dance-lessons-coach/adr/0024-bdd-test-organization-and-isolation.md

ADR 0024: BDD Test Organization and Isolation Strategy

Status

Accepted ✅

Context

As the dance-lessons-coach project grows, our BDD test suite has encountered several challenges. While we initially followed basic Godog patterns, we need to evolve our organization to handle complex scenarios like config hot reloading while maintaining test reliability.

Current Issues

1. Test Interdependence: Tests affect each other through shared state (config files, database)
2. Timing Issues: Config reloading and server restarts cause race conditions
3. Cognitive Load: Large test files with many scenarios are hard to maintain
4. Flaky Tests: Tests pass individually but fail when run together
5. Edge Case Handling: Special setup/teardown requirements for certain tests

Godog Best Practices Alignment

According to Godog documentation and community best practices, our current organization partially follows recommendations but needs improvement in:

• Feature Granularity: Some files contain multiple unrelated features
• Step Organization: Steps could be better grouped by domain
• Context Management: Need better state isolation between scenarios
• Tagging Strategy: Currently missing tag-based test selection

Decision

Adopt a modular, isolated test suite architecture with the following principles:

1. Test Organization by Feature (Godog-Aligned)

Following Godog best practices, we organize tests by business domain with proper feature granularity:

features/
├── auth/                          # Business domain
│   ├── authentication.feature     # Single feature per file
│   ├── password_reset.feature     # Single feature per file
│   └── user_management.feature    # Single feature per file
├── config/                        # Business domain
│   ├── hot_reloading.feature       # Single feature per file
│   └── validation.feature          # Single feature per file
├── greet/                         # Business domain
│   ├── v1_greeting.feature         # Single feature per file
│   └── v2_greeting.feature         # Single feature per file
├── health/                        # Business domain
│   └── health_check.feature        # Single feature per file
└── jwt/                           # Business domain
    ├── secret_rotation.feature     # Single feature per file
    └── retention_policy.feature   # Single feature per file

Key Improvements over current structure:

• ✅ Single responsibility: One feature per file
• ✅ Business alignment: Grouped by domain, not technical concerns
• ✅ Scalability: Easy to add new features without bloating files

2. Isolation Strategies

A. Config File Isolation

• Each feature directory has its own config file pattern
• Config files are cleaned up after each feature test run
• Example: features/auth/auth-test-config.yaml

B. Database Isolation

• Use separate database schemas or suffixes per feature
• Example: dance_lessons_coach_auth_test, dance_lessons_coach_greet_test

C. Server Port Isolation

• Assign different ports to different test groups
• Prevents port conflicts during parallel testing

3. Test Execution Strategy

Option 1: Sequential Feature Testing (Recommended)

# Run tests by feature group
./scripts/test-feature.sh auth
./scripts/test-feature.sh config
./scripts/test-feature.sh greet

Option 2: Parallel Feature Testing (Advanced)

# Run features in parallel with isolation
./scripts/test-all-features-parallel.sh

4. Test Synchronization (Godog Best Practices)

A. Explicit Waits with Timeouts

Following Godog's arrange-act-assert pattern:

// Instead of fixed sleep times
func waitForServerReady(maxAttempts int, delay time.Duration) error {
    for i := 0; i < maxAttempts; i++ {
        if serverIsReady() {
            return nil
        }
        time.Sleep(delay)
    }
    return fmt.Errorf("server not ready after %d attempts", maxAttempts)
}

B. Godog Context Management

Implement proper context structs as recommended by Godog:

// Feature-specific context for isolation
type AuthContext struct {
    client *testserver.Client
    db     *sql.DB
    users  map[string]UserData
}

func InitializeAuthContext() *AuthContext {
    return &AuthContext{
        client: testserver.NewClient(),
        db:     connectToFeatureDB("auth"),
        users:  make(map[string]UserData),
    }
}

func CleanupAuthContext(ctx *AuthContext) {
    // Cleanup resources
    ctx.db.Close()
}

C. Tag-Based Test Selection

Add Godog tag support for selective test execution:

// In feature files
@smoke @auth
Scenario: Successful user authentication
  Given the server is running
  When I authenticate with valid credentials
  Then the authentication should be successful

// Run specific tags
go test ./features/... -tags=smoke
godog --tags=@auth features/

B. Event-Based Synchronization

// Use server lifecycle events
func waitForConfigReload() error {
    return waitForEvent("config_reloaded", 30*time.Second)
}

C. Test Hooks with Timeouts

// In test setup
ctx.Step("^I wait for v2 API to be enabled$", func() error {
    return waitForCondition(30*time.Second, func() bool {
        return v2EndpointAvailable()
    })
})

5. Test Lifecycle Management

Before Suite (Feature Level)

func InitializeFeatureSuite(featureName string) {
    // Setup feature-specific resources
    initDatabaseForFeature(featureName)
    createFeatureConfigFile(featureName)
    startIsolatedServer(featureName)
}

After Suite (Feature Level)

func CleanupFeatureSuite(featureName string) {
    // Cleanup feature-specific resources
    cleanupDatabaseForFeature(featureName)
    removeFeatureConfigFile(featureName)
    stopIsolatedServer(featureName)
}

6. Shell Script Integration

Create feature-specific test scripts:

# scripts/test-feature.sh
#!/bin/bash

FEATURE=$1
DATABASE="dance_lessons_coach_${FEATURE}_test"
CONFIG="features/${FEATURE}/${FEATURE}-test-config.yaml"

# Setup
setup_feature_environment() {
    echo "🧪 Setting up ${FEATURE} feature tests..."
    create_database ${DATABASE}
    generate_config ${CONFIG}
}

# Run tests
run_feature_tests() {
    echo "🚀 Running ${FEATURE} feature tests..."
    DLC_DATABASE_NAME=${DATABASE} \
    DLC_CONFIG_FILE=${CONFIG} \
    go test ./features/${FEATURE}/... -v
}

# Teardown
cleanup_feature_environment() {
    echo "🧹 Cleaning up ${FEATURE} feature tests..."
    drop_database ${DATABASE}
    remove_config ${CONFIG}
}

# Main execution
setup_feature_environment
run_feature_tests
cleanup_feature_environment

7. Configuration Management

Feature-Specific Config Files

# features/auth/auth-test-config.yaml
server:
  host: "127.0.0.1"
  port: 9192  # Feature-specific port

database:
  name: "dance_lessons_coach_auth_test"  # Feature-specific database

api:
  v2_enabled: true  # Feature-specific settings

auth:
  jwt:
    ttl: 1h

8. Test Data Management

A. Feature-Scoped Data

• Each feature gets its own data namespace
• Example: auth_user_*, greet_message_* prefixes

B. Automatic Cleanup

func CleanupFeatureData(featureName string) {
    // Remove all data created by this feature
    db.Exec(fmt.Sprintf("DELETE FROM %s_* WHERE feature = '%s'", featureName, featureName))
}

Consequences

Positive

1. Improved Test Reliability: Tests don't interfere with each other
2. Better Maintainability: Smaller, focused test files
3. Faster Development: Run only relevant tests during feature development
4. Easier Debugging: Isolate issues to specific features
5. Parallel Testing: Enable safe parallel execution
6. SOLID Compliance: Single responsibility for test files

Negative

1. Increased Complexity: More moving parts in test infrastructure
2. Resource Usage: Multiple databases/servers consume more resources
3. Setup Time: Initial test runs may be slower due to setup
4. Learning Curve: Team needs to understand the isolation patterns

Neutral

1. Test Execution Time: May increase or decrease depending on parallelization
2. CI/CD Changes: Pipeline needs adaptation for new test organization

Implementation Plan

Phase 1: Refactor Current Tests (1-2 weeks)

1. Split monolithic feature files into feature directories
2. Create feature-specific test scripts
3. Implement basic isolation (config files, database names)

Phase 2: Enhance Test Infrastructure (2-3 weeks)

1. Add synchronization helpers to test framework
2. Implement server lifecycle management
3. Create comprehensive cleanup routines

Phase 3: Parallel Testing (Optional)

1. Add parallel test execution capability
2. Implement port management for parallel runs
3. Add resource monitoring

Alternatives Considered

1. Single Test Suite with Better Cleanup

Rejected because: Doesn't solve fundamental interdependence issues

2. Docker-Based Isolation

Rejected because: Too heavyweight for local development

3. Test Virtualization

Rejected because: Overkill for current project size

Success Metrics

1. Test Reliability: >95% pass rate in CI/CD
2. Test Isolation: Ability to run any single feature test independently
3. Developer Experience: Feature tests run in <30 seconds locally
4. Maintainability: New team members can understand test structure in <1 hour

References

Godog Official Resources

• Godog GitHub Repository
• Godog Documentation

BDD Best Practices

• BDD Best Practices
• Alice GG • BDD in Golang
• Scrap Your TDD for BDD: Part II

Test Organization Patterns

• Test Server Implementation
• Optimizing Godog Test Execution

Revision History

• 2026-04-09: Initial draft based on BDD test challenges
• 2026-04-09: Added implementation details and examples

Decision Makers

• Approved by: Gabriel Radureau
• Consulted: AI Agent (Mistral Vibe)
• Informed: Development Team

Future Considerations

1. Test Impact Analysis: Track which tests are affected by code changes
2. Flaky Test Detection: Automatically identify and quarantine flaky tests
3. Performance Benchmarking: Monitor test execution times over time
4. Test Coverage Visualization: Feature-level coverage reports

---

Status: 🟡 Proposed → Ready for team review and implementation

Note: This ADR complements ADR 0023 (Config Hot Reloading) by addressing the test organization aspects of hot reloading functionality.