My point with those comparisons is that you have not bothered to define a reasonable and verifiable standard for what counts as "contrived" code - which is what ultimately seems to determine whether a language is memory safe, according to your definition.