Then even in the worst case scenario, they were addressing this issue two days after it was publicly disclosed. So this wasn't a "rush to fix the zero day ASAP" scenario, which makes it harder to justify ignoring errors that started occuring in a small scale rollout.