Hi Thomas, > I suspected a testism indeed. Could we have an ARM specific expected result? I > am concerned that an XFAIL would mask a real regression in the future. Sure - but could you do it ? I am a bit swamped at the moment. Cheers Nick