # CASE 1
469 \& ${$hashref{"KEY"}} = "VALUE"; # CASE 2
470 \& ${$hashref\->{"KEY"}} = "VALUE"; # CASE 3
471 .Ve
472 .PP
473 Case 2 is also deceptive in that you're accessing a variable
474 called \f(CW%hashref\fR, not dereferencing through \f(CW$hashref\fR to the hash
475 it's presumably referencing. That would be case 3.
476 .PP
477 \fIArrow Notation\fR
478 .IX Subsection "Arrow Notation"
479 .PP
480 Subroutine calls and lookups of individual array elements arise often
481 enough that it gets cumbersome to use method 2. As a form of
482 syntactic sugar, the examples for method 2 may be written:
483 .PP
484 .Vb 3
485 \& $arrayref\->[0] = "January"; # Array element
486 \& $hashref\->{"KEY"} = "VALUE"; # Hash element
487 \& $coderef\->(1,2,3); # Subroutine call
488 .Ve
489 .PP
490 The left side of the arrow can be any expression returning a reference,
491 including a previous dereference. Note that \f(CW$array[$x]\fR is \fInot\fR the
492 same thing as \f(CW\*(C`$array\->[$x]\*(C'\fR here:
493 .PP
494 .Vb 1
495 \& $array[$x]\->{"foo"}\->[0] = "January";
496 .Ve
497 .PP
498 This is one of the cases we mentioned earlier in which references could
499 spring into existence when in an lvalue context. Before this
500 statement, \f(CW$array[$x]\fR may have been undefined. If so, it's
501 automatically defined with a hash reference so that we can look up
502 \&\f(CW\*(C`{"foo"}\*(C'\fR in it. Likewise \f(CW\*(C`$array[$x]\->{"foo"}\*(C'\fR will automatically get
503 defined with an array reference so that we can look up \f(CW\*(C`[0]\*(C'\fR in it.
504 This process is called \fIautovivification\fR.
505 .PP
506 One more thing here. The arrow is optional \fIbetween\fR brackets
507 subscripts, so you can shrink the above down to
508 .PP
509 .Vb 1
510 \& $array[$x]{"foo"}[0] = "January";
511 .Ve
512 .PP
513 Which, in the degenerate case of using only ordinary arrays, gives you
514 multidimensional arrays just like C's:
515 .PP
516 .Vb 1
517 \& $score[$x][$y][$z] += 42;
518 .Ve
519 .PP
520 Well, okay, not entirely like C's arrays, actually. C doesn't know how
521 to grow its arrays on demand. Perl does.
522 .PP
523 \fIObjects\fR
524 .IX Subsection "Objects"
525 .PP
526 If a reference happens to be a reference to an object, then there are
527 probably methods to access the things referred to, and you should probably
528 stick to those methods unless you're in the class package that defines the
529 object's methods. In other words, be nice, and don't violate the object's
530 encapsulation without a very good reason. Perl does not enforce
531 encapsulation. We are not totalitarians here. We do expect some basic
532 civility though.
533 .PP
534 \fIMiscellaneous Usage\fR
535 .IX Subsection "Miscellaneous Usage"
536 .PP
537 Using a string or number as a reference produces a symbolic reference,
538 as explained above. Using a reference as a number produces an
539 integer representing its storage location in memory. The only
540 useful thing to be done with this is to compare two references
541 numerically to see whether they refer to the same location.
542 .IX Xref "reference, numeric context"
543 .PP
544 .Vb 3
545 \& if ($ref1 == $ref2) { # cheap numeric compare of references
546 \& print "refs 1 and 2 refer to the same thing\en";
547 \& }
548 .Ve
549 .PP
550 Using a reference as a string produces both its referent's type,
551 including any package blessing as described in perlobj, as well
552 as the numeric address expressed in hex. The \fBref()\fR operator returns
553 just the type of thing the reference is pointing to, without the
554 address. See "ref" in perlfunc for details and examples of its use.
555 .IX Xref "reference, string context"
556 .PP
557 The \fBbless()\fR operator may be used to associate the object a reference
558 points to with a package functioning as an object class. See perlobj.
559 .PP
560 A typeglob may be dereferenced the same way a reference can, because
561 the dereference syntax always indicates the type of reference desired.
562 So \f(CW\*(C`${*foo}\*(C'\fR and \f(CW\*(C`${\e$foo}\*(C'\fR both indicate the same scalar variable.
563 .PP
564 Here's a trick for interpolating a subroutine call into a string:
565 .PP
566 .Vb 1
567 \& print "My sub returned @{[mysub(1,2,3)]} that time.\en";
568 .Ve
569 .PP
570 The way it works is that when the \f(CW\*(C`@{...}\*(C'\fR is seen in the double-quoted
571 string, it's evaluated as a block. The block creates a reference to an
572 anonymous array containing the results of the call to \f(CW\*(C`mysub(1,2,3)\*(C'\fR. So
573 the whole block returns a reference to an array, which is then
574 dereferenced by \f(CW\*(C`@{...}\*(C'\fR and stuck into the double-quoted string. This
575 chicanery is also useful for arbitrary expressions:
576 .PP
577 .Vb 1
578 \& print "That yields @{[$n + 5]} widgets\en";
579 .Ve
580 .PP
581 Similarly, an expression that returns a reference to a scalar can be
582 dereferenced via \f(CW\*(C`${...}\*(C'\fR. Thus, the above expression may be written
583 as:
584 .PP
585 .Vb 1
586 \& print "That yields ${\e($n + 5)} widgets\en";
587 .Ve
588 .SS "Circular References"
589 .IX Xref "circular reference reference, circular"
590 .IX Subsection "Circular References"
591 It is possible to create a "circular reference" in Perl, which can lead
592 to memory leaks. A circular reference occurs when two references
593 contain a reference to each other, like this:
594 .PP
595 .Vb 3
596 \& my $foo = {};
597 \& my $bar = { foo => $foo };
598 \& $foo\->{bar} = $bar;
599 .Ve
600 .PP
601 You can also create a circular reference with a single variable:
602 .PP
603 .Vb 2
604 \& my $foo;
605 \& $foo = \e$foo;
606 .Ve
607 .PP
608 In this case, the reference count for the variables will never reach 0,
609 and the references will never be garbage-collected. This can lead to
610 memory leaks.
611 .PP
612 Because objects in Perl are implemented as references, it's possible to
613 have circular references with objects as well. Imagine a TreeNode class
614 where each node references its parent and child nodes. Any node with a
615 parent will be part of a circular reference.
616 .PP
617 You can break circular references by creating a "weak reference". A
618 weak reference does not increment the reference count for a variable,
619 which means that the object can go out of scope and be destroyed. You
620 can weaken a reference with the \f(CW\*(C`weaken\*(C'\fR function exported by the
621 Scalar::Util module, or available as \f(CW\*(C`builtin::weaken\*(C'\fR directly in
622 Perl version 5.35.7 or later.
623 .PP
624 Here's how we can make the first example safer:
625 .PP
626 .Vb 1
627 \& use Scalar::Util \*(Aqweaken\*(Aq;
628 \&
629 \& my $foo = {};
630 \& my $bar = { foo => $foo };
631 \& $foo\->{bar} = $bar;
632 \&
633 \& weaken $foo\->{bar};
634 .Ve
635 .PP
636 The reference from \f(CW$foo\fR to \f(CW$bar\fR has been weakened. When the
637 \&\f(CW$bar\fR variable goes out of scope, it will be garbage-collected. The
638 next time you look at the value of the \f(CW\*(C`$foo\->{bar}\*(C'\fR key, it will
639 be \f(CW\*(C`undef\*(C'\fR.
640 .PP
641 This action at a distance can be confusing, so you should be careful
642 with your use of weaken. You should weaken the reference in the
643 variable that will go out of scope \fIfirst\fR. That way, the longer-lived
644 variable will contain the expected reference until it goes out of
645 scope.
646 .SS "Symbolic references"
647 .IX Xref "reference, symbolic reference, soft symbolic reference soft reference"
648 .IX Subsection "Symbolic references"
649 We said that references spring into existence as necessary if they are
650 undefined, but we didn't say what happens if a value used as a
651 reference is already defined, but \fIisn't\fR a hard reference. If you
652 use it as a reference, it'll be treated as a symbolic
653 reference. That is, the value of the scalar is taken to be the \fIname\fR
654 of a variable, rather than a direct link to a (possibly) anonymous
655 value.
656 .PP
657 People frequently expect it to work like this. So it does.
658 .PP
659 .Vb 9
660 \& $name = "foo";
661 \& $$name = 1; # Sets $foo
662 \& ${$name} = 2; # Sets $foo
663 \& ${$name x 2} = 3; # Sets $foofoo
664 \& $name\->[0] = 4; # Sets $foo[0]
665 \& @$name = (); # Clears @foo
666 \& &$name(); # Calls &foo()
667 \& $pack = "THAT";
668 \& ${"${pack}::$name"} = 5; # Sets $THAT::foo without eval
669 .Ve
670 .PP
671 This is powerful, and slightly dangerous, in that it's possible
672 to intend (with the utmost sincerity) to use a hard reference, and
673 accidentally use a symbolic reference instead. To protect against
674 that, you can say
675 .PP
676 .Vb 1
677 \& use strict \*(Aqrefs\*(Aq;
678 .Ve
679 .PP
680 and then only hard references will be allowed for the rest of the enclosing
681 block. An inner block may countermand that with
682 .PP
683 .Vb 1
684 \& no strict \*(Aqrefs\*(Aq;
685 .Ve
686 .PP
687 Only package variables (globals, even if localized) are visible to
688 symbolic references. Lexical variables (declared with \fBmy()\fR) aren't in
689 a symbol table, and thus are invisible to this mechanism. For example:
690 .PP
691 .Vb 6
692 \& local $value = 10;
693 \& $ref = "value";
694 \& {
695 \& my $value = 20;
696 \& print $$ref;
697 \& }
698 .Ve
699 .PP
700 This will still print 10, not 20. Remember that \fBlocal()\fR affects package
701 variables, which are all "global" to the package.
702 .SS "Not-so-symbolic references"
703 .IX Subsection "Not-so-symbolic references"
704 Brackets around a symbolic reference can simply
705 serve to isolate an identifier or variable name from the rest of an
706 expression, just as they always have within a string. For example,
707 .PP
708 .Vb 2
709 \& $push = "pop on ";
710 \& print "${push}over";
711 .Ve
712 .PP
713 has always meant to print "pop on over", even though push is
714 a reserved word. This is generalized to work the same
715 without the enclosing double quotes, so that
716 .PP
717 .Vb 1
718 \& print ${push} . "over";
719 .Ve
720 .PP
721 and even
722 .PP
723 .Vb 1
724 \& print ${ push } . "over";
725 .Ve
726 .PP
727 will have the same effect. This
728 construct is \fInot\fR considered to be a symbolic reference when you're
729 using strict refs:
730 .PP
731 .Vb 3
732 \& use strict \*(Aqrefs\*(Aq;
733 \& ${ bareword }; # Okay, means $bareword.
734 \& ${ "bareword" }; # Error, symbolic reference.
735 .Ve
736 .PP
737 Similarly, because of all the subscripting that is done using single words,
738 the same rule applies to any bareword that is used for subscripting a hash.
739 So now, instead of writing
740 .PP
741 .Vb 1
742 \& $hash{ "aaa" }{ "bbb" }{ "ccc" }
743 .Ve
744 .PP
745 you can write just
746 .PP
747 .Vb 1
748 \& $hash{ aaa }{ bbb }{ ccc }
749 .Ve
750 .PP
751 and not worry about whether the subscripts are reserved words. In the
752 rare event that you do wish to do something like
753 .PP
754 .Vb 1
755 \& $hash{ shift }
756 .Ve
757 .PP
758 you can force interpretation as a reserved word by adding anything that
759 makes it more than a bareword:
760 .PP
761 .Vb 3
762 \& $hash{ shift() }
763 \& $hash{ +shift }
764 \& $hash{ shift @_ }
765 .Ve
766 .PP
767 The \f(CW\*(C`use warnings\*(C'\fR pragma or the \fB\-w\fR switch will warn you if it
768 interprets a reserved word as a string.
769 But it will no longer warn you about using lowercase words, because the
770 string is effectively quoted.
771 .SS "Pseudo-hashes: Using an array as a hash"
772 .IX Xref "pseudo-hash pseudo hash pseudohash"
773 .IX Subsection "Pseudo-hashes: Using an array as a hash"
774 Pseudo-hashes have been removed from Perl. The 'fields' pragma
775 remains available.
776 .SS "Function Templates"
777 .IX Xref "scope, lexical closure lexical lexical scope subroutine, nested sub, nested subroutine, local sub, local"
778 .IX Subsection "Function Templates"
779 As explained above, an anonymous function with access to the lexical
780 variables visible when that function was compiled, creates a closure. It
781 retains access to those variables even though it doesn't get run until
782 later, such as in a signal handler or a Tk callback.
783 .PP
784 Using a closure as a function template allows us to generate many functions
785 that act similarly. Suppose you wanted functions named after the colors
786 that generated HTML font changes for the various colors:
787 .PP
788 .Vb 1
789 \& print "Be ", red("careful"), "with that ", green("light");
790 .Ve
791 .PP
792 The \fBred()\fR and \fBgreen()\fR functions would be similar. To create these,
793 we'll assign a closure to a typeglob of the name of the function we're
794 trying to build.
795 .PP
796 .Vb 5
797 \& @colors = qw(red blue green yellow orange purple violet);
798 \& for my $name (@colors) {
799 \& no strict \*(Aqrefs\*(Aq; # allow symbol table manipulation
800 \& *$name = *{uc $name} = sub { "<FONT COLOR=\*(Aq$name\*(Aq>@_</FONT>" };
801 \& }
802 .Ve
803 .PP
804 Now all those different functions appear to exist independently. You can
805 call \fBred()\fR, \fBRED()\fR, \fBblue()\fR, \fBBLUE()\fR, \fBgreen()\fR, etc. This technique saves on
806 both compile time and memory use, and is less error-prone as well, since
807 syntax checks happen at compile time. It's critical that any variables in
808 the anonymous subroutine be lexicals in order to create a proper closure.
809 That's the reasons for the \f(CW\*(C`my\*(C'\fR on the loop iteration variable.
810 .PP
811 This is one of the only places where giving a prototype to a closure makes
812 much sense. If you wanted to impose scalar context on the arguments of
813 these functions (probably not a wise idea for this particular example),
814 you could have written it this way instead:
815 .PP
816 .Vb 1
817 \& *$name = sub ($) { "<FONT COLOR=\*(Aq$name\*(Aq>$_[0]</FONT>" };
818 .Ve
819 .PP
820 However, since prototype checking happens at compile time, the assignment
821 above happens too late to be of much use. You could address this by
822 putting the whole loop of assignments within a BEGIN block, forcing it
823 to occur during compilation.
824 .PP
825 Access to lexicals that change over time\-\-like those in the \f(CW\*(C`for\*(C'\fR loop
826 above, basically aliases to elements from the surrounding lexical scopes\-\-
827 only works with anonymous subs, not with named subroutines. Generally
828 said, named subroutines do not nest properly and should only be declared
829 in the main package scope.
830 .PP
831 This is because named subroutines are created at compile time so their
832 lexical variables get assigned to the parent lexicals from the first
833 execution of the parent block. If a parent scope is entered a second
834 time, its lexicals are created again, while the nested subs still
835 reference the old ones.
836 .PP
837 Anonymous subroutines get to capture each time you execute the \f(CW\*(C`sub\*(C'\fR
838 operator, as they are created on the fly. If you are accustomed to using
839 nested subroutines in other programming languages with their own private
840 variables, you'll have to work at it a bit in Perl. The intuitive coding
841 of this type of thing incurs mysterious warnings about "will not stay
842 shared" due to the reasons explained above.
843 For example, this won't work:
844 .PP
845 .Vb 5
846 \& sub outer {
847 \& my $x = $_[0] + 35;
848 \& sub inner { return $x * 19 } # WRONG
849 \& return $x + inner();
850 \& }
851 .Ve
852 .PP
853 A work-around is the following:
854 .PP
855 .Vb 5
856 \& sub outer {
857 \& my $x = $_[0] + 35;
858 \& local *inner = sub { return $x * 19 };
859 \& return $x + inner();
860 \& }
861 .Ve
862 .PP
863 Now \fBinner()\fR can only be called from within \fBouter()\fR, because of the
864 temporary assignments of the anonymous subroutine. But when it does,
865 it has normal access to the lexical variable \f(CW$x\fR from the scope of
866 \&\fBouter()\fR at the time outer is invoked.
867 .PP
868 This has the interesting effect of creating a function local to another
869 function, something not normally supported in Perl.
870 .SS "Postfix Dereference Syntax"
871 .IX Subsection "Postfix Dereference Syntax"
872 Beginning in v5.20.0, a postfix syntax for using references is
873 available. It behaves as described in "Using References", but instead
874 of a prefixed sigil, a postfixed sigil-and-star is used.
875 .PP
876 For example:
877 .PP
878 .Vb 2
879 \& $r = \e@a;
880 \& @b = $r\->@*; # equivalent to @$r or @{ $r }
881 \&
882 \& $r = [ 1, [ 2, 3 ], 4 ];
883 \& $r\->[1]\->@*; # equivalent to @{ $r\->[1] }
884 .Ve
885 .PP
886 In Perl 5.20 and 5.22, this syntax must be enabled with \f(CWuse feature
887 \&\*(Aqpostderef\*(Aq\fR. As of Perl 5.24, no feature declarations are required to make
888 it available.
889 .PP
890 Postfix dereference should work in all circumstances where block
891 (circumfix) dereference worked, and should be entirely equivalent. This
892 syntax allows dereferencing to be written and read entirely
893 left-to-right. The following equivalencies are defined:
894 .PP
895 .Vb 6
896 \& $sref\->$*; # same as ${ $sref }
897 \& $aref\->@*; # same as @{ $aref }
898 \& $aref\->$#*; # same as $#{ $aref }
899 \& $href\->%*; # same as %{ $href }
900 \& $cref\->&*; # same as &{ $cref }
901 \& $gref\->**; # same as *{ $gref }
902 .Ve
903 .PP
904 Note especially that \f(CW\*(C`$cref\->&*\*(C'\fR is \fInot\fR equivalent to \f(CW$cref\->()\fR, and can serve different purposes.
905 .PP
906 Glob elements can be extracted through the postfix dereferencing feature:
907 .PP
908 .Vb 1
909 \& $gref\->*{SCALAR}; # same as *{ $gref }{SCALAR}
910 .Ve
911 .PP
912 Postfix array and scalar dereferencing \fIcan\fR be used in interpolating
913 strings (double quotes or the \f(CW\*(C`qq\*(C'\fR operator), but only if the
914 \&\f(CW\*(C`postderef_qq\*(C'\fR feature is enabled. Interpolation of postfix array highest index
915 access (\f(CW\*(C`\->$#*\*(C'\fR) is also supported when the \f(CW\*(C`postderef_qq\*(C'\fR feature is
916 enabled.
917 .SS "Postfix Reference Slicing"
918 .IX Subsection "Postfix Reference Slicing"
919 Value slices of arrays and hashes may also be taken with postfix
920 dereferencing notation, with the following equivalencies:
921 .PP
922 .Vb 2
923 \& $aref\->@[ ... ]; # same as @$aref[ ... ]
924 \& $href\->@{ ... }; # same as @$href{ ... }
925 .Ve
926 .PP
927 Postfix key/value pair slicing, added in 5.20.0 and documented in
928 the Key/Value Hash Slices section of perldata, also behaves as expected:
929 .PP
930 .Vb 2
931 \& $aref\->%[ ... ]; # same as %$aref[ ... ]
932 \& $href\->%{ ... }; # same as %$href{ ... }
933 .Ve
934 .PP
935 As with postfix array, postfix value slice dereferencing \fIcan\fR be used
936 in interpolating strings (double quotes or the \f(CW\*(C`qq\*(C'\fR operator), but only
937 if the \f(CW\*(C`postderef_qq\*(C'\fR feature is enabled.
938 .SS "Assigning to References"
939 .IX Subsection "Assigning to References"
940 Beginning in v5.22.0, the referencing operator can be assigned to. It
941 performs an aliasing operation, so that the variable name referenced on the
942 left-hand side becomes an alias for the thing referenced on the right-hand
943 side:
944 .PP
945 .Vb 2
946 \& \e$x = \e$y; # $x and $y now point to the same scalar
947 \& \e&foo = \e&bar; # foo() now means bar()
948 .Ve
949 .PP
950 This syntax must be enabled with \f(CW\*(C`use feature \*(Aqrefaliasing\*(Aq\*(C'\fR. It is
951 experimental, and will warn by default unless \f(CWno warnings
952 \&\*(Aqexperimental::refaliasing\*(Aq\fR is in effect.
953 .PP
954 These forms may be assigned to, and cause the right-hand side to be
955 evaluated in scalar context:
956 .PP
957 .Vb 10
958 \& \e$scalar
959 \& \e@array
960 \& \e%hash
961 \& \e&sub
962 \& \emy $scalar
963 \& \emy @array
964 \& \emy %hash
965 \& \estate $scalar # or @array, etc.
966 \& \eour $scalar # etc.
967 \& \elocal $scalar # etc.
968 \& \elocal our $scalar # etc.
969 \& \e$some_array[$index]
970 \& \e$some_hash{$key}
971 \& \elocal $some_array[$index]
972 \& \elocal $some_hash{$key}
973 \& condition ? \e$this : \e$that[0] # etc.
974 .Ve
975 .PP
976 Slicing operations and parentheses cause
977 the right-hand side to be evaluated in
978 list context:
979 .PP
980 .Vb 10
981 \& \e@array[5..7]
982 \& (\e@array[5..7])
983 \& \e(@array[5..7])
984 \& \e@hash{\*(Aqfoo\*(Aq,\*(Aqbar\*(Aq}
985 \& (\e@hash{\*(Aqfoo\*(Aq,\*(Aqbar\*(Aq})
986 \& \e(@hash{\*(Aqfoo\*(Aq,\*(Aqbar\*(Aq})
987 \& (\e$scalar)
988 \& \e($scalar)
989 \& \e(my $scalar)
990 \& \emy($scalar)
991 \& (\e@array)
992 \& (\e%hash)
993 \& (\e&sub)
994 \& \e(&sub)
995 \& \e($foo, @bar, %baz)
996 \& (\e$foo, \e@bar, \e%baz)
997 .Ve
998 .PP
999 Each element on the right-hand side must be a reference to a datum of the
1000 right type. Parentheses immediately surrounding an array (and possibly
1001 also \f(CW\*(C`my\*(C'\fR/\f(CW\*(C`state\*(C'\fR/\f(CW\*(C`our\*(C'\fR/\f(CW\*(C`local\*(C'\fR) will make each element of the array an
1002 alias to the corresponding scalar referenced on the right-hand side:
1003 .PP
1004 .Vb 5
1005 \& \e(@x) = \e(@y); # @x and @y now have the same elements
1006 \& \emy(@x) = \e(@y); # likewise
1007 \& \e(my @x) = \e(@y); # likewise
1008 \& push @x, 3; # but now @x has an extra element that @y lacks
1009 \& \e(@x) = (\e$x, \e$y, \e$z); # @x now contains $x, $y, and $z
1010 .Ve
1011 .PP
1012 Combining that form with \f(CW\*(C`local\*(C'\fR and putting parentheses immediately
1013 around a hash are forbidden (because it is not clear what they should do):
1014 .PP
1015 .Vb 2
1016 \& \elocal(@array) = foo(); # WRONG
1017 \& \e(%hash) = bar(); # WRONG
1018 .Ve
1019 .PP
1020 Assignment to references and non-references may be combined in lists and
1021 conditional ternary expressions, as long as the values on the right-hand
1022 side are the right type for each element on the left, though this may make
1023 for obfuscated code:
1024 .PP
1025 .Vb 4
1026 \& (my $tom, \emy $dick, \emy @harry) = (\e1, \e2, [1..3]);
1027 \& # $tom is now \e1
1028 \& # $dick is now 2 (read\-only)
1029 \& # @harry is (1,2,3)
1030 \&
1031 \& my $type = ref $thingy;
1032 \& ($type ? $type eq \*(AqARRAY\*(Aq ? \e@foo : \e$bar : $baz) = $thingy;
1033 .Ve
1034 .PP
1035 A reference assignment in list context returns a list of references to
1036 each left-hand value. So for example
1037 .PP
1038 .Vb 1
1039 \& @b = ((\e$l1, \e$l2, \e(@a)) = (\e$r1, \e$r2, $\er3, \e$4));
1040 .Ve
1041 .PP
1042 is equivalent to
1043 .PP
1044 .Vb 2
1045 \& (\e$l1, \e$l2, \e(@a)) = (\e$r1, \e$r2, $\er3, \e$4);
1046 \& @b = (\e$l1, \e$l2, \e$a[0], \e$a[1]);
1047 .Ve
1048 .PP
1049 The \f(CW\*(C`foreach\*(C'\fR loop can also take a reference constructor for its loop
1050 variable, though the syntax is limited to one of the following, with an
1051 optional \f(CW\*(C`my\*(C'\fR, \f(CW\*(C`state\*(C'\fR, or \f(CW\*(C`our\*(C'\fR after the backslash:
1052 .PP
1053 .Vb 4
1054 \& \e$s
1055 \& \e@a
1056 \& \e%h
1057 \& \e&c
1058 .Ve
1059 .PP
1060 No parentheses are permitted. This feature is particularly useful for
1061 arrays-of-arrays, or arrays-of-hashes:
1062 .PP
1063 .Vb 3
1064 \& foreach \emy @a (@array_of_arrays) {
1065 \& frobnicate($a[0], $a[\-1]);
1066 \& }
1067 \&
1068 \& foreach \emy %h (@array_of_hashes) {
1069 \& $h{gelastic}++ if $h{type} eq \*(Aqfunny\*(Aq;
1070 \& }
1071 .Ve
1072 .PP
1073 \&\fBCAVEAT:\fR Aliasing does not work correctly with closures. If you try to
1074 alias lexical variables from an inner subroutine or \f(CW\*(C`eval\*(C'\fR, the aliasing
1075 will only be visible within that inner sub, and will not affect the outer
1076 subroutine where the variables are declared. This bizarre behavior is
1077 subject to change.
1078 .SS "Declaring a Reference to a Variable"
1079 .IX Subsection "Declaring a Reference to a Variable"
1080 Beginning in v5.26.0, the referencing operator can come after \f(CW\*(C`my\*(C'\fR,
1081 \&\f(CW\*(C`state\*(C'\fR, \f(CW\*(C`our\*(C'\fR, or \f(CW\*(C`local\*(C'\fR. This syntax must be enabled with \f(CW\*(C`use
1082 feature \*(Aqdeclared_refs\*(Aq\*(C'\fR. It is experimental, and will warn by default
1083 unless \f(CW\*(C`no warnings \*(Aqexperimental::refaliasing\*(Aq\*(C'\fR is in effect.
1084 .PP
1085 This feature makes these:
1086 .PP
1087 .Vb 2
1088 \& my \e$x;
1089 \& our \e$y;
1090 .Ve
1091 .PP
1092 equivalent to:
1093 .PP
1094 .Vb 2
1095 \& \emy $x;
1096 \& \eour $x;
1097 .Ve
1098 .PP
1099 It is intended mainly for use in assignments to references (see
1100 "Assigning to References", above). It also allows the backslash to be
1101 used on just some items in a list of declared variables:
1102 .PP
1103 .Vb 1
1104 \& my ($foo, \e@bar, \e%baz); # equivalent to: my $foo, \emy(@bar, %baz);
1105 .Ve
1106 .SH "WARNING: Don't use references as hash keys"
1107 .IX Xref "reference, string context reference, use as hash key"
1108 .IX Header "WARNING: Don't use references as hash keys"
1109 You may not (usefully) use a reference as the key to a hash. It will be
1110 converted into a string:
1111 .PP
1112 .Vb 1
1113 \& $x{ \e$x } = $x;
1114 .Ve
1115 .PP
1116 If you try to dereference the key, it won't do a hard dereference, and
1117 you won't accomplish what you're attempting. You might want to do something
1118 more like
1119 .PP
1120 .Vb 2
1121 \& $r = \e@a;
1122 \& $x{ $r } = $r;
1123 .Ve
1124 .PP
1125 And then at least you can use the \fBvalues()\fR, which will be
1126 real refs, instead of the \fBkeys()\fR, which won't.
1127 .PP
1128 The standard Tie::RefHash module provides a convenient workaround to this.
1129 .SH "SEE ALSO"
1130 .IX Header "SEE ALSO"
1131 Besides the obvious documents, source code can be instructive.
1132 Some pathological examples of the use of references can be found
1133 in the \fIt/op/ref.t\fR regression test in the Perl source directory.
1134 .PP
1135 See also perldsc and perllol for how to use references to create
1136 complex data structures, and perlootut and perlobj
1137 for how to use them to create objects.