Get *uint8 pointer to a first character in a string

Question

I am trying to port a hash function from C to Go using unsafe.Pointer. I need to get the pointer to the first character of the key string.

I tried &(hello[0]) and &hello[0] and failed.

You most likely have an xy problem. If you need to iterate through the bytes in a string, there are nice go-ways to do it than without using a pointer. — nos, Feb 28 '17 at 08:41
@nos unfortunately not, there is a heavy pointer arithmetic that deals with the caracters in a string as bytes, words and dwords. — exebook, Feb 28 '17 at 08:45
You cannot deal with "characters" in a string as bytes, at least not in the Unicode world. — Volker, Feb 28 '17 at 08:54
@exebook Then that is of vital importance to your question, and you should mention explicitly what you need to do. There are ways to do what you need, but your question is too open ended. It's better that you give concrete problems or code snippets, so you can get concrete answers . You still might need to re-write the code and not do a 1:1 porting from C though. — nos, Feb 28 '17 at 11:45
@nos thank you, created a specific question http://stackoverflow.com/questions/42507846/porting-meiyan-hash-function-to-go — exebook, Feb 28 '17 at 11:47

score 2 · Answer 1 · edited May 23 '17 at 12:09

string values in Go are immutable, and therefore not addressable; so you cannot take the address of its runes or bytes. If you could, you could modify the pointed value and thus modify the content of the string value. In Go there is also no pointer arithmetic.

What you can do is simply index the string, indexing a string indexes its bytes (in UTF-8 encoded form, this is how Go stores string values in memory). You can "mimic" the pointer arithmetic with "index arithmetic".

See this example:

s := "Hello World"
i := 0
fmt.Println(s[i], string(s[i]))
i += 4
fmt.Println(s[i], string(s[i]))

Output (try it on the Go Playground):

72 H
111 o

You may also convert the string to a []byte or []rune, should you have the need to do so (e.g. if you need to modify the bytes), and you can work with that. Slices are also addressable. Note that converting a string value to the mentioned slice types will make a copy of the string content (to preserve the immutability of the string).

While it's the rune type that represents a Unicode character, if you're trying to reproduce a hash algorithm, you most likely want to work with bytes, so convert it to a byte slice (which is the UTF-8 encoded byte sequence of the string):

s := "Hello World"
b := []byte(s)
// Here you can work with b which is addressable

Note that in Go, the types byte and uint8 are "identical", byte is an alias for uint8, you can also write:

s := "Hello World"
b := []uint8(s)

score 0 · Answer 2 · answered Feb 28 '17 at 08:54

In golang string it's a massive of bytes, need convert string to rune

package main

import (
    "fmt"
)

func test(r *rune) {
    *r = rune('L')
}

func main() {
    s := []rune("Hello")
    fmt.Println(string(s))
    fmt.Println(&s[0])
    test(&s[0])
    fmt.Println(string(s))
}

Example playground golang

Get *uint8 pointer to a first character in a string

2 Answers2