proposals VS binaryen

Memory64 is supported by a lot of runtimes now, although it isn't fully standardized yet (see https://github.com/WebAssembly/proposals), so no idea how reliable the implementations actually are, since I haven't had a need for that much memory yet

Block only. There’s a tail call proposal[1] that’s in phase 4 (nearly standardized).

[1]: https://github.com/WebAssembly/proposals

While it'd be a nice addition, I wouldn't expect it any time soon.

It's currently still a stage 1 proposal, while we've been waiting for years for other proposals to be merged. The last time a proposal was actually finished was over 2 years ago.

https://github.com/WebAssembly/proposals

https://github.com/WebAssembly/proposals/blob/main/finished-...

Do you know anything about any WASM developments that will enable pure WASM interaction with browser's Web-APIs at no or at a low cost without the JS layer? I'm looking at https://github.com/WebAssembly/proposals and it's very confusing. There are type imports, almost complete GC proposal(which apparently only for GCd languages, but not for anything browser<->wasm), the component model(which looks and sounds as something not for the browser use case), JS String Builtins (which will provide faster JS strings, but not DOM) and ECMAScript module integration (which will turn WASM modules into ES modules, but Web-APIs aren't ES modules so no luck). Sometimes I read contributor interactions and it looks as if providing such functionality isn't their priority or even in their plans, and WASI + component model for cloud and similar use cases are more important.

It's already in Phase 4, so close: https://github.com/WebAssembly/proposals#phase-4---standardi...

Threads are Phase 3

https://github.com/WebAssembly/proposals

You can also check out:

https://webassembly.org/roadmap/

And for Go, the proposal project on Github has many interesting conversations from the devs.

And as a reminder to anyone interested in using Go WASM, it’s experimental and does not come with the same compatibility promise as Go itself:

https://github.com/golang/go/wiki/WebAssembly

GC proposal is from 2018: https://github.com/WebAssembly/proposals/issues/16 and there’s code: https://github.com/WebAssembly/gc/blob/master/proposals/gc/O...

Seems like an awefully long time for progress to be made, given all the possibilities it would unlock.

The proposal was recently bumped to stage 4 (the penultimate stage) with at least a couple of runtimes working on implementing (besides v8, which has supported it for quite awhile now)

https://github.com/WebAssembly/proposals

But basically it boils down to the memory control proposal, which can be found here and is not very far along; Webassembly proposals lists it in stage one.

I've encountered a building problem when using asterius to compile a multi-packages cabal project, the detail could be found here, any suggestions?

ghcjs is the way to go for you, and soon it might be asterius. i do not know how hard it is to set ghcjs up without a framework. but frameworks like obelisk (based on reflex-dom), shpadoinkle, and miso automate that for. i personally like obelisk for its functional reactive programming but it can get awkward and get in your way. so if gui programming is just a means to the end of this one small application and you are not really interested in it nor functional reactive programming, shpadoinkle or miso might suit you better. miso implements the elm architecture (also "TEA", "functional model view controller") and shpadoinkle implements something directly equivalent to the elm architecture. but shpadoinkle achieves more composable widgets by minimalizing the elm architecture. so i recommend shpadoinkle for its better concept although miso is more mature.

Also, hint uses unsafeCoerce, and thus implicitly relies on an assumption about how values are represented at runtime. Namely, if a program P is interpreting an expression E of type A, hint assumes that the value of type A produced by the ghc interpreter has the same representation as the values of type A which are manipulated by program P. This is not guaranteed to be the case, since P has been compiled by the compiler portion of ghc while E has been evaluated by the interpreter portion of ghc. This means the ghc devs had to carefully craft their compiler and interpreters to match. When targetting the browser, a Haskell-to-js or Haskell-to-wasm compiler such as Asterius modifies ghc's code-generator so it produces js or wasm code. You would thus also need to tweak the interpreter so that it produces js or wasm values which match what the modified code-generator outputs. Or you could restrict yourself to the hint's less expressive eval :: String -> String API.

We found a post where someone had a similar issue (here), but the fix in that issue doesn't help: using `ar` from `binutils` causes link errors like this instead:

https://github.com/tweag/asterius/issues/851 (asterius has a demo, but no source, and I -assume no lua filter support)

All modern ones support Haskell: https://github.com/tweag/asterius

Another option is Asterius. I'm not familiar with the current state, and it's not had active development for about 3 months now, either, so it may be in the same boat? But I think the big disadvantage of Asterius is that there's just a lot less usage, and therefore a lot less testing with the whole Haskell ecosystem, versus GHCJS which has been a fixture for a while and where loads of people have thought about compatibility for years.

Check out asterius

I've played around with Haskell via the Asteruis project : https://github.com/tweag/asterius

Also emscripten of course, for C/C++.