Unused emitted_events and return_values in Private Kernel Circuit?

[suyash67]

We have emitted_events and return_values as a part of the PrivateCircuitPublicInputs:

aztec-packages/circuits/cpp/src/aztec3/circuits/abis/private_circuit_public_inputs.hpp

Lines 31 to 33 in ab006b0

	std::array<fr, RETURN_VALUES_LENGTH> return_values = zero_array<fr, RETURN_VALUES_LENGTH>();
	std::array<fr, EMITTED_EVENTS_LENGTH> emitted_events = zero_array<fr, EMITTED_EVENTS_LENGTH>();

The emitted_events are probably copied over to the OptionallyRevealedData struct in the PKC but the return_values aren't. Need to check if this a bug.

aztec-packages/circuits/cpp/src/aztec3/circuits/kernel/private/native_private_kernel_circuit_inner.cpp

Line 61 in ab006b0

end.optionally_revealed_data = start.optionally_revealed_data;

aztec-packages/circuits/cpp/src/aztec3/circuits/abis/optionally_revealed_data.hpp

Line 23 in ab006b0

std::array<fr, EMITTED_EVENTS_LENGTH> emitted_events = zero_array<fr, EMITTED_EVENTS_LENGTH>();

[benesjan]

emitted_events will be replaced in this PR.

[benesjan]

The return values are being used as described by Mike on slack so this issue can be closed.

Mike's response on slack:

The return values of a private function call should be constrained by the private kernel circuit. At the very least, they’re the public inputs of the ‘private call’ proof which will be verified by the kernel circuit (when we do proper, non-simulated proofs).
I think the return values are also implicitly checked to be consistent (between the function making the call, and the called function) due to the fact the public inputs (which include the return values) are included in the call stack item hashes, which are pushed to the call stack in one kernel iteration, and then popped from the call stack in a later iteration.