1 /* zcat : stripped version based on gzip sources
2 Sven Rudolph <sr1@inf.tu-dresden.de>
7 static const char gunzip_usage[] =
8 "gunzip [OPTION]... FILE\n\n"
9 "Uncompress FILE (or standard input if FILE is '-').\n\n"
11 "\t-c\tWrite output to standard output\n"
12 "\t-t\tTest compressed file integrity\n";
14 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
15 * Copyright (C) 1992-1993 Jean-loup Gailly
16 * The unzip code was written and put in the public domain by Mark Adler.
17 * Portions of the lzw code are derived from the public domain 'compress'
18 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
19 * Ken Turkowski, Dave Mack and Peter Jannesen.
21 * See the license_msg below and the file COPYING for the software license.
22 * See the file algorithm.doc for the compression algorithms and file formats.
26 static char *license_msg[] = {
27 " Copyright (C) 1992-1993 Jean-loup Gailly",
28 " This program is free software; you can redistribute it and/or modify",
29 " it under the terms of the GNU General Public License as published by",
30 " the Free Software Foundation; either version 2, or (at your option)",
31 " any later version.",
33 " This program is distributed in the hope that it will be useful,",
34 " but WITHOUT ANY WARRANTY; without even the implied warranty of",
35 " MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the",
36 " GNU General Public License for more details.",
38 " You should have received a copy of the GNU General Public License",
39 " along with this program; if not, write to the Free Software",
40 " Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.",
44 /* Compress files with zip algorithm and 'compress' interface.
45 * See usage() and help() functions below for all options.
47 * file.gz: compressed file with same mode, owner, and utimes
48 * or stdout with -c option or if stdin used as input.
49 * If the output file name had to be truncated, the original name is kept
50 * in the compressed file.
51 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
53 * Using gz on MSDOS would create too many file name conflicts. For
54 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
55 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
56 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
57 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
59 * For the meaning of all compilation flags, see comments in Makefile.in.
63 #include <sys/types.h>
68 /* #include "tailor.h" */
70 /* tailor.h -- target dependent definitions
71 * Copyright (C) 1992-1993 Jean-loup Gailly.
72 * This is free software; you can redistribute it and/or modify it under the
73 * terms of the GNU General Public License, see the file COPYING.
76 /* The target dependent definitions should be defined here only.
77 * The target dependent functions should be defined in tailor.c.
82 #define get_char() get_byte()
83 #define put_char(c) put_byte(c)
85 /* #include "gzip.h" */
87 /* gzip.h -- common declarations for all gzip modules
88 * Copyright (C) 1992-1993 Jean-loup Gailly.
89 * This is free software; you can redistribute it and/or modify it under the
90 * terms of the GNU General Public License, see the file COPYING.
93 #if defined(__STDC__) || defined(PROTO)
94 # define OF(args) args
105 /* I don't like nested includes, but the string and io functions are used
109 #if !defined(NO_STRING_H) || defined(STDC_HEADERS)
111 # if !defined(STDC_HEADERS) && !defined(NO_MEMORY_H) && !defined(__GNUC__)
114 # define memzero(s, n) memset ((voidp)(s), 0, (n))
116 # include <strings.h>
117 # define strchr index
118 # define strrchr rindex
119 # define memcpy(d, s, n) bcopy((s), (d), (n))
120 # define memcmp(s1, s2, n) bcmp((s1), (s2), (n))
121 # define memzero(s, n) bzero((s), (n))
125 # define RETSIGTYPE void
130 typedef unsigned char uch;
131 typedef unsigned short ush;
132 typedef unsigned long ulg;
134 /* Return codes from gzip */
139 /* Compression methods (see algorithm.doc) */
142 extern int method; /* compression method */
144 /* To save memory for 16 bit systems, some arrays are overlaid between
145 * the various modules:
146 * deflate: prev+head window d_buf l_buf outbuf
147 * unlzw: tab_prefix tab_suffix stack inbuf outbuf
148 * inflate: window inbuf
149 * unpack: window inbuf prefix_len
150 * unlzh: left+right window c_table inbuf c_len
151 * For compression, input is done in window[]. For decompression, output
152 * is done in window except for unlzw.
157 # define INBUFSIZ 0x2000 /* input buffer size */
159 # define INBUFSIZ 0x8000 /* input buffer size */
162 #define INBUF_EXTRA 64 /* required by unlzw() */
166 # define OUTBUFSIZ 8192 /* output buffer size */
168 # define OUTBUFSIZ 16384 /* output buffer size */
171 #define OUTBUF_EXTRA 2048 /* required by unlzw() */
177 # define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
179 # define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
183 /*#define DYN_ALLOC*/
186 # define EXTERN(type, array) extern type * array
187 # define DECLARE(type, array, size) type * array
188 # define ALLOC(type, array, size) { \
189 array = (type*)calloc((size_t)(((size)+1L)/2), 2*sizeof(type)); \
190 if (array == NULL) error("insufficient memory"); \
192 # define FREE(array) {if (array != NULL) free(array), array=NULL;}
194 # define EXTERN(type, array) extern type array[]
195 # define DECLARE(type, array, size) type array[size]
196 # define ALLOC(type, array, size)
200 EXTERN(uch, inbuf); /* input buffer */
201 EXTERN(uch, outbuf); /* output buffer */
202 EXTERN(ush, d_buf); /* buffer for distances, see trees.c */
203 EXTERN(uch, window); /* Sliding window and suffix table (unlzw) */
204 #define tab_suffix window
206 # define tab_prefix prev /* hash link (see deflate.c) */
207 # define head (prev+WSIZE) /* hash head (see deflate.c) */
208 EXTERN(ush, tab_prefix); /* prefix code (see unlzw.c) */
210 # define tab_prefix0 prev
211 # define head tab_prefix1
212 EXTERN(ush, tab_prefix0); /* prefix for even codes */
213 EXTERN(ush, tab_prefix1); /* prefix for odd codes */
216 extern unsigned insize; /* valid bytes in inbuf */
217 extern unsigned inptr; /* index of next byte to be processed in inbuf */
218 extern unsigned outcnt; /* bytes in output buffer */
220 extern long bytes_in; /* number of input bytes */
221 extern long bytes_out; /* number of output bytes */
222 extern long header_bytes;/* number of bytes in gzip header */
224 extern long ifile_size; /* input file size, -1 for devices (debug only) */
226 typedef int file_t; /* Do not use stdio */
227 #define NO_FILE (-1) /* in memory compression */
230 #define GZIP_MAGIC "\037\213" /* Magic header for gzip files, 1F 8B */
233 #define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
234 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
235 #define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
236 #define ORIG_NAME 0x08 /* bit 3 set: original file name present */
237 #define COMMENT 0x10 /* bit 4 set: file comment present */
238 #define ENCRYPTED 0x20 /* bit 5 set: file is encrypted */
239 #define RESERVED 0xC0 /* bit 6,7: reserved */
242 # define WSIZE 0x8000 /* window size--must be a power of two, and */
243 #endif /* at least 32K for zip's deflate method */
246 #define MAX_MATCH 258
247 /* The minimum and maximum match lengths */
249 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
250 /* Minimum amount of lookahead, except at the end of the input file.
251 * See deflate.c for comments about the MIN_MATCH+1.
254 #define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
255 /* In order to simplify the code, particularly on 16 bit machines, match
256 * distances are limited to MAX_DIST instead of WSIZE.
259 extern int exit_code; /* program exit code */
260 extern int verbose; /* be verbose (-v) */
261 extern int level; /* compression level */
262 extern int test; /* check .z file integrity */
263 extern int save_orig_name; /* set if original name must be saved */
265 #define get_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(0))
266 #define try_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(1))
268 /* put_byte is used for the compressed output, put_ubyte for the
269 * uncompressed output. However unlzw() uses window for its
270 * suffix table instead of its output buffer, so it does not use put_ubyte
271 * (to be cleaned up).
273 #define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
275 #define put_ubyte(c) {window[outcnt++]=(uch)(c); if (outcnt==WSIZE)\
278 /* Output a 16 bit value, lsb first */
279 #define put_short(w) \
280 { if (outcnt < OUTBUFSIZ-2) { \
281 outbuf[outcnt++] = (uch) ((w) & 0xff); \
282 outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
284 put_byte((uch)((w) & 0xff)); \
285 put_byte((uch)((ush)(w) >> 8)); \
289 /* Output a 32 bit value to the bit stream, lsb first */
290 #define put_long(n) { \
291 put_short((n) & 0xffff); \
292 put_short(((ulg)(n)) >> 16); \
295 #define seekable() 0 /* force sequential output */
296 #define translate_eol 0 /* no option -a yet */
298 #define tolow(c) (isupper(c) ? (c)-'A'+'a' : (c)) /* force to lower case */
300 /* Macros for getting two-byte and four-byte header values */
301 #define SH(p) ((ush)(uch)((p)[0]) | ((ush)(uch)((p)[1]) << 8))
302 #define LG(p) ((ulg)(SH(p)) | ((ulg)(SH((p)+2)) << 16))
304 /* Diagnostic functions */
306 # define Assert(cond,msg) {if(!(cond)) error(msg);}
307 # define Trace(x) fprintf x
308 # define Tracev(x) {if (verbose) fprintf x ;}
309 # define Tracevv(x) {if (verbose>1) fprintf x ;}
310 # define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
311 # define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
313 # define Assert(cond,msg)
318 # define Tracecv(c,x)
321 #define WARN(msg) {fprintf msg ; \
322 if (exit_code == OK) exit_code = WARNING;}
324 #define do_exit(c) exit(c)
328 extern int unzip OF((int in, int out));
331 RETSIGTYPE abort_gzip OF((void));
334 void lm_init OF((int pack_level, ush *flags));
335 ulg deflate OF((void));
338 void ct_init OF((ush *attr, int *method));
339 int ct_tally OF((int dist, int lc));
340 ulg flush_block OF((char *buf, ulg stored_len, int eof));
343 void bi_init OF((file_t zipfile));
344 void send_bits OF((int value, int length));
345 unsigned bi_reverse OF((unsigned value, int length));
346 void bi_windup OF((void));
347 void copy_block OF((char *buf, unsigned len, int header));
348 extern int (*read_buf) OF((char *buf, unsigned size));
351 extern int copy OF((int in, int out));
352 extern ulg updcrc OF((uch *s, unsigned n));
353 extern void clear_bufs OF((void));
354 extern int fill_inbuf OF((int eof_ok));
355 extern void flush_outbuf OF((void));
356 extern void flush_window OF((void));
357 extern void write_buf OF((int fd, voidp buf, unsigned cnt));
359 extern char *basename OF((char *fname));
360 #endif /* not __linux__ */
361 extern void error OF((char *m));
362 extern void warn OF((char *a, char *b));
363 extern void read_error OF((void));
364 extern void write_error OF((void));
367 extern int inflate OF((void));
369 /* #include "lzw.h" */
371 /* lzw.h -- define the lzw functions.
372 * Copyright (C) 1992-1993 Jean-loup Gailly.
373 * This is free software; you can redistribute it and/or modify it under the
374 * terms of the GNU General Public License, see the file COPYING.
377 #if !defined(OF) && defined(lint)
384 #define INIT_BITS 9 /* Initial number of bits per code */
386 #define LZW_MAGIC "\037\235" /* Magic header for lzw files, 1F 9D */
388 #define BIT_MASK 0x1f /* Mask for 'number of compression bits' */
389 /* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
390 * It's a pity that old uncompress does not check bit 0x20. That makes
391 * extension of the format actually undesirable because old compress
392 * would just crash on the new format instead of giving a meaningful
393 * error message. It does check the number of bits, but it's more
394 * helpful to say "unsupported format, get a new version" than
395 * "can only handle 16 bits".
398 #define BLOCK_MODE 0x80
399 /* Block compression: if table is full and compression rate is dropping,
400 * clear the dictionary.
403 #define LZW_RESERVED 0x60 /* reserved bits */
405 #define CLEAR 256 /* flush the dictionary */
406 #define FIRST (CLEAR+1) /* first free entry */
408 extern int maxbits; /* max bits per code for LZW */
409 extern int block_mode; /* block compress mode -C compatible with 2.0 */
411 extern int lzw OF((int in, int out));
412 extern int unlzw OF((int in, int out));
415 /* #include "revision.h" */
417 /* revision.h -- define the version number
418 * Copyright (C) 1992-1993 Jean-loup Gailly.
419 * This is free software; you can redistribute it and/or modify it under the
420 * terms of the GNU General Public License, see the file COPYING.
423 #define VERSION "1.2.4"
425 #define REVDATE "18 Aug 93"
427 /* This version does not support compression into old compress format: */
432 /* #include "getopt.h" */
434 /* Declarations for getopt.
435 Copyright (C) 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
437 This program is free software; you can redistribute it and/or modify it
438 under the terms of the GNU General Public License as published by the
439 Free Software Foundation; either version 2, or (at your option) any
442 This program is distributed in the hope that it will be useful,
443 but WITHOUT ANY WARRANTY; without even the implied warranty of
444 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
445 GNU General Public License for more details.
447 You should have received a copy of the GNU General Public License
448 along with this program; if not, write to the Free Software
449 Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
458 /* For communication from `getopt' to the caller.
459 When `getopt' finds an option that takes an argument,
460 the argument value is returned here.
461 Also, when `ordering' is RETURN_IN_ORDER,
462 each non-option ARGV-element is returned here. */
466 /* Index in ARGV of the next element to be scanned.
467 This is used for communication to and from the caller
468 and for communication between successive calls to `getopt'.
470 On entry to `getopt', zero means this is the first call; initialize.
472 When `getopt' returns EOF, this is the index of the first of the
473 non-option elements that the caller should itself scan.
475 Otherwise, `optind' communicates from one call to the next
476 how much of ARGV has been scanned so far. */
480 /* Callers store zero here to inhibit the error message `getopt' prints
481 for unrecognized options. */
485 /* Set to an option character which was unrecognized. */
489 /* Describe the long-named options requested by the application.
490 The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector
491 of `struct option' terminated by an element containing a name which is
494 The field `has_arg' is:
495 no_argument (or 0) if the option does not take an argument,
496 required_argument (or 1) if the option requires an argument,
497 optional_argument (or 2) if the option takes an optional argument.
499 If the field `flag' is not NULL, it points to a variable that is set
500 to the value given in the field `val' when the option is found, but
501 left unchanged if the option is not found.
503 To have a long-named option do something other than set an `int' to
504 a compiled-in constant, such as set a value from `optarg', set the
505 option's `flag' field to zero and its `val' field to a nonzero
506 value (the equivalent single-letter option character, if there is
507 one). For long options that have a zero `flag' field, `getopt'
508 returns the contents of the `val' field. */
517 /* has_arg can't be an enum because some compilers complain about
518 type mismatches in all the code that assumes it is an int. */
524 /* Names for the values of the `has_arg' field of `struct option'. */
526 #define no_argument 0
527 #define required_argument 1
528 #define optional_argument 2
530 #if __STDC__ || defined(PROTO)
531 #if defined(__GNU_LIBRARY__)
532 /* Many other libraries have conflicting prototypes for getopt, with
533 differences in the consts, in stdlib.h. To avoid compilation
534 errors, only prototype getopt for the GNU C library. */
535 extern int getopt (int argc, char *const *argv, const char *shortopts);
536 #endif /* not __GNU_LIBRARY__ */
537 extern int getopt_long (int argc, char *const *argv, const char *shortopts,
538 const struct option *longopts, int *longind);
539 extern int getopt_long_only (int argc, char *const *argv,
540 const char *shortopts,
541 const struct option *longopts, int *longind);
543 /* Internal only. Users should not call this directly. */
544 extern int _getopt_internal (int argc, char *const *argv,
545 const char *shortopts,
546 const struct option *longopts, int *longind,
548 #else /* not __STDC__ */
549 extern int getopt ();
550 extern int getopt_long ();
551 extern int getopt_long_only ();
553 extern int _getopt_internal ();
554 #endif /* not __STDC__ */
560 #endif /* _GETOPT_H */
571 typedef struct dirent dir_type;
572 # define NLENGTH(dirent) ((int)strlen((dirent)->d_name))
573 # define DIR_OPT "DIRENT"
575 # define NLENGTH(dirent) ((dirent)->d_namlen)
577 # include <sys/dir.h>
578 typedef struct direct dir_type;
579 # define DIR_OPT "SYSDIR"
582 # include <sys/ndir.h>
583 typedef struct direct dir_type;
584 # define DIR_OPT "SYSNDIR"
588 typedef struct direct dir_type;
589 # define DIR_OPT "NDIR"
592 # define DIR_OPT "NO_DIR"
598 #if !defined(S_ISDIR) && defined(S_IFDIR)
599 # define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR)
601 #if !defined(S_ISREG) && defined(S_IFREG)
602 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
605 typedef RETSIGTYPE (*sig_type) OF((int));
608 # define O_BINARY 0 /* creation mode for open() */
612 /* Pure BSD system? */
613 # include <sys/file.h>
615 # define O_CREAT FCREAT
618 # define O_EXCL FEXCL
623 # define S_IRUSR 0400
626 # define S_IWUSR 0200
628 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
631 # define MAX_PATH_LEN 1024 /* max pathname length */
640 off_t lseek OF((int fd, off_t offset, int whence));
646 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
647 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
648 DECLARE(ush, d_buf, DIST_BUFSIZE);
649 DECLARE(uch, window, 2L*WSIZE);
651 DECLARE(ush, tab_prefix, 1L<<BITS);
653 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
654 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
657 /* local variables */
659 int test_mode = 0; /* check file integrity option */
660 int foreground; /* set if program run in foreground */
661 int maxbits = BITS; /* max bits per code for LZW */
662 int method = DEFLATED;/* compression method */
663 int exit_code = OK; /* program exit code */
664 int last_member; /* set for .zip and .Z files */
665 int part_nb; /* number of parts in .gz file */
666 long ifile_size; /* input file size, -1 for devices (debug only) */
668 long bytes_in; /* number of input bytes */
669 long bytes_out; /* number of output bytes */
670 long total_in = 0; /* input bytes for all files */
671 long total_out = 0; /* output bytes for all files */
672 struct stat istat; /* status for input file */
673 int ifd; /* input file descriptor */
674 int ofd; /* output file descriptor */
675 unsigned insize; /* valid bytes in inbuf */
676 unsigned inptr; /* index of next byte to be processed in inbuf */
677 unsigned outcnt; /* bytes in output buffer */
679 long header_bytes; /* number of bytes in gzip header */
681 /* local functions */
683 local int get_method OF((int in));
685 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
687 /* ======================================================================== */
688 int gunzip_main (int argc, char** argv)
690 int file_count; /* number of files to precess */
699 char ifname[MAX_PATH_LEN]; /* input file name */
700 char ofname[MAX_PATH_LEN]; /* output file name */
705 if (strcmp(*argv, "zcat")==0)
708 /* Parse any options */
709 while (--argc > 0 && **(++argv) == '-') {
710 if (*((*argv)+1) == '\0') {
714 while (*(++(*argv))) {
729 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
731 (void) signal (SIGINT, (sig_type)abort_gzip);
734 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
735 (void) signal(SIGTERM, (sig_type)abort_gzip);
739 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
740 (void) signal(SIGHUP, (sig_type)abort_gzip);
744 file_count = argc - optind;
746 /* Allocate all global buffers (for DYN_ALLOC option) */
747 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
748 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
749 ALLOC(ush, d_buf, DIST_BUFSIZE);
750 ALLOC(uch, window, 2L*WSIZE);
752 ALLOC(ush, tab_prefix, 1L<<BITS);
754 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
755 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
759 strcpy(ofname, "stdin");
761 inFileNum=fileno(stdin);
762 ifile_size = -1L; /* convention for unknown size */
764 /* Open up the input file */
767 strncpy(ifname, *argv, MAX_PATH_LEN);
769 /* Open input fille */
770 inFileNum=open( ifname, O_RDONLY);
775 /* Get the time stamp on the input file. */
776 result = stat(ifname, &statBuf);
781 ifile_size = statBuf.st_size;
785 /* And get to work */
786 strcpy(ofname, "stdout");
787 outFileNum=fileno(stdout);
789 clear_bufs(); /* clear input and output buffers */
792 /* Actually do the compression/decompression. */
793 unzip(inFileNum, outFileNum);
795 } else if (test_mode) {
796 /* Actually do the compression/decompression. */
801 /* And get to work */
802 strncpy(ofname, ifname, MAX_PATH_LEN-4);
803 pos=strstr(ofname, ".gz");
808 pos=strstr(ofname, ".tgz");
811 strcat( pos, ".tar");
816 /* Open output fille */
817 #if (__GLIBC__ >= 2) && (__GLIBC_MINOR__ >= 1)
818 outFileNum=open( ofname, O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW);
820 outFileNum=open( ofname, O_RDWR|O_CREAT|O_EXCL);
822 if (outFileNum < 0) {
826 /* Set permissions on the file */
827 fchmod(outFileNum, statBuf.st_mode);
829 clear_bufs(); /* clear input and output buffers */
832 /* Actually do the compression/decompression. */
833 result=unzip(inFileNum, outFileNum);
837 /* Delete the original file */
843 if (delInputFile == 1 && unlink (delFileName) < 0) {
844 perror (delFileName);
852 /* ========================================================================
853 * Check the magic number of the input file and update ofname if an
854 * original name was given and to_stdout is not set.
855 * Return the compression method, -1 for error, -2 for warning.
856 * Set inptr to the offset of the next byte to be processed.
857 * Updates time_stamp if there is one and --no-time is not used.
858 * This function may be called repeatedly for an input file consisting
859 * of several contiguous gzip'ed members.
860 * IN assertions: there is at least one remaining compressed member.
861 * If the member is a zip file, it must be the only one.
863 local int get_method(in)
864 int in; /* input file descriptor */
866 uch flags; /* compression flags */
867 char magic[2]; /* magic header */
869 magic[0] = (char)get_byte();
870 magic[1] = (char)get_byte();
871 method = -1; /* unknown yet */
872 part_nb++; /* number of parts in gzip file */
874 last_member = RECORD_IO;
875 /* assume multiple members in gzip file except for record oriented I/O */
877 if (memcmp(magic, GZIP_MAGIC, 2) == 0) {
879 method = (int)get_byte();
880 if (method != DEFLATED) {
882 "unknown method %d -- get newer version of gzip\n",
887 flags = (uch)get_byte();
889 (ulg)get_byte(); /* Ignore time stamp */
894 (void)get_byte(); /* Ignore extra flags for the moment */
895 (void)get_byte(); /* Ignore OS type for the moment */
897 if ((flags & EXTRA_FIELD) != 0) {
898 unsigned len = (unsigned)get_byte();
899 len |= ((unsigned)get_byte())<<8;
901 while (len--) (void)get_byte();
904 /* Discard original name if any */
905 if ((flags & ORIG_NAME) != 0) {
906 while (get_char() != 0) /* null */ ;
909 /* Discard file comment if any */
910 if ((flags & COMMENT) != 0) {
911 while (get_char() != 0) /* null */ ;
914 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
919 if (method >= 0) return method;
922 fprintf(stderr, "\nnot in gzip format\n");
926 WARN((stderr, "\ndecompression OK, trailing garbage ignored\n"));
931 /* ========================================================================
932 * Signal and error handler.
934 RETSIGTYPE abort_gzip()
938 /* unzip.c -- decompress files in gzip or pkzip format.
939 * Copyright (C) 1992-1993 Jean-loup Gailly
940 * This is free software; you can redistribute it and/or modify it under the
941 * terms of the GNU General Public License, see the file COPYING.
943 * The code in this file is derived from the file funzip.c written
944 * and put in the public domain by Mark Adler.
948 This version can extract files in gzip or pkzip format.
949 For the latter, only the first entry is extracted, and it has to be
950 either deflated or stored.
953 /* #include "crypt.h" */
955 /* crypt.h (dummy version) -- do not perform encryption
956 * Hardly worth copyrighting :-)
960 # undef CRYPT /* dummy version */
963 #define RAND_HEAD_LEN 12 /* length of encryption random header */
968 /* PKZIP header definitions */
969 #define LOCSIG 0x04034b50L /* four-byte lead-in (lsb first) */
970 #define LOCFLG 6 /* offset of bit flag */
971 #define CRPFLG 1 /* bit for encrypted entry */
972 #define EXTFLG 8 /* bit for extended local header */
973 #define LOCHOW 8 /* offset of compression method */
974 #define LOCTIM 10 /* file mod time (for decryption) */
975 #define LOCCRC 14 /* offset of crc */
976 #define LOCSIZ 18 /* offset of compressed size */
977 #define LOCLEN 22 /* offset of uncompressed length */
978 #define LOCFIL 26 /* offset of file name field length */
979 #define LOCEXT 28 /* offset of extra field length */
980 #define LOCHDR 30 /* size of local header, including sig */
981 #define EXTHDR 16 /* size of extended local header, inc sig */
986 char *key; /* not used--needed to link crypt.c */
987 int pkzip = 0; /* set for a pkzip file */
988 int ext_header = 0; /* set if extended local header */
990 /* ===========================================================================
991 * Unzip in to out. This routine works on both gzip and pkzip files.
993 * IN assertions: the buffer inbuf contains already the beginning of
994 * the compressed data, from offsets inptr to insize-1 included.
995 * The magic header has already been checked. The output buffer is cleared.
998 int in, out; /* input and output file descriptors */
1000 ulg orig_crc = 0; /* original crc */
1001 ulg orig_len = 0; /* original uncompressed length */
1003 uch buf[EXTHDR]; /* extended local header */
1007 method = get_method(ifd);
1009 do_exit(exit_code); /* error message already emitted */
1012 updcrc(NULL, 0); /* initialize crc */
1014 if (pkzip && !ext_header) { /* crc and length at the end otherwise */
1015 orig_crc = LG(inbuf + LOCCRC);
1016 orig_len = LG(inbuf + LOCLEN);
1020 if (method == DEFLATED) {
1022 int res = inflate();
1025 error("out of memory");
1026 } else if (res != 0) {
1027 error("invalid compressed data--format violated");
1031 error("internal error, invalid method");
1034 /* Get the crc and original length */
1036 /* crc32 (see algorithm.doc)
1037 * uncompressed input size modulo 2^32
1039 for (n = 0; n < 8; n++) {
1040 buf[n] = (uch)get_byte(); /* may cause an error if EOF */
1043 orig_len = LG(buf+4);
1045 } else if (ext_header) { /* If extended header, check it */
1046 /* signature - 4bytes: 0x50 0x4b 0x07 0x08
1048 * compressed size 4-bytes
1049 * uncompressed size 4-bytes
1051 for (n = 0; n < EXTHDR; n++) {
1052 buf[n] = (uch)get_byte(); /* may cause an error if EOF */
1054 orig_crc = LG(buf+4);
1055 orig_len = LG(buf+12);
1058 /* Validate decompression */
1059 if (orig_crc != updcrc(outbuf, 0)) {
1060 error("invalid compressed data--crc error");
1062 if (orig_len != (ulg)bytes_out) {
1063 error("invalid compressed data--length error");
1066 /* Check if there are more entries in a pkzip file */
1067 if (pkzip && inptr + 4 < insize && LG(inbuf+inptr) == LOCSIG) {
1068 WARN((stderr,"has more than one entry--rest ignored\n"));
1070 ext_header = pkzip = 0; /* for next file */
1073 /* util.c -- utility functions for gzip support
1074 * Copyright (C) 1992-1993 Jean-loup Gailly
1075 * This is free software; you can redistribute it and/or modify it under the
1076 * terms of the GNU General Public License, see the file COPYING.
1081 #include <sys/types.h>
1083 #ifdef HAVE_UNISTD_H
1084 # include <unistd.h>
1090 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1091 # include <stdlib.h>
1096 static const ulg crc_32_tab[]; /* crc table, defined below */
1098 /* ===========================================================================
1099 * Run a set of bytes through the crc shift register. If s is a NULL
1100 * pointer, then initialize the crc shift register contents instead.
1101 * Return the current crc in either case.
1104 uch *s; /* pointer to bytes to pump through */
1105 unsigned n; /* number of bytes in s[] */
1107 register ulg c; /* temporary variable */
1109 static ulg crc = (ulg)0xffffffffL; /* shift register contents */
1116 c = crc_32_tab[((int)c ^ (*s++)) & 0xff] ^ (c >> 8);
1120 return c ^ 0xffffffffL; /* (instead of ~c for 64-bit machines) */
1123 /* ===========================================================================
1124 * Clear input and output buffers
1130 bytes_in = bytes_out = 0L;
1133 /* ===========================================================================
1134 * Fill the input buffer. This is called only when the buffer is empty.
1136 int fill_inbuf(eof_ok)
1137 int eof_ok; /* set if EOF acceptable as a result */
1141 /* Read as much as possible */
1145 len = read(ifd, (char*)inbuf+insize, INBUFSIZ-insize);
1146 if (len == 0 || len == EOF) break;
1148 } while (insize < INBUFSIZ);
1151 if (eof_ok) return EOF;
1154 bytes_in += (ulg)insize;
1159 /* ===========================================================================
1160 * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
1161 * (used for the compressed data only)
1165 if (outcnt == 0) return;
1168 write_buf(ofd, (char *)outbuf, outcnt);
1169 bytes_out += (ulg)outcnt;
1173 /* ===========================================================================
1174 * Write the output window window[0..outcnt-1] and update crc and bytes_out.
1175 * (Used for the decompressed data only.)
1179 if (outcnt == 0) return;
1180 updcrc(window, outcnt);
1183 write_buf(ofd, (char *)window, outcnt);
1184 bytes_out += (ulg)outcnt;
1188 /* ===========================================================================
1189 * Does the same as write(), but also handles partial pipe writes and checks
1192 void write_buf(fd, buf, cnt)
1199 while ((n = write(fd, buf, cnt)) != cnt) {
1200 if (n == (unsigned)(-1)) {
1204 buf = (voidp)((char*)buf+n);
1208 #if defined(NO_STRING_H) && !defined(STDC_HEADERS)
1210 /* Provide missing strspn and strcspn functions. */
1216 int strspn OF((const char *s, const char *accept));
1217 int strcspn OF((const char *s, const char *reject));
1219 /* ========================================================================
1220 * Return the length of the maximum initial segment
1221 * of s which contains only characters in accept.
1223 int strspn(s, accept)
1227 register const char *p;
1228 register const char *a;
1229 register int count = 0;
1231 for (p = s; *p != '\0'; ++p) {
1232 for (a = accept; *a != '\0'; ++a) {
1233 if (*p == *a) break;
1235 if (*a == '\0') return count;
1241 /* ========================================================================
1242 * Return the length of the maximum inital segment of s
1243 * which contains no characters from reject.
1245 int strcspn(s, reject)
1249 register int count = 0;
1251 while (*s != '\0') {
1252 if (strchr(reject, *s++) != NULL) return count;
1258 #endif /* NO_STRING_H */
1261 /* ========================================================================
1265 char *a, *b; /* message strings juxtaposed in output */
1267 WARN((stderr, "warning: %s%s\n", a, b));
1272 fprintf(stderr, "\n");
1276 fprintf(stderr, "unexpected end of file\n");
1283 fprintf(stderr, "\n");
1289 /* ========================================================================
1290 * Table of CRC-32's of all single-byte values (made by makecrc.c)
1292 static const ulg crc_32_tab[] = {
1293 0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
1294 0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
1295 0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
1296 0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
1297 0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
1298 0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
1299 0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
1300 0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
1301 0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
1302 0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
1303 0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
1304 0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
1305 0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
1306 0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
1307 0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
1308 0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
1309 0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
1310 0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
1311 0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
1312 0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
1313 0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
1314 0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
1315 0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
1316 0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
1317 0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
1318 0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
1319 0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
1320 0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
1321 0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
1322 0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
1323 0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
1324 0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
1325 0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
1326 0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
1327 0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
1328 0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
1329 0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
1330 0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
1331 0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
1332 0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
1333 0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
1334 0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
1335 0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
1336 0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
1337 0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
1338 0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
1339 0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
1340 0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
1341 0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
1342 0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
1343 0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
1346 /* inflate.c -- Not copyrighted 1992 by Mark Adler
1347 version c10p1, 10 January 1993 */
1349 /* You can do whatever you like with this source file, though I would
1350 prefer that if you modify it and redistribute it that you include
1351 comments to that effect with your name and the date. Thank you.
1352 [The history has been moved to the file ChangeLog.]
1356 Inflate deflated (PKZIP's method 8 compressed) data. The compression
1357 method searches for as much of the current string of bytes (up to a
1358 length of 258) in the previous 32K bytes. If it doesn't find any
1359 matches (of at least length 3), it codes the next byte. Otherwise, it
1360 codes the length of the matched string and its distance backwards from
1361 the current position. There is a single Huffman code that codes both
1362 single bytes (called "literals") and match lengths. A second Huffman
1363 code codes the distance information, which follows a length code. Each
1364 length or distance code actually represents a base value and a number
1365 of "extra" (sometimes zero) bits to get to add to the base value. At
1366 the end of each deflated block is a special end-of-block (EOB) literal/
1367 length code. The decoding process is basically: get a literal/length
1368 code; if EOB then done; if a literal, emit the decoded byte; if a
1369 length then get the distance and emit the referred-to bytes from the
1370 sliding window of previously emitted data.
1372 There are (currently) three kinds of inflate blocks: stored, fixed, and
1373 dynamic. The compressor deals with some chunk of data at a time, and
1374 decides which method to use on a chunk-by-chunk basis. A chunk might
1375 typically be 32K or 64K. If the chunk is uncompressible, then the
1376 "stored" method is used. In this case, the bytes are simply stored as
1377 is, eight bits per byte, with none of the above coding. The bytes are
1378 preceded by a count, since there is no longer an EOB code.
1380 If the data is compressible, then either the fixed or dynamic methods
1381 are used. In the dynamic method, the compressed data is preceded by
1382 an encoding of the literal/length and distance Huffman codes that are
1383 to be used to decode this block. The representation is itself Huffman
1384 coded, and so is preceded by a description of that code. These code
1385 descriptions take up a little space, and so for small blocks, there is
1386 a predefined set of codes, called the fixed codes. The fixed method is
1387 used if the block codes up smaller that way (usually for quite small
1388 chunks), otherwise the dynamic method is used. In the latter case, the
1389 codes are customized to the probabilities in the current block, and so
1390 can code it much better than the pre-determined fixed codes.
1392 The Huffman codes themselves are decoded using a mutli-level table
1393 lookup, in order to maximize the speed of decoding plus the speed of
1394 building the decoding tables. See the comments below that precede the
1395 lbits and dbits tuning parameters.
1400 Notes beyond the 1.93a appnote.txt:
1402 1. Distance pointers never point before the beginning of the output
1404 2. Distance pointers can point back across blocks, up to 32k away.
1405 3. There is an implied maximum of 7 bits for the bit length table and
1406 15 bits for the actual data.
1407 4. If only one code exists, then it is encoded using one bit. (Zero
1408 would be more efficient, but perhaps a little confusing.) If two
1409 codes exist, they are coded using one bit each (0 and 1).
1410 5. There is no way of sending zero distance codes--a dummy must be
1411 sent if there are none. (History: a pre 2.0 version of PKZIP would
1412 store blocks with no distance codes, but this was discovered to be
1413 too harsh a criterion.) Valid only for 1.93a. 2.04c does allow
1414 zero distance codes, which is sent as one code of zero bits in
1416 6. There are up to 286 literal/length codes. Code 256 represents the
1417 end-of-block. Note however that the static length tree defines
1418 288 codes just to fill out the Huffman codes. Codes 286 and 287
1419 cannot be used though, since there is no length base or extra bits
1420 defined for them. Similarly, there are up to 30 distance codes.
1421 However, static trees define 32 codes (all 5 bits) to fill out the
1422 Huffman codes, but the last two had better not show up in the data.
1423 7. Unzip can check dynamic Huffman blocks for complete code sets.
1424 The exception is that a single code would not be complete (see #4).
1425 8. The five bits following the block type is really the number of
1426 literal codes sent minus 257.
1427 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
1428 (1+6+6). Therefore, to output three times the length, you output
1429 three codes (1+1+1), whereas to output four times the same length,
1430 you only need two codes (1+3). Hmm.
1431 10. In the tree reconstruction algorithm, Code = Code + Increment
1432 only if BitLength(i) is not zero. (Pretty obvious.)
1433 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19)
1434 12. Note: length code 284 can represent 227-258, but length code 285
1435 really is 258. The last length deserves its own, short code
1436 since it gets used a lot in very redundant files. The length
1437 258 is special since 258 - 3 (the min match length) is 255.
1438 13. The literal/length and distance code bit lengths are read as a
1439 single stream of lengths. It is possible (and advantageous) for
1440 a repeat code (16, 17, or 18) to go across the boundary between
1441 the two sets of lengths.
1444 #include <sys/types.h>
1446 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1447 # include <stdlib.h>
1451 #define slide window
1453 /* Huffman code lookup table entry--this entry is four bytes for machines
1454 that have 16-bit pointers (e.g. PC's in the small or medium model).
1455 Valid extra bits are 0..13. e == 15 is EOB (end of block), e == 16
1456 means that v is a literal, 16 < e < 32 means that v is a pointer to
1457 the next table, which codes e - 16 bits, and lastly e == 99 indicates
1458 an unused code. If a code with e == 99 is looked up, this implies an
1459 error in the data. */
1461 uch e; /* number of extra bits or operation */
1462 uch b; /* number of bits in this code or subcode */
1464 ush n; /* literal, length base, or distance base */
1465 struct huft *t; /* pointer to next level of table */
1470 /* Function prototypes */
1471 int huft_build OF((unsigned *, unsigned, unsigned, ush *, ush *,
1472 struct huft **, int *));
1473 int huft_free OF((struct huft *));
1474 int inflate_codes OF((struct huft *, struct huft *, int, int));
1475 int inflate_stored OF((void));
1476 int inflate_fixed OF((void));
1477 int inflate_dynamic OF((void));
1478 int inflate_block OF((int *));
1479 int inflate OF((void));
1482 /* The inflate algorithm uses a sliding 32K byte window on the uncompressed
1483 stream to find repeated byte strings. This is implemented here as a
1484 circular buffer. The index is updated simply by incrementing and then
1485 and'ing with 0x7fff (32K-1). */
1486 /* It is left to other modules to supply the 32K area. It is assumed
1487 to be usable as if it were declared "uch slide[32768];" or as just
1488 "uch *slide;" and then malloc'ed in the latter case. The definition
1489 must be in unzip.h, included above. */
1490 /* unsigned wp; current position in slide */
1492 #define flush_output(w) (wp=(w),flush_window())
1494 /* Tables for deflate from PKZIP's appnote.txt. */
1495 static unsigned border[] = { /* Order of the bit length code lengths */
1496 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
1497 static ush cplens[] = { /* Copy lengths for literal codes 257..285 */
1498 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
1499 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
1500 /* note: see note #13 above about the 258 in this list. */
1501 static ush cplext[] = { /* Extra bits for literal codes 257..285 */
1502 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
1503 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99}; /* 99==invalid */
1504 static ush cpdist[] = { /* Copy offsets for distance codes 0..29 */
1505 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
1506 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
1507 8193, 12289, 16385, 24577};
1508 static ush cpdext[] = { /* Extra bits for distance codes */
1509 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
1510 7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
1515 /* Macros for inflate() bit peeking and grabbing.
1519 x = b & mask_bits[j];
1522 where NEEDBITS makes sure that b has at least j bits in it, and
1523 DUMPBITS removes the bits from b. The macros use the variable k
1524 for the number of bits in b. Normally, b and k are register
1525 variables for speed, and are initialized at the beginning of a
1526 routine that uses these macros from a global bit buffer and count.
1528 If we assume that EOB will be the longest code, then we will never
1529 ask for bits with NEEDBITS that are beyond the end of the stream.
1530 So, NEEDBITS should not read any more bytes than are needed to
1531 meet the request. Then no bytes need to be "returned" to the buffer
1532 at the end of the last block.
1534 However, this assumption is not true for fixed blocks--the EOB code
1535 is 7 bits, but the other literal/length codes can be 8 or 9 bits.
1536 (The EOB code is shorter than other codes because fixed blocks are
1537 generally short. So, while a block always has an EOB, many other
1538 literal/length codes have a significantly lower probability of
1539 showing up at all.) However, by making the first table have a
1540 lookup of seven bits, the EOB code will be found in that first
1541 lookup, and so will not require that too many bits be pulled from
1545 ulg bb; /* bit buffer */
1546 unsigned bk; /* bits in bit buffer */
1550 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
1551 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
1556 # define NEXTBYTE() (cc = get_byte(), zdecode(cc), cc)
1558 # define NEXTBYTE() (uch)get_byte()
1560 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
1561 #define DUMPBITS(n) {b>>=(n);k-=(n);}
1565 Huffman code decoding is performed using a multi-level table lookup.
1566 The fastest way to decode is to simply build a lookup table whose
1567 size is determined by the longest code. However, the time it takes
1568 to build this table can also be a factor if the data being decoded
1569 is not very long. The most common codes are necessarily the
1570 shortest codes, so those codes dominate the decoding time, and hence
1571 the speed. The idea is you can have a shorter table that decodes the
1572 shorter, more probable codes, and then point to subsidiary tables for
1573 the longer codes. The time it costs to decode the longer codes is
1574 then traded against the time it takes to make longer tables.
1576 This results of this trade are in the variables lbits and dbits
1577 below. lbits is the number of bits the first level table for literal/
1578 length codes can decode in one step, and dbits is the same thing for
1579 the distance codes. Subsequent tables are also less than or equal to
1580 those sizes. These values may be adjusted either when all of the
1581 codes are shorter than that, in which case the longest code length in
1582 bits is used, or when the shortest code is *longer* than the requested
1583 table size, in which case the length of the shortest code in bits is
1586 There are two different values for the two tables, since they code a
1587 different number of possibilities each. The literal/length table
1588 codes 286 possible values, or in a flat code, a little over eight
1589 bits. The distance table codes 30 possible values, or a little less
1590 than five bits, flat. The optimum values for speed end up being
1591 about one bit more than those, so lbits is 8+1 and dbits is 5+1.
1592 The optimum values may differ though from machine to machine, and
1593 possibly even between compilers. Your mileage may vary.
1597 int lbits = 9; /* bits in base literal/length lookup table */
1598 int dbits = 6; /* bits in base distance lookup table */
1601 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
1602 #define BMAX 16 /* maximum bit length of any code (16 for explode) */
1603 #define N_MAX 288 /* maximum number of codes in any set */
1606 unsigned hufts; /* track memory usage */
1609 int huft_build(b, n, s, d, e, t, m)
1610 unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
1611 unsigned n; /* number of codes (assumed <= N_MAX) */
1612 unsigned s; /* number of simple-valued codes (0..s-1) */
1613 ush *d; /* list of base values for non-simple codes */
1614 ush *e; /* list of extra bits for non-simple codes */
1615 struct huft **t; /* result: starting table */
1616 int *m; /* maximum lookup bits, returns actual */
1617 /* Given a list of code lengths and a maximum table size, make a set of
1618 tables to decode that set of codes. Return zero on success, one if
1619 the given code set is incomplete (the tables are still built in this
1620 case), two if the input is invalid (all zero length codes or an
1621 oversubscribed set of lengths), and three if not enough memory. */
1623 unsigned a; /* counter for codes of length k */
1624 unsigned c[BMAX+1]; /* bit length count table */
1625 unsigned f; /* i repeats in table every f entries */
1626 int g; /* maximum code length */
1627 int h; /* table level */
1628 register unsigned i; /* counter, current code */
1629 register unsigned j; /* counter */
1630 register int k; /* number of bits in current code */
1631 int l; /* bits per table (returned in m) */
1632 register unsigned *p; /* pointer into c[], b[], or v[] */
1633 register struct huft *q; /* points to current table */
1634 struct huft r; /* table entry for structure assignment */
1635 struct huft *u[BMAX]; /* table stack */
1636 unsigned v[N_MAX]; /* values in order of bit length */
1637 register int w; /* bits before this table == (l * h) */
1638 unsigned x[BMAX+1]; /* bit offsets, then code stack */
1639 unsigned *xp; /* pointer into x */
1640 int y; /* number of dummy codes added */
1641 unsigned z; /* number of entries in current table */
1644 /* Generate counts for each bit length */
1645 memzero(c, sizeof(c));
1648 Tracecv(*p, (stderr, (n-i >= ' ' && n-i <= '~' ? "%c %d\n" : "0x%x %d\n"),
1650 c[*p]++; /* assume all entries <= BMAX */
1651 p++; /* Can't combine with above line (Solaris bug) */
1653 if (c[0] == n) /* null input--all zero length codes */
1655 *t = (struct huft *)NULL;
1661 /* Find minimum and maximum length, bound *m by those */
1663 for (j = 1; j <= BMAX; j++)
1666 k = j; /* minimum code length */
1667 if ((unsigned)l < j)
1669 for (i = BMAX; i; i--)
1672 g = i; /* maximum code length */
1673 if ((unsigned)l > i)
1678 /* Adjust last length count to fill out codes, if needed */
1679 for (y = 1 << j; j < i; j++, y <<= 1)
1680 if ((y -= c[j]) < 0)
1681 return 2; /* bad input: more codes than bits */
1682 if ((y -= c[i]) < 0)
1687 /* Generate starting offsets into the value table for each length */
1689 p = c + 1; xp = x + 2;
1690 while (--i) { /* note that i == g from above */
1691 *xp++ = (j += *p++);
1695 /* Make a table of values in order of bit lengths */
1698 if ((j = *p++) != 0)
1703 /* Generate the Huffman codes and for each, make the table entries */
1704 x[0] = i = 0; /* first Huffman code is zero */
1705 p = v; /* grab values in bit order */
1706 h = -1; /* no tables yet--level -1 */
1707 w = -l; /* bits decoded == (l * h) */
1708 u[0] = (struct huft *)NULL; /* just to keep compilers happy */
1709 q = (struct huft *)NULL; /* ditto */
1712 /* go through the bit lengths (k already is bits in shortest code) */
1718 /* here i is the Huffman code of length k bits for value *p */
1719 /* make tables up to required level */
1723 w += l; /* previous table always l bits */
1725 /* compute minimum size table less than or equal to l bits */
1726 z = (z = g - w) > (unsigned)l ? l : z; /* upper limit on table size */
1727 if ((f = 1 << (j = k - w)) > a + 1) /* try a k-w bit table */
1728 { /* too few codes for k-w bit table */
1729 f -= a + 1; /* deduct codes from patterns left */
1731 while (++j < z) /* try smaller tables up to z bits */
1733 if ((f <<= 1) <= *++xp)
1734 break; /* enough codes to use up j bits */
1735 f -= *xp; /* else deduct codes from patterns */
1738 z = 1 << j; /* table entries for j-bit table */
1740 /* allocate and link in new table */
1741 if ((q = (struct huft *)malloc((z + 1)*sizeof(struct huft))) ==
1742 (struct huft *)NULL)
1746 return 3; /* not enough memory */
1748 hufts += z + 1; /* track memory usage */
1749 *t = q + 1; /* link to list for huft_free() */
1750 *(t = &(q->v.t)) = (struct huft *)NULL;
1751 u[h] = ++q; /* table starts after link */
1753 /* connect to last table, if there is one */
1756 x[h] = i; /* save pattern for backing up */
1757 r.b = (uch)l; /* bits to dump before this table */
1758 r.e = (uch)(16 + j); /* bits in this table */
1759 r.v.t = q; /* pointer to this table */
1760 j = i >> (w - l); /* (get around Turbo C bug) */
1761 u[h-1][j] = r; /* connect to last table */
1765 /* set up table entry in r */
1768 r.e = 99; /* out of values--invalid code */
1771 r.e = (uch)(*p < 256 ? 16 : 15); /* 256 is end-of-block code */
1772 r.v.n = (ush)(*p); /* simple code is just the value */
1773 p++; /* one compiler does not like *p++ */
1777 r.e = (uch)e[*p - s]; /* non-simple--look up in lists */
1778 r.v.n = d[*p++ - s];
1781 /* fill code-like entries with r */
1783 for (j = i >> w; j < z; j += f)
1786 /* backwards increment the k-bit code i */
1787 for (j = 1 << (k - 1); i & j; j >>= 1)
1791 /* backup over finished tables */
1792 while ((i & ((1 << w) - 1)) != x[h])
1794 h--; /* don't need to update q */
1801 /* Return true (1) if we were given an incomplete table */
1802 return y != 0 && g != 1;
1808 struct huft *t; /* table to free */
1809 /* Free the malloc'ed tables built by huft_build(), which makes a linked
1810 list of the tables it made, with the links in a dummy first entry of
1813 register struct huft *p, *q;
1816 /* Go through linked list, freeing from the malloced (t[-1]) address. */
1818 while (p != (struct huft *)NULL)
1828 int inflate_codes(tl, td, bl, bd)
1829 struct huft *tl, *td; /* literal/length and distance decoder tables */
1830 int bl, bd; /* number of bits decoded by tl[] and td[] */
1831 /* inflate (decompress) the codes in a deflated (compressed) block.
1832 Return an error code or zero if it all goes ok. */
1834 register unsigned e; /* table entry flag/number of extra bits */
1835 unsigned n, d; /* length and index for copy */
1836 unsigned w; /* current window position */
1837 struct huft *t; /* pointer to table entry */
1838 unsigned ml, md; /* masks for bl and bd bits */
1839 register ulg b; /* bit buffer */
1840 register unsigned k; /* number of bits in bit buffer */
1843 /* make local copies of globals */
1844 b = bb; /* initialize bit buffer */
1846 w = wp; /* initialize window position */
1848 /* inflate the coded data */
1849 ml = mask_bits[bl]; /* precompute masks for speed */
1851 for (;;) /* do until end of block */
1853 NEEDBITS((unsigned)bl)
1854 if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
1861 } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
1863 if (e == 16) /* then it's a literal */
1865 slide[w++] = (uch)t->v.n;
1866 Tracevv((stderr, "%c", slide[w-1]));
1873 else /* it's an EOB or a length */
1875 /* exit if end of block */
1879 /* get length of block to copy */
1881 n = t->v.n + ((unsigned)b & mask_bits[e]);
1884 /* decode distance of block to copy */
1885 NEEDBITS((unsigned)bd)
1886 if ((e = (t = td + ((unsigned)b & md))->e) > 16)
1893 } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
1896 d = w - t->v.n - ((unsigned)b & mask_bits[e]);
1898 Tracevv((stderr,"\\[%d,%d]", w-d, n));
1902 n -= (e = (e = WSIZE - ((d &= WSIZE-1) > w ? d : w)) > n ? n : e);
1903 #if !defined(NOMEMCPY) && !defined(DEBUG)
1904 if (w - d >= e) /* (this test assumes unsigned comparison) */
1906 memcpy(slide + w, slide + d, e);
1910 else /* do it slow to avoid memcpy() overlap */
1911 #endif /* !NOMEMCPY */
1913 slide[w++] = slide[d++];
1914 Tracevv((stderr, "%c", slide[w-1]));
1926 /* restore the globals from the locals */
1927 wp = w; /* restore global window pointer */
1928 bb = b; /* restore global bit buffer */
1937 int inflate_stored()
1938 /* "decompress" an inflated type 0 (stored) block. */
1940 unsigned n; /* number of bytes in block */
1941 unsigned w; /* current window position */
1942 register ulg b; /* bit buffer */
1943 register unsigned k; /* number of bits in bit buffer */
1946 /* make local copies of globals */
1947 b = bb; /* initialize bit buffer */
1949 w = wp; /* initialize window position */
1952 /* go to byte boundary */
1957 /* get the length and its complement */
1959 n = ((unsigned)b & 0xffff);
1962 if (n != (unsigned)((~b) & 0xffff))
1963 return 1; /* error in compressed data */
1967 /* read and output the compressed data */
1971 slide[w++] = (uch)b;
1981 /* restore the globals from the locals */
1982 wp = w; /* restore global window pointer */
1983 bb = b; /* restore global bit buffer */
1991 /* decompress an inflated type 1 (fixed Huffman codes) block. We should
1992 either replace this with a custom decoder, or at least precompute the
1995 int i; /* temporary variable */
1996 struct huft *tl; /* literal/length code table */
1997 struct huft *td; /* distance code table */
1998 int bl; /* lookup bits for tl */
1999 int bd; /* lookup bits for td */
2000 unsigned l[288]; /* length list for huft_build */
2003 /* set up literal table */
2004 for (i = 0; i < 144; i++)
2006 for (; i < 256; i++)
2008 for (; i < 280; i++)
2010 for (; i < 288; i++) /* make a complete, but wrong code set */
2013 if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
2017 /* set up distance table */
2018 for (i = 0; i < 30; i++) /* make an incomplete code set */
2021 if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1)
2028 /* decompress until an end-of-block code */
2029 if (inflate_codes(tl, td, bl, bd))
2033 /* free the decoding tables, return */
2041 int inflate_dynamic()
2042 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
2044 int i; /* temporary variables */
2046 unsigned l; /* last length */
2047 unsigned m; /* mask for bit lengths table */
2048 unsigned n; /* number of lengths to get */
2049 struct huft *tl; /* literal/length code table */
2050 struct huft *td; /* distance code table */
2051 int bl; /* lookup bits for tl */
2052 int bd; /* lookup bits for td */
2053 unsigned nb; /* number of bit length codes */
2054 unsigned nl; /* number of literal/length codes */
2055 unsigned nd; /* number of distance codes */
2056 #ifdef PKZIP_BUG_WORKAROUND
2057 unsigned ll[288+32]; /* literal/length and distance code lengths */
2059 unsigned ll[286+30]; /* literal/length and distance code lengths */
2061 register ulg b; /* bit buffer */
2062 register unsigned k; /* number of bits in bit buffer */
2065 /* make local bit buffer */
2070 /* read in table lengths */
2072 nl = 257 + ((unsigned)b & 0x1f); /* number of literal/length codes */
2075 nd = 1 + ((unsigned)b & 0x1f); /* number of distance codes */
2078 nb = 4 + ((unsigned)b & 0xf); /* number of bit length codes */
2080 #ifdef PKZIP_BUG_WORKAROUND
2081 if (nl > 288 || nd > 32)
2083 if (nl > 286 || nd > 30)
2085 return 1; /* bad lengths */
2088 /* read in bit-length-code lengths */
2089 for (j = 0; j < nb; j++)
2092 ll[border[j]] = (unsigned)b & 7;
2099 /* build decoding table for trees--single level, 7 bit lookup */
2101 if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0)
2105 return i; /* incomplete code set */
2109 /* read in literal and distance code lengths */
2113 while ((unsigned)i < n)
2115 NEEDBITS((unsigned)bl)
2116 j = (td = tl + ((unsigned)b & m))->b;
2119 if (j < 16) /* length of code in bits (0..15) */
2120 ll[i++] = l = j; /* save last length in l */
2121 else if (j == 16) /* repeat last length 3 to 6 times */
2124 j = 3 + ((unsigned)b & 3);
2126 if ((unsigned)i + j > n)
2131 else if (j == 17) /* 3 to 10 zero length codes */
2134 j = 3 + ((unsigned)b & 7);
2136 if ((unsigned)i + j > n)
2142 else /* j == 18: 11 to 138 zero length codes */
2145 j = 11 + ((unsigned)b & 0x7f);
2147 if ((unsigned)i + j > n)
2156 /* free decoding table for trees */
2160 /* restore the global bit buffer */
2165 /* build the decoding tables for literal/length and distance codes */
2167 if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0)
2170 fprintf(stderr, " incomplete literal tree\n");
2173 return i; /* incomplete code set */
2176 if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0)
2179 fprintf(stderr, " incomplete distance tree\n");
2180 #ifdef PKZIP_BUG_WORKAROUND
2187 return i; /* incomplete code set */
2192 /* decompress until an end-of-block code */
2193 if (inflate_codes(tl, td, bl, bd))
2197 /* free the decoding tables, return */
2205 int inflate_block(e)
2206 int *e; /* last block flag */
2207 /* decompress an inflated block */
2209 unsigned t; /* block type */
2210 register ulg b; /* bit buffer */
2211 register unsigned k; /* number of bits in bit buffer */
2214 /* make local bit buffer */
2219 /* read in last block bit */
2225 /* read in block type */
2227 t = (unsigned)b & 3;
2231 /* restore the global bit buffer */
2236 /* inflate that block type */
2238 return inflate_dynamic();
2240 return inflate_stored();
2242 return inflate_fixed();
2245 /* bad block type */
2252 /* decompress an inflated entry */
2254 int e; /* last block flag */
2255 int r; /* result code */
2256 unsigned h; /* maximum struct huft's malloc'ed */
2259 /* initialize window, bit buffer */
2265 /* decompress until the last block */
2269 if ((r = inflate_block(&e)) != 0)
2275 /* Undo too much lookahead. The next read will be byte aligned so we
2276 * can discard unused bits in the last meaningful byte.
2283 /* flush out slide */
2287 /* return success */
2289 fprintf(stderr, "<%u> ", h);