1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001-2002, 2006-2007, 2009-2012 Free Software
5 Copyright (C) 1992-1993 Jean-loup Gailly
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 3, or (at your option)
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
17 You should have received a copy of the GNU General Public License
18 along with this program; if not, write to the Free Software Foundation,
19 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
22 * The unzip code was written and put in the public domain by Mark Adler.
23 * Portions of the lzw code are derived from the public domain 'compress'
24 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
25 * Ken Turkowski, Dave Mack and Peter Jannesen.
27 * See the license_msg below and the file COPYING for the software license.
28 * See the file algorithm.doc for the compression algorithms and file formats.
31 static char const *const license_msg[] = {
32 "Copyright (C) 2007, 2010, 2011 Free Software Foundation, Inc.",
33 "Copyright (C) 1993 Jean-loup Gailly.",
34 "This is free software. You may redistribute copies of it under the terms of",
35 "the GNU General Public License <http://www.gnu.org/licenses/gpl.html>.",
36 "There is NO WARRANTY, to the extent permitted by law.",
39 /* Compress files with zip algorithm and 'compress' interface.
40 * See help() function below for all options.
42 * file.gz: compressed file with same mode, owner, and utimes
43 * or stdout with -c option or if stdin used as input.
44 * If the output file name had to be truncated, the original name is kept
45 * in the compressed file.
46 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
48 * Using gz on MSDOS would create too many file name conflicts. For
49 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
50 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
51 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
52 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
54 * For the meaning of all compilation flags, see comments in Makefile.in.
59 #include <sys/types.h>
72 #include "fcntl-safer.h"
74 #include "ignore-value.h"
75 #include "stat-time.h"
98 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
101 # define MAX_PATH_LEN 1024 /* max pathname length */
113 off_t lseek (int fd, off_t offset, int whence);
117 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
121 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
124 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
127 # define SA_NOCLDSTOP 0
128 # define sigprocmask(how, set, oset) /* empty */
129 # define sigset_t int
130 # if ! HAVE_SIGINTERRUPT
131 # define siginterrupt(sig, flag) /* empty */
135 #ifndef HAVE_WORKING_O_NOFOLLOW
136 # define HAVE_WORKING_O_NOFOLLOW 0
139 /* Separator for file name parts (see shorten_name()) */
140 #ifdef NO_MULTIPLE_DOTS
141 # define PART_SEP "-"
143 # define PART_SEP "."
148 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
149 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
150 DECLARE(ush, d_buf, DIST_BUFSIZE);
151 DECLARE(uch, window, 2L*WSIZE);
153 DECLARE(ush, tab_prefix, 1L<<BITS);
155 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
156 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
159 /* local variables */
161 /* If true, pretend that standard input is a tty. This option
162 is deliberately not documented, and only for testing. */
163 static bool presume_input_tty;
165 static int ascii = 0; /* convert end-of-lines to local OS conventions */
166 int to_stdout = 0; /* output to stdout (-c) */
167 static int decompress = 0; /* decompress (-d) */
168 static int force = 0; /* don't ask questions, compress links (-f) */
169 static int no_name = -1; /* don't save or restore the original file name */
170 static int no_time = -1; /* don't save or restore the original file time */
171 static int recursive = 0; /* recurse through directories (-r) */
172 static int list = 0; /* list the file contents (-l) */
173 int verbose = 0; /* be verbose (-v) */
174 int quiet = 0; /* be very quiet (-q) */
175 static int do_lzw = 0; /* generate output compatible with old compress (-Z) */
176 int test = 0; /* test .gz file integrity */
177 static int foreground = 0; /* set if program run in foreground */
178 char *program_name; /* program name */
179 int maxbits = BITS; /* max bits per code for LZW */
180 int method = DEFLATED;/* compression method */
181 int level = 6; /* compression level */
182 int exit_code = OK; /* program exit code */
183 int save_orig_name; /* set if original name must be saved */
184 static int last_member; /* set for .zip and .Z files */
185 static int part_nb; /* number of parts in .gz file */
186 struct timespec time_stamp; /* original time stamp (modification time) */
187 off_t ifile_size; /* input file size, -1 for devices (debug only) */
188 static char *env; /* contents of GZIP env variable */
189 static char **args = NULL; /* argv pointer if GZIP env variable defined */
190 static char const *z_suffix; /* default suffix (can be set with --suffix) */
191 static size_t z_len; /* strlen(z_suffix) */
193 /* The set of signals that are caught. */
194 static sigset_t caught_signals;
196 /* If nonzero then exit with status WARNING, rather than with the usual
197 signal status, on receipt of a signal with this value. This
198 suppresses a "Broken Pipe" message with some shells. */
199 static int volatile exiting_signal;
201 /* If nonnegative, close this file descriptor and unlink ofname on error. */
202 static int volatile remove_ofname_fd = -1;
204 off_t bytes_in; /* number of input bytes */
205 off_t bytes_out; /* number of output bytes */
206 static off_t total_in; /* input bytes for all files */
207 static off_t total_out; /* output bytes for all files */
208 char ifname[MAX_PATH_LEN]; /* input file name */
209 char ofname[MAX_PATH_LEN]; /* output file name */
210 static struct stat istat; /* status for input file */
211 int ifd; /* input file descriptor */
212 int ofd; /* output file descriptor */
213 unsigned insize; /* valid bytes in inbuf */
214 unsigned inptr; /* index of next byte to be processed in inbuf */
215 unsigned outcnt; /* bytes in output buffer */
217 static int handled_sig[] =
219 /* SIGINT must be first, as 'foreground' depends on it. */
239 /* For long options that have no equivalent short option, use a
240 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
243 PRESUME_INPUT_TTY_OPTION = CHAR_MAX + 1
246 static const struct option longopts[] =
248 /* { name has_arg *flag val } */
249 {"ascii", 0, 0, 'a'}, /* ascii text mode */
250 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
251 {"stdout", 0, 0, 'c'}, /* write output on standard output */
252 {"decompress", 0, 0, 'd'}, /* decompress */
253 {"uncompress", 0, 0, 'd'}, /* decompress */
254 /* {"encrypt", 0, 0, 'e'}, encrypt */
255 {"force", 0, 0, 'f'}, /* force overwrite of output file */
256 {"help", 0, 0, 'h'}, /* give help */
257 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
258 {"list", 0, 0, 'l'}, /* list .gz file contents */
259 {"license", 0, 0, 'L'}, /* display software license */
260 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
261 {"name", 0, 0, 'N'}, /* save or restore original name & time */
262 {"-presume-input-tty", no_argument, NULL, PRESUME_INPUT_TTY_OPTION},
263 {"quiet", 0, 0, 'q'}, /* quiet mode */
264 {"silent", 0, 0, 'q'}, /* quiet mode */
265 {"recursive", 0, 0, 'r'}, /* recurse through directories */
266 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
267 {"test", 0, 0, 't'}, /* test compressed file integrity */
268 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
269 {"verbose", 0, 0, 'v'}, /* verbose mode */
270 {"version", 0, 0, 'V'}, /* display version number */
271 {"fast", 0, 0, '1'}, /* compress faster */
272 {"best", 0, 0, '9'}, /* compress better */
273 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
274 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
279 /* local functions */
281 local void try_help (void) ATTRIBUTE_NORETURN;
282 local void help (void);
283 local void license (void);
284 local void version (void);
285 local int input_eof (void);
286 local void treat_stdin (void);
287 local void treat_file (char *iname);
288 local int create_outfile (void);
289 local char *get_suffix (char *name);
290 local int open_input_file (char *iname, struct stat *sbuf);
291 local void discard_input_bytes (size_t nbytes, unsigned int flags);
292 local int make_ofname (void);
293 local void shorten_name (char *name);
294 local int get_method (int in);
295 local void do_list (int ifd, int method);
296 local int check_ofname (void);
297 local void copy_stat (struct stat *ifstat);
298 local void install_signal_handlers (void);
299 local void remove_output_file (void);
300 local RETSIGTYPE abort_gzip_signal (int);
301 local void do_exit (int exitcode) ATTRIBUTE_NORETURN;
302 int main (int argc, char **argv);
303 static int (*work) (int infile, int outfile) = zip; /* function to call */
306 local void treat_dir (int fd, char *dir);
309 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
314 fprintf (stderr, "Try `%s --help' for more information.\n",
319 /* ======================================================================== */
322 static char const* const help_msg[] = {
323 "Compress or uncompress FILEs (by default, compress FILES in-place).",
325 "Mandatory arguments to long options are mandatory for short options too.",
328 " -a, --ascii ascii text; convert end-of-line using local conventions",
330 " -c, --stdout write on standard output, keep original files unchanged",
331 " -d, --decompress decompress",
332 /* -e, --encrypt encrypt */
333 " -f, --force force overwrite of output file and compress links",
334 " -h, --help give this help",
335 /* -k, --pkzip force output in pkzip format */
336 " -l, --list list compressed file contents",
337 " -L, --license display software license",
339 " -m, --no-time do not save or restore the original modification time",
340 " -M, --time save or restore the original modification time",
342 " -n, --no-name do not save or restore the original name and time stamp",
343 " -N, --name save or restore the original name and time stamp",
344 " -q, --quiet suppress all warnings",
346 " -r, --recursive operate recursively on directories",
348 " -S, --suffix=SUF use suffix SUF on compressed files",
349 " -t, --test test compressed file integrity",
350 " -v, --verbose verbose mode",
351 " -V, --version display version number",
352 " -1, --fast compress faster",
353 " -9, --best compress better",
355 " -Z, --lzw produce output compatible with old compress",
356 " -b, --bits=BITS max number of bits per code (implies -Z)",
359 "With no FILE, or when FILE is -, read standard input.",
361 "Report bugs to <bug-gzip@gnu.org>.",
363 char const *const *p = help_msg;
365 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
366 while (*p) printf ("%s\n", *p++);
369 /* ======================================================================== */
372 char const *const *p = license_msg;
374 printf ("%s %s\n", program_name, Version);
375 while (*p) printf ("%s\n", *p++);
378 /* ======================================================================== */
383 printf ("Written by Jean-loup Gailly.\n");
386 local void progerror (char const *string)
389 fprintf (stderr, "%s: ", program_name);
395 /* ======================================================================== */
396 int main (int argc, char **argv)
398 int file_count; /* number of files to process */
399 size_t proglen; /* length of program_name */
400 int optc; /* current option */
402 EXPAND(argc, argv); /* wild card expansion if necessary */
404 program_name = gzip_base_name (argv[0]);
405 proglen = strlen (program_name);
407 atexit (close_stdin);
409 /* Suppress .exe for MSDOS, OS/2 and VMS: */
410 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
411 program_name[proglen - 4] = '\0';
413 /* Add options in GZIP environment variable if there is one */
414 env = add_envopt(&argc, &argv, OPTIONS_VAR);
415 if (env != NULL) args = argv;
418 # define GNU_STANDARD 1
421 /* For compatibility with old compress, use program name as an option.
422 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
423 * gzip even if it is invoked under the name gunzip or zcat.
425 * Systems which do not support links can still use -d or -dc.
426 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
428 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
429 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
431 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
432 || strequ (program_name, "gzcat")) /* gzcat */
433 decompress = to_stdout = 1;
437 z_len = strlen(z_suffix);
439 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
440 longopts, (int *)0)) != -1) {
445 maxbits = atoi(optarg);
446 for (; *optarg; optarg++)
447 if (! ('0' <= *optarg && *optarg <= '9'))
449 fprintf (stderr, "%s: -b operand is not an integer\n",
455 to_stdout = 1; break;
457 decompress = 1; break;
461 help(); do_exit(OK); break;
463 list = decompress = to_stdout = 1; break;
465 license(); do_exit(OK); break;
466 case 'm': /* undocumented, may change later */
468 case 'M': /* undocumented, may change later */
471 no_name = no_time = 1; break;
473 no_name = no_time = 0; break;
474 case PRESUME_INPUT_TTY_OPTION:
475 presume_input_tty = true; break;
477 quiet = 1; verbose = 0; break;
480 fprintf (stderr, "%s: -r not supported on this system\n",
488 #ifdef NO_MULTIPLE_DOTS
489 if (*optarg == '.') optarg++;
491 z_len = strlen(optarg);
495 test = decompress = to_stdout = 1;
498 verbose++; quiet = 0; break;
500 version(); do_exit(OK); break;
505 fprintf(stderr, "%s: -Z not supported in this version\n",
510 case '1': case '2': case '3': case '4':
511 case '5': case '6': case '7': case '8': case '9':
515 /* Error message already emitted by getopt_long. */
518 } /* loop on all arguments */
520 /* By default, save name and timestamp on compression but do not
521 * restore them on decompression.
523 if (no_time < 0) no_time = decompress;
524 if (no_name < 0) no_name = decompress;
526 file_count = argc - optind;
530 if (ascii && !quiet) {
531 fprintf(stderr, "%s: option --ascii ignored on this system\n",
535 if (z_len == 0 || z_len > MAX_SUFFIX) {
536 fprintf(stderr, "%s: invalid suffix '%s'\n", program_name, z_suffix);
540 if (do_lzw && !decompress) work = lzw;
542 /* Allocate all global buffers (for DYN_ALLOC option) */
543 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
544 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
545 ALLOC(ush, d_buf, DIST_BUFSIZE);
546 ALLOC(uch, window, 2L*WSIZE);
548 ALLOC(ush, tab_prefix, 1L<<BITS);
550 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
551 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
554 exiting_signal = quiet ? SIGPIPE : 0;
555 install_signal_handlers ();
557 /* And get to work */
558 if (file_count != 0) {
559 if (to_stdout && !test && !list && (!decompress || !ascii)) {
560 SET_BINARY_MODE(fileno(stdout));
562 while (optind < argc) {
563 treat_file(argv[optind++]);
565 } else { /* Standard input */
568 if (list && !quiet && file_count > 1) {
569 do_list(-1, -1); /* print totals */
572 return exit_code; /* just to avoid lint warning */
575 /* Return nonzero when at end of file on input. */
579 if (!decompress || last_member)
584 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
587 /* Unget the char that fill_inbuf got. */
594 /* ========================================================================
595 * Compress or decompress stdin
597 local void treat_stdin()
600 && (presume_input_tty
601 || isatty(fileno((FILE *)(decompress ? stdin : stdout))))) {
602 /* Do not send compressed data to the terminal or read it from
603 * the terminal. We get here when user invoked the program
604 * without parameters, so be helpful. According to the GNU standards:
606 * If there is one behavior you think is most useful when the output
607 * is to a terminal, and another that you think is most useful when
608 * the output is a file or a pipe, then it is usually best to make
609 * the default behavior the one that is useful with output to a
610 * terminal, and have an option for the other behavior.
612 * Here we use the --force option to get the other behavior.
616 ("%s: compressed data not %s a terminal."
617 " Use -f to force %scompression.\n"
618 "For help, type: %s -h\n"),
620 decompress ? "read from" : "written to",
621 decompress ? "de" : "",
626 if (decompress || !ascii) {
627 SET_BINARY_MODE(fileno(stdin));
629 if (!test && !list && (!decompress || !ascii)) {
630 SET_BINARY_MODE(fileno(stdout));
632 strcpy(ifname, "stdin");
633 strcpy(ofname, "stdout");
635 /* Get the file's time stamp and size. */
636 if (fstat (fileno (stdin), &istat) != 0)
638 progerror ("standard input");
641 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
642 time_stamp.tv_nsec = -1;
643 if (!no_time || list)
645 if (S_ISREG (istat.st_mode))
646 time_stamp = get_stat_mtime (&istat);
648 gettime (&time_stamp);
651 clear_bufs(); /* clear input and output buffers */
657 method = get_method(ifd);
659 do_exit(exit_code); /* error message already emitted */
663 do_list(ifd, method);
667 /* Actually do the compression/decompression. Loop over zipped members.
670 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
675 method = get_method(ifd);
676 if (method < 0) return; /* error message already emitted */
677 bytes_out = 0; /* required for length check */
682 fprintf(stderr, " OK\n");
684 } else if (!decompress) {
685 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
686 fprintf(stderr, "\n");
687 #ifdef DISPLAY_STDIN_RATIO
689 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
690 fprintf(stderr, "\n");
696 /* ========================================================================
697 * Compress or decompress the given file
699 local void treat_file(iname)
702 /* Accept "-" as synonym for stdin */
703 if (strequ(iname, "-")) {
704 int cflag = to_stdout;
710 /* Check if the input file is present, set ifname and istat: */
711 ifd = open_input_file (iname, &istat);
715 /* If the input name is that of a directory, recurse or ignore: */
716 if (S_ISDIR(istat.st_mode)) {
719 treat_dir (ifd, iname);
720 /* Warning: ifname is now garbage */
725 WARN ((stderr, "%s: %s is a directory -- ignored\n",
726 program_name, ifname));
732 if (! S_ISREG (istat.st_mode))
735 "%s: %s is not a directory or a regular file - ignored\n",
736 program_name, ifname));
740 if (istat.st_mode & S_ISUID)
742 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
743 program_name, ifname));
747 if (istat.st_mode & S_ISGID)
749 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
750 program_name, ifname));
757 if (istat.st_mode & S_ISVTX)
760 "%s: %s has the sticky bit set - file ignored\n",
761 program_name, ifname));
765 if (2 <= istat.st_nlink)
767 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
768 program_name, ifname,
769 (unsigned long int) istat.st_nlink - 1,
770 istat.st_nlink == 2 ? ' ' : 's'));
777 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
778 time_stamp.tv_nsec = -1;
779 if (!no_time || list)
780 time_stamp = get_stat_mtime (&istat);
782 /* Generate output file name. For -r and (-t or -l), skip files
783 * without a valid gzip suffix (check done in make_ofname).
785 if (to_stdout && !list && !test) {
786 strcpy(ofname, "stdout");
788 } else if (make_ofname() != OK) {
793 clear_bufs(); /* clear input and output buffers */
797 method = get_method(ifd); /* updates ofname if original given */
800 return; /* error message already emitted */
804 do_list(ifd, method);
805 if (close (ifd) != 0)
810 /* If compressing to a file, check if ofname is not ambiguous
811 * because the operating system truncates names. Otherwise, generate
812 * a new ofname and save the original name in the compressed file.
815 ofd = fileno(stdout);
816 /* Keep remove_ofname_fd negative. */
818 if (create_outfile() != OK) return;
820 if (!decompress && save_orig_name && !verbose && !quiet) {
821 fprintf(stderr, "%s: %s compressed to %s\n",
822 program_name, ifname, ofname);
825 /* Keep the name even if not truncated except with --no-name: */
826 if (!save_orig_name) save_orig_name = !no_name;
829 fprintf(stderr, "%s:\t", ifname);
832 /* Actually do the compression/decompression. Loop over zipped members.
835 if ((*work)(ifd, ofd) != OK) {
836 method = -1; /* force cleanup */
843 method = get_method(ifd);
844 if (method < 0) break; /* error message already emitted */
845 bytes_out = 0; /* required for length check */
848 if (close (ifd) != 0)
857 if (close (ofd) != 0)
860 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
861 remove_ofname_fd = -1;
862 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
863 sigprocmask (SIG_SETMASK, &oldset, NULL);
867 WARN ((stderr, "%s: ", program_name));
870 errno = unlink_errno;
878 remove_output_file ();
882 /* Display statistics */
885 fprintf(stderr, " OK");
886 } else if (decompress) {
887 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
889 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
891 if (!test && !to_stdout) {
892 fprintf(stderr, " -- replaced with %s", ofname);
894 fprintf(stderr, "\n");
898 /* ========================================================================
899 * Create the output file. Return OK or ERROR.
900 * Try several times if necessary to avoid truncating the z_suffix. For
901 * example, do not create a compressed file of name "1234567890123."
902 * Sets save_orig_name to true if the file name has been truncated.
903 * IN assertions: the input file has already been open (ifd is set) and
904 * ofname has already been updated if there was an original name.
905 * OUT assertions: ifd and ofd are closed in case of error.
907 local int create_outfile()
909 int name_shortened = 0;
910 int flags = (O_WRONLY | O_CREAT | O_EXCL
911 | (ascii && decompress ? 0 : O_BINARY));
918 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
919 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
921 sigprocmask (SIG_SETMASK, &oldset, NULL);
930 shorten_name (ofname);
936 if (check_ofname () != OK)
950 if (name_shortened && decompress)
952 /* name might be too long if an original name was saved */
953 WARN ((stderr, "%s: %s: warning, name truncated\n",
954 program_name, ofname));
960 /* ========================================================================
961 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
962 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
963 * accepted suffixes, in addition to the value of the --suffix option.
964 * ".tgz" is a useful convention for tar.z files on systems limited
965 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
966 * also accepted suffixes. For Unix, we do not want to accept any
967 * .??z suffix as indicating a compressed file; some people use .xyz
968 * to denote volume data.
969 * On systems allowing multiple versions of the same file (such as VMS),
970 * this function removes any version suffix in the given name.
972 local char *get_suffix(name)
976 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
977 static char const *known_suffixes[] =
978 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
983 char const **suf = known_suffixes;
986 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
989 /* strip a version number from the file name */
991 char *v = strrchr(name, SUFFIX_SEP);
992 if (v != NULL) *v = '\0';
996 if (nlen <= MAX_SUFFIX+2) {
997 strcpy(suffix, name);
999 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1002 slen = strlen(suffix);
1004 int s = strlen(*suf);
1005 if (slen > s && suffix[slen-s-1] != PATH_SEP
1006 && strequ(suffix + slen - s, *suf)) {
1009 } while (*++suf != NULL);
1015 /* Open file NAME with the given flags and mode and store its status
1016 into *ST. Return a file descriptor to the newly opened file, or -1
1017 (setting errno) on failure. */
1019 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1023 /* Refuse to follow symbolic links unless -c or -f. */
1024 if (!to_stdout && !force)
1026 if (HAVE_WORKING_O_NOFOLLOW)
1027 flags |= O_NOFOLLOW;
1030 #if HAVE_LSTAT || defined lstat
1031 if (lstat (name, st) != 0)
1033 else if (S_ISLNK (st->st_mode))
1042 fd = OPEN (name, flags, mode);
1043 if (0 <= fd && fstat (fd, st) != 0)
1054 /* ========================================================================
1055 * Set ifname to the input file name (with a suffix appended if necessary)
1056 * and istat to its stats. For decompression, if no file exists with the
1057 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1058 * For MSDOS, we try only z_suffix and z.
1059 * Return an open file descriptor or -1.
1062 open_input_file (iname, sbuf)
1066 int ilen; /* strlen(ifname) */
1067 int z_suffix_errno = 0;
1068 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1069 char const **suf = suffixes;
1071 #ifdef NO_MULTIPLE_DOTS
1072 char *dot; /* pointer to ifname extension, or NULL */
1075 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1076 | (ascii && !decompress ? 0 : O_BINARY));
1080 if (sizeof ifname - 1 <= strlen (iname))
1083 strcpy(ifname, iname);
1085 /* If input file exists, return OK. */
1086 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1090 if (!decompress || errno != ENOENT) {
1094 /* file.ext doesn't exist, try adding a suffix (after removing any
1095 * version number for VMS).
1097 s = get_suffix(ifname);
1099 progerror(ifname); /* ifname already has z suffix and does not exist */
1102 #ifdef NO_MULTIPLE_DOTS
1103 dot = strrchr(ifname, '.');
1105 strcat(ifname, ".");
1106 dot = strrchr(ifname, '.');
1109 ilen = strlen(ifname);
1110 if (strequ(z_suffix, ".gz")) suf++;
1112 /* Search for all suffixes */
1114 char const *s0 = s = *suf;
1115 strcpy (ifname, iname);
1116 #ifdef NO_MULTIPLE_DOTS
1118 if (*dot == '\0') strcpy (dot, ".");
1120 #ifdef MAX_EXT_CHARS
1121 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1122 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1124 if (sizeof ifname <= ilen + strlen (s))
1127 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1130 if (errno != ENOENT)
1135 if (strequ (s0, z_suffix))
1136 z_suffix_errno = errno;
1137 } while (*++suf != NULL);
1139 /* No suffix found, complain using z_suffix: */
1140 strcpy(ifname, iname);
1141 #ifdef NO_MULTIPLE_DOTS
1142 if (*dot == '\0') strcpy(dot, ".");
1144 #ifdef MAX_EXT_CHARS
1145 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1146 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1148 strcat(ifname, z_suffix);
1149 errno = z_suffix_errno;
1154 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1159 /* ========================================================================
1160 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1161 * Sets save_orig_name to true if the file name has been truncated.
1163 local int make_ofname()
1165 char *suff; /* ofname z suffix */
1167 strcpy(ofname, ifname);
1168 /* strip a version number if any and get the gzip suffix if present: */
1169 suff = get_suffix(ofname);
1173 /* With -t or -l, try all files (even without .gz suffix)
1174 * except with -r (behave as with just -dr).
1176 if (!recursive && (list || test)) return OK;
1178 /* Avoid annoying messages with -r */
1179 if (verbose || (!recursive && !quiet)) {
1180 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1181 program_name, ifname));
1185 /* Make a special case for .tgz and .taz: */
1187 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1188 strcpy(suff, ".tar");
1190 *suff = '\0'; /* strip the z suffix */
1192 /* ofname might be changed later if infile contains an original name */
1194 } else if (suff && ! force) {
1195 /* Avoid annoying messages with -r (see treat_dir()) */
1196 if (verbose || (!recursive && !quiet)) {
1197 /* Don't use WARN, as it affects exit status. */
1198 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1199 program_name, ifname, suff);
1205 #ifdef NO_MULTIPLE_DOTS
1206 suff = strrchr(ofname, '.');
1208 if (sizeof ofname <= strlen (ofname) + 1)
1210 strcat(ofname, ".");
1211 # ifdef MAX_EXT_CHARS
1212 if (strequ(z_suffix, "z")) {
1213 if (sizeof ofname <= strlen (ofname) + 2)
1215 strcat(ofname, "gz"); /* enough room */
1218 /* On the Atari and some versions of MSDOS,
1219 * ENAMETOOLONG does not work correctly. So we
1220 * must truncate here.
1222 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1223 suff[MAX_SUFFIX+1-z_len] = '\0';
1227 #endif /* NO_MULTIPLE_DOTS */
1228 if (sizeof ofname <= strlen (ofname) + z_len)
1230 strcat(ofname, z_suffix);
1232 } /* decompress ? */
1236 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1240 /* Discard NBYTES input bytes from the input, or up through the next
1241 zero byte if NBYTES == (size_t) -1. If FLAGS say that the header
1242 CRC should be computed, update the CRC accordingly. */
1244 discard_input_bytes (nbytes, flags)
1250 uch c = get_byte ();
1251 if (flags & HEADER_CRC)
1253 if (nbytes != (size_t) -1)
1260 /* ========================================================================
1261 * Check the magic number of the input file and update ofname if an
1262 * original name was given and to_stdout is not set.
1263 * Return the compression method, -1 for error, -2 for warning.
1264 * Set inptr to the offset of the next byte to be processed.
1265 * Updates time_stamp if there is one and --no-time is not used.
1266 * This function may be called repeatedly for an input file consisting
1267 * of several contiguous gzip'ed members.
1268 * IN assertions: there is at least one remaining compressed member.
1269 * If the member is a zip file, it must be the only one.
1271 local int get_method(in)
1272 int in; /* input file descriptor */
1274 uch flags; /* compression flags */
1275 uch magic[10]; /* magic header */
1276 int imagic0; /* first magic byte or EOF */
1277 int imagic1; /* like magic[1], but can represent EOF */
1278 ulg stamp; /* time stamp */
1280 /* If --force and --stdout, zcat == cat, so do not complain about
1281 * premature end of file: use try_byte instead of get_byte.
1283 if (force && to_stdout) {
1284 imagic0 = try_byte();
1286 imagic1 = try_byte ();
1288 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1290 magic[0] = get_byte ();
1293 magic[1] = get_byte ();
1294 imagic1 = 0; /* avoid lint warning */
1296 imagic1 = try_byte ();
1300 method = -1; /* unknown yet */
1301 part_nb++; /* number of parts in gzip file */
1303 last_member = RECORD_IO;
1304 /* assume multiple members in gzip file except for record oriented I/O */
1306 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1307 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1309 method = (int)get_byte();
1310 if (method != DEFLATED) {
1312 "%s: %s: unknown method %d -- not supported\n",
1313 program_name, ifname, method);
1318 flags = (uch)get_byte();
1320 if ((flags & ENCRYPTED) != 0) {
1322 "%s: %s is encrypted -- not supported\n",
1323 program_name, ifname);
1327 if ((flags & RESERVED) != 0) {
1329 "%s: %s has flags 0x%x -- not supported\n",
1330 program_name, ifname, flags);
1332 if (force <= 1) return -1;
1334 stamp = (ulg)get_byte();
1335 stamp |= ((ulg)get_byte()) << 8;
1336 stamp |= ((ulg)get_byte()) << 16;
1337 stamp |= ((ulg)get_byte()) << 24;
1338 if (stamp != 0 && !no_time)
1340 time_stamp.tv_sec = stamp;
1341 time_stamp.tv_nsec = 0;
1344 magic[8] = get_byte (); /* Ignore extra flags. */
1345 magic[9] = get_byte (); /* Ignore OS type. */
1347 if (flags & HEADER_CRC)
1349 magic[2] = DEFLATED;
1351 magic[4] = stamp & 0xff;
1352 magic[5] = (stamp >> 8) & 0xff;
1353 magic[6] = (stamp >> 16) & 0xff;
1354 magic[7] = stamp >> 24;
1359 if ((flags & EXTRA_FIELD) != 0) {
1361 unsigned int len = lenbuf[0] = get_byte ();
1362 len |= (lenbuf[1] = get_byte ()) << 8;
1364 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1365 program_name, ifname, len);
1367 if (flags & HEADER_CRC)
1369 discard_input_bytes (len, flags);
1372 /* Get original file name if it was truncated */
1373 if ((flags & ORIG_NAME) != 0) {
1374 if (no_name || (to_stdout && !list) || part_nb > 1) {
1375 /* Discard the old name */
1376 discard_input_bytes (-1, flags);
1378 /* Copy the base name. Keep a directory prefix intact. */
1379 char *p = gzip_base_name (ofname);
1382 *p = (char) get_byte ();
1383 if (*p++ == '\0') break;
1384 if (p >= ofname+sizeof(ofname)) {
1385 gzip_error ("corrupted input -- file name too large");
1388 if (flags & HEADER_CRC)
1389 updcrc ((uch *) base, p - base);
1390 p = gzip_base_name (base);
1391 memmove (base, p, strlen (p) + 1);
1392 /* If necessary, adapt the name to local OS conventions: */
1394 MAKE_LEGAL_NAME(base);
1395 if (base) list=0; /* avoid warning about unused variable */
1397 } /* no_name || to_stdout */
1400 /* Discard file comment if any */
1401 if ((flags & COMMENT) != 0) {
1402 discard_input_bytes (-1, flags);
1405 if (flags & HEADER_CRC)
1407 unsigned int crc16 = updcrc (magic, 0) & 0xffff;
1408 unsigned int header16 = get_byte ();
1409 header16 |= ((unsigned int) get_byte ()) << 8;
1410 if (header16 != crc16)
1413 "%s: %s: header checksum 0x%04x != computed checksum 0x%04x\n",
1414 program_name, ifname, header16, crc16);
1422 header_bytes = inptr + 2*4; /* include crc and size */
1425 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1426 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1427 /* To simplify the code, we support a zip file when alone only.
1428 * We are thus guaranteed that the entire local header fits in inbuf.
1432 if (check_zipfile(in) != OK) return -1;
1433 /* check_zipfile may get ofname from the local header */
1436 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1440 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1442 method = COMPRESSED;
1445 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1450 } else if (force && to_stdout && !list) { /* pass input unchanged */
1456 if (imagic0 != EOF) {
1457 write_buf(fileno(stdout), magic, 1);
1461 if (method >= 0) return method;
1464 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1465 program_name, ifname);
1472 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1477 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1478 program_name, ifname));
1483 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1484 program_name, ifname));
1489 /* ========================================================================
1490 * Display the characteristics of the compressed file.
1491 * If the given method is < 0, display the accumulated totals.
1492 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1494 local void do_list(ifd, method)
1495 int ifd; /* input file descriptor */
1496 int method; /* compression method */
1498 ulg crc; /* original crc */
1499 static int first_time = 1;
1500 static char const *const methods[MAX_METHODS] = {
1505 "", "", "", "", /* 4 to 7 reserved */
1507 int positive_off_t_width = 1;
1510 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1511 positive_off_t_width++;
1514 if (first_time && method >= 0) {
1517 printf("method crc date time ");
1520 printf("%*.*s %*.*s ratio uncompressed_name\n",
1521 positive_off_t_width, positive_off_t_width, "compressed",
1522 positive_off_t_width, positive_off_t_width, "uncompressed");
1524 } else if (method < 0) {
1525 if (total_in <= 0 || total_out <= 0) return;
1529 if (verbose || !quiet) {
1530 fprint_off(stdout, total_in, positive_off_t_width);
1532 fprint_off(stdout, total_out, positive_off_t_width);
1535 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1536 /* header_bytes is not meaningful but used to ensure the same
1537 * ratio if there is a single file.
1539 printf(" (totals)\n");
1542 crc = (ulg)~0; /* unknown */
1544 bytes_in = ifile_size;
1546 if (!RECORD_IO && method == DEFLATED && !last_member) {
1547 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1548 * If the lseek fails, we could use read() to get to the end, but
1549 * --list is used to get quick results.
1550 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1551 * you are not concerned about speed.
1553 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1554 if (bytes_in != -1L) {
1557 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1561 bytes_out = LG(buf+4);
1567 struct tm *tm = localtime (&time_stamp.tv_sec);
1568 printf ("%5s %08lx ", methods[method], crc);
1570 printf ("%s%3d %02d:%02d ",
1571 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1573 tm->tm_mday, tm->tm_hour, tm->tm_min);
1575 printf ("??? ?? ??:?? ");
1577 fprint_off(stdout, bytes_in, positive_off_t_width);
1579 fprint_off(stdout, bytes_out, positive_off_t_width);
1581 if (bytes_in == -1L) {
1583 bytes_in = bytes_out = header_bytes = 0;
1584 } else if (total_in >= 0) {
1585 total_in += bytes_in;
1587 if (bytes_out == -1L) {
1589 bytes_in = bytes_out = header_bytes = 0;
1590 } else if (total_out >= 0) {
1591 total_out += bytes_out;
1593 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1594 printf(" %s\n", ofname);
1597 /* ========================================================================
1598 * Shorten the given name by one character, or replace a .tar extension
1599 * with .tgz. Truncate the last part of the name which is longer than
1600 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1601 * has only parts shorter than MIN_PART truncate the longest part.
1602 * For decompression, just remove the last character of the name.
1604 * IN assertion: for compression, the suffix of the given name is z_suffix.
1606 local void shorten_name(name)
1609 int len; /* length of name without z_suffix */
1610 char *trunc = NULL; /* character to be truncated */
1611 int plen; /* current part length */
1612 int min_part = MIN_PART; /* current minimum part length */
1618 gzip_error ("name too short");
1622 p = get_suffix(name);
1624 gzip_error ("can't recover suffix\n");
1628 /* compress 1234567890.tar to 1234567890.tgz */
1629 if (len > 4 && strequ(p-4, ".tar")) {
1630 strcpy(p-4, ".tgz");
1633 /* Try keeping short extensions intact:
1634 * 1234.678.012.gz -> 123.678.012.gz
1637 p = strrchr(name, PATH_SEP);
1640 plen = strcspn(p, PART_SEP);
1642 if (plen > min_part) trunc = p-1;
1645 } while (trunc == NULL && --min_part != 0);
1647 if (trunc != NULL) {
1649 trunc[0] = trunc[1];
1653 trunc = strrchr(name, PART_SEP[0]);
1655 gzip_error ("internal error in shorten_name");
1656 if (trunc[1] == '\0') trunc--; /* force truncation */
1658 strcpy(trunc, z_suffix);
1661 /* ========================================================================
1662 * The compressed file already exists, so ask for confirmation.
1663 * Return ERROR if the file must be skipped.
1665 local int check_ofname()
1667 /* Ask permission to overwrite the existing file */
1670 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1671 if (foreground && (presume_input_tty || isatty(fileno(stdin)))) {
1672 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1677 fprintf(stderr, "\tnot overwritten\n");
1678 if (exit_code == OK) exit_code = WARNING;
1682 if (xunlink (ofname)) {
1690 /* ========================================================================
1691 * Copy modes, times, ownership from input file to output file.
1692 * IN assertion: to_stdout is false.
1694 local void copy_stat(ifstat)
1695 struct stat *ifstat;
1697 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1701 struct timespec timespec[2];
1702 timespec[0] = get_stat_atime (ifstat);
1703 timespec[1] = get_stat_mtime (ifstat);
1705 if (decompress && 0 <= time_stamp.tv_nsec
1706 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1707 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1709 timespec[1] = time_stamp;
1711 fprintf(stderr, "%s: time stamp restored\n", ofname);
1715 if (fdutimens (ofd, ofname, timespec) != 0)
1718 WARN ((stderr, "%s: ", program_name));
1728 /* Copy ownership */
1730 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1732 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1736 /* Copy the protection modes */
1738 r = fchmod (ofd, mode);
1740 r = chmod (ofname, mode);
1744 WARN ((stderr, "%s: ", program_name));
1754 /* ========================================================================
1755 * Recurse through the given directory.
1757 local void treat_dir (fd, dir)
1762 char nbuf[MAX_PATH_LEN];
1767 dirp = fdopendir (fd);
1775 entries = streamsavedir (dirp);
1778 if (closedir (dirp) != 0)
1783 for (entry = entries; *entry; entry += entrylen + 1) {
1784 size_t len = strlen (dir);
1785 entrylen = strlen (entry);
1786 if (strequ (entry, ".") || strequ (entry, ".."))
1788 if (len + entrylen < MAX_PATH_LEN - 2) {
1790 if (len != 0 /* dir = "" means current dir on Amiga */
1792 && dir[len-1] != PATH_SEP2
1795 && dir[len-1] != PATH_SEP3
1798 nbuf[len++] = PATH_SEP;
1800 strcpy (nbuf + len, entry);
1803 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1804 program_name, dir, entry);
1810 #endif /* ! NO_DIR */
1812 /* Make sure signals get handled properly. */
1815 install_signal_handlers ()
1817 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1821 struct sigaction act;
1823 sigemptyset (&caught_signals);
1824 for (i = 0; i < nsigs; i++)
1826 sigaction (handled_sig[i], NULL, &act);
1827 if (act.sa_handler != SIG_IGN)
1828 sigaddset (&caught_signals, handled_sig[i]);
1831 act.sa_handler = abort_gzip_signal;
1832 act.sa_mask = caught_signals;
1835 for (i = 0; i < nsigs; i++)
1836 if (sigismember (&caught_signals, handled_sig[i]))
1840 sigaction (handled_sig[i], &act, NULL);
1843 for (i = 0; i < nsigs; i++)
1844 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1848 signal (handled_sig[i], abort_gzip_signal);
1849 siginterrupt (handled_sig[i], 1);
1854 /* ========================================================================
1855 * Free all dynamically allocated variables and exit with the given code.
1857 local void do_exit(exitcode)
1860 static int in_exit = 0;
1862 if (in_exit) exit(exitcode);
1881 /* ========================================================================
1882 * Close and unlink the output file.
1885 remove_output_file ()
1890 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1891 fd = remove_ofname_fd;
1894 remove_ofname_fd = -1;
1898 sigprocmask (SIG_SETMASK, &oldset, NULL);
1901 /* ========================================================================
1907 remove_output_file ();
1911 /* ========================================================================
1915 abort_gzip_signal (sig)
1919 signal (sig, SIG_IGN);
1920 remove_output_file ();
1921 if (sig == exiting_signal)
1923 signal (sig, SIG_DFL);