charinfo.java

来自「JAVA 所有包」· Java 代码 · 共 745 行 · 第 1/2 页

     * but additional mappings defined through calls to defineChar2String()
     * are possible. Such entity reference mappings could be over-ridden.
     *
     * This is reusing a stored key object, in an effort to avoid
     * heap activity. Unfortunately, that introduces a threading risk.
     * Simplest fix for now is to make it a synchronized method, or to give
     * up the reuse; I see very little performance difference between them.
     * Long-term solution would be to replace the hashtable with a sparse array
     * keyed directly from the character's integer value; see DTM's
     * string pool for a related solution.
     *
     * @param value The character that should be resolved to
     * a String, e.g. resolve '>' to  "<".
     *
     * @return The String that the character is mapped to, or null if not found.
     * @xsl.usage internal
     */
    synchronized String getOutputStringForChar(char value)
    {
        // CharKey m_charKey = new CharKey(); //Alternative to synchronized
        m_charKey.setChar(value);
        return (String) m_charToString.get(m_charKey);
    }
    
    /**
     * Tell if the character argument that is from
     * an attribute value should have special treatment.
     * 
     * @param value the value of a character that is in an attribute value
     * @return true if the character should have any special treatment, 
     * such as when writing out attribute values, 
     * or entity references.
     * @xsl.usage internal
     */
    final boolean isSpecialAttrChar(int value)
    {
        // for performance try the values in the boolean array first,
        // this is faster access than the BitSet for common ASCII values

        if (value < ASCII_MAX)
            return isSpecialAttrASCII[value];

        // rather than java.util.BitSet, our private
        // implementation is faster (and less general).
        return get(value);
    }    

    /**
     * Tell if the character argument that is from a 
     * text node should have special treatment.
     * 
     * @param value the value of a character that is in a text node
     * @return true if the character should have any special treatment, 
     * such as when writing out attribute values, 
     * or entity references.
     * @xsl.usage internal
     */
    final boolean isSpecialTextChar(int value)
    {
        // for performance try the values in the boolean array first,
        // this is faster access than the BitSet for common ASCII values

        if (value < ASCII_MAX)
            return isSpecialTextASCII[value];

        // rather than java.util.BitSet, our private
        // implementation is faster (and less general).
        return get(value);
    }
    
    /**
     * This method is used to determine if an ASCII character in
     * a text node (not an attribute value) is "clean".
     * @param value the character to check (0 to 127).
     * @return true if the character can go to the writer as-is
     * @xsl.usage internal
     */
    final boolean isTextASCIIClean(int value)
    {
        return isCleanTextASCII[value];
    }
    
//  In the future one might want to use the array directly and avoid
//  the method call, but I think the JIT alreay inlines this well enough
//  so don't do it (for now) - bjm    
//    public final boolean[] getASCIIClean()
//    {
//        return isCleanTextASCII;
//    }

     
    private static CharInfo getCharInfoBasedOnPrivilege(
        final String entitiesFileName, final String method, 
        final boolean internal){
            return (CharInfo) AccessController.doPrivileged(
                new PrivilegedAction() {
                        public Object run() {
                            return new CharInfo(entitiesFileName, 
                              method, internal);}
            });            
    }
     
    /**
     * Factory that reads in a resource file that describes the mapping of
     * characters to entity references.
     *
     * Resource files must be encoded in UTF-8 and have a format like:
     * <pre>
     * # First char # is a comment
     * Entity numericValue
     * quot 34
     * amp 38
     * </pre>
     * (Note: Why don't we just switch to .properties files? Oct-01 -sc)
     *
     * @param entitiesResource Name of entities resource file that should
     * be loaded, which describes that mapping of characters to entity references.
     * @param method the output method type, which should be one of "xml", "html", "text"...
     * 
     * @xsl.usage internal
     */
    static CharInfo getCharInfo(String entitiesFileName, String method)
    {
        CharInfo charInfo = (CharInfo) m_getCharInfoCache.get(entitiesFileName);
        if (charInfo != null) {
            return charInfo;
        }

        // try to load it internally - cache
        try {
            charInfo = getCharInfoBasedOnPrivilege(entitiesFileName, 
                                        method, true);
            m_getCharInfoCache.put(entitiesFileName, charInfo);
            return charInfo;
        } catch (Exception e) {}

        // try to load it externally - do not cache
        try {
            return getCharInfoBasedOnPrivilege(entitiesFileName, 
                                method, false);
        } catch (Exception e) {}

        String absoluteEntitiesFileName;

        if (entitiesFileName.indexOf(':') < 0) {
            absoluteEntitiesFileName =
                SystemIDResolver.getAbsoluteURIFromRelative(entitiesFileName);
        } else {
            try {
                absoluteEntitiesFileName =
                    SystemIDResolver.getAbsoluteURI(entitiesFileName, null);
            } catch (TransformerException te) {
                throw new WrappedRuntimeException(te);
            }
        }

        return getCharInfoBasedOnPrivilege(entitiesFileName, 
                                method, false);
    }

    /** Table of user-specified char infos. */
    private static Hashtable m_getCharInfoCache = new Hashtable();

    /**
     * Returns the array element holding the bit value for the
     * given integer
     * @param i the integer that might be in the set of integers
     * 
     */
    private static int arrayIndex(int i) {
        return (i >> SHIFT_PER_WORD);
    }

    /**
     * For a given integer in the set it returns the single bit
     * value used within a given word that represents whether
     * the integer is in the set or not.
     */
    private static int bit(int i) {
        int ret = (1 << (i & LOW_ORDER_BITMASK));
        return ret;
    }

    /**
     * Creates a new empty set of integers (characters)
     * @param max the maximum integer to be in the set.
     */
    private int[] createEmptySetOfIntegers(int max) {
        firstWordNotUsed = 0; // an optimization 

        int[] arr = new int[arrayIndex(max - 1) + 1];
            return arr;
 
    }

    /**
     * Adds the integer (character) to the set of integers.
     * @param i the integer to add to the set, valid values are 
     * 0, 1, 2 ... up to the maximum that was specified at
     * the creation of the set.
     */
    private final void set(int i) {   
        setASCIIdirty(i);
             
        int j = (i >> SHIFT_PER_WORD); // this word is used
        int k = j + 1;       
        
        if(firstWordNotUsed < k) // for optimization purposes.
            firstWordNotUsed = k;
            
        array_of_bits[j] |= (1 << (i & LOW_ORDER_BITMASK));
    }


    /**
     * Return true if the integer (character)is in the set of integers.
     * 
     * This implementation uses an array of integers with 32 bits per
     * integer.  If a bit is set to 1 the corresponding integer is 
     * in the set of integers.
     * 
     * @param i an integer that is tested to see if it is the
     * set of integers, or not.
     */
    private final boolean get(int i) {

        boolean in_the_set = false;
        int j = (i >> SHIFT_PER_WORD); // wordIndex(i)
        // an optimization here, ... a quick test to see
        // if this integer is beyond any of the words in use
        if(j < firstWordNotUsed)
            in_the_set = (array_of_bits[j] & 
                          (1 << (i & LOW_ORDER_BITMASK))
            ) != 0;  // 0L for 64 bit words
        return in_the_set;
    }
    
    // record if there are any entities other than
    // quot, amp, lt, gt  (probably user defined)
    /**
     * @return true if the entity 
     * @param code The value of the character that has an entity defined
     * for it.
     */
    private boolean extraEntity(int entityValue)
    {
        boolean extra = false;
        if (entityValue < 128)
        {
            switch (entityValue)
            {
                case 34 : // quot
                case 38 : // amp
                case 60 : // lt
                case 62 : // gt
                    break;
                default : // other entity in range 0 to 127  
                    extra = true;
            }
        }
        return extra;
    }    
    
    /**
     * If the character is a printable ASCII character then
     * mark it as not clean and needing replacement with
     * a String on output.
     * @param ch
     */
    private void setASCIIdirty(int j) 
    {
        if (0 <= j && j < ASCII_MAX) 
        {
            isCleanTextASCII[j] = false;
            isSpecialTextASCII[j] = true;
        } 
    }

    /**
     * If the character is a printable ASCII character then
     * mark it as and not needing replacement with
     * a String on output.
     * @param ch
     */    
    private void setASCIIclean(int j)
    {
        if (0 <= j && j < ASCII_MAX) 
        {        
            isCleanTextASCII[j] = true;
            isSpecialTextASCII[j] = false;
        }
    }
    
    private void defineChar2StringMapping(String outputString, char inputChar) 
    {
        CharKey character = new CharKey(inputChar);
        m_charToString.put(character, outputString);
        set(inputChar);        
    }

    /**
     * Simple class for fast lookup of char values, when used with
     * hashtables.  You can set the char, then use it as a key.
     * 
     * This class is a copy of the one in com.sun.org.apache.xml.internal.utils. 
     * It exists to cut the serializers dependancy on that package.
     *  
     * @xsl.usage internal
     */
    private static class CharKey extends Object
    {

      /** String value          */
      private char m_char;

      /**
       * Constructor CharKey
       *
       * @param key char value of this object.
       */
      public CharKey(char key)
      {
        m_char = key;
      }
  
      /**
       * Default constructor for a CharKey.
       *
       * @param key char value of this object.
       */
      public CharKey()
      {
      }
  
      /**
       * Get the hash value of the character.  
       *
       * @return hash value of the character.
       */
      public final void setChar(char c)
      {
        m_char = c;
      }



      /**
       * Get the hash value of the character.  
       *
       * @return hash value of the character.
       */
      public final int hashCode()
      {
        return (int)m_char;
      }

      /**
       * Override of equals() for this object 
       *
       * @param obj to compare to
       *
       * @return True if this object equals this string value 
       */
      public final boolean equals(Object obj)
      {
        return ((CharKey)obj).m_char == m_char;
      }
    }
   

}