1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001, 2002, 2006, 2007 Free Software Foundation, Inc.
4 Copyright (C) 1992-1993 Jean-loup Gailly
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
21 * The unzip code was written and put in the public domain by Mark Adler.
22 * Portions of the lzw code are derived from the public domain 'compress'
23 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
24 * Ken Turkowski, Dave Mack and Peter Jannesen.
26 * See the license_msg below and the file COPYING for the software license.
27 * See the file algorithm.doc for the compression algorithms and file formats.
30 static char *license_msg[] = {
31 "Copyright (C) 2007 Free Software Foundation, Inc.",
32 "Copyright (C) 1993 Jean-loup Gailly.",
33 "This is free software. You may redistribute copies of it under the terms of",
34 "the GNU General Public License <http://www.gnu.org/licenses/gpl.html>.",
35 "There is NO WARRANTY, to the extent permitted by law.",
38 /* Compress files with zip algorithm and 'compress' interface.
39 * See help() function below for all options.
41 * file.gz: compressed file with same mode, owner, and utimes
42 * or stdout with -c option or if stdin used as input.
43 * If the output file name had to be truncated, the original name is kept
44 * in the compressed file.
45 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
47 * Using gz on MSDOS would create too many file name conflicts. For
48 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
49 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
50 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
51 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
53 * For the meaning of all compilation flags, see comments in Makefile.in.
57 static char rcsid[] = "$Id: gzip.c,v 1.16 2007/03/20 05:09:51 eggert Exp $";
62 #include <sys/types.h>
72 #include "fcntl-safer.h"
74 #include "stat-time.h"
90 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
101 # ifndef _D_EXACT_NAMLEN
102 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
107 # define CLOSEDIR(d) (closedir(d), 0)
109 # define CLOSEDIR(d) closedir(d)
113 # include <utimens.h>
116 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
119 # define MAX_PATH_LEN 1024 /* max pathname length */
131 off_t lseek OF((int fd, off_t offset, int whence));
135 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
139 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
142 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
145 # define SA_NOCLDSTOP 0
146 # define sigprocmask(how, set, oset) /* empty */
147 # define sigset_t int
148 # if ! HAVE_SIGINTERRUPT
149 # define siginterrupt(sig, flag) /* empty */
153 #ifndef HAVE_WORKING_O_NOFOLLOW
154 # define HAVE_WORKING_O_NOFOLLOW 0
158 # define ELOOP EINVAL
161 /* Separator for file name parts (see shorten_name()) */
162 #ifdef NO_MULTIPLE_DOTS
163 # define PART_SEP "-"
165 # define PART_SEP "."
170 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
171 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
172 DECLARE(ush, d_buf, DIST_BUFSIZE);
173 DECLARE(uch, window, 2L*WSIZE);
175 DECLARE(ush, tab_prefix, 1L<<BITS);
177 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
178 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
181 /* local variables */
183 int ascii = 0; /* convert end-of-lines to local OS conventions */
184 int to_stdout = 0; /* output to stdout (-c) */
185 int decompress = 0; /* decompress (-d) */
186 int force = 0; /* don't ask questions, compress links (-f) */
187 int no_name = -1; /* don't save or restore the original file name */
188 int no_time = -1; /* don't save or restore the original file time */
189 int recursive = 0; /* recurse through directories (-r) */
190 int list = 0; /* list the file contents (-l) */
191 int verbose = 0; /* be verbose (-v) */
192 int quiet = 0; /* be very quiet (-q) */
193 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
194 int test = 0; /* test .gz file integrity */
195 int foreground = 0; /* set if program run in foreground */
196 char *program_name; /* program name */
197 int maxbits = BITS; /* max bits per code for LZW */
198 int method = DEFLATED;/* compression method */
199 int level = 6; /* compression level */
200 int exit_code = OK; /* program exit code */
201 int save_orig_name; /* set if original name must be saved */
202 int last_member; /* set for .zip and .Z files */
203 int part_nb; /* number of parts in .gz file */
204 struct timespec time_stamp; /* original time stamp (modification time) */
205 off_t ifile_size; /* input file size, -1 for devices (debug only) */
206 char *env; /* contents of GZIP env variable */
207 char **args = NULL; /* argv pointer if GZIP env variable defined */
208 char *z_suffix; /* default suffix (can be set with --suffix) */
209 size_t z_len; /* strlen(z_suffix) */
211 /* The set of signals that are caught. */
212 static sigset_t caught_signals;
214 /* If nonzero then exit with status WARNING, rather than with the usual
215 signal status, on receipt of a signal with this value. This
216 suppresses a "Broken Pipe" message with some shells. */
217 static int volatile exiting_signal;
219 /* If nonnegative, close this file descriptor and unlink ofname on error. */
220 static int volatile remove_ofname_fd = -1;
222 off_t bytes_in; /* number of input bytes */
223 off_t bytes_out; /* number of output bytes */
224 off_t total_in; /* input bytes for all files */
225 off_t total_out; /* output bytes for all files */
226 char ifname[MAX_PATH_LEN]; /* input file name */
227 char ofname[MAX_PATH_LEN]; /* output file name */
228 struct stat istat; /* status for input file */
229 int ifd; /* input file descriptor */
230 int ofd; /* output file descriptor */
231 unsigned insize; /* valid bytes in inbuf */
232 unsigned inptr; /* index of next byte to be processed in inbuf */
233 unsigned outcnt; /* bytes in output buffer */
234 int rsync = 0; /* make ryncable chunks */
236 static int handled_sig[] =
238 /* SIGINT must be first, as 'foreground' depends on it. */
260 struct option longopts[] =
262 /* { name has_arg *flag val } */
263 {"ascii", 0, 0, 'a'}, /* ascii text mode */
264 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
265 {"stdout", 0, 0, 'c'}, /* write output on standard output */
266 {"decompress", 0, 0, 'd'}, /* decompress */
267 {"uncompress", 0, 0, 'd'}, /* decompress */
268 /* {"encrypt", 0, 0, 'e'}, encrypt */
269 {"force", 0, 0, 'f'}, /* force overwrite of output file */
270 {"help", 0, 0, 'h'}, /* give help */
271 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
272 {"list", 0, 0, 'l'}, /* list .gz file contents */
273 {"license", 0, 0, 'L'}, /* display software license */
274 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
275 {"name", 0, 0, 'N'}, /* save or restore original name & time */
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) */
288 {"rsyncable", 0, 0, 'R'}, /* make rsync-friendly archive */
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 *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)",
370 " --rsyncable Make rsync-friendly archive",
372 "With no FILE, or when FILE is -, read standard input.",
374 "Report bugs to <bug-gzip@gnu.org>.",
378 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
379 while (*p) printf ("%s\n", *p++);
382 /* ======================================================================== */
385 char **p = license_msg;
387 printf ("%s %s\n", program_name, VERSION);
388 while (*p) printf ("%s\n", *p++);
391 /* ======================================================================== */
396 printf ("Written by Jean-loup Gailly.\n");
399 local void progerror (string)
403 fprintf (stderr, "%s: ", program_name);
409 /* ======================================================================== */
410 int main (argc, argv)
414 int file_count; /* number of files to process */
415 size_t proglen; /* length of program_name */
416 int optc; /* current option */
418 EXPAND(argc, argv); /* wild card expansion if necessary */
420 program_name = gzip_base_name (argv[0]);
421 proglen = strlen (program_name);
423 /* Suppress .exe for MSDOS, OS/2 and VMS: */
424 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
425 program_name[proglen - 4] = '\0';
427 /* Add options in GZIP environment variable if there is one */
428 env = add_envopt(&argc, &argv, OPTIONS_VAR);
429 if (env != NULL) args = argv;
432 # define GNU_STANDARD 1
435 /* For compatibility with old compress, use program name as an option.
436 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
437 * gzip even if it is invoked under the name gunzip or zcat.
439 * Systems which do not support links can still use -d or -dc.
440 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
442 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
443 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
445 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
446 || strequ (program_name, "gzcat")) /* gzcat */
447 decompress = to_stdout = 1;
451 z_len = strlen(z_suffix);
453 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
454 longopts, (int *)0)) != -1) {
459 maxbits = atoi(optarg);
460 for (; *optarg; optarg++)
461 if (! ('0' <= *optarg && *optarg <= '9'))
463 fprintf (stderr, "%s: -b operand is not an integer\n",
469 to_stdout = 1; break;
471 decompress = 1; break;
475 help(); do_exit(OK); break;
477 list = decompress = to_stdout = 1; break;
479 license(); do_exit(OK); break;
480 case 'm': /* undocumented, may change later */
482 case 'M': /* undocumented, may change later */
485 no_name = no_time = 1; break;
487 no_name = no_time = 0; break;
489 quiet = 1; verbose = 0; break;
492 fprintf (stderr, "%s: -r not supported on this system\n",
502 #ifdef NO_MULTIPLE_DOTS
503 if (*optarg == '.') optarg++;
505 z_len = strlen(optarg);
509 test = decompress = to_stdout = 1;
512 verbose++; quiet = 0; break;
514 version(); do_exit(OK); break;
519 fprintf(stderr, "%s: -Z not supported in this version\n",
524 case '1': case '2': case '3': case '4':
525 case '5': case '6': case '7': case '8': case '9':
529 /* Error message already emitted by getopt_long. */
532 } /* loop on all arguments */
534 /* By default, save name and timestamp on compression but do not
535 * restore them on decompression.
537 if (no_time < 0) no_time = decompress;
538 if (no_name < 0) no_name = decompress;
540 file_count = argc - optind;
544 if (ascii && !quiet) {
545 fprintf(stderr, "%s: option --ascii ignored on this system\n",
549 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
550 fprintf(stderr, "%s: incorrect suffix '%s'\n",
551 program_name, z_suffix);
554 if (do_lzw && !decompress) work = lzw;
556 /* Allocate all global buffers (for DYN_ALLOC option) */
557 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
558 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
559 ALLOC(ush, d_buf, DIST_BUFSIZE);
560 ALLOC(uch, window, 2L*WSIZE);
562 ALLOC(ush, tab_prefix, 1L<<BITS);
564 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
565 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
568 exiting_signal = quiet ? SIGPIPE : 0;
569 install_signal_handlers ();
571 /* And get to work */
572 if (file_count != 0) {
573 if (to_stdout && !test && !list && (!decompress || !ascii)) {
574 SET_BINARY_MODE(fileno(stdout));
576 while (optind < argc) {
577 treat_file(argv[optind++]);
579 } else { /* Standard input */
582 if (list && !quiet && file_count > 1) {
583 do_list(-1, -1); /* print totals */
586 return exit_code; /* just to avoid lint warning */
589 /* Return nonzero when at end of file on input. */
593 if (!decompress || last_member)
598 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
601 /* Unget the char that fill_inbuf got. */
608 /* ========================================================================
609 * Compress or decompress stdin
611 local void treat_stdin()
613 if (!force && !list &&
614 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
615 /* Do not send compressed data to the terminal or read it from
616 * the terminal. We get here when user invoked the program
617 * without parameters, so be helpful. According to the GNU standards:
619 * If there is one behavior you think is most useful when the output
620 * is to a terminal, and another that you think is most useful when
621 * the output is a file or a pipe, then it is usually best to make
622 * the default behavior the one that is useful with output to a
623 * terminal, and have an option for the other behavior.
625 * Here we use the --force option to get the other behavior.
628 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
629 program_name, decompress ? "read from" : "written to",
630 decompress ? "de" : "");
631 fprintf (stderr, "For help, type: %s -h\n", program_name);
635 if (decompress || !ascii) {
636 SET_BINARY_MODE(fileno(stdin));
638 if (!test && !list && (!decompress || !ascii)) {
639 SET_BINARY_MODE(fileno(stdout));
641 strcpy(ifname, "stdin");
642 strcpy(ofname, "stdout");
644 /* Get the file's time stamp and size. */
645 if (fstat (fileno (stdin), &istat) != 0)
647 progerror ("standard input");
650 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
651 time_stamp.tv_nsec = -1;
652 if (!no_time || list)
653 time_stamp = get_stat_mtime (&istat);
655 clear_bufs(); /* clear input and output buffers */
660 method = get_method(ifd);
662 do_exit(exit_code); /* error message already emitted */
666 do_list(ifd, method);
670 /* Actually do the compression/decompression. Loop over zipped members.
673 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
678 method = get_method(ifd);
679 if (method < 0) return; /* error message already emitted */
680 bytes_out = 0; /* required for length check */
685 fprintf(stderr, " OK\n");
687 } else if (!decompress) {
688 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
689 fprintf(stderr, "\n");
690 #ifdef DISPLAY_STDIN_RATIO
692 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
693 fprintf(stderr, "\n");
699 /* ========================================================================
700 * Compress or decompress the given file
702 local void treat_file(iname)
705 /* Accept "-" as synonym for stdin */
706 if (strequ(iname, "-")) {
707 int cflag = to_stdout;
713 /* Check if the input file is present, set ifname and istat: */
714 ifd = open_input_file (iname, &istat);
718 /* If the input name is that of a directory, recurse or ignore: */
719 if (S_ISDIR(istat.st_mode)) {
722 treat_dir (ifd, iname);
723 /* Warning: ifname is now garbage */
728 WARN ((stderr, "%s: %s is a directory -- ignored\n",
729 program_name, ifname));
735 if (! S_ISREG (istat.st_mode))
738 "%s: %s is not a directory or a regular file - ignored\n",
739 program_name, ifname));
743 if (istat.st_mode & S_ISUID)
745 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
746 program_name, ifname));
750 if (istat.st_mode & S_ISGID)
752 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
753 program_name, ifname));
760 if (istat.st_mode & S_ISVTX)
763 "%s: %s has the sticky bit set - file ignored\n",
764 program_name, ifname));
768 if (2 <= istat.st_nlink)
770 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
771 program_name, ifname,
772 (unsigned long int) istat.st_nlink - 1,
773 istat.st_nlink == 2 ? ' ' : 's'));
780 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
781 time_stamp.tv_nsec = -1;
782 if (!no_time || list)
783 time_stamp = get_stat_mtime (&istat);
785 /* Generate output file name. For -r and (-t or -l), skip files
786 * without a valid gzip suffix (check done in make_ofname).
788 if (to_stdout && !list && !test) {
789 strcpy(ofname, "stdout");
791 } else if (make_ofname() != OK) {
796 clear_bufs(); /* clear input and output buffers */
800 method = get_method(ifd); /* updates ofname if original given */
803 return; /* error message already emitted */
807 do_list(ifd, method);
808 if (close (ifd) != 0)
813 /* If compressing to a file, check if ofname is not ambiguous
814 * because the operating system truncates names. Otherwise, generate
815 * a new ofname and save the original name in the compressed file.
818 ofd = fileno(stdout);
819 /* Keep remove_ofname_fd negative. */
821 if (create_outfile() != OK) return;
823 if (!decompress && save_orig_name && !verbose && !quiet) {
824 fprintf(stderr, "%s: %s compressed to %s\n",
825 program_name, ifname, ofname);
828 /* Keep the name even if not truncated except with --no-name: */
829 if (!save_orig_name) save_orig_name = !no_name;
832 fprintf(stderr, "%s:\t", ifname);
835 /* Actually do the compression/decompression. Loop over zipped members.
838 if ((*work)(ifd, ofd) != OK) {
839 method = -1; /* force cleanup */
846 method = get_method(ifd);
847 if (method < 0) break; /* error message already emitted */
848 bytes_out = 0; /* required for length check */
851 if (close (ifd) != 0)
860 if (close (ofd) != 0)
863 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
864 remove_ofname_fd = -1;
865 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
866 sigprocmask (SIG_SETMASK, &oldset, NULL);
870 WARN ((stderr, "%s: ", program_name));
873 errno = unlink_errno;
881 remove_output_file ();
885 /* Display statistics */
888 fprintf(stderr, " OK");
889 } else if (decompress) {
890 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
892 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
894 if (!test && !to_stdout) {
895 fprintf(stderr, " -- replaced with %s", ofname);
897 fprintf(stderr, "\n");
901 /* ========================================================================
902 * Create the output file. Return OK or ERROR.
903 * Try several times if necessary to avoid truncating the z_suffix. For
904 * example, do not create a compressed file of name "1234567890123."
905 * Sets save_orig_name to true if the file name has been truncated.
906 * IN assertions: the input file has already been open (ifd is set) and
907 * ofname has already been updated if there was an original name.
908 * OUT assertions: ifd and ofd are closed in case of error.
910 local int create_outfile()
912 int name_shortened = 0;
913 int flags = (O_WRONLY | O_CREAT | O_EXCL
914 | (ascii && decompress ? 0 : O_BINARY));
921 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
922 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
924 sigprocmask (SIG_SETMASK, &oldset, NULL);
933 shorten_name (ofname);
939 if (check_ofname () != OK)
953 if (name_shortened && decompress)
955 /* name might be too long if an original name was saved */
956 WARN ((stderr, "%s: %s: warning, name truncated\n",
957 program_name, ofname));
963 /* ========================================================================
964 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
965 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
966 * accepted suffixes, in addition to the value of the --suffix option.
967 * ".tgz" is a useful convention for tar.z files on systems limited
968 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
969 * also accepted suffixes. For Unix, we do not want to accept any
970 * .??z suffix as indicating a compressed file; some people use .xyz
971 * to denote volume data.
972 * On systems allowing multiple versions of the same file (such as VMS),
973 * this function removes any version suffix in the given name.
975 local char *get_suffix(name)
979 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
980 static char *known_suffixes[] =
981 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
986 char **suf = known_suffixes;
989 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
992 /* strip a version number from the file name */
994 char *v = strrchr(name, SUFFIX_SEP);
995 if (v != NULL) *v = '\0';
999 if (nlen <= MAX_SUFFIX+2) {
1000 strcpy(suffix, name);
1002 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1005 slen = strlen(suffix);
1007 int s = strlen(*suf);
1008 if (slen > s && suffix[slen-s-1] != PATH_SEP
1009 && strequ(suffix + slen - s, *suf)) {
1012 } while (*++suf != NULL);
1018 /* Open file NAME with the given flags and mode and store its status
1019 into *ST. Return a file descriptor to the newly opened file, or -1
1020 (setting errno) on failure. */
1022 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1026 /* Refuse to follow symbolic links unless -c or -f. */
1027 if (!to_stdout && !force)
1029 if (HAVE_WORKING_O_NOFOLLOW)
1030 flags |= O_NOFOLLOW;
1033 #if HAVE_LSTAT || defined lstat
1034 if (lstat (name, st) != 0)
1036 else if (S_ISLNK (st->st_mode))
1045 fd = OPEN (name, flags, mode);
1046 if (0 <= fd && fstat (fd, st) != 0)
1057 /* ========================================================================
1058 * Set ifname to the input file name (with a suffix appended if necessary)
1059 * and istat to its stats. For decompression, if no file exists with the
1060 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1061 * For MSDOS, we try only z_suffix and z.
1062 * Return an open file descriptor or -1.
1065 open_input_file (iname, sbuf)
1069 int ilen; /* strlen(ifname) */
1070 int z_suffix_errno = 0;
1071 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1072 char **suf = suffixes;
1074 #ifdef NO_MULTIPLE_DOTS
1075 char *dot; /* pointer to ifname extension, or NULL */
1078 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1079 | (ascii && !decompress ? 0 : O_BINARY));
1083 if (sizeof ifname - 1 <= strlen (iname))
1086 strcpy(ifname, iname);
1088 /* If input file exists, return OK. */
1089 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1093 if (!decompress || errno != ENOENT) {
1097 /* file.ext doesn't exist, try adding a suffix (after removing any
1098 * version number for VMS).
1100 s = get_suffix(ifname);
1102 progerror(ifname); /* ifname already has z suffix and does not exist */
1105 #ifdef NO_MULTIPLE_DOTS
1106 dot = strrchr(ifname, '.');
1108 strcat(ifname, ".");
1109 dot = strrchr(ifname, '.');
1112 ilen = strlen(ifname);
1113 if (strequ(z_suffix, ".gz")) suf++;
1115 /* Search for all suffixes */
1117 char *s0 = s = *suf;
1118 strcpy (ifname, iname);
1119 #ifdef NO_MULTIPLE_DOTS
1121 if (*dot == '\0') strcpy (dot, ".");
1123 #ifdef MAX_EXT_CHARS
1124 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1125 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1127 if (sizeof ifname <= ilen + strlen (s))
1130 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1133 if (errno != ENOENT)
1138 if (strequ (s0, z_suffix))
1139 z_suffix_errno = errno;
1140 } while (*++suf != NULL);
1142 /* No suffix found, complain using z_suffix: */
1143 strcpy(ifname, iname);
1144 #ifdef NO_MULTIPLE_DOTS
1145 if (*dot == '\0') strcpy(dot, ".");
1147 #ifdef MAX_EXT_CHARS
1148 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1149 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1151 strcat(ifname, z_suffix);
1152 errno = z_suffix_errno;
1157 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1162 /* ========================================================================
1163 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1164 * Sets save_orig_name to true if the file name has been truncated.
1166 local int make_ofname()
1168 char *suff; /* ofname z suffix */
1170 strcpy(ofname, ifname);
1171 /* strip a version number if any and get the gzip suffix if present: */
1172 suff = get_suffix(ofname);
1176 /* With -t or -l, try all files (even without .gz suffix)
1177 * except with -r (behave as with just -dr).
1179 if (!recursive && (list || test)) return OK;
1181 /* Avoid annoying messages with -r */
1182 if (verbose || (!recursive && !quiet)) {
1183 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1184 program_name, ifname));
1188 /* Make a special case for .tgz and .taz: */
1190 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1191 strcpy(suff, ".tar");
1193 *suff = '\0'; /* strip the z suffix */
1195 /* ofname might be changed later if infile contains an original name */
1197 } else if (suff != NULL) {
1198 /* Avoid annoying messages with -r (see treat_dir()) */
1199 if (verbose || (!recursive && !quiet)) {
1200 /* Don't use WARN, as it affects exit status. */
1201 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1202 program_name, ifname, suff);
1208 #ifdef NO_MULTIPLE_DOTS
1209 suff = strrchr(ofname, '.');
1211 if (sizeof ofname <= strlen (ofname) + 1)
1213 strcat(ofname, ".");
1214 # ifdef MAX_EXT_CHARS
1215 if (strequ(z_suffix, "z")) {
1216 if (sizeof ofname <= strlen (ofname) + 2)
1218 strcat(ofname, "gz"); /* enough room */
1221 /* On the Atari and some versions of MSDOS,
1222 * ENAMETOOLONG does not work correctly. So we
1223 * must truncate here.
1225 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1226 suff[MAX_SUFFIX+1-z_len] = '\0';
1230 #endif /* NO_MULTIPLE_DOTS */
1231 if (sizeof ofname <= strlen (ofname) + z_len)
1233 strcat(ofname, z_suffix);
1235 } /* decompress ? */
1239 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1244 /* ========================================================================
1245 * Check the magic number of the input file and update ofname if an
1246 * original name was given and to_stdout is not set.
1247 * Return the compression method, -1 for error, -2 for warning.
1248 * Set inptr to the offset of the next byte to be processed.
1249 * Updates time_stamp if there is one and --no-time is not used.
1250 * This function may be called repeatedly for an input file consisting
1251 * of several contiguous gzip'ed members.
1252 * IN assertions: there is at least one remaining compressed member.
1253 * If the member is a zip file, it must be the only one.
1255 local int get_method(in)
1256 int in; /* input file descriptor */
1258 uch flags; /* compression flags */
1259 char magic[2]; /* magic header */
1260 int imagic1; /* like magic[1], but can represent EOF */
1261 ulg stamp; /* time stamp */
1263 /* If --force and --stdout, zcat == cat, so do not complain about
1264 * premature end of file: use try_byte instead of get_byte.
1266 if (force && to_stdout) {
1267 magic[0] = (char)try_byte();
1268 imagic1 = try_byte ();
1269 magic[1] = (char) imagic1;
1270 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1272 magic[0] = (char)get_byte();
1273 magic[1] = (char)get_byte();
1274 imagic1 = 0; /* avoid lint warning */
1276 method = -1; /* unknown yet */
1277 part_nb++; /* number of parts in gzip file */
1279 last_member = RECORD_IO;
1280 /* assume multiple members in gzip file except for record oriented I/O */
1282 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1283 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1285 method = (int)get_byte();
1286 if (method != DEFLATED) {
1288 "%s: %s: unknown method %d -- not supported\n",
1289 program_name, ifname, method);
1294 flags = (uch)get_byte();
1296 if ((flags & ENCRYPTED) != 0) {
1298 "%s: %s is encrypted -- not supported\n",
1299 program_name, ifname);
1303 if ((flags & CONTINUATION) != 0) {
1305 "%s: %s is a multi-part gzip file -- not supported\n",
1306 program_name, ifname);
1308 if (force <= 1) return -1;
1310 if ((flags & RESERVED) != 0) {
1312 "%s: %s has flags 0x%x -- not supported\n",
1313 program_name, ifname, flags);
1315 if (force <= 1) return -1;
1317 stamp = (ulg)get_byte();
1318 stamp |= ((ulg)get_byte()) << 8;
1319 stamp |= ((ulg)get_byte()) << 16;
1320 stamp |= ((ulg)get_byte()) << 24;
1321 if (stamp != 0 && !no_time)
1323 time_stamp.tv_sec = stamp;
1324 time_stamp.tv_nsec = 0;
1327 (void)get_byte(); /* Ignore extra flags for the moment */
1328 (void)get_byte(); /* Ignore OS type for the moment */
1330 if ((flags & CONTINUATION) != 0) {
1331 unsigned part = (unsigned)get_byte();
1332 part |= ((unsigned)get_byte())<<8;
1334 fprintf(stderr,"%s: %s: part number %u\n",
1335 program_name, ifname, part);
1338 if ((flags & EXTRA_FIELD) != 0) {
1339 unsigned len = (unsigned)get_byte();
1340 len |= ((unsigned)get_byte())<<8;
1342 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1343 program_name, ifname, len);
1345 while (len--) (void)get_byte();
1348 /* Get original file name if it was truncated */
1349 if ((flags & ORIG_NAME) != 0) {
1350 if (no_name || (to_stdout && !list) || part_nb > 1) {
1351 /* Discard the old name */
1352 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1353 do {c=get_byte();} while (c != 0);
1355 /* Copy the base name. Keep a directory prefix intact. */
1356 char *p = gzip_base_name (ofname);
1359 *p = (char)get_char();
1360 if (*p++ == '\0') break;
1361 if (p >= ofname+sizeof(ofname)) {
1362 gzip_error ("corrupted input -- file name too large");
1365 p = gzip_base_name (base);
1366 memmove (base, p, strlen (p) + 1);
1367 /* If necessary, adapt the name to local OS conventions: */
1369 MAKE_LEGAL_NAME(base);
1370 if (base) list=0; /* avoid warning about unused variable */
1372 } /* no_name || to_stdout */
1375 /* Discard file comment if any */
1376 if ((flags & COMMENT) != 0) {
1377 while (get_char() != 0) /* null */ ;
1380 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1383 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1384 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1385 /* To simplify the code, we support a zip file when alone only.
1386 * We are thus guaranteed that the entire local header fits in inbuf.
1390 if (check_zipfile(in) != OK) return -1;
1391 /* check_zipfile may get ofname from the local header */
1394 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1398 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1400 method = COMPRESSED;
1403 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1408 } else if (force && to_stdout && !list) { /* pass input unchanged */
1414 if (method >= 0) return method;
1417 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1418 program_name, ifname);
1425 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1430 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1431 program_name, ifname));
1436 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1437 program_name, ifname));
1442 /* ========================================================================
1443 * Display the characteristics of the compressed file.
1444 * If the given method is < 0, display the accumulated totals.
1445 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1447 local void do_list(ifd, method)
1448 int ifd; /* input file descriptor */
1449 int method; /* compression method */
1451 ulg crc; /* original crc */
1452 static int first_time = 1;
1453 static char* methods[MAX_METHODS] = {
1458 "", "", "", "", /* 4 to 7 reserved */
1460 int positive_off_t_width = 1;
1463 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1464 positive_off_t_width++;
1467 if (first_time && method >= 0) {
1470 printf("method crc date time ");
1473 printf("%*.*s %*.*s ratio uncompressed_name\n",
1474 positive_off_t_width, positive_off_t_width, "compressed",
1475 positive_off_t_width, positive_off_t_width, "uncompressed");
1477 } else if (method < 0) {
1478 if (total_in <= 0 || total_out <= 0) return;
1482 if (verbose || !quiet) {
1483 fprint_off(stdout, total_in, positive_off_t_width);
1485 fprint_off(stdout, total_out, positive_off_t_width);
1488 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1489 /* header_bytes is not meaningful but used to ensure the same
1490 * ratio if there is a single file.
1492 printf(" (totals)\n");
1495 crc = (ulg)~0; /* unknown */
1497 bytes_in = ifile_size;
1500 if (method == DEFLATED && !last_member) {
1501 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1502 * If the lseek fails, we could use read() to get to the end, but
1503 * --list is used to get quick results.
1504 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1505 * you are not concerned about speed.
1507 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1508 if (bytes_in != -1L) {
1511 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1515 bytes_out = LG(buf+4);
1518 #endif /* RECORD_IO */
1521 struct tm *tm = localtime (&time_stamp.tv_sec);
1522 printf ("%5s %08lx ", methods[method], crc);
1524 printf ("%s%3d %02d:%02d ",
1525 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1527 tm->tm_mday, tm->tm_hour, tm->tm_min);
1529 printf ("??? ?? ??:?? ");
1531 fprint_off(stdout, bytes_in, positive_off_t_width);
1533 fprint_off(stdout, bytes_out, positive_off_t_width);
1535 if (bytes_in == -1L) {
1537 bytes_in = bytes_out = header_bytes = 0;
1538 } else if (total_in >= 0) {
1539 total_in += bytes_in;
1541 if (bytes_out == -1L) {
1543 bytes_in = bytes_out = header_bytes = 0;
1544 } else if (total_out >= 0) {
1545 total_out += bytes_out;
1547 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1548 printf(" %s\n", ofname);
1551 /* ========================================================================
1552 * Shorten the given name by one character, or replace a .tar extension
1553 * with .tgz. Truncate the last part of the name which is longer than
1554 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1555 * has only parts shorter than MIN_PART truncate the longest part.
1556 * For decompression, just remove the last character of the name.
1558 * IN assertion: for compression, the suffix of the given name is z_suffix.
1560 local void shorten_name(name)
1563 int len; /* length of name without z_suffix */
1564 char *trunc = NULL; /* character to be truncated */
1565 int plen; /* current part length */
1566 int min_part = MIN_PART; /* current minimum part length */
1572 gzip_error ("name too short");
1576 p = get_suffix(name);
1578 gzip_error ("can't recover suffix\n");
1582 /* compress 1234567890.tar to 1234567890.tgz */
1583 if (len > 4 && strequ(p-4, ".tar")) {
1584 strcpy(p-4, ".tgz");
1587 /* Try keeping short extensions intact:
1588 * 1234.678.012.gz -> 123.678.012.gz
1591 p = strrchr(name, PATH_SEP);
1594 plen = strcspn(p, PART_SEP);
1596 if (plen > min_part) trunc = p-1;
1599 } while (trunc == NULL && --min_part != 0);
1601 if (trunc != NULL) {
1603 trunc[0] = trunc[1];
1607 trunc = strrchr(name, PART_SEP[0]);
1609 gzip_error ("internal error in shorten_name");
1610 if (trunc[1] == '\0') trunc--; /* force truncation */
1612 strcpy(trunc, z_suffix);
1615 /* ========================================================================
1616 * The compressed file already exists, so ask for confirmation.
1617 * Return ERROR if the file must be skipped.
1619 local int check_ofname()
1621 /* Ask permission to overwrite the existing file */
1624 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1625 if (foreground && isatty(fileno(stdin))) {
1626 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1631 fprintf(stderr, "\tnot overwritten\n");
1632 if (exit_code == OK) exit_code = WARNING;
1636 if (xunlink (ofname)) {
1644 /* ========================================================================
1645 * Copy modes, times, ownership from input file to output file.
1646 * IN assertion: to_stdout is false.
1648 local void copy_stat(ifstat)
1649 struct stat *ifstat;
1651 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1655 struct timespec timespec[2];
1656 timespec[0] = get_stat_atime (ifstat);
1657 timespec[1] = get_stat_mtime (ifstat);
1659 if (decompress && 0 <= time_stamp.tv_nsec
1660 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1661 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1663 timespec[1] = time_stamp;
1665 fprintf(stderr, "%s: time stamp restored\n", ofname);
1669 if (gz_futimens (ofd, ofname, timespec) != 0)
1672 WARN ((stderr, "%s: ", program_name));
1683 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1685 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1689 /* Copy the protection modes */
1691 r = fchmod (ofd, mode);
1693 r = chmod (ofname, mode);
1697 WARN ((stderr, "%s: ", program_name));
1707 /* ========================================================================
1708 * Recurse through the given directory. This code is taken from ncompress.
1710 local void treat_dir (fd, dir)
1716 char nbuf[MAX_PATH_LEN];
1720 dirp = fdopendir (fd);
1723 dirp = opendir(dir);
1734 ** WARNING: the following algorithm could occasionally cause
1735 ** compress to produce error warnings of the form "<filename>.gz
1736 ** already has .gz suffix - ignored". This occurs when the
1737 ** .gz output file is inserted into the directory below
1738 ** readdir's current pointer.
1739 ** These warnings are harmless but annoying, so they are suppressed
1740 ** with option -r (except when -v is on). An alternative
1741 ** to allowing this would be to store the entire directory
1742 ** list in memory, then compress the entries in the stored
1743 ** list. Given the depth-first recursive algorithm used here,
1744 ** this could use up a tremendous amount of memory. I don't
1745 ** think it's worth it. -- Dave Mack
1746 ** (An other alternative might be two passes to avoid depth-first.)
1749 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1751 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1755 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1757 if (len != 0 /* dir = "" means current dir on Amiga */
1759 && dir[len-1] != PATH_SEP2
1762 && dir[len-1] != PATH_SEP3
1765 nbuf[len++] = PATH_SEP;
1767 strcpy(nbuf+len, dp->d_name);
1770 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1771 program_name, dir, dp->d_name);
1777 if (CLOSEDIR(dirp) != 0)
1780 #endif /* ! NO_DIR */
1782 /* Make sure signals get handled properly. */
1785 install_signal_handlers ()
1787 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1791 struct sigaction act;
1793 sigemptyset (&caught_signals);
1794 for (i = 0; i < nsigs; i++)
1796 sigaction (handled_sig[i], NULL, &act);
1797 if (act.sa_handler != SIG_IGN)
1798 sigaddset (&caught_signals, handled_sig[i]);
1801 act.sa_handler = abort_gzip_signal;
1802 act.sa_mask = caught_signals;
1805 for (i = 0; i < nsigs; i++)
1806 if (sigismember (&caught_signals, handled_sig[i]))
1810 sigaction (handled_sig[i], &act, NULL);
1813 for (i = 0; i < nsigs; i++)
1814 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1818 signal (handled_sig[i], abort_gzip_signal);
1819 siginterrupt (handled_sig[i], 1);
1824 /* ========================================================================
1825 * Free all dynamically allocated variables and exit with the given code.
1827 local void do_exit(exitcode)
1830 static int in_exit = 0;
1832 if (in_exit) exit(exitcode);
1834 if (env != NULL) free(env), env = NULL;
1835 if (args != NULL) free((char*)args), args = NULL;
1849 /* ========================================================================
1850 * Close and unlink the output file.
1853 remove_output_file ()
1858 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1859 fd = remove_ofname_fd;
1862 remove_ofname_fd = -1;
1866 sigprocmask (SIG_SETMASK, &oldset, NULL);
1869 /* ========================================================================
1875 remove_output_file ();
1879 /* ========================================================================
1883 abort_gzip_signal (sig)
1887 signal (sig, SIG_IGN);
1888 remove_output_file ();
1889 if (sig == exiting_signal)
1891 signal (sig, SIG_DFL);