https://gcc.gnu.org/bugzilla/show_bug.cgi?id=30484
--- Comment #13 from joseph at codesourcery dot com <joseph at codesourcery dot com> --- For -fwrapv, the mathematical values of INT_MIN / -1 and INT_MIN % -1 should be reduced using modulo arithmetic, so both operations are well-defined, and there is a bug then either operation causes SIGFPE or fails to produce the correct modulo result. Without -fwrapv, INT_MIN / -1 is not representable, so "otherwise, the behavior of both a/b and a%b is undefined" (C11 and later, considered as a defect fix for older standard revisions) applies to make both operations undefined, and there is no bug. That is, there is a bug in the -fwrapv case (when the operations should give well-defined results but GCC fails to implement that), but not in the -fno-wrapv case because both operations are undefined if either one is.
