Home/ Questions/Q 4036426
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-20T12:13:18+00:00

I have been trying to use XStreamMarshaller to generate XML output in my Java

  • 0

I have been trying to use XStreamMarshaller to generate XML output in my Java Spring project. The XML I am generating has CDATA values in the element text. I am manually creating this CDATA text in the command object like this: 

 f.setText("<![CDATA[cdata-text]]>");

The XStreamMarshaller generated the element(text-data below is an alias) as: 

<text-data><![CDATA[cdata-text]]></text-data>

The above XML display is as expected (Please ignore the back slash in the above element name: forum formatting). But when I do a View Source on the XML output generated I see this for the element: <text-data>&lt;![CDATA[cdata-text]]&gt;</text-data>.

Issue:

As you can see the less than and greater than characters have been replaced by &lt; and &gt; in the View Source. I need my client to read the source and identify CDATA section from the XML output which it will not in the above scenario.

Is there a way I can get the XStreamMarshaller to escape special characters in the text I provided?

I have set the encoding of the Marshaller to ISO-8859-1 but that does not work either. If the above cannot be done by XStreamMarshaller can you please suggest alternate marshallers/unmarshallers that can do this for me?

// Displaying my XML and View Source as suggested by Paŭlo Ebermann below:

XML View (as displayed in IE):

An invalid character was found in text content. Error processing resource 'http://localhost:8080/file-service-framework/fil...

Los t

View Source:

<service id="file-text"><text-data>&lt;![CDATA[
Los túneles a través de las montañas hacen más fácil viajar por carretera.
]]&gt;</text-data></service>

Thanks you very much.

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    Generating CDATA sections is the task of your XML-generating library, not of its client. So you should simply have to write

    f.setText("cdata-text");

    and then the library can decide whether to use <![CDATA[…]]> or the &lt;-escaping for its contents. It should make no difference for the receiver.

    Edit:

    Looking at your output, it looks right (apart from the CDATA) – here you must work on your input, as said.

    If IE throws an error here, most probably you don’t have declared the right encoding.

    I don’t really know much about the Spring framework, but the encoding used by the Marshaller should be the same encoding as the encoding sent in either the HTTP header (Content-Type: ... ;charset=...) or the <?xml version="1.0" encoding="..." ?> XML prologue (these two should not differ, too).

    I would recommend UTF-8 as encoding everywhere, as this can represent all characters, not only the Latin-1 ones.

    • 0