1Moose::Manual::FAQ(3) User Contributed Perl DocumentationMoose::Manual::FAQ(3)
2
3
4
6 Moose::Manual::FAQ - Frequently asked questions about Moose
7
9 version 2.2013
10
12 Module Stability
13 Is Moose "production ready"?
14
15 Yes! Many sites with household names are using Moose to build high-
16 traffic services. Countless others are using Moose in production. See
17 <http://moose.iinteractive.com/about.html#organizations> for a partial
18 list.
19
20 As of this writing, Moose is a dependency of several hundred CPAN
21 modules. <https://metacpan.org/requires/module/Moose>
22
23 Is Moose's API stable?
24
25 Yes. The sugary API, the one 95% of users will interact with, is very
26 stable. Any changes will be 100% backwards compatible.
27
28 The meta API is less set in stone. We reserve the right to tweak parts
29 of it to improve efficiency or consistency. This will not be done
30 lightly. We do perform deprecation cycles. We really do not like making
31 ourselves look bad by breaking your code. Submitting test cases is the
32 best way to ensure that your code is not inadvertently broken by
33 refactoring.
34
35 I heard Moose is slow, is this true?
36
37 Again, this one is tricky, so Yes and No.
38
39 Firstly, nothing in life is free, and some Moose features do cost more
40 than others. It is also the policy of Moose to only charge you for the
41 features you use, and to do our absolute best to not place any extra
42 burdens on the execution of your code for features you are not using.
43 Of course using Moose itself does involve some overhead, but it is
44 mostly compile time. At this point we do have some options available
45 for getting the speed you need.
46
47 Currently we provide the option of making your classes immutable as a
48 means of boosting speed. This will mean a slightly larger compile time
49 cost, but the runtime speed increase (especially in object
50 construction) is pretty significant. This can be done with the
51 following code:
52
53 MyClass->meta->make_immutable();
54
55 Constructors
56 How do I write custom constructors with Moose?
57
58 Ideally, you should never write your own "new" method, and should use
59 Moose's other features to handle your specific object construction
60 needs. Here are a few scenarios, and the Moose way to solve them;
61
62 If you need to call initialization code post instance construction,
63 then use the "BUILD" method. This feature is taken directly from Perl
64 6. Every "BUILD" method in your inheritance chain is called (in the
65 correct order) immediately after the instance is constructed. This
66 allows you to ensure that all your superclasses are initialized
67 properly as well. This is the best approach to take (when possible)
68 because it makes subclassing your class much easier.
69
70 If you need to affect the constructor's parameters prior to the
71 instance actually being constructed, you have a number of options.
72
73 To change the parameter processing as a whole, you can use the
74 "BUILDARGS" method. The default implementation accepts key/value pairs
75 or a hash reference. You can override it to take positional args, or
76 any other format
77
78 To change the handling of individual parameters, there are coercions
79 (See the Moose::Cookbook::Basics::HTTP_SubtypesAndCoercion for a
80 complete example and explanation of coercions). With coercions it is
81 possible to morph argument values into the correct expected types. This
82 approach is the most flexible and robust, but does have a slightly
83 higher learning curve.
84
85 How do I make non-Moose constructors work with Moose?
86
87 Usually the correct approach to subclassing a non-Moose class is
88 delegation. Moose makes this easy using the "handles" keyword,
89 coercions, and "lazy_build", so subclassing is often not the ideal
90 route.
91
92 That said, if you really need to inherit from a non-Moose class, see
93 Moose::Cookbook::Basics::DateTime_ExtendingNonMooseParent for an
94 example of how to do it, or take a look at "MooseX::NonMoose" in
95 Moose::Manual::MooseX.
96
97 Accessors
98 How do I tell Moose to use get/set accessors?
99
100 The easiest way to accomplish this is to use the "reader" and "writer"
101 attribute options:
102
103 has 'bar' => (
104 isa => 'Baz',
105 reader => 'get_bar',
106 writer => 'set_bar',
107 );
108
109 Moose will still take advantage of type constraints, triggers, etc.
110 when creating these methods.
111
112 If you do not like this much typing, and wish it to be a default for
113 your classes, please see MooseX::FollowPBP. This extension will allow
114 you to write:
115
116 has 'bar' => (
117 isa => 'Baz',
118 is => 'rw',
119 );
120
121 Moose will create separate "get_bar" and "set_bar" methods instead of a
122 single "bar" method.
123
124 If you like "bar" and "set_bar", see MooseX::SemiAffordanceAccessor.
125
126 NOTE: This cannot be set globally in Moose, as that would break other
127 classes which are built with Moose. You can still save on typing by
128 defining a new "MyApp::Moose" that exports Moose's sugar and then turns
129 on MooseX::FollowPBP. See
130 Moose::Cookbook::Extending::Mooseish_MooseSugar.
131
132 How can I inflate/deflate values in accessors?
133
134 Well, the first question to ask is if you actually need both inflate
135 and deflate.
136
137 If you only need to inflate, then we suggest using coercions. Here is
138 some basic sample code for inflating a DateTime object:
139
140 class_type 'DateTime';
141
142 coerce 'DateTime'
143 => from 'Str'
144 => via { DateTime::Format::MySQL->parse_datetime($_) };
145
146 has 'timestamp' => (is => 'rw', isa => 'DateTime', coerce => 1);
147
148 This creates a custom type for DateTime objects, then attaches a
149 coercion to that type. The "timestamp" attribute is then told to expect
150 a "DateTime" type, and to try to coerce it. When a "Str" type is given
151 to the "timestamp" accessor, it will attempt to coerce the value into a
152 "DateTime" object using the code in found in the "via" block.
153
154 For a more comprehensive example of using coercions, see the
155 Moose::Cookbook::Basics::HTTP_SubtypesAndCoercion.
156
157 If you need to deflate your attribute's value, the current best
158 practice is to add an "around" modifier to your accessor:
159
160 # a timestamp which stores as
161 # seconds from the epoch
162 has 'timestamp' => (is => 'rw', isa => 'Int');
163
164 around 'timestamp' => sub {
165 my $next = shift;
166 my $self = shift;
167
168 return $self->$next unless @_;
169
170 # assume we get a DateTime object ...
171 my $timestamp = shift;
172 return $self->$next( $timestamp->epoch );
173 };
174
175 It is also possible to do deflation using coercion, but this tends to
176 get quite complex and require many subtypes. An example of this is
177 outside the scope of this document, ask on #moose or send a mail to the
178 list.
179
180 Still another option is to write a custom attribute metaclass, which is
181 also outside the scope of this document, but we would be happy to
182 explain it on #moose or the mailing list.
183
184 Method Modifiers
185 How can I affect the values in @_ using "before"?
186
187 You can't, actually: "before" only runs before the main method, and it
188 cannot easily affect the method's execution.
189
190 You similarly can't use "after" to affect the return value of a method.
191
192 We limit "before" and "after" because this lets you write more concise
193 code. You do not have to worry about passing @_ to the original method,
194 or forwarding its return value (being careful to preserve context).
195
196 The "around" method modifier has neither of these limitations, but is a
197 little more verbose.
198
199 Alternatively, the MooseX::Mangle extension provides the "mangle_args"
200 function, which does allow you to affect @_.
201
202 Can I use "before" to stop execution of a method?
203
204 Yes, but only if you throw an exception. If this is too drastic a
205 measure then we suggest using "around" instead. The "around" method
206 modifier is the only modifier which can gracefully prevent execution of
207 the main method. Here is an example:
208
209 around 'baz' => sub {
210 my $next = shift;
211 my ($self, %options) = @_;
212 unless ($options->{bar} eq 'foo') {
213 return 'bar';
214 }
215 $self->$next(%options);
216 };
217
218 By choosing not to call the $next method, you can stop the execution of
219 the main method.
220
221 Alternatively, the MooseX::Mangle extension provides the "guard"
222 function, which will conditionally prevent execution of the original
223 method.
224
225 Why can't I see return values in an "after" modifier?
226
227 As with the "before" modifier, the "after" modifier is simply called
228 after the main method. It is passed the original contents of @_ and not
229 the return values of the main method.
230
231 Again, the arguments are too lengthy as to why this has to be. And as
232 with "before" I recommend using an "around" modifier instead. Here is
233 some sample code:
234
235 around 'foo' => sub {
236 my $next = shift;
237 my ($self, @args) = @_;
238 my @rv = $next->($self, @args);
239 # do something silly with the return values
240 return reverse @rv;
241 };
242
243 Alternatively, the MooseX::Mangle extension provides the
244 "mangle_return" function, which allows modifying the return values of
245 the original method.
246
247 Type Constraints
248 How can I provide a custom error message for a type constraint?
249
250 Use the "message" option when building the subtype:
251
252 subtype 'NaturalLessThanTen'
253 => as 'Natural'
254 => where { $_ < 10 }
255 => message { "This number ($_) is not less than ten!" };
256
257 This "message" block will be called when a value fails to pass the
258 "NaturalLessThanTen" constraint check.
259
260 Can I turn off type constraint checking?
261
262 There's no support for it in the core of Moose yet. This option may
263 come in a future release.
264
265 Meanwhile there's a MooseX extension that allows you to do this on a
266 per-attribute basis, and if it doesn't do what you it's easy to write
267 one that fits your use case.
268
269 My coercions stopped working with recent Moose, why did you break it?
270
271 Moose 0.76 fixed a case where coercions were being applied even if the
272 original constraint passed. This has caused some edge cases to fail
273 where people were doing something like
274
275 subtype 'Address', as 'Str';
276 coerce 'Address', from 'Str', via { get_address($_) };
277
278 This is not what they intended, because the type constraint "Address"
279 is too loose in this case. It is saying that all strings are Addresses,
280 which is obviously not the case. The solution is to provide a "where"
281 clause that properly restricts the type constraint:
282
283 subtype 'Address', as 'Str', where { looks_like_address($_) };
284
285 This will allow the coercion to apply only to strings that fail to look
286 like an Address.
287
288 Roles
289 Why is BUILD not called for my composed roles?
290
291 "BUILD" is never called in composed roles. The primary reason is that
292 roles are not order sensitive. Roles are composed in such a way that
293 the order of composition does not matter (for information on the deeper
294 theory of this read the original traits papers here
295 <http://www.iam.unibe.ch/~scg/Research/Traits/>).
296
297 Because roles are essentially unordered, it would be impossible to
298 determine the order in which to execute the "BUILD" methods.
299
300 As for alternate solutions, there are a couple.
301
302 · Using a combination of lazy and default in your attributes to defer
303 initialization (see the Binary Tree example in the cookbook for a
304 good example of lazy/default usage
305 Moose::Cookbook::Basics::BinaryTree_AttributeFeatures)
306
307 · Use attribute triggers, which fire after an attribute is set, to
308 facilitate initialization. These are described in the Moose docs,
309 and examples can be found in the test suite.
310
311 In general, roles should not require initialization; they should either
312 provide sane defaults or should be documented as needing specific
313 initialization. One such way to "document" this is to have a separate
314 attribute initializer which is required for the role. Here is an
315 example of how to do this:
316
317 package My::Role;
318 use Moose::Role;
319
320 has 'height' => (
321 is => 'rw',
322 isa => 'Int',
323 lazy => 1,
324 default => sub {
325 my $self = shift;
326 $self->init_height;
327 }
328 );
329
330 requires 'init_height';
331
332 In this example, the role will not compose successfully unless the
333 class provides a "init_height" method.
334
335 If none of those solutions work, then it is possible that a role is not
336 the best tool for the job, and you really should be using classes. Or,
337 at the very least, you should reduce the amount of functionality in
338 your role so that it does not require initialization.
339
340 What are traits, and how are they different from roles?
341
342 In Moose, a trait is almost exactly the same thing as a role, except
343 that traits typically register themselves, which allows you to refer to
344 them by a short name ("Big" vs "MyApp::Role::Big").
345
346 In Moose-speak, a Role is usually composed into a class at compile
347 time, whereas a Trait is usually composed into an instance of a class
348 at runtime to add or modify the behavior of just that instance.
349
350 Outside the context of Moose, traits and roles generally mean exactly
351 the same thing. The original paper called them traits, but Perl 6 will
352 call them roles.
353
354 Can an attribute-generated method (e.g. an accessor) satisfy requires?
355
356 Yes, just be sure to consume the role after declaring your attribute.
357 "Required Attributes" in Moose::Manual::Roles provides an example:
358
359 package Breakable;
360 use Moose::Role;
361 requires 'stress';
362
363 package Car;
364 use Moose;
365 has 'stress' => ( is => 'rw', isa => 'Int' );
366 with 'Breakable';
367
368 If you mistakenly consume the "Breakable" role before declaring your
369 "stress" attribute, you would see an error like this:
370
371 'Breakable' requires the method 'stress' to be implemented by 'Car' at...
372
373 Moose and Subroutine Attributes
374 Why don't subroutine attributes I inherited from a superclass work?
375
376 Currently when subclassing a module is done at runtime with the
377 "extends" keyword, but attributes are checked at compile time by Perl.
378 To make attributes work, you must place "extends" in a "BEGIN" block so
379 that the attribute handlers will be available at compile time, like
380 this:
381
382 BEGIN { extends qw/Foo/ }
383
384 Note that we're talking about Perl's subroutine attributes here, not
385 Moose attributes:
386
387 sub foo : Bar(27) { ... }
388
390 · Stevan Little <stevan.little@iinteractive.com>
391
392 · Dave Rolsky <autarch@urth.org>
393
394 · Jesse Luehrs <doy@tozt.net>
395
396 · Shawn M Moore <code@sartak.org>
397
398 · יובל קוג'מן (Yuval Kogman) <nothingmuch@woobling.org>
399
400 · Karen Etheridge <ether@cpan.org>
401
402 · Florian Ragwitz <rafl@debian.org>
403
404 · Hans Dieter Pearcey <hdp@weftsoar.net>
405
406 · Chris Prather <chris@prather.org>
407
408 · Matt S Trout <mst@shadowcat.co.uk>
409
411 This software is copyright (c) 2006 by Infinity Interactive, Inc.
412
413 This is free software; you can redistribute it and/or modify it under
414 the same terms as the Perl 5 programming language system itself.
415
416
417
418perl v5.32.0 2020-07-28 Moose::Manual::FAQ(3)