Is there a reason to not do that? Other than six-folding the test-time for g++.dg parts, I mean. If not, patch at 11. I imagine there are (at least) lots of of hidden eh-with-optimization bugs crawling there, ready to be eaten^Wsquished. brgds, H-P