1PERLTOOT(1) Perl Programmers Reference Guide PERLTOOT(1)
2
6 perltoot - Tom's object-oriented tutorial for perl
7
9 Object-oriented programming is a big seller these days. Some managers
10 would rather have objects than sliced bread. Why is that? What's so
11 special about an object? Just what is an object anyway?
12 An object is nothing but a way of tucking away complex behaviours into
14 a neat little easy-to-use bundle. (This is what professors call
15 abstraction.) Smart people who have nothing to do but sit around for
16 weeks on end figuring out really hard problems make these nifty objects
17 that even regular people can use. (This is what professors call
18 software reuse.) Users (well, programmers) can play with this little
19 bundle all they want, but they aren't to open it up and mess with the
20 insides. Just like an expensive piece of hardware, the contract says
21 that you void the warranty if you muck with the cover. So don't do
22 that.
23 The heart of objects is the class, a protected little private namespace
25 full of data and functions. A class is a set of related routines that
26 addresses some problem area. You can think of it as a user-defined
27 type. The Perl package mechanism, also used for more traditional
28 modules, is used for class modules as well. Objects "live" in a class,
29 meaning that they belong to some package.
30 More often than not, the class provides the user with little bundles.
32 These bundles are objects. They know whose class they belong to, and
33 how to behave. Users ask the class to do something, like "give me an
34 object." Or they can ask one of these objects to do something. Asking
35 a class to do something for you is calling a class method. Asking an
36 object to do something for you is calling an object method. Asking
37 either a class (usually) or an object (sometimes) to give you back an
38 object is calling a constructor, which is just a kind of method.
39 That's all well and good, but how is an object different from any other
41 Perl data type? Just what is an object really; that is, what's its
42 fundamental type? The answer to the first question is easy. An object
43 is different from any other data type in Perl in one and only one way:
44 you may dereference it using not merely string or numeric subscripts as
45 with simple arrays and hashes, but with named subroutine calls. In a
46 word, with methods.
47 The answer to the second question is that it's a reference, and not
49 just any reference, mind you, but one whose referent has been bless()ed
50 into a particular class (read: package). What kind of reference?
51 Well, the answer to that one is a bit less concrete. That's because in
52 Perl the designer of the class can employ any sort of reference they'd
53 like as the underlying intrinsic data type. It could be a scalar, an
54 array, or a hash reference. It could even be a code reference. But
55 because of its inherent flexibility, an object is usually a hash
56 reference.
57
59 Before you create a class, you need to decide what to name it. That's
60 because the class (package) name governs the name of the file used to
61 house it, just as with regular modules. Then, that class (package)
62 should provide one or more ways to generate objects. Finally, it
63 should provide mechanisms to allow users of its objects to indirectly
64 manipulate these objects from a distance.
65 For example, let's make a simple Person class module. It gets stored
67 in the file Person.pm. If it were called a Happy::Person class, it
68 would be stored in the file Happy/Person.pm, and its package would
69 become Happy::Person instead of just Person. (On a personal computer
70 not running Unix or Plan 9, but something like Mac OS or VMS, the
71 directory separator may be different, but the principle is the same.)
72 Do not assume any formal relationship between modules based on their
73 directory names. This is merely a grouping convenience, and has no
74 effect on inheritance, variable accessibility, or anything else.
75 For this module we aren't going to use Exporter, because we're a well-
77 behaved class module that doesn't export anything at all. In order to
78 manufacture objects, a class needs to have a constructor method. A
79 constructor gives you back not just a regular data type, but a brand-
80 new object in that class. This magic is taken care of by the bless()
81 function, whose sole purpose is to enable its referent to be used as an
82 object. Remember: being an object really means nothing more than that
83 methods may now be called against it.
84 While a constructor may be named anything you'd like, most Perl
86 programmers seem to like to call theirs new(). However, new() is not a
87 reserved word, and a class is under no obligation to supply such. Some
88 programmers have also been known to use a function with the same name
89 as the class as the constructor.
90 Object Representation
92 By far the most common mechanism used in Perl to represent a Pascal
93 record, a C struct, or a C++ class is an anonymous hash. That's
94 because a hash has an arbitrary number of data fields, each
95 conveniently accessed by an arbitrary name of your own devising.
96 If you were just doing a simple struct-like emulation, you would likely
98 go about it something like this:
99 $rec = {
101 name => "Jason",
102 age => 23,
103 peers => [ "Norbert", "Rhys", "Phineas"],
104 };
105 If you felt like it, you could add a bit of visual distinction by up-
107 casing the hash keys:
108 $rec = {
110 NAME => "Jason",
111 AGE => 23,
112 PEERS => [ "Norbert", "Rhys", "Phineas"],
113 };
114 And so you could get at "$rec->{NAME}" to find "Jason", or "@{
116 $rec->{PEERS} }" to get at "Norbert", "Rhys", and "Phineas". (Have you
117 ever noticed how many 23-year-old programmers seem to be named "Jason"
118 these days? :-)
119 This same model is often used for classes, although it is not
121 considered the pinnacle of programming propriety for folks from outside
122 the class to come waltzing into an object, brazenly accessing its data
123 members directly. Generally speaking, an object should be considered
124 an opaque cookie that you use object methods to access. Visually,
125 methods look like you're dereffing a reference using a function name
126 instead of brackets or braces.
127 Class Interface
129 Some languages provide a formal syntactic interface to a class's
130 methods, but Perl does not. It relies on you to read the documentation
131 of each class. If you try to call an undefined method on an object,
132 Perl won't complain, but the program will trigger an exception while
133 it's running. Likewise, if you call a method expecting a prime number
134 as its argument with a non-prime one instead, you can't expect the
135 compiler to catch this. (Well, you can expect it all you like, but
136 it's not going to happen.)
137 Let's suppose you have a well-educated user of your Person class,
139 someone who has read the docs that explain the prescribed interface.
140 Here's how they might use the Person class:
141 use Person;
143 $him = Person->new();
145 $him->name("Jason");
146 $him->age(23);
147 $him->peers( "Norbert", "Rhys", "Phineas" );
148 push @All_Recs, $him; # save object in array for later
150 printf "%s is %d years old.\n", $him->name, $him->age;
152 print "His peers are: ", join(", ", $him->peers), "\n";
153 printf "Last rec's name is %s\n", $All_Recs[-1]->name;
155 As you can see, the user of the class doesn't know (or at least, has no
157 business paying attention to the fact) that the object has one
158 particular implementation or another. The interface to the class and
159 its objects is exclusively via methods, and that's all the user of the
160 class should ever play with.
161 Constructors and Instance Methods
163 Still, someone has to know what's in the object. And that someone is
164 the class. It implements methods that the programmer uses to access
165 the object. Here's how to implement the Person class using the
166 standard hash-ref-as-an-object idiom. We'll make a class method called
167 new() to act as the constructor, and three object methods called
168 name(), age(), and peers() to get at per-object data hidden away in our
169 anonymous hash.
170 package Person;
172 use strict;
173 ##################################################
175 ## the object constructor (simplistic version) ##
176 ##################################################
177 sub new {
178 my $self = {};
179 $self->{NAME} = undef;
180 $self->{AGE} = undef;
181 $self->{PEERS} = [];
182 bless($self); # but see below
183 return $self;
184 }
185 ##############################################
187 ## methods to access per-object data ##
188 ## ##
189 ## With args, they set the value. Without ##
190 ## any, they only retrieve it/them. ##
191 ##############################################
192 sub name {
194 my $self = shift;
195 if (@_) { $self->{NAME} = shift }
196 return $self->{NAME};
197 }
198 sub age {
200 my $self = shift;
201 if (@_) { $self->{AGE} = shift }
202 return $self->{AGE};
203 }
204 sub peers {
206 my $self = shift;
207 if (@_) { @{ $self->{PEERS} } = @_ }
208 return @{ $self->{PEERS} };
209 }
210 1; # so the require or use succeeds
212 We've created three methods to access an object's data, name(), age(),
214 and peers(). These are all substantially similar. If called with an
215 argument, they set the appropriate field; otherwise they return the
216 value held by that field, meaning the value of that hash key.
217 Planning for the Future: Better Constructors
219 Even though at this point you may not even know what it means, someday
220 you're going to worry about inheritance. (You can safely ignore this
221 for now and worry about it later if you'd like.) To ensure that this
222 all works out smoothly, you must use the double-argument form of
223 bless(). The second argument is the class into which the referent will
224 be blessed. By not assuming our own class as the default second
225 argument and instead using the class passed into us, we make our
226 constructor inheritable.
227 sub new {
229 my $class = shift;
230 my $self = {};
231 $self->{NAME} = undef;
232 $self->{AGE} = undef;
233 $self->{PEERS} = [];
234 bless ($self, $class);
235 return $self;
236 }
237 That's about all there is for constructors. These methods bring
239 objects to life, returning neat little opaque bundles to the user to be
240 used in subsequent method calls.
241 Destructors
243 Every story has a beginning and an end. The beginning of the object's
244 story is its constructor, explicitly called when the object comes into
245 existence. But the ending of its story is the destructor, a method
246 implicitly called when an object leaves this life. Any per-object
247 clean-up code is placed in the destructor, which must (in Perl) be
248 called DESTROY.
249 If constructors can have arbitrary names, then why not destructors?
251 Because while a constructor is explicitly called, a destructor is not.
252 Destruction happens automatically via Perl's garbage collection (GC)
253 system, which is a quick but somewhat lazy reference-based GC system.
254 To know what to call, Perl insists that the destructor be named
255 DESTROY. Perl's notion of the right time to call a destructor is not
256 well-defined currently, which is why your destructors should not rely
257 on when they are called.
258 Why is DESTROY in all caps? Perl on occasion uses purely uppercase
260 function names as a convention to indicate that the function will be
261 automatically called by Perl in some way. Others that are called
262 implicitly include BEGIN, END, AUTOLOAD, plus all methods used by tied
263 objects, described in perltie.
264 In really good object-oriented programming languages, the user doesn't
266 care when the destructor is called. It just happens when it's supposed
267 to. In low-level languages without any GC at all, there's no way to
268 depend on this happening at the right time, so the programmer must
269 explicitly call the destructor to clean up memory and state, crossing
270 their fingers that it's the right time to do so. Unlike C++, an
271 object destructor is nearly never needed in Perl, and even when it is,
272 explicit invocation is uncalled for. In the case of our Person class,
273 we don't need a destructor because Perl takes care of simple matters
274 like memory deallocation.
275 The only situation where Perl's reference-based GC won't work is when
277 there's a circularity in the data structure, such as:
278 $this->{WHATEVER} = $this;
280 In that case, you must delete the self-reference manually if you expect
282 your program not to leak memory. While admittedly error-prone, this is
283 the best we can do right now. Nonetheless, rest assured that when your
284 program is finished, its objects' destructors are all duly called. So
285 you are guaranteed that an object eventually gets properly destroyed,
286 except in the unique case of a program that never exits. (If you're
287 running Perl embedded in another application, this full GC pass happens
288 a bit more frequently--whenever a thread shuts down.)
289 Other Object Methods
291 The methods we've talked about so far have either been constructors or
292 else simple "data methods", interfaces to data stored in the object.
293 These are a bit like an object's data members in the C++ world, except
294 that strangers don't access them as data. Instead, they should only
295 access the object's data indirectly via its methods. This is an
296 important rule: in Perl, access to an object's data should only be made
297 through methods.
298 Perl doesn't impose restrictions on who gets to use which methods. The
300 public-versus-private distinction is by convention, not syntax. (Well,
301 unless you use the Alias module described below in "Data Members as
302 Variables".) Occasionally you'll see method names beginning or ending
303 with an underscore or two. This marking is a convention indicating
304 that the methods are private to that class alone and sometimes to its
305 closest acquaintances, its immediate subclasses. But this distinction
306 is not enforced by Perl itself. It's up to the programmer to behave.
307 There's no reason to limit methods to those that simply access data.
309 Methods can do anything at all. The key point is that they're invoked
310 against an object or a class. Let's say we'd like object methods that
311 do more than fetch or set one particular field.
312 sub exclaim {
314 my $self = shift;
315 return sprintf "Hi, I'm %s, age %d, working with %s",
316 $self->{NAME}, $self->{AGE}, join(", ", @{$self->{PEERS}});
317 }
318 Or maybe even one like this:
320 sub happy_birthday {
322 my $self = shift;
323 return ++$self->{AGE};
324 }
325 Some might argue that one should go at these this way:
327 sub exclaim {
329 my $self = shift;
330 return sprintf "Hi, I'm %s, age %d, working with %s",
331 $self->name, $self->age, join(", ", $self->peers);
332 }
333 sub happy_birthday {
335 my $self = shift;
336 return $self->age( $self->age() + 1 );
337 }
338 But since these methods are all executing in the class itself, this may
340 not be critical. There are tradeoffs to be made. Using direct hash
341 access is faster (about an order of magnitude faster, in fact), and
342 it's more convenient when you want to interpolate in strings. But
343 using methods (the external interface) internally shields not just the
344 users of your class but even you yourself from changes in your data
345 representation.
346
348 What about "class data", data items common to each object in a class?
349 What would you want that for? Well, in your Person class, you might
350 like to keep track of the total people alive. How do you implement
351 that?
352 You could make it a global variable called $Person::Census. But about
354 only reason you'd do that would be if you wanted people to be able to
355 get at your class data directly. They could just say $Person::Census
356 and play around with it. Maybe this is ok in your design scheme. You
357 might even conceivably want to make it an exported variable. To be
358 exportable, a variable must be a (package) global. If this were a
359 traditional module rather than an object-oriented one, you might do
360 that.
361 While this approach is expected in most traditional modules, it's
363 generally considered rather poor form in most object modules. In an
364 object module, you should set up a protective veil to separate
365 interface from implementation. So provide a class method to access
366 class data just as you provide object methods to access object data.
367 So, you could still keep $Census as a package global and rely upon
369 others to honor the contract of the module and therefore not play
370 around with its implementation. You could even be supertricky and make
371 $Census a tied object as described in perltie, thereby intercepting all
372 accesses.
373 But more often than not, you just want to make your class data a file-
375 scoped lexical. To do so, simply put this at the top of the file:
376 my $Census = 0;
378 Even though the scope of a my() normally expires when the block in
380 which it was declared is done (in this case the whole file being
381 required or used), Perl's deep binding of lexical variables guarantees
382 that the variable will not be deallocated, remaining accessible to
383 functions declared within that scope. This doesn't work with global
384 variables given temporary values via local(), though.
385 Irrespective of whether you leave $Census a package global or make it
387 instead a file-scoped lexical, you should make these changes to your
388 Person::new() constructor:
389 sub new {
391 my $class = shift;
392 my $self = {};
393 $Census++;
394 $self->{NAME} = undef;
395 $self->{AGE} = undef;
396 $self->{PEERS} = [];
397 bless ($self, $class);
398 return $self;
399 }
400 sub population {
402 return $Census;
403 }
404 Now that we've done this, we certainly do need a destructor so that
406 when Person is destroyed, the $Census goes down. Here's how this could
407 be done:
408 sub DESTROY { --$Census }
410 Notice how there's no memory to deallocate in the destructor? That's
412 something that Perl takes care of for you all by itself.
413 Alternatively, you could use the Class::Data::Inheritable module from
415 CPAN.
416 Accessing Class Data
418 It turns out that this is not really a good way to go about handling
419 class data. A good scalable rule is that you must never reference
420 class data directly from an object method. Otherwise you aren't
421 building a scalable, inheritable class. The object must be the
422 rendezvous point for all operations, especially from an object method.
423 The globals (class data) would in some sense be in the "wrong" package
424 in your derived classes. In Perl, methods execute in the context of
425 the class they were defined in, not that of the object that triggered
426 them. Therefore, namespace visibility of package globals in methods is
427 unrelated to inheritance.
428 Got that? Maybe not. Ok, let's say that some other class "borrowed"
430 (well, inherited) the DESTROY method as it was defined above. When
431 those objects are destroyed, the original $Census variable will be
432 altered, not the one in the new class's package namespace. Perhaps
433 this is what you want, but probably it isn't.
434 Here's how to fix this. We'll store a reference to the data in the
436 value accessed by the hash key "_CENSUS". Why the underscore? Well,
437 mostly because an initial underscore already conveys strong feelings of
438 magicalness to a C programmer. It's really just a mnemonic device to
439 remind ourselves that this field is special and not to be used as a
440 public data member in the same way that NAME, AGE, and PEERS are.
441 (Because we've been developing this code under the strict pragma, prior
442 to perl version 5.004 we'll have to quote the field name.)
443 sub new {
445 my $class = shift;
446 my $self = {};
447 $self->{NAME} = undef;
448 $self->{AGE} = undef;
449 $self->{PEERS} = [];
450 # "private" data
451 $self->{"_CENSUS"} = \$Census;
452 bless ($self, $class);
453 ++ ${ $self->{"_CENSUS"} };
454 return $self;
455 }
456 sub population {
458 my $self = shift;
459 if (ref $self) {
460 return ${ $self->{"_CENSUS"} };
461 } else {
462 return $Census;
463 }
464 }
465 sub DESTROY {
467 my $self = shift;
468 -- ${ $self->{"_CENSUS"} };
469 }
470 Debugging Methods
472 It's common for a class to have a debugging mechanism. For example,
473 you might want to see when objects are created or destroyed. To do
474 that, add a debugging variable as a file-scoped lexical. For this,
475 we'll pull in the standard Carp module to emit our warnings and fatal
476 messages. That way messages will come out with the caller's filename
477 and line number instead of our own; if we wanted them to be from our
478 own perspective, we'd just use die() and warn() directly instead of
479 croak() and carp() respectively.
480 use Carp;
482 my $Debugging = 0;
483 Now add a new class method to access the variable.
485 sub debug {
487 my $class = shift;
488 if (ref $class) { confess "Class method called as object method" }
489 unless (@_ == 1) { confess "usage: CLASSNAME->debug(level)" }
490 $Debugging = shift;
491 }
492 Now fix up DESTROY to murmur a bit as the moribund object expires:
494 sub DESTROY {
496 my $self = shift;
497 if ($Debugging) { carp "Destroying $self " . $self->name }
498 -- ${ $self->{"_CENSUS"} };
499 }
500 One could conceivably make a per-object debug state. That way you
502 could call both of these:
503 Person->debug(1); # entire class
505 $him->debug(1); # just this object
506 To do so, we need our debugging method to be a "bimodal" one, one that
508 works on both classes and objects. Therefore, adjust the debug() and
509 DESTROY methods as follows:
510 sub debug {
512 my $self = shift;
513 confess "usage: thing->debug(level)" unless @_ == 1;
514 my $level = shift;
515 if (ref($self)) {
516 $self->{"_DEBUG"} = $level; # just myself
517 } else {
518 $Debugging = $level; # whole class
519 }
520 }
521 sub DESTROY {
523 my $self = shift;
524 if ($Debugging || $self->{"_DEBUG"}) {
525 carp "Destroying $self " . $self->name;
526 }
527 -- ${ $self->{"_CENSUS"} };
528 }
529 What happens if a derived class (which we'll call Employee) inherits
531 methods from this Person base class? Then "Employee->debug()", when
532 called as a class method, manipulates $Person::Debugging not
533 $Employee::Debugging.
534 Class Destructors
536 The object destructor handles the death of each distinct object. But
537 sometimes you want a bit of cleanup when the entire class is shut down,
538 which currently only happens when the program exits. To make such a
539 class destructor, create a function in that class's package named END.
540 This works just like the END function in traditional modules, meaning
541 that it gets called whenever your program exits unless it execs or dies
542 of an uncaught signal. For example,
543 sub END {
545 if ($Debugging) {
546 print "All persons are going away now.\n";
547 }
548 }
549 When the program exits, all the class destructors (END functions) are
551 be called in the opposite order that they were loaded in (LIFO order).
552 Documenting the Interface
554 And there you have it: we've just shown you the implementation of this
555 Person class. Its interface would be its documentation. Usually this
556 means putting it in pod ("plain old documentation") format right there
557 in the same file. In our Person example, we would place the following
558 docs anywhere in the Person.pm file. Even though it looks mostly like
559 code, it's not. It's embedded documentation such as would be used by
560 the pod2man, pod2html, or pod2text programs. The Perl compiler ignores
561 pods entirely, just as the translators ignore code. Here's an example
562 of some pods describing the informal interface:
563 =head1 NAME
565 Person - class to implement people
567 =head1 SYNOPSIS
569 use Person;
571 #################
573 # class methods #
574 #################
575 $ob = Person->new;
576 $count = Person->population;
577 #######################
579 # object data methods #
580 #######################
581 ### get versions ###
583 $who = $ob->name;
584 $years = $ob->age;
585 @pals = $ob->peers;
586 ### set versions ###
588 $ob->name("Jason");
589 $ob->age(23);
590 $ob->peers( "Norbert", "Rhys", "Phineas" );
591 ########################
593 # other object methods #
594 ########################
595 $phrase = $ob->exclaim;
597 $ob->happy_birthday;
598 =head1 DESCRIPTION
600 The Person class implements dah dee dah dee dah....
602 That's all there is to the matter of interface versus implementation.
604 A programmer who opens up the module and plays around with all the
605 private little shiny bits that were safely locked up behind the
606 interface contract has voided the warranty, and you shouldn't worry
607 about their fate.
608
610 Suppose you later want to change the class to implement better names.
611 Perhaps you'd like to support both given names (called Christian names,
612 irrespective of one's religion) and family names (called surnames),
613 plus nicknames and titles. If users of your Person class have been
614 properly accessing it through its documented interface, then you can
615 easily change the underlying implementation. If they haven't, then
616 they lose and it's their fault for breaking the contract and voiding
617 their warranty.
618 To do this, we'll make another class, this one called Fullname. What's
620 the Fullname class look like? To answer that question, you have to
621 first figure out how you want to use it. How about we use it this way:
622 $him = Person->new();
624 $him->fullname->title("St");
625 $him->fullname->christian("Thomas");
626 $him->fullname->surname("Aquinas");
627 $him->fullname->nickname("Tommy");
628 printf "His normal name is %s\n", $him->name;
629 printf "But his real name is %s\n", $him->fullname->as_string;
630 Ok. To do this, we'll change Person::new() so that it supports a full
632 name field this way:
633 sub new {
635 my $class = shift;
636 my $self = {};
637 $self->{FULLNAME} = Fullname->new();
638 $self->{AGE} = undef;
639 $self->{PEERS} = [];
640 $self->{"_CENSUS"} = \$Census;
641 bless ($self, $class);
642 ++ ${ $self->{"_CENSUS"} };
643 return $self;
644 }
645 sub fullname {
647 my $self = shift;
648 return $self->{FULLNAME};
649 }
650 Then to support old code, define Person::name() this way:
652 sub name {
654 my $self = shift;
655 return $self->{FULLNAME}->nickname(@_)
656 || $self->{FULLNAME}->christian(@_);
657 }
658 Here's the Fullname class. We'll use the same technique of using a
660 hash reference to hold data fields, and methods by the appropriate name
661 to access them:
662 package Fullname;
664 use strict;
665 sub new {
667 my $class = shift;
668 my $self = {
669 TITLE => undef,
670 CHRISTIAN => undef,
671 SURNAME => undef,
672 NICK => undef,
673 };
674 bless ($self, $class);
675 return $self;
676 }
677 sub christian {
679 my $self = shift;
680 if (@_) { $self->{CHRISTIAN} = shift }
681 return $self->{CHRISTIAN};
682 }
683 sub surname {
685 my $self = shift;
686 if (@_) { $self->{SURNAME} = shift }
687 return $self->{SURNAME};
688 }
689 sub nickname {
691 my $self = shift;
692 if (@_) { $self->{NICK} = shift }
693 return $self->{NICK};
694 }
695 sub title {
697 my $self = shift;
698 if (@_) { $self->{TITLE} = shift }
699 return $self->{TITLE};
700 }
701 sub as_string {
703 my $self = shift;
704 my $name = join(" ", @$self{'CHRISTIAN', 'SURNAME'});
705 if ($self->{TITLE}) {
706 $name = $self->{TITLE} . " " . $name;
707 }
708 return $name;
709 }
710 1;
712 Finally, here's the test program:
714 #!/usr/bin/perl -w
716 use strict;
717 use Person;
718 sub END { show_census() }
719 sub show_census () {
721 printf "Current population: %d\n", Person->population;
722 }
723 Person->debug(1);
725 show_census();
727 my $him = Person->new();
729 $him->fullname->christian("Thomas");
731 $him->fullname->surname("Aquinas");
732 $him->fullname->nickname("Tommy");
733 $him->fullname->title("St");
734 $him->age(1);
735 printf "%s is really %s.\n", $him->name, $him->fullname->as_string;
737 printf "%s's age: %d.\n", $him->name, $him->age;
738 $him->happy_birthday;
739 printf "%s's age: %d.\n", $him->name, $him->age;
740 show_census();
742
744 Object-oriented programming systems all support some notion of
745 inheritance. Inheritance means allowing one class to piggy-back on top
746 of another one so you don't have to write the same code again and
747 again. It's about software reuse, and therefore related to Laziness,
748 the principal virtue of a programmer. (The import/export mechanisms in
749 traditional modules are also a form of code reuse, but a simpler one
750 than the true inheritance that you find in object modules.)
751 Sometimes the syntax of inheritance is built into the core of the
753 language, and sometimes it's not. Perl has no special syntax for
754 specifying the class (or classes) to inherit from. Instead, it's all
755 strictly in the semantics. Each package can have a variable called
756 @ISA, which governs (method) inheritance. If you try to call a method
757 on an object or class, and that method is not found in that object's
758 package, Perl then looks to @ISA for other packages to go looking
759 through in search of the missing method.
760 Like the special per-package variables recognized by Exporter (such as
762 @EXPORT, @EXPORT_OK, @EXPORT_FAIL, %EXPORT_TAGS, and $VERSION), the
763 @ISA array must be a package-scoped global and not a file-scoped
764 lexical created via my(). Most classes have just one item in their
765 @ISA array. In this case, we have what's called "single inheritance",
766 or SI for short.
767 Consider this class:
769 package Employee;
771 use Person;
772 @ISA = ("Person");
773 1;
774 Not a lot to it, eh? All it's doing so far is loading in another class
776 and stating that this one will inherit methods from that other class if
777 need be. We have given it none of its own methods. We rely upon an
778 Employee to behave just like a Person.
779 Setting up an empty class like this is called the "empty subclass
781 test"; that is, making a derived class that does nothing but inherit
782 from a base class. If the original base class has been designed
783 properly, then the new derived class can be used as a drop-in
784 replacement for the old one. This means you should be able to write a
785 program like this:
786 use Employee;
788 my $empl = Employee->new();
789 $empl->name("Jason");
790 $empl->age(23);
791 printf "%s is age %d.\n", $empl->name, $empl->age;
792 By proper design, we mean always using the two-argument form of
794 bless(), avoiding direct access of global data, and not exporting
795 anything. If you look back at the Person::new() function we defined
796 above, we were careful to do that. There's a bit of package data used
797 in the constructor, but the reference to this is stored on the object
798 itself and all other methods access package data via that reference, so
799 we should be ok.
800 What do we mean by the Person::new() function -- isn't that actually a
802 method? Well, in principle, yes. A method is just a function that
803 expects as its first argument a class name (package) or object (blessed
804 reference). Person::new() is the function that both the
805 "Person->new()" method and the "Employee->new()" method end up calling.
806 Understand that while a method call looks a lot like a function call,
807 they aren't really quite the same, and if you treat them as the same,
808 you'll very soon be left with nothing but broken programs. First, the
809 actual underlying calling conventions are different: method calls get
810 an extra argument. Second, function calls don't do inheritance, but
811 methods do.
812 Method Call Resulting Function Call
814 ----------- ------------------------
815 Person->new() Person::new("Person")
816 Employee->new() Person::new("Employee")
817 So don't use function calls when you mean to call a method.
819 If an employee is just a Person, that's not all too very interesting.
821 So let's add some other methods. We'll give our employee data fields
822 to access their salary, their employee ID, and their start date.
823 If you're getting a little tired of creating all these nearly identical
825 methods just to get at the object's data, do not despair. Later, we'll
826 describe several different convenience mechanisms for shortening this
827 up. Meanwhile, here's the straight-forward way:
828 sub salary {
830 my $self = shift;
831 if (@_) { $self->{SALARY} = shift }
832 return $self->{SALARY};
833 }
834 sub id_number {
836 my $self = shift;
837 if (@_) { $self->{ID} = shift }
838 return $self->{ID};
839 }
840 sub start_date {
842 my $self = shift;
843 if (@_) { $self->{START_DATE} = shift }
844 return $self->{START_DATE};
845 }
846 Overridden Methods
848 What happens when both a derived class and its base class have the same
849 method defined? Well, then you get the derived class's version of that
850 method. For example, let's say that we want the peers() method called
851 on an employee to act a bit differently. Instead of just returning the
852 list of peer names, let's return slightly different strings. So doing
853 this:
854 $empl->peers("Peter", "Paul", "Mary");
856 printf "His peers are: %s\n", join(", ", $empl->peers);
857 will produce:
859 His peers are: PEON=PETER, PEON=PAUL, PEON=MARY
861 To do this, merely add this definition into the Employee.pm file:
863 sub peers {
865 my $self = shift;
866 if (@_) { @{ $self->{PEERS} } = @_ }
867 return map { "PEON=\U$_" } @{ $self->{PEERS} };
868 }
869 There, we've just demonstrated the high-falutin' concept known in
871 certain circles as polymorphism. We've taken on the form and behaviour
872 of an existing object, and then we've altered it to suit our own
873 purposes. This is a form of Laziness. (Getting polymorphed is also
874 what happens when the wizard decides you'd look better as a frog.)
875 Every now and then you'll want to have a method call trigger both its
877 derived class (also known as "subclass") version as well as its base
878 class (also known as "superclass") version. In practice, constructors
879 and destructors are likely to want to do this, and it probably also
880 makes sense in the debug() method we showed previously.
881 To do this, add this to Employee.pm:
883 use Carp;
885 my $Debugging = 0;
886 sub debug {
888 my $self = shift;
889 confess "usage: thing->debug(level)" unless @_ == 1;
890 my $level = shift;
891 if (ref($self)) {
892 $self->{"_DEBUG"} = $level;
893 } else {
894 $Debugging = $level; # whole class
895 }
896 Person::debug($self, $Debugging); # don't really do this
897 }
898 As you see, we turn around and call the Person package's debug()
900 function. But this is far too fragile for good design. What if Person
901 doesn't have a debug() function, but is inheriting its debug() method
902 from elsewhere? It would have been slightly better to say
903 Person->debug($Debugging);
905 But even that's got too much hard-coded. It's somewhat better to say
907 $self->Person::debug($Debugging);
909 Which is a funny way to say to start looking for a debug() method up in
911 Person. This strategy is more often seen on overridden object methods
912 than on overridden class methods.
913 There is still something a bit off here. We've hard-coded our
915 superclass's name. This in particular is bad if you change which
916 classes you inherit from, or add others. Fortunately, the pseudoclass
917 SUPER comes to the rescue here.
918 $self->SUPER::debug($Debugging);
920 This way it starts looking in my class's @ISA. This only makes sense
922 from within a method call, though. Don't try to access anything in
923 SUPER:: from anywhere else, because it doesn't exist outside an
924 overridden method call. Note that "SUPER" refers to the superclass of
925 the current package, not to the superclass of $self.
926 Things are getting a bit complicated here. Have we done anything we
928 shouldn't? As before, one way to test whether we're designing a decent
929 class is via the empty subclass test. Since we already have an
930 Employee class that we're trying to check, we'd better get a new empty
931 subclass that can derive from Employee. Here's one:
932 package Boss;
934 use Employee; # :-)
935 @ISA = qw(Employee);
936 And here's the test program:
938 #!/usr/bin/perl -w
940 use strict;
941 use Boss;
942 Boss->debug(1);
943 my $boss = Boss->new();
945 $boss->fullname->title("Don");
947 $boss->fullname->surname("Pichon Alvarez");
948 $boss->fullname->christian("Federico Jesus");
949 $boss->fullname->nickname("Fred");
950 $boss->age(47);
952 $boss->peers("Frank", "Felipe", "Faust");
953 printf "%s is age %d.\n", $boss->fullname->as_string, $boss->age;
955 printf "His peers are: %s\n", join(", ", $boss->peers);
956 Running it, we see that we're still ok. If you'd like to dump out your
958 object in a nice format, somewhat like the way the 'x' command works in
959 the debugger, you could use the Data::Dumper module from CPAN this way:
960 use Data::Dumper;
962 print "Here's the boss:\n";
963 print Dumper($boss);
964 Which shows us something like this:
966 Here's the boss:
968 $VAR1 = bless( {
969 _CENSUS => \1,
970 FULLNAME => bless( {
971 TITLE => 'Don',
972 SURNAME => 'Pichon Alvarez',
973 NICK => 'Fred',
974 CHRISTIAN => 'Federico Jesus'
975 }, 'Fullname' ),
976 AGE => 47,
977 PEERS => [
978 'Frank',
979 'Felipe',
980 'Faust'
981 ]
982 }, 'Boss' );
983 Hm.... something's missing there. What about the salary, start date,
985 and ID fields? Well, we never set them to anything, even undef, so
986 they don't show up in the hash's keys. The Employee class has no new()
987 method of its own, and the new() method in Person doesn't know about
988 Employees. (Nor should it: proper OO design dictates that a subclass
989 be allowed to know about its immediate superclass, but never vice-
990 versa.) So let's fix up Employee::new() this way:
991 sub new {
993 my $class = shift;
994 my $self = $class->SUPER::new();
995 $self->{SALARY} = undef;
996 $self->{ID} = undef;
997 $self->{START_DATE} = undef;
998 bless ($self, $class); # reconsecrate
999 return $self;
1000 }
1001 Now if you dump out an Employee or Boss object, you'll find that new
1003 fields show up there now.
1004 Multiple Inheritance
1006 Ok, at the risk of confusing beginners and annoying OO gurus, it's time
1007 to confess that Perl's object system includes that controversial notion
1008 known as multiple inheritance, or MI for short. All this means is that
1009 rather than having just one parent class who in turn might itself have
1010 a parent class, etc., that you can directly inherit from two or more
1011 parents. It's true that some uses of MI can get you into trouble,
1012 although hopefully not quite so much trouble with Perl as with
1013 dubiously-OO languages like C++.
1014 The way it works is actually pretty simple: just put more than one
1016 package name in your @ISA array. When it comes time for Perl to go
1017 finding methods for your object, it looks at each of these packages in
1018 order. Well, kinda. It's actually a fully recursive, depth-first
1019 order by default (see mro for alternate method resolution orders).
1020 Consider a bunch of @ISA arrays like this:
1021 @First::ISA = qw( Alpha );
1023 @Second::ISA = qw( Beta );
1024 @Third::ISA = qw( First Second );
1025 If you have an object of class Third:
1027 my $ob = Third->new();
1029 $ob->spin();
1030 How do we find a spin() method (or a new() method for that matter)?
1032 Because the search is depth-first, classes will be looked up in the
1033 following order: Third, First, Alpha, Second, and Beta.
1034 In practice, few class modules have been seen that actually make use of
1036 MI. One nearly always chooses simple containership of one class within
1037 another over MI. That's why our Person object contained a Fullname
1038 object. That doesn't mean it was one.
1039 However, there is one particular area where MI in Perl is rampant:
1041 borrowing another class's class methods. This is rather common,
1042 especially with some bundled "objectless" classes, like Exporter,
1043 DynaLoader, AutoLoader, and SelfLoader. These classes do not provide
1044 constructors; they exist only so you may inherit their class methods.
1045 (It's not entirely clear why inheritance was done here rather than
1046 traditional module importation.)
1047 For example, here is the POSIX module's @ISA:
1049 package POSIX;
1051 @ISA = qw(Exporter DynaLoader);
1052 The POSIX module isn't really an object module, but then, neither are
1054 Exporter or DynaLoader. They're just lending their classes' behaviours
1055 to POSIX.
1056 Why don't people use MI for object methods much? One reason is that it
1058 can have complicated side-effects. For one thing, your inheritance
1059 graph (no longer a tree) might converge back to the same base class.
1060 Although Perl guards against recursive inheritance, merely having
1061 parents who are related to each other via a common ancestor, incestuous
1062 though it sounds, is not forbidden. What if in our Third class shown
1063 above we wanted its new() method to also call both overridden
1064 constructors in its two parent classes? The SUPER notation would only
1065 find the first one. Also, what about if the Alpha and Beta classes
1066 both had a common ancestor, like Nought? If you kept climbing up the
1067 inheritance tree calling overridden methods, you'd end up calling
1068 Nought::new() twice, which might well be a bad idea.
1069 UNIVERSAL: The Root of All Objects
1071 Wouldn't it be convenient if all objects were rooted at some ultimate
1072 base class? That way you could give every object common methods
1073 without having to go and add it to each and every @ISA. Well, it turns
1074 out that you can. You don't see it, but Perl tacitly and irrevocably
1075 assumes that there's an extra element at the end of @ISA: the class
1076 UNIVERSAL. In version 5.003, there were no predefined methods there,
1077 but you could put whatever you felt like into it.
1078 However, as of version 5.004 (or some subversive releases, like
1080 5.003_08), UNIVERSAL has some methods in it already. These are builtin
1081 to your Perl binary, so they don't take any extra time to load.
1082 Predefined methods include isa(), can(), and VERSION(). isa() tells
1083 you whether an object or class "is" another one without having to
1084 traverse the hierarchy yourself:
1085 $has_io = $fd->isa("IO::Handle");
1087 $itza_handle = IO::Socket->isa("IO::Handle");
1088 The can() method, called against that object or class, reports back
1090 whether its string argument is a callable method name in that class.
1091 In fact, it gives you back a function reference to that method:
1092 $his_print_method = $obj->can('as_string');
1094 Finally, the VERSION method checks whether the class (or the object's
1096 class) has a package global called $VERSION that's high enough, as in:
1097 Some_Module->VERSION(3.0);
1099 $his_vers = $ob->VERSION();
1100 However, we don't usually call VERSION ourselves. (Remember that an
1102 all uppercase function name is a Perl convention that indicates that
1103 the function will be automatically used by Perl in some way.) In this
1104 case, it happens when you say
1105 use Some_Module 3.0;
1107 If you wanted to add version checking to your Person class explained
1109 above, just add this to Person.pm:
1110 our $VERSION = '1.1';
1112 and then in Employee.pm you can say
1114 use Person 1.1;
1116 And it would make sure that you have at least that version number or
1118 higher available. This is not the same as loading in that exact
1119 version number. No mechanism currently exists for concurrent
1120 installation of multiple versions of a module. Lamentably.
1121 Deeper UNIVERSAL details
1123 It is also valid (though perhaps unwise in most cases) to put other
1124 packages' names in @UNIVERSAL::ISA. These packages will also be
1125 implicitly inherited by all classes, just as UNIVERSAL itself is.
1126 However, neither UNIVERSAL nor any of its parents from the @ISA tree
1127 are explicit base classes of all objects. To clarify, given the
1128 following:
1129 @UNIVERSAL::ISA = ('REALLYUNIVERSAL');
1131 package REALLYUNIVERSAL;
1133 sub special_method { return "123" }
1134 package Foo;
1136 sub normal_method { return "321" }
1137 Calling Foo->special_method() will return "123", but calling
1139 Foo->isa('REALLYUNIVERSAL') or Foo->isa('UNIVERSAL') will return false.
1140 If your class is using an alternate mro like C3 (see mro), method
1142 resolution within UNIVERSAL / @UNIVERSAL::ISA will still occur in the
1143 default depth-first left-to-right manner, after the class's C3 mro is
1144 exhausted.
1145 All of the above is made more intuitive by realizing what really
1147 happens during method lookup, which is roughly like this ugly pseudo-
1148 code:
1149 get_mro(class) {
1151 # recurses down the @ISA's starting at class,
1152 # builds a single linear array of all
1153 # classes to search in the appropriate order.
1154 # The method resolution order (mro) to use
1155 # for the ordering is whichever mro "class"
1156 # has set on it (either default (depth first
1157 # l-to-r) or C3 ordering).
1158 # The first entry in the list is the class
1159 # itself.
1160 }
1161 find_method(class, methname) {
1163 foreach $class (get_mro(class)) {
1164 if($class->has_method(methname)) {
1165 return ref_to($class->$methname);
1166 }
1167 }
1168 foreach $class (get_mro(UNIVERSAL)) {
1169 if($class->has_method(methname)) {
1170 return ref_to($class->$methname);
1171 }
1172 }
1173 return undef;
1174 }
1175 However the code that implements UNIVERSAL::isa does not search in
1177 UNIVERSAL itself, only in the package's actual @ISA.
1178
1180 Nothing requires objects to be implemented as hash references. An
1181 object can be any sort of reference so long as its referent has been
1182 suitably blessed. That means scalar, array, and code references are
1183 also fair game.
1184 A scalar would work if the object has only one datum to hold. An array
1186 would work for most cases, but makes inheritance a bit dodgy because
1187 you have to invent new indices for the derived classes.
1188 Arrays as Objects
1190 If the user of your class honors the contract and sticks to the
1191 advertised interface, then you can change its underlying interface if
1192 you feel like it. Here's another implementation that conforms to the
1193 same interface specification. This time we'll use an array reference
1194 instead of a hash reference to represent the object.
1195 package Person;
1197 use strict;
1198 my($NAME, $AGE, $PEERS) = ( 0 .. 2 );
1200 ############################################
1202 ## the object constructor (array version) ##
1203 ############################################
1204 sub new {
1205 my $self = [];
1206 $self->[$NAME] = undef; # this is unnecessary
1207 $self->[$AGE] = undef; # as is this
1208 $self->[$PEERS] = []; # but this isn't, really
1209 bless($self);
1210 return $self;
1211 }
1212 sub name {
1214 my $self = shift;
1215 if (@_) { $self->[$NAME] = shift }
1216 return $self->[$NAME];
1217 }
1218 sub age {
1220 my $self = shift;
1221 if (@_) { $self->[$AGE] = shift }
1222 return $self->[$AGE];
1223 }
1224 sub peers {
1226 my $self = shift;
1227 if (@_) { @{ $self->[$PEERS] } = @_ }
1228 return @{ $self->[$PEERS] };
1229 }
1230 1; # so the require or use succeeds
1232 You might guess that the array access would be a lot faster than the
1234 hash access, but they're actually comparable. The array is a little
1235 bit faster, but not more than ten or fifteen percent, even when you
1236 replace the variables above like $AGE with literal numbers, like 1. A
1237 bigger difference between the two approaches can be found in memory
1238 use. A hash representation takes up more memory than an array
1239 representation because you have to allocate memory for the keys as well
1240 as for the values. However, it really isn't that bad, especially since
1241 as of version 5.004, memory is only allocated once for a given hash
1242 key, no matter how many hashes have that key. It's expected that
1243 sometime in the future, even these differences will fade into obscurity
1244 as more efficient underlying representations are devised.
1245 Still, the tiny edge in speed (and somewhat larger one in memory) is
1247 enough to make some programmers choose an array representation for
1248 simple classes. There's still a little problem with scalability,
1249 though, because later in life when you feel like creating subclasses,
1250 you'll find that hashes just work out better.
1251 Closures as Objects
1253 Using a code reference to represent an object offers some fascinating
1254 possibilities. We can create a new anonymous function (closure) who
1255 alone in all the world can see the object's data. This is because we
1256 put the data into an anonymous hash that's lexically visible only to
1257 the closure we create, bless, and return as the object. This object's
1258 methods turn around and call the closure as a regular subroutine call,
1259 passing it the field we want to affect. (Yes, the double-function call
1260 is slow, but if you wanted fast, you wouldn't be using objects at all,
1261 eh? :-)
1262 Use would be similar to before:
1264 use Person;
1266 $him = Person->new();
1267 $him->name("Jason");
1268 $him->age(23);
1269 $him->peers( [ "Norbert", "Rhys", "Phineas" ] );
1270 printf "%s is %d years old.\n", $him->name, $him->age;
1271 print "His peers are: ", join(", ", @{$him->peers}), "\n";
1272 but the implementation would be radically, perhaps even sublimely
1274 different:
1275 package Person;
1277 sub new {
1279 my $class = shift;
1280 my $self = {
1281 NAME => undef,
1282 AGE => undef,
1283 PEERS => [],
1284 };
1285 my $closure = sub {
1286 my $field = shift;
1287 if (@_) { $self->{$field} = shift }
1288 return $self->{$field};
1289 };
1290 bless($closure, $class);
1291 return $closure;
1292 }
1293 sub name { &{ $_[0] }("NAME", @_[ 1 .. $#_ ] ) }
1295 sub age { &{ $_[0] }("AGE", @_[ 1 .. $#_ ] ) }
1296 sub peers { &{ $_[0] }("PEERS", @_[ 1 .. $#_ ] ) }
1297 1;
1299 Because this object is hidden behind a code reference, it's probably a
1301 bit mysterious to those whose background is more firmly rooted in
1302 standard procedural or object-based programming languages than in
1303 functional programming languages whence closures derive. The object
1304 created and returned by the new() method is itself not a data reference
1305 as we've seen before. It's an anonymous code reference that has within
1306 it access to a specific version (lexical binding and instantiation) of
1307 the object's data, which are stored in the private variable $self.
1308 Although this is the same function each time, it contains a different
1309 version of $self.
1310 When a method like "$him->name("Jason")" is called, its implicit zeroth
1312 argument is the invoking object--just as it is with all method calls.
1313 But in this case, it's our code reference (something like a function
1314 pointer in C++, but with deep binding of lexical variables). There's
1315 not a lot to be done with a code reference beyond calling it, so that's
1316 just what we do when we say "&{$_[0]}". This is just a regular
1317 function call, not a method call. The initial argument is the string
1318 "NAME", and any remaining arguments are whatever had been passed to the
1319 method itself.
1320 Once we're executing inside the closure that had been created in new(),
1322 the $self hash reference suddenly becomes visible. The closure grabs
1323 its first argument ("NAME" in this case because that's what the name()
1324 method passed it), and uses that string to subscript into the private
1325 hash hidden in its unique version of $self.
1326 Nothing under the sun will allow anyone outside the executing method to
1328 be able to get at this hidden data. Well, nearly nothing. You could
1329 single step through the program using the debugger and find out the
1330 pieces while you're in the method, but everyone else is out of luck.
1331 There, if that doesn't excite the Scheme folks, then I just don't know
1333 what will. Translation of this technique into C++, Java, or any other
1334 braindead-static language is left as a futile exercise for aficionados
1335 of those camps.
1336 You could even add a bit of nosiness via the caller() function and make
1338 the closure refuse to operate unless called via its own package. This
1339 would no doubt satisfy certain fastidious concerns of programming
1340 police and related puritans.
1341 If you were wondering when Hubris, the third principle virtue of a
1343 programmer, would come into play, here you have it. (More seriously,
1344 Hubris is just the pride in craftsmanship that comes from having
1345 written a sound bit of well-designed code.)
1346
1348 Autoloading is a way to intercept calls to undefined methods. An
1349 autoload routine may choose to create a new function on the fly, either
1350 loaded from disk or perhaps just eval()ed right there. This define-on-
1351 the-fly strategy is why it's called autoloading.
1352 But that's only one possible approach. Another one is to just have the
1354 autoloaded method itself directly provide the requested service. When
1355 used in this way, you may think of autoloaded methods as "proxy"
1356 methods.
1357 When Perl tries to call an undefined function in a particular package
1359 and that function is not defined, it looks for a function in that same
1360 package called AUTOLOAD. If one exists, it's called with the same
1361 arguments as the original function would have had. The fully-qualified
1362 name of the function is stored in that package's global variable
1363 $AUTOLOAD. Once called, the function can do anything it would like,
1364 including defining a new function by the right name, and then doing a
1365 really fancy kind of "goto" right to it, erasing itself from the call
1366 stack.
1367 What does this have to do with objects? After all, we keep talking
1369 about functions, not methods. Well, since a method is just a function
1370 with an extra argument and some fancier semantics about where it's
1371 found, we can use autoloading for methods, too. Perl doesn't start
1372 looking for an AUTOLOAD method until it has exhausted the recursive
1373 hunt up through @ISA, though. Some programmers have even been known to
1374 define a UNIVERSAL::AUTOLOAD method to trap unresolved method calls to
1375 any kind of object.
1376 Autoloaded Data Methods
1378 You probably began to get a little suspicious about the duplicated code
1379 way back earlier when we first showed you the Person class, and then
1380 later the Employee class. Each method used to access the hash fields
1381 looked virtually identical. This should have tickled that great
1382 programming virtue, Impatience, but for the time, we let Laziness win
1383 out, and so did nothing. Proxy methods can cure this.
1384 Instead of writing a new function every time we want a new data field,
1386 we'll use the autoload mechanism to generate (actually, mimic) methods
1387 on the fly. To verify that we're accessing a valid member, we will
1388 check against an "_permitted" (pronounced "under-permitted") field,
1389 which is a reference to a file-scoped lexical (like a C file static)
1390 hash of permitted fields in this record called %fields. Why the
1391 underscore? For the same reason as the _CENSUS field we once used: as
1392 a marker that means "for internal use only".
1393 Here's what the module initialization code and class constructor will
1395 look like when taking this approach:
1396 package Person;
1398 use Carp;
1399 our $AUTOLOAD; # it's a package global
1400 my %fields = (
1402 name => undef,
1403 age => undef,
1404 peers => undef,
1405 );
1406 sub new {
1408 my $class = shift;
1409 my $self = {
1410 _permitted => \%fields,
1411 %fields,
1412 };
1413 bless $self, $class;
1414 return $self;
1415 }
1416 If we wanted our record to have default values, we could fill those in
1418 where current we have "undef" in the %fields hash.
1419 Notice how we saved a reference to our class data on the object itself?
1421 Remember that it's important to access class data through the object
1422 itself instead of having any method reference %fields directly, or else
1423 you won't have a decent inheritance.
1424 The real magic, though, is going to reside in our proxy method, which
1426 will handle all calls to undefined methods for objects of class Person
1427 (or subclasses of Person). It has to be called AUTOLOAD. Again, it's
1428 all caps because it's called for us implicitly by Perl itself, not by a
1429 user directly.
1430 sub AUTOLOAD {
1432 my $self = shift;
1433 my $type = ref($self)
1434 or croak "$self is not an object";
1435 my $name = $AUTOLOAD;
1437 $name =~ s/.*://; # strip fully-qualified portion
1438 unless (exists $self->{_permitted}->{$name} ) {
1440 croak "Can't access `$name' field in class $type";
1441 }
1442 if (@_) {
1444 return $self->{$name} = shift;
1445 } else {
1446 return $self->{$name};
1447 }
1448 }
1449 Pretty nifty, eh? All we have to do to add new data fields is modify
1451 %fields. No new functions need be written.
1452 I could have avoided the "_permitted" field entirely, but I wanted to
1454 demonstrate how to store a reference to class data on the object so you
1455 wouldn't have to access that class data directly from an object method.
1456 Inherited Autoloaded Data Methods
1458 But what about inheritance? Can we define our Employee class
1459 similarly? Yes, so long as we're careful enough.
1460 Here's how to be careful:
1462 package Employee;
1464 use Person;
1465 use strict;
1466 our @ISA = qw(Person);
1467 my %fields = (
1469 id => undef,
1470 salary => undef,
1471 );
1472 sub new {
1474 my $class = shift;
1475 my $self = $class->SUPER::new();
1476 my($element);
1477 foreach $element (keys %fields) {
1478 $self->{_permitted}->{$element} = $fields{$element};
1479 }
1480 @{$self}{keys %fields} = values %fields;
1481 return $self;
1482 }
1483 Once we've done this, we don't even need to have an AUTOLOAD function
1485 in the Employee package, because we'll grab Person's version of that
1486 via inheritance, and it will all work out just fine.
1487
1489 Even though proxy methods can provide a more convenient approach to
1490 making more struct-like classes than tediously coding up data methods
1491 as functions, it still leaves a bit to be desired. For one thing, it
1492 means you have to handle bogus calls that you don't mean to trap via
1493 your proxy. It also means you have to be quite careful when dealing
1494 with inheritance, as detailed above.
1495 Perl programmers have responded to this by creating several different
1497 class construction classes. These metaclasses are classes that create
1498 other classes. A couple worth looking at are Class::Struct and Alias.
1499 These and other related metaclasses can be found in the modules
1500 directory on CPAN.
1501 Class::Struct
1503 One of the older ones is Class::Struct. In fact, its syntax and
1504 interface were sketched out long before perl5 even solidified into a
1505 real thing. What it does is provide you a way to "declare" a class as
1506 having objects whose fields are of a specific type. The function that
1507 does this is called, not surprisingly enough, struct(). Because
1508 structures or records are not base types in Perl, each time you want to
1509 create a class to provide a record-like data object, you yourself have
1510 to define a new() method, plus separate data-access methods for each of
1511 that record's fields. You'll quickly become bored with this process.
1512 The Class::Struct::struct() function alleviates this tedium.
1513 Here's a simple example of using it:
1515 use Class::Struct qw(struct);
1517 use Jobbie; # user-defined; see below
1518 struct 'Fred' => {
1520 one => '$',
1521 many => '@',
1522 profession => 'Jobbie', # does not call Jobbie->new()
1523 };
1524 $ob = Fred->new(profession => Jobbie->new());
1526 $ob->one("hmmmm");
1527 $ob->many(0, "here");
1529 $ob->many(1, "you");
1530 $ob->many(2, "go");
1531 print "Just set: ", $ob->many(2), "\n";
1532 $ob->profession->salary(10_000);
1534 You can declare types in the struct to be basic Perl types, or user-
1536 defined types (classes). User types will be initialized by calling
1537 that class's new() method.
1538 Take care that the "Jobbie" object is not created automatically by the
1540 "Fred" class's new() method, so you should specify a "Jobbie" object
1541 when you create an instance of "Fred".
1542 Here's a real-world example of using struct generation. Let's say you
1544 wanted to override Perl's idea of gethostbyname() and gethostbyaddr()
1545 so that they would return objects that acted like C structures. We
1546 don't care about high-falutin' OO gunk. All we want is for these
1547 objects to act like structs in the C sense.
1548 use Socket;
1550 use Net::hostent;
1551 $h = gethostbyname("perl.com"); # object return
1552 printf "perl.com's real name is %s, address %s\n",
1553 $h->name, inet_ntoa($h->addr);
1554 Here's how to do this using the Class::Struct module. The crux is
1556 going to be this call:
1557 struct 'Net::hostent' => [ # note bracket
1559 name => '$',
1560 aliases => '@',
1561 addrtype => '$',
1562 'length' => '$',
1563 addr_list => '@',
1564 ];
1565 Which creates object methods of those names and types. It even creates
1567 a new() method for us.
1568 We could also have implemented our object this way:
1570 struct 'Net::hostent' => { # note brace
1572 name => '$',
1573 aliases => '@',
1574 addrtype => '$',
1575 'length' => '$',
1576 addr_list => '@',
1577 };
1578 and then Class::Struct would have used an anonymous hash as the object
1580 type, instead of an anonymous array. The array is faster and smaller,
1581 but the hash works out better if you eventually want to do inheritance.
1582 Since for this struct-like object we aren't planning on inheritance,
1583 this time we'll opt for better speed and size over better flexibility.
1584 Here's the whole implementation:
1586 package Net::hostent;
1588 use strict;
1589 BEGIN {
1591 use Exporter ();
1592 our @EXPORT = qw(gethostbyname gethostbyaddr gethost);
1593 our @EXPORT_OK = qw(
1594 $h_name @h_aliases
1595 $h_addrtype $h_length
1596 @h_addr_list $h_addr
1597 );
1598 our %EXPORT_TAGS = ( FIELDS => [ @EXPORT_OK, @EXPORT ] );
1599 }
1600 our @EXPORT_OK;
1601 # Class::Struct forbids use of @ISA
1603 sub import { goto &Exporter::import }
1604 use Class::Struct qw(struct);
1606 struct 'Net::hostent' => [
1607 name => '$',
1608 aliases => '@',
1609 addrtype => '$',
1610 'length' => '$',
1611 addr_list => '@',
1612 ];
1613 sub addr { shift->addr_list->[0] }
1615 sub populate (@) {
1617 return unless @_;
1618 my $hob = new(); # Class::Struct made this!
1619 $h_name = $hob->[0] = $_[0];
1620 @h_aliases = @{ $hob->[1] } = split ' ', $_[1];
1621 $h_addrtype = $hob->[2] = $_[2];
1622 $h_length = $hob->[3] = $_[3];
1623 $h_addr = $_[4];
1624 @h_addr_list = @{ $hob->[4] } = @_[ (4 .. $#_) ];
1625 return $hob;
1626 }
1627 sub gethostbyname ($) { populate(CORE::gethostbyname(shift)) }
1629 sub gethostbyaddr ($;$) {
1631 my ($addr, $addrtype);
1632 $addr = shift;
1633 require Socket unless @_;
1634 $addrtype = @_ ? shift : Socket::AF_INET();
1635 populate(CORE::gethostbyaddr($addr, $addrtype))
1636 }
1637 sub gethost($) {
1639 if ($_[0] =~ /^\d+(?:\.\d+(?:\.\d+(?:\.\d+)?)?)?$/) {
1640 require Socket;
1641 &gethostbyaddr(Socket::inet_aton(shift));
1642 } else {
1643 &gethostbyname;
1644 }
1645 }
1646 1;
1648 We've snuck in quite a fair bit of other concepts besides just dynamic
1650 class creation, like overriding core functions, import/export bits,
1651 function prototyping, short-cut function call via &whatever, and
1652 function replacement with "goto &whatever". These all mostly make
1653 sense from the perspective of a traditional module, but as you can see,
1654 we can also use them in an object module.
1655 You can look at other object-based, struct-like overrides of core
1657 functions in the 5.004 release of Perl in File::stat, Net::hostent,
1658 Net::netent, Net::protoent, Net::servent, Time::gmtime,
1659 Time::localtime, User::grent, and User::pwent. These modules have a
1660 final component that's all lowercase, by convention reserved for
1661 compiler pragmas, because they affect the compilation and change a
1662 builtin function. They also have the type names that a C programmer
1663 would most expect.
1664 Data Members as Variables
1666 If you're used to C++ objects, then you're accustomed to being able to
1667 get at an object's data members as simple variables from within a
1668 method. The Alias module provides for this, as well as a good bit
1669 more, such as the possibility of private methods that the object can
1670 call but folks outside the class cannot.
1671 Here's an example of creating a Person using the Alias module. When
1673 you update these magical instance variables, you automatically update
1674 value fields in the hash. Convenient, eh?
1675 package Person;
1677 # this is the same as before...
1679 sub new {
1680 my $class = shift;
1681 my $self = {
1682 NAME => undef,
1683 AGE => undef,
1684 PEERS => [],
1685 };
1686 bless($self, $class);
1687 return $self;
1688 }
1689 use Alias qw(attr);
1691 our ($NAME, $AGE, $PEERS);
1692 sub name {
1694 my $self = attr shift;
1695 if (@_) { $NAME = shift; }
1696 return $NAME;
1697 }
1698 sub age {
1700 my $self = attr shift;
1701 if (@_) { $AGE = shift; }
1702 return $AGE;
1703 }
1704 sub peers {
1706 my $self = attr shift;
1707 if (@_) { @PEERS = @_; }
1708 return @PEERS;
1709 }
1710 sub exclaim {
1712 my $self = attr shift;
1713 return sprintf "Hi, I'm %s, age %d, working with %s",
1714 $NAME, $AGE, join(", ", @PEERS);
1715 }
1716 sub happy_birthday {
1718 my $self = attr shift;
1719 return ++$AGE;
1720 }
1721 The need for the "our" declaration is because what Alias does is play
1723 with package globals with the same name as the fields. To use globals
1724 while "use strict" is in effect, you have to predeclare them. These
1725 package variables are localized to the block enclosing the attr() call
1726 just as if you'd used a local() on them. However, that means that
1727 they're still considered global variables with temporary values, just
1728 as with any other local().
1729 It would be nice to combine Alias with something like Class::Struct or
1731 Class::MethodMaker.
1732
1734 Object Terminology
1735 In the various OO literature, it seems that a lot of different words
1736 are used to describe only a few different concepts. If you're not
1737 already an object programmer, then you don't need to worry about all
1738 these fancy words. But if you are, then you might like to know how to
1739 get at the same concepts in Perl.
1740 For example, it's common to call an object an instance of a class and
1742 to call those objects' methods instance methods. Data fields peculiar
1743 to each object are often called instance data or object attributes, and
1744 data fields common to all members of that class are class data, class
1745 attributes, or static data members.
1746 Also, base class, generic class, and superclass all describe the same
1748 notion, whereas derived class, specific class, and subclass describe
1749 the other related one.
1750 C++ programmers have static methods and virtual methods, but Perl only
1752 has class methods and object methods. Actually, Perl only has methods.
1753 Whether a method gets used as a class or object method is by usage
1754 only. You could accidentally call a class method (one expecting a
1755 string argument) on an object (one expecting a reference), or vice
1756 versa.
1757 From the C++ perspective, all methods in Perl are virtual. This, by
1759 the way, is why they are never checked for function prototypes in the
1760 argument list as regular builtin and user-defined functions can be.
1761 Because a class is itself something of an object, Perl's classes can be
1763 taken as describing both a "class as meta-object" (also called object
1764 factory) philosophy and the "class as type definition" (declaring
1765 behaviour, not defining mechanism) idea. C++ supports the latter
1766 notion, but not the former.
1767
1769 The following manpages will doubtless provide more background for this
1770 one: perlmod, perlref, perlobj, perlbot, perltie, and overload.
1771 perlboot is a kinder, gentler introduction to object-oriented
1773 programming.
1774 perltooc provides more detail on class data.
1776 Some modules which might prove interesting are Class::Accessor,
1778 Class::Class, Class::Contract, Class::Data::Inheritable,
1779 Class::MethodMaker and Tie::SecureHash
1780
1782 Copyright (c) 1997, 1998 Tom Christiansen All rights reserved.
1783 This documentation is free; you can redistribute it and/or modify it
1785 under the same terms as Perl itself.
1786 Irrespective of its distribution, all code examples in this file are
1788 hereby placed into the public domain. You are permitted and encouraged
1789 to use this code in your own programs for fun or for profit as you see
1790 fit. A simple comment in the code giving credit would be courteous but
1791 is not required.
1792
1794 Acknowledgments
1795 Thanks to Larry Wall, Roderick Schertler, Gurusamy Sarathy, Dean
1796 Roehrich, Raphael Manfredi, Brent Halsey, Greg Bacon, Brad Appleton,
1797 and many others for their helpful comments.
1798perl v5.10.1 2009-02-12 PERLTOOT(1)