Implement the Wasm component-model proposal

[Robbepop]

This is about adding support for the component-model Wasm proposal.

The component-model Wasm proposal is still in early Phase 1 and thus stabilization cannot be expected to happen soon, however, the people working on the Wasm standard are extremely convinced that it is going to become the defacto standard for all future Wasm runtimes to be supported since it is a huge enabler for the multi-language interopt vision of Wasm.

[syrusakbary]

Note: component model is in Phase 1 (not 4):
https://github.com/WebAssembly/proposals?tab=readme-ov-file#phase-1---feature-proposal-cg

From what other browsers have inputted, I believe is unlikely to transition phase any time in the close horizon

[Robbepop]

@syrusakbary Thank you for the correction. I fixed the typo to avoid confusion.

I usually try to only implement phase 4+ proposals in Wasmi since that provides safety to avoid maintenance burden with changes to the spec as is common for early phase proposals. However, with the component-model I feel people are really serious about its inclusion into the standard and it seems to be in a pretty developed state already for a phase 1 proposal. I might be wrong about this though.

[juntyr]

It is already possible to use the component model with wasmi (as long as no multi-memory is required, see #776) using the wasm_component_layer and wasmi_runtime_layer crates (the former provides component model emulation using multi-memory, the later is an adapter for the wasm_runtime_layer abstraction over various WASM runtimes).

[ZheniaZuser]

Is this still planned? I see its removed from readme as planned.

[Robbepop]

Is this still planned? I see its removed from readme as planned.

It is not planned nor unplanned. There are just many Wasm proposals that also need to be implemented and that have higher priority right now. Furthermore, the Wasm component-model is still not really stable and thus it won't make a lot of sense to be implemented at the moment.

The list of Wasm proposals that are planned to be implemented more eagerly is roughly the following:

1. memory64
2. simd
3. wide-arithmetic (mostly because it is so smol)
4. function-references
5. gc
6. and then maybe the component-model

I just removed the component-model proposal from the list because it is not expected to be implemented very soon but eventually.

Though according to @juntyr it is already possible to somehow use Wasmi with the component-model as can be read here.