UTF16、UTF8 和字节顺序标记由Unicode 联盟定义：UTF-16 FAQ、UTF-8 FAQ和字节顺序标记 (BOM) FAQ。
问题 4802：bufio：读行太麻烦
在 Go 中从文件中读取行太麻烦了。
人们经常被 bufio.Reader.ReadLine 所吸引，因为它的名字，但是它有一个奇怪的签名，返回 (line []byte, isPrefix bool, err error)，并且需要大量的工作。
ReadSlice 和 ReadString 需要一个分隔字节，它几乎总是明显且难看的 '\n'，也可以返回一行和一个 EOF
修订： f685026a2d38
bufio：新的扫描仪界面
添加一个新的、简单的界面，用于扫描（可能是文本）数据，基于一种名为 Scanner 的新类型。它有自己的内部缓冲，因此即使没有注入 bufio.Reader 也应该是合理有效的。输入的格式由“拆分函数”定义，默认情况下拆分为行。
go1.1beta1 发布
您可以从通常的位置下载二进制和源代码分发：
https ://code.google.com/p/go/downloads/list?q=go1.1beta1
这是一个使用 Unicode 规则将 UTF16 文本文件行转换为 Go UTF8 编码字符串的程序。代码已被修改以利用bufio.ScannerGo 1.1 中的新接口。
package main
import (
"bufio"
"bytes"
"encoding/binary"
"fmt"
"os"
"runtime"
"unicode/utf16"
"unicode/utf8"
)
// UTF16BytesToString converts UTF-16 encoded bytes, in big or little endian byte order,
// to a UTF-8 encoded string.
func UTF16BytesToString(b []byte, o binary.ByteOrder) string {
utf := make([]uint16, (len(b)+(2-1))/2)
for i := 0; i+(2-1) < len(b); i += 2 {
utf[i/2] = o.Uint16(b[i:])
}
if len(b)/2 < len(utf) {
utf[len(utf)-1] = utf8.RuneError
}
return string(utf16.Decode(utf))
}
// UTF-16 endian byte order
const (
unknownEndian = iota
bigEndian
littleEndian
)
// dropCREndian drops a terminal \r from the endian data.
func dropCREndian(data []byte, t1, t2 byte) []byte {
if len(data) > 1 {
if data[len(data)-2] == t1 && data[len(data)-1] == t2 {
return data[0 : len(data)-2]
}
}
return data
}
// dropCRBE drops a terminal \r from the big endian data.
func dropCRBE(data []byte) []byte {
return dropCREndian(data, '\x00', '\r')
}
// dropCRLE drops a terminal \r from the little endian data.
func dropCRLE(data []byte) []byte {
return dropCREndian(data, '\r', '\x00')
}
// dropCR drops a terminal \r from the data.
func dropCR(data []byte) ([]byte, int) {
var endian = unknownEndian
switch ld := len(data); {
case ld != len(dropCRLE(data)):
endian = littleEndian
case ld != len(dropCRBE(data)):
endian = bigEndian
}
return data, endian
}
// SplitFunc is a split function for a Scanner that returns each line of
// text, stripped of any trailing end-of-line marker. The returned line may
// be empty. The end-of-line marker is one optional carriage return followed
// by one mandatory newline. In regular expression notation, it is `\r?\n`.
// The last non-empty line of input will be returned even if it has no
// newline.
func ScanUTF16LinesFunc(byteOrder binary.ByteOrder) (bufio.SplitFunc, func() binary.ByteOrder) {
// Function closure variables
var endian = unknownEndian
switch byteOrder {
case binary.BigEndian:
endian = bigEndian
case binary.LittleEndian:
endian = littleEndian
}
const bom = 0xFEFF
var checkBOM bool = endian == unknownEndian
// Scanner split function
splitFunc := func(data []byte, atEOF bool) (advance int, token []byte, err error) {
if atEOF && len(data) == 0 {
return 0, nil, nil
}
if checkBOM {
checkBOM = false
if len(data) > 1 {
switch uint16(bom) {
case uint16(data[0])<<8 | uint16(data[1]):
endian = bigEndian
return 2, nil, nil
case uint16(data[1])<<8 | uint16(data[0]):
endian = littleEndian
return 2, nil, nil
}
}
}
// Scan for newline-terminated lines.
i := 0
for {
j := bytes.IndexByte(data[i:], '\n')
if j < 0 {
break
}
i += j
switch e := i % 2; e {
case 1: // UTF-16BE
if endian != littleEndian {
if i > 1 {
if data[i-1] == '\x00' {
endian = bigEndian
// We have a full newline-terminated line.
return i + 1, dropCRBE(data[0 : i-1]), nil
}
}
}
case 0: // UTF-16LE
if endian != bigEndian {
if i+1 < len(data) {
i++
if data[i] == '\x00' {
endian = littleEndian
// We have a full newline-terminated line.
return i + 1, dropCRLE(data[0 : i-1]), nil
}
}
}
}
i++
}
// If we're at EOF, we have a final, non-terminated line. Return it.
if atEOF {
// drop CR.
advance = len(data)
switch endian {
case bigEndian:
data = dropCRBE(data)
case littleEndian:
data = dropCRLE(data)
default:
data, endian = dropCR(data)
}
if endian == unknownEndian {
if runtime.GOOS == "windows" {
endian = littleEndian
} else {
endian = bigEndian
}
}
return advance, data, nil
}
// Request more data.
return 0, nil, nil
}
// Endian byte order function
orderFunc := func() (byteOrder binary.ByteOrder) {
switch endian {
case bigEndian:
byteOrder = binary.BigEndian
case littleEndian:
byteOrder = binary.LittleEndian
}
return byteOrder
}
return splitFunc, orderFunc
}
func main() {
file, err := os.Open("utf16.le.txt")
if err != nil {
fmt.Println(err)
os.Exit(1)
}
defer file.Close()
fmt.Println(file.Name())
rdr := bufio.NewReader(file)
scanner := bufio.NewScanner(rdr)
var bo binary.ByteOrder // unknown, infer from data
// bo = binary.LittleEndian // windows
splitFunc, orderFunc := ScanUTF16LinesFunc(bo)
scanner.Split(splitFunc)
for scanner.Scan() {
b := scanner.Bytes()
s := UTF16BytesToString(b, orderFunc())
fmt.Println(len(s), s)
fmt.Println(len(b), b)
}
fmt.Println(orderFunc())
if err := scanner.Err(); err != nil {
fmt.Println(err)
}
}
输出：
utf16.le.txt
15 "Hello, 世界"
22 [34 0 72 0 101 0 108 0 108 0 111 0 44 0 32 0 22 78 76 117 34 0]
0
0 []
15 "Hello, 世界"
22 [34 0 72 0 101 0 108 0 108 0 111 0 44 0 32 0 22 78 76 117 34 0]
LittleEndian
utf16.be.txt
15 "Hello, 世界"
22 [0 34 0 72 0 101 0 108 0 108 0 111 0 44 0 32 78 22 117 76 0 34]
0
0 []
15 "Hello, 世界"
22 [0 34 0 72 0 101 0 108 0 108 0 111 0 44 0 32 78 22 117 76 0 34]
BigEndian