1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001-2002, 2006-2007, 2009-2010 Free Software
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 "ignore-value.h"
73 #include "stat-time.h"
88 # ifndef _D_EXACT_NAMLEN
89 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
94 # define CLOSEDIR(d) (closedir(d), 0)
96 # define CLOSEDIR(d) closedir(d)
100 # include <utimens.h>
103 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
106 # define MAX_PATH_LEN 1024 /* max pathname length */
118 off_t lseek OF((int fd, off_t offset, int whence));
122 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
126 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
129 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
132 # define SA_NOCLDSTOP 0
133 # define sigprocmask(how, set, oset) /* empty */
134 # define sigset_t int
135 # if ! HAVE_SIGINTERRUPT
136 # define siginterrupt(sig, flag) /* empty */
140 #ifndef HAVE_WORKING_O_NOFOLLOW
141 # define HAVE_WORKING_O_NOFOLLOW 0
145 # define ELOOP EINVAL
148 /* Separator for file name parts (see shorten_name()) */
149 #ifdef NO_MULTIPLE_DOTS
150 # define PART_SEP "-"
152 # define PART_SEP "."
157 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
158 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
159 DECLARE(ush, d_buf, DIST_BUFSIZE);
160 DECLARE(uch, window, 2L*WSIZE);
162 DECLARE(ush, tab_prefix, 1L<<BITS);
164 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
165 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
168 /* local variables */
170 int ascii = 0; /* convert end-of-lines to local OS conventions */
171 int to_stdout = 0; /* output to stdout (-c) */
172 int decompress = 0; /* decompress (-d) */
173 int force = 0; /* don't ask questions, compress links (-f) */
174 int no_name = -1; /* don't save or restore the original file name */
175 int no_time = -1; /* don't save or restore the original file time */
176 int recursive = 0; /* recurse through directories (-r) */
177 int list = 0; /* list the file contents (-l) */
178 int verbose = 0; /* be verbose (-v) */
179 int quiet = 0; /* be very quiet (-q) */
180 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
181 int test = 0; /* test .gz file integrity */
182 int foreground = 0; /* set if program run in foreground */
183 char *program_name; /* program name */
184 int maxbits = BITS; /* max bits per code for LZW */
185 int method = DEFLATED;/* compression method */
186 int level = 6; /* compression level */
187 int exit_code = OK; /* program exit code */
188 int save_orig_name; /* set if original name must be saved */
189 int last_member; /* set for .zip and .Z files */
190 int part_nb; /* number of parts in .gz file */
191 struct timespec time_stamp; /* original time stamp (modification time) */
192 off_t ifile_size; /* input file size, -1 for devices (debug only) */
193 char *env; /* contents of GZIP env variable */
194 char **args = NULL; /* argv pointer if GZIP env variable defined */
195 char const *z_suffix; /* default suffix (can be set with --suffix) */
196 size_t z_len; /* strlen(z_suffix) */
198 /* The set of signals that are caught. */
199 static sigset_t caught_signals;
201 /* If nonzero then exit with status WARNING, rather than with the usual
202 signal status, on receipt of a signal with this value. This
203 suppresses a "Broken Pipe" message with some shells. */
204 static int volatile exiting_signal;
206 /* If nonnegative, close this file descriptor and unlink ofname on error. */
207 static int volatile remove_ofname_fd = -1;
209 off_t bytes_in; /* number of input bytes */
210 off_t bytes_out; /* number of output bytes */
211 off_t total_in; /* input bytes for all files */
212 off_t total_out; /* output bytes for all files */
213 char ifname[MAX_PATH_LEN]; /* input file name */
214 char ofname[MAX_PATH_LEN]; /* output file name */
215 struct stat istat; /* status for input file */
216 int ifd; /* input file descriptor */
217 int ofd; /* output file descriptor */
218 unsigned insize; /* valid bytes in inbuf */
219 unsigned inptr; /* index of next byte to be processed in inbuf */
220 unsigned outcnt; /* bytes in output buffer */
221 int rsync = 0; /* make ryncable chunks */
223 static int handled_sig[] =
225 /* SIGINT must be first, as 'foreground' depends on it. */
247 static int handled_sig[] =
249 /* SIGINT must be first, as 'foreground' depends on it. */
271 struct option longopts[] =
273 /* { name has_arg *flag val } */
274 {"ascii", 0, 0, 'a'}, /* ascii text mode */
275 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
276 {"stdout", 0, 0, 'c'}, /* write output on standard output */
277 {"decompress", 0, 0, 'd'}, /* decompress */
278 {"uncompress", 0, 0, 'd'}, /* decompress */
279 /* {"encrypt", 0, 0, 'e'}, encrypt */
280 {"force", 0, 0, 'f'}, /* force overwrite of output file */
281 {"help", 0, 0, 'h'}, /* give help */
282 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
283 {"list", 0, 0, 'l'}, /* list .gz file contents */
284 {"license", 0, 0, 'L'}, /* display software license */
285 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
286 {"name", 0, 0, 'N'}, /* save or restore original name & time */
287 {"quiet", 0, 0, 'q'}, /* quiet mode */
288 {"silent", 0, 0, 'q'}, /* quiet mode */
289 {"recursive", 0, 0, 'r'}, /* recurse through directories */
290 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
291 {"test", 0, 0, 't'}, /* test compressed file integrity */
292 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
293 {"verbose", 0, 0, 'v'}, /* verbose mode */
294 {"version", 0, 0, 'V'}, /* display version number */
295 {"fast", 0, 0, '1'}, /* compress faster */
296 {"best", 0, 0, '9'}, /* compress better */
297 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
298 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
299 {"rsyncable", 0, 0, 'R'}, /* make rsync-friendly archive */
303 /* local functions */
305 local void try_help OF((void)) ATTRIBUTE_NORETURN;
306 local void help OF((void));
307 local void license OF((void));
308 local void version OF((void));
309 local int input_eof OF((void));
310 local void treat_stdin OF((void));
311 local void treat_file OF((char *iname));
312 local int create_outfile OF((void));
313 local char *get_suffix OF((char *name));
314 local int open_input_file OF((char *iname, struct stat *sbuf));
315 local int make_ofname OF((void));
316 local void shorten_name OF((char *name));
317 local int get_method OF((int in));
318 local void do_list OF((int ifd, int method));
319 local int check_ofname OF((void));
320 local void copy_stat OF((struct stat *ifstat));
321 local void install_signal_handlers OF((void));
322 local void remove_output_file OF((void));
323 local RETSIGTYPE abort_gzip_signal OF((int));
324 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
325 int main OF((int argc, char **argv));
326 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
329 local void treat_dir OF((int fd, char *dir));
332 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
337 fprintf (stderr, "Try `%s --help' for more information.\n",
342 /* ======================================================================== */
345 static char const* const help_msg[] = {
346 "Compress or uncompress FILEs (by default, compress FILES in-place).",
348 "Mandatory arguments to long options are mandatory for short options too.",
351 " -a, --ascii ascii text; convert end-of-line using local conventions",
353 " -c, --stdout write on standard output, keep original files unchanged",
354 " -d, --decompress decompress",
355 /* -e, --encrypt encrypt */
356 " -f, --force force overwrite of output file and compress links",
357 " -h, --help give this help",
358 /* -k, --pkzip force output in pkzip format */
359 " -l, --list list compressed file contents",
360 " -L, --license display software license",
362 " -m, --no-time do not save or restore the original modification time",
363 " -M, --time save or restore the original modification time",
365 " -n, --no-name do not save or restore the original name and time stamp",
366 " -N, --name save or restore the original name and time stamp",
367 " -q, --quiet suppress all warnings",
369 " -r, --recursive operate recursively on directories",
371 " -S, --suffix=SUF use suffix SUF on compressed files",
372 " -t, --test test compressed file integrity",
373 " -v, --verbose verbose mode",
374 " -V, --version display version number",
375 " -1, --fast compress faster",
376 " -9, --best compress better",
378 " -Z, --lzw produce output compatible with old compress",
379 " -b, --bits=BITS max number of bits per code (implies -Z)",
381 " --rsyncable Make rsync-friendly archive",
383 "With no FILE, or when FILE is -, read standard input.",
385 "Report bugs to <bug-gzip@gnu.org>.",
387 char const *const *p = help_msg;
389 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
390 while (*p) printf ("%s\n", *p++);
393 /* ======================================================================== */
396 char const *const *p = license_msg;
398 printf ("%s %s\n", program_name, VERSION);
399 while (*p) printf ("%s\n", *p++);
402 /* ======================================================================== */
407 printf ("Written by Jean-loup Gailly.\n");
410 local void progerror (char const *string)
413 fprintf (stderr, "%s: ", program_name);
419 /* ======================================================================== */
420 int main (int argc, char **argv)
422 int file_count; /* number of files to process */
423 size_t proglen; /* length of program_name */
424 int optc; /* current option */
426 EXPAND(argc, argv); /* wild card expansion if necessary */
428 program_name = gzip_base_name (argv[0]);
429 proglen = strlen (program_name);
431 atexit (close_stdin);
433 /* Suppress .exe for MSDOS, OS/2 and VMS: */
434 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
435 program_name[proglen - 4] = '\0';
437 /* Add options in GZIP environment variable if there is one */
438 env = add_envopt(&argc, &argv, OPTIONS_VAR);
439 if (env != NULL) args = argv;
442 # define GNU_STANDARD 1
445 /* For compatibility with old compress, use program name as an option.
446 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
447 * gzip even if it is invoked under the name gunzip or zcat.
449 * Systems which do not support links can still use -d or -dc.
450 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
452 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
453 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
455 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
456 || strequ (program_name, "gzcat")) /* gzcat */
457 decompress = to_stdout = 1;
461 z_len = strlen(z_suffix);
463 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
464 longopts, (int *)0)) != -1) {
469 maxbits = atoi(optarg);
470 for (; *optarg; optarg++)
471 if (! ('0' <= *optarg && *optarg <= '9'))
473 fprintf (stderr, "%s: -b operand is not an integer\n",
479 to_stdout = 1; break;
481 decompress = 1; break;
485 help(); do_exit(OK); break;
487 list = decompress = to_stdout = 1; break;
489 license(); do_exit(OK); break;
490 case 'm': /* undocumented, may change later */
492 case 'M': /* undocumented, may change later */
495 no_name = no_time = 1; break;
497 no_name = no_time = 0; break;
499 quiet = 1; verbose = 0; break;
502 fprintf (stderr, "%s: -r not supported on this system\n",
512 #ifdef NO_MULTIPLE_DOTS
513 if (*optarg == '.') optarg++;
515 z_len = strlen(optarg);
519 test = decompress = to_stdout = 1;
522 verbose++; quiet = 0; break;
524 version(); do_exit(OK); break;
529 fprintf(stderr, "%s: -Z not supported in this version\n",
534 case '1': case '2': case '3': case '4':
535 case '5': case '6': case '7': case '8': case '9':
539 /* Error message already emitted by getopt_long. */
542 } /* loop on all arguments */
544 /* By default, save name and timestamp on compression but do not
545 * restore them on decompression.
547 if (no_time < 0) no_time = decompress;
548 if (no_name < 0) no_name = decompress;
550 file_count = argc - optind;
554 if (ascii && !quiet) {
555 fprintf(stderr, "%s: option --ascii ignored on this system\n",
559 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
560 fprintf(stderr, "%s: incorrect suffix '%s'\n",
561 program_name, z_suffix);
564 if (do_lzw && !decompress) work = lzw;
566 /* Allocate all global buffers (for DYN_ALLOC option) */
567 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
568 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
569 ALLOC(ush, d_buf, DIST_BUFSIZE);
570 ALLOC(uch, window, 2L*WSIZE);
572 ALLOC(ush, tab_prefix, 1L<<BITS);
574 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
575 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
578 exiting_signal = quiet ? SIGPIPE : 0;
579 install_signal_handlers ();
581 /* And get to work */
582 if (file_count != 0) {
583 if (to_stdout && !test && !list && (!decompress || !ascii)) {
584 SET_BINARY_MODE(fileno(stdout));
586 while (optind < argc) {
587 treat_file(argv[optind++]);
589 } else { /* Standard input */
592 if (list && !quiet && file_count > 1) {
593 do_list(-1, -1); /* print totals */
596 return exit_code; /* just to avoid lint warning */
599 /* Return nonzero when at end of file on input. */
603 if (!decompress || last_member)
608 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
611 /* Unget the char that fill_inbuf got. */
618 /* ========================================================================
619 * Compress or decompress stdin
621 local void treat_stdin()
623 if (!force && !list &&
624 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
625 /* Do not send compressed data to the terminal or read it from
626 * the terminal. We get here when user invoked the program
627 * without parameters, so be helpful. According to the GNU standards:
629 * If there is one behavior you think is most useful when the output
630 * is to a terminal, and another that you think is most useful when
631 * the output is a file or a pipe, then it is usually best to make
632 * the default behavior the one that is useful with output to a
633 * terminal, and have an option for the other behavior.
635 * Here we use the --force option to get the other behavior.
638 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
639 program_name, decompress ? "read from" : "written to",
640 decompress ? "de" : "");
641 fprintf (stderr, "For help, type: %s -h\n", program_name);
645 if (decompress || !ascii) {
646 SET_BINARY_MODE(fileno(stdin));
648 if (!test && !list && (!decompress || !ascii)) {
649 SET_BINARY_MODE(fileno(stdout));
651 strcpy(ifname, "stdin");
652 strcpy(ofname, "stdout");
654 /* Get the file's time stamp and size. */
655 if (fstat (fileno (stdin), &istat) != 0)
657 progerror ("standard input");
660 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
661 time_stamp.tv_nsec = -1;
662 if (!no_time || list)
663 time_stamp = get_stat_mtime (&istat);
665 clear_bufs(); /* clear input and output buffers */
671 method = get_method(ifd);
673 do_exit(exit_code); /* error message already emitted */
677 do_list(ifd, method);
681 /* Actually do the compression/decompression. Loop over zipped members.
684 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
689 method = get_method(ifd);
690 if (method < 0) return; /* error message already emitted */
691 bytes_out = 0; /* required for length check */
696 fprintf(stderr, " OK\n");
698 } else if (!decompress) {
699 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
700 fprintf(stderr, "\n");
701 #ifdef DISPLAY_STDIN_RATIO
703 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
704 fprintf(stderr, "\n");
710 /* ========================================================================
711 * Compress or decompress the given file
713 local void treat_file(iname)
716 /* Accept "-" as synonym for stdin */
717 if (strequ(iname, "-")) {
718 int cflag = to_stdout;
724 /* Check if the input file is present, set ifname and istat: */
725 ifd = open_input_file (iname, &istat);
729 /* If the input name is that of a directory, recurse or ignore: */
730 if (S_ISDIR(istat.st_mode)) {
733 treat_dir (ifd, iname);
734 /* Warning: ifname is now garbage */
739 WARN ((stderr, "%s: %s is a directory -- ignored\n",
740 program_name, ifname));
746 if (! S_ISREG (istat.st_mode))
749 "%s: %s is not a directory or a regular file - ignored\n",
750 program_name, ifname));
754 if (istat.st_mode & S_ISUID)
756 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
757 program_name, ifname));
761 if (istat.st_mode & S_ISGID)
763 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
764 program_name, ifname));
771 if (istat.st_mode & S_ISVTX)
774 "%s: %s has the sticky bit set - file ignored\n",
775 program_name, ifname));
779 if (2 <= istat.st_nlink)
781 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
782 program_name, ifname,
783 (unsigned long int) istat.st_nlink - 1,
784 istat.st_nlink == 2 ? ' ' : 's'));
791 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
792 time_stamp.tv_nsec = -1;
793 if (!no_time || list)
794 time_stamp = get_stat_mtime (&istat);
796 /* Generate output file name. For -r and (-t or -l), skip files
797 * without a valid gzip suffix (check done in make_ofname).
799 if (to_stdout && !list && !test) {
800 strcpy(ofname, "stdout");
802 } else if (make_ofname() != OK) {
807 clear_bufs(); /* clear input and output buffers */
811 method = get_method(ifd); /* updates ofname if original given */
814 return; /* error message already emitted */
818 do_list(ifd, method);
819 if (close (ifd) != 0)
824 /* If compressing to a file, check if ofname is not ambiguous
825 * because the operating system truncates names. Otherwise, generate
826 * a new ofname and save the original name in the compressed file.
829 ofd = fileno(stdout);
830 /* Keep remove_ofname_fd negative. */
832 if (create_outfile() != OK) return;
834 if (!decompress && save_orig_name && !verbose && !quiet) {
835 fprintf(stderr, "%s: %s compressed to %s\n",
836 program_name, ifname, ofname);
839 /* Keep the name even if not truncated except with --no-name: */
840 if (!save_orig_name) save_orig_name = !no_name;
843 fprintf(stderr, "%s:\t", ifname);
846 /* Actually do the compression/decompression. Loop over zipped members.
849 if ((*work)(ifd, ofd) != OK) {
850 method = -1; /* force cleanup */
857 method = get_method(ifd);
858 if (method < 0) break; /* error message already emitted */
859 bytes_out = 0; /* required for length check */
862 if (close (ifd) != 0)
871 if (close (ofd) != 0)
874 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
875 remove_ofname_fd = -1;
876 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
877 sigprocmask (SIG_SETMASK, &oldset, NULL);
881 WARN ((stderr, "%s: ", program_name));
884 errno = unlink_errno;
892 remove_output_file ();
896 /* Display statistics */
899 fprintf(stderr, " OK");
900 } else if (decompress) {
901 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
903 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
905 if (!test && !to_stdout) {
906 fprintf(stderr, " -- replaced with %s", ofname);
908 fprintf(stderr, "\n");
912 /* ========================================================================
913 * Create the output file. Return OK or ERROR.
914 * Try several times if necessary to avoid truncating the z_suffix. For
915 * example, do not create a compressed file of name "1234567890123."
916 * Sets save_orig_name to true if the file name has been truncated.
917 * IN assertions: the input file has already been open (ifd is set) and
918 * ofname has already been updated if there was an original name.
919 * OUT assertions: ifd and ofd are closed in case of error.
921 local int create_outfile()
923 int name_shortened = 0;
924 int flags = (O_WRONLY | O_CREAT | O_EXCL
925 | (ascii && decompress ? 0 : O_BINARY));
932 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
933 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
935 sigprocmask (SIG_SETMASK, &oldset, NULL);
944 shorten_name (ofname);
950 if (check_ofname () != OK)
964 if (name_shortened && decompress)
966 /* name might be too long if an original name was saved */
967 WARN ((stderr, "%s: %s: warning, name truncated\n",
968 program_name, ofname));
974 /* ========================================================================
975 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
976 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
977 * accepted suffixes, in addition to the value of the --suffix option.
978 * ".tgz" is a useful convention for tar.z files on systems limited
979 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
980 * also accepted suffixes. For Unix, we do not want to accept any
981 * .??z suffix as indicating a compressed file; some people use .xyz
982 * to denote volume data.
983 * On systems allowing multiple versions of the same file (such as VMS),
984 * this function removes any version suffix in the given name.
986 local char *get_suffix(name)
990 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
991 static char const *known_suffixes[] =
992 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
997 char const **suf = known_suffixes;
1000 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
1003 /* strip a version number from the file name */
1005 char *v = strrchr(name, SUFFIX_SEP);
1006 if (v != NULL) *v = '\0';
1009 nlen = strlen(name);
1010 if (nlen <= MAX_SUFFIX+2) {
1011 strcpy(suffix, name);
1013 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1016 slen = strlen(suffix);
1018 int s = strlen(*suf);
1019 if (slen > s && suffix[slen-s-1] != PATH_SEP
1020 && strequ(suffix + slen - s, *suf)) {
1023 } while (*++suf != NULL);
1029 /* Open file NAME with the given flags and mode and store its status
1030 into *ST. Return a file descriptor to the newly opened file, or -1
1031 (setting errno) on failure. */
1033 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1037 /* Refuse to follow symbolic links unless -c or -f. */
1038 if (!to_stdout && !force)
1040 if (HAVE_WORKING_O_NOFOLLOW)
1041 flags |= O_NOFOLLOW;
1044 #if HAVE_LSTAT || defined lstat
1045 if (lstat (name, st) != 0)
1047 else if (S_ISLNK (st->st_mode))
1056 fd = OPEN (name, flags, mode);
1057 if (0 <= fd && fstat (fd, st) != 0)
1068 /* ========================================================================
1069 * Set ifname to the input file name (with a suffix appended if necessary)
1070 * and istat to its stats. For decompression, if no file exists with the
1071 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1072 * For MSDOS, we try only z_suffix and z.
1073 * Return an open file descriptor or -1.
1076 open_input_file (iname, sbuf)
1080 int ilen; /* strlen(ifname) */
1081 int z_suffix_errno = 0;
1082 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1083 char const **suf = suffixes;
1085 #ifdef NO_MULTIPLE_DOTS
1086 char *dot; /* pointer to ifname extension, or NULL */
1089 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1090 | (ascii && !decompress ? 0 : O_BINARY));
1094 if (sizeof ifname - 1 <= strlen (iname))
1097 strcpy(ifname, iname);
1099 /* If input file exists, return OK. */
1100 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1104 if (!decompress || errno != ENOENT) {
1108 /* file.ext doesn't exist, try adding a suffix (after removing any
1109 * version number for VMS).
1111 s = get_suffix(ifname);
1113 progerror(ifname); /* ifname already has z suffix and does not exist */
1116 #ifdef NO_MULTIPLE_DOTS
1117 dot = strrchr(ifname, '.');
1119 strcat(ifname, ".");
1120 dot = strrchr(ifname, '.');
1123 ilen = strlen(ifname);
1124 if (strequ(z_suffix, ".gz")) suf++;
1126 /* Search for all suffixes */
1128 char const *s0 = s = *suf;
1129 strcpy (ifname, iname);
1130 #ifdef NO_MULTIPLE_DOTS
1132 if (*dot == '\0') strcpy (dot, ".");
1134 #ifdef MAX_EXT_CHARS
1135 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1136 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1138 if (sizeof ifname <= ilen + strlen (s))
1141 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1144 if (errno != ENOENT)
1149 if (strequ (s0, z_suffix))
1150 z_suffix_errno = errno;
1151 } while (*++suf != NULL);
1153 /* No suffix found, complain using z_suffix: */
1154 strcpy(ifname, iname);
1155 #ifdef NO_MULTIPLE_DOTS
1156 if (*dot == '\0') strcpy(dot, ".");
1158 #ifdef MAX_EXT_CHARS
1159 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1160 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1162 strcat(ifname, z_suffix);
1163 errno = z_suffix_errno;
1168 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1173 /* ========================================================================
1174 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1175 * Sets save_orig_name to true if the file name has been truncated.
1177 local int make_ofname()
1179 char *suff; /* ofname z suffix */
1181 strcpy(ofname, ifname);
1182 /* strip a version number if any and get the gzip suffix if present: */
1183 suff = get_suffix(ofname);
1187 /* With -t or -l, try all files (even without .gz suffix)
1188 * except with -r (behave as with just -dr).
1190 if (!recursive && (list || test)) return OK;
1192 /* Avoid annoying messages with -r */
1193 if (verbose || (!recursive && !quiet)) {
1194 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1195 program_name, ifname));
1199 /* Make a special case for .tgz and .taz: */
1201 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1202 strcpy(suff, ".tar");
1204 *suff = '\0'; /* strip the z suffix */
1206 /* ofname might be changed later if infile contains an original name */
1208 } else if (suff && ! force) {
1209 /* Avoid annoying messages with -r (see treat_dir()) */
1210 if (verbose || (!recursive && !quiet)) {
1211 /* Don't use WARN, as it affects exit status. */
1212 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1213 program_name, ifname, suff);
1219 #ifdef NO_MULTIPLE_DOTS
1220 suff = strrchr(ofname, '.');
1222 if (sizeof ofname <= strlen (ofname) + 1)
1224 strcat(ofname, ".");
1225 # ifdef MAX_EXT_CHARS
1226 if (strequ(z_suffix, "z")) {
1227 if (sizeof ofname <= strlen (ofname) + 2)
1229 strcat(ofname, "gz"); /* enough room */
1232 /* On the Atari and some versions of MSDOS,
1233 * ENAMETOOLONG does not work correctly. So we
1234 * must truncate here.
1236 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1237 suff[MAX_SUFFIX+1-z_len] = '\0';
1241 #endif /* NO_MULTIPLE_DOTS */
1242 if (sizeof ofname <= strlen (ofname) + z_len)
1244 strcat(ofname, z_suffix);
1246 } /* decompress ? */
1250 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1255 /* ========================================================================
1256 * Check the magic number of the input file and update ofname if an
1257 * original name was given and to_stdout is not set.
1258 * Return the compression method, -1 for error, -2 for warning.
1259 * Set inptr to the offset of the next byte to be processed.
1260 * Updates time_stamp if there is one and --no-time is not used.
1261 * This function may be called repeatedly for an input file consisting
1262 * of several contiguous gzip'ed members.
1263 * IN assertions: there is at least one remaining compressed member.
1264 * If the member is a zip file, it must be the only one.
1266 local int get_method(in)
1267 int in; /* input file descriptor */
1269 uch flags; /* compression flags */
1270 char magic[2]; /* magic header */
1271 int imagic1; /* like magic[1], but can represent EOF */
1272 ulg stamp; /* time stamp */
1274 /* If --force and --stdout, zcat == cat, so do not complain about
1275 * premature end of file: use try_byte instead of get_byte.
1277 if (force && to_stdout) {
1278 magic[0] = (char)try_byte();
1279 imagic1 = try_byte ();
1280 magic[1] = (char) imagic1;
1281 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1283 magic[0] = (char)get_byte();
1285 magic[1] = (char)get_byte();
1286 imagic1 = 0; /* avoid lint warning */
1288 imagic1 = try_byte ();
1289 magic[1] = (char) imagic1;
1292 method = -1; /* unknown yet */
1293 part_nb++; /* number of parts in gzip file */
1295 last_member = RECORD_IO;
1296 /* assume multiple members in gzip file except for record oriented I/O */
1298 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1299 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1301 method = (int)get_byte();
1302 if (method != DEFLATED) {
1304 "%s: %s: unknown method %d -- not supported\n",
1305 program_name, ifname, method);
1310 flags = (uch)get_byte();
1312 if ((flags & ENCRYPTED) != 0) {
1314 "%s: %s is encrypted -- not supported\n",
1315 program_name, ifname);
1319 if ((flags & CONTINUATION) != 0) {
1321 "%s: %s is a multi-part gzip file -- not supported\n",
1322 program_name, ifname);
1324 if (force <= 1) return -1;
1326 if ((flags & RESERVED) != 0) {
1328 "%s: %s has flags 0x%x -- not supported\n",
1329 program_name, ifname, flags);
1331 if (force <= 1) return -1;
1333 stamp = (ulg)get_byte();
1334 stamp |= ((ulg)get_byte()) << 8;
1335 stamp |= ((ulg)get_byte()) << 16;
1336 stamp |= ((ulg)get_byte()) << 24;
1337 if (stamp != 0 && !no_time)
1339 time_stamp.tv_sec = stamp;
1340 time_stamp.tv_nsec = 0;
1343 (void)get_byte(); /* Ignore extra flags for the moment */
1344 (void)get_byte(); /* Ignore OS type for the moment */
1346 if ((flags & CONTINUATION) != 0) {
1347 unsigned part = (unsigned)get_byte();
1348 part |= ((unsigned)get_byte())<<8;
1350 fprintf(stderr,"%s: %s: part number %u\n",
1351 program_name, ifname, part);
1354 if ((flags & EXTRA_FIELD) != 0) {
1355 unsigned len = (unsigned)get_byte();
1356 len |= ((unsigned)get_byte())<<8;
1358 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1359 program_name, ifname, len);
1361 while (len--) (void)get_byte();
1364 /* Get original file name if it was truncated */
1365 if ((flags & ORIG_NAME) != 0) {
1366 if (no_name || (to_stdout && !list) || part_nb > 1) {
1367 /* Discard the old name */
1368 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1369 do {c=get_byte();} while (c != 0);
1371 /* Copy the base name. Keep a directory prefix intact. */
1372 char *p = gzip_base_name (ofname);
1375 *p = (char)get_char();
1376 if (*p++ == '\0') break;
1377 if (p >= ofname+sizeof(ofname)) {
1378 gzip_error ("corrupted input -- file name too large");
1381 p = gzip_base_name (base);
1382 memmove (base, p, strlen (p) + 1);
1383 /* If necessary, adapt the name to local OS conventions: */
1385 MAKE_LEGAL_NAME(base);
1386 if (base) list=0; /* avoid warning about unused variable */
1388 } /* no_name || to_stdout */
1391 /* Discard file comment if any */
1392 if ((flags & COMMENT) != 0) {
1393 while (get_char() != 0) /* null */ ;
1396 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1399 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1400 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1401 /* To simplify the code, we support a zip file when alone only.
1402 * We are thus guaranteed that the entire local header fits in inbuf.
1406 if (check_zipfile(in) != OK) return -1;
1407 /* check_zipfile may get ofname from the local header */
1410 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1414 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1416 method = COMPRESSED;
1419 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1424 } else if (force && to_stdout && !list) { /* pass input unchanged */
1430 if (method >= 0) return method;
1433 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1434 program_name, ifname);
1441 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1446 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1447 program_name, ifname));
1452 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1453 program_name, ifname));
1458 /* ========================================================================
1459 * Display the characteristics of the compressed file.
1460 * If the given method is < 0, display the accumulated totals.
1461 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1463 local void do_list(ifd, method)
1464 int ifd; /* input file descriptor */
1465 int method; /* compression method */
1467 ulg crc; /* original crc */
1468 static int first_time = 1;
1469 static char const *const methods[MAX_METHODS] = {
1474 "", "", "", "", /* 4 to 7 reserved */
1476 int positive_off_t_width = 1;
1479 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1480 positive_off_t_width++;
1483 if (first_time && method >= 0) {
1486 printf("method crc date time ");
1489 printf("%*.*s %*.*s ratio uncompressed_name\n",
1490 positive_off_t_width, positive_off_t_width, "compressed",
1491 positive_off_t_width, positive_off_t_width, "uncompressed");
1493 } else if (method < 0) {
1494 if (total_in <= 0 || total_out <= 0) return;
1498 if (verbose || !quiet) {
1499 fprint_off(stdout, total_in, positive_off_t_width);
1501 fprint_off(stdout, total_out, positive_off_t_width);
1504 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1505 /* header_bytes is not meaningful but used to ensure the same
1506 * ratio if there is a single file.
1508 printf(" (totals)\n");
1511 crc = (ulg)~0; /* unknown */
1513 bytes_in = ifile_size;
1516 if (method == DEFLATED && !last_member) {
1517 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1518 * If the lseek fails, we could use read() to get to the end, but
1519 * --list is used to get quick results.
1520 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1521 * you are not concerned about speed.
1523 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1524 if (bytes_in != -1L) {
1527 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1531 bytes_out = LG(buf+4);
1534 #endif /* RECORD_IO */
1537 struct tm *tm = localtime (&time_stamp.tv_sec);
1538 printf ("%5s %08lx ", methods[method], crc);
1540 printf ("%s%3d %02d:%02d ",
1541 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1543 tm->tm_mday, tm->tm_hour, tm->tm_min);
1545 printf ("??? ?? ??:?? ");
1547 fprint_off(stdout, bytes_in, positive_off_t_width);
1549 fprint_off(stdout, bytes_out, positive_off_t_width);
1551 if (bytes_in == -1L) {
1553 bytes_in = bytes_out = header_bytes = 0;
1554 } else if (total_in >= 0) {
1555 total_in += bytes_in;
1557 if (bytes_out == -1L) {
1559 bytes_in = bytes_out = header_bytes = 0;
1560 } else if (total_out >= 0) {
1561 total_out += bytes_out;
1563 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1564 printf(" %s\n", ofname);
1567 /* ========================================================================
1568 * Shorten the given name by one character, or replace a .tar extension
1569 * with .tgz. Truncate the last part of the name which is longer than
1570 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1571 * has only parts shorter than MIN_PART truncate the longest part.
1572 * For decompression, just remove the last character of the name.
1574 * IN assertion: for compression, the suffix of the given name is z_suffix.
1576 local void shorten_name(name)
1579 int len; /* length of name without z_suffix */
1580 char *trunc = NULL; /* character to be truncated */
1581 int plen; /* current part length */
1582 int min_part = MIN_PART; /* current minimum part length */
1588 gzip_error ("name too short");
1592 p = get_suffix(name);
1594 gzip_error ("can't recover suffix\n");
1598 /* compress 1234567890.tar to 1234567890.tgz */
1599 if (len > 4 && strequ(p-4, ".tar")) {
1600 strcpy(p-4, ".tgz");
1603 /* Try keeping short extensions intact:
1604 * 1234.678.012.gz -> 123.678.012.gz
1607 p = strrchr(name, PATH_SEP);
1610 plen = strcspn(p, PART_SEP);
1612 if (plen > min_part) trunc = p-1;
1615 } while (trunc == NULL && --min_part != 0);
1617 if (trunc != NULL) {
1619 trunc[0] = trunc[1];
1623 trunc = strrchr(name, PART_SEP[0]);
1625 gzip_error ("internal error in shorten_name");
1626 if (trunc[1] == '\0') trunc--; /* force truncation */
1628 strcpy(trunc, z_suffix);
1631 /* ========================================================================
1632 * The compressed file already exists, so ask for confirmation.
1633 * Return ERROR if the file must be skipped.
1635 local int check_ofname()
1637 /* Ask permission to overwrite the existing file */
1640 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1641 if (foreground && isatty(fileno(stdin))) {
1642 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1647 fprintf(stderr, "\tnot overwritten\n");
1648 if (exit_code == OK) exit_code = WARNING;
1652 if (xunlink (ofname)) {
1660 /* ========================================================================
1661 * Copy modes, times, ownership from input file to output file.
1662 * IN assertion: to_stdout is false.
1664 local void copy_stat(ifstat)
1665 struct stat *ifstat;
1667 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1671 struct timespec timespec[2];
1672 timespec[0] = get_stat_atime (ifstat);
1673 timespec[1] = get_stat_mtime (ifstat);
1675 if (decompress && 0 <= time_stamp.tv_nsec
1676 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1677 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1679 timespec[1] = time_stamp;
1681 fprintf(stderr, "%s: time stamp restored\n", ofname);
1685 if (gl_futimens (ofd, ofname, timespec) != 0)
1688 WARN ((stderr, "%s: ", program_name));
1698 /* Copy ownership */
1700 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1702 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1706 /* Copy the protection modes */
1708 r = fchmod (ofd, mode);
1710 r = chmod (ofname, mode);
1714 WARN ((stderr, "%s: ", program_name));
1724 /* ========================================================================
1725 * Recurse through the given directory. This code is taken from ncompress.
1727 local void treat_dir (fd, dir)
1733 char nbuf[MAX_PATH_LEN];
1736 dirp = fdopendir (fd);
1744 ** WARNING: the following algorithm could occasionally cause
1745 ** compress to produce error warnings of the form "<filename>.gz
1746 ** already has .gz suffix - ignored". This occurs when the
1747 ** .gz output file is inserted into the directory below
1748 ** readdir's current pointer.
1749 ** These warnings are harmless but annoying, so they are suppressed
1750 ** with option -r (except when -v is on). An alternative
1751 ** to allowing this would be to store the entire directory
1752 ** list in memory, then compress the entries in the stored
1753 ** list. Given the depth-first recursive algorithm used here,
1754 ** this could use up a tremendous amount of memory. I don't
1755 ** think it's worth it. -- Dave Mack
1756 ** (An other alternative might be two passes to avoid depth-first.)
1759 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1761 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1765 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1767 if (len != 0 /* dir = "" means current dir on Amiga */
1769 && dir[len-1] != PATH_SEP2
1772 && dir[len-1] != PATH_SEP3
1775 nbuf[len++] = PATH_SEP;
1777 strcpy(nbuf+len, dp->d_name);
1780 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1781 program_name, dir, dp->d_name);
1787 if (CLOSEDIR(dirp) != 0)
1790 #endif /* ! NO_DIR */
1792 /* Make sure signals get handled properly. */
1795 install_signal_handlers ()
1797 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1801 struct sigaction act;
1803 sigemptyset (&caught_signals);
1804 for (i = 0; i < nsigs; i++)
1806 sigaction (handled_sig[i], NULL, &act);
1807 if (act.sa_handler != SIG_IGN)
1808 sigaddset (&caught_signals, handled_sig[i]);
1811 act.sa_handler = abort_gzip_signal;
1812 act.sa_mask = caught_signals;
1815 for (i = 0; i < nsigs; i++)
1816 if (sigismember (&caught_signals, handled_sig[i]))
1820 sigaction (handled_sig[i], &act, NULL);
1823 for (i = 0; i < nsigs; i++)
1824 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1828 signal (handled_sig[i], abort_gzip_signal);
1829 siginterrupt (handled_sig[i], 1);
1834 /* ========================================================================
1835 * Free all dynamically allocated variables and exit with the given code.
1837 local void do_exit(exitcode)
1840 static int in_exit = 0;
1842 if (in_exit) exit(exitcode);
1861 /* ========================================================================
1862 * Close and unlink the output file.
1865 remove_output_file ()
1870 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1871 fd = remove_ofname_fd;
1874 remove_ofname_fd = -1;
1878 sigprocmask (SIG_SETMASK, &oldset, NULL);
1881 /* ========================================================================
1887 remove_output_file ();
1891 /* ========================================================================
1895 abort_gzip_signal (sig)
1899 signal (sig, SIG_IGN);
1900 remove_output_file ();
1901 if (sig == exiting_signal)
1903 signal (sig, SIG_DFL);