dom2dtm.java

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

      }
     boolean b = buf.isWhitespace(0, buf.length());
      StringBufferPool.free(buf);
     return b;
    }
    return false;
  }
  
  /**
   * Retrieve the text content of a DOM subtree, appending it into a
   * user-supplied FastStringBuffer object. Note that attributes are
   * not considered part of the content of an element.
   * <p>
   * There are open questions regarding whitespace stripping. 
   * Currently we make no special effort in that regard, since the standard
   * DOM doesn't yet provide DTD-based information to distinguish
   * whitespace-in-element-context from genuine #PCDATA. Note that we
   * should probably also consider xml:space if/when we address this.
   * DOM Level 3 may solve the problem for us.
   * <p>
   * %REVIEW% Actually, since this method operates on the DOM side of the
   * fence rather than the DTM side, it SHOULDN'T do
   * any special handling. The DOM does what the DOM does; if you want
   * DTM-level abstractions, use DTM-level methods.
   *
   * @param node Node whose subtree is to be walked, gathering the
   * contents of all Text or CDATASection nodes.
   * @param buf FastStringBuffer into which the contents of the text
   * nodes are to be concatenated.
   */
  protected static void getNodeData(Node node, FastStringBuffer buf)
  {

    switch (node.getNodeType())
    {
    case Node.DOCUMENT_FRAGMENT_NODE :
    case Node.DOCUMENT_NODE :
    case Node.ELEMENT_NODE :
    {
      for (Node child = node.getFirstChild(); null != child;
              child = child.getNextSibling())
      {
        getNodeData(child, buf);
      }
    }
    break;
    case Node.TEXT_NODE :
    case Node.CDATA_SECTION_NODE :
    case Node.ATTRIBUTE_NODE :	// Never a child but might be our starting node
      buf.append(node.getNodeValue());
      break;
    case Node.PROCESSING_INSTRUCTION_NODE :
      // warning(XPATHErrorResources.WG_PARSING_AND_PREPARING);        
      break;
    default :
      // ignore
      break;
    }
  }

  /**
   * Given a node handle, return its DOM-style node name. This will
   * include names such as #text or #document.
   *
   * @param nodeHandle the id of the node.
   * @return String Name of this node, which may be an empty string.
   * %REVIEW% Document when empty string is possible...
   * %REVIEW-COMMENT% It should never be empty, should it?
   */
  public String getNodeName(int nodeHandle)
  {

    Node node = getNode(nodeHandle);

    // Assume non-null.
    return node.getNodeName();
  }

  /**
   * Given a node handle, return the XPath node name.  This should be
   * the name as described by the XPath data model, NOT the DOM-style
   * name.
   *
   * @param nodeHandle the id of the node.
   * @return String Name of this node, which may be an empty string.
   */
  public String getNodeNameX(int nodeHandle)
  {

    String name;
    short type = getNodeType(nodeHandle);

    switch (type)
    {
    case DTM.NAMESPACE_NODE :
    {
      Node node = getNode(nodeHandle);

      // assume not null.
      name = node.getNodeName();
      if(name.startsWith("xmlns:"))
      {
        name = QName.getLocalPart(name);
      }
      else if(name.equals("xmlns"))
      {
        name = "";
      }
    }
    break;
    case DTM.ATTRIBUTE_NODE :
    case DTM.ELEMENT_NODE :
    case DTM.ENTITY_REFERENCE_NODE :
    case DTM.PROCESSING_INSTRUCTION_NODE :
    {
      Node node = getNode(nodeHandle);

      // assume not null.
      name = node.getNodeName();
    }
    break;
    default :
      name = "";
    }

    return name;
  }

  /**
   * Given a node handle, return its XPath-style localname.
   * (As defined in Namespaces, this is the portion of the name after any
   * colon character).
   *
   * @param nodeHandle the id of the node.
   * @return String Local name of this node.
   */
  public String getLocalName(int nodeHandle)
  {
    if(JJK_NEWCODE)
    {
      int id=makeNodeIdentity(nodeHandle);
      if(NULL==id) return null;
      Node newnode=(Node)m_nodes.elementAt(id);
      String newname=newnode.getLocalName();
      if (null == newname)
      {
	// XSLT treats PIs, and possibly other things, as having QNames.
	String qname = newnode.getNodeName();
	if('#'==qname.charAt(0))
	{
	  //  Match old default for this function
	  // This conversion may or may not be necessary
	  newname="";
	}
	else
	{
	  int index = qname.indexOf(':');
	  newname = (index < 0) ? qname : qname.substring(index + 1);
	}
      }
      return newname;
    }
    else
    {
      String name;
      short type = getNodeType(nodeHandle);
      switch (type)
      {
      case DTM.ATTRIBUTE_NODE :
      case DTM.ELEMENT_NODE :
      case DTM.ENTITY_REFERENCE_NODE :
      case DTM.NAMESPACE_NODE :
      case DTM.PROCESSING_INSTRUCTION_NODE :
	{
	  Node node = getNode(nodeHandle);
	  
	  // assume not null.
	  name = node.getLocalName();
	  
	  if (null == name)
	  {
	    String qname = node.getNodeName();
	    int index = qname.indexOf(':');
	    
	    name = (index < 0) ? qname : qname.substring(index + 1);
	  }
	}
	break;
      default :
	name = "";
      }
      return name;
    }
  }

  /**
   * Given a namespace handle, return the prefix that the namespace decl is
   * mapping.
   * Given a node handle, return the prefix used to map to the namespace.
   *
   * <p> %REVIEW% Are you sure you want "" for no prefix?  </p>
   * <p> %REVIEW-COMMENT% I think so... not totally sure. -sb  </p>
   *
   * @param nodeHandle the id of the node.
   * @return String prefix of this node's name, or "" if no explicit
   * namespace prefix was given.
   */
  public String getPrefix(int nodeHandle)
  {

    String prefix;
    short type = getNodeType(nodeHandle);

    switch (type)
    {
    case DTM.NAMESPACE_NODE :
    {
      Node node = getNode(nodeHandle);

      // assume not null.
      String qname = node.getNodeName();
      int index = qname.indexOf(':');

      prefix = (index < 0) ? "" : qname.substring(index + 1);
    }
    break;
    case DTM.ATTRIBUTE_NODE :
    case DTM.ELEMENT_NODE :
    {
      Node node = getNode(nodeHandle);

      // assume not null.
      String qname = node.getNodeName();
      int index = qname.indexOf(':');

      prefix = (index < 0) ? "" : qname.substring(0, index);
    }
    break;
    default :
      prefix = "";
    }

    return prefix;
  }

  /**
   * Given a node handle, return its DOM-style namespace URI
   * (As defined in Namespaces, this is the declared URI which this node's
   * prefix -- or default in lieu thereof -- was mapped to.)
   *
   * <p>%REVIEW% Null or ""? -sb</p>
   *
   * @param nodeHandle the id of the node.
   * @return String URI value of this node's namespace, or null if no
   * namespace was resolved.
   */
  public String getNamespaceURI(int nodeHandle)
  {
    if(JJK_NEWCODE)
    {
      int id=makeNodeIdentity(nodeHandle);
      if(id==NULL) return null;
      Node node=(Node)m_nodes.elementAt(id);
      return node.getNamespaceURI();
    }
    else
    {
      String nsuri;
      short type = getNodeType(nodeHandle);
      
      switch (type)
      {
      case DTM.ATTRIBUTE_NODE :
      case DTM.ELEMENT_NODE :
      case DTM.ENTITY_REFERENCE_NODE :
      case DTM.NAMESPACE_NODE :
      case DTM.PROCESSING_INSTRUCTION_NODE :
	{
	  Node node = getNode(nodeHandle);
	  
	  // assume not null.
	  nsuri = node.getNamespaceURI();
	  
	  // %TBD% Handle DOM1?
	}
	break;
      default :
	nsuri = null;
      }

      return nsuri;
    }
    
  }
  
  /** Utility function: Given a DOM Text node, determine whether it is
   * logically followed by another Text or CDATASection node. This may
   * involve traversing into Entity References.
   * 
   * %REVIEW% DOM Level 3 is expected to add functionality which may 
   * allow us to retire this.
   */
  private Node logicalNextDOMTextNode(Node n)
  {
        Node p=n.getNextSibling();
        if(p==null)
        {
                // Walk out of any EntityReferenceNodes that ended with text
                for(n=n.getParentNode();
                        n!=null && ENTITY_REFERENCE_NODE == n.getNodeType();
                        n=n.getParentNode())
                {
                        p=n.getNextSibling();
                        if(p!=null)
                                break;
                }
        }
        n=p;
        while(n!=null && ENTITY_REFERENCE_NODE == n.getNodeType())
        {
                // Walk into any EntityReferenceNodes that start with text
                if(n.hasChildNodes())
                        n=n.getFirstChild();
                else
                        n=n.getNextSibling();
        }
        if(n!=null)
        {
                // Found a logical next sibling. Is it text?
                int ntype=n.getNodeType();
                if(TEXT_NODE != ntype && CDATA_SECTION_NODE != ntype)
                        n=null;
        }
        return n;
  }

  /**
   * Given a node handle, return its node value. This is mostly
   * as defined by the DOM, but may ignore some conveniences.
   * <p>
   *
   * @param nodeHandle The node id.
   * @return String Value of this node, or null if not
   * meaningful for this node type.
   */
  public String getNodeValue(int nodeHandle)
  {
    // The _type(nodeHandle) call was taking the lion's share of our
    // time, and was wrong anyway since it wasn't coverting handle to
    // identity. Inlined it.
    int type = _exptype(makeNodeIdentity(nodeHandle));
    type=(NULL != type) ? getNodeType(nodeHandle) : NULL;
    
    if(TEXT_NODE!=type && CDATA_SECTION_NODE!=type)
      return getNode(nodeHandle).getNodeValue();
    
    // 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.
    Node node = getNode(nodeHandle);
    Node n=logicalNextDOMTextNode(node);
    if(n==null)
      return node.getNodeValue();
    
    FastStringBuffer buf = StringBufferPool.get();
        buf.append(node.getNodeValue());
    while(n!=null)
    {
      buf.append(n.getNodeValue());
      n=logicalNextDOMTextNode(n);
    }
    String s = (buf.length() > 0) ? buf.toString() : "";
    StringBufferPool.free(buf);
    return s;
  }

  /**
   *   A document type declaration information item has the following properties:
   *
   *     1. [system identifier] The system identifier of the external subset, if
   *        it exists. Otherwise this property has no value.
   *
   * @return the system identifier String object, or null if there is none.
   */
  public String getDocumentTypeDeclarationSystemIdentifier()
  {

    Document doc;

    if (m_root.getNodeType() == Node.DOCUMENT_NODE)
      doc = (Document) m_root;
    else
      doc = m_root.getOwnerDocument();

    if (null != doc)
    {
      DocumentType dtd = doc.getDoctype();

      if (null != dtd)
      {
        return dtd.getSystemId();
      }
    }

    return null;
  }

  /**
   * Return the public identifier of the external subset,
   * normalized as described in 4.2.2 External Entities [XML]. If there is
   * no external subset or if it has no public identifier, this property
   * has no value.
   *
   * @return the public identifier String object, or null if there is none.
   */
  public String getDocumentTypeDeclarationPublicIdentifier()
  {

    Document doc;

    if (m_root.getNodeType() == Node.DOCUMENT_NODE)
      doc = (Document) m_root;
    else
      doc = m_root.getOwnerDocument();

    if (null != doc)
    {
      DocumentType dtd = doc.getDoctype();

      if (null != dtd)
      {
        return dtd.getPublicId();
      }
    }

    return null;
  }