🔍 Exclude Generated Code and Mocks from Test Coverage #235

New issue

Open

opened 2025-08-08 07:15:28 +00:00 by goern · 0 comments

goern commented

2025-08-08 07:15:28 +00:00

(Migrated from codeberg.org)

🎯 Problem Statement

Currently, our test coverage includes generated code and mock implementations, which inflates coverage metrics and doesn't reflect the actual quality of our hand-written code.

📋 Requirements

Code Types to Exclude from Coverage:

1. Generated Code

Prisma generated files: src/lib/generated/**
Any codegen output: **/*.generated.{js,jsx,ts,tsx}
GraphQL codegen (if used in future)
Protobuf generated code (if used)

2. Mock Implementations

Test mocks: **/__mocks__/**
Development mocks: src/mocks/**
Jest setup mocks: tests/**/__mocks__/**

3. Framework Boilerplate

Next.js default files: layout.tsx, error.tsx, not-found.tsx
Page components: page.tsx (tested via E2E instead)
Type-only files: Pure TypeScript definition files

4. Third-party Integration Code

Vendor code: src/vendor/**
External integrations: src/external/**

🔧 Implementation Tasks

Phase 1: Update Jest Configuration

Update jest.config.base.ts with comprehensive exclusion patterns
Test exclusions work correctly
Verify coverage reports exclude specified files

Phase 2: Update CI/CD Pipeline

Update .forgejo/workflows/test-coverage.yml
Ensure consistent exclusions across environments
Validate coverage thresholds still meaningful

Phase 3: Documentation

Update testing strategy documentation
Create exclusion policy guidelines
Document rationale for each exclusion category

✅ Success Criteria

Coverage reports only include hand-written code
Generated files have 0% coverage impact
Mock files are excluded from coverage metrics
Coverage thresholds remain achievable and meaningful
CI/CD pipeline respects all exclusions

📊 Expected Impact

More accurate coverage metrics reflecting actual code quality
Reduced false positives in coverage reports
Clearer focus on testing business logic vs. generated code
Improved developer experience with realistic coverage targets

#234 - Increase Test Coverage to 80% Global Target
Related to overall testing strategy improvements

📝 Notes

This follows industry best practices where:

Generated code is assumed to be tested by the generator
Mock code is implementation detail, not business logic
Framework boilerplate is tested by the framework itself

Priority: High - Should be implemented before finalizing coverage targets

## 🎯 Problem Statement Currently, our test coverage includes generated code and mock implementations, which inflates coverage metrics and doesn't reflect the actual quality of our hand-written code. ## 📋 Requirements ### Code Types to Exclude from Coverage: #### 1. Generated Code - **Prisma generated files**: `src/lib/generated/**` - **Any codegen output**: `**/*.generated.{js,jsx,ts,tsx}` - **GraphQL codegen** (if used in future) - **Protobuf generated code** (if used) #### 2. Mock Implementations - **Test mocks**: `**/__mocks__/**` - **Development mocks**: `src/mocks/**` - **Jest setup mocks**: `tests/**/__mocks__/**` #### 3. Framework Boilerplate - **Next.js default files**: `layout.tsx`, `error.tsx`, `not-found.tsx` - **Page components**: `page.tsx` (tested via E2E instead) - **Type-only files**: Pure TypeScript definition files #### 4. Third-party Integration Code - **Vendor code**: `src/vendor/**` - **External integrations**: `src/external/**` ## 🔧 Implementation Tasks ### Phase 1: Update Jest Configuration - [ ] Update `jest.config.base.ts` with comprehensive exclusion patterns - [ ] Test exclusions work correctly - [ ] Verify coverage reports exclude specified files ### Phase 2: Update CI/CD Pipeline - [ ] Update `.forgejo/workflows/test-coverage.yml` - [ ] Ensure consistent exclusions across environments - [ ] Validate coverage thresholds still meaningful ### Phase 3: Documentation - [ ] Update testing strategy documentation - [ ] Create exclusion policy guidelines - [ ] Document rationale for each exclusion category ## ✅ Success Criteria 1. **Coverage reports only include hand-written code** 2. **Generated files have 0% coverage impact** 3. **Mock files are excluded from coverage metrics** 4. **Coverage thresholds remain achievable and meaningful** 5. **CI/CD pipeline respects all exclusions** ## 📊 Expected Impact - **More accurate coverage metrics** reflecting actual code quality - **Reduced false positives** in coverage reports - **Clearer focus** on testing business logic vs. generated code - **Improved developer experience** with realistic coverage targets ## 🔗 Related Issues - #234 - Increase Test Coverage to 80% Global Target - Related to overall testing strategy improvements ## 📝 Notes This follows industry best practices where: - Generated code is assumed to be tested by the generator - Mock code is implementation detail, not business logic - Framework boilerplate is tested by the framework itself **Priority**: High - Should be implemented before finalizing coverage targets