1PERLOOTUT(1) Perl Programmers Reference Guide PERLOOTUT(1)
2
3
4
6 perlootut - Object-Oriented Programming in Perl Tutorial
7
9 This document was created in February, 2011, and the last major
10 revision was in February, 2013.
11
12 If you are reading this in the future then it's possible that the state
13 of the art has changed. We recommend you start by reading the perlootut
14 document in the latest stable release of Perl, rather than this
15 version.
16
18 This document provides an introduction to object-oriented programming
19 in Perl. It begins with a brief overview of the concepts behind object
20 oriented design. Then it introduces several different OO systems from
21 CPAN <https://www.cpan.org> which build on top of what Perl provides.
22
23 By default, Perl's built-in OO system is very minimal, leaving you to
24 do most of the work. This minimalism made a lot of sense in 1994, but
25 in the years since Perl 5.0 we've seen a number of common patterns
26 emerge in Perl OO. Fortunately, Perl's flexibility has allowed a rich
27 ecosystem of Perl OO systems to flourish.
28
29 If you want to know how Perl OO works under the hood, the perlobj
30 document explains the nitty gritty details.
31
32 This document assumes that you already understand the basics of Perl
33 syntax, variable types, operators, and subroutine calls. If you don't
34 understand these concepts yet, please read perlintro first. You should
35 also read the perlsyn, perlop, and perlsub documents.
36
38 Most object systems share a number of common concepts. You've probably
39 heard terms like "class", "object, "method", and "attribute" before.
40 Understanding the concepts will make it much easier to read and write
41 object-oriented code. If you're already familiar with these terms, you
42 should still skim this section, since it explains each concept in terms
43 of Perl's OO implementation.
44
45 Perl's OO system is class-based. Class-based OO is fairly common. It's
46 used by Java, C++, C#, Python, Ruby, and many other languages. There
47 are other object orientation paradigms as well. JavaScript is the most
48 popular language to use another paradigm. JavaScript's OO system is
49 prototype-based.
50
51 Object
52 An object is a data structure that bundles together data and
53 subroutines which operate on that data. An object's data is called
54 attributes, and its subroutines are called methods. An object can be
55 thought of as a noun (a person, a web service, a computer).
56
57 An object represents a single discrete thing. For example, an object
58 might represent a file. The attributes for a file object might include
59 its path, content, and last modification time. If we created an object
60 to represent /etc/hostname on a machine named "foo.example.com", that
61 object's path would be "/etc/hostname", its content would be "foo\n",
62 and it's last modification time would be 1304974868 seconds since the
63 beginning of the epoch.
64
65 The methods associated with a file might include "rename()" and
66 "write()".
67
68 In Perl most objects are hashes, but the OO systems we recommend keep
69 you from having to worry about this. In practice, it's best to consider
70 an object's internal data structure opaque.
71
72 Class
73 A class defines the behavior of a category of objects. A class is a
74 name for a category (like "File"), and a class also defines the
75 behavior of objects in that category.
76
77 All objects belong to a specific class. For example, our /etc/hostname
78 object belongs to the "File" class. When we want to create a specific
79 object, we start with its class, and construct or instantiate an
80 object. A specific object is often referred to as an instance of a
81 class.
82
83 In Perl, any package can be a class. The difference between a package
84 which is a class and one which isn't is based on how the package is
85 used. Here's our "class declaration" for the "File" class:
86
87 package File;
88
89 In Perl, there is no special keyword for constructing an object.
90 However, most OO modules on CPAN use a method named "new()" to
91 construct a new object:
92
93 my $hostname = File->new(
94 path => '/etc/hostname',
95 content => "foo\n",
96 last_mod_time => 1304974868,
97 );
98
99 (Don't worry about that "->" operator, it will be explained later.)
100
101 Blessing
102
103 As we said earlier, most Perl objects are hashes, but an object can be
104 an instance of any Perl data type (scalar, array, etc.). Turning a
105 plain data structure into an object is done by blessing that data
106 structure using Perl's "bless" function.
107
108 While we strongly suggest you don't build your objects from scratch,
109 you should know the term bless. A blessed data structure (aka "a
110 referent") is an object. We sometimes say that an object has been
111 "blessed into a class".
112
113 Once a referent has been blessed, the "blessed" function from the
114 Scalar::Util core module can tell us its class name. This subroutine
115 returns an object's class when passed an object, and false otherwise.
116
117 use Scalar::Util 'blessed';
118
119 print blessed($hash); # undef
120 print blessed($hostname); # File
121
122 Constructor
123
124 A constructor creates a new object. In Perl, a class's constructor is
125 just another method, unlike some other languages, which provide syntax
126 for constructors. Most Perl classes use "new" as the name for their
127 constructor:
128
129 my $file = File->new(...);
130
131 Methods
132 You already learned that a method is a subroutine that operates on an
133 object. You can think of a method as the things that an object can do.
134 If an object is a noun, then methods are its verbs (save, print, open).
135
136 In Perl, methods are simply subroutines that live in a class's package.
137 Methods are always written to receive the object as their first
138 argument:
139
140 sub print_info {
141 my $self = shift;
142
143 print "This file is at ", $self->path, "\n";
144 }
145
146 $file->print_info;
147 # The file is at /etc/hostname
148
149 What makes a method special is how it's called. The arrow operator
150 ("->") tells Perl that we are calling a method.
151
152 When we make a method call, Perl arranges for the method's invocant to
153 be passed as the first argument. Invocant is a fancy name for the thing
154 on the left side of the arrow. The invocant can either be a class name
155 or an object. We can also pass additional arguments to the method:
156
157 sub print_info {
158 my $self = shift;
159 my $prefix = shift // "This file is at ";
160
161 print $prefix, ", ", $self->path, "\n";
162 }
163
164 $file->print_info("The file is located at ");
165 # The file is located at /etc/hostname
166
167 Attributes
168 Each class can define its attributes. When we instantiate an object, we
169 assign values to those attributes. For example, every "File" object has
170 a path. Attributes are sometimes called properties.
171
172 Perl has no special syntax for attributes. Under the hood, attributes
173 are often stored as keys in the object's underlying hash, but don't
174 worry about this.
175
176 We recommend that you only access attributes via accessor methods.
177 These are methods that can get or set the value of each attribute. We
178 saw this earlier in the "print_info()" example, which calls
179 "$self->path".
180
181 You might also see the terms getter and setter. These are two types of
182 accessors. A getter gets the attribute's value, while a setter sets it.
183 Another term for a setter is mutator
184
185 Attributes are typically defined as read-only or read-write. Read-only
186 attributes can only be set when the object is first created, while
187 read-write attributes can be altered at any time.
188
189 The value of an attribute may itself be another object. For example,
190 instead of returning its last mod time as a number, the "File" class
191 could return a DateTime object representing that value.
192
193 It's possible to have a class that does not expose any publicly
194 settable attributes. Not every class has attributes and methods.
195
196 Polymorphism
197 Polymorphism is a fancy way of saying that objects from two different
198 classes share an API. For example, we could have "File" and "WebPage"
199 classes which both have a "print_content()" method. This method might
200 produce different output for each class, but they share a common
201 interface.
202
203 While the two classes may differ in many ways, when it comes to the
204 "print_content()" method, they are the same. This means that we can try
205 to call the "print_content()" method on an object of either class, and
206 we don't have to know what class the object belongs to!
207
208 Polymorphism is one of the key concepts of object-oriented design.
209
210 Inheritance
211 Inheritance lets you create a specialized version of an existing class.
212 Inheritance lets the new class reuse the methods and attributes of
213 another class.
214
215 For example, we could create an "File::MP3" class which inherits from
216 "File". An "File::MP3" is-a more specific type of "File". All mp3
217 files are files, but not all files are mp3 files.
218
219 We often refer to inheritance relationships as parent-child or
220 "superclass"/"subclass" relationships. Sometimes we say that the child
221 has an is-a relationship with its parent class.
222
223 "File" is a superclass of "File::MP3", and "File::MP3" is a subclass of
224 "File".
225
226 package File::MP3;
227
228 use parent 'File';
229
230 The parent module is one of several ways that Perl lets you define
231 inheritance relationships.
232
233 Perl allows multiple inheritance, which means that a class can inherit
234 from multiple parents. While this is possible, we strongly recommend
235 against it. Generally, you can use roles to do everything you can do
236 with multiple inheritance, but in a cleaner way.
237
238 Note that there's nothing wrong with defining multiple subclasses of a
239 given class. This is both common and safe. For example, we might define
240 "File::MP3::FixedBitrate" and "File::MP3::VariableBitrate" classes to
241 distinguish between different types of mp3 file.
242
243 Overriding methods and method resolution
244
245 Inheritance allows two classes to share code. By default, every method
246 in the parent class is also available in the child. The child can
247 explicitly override a parent's method to provide its own
248 implementation. For example, if we have an "File::MP3" object, it has
249 the "print_info()" method from "File":
250
251 my $cage = File::MP3->new(
252 path => 'mp3s/My-Body-Is-a-Cage.mp3',
253 content => $mp3_data,
254 last_mod_time => 1304974868,
255 title => 'My Body Is a Cage',
256 );
257
258 $cage->print_info;
259 # The file is at mp3s/My-Body-Is-a-Cage.mp3
260
261 If we wanted to include the mp3's title in the greeting, we could
262 override the method:
263
264 package File::MP3;
265
266 use parent 'File';
267
268 sub print_info {
269 my $self = shift;
270
271 print "This file is at ", $self->path, "\n";
272 print "Its title is ", $self->title, "\n";
273 }
274
275 $cage->print_info;
276 # The file is at mp3s/My-Body-Is-a-Cage.mp3
277 # Its title is My Body Is a Cage
278
279 The process of determining what method should be used is called method
280 resolution. What Perl does is look at the object's class first
281 ("File::MP3" in this case). If that class defines the method, then that
282 class's version of the method is called. If not, Perl looks at each
283 parent class in turn. For "File::MP3", its only parent is "File". If
284 "File::MP3" does not define the method, but "File" does, then Perl
285 calls the method in "File".
286
287 If "File" inherited from "DataSource", which inherited from "Thing",
288 then Perl would keep looking "up the chain" if necessary.
289
290 It is possible to explicitly call a parent method from a child:
291
292 package File::MP3;
293
294 use parent 'File';
295
296 sub print_info {
297 my $self = shift;
298
299 $self->SUPER::print_info();
300 print "Its title is ", $self->title, "\n";
301 }
302
303 The "SUPER::" bit tells Perl to look for the "print_info()" in the
304 "File::MP3" class's inheritance chain. When it finds the parent class
305 that implements this method, the method is called.
306
307 We mentioned multiple inheritance earlier. The main problem with
308 multiple inheritance is that it greatly complicates method resolution.
309 See perlobj for more details.
310
311 Encapsulation
312 Encapsulation is the idea that an object is opaque. When another
313 developer uses your class, they don't need to know how it is
314 implemented, they just need to know what it does.
315
316 Encapsulation is important for several reasons. First, it allows you to
317 separate the public API from the private implementation. This means you
318 can change that implementation without breaking the API.
319
320 Second, when classes are well encapsulated, they become easier to
321 subclass. Ideally, a subclass uses the same APIs to access object data
322 that its parent class uses. In reality, subclassing sometimes involves
323 violating encapsulation, but a good API can minimize the need to do
324 this.
325
326 We mentioned earlier that most Perl objects are implemented as hashes
327 under the hood. The principle of encapsulation tells us that we should
328 not rely on this. Instead, we should use accessor methods to access the
329 data in that hash. The object systems that we recommend below all
330 automate the generation of accessor methods. If you use one of them,
331 you should never have to access the object as a hash directly.
332
333 Composition
334 In object-oriented code, we often find that one object references
335 another object. This is called composition, or a has-a relationship.
336
337 Earlier, we mentioned that the "File" class's "last_mod_time" accessor
338 could return a DateTime object. This is a perfect example of
339 composition. We could go even further, and make the "path" and
340 "content" accessors return objects as well. The "File" class would then
341 be composed of several other objects.
342
343 Roles
344 Roles are something that a class does, rather than something that it
345 is. Roles are relatively new to Perl, but have become rather popular.
346 Roles are applied to classes. Sometimes we say that classes consume
347 roles.
348
349 Roles are an alternative to inheritance for providing polymorphism.
350 Let's assume we have two classes, "Radio" and "Computer". Both of these
351 things have on/off switches. We want to model that in our class
352 definitions.
353
354 We could have both classes inherit from a common parent, like
355 "Machine", but not all machines have on/off switches. We could create a
356 parent class called "HasOnOffSwitch", but that is very artificial.
357 Radios and computers are not specializations of this parent. This
358 parent is really a rather ridiculous creation.
359
360 This is where roles come in. It makes a lot of sense to create a
361 "HasOnOffSwitch" role and apply it to both classes. This role would
362 define a known API like providing "turn_on()" and "turn_off()" methods.
363
364 Perl does not have any built-in way to express roles. In the past,
365 people just bit the bullet and used multiple inheritance. Nowadays,
366 there are several good choices on CPAN for using roles.
367
368 When to Use OO
369 Object Orientation is not the best solution to every problem. In Perl
370 Best Practices (copyright 2004, Published by O'Reilly Media, Inc.),
371 Damian Conway provides a list of criteria to use when deciding if OO is
372 the right fit for your problem:
373
374 • The system being designed is large, or is likely to become large.
375
376 • The data can be aggregated into obvious structures, especially if
377 there's a large amount of data in each aggregate.
378
379 • The various types of data aggregate form a natural hierarchy that
380 facilitates the use of inheritance and polymorphism.
381
382 • You have a piece of data on which many different operations are
383 applied.
384
385 • You need to perform the same general operations on related types of
386 data, but with slight variations depending on the specific type of
387 data the operations are applied to.
388
389 • It's likely you'll have to add new data types later.
390
391 • The typical interactions between pieces of data are best
392 represented by operators.
393
394 • The implementation of individual components of the system is likely
395 to change over time.
396
397 • The system design is already object-oriented.
398
399 • Large numbers of other programmers will be using your code modules.
400
402 As we mentioned before, Perl's built-in OO system is very minimal, but
403 also quite flexible. Over the years, many people have developed systems
404 which build on top of Perl's built-in system to provide more features
405 and convenience.
406
407 We strongly recommend that you use one of these systems. Even the most
408 minimal of them eliminates a lot of repetitive boilerplate. There's
409 really no good reason to write your classes from scratch in Perl.
410
411 If you are interested in the guts underlying these systems, check out
412 perlobj.
413
414 Moose
415 Moose bills itself as a "postmodern object system for Perl 5". Don't be
416 scared, the "postmodern" label is a callback to Larry's description of
417 Perl as "the first postmodern computer language".
418
419 "Moose" provides a complete, modern OO system. Its biggest influence is
420 the Common Lisp Object System, but it also borrows ideas from Smalltalk
421 and several other languages. "Moose" was created by Stevan Little, and
422 draws heavily from his work on the Raku OO design.
423
424 Here is our "File" class using "Moose":
425
426 package File;
427 use Moose;
428
429 has path => ( is => 'ro' );
430 has content => ( is => 'ro' );
431 has last_mod_time => ( is => 'ro' );
432
433 sub print_info {
434 my $self = shift;
435
436 print "This file is at ", $self->path, "\n";
437 }
438
439 "Moose" provides a number of features:
440
441 • Declarative sugar
442
443 "Moose" provides a layer of declarative "sugar" for defining
444 classes. That sugar is just a set of exported functions that make
445 declaring how your class works simpler and more palatable. This
446 lets you describe what your class is, rather than having to tell
447 Perl how to implement your class.
448
449 The "has()" subroutine declares an attribute, and "Moose"
450 automatically creates accessors for these attributes. It also takes
451 care of creating a "new()" method for you. This constructor knows
452 about the attributes you declared, so you can set them when
453 creating a new "File".
454
455 • Roles built-in
456
457 "Moose" lets you define roles the same way you define classes:
458
459 package HasOnOffSwitch;
460 use Moose::Role;
461
462 has is_on => (
463 is => 'rw',
464 isa => 'Bool',
465 );
466
467 sub turn_on {
468 my $self = shift;
469 $self->is_on(1);
470 }
471
472 sub turn_off {
473 my $self = shift;
474 $self->is_on(0);
475 }
476
477 • A miniature type system
478
479 In the example above, you can see that we passed "isa => 'Bool'" to
480 "has()" when creating our "is_on" attribute. This tells "Moose"
481 that this attribute must be a boolean value. If we try to set it to
482 an invalid value, our code will throw an error.
483
484 • Full introspection and manipulation
485
486 Perl's built-in introspection features are fairly minimal. "Moose"
487 builds on top of them and creates a full introspection layer for
488 your classes. This lets you ask questions like "what methods does
489 the File class implement?" It also lets you modify your classes
490 programmatically.
491
492 • Self-hosted and extensible
493
494 "Moose" describes itself using its own introspection API. Besides
495 being a cool trick, this means that you can extend "Moose" using
496 "Moose" itself.
497
498 • Rich ecosystem
499
500 There is a rich ecosystem of "Moose" extensions on CPAN under the
501 MooseX <https://metacpan.org/search?q=MooseX> namespace. In
502 addition, many modules on CPAN already use "Moose", providing you
503 with lots of examples to learn from.
504
505 • Many more features
506
507 "Moose" is a very powerful tool, and we can't cover all of its
508 features here. We encourage you to learn more by reading the
509 "Moose" documentation, starting with Moose::Manual
510 <https://metacpan.org/pod/Moose::Manual>.
511
512 Of course, "Moose" isn't perfect.
513
514 "Moose" can make your code slower to load. "Moose" itself is not small,
515 and it does a lot of code generation when you define your class. This
516 code generation means that your runtime code is as fast as it can be,
517 but you pay for this when your modules are first loaded.
518
519 This load time hit can be a problem when startup speed is important,
520 such as with a command-line script or a "plain vanilla" CGI script that
521 must be loaded each time it is executed.
522
523 Before you panic, know that many people do use "Moose" for command-line
524 tools and other startup-sensitive code. We encourage you to try "Moose"
525 out first before worrying about startup speed.
526
527 "Moose" also has several dependencies on other modules. Most of these
528 are small stand-alone modules, a number of which have been spun off
529 from "Moose". "Moose" itself, and some of its dependencies, require a
530 compiler. If you need to install your software on a system without a
531 compiler, or if having any dependencies is a problem, then "Moose" may
532 not be right for you.
533
534 Moo
535
536 If you try "Moose" and find that one of these issues is preventing you
537 from using "Moose", we encourage you to consider Moo next. "Moo"
538 implements a subset of "Moose"'s functionality in a simpler package.
539 For most features that it does implement, the end-user API is identical
540 to "Moose", meaning you can switch from "Moo" to "Moose" quite easily.
541
542 "Moo" does not implement most of "Moose"'s introspection API, so it's
543 often faster when loading your modules. Additionally, none of its
544 dependencies require XS, so it can be installed on machines without a
545 compiler.
546
547 One of "Moo"'s most compelling features is its interoperability with
548 "Moose". When someone tries to use "Moose"'s introspection API on a
549 "Moo" class or role, it is transparently inflated into a "Moose" class
550 or role. This makes it easier to incorporate "Moo"-using code into a
551 "Moose" code base and vice versa.
552
553 For example, a "Moose" class can subclass a "Moo" class using "extends"
554 or consume a "Moo" role using "with".
555
556 The "Moose" authors hope that one day "Moo" can be made obsolete by
557 improving "Moose" enough, but for now it provides a worthwhile
558 alternative to "Moose".
559
560 Class::Accessor
561 Class::Accessor is the polar opposite of "Moose". It provides very few
562 features, nor is it self-hosting.
563
564 It is, however, very simple, pure Perl, and it has no non-core
565 dependencies. It also provides a "Moose-like" API on demand for the
566 features it supports.
567
568 Even though it doesn't do much, it is still preferable to writing your
569 own classes from scratch.
570
571 Here's our "File" class with "Class::Accessor":
572
573 package File;
574 use Class::Accessor 'antlers';
575
576 has path => ( is => 'ro' );
577 has content => ( is => 'ro' );
578 has last_mod_time => ( is => 'ro' );
579
580 sub print_info {
581 my $self = shift;
582
583 print "This file is at ", $self->path, "\n";
584 }
585
586 The "antlers" import flag tells "Class::Accessor" that you want to
587 define your attributes using "Moose"-like syntax. The only parameter
588 that you can pass to "has" is "is". We recommend that you use this
589 Moose-like syntax if you choose "Class::Accessor" since it means you
590 will have a smoother upgrade path if you later decide to move to
591 "Moose".
592
593 Like "Moose", "Class::Accessor" generates accessor methods and a
594 constructor for your class.
595
596 Class::Tiny
597 Finally, we have Class::Tiny. This module truly lives up to its name.
598 It has an incredibly minimal API and absolutely no dependencies on any
599 recent Perl. Still, we think it's a lot easier to use than writing your
600 own OO code from scratch.
601
602 Here's our "File" class once more:
603
604 package File;
605 use Class::Tiny qw( path content last_mod_time );
606
607 sub print_info {
608 my $self = shift;
609
610 print "This file is at ", $self->path, "\n";
611 }
612
613 That's it!
614
615 With "Class::Tiny", all accessors are read-write. It generates a
616 constructor for you, as well as the accessors you define.
617
618 You can also use Class::Tiny::Antlers for "Moose"-like syntax.
619
620 Role::Tiny
621 As we mentioned before, roles provide an alternative to inheritance,
622 but Perl does not have any built-in role support. If you choose to use
623 Moose, it comes with a full-fledged role implementation. However, if
624 you use one of our other recommended OO modules, you can still use
625 roles with Role::Tiny
626
627 "Role::Tiny" provides some of the same features as Moose's role system,
628 but in a much smaller package. Most notably, it doesn't support any
629 sort of attribute declaration, so you have to do that by hand. Still,
630 it's useful, and works well with "Class::Accessor" and "Class::Tiny"
631
632 OO System Summary
633 Here's a brief recap of the options we covered:
634
635 • Moose
636
637 "Moose" is the maximal option. It has a lot of features, a big
638 ecosystem, and a thriving user base. We also covered Moo briefly.
639 "Moo" is "Moose" lite, and a reasonable alternative when Moose
640 doesn't work for your application.
641
642 • Class::Accessor
643
644 "Class::Accessor" does a lot less than "Moose", and is a nice
645 alternative if you find "Moose" overwhelming. It's been around a
646 long time and is well battle-tested. It also has a minimal "Moose"
647 compatibility mode which makes moving from "Class::Accessor" to
648 "Moose" easy.
649
650 • Class::Tiny
651
652 "Class::Tiny" is the absolute minimal option. It has no
653 dependencies, and almost no syntax to learn. It's a good option for
654 a super minimal environment and for throwing something together
655 quickly without having to worry about details.
656
657 • Role::Tiny
658
659 Use "Role::Tiny" with "Class::Accessor" or "Class::Tiny" if you
660 find yourself considering multiple inheritance. If you go with
661 "Moose", it comes with its own role implementation.
662
663 Other OO Systems
664 There are literally dozens of other OO-related modules on CPAN besides
665 those covered here, and you're likely to run across one or more of them
666 if you work with other people's code.
667
668 In addition, plenty of code in the wild does all of its OO "by hand",
669 using just the Perl built-in OO features. If you need to maintain such
670 code, you should read perlobj to understand exactly how Perl's built-in
671 OO works.
672
674 As we said before, Perl's minimal OO system has led to a profusion of
675 OO systems on CPAN. While you can still drop down to the bare metal and
676 write your classes by hand, there's really no reason to do that with
677 modern Perl.
678
679 For small systems, Class::Tiny and Class::Accessor both provide minimal
680 object systems that take care of basic boilerplate for you.
681
682 For bigger projects, Moose provides a rich set of features that will
683 let you focus on implementing your business logic. Moo provides a nice
684 alternative to Moose when you want a lot of features but need faster
685 compile time or to avoid XS.
686
687 We encourage you to play with and evaluate Moose, Moo, Class::Accessor,
688 and Class::Tiny to see which OO system is right for you.
689
690
691
692perl v5.36.0 2022-08-30 PERLOOTUT(1)