1 /* zcat : stripped version based on gzip sources
2 Sven Rudolph <sr1@inf.tu-dresden.de>
6 #define bb_need_name_too_long
7 #define BB_DECLARE_EXTERN
10 static const char gunzip_usage[] =
11 "gunzip [OPTION]... FILE\n\n"
12 "Uncompress FILE (or standard input if FILE is '-').\n\n"
14 "\t-c\tWrite output to standard output\n"
15 "\t-t\tTest compressed file integrity\n";
17 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
18 * Copyright (C) 1992-1993 Jean-loup Gailly
19 * The unzip code was written and put in the public domain by Mark Adler.
20 * Portions of the lzw code are derived from the public domain 'compress'
21 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
22 * Ken Turkowski, Dave Mack and Peter Jannesen.
24 * See the license_msg below and the file COPYING for the software license.
25 * See the file algorithm.doc for the compression algorithms and file formats.
29 static char *license_msg[] = {
30 " Copyright (C) 1992-1993 Jean-loup Gailly",
31 " This program is free software; you can redistribute it and/or modify",
32 " it under the terms of the GNU General Public License as published by",
33 " the Free Software Foundation; either version 2, or (at your option)",
34 " any later version.",
36 " This program is distributed in the hope that it will be useful,",
37 " but WITHOUT ANY WARRANTY; without even the implied warranty of",
38 " MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the",
39 " GNU General Public License for more details.",
41 " You should have received a copy of the GNU General Public License",
42 " along with this program; if not, write to the Free Software",
43 " Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.",
47 /* Compress files with zip algorithm and 'compress' interface.
48 * See usage() and help() functions below for all options.
50 * file.gz: compressed file with same mode, owner, and utimes
51 * or stdout with -c option or if stdin used as input.
52 * If the output file name had to be truncated, the original name is kept
53 * in the compressed file.
54 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
56 * Using gz on MSDOS would create too many file name conflicts. For
57 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
58 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
59 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
60 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
62 * For the meaning of all compilation flags, see comments in Makefile.in.
66 #include <sys/types.h>
70 #include <sys/param.h> /* for PATH_MAX */
72 /* #include "tailor.h" */
74 /* tailor.h -- target dependent definitions
75 * Copyright (C) 1992-1993 Jean-loup Gailly.
76 * This is free software; you can redistribute it and/or modify it under the
77 * terms of the GNU General Public License, see the file COPYING.
80 /* The target dependent definitions should be defined here only.
81 * The target dependent functions should be defined in tailor.c.
86 #define get_char() get_byte()
87 #define put_char(c) put_byte(c)
89 /* #include "gzip.h" */
91 /* gzip.h -- common declarations for all gzip modules
92 * Copyright (C) 1992-1993 Jean-loup Gailly.
93 * This is free software; you can redistribute it and/or modify it under the
94 * terms of the GNU General Public License, see the file COPYING.
97 #if defined(__STDC__) || defined(PROTO)
98 # define OF(args) args
109 /* I don't like nested includes, but the string and io functions are used
113 #if !defined(NO_STRING_H) || defined(STDC_HEADERS)
115 # if !defined(STDC_HEADERS) && !defined(NO_MEMORY_H) && !defined(__GNUC__)
118 # define memzero(s, n) memset ((voidp)(s), 0, (n))
120 # include <strings.h>
121 # define strchr index
122 # define strrchr rindex
123 # define memcpy(d, s, n) bcopy((s), (d), (n))
124 # define memcmp(s1, s2, n) bcmp((s1), (s2), (n))
125 # define memzero(s, n) bzero((s), (n))
129 # define RETSIGTYPE void
134 typedef unsigned char uch;
135 typedef unsigned short ush;
136 typedef unsigned long ulg;
138 /* Return codes from gzip */
143 /* Compression methods (see algorithm.doc) */
146 extern int method; /* compression method */
148 /* To save memory for 16 bit systems, some arrays are overlaid between
149 * the various modules:
150 * deflate: prev+head window d_buf l_buf outbuf
151 * unlzw: tab_prefix tab_suffix stack inbuf outbuf
152 * inflate: window inbuf
153 * unpack: window inbuf prefix_len
154 * unlzh: left+right window c_table inbuf c_len
155 * For compression, input is done in window[]. For decompression, output
156 * is done in window except for unlzw.
161 # define INBUFSIZ 0x2000 /* input buffer size */
163 # define INBUFSIZ 0x8000 /* input buffer size */
166 #define INBUF_EXTRA 64 /* required by unlzw() */
170 # define OUTBUFSIZ 8192 /* output buffer size */
172 # define OUTBUFSIZ 16384 /* output buffer size */
175 #define OUTBUF_EXTRA 2048 /* required by unlzw() */
181 # define DIST_BUFSIZE 0x2000 /* buffer for distances, see trees.c */
183 # define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
187 /*#define DYN_ALLOC*/
190 # define EXTERN(type, array) extern type * array
191 # define DECLARE(type, array, size) type * array
192 # define ALLOC(type, array, size) { \
193 array = (type*)calloc((size_t)(((size)+1L)/2), 2*sizeof(type)); \
194 if (array == NULL) error("insufficient memory"); \
196 # define FREE(array) {if (array != NULL) free(array), array=NULL;}
198 # define EXTERN(type, array) extern type array[]
199 # define DECLARE(type, array, size) type array[size]
200 # define ALLOC(type, array, size)
204 EXTERN(uch, inbuf); /* input buffer */
205 EXTERN(uch, outbuf); /* output buffer */
206 EXTERN(ush, d_buf); /* buffer for distances, see trees.c */
207 EXTERN(uch, window); /* Sliding window and suffix table (unlzw) */
208 #define tab_suffix window
210 # define tab_prefix prev /* hash link (see deflate.c) */
211 # define head (prev+WSIZE) /* hash head (see deflate.c) */
212 EXTERN(ush, tab_prefix); /* prefix code (see unlzw.c) */
214 # define tab_prefix0 prev
215 # define head tab_prefix1
216 EXTERN(ush, tab_prefix0); /* prefix for even codes */
217 EXTERN(ush, tab_prefix1); /* prefix for odd codes */
220 extern unsigned insize; /* valid bytes in inbuf */
221 extern unsigned inptr; /* index of next byte to be processed in inbuf */
222 extern unsigned outcnt; /* bytes in output buffer */
224 extern long bytes_in; /* number of input bytes */
225 extern long bytes_out; /* number of output bytes */
226 extern long header_bytes;/* number of bytes in gzip header */
228 extern long ifile_size; /* input file size, -1 for devices (debug only) */
230 typedef int file_t; /* Do not use stdio */
231 #define NO_FILE (-1) /* in memory compression */
234 #define GZIP_MAGIC "\037\213" /* Magic header for gzip files, 1F 8B */
237 #define ASCII_FLAG 0x01 /* bit 0 set: file probably ascii text */
238 #define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
239 #define EXTRA_FIELD 0x04 /* bit 2 set: extra field present */
240 #define ORIG_NAME 0x08 /* bit 3 set: original file name present */
241 #define COMMENT 0x10 /* bit 4 set: file comment present */
242 #define ENCRYPTED 0x20 /* bit 5 set: file is encrypted */
243 #define RESERVED 0xC0 /* bit 6,7: reserved */
246 # define WSIZE 0x8000 /* window size--must be a power of two, and */
247 #endif /* at least 32K for zip's deflate method */
250 #define MAX_MATCH 258
251 /* The minimum and maximum match lengths */
253 #define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
254 /* Minimum amount of lookahead, except at the end of the input file.
255 * See deflate.c for comments about the MIN_MATCH+1.
258 #define MAX_DIST (WSIZE-MIN_LOOKAHEAD)
259 /* In order to simplify the code, particularly on 16 bit machines, match
260 * distances are limited to MAX_DIST instead of WSIZE.
263 extern int exit_code; /* program exit code */
264 extern int verbose; /* be verbose (-v) */
265 extern int level; /* compression level */
266 extern int test; /* check .z file integrity */
267 extern int save_orig_name; /* set if original name must be saved */
269 #define get_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(0))
270 #define try_byte() (inptr < insize ? inbuf[inptr++] : fill_inbuf(1))
272 /* put_byte is used for the compressed output, put_ubyte for the
273 * uncompressed output. However unlzw() uses window for its
274 * suffix table instead of its output buffer, so it does not use put_ubyte
275 * (to be cleaned up).
277 #define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
279 #define put_ubyte(c) {window[outcnt++]=(uch)(c); if (outcnt==WSIZE)\
282 /* Output a 16 bit value, lsb first */
283 #define put_short(w) \
284 { if (outcnt < OUTBUFSIZ-2) { \
285 outbuf[outcnt++] = (uch) ((w) & 0xff); \
286 outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
288 put_byte((uch)((w) & 0xff)); \
289 put_byte((uch)((ush)(w) >> 8)); \
293 /* Output a 32 bit value to the bit stream, lsb first */
294 #define put_long(n) { \
295 put_short((n) & 0xffff); \
296 put_short(((ulg)(n)) >> 16); \
299 #define seekable() 0 /* force sequential output */
300 #define translate_eol 0 /* no option -a yet */
302 #define tolow(c) (isupper(c) ? (c)-'A'+'a' : (c)) /* force to lower case */
304 /* Macros for getting two-byte and four-byte header values */
305 #define SH(p) ((ush)(uch)((p)[0]) | ((ush)(uch)((p)[1]) << 8))
306 #define LG(p) ((ulg)(SH(p)) | ((ulg)(SH((p)+2)) << 16))
308 /* Diagnostic functions */
310 # define Assert(cond,msg) {if(!(cond)) error(msg);}
311 # define Trace(x) fprintf x
312 # define Tracev(x) {if (verbose) fprintf x ;}
313 # define Tracevv(x) {if (verbose>1) fprintf x ;}
314 # define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
315 # define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
317 # define Assert(cond,msg)
322 # define Tracecv(c,x)
325 #define WARN(msg) {fprintf msg ; \
326 if (exit_code == OK) exit_code = WARNING;}
328 #define do_exit(c) exit(c)
332 extern int unzip OF((int in, int out));
335 RETSIGTYPE abort_gzip OF((void));
338 void lm_init OF((int pack_level, ush *flags));
339 ulg deflate OF((void));
342 void ct_init OF((ush *attr, int *method));
343 int ct_tally OF((int dist, int lc));
344 ulg flush_block OF((char *buf, ulg stored_len, int eof));
347 void bi_init OF((file_t zipfile));
348 void send_bits OF((int value, int length));
349 unsigned bi_reverse OF((unsigned value, int length));
350 void bi_windup OF((void));
351 void copy_block OF((char *buf, unsigned len, int header));
352 extern int (*read_buf) OF((char *buf, unsigned size));
355 extern int copy OF((int in, int out));
356 extern ulg updcrc OF((uch *s, unsigned n));
357 extern void clear_bufs OF((void));
358 extern int fill_inbuf OF((int eof_ok));
359 extern void flush_outbuf OF((void));
360 extern void flush_window OF((void));
361 extern void write_buf OF((int fd, voidp buf, unsigned cnt));
363 extern char *basename OF((char *fname));
364 #endif /* not __linux__ */
365 extern void error OF((char *m));
366 extern void warn OF((char *a, char *b));
367 extern void read_error OF((void));
368 extern void write_error OF((void));
371 extern int inflate OF((void));
373 /* #include "lzw.h" */
375 /* lzw.h -- define the lzw functions.
376 * Copyright (C) 1992-1993 Jean-loup Gailly.
377 * This is free software; you can redistribute it and/or modify it under the
378 * terms of the GNU General Public License, see the file COPYING.
381 #if !defined(OF) && defined(lint)
388 #define INIT_BITS 9 /* Initial number of bits per code */
390 #define LZW_MAGIC "\037\235" /* Magic header for lzw files, 1F 9D */
392 #define BIT_MASK 0x1f /* Mask for 'number of compression bits' */
393 /* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
394 * It's a pity that old uncompress does not check bit 0x20. That makes
395 * extension of the format actually undesirable because old compress
396 * would just crash on the new format instead of giving a meaningful
397 * error message. It does check the number of bits, but it's more
398 * helpful to say "unsupported format, get a new version" than
399 * "can only handle 16 bits".
402 #define BLOCK_MODE 0x80
403 /* Block compression: if table is full and compression rate is dropping,
404 * clear the dictionary.
407 #define LZW_RESERVED 0x60 /* reserved bits */
409 #define CLEAR 256 /* flush the dictionary */
410 #define FIRST (CLEAR+1) /* first free entry */
412 extern int maxbits; /* max bits per code for LZW */
413 extern int block_mode; /* block compress mode -C compatible with 2.0 */
415 extern int lzw OF((int in, int out));
416 extern int unlzw OF((int in, int out));
419 /* #include "revision.h" */
421 /* revision.h -- define the version number
422 * Copyright (C) 1992-1993 Jean-loup Gailly.
423 * This is free software; you can redistribute it and/or modify it under the
424 * terms of the GNU General Public License, see the file COPYING.
427 #define VERSION "1.2.4"
429 #define REVDATE "18 Aug 93"
431 /* This version does not support compression into old compress format: */
436 /* #include "getopt.h" */
438 /* Declarations for getopt.
439 Copyright (C) 1989, 1990, 1991, 1992, 1993 Free Software Foundation, Inc.
441 This program is free software; you can redistribute it and/or modify it
442 under the terms of the GNU General Public License as published by the
443 Free Software Foundation; either version 2, or (at your option) any
446 This program is distributed in the hope that it will be useful,
447 but WITHOUT ANY WARRANTY; without even the implied warranty of
448 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
449 GNU General Public License for more details.
451 You should have received a copy of the GNU General Public License
452 along with this program; if not, write to the Free Software
453 Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
462 /* For communication from `getopt' to the caller.
463 When `getopt' finds an option that takes an argument,
464 the argument value is returned here.
465 Also, when `ordering' is RETURN_IN_ORDER,
466 each non-option ARGV-element is returned here. */
470 /* Index in ARGV of the next element to be scanned.
471 This is used for communication to and from the caller
472 and for communication between successive calls to `getopt'.
474 On entry to `getopt', zero means this is the first call; initialize.
476 When `getopt' returns EOF, this is the index of the first of the
477 non-option elements that the caller should itself scan.
479 Otherwise, `optind' communicates from one call to the next
480 how much of ARGV has been scanned so far. */
484 /* Callers store zero here to inhibit the error message `getopt' prints
485 for unrecognized options. */
489 /* Set to an option character which was unrecognized. */
493 /* Describe the long-named options requested by the application.
494 The LONG_OPTIONS argument to getopt_long or getopt_long_only is a vector
495 of `struct option' terminated by an element containing a name which is
498 The field `has_arg' is:
499 no_argument (or 0) if the option does not take an argument,
500 required_argument (or 1) if the option requires an argument,
501 optional_argument (or 2) if the option takes an optional argument.
503 If the field `flag' is not NULL, it points to a variable that is set
504 to the value given in the field `val' when the option is found, but
505 left unchanged if the option is not found.
507 To have a long-named option do something other than set an `int' to
508 a compiled-in constant, such as set a value from `optarg', set the
509 option's `flag' field to zero and its `val' field to a nonzero
510 value (the equivalent single-letter option character, if there is
511 one). For long options that have a zero `flag' field, `getopt'
512 returns the contents of the `val' field. */
521 /* has_arg can't be an enum because some compilers complain about
522 type mismatches in all the code that assumes it is an int. */
528 /* Names for the values of the `has_arg' field of `struct option'. */
530 #define no_argument 0
531 #define required_argument 1
532 #define optional_argument 2
534 #if __STDC__ || defined(PROTO)
535 #if defined(__GNU_LIBRARY__)
536 /* Many other libraries have conflicting prototypes for getopt, with
537 differences in the consts, in stdlib.h. To avoid compilation
538 errors, only prototype getopt for the GNU C library. */
539 extern int getopt (int argc, char *const *argv, const char *shortopts);
540 #endif /* not __GNU_LIBRARY__ */
541 extern int getopt_long (int argc, char *const *argv, const char *shortopts,
542 const struct option *longopts, int *longind);
543 extern int getopt_long_only (int argc, char *const *argv,
544 const char *shortopts,
545 const struct option *longopts, int *longind);
547 /* Internal only. Users should not call this directly. */
548 extern int _getopt_internal (int argc, char *const *argv,
549 const char *shortopts,
550 const struct option *longopts, int *longind,
552 #else /* not __STDC__ */
553 extern int getopt ();
554 extern int getopt_long ();
555 extern int getopt_long_only ();
557 extern int _getopt_internal ();
558 #endif /* not __STDC__ */
564 #endif /* _GETOPT_H */
575 typedef struct dirent dir_type;
576 # define NLENGTH(dirent) ((int)strlen((dirent)->d_name))
577 # define DIR_OPT "DIRENT"
579 # define NLENGTH(dirent) ((dirent)->d_namlen)
581 # include <sys/dir.h>
582 typedef struct direct dir_type;
583 # define DIR_OPT "SYSDIR"
586 # include <sys/ndir.h>
587 typedef struct direct dir_type;
588 # define DIR_OPT "SYSNDIR"
592 typedef struct direct dir_type;
593 # define DIR_OPT "NDIR"
596 # define DIR_OPT "NO_DIR"
602 #if !defined(S_ISDIR) && defined(S_IFDIR)
603 # define S_ISDIR(m) (((m) & S_IFMT) == S_IFDIR)
605 #if !defined(S_ISREG) && defined(S_IFREG)
606 # define S_ISREG(m) (((m) & S_IFMT) == S_IFREG)
609 typedef RETSIGTYPE (*sig_type) OF((int));
612 # define O_BINARY 0 /* creation mode for open() */
616 /* Pure BSD system? */
617 # include <sys/file.h>
619 # define O_CREAT FCREAT
622 # define O_EXCL FEXCL
627 # define S_IRUSR 0400
630 # define S_IWUSR 0200
632 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
634 #ifndef MAX_PATH_LEN /* max pathname length */
636 # define MAX_PATH_LEN PATH_MAX
638 # define MAX_PATH_LEN 1024
648 off_t lseek OF((int fd, off_t offset, int whence));
654 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
655 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
656 DECLARE(ush, d_buf, DIST_BUFSIZE);
657 DECLARE(uch, window, 2L*WSIZE);
659 DECLARE(ush, tab_prefix, 1L<<BITS);
661 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
662 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
665 /* local variables */
667 int test_mode = 0; /* check file integrity option */
668 int foreground; /* set if program run in foreground */
669 int maxbits = BITS; /* max bits per code for LZW */
670 int method = DEFLATED;/* compression method */
671 int exit_code = OK; /* program exit code */
672 int last_member; /* set for .zip and .Z files */
673 int part_nb; /* number of parts in .gz file */
674 long ifile_size; /* input file size, -1 for devices (debug only) */
676 long bytes_in; /* number of input bytes */
677 long bytes_out; /* number of output bytes */
678 long total_in = 0; /* input bytes for all files */
679 long total_out = 0; /* output bytes for all files */
680 struct stat istat; /* status for input file */
681 int ifd; /* input file descriptor */
682 int ofd; /* output file descriptor */
683 unsigned insize; /* valid bytes in inbuf */
684 unsigned inptr; /* index of next byte to be processed in inbuf */
685 unsigned outcnt; /* bytes in output buffer */
687 long header_bytes; /* number of bytes in gzip header */
689 /* local functions */
691 local int get_method OF((int in));
693 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
695 /* ======================================================================== */
696 int gunzip_main (int argc, char** argv)
698 int file_count; /* number of files to precess */
707 char ifname[MAX_PATH_LEN + 1]; /* input file name */
708 char ofname[MAX_PATH_LEN + 1]; /* output file name */
713 if (strcmp(*argv, "zcat")==0)
716 /* Parse any options */
717 while (--argc > 0 && **(++argv) == '-') {
718 if (*((*argv)+1) == '\0') {
722 while (*(++(*argv))) {
737 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
739 (void) signal (SIGINT, (sig_type)abort_gzip);
742 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
743 (void) signal(SIGTERM, (sig_type)abort_gzip);
747 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
748 (void) signal(SIGHUP, (sig_type)abort_gzip);
752 file_count = argc - optind;
754 /* Allocate all global buffers (for DYN_ALLOC option) */
755 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
756 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
757 ALLOC(ush, d_buf, DIST_BUFSIZE);
758 ALLOC(uch, window, 2L*WSIZE);
760 ALLOC(ush, tab_prefix, 1L<<BITS);
762 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
763 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
767 strcpy(ofname, "stdin");
769 inFileNum=fileno(stdin);
770 ifile_size = -1L; /* convention for unknown size */
772 /* Open up the input file */
775 if (strlen(*argv) > MAX_PATH_LEN) {
776 fprintf(stderr, name_too_long, "gunzip");
779 strcpy(ifname, *argv);
781 /* Open input fille */
782 inFileNum=open( ifname, O_RDONLY);
787 /* Get the time stamp on the input file. */
788 result = stat(ifname, &statBuf);
793 ifile_size = statBuf.st_size;
797 /* And get to work */
798 strcpy(ofname, "stdout");
799 outFileNum=fileno(stdout);
801 clear_bufs(); /* clear input and output buffers */
804 /* Actually do the compression/decompression. */
805 unzip(inFileNum, outFileNum);
807 } else if (test_mode) {
808 /* Actually do the compression/decompression. */
813 /* And get to work */
814 if (strlen(ifname) > MAX_PATH_LEN - 4) {
815 fprintf(stderr, name_too_long, "gunzip");
818 strcpy(ofname, ifname);
819 pos=strstr(ofname, ".gz");
824 pos=strstr(ofname, ".tgz");
827 strcat( pos, ".tar");
832 /* Open output fille */
833 #if (__GLIBC__ >= 2) && (__GLIBC_MINOR__ >= 1)
834 outFileNum=open( ofname, O_RDWR|O_CREAT|O_EXCL|O_NOFOLLOW);
836 outFileNum=open( ofname, O_RDWR|O_CREAT|O_EXCL);
838 if (outFileNum < 0) {
842 /* Set permissions on the file */
843 fchmod(outFileNum, statBuf.st_mode);
845 clear_bufs(); /* clear input and output buffers */
848 /* Actually do the compression/decompression. */
849 result=unzip(inFileNum, outFileNum);
853 /* Delete the original file */
859 if (delInputFile == 1 && unlink (delFileName) < 0) {
860 perror (delFileName);
868 /* ========================================================================
869 * Check the magic number of the input file and update ofname if an
870 * original name was given and to_stdout is not set.
871 * Return the compression method, -1 for error, -2 for warning.
872 * Set inptr to the offset of the next byte to be processed.
873 * Updates time_stamp if there is one and --no-time is not used.
874 * This function may be called repeatedly for an input file consisting
875 * of several contiguous gzip'ed members.
876 * IN assertions: there is at least one remaining compressed member.
877 * If the member is a zip file, it must be the only one.
879 local int get_method(in)
880 int in; /* input file descriptor */
882 uch flags; /* compression flags */
883 char magic[2]; /* magic header */
885 magic[0] = (char)get_byte();
886 magic[1] = (char)get_byte();
887 method = -1; /* unknown yet */
888 part_nb++; /* number of parts in gzip file */
890 last_member = RECORD_IO;
891 /* assume multiple members in gzip file except for record oriented I/O */
893 if (memcmp(magic, GZIP_MAGIC, 2) == 0) {
895 method = (int)get_byte();
896 if (method != DEFLATED) {
898 "unknown method %d -- get newer version of gzip\n",
903 flags = (uch)get_byte();
905 (ulg)get_byte(); /* Ignore time stamp */
910 (void)get_byte(); /* Ignore extra flags for the moment */
911 (void)get_byte(); /* Ignore OS type for the moment */
913 if ((flags & EXTRA_FIELD) != 0) {
914 unsigned len = (unsigned)get_byte();
915 len |= ((unsigned)get_byte())<<8;
917 while (len--) (void)get_byte();
920 /* Discard original name if any */
921 if ((flags & ORIG_NAME) != 0) {
922 while (get_char() != 0) /* null */ ;
925 /* Discard file comment if any */
926 if ((flags & COMMENT) != 0) {
927 while (get_char() != 0) /* null */ ;
930 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
935 if (method >= 0) return method;
938 fprintf(stderr, "\nnot in gzip format\n");
942 WARN((stderr, "\ndecompression OK, trailing garbage ignored\n"));
947 /* ========================================================================
948 * Signal and error handler.
950 RETSIGTYPE abort_gzip()
954 /* unzip.c -- decompress files in gzip or pkzip format.
955 * Copyright (C) 1992-1993 Jean-loup Gailly
956 * This is free software; you can redistribute it and/or modify it under the
957 * terms of the GNU General Public License, see the file COPYING.
959 * The code in this file is derived from the file funzip.c written
960 * and put in the public domain by Mark Adler.
964 This version can extract files in gzip or pkzip format.
965 For the latter, only the first entry is extracted, and it has to be
966 either deflated or stored.
969 /* #include "crypt.h" */
971 /* crypt.h (dummy version) -- do not perform encryption
972 * Hardly worth copyrighting :-)
976 # undef CRYPT /* dummy version */
979 #define RAND_HEAD_LEN 12 /* length of encryption random header */
984 /* PKZIP header definitions */
985 #define LOCSIG 0x04034b50L /* four-byte lead-in (lsb first) */
986 #define LOCFLG 6 /* offset of bit flag */
987 #define CRPFLG 1 /* bit for encrypted entry */
988 #define EXTFLG 8 /* bit for extended local header */
989 #define LOCHOW 8 /* offset of compression method */
990 #define LOCTIM 10 /* file mod time (for decryption) */
991 #define LOCCRC 14 /* offset of crc */
992 #define LOCSIZ 18 /* offset of compressed size */
993 #define LOCLEN 22 /* offset of uncompressed length */
994 #define LOCFIL 26 /* offset of file name field length */
995 #define LOCEXT 28 /* offset of extra field length */
996 #define LOCHDR 30 /* size of local header, including sig */
997 #define EXTHDR 16 /* size of extended local header, inc sig */
1002 char *key; /* not used--needed to link crypt.c */
1003 int pkzip = 0; /* set for a pkzip file */
1004 int ext_header = 0; /* set if extended local header */
1006 /* ===========================================================================
1007 * Unzip in to out. This routine works on both gzip and pkzip files.
1009 * IN assertions: the buffer inbuf contains already the beginning of
1010 * the compressed data, from offsets inptr to insize-1 included.
1011 * The magic header has already been checked. The output buffer is cleared.
1014 int in, out; /* input and output file descriptors */
1016 ulg orig_crc = 0; /* original crc */
1017 ulg orig_len = 0; /* original uncompressed length */
1019 uch buf[EXTHDR]; /* extended local header */
1023 method = get_method(ifd);
1025 do_exit(exit_code); /* error message already emitted */
1028 updcrc(NULL, 0); /* initialize crc */
1030 if (pkzip && !ext_header) { /* crc and length at the end otherwise */
1031 orig_crc = LG(inbuf + LOCCRC);
1032 orig_len = LG(inbuf + LOCLEN);
1036 if (method == DEFLATED) {
1038 int res = inflate();
1041 error("out of memory");
1042 } else if (res != 0) {
1043 error("invalid compressed data--format violated");
1047 error("internal error, invalid method");
1050 /* Get the crc and original length */
1052 /* crc32 (see algorithm.doc)
1053 * uncompressed input size modulo 2^32
1055 for (n = 0; n < 8; n++) {
1056 buf[n] = (uch)get_byte(); /* may cause an error if EOF */
1059 orig_len = LG(buf+4);
1061 } else if (ext_header) { /* If extended header, check it */
1062 /* signature - 4bytes: 0x50 0x4b 0x07 0x08
1064 * compressed size 4-bytes
1065 * uncompressed size 4-bytes
1067 for (n = 0; n < EXTHDR; n++) {
1068 buf[n] = (uch)get_byte(); /* may cause an error if EOF */
1070 orig_crc = LG(buf+4);
1071 orig_len = LG(buf+12);
1074 /* Validate decompression */
1075 if (orig_crc != updcrc(outbuf, 0)) {
1076 error("invalid compressed data--crc error");
1078 if (orig_len != (ulg)bytes_out) {
1079 error("invalid compressed data--length error");
1082 /* Check if there are more entries in a pkzip file */
1083 if (pkzip && inptr + 4 < insize && LG(inbuf+inptr) == LOCSIG) {
1084 WARN((stderr,"has more than one entry--rest ignored\n"));
1086 ext_header = pkzip = 0; /* for next file */
1089 /* util.c -- utility functions for gzip support
1090 * Copyright (C) 1992-1993 Jean-loup Gailly
1091 * This is free software; you can redistribute it and/or modify it under the
1092 * terms of the GNU General Public License, see the file COPYING.
1097 #include <sys/types.h>
1099 #ifdef HAVE_UNISTD_H
1100 # include <unistd.h>
1106 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1107 # include <stdlib.h>
1112 static const ulg crc_32_tab[]; /* crc table, defined below */
1114 /* ===========================================================================
1115 * Run a set of bytes through the crc shift register. If s is a NULL
1116 * pointer, then initialize the crc shift register contents instead.
1117 * Return the current crc in either case.
1120 uch *s; /* pointer to bytes to pump through */
1121 unsigned n; /* number of bytes in s[] */
1123 register ulg c; /* temporary variable */
1125 static ulg crc = (ulg)0xffffffffL; /* shift register contents */
1132 c = crc_32_tab[((int)c ^ (*s++)) & 0xff] ^ (c >> 8);
1136 return c ^ 0xffffffffL; /* (instead of ~c for 64-bit machines) */
1139 /* ===========================================================================
1140 * Clear input and output buffers
1146 bytes_in = bytes_out = 0L;
1149 /* ===========================================================================
1150 * Fill the input buffer. This is called only when the buffer is empty.
1152 int fill_inbuf(eof_ok)
1153 int eof_ok; /* set if EOF acceptable as a result */
1157 /* Read as much as possible */
1161 len = read(ifd, (char*)inbuf+insize, INBUFSIZ-insize);
1162 if (len == 0 || len == EOF) break;
1164 } while (insize < INBUFSIZ);
1167 if (eof_ok) return EOF;
1170 bytes_in += (ulg)insize;
1175 /* ===========================================================================
1176 * Write the output buffer outbuf[0..outcnt-1] and update bytes_out.
1177 * (used for the compressed data only)
1181 if (outcnt == 0) return;
1184 write_buf(ofd, (char *)outbuf, outcnt);
1185 bytes_out += (ulg)outcnt;
1189 /* ===========================================================================
1190 * Write the output window window[0..outcnt-1] and update crc and bytes_out.
1191 * (Used for the decompressed data only.)
1195 if (outcnt == 0) return;
1196 updcrc(window, outcnt);
1199 write_buf(ofd, (char *)window, outcnt);
1200 bytes_out += (ulg)outcnt;
1204 /* ===========================================================================
1205 * Does the same as write(), but also handles partial pipe writes and checks
1208 void write_buf(fd, buf, cnt)
1215 while ((n = write(fd, buf, cnt)) != cnt) {
1216 if (n == (unsigned)(-1)) {
1220 buf = (voidp)((char*)buf+n);
1224 #if defined(NO_STRING_H) && !defined(STDC_HEADERS)
1226 /* Provide missing strspn and strcspn functions. */
1232 int strspn OF((const char *s, const char *accept));
1233 int strcspn OF((const char *s, const char *reject));
1235 /* ========================================================================
1236 * Return the length of the maximum initial segment
1237 * of s which contains only characters in accept.
1239 int strspn(s, accept)
1243 register const char *p;
1244 register const char *a;
1245 register int count = 0;
1247 for (p = s; *p != '\0'; ++p) {
1248 for (a = accept; *a != '\0'; ++a) {
1249 if (*p == *a) break;
1251 if (*a == '\0') return count;
1257 /* ========================================================================
1258 * Return the length of the maximum inital segment of s
1259 * which contains no characters from reject.
1261 int strcspn(s, reject)
1265 register int count = 0;
1267 while (*s != '\0') {
1268 if (strchr(reject, *s++) != NULL) return count;
1274 #endif /* NO_STRING_H */
1277 /* ========================================================================
1281 char *a, *b; /* message strings juxtaposed in output */
1283 WARN((stderr, "warning: %s%s\n", a, b));
1288 fprintf(stderr, "\n");
1292 fprintf(stderr, "unexpected end of file\n");
1299 fprintf(stderr, "\n");
1305 /* ========================================================================
1306 * Table of CRC-32's of all single-byte values (made by makecrc.c)
1308 static const ulg crc_32_tab[] = {
1309 0x00000000L, 0x77073096L, 0xee0e612cL, 0x990951baL, 0x076dc419L,
1310 0x706af48fL, 0xe963a535L, 0x9e6495a3L, 0x0edb8832L, 0x79dcb8a4L,
1311 0xe0d5e91eL, 0x97d2d988L, 0x09b64c2bL, 0x7eb17cbdL, 0xe7b82d07L,
1312 0x90bf1d91L, 0x1db71064L, 0x6ab020f2L, 0xf3b97148L, 0x84be41deL,
1313 0x1adad47dL, 0x6ddde4ebL, 0xf4d4b551L, 0x83d385c7L, 0x136c9856L,
1314 0x646ba8c0L, 0xfd62f97aL, 0x8a65c9ecL, 0x14015c4fL, 0x63066cd9L,
1315 0xfa0f3d63L, 0x8d080df5L, 0x3b6e20c8L, 0x4c69105eL, 0xd56041e4L,
1316 0xa2677172L, 0x3c03e4d1L, 0x4b04d447L, 0xd20d85fdL, 0xa50ab56bL,
1317 0x35b5a8faL, 0x42b2986cL, 0xdbbbc9d6L, 0xacbcf940L, 0x32d86ce3L,
1318 0x45df5c75L, 0xdcd60dcfL, 0xabd13d59L, 0x26d930acL, 0x51de003aL,
1319 0xc8d75180L, 0xbfd06116L, 0x21b4f4b5L, 0x56b3c423L, 0xcfba9599L,
1320 0xb8bda50fL, 0x2802b89eL, 0x5f058808L, 0xc60cd9b2L, 0xb10be924L,
1321 0x2f6f7c87L, 0x58684c11L, 0xc1611dabL, 0xb6662d3dL, 0x76dc4190L,
1322 0x01db7106L, 0x98d220bcL, 0xefd5102aL, 0x71b18589L, 0x06b6b51fL,
1323 0x9fbfe4a5L, 0xe8b8d433L, 0x7807c9a2L, 0x0f00f934L, 0x9609a88eL,
1324 0xe10e9818L, 0x7f6a0dbbL, 0x086d3d2dL, 0x91646c97L, 0xe6635c01L,
1325 0x6b6b51f4L, 0x1c6c6162L, 0x856530d8L, 0xf262004eL, 0x6c0695edL,
1326 0x1b01a57bL, 0x8208f4c1L, 0xf50fc457L, 0x65b0d9c6L, 0x12b7e950L,
1327 0x8bbeb8eaL, 0xfcb9887cL, 0x62dd1ddfL, 0x15da2d49L, 0x8cd37cf3L,
1328 0xfbd44c65L, 0x4db26158L, 0x3ab551ceL, 0xa3bc0074L, 0xd4bb30e2L,
1329 0x4adfa541L, 0x3dd895d7L, 0xa4d1c46dL, 0xd3d6f4fbL, 0x4369e96aL,
1330 0x346ed9fcL, 0xad678846L, 0xda60b8d0L, 0x44042d73L, 0x33031de5L,
1331 0xaa0a4c5fL, 0xdd0d7cc9L, 0x5005713cL, 0x270241aaL, 0xbe0b1010L,
1332 0xc90c2086L, 0x5768b525L, 0x206f85b3L, 0xb966d409L, 0xce61e49fL,
1333 0x5edef90eL, 0x29d9c998L, 0xb0d09822L, 0xc7d7a8b4L, 0x59b33d17L,
1334 0x2eb40d81L, 0xb7bd5c3bL, 0xc0ba6cadL, 0xedb88320L, 0x9abfb3b6L,
1335 0x03b6e20cL, 0x74b1d29aL, 0xead54739L, 0x9dd277afL, 0x04db2615L,
1336 0x73dc1683L, 0xe3630b12L, 0x94643b84L, 0x0d6d6a3eL, 0x7a6a5aa8L,
1337 0xe40ecf0bL, 0x9309ff9dL, 0x0a00ae27L, 0x7d079eb1L, 0xf00f9344L,
1338 0x8708a3d2L, 0x1e01f268L, 0x6906c2feL, 0xf762575dL, 0x806567cbL,
1339 0x196c3671L, 0x6e6b06e7L, 0xfed41b76L, 0x89d32be0L, 0x10da7a5aL,
1340 0x67dd4accL, 0xf9b9df6fL, 0x8ebeeff9L, 0x17b7be43L, 0x60b08ed5L,
1341 0xd6d6a3e8L, 0xa1d1937eL, 0x38d8c2c4L, 0x4fdff252L, 0xd1bb67f1L,
1342 0xa6bc5767L, 0x3fb506ddL, 0x48b2364bL, 0xd80d2bdaL, 0xaf0a1b4cL,
1343 0x36034af6L, 0x41047a60L, 0xdf60efc3L, 0xa867df55L, 0x316e8eefL,
1344 0x4669be79L, 0xcb61b38cL, 0xbc66831aL, 0x256fd2a0L, 0x5268e236L,
1345 0xcc0c7795L, 0xbb0b4703L, 0x220216b9L, 0x5505262fL, 0xc5ba3bbeL,
1346 0xb2bd0b28L, 0x2bb45a92L, 0x5cb36a04L, 0xc2d7ffa7L, 0xb5d0cf31L,
1347 0x2cd99e8bL, 0x5bdeae1dL, 0x9b64c2b0L, 0xec63f226L, 0x756aa39cL,
1348 0x026d930aL, 0x9c0906a9L, 0xeb0e363fL, 0x72076785L, 0x05005713L,
1349 0x95bf4a82L, 0xe2b87a14L, 0x7bb12baeL, 0x0cb61b38L, 0x92d28e9bL,
1350 0xe5d5be0dL, 0x7cdcefb7L, 0x0bdbdf21L, 0x86d3d2d4L, 0xf1d4e242L,
1351 0x68ddb3f8L, 0x1fda836eL, 0x81be16cdL, 0xf6b9265bL, 0x6fb077e1L,
1352 0x18b74777L, 0x88085ae6L, 0xff0f6a70L, 0x66063bcaL, 0x11010b5cL,
1353 0x8f659effL, 0xf862ae69L, 0x616bffd3L, 0x166ccf45L, 0xa00ae278L,
1354 0xd70dd2eeL, 0x4e048354L, 0x3903b3c2L, 0xa7672661L, 0xd06016f7L,
1355 0x4969474dL, 0x3e6e77dbL, 0xaed16a4aL, 0xd9d65adcL, 0x40df0b66L,
1356 0x37d83bf0L, 0xa9bcae53L, 0xdebb9ec5L, 0x47b2cf7fL, 0x30b5ffe9L,
1357 0xbdbdf21cL, 0xcabac28aL, 0x53b39330L, 0x24b4a3a6L, 0xbad03605L,
1358 0xcdd70693L, 0x54de5729L, 0x23d967bfL, 0xb3667a2eL, 0xc4614ab8L,
1359 0x5d681b02L, 0x2a6f2b94L, 0xb40bbe37L, 0xc30c8ea1L, 0x5a05df1bL,
1362 /* inflate.c -- Not copyrighted 1992 by Mark Adler
1363 version c10p1, 10 January 1993 */
1365 /* You can do whatever you like with this source file, though I would
1366 prefer that if you modify it and redistribute it that you include
1367 comments to that effect with your name and the date. Thank you.
1368 [The history has been moved to the file ChangeLog.]
1372 Inflate deflated (PKZIP's method 8 compressed) data. The compression
1373 method searches for as much of the current string of bytes (up to a
1374 length of 258) in the previous 32K bytes. If it doesn't find any
1375 matches (of at least length 3), it codes the next byte. Otherwise, it
1376 codes the length of the matched string and its distance backwards from
1377 the current position. There is a single Huffman code that codes both
1378 single bytes (called "literals") and match lengths. A second Huffman
1379 code codes the distance information, which follows a length code. Each
1380 length or distance code actually represents a base value and a number
1381 of "extra" (sometimes zero) bits to get to add to the base value. At
1382 the end of each deflated block is a special end-of-block (EOB) literal/
1383 length code. The decoding process is basically: get a literal/length
1384 code; if EOB then done; if a literal, emit the decoded byte; if a
1385 length then get the distance and emit the referred-to bytes from the
1386 sliding window of previously emitted data.
1388 There are (currently) three kinds of inflate blocks: stored, fixed, and
1389 dynamic. The compressor deals with some chunk of data at a time, and
1390 decides which method to use on a chunk-by-chunk basis. A chunk might
1391 typically be 32K or 64K. If the chunk is uncompressible, then the
1392 "stored" method is used. In this case, the bytes are simply stored as
1393 is, eight bits per byte, with none of the above coding. The bytes are
1394 preceded by a count, since there is no longer an EOB code.
1396 If the data is compressible, then either the fixed or dynamic methods
1397 are used. In the dynamic method, the compressed data is preceded by
1398 an encoding of the literal/length and distance Huffman codes that are
1399 to be used to decode this block. The representation is itself Huffman
1400 coded, and so is preceded by a description of that code. These code
1401 descriptions take up a little space, and so for small blocks, there is
1402 a predefined set of codes, called the fixed codes. The fixed method is
1403 used if the block codes up smaller that way (usually for quite small
1404 chunks), otherwise the dynamic method is used. In the latter case, the
1405 codes are customized to the probabilities in the current block, and so
1406 can code it much better than the pre-determined fixed codes.
1408 The Huffman codes themselves are decoded using a mutli-level table
1409 lookup, in order to maximize the speed of decoding plus the speed of
1410 building the decoding tables. See the comments below that precede the
1411 lbits and dbits tuning parameters.
1416 Notes beyond the 1.93a appnote.txt:
1418 1. Distance pointers never point before the beginning of the output
1420 2. Distance pointers can point back across blocks, up to 32k away.
1421 3. There is an implied maximum of 7 bits for the bit length table and
1422 15 bits for the actual data.
1423 4. If only one code exists, then it is encoded using one bit. (Zero
1424 would be more efficient, but perhaps a little confusing.) If two
1425 codes exist, they are coded using one bit each (0 and 1).
1426 5. There is no way of sending zero distance codes--a dummy must be
1427 sent if there are none. (History: a pre 2.0 version of PKZIP would
1428 store blocks with no distance codes, but this was discovered to be
1429 too harsh a criterion.) Valid only for 1.93a. 2.04c does allow
1430 zero distance codes, which is sent as one code of zero bits in
1432 6. There are up to 286 literal/length codes. Code 256 represents the
1433 end-of-block. Note however that the static length tree defines
1434 288 codes just to fill out the Huffman codes. Codes 286 and 287
1435 cannot be used though, since there is no length base or extra bits
1436 defined for them. Similarly, there are up to 30 distance codes.
1437 However, static trees define 32 codes (all 5 bits) to fill out the
1438 Huffman codes, but the last two had better not show up in the data.
1439 7. Unzip can check dynamic Huffman blocks for complete code sets.
1440 The exception is that a single code would not be complete (see #4).
1441 8. The five bits following the block type is really the number of
1442 literal codes sent minus 257.
1443 9. Length codes 8,16,16 are interpreted as 13 length codes of 8 bits
1444 (1+6+6). Therefore, to output three times the length, you output
1445 three codes (1+1+1), whereas to output four times the same length,
1446 you only need two codes (1+3). Hmm.
1447 10. In the tree reconstruction algorithm, Code = Code + Increment
1448 only if BitLength(i) is not zero. (Pretty obvious.)
1449 11. Correction: 4 Bits: # of Bit Length codes - 4 (4 - 19)
1450 12. Note: length code 284 can represent 227-258, but length code 285
1451 really is 258. The last length deserves its own, short code
1452 since it gets used a lot in very redundant files. The length
1453 258 is special since 258 - 3 (the min match length) is 255.
1454 13. The literal/length and distance code bit lengths are read as a
1455 single stream of lengths. It is possible (and advantageous) for
1456 a repeat code (16, 17, or 18) to go across the boundary between
1457 the two sets of lengths.
1460 #include <sys/types.h>
1462 #if defined(STDC_HEADERS) || !defined(NO_STDLIB_H)
1463 # include <stdlib.h>
1467 #define slide window
1469 /* Huffman code lookup table entry--this entry is four bytes for machines
1470 that have 16-bit pointers (e.g. PC's in the small or medium model).
1471 Valid extra bits are 0..13. e == 15 is EOB (end of block), e == 16
1472 means that v is a literal, 16 < e < 32 means that v is a pointer to
1473 the next table, which codes e - 16 bits, and lastly e == 99 indicates
1474 an unused code. If a code with e == 99 is looked up, this implies an
1475 error in the data. */
1477 uch e; /* number of extra bits or operation */
1478 uch b; /* number of bits in this code or subcode */
1480 ush n; /* literal, length base, or distance base */
1481 struct huft *t; /* pointer to next level of table */
1486 /* Function prototypes */
1487 int huft_build OF((unsigned *, unsigned, unsigned, ush *, ush *,
1488 struct huft **, int *));
1489 int huft_free OF((struct huft *));
1490 int inflate_codes OF((struct huft *, struct huft *, int, int));
1491 int inflate_stored OF((void));
1492 int inflate_fixed OF((void));
1493 int inflate_dynamic OF((void));
1494 int inflate_block OF((int *));
1495 int inflate OF((void));
1498 /* The inflate algorithm uses a sliding 32K byte window on the uncompressed
1499 stream to find repeated byte strings. This is implemented here as a
1500 circular buffer. The index is updated simply by incrementing and then
1501 and'ing with 0x7fff (32K-1). */
1502 /* It is left to other modules to supply the 32K area. It is assumed
1503 to be usable as if it were declared "uch slide[32768];" or as just
1504 "uch *slide;" and then malloc'ed in the latter case. The definition
1505 must be in unzip.h, included above. */
1506 /* unsigned wp; current position in slide */
1508 #define flush_output(w) (wp=(w),flush_window())
1510 /* Tables for deflate from PKZIP's appnote.txt. */
1511 static unsigned border[] = { /* Order of the bit length code lengths */
1512 16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
1513 static ush cplens[] = { /* Copy lengths for literal codes 257..285 */
1514 3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
1515 35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
1516 /* note: see note #13 above about the 258 in this list. */
1517 static ush cplext[] = { /* Extra bits for literal codes 257..285 */
1518 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
1519 3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99}; /* 99==invalid */
1520 static ush cpdist[] = { /* Copy offsets for distance codes 0..29 */
1521 1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
1522 257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
1523 8193, 12289, 16385, 24577};
1524 static ush cpdext[] = { /* Extra bits for distance codes */
1525 0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
1526 7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
1531 /* Macros for inflate() bit peeking and grabbing.
1535 x = b & mask_bits[j];
1538 where NEEDBITS makes sure that b has at least j bits in it, and
1539 DUMPBITS removes the bits from b. The macros use the variable k
1540 for the number of bits in b. Normally, b and k are register
1541 variables for speed, and are initialized at the beginning of a
1542 routine that uses these macros from a global bit buffer and count.
1544 If we assume that EOB will be the longest code, then we will never
1545 ask for bits with NEEDBITS that are beyond the end of the stream.
1546 So, NEEDBITS should not read any more bytes than are needed to
1547 meet the request. Then no bytes need to be "returned" to the buffer
1548 at the end of the last block.
1550 However, this assumption is not true for fixed blocks--the EOB code
1551 is 7 bits, but the other literal/length codes can be 8 or 9 bits.
1552 (The EOB code is shorter than other codes because fixed blocks are
1553 generally short. So, while a block always has an EOB, many other
1554 literal/length codes have a significantly lower probability of
1555 showing up at all.) However, by making the first table have a
1556 lookup of seven bits, the EOB code will be found in that first
1557 lookup, and so will not require that too many bits be pulled from
1561 ulg bb; /* bit buffer */
1562 unsigned bk; /* bits in bit buffer */
1566 0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
1567 0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
1572 # define NEXTBYTE() (cc = get_byte(), zdecode(cc), cc)
1574 # define NEXTBYTE() (uch)get_byte()
1576 #define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
1577 #define DUMPBITS(n) {b>>=(n);k-=(n);}
1581 Huffman code decoding is performed using a multi-level table lookup.
1582 The fastest way to decode is to simply build a lookup table whose
1583 size is determined by the longest code. However, the time it takes
1584 to build this table can also be a factor if the data being decoded
1585 is not very long. The most common codes are necessarily the
1586 shortest codes, so those codes dominate the decoding time, and hence
1587 the speed. The idea is you can have a shorter table that decodes the
1588 shorter, more probable codes, and then point to subsidiary tables for
1589 the longer codes. The time it costs to decode the longer codes is
1590 then traded against the time it takes to make longer tables.
1592 This results of this trade are in the variables lbits and dbits
1593 below. lbits is the number of bits the first level table for literal/
1594 length codes can decode in one step, and dbits is the same thing for
1595 the distance codes. Subsequent tables are also less than or equal to
1596 those sizes. These values may be adjusted either when all of the
1597 codes are shorter than that, in which case the longest code length in
1598 bits is used, or when the shortest code is *longer* than the requested
1599 table size, in which case the length of the shortest code in bits is
1602 There are two different values for the two tables, since they code a
1603 different number of possibilities each. The literal/length table
1604 codes 286 possible values, or in a flat code, a little over eight
1605 bits. The distance table codes 30 possible values, or a little less
1606 than five bits, flat. The optimum values for speed end up being
1607 about one bit more than those, so lbits is 8+1 and dbits is 5+1.
1608 The optimum values may differ though from machine to machine, and
1609 possibly even between compilers. Your mileage may vary.
1613 int lbits = 9; /* bits in base literal/length lookup table */
1614 int dbits = 6; /* bits in base distance lookup table */
1617 /* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
1618 #define BMAX 16 /* maximum bit length of any code (16 for explode) */
1619 #define N_MAX 288 /* maximum number of codes in any set */
1622 unsigned hufts; /* track memory usage */
1625 int huft_build(b, n, s, d, e, t, m)
1626 unsigned *b; /* code lengths in bits (all assumed <= BMAX) */
1627 unsigned n; /* number of codes (assumed <= N_MAX) */
1628 unsigned s; /* number of simple-valued codes (0..s-1) */
1629 ush *d; /* list of base values for non-simple codes */
1630 ush *e; /* list of extra bits for non-simple codes */
1631 struct huft **t; /* result: starting table */
1632 int *m; /* maximum lookup bits, returns actual */
1633 /* Given a list of code lengths and a maximum table size, make a set of
1634 tables to decode that set of codes. Return zero on success, one if
1635 the given code set is incomplete (the tables are still built in this
1636 case), two if the input is invalid (all zero length codes or an
1637 oversubscribed set of lengths), and three if not enough memory. */
1639 unsigned a; /* counter for codes of length k */
1640 unsigned c[BMAX+1]; /* bit length count table */
1641 unsigned f; /* i repeats in table every f entries */
1642 int g; /* maximum code length */
1643 int h; /* table level */
1644 register unsigned i; /* counter, current code */
1645 register unsigned j; /* counter */
1646 register int k; /* number of bits in current code */
1647 int l; /* bits per table (returned in m) */
1648 register unsigned *p; /* pointer into c[], b[], or v[] */
1649 register struct huft *q; /* points to current table */
1650 struct huft r; /* table entry for structure assignment */
1651 struct huft *u[BMAX]; /* table stack */
1652 unsigned v[N_MAX]; /* values in order of bit length */
1653 register int w; /* bits before this table == (l * h) */
1654 unsigned x[BMAX+1]; /* bit offsets, then code stack */
1655 unsigned *xp; /* pointer into x */
1656 int y; /* number of dummy codes added */
1657 unsigned z; /* number of entries in current table */
1660 /* Generate counts for each bit length */
1661 memzero(c, sizeof(c));
1664 Tracecv(*p, (stderr, (n-i >= ' ' && n-i <= '~' ? "%c %d\n" : "0x%x %d\n"),
1666 c[*p]++; /* assume all entries <= BMAX */
1667 p++; /* Can't combine with above line (Solaris bug) */
1669 if (c[0] == n) /* null input--all zero length codes */
1671 *t = (struct huft *)NULL;
1677 /* Find minimum and maximum length, bound *m by those */
1679 for (j = 1; j <= BMAX; j++)
1682 k = j; /* minimum code length */
1683 if ((unsigned)l < j)
1685 for (i = BMAX; i; i--)
1688 g = i; /* maximum code length */
1689 if ((unsigned)l > i)
1694 /* Adjust last length count to fill out codes, if needed */
1695 for (y = 1 << j; j < i; j++, y <<= 1)
1696 if ((y -= c[j]) < 0)
1697 return 2; /* bad input: more codes than bits */
1698 if ((y -= c[i]) < 0)
1703 /* Generate starting offsets into the value table for each length */
1705 p = c + 1; xp = x + 2;
1706 while (--i) { /* note that i == g from above */
1707 *xp++ = (j += *p++);
1711 /* Make a table of values in order of bit lengths */
1714 if ((j = *p++) != 0)
1719 /* Generate the Huffman codes and for each, make the table entries */
1720 x[0] = i = 0; /* first Huffman code is zero */
1721 p = v; /* grab values in bit order */
1722 h = -1; /* no tables yet--level -1 */
1723 w = -l; /* bits decoded == (l * h) */
1724 u[0] = (struct huft *)NULL; /* just to keep compilers happy */
1725 q = (struct huft *)NULL; /* ditto */
1728 /* go through the bit lengths (k already is bits in shortest code) */
1734 /* here i is the Huffman code of length k bits for value *p */
1735 /* make tables up to required level */
1739 w += l; /* previous table always l bits */
1741 /* compute minimum size table less than or equal to l bits */
1742 z = (z = g - w) > (unsigned)l ? l : z; /* upper limit on table size */
1743 if ((f = 1 << (j = k - w)) > a + 1) /* try a k-w bit table */
1744 { /* too few codes for k-w bit table */
1745 f -= a + 1; /* deduct codes from patterns left */
1747 while (++j < z) /* try smaller tables up to z bits */
1749 if ((f <<= 1) <= *++xp)
1750 break; /* enough codes to use up j bits */
1751 f -= *xp; /* else deduct codes from patterns */
1754 z = 1 << j; /* table entries for j-bit table */
1756 /* allocate and link in new table */
1757 if ((q = (struct huft *)malloc((z + 1)*sizeof(struct huft))) ==
1758 (struct huft *)NULL)
1762 return 3; /* not enough memory */
1764 hufts += z + 1; /* track memory usage */
1765 *t = q + 1; /* link to list for huft_free() */
1766 *(t = &(q->v.t)) = (struct huft *)NULL;
1767 u[h] = ++q; /* table starts after link */
1769 /* connect to last table, if there is one */
1772 x[h] = i; /* save pattern for backing up */
1773 r.b = (uch)l; /* bits to dump before this table */
1774 r.e = (uch)(16 + j); /* bits in this table */
1775 r.v.t = q; /* pointer to this table */
1776 j = i >> (w - l); /* (get around Turbo C bug) */
1777 u[h-1][j] = r; /* connect to last table */
1781 /* set up table entry in r */
1784 r.e = 99; /* out of values--invalid code */
1787 r.e = (uch)(*p < 256 ? 16 : 15); /* 256 is end-of-block code */
1788 r.v.n = (ush)(*p); /* simple code is just the value */
1789 p++; /* one compiler does not like *p++ */
1793 r.e = (uch)e[*p - s]; /* non-simple--look up in lists */
1794 r.v.n = d[*p++ - s];
1797 /* fill code-like entries with r */
1799 for (j = i >> w; j < z; j += f)
1802 /* backwards increment the k-bit code i */
1803 for (j = 1 << (k - 1); i & j; j >>= 1)
1807 /* backup over finished tables */
1808 while ((i & ((1 << w) - 1)) != x[h])
1810 h--; /* don't need to update q */
1817 /* Return true (1) if we were given an incomplete table */
1818 return y != 0 && g != 1;
1824 struct huft *t; /* table to free */
1825 /* Free the malloc'ed tables built by huft_build(), which makes a linked
1826 list of the tables it made, with the links in a dummy first entry of
1829 register struct huft *p, *q;
1832 /* Go through linked list, freeing from the malloced (t[-1]) address. */
1834 while (p != (struct huft *)NULL)
1844 int inflate_codes(tl, td, bl, bd)
1845 struct huft *tl, *td; /* literal/length and distance decoder tables */
1846 int bl, bd; /* number of bits decoded by tl[] and td[] */
1847 /* inflate (decompress) the codes in a deflated (compressed) block.
1848 Return an error code or zero if it all goes ok. */
1850 register unsigned e; /* table entry flag/number of extra bits */
1851 unsigned n, d; /* length and index for copy */
1852 unsigned w; /* current window position */
1853 struct huft *t; /* pointer to table entry */
1854 unsigned ml, md; /* masks for bl and bd bits */
1855 register ulg b; /* bit buffer */
1856 register unsigned k; /* number of bits in bit buffer */
1859 /* make local copies of globals */
1860 b = bb; /* initialize bit buffer */
1862 w = wp; /* initialize window position */
1864 /* inflate the coded data */
1865 ml = mask_bits[bl]; /* precompute masks for speed */
1867 for (;;) /* do until end of block */
1869 NEEDBITS((unsigned)bl)
1870 if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
1877 } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
1879 if (e == 16) /* then it's a literal */
1881 slide[w++] = (uch)t->v.n;
1882 Tracevv((stderr, "%c", slide[w-1]));
1889 else /* it's an EOB or a length */
1891 /* exit if end of block */
1895 /* get length of block to copy */
1897 n = t->v.n + ((unsigned)b & mask_bits[e]);
1900 /* decode distance of block to copy */
1901 NEEDBITS((unsigned)bd)
1902 if ((e = (t = td + ((unsigned)b & md))->e) > 16)
1909 } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
1912 d = w - t->v.n - ((unsigned)b & mask_bits[e]);
1914 Tracevv((stderr,"\\[%d,%d]", w-d, n));
1918 n -= (e = (e = WSIZE - ((d &= WSIZE-1) > w ? d : w)) > n ? n : e);
1919 #if !defined(NOMEMCPY) && !defined(DEBUG)
1920 if (w - d >= e) /* (this test assumes unsigned comparison) */
1922 memcpy(slide + w, slide + d, e);
1926 else /* do it slow to avoid memcpy() overlap */
1927 #endif /* !NOMEMCPY */
1929 slide[w++] = slide[d++];
1930 Tracevv((stderr, "%c", slide[w-1]));
1942 /* restore the globals from the locals */
1943 wp = w; /* restore global window pointer */
1944 bb = b; /* restore global bit buffer */
1953 int inflate_stored()
1954 /* "decompress" an inflated type 0 (stored) block. */
1956 unsigned n; /* number of bytes in block */
1957 unsigned w; /* current window position */
1958 register ulg b; /* bit buffer */
1959 register unsigned k; /* number of bits in bit buffer */
1962 /* make local copies of globals */
1963 b = bb; /* initialize bit buffer */
1965 w = wp; /* initialize window position */
1968 /* go to byte boundary */
1973 /* get the length and its complement */
1975 n = ((unsigned)b & 0xffff);
1978 if (n != (unsigned)((~b) & 0xffff))
1979 return 1; /* error in compressed data */
1983 /* read and output the compressed data */
1987 slide[w++] = (uch)b;
1997 /* restore the globals from the locals */
1998 wp = w; /* restore global window pointer */
1999 bb = b; /* restore global bit buffer */
2007 /* decompress an inflated type 1 (fixed Huffman codes) block. We should
2008 either replace this with a custom decoder, or at least precompute the
2011 int i; /* temporary variable */
2012 struct huft *tl; /* literal/length code table */
2013 struct huft *td; /* distance code table */
2014 int bl; /* lookup bits for tl */
2015 int bd; /* lookup bits for td */
2016 unsigned l[288]; /* length list for huft_build */
2019 /* set up literal table */
2020 for (i = 0; i < 144; i++)
2022 for (; i < 256; i++)
2024 for (; i < 280; i++)
2026 for (; i < 288; i++) /* make a complete, but wrong code set */
2029 if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
2033 /* set up distance table */
2034 for (i = 0; i < 30; i++) /* make an incomplete code set */
2037 if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1)
2044 /* decompress until an end-of-block code */
2045 if (inflate_codes(tl, td, bl, bd))
2049 /* free the decoding tables, return */
2057 int inflate_dynamic()
2058 /* decompress an inflated type 2 (dynamic Huffman codes) block. */
2060 int i; /* temporary variables */
2062 unsigned l; /* last length */
2063 unsigned m; /* mask for bit lengths table */
2064 unsigned n; /* number of lengths to get */
2065 struct huft *tl; /* literal/length code table */
2066 struct huft *td; /* distance code table */
2067 int bl; /* lookup bits for tl */
2068 int bd; /* lookup bits for td */
2069 unsigned nb; /* number of bit length codes */
2070 unsigned nl; /* number of literal/length codes */
2071 unsigned nd; /* number of distance codes */
2072 #ifdef PKZIP_BUG_WORKAROUND
2073 unsigned ll[288+32]; /* literal/length and distance code lengths */
2075 unsigned ll[286+30]; /* literal/length and distance code lengths */
2077 register ulg b; /* bit buffer */
2078 register unsigned k; /* number of bits in bit buffer */
2081 /* make local bit buffer */
2086 /* read in table lengths */
2088 nl = 257 + ((unsigned)b & 0x1f); /* number of literal/length codes */
2091 nd = 1 + ((unsigned)b & 0x1f); /* number of distance codes */
2094 nb = 4 + ((unsigned)b & 0xf); /* number of bit length codes */
2096 #ifdef PKZIP_BUG_WORKAROUND
2097 if (nl > 288 || nd > 32)
2099 if (nl > 286 || nd > 30)
2101 return 1; /* bad lengths */
2104 /* read in bit-length-code lengths */
2105 for (j = 0; j < nb; j++)
2108 ll[border[j]] = (unsigned)b & 7;
2115 /* build decoding table for trees--single level, 7 bit lookup */
2117 if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0)
2121 return i; /* incomplete code set */
2125 /* read in literal and distance code lengths */
2129 while ((unsigned)i < n)
2131 NEEDBITS((unsigned)bl)
2132 j = (td = tl + ((unsigned)b & m))->b;
2135 if (j < 16) /* length of code in bits (0..15) */
2136 ll[i++] = l = j; /* save last length in l */
2137 else if (j == 16) /* repeat last length 3 to 6 times */
2140 j = 3 + ((unsigned)b & 3);
2142 if ((unsigned)i + j > n)
2147 else if (j == 17) /* 3 to 10 zero length codes */
2150 j = 3 + ((unsigned)b & 7);
2152 if ((unsigned)i + j > n)
2158 else /* j == 18: 11 to 138 zero length codes */
2161 j = 11 + ((unsigned)b & 0x7f);
2163 if ((unsigned)i + j > n)
2172 /* free decoding table for trees */
2176 /* restore the global bit buffer */
2181 /* build the decoding tables for literal/length and distance codes */
2183 if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0)
2186 fprintf(stderr, " incomplete literal tree\n");
2189 return i; /* incomplete code set */
2192 if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0)
2195 fprintf(stderr, " incomplete distance tree\n");
2196 #ifdef PKZIP_BUG_WORKAROUND
2203 return i; /* incomplete code set */
2208 /* decompress until an end-of-block code */
2209 if (inflate_codes(tl, td, bl, bd))
2213 /* free the decoding tables, return */
2221 int inflate_block(e)
2222 int *e; /* last block flag */
2223 /* decompress an inflated block */
2225 unsigned t; /* block type */
2226 register ulg b; /* bit buffer */
2227 register unsigned k; /* number of bits in bit buffer */
2230 /* make local bit buffer */
2235 /* read in last block bit */
2241 /* read in block type */
2243 t = (unsigned)b & 3;
2247 /* restore the global bit buffer */
2252 /* inflate that block type */
2254 return inflate_dynamic();
2256 return inflate_stored();
2258 return inflate_fixed();
2261 /* bad block type */
2268 /* decompress an inflated entry */
2270 int e; /* last block flag */
2271 int r; /* result code */
2272 unsigned h; /* maximum struct huft's malloc'ed */
2275 /* initialize window, bit buffer */
2281 /* decompress until the last block */
2285 if ((r = inflate_block(&e)) != 0)
2291 /* Undo too much lookahead. The next read will be byte aligned so we
2292 * can discard unused bits in the last meaningful byte.
2299 /* flush out slide */
2303 /* return success */
2305 fprintf(stderr, "<%u> ", h);