Skip to content

Frequently Asked Questions

What is the difference between emnapi and native Node-API

  • Most emnapi's APIs are implemented by JavaScript, while native Node-API is implemented by Node.js C++ code and has full access to V8 engine.

  • emnapi is compiled to WebAssembly by Emscripten, while native Node-API is compiled to Node.js addon (.node is operating system shared library) and can use operating system APIs.

  • emnapi has no Node.js specific APIs, some of APIs depends the host JavaScript environment. In particular, WebAssembly does not share heap memory with the host, and the API related to ArrayBuffer has certain limitations. see API List

  • Using emnapi is able to use the builtin node-addon-api out of box if the if the runtime support FinalizationRegistry and WeakRef.

When should I use emnapi

  • You prefer Node-API, and you are more familiar with Node-API than embind.
  • You want to port your (or existing) Node.js addon written in Node-API to WebAssembly.
  • You want your native module npm package to be installed smoothly by users without having to deal with node-gyp.

How to port existing Node.js addon written in Node-API

WARNING

It is very difficult to port a native addon which has heavy use of operating system APIs (especially heavy use of Windows.h).

You can not use node-addon-api if the runtime does not support FinalizationRegistry and WeakRef.

  1. Checking if all APIs used in the addon are implemented in emnapi. See API List.
  2. Checking runtime weak reference support by using emnapi_is_support_weakref before calling napi_create_reference / napi_wrap.
  3. Checking if it is necessary to sync wasm memory according to ArrayBuffer Related API
  4. Writing CMakeLists. node-gyp is not supported.

Released under the MIT License.