org.apache.xml.serialize.XMLSerializer problem with UTF-8

[Jim Cobban]

I must be missing something.

I am using org.apache.xml.serialize.XMLSerializer to save a DOM but I am not
getting non-basic characters converted to UTF-8.

I create Text nodes in the DOM by, for example:

Document doc;
JTextArea textPrompt;
Text newTextNode;
Element descElt;
....
newTextNode = doc.createTextNode(textPrompt.getText());
descElt.appendChild(newTextNode);

The code to serialize the DOM is:

private void saveXml(Document document)
{
// rename the existing layout file
new File(fileName).renameTo(new File(fileName + "~"));
// write the document out
OutputFormat format = new OutputFormat(document);
format.setIndenting(true);
format.setLineWidth(0);
format.setPreserveSpace(true);
try {
XMLSerializer serializer;
serializer = new XMLSerializer (
new FileWriter(fileName),
format);
serializer.asDOMSerializer();
serializer.serialize(document);
}
catch (IOException ioe)
{
....
}
}

If I enter a character such as e' (e with acute accent) into the JTextArea
and I look at the XML file using a non-UTF-8-aware editor I see that the e'
has been inserted as a single byte, not as the 2 character UTF-8 escaped
value. If I subsequently try to read the XML file using XERCES it blows up
because of the invalid escape sequence.

How do I get a valid serialization of this DOM into XML using UTF-8?


--
Jim Cobban (e-mail address removed)
34 Palomino Dr.
Kanata, ON, CANADA
K2M 1M1
+1-613-592-9438

 

[Soren Kuula]

I must be missing something.

XMLSerializer serializer;
serializer = new XMLSerializer (
new FileWriter(fileName),
format);
serializer.asDOMSerializer();
If I enter a character such as e' (e with acute accent) into the JTextArea
and I look at the XML file using a non-UTF-8-aware editor I see that the e'
has been inserted as a single byte, not as the 2 character UTF-8 escaped
value. If I subsequently try to read the XML file using XERCES it blows up
because of the invalid escape sequence.

How do I get a valid serialization of this DOM into XML using UTF-8?

As far as I know it is the Writer responsible for the encoding.

From FileWriter API doc:

public class FileWriter
extends OutputStreamWriter

Convenience class for writing character files. The constructors of this
class assume that the default character encoding and the default
byte-buffer size are acceptable. To specify these values yourself,
construct an OutputStreamWriter on a FileOutputStream.


- try that.

Soren

 

[Jim Cobban]

As far as I know it is the Writer responsible for the encoding.

From FileWriter API doc:

public class FileWriter
extends OutputStreamWriter

Convenience class for writing character files. The constructors of this
class assume that the default character encoding and the default
byte-buffer size are acceptable. To specify these values yourself,
construct an OutputStreamWriter on a FileOutputStream.

Thank you.

The problem was that I copied the code from one of the examples that came
with Xerces. It was that example which constructed the default FileWriter.
Since their is a version of the XMLSerializer constructor which takes an
OutpuStream and internally constructs a Writer with the correct "utf-8"
encoding, that is the form of the constructor which I needed to use. I
should have read the documentation in more detail rather than trusting that
the example had been written correctly.