Confusing Behaviour of Const_Get in Ruby

Confusing behaviour of const_get in Ruby?

You are correct to be confused... The doc didn't state that Ruby makes a special case for lookup of constants in Modules and has been modified to state this explicitly. If the constant has not been found in the normal hierarchy, Ruby restarts the lookup from Object, as can be found in the source.

Constant lookup by itself can be bit confusing. Take the following example:

module M
  Foo = :bar
  module N
    # Accessing Foo here is fine:
    p Foo # => bar
  end
end
module M::N
  # Accessing Foo here isn't
  p Foo  # => uninitialized constant M::N::Foo
end
p M::N.const_get :Foo  # => uninitialized constant M::N::Foo

In both places, though, accessing Object level constants like Array is fine (thank god!). What's going on is that Ruby maintains a list of "opened Module definitions". If a constant has an explicit scope, say LookHereOnly::Foo, then only LookHereOnly and its included modules will be searched. If no scope is specified (like Foo in the example above), Ruby will look through the opened module definitions to find the constant Foo: M::N, then M and finally Object. The topmost opened module definition is always Object.

So M::N.const_get :Foo is equivalent to accessing Foo when the opened classes are only M::N and Object, like in the last part of my example.

I hope I got this right, coz I'm still confused by constant lookups myself :-)

Confusing behaviour between class and object

Here is a little different way to define a classes/objects:

klass = Class.new {
  puts self
  def bar
    puts self
  end
}
# #<Class:0x3fbdbb8>

As you can see, it outputs #<Class:0x3fbdbb8>. So its class is Class. You can check it via class method:

klass.class
# => Class

When you name your class with uppercase letter (for example Example) for the first time, it use it as output instead of cryptic name like: #<Class:0x3fbdbb8>.

Foo = Class.new {
  puts self
  def bar
    puts self
  end
}

Still outputs cryptic name because Foo ='s part hasn't been evaluated yet.

puts Foo

Outputs correct name - Foo

What if I name it again?

Qux = Foo
# => Foo

Nothing. Foo will be Foo forever.

Strange ruby behaviour with __FILE__ constant?

What you want is to expand the path properly:

# Affected by the current working directory, etc.
puts __FILE__

# Always an absolute path
puts File.expand_path(__FILE__, Dir.getwd)

This takes your current working directory into account.

How to judge whether a Constant has been defined within a module namespace, not a global one?

Indeed, const_defined? and const_get go through the hierarchy, which for modules include (artificially) the Object class. You can use Module#constants to avoid this, though:

module Admin  
  A = 'A Const with in the Admin namespace' unless constants.include?(:A)
end

Note: In Ruby 1.8, you check against "A", not :A

Confusing behavior of File.dirname

File.dirname should return an absolute path if given an absolute path, and a relative path if given a relative path:

File.dirname('/home/jon/test.rb') # => '/home/jon'
File.dirname('test.rb') # => '.'

__FILE__ returns the name of the current script, which is therefore a relative path from the current directory. That means you should always use expand_path if you want to get the absolute path with File.dirname(__FILE__).

NB Ruby 2.0.0 introduces the __dir__ constant

Why doesn't the Digest gem show all the class as constants?

The magic is in the const_missing singleton method in the Digest module. @knugje has pointed out a very useful link in his answer which makes in clearer.

ri documentation about const_missing says:

Invoked when a reference is made to an undefined constant in mod

So here's the trick:

module X
    def self.const_missing(name)
        name
    end
end

p X.constants        # => []
p X::Y               # => :Y
p X.constants        # => []

So the code doesn't define the constant in the module. So we can use a fairly simple trick:

module X
    define_singleton_method(:const_missing) { |n| const_set(n, n.id2name) }
end

p X.constants      # => []
p X::Y             # => "Y"
p X::Z             # => "Z"
p X.constants      # => [:Y, :Z]
p X::Z             # "Z"

That's how const_missing is used here. But Instead of const_set, what digest does is loading up a file with the name and do some exception handling and checking if the constants exist on the file.

Personally speaking, this may create confusion. For example, when you have no document to offline and online documents, you can't just simply run Digest.constants to list the available constants.

As knugie pointed out, you can read about the const_missing implementation in the Digest module here.

Why do Global Variables exist in Ruby?

Global variables are not bad. They're not evil. They're just incredibly, incredibly powerful. Which is why you shouldn't use them.

Global variables are global- they can be accessed and modified anywhere in the code. A single global variable has the potential to affect all of your classes, all of your functions, all of the classes and functions of every single library or dependency you load into your project, and all of the classes and functions of every single project which loads your project as a dependency, as well as the projects that load those projects, and so and and so forth, for ever and always, for the rest of time.

The second people start feeling comfortable using global variables, the namespace gets insanely cluttered and we get conflicts left and right and the stability of the programming language itself is threatened. Which is why the use of global variables is emphatically and repeatedly discouraged.

But global variables are not bad. They're like the highway lanes labeled "for emergency vehicles only," or like those fire-axes behind glass labeled "break glass in case of emergency."

It's entirely possible that at some point, in the distant future, you will have an incredibly unusual situation which merits the use of a single global variable. But that day is not today. And it is probably not tomorrow, or a month from now, or a year from now. Daily life, daily code- it just doesn't call for the unbridled power of a global variable.

---

$stdout is a great example of why global variables are sometimes important. $stdout is the default stream in ruby- the one where things will print if no other stream is specified. $stdout should be accessible from every class and every function in every library because it acts like a giant funnel, shoveling all output to a single location. The whole world knows and agrees that $stdout exists in ruby, and its uses are well-documented, so its power is well-managed.

This isn't to be confused with STDOUT which is a constant representing the actual pipe which sets up a stream between ruby and its parent program (usually a terminal). $stdout = STDOUT by default, but $stdout can be changed to anything. If you want your program to print to a file, you can change $stdout to a file stream.

I don't think this name choice is confusing for a seasoned rubyist. A variable is designed to be modified and a constant is designed to be constant. The difference between $stdout and STDOUT is that the former can be modified to change the standard output location of your program, and the latter is a constant, always pointing to the stdout stream. The capitalization makes a world of difference and conveys very different meanings.

---

As for why global constants are uninitialized and global variables are nil, that actually has nothing to do with globals. Ruby automatically initializes all variables as nil. You can easily see this with instance variables such as @foo or @@foo. In almost every situation, an undefined local variable will throw a NameError because ruby cannot tell whether it is a variable or a method. But in strange situations, they too are initialized as nil:

puts foo # => NameError: undefined local variable or method 'foo'
foo = 42 if false
puts foo # => nil

puts bar # => NameError
bar = bar
puts bar # => nil

It was a conscious design choice in Ruby not to automatically initialize constants. Because a constant is, by definition, something which is initialized once and then never changed, it would break the definition for a constant to be nil at first and then a different value later in the code.

---

I should also mention that global constants are considered acceptable, even among people who tout global variables as bad. The difference is that constants can only be assigned once and generally throw a warning or error if they get assigned again. This protects programmers from situations where conflicting global constants might cause problems.

---

Related Topics