From wikipedia article we know that a GBK encoding character include one or two bytes. If the first character is within the range 0x00~0x7f, it presents a character in ascii encoding table, we can convert it to ascii and do whatever we want to the converted character.

Otherwise, the first character is not within the bound 0x00~0x7f, which means it has the most significant bit set, then the first character and the second character together denote a GBK encoding character. We can construct a valid and completed GBK encoding character and do whatever we want to this character from these two bytes.

To show how it works, follow the steps below.

  1. obtain a GBK encoding file

    I use the below command to create an GBK encoding file

    $ vim myfile-utf8
# type anything GBK character
# quit vim
$ iconv -f utf8 -t gbk myfile-utf8 > myfile-gbk

  2. Now we can use the simple rule described above to parse the GBK encoding file

    fun main(args: Array<String>) {
   
    val inputStream = FileInputStream(File("/home/mutter/myfile-gbk"))
   
    val byteArray = ByteArray(inputStream.available()).apply { inputStream.read(this) }
   
    val iterator = byteArray.iterator()
   
    val stringBuilder = StringBuilder()
   
    while (true) {
        if (iterator.hasNext()) {
            val next = iterator.nextByte()
   
            // it's two bytes gbk encoding if the most significant bit(MSB) is 1
            if (next.and(0x80.toByte()) == 0x80.toByte()) {
                if (iterator.hasNext()) {
                    val nn = iterator.nextByte()
                    val gbkString = String(byteArrayOf(next, nn),
                            Charset.forName("GBK"))
                    stringBuilder.append(gbkString)
                } else {
                    break
                }
            } else {
                stringBuilder.append(String(byteArrayOf(next), Charsets.US_ASCII))
            }
        } else {
            break
        }
    }
    inputStream.close()
   
    println(stringBuilder.toString())
}