How should I deal with linebreaks in strings I want to marshal in Java to XML?

Question

How should I deal with linebreaks in strings I want to marshal to XML?

I am having difficulty using Java and JAXB to handle putting strings in XML files that have linefeeds in them. The data is being pulled from a database with the actual line feed characters in them.

Foo <LF>
bar

Or an additional example:

Foo\r\n\r\nBar

Yields:

Foo

Bar

If I just marshal this data into XML, I get literal line feed characters in the output. This is apparently against XML standards where the characters should be encoded to 
. Ie in the XML file output I should see:

Foo 
bar

But if I try and do this manually, I end up with my ampersand getting encoded!

Foo 
bar

This is pretty ironic because the process which is apparently supposed to encode the linebreaks in the first place and is not, is foiling my attempts to encode it manually.

Answer 1

Below is an example of JAXB's default behaviour regarding \n and \r:

Java Model (Root)

import javax.xml.bind.annotation.XmlRootElement;

@XmlRootElement
public class Root {

    private String foo;
    private String bar;

    public String getFoo() {
        return foo;
    }

    public void setFoo(String foo) {
        this.foo = foo;
    }

    public String getBar() {
        return bar;
    }

    public void setBar(String bar) {
        this.bar = bar;
    }

}

Demo Code

import javax.xml.bind.*;

public class Demo {

    public static void main(String[] args) throws Exception {
        JAXBContext jc = JAXBContext.newInstance(Root.class);

        Root root = new Root();
        root.setFoo("Hello\rWorld");
        root.setBar("Hello\nWorld");

        Marshaller marshaller = jc.createMarshaller();
        marshaller.marshal(root, System.out);
    }

}

Output

<?xml version="1.0" encoding="UTF-8" standalone="yes"?><root><bar>Hello
World</bar><foo>Hello&#xD;World</foo></root>

---

UPDATE

Below are some additional details based on some investigation that I did.

Common to All JAXB (JSR-222) Implementations

• If you are marshalling to an XMLStreamWriter or XMLEventWriter directly (via Marshaller) or indirectly (via potentially a JAX-RS or JAX-WS provider) then the escaping will be based on the StAX implementation. Woodstox appears to escape things correctly, but the StAX implementation in the JDK I'm using did not.

EclipseLink JAXB (MOXy)

• There is a bug in MOXy related to escaping \r that I am currently in the process of fixing (see: http://bugs.eclipse.org/414608)

JAXB Reference Implementation

• The JAXB reference implementation will properly escape '\r' when marshalling to an OutputStream, but not to a Writer atleast in the JDK I'm using.