剪贴板内容在从 Firefox 复制并在 Ubuntu 中使用 Java 读取时会弄乱

在看了相当多的这一点之后，看起来这是Java的一个长期存在的错误（这里甚至更旧的报告）。

看起来X11 Java组件期望剪贴板数据始终采用UTF-8编码，Firefox使用UTF-16编码数据。由于这些假设，Java 通过强制将 UTF-16 解析为 UTF-8 来破坏文本。我试过了，但找不到绕过这个问题的好方法。“text/html”的“text”部分似乎向Java表明，从剪贴板接收的字节应始终首先被解释为文本，然后以各种风格提供。我找不到任何直接的方法从X11访问预先转换的字节数组。

由于直到现在还没有有价值的答案，似乎我们需要一个丑陋的解决方法来使用.非常可惜。O tempora， o mores.我们生活在一个使用Unicode编码比现在更好的时代。UbuntuJavaWindowsUbuntuLinux

我们所知道的已经在答案中陈述过了。因此，我们有一个正确的编码结果，但结果是混乱的。我们知道结果是如何搞砸的。text/plaintext/htmltext/html

因此，我们可以做的是“修复”错误编码的HTML，首先将所有混乱的字符替换为正确的替换字符。然后，我们可以将替换字符替换为从正确编码的纯文本中获得的正确字符。当然，这只能对HTML中可见文本的部分而不是在属性中完成。因为属性内容当然不在纯文本中。

解决方法：

import java.io.*;

import java.awt.Toolkit;
import java.awt.datatransfer.Clipboard;
import java.awt.datatransfer.DataFlavor;

import java.nio.charset.Charset;

public class WorkingWithClipboadDataBytesUTF8 {

 static byte[] repairUTF8HTMLDataBytes(byte[] plainDataBytes, byte[] htmlDataBytes) throws Exception {

  //get all the not ASCII characters from plainDataBytes
  //we need them for replacement later
  String plain = new String(plainDataBytes, Charset.forName("UTF-8"));
  char[] chars = plain.toCharArray();
  StringBuffer unicodeChars = new StringBuffer();
  for (int i = 0; i < chars.length; i++) {
   if (chars[i] > 127) unicodeChars.append(chars[i]);
  }
System.out.println(unicodeChars);

  //ommit the first 6 bytes from htmlDataBytes which are the wrong BOM
  htmlDataBytes = java.util.Arrays.copyOfRange(htmlDataBytes, 6, htmlDataBytes.length);

  //The wrong UTF-8 encoded single bytes which are not replaced by `0xefbfbd` 
  //are coincidentally UTF-16LE if two bytes immediately following each other.
  //So we are "repairing" this accordingly. 
  //Goal: all garbage shall be the replacement character 0xFFFD.

  //replace parts of a surrogate pair with 0xFFFD
  //replace the wrong UFT-8 bytes 0xefbfbd for replacement character with 0xFFFD
  ByteArrayInputStream in = new ByteArrayInputStream(htmlDataBytes);
  ByteArrayOutputStream out = new ByteArrayOutputStream();
  int b = -1;
  int[] btmp = new int[6];
  while ((b = in.read()) != -1) {
   btmp[0] = b;
   btmp[1] = in.read(); //there must always be two bytes because of wron encoding 16 bit Unicode
   if (btmp[0] != 0xef && btmp[1] != 0xef) { // not a replacement character
    if (btmp[1] > 0xd7 && btmp[1] < 0xe0) { // part of a surrogate pair
     out.write(0xFD); out.write(0xFF);
    } else {
     out.write(btmp[0]); out.write(btmp[1]); //two default bytes
    }
   } else { // at least one must be the replacelement 0xefbfbd
    btmp[2] = in.read(); btmp[3] = in.read(); //there must be at least two further bytes
    if (btmp[0] != 0xef && btmp[1] == 0xef && btmp[2] == 0xbf && btmp[3] == 0xbd ||
        btmp[0] == 0xef && btmp[1] == 0xbf && btmp[2] == 0xbd && btmp[3] != 0xef) {
     out.write(0xFD); out.write(0xFF);
    } else if (btmp[0] == 0xef && btmp[1] == 0xbf && btmp[2] == 0xbd && btmp[3] == 0xef) {
     btmp[4] = in.read(); btmp[5] = in.read();
     if (btmp[4] == 0xbf &&  btmp[5] == 0xbd) {
      out.write(0xFD); out.write(0xFF);
     } else {
      throw new Exception("Wrong byte sequence: "
      + String.format("%02X%02X%02X%02X%02X%02X", btmp[0], btmp[1], btmp[2], btmp[3], btmp[4], btmp[5]), 
      new Throwable().fillInStackTrace());
     }
    } else {
     throw new Exception("Wrong byte sequence: " 
      + String.format("%02X%02X%02X%02X%02X%02X", btmp[0], btmp[1], btmp[2], btmp[3], btmp[4], btmp[5]),
      new Throwable().fillInStackTrace());
    }
   }
  }

  htmlDataBytes = out.toByteArray();

  //now get this as UTF_16LE (2 byte for each character, little endian)
  String html = new String(htmlDataBytes, Charset.forName("UTF-16LE"));
System.out.println(html);

  //replace all of the wrongUnicode with the unicodeChars selected from plainDataBytes
  boolean insideTag = false;
  int unicodeCharCount = 0;
  char[] textChars = html.toCharArray();
  StringBuffer newHTML = new StringBuffer();
  for (int i = 0; i < textChars.length; i++) {
   if (textChars[i] == '<') insideTag = true;
   if (textChars[i] == '>') insideTag = false;
   if (!insideTag && textChars[i] > 127) {
    if (unicodeCharCount >= unicodeChars.length()) 
     throw new Exception("Unicode chars count don't match. " 
      + "We got from plain text " + unicodeChars.length() + " chars. Text until now:\n" + newHTML,
      new Throwable().fillInStackTrace());

    newHTML.append(unicodeChars.charAt(unicodeCharCount++));
   } else {
    newHTML.append(textChars[i]);
   }
  }

  html = newHTML.toString();
System.out.println(html);

  return html.getBytes("UTF-8");

 }

 static void doSomethingWithUTF8BytesFromClipboard(byte[] plainDataBytes, byte[] htmlDataBytes) throws Exception {

  if (plainDataBytes != null && htmlDataBytes != null) {

   String fileName; 
   OutputStream fileOut;

   fileName = "ResultPlainText.txt";
   fileOut = new FileOutputStream(fileName);
   fileOut.write(plainDataBytes, 0, plainDataBytes.length);
   fileOut.close();

   fileName = "ResultHTMLRaw.txt";
   fileOut = new FileOutputStream(fileName);
   fileOut.write(htmlDataBytes, 0, htmlDataBytes.length);
   fileOut.close();

   //do we have wrong encoded UTF-8 in htmlDataBytes?
   if (htmlDataBytes[0] == (byte)0xef && htmlDataBytes[1] == (byte)0xbf && htmlDataBytes[2] == (byte)0xbd 
    && htmlDataBytes[3] == (byte)0xef && htmlDataBytes[4] == (byte)0xbf && htmlDataBytes[5] == (byte)0xbd) {
    //try repair the UTF-8 HTML data bytes
    htmlDataBytes = repairUTF8HTMLDataBytes(plainDataBytes, htmlDataBytes);
          //do we have additional 0x00 byte at the end?
   } else if (htmlDataBytes[htmlDataBytes.length-1] == (byte)0x00) {
    //do repair this
    htmlDataBytes = java.util.Arrays.copyOf(htmlDataBytes, htmlDataBytes.length-1);
   }

   fileName = "ResultHTML.txt";
   fileOut = new FileOutputStream(fileName);
   fileOut.write(htmlDataBytes, 0, htmlDataBytes.length);
   fileOut.close();

  }

 }

 public static void main(String[] args) throws Exception {

  Clipboard clipboard = Toolkit.getDefaultToolkit().getSystemClipboard();

  byte[] htmlDataBytes = null;
  byte[] plainDataBytes = null;

  for (DataFlavor dataFlavor : clipboard.getAvailableDataFlavors()) {

   String mimeType = dataFlavor.getHumanPresentableName();

   if ("text/html".equalsIgnoreCase(mimeType)) {
    String paramClass = dataFlavor.getParameter("class");
    if ("[B".equals(paramClass)) {
     String paramCharset = dataFlavor.getParameter("charset");
     if (paramCharset != null  && "UTF-8".equalsIgnoreCase(paramCharset)) {

      htmlDataBytes = (byte[])clipboard.getData(dataFlavor);

     }
    } //else if("java.io.InputStream".equals(paramClass)) ...

   } else if ("text/plain".equalsIgnoreCase(mimeType)) {
    String paramClass = dataFlavor.getParameter("class");
    if ("[B".equals(paramClass)) {
     String paramCharset = dataFlavor.getParameter("charset");
     if (paramCharset != null  && "UTF-8".equalsIgnoreCase(paramCharset)) {

      plainDataBytes = (byte[])clipboard.getData(dataFlavor);

     }
    } //else if("java.io.InputStream".equals(paramClass)) ...
   }
  }

  doSomethingWithUTF8BytesFromClipboard(plainDataBytes, htmlDataBytes);

 }

}