1 /* vi: set sw=4 ts=4: */
3 * Gzip implementation for busybox
5 * Based on GNU gzip Copyright (C) 1992-1993 Jean-loup Gailly.
7 * Originally adjusted for busybox by Sven Rudolph <sr1@inf.tu-dresden.de>
8 * based on gzip sources
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.
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.
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.
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
33 static const char gunzip_usage[] =
34 "gunzip [OPTION]... FILE\n"
35 #ifndef BB_FEATURE_TRIVIAL_HELP
36 "\nUncompress FILE (or standard input if FILE is '-').\n\n"
39 "\t-c\tWrite output to standard output\n"
40 "\t-t\tTest compressed file integrity\n"
45 /* These defines are very important for BusyBox. Without these,
46 * huge chunks of ram are pre-allocated making the BusyBox bss
47 * size Freaking Huge(tm), which is a bad thing.*/
51 #define BB_DECLARE_EXTERN
52 #define bb_need_memory_exhausted
53 #define bb_need_name_too_long
57 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
58 * Copyright (C) 1992-1993 Jean-loup Gailly
59 * The unzip code was written and put in the public domain by Mark Adler.
60 * Portions of the lzw code are derived from the public domain 'compress'
61 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
62 * Ken Turkowski, Dave Mack and Peter Jannesen.
64 * See the license_msg below and the file COPYING for the software license.
65 * See the file algorithm.doc for the compression algorithms and file formats.
69 static char *license_msg[] = {
70 " Copyright (C) 1992-1993 Jean-loup Gailly",
71 " This program is free software; you can redistribute it and/or modify",
72 " it under the terms of the GNU General Public License as published by",
73 " the Free Software Foundation; either version 2, or (at your option)",
74 " any later version.",
76 " This program is distributed in the hope that it will be useful,",
77 " but WITHOUT ANY WARRANTY; without even the implied warranty of",
78 " MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the",
79 " GNU General Public License for more details.",
81 " You should have received a copy of the GNU General Public License",
82 " along with this program; if not, write to the Free Software",
83 " Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.",
88 /* Compress files with zip algorithm and 'compress' interface.
89 * See usage() and help() functions below for all options.
91 * file.gz: compressed file with same mode, owner, and utimes
92 * or stdout with -c option or if stdin used as input.
93 * If the output file name had to be truncated, the original name is kept
94 * in the compressed file.
95 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
97 * Using gz on MSDOS would create too many file name conflicts. For
98 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
99 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
100 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
101 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
103 * For the meaning of all compilation flags, see comments in Makefile.in.
107 #include <sys/types.h>
109 #include <sys/stat.h>
112 /* #include "tailor.h" */
114 /* tailor.h -- target dependent definitions
115 * Copyright (C) 1992-1993 Jean-loup Gailly.
116 * This is free software; you can redistribute it and/or modify it under the
117 * terms of the GNU General Public License, see the file COPYING.
120 /* The target dependent definitions should be defined here only.
121 * The target dependent functions should be defined in tailor.c.
126 #define get_char() get_byte()
127 #define put_char(c) put_byte(c)
130 /* I don't like nested includes, but the string and io functions are used
134 #if !defined(NO_STRING_H) || defined(STDC_HEADERS)
136 # if !defined(STDC_HEADERS) && !defined(NO_MEMORY_H) && !defined(__GNUC__)
139 # define memzero(s, n) memset ((void *)(s), 0, (n))
141 # include <strings.h>
142 # define strchr index
143 # define strrchr rindex
144 # define memcpy(d, s, n) bcopy((s), (d), (n))
145 # define memcmp(s1, s2, n) bcmp((s1), (s2), (n))
146 # define memzero(s, n) bzero((s), (n))
150 # define RETSIGTYPE void
155 typedef unsigned char uch;
156 typedef unsigned short ush;
157 typedef unsigned long ulg;
159 /* Return codes from gzip */
164 /* Compression methods (see algorithm.doc) */
167 extern int method; /* compression method */
169 /* To save memory for 16 bit systems, some arrays are overlaid between
170 * the various modules:
171 * deflate: prev+head window d_buf l_buf outbuf
172 * unlzw: tab_prefix tab_suffix stack inbuf outbuf
173 * inflate: window inbuf
174 * unpack: window inbuf prefix_len
175 * unlzh: left+right window c_table inbuf c_len
176 * For compression, input is done in window[]. For decompression, output
177 * is done in window except for unlzw.
182 # define INBUFSIZ 0x2000 /* input buffer size */
184 # define INBUFSIZ 0x8000 /* input buffer size */
187 #define INBUF_EXTRA 64 /* required by unlzw() */
191 # define OUTBUFSIZ 8192 /* output buffer size */
193 # define OUTBUFSIZ 16384 /* output buffer size */
196 #define OUTBUF_EXTRA 2048 /* required by unlzw() */
202 # define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
204 # define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
208 /*#define DYN_ALLOC*/
211 # define EXTERN(type, array) extern type * array
212 # define DECLARE(type, array, size) type * array
213 # define ALLOC(type, array, size) { \
214 array = (type*)calloc((size_t)(((size)+1L)/2), 2*sizeof(type)); \
215 if (array == NULL) errorMsg(memory_exhausted, "gunzip"); \
217 # define FREE(array) {if (array != NULL) free(array), array=NULL;}
219 # define EXTERN(type, array) extern type array[]
220 # define DECLARE(type, array, size) type array[size]
221 # define ALLOC(type, array, size)
225 EXTERN(uch, inbuf); /* input buffer */
226 EXTERN(uch, outbuf); /* output buffer */
227 EXTERN(ush, d_buf); /* buffer for distances, see trees.c */
228 EXTERN(uch, window); /* Sliding window and suffix table (unlzw) */
229 #define tab_suffix window
231 # define tab_prefix prev /* hash link (see deflate.c) */
232 # define head (prev+WSIZE) /* hash head (see deflate.c) */
233 EXTERN(ush, tab_prefix); /* prefix code (see unlzw.c) */
235 # define tab_prefix0 prev
236 # define head tab_prefix1
237 EXTERN(ush, tab_prefix0); /* prefix for even codes */
238 EXTERN(ush, tab_prefix1); /* prefix for odd codes */
241 extern unsigned insize; /* valid bytes in inbuf */
242 extern unsigned inptr; /* index of next byte to be processed in inbuf */
243 extern unsigned outcnt; /* bytes in output buffer */
245 extern long bytes_in; /* number of input bytes */
246 extern long bytes_out; /* number of output bytes */
247 extern long header_bytes; /* number of bytes in gzip header */
249 extern long ifile_size; /* input file size, -1 for devices (debug only) */
251 typedef int file_t; /* Do not use stdio */
253 #define NO_FILE (-1) /* in memory compression */
256 #define GZIP_MAGIC "\037\213" /* Magic header for gzip files, 1F 8B */
259 #define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
260 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
261 #define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
262 #define ORIG_NAME 0x08 /* bit 3 set: original file name present */
263 #define COMMENT 0x10 /* bit 4 set: file comment present */
264 #define ENCRYPTED 0x20 /* bit 5 set: file is encrypted */
265 #define RESERVED 0xC0 /* bit 6,7: reserved */
268 # define WSIZE 0x8000 /* window size--must be a power of two, and */
269 #endif /* at least 32K for zip's deflate method */
272 #define MAX_MATCH 258
273 /* The minimum and maximum match lengths */
275 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
276 /* Minimum amount of lookahead, except at the end of the input file.
277 * See deflate.c for comments about the MIN_MATCH+1.
280 #define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
281 /* In order to simplify the code, particularly on 16 bit machines, match
282 * distances are limited to MAX_DIST instead of WSIZE.
285 extern int exit_code; /* program exit code */
286 extern int verbose; /* be verbose (-v) */
287 extern int level; /* compression level */
288 extern int test; /* check .z file integrity */
289 extern int save_orig_name; /* set if original name must be saved */
291 #define get_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(0))
292 #define try_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(1))
294 /* put_byte is used for the compressed output, put_ubyte for the
295 * uncompressed output. However unlzw() uses window for its
296 * suffix table instead of its output buffer, so it does not use put_ubyte
297 * (to be cleaned up).
299 #define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
301 #define put_ubyte(c) {window[outcnt++]=(uch)(c); if (outcnt==WSIZE)\
304 /* Output a 16 bit value, lsb first */
305 #define put_short(w) \
306 { if (outcnt < OUTBUFSIZ-2) { \
307 outbuf[outcnt++] = (uch) ((w) & 0xff); \
308 outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
310 put_byte((uch)((w) & 0xff)); \
311 put_byte((uch)((ush)(w) >> 8)); \
315 /* Output a 32 bit value to the bit stream, lsb first */
316 #define put_long(n) { \
317 put_short((n) & 0xffff); \
318 put_short(((ulg)(n)) >> 16); \
321 #define seekable() 0 /* force sequential output */
322 #define translate_eol 0 /* no option -a yet */
324 #define tolow(c) (isupper(c) ? (c)-'A'+'a' : (c)) /* force to lower case */
326 /* Macros for getting two-byte and four-byte header values */
327 #define SH(p) ((ush)(uch)((p)[0]) | ((ush)(uch)((p)[1]) << 8))
328 #define LG(p) ((ulg)(SH(p)) | ((ulg)(SH((p)+2)) << 16))
330 /* Diagnostic functions */
332 # define Assert(cond,msg) {if(!(cond)) errorMsg(msg);}
333 # define Trace(x) fprintf x
334 # define Tracev(x) {if (verbose) fprintf x ;}
335 # define Tracevv(x) {if (verbose>1) fprintf x ;}
336 # define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
337 # define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
339 # define Assert(cond,msg)
344 # define Tracecv(c,x)
347 #define WARN(msg) {fprintf msg ; \
348 if (exit_code == OK) exit_code = WARNING;}
351 extern int unzip (int in, int out);
354 RETSIGTYPE abort_gzip (void);
357 void lm_init (int pack_level, ush * flags);
361 void ct_init (ush * attr, int *method);
362 int ct_tally (int dist, int lc);
363 ulg flush_block (char *buf, ulg stored_len, int eof);
366 void bi_init (file_t zipfile);
367 void send_bits (int value, int length);
368 unsigned bi_reverse (unsigned value, int length);
369 void bi_windup (void);
370 void copy_block (char *buf, unsigned len, int header);
373 extern ulg updcrc (uch * s, unsigned n);
374 extern void clear_bufs (void);
375 static int fill_inbuf (int eof_ok);
376 extern void flush_outbuf (void);
377 static void flush_window (void);
378 extern void write_buf (int fd, void * buf, unsigned cnt);
381 static char *basename (char *fname);
382 #endif /* not __linux__ */
383 void read_error_msg (void);
384 void write_error_msg (void);
387 static int inflate (void);
389 /* #include "lzw.h" */
391 /* lzw.h -- define the lzw functions.
392 * Copyright (C) 1992-1993 Jean-loup Gailly.
393 * This is free software; you can redistribute it and/or modify it under the
394 * terms of the GNU General Public License, see the file COPYING.
397 #if !defined(OF) && defined(lint)
404 #define INIT_BITS 9 /* Initial number of bits per code */
406 #define LZW_MAGIC "\037\235" /* Magic header for lzw files, 1F 9D */
408 #define BIT_MASK 0x1f /* Mask for 'number of compression bits' */
409 /* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
410 * It's a pity that old uncompress does not check bit 0x20. That makes
411 * extension of the format actually undesirable because old compress
412 * would just crash on the new format instead of giving a meaningful
413 * error message. It does check the number of bits, but it's more
414 * helpful to say "unsupported format, get a new version" than
415 * "can only handle 16 bits".
418 #define BLOCK_MODE 0x80
419 /* Block compression: if table is full and compression rate is dropping,
420 * clear the dictionary.
423 #define LZW_RESERVED 0x60 /* reserved bits */
425 #define CLEAR 256 /* flush the dictionary */
426 #define FIRST (CLEAR+1) /* first free entry */
428 extern int maxbits; /* max bits per code for LZW */
429 extern int block_mode; /* block compress mode -C compatible with 2.0 */
431 extern int lzw (int in, int out);
432 extern int unlzw (int in, int out);
435 /* #include "revision.h" */
437 /* revision.h -- define the version number
438 * Copyright (C) 1992-1993 Jean-loup Gailly.
439 * This is free software; you can redistribute it and/or modify it under the
440 * terms of the GNU General Public License, see the file COPYING.
443 #define VERSION "1.2.4"
445 #define REVDATE "18 Aug 93"
447 /* This version does not support compression into old compress format: */
458 typedef struct dirent dir_type;
460 # define NLENGTH(dirent) ((int)strlen((dirent)->d_name))
461 # define DIR_OPT "DIRENT"
463 # define NLENGTH(dirent) ((dirent)->d_namlen)
465 # include <sys/dir.h>
466 typedef struct direct dir_type;
468 # define DIR_OPT "SYSDIR"
471 # include <sys/ndir.h>
472 typedef struct direct dir_type;
474 # define DIR_OPT "SYSNDIR"
478 typedef struct direct dir_type;
480 # define DIR_OPT "NDIR"
483 # define DIR_OPT "NO_DIR"
488 #if !defined(S_ISDIR) && defined(S_IFDIR)
489 # define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR)
491 #if !defined(S_ISREG) && defined(S_IFREG)
492 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
494 typedef RETSIGTYPE(*sig_type) (int);
497 # define O_BINARY 0 /* creation mode for open() */
501 /* Pure BSD system? */
502 # include <sys/file.h>
504 # define O_CREAT FCREAT
507 # define O_EXCL FEXCL
512 # define S_IRUSR 0400
515 # define S_IWUSR 0200
517 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
519 #ifndef MAX_PATH_LEN /* max pathname length */
521 # define MAX_PATH_LEN BUFSIZ
523 # define MAX_PATH_LEN 1024
533 off_t lseek (int fd, off_t offset, int whence);
539 DECLARE(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
540 DECLARE(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
541 DECLARE(ush, d_buf, DIST_BUFSIZE);
542 DECLARE(uch, window, 2L * WSIZE);
544 DECLARE(ush, tab_prefix, 1L << BITS);
546 DECLARE(ush, tab_prefix0, 1L << (BITS - 1));
547 DECLARE(ush, tab_prefix1, 1L << (BITS - 1));
550 /* local variables */
552 int test_mode = 0; /* check file integrity option */
553 int foreground; /* set if program run in foreground */
554 int maxbits = BITS; /* max bits per code for LZW */
555 int method = DEFLATED; /* compression method */
556 int exit_code = OK; /* program exit code */
557 int last_member; /* set for .zip and .Z files */
558 int part_nb; /* number of parts in .gz file */
559 long ifile_size; /* input file size, -1 for devices (debug only) */
561 long bytes_in; /* number of input bytes */
562 long bytes_out; /* number of output bytes */
563 long total_in = 0; /* input bytes for all files */
564 long total_out = 0; /* output bytes for all files */
565 struct stat istat; /* status for input file */
566 int ifd; /* input file descriptor */
567 int ofd; /* output file descriptor */
568 unsigned insize; /* valid bytes in inbuf */
569 unsigned inptr; /* index of next byte to be processed in inbuf */
570 unsigned outcnt; /* bytes in output buffer */
572 long header_bytes; /* number of bytes in gzip header */
574 /* local functions */
576 local int get_method (int in);
578 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
580 /* ======================================================================== */
581 int gunzip_main(int argc, char **argv)
583 int file_count; /* number of files to precess */
589 int delInputFile = 0;
592 char ifname[MAX_PATH_LEN + 1]; /* input file name */
593 char ofname[MAX_PATH_LEN + 1]; /* output file name */
595 if (strcmp(applet_name, "zcat") == 0) {
602 /* Parse any options */
603 while (--argc > 0 && **(++argv) == '-') {
604 if (*((*argv) + 1) == '\0') {
608 while (*(++(*argv))) {
623 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
625 (void) signal(SIGINT, (sig_type) abort_gzip);
628 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
629 (void) signal(SIGTERM, (sig_type) abort_gzip);
633 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
634 (void) signal(SIGHUP, (sig_type) abort_gzip);
638 file_count = argc - optind;
640 /* Allocate all global buffers (for DYN_ALLOC option) */
641 ALLOC(uch, inbuf, INBUFSIZ + INBUF_EXTRA);
642 ALLOC(uch, outbuf, OUTBUFSIZ + OUTBUF_EXTRA);
643 ALLOC(ush, d_buf, DIST_BUFSIZE);
644 ALLOC(uch, window, 2L * WSIZE);
646 ALLOC(ush, tab_prefix, 1L << BITS);
648 ALLOC(ush, tab_prefix0, 1L << (BITS - 1));
649 ALLOC(ush, tab_prefix1, 1L << (BITS - 1));
652 if (fromstdin == 1) {
653 strcpy(ofname, "stdin");
655 inFileNum = fileno(stdin);
656 ifile_size = -1L; /* convention for unknown size */
658 /* Open up the input file */
661 if (strlen(*argv) > MAX_PATH_LEN) {
662 fprintf(stderr, name_too_long, "gunzip");
665 strcpy(ifname, *argv);
667 /* Open input fille */
668 inFileNum = open(ifname, O_RDONLY);
673 /* Get the time stamp on the input file. */
674 result = stat(ifname, &statBuf);
679 ifile_size = statBuf.st_size;
682 if (to_stdout == 1) {
683 /* And get to work */
684 strcpy(ofname, "stdout");
685 outFileNum = fileno(stdout);
687 clear_bufs(); /* clear input and output buffers */
690 /* Actually do the compression/decompression. */
691 unzip(inFileNum, outFileNum);
693 } else if (test_mode) {
694 /* Actually do the compression/decompression. */
699 /* And get to work */
700 if (strlen(ifname) > MAX_PATH_LEN - 4) {
701 fprintf(stderr, name_too_long, "gunzip");
704 strcpy(ofname, ifname);
705 pos = strstr(ofname, ".gz");
710 pos = strstr(ofname, ".tgz");
718 /* Open output fille */
719 #if (__GLIBC__ >= 2) && (__GLIBC_MINOR__ >= 1)
720 outFileNum = open(ofname, O_RDWR | O_CREAT | O_EXCL | O_NOFOLLOW);
722 outFileNum = open(ofname, O_RDWR | O_CREAT | O_EXCL);
724 if (outFileNum < 0) {
728 /* Set permissions on the file */
729 fchmod(outFileNum, statBuf.st_mode);
731 clear_bufs(); /* clear input and output buffers */
734 /* Actually do the compression/decompression. */
735 result = unzip(inFileNum, outFileNum);
739 /* Delete the original file */
741 delFileName = ifname;
743 delFileName = ofname;
745 if (delInputFile == 1 && unlink(delFileName) < 0) {
754 /* ========================================================================
755 * Check the magic number of the input file and update ofname if an
756 * original name was given and to_stdout is not set.
757 * Return the compression method, -1 for error, -2 for warning.
758 * Set inptr to the offset of the next byte to be processed.
759 * Updates time_stamp if there is one and --no-time is not used.
760 * This function may be called repeatedly for an input file consisting
761 * of several contiguous gzip'ed members.
762 * IN assertions: there is at least one remaining compressed member.
763 * If the member is a zip file, it must be the only one.
765 local int get_method(in)
766 int in; /* input file descriptor */
768 uch flags; /* compression flags */
769 char magic[2]; /* magic header */
771 magic[0] = (char) get_byte();
772 magic[1] = (char) get_byte();
773 method = -1; /* unknown yet */
774 part_nb++; /* number of parts in gzip file */
776 last_member = RECORD_IO;
777 /* assume multiple members in gzip file except for record oriented I/O */
779 if (memcmp(magic, GZIP_MAGIC, 2) == 0) {
781 method = (int) get_byte();
782 if (method != DEFLATED) {
784 "unknown method %d -- get newer version of gzip\n",
789 flags = (uch) get_byte();
791 (ulg) get_byte(); /* Ignore time stamp */
796 (void) get_byte(); /* Ignore extra flags for the moment */
797 (void) get_byte(); /* Ignore OS type for the moment */
799 if ((flags & EXTRA_FIELD) != 0) {
800 unsigned len = (unsigned) get_byte();
802 len |= ((unsigned) get_byte()) << 8;
808 /* Discard original name if any */
809 if ((flags & ORIG_NAME) != 0) {
810 while (get_char() != 0) /* null */
814 /* Discard file comment if any */
815 if ((flags & COMMENT) != 0) {
816 while (get_char() != 0) /* null */
820 header_bytes = inptr + 2 * sizeof(long); /* include crc and size */
829 fprintf(stderr, "\nnot in gzip format\n");
833 WARN((stderr, "\ndecompression OK, trailing garbage ignored\n"));
838 /* ========================================================================
839 * Signal and error handler.
841 RETSIGTYPE abort_gzip()
846 /* unzip.c -- decompress files in gzip or pkzip format.
847 * Copyright (C) 1992-1993 Jean-loup Gailly
848 * This is free software; you can redistribute it and/or modify it under the
849 * terms of the GNU General Public License, see the file COPYING.
851 * The code in this file is derived from the file funzip.c written
852 * and put in the public domain by Mark Adler.
856 This version can extract files in gzip or pkzip format.
857 For the latter, only the first entry is extracted, and it has to be
858 either deflated or stored.
861 /* #include "crypt.h" */
863 /* crypt.h (dummy version) -- do not perform encryption
864 * Hardly worth copyrighting :-)
868 # undef CRYPT /* dummy version */
871 #define RAND_HEAD_LEN 12 /* length of encryption random header */
876 /* PKZIP header definitions */
877 #define LOCSIG 0x04034b50L /* four-byte lead-in (lsb first) */
878 #define LOCFLG 6 /* offset of bit flag */
879 #define CRPFLG 1 /* bit for encrypted entry */
880 #define EXTFLG 8 /* bit for extended local header */
881 #define LOCHOW 8 /* offset of compression method */
882 #define LOCTIM 10 /* file mod time (for decryption) */
883 #define LOCCRC 14 /* offset of crc */
884 #define LOCSIZ 18 /* offset of compressed size */
885 #define LOCLEN 22 /* offset of uncompressed length */
886 #define LOCFIL 26 /* offset of file name field length */
887 #define LOCEXT 28 /* offset of extra field length */
888 #define LOCHDR 30 /* size of local header, including sig */
889 #define EXTHDR 16 /* size of extended local header, inc sig */
894 char *key; /* not used--needed to link crypt.c */
895 int pkzip = 0; /* set for a pkzip file */
896 int ext_header = 0; /* set if extended local header */
898 /* ===========================================================================
899 * Unzip in to out. This routine works on both gzip and pkzip files.
901 * IN assertions: the buffer inbuf contains already the beginning of
902 * the compressed data, from offsets inptr to insize-1 included.
903 * The magic header has already been checked. The output buffer is cleared.
906 int in, out; /* input and output file descriptors */
908 ulg orig_crc = 0; /* original crc */
909 ulg orig_len = 0; /* original uncompressed length */
911 uch buf[EXTHDR]; /* extended local header */
915 method = get_method(ifd);
917 exit(exit_code); /* error message already emitted */
920 updcrc(NULL, 0); /* initialize crc */
922 if (pkzip && !ext_header) { /* crc and length at the end otherwise */
923 orig_crc = LG(inbuf + LOCCRC);
924 orig_len = LG(inbuf + LOCLEN);
928 if (method == DEFLATED) {
933 errorMsg(memory_exhausted, "gunzip");
934 } else if (res != 0) {
935 errorMsg("invalid compressed data--format violated");
939 errorMsg("internal error, invalid method");
942 /* Get the crc and original length */
944 /* crc32 (see algorithm.doc)
945 * uncompressed input size modulo 2^32
947 for (n = 0; n < 8; n++) {
948 buf[n] = (uch) get_byte(); /* may cause an error if EOF */
951 orig_len = LG(buf + 4);
953 } else if (ext_header) { /* If extended header, check it */
954 /* signature - 4bytes: 0x50 0x4b 0x07 0x08
956 * compressed size 4-bytes
957 * uncompressed size 4-bytes
959 for (n = 0; n < EXTHDR; n++) {
960 buf[n] = (uch) get_byte(); /* may cause an error if EOF */
962 orig_crc = LG(buf + 4);
963 orig_len = LG(buf + 12);
966 /* Validate decompression */
967 if (orig_crc != updcrc(outbuf, 0)) {
968 errorMsg("invalid compressed data--crc error");
970 if (orig_len != (ulg) bytes_out) {
971 errorMsg("invalid compressed data--length error");
974 /* Check if there are more entries in a pkzip file */
975 if (pkzip && inptr + 4 < insize && LG(inbuf + inptr) == LOCSIG) {
976 WARN((stderr, "has more than one entry--rest ignored\n"));
978 ext_header = pkzip = 0; /* for next file */
982 /* util.c -- utility functions for gzip support
983 * Copyright (C) 1992-1993 Jean-loup Gailly
984 * This is free software; you can redistribute it and/or modify it under the
985 * terms of the GNU General Public License, see the file COPYING.
990 #include <sys/types.h>
999 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1000 # include <stdlib.h>
1005 static const ulg crc_32_tab[]; /* crc table, defined below */
1007 /* ===========================================================================
1008 * Run a set of bytes through the crc shift register. If s is a NULL
1009 * pointer, then initialize the crc shift register contents instead.
1010 * Return the current crc in either case.
1013 uch *s; /* pointer to bytes to pump through */
1014 unsigned n; /* number of bytes in s[] */
1016 register ulg c; /* temporary variable */
1018 static ulg crc = (ulg) 0xffffffffL; /* shift register contents */
1026 c = crc_32_tab[((int) c ^ (*s++)) & 0xff] ^ (c >> 8);
1030 return c ^ 0xffffffffL; /* (instead of ~c for 64-bit machines) */
1033 /* ===========================================================================
1034 * Clear input and output buffers
1036 void clear_bufs(void)
1040 bytes_in = bytes_out = 0L;
1043 /* ===========================================================================
1044 * Fill the input buffer. This is called only when the buffer is empty.
1046 int fill_inbuf(eof_ok)
1047 int eof_ok; /* set if EOF acceptable as a result */
1051 /* Read as much as possible */
1055 len = read(ifd, (char *) inbuf + insize, INBUFSIZ - insize);
1056 if (len == 0 || len == EOF)
1059 } while (insize < INBUFSIZ);
1066 bytes_in += (ulg) insize;
1071 /* ===========================================================================
1072 * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
1073 * (used for the compressed data only)
1081 write_buf(ofd, (char *) outbuf, outcnt);
1082 bytes_out += (ulg) outcnt;
1086 /* ===========================================================================
1087 * Write the output window window[0..outcnt-1] and update crc and bytes_out.
1088 * (Used for the decompressed data only.)
1094 updcrc(window, outcnt);
1097 write_buf(ofd, (char *) window, outcnt);
1098 bytes_out += (ulg) outcnt;
1102 /* ===========================================================================
1103 * Does the same as write(), but also handles partial pipe writes and checks
1106 void write_buf(fd, buf, cnt)
1113 while ((n = write(fd, buf, cnt)) != cnt) {
1114 if (n == (unsigned) (-1)) {
1118 buf = (void *) ((char *) buf + n);
1122 #if defined(NO_STRING_H) && !defined(STDC_HEADERS)
1124 /* Provide missing strspn and strcspn functions. */
1130 int strspn (const char *s, const char *accept);
1131 int strcspn (const char *s, const char *reject);
1133 /* ========================================================================
1134 * Return the length of the maximum initial segment
1135 * of s which contains only characters in accept.
1137 int strspn(s, accept)
1141 register const char *p;
1142 register const char *a;
1143 register int count = 0;
1145 for (p = s; *p != '\0'; ++p) {
1146 for (a = accept; *a != '\0'; ++a) {
1157 /* ========================================================================
1158 * Return the length of the maximum inital segment of s
1159 * which contains no characters from reject.
1161 int strcspn(s, reject)
1165 register int count = 0;
1167 while (*s != '\0') {
1168 if (strchr(reject, *s++) != NULL)
1175 #endif /* NO_STRING_H */
1178 /* ========================================================================
1181 void read_error_msg()
1183 fprintf(stderr, "\n");
1187 fprintf(stderr, "unexpected end of file\n");
1192 void write_error_msg()
1194 fprintf(stderr, "\n");
1200 /* ========================================================================
1201 * Table of CRC-32's of all single-byte values (made by makecrc.c)
1203 static const ulg crc_32_tab[] = {
1204 0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
1205 0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
1206 0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
1207 0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
1208 0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
1209 0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
1210 0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
1211 0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
1212 0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
1213 0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
1214 0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
1215 0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
1216 0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
1217 0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
1218 0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
1219 0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
1220 0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
1221 0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
1222 0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
1223 0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
1224 0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
1225 0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
1226 0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
1227 0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
1228 0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
1229 0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
1230 0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
1231 0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
1232 0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
1233 0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
1234 0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
1235 0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
1236 0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
1237 0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
1238 0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
1239 0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
1240 0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
1241 0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
1242 0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
1243 0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
1244 0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
1245 0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
1246 0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
1247 0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
1248 0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
1249 0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
1250 0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
1251 0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
1252 0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
1253 0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
1254 0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
1258 /* inflate.c -- Not copyrighted 1992 by Mark Adler
1259 version c10p1, 10 January 1993 */
1261 /* You can do whatever you like with this source file, though I would
1262 prefer that if you modify it and redistribute it that you include
1263 comments to that effect with your name and the date. Thank you.
1264 [The history has been moved to the file ChangeLog.]
1268 Inflate deflated (PKZIP's method 8 compressed) data. The compression
1269 method searches for as much of the current string of bytes (up to a
1270 length of 258) in the previous 32K bytes. If it doesn't find any
1271 matches (of at least length 3), it codes the next byte. Otherwise, it
1272 codes the length of the matched string and its distance backwards from
1273 the current position. There is a single Huffman code that codes both
1274 single bytes (called "literals") and match lengths. A second Huffman
1275 code codes the distance information, which follows a length code. Each
1276 length or distance code actually represents a base value and a number
1277 of "extra" (sometimes zero) bits to get to add to the base value. At
1278 the end of each deflated block is a special end-of-block (EOB) literal/
1279 length code. The decoding process is basically: get a literal/length
1280 code; if EOB then done; if a literal, emit the decoded byte; if a
1281 length then get the distance and emit the referred-to bytes from the
1282 sliding window of previously emitted data.
1284 There are (currently) three kinds of inflate blocks: stored, fixed, and
1285 dynamic. The compressor deals with some chunk of data at a time, and
1286 decides which method to use on a chunk-by-chunk basis. A chunk might
1287 typically be 32K or 64K. If the chunk is uncompressible, then the
1288 "stored" method is used. In this case, the bytes are simply stored as
1289 is, eight bits per byte, with none of the above coding. The bytes are
1290 preceded by a count, since there is no longer an EOB code.
1292 If the data is compressible, then either the fixed or dynamic methods
1293 are used. In the dynamic method, the compressed data is preceded by
1294 an encoding of the literal/length and distance Huffman codes that are
1295 to be used to decode this block. The representation is itself Huffman
1296 coded, and so is preceded by a description of that code. These code
1297 descriptions take up a little space, and so for small blocks, there is
1298 a predefined set of codes, called the fixed codes. The fixed method is
1299 used if the block codes up smaller that way (usually for quite small
1300 chunks), otherwise the dynamic method is used. In the latter case, the
1301 codes are customized to the probabilities in the current block, and so
1302 can code it much better than the pre-determined fixed codes.
1304 The Huffman codes themselves are decoded using a mutli-level table
1305 lookup, in order to maximize the speed of decoding plus the speed of
1306 building the decoding tables. See the comments below that precede the
1307 lbits and dbits tuning parameters.
1312 Notes beyond the 1.93a appnote.txt:
1314 1. Distance pointers never point before the beginning of the output
1316 2. Distance pointers can point back across blocks, up to 32k away.
1317 3. There is an implied maximum of 7 bits for the bit length table and
1318 15 bits for the actual data.
1319 4. If only one code exists, then it is encoded using one bit. (Zero
1320 would be more efficient, but perhaps a little confusing.) If two
1321 codes exist, they are coded using one bit each (0 and 1).
1322 5. There is no way of sending zero distance codes--a dummy must be
1323 sent if there are none. (History: a pre 2.0 version of PKZIP would
1324 store blocks with no distance codes, but this was discovered to be
1325 too harsh a criterion.) Valid only for 1.93a. 2.04c does allow
1326 zero distance codes, which is sent as one code of zero bits in
1328 6. There are up to 286 literal/length codes. Code 256 represents the
1329 end-of-block. Note however that the static length tree defines
1330 288 codes just to fill out the Huffman codes. Codes 286 and 287
1331 cannot be used though, since there is no length base or extra bits
1332 defined for them. Similarly, there are up to 30 distance codes.
1333 However, static trees define 32 codes (all 5 bits) to fill out the
1334 Huffman codes, but the last two had better not show up in the data.
1335 7. Unzip can check dynamic Huffman blocks for complete code sets.
1336 The exception is that a single code would not be complete (see #4).
1337 8. The five bits following the block type is really the number of
1338 literal codes sent minus 257.
1339 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
1340 (1+6+6). Therefore, to output three times the length, you output
1341 three codes (1+1+1), whereas to output four times the same length,
1342 you only need two codes (1+3). Hmm.
1343 10. In the tree reconstruction algorithm, Code = Code + Increment
1344 only if BitLength(i) is not zero. (Pretty obvious.)
1345 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19)
1346 12. Note: length code 284 can represent 227-258, but length code 285
1347 really is 258. The last length deserves its own, short code
1348 since it gets used a lot in very redundant files. The length
1349 258 is special since 258 - 3 (the min match length) is 255.
1350 13. The literal/length and distance code bit lengths are read as a
1351 single stream of lengths. It is possible (and advantageous) for
1352 a repeat code (16, 17, or 18) to go across the boundary between
1353 the two sets of lengths.
1356 #include <sys/types.h>
1358 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1359 # include <stdlib.h>
1363 #define slide window
1365 /* Huffman code lookup table entry--this entry is four bytes for machines
1366 that have 16-bit pointers (e.g. PC's in the small or medium model).
1367 Valid extra bits are 0..13. e == 15 is EOB (end of block), e == 16
1368 means that v is a literal, 16 < e < 32 means that v is a pointer to
1369 the next table, which codes e - 16 bits, and lastly e == 99 indicates
1370 an unused code. If a code with e == 99 is looked up, this implies an
1371 error in the data. */
1373 uch e; /* number of extra bits or operation */
1374 uch b; /* number of bits in this code or subcode */
1376 ush n; /* literal, length base, or distance base */
1377 struct huft *t; /* pointer to next level of table */
1382 /* Function prototypes */
1383 int huft_build (unsigned *, unsigned, unsigned, ush *, ush *,
1384 struct huft **, int *);
1385 int huft_free (struct huft *);
1386 int inflate_codes (struct huft *, struct huft *, int, int);
1387 int inflate_stored (void);
1388 int inflate_fixed (void);
1389 int inflate_dynamic (void);
1390 int inflate_block (int *);
1394 /* The inflate algorithm uses a sliding 32K byte window on the uncompressed
1395 stream to find repeated byte strings. This is implemented here as a
1396 circular buffer. The index is updated simply by incrementing and then
1397 and'ing with 0x7fff (32K-1). */
1398 /* It is left to other modules to supply the 32K area. It is assumed
1399 to be usable as if it were declared "uch slide[32768];" or as just
1400 "uch *slide;" and then malloc'ed in the latter case. The definition
1401 must be in unzip.h, included above. */
1402 /* unsigned wp; current position in slide */
1404 #define flush_output(w) (wp=(w),flush_window())
1406 /* Tables for deflate from PKZIP's appnote.txt. */
1407 static unsigned border[] = { /* Order of the bit length code lengths */
1408 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15
1410 static ush cplens[] = { /* Copy lengths for literal codes 257..285 */
1411 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
1412 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0
1415 /* note: see note #13 above about the 258 in this list. */
1416 static ush cplext[] = { /* Extra bits for literal codes 257..285 */
1417 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
1418 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99
1419 }; /* 99==invalid */
1420 static ush cpdist[] = { /* Copy offsets for distance codes 0..29 */
1421 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
1422 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
1423 8193, 12289, 16385, 24577
1425 static ush cpdext[] = { /* Extra bits for distance codes */
1426 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
1427 7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
1433 /* Macros for inflate() bit peeking and grabbing.
1437 x = b & mask_bits[j];
1440 where NEEDBITS makes sure that b has at least j bits in it, and
1441 DUMPBITS removes the bits from b. The macros use the variable k
1442 for the number of bits in b. Normally, b and k are register
1443 variables for speed, and are initialized at the beginning of a
1444 routine that uses these macros from a global bit buffer and count.
1446 If we assume that EOB will be the longest code, then we will never
1447 ask for bits with NEEDBITS that are beyond the end of the stream.
1448 So, NEEDBITS should not read any more bytes than are needed to
1449 meet the request. Then no bytes need to be "returned" to the buffer
1450 at the end of the last block.
1452 However, this assumption is not true for fixed blocks--the EOB code
1453 is 7 bits, but the other literal/length codes can be 8 or 9 bits.
1454 (The EOB code is shorter than other codes because fixed blocks are
1455 generally short. So, while a block always has an EOB, many other
1456 literal/length codes have a significantly lower probability of
1457 showing up at all.) However, by making the first table have a
1458 lookup of seven bits, the EOB code will be found in that first
1459 lookup, and so will not require that too many bits be pulled from
1463 ulg bb; /* bit buffer */
1464 unsigned bk; /* bits in bit buffer */
1468 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
1469 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
1475 # define NEXTBYTE() (cc = get_byte(), zdecode(cc), cc)
1477 # define NEXTBYTE() (uch)get_byte()
1479 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
1480 #define DUMPBITS(n) {b>>=(n);k-=(n);}
1484 Huffman code decoding is performed using a multi-level table lookup.
1485 The fastest way to decode is to simply build a lookup table whose
1486 size is determined by the longest code. However, the time it takes
1487 to build this table can also be a factor if the data being decoded
1488 is not very long. The most common codes are necessarily the
1489 shortest codes, so those codes dominate the decoding time, and hence
1490 the speed. The idea is you can have a shorter table that decodes the
1491 shorter, more probable codes, and then point to subsidiary tables for
1492 the longer codes. The time it costs to decode the longer codes is
1493 then traded against the time it takes to make longer tables.
1495 This results of this trade are in the variables lbits and dbits
1496 below. lbits is the number of bits the first level table for literal/
1497 length codes can decode in one step, and dbits is the same thing for
1498 the distance codes. Subsequent tables are also less than or equal to
1499 those sizes. These values may be adjusted either when all of the
1500 codes are shorter than that, in which case the longest code length in
1501 bits is used, or when the shortest code is *longer* than the requested
1502 table size, in which case the length of the shortest code in bits is
1505 There are two different values for the two tables, since they code a
1506 different number of possibilities each. The literal/length table
1507 codes 286 possible values, or in a flat code, a little over eight
1508 bits. The distance table codes 30 possible values, or a little less
1509 than five bits, flat. The optimum values for speed end up being
1510 about one bit more than those, so lbits is 8+1 and dbits is 5+1.
1511 The optimum values may differ though from machine to machine, and
1512 possibly even between compilers. Your mileage may vary.
1516 int lbits = 9; /* bits in base literal/length lookup table */
1517 int dbits = 6; /* bits in base distance lookup table */
1520 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
1521 #define BMAX 16 /* maximum bit length of any code (16 for explode) */
1522 #define N_MAX 288 /* maximum number of codes in any set */
1525 unsigned hufts; /* track memory usage */
1528 int huft_build(b, n, s, d, e, t, m)
1529 unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
1530 unsigned n; /* number of codes (assumed <= N_MAX) */
1531 unsigned s; /* number of simple-valued codes (0..s-1) */
1532 ush *d; /* list of base values for non-simple codes */
1533 ush *e; /* list of extra bits for non-simple codes */
1534 struct huft **t; /* result: starting table */
1535 int *m; /* maximum lookup bits, returns actual */
1537 /* Given a list of code lengths and a maximum table size, make a set of
1538 tables to decode that set of codes. Return zero on success, one if
1539 the given code set is incomplete (the tables are still built in this
1540 case), two if the input is invalid (all zero length codes or an
1541 oversubscribed set of lengths), and three if not enough memory. */
1543 unsigned a; /* counter for codes of length k */
1544 unsigned c[BMAX + 1]; /* bit length count table */
1545 unsigned f; /* i repeats in table every f entries */
1546 int g; /* maximum code length */
1547 int h; /* table level */
1548 register unsigned i; /* counter, current code */
1549 register unsigned j; /* counter */
1550 register int k; /* number of bits in current code */
1551 int l; /* bits per table (returned in m) */
1552 register unsigned *p; /* pointer into c[], b[], or v[] */
1553 register struct huft *q; /* points to current table */
1554 struct huft r; /* table entry for structure assignment */
1555 struct huft *u[BMAX]; /* table stack */
1556 unsigned v[N_MAX]; /* values in order of bit length */
1557 register int w; /* bits before this table == (l * h) */
1558 unsigned x[BMAX + 1]; /* bit offsets, then code stack */
1559 unsigned *xp; /* pointer into x */
1560 int y; /* number of dummy codes added */
1561 unsigned z; /* number of entries in current table */
1564 /* Generate counts for each bit length */
1565 memzero(c, sizeof(c));
1572 && n - i <= '~' ? "%c %d\n" : "0x%x %d\n"), n - i, *p));
1573 c[*p]++; /* assume all entries <= BMAX */
1574 p++; /* Can't combine with above line (Solaris bug) */
1576 if (c[0] == n) { /* null input--all zero length codes */
1577 *t = (struct huft *) NULL;
1583 /* Find minimum and maximum length, bound *m by those */
1585 for (j = 1; j <= BMAX; j++)
1588 k = j; /* minimum code length */
1589 if ((unsigned) l < j)
1591 for (i = BMAX; i; i--)
1594 g = i; /* maximum code length */
1595 if ((unsigned) l > i)
1600 /* Adjust last length count to fill out codes, if needed */
1601 for (y = 1 << j; j < i; j++, y <<= 1)
1602 if ((y -= c[j]) < 0)
1603 return 2; /* bad input: more codes than bits */
1604 if ((y -= c[i]) < 0)
1609 /* Generate starting offsets into the value table for each length */
1613 while (--i) { /* note that i == g from above */
1614 *xp++ = (j += *p++);
1618 /* Make a table of values in order of bit lengths */
1622 if ((j = *p++) != 0)
1627 /* Generate the Huffman codes and for each, make the table entries */
1628 x[0] = i = 0; /* first Huffman code is zero */
1629 p = v; /* grab values in bit order */
1630 h = -1; /* no tables yet--level -1 */
1631 w = -l; /* bits decoded == (l * h) */
1632 u[0] = (struct huft *) NULL; /* just to keep compilers happy */
1633 q = (struct huft *) NULL; /* ditto */
1636 /* go through the bit lengths (k already is bits in shortest code) */
1637 for (; k <= g; k++) {
1640 /* here i is the Huffman code of length k bits for value *p */
1641 /* make tables up to required level */
1644 w += l; /* previous table always l bits */
1646 /* compute minimum size table less than or equal to l bits */
1647 z = (z = g - w) > (unsigned) l ? l : z; /* upper limit on table size */
1648 if ((f = 1 << (j = k - w)) > a + 1) { /* try a k-w bit table *//* too few codes for k-w bit table */
1649 f -= a + 1; /* deduct codes from patterns left */
1651 while (++j < z) { /* try smaller tables up to z bits */
1652 if ((f <<= 1) <= *++xp)
1653 break; /* enough codes to use up j bits */
1654 f -= *xp; /* else deduct codes from patterns */
1657 z = 1 << j; /* table entries for j-bit table */
1659 /* allocate and link in new table */
1662 (struct huft *) malloc((z + 1) *
1663 sizeof(struct huft))) ==
1664 (struct huft *) NULL) {
1667 return 3; /* not enough memory */
1669 hufts += z + 1; /* track memory usage */
1670 *t = q + 1; /* link to list for huft_free() */
1671 *(t = &(q->v.t)) = (struct huft *) NULL;
1672 u[h] = ++q; /* table starts after link */
1674 /* connect to last table, if there is one */
1676 x[h] = i; /* save pattern for backing up */
1677 r.b = (uch) l; /* bits to dump before this table */
1678 r.e = (uch) (16 + j); /* bits in this table */
1679 r.v.t = q; /* pointer to this table */
1680 j = i >> (w - l); /* (get around Turbo C bug) */
1681 u[h - 1][j] = r; /* connect to last table */
1685 /* set up table entry in r */
1686 r.b = (uch) (k - w);
1688 r.e = 99; /* out of values--invalid code */
1690 r.e = (uch) (*p < 256 ? 16 : 15); /* 256 is end-of-block code */
1691 r.v.n = (ush) (*p); /* simple code is just the value */
1692 p++; /* one compiler does not like *p++ */
1694 r.e = (uch) e[*p - s]; /* non-simple--look up in lists */
1695 r.v.n = d[*p++ - s];
1698 /* fill code-like entries with r */
1700 for (j = i >> w; j < z; j += f)
1703 /* backwards increment the k-bit code i */
1704 for (j = 1 << (k - 1); i & j; j >>= 1)
1708 /* backup over finished tables */
1709 while ((i & ((1 << w) - 1)) != x[h]) {
1710 h--; /* don't need to update q */
1717 /* Return true (1) if we were given an incomplete table */
1718 return y != 0 && g != 1;
1724 struct huft *t; /* table to free */
1726 /* Free the malloc'ed tables built by huft_build(), which makes a linked
1727 list of the tables it made, with the links in a dummy first entry of
1730 register struct huft *p, *q;
1733 /* Go through linked list, freeing from the malloced (t[-1]) address. */
1735 while (p != (struct huft *) NULL) {
1744 int inflate_codes(tl, td, bl, bd)
1745 struct huft *tl, *td; /* literal/length and distance decoder tables */
1746 int bl, bd; /* number of bits decoded by tl[] and td[] */
1748 /* inflate (decompress) the codes in a deflated (compressed) block.
1749 Return an error code or zero if it all goes ok. */
1751 register unsigned e; /* table entry flag/number of extra bits */
1752 unsigned n, d; /* length and index for copy */
1753 unsigned w; /* current window position */
1754 struct huft *t; /* pointer to table entry */
1755 unsigned ml, md; /* masks for bl and bd bits */
1756 register ulg b; /* bit buffer */
1757 register unsigned k; /* number of bits in bit buffer */
1760 /* make local copies of globals */
1761 b = bb; /* initialize bit buffer */
1763 w = wp; /* initialize window position */
1765 /* inflate the coded data */
1766 ml = mask_bits[bl]; /* precompute masks for speed */
1768 for (;;) { /* do until end of block */
1769 NEEDBITS((unsigned) bl)
1770 if ((e = (t = tl + ((unsigned) b & ml))->e) > 16)
1777 } while ((e = (t = t->v.t + ((unsigned) b & mask_bits[e]))->e)
1780 if (e == 16) { /* then it's a literal */
1781 slide[w++] = (uch) t->v.n;
1782 Tracevv((stderr, "%c", slide[w - 1]));
1787 } else { /* it's an EOB or a length */
1789 /* exit if end of block */
1793 /* get length of block to copy */
1795 n = t->v.n + ((unsigned) b & mask_bits[e]);
1798 /* decode distance of block to copy */
1799 NEEDBITS((unsigned) bd)
1800 if ((e = (t = td + ((unsigned) b & md))->e) > 16)
1811 t->v.t + ((unsigned) b & mask_bits[e]))->e) >
1815 d = w - t->v.n - ((unsigned) b & mask_bits[e]);
1817 Tracevv((stderr, "\\[%d,%d]", w - d, n));
1823 WSIZE - ((d &= WSIZE - 1) > w ? d : w)) >
1825 #if !defined(NOMEMCPY) && !defined(DEBUG)
1826 if (w - d >= e) { /* (this test assumes unsigned comparison) */
1827 memcpy(slide + w, slide + d, e);
1830 } else /* do it slow to avoid memcpy() overlap */
1831 #endif /* !NOMEMCPY */
1833 slide[w++] = slide[d++];
1834 Tracevv((stderr, "%c", slide[w - 1]));
1845 /* restore the globals from the locals */
1846 wp = w; /* restore global window pointer */
1847 bb = b; /* restore global bit buffer */
1856 int inflate_stored()
1857 /* "decompress" an inflated type 0 (stored) block. */
1859 unsigned n; /* number of bytes in block */
1860 unsigned w; /* current window position */
1861 register ulg b; /* bit buffer */
1862 register unsigned k; /* number of bits in bit buffer */
1865 /* make local copies of globals */
1866 b = bb; /* initialize bit buffer */
1868 w = wp; /* initialize window position */
1871 /* go to byte boundary */
1876 /* get the length and its complement */
1878 n = ((unsigned) b & 0xffff);
1881 if (n != (unsigned) ((~b) & 0xffff))
1882 return 1; /* error in compressed data */
1886 /* read and output the compressed data */
1889 slide[w++] = (uch) b;
1898 /* restore the globals from the locals */
1899 wp = w; /* restore global window pointer */
1900 bb = b; /* restore global bit buffer */
1908 /* decompress an inflated type 1 (fixed Huffman codes) block. We should
1909 either replace this with a custom decoder, or at least precompute the
1912 int i; /* temporary variable */
1913 struct huft *tl; /* literal/length code table */
1914 struct huft *td; /* distance code table */
1915 int bl; /* lookup bits for tl */
1916 int bd; /* lookup bits for td */
1917 unsigned l[288]; /* length list for huft_build */
1920 /* set up literal table */
1921 for (i = 0; i < 144; i++)
1923 for (; i < 256; i++)
1925 for (; i < 280; i++)
1927 for (; i < 288; i++) /* make a complete, but wrong code set */
1930 if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
1934 /* set up distance table */
1935 for (i = 0; i < 30; i++) /* make an incomplete code set */
1938 if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1) {
1944 /* decompress until an end-of-block code */
1945 if (inflate_codes(tl, td, bl, bd))
1949 /* free the decoding tables, return */
1957 int inflate_dynamic()
1958 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
1960 int i; /* temporary variables */
1962 unsigned l; /* last length */
1963 unsigned m; /* mask for bit lengths table */
1964 unsigned n; /* number of lengths to get */
1965 struct huft *tl; /* literal/length code table */
1966 struct huft *td; /* distance code table */
1967 int bl; /* lookup bits for tl */
1968 int bd; /* lookup bits for td */
1969 unsigned nb; /* number of bit length codes */
1970 unsigned nl; /* number of literal/length codes */
1971 unsigned nd; /* number of distance codes */
1973 #ifdef PKZIP_BUG_WORKAROUND
1974 unsigned ll[288 + 32]; /* literal/length and distance code lengths */
1976 unsigned ll[286 + 30]; /* literal/length and distance code lengths */
1978 register ulg b; /* bit buffer */
1979 register unsigned k; /* number of bits in bit buffer */
1982 /* make local bit buffer */
1987 /* read in table lengths */
1989 nl = 257 + ((unsigned) b & 0x1f); /* number of literal/length codes */
1992 nd = 1 + ((unsigned) b & 0x1f); /* number of distance codes */
1995 nb = 4 + ((unsigned) b & 0xf); /* number of bit length codes */
1997 #ifdef PKZIP_BUG_WORKAROUND
1998 if (nl > 288 || nd > 32)
2000 if (nl > 286 || nd > 30)
2002 return 1; /* bad lengths */
2005 /* read in bit-length-code lengths */
2006 for (j = 0; j < nb; j++) {
2008 ll[border[j]] = (unsigned) b & 7;
2015 /* build decoding table for trees--single level, 7 bit lookup */
2017 if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0) {
2020 return i; /* incomplete code set */
2024 /* read in literal and distance code lengths */
2028 while ((unsigned) i < n) {
2029 NEEDBITS((unsigned) bl)
2030 j = (td = tl + ((unsigned) b & m))->b;
2033 if (j < 16) /* length of code in bits (0..15) */
2034 ll[i++] = l = j; /* save last length in l */
2035 else if (j == 16) { /* repeat last length 3 to 6 times */
2037 j = 3 + ((unsigned) b & 3);
2039 if ((unsigned) i + j > n)
2043 } else if (j == 17) { /* 3 to 10 zero length codes */
2045 j = 3 + ((unsigned) b & 7);
2047 if ((unsigned) i + j > n)
2052 } else { /* j == 18: 11 to 138 zero length codes */
2055 j = 11 + ((unsigned) b & 0x7f);
2057 if ((unsigned) i + j > n)
2066 /* free decoding table for trees */
2070 /* restore the global bit buffer */
2075 /* build the decoding tables for literal/length and distance codes */
2077 if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0) {
2079 fprintf(stderr, " incomplete literal tree\n");
2082 return i; /* incomplete code set */
2085 if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0) {
2087 fprintf(stderr, " incomplete distance tree\n");
2088 #ifdef PKZIP_BUG_WORKAROUND
2095 return i; /* incomplete code set */
2100 /* decompress until an end-of-block code */
2101 if (inflate_codes(tl, td, bl, bd))
2105 /* free the decoding tables, return */
2113 int inflate_block(e)
2114 int *e; /* last block flag */
2116 /* decompress an inflated block */
2118 unsigned t; /* block type */
2119 register ulg b; /* bit buffer */
2120 register unsigned k; /* number of bits in bit buffer */
2123 /* make local bit buffer */
2128 /* read in last block bit */
2134 /* read in block type */
2136 t = (unsigned) b & 3;
2140 /* restore the global bit buffer */
2145 /* inflate that block type */
2147 return inflate_dynamic();
2149 return inflate_stored();
2151 return inflate_fixed();
2154 /* bad block type */
2161 /* decompress an inflated entry */
2163 int e; /* last block flag */
2164 int r; /* result code */
2165 unsigned h; /* maximum struct huft's malloc'ed */
2168 /* initialize window, bit buffer */
2174 /* decompress until the last block */
2178 if ((r = inflate_block(&e)) != 0)
2184 /* Undo too much lookahead. The next read will be byte aligned so we
2185 * can discard unused bits in the last meaningful byte.
2192 /* flush out slide */
2196 /* return success */
2198 fprintf(stderr, "<%u> ", h);