Mojo2Wasm

An experimental, self-hosted compiler frontend written entirely in Mojo to compile Mojo source code into WebAssembly (Wasm).

The goal of this project is to bridge the gap between Mojo's high-performance systems programming capabilities and the web/edge ecosystem.

Why Mojo2Wasm?

While official WebAssembly support is still maturing in the core toolchain, mojo2wasm takes a grassroots, native approach. By building the compiler in Mojo itself, this project dogfoods the language's strict memory management, powerful struct system, and performance capabilities to generate optimized Wasm binaries.

How It Works (Architecture)

The compiler bypasses heavy toolchain dependencies by implementing a custom compilation pipeline:

1. Lexing & Parsing: Scans native Mojo source code and converts it into a structured Abstract Syntax Tree (AST) using highly efficient Mojo data structures.
2. Type Checking & Lowering: (In Progress) Maps Mojo's strict type system and object model down to WebAssembly primitives.
3. Code Generation: Translates the valid AST nodes directly into WebAssembly format, utilizing Wasm's linear memory model.

📁 Repository Structure

• /src: Core compiler source code (Lexer, Parser, AST structures, and Code Generator).
• /examples: Sample Mojo files intended to be compiled to Wasm.
• /tests: Test suites verifying AST generation and Wasm output compliance.
• main.mojo: The primary entry point for the CLI tool.

🚧 Current Project Status

This project is a Work in Progress (WIP).

Roadmap:

• Lexer & Tokenizer for basic Mojo syntax
• AST Node definitions for variables, functions, and control flow
• WebAssembly Text Format (.wat) emitter
• WebAssembly Binary Format (.wasm) encoder
• Memory allocator mapping (Mojo pointers to Wasm linear memory)
• Support for Mojo fn and struct definitions

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.