What does this integer puzzle print?

var x int8 = 127
x++
fmt.Println(x)        // ?

var u uint8 = 0
u--
fmt.Println(u)        // ?

m := int8(-128)
fmt.Println(-m)       // ?

fmt.Println(-7/2, -7%2) // ?

// const big = 1 << 62
// fmt.Println(big * 4) // ?

Output: -128, 255, -128, -3 -1. The commented-out line is a compile error (cannot use big * 4 (untyped int constant 18446744073709551616) as int value in argument to fmt.Println (overflows)).

Explanation: at runtime, signed and unsigned integer arithmetic wraps using two's complement. There is no panic and no error. Negating the minimum value gives the minimum value back, so a naive abs is wrong for math.MinInt64. Integer division truncates toward zero, and % takes the sign of the dividend, so use ((a % n) + n) % n for a true modulo. Constants are arbitrary-precision and checked at compile time, so overflow shows up only when the constant is converted to a typed value.

A real bug: len(s)-1 with an unsigned index type when s is empty, or int32(bigInt64) silently truncating an ID. Check with math/bits.Add64 or with explicit bounds checks.