http://open-std.org/JTC1/SC22/WG21/docs/papers/2014/n3839.pdf

http://open-std.org/JTC1/SC22/WG21/docs/papers/2013/n3578.pdf

Bristol 2013: Considered NAD, too early for such a breaking change after C++11, breaks valid programs that use C++11 semantics (defaulted destructor outside class definition, otherwise generated members, used with various smart pointer members).

Issaquah 2014: Brown pointed out that he wasn't in Bristol. Vandevoorde explained that users rely on the current semantics, and that there's existing code in various library headers that would break with this change, and users cannot necessarily fix that breakage due to licensing issues that prevent modifying such code. Voutilainen pointed out that while there are cases where not suppressing copy operations when a destructor is declared leads to incorrect code, there's also quite many cases where the code is correct. Brown opined that he doesn't want to close the door for this kind of changes forever, but EWG didn't see a practical way to be able to enable this rule. The issue therefore stays closed.