https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118868
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> --- (In reply to Richard Biener from comment #3) > Hmm, I'm not sure the GIMPLE verification is sensible, I'd have simply > required compatible types. Right, I agree; I suspect I missed that PAREN_EXPR is not fully an unary operator and just a pass through one. I am still deciding if we should reject __builtin_assoc_barrier for pointers or not.
