filcab added a comment. Please make the tests tighter using `CHECK-NEXT` when possible. Much easier if later anyone needs to debug differences in IR.
================ Comment at: docs/UndefinedBehaviorSanitizer.rst:102 + violating nullability rules does not result in undefined behavior, it + is often unintentional, so UBSan offers to catch it. - ``-fsanitize=object-size``: An attempt to potentially use bytes which ---------------- This is weird. Please just document each of the `nullability-*` in isolation. None of the other flags is documenting more than one flag per bullet point, and I think splitting them up here is easy. It also makes it easier to read. You already added the `nullability` group below, which mentions it adds the three checks. ================ Comment at: docs/UndefinedBehaviorSanitizer.rst:141 - ``-fsanitize=undefined``: All of the checks listed above other than - ``unsigned-integer-overflow``. + ``unsigned-integer-overflow`` and ``nullability``. - ``-fsanitize=undefined-trap``: Deprecated alias of ---------------- ```... and the ``nullability`` group.``` ================ Comment at: test/CodeGenObjC/ubsan-nullability.m:114 + // CHECK: [[NONULL]]: + // CHECK: ret i32* +} ---------------- `CHECK-NEXT`? https://reviews.llvm.org/D30762 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits