As of now, the implementation of the gio function in the DaveConsensus contract uses a hard-coded tree height that covers 2 MB, but what we should do, instead is:

• Compute the Merkle tree of the smallest tree that fits the response data
• This tree must encompass at least 32 bytes
• The function should return, besides the Merkle root, the size of the response data (in bytes)

As it is already implemented, when computing the Merkle root of the response data, we should pad it with zeros to the next power of 2.