> Did anyone try to map this code back to the original codebase to see if this memory problem exists in the original codebase?

FWIW what is being discussed is not memory problems, it's breaking rust invariants (the unsafe code has to follow specific rules, e.g. annotate lifetimes properly).