datainputstream.java

gcc的组建

            break;
	strb.append((char) c);
      }

    return strb.length() > 0 ? strb.toString() : "";
  }

  /**
   * This method reads a Java <code>long</code> value from an input stream
   * It operates by reading eight bytes from the stream and converting them to
   * a single Java <code>long</code>.  The bytes are stored most
   * significant byte first (i.e., "big endian") regardless of the native
   * host byte ordering.
   * <p>
   * As an example, if <code>byte1</code> through <code>byte8</code> represent
   * the first eight bytes read from the stream, they will be
   * transformed to an <code>long</code> in the following manner:
   * <p>
   * <code>(long)(((byte1 &amp; 0xFF) &lt;&lt; 56) + ((byte2 &amp; 0xFF) &lt;&lt; 48) +
   * ((byte3 &amp; 0xFF) &lt;&lt; 40) + ((byte4 &amp; 0xFF) &lt;&lt; 32) +
   * ((byte5 &amp; 0xFF) &lt;&lt; 24) + ((byte6 &amp; 0xFF) &lt;&lt; 16) +
   * ((byte7 &amp; 0xFF) &lt;&lt; 8) + (byte8 &amp; 0xFF)))
   * </code>
   * <p>
   * The value returned is in the range of -9223372036854775808 to
   * 9223372036854775807.
   * <p>
   * This method can read an <code>long</code> written by an object
   * implementing the <code>writeLong()</code> method in the
   * <code>DataOutput</code> interface.
   *
   * @return The <code>long</code> value read
   *
   * @exception EOFException If end of file is reached before reading the long
   * @exception IOException If any other error occurs
   *
   * @see DataOutput#writeLong
   */
  public final long readLong () throws IOException
  {
    readFully (buf, 0, 8);
    return convertToLong (buf);
  }

  /**
   * This method reads a signed 16-bit value into a Java in from the
   * stream.  It operates by reading two bytes from the stream and
   * converting them to a single 16-bit Java <code>short</code>.  The
   * two bytes are stored most significant byte first (i.e., "big
   * endian") regardless of the native host byte ordering.
   * <p>
   * As an example, if <code>byte1</code> and <code>byte2</code>
   * represent the first and second byte read from the stream
   * respectively, they will be transformed to a <code>short</code>. in
   * the following manner:
   * <p>
   * <code>(short)(((byte1 &amp; 0xFF) &lt;&lt; 8) | (byte2 &amp; 0xFF))</code>
   * <p>
   * The value returned is in the range of -32768 to 32767.
   * <p>
   * This method can read a <code>short</code> written by an object
   * implementing the <code>writeShort()</code> method in the
   * <code>DataOutput</code> interface.
   *
   * @return The <code>short</code> value read
   *
   * @exception EOFException If end of file is reached before reading the value
   * @exception IOException If any other error occurs
   *
   * @see DataOutput#writeShort
   */
  public final short readShort () throws IOException
  {
    readFully (buf, 0, 2);
    return convertToShort (buf);
  }
  
  /**
   * This method reads 8 unsigned bits into a Java <code>int</code>
   * value from the stream. The value returned is in the range of 0 to
   * 255.
   * <p>
   * This method can read an unsigned byte written by an object
   * implementing the <code>writeUnsignedByte()</code> method in the
   * <code>DataOutput</code> interface.
   *
   * @return The unsigned bytes value read as a Java <code>int</code>.
   *
   * @exception EOFException If end of file is reached before reading the value
   * @exception IOException If any other error occurs
   *
   * @see DataOutput#writeByte
   */
  public final int readUnsignedByte () throws IOException
  {
    return convertToUnsignedByte (in.read ());
  }

  /**
   * This method reads 16 unsigned bits into a Java int value from the stream.
   * It operates by reading two bytes from the stream and converting them to 
   * a single Java <code>int</code>  The two bytes are stored most
   * significant byte first (i.e., "big endian") regardless of the native
   * host byte ordering. 
   * <p>
   * As an example, if <code>byte1</code> and <code>byte2</code>
   * represent the first and second byte read from the stream
   * respectively, they will be transformed to an <code>int</code> in
   * the following manner:
   * <p>
   * <code>(int)(((byte1 &amp; 0xFF) &lt;&lt; 8) + (byte2 &amp; 0xFF))</code>
   * <p>
   * The value returned is in the range of 0 to 65535.
   * <p>
   * This method can read an unsigned short written by an object
   * implementing the <code>writeUnsignedShort()</code> method in the
   * <code>DataOutput</code> interface.
   *
   * @return The unsigned short value read as a Java <code>int</code>
   *
   * @exception EOFException If end of file is reached before reading the value
   * @exception IOException If any other error occurs
   *
   * @see DataOutput#writeShort
   */
  public final int readUnsignedShort () throws IOException
  {
    readFully (buf, 0, 2);
    return convertToUnsignedShort (buf);
  }

  /**
   * This method reads a <code>String</code> from an input stream that
   * is encoded in a modified UTF-8 format.  This format has a leading
   * two byte sequence that contains the remaining number of bytes to
   * read.  This two byte sequence is read using the
   * <code>readUnsignedShort()</code> method of this interface.
   * <p>
   * After the number of remaining bytes have been determined, these
   * bytes are read an transformed into <code>char</code> values.
   * These <code>char</code> values are encoded in the stream using
   * either a one, two, or three byte format.  The particular format
   * in use can be determined by examining the first byte read.
   * <p>
   * If the first byte has a high order bit of 0, then that character
   * consists on only one byte.  This character value consists of
   * seven bits that are at positions 0 through 6 of the byte.  As an
   * example, if <code>byte1</code> is the byte read from the stream,
   * it would be converted to a <code>char</code> like so:
   * <p>
   * <code>(char)byte1</code>
   * <p>
   * If the first byte has 110 as its high order bits, then the 
   * character consists of two bytes.  The bits that make up the character
   * value are in positions 0 through 4 of the first byte and bit positions
   * 0 through 5 of the second byte.  (The second byte should have 
   * 10 as its high order bits).  These values are in most significant
   * byte first (i.e., "big endian") order.
   * <p>
   * As an example, if <code>byte1</code> and <code>byte2</code> are
   * the first two bytes read respectively, and the high order bits of
   * them match the patterns which indicate a two byte character
   * encoding, then they would be converted to a Java
   * <code>char</code> like so:
   * <p>
   * <code>(char)(((byte1 & 0x1F) << 6) | (byte2 & 0x3F))</code>
   * <p>
   * If the first byte has a 1110 as its high order bits, then the
   * character consists of three bytes.  The bits that make up the character
   * value are in positions 0 through 3 of the first byte and bit positions
   * 0 through 5 of the other two bytes.  (The second and third bytes should
   * have 10 as their high order bits).  These values are in most
   * significant byte first (i.e., "big endian") order.
   * <p>
   * As an example, if <code>byte1</code> <code>byte2</code> and
   * <code>byte3</code> are the three bytes read, and the high order
   * bits of them match the patterns which indicate a three byte
   * character encoding, then they would be converted to a Java
   * <code>char</code> like so:
   * <p>
   * <code>(char)(((byte1 & 0x0F) << 12) | ((byte2 & 0x3F) << 6) | 
   * (byte3 & 0x3F))</code>
   * <p>
   * Note that all characters are encoded in the method that requires
   * the fewest number of bytes with the exception of the character
   * with the value of <code>&#92;u0000</code> which is encoded as two
   * bytes.  This is a modification of the UTF standard used to
   * prevent C language style <code>NUL</code> values from appearing
   * in the byte stream.
   * <p>
   * This method can read data that was written by an object implementing the
   * <code>writeUTF()</code> method in <code>DataOutput</code>
   * 
   * @return The <code>String</code> read
   *
   * @exception EOFException If end of file is reached before reading
   * the String
   * @exception UTFDataFormatException If the data is not in UTF-8 format
   * @exception IOException If any other error occurs
   *
   * @see DataOutput#writeUTF
   */
  public final String readUTF () throws IOException
  {
    return readUTF (this);
  }

  /**
   * This method reads a String encoded in UTF-8 format from the 
   * specified <code>DataInput</code> source.
   *
   * @param in The <code>DataInput</code> source to read from
   *
   * @return The String read from the source
   *
   * @exception IOException If an error occurs
   *
   * @see DataInput#readUTF
   */
  public static final String readUTF(DataInput in) throws IOException
  {
    final int UTFlen = in.readUnsignedShort ();
    byte[] buf = new byte [UTFlen];

    // This blocks until the entire string is available rather than
    // doing partial processing on the bytes that are available and then
    // blocking.  An advantage of the latter is that Exceptions
    // could be thrown earlier.  The former is a bit cleaner.
    in.readFully (buf, 0, UTFlen);

    return convertFromUTF (buf);
  }

  /**
   * This method attempts to skip and discard the specified number of bytes 
   * in the input stream.  It may actually skip fewer bytes than requested. 
   * This method will not skip any bytes if passed a negative number of bytes 
   * to skip. 
   *
   * @param n The requested number of bytes to skip.
   *
   * @return The requested number of bytes to skip.
   *
   * @exception IOException If an error occurs.
   * @specnote The JDK docs claim that this returns the number of bytes 
   *  actually skipped. The JCL claims that this method can throw an 
   *  EOFException. Neither of these appear to be true in the JDK 1.3's
   *  implementation. This tries to implement the actual JDK behaviour.
   */
  public final int skipBytes (int n) throws IOException
  {
    if (n <= 0)
      return 0;    
    try
      {
        return (int) in.skip (n);
      }
    catch (EOFException x)
      {
        // do nothing.
      }         
    return n;
  }
  
  static boolean convertToBoolean (int b) throws EOFException
  {
    if (b < 0)
      throw new EOFException ();
    
    return (b != 0);
  }

  static byte convertToByte (int i) throws EOFException
  {
    if (i < 0)
      throw new EOFException ();
    
    return (byte) i;
  }

  static int convertToUnsignedByte (int i) throws EOFException
  {
    if (i < 0)
      throw new EOFException ();
    
    return (i & 0xFF);
  }

  static char convertToChar (byte[] buf)
  {
    return (char) ((buf [0] << 8)
		    | (buf [1] & 0xff));  
  }  

  static short convertToShort (byte[] buf)
  {
    return (short) ((buf [0] << 8)
		    | (buf [1] & 0xff));  
  }  

  static int convertToUnsignedShort (byte[] buf)
  {
    return (((buf [0] & 0xff) << 8)
	    | (buf [1] & 0xff));  
  }

  static int convertToInt (byte[] buf)
  {
    return (((buf [0] & 0xff) << 24)
	    | ((buf [1] & 0xff) << 16)
	    | ((buf [2] & 0xff) << 8)
	    | (buf [3] & 0xff));  
  }

  static long convertToLong (byte[] buf)
  {
    return (((long)(buf [0] & 0xff) << 56) |
	    ((long)(buf [1] & 0xff) << 48) |
	    ((long)(buf [2] & 0xff) << 40) |
	    ((long)(buf [3] & 0xff) << 32) |
	    ((long)(buf [4] & 0xff) << 24) |
	    ((long)(buf [5] & 0xff) << 16) |
	    ((long)(buf [6] & 0xff) <<  8) |
	    ((long)(buf [7] & 0xff)));  
  }

  // FIXME: This method should be re-thought.  I suspect we have multiple
  // UTF-8 decoders floating around.  We should use the standard charset
  // converters, maybe and adding a direct call into one of the new
  // NIO converters for a super-fast UTF8 decode.
  static String convertFromUTF (byte[] buf) 
    throws EOFException, UTFDataFormatException
  {
    // Give StringBuffer an initial estimated size to avoid 
    // enlarge buffer frequently
    StringBuffer strbuf = new StringBuffer (buf.length / 2 + 2);

    for (int i = 0; i < buf.length; )
      {
	if ((buf [i] & 0x80) == 0)		// bit pattern 0xxxxxxx
	  strbuf.append ((char) (buf [i++] & 0xFF));
	else if ((buf [i] & 0xE0) == 0xC0)	// bit pattern 110xxxxx
	  {
	    if (i + 1 >= buf.length
		|| (buf [i + 1] & 0xC0) != 0x80)
	      throw new UTFDataFormatException ();

	    strbuf.append((char) (((buf [i++] & 0x1F) << 6)
				  | (buf [i++] & 0x3F)));
	  }
	else if ((buf [i] & 0xF0) == 0xE0)	// bit pattern 1110xxxx
	  {
	    if (i + 2 >= buf.length
		|| (buf [i + 1] & 0xC0) != 0x80
		|| (buf [i + 2] & 0xC0) != 0x80)
	      throw new UTFDataFormatException ();

	    strbuf.append ((char) (((buf [i++] & 0x0F) << 12)
				   | ((buf [i++] & 0x3F) << 6)
				   | (buf [i++] & 0x3F)));
	  }
	else // must be ((buf [i] & 0xF0) == 0xF0 || (buf [i] & 0xC0) == 0x80)
	  throw new UTFDataFormatException ();	// bit patterns 1111xxxx or
						// 		10xxxxxx
      }

    return strbuf.toString ();
  }
}