#MoonMark - High-Performance MoonBit Markdown Parser

#🌟 Key Features & Implementation Guarantees

#1. 🚀 StringView-Backed Parsing Architecture

• Memory View Design: Text payloads in AST nodes (Text, CodeSpan, Link, Image, CodeBlock) use native MoonBit StringView slices when the parser can preserve the source slice.
• Measured, Not Guaranteed: Flat block/inline parsing avoids copying text spans; normalized nested blocks may allocate temporary source buffers. Benchmark results depend on the toolchain and machine.

#2. 📍 Precise Source Mapping (Span Tracking)

• The Position struct tracks 1-based line numbers, 1-based columns, and 0-based offsets in the source string index model used by the parser; columns are not grapheme-cluster counts.
• Every Block and Inline AST node contains exact span : Span metadata (start and end), ideal for IDE syntax highlighting, diagnostic linters, and refactoring tools.

#3. 🛡️ Defensive HTML Rendering

• HTML Entity Escaping: Strict escaping for < (&lt;), > (&gt;), & (&amp;), " (&quot;), and ' (&#39;).
• URL Protocol Sanitization (sanitize_url): Scans Link and Image schemes, rejects javascript:, vbscript:, file:, and unsafe data: payloads, and escapes safe values before placing them in attributes. This is a renderer defense layer, not a replacement for a Content Security Policy.

#4. 📦 Batch Conversion & Powerful CLI

• Out-of-the-Box CLI: Supports single-file conversion, multi-file batch conversion (--batch), and recursive directory batch processing (--dir) while preserving relative subdirectories.
• Auto Directory Creation: Automatically invokes filesystem APIs to create missing target directories.

#Scope and Current Limits

• Supported constructs in this revision include headings, paragraphs, blockquotes, fenced code blocks, ordered and unordered lists, emphasis, strong emphasis, code spans, links, and images.
• Raw HTML is rendered as escaped text. This project does not claim full CommonMark conformance, raw-HTML passthrough, or extension-syntax compatibility.
• URL filtering is a defensive renderer layer. Applications should still validate untrusted input and deploy an appropriate Content Security Policy (CSP).
• --dir processes .md and .markdown files recursively and writes .html files while preserving relative directories; other file extensions are ignored.

#🏗️ Architecture Overview

moonmark/
├── src/ast/       # Core AST definitions using StringView and Span
├── src/parser/    # State-machine Lexer and recursive descent Parser
├── src/html/      # HTML escaper, URL sanitizer, and Renderer
├── src/cli/       # CLI application with batch conversion modes
└── benches/       # Performance benchmarking suite

#🛠️ Installation & Reproduction Steps

#1. Install MoonBit Toolchain

moon version

#2. Clone Repository & Run Checks

git clone https://github.com/zmjknn/zxnMoonbit.git
cd zxnMoonbit

# 1. Type check
moon check --deny-warn

# 2. Run unit tests
moon test

# 3. Benchmark suite
moon bench

# 4. Format check
moon fmt --check

# 5. Regenerate public interface files (review the generated diff)
moon info

# 6. CLI smoke test
moon run src/cli -- --help

# 7. Check patch formatting
git diff --check

#💻 CLI User Guide

#1. Demo Mode

moon run src/cli

#2. Single File Conversion

moon run src/cli -- input.md -o output.html

#3. Multi-File Batch Conversion (--batch / -b)

moon run src/cli -- --batch README.md README_zh.md -o dist/

#4. Directory Batch Conversion (--dir / -d)

moon run src/cli -- --dir docs/ -o dist/
# docs/guide/start.md -> dist/guide/start.html

#5. Help Menu

moon run src/cli -- --help

#Validation Snapshot (this revision)

#📚 Prior Art & References

1. marked.js (MIT License)
   • Reference Scope: AST node taxonomy (Block/Inline division), HTML escaping rules, and CLI flag patterns.
2. pulldown-cmark (MIT License)
   • Reference Scope: Zero-allocation state-machine lexing model using string slices, event-driven parsing flow, and CommonMark compliance testing strategies.

#📄 License