1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001, 2002, 2006 Free Software Foundation, Inc.
4 Copyright (C) 1992-1993 Jean-loup Gailly
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
21 * The unzip code was written and put in the public domain by Mark Adler.
22 * Portions of the lzw code are derived from the public domain 'compress'
23 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
24 * Ken Turkowski, Dave Mack and Peter Jannesen.
26 * See the license_msg below and the file COPYING for the software license.
27 * See the file algorithm.doc for the compression algorithms and file formats.
30 static char *license_msg[] = {
31 "Copyright (C) 2006 Free Software Foundation, Inc.",
32 "Copyright (C) 1993 Jean-loup Gailly.",
33 "This is free software. You may redistribute copies of it under the terms of",
34 "the GNU General Public License <http://www.gnu.org/licenses/gpl.html>.",
35 "There is NO WARRANTY, to the extent permitted by law.",
38 /* Compress files with zip algorithm and 'compress' interface.
39 * See help() function below for all options.
41 * file.gz: compressed file with same mode, owner, and utimes
42 * or stdout with -c option or if stdin used as input.
43 * If the output file name had to be truncated, the original name is kept
44 * in the compressed file.
45 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
47 * Using gz on MSDOS would create too many file name conflicts. For
48 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
49 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
50 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
51 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
53 * For the meaning of all compilation flags, see comments in Makefile.in.
57 static char rcsid[] = "$Id$";
62 #include <sys/types.h>
72 #include "fcntl-safer.h"
75 #include "stat-macros.h"
76 #include "stat-time.h"
83 # include <sys/time.h>
98 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
109 # ifndef _D_EXACT_NAMLEN
110 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
115 # define CLOSEDIR(d) (closedir(d), 0)
117 # define CLOSEDIR(d) closedir(d)
120 #if !defined(HAVE_LSTAT) && !defined(lstat)
121 # define lstat(name, buf) stat(name, buf)
125 # include <utimens.h>
128 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
131 # define MAX_PATH_LEN 1024 /* max pathname length */
143 off_t lseek OF((int fd, off_t offset, int whence));
147 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
151 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
155 # define ELOOP EINVAL
158 /* Separator for file name parts (see shorten_name()) */
159 #ifdef NO_MULTIPLE_DOTS
160 # define PART_SEP "-"
162 # define PART_SEP "."
167 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
168 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
169 DECLARE(ush, d_buf, DIST_BUFSIZE);
170 DECLARE(uch, window, 2L*WSIZE);
172 DECLARE(ush, tab_prefix, 1L<<BITS);
174 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
175 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
178 /* local variables */
180 int ascii = 0; /* convert end-of-lines to local OS conventions */
181 int to_stdout = 0; /* output to stdout (-c) */
182 int decompress = 0; /* decompress (-d) */
183 int force = 0; /* don't ask questions, compress links (-f) */
184 int no_name = -1; /* don't save or restore the original file name */
185 int no_time = -1; /* don't save or restore the original file time */
186 int recursive = 0; /* recurse through directories (-r) */
187 int list = 0; /* list the file contents (-l) */
188 int verbose = 0; /* be verbose (-v) */
189 int quiet = 0; /* be very quiet (-q) */
190 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
191 int test = 0; /* test .gz file integrity */
192 int foreground; /* set if program run in foreground */
193 char *program_name; /* program name */
194 int maxbits = BITS; /* max bits per code for LZW */
195 int method = DEFLATED;/* compression method */
196 int level = 6; /* compression level */
197 int exit_code = OK; /* program exit code */
198 int save_orig_name; /* set if original name must be saved */
199 int last_member; /* set for .zip and .Z files */
200 int part_nb; /* number of parts in .gz file */
201 struct timespec time_stamp; /* original time stamp (modification time) */
202 off_t ifile_size; /* input file size, -1 for devices (debug only) */
203 char *env; /* contents of GZIP env variable */
204 char **args = NULL; /* argv pointer if GZIP env variable defined */
205 char *z_suffix; /* default suffix (can be set with --suffix) */
206 size_t z_len; /* strlen(z_suffix) */
208 off_t bytes_in; /* number of input bytes */
209 off_t bytes_out; /* number of output bytes */
210 off_t total_in; /* input bytes for all files */
211 off_t total_out; /* output bytes for all files */
212 char ifname[MAX_PATH_LEN]; /* input file name */
213 char ofname[MAX_PATH_LEN]; /* output file name */
214 int volatile remove_ofname = 0; /* remove output file on error */
215 struct stat istat; /* status for input file */
216 int ifd; /* input file descriptor */
217 int ofd; /* output file descriptor */
218 unsigned insize; /* valid bytes in inbuf */
219 unsigned inptr; /* index of next byte to be processed in inbuf */
220 unsigned outcnt; /* bytes in output buffer */
222 struct option longopts[] =
224 /* { name has_arg *flag val } */
225 {"ascii", 0, 0, 'a'}, /* ascii text mode */
226 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
227 {"stdout", 0, 0, 'c'}, /* write output on standard output */
228 {"decompress", 0, 0, 'd'}, /* decompress */
229 {"uncompress", 0, 0, 'd'}, /* decompress */
230 /* {"encrypt", 0, 0, 'e'}, encrypt */
231 {"force", 0, 0, 'f'}, /* force overwrite of output file */
232 {"help", 0, 0, 'h'}, /* give help */
233 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
234 {"list", 0, 0, 'l'}, /* list .gz file contents */
235 {"license", 0, 0, 'L'}, /* display software license */
236 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
237 {"name", 0, 0, 'N'}, /* save or restore original name & time */
238 {"quiet", 0, 0, 'q'}, /* quiet mode */
239 {"silent", 0, 0, 'q'}, /* quiet mode */
240 {"recursive", 0, 0, 'r'}, /* recurse through directories */
241 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
242 {"test", 0, 0, 't'}, /* test compressed file integrity */
243 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
244 {"verbose", 0, 0, 'v'}, /* verbose mode */
245 {"version", 0, 0, 'V'}, /* display version number */
246 {"fast", 0, 0, '1'}, /* compress faster */
247 {"best", 0, 0, '9'}, /* compress better */
248 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
249 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
253 /* local functions */
255 local void try_help OF((void)) ATTRIBUTE_NORETURN;
256 local void help OF((void));
257 local void license OF((void));
258 local void version OF((void));
259 local int input_eof OF((void));
260 local void treat_stdin OF((void));
261 local void treat_file OF((char *iname));
262 local int create_outfile OF((void));
263 local char *get_suffix OF((char *name));
264 local int open_input_file OF((char *iname, struct stat *sbuf));
265 local int make_ofname OF((void));
266 local int same_file OF((struct stat *stat1, struct stat *stat2));
267 local int name_too_long OF((char *name, struct stat *statb));
268 local void shorten_name OF((char *name));
269 local int get_method OF((int in));
270 local void do_list OF((int ifd, int method));
271 local int check_ofname OF((void));
272 local void copy_stat OF((struct stat *ifstat));
273 local void remove_output_file OF((void));
274 local RETSIGTYPE abort_gzip_signal OF((int));
275 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
276 int main OF((int argc, char **argv));
277 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
280 local void treat_dir OF((int fd, char *dir));
283 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
288 fprintf (stderr, "Try `%s --help' for more information.\n",
293 /* ======================================================================== */
296 static char *help_msg[] = {
297 "Compress or uncompress FILEs (by default, compress FILES in-place).",
299 "Mandatory arguments to long options are mandatory for short options too.",
302 " -a, --ascii ascii text; convert end-of-line using local conventions",
304 " -c, --stdout write on standard output, keep original files unchanged",
305 " -d, --decompress decompress",
306 /* -e, --encrypt encrypt */
307 " -f, --force force overwrite of output file and compress links",
308 " -h, --help give this help",
309 /* -k, --pkzip force output in pkzip format */
310 " -l, --list list compressed file contents",
311 " -L, --license display software license",
313 " -m, --no-time do not save or restore the original modification time",
314 " -M, --time save or restore the original modification time",
316 " -n, --no-name do not save or restore the original name and time stamp",
317 " -N, --name save or restore the original name and time stamp",
318 " -q, --quiet suppress all warnings",
320 " -r, --recursive operate recursively on directories",
322 " -S, --suffix=SUF use suffix SUF on compressed files",
323 " -t, --test test compressed file integrity",
324 " -v, --verbose verbose mode",
325 " -V, --version display version number",
326 " -1, --fast compress faster",
327 " -9, --best compress better",
329 " -Z, --lzw produce output compatible with old compress",
330 " -b, --bits=BITS max number of bits per code (implies -Z)",
333 "With no FILE, or when FILE is -, read standard input.",
335 "Report bugs to <bug-gzip@gnu.org>.",
339 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
340 while (*p) printf ("%s\n", *p++);
343 /* ======================================================================== */
346 char **p = license_msg;
348 printf ("%s %s\n", program_name, VERSION);
349 while (*p) printf ("%s\n", *p++);
352 /* ======================================================================== */
357 printf ("Written by Jean-loup Gailly.\n");
360 local void progerror (string)
364 fprintf (stderr, "%s: ", program_name);
370 /* ======================================================================== */
371 int main (argc, argv)
375 int file_count; /* number of files to process */
376 size_t proglen; /* length of program_name */
377 int optc; /* current option */
379 EXPAND(argc, argv); /* wild card expansion if necessary */
381 program_name = gzip_base_name (argv[0]);
382 proglen = strlen (program_name);
384 /* Suppress .exe for MSDOS, OS/2 and VMS: */
385 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
386 program_name[proglen - 4] = '\0';
388 /* Add options in GZIP environment variable if there is one */
389 env = add_envopt(&argc, &argv, OPTIONS_VAR);
390 if (env != NULL) args = argv;
392 foreground = signal(SIGINT, SIG_IGN) != SIG_IGN;
394 signal (SIGINT, abort_gzip_signal);
397 if (signal(SIGTERM, SIG_IGN) != SIG_IGN) {
398 signal (SIGTERM, abort_gzip_signal);
402 if (signal(SIGHUP, SIG_IGN) != SIG_IGN) {
403 signal (SIGHUP, abort_gzip_signal);
408 /* For compatibility with old compress, use program name as an option.
409 * If you compile with -DGNU_STANDARD, this program will behave as
410 * gzip even if it is invoked under the name gunzip or zcat.
412 * Systems which do not support links can still use -d or -dc.
413 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
415 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
416 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
418 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
419 || strequ (program_name, "gzcat")) /* gzcat */
420 decompress = to_stdout = 1;
424 z_len = strlen(z_suffix);
426 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
427 longopts, (int *)0)) != -1) {
432 maxbits = atoi(optarg);
433 for (; *optarg; optarg++)
434 if (! ('0' <= *optarg && *optarg <= '9'))
436 fprintf (stderr, "%s: -b operand is not an integer\n",
442 to_stdout = 1; break;
444 decompress = 1; break;
448 help(); do_exit(OK); break;
450 list = decompress = to_stdout = 1; break;
452 license(); do_exit(OK); break;
453 case 'm': /* undocumented, may change later */
455 case 'M': /* undocumented, may change later */
458 no_name = no_time = 1; break;
460 no_name = no_time = 0; break;
462 quiet = 1; verbose = 0; break;
465 fprintf (stderr, "%s: -r not supported on this system\n",
473 #ifdef NO_MULTIPLE_DOTS
474 if (*optarg == '.') optarg++;
476 z_len = strlen(optarg);
480 test = decompress = to_stdout = 1;
483 verbose++; quiet = 0; break;
485 version(); do_exit(OK); break;
490 fprintf(stderr, "%s: -Z not supported in this version\n",
495 case '1': case '2': case '3': case '4':
496 case '5': case '6': case '7': case '8': case '9':
500 /* Error message already emitted by getopt_long. */
503 } /* loop on all arguments */
506 /* Ignore "Broken Pipe" message with --quiet */
507 if (quiet && signal (SIGPIPE, SIG_IGN) != SIG_IGN)
508 signal (SIGPIPE, abort_gzip_signal);
511 /* By default, save name and timestamp on compression but do not
512 * restore them on decompression.
514 if (no_time < 0) no_time = decompress;
515 if (no_name < 0) no_name = decompress;
517 file_count = argc - optind;
521 if (ascii && !quiet) {
522 fprintf(stderr, "%s: option --ascii ignored on this system\n",
526 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
527 fprintf(stderr, "%s: incorrect suffix '%s'\n",
528 program_name, z_suffix);
531 if (do_lzw && !decompress) work = lzw;
533 /* Allocate all global buffers (for DYN_ALLOC option) */
534 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
535 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
536 ALLOC(ush, d_buf, DIST_BUFSIZE);
537 ALLOC(uch, window, 2L*WSIZE);
539 ALLOC(ush, tab_prefix, 1L<<BITS);
541 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
542 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
545 /* And get to work */
546 if (file_count != 0) {
547 if (to_stdout && !test && !list && (!decompress || !ascii)) {
548 SET_BINARY_MODE(fileno(stdout));
550 while (optind < argc) {
551 treat_file(argv[optind++]);
553 } else { /* Standard input */
556 if (list && !quiet && file_count > 1) {
557 do_list(-1, -1); /* print totals */
560 return exit_code; /* just to avoid lint warning */
563 /* Return nonzero when at end of file on input. */
567 if (!decompress || last_member)
572 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
575 /* Unget the char that fill_inbuf got. */
582 /* ========================================================================
583 * Compress or decompress stdin
585 local void treat_stdin()
587 if (!force && !list &&
588 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
589 /* Do not send compressed data to the terminal or read it from
590 * the terminal. We get here when user invoked the program
591 * without parameters, so be helpful. According to the GNU standards:
593 * If there is one behavior you think is most useful when the output
594 * is to a terminal, and another that you think is most useful when
595 * the output is a file or a pipe, then it is usually best to make
596 * the default behavior the one that is useful with output to a
597 * terminal, and have an option for the other behavior.
599 * Here we use the --force option to get the other behavior.
602 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
603 program_name, decompress ? "read from" : "written to",
604 decompress ? "de" : "");
605 fprintf (stderr, "For help, type: %s -h\n", program_name);
609 if (decompress || !ascii) {
610 SET_BINARY_MODE(fileno(stdin));
612 if (!test && !list && (!decompress || !ascii)) {
613 SET_BINARY_MODE(fileno(stdout));
615 strcpy(ifname, "stdin");
616 strcpy(ofname, "stdout");
618 /* Get the time stamp on the input file. */
619 time_stamp.tv_nsec = -1; /* The time is unknown by default. */
621 #ifndef NO_STDIN_FSTAT
622 if (list || !no_time) {
623 if (fstat(fileno(stdin), &istat) != 0) {
624 progerror("standard input");
627 # ifdef NO_PIPE_TIMESTAMP
628 if (S_ISREG(istat.st_mode))
630 time_stamp = get_stat_mtime (&istat);
631 #endif /* NO_STDIN_FSTAT */
633 ifile_size = -1L; /* convention for unknown size */
635 clear_bufs(); /* clear input and output buffers */
640 method = get_method(ifd);
642 do_exit(exit_code); /* error message already emitted */
646 do_list(ifd, method);
650 /* Actually do the compression/decompression. Loop over zipped members.
653 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
658 method = get_method(ifd);
659 if (method < 0) return; /* error message already emitted */
660 bytes_out = 0; /* required for length check */
665 fprintf(stderr, " OK\n");
667 } else if (!decompress) {
668 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
669 fprintf(stderr, "\n");
670 #ifdef DISPLAY_STDIN_RATIO
672 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
673 fprintf(stderr, "\n");
679 /* ========================================================================
680 * Compress or decompress the given file
682 local void treat_file(iname)
685 /* Accept "-" as synonym for stdin */
686 if (strequ(iname, "-")) {
687 int cflag = to_stdout;
693 /* Check if the input file is present, set ifname and istat: */
694 ifd = open_input_file (iname, &istat);
698 /* If the input name is that of a directory, recurse or ignore: */
699 if (S_ISDIR(istat.st_mode)) {
702 treat_dir (ifd, iname);
703 /* Warning: ifname is now garbage */
708 WARN ((stderr, "%s: %s is a directory -- ignored\n",
709 program_name, ifname));
712 if (!S_ISREG(istat.st_mode)) {
714 "%s: %s is not a directory or a regular file - ignored\n",
715 program_name, ifname));
719 if (istat.st_nlink > 1 && !to_stdout && !force) {
720 WARN((stderr, "%s: %s has %lu other link%c -- unchanged\n",
721 program_name, ifname, (unsigned long) istat.st_nlink - 1,
722 istat.st_nlink > 2 ? 's' : ' '));
727 ifile_size = istat.st_size;
728 if (no_time && !list)
729 time_stamp.tv_nsec = -1;
731 time_stamp = get_stat_mtime (&istat);
733 /* Generate output file name. For -r and (-t or -l), skip files
734 * without a valid gzip suffix (check done in make_ofname).
736 if (to_stdout && !list && !test) {
737 strcpy(ofname, "stdout");
739 } else if (make_ofname() != OK) {
744 clear_bufs(); /* clear input and output buffers */
748 method = get_method(ifd); /* updates ofname if original given */
751 return; /* error message already emitted */
755 do_list(ifd, method);
760 /* If compressing to a file, check if ofname is not ambiguous
761 * because the operating system truncates names. Otherwise, generate
762 * a new ofname and save the original name in the compressed file.
765 ofd = fileno(stdout);
766 /* keep remove_ofname as zero */
768 if (create_outfile() != OK) return;
770 if (!decompress && save_orig_name && !verbose && !quiet) {
771 fprintf(stderr, "%s: %s compressed to %s\n",
772 program_name, ifname, ofname);
775 /* Keep the name even if not truncated except with --no-name: */
776 if (!save_orig_name) save_orig_name = !no_name;
779 fprintf(stderr, "%s:\t", ifname);
782 /* Actually do the compression/decompression. Loop over zipped members.
785 if ((*work)(ifd, ofd) != OK) {
786 method = -1; /* force cleanup */
793 method = get_method(ifd);
794 if (method < 0) break; /* error message already emitted */
795 bytes_out = 0; /* required for length check */
803 if (close (ofd) != 0)
807 /* It's now safe to remove the input file. */
808 if (xunlink (ifname) != 0)
811 WARN ((stderr, "%s: ", program_name));
821 if (!to_stdout) xunlink (ofname);
824 /* Display statistics */
827 fprintf(stderr, " OK");
828 } else if (decompress) {
829 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
831 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
833 if (!test && !to_stdout) {
834 fprintf(stderr, " -- replaced with %s", ofname);
836 fprintf(stderr, "\n");
840 /* ========================================================================
841 * Create the output file. Return OK or ERROR.
842 * Try several times if necessary to avoid truncating the z_suffix. For
843 * example, do not create a compressed file of name "1234567890123."
844 * Sets save_orig_name to true if the file name has been truncated.
845 * IN assertions: the input file has already been open (ifd is set) and
846 * ofname has already been updated if there was an original name.
847 * OUT assertions: ifd and ofd are closed in case of error.
849 local int create_outfile()
851 struct stat ostat; /* stat for ofname */
852 int flags = O_WRONLY | O_CREAT | O_EXCL | O_BINARY;
854 if (ascii && decompress) {
855 flags &= ~O_BINARY; /* force ascii text mode */
858 /* Make sure that ofname is not an existing file */
859 if (check_ofname() != OK) {
863 /* Create the output file */
865 ofd = OPEN(ofname, flags, RW_USER);
872 /* Check for name truncation on new file (1234567890123.gz) */
874 if (stat(ofname, &ostat) != 0) {
876 if (fstat(ofd, &ostat) != 0) {
879 close(ifd); close(ofd);
883 if (!name_too_long(ofname, &ostat)) return OK;
886 /* name might be too long if an original name was saved */
887 WARN((stderr, "%s: %s: warning, name truncated\n",
888 program_name, ofname));
893 #ifdef NO_MULTIPLE_DOTS
894 /* Should never happen, see check_ofname() */
895 fprintf (stderr, "%s: %s: name too long\n", program_name, ofname);
898 shorten_name(ofname);
902 /* ========================================================================
903 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
904 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
905 * accepted suffixes, in addition to the value of the --suffix option.
906 * ".tgz" is a useful convention for tar.z files on systems limited
907 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
908 * also accepted suffixes. For Unix, we do not want to accept any
909 * .??z suffix as indicating a compressed file; some people use .xyz
910 * to denote volume data.
911 * On systems allowing multiple versions of the same file (such as VMS),
912 * this function removes any version suffix in the given name.
914 local char *get_suffix(name)
918 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
919 static char *known_suffixes[] =
920 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
925 char **suf = known_suffixes;
928 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
931 /* strip a version number from the file name */
933 char *v = strrchr(name, SUFFIX_SEP);
934 if (v != NULL) *v = '\0';
938 if (nlen <= MAX_SUFFIX+2) {
939 strcpy(suffix, name);
941 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
944 slen = strlen(suffix);
946 int s = strlen(*suf);
947 if (slen > s && suffix[slen-s-1] != PATH_SEP
948 && strequ(suffix + slen - s, *suf)) {
951 } while (*++suf != NULL);
957 /* Open file NAME with the given flags and mode and store its status
958 into *ST. Return a file descriptor to the newly opened file, or -1
959 (setting errno) on failure. */
961 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
965 /* Refuse to follow symbolic links unless -c or -f. */
966 if (!to_stdout && !force)
970 else if (lstat (name, st) == 0 && S_ISLNK (st->st_mode))
977 fd = OPEN (name, flags, mode);
978 if (0 <= fd && fstat (fd, st) != 0)
989 /* ========================================================================
990 * Set ifname to the input file name (with a suffix appended if necessary)
991 * and istat to its stats. For decompression, if no file exists with the
992 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
993 * For MSDOS, we try only z_suffix and z.
994 * Return an open file descriptor or -1.
997 open_input_file (iname, sbuf)
1001 int ilen; /* strlen(ifname) */
1002 int z_suffix_errno = 0;
1003 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1004 char **suf = suffixes;
1006 #ifdef NO_MULTIPLE_DOTS
1007 char *dot; /* pointer to ifname extension, or NULL */
1010 int open_flags = ascii && !decompress ? O_RDONLY : O_RDONLY | O_BINARY;
1014 if (sizeof ifname - 1 <= strlen (iname))
1017 strcpy(ifname, iname);
1019 /* If input file exists, return OK. */
1020 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1024 if (!decompress || errno != ENOENT) {
1028 /* file.ext doesn't exist, try adding a suffix (after removing any
1029 * version number for VMS).
1031 s = get_suffix(ifname);
1033 progerror(ifname); /* ifname already has z suffix and does not exist */
1036 #ifdef NO_MULTIPLE_DOTS
1037 dot = strrchr(ifname, '.');
1039 strcat(ifname, ".");
1040 dot = strrchr(ifname, '.');
1043 ilen = strlen(ifname);
1044 if (strequ(z_suffix, ".gz")) suf++;
1046 /* Search for all suffixes */
1048 char *s0 = s = *suf;
1049 strcpy (ifname, iname);
1050 #ifdef NO_MULTIPLE_DOTS
1052 if (*dot == '\0') strcpy (dot, ".");
1054 #ifdef MAX_EXT_CHARS
1055 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1056 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1058 if (sizeof ifname <= ilen + strlen (s))
1061 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1064 if (errno != ENOENT)
1069 if (strequ (s0, z_suffix))
1070 z_suffix_errno = errno;
1071 } while (*++suf != NULL);
1073 /* No suffix found, complain using z_suffix: */
1074 strcpy(ifname, iname);
1075 #ifdef NO_MULTIPLE_DOTS
1076 if (*dot == '\0') strcpy(dot, ".");
1078 #ifdef MAX_EXT_CHARS
1079 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1080 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1082 strcat(ifname, z_suffix);
1083 errno = z_suffix_errno;
1088 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1093 /* ========================================================================
1094 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1095 * Sets save_orig_name to true if the file name has been truncated.
1097 local int make_ofname()
1099 char *suff; /* ofname z suffix */
1101 strcpy(ofname, ifname);
1102 /* strip a version number if any and get the gzip suffix if present: */
1103 suff = get_suffix(ofname);
1107 /* With -t or -l, try all files (even without .gz suffix)
1108 * except with -r (behave as with just -dr).
1110 if (!recursive && (list || test)) return OK;
1112 /* Avoid annoying messages with -r */
1113 if (verbose || (!recursive && !quiet)) {
1114 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1115 program_name, ifname));
1119 /* Make a special case for .tgz and .taz: */
1121 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1122 strcpy(suff, ".tar");
1124 *suff = '\0'; /* strip the z suffix */
1126 /* ofname might be changed later if infile contains an original name */
1128 } else if (suff != NULL) {
1129 /* Avoid annoying messages with -r (see treat_dir()) */
1130 if (verbose || (!recursive && !quiet)) {
1131 /* Don't use WARN, as it affects exit status. */
1132 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1133 program_name, ifname, suff);
1139 #ifdef NO_MULTIPLE_DOTS
1140 suff = strrchr(ofname, '.');
1142 if (sizeof ofname <= strlen (ofname) + 1)
1144 strcat(ofname, ".");
1145 # ifdef MAX_EXT_CHARS
1146 if (strequ(z_suffix, "z")) {
1147 if (sizeof ofname <= strlen (ofname) + 2)
1149 strcat(ofname, "gz"); /* enough room */
1152 /* On the Atari and some versions of MSDOS, name_too_long()
1153 * does not work correctly because of a bug in stat(). So we
1154 * must truncate here.
1156 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1157 suff[MAX_SUFFIX+1-z_len] = '\0';
1161 #endif /* NO_MULTIPLE_DOTS */
1162 if (sizeof ofname <= strlen (ofname) + z_len)
1164 strcat(ofname, z_suffix);
1166 } /* decompress ? */
1170 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1175 /* ========================================================================
1176 * Check the magic number of the input file and update ofname if an
1177 * original name was given and to_stdout is not set.
1178 * Return the compression method, -1 for error, -2 for warning.
1179 * Set inptr to the offset of the next byte to be processed.
1180 * Updates time_stamp if there is one and --no-time is not used.
1181 * This function may be called repeatedly for an input file consisting
1182 * of several contiguous gzip'ed members.
1183 * IN assertions: there is at least one remaining compressed member.
1184 * If the member is a zip file, it must be the only one.
1186 local int get_method(in)
1187 int in; /* input file descriptor */
1189 uch flags; /* compression flags */
1190 char magic[2]; /* magic header */
1191 int imagic1; /* like magic[1], but can represent EOF */
1192 ulg stamp; /* time stamp */
1194 /* If --force and --stdout, zcat == cat, so do not complain about
1195 * premature end of file: use try_byte instead of get_byte.
1197 if (force && to_stdout) {
1198 magic[0] = (char)try_byte();
1199 imagic1 = try_byte ();
1200 magic[1] = (char) imagic1;
1201 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1203 magic[0] = (char)get_byte();
1204 magic[1] = (char)get_byte();
1205 imagic1 = 0; /* avoid lint warning */
1207 method = -1; /* unknown yet */
1208 part_nb++; /* number of parts in gzip file */
1210 last_member = RECORD_IO;
1211 /* assume multiple members in gzip file except for record oriented I/O */
1213 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1214 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1216 method = (int)get_byte();
1217 if (method != DEFLATED) {
1219 "%s: %s: unknown method %d -- not supported\n",
1220 program_name, ifname, method);
1225 flags = (uch)get_byte();
1227 if ((flags & ENCRYPTED) != 0) {
1229 "%s: %s is encrypted -- not supported\n",
1230 program_name, ifname);
1234 if ((flags & CONTINUATION) != 0) {
1236 "%s: %s is a a multi-part gzip file -- not supported\n",
1237 program_name, ifname);
1239 if (force <= 1) return -1;
1241 if ((flags & RESERVED) != 0) {
1243 "%s: %s has flags 0x%x -- not supported\n",
1244 program_name, ifname, flags);
1246 if (force <= 1) return -1;
1248 stamp = (ulg)get_byte();
1249 stamp |= ((ulg)get_byte()) << 8;
1250 stamp |= ((ulg)get_byte()) << 16;
1251 stamp |= ((ulg)get_byte()) << 24;
1252 if (stamp != 0 && !no_time)
1254 time_stamp.tv_sec = stamp;
1255 time_stamp.tv_nsec = 0;
1258 (void)get_byte(); /* Ignore extra flags for the moment */
1259 (void)get_byte(); /* Ignore OS type for the moment */
1261 if ((flags & CONTINUATION) != 0) {
1262 unsigned part = (unsigned)get_byte();
1263 part |= ((unsigned)get_byte())<<8;
1265 fprintf(stderr,"%s: %s: part number %u\n",
1266 program_name, ifname, part);
1269 if ((flags & EXTRA_FIELD) != 0) {
1270 unsigned len = (unsigned)get_byte();
1271 len |= ((unsigned)get_byte())<<8;
1273 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1274 program_name, ifname, len);
1276 while (len--) (void)get_byte();
1279 /* Get original file name if it was truncated */
1280 if ((flags & ORIG_NAME) != 0) {
1281 if (no_name || (to_stdout && !list) || part_nb > 1) {
1282 /* Discard the old name */
1283 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1284 do {c=get_byte();} while (c != 0);
1286 /* Copy the base name. Keep a directory prefix intact. */
1287 char *p = gzip_base_name (ofname);
1290 *p = (char)get_char();
1291 if (*p++ == '\0') break;
1292 if (p >= ofname+sizeof(ofname)) {
1293 gzip_error ("corrupted input -- file name too large");
1296 p = gzip_base_name (base);
1297 memmove (base, p, strlen (p) + 1);
1298 /* If necessary, adapt the name to local OS conventions: */
1300 MAKE_LEGAL_NAME(base);
1301 if (base) list=0; /* avoid warning about unused variable */
1303 } /* no_name || to_stdout */
1306 /* Discard file comment if any */
1307 if ((flags & COMMENT) != 0) {
1308 while (get_char() != 0) /* null */ ;
1311 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1314 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1315 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1316 /* To simplify the code, we support a zip file when alone only.
1317 * We are thus guaranteed that the entire local header fits in inbuf.
1321 if (check_zipfile(in) != OK) return -1;
1322 /* check_zipfile may get ofname from the local header */
1325 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1329 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1331 method = COMPRESSED;
1334 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1339 } else if (force && to_stdout && !list) { /* pass input unchanged */
1345 if (method >= 0) return method;
1348 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1349 program_name, ifname);
1356 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1361 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1362 program_name, ifname));
1367 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1368 program_name, ifname));
1373 /* ========================================================================
1374 * Display the characteristics of the compressed file.
1375 * If the given method is < 0, display the accumulated totals.
1376 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1378 local void do_list(ifd, method)
1379 int ifd; /* input file descriptor */
1380 int method; /* compression method */
1382 ulg crc; /* original crc */
1383 static int first_time = 1;
1384 static char* methods[MAX_METHODS] = {
1389 "", "", "", "", /* 4 to 7 reserved */
1391 int positive_off_t_width = 1;
1394 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1395 positive_off_t_width++;
1398 if (first_time && method >= 0) {
1401 printf("method crc date time ");
1404 printf("%*.*s %*.*s ratio uncompressed_name\n",
1405 positive_off_t_width, positive_off_t_width, "compressed",
1406 positive_off_t_width, positive_off_t_width, "uncompressed");
1408 } else if (method < 0) {
1409 if (total_in <= 0 || total_out <= 0) return;
1413 if (verbose || !quiet) {
1414 fprint_off(stdout, total_in, positive_off_t_width);
1416 fprint_off(stdout, total_out, positive_off_t_width);
1419 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1420 /* header_bytes is not meaningful but used to ensure the same
1421 * ratio if there is a single file.
1423 printf(" (totals)\n");
1426 crc = (ulg)~0; /* unknown */
1428 bytes_in = ifile_size;
1431 if (method == DEFLATED && !last_member) {
1432 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1433 * If the lseek fails, we could use read() to get to the end, but
1434 * --list is used to get quick results.
1435 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1436 * you are not concerned about speed.
1438 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1439 if (bytes_in != -1L) {
1442 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1446 bytes_out = LG(buf+4);
1449 #endif /* RECORD_IO */
1452 struct tm *tm = localtime (&time_stamp.tv_sec);
1453 printf ("%5s %08lx ", methods[method], crc);
1455 printf ("%s%3d %02d:%02d ",
1456 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1458 tm->tm_mday, tm->tm_hour, tm->tm_min);
1460 printf ("??? ?? ??:?? ");
1462 fprint_off(stdout, bytes_in, positive_off_t_width);
1464 fprint_off(stdout, bytes_out, positive_off_t_width);
1466 if (bytes_in == -1L) {
1468 bytes_in = bytes_out = header_bytes = 0;
1469 } else if (total_in >= 0) {
1470 total_in += bytes_in;
1472 if (bytes_out == -1L) {
1474 bytes_in = bytes_out = header_bytes = 0;
1475 } else if (total_out >= 0) {
1476 total_out += bytes_out;
1478 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1479 printf(" %s\n", ofname);
1482 /* ========================================================================
1483 * Return true if the two stat structures correspond to the same file.
1485 local int same_file(stat1, stat2)
1489 return stat1->st_ino == stat2->st_ino
1490 && stat1->st_dev == stat2->st_dev
1492 /* Can't rely on st_ino and st_dev, use other fields: */
1493 && stat1->st_mode == stat2->st_mode
1494 && stat1->st_uid == stat2->st_uid
1495 && stat1->st_gid == stat2->st_gid
1496 && stat1->st_size == stat2->st_size
1497 && stat1->st_atime == stat2->st_atime
1498 && stat1->st_mtime == stat2->st_mtime
1499 && stat1->st_ctime == stat2->st_ctime
1504 /* ========================================================================
1505 * Return true if a file name is ambiguous because the operating system
1506 * truncates file names.
1508 local int name_too_long(name, statb)
1509 char *name; /* file name to check */
1510 struct stat *statb; /* stat buf for this file name */
1512 int s = strlen(name);
1514 struct stat tstat; /* stat for truncated name */
1517 tstat = *statb; /* Just in case OS does not fill all fields */
1519 res = lstat(name, &tstat) == 0 && same_file(statb, &tstat);
1521 Trace((stderr, " too_long(%s) => %d\n", name, res));
1525 /* ========================================================================
1526 * Shorten the given name by one character, or replace a .tar extension
1527 * with .tgz. Truncate the last part of the name which is longer than
1528 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1529 * has only parts shorter than MIN_PART truncate the longest part.
1530 * For decompression, just remove the last character of the name.
1532 * IN assertion: for compression, the suffix of the given name is z_suffix.
1534 local void shorten_name(name)
1537 int len; /* length of name without z_suffix */
1538 char *trunc = NULL; /* character to be truncated */
1539 int plen; /* current part length */
1540 int min_part = MIN_PART; /* current minimum part length */
1546 gzip_error ("name too short");
1550 p = get_suffix(name);
1552 gzip_error ("can't recover suffix\n");
1556 /* compress 1234567890.tar to 1234567890.tgz */
1557 if (len > 4 && strequ(p-4, ".tar")) {
1558 strcpy(p-4, ".tgz");
1561 /* Try keeping short extensions intact:
1562 * 1234.678.012.gz -> 123.678.012.gz
1565 p = strrchr(name, PATH_SEP);
1568 plen = strcspn(p, PART_SEP);
1570 if (plen > min_part) trunc = p-1;
1573 } while (trunc == NULL && --min_part != 0);
1575 if (trunc != NULL) {
1577 trunc[0] = trunc[1];
1581 trunc = strrchr(name, PART_SEP[0]);
1583 gzip_error ("internal error in shorten_name");
1584 if (trunc[1] == '\0') trunc--; /* force truncation */
1586 strcpy(trunc, z_suffix);
1589 /* ========================================================================
1590 * If compressing to a file, check if ofname is not ambiguous
1591 * because the operating system truncates names. Otherwise, generate
1592 * a new ofname and save the original name in the compressed file.
1593 * If the compressed file already exists, ask for confirmation.
1594 * The check for name truncation is made dynamically, because different
1595 * file systems on the same OS might use different truncation rules (on SVR4
1596 * s5 truncates to 14 chars and ufs does not truncate).
1597 * This function returns -1 if the file must be skipped, and
1598 * updates save_orig_name if necessary.
1599 * IN assertions: save_orig_name is already set if ofname has been
1600 * already truncated because of NO_MULTIPLE_DOTS. The input file has
1601 * already been open and istat is set.
1603 local int check_ofname()
1605 struct stat ostat; /* stat for ofname */
1608 /* Check for strictly conforming Posix systems (which return ENAMETOOLONG
1609 * instead of silently truncating filenames).
1612 while (lstat(ofname, &ostat) != 0) {
1613 if (errno != ENAMETOOLONG) return 0; /* ofname does not exist */
1614 shorten_name(ofname);
1617 if (lstat(ofname, &ostat) != 0) return 0;
1619 /* Check for name truncation on existing file. Do this even on systems
1620 * defining ENAMETOOLONG, because on most systems the strict Posix
1621 * behavior is disabled by default (silent name truncation allowed).
1623 if (!decompress && name_too_long(ofname, &ostat)) {
1624 shorten_name(ofname);
1625 if (lstat(ofname, &ostat) != 0) return 0;
1628 /* Check that the input and output files are different (could be
1629 * the same by name truncation or links).
1631 if (same_file(&istat, &ostat)) {
1632 if (strequ(ifname, ofname)) {
1633 fprintf(stderr, "%s: %s: cannot %scompress onto itself\n",
1634 program_name, ifname, decompress ? "de" : "");
1636 fprintf(stderr, "%s: %s and %s are the same file\n",
1637 program_name, ifname, ofname);
1642 /* Ask permission to overwrite the existing file */
1645 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1646 if (foreground && isatty(fileno(stdin))) {
1647 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1652 fprintf(stderr, "\tnot overwritten\n");
1653 if (exit_code == OK) exit_code = WARNING;
1657 if (xunlink (ofname)) {
1665 /* ========================================================================
1666 * Copy modes, times, ownership from input file to output file.
1667 * IN assertion: to_stdout is false.
1669 local void copy_stat(ifstat)
1670 struct stat *ifstat;
1672 mode_t mode = ifstat->st_mode & CHMOD_MODE_BITS;
1676 struct timespec timespec[2];
1677 timespec[0] = get_stat_atime (ifstat);
1678 timespec[1] = get_stat_mtime (ifstat);
1680 if (decompress && 0 <= time_stamp.tv_nsec
1681 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1682 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1684 timespec[1] = time_stamp;
1686 fprintf(stderr, "%s: time stamp restored\n", ofname);
1690 if (futimens (ofd, ofname, timespec) != 0)
1693 WARN ((stderr, "%s: ", program_name));
1701 /* Copy the protection modes */
1703 r = fchmod (ofd, mode);
1705 r = chmod (ofname, mode);
1709 WARN ((stderr, "%s: ", program_name));
1717 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1719 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1726 /* ========================================================================
1727 * Recurse through the given directory. This code is taken from ncompress.
1729 local void treat_dir (fd, dir)
1735 char nbuf[MAX_PATH_LEN];
1738 dirp = fdopendir (fd);
1746 ** WARNING: the following algorithm could occasionally cause
1747 ** compress to produce error warnings of the form "<filename>.gz
1748 ** already has .gz suffix - ignored". This occurs when the
1749 ** .gz output file is inserted into the directory below
1750 ** readdir's current pointer.
1751 ** These warnings are harmless but annoying, so they are suppressed
1752 ** with option -r (except when -v is on). An alternative
1753 ** to allowing this would be to store the entire directory
1754 ** list in memory, then compress the entries in the stored
1755 ** list. Given the depth-first recursive algorithm used here,
1756 ** this could use up a tremendous amount of memory. I don't
1757 ** think it's worth it. -- Dave Mack
1758 ** (An other alternative might be two passes to avoid depth-first.)
1761 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1763 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1767 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1769 if (len != 0 /* dir = "" means current dir on Amiga */
1771 && dir[len-1] != PATH_SEP2
1774 && dir[len-1] != PATH_SEP3
1777 nbuf[len++] = PATH_SEP;
1779 strcpy(nbuf+len, dp->d_name);
1782 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1783 program_name, dir, dp->d_name);
1789 if (CLOSEDIR(dirp) != 0)
1792 #endif /* ! NO_DIR */
1794 /* ========================================================================
1795 * Free all dynamically allocated variables and exit with the given code.
1797 local void do_exit(exitcode)
1800 static int in_exit = 0;
1802 if (in_exit) exit(exitcode);
1804 if (env != NULL) free(env), env = NULL;
1805 if (args != NULL) free((char*)args), args = NULL;
1819 /* ========================================================================
1820 * Close and unlink the output file.
1823 remove_output_file ()
1825 if (remove_ofname) {
1831 /* ========================================================================
1837 remove_output_file ();
1841 /* ========================================================================
1845 abort_gzip_signal (sig)
1848 remove_output_file ();
1849 signal (sig, SIG_DFL);