dom2dtm.java

来自「JAVA 所有包」· Java 代码 · 共 1,764 行 · 第 1/4 页

        
    // Did we run out of the tree?
    if(next==null)
      {
        m_nextsib.setElementAt(NULL,0);
        m_nodesAreProcessed = true;
        m_pos=null;
                
        if(JJK_DEBUG)
          {
            System.out.println("***** DOM2DTM Crosscheck:");
            for(int i=0;i<m_nodes.size();++i)
              System.out.println(i+":\t"+m_firstch.elementAt(i)+"\t"+m_nextsib.elementAt(i));
          }
                
        return false;
      }

    // Text needs some special handling:
    //
    // DTM may skip whitespace. This is handled by the suppressNode flag, which
    // when true will keep the DTM node from being created.
    //
    // DTM only directly records the first DOM node of any logically-contiguous
    // sequence. The lastTextNode value will be set to the last node in the 
    // contiguous sequence, and -- AFTER the DTM addNode -- can be used to 
    // advance next over this whole block. Should be simpler than special-casing
    // the above loop for "Was the logically-preceeding sibling a text node".
    // 
    // Finally, a DTM node should be considered a CDATASection only if all the
    // contiguous text it covers is CDATASections. The first Text should
    // force DTM to Text.
        
    boolean suppressNode=false;
    Node lastTextNode=null;

    nexttype=next.getNodeType();
        
    // nexttype=pos.getNodeType();
    if(TEXT_NODE == nexttype || CDATA_SECTION_NODE == nexttype)
      {
        // If filtering, initially assume we're going to suppress the node
        suppressNode=((null != m_wsfilter) && getShouldStripWhitespace());

        // Scan logically contiguous text (siblings, plus "flattening"
        // of entity reference boundaries).
        Node n=next;
        while(n!=null)
          {
            lastTextNode=n;
            // Any Text node means DTM considers it all Text
            if(TEXT_NODE == n.getNodeType())
              nexttype=TEXT_NODE;
            // Any non-whitespace in this sequence blocks whitespace
            // suppression
            suppressNode &=
              XMLCharacterRecognizer.isWhiteSpace(n.getNodeValue());
                        
            n=logicalNextDOMTextNode(n);
          }
      }
        
    // Special handling for PIs: Some DOMs represent the XML
    // Declaration as a PI. This is officially incorrect, per the DOM
    // spec, but is considered a "wrong but tolerable" temporary
    // workaround pending proper handling of these fields in DOM Level
    // 3. We want to recognize and reject that case.
    else if(PROCESSING_INSTRUCTION_NODE==nexttype)
      {
        suppressNode = (pos.getNodeName().toLowerCase().equals("xml"));
      }
        
        
    if(!suppressNode)
      {
        // Inserting next. NOTE that we force the node type; for
        // coalesced Text, this records CDATASections adjacent to
        // ordinary Text as Text.
        int nextindex=addNode(next,m_last_parent,m_last_kid,
			      nexttype);
	
        m_last_kid=nextindex;

        if(ELEMENT_NODE == nexttype)
          {
            int attrIndex=NULL; // start with no previous sib
            // Process attributes _now_, rather than waiting.
            // Simpler control flow, makes NS cache available immediately.
            NamedNodeMap attrs=next.getAttributes();
            int attrsize=(attrs==null) ? 0 : attrs.getLength();
            if(attrsize>0)
              {
                for(int i=0;i<attrsize;++i)
                  {
                    // No need to force nodetype in this case;
                    // addNode() will take care of switching it from
                    // Attr to Namespace if necessary.
                    attrIndex=addNode(attrs.item(i),
                                      nextindex,attrIndex,NULL);
                    m_firstch.setElementAt(DTM.NULL,attrIndex);

                    // If the xml: prefix is explicitly declared
                    // we don't need to synthesize one.
		    //
		    // NOTE that XML Namespaces were not originally
		    // defined as being namespace-aware (grrr), and
		    // while the W3C is planning to fix this it's
		    // safer for now to test the QName and trust the
		    // parsers to prevent anyone from redefining the
		    // reserved xmlns: prefix
                    if(!m_processedFirstElement
                       && "xmlns:xml".equals(attrs.item(i).getNodeName()))
                      m_processedFirstElement=true; 
                  }
                // Terminate list of attrs, and make sure they aren't
                // considered children of the element
              } // if attrs exist
            if(!m_processedFirstElement)
            {
              // The DOM might not have an explicit declaration for the
              // implicit "xml:" prefix, but the XPath data model
              // requires that this appear as a Namespace Node so we
              // have to synthesize one. You can think of this as
              // being a default attribute defined by the XML
              // Namespaces spec rather than by the DTD.
              attrIndex=addNode(new DOM2DTMdefaultNamespaceDeclarationNode(
																	(Element)next,"xml",NAMESPACE_DECL_NS,
																	makeNodeHandle(((attrIndex==NULL)?nextindex:attrIndex)+1)
																	),
                                nextindex,attrIndex,NULL);      
              m_firstch.setElementAt(DTM.NULL,attrIndex);
              m_processedFirstElement=true;
            }
            if(attrIndex!=NULL)
              m_nextsib.setElementAt(DTM.NULL,attrIndex);
          } //if(ELEMENT_NODE)
      } // (if !suppressNode)

    // Text postprocessing: Act on values stored above
    if(TEXT_NODE == nexttype || CDATA_SECTION_NODE == nexttype)
      {
        // %TBD% If nexttype was forced to TEXT, patch the DTM node
                
        next=lastTextNode;      // Advance the DOM cursor over contiguous text
      }
        
    // Remember where we left off.
    m_pos=next;
    return true;
  }  


  /**
   * Return an DOM node for the given node.
   *
   * @param nodeHandle The node ID.
   *
   * @return A node representation of the DTM node.
   */
  public Node getNode(int nodeHandle)
  {

    int identity = makeNodeIdentity(nodeHandle);

    return (Node) m_nodes.elementAt(identity);
  }

  /**
   * Get a Node from an identity index.
   *
   * NEEDSDOC @param nodeIdentity
   *
   * NEEDSDOC ($objectName$) @return
   */
  protected Node lookupNode(int nodeIdentity)
  {
    return (Node) m_nodes.elementAt(nodeIdentity);
  }

  /**
   * Get the next node identity value in the list, and call the iterator
   * if it hasn't been added yet.
   *
   * @param identity The node identity (index).
   * @return identity+1, or DTM.NULL.
   */
  protected int getNextNodeIdentity(int identity)
  {

    identity += 1;

    if (identity >= m_nodes.size())
    {
      if (!nextNode())
        identity = DTM.NULL;
    }

    return identity;
  }

  /**
   * Get the handle from a Node.
   * <p>%OPT% This will be pretty slow.</p>
   *
   * <p>%OPT% An XPath-like search (walk up DOM to root, tracking path;
   * walk down DTM reconstructing path) might be considerably faster
   * on later nodes in large documents. That might also imply improving
   * this call to handle nodes which would be in this DTM but
   * have not yet been built, which might or might not be a Good Thing.</p>
   * 
   * %REVIEW% This relies on being able to test node-identity via
   * object-identity. DTM2DOM proxying is a great example of a case where
   * that doesn't work. DOM Level 3 will provide the isSameNode() method
   * to fix that, but until then this is going to be flaky.
   *
   * @param node A node, which may be null.
   *
   * @return The node handle or <code>DTM.NULL</code>.
   */
  private int getHandleFromNode(Node node)
  {
    if (null != node)
    {
      int len = m_nodes.size();        
      boolean isMore;
      int i = 0;
      do
      {          
        for (; i < len; i++)
        {
          if (m_nodes.elementAt(i) == node)
            return makeNodeHandle(i);
        }

        isMore = nextNode();
  
        len = m_nodes.size();
            
      } 
      while(isMore || i < len);
    }
    
    return DTM.NULL;
  }

  /** Get the handle from a Node. This is a more robust version of
   * getHandleFromNode, intended to be usable by the public.
   *
   * <p>%OPT% This will be pretty slow.</p>
   * 
   * %REVIEW% This relies on being able to test node-identity via
   * object-identity. DTM2DOM proxying is a great example of a case where
   * that doesn't work. DOM Level 3 will provide the isSameNode() method
   * to fix that, but until then this is going to be flaky.
   *
   * @param node A node, which may be null.
   *
   * @return The node handle or <code>DTM.NULL</code>.  */
  public int getHandleOfNode(Node node)
  {
    if (null != node)
    {
      // Is Node actually within the same document? If not, don't search!
      // This would be easier if m_root was always the Document node, but
      // we decided to allow wrapping a DTM around a subtree.
      if((m_root==node) ||
         (m_root.getNodeType()==DOCUMENT_NODE &&
          m_root==node.getOwnerDocument()) ||
         (m_root.getNodeType()!=DOCUMENT_NODE &&
          m_root.getOwnerDocument()==node.getOwnerDocument())
         )
        {
          // If node _is_ in m_root's tree, find its handle
          //
          // %OPT% This check may be improved significantly when DOM
          // Level 3 nodeKey and relative-order tests become
          // available!
          for(Node cursor=node;
              cursor!=null;
              cursor=
                (cursor.getNodeType()!=ATTRIBUTE_NODE)
                ? cursor.getParentNode()
                : ((org.w3c.dom.Attr)cursor).getOwnerElement())
            {
              if(cursor==m_root)
                // We know this node; find its handle.
                return getHandleFromNode(node); 
            } // for ancestors of node
        } // if node and m_root in same Document
    } // if node!=null

    return DTM.NULL;
  }

  /**
   * Retrieves an attribute node by by qualified name and namespace URI.
   *
   * @param nodeHandle int Handle of the node upon which to look up this attribute..
   * @param namespaceURI The namespace URI of the attribute to
   *   retrieve, or null.
   * @param name The local name of the attribute to
   *   retrieve.
   * @return The attribute node handle with the specified name (
   *   <code>nodeName</code>) or <code>DTM.NULL</code> if there is no such
   *   attribute.
   */
  public int getAttributeNode(int nodeHandle, String namespaceURI,
                              String name)
  {

    // %OPT% This is probably slower than it needs to be.
    if (null == namespaceURI)
      namespaceURI = "";

    int type = getNodeType(nodeHandle);

    if (DTM.ELEMENT_NODE == type)
    {

      // Assume that attributes immediately follow the element.
      int identity = makeNodeIdentity(nodeHandle);

      while (DTM.NULL != (identity = getNextNodeIdentity(identity)))
      {
        // Assume this can not be null.
        type = _type(identity);

				// %REVIEW%
				// Should namespace nodes be retrievable DOM-style as attrs?
				// If not we need a separate function... which may be desirable
				// architecturally, but which is ugly from a code point of view.
				// (If we REALLY insist on it, this code should become a subroutine
				// of both -- retrieve the node, then test if the type matches
				// what you're looking for.)
        if (type == DTM.ATTRIBUTE_NODE || type==DTM.NAMESPACE_NODE)
        {
          Node node = lookupNode(identity);
          String nodeuri = node.getNamespaceURI();

          if (null == nodeuri)
            nodeuri = "";

          String nodelocalname = node.getLocalName();

          if (nodeuri.equals(namespaceURI) && name.equals(nodelocalname))
            return makeNodeHandle(identity);
        }
				
        else // if (DTM.NAMESPACE_NODE != type)
        {
          break;
        }
      }
    }

    return DTM.NULL;
  }

  /**
   * Get the string-value of a node as a String object
   * (see http://www.w3.org/TR/xpath#data-model
   * for the definition of a node's string-value).
   *
   * @param nodeHandle The node ID.
   *
   * @return A string object that represents the string-value of the given node.
   */
  public XMLString getStringValue(int nodeHandle)
  {

    int type = getNodeType(nodeHandle);
    Node node = getNode(nodeHandle);
    // %TBD% If an element only has one text node, we should just use it 
    // directly.
    if(DTM.ELEMENT_NODE == type || DTM.DOCUMENT_NODE == type 
    || DTM.DOCUMENT_FRAGMENT_NODE == type)
    {
      FastStringBuffer buf = StringBufferPool.get();
      String s;
  
      try
      {
        getNodeData(node, buf);
  
        s = (buf.length() > 0) ? buf.toString() : "";
      }
      finally
      {
        StringBufferPool.free(buf);
      }
  
      return m_xstrf.newstr( s );
    }
    else if(TEXT_NODE == type || CDATA_SECTION_NODE == type)
    {
      // If this is a DTM text node, it may be made of multiple DOM text
      // nodes -- including navigating into Entity References. DOM2DTM
      // records the first node in the sequence and requires that we
      // pick up the others when we retrieve the DTM node's value.
      //
      // %REVIEW% DOM Level 3 is expected to add a "whole text"
      // retrieval method which performs this function for us.
      FastStringBuffer buf = StringBufferPool.get();
      while(node!=null)
      {
        buf.append(node.getNodeValue());
        node=logicalNextDOMTextNode(node);
      }
      String s=(buf.length() > 0) ? buf.toString() : "";
      StringBufferPool.free(buf);
      return m_xstrf.newstr( s );
    }
    else
      return m_xstrf.newstr( node.getNodeValue() );
  }
  
  /**
   * Determine if the string-value of a node is whitespace
   *
   * @param nodeHandle The node Handle.
   *
   * @return Return true if the given node is whitespace.
   */
  public boolean isWhitespace(int nodeHandle)
  {
  	int type = getNodeType(nodeHandle);
    Node node = getNode(nodeHandle);
  	if(TEXT_NODE == type || CDATA_SECTION_NODE == type)
    {
      // If this is a DTM text node, it may be made of multiple DOM text
      // nodes -- including navigating into Entity References. DOM2DTM
      // records the first node in the sequence and requires that we
      // pick up the others when we retrieve the DTM node's value.
      //
      // %REVIEW% DOM Level 3 is expected to add a "whole text"
      // retrieval method which performs this function for us.
      FastStringBuffer buf = StringBufferPool.get();
      while(node!=null)
      {
        buf.append(node.getNodeValue());
        node=logicalNextDOMTextNode(node);