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 struct option longopts[] =
238 /* { name has_arg *flag val } */
239 {"ascii", 0, 0, 'a'}, /* ascii text mode */
240 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
241 {"stdout", 0, 0, 'c'}, /* write output on standard output */
242 {"decompress", 0, 0, 'd'}, /* decompress */
243 {"uncompress", 0, 0, 'd'}, /* decompress */
244 /* {"encrypt", 0, 0, 'e'}, encrypt */
245 {"force", 0, 0, 'f'}, /* force overwrite of output file */
246 {"help", 0, 0, 'h'}, /* give help */
247 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
248 {"list", 0, 0, 'l'}, /* list .gz file contents */
249 {"license", 0, 0, 'L'}, /* display software license */
250 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
251 {"name", 0, 0, 'N'}, /* save or restore original name & time */
252 {"quiet", 0, 0, 'q'}, /* quiet mode */
253 {"silent", 0, 0, 'q'}, /* quiet mode */
254 {"recursive", 0, 0, 'r'}, /* recurse through directories */
255 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
256 {"test", 0, 0, 't'}, /* test compressed file integrity */
257 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
258 {"verbose", 0, 0, 'v'}, /* verbose mode */
259 {"version", 0, 0, 'V'}, /* display version number */
260 {"fast", 0, 0, '1'}, /* compress faster */
261 {"best", 0, 0, '9'}, /* compress better */
262 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
263 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
264 {"rsyncable", 0, 0, 'R'}, /* make rsync-friendly archive */
268 /* local functions */
270 local void try_help OF((void)) ATTRIBUTE_NORETURN;
271 local void help OF((void));
272 local void license OF((void));
273 local void version OF((void));
274 local int input_eof OF((void));
275 local void treat_stdin OF((void));
276 local void treat_file OF((char *iname));
277 local int create_outfile OF((void));
278 local char *get_suffix OF((char *name));
279 local int open_input_file OF((char *iname, struct stat *sbuf));
280 local int make_ofname OF((void));
281 local void shorten_name OF((char *name));
282 local int get_method OF((int in));
283 local void do_list OF((int ifd, int method));
284 local int check_ofname OF((void));
285 local void copy_stat OF((struct stat *ifstat));
286 local void install_signal_handlers OF((void));
287 local void remove_output_file OF((void));
288 local RETSIGTYPE abort_gzip_signal OF((int));
289 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
290 int main OF((int argc, char **argv));
291 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
294 local void treat_dir OF((int fd, char *dir));
297 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
302 fprintf (stderr, "Try `%s --help' for more information.\n",
307 /* ======================================================================== */
310 static char *help_msg[] = {
311 "Compress or uncompress FILEs (by default, compress FILES in-place).",
313 "Mandatory arguments to long options are mandatory for short options too.",
316 " -a, --ascii ascii text; convert end-of-line using local conventions",
318 " -c, --stdout write on standard output, keep original files unchanged",
319 " -d, --decompress decompress",
320 /* -e, --encrypt encrypt */
321 " -f, --force force overwrite of output file and compress links",
322 " -h, --help give this help",
323 /* -k, --pkzip force output in pkzip format */
324 " -l, --list list compressed file contents",
325 " -L, --license display software license",
327 " -m, --no-time do not save or restore the original modification time",
328 " -M, --time save or restore the original modification time",
330 " -n, --no-name do not save or restore the original name and time stamp",
331 " -N, --name save or restore the original name and time stamp",
332 " -q, --quiet suppress all warnings",
334 " -r, --recursive operate recursively on directories",
336 " -S, --suffix=SUF use suffix SUF on compressed files",
337 " -t, --test test compressed file integrity",
338 " -v, --verbose verbose mode",
339 " -V, --version display version number",
340 " -1, --fast compress faster",
341 " -9, --best compress better",
343 " -Z, --lzw produce output compatible with old compress",
344 " -b, --bits=BITS max number of bits per code (implies -Z)",
346 " --rsyncable Make rsync-friendly archive",
348 "With no FILE, or when FILE is -, read standard input.",
350 "Report bugs to <bug-gzip@gnu.org>.",
354 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
355 while (*p) printf ("%s\n", *p++);
358 /* ======================================================================== */
361 char **p = license_msg;
363 printf ("%s %s\n", program_name, VERSION);
364 while (*p) printf ("%s\n", *p++);
367 /* ======================================================================== */
372 printf ("Written by Jean-loup Gailly.\n");
375 local void progerror (string)
379 fprintf (stderr, "%s: ", program_name);
385 /* ======================================================================== */
386 int main (argc, argv)
390 int file_count; /* number of files to process */
391 size_t proglen; /* length of program_name */
392 int optc; /* current option */
394 EXPAND(argc, argv); /* wild card expansion if necessary */
396 program_name = gzip_base_name (argv[0]);
397 proglen = strlen (program_name);
399 /* Suppress .exe for MSDOS, OS/2 and VMS: */
400 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
401 program_name[proglen - 4] = '\0';
403 /* Add options in GZIP environment variable if there is one */
404 env = add_envopt(&argc, &argv, OPTIONS_VAR);
405 if (env != NULL) args = argv;
408 # define GNU_STANDARD 1
411 /* For compatibility with old compress, use program name as an option.
412 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
413 * gzip even if it is invoked under the name gunzip or zcat.
415 * Systems which do not support links can still use -d or -dc.
416 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
418 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
419 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
421 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
422 || strequ (program_name, "gzcat")) /* gzcat */
423 decompress = to_stdout = 1;
427 z_len = strlen(z_suffix);
429 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
430 longopts, (int *)0)) != -1) {
435 maxbits = atoi(optarg);
436 for (; *optarg; optarg++)
437 if (! ('0' <= *optarg && *optarg <= '9'))
439 fprintf (stderr, "%s: -b operand is not an integer\n",
445 to_stdout = 1; break;
447 decompress = 1; break;
451 help(); do_exit(OK); break;
453 list = decompress = to_stdout = 1; break;
455 license(); do_exit(OK); break;
456 case 'm': /* undocumented, may change later */
458 case 'M': /* undocumented, may change later */
461 no_name = no_time = 1; break;
463 no_name = no_time = 0; break;
465 quiet = 1; verbose = 0; break;
468 fprintf (stderr, "%s: -r not supported on this system\n",
478 #ifdef NO_MULTIPLE_DOTS
479 if (*optarg == '.') optarg++;
481 z_len = strlen(optarg);
485 test = decompress = to_stdout = 1;
488 verbose++; quiet = 0; break;
490 version(); do_exit(OK); break;
495 fprintf(stderr, "%s: -Z not supported in this version\n",
500 case '1': case '2': case '3': case '4':
501 case '5': case '6': case '7': case '8': case '9':
505 /* Error message already emitted by getopt_long. */
508 } /* loop on all arguments */
510 /* By default, save name and timestamp on compression but do not
511 * restore them on decompression.
513 if (no_time < 0) no_time = decompress;
514 if (no_name < 0) no_name = decompress;
516 file_count = argc - optind;
520 if (ascii && !quiet) {
521 fprintf(stderr, "%s: option --ascii ignored on this system\n",
525 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
526 fprintf(stderr, "%s: incorrect suffix '%s'\n",
527 program_name, z_suffix);
530 if (do_lzw && !decompress) work = lzw;
532 /* Allocate all global buffers (for DYN_ALLOC option) */
533 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
534 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
535 ALLOC(ush, d_buf, DIST_BUFSIZE);
536 ALLOC(uch, window, 2L*WSIZE);
538 ALLOC(ush, tab_prefix, 1L<<BITS);
540 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
541 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
544 exiting_signal = quiet ? SIGPIPE : 0;
545 install_signal_handlers ();
547 /* And get to work */
548 if (file_count != 0) {
549 if (to_stdout && !test && !list && (!decompress || !ascii)) {
550 SET_BINARY_MODE(fileno(stdout));
552 while (optind < argc) {
553 treat_file(argv[optind++]);
555 } else { /* Standard input */
558 if (list && !quiet && file_count > 1) {
559 do_list(-1, -1); /* print totals */
562 return exit_code; /* just to avoid lint warning */
565 /* Return nonzero when at end of file on input. */
569 if (!decompress || last_member)
574 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
577 /* Unget the char that fill_inbuf got. */
584 /* ========================================================================
585 * Compress or decompress stdin
587 local void treat_stdin()
589 if (!force && !list &&
590 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
591 /* Do not send compressed data to the terminal or read it from
592 * the terminal. We get here when user invoked the program
593 * without parameters, so be helpful. According to the GNU standards:
595 * If there is one behavior you think is most useful when the output
596 * is to a terminal, and another that you think is most useful when
597 * the output is a file or a pipe, then it is usually best to make
598 * the default behavior the one that is useful with output to a
599 * terminal, and have an option for the other behavior.
601 * Here we use the --force option to get the other behavior.
604 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
605 program_name, decompress ? "read from" : "written to",
606 decompress ? "de" : "");
607 fprintf (stderr, "For help, type: %s -h\n", program_name);
611 if (decompress || !ascii) {
612 SET_BINARY_MODE(fileno(stdin));
614 if (!test && !list && (!decompress || !ascii)) {
615 SET_BINARY_MODE(fileno(stdout));
617 strcpy(ifname, "stdin");
618 strcpy(ofname, "stdout");
620 /* Get the file's time stamp and size. */
621 if (fstat (fileno (stdin), &istat) != 0)
623 progerror ("standard input");
626 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
627 time_stamp.tv_nsec = -1;
628 if (!no_time || list)
629 time_stamp = get_stat_mtime (&istat);
631 clear_bufs(); /* clear input and output buffers */
636 method = get_method(ifd);
638 do_exit(exit_code); /* error message already emitted */
642 do_list(ifd, method);
646 /* Actually do the compression/decompression. Loop over zipped members.
649 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
654 method = get_method(ifd);
655 if (method < 0) return; /* error message already emitted */
656 bytes_out = 0; /* required for length check */
661 fprintf(stderr, " OK\n");
663 } else if (!decompress) {
664 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
665 fprintf(stderr, "\n");
666 #ifdef DISPLAY_STDIN_RATIO
668 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
669 fprintf(stderr, "\n");
675 /* ========================================================================
676 * Compress or decompress the given file
678 local void treat_file(iname)
681 /* Accept "-" as synonym for stdin */
682 if (strequ(iname, "-")) {
683 int cflag = to_stdout;
689 /* Check if the input file is present, set ifname and istat: */
690 ifd = open_input_file (iname, &istat);
694 /* If the input name is that of a directory, recurse or ignore: */
695 if (S_ISDIR(istat.st_mode)) {
698 treat_dir (ifd, iname);
699 /* Warning: ifname is now garbage */
704 WARN ((stderr, "%s: %s is a directory -- ignored\n",
705 program_name, ifname));
711 if (! S_ISREG (istat.st_mode))
714 "%s: %s is not a directory or a regular file - ignored\n",
715 program_name, ifname));
719 if (istat.st_mode & S_ISUID)
721 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
722 program_name, ifname));
726 if (istat.st_mode & S_ISGID)
728 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
729 program_name, ifname));
736 if (istat.st_mode & S_ISVTX)
739 "%s: %s has the sticky bit set - file ignored\n",
740 program_name, ifname));
744 if (2 <= istat.st_nlink)
746 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
747 program_name, ifname,
748 (unsigned long int) istat.st_nlink - 1,
749 istat.st_nlink == 2 ? ' ' : 's'));
756 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
757 time_stamp.tv_nsec = -1;
758 if (!no_time || list)
759 time_stamp = get_stat_mtime (&istat);
761 /* Generate output file name. For -r and (-t or -l), skip files
762 * without a valid gzip suffix (check done in make_ofname).
764 if (to_stdout && !list && !test) {
765 strcpy(ofname, "stdout");
767 } else if (make_ofname() != OK) {
772 clear_bufs(); /* clear input and output buffers */
776 method = get_method(ifd); /* updates ofname if original given */
779 return; /* error message already emitted */
783 do_list(ifd, method);
784 if (close (ifd) != 0)
789 /* If compressing to a file, check if ofname is not ambiguous
790 * because the operating system truncates names. Otherwise, generate
791 * a new ofname and save the original name in the compressed file.
794 ofd = fileno(stdout);
795 /* Keep remove_ofname_fd negative. */
797 if (create_outfile() != OK) return;
799 if (!decompress && save_orig_name && !verbose && !quiet) {
800 fprintf(stderr, "%s: %s compressed to %s\n",
801 program_name, ifname, ofname);
804 /* Keep the name even if not truncated except with --no-name: */
805 if (!save_orig_name) save_orig_name = !no_name;
808 fprintf(stderr, "%s:\t", ifname);
811 /* Actually do the compression/decompression. Loop over zipped members.
814 if ((*work)(ifd, ofd) != OK) {
815 method = -1; /* force cleanup */
822 method = get_method(ifd);
823 if (method < 0) break; /* error message already emitted */
824 bytes_out = 0; /* required for length check */
827 if (close (ifd) != 0)
836 if (close (ofd) != 0)
839 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
840 remove_ofname_fd = -1;
841 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
842 sigprocmask (SIG_SETMASK, &oldset, NULL);
846 WARN ((stderr, "%s: ", program_name));
849 errno = unlink_errno;
857 remove_output_file ();
861 /* Display statistics */
864 fprintf(stderr, " OK");
865 } else if (decompress) {
866 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
868 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
870 if (!test && !to_stdout) {
871 fprintf(stderr, " -- replaced with %s", ofname);
873 fprintf(stderr, "\n");
877 /* ========================================================================
878 * Create the output file. Return OK or ERROR.
879 * Try several times if necessary to avoid truncating the z_suffix. For
880 * example, do not create a compressed file of name "1234567890123."
881 * Sets save_orig_name to true if the file name has been truncated.
882 * IN assertions: the input file has already been open (ifd is set) and
883 * ofname has already been updated if there was an original name.
884 * OUT assertions: ifd and ofd are closed in case of error.
886 local int create_outfile()
888 int name_shortened = 0;
889 int flags = (O_WRONLY | O_CREAT | O_EXCL
890 | (ascii && decompress ? 0 : O_BINARY));
897 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
898 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
900 sigprocmask (SIG_SETMASK, &oldset, NULL);
909 shorten_name (ofname);
915 if (check_ofname () != OK)
929 if (name_shortened && decompress)
931 /* name might be too long if an original name was saved */
932 WARN ((stderr, "%s: %s: warning, name truncated\n",
933 program_name, ofname));
939 /* ========================================================================
940 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
941 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
942 * accepted suffixes, in addition to the value of the --suffix option.
943 * ".tgz" is a useful convention for tar.z files on systems limited
944 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
945 * also accepted suffixes. For Unix, we do not want to accept any
946 * .??z suffix as indicating a compressed file; some people use .xyz
947 * to denote volume data.
948 * On systems allowing multiple versions of the same file (such as VMS),
949 * this function removes any version suffix in the given name.
951 local char *get_suffix(name)
955 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
956 static char *known_suffixes[] =
957 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
962 char **suf = known_suffixes;
965 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
968 /* strip a version number from the file name */
970 char *v = strrchr(name, SUFFIX_SEP);
971 if (v != NULL) *v = '\0';
975 if (nlen <= MAX_SUFFIX+2) {
976 strcpy(suffix, name);
978 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
981 slen = strlen(suffix);
983 int s = strlen(*suf);
984 if (slen > s && suffix[slen-s-1] != PATH_SEP
985 && strequ(suffix + slen - s, *suf)) {
988 } while (*++suf != NULL);
994 /* Open file NAME with the given flags and mode and store its status
995 into *ST. Return a file descriptor to the newly opened file, or -1
996 (setting errno) on failure. */
998 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1002 /* Refuse to follow symbolic links unless -c or -f. */
1003 if (!to_stdout && !force)
1005 if (HAVE_WORKING_O_NOFOLLOW)
1006 flags |= O_NOFOLLOW;
1009 #if HAVE_LSTAT || defined lstat
1010 if (lstat (name, st) != 0)
1012 else if (S_ISLNK (st->st_mode))
1021 fd = OPEN (name, flags, mode);
1022 if (0 <= fd && fstat (fd, st) != 0)
1033 /* ========================================================================
1034 * Set ifname to the input file name (with a suffix appended if necessary)
1035 * and istat to its stats. For decompression, if no file exists with the
1036 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1037 * For MSDOS, we try only z_suffix and z.
1038 * Return an open file descriptor or -1.
1041 open_input_file (iname, sbuf)
1045 int ilen; /* strlen(ifname) */
1046 int z_suffix_errno = 0;
1047 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1048 char **suf = suffixes;
1050 #ifdef NO_MULTIPLE_DOTS
1051 char *dot; /* pointer to ifname extension, or NULL */
1054 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1055 | (ascii && !decompress ? 0 : O_BINARY));
1059 if (sizeof ifname - 1 <= strlen (iname))
1062 strcpy(ifname, iname);
1064 /* If input file exists, return OK. */
1065 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1069 if (!decompress || errno != ENOENT) {
1073 /* file.ext doesn't exist, try adding a suffix (after removing any
1074 * version number for VMS).
1076 s = get_suffix(ifname);
1078 progerror(ifname); /* ifname already has z suffix and does not exist */
1081 #ifdef NO_MULTIPLE_DOTS
1082 dot = strrchr(ifname, '.');
1084 strcat(ifname, ".");
1085 dot = strrchr(ifname, '.');
1088 ilen = strlen(ifname);
1089 if (strequ(z_suffix, ".gz")) suf++;
1091 /* Search for all suffixes */
1093 char *s0 = s = *suf;
1094 strcpy (ifname, iname);
1095 #ifdef NO_MULTIPLE_DOTS
1097 if (*dot == '\0') strcpy (dot, ".");
1099 #ifdef MAX_EXT_CHARS
1100 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1101 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1103 if (sizeof ifname <= ilen + strlen (s))
1106 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1109 if (errno != ENOENT)
1114 if (strequ (s0, z_suffix))
1115 z_suffix_errno = errno;
1116 } while (*++suf != NULL);
1118 /* No suffix found, complain using z_suffix: */
1119 strcpy(ifname, iname);
1120 #ifdef NO_MULTIPLE_DOTS
1121 if (*dot == '\0') strcpy(dot, ".");
1123 #ifdef MAX_EXT_CHARS
1124 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1125 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1127 strcat(ifname, z_suffix);
1128 errno = z_suffix_errno;
1133 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1138 /* ========================================================================
1139 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1140 * Sets save_orig_name to true if the file name has been truncated.
1142 local int make_ofname()
1144 char *suff; /* ofname z suffix */
1146 strcpy(ofname, ifname);
1147 /* strip a version number if any and get the gzip suffix if present: */
1148 suff = get_suffix(ofname);
1152 /* With -t or -l, try all files (even without .gz suffix)
1153 * except with -r (behave as with just -dr).
1155 if (!recursive && (list || test)) return OK;
1157 /* Avoid annoying messages with -r */
1158 if (verbose || (!recursive && !quiet)) {
1159 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1160 program_name, ifname));
1164 /* Make a special case for .tgz and .taz: */
1166 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1167 strcpy(suff, ".tar");
1169 *suff = '\0'; /* strip the z suffix */
1171 /* ofname might be changed later if infile contains an original name */
1173 } else if (suff != NULL) {
1174 /* Avoid annoying messages with -r (see treat_dir()) */
1175 if (verbose || (!recursive && !quiet)) {
1176 /* Don't use WARN, as it affects exit status. */
1177 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1178 program_name, ifname, suff);
1184 #ifdef NO_MULTIPLE_DOTS
1185 suff = strrchr(ofname, '.');
1187 if (sizeof ofname <= strlen (ofname) + 1)
1189 strcat(ofname, ".");
1190 # ifdef MAX_EXT_CHARS
1191 if (strequ(z_suffix, "z")) {
1192 if (sizeof ofname <= strlen (ofname) + 2)
1194 strcat(ofname, "gz"); /* enough room */
1197 /* On the Atari and some versions of MSDOS,
1198 * ENAMETOOLONG does not work correctly. So we
1199 * must truncate here.
1201 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1202 suff[MAX_SUFFIX+1-z_len] = '\0';
1206 #endif /* NO_MULTIPLE_DOTS */
1207 if (sizeof ofname <= strlen (ofname) + z_len)
1209 strcat(ofname, z_suffix);
1211 } /* decompress ? */
1215 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1220 /* ========================================================================
1221 * Check the magic number of the input file and update ofname if an
1222 * original name was given and to_stdout is not set.
1223 * Return the compression method, -1 for error, -2 for warning.
1224 * Set inptr to the offset of the next byte to be processed.
1225 * Updates time_stamp if there is one and --no-time is not used.
1226 * This function may be called repeatedly for an input file consisting
1227 * of several contiguous gzip'ed members.
1228 * IN assertions: there is at least one remaining compressed member.
1229 * If the member is a zip file, it must be the only one.
1231 local int get_method(in)
1232 int in; /* input file descriptor */
1234 uch flags; /* compression flags */
1235 char magic[2]; /* magic header */
1236 int imagic1; /* like magic[1], but can represent EOF */
1237 ulg stamp; /* time stamp */
1239 /* If --force and --stdout, zcat == cat, so do not complain about
1240 * premature end of file: use try_byte instead of get_byte.
1242 if (force && to_stdout) {
1243 magic[0] = (char)try_byte();
1244 imagic1 = try_byte ();
1245 magic[1] = (char) imagic1;
1246 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1248 magic[0] = (char)get_byte();
1249 magic[1] = (char)get_byte();
1250 imagic1 = 0; /* avoid lint warning */
1252 method = -1; /* unknown yet */
1253 part_nb++; /* number of parts in gzip file */
1255 last_member = RECORD_IO;
1256 /* assume multiple members in gzip file except for record oriented I/O */
1258 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1259 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1261 method = (int)get_byte();
1262 if (method != DEFLATED) {
1264 "%s: %s: unknown method %d -- not supported\n",
1265 program_name, ifname, method);
1270 flags = (uch)get_byte();
1272 if ((flags & ENCRYPTED) != 0) {
1274 "%s: %s is encrypted -- not supported\n",
1275 program_name, ifname);
1279 if ((flags & CONTINUATION) != 0) {
1281 "%s: %s is a multi-part gzip file -- not supported\n",
1282 program_name, ifname);
1284 if (force <= 1) return -1;
1286 if ((flags & RESERVED) != 0) {
1288 "%s: %s has flags 0x%x -- not supported\n",
1289 program_name, ifname, flags);
1291 if (force <= 1) return -1;
1293 stamp = (ulg)get_byte();
1294 stamp |= ((ulg)get_byte()) << 8;
1295 stamp |= ((ulg)get_byte()) << 16;
1296 stamp |= ((ulg)get_byte()) << 24;
1297 if (stamp != 0 && !no_time)
1299 time_stamp.tv_sec = stamp;
1300 time_stamp.tv_nsec = 0;
1303 (void)get_byte(); /* Ignore extra flags for the moment */
1304 (void)get_byte(); /* Ignore OS type for the moment */
1306 if ((flags & CONTINUATION) != 0) {
1307 unsigned part = (unsigned)get_byte();
1308 part |= ((unsigned)get_byte())<<8;
1310 fprintf(stderr,"%s: %s: part number %u\n",
1311 program_name, ifname, part);
1314 if ((flags & EXTRA_FIELD) != 0) {
1315 unsigned len = (unsigned)get_byte();
1316 len |= ((unsigned)get_byte())<<8;
1318 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1319 program_name, ifname, len);
1321 while (len--) (void)get_byte();
1324 /* Get original file name if it was truncated */
1325 if ((flags & ORIG_NAME) != 0) {
1326 if (no_name || (to_stdout && !list) || part_nb > 1) {
1327 /* Discard the old name */
1328 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1329 do {c=get_byte();} while (c != 0);
1331 /* Copy the base name. Keep a directory prefix intact. */
1332 char *p = gzip_base_name (ofname);
1335 *p = (char)get_char();
1336 if (*p++ == '\0') break;
1337 if (p >= ofname+sizeof(ofname)) {
1338 gzip_error ("corrupted input -- file name too large");
1341 p = gzip_base_name (base);
1342 memmove (base, p, strlen (p) + 1);
1343 /* If necessary, adapt the name to local OS conventions: */
1345 MAKE_LEGAL_NAME(base);
1346 if (base) list=0; /* avoid warning about unused variable */
1348 } /* no_name || to_stdout */
1351 /* Discard file comment if any */
1352 if ((flags & COMMENT) != 0) {
1353 while (get_char() != 0) /* null */ ;
1356 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1359 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1360 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1361 /* To simplify the code, we support a zip file when alone only.
1362 * We are thus guaranteed that the entire local header fits in inbuf.
1366 if (check_zipfile(in) != OK) return -1;
1367 /* check_zipfile may get ofname from the local header */
1370 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1374 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1376 method = COMPRESSED;
1379 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1384 } else if (force && to_stdout && !list) { /* pass input unchanged */
1390 if (method >= 0) return method;
1393 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1394 program_name, ifname);
1401 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1406 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1407 program_name, ifname));
1412 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1413 program_name, ifname));
1418 /* ========================================================================
1419 * Display the characteristics of the compressed file.
1420 * If the given method is < 0, display the accumulated totals.
1421 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1423 local void do_list(ifd, method)
1424 int ifd; /* input file descriptor */
1425 int method; /* compression method */
1427 ulg crc; /* original crc */
1428 static int first_time = 1;
1429 static char* methods[MAX_METHODS] = {
1434 "", "", "", "", /* 4 to 7 reserved */
1436 int positive_off_t_width = 1;
1439 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1440 positive_off_t_width++;
1443 if (first_time && method >= 0) {
1446 printf("method crc date time ");
1449 printf("%*.*s %*.*s ratio uncompressed_name\n",
1450 positive_off_t_width, positive_off_t_width, "compressed",
1451 positive_off_t_width, positive_off_t_width, "uncompressed");
1453 } else if (method < 0) {
1454 if (total_in <= 0 || total_out <= 0) return;
1458 if (verbose || !quiet) {
1459 fprint_off(stdout, total_in, positive_off_t_width);
1461 fprint_off(stdout, total_out, positive_off_t_width);
1464 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1465 /* header_bytes is not meaningful but used to ensure the same
1466 * ratio if there is a single file.
1468 printf(" (totals)\n");
1471 crc = (ulg)~0; /* unknown */
1473 bytes_in = ifile_size;
1476 if (method == DEFLATED && !last_member) {
1477 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1478 * If the lseek fails, we could use read() to get to the end, but
1479 * --list is used to get quick results.
1480 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1481 * you are not concerned about speed.
1483 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1484 if (bytes_in != -1L) {
1487 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1491 bytes_out = LG(buf+4);
1494 #endif /* RECORD_IO */
1497 struct tm *tm = localtime (&time_stamp.tv_sec);
1498 printf ("%5s %08lx ", methods[method], crc);
1500 printf ("%s%3d %02d:%02d ",
1501 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1503 tm->tm_mday, tm->tm_hour, tm->tm_min);
1505 printf ("??? ?? ??:?? ");
1507 fprint_off(stdout, bytes_in, positive_off_t_width);
1509 fprint_off(stdout, bytes_out, positive_off_t_width);
1511 if (bytes_in == -1L) {
1513 bytes_in = bytes_out = header_bytes = 0;
1514 } else if (total_in >= 0) {
1515 total_in += bytes_in;
1517 if (bytes_out == -1L) {
1519 bytes_in = bytes_out = header_bytes = 0;
1520 } else if (total_out >= 0) {
1521 total_out += bytes_out;
1523 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1524 printf(" %s\n", ofname);
1527 /* ========================================================================
1528 * Shorten the given name by one character, or replace a .tar extension
1529 * with .tgz. Truncate the last part of the name which is longer than
1530 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1531 * has only parts shorter than MIN_PART truncate the longest part.
1532 * For decompression, just remove the last character of the name.
1534 * IN assertion: for compression, the suffix of the given name is z_suffix.
1536 local void shorten_name(name)
1539 int len; /* length of name without z_suffix */
1540 char *trunc = NULL; /* character to be truncated */
1541 int plen; /* current part length */
1542 int min_part = MIN_PART; /* current minimum part length */
1548 gzip_error ("name too short");
1552 p = get_suffix(name);
1554 gzip_error ("can't recover suffix\n");
1558 /* compress 1234567890.tar to 1234567890.tgz */
1559 if (len > 4 && strequ(p-4, ".tar")) {
1560 strcpy(p-4, ".tgz");
1563 /* Try keeping short extensions intact:
1564 * 1234.678.012.gz -> 123.678.012.gz
1567 p = strrchr(name, PATH_SEP);
1570 plen = strcspn(p, PART_SEP);
1572 if (plen > min_part) trunc = p-1;
1575 } while (trunc == NULL && --min_part != 0);
1577 if (trunc != NULL) {
1579 trunc[0] = trunc[1];
1583 trunc = strrchr(name, PART_SEP[0]);
1585 gzip_error ("internal error in shorten_name");
1586 if (trunc[1] == '\0') trunc--; /* force truncation */
1588 strcpy(trunc, z_suffix);
1591 /* ========================================================================
1592 * The compressed file already exists, so ask for confirmation.
1593 * Return ERROR if the file must be skipped.
1595 local int check_ofname()
1597 /* Ask permission to overwrite the existing file */
1600 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1601 if (foreground && isatty(fileno(stdin))) {
1602 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1607 fprintf(stderr, "\tnot overwritten\n");
1608 if (exit_code == OK) exit_code = WARNING;
1612 if (xunlink (ofname)) {
1620 /* ========================================================================
1621 * Copy modes, times, ownership from input file to output file.
1622 * IN assertion: to_stdout is false.
1624 local void copy_stat(ifstat)
1625 struct stat *ifstat;
1627 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1631 struct timespec timespec[2];
1632 timespec[0] = get_stat_atime (ifstat);
1633 timespec[1] = get_stat_mtime (ifstat);
1635 if (decompress && 0 <= time_stamp.tv_nsec
1636 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1637 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1639 timespec[1] = time_stamp;
1641 fprintf(stderr, "%s: time stamp restored\n", ofname);
1645 if (gz_futimens (ofd, ofname, timespec) != 0)
1648 WARN ((stderr, "%s: ", program_name));
1659 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1661 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1665 /* Copy the protection modes */
1667 r = fchmod (ofd, mode);
1669 r = chmod (ofname, mode);
1673 WARN ((stderr, "%s: ", program_name));
1683 /* ========================================================================
1684 * Recurse through the given directory. This code is taken from ncompress.
1686 local void treat_dir (fd, dir)
1692 char nbuf[MAX_PATH_LEN];
1696 dirp = fdopendir (fd);
1699 dirp = opendir(dir);
1710 ** WARNING: the following algorithm could occasionally cause
1711 ** compress to produce error warnings of the form "<filename>.gz
1712 ** already has .gz suffix - ignored". This occurs when the
1713 ** .gz output file is inserted into the directory below
1714 ** readdir's current pointer.
1715 ** These warnings are harmless but annoying, so they are suppressed
1716 ** with option -r (except when -v is on). An alternative
1717 ** to allowing this would be to store the entire directory
1718 ** list in memory, then compress the entries in the stored
1719 ** list. Given the depth-first recursive algorithm used here,
1720 ** this could use up a tremendous amount of memory. I don't
1721 ** think it's worth it. -- Dave Mack
1722 ** (An other alternative might be two passes to avoid depth-first.)
1725 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1727 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1731 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1733 if (len != 0 /* dir = "" means current dir on Amiga */
1735 && dir[len-1] != PATH_SEP2
1738 && dir[len-1] != PATH_SEP3
1741 nbuf[len++] = PATH_SEP;
1743 strcpy(nbuf+len, dp->d_name);
1746 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1747 program_name, dir, dp->d_name);
1753 if (CLOSEDIR(dirp) != 0)
1756 #endif /* ! NO_DIR */
1758 /* Make sure signals get handled properly. */
1761 install_signal_handlers ()
1765 /* SIGINT must be first, as 'foreground' depends on it. */
1786 int nsigs = sizeof sig / sizeof sig[0];
1790 struct sigaction act;
1792 sigemptyset (&caught_signals);
1793 for (i = 0; i < nsigs; i++)
1795 sigaction (sig[i], NULL, &act);
1796 if (act.sa_handler != SIG_IGN)
1797 sigaddset (&caught_signals, sig[i]);
1800 act.sa_handler = abort_gzip_signal;
1801 act.sa_mask = caught_signals;
1804 for (i = 0; i < nsigs; i++)
1805 if (sigismember (&caught_signals, sig[i]))
1809 sigaction (sig[i], &act, NULL);
1812 for (i = 0; i < nsigs; i++)
1813 if (signal (sig[i], SIG_IGN) != SIG_IGN)
1817 signal (sig[i], abort_gzip_signal);
1818 siginterrupt (sig[i], 1);
1823 /* ========================================================================
1824 * Free all dynamically allocated variables and exit with the given code.
1826 local void do_exit(exitcode)
1829 static int in_exit = 0;
1831 if (in_exit) exit(exitcode);
1833 if (env != NULL) free(env), env = NULL;
1834 if (args != NULL) free((char*)args), args = NULL;
1848 /* ========================================================================
1849 * Close and unlink the output file.
1852 remove_output_file ()
1857 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1858 fd = remove_ofname_fd;
1861 remove_ofname_fd = -1;
1865 sigprocmask (SIG_SETMASK, &oldset, NULL);
1868 /* ========================================================================
1874 remove_output_file ();
1878 /* ========================================================================
1882 abort_gzip_signal (sig)
1886 signal (sig, SIG_IGN);
1887 remove_output_file ();
1888 if (sig == exiting_signal)
1890 signal (sig, SIG_DFL);