You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

90 lines
1.9 KiB
Go

3 years ago
package startext
import "golang.org/x/text/encoding/simplifiedchinese"
func IsUtf8(data []byte) bool {
return isUtf8(data)
}
func IsGBK(data []byte) bool {
return (!isUtf8(data)) && isGBK(data)
}
func isGBK(data []byte) bool {
length := len(data)
var i int = 0
for i < length {
if data[i] <= 0x7f {
//编码0~127,只有一个字节的编码兼容ASCII码
i++
continue
} else {
//大于127的使用双字节编码落在gbk编码范围内的字符
if data[i] >= 0x81 &&
data[i] <= 0xfe &&
data[i+1] >= 0x40 &&
data[i+1] <= 0xfe &&
data[i+1] != 0x7f {
i += 2
continue
} else {
return false
}
}
}
return true
}
//UTF-8编码格式的判断
func preNUm(data byte) int {
var mask byte = 0x80
var num int = 0
//8bit中首个0bit前有多少个1bits
for i := 0; i < 8; i++ {
if (data & mask) == mask {
num++
mask = mask >> 1
} else {
break
}
}
return num
}
func isUtf8(data []byte) bool {
i := 0
for i < len(data) {
if (data[i] & 0x80) == 0x00 {
// 0XXX_XXXX
i++
continue
} else if num := preNUm(data[i]); num > 2 {
// 110X_XXXX 10XX_XXXX
// 1110_XXXX 10XX_XXXX 10XX_XXXX
// 1111_0XXX 10XX_XXXX 10XX_XXXX 10XX_XXXX
// 1111_10XX 10XX_XXXX 10XX_XXXX 10XX_XXXX 10XX_XXXX
// 1111_110X 10XX_XXXX 10XX_XXXX 10XX_XXXX 10XX_XXXX 10XX_XXXX
// preNUm() 返回首个字节的8个bits中首个0bit前面1bit的个数该数量也是该字符所使用的字节数
i++
for j := 0; j < num-1; j++ {
//判断后面的 num - 1 个字节是不是都是10开头
if (data[i] & 0xc0) != 0x80 {
return false
}
i++
}
} else {
//其他情况说明不是utf-8
return false
}
}
return true
}
func GBK2UTF8(data []byte) ([]byte,error) {
return simplifiedchinese.GBK.NewDecoder().Bytes(data)
}
func UTF82GBK(data []byte) ([]byte,error) {
return simplifiedchinese.GBK.NewEncoder().Bytes(data)
}