1
/* Inflater.java - Decompress a data stream
2
Copyright (C) 1999, 2000, 2001, 2003, 2005 Free Software Foundation, Inc.
4
This file is part of GNU Classpath.
6
GNU Classpath is free software; you can redistribute it and/or modify
7
it under the terms of the GNU General Public License as published by
8
the Free Software Foundation; either version 2, or (at your option)
11
GNU Classpath is distributed in the hope that it will be useful, but
12
WITHOUT ANY WARRANTY; without even the implied warranty of
13
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14
General Public License for more details.
16
You should have received a copy of the GNU General Public License
17
along with GNU Classpath; see the file COPYING. If not, write to the
18
Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
21
Linking this library statically or dynamically with other modules is
22
making a combined work based on this library. Thus, the terms and
23
conditions of the GNU General Public License cover the whole
26
As a special exception, the copyright holders of this library give you
27
permission to link this library with independent modules to produce an
28
executable, regardless of the license terms of these independent
29
modules, and to copy and distribute the resulting executable under
30
terms of your choice, provided that you also meet, for each linked
31
independent module, the terms and conditions of the license of that
32
module. An independent module is a module which is not derived from
33
or based on this library. If you modify this library, you may extend
34
this exception to your version of the library, but you are not
35
obligated to do so. If you do not wish to do so, delete this
36
exception statement from your version. */
38
package java.util.zip;
40
/* Written using on-line Java Platform 1.2 API Specification
42
* Believed complete and correct.
46
* Inflater is used to decompress data that has been compressed according
47
* to the "deflate" standard described in rfc1950.
49
* The usage is as following. First you have to set some input with
50
* <code>setInput()</code>, then inflate() it. If inflate doesn't
51
* inflate any bytes there may be three reasons:
53
* <li>needsInput() returns true because the input buffer is empty.
54
* You have to provide more input with <code>setInput()</code>.
55
* NOTE: needsInput() also returns true when, the stream is finished.
57
* <li>needsDictionary() returns true, you have to provide a preset
58
* dictionary with <code>setDictionary()</code>.</li>
59
* <li>finished() returns true, the inflater has finished.</li>
61
* Once the first output byte is produced, a dictionary will not be
62
* needed at a later stage.
64
* @author John Leuner, Jochen Hoenicke
71
/* Copy lengths for literal codes 257..285 */
72
private static final int CPLENS[] =
74
3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
75
35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258
78
/* Extra bits for literal codes 257..285 */
79
private static final int CPLEXT[] =
81
0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
82
3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0
85
/* Copy offsets for distance codes 0..29 */
86
private static final int CPDIST[] = {
87
1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
88
257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
89
8193, 12289, 16385, 24577
92
/* Extra bits for distance codes */
93
private static final int CPDEXT[] = {
94
0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
95
7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
99
/* This are the state in which the inflater can be. */
100
private static final int DECODE_HEADER = 0;
101
private static final int DECODE_DICT = 1;
102
private static final int DECODE_BLOCKS = 2;
103
private static final int DECODE_STORED_LEN1 = 3;
104
private static final int DECODE_STORED_LEN2 = 4;
105
private static final int DECODE_STORED = 5;
106
private static final int DECODE_DYN_HEADER = 6;
107
private static final int DECODE_HUFFMAN = 7;
108
private static final int DECODE_HUFFMAN_LENBITS = 8;
109
private static final int DECODE_HUFFMAN_DIST = 9;
110
private static final int DECODE_HUFFMAN_DISTBITS = 10;
111
private static final int DECODE_CHKSUM = 11;
112
private static final int FINISHED = 12;
114
/** This variable contains the current state. */
118
* The adler checksum of the dictionary or of the decompressed
119
* stream, as it is written in the header resp. footer of the
120
* compressed stream. <br>
122
* Only valid if mode is DECODE_DICT or DECODE_CHKSUM.
124
private int readAdler;
126
* The number of bits needed to complete the current state. This
127
* is valid, if mode is DECODE_DICT, DECODE_CHKSUM,
128
* DECODE_HUFFMAN_LENBITS or DECODE_HUFFMAN_DISTBITS.
130
private int neededBits;
131
private int repLength, repDist;
132
private int uncomprLen;
134
* True, if the last block flag was set in the last block of the
135
* inflated stream. This means that the stream ends after the
138
private boolean isLastBlock;
141
* The total number of inflated bytes.
143
private long totalOut;
145
* The total number of bytes set with setInput(). This is not the
146
* value returned by getTotalIn(), since this also includes the
149
private long totalIn;
151
* This variable stores the nowrap flag that was given to the constructor.
152
* True means, that the inflated stream doesn't contain a header nor the
153
* checksum in the footer.
155
private boolean nowrap;
157
private StreamManipulator input;
158
private OutputWindow outputWindow;
159
private InflaterDynHeader dynHeader;
160
private InflaterHuffmanTree litlenTree, distTree;
161
private Adler32 adler;
164
* Creates a new inflater.
172
* Creates a new inflater.
173
* @param nowrap true if no header and checksum field appears in the
174
* stream. This is used for GZIPed input. For compatibility with
175
* Sun JDK you should provide one byte of input more than needed in
178
public Inflater (boolean nowrap)
180
this.nowrap = nowrap;
181
this.adler = new Adler32();
182
input = new StreamManipulator();
183
outputWindow = new OutputWindow();
184
mode = nowrap ? DECODE_BLOCKS : DECODE_HEADER;
188
* Finalizes this object.
190
protected void finalize ()
192
/* Exists only for compatibility */
196
* Frees all objects allocated by the inflater. There's no reason
197
* to call this, since you can just rely on garbage collection (even
198
* for the Sun implementation). Exists only for compatibility
199
* with Sun's JDK, where the compressor allocates native memory.
200
* If you call any method (even reset) afterwards the behaviour is
214
* Returns true, if the inflater has finished. This means, that no
215
* input is needed and no output can be produced.
217
public boolean finished()
219
return mode == FINISHED
220
&& (outputWindow == null || outputWindow.getAvailable() == 0);
224
* Gets the adler checksum. This is either the checksum of all
225
* uncompressed bytes returned by inflate(), or if needsDictionary()
226
* returns true (and thus no output was yet produced) this is the
227
* adler checksum of the expected dictionary.
228
* @returns the adler checksum.
230
public int getAdler()
232
return needsDictionary() ? readAdler : (int) adler.getValue();
236
* Gets the number of unprocessed input. Useful, if the end of the
237
* stream is reached and you want to further process the bytes after
238
* the deflate stream.
239
* @return the number of bytes of the input which were not processed.
241
public int getRemaining()
243
return input.getAvailableBytes();
247
* Gets the total number of processed compressed input bytes.
248
* @return the total number of bytes of processed input bytes.
250
public int getTotalIn()
252
return (int) (totalIn - getRemaining());
256
* Gets the total number of processed compressed input bytes.
257
* @return the total number of bytes of processed input bytes.
260
public long getBytesRead()
262
return totalIn - getRemaining();
266
* Gets the total number of output bytes returned by inflate().
267
* @return the total number of output bytes.
269
public int getTotalOut()
271
return (int) totalOut;
275
* Gets the total number of output bytes returned by inflate().
276
* @return the total number of output bytes.
279
public long getBytesWritten()
285
* Inflates the compressed stream to the output buffer. If this
286
* returns 0, you should check, whether needsDictionary(),
287
* needsInput() or finished() returns true, to determine why no
288
* further output is produced.
289
* @param buf the output buffer.
290
* @return the number of bytes written to the buffer, 0 if no further
291
* output can be produced.
292
* @exception DataFormatException if deflated stream is invalid.
293
* @exception IllegalArgumentException if buf has length 0.
295
public int inflate (byte[] buf) throws DataFormatException
297
return inflate (buf, 0, buf.length);
301
* Inflates the compressed stream to the output buffer. If this
302
* returns 0, you should check, whether needsDictionary(),
303
* needsInput() or finished() returns true, to determine why no
304
* further output is produced.
305
* @param buf the output buffer.
306
* @param off the offset into buffer where the output should start.
307
* @param len the maximum length of the output.
308
* @return the number of bytes written to the buffer, 0 if no further
309
* output can be produced.
310
* @exception DataFormatException if deflated stream is invalid.
311
* @exception IndexOutOfBoundsException if the off and/or len are wrong.
313
public int inflate (byte[] buf, int off, int len) throws DataFormatException
315
/* Check for correct buff, off, len triple */
316
if (0 > off || off > off + len || off + len > buf.length)
317
throw new ArrayIndexOutOfBoundsException();
321
if (outputWindow.getAvailable() == 0)
328
int more = outputWindow.copyOutput(buf, off, len);
329
adler.update(buf, off, more);
342
* Returns true, if a preset dictionary is needed to inflate the input.
344
public boolean needsDictionary ()
346
return mode == DECODE_DICT && neededBits == 0;
350
* Returns true, if the input buffer is empty.
351
* You should then call setInput(). <br>
353
* <em>NOTE</em>: This method also returns true when the stream is finished.
355
public boolean needsInput ()
357
return input == null || input.needsInput ();
361
* Resets the inflater so that a new stream can be decompressed. All
362
* pending input and output will be discarded.
366
mode = nowrap ? DECODE_BLOCKS : DECODE_HEADER;
367
totalIn = totalOut = 0;
369
outputWindow.reset();
378
* Sets the preset dictionary. This should only be called, if
379
* needsDictionary() returns true and it should set the same
380
* dictionary, that was used for deflating. The getAdler()
381
* function returns the checksum of the dictionary needed.
382
* @param buffer the dictionary.
383
* @exception IllegalStateException if no dictionary is needed.
384
* @exception IllegalArgumentException if the dictionary checksum is
387
public void setDictionary (byte[] buffer)
389
setDictionary(buffer, 0, buffer.length);
393
* Sets the preset dictionary. This should only be called, if
394
* needsDictionary() returns true and it should set the same
395
* dictionary, that was used for deflating. The getAdler()
396
* function returns the checksum of the dictionary needed.
397
* @param buffer the dictionary.
398
* @param off the offset into buffer where the dictionary starts.
399
* @param len the length of the dictionary.
400
* @exception IllegalStateException if no dictionary is needed.
401
* @exception IllegalArgumentException if the dictionary checksum is
403
* @exception IndexOutOfBoundsException if the off and/or len are wrong.
405
public void setDictionary (byte[] buffer, int off, int len)
407
if (!needsDictionary())
408
throw new IllegalStateException();
410
adler.update(buffer, off, len);
411
if ((int) adler.getValue() != readAdler)
412
throw new IllegalArgumentException("Wrong adler checksum");
414
outputWindow.copyDict(buffer, off, len);
415
mode = DECODE_BLOCKS;
419
* Sets the input. This should only be called, if needsInput()
421
* @param buf the input.
422
* @exception IllegalStateException if no input is needed.
424
public void setInput (byte[] buf)
426
setInput (buf, 0, buf.length);
430
* Sets the input. This should only be called, if needsInput()
432
* @param buf the input.
433
* @param off the offset into buffer where the input starts.
434
* @param len the length of the input.
435
* @exception IllegalStateException if no input is needed.
436
* @exception IndexOutOfBoundsException if the off and/or len are wrong.
438
public void setInput (byte[] buf, int off, int len)
440
input.setInput (buf, off, len);
445
* Decodes the deflate header.
446
* @return false if more input is needed.
447
* @exception DataFormatException if header is invalid.
449
private boolean decodeHeader () throws DataFormatException
451
int header = input.peekBits(16);
456
/* The header is written in "wrong" byte order */
457
header = ((header << 8) | (header >> 8)) & 0xffff;
458
if (header % 31 != 0)
459
throw new DataFormatException("Header checksum illegal");
461
if ((header & 0x0f00) != (Deflater.DEFLATED << 8))
462
throw new DataFormatException("Compression Method unknown");
464
/* Maximum size of the backwards window in bits.
465
* We currently ignore this, but we could use it to make the
466
* inflater window more space efficient. On the other hand the
467
* full window (15 bits) is needed most times, anyway.
468
int max_wbits = ((header & 0x7000) >> 12) + 8;
471
if ((header & 0x0020) == 0) // Dictionary flag?
473
mode = DECODE_BLOCKS;
484
* Decodes the dictionary checksum after the deflate header.
485
* @return false if more input is needed.
487
private boolean decodeDict ()
489
while (neededBits > 0)
491
int dictByte = input.peekBits(8);
495
readAdler = (readAdler << 8) | dictByte;
502
* Decodes the huffman encoded symbols in the input stream.
503
* @return false if more input is needed, true if output window is
504
* full or the current block ends.
505
* @exception DataFormatException if deflated stream is invalid.
507
private boolean decodeHuffman () throws DataFormatException
509
int free = outputWindow.getFreeSpace();
516
/* This is the inner loop so it is optimized a bit */
517
while (((symbol = litlenTree.getSymbol(input)) & ~0xff) == 0)
519
outputWindow.write(symbol);
529
/* symbol == 256: end of block */
532
mode = DECODE_BLOCKS;
539
repLength = CPLENS[symbol - 257];
540
neededBits = CPLEXT[symbol - 257];
542
catch (ArrayIndexOutOfBoundsException ex)
544
throw new DataFormatException("Illegal rep length code");
547
case DECODE_HUFFMAN_LENBITS:
550
mode = DECODE_HUFFMAN_LENBITS;
551
int i = input.peekBits(neededBits);
554
input.dropBits(neededBits);
557
mode = DECODE_HUFFMAN_DIST;
559
case DECODE_HUFFMAN_DIST:
560
symbol = distTree.getSymbol(input);
565
repDist = CPDIST[symbol];
566
neededBits = CPDEXT[symbol];
568
catch (ArrayIndexOutOfBoundsException ex)
570
throw new DataFormatException("Illegal rep dist code");
573
case DECODE_HUFFMAN_DISTBITS:
576
mode = DECODE_HUFFMAN_DISTBITS;
577
int i = input.peekBits(neededBits);
580
input.dropBits(neededBits);
583
outputWindow.repeat(repLength, repDist);
585
mode = DECODE_HUFFMAN;
588
throw new IllegalStateException();
595
* Decodes the adler checksum after the deflate stream.
596
* @return false if more input is needed.
597
* @exception DataFormatException if checksum doesn't match.
599
private boolean decodeChksum () throws DataFormatException
601
while (neededBits > 0)
603
int chkByte = input.peekBits(8);
607
readAdler = (readAdler << 8) | chkByte;
610
if ((int) adler.getValue() != readAdler)
611
throw new DataFormatException("Adler chksum doesn't match: "
612
+Integer.toHexString((int)adler.getValue())
613
+" vs. "+Integer.toHexString(readAdler));
619
* Decodes the deflated stream.
620
* @return false if more input is needed, or if finished.
621
* @exception DataFormatException if deflated stream is invalid.
623
private boolean decode () throws DataFormatException
628
return decodeHeader();
632
return decodeChksum();
644
input.skipToByteBoundary();
646
mode = DECODE_CHKSUM;
651
int type = input.peekBits(3);
660
case DeflaterConstants.STORED_BLOCK:
661
input.skipToByteBoundary();
662
mode = DECODE_STORED_LEN1;
664
case DeflaterConstants.STATIC_TREES:
665
litlenTree = InflaterHuffmanTree.defLitLenTree;
666
distTree = InflaterHuffmanTree.defDistTree;
667
mode = DECODE_HUFFMAN;
669
case DeflaterConstants.DYN_TREES:
670
dynHeader = new InflaterDynHeader();
671
mode = DECODE_DYN_HEADER;
674
throw new DataFormatException("Unknown block type "+type);
678
case DECODE_STORED_LEN1:
680
if ((uncomprLen = input.peekBits(16)) < 0)
683
mode = DECODE_STORED_LEN2;
686
case DECODE_STORED_LEN2:
688
int nlen = input.peekBits(16);
692
if (nlen != (uncomprLen ^ 0xffff))
693
throw new DataFormatException("broken uncompressed block");
694
mode = DECODE_STORED;
699
int more = outputWindow.copyStored(input, uncomprLen);
703
mode = DECODE_BLOCKS;
706
return !input.needsInput();
709
case DECODE_DYN_HEADER:
710
if (!dynHeader.decode(input))
712
litlenTree = dynHeader.buildLitLenTree();
713
distTree = dynHeader.buildDistTree();
714
mode = DECODE_HUFFMAN;
717
case DECODE_HUFFMAN_LENBITS:
718
case DECODE_HUFFMAN_DIST:
719
case DECODE_HUFFMAN_DISTBITS:
720
return decodeHuffman();
724
throw new IllegalStateException();