Commit | Line | Data |
ecd7cc3e |
1 | =pod |
2 | |
3 | =head1 NAME |
4 | |
d67ce58f |
5 | Moose::Manual::Types - Moose's type system |
ecd7cc3e |
6 | |
7 | =head1 TYPES IN PERL? |
8 | |
0c551c67 |
9 | Moose provides its own type system for attributes. You can also use |
10 | these types to validate method parameters with the help of a MooseX |
11 | module. |
ecd7cc3e |
12 | |
13 | Moose's type system is based on a combination of Perl 5's own |
909103e1 |
14 | I<implicit> types and some Perl 6 concepts. You can create your |
0c551c67 |
15 | own subtypes with custom constraints, making it easy to express any |
16 | sort of validation. |
ecd7cc3e |
17 | |
0c551c67 |
18 | Types have names, and you can re-use them by name, making it easy to |
ecd7cc3e |
19 | share types throughout a large application. |
20 | |
909103e1 |
21 | However, this is not a "real" type system. Moose does not magically make Perl |
22 | start associating types with variables. This is just an advanced parameter |
23 | checking system which allows you to associate a name with a constraint. |
ecd7cc3e |
24 | |
25 | That said, it's still pretty damn useful, and we think it's one of the |
26 | things that makes Moose both fun and powerful. Taking advantage of the |
27 | type system makes it much easier to ensure that you are getting valid |
28 | data, and it also contributes greatly to code maintainability. |
29 | |
30 | =head1 THE TYPES |
31 | |
32 | The basic Moose type hierarchy looks like this |
33 | |
34 | Any |
35 | Item |
36 | Bool |
37 | Maybe[`a] |
38 | Undef |
39 | Defined |
40 | Value |
ecd7cc3e |
41 | Str |
909103e1 |
42 | Num |
43 | Int |
44 | ClassName |
45 | RoleName |
ecd7cc3e |
46 | Ref |
977a5e90 |
47 | ScalarRef[`a] |
ecd7cc3e |
48 | ArrayRef[`a] |
49 | HashRef[`a] |
50 | CodeRef |
51 | RegexpRef |
52 | GlobRef |
909103e1 |
53 | FileHandle |
ecd7cc3e |
54 | Object |
ecd7cc3e |
55 | |
56 | In practice, the only difference between C<Any> and C<Item> is |
57 | conceptual. C<Item> is used as the top-level type in the hierarchy. |
58 | |
f91b2a92 |
59 | The rest of these types correspond to existing Perl concepts. |
60 | In particular: |
61 | |
62 | =over 4 |
63 | |
909103e1 |
64 | =item |
f91b2a92 |
65 | |
909103e1 |
66 | C<Bool> accepts C<1> for true, and undef, 0, or the empty string as false. |
f91b2a92 |
67 | |
909103e1 |
68 | =item |
f91b2a92 |
69 | |
909103e1 |
70 | C<Maybe[`a]> accepts either C<`a> or C<undef>. |
f91b2a92 |
71 | |
909103e1 |
72 | =item |
f91b2a92 |
73 | |
909103e1 |
74 | C<Num> accepts anything that perl thinks looks like a number (see L<Scalar::Util/looks_like_number>). |
75 | |
76 | =item |
77 | |
78 | C<ClassName> and C<RoleName> accept strings that are either the name of a class or the name of a role. The class/role must already be loaded when the constraint is checked. |
79 | |
80 | =item |
81 | |
82 | C<FileHandle> accepts either an L<IO::Handle> object or a builtin perl filehandle (see L<Scalar::Util/openhandle>). |
83 | |
84 | =item |
85 | |
86 | C<Object> accepts any blessed reference. |
f91b2a92 |
87 | |
88 | =back |
ecd7cc3e |
89 | |
90 | The types followed by "[`a]" can be parameterized. So instead of just |
91 | plain C<ArrayRef> we can say that we want C<ArrayRef[Int]> instead. We |
92 | can even do something like C<HashRef[ArrayRef[Str]]>. |
93 | |
94 | The C<Maybe[`a]> type deserves a special mention. Used by itself, it |
95 | doesn't really mean anything (and is equivalent to C<Item>). When it |
96 | is parameterized, it means that the value is either C<undef> or the |
5a3fb5fc |
97 | parameterized type. So C<Maybe[Int]> means an integer or C<undef>. |
ecd7cc3e |
98 | |
99 | For more details on the type hierarchy, see |
100 | L<Moose::Util::TypeConstraints>. |
101 | |
102 | =head1 WHAT IS A TYPE? |
103 | |
104 | It's important to realize that types are not classes (or |
105 | packages). Types are just objects (L<Moose::Meta::TypeConstraint> |
0c551c67 |
106 | objects, to be exact) with a name and a constraint. Moose maintains a |
6549b0d1 |
107 | global type registry that lets it convert names like C<Num> into the |
0c551c67 |
108 | appropriate object. |
ecd7cc3e |
109 | |
110 | However, class names I<can be> type names. When you define a new class |
111 | using Moose, it defines an associated type name behind the scenes: |
112 | |
113 | package MyApp::User; |
114 | |
115 | use Moose; |
116 | |
117 | Now you can use C<'MyApp::User'> as a type name: |
118 | |
119 | has creator => ( |
6c5b976f |
120 | is => 'ro', |
ecd7cc3e |
121 | isa => 'MyApp::User', |
122 | ); |
123 | |
124 | However, for non-Moose classes there's no magic. You may have to |
125 | explicitly declare the class type. This is a bit muddled because Moose |
126 | assumes that any unknown type name passed as the C<isa> value for an |
127 | attribute is a class. So this works: |
128 | |
129 | has 'birth_date' => ( |
6c5b976f |
130 | is => 'ro', |
ecd7cc3e |
131 | isa => 'DateTime', |
132 | ); |
133 | |
134 | In general, when Moose is presented with an unknown name, it assumes |
135 | that the name is a class: |
136 | |
137 | subtype 'ModernDateTime' |
138 | => as 'DateTime' |
139 | => where { $_->year() >= 1980 } |
140 | => message { 'The date you provided is not modern enough' }; |
141 | |
142 | has 'valid_dates' => ( |
143 | is => 'ro', |
144 | isa => 'ArrayRef[DateTime]', |
145 | ); |
146 | |
0c39debe |
147 | Moose will assume that C<DateTime> is a class name in both of these |
0c551c67 |
148 | instances. |
ecd7cc3e |
149 | |
150 | =head1 SUBTYPES |
151 | |
dab94063 |
152 | Moose uses subtypes in its built-in hierarchy. For example, C<Int> is |
153 | a child of C<Num>. |
ecd7cc3e |
154 | |
155 | A subtype is defined in terms of a parent type and a constraint. Any |
dab94063 |
156 | constraints defined by the parent(s) will be checked first, followed by |
157 | constraints defined by the subtype. A value must pass I<all> of these |
158 | checks to be valid for the subtype. |
ecd7cc3e |
159 | |
0c551c67 |
160 | Typically, a subtype takes the parent's constraint and makes it more |
ecd7cc3e |
161 | specific. |
162 | |
163 | A subtype can also define its own constraint failure message. This |
164 | lets you do things like have an error "The value you provided (20), |
165 | was not a valid rating, which must be a number from 1-10." This is |
166 | much friendlier than the default error, which just says that the value |
167 | failed a validation check for the type. |
168 | |
169 | Here's a simple (and useful) subtype example: |
170 | |
171 | subtype 'PositiveInt' |
172 | => as 'Int' |
173 | => where { $_ > 0 } |
174 | => message { "The number you provided, $_, was not a positive number" } |
175 | |
176 | Note that the sugar functions for working with types are all exported |
177 | by L<Moose::Util::TypeConstraints>. |
178 | |
ecd7cc3e |
179 | =head1 TYPE NAMES |
180 | |
181 | Type names are global throughout the current Perl |
0c551c67 |
182 | interpreter. Internally, Moose maps names to type objects via a |
183 | L<registry|Moose::Meta::TypeConstraint::Registry>. |
ecd7cc3e |
184 | |
185 | If you have multiple apps or libraries all using Moose in the same |
186 | process, you could have problems with collisions. We recommend that |
187 | you prefix names with some sort of namespace indicator to prevent |
188 | these sorts of collisions. |
189 | |
190 | For example, instead of calling a type "PositiveInt", call it |
bcc22289 |
191 | "MyApp::Type::PositiveInt" or "MyApp::Types::PositiveInt". We |
192 | recommend that you centralize all of these definitions in a single |
193 | package, C<MyApp::Types>, which can be loaded by other classes in your |
194 | application. |
ecd7cc3e |
195 | |
909103e1 |
196 | However, before you do this, you should look at the L<MooseX::Types> |
197 | module. This module makes it easy to create a "type library" module, which can |
198 | export your types as perl constants. |
06712014 |
199 | |
200 | has 'counter' => (is => 'rw', isa => PositiveInt); |
201 | |
99f1bb40 |
202 | This lets you use a short name rather than needing to fully qualify the name |
203 | everywhere. It also allows you to write easily create parameterized types: |
06712014 |
204 | |
205 | has 'counts' => (is => 'ro', isa => HashRef[PositiveInt]); |
206 | |
99f1bb40 |
207 | This module will check your names at compile time, and is generally more |
208 | robust than the string type parsing for complex cases. |
06712014 |
209 | |
ecd7cc3e |
210 | =head1 COERCION |
211 | |
909103e1 |
212 | A coercion lets you tell Moose to automatically convert one type to another. |
ecd7cc3e |
213 | |
214 | subtype 'ArrayRefOfInts' |
215 | => as 'ArrayRef[Int]'; |
216 | |
217 | coerce 'ArrayRefOfInts' |
218 | => from 'Int' |
219 | => via { [ $_ ] }; |
220 | |
909103e1 |
221 | You'll note that we created a subtype rather than coercing C<ArrayRef[Int]> |
222 | directly. It's a bad idea to add coercions to the raw built in |
223 | types. |
224 | |
225 | Coercions are global, just like type names, so a coercion applied to a built |
226 | in type is seen by all modules using Moose types. This is I<another> reason |
227 | why it is good to namespace your types. |
ecd7cc3e |
228 | |
909103e1 |
229 | Moose will I<never> try to coerce a value unless you explicitly ask for |
230 | it. This is done by setting the C<coerce> attribute option to a true value: |
ecd7cc3e |
231 | |
232 | package Foo; |
233 | |
234 | has 'sizes' => ( |
6c5b976f |
235 | is => 'ro', |
ecd7cc3e |
236 | isa => 'ArrayRefOfInts', |
237 | coerce => 1, |
238 | ); |
239 | |
240 | Foo->new( sizes => 42 ); |
241 | |
242 | This code example will do the right thing, and the newly created |
243 | object will have C<[ 42 ]> as its C<sizes> attribute. |
244 | |
d67ce58f |
245 | =head2 Deep coercion |
ecd7cc3e |
246 | |
247 | Deep coercion is the coercion of type parameters for parameterized |
248 | types. Let's take these types as an example: |
249 | |
250 | subtype 'HexNum' |
251 | => as 'Str' |
252 | => where { /[a-f0-9]/i }; |
253 | |
254 | coerce 'Int' |
255 | => from 'HexNum' |
256 | => via { hex $_ }; |
257 | |
258 | has 'sizes' => ( |
6c5b976f |
259 | is => 'ro', |
ecd7cc3e |
260 | isa => 'ArrayRef[Int]', |
261 | coerce => 1, |
262 | ); |
263 | |
264 | If we try passing an array reference of hex numbers for the C<sizes> |
0c551c67 |
265 | attribute, Moose will not do any coercion. |
ecd7cc3e |
266 | |
0c551c67 |
267 | However, you can define a set of subtypes to enable coercion between |
268 | two parameterized types. |
ecd7cc3e |
269 | |
270 | subtype 'ArrayRefOfHexNums' |
271 | => as 'ArrayRef[HexNum]'; |
272 | |
273 | subtype 'ArrayRefOfInts' |
274 | => as 'ArrayRef[Int]'; |
275 | |
276 | coerce 'ArrayRefOfInts' |
277 | => from 'ArrayRefOfHexNums' |
278 | => via { [ map { hex } @{$_} ] }; |
279 | |
280 | Foo->new( sizes => [ 'a1', 'ff', '22' ] ); |
281 | |
282 | Now Moose will coerce the hex numbers to integers. |
283 | |
909103e1 |
284 | Moose does not attempt to chain coercions, so it will not |
0c551c67 |
285 | coerce a single hex number. To do that, we need to define a separate |
286 | coercion: |
ecd7cc3e |
287 | |
288 | coerce 'ArrayRefOfInts' |
289 | => from 'HexNum' |
290 | => via { [ hex $_ ] }; |
291 | |
292 | Yes, this can all get verbose, but coercion is tricky magic, and we |
0c551c67 |
293 | think it's best to make it explicit. |
ecd7cc3e |
294 | |
295 | =head1 TYPE UNIONS |
296 | |
297 | Moose allows you to say that an attribute can be of two or more |
298 | disparate types. For example, we might allow an C<Object> or |
299 | C<FileHandle>: |
300 | |
301 | has 'output' => ( |
302 | is => 'rw', |
303 | isa => 'Object | FileHandle', |
304 | ); |
305 | |
306 | Moose actually parses that string and recognizes that you are creating |
307 | a type union. The C<output> attribute will accept any sort of object, |
308 | as well as an unblessed file handle. It is up to you to do the right |
309 | thing for each of them in your code. |
310 | |
0c551c67 |
311 | Whenever you use a type union, you should consider whether or not |
312 | coercion might be a better answer. |
ecd7cc3e |
313 | |
314 | For our example above, we might want to be more specific, and insist |
315 | that output be an object with a C<print> method: |
316 | |
317 | subtype 'CanPrint' |
318 | => as 'Object' |
319 | => where { $_->can('print') }; |
320 | |
321 | We can coerce file handles to an object that satisfies this condition |
322 | with a simple wrapper class: |
323 | |
324 | package FHWrapper; |
325 | |
326 | use Moose; |
327 | |
328 | has 'handle' => ( |
6c5b976f |
329 | is => 'rw', |
ecd7cc3e |
330 | isa => 'FileHandle', |
331 | ); |
332 | |
333 | sub print { |
334 | my $self = shift; |
335 | my $fh = $self->handle(); |
336 | |
909103e1 |
337 | print {$fh} @_; |
ecd7cc3e |
338 | } |
339 | |
340 | Now we can define a coercion from C<FileHandle> to our wrapper class: |
341 | |
0c551c67 |
342 | coerce 'CanPrint' |
ecd7cc3e |
343 | => from 'FileHandle' |
344 | => via { FHWrapper->new( handle => $_ ) }; |
345 | |
346 | has 'output' => ( |
347 | is => 'rw', |
348 | isa => 'CanPrint', |
349 | coerce => 1, |
350 | ); |
351 | |
0c551c67 |
352 | This pattern of using a coercion instead of a type union will help |
353 | make your class internals simpler. |
ecd7cc3e |
354 | |
355 | =head1 TYPE CREATION HELPERS |
356 | |
357 | The L<Moose::Util::TypeConstraints> module exports a number of helper |
358 | functions for creating specific kinds of types. These include |
359 | C<class_type>, C<role_type>, and C<maybe_type>. See the docs for |
360 | details. |
361 | |
362 | One helper worth noting is C<enum>, which allows you to create a |
363 | subtype of C<Str> that only allows the specified values: |
364 | |
365 | enum 'RGB' => qw( red green blue ); |
366 | |
dab94063 |
367 | This creates a type named C<RGB>. |
ecd7cc3e |
368 | |
369 | =head1 ANONYMOUS TYPES |
370 | |
371 | All of the type creation functions return a type object. This type |
372 | object can be used wherever you would use a type name, as a parent |
16fb3624 |
373 | type, or as the value for an attribute's C<isa> option: |
ecd7cc3e |
374 | |
375 | has 'size' => ( |
909103e1 |
376 | is => 'ro', |
377 | isa => subtype( 'Int' => where { $_ > 0 } ), |
ecd7cc3e |
378 | ); |
379 | |
380 | This is handy when you want to create a one-off type and don't want to |
381 | "pollute" the global namespace registry. |
382 | |
383 | =head1 VALIDATING METHOD PARAMETERS |
384 | |
385 | Moose does not provide any means of validating method |
386 | parameters. However, there are several MooseX extensions on CPAN which |
387 | let you do this. |
388 | |
0c39debe |
389 | The simplest and least sugary is L<MooseX::Params::Validate>. This |
ecd7cc3e |
390 | lets you validate a set of named parameters using Moose types: |
391 | |
392 | use Moose; |
393 | use MooseX::Params::Validate; |
394 | |
395 | sub foo { |
396 | my $self = shift; |
0c551c67 |
397 | my %params = validated_hash( |
ecd7cc3e |
398 | \@_, |
399 | bar => { isa => 'Str', default => 'Moose' }, |
400 | ); |
401 | ... |
402 | } |
403 | |
0c39debe |
404 | L<MooseX::Params::Validate> also supports coercions. |
ecd7cc3e |
405 | |
406 | There are several more powerful extensions that support method |
407 | parameter validation using Moose types, including |
0c39debe |
408 | L<MooseX::Method::Signatures>, which gives you a full-blown C<method> |
ecd7cc3e |
409 | keyword. |
410 | |
909103e1 |
411 | method morning ( Str $name ) { |
ecd7cc3e |
412 | $self->say("Good morning ${name}!"); |
413 | } |
414 | |
ca680d1f |
415 | =head1 LOAD ORDER ISSUES |
416 | |
0c551c67 |
417 | Because Moose types are defined at runtime, you may run into load |
418 | order problems. In particular, you may want to use a class's type |
419 | constraint before that type has been defined. |
ca680d1f |
420 | |
909103e1 |
421 | In order to ameliorate this problem, we recommend defining I<all> of your |
422 | custom types in one module, C<MyApp::Types>, and then loading this module in |
423 | all of your other modules. |
ca680d1f |
424 | |
ecd7cc3e |
425 | =head1 AUTHOR |
426 | |
427 | Dave Rolsky E<lt>autarch@urth.orgE<gt> |
428 | |
429 | =head1 COPYRIGHT AND LICENSE |
430 | |
2840a3b2 |
431 | Copyright 2009 by Infinity Interactive, Inc. |
ecd7cc3e |
432 | |
433 | L<http://www.iinteractive.com> |
434 | |
435 | This library is free software; you can redistribute it and/or modify |
436 | it under the same terms as Perl itself. |
437 | |
438 | =cut |