Classes are the basic template from which object instances are created. A class is made up of a collection of variables representing internal state and methods providing behaviours that operate on that state.
Classes are defined in Ruby using the
class keyword followed by a name. The name must begin with a capital letter and by convention names that contain more than one word are run together with each word capitalized and no separating characters (CamelCase). The class definition may contain method, class variable, and instance variable declarations as well as calls to methods that execute in the class context at read time, such as attr_accessor. The class declaration is terminated by the
class MyClass def some_method end end
Instance variables are created for each class instance and are accessible only within that instance. They are accessed using the @ operator. Outside of the class definition, the value of an instance variable can only be read or modified via that instance's public methods.
class MyClass @one = 1 def do_something @one = 2 end def output puts @one end end instance = MyClass.new instance.output instance.do_something instance.output
Surprisingly, this outputs:
This happens (nil in the first output line) because @one defined below class MyClass is an instance variable belonging to the class object (note this is not the same as a class variable and could not be referred to as @@one), whereas @one defined inside the do_something method is an instance variable belonging to instances of MyClass. They are two distinct variables and the first is accessible only in a class method.
As noted in the previous section, an instance variable can only be directly accessed or modified within an instance method definition. If you want to provide access to it from outside, you need to define public accessor methods, for example
class MyClass def initialize @foo = 28 end def foo return @foo end def foo=(value) @foo = value end end instance = MyClass.new puts instance.foo instance.foo = 496 puts instance.foo
Note that ruby provides a bit of syntactic sugar to make it look like you are getting and setting a variable directly; under the hood
a = instance.foo instance.foo = b
are calls to the foo and foo= methods
a = instance.foo() instance.foo=(b)
Since this is such a common use case, there is also a convenience method to autogenerate these getters and setters:
class MyClass attr_accessor :foo def initialize @foo = 28 end end instance = MyClass.new puts instance.foo instance.foo = 496 puts instance.foo
does the same thing as the above program. The attr_accessor method is run at read time, when ruby is constructing the class object, and it generates the foo and foo= methods.
However, there is no requirement for the accessor methods to simply transparently access the instance variable. For example, we could ensure that all values are rounded before being stored in foo:
class MyClass def initialize @foo = 28 end def foo return @foo end def foo=(value) @foo = value.round end end instance = MyClass.new puts instance.foo instance.foo = 496.2 puts instance.foo #=> 496
Class variables are accessed using the @@ operator. These variables are associated with the class hierarchy rather than any object instance of the class and are the same across all object instances. (These are similar to class "static" variables in Java or C++).
class MyClass @@value = 1 def add_one @@value= @@value + 1 end def value @@value end end instanceOne = MyClass.new instanceTwo = MyClass.new puts instanceOne.value instanceOne.add_one puts instanceOne.value puts instanceTwo.value
1 2 2
Class Instance Variables
Classes can have instance variables. This gives each class a variable that is not shared by other classes in the inheritance chain.
class Employee class << self; attr_accessor :instances; end def store self.class.instances ||=  self.class.instances << self end def initialize name @name = name end end class Overhead < Employee; end class Programmer < Employee; end Overhead.new('Martin').store Overhead.new('Roy').store Programmer.new('Erik').store puts Overhead.instances.size # => 2 puts Programmer.instances.size # => 1
For more details, see MF Bliki: ClassInstanceVariables
Class methods are declared the same way as normal methods, except that they are prefixed by
self, or the class name, followed by a period. These methods are executed at the Class level and may be called without an object instance. They cannot access instance variables but do have access to class variables.
class MyClass def self.some_method puts 'something' end end MyClass.some_method
An object instance is created from a class through the a process called instantiation. In Ruby this takes place through the Class method
anObject = MyClass.new(parameters)
This function sets up the object in memory and then delegates control to the initialize function of the class if it is present. Parameters passed to the new function are passed into the
class MyClass def initialize(parameters) end end
By default, all methods in Ruby classes are public - accessible by anyone. There are, nonetheless, only two exceptions for this rule: the global methods defined under the Object class, and the initialize method for any class. Both of them are implicitly private.
If desired, the access for methods can be restricted by public, private, protected object methods.
It is interesting that these are not actually keywords, but actual methods that operate on the class, dynamically altering the visibility of the methods, and as a result, these 'keywords' influence the visibility of all following declarations until a new visibility is set or the end of the declaration-body is reached.
class Example def methodA end private # all methods that follow will be made private: not accessible for outside objects def methodP end end
If private is invoked without arguments, it sets access to private for all subsequent methods. It can also be invoked with named arguments.
Named private method example:
class Example def methodA end def methodP end private :methodP end
Here private was invoked with an argument, altering the visibility of methodP to private.
Note for class methods (those that are declared using def ClassName.method_name), you need to use another function: private_class_method
Common usage of private_class_method is to make the "new" method (constructor) inaccessible, to force access to an object through some getter function. A typical Singleton implementation is an obvious example.
class SingletonLike private_class_method :new def SingletonLike.create(*args, &block) @@inst = new(*args, &block) unless @@inst return @@inst end end
Note : another popular way to code the same declaration
class SingletonLike private_class_method :new def SingletonLike.create(*args, &block) @@inst ||= new(*args, &block) end end
More info about the difference between C++ and Ruby private/protected: http://lylejohnson.name/blog/?p=5
One person summed up the distinctions by saying that in C++, “private” means “private to this class”, while in Ruby it means “private to this instance”. What this means, in C++ from code in class A, you can access any private method for any other object of type A. In Ruby, you can not: you can only access private methods for your instance of object, and not for any other object instance (of class A).
Ruby folks keep saying "private means you cannot specify the receiver". What they are saying, if method is private, in your code you can say:
class AccessPrivate def a end private :a # a is private method def accessing_private a # sure! self.a # nope! private methods cannot be called with an explicit receiver at all, even if that receiver is "self" other_object.a # nope, a is private, you can't get it (but if it was protected, you could!) end end
Here, "other_object" is the "receiver" that method "a" is invoked on. For private methods, it does not work. However, that is what "protected" visibility will allow.
Public is default accessibility level for class methods. I am not sure why this is specified - maybe for completeness, maybe so that you could dynamically make some method private at some point, and later - public.
In Ruby, visibility is completely dynamic. You can change method visibility at runtime!
Now, “protected” deserves more discussion. Those of you coming from Java or C++ have learned that in those languages, if a method is “private”, its visibility is restricted to the declaring class, and if the method is “protected”, it will be accessible to children of the class (classes that inherit from parent) or other classes in that package.
In Ruby, “private” visibility is similar to what “protected” is in Java. Private methods in Ruby are accessible from children. You can’t have truly private methods in Ruby; you can’t completely hide a method.
The difference between protected and private is subtle. If a method is protected, it may be called by any instance of the defining class or its subclasses. If a method is private, it may be called only within the context of the calling object---it is never possible to access another object instance's private methods directly, even if the object is of the same class as the caller. For protected methods, they are accessible from objects of the same class (or children).
So, from within an object "a1" (an instance of Class A), you can call private methods only for instance of "a1" (self). And you can not call private methods of object "a2" (that also is of class A) - they are private to a2. But you can call protected methods of object "a2" since objects a1 and a2 are both of class A.
Ruby FAQ gives following example - implementing an operator that compares one internal variable with a variable from the same class (for purposes of comparing the objects):
def <=>(other) self.age <=> other.age end
If age is private, this method will not work, because other.age is not accessible. If "age" is protected, this will work fine, because self and other are of same class, and can access each other's protected methods.
To think of this, protected actually reminds me of the "internal" accessibility modifier in C# or "default" accessibility in Java (when no accessibility keword is set on method or variable): method is accessible just as "public", but only for classes inside the same package.
Note that object instance variables are not really private, you just can't see them. To access an instance variable, you need to create a getter and setter.
Like this (no, don't do this by hand! See below):
class GotAccessor def initialize(size) @size = size end def size @size end def size=(val) @size = val end end # you could the access @size variable as # a = GotAccessor.new(5) # x = a.size # a.size = y
Luckily, we have special functions to do just that: attr_accessor, attr_reader, attr_writer. attr_accessor will give you get/set functionality, reader will give only getter and writer will give only setter.
Now reduced to:
class GotAccessor def initialize(size) @size = size end attr_accessor :size end # attr_accessor generates variable @size accessor methods automatically: # a = GotAccessor.new(5) # x = a.size # a.size = y
A class can inherit functionality and variables from a superclass, sometimes referred to as a parent class or base class. Ruby does not support multiple inheritance and so a class in Ruby can have only one superclass. The syntax is as follows:
class ParentClass def a_method puts 'b' end end class SomeClass < ParentClass # < means inherit (or "extends" if you are from Java background) def another_method puts 'a' end end instance = SomeClass.new instance.another_method instance.a_method
All non-private variables and functions are inherited by the child class from the superclass.
If your class overrides a method from parent class (superclass), you still can access the parent's method by using 'super' keyword.
class ParentClass def a_method puts 'b' end end class SomeClass < ParentClass def a_method super puts 'a' end end instance = SomeClass.new instance.a_method
(because a_method also did invoke the method from parent class).
If you have a deep inheritance line, and still want to access some parent class (superclass) methods directly, you can't. super only gets you a direct parent's method. But there is a workaround! When inheriting from a class, you can alias parent class method to a different name. Then you can access methods by alias.
class X def foo "hello" end end class Y < X alias xFoo foo def foo xFoo + "y" end end class Z < Y def foo xFoo + "z" end end puts X.new.foo puts Y.new.foo puts Z.new.foo
hello helloy helloz
Mixing in Modules
First, you need to read up on modules Ruby modules. Modules are a way of grouping together some functions and variables and classes, somewhat like classes, but more like namespaces. So a module is not really a class. You can't instantiate a Module, and thus it does not have self.
This trait, however, allows us to include the module into a class. Mix it in, so to speak.
module A def a1 puts 'a1 is called' end end module B def b1 puts 'b1 is called' end end module C def c1 puts 'c1 is called' end end class Test include A include B include C def display puts 'Modules are included' end end object=Test.new object.display object.a1 object.b1 object.c1
Modules are included a1 is called b1 is called c1 is called
The code shows Multiple Inheritance using modules.
Ruby Class Meta-Model
In keeping with the Ruby principle that everything is an object, classes are themselves instances of the class Class. They are stored in constants under the scope of the module in which they are declared. A call to a method on an object instance is delegated to a variable inside the object that contains a reference to the class of that object. The method implementation exists on the Class instance object itself. Class methods are implemented on meta-classes that are linked to the existing class instance objects in the same way that those classes instances are linked to them. These meta-classes are hidden from most Ruby functions.