I wonder if most failures would occur in a way that would not harm the shooter before the rifle became unusable.

Been a while since I read his description, but I think they had to be hammered open to continue the test long before anything catastrophic happened. I think the lugs held and never did shear or let go.