I’m learning about two’s complement and wanted to see what the negation of TMIN (the minimum value for a signed type) is. I wrote the following code for int:

#include <stdio.h>
int main()
{
    int i = -2147483648;
    printf("%dn", i);
    printf("Size of int: %zu bytesn", sizeof(int));

    if (i < 0)
    {
        printf("%dn", -i);
    }
    else
    {
        printf("%dn", i);
    }
    return 0;
}

The output is as expected:
-2147483648

Size of int: 4 bytes

-2147483648

This indicates that the negation of -2147483648 is still -2147483648.

However, when I tried this with char, the behavior was different:

#include <stdio.h>

int main()
{
    char i = -128;
    printf("%dn", i);
    printf("Size of int: %zu bytesn", sizeof(char));

    if (i < 0)
    {
        printf("%dn", -i);
    }
    else
    {
        printf("%dn", i);
    }
    return 0;
}

The output was:
-128

Size of int: 1 byte

128

I’m using Clang on a Macbook M1.
Why does the negation of -128 for char result in 128 instead of the -128 I was expecting? Is it a compiler problem or a special feature of char?