feat(executors): Add Java connector support with automatic JRE management

Summary

This PR adds comprehensive Java connector support to PyAirbyte by implementing a new JavaExecutor class that automatically downloads and manages Zulu JRE installations. The implementation follows the working logic from the Airbyte shell script and integrates seamlessly with PyAirbyte's existing executor framework.

Key Features:

• Automatic JRE download and caching in ~/.airbyte/java/{os}-{arch}/ directories
• Platform detection for Linux/macOS and x64/aarch64 architectures
• Integration with Azul's API for Zulu 21 JRE downloads
• Support for use_java_tar parameter with Path, str, bool, and None types
• Fallback to system JAVA_HOME if available
• Proper error handling and user-friendly messages

API Usage:

# With explicit tar path
source = ab.get_source("source-mssql", use_java_tar="/path/to/connector.tar")

# With auto-location (when implemented)
source = ab.get_source("source-mssql", use_java_tar=True)

Review & Testing Checklist for Human

This is a medium-risk change introducing new platform-specific functionality:

• End-to-end testing: Build a Java connector (e.g., source-mssql) using gradlew and test the complete flow from tar installation to --spec execution
• JRE download verification: Test JRE download on both Linux and macOS to ensure platform detection and Azul API integration work correctly
• Parameter type handling: Verify all use_java_tar parameter types work as expected, especially the bool=True case for auto-location
• Error handling validation: Test edge cases like network failures, invalid tar files, and unsupported platforms to ensure user-friendly error messages
• Registry integration: Verify that Java connectors in the registry properly trigger JavaExecutor usage

Recommended test plan:

1. Build source-mssql connector: cd ~/repos/airbyte && ./gradlew :airbyte-integrations:connectors:source-mssql:build
2. Test with explicit path: ab.get_source("source-mssql", use_java_tar="path/to/airbyte-app.tar").spec()
3. Verify JRE caching works by running twice and confirming no re-download
4. Test on clean environment to validate full JRE download flow

Diagram

%%{ init : { "theme" : "default" }}%%
graph TD
    User["User calls<br/>ab.get_source()"]:::context
    Factory["airbyte/_executors/<br/>util.py"]:::major-edit
    JavaExec["airbyte/_executors/<br/>java.py"]:::major-edit
    Init["airbyte/_executors/<br/>__init__.py"]:::minor-edit
    
    AzulAPI["Azul JRE API"]:::context
    JRECache["~/.airbyte/java/<br/>{os}-{arch}/"]:::context
    ConnectorTar["Connector TAR file"]:::context
    
    User -->|"use_java_tar param"| Factory
    Factory -->|"creates JavaExecutor"| JavaExec
    JavaExec -->|"downloads JRE"| AzulAPI
    JavaExec -->|"caches to"| JRECache
    JavaExec -->|"extracts & runs"| ConnectorTar
    
    subgraph Legend
        L1[Major Edit]:::major-edit
        L2[Minor Edit]:::minor-edit  
        L3[Context/No Edit]:::context
    end
    
    classDef major-edit fill:#90EE90
    classDef minor-edit fill:#87CEEB
    classDef context fill:#FFFFFF

Notes

• Platform support: Currently limited to Linux/macOS with x64/aarch64 architectures. Windows support would require additional implementation.
• JRE version: Hardcoded to Zulu 21 based on Airbyte's current requirements. Future versions may need parameterization.
• Auto-location: The use_java_tar=True case currently sets path to None - full auto-location logic may need additional implementation based on registry metadata.
• Performance: JRE downloads are ~50-100MB and cached permanently. First-time setup will be slower but subsequent runs should be fast.

Session Info:

• Link to Devin run: https://app.devin.ai/sessions/20038c70b8be44b09cd952cef2e2cb1d
• Requested by: @aaronsteers
• Based on shell script logic from: https://github.com/airbytehq/airbyte/pull/63367/files