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"
95 # define CLOSEDIR(d) (closedir(d), 0)
97 # define CLOSEDIR(d) closedir(d)
101 # include <utimens.h>
104 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
107 # define MAX_PATH_LEN 1024 /* max pathname length */
119 off_t lseek (int fd, off_t offset, int whence);
123 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
127 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
130 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
133 # define SA_NOCLDSTOP 0
134 # define sigprocmask(how, set, oset) /* empty */
135 # define sigset_t int
136 # if ! HAVE_SIGINTERRUPT
137 # define siginterrupt(sig, flag) /* empty */
141 #ifndef HAVE_WORKING_O_NOFOLLOW
142 # define HAVE_WORKING_O_NOFOLLOW 0
145 /* Separator for file name parts (see shorten_name()) */
146 #ifdef NO_MULTIPLE_DOTS
147 # define PART_SEP "-"
149 # define PART_SEP "."
154 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
155 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
156 DECLARE(ush, d_buf, DIST_BUFSIZE);
157 DECLARE(uch, window, 2L*WSIZE);
159 DECLARE(ush, tab_prefix, 1L<<BITS);
161 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
162 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
165 /* local variables */
167 /* If true, pretend that standard input is a tty. This option
168 is deliberately not documented, and only for testing. */
169 static bool presume_input_tty;
171 static int ascii = 0; /* convert end-of-lines to local OS conventions */
172 int to_stdout = 0; /* output to stdout (-c) */
173 static int decompress = 0; /* decompress (-d) */
174 static int force = 0; /* don't ask questions, compress links (-f) */
175 static int no_name = -1; /* don't save or restore the original file name */
176 static int no_time = -1; /* don't save or restore the original file time */
177 static int recursive = 0; /* recurse through directories (-r) */
178 static int list = 0; /* list the file contents (-l) */
179 int verbose = 0; /* be verbose (-v) */
180 int quiet = 0; /* be very quiet (-q) */
181 static int do_lzw = 0; /* generate output compatible with old compress (-Z) */
182 int test = 0; /* test .gz file integrity */
183 static int foreground = 0; /* set if program run in foreground */
184 char *program_name; /* program name */
185 int maxbits = BITS; /* max bits per code for LZW */
186 int method = DEFLATED;/* compression method */
187 int level = 6; /* compression level */
188 int exit_code = OK; /* program exit code */
189 int save_orig_name; /* set if original name must be saved */
190 static int last_member; /* set for .zip and .Z files */
191 static int part_nb; /* number of parts in .gz file */
192 struct timespec time_stamp; /* original time stamp (modification time) */
193 off_t ifile_size; /* input file size, -1 for devices (debug only) */
194 static char *env; /* contents of GZIP env variable */
195 static char **args = NULL; /* argv pointer if GZIP env variable defined */
196 static char const *z_suffix; /* default suffix (can be set with --suffix) */
197 static size_t z_len; /* strlen(z_suffix) */
199 /* The set of signals that are caught. */
200 static sigset_t caught_signals;
202 /* If nonzero then exit with status WARNING, rather than with the usual
203 signal status, on receipt of a signal with this value. This
204 suppresses a "Broken Pipe" message with some shells. */
205 static int volatile exiting_signal;
207 /* If nonnegative, close this file descriptor and unlink ofname on error. */
208 static int volatile remove_ofname_fd = -1;
210 off_t bytes_in; /* number of input bytes */
211 off_t bytes_out; /* number of output bytes */
212 static off_t total_in; /* input bytes for all files */
213 static off_t total_out; /* output bytes for all files */
214 char ifname[MAX_PATH_LEN]; /* input file name */
215 char ofname[MAX_PATH_LEN]; /* output file name */
216 static struct stat istat; /* status for input file */
217 int ifd; /* input file descriptor */
218 int ofd; /* output file descriptor */
219 unsigned insize; /* valid bytes in inbuf */
220 unsigned inptr; /* index of next byte to be processed in inbuf */
221 unsigned outcnt; /* bytes in output buffer */
223 static int handled_sig[] =
225 /* SIGINT must be first, as 'foreground' depends on it. */
243 /* For long options that have no equivalent short option, use a
244 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
247 PRESUME_INPUT_TTY_OPTION = CHAR_MAX + 1
250 static const struct option longopts[] =
252 /* { name has_arg *flag val } */
253 {"ascii", 0, 0, 'a'}, /* ascii text mode */
254 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
255 {"stdout", 0, 0, 'c'}, /* write output on standard output */
256 {"decompress", 0, 0, 'd'}, /* decompress */
257 {"uncompress", 0, 0, 'd'}, /* decompress */
258 /* {"encrypt", 0, 0, 'e'}, encrypt */
259 {"force", 0, 0, 'f'}, /* force overwrite of output file */
260 {"help", 0, 0, 'h'}, /* give help */
261 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
262 {"list", 0, 0, 'l'}, /* list .gz file contents */
263 {"license", 0, 0, 'L'}, /* display software license */
264 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
265 {"name", 0, 0, 'N'}, /* save or restore original name & time */
266 {"-presume-input-tty", no_argument, NULL, PRESUME_INPUT_TTY_OPTION},
267 {"quiet", 0, 0, 'q'}, /* quiet mode */
268 {"silent", 0, 0, 'q'}, /* quiet mode */
269 {"recursive", 0, 0, 'r'}, /* recurse through directories */
270 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
271 {"test", 0, 0, 't'}, /* test compressed file integrity */
272 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
273 {"verbose", 0, 0, 'v'}, /* verbose mode */
274 {"version", 0, 0, 'V'}, /* display version number */
275 {"fast", 0, 0, '1'}, /* compress faster */
276 {"best", 0, 0, '9'}, /* compress better */
277 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
278 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
283 /* local functions */
285 local void try_help (void) ATTRIBUTE_NORETURN;
286 local void help (void);
287 local void license (void);
288 local void version (void);
289 local int input_eof (void);
290 local void treat_stdin (void);
291 local void treat_file (char *iname);
292 local int create_outfile (void);
293 local char *get_suffix (char *name);
294 local int open_input_file (char *iname, struct stat *sbuf);
295 local void discard_input_bytes (size_t nbytes, unsigned int flags);
296 local int make_ofname (void);
297 local void shorten_name (char *name);
298 local int get_method (int in);
299 local void do_list (int ifd, int method);
300 local int check_ofname (void);
301 local void copy_stat (struct stat *ifstat);
302 local void install_signal_handlers (void);
303 local void remove_output_file (void);
304 local RETSIGTYPE abort_gzip_signal (int);
305 local void do_exit (int exitcode) ATTRIBUTE_NORETURN;
306 int main (int argc, char **argv);
307 static int (*work) (int infile, int outfile) = zip; /* function to call */
310 local void treat_dir (int fd, char *dir);
313 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
318 fprintf (stderr, "Try `%s --help' for more information.\n",
323 /* ======================================================================== */
326 static char const* const help_msg[] = {
327 "Compress or uncompress FILEs (by default, compress FILES in-place).",
329 "Mandatory arguments to long options are mandatory for short options too.",
332 " -a, --ascii ascii text; convert end-of-line using local conventions",
334 " -c, --stdout write on standard output, keep original files unchanged",
335 " -d, --decompress decompress",
336 /* -e, --encrypt encrypt */
337 " -f, --force force overwrite of output file and compress links",
338 " -h, --help give this help",
339 /* -k, --pkzip force output in pkzip format */
340 " -l, --list list compressed file contents",
341 " -L, --license display software license",
343 " -m, --no-time do not save or restore the original modification time",
344 " -M, --time save or restore the original modification time",
346 " -n, --no-name do not save or restore the original name and time stamp",
347 " -N, --name save or restore the original name and time stamp",
348 " -q, --quiet suppress all warnings",
350 " -r, --recursive operate recursively on directories",
352 " -S, --suffix=SUF use suffix SUF on compressed files",
353 " -t, --test test compressed file integrity",
354 " -v, --verbose verbose mode",
355 " -V, --version display version number",
356 " -1, --fast compress faster",
357 " -9, --best compress better",
359 " -Z, --lzw produce output compatible with old compress",
360 " -b, --bits=BITS max number of bits per code (implies -Z)",
363 "With no FILE, or when FILE is -, read standard input.",
365 "Report bugs to <bug-gzip@gnu.org>.",
367 char const *const *p = help_msg;
369 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
370 while (*p) printf ("%s\n", *p++);
373 /* ======================================================================== */
376 char const *const *p = license_msg;
378 printf ("%s %s\n", program_name, Version);
379 while (*p) printf ("%s\n", *p++);
382 /* ======================================================================== */
387 printf ("Written by Jean-loup Gailly.\n");
390 local void progerror (char const *string)
393 fprintf (stderr, "%s: ", program_name);
399 /* ======================================================================== */
400 int main (int argc, char **argv)
402 int file_count; /* number of files to process */
403 size_t proglen; /* length of program_name */
404 int optc; /* current option */
406 EXPAND(argc, argv); /* wild card expansion if necessary */
408 program_name = gzip_base_name (argv[0]);
409 proglen = strlen (program_name);
411 atexit (close_stdin);
413 /* Suppress .exe for MSDOS, OS/2 and VMS: */
414 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
415 program_name[proglen - 4] = '\0';
417 /* Add options in GZIP environment variable if there is one */
418 env = add_envopt(&argc, &argv, OPTIONS_VAR);
419 if (env != NULL) args = argv;
422 # define GNU_STANDARD 1
425 /* For compatibility with old compress, use program name as an option.
426 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
427 * gzip even if it is invoked under the name gunzip or zcat.
429 * Systems which do not support links can still use -d or -dc.
430 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
432 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
433 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
435 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
436 || strequ (program_name, "gzcat")) /* gzcat */
437 decompress = to_stdout = 1;
441 z_len = strlen(z_suffix);
443 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
444 longopts, (int *)0)) != -1) {
449 maxbits = atoi(optarg);
450 for (; *optarg; optarg++)
451 if (! ('0' <= *optarg && *optarg <= '9'))
453 fprintf (stderr, "%s: -b operand is not an integer\n",
459 to_stdout = 1; break;
461 decompress = 1; break;
465 help(); do_exit(OK); break;
467 list = decompress = to_stdout = 1; break;
469 license(); do_exit(OK); break;
470 case 'm': /* undocumented, may change later */
472 case 'M': /* undocumented, may change later */
475 no_name = no_time = 1; break;
477 no_name = no_time = 0; break;
478 case PRESUME_INPUT_TTY_OPTION:
479 presume_input_tty = true; break;
481 quiet = 1; verbose = 0; break;
484 fprintf (stderr, "%s: -r not supported on this system\n",
492 #ifdef NO_MULTIPLE_DOTS
493 if (*optarg == '.') optarg++;
495 z_len = strlen(optarg);
499 test = decompress = to_stdout = 1;
502 verbose++; quiet = 0; break;
504 version(); do_exit(OK); break;
509 fprintf(stderr, "%s: -Z not supported in this version\n",
514 case '1': case '2': case '3': case '4':
515 case '5': case '6': case '7': case '8': case '9':
519 /* Error message already emitted by getopt_long. */
522 } /* loop on all arguments */
524 /* By default, save name and timestamp on compression but do not
525 * restore them on decompression.
527 if (no_time < 0) no_time = decompress;
528 if (no_name < 0) no_name = decompress;
530 file_count = argc - optind;
534 if (ascii && !quiet) {
535 fprintf(stderr, "%s: option --ascii ignored on this system\n",
539 if (z_len == 0 || z_len > MAX_SUFFIX) {
540 fprintf(stderr, "%s: invalid suffix '%s'\n", program_name, z_suffix);
544 if (do_lzw && !decompress) work = lzw;
546 /* Allocate all global buffers (for DYN_ALLOC option) */
547 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
548 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
549 ALLOC(ush, d_buf, DIST_BUFSIZE);
550 ALLOC(uch, window, 2L*WSIZE);
552 ALLOC(ush, tab_prefix, 1L<<BITS);
554 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
555 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
558 exiting_signal = quiet ? SIGPIPE : 0;
559 install_signal_handlers ();
561 /* And get to work */
562 if (file_count != 0) {
563 if (to_stdout && !test && !list && (!decompress || !ascii)) {
564 SET_BINARY_MODE(fileno(stdout));
566 while (optind < argc) {
567 treat_file(argv[optind++]);
569 } else { /* Standard input */
572 if (list && !quiet && file_count > 1) {
573 do_list(-1, -1); /* print totals */
576 return exit_code; /* just to avoid lint warning */
579 /* Return nonzero when at end of file on input. */
583 if (!decompress || last_member)
588 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
591 /* Unget the char that fill_inbuf got. */
598 /* ========================================================================
599 * Compress or decompress stdin
601 local void treat_stdin()
604 && (presume_input_tty
605 || isatty(fileno((FILE *)(decompress ? stdin : stdout))))) {
606 /* Do not send compressed data to the terminal or read it from
607 * the terminal. We get here when user invoked the program
608 * without parameters, so be helpful. According to the GNU standards:
610 * If there is one behavior you think is most useful when the output
611 * is to a terminal, and another that you think is most useful when
612 * the output is a file or a pipe, then it is usually best to make
613 * the default behavior the one that is useful with output to a
614 * terminal, and have an option for the other behavior.
616 * Here we use the --force option to get the other behavior.
620 ("%s: compressed data not %s a terminal."
621 " Use -f to force %scompression.\n"
622 "For help, type: %s -h\n"),
624 decompress ? "read from" : "written to",
625 decompress ? "de" : "",
630 if (decompress || !ascii) {
631 SET_BINARY_MODE(fileno(stdin));
633 if (!test && !list && (!decompress || !ascii)) {
634 SET_BINARY_MODE(fileno(stdout));
636 strcpy(ifname, "stdin");
637 strcpy(ofname, "stdout");
639 /* Get the file's time stamp and size. */
640 if (fstat (fileno (stdin), &istat) != 0)
642 progerror ("standard input");
645 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
646 time_stamp.tv_nsec = -1;
647 if (!no_time || list)
649 if (S_ISREG (istat.st_mode))
650 time_stamp = get_stat_mtime (&istat);
652 gettime (&time_stamp);
655 clear_bufs(); /* clear input and output buffers */
661 method = get_method(ifd);
663 do_exit(exit_code); /* error message already emitted */
667 do_list(ifd, method);
671 /* Actually do the compression/decompression. Loop over zipped members.
674 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
679 method = get_method(ifd);
680 if (method < 0) return; /* error message already emitted */
681 bytes_out = 0; /* required for length check */
686 fprintf(stderr, " OK\n");
688 } else if (!decompress) {
689 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
690 fprintf(stderr, "\n");
691 #ifdef DISPLAY_STDIN_RATIO
693 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
694 fprintf(stderr, "\n");
700 /* ========================================================================
701 * Compress or decompress the given file
703 local void treat_file(iname)
706 /* Accept "-" as synonym for stdin */
707 if (strequ(iname, "-")) {
708 int cflag = to_stdout;
714 /* Check if the input file is present, set ifname and istat: */
715 ifd = open_input_file (iname, &istat);
719 /* If the input name is that of a directory, recurse or ignore: */
720 if (S_ISDIR(istat.st_mode)) {
723 treat_dir (ifd, iname);
724 /* Warning: ifname is now garbage */
729 WARN ((stderr, "%s: %s is a directory -- ignored\n",
730 program_name, ifname));
736 if (! S_ISREG (istat.st_mode))
739 "%s: %s is not a directory or a regular file - ignored\n",
740 program_name, ifname));
744 if (istat.st_mode & S_ISUID)
746 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
747 program_name, ifname));
751 if (istat.st_mode & S_ISGID)
753 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
754 program_name, ifname));
761 if (istat.st_mode & S_ISVTX)
764 "%s: %s has the sticky bit set - file ignored\n",
765 program_name, ifname));
769 if (2 <= istat.st_nlink)
771 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
772 program_name, ifname,
773 (unsigned long int) istat.st_nlink - 1,
774 istat.st_nlink == 2 ? ' ' : 's'));
781 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
782 time_stamp.tv_nsec = -1;
783 if (!no_time || list)
784 time_stamp = get_stat_mtime (&istat);
786 /* Generate output file name. For -r and (-t or -l), skip files
787 * without a valid gzip suffix (check done in make_ofname).
789 if (to_stdout && !list && !test) {
790 strcpy(ofname, "stdout");
792 } else if (make_ofname() != OK) {
797 clear_bufs(); /* clear input and output buffers */
801 method = get_method(ifd); /* updates ofname if original given */
804 return; /* error message already emitted */
808 do_list(ifd, method);
809 if (close (ifd) != 0)
814 /* If compressing to a file, check if ofname is not ambiguous
815 * because the operating system truncates names. Otherwise, generate
816 * a new ofname and save the original name in the compressed file.
819 ofd = fileno(stdout);
820 /* Keep remove_ofname_fd negative. */
822 if (create_outfile() != OK) return;
824 if (!decompress && save_orig_name && !verbose && !quiet) {
825 fprintf(stderr, "%s: %s compressed to %s\n",
826 program_name, ifname, ofname);
829 /* Keep the name even if not truncated except with --no-name: */
830 if (!save_orig_name) save_orig_name = !no_name;
833 fprintf(stderr, "%s:\t", ifname);
836 /* Actually do the compression/decompression. Loop over zipped members.
839 if ((*work)(ifd, ofd) != OK) {
840 method = -1; /* force cleanup */
847 method = get_method(ifd);
848 if (method < 0) break; /* error message already emitted */
849 bytes_out = 0; /* required for length check */
852 if (close (ifd) != 0)
861 if (close (ofd) != 0)
864 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
865 remove_ofname_fd = -1;
866 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
867 sigprocmask (SIG_SETMASK, &oldset, NULL);
871 WARN ((stderr, "%s: ", program_name));
874 errno = unlink_errno;
882 remove_output_file ();
886 /* Display statistics */
889 fprintf(stderr, " OK");
890 } else if (decompress) {
891 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
893 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
895 if (!test && !to_stdout) {
896 fprintf(stderr, " -- replaced with %s", ofname);
898 fprintf(stderr, "\n");
902 /* ========================================================================
903 * Create the output file. Return OK or ERROR.
904 * Try several times if necessary to avoid truncating the z_suffix. For
905 * example, do not create a compressed file of name "1234567890123."
906 * Sets save_orig_name to true if the file name has been truncated.
907 * IN assertions: the input file has already been open (ifd is set) and
908 * ofname has already been updated if there was an original name.
909 * OUT assertions: ifd and ofd are closed in case of error.
911 local int create_outfile()
913 int name_shortened = 0;
914 int flags = (O_WRONLY | O_CREAT | O_EXCL
915 | (ascii && decompress ? 0 : O_BINARY));
922 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
923 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
925 sigprocmask (SIG_SETMASK, &oldset, NULL);
934 shorten_name (ofname);
940 if (check_ofname () != OK)
954 if (name_shortened && decompress)
956 /* name might be too long if an original name was saved */
957 WARN ((stderr, "%s: %s: warning, name truncated\n",
958 program_name, ofname));
964 /* ========================================================================
965 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
966 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
967 * accepted suffixes, in addition to the value of the --suffix option.
968 * ".tgz" is a useful convention for tar.z files on systems limited
969 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
970 * also accepted suffixes. For Unix, we do not want to accept any
971 * .??z suffix as indicating a compressed file; some people use .xyz
972 * to denote volume data.
973 * On systems allowing multiple versions of the same file (such as VMS),
974 * this function removes any version suffix in the given name.
976 local char *get_suffix(name)
980 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
981 static char const *known_suffixes[] =
982 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
987 char const **suf = known_suffixes;
990 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
993 /* strip a version number from the file name */
995 char *v = strrchr(name, SUFFIX_SEP);
996 if (v != NULL) *v = '\0';
1000 if (nlen <= MAX_SUFFIX+2) {
1001 strcpy(suffix, name);
1003 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1006 slen = strlen(suffix);
1008 int s = strlen(*suf);
1009 if (slen > s && suffix[slen-s-1] != PATH_SEP
1010 && strequ(suffix + slen - s, *suf)) {
1013 } while (*++suf != NULL);
1019 /* Open file NAME with the given flags and mode and store its status
1020 into *ST. Return a file descriptor to the newly opened file, or -1
1021 (setting errno) on failure. */
1023 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1027 /* Refuse to follow symbolic links unless -c or -f. */
1028 if (!to_stdout && !force)
1030 if (HAVE_WORKING_O_NOFOLLOW)
1031 flags |= O_NOFOLLOW;
1034 #if HAVE_LSTAT || defined lstat
1035 if (lstat (name, st) != 0)
1037 else if (S_ISLNK (st->st_mode))
1046 fd = OPEN (name, flags, mode);
1047 if (0 <= fd && fstat (fd, st) != 0)
1058 /* ========================================================================
1059 * Set ifname to the input file name (with a suffix appended if necessary)
1060 * and istat to its stats. For decompression, if no file exists with the
1061 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1062 * For MSDOS, we try only z_suffix and z.
1063 * Return an open file descriptor or -1.
1066 open_input_file (iname, sbuf)
1070 int ilen; /* strlen(ifname) */
1071 int z_suffix_errno = 0;
1072 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1073 char const **suf = suffixes;
1075 #ifdef NO_MULTIPLE_DOTS
1076 char *dot; /* pointer to ifname extension, or NULL */
1079 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1080 | (ascii && !decompress ? 0 : O_BINARY));
1084 if (sizeof ifname - 1 <= strlen (iname))
1087 strcpy(ifname, iname);
1089 /* If input file exists, return OK. */
1090 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1094 if (!decompress || errno != ENOENT) {
1098 /* file.ext doesn't exist, try adding a suffix (after removing any
1099 * version number for VMS).
1101 s = get_suffix(ifname);
1103 progerror(ifname); /* ifname already has z suffix and does not exist */
1106 #ifdef NO_MULTIPLE_DOTS
1107 dot = strrchr(ifname, '.');
1109 strcat(ifname, ".");
1110 dot = strrchr(ifname, '.');
1113 ilen = strlen(ifname);
1114 if (strequ(z_suffix, ".gz")) suf++;
1116 /* Search for all suffixes */
1118 char const *s0 = s = *suf;
1119 strcpy (ifname, iname);
1120 #ifdef NO_MULTIPLE_DOTS
1122 if (*dot == '\0') strcpy (dot, ".");
1124 #ifdef MAX_EXT_CHARS
1125 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1126 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1128 if (sizeof ifname <= ilen + strlen (s))
1131 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1134 if (errno != ENOENT)
1139 if (strequ (s0, z_suffix))
1140 z_suffix_errno = errno;
1141 } while (*++suf != NULL);
1143 /* No suffix found, complain using z_suffix: */
1144 strcpy(ifname, iname);
1145 #ifdef NO_MULTIPLE_DOTS
1146 if (*dot == '\0') strcpy(dot, ".");
1148 #ifdef MAX_EXT_CHARS
1149 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1150 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1152 strcat(ifname, z_suffix);
1153 errno = z_suffix_errno;
1158 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1163 /* ========================================================================
1164 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1165 * Sets save_orig_name to true if the file name has been truncated.
1167 local int make_ofname()
1169 char *suff; /* ofname z suffix */
1171 strcpy(ofname, ifname);
1172 /* strip a version number if any and get the gzip suffix if present: */
1173 suff = get_suffix(ofname);
1177 /* With -t or -l, try all files (even without .gz suffix)
1178 * except with -r (behave as with just -dr).
1180 if (!recursive && (list || test)) return OK;
1182 /* Avoid annoying messages with -r */
1183 if (verbose || (!recursive && !quiet)) {
1184 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1185 program_name, ifname));
1189 /* Make a special case for .tgz and .taz: */
1191 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1192 strcpy(suff, ".tar");
1194 *suff = '\0'; /* strip the z suffix */
1196 /* ofname might be changed later if infile contains an original name */
1198 } else if (suff && ! force) {
1199 /* Avoid annoying messages with -r (see treat_dir()) */
1200 if (verbose || (!recursive && !quiet)) {
1201 /* Don't use WARN, as it affects exit status. */
1202 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1203 program_name, ifname, suff);
1209 #ifdef NO_MULTIPLE_DOTS
1210 suff = strrchr(ofname, '.');
1212 if (sizeof ofname <= strlen (ofname) + 1)
1214 strcat(ofname, ".");
1215 # ifdef MAX_EXT_CHARS
1216 if (strequ(z_suffix, "z")) {
1217 if (sizeof ofname <= strlen (ofname) + 2)
1219 strcat(ofname, "gz"); /* enough room */
1222 /* On the Atari and some versions of MSDOS,
1223 * ENAMETOOLONG does not work correctly. So we
1224 * must truncate here.
1226 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1227 suff[MAX_SUFFIX+1-z_len] = '\0';
1231 #endif /* NO_MULTIPLE_DOTS */
1232 if (sizeof ofname <= strlen (ofname) + z_len)
1234 strcat(ofname, z_suffix);
1236 } /* decompress ? */
1240 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1244 /* Discard NBYTES input bytes from the input, or up through the next
1245 zero byte if NBYTES == (size_t) -1. If FLAGS say that the header
1246 CRC should be computed, update the CRC accordingly. */
1248 discard_input_bytes (nbytes, flags)
1254 uch c = get_byte ();
1255 if (flags & HEADER_CRC)
1257 if (nbytes != (size_t) -1)
1264 /* ========================================================================
1265 * Check the magic number of the input file and update ofname if an
1266 * original name was given and to_stdout is not set.
1267 * Return the compression method, -1 for error, -2 for warning.
1268 * Set inptr to the offset of the next byte to be processed.
1269 * Updates time_stamp if there is one and --no-time is not used.
1270 * This function may be called repeatedly for an input file consisting
1271 * of several contiguous gzip'ed members.
1272 * IN assertions: there is at least one remaining compressed member.
1273 * If the member is a zip file, it must be the only one.
1275 local int get_method(in)
1276 int in; /* input file descriptor */
1278 uch flags; /* compression flags */
1279 uch magic[10]; /* magic header */
1280 int imagic0; /* first magic byte or EOF */
1281 int imagic1; /* like magic[1], but can represent EOF */
1282 ulg stamp; /* time stamp */
1284 /* If --force and --stdout, zcat == cat, so do not complain about
1285 * premature end of file: use try_byte instead of get_byte.
1287 if (force && to_stdout) {
1288 imagic0 = try_byte();
1290 imagic1 = try_byte ();
1292 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1294 magic[0] = get_byte ();
1297 magic[1] = get_byte ();
1298 imagic1 = 0; /* avoid lint warning */
1300 imagic1 = try_byte ();
1304 method = -1; /* unknown yet */
1305 part_nb++; /* number of parts in gzip file */
1307 last_member = RECORD_IO;
1308 /* assume multiple members in gzip file except for record oriented I/O */
1310 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1311 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1313 method = (int)get_byte();
1314 if (method != DEFLATED) {
1316 "%s: %s: unknown method %d -- not supported\n",
1317 program_name, ifname, method);
1322 flags = (uch)get_byte();
1324 if ((flags & ENCRYPTED) != 0) {
1326 "%s: %s is encrypted -- not supported\n",
1327 program_name, ifname);
1331 if ((flags & RESERVED) != 0) {
1333 "%s: %s has flags 0x%x -- not supported\n",
1334 program_name, ifname, flags);
1336 if (force <= 1) return -1;
1338 stamp = (ulg)get_byte();
1339 stamp |= ((ulg)get_byte()) << 8;
1340 stamp |= ((ulg)get_byte()) << 16;
1341 stamp |= ((ulg)get_byte()) << 24;
1342 if (stamp != 0 && !no_time)
1344 time_stamp.tv_sec = stamp;
1345 time_stamp.tv_nsec = 0;
1348 magic[8] = get_byte (); /* Ignore extra flags. */
1349 magic[9] = get_byte (); /* Ignore OS type. */
1351 if (flags & HEADER_CRC)
1353 magic[2] = DEFLATED;
1355 magic[4] = stamp & 0xff;
1356 magic[5] = (stamp >> 8) & 0xff;
1357 magic[6] = (stamp >> 16) & 0xff;
1358 magic[7] = stamp >> 24;
1363 if ((flags & EXTRA_FIELD) != 0) {
1365 unsigned int len = lenbuf[0] = get_byte ();
1366 len |= (lenbuf[1] = get_byte ()) << 8;
1368 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1369 program_name, ifname, len);
1371 if (flags & HEADER_CRC)
1373 discard_input_bytes (len, flags);
1376 /* Get original file name if it was truncated */
1377 if ((flags & ORIG_NAME) != 0) {
1378 if (no_name || (to_stdout && !list) || part_nb > 1) {
1379 /* Discard the old name */
1380 discard_input_bytes (-1, flags);
1382 /* Copy the base name. Keep a directory prefix intact. */
1383 char *p = gzip_base_name (ofname);
1386 *p = (char) get_byte ();
1387 if (*p++ == '\0') break;
1388 if (p >= ofname+sizeof(ofname)) {
1389 gzip_error ("corrupted input -- file name too large");
1392 if (flags & HEADER_CRC)
1393 updcrc ((uch *) base, p - base);
1394 p = gzip_base_name (base);
1395 memmove (base, p, strlen (p) + 1);
1396 /* If necessary, adapt the name to local OS conventions: */
1398 MAKE_LEGAL_NAME(base);
1399 if (base) list=0; /* avoid warning about unused variable */
1401 } /* no_name || to_stdout */
1404 /* Discard file comment if any */
1405 if ((flags & COMMENT) != 0) {
1406 discard_input_bytes (-1, flags);
1409 if (flags & HEADER_CRC)
1411 unsigned int crc16 = updcrc (magic, 0) & 0xffff;
1412 unsigned int header16 = get_byte ();
1413 header16 |= ((unsigned int) get_byte ()) << 8;
1414 if (header16 != crc16)
1417 "%s: %s: header checksum 0x%04x != computed checksum 0x%04x\n",
1418 program_name, ifname, header16, crc16);
1426 header_bytes = inptr + 2*4; /* include crc and size */
1429 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1430 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1431 /* To simplify the code, we support a zip file when alone only.
1432 * We are thus guaranteed that the entire local header fits in inbuf.
1436 if (check_zipfile(in) != OK) return -1;
1437 /* check_zipfile may get ofname from the local header */
1440 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1444 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1446 method = COMPRESSED;
1449 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1454 } else if (force && to_stdout && !list) { /* pass input unchanged */
1460 if (imagic0 != EOF) {
1461 write_buf(fileno(stdout), magic, 1);
1465 if (method >= 0) return method;
1468 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1469 program_name, ifname);
1476 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1481 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1482 program_name, ifname));
1487 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1488 program_name, ifname));
1493 /* ========================================================================
1494 * Display the characteristics of the compressed file.
1495 * If the given method is < 0, display the accumulated totals.
1496 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1498 local void do_list(ifd, method)
1499 int ifd; /* input file descriptor */
1500 int method; /* compression method */
1502 ulg crc; /* original crc */
1503 static int first_time = 1;
1504 static char const *const methods[MAX_METHODS] = {
1509 "", "", "", "", /* 4 to 7 reserved */
1511 int positive_off_t_width = 1;
1514 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1515 positive_off_t_width++;
1518 if (first_time && method >= 0) {
1521 printf("method crc date time ");
1524 printf("%*.*s %*.*s ratio uncompressed_name\n",
1525 positive_off_t_width, positive_off_t_width, "compressed",
1526 positive_off_t_width, positive_off_t_width, "uncompressed");
1528 } else if (method < 0) {
1529 if (total_in <= 0 || total_out <= 0) return;
1533 if (verbose || !quiet) {
1534 fprint_off(stdout, total_in, positive_off_t_width);
1536 fprint_off(stdout, total_out, positive_off_t_width);
1539 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1540 /* header_bytes is not meaningful but used to ensure the same
1541 * ratio if there is a single file.
1543 printf(" (totals)\n");
1546 crc = (ulg)~0; /* unknown */
1548 bytes_in = ifile_size;
1550 if (!RECORD_IO && method == DEFLATED && !last_member) {
1551 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1552 * If the lseek fails, we could use read() to get to the end, but
1553 * --list is used to get quick results.
1554 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1555 * you are not concerned about speed.
1557 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1558 if (bytes_in != -1L) {
1561 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1565 bytes_out = LG(buf+4);
1571 struct tm *tm = localtime (&time_stamp.tv_sec);
1572 printf ("%5s %08lx ", methods[method], crc);
1574 printf ("%s%3d %02d:%02d ",
1575 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1577 tm->tm_mday, tm->tm_hour, tm->tm_min);
1579 printf ("??? ?? ??:?? ");
1581 fprint_off(stdout, bytes_in, positive_off_t_width);
1583 fprint_off(stdout, bytes_out, positive_off_t_width);
1585 if (bytes_in == -1L) {
1587 bytes_in = bytes_out = header_bytes = 0;
1588 } else if (total_in >= 0) {
1589 total_in += bytes_in;
1591 if (bytes_out == -1L) {
1593 bytes_in = bytes_out = header_bytes = 0;
1594 } else if (total_out >= 0) {
1595 total_out += bytes_out;
1597 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1598 printf(" %s\n", ofname);
1601 /* ========================================================================
1602 * Shorten the given name by one character, or replace a .tar extension
1603 * with .tgz. Truncate the last part of the name which is longer than
1604 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1605 * has only parts shorter than MIN_PART truncate the longest part.
1606 * For decompression, just remove the last character of the name.
1608 * IN assertion: for compression, the suffix of the given name is z_suffix.
1610 local void shorten_name(name)
1613 int len; /* length of name without z_suffix */
1614 char *trunc = NULL; /* character to be truncated */
1615 int plen; /* current part length */
1616 int min_part = MIN_PART; /* current minimum part length */
1622 gzip_error ("name too short");
1626 p = get_suffix(name);
1628 gzip_error ("can't recover suffix\n");
1632 /* compress 1234567890.tar to 1234567890.tgz */
1633 if (len > 4 && strequ(p-4, ".tar")) {
1634 strcpy(p-4, ".tgz");
1637 /* Try keeping short extensions intact:
1638 * 1234.678.012.gz -> 123.678.012.gz
1641 p = strrchr(name, PATH_SEP);
1644 plen = strcspn(p, PART_SEP);
1646 if (plen > min_part) trunc = p-1;
1649 } while (trunc == NULL && --min_part != 0);
1651 if (trunc != NULL) {
1653 trunc[0] = trunc[1];
1657 trunc = strrchr(name, PART_SEP[0]);
1659 gzip_error ("internal error in shorten_name");
1660 if (trunc[1] == '\0') trunc--; /* force truncation */
1662 strcpy(trunc, z_suffix);
1665 /* ========================================================================
1666 * The compressed file already exists, so ask for confirmation.
1667 * Return ERROR if the file must be skipped.
1669 local int check_ofname()
1671 /* Ask permission to overwrite the existing file */
1674 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1675 if (foreground && (presume_input_tty || isatty(fileno(stdin)))) {
1676 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1681 fprintf(stderr, "\tnot overwritten\n");
1682 if (exit_code == OK) exit_code = WARNING;
1686 if (xunlink (ofname)) {
1694 /* ========================================================================
1695 * Copy modes, times, ownership from input file to output file.
1696 * IN assertion: to_stdout is false.
1698 local void copy_stat(ifstat)
1699 struct stat *ifstat;
1701 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1705 struct timespec timespec[2];
1706 timespec[0] = get_stat_atime (ifstat);
1707 timespec[1] = get_stat_mtime (ifstat);
1709 if (decompress && 0 <= time_stamp.tv_nsec
1710 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1711 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1713 timespec[1] = time_stamp;
1715 fprintf(stderr, "%s: time stamp restored\n", ofname);
1719 if (fdutimens (ofd, ofname, timespec) != 0)
1722 WARN ((stderr, "%s: ", program_name));
1732 /* Copy ownership */
1734 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1736 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1740 /* Copy the protection modes */
1742 r = fchmod (ofd, mode);
1744 r = chmod (ofname, mode);
1748 WARN ((stderr, "%s: ", program_name));
1758 /* ========================================================================
1759 * Recurse through the given directory.
1761 local void treat_dir (fd, dir)
1766 char nbuf[MAX_PATH_LEN];
1771 dirp = fdopendir (fd);
1779 entries = streamsavedir (dirp);
1782 if (closedir (dirp) != 0)
1787 for (entry = entries; *entry; entry += entrylen + 1) {
1788 size_t len = strlen (dir);
1789 entrylen = strlen (entry);
1790 if (strequ (entry, ".") || strequ (entry, ".."))
1792 if (len + entrylen < MAX_PATH_LEN - 2) {
1794 if (len != 0 /* dir = "" means current dir on Amiga */
1796 && dir[len-1] != PATH_SEP2
1799 && dir[len-1] != PATH_SEP3
1802 nbuf[len++] = PATH_SEP;
1804 strcpy (nbuf + len, entry);
1807 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1808 program_name, dir, entry);
1814 #endif /* ! NO_DIR */
1816 /* Make sure signals get handled properly. */
1819 install_signal_handlers ()
1821 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1825 struct sigaction act;
1827 sigemptyset (&caught_signals);
1828 for (i = 0; i < nsigs; i++)
1830 sigaction (handled_sig[i], NULL, &act);
1831 if (act.sa_handler != SIG_IGN)
1832 sigaddset (&caught_signals, handled_sig[i]);
1835 act.sa_handler = abort_gzip_signal;
1836 act.sa_mask = caught_signals;
1839 for (i = 0; i < nsigs; i++)
1840 if (sigismember (&caught_signals, handled_sig[i]))
1844 sigaction (handled_sig[i], &act, NULL);
1847 for (i = 0; i < nsigs; i++)
1848 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1852 signal (handled_sig[i], abort_gzip_signal);
1853 siginterrupt (handled_sig[i], 1);
1858 /* ========================================================================
1859 * Free all dynamically allocated variables and exit with the given code.
1861 local void do_exit(exitcode)
1864 static int in_exit = 0;
1866 if (in_exit) exit(exitcode);
1885 /* ========================================================================
1886 * Close and unlink the output file.
1889 remove_output_file ()
1894 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1895 fd = remove_ofname_fd;
1898 remove_ofname_fd = -1;
1902 sigprocmask (SIG_SETMASK, &oldset, NULL);
1905 /* ========================================================================
1911 remove_output_file ();
1915 /* ========================================================================
1919 abort_gzip_signal (sig)
1923 signal (sig, SIG_IGN);
1924 remove_output_file ();
1925 if (sig == exiting_signal)
1927 signal (sig, SIG_DFL);