Why Test doubles concept in Ruby? - Purpose & Use Cases

The Big Idea

What if you could test your code without waiting for slow or tricky parts every time?

The Scenario

Imagine you are testing a car's braking system, but you have to test it on a real car every time. You need the car, the road, and a safe place to stop. This is like testing code that depends on other parts that are hard to use or slow.

The Problem

Testing with real parts can be slow, risky, and sometimes impossible. You might wait a long time for a response or cause errors that are hard to fix. It's like waiting for a friend to answer the phone every time you want to talk.

The Solution

Test doubles act like pretend parts or friends. They stand in for the real things during tests, so you can check your code quickly and safely without needing the real parts every time.

Before vs After

✗ Before

real_car.brake
assert real_car.stopped?

✓ After

fake_car = TestDouble.new
fake_car.brake
assert fake_car.stopped?

What It Enables

It lets you test your code faster and more reliably by using simple stand-ins instead of the real, complex parts.

Real Life Example

When building a website that sends emails, you don't want to send real emails every time you test. Using test doubles, you can pretend to send emails and check if your code tries to send them correctly.

Key Takeaways

Manual testing with real parts is slow and risky.

Test doubles replace real parts with simple stand-ins.

This makes testing faster, safer, and easier to control.

Practice

1.

What is the main purpose of using test doubles in Ruby testing?

easy

A. To permanently change the behavior of classes

B. To speed up the program execution in production

C. To replace real objects with fake ones to isolate tests

D. To automatically generate user interface elements

Solution

1. Step 1: Understand what test doubles do

   Test doubles are fake objects used in tests to replace real parts of the program.

2. Step 2: Identify the main goal

   The goal is to isolate the part being tested by removing outside effects from real objects.

3. Final Answer:

   To replace real objects with fake ones to isolate tests -> Option C

4. Quick Check:

   Test doubles isolate tests = A [OK]

Hint: Test doubles fake objects to isolate tests quickly [OK]

Common Mistakes:

• Thinking test doubles speed up production code
• Believing test doubles permanently change classes
• Confusing test doubles with UI tools

2.

Which of the following is the correct way to create a test double named user in Ruby using RSpec?

?

easy

A. user = fake('user')

B. user = double('user')

C. user = stub('user')

D. user = mock_object('user')

Solution

1. Step 1: Recall RSpec syntax for test doubles

   RSpec uses the method double to create test doubles.

2. Step 2: Match the correct method call

   The correct syntax is double('name'), so double('user') is right.

3. Final Answer:

   user = double('user') -> Option B

4. Quick Check:

   RSpec double creation = A [OK]

Hint: Use double('name') to create test doubles in RSpec [OK]

Common Mistakes:

• Using fake or stub instead of double
• Confusing mock_object with double
• Missing quotes around the name

3.

What will be the output of this Ruby code using a test double?

user = double('user')
allow(user).to receive(:name).and_return('Alice')
puts user.name

medium

A. Alice

B. nil

C. user

D. NoMethodError

Solution

1. Step 1: Understand the double setup

   The double user is told to respond to name with 'Alice'.

2. Step 2: Check the method call output

   Calling user.name returns 'Alice' as set by and_return.

3. Final Answer:

   Alice -> Option A

4. Quick Check:

   Double returns stubbed value = C [OK]

Hint: Stub method with and_return to control output [OK]

Common Mistakes:

• Expecting the double name instead of stubbed return
• Thinking it returns nil by default
• Assuming NoMethodError without stub

4.

Identify the error in this test double usage and how to fix it:

order = double('order')
allow(order).to receive(:total)
puts order.total

medium

A. Missing and_return causes order.total to return nil, fix by adding and_return

B. Double name should be a symbol, fix by using double(:order)

C. Cannot call allow on a double, fix by using expect

D. Missing puts before order.total

Solution

1. Step 1: Analyze the stub setup

   The method total is allowed but no return value is set, so it returns nil.

2. Step 2: Fix by adding return value

   Add and_return(value) to specify what total should return.

3. Final Answer:

   Missing and_return causes order.total to return nil, fix by adding and_return -> Option A

4. Quick Check:

   Stub needs return value = B [OK]

Hint: Always add and_return to stub methods for expected output [OK]

Common Mistakes:

• Thinking double name must be symbol
• Confusing allow with expect usage
• Forgetting to print output

5.

You want to test a PaymentProcessor class that calls an external Gateway service. How can you use a test double to ensure your test does not call the real Gateway?

class PaymentProcessor
  def initialize(gateway)
    @gateway = gateway
  end

  def pay(amount)
    @gateway.charge(amount)
  end
end

# Test code here?

hard

A. Remove the gateway parameter and hardcode the call inside PaymentProcessor

B. Call the real Gateway class directly in the test

C. Use allow_any_instance_of(Gateway) to stub charge

D. Create a double for gateway that expects charge with amount and returns true

Solution

1. Step 1: Understand dependency injection

   The PaymentProcessor takes a gateway object to call charge.

2. Step 2: Use a test double for gateway

   Create a double that expects charge with the amount and returns true to simulate success.

3. Step 3: Avoid calling real external service

   This prevents the test from calling the real Gateway, isolating the test.

4. Final Answer:

   Create a double for gateway that expects charge with amount and returns true -> Option D

5. Quick Check:

   Use doubles to isolate external calls = D [OK]

Hint: Inject doubles to replace external services in tests [OK]

Common Mistakes:

• Calling real external services in tests
• Hardcoding dependencies inside classes
• Using allow_any_instance_of unnecessarily