Files
2125_GCE/integration_tests/README.md
T
2026-07-09 12:54:44 -04:00

131 lines
3.4 KiB
Markdown
Executable File

# GlyphRunner Integration Test Suite
Comprehensive tests for the GlyphRunner pipeline: compilation, execution, inspection, and summarization.
## Tests
### test_compile.py
**Coverage**: Source → .gx compilation pipeline
- Basic compile with explicit output path
- Auto output naming (derive .gx from .py)
- Verify magic bytes (XIC header)
- File creation and size validation
**Run**: `python3 test_compile.py`
### test_run.py
**Coverage**: Execution and code verification
- Basic .gx execution
- Code execution verification (variables, output capture)
- Segment counting and timing
**Run**: `python3 test_run.py`
### test_inspect.py
**Coverage**: .gx file inspection
- Manifest section display
- Segments section display
- Payload information
- All required manifest fields (version, origin, source_file, etc.)
**Run**: `python3 test_inspect.py`
### test_summary.py
**Coverage**: Human-readable summary generation
- Summary basic output
- All expected fields (GX File, Source, Type, Segments, Compressed, Version)
- Human-readable formatting
**Run**: `python3 test_summary.py`
### test_errors.py
**Coverage**: Error handling and graceful failure
- Compile with missing source file
- Run with missing .gx file
- Inspect with missing .gx file
- Summary with missing .gx file
- No command (help text)
All error cases return exit code 1 and print helpful error messages.
**Run**: `python3 test_errors.py`
### test_determinism.py
**Coverage**: Deterministic output and reproducibility
- Manifest structure consistency across runs
- File size determinism
- Magic bytes (XIC header) consistency
- Compressed payload determinism
**Note**: Full byte-for-byte determinism is not guaranteed due to timestamp fields in manifests. This test verifies structural and payload determinism instead.
**Run**: `python3 test_determinism.py`
## Running All Tests
**Complete test suite**:
```bash
python3 run_all_tests.py
```
This runs all test suites in sequence and prints a summary.
## Test Design
- **Plain Python**: No pytest/unittest frameworks, just subprocess and assertions
- **Exit codes**: Tests exit 0 on all-pass, non-zero on any failure
- **Clear output**: Each test prints PASS/FAIL per assertion
- **Isolated**: Tests use /tmp for output, don't interfere with each other
- **Independent**: Each test can be run individually
## Sample Coverage
All tests use `/home/dave/sample_code.py` as the test source:
- Real Python code (functions, classes, control flow)
- ~600 bytes source
- Compiles to ~960 byte .gx file
- 6 segments identified
- ~280 bytes compressed
## Expected Results
All 6 test suites should pass with ~25 individual test cases across:
- 2 compile tests
- 2 run tests
- 3 inspect tests
- 3 summary tests
- 5 error handling tests
- 3 determinism tests
## Pipeline Verification
The test suite exercises the full pipeline:
```
sample_code.py (source)
gx_cli compile (via commands.cmd_compile)
gx_compiler.compiler.GXCompiler
sample_code.gx (compiled)
gx_cli run/inspect/summary
runtime_executor (gx_loader, execution, integration)
Output + Execution Results
```
## Constraints Satisfied
✓ Tests in superdave/integration_tests/
✓ Plain Python scripts (no frameworks)
✓ Runnable as `python3 test_xxx.py`
✓ Exit non-zero on failure
✓ Clear PASS/FAIL messages
✓ Only touch /home/dave/superdave and /tmp
✓ Happy path with real sample_code.py
✓ Error handling (missing files, bad paths)
✓ Determinism verification