迭代边界与数据类型相同

Question

A function I have takes min, max uint16 parameters and at some point iterates over the numeric range. However, if max happens to be 2^16-1 (and it is a valid use case), then overflow breaks the loop logic. Here is an example code demonstrating the problem with uint8:

package main

import "fmt"

func iter(min, max uint8) {
    for i := min; i <= max; i++ {
        fmt.Printf("%d, ", i)
    }
}

func main() {
    iter(0, 255)
}

As you can see, the program never ends. A similar question was asked at another question but the solution exactly exhibits the same problem I have.

My thinking for now is to convert the loop variable to uint32, similar to this:

package main

import "fmt"

func iter(min, max uint8) {
    for i := uint16(min); i <= uint16(max); i++ {
        fmt.Printf("%d, ", i)
    }
}

func main() {
    iter(0, 255)
}

However, this seems to be a clumsy solution, which is not going to work for uint64 or whatever biggest uintN type. Feels like I am missing something very basic. Guidance?

I am aware of Brad Fitz's Iter solution, but it seems to add unneeded overhead. Is that true as well?

Answer 1

For example, for uint8,

package main

import "fmt"

func iter(min, max uint8) {
    {
        min, max := uint(min), uint(max)
        for i := min; i <= max; i++ {
            fmt.Printf("%d, ", i)
        }
    }
}

func main() {
    iter(0, 255)
}

For uint64,

package main

import "fmt"

func iter(min, max uint64) {
    for i := min; i <= max; i++ {
        fmt.Printf("%d, ", i)
        if i == max {
            break
        }
    }
}

func main() {
    iter(^uint64(0)-2, ^uint64(0))
}

Output:

18446744073709551613, 18446744073709551614, 18446744073709551615

Addendum:

Here's my version of Dave C's suggestion.

package main

import "fmt"

func iter(min, max uint64) {
    for i, next := min, min <= max; next; i, next = i+1, i < max {
        fmt.Printf("%#016[1]x ", i)
    }
    fmt.Println()
}

func main() {
    const maxUint64 = ^uint64(0)
    iter(0, 3)
    iter(10, 9)
    iter(maxUint64-2, maxUint64)
}

Output:

0x0000000000000000 0x0000000000000001 0x0000000000000002 0x0000000000000003 
0xfffffffffffffffd 0xfffffffffffffffe 0xffffffffffffffff