General Ruby question:
In Ruby, I frequently see code that's inside a class, but not part of a method. For example:
class DooDad
attr_accessor :foo
end
or
class Teacher < ActiveRecord::Base
has_many :students
end
I think attr_accessor and has_many are methods getting invoked with the :foo or :students arguments, respectively, is that right? If so, when do these kinds of statements get executed. I tried this:
class DooDad
attr_accessor :foo
puts "I happened!"
#foo = 7
end
It doesn't seem to run these a part of the new method:
dd = DooDad.new
dd.foo
outputs nil, and never spits out any puts stuff
How exactly does all that work?
Methods like attr_accessor and has_many are often called "mimic methods" because they kinda look like ruby keywords (mimic them) but they're in fact, as you and others have correctly pointed out, method calls.
dd = DooDad.new
dd.foo
outputs nil, and never spits out any puts stuff
How exactly does all that work?
When you're inside of a class definition the implicit receiver of all method calls and "variable definitions" is self, which in your case is DooDad.
So when you're writing
class DooDad
#foo = 1
end
you're actually defining an instance variable on self, that happens to be the class itself , since you're inside of that classes definition. (and outside of any other class, module or method definitions)
The method attr_accessor on the other hand generates, with some help of metaprogramming, accessor methods for an instance variable of the objects that are instantiated from class DooDad.
Back to your example:
class DooDad
attr_accessor :foo
puts "I happened!"
#foo = 7
end
With the stuff mentioned above, you should now understand that you're dealing with two different #foo variables, one for instances of class DooDad (i.e DooDad.new), the other (the one you created by writing #foo = 7) for the class DooDad itself!
When calling the new method on a class, you create an instance of it.
dd = DooDad.new
#=> dd is now an object of class DooDad
dd.foo
#=> You just called the "getter" method for an instance variable foo of object dd, which was never defined before, that's why it's returning nil.
The puts "I happened!" statement, just as the other two in fact, gets evaluated as soon as the class is loaded, but not when you call new on it.
If you want the behaviour you described (doing stuff when calling new), I suggest implementing an initialize() method for DooDad, which will get called when you call new:
class DooDad
attr_accessor :foo
def initialize()
puts "I happened!"
#foo = 7
end
end
dd = DooDad.new
#=> outputs "I happened!" and sets dd.foo = 7
dd.foo
#=> 7
But why does #foo = 7 now set the instance variable of dd and not DooDad?
When you define a method with the keyword def, you enter a new scope (you pass a scope gate). self is now no longer the class, but instead an instance of that class, that you created with new, just like dd. So when you're writing #foo = 7 inside of a method definition, you're talking about variables for an instance of class DooDad, not the class itself.
This post is probably way too long and might not even satisfy as an answer, but I hope it was somewhat comprehensive.
Methods like has_many and attr_accessor are actually methods on the Module or Class. You are absolutely right about them being normal methods, invoked with arguments just like any other. When a method is called directly in the class (outside of a method definition), it is being called on the class itself.
Here are the docs for attr_accessor.
They are syntactic sugar built into the language, so it is not the same as your test. When you write (for example)
class Foo
attr_accessor :bar
end
It is actually shorthand for...
class Foo
def bar
#bar
end
def bar=(value)
#bar = value
end
end
I am unsure about the difference between this.
def String.hello
puts "hello there"
end
and
x = Person.new
def x.hello
puts "hello there"
end
From my understanding the second code block will create an object of class Person. When I do the def x.hello it creates an anonymous class (singleton class) that will be checked first for methods when sending a message to the x object.
Is this the same case for the def String.hello? String is just an instance of class Class correct? I have read that doing def String.hello will add the method as one of String's class methods.... this would be different than an anonymous class being created that sits in between the object and its class where it gets its instance methods.
What happens with both blocks of code above?
I love this part of ruby. There is this beautiful symmetry where most of the core functionality is just sugar over the advanced functionality, so once you fully grok a concept, you can apply that understanding to a lot of the language.
Is this the same case for the def String.hello? String is just an instance of class Class correct?
Yes, you are creating an instance of Class, and assigning it to a constant.
I have read that doing def String.hello will add the method as one of String's class methods.... this would be different than an anonymous class being created that sits in between the object and its class where it gets its instance methods.
Nope, the piece you are missing is thinking its possible to have a class level method WITHOUT adding it to a singleton class. What you have is an object that is an instance of Class, and you are adding methods to an implicit class that sits inbetween it and Class. You will also see this syntax sometimes
class << self
def method
end
end
That is doing the same thing, just being very explicit about it.
Just to add to the Matt's answer:
Both examples do the same thing, Writting them in another way:
String = Class.new # < done inside ruby initialization
def String.hello
puts "hello there"
end
and
x = Person.new
def x.hello
puts "hello there"
end
On Ruby you can add methods to a class, created with A = Class.new or with the syntax sugar class A; ...; end, or to a Eigenclass, that exists for every object. Class methods are, on really, methods of the Eigenclass of an instance of Class, think about what is "self" in def self.method; ...; end. Eigenclasses can be opened with this sintax:
x = Person.new
class << x
# ...
end
As Eigenclasses are also instances of class (try to add p self.class on last example) they also have Eigenclasses and so on. If it appears to be confusing, just remember that Object is a class and Class is an object. This is why I love Ruby!
The following code will add the "hello" method to the String class, that way all following strings will have the "hello" method:
def String.hello
puts "hello there"
end
Whilst the following code, will add the "hello" method, to the instance x of the class Person. If you create a new Person, that object will not have the "hello" method.
x = Person.new
def x.hello
puts "hello there"
end
That is the main difference, as I understand it.
Does this help?
I am having a hard time understanding attr_accessor in Ruby.
Can someone explain this to me?
Let's say you have a class Person.
class Person
end
person = Person.new
person.name # => no method error
Obviously we never defined method name. Let's do that.
class Person
def name
#name # simply returning an instance variable #name
end
end
person = Person.new
person.name # => nil
person.name = "Dennis" # => no method error
Aha, we can read the name, but that doesn't mean we can assign the name. Those are two different methods. The former is called reader and latter is called writer. We didn't create the writer yet so let's do that.
class Person
def name
#name
end
def name=(str)
#name = str
end
end
person = Person.new
person.name = 'Dennis'
person.name # => "Dennis"
Awesome. Now we can write and read instance variable #name using reader and writer methods. Except, this is done so frequently, why waste time writing these methods every time? We can do it easier.
class Person
attr_reader :name
attr_writer :name
end
Even this can get repetitive. When you want both reader and writer just use accessor!
class Person
attr_accessor :name
end
person = Person.new
person.name = "Dennis"
person.name # => "Dennis"
Works the same way! And guess what: the instance variable #name in our person object will be set just like when we did it manually, so you can use it in other methods.
class Person
attr_accessor :name
def greeting
"Hello #{#name}"
end
end
person = Person.new
person.name = "Dennis"
person.greeting # => "Hello Dennis"
That's it. In order to understand how attr_reader, attr_writer, and attr_accessor methods actually generate methods for you, read other answers, books, ruby docs.
attr_accessor is just a method. (The link should provide more insight with how it works - look at the pairs of methods generated, and a tutorial should show you how to use it.)
The trick is that class is not a definition in Ruby (it is "just a definition" in languages like C++ and Java), but it is an expression that evaluates. It is during this evaluation when the attr_accessor method is invoked which in turn modifies the current class - remember the implicit receiver: self.attr_accessor, where self is the "open" class object at this point.
The need for attr_accessor and friends, is, well:
Ruby, like Smalltalk, does not allow instance variables to be accessed outside of methods1 for that object. That is, instance variables cannot be accessed in the x.y form as is common in say, Java or even Python. In Ruby y is always taken as a message to send (or "method to call"). Thus the attr_* methods create wrappers which proxy the instance #variable access through dynamically created methods.
Boilerplate sucks
Hope this clarifies some of the little details. Happy coding.
1 This isn't strictly true and there are some "techniques" around this, but there is no syntax support for "public instance variable" access.
attr_accessor is (as #pst stated) just a method. What it does is create more methods for you.
So this code here:
class Foo
attr_accessor :bar
end
is equivalent to this code:
class Foo
def bar
#bar
end
def bar=( new_value )
#bar = new_value
end
end
You can write this sort of method yourself in Ruby:
class Module
def var( method_name )
inst_variable_name = "##{method_name}".to_sym
define_method method_name do
instance_variable_get inst_variable_name
end
define_method "#{method_name}=" do |new_value|
instance_variable_set inst_variable_name, new_value
end
end
end
class Foo
var :bar
end
f = Foo.new
p f.bar #=> nil
f.bar = 42
p f.bar #=> 42
attr_accessor is very simple:
attr_accessor :foo
is a shortcut for:
def foo=(val)
#foo = val
end
def foo
#foo
end
it is nothing more than a getter/setter for an object
Basically they fake publicly accessible data attributes, which Ruby doesn't have.
It is just a method that defines getter and setter methods for instance variables. An example implementation would be:
def self.attr_accessor(*names)
names.each do |name|
define_method(name) {instance_variable_get("##{name}")} # This is the getter
define_method("#{name}=") {|arg| instance_variable_set("##{name}", arg)} # This is the setter
end
end
If you are familiar with OOP concept, You must familiar with getter and setter method.
attr_accessor does the same in Ruby.
Getter and Setter in General Way
class Person
def name
#name
end
def name=(str)
#name = str
end
end
person = Person.new
person.name = 'Eshaan'
person.name # => "Eshaan"
Setter Method
def name=(val)
#name = val
end
Getter method
def name
#name
end
Getter and Setter method in Ruby
class Person
attr_accessor :name
end
person = Person.new
person.name = "Eshaan"
person.name # => "Eshaan"
Simple Explanation Without Any Code
Most of the above answers use code. This explanation attempts to answer it without using any, via an analogy/story:
Outside parties cannot access internal CIA secrets
Let's imagine a really secret place: the CIA. Nobody knows what's happening in the CIA apart from the people inside the CIA. In other words, external people cannot access any information in the CIA. But because it's no good having an organisation that is completely secret, certain information is made available to the outside world - only things that the CIA wants everyone to know about of course: e.g. the Director of the CIA, how environmentally friendly this department is compared to all other government departments etc. Other information: e.g. who are its covert operatives in Iraq or Afghanistan - these types of things will probably remain a secret for the next 150 years.
If you're outside the CIA you can only access the information that it has made available to the public. Or to use CIA parlance you can only access information that is "cleared".
The information that the CIA wants to make available to the general public outside the CIA are called: attributes.
The meaning of read and write attributes:
In the case of the CIA, most attributes are "read only". This means if you are a party external to the CIA, you can ask: "who is the director of the CIA?" and you will get a straight answer. But what you cannot do with "read only" attributes is to make changes changes in the CIA. e.g. you cannot make a phone call and suddenly decide that you want Kim Kardashian to be the Director, or that you want Paris Hilton to be the Commander in Chief.
If the attributes gave you "write" access, then you could make changes if you want to, even if you were outside. Otherwise, the only thing you can do is read.
In other words accessors allow you to make inquiries, or to make changes, to organisations that otherwise do not let external people in, depending on whether the accessors are read or write accessors.
Objects inside a class can easily access each other
On the other hand, if you were already inside the CIA, then you could easily call up your CIA operative in Kabul because this information is easily accessible given you are already inside. But if you're outside the CIA, you simply will not be given access: you will not be able to know who they are (read access), and you will not be able to change their mission (write access).
Exact same thing with classes and your ability to access variables, properties and methods within them. HTH! Any questions, please ask and I hope i can clarify.
I faced this problem as well and wrote a somewhat lengthy answer to this question. There are some great answers on this already, but anyone looking for more clarification, I hope my answer can help
Initialize Method
Initialize allows you to set data to an instance of an object upon creation of the instance rather than having to set them on a separate line in your code each time you create a new instance of the class.
class Person
def initialize(name)
#name = name
end
def greeting
"Hello #{#name}"
end
end
person = Person.new("Denis")
puts person.greeting
In the code above we are setting the name “Denis” using the initialize method by passing Dennis through the parameter in Initialize. If we wanted to set the name without the initialize method we could do so like this:
class Person
attr_accessor :name
# def initialize(name)
# #name = name
# end
def greeting
"Hello #{name}"
end
end
person = Person.new
person.name = "Dennis"
puts person.greeting
In the code above, we set the name by calling on the attr_accessor setter method using person.name, rather than setting the values upon initialization of the object.
Both “methods” of doing this work, but initialize saves us time and lines of code.
This is the only job of initialize. You cannot call on initialize as a method. To actually get the values of an instance object you need to use getters and setters (attr_reader (get), attr_writer(set), and attr_accessor(both)). See below for more detail on those.
Getters, Setters (attr_reader, attr_writer, attr_accessor)
Getters, attr_reader: The entire purpose of a getter is to return the value of a particular instance variable. Visit the sample code below for a breakdown on this.
class Item
def initialize(item_name, quantity)
#item_name = item_name
#quantity = quantity
end
def item_name
#item_name
end
def quantity
#quantity
end
end
example = Item.new("TV",2)
puts example.item_name
puts example.quantity
In the code above you are calling the methods “item_name” and “quantity” on the instance of Item “example”. The “puts example.item_name” and “example.quantity” will return (or “get”) the value for the parameters that were passed into the “example” and display them to the screen.
Luckily in Ruby there is an inherent method that allows us to write this code more succinctly; the attr_reader method. See the code below;
class Item
attr_reader :item_name, :quantity
def initialize(item_name, quantity)
#item_name = item_name
#quantity = quantity
end
end
item = Item.new("TV",2)
puts item.item_name
puts item.quantity
This syntax works exactly the same way, only it saves us six lines of code. Imagine if you had 5 more state attributable to the Item class? The code would get long quickly.
Setters, attr_writer: What crossed me up at first with setter methods is that in my eyes it seemed to perform an identical function to the initialize method. Below I explain the difference based on my understanding;
As stated before, the initialize method allows you to set the values for an instance of an object upon object creation.
But what if you wanted to set the values later, after the instance was created, or change them after they have been initialized? This would be a scenario where you would use a setter method. THAT IS THE DIFFERENCE. You don’t have to “set” a particular state when you are using the attr_writer method initially.
The code below is an example of using a setter method to declare the value item_name for this instance of the Item class. Notice that we continue to use the getter method attr_reader so that we can get the values and print them to the screen, just in case you want to test the code on your own.
class Item
attr_reader :item_name
def item_name=(str)
#item_name = (str)
end
end
The code below is an example of using attr_writer to once again shorten our code and save us time.
class Item
attr_reader :item_name
attr_writer :item_name
end
item = Item.new
puts item.item_name = "TV"
The code below is a reiteration of the initialize example above of where we are using initialize to set the objects value of item_name upon creation.
class Item
attr_reader :item_name
def initialize(item_name)
#item_name = item_name
end
end
item = Item.new("TV")
puts item.item_name
attr_accessor: Performs the functions of both attr_reader and attr_writer, saving you one more line of code.
I think part of what confuses new Rubyists/programmers (like myself) is:
"Why can't I just tell the instance it has any given attribute (e.g., name) and give that attribute a value all in one swoop?"
A little more generalized, but this is how it clicked for me:
Given:
class Person
end
We haven't defined Person as something that can have a name or any other attributes for that matter.
So if we then:
baby = Person.new
...and try to give them a name...
baby.name = "Ruth"
We get an error because, in Rubyland, a Person class of object is not something that is associated with or capable of having a "name" ... yet!
BUT we can use any of the given methods (see previous answers) as a way to say, "An instance of a Person class (baby) can now have an attribute called 'name', therefore we not only have a syntactical way of getting and setting that name, but it makes sense for us to do so."
Again, hitting this question from a slightly different and more general angle, but I hope this helps the next instance of class Person who finds their way to this thread.
Simply put it will define a setter and getter for the class.
Note that
attr_reader :v is equivalant to
def v
#v
end
attr_writer :v is equivalant to
def v=(value)
#v=value
end
So
attr_accessor :v which means
attr_reader :v; attr_writer :v
are equivalant to define a setter and getter for the class.
Simply attr-accessor creates the getter and setter methods for the specified attributes
Another way to understand it is to figure out what error code it eliminates by having attr_accessor.
Example:
class BankAccount
def initialize( account_owner )
#owner = account_owner
#balance = 0
end
def deposit( amount )
#balance = #balance + amount
end
def withdraw( amount )
#balance = #balance - amount
end
end
The following methods are available:
$ bankie = BankAccout.new("Iggy")
$ bankie
$ bankie.deposit(100)
$ bankie.withdraw(5)
The following methods throws error:
$ bankie.owner #undefined method `owner'...
$ bankie.balance #undefined method `balance'...
owner and balance are not, technically, a method, but an attribute. BankAccount class does not have def owner and def balance. If it does, then you can use the two commands below. But those two methods aren't there. However, you can access attributes as if you'd access a method via attr_accessor!! Hence the word attr_accessor. Attribute. Accessor. It accesses attributes like you would access a method.
Adding attr_accessor :balance, :owner allows you to read and write balance and owner "method". Now you can use the last 2 methods.
$ bankie.balance
$ bankie.owner
Despite the large number of existing answers, none of them seems to me to explain the actual mechanism involved here. It's metaprogramming; it takes advantage of the following two facts:
You can modify a module / class on the fly
A module / class declaration is itself executable code
Okay, so imagine the following:
class Nameable
def self.named(whatvalue)
define_method :name do whatvalue end
end
end
We are declaring a class method named which, when called with a value, creates an instance method called name which returns that value. That is the metaprogramming part.
Now we'll subclass that class:
class Dog < Nameable
named "Fido"
end
What on earth did we just do? Well, in the class declaration, executable code executes with reference to the class. So the bare word named is actually a call to the class method named, which we inherited from Nameable; and we are passing the string "Fido" as the argument.
And what does the class method named do? It creates an instance method called name, which returns that value. So now, behind the scenes, Dog has a method that looks like this:
def name
"Fido"
end
Don't believe me? Then watch this little move:
puts Dog.new.name #=> Fido
Why did I tell you all that? Because what I just did with named for Nameable is almost exactly what attr_accessor does for Module. When you say attr_accessor you are calling a class method (inherited from Module) that creates instance methods. In particular, it creates a getter and setter method for the instance property whose name you provide as argument, so that you don't have to write those getter and setter methods yourself.
Defines a named attribute for this module, where the name is symbol.id2name, creating an instance variable (#name) and a corresponding access method to read it. Also creates a method called name= to set the attribute.
module Mod
attr_accessor(:one, :two)
end
Mod.instance_methods.sort #=> [:one, :one=, :two, :two=]
To summarize an attribute accessor aka attr_accessor gives you two free methods.
Like in Java they get called getters and setters.
Many answers have shown good examples so I'm just going to be brief.
#the_attribute
and
#the_attribute=
In the old ruby docs a hash tag # means a method.
It could also include a class name prefix...
MyClass#my_method
I am new to ruby and had to just deal with understanding the following weirdness. Might help out someone else in the future. In the end it is as was mentioned above, where 2 functions (def myvar, def myvar=) both get implicitly for accessing #myvar, but these methods can be overridden by local declarations.
class Foo
attr_accessor 'myvar'
def initialize
#myvar = "A"
myvar = "B"
puts #myvar # A
puts myvar # B - myvar declared above overrides myvar method
end
def test
puts #myvar # A
puts myvar # A - coming from myvar accessor
myvar = "C" # local myvar overrides accessor
puts #myvar # A
puts myvar # C
send "myvar=", "E" # not running "myvar =", but instead calls setter for #myvar
puts #myvar # E
puts myvar # C
end
end
Attributes and accessor methods
Attributes are class components that can be accessed from outside the object. They are known as properties in many other programming languages. Their values are accessible by using the "dot notation", as in object_name.attribute_name. Unlike Python and a few other languages, Ruby does not allow instance variables to be accessed directly from outside the object.
class Car
def initialize
#wheels = 4 # This is an instance variable
end
end
c = Car.new
c.wheels # Output: NoMethodError: undefined method `wheels' for #<Car:0x00000000d43500>
In the above example, c is an instance (object) of the Car class. We tried unsuccessfully to read the value of the wheels instance variable from outside the object. What happened is that Ruby attempted to call a method named wheels within the c object, but no such method was defined. In short, object_name.attribute_name tries to call a method named attribute_name within the object. To access the value of the wheels variable from the outside, we need to implement an instance method by that name, which will return the value of that variable when called. That's called an accessor method. In the general programming context, the usual way to access an instance variable from outside the object is to implement accessor methods, also known as getter and setter methods. A getter allows the value of a variable defined within a class to be read from the outside and a setter allows it to be written from the outside.
In the following example, we have added getter and setter methods to the Car class to access the wheels variable from outside the object. This is not the "Ruby way" of defining getters and setters; it serves only to illustrate what getter and setter methods do.
class Car
def wheels # getter method
#wheels
end
def wheels=(val) # setter method
#wheels = val
end
end
f = Car.new
f.wheels = 4 # The setter method was invoked
f.wheels # The getter method was invoked
# Output: => 4
The above example works and similar code is commonly used to create getter and setter methods in other languages. However, Ruby provides a simpler way to do this: three built-in methods called attr_reader, attr_writer and attr_acessor. The attr_reader method makes an instance variable readable from the outside, attr_writer makes it writeable, and attr_acessor makes it readable and writeable.
The above example can be rewritten like this.
class Car
attr_accessor :wheels
end
f = Car.new
f.wheels = 4
f.wheels # Output: => 4
In the above example, the wheels attribute will be readable and writable from outside the object. If instead of attr_accessor, we used attr_reader, it would be read-only. If we used attr_writer, it would be write-only. Those three methods are not getters and setters in themselves but, when called, they create getter and setter methods for us. They are methods that dynamically (programmatically) generate other methods; that's called metaprogramming.
The first (longer) example, which does not employ Ruby's built-in methods, should only be used when additional code is required in the getter and setter methods. For instance, a setter method may need to validate data or do some calculation before assigning a value to an instance variable.
It is possible to access (read and write) instance variables from outside the object, by using the instance_variable_get and instance_variable_set built-in methods. However, this is rarely justifiable and usually a bad idea, as bypassing encapsulation tends to wreak all sorts of havoc.
Hmmm. Lots of good answers. Here is my few cents on it.
attr_accessor is a simple method that helps us in cleaning(DRY-ing) up the repeating getter and setter methods.
So that we can focus more on writing business logic and not worry about the setters and getters.
The main functionality of attr_accessor over the other ones is the capability of accessing data from other files.
So you usually would have attr_reader or attr_writer but the good news is that Ruby lets you combine these two together with attr_accessor. I think of it as my to go method because it is more well rounded or versatile.
Also, peep in mind that in Rails, this is eliminated because it does it for you in the back end. So in other words: you are better off using attr_acessor over the other two because you don't have to worry about being to specific, the accessor covers it all. I know this is more of a general explanation but it helped me as a beginner.
Hope this helped!
I'm reading the Metaprogramming section of Programming Ruby 1.9 and I'm having trouble understanding what's going on internally between class_eval/class_exec vs. instance_eval/instance_exec.
So first of all, my understanding is that def adds a method to the method table of self (the class object):
class A
puts self # => A
def foo; 42; end # added to the method table of self, so becomes an instance method
end
A.new.foo # => 42
And if we use class_eval, we get the same behavior:
A.class_eval do
puts self # => A
def bar; 42; end # same as above
end
A.new.bar # => 42
But somehow in the instance_eval case, things are different:
A.instance_eval do
puts self # => A
def baz; 42; end # added to the method table of an anonymous
# singleton class of self, so becomes a class method
end
puts A.baz # => 42
s = 'string'
s.instance_eval do ... end # same behavior, so now def creates an instance method
So I understand the functional difference between class_eval and instance_eval.
But the contexts inside the class_eval and instance_eval blocks look exactly the same to me -- in particular, self points to the same object, and the local_variables are the same. So what's going on inside the blocks (internally) that's making def act different?
Is there some piece of documentation I could read? The RDoc for instance_eval and class_eval doesn't help. Looking at the source, instance_eval seems to set up a singleton class object whereas class_eval doesn't -- but is this difference visible outside the C code, on the Ruby level?
I think your confusion comes from the fact that def does not depend on the current self, you might think about it as being a "current class" that has it's own rules.
Following your examples:
class A
# defs here go to A
puts self # => A
class << self
#defs here go to A's eigenclass
end
end
A.class_eval do
#defs here go to A
end
A.instance_eval do
#defs here go to A's eigenclass
end
s = "Hello World"
class << s
#defs here go to s's eigenclass
end
Here's the portion of the chapter that talks about the issue and it's pretty clear about the behaviour
class_eval and instance_eval both set
self for the duration of the block.
However, they differ in the way they
set up the environment for method
definition. class_eval sets things up
as if you were in the body of a class
definition, so method definitions will
define instance methods In contrast,
calling instance_eval on a class acts
as if you were working inside the
singleton class of self. Therefore,
any methods you define will become
class methods.
The only thing I think is worth adding is that you can call instance_eval in any object, not just classes, and the behaviour doesn't change but has different consequences.
Some relevant reading:
Ruby: instance_eval and class_eval method definitions
Chapter 4 of this most excelent series
Just to add to #krusty.ar's answer: def and define_method add methods to the current method definition context (I believe that's what it's called, I'm not sure), not to the current self.
It's just that inside of a module, class or singleton class body, those two happen to be the same.
But, for example, in a script body (aka top-level), self is the main object, but the current method definition context is Object.