htmlpolicy.java

来自「RESIN 3.2 最新源码」· Java 代码 · 共 704 行 · 第 1/2 页

/*
 * Copyright (c) 1998-2008 Caucho Technology -- all rights reserved
 *
 * This file is part of Resin(R) Open Source
 *
 * Each copy or derived work must preserve the copyright notice and this
 * notice unmodified.
 *
 * Resin Open Source is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * Resin Open Source is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE, or any warranty
 * of NON-INFRINGEMENT.  See the GNU General Public License for more
 * details.
 *
 * You should have received a copy of the GNU General Public License
 * along with Resin Open Source; if not, write to the
 *   Free SoftwareFoundation, Inc.
 *   59 Temple Place, Suite 330
 *   Boston, MA 02111-1307  USA
 *
 * @author Scott Ferguson
 */

package com.caucho.xml;

import com.caucho.util.CharBuffer;
import com.caucho.util.CharCursor;
import com.caucho.util.CharScanner;
import com.caucho.util.IntMap;
import com.caucho.util.StringCharCursor;

import org.w3c.dom.Element;

import java.io.IOException;

/**
 * Policy for parsing an HTML file.
 */
class HtmlPolicy extends Policy {
  static final int DOCUMENT = 1;
  static final int COMMENT = DOCUMENT + 1;
  static final int TEXT = COMMENT + 1;
  static final int JSP = TEXT + 1;
  static final int WHITESPACE = JSP + 1;

  static final int HTML = WHITESPACE + 1;
  static final int HEAD = HTML + 1;
  static final int TITLE = HEAD + 1;
  static final int ISINDEX = TITLE + 1;
  static final int BASE = ISINDEX + 1;
  static final int SCRIPT = BASE + 1;
  static final int STYLE = SCRIPT + 1;
  static final int META = STYLE + 1;
  static final int LINK = META + 1;
  static final int OBJECT = LINK + 1;

  static final int BODY = OBJECT + 1;

  static final int BASEFONT = BODY + 1;
  static final int BR = BASEFONT + 1;
  static final int AREA = BR + 1;
  static final int IMG = AREA + 1;
  static final int PARAM = IMG + 1;
  static final int HR = PARAM + 1;
  static final int INPUT = HR + 1;

  static final int P = INPUT + 1;
  static final int DT = P + 1;
  static final int DD = DT + 1;
  static final int LI = DD + 1;
  static final int OPTION = LI + 1;

  static final int TABLE = OPTION + 1;
  static final int CAPTION = TABLE + 1;
  static final int THEAD = CAPTION + 1;
  static final int TFOOT = THEAD + 1;
  static final int COL = TFOOT + 1;
  static final int COLGROUP = COL + 1;
  static final int TBODY = COLGROUP + 1;
  static final int TR = TBODY + 1;
  static final int TD = TR + 1;
  static final int TH = TD + 1;

  static final int FRAME = TH + 1;
  static final int FRAMESET = FRAME + 1;

  static final int BLOCK = FRAMESET + 1;
  static final int INLINE = BLOCK + 1;

  static IntMap names;
  static IntMap cbNames;
  
  static QName htmlName = new QName(null, "html", null);
  static QName headName = new QName(null, "head", null);
  static QName bodyName = new QName(null, "body", null);

  boolean toLower = true;
  boolean isJsp = false;
  boolean autoHtml = false;
  boolean hasBody = false;
  boolean autoHead = false;
  
  CharBuffer cb = new CharBuffer();

  public void init()
  {
    toLower = true;
    isJsp = false;
    autoHtml = false;
    hasBody = false;
    autoHead = false;
  }

  /**
   * When true, HTML parsing normalizes HTML tags to lower case.
   */
  public void setToLower(boolean toLower)
  {
    this.toLower = toLower;
  }

  /**
   * When true, treat text before HTML specially.
   */
  public void setJsp(boolean isJsp)
  {
    this.isJsp = isJsp;
  }

  /**
   * Return the normalized name.
   *
   * @param tag the raw name in the XML file.
   *
   * @return the normalized name.
   */
  QName getName(CharBuffer tag)
  {
    if (! toLower)
      return super.getName(tag);
    
    cb.clear();
    cb.append(tag);
    cb.toLowerCase();

    int name = cbNames.get(cb);

    if (name >= 0)
      return super.getName(cb);
    else
      return super.getName(tag);
  }

  QName getAttributeName(CharBuffer eltName, CharBuffer source)
  {
    if (! toLower)
      return super.getName(source);
    
    cb.clear();
    cb.append(eltName);
    cb.toLowerCase();
    int name = cbNames.get(cb);

    if (name < 0)
      return super.getName(source);
    else {
      source.toLowerCase();
      return super.getName(source);
    }
  }

  /**
   * Returns the appropriate action when opening a HTML tag.
   *
   * @param parser the XML parser
   * @param node the parent node
   * @param next the next child
   * @return the action code
   */
  int openAction(XmlParser parser, QName node, QName next)
    throws XmlParseException
  {
    String nodeName = node == null ? "#document" : node.getName();
    String nextName = next.getName();

    int nextCode = names.get(nextName);

    switch (names.get(nodeName)) {
    case DOCUMENT:
      switch (nextCode) {
      case HTML:
	return PUSH;

      case COMMENT:
        return PUSH;

      case HEAD: case TITLE: case ISINDEX: case BASE: case SCRIPT: 
      case STYLE: case META: case LINK: case OBJECT:
	opt = htmlName;
	return PUSH_OPT;

      case WHITESPACE:
        return IGNORE;

      case JSP:
        return PUSH;

      default:
        if (autoHtml)
          return PUSH;
        
        autoHtml = true;
	opt = htmlName;
	return PUSH_OPT;
      }

    case HTML:
      switch (nextCode) {
      case HTML:
	return ERROR;

      case HEAD:
      case COMMENT:
      case FRAMESET:
	return PUSH;
        
      case BODY:
        hasBody = true;
	return PUSH;

      case TITLE: case ISINDEX: case BASE: case SCRIPT: 
      case STYLE: case META: case LINK: case OBJECT:
	opt = headName;
        autoHead = true;
	return PUSH_OPT;

      case WHITESPACE:
        return PUSH;

      case JSP:
        return PUSH;

      default:
        if (hasBody)
          return PUSH;
        
        hasBody = true;
	opt = bodyName;
	return PUSH_OPT;
      }

    case HEAD:
      switch (nextCode) {
      case META:
	// checkMetaEncoding((Element) next);
	return PUSH_EMPTY;

      case LINK: case ISINDEX: case BASE: 
	return PUSH_EMPTY;
        
      case SCRIPT: case STYLE:
        return PUSH_VERBATIM;
        
      case TITLE:
      case OBJECT:
	return PUSH;

      case WHITESPACE:
        return PUSH;
        
      case JSP:
      case TEXT:
        if (autoHead)
          return POP;
        else
          return PUSH;

      default:
	return POP;
      }

    case LI:
      switch (nextCode) {
      case LI:
	return POP;

      case BASEFONT: case BR: case AREA: case LINK: case IMG: case PARAM: 
      case HR: case INPUT: case COL: case FRAME: case ISINDEX: 
      case BASE: case META:
	return PUSH_EMPTY;

      case SCRIPT: case STYLE:
        return PUSH_VERBATIM;

      default:
	return PUSH;
      }

    case OPTION:
      switch (nextCode) {
      case WHITESPACE:
      case TEXT:
        return PUSH;

      default:
	return POP;
      }

    case DD:
      switch (nextCode) {
      case DD: case DT:
	return POP;

      case BASEFONT: case BR: case AREA: case LINK: case IMG: case PARAM: 
      case HR: case INPUT: case COL: case FRAME: case ISINDEX: 
      case BASE: case META:
	return PUSH_EMPTY;

      case SCRIPT: case STYLE:
        return PUSH_VERBATIM;

      default:
	return PUSH;
      }

    case THEAD: case TFOOT: case COLGROUP:
      switch (nextCode) {
      case THEAD: case TFOOT: case TBODY: case COLGROUP: case COL:
	return POP;

      case BASEFONT: case BR: case AREA: case LINK: case IMG: case PARAM: 
      case HR: case INPUT: case FRAME: case ISINDEX: 
      case BASE: case META:
	return PUSH_EMPTY;

      case SCRIPT: case STYLE:
        return PUSH_VERBATIM;

      default:
	return PUSH;
      }

    case TR:
      switch (nextCode) {
      case THEAD: case TFOOT: case TBODY: case COLGROUP: case COL: case TR:
	return POP;