On Wed, Dec 14, 2011 at 4:10 PM, Jason Merrill <ja...@redhat.com> wrote: > Here we were assuming that the type of a POINTER_PLUS_EXPR would reflect the > underlying type of the object, which is not the case with this testcase. So > now we keep looking until we find the ADDR_EXPR. > > Tested x86_64-pc-linux-gnu, applying to trunk.
my fault. Thanks.
