driftjs-compiler

Ahead-of-time (AOT) compiler for DriftJS .drift Single File Components (SFCs).

Converts .drift template markup and <script> blocks into compact, linear 32-bit Uint32Array register VM bytecode instructions and optimized javascript thunk arrays.

Installation

pnpm add driftjs-compiler
# or
npm install driftjs-compiler

Features

• Lexer: Fast tokenization with nested template expression and string literal tracking.
• Parser: Parses token streams into structured AST nodes (ElementNode, TextNode, InterpolationNode, ScriptNode).
• Analyzer: Analyzes AST using Acorn, tracks reactive variable scope, transforms state mutations into zero-overhead dirty mask markers (markDirty), and assigns 32-bit register indexes.
• Generator: Emits binary bytecode, serialized thunks, and update block offsets for execution by the DriftJS VM.

API Usage

import { DriftJSLexer, DriftJSParser, DriftJSAnalyzer, DriftJSGenerator, generate } from 'driftjs-compiler';

const source = `
<script>
  let count = 0;
  function increment() { count++; }
</script>

<button onclick={increment}>Count: {count}</button>
`;

// One-shot compilation helper
const program = generate(source);

console.log(program.bytecode);          // Uint32Array instruction stream
console.log(program.constants);         // Constants pool & compiled thunks
console.log(program.updateBlockOffset); // Bytecode offset for reactive updates

Architecture

The compiler pipeline runs as follows:

.drift Source -> Lexer -> Tokens -> Parser -> AST -> Analyzer -> AnalysisResult -> Generator -> CompiledProgram

License

MIT © Hrutav Modha