On Jan 24, 7:51pm, Martin von Loewis wrote: > Did you mean ::operator new[](std::size_t, void*), here? That is the > one where special-casing is useful. Oops, yes. That was a typo or something. My "Proposal A" is to have no special cases, and my "Proposal B" is to treat ::operator new[](std::size_t, void*) as a special case. --Matt