Add tracers to zk server

[gianbelinche]

What ❔

This PR adds tracers for the zk server, it is basically a merge between branches era_vm_zk_server_v2 and implement_tracers with a few changes to make it work.

All tests under zk test i server pass.

Note: A couple of tracers, StorageInvocations and ValidationTracer, were not needed to make te tests pass, so they were not implemented for era vm, TODO comments were left for them in case we need them in the future.

The timeout for tests was increased in order to make the tests pass, however this is due to the era vm being too slow, this could be changed later once it is well optimized.

Related to:
lambdaclass/era-compiler-tester#24
lambdaclass/era_vm#221

Why ❔

Checklist

• PR title corresponds to the body of PR (we generate changelog entries from PRs).
• Tests for the changes have been added / updated.
• Documentation comments have been added / updated.
• Code has been formatted via zk fmt and zk lint.

[fkrause98]

Consider this change, I think it's clearer:

Suggested change

	let output = if fat_data_pointer.is_pointer {
	let fat_data_pointer = FatPointer::decode(fat_data_pointer.value);
	if fat_data_pointer.len == 0 && fat_data_pointer.offset == 0 {
	Some(
	execution
	.heaps
	.get(fat_data_pointer.page)
	.unwrap()
	.read_unaligned_from_pointer(&fat_data_pointer)
	.unwrap(),
	)
	} else {
	None
	}
	} else {
	None
	};
	let output = match fat_data_pointer.is_pointer {
	true => {
	let fat_data_pointer = FatPointer::decode(fat_data_pointer.value);
	match (fat_data_pointer.len, fat_data_pointer.offset) {
	(0, 0) => execution
	.heaps
	.get(fat_data_pointer.page)
	.and_then(|ptr| ptr.read_unaligned_from_pointer(&fat_data_pointer).ok()),
	_ => None,
	}
	}
	_ => None,
	};

[fkrause98]

I don't like borrowing refcells outside of the implant block for this struct:
The reason is that this will panic in runtime if there is another living reference to it, which it could possibly happen given that this field is public, let's change this, do we need to change something else in our vm?

[Oppen]

You can reduce nesting a bit by using let <destructuring pattern> else here:

Suggested change

	let call_type = if let CallType::Call(far_call) = current_call.r#type {
	if matches!(far_call, FarCallOpcode::Mimic) {
	let previous_caller = execution
	.running_contexts
	.first()
	.map(|call| call.caller)
	// Actually it's safe to just unwrap here, because we have at least one call in the stack
	// But i want to be sure that we will not have any problems in the future
	.unwrap_or(current.caller);
	if previous_caller == CONTRACT_DEPLOYER_ADDRESS {
	CallType::Create
	} else {
	CallType::Call(far_call)
	}
	} else {
	CallType::Call(far_call)
	}
	} else {
	return;
	};
	let CallType::Call(far_call) = current_call.r#type else {
	return;
	}
	let call_type = if matches!(far_call, FarCallOpcode::Mimic) {
	let previous_caller = execution
	.running_contexts
	.first()
	.map(|call| call.caller)
	// Actually it's safe to just unwrap here, because we have at least one call in the stack
	// But i want to be sure that we will not have any problems in the future
	.unwrap_or(current.caller);
	if previous_caller == CONTRACT_DEPLOYER_ADDRESS {
	CallType::Create
	} else {
	CallType::Call(far_call)
	}
	} else {
	CallType::Call(far_call)
	};

[gianbelinche]

The thing is that this code is exactly as is for the other vms, so I think it is better to leave it as is. The only change was the one from the unreachable, so we don't stop execution due to a tracer (even though in the other vms they still have it)