http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49395
Jason Merrill <jason at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |ASSIGNED Last reconfirmed| |2011.06.22 20:36:09 CC| |jason at gcc dot gnu.org AssignedTo|unassigned at gcc dot |jason at gcc dot gnu.org |gnu.org | Ever Confirmed|0 |1 --- Comment #1 from Jason Merrill <jason at gcc dot gnu.org> 2011-06-22 20:36:09 UTC --- I think line 10 is definitely a bug, line 8 may be a standards issue; I notice that EDG also complains about that line. Since A().i is a subobject of a class prvalue, we might want to treat it differently from a normal int prvalue.