001 /* DataInputStream.java -- FilteredInputStream that implements DataInput 002 Copyright (C) 1998, 1999, 2000, 2001, 2003, 2005 Free Software Foundation 003 004 This file is part of GNU Classpath. 005 006 GNU Classpath is free software; you can redistribute it and/or modify 007 it under the terms of the GNU General Public License as published by 008 the Free Software Foundation; either version 2, or (at your option) 009 any later version. 010 011 GNU Classpath is distributed in the hope that it will be useful, but 012 WITHOUT ANY WARRANTY; without even the implied warranty of 013 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 014 General Public License for more details. 015 016 You should have received a copy of the GNU General Public License 017 along with GNU Classpath; see the file COPYING. If not, write to the 018 Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 019 02110-1301 USA. 020 021 Linking this library statically or dynamically with other modules is 022 making a combined work based on this library. Thus, the terms and 023 conditions of the GNU General Public License cover the whole 024 combination. 025 026 As a special exception, the copyright holders of this library give you 027 permission to link this library with independent modules to produce an 028 executable, regardless of the license terms of these independent 029 modules, and to copy and distribute the resulting executable under 030 terms of your choice, provided that you also meet, for each linked 031 independent module, the terms and conditions of the license of that 032 module. An independent module is a module which is not derived from 033 or based on this library. If you modify this library, you may extend 034 this exception to your version of the library, but you are not 035 obligated to do so. If you do not wish to do so, delete this 036 exception statement from your version. */ 037 038 package java.io; 039 040 /* Written using "Java Class Libraries", 2nd edition, ISBN 0-201-31002-3 041 * "The Java Language Specification", ISBN 0-201-63451-1 042 * plus online API docs for JDK 1.2 beta from http://www.javasoft.com. 043 * Status: Believed complete and correct. 044 */ 045 046 /** 047 * This subclass of <code>FilteredInputStream</code> implements the 048 * <code>DataInput</code> interface that provides method for reading primitive 049 * Java data types from a stream. 050 * 051 * @see DataInput 052 * 053 * @author Warren Levy (warrenl@cygnus.com) 054 * @author Aaron M. Renn (arenn@urbanophile.com) 055 * @date October 20, 1998. 056 */ 057 public class DataInputStream extends FilterInputStream implements DataInput 058 { 059 // Byte buffer, used to make primitive read calls more efficient. 060 byte[] buf = new byte [8]; 061 062 /** 063 * This constructor initializes a new <code>DataInputStream</code> 064 * to read from the specified subordinate stream. 065 * 066 * @param in The subordinate <code>InputStream</code> to read from 067 */ 068 public DataInputStream (InputStream in) 069 { 070 super (in); 071 } 072 073 /** 074 * This method reads bytes from the underlying stream into the specified 075 * byte array buffer. It will attempt to fill the buffer completely, but 076 * may return a short count if there is insufficient data remaining to be 077 * read to fill the buffer. 078 * 079 * @param b The buffer into which bytes will be read. 080 * 081 * @return The actual number of bytes read, or -1 if end of stream reached 082 * before reading any bytes. 083 * 084 * @exception IOException If an error occurs. 085 */ 086 public final int read (byte[] b) throws IOException 087 { 088 return in.read (b, 0, b.length); 089 } 090 091 /** 092 * This method reads bytes from the underlying stream into the specified 093 * byte array buffer. It will attempt to read <code>len</code> bytes and 094 * will start storing them at position <code>off</code> into the buffer. 095 * This method can return a short count if there is insufficient data 096 * remaining to be read to complete the desired read length. 097 * 098 * @param b The buffer into which bytes will be read. 099 * @param off The offset into the buffer to start storing bytes. 100 * @param len The requested number of bytes to read. 101 * 102 * @return The actual number of bytes read, or -1 if end of stream reached 103 * before reading any bytes. 104 * 105 * @exception IOException If an error occurs. 106 */ 107 public final int read (byte[] b, int off, int len) throws IOException 108 { 109 return in.read (b, off, len); 110 } 111 112 /** 113 * This method reads a Java boolean value from an input stream. It does 114 * so by reading a single byte of data. If that byte is zero, then the 115 * value returned is <code>false</code>. If the byte is non-zero, then 116 * the value returned is <code>true</code>. 117 * <p> 118 * This method can read a <code>boolean</code> written by an object 119 * implementing the <code>writeBoolean()</code> method in the 120 * <code>DataOutput</code> interface. 121 * 122 * @return The <code>boolean</code> value read 123 * 124 * @exception EOFException If end of file is reached before reading 125 * the boolean 126 * @exception IOException If any other error occurs 127 * 128 * @see DataOutput#writeBoolean 129 */ 130 public final boolean readBoolean () throws IOException 131 { 132 return convertToBoolean (in.read ()); 133 } 134 135 /** 136 * This method reads a Java byte value from an input stream. The value 137 * is in the range of -128 to 127. 138 * <p> 139 * This method can read a <code>byte</code> written by an object 140 * implementing the <code>writeByte()</code> method in the 141 * <code>DataOutput</code> interface. 142 * 143 * @return The <code>byte</code> value read 144 * 145 * @exception EOFException If end of file is reached before reading the byte 146 * @exception IOException If any other error occurs 147 * 148 * @see DataOutput#writeByte 149 */ 150 public final byte readByte () throws IOException 151 { 152 return convertToByte (in.read ()); 153 } 154 155 /** 156 * This method reads a Java <code>char</code> value from an input stream. 157 * It operates by reading two bytes from the stream and converting them to 158 * a single 16-bit Java <code>char</code>. The two bytes are stored most 159 * significant byte first (i.e., "big endian") regardless of the native 160 * host byte ordering. 161 * <p> 162 * As an example, if <code>byte1</code> and <code>byte2</code> 163 * represent the first and second byte read from the stream 164 * respectively, they will be transformed to a <code>char</code> in 165 * the following manner: 166 * <p> 167 * <code>(char)(((byte1 & 0xFF) << 8) | (byte2 & 0xFF)</code> 168 * <p> 169 * This method can read a <code>char</code> written by an object 170 * implementing the <code>writeChar()</code> method in the 171 * <code>DataOutput</code> interface. 172 * 173 * @return The <code>char</code> value read 174 * 175 * @exception EOFException If end of file is reached before reading the char 176 * @exception IOException If any other error occurs 177 * 178 * @see DataOutput#writeChar 179 */ 180 public final char readChar () throws IOException 181 { 182 readFully (buf, 0, 2); 183 return convertToChar (buf); 184 } 185 186 /** 187 * This method reads a Java double value from an input stream. It operates 188 * by first reading a <code>long</code> value from the stream by calling the 189 * <code>readLong()</code> method in this interface, then converts 190 * that <code>long</code> to a <code>double</code> using the 191 * <code>longBitsToDouble</code> method in the class 192 * <code>java.lang.Double</code> 193 * <p> 194 * This method can read a <code>double</code> written by an object 195 * implementing the <code>writeDouble()</code> method in the 196 * <code>DataOutput</code> interface. 197 * 198 * @return The <code>double</code> value read 199 * 200 * @exception EOFException If end of file is reached before reading 201 * the double 202 * @exception IOException If any other error occurs 203 * 204 * @see DataOutput#writeDouble 205 * @see java.lang.Double#longBitsToDouble 206 */ 207 public final double readDouble () throws IOException 208 { 209 return Double.longBitsToDouble (readLong ()); 210 } 211 212 /** 213 * This method reads a Java float value from an input stream. It 214 * operates by first reading an <code>int</code> value from the 215 * stream by calling the <code>readInt()</code> method in this 216 * interface, then converts that <code>int</code> to a 217 * <code>float</code> using the <code>intBitsToFloat</code> method 218 * in the class <code>java.lang.Float</code> 219 * <p> 220 * This method can read a <code>float</code> written by an object 221 * implementing the <code>writeFloat()</code> method in the 222 * <code>DataOutput</code> interface. 223 * 224 * @return The <code>float</code> value read 225 * 226 * @exception EOFException If end of file is reached before reading the float 227 * @exception IOException If any other error occurs 228 * 229 * @see DataOutput#writeFloat 230 * @see java.lang.Float#intBitsToFloat 231 */ 232 public final float readFloat () throws IOException 233 { 234 return Float.intBitsToFloat (readInt ()); 235 } 236 237 /** 238 * This method reads raw bytes into the passed array until the array is 239 * full. Note that this method blocks until the data is available and 240 * throws an exception if there is not enough data left in the stream to 241 * fill the buffer. Note also that zero length buffers are permitted. 242 * In this case, the method will return immediately without reading any 243 * bytes from the stream. 244 * 245 * @param b The buffer into which to read the data 246 * 247 * @exception EOFException If end of file is reached before filling the 248 * buffer 249 * @exception IOException If any other error occurs 250 */ 251 public final void readFully (byte[] b) throws IOException 252 { 253 readFully (b, 0, b.length); 254 } 255 256 /** 257 * This method reads raw bytes into the passed array <code>buf</code> 258 * starting 259 * <code>offset</code> bytes into the buffer. The number of bytes read 260 * will be 261 * exactly <code>len</code>. Note that this method blocks until the data is 262 * available and throws an exception if there is not enough data left in 263 * the stream to read <code>len</code> bytes. Note also that zero length 264 * buffers are permitted. In this case, the method will return immediately 265 * without reading any bytes from the stream. 266 * 267 * @param buf The buffer into which to read the data 268 * @param offset The offset into the buffer to start storing data 269 * @param len The number of bytes to read into the buffer 270 * 271 * @exception EOFException If end of file is reached before filling the 272 * buffer 273 * @exception IOException If any other error occurs 274 */ 275 public final void readFully (byte[] buf, int offset, int len) throws IOException 276 { 277 if (len < 0) 278 throw new IndexOutOfBoundsException("Negative length: " + len); 279 280 while (len > 0) 281 { 282 // in.read will block until some data is available. 283 int numread = in.read (buf, offset, len); 284 if (numread < 0) 285 throw new EOFException (); 286 len -= numread; 287 offset += numread; 288 } 289 } 290 291 /** 292 * This method reads a Java <code>int</code> value from an input stream 293 * It operates by reading four bytes from the stream and converting them to 294 * a single Java <code>int</code>. The bytes are stored most 295 * significant byte first (i.e., "big endian") regardless of the native 296 * host byte ordering. 297 * <p> 298 * As an example, if <code>byte1</code> through <code>byte4</code> represent 299 * the first four bytes read from the stream, they will be 300 * transformed to an <code>int</code> in the following manner: 301 * <p> 302 * <code>(int)(((byte1 & 0xFF) << 24) + ((byte2 & 0xFF) << 16) + 303 * ((byte3 & 0xFF)<< 8) + (byte4 & 0xFF)))</code> 304 * <p> 305 * The value returned is in the range of -2147483648 to 2147483647. 306 * <p> 307 * This method can read an <code>int</code> written by an object 308 * implementing the <code>writeInt()</code> method in the 309 * <code>DataOutput</code> interface. 310 * 311 * @return The <code>int</code> value read 312 * 313 * @exception EOFException If end of file is reached before reading the int 314 * @exception IOException If any other error occurs 315 * 316 * @see DataOutput#writeInt 317 */ 318 public final int readInt () throws IOException 319 { 320 readFully (buf, 0, 4); 321 return convertToInt (buf); 322 } 323 324 /** 325 * This method reads the next line of text data from an input 326 * stream. It operates by reading bytes and converting those bytes 327 * to <code>char</code> values by treating the byte read as the low 328 * eight bits of the <code>char</code> and using 0 as the high eight 329 * bits. Because of this, it does not support the full 16-bit 330 * Unicode character set. 331 * <p> 332 * The reading of bytes ends when either the end of file or a line 333 * terminator is encountered. The bytes read are then returned as a 334 * <code>String</code> A line terminator is a byte sequence 335 * consisting of either <code>\r</code>, <code>\n</code> or 336 * <code>\r\n</code>. These termination charaters are discarded and 337 * are not returned as part of the string. 338 * <p> 339 * This method can read data that was written by an object implementing the 340 * <code>writeLine()</code> method in <code>DataOutput</code>. 341 * 342 * @return The line read as a <code>String</code> 343 * 344 * @exception IOException If an error occurs 345 * 346 * @see DataOutput 347 * 348 * @deprecated 349 */ 350 public final String readLine() throws IOException 351 { 352 StringBuffer strb = new StringBuffer(); 353 354 while (true) 355 { 356 int c = in.read(); 357 if (c == -1) // got an EOF 358 return strb.length() > 0 ? strb.toString() : null; 359 if (c == '\r') 360 { 361 int next_c = in.read(); 362 if (next_c != '\n' && next_c != -1) 363 { 364 if (!(in instanceof PushbackInputStream)) 365 in = new PushbackInputStream(in); 366 ((PushbackInputStream) in).unread(next_c); 367 } 368 break; 369 } 370 if (c == '\n') 371 break; 372 strb.append((char) c); 373 } 374 375 return strb.length() > 0 ? strb.toString() : ""; 376 } 377 378 /** 379 * This method reads a Java <code>long</code> value from an input stream 380 * It operates by reading eight bytes from the stream and converting them to 381 * a single Java <code>long</code>. The bytes are stored most 382 * significant byte first (i.e., "big endian") regardless of the native 383 * host byte ordering. 384 * <p> 385 * As an example, if <code>byte1</code> through <code>byte8</code> represent 386 * the first eight bytes read from the stream, they will be 387 * transformed to an <code>long</code> in the following manner: 388 * <p> 389 * <code>(long)(((byte1 & 0xFF) << 56) + ((byte2 & 0xFF) << 48) + 390 * ((byte3 & 0xFF) << 40) + ((byte4 & 0xFF) << 32) + 391 * ((byte5 & 0xFF) << 24) + ((byte6 & 0xFF) << 16) + 392 * ((byte7 & 0xFF) << 8) + (byte8 & 0xFF))) 393 * </code> 394 * <p> 395 * The value returned is in the range of -9223372036854775808 to 396 * 9223372036854775807. 397 * <p> 398 * This method can read an <code>long</code> written by an object 399 * implementing the <code>writeLong()</code> method in the 400 * <code>DataOutput</code> interface. 401 * 402 * @return The <code>long</code> value read 403 * 404 * @exception EOFException If end of file is reached before reading the long 405 * @exception IOException If any other error occurs 406 * 407 * @see DataOutput#writeLong 408 */ 409 public final long readLong () throws IOException 410 { 411 readFully (buf, 0, 8); 412 return convertToLong (buf); 413 } 414 415 /** 416 * This method reads a signed 16-bit value into a Java in from the 417 * stream. It operates by reading two bytes from the stream and 418 * converting them to a single 16-bit Java <code>short</code>. The 419 * two bytes are stored most significant byte first (i.e., "big 420 * endian") regardless of the native host byte ordering. 421 * <p> 422 * As an example, if <code>byte1</code> and <code>byte2</code> 423 * represent the first and second byte read from the stream 424 * respectively, they will be transformed to a <code>short</code>. in 425 * the following manner: 426 * <p> 427 * <code>(short)(((byte1 & 0xFF) << 8) | (byte2 & 0xFF))</code> 428 * <p> 429 * The value returned is in the range of -32768 to 32767. 430 * <p> 431 * This method can read a <code>short</code> written by an object 432 * implementing the <code>writeShort()</code> method in the 433 * <code>DataOutput</code> interface. 434 * 435 * @return The <code>short</code> value read 436 * 437 * @exception EOFException If end of file is reached before reading the value 438 * @exception IOException If any other error occurs 439 * 440 * @see DataOutput#writeShort 441 */ 442 public final short readShort () throws IOException 443 { 444 readFully (buf, 0, 2); 445 return convertToShort (buf); 446 } 447 448 /** 449 * This method reads 8 unsigned bits into a Java <code>int</code> 450 * value from the stream. The value returned is in the range of 0 to 451 * 255. 452 * <p> 453 * This method can read an unsigned byte written by an object 454 * implementing the <code>writeUnsignedByte()</code> method in the 455 * <code>DataOutput</code> interface. 456 * 457 * @return The unsigned bytes value read as a Java <code>int</code>. 458 * 459 * @exception EOFException If end of file is reached before reading the value 460 * @exception IOException If any other error occurs 461 * 462 * @see DataOutput#writeByte 463 */ 464 public final int readUnsignedByte () throws IOException 465 { 466 return convertToUnsignedByte (in.read ()); 467 } 468 469 /** 470 * This method reads 16 unsigned bits into a Java int value from the stream. 471 * It operates by reading two bytes from the stream and converting them to 472 * a single Java <code>int</code> The two bytes are stored most 473 * significant byte first (i.e., "big endian") regardless of the native 474 * host byte ordering. 475 * <p> 476 * As an example, if <code>byte1</code> and <code>byte2</code> 477 * represent the first and second byte read from the stream 478 * respectively, they will be transformed to an <code>int</code> in 479 * the following manner: 480 * <p> 481 * <code>(int)(((byte1 & 0xFF) << 8) + (byte2 & 0xFF))</code> 482 * <p> 483 * The value returned is in the range of 0 to 65535. 484 * <p> 485 * This method can read an unsigned short written by an object 486 * implementing the <code>writeUnsignedShort()</code> method in the 487 * <code>DataOutput</code> interface. 488 * 489 * @return The unsigned short value read as a Java <code>int</code> 490 * 491 * @exception EOFException If end of file is reached before reading the value 492 * @exception IOException If any other error occurs 493 * 494 * @see DataOutput#writeShort 495 */ 496 public final int readUnsignedShort () throws IOException 497 { 498 readFully (buf, 0, 2); 499 return convertToUnsignedShort (buf); 500 } 501 502 /** 503 * This method reads a <code>String</code> from an input stream that 504 * is encoded in a modified UTF-8 format. This format has a leading 505 * two byte sequence that contains the remaining number of bytes to 506 * read. This two byte sequence is read using the 507 * <code>readUnsignedShort()</code> method of this interface. 508 * <p> 509 * After the number of remaining bytes have been determined, these 510 * bytes are read an transformed into <code>char</code> values. 511 * These <code>char</code> values are encoded in the stream using 512 * either a one, two, or three byte format. The particular format 513 * in use can be determined by examining the first byte read. 514 * <p> 515 * If the first byte has a high order bit of 0, then that character 516 * consists on only one byte. This character value consists of 517 * seven bits that are at positions 0 through 6 of the byte. As an 518 * example, if <code>byte1</code> is the byte read from the stream, 519 * it would be converted to a <code>char</code> like so: 520 * <p> 521 * <code>(char)byte1</code> 522 * <p> 523 * If the first byte has 110 as its high order bits, then the 524 * character consists of two bytes. The bits that make up the character 525 * value are in positions 0 through 4 of the first byte and bit positions 526 * 0 through 5 of the second byte. (The second byte should have 527 * 10 as its high order bits). These values are in most significant 528 * byte first (i.e., "big endian") order. 529 * <p> 530 * As an example, if <code>byte1</code> and <code>byte2</code> are 531 * the first two bytes read respectively, and the high order bits of 532 * them match the patterns which indicate a two byte character 533 * encoding, then they would be converted to a Java 534 * <code>char</code> like so: 535 * <p> 536 * <code>(char)(((byte1 & 0x1F) << 6) | (byte2 & 0x3F))</code> 537 * <p> 538 * If the first byte has a 1110 as its high order bits, then the 539 * character consists of three bytes. The bits that make up the character 540 * value are in positions 0 through 3 of the first byte and bit positions 541 * 0 through 5 of the other two bytes. (The second and third bytes should 542 * have 10 as their high order bits). These values are in most 543 * significant byte first (i.e., "big endian") order. 544 * <p> 545 * As an example, if <code>byte1</code> <code>byte2</code> and 546 * <code>byte3</code> are the three bytes read, and the high order 547 * bits of them match the patterns which indicate a three byte 548 * character encoding, then they would be converted to a Java 549 * <code>char</code> like so: 550 * <p> 551 * <code>(char)(((byte1 & 0x0F) << 12) | ((byte2 & 0x3F) << 6) | 552 * (byte3 & 0x3F))</code> 553 * <p> 554 * Note that all characters are encoded in the method that requires 555 * the fewest number of bytes with the exception of the character 556 * with the value of <code>\u0000</code> which is encoded as two 557 * bytes. This is a modification of the UTF standard used to 558 * prevent C language style <code>NUL</code> values from appearing 559 * in the byte stream. 560 * <p> 561 * This method can read data that was written by an object implementing the 562 * <code>writeUTF()</code> method in <code>DataOutput</code> 563 * 564 * @return The <code>String</code> read 565 * 566 * @exception EOFException If end of file is reached before reading 567 * the String 568 * @exception UTFDataFormatException If the data is not in UTF-8 format 569 * @exception IOException If any other error occurs 570 * 571 * @see DataOutput#writeUTF 572 */ 573 public final String readUTF () throws IOException 574 { 575 return readUTF (this); 576 } 577 578 /** 579 * This method reads a String encoded in UTF-8 format from the 580 * specified <code>DataInput</code> source. 581 * 582 * @param in The <code>DataInput</code> source to read from 583 * 584 * @return The String read from the source 585 * 586 * @exception IOException If an error occurs 587 * 588 * @see DataInput#readUTF 589 */ 590 public static final String readUTF(DataInput in) throws IOException 591 { 592 final int UTFlen = in.readUnsignedShort (); 593 byte[] buf = new byte [UTFlen]; 594 595 // This blocks until the entire string is available rather than 596 // doing partial processing on the bytes that are available and then 597 // blocking. An advantage of the latter is that Exceptions 598 // could be thrown earlier. The former is a bit cleaner. 599 in.readFully (buf, 0, UTFlen); 600 601 return convertFromUTF (buf); 602 } 603 604 /** 605 * This method attempts to skip and discard the specified number of bytes 606 * in the input stream. It may actually skip fewer bytes than requested. 607 * This method will not skip any bytes if passed a negative number of bytes 608 * to skip. 609 * 610 * @param n The requested number of bytes to skip. 611 * 612 * @return The requested number of bytes to skip. 613 * 614 * @exception IOException If an error occurs. 615 * @specnote The JDK docs claim that this returns the number of bytes 616 * actually skipped. The JCL claims that this method can throw an 617 * EOFException. Neither of these appear to be true in the JDK 1.3's 618 * implementation. This tries to implement the actual JDK behaviour. 619 */ 620 public final int skipBytes (int n) throws IOException 621 { 622 if (n <= 0) 623 return 0; 624 try 625 { 626 return (int) in.skip (n); 627 } 628 catch (EOFException x) 629 { 630 // do nothing. 631 } 632 return n; 633 } 634 635 static boolean convertToBoolean (int b) throws EOFException 636 { 637 if (b < 0) 638 throw new EOFException (); 639 640 return (b != 0); 641 } 642 643 static byte convertToByte (int i) throws EOFException 644 { 645 if (i < 0) 646 throw new EOFException (); 647 648 return (byte) i; 649 } 650 651 static int convertToUnsignedByte (int i) throws EOFException 652 { 653 if (i < 0) 654 throw new EOFException (); 655 656 return (i & 0xFF); 657 } 658 659 static char convertToChar (byte[] buf) 660 { 661 return (char) ((buf [0] << 8) 662 | (buf [1] & 0xff)); 663 } 664 665 static short convertToShort (byte[] buf) 666 { 667 return (short) ((buf [0] << 8) 668 | (buf [1] & 0xff)); 669 } 670 671 static int convertToUnsignedShort (byte[] buf) 672 { 673 return (((buf [0] & 0xff) << 8) 674 | (buf [1] & 0xff)); 675 } 676 677 static int convertToInt (byte[] buf) 678 { 679 return (((buf [0] & 0xff) << 24) 680 | ((buf [1] & 0xff) << 16) 681 | ((buf [2] & 0xff) << 8) 682 | (buf [3] & 0xff)); 683 } 684 685 static long convertToLong (byte[] buf) 686 { 687 return (((long)(buf [0] & 0xff) << 56) | 688 ((long)(buf [1] & 0xff) << 48) | 689 ((long)(buf [2] & 0xff) << 40) | 690 ((long)(buf [3] & 0xff) << 32) | 691 ((long)(buf [4] & 0xff) << 24) | 692 ((long)(buf [5] & 0xff) << 16) | 693 ((long)(buf [6] & 0xff) << 8) | 694 ((long)(buf [7] & 0xff))); 695 } 696 697 // FIXME: This method should be re-thought. I suspect we have multiple 698 // UTF-8 decoders floating around. We should use the standard charset 699 // converters, maybe and adding a direct call into one of the new 700 // NIO converters for a super-fast UTF8 decode. 701 static String convertFromUTF (byte[] buf) 702 throws EOFException, UTFDataFormatException 703 { 704 // Give StringBuffer an initial estimated size to avoid 705 // enlarge buffer frequently 706 StringBuffer strbuf = new StringBuffer (buf.length / 2 + 2); 707 708 for (int i = 0; i < buf.length; ) 709 { 710 if ((buf [i] & 0x80) == 0) // bit pattern 0xxxxxxx 711 strbuf.append ((char) (buf [i++] & 0xFF)); 712 else if ((buf [i] & 0xE0) == 0xC0) // bit pattern 110xxxxx 713 { 714 if (i + 1 >= buf.length 715 || (buf [i + 1] & 0xC0) != 0x80) 716 throw new UTFDataFormatException (); 717 718 strbuf.append((char) (((buf [i++] & 0x1F) << 6) 719 | (buf [i++] & 0x3F))); 720 } 721 else if ((buf [i] & 0xF0) == 0xE0) // bit pattern 1110xxxx 722 { 723 if (i + 2 >= buf.length 724 || (buf [i + 1] & 0xC0) != 0x80 725 || (buf [i + 2] & 0xC0) != 0x80) 726 throw new UTFDataFormatException (); 727 728 strbuf.append ((char) (((buf [i++] & 0x0F) << 12) 729 | ((buf [i++] & 0x3F) << 6) 730 | (buf [i++] & 0x3F))); 731 } 732 else // must be ((buf [i] & 0xF0) == 0xF0 || (buf [i] & 0xC0) == 0x80) 733 throw new UTFDataFormatException (); // bit patterns 1111xxxx or 734 // 10xxxxxx 735 } 736 737 return strbuf.toString (); 738 } 739 }