https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112942
--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #4) > Just giving background information of the wording changes to C++17 on > std::variant and std::swap here and about the testcase you pointed to and > about when it was added specifcially while changing for that defect report. The test case verifies the LWG 2766 behaviour, *not* the LWG 2749 behaviour.
