123456789_123456789_123456789_123456789_123456789_

RDoc Project Guide for AI Agents

Project Overview

RDoc is Ruby's default documentation generation tool that produces HTML and command-line documentation for Ruby projects. It parses Ruby source code, C extensions, and markup files to generate documentation.

Key Development Commands

Testing

# Run all tests (default task)
bundle exec rake

# Run unit tests only (excludes RubyGems integration)
bundle exec rake normal_test

# Run RubyGems integration tests only
bundle exec rake rubygems_test

# Verify generated parser files are current (CI check)
bundle exec rake verify_generated

Test Framework: Test::Unit with test-unit gem Test Location: test/ directory Test Helper: test/lib/helper.rb

Linting

RuboCop (Ruby Linting)

# Check Ruby code style
bundle exec rubocop

# Auto-fix style issues
bundle exec rubocop -A

Configuration: .rubocop.yml

Herb Linter (ERB/RHTML Files)

# Lint ERB template files
npx @herb-tools/linter "**/*.rhtml"

# Lint specific directory
npx @herb-tools/linter "lib/**/*.rhtml"

Template Location: lib/rdoc/generator/template/**/*.rhtml CI Workflow: .github/workflows/lint.yml

Stylelint (CSS Files)

# Lint CSS files
npm run lint:css

# Auto-fix style issues
npm run lint:css -- --fix

# Lint specific file
npx stylelint "lib/rdoc/generator/template/aliki/css/rdoc.css"

Configuration: .stylelintrc.json Features:

Type annotations

Annotate method types using Sorbet flavored RBS in inline comments. For more information about RBS syntax, see the [documentation]).

A few examples:

# Method that receives an integer and doesn't return anything
#: (Integer) -> void
def foo(something); end

# Method that receives a string and returns an integer
#: (String) -> Integer
def bar(something)
   123
end

# Method that doesn't accept arguments and returns a hash of symbol to string
#: () -> Hash[Symbol, String]
def bar
   { key: "value" }
end

# Method that accepts a block, which yields a single integer argument and returns whatever the block returns
#: [T] () { (Integer) -> T } -> T
def bar
   yield(5)
end

Documentation Generation

# Generate documentation (creates _site directory)
bundle exec rake rdoc

# Force regenerate documentation
bundle exec rake rerdoc

# Show documentation coverage
bundle exec rake rdoc:coverage
bundle exec rake coverage

Output Directory: _site/ (GitHub Pages compatible) Configuration: .rdoc_options

Parser Generation

RDoc uses generated parsers for Markdown and RD formats:

# Generate all parser files from sources
bundle exec rake generate

# Remove generated parser files
bundle exec rake clean

Generated Files:

Note: These files are auto-generated and should not be edited manually. Always regenerate after modifying source .ry or .kpeg files.

Building and Releasing

# Build gem package
bundle exec rake build

# Install gem locally
bundle exec rake install

# Create tag and push to rubygems.org
bundle exec rake release

Project Structure

lib/rdoc/
├── rdoc.rb                    # Main entry point (RDoc::RDoc class)
├── version.rb                 # Version constant
├── task.rb                    # Rake task integration
├── parser/                    # Source code parsers (Ruby, C, Markdown, RD)
│   ├── ruby.rb                # Ruby code parser
│   ├── c.rb                   # C extension parser
│   ├── prism_ruby.rb          # Prism-based Ruby parser
│   └── ...
├── generator/                 # Documentation generators
│   ├── aliki.rb               # HTML generator (default theme)
│   ├── darkfish.rb            # HTML generator (deprecated, will be removed in v8.0)
│   ├── markup.rb              # Markup format generator
│   ├── ri.rb                  # RI command generator
│   └── template/              # ERB templates (.rhtml files)
│       ├── aliki/             # Aliki theme (default)
│       └── darkfish/          # Darkfish theme (deprecated)
├── markup/                    # Markup parsing and formatting
├── code_object/               # AST objects for documented items
├── markdown/                  # Markdown parsing
├── rd/                        # RD format parsing
└── ri/                        # RI (Ruby Info) tool

test/                          # 79 test files
├── lib/helper.rb              # Test helpers
└── rdoc/                      # Main test directory

exe/
├── rdoc                       # rdoc command executable
└── ri                         # ri command executable

Important Files

Configuration

CI/CD

Documentation

Architecture Notes

Pluggable System

Common Workflows

Do NOT commit anything. Ask the developer to review the changes after tasks are finished.

NEVER pushes code to any repositories.

Making Code Changes

Use Red, Green, Refactor approach:

  1. Ensure Ruby version: Verify you're using Ruby 3.3.0+ (prepend chruby <ruby version> if needed)
  2. Red - Write failing tests: Add tests that fail for the new behavior
  3. Verify failure: Run bundle exec rake to confirm tests fail as expected
  4. Green - Make it work: Implement the minimum code to make tests pass
  5. Refactor - Make it right: Improve code quality while keeping tests green
    • Run bundle exec rake after each refactor to ensure tests still pass
    • Iterate on steps 4-5 as needed
  6. Lint your changes:
    • Ruby code: bundle exec rubocop -A (auto-fix when possible)
    • ERB templates: npx @herb-tools/linter "**/*.rhtml" (if modified)
    • CSS files: npm run lint:css -- --fix (if modified)

Modifying Parsers

  1. Edit source files (.ry or .kpeg)
  2. Regenerate: bundle exec rake generate
  3. Verify: bundle exec rake verify_generated
  4. Run tests: bundle exec rake

Updating Documentation

  1. Modify documentation comments in source
  2. Regenerate: bundle exec rake rerdoc
  3. Check output in _site/ directory
  4. Check coverage: bundle exec rake coverage

Modifying Markup Reference Documentation

When editing markup reference documentation, such as doc/markup_reference/markdown.md and doc/markup_reference/rdoc.rdoc:

  1. Always verify rendering - After making changes, test that the content renders correctly using Ruby:

For Markdown files:

ruby -r rdoc -r rdoc/markdown -e '
md = RDoc::Markdown.new
doc = md.parse("YOUR CONTENT HERE")
formatter = RDoc::Markup::ToHtml.new(RDoc::Options.new)
puts formatter.convert(doc)
'

For RDoc files:

ruby -r rdoc -e '
parser = RDoc::Markup::Parser.new
doc = parser.parse("YOUR CONTENT HERE")
formatter = RDoc::Markup::ToHtml.new(RDoc::Options.new)
puts formatter.convert(doc)
'
  1. Watch for rendering issues:

    • Backtick escaping (especially nested code blocks)
    • Tilde characters being interpreted as strikethrough
    • Special characters in examples
    • Anchor links pointing to correct headings
  2. Known RDoc Markdown limitations:

    • Only triple backticks for fenced code blocks (no tildes, no quad-backticks)
    • Tilde fences (~~~) conflict with strikethrough syntax
    • Use 4-space indentation to show literal code fence examples
  3. Full verification: Generate documentation and inspect the HTML output:

   bundle exec rake rerdoc
   # Inspect the generated HTML file directly
   grep -A5 "your content" _site/path/to/file.html

Modifying Themes/Styling

When making changes to theme CSS or templates (e.g., Darkfish or Aliki themes):

  1. Generate documentation: Run bundle exec rake rerdoc to create baseline
  2. Start HTTP server: Run cd _site && python3 -m http.server 8000 (use different port if 8000 is in use)
  3. Investigate with Playwright: Ask the AI assistant to take screenshots and inspect the documentation visually
    • Example: "Navigate to the docs at localhost:8000 and screenshot the RDoc module page"
    • See "Playwright MCP for Testing Generated Documentation" section below for details
  4. Make changes: Edit files in lib/rdoc/generator/template/<theme>/ as needed
  5. Regenerate: Run bundle exec rake rerdoc to rebuild documentation with changes
  6. Verify with Playwright: Take new screenshots and compare to original issues
  7. Lint changes (if modified):
    • ERB templates: npx @herb-tools/linter "lib/rdoc/generator/template/**/*.rhtml"
    • CSS files: npm run lint:css -- --fix
  8. Stop server: Kill the HTTP server process when done

Tip: Keep HTTP server running during iteration. Just regenerate with bundle exec rake rerdoc between changes.

Notes for AI Agents

  1. Always run tests after making changes: bundle exec rake
  2. Lint your changes:
    • RuboCop for Ruby: bundle exec rubocop -A
    • Herb for ERB templates: npx @herb-tools/linter "**/*.rhtml"
    • Stylelint for CSS: npm run lint:css -- --fix
  3. Regenerate parsers if you modify .ry or .kpeg files
  4. Use rake rerdoc to regenerate documentation (not just rdoc)
  5. Verify generated files with rake verify_generated
  6. Don't edit generated files directly (in lib/rdoc/markdown/ and lib/rdoc/rd/)

Playwright MCP for Testing Generated Documentation

The Playwright MCP server enables visual inspection and interaction with generated HTML documentation. This is useful for verifying CSS styling, layout issues, and overall appearance.

MCP Server: @playwright/mcp (Microsoft's official browser automation server)

Setup

The Playwright MCP server can be used with any MCP-compatible AI tool (Claude Code, Cursor, GitHub Copilot, OpenAI Agents, etc.).

Claude Code:

/plugin playwright

Other MCP-compatible tools:

npx @playwright/mcp@latest

Configure your tool to connect to this MCP server. Playwright launches its own browser instance automatically - no manual browser setup or extensions required.

Troubleshooting: Chrome Remote Debugging Blocked

If you encounter DevTools remote debugging is disallowed by the system admin, Chrome's debugging is blocked by the machine's policy. Use Firefox instead:

# Install Firefox for Playwright
npx playwright install firefox

# Add Playwright MCP with Firefox to your project (creates/updates .mcp.json)
claude mcp add playwright --scope project -- npx -y @playwright/mcp@latest --browser firefox

Restart Claude Code after running these commands.

Testing Generated Documentation

To test the generated documentation:

# Generate documentation
bundle exec rake rerdoc

# Start a simple HTTP server in the _site directory (use an available port)
cd _site && python3 -m http.server 8000

If port 8000 is already in use, try another port (e.g., python3 -m http.server 9000).

Then ask the AI assistant to inspect the documentation. It will use the appropriate Playwright tools (browser_navigate, browser_snapshot, browser_take_screenshot, etc.) based on your request.

Example requests: