ValidationContext::compute_packet_commitment should be a standalone exported function

[plafer]

There's only one correct way to compute a commitment. This would also make it available for implementors e.g. of TokenTransferValidationContext.

[Farhad-Shabani]

When you say "one correct way", do you mean it's required by the protocol? Wouldn't be a different implementation for a non-tendermint chain?
If that's the case, I think it would make sense to apply this to the ack_commitment() method as well.

[plafer]

When you say "one correct way", do you mean it's required by the protocol?

Yes; if a chain computes it differently from any other chain, the Merkle proof for checking that the packet commitment is there will fail. This is independent from the host consensus algorithm.

If that's the case, I think it would make sense to apply this to the ack_commitment() method as well.

Right, we should also export it as standalone. Related: #375