1. Range on a string iterates over Unicode code points (runes):

    s := "Какая-то строка"
    for _, rune := range s {
        // do something with rune 'К', 'a', ...
    }

2. Converting to []int gives you a slice of runes:

   s := "Какая-то строка"
   runes := []int(s)

   sub := string(runes[:8]) // "Какая-то"

however, slicing a string directly will slice it by byte:

   s[:8] // "Кака"

3. With package utf8 (http://golang.org/pkg/utf8/) you can manipulate runes manually.

While this is all not intuitive (you have to know what does what), I find it rather easy.

Iteration over strings is rune-by-rune in Go. However, (somewhat counter-intuitively) string indexing/slicing is byte-by-byte, so you can't just go "str[0:10]" and get the first 10 code points. Then again, that's true in utf16 also, if I'm not mistaken. But if you want an array of runes instead of a utf8 encoded string, you can just do "[]int(mystring)" and it'll do the conversion for you.

Same in Java, C# and Python. To get the codepoints out or access the nth codepoint you need to iterate over the string if you want your code to be correct and safe.