McRogueFace/tests/KNOWN_ISSUES.md
Frick 4528ece0a7 Refactor timing tests to use mcrfpy.step() for synchronous execution
Converts tests from Timer-based async patterns to step()-based sync
patterns, eliminating timeout issues in headless testing.

Refactored tests:
- simple_timer_screenshot_test.py
- test_animation_callback_simple.py
- test_animation_property_locking.py
- test_animation_raii.py
- test_animation_removal.py
- test_timer_callback.py

Also updates KNOWN_ISSUES.md with comprehensive documentation on
the step()-based testing pattern including examples and best practices.

🤖 Generated with Claude Code (https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
2026-01-14 02:56:21 +00:00

3.4 KiB

McRogueFace Test Suite - Known Issues

Test Results Summary

As of 2026-01-14, with --mcrf-timeout=5:

  • 120 passed (67%)
  • 59 failed (33%)
    • 40 timeout failures (tests requiring timers/animations)
    • 19 actual failures (API changes, missing features, or bugs)

Synchronous Testing with mcrfpy.step()

RECOMMENDED: Use mcrfpy.step(t) to advance simulation time synchronously instead of relying on Timer callbacks and the game loop. This eliminates timeout issues and makes tests deterministic.

Old Pattern (Timer-based, async)

# OLD: Requires game loop, subject to timeouts
def run_tests(timer, runtime):
    # tests here
    sys.exit(0)

mcrfpy.Timer("run", run_tests, 100, once=True)
# Script ends, game loop runs, timer eventually fires

New Pattern (step-based, sync)

# NEW: Synchronous, no timeouts
import mcrfpy
import sys

# Setup scene
scene = mcrfpy.Scene("test")
scene.activate()
mcrfpy.step(0.1)  # Initialize scene

# Run tests directly
frame = mcrfpy.Frame(pos=(100, 100), size=(100, 100))
scene.children.append(frame)

# Start animation
anim = mcrfpy.Animation("x", 500.0, 1.0, "linear")
anim.start(frame)

# Advance simulation to complete animation
mcrfpy.step(1.5)  # Advances 1.5 seconds synchronously

# Verify results
if frame.x == 500.0:
    print("PASS")
    sys.exit(0)
else:
    print("FAIL")
    sys.exit(1)

Key Differences

Aspect Timer-based step()-based
Execution Async (game loop) Sync (immediate)
Timeout risk High None
Determinism Variable Consistent
Timer firing Once per step() call Per elapsed interval

Timer Behavior with step()

  • Timers fire once per step() call if their interval has elapsed
  • To fire a timer multiple times, call step() multiple times:
# Timer fires every 100ms
timer = mcrfpy.Timer("tick", callback, 100)

# This fires the timer ~6 times
for i in range(6):
    mcrfpy.step(0.1)  # Each step processes timers once

Refactored Tests

The following tests have been converted to use mcrfpy.step():

  • simple_timer_screenshot_test.py
  • test_animation_callback_simple.py
  • test_animation_property_locking.py
  • test_animation_raii.py
  • test_animation_removal.py
  • test_timer_callback.py
  • test_timer_once.py

Remaining Timeout Failures

These tests still use Timer-based async patterns:

  • WORKING_automation_test_example.py
  • benchmark_logging_test.py
  • keypress_scene_validation_test.py
  • test_empty_animation_manager.py
  • test_simple_callback.py

Headless mode tests:

  • test_headless_detection.py
  • test_headless_modes.py

Other timing-dependent:

  • test_color_helpers.py
  • test_frame_clipping.py
  • test_frame_clipping_advanced.py
  • test_grid_children.py
  • test_no_arg_constructors.py
  • test_properties_quick.py
  • test_python_object_cache.py
  • test_simple_drawable.py

Running Tests

# Quick run (5s timeout, many timeouts expected)
pytest tests/ -q --mcrf-timeout=5

# Full run (30s timeout, should pass most timing tests)
pytest tests/ -q --mcrf-timeout=30

# Filter by pattern
pytest tests/ -k "bsp" -q

# Run original runner
python3 tests/run_tests.py -q

Recommendations

  1. For CI: Use --mcrf-timeout=10 and accept ~30% timeout failures
  2. For local dev: Use --mcrf-timeout=30 for comprehensive testing
  3. For quick validation: Use -k "not animation and not timer" to skip slow tests