autusm/test_extraction_consistency.py

#!/usr/bin/env python3
"""
Test script to verify extraction consistency between autusm and USM acquire script.

This script tests that both autusm's initial extraction and the acquire script
produce the same directory structure (flattened to the output directory).
"""

import os
import sys
import tempfile
import shutil
from pathlib import Path

# Add src to path to import autusm modules
sys.path.insert(0, str(Path(__file__).parent / "src"))

from autusm.extractor import ArchiveExtractor
from autusm.generator import ScriptGenerator
from autusm.models import PackageInfo, BuildSystem, BuildSystemType


def test_extraction_consistency():
    """Test that extraction behavior is consistent between autusm and acquire script."""
    
    # Test URL for hello-2.12.tar.gz
    test_url = "https://mirror.freedif.org/GNU/hello/hello-2.12.tar.gz"
    package_name = "hello"
    
    print("Testing extraction consistency...")
    print(f"Test URL: {test_url}")
    print(f"Package: {package_name}")
    print()
    
    # Create test directories
    with tempfile.TemporaryDirectory() as temp_dir:
        temp_path = Path(temp_dir)
        
        # Create output directories for both tests
        autusm_output = temp_path / "autusm_output"
        usm_output = temp_path / "usm_output"
        autusm_output.mkdir()
        usm_output.mkdir()
        
        # Download the test archive
        print("Downloading test archive...")
        from autusm.download import DownloadManager
        download_manager = DownloadManager()
        archive_path = download_manager.download(test_url, temp_path)
        print(f"Downloaded to: {archive_path}")
        print()
        
        # Test 1: autusm extraction
        print("Test 1: autusm extraction")
        extractor = ArchiveExtractor()
        autusm_source_dir = extractor.extract(archive_path, autusm_output)
        print(f"autusm extracted to: {autusm_source_dir}")
        
        # List the structure
        print("autusm directory structure:")
        for item in autusm_output.rglob("*"):
            if item.is_file():
                rel_path = item.relative_to(autusm_output)
                print(f"  {rel_path}")
        print()
        
        # Test 2: Simulate USM acquire script extraction
        print("Test 2: Simulating USM acquire script extraction")
        
        # Create a test package info
        package_info = PackageInfo(
            name=package_name,
            version="2.12",
            summary="Test package",
            url=test_url
        )
        
        # Create a test build system
        build_system = BuildSystem(
            type=BuildSystemType.AUTOTOOLS,
            config_files=[],
            build_files=[],
            detected_commands=[],
            custom_args={}
        )
        
        # Generate the acquire script
        script_generator = ScriptGenerator()
        acquire_script = script_generator._common_acquire_template(package_info, build_system)
        
        # Write the acquire script to a file
        acquire_script_path = usm_output / "acquire"
        with open(acquire_script_path, "w") as f:
            f.write(acquire_script)
        os.chmod(acquire_script_path, 0o755)
        
        # Run the acquire script
        import subprocess
        result = subprocess.run(
            ["./acquire"],
            cwd=usm_output,
            capture_output=True,
            text=True
        )
        
        if result.returncode != 0:
            print(f"Error running acquire script: {result.stderr}")
            return False
        
        print("USM acquire script executed successfully")
        
        # List the structure
        print("USM directory structure:")
        for item in usm_output.rglob("*"):
            if item.is_file() and item.name != "acquire":
                rel_path = item.relative_to(usm_output)
                print(f"  {rel_path}")
        print()
        
        # Compare the directory structures
        print("Comparing directory structures...")
        
        # Get file lists (excluding the acquire script)
        autusm_files = set()
        for item in autusm_output.rglob("*"):
            if item.is_file():
                rel_path = item.relative_to(autusm_output)
                autusm_files.add(str(rel_path))
        
        usm_files = set()
        for item in usm_output.rglob("*"):
            if item.is_file() and item.name != "acquire":
                rel_path = item.relative_to(usm_output)
                usm_files.add(str(rel_path))
        
        # Check if they match
        if autusm_files == usm_files:
            print("✓ SUCCESS: Directory structures match!")
            print(f"Both contain {len(autusm_files)} files")
            return True
        else:
            print("✗ FAILURE: Directory structures differ!")
            
            print("Files only in autusm:")
            for f in sorted(autusm_files - usm_files):
                print(f"  {f}")
            
            print("Files only in USM:")
            for f in sorted(usm_files - autusm_files):
                print(f"  {f}")
            
            return False


if __name__ == "__main__":
    success = test_extraction_consistency()
    sys.exit(0 if success else 1)