com.sun.org.apache.xerces.internal.impl.io.MalformedByteSequenceException解析POST中的XML（Java Servlet）

Question

I am developing a Servlet which uses a POST to send the information. I am getting an exception depending where the servlet is deployed. If I deploy on a local JBoss EAP 7 it works like a charm. But, if I deploy into a remote server (JBoss EAP 7 too) i am getting the following error:

CODE

final InputStream is = new ByteArrayInputStream(xml);
doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(is);

ERROR

<message>
Failed to load XML: com.sun.org.apache.xerces.internal.impl.io.MalformedByteSequenceException: Byte no válido 2 de la secuencia UTF-8 de 4 bytes

~ôJ뢺ÞjX³jwh•ªkyø«™§¬‚†Ú~*æiÊ+x„ÅÇ©¶*'ºè­©`zw«j¹Z~*æip(Ú½©Z </message>

For more information the servlet has a jboss-deployment-structure.xml

<jboss-deployment-structure xmlns="urn:jboss:deployment-structure:1.1">
  <deployment>
    <dependencies>
      <module name="javax.api"/>
      <module name="org.apache.santuario.xmlsec"/>
      <module name="org.apache.xerces" />
      <system export="true">
        <paths> 
          <path name="com/sun/org/apache/xerces/internal/dom"/>
        </paths>
      </system>
    </dependencies>
  </deployment>
</jboss-deployment-structure>

Where the xerces library is implicated. I don´t understand if my problem is related to that xml , or I should change some encoding in the server/servlet code. Thnaks in advance.

Answer 1

Badulake,
What you are hitting looks like a encoding issue. When you don't specify encoding while converting a byte array to Java string, JVM uses the default encoding of the platform. In most cases the default encoding is utf-8 and that works great. However I have seen many OS installations where the encoding is simply US-ASCII.

If this is your issue, you can either set the default encoding as a JVM option on the Java command line or change the encoding on the platform to utf-8.