feat: migrate from Yeoman to Plop for modern scaffolding

[prometheas]

Summary

Migrate the project from Yeoman Generator to Plop for a more modern, performant, and maintainable scaffolding solution.

Background

The project currently uses Yeoman Generator (yeoman-generator@5.10.0) which, while functional, has several limitations:

• Complex class-based architecture making testing difficult
• Heavy dependency footprint
• Slower performance due to filesystem ceremony
• Testing challenges (complex mocking, temp directories, slow execution)
• Aging ecosystem with less active development

Proposed Solution: Migrate to Plop

Plop is a modern micro-generator framework that's perfect for our use case:

Benefits

• Modern Architecture: Simple function-based generators vs complex classes
• Better Performance: Lightweight with minimal dependencies
• Improved Testing: Easy to unit test, supports in-memory operations
• TypeScript Support: Built-in TypeScript support with proper types
• Active Maintenance: Modern, well-maintained project
• Simpler API: More intuitive than Yeoman's class hierarchy

Migration Scope

Phase 1: Core Migration

• Add Plop as dependency, remove Yeoman Generator
• Convert main generator (app/index.js) to Plop format (plopfile.js)
• Migrate prompts (should be nearly 1:1)
• Convert file copy/template operations to Plop actions
• Update CLI interface and package.json scripts

Phase 2: Templates & Testing

• Ensure all template stacks work correctly (js-vitest, php-phpspec, php-phpunit)
• Migrate test suite to use node-plop for programmatic testing
• Implement faster, more reliable testing with in-memory operations
• Update documentation and README

Phase 3: Enhancements (Optional)

• Add TypeScript for better developer experience
• Implement additional Handlebars helpers for template flexibility
• Add custom actions for advanced file operations if needed

Technical Comparison

Current Yeoman Structure:

module.exports = class extends Generator {
  async prompting() {
    const prompts = [/* prompts array */];
    const props = await this.prompt(prompts);
    // set properties
  }
  
  writing() {
    this.fs.copyTpl(/* template operations */);
  }
};

Proposed Plop Structure:

export default function (plop) {
  plop.setGenerator('tdd-kata', {
    description: 'Generate TDD kata scaffolds',
    prompts: [/* same prompts array */],
    actions: [
      {
        type: 'add',
        path: 'README.md', 
        templateFile: 'templates/katas/{{kata}}.md'
      },
      // dynamic actions based on stack choice
    ]
  });
}

Testing Improvements

One of the major benefits will be drastically improved testing:

Current (Yeoman):

• Complex yeoman-test setup with temp directories
• Slow filesystem operations
• Difficult to mock and isolate
• Shell execution required for integration tests

Proposed (Plop):

• Simple programmatic testing with node-plop
• In-memory operations possible
• Easy mocking and isolation
• Much faster test execution

Breaking Changes

• CLI interface will change from yo generator-multistack-tdd-kata to plop (or custom CLI name)
• Generated project structure remains identical (no user impact)
• Development/contribution workflow changes

Success Criteria

• All existing functionality preserved
• Generated projects identical to current output
• Test suite runs faster (target: 50%+ improvement)
• Simpler codebase with better maintainability
• Documentation updated
• CLI remains user-friendly

Implementation Notes

• Consider keeping Yeoman temporarily during migration for comparison
• Ensure template compatibility (currently using basic templating, should work with Handlebars)
• May need custom actions for complex file operations currently done in writeProjectFiles()
• Test the migration thoroughly with all stack combinations

Related Issues

This addresses testing performance issues and modernizes the codebase for better maintainability.

---

Priority: Medium (good improvement but not urgent)
Estimated Effort: 1-2 days for core migration + testing