Can this be replaced by webassembly version?

[lygstate]

[lpinca]

Yes, probably, but it must be faster otherwise it does not make sense.

[lygstate]

Yes, probably, but it must be faster otherwise it does not make sense.

It's make sense when I wanna webpack ws, does not needs to be faster, equal performance is enough

[lpinca]

You can connect native modules with webpack.

[lygstate]

You can connect native modules with webpack.

The webpack result are not portable, for example, it's can not running under both windows/linux/osx

[lpinca]

The webpack result are not portable, for example, it's can not running under both windows/linux/osx

I think you can produce different bundles picking the correct prebuilt binary and use the correct bundle at runtime. I see your point but I'm not sure if it is a common use case.

[lygstate]

The webpack result are not portable, for example, it's can not running under both windows/linux/osx

I think you can produce different bundles picking the correct prebuilt binary and use the correct bundle at runtime. I see your point but I'm not sure if it is a common use case.

cross-compiling is a nightmare, for example for windows, we have MSVC,GNU,Electron, for Linux we have MUSL,GLIBC and different architecture

[lpinca]

There are already prebuilt binaries for Windows, macOS, and Linux (glibc) that cover 99% of use cases. Anyway feel free to open a PR.

[jerch]

To me this seems to be a perfect wasm-simd candidate.

[ChALkeR]

The js version can be made significantly faster, if needed. This likely doesn't need a webassembly version

[lpinca]

See https://github.com/tsctx/ws-masking for a WebAssembly variant.

[ChALkeR]

Yeah, that also has an optimized pure js impl which is much faster than simple one. It's unpublished from npm though.