Skip to content

README.md

Latest commit

 

History

History
192 lines (149 loc) · 6.5 KB

README.md

File metadata and controls

192 lines (149 loc) · 6.5 KB

🐙 GitHub Integration

This chapter provides technical guidance for integrating mainframe modernization workflows with GitHub using Azure AI Foundry, supporting multiple mainframe platforms (IBM z/OS, Unisys ClearPath, Bull GCOS, NEC ACOS).

📋 Overview

GitHub integration enables modern DevOps practices for mainframe application development, providing version control, collaboration tools, automated workflows, and continuous integration capabilities. Azure AI Foundry enhances this integration with AI-powered capabilities specifically designed for mainframe code across different platforms.

🎯 Objectives

Objective Description
Repository Setup Set up GitHub repositories for mainframe code management
Git Configuration Configure appropriate Git settings for mainframe code files
Workflows Implement GitHub Actions workflows for mainframe CI/CD
Collaboration Enable collaborative development practices for mainframe teams
AI Integration Integrate AI-powered code analysis and transformation with GitHub
Multi-Platform Support Support various mainframe platforms with specific configurations

📦 GitHub Repository Setup

A well-structured repository helps organize mainframe code effectively:

mainframe-app/
├── src/
│   ├── cobol/                  # COBOL source code
│   ├── pl1/                    # PL/I source code
│   ├── natural/                # Natural source code
│   ├── jcl/                    # JCL procedures
│   ├── copybooks/              # Copybooks
│   └── platform-specific/      # Platform-specific code
│       ├── zos/                # IBM z/OS specific code
│       ├── clearpath/          # Unisys ClearPath specific code
│       ├── gcos/               # Bull GCOS specific code
│       └── acos/               # NEC ACOS specific code
├── tests/                      # Test resources
├── docs/                       # Documentation
├── .github/workflows/          # GitHub Actions workflow definitions
├── .gitattributes              # Git attributes for mainframe files
└── README.md                   # Repository documentation

Configure Git for mainframe files by creating a .gitattributes file with platform-specific settings:

# Default handling of line endings
* text=auto

# IBM z/OS specific files
*.cbl text eol=lf working-tree-encoding=ibm1047 zos-working-tree-encoding=ibm1047
*.cpy text eol=lf working-tree-encoding=ibm1047 zos-working-tree-encoding=ibm1047
*.jcl text eol=lf working-tree-encoding=ibm1047 zos-working-tree-encoding=ibm1047

# Unisys ClearPath specific files
*.cob text eol=lf
*.wrk text eol=lf

# Bull GCOS specific files
*.gcos text eol=lf

# NEC ACOS specific files
*.acos text eol=lf

🔄 GitHub Actions Workflows

Create a basic mainframe CI workflow with platform-specific jobs (.github/workflows/mainframe-ci.yml):

name: Mainframe CI

on:
  push:
    branches: [ main, develop ]
  pull_request:
    branches: [ main, develop ]

jobs:
  analyze:
    runs-on: ubuntu-latest
    strategy:
      matrix:
        platform: [zos, clearpath, gcos, acos]
    
    steps:
    - uses: actions/checkout@v3
    
    - name: AI-Powered Code Analysis
      uses: azure/ai-foundry-code-analysis@v1
      with:
        source-dir: ./src
        platform: ${{ matrix.platform }}
        language: auto-detect
        output-dir: ./analysis-results/${{ matrix.platform }}
  
  build:
    needs: analyze
    runs-on: ubuntu-latest
    
    steps:
    - uses: actions/checkout@v3
    
    - name: Download Analysis Results
      uses: actions/download-artifact@v3
      with:
        name: analysis-results
        path: ./analysis-results
        
    - name: Compile Programs
      run: |
        # Platform-specific compilation commands
        if [ -d "src/cobol" ]; then
          echo "Compiling COBOL programs..."
        for file in src/cobol/*.cbl; do
          cobol-compiler "$file" -I src/copybooks -o "$(basename "$file" .cbl).so"
        done
        fi
        
        if [ -d "src/pl1" ]; then
          echo "Compiling PL/I programs..."
          for file in src/pl1/*.pli; do
            pli-compiler "$file" -o "$(basename "$file" .pli).so"
          done
        fi
        
    - name: Run Tests
      run: |
        mainframe-test-runner ./tests

🔍 Platform-Specific Considerations

IBM z/OS Integration

For IBM z/OS integration, consider:

  • EBCDIC to ASCII conversions
  • Dataset naming conventions
  • JCL handling
  • COBOL copybook management

Unisys ClearPath Integration

For Unisys ClearPath integration, focus on:

  • MCP or OS 2200 specific configurations
  • COBOL and Algol source handling
  • WFL workflow definitions
  • DMSII database integration

Bull GCOS Integration

For Bull GCOS integration, address:

  • JCL equivalents
  • TP8 considerations
  • COBOL dialect differences
  • Database access methods

NEC ACOS Integration

For NEC ACOS integration, manage:

  • ACOS-specific language features
  • Transaction processing integration
  • Database access patterns
  • Special character handling

✅ Best Practices

Practice Description
Incremental Migration Start with small, well-bounded components
Consistent Environments Use development containers for consistent experiences
Comprehensive Testing Implement automated testing for mainframe code
Security First Protect mainframe credentials with secure practices
Knowledge Sharing Document mainframe-specific workflows and considerations
Platform-Specific Branching Consider separate branches for platform-specific code
Multi-Platform Testing Test across platforms for cross-platform components

➡️ Next Steps

After setting up GitHub integration:

  1. Implement 🔄 Azure DevOps Integration for project management
  2. Set up 🧠 AI-Powered Transformation workflows
  3. Establish comprehensive 📦 CI/CD Implementation

📚 References

Resource Description
Azure AI Foundry Documentation Official documentation for Azure AI Foundry
GitHub Actions Documentation Documentation for GitHub Actions
Mainframe DevOps Best Practices Best practices for mainframe DevOps