Comment by titzer

Wasm evolution happens in fits and starts. There's a lot of nested chicken-and-egg problems. We first started on Wasm GC less than two years after MVP, and then we didn't have any language targeting it, so we had to bootstrap it. Now, Java, Dart, Kotlin, and Scala all target Wasm GC (and Virgil too :-)). The interior pointer is on people's radar.

The next big feature coming is stack switching. It works best with unboxed continuations, which necessitates a fat pointer representation in the engine. Once the engine supports fat pointers, then interior pointers will be an easier sell. It might take several years to get there, but Wasm evolves slowly and deliberately, and IMO hasn't made any massive fatal design errors yet.