[RPC] RPC server handle panic exceptions to output error msg

[baichuan3]

In #774 for implementing MoveFunctionCall for the RPC Server, the solution is to introduce another tokio Runtime.

Then it is discovered that the RPC server may not be able to directly capture and process an error. The phenomenon is that the call RPC returns a blank, but there is an error log in the console output.

The tokio Runtime implemented in AggrateService, both AggrateService and RPCService will occur the above situation.

Exception case

let coin_store_handle = coin_module.coin_store_handle(account_addr)?.except("Coin store handle does not exist");

error stack

thread 'tokio-runtime-worker' panicked at 'Coin store handle does not exist', crates/rooch-rpc-server/src/service/aggregate_service.rs:73:31

Normal case after rewriting

let coin_store_handle = coin_module.coin_store_handle(account_addr)?;
// Handle error due to `coin_store_handle` executed in another tokio Runtime
let coin_store_handle = match coin_store_handle {
    Some(v) => v,
    None => anyhow::bail!("Coin store handle does not exist"),
};

Relative #817

[baichuan3]

It is determined that it is not caused by the introduction of dual tikio runtime. The same problem existed in the earlier RPC Server. The reason is that methods such as expect or unwrap will cause panic exceptions, causing the program to exit directly.

[jolestar]

I take this.

[jolestar]

In the multi-thread executor, if a thread panics, the RPC will return nothing.
We need to fix the panic bug or convert it to an error.

This is not a bug. I close this issue. and record two:

1. [Rpc server] Eliminate the static tokio runtime #826
2. Check all expect and unwrap #827

[baichuan3]

I have almost handled over the expect and unwrap in RPC.