I would expect these bugs to keep occurring, because they're still approaching the development of the consensus algorithm incorrectly. Instead of formally proving correctness, they're doing something that seems right and then patching bugs when they are discovered.