2 * Licensed to the Apache Software Foundation (ASF) under one
3 * or more contributor license agreements. See the NOTICE file
4 * distributed with this work for additional information
5 * regarding copyright ownership. The ASF licenses this file
6 * to you under the Apache License, Version 2.0 (the "License");
7 * you may not use this file except in compliance with the License.
8 * You may obtain a copy of the License at
10 * http://www.apache.org/licenses/LICENSE-2.0
12 * Unless required by applicable law or agreed to in writing, software
13 * distributed under the License is distributed on an "AS IS" BASIS,
14 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15 * See the License for the specific language governing permissions and
16 * limitations under the License.
19 * $Id: DTMDefaultBase.java 468653 2006-10-28 07:07:05Z minchau $
21 package org.apache.xml.dtm.ref;
23 import org.apache.xml.dtm.*;
24 import org.apache.xml.utils.SuballocatedIntVector;
25 import org.apache.xml.utils.BoolStack;
27 import java.util.Vector;
29 import javax.xml.transform.Source;
31 import org.apache.xml.utils.XMLString;
32 import org.apache.xml.utils.XMLStringFactory;
34 import org.apache.xml.res.XMLMessages;
35 import org.apache.xml.res.XMLErrorResources;
37 import java.io.*; // for dumpDTM
40 * The <code>DTMDefaultBase</code> class serves as a helper base for DTMs.
41 * It sets up structures for navigation and type, while leaving data
42 * management and construction to the derived classes.
44 public abstract class DTMDefaultBase implements DTM
46 static final boolean JJK_DEBUG=false;
48 // This constant is likely to be removed in the future. Use the
49 // getDocument() method instead of ROOTNODE to get at the root
51 /** The identity of the root node. */
52 public static final int ROOTNODE = 0;
55 * The number of nodes, which is also used to determine the next
58 protected int m_size = 0;
60 /** The expanded names, one array element for each node. */
61 protected SuballocatedIntVector m_exptype;
63 /** First child values, one array element for each node. */
64 protected SuballocatedIntVector m_firstch;
66 /** Next sibling values, one array element for each node. */
67 protected SuballocatedIntVector m_nextsib;
69 /** Previous sibling values, one array element for each node. */
70 protected SuballocatedIntVector m_prevsib;
72 /** Previous sibling values, one array element for each node. */
73 protected SuballocatedIntVector m_parent;
75 /** Vector of SuballocatedIntVectors of NS decl sets */
76 protected Vector m_namespaceDeclSets = null;
78 /** SuballocatedIntVector of elements at which corresponding
79 * namespaceDeclSets were defined */
80 protected SuballocatedIntVector m_namespaceDeclSetElements = null;
83 * These hold indexes to elements based on namespace and local name.
84 * The base lookup is the the namespace. The second lookup is the local
85 * name, and the last array contains the the first free element
86 * at the start, and the list of element handles following.
88 protected int[][][] m_elemIndexes;
90 /** The default block size of the node arrays */
91 public static final int DEFAULT_BLOCKSIZE = 512; // favor small docs.
93 /** The number of blocks for the node arrays */
94 public static final int DEFAULT_NUMBLOCKS = 32;
96 /** The number of blocks used for small documents & RTFs */
97 public static final int DEFAULT_NUMBLOCKS_SMALL = 4;
99 /** The block size of the node arrays */
100 //protected final int m_blocksize;
103 * The value to use when the information has not been built yet.
105 protected static final int NOTPROCESSED = DTM.NULL - 1;
108 * The DTM manager who "owns" this DTM.
111 public DTMManager m_mgr;
114 * m_mgr cast to DTMManagerDefault, or null if it isn't an instance
117 protected DTMManagerDefault m_mgrDefault=null;
120 /** The document identity number(s). If we have overflowed the addressing
121 * range of the first that was assigned to us, we may add others. */
122 protected SuballocatedIntVector m_dtmIdent;
124 /** The mask for the identity.
125 %REVIEW% Should this really be set to the _DEFAULT? What if
126 a particular DTM wanted to use another value? */
127 //protected final static int m_mask = DTMManager.IDENT_NODE_DEFAULT;
129 /** The base URI for this document. */
130 protected String m_documentBaseURI;
133 * The whitespace filter that enables elements to strip whitespace or not.
135 protected DTMWSFilter m_wsfilter;
137 /** Flag indicating whether to strip whitespace nodes */
138 protected boolean m_shouldStripWS = false;
140 /** Stack of flags indicating whether to strip whitespace nodes */
141 protected BoolStack m_shouldStripWhitespaceStack;
143 /** The XMLString factory for creating XMLStrings. */
144 protected XMLStringFactory m_xstrf;
147 * The table for exandedNameID lookups. This may or may not be the same
148 * table as is contained in the DTMManagerDefault.
150 protected ExpandedNameTable m_expandedNameTable;
152 /** true if indexing is turned on. */
153 protected boolean m_indexing;
156 * Construct a DTMDefaultBase object using the default block size.
158 * @param mgr The DTMManager who owns this DTM.
159 * @param source The object that is used to specify the construction source.
160 * @param dtmIdentity The DTM identity ID for this DTM.
161 * @param whiteSpaceFilter The white space filter for this DTM, which may
163 * @param xstringfactory The factory to use for creating XMLStrings.
164 * @param doIndexing true if the caller considers it worth it to use
167 public DTMDefaultBase(DTMManager mgr, Source source, int dtmIdentity,
168 DTMWSFilter whiteSpaceFilter,
169 XMLStringFactory xstringfactory, boolean doIndexing)
171 this(mgr, source, dtmIdentity, whiteSpaceFilter, xstringfactory,
172 doIndexing, DEFAULT_BLOCKSIZE, true, false);
176 * Construct a DTMDefaultBase object from a DOM node.
178 * @param mgr The DTMManager who owns this DTM.
179 * @param source The object that is used to specify the construction source.
180 * @param dtmIdentity The DTM identity ID for this DTM.
181 * @param whiteSpaceFilter The white space filter for this DTM, which may
183 * @param xstringfactory The factory to use for creating XMLStrings.
184 * @param doIndexing true if the caller considers it worth it to use
186 * @param blocksize The block size of the DTM.
187 * @param usePrevsib true if we want to build the previous sibling node array.
188 * @param newNameTable true if we want to use a new ExpandedNameTable for this DTM.
190 public DTMDefaultBase(DTMManager mgr, Source source, int dtmIdentity,
191 DTMWSFilter whiteSpaceFilter,
192 XMLStringFactory xstringfactory, boolean doIndexing,
193 int blocksize, boolean usePrevsib,
194 boolean newNameTable)
196 // Use smaller sizes for the internal node arrays if the block size
201 numblocks = DEFAULT_NUMBLOCKS_SMALL;
202 m_dtmIdent= new SuballocatedIntVector(4, 1);
206 numblocks = DEFAULT_NUMBLOCKS;
207 m_dtmIdent= new SuballocatedIntVector(32);
210 m_exptype = new SuballocatedIntVector(blocksize, numblocks);
211 m_firstch = new SuballocatedIntVector(blocksize, numblocks);
212 m_nextsib = new SuballocatedIntVector(blocksize, numblocks);
213 m_parent = new SuballocatedIntVector(blocksize, numblocks);
215 // Only create the m_prevsib array if the usePrevsib flag is true.
216 // Some DTM implementations (e.g. SAXImpl) do not need this array.
217 // We can save the time to build it in those cases.
219 m_prevsib = new SuballocatedIntVector(blocksize, numblocks);
222 if(mgr instanceof DTMManagerDefault)
223 m_mgrDefault=(DTMManagerDefault)mgr;
225 m_documentBaseURI = (null != source) ? source.getSystemId() : null;
226 m_dtmIdent.setElementAt(dtmIdentity,0);
227 m_wsfilter = whiteSpaceFilter;
228 m_xstrf = xstringfactory;
229 m_indexing = doIndexing;
233 m_expandedNameTable = new ExpandedNameTable();
237 // Note that this fails if we aren't talking to an instance of
239 m_expandedNameTable = m_mgrDefault.getExpandedNameTable(this);
242 if (null != whiteSpaceFilter)
244 m_shouldStripWhitespaceStack = new BoolStack();
246 pushShouldStripWhitespace(false);
251 * Ensure that the size of the element indexes can hold the information.
253 * @param namespaceID Namespace ID index.
254 * @param LocalNameID Local name ID.
256 protected void ensureSizeOfIndex(int namespaceID, int LocalNameID)
259 if (null == m_elemIndexes)
261 m_elemIndexes = new int[namespaceID + 20][][];
263 else if (m_elemIndexes.length <= namespaceID)
265 int[][][] indexes = m_elemIndexes;
267 m_elemIndexes = new int[namespaceID + 20][][];
269 System.arraycopy(indexes, 0, m_elemIndexes, 0, indexes.length);
272 int[][] localNameIndex = m_elemIndexes[namespaceID];
274 if (null == localNameIndex)
276 localNameIndex = new int[LocalNameID + 100][];
277 m_elemIndexes[namespaceID] = localNameIndex;
279 else if (localNameIndex.length <= LocalNameID)
281 int[][] indexes = localNameIndex;
283 localNameIndex = new int[LocalNameID + 100][];
285 System.arraycopy(indexes, 0, localNameIndex, 0, indexes.length);
287 m_elemIndexes[namespaceID] = localNameIndex;
290 int[] elemHandles = localNameIndex[LocalNameID];
292 if (null == elemHandles)
294 elemHandles = new int[128];
295 localNameIndex[LocalNameID] = elemHandles;
298 else if (elemHandles.length <= elemHandles[0] + 1)
300 int[] indexes = elemHandles;
302 elemHandles = new int[elemHandles[0] + 1024];
304 System.arraycopy(indexes, 0, elemHandles, 0, indexes.length);
306 localNameIndex[LocalNameID] = elemHandles;
311 * Add a node to the element indexes. The node will not be added unless
314 * @param expandedTypeID The expanded type ID of the node.
315 * @param identity The node identity index.
317 protected void indexNode(int expandedTypeID, int identity)
320 ExpandedNameTable ent = m_expandedNameTable;
321 short type = ent.getType(expandedTypeID);
323 if (DTM.ELEMENT_NODE == type)
325 int namespaceID = ent.getNamespaceID(expandedTypeID);
326 int localNameID = ent.getLocalNameID(expandedTypeID);
328 ensureSizeOfIndex(namespaceID, localNameID);
330 int[] index = m_elemIndexes[namespaceID][localNameID];
332 index[index[0]] = identity;
339 * Find the first index that occurs in the list that is greater than or
340 * equal to the given value.
342 * @param list A list of integers.
343 * @param start The start index to begin the search.
344 * @param len The number of items to search.
345 * @param value Find the slot that has a value that is greater than or
346 * identical to this argument.
348 * @return The index in the list of the slot that is higher or identical
349 * to the identity argument, or -1 if no node is higher or equal.
351 protected int findGTE(int[] list, int start, int len, int value)
355 int high = start + (len - 1);
360 int mid = (low + high) / 2;
371 return (low <= end && list[low] > value) ? low : -1;
375 * Find the first matching element from the index at or after the
378 * @param nsIndex The namespace index lookup.
379 * @param lnIndex The local name index lookup.
380 * @param firstPotential The first potential match that is worth looking at.
382 * @return The first node that is greater than or equal to the
383 * firstPotential argument, or DTM.NOTPROCESSED if not found.
385 int findElementFromIndex(int nsIndex, int lnIndex, int firstPotential)
388 int[][][] indexes = m_elemIndexes;
390 if (null != indexes && nsIndex < indexes.length)
392 int[][] lnIndexs = indexes[nsIndex];
394 if (null != lnIndexs && lnIndex < lnIndexs.length)
396 int[] elems = lnIndexs[lnIndex];
400 int pos = findGTE(elems, 1, elems[0], firstPotential);
414 * Get the next node identity value in the list, and call the iterator
415 * if it hasn't been added yet.
417 * @param identity The node identity (index).
418 * @return identity+1, or DTM.NULL.
420 protected abstract int getNextNodeIdentity(int identity);
423 * This method should try and build one or more nodes in the table.
425 * @return The true if a next node is found or false if
426 * there are no more nodes.
428 protected abstract boolean nextNode();
431 * Get the number of nodes that have been added.
433 * @return the number of nodes that have been mapped.
435 protected abstract int getNumberOfNodes();
437 /** Stateless axis traversers, lazely built. */
438 protected DTMAxisTraverser[] m_traversers;
441 // * Ensure that the size of the information arrays can hold another entry
442 // * at the given index.
444 // * @param index On exit from this function, the information arrays sizes must be
445 // * at least index+1.
447 // protected void ensureSize(int index)
449 // // We've cut over to Suballocated*Vector, which are self-sizing.
453 * Get the simple type ID for the given node identity.
455 * @param identity The node identity.
457 * @return The simple type ID, or DTM.NULL.
459 protected short _type(int identity)
462 int info = _exptype(identity);
465 return m_expandedNameTable.getType(info);
471 * Get the expanded type ID for the given node identity.
473 * @param identity The node identity.
475 * @return The expanded type ID, or DTM.NULL.
477 protected int _exptype(int identity)
479 if (identity == DTM.NULL)
481 // Reorganized test and loop into single flow
482 // Tiny performance improvement, saves a few bytes of code, clearer.
483 // %OPT% Other internal getters could be treated simliarly
484 while (identity>=m_size)
486 if (!nextNode() && identity >= m_size)
489 return m_exptype.elementAt(identity);
494 * Get the level in the tree for the given node identity.
496 * @param identity The node identity.
498 * @return The tree level, or DTM.NULL.
500 protected int _level(int identity)
502 while (identity>=m_size)
504 boolean isMore = nextNode();
505 if (!isMore && identity >= m_size)
510 while(NULL != (identity=_parent(identity)))
516 * Get the first child for the given node identity.
518 * @param identity The node identity.
520 * @return The first child identity, or DTM.NULL.
522 protected int _firstch(int identity)
525 // Boiler-plate code for each of the _xxx functions, except for the array.
526 int info = (identity >= m_size) ? NOTPROCESSED : m_firstch.elementAt(identity);
528 // Check to see if the information requested has been processed, and,
529 // if not, advance the iterator until we the information has been
531 while (info == NOTPROCESSED)
533 boolean isMore = nextNode();
535 if (identity >= m_size &&!isMore)
539 info = m_firstch.elementAt(identity);
540 if(info == NOTPROCESSED && !isMore)
549 * Get the next sibling for the given node identity.
551 * @param identity The node identity.
553 * @return The next sibling identity, or DTM.NULL.
555 protected int _nextsib(int identity)
557 // Boiler-plate code for each of the _xxx functions, except for the array.
558 int info = (identity >= m_size) ? NOTPROCESSED : m_nextsib.elementAt(identity);
560 // Check to see if the information requested has been processed, and,
561 // if not, advance the iterator until we the information has been
563 while (info == NOTPROCESSED)
565 boolean isMore = nextNode();
567 if (identity >= m_size &&!isMore)
571 info = m_nextsib.elementAt(identity);
572 if(info == NOTPROCESSED && !isMore)
581 * Get the previous sibling for the given node identity.
583 * @param identity The node identity.
585 * @return The previous sibling identity, or DTM.NULL.
587 protected int _prevsib(int identity)
590 if (identity < m_size)
591 return m_prevsib.elementAt(identity);
593 // Check to see if the information requested has been processed, and,
594 // if not, advance the iterator until we the information has been
598 boolean isMore = nextNode();
600 if (identity >= m_size && !isMore)
602 else if (identity < m_size)
603 return m_prevsib.elementAt(identity);
608 * Get the parent for the given node identity.
610 * @param identity The node identity.
612 * @return The parent identity, or DTM.NULL.
614 protected int _parent(int identity)
617 if (identity < m_size)
618 return m_parent.elementAt(identity);
620 // Check to see if the information requested has been processed, and,
621 // if not, advance the iterator until we the information has been
625 boolean isMore = nextNode();
627 if (identity >= m_size && !isMore)
629 else if (identity < m_size)
630 return m_parent.elementAt(identity);
635 * Diagnostics function to dump the DTM.
637 public void dumpDTM(OutputStream os)
643 File f = new File("DTMDump"+((Object)this).hashCode()+".txt");
644 System.err.println("Dumping... "+f.getAbsolutePath());
645 os=new FileOutputStream(f);
647 PrintStream ps = new PrintStream(os);
651 int nRecords = m_size;
653 ps.println("Total nodes: " + nRecords);
655 for (int index = 0; index < nRecords; ++index)
657 int i=makeNodeHandle(index);
658 ps.println("=========== index=" + index + " handle=" + i + " ===========");
659 ps.println("NodeName: " + getNodeName(i));
660 ps.println("NodeNameX: " + getNodeNameX(i));
661 ps.println("LocalName: " + getLocalName(i));
662 ps.println("NamespaceURI: " + getNamespaceURI(i));
663 ps.println("Prefix: " + getPrefix(i));
665 int exTypeID = _exptype(index);
667 ps.println("Expanded Type ID: "
668 + Integer.toHexString(exTypeID));
670 int type = _type(index);
675 case DTM.ATTRIBUTE_NODE :
676 typestring = "ATTRIBUTE_NODE";
678 case DTM.CDATA_SECTION_NODE :
679 typestring = "CDATA_SECTION_NODE";
681 case DTM.COMMENT_NODE :
682 typestring = "COMMENT_NODE";
684 case DTM.DOCUMENT_FRAGMENT_NODE :
685 typestring = "DOCUMENT_FRAGMENT_NODE";
687 case DTM.DOCUMENT_NODE :
688 typestring = "DOCUMENT_NODE";
690 case DTM.DOCUMENT_TYPE_NODE :
691 typestring = "DOCUMENT_NODE";
693 case DTM.ELEMENT_NODE :
694 typestring = "ELEMENT_NODE";
696 case DTM.ENTITY_NODE :
697 typestring = "ENTITY_NODE";
699 case DTM.ENTITY_REFERENCE_NODE :
700 typestring = "ENTITY_REFERENCE_NODE";
702 case DTM.NAMESPACE_NODE :
703 typestring = "NAMESPACE_NODE";
705 case DTM.NOTATION_NODE :
706 typestring = "NOTATION_NODE";
711 case DTM.PROCESSING_INSTRUCTION_NODE :
712 typestring = "PROCESSING_INSTRUCTION_NODE";
715 typestring = "TEXT_NODE";
718 typestring = "Unknown!";
722 ps.println("Type: " + typestring);
724 int firstChild = _firstch(index);
726 if (DTM.NULL == firstChild)
727 ps.println("First child: DTM.NULL");
728 else if (NOTPROCESSED == firstChild)
729 ps.println("First child: NOTPROCESSED");
731 ps.println("First child: " + firstChild);
733 if (m_prevsib != null)
735 int prevSibling = _prevsib(index);
737 if (DTM.NULL == prevSibling)
738 ps.println("Prev sibling: DTM.NULL");
739 else if (NOTPROCESSED == prevSibling)
740 ps.println("Prev sibling: NOTPROCESSED");
742 ps.println("Prev sibling: " + prevSibling);
745 int nextSibling = _nextsib(index);
747 if (DTM.NULL == nextSibling)
748 ps.println("Next sibling: DTM.NULL");
749 else if (NOTPROCESSED == nextSibling)
750 ps.println("Next sibling: NOTPROCESSED");
752 ps.println("Next sibling: " + nextSibling);
754 int parent = _parent(index);
756 if (DTM.NULL == parent)
757 ps.println("Parent: DTM.NULL");
758 else if (NOTPROCESSED == parent)
759 ps.println("Parent: NOTPROCESSED");
761 ps.println("Parent: " + parent);
763 int level = _level(index);
765 ps.println("Level: " + level);
766 ps.println("Node Value: " + getNodeValue(i));
767 ps.println("String Value: " + getStringValue(i));
770 catch(IOException ioe)
772 ioe.printStackTrace(System.err);
773 throw new RuntimeException(ioe.getMessage());
778 * Diagnostics function to dump a single node.
780 * %REVIEW% KNOWN GLITCH: If you pass it a node index rather than a
781 * node handle, it works just fine... but the displayed identity
782 * number before the colon is different, which complicates comparing
783 * it with nodes printed the other way. We could always OR the DTM ID
784 * into the value, to suppress that distinction...
786 * %REVIEW% This might want to be moved up to DTMDefaultBase, or possibly
787 * DTM itself, since it's a useful diagnostic and uses only DTM's public
790 public String dumpNode(int nodeHandle)
792 if(nodeHandle==DTM.NULL)
796 switch (getNodeType(nodeHandle))
798 case DTM.ATTRIBUTE_NODE :
801 case DTM.CDATA_SECTION_NODE :
802 typestring = "CDATA";
804 case DTM.COMMENT_NODE :
805 typestring = "COMMENT";
807 case DTM.DOCUMENT_FRAGMENT_NODE :
808 typestring = "DOC_FRAG";
810 case DTM.DOCUMENT_NODE :
813 case DTM.DOCUMENT_TYPE_NODE :
814 typestring = "DOC_TYPE";
816 case DTM.ELEMENT_NODE :
817 typestring = "ELEMENT";
819 case DTM.ENTITY_NODE :
820 typestring = "ENTITY";
822 case DTM.ENTITY_REFERENCE_NODE :
823 typestring = "ENT_REF";
825 case DTM.NAMESPACE_NODE :
826 typestring = "NAMESPACE";
828 case DTM.NOTATION_NODE :
829 typestring = "NOTATION";
834 case DTM.PROCESSING_INSTRUCTION_NODE :
841 typestring = "Unknown!";
845 StringBuffer sb=new StringBuffer();
846 sb.append("["+nodeHandle+": "+typestring+
847 "(0x"+Integer.toHexString(getExpandedTypeID(nodeHandle))+") "+
848 getNodeNameX(nodeHandle)+" {"+getNamespaceURI(nodeHandle)+"}"+
849 "=\""+ getNodeValue(nodeHandle)+"\"]");
850 return sb.toString();
853 // ========= DTM Implementation Control Functions. ==============
856 * Set an implementation dependent feature.
858 * %REVIEW% Do we really expect to set features on DTMs?
860 * @param featureId A feature URL.
861 * @param state true if this feature should be on, false otherwise.
863 public void setFeature(String featureId, boolean state){}
865 // ========= Document Navigation Functions =========
868 * Given a node handle, test if it has child nodes.
869 * <p> %REVIEW% This is obviously useful at the DOM layer, where it
870 * would permit testing this without having to create a proxy
871 * node. It's less useful in the DTM API, where
872 * (dtm.getFirstChild(nodeHandle)!=DTM.NULL) is just as fast and
873 * almost as self-evident. But it's a convenience, and eases porting
874 * of DOM code to DTM. </p>
876 * @param nodeHandle int Handle of the node.
877 * @return int true if the given node has child nodes.
879 public boolean hasChildNodes(int nodeHandle)
882 int identity = makeNodeIdentity(nodeHandle);
883 int firstChild = _firstch(identity);
885 return firstChild != DTM.NULL;
888 /** Given a node identity, return a node handle. If extended addressing
889 * has been used (multiple DTM IDs), we need to map the high bits of the
890 * identity into the proper DTM ID.
892 * This has been made FINAL to facilitate inlining, since we do not expect
893 * any subclass of DTMDefaultBase to ever change the algorithm. (I don't
894 * really like doing so, and would love to have an excuse not to...)
896 * %REVIEW% Is it worth trying to specialcase small documents?
897 * %REVIEW% Should this be exposed at the package/public layers?
899 * @param nodeIdentity Internal offset to this node's records.
900 * @return NodeHandle (external representation of node)
902 final public int makeNodeHandle(int nodeIdentity)
904 if(NULL==nodeIdentity) return NULL;
906 if(JJK_DEBUG && nodeIdentity>DTMManager.IDENT_NODE_DEFAULT)
907 System.err.println("GONK! (only useful in limited situations)");
909 return m_dtmIdent.elementAt(nodeIdentity >>> DTMManager.IDENT_DTM_NODE_BITS)
910 + (nodeIdentity & DTMManager.IDENT_NODE_DEFAULT) ;
913 /** Given a node handle, return a node identity. If extended addressing
914 * has been used (multiple DTM IDs), we need to map the high bits of the
915 * identity into the proper DTM ID and thence find the proper offset
916 * to add to the low bits of the identity
918 * This has been made FINAL to facilitate inlining, since we do not expect
919 * any subclass of DTMDefaultBase to ever change the algorithm. (I don't
920 * really like doing so, and would love to have an excuse not to...)
922 * %OPT% Performance is critical for this operation.
924 * %REVIEW% Should this be exposed at the package/public layers?
926 * @param nodeHandle (external representation of node)
927 * @return nodeIdentity Internal offset to this node's records.
929 final public int makeNodeIdentity(int nodeHandle)
931 if(NULL==nodeHandle) return NULL;
933 if(m_mgrDefault!=null)
935 // Optimization: use the DTMManagerDefault's fast DTMID-to-offsets
936 // table. I'm not wild about this solution but this operation
937 // needs need extreme speed.
939 int whichDTMindex=nodeHandle>>>DTMManager.IDENT_DTM_NODE_BITS;
941 // %REVIEW% Wish I didn't have to perform the pre-test, but
942 // someone is apparently asking DTMs whether they contain nodes
943 // which really don't belong to them. That's probably a bug
944 // which should be fixed, but until it is:
945 if(m_mgrDefault.m_dtms[whichDTMindex]!=this)
949 m_mgrDefault.m_dtm_offsets[whichDTMindex]
950 | (nodeHandle & DTMManager.IDENT_NODE_DEFAULT);
953 int whichDTMid=m_dtmIdent.indexOf(nodeHandle & DTMManager.IDENT_DTM_DEFAULT);
954 return (whichDTMid==NULL)
956 : (whichDTMid << DTMManager.IDENT_DTM_NODE_BITS)
957 + (nodeHandle & DTMManager.IDENT_NODE_DEFAULT);
962 * Given a node handle, get the handle of the node's first child.
963 * If not yet resolved, waits for more nodes to be added to the document and
966 * @param nodeHandle int Handle of the node.
967 * @return int DTM node-number of first child, or DTM.NULL to indicate none exists.
969 public int getFirstChild(int nodeHandle)
972 int identity = makeNodeIdentity(nodeHandle);
973 int firstChild = _firstch(identity);
975 return makeNodeHandle(firstChild);
979 * Given a node handle, get the handle of the node's first child.
980 * If not yet resolved, waits for more nodes to be added to the document and
983 * @param nodeHandle int Handle of the node.
984 * @return int DTM node-number of first child, or DTM.NULL to indicate none exists.
986 public int getTypedFirstChild(int nodeHandle, int nodeType)
989 int firstChild, eType;
990 if (nodeType < DTM.NTYPES) {
991 for (firstChild = _firstch(makeNodeIdentity(nodeHandle));
992 firstChild != DTM.NULL;
993 firstChild = _nextsib(firstChild)) {
994 eType = _exptype(firstChild);
995 if (eType == nodeType
996 || (eType >= DTM.NTYPES
997 && m_expandedNameTable.getType(eType) == nodeType)) {
998 return makeNodeHandle(firstChild);
1002 for (firstChild = _firstch(makeNodeIdentity(nodeHandle));
1003 firstChild != DTM.NULL;
1004 firstChild = _nextsib(firstChild)) {
1005 if (_exptype(firstChild) == nodeType) {
1006 return makeNodeHandle(firstChild);
1014 * Given a node handle, advance to its last child.
1015 * If not yet resolved, waits for more nodes to be added to the document and
1018 * @param nodeHandle int Handle of the node.
1019 * @return int Node-number of last child,
1020 * or DTM.NULL to indicate none exists.
1022 public int getLastChild(int nodeHandle)
1025 int identity = makeNodeIdentity(nodeHandle);
1026 int child = _firstch(identity);
1027 int lastChild = DTM.NULL;
1029 while (child != DTM.NULL)
1032 child = _nextsib(child);
1035 return makeNodeHandle(lastChild);
1039 * Retrieves an attribute node by by qualified name and namespace URI.
1041 * @param nodeHandle int Handle of the node upon which to look up this attribute..
1042 * @param namespaceURI The namespace URI of the attribute to
1043 * retrieve, or null.
1044 * @param name The local name of the attribute to
1046 * @return The attribute node handle with the specified name (
1047 * <code>nodeName</code>) or <code>DTM.NULL</code> if there is no such
1050 public abstract int getAttributeNode(int nodeHandle, String namespaceURI,
1054 * Given a node handle, get the index of the node's first attribute.
1056 * @param nodeHandle int Handle of the node.
1057 * @return Handle of first attribute, or DTM.NULL to indicate none exists.
1059 public int getFirstAttribute(int nodeHandle)
1061 int nodeID = makeNodeIdentity(nodeHandle);
1063 return makeNodeHandle(getFirstAttributeIdentity(nodeID));
1067 * Given a node identity, get the index of the node's first attribute.
1069 * @param identity int identity of the node.
1070 * @return Identity of first attribute, or DTM.NULL to indicate none exists.
1072 protected int getFirstAttributeIdentity(int identity) {
1073 int type = _type(identity);
1075 if (DTM.ELEMENT_NODE == type)
1077 // Assume that attributes and namespaces immediately follow the element.
1078 while (DTM.NULL != (identity = getNextNodeIdentity(identity)))
1081 // Assume this can not be null.
1082 type = _type(identity);
1084 if (type == DTM.ATTRIBUTE_NODE)
1088 else if (DTM.NAMESPACE_NODE != type)
1099 * Given a node handle and an expanded type ID, get the index of the node's
1100 * attribute of that type, if any.
1102 * @param nodeHandle int Handle of the node.
1103 * @param attType int expanded type ID of the required attribute.
1104 * @return Handle of attribute of the required type, or DTM.NULL to indicate
1107 protected int getTypedAttribute(int nodeHandle, int attType) {
1108 int type = getNodeType(nodeHandle);
1109 if (DTM.ELEMENT_NODE == type) {
1110 int identity = makeNodeIdentity(nodeHandle);
1112 while (DTM.NULL != (identity = getNextNodeIdentity(identity)))
1114 type = _type(identity);
1116 if (type == DTM.ATTRIBUTE_NODE)
1118 if (_exptype(identity) == attType) return makeNodeHandle(identity);
1120 else if (DTM.NAMESPACE_NODE != type)
1131 * Given a node handle, advance to its next sibling.
1132 * If not yet resolved, waits for more nodes to be added to the document and
1134 * @param nodeHandle int Handle of the node.
1135 * @return int Node-number of next sibling,
1136 * or DTM.NULL to indicate none exists.
1138 public int getNextSibling(int nodeHandle)
1140 if (nodeHandle == DTM.NULL)
1142 return makeNodeHandle(_nextsib(makeNodeIdentity(nodeHandle)));
1146 * Given a node handle, advance to its next sibling.
1147 * If not yet resolved, waits for more nodes to be added to the document and
1149 * @param nodeHandle int Handle of the node.
1150 * @return int Node-number of next sibling,
1151 * or DTM.NULL to indicate none exists.
1153 public int getTypedNextSibling(int nodeHandle, int nodeType)
1155 if (nodeHandle == DTM.NULL)
1157 int node = makeNodeIdentity(nodeHandle);
1159 while ((node = _nextsib(node)) != DTM.NULL &&
1160 ((eType = _exptype(node)) != nodeType &&
1161 m_expandedNameTable.getType(eType)!= nodeType));
1162 //_type(node) != nodeType));
1164 return (node == DTM.NULL ? DTM.NULL : makeNodeHandle(node));
1168 * Given a node handle, find its preceeding sibling.
1169 * WARNING: DTM is asymmetric; this operation is resolved by search, and is
1170 * relatively expensive.
1172 * @param nodeHandle the id of the node.
1173 * @return int Node-number of the previous sib,
1174 * or DTM.NULL to indicate none exists.
1176 public int getPreviousSibling(int nodeHandle)
1178 if (nodeHandle == DTM.NULL)
1181 if (m_prevsib != null)
1182 return makeNodeHandle(_prevsib(makeNodeIdentity(nodeHandle)));
1185 // If the previous sibling array is not built, we get at
1186 // the previous sibling using the parent, firstch and
1188 int nodeID = makeNodeIdentity(nodeHandle);
1189 int parent = _parent(nodeID);
1190 int node = _firstch(parent);
1191 int result = DTM.NULL;
1192 while (node != nodeID)
1195 node = _nextsib(node);
1197 return makeNodeHandle(result);
1202 * Given a node handle, advance to the next attribute.
1203 * If an attr, we advance to
1204 * the next attr on the same node. If not an attribute, we return NULL.
1206 * @param nodeHandle int Handle of the node.
1207 * @return int DTM node-number of the resolved attr,
1208 * or DTM.NULL to indicate none exists.
1210 public int getNextAttribute(int nodeHandle) {
1211 int nodeID = makeNodeIdentity(nodeHandle);
1213 if (_type(nodeID) == DTM.ATTRIBUTE_NODE) {
1214 return makeNodeHandle(getNextAttributeIdentity(nodeID));
1221 * Given a node identity for an attribute, advance to the next attribute.
1223 * @param identity int identity of the attribute node. This
1224 * <strong>must</strong> be an attribute node.
1226 * @return int DTM node-identity of the resolved attr,
1227 * or DTM.NULL to indicate none exists.
1230 protected int getNextAttributeIdentity(int identity) {
1231 // Assume that attributes and namespace nodes immediately follow the element
1232 while (DTM.NULL != (identity = getNextNodeIdentity(identity))) {
1233 int type = _type(identity);
1235 if (type == DTM.ATTRIBUTE_NODE) {
1237 } else if (type != DTM.NAMESPACE_NODE) {
1245 /** Lazily created namespace lists. */
1246 private Vector m_namespaceLists = null; // on demand
1249 /** Build table of namespace declaration
1250 * locations during DTM construction. Table is a Vector of
1251 * SuballocatedIntVectors containing the namespace node HANDLES declared at
1252 * that ID, plus an SuballocatedIntVector of the element node INDEXES at which
1253 * these declarations appeared.
1255 * NOTE: Since this occurs during model build, nodes will be encountered
1256 * in doucment order and thus the table will be ordered by element,
1257 * permitting binary-search as a possible retrieval optimization.
1259 * %REVIEW% Directly managed arrays rather than vectors?
1260 * %REVIEW% Handles or IDs? Given usage, I think handles.
1262 protected void declareNamespaceInContext(int elementNodeIndex,int namespaceNodeIndex)
1264 SuballocatedIntVector nsList=null;
1265 if(m_namespaceDeclSets==null)
1269 m_namespaceDeclSetElements=new SuballocatedIntVector(32);
1270 m_namespaceDeclSetElements.addElement(elementNodeIndex);
1271 m_namespaceDeclSets=new Vector();
1272 nsList=new SuballocatedIntVector(32);
1273 m_namespaceDeclSets.addElement(nsList);
1277 // Most recent. May be -1 (none) if DTM was pruned.
1278 // %OPT% Is there a lastElement() method? Should there be?
1279 int last=m_namespaceDeclSetElements.size()-1;
1281 if(last>=0 && elementNodeIndex==m_namespaceDeclSetElements.elementAt(last))
1283 nsList=(SuballocatedIntVector)m_namespaceDeclSets.elementAt(last);
1288 m_namespaceDeclSetElements.addElement(elementNodeIndex);
1290 SuballocatedIntVector inherited =
1291 findNamespaceContext(_parent(elementNodeIndex));
1293 if (inherited!=null) {
1294 // %OPT% Count-down might be faster, but debuggability may
1295 // be better this way, and if we ever decide we want to
1296 // keep this ordered by expanded-type...
1297 int isize=inherited.size();
1299 // Base the size of a new namespace list on the
1300 // size of the inherited list - but within reason!
1301 nsList=new SuballocatedIntVector(Math.max(Math.min(isize+16,2048),
1304 for(int i=0;i<isize;++i)
1306 nsList.addElement(inherited.elementAt(i));
1309 nsList=new SuballocatedIntVector(32);
1312 m_namespaceDeclSets.addElement(nsList);
1315 // Handle overwriting inherited.
1316 // %OPT% Keep sorted? (By expanded-name rather than by doc order...)
1317 // Downside: Would require insertElementAt if not found,
1318 // which has recopying costs. But these are generally short lists...
1319 int newEType=_exptype(namespaceNodeIndex);
1321 for(int i=nsList.size()-1;i>=0;--i)
1323 if(newEType==getExpandedTypeID(nsList.elementAt(i)))
1325 nsList.setElementAt(makeNodeHandle(namespaceNodeIndex),i);
1329 nsList.addElement(makeNodeHandle(namespaceNodeIndex));
1332 /** Retrieve list of namespace declaration locations
1333 * active at this node. List is an SuballocatedIntVector whose
1334 * entries are the namespace node HANDLES declared at that ID.
1336 * %REVIEW% Directly managed arrays rather than vectors?
1337 * %REVIEW% Handles or IDs? Given usage, I think handles.
1339 protected SuballocatedIntVector findNamespaceContext(int elementNodeIndex)
1341 if (null!=m_namespaceDeclSetElements)
1343 // %OPT% Is binary-search really saving us a lot versus linear?
1344 // (... It may be, in large docs with many NS decls.)
1345 int wouldBeAt=findInSortedSuballocatedIntVector(m_namespaceDeclSetElements,
1347 if(wouldBeAt>=0) // Found it
1348 return (SuballocatedIntVector) m_namespaceDeclSets.elementAt(wouldBeAt);
1349 if(wouldBeAt == -1) // -1-wouldbeat == 0
1350 return null; // Not after anything; definitely not found
1352 // Not found, but we know where it should have been.
1353 // Search back until we find an ancestor or run out.
1354 wouldBeAt=-1-wouldBeAt;
1356 // Decrement wouldBeAt to find last possible ancestor
1357 int candidate=m_namespaceDeclSetElements.elementAt(-- wouldBeAt);
1358 int ancestor=_parent(elementNodeIndex);
1360 // Special case: if the candidate is before the given node, and
1361 // is in the earliest possible position in the document, it
1362 // must have the namespace declarations we're interested in.
1363 if (wouldBeAt == 0 && candidate < ancestor) {
1364 int rootHandle = getDocumentRoot(makeNodeHandle(elementNodeIndex));
1365 int rootID = makeNodeIdentity(rootHandle);
1366 int uppermostNSCandidateID;
1368 if (getNodeType(rootHandle) == DTM.DOCUMENT_NODE) {
1369 int ch = _firstch(rootID);
1370 uppermostNSCandidateID = (ch != DTM.NULL) ? ch : rootID;
1372 uppermostNSCandidateID = rootID;
1375 if (candidate == uppermostNSCandidateID) {
1376 return (SuballocatedIntVector)m_namespaceDeclSets.elementAt(wouldBeAt);
1380 while(wouldBeAt>=0 && ancestor>0)
1382 if (candidate==ancestor) {
1383 // Found ancestor in list
1384 return (SuballocatedIntVector)m_namespaceDeclSets.elementAt(wouldBeAt);
1385 } else if (candidate<ancestor) {
1388 ancestor=_parent(ancestor);
1389 } while (candidate < ancestor);
1390 } else if(wouldBeAt > 0){
1392 candidate=m_namespaceDeclSetElements.elementAt(--wouldBeAt);
1399 return null; // No namespaces known at this node
1403 * Subroutine: Locate the specified node within
1404 * m_namespaceDeclSetElements, or the last element which
1405 * preceeds it in document order
1407 * %REVIEW% Inlne this into findNamespaceContext? Create SortedSuballocatedIntVector type?
1409 * @return If positive or zero, the index of the found item.
1410 * If negative, index of the point at which it would have appeared,
1411 * encoded as -1-index and hence reconvertable by subtracting
1412 * it from -1. (Encoding because I don't want to recompare the strings
1413 * but don't want to burn bytes on a datatype to hold a flagged value.)
1415 protected int findInSortedSuballocatedIntVector(SuballocatedIntVector vector, int lookfor)
1419 if(vector != null) {
1421 int last = vector.size() - 1;
1423 while (first <= last) {
1424 i = (first + last) / 2;
1425 int test = lookfor-vector.elementAt(i);
1427 return i; // Name found
1429 else if (test < 0) {
1430 last = i - 1; // looked too late
1433 first = i + 1; // looked ot early
1438 i = first; // Clean up at loop end
1442 return -1 - i; // not-found has to be encoded.
1447 * Given a node handle, get the index of the node's first child.
1448 * If not yet resolved, waits for more nodes to be added to the document and
1451 * @param nodeHandle handle to node, which should probably be an element
1452 * node, but need not be.
1454 * @param inScope true if all namespaces in scope should be returned,
1455 * false if only the namespace declarations should be
1457 * @return handle of first namespace, or DTM.NULL to indicate none exists.
1459 public int getFirstNamespaceNode(int nodeHandle, boolean inScope)
1463 int identity = makeNodeIdentity(nodeHandle);
1464 if (_type(identity) == DTM.ELEMENT_NODE)
1466 SuballocatedIntVector nsContext=findNamespaceContext(identity);
1467 if(nsContext==null || nsContext.size()<1)
1470 return nsContext.elementAt(0);
1477 // Assume that attributes and namespaces immediately
1478 // follow the element.
1480 // %OPT% Would things be faster if all NS nodes were built
1481 // before all Attr nodes? Some costs at build time for 2nd
1483 int identity = makeNodeIdentity(nodeHandle);
1484 if (_type(identity) == DTM.ELEMENT_NODE)
1486 while (DTM.NULL != (identity = getNextNodeIdentity(identity)))
1488 int type = _type(identity);
1489 if (type == DTM.NAMESPACE_NODE)
1490 return makeNodeHandle(identity);
1491 else if (DTM.ATTRIBUTE_NODE != type)
1502 * Given a namespace handle, advance to the next namespace.
1504 * @param baseHandle handle to original node from where the first namespace
1505 * was relative to (needed to return nodes in document order).
1506 * @param nodeHandle A namespace handle for which we will find the next node.
1507 * @param inScope true if all namespaces that are in scope should be processed,
1508 * otherwise just process the nodes in the given element handle.
1509 * @return handle of next namespace, or DTM.NULL to indicate none exists.
1511 public int getNextNamespaceNode(int baseHandle, int nodeHandle,
1516 //Since we've been given the base, try direct lookup
1517 //(could look from nodeHandle but this is at least one
1518 //comparison/get-parent faster)
1519 //SuballocatedIntVector nsContext=findNamespaceContext(nodeHandle & m_mask);
1521 SuballocatedIntVector nsContext=findNamespaceContext(makeNodeIdentity(baseHandle));
1525 int i=1 + nsContext.indexOf(nodeHandle);
1526 if(i<=0 || i==nsContext.size())
1529 return nsContext.elementAt(i);
1533 // Assume that attributes and namespace nodes immediately follow the element.
1534 int identity = makeNodeIdentity(nodeHandle);
1535 while (DTM.NULL != (identity = getNextNodeIdentity(identity)))
1537 int type = _type(identity);
1538 if (type == DTM.NAMESPACE_NODE)
1540 return makeNodeHandle(identity);
1542 else if (type != DTM.ATTRIBUTE_NODE)
1552 * Given a node handle, find its parent node.
1554 * @param nodeHandle the id of the node.
1555 * @return int Node-number of parent,
1556 * or DTM.NULL to indicate none exists.
1558 public int getParent(int nodeHandle)
1561 int identity = makeNodeIdentity(nodeHandle);
1564 return makeNodeHandle(_parent(identity));
1570 * Find the Document node handle for the document currently under construction.
1571 * PLEASE NOTE that most people should use getOwnerDocument(nodeHandle) instead;
1572 * this version of the operation is primarily intended for use during negotiation
1573 * with the DTM Manager.
1575 * @return int Node handle of document, which should always be valid.
1577 public int getDocument()
1579 return m_dtmIdent.elementAt(0); // makeNodeHandle(0)
1583 * Given a node handle, find the owning document node. This has the exact
1584 * same semantics as the DOM Document method of the same name, in that if
1585 * the nodeHandle is a document node, it will return NULL.
1587 * <p>%REVIEW% Since this is DOM-specific, it may belong at the DOM
1588 * binding layer. Included here as a convenience function and to
1589 * aid porting of DOM code to DTM.</p>
1591 * @param nodeHandle the id of the node.
1592 * @return int Node handle of owning document, or -1 if the node was a Docment
1594 public int getOwnerDocument(int nodeHandle)
1597 if (DTM.DOCUMENT_NODE == getNodeType(nodeHandle))
1600 return getDocumentRoot(nodeHandle);
1604 * Given a node handle, find the owning document node. Unlike the DOM,
1605 * this considers the owningDocument of a Document to be itself.
1607 * @param nodeHandle the id of the node.
1608 * @return int Node handle of owning document, or the nodeHandle if it is
1611 public int getDocumentRoot(int nodeHandle)
1613 return getManager().getDTM(nodeHandle).getDocument();
1617 * Get the string-value of a node as a String object
1618 * (see http://www.w3.org/TR/xpath#data-model
1619 * for the definition of a node's string-value).
1621 * @param nodeHandle The node ID.
1623 * @return A string object that represents the string-value of the given node.
1625 public abstract XMLString getStringValue(int nodeHandle);
1628 * Get number of character array chunks in
1629 * the string-value of a node.
1630 * (see http://www.w3.org/TR/xpath#data-model
1631 * for the definition of a node's string-value).
1632 * Note that a single text node may have multiple text chunks.
1634 * @param nodeHandle The node ID.
1636 * @return number of character array chunks in
1637 * the string-value of a node.
1639 public int getStringValueChunkCount(int nodeHandle)
1643 error(XMLMessages.createXMLMessage(XMLErrorResources.ER_METHOD_NOT_SUPPORTED, null));//("getStringValueChunkCount not yet supported!");
1649 * Get a character array chunk in the string-value of a node.
1650 * (see http://www.w3.org/TR/xpath#data-model
1651 * for the definition of a node's string-value).
1652 * Note that a single text node may have multiple text chunks.
1654 * @param nodeHandle The node ID.
1655 * @param chunkIndex Which chunk to get.
1656 * @param startAndLen An array of 2 where the start position and length of
1657 * the chunk will be returned.
1659 * @return The character array reference where the chunk occurs.
1661 public char[] getStringValueChunk(int nodeHandle, int chunkIndex,
1666 error(XMLMessages.createXMLMessage(XMLErrorResources.ER_METHOD_NOT_SUPPORTED, null));//"getStringValueChunk not yet supported!");
1672 * Given a node handle, return an ID that represents the node's expanded name.
1674 * @param nodeHandle The handle to the node in question.
1676 * @return the expanded-name id of the node.
1678 public int getExpandedTypeID(int nodeHandle)
1680 // %REVIEW% This _should_ only be null if someone asked the wrong DTM about the node...
1681 // which one would hope would never happen...
1682 int id=makeNodeIdentity(nodeHandle);
1685 return _exptype(id);
1689 * Given an expanded name, return an ID. If the expanded-name does not
1690 * exist in the internal tables, the entry will be created, and the ID will
1691 * be returned. Any additional nodes that are created that have this
1692 * expanded name will use this ID.
1694 * @param type The simple type, i.e. one of ELEMENT, ATTRIBUTE, etc.
1696 * @param namespace The namespace URI, which may be null, may be an empty
1697 * string (which will be the same as null), or may be a
1699 * @param localName The local name string, which must be a valid
1700 * <a href="http://www.w3.org/TR/REC-xml-names/">NCName</a>.
1702 * @return the expanded-name id of the node.
1704 public int getExpandedTypeID(String namespace, String localName, int type)
1707 ExpandedNameTable ent = m_expandedNameTable;
1709 return ent.getExpandedTypeID(namespace, localName, type);
1713 * Given an expanded-name ID, return the local name part.
1715 * @param expandedNameID an ID that represents an expanded-name.
1716 * @return String Local name of this node.
1718 public String getLocalNameFromExpandedNameID(int expandedNameID)
1720 return m_expandedNameTable.getLocalName(expandedNameID);
1724 * Given an expanded-name ID, return the namespace URI part.
1726 * @param expandedNameID an ID that represents an expanded-name.
1727 * @return String URI value of this node's namespace, or null if no
1728 * namespace was resolved.
1730 public String getNamespaceFromExpandedNameID(int expandedNameID)
1732 return m_expandedNameTable.getNamespace(expandedNameID);
1736 * Returns the namespace type of a specific node
1737 * @param nodeHandle the id of the node.
1738 * @return the ID of the namespace.
1740 public int getNamespaceType(final int nodeHandle)
1743 int identity = makeNodeIdentity(nodeHandle);
1744 int expandedNameID = _exptype(identity);
1746 return m_expandedNameTable.getNamespaceID(expandedNameID);
1750 * Given a node handle, return its DOM-style node name. This will
1751 * include names such as #text or #document.
1753 * @param nodeHandle the id of the node.
1754 * @return String Name of this node, which may be an empty string.
1755 * %REVIEW% Document when empty string is possible...
1756 * %REVIEW-COMMENT% It should never be empty, should it?
1758 public abstract String getNodeName(int nodeHandle);
1761 * Given a node handle, return the XPath node name. This should be
1762 * the name as described by the XPath data model, NOT the DOM-style
1765 * @param nodeHandle the id of the node.
1766 * @return String Name of this node, which may be an empty string.
1768 public String getNodeNameX(int nodeHandle)
1771 /** @todo: implement this org.apache.xml.dtm.DTMDefaultBase abstract method */
1772 error(XMLMessages.createXMLMessage(XMLErrorResources.ER_METHOD_NOT_SUPPORTED, null));//"Not yet supported!");
1778 * Given a node handle, return its XPath-style localname.
1779 * (As defined in Namespaces, this is the portion of the name after any
1782 * @param nodeHandle the id of the node.
1783 * @return String Local name of this node.
1785 public abstract String getLocalName(int nodeHandle);
1788 * Given a namespace handle, return the prefix that the namespace decl is
1790 * Given a node handle, return the prefix used to map to the namespace.
1792 * <p> %REVIEW% Are you sure you want "" for no prefix? </p>
1793 * <p> %REVIEW-COMMENT% I think so... not totally sure. -sb </p>
1795 * @param nodeHandle the id of the node.
1796 * @return String prefix of this node's name, or "" if no explicit
1797 * namespace prefix was given.
1799 public abstract String getPrefix(int nodeHandle);
1802 * Given a node handle, return its DOM-style namespace URI
1803 * (As defined in Namespaces, this is the declared URI which this node's
1804 * prefix -- or default in lieu thereof -- was mapped to.)
1806 * <p>%REVIEW% Null or ""? -sb</p>
1808 * @param nodeHandle the id of the node.
1809 * @return String URI value of this node's namespace, or null if no
1810 * namespace was resolved.
1812 public abstract String getNamespaceURI(int nodeHandle);
1815 * Given a node handle, return its node value. This is mostly
1816 * as defined by the DOM, but may ignore some conveniences.
1819 * @param nodeHandle The node id.
1820 * @return String Value of this node, or null if not
1821 * meaningful for this node type.
1823 public abstract String getNodeValue(int nodeHandle);
1826 * Given a node handle, return its DOM-style node type.
1828 * %REVIEW% Generally, returning short is false economy. Return int?
1829 * %REVIEW% Make assumption that node has already arrived. Is OK?
1831 * @param nodeHandle The node id.
1832 * @return int Node type, as per the DOM's Node._NODE constants.
1834 public short getNodeType(int nodeHandle)
1836 if (nodeHandle == DTM.NULL)
1838 return m_expandedNameTable.getType(_exptype(makeNodeIdentity(nodeHandle)));
1842 * Get the depth level of this node in the tree (equals 1 for
1843 * a parentless node).
1845 * @param nodeHandle The node id.
1846 * @return the number of ancestors, plus one
1847 * @xsl.usage internal
1849 public short getLevel(int nodeHandle)
1851 // Apparently, the axis walker stuff requires levels to count from 1.
1852 int identity = makeNodeIdentity(nodeHandle);
1853 return (short) (_level(identity) + 1);
1857 * Get the identity of this node in the tree
1859 * @param nodeHandle The node handle.
1860 * @return the node identity
1861 * @xsl.usage internal
1863 public int getNodeIdent(int nodeHandle)
1865 /*if (nodeHandle != DTM.NULL)
1866 return nodeHandle & m_mask;
1870 return makeNodeIdentity(nodeHandle);
1874 * Get the handle of this node in the tree
1876 * @param nodeId The node identity.
1877 * @return the node handle
1878 * @xsl.usage internal
1880 public int getNodeHandle(int nodeId)
1882 /*if (nodeId != DTM.NULL)
1883 return nodeId | m_dtmIdent;
1887 return makeNodeHandle(nodeId);
1890 // ============== Document query functions ==============
1893 * Tests whether DTM DOM implementation implements a specific feature and
1894 * that feature is supported by this node.
1896 * @param feature The name of the feature to test.
1897 * @param version This is the version number of the feature to test.
1898 * If the version is not
1899 * specified, supporting any version of the feature will cause the
1900 * method to return <code>true</code>.
1901 * @return Returns <code>true</code> if the specified feature is
1902 * supported on this node, <code>false</code> otherwise.
1904 public boolean isSupported(String feature, String version)
1912 * Return the base URI of the document entity. If it is not known
1913 * (because the document was parsed from a socket connection or from
1914 * standard input, for example), the value of this property is unknown.
1916 * @return the document base URI String object or null if unknown.
1918 public String getDocumentBaseURI()
1920 return m_documentBaseURI;
1924 * Set the base URI of the document entity.
1926 * @param baseURI the document base URI String object or null if unknown.
1928 public void setDocumentBaseURI(String baseURI)
1930 m_documentBaseURI = baseURI;
1934 * Return the system identifier of the document entity. If
1935 * it is not known, the value of this property is unknown.
1937 * @param nodeHandle The node id, which can be any valid node handle.
1938 * @return the system identifier String object or null if unknown.
1940 public String getDocumentSystemIdentifier(int nodeHandle)
1944 return m_documentBaseURI;
1948 * Return the name of the character encoding scheme
1949 * in which the document entity is expressed.
1951 * @param nodeHandle The node id, which can be any valid node handle.
1952 * @return the document encoding String object.
1953 * @xsl.usage internal
1955 public String getDocumentEncoding(int nodeHandle)
1958 // %REVIEW% OK?? -sb
1963 * Return an indication of the standalone status of the document,
1964 * either "yes" or "no". This property is derived from the optional
1965 * standalone document declaration in the XML declaration at the
1966 * beginning of the document entity, and has no value if there is no
1967 * standalone document declaration.
1969 * @param nodeHandle The node id, which can be any valid node handle.
1970 * @return the document standalone String object, either "yes", "no", or null.
1972 public String getDocumentStandalone(int nodeHandle)
1978 * Return a string representing the XML version of the document. This
1979 * property is derived from the XML declaration optionally present at the
1980 * beginning of the document entity, and has no value if there is no XML
1983 * @param documentHandle The document handle
1985 * @return the document version String object.
1987 public String getDocumentVersion(int documentHandle)
1993 * Return an indication of
1994 * whether the processor has read the complete DTD. Its value is a
1995 * boolean. If it is false, then certain properties (indicated in their
1996 * descriptions below) may be unknown. If it is true, those properties
1997 * are never unknown.
1999 * @return <code>true</code> if all declarations were processed;
2000 * <code>false</code> otherwise.
2002 public boolean getDocumentAllDeclarationsProcessed()
2010 * A document type declaration information item has the following properties:
2012 * 1. [system identifier] The system identifier of the external subset, if
2013 * it exists. Otherwise this property has no value.
2015 * @return the system identifier String object, or null if there is none.
2017 public abstract String getDocumentTypeDeclarationSystemIdentifier();
2020 * Return the public identifier of the external subset,
2021 * normalized as described in 4.2.2 External Entities [XML]. If there is
2022 * no external subset or if it has no public identifier, this property
2025 * @return the public identifier String object, or null if there is none.
2027 public abstract String getDocumentTypeDeclarationPublicIdentifier();
2030 * Returns the <code>Element</code> whose <code>ID</code> is given by
2031 * <code>elementId</code>. If no such element exists, returns
2032 * <code>DTM.NULL</code>. Behavior is not defined if more than one element
2033 * has this <code>ID</code>. Attributes (including those
2034 * with the name "ID") are not of type ID unless so defined by DTD/Schema
2035 * information available to the DTM implementation.
2036 * Implementations that do not know whether attributes are of type ID or
2037 * not are expected to return <code>DTM.NULL</code>.
2039 * <p>%REVIEW% Presumably IDs are still scoped to a single document,
2040 * and this operation searches only within a single document, right?
2041 * Wouldn't want collisions between DTMs in the same process.</p>
2043 * @param elementId The unique <code>id</code> value for an element.
2044 * @return The handle of the matching element.
2046 public abstract int getElementById(String elementId);
2049 * The getUnparsedEntityURI function returns the URI of the unparsed
2050 * entity with the specified name in the same document as the context
2051 * node (see [3.3 Unparsed Entities]). It returns the empty string if
2052 * there is no such entity.
2054 * XML processors may choose to use the System Identifier (if one
2055 * is provided) to resolve the entity, rather than the URI in the
2056 * Public Identifier. The details are dependent on the processor, and
2057 * we would have to support some form of plug-in resolver to handle
2058 * this properly. Currently, we simply return the System Identifier if
2059 * present, and hope that it a usable URI or that our caller can
2061 * TODO: Resolve Public Identifiers... or consider changing function name.
2063 * If we find a relative URI
2064 * reference, XML expects it to be resolved in terms of the base URI
2065 * of the document. The DOM doesn't do that for us, and it isn't
2066 * entirely clear whether that should be done here; currently that's
2067 * pushed up to a higher level of our application. (Note that DOM Level
2068 * 1 didn't store the document's base URI.)
2069 * TODO: Consider resolving Relative URIs.
2071 * (The DOM's statement that "An XML processor may choose to
2072 * completely expand entities before the structure model is passed
2073 * to the DOM" refers only to parsed entities, not unparsed, and hence
2074 * doesn't affect this function.)
2076 * @param name A string containing the Entity Name of the unparsed
2079 * @return String containing the URI of the Unparsed Entity, or an
2080 * empty string if no such entity exists.
2082 public abstract String getUnparsedEntityURI(String name);
2084 // ============== Boolean methods ================
2087 * Return true if the xsl:strip-space or xsl:preserve-space was processed
2088 * during construction of the DTM document.
2090 * @return true if this DTM supports prestripping.
2092 public boolean supportsPreStripping()
2098 * Figure out whether nodeHandle2 should be considered as being later
2099 * in the document than nodeHandle1, in Document Order as defined
2100 * by the XPath model. This may not agree with the ordering defined
2101 * by other XML applications.
2103 * There are some cases where ordering isn't defined, and neither are
2104 * the results of this function -- though we'll generally return false.
2106 * @param nodeHandle1 Node handle to perform position comparison on.
2107 * @param nodeHandle2 Second Node handle to perform position comparison on .
2109 * @return true if node1 comes before node2, otherwise return false.
2110 * You can think of this as
2111 * <code>(node1.documentOrderPosition <= node2.documentOrderPosition)</code>.
2113 public boolean isNodeAfter(int nodeHandle1, int nodeHandle2)
2115 // These return NULL if the node doesn't belong to this document.
2116 int index1 = makeNodeIdentity(nodeHandle1);
2117 int index2 = makeNodeIdentity(nodeHandle2);
2119 return index1!=NULL && index2!=NULL && index1 <= index2;
2123 * 2. [element content whitespace] A boolean indicating whether the
2124 * character is white space appearing within element content (see [XML],
2125 * 2.10 "White Space Handling"). Note that validating XML processors are
2126 * required by XML 1.0 to provide this information. If there is no
2127 * declaration for the containing element, this property has no value for
2128 * white space characters. If no declaration has been read, but the [all
2129 * declarations processed] property of the document information item is
2130 * false (so there may be an unread declaration), then the value of this
2131 * property is unknown for white space characters. It is always false for
2132 * characters that are not white space.
2134 * @param nodeHandle the node ID.
2135 * @return <code>true</code> if the character data is whitespace;
2136 * <code>false</code> otherwise.
2138 public boolean isCharacterElementContentWhitespace(int nodeHandle)
2146 * 10. [all declarations processed] This property is not strictly speaking
2147 * part of the infoset of the document. Rather it is an indication of
2148 * whether the processor has read the complete DTD. Its value is a
2149 * boolean. If it is false, then certain properties (indicated in their
2150 * descriptions below) may be unknown. If it is true, those properties
2151 * are never unknown.
2153 * @param documentHandle A node handle that must identify a document.
2154 * @return <code>true</code> if all declarations were processed;
2155 * <code>false</code> otherwise.
2157 public boolean isDocumentAllDeclarationsProcessed(int documentHandle)
2163 * 5. [specified] A flag indicating whether this attribute was actually
2164 * specified in the start-tag of its element, or was defaulted from the
2167 * @param attributeHandle The attribute handle in question.
2169 * @return <code>true</code> if the attribute was specified;
2170 * <code>false</code> if it was defaulted.
2172 public abstract boolean isAttributeSpecified(int attributeHandle);
2174 // ========== Direct SAX Dispatch, for optimization purposes ========
2178 * characters method on the passed ContentHandler for the
2179 * string-value of the given node (see http://www.w3.org/TR/xpath#data-model
2180 * for the definition of a node's string-value). Multiple calls to the
2181 * ContentHandler's characters methods may well occur for a single call to
2184 * @param nodeHandle The node ID.
2185 * @param ch A non-null reference to a ContentHandler.
2186 * @param normalize true if the content should be normalized according to
2187 * the rules for the XPath
2188 * <a href="http://www.w3.org/TR/xpath#function-normalize-space">normalize-space</a>
2191 * @throws org.xml.sax.SAXException
2193 public abstract void dispatchCharactersEvents(
2194 int nodeHandle, org.xml.sax.ContentHandler ch, boolean normalize)
2195 throws org.xml.sax.SAXException;
2198 * Directly create SAX parser events from a subtree.
2200 * @param nodeHandle The node ID.
2201 * @param ch A non-null reference to a ContentHandler.
2203 * @throws org.xml.sax.SAXException
2205 public abstract void dispatchToEvents(
2206 int nodeHandle, org.xml.sax.ContentHandler ch)
2207 throws org.xml.sax.SAXException;
2210 * Return an DOM node for the given node.
2212 * @param nodeHandle The node ID.
2214 * @return A node representation of the DTM node.
2216 public org.w3c.dom.Node getNode(int nodeHandle)
2218 return new DTMNodeProxy(this, nodeHandle);
2221 // ==== Construction methods (may not be supported by some implementations!) =====
2224 * Append a child to the end of the document. Please note that the node
2225 * is always cloned if it is owned by another document.
2227 * <p>%REVIEW% "End of the document" needs to be defined more clearly.
2228 * Does it become the last child of the Document? Of the root element?</p>
2230 * @param newChild Must be a valid new node handle.
2231 * @param clone true if the child should be cloned into the document.
2232 * @param cloneDepth if the clone argument is true, specifies that the
2233 * clone should include all it's children.
2235 public void appendChild(int newChild, boolean clone, boolean cloneDepth)
2237 error(XMLMessages.createXMLMessage(XMLErrorResources.ER_METHOD_NOT_SUPPORTED, null));//"appendChild not yet supported!");
2241 * Append a text node child that will be constructed from a string,
2242 * to the end of the document.
2244 * <p>%REVIEW% "End of the document" needs to be defined more clearly.
2245 * Does it become the last child of the Document? Of the root element?</p>
2247 * @param str Non-null reverence to a string.
2249 public void appendTextChild(String str)
2251 error(XMLMessages.createXMLMessage(XMLErrorResources.ER_METHOD_NOT_SUPPORTED, null));//"appendTextChild not yet supported!");
2255 * Simple error for asserts and the like.
2257 * @param msg Error message to report.
2259 protected void error(String msg)
2261 throw new DTMException(msg);
2265 * Find out whether or not to strip whispace nodes.
2268 * @return whether or not to strip whispace nodes.
2270 protected boolean getShouldStripWhitespace()
2272 return m_shouldStripWS;
2276 * Set whether to strip whitespaces and push in current value of
2277 * m_shouldStripWS in m_shouldStripWhitespaceStack.
2279 * @param shouldStrip Flag indicating whether to strip whitespace nodes
2281 protected void pushShouldStripWhitespace(boolean shouldStrip)
2284 m_shouldStripWS = shouldStrip;
2286 if (null != m_shouldStripWhitespaceStack)
2287 m_shouldStripWhitespaceStack.push(shouldStrip);
2291 * Set whether to strip whitespaces at this point by popping out
2292 * m_shouldStripWhitespaceStack.
2295 protected void popShouldStripWhitespace()
2297 if (null != m_shouldStripWhitespaceStack)
2298 m_shouldStripWS = m_shouldStripWhitespaceStack.popAndTop();
2302 * Set whether to strip whitespaces and set the top of the stack to
2303 * the current value of m_shouldStripWS.
2306 * @param shouldStrip Flag indicating whether to strip whitespace nodes
2308 protected void setShouldStripWhitespace(boolean shouldStrip)
2311 m_shouldStripWS = shouldStrip;
2313 if (null != m_shouldStripWhitespaceStack)
2314 m_shouldStripWhitespaceStack.setTop(shouldStrip);
2318 * A dummy routine to satisify the abstract interface. If the DTM
2319 * implememtation that extends the default base requires notification
2320 * of registration, they can override this method.
2322 public void documentRegistration()
2327 * A dummy routine to satisify the abstract interface. If the DTM
2328 * implememtation that extends the default base requires notification
2329 * when the document is being released, they can override this method
2331 public void documentRelease()
2336 * Migrate a DTM built with an old DTMManager to a new DTMManager.
2337 * After the migration, the new DTMManager will treat the DTM as
2338 * one that is built by itself.
2339 * This is used to support DTM sharing between multiple transformations.
2340 * @param mgr the DTMManager
2342 public void migrateTo(DTMManager mgr)
2345 if(mgr instanceof DTMManagerDefault)
2346 m_mgrDefault=(DTMManagerDefault)mgr;
2349 /** Query which DTMManager this DTM is currently being handled by.
2351 * %REVEW% Should this become part of the base DTM API?
2353 * @return a DTMManager, or null if this is a "stand-alone" DTM.
2355 public DTMManager getManager()
2360 /** Query which DTMIDs this DTM is currently using within the DTMManager.
2362 * %REVEW% Should this become part of the base DTM API?
2364 * @return an IntVector, or null if this is a "stand-alone" DTM.
2366 public SuballocatedIntVector getDTMIDs()
2368 if(m_mgr==null) return null;