Kaveh R. Ghazi writes: > PS: I also saw a few tree checking failures at the time, which are > also documented in GNATS. Were you able to reproduce those? No. I put this particular PR in analyzed mode two/three weeks ago (2697 if memory serves well.) ./A