blob: b27ebeda8cebf0f51e2933bb76b530c7fdb45fd1 [file] [log] [blame]
Erik Andersene49d5ec2000-02-08 19:58:47 +00001/* vi: set sw=4 ts=4: */
Erik Andersen61677fe2000-04-13 01:18:56 +00002/*
3 * Gzip implementation for busybox
4 *
5 * Based on GNU gzip Copyright (C) 1992-1993 Jean-loup Gailly.
6 *
7 * Originally adjusted for busybox by Sven Rudolph <sr1@inf.tu-dresden.de>
8 * based on gzip sources
9 *
10 * Adjusted further by Erik Andersen <andersen@lineo.com>, <andersee@debian.org>
11 * to support files as well as stdin/stdout, and to generally behave itself wrt
12 * command line handling.
13 *
14 * This program is free software; you can redistribute it and/or modify
15 * it under the terms of the GNU General Public License as published by
16 * the Free Software Foundation; either version 2 of the License, or
17 * (at your option) any later version.
18 *
19 * This program is distributed in the hope that it will be useful,
20 * but WITHOUT ANY WARRANTY; without even the implied warranty of
21 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
22 * General Public License for more details.
23 *
24 * You should have received a copy of the GNU General Public License
25 * along with this program; if not, write to the Free Software
26 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
27 *
28 */
Eric Andersenb052b471999-11-18 00:21:59 +000029
30#include "internal.h"
Eric Andersen1e03add2000-07-06 09:56:35 +000031#include <getopt.h>
Erik Andersen7ab9c7e2000-05-12 19:41:47 +000032
Matt Kraaibf181b92000-07-16 20:57:15 +000033/* These defines are very important for BusyBox. Without these,
Erik Andersen61677fe2000-04-13 01:18:56 +000034 * huge chunks of ram are pre-allocated making the BusyBox bss
35 * size Freaking Huge(tm), which is a bad thing.*/
36#define SMALL_MEM
37#define DYN_ALLOC
38
Erik Andersen61677fe2000-04-13 01:18:56 +000039#define BB_DECLARE_EXTERN
Erik Andersen7ab9c7e2000-05-12 19:41:47 +000040#define bb_need_memory_exhausted
41#define bb_need_name_too_long
Erik Andersen61677fe2000-04-13 01:18:56 +000042#include "messages.c"
43
44
Eric Andersenb052b471999-11-18 00:21:59 +000045/* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
46 * Copyright (C) 1992-1993 Jean-loup Gailly
47 * The unzip code was written and put in the public domain by Mark Adler.
48 * Portions of the lzw code are derived from the public domain 'compress'
49 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
50 * Ken Turkowski, Dave Mack and Peter Jannesen.
51 *
52 * See the license_msg below and the file COPYING for the software license.
53 * See the file algorithm.doc for the compression algorithms and file formats.
54 */
55
56#if 0
Erik Andersene49d5ec2000-02-08 19:58:47 +000057static char *license_msg[] = {
58 " Copyright (C) 1992-1993 Jean-loup Gailly",
59 " This program is free software; you can redistribute it and/or modify",
60 " it under the terms of the GNU General Public License as published by",
61 " the Free Software Foundation; either version 2, or (at your option)",
62 " any later version.",
63 "",
64 " This program is distributed in the hope that it will be useful,",
65 " but WITHOUT ANY WARRANTY; without even the implied warranty of",
66 " MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the",
67 " GNU General Public License for more details.",
68 "",
69 " You should have received a copy of the GNU General Public License",
70 " along with this program; if not, write to the Free Software",
71 " Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.",
72 0
73};
Eric Andersenb052b471999-11-18 00:21:59 +000074#endif
75
76/* Compress files with zip algorithm and 'compress' interface.
77 * See usage() and help() functions below for all options.
78 * Outputs:
79 * file.gz: compressed file with same mode, owner, and utimes
80 * or stdout with -c option or if stdin used as input.
81 * If the output file name had to be truncated, the original name is kept
82 * in the compressed file.
83 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
84 *
85 * Using gz on MSDOS would create too many file name conflicts. For
86 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
87 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
88 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
89 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
90 *
91 * For the meaning of all compilation flags, see comments in Makefile.in.
92 */
93
94#include <ctype.h>
95#include <sys/types.h>
96#include <signal.h>
Eric Andersenb052b471999-11-18 00:21:59 +000097#include <errno.h>
98
99/* #include "tailor.h" */
100
101/* tailor.h -- target dependent definitions
102 * Copyright (C) 1992-1993 Jean-loup Gailly.
103 * This is free software; you can redistribute it and/or modify it under the
104 * terms of the GNU General Public License, see the file COPYING.
105 */
106
107/* The target dependent definitions should be defined here only.
108 * The target dependent functions should be defined in tailor.c.
109 */
110
111#define RECORD_IO 0
112
113#define get_char() get_byte()
114#define put_char(c) put_byte(c)
115
Eric Andersenb052b471999-11-18 00:21:59 +0000116
117/* I don't like nested includes, but the string and io functions are used
118 * too often
119 */
120#include <stdio.h>
121#if !defined(NO_STRING_H) || defined(STDC_HEADERS)
122# include <string.h>
123# if !defined(STDC_HEADERS) && !defined(NO_MEMORY_H) && !defined(__GNUC__)
124# include <memory.h>
125# endif
Erik Andersen61677fe2000-04-13 01:18:56 +0000126# define memzero(s, n) memset ((void *)(s), 0, (n))
Eric Andersenb052b471999-11-18 00:21:59 +0000127#else
128# include <strings.h>
Erik Andersene49d5ec2000-02-08 19:58:47 +0000129# define strchr index
Eric Andersenb052b471999-11-18 00:21:59 +0000130# define strrchr rindex
Erik Andersene49d5ec2000-02-08 19:58:47 +0000131# define memcpy(d, s, n) bcopy((s), (d), (n))
132# define memcmp(s1, s2, n) bcmp((s1), (s2), (n))
Eric Andersenb052b471999-11-18 00:21:59 +0000133# define memzero(s, n) bzero((s), (n))
134#endif
135
136#ifndef RETSIGTYPE
137# define RETSIGTYPE void
138#endif
139
140#define local static
141
Erik Andersene49d5ec2000-02-08 19:58:47 +0000142typedef unsigned char uch;
Eric Andersenb052b471999-11-18 00:21:59 +0000143typedef unsigned short ush;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000144typedef unsigned long ulg;
Eric Andersenb052b471999-11-18 00:21:59 +0000145
146/* Return codes from gzip */
147#define OK 0
148#define ERROR 1
149#define WARNING 2
150
151/* Compression methods (see algorithm.doc) */
152#define DEFLATED 8
153
Erik Andersene49d5ec2000-02-08 19:58:47 +0000154extern int method; /* compression method */
Eric Andersenb052b471999-11-18 00:21:59 +0000155
156/* To save memory for 16 bit systems, some arrays are overlaid between
157 * the various modules:
158 * deflate: prev+head window d_buf l_buf outbuf
159 * unlzw: tab_prefix tab_suffix stack inbuf outbuf
160 * inflate: window inbuf
161 * unpack: window inbuf prefix_len
162 * unlzh: left+right window c_table inbuf c_len
163 * For compression, input is done in window[]. For decompression, output
164 * is done in window except for unlzw.
165 */
166
167#ifndef INBUFSIZ
168# ifdef SMALL_MEM
Erik Andersene49d5ec2000-02-08 19:58:47 +0000169# define INBUFSIZ 0x2000 /* input buffer size */
Eric Andersenb052b471999-11-18 00:21:59 +0000170# else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000171# define INBUFSIZ 0x8000 /* input buffer size */
Eric Andersenb052b471999-11-18 00:21:59 +0000172# endif
173#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +0000174#define INBUF_EXTRA 64 /* required by unlzw() */
Eric Andersenb052b471999-11-18 00:21:59 +0000175
176#ifndef OUTBUFSIZ
177# ifdef SMALL_MEM
Erik Andersene49d5ec2000-02-08 19:58:47 +0000178# define OUTBUFSIZ 8192 /* output buffer size */
Eric Andersenb052b471999-11-18 00:21:59 +0000179# else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000180# define OUTBUFSIZ 16384 /* output buffer size */
Eric Andersenb052b471999-11-18 00:21:59 +0000181# endif
182#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +0000183#define OUTBUF_EXTRA 2048 /* required by unlzw() */
Eric Andersenb052b471999-11-18 00:21:59 +0000184
185#define SMALL_MEM
186
187#ifndef DIST_BUFSIZE
188# ifdef SMALL_MEM
Erik Andersene49d5ec2000-02-08 19:58:47 +0000189# define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
Eric Andersenb052b471999-11-18 00:21:59 +0000190# else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000191# define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
Eric Andersenb052b471999-11-18 00:21:59 +0000192# endif
193#endif
194
195/*#define DYN_ALLOC*/
196
197#ifdef DYN_ALLOC
198# define EXTERN(type, array) extern type * array
199# define DECLARE(type, array, size) type * array
200# define ALLOC(type, array, size) { \
201 array = (type*)calloc((size_t)(((size)+1L)/2), 2*sizeof(type)); \
Matt Kraaibe84cd42000-07-12 17:02:35 +0000202 if (array == NULL) errorMsg(memory_exhausted); \
Eric Andersenb052b471999-11-18 00:21:59 +0000203 }
204# define FREE(array) {if (array != NULL) free(array), array=NULL;}
205#else
206# define EXTERN(type, array) extern type array[]
207# define DECLARE(type, array, size) type array[size]
208# define ALLOC(type, array, size)
209# define FREE(array)
210#endif
211
Erik Andersene49d5ec2000-02-08 19:58:47 +0000212EXTERN(uch, inbuf); /* input buffer */
213EXTERN(uch, outbuf); /* output buffer */
214EXTERN(ush, d_buf); /* buffer for distances, see trees.c */
215EXTERN(uch, window); /* Sliding window and suffix table (unlzw) */
Eric Andersenb052b471999-11-18 00:21:59 +0000216#define tab_suffix window
217#ifndef MAXSEG_64K
Erik Andersene49d5ec2000-02-08 19:58:47 +0000218# define tab_prefix prev /* hash link (see deflate.c) */
219# define head (prev+WSIZE) /* hash head (see deflate.c) */
220EXTERN(ush, tab_prefix); /* prefix code (see unlzw.c) */
Eric Andersenb052b471999-11-18 00:21:59 +0000221#else
222# define tab_prefix0 prev
223# define head tab_prefix1
Erik Andersene49d5ec2000-02-08 19:58:47 +0000224EXTERN(ush, tab_prefix0); /* prefix for even codes */
225EXTERN(ush, tab_prefix1); /* prefix for odd codes */
Eric Andersenb052b471999-11-18 00:21:59 +0000226#endif
227
Erik Andersene49d5ec2000-02-08 19:58:47 +0000228extern unsigned insize; /* valid bytes in inbuf */
229extern unsigned inptr; /* index of next byte to be processed in inbuf */
230extern unsigned outcnt; /* bytes in output buffer */
Eric Andersenb052b471999-11-18 00:21:59 +0000231
Erik Andersene49d5ec2000-02-08 19:58:47 +0000232extern long bytes_in; /* number of input bytes */
233extern long bytes_out; /* number of output bytes */
234extern long header_bytes; /* number of bytes in gzip header */
Eric Andersenb052b471999-11-18 00:21:59 +0000235
Erik Andersene49d5ec2000-02-08 19:58:47 +0000236extern long ifile_size; /* input file size, -1 for devices (debug only) */
Eric Andersenb052b471999-11-18 00:21:59 +0000237
Erik Andersene49d5ec2000-02-08 19:58:47 +0000238typedef int file_t; /* Do not use stdio */
239
240#define NO_FILE (-1) /* in memory compression */
Eric Andersenb052b471999-11-18 00:21:59 +0000241
242
Erik Andersene49d5ec2000-02-08 19:58:47 +0000243#define GZIP_MAGIC "\037\213" /* Magic header for gzip files, 1F 8B */
Eric Andersenb052b471999-11-18 00:21:59 +0000244
245/* gzip flag byte */
Erik Andersene49d5ec2000-02-08 19:58:47 +0000246#define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
247#define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
248#define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
249#define ORIG_NAME 0x08 /* bit 3 set: original file name present */
250#define COMMENT 0x10 /* bit 4 set: file comment present */
251#define ENCRYPTED 0x20 /* bit 5 set: file is encrypted */
252#define RESERVED 0xC0 /* bit 6,7: reserved */
Eric Andersenb052b471999-11-18 00:21:59 +0000253
254#ifndef WSIZE
Erik Andersene49d5ec2000-02-08 19:58:47 +0000255# define WSIZE 0x8000 /* window size--must be a power of two, and */
256#endif /* at least 32K for zip's deflate method */
Eric Andersenb052b471999-11-18 00:21:59 +0000257
258#define MIN_MATCH 3
259#define MAX_MATCH 258
260/* The minimum and maximum match lengths */
261
262#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
263/* Minimum amount of lookahead, except at the end of the input file.
264 * See deflate.c for comments about the MIN_MATCH+1.
265 */
266
267#define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
268/* In order to simplify the code, particularly on 16 bit machines, match
269 * distances are limited to MAX_DIST instead of WSIZE.
270 */
271
Erik Andersene49d5ec2000-02-08 19:58:47 +0000272extern int exit_code; /* program exit code */
273extern int verbose; /* be verbose (-v) */
274extern int level; /* compression level */
275extern int test; /* check .z file integrity */
276extern int save_orig_name; /* set if original name must be saved */
Eric Andersenb052b471999-11-18 00:21:59 +0000277
278#define get_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(0))
279#define try_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(1))
280
281/* put_byte is used for the compressed output, put_ubyte for the
282 * uncompressed output. However unlzw() uses window for its
283 * suffix table instead of its output buffer, so it does not use put_ubyte
284 * (to be cleaned up).
285 */
286#define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
287 flush_outbuf();}
288#define put_ubyte(c) {window[outcnt++]=(uch)(c); if (outcnt==WSIZE)\
289 flush_window();}
290
291/* Output a 16 bit value, lsb first */
292#define put_short(w) \
293{ if (outcnt < OUTBUFSIZ-2) { \
294 outbuf[outcnt++] = (uch) ((w) & 0xff); \
295 outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
296 } else { \
297 put_byte((uch)((w) & 0xff)); \
298 put_byte((uch)((ush)(w) >> 8)); \
299 } \
300}
301
302/* Output a 32 bit value to the bit stream, lsb first */
303#define put_long(n) { \
304 put_short((n) & 0xffff); \
305 put_short(((ulg)(n)) >> 16); \
306}
307
Erik Andersene49d5ec2000-02-08 19:58:47 +0000308#define seekable() 0 /* force sequential output */
309#define translate_eol 0 /* no option -a yet */
Eric Andersenb052b471999-11-18 00:21:59 +0000310
Erik Andersene49d5ec2000-02-08 19:58:47 +0000311#define tolow(c) (isupper(c) ? (c)-'A'+'a' : (c)) /* force to lower case */
Eric Andersenb052b471999-11-18 00:21:59 +0000312
313/* Macros for getting two-byte and four-byte header values */
314#define SH(p) ((ush)(uch)((p)[0]) | ((ush)(uch)((p)[1]) << 8))
315#define LG(p) ((ulg)(SH(p)) | ((ulg)(SH((p)+2)) << 16))
316
317/* Diagnostic functions */
318#ifdef DEBUG
Erik Andersen9ffdaa62000-02-11 21:55:04 +0000319# define Assert(cond,msg) {if(!(cond)) errorMsg(msg);}
Eric Andersenb052b471999-11-18 00:21:59 +0000320# define Trace(x) fprintf x
321# define Tracev(x) {if (verbose) fprintf x ;}
322# define Tracevv(x) {if (verbose>1) fprintf x ;}
323# define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
324# define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
325#else
326# define Assert(cond,msg)
327# define Trace(x)
328# define Tracev(x)
329# define Tracevv(x)
330# define Tracec(c,x)
331# define Tracecv(c,x)
332#endif
333
334#define WARN(msg) {fprintf msg ; \
335 if (exit_code == OK) exit_code = WARNING;}
336
337 /* in unzip.c */
Erik Andersen61677fe2000-04-13 01:18:56 +0000338extern int unzip (int in, int out);
Eric Andersenb052b471999-11-18 00:21:59 +0000339
340 /* in gzip.c */
Erik Andersen61677fe2000-04-13 01:18:56 +0000341RETSIGTYPE abort_gzip (void);
Eric Andersenb052b471999-11-18 00:21:59 +0000342
Erik Andersene49d5ec2000-02-08 19:58:47 +0000343 /* in deflate.c */
Erik Andersen61677fe2000-04-13 01:18:56 +0000344void lm_init (int pack_level, ush * flags);
345ulg deflate (void);
Eric Andersenb052b471999-11-18 00:21:59 +0000346
Erik Andersene49d5ec2000-02-08 19:58:47 +0000347 /* in trees.c */
Erik Andersen61677fe2000-04-13 01:18:56 +0000348void ct_init (ush * attr, int *method);
349int ct_tally (int dist, int lc);
350ulg flush_block (char *buf, ulg stored_len, int eof);
Eric Andersenb052b471999-11-18 00:21:59 +0000351
Erik Andersene49d5ec2000-02-08 19:58:47 +0000352 /* in bits.c */
Erik Andersen61677fe2000-04-13 01:18:56 +0000353void bi_init (file_t zipfile);
354void send_bits (int value, int length);
355unsigned bi_reverse (unsigned value, int length);
356void bi_windup (void);
357void copy_block (char *buf, unsigned len, int header);
Eric Andersenb052b471999-11-18 00:21:59 +0000358
359 /* in util.c: */
Erik Andersen61677fe2000-04-13 01:18:56 +0000360extern ulg updcrc (uch * s, unsigned n);
361extern void clear_bufs (void);
Erik Andersen330fd2b2000-05-19 05:35:19 +0000362static int fill_inbuf (int eof_ok);
Erik Andersen61677fe2000-04-13 01:18:56 +0000363extern void flush_outbuf (void);
Erik Andersen330fd2b2000-05-19 05:35:19 +0000364static void flush_window (void);
Erik Andersen61677fe2000-04-13 01:18:56 +0000365extern void write_buf (int fd, void * buf, unsigned cnt);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000366
Erik Andersen330fd2b2000-05-19 05:35:19 +0000367void read_error_msg (void);
368void write_error_msg (void);
Eric Andersenb052b471999-11-18 00:21:59 +0000369
370 /* in inflate.c */
Erik Andersen330fd2b2000-05-19 05:35:19 +0000371static int inflate (void);
Eric Andersenb052b471999-11-18 00:21:59 +0000372
373/* #include "lzw.h" */
374
375/* lzw.h -- define the lzw functions.
376 * Copyright (C) 1992-1993 Jean-loup Gailly.
377 * This is free software; you can redistribute it and/or modify it under the
378 * terms of the GNU General Public License, see the file COPYING.
379 */
380
381#if !defined(OF) && defined(lint)
382# include "gzip.h"
383#endif
384
385#ifndef BITS
386# define BITS 16
387#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +0000388#define INIT_BITS 9 /* Initial number of bits per code */
Eric Andersenb052b471999-11-18 00:21:59 +0000389
Erik Andersene49d5ec2000-02-08 19:58:47 +0000390#define LZW_MAGIC "\037\235" /* Magic header for lzw files, 1F 9D */
Eric Andersenb052b471999-11-18 00:21:59 +0000391
Erik Andersene49d5ec2000-02-08 19:58:47 +0000392#define BIT_MASK 0x1f /* Mask for 'number of compression bits' */
Eric Andersenb052b471999-11-18 00:21:59 +0000393/* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
394 * It's a pity that old uncompress does not check bit 0x20. That makes
395 * extension of the format actually undesirable because old compress
396 * would just crash on the new format instead of giving a meaningful
397 * error message. It does check the number of bits, but it's more
398 * helpful to say "unsupported format, get a new version" than
399 * "can only handle 16 bits".
400 */
401
402#define BLOCK_MODE 0x80
403/* Block compression: if table is full and compression rate is dropping,
404 * clear the dictionary.
405 */
406
Erik Andersene49d5ec2000-02-08 19:58:47 +0000407#define LZW_RESERVED 0x60 /* reserved bits */
Eric Andersenb052b471999-11-18 00:21:59 +0000408
Erik Andersene49d5ec2000-02-08 19:58:47 +0000409#define CLEAR 256 /* flush the dictionary */
410#define FIRST (CLEAR+1) /* first free entry */
Eric Andersenb052b471999-11-18 00:21:59 +0000411
Erik Andersene49d5ec2000-02-08 19:58:47 +0000412extern int maxbits; /* max bits per code for LZW */
413extern int block_mode; /* block compress mode -C compatible with 2.0 */
Eric Andersenb052b471999-11-18 00:21:59 +0000414
Erik Andersen61677fe2000-04-13 01:18:56 +0000415extern int lzw (int in, int out);
416extern int unlzw (int in, int out);
Eric Andersenb052b471999-11-18 00:21:59 +0000417
418
419/* #include "revision.h" */
420
421/* revision.h -- define the version number
422 * Copyright (C) 1992-1993 Jean-loup Gailly.
423 * This is free software; you can redistribute it and/or modify it under the
424 * terms of the GNU General Public License, see the file COPYING.
425 */
426
427#define VERSION "1.2.4"
428#define PATCHLEVEL 0
429#define REVDATE "18 Aug 93"
430
431/* This version does not support compression into old compress format: */
432#ifdef LZW
433# undef LZW
434#endif
435
Eric Andersenb052b471999-11-18 00:21:59 +0000436#include <time.h>
437#include <fcntl.h>
438#include <unistd.h>
Eric Andersenb052b471999-11-18 00:21:59 +0000439#include <stdlib.h>
Eric Andersenb052b471999-11-18 00:21:59 +0000440#if defined(DIRENT)
441# include <dirent.h>
Erik Andersene49d5ec2000-02-08 19:58:47 +0000442typedef struct dirent dir_type;
443
Eric Andersenb052b471999-11-18 00:21:59 +0000444# define NLENGTH(dirent) ((int)strlen((dirent)->d_name))
445# define DIR_OPT "DIRENT"
446#else
447# define NLENGTH(dirent) ((dirent)->d_namlen)
448# ifdef SYSDIR
449# include <sys/dir.h>
Erik Andersene49d5ec2000-02-08 19:58:47 +0000450typedef struct direct dir_type;
451
Eric Andersenb052b471999-11-18 00:21:59 +0000452# define DIR_OPT "SYSDIR"
453# else
454# ifdef SYSNDIR
455# include <sys/ndir.h>
Erik Andersene49d5ec2000-02-08 19:58:47 +0000456typedef struct direct dir_type;
457
Eric Andersenb052b471999-11-18 00:21:59 +0000458# define DIR_OPT "SYSNDIR"
459# else
460# ifdef NDIR
461# include <ndir.h>
Erik Andersene49d5ec2000-02-08 19:58:47 +0000462typedef struct direct dir_type;
463
Eric Andersenb052b471999-11-18 00:21:59 +0000464# define DIR_OPT "NDIR"
465# else
466# define NO_DIR
467# define DIR_OPT "NO_DIR"
468# endif
469# endif
470# endif
471#endif
Eric Andersenb052b471999-11-18 00:21:59 +0000472#if !defined(S_ISDIR) && defined(S_IFDIR)
473# define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR)
474#endif
475#if !defined(S_ISREG) && defined(S_IFREG)
476# define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
477#endif
Erik Andersen61677fe2000-04-13 01:18:56 +0000478typedef RETSIGTYPE(*sig_type) (int);
Eric Andersenb052b471999-11-18 00:21:59 +0000479
480#ifndef O_BINARY
Erik Andersene49d5ec2000-02-08 19:58:47 +0000481# define O_BINARY 0 /* creation mode for open() */
Eric Andersenb052b471999-11-18 00:21:59 +0000482#endif
483
484#ifndef O_CREAT
485 /* Pure BSD system? */
486# include <sys/file.h>
487# ifndef O_CREAT
488# define O_CREAT FCREAT
489# endif
490# ifndef O_EXCL
491# define O_EXCL FEXCL
492# endif
493#endif
494
495#ifndef S_IRUSR
496# define S_IRUSR 0400
497#endif
498#ifndef S_IWUSR
499# define S_IWUSR 0200
500#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +0000501#define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
Eric Andersenb052b471999-11-18 00:21:59 +0000502
Erik Andersene49d5ec2000-02-08 19:58:47 +0000503#ifndef MAX_PATH_LEN /* max pathname length */
Erik Andersen4f3f7572000-04-28 00:18:56 +0000504# ifdef BUFSIZ
505# define MAX_PATH_LEN BUFSIZ
Erik Andersenfac10d72000-02-07 05:29:42 +0000506# else
507# define MAX_PATH_LEN 1024
508# endif
Eric Andersenb052b471999-11-18 00:21:59 +0000509#endif
510
511#ifndef SEEK_END
512# define SEEK_END 2
513#endif
514
515#ifdef NO_OFF_T
Erik Andersene49d5ec2000-02-08 19:58:47 +0000516typedef long off_t;
Erik Andersen61677fe2000-04-13 01:18:56 +0000517off_t lseek (int fd, off_t offset, int whence);
Eric Andersenb052b471999-11-18 00:21:59 +0000518#endif
519
520
521 /* global buffers */
522
Erik Andersene49d5ec2000-02-08 19:58:47 +0000523DECLARE(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
524DECLARE(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
525DECLARE(ush, d_buf, DIST_BUFSIZE);
526DECLARE(uch, window, 2L * WSIZE);
Eric Andersenb052b471999-11-18 00:21:59 +0000527#ifndef MAXSEG_64K
Erik Andersene49d5ec2000-02-08 19:58:47 +0000528DECLARE(ush, tab_prefix, 1L << BITS);
Eric Andersenb052b471999-11-18 00:21:59 +0000529#else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000530DECLARE(ush, tab_prefix0, 1L << (BITS - 1));
531DECLARE(ush, tab_prefix1, 1L << (BITS - 1));
Eric Andersenb052b471999-11-18 00:21:59 +0000532#endif
533
534 /* local variables */
535
Erik Andersene49d5ec2000-02-08 19:58:47 +0000536int test_mode = 0; /* check file integrity option */
537int foreground; /* set if program run in foreground */
538int maxbits = BITS; /* max bits per code for LZW */
539int method = DEFLATED; /* compression method */
540int exit_code = OK; /* program exit code */
541int last_member; /* set for .zip and .Z files */
542int part_nb; /* number of parts in .gz file */
543long ifile_size; /* input file size, -1 for devices (debug only) */
Eric Andersenb052b471999-11-18 00:21:59 +0000544
Erik Andersene49d5ec2000-02-08 19:58:47 +0000545long bytes_in; /* number of input bytes */
546long bytes_out; /* number of output bytes */
547long total_in = 0; /* input bytes for all files */
548long total_out = 0; /* output bytes for all files */
549struct stat istat; /* status for input file */
550int ifd; /* input file descriptor */
551int ofd; /* output file descriptor */
552unsigned insize; /* valid bytes in inbuf */
553unsigned inptr; /* index of next byte to be processed in inbuf */
554unsigned outcnt; /* bytes in output buffer */
Eric Andersenb052b471999-11-18 00:21:59 +0000555
Erik Andersene49d5ec2000-02-08 19:58:47 +0000556long header_bytes; /* number of bytes in gzip header */
Eric Andersenb052b471999-11-18 00:21:59 +0000557
558/* local functions */
559
Erik Andersen61677fe2000-04-13 01:18:56 +0000560local int get_method (int in);
Eric Andersenb052b471999-11-18 00:21:59 +0000561
562#define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
563
564/* ======================================================================== */
Erik Andersene49d5ec2000-02-08 19:58:47 +0000565int gunzip_main(int argc, char **argv)
Eric Andersenb052b471999-11-18 00:21:59 +0000566{
Erik Andersene49d5ec2000-02-08 19:58:47 +0000567 int file_count; /* number of files to precess */
Eric Andersenea824fb2000-07-21 22:17:39 +0000568 int tostdout = 0;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000569 int fromstdin = 0;
570 int result;
571 int inFileNum;
572 int outFileNum;
573 int delInputFile = 0;
Eric Andersenea824fb2000-07-21 22:17:39 +0000574 int force = 0;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000575 struct stat statBuf;
576 char *delFileName;
577 char ifname[MAX_PATH_LEN + 1]; /* input file name */
578 char ofname[MAX_PATH_LEN + 1]; /* output file name */
Eric Andersenb052b471999-11-18 00:21:59 +0000579
Matt Kraaie58771e2000-07-12 15:38:49 +0000580 if (strcmp(applet_name, "zcat") == 0) {
Eric Andersene99674a2000-09-01 00:41:10 +0000581 force = 1;
Eric Andersen73bdf132000-09-01 16:25:33 +0000582 tostdout = 1;
Erik Andersen59b9e872000-05-10 05:05:45 +0000583 }
Erik Andersene49d5ec2000-02-08 19:58:47 +0000584
585 /* Parse any options */
586 while (--argc > 0 && **(++argv) == '-') {
587 if (*((*argv) + 1) == '\0') {
Eric Andersenea824fb2000-07-21 22:17:39 +0000588 tostdout = 1;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000589 }
590 while (*(++(*argv))) {
591 switch (**argv) {
592 case 'c':
Eric Andersenea824fb2000-07-21 22:17:39 +0000593 tostdout = 1;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000594 break;
595 case 't':
596 test_mode = 1;
597 break;
Eric Andersenea824fb2000-07-21 22:17:39 +0000598 case 'f':
599 force = 1;
600 break;
Erik Andersene49d5ec2000-02-08 19:58:47 +0000601 default:
602 usage(gunzip_usage);
603 }
604 }
605 }
Eric Andersene99674a2000-09-01 00:41:10 +0000606
607 if (argc <= 0) {
608 tostdout = 1;
Eric Andersenea824fb2000-07-21 22:17:39 +0000609 fromstdin = 1;
Eric Andersene99674a2000-09-01 00:41:10 +0000610 }
Eric Andersenea824fb2000-07-21 22:17:39 +0000611
612 if (isatty(fileno(stdin)) && fromstdin==1 && force==0)
613 fatalError( "data not read from terminal. Use -f to force it.\n");
614 if (isatty(fileno(stdout)) && tostdout==1 && force==0)
615 fatalError( "data not written to terminal. Use -f to force it.\n");
616
Erik Andersene49d5ec2000-02-08 19:58:47 +0000617
618 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
619 if (foreground) {
620 (void) signal(SIGINT, (sig_type) abort_gzip);
621 }
Eric Andersenb052b471999-11-18 00:21:59 +0000622#ifdef SIGTERM
Erik Andersene49d5ec2000-02-08 19:58:47 +0000623 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
624 (void) signal(SIGTERM, (sig_type) abort_gzip);
625 }
Eric Andersenb052b471999-11-18 00:21:59 +0000626#endif
627#ifdef SIGHUP
Erik Andersene49d5ec2000-02-08 19:58:47 +0000628 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
629 (void) signal(SIGHUP, (sig_type) abort_gzip);
630 }
Eric Andersenb052b471999-11-18 00:21:59 +0000631#endif
632
Erik Andersene49d5ec2000-02-08 19:58:47 +0000633 file_count = argc - optind;
Eric Andersenb052b471999-11-18 00:21:59 +0000634
Erik Andersene49d5ec2000-02-08 19:58:47 +0000635 /* Allocate all global buffers (for DYN_ALLOC option) */
636 ALLOC(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
637 ALLOC(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
638 ALLOC(ush, d_buf, DIST_BUFSIZE);
639 ALLOC(uch, window, 2L * WSIZE);
Eric Andersenb052b471999-11-18 00:21:59 +0000640#ifndef MAXSEG_64K
Erik Andersene49d5ec2000-02-08 19:58:47 +0000641 ALLOC(ush, tab_prefix, 1L << BITS);
Eric Andersenb052b471999-11-18 00:21:59 +0000642#else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000643 ALLOC(ush, tab_prefix0, 1L << (BITS - 1));
644 ALLOC(ush, tab_prefix1, 1L << (BITS - 1));
Eric Andersenb052b471999-11-18 00:21:59 +0000645#endif
646
Erik Andersene49d5ec2000-02-08 19:58:47 +0000647 if (fromstdin == 1) {
648 strcpy(ofname, "stdin");
Eric Andersenb052b471999-11-18 00:21:59 +0000649
Erik Andersene49d5ec2000-02-08 19:58:47 +0000650 inFileNum = fileno(stdin);
651 ifile_size = -1L; /* convention for unknown size */
Eric Andersenb052b471999-11-18 00:21:59 +0000652 } else {
Erik Andersene49d5ec2000-02-08 19:58:47 +0000653 /* Open up the input file */
Eric Andersenea824fb2000-07-21 22:17:39 +0000654 if (argc <= 0)
Erik Andersene49d5ec2000-02-08 19:58:47 +0000655 usage(gunzip_usage);
656 if (strlen(*argv) > MAX_PATH_LEN) {
Matt Kraaid537a952000-07-14 01:51:25 +0000657 errorMsg(name_too_long);
Erik Andersen61677fe2000-04-13 01:18:56 +0000658 exit(WARNING);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000659 }
660 strcpy(ifname, *argv);
661
662 /* Open input fille */
663 inFileNum = open(ifname, O_RDONLY);
664 if (inFileNum < 0) {
665 perror(ifname);
Erik Andersen61677fe2000-04-13 01:18:56 +0000666 exit(WARNING);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000667 }
668 /* Get the time stamp on the input file. */
669 result = stat(ifname, &statBuf);
670 if (result < 0) {
671 perror(ifname);
Erik Andersen61677fe2000-04-13 01:18:56 +0000672 exit(WARNING);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000673 }
674 ifile_size = statBuf.st_size;
Eric Andersenb052b471999-11-18 00:21:59 +0000675 }
676
Eric Andersenea824fb2000-07-21 22:17:39 +0000677 if (tostdout == 1) {
Erik Andersene49d5ec2000-02-08 19:58:47 +0000678 /* And get to work */
679 strcpy(ofname, "stdout");
680 outFileNum = fileno(stdout);
681
682 clear_bufs(); /* clear input and output buffers */
683 part_nb = 0;
684
685 /* Actually do the compression/decompression. */
686 unzip(inFileNum, outFileNum);
687
688 } else if (test_mode) {
689 /* Actually do the compression/decompression. */
690 unzip(inFileNum, 2);
691 } else {
692 char *pos;
693
694 /* And get to work */
695 if (strlen(ifname) > MAX_PATH_LEN - 4) {
Matt Kraaid537a952000-07-14 01:51:25 +0000696 errorMsg(name_too_long);
Erik Andersen61677fe2000-04-13 01:18:56 +0000697 exit(WARNING);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000698 }
699 strcpy(ofname, ifname);
700 pos = strstr(ofname, ".gz");
701 if (pos != NULL) {
702 *pos = '\0';
703 delInputFile = 1;
704 } else {
705 pos = strstr(ofname, ".tgz");
706 if (pos != NULL) {
707 *pos = '\0';
708 strcat(pos, ".tar");
709 delInputFile = 1;
710 }
711 }
712
713 /* Open output fille */
Erik Andersen4d1d0111999-12-17 18:44:15 +0000714#if (__GLIBC__ >= 2) && (__GLIBC_MINOR__ >= 1)
Erik Andersene49d5ec2000-02-08 19:58:47 +0000715 outFileNum = open(ofname, O_RDWR | O_CREAT | O_EXCL | O_NOFOLLOW);
Erik Andersen4d1d0111999-12-17 18:44:15 +0000716#else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000717 outFileNum = open(ofname, O_RDWR | O_CREAT | O_EXCL);
Erik Andersen4d1d0111999-12-17 18:44:15 +0000718#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +0000719 if (outFileNum < 0) {
720 perror(ofname);
Erik Andersen61677fe2000-04-13 01:18:56 +0000721 exit(WARNING);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000722 }
723 /* Set permissions on the file */
724 fchmod(outFileNum, statBuf.st_mode);
725
726 clear_bufs(); /* clear input and output buffers */
727 part_nb = 0;
728
729 /* Actually do the compression/decompression. */
730 result = unzip(inFileNum, outFileNum);
731
732 close(outFileNum);
733 close(inFileNum);
734 /* Delete the original file */
735 if (result == OK)
736 delFileName = ifname;
737 else
738 delFileName = ofname;
739
740 if (delInputFile == 1 && unlink(delFileName) < 0) {
741 perror(delFileName);
742 exit(FALSE);
743 }
Eric Andersenb052b471999-11-18 00:21:59 +0000744 }
Eric Andersenb6106152000-06-19 17:25:40 +0000745 return(exit_code);
Eric Andersenb052b471999-11-18 00:21:59 +0000746}
747
748
749/* ========================================================================
750 * Check the magic number of the input file and update ofname if an
Eric Andersenea824fb2000-07-21 22:17:39 +0000751 * original name was given and tostdout is not set.
Eric Andersenb052b471999-11-18 00:21:59 +0000752 * Return the compression method, -1 for error, -2 for warning.
753 * Set inptr to the offset of the next byte to be processed.
754 * Updates time_stamp if there is one and --no-time is not used.
755 * This function may be called repeatedly for an input file consisting
756 * of several contiguous gzip'ed members.
757 * IN assertions: there is at least one remaining compressed member.
758 * If the member is a zip file, it must be the only one.
759 */
760local int get_method(in)
Erik Andersene49d5ec2000-02-08 19:58:47 +0000761int in; /* input file descriptor */
Eric Andersenb052b471999-11-18 00:21:59 +0000762{
Erik Andersene49d5ec2000-02-08 19:58:47 +0000763 uch flags; /* compression flags */
764 char magic[2]; /* magic header */
Eric Andersenb052b471999-11-18 00:21:59 +0000765
Erik Andersene49d5ec2000-02-08 19:58:47 +0000766 magic[0] = (char) get_byte();
767 magic[1] = (char) get_byte();
768 method = -1; /* unknown yet */
769 part_nb++; /* number of parts in gzip file */
770 header_bytes = 0;
771 last_member = RECORD_IO;
772 /* assume multiple members in gzip file except for record oriented I/O */
Eric Andersenb052b471999-11-18 00:21:59 +0000773
Erik Andersene49d5ec2000-02-08 19:58:47 +0000774 if (memcmp(magic, GZIP_MAGIC, 2) == 0) {
Eric Andersenb052b471999-11-18 00:21:59 +0000775
Erik Andersene49d5ec2000-02-08 19:58:47 +0000776 method = (int) get_byte();
777 if (method != DEFLATED) {
Matt Kraaid537a952000-07-14 01:51:25 +0000778 errorMsg("unknown method %d -- get newer version of gzip\n",
Erik Andersene49d5ec2000-02-08 19:58:47 +0000779 method);
780 exit_code = ERROR;
781 return -1;
782 }
783 flags = (uch) get_byte();
Eric Andersenb052b471999-11-18 00:21:59 +0000784
Erik Andersene49d5ec2000-02-08 19:58:47 +0000785 (ulg) get_byte(); /* Ignore time stamp */
786 (ulg) get_byte();
787 (ulg) get_byte();
788 (ulg) get_byte();
Eric Andersenb052b471999-11-18 00:21:59 +0000789
Erik Andersene49d5ec2000-02-08 19:58:47 +0000790 (void) get_byte(); /* Ignore extra flags for the moment */
791 (void) get_byte(); /* Ignore OS type for the moment */
Eric Andersenb052b471999-11-18 00:21:59 +0000792
Erik Andersene49d5ec2000-02-08 19:58:47 +0000793 if ((flags & EXTRA_FIELD) != 0) {
794 unsigned len = (unsigned) get_byte();
Eric Andersenb052b471999-11-18 00:21:59 +0000795
Erik Andersene49d5ec2000-02-08 19:58:47 +0000796 len |= ((unsigned) get_byte()) << 8;
797
798 while (len--)
799 (void) get_byte();
800 }
801
802 /* Discard original name if any */
803 if ((flags & ORIG_NAME) != 0) {
804 while (get_char() != 0) /* null */
805 ;
806 }
807
808 /* Discard file comment if any */
809 if ((flags & COMMENT) != 0) {
810 while (get_char() != 0) /* null */
811 ;
812 }
813 if (part_nb == 1) {
814 header_bytes = inptr + 2 * sizeof(long); /* include crc and size */
815 }
816
Eric Andersenb052b471999-11-18 00:21:59 +0000817 }
818
Erik Andersene49d5ec2000-02-08 19:58:47 +0000819 if (method >= 0)
820 return method;
Eric Andersenb052b471999-11-18 00:21:59 +0000821
Eric Andersenb052b471999-11-18 00:21:59 +0000822 if (part_nb == 1) {
Erik Andersene49d5ec2000-02-08 19:58:47 +0000823 fprintf(stderr, "\nnot in gzip format\n");
824 exit_code = ERROR;
825 return -1;
826 } else {
827 WARN((stderr, "\ndecompression OK, trailing garbage ignored\n"));
828 return -2;
Eric Andersenb052b471999-11-18 00:21:59 +0000829 }
Eric Andersenb052b471999-11-18 00:21:59 +0000830}
831
Eric Andersenb052b471999-11-18 00:21:59 +0000832/* ========================================================================
833 * Signal and error handler.
834 */
835RETSIGTYPE abort_gzip()
836{
Erik Andersen61677fe2000-04-13 01:18:56 +0000837 exit(ERROR);
Eric Andersenb052b471999-11-18 00:21:59 +0000838}
Erik Andersene49d5ec2000-02-08 19:58:47 +0000839
Eric Andersenb052b471999-11-18 00:21:59 +0000840/* unzip.c -- decompress files in gzip or pkzip format.
841 * Copyright (C) 1992-1993 Jean-loup Gailly
842 * This is free software; you can redistribute it and/or modify it under the
843 * terms of the GNU General Public License, see the file COPYING.
844 *
845 * The code in this file is derived from the file funzip.c written
846 * and put in the public domain by Mark Adler.
847 */
848
849/*
850 This version can extract files in gzip or pkzip format.
851 For the latter, only the first entry is extracted, and it has to be
852 either deflated or stored.
853 */
854
855/* #include "crypt.h" */
856
857/* crypt.h (dummy version) -- do not perform encryption
858 * Hardly worth copyrighting :-)
859 */
860
861#ifdef CRYPT
Erik Andersene49d5ec2000-02-08 19:58:47 +0000862# undef CRYPT /* dummy version */
Eric Andersenb052b471999-11-18 00:21:59 +0000863#endif
864
Erik Andersene49d5ec2000-02-08 19:58:47 +0000865#define RAND_HEAD_LEN 12 /* length of encryption random header */
Eric Andersenb052b471999-11-18 00:21:59 +0000866
867#define zencode
868#define zdecode
869
870/* PKZIP header definitions */
Erik Andersene49d5ec2000-02-08 19:58:47 +0000871#define LOCSIG 0x04034b50L /* four-byte lead-in (lsb first) */
872#define LOCFLG 6 /* offset of bit flag */
873#define CRPFLG 1 /* bit for encrypted entry */
874#define EXTFLG 8 /* bit for extended local header */
875#define LOCHOW 8 /* offset of compression method */
876#define LOCTIM 10 /* file mod time (for decryption) */
877#define LOCCRC 14 /* offset of crc */
878#define LOCSIZ 18 /* offset of compressed size */
879#define LOCLEN 22 /* offset of uncompressed length */
880#define LOCFIL 26 /* offset of file name field length */
881#define LOCEXT 28 /* offset of extra field length */
882#define LOCHDR 30 /* size of local header, including sig */
883#define EXTHDR 16 /* size of extended local header, inc sig */
Eric Andersenb052b471999-11-18 00:21:59 +0000884
885
886/* Globals */
887
Erik Andersene49d5ec2000-02-08 19:58:47 +0000888char *key; /* not used--needed to link crypt.c */
889int pkzip = 0; /* set for a pkzip file */
890int ext_header = 0; /* set if extended local header */
Eric Andersenb052b471999-11-18 00:21:59 +0000891
892/* ===========================================================================
893 * Unzip in to out. This routine works on both gzip and pkzip files.
894 *
895 * IN assertions: the buffer inbuf contains already the beginning of
896 * the compressed data, from offsets inptr to insize-1 included.
897 * The magic header has already been checked. The output buffer is cleared.
898 */
899int unzip(in, out)
Erik Andersene49d5ec2000-02-08 19:58:47 +0000900int in, out; /* input and output file descriptors */
Eric Andersenb052b471999-11-18 00:21:59 +0000901{
Erik Andersene49d5ec2000-02-08 19:58:47 +0000902 ulg orig_crc = 0; /* original crc */
903 ulg orig_len = 0; /* original uncompressed length */
904 int n;
905 uch buf[EXTHDR]; /* extended local header */
Eric Andersenb052b471999-11-18 00:21:59 +0000906
Erik Andersene49d5ec2000-02-08 19:58:47 +0000907 ifd = in;
908 ofd = out;
909 method = get_method(ifd);
910 if (method < 0) {
Erik Andersen61677fe2000-04-13 01:18:56 +0000911 exit(exit_code); /* error message already emitted */
Eric Andersenb052b471999-11-18 00:21:59 +0000912 }
913
Erik Andersene49d5ec2000-02-08 19:58:47 +0000914 updcrc(NULL, 0); /* initialize crc */
Eric Andersenb052b471999-11-18 00:21:59 +0000915
Erik Andersene49d5ec2000-02-08 19:58:47 +0000916 if (pkzip && !ext_header) { /* crc and length at the end otherwise */
917 orig_crc = LG(inbuf + LOCCRC);
918 orig_len = LG(inbuf + LOCLEN);
Eric Andersenb052b471999-11-18 00:21:59 +0000919 }
Eric Andersenb052b471999-11-18 00:21:59 +0000920
Erik Andersene49d5ec2000-02-08 19:58:47 +0000921 /* Decompress */
922 if (method == DEFLATED) {
923
924 int res = inflate();
925
926 if (res == 3) {
Matt Kraaibe84cd42000-07-12 17:02:35 +0000927 errorMsg(memory_exhausted);
Erik Andersene49d5ec2000-02-08 19:58:47 +0000928 } else if (res != 0) {
Erik Andersen9ffdaa62000-02-11 21:55:04 +0000929 errorMsg("invalid compressed data--format violated");
Erik Andersene49d5ec2000-02-08 19:58:47 +0000930 }
931
932 } else {
Erik Andersen9ffdaa62000-02-11 21:55:04 +0000933 errorMsg("internal error, invalid method");
Eric Andersenb052b471999-11-18 00:21:59 +0000934 }
Eric Andersenb052b471999-11-18 00:21:59 +0000935
Erik Andersene49d5ec2000-02-08 19:58:47 +0000936 /* Get the crc and original length */
937 if (!pkzip) {
938 /* crc32 (see algorithm.doc)
939 * uncompressed input size modulo 2^32
940 */
941 for (n = 0; n < 8; n++) {
942 buf[n] = (uch) get_byte(); /* may cause an error if EOF */
943 }
944 orig_crc = LG(buf);
945 orig_len = LG(buf + 4);
Eric Andersenb052b471999-11-18 00:21:59 +0000946
Erik Andersene49d5ec2000-02-08 19:58:47 +0000947 } else if (ext_header) { /* If extended header, check it */
948 /* signature - 4bytes: 0x50 0x4b 0x07 0x08
949 * CRC-32 value
950 * compressed size 4-bytes
951 * uncompressed size 4-bytes
952 */
953 for (n = 0; n < EXTHDR; n++) {
954 buf[n] = (uch) get_byte(); /* may cause an error if EOF */
955 }
956 orig_crc = LG(buf + 4);
957 orig_len = LG(buf + 12);
958 }
959
960 /* Validate decompression */
961 if (orig_crc != updcrc(outbuf, 0)) {
Erik Andersen9ffdaa62000-02-11 21:55:04 +0000962 errorMsg("invalid compressed data--crc error");
Erik Andersene49d5ec2000-02-08 19:58:47 +0000963 }
964 if (orig_len != (ulg) bytes_out) {
Erik Andersen9ffdaa62000-02-11 21:55:04 +0000965 errorMsg("invalid compressed data--length error");
Erik Andersene49d5ec2000-02-08 19:58:47 +0000966 }
967
968 /* Check if there are more entries in a pkzip file */
969 if (pkzip && inptr + 4 < insize && LG(inbuf + inptr) == LOCSIG) {
970 WARN((stderr, "has more than one entry--rest ignored\n"));
971 }
972 ext_header = pkzip = 0; /* for next file */
973 return OK;
Eric Andersenb052b471999-11-18 00:21:59 +0000974}
Erik Andersene49d5ec2000-02-08 19:58:47 +0000975
Eric Andersenb052b471999-11-18 00:21:59 +0000976/* util.c -- utility functions for gzip support
977 * Copyright (C) 1992-1993 Jean-loup Gailly
978 * This is free software; you can redistribute it and/or modify it under the
979 * terms of the GNU General Public License, see the file COPYING.
980 */
981
982#include <ctype.h>
983#include <errno.h>
984#include <sys/types.h>
985
986#ifdef HAVE_UNISTD_H
987# include <unistd.h>
988#endif
989#ifndef NO_FCNTL_H
990# include <fcntl.h>
991#endif
992
993#if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
994# include <stdlib.h>
995#else
Erik Andersene49d5ec2000-02-08 19:58:47 +0000996extern int errno;
Eric Andersenb052b471999-11-18 00:21:59 +0000997#endif
998
Erik Andersene49d5ec2000-02-08 19:58:47 +0000999static const ulg crc_32_tab[]; /* crc table, defined below */
Eric Andersenb052b471999-11-18 00:21:59 +00001000
1001/* ===========================================================================
1002 * Run a set of bytes through the crc shift register. If s is a NULL
1003 * pointer, then initialize the crc shift register contents instead.
1004 * Return the current crc in either case.
1005 */
1006ulg updcrc(s, n)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001007uch *s; /* pointer to bytes to pump through */
1008unsigned n; /* number of bytes in s[] */
Eric Andersenb052b471999-11-18 00:21:59 +00001009{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001010 register ulg c; /* temporary variable */
Eric Andersenb052b471999-11-18 00:21:59 +00001011
Erik Andersene49d5ec2000-02-08 19:58:47 +00001012 static ulg crc = (ulg) 0xffffffffL; /* shift register contents */
Eric Andersenb052b471999-11-18 00:21:59 +00001013
Erik Andersene49d5ec2000-02-08 19:58:47 +00001014 if (s == NULL) {
1015 c = 0xffffffffL;
1016 } else {
1017 c = crc;
1018 if (n)
1019 do {
1020 c = crc_32_tab[((int) c ^ (*s++)) & 0xff] ^ (c >> 8);
1021 } while (--n);
1022 }
1023 crc = c;
1024 return c ^ 0xffffffffL; /* (instead of ~c for 64-bit machines) */
Eric Andersenb052b471999-11-18 00:21:59 +00001025}
1026
1027/* ===========================================================================
1028 * Clear input and output buffers
1029 */
Erik Andersen330fd2b2000-05-19 05:35:19 +00001030void clear_bufs(void)
Eric Andersenb052b471999-11-18 00:21:59 +00001031{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001032 outcnt = 0;
1033 insize = inptr = 0;
1034 bytes_in = bytes_out = 0L;
Eric Andersenb052b471999-11-18 00:21:59 +00001035}
1036
1037/* ===========================================================================
1038 * Fill the input buffer. This is called only when the buffer is empty.
1039 */
1040int fill_inbuf(eof_ok)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001041int eof_ok; /* set if EOF acceptable as a result */
Eric Andersenb052b471999-11-18 00:21:59 +00001042{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001043 int len;
Eric Andersenb052b471999-11-18 00:21:59 +00001044
Erik Andersene49d5ec2000-02-08 19:58:47 +00001045 /* Read as much as possible */
1046 insize = 0;
1047 errno = 0;
1048 do {
1049 len = read(ifd, (char *) inbuf + insize, INBUFSIZ - insize);
1050 if (len == 0 || len == EOF)
1051 break;
1052 insize += len;
1053 } while (insize < INBUFSIZ);
Eric Andersenb052b471999-11-18 00:21:59 +00001054
Erik Andersene49d5ec2000-02-08 19:58:47 +00001055 if (insize == 0) {
1056 if (eof_ok)
1057 return EOF;
Erik Andersen330fd2b2000-05-19 05:35:19 +00001058 read_error_msg();
Erik Andersene49d5ec2000-02-08 19:58:47 +00001059 }
1060 bytes_in += (ulg) insize;
1061 inptr = 1;
1062 return inbuf[0];
Eric Andersenb052b471999-11-18 00:21:59 +00001063}
1064
1065/* ===========================================================================
1066 * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
1067 * (used for the compressed data only)
1068 */
1069void flush_outbuf()
1070{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001071 if (outcnt == 0)
1072 return;
Eric Andersenb052b471999-11-18 00:21:59 +00001073
Erik Andersene49d5ec2000-02-08 19:58:47 +00001074 if (!test_mode)
1075 write_buf(ofd, (char *) outbuf, outcnt);
1076 bytes_out += (ulg) outcnt;
1077 outcnt = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001078}
1079
1080/* ===========================================================================
1081 * Write the output window window[0..outcnt-1] and update crc and bytes_out.
1082 * (Used for the decompressed data only.)
1083 */
1084void flush_window()
1085{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001086 if (outcnt == 0)
1087 return;
1088 updcrc(window, outcnt);
Eric Andersenb052b471999-11-18 00:21:59 +00001089
Erik Andersene49d5ec2000-02-08 19:58:47 +00001090 if (!test_mode)
1091 write_buf(ofd, (char *) window, outcnt);
1092 bytes_out += (ulg) outcnt;
1093 outcnt = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001094}
1095
1096/* ===========================================================================
1097 * Does the same as write(), but also handles partial pipe writes and checks
1098 * for error return.
1099 */
1100void write_buf(fd, buf, cnt)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001101int fd;
Erik Andersen61677fe2000-04-13 01:18:56 +00001102void * buf;
Erik Andersene49d5ec2000-02-08 19:58:47 +00001103unsigned cnt;
Eric Andersenb052b471999-11-18 00:21:59 +00001104{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001105 unsigned n;
Eric Andersenb052b471999-11-18 00:21:59 +00001106
Erik Andersene49d5ec2000-02-08 19:58:47 +00001107 while ((n = write(fd, buf, cnt)) != cnt) {
1108 if (n == (unsigned) (-1)) {
Erik Andersen330fd2b2000-05-19 05:35:19 +00001109 write_error_msg();
Erik Andersene49d5ec2000-02-08 19:58:47 +00001110 }
1111 cnt -= n;
Erik Andersen61677fe2000-04-13 01:18:56 +00001112 buf = (void *) ((char *) buf + n);
Eric Andersenb052b471999-11-18 00:21:59 +00001113 }
Eric Andersenb052b471999-11-18 00:21:59 +00001114}
1115
1116#if defined(NO_STRING_H) && !defined(STDC_HEADERS)
1117
1118/* Provide missing strspn and strcspn functions. */
1119
1120# ifndef __STDC__
1121# define const
1122# endif
1123
Erik Andersen61677fe2000-04-13 01:18:56 +00001124int strspn (const char *s, const char *accept);
1125int strcspn (const char *s, const char *reject);
Eric Andersenb052b471999-11-18 00:21:59 +00001126
1127/* ========================================================================
1128 * Return the length of the maximum initial segment
1129 * of s which contains only characters in accept.
1130 */
1131int strspn(s, accept)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001132const char *s;
1133const char *accept;
Eric Andersenb052b471999-11-18 00:21:59 +00001134{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001135 register const char *p;
1136 register const char *a;
1137 register int count = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001138
Erik Andersene49d5ec2000-02-08 19:58:47 +00001139 for (p = s; *p != '\0'; ++p) {
1140 for (a = accept; *a != '\0'; ++a) {
1141 if (*p == *a)
1142 break;
1143 }
1144 if (*a == '\0')
1145 return count;
1146 ++count;
Eric Andersenb052b471999-11-18 00:21:59 +00001147 }
Erik Andersene49d5ec2000-02-08 19:58:47 +00001148 return count;
Eric Andersenb052b471999-11-18 00:21:59 +00001149}
1150
1151/* ========================================================================
1152 * Return the length of the maximum inital segment of s
1153 * which contains no characters from reject.
1154 */
1155int strcspn(s, reject)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001156const char *s;
1157const char *reject;
Eric Andersenb052b471999-11-18 00:21:59 +00001158{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001159 register int count = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001160
Erik Andersene49d5ec2000-02-08 19:58:47 +00001161 while (*s != '\0') {
1162 if (strchr(reject, *s++) != NULL)
1163 return count;
1164 ++count;
1165 }
1166 return count;
Eric Andersenb052b471999-11-18 00:21:59 +00001167}
1168
Erik Andersene49d5ec2000-02-08 19:58:47 +00001169#endif /* NO_STRING_H */
Eric Andersenb052b471999-11-18 00:21:59 +00001170
1171
1172/* ========================================================================
1173 * Error handlers.
1174 */
Erik Andersen330fd2b2000-05-19 05:35:19 +00001175void read_error_msg()
Eric Andersenb052b471999-11-18 00:21:59 +00001176{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001177 fprintf(stderr, "\n");
1178 if (errno != 0) {
1179 perror("");
1180 } else {
1181 fprintf(stderr, "unexpected end of file\n");
1182 }
1183 abort_gzip();
Eric Andersenb052b471999-11-18 00:21:59 +00001184}
1185
Erik Andersen330fd2b2000-05-19 05:35:19 +00001186void write_error_msg()
Eric Andersenb052b471999-11-18 00:21:59 +00001187{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001188 fprintf(stderr, "\n");
1189 perror("");
1190 abort_gzip();
Eric Andersenb052b471999-11-18 00:21:59 +00001191}
1192
1193
1194/* ========================================================================
Eric Andersenb052b471999-11-18 00:21:59 +00001195 * Table of CRC-32's of all single-byte values (made by makecrc.c)
1196 */
1197static const ulg crc_32_tab[] = {
Erik Andersene49d5ec2000-02-08 19:58:47 +00001198 0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
1199 0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
1200 0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
1201 0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
1202 0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
1203 0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
1204 0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
1205 0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
1206 0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
1207 0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
1208 0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
1209 0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
1210 0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
1211 0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
1212 0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
1213 0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
1214 0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
1215 0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
1216 0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
1217 0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
1218 0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
1219 0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
1220 0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
1221 0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
1222 0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
1223 0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
1224 0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
1225 0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
1226 0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
1227 0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
1228 0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
1229 0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
1230 0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
1231 0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
1232 0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
1233 0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
1234 0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
1235 0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
1236 0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
1237 0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
1238 0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
1239 0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
1240 0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
1241 0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
1242 0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
1243 0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
1244 0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
1245 0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
1246 0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
1247 0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
1248 0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
1249 0x2d02ef8dL
Eric Andersenb052b471999-11-18 00:21:59 +00001250};
Erik Andersene49d5ec2000-02-08 19:58:47 +00001251
Eric Andersenb052b471999-11-18 00:21:59 +00001252/* inflate.c -- Not copyrighted 1992 by Mark Adler
1253 version c10p1, 10 January 1993 */
1254
1255/* You can do whatever you like with this source file, though I would
1256 prefer that if you modify it and redistribute it that you include
1257 comments to that effect with your name and the date. Thank you.
1258 [The history has been moved to the file ChangeLog.]
1259 */
1260
1261/*
1262 Inflate deflated (PKZIP's method 8 compressed) data. The compression
1263 method searches for as much of the current string of bytes (up to a
1264 length of 258) in the previous 32K bytes. If it doesn't find any
1265 matches (of at least length 3), it codes the next byte. Otherwise, it
1266 codes the length of the matched string and its distance backwards from
1267 the current position. There is a single Huffman code that codes both
1268 single bytes (called "literals") and match lengths. A second Huffman
1269 code codes the distance information, which follows a length code. Each
1270 length or distance code actually represents a base value and a number
1271 of "extra" (sometimes zero) bits to get to add to the base value. At
1272 the end of each deflated block is a special end-of-block (EOB) literal/
1273 length code. The decoding process is basically: get a literal/length
1274 code; if EOB then done; if a literal, emit the decoded byte; if a
1275 length then get the distance and emit the referred-to bytes from the
1276 sliding window of previously emitted data.
1277
1278 There are (currently) three kinds of inflate blocks: stored, fixed, and
1279 dynamic. The compressor deals with some chunk of data at a time, and
1280 decides which method to use on a chunk-by-chunk basis. A chunk might
1281 typically be 32K or 64K. If the chunk is uncompressible, then the
1282 "stored" method is used. In this case, the bytes are simply stored as
1283 is, eight bits per byte, with none of the above coding. The bytes are
1284 preceded by a count, since there is no longer an EOB code.
1285
1286 If the data is compressible, then either the fixed or dynamic methods
1287 are used. In the dynamic method, the compressed data is preceded by
1288 an encoding of the literal/length and distance Huffman codes that are
1289 to be used to decode this block. The representation is itself Huffman
1290 coded, and so is preceded by a description of that code. These code
1291 descriptions take up a little space, and so for small blocks, there is
1292 a predefined set of codes, called the fixed codes. The fixed method is
1293 used if the block codes up smaller that way (usually for quite small
1294 chunks), otherwise the dynamic method is used. In the latter case, the
1295 codes are customized to the probabilities in the current block, and so
1296 can code it much better than the pre-determined fixed codes.
1297
1298 The Huffman codes themselves are decoded using a mutli-level table
1299 lookup, in order to maximize the speed of decoding plus the speed of
1300 building the decoding tables. See the comments below that precede the
1301 lbits and dbits tuning parameters.
1302 */
1303
1304
1305/*
1306 Notes beyond the 1.93a appnote.txt:
1307
1308 1. Distance pointers never point before the beginning of the output
1309 stream.
1310 2. Distance pointers can point back across blocks, up to 32k away.
1311 3. There is an implied maximum of 7 bits for the bit length table and
1312 15 bits for the actual data.
1313 4. If only one code exists, then it is encoded using one bit. (Zero
1314 would be more efficient, but perhaps a little confusing.) If two
1315 codes exist, they are coded using one bit each (0 and 1).
1316 5. There is no way of sending zero distance codes--a dummy must be
1317 sent if there are none. (History: a pre 2.0 version of PKZIP would
1318 store blocks with no distance codes, but this was discovered to be
1319 too harsh a criterion.) Valid only for 1.93a. 2.04c does allow
1320 zero distance codes, which is sent as one code of zero bits in
1321 length.
1322 6. There are up to 286 literal/length codes. Code 256 represents the
1323 end-of-block. Note however that the static length tree defines
1324 288 codes just to fill out the Huffman codes. Codes 286 and 287
1325 cannot be used though, since there is no length base or extra bits
1326 defined for them. Similarly, there are up to 30 distance codes.
1327 However, static trees define 32 codes (all 5 bits) to fill out the
1328 Huffman codes, but the last two had better not show up in the data.
1329 7. Unzip can check dynamic Huffman blocks for complete code sets.
1330 The exception is that a single code would not be complete (see #4).
1331 8. The five bits following the block type is really the number of
1332 literal codes sent minus 257.
1333 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
1334 (1+6+6). Therefore, to output three times the length, you output
1335 three codes (1+1+1), whereas to output four times the same length,
1336 you only need two codes (1+3). Hmm.
1337 10. In the tree reconstruction algorithm, Code = Code + Increment
1338 only if BitLength(i) is not zero. (Pretty obvious.)
1339 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19)
1340 12. Note: length code 284 can represent 227-258, but length code 285
1341 really is 258. The last length deserves its own, short code
1342 since it gets used a lot in very redundant files. The length
1343 258 is special since 258 - 3 (the min match length) is 255.
1344 13. The literal/length and distance code bit lengths are read as a
1345 single stream of lengths. It is possible (and advantageous) for
1346 a repeat code (16, 17, or 18) to go across the boundary between
1347 the two sets of lengths.
1348 */
1349
1350#include <sys/types.h>
1351
1352#if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1353# include <stdlib.h>
1354#endif
1355
1356
1357#define slide window
1358
1359/* Huffman code lookup table entry--this entry is four bytes for machines
1360 that have 16-bit pointers (e.g. PC's in the small or medium model).
1361 Valid extra bits are 0..13. e == 15 is EOB (end of block), e == 16
1362 means that v is a literal, 16 < e < 32 means that v is a pointer to
1363 the next table, which codes e - 16 bits, and lastly e == 99 indicates
1364 an unused code. If a code with e == 99 is looked up, this implies an
1365 error in the data. */
1366struct huft {
Erik Andersene49d5ec2000-02-08 19:58:47 +00001367 uch e; /* number of extra bits or operation */
1368 uch b; /* number of bits in this code or subcode */
1369 union {
1370 ush n; /* literal, length base, or distance base */
1371 struct huft *t; /* pointer to next level of table */
1372 } v;
Eric Andersenb052b471999-11-18 00:21:59 +00001373};
1374
1375
1376/* Function prototypes */
Erik Andersen61677fe2000-04-13 01:18:56 +00001377int huft_build (unsigned *, unsigned, unsigned, ush *, ush *,
1378 struct huft **, int *);
1379int huft_free (struct huft *);
1380int inflate_codes (struct huft *, struct huft *, int, int);
1381int inflate_stored (void);
1382int inflate_fixed (void);
1383int inflate_dynamic (void);
1384int inflate_block (int *);
1385int inflate (void);
Eric Andersenb052b471999-11-18 00:21:59 +00001386
1387
1388/* The inflate algorithm uses a sliding 32K byte window on the uncompressed
1389 stream to find repeated byte strings. This is implemented here as a
1390 circular buffer. The index is updated simply by incrementing and then
1391 and'ing with 0x7fff (32K-1). */
1392/* It is left to other modules to supply the 32K area. It is assumed
1393 to be usable as if it were declared "uch slide[32768];" or as just
1394 "uch *slide;" and then malloc'ed in the latter case. The definition
1395 must be in unzip.h, included above. */
1396/* unsigned wp; current position in slide */
1397#define wp outcnt
1398#define flush_output(w) (wp=(w),flush_window())
1399
1400/* Tables for deflate from PKZIP's appnote.txt. */
Erik Andersene49d5ec2000-02-08 19:58:47 +00001401static unsigned border[] = { /* Order of the bit length code lengths */
1402 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15
1403};
1404static ush cplens[] = { /* Copy lengths for literal codes 257..285 */
1405 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
1406 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0
1407};
1408
1409 /* note: see note #13 above about the 258 in this list. */
1410static ush cplext[] = { /* Extra bits for literal codes 257..285 */
1411 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
1412 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99
1413}; /* 99==invalid */
1414static ush cpdist[] = { /* Copy offsets for distance codes 0..29 */
1415 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
1416 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
1417 8193, 12289, 16385, 24577
1418};
1419static ush cpdext[] = { /* Extra bits for distance codes */
1420 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
1421 7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
1422 12, 12, 13, 13
1423};
Eric Andersenb052b471999-11-18 00:21:59 +00001424
1425
1426
1427/* Macros for inflate() bit peeking and grabbing.
1428 The usage is:
1429
1430 NEEDBITS(j)
1431 x = b & mask_bits[j];
1432 DUMPBITS(j)
1433
1434 where NEEDBITS makes sure that b has at least j bits in it, and
1435 DUMPBITS removes the bits from b. The macros use the variable k
1436 for the number of bits in b. Normally, b and k are register
1437 variables for speed, and are initialized at the beginning of a
1438 routine that uses these macros from a global bit buffer and count.
1439
1440 If we assume that EOB will be the longest code, then we will never
1441 ask for bits with NEEDBITS that are beyond the end of the stream.
1442 So, NEEDBITS should not read any more bytes than are needed to
1443 meet the request. Then no bytes need to be "returned" to the buffer
1444 at the end of the last block.
1445
1446 However, this assumption is not true for fixed blocks--the EOB code
1447 is 7 bits, but the other literal/length codes can be 8 or 9 bits.
1448 (The EOB code is shorter than other codes because fixed blocks are
1449 generally short. So, while a block always has an EOB, many other
1450 literal/length codes have a significantly lower probability of
1451 showing up at all.) However, by making the first table have a
1452 lookup of seven bits, the EOB code will be found in that first
1453 lookup, and so will not require that too many bits be pulled from
1454 the stream.
1455 */
1456
Erik Andersene49d5ec2000-02-08 19:58:47 +00001457ulg bb; /* bit buffer */
1458unsigned bk; /* bits in bit buffer */
Eric Andersenb052b471999-11-18 00:21:59 +00001459
1460ush mask_bits[] = {
Erik Andersene49d5ec2000-02-08 19:58:47 +00001461 0x0000,
1462 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
1463 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
Eric Andersenb052b471999-11-18 00:21:59 +00001464};
1465
1466#ifdef CRYPT
Erik Andersene49d5ec2000-02-08 19:58:47 +00001467uch cc;
1468
Eric Andersenb052b471999-11-18 00:21:59 +00001469# define NEXTBYTE() (cc = get_byte(), zdecode(cc), cc)
1470#else
1471# define NEXTBYTE() (uch)get_byte()
1472#endif
1473#define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
1474#define DUMPBITS(n) {b>>=(n);k-=(n);}
1475
1476
1477/*
1478 Huffman code decoding is performed using a multi-level table lookup.
1479 The fastest way to decode is to simply build a lookup table whose
1480 size is determined by the longest code. However, the time it takes
1481 to build this table can also be a factor if the data being decoded
1482 is not very long. The most common codes are necessarily the
1483 shortest codes, so those codes dominate the decoding time, and hence
1484 the speed. The idea is you can have a shorter table that decodes the
1485 shorter, more probable codes, and then point to subsidiary tables for
1486 the longer codes. The time it costs to decode the longer codes is
1487 then traded against the time it takes to make longer tables.
1488
1489 This results of this trade are in the variables lbits and dbits
1490 below. lbits is the number of bits the first level table for literal/
1491 length codes can decode in one step, and dbits is the same thing for
1492 the distance codes. Subsequent tables are also less than or equal to
1493 those sizes. These values may be adjusted either when all of the
1494 codes are shorter than that, in which case the longest code length in
1495 bits is used, or when the shortest code is *longer* than the requested
1496 table size, in which case the length of the shortest code in bits is
1497 used.
1498
1499 There are two different values for the two tables, since they code a
1500 different number of possibilities each. The literal/length table
1501 codes 286 possible values, or in a flat code, a little over eight
1502 bits. The distance table codes 30 possible values, or a little less
1503 than five bits, flat. The optimum values for speed end up being
1504 about one bit more than those, so lbits is 8+1 and dbits is 5+1.
1505 The optimum values may differ though from machine to machine, and
1506 possibly even between compilers. Your mileage may vary.
1507 */
1508
1509
Erik Andersene49d5ec2000-02-08 19:58:47 +00001510int lbits = 9; /* bits in base literal/length lookup table */
1511int dbits = 6; /* bits in base distance lookup table */
Eric Andersenb052b471999-11-18 00:21:59 +00001512
1513
1514/* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
Erik Andersene49d5ec2000-02-08 19:58:47 +00001515#define BMAX 16 /* maximum bit length of any code (16 for explode) */
1516#define N_MAX 288 /* maximum number of codes in any set */
Eric Andersenb052b471999-11-18 00:21:59 +00001517
1518
Erik Andersene49d5ec2000-02-08 19:58:47 +00001519unsigned hufts; /* track memory usage */
Eric Andersenb052b471999-11-18 00:21:59 +00001520
1521
1522int huft_build(b, n, s, d, e, t, m)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001523unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
1524unsigned n; /* number of codes (assumed <= N_MAX) */
1525unsigned s; /* number of simple-valued codes (0..s-1) */
1526ush *d; /* list of base values for non-simple codes */
1527ush *e; /* list of extra bits for non-simple codes */
1528struct huft **t; /* result: starting table */
1529int *m; /* maximum lookup bits, returns actual */
1530
Eric Andersenb052b471999-11-18 00:21:59 +00001531/* Given a list of code lengths and a maximum table size, make a set of
1532 tables to decode that set of codes. Return zero on success, one if
1533 the given code set is incomplete (the tables are still built in this
1534 case), two if the input is invalid (all zero length codes or an
1535 oversubscribed set of lengths), and three if not enough memory. */
1536{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001537 unsigned a; /* counter for codes of length k */
1538 unsigned c[BMAX + 1]; /* bit length count table */
1539 unsigned f; /* i repeats in table every f entries */
1540 int g; /* maximum code length */
1541 int h; /* table level */
1542 register unsigned i; /* counter, current code */
1543 register unsigned j; /* counter */
1544 register int k; /* number of bits in current code */
1545 int l; /* bits per table (returned in m) */
1546 register unsigned *p; /* pointer into c[], b[], or v[] */
1547 register struct huft *q; /* points to current table */
1548 struct huft r; /* table entry for structure assignment */
1549 struct huft *u[BMAX]; /* table stack */
1550 unsigned v[N_MAX]; /* values in order of bit length */
1551 register int w; /* bits before this table == (l * h) */
1552 unsigned x[BMAX + 1]; /* bit offsets, then code stack */
1553 unsigned *xp; /* pointer into x */
1554 int y; /* number of dummy codes added */
1555 unsigned z; /* number of entries in current table */
Eric Andersenb052b471999-11-18 00:21:59 +00001556
1557
Erik Andersene49d5ec2000-02-08 19:58:47 +00001558 /* Generate counts for each bit length */
1559 memzero(c, sizeof(c));
1560 p = b;
1561 i = n;
1562 do {
1563 Tracecv(*p,
1564 (stderr,
1565 (n - i >= ' '
1566 && n - i <= '~' ? "%c %d\n" : "0x%x %d\n"), n - i, *p));
1567 c[*p]++; /* assume all entries <= BMAX */
1568 p++; /* Can't combine with above line (Solaris bug) */
1569 } while (--i);
1570 if (c[0] == n) { /* null input--all zero length codes */
1571 *t = (struct huft *) NULL;
1572 *m = 0;
1573 return 0;
1574 }
Eric Andersenb052b471999-11-18 00:21:59 +00001575
1576
Erik Andersene49d5ec2000-02-08 19:58:47 +00001577 /* Find minimum and maximum length, bound *m by those */
1578 l = *m;
1579 for (j = 1; j <= BMAX; j++)
1580 if (c[j])
1581 break;
1582 k = j; /* minimum code length */
1583 if ((unsigned) l < j)
1584 l = j;
1585 for (i = BMAX; i; i--)
1586 if (c[i])
1587 break;
1588 g = i; /* maximum code length */
1589 if ((unsigned) l > i)
1590 l = i;
1591 *m = l;
Eric Andersenb052b471999-11-18 00:21:59 +00001592
1593
Erik Andersene49d5ec2000-02-08 19:58:47 +00001594 /* Adjust last length count to fill out codes, if needed */
1595 for (y = 1 << j; j < i; j++, y <<= 1)
1596 if ((y -= c[j]) < 0)
1597 return 2; /* bad input: more codes than bits */
1598 if ((y -= c[i]) < 0)
1599 return 2;
1600 c[i] += y;
Eric Andersenb052b471999-11-18 00:21:59 +00001601
1602
Erik Andersene49d5ec2000-02-08 19:58:47 +00001603 /* Generate starting offsets into the value table for each length */
1604 x[1] = j = 0;
1605 p = c + 1;
1606 xp = x + 2;
1607 while (--i) { /* note that i == g from above */
1608 *xp++ = (j += *p++);
1609 }
Eric Andersenb052b471999-11-18 00:21:59 +00001610
1611
Erik Andersene49d5ec2000-02-08 19:58:47 +00001612 /* Make a table of values in order of bit lengths */
1613 p = b;
1614 i = 0;
1615 do {
1616 if ((j = *p++) != 0)
1617 v[x[j]++] = i;
1618 } while (++i < n);
Eric Andersenb052b471999-11-18 00:21:59 +00001619
1620
Erik Andersene49d5ec2000-02-08 19:58:47 +00001621 /* Generate the Huffman codes and for each, make the table entries */
1622 x[0] = i = 0; /* first Huffman code is zero */
1623 p = v; /* grab values in bit order */
1624 h = -1; /* no tables yet--level -1 */
1625 w = -l; /* bits decoded == (l * h) */
1626 u[0] = (struct huft *) NULL; /* just to keep compilers happy */
1627 q = (struct huft *) NULL; /* ditto */
1628 z = 0; /* ditto */
Eric Andersenb052b471999-11-18 00:21:59 +00001629
Erik Andersene49d5ec2000-02-08 19:58:47 +00001630 /* go through the bit lengths (k already is bits in shortest code) */
1631 for (; k <= g; k++) {
1632 a = c[k];
1633 while (a--) {
1634 /* here i is the Huffman code of length k bits for value *p */
1635 /* make tables up to required level */
1636 while (k > w + l) {
1637 h++;
1638 w += l; /* previous table always l bits */
Eric Andersenb052b471999-11-18 00:21:59 +00001639
Erik Andersene49d5ec2000-02-08 19:58:47 +00001640 /* compute minimum size table less than or equal to l bits */
1641 z = (z = g - w) > (unsigned) l ? l : z; /* upper limit on table size */
1642 if ((f = 1 << (j = k - w)) > a + 1) { /* try a k-w bit table *//* too few codes for k-w bit table */
1643 f -= a + 1; /* deduct codes from patterns left */
1644 xp = c + k;
1645 while (++j < z) { /* try smaller tables up to z bits */
1646 if ((f <<= 1) <= *++xp)
1647 break; /* enough codes to use up j bits */
1648 f -= *xp; /* else deduct codes from patterns */
1649 }
1650 }
1651 z = 1 << j; /* table entries for j-bit table */
Eric Andersenb052b471999-11-18 00:21:59 +00001652
Erik Andersene49d5ec2000-02-08 19:58:47 +00001653 /* allocate and link in new table */
1654 if (
1655 (q =
1656 (struct huft *) malloc((z + 1) *
1657 sizeof(struct huft))) ==
1658 (struct huft *) NULL) {
1659 if (h)
1660 huft_free(u[0]);
1661 return 3; /* not enough memory */
1662 }
1663 hufts += z + 1; /* track memory usage */
1664 *t = q + 1; /* link to list for huft_free() */
1665 *(t = &(q->v.t)) = (struct huft *) NULL;
1666 u[h] = ++q; /* table starts after link */
Eric Andersenb052b471999-11-18 00:21:59 +00001667
Erik Andersene49d5ec2000-02-08 19:58:47 +00001668 /* connect to last table, if there is one */
1669 if (h) {
1670 x[h] = i; /* save pattern for backing up */
1671 r.b = (uch) l; /* bits to dump before this table */
1672 r.e = (uch) (16 + j); /* bits in this table */
1673 r.v.t = q; /* pointer to this table */
1674 j = i >> (w - l); /* (get around Turbo C bug) */
1675 u[h - 1][j] = r; /* connect to last table */
1676 }
1677 }
Eric Andersenb052b471999-11-18 00:21:59 +00001678
Erik Andersene49d5ec2000-02-08 19:58:47 +00001679 /* set up table entry in r */
1680 r.b = (uch) (k - w);
1681 if (p >= v + n)
1682 r.e = 99; /* out of values--invalid code */
1683 else if (*p < s) {
1684 r.e = (uch) (*p < 256 ? 16 : 15); /* 256 is end-of-block code */
1685 r.v.n = (ush) (*p); /* simple code is just the value */
1686 p++; /* one compiler does not like *p++ */
1687 } else {
1688 r.e = (uch) e[*p - s]; /* non-simple--look up in lists */
1689 r.v.n = d[*p++ - s];
1690 }
Eric Andersenb052b471999-11-18 00:21:59 +00001691
Erik Andersene49d5ec2000-02-08 19:58:47 +00001692 /* fill code-like entries with r */
1693 f = 1 << (k - w);
1694 for (j = i >> w; j < z; j += f)
1695 q[j] = r;
Eric Andersenb052b471999-11-18 00:21:59 +00001696
Erik Andersene49d5ec2000-02-08 19:58:47 +00001697 /* backwards increment the k-bit code i */
1698 for (j = 1 << (k - 1); i & j; j >>= 1)
1699 i ^= j;
1700 i ^= j;
Eric Andersenb052b471999-11-18 00:21:59 +00001701
Erik Andersene49d5ec2000-02-08 19:58:47 +00001702 /* backup over finished tables */
1703 while ((i & ((1 << w) - 1)) != x[h]) {
1704 h--; /* don't need to update q */
1705 w -= l;
1706 }
1707 }
1708 }
Eric Andersenb052b471999-11-18 00:21:59 +00001709
1710
Erik Andersene49d5ec2000-02-08 19:58:47 +00001711 /* Return true (1) if we were given an incomplete table */
1712 return y != 0 && g != 1;
Eric Andersenb052b471999-11-18 00:21:59 +00001713}
1714
1715
1716
1717int huft_free(t)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001718struct huft *t; /* table to free */
1719
Eric Andersenb052b471999-11-18 00:21:59 +00001720/* Free the malloc'ed tables built by huft_build(), which makes a linked
1721 list of the tables it made, with the links in a dummy first entry of
1722 each table. */
1723{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001724 register struct huft *p, *q;
Eric Andersenb052b471999-11-18 00:21:59 +00001725
1726
Erik Andersene49d5ec2000-02-08 19:58:47 +00001727 /* Go through linked list, freeing from the malloced (t[-1]) address. */
1728 p = t;
1729 while (p != (struct huft *) NULL) {
1730 q = (--p)->v.t;
1731 free((char *) p);
1732 p = q;
1733 }
1734 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001735}
1736
1737
1738int inflate_codes(tl, td, bl, bd)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001739struct huft *tl, *td; /* literal/length and distance decoder tables */
1740int bl, bd; /* number of bits decoded by tl[] and td[] */
1741
Eric Andersenb052b471999-11-18 00:21:59 +00001742/* inflate (decompress) the codes in a deflated (compressed) block.
1743 Return an error code or zero if it all goes ok. */
1744{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001745 register unsigned e; /* table entry flag/number of extra bits */
1746 unsigned n, d; /* length and index for copy */
1747 unsigned w; /* current window position */
1748 struct huft *t; /* pointer to table entry */
1749 unsigned ml, md; /* masks for bl and bd bits */
1750 register ulg b; /* bit buffer */
1751 register unsigned k; /* number of bits in bit buffer */
Eric Andersenb052b471999-11-18 00:21:59 +00001752
1753
Erik Andersene49d5ec2000-02-08 19:58:47 +00001754 /* make local copies of globals */
1755 b = bb; /* initialize bit buffer */
1756 k = bk;
1757 w = wp; /* initialize window position */
Eric Andersenb052b471999-11-18 00:21:59 +00001758
Erik Andersene49d5ec2000-02-08 19:58:47 +00001759 /* inflate the coded data */
1760 ml = mask_bits[bl]; /* precompute masks for speed */
1761 md = mask_bits[bd];
1762 for (;;) { /* do until end of block */
1763 NEEDBITS((unsigned) bl)
1764 if ((e = (t = tl + ((unsigned) b & ml))->e) > 16)
1765 do {
1766 if (e == 99)
1767 return 1;
1768 DUMPBITS(t->b)
1769 e -= 16;
1770 NEEDBITS(e)
1771 } while ((e = (t = t->v.t + ((unsigned) b & mask_bits[e]))->e)
1772 > 16);
1773 DUMPBITS(t->b)
1774 if (e == 16) { /* then it's a literal */
1775 slide[w++] = (uch) t->v.n;
1776 Tracevv((stderr, "%c", slide[w - 1]));
1777 if (w == WSIZE) {
1778 flush_output(w);
1779 w = 0;
1780 }
1781 } else { /* it's an EOB or a length */
Eric Andersenb052b471999-11-18 00:21:59 +00001782
Erik Andersene49d5ec2000-02-08 19:58:47 +00001783 /* exit if end of block */
1784 if (e == 15)
1785 break;
Eric Andersenb052b471999-11-18 00:21:59 +00001786
Erik Andersene49d5ec2000-02-08 19:58:47 +00001787 /* get length of block to copy */
1788 NEEDBITS(e)
1789 n = t->v.n + ((unsigned) b & mask_bits[e]);
1790 DUMPBITS(e);
Eric Andersenb052b471999-11-18 00:21:59 +00001791
Erik Andersene49d5ec2000-02-08 19:58:47 +00001792 /* decode distance of block to copy */
1793 NEEDBITS((unsigned) bd)
1794 if ((e = (t = td + ((unsigned) b & md))->e) > 16)
1795 do {
1796 if (e == 99)
1797 return 1;
1798 DUMPBITS(t->b)
1799 e -= 16;
1800 NEEDBITS(e)
1801 }
1802 while (
1803 (e =
1804 (t =
1805 t->v.t + ((unsigned) b & mask_bits[e]))->e) >
1806 16);
1807 DUMPBITS(t->b)
1808 NEEDBITS(e)
1809 d = w - t->v.n - ((unsigned) b & mask_bits[e]);
1810 DUMPBITS(e)
1811 Tracevv((stderr, "\\[%d,%d]", w - d, n));
1812
1813 /* do the copy */
1814 do {
1815 n -= (e =
1816 (e =
1817 WSIZE - ((d &= WSIZE - 1) > w ? d : w)) >
1818 n ? n : e);
Eric Andersenb052b471999-11-18 00:21:59 +00001819#if !defined(NOMEMCPY) && !defined(DEBUG)
Erik Andersene49d5ec2000-02-08 19:58:47 +00001820 if (w - d >= e) { /* (this test assumes unsigned comparison) */
1821 memcpy(slide + w, slide + d, e);
1822 w += e;
1823 d += e;
1824 } else /* do it slow to avoid memcpy() overlap */
1825#endif /* !NOMEMCPY */
1826 do {
1827 slide[w++] = slide[d++];
1828 Tracevv((stderr, "%c", slide[w - 1]));
1829 } while (--e);
1830 if (w == WSIZE) {
1831 flush_output(w);
1832 w = 0;
1833 }
1834 } while (n);
1835 }
1836 }
Eric Andersenb052b471999-11-18 00:21:59 +00001837
1838
Erik Andersene49d5ec2000-02-08 19:58:47 +00001839 /* restore the globals from the locals */
1840 wp = w; /* restore global window pointer */
1841 bb = b; /* restore global bit buffer */
1842 bk = k;
Eric Andersenb052b471999-11-18 00:21:59 +00001843
Erik Andersene49d5ec2000-02-08 19:58:47 +00001844 /* done */
1845 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001846}
1847
1848
1849
1850int inflate_stored()
1851/* "decompress" an inflated type 0 (stored) block. */
1852{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001853 unsigned n; /* number of bytes in block */
1854 unsigned w; /* current window position */
1855 register ulg b; /* bit buffer */
1856 register unsigned k; /* number of bits in bit buffer */
Eric Andersenb052b471999-11-18 00:21:59 +00001857
1858
Erik Andersene49d5ec2000-02-08 19:58:47 +00001859 /* make local copies of globals */
1860 b = bb; /* initialize bit buffer */
1861 k = bk;
1862 w = wp; /* initialize window position */
Eric Andersenb052b471999-11-18 00:21:59 +00001863
1864
Erik Andersene49d5ec2000-02-08 19:58:47 +00001865 /* go to byte boundary */
1866 n = k & 7;
1867 DUMPBITS(n);
Eric Andersenb052b471999-11-18 00:21:59 +00001868
1869
Erik Andersene49d5ec2000-02-08 19:58:47 +00001870 /* get the length and its complement */
1871 NEEDBITS(16)
1872 n = ((unsigned) b & 0xffff);
1873 DUMPBITS(16)
1874 NEEDBITS(16)
1875 if (n != (unsigned) ((~b) & 0xffff))
1876 return 1; /* error in compressed data */
1877 DUMPBITS(16)
Eric Andersenb052b471999-11-18 00:21:59 +00001878
1879
Erik Andersene49d5ec2000-02-08 19:58:47 +00001880 /* read and output the compressed data */
1881 while (n--) {
1882 NEEDBITS(8)
1883 slide[w++] = (uch) b;
1884 if (w == WSIZE) {
1885 flush_output(w);
1886 w = 0;
1887 }
1888 DUMPBITS(8)
1889 }
Eric Andersenb052b471999-11-18 00:21:59 +00001890
1891
Erik Andersene49d5ec2000-02-08 19:58:47 +00001892 /* restore the globals from the locals */
1893 wp = w; /* restore global window pointer */
1894 bb = b; /* restore global bit buffer */
1895 bk = k;
1896 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001897}
1898
1899
1900
1901int inflate_fixed()
1902/* decompress an inflated type 1 (fixed Huffman codes) block. We should
1903 either replace this with a custom decoder, or at least precompute the
1904 Huffman tables. */
1905{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001906 int i; /* temporary variable */
1907 struct huft *tl; /* literal/length code table */
1908 struct huft *td; /* distance code table */
1909 int bl; /* lookup bits for tl */
1910 int bd; /* lookup bits for td */
1911 unsigned l[288]; /* length list for huft_build */
Eric Andersenb052b471999-11-18 00:21:59 +00001912
1913
Erik Andersene49d5ec2000-02-08 19:58:47 +00001914 /* set up literal table */
1915 for (i = 0; i < 144; i++)
1916 l[i] = 8;
1917 for (; i < 256; i++)
1918 l[i] = 9;
1919 for (; i < 280; i++)
1920 l[i] = 7;
1921 for (; i < 288; i++) /* make a complete, but wrong code set */
1922 l[i] = 8;
1923 bl = 7;
1924 if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
1925 return i;
Eric Andersenb052b471999-11-18 00:21:59 +00001926
1927
Erik Andersene49d5ec2000-02-08 19:58:47 +00001928 /* set up distance table */
1929 for (i = 0; i < 30; i++) /* make an incomplete code set */
1930 l[i] = 5;
1931 bd = 5;
1932 if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1) {
1933 huft_free(tl);
1934 return i;
1935 }
Eric Andersenb052b471999-11-18 00:21:59 +00001936
1937
Erik Andersene49d5ec2000-02-08 19:58:47 +00001938 /* decompress until an end-of-block code */
1939 if (inflate_codes(tl, td, bl, bd))
1940 return 1;
Eric Andersenb052b471999-11-18 00:21:59 +00001941
1942
Erik Andersene49d5ec2000-02-08 19:58:47 +00001943 /* free the decoding tables, return */
1944 huft_free(tl);
1945 huft_free(td);
1946 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00001947}
1948
1949
1950
1951int inflate_dynamic()
1952/* decompress an inflated type 2 (dynamic Huffman codes) block. */
1953{
Erik Andersene49d5ec2000-02-08 19:58:47 +00001954 int i; /* temporary variables */
1955 unsigned j;
1956 unsigned l; /* last length */
1957 unsigned m; /* mask for bit lengths table */
1958 unsigned n; /* number of lengths to get */
1959 struct huft *tl; /* literal/length code table */
1960 struct huft *td; /* distance code table */
1961 int bl; /* lookup bits for tl */
1962 int bd; /* lookup bits for td */
1963 unsigned nb; /* number of bit length codes */
1964 unsigned nl; /* number of literal/length codes */
1965 unsigned nd; /* number of distance codes */
1966
Eric Andersenb052b471999-11-18 00:21:59 +00001967#ifdef PKZIP_BUG_WORKAROUND
Erik Andersene49d5ec2000-02-08 19:58:47 +00001968 unsigned ll[288 + 32]; /* literal/length and distance code lengths */
Eric Andersenb052b471999-11-18 00:21:59 +00001969#else
Erik Andersene49d5ec2000-02-08 19:58:47 +00001970 unsigned ll[286 + 30]; /* literal/length and distance code lengths */
Eric Andersenb052b471999-11-18 00:21:59 +00001971#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +00001972 register ulg b; /* bit buffer */
1973 register unsigned k; /* number of bits in bit buffer */
Eric Andersenb052b471999-11-18 00:21:59 +00001974
1975
Erik Andersene49d5ec2000-02-08 19:58:47 +00001976 /* make local bit buffer */
1977 b = bb;
1978 k = bk;
Eric Andersenb052b471999-11-18 00:21:59 +00001979
1980
Erik Andersene49d5ec2000-02-08 19:58:47 +00001981 /* read in table lengths */
1982 NEEDBITS(5)
1983 nl = 257 + ((unsigned) b & 0x1f); /* number of literal/length codes */
1984 DUMPBITS(5)
1985 NEEDBITS(5)
1986 nd = 1 + ((unsigned) b & 0x1f); /* number of distance codes */
1987 DUMPBITS(5)
1988 NEEDBITS(4)
1989 nb = 4 + ((unsigned) b & 0xf); /* number of bit length codes */
1990 DUMPBITS(4)
Eric Andersenb052b471999-11-18 00:21:59 +00001991#ifdef PKZIP_BUG_WORKAROUND
Erik Andersene49d5ec2000-02-08 19:58:47 +00001992 if (nl > 288 || nd > 32)
Eric Andersenb052b471999-11-18 00:21:59 +00001993#else
Erik Andersene49d5ec2000-02-08 19:58:47 +00001994 if (nl > 286 || nd > 30)
Eric Andersenb052b471999-11-18 00:21:59 +00001995#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +00001996 return 1; /* bad lengths */
Eric Andersenb052b471999-11-18 00:21:59 +00001997
1998
Erik Andersene49d5ec2000-02-08 19:58:47 +00001999 /* read in bit-length-code lengths */
2000 for (j = 0; j < nb; j++) {
2001 NEEDBITS(3)
2002 ll[border[j]] = (unsigned) b & 7;
2003 DUMPBITS(3)
2004 }
2005 for (; j < 19; j++)
2006 ll[border[j]] = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00002007
2008
Erik Andersene49d5ec2000-02-08 19:58:47 +00002009 /* build decoding table for trees--single level, 7 bit lookup */
2010 bl = 7;
2011 if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0) {
2012 if (i == 1)
2013 huft_free(tl);
2014 return i; /* incomplete code set */
2015 }
Eric Andersenb052b471999-11-18 00:21:59 +00002016
2017
Erik Andersene49d5ec2000-02-08 19:58:47 +00002018 /* read in literal and distance code lengths */
2019 n = nl + nd;
2020 m = mask_bits[bl];
2021 i = l = 0;
2022 while ((unsigned) i < n) {
2023 NEEDBITS((unsigned) bl)
2024 j = (td = tl + ((unsigned) b & m))->b;
2025 DUMPBITS(j)
2026 j = td->v.n;
2027 if (j < 16) /* length of code in bits (0..15) */
2028 ll[i++] = l = j; /* save last length in l */
2029 else if (j == 16) { /* repeat last length 3 to 6 times */
2030 NEEDBITS(2)
2031 j = 3 + ((unsigned) b & 3);
2032 DUMPBITS(2)
2033 if ((unsigned) i + j > n)
2034 return 1;
2035 while (j--)
2036 ll[i++] = l;
2037 } else if (j == 17) { /* 3 to 10 zero length codes */
2038 NEEDBITS(3)
2039 j = 3 + ((unsigned) b & 7);
2040 DUMPBITS(3)
2041 if ((unsigned) i + j > n)
2042 return 1;
2043 while (j--)
2044 ll[i++] = 0;
2045 l = 0;
2046 } else { /* j == 18: 11 to 138 zero length codes */
2047
2048 NEEDBITS(7)
2049 j = 11 + ((unsigned) b & 0x7f);
2050 DUMPBITS(7)
2051 if ((unsigned) i + j > n)
2052 return 1;
2053 while (j--)
2054 ll[i++] = 0;
2055 l = 0;
2056 }
2057 }
Eric Andersenb052b471999-11-18 00:21:59 +00002058
2059
Erik Andersene49d5ec2000-02-08 19:58:47 +00002060 /* free decoding table for trees */
2061 huft_free(tl);
Eric Andersenb052b471999-11-18 00:21:59 +00002062
2063
Erik Andersene49d5ec2000-02-08 19:58:47 +00002064 /* restore the global bit buffer */
2065 bb = b;
2066 bk = k;
Eric Andersenb052b471999-11-18 00:21:59 +00002067
2068
Erik Andersene49d5ec2000-02-08 19:58:47 +00002069 /* build the decoding tables for literal/length and distance codes */
2070 bl = lbits;
2071 if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0) {
2072 if (i == 1) {
2073 fprintf(stderr, " incomplete literal tree\n");
2074 huft_free(tl);
2075 }
2076 return i; /* incomplete code set */
2077 }
2078 bd = dbits;
2079 if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0) {
2080 if (i == 1) {
2081 fprintf(stderr, " incomplete distance tree\n");
Eric Andersenb052b471999-11-18 00:21:59 +00002082#ifdef PKZIP_BUG_WORKAROUND
Erik Andersene49d5ec2000-02-08 19:58:47 +00002083 i = 0;
2084 }
Eric Andersenb052b471999-11-18 00:21:59 +00002085#else
Erik Andersene49d5ec2000-02-08 19:58:47 +00002086 huft_free(td);
2087 }
2088 huft_free(tl);
2089 return i; /* incomplete code set */
Eric Andersenb052b471999-11-18 00:21:59 +00002090#endif
Erik Andersene49d5ec2000-02-08 19:58:47 +00002091 }
Eric Andersenb052b471999-11-18 00:21:59 +00002092
2093
Erik Andersene49d5ec2000-02-08 19:58:47 +00002094 /* decompress until an end-of-block code */
2095 if (inflate_codes(tl, td, bl, bd))
2096 return 1;
Eric Andersenb052b471999-11-18 00:21:59 +00002097
2098
Erik Andersene49d5ec2000-02-08 19:58:47 +00002099 /* free the decoding tables, return */
2100 huft_free(tl);
2101 huft_free(td);
2102 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00002103}
2104
2105
2106
2107int inflate_block(e)
Erik Andersene49d5ec2000-02-08 19:58:47 +00002108int *e; /* last block flag */
2109
Eric Andersenb052b471999-11-18 00:21:59 +00002110/* decompress an inflated block */
2111{
Erik Andersene49d5ec2000-02-08 19:58:47 +00002112 unsigned t; /* block type */
2113 register ulg b; /* bit buffer */
2114 register unsigned k; /* number of bits in bit buffer */
Eric Andersenb052b471999-11-18 00:21:59 +00002115
2116
Erik Andersene49d5ec2000-02-08 19:58:47 +00002117 /* make local bit buffer */
2118 b = bb;
2119 k = bk;
Eric Andersenb052b471999-11-18 00:21:59 +00002120
2121
Erik Andersene49d5ec2000-02-08 19:58:47 +00002122 /* read in last block bit */
2123 NEEDBITS(1)
2124 * e = (int) b & 1;
2125 DUMPBITS(1)
Eric Andersenb052b471999-11-18 00:21:59 +00002126
2127
Erik Andersene49d5ec2000-02-08 19:58:47 +00002128 /* read in block type */
2129 NEEDBITS(2)
2130 t = (unsigned) b & 3;
2131 DUMPBITS(2)
Eric Andersenb052b471999-11-18 00:21:59 +00002132
2133
Erik Andersene49d5ec2000-02-08 19:58:47 +00002134 /* restore the global bit buffer */
2135 bb = b;
2136 bk = k;
Eric Andersenb052b471999-11-18 00:21:59 +00002137
2138
Erik Andersene49d5ec2000-02-08 19:58:47 +00002139 /* inflate that block type */
2140 if (t == 2)
2141 return inflate_dynamic();
2142 if (t == 0)
2143 return inflate_stored();
2144 if (t == 1)
2145 return inflate_fixed();
Eric Andersenb052b471999-11-18 00:21:59 +00002146
2147
Erik Andersene49d5ec2000-02-08 19:58:47 +00002148 /* bad block type */
2149 return 2;
Eric Andersenb052b471999-11-18 00:21:59 +00002150}
2151
2152
2153
2154int inflate()
2155/* decompress an inflated entry */
2156{
Erik Andersene49d5ec2000-02-08 19:58:47 +00002157 int e; /* last block flag */
2158 int r; /* result code */
2159 unsigned h; /* maximum struct huft's malloc'ed */
Eric Andersenb052b471999-11-18 00:21:59 +00002160
2161
Erik Andersene49d5ec2000-02-08 19:58:47 +00002162 /* initialize window, bit buffer */
2163 wp = 0;
2164 bk = 0;
2165 bb = 0;
Eric Andersenb052b471999-11-18 00:21:59 +00002166
2167
Erik Andersene49d5ec2000-02-08 19:58:47 +00002168 /* decompress until the last block */
2169 h = 0;
2170 do {
2171 hufts = 0;
2172 if ((r = inflate_block(&e)) != 0)
2173 return r;
2174 if (hufts > h)
2175 h = hufts;
2176 } while (!e);
Eric Andersenb052b471999-11-18 00:21:59 +00002177
Erik Andersene49d5ec2000-02-08 19:58:47 +00002178 /* Undo too much lookahead. The next read will be byte aligned so we
2179 * can discard unused bits in the last meaningful byte.
2180 */
2181 while (bk >= 8) {
2182 bk -= 8;
2183 inptr--;
2184 }
Eric Andersenb052b471999-11-18 00:21:59 +00002185
Erik Andersene49d5ec2000-02-08 19:58:47 +00002186 /* flush out slide */
2187 flush_output(wp);
Eric Andersenb052b471999-11-18 00:21:59 +00002188
2189
Erik Andersene49d5ec2000-02-08 19:58:47 +00002190 /* return success */
Eric Andersenb052b471999-11-18 00:21:59 +00002191#ifdef DEBUG
Erik Andersene49d5ec2000-02-08 19:58:47 +00002192 fprintf(stderr, "<%u> ", h);
2193#endif /* DEBUG */
2194 return 0;
Eric Andersenb052b471999-11-18 00:21:59 +00002195}