End-of-Line Characters Supported in XML 1.1

XML Tutorials - Herong's Tutorial Examples

∟End-of-Line Characters Supported in XML 1.1

This section provides a tutorial example showing that two more end-of-line characters, #x85 and #x2028, are supported in XML 1.1.

Next, let's prove that end-of-line characters are handled differently in XML 1.1 than XML 1.0. Since end-of-line characters are not easy to be presented as text files, I created this test program, EndOfLineXml.java:

/* EndOfLineXml.java
 * Copyright (c) 2002-2018 HerongYang.com. All Rights Reserved.
 */
import java.io.*;
import java.math.*;
import javax.xml.parsers.*;
import org.w3c.dom.*;
class EndOfLineXml {
   public static void main(String[] args) {
      try {
         String ver = args[0];
         char u000D = 0x000D;
         char u000A = 0x000A;
         char u0085 = 0x0085;
         char u2028 = 0x2028;
         String xmlString =
            "<?xml version=\""+ver+"\" encoding=\"UTF-16BE\"?><ul>"
            +"<li>1111"+u000D+u000A+"1111</li>"
            +"<li>2222"+u000D+u0085+"2222</li>"
            +"<li>3333"+u0085+"03333</li>"
            +"<li>4444"+u2028+"04444</li>"
            +"<li>5555"+u000D+"05555</li>"
            +"<li>6666"+u000A+"06666</li>"
            +"</ul>";
         File xmlFile = new File(args[1]);
         FileOutputStream fos = new FileOutputStream(xmlFile);
         OutputStreamWriter osw =
            new OutputStreamWriter(fos,"UTF-16BE");
         osw.write(xmlString);
         osw.close();

         DocumentBuilderFactory fct
            = DocumentBuilderFactory.newInstance();
         DocumentBuilder bld = fct.newDocumentBuilder();
         Document doc = bld.parse(xmlFile);
         dumpNode(doc, "");

      } catch (Exception e) {
         System.out.println(e.toString());
      }
   }
   static void dumpNode(Node n, String p) throws Exception {
      NodeList l = n.getChildNodes();
      NamedNodeMap m = n.getAttributes();
      int ml = -1;
      if (m!=null) ml = m.getLength();
      System.out.println(p+n.getNodeName()+": "+n.getNodeType()+", "
         +l.getLength()+", "+ml+", "+str10ToHex(n.getNodeValue()));
      for (int i=0; i<ml; i++) {
         Node c = m.item(i);
         dumpNode(c,p+" |-");
      }
      for (int i=0; i<l.getLength(); i++) {
         Node c = l.item(i);
         dumpNode(c,p+" ");
      }
   }
   static String str10ToHex(String str) throws Exception {
      if (str!=null) {
         return String.format("%040X",
            new BigInteger(1,str.getBytes("UTF-16BE")));
      } else {
         return "NULL";
      }
   }
}

Some notes on EndOfLineXml.java:

The XML document is created inside the program as a String object, instead of loaded from an input file.
The XML document has 6 "li" elements with different end-of-line character combinations.
You should not use the \u format for end-of-line characters like \u000D or \u000A in Java source code. The compiler will treat it as end-of-line character and break a quoted string literal into 2 lines causing compiler errors.
In order to see how end-of-line characters are parsed as element values, The program dumps element values are double-byte character hex values.
The BigInteger class is used to convert a string with up to 10 characters into 40 Hexadecimal digits. This is a quicker way to convert string into Hexadecimal digits without using a loop.

Let's try XML 1.0 first with JDK 10 and 1.8:

herong> java EndOfLineXml 1.0 end-of-line-1-0.xml

#document: 9, 1, -1, NULL
 ul: 1, 6, 0, NULL
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 00000031003100310031000A0031003100310031
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0032003200320032000A00850032003200320032
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0033003300330033008500300033003300330033
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0034003400340034202800300034003400340034
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0035003500350035000A00300035003500350035
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0036003600360036000A00300036003600360036

The output confirms that end-of-line characters are handled in XML 1.0 as below:

#x00D#x000A is replaced by #x0A, see the node value in line 1. Note that the str10ToHex() method padded '0000' on the left, because the node value only has 9 characters.
#x0085 is kept as is, see the node value in line 2 and 3
#x2028 is kept as is, see the node value in line 4.
#x000D is replaced by #x0A, see the node value in line 5.

Here is the output of XML 1.1:

herong> java EndOfLineXml 1.1 end-of-line-1-1.xml

#document: 9, 1, -1, NULL
 ul: 1, 6, 0, NULL
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 00000031003100310031000A0031003100310031
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 00000032003200320032000A0032003200320032
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0033003300330033000A00300033003300330033
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0034003400340034000A00300034003400340034
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0035003500350035000A00300035003500350035
  li: 1, 1, 0, NULL
   #text: 3, 0, -1, 0036003600360036000A00300036003600360036

The output confirms that end-of-line characters are handled in XML 1.1 as below:

#x00D#x000A is replaced by #x0A, see the node value in line 1. Note that the str10ToHex() method padded '0000' on the left, because the node value only has 9 characters.
#x00D#x0085 is replaced by #x0A, see the node value in line 2.
#x0085 is replaced by #x0A, see the node value in line 3.
#x2028 is replaced by #x0A, see the node value in line 4
#x000D is replaced by #x0A, see the node value in line 5.