Testing Procedures

1. Build Commands

Windows

scripts/build_windows.bat Debug

Linux

scripts/build_linux.sh

NEVER

• Run cmake directly
• Use WSL for Windows builds

2. Test Execution

Windows

./build-windows/bin/kalahari-tests.exe

Linux

./build-linux/bin/kalahari-tests

3. Test Framework

• Framework: Catch2 v3
• Style: BDD (Behavior-Driven Development)
• Location: tests/

4. Result Interpretation

Output format

[PASS] TestName
[FAIL] TestName - expected X, got Y

Summary

===============================================
All tests passed (42 assertions in 15 test cases)

or

===============================================
test cases: 15 | 14 passed | 1 failed
assertions: 42 | 40 passed | 2 failed

5. What to Check

After implementation

• All existing tests still pass?
• No regressions (previously passing tests)?
• New tests added for new functionality?

Test coverage

• Core business logic tested?
• Edge cases covered?
• Error handling tested?

6. Manual Testing

When needed

• UI changes (visual verification)
• User interaction flows
• Theme switching
• Panel resizing/docking

Steps

1. Run application: ./build-windows/bin/kalahari.exe
2. Test the specific feature
3. Verify visual appearance
4. Check responsiveness

7. Reporting Results

Pass

{
  "decision": "pass",
  "tests": "42/42 passed",
  "summary": "All tests pass"
}

Fail

{
  "decision": "fail",
  "tests": "40/42 passed",
  "failures": [
    "TestSettings::save - expected true, got false",
    "TestDocument::load - file not found"
  ],
  "summary": "2 tests failed"
}

8. Common Issues

Build fails

1. Check CMake output for errors
2. Verify vcpkg dependencies
3. Check for missing includes

Tests fail

1. Read failure message carefully
2. Check if test assumptions still valid
3. Verify test data/fixtures

Flaky tests

1. Check for timing dependencies
2. Look for shared state between tests
3. Verify test isolation