------- Comment #15 from jason at gcc dot gnu dot org 2006-07-05 00:46 ------- (In reply to comment #14) > Marks's testcase from comment #6 still ICEs.
True, but it's a contrived testcase; I doubt anyone deliberately tries to force a load of a non-POD object for its own sake. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=26577
