feat: impl LongRefFromWasmAbi for T where T: Tsify + FromWasmAbi

[Pantamis]

Implement #34

I tried to make a test of the behavior by building a test very similar to #33 using async function for validate and it fails 😧...

which is expected because Rust must be responsible for dropping the reference to the closure here so you cannot invoke the closure twice and we have no way to check it in Rust according to rustwasm/wasm-bindgen#3178 so this suggest my implementation is right actually 😃 ...

But it also fails if I use the same implementation as RefFromWasmAbi for the LongRefFromWasmAbi 😧, which normally does not drop the reference at the end of the first Rust invocation and so allow the second invocation to be processed.

In summary, I am not sure how to test that my implementation is the correct one 😕

I leave this PR here for people who are maybe more knowledgeable building a test for this. My main issue is that we cannot take an async closure as function argument to reproduce a test like in aaaa398 where validation is async

[siefkenj]

It looks like the expansion tests are the only ones that are failing? Or did you remove a different failing test?

[Pantamis]

Yeah I forgot to change expansion results in test, I am on it !

[Pantamis]

All good for the CI now

I did not include the test using an async function for now because I am not really sure about what it is really testing 😅 (so tests did not change at all, I just completed expansion results)

If my implementation is bad, the worst case should be that we leak memory in the ABI heap because Rust may not clean the imported JsValue when dropping it (obviously, that's not what I have done and I want to test I don't do it), this may happen if I copy-paste the implementation of RefFromWasmAbi in the implementation of LongRefFromWasmAbi but I am not able to make a test that show this kind of mistake clearly.

Leaking memory is safe but ... yeah we should avoid it so it would be nice to test that a correct implementation avoid it 😒