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[] = "$Id$";
62 #include <sys/types.h>
72 #include "fcntl-safer.h"
74 #include "stat-time.h"
81 # include <sys/time.h>
92 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
103 # ifndef _D_EXACT_NAMLEN
104 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
109 # define CLOSEDIR(d) (closedir(d), 0)
111 # define CLOSEDIR(d) closedir(d)
115 # include <utimens.h>
118 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
121 # define MAX_PATH_LEN 1024 /* max pathname length */
133 off_t lseek OF((int fd, off_t offset, int whence));
137 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
141 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
144 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
147 # define SA_NOCLDSTOP 0
148 # define sigprocmask(how, set, oset) /* empty */
149 # define sigset_t int
150 # if ! HAVE_SIGINTERRUPT
151 # define siginterrupt(sig, flag) /* empty */
155 #ifndef HAVE_WORKING_O_NOFOLLOW
156 # define HAVE_WORKING_O_NOFOLLOW 0
160 # define ELOOP EINVAL
163 /* Separator for file name parts (see shorten_name()) */
164 #ifdef NO_MULTIPLE_DOTS
165 # define PART_SEP "-"
167 # define PART_SEP "."
172 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
173 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
174 DECLARE(ush, d_buf, DIST_BUFSIZE);
175 DECLARE(uch, window, 2L*WSIZE);
177 DECLARE(ush, tab_prefix, 1L<<BITS);
179 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
180 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
183 /* local variables */
185 int ascii = 0; /* convert end-of-lines to local OS conventions */
186 int to_stdout = 0; /* output to stdout (-c) */
187 int decompress = 0; /* decompress (-d) */
188 int force = 0; /* don't ask questions, compress links (-f) */
189 int no_name = -1; /* don't save or restore the original file name */
190 int no_time = -1; /* don't save or restore the original file time */
191 int recursive = 0; /* recurse through directories (-r) */
192 int list = 0; /* list the file contents (-l) */
193 int verbose = 0; /* be verbose (-v) */
194 int quiet = 0; /* be very quiet (-q) */
195 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
196 int test = 0; /* test .gz file integrity */
197 int foreground; /* set if program run in foreground */
198 char *program_name; /* program name */
199 int maxbits = BITS; /* max bits per code for LZW */
200 int method = DEFLATED;/* compression method */
201 int level = 6; /* compression level */
202 int exit_code = OK; /* program exit code */
203 int save_orig_name; /* set if original name must be saved */
204 int last_member; /* set for .zip and .Z files */
205 int part_nb; /* number of parts in .gz file */
206 struct timespec time_stamp; /* original time stamp (modification time) */
207 off_t ifile_size; /* input file size, -1 for devices (debug only) */
208 char *env; /* contents of GZIP env variable */
209 char **args = NULL; /* argv pointer if GZIP env variable defined */
210 char *z_suffix; /* default suffix (can be set with --suffix) */
211 size_t z_len; /* strlen(z_suffix) */
213 /* The set of signals that are caught. */
214 static sigset_t caught_signals;
216 /* If nonzero then exit with status 1, rather than with the usual
217 signal status, on receipt of a signal with this value. This
218 suppresses a "Broken Pipe" message with some shells. */
219 static int volatile exiting_signal;
221 /* If nonnegative, close this file descriptor and unlink ofname on error. */
222 static int volatile remove_ofname_fd = -1;
224 off_t bytes_in; /* number of input bytes */
225 off_t bytes_out; /* number of output bytes */
226 off_t total_in; /* input bytes for all files */
227 off_t total_out; /* output bytes for all files */
228 char ifname[MAX_PATH_LEN]; /* input file name */
229 char ofname[MAX_PATH_LEN]; /* output file name */
230 struct stat istat; /* status for input file */
231 int ifd; /* input file descriptor */
232 int ofd; /* output file descriptor */
233 unsigned insize; /* valid bytes in inbuf */
234 unsigned inptr; /* index of next byte to be processed in inbuf */
235 unsigned outcnt; /* bytes in output buffer */
237 struct option longopts[] =
239 /* { name has_arg *flag val } */
240 {"ascii", 0, 0, 'a'}, /* ascii text mode */
241 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
242 {"stdout", 0, 0, 'c'}, /* write output on standard output */
243 {"decompress", 0, 0, 'd'}, /* decompress */
244 {"uncompress", 0, 0, 'd'}, /* decompress */
245 /* {"encrypt", 0, 0, 'e'}, encrypt */
246 {"force", 0, 0, 'f'}, /* force overwrite of output file */
247 {"help", 0, 0, 'h'}, /* give help */
248 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
249 {"list", 0, 0, 'l'}, /* list .gz file contents */
250 {"license", 0, 0, 'L'}, /* display software license */
251 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
252 {"name", 0, 0, 'N'}, /* save or restore original name & time */
253 {"quiet", 0, 0, 'q'}, /* quiet mode */
254 {"silent", 0, 0, 'q'}, /* quiet mode */
255 {"recursive", 0, 0, 'r'}, /* recurse through directories */
256 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
257 {"test", 0, 0, 't'}, /* test compressed file integrity */
258 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
259 {"verbose", 0, 0, 'v'}, /* verbose mode */
260 {"version", 0, 0, 'V'}, /* display version number */
261 {"fast", 0, 0, '1'}, /* compress faster */
262 {"best", 0, 0, '9'}, /* compress better */
263 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
264 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
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)",
347 "With no FILE, or when FILE is -, read standard input.",
349 "Report bugs to <bug-gzip@gnu.org>.",
353 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
354 while (*p) printf ("%s\n", *p++);
357 /* ======================================================================== */
360 char **p = license_msg;
362 printf ("%s %s\n", program_name, VERSION);
363 while (*p) printf ("%s\n", *p++);
366 /* ======================================================================== */
371 printf ("Written by Jean-loup Gailly.\n");
374 local void progerror (string)
378 fprintf (stderr, "%s: ", program_name);
384 /* ======================================================================== */
385 int main (argc, argv)
389 int file_count; /* number of files to process */
390 size_t proglen; /* length of program_name */
391 int optc; /* current option */
393 EXPAND(argc, argv); /* wild card expansion if necessary */
395 program_name = gzip_base_name (argv[0]);
396 proglen = strlen (program_name);
398 /* Suppress .exe for MSDOS, OS/2 and VMS: */
399 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
400 program_name[proglen - 4] = '\0';
402 /* Add options in GZIP environment variable if there is one */
403 env = add_envopt(&argc, &argv, OPTIONS_VAR);
404 if (env != NULL) args = argv;
407 /* For compatibility with old compress, use program name as an option.
408 * If you compile with -DGNU_STANDARD, this program will behave as
409 * gzip even if it is invoked under the name gunzip or zcat.
411 * Systems which do not support links can still use -d or -dc.
412 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
414 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
415 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
417 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
418 || strequ (program_name, "gzcat")) /* gzcat */
419 decompress = to_stdout = 1;
423 z_len = strlen(z_suffix);
425 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
426 longopts, (int *)0)) != -1) {
431 maxbits = atoi(optarg);
432 for (; *optarg; optarg++)
433 if (! ('0' <= *optarg && *optarg <= '9'))
435 fprintf (stderr, "%s: -b operand is not an integer\n",
441 to_stdout = 1; break;
443 decompress = 1; break;
447 help(); do_exit(OK); break;
449 list = decompress = to_stdout = 1; break;
451 license(); do_exit(OK); break;
452 case 'm': /* undocumented, may change later */
454 case 'M': /* undocumented, may change later */
457 no_name = no_time = 1; break;
459 no_name = no_time = 0; break;
461 quiet = 1; verbose = 0; break;
464 fprintf (stderr, "%s: -r not supported on this system\n",
472 #ifdef NO_MULTIPLE_DOTS
473 if (*optarg == '.') optarg++;
475 z_len = strlen(optarg);
479 test = decompress = to_stdout = 1;
482 verbose++; quiet = 0; break;
484 version(); do_exit(OK); break;
489 fprintf(stderr, "%s: -Z not supported in this version\n",
494 case '1': case '2': case '3': case '4':
495 case '5': case '6': case '7': case '8': case '9':
499 /* Error message already emitted by getopt_long. */
502 } /* loop on all arguments */
504 /* By default, save name and timestamp on compression but do not
505 * restore them on decompression.
507 if (no_time < 0) no_time = decompress;
508 if (no_name < 0) no_name = decompress;
510 file_count = argc - optind;
514 if (ascii && !quiet) {
515 fprintf(stderr, "%s: option --ascii ignored on this system\n",
519 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
520 fprintf(stderr, "%s: incorrect suffix '%s'\n",
521 program_name, z_suffix);
524 if (do_lzw && !decompress) work = lzw;
526 /* Allocate all global buffers (for DYN_ALLOC option) */
527 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
528 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
529 ALLOC(ush, d_buf, DIST_BUFSIZE);
530 ALLOC(uch, window, 2L*WSIZE);
532 ALLOC(ush, tab_prefix, 1L<<BITS);
534 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
535 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
538 exiting_signal = quiet ? SIGPIPE : 0;
539 install_signal_handlers ();
541 /* And get to work */
542 if (file_count != 0) {
543 if (to_stdout && !test && !list && (!decompress || !ascii)) {
544 SET_BINARY_MODE(fileno(stdout));
546 while (optind < argc) {
547 treat_file(argv[optind++]);
549 } else { /* Standard input */
552 if (list && !quiet && file_count > 1) {
553 do_list(-1, -1); /* print totals */
556 return exit_code; /* just to avoid lint warning */
559 /* Return nonzero when at end of file on input. */
563 if (!decompress || last_member)
568 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
571 /* Unget the char that fill_inbuf got. */
578 /* ========================================================================
579 * Compress or decompress stdin
581 local void treat_stdin()
583 if (!force && !list &&
584 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
585 /* Do not send compressed data to the terminal or read it from
586 * the terminal. We get here when user invoked the program
587 * without parameters, so be helpful. According to the GNU standards:
589 * If there is one behavior you think is most useful when the output
590 * is to a terminal, and another that you think is most useful when
591 * the output is a file or a pipe, then it is usually best to make
592 * the default behavior the one that is useful with output to a
593 * terminal, and have an option for the other behavior.
595 * Here we use the --force option to get the other behavior.
598 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
599 program_name, decompress ? "read from" : "written to",
600 decompress ? "de" : "");
601 fprintf (stderr, "For help, type: %s -h\n", program_name);
605 if (decompress || !ascii) {
606 SET_BINARY_MODE(fileno(stdin));
608 if (!test && !list && (!decompress || !ascii)) {
609 SET_BINARY_MODE(fileno(stdout));
611 strcpy(ifname, "stdin");
612 strcpy(ofname, "stdout");
614 /* Get the time stamp on the input file. */
615 time_stamp.tv_nsec = -1; /* The time is unknown by default. */
617 #ifndef NO_STDIN_FSTAT
618 if (list || !no_time) {
619 if (fstat(fileno(stdin), &istat) != 0) {
620 progerror("standard input");
623 # ifdef NO_PIPE_TIMESTAMP
624 if (S_ISREG(istat.st_mode))
626 time_stamp = get_stat_mtime (&istat);
627 #endif /* NO_STDIN_FSTAT */
629 ifile_size = -1L; /* convention for unknown size */
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));
708 if (!S_ISREG(istat.st_mode)) {
710 "%s: %s is not a directory or a regular file - ignored\n",
711 program_name, ifname));
716 if (istat.st_mode & S_ISUID)
718 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
719 program_name, ifname));
723 if (istat.st_mode & S_ISGID)
725 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
726 program_name, ifname));
730 if (istat.st_mode & S_ISVTX)
732 WARN ((stderr, "%s: %s has the sticky bit set - file ignored\n",
733 program_name, ifname));
738 if (istat.st_nlink > 1 && !to_stdout && !force) {
739 WARN((stderr, "%s: %s has %lu other link%c -- unchanged\n",
740 program_name, ifname, (unsigned long) istat.st_nlink - 1,
741 istat.st_nlink > 2 ? 's' : ' '));
746 ifile_size = istat.st_size;
747 if (no_time && !list)
748 time_stamp.tv_nsec = -1;
750 time_stamp = get_stat_mtime (&istat);
752 /* Generate output file name. For -r and (-t or -l), skip files
753 * without a valid gzip suffix (check done in make_ofname).
755 if (to_stdout && !list && !test) {
756 strcpy(ofname, "stdout");
758 } else if (make_ofname() != OK) {
763 clear_bufs(); /* clear input and output buffers */
767 method = get_method(ifd); /* updates ofname if original given */
770 return; /* error message already emitted */
774 do_list(ifd, method);
775 if (close (ifd) != 0)
780 /* If compressing to a file, check if ofname is not ambiguous
781 * because the operating system truncates names. Otherwise, generate
782 * a new ofname and save the original name in the compressed file.
785 ofd = fileno(stdout);
786 /* Keep remove_ofname_fd negative. */
788 if (create_outfile() != OK) return;
790 if (!decompress && save_orig_name && !verbose && !quiet) {
791 fprintf(stderr, "%s: %s compressed to %s\n",
792 program_name, ifname, ofname);
795 /* Keep the name even if not truncated except with --no-name: */
796 if (!save_orig_name) save_orig_name = !no_name;
799 fprintf(stderr, "%s:\t", ifname);
802 /* Actually do the compression/decompression. Loop over zipped members.
805 if ((*work)(ifd, ofd) != OK) {
806 method = -1; /* force cleanup */
813 method = get_method(ifd);
814 if (method < 0) break; /* error message already emitted */
815 bytes_out = 0; /* required for length check */
818 if (close (ifd) != 0)
827 if (close (ofd) != 0)
830 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
831 remove_ofname_fd = -1;
832 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
833 sigprocmask (SIG_SETMASK, &oldset, NULL);
837 WARN ((stderr, "%s: ", program_name));
840 errno = unlink_errno;
848 remove_output_file ();
852 /* Display statistics */
855 fprintf(stderr, " OK");
856 } else if (decompress) {
857 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
859 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
861 if (!test && !to_stdout) {
862 fprintf(stderr, " -- replaced with %s", ofname);
864 fprintf(stderr, "\n");
868 /* ========================================================================
869 * Create the output file. Return OK or ERROR.
870 * Try several times if necessary to avoid truncating the z_suffix. For
871 * example, do not create a compressed file of name "1234567890123."
872 * Sets save_orig_name to true if the file name has been truncated.
873 * IN assertions: the input file has already been open (ifd is set) and
874 * ofname has already been updated if there was an original name.
875 * OUT assertions: ifd and ofd are closed in case of error.
877 local int create_outfile()
879 int name_shortened = 0;
880 int flags = (O_WRONLY | O_CREAT | O_EXCL
881 | (ascii && decompress ? 0 : O_BINARY));
888 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
889 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
891 sigprocmask (SIG_SETMASK, &oldset, NULL);
900 shorten_name (ofname);
906 if (check_ofname () != OK)
920 if (name_shortened && decompress)
922 /* name might be too long if an original name was saved */
923 WARN ((stderr, "%s: %s: warning, name truncated\n",
924 program_name, ofname));
930 /* ========================================================================
931 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
932 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
933 * accepted suffixes, in addition to the value of the --suffix option.
934 * ".tgz" is a useful convention for tar.z files on systems limited
935 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
936 * also accepted suffixes. For Unix, we do not want to accept any
937 * .??z suffix as indicating a compressed file; some people use .xyz
938 * to denote volume data.
939 * On systems allowing multiple versions of the same file (such as VMS),
940 * this function removes any version suffix in the given name.
942 local char *get_suffix(name)
946 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
947 static char *known_suffixes[] =
948 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
953 char **suf = known_suffixes;
956 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
959 /* strip a version number from the file name */
961 char *v = strrchr(name, SUFFIX_SEP);
962 if (v != NULL) *v = '\0';
966 if (nlen <= MAX_SUFFIX+2) {
967 strcpy(suffix, name);
969 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
972 slen = strlen(suffix);
974 int s = strlen(*suf);
975 if (slen > s && suffix[slen-s-1] != PATH_SEP
976 && strequ(suffix + slen - s, *suf)) {
979 } while (*++suf != NULL);
985 /* Open file NAME with the given flags and mode and store its status
986 into *ST. Return a file descriptor to the newly opened file, or -1
987 (setting errno) on failure. */
989 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
993 /* Refuse to follow symbolic links unless -c or -f. */
994 if (!to_stdout && !force)
996 if (HAVE_WORKING_O_NOFOLLOW)
1000 #if HAVE_LSTAT || defined lstat
1001 if (lstat (name, st) != 0)
1003 else if (S_ISLNK (st->st_mode))
1012 fd = OPEN (name, flags, mode);
1013 if (0 <= fd && fstat (fd, st) != 0)
1024 /* ========================================================================
1025 * Set ifname to the input file name (with a suffix appended if necessary)
1026 * and istat to its stats. For decompression, if no file exists with the
1027 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1028 * For MSDOS, we try only z_suffix and z.
1029 * Return an open file descriptor or -1.
1032 open_input_file (iname, sbuf)
1036 int ilen; /* strlen(ifname) */
1037 int z_suffix_errno = 0;
1038 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1039 char **suf = suffixes;
1041 #ifdef NO_MULTIPLE_DOTS
1042 char *dot; /* pointer to ifname extension, or NULL */
1045 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1046 | (ascii && !decompress ? 0 : O_BINARY));
1050 if (sizeof ifname - 1 <= strlen (iname))
1053 strcpy(ifname, iname);
1055 /* If input file exists, return OK. */
1056 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1060 if (!decompress || errno != ENOENT) {
1064 /* file.ext doesn't exist, try adding a suffix (after removing any
1065 * version number for VMS).
1067 s = get_suffix(ifname);
1069 progerror(ifname); /* ifname already has z suffix and does not exist */
1072 #ifdef NO_MULTIPLE_DOTS
1073 dot = strrchr(ifname, '.');
1075 strcat(ifname, ".");
1076 dot = strrchr(ifname, '.');
1079 ilen = strlen(ifname);
1080 if (strequ(z_suffix, ".gz")) suf++;
1082 /* Search for all suffixes */
1084 char *s0 = s = *suf;
1085 strcpy (ifname, iname);
1086 #ifdef NO_MULTIPLE_DOTS
1088 if (*dot == '\0') strcpy (dot, ".");
1090 #ifdef MAX_EXT_CHARS
1091 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1092 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1094 if (sizeof ifname <= ilen + strlen (s))
1097 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1100 if (errno != ENOENT)
1105 if (strequ (s0, z_suffix))
1106 z_suffix_errno = errno;
1107 } while (*++suf != NULL);
1109 /* No suffix found, complain using z_suffix: */
1110 strcpy(ifname, iname);
1111 #ifdef NO_MULTIPLE_DOTS
1112 if (*dot == '\0') strcpy(dot, ".");
1114 #ifdef MAX_EXT_CHARS
1115 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1116 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1118 strcat(ifname, z_suffix);
1119 errno = z_suffix_errno;
1124 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1129 /* ========================================================================
1130 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1131 * Sets save_orig_name to true if the file name has been truncated.
1133 local int make_ofname()
1135 char *suff; /* ofname z suffix */
1137 strcpy(ofname, ifname);
1138 /* strip a version number if any and get the gzip suffix if present: */
1139 suff = get_suffix(ofname);
1143 /* With -t or -l, try all files (even without .gz suffix)
1144 * except with -r (behave as with just -dr).
1146 if (!recursive && (list || test)) return OK;
1148 /* Avoid annoying messages with -r */
1149 if (verbose || (!recursive && !quiet)) {
1150 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1151 program_name, ifname));
1155 /* Make a special case for .tgz and .taz: */
1157 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1158 strcpy(suff, ".tar");
1160 *suff = '\0'; /* strip the z suffix */
1162 /* ofname might be changed later if infile contains an original name */
1164 } else if (suff != NULL) {
1165 /* Avoid annoying messages with -r (see treat_dir()) */
1166 if (verbose || (!recursive && !quiet)) {
1167 /* Don't use WARN, as it affects exit status. */
1168 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1169 program_name, ifname, suff);
1175 #ifdef NO_MULTIPLE_DOTS
1176 suff = strrchr(ofname, '.');
1178 if (sizeof ofname <= strlen (ofname) + 1)
1180 strcat(ofname, ".");
1181 # ifdef MAX_EXT_CHARS
1182 if (strequ(z_suffix, "z")) {
1183 if (sizeof ofname <= strlen (ofname) + 2)
1185 strcat(ofname, "gz"); /* enough room */
1188 /* On the Atari and some versions of MSDOS,
1189 * ENAMETOOLONG does not work correctly. So we
1190 * must truncate here.
1192 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1193 suff[MAX_SUFFIX+1-z_len] = '\0';
1197 #endif /* NO_MULTIPLE_DOTS */
1198 if (sizeof ofname <= strlen (ofname) + z_len)
1200 strcat(ofname, z_suffix);
1202 } /* decompress ? */
1206 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1211 /* ========================================================================
1212 * Check the magic number of the input file and update ofname if an
1213 * original name was given and to_stdout is not set.
1214 * Return the compression method, -1 for error, -2 for warning.
1215 * Set inptr to the offset of the next byte to be processed.
1216 * Updates time_stamp if there is one and --no-time is not used.
1217 * This function may be called repeatedly for an input file consisting
1218 * of several contiguous gzip'ed members.
1219 * IN assertions: there is at least one remaining compressed member.
1220 * If the member is a zip file, it must be the only one.
1222 local int get_method(in)
1223 int in; /* input file descriptor */
1225 uch flags; /* compression flags */
1226 char magic[2]; /* magic header */
1227 int imagic1; /* like magic[1], but can represent EOF */
1228 ulg stamp; /* time stamp */
1230 /* If --force and --stdout, zcat == cat, so do not complain about
1231 * premature end of file: use try_byte instead of get_byte.
1233 if (force && to_stdout) {
1234 magic[0] = (char)try_byte();
1235 imagic1 = try_byte ();
1236 magic[1] = (char) imagic1;
1237 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1239 magic[0] = (char)get_byte();
1240 magic[1] = (char)get_byte();
1241 imagic1 = 0; /* avoid lint warning */
1243 method = -1; /* unknown yet */
1244 part_nb++; /* number of parts in gzip file */
1246 last_member = RECORD_IO;
1247 /* assume multiple members in gzip file except for record oriented I/O */
1249 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1250 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1252 method = (int)get_byte();
1253 if (method != DEFLATED) {
1255 "%s: %s: unknown method %d -- not supported\n",
1256 program_name, ifname, method);
1261 flags = (uch)get_byte();
1263 if ((flags & ENCRYPTED) != 0) {
1265 "%s: %s is encrypted -- not supported\n",
1266 program_name, ifname);
1270 if ((flags & CONTINUATION) != 0) {
1272 "%s: %s is a a multi-part gzip file -- not supported\n",
1273 program_name, ifname);
1275 if (force <= 1) return -1;
1277 if ((flags & RESERVED) != 0) {
1279 "%s: %s has flags 0x%x -- not supported\n",
1280 program_name, ifname, flags);
1282 if (force <= 1) return -1;
1284 stamp = (ulg)get_byte();
1285 stamp |= ((ulg)get_byte()) << 8;
1286 stamp |= ((ulg)get_byte()) << 16;
1287 stamp |= ((ulg)get_byte()) << 24;
1288 if (stamp != 0 && !no_time)
1290 time_stamp.tv_sec = stamp;
1291 time_stamp.tv_nsec = 0;
1294 (void)get_byte(); /* Ignore extra flags for the moment */
1295 (void)get_byte(); /* Ignore OS type for the moment */
1297 if ((flags & CONTINUATION) != 0) {
1298 unsigned part = (unsigned)get_byte();
1299 part |= ((unsigned)get_byte())<<8;
1301 fprintf(stderr,"%s: %s: part number %u\n",
1302 program_name, ifname, part);
1305 if ((flags & EXTRA_FIELD) != 0) {
1306 unsigned len = (unsigned)get_byte();
1307 len |= ((unsigned)get_byte())<<8;
1309 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1310 program_name, ifname, len);
1312 while (len--) (void)get_byte();
1315 /* Get original file name if it was truncated */
1316 if ((flags & ORIG_NAME) != 0) {
1317 if (no_name || (to_stdout && !list) || part_nb > 1) {
1318 /* Discard the old name */
1319 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1320 do {c=get_byte();} while (c != 0);
1322 /* Copy the base name. Keep a directory prefix intact. */
1323 char *p = gzip_base_name (ofname);
1326 *p = (char)get_char();
1327 if (*p++ == '\0') break;
1328 if (p >= ofname+sizeof(ofname)) {
1329 gzip_error ("corrupted input -- file name too large");
1332 p = gzip_base_name (base);
1333 memmove (base, p, strlen (p) + 1);
1334 /* If necessary, adapt the name to local OS conventions: */
1336 MAKE_LEGAL_NAME(base);
1337 if (base) list=0; /* avoid warning about unused variable */
1339 } /* no_name || to_stdout */
1342 /* Discard file comment if any */
1343 if ((flags & COMMENT) != 0) {
1344 while (get_char() != 0) /* null */ ;
1347 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1350 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1351 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1352 /* To simplify the code, we support a zip file when alone only.
1353 * We are thus guaranteed that the entire local header fits in inbuf.
1357 if (check_zipfile(in) != OK) return -1;
1358 /* check_zipfile may get ofname from the local header */
1361 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1365 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1367 method = COMPRESSED;
1370 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1375 } else if (force && to_stdout && !list) { /* pass input unchanged */
1381 if (method >= 0) return method;
1384 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1385 program_name, ifname);
1392 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1397 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1398 program_name, ifname));
1403 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1404 program_name, ifname));
1409 /* ========================================================================
1410 * Display the characteristics of the compressed file.
1411 * If the given method is < 0, display the accumulated totals.
1412 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1414 local void do_list(ifd, method)
1415 int ifd; /* input file descriptor */
1416 int method; /* compression method */
1418 ulg crc; /* original crc */
1419 static int first_time = 1;
1420 static char* methods[MAX_METHODS] = {
1425 "", "", "", "", /* 4 to 7 reserved */
1427 int positive_off_t_width = 1;
1430 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1431 positive_off_t_width++;
1434 if (first_time && method >= 0) {
1437 printf("method crc date time ");
1440 printf("%*.*s %*.*s ratio uncompressed_name\n",
1441 positive_off_t_width, positive_off_t_width, "compressed",
1442 positive_off_t_width, positive_off_t_width, "uncompressed");
1444 } else if (method < 0) {
1445 if (total_in <= 0 || total_out <= 0) return;
1449 if (verbose || !quiet) {
1450 fprint_off(stdout, total_in, positive_off_t_width);
1452 fprint_off(stdout, total_out, positive_off_t_width);
1455 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1456 /* header_bytes is not meaningful but used to ensure the same
1457 * ratio if there is a single file.
1459 printf(" (totals)\n");
1462 crc = (ulg)~0; /* unknown */
1464 bytes_in = ifile_size;
1467 if (method == DEFLATED && !last_member) {
1468 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1469 * If the lseek fails, we could use read() to get to the end, but
1470 * --list is used to get quick results.
1471 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1472 * you are not concerned about speed.
1474 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1475 if (bytes_in != -1L) {
1478 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1482 bytes_out = LG(buf+4);
1485 #endif /* RECORD_IO */
1488 struct tm *tm = localtime (&time_stamp.tv_sec);
1489 printf ("%5s %08lx ", methods[method], crc);
1491 printf ("%s%3d %02d:%02d ",
1492 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1494 tm->tm_mday, tm->tm_hour, tm->tm_min);
1496 printf ("??? ?? ??:?? ");
1498 fprint_off(stdout, bytes_in, positive_off_t_width);
1500 fprint_off(stdout, bytes_out, positive_off_t_width);
1502 if (bytes_in == -1L) {
1504 bytes_in = bytes_out = header_bytes = 0;
1505 } else if (total_in >= 0) {
1506 total_in += bytes_in;
1508 if (bytes_out == -1L) {
1510 bytes_in = bytes_out = header_bytes = 0;
1511 } else if (total_out >= 0) {
1512 total_out += bytes_out;
1514 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1515 printf(" %s\n", ofname);
1518 /* ========================================================================
1519 * Shorten the given name by one character, or replace a .tar extension
1520 * with .tgz. Truncate the last part of the name which is longer than
1521 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1522 * has only parts shorter than MIN_PART truncate the longest part.
1523 * For decompression, just remove the last character of the name.
1525 * IN assertion: for compression, the suffix of the given name is z_suffix.
1527 local void shorten_name(name)
1530 int len; /* length of name without z_suffix */
1531 char *trunc = NULL; /* character to be truncated */
1532 int plen; /* current part length */
1533 int min_part = MIN_PART; /* current minimum part length */
1539 gzip_error ("name too short");
1543 p = get_suffix(name);
1545 gzip_error ("can't recover suffix\n");
1549 /* compress 1234567890.tar to 1234567890.tgz */
1550 if (len > 4 && strequ(p-4, ".tar")) {
1551 strcpy(p-4, ".tgz");
1554 /* Try keeping short extensions intact:
1555 * 1234.678.012.gz -> 123.678.012.gz
1558 p = strrchr(name, PATH_SEP);
1561 plen = strcspn(p, PART_SEP);
1563 if (plen > min_part) trunc = p-1;
1566 } while (trunc == NULL && --min_part != 0);
1568 if (trunc != NULL) {
1570 trunc[0] = trunc[1];
1574 trunc = strrchr(name, PART_SEP[0]);
1576 gzip_error ("internal error in shorten_name");
1577 if (trunc[1] == '\0') trunc--; /* force truncation */
1579 strcpy(trunc, z_suffix);
1582 /* ========================================================================
1583 * The compressed file already exists, so ask for confirmation.
1584 * Return ERROR if the file must be skipped.
1586 local int check_ofname()
1588 /* Ask permission to overwrite the existing file */
1591 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1592 if (foreground && isatty(fileno(stdin))) {
1593 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1598 fprintf(stderr, "\tnot overwritten\n");
1599 if (exit_code == OK) exit_code = WARNING;
1603 if (xunlink (ofname)) {
1611 /* ========================================================================
1612 * Copy modes, times, ownership from input file to output file.
1613 * IN assertion: to_stdout is false.
1615 local void copy_stat(ifstat)
1616 struct stat *ifstat;
1618 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1622 struct timespec timespec[2];
1623 timespec[0] = get_stat_atime (ifstat);
1624 timespec[1] = get_stat_mtime (ifstat);
1626 if (decompress && 0 <= time_stamp.tv_nsec
1627 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1628 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1630 timespec[1] = time_stamp;
1632 fprintf(stderr, "%s: time stamp restored\n", ofname);
1636 if (futimens (ofd, ofname, timespec) != 0)
1639 WARN ((stderr, "%s: ", program_name));
1650 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1652 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1656 /* Copy the protection modes */
1658 r = fchmod (ofd, mode);
1660 r = chmod (ofname, mode);
1664 WARN ((stderr, "%s: ", program_name));
1674 /* ========================================================================
1675 * Recurse through the given directory. This code is taken from ncompress.
1677 local void treat_dir (fd, dir)
1683 char nbuf[MAX_PATH_LEN];
1687 dirp = fdopendir (fd);
1690 dirp = opendir(dir);
1701 ** WARNING: the following algorithm could occasionally cause
1702 ** compress to produce error warnings of the form "<filename>.gz
1703 ** already has .gz suffix - ignored". This occurs when the
1704 ** .gz output file is inserted into the directory below
1705 ** readdir's current pointer.
1706 ** These warnings are harmless but annoying, so they are suppressed
1707 ** with option -r (except when -v is on). An alternative
1708 ** to allowing this would be to store the entire directory
1709 ** list in memory, then compress the entries in the stored
1710 ** list. Given the depth-first recursive algorithm used here,
1711 ** this could use up a tremendous amount of memory. I don't
1712 ** think it's worth it. -- Dave Mack
1713 ** (An other alternative might be two passes to avoid depth-first.)
1716 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1718 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1722 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1724 if (len != 0 /* dir = "" means current dir on Amiga */
1726 && dir[len-1] != PATH_SEP2
1729 && dir[len-1] != PATH_SEP3
1732 nbuf[len++] = PATH_SEP;
1734 strcpy(nbuf+len, dp->d_name);
1737 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1738 program_name, dir, dp->d_name);
1744 if (CLOSEDIR(dirp) != 0)
1747 #endif /* ! NO_DIR */
1749 /* Make sure signals get handled properly. */
1752 install_signal_handlers ()
1756 /* SIGINT must be first, as 'foreground' depends on it. */
1777 int nsigs = sizeof sig / sizeof sig[0];
1781 struct sigaction act;
1783 sigemptyset (&caught_signals);
1784 for (i = 0; i < nsigs; i++)
1786 sigaction (sig[i], NULL, &act);
1787 if (act.sa_handler != SIG_IGN)
1788 sigaddset (&caught_signals, sig[i]);
1791 act.sa_handler = abort_gzip_signal;
1792 act.sa_mask = caught_signals;
1795 for (i = 0; i < nsigs; i++)
1796 if (sigismember (&caught_signals, sig[i]))
1800 sigaction (sig[i], &act, NULL);
1803 for (i = 0; i < nsigs; i++)
1804 if (signal (sig[i], SIG_IGN) != SIG_IGN)
1808 signal (sig[i], abort_gzip_signal);
1809 siginterrupt (sig[i], 1);
1814 /* ========================================================================
1815 * Free all dynamically allocated variables and exit with the given code.
1817 local void do_exit(exitcode)
1820 static int in_exit = 0;
1822 if (in_exit) exit(exitcode);
1824 if (env != NULL) free(env), env = NULL;
1825 if (args != NULL) free((char*)args), args = NULL;
1839 /* ========================================================================
1840 * Close and unlink the output file.
1843 remove_output_file ()
1848 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1849 fd = remove_ofname_fd;
1852 remove_ofname_fd = -1;
1856 sigprocmask (SIG_SETMASK, &oldset, NULL);
1859 /* ========================================================================
1865 remove_output_file ();
1869 /* ========================================================================
1873 abort_gzip_signal (sig)
1877 signal (sig, SIG_IGN);
1878 remove_output_file ();
1879 if (sig == exiting_signal)
1881 signal (sig, SIG_DFL);