adjust for div/rem overflow being UB

Author: RalfJung

tests/compile-fail/intrinsics/exact_div4.rs

@@ -1,5 +1,5 @@
 #![feature(core_intrinsics)]
 fn main() {
     // divison of MIN by -1
-    unsafe { std::intrinsics::exact_div(i64::MIN, -1); } //~ ERROR result of dividing MIN by -1 cannot be represented
+    unsafe { std::intrinsics::exact_div(i64::MIN, -1); } //~ ERROR overflow in signed remainder (dividing MIN by -1)
 }

tests/compile-fail/intrinsics/unchecked_div1.rs

@@ -1,5 +1,5 @@
 #![feature(core_intrinsics)]
 fn main() {
     // MIN/-1 cannot be represented
-    unsafe { std::intrinsics::unchecked_div(i16::MIN, -1); } //~ ERROR overflow executing `unchecked_div`
+    unsafe { std::intrinsics::unchecked_div(i16::MIN, -1); } //~ ERROR overflow in signed division (dividing MIN by -1)
 }

tests/run-pass/integer-ops.rs

@@ -116,6 +116,9 @@ pub fn main() {
     assert_eq!(100i8.wrapping_rem(10), 0);
     assert_eq!((-128i8).wrapping_rem(-1), 0);
 
+    assert_eq!(i32::MIN.wrapping_div(-1), i32::MIN);
+    assert_eq!(i32::MIN.wrapping_rem(-1), 0);
+
     assert_eq!(100i8.wrapping_neg(), -100);
     assert_eq!((-128i8).wrapping_neg(), -128);