class << self idiom in Ruby - ruby

What does class << self do in Ruby?

First, the class << foo syntax opens up foo's singleton class (eigenclass). This allows you to specialise the behaviour of methods called on that specific object.
a = 'foo'
class << a
def inspect
'"bar"'
end
end
a.inspect # => "bar"
a = 'foo' # new object, new singleton class
a.inspect # => "foo"
Now, to answer the question: class << self opens up self's singleton class, so that methods can be redefined for the current self object (which inside a class or module body is the class or module itself). Usually, this is used to define class/module ("static") methods:
class String
class << self
def value_of obj
obj.to_s
end
end
end
String.value_of 42 # => "42"
This can also be written as a shorthand:
class String
def self.value_of obj
obj.to_s
end
end
Or even shorter:
def String.value_of obj
obj.to_s
end
When inside a function definition, self refers to the object the function is being called with. In this case, class << self opens the singleton class for that object; one use of that is to implement a poor man's state machine:
class StateMachineExample
def process obj
process_hook obj
end
private
def process_state_1 obj
# ...
class << self
alias process_hook process_state_2
end
end
def process_state_2 obj
# ...
class << self
alias process_hook process_state_1
end
end
# Set up initial state
alias process_hook process_state_1
end
So, in the example above, each instance of StateMachineExample has process_hook aliased to process_state_1, but note how in the latter, it can redefine process_hook (for self only, not affecting other StateMachineExample instances) to process_state_2. So, each time a caller calls the process method (which calls the redefinable process_hook), the behaviour changes depending on what state it's in.

I found a super simple explanation about class << self , Eigenclass and different type of methods.
In Ruby, there are three types of methods that can be applied to a class:
Instance methods
Singleton methods
Class methods
Instance methods and class methods are almost similar to their homonymous in other programming languages.
class Foo
def an_instance_method
puts "I am an instance method"
end
def self.a_class_method
puts "I am a class method"
end
end
foo = Foo.new
def foo.a_singleton_method
puts "I am a singletone method"
end
Another way of accessing an Eigenclass(which includes singleton methods) is with the following syntax (class <<):
foo = Foo.new
class << foo
def a_singleton_method
puts "I am a singleton method"
end
end
now you can define a singleton method for self which is the class Foo itself in this context:
class Foo
class << self
def a_singleton_and_class_method
puts "I am a singleton method for self and a class method for Foo"
end
end
end

Usually, instance methods are global methods. That means they are available in all instances of the class on which they were defined. In contrast, a singleton method is implemented on a single object.
Ruby stores methods in classes and all methods must be associated with a class. The object on which a singleton method is defined is not a class (it is an instance of a class). If only classes can store methods, how can an object store a singleton method? When a singleton method is created, Ruby automatically creates an anonymous class to store that method. These anonymous classes are called metaclasses, also known as singleton classes or eigenclasses. The singleton method is associated with the metaclass which, in turn, is associated with the object on which the singleton method was defined.
If multiple singleton methods are defined within a single object, they are all stored in the same metaclass.
class Zen
end
z1 = Zen.new
z2 = Zen.new
class << z1
def say_hello
puts "Hello!"
end
end
z1.say_hello # Output: Hello!
z2.say_hello # Output: NoMethodError: undefined method `say_hello'…
In the above example, class << z1 changes the current self to point to the metaclass of the z1 object; then, it defines the say_hello method within the metaclass.
Classes are also objects (instances of the built-in class called Class). Class methods are nothing more than singleton methods associated with a class object.
class Zabuton
class << self
def stuff
puts "Stuffing zabuton…"
end
end
end
All objects may have metaclasses. That means classes can also have metaclasses. In the above example, class << self modifies self so it points to the metaclass of the Zabuton class. When a method is defined without an explicit receiver (the class/object on which the method will be defined), it is implicitly defined within the current scope, that is, the current value of self. Hence, the stuff method is defined within the metaclass of the Zabuton class. The above example is just another way to define a class method. IMHO, it's better to use the def self.my_new_clas_method syntax to define class methods, as it makes the code easier to understand. The above example was included so we understand what's happening when we come across the class << self syntax.
Additional info can be found at this post about Ruby Classes.

What class << thing does:
class Hi
self #=> Hi
class << self #same as 'class << Hi'
self #=> #<Class:Hi>
self == Hi.singleton_class #=> true
end
end
[it makes self == thing.singleton_class in the context of its block].
What is thing.singleton_class?
hi = String.new
def hi.a
end
hi.class.instance_methods.include? :a #=> false
hi.singleton_class.instance_methods.include? :a #=> true
hi object inherits its #methods from its #singleton_class.instance_methods and then from its #class.instance_methods.
Here we gave hi's singleton class instance method :a. It could have been done with class << hi instead.
hi's #singleton_class has all instance methods hi's #class has, and possibly some more (:a here).
[instance methods of thing's #class and #singleton_class can be applied directly to thing. when ruby sees thing.a, it first looks for :a method definition in thing.singleton_class.instance_methods and then in thing.class.instance_methods]
By the way - they call object's singleton class == metaclass == eigenclass.

А singleton method is a method that is defined only for a single object.
Example:
class SomeClass
class << self
def test
end
end
end
test_obj = SomeClass.new
def test_obj.test_2
end
class << test_obj
def test_3
end
end
puts "Singleton's methods of SomeClass"
puts SomeClass.singleton_methods
puts '------------------------------------------'
puts "Singleton's methods of test_obj"
puts test_obj.singleton_methods
Singleton's methods of SomeClass
test
Singleton's methods of test_obj
test_2
test_3

In fact if you write any C extensions for your Ruby projects there is really only one way to define a Module method.
rb_define_singleton_method
I know this self business just opens up all kinds of other questions so you could do better by searching each part.
Objects first.
foo = Object.new
Can I make a method for foo?
Sure
def foo.hello
'hello'
end
What do I do with it?
foo.hello
==>"hello"
Just another object.
foo.methods
You get all the Object methods plus your new one.
def foo.self
self
end
foo.self
Just the foo Object.
Try to see what happens if you make foo from other Objects like Class and Module. The examples from all the answers are nice to play with but you have to work with different ideas or concepts to really understand what is going on with the way the code is written. So now you have lots of terms to go look at.
Singleton,
Class,
Module,
self,
Object,
and Eigenclass was brought up but Ruby doesn't name Object Models that way. It's more like Metaclass.
Richard or __why shows you the idea here.
http://viewsourcecode.org/why/hacking/seeingMetaclassesClearly.html
And if the blows you away then try looking up Ruby Object Model in search.
Two videos that I know of on YouTube are Dave Thomas and Peter Cooper. They try to explain that concept too. It took Dave a long time to get it so don't worry.
I'm still working on it too. Why else would I be here?
Thanks for your question.
Also take a look at the standard library. It has a Singleton Module just as an FYI.
This is pretty good.
https://www.youtube.com/watch?v=i4uiyWA8eFk

Related

class method `self.` within class methods block `class << self` in Ruby

context: I'm currently working with the parser gem and trying to handle all cases of what is a public method.
I've written this next code, hoping it will fail on runtime. But it doesn't.
class Foo
class << self
def self.met
puts "I'm just a troll"
end
class << self
def mut
puts "Try and find me"
end
end
end
end
So I'm wondering where is met callable (Foo.met would raise a NoMethodError)? And is this a useful Ruby pattern or just something I should not do, neither care about?
Each object in Ruby has its own singleton class. This is where all the methods of instances are defined.
Consider the following example.
class C; end
c1, c2 = C.new, C.new
c1.extend(Module.new { def m1; 42; end })
c1.m1
#⇒ 42
c2.m1
#⇒ NoMethodError: undefined method `m1' for #<C:0x000055cb062e6888>
c1.singleton_class.instance_methods.grep /m1/
#⇒ [:m1]
c2.singleton_class.instance_methods.grep /m1/
#⇒ []
Singleton class is needed to make it possible to extend objects etc.
In Ruby, everything is object. Classes are indeed objects as well. That’s why each class has its own singleton class. And each singleton class has its singleton class.
c1.singleton_class.singleton_class.singleton_class.singleton_class
#⇒ #<Class:#<Class:#<Class:#<Class:#<C:0x000055cb0459c700>>>>>
Methods defined on foo are stored in the singleton class of foo. Methods defined on foo’s singleton class are stored in the singleton class of the singleton class of foo. And so forth.
It’s not quite practical, but it’s still possible due to how Ruby treats everything as Object.

"self" in normal method vs. "self" in singleton method?

Is self the same in both of these cases?
class Person
def who_am_i?
puts self # self?
end
end
ted = Person.new
def ted.singleton_who_am_i?
puts self # self?
end
ted.who_am_i?
ted.singleton_who_am_i?
Yes, it appears so:
class Person
def who_am_i?
puts self.to_yaml
end
end
ted = Person.new
def ted.singleton_who_am_i?
puts self.to_yaml
end
ted.who_am_i?
--- !ruby/object:Person {}
ted.singleton_who_am_i?
--- !ruby/object:Person {}
class Foo
def bar
p self
end
end
class << Foo
def bar
p self
end
end
Foo.bar #Foo (the class itself)
Foo.new.bar #<Foo:0x10acc70> (a Foo class object)
When the context is in a object, self is the object.
When the context is in the class, self is the class.
Also I do know if "singleton" is a good word for Ruby because in Ruby, even a class is an object, and "singleton" simply adds method to an existing object.
Singleton methods are instance methods only a certain object responds to. Here's how this is implemented:
The object's singleton class is opened;
Ruby defines your singleton method as an instance method of the singleton class.
What is not apparent here is the actual relationship between the actual class and singleton class. One could say it sits between the class and the object.
More precisely, the singleton class inherits from the actual class, and actually is the object's class.
Why doesn't Object#class return the singleton class, you ask? Well, it simply skips over it. Yeah.
Here's an example. Given the following class:
class Person
def instance; self end # self is the Person instance here
end
person = Person.new
If we wanted to define a singleton method for person, we could write this:
class << person
def singleton; self end # What is self in this context?
end
This is roughly equivalent to this:
class SingletonPerson < Person
def singleton; self end # self is the Person instance here too!
end
person = SingletonPerson.new
The main difference is that Ruby knows that SingletonPerson is a singleton class, so when you call person.class you actually get Person and not SingletonPerson.
This effectively hides all this complexity from you, which is a great thing. However, it is also great to understand how things work under the hood. The same logic here applies to class methods, which are really just singleton methods of a Class or Module instance.

Ruby metaclass madness

I'm stuck. I'm trying to dynamically define a class method and I can't wrap my head around the ruby metaclass model. Consider the following class:
class Example
def self.meta; (class << self; self; end); end
def self.class_instance; self; end
end
Example.class_instance.class # => Class
Example.meta.class # => Class
Example.class_instance == Example # => true
Example.class_instance == Example.meta # => false
Obviously both methods return an instance of Class. But these two instances
are not the same. They also have different ancestors:
Example.meta.ancestors # => [Class, Module, Object, Kernel]
Example.class_instance.ancestors # => [Example, Object, Kernel]
What's the point in making a difference between the metaclass and the class instance?
I figured out, that I can send :define_method to the metaclass to dynamically define a method, but if I try to send it to the class instance it won't work. At least I could solve my problem, but I still want to understand why it is working this way.
Update Mar 15, 2010 13:40
Are the following assumptions correct.
If I have an instance method which calls self.instance_eval and defines a method, it will only affect the particular instance of that class.
If I have an instance method which calls self.class.instance_eval (which would be the same as calling class_eval) and defines a method it will affect all instances of that particular class resulting in a new instance method.
If I have a class method which calls instance_eval and defines a method it will result in a new instance method for all instances.
If I have a class method which calls instance_eval on the meta/eigen class and defines a method it will result in a class method.
I think it starts to make sense to me. It would certainly limit your possibilities if self inside an class method would point to the eigen class. If so it would not be possible to define an instance method from inside a class method. Is that correct?
Defining a singleton method dynamically is simple when you use instance_eval:
Example.instance_eval{ def square(n); n*n; end }
Example.square(2) #=> 4
# you can pass instance_eval a string as well.
Example.instance_eval "def multiply(x,y); x*y; end"
Example.multiply(3,9) #=> 27
As for the difference above, you are confusing 2 things:
The meta class defined by you, is what called in Ruby community as singelton class or eigen class. That singleton class is the class that you can add class(singleton) methods to.
As for the class instance you are trying to define using the class_instance method, is nothing but the class itself, to prove it, just try adding an instance method to the class Example and check if the class_instance method defined by you returns the class Example itself by checking the existence of that method:
class Example
def self.meta; (class << self; self; end); end
def self.class_instance; self; end
def hey; puts hey; end
end
Example.class_instance.instance_methods(false) #=> ['hey']
Anyway to sum it for you, when you want to add class methods, just add them to that meta class. As for the class_instance method is useless, just remove it.
Anyway I suggest you read this post to grasp some concepts of Ruby reflection system.
UPDATE
I suggest you read this nice post: Fun with Ruby's instance_eval and class_eval,
Unfortunately class_eval and instance_eval are confusing because they somehow work against their naming!
Use ClassName.instance_eval to define class methods.
Use ClassName.class_eval to define instance methods.
Now answering your assumptions:
If I have an instance method which
calls self.instance_eval and defines a
method, it will only affect the
particular instance of that class.
yes:
class Foo
def assumption1()
self.instance_eval("def test_assumption_1; puts 'works'; end")
end
end
f1 = Foo.new
f1.assumption1
f1.methods(false) #=> ["test_assumption_1"]
f2 = Foo.new.methods(false) #=> []
If I have an instance method which
calls self.class.instance_eval (which
would be the same as calling
class_eval) and defines a method it
will affect all instances of that
particular class resulting in a new
instance method.
no instance_eval in that context will define singleton methods(not instance ones) on the class itself:
class Foo
def assumption2()
self.class.instance_eval("def test_assumption_2; puts 'works'; end")
end
end
f3 = Foo.new
f3.assumption2
f3.methods(false) #=> []
Foo.singleton_methods(false) #=> ["test_assumption_2"]
For that to work replace instance_eval with class_eval above.
If I have a class method which calls
instance_eval and defines a method it
will result in a new instance method
for all instances.
Nope:
class Foo
instance_eval do
def assumption3()
puts 'works'
end
end
end
Foo.instance_methods(false) #=> []
Foo.singleton_methods(false) #=> ["assumption_3"]
That will make singleton methods, not instance methods. For that to work replace instance_eval with class_eval above.
If I have a class method which calls
instance_eval on the meta/eigen class
and defines a method it will result in
a class method.
well no, that will make so sophisticated stuff, as it will add singleton method to the singleton class, I don't think that will have any practical use.
If you define a method on a class, it can be invoked on its objects. It is an instance method.
class Example
end
Example.send :define_method, :foo do
puts "foo"
end
Example.new.foo
#=> "foo"
If you define a method on a metaclass, it can be invoked on the class. This is similar to the concept of a class method or static method in other languages.
class Example
def self.metaclass
class << self
self
end
end
end
Example.metaclass.send :define_method, :bar do
puts "bar"
end
Example.bar
#=> "bar"
The reason that metaclasses exist is because you can do this in Ruby:
str = "hello"
class << str
def output
puts self
end
end
str.output
#=> "hello"
"hi".output
# NoMethodError
As you can see, we defined a method that is only available to one instance of a String. The thing that we defined this method on is called the metaclass. In the method lookup chain, the metaclass is accessed first before searching the object's class.
If we replace the object of type String with an object of type Class, you can imagine why this means we're only defining a method on a specific class, not on all classes.
The differences between the current context and self are subtle, you can read more if you're interested.

Why isn't the eigenclass equivalent to self.class, when it looks so similar?

I've missed the memo somewhere, and I hope you'll explain this to me.
Why is the eigenclass of an object different from self.class?
class Foo
def initialize(symbol)
eigenclass = class << self
self
end
eigenclass.class_eval do
attr_accessor symbol
end
end
end
My train of logic that equates the eigenclass with class.self is rather simple:
class << self is a way of declaring class methods, rather than instance methods. It's a shortcut to def Foo.bar.
So within the reference to the class object, returning self should be identical to self.class. This is because class << self would set self to Foo.class for definition of class methods/attributes.
Am I just confused? Or, is this a sneaky trick of Ruby meta-programming?
class << self is more than just a way of declaring class methods (though it can be used that way). Probably you've seen some usage like:
class Foo
class << self
def a
print "I could also have been defined as def Foo.a."
end
end
end
This works, and is equivalent to def Foo.a, but the way it works is a little subtle. The secret is that self, in that context, refers to the object Foo, whose class is a unique, anonymous subclass of Class. This subclass is called Foo's eigenclass. So def a creates a new method called a in Foo's eigenclass, accessible by the normal method call syntax: Foo.a.
Now let's look at a different example:
str = "abc"
other_str = "def"
class << str
def frob
return self + "d"
end
end
print str.frob # => "abcd"
print other_str.frob # => raises an exception, 'frob' is not defined on other_str
This example is the same as the last one, though it may be hard to tell at first. frob is defined, not on the String class, but on the eigenclass of str, a unique anonymous subclass of String. So str has a frob method, but instances of String in general do not. We could also have overridden methods of String (very useful in certain tricky testing scenarios).
Now we're equipped to understand your original example. Inside Foo's initialize method, self refers not to the class Foo, but to some particular instance of Foo. Its eigenclass is a subclass of Foo, but it is not Foo; it couldn't be, or else the trick we saw in the second example couldn't work. So to continue your example:
f1 = Foo.new(:weasels)
f2 = Foo.new(:monkeys)
f1.weasels = 4 # Fine
f2.monkeys = 5 # Also ok
print(f1.monkeys) # Doesn't work, f1 doesn't have a 'monkeys' method.
The simplest answer: the eigenclass can't be instantiated.
class F
def eigen
class << self
self
end
end
end
F.new.eigen.new #=> TypeError: can't create instance of virtual class
Yehuda Katz does a pretty good job of explaining the subtleties in "Metaprogramming in Ruby: It's All About the Self"

In Ruby are there any related applications of the syntax: class << self ... end

class << self
attr_accessor :n, :totalX, :totalY
end
The syntax above is used for defining class instance variables. But when I think about what syntax implies, it doesn't make any sense to me, so I'm wondering if this type of syntax is used for any other types of definitions. My point of confusion here is this:
class << self
The append operator normally means "add what's on the right to the object on the left". But in the context of this block, how does that add up to "put the contents of this block into the definition of the class instance rather than the instance"?
For the same reason I'm confused as to why in one context class << self can define class instance variables while in another it seems to create class variables such as here:
class Point
# Instance methods go here
class << self
# Class methods go here
end
end
in Ruby you can reopen existing classes and add methods. That is, you can say:
class Foo
def bob
return "hello from bob"
end
end
these methods get stored somewhere in an internal dictionary (maybe an instance variable) of the Foo-class (which is just an instance of the Class-class and therefore has instance variables)
But the suprising thing is, that you can also add methods to instances of existing objects
foo = Foo.new
foo2 = Foo.new
def foo.fred
return "I am fred"
end
foo.fred #=> "I am fred"
foo2.fred #=> NoMethodError
but Where is this method actually stored?
Turns out Ruby creates a new class behind the scenes (sometimes called singleton class, metaclass or eigenclass) which gets inserted in the inheritance heirarchy between the Foo-class and its instance.
So the inheritance relationship looks like that:
foo < (eigenclass of foo) < Foo < Class
(if you say foo.superclass you will not see the singleton class)
the class << X-syntax is a way to get to this special class, so that you can manipulate it directly. The following code blocks are exactly equivalent:
def foo.bar
return "xy"
end
# is exactly the same as
class << foo
def bar
return "xy"
end
end
So the similarity between class Foo < Bar and class << Foo is not accidental, there is inheritance going on in both.
Think of class << X as "open up the metaclass of X"
The thing to remember in Ruby is that classes themselves are just objects. (Instances of the class Class) so if you say:
class Foo
class << self
def k
return "x"
end
end
end
(self is bound to Foo in this code block) so k is an instance method of the eigenclass of Foo, which makes it a class method for Foo
all this is explained more clearly in the chapter about classes of the Pickaxe (the web version does not contain the diagrams, unfortunately) and _whys Seeing Metaclasses Clearly
Think of the class as containing a dictionary of members including all the accessors and instance variables. When you tell the class to "append" to "itself", you're saying "add these to the dictionary of class members."
I'll grant the notation is a little hinky, though.
it's actually confusing to think of it in terms of an "append" operator. a better way to look at it is that just as class Foo opens up class Foo, that is, it sets 'self' to the class object Foo, creating it if necessary, so class << self opens up the eigenclass of the current 'self' object. note that it is not limited to self - for any object bar, you can say class << bar to open up the eigenclass of that object.
class A
def hello
print "hello world"
end
end
a = A.new
b = A.new
class << a
def goodbye
print "goodbye cruel world"
end
end
a.hello
b.hello
a.goodbye
b.goodbye

Resources