feat(native): enhance traceability with additional logs for better debugging

[goshawk-3]

Summary

Simplifies debugging by adding trace logs for key actions such as execute, commit, queue, flush, and compute.

Changes

• Most traces are now logged at the log.debug level with a few in log.info.
• In Geth, enable debug-level logging using the --verbosity 4 CLI parameter.

Examples for Filtering Logs

• grep fhevm: Display all FHEVM-related logs.
• grep handle=59b4f9: Filter logs by specific handle.
• grep 'sync compute': Display input and output of grpc.SyncCompute.

DEBUG[12-18|12:21:25.859] Add operation                            module="fhevm:sync compute" op=FHE_TRIVIAL_ENCRYPT handle=59b4f9..500300
INFO [12-18|12:21:25.859] Sending grpc request                     module="fhevm:sync compute" computations=1 "compressed ciphertexts"=0
INFO [12-18|12:21:25.864] Response                                 module="fhevm:sync compute" "ciphertexts count"=1
DEBUG[12-18|12:21:25.864] Response ciphertext                      module="fhevm:sync compute" handle=59b4f9..500300 len=1866
DEBUG[12-18|12:21:25.864] Computations completed                   module="fhevm:sync compute" duration=5ms

• grep 'Insert computation' fhevm.log | grep 'CommitBlockId: 15' - Display computations inserted in Block height 10

[david-zk]

I think we need to ensure that whatever logger the integrator uses it should work for him

[david-zk]

Do we know that this is the logging library that every integrator will use?

We had logging before in avalanche integration, so we had to expose in the api so user could pass in his own logger interface and we'd use that. I'm not sure we can assume logging implementation users will use, that's why this PR was not trivial to implement right away

[david-zk]

This is how we did the logger before https://github.com/zama-ai/fhevm-go/blob/53d5dae4d6a4c9d887ce6cb484b642b964efc690/fhevm/interface.go#L26 we allow user to provide interface object as part of the environment

[david-zk]

Maybe the perfect place to pass in the logger would be once in adding argument to func InitExecutor ? CreateSession happens many times in every EVM call, we could keep one pointer to logger after InitExecutor and keep that for the entirety of application

[goshawk-3]

Maybe the perfect place to pass in the logger would be once in adding argument to func InitExecutor ? CreateSession happens many times in every EVM call, we could keep one pointer to logger after InitExecutor and keep that for the entirety of application

Makes sense!

For the record, InitLogger( hostLogger) is now called during the blockchain host initialization.