WASM is in a questionable state

[FreezyLemon]

• There are wasm_bindgen dependencies in the manifest even though it's never used anywhere (there are 0 wasm_bindgen tests). This also makes the CI job running WASM tests redundant.
• The DATA_ALIGNMENT change was introduced to rav1e in early 2020 and I'm not sure if it's still needed
• more WASM targets exist now (wasip1, wasip2, wasip3) and I'm not 100% sure what to target as the minimum/what to support
  • wasm32-unknown-unknown/wasm32-unknown-emscripten are the old "base" targets
  • the p1/p2/p3 extensions kind of build on top of it, p2 introduces a big change with the component model thing (and way more of std actually works)
  • p1 and higher might support higher alignments?
  • tooling seems to be very different between WASI and older targets

I don't know much about WASM but this might need some investigation and changes.

[FreezyLemon]

WASI seems fine now, but wasm32-unknown-unknown and wasm32-unknown-emscripten are still in a bit of an unknown state.

I think it's better to not create any WASM bindings in v_frame and let downstreams handle it instead, so removal of wasm_bindgen makes sense in hindsight. Maybe an integration test could be added just for WASM...