Add "Writing-Tests"

John McCardle 2025-10-25 22:38:06 +00:00
parent 3b9f342ef1
commit ea76c22954
1 changed files with 512 additions and 0 deletions

512
Writing-Tests.-.md Normal file

@ -0,0 +1,512 @@
# Writing Tests
Guide to creating automated tests for McRogueFace using Python and the automation API.
## Quick Reference
**Test Location:** `tests/` directory (NOT `build/tests/` - that gets shipped!)
**Test Types:**
1. **Direct Execution** - No game loop, immediate results
2. **Timer-Based** - Requires rendering/game loop
**Key Tools:**
- `mcrfpy.automation` - Screenshot and input automation
- `--headless` flag - Run without display
- `--exec` flag - Execute specific script
**Format:** See CLAUDE.md "Testing Guidelines" section
---
## Test Type 1: Direct Execution
For tests that don't need rendering or game state.
### Template
```python
#!/usr/bin/env python3
"""Test description goes here."""
import mcrfpy
import sys
def test_feature():
# Setup
obj = mcrfpy.SomeClass()
# Test
obj.property = 42
# Verify
assert obj.property == 42, f"Expected 42, got {obj.property}"
# More tests...
return True
if __name__ == "__main__":
try:
if test_feature():
print("PASS")
sys.exit(0)
else:
print("FAIL")
sys.exit(1)
except Exception as e:
print(f"FAIL: {e}")
import traceback
traceback.print_exc()
sys.exit(1)
```
### Running
```bash
cd build
./mcrogueface --headless --exec ../tests/test_myfeature.py
```
### Example: Testing Vector Operations
```python
import mcrfpy
import sys
def test_vector():
# Create vector
v = mcrfpy.Vector(3.0, 4.0)
# Test properties
assert v.x == 3.0
assert v.y == 4.0
# Test tuple access
assert v[0] == 3.0
assert v[1] == 4.0
# Test modification
v.x = 10.0
assert v.x == 10.0
print("All vector tests passed")
return True
if __name__ == "__main__":
if test_vector():
print("PASS")
sys.exit(0)
else:
print("FAIL")
sys.exit(1)
```
---
## Test Type 2: Timer-Based Tests
For tests requiring rendering, screenshots, or game loop.
### Template
```python
#!/usr/bin/env python3
"""Test description - requires game loop."""
import mcrfpy
from mcrfpy import automation
import sys
# Setup scene BEFORE game loop starts
mcrfpy.createScene("test")
# Create UI elements
frame = mcrfpy.Frame(100, 100, 200, 150)
frame.fill_color = mcrfpy.Color(255, 0, 0)
mcrfpy.sceneUI("test").append(frame)
# Switch to test scene
mcrfpy.setScene("test")
def run_test(runtime_ms):
"""Timer callback - runs AFTER game loop starts."""
# Take screenshot
automation.screenshot("test_output.png")
# Perform interactions
automation.click(150, 150) # Click on frame
# Verify results
# ... check state, take more screenshots, etc ...
# Report results
print("PASS")
sys.exit(0) # MUST exit!
# Schedule test to run after game loop starts
mcrfpy.setTimer("test_runner", run_test, 100) # 100ms delay
```
### Running
```bash
cd build
./mcrogueface --headless --exec ../tests/test_rendering.py
```
**Important:** Timer callbacks are essential! Screenshots only work after rendering starts.
### Example: Testing Click Events
```python
import mcrfpy
from mcrfpy import automation
import sys
# Track clicks
clicks_received = []
def on_frame_click(x, y, button, state):
clicks_received.append((x, y, button, state))
# Setup scene
mcrfpy.createScene("test")
frame = mcrfpy.Frame(100, 100, 200, 150)
frame.fill_color = mcrfpy.Color(0, 255, 0)
frame.click = on_frame_click
mcrfpy.sceneUI("test").append(frame)
mcrfpy.setScene("test")
def run_test(runtime_ms):
# Simulate click on frame center
automation.click(200, 175)
# Give it a frame to process
mcrfpy.setTimer("verify", verify_results, 32)
def verify_results(runtime_ms):
# Check click was received
assert len(clicks_received) > 0, "No clicks received!"
x, y, button, state = clicks_received[0]
assert button == 1, f"Wrong button: {button}"
assert state == True, "Wrong state"
print(f"Click received at ({x}, {y})")
print("PASS")
sys.exit(0)
mcrfpy.setTimer("test", run_test, 100)
```
---
## Automation API
### Screenshots
```python
from mcrfpy import automation
# Take screenshot
automation.screenshot("output.png")
# Screenshots saved to current directory
# Use for visual regression testing
```
### Mouse Input
```python
from mcrfpy import automation
# Click at position
automation.click(x, y, button=1) # button: 1=left, 2=middle, 3=right
# Move mouse
automation.move(x, y)
```
### Keyboard Input
```python
from mcrfpy import automation
# Key press
automation.keypress(key_code, pressed=True)
# Key release
automation.keypress(key_code, pressed=False)
```
**Key codes:** See `mcrfpy.Key.*` constants
---
## Test-Driven Development (TDD)
### TDD Workflow
1. **Write failing test** - Demonstrates the bug/missing feature
2. **Run test** - Verify it fails
3. **Implement fix** - Make minimum change to pass
4. **Run test** - Verify it passes
5. **Refactor** - Clean up if needed
6. **Run test** - Verify still passes
### Example: TDD for New Feature
**Step 1: Write failing test**
`tests/test_sprite_rotation.py`:
```python
import mcrfpy
def test_sprite_rotation():
sprite = mcrfpy.Sprite("test.png", 0, 0)
# This should work but doesn't yet
sprite.rotation = 90
assert sprite.rotation == 90
return True
if __name__ == "__main__":
if test_sprite_rotation():
print("PASS")
else:
print("FAIL")
```
**Step 2: Run test - it fails**
```bash
$ ./mcrogueface --headless --exec tests/test_sprite_rotation.py
AttributeError: 'Sprite' object has no attribute 'rotation'
FAIL
```
**Step 3: Implement feature**
Add `rotation` property to `src/UISprite.cpp` (see [[Adding-Python-Bindings]])
**Step 4: Run test - it passes**
```bash
$ ./mcrogueface --headless --exec tests/test_sprite_rotation.py
PASS
```
---
## Testing Best Practices
### DO:
✅ **Test one thing at a time**
```python
def test_grid_walkable_setter():
grid = mcrfpy.Grid(10, 10, 16, 16)
grid.walkable((5, 5), True)
# Test only walkable setter, nothing else
```
✅ **Use descriptive test names**
```python
# Good
def test_entity_moves_to_valid_position()
# Bad
def test1()
```
✅ **Clean up after tests**
```python
def test_feature():
# Setup
obj = create_test_object()
try:
# Test
obj.do_thing()
finally:
# Cleanup
obj.cleanup()
```
✅ **Test edge cases**
```python
def test_grid_bounds():
grid = mcrfpy.Grid(10, 10, 16, 16)
# Test boundaries
grid.at((0, 0)) # Min
grid.at((9, 9)) # Max
# Test out of bounds
try:
grid.at((10, 10))
assert False, "Should have raised exception"
except ValueError:
pass # Expected
```
### DON'T:
❌ **Test multiple unrelated things**
```python
# Bad - testing grid, entity, AND animation
def test_everything():
grid = mcrfpy.Grid(10, 10, 16, 16)
entity = mcrfpy.Entity(5, 5, 0)
mcrfpy.animate(entity, "x", 10, 1000, "linear")
# Too much!
```
❌ **Forget to exit in timer-based tests**
```python
def run_test(runtime_ms):
automation.screenshot("test.png")
print("PASS")
# MISSING: sys.exit(0) - will hang!
```
❌ **Rely on timing**
```python
# Bad - fragile
time.sleep(0.5) # Hope animation finished
assert entity.x == 100
# Good - check state
while entity.x < 100 and timeout < 1000:
time.sleep(0.016)
timeout += 16
assert entity.x == 100
```
---
## Common Testing Patterns
### Pattern 1: Property Round-Trip
```python
def test_property_roundtrip():
obj = mcrfpy.Frame(0, 0, 100, 100)
# Set various values
test_values = [0, 50, 100, 255, 127]
for value in test_values:
obj.x = value
assert obj.x == value, f"Failed for {value}"
```
### Pattern 2: Visual Regression
```python
def run_test(runtime_ms):
# Render scene
automation.screenshot("current.png")
# Compare with golden image
# (Manual comparison or use image diff tool)
print("Check current.png against golden/expected.png")
sys.exit(0)
```
### Pattern 3: Exception Testing
```python
def test_invalid_input():
grid = mcrfpy.Grid(10, 10, 16, 16)
# Test exception is raised
try:
grid.at((-1, -1)) # Invalid coords
assert False, "Should have raised ValueError"
except ValueError as e:
assert "out of bounds" in str(e).lower()
```
### Pattern 4: State Machine Testing
```python
def test_entity_states():
entity = mcrfpy.Entity(0, 0, 0)
# Initial state
assert entity.state == "idle"
# Transition
entity.attack()
assert entity.state == "attacking"
# Complete action
entity.finish_attack()
assert entity.state == "idle"
```
---
## Debugging Failed Tests
### Get More Information
```python
def test_feature():
try:
obj.do_thing()
except Exception as e:
print(f"Exception: {e}")
print(f"Object state: {obj}")
print(f"Type: {type(obj)}")
raise # Re-raise to fail test
```
### Visual Debugging
```python
def run_test(runtime_ms):
# Take screenshot before action
automation.screenshot("before.png")
# Perform action
do_something()
# Take screenshot after
automation.screenshot("after.png")
# Now you can compare visually
```
### Print Debugging
```python
def test_complex_operation():
print("Starting test...")
obj = create_object()
print(f"Created: {obj}")
result = obj.process()
print(f"Result: {result}")
assert result == expected
```
---
## Related Documentation
- CLAUDE.md - Testing guidelines section
- [[Performance-Optimization-Workflow]] - Creating benchmarks
- `tests/` directory - Example tests
**Testing Pattern Examples:**
- `tests/test_grid_operations.py` - Direct execution
- `tests/issue_78_middle_click_test.py` - Timer-based with automation