This is great, and pretty damning evidence of traditional, unsafe languages.
Do we have any data on how much of iOS is implemented in unsafe code vs in macOS? In particular, I want to know if the difference in memory safety ratio correlates with a language choice ratio. I don’t think sufficient information to make that determination is public, but I would love to be wrong about that.
(I guess one of the problems with that measurement would be is that it’s hard to tell if an ObjectiveC file has parts of straight C in it, which you’d expect more of in system code.)
