org.apache.xml.dtm.ref.sax2dtm
Class SAX2RTFDTM

java.lang.Object
  org.apache.xml.dtm.ref.DTMDefaultBase
      org.apache.xml.dtm.ref.DTMDefaultBaseTraversers
          org.apache.xml.dtm.ref.DTMDefaultBaseIterators
              org.apache.xml.dtm.ref.sax2dtm.SAX2DTM
                  org.apache.xml.dtm.ref.sax2dtm.SAX2RTFDTM

All Implemented Interfaces:

DTM, org.xml.sax.ContentHandler, org.xml.sax.DTDHandler, org.xml.sax.EntityResolver, org.xml.sax.ErrorHandler, org.xml.sax.ext.DeclHandler, org.xml.sax.ext.LexicalHandler

public class SAX2RTFDTM
extends SAX2DTM

This is a subclass of SAX2DTM which has been modified to meet the needs of Result Tree Frameworks (RTFs). The differences are: 1) Multiple XML trees may be appended to the single DTM. This means that the root node of each document is _not_ node 0. Some code has had to be deoptimized to support this mode of operation, and an explicit mechanism for obtaining the Node Handle of the root node has been provided. 2) A stack of these documents is maintained, allowing us to "tail-prune" the most recently added trees off the end of the DTM as stylesheet elements (and thus variable contexts) are exited. PLEASE NOTE that this class may be _heavily_ dependent upon the internals of the SAX2DTM superclass, and must be maintained in parallel with that code. Arguably, they should be conditionals within a single class... but they have deen separated for performance reasons. (In fact, one could even argue about which is the superclass and which is the subclass; the current arrangement is as much about preserving stability of existing code during development as anything else.) %REVIEW% In fact, since the differences are so minor, I think it may be possible/practical to fold them back into the base SAX2DTM. Consider that as a future code-size optimization.

Nested Class Summary

Nested classes/interfaces inherited from class org.apache.xml.dtm.ref.DTMDefaultBaseIterators

DTMDefaultBaseIterators.AncestorIterator, DTMDefaultBaseIterators.AttributeIterator, DTMDefaultBaseIterators.ChildrenIterator, DTMDefaultBaseIterators.DescendantIterator, DTMDefaultBaseIterators.FollowingIterator, DTMDefaultBaseIterators.FollowingSiblingIterator, DTMDefaultBaseIterators.InternalAxisIteratorBase, DTMDefaultBaseIterators.NamespaceAttributeIterator, DTMDefaultBaseIterators.NamespaceChildrenIterator, DTMDefaultBaseIterators.NamespaceIterator, DTMDefaultBaseIterators.NthDescendantIterator, DTMDefaultBaseIterators.ParentIterator, DTMDefaultBaseIterators.PrecedingIterator, DTMDefaultBaseIterators.PrecedingSiblingIterator, DTMDefaultBaseIterators.RootIterator, DTMDefaultBaseIterators.SingletonIterator, DTMDefaultBaseIterators.TypedAncestorIterator, DTMDefaultBaseIterators.TypedAttributeIterator, DTMDefaultBaseIterators.TypedChildrenIterator, DTMDefaultBaseIterators.TypedDescendantIterator, DTMDefaultBaseIterators.TypedFollowingIterator, DTMDefaultBaseIterators.TypedFollowingSiblingIterator, DTMDefaultBaseIterators.TypedNamespaceIterator, DTMDefaultBaseIterators.TypedPrecedingIterator, DTMDefaultBaseIterators.TypedPrecedingSiblingIterator, DTMDefaultBaseIterators.TypedRootIterator, DTMDefaultBaseIterators.TypedSingletonIterator

Field Summary

private static boolean	DEBUG Set true to monitor SAX events and similar diagnostic info.
private int	m_currentDocumentNode Most recently started Document, or null if the DTM is empty.
(package private) int	m_emptyCharsCount Tail-pruning mark: initial number of characters in use
(package private) int	m_emptyDataCount Tail-pruning mark: initial number of data items in use
(package private) int	m_emptyDataQNCount Tail-pruning mark: default initial number of dataOrQName slots in use
(package private) int	m_emptyNodeCount Tail-pruning mark: initial number of nodes in use
(package private) int	m_emptyNSDeclSetCount Tail-pruning mark: initial number of namespace declaration sets
(package private) int	m_emptyNSDeclSetElemsCount Tail-pruning mark: initial number of namespace declaration elements
(package private) IntStack	mark_char_size Tail-pruning mark: Number of size-of-data fields in use
(package private) IntStack	mark_data_size Tail-pruning mark: Number of data items in use
(package private) IntStack	mark_doq_size Tail-pruning mark: Number of dataOrQName slots in use
(package private) IntStack	mark_nsdeclelem_size Tail-pruning mark: Number of naespace declaration elements in use %REVIEW% I don't think number of NS sets is ever different from number of NS elements.
(package private) IntStack	mark_nsdeclset_size Tail-pruning mark: Number of namespace declaration sets in use %REVIEW% I don't think number of NS sets is ever different from number of NS elements.
(package private) IntStack	mark_size Tail-pruning mark: Number of nodes in use

Fields inherited from class org.apache.xml.dtm.ref.sax2dtm.SAX2DTM

m_chars, m_coalescedTextType, m_contextIndexes, m_data, m_dataOrQName, m_endDocumentOccured, m_idAttributes, m_insideDTD, m_locator, m_parents, m_pastFirstElement, m_prefixMappings, m_previous, m_sourceColumn, m_sourceLine, m_sourceSystemId, m_textPendingStart, m_textType, m_useSourceLocationProperty, m_valuesOrPrefixes, m_walker

Fields inherited from class org.apache.xml.dtm.ref.DTMDefaultBase

DEFAULT_BLOCKSIZE, DEFAULT_NUMBLOCKS, DEFAULT_NUMBLOCKS_SMALL, m_documentBaseURI, m_dtmIdent, m_elemIndexes, m_expandedNameTable, m_exptype, m_firstch, m_indexing, m_mgr, m_mgrDefault, m_namespaceDeclSetElements, m_namespaceDeclSets, m_nextsib, m_parent, m_prevsib, m_shouldStripWhitespaceStack, m_shouldStripWS, m_size, m_traversers, m_wsfilter, m_xstrf, NOTPROCESSED, ROOTNODE

Fields inherited from interface org.apache.xml.dtm.DTM

ATTRIBUTE_NODE, CDATA_SECTION_NODE, COMMENT_NODE, DOCUMENT_FRAGMENT_NODE, DOCUMENT_NODE, DOCUMENT_TYPE_NODE, ELEMENT_NODE, ENTITY_NODE, ENTITY_REFERENCE_NODE, NAMESPACE_NODE, NOTATION_NODE, NTYPES, NULL, PROCESSING_INSTRUCTION_NODE, ROOT_NODE, TEXT_NODE

Constructor Summary

SAX2RTFDTM(DTMManager mgr, javax.xml.transform.Source source, int dtmIdentity, DTMWSFilter whiteSpaceFilter, XMLStringFactory xstringfactory, boolean doIndexing)

Method Summary

protected int	_documentRoot(int nodeIdentifier) Given a node identifier, find the owning document node.
void	endDocument() Receive notification of the end of the document.
int	getDocument() Given a DTM, find the owning document node.
int	getDocumentRoot(int nodeHandle) Given a node handle, find the owning document node, using DTM semantics (Document owns itself) rather than DOM semantics (Document has no owner).
boolean	isTreeIncomplete()
boolean	popRewindMark() "Tail-pruning" support for RTFs.
void	pushRewindMark() "Tail-pruning" support for RTFs.
void	startDocument() Receive notification of the beginning of a new RTF document.

Methods inherited from class org.apache.xml.dtm.ref.sax2dtm.SAX2DTM

_dataOrQName, addNewDTMID, addNode, attributeDecl, characters, charactersFlush, clearCoRoutine, clearCoRoutine, comment, declAlreadyDeclared, dispatchCharactersEvents, dispatchToEvents, elementDecl, endCDATA, endDTD, endElement, endEntity, endPrefixMapping, error, externalEntityDecl, fatalError, getAttributeNode, getContentHandler, getDeclHandler, getDocumentTypeDeclarationPublicIdentifier, getDocumentTypeDeclarationSystemIdentifier, getDTDHandler, getElementById, getEntityResolver, getErrorHandler, getFixedNames, getIdForNamespace, getLexicalHandler, getLocalName, getNamespaceURI, getNamespaceURI, getNextNodeIdentity, getNodeName, getNodeNameX, getNodeValue, getNumberOfNodes, getPrefix, getPrefix, getSourceLocatorFor, getStringValue, getUnparsedEntityURI, ignorableWhitespace, internalEntityDecl, isAttributeSpecified, isWhitespace, migrateTo, needsTwoThreads, nextNode, notationDecl, processingInstruction, resolveEntity, setDocumentLocator, setIDAttribute, setIncrementalSAXSource, setProperty, setSourceLocation, setUseSourceLocation, skippedEntity, startCDATA, startDTD, startElement, startEntity, startPrefixMapping, unparsedEntityDecl, warning

Methods inherited from class org.apache.xml.dtm.ref.DTMDefaultBaseIterators

getAxisIterator, getTypedAxisIterator

Methods inherited from class org.apache.xml.dtm.ref.DTMDefaultBaseTraversers

getAxisTraverser

Methods inherited from class org.apache.xml.dtm.ref.DTMDefaultBase

_exptype, _firstch, _level, _nextsib, _parent, _prevsib, _type, appendChild, appendTextChild, declareNamespaceInContext, documentRegistration, documentRelease, dumpDTM, dumpNode, ensureSizeOfIndex, error, findGTE, findInSortedSuballocatedIntVector, findNamespaceContext, getDocumentAllDeclarationsProcessed, getDocumentBaseURI, getDocumentEncoding, getDocumentStandalone, getDocumentSystemIdentifier, getDocumentVersion, getDTMIDs, getExpandedTypeID, getExpandedTypeID, getFirstAttribute, getFirstAttributeIdentity, getFirstChild, getFirstNamespaceNode, getLastChild, getLevel, getLocalNameFromExpandedNameID, getManager, getNamespaceFromExpandedNameID, getNamespaceType, getNextAttribute, getNextAttributeIdentity, getNextNamespaceNode, getNextSibling, getNode, getNodeHandle, getNodeIdent, getNodeType, getOwnerDocument, getParent, getPreviousSibling, getShouldStripWhitespace, getStringValueChunk, getStringValueChunkCount, getTypedAttribute, getTypedFirstChild, getTypedNextSibling, hasChildNodes, indexNode, isCharacterElementContentWhitespace, isDocumentAllDeclarationsProcessed, isNodeAfter, isSupported, makeNodeHandle, makeNodeIdentity, popShouldStripWhitespace, pushShouldStripWhitespace, setDocumentBaseURI, setFeature, setShouldStripWhitespace, supportsPreStripping

Methods inherited from class java.lang.Object

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

Field Detail

DEBUG

private static final boolean DEBUG

Set true to monitor SAX events and similar diagnostic info.

See Also:

Constant Field Values

m_currentDocumentNode

private int m_currentDocumentNode

Most recently started Document, or null if the DTM is empty.

mark_size

IntStack mark_size

Tail-pruning mark: Number of nodes in use

mark_data_size

IntStack mark_data_size

Tail-pruning mark: Number of data items in use

mark_char_size

IntStack mark_char_size

Tail-pruning mark: Number of size-of-data fields in use

mark_doq_size

IntStack mark_doq_size

Tail-pruning mark: Number of dataOrQName slots in use

mark_nsdeclset_size

IntStack mark_nsdeclset_size

Tail-pruning mark: Number of namespace declaration sets in use %REVIEW% I don't think number of NS sets is ever different from number of NS elements. We can probabably reduce these to a single stack and save some storage.

mark_nsdeclelem_size

IntStack mark_nsdeclelem_size

Tail-pruning mark: Number of naespace declaration elements in use %REVIEW% I don't think number of NS sets is ever different from number of NS elements. We can probabably reduce these to a single stack and save some storage.

m_emptyNodeCount

int m_emptyNodeCount

Tail-pruning mark: initial number of nodes in use

m_emptyNSDeclSetCount

int m_emptyNSDeclSetCount

Tail-pruning mark: initial number of namespace declaration sets

m_emptyNSDeclSetElemsCount

int m_emptyNSDeclSetElemsCount

Tail-pruning mark: initial number of namespace declaration elements

m_emptyDataCount

int m_emptyDataCount

Tail-pruning mark: initial number of data items in use

m_emptyCharsCount

int m_emptyCharsCount

Tail-pruning mark: initial number of characters in use

m_emptyDataQNCount

int m_emptyDataQNCount

Tail-pruning mark: default initial number of dataOrQName slots in use

Constructor Detail

SAX2RTFDTM

public SAX2RTFDTM(DTMManager mgr,
                  javax.xml.transform.Source source,
                  int dtmIdentity,
                  DTMWSFilter whiteSpaceFilter,
                  XMLStringFactory xstringfactory,
                  boolean doIndexing)

Method Detail

getDocument

public int getDocument()

Given a DTM, find the owning document node. In the case of SAX2RTFDTM, which may contain multiple documents, this returns the most recently started document, or null if the DTM is empty or no document is currently under construction. %REVIEW% Should we continue to report the most recent after construction has ended? I think not, given that it may have been tail-pruned.

Specified by:

getDocument in interface DTM

Overrides:

getDocument in class DTMDefaultBase

Returns:

int Node handle of Document node, or null if this DTM does not contain an "active" document.

getDocumentRoot

public int getDocumentRoot(int nodeHandle)

Given a node handle, find the owning document node, using DTM semantics (Document owns itself) rather than DOM semantics (Document has no owner). (I'm counting on the fact that getOwnerDocument() is implemented on top of this call, in the superclass, to avoid having to rewrite that one. Be careful if that code changes!)

Specified by:

getDocumentRoot in interface DTM

Overrides:

getDocumentRoot in class DTMDefaultBase

Parameters:

nodeHandle - the id of the node.

Returns:

int Node handle of owning document

See Also:

DTM.getOwnerDocument(int nodeHandle)

_documentRoot

protected int _documentRoot(int nodeIdentifier)

Given a node identifier, find the owning document node. Unlike the DOM, this considers the owningDocument of a Document to be itself. Note that in shared DTMs this may not be zero.

Parameters:

nodeIdentifier - the id of the starting node.

Returns:

int Node identifier of the root of this DTM tree

startDocument

public void startDocument()
                   throws org.xml.sax.SAXException

Receive notification of the beginning of a new RTF document. %REVIEW% Y'know, this isn't all that much of a deoptimization. We might want to consider folding the start/endDocument changes back into the main SAX2DTM so we don't have to expose so many fields (even as Protected) and carry the additional code.

Specified by:

startDocument in interface org.xml.sax.ContentHandler

Overrides:

startDocument in class SAX2DTM

Throws:

org.xml.sax.SAXException - Any SAX exception, possibly wrapping another exception.

See Also:

ContentHandler.startDocument()

endDocument

public void endDocument()
                 throws org.xml.sax.SAXException

Receive notification of the end of the document. %REVIEW% Y'know, this isn't all that much of a deoptimization. We might want to consider folding the start/endDocument changes back into the main SAX2DTM so we don't have to expose so many fields (even as Protected).

Specified by:

endDocument in interface org.xml.sax.ContentHandler

Overrides:

endDocument in class SAX2DTM

Throws:

org.xml.sax.SAXException - Any SAX exception, possibly wrapping another exception.

See Also:

ContentHandler.endDocument()

pushRewindMark

public void pushRewindMark()

"Tail-pruning" support for RTFs. This function pushes information about the current size of the DTM's data structures onto a stack, for use by popRewindMark() (which see). %REVIEW% I have no idea how to rewind m_elemIndexes. However, RTFs will not be indexed, so I can simply panic if that case arises. Hey, it works...

popRewindMark

public boolean popRewindMark()

"Tail-pruning" support for RTFs. This function pops the information previously saved by pushRewindMark (which see) and uses it to discard all nodes added to the DTM after that time. We expect that this will allow us to reuse storage more effectively. This is _not_ intended to be called while a document is still being constructed -- only between endDocument and the next startDocument %REVIEW% WARNING: This is the first use of some of the truncation methods. If Xalan blows up after this is called, that's a likely place to check. %REVIEW% Our original design for DTMs permitted them to share string pools. If there any risk that this might be happening, we can _not_ rewind and recover the string storage. One solution might to assert that DTMs used for RTFs Must Not take advantage of that feature, but this seems excessively fragile. Another, much less attractive, would be to just let them leak... Nah.

Returns:

true if and only if the pop completely emptied the RTF. That response is used when determining how to unspool RTF-started-while-RTF-open situations.

isTreeIncomplete

public boolean isTreeIncomplete()

Returns:

true if a DTM tree is currently under construction.