Implement missing string ops

[quasilyte]

Operations:

s[index] -> nth string byte
len(s) -> string byte length
for K, V := range s 
  -> K is byteN
  -> V is rune

Possible representations:

• unibyte strings. Easier to implement s[index], harder to implement for/range.
• multibyte strings. Easier to implement for/range, harder to implement s[index]
• vector of chars. Like unibyte strings, but can be less efficient

[quasilyte]

Useful blog entry: http://nullprogram.com/blog/2014/06/13/
Asked a question at https://emacs.stackexchange.com/questions/33970/get-multibyte-string-byte

[quasilyte]

Can affect #26.

[quasilyte]

Still missing:

• []byte(string) conversion
• for/range over string