On 6/7/2016 3:23 PM, Timon Gehr wrote: > Obviously they proved the virtual machine itself memory safe, As I recall, the proof was broken, not the implementation. People do make mistakes and overlook cases with proofs. There's nothing magical about them.