utf8 包源码解读

utf8 包用到的常量定义如下：

const (
 RuneError = '\uFFFD'     // the "error" Rune or "Unicode replacement character"
 RuneSelf  = 0x80         // characters below Runeself are represented as themselves in a single byte.
 MaxRune   = '\U0010FFFF' // Maximum valid Unicode code point.
 UTFMax    = 4            // maximum number of bytes of a UTF-8 encoded Unicode character.
)

// Code points in the surrogate range are not valid for UTF-8.
const (
 surrogateMin = 0xD800
 surrogateMax = 0xDFFF
)

const (
 t1 = 0x00 // 0000 0000
 tx = 0x80 // 1000 0000
 t2 = 0xC0 // 1100 0000
 t3 = 0xE0 // 1110 0000
 t4 = 0xF0 // 1111 0000
 t5 = 0xF8 // 1111 1000

 maskx = 0x3F // 0011 1111
 mask2 = 0x1F // 0001 1111
 mask3 = 0x0F // 0000 1111
 mask4 = 0x07 // 0000 0111

 rune1Max = 1<<7 - 1
 rune2Max = 1<<11 - 1
 rune3Max = 1<<16 - 1
)

导出函数解读：

1、DecodeLastRune(p []byte)

    a、 函数功能说明：返回p中最后一个rune和它占用的字节数，若果p最后一个rune是非法字符的话，返回\uFFFD 和0 .

         【其中\uFFFD表示：当从某语言向Unicode转化时，如果在某语言中没有该字符，得到的将是Unicode的代码“\uffffd”（“\u”表示是Unicode编码，）】

   b、 函数源码：

         func DecodeLastRune(p []byte) (r rune, size int) {
         end := len(p)
         if end == 0 {
         return RuneError, 0
        }
       start := end - 1
        r = rune(p[start])
 if r < RuneSelf {
  return r, 1
 }
 // guard against O(n^2) behavior when traversing
 // backwards through strings with long sequences of
 // invalid UTF-8.
 lim := end - UTFMax
 if lim < 0 {
  lim = 0
 }
 for start--; start >= lim; start-- {
  if RuneStart(p[start]) {
   break
  }
 }
 if start < 0 {
  start = 0
 }
 r, size = DecodeRune(p[start:end])
 if start+size != end {
  return RuneError, 1
 }
 return r, size
}