1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001-2002, 2006-2007, 2009 Free Software Foundation,
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 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>
70 #include "fcntl-safer.h"
72 #include "stat-time.h"
88 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
99 # ifndef _D_EXACT_NAMLEN
100 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
105 # define CLOSEDIR(d) (closedir(d), 0)
107 # define CLOSEDIR(d) closedir(d)
111 # include <utimens.h>
114 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
117 # define MAX_PATH_LEN 1024 /* max pathname length */
129 off_t lseek OF((int fd, off_t offset, int whence));
133 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
137 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
140 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
143 # define SA_NOCLDSTOP 0
144 # define sigprocmask(how, set, oset) /* empty */
145 # define sigset_t int
146 # if ! HAVE_SIGINTERRUPT
147 # define siginterrupt(sig, flag) /* empty */
151 #ifndef HAVE_WORKING_O_NOFOLLOW
152 # define HAVE_WORKING_O_NOFOLLOW 0
156 # define ELOOP EINVAL
159 /* Separator for file name parts (see shorten_name()) */
160 #ifdef NO_MULTIPLE_DOTS
161 # define PART_SEP "-"
163 # define PART_SEP "."
168 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
169 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
170 DECLARE(ush, d_buf, DIST_BUFSIZE);
171 DECLARE(uch, window, 2L*WSIZE);
173 DECLARE(ush, tab_prefix, 1L<<BITS);
175 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
176 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
179 /* local variables */
181 int ascii = 0; /* convert end-of-lines to local OS conventions */
182 int to_stdout = 0; /* output to stdout (-c) */
183 int decompress = 0; /* decompress (-d) */
184 int force = 0; /* don't ask questions, compress links (-f) */
185 int no_name = -1; /* don't save or restore the original file name */
186 int no_time = -1; /* don't save or restore the original file time */
187 int recursive = 0; /* recurse through directories (-r) */
188 int list = 0; /* list the file contents (-l) */
189 int verbose = 0; /* be verbose (-v) */
190 int quiet = 0; /* be very quiet (-q) */
191 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
192 int test = 0; /* test .gz file integrity */
193 int foreground = 0; /* set if program run in foreground */
194 char *program_name; /* program name */
195 int maxbits = BITS; /* max bits per code for LZW */
196 int method = DEFLATED;/* compression method */
197 int level = 6; /* compression level */
198 int exit_code = OK; /* program exit code */
199 int save_orig_name; /* set if original name must be saved */
200 int last_member; /* set for .zip and .Z files */
201 int part_nb; /* number of parts in .gz file */
202 struct timespec time_stamp; /* original time stamp (modification time) */
203 off_t ifile_size; /* input file size, -1 for devices (debug only) */
204 char *env; /* contents of GZIP env variable */
205 char **args = NULL; /* argv pointer if GZIP env variable defined */
206 char const *z_suffix; /* default suffix (can be set with --suffix) */
207 size_t z_len; /* strlen(z_suffix) */
209 /* The set of signals that are caught. */
210 static sigset_t caught_signals;
212 /* If nonzero then exit with status WARNING, rather than with the usual
213 signal status, on receipt of a signal with this value. This
214 suppresses a "Broken Pipe" message with some shells. */
215 static int volatile exiting_signal;
217 /* If nonnegative, close this file descriptor and unlink ofname on error. */
218 static int volatile remove_ofname_fd = -1;
220 off_t bytes_in; /* number of input bytes */
221 off_t bytes_out; /* number of output bytes */
222 off_t total_in; /* input bytes for all files */
223 off_t total_out; /* output bytes for all files */
224 char ifname[MAX_PATH_LEN]; /* input file name */
225 char ofname[MAX_PATH_LEN]; /* output file name */
226 struct stat istat; /* status for input file */
227 int ifd; /* input file descriptor */
228 int ofd; /* output file descriptor */
229 unsigned insize; /* valid bytes in inbuf */
230 unsigned inptr; /* index of next byte to be processed in inbuf */
231 unsigned outcnt; /* bytes in output buffer */
233 static int handled_sig[] =
235 /* SIGINT must be first, as 'foreground' depends on it. */
257 struct option longopts[] =
259 /* { name has_arg *flag val } */
260 {"ascii", 0, 0, 'a'}, /* ascii text mode */
261 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
262 {"stdout", 0, 0, 'c'}, /* write output on standard output */
263 {"decompress", 0, 0, 'd'}, /* decompress */
264 {"uncompress", 0, 0, 'd'}, /* decompress */
265 /* {"encrypt", 0, 0, 'e'}, encrypt */
266 {"force", 0, 0, 'f'}, /* force overwrite of output file */
267 {"help", 0, 0, 'h'}, /* give help */
268 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
269 {"list", 0, 0, 'l'}, /* list .gz file contents */
270 {"license", 0, 0, 'L'}, /* display software license */
271 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
272 {"name", 0, 0, 'N'}, /* save or restore original name & time */
273 {"quiet", 0, 0, 'q'}, /* quiet mode */
274 {"silent", 0, 0, 'q'}, /* quiet mode */
275 {"recursive", 0, 0, 'r'}, /* recurse through directories */
276 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
277 {"test", 0, 0, 't'}, /* test compressed file integrity */
278 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
279 {"verbose", 0, 0, 'v'}, /* verbose mode */
280 {"version", 0, 0, 'V'}, /* display version number */
281 {"fast", 0, 0, '1'}, /* compress faster */
282 {"best", 0, 0, '9'}, /* compress better */
283 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
284 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
288 /* local functions */
290 local void try_help OF((void)) ATTRIBUTE_NORETURN;
291 local void help OF((void));
292 local void license OF((void));
293 local void version OF((void));
294 local int input_eof OF((void));
295 local void treat_stdin OF((void));
296 local void treat_file OF((char *iname));
297 local int create_outfile OF((void));
298 local char *get_suffix OF((char *name));
299 local int open_input_file OF((char *iname, struct stat *sbuf));
300 local int make_ofname OF((void));
301 local void shorten_name OF((char *name));
302 local int get_method OF((int in));
303 local void do_list OF((int ifd, int method));
304 local int check_ofname OF((void));
305 local void copy_stat OF((struct stat *ifstat));
306 local void install_signal_handlers OF((void));
307 local void remove_output_file OF((void));
308 local RETSIGTYPE abort_gzip_signal OF((int));
309 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
310 int main OF((int argc, char **argv));
311 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
314 local void treat_dir OF((int fd, char *dir));
317 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
322 fprintf (stderr, "Try `%s --help' for more information.\n",
327 /* ======================================================================== */
330 static char const* const help_msg[] = {
331 "Compress or uncompress FILEs (by default, compress FILES in-place).",
333 "Mandatory arguments to long options are mandatory for short options too.",
336 " -a, --ascii ascii text; convert end-of-line using local conventions",
338 " -c, --stdout write on standard output, keep original files unchanged",
339 " -d, --decompress decompress",
340 /* -e, --encrypt encrypt */
341 " -f, --force force overwrite of output file and compress links",
342 " -h, --help give this help",
343 /* -k, --pkzip force output in pkzip format */
344 " -l, --list list compressed file contents",
345 " -L, --license display software license",
347 " -m, --no-time do not save or restore the original modification time",
348 " -M, --time save or restore the original modification time",
350 " -n, --no-name do not save or restore the original name and time stamp",
351 " -N, --name save or restore the original name and time stamp",
352 " -q, --quiet suppress all warnings",
354 " -r, --recursive operate recursively on directories",
356 " -S, --suffix=SUF use suffix SUF on compressed files",
357 " -t, --test test compressed file integrity",
358 " -v, --verbose verbose mode",
359 " -V, --version display version number",
360 " -1, --fast compress faster",
361 " -9, --best compress better",
363 " -Z, --lzw produce output compatible with old compress",
364 " -b, --bits=BITS max number of bits per code (implies -Z)",
367 "With no FILE, or when FILE is -, read standard input.",
369 "Report bugs to <bug-gzip@gnu.org>.",
371 char const *const *p = help_msg;
373 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
374 while (*p) printf ("%s\n", *p++);
377 /* ======================================================================== */
380 char const *const *p = license_msg;
382 printf ("%s %s\n", program_name, VERSION);
383 while (*p) printf ("%s\n", *p++);
386 /* ======================================================================== */
391 printf ("Written by Jean-loup Gailly.\n");
394 local void progerror (char const *string)
397 fprintf (stderr, "%s: ", program_name);
403 /* ======================================================================== */
404 int main (int argc, char **argv)
406 int file_count; /* number of files to process */
407 size_t proglen; /* length of program_name */
408 int optc; /* current option */
410 EXPAND(argc, argv); /* wild card expansion if necessary */
412 program_name = gzip_base_name (argv[0]);
413 proglen = strlen (program_name);
415 atexit (close_stdin);
417 /* Suppress .exe for MSDOS, OS/2 and VMS: */
418 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
419 program_name[proglen - 4] = '\0';
421 /* Add options in GZIP environment variable if there is one */
422 env = add_envopt(&argc, &argv, OPTIONS_VAR);
423 if (env != NULL) args = argv;
426 # define GNU_STANDARD 1
429 /* For compatibility with old compress, use program name as an option.
430 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
431 * gzip even if it is invoked under the name gunzip or zcat.
433 * Systems which do not support links can still use -d or -dc.
434 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
436 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
437 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
439 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
440 || strequ (program_name, "gzcat")) /* gzcat */
441 decompress = to_stdout = 1;
445 z_len = strlen(z_suffix);
447 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
448 longopts, (int *)0)) != -1) {
453 maxbits = atoi(optarg);
454 for (; *optarg; optarg++)
455 if (! ('0' <= *optarg && *optarg <= '9'))
457 fprintf (stderr, "%s: -b operand is not an integer\n",
463 to_stdout = 1; break;
465 decompress = 1; break;
469 help(); do_exit(OK); break;
471 list = decompress = to_stdout = 1; break;
473 license(); do_exit(OK); break;
474 case 'm': /* undocumented, may change later */
476 case 'M': /* undocumented, may change later */
479 no_name = no_time = 1; break;
481 no_name = no_time = 0; break;
483 quiet = 1; verbose = 0; break;
486 fprintf (stderr, "%s: -r not supported on this system\n",
494 #ifdef NO_MULTIPLE_DOTS
495 if (*optarg == '.') optarg++;
497 z_len = strlen(optarg);
501 test = decompress = to_stdout = 1;
504 verbose++; quiet = 0; break;
506 version(); do_exit(OK); break;
511 fprintf(stderr, "%s: -Z not supported in this version\n",
516 case '1': case '2': case '3': case '4':
517 case '5': case '6': case '7': case '8': case '9':
521 /* Error message already emitted by getopt_long. */
524 } /* loop on all arguments */
526 /* By default, save name and timestamp on compression but do not
527 * restore them on decompression.
529 if (no_time < 0) no_time = decompress;
530 if (no_name < 0) no_name = decompress;
532 file_count = argc - optind;
536 if (ascii && !quiet) {
537 fprintf(stderr, "%s: option --ascii ignored on this system\n",
541 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
542 fprintf(stderr, "%s: incorrect suffix '%s'\n",
543 program_name, z_suffix);
546 if (do_lzw && !decompress) work = lzw;
548 /* Allocate all global buffers (for DYN_ALLOC option) */
549 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
550 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
551 ALLOC(ush, d_buf, DIST_BUFSIZE);
552 ALLOC(uch, window, 2L*WSIZE);
554 ALLOC(ush, tab_prefix, 1L<<BITS);
556 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
557 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
560 exiting_signal = quiet ? SIGPIPE : 0;
561 install_signal_handlers ();
563 /* And get to work */
564 if (file_count != 0) {
565 if (to_stdout && !test && !list && (!decompress || !ascii)) {
566 SET_BINARY_MODE(fileno(stdout));
568 while (optind < argc) {
569 treat_file(argv[optind++]);
571 } else { /* Standard input */
574 if (list && !quiet && file_count > 1) {
575 do_list(-1, -1); /* print totals */
578 return exit_code; /* just to avoid lint warning */
581 /* Return nonzero when at end of file on input. */
585 if (!decompress || last_member)
590 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
593 /* Unget the char that fill_inbuf got. */
600 /* ========================================================================
601 * Compress or decompress stdin
603 local void treat_stdin()
605 if (!force && !list &&
606 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
607 /* Do not send compressed data to the terminal or read it from
608 * the terminal. We get here when user invoked the program
609 * without parameters, so be helpful. According to the GNU standards:
611 * If there is one behavior you think is most useful when the output
612 * is to a terminal, and another that you think is most useful when
613 * the output is a file or a pipe, then it is usually best to make
614 * the default behavior the one that is useful with output to a
615 * terminal, and have an option for the other behavior.
617 * Here we use the --force option to get the other behavior.
620 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
621 program_name, decompress ? "read from" : "written to",
622 decompress ? "de" : "");
623 fprintf (stderr, "For help, type: %s -h\n", program_name);
627 if (decompress || !ascii) {
628 SET_BINARY_MODE(fileno(stdin));
630 if (!test && !list && (!decompress || !ascii)) {
631 SET_BINARY_MODE(fileno(stdout));
633 strcpy(ifname, "stdin");
634 strcpy(ofname, "stdout");
636 /* Get the file's time stamp and size. */
637 if (fstat (fileno (stdin), &istat) != 0)
639 progerror ("standard input");
642 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
643 time_stamp.tv_nsec = -1;
644 if (!no_time || list)
645 time_stamp = get_stat_mtime (&istat);
647 clear_bufs(); /* clear input and output buffers */
653 method = get_method(ifd);
655 do_exit(exit_code); /* error message already emitted */
659 do_list(ifd, method);
663 /* Actually do the compression/decompression. Loop over zipped members.
666 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
671 method = get_method(ifd);
672 if (method < 0) return; /* error message already emitted */
673 bytes_out = 0; /* required for length check */
678 fprintf(stderr, " OK\n");
680 } else if (!decompress) {
681 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
682 fprintf(stderr, "\n");
683 #ifdef DISPLAY_STDIN_RATIO
685 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
686 fprintf(stderr, "\n");
692 /* ========================================================================
693 * Compress or decompress the given file
695 local void treat_file(iname)
698 /* Accept "-" as synonym for stdin */
699 if (strequ(iname, "-")) {
700 int cflag = to_stdout;
706 /* Check if the input file is present, set ifname and istat: */
707 ifd = open_input_file (iname, &istat);
711 /* If the input name is that of a directory, recurse or ignore: */
712 if (S_ISDIR(istat.st_mode)) {
715 treat_dir (ifd, iname);
716 /* Warning: ifname is now garbage */
721 WARN ((stderr, "%s: %s is a directory -- ignored\n",
722 program_name, ifname));
728 if (! S_ISREG (istat.st_mode))
731 "%s: %s is not a directory or a regular file - ignored\n",
732 program_name, ifname));
736 if (istat.st_mode & S_ISUID)
738 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
739 program_name, ifname));
743 if (istat.st_mode & S_ISGID)
745 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
746 program_name, ifname));
753 if (istat.st_mode & S_ISVTX)
756 "%s: %s has the sticky bit set - file ignored\n",
757 program_name, ifname));
761 if (2 <= istat.st_nlink)
763 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
764 program_name, ifname,
765 (unsigned long int) istat.st_nlink - 1,
766 istat.st_nlink == 2 ? ' ' : 's'));
773 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
774 time_stamp.tv_nsec = -1;
775 if (!no_time || list)
776 time_stamp = get_stat_mtime (&istat);
778 /* Generate output file name. For -r and (-t or -l), skip files
779 * without a valid gzip suffix (check done in make_ofname).
781 if (to_stdout && !list && !test) {
782 strcpy(ofname, "stdout");
784 } else if (make_ofname() != OK) {
789 clear_bufs(); /* clear input and output buffers */
793 method = get_method(ifd); /* updates ofname if original given */
796 return; /* error message already emitted */
800 do_list(ifd, method);
801 if (close (ifd) != 0)
806 /* If compressing to a file, check if ofname is not ambiguous
807 * because the operating system truncates names. Otherwise, generate
808 * a new ofname and save the original name in the compressed file.
811 ofd = fileno(stdout);
812 /* Keep remove_ofname_fd negative. */
814 if (create_outfile() != OK) return;
816 if (!decompress && save_orig_name && !verbose && !quiet) {
817 fprintf(stderr, "%s: %s compressed to %s\n",
818 program_name, ifname, ofname);
821 /* Keep the name even if not truncated except with --no-name: */
822 if (!save_orig_name) save_orig_name = !no_name;
825 fprintf(stderr, "%s:\t", ifname);
828 /* Actually do the compression/decompression. Loop over zipped members.
831 if ((*work)(ifd, ofd) != OK) {
832 method = -1; /* force cleanup */
839 method = get_method(ifd);
840 if (method < 0) break; /* error message already emitted */
841 bytes_out = 0; /* required for length check */
844 if (close (ifd) != 0)
853 if (close (ofd) != 0)
856 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
857 remove_ofname_fd = -1;
858 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
859 sigprocmask (SIG_SETMASK, &oldset, NULL);
863 WARN ((stderr, "%s: ", program_name));
866 errno = unlink_errno;
874 remove_output_file ();
878 /* Display statistics */
881 fprintf(stderr, " OK");
882 } else if (decompress) {
883 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
885 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
887 if (!test && !to_stdout) {
888 fprintf(stderr, " -- replaced with %s", ofname);
890 fprintf(stderr, "\n");
894 /* ========================================================================
895 * Create the output file. Return OK or ERROR.
896 * Try several times if necessary to avoid truncating the z_suffix. For
897 * example, do not create a compressed file of name "1234567890123."
898 * Sets save_orig_name to true if the file name has been truncated.
899 * IN assertions: the input file has already been open (ifd is set) and
900 * ofname has already been updated if there was an original name.
901 * OUT assertions: ifd and ofd are closed in case of error.
903 local int create_outfile()
905 int name_shortened = 0;
906 int flags = (O_WRONLY | O_CREAT | O_EXCL
907 | (ascii && decompress ? 0 : O_BINARY));
914 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
915 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
917 sigprocmask (SIG_SETMASK, &oldset, NULL);
926 shorten_name (ofname);
932 if (check_ofname () != OK)
946 if (name_shortened && decompress)
948 /* name might be too long if an original name was saved */
949 WARN ((stderr, "%s: %s: warning, name truncated\n",
950 program_name, ofname));
956 /* ========================================================================
957 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
958 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
959 * accepted suffixes, in addition to the value of the --suffix option.
960 * ".tgz" is a useful convention for tar.z files on systems limited
961 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
962 * also accepted suffixes. For Unix, we do not want to accept any
963 * .??z suffix as indicating a compressed file; some people use .xyz
964 * to denote volume data.
965 * On systems allowing multiple versions of the same file (such as VMS),
966 * this function removes any version suffix in the given name.
968 local char *get_suffix(name)
972 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
973 static char const *known_suffixes[] =
974 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
979 char const **suf = known_suffixes;
982 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
985 /* strip a version number from the file name */
987 char *v = strrchr(name, SUFFIX_SEP);
988 if (v != NULL) *v = '\0';
992 if (nlen <= MAX_SUFFIX+2) {
993 strcpy(suffix, name);
995 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
998 slen = strlen(suffix);
1000 int s = strlen(*suf);
1001 if (slen > s && suffix[slen-s-1] != PATH_SEP
1002 && strequ(suffix + slen - s, *suf)) {
1005 } while (*++suf != NULL);
1011 /* Open file NAME with the given flags and mode and store its status
1012 into *ST. Return a file descriptor to the newly opened file, or -1
1013 (setting errno) on failure. */
1015 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1019 /* Refuse to follow symbolic links unless -c or -f. */
1020 if (!to_stdout && !force)
1022 if (HAVE_WORKING_O_NOFOLLOW)
1023 flags |= O_NOFOLLOW;
1026 #if HAVE_LSTAT || defined lstat
1027 if (lstat (name, st) != 0)
1029 else if (S_ISLNK (st->st_mode))
1038 fd = OPEN (name, flags, mode);
1039 if (0 <= fd && fstat (fd, st) != 0)
1050 /* ========================================================================
1051 * Set ifname to the input file name (with a suffix appended if necessary)
1052 * and istat to its stats. For decompression, if no file exists with the
1053 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1054 * For MSDOS, we try only z_suffix and z.
1055 * Return an open file descriptor or -1.
1058 open_input_file (iname, sbuf)
1062 int ilen; /* strlen(ifname) */
1063 int z_suffix_errno = 0;
1064 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1065 char const **suf = suffixes;
1067 #ifdef NO_MULTIPLE_DOTS
1068 char *dot; /* pointer to ifname extension, or NULL */
1071 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1072 | (ascii && !decompress ? 0 : O_BINARY));
1076 if (sizeof ifname - 1 <= strlen (iname))
1079 strcpy(ifname, iname);
1081 /* If input file exists, return OK. */
1082 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1086 if (!decompress || errno != ENOENT) {
1090 /* file.ext doesn't exist, try adding a suffix (after removing any
1091 * version number for VMS).
1093 s = get_suffix(ifname);
1095 progerror(ifname); /* ifname already has z suffix and does not exist */
1098 #ifdef NO_MULTIPLE_DOTS
1099 dot = strrchr(ifname, '.');
1101 strcat(ifname, ".");
1102 dot = strrchr(ifname, '.');
1105 ilen = strlen(ifname);
1106 if (strequ(z_suffix, ".gz")) suf++;
1108 /* Search for all suffixes */
1110 char const *s0 = s = *suf;
1111 strcpy (ifname, iname);
1112 #ifdef NO_MULTIPLE_DOTS
1114 if (*dot == '\0') strcpy (dot, ".");
1116 #ifdef MAX_EXT_CHARS
1117 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1118 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1120 if (sizeof ifname <= ilen + strlen (s))
1123 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1126 if (errno != ENOENT)
1131 if (strequ (s0, z_suffix))
1132 z_suffix_errno = errno;
1133 } while (*++suf != NULL);
1135 /* No suffix found, complain using z_suffix: */
1136 strcpy(ifname, iname);
1137 #ifdef NO_MULTIPLE_DOTS
1138 if (*dot == '\0') strcpy(dot, ".");
1140 #ifdef MAX_EXT_CHARS
1141 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1142 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1144 strcat(ifname, z_suffix);
1145 errno = z_suffix_errno;
1150 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1155 /* ========================================================================
1156 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1157 * Sets save_orig_name to true if the file name has been truncated.
1159 local int make_ofname()
1161 char *suff; /* ofname z suffix */
1163 strcpy(ofname, ifname);
1164 /* strip a version number if any and get the gzip suffix if present: */
1165 suff = get_suffix(ofname);
1169 /* With -t or -l, try all files (even without .gz suffix)
1170 * except with -r (behave as with just -dr).
1172 if (!recursive && (list || test)) return OK;
1174 /* Avoid annoying messages with -r */
1175 if (verbose || (!recursive && !quiet)) {
1176 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1177 program_name, ifname));
1181 /* Make a special case for .tgz and .taz: */
1183 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1184 strcpy(suff, ".tar");
1186 *suff = '\0'; /* strip the z suffix */
1188 /* ofname might be changed later if infile contains an original name */
1190 } else if (suff && ! force) {
1191 /* Avoid annoying messages with -r (see treat_dir()) */
1192 if (verbose || (!recursive && !quiet)) {
1193 /* Don't use WARN, as it affects exit status. */
1194 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1195 program_name, ifname, suff);
1201 #ifdef NO_MULTIPLE_DOTS
1202 suff = strrchr(ofname, '.');
1204 if (sizeof ofname <= strlen (ofname) + 1)
1206 strcat(ofname, ".");
1207 # ifdef MAX_EXT_CHARS
1208 if (strequ(z_suffix, "z")) {
1209 if (sizeof ofname <= strlen (ofname) + 2)
1211 strcat(ofname, "gz"); /* enough room */
1214 /* On the Atari and some versions of MSDOS,
1215 * ENAMETOOLONG does not work correctly. So we
1216 * must truncate here.
1218 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1219 suff[MAX_SUFFIX+1-z_len] = '\0';
1223 #endif /* NO_MULTIPLE_DOTS */
1224 if (sizeof ofname <= strlen (ofname) + z_len)
1226 strcat(ofname, z_suffix);
1228 } /* decompress ? */
1232 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1237 /* ========================================================================
1238 * Check the magic number of the input file and update ofname if an
1239 * original name was given and to_stdout is not set.
1240 * Return the compression method, -1 for error, -2 for warning.
1241 * Set inptr to the offset of the next byte to be processed.
1242 * Updates time_stamp if there is one and --no-time is not used.
1243 * This function may be called repeatedly for an input file consisting
1244 * of several contiguous gzip'ed members.
1245 * IN assertions: there is at least one remaining compressed member.
1246 * If the member is a zip file, it must be the only one.
1248 local int get_method(in)
1249 int in; /* input file descriptor */
1251 uch flags; /* compression flags */
1252 char magic[2]; /* magic header */
1253 int imagic1; /* like magic[1], but can represent EOF */
1254 ulg stamp; /* time stamp */
1256 /* If --force and --stdout, zcat == cat, so do not complain about
1257 * premature end of file: use try_byte instead of get_byte.
1259 if (force && to_stdout) {
1260 magic[0] = (char)try_byte();
1261 imagic1 = try_byte ();
1262 magic[1] = (char) imagic1;
1263 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1265 magic[0] = (char)get_byte();
1267 magic[1] = (char)get_byte();
1268 imagic1 = 0; /* avoid lint warning */
1270 imagic1 = try_byte ();
1271 magic[1] = (char) imagic1;
1274 method = -1; /* unknown yet */
1275 part_nb++; /* number of parts in gzip file */
1277 last_member = RECORD_IO;
1278 /* assume multiple members in gzip file except for record oriented I/O */
1280 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1281 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1283 method = (int)get_byte();
1284 if (method != DEFLATED) {
1286 "%s: %s: unknown method %d -- not supported\n",
1287 program_name, ifname, method);
1292 flags = (uch)get_byte();
1294 if ((flags & ENCRYPTED) != 0) {
1296 "%s: %s is encrypted -- not supported\n",
1297 program_name, ifname);
1301 if ((flags & CONTINUATION) != 0) {
1303 "%s: %s is a multi-part gzip file -- not supported\n",
1304 program_name, ifname);
1306 if (force <= 1) return -1;
1308 if ((flags & RESERVED) != 0) {
1310 "%s: %s has flags 0x%x -- not supported\n",
1311 program_name, ifname, flags);
1313 if (force <= 1) return -1;
1315 stamp = (ulg)get_byte();
1316 stamp |= ((ulg)get_byte()) << 8;
1317 stamp |= ((ulg)get_byte()) << 16;
1318 stamp |= ((ulg)get_byte()) << 24;
1319 if (stamp != 0 && !no_time)
1321 time_stamp.tv_sec = stamp;
1322 time_stamp.tv_nsec = 0;
1325 (void)get_byte(); /* Ignore extra flags for the moment */
1326 (void)get_byte(); /* Ignore OS type for the moment */
1328 if ((flags & CONTINUATION) != 0) {
1329 unsigned part = (unsigned)get_byte();
1330 part |= ((unsigned)get_byte())<<8;
1332 fprintf(stderr,"%s: %s: part number %u\n",
1333 program_name, ifname, part);
1336 if ((flags & EXTRA_FIELD) != 0) {
1337 unsigned len = (unsigned)get_byte();
1338 len |= ((unsigned)get_byte())<<8;
1340 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1341 program_name, ifname, len);
1343 while (len--) (void)get_byte();
1346 /* Get original file name if it was truncated */
1347 if ((flags & ORIG_NAME) != 0) {
1348 if (no_name || (to_stdout && !list) || part_nb > 1) {
1349 /* Discard the old name */
1350 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1351 do {c=get_byte();} while (c != 0);
1353 /* Copy the base name. Keep a directory prefix intact. */
1354 char *p = gzip_base_name (ofname);
1357 *p = (char)get_char();
1358 if (*p++ == '\0') break;
1359 if (p >= ofname+sizeof(ofname)) {
1360 gzip_error ("corrupted input -- file name too large");
1363 p = gzip_base_name (base);
1364 memmove (base, p, strlen (p) + 1);
1365 /* If necessary, adapt the name to local OS conventions: */
1367 MAKE_LEGAL_NAME(base);
1368 if (base) list=0; /* avoid warning about unused variable */
1370 } /* no_name || to_stdout */
1373 /* Discard file comment if any */
1374 if ((flags & COMMENT) != 0) {
1375 while (get_char() != 0) /* null */ ;
1378 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1381 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1382 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1383 /* To simplify the code, we support a zip file when alone only.
1384 * We are thus guaranteed that the entire local header fits in inbuf.
1388 if (check_zipfile(in) != OK) return -1;
1389 /* check_zipfile may get ofname from the local header */
1392 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1396 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1398 method = COMPRESSED;
1401 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1406 } else if (force && to_stdout && !list) { /* pass input unchanged */
1412 if (method >= 0) return method;
1415 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1416 program_name, ifname);
1423 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1428 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1429 program_name, ifname));
1434 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1435 program_name, ifname));
1440 /* ========================================================================
1441 * Display the characteristics of the compressed file.
1442 * If the given method is < 0, display the accumulated totals.
1443 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1445 local void do_list(ifd, method)
1446 int ifd; /* input file descriptor */
1447 int method; /* compression method */
1449 ulg crc; /* original crc */
1450 static int first_time = 1;
1451 static char const *const methods[MAX_METHODS] = {
1456 "", "", "", "", /* 4 to 7 reserved */
1458 int positive_off_t_width = 1;
1461 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1462 positive_off_t_width++;
1465 if (first_time && method >= 0) {
1468 printf("method crc date time ");
1471 printf("%*.*s %*.*s ratio uncompressed_name\n",
1472 positive_off_t_width, positive_off_t_width, "compressed",
1473 positive_off_t_width, positive_off_t_width, "uncompressed");
1475 } else if (method < 0) {
1476 if (total_in <= 0 || total_out <= 0) return;
1480 if (verbose || !quiet) {
1481 fprint_off(stdout, total_in, positive_off_t_width);
1483 fprint_off(stdout, total_out, positive_off_t_width);
1486 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1487 /* header_bytes is not meaningful but used to ensure the same
1488 * ratio if there is a single file.
1490 printf(" (totals)\n");
1493 crc = (ulg)~0; /* unknown */
1495 bytes_in = ifile_size;
1498 if (method == DEFLATED && !last_member) {
1499 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1500 * If the lseek fails, we could use read() to get to the end, but
1501 * --list is used to get quick results.
1502 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1503 * you are not concerned about speed.
1505 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1506 if (bytes_in != -1L) {
1509 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1513 bytes_out = LG(buf+4);
1516 #endif /* RECORD_IO */
1519 struct tm *tm = localtime (&time_stamp.tv_sec);
1520 printf ("%5s %08lx ", methods[method], crc);
1522 printf ("%s%3d %02d:%02d ",
1523 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1525 tm->tm_mday, tm->tm_hour, tm->tm_min);
1527 printf ("??? ?? ??:?? ");
1529 fprint_off(stdout, bytes_in, positive_off_t_width);
1531 fprint_off(stdout, bytes_out, positive_off_t_width);
1533 if (bytes_in == -1L) {
1535 bytes_in = bytes_out = header_bytes = 0;
1536 } else if (total_in >= 0) {
1537 total_in += bytes_in;
1539 if (bytes_out == -1L) {
1541 bytes_in = bytes_out = header_bytes = 0;
1542 } else if (total_out >= 0) {
1543 total_out += bytes_out;
1545 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1546 printf(" %s\n", ofname);
1549 /* ========================================================================
1550 * Shorten the given name by one character, or replace a .tar extension
1551 * with .tgz. Truncate the last part of the name which is longer than
1552 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1553 * has only parts shorter than MIN_PART truncate the longest part.
1554 * For decompression, just remove the last character of the name.
1556 * IN assertion: for compression, the suffix of the given name is z_suffix.
1558 local void shorten_name(name)
1561 int len; /* length of name without z_suffix */
1562 char *trunc = NULL; /* character to be truncated */
1563 int plen; /* current part length */
1564 int min_part = MIN_PART; /* current minimum part length */
1570 gzip_error ("name too short");
1574 p = get_suffix(name);
1576 gzip_error ("can't recover suffix\n");
1580 /* compress 1234567890.tar to 1234567890.tgz */
1581 if (len > 4 && strequ(p-4, ".tar")) {
1582 strcpy(p-4, ".tgz");
1585 /* Try keeping short extensions intact:
1586 * 1234.678.012.gz -> 123.678.012.gz
1589 p = strrchr(name, PATH_SEP);
1592 plen = strcspn(p, PART_SEP);
1594 if (plen > min_part) trunc = p-1;
1597 } while (trunc == NULL && --min_part != 0);
1599 if (trunc != NULL) {
1601 trunc[0] = trunc[1];
1605 trunc = strrchr(name, PART_SEP[0]);
1607 gzip_error ("internal error in shorten_name");
1608 if (trunc[1] == '\0') trunc--; /* force truncation */
1610 strcpy(trunc, z_suffix);
1613 /* ========================================================================
1614 * The compressed file already exists, so ask for confirmation.
1615 * Return ERROR if the file must be skipped.
1617 local int check_ofname()
1619 /* Ask permission to overwrite the existing file */
1622 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1623 if (foreground && isatty(fileno(stdin))) {
1624 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1629 fprintf(stderr, "\tnot overwritten\n");
1630 if (exit_code == OK) exit_code = WARNING;
1634 if (xunlink (ofname)) {
1642 /* ========================================================================
1643 * Copy modes, times, ownership from input file to output file.
1644 * IN assertion: to_stdout is false.
1646 local void copy_stat(ifstat)
1647 struct stat *ifstat;
1649 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1653 struct timespec timespec[2];
1654 timespec[0] = get_stat_atime (ifstat);
1655 timespec[1] = get_stat_mtime (ifstat);
1657 if (decompress && 0 <= time_stamp.tv_nsec
1658 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1659 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1661 timespec[1] = time_stamp;
1663 fprintf(stderr, "%s: time stamp restored\n", ofname);
1667 if (gl_futimens (ofd, ofname, timespec) != 0)
1670 WARN ((stderr, "%s: ", program_name));
1681 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1683 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1687 /* Copy the protection modes */
1689 r = fchmod (ofd, mode);
1691 r = chmod (ofname, mode);
1695 WARN ((stderr, "%s: ", program_name));
1705 /* ========================================================================
1706 * Recurse through the given directory. This code is taken from ncompress.
1708 local void treat_dir (fd, dir)
1714 char nbuf[MAX_PATH_LEN];
1717 dirp = fdopendir (fd);
1725 ** WARNING: the following algorithm could occasionally cause
1726 ** compress to produce error warnings of the form "<filename>.gz
1727 ** already has .gz suffix - ignored". This occurs when the
1728 ** .gz output file is inserted into the directory below
1729 ** readdir's current pointer.
1730 ** These warnings are harmless but annoying, so they are suppressed
1731 ** with option -r (except when -v is on). An alternative
1732 ** to allowing this would be to store the entire directory
1733 ** list in memory, then compress the entries in the stored
1734 ** list. Given the depth-first recursive algorithm used here,
1735 ** this could use up a tremendous amount of memory. I don't
1736 ** think it's worth it. -- Dave Mack
1737 ** (An other alternative might be two passes to avoid depth-first.)
1740 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1742 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1746 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1748 if (len != 0 /* dir = "" means current dir on Amiga */
1750 && dir[len-1] != PATH_SEP2
1753 && dir[len-1] != PATH_SEP3
1756 nbuf[len++] = PATH_SEP;
1758 strcpy(nbuf+len, dp->d_name);
1761 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1762 program_name, dir, dp->d_name);
1768 if (CLOSEDIR(dirp) != 0)
1771 #endif /* ! NO_DIR */
1773 /* Make sure signals get handled properly. */
1776 install_signal_handlers ()
1778 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1782 struct sigaction act;
1784 sigemptyset (&caught_signals);
1785 for (i = 0; i < nsigs; i++)
1787 sigaction (handled_sig[i], NULL, &act);
1788 if (act.sa_handler != SIG_IGN)
1789 sigaddset (&caught_signals, handled_sig[i]);
1792 act.sa_handler = abort_gzip_signal;
1793 act.sa_mask = caught_signals;
1796 for (i = 0; i < nsigs; i++)
1797 if (sigismember (&caught_signals, handled_sig[i]))
1801 sigaction (handled_sig[i], &act, NULL);
1804 for (i = 0; i < nsigs; i++)
1805 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1809 signal (handled_sig[i], abort_gzip_signal);
1810 siginterrupt (handled_sig[i], 1);
1815 /* ========================================================================
1816 * Free all dynamically allocated variables and exit with the given code.
1818 local void do_exit(exitcode)
1821 static int in_exit = 0;
1823 if (in_exit) exit(exitcode);
1842 /* ========================================================================
1843 * Close and unlink the output file.
1846 remove_output_file ()
1851 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1852 fd = remove_ofname_fd;
1855 remove_ofname_fd = -1;
1859 sigprocmask (SIG_SETMASK, &oldset, NULL);
1862 /* ========================================================================
1868 remove_output_file ();
1872 /* ========================================================================
1876 abort_gzip_signal (sig)
1880 signal (sig, SIG_IGN);
1881 remove_output_file ();
1882 if (sig == exiting_signal)
1884 signal (sig, SIG_DFL);