fix the overflow warning.

benchmark result:
```
$ cargo bench
   Compiling div-euclid v0.1.0 (/me/div-euclid)
    Finished bench [optimized] target(s) in 1.01s
     Running unittests src/lib.rs (target/release/deps/div_euclid-7a4530ca7817d1ef)

running 7 tests
test tests::it_works ... ignored
test tests::bench_aaabs     ... bench:  10,498,793 ns/iter (+/- 104,360)
test tests::bench_aadefault ... bench:  11,061,862 ns/iter (+/- 94,107)
test tests::bench_abs       ... bench:  10,477,193 ns/iter (+/- 81,942)
test tests::bench_default   ... bench:  10,622,983 ns/iter (+/- 25,119)
test tests::bench_zzabs     ... bench:  10,481,971 ns/iter (+/- 43,787)
test tests::bench_zzdefault ... bench:  11,074,976 ns/iter (+/- 29,633)

test result: ok. 0 passed; 0 failed; 1 ignored; 6 measured; 0 filtered out; finished in 19.35s
```
benchmark code:
```rust
#![feature(test)]
extern crate test;

#[inline(always)]
fn rem_euclid(a:i32,rhs:i32)->i32{
    let r = a % rhs;
    if r < 0 {
        // if rhs is `integer::MIN`, rhs.wrapping_abs() == rhs.wrapping_abs,
        // thus r.wrapping_add(rhs.wrapping_abs()) == r.wrapping_add(rhs) == r - rhs,
        // which suits our need.
        // otherwise, rhs.wrapping_abs() == -rhs, which won't overflow since r is negative.
        r.wrapping_add(rhs.wrapping_abs())
    } else {
        r
    }
}

#[cfg(test)]
mod tests {
    use super::*;
    use test::Bencher;
    use rand::prelude::*;
    use rand::rngs::SmallRng;
    const N:i32=1000;
    #[test]
    fn it_works() {
        let a: i32 = 7; // or any other integer type
        let b = 4;

        let d:Vec<i32>=(-N..=N).collect();
        let n:Vec<i32>=(-N..0).chain(1..=N).collect();

        for i in &d {
            for j in &n {
                assert_eq!(i.rem_euclid(*j),rem_euclid(*i,*j));
            }
        }

        assert_eq!(rem_euclid(a,b), 3);
        assert_eq!(rem_euclid(-a,b), 1);
        assert_eq!(rem_euclid(a,-b), 3);
        assert_eq!(rem_euclid(-a,-b), 1);
    }


    #[bench]
    fn bench_aaabs(b: &mut Bencher) {
        let mut d:Vec<i32>=(-N..=N).collect();
        let mut n:Vec<i32>=(-N..0).chain(1..=N).collect();
        let mut rng=SmallRng::from_seed([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,21]);
        n.shuffle(&mut rng);
        d.shuffle(&mut rng);
        n.shuffle(&mut rng);
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=rem_euclid(*i,*j);
                }
            }
            res
        });
    }
    #[bench]
    fn bench_aadefault(b: &mut Bencher) {
        let mut d:Vec<i32>=(-N..=N).collect();
        let mut n:Vec<i32>=(-N..0).chain(1..=N).collect();
        let mut rng=SmallRng::from_seed([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,21]);
        n.shuffle(&mut rng);
        d.shuffle(&mut rng);
        n.shuffle(&mut rng);
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=i.rem_euclid(*j);
                }
            }
            res
        });
    }

    #[bench]
    fn bench_abs(b: &mut Bencher) {
        let d:Vec<i32>=(-N..=N).collect();
        let n:Vec<i32>=(-N..0).chain(1..=N).collect();
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=rem_euclid(*i,*j);
                }
            }
            res
        });
    }
    #[bench]
    fn bench_default(b: &mut Bencher) {
        let d:Vec<i32>=(-N..=N).collect();
        let n:Vec<i32>=(-N..0).chain(1..=N).collect();
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=i.rem_euclid(*j);
                }
            }
            res
        });
    }

    #[bench]
    fn bench_zzabs(b: &mut Bencher) {
        let mut d:Vec<i32>=(-N..=N).collect();
        let mut n:Vec<i32>=(-N..0).chain(1..=N).collect();
        let mut rng=SmallRng::from_seed([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,21]);
        d.shuffle(&mut rng);
        n.shuffle(&mut rng);
        d.shuffle(&mut rng);
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=rem_euclid(*i,*j);
                }
            }
            res
        });
    }
    #[bench]
    fn bench_zzdefault(b: &mut Bencher) {
        let mut d:Vec<i32>=(-N..=N).collect();
        let mut n:Vec<i32>=(-N..0).chain(1..=N).collect();
        let mut rng=SmallRng::from_seed([1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,21]);
        d.shuffle(&mut rng);
        n.shuffle(&mut rng);
        d.shuffle(&mut rng);
        b.iter(||{
            let mut res=0;
            for i in &d {
                for j in &n {
                    res+=i.rem_euclid(*j);
                }
            }
            res
        });
    }
}
```

library/core/src/num/int_macros.rs

@@ -2067,7 +2067,15 @@ macro_rules! int_impl {
         #[rustc_inherit_overflow_checks]
         pub const fn rem_euclid(self, rhs: Self) -> Self {
             let r = self % rhs;
-            if r < 0 { r + rhs.abs() } else { r }
+            if r < 0 {
+                // if rhs is `integer::MIN`, rhs.wrapping_abs() == rhs.wrapping_abs,
+                // thus r.wrapping_add(rhs.wrapping_abs()) == r.wrapping_add(rhs) == r - rhs,
+                // which suits our need.
+                // otherwise, rhs.wrapping_abs() == -rhs, which won't overflow since r is negative.
+                r.wrapping_add(rhs.wrapping_abs())
+            } else {
+                r
+            }
         }
 
         /// Calculates the quotient of `self` and `rhs`, rounding the result towards negative infinity.