domhelper.java

来自「JAVA 所有包」· Java 代码 · 共 1,333 行 · 第 1/3 页

  public boolean isIgnorableWhitespace(Text node)
  {

    boolean isIgnorable = false;  // return value

    // TODO: I can probably do something to figure out if this 
    // space is ignorable from just the information in
    // the DOM tree.
        // -- You need to be able to distinguish whitespace
        // that is #PCDATA from whitespace that isn't.  That requires
        // DTD support, which won't be standardized until DOM Level 3.
    return isIgnorable;
  }

  /**
   * Get the first unparented node in the ancestor chain.
   * @deprecated
   *
   * @param node Starting node, to specify which chain to chase
   *
   * @return the topmost ancestor.
   */
  public Node getRoot(Node node)
  {

    Node root = null;

    while (node != null)
    {
      root = node;
      node = getParentOfNode(node);
    }

    return root;
  }

  /**
   * Get the root node of the document tree, regardless of
   * whether or not the node passed in is a document node.
   * <p>
   * TODO: This doesn't handle DocumentFragments or "orphaned" subtrees
   * -- it's currently returning ownerDocument even when the tree is
   * not actually part of the main Document tree. We should either
   * rewrite the description to say that it finds the Document node,
   * or change the code to walk up the ancestor chain.

   *
   * @param n Node to be examined
   *
   * @return the Document node. Note that this is not the correct answer
   * if n was (or was a child of) a DocumentFragment or an orphaned node,
   * as can arise if the DOM has been edited rather than being generated
   * by a parser.
   */
  public Node getRootNode(Node n)
  {
    int nt = n.getNodeType();
    return ( (Node.DOCUMENT_NODE == nt) || (Node.DOCUMENT_FRAGMENT_NODE == nt) ) 
           ? n : n.getOwnerDocument();
  }

  /**
   * Test whether the given node is a namespace decl node. In DOM Level 2
   * this can be done in a namespace-aware manner, but in Level 1 DOMs
   * it has to be done by testing the node name.
   *
   * @param n Node to be examined.
   *
   * @return boolean -- true iff the node is an Attr whose name is 
   * "xmlns" or has the "xmlns:" prefix.
   */
  public boolean isNamespaceNode(Node n)
  {

    if (Node.ATTRIBUTE_NODE == n.getNodeType())
    {
      String attrName = n.getNodeName();

      return (attrName.startsWith("xmlns:") || attrName.equals("xmlns"));
    }

    return false;
  }

  /**
   * Obtain the XPath-model parent of a DOM node -- ownerElement for Attrs,
   * parent for other nodes. 
   * <p>
   * Background: The DOM believes that you must be your Parent's
   * Child, and thus Attrs don't have parents. XPath said that Attrs
   * do have their owning Element as their parent. This function
   * bridges the difference, either by using the DOM Level 2 ownerElement
   * function or by using a "silly and expensive function" in Level 1
   * DOMs.
   * <p>
   * (There's some discussion of future DOMs generalizing ownerElement 
   * into ownerNode and making it work on all types of nodes. This
   * still wouldn't help the users of Level 1 or Level 2 DOMs)
   * <p>
   *
   * @param node Node whose XPath parent we want to obtain
   *
   * @return the parent of the node, or the ownerElement if it's an
   * Attr node, or null if the node is an orphan.
   *
   * @throws RuntimeException if the Document has no root element.
   * This can't arise if the Document was created
   * via the DOM Level 2 factory methods, but is possible if other
   * mechanisms were used to obtain it
   */
  public static Node getParentOfNode(Node node) throws RuntimeException
  {
    Node parent;
    short nodeType = node.getNodeType();

    if (Node.ATTRIBUTE_NODE == nodeType)
    {
      Document doc = node.getOwnerDocument();
          /*
      TBD:
      if(null == doc)
      {
        throw new RuntimeException(XSLMessages.createXPATHMessage(XPATHErrorResources.ER_CHILD_HAS_NO_OWNER_DOCUMENT, null));//"Attribute child does not have an owner document!");
      }
      */

          // Given how expensive the tree walk may be, we should first ask 
          // whether this DOM can answer the question for us. The additional
          // test does slow down Level 1 DOMs slightly. DOMHelper2, which
          // is currently specialized for Xerces, assumes it can use the
          // Level 2 solution. We might want to have an intermediate stage,
          // which would assume DOM Level 2 but not assume Xerces.
          //
          // (Shouldn't have to check whether impl is null in a compliant DOM,
          // but let's be paranoid for a moment...)
          DOMImplementation impl=doc.getImplementation();
          if(impl!=null && impl.hasFeature("Core","2.0"))
          {
                  parent=((Attr)node).getOwnerElement();
                  return parent;
          }

          // DOM Level 1 solution, as fallback. Hugely expensive. 

      Element rootElem = doc.getDocumentElement();

      if (null == rootElem)
      {
        throw new RuntimeException(
          XMLMessages.createXMLMessage(
            XMLErrorResources.ER_CHILD_HAS_NO_OWNER_DOCUMENT_ELEMENT,
            null));  //"Attribute child does not have an owner document element!");
      }

      parent = locateAttrParent(rootElem, node);

        }
    else
    {
      parent = node.getParentNode();

      // if((Node.DOCUMENT_NODE != nodeType) && (null == parent))
      // {
      //   throw new RuntimeException("Child does not have parent!");
      // }
    }

    return parent;
  }

  /**
   * Given an ID, return the element. This can work only if the document
   * is interpreted in the context of a DTD or Schema, since otherwise
   * we don't know which attributes are or aren't IDs.
   * <p>
   * Note that DOM Level 1 had no ability to retrieve this information.
   * DOM Level 2 introduced it but does not promise that it will be
   * supported in all DOMs; those which can't support it will always
   * return null.
   * <p>
   * TODO: getElementByID is currently unimplemented. Support DOM Level 2?
   *
   * @param id The unique identifier to be searched for.
   * @param doc The document to search within.
   * @return CURRENTLY HARDCODED TO NULL, but it should be:
   * The node which has this unique identifier, or null if there
   * is no such node or this DOM can't reliably recognize it.
   */
  public Element getElementByID(String id, Document doc)
  {
    return null;
  }

  /**
   * The getUnparsedEntityURI function returns the URI of the unparsed
   * entity with the specified name in the same document as the context
   * node (see [3.3 Unparsed Entities]). It returns the empty string if
   * there is no such entity.
   * <p>
   * XML processors may choose to use the System Identifier (if one
   * is provided) to resolve the entity, rather than the URI in the
   * Public Identifier. The details are dependent on the processor, and
   * we would have to support some form of plug-in resolver to handle
   * this properly. Currently, we simply return the System Identifier if
   * present, and hope that it a usable URI or that our caller can
   * map it to one.
   * TODO: Resolve Public Identifiers... or consider changing function name.
   * <p>
   * If we find a relative URI 
   * reference, XML expects it to be resolved in terms of the base URI 
   * of the document. The DOM doesn't do that for us, and it isn't 
   * entirely clear whether that should be done here; currently that's
   * pushed up to a higher levelof our application. (Note that DOM Level 
   * 1 didn't store the document's base URI.)
   * TODO: Consider resolving Relative URIs.
   * <p>
   * (The DOM's statement that "An XML processor may choose to
   * completely expand entities before the structure model is passed
   * to the DOM" refers only to parsed entities, not unparsed, and hence
   * doesn't affect this function.)
   *
   * @param name A string containing the Entity Name of the unparsed
   * entity.
   * @param doc Document node for the document to be searched.
   *
   * @return String containing the URI of the Unparsed Entity, or an
   * empty string if no such entity exists.
   */
  public String getUnparsedEntityURI(String name, Document doc)
  {

    String url = "";
    DocumentType doctype = doc.getDoctype();

    if (null != doctype)
    {
      NamedNodeMap entities = doctype.getEntities();
      if(null == entities)
        return url;
      Entity entity = (Entity) entities.getNamedItem(name);
      if(null == entity)
        return url;
      
      String notationName = entity.getNotationName();

      if (null != notationName)  // then it's unparsed
      {
        // The draft says: "The XSLT processor may use the public 
        // identifier to generate a URI for the entity instead of the URI 
        // specified in the system identifier. If the XSLT processor does 
        // not use the public identifier to generate the URI, it must use 
        // the system identifier; if the system identifier is a relative 
        // URI, it must be resolved into an absolute URI using the URI of 
        // the resource containing the entity declaration as the base 
        // URI [RFC2396]."
        // So I'm falling a bit short here.
        url = entity.getSystemId();

        if (null == url)
        {
          url = entity.getPublicId();
        }
        else
        {
          // This should be resolved to an absolute URL, but that's hard 
          // to do from here.
        }        
      }
    }

    return url;
  }

  /**
   * Support for getParentOfNode; walks a DOM tree until it finds
   * the Element which owns the Attr. This is hugely expensive, and
   * if at all possible you should use the DOM Level 2 Attr.ownerElement()
   * method instead.
   *  <p>
   * The DOM Level 1 developers expected that folks would keep track
   * of the last Element they'd seen and could recover the info from
   * that source. Obviously that doesn't work very well if the only
   * information you've been presented with is the Attr. The DOM Level 2
   * getOwnerElement() method fixes that, but only for Level 2 and
   * later DOMs.
   *
   * @param elem Element whose subtree is to be searched for this Attr
   * @param attr Attr whose owner is to be located.
   *
   * @return the first Element whose attribute list includes the provided
   * attr. In modern DOMs, this will also be the only such Element. (Early
   * DOMs had some hope that Attrs might be sharable, but this idea has
   * been abandoned.)
   */
  private static Node locateAttrParent(Element elem, Node attr)
  {

    Node parent = null;

        // This should only be called for Level 1 DOMs, so we don't have to
        // worry about namespace issues. In later levels, it's possible
        // for a DOM to have two Attrs with the same NodeName but
        // different namespaces, and we'd need to get getAttributeNodeNS...
        // but later levels also have Attr.getOwnerElement.
        Attr check=elem.getAttributeNode(attr.getNodeName());
        if(check==attr)
                parent = elem;

    if (null == parent)
    {
      for (Node node = elem.getFirstChild(); null != node;
              node = node.getNextSibling())
      {
        if (Node.ELEMENT_NODE == node.getNodeType())
        {
          parent = locateAttrParent((Element) node, attr);

          if (null != parent)
            break;
        }
      }
    }

    return parent;
  }

  /**
   * The factory object used for creating nodes
   * in the result tree.
   */
  protected Document m_DOMFactory = null;

  /**
   * Store the factory object required to create DOM nodes
   * in the result tree. In fact, that's just the result tree's
   * Document node...
   *
   * @param domFactory The DOM Document Node within whose context
   * the result tree will be built.
   */
  public void setDOMFactory(Document domFactory)
  {
    this.m_DOMFactory = domFactory;
  }

  /**
   * Retrieve the factory object required to create DOM nodes
   * in the result tree.
   *
   * @return The result tree's DOM Document Node.
   */
  public Document getDOMFactory()
  {

    if (null == this.m_DOMFactory)
    {
      this.m_DOMFactory = createDocument();
    }

    return this.m_DOMFactory;
  }

  /**
   * Get the textual contents of the node. See
   * getNodeData(Node,FastStringBuffer) for discussion of how
   * whitespace nodes are handled.
   *
   * @param node DOM Node to be examined
   * @return String containing a concatenation of all the 
   * textual content within that node. 
   * @see #getNodeData(Node,FastStringBuffer)
   * 
   */
  public static String getNodeData(Node node)
  {

    FastStringBuffer buf = StringBufferPool.get();
    String s;

    try
    {
      getNodeData(node, buf);

      s = (buf.length() > 0) ? buf.toString() : "";
    }
    finally
    {
      StringBufferPool.free(buf);
    }

    return s;
  }

  /**
   * 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.
   *
   * @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.
   */
  public 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 :
      buf.append(node.getNodeValue());
      break;
    case Node.ATTRIBUTE_NODE :
      buf.append(node.getNodeValue());
      break;
    case Node.PROCESSING_INSTRUCTION_NODE :
      // warning(XPATHErrorResources.WG_PARSING_AND_PREPARING);        
      break;
    default :
      // ignore
      break;
    }
  }
}