Mutation Testing

Technique to measure how well tests detect bugs by injecting mutations.

Context

You are using mutation testing to validate test quality. High mutation score means tests actually verify behavior.

Domain Context

• Mutation: Small code change; operator flip, boundary change, constant modification
• Killed: Test fails on mutation; test is effective
• Survived: Test passes on mutation; gap in test coverage
• Goal: 80%+ mutation score indicates strong tests
• Tools: PIT (Java), Stryker (JavaScript), Mutant (Ruby)

Instructions

1. Run Baseline: Execute test suite; note coverage
2. Inject Mutations: Tool modifies code (< becomes <=, true becomes false)
3. Run Tests: Check if tests catch each mutation
4. Identify Survivors: Which mutations did tests miss? Why?
5. Strengthen Tests: Add tests for survivor scenarios
6. Repeat: Run mutation testing again; iterate until satisfied

Anti-Patterns

• Ignoring mutation testing; high code coverage doesn't guarantee good tests
• Trying to reach 100% mutation score; diminishing returns after 80%
• Adding tests just to kill mutations; tests should be meaningful
• Slow mutation testing blocking development; run on critical paths
• Not understanding why mutations survived; just adding tests blindly doesn't help

Further Reading

• Richard Lipton, "The Fault-Based Testing (FBT) Paradigm"
• PIT documentation (Java mutation testing)
• Stryker documentation (JavaScript/TypeScript mutation testing)