(I paste the comment there as there's no such an issue about WebAssembly support.)
Maybe we could have some collaboration between the existing efforts for compiling Haskell programs to the Web. The community already has ghcjs, webghc and Asterius. And the third-party toolchains and infras are becoming mature in past years, e.g., the LLVM support for wasm32/wasm64 backends, binaryen, and many faster wasm runtimes.
Maintaining both backend for JS and wasm looks not a good idea, and I'm also wondering if ghc could accept a progressive approach for supporting compiling to the Web, e.g.,
I think for all existing efforts, it is non-trivial to answer all questions for full-featured support of js/wasm in ghc, but all existing projects have reached a status that can run some non-trivial applications but diverge from the ghc main branch after months of development and becoming a monster MR and hard to gain consensus to be landed.
Thus I'm wondering if a progressive manner is acceptable for compiling for the Web.
(I'm recently also working on adding wasm32-unknown-wasi target for ghc, but with the LLVM backend. I think there should be other undisclosed efforts towards the support of js/wasm in ghc. And all projects should face the same problems (FFI, js interop with js, TH, interpreter, threaded rts, etc.) after supporting compiling a "hello-world" or "fibonacci").