[FEATURE] Support unique message identifier in `computeL2ToL1MembershipWitness`

[jasonxh]

Problem Statement

There has been this lingering TODO in the code for a while now:

aztec-packages/yarn-project/stdlib/src/messaging/l2_to_l1_membership.ts

Lines 118 to 120 in e6dc4a6

	// TODO: Allow to specify the message to consume by its index or by an offset, in case there are multiple messages with
	// the same value.
	export function computeL2ToL1MembershipWitnessFromMessagesInEpoch(

With the v4 update, L2 -> L1 messages rollup has been changed from block level to epoch level, which further exacerbated this problem - it's much more likely to have duplicate L2 -> L1 messages in an epoch than in a block. As is, there doesn't seem to be a way to compute a message witness for any duplicate messages except the first one.

Proposed Solution

Implement a more intuitive and uniquely identifying computeL2ToL1MembershipWitness signature, such as

export type L2ToL1MembershipWitness = {
  epoch: EpochNumber;  // Include this in output as opposed to input
  root: Fr;
  leafIndex: bigint;
  siblingPath: SiblingPath<number>;
};

export async function computeL2ToL1MembershipWitness(
  node: AztecNode,
  txHash: TxHash,
  messageIndexInTx: number,
): Promise<L2ToL1MembershipWitness | undefined>;

which internally figures out the message's full coordinates, i.e. (epoch, checkpointIndex, blockIndex, txIndex, messageIndexInTx).

Example Use Case

No response

Alternative Solutions

No response

Additional Context

No response