Reorder arguments for multiProof functions

[frangio]

Resolves part of #3441.

Reorders arguments for consistency with simple proofs.

[Amxx]

While I understand the logic here, I must say I'm not a fan of this new format.

In particular because it separates the proof and the proofFlag arguments. The thing is, in multiProofVerify, the proof array should possibly be called "proofSiblings", and it is useless without the "proofFlags" array.

Afaik, the idea for having the proof being the first argument of the single version was to be able to do:

using Merkle for bytes32[];
...
proof.verify(root, leaf)

IMO this was a bad design from the start. Rater then proving a property on the proof, using the root, I would rather prove a property on the tree itself (namely that it contains a leaf) using the proof as an argument.

So I would personally have preferred

• root.verifyInclusion(leaf, proof)
• leaf.rebuildRoot(proof)
• root.verifyIncusion(leaves, proofSiblings, proofFlags) (overloaded)
• leaves.rebuildRoot(proofSiblings, proofFlags) (overloaded)

[frangio]

Reordered to proof, proofFlags, root, leaves. I also edited the documentation.