webgl2 0.1.1

WebGL2 shader compiler, emulator, and debugger
Documentation

WebGL2 Development Platform: VISION

A Rust + WASM based toolkit for debugging GLSL shaders and generating ergonomic WebGL2 bindings.

🎯 Quick Start

# Build the project
cargo build --release

# Compile a GLSL shader to WASM with debug info
cargo run --bin webgl2 -- compile tests/fixtures/simple.vert --debug

# Validate a shader
cargo run --bin webgl2 -- validate tests/fixtures/simple.frag

# Generate TypeScript harness
cargo run --bin webgl2 -- codegen tests/fixtures/simple.vert -o output.ts

Build via npm

This repository is both a Rust workspace and an npm package. Run the npm build helper to build the Rust workspace for the WASM target and copy produced .wasm files into runners/wasm/:

# from repository root
npm run build

Notes:

  • The script runs cargo build --target wasm32-unknown-unknown --release and copies any .wasm files from target/wasm32-unknown-unknown/release/ into runners/wasm/.
  • If you need wasm-bindgen output (JS glue), run wasm-bindgen manually on the produced .wasm files; adding automated wasm-bindgen support is a follow-up task.

πŸš€ Project Overview and Goals

The project aims to create a Composite WebGL2 Development Platform built with Rust and WASM. The primary objective is to significantly improve the developer experience by introducing standard software engineering practicesβ€”specifically, robust debugging and streamlined resource managementβ€”into the WebGL/GLSL workflow, which is currently hindered by platform-specific API complexity and opaque GPU execution.

Key Goals Target Block Value Proposition
GPU Debugging Block 1 (Emulator) Enable step-through debugging, breakpoints, and variable inspection for GLSL code.
Unit Testing Block 1 (Emulator) Provide a stable, deterministic environment for automated testing of graphics logic.
API Ergonomics Block 2 (Codegen) Automate boilerplate code for resource binding, attribute setup, and uniform linking.
Tech Stack Both Utilize Rust for safety and WASM for high-performance cross-platform execution in the browser.

πŸ› οΈ Functional Block 1: WebGL2 Software Rendering Pipeline (The Debugger)

This block provides the deterministic, inspectable execution environment for GLSL logic.

1. Core Component: Rust-based WebGL2 Emulator (wasm-gl-emu)

  • State Machine Emulation: Implement a full software model of the WebGL2 state (e.g., Framebuffer Objects, Renderbuffers, Texture Units, Vertex Array Objects, current programs, depth/stencil settings). This component will track all API calls and maintain a CPU-accessible copy of all state.
  • Rasterization Logic: Implement CPU-based vertex and fragment processing. This logic must precisely follow the WebGL2/OpenGL ES 3.0 specification, including clipping, primitive assembly, and the fragment pipeline (culling, depth test, blending).
  • Input/Output: Expose the emulator's state and rendering results via a standard Rust interface that can be compiled to WASM and wrapped for JavaScript access.

2. GLSL Translation and Debugging Integration (glsl-to-wasm)

  • GLSL Frontend: Use a Rust-based GLSL parser (e.g., based on naga or a custom parser) to convert shader source code into an Intermediate Representation (IR).
  • WASM Backend: Compile the IR into WASM module functions. Each shader stage (Vertex, Fragment) will be converted into a function that executes the shader logic on a single vertex or fragment input.
  • Source Map Generation: The crucial step is to generate high-quality source maps that link the generated WASM instructions back to the original GLSL source code line and variable names. This enables DevTools/IDE to pause WASM execution and display the corresponding GLSL line and variable values.
  • JIT Compilation: For performance during debugging, the WASM modules can be compiled using a fast JIT compiler (potentially leveraging existing browser capabilities or a custom runtime) to execute the many vertex/fragment calls.

3. Debugging and Testing Harness

  • Test Runner: Develop a testing harness in Rust/WASM that can execute a defined set of inputs (e.g., vertex attributes, uniform values) against the emulated pipeline and assert on the final pixel output or intermediate variable state.
  • Step-Through Integration: The WASM modules, when run by the browser's JavaScript engine, will expose the generated source maps, allowing the developer to naturally set breakpoints within the GLSL source file viewed in the browser's DevTools (or a connected IDE).

βš™οΈ Functional Block 2: Introspection and Codegen Tool (The Ergonomics Layer)

This block automates the error-prone, repetitive JavaScript/TypeScript code required for WebGL2 resource setup.

1. GLSL Introspection Engine (glsl-parser-rs)

  • Parsing: Use a dedicated Rust parser to analyze the GLSL source files (both Vertex and Fragment shaders).
  • Annotation Extraction: The parser must identify and extract discardable annotations embedded in the GLSL source.
    • Example Annotation:
      //! @buffer_layout MeshData
      layout(location = 0) in vec3 a_position;
      // JSDoc-like for resource description
      /** @uniform_group Camera @semantic ProjectionMatrix */
      uniform mat4 u_projection;
      
  • Resource Mapping: Generate a structured data model (e.g., JSON or Rust structs) that lists all attributes, uniforms, uniform blocks, and their associated types, locations, and extracted metadata (like semantic hints from the annotations).

2. Code Generation Module (js-harness-codegen)

  • Harness Template: Define a set of customizable templates (e.g., Handlebars, Tera) for generating the target JavaScript/TypeScript harness code.
  • Code Generation: Using the resource map data from the parser, the module will generate files that:
    • Define JavaScript/TypeScript classes for each shader program.
    • Provide methods for binding resource objects (e.g., program.setUniforms(cameraData)).
    • Implement all the necessary gl.bindBuffer(), gl.getUniformLocation(), gl.vertexAttribPointer(), and gl.uniformX() calls.
    • Goal: Reduce the developer's WebGL setup code to simple, type-safe resource assignments.

3. Tool Integration

  • Package the Rust component as a Command-Line Interface (CLI) tool (or an accompanying library) that runs during the build step, consuming GLSL files and outputting the JavaScript/TypeScript harness code. This integrates smoothly into modern build pipelines (e.g., Webpack, Vite).

⏳ Project Phases and Deliverables

Phase Duration Focus Key Deliverables
Phase 1 3 Months Core Compiler & Codegen Functional GLSL parser (Block 2). CLI tool for JS harness generation (Block 2). Prototype glsl-to-wasm compiler (Block 1).
Phase 2 4 Months Core Emulator Implementation Basic WebGL2 State Machine and Triangle Rasterizer (Block 1). Source Map integration (GLSL <-> WASM <-> DevTools) (Block 1).
Phase 3 3 Months Feature Completion & Polishing Full WebGL2 feature set support in emulator (textures, complex blending, stencil). Robustness testing and documentation.
Phase 4 2 Months Integration & Release Unit testing framework integration. Final documentation and developer tutorials. Full platform release.

οΏ½ Project Structure

webgl2/
β”œβ”€β”€ crates/
β”‚   β”œβ”€β”€ naga-wasm-backend/    # Naga IR β†’ WASM compiler with DWARF
β”‚   β”œβ”€β”€ wasm-gl-emu/          # Software rasterizer & WASM runtime
β”‚   β”œβ”€β”€ glsl-introspection/   # GLSL parser + annotation extraction
β”‚   β”œβ”€β”€ js-codegen/           # TypeScript harness generator
β”‚   └── webgl2-cli/           # Command-line interface
β”œβ”€β”€ tests/fixtures/           # Test shaders
β”œβ”€β”€ docs/                     # Detailed documentation
β”‚   β”œβ”€β”€ 1-plan.md            # Original project plan
β”‚   └── 1.1-ir-wasm.md       # Naga IR β†’ WASM architecture
└── external/                 # Reference implementations (naga, wgpu, servo)

πŸ—οΈ Architecture

This project uses Naga (from the wgpu/WebGPU ecosystem) as the shader IR, rather than building a custom IR from scratch. This significantly reduces complexity while providing a proven, well-maintained foundation.

Key Components

  1. naga-wasm-backend: Translates Naga IR to WebAssembly with DWARF debug information
  2. wasm-gl-emu: Executes WASM shaders in a software rasterizer for debugging
  3. glsl-introspection: Parses GLSL and extracts resource metadata
  4. js-codegen: Generates ergonomic TypeScript bindings
  5. webgl2-cli: Unified command-line tool

See docs/1.1-ir-wasm.md for detailed architecture documentation.

πŸ”§ Development Status

Current Phase: Phase 0 - Foundation Setup βœ…

  • Workspace structure created
  • Core crate skeletons implemented
  • Basic WASM backend (emits empty functions)
  • Runtime structure (wasmtime integration)
  • CLI tool with compile/validate/codegen/run commands
  • DWARF debug information generation (in progress)
  • Browser DevTools integration validation

πŸ“š Documentation

⏳ Project Phases

Phase Duration Focus Status
Phase 0 2 Weeks Foundation & DWARF Validation In Progress
Phase 1 8 Weeks Core Backend (scalars, vectors, control flow) Not Started
Phase 2 6 Weeks Advanced Features (uniforms, textures, matrices) Not Started
Phase 3 4 Weeks Software Rasterizer Integration Not Started
Phase 4 8 Weeks Codegen Tool & Polish Not Started

πŸ§ͺ Testing

# Run all tests
cargo test

# Test with a simple shader
cargo run --bin webgl2 -- compile tests/fixtures/simple.vert --debug -o output.wasm
cargo run --bin webgl2 -- run output.wasm

🀝 Contributing

This project is in early development. Contributions are welcome once Phase 0 is complete and the architecture is validated.

πŸ“„ License

MIT OR Apache-2.0


οΏ½πŸ’° Resource Requirements

The project will require specialized expertise in:

  • Rust and WASM Development
  • Computer Graphics / GPU Pipeline Implementation (for the emulator)
  • Compiler Design / Language Tooling (for GLSL parsing and source map generation)

Would you like to discuss the specific toolchain recommendations for the GLSL-to-WASM compilation process?