org.apache.axiom.om.ds

Class AbstractPullOMDataSource

java.lang.Object

org.apache.axiom.om.ds.AbstractOMDataSource

org.apache.axiom.om.ds.AbstractPullOMDataSource

All Implemented Interfaces:

OMDataSource, OMDataSourceExt

Direct Known Subclasses:

WrappedTextNodeOMDataSource

public abstract class AbstractPullOMDataSource
extends AbstractOMDataSource

Base class for OMDataSourceExt implementations that can easily produce the content as an XMLStreamReader and that don't implement any kind of optimization for serializing the content.

Field Summary

Fields inherited from interface org.apache.axiom.om.OMDataSourceExt

LOSSY_PREFIX

Constructor Summary

Constructors

Constructor and Description
AbstractPullOMDataSource()

Method Summary

Modifier and Type	Method and Description
boolean	isDestructiveWrite() Returns true if writing the backing object is destructive.
void	serialize(javax.xml.stream.XMLStreamWriter xmlWriter) Serializes element data directly to StAX writer.

Methods inherited from class org.apache.axiom.om.ds.AbstractOMDataSource

close, copy, getObject, getProperty, getXMLBytes, getXMLInputStream, hasProperty, serialize, serialize, setProperty

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.apache.axiom.om.OMDataSourceExt

isDestructiveRead

Methods inherited from interface org.apache.axiom.om.OMDataSource

getReader

Constructor Detail

AbstractPullOMDataSource

public AbstractPullOMDataSource()

Method Detail

isDestructiveWrite

public final boolean isDestructiveWrite()

Description copied from interface: OMDataSourceExt

Returns true if writing the backing object is destructive. An example of an object with a destructive write is an InputStream. The owning OMSourcedElement uses this information to detemine if OM tree expansion is needed when writing the OMDataSourceExt.

Returns:

boolean

serialize

public final void serialize(javax.xml.stream.XMLStreamWriter xmlWriter)
                     throws javax.xml.stream.XMLStreamException

Description copied from interface: OMDataSource

Serializes element data directly to StAX writer.

The implementation of this method must satisfy the following requirements:

• The implementation MUST NOT not write any start document or end document event, i.e. it MUST NOT use XMLStreamWriter.writeStartDocument(), XMLStreamWriter.writeStartDocument(String), XMLStreamWriter.writeStartDocument(String, String) or XMLStreamWriter.writeEndElement().
• The implementation MUST output a single element (hereafter called the root element). It MUST NOT output any other content before or after that element.
• The implementation MUST specify the namespace URI when writing an element, i.e. it MUST NOT use the namespace unaware methods XMLStreamWriter.writeStartElement(String) or XMLStreamWriter.writeEmptyElement(String). On the other hand, it MAY use the namespace unaware XMLStreamWriter.writeAttribute(String, String) method, provided that the specified name is an NCName.
• The implementation MUST ensure that the produced XML is well formed with respect to namespaces, i.e. it MUST generate the required namespace declarations using XMLStreamWriter.writeNamespace(String, String) and XMLStreamWriter.writeDefaultNamespace(String). It MUST NOT assume that the XMLStreamWriter performs any kind of namespace repairing (although that may be the case).
• In addition the implementation MAY use XMLStreamWriter.setPrefix(String, String) and XMLStreamWriter.setDefaultNamespace(String) to track the namespace declarations it generates, so that the namespace context maintained by the XMLStreamWriter accurately reflects the namespace context in the output document. However, it MUST NOT call these methods before the start of the root element or after the end of the root element.
• Since the element may be serialized as part of a larger document, the implementation MUST take into account the pre-existing namespace context (which can be queried using XMLStreamWriter.getPrefix(String) and XMLStreamWriter.getNamespaceContext()). This means that the implementation MUST NOT assume that the empty prefix is bound to the empty namespace URI. Therefore if the implementation outputs elements that have no namespace, it MUST generate namespace declarations of the form xmlns="" in the appropriate locations. In addition it MAY use the namespace context information to minimize the number of generated namespace declarations (by reusing already bound prefixes).
• If the implementation produces base64 binary data (that could be optimized using XOP/MTOM), then it SHOULD use XMLStreamWriterUtils.writeDataHandler(XMLStreamWriter, DataHandler, String, boolean) or XMLStreamWriterUtils.writeDataHandler(XMLStreamWriter, DataHandlerProvider, String, boolean) to write the data to the stream. If this is not possible (e.g. because the content is produced by a third party library that is not aware of these APIs), then the implementation MUST use the following approach:
  • If the XMLStreamWriter is an MTOMXMLStreamWriter, then the implementation MAY use MTOMXMLStreamWriter.prepareDataHandler(DataHandler) and generate the necessary xop:Include elements itself. In this case, the implementation MAY use MTOMXMLStreamWriter.isOptimized() to check if XOP/MTOM is enabled at all. Alternatively, instead of handling MTOMXMLStreamWriter in a special way, the implementation MAY use the approach described in the next item. This works because MTOMXMLStreamWriter exposes the DataHandlerWriter extension. However, this causes a slight overhead because the stream is first XOP decoded and then reencoded again.
  • If the XMLStreamWriter exposes the DataHandlerWriter extension, but is not an MTOMXMLStreamWriter (or is an MTOMXMLStreamWriter, but the implementation doesn't implement the approach described in the previous item), then the implementation MUST wrap the XMLStreamWriter in an XOPDecodingStreamWriter and write xop:Include elements to that wrapper, so that they can be translated into appropriate calls to the DataHandlerWriter. This requirement is important for two reasons:
    • It allows Axiom to respect the contract of the OMSerializable.serialize(XMLStreamWriter, boolean) method.
    • If the OMDataSource is push-only (see AbstractPushOMDataSource), then it enables OMSourcedElement to create OMText nodes for the binary content in an efficient way.
  • In all other cases, the implementation MUST use XMLStreamWriter.writeCharacters(String) or XMLStreamWriter.writeCharacters(char[], int, int) to write the base64 encoded data to the stream.

On the other hand, the caller of this method (typically an OMSourcedElement instance) must ensure that the following requirements are satisfied:

• The namespace context information provided by XMLStreamWriter.getPrefix(String) and XMLStreamWriter.getNamespaceContext() MUST accurately reflect the actual namespace context at the location in the output document where the root element is serialized. Note that this requirement may be relaxed if the caller implements some form of namespace repairing.

It is assumed that this method consumes the content (i.e. destroys the backing object) unless the data source also implements OMDataSourceExt and OMDataSourceExt.isDestructiveWrite() returns false.

Parameters:

xmlWriter - destination writer

Throws:

javax.xml.stream.XMLStreamException