#js_engine

• Conformance on test262: each file is run in strict and non-strict modes and reported per mode. Do not sum the modes. Generate current numbers from CI artifacts with make test262-report; see docs/TEST262.md.
• Cross-target embedding: the same stateful Engine API is tested on native, JavaScript, Wasm, and Wasm-GC.
• Benchmark dashboard: https://dowdiness.github.io/js_engine/benchmarks/

#Quick Start

#CLI

moon run cmd/main -- 'console.log(1 + 2)'
# 3

moon run cmd/main -- '
function fib(n) {
  if (n <= 1) { return n; }
  return fib(n - 1) + fib(n - 2);
}
console.log(fib(10));
'
# 55

#As a Library

///|
test "README stateful rule engine" {
  let engine = @js_engine.Engine()
  let source =
    #|let evaluations = 0;
    #|function allow(request) {
    #|  evaluations += 1;
    #|  return { allowed: request.role === "admin", evaluations };
    #|}
  engine.eval(source)
  let admin = Json::object({ "role": Json::string("admin") })
  let member_request = Json::object({ "role": Json::string("member") })
  json_inspect(engine.call_json("allow", [admin]), content={
    "allowed": true,
    "evaluations": 1,
  })
  json_inspect(engine.call_json("allow", [member_request]), content={
    "allowed": false,
    "evaluations": 2,
  })
}

///|
test "README one-shot facade" {
  let (output, _) = @js_engine.run("console.log(1 + 2)")
  json_inspect(output, content=["3"])
}

• Stable embedding: run, Engine, EngineError, and the unbounded persistent Engine methods listed in the guide.
• Staged Stage 4 availability: Engine::eval_bounded, Engine::call_json_bounded, Engine::run_microtask_checkpoint_bounded, Engine::run_timer_checkpoint_bounded, ExecutionPolicy, ExecutionPolicyError, and InterruptionHandle.
• Compatibility: run_module / run_modules; their export maps expose raw runtime values.
• Advanced/internal: run_compiled and the module-level event-loop APIs that expose or accept a raw interpreter.

#Embedding (custom host objects)

let interp = @interpreter.new_interpreter()
// Build query_selector with realm_state=Some(interp.realm_state) — see guide.
let document = @runtime.make_host_object(
  name="Document",
  proto=@runtime.get_obj_proto(realm_state=Some(interp.realm_state)),
  methods={ "querySelector": query_selector },
)
interp.global.def_builtin("document", document)
// Then parse, interp.run, interp.run_microtasks(), interp.run_timers().

#Supported Language

#Conformance

#strict

#non-strict

#Package Structure

token/          Token types and source locations
errors/         JavaScript error variants and formatting helpers
lexer/          Tokenizer
ast/            AST node definitions
parser/         Recursive descent parser with Pratt precedence
static_semantics/  Early-error and declaration-fact analysis
compiler/       Opt-in closure-conversion prototype
interpreter/    Wiring layer for runtime + standard library
interpreter/runtime/  Tree-walking evaluator, value model, host state
interpreter/stdlib/   JavaScript built-ins
cmd/main/       CLI entry point
cmd/test262_runner/  Native test262 runner
cmd/report_test262/  CI artifact report generator
benchmarks/     Benchmark workloads and runner
example/rule_engine/  Canonical stateful JSON rule-engine embedding

#Development

moon check        # Type check
moon test         # Run unit tests
moon fmt          # Format code
moon info         # Update .mbti interface files
moon build        # Build

#Documentation

• docs/README.md — start here for deeper material
• docs/development.md — maintainer workflow and generated files
• docs/ROADMAP.md — current status and active roadmap
• docs/supported-features.md — per-category conformance, Annex B, and missing features
• docs/GLOSSARY.md — terminology used in the code and docs
• AGENTS.md — MoonBit coding conventions (also used by AI agents)

#License