Monte-to-WASM compiler #83
Description
@erights asked me how hard it would it be to compile Monte to wasm.
I said the anf branch looks like the hard part is done, though the issue of GC (and runtime) remains.
Meanwhile, I see this discussion (I presume it's OK to publish...):
dustyweb (aka @cwebber):
WASM-GC will (when it eventually happens, which may take 2 to 3 years) have unforgeable references and all the things #erights probably rwants :)
https://github.com/WebAssembly/gc/blob/master/proposals/gc/Overview.md the right thing to read if you want to read it
simpson (aka @MostAwesomeDude):
dustyweb: I may very well have written a Monte-to-WASM compiler by that point, although probably not.
dustyweb
simpson: cool!
I sat in on the WASM group during TPAC btw
I was... super impressed.
simpson:
"Efficient interoperability with embedder heap" ah, yes, I too would like to ensure that my GC'd high-level languages retain the ability to spray the browser's heap.
Er, I meant for a ~ sarcasm mark.
dustyweb
simpson: in the WASM-GC proposal memory would be managed by the browser though
so it wouldn't be "dangerous"
but yes it may be a spray :)
simpson
"an untyped language (e.g., a subset of Scheme or Python or something else)" Hmm, I wonder what people are thinking of. Hopefully a member of the E family!
Yeah, I know it's up to the boundary of the browser, it's just that none of the browsers are built cap-safe internally yet.