Potential problem in Secure Message Delivery

[fizyk20]

Consider the following scenario:

• A section X recently split into X0 and X1.
• X0 sends two messages to Y in close succession. Both contain old_prefix = X and current_prefix = X0.
• Upon receipt of the first message, Y updates the map of the keys, replacing prefix X with X0 and X1.
• Upon receipt of the second message, Y looks for a key for prefix X, but there is no such prefix in the map anymore.

This case illustrates that we would still need to look for compatible prefixes in the map, even if we have two prefixes in the message, so it might make sense to only include a single prefix (the current one) in the message.