Is there a reason that it matters what the reference implementation of Libra is written in?
A blockchain network is usually not considered healthy until it has multiple implementations of nodes running non-negligible parts of the network. I would say that Libra having such alternate impls is almost inevitable.
Exploiting any flaw in a particular blockchain node impl, when there are multiple such impls, would then require either finding the same vulnerability in all other clients; or else, attacking that one client, forcing a hard fork, and then having that client’s version of events “win” such that other node software choose to adopt it by hardcoding a switchover. Neither is ever very likely.
A blockchain network is usually not considered healthy until it has multiple implementations of nodes running non-negligible parts of the network. I would say that Libra having such alternate impls is almost inevitable.
Exploiting any flaw in a particular blockchain node impl, when there are multiple such impls, would then require either finding the same vulnerability in all other clients; or else, attacking that one client, forcing a hard fork, and then having that client’s version of events “win” such that other node software choose to adopt it by hardcoding a switchover. Neither is ever very likely.