On Thu, Apr 24, 2014 at 1:42 PM, Matt Godbolt <matt@xxxxxxxxxxx> wrote: > I will do so and follow up here when I have a minimal example. > As promised: PR filed as: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60955 Minimal test case: --- unsigned int erroneous_warning(register int a) { if ((a) & 0xff) return 1; else return 0; } --- it's a combination of std=c++1y, register, and double parens around the register'd variable. Very peculiar: I can't think of what c++1y feature this might be tickling to cause the problem! Thanks again, Matt
