1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001-2002, 2006-2007, 2009-2010 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 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"
90 # ifndef _D_EXACT_NAMLEN
91 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
96 # define CLOSEDIR(d) (closedir(d), 0)
98 # define CLOSEDIR(d) closedir(d)
102 # include <utimens.h>
105 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
108 # define MAX_PATH_LEN 1024 /* max pathname length */
120 off_t lseek OF((int fd, off_t offset, int whence));
124 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
128 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
131 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
134 # define SA_NOCLDSTOP 0
135 # define sigprocmask(how, set, oset) /* empty */
136 # define sigset_t int
137 # if ! HAVE_SIGINTERRUPT
138 # define siginterrupt(sig, flag) /* empty */
142 #ifndef HAVE_WORKING_O_NOFOLLOW
143 # define HAVE_WORKING_O_NOFOLLOW 0
147 # define ELOOP EINVAL
150 /* Separator for file name parts (see shorten_name()) */
151 #ifdef NO_MULTIPLE_DOTS
152 # define PART_SEP "-"
154 # define PART_SEP "."
159 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
160 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
161 DECLARE(ush, d_buf, DIST_BUFSIZE);
162 DECLARE(uch, window, 2L*WSIZE);
164 DECLARE(ush, tab_prefix, 1L<<BITS);
166 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
167 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
170 /* local variables */
172 /* If true, pretend that standard input is a tty. This option
173 is deliberately not documented, and only for testing. */
174 static bool presume_input_tty;
176 int ascii = 0; /* convert end-of-lines to local OS conventions */
177 int to_stdout = 0; /* output to stdout (-c) */
178 int decompress = 0; /* decompress (-d) */
179 int force = 0; /* don't ask questions, compress links (-f) */
180 int no_name = -1; /* don't save or restore the original file name */
181 int no_time = -1; /* don't save or restore the original file time */
182 int recursive = 0; /* recurse through directories (-r) */
183 int list = 0; /* list the file contents (-l) */
184 int verbose = 0; /* be verbose (-v) */
185 int quiet = 0; /* be very quiet (-q) */
186 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
187 int test = 0; /* test .gz file integrity */
188 int foreground = 0; /* set if program run in foreground */
189 char *program_name; /* program name */
190 int maxbits = BITS; /* max bits per code for LZW */
191 int method = DEFLATED;/* compression method */
192 int level = 6; /* compression level */
193 int exit_code = OK; /* program exit code */
194 int save_orig_name; /* set if original name must be saved */
195 int last_member; /* set for .zip and .Z files */
196 int part_nb; /* number of parts in .gz file */
197 struct timespec time_stamp; /* original time stamp (modification time) */
198 off_t ifile_size; /* input file size, -1 for devices (debug only) */
199 char *env; /* contents of GZIP env variable */
200 char **args = NULL; /* argv pointer if GZIP env variable defined */
201 char const *z_suffix; /* default suffix (can be set with --suffix) */
202 size_t z_len; /* strlen(z_suffix) */
204 /* The set of signals that are caught. */
205 static sigset_t caught_signals;
207 /* If nonzero then exit with status WARNING, rather than with the usual
208 signal status, on receipt of a signal with this value. This
209 suppresses a "Broken Pipe" message with some shells. */
210 static int volatile exiting_signal;
212 /* If nonnegative, close this file descriptor and unlink ofname on error. */
213 static int volatile remove_ofname_fd = -1;
215 off_t bytes_in; /* number of input bytes */
216 off_t bytes_out; /* number of output bytes */
217 off_t total_in; /* input bytes for all files */
218 off_t total_out; /* output bytes for all files */
219 char ifname[MAX_PATH_LEN]; /* input file name */
220 char ofname[MAX_PATH_LEN]; /* output file name */
221 struct stat istat; /* status for input file */
222 int ifd; /* input file descriptor */
223 int ofd; /* output file descriptor */
224 unsigned insize; /* valid bytes in inbuf */
225 unsigned inptr; /* index of next byte to be processed in inbuf */
226 unsigned outcnt; /* bytes in output buffer */
228 static int handled_sig[] =
230 /* SIGINT must be first, as 'foreground' depends on it. */
252 /* For long options that have no equivalent short option, use a
253 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
256 PRESUME_INPUT_TTY_OPTION = CHAR_MAX + 1
259 struct option longopts[] =
261 /* { name has_arg *flag val } */
262 {"ascii", 0, 0, 'a'}, /* ascii text mode */
263 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
264 {"stdout", 0, 0, 'c'}, /* write output on standard output */
265 {"decompress", 0, 0, 'd'}, /* decompress */
266 {"uncompress", 0, 0, 'd'}, /* decompress */
267 /* {"encrypt", 0, 0, 'e'}, encrypt */
268 {"force", 0, 0, 'f'}, /* force overwrite of output file */
269 {"help", 0, 0, 'h'}, /* give help */
270 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
271 {"list", 0, 0, 'l'}, /* list .gz file contents */
272 {"license", 0, 0, 'L'}, /* display software license */
273 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
274 {"name", 0, 0, 'N'}, /* save or restore original name & time */
275 {"-presume-input-tty", no_argument, NULL, PRESUME_INPUT_TTY_OPTION},
276 {"quiet", 0, 0, 'q'}, /* quiet mode */
277 {"silent", 0, 0, 'q'}, /* quiet mode */
278 {"recursive", 0, 0, 'r'}, /* recurse through directories */
279 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
280 {"test", 0, 0, 't'}, /* test compressed file integrity */
281 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
282 {"verbose", 0, 0, 'v'}, /* verbose mode */
283 {"version", 0, 0, 'V'}, /* display version number */
284 {"fast", 0, 0, '1'}, /* compress faster */
285 {"best", 0, 0, '9'}, /* compress better */
286 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
287 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
292 /* local functions */
294 local void try_help OF((void)) ATTRIBUTE_NORETURN;
295 local void help OF((void));
296 local void license OF((void));
297 local void version OF((void));
298 local int input_eof OF((void));
299 local void treat_stdin OF((void));
300 local void treat_file OF((char *iname));
301 local int create_outfile OF((void));
302 local char *get_suffix OF((char *name));
303 local int open_input_file OF((char *iname, struct stat *sbuf));
304 local int make_ofname OF((void));
305 local void shorten_name OF((char *name));
306 local int get_method OF((int in));
307 local void do_list OF((int ifd, int method));
308 local int check_ofname OF((void));
309 local void copy_stat OF((struct stat *ifstat));
310 local void install_signal_handlers OF((void));
311 local void remove_output_file OF((void));
312 local RETSIGTYPE abort_gzip_signal OF((int));
313 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
314 int main OF((int argc, char **argv));
315 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
318 local void treat_dir OF((int fd, char *dir));
321 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
326 fprintf (stderr, "Try `%s --help' for more information.\n",
331 /* ======================================================================== */
334 static char const* const help_msg[] = {
335 "Compress or uncompress FILEs (by default, compress FILES in-place).",
337 "Mandatory arguments to long options are mandatory for short options too.",
340 " -a, --ascii ascii text; convert end-of-line using local conventions",
342 " -c, --stdout write on standard output, keep original files unchanged",
343 " -d, --decompress decompress",
344 /* -e, --encrypt encrypt */
345 " -f, --force force overwrite of output file and compress links",
346 " -h, --help give this help",
347 /* -k, --pkzip force output in pkzip format */
348 " -l, --list list compressed file contents",
349 " -L, --license display software license",
351 " -m, --no-time do not save or restore the original modification time",
352 " -M, --time save or restore the original modification time",
354 " -n, --no-name do not save or restore the original name and time stamp",
355 " -N, --name save or restore the original name and time stamp",
356 " -q, --quiet suppress all warnings",
358 " -r, --recursive operate recursively on directories",
360 " -S, --suffix=SUF use suffix SUF on compressed files",
361 " -t, --test test compressed file integrity",
362 " -v, --verbose verbose mode",
363 " -V, --version display version number",
364 " -1, --fast compress faster",
365 " -9, --best compress better",
367 " -Z, --lzw produce output compatible with old compress",
368 " -b, --bits=BITS max number of bits per code (implies -Z)",
371 "With no FILE, or when FILE is -, read standard input.",
373 "Report bugs to <bug-gzip@gnu.org>.",
375 char const *const *p = help_msg;
377 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
378 while (*p) printf ("%s\n", *p++);
381 /* ======================================================================== */
384 char const *const *p = license_msg;
386 printf ("%s %s\n", program_name, VERSION);
387 while (*p) printf ("%s\n", *p++);
390 /* ======================================================================== */
395 printf ("Written by Jean-loup Gailly.\n");
398 local void progerror (char const *string)
401 fprintf (stderr, "%s: ", program_name);
407 /* ======================================================================== */
408 int main (int argc, char **argv)
410 int file_count; /* number of files to process */
411 size_t proglen; /* length of program_name */
412 int optc; /* current option */
414 EXPAND(argc, argv); /* wild card expansion if necessary */
416 program_name = gzip_base_name (argv[0]);
417 proglen = strlen (program_name);
419 atexit (close_stdin);
421 /* Suppress .exe for MSDOS, OS/2 and VMS: */
422 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
423 program_name[proglen - 4] = '\0';
425 /* Add options in GZIP environment variable if there is one */
426 env = add_envopt(&argc, &argv, OPTIONS_VAR);
427 if (env != NULL) args = argv;
430 # define GNU_STANDARD 1
433 /* For compatibility with old compress, use program name as an option.
434 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
435 * gzip even if it is invoked under the name gunzip or zcat.
437 * Systems which do not support links can still use -d or -dc.
438 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
440 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
441 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
443 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
444 || strequ (program_name, "gzcat")) /* gzcat */
445 decompress = to_stdout = 1;
449 z_len = strlen(z_suffix);
451 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
452 longopts, (int *)0)) != -1) {
457 maxbits = atoi(optarg);
458 for (; *optarg; optarg++)
459 if (! ('0' <= *optarg && *optarg <= '9'))
461 fprintf (stderr, "%s: -b operand is not an integer\n",
467 to_stdout = 1; break;
469 decompress = 1; break;
473 help(); do_exit(OK); break;
475 list = decompress = to_stdout = 1; break;
477 license(); do_exit(OK); break;
478 case 'm': /* undocumented, may change later */
480 case 'M': /* undocumented, may change later */
483 no_name = no_time = 1; break;
485 no_name = no_time = 0; break;
486 case PRESUME_INPUT_TTY_OPTION:
487 presume_input_tty = true; break;
489 quiet = 1; verbose = 0; break;
492 fprintf (stderr, "%s: -r not supported on this system\n",
500 #ifdef NO_MULTIPLE_DOTS
501 if (*optarg == '.') optarg++;
503 z_len = strlen(optarg);
507 test = decompress = to_stdout = 1;
510 verbose++; quiet = 0; break;
512 version(); do_exit(OK); break;
517 fprintf(stderr, "%s: -Z not supported in this version\n",
522 case '1': case '2': case '3': case '4':
523 case '5': case '6': case '7': case '8': case '9':
527 /* Error message already emitted by getopt_long. */
530 } /* loop on all arguments */
532 /* By default, save name and timestamp on compression but do not
533 * restore them on decompression.
535 if (no_time < 0) no_time = decompress;
536 if (no_name < 0) no_name = decompress;
538 file_count = argc - optind;
542 if (ascii && !quiet) {
543 fprintf(stderr, "%s: option --ascii ignored on this system\n",
547 if (z_len == 0 || z_len > MAX_SUFFIX) {
548 fprintf(stderr, "%s: invalid suffix '%s'\n", program_name, z_suffix);
552 if (do_lzw && !decompress) work = lzw;
554 /* Allocate all global buffers (for DYN_ALLOC option) */
555 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
556 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
557 ALLOC(ush, d_buf, DIST_BUFSIZE);
558 ALLOC(uch, window, 2L*WSIZE);
560 ALLOC(ush, tab_prefix, 1L<<BITS);
562 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
563 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
566 exiting_signal = quiet ? SIGPIPE : 0;
567 install_signal_handlers ();
569 /* And get to work */
570 if (file_count != 0) {
571 if (to_stdout && !test && !list && (!decompress || !ascii)) {
572 SET_BINARY_MODE(fileno(stdout));
574 while (optind < argc) {
575 treat_file(argv[optind++]);
577 } else { /* Standard input */
580 if (list && !quiet && file_count > 1) {
581 do_list(-1, -1); /* print totals */
584 return exit_code; /* just to avoid lint warning */
587 /* Return nonzero when at end of file on input. */
591 if (!decompress || last_member)
596 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
599 /* Unget the char that fill_inbuf got. */
606 /* ========================================================================
607 * Compress or decompress stdin
609 local void treat_stdin()
612 && (presume_input_tty
613 || isatty(fileno((FILE *)(decompress ? stdin : stdout))))) {
614 /* Do not send compressed data to the terminal or read it from
615 * the terminal. We get here when user invoked the program
616 * without parameters, so be helpful. According to the GNU standards:
618 * If there is one behavior you think is most useful when the output
619 * is to a terminal, and another that you think is most useful when
620 * the output is a file or a pipe, then it is usually best to make
621 * the default behavior the one that is useful with output to a
622 * terminal, and have an option for the other behavior.
624 * Here we use the --force option to get the other behavior.
627 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
628 program_name, decompress ? "read from" : "written to",
629 decompress ? "de" : "");
630 fprintf (stderr, "For help, type: %s -h\n", program_name);
634 if (decompress || !ascii) {
635 SET_BINARY_MODE(fileno(stdin));
637 if (!test && !list && (!decompress || !ascii)) {
638 SET_BINARY_MODE(fileno(stdout));
640 strcpy(ifname, "stdin");
641 strcpy(ofname, "stdout");
643 /* Get the file's time stamp and size. */
644 if (fstat (fileno (stdin), &istat) != 0)
646 progerror ("standard input");
649 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
650 time_stamp.tv_nsec = -1;
651 if (!no_time || list)
653 if (S_ISREG (istat.st_mode))
654 time_stamp = get_stat_mtime (&istat);
656 gettime (&time_stamp);
659 clear_bufs(); /* clear input and output buffers */
665 method = get_method(ifd);
667 do_exit(exit_code); /* error message already emitted */
671 do_list(ifd, method);
675 /* Actually do the compression/decompression. Loop over zipped members.
678 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
683 method = get_method(ifd);
684 if (method < 0) return; /* error message already emitted */
685 bytes_out = 0; /* required for length check */
690 fprintf(stderr, " OK\n");
692 } else if (!decompress) {
693 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
694 fprintf(stderr, "\n");
695 #ifdef DISPLAY_STDIN_RATIO
697 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
698 fprintf(stderr, "\n");
704 /* ========================================================================
705 * Compress or decompress the given file
707 local void treat_file(iname)
710 /* Accept "-" as synonym for stdin */
711 if (strequ(iname, "-")) {
712 int cflag = to_stdout;
718 /* Check if the input file is present, set ifname and istat: */
719 ifd = open_input_file (iname, &istat);
723 /* If the input name is that of a directory, recurse or ignore: */
724 if (S_ISDIR(istat.st_mode)) {
727 treat_dir (ifd, iname);
728 /* Warning: ifname is now garbage */
733 WARN ((stderr, "%s: %s is a directory -- ignored\n",
734 program_name, ifname));
740 if (! S_ISREG (istat.st_mode))
743 "%s: %s is not a directory or a regular file - ignored\n",
744 program_name, ifname));
748 if (istat.st_mode & S_ISUID)
750 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
751 program_name, ifname));
755 if (istat.st_mode & S_ISGID)
757 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
758 program_name, ifname));
765 if (istat.st_mode & S_ISVTX)
768 "%s: %s has the sticky bit set - file ignored\n",
769 program_name, ifname));
773 if (2 <= istat.st_nlink)
775 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
776 program_name, ifname,
777 (unsigned long int) istat.st_nlink - 1,
778 istat.st_nlink == 2 ? ' ' : 's'));
785 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
786 time_stamp.tv_nsec = -1;
787 if (!no_time || list)
788 time_stamp = get_stat_mtime (&istat);
790 /* Generate output file name. For -r and (-t or -l), skip files
791 * without a valid gzip suffix (check done in make_ofname).
793 if (to_stdout && !list && !test) {
794 strcpy(ofname, "stdout");
796 } else if (make_ofname() != OK) {
801 clear_bufs(); /* clear input and output buffers */
805 method = get_method(ifd); /* updates ofname if original given */
808 return; /* error message already emitted */
812 do_list(ifd, method);
813 if (close (ifd) != 0)
818 /* If compressing to a file, check if ofname is not ambiguous
819 * because the operating system truncates names. Otherwise, generate
820 * a new ofname and save the original name in the compressed file.
823 ofd = fileno(stdout);
824 /* Keep remove_ofname_fd negative. */
826 if (create_outfile() != OK) return;
828 if (!decompress && save_orig_name && !verbose && !quiet) {
829 fprintf(stderr, "%s: %s compressed to %s\n",
830 program_name, ifname, ofname);
833 /* Keep the name even if not truncated except with --no-name: */
834 if (!save_orig_name) save_orig_name = !no_name;
837 fprintf(stderr, "%s:\t", ifname);
840 /* Actually do the compression/decompression. Loop over zipped members.
843 if ((*work)(ifd, ofd) != OK) {
844 method = -1; /* force cleanup */
851 method = get_method(ifd);
852 if (method < 0) break; /* error message already emitted */
853 bytes_out = 0; /* required for length check */
856 if (close (ifd) != 0)
865 if (close (ofd) != 0)
868 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
869 remove_ofname_fd = -1;
870 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
871 sigprocmask (SIG_SETMASK, &oldset, NULL);
875 WARN ((stderr, "%s: ", program_name));
878 errno = unlink_errno;
886 remove_output_file ();
890 /* Display statistics */
893 fprintf(stderr, " OK");
894 } else if (decompress) {
895 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
897 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
899 if (!test && !to_stdout) {
900 fprintf(stderr, " -- replaced with %s", ofname);
902 fprintf(stderr, "\n");
906 /* ========================================================================
907 * Create the output file. Return OK or ERROR.
908 * Try several times if necessary to avoid truncating the z_suffix. For
909 * example, do not create a compressed file of name "1234567890123."
910 * Sets save_orig_name to true if the file name has been truncated.
911 * IN assertions: the input file has already been open (ifd is set) and
912 * ofname has already been updated if there was an original name.
913 * OUT assertions: ifd and ofd are closed in case of error.
915 local int create_outfile()
917 int name_shortened = 0;
918 int flags = (O_WRONLY | O_CREAT | O_EXCL
919 | (ascii && decompress ? 0 : O_BINARY));
926 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
927 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
929 sigprocmask (SIG_SETMASK, &oldset, NULL);
938 shorten_name (ofname);
944 if (check_ofname () != OK)
958 if (name_shortened && decompress)
960 /* name might be too long if an original name was saved */
961 WARN ((stderr, "%s: %s: warning, name truncated\n",
962 program_name, ofname));
968 /* ========================================================================
969 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
970 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
971 * accepted suffixes, in addition to the value of the --suffix option.
972 * ".tgz" is a useful convention for tar.z files on systems limited
973 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
974 * also accepted suffixes. For Unix, we do not want to accept any
975 * .??z suffix as indicating a compressed file; some people use .xyz
976 * to denote volume data.
977 * On systems allowing multiple versions of the same file (such as VMS),
978 * this function removes any version suffix in the given name.
980 local char *get_suffix(name)
984 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
985 static char const *known_suffixes[] =
986 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
991 char const **suf = known_suffixes;
994 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
997 /* strip a version number from the file name */
999 char *v = strrchr(name, SUFFIX_SEP);
1000 if (v != NULL) *v = '\0';
1003 nlen = strlen(name);
1004 if (nlen <= MAX_SUFFIX+2) {
1005 strcpy(suffix, name);
1007 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1010 slen = strlen(suffix);
1012 int s = strlen(*suf);
1013 if (slen > s && suffix[slen-s-1] != PATH_SEP
1014 && strequ(suffix + slen - s, *suf)) {
1017 } while (*++suf != NULL);
1023 /* Open file NAME with the given flags and mode and store its status
1024 into *ST. Return a file descriptor to the newly opened file, or -1
1025 (setting errno) on failure. */
1027 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1031 /* Refuse to follow symbolic links unless -c or -f. */
1032 if (!to_stdout && !force)
1034 if (HAVE_WORKING_O_NOFOLLOW)
1035 flags |= O_NOFOLLOW;
1038 #if HAVE_LSTAT || defined lstat
1039 if (lstat (name, st) != 0)
1041 else if (S_ISLNK (st->st_mode))
1050 fd = OPEN (name, flags, mode);
1051 if (0 <= fd && fstat (fd, st) != 0)
1062 /* ========================================================================
1063 * Set ifname to the input file name (with a suffix appended if necessary)
1064 * and istat to its stats. For decompression, if no file exists with the
1065 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1066 * For MSDOS, we try only z_suffix and z.
1067 * Return an open file descriptor or -1.
1070 open_input_file (iname, sbuf)
1074 int ilen; /* strlen(ifname) */
1075 int z_suffix_errno = 0;
1076 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1077 char const **suf = suffixes;
1079 #ifdef NO_MULTIPLE_DOTS
1080 char *dot; /* pointer to ifname extension, or NULL */
1083 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1084 | (ascii && !decompress ? 0 : O_BINARY));
1088 if (sizeof ifname - 1 <= strlen (iname))
1091 strcpy(ifname, iname);
1093 /* If input file exists, return OK. */
1094 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1098 if (!decompress || errno != ENOENT) {
1102 /* file.ext doesn't exist, try adding a suffix (after removing any
1103 * version number for VMS).
1105 s = get_suffix(ifname);
1107 progerror(ifname); /* ifname already has z suffix and does not exist */
1110 #ifdef NO_MULTIPLE_DOTS
1111 dot = strrchr(ifname, '.');
1113 strcat(ifname, ".");
1114 dot = strrchr(ifname, '.');
1117 ilen = strlen(ifname);
1118 if (strequ(z_suffix, ".gz")) suf++;
1120 /* Search for all suffixes */
1122 char const *s0 = s = *suf;
1123 strcpy (ifname, iname);
1124 #ifdef NO_MULTIPLE_DOTS
1126 if (*dot == '\0') strcpy (dot, ".");
1128 #ifdef MAX_EXT_CHARS
1129 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1130 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1132 if (sizeof ifname <= ilen + strlen (s))
1135 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1138 if (errno != ENOENT)
1143 if (strequ (s0, z_suffix))
1144 z_suffix_errno = errno;
1145 } while (*++suf != NULL);
1147 /* No suffix found, complain using z_suffix: */
1148 strcpy(ifname, iname);
1149 #ifdef NO_MULTIPLE_DOTS
1150 if (*dot == '\0') strcpy(dot, ".");
1152 #ifdef MAX_EXT_CHARS
1153 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1154 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1156 strcat(ifname, z_suffix);
1157 errno = z_suffix_errno;
1162 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1167 /* ========================================================================
1168 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1169 * Sets save_orig_name to true if the file name has been truncated.
1171 local int make_ofname()
1173 char *suff; /* ofname z suffix */
1175 strcpy(ofname, ifname);
1176 /* strip a version number if any and get the gzip suffix if present: */
1177 suff = get_suffix(ofname);
1181 /* With -t or -l, try all files (even without .gz suffix)
1182 * except with -r (behave as with just -dr).
1184 if (!recursive && (list || test)) return OK;
1186 /* Avoid annoying messages with -r */
1187 if (verbose || (!recursive && !quiet)) {
1188 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1189 program_name, ifname));
1193 /* Make a special case for .tgz and .taz: */
1195 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1196 strcpy(suff, ".tar");
1198 *suff = '\0'; /* strip the z suffix */
1200 /* ofname might be changed later if infile contains an original name */
1202 } else if (suff && ! force) {
1203 /* Avoid annoying messages with -r (see treat_dir()) */
1204 if (verbose || (!recursive && !quiet)) {
1205 /* Don't use WARN, as it affects exit status. */
1206 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1207 program_name, ifname, suff);
1213 #ifdef NO_MULTIPLE_DOTS
1214 suff = strrchr(ofname, '.');
1216 if (sizeof ofname <= strlen (ofname) + 1)
1218 strcat(ofname, ".");
1219 # ifdef MAX_EXT_CHARS
1220 if (strequ(z_suffix, "z")) {
1221 if (sizeof ofname <= strlen (ofname) + 2)
1223 strcat(ofname, "gz"); /* enough room */
1226 /* On the Atari and some versions of MSDOS,
1227 * ENAMETOOLONG does not work correctly. So we
1228 * must truncate here.
1230 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1231 suff[MAX_SUFFIX+1-z_len] = '\0';
1235 #endif /* NO_MULTIPLE_DOTS */
1236 if (sizeof ofname <= strlen (ofname) + z_len)
1238 strcat(ofname, z_suffix);
1240 } /* decompress ? */
1244 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1249 /* ========================================================================
1250 * Check the magic number of the input file and update ofname if an
1251 * original name was given and to_stdout is not set.
1252 * Return the compression method, -1 for error, -2 for warning.
1253 * Set inptr to the offset of the next byte to be processed.
1254 * Updates time_stamp if there is one and --no-time is not used.
1255 * This function may be called repeatedly for an input file consisting
1256 * of several contiguous gzip'ed members.
1257 * IN assertions: there is at least one remaining compressed member.
1258 * If the member is a zip file, it must be the only one.
1260 local int get_method(in)
1261 int in; /* input file descriptor */
1263 uch flags; /* compression flags */
1264 char magic[2]; /* magic header */
1265 int imagic0; /* first magic byte or EOF */
1266 int imagic1; /* like magic[1], but can represent EOF */
1267 ulg stamp; /* time stamp */
1269 /* If --force and --stdout, zcat == cat, so do not complain about
1270 * premature end of file: use try_byte instead of get_byte.
1272 if (force && to_stdout) {
1273 imagic0 = try_byte();
1274 magic[0] = (char) imagic0;
1275 imagic1 = try_byte ();
1276 magic[1] = (char) imagic1;
1277 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1279 magic[0] = (char)get_byte();
1282 magic[1] = (char)get_byte();
1283 imagic1 = 0; /* avoid lint warning */
1285 imagic1 = try_byte ();
1286 magic[1] = (char) imagic1;
1289 method = -1; /* unknown yet */
1290 part_nb++; /* number of parts in gzip file */
1292 last_member = RECORD_IO;
1293 /* assume multiple members in gzip file except for record oriented I/O */
1295 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1296 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1298 method = (int)get_byte();
1299 if (method != DEFLATED) {
1301 "%s: %s: unknown method %d -- not supported\n",
1302 program_name, ifname, method);
1307 flags = (uch)get_byte();
1309 if ((flags & ENCRYPTED) != 0) {
1311 "%s: %s is encrypted -- not supported\n",
1312 program_name, ifname);
1316 if ((flags & CONTINUATION) != 0) {
1318 "%s: %s is a multi-part gzip file -- not supported\n",
1319 program_name, ifname);
1321 if (force <= 1) return -1;
1323 if ((flags & RESERVED) != 0) {
1325 "%s: %s has flags 0x%x -- not supported\n",
1326 program_name, ifname, flags);
1328 if (force <= 1) return -1;
1330 stamp = (ulg)get_byte();
1331 stamp |= ((ulg)get_byte()) << 8;
1332 stamp |= ((ulg)get_byte()) << 16;
1333 stamp |= ((ulg)get_byte()) << 24;
1334 if (stamp != 0 && !no_time)
1336 time_stamp.tv_sec = stamp;
1337 time_stamp.tv_nsec = 0;
1340 (void)get_byte(); /* Ignore extra flags for the moment */
1341 (void)get_byte(); /* Ignore OS type for the moment */
1343 if ((flags & CONTINUATION) != 0) {
1344 unsigned part = (unsigned)get_byte();
1345 part |= ((unsigned)get_byte())<<8;
1347 fprintf(stderr,"%s: %s: part number %u\n",
1348 program_name, ifname, part);
1351 if ((flags & EXTRA_FIELD) != 0) {
1352 unsigned len = (unsigned)get_byte();
1353 len |= ((unsigned)get_byte())<<8;
1355 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1356 program_name, ifname, len);
1358 while (len--) (void)get_byte();
1361 /* Get original file name if it was truncated */
1362 if ((flags & ORIG_NAME) != 0) {
1363 if (no_name || (to_stdout && !list) || part_nb > 1) {
1364 /* Discard the old name */
1365 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1366 do {c=get_byte();} while (c != 0);
1368 /* Copy the base name. Keep a directory prefix intact. */
1369 char *p = gzip_base_name (ofname);
1372 *p = (char)get_char();
1373 if (*p++ == '\0') break;
1374 if (p >= ofname+sizeof(ofname)) {
1375 gzip_error ("corrupted input -- file name too large");
1378 p = gzip_base_name (base);
1379 memmove (base, p, strlen (p) + 1);
1380 /* If necessary, adapt the name to local OS conventions: */
1382 MAKE_LEGAL_NAME(base);
1383 if (base) list=0; /* avoid warning about unused variable */
1385 } /* no_name || to_stdout */
1388 /* Discard file comment if any */
1389 if ((flags & COMMENT) != 0) {
1390 while (get_char() != 0) /* null */ ;
1393 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1396 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1397 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1398 /* To simplify the code, we support a zip file when alone only.
1399 * We are thus guaranteed that the entire local header fits in inbuf.
1403 if (check_zipfile(in) != OK) return -1;
1404 /* check_zipfile may get ofname from the local header */
1407 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1411 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1413 method = COMPRESSED;
1416 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1421 } else if (force && to_stdout && !list) { /* pass input unchanged */
1427 if (imagic0 != EOF) {
1428 write_buf(fileno(stdout), magic, 1);
1432 if (method >= 0) return method;
1435 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1436 program_name, ifname);
1443 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1448 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1449 program_name, ifname));
1454 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1455 program_name, ifname));
1460 /* ========================================================================
1461 * Display the characteristics of the compressed file.
1462 * If the given method is < 0, display the accumulated totals.
1463 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1465 local void do_list(ifd, method)
1466 int ifd; /* input file descriptor */
1467 int method; /* compression method */
1469 ulg crc; /* original crc */
1470 static int first_time = 1;
1471 static char const *const methods[MAX_METHODS] = {
1476 "", "", "", "", /* 4 to 7 reserved */
1478 int positive_off_t_width = 1;
1481 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1482 positive_off_t_width++;
1485 if (first_time && method >= 0) {
1488 printf("method crc date time ");
1491 printf("%*.*s %*.*s ratio uncompressed_name\n",
1492 positive_off_t_width, positive_off_t_width, "compressed",
1493 positive_off_t_width, positive_off_t_width, "uncompressed");
1495 } else if (method < 0) {
1496 if (total_in <= 0 || total_out <= 0) return;
1500 if (verbose || !quiet) {
1501 fprint_off(stdout, total_in, positive_off_t_width);
1503 fprint_off(stdout, total_out, positive_off_t_width);
1506 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1507 /* header_bytes is not meaningful but used to ensure the same
1508 * ratio if there is a single file.
1510 printf(" (totals)\n");
1513 crc = (ulg)~0; /* unknown */
1515 bytes_in = ifile_size;
1518 if (method == DEFLATED && !last_member) {
1519 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1520 * If the lseek fails, we could use read() to get to the end, but
1521 * --list is used to get quick results.
1522 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1523 * you are not concerned about speed.
1525 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1526 if (bytes_in != -1L) {
1529 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1533 bytes_out = LG(buf+4);
1536 #endif /* RECORD_IO */
1539 struct tm *tm = localtime (&time_stamp.tv_sec);
1540 printf ("%5s %08lx ", methods[method], crc);
1542 printf ("%s%3d %02d:%02d ",
1543 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1545 tm->tm_mday, tm->tm_hour, tm->tm_min);
1547 printf ("??? ?? ??:?? ");
1549 fprint_off(stdout, bytes_in, positive_off_t_width);
1551 fprint_off(stdout, bytes_out, positive_off_t_width);
1553 if (bytes_in == -1L) {
1555 bytes_in = bytes_out = header_bytes = 0;
1556 } else if (total_in >= 0) {
1557 total_in += bytes_in;
1559 if (bytes_out == -1L) {
1561 bytes_in = bytes_out = header_bytes = 0;
1562 } else if (total_out >= 0) {
1563 total_out += bytes_out;
1565 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1566 printf(" %s\n", ofname);
1569 /* ========================================================================
1570 * Shorten the given name by one character, or replace a .tar extension
1571 * with .tgz. Truncate the last part of the name which is longer than
1572 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1573 * has only parts shorter than MIN_PART truncate the longest part.
1574 * For decompression, just remove the last character of the name.
1576 * IN assertion: for compression, the suffix of the given name is z_suffix.
1578 local void shorten_name(name)
1581 int len; /* length of name without z_suffix */
1582 char *trunc = NULL; /* character to be truncated */
1583 int plen; /* current part length */
1584 int min_part = MIN_PART; /* current minimum part length */
1590 gzip_error ("name too short");
1594 p = get_suffix(name);
1596 gzip_error ("can't recover suffix\n");
1600 /* compress 1234567890.tar to 1234567890.tgz */
1601 if (len > 4 && strequ(p-4, ".tar")) {
1602 strcpy(p-4, ".tgz");
1605 /* Try keeping short extensions intact:
1606 * 1234.678.012.gz -> 123.678.012.gz
1609 p = strrchr(name, PATH_SEP);
1612 plen = strcspn(p, PART_SEP);
1614 if (plen > min_part) trunc = p-1;
1617 } while (trunc == NULL && --min_part != 0);
1619 if (trunc != NULL) {
1621 trunc[0] = trunc[1];
1625 trunc = strrchr(name, PART_SEP[0]);
1627 gzip_error ("internal error in shorten_name");
1628 if (trunc[1] == '\0') trunc--; /* force truncation */
1630 strcpy(trunc, z_suffix);
1633 /* ========================================================================
1634 * The compressed file already exists, so ask for confirmation.
1635 * Return ERROR if the file must be skipped.
1637 local int check_ofname()
1639 /* Ask permission to overwrite the existing file */
1642 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1643 if (foreground && (presume_input_tty || isatty(fileno(stdin)))) {
1644 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1649 fprintf(stderr, "\tnot overwritten\n");
1650 if (exit_code == OK) exit_code = WARNING;
1654 if (xunlink (ofname)) {
1662 /* ========================================================================
1663 * Copy modes, times, ownership from input file to output file.
1664 * IN assertion: to_stdout is false.
1666 local void copy_stat(ifstat)
1667 struct stat *ifstat;
1669 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1673 struct timespec timespec[2];
1674 timespec[0] = get_stat_atime (ifstat);
1675 timespec[1] = get_stat_mtime (ifstat);
1677 if (decompress && 0 <= time_stamp.tv_nsec
1678 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1679 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1681 timespec[1] = time_stamp;
1683 fprintf(stderr, "%s: time stamp restored\n", ofname);
1687 if (gl_futimens (ofd, ofname, timespec) != 0)
1690 WARN ((stderr, "%s: ", program_name));
1700 /* Copy ownership */
1702 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1704 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1708 /* Copy the protection modes */
1710 r = fchmod (ofd, mode);
1712 r = chmod (ofname, mode);
1716 WARN ((stderr, "%s: ", program_name));
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 /* Make sure signals get handled properly. */
1797 install_signal_handlers ()
1799 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1803 struct sigaction act;
1805 sigemptyset (&caught_signals);
1806 for (i = 0; i < nsigs; i++)
1808 sigaction (handled_sig[i], NULL, &act);
1809 if (act.sa_handler != SIG_IGN)
1810 sigaddset (&caught_signals, handled_sig[i]);
1813 act.sa_handler = abort_gzip_signal;
1814 act.sa_mask = caught_signals;
1817 for (i = 0; i < nsigs; i++)
1818 if (sigismember (&caught_signals, handled_sig[i]))
1822 sigaction (handled_sig[i], &act, NULL);
1825 for (i = 0; i < nsigs; i++)
1826 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1830 signal (handled_sig[i], abort_gzip_signal);
1831 siginterrupt (handled_sig[i], 1);
1836 /* ========================================================================
1837 * Free all dynamically allocated variables and exit with the given code.
1839 local void do_exit(exitcode)
1842 static int in_exit = 0;
1844 if (in_exit) exit(exitcode);
1863 /* ========================================================================
1864 * Close and unlink the output file.
1867 remove_output_file ()
1872 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1873 fd = remove_ofname_fd;
1876 remove_ofname_fd = -1;
1880 sigprocmask (SIG_SETMASK, &oldset, NULL);
1883 /* ========================================================================
1889 remove_output_file ();
1893 /* ========================================================================
1897 abort_gzip_signal (sig)
1901 signal (sig, SIG_IGN);
1902 remove_output_file ();
1903 if (sig == exiting_signal)
1905 signal (sig, SIG_DFL);