Why does the patch and fix have any architecture or OS-dependency? - David
On Thu, Aug 8, 2013 at 1:57 PM, Paolo Carlini <paolo.carl...@oracle.com> wrote: > On 08/08/2013 07:35 PM, David Edelsohn wrote: >> >> I'm now seeing a new testsuite failure: >> >> FAIL: g++.dg/lookup/typedef2.C -std=c++98 (test for excess errors) >> Excess errors: >> /nasfarm/edelsohn/src/src/gcc/testsuite/g++.dg/lookup/typedef2.C:8:12: >> error: using typedef-name 'Foo1::Bar' after 'struct' >> /nasfarm/edelsohn/src/src/gcc/testsuite/g++.dg/lookup/typedef2.C:8:19: >> error: invalid type in declaration before ';' token >> /nasfarm/edelsohn/src/src/gcc/testsuite/g++.dg/lookup/typedef2.C:18:12: >> error: using typedef-name 'Foo2::Bar' after 'struct' >> /nasfarm/edelsohn/src/src/gcc/testsuite/g++.dg/lookup/typedef2.C:18:19: >> error: invalid type in declaration before ';' token > > Which essentially means that my patchlet fixes the problem on x86_64-linux, > but not on AIX. Because you would see the exact same fails if I revert the > patch and keep the testcase in. > > I'm ready to revert the patch and re-opening or xfail-ing on AIX. Jason call > I guess. > > Paolo.
