nan

On the other hand, their actual dependency list looks closer to this, and this is definitely not comprehensive: https://github.com/ionic-team/capacitor https://github.com/Microsoft/tslib https://github.com/codemirror https://github.com/codemirror/autocomplete https://github.com/codemirror/language https://github.com/marijnh/style-mod https://github.com/marijnh/crelt https://github.com/marijnh/find-cluster-break https://github.com/marijnh/w3c-keyname https://github.com/cure53/DOMPurify https://github.com/electron/electron https://github.com/electron/get https://github.com/debug-js/debug https://github.com/vercel/ms https://github.com/sindresorhus/env-paths https://github.com/sindresorhus/got https://github.com/sindresorhus/is https://github.com/visionmedia/node-progress https://github.com/npm/node-semver https://github.com/malept/sumchecke

Or use the relatively new NaN helper TypedArrayContents: https://github.com/nodejs/nan/blob/master/doc/v8_misc.md#api...

It looks like that one is pure JS, using native node crypt functionality for randomness. The one recommended in the article[0] actually includes a blowfish C++ binding, with both synchronous and asynchronous methods for comparing and hashing. That binding uses NAN[1] to queue AsyncWorkers. NAN uses libuv[2] behind the scenes to launch a thread[3]. 0. https://www.npmjs.com/package/bcrypt 1. https://github.com/nodejs/nan 2. https://github.com/libuv/libuv 3. https://blog.scottfrees.com/building-an-asynchronous-c-addon...

And it (and native modules) do this pretty well via nan (https://github.com/nodejs/nan).

Node's and V8's APIs changed so frequently that someone created an abstraction layer called NaN ("Native Abstractions for Node.js") for people developing Node binary plugins. https://github.com/nodejs/nan

Its a long goal as of now that everyone want to achieve. You can follow this discussion at https://github.com/nodejs/nan/issues/349

This is definitely a problem today, but NAN (https://github.com/nodejs/nan) which is now the recommended way to build native modules, offers a way out. That project provides a stable API to insulate module developers from changes between v8 versions. Sometime down the line, a NAN-spidermonkey or NAN-chakra project might become feasible.

>The only sensible way to fix this that I can think of is to build in an extension layer into node so that binary packages don't use raw v8 APIs (since those seem to change a lot). Then that layer can be made (more) stable, even in the face of constant v8 upgrading. That already exists, it's called nan: https://github.com/nodejs/nan

JXcore is fork of Node that supports either V8 or SpiderMonkey. I suspect neither the Node or io.js developers are interested because they don't want the extra maintenance costs of supporting more platforms and abstraction layers. However, decoupling your product's core from its third-party dependencies is usually a good idea. V8 API and ABI changes have been enough a problem that someone created the NaN library (Native Abstractions for Node.js) to help Node plugin maintainers. https://github.com/jxcore/jxcore https://github.com/rvagg/nan

The io.js team is working alongside the V8 team to make sure that the upgrades are as easy as possible. Rod Vagg has a project called NAN[0] to try and allow node modules to compile across node versions. [0] https://github.com/rvagg/nan

In the name of competition and architectural flexibility, perhaps the io.js team will be more receptive. :) AFAIK, though, V8 is so tightly integrated into Node core that a third-party library, NaN ("Native Abstractions for Node.js"), was created to improve API compatibility for Node binary addons: https://github.com/rvagg/nan

V8 updates bring incompatible changes to the V8 API, and many node extensions written in C++ need to be updated. Hence the decision to stick with an old V8, even though it's no longer maintained by Google. There is a compatibility layer to assist with making these modules work across different V8 versions: https://github.com/rvagg/nan