Use fixtures in testing environment to reduce errors and create consistent and repeatable tests.

fixture = Fixture.define do
  def self.example_class
    define_class do
      def foo
        "bar"
      end
    end
  end

  def self.example
    example_class.new
  end
end

some_fixture = fixture.example

some_fixture.foo == "bar"
# => true

In the above code, some_fixture is an instance of a class with an outermost namespace of Fixture, a randomly assigned UUID as a module within Fixture, and a randomly assigned UUID as a class within the module, i.e.

puts fixture.example
# => <Fixture::M1bd19401d8674d54a87f7871511f0daf::
# C9980bf7c3cbd42df88cc9bc35b251d50:0x26f67b76>

This allows each set of tests to run independently of one another, but still contains all test data within the Fixture namespace.

fixture = Fixture.define(:SomeModule) do
  def self.naming_example_class
    define_class(:SomeClass) do
      def foo
        "bar"
      end
    end
  end

  def self.naming_example
    naming_example_class.new
  end
end

In this case, a module name SomeModule and a class name SomeClass are explicitly passed in to the Fixture.

puts fixture.naming_example
# => <Fixture::Mdaa0e47242934a8db244fc91363a4e0d::SomeModule::SomeClass:0x26f67b76>

The fixture retains the randomly assigned UUID module namespace directly beneath the Fixture namespace so that the test environments are still guaranteed to be independent. The SomeModule namespace and the SomeClass class name follow.

The fixture library is released under the MIT License.