public abstract class AbstractPullOMDataSource extends AbstractOMDataSource
OMDataSourceExt
implementations that can easily produce the content as an
XMLStreamReader
and that don't implement any kind of optimization for serializing the
content.LOSSY_PREFIX
Constructor and Description |
---|
AbstractPullOMDataSource() |
Modifier and Type | Method and Description |
---|---|
boolean |
isDestructiveWrite()
Returns true if writing the backing object is destructive.
|
void |
serialize(XMLStreamWriter xmlWriter)
Serializes element data directly to StAX writer.
|
close, copy, getObject, getProperty, getXMLBytes, getXMLInputStream, hasProperty, serialize, serialize, setProperty
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
isDestructiveRead
getReader
public final boolean isDestructiveWrite()
OMDataSourceExt
public final void serialize(XMLStreamWriter xmlWriter) throws XMLStreamException
OMDataSource
The implementation of this method must satisfy the following requirements:
XMLStreamWriter.writeStartDocument()
,
XMLStreamWriter.writeStartDocument(String)
,
XMLStreamWriter.writeStartDocument(String, String)
or
XMLStreamWriter.writeEndElement()
.
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.
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).
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.
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).
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:
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.
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:
OMSerializable.serialize(XMLStreamWriter, boolean)
method.
OMDataSource
is push-only (see AbstractPushOMDataSource
), then it
enables OMSourcedElement
to create OMText
nodes for the binary content in an
efficient way.
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:
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
.
xmlWriter
- destination writerXMLStreamException
Copyright © 2004–2024 The Apache Software Foundation. All rights reserved.