analyze-test-failures

Analyze Test Failures

Analyze failing test cases with a balanced, investigative approach.

Context

When tests fail, there are two primary possibilities:

1. False positive: The test itself is incorrect
2. True positive: The test discovered a genuine bug

Assuming tests are wrong by default is a dangerous anti-pattern that defeats the purpose of testing.

Analysis Process

1. Initial Analysis

• Read the failing test carefully, understanding its intent
• Examine the test's assertions and expected behavior
• Review the error message and stack trace

2. Investigate the Implementation

• Check the actual implementation being tested
• Trace through the code path that leads to the failure
• Verify that implementation matches documented behavior

3. Apply Critical Thinking

For each failing test, ask:

• What behavior is the test trying to verify?
• Is this behavior clearly documented or implied by the API design?
• Does the current implementation actually provide this behavior?
• Could this be an edge case the implementation missed?

4. Make a Determination

Classify the failure as one of:

Classification	Meaning
Test Bug	Test's expectations are incorrect
Implementation Bug	Code doesn't behave as it should
Ambiguous	Intended behavior is unclear

5. Document Reasoning

Provide clear explanation including:

• Evidence supporting the conclusion
• Specific mismatch between expectation and reality
• Recommended fix (to test or implementation)

Example Analyses

Example 1: Ambiguous Behavior

Scenario: Test expects calculateDiscount(100, 0.2) to return 20, but it returns 80

Analysis:

• Test assumes function returns discount amount
• Implementation returns price after discount
• Function name is ambiguous

Determination: Ambiguous Recommendation: Check documentation or clarify intended behavior

Example 2: Implementation Bug

Scenario: Test expects validateEmail("user@example.com") to return true, but it returns false

Analysis:

• Test provides a valid email format
• Implementation regex is missing support for dots in domain
• Other valid emails also fail

Determination: Implementation Bug Recommendation: Fix the regex to properly validate email addresses per RFC standards

Example 3: Test Bug

Scenario: Test expects divide(10, 0) to return 0, but it throws an error

Analysis:

• Test assumes division by zero returns 0
• Implementation throws DivisionByZeroError
• Standard mathematical behavior is to treat as undefined/error

Determination: Test Bug Recommendation: Update test to expect an error, not 0

Output Format

For each failing test, provide:

Test: [test name/description]
Failure: [what failed and how]

Investigation:
- Test expects: [expected behavior]
- Implementation does: [actual behavior]
- Root cause: [why they differ]

Determination: [Test Bug | Implementation Bug | Ambiguous]

Recommendation:
[Specific fix to either test or implementation]

Key Principles

• NEVER automatically assume the test is wrong
• ALWAYS consider that the test might have found a real bug
• When uncertain, lean toward investigating the implementation
• Tests are often your specification - they define expected behavior
• A failing test is a gift - it's either catching a bug or clarifying requirements

Related Skills

• test-failure-mindset: Set investigative approach for session
• comprehensive-test-review: Full test suite review