Geoffrey Keating <geoffk@geoffk.org> writes: > For malloc-ed memory, I strongly support the > idea that the first store should determine its type and the type can't > change after that. If I understand that correctly, that would break many existing programs, including BFD. Ian