https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119876
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> --- Note the difference between what I get with PR 119920 and what LLVM produces is due to the generation of `b[i] > 0 ? 1 : 2` for GCC; should we file this seperately? Jan where you did notice this?
