1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001, 2002, 2006 Free Software Foundation, Inc.
4 Copyright (C) 1992-1993 Jean-loup Gailly
6 This program is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 This program is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with this program; if not, write to the Free Software Foundation,
18 Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. */
21 * The unzip code was written and put in the public domain by Mark Adler.
22 * Portions of the lzw code are derived from the public domain 'compress'
23 * written by Spencer Thomas, Joe Orost, James Woods, Jim McKie, Steve Davies,
24 * Ken Turkowski, Dave Mack and Peter Jannesen.
26 * See the license_msg below and the file COPYING for the software license.
27 * See the file algorithm.doc for the compression algorithms and file formats.
30 static char *license_msg[] = {
31 "Copyright (C) 2006 Free Software Foundation, Inc.",
32 "Copyright (C) 1993 Jean-loup Gailly.",
33 "This is free software. You may redistribute copies of it under the terms of",
34 "the GNU General Public License <http://www.gnu.org/licenses/gpl.html>.",
35 "There is NO WARRANTY, to the extent permitted by law.",
38 /* Compress files with zip algorithm and 'compress' interface.
39 * See help() function below for all options.
41 * file.gz: compressed file with same mode, owner, and utimes
42 * or stdout with -c option or if stdin used as input.
43 * If the output file name had to be truncated, the original name is kept
44 * in the compressed file.
45 * On MSDOS, file.tmp -> file.tmz. On VMS, file.tmp -> file.tmp-gz.
47 * Using gz on MSDOS would create too many file name conflicts. For
48 * example, foo.txt -> foo.tgz (.tgz must be reserved as shorthand for
49 * tar.gz). Similarly, foo.dir and foo.doc would both be mapped to foo.dgz.
50 * I also considered 12345678.txt -> 12345txt.gz but this truncates the name
51 * too heavily. There is no ideal solution given the MSDOS 8+3 limitation.
53 * For the meaning of all compilation flags, see comments in Makefile.in.
57 static char rcsid[] = "$Id: gzip.c,v 1.11 2006/12/12 00:03:17 eggert Exp $";
62 #include <sys/types.h>
72 #include "fcntl-safer.h"
74 #include "stat-time.h"
82 # include <sys/time.h>
97 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
108 # ifndef _D_EXACT_NAMLEN
109 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
114 # define CLOSEDIR(d) (closedir(d), 0)
116 # define CLOSEDIR(d) closedir(d)
120 # include <utimens.h>
123 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
126 # define MAX_PATH_LEN 1024 /* max pathname length */
138 off_t lseek OF((int fd, off_t offset, int whence));
142 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
146 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
149 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
152 # define SA_NOCLDSTOP 0
153 # define sigprocmask(how, set, oset) /* empty */
154 # define sigset_t int
155 # if ! HAVE_SIGINTERRUPT
156 # define siginterrupt(sig, flag) /* empty */
160 #ifndef HAVE_WORKING_O_NOFOLLOW
161 # define HAVE_WORKING_O_NOFOLLOW 0
165 # define ELOOP EINVAL
168 /* Separator for file name parts (see shorten_name()) */
169 #ifdef NO_MULTIPLE_DOTS
170 # define PART_SEP "-"
172 # define PART_SEP "."
177 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
178 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
179 DECLARE(ush, d_buf, DIST_BUFSIZE);
180 DECLARE(uch, window, 2L*WSIZE);
182 DECLARE(ush, tab_prefix, 1L<<BITS);
184 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
185 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
188 /* local variables */
190 int ascii = 0; /* convert end-of-lines to local OS conventions */
191 int to_stdout = 0; /* output to stdout (-c) */
192 int decompress = 0; /* decompress (-d) */
193 int force = 0; /* don't ask questions, compress links (-f) */
194 int no_name = -1; /* don't save or restore the original file name */
195 int no_time = -1; /* don't save or restore the original file time */
196 int recursive = 0; /* recurse through directories (-r) */
197 int list = 0; /* list the file contents (-l) */
198 int verbose = 0; /* be verbose (-v) */
199 int quiet = 0; /* be very quiet (-q) */
200 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
201 int test = 0; /* test .gz file integrity */
202 int foreground; /* set if program run in foreground */
203 char *program_name; /* program name */
204 int maxbits = BITS; /* max bits per code for LZW */
205 int method = DEFLATED;/* compression method */
206 int level = 6; /* compression level */
207 int exit_code = OK; /* program exit code */
208 int save_orig_name; /* set if original name must be saved */
209 int last_member; /* set for .zip and .Z files */
210 int part_nb; /* number of parts in .gz file */
211 struct timespec time_stamp; /* original time stamp (modification time) */
212 off_t ifile_size; /* input file size, -1 for devices (debug only) */
213 char *env; /* contents of GZIP env variable */
214 char **args = NULL; /* argv pointer if GZIP env variable defined */
215 char *z_suffix; /* default suffix (can be set with --suffix) */
216 size_t z_len; /* strlen(z_suffix) */
218 /* The set of signals that are caught. */
219 static sigset_t caught_signals;
221 /* If nonzero then exit with status 1, rather than with the usual
222 signal status, on receipt of a signal with this value. This
223 suppresses a "Broken Pipe" message with some shells. */
224 static int volatile exiting_signal;
226 /* If nonnegative, close this file descriptor and unlink ofname on error. */
227 static int volatile remove_ofname_fd = -1;
229 off_t bytes_in; /* number of input bytes */
230 off_t bytes_out; /* number of output bytes */
231 off_t total_in; /* input bytes for all files */
232 off_t total_out; /* output bytes for all files */
233 char ifname[MAX_PATH_LEN]; /* input file name */
234 char ofname[MAX_PATH_LEN]; /* output file name */
235 struct stat istat; /* status for input file */
236 int ifd; /* input file descriptor */
237 int ofd; /* output file descriptor */
238 unsigned insize; /* valid bytes in inbuf */
239 unsigned inptr; /* index of next byte to be processed in inbuf */
240 unsigned outcnt; /* bytes in output buffer */
241 int rsync = 0; /* make ryncable chunks */
243 struct option longopts[] =
245 /* { name has_arg *flag val } */
246 {"ascii", 0, 0, 'a'}, /* ascii text mode */
247 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
248 {"stdout", 0, 0, 'c'}, /* write output on standard output */
249 {"decompress", 0, 0, 'd'}, /* decompress */
250 {"uncompress", 0, 0, 'd'}, /* decompress */
251 /* {"encrypt", 0, 0, 'e'}, encrypt */
252 {"force", 0, 0, 'f'}, /* force overwrite of output file */
253 {"help", 0, 0, 'h'}, /* give help */
254 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
255 {"list", 0, 0, 'l'}, /* list .gz file contents */
256 {"license", 0, 0, 'L'}, /* display software license */
257 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
258 {"name", 0, 0, 'N'}, /* save or restore original name & time */
259 {"quiet", 0, 0, 'q'}, /* quiet mode */
260 {"silent", 0, 0, 'q'}, /* quiet mode */
261 {"recursive", 0, 0, 'r'}, /* recurse through directories */
262 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
263 {"test", 0, 0, 't'}, /* test compressed file integrity */
264 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
265 {"verbose", 0, 0, 'v'}, /* verbose mode */
266 {"version", 0, 0, 'V'}, /* display version number */
267 {"fast", 0, 0, '1'}, /* compress faster */
268 {"best", 0, 0, '9'}, /* compress better */
269 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
270 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
271 {"rsyncable", 0, 0, 'R'}, /* make rsync-friendly archive */
275 /* local functions */
277 local void try_help OF((void)) ATTRIBUTE_NORETURN;
278 local void help OF((void));
279 local void license OF((void));
280 local void version OF((void));
281 local int input_eof OF((void));
282 local void treat_stdin OF((void));
283 local void treat_file OF((char *iname));
284 local int create_outfile OF((void));
285 local char *get_suffix OF((char *name));
286 local int open_input_file OF((char *iname, struct stat *sbuf));
287 local int make_ofname OF((void));
288 local void shorten_name OF((char *name));
289 local int get_method OF((int in));
290 local void do_list OF((int ifd, int method));
291 local int check_ofname OF((void));
292 local void copy_stat OF((struct stat *ifstat));
293 local void install_signal_handlers OF((void));
294 local void remove_output_file OF((void));
295 local RETSIGTYPE abort_gzip_signal OF((int));
296 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
297 int main OF((int argc, char **argv));
298 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
301 local void treat_dir OF((int fd, char *dir));
304 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
309 fprintf (stderr, "Try `%s --help' for more information.\n",
314 /* ======================================================================== */
317 static char *help_msg[] = {
318 "Compress or uncompress FILEs (by default, compress FILES in-place).",
320 "Mandatory arguments to long options are mandatory for short options too.",
323 " -a, --ascii ascii text; convert end-of-line using local conventions",
325 " -c, --stdout write on standard output, keep original files unchanged",
326 " -d, --decompress decompress",
327 /* -e, --encrypt encrypt */
328 " -f, --force force overwrite of output file and compress links",
329 " -h, --help give this help",
330 /* -k, --pkzip force output in pkzip format */
331 " -l, --list list compressed file contents",
332 " -L, --license display software license",
334 " -m, --no-time do not save or restore the original modification time",
335 " -M, --time save or restore the original modification time",
337 " -n, --no-name do not save or restore the original name and time stamp",
338 " -N, --name save or restore the original name and time stamp",
339 " -q, --quiet suppress all warnings",
341 " -r, --recursive operate recursively on directories",
343 " -S, --suffix=SUF use suffix SUF on compressed files",
344 " -t, --test test compressed file integrity",
345 " -v, --verbose verbose mode",
346 " -V, --version display version number",
347 " -1, --fast compress faster",
348 " -9, --best compress better",
350 " -Z, --lzw produce output compatible with old compress",
351 " -b, --bits=BITS max number of bits per code (implies -Z)",
353 " --rsyncable Make rsync-friendly archive",
355 "With no FILE, or when FILE is -, read standard input.",
357 "Report bugs to <bug-gzip@gnu.org>.",
361 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
362 while (*p) printf ("%s\n", *p++);
365 /* ======================================================================== */
368 char **p = license_msg;
370 printf ("%s %s\n", program_name, VERSION);
371 while (*p) printf ("%s\n", *p++);
374 /* ======================================================================== */
379 printf ("Written by Jean-loup Gailly.\n");
382 local void progerror (string)
386 fprintf (stderr, "%s: ", program_name);
392 /* ======================================================================== */
393 int main (argc, argv)
397 int file_count; /* number of files to process */
398 size_t proglen; /* length of program_name */
399 int optc; /* current option */
401 EXPAND(argc, argv); /* wild card expansion if necessary */
403 program_name = gzip_base_name (argv[0]);
404 proglen = strlen (program_name);
406 /* Suppress .exe for MSDOS, OS/2 and VMS: */
407 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
408 program_name[proglen - 4] = '\0';
410 /* Add options in GZIP environment variable if there is one */
411 env = add_envopt(&argc, &argv, OPTIONS_VAR);
412 if (env != NULL) args = argv;
415 /* For compatibility with old compress, use program name as an option.
416 * If you compile with -DGNU_STANDARD, this program will behave as
417 * gzip even if it is invoked under the name gunzip or zcat.
419 * Systems which do not support links can still use -d or -dc.
420 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
422 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
423 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
425 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
426 || strequ (program_name, "gzcat")) /* gzcat */
427 decompress = to_stdout = 1;
431 z_len = strlen(z_suffix);
433 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
434 longopts, (int *)0)) != -1) {
439 maxbits = atoi(optarg);
440 for (; *optarg; optarg++)
441 if (! ('0' <= *optarg && *optarg <= '9'))
443 fprintf (stderr, "%s: -b operand is not an integer\n",
449 to_stdout = 1; break;
451 decompress = 1; break;
455 help(); do_exit(OK); break;
457 list = decompress = to_stdout = 1; break;
459 license(); do_exit(OK); break;
460 case 'm': /* undocumented, may change later */
462 case 'M': /* undocumented, may change later */
465 no_name = no_time = 1; break;
467 no_name = no_time = 0; break;
469 quiet = 1; verbose = 0; break;
472 fprintf (stderr, "%s: -r not supported on this system\n",
482 #ifdef NO_MULTIPLE_DOTS
483 if (*optarg == '.') optarg++;
485 z_len = strlen(optarg);
489 test = decompress = to_stdout = 1;
492 verbose++; quiet = 0; break;
494 version(); do_exit(OK); break;
499 fprintf(stderr, "%s: -Z not supported in this version\n",
504 case '1': case '2': case '3': case '4':
505 case '5': case '6': case '7': case '8': case '9':
509 /* Error message already emitted by getopt_long. */
512 } /* loop on all arguments */
514 /* By default, save name and timestamp on compression but do not
515 * restore them on decompression.
517 if (no_time < 0) no_time = decompress;
518 if (no_name < 0) no_name = decompress;
520 file_count = argc - optind;
524 if (ascii && !quiet) {
525 fprintf(stderr, "%s: option --ascii ignored on this system\n",
529 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
530 fprintf(stderr, "%s: incorrect suffix '%s'\n",
531 program_name, z_suffix);
534 if (do_lzw && !decompress) work = lzw;
536 /* Allocate all global buffers (for DYN_ALLOC option) */
537 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
538 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
539 ALLOC(ush, d_buf, DIST_BUFSIZE);
540 ALLOC(uch, window, 2L*WSIZE);
542 ALLOC(ush, tab_prefix, 1L<<BITS);
544 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
545 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
548 exiting_signal = quiet ? SIGPIPE : 0;
549 install_signal_handlers ();
551 /* And get to work */
552 if (file_count != 0) {
553 if (to_stdout && !test && !list && (!decompress || !ascii)) {
554 SET_BINARY_MODE(fileno(stdout));
556 while (optind < argc) {
557 treat_file(argv[optind++]);
559 } else { /* Standard input */
562 if (list && !quiet && file_count > 1) {
563 do_list(-1, -1); /* print totals */
566 return exit_code; /* just to avoid lint warning */
569 /* Return nonzero when at end of file on input. */
573 if (!decompress || last_member)
578 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
581 /* Unget the char that fill_inbuf got. */
588 /* ========================================================================
589 * Compress or decompress stdin
591 local void treat_stdin()
593 if (!force && !list &&
594 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
595 /* Do not send compressed data to the terminal or read it from
596 * the terminal. We get here when user invoked the program
597 * without parameters, so be helpful. According to the GNU standards:
599 * If there is one behavior you think is most useful when the output
600 * is to a terminal, and another that you think is most useful when
601 * the output is a file or a pipe, then it is usually best to make
602 * the default behavior the one that is useful with output to a
603 * terminal, and have an option for the other behavior.
605 * Here we use the --force option to get the other behavior.
608 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
609 program_name, decompress ? "read from" : "written to",
610 decompress ? "de" : "");
611 fprintf (stderr, "For help, type: %s -h\n", program_name);
615 if (decompress || !ascii) {
616 SET_BINARY_MODE(fileno(stdin));
618 if (!test && !list && (!decompress || !ascii)) {
619 SET_BINARY_MODE(fileno(stdout));
621 strcpy(ifname, "stdin");
622 strcpy(ofname, "stdout");
624 /* Get the time stamp on the input file. */
625 time_stamp.tv_nsec = -1; /* The time is unknown by default. */
627 #ifndef NO_STDIN_FSTAT
628 if (list || !no_time) {
629 if (fstat(fileno(stdin), &istat) != 0) {
630 progerror("standard input");
633 # ifdef NO_PIPE_TIMESTAMP
634 if (S_ISREG(istat.st_mode))
636 time_stamp = get_stat_mtime (&istat);
637 #endif /* NO_STDIN_FSTAT */
639 ifile_size = -1L; /* convention for unknown size */
641 clear_bufs(); /* clear input and output buffers */
646 method = get_method(ifd);
648 do_exit(exit_code); /* error message already emitted */
652 do_list(ifd, method);
656 /* Actually do the compression/decompression. Loop over zipped members.
659 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
664 method = get_method(ifd);
665 if (method < 0) return; /* error message already emitted */
666 bytes_out = 0; /* required for length check */
671 fprintf(stderr, " OK\n");
673 } else if (!decompress) {
674 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
675 fprintf(stderr, "\n");
676 #ifdef DISPLAY_STDIN_RATIO
678 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
679 fprintf(stderr, "\n");
685 /* ========================================================================
686 * Compress or decompress the given file
688 local void treat_file(iname)
691 /* Accept "-" as synonym for stdin */
692 if (strequ(iname, "-")) {
693 int cflag = to_stdout;
699 /* Check if the input file is present, set ifname and istat: */
700 ifd = open_input_file (iname, &istat);
704 /* If the input name is that of a directory, recurse or ignore: */
705 if (S_ISDIR(istat.st_mode)) {
708 treat_dir (ifd, iname);
709 /* Warning: ifname is now garbage */
714 WARN ((stderr, "%s: %s is a directory -- ignored\n",
715 program_name, ifname));
718 if (!S_ISREG(istat.st_mode)) {
720 "%s: %s is not a directory or a regular file - ignored\n",
721 program_name, ifname));
726 if (istat.st_mode & S_ISUID)
728 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
729 program_name, ifname));
733 if (istat.st_mode & S_ISGID)
735 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
736 program_name, ifname));
740 if (istat.st_mode & S_ISVTX)
742 WARN ((stderr, "%s: %s has the sticky bit set - file ignored\n",
743 program_name, ifname));
748 if (istat.st_nlink > 1 && !to_stdout && !force) {
749 WARN((stderr, "%s: %s has %lu other link%c -- unchanged\n",
750 program_name, ifname, (unsigned long) istat.st_nlink - 1,
751 istat.st_nlink > 2 ? 's' : ' '));
756 ifile_size = istat.st_size;
757 if (no_time && !list)
758 time_stamp.tv_nsec = -1;
760 time_stamp = get_stat_mtime (&istat);
762 /* Generate output file name. For -r and (-t or -l), skip files
763 * without a valid gzip suffix (check done in make_ofname).
765 if (to_stdout && !list && !test) {
766 strcpy(ofname, "stdout");
768 } else if (make_ofname() != OK) {
773 clear_bufs(); /* clear input and output buffers */
777 method = get_method(ifd); /* updates ofname if original given */
780 return; /* error message already emitted */
784 do_list(ifd, method);
785 if (close (ifd) != 0)
790 /* If compressing to a file, check if ofname is not ambiguous
791 * because the operating system truncates names. Otherwise, generate
792 * a new ofname and save the original name in the compressed file.
795 ofd = fileno(stdout);
796 /* Keep remove_ofname_fd negative. */
798 if (create_outfile() != OK) return;
800 if (!decompress && save_orig_name && !verbose && !quiet) {
801 fprintf(stderr, "%s: %s compressed to %s\n",
802 program_name, ifname, ofname);
805 /* Keep the name even if not truncated except with --no-name: */
806 if (!save_orig_name) save_orig_name = !no_name;
809 fprintf(stderr, "%s:\t", ifname);
812 /* Actually do the compression/decompression. Loop over zipped members.
815 if ((*work)(ifd, ofd) != OK) {
816 method = -1; /* force cleanup */
823 method = get_method(ifd);
824 if (method < 0) break; /* error message already emitted */
825 bytes_out = 0; /* required for length check */
828 if (close (ifd) != 0)
837 if (close (ofd) != 0)
840 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
841 remove_ofname_fd = -1;
842 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
843 sigprocmask (SIG_SETMASK, &oldset, NULL);
847 WARN ((stderr, "%s: ", program_name));
850 errno = unlink_errno;
858 remove_output_file ();
862 /* Display statistics */
865 fprintf(stderr, " OK");
866 } else if (decompress) {
867 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
869 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
871 if (!test && !to_stdout) {
872 fprintf(stderr, " -- replaced with %s", ofname);
874 fprintf(stderr, "\n");
878 /* ========================================================================
879 * Create the output file. Return OK or ERROR.
880 * Try several times if necessary to avoid truncating the z_suffix. For
881 * example, do not create a compressed file of name "1234567890123."
882 * Sets save_orig_name to true if the file name has been truncated.
883 * IN assertions: the input file has already been open (ifd is set) and
884 * ofname has already been updated if there was an original name.
885 * OUT assertions: ifd and ofd are closed in case of error.
887 local int create_outfile()
889 int name_shortened = 0;
890 int flags = (O_WRONLY | O_CREAT | O_EXCL
891 | (ascii && decompress ? 0 : O_BINARY));
898 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
899 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
901 sigprocmask (SIG_SETMASK, &oldset, NULL);
910 shorten_name (ofname);
916 if (check_ofname () != OK)
930 if (name_shortened && decompress)
932 /* name might be too long if an original name was saved */
933 WARN ((stderr, "%s: %s: warning, name truncated\n",
934 program_name, ofname));
940 /* ========================================================================
941 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
942 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
943 * accepted suffixes, in addition to the value of the --suffix option.
944 * ".tgz" is a useful convention for tar.z files on systems limited
945 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
946 * also accepted suffixes. For Unix, we do not want to accept any
947 * .??z suffix as indicating a compressed file; some people use .xyz
948 * to denote volume data.
949 * On systems allowing multiple versions of the same file (such as VMS),
950 * this function removes any version suffix in the given name.
952 local char *get_suffix(name)
956 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
957 static char *known_suffixes[] =
958 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
963 char **suf = known_suffixes;
966 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
969 /* strip a version number from the file name */
971 char *v = strrchr(name, SUFFIX_SEP);
972 if (v != NULL) *v = '\0';
976 if (nlen <= MAX_SUFFIX+2) {
977 strcpy(suffix, name);
979 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
982 slen = strlen(suffix);
984 int s = strlen(*suf);
985 if (slen > s && suffix[slen-s-1] != PATH_SEP
986 && strequ(suffix + slen - s, *suf)) {
989 } while (*++suf != NULL);
995 /* Open file NAME with the given flags and mode and store its status
996 into *ST. Return a file descriptor to the newly opened file, or -1
997 (setting errno) on failure. */
999 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1003 /* Refuse to follow symbolic links unless -c or -f. */
1004 if (!to_stdout && !force)
1006 if (HAVE_WORKING_O_NOFOLLOW)
1007 flags |= O_NOFOLLOW;
1010 #if HAVE_LSTAT || defined lstat
1011 if (lstat (name, st) != 0)
1013 else if (S_ISLNK (st->st_mode))
1022 fd = OPEN (name, flags, mode);
1023 if (0 <= fd && fstat (fd, st) != 0)
1034 /* ========================================================================
1035 * Set ifname to the input file name (with a suffix appended if necessary)
1036 * and istat to its stats. For decompression, if no file exists with the
1037 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1038 * For MSDOS, we try only z_suffix and z.
1039 * Return an open file descriptor or -1.
1042 open_input_file (iname, sbuf)
1046 int ilen; /* strlen(ifname) */
1047 int z_suffix_errno = 0;
1048 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1049 char **suf = suffixes;
1051 #ifdef NO_MULTIPLE_DOTS
1052 char *dot; /* pointer to ifname extension, or NULL */
1055 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1056 | (ascii && !decompress ? 0 : O_BINARY));
1060 if (sizeof ifname - 1 <= strlen (iname))
1063 strcpy(ifname, iname);
1065 /* If input file exists, return OK. */
1066 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1070 if (!decompress || errno != ENOENT) {
1074 /* file.ext doesn't exist, try adding a suffix (after removing any
1075 * version number for VMS).
1077 s = get_suffix(ifname);
1079 progerror(ifname); /* ifname already has z suffix and does not exist */
1082 #ifdef NO_MULTIPLE_DOTS
1083 dot = strrchr(ifname, '.');
1085 strcat(ifname, ".");
1086 dot = strrchr(ifname, '.');
1089 ilen = strlen(ifname);
1090 if (strequ(z_suffix, ".gz")) suf++;
1092 /* Search for all suffixes */
1094 char *s0 = s = *suf;
1095 strcpy (ifname, iname);
1096 #ifdef NO_MULTIPLE_DOTS
1098 if (*dot == '\0') strcpy (dot, ".");
1100 #ifdef MAX_EXT_CHARS
1101 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1102 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1104 if (sizeof ifname <= ilen + strlen (s))
1107 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1110 if (errno != ENOENT)
1115 if (strequ (s0, z_suffix))
1116 z_suffix_errno = errno;
1117 } while (*++suf != NULL);
1119 /* No suffix found, complain using z_suffix: */
1120 strcpy(ifname, iname);
1121 #ifdef NO_MULTIPLE_DOTS
1122 if (*dot == '\0') strcpy(dot, ".");
1124 #ifdef MAX_EXT_CHARS
1125 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1126 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1128 strcat(ifname, z_suffix);
1129 errno = z_suffix_errno;
1134 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1139 /* ========================================================================
1140 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1141 * Sets save_orig_name to true if the file name has been truncated.
1143 local int make_ofname()
1145 char *suff; /* ofname z suffix */
1147 strcpy(ofname, ifname);
1148 /* strip a version number if any and get the gzip suffix if present: */
1149 suff = get_suffix(ofname);
1153 /* With -t or -l, try all files (even without .gz suffix)
1154 * except with -r (behave as with just -dr).
1156 if (!recursive && (list || test)) return OK;
1158 /* Avoid annoying messages with -r */
1159 if (verbose || (!recursive && !quiet)) {
1160 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1161 program_name, ifname));
1165 /* Make a special case for .tgz and .taz: */
1167 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1168 strcpy(suff, ".tar");
1170 *suff = '\0'; /* strip the z suffix */
1172 /* ofname might be changed later if infile contains an original name */
1174 } else if (suff != NULL) {
1175 /* Avoid annoying messages with -r (see treat_dir()) */
1176 if (verbose || (!recursive && !quiet)) {
1177 /* Don't use WARN, as it affects exit status. */
1178 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1179 program_name, ifname, suff);
1185 #ifdef NO_MULTIPLE_DOTS
1186 suff = strrchr(ofname, '.');
1188 if (sizeof ofname <= strlen (ofname) + 1)
1190 strcat(ofname, ".");
1191 # ifdef MAX_EXT_CHARS
1192 if (strequ(z_suffix, "z")) {
1193 if (sizeof ofname <= strlen (ofname) + 2)
1195 strcat(ofname, "gz"); /* enough room */
1198 /* On the Atari and some versions of MSDOS,
1199 * ENAMETOOLONG does not work correctly. So we
1200 * must truncate here.
1202 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1203 suff[MAX_SUFFIX+1-z_len] = '\0';
1207 #endif /* NO_MULTIPLE_DOTS */
1208 if (sizeof ofname <= strlen (ofname) + z_len)
1210 strcat(ofname, z_suffix);
1212 } /* decompress ? */
1216 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1221 /* ========================================================================
1222 * Check the magic number of the input file and update ofname if an
1223 * original name was given and to_stdout is not set.
1224 * Return the compression method, -1 for error, -2 for warning.
1225 * Set inptr to the offset of the next byte to be processed.
1226 * Updates time_stamp if there is one and --no-time is not used.
1227 * This function may be called repeatedly for an input file consisting
1228 * of several contiguous gzip'ed members.
1229 * IN assertions: there is at least one remaining compressed member.
1230 * If the member is a zip file, it must be the only one.
1232 local int get_method(in)
1233 int in; /* input file descriptor */
1235 uch flags; /* compression flags */
1236 char magic[2]; /* magic header */
1237 int imagic1; /* like magic[1], but can represent EOF */
1238 ulg stamp; /* time stamp */
1240 /* If --force and --stdout, zcat == cat, so do not complain about
1241 * premature end of file: use try_byte instead of get_byte.
1243 if (force && to_stdout) {
1244 magic[0] = (char)try_byte();
1245 imagic1 = try_byte ();
1246 magic[1] = (char) imagic1;
1247 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1249 magic[0] = (char)get_byte();
1250 magic[1] = (char)get_byte();
1251 imagic1 = 0; /* avoid lint warning */
1253 method = -1; /* unknown yet */
1254 part_nb++; /* number of parts in gzip file */
1256 last_member = RECORD_IO;
1257 /* assume multiple members in gzip file except for record oriented I/O */
1259 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1260 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1262 method = (int)get_byte();
1263 if (method != DEFLATED) {
1265 "%s: %s: unknown method %d -- not supported\n",
1266 program_name, ifname, method);
1271 flags = (uch)get_byte();
1273 if ((flags & ENCRYPTED) != 0) {
1275 "%s: %s is encrypted -- not supported\n",
1276 program_name, ifname);
1280 if ((flags & CONTINUATION) != 0) {
1282 "%s: %s is a a multi-part gzip file -- not supported\n",
1283 program_name, ifname);
1285 if (force <= 1) return -1;
1287 if ((flags & RESERVED) != 0) {
1289 "%s: %s has flags 0x%x -- not supported\n",
1290 program_name, ifname, flags);
1292 if (force <= 1) return -1;
1294 stamp = (ulg)get_byte();
1295 stamp |= ((ulg)get_byte()) << 8;
1296 stamp |= ((ulg)get_byte()) << 16;
1297 stamp |= ((ulg)get_byte()) << 24;
1298 if (stamp != 0 && !no_time)
1300 time_stamp.tv_sec = stamp;
1301 time_stamp.tv_nsec = 0;
1304 (void)get_byte(); /* Ignore extra flags for the moment */
1305 (void)get_byte(); /* Ignore OS type for the moment */
1307 if ((flags & CONTINUATION) != 0) {
1308 unsigned part = (unsigned)get_byte();
1309 part |= ((unsigned)get_byte())<<8;
1311 fprintf(stderr,"%s: %s: part number %u\n",
1312 program_name, ifname, part);
1315 if ((flags & EXTRA_FIELD) != 0) {
1316 unsigned len = (unsigned)get_byte();
1317 len |= ((unsigned)get_byte())<<8;
1319 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1320 program_name, ifname, len);
1322 while (len--) (void)get_byte();
1325 /* Get original file name if it was truncated */
1326 if ((flags & ORIG_NAME) != 0) {
1327 if (no_name || (to_stdout && !list) || part_nb > 1) {
1328 /* Discard the old name */
1329 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1330 do {c=get_byte();} while (c != 0);
1332 /* Copy the base name. Keep a directory prefix intact. */
1333 char *p = gzip_base_name (ofname);
1336 *p = (char)get_char();
1337 if (*p++ == '\0') break;
1338 if (p >= ofname+sizeof(ofname)) {
1339 gzip_error ("corrupted input -- file name too large");
1342 p = gzip_base_name (base);
1343 memmove (base, p, strlen (p) + 1);
1344 /* If necessary, adapt the name to local OS conventions: */
1346 MAKE_LEGAL_NAME(base);
1347 if (base) list=0; /* avoid warning about unused variable */
1349 } /* no_name || to_stdout */
1352 /* Discard file comment if any */
1353 if ((flags & COMMENT) != 0) {
1354 while (get_char() != 0) /* null */ ;
1357 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1360 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1361 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1362 /* To simplify the code, we support a zip file when alone only.
1363 * We are thus guaranteed that the entire local header fits in inbuf.
1367 if (check_zipfile(in) != OK) return -1;
1368 /* check_zipfile may get ofname from the local header */
1371 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1375 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1377 method = COMPRESSED;
1380 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1385 } else if (force && to_stdout && !list) { /* pass input unchanged */
1391 if (method >= 0) return method;
1394 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1395 program_name, ifname);
1402 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1407 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1408 program_name, ifname));
1413 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1414 program_name, ifname));
1419 /* ========================================================================
1420 * Display the characteristics of the compressed file.
1421 * If the given method is < 0, display the accumulated totals.
1422 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1424 local void do_list(ifd, method)
1425 int ifd; /* input file descriptor */
1426 int method; /* compression method */
1428 ulg crc; /* original crc */
1429 static int first_time = 1;
1430 static char* methods[MAX_METHODS] = {
1435 "", "", "", "", /* 4 to 7 reserved */
1437 int positive_off_t_width = 1;
1440 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1441 positive_off_t_width++;
1444 if (first_time && method >= 0) {
1447 printf("method crc date time ");
1450 printf("%*.*s %*.*s ratio uncompressed_name\n",
1451 positive_off_t_width, positive_off_t_width, "compressed",
1452 positive_off_t_width, positive_off_t_width, "uncompressed");
1454 } else if (method < 0) {
1455 if (total_in <= 0 || total_out <= 0) return;
1459 if (verbose || !quiet) {
1460 fprint_off(stdout, total_in, positive_off_t_width);
1462 fprint_off(stdout, total_out, positive_off_t_width);
1465 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1466 /* header_bytes is not meaningful but used to ensure the same
1467 * ratio if there is a single file.
1469 printf(" (totals)\n");
1472 crc = (ulg)~0; /* unknown */
1474 bytes_in = ifile_size;
1477 if (method == DEFLATED && !last_member) {
1478 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1479 * If the lseek fails, we could use read() to get to the end, but
1480 * --list is used to get quick results.
1481 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1482 * you are not concerned about speed.
1484 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1485 if (bytes_in != -1L) {
1488 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1492 bytes_out = LG(buf+4);
1495 #endif /* RECORD_IO */
1498 struct tm *tm = localtime (&time_stamp.tv_sec);
1499 printf ("%5s %08lx ", methods[method], crc);
1501 printf ("%s%3d %02d:%02d ",
1502 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1504 tm->tm_mday, tm->tm_hour, tm->tm_min);
1506 printf ("??? ?? ??:?? ");
1508 fprint_off(stdout, bytes_in, positive_off_t_width);
1510 fprint_off(stdout, bytes_out, positive_off_t_width);
1512 if (bytes_in == -1L) {
1514 bytes_in = bytes_out = header_bytes = 0;
1515 } else if (total_in >= 0) {
1516 total_in += bytes_in;
1518 if (bytes_out == -1L) {
1520 bytes_in = bytes_out = header_bytes = 0;
1521 } else if (total_out >= 0) {
1522 total_out += bytes_out;
1524 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1525 printf(" %s\n", ofname);
1528 /* ========================================================================
1529 * Shorten the given name by one character, or replace a .tar extension
1530 * with .tgz. Truncate the last part of the name which is longer than
1531 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1532 * has only parts shorter than MIN_PART truncate the longest part.
1533 * For decompression, just remove the last character of the name.
1535 * IN assertion: for compression, the suffix of the given name is z_suffix.
1537 local void shorten_name(name)
1540 int len; /* length of name without z_suffix */
1541 char *trunc = NULL; /* character to be truncated */
1542 int plen; /* current part length */
1543 int min_part = MIN_PART; /* current minimum part length */
1549 gzip_error ("name too short");
1553 p = get_suffix(name);
1555 gzip_error ("can't recover suffix\n");
1559 /* compress 1234567890.tar to 1234567890.tgz */
1560 if (len > 4 && strequ(p-4, ".tar")) {
1561 strcpy(p-4, ".tgz");
1564 /* Try keeping short extensions intact:
1565 * 1234.678.012.gz -> 123.678.012.gz
1568 p = strrchr(name, PATH_SEP);
1571 plen = strcspn(p, PART_SEP);
1573 if (plen > min_part) trunc = p-1;
1576 } while (trunc == NULL && --min_part != 0);
1578 if (trunc != NULL) {
1580 trunc[0] = trunc[1];
1584 trunc = strrchr(name, PART_SEP[0]);
1586 gzip_error ("internal error in shorten_name");
1587 if (trunc[1] == '\0') trunc--; /* force truncation */
1589 strcpy(trunc, z_suffix);
1592 /* ========================================================================
1593 * The compressed file already exists, so ask for confirmation.
1594 * Return ERROR if the file must be skipped.
1596 local int check_ofname()
1598 /* Ask permission to overwrite the existing file */
1601 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1602 if (foreground && isatty(fileno(stdin))) {
1603 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1608 fprintf(stderr, "\tnot overwritten\n");
1609 if (exit_code == OK) exit_code = WARNING;
1613 if (xunlink (ofname)) {
1621 /* ========================================================================
1622 * Copy modes, times, ownership from input file to output file.
1623 * IN assertion: to_stdout is false.
1625 local void copy_stat(ifstat)
1626 struct stat *ifstat;
1628 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1632 struct timespec timespec[2];
1633 timespec[0] = get_stat_atime (ifstat);
1634 timespec[1] = get_stat_mtime (ifstat);
1636 if (decompress && 0 <= time_stamp.tv_nsec
1637 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1638 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1640 timespec[1] = time_stamp;
1642 fprintf(stderr, "%s: time stamp restored\n", ofname);
1646 if (futimens (ofd, ofname, timespec) != 0)
1649 WARN ((stderr, "%s: ", program_name));
1660 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1662 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1666 /* Copy the protection modes */
1668 r = fchmod (ofd, mode);
1670 r = chmod (ofname, mode);
1674 WARN ((stderr, "%s: ", program_name));
1684 /* ========================================================================
1685 * Recurse through the given directory. This code is taken from ncompress.
1687 local void treat_dir (fd, dir)
1693 char nbuf[MAX_PATH_LEN];
1697 dirp = fdopendir (fd);
1700 dirp = opendir(dir);
1711 ** WARNING: the following algorithm could occasionally cause
1712 ** compress to produce error warnings of the form "<filename>.gz
1713 ** already has .gz suffix - ignored". This occurs when the
1714 ** .gz output file is inserted into the directory below
1715 ** readdir's current pointer.
1716 ** These warnings are harmless but annoying, so they are suppressed
1717 ** with option -r (except when -v is on). An alternative
1718 ** to allowing this would be to store the entire directory
1719 ** list in memory, then compress the entries in the stored
1720 ** list. Given the depth-first recursive algorithm used here,
1721 ** this could use up a tremendous amount of memory. I don't
1722 ** think it's worth it. -- Dave Mack
1723 ** (An other alternative might be two passes to avoid depth-first.)
1726 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1728 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1732 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1734 if (len != 0 /* dir = "" means current dir on Amiga */
1736 && dir[len-1] != PATH_SEP2
1739 && dir[len-1] != PATH_SEP3
1742 nbuf[len++] = PATH_SEP;
1744 strcpy(nbuf+len, dp->d_name);
1747 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1748 program_name, dir, dp->d_name);
1754 if (CLOSEDIR(dirp) != 0)
1757 #endif /* ! NO_DIR */
1759 /* Make sure signals get handled properly. */
1762 install_signal_handlers ()
1766 /* SIGINT must be first, as 'foreground' depends on it. */
1787 int nsigs = sizeof sig / sizeof sig[0];
1791 struct sigaction act;
1793 sigemptyset (&caught_signals);
1794 for (i = 0; i < nsigs; i++)
1796 sigaction (sig[i], NULL, &act);
1797 if (act.sa_handler != SIG_IGN)
1798 sigaddset (&caught_signals, 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, sig[i]))
1810 sigaction (sig[i], &act, NULL);
1813 for (i = 0; i < nsigs; i++)
1814 if (signal (sig[i], SIG_IGN) != SIG_IGN)
1818 signal (sig[i], abort_gzip_signal);
1819 siginterrupt (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);