1 /* gzip (GNU zip) -- compress files with zip algorithm and 'compress' interface
3 Copyright (C) 1999, 2001-2002, 2006-2007, 2009-2013 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, 2010, 2011 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>
72 #include "fcntl-safer.h"
74 #include "ignore-value.h"
75 #include "stat-time.h"
99 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
102 # define MAX_PATH_LEN 1024 /* max pathname length */
114 off_t lseek (int fd, off_t offset, int whence);
118 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
122 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
125 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
128 # define SA_NOCLDSTOP 0
129 # define sigprocmask(how, set, oset) /* empty */
130 # define sigset_t int
131 # if ! HAVE_SIGINTERRUPT
132 # define siginterrupt(sig, flag) /* empty */
136 #ifndef HAVE_WORKING_O_NOFOLLOW
137 # define HAVE_WORKING_O_NOFOLLOW 0
140 /* Separator for file name parts (see shorten_name()) */
141 #ifdef NO_MULTIPLE_DOTS
142 # define PART_SEP "-"
144 # define PART_SEP "."
149 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
150 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
151 DECLARE(ush, d_buf, DIST_BUFSIZE);
152 DECLARE(uch, window, 2L*WSIZE);
154 DECLARE(ush, tab_prefix, 1L<<BITS);
156 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
157 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
160 /* local variables */
162 /* If true, pretend that standard input is a tty. This option
163 is deliberately not documented, and only for testing. */
164 static bool presume_input_tty;
166 static int ascii = 0; /* convert end-of-lines to local OS conventions */
167 int to_stdout = 0; /* output to stdout (-c) */
168 static int decompress = 0; /* decompress (-d) */
169 static int force = 0; /* don't ask questions, compress links (-f) */
170 static int no_name = -1; /* don't save or restore the original file name */
171 static int no_time = -1; /* don't save or restore the original file time */
172 static int recursive = 0; /* recurse through directories (-r) */
173 static int list = 0; /* list the file contents (-l) */
174 int verbose = 0; /* be verbose (-v) */
175 int quiet = 0; /* be very quiet (-q) */
176 static int do_lzw = 0; /* generate output compatible with old compress (-Z) */
177 int test = 0; /* test .gz file integrity */
178 static int foreground = 0; /* set if program run in foreground */
179 char *program_name; /* program name */
180 int maxbits = BITS; /* max bits per code for LZW */
181 int method = DEFLATED;/* compression method */
182 int level = 6; /* compression level */
183 int exit_code = OK; /* program exit code */
184 int save_orig_name; /* set if original name must be saved */
185 static int last_member; /* set for .zip and .Z files */
186 static int part_nb; /* number of parts in .gz file */
187 struct timespec time_stamp; /* original time stamp (modification time) */
188 off_t ifile_size; /* input file size, -1 for devices (debug only) */
189 static char *env; /* contents of GZIP env variable */
190 static char **args = NULL; /* argv pointer if GZIP env variable defined */
191 static char const *z_suffix; /* default suffix (can be set with --suffix) */
192 static size_t z_len; /* strlen(z_suffix) */
194 /* The set of signals that are caught. */
195 static sigset_t caught_signals;
197 /* If nonzero then exit with status WARNING, rather than with the usual
198 signal status, on receipt of a signal with this value. This
199 suppresses a "Broken Pipe" message with some shells. */
200 static int volatile exiting_signal;
202 /* If nonnegative, close this file descriptor and unlink ofname on error. */
203 static int volatile remove_ofname_fd = -1;
205 off_t bytes_in; /* number of input bytes */
206 off_t bytes_out; /* number of output bytes */
207 static off_t total_in; /* input bytes for all files */
208 static off_t total_out; /* output bytes for all files */
209 char ifname[MAX_PATH_LEN]; /* input file name */
210 char ofname[MAX_PATH_LEN]; /* output file name */
211 static struct stat istat; /* status for input file */
212 int ifd; /* input file descriptor */
213 int ofd; /* output file descriptor */
214 unsigned insize; /* valid bytes in inbuf */
215 unsigned inptr; /* index of next byte to be processed in inbuf */
216 unsigned outcnt; /* bytes in output buffer */
218 static int handled_sig[] =
220 /* SIGINT must be first, as 'foreground' depends on it. */
240 /* For long options that have no equivalent short option, use a
241 non-character as a pseudo short option, starting with CHAR_MAX + 1. */
244 PRESUME_INPUT_TTY_OPTION = CHAR_MAX + 1
247 static const struct option longopts[] =
249 /* { name has_arg *flag val } */
250 {"ascii", 0, 0, 'a'}, /* ascii text mode */
251 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
252 {"stdout", 0, 0, 'c'}, /* write output on standard output */
253 {"decompress", 0, 0, 'd'}, /* decompress */
254 {"uncompress", 0, 0, 'd'}, /* decompress */
255 /* {"encrypt", 0, 0, 'e'}, encrypt */
256 {"force", 0, 0, 'f'}, /* force overwrite of output file */
257 {"help", 0, 0, 'h'}, /* give help */
258 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
259 {"list", 0, 0, 'l'}, /* list .gz file contents */
260 {"license", 0, 0, 'L'}, /* display software license */
261 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
262 {"name", 0, 0, 'N'}, /* save or restore original name & time */
263 {"-presume-input-tty", no_argument, NULL, PRESUME_INPUT_TTY_OPTION},
264 {"quiet", 0, 0, 'q'}, /* quiet mode */
265 {"silent", 0, 0, 'q'}, /* quiet mode */
266 {"recursive", 0, 0, 'r'}, /* recurse through directories */
267 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
268 {"test", 0, 0, 't'}, /* test compressed file integrity */
269 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
270 {"verbose", 0, 0, 'v'}, /* verbose mode */
271 {"version", 0, 0, 'V'}, /* display version number */
272 {"fast", 0, 0, '1'}, /* compress faster */
273 {"best", 0, 0, '9'}, /* compress better */
274 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
275 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
280 /* local functions */
282 local void try_help (void) ATTRIBUTE_NORETURN;
283 local void help (void);
284 local void license (void);
285 local void version (void);
286 local int input_eof (void);
287 local void treat_stdin (void);
288 local void treat_file (char *iname);
289 local int create_outfile (void);
290 local char *get_suffix (char *name);
291 local int open_input_file (char *iname, struct stat *sbuf);
292 local void discard_input_bytes (size_t nbytes, unsigned int flags);
293 local int make_ofname (void);
294 local void shorten_name (char *name);
295 local int get_method (int in);
296 local void do_list (int ifd, int method);
297 local int check_ofname (void);
298 local void copy_stat (struct stat *ifstat);
299 local void install_signal_handlers (void);
300 local void remove_output_file (void);
301 local RETSIGTYPE abort_gzip_signal (int);
302 local void do_exit (int exitcode) ATTRIBUTE_NORETURN;
303 int main (int argc, char **argv);
304 static int (*work) (int infile, int outfile) = zip; /* function to call */
307 local void treat_dir (int fd, char *dir);
310 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
315 fprintf (stderr, "Try `%s --help' for more information.\n",
320 /* ======================================================================== */
323 static char const* const help_msg[] = {
324 "Compress or uncompress FILEs (by default, compress FILES in-place).",
326 "Mandatory arguments to long options are mandatory for short options too.",
329 " -a, --ascii ascii text; convert end-of-line using local conventions",
331 " -c, --stdout write on standard output, keep original files unchanged",
332 " -d, --decompress decompress",
333 /* -e, --encrypt encrypt */
334 " -f, --force force overwrite of output file and compress links",
335 " -h, --help give this help",
336 /* -k, --pkzip force output in pkzip format */
337 " -l, --list list compressed file contents",
338 " -L, --license display software license",
340 " -m, --no-time do not save or restore the original modification time",
341 " -M, --time save or restore the original modification time",
343 " -n, --no-name do not save or restore the original name and time stamp",
344 " -N, --name save or restore the original name and time stamp",
345 " -q, --quiet suppress all warnings",
347 " -r, --recursive operate recursively on directories",
349 " -S, --suffix=SUF use suffix SUF on compressed files",
350 " -t, --test test compressed file integrity",
351 " -v, --verbose verbose mode",
352 " -V, --version display version number",
353 " -1, --fast compress faster",
354 " -9, --best compress better",
356 " -Z, --lzw produce output compatible with old compress",
357 " -b, --bits=BITS max number of bits per code (implies -Z)",
360 "With no FILE, or when FILE is -, read standard input.",
362 "Report bugs to <bug-gzip@gnu.org>.",
364 char const *const *p = help_msg;
366 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
367 while (*p) printf ("%s\n", *p++);
370 /* ======================================================================== */
373 char const *const *p = license_msg;
375 printf ("%s %s\n", program_name, Version);
376 while (*p) printf ("%s\n", *p++);
379 /* ======================================================================== */
384 printf ("Written by Jean-loup Gailly.\n");
387 local void progerror (char const *string)
390 fprintf (stderr, "%s: ", program_name);
396 /* ======================================================================== */
397 int main (int argc, char **argv)
399 int file_count; /* number of files to process */
400 size_t proglen; /* length of program_name */
401 int optc; /* current option */
403 EXPAND(argc, argv); /* wild card expansion if necessary */
405 program_name = gzip_base_name (argv[0]);
406 proglen = strlen (program_name);
408 atexit (close_stdin);
410 /* Suppress .exe for MSDOS, OS/2 and VMS: */
411 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
412 program_name[proglen - 4] = '\0';
414 /* Add options in GZIP environment variable if there is one */
415 env = add_envopt(&argc, &argv, OPTIONS_VAR);
416 if (env != NULL) args = argv;
419 # define GNU_STANDARD 1
422 /* For compatibility with old compress, use program name as an option.
423 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
424 * gzip even if it is invoked under the name gunzip or zcat.
426 * Systems which do not support links can still use -d or -dc.
427 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
429 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
430 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
432 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
433 || strequ (program_name, "gzcat")) /* gzcat */
434 decompress = to_stdout = 1;
438 z_len = strlen(z_suffix);
440 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
441 longopts, (int *)0)) != -1) {
446 maxbits = atoi(optarg);
447 for (; *optarg; optarg++)
448 if (! ('0' <= *optarg && *optarg <= '9'))
450 fprintf (stderr, "%s: -b operand is not an integer\n",
456 to_stdout = 1; break;
458 decompress = 1; break;
462 help(); do_exit(OK); break;
464 list = decompress = to_stdout = 1; break;
466 license(); do_exit(OK); break;
467 case 'm': /* undocumented, may change later */
469 case 'M': /* undocumented, may change later */
472 no_name = no_time = 1; break;
474 no_name = no_time = 0; break;
475 case PRESUME_INPUT_TTY_OPTION:
476 presume_input_tty = true; break;
478 quiet = 1; verbose = 0; break;
481 fprintf (stderr, "%s: -r not supported on this system\n",
489 #ifdef NO_MULTIPLE_DOTS
490 if (*optarg == '.') optarg++;
492 z_len = strlen(optarg);
496 test = decompress = to_stdout = 1;
499 verbose++; quiet = 0; break;
501 version(); do_exit(OK); break;
506 fprintf(stderr, "%s: -Z not supported in this version\n",
511 case '1': case '2': case '3': case '4':
512 case '5': case '6': case '7': case '8': case '9':
516 /* Error message already emitted by getopt_long. */
519 } /* loop on all arguments */
521 /* By default, save name and timestamp on compression but do not
522 * restore them on decompression.
524 if (no_time < 0) no_time = decompress;
525 if (no_name < 0) no_name = decompress;
527 file_count = argc - optind;
531 if (ascii && !quiet) {
532 fprintf(stderr, "%s: option --ascii ignored on this system\n",
536 if (z_len == 0 || z_len > MAX_SUFFIX) {
537 fprintf(stderr, "%s: invalid suffix '%s'\n", program_name, z_suffix);
541 if (do_lzw && !decompress) work = lzw;
543 /* Allocate all global buffers (for DYN_ALLOC option) */
544 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
545 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
546 ALLOC(ush, d_buf, DIST_BUFSIZE);
547 ALLOC(uch, window, 2L*WSIZE);
549 ALLOC(ush, tab_prefix, 1L<<BITS);
551 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
552 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
555 exiting_signal = quiet ? SIGPIPE : 0;
556 install_signal_handlers ();
558 /* And get to work */
559 if (file_count != 0) {
560 if (to_stdout && !test && !list && (!decompress || !ascii)) {
561 SET_BINARY_MODE(fileno(stdout));
563 while (optind < argc) {
564 treat_file(argv[optind++]);
566 } else { /* Standard input */
569 if (list && !quiet && file_count > 1) {
570 do_list(-1, -1); /* print totals */
573 return exit_code; /* just to avoid lint warning */
576 /* Return nonzero when at end of file on input. */
580 if (!decompress || last_member)
585 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
588 /* Unget the char that fill_inbuf got. */
595 /* ========================================================================
596 * Compress or decompress stdin
598 local void treat_stdin()
601 && (presume_input_tty
602 || isatty(fileno((FILE *)(decompress ? stdin : stdout))))) {
603 /* Do not send compressed data to the terminal or read it from
604 * the terminal. We get here when user invoked the program
605 * without parameters, so be helpful. According to the GNU standards:
607 * If there is one behavior you think is most useful when the output
608 * is to a terminal, and another that you think is most useful when
609 * the output is a file or a pipe, then it is usually best to make
610 * the default behavior the one that is useful with output to a
611 * terminal, and have an option for the other behavior.
613 * Here we use the --force option to get the other behavior.
617 ("%s: compressed data not %s a terminal."
618 " Use -f to force %scompression.\n"
619 "For help, type: %s -h\n"),
621 decompress ? "read from" : "written to",
622 decompress ? "de" : "",
627 if (decompress || !ascii) {
628 SET_BINARY_MODE(fileno(stdin));
630 if (!test && !list && (!decompress || !ascii)) {
631 SET_BINARY_MODE(fileno(stdout));
633 strcpy(ifname, "stdin");
634 strcpy(ofname, "stdout");
636 /* Get the file's time stamp and size. */
637 if (fstat (fileno (stdin), &istat) != 0)
639 progerror ("standard input");
642 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
643 time_stamp.tv_nsec = -1;
644 if (!no_time || list)
646 if (S_ISREG (istat.st_mode))
647 time_stamp = get_stat_mtime (&istat);
649 gettime (&time_stamp);
652 clear_bufs(); /* clear input and output buffers */
658 method = get_method(ifd);
660 do_exit(exit_code); /* error message already emitted */
664 do_list(ifd, method);
668 /* Actually do the compression/decompression. Loop over zipped members.
671 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
676 method = get_method(ifd);
677 if (method < 0) return; /* error message already emitted */
678 bytes_out = 0; /* required for length check */
683 fprintf(stderr, " OK\n");
685 } else if (!decompress) {
686 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
687 fprintf(stderr, "\n");
688 #ifdef DISPLAY_STDIN_RATIO
690 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
691 fprintf(stderr, "\n");
697 /* ========================================================================
698 * Compress or decompress the given file
700 local void treat_file(iname)
703 /* Accept "-" as synonym for stdin */
704 if (strequ(iname, "-")) {
705 int cflag = to_stdout;
711 /* Check if the input file is present, set ifname and istat: */
712 ifd = open_input_file (iname, &istat);
716 /* If the input name is that of a directory, recurse or ignore: */
717 if (S_ISDIR(istat.st_mode)) {
720 treat_dir (ifd, iname);
721 /* Warning: ifname is now garbage */
726 WARN ((stderr, "%s: %s is a directory -- ignored\n",
727 program_name, ifname));
733 if (! S_ISREG (istat.st_mode))
736 "%s: %s is not a directory or a regular file - ignored\n",
737 program_name, ifname));
741 if (istat.st_mode & S_ISUID)
743 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
744 program_name, ifname));
748 if (istat.st_mode & S_ISGID)
750 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
751 program_name, ifname));
758 if (istat.st_mode & S_ISVTX)
761 "%s: %s has the sticky bit set - file ignored\n",
762 program_name, ifname));
766 if (2 <= istat.st_nlink)
768 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
769 program_name, ifname,
770 (unsigned long int) istat.st_nlink - 1,
771 istat.st_nlink == 2 ? ' ' : 's'));
778 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
779 time_stamp.tv_nsec = -1;
780 if (!no_time || list)
781 time_stamp = get_stat_mtime (&istat);
783 /* Generate output file name. For -r and (-t or -l), skip files
784 * without a valid gzip suffix (check done in make_ofname).
786 if (to_stdout && !list && !test) {
787 strcpy(ofname, "stdout");
789 } else if (make_ofname() != OK) {
794 clear_bufs(); /* clear input and output buffers */
798 method = get_method(ifd); /* updates ofname if original given */
801 return; /* error message already emitted */
805 do_list(ifd, method);
806 if (close (ifd) != 0)
811 /* If compressing to a file, check if ofname is not ambiguous
812 * because the operating system truncates names. Otherwise, generate
813 * a new ofname and save the original name in the compressed file.
816 ofd = fileno(stdout);
817 /* Keep remove_ofname_fd negative. */
819 if (create_outfile() != OK) return;
821 if (!decompress && save_orig_name && !verbose && !quiet) {
822 fprintf(stderr, "%s: %s compressed to %s\n",
823 program_name, ifname, ofname);
826 /* Keep the name even if not truncated except with --no-name: */
827 if (!save_orig_name) save_orig_name = !no_name;
830 fprintf(stderr, "%s:\t", ifname);
833 /* Actually do the compression/decompression. Loop over zipped members.
836 if ((*work)(ifd, ofd) != OK) {
837 method = -1; /* force cleanup */
844 method = get_method(ifd);
845 if (method < 0) break; /* error message already emitted */
846 bytes_out = 0; /* required for length check */
849 if (close (ifd) != 0)
858 if (close (ofd) != 0)
861 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
862 remove_ofname_fd = -1;
863 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
864 sigprocmask (SIG_SETMASK, &oldset, NULL);
868 WARN ((stderr, "%s: ", program_name));
871 errno = unlink_errno;
879 remove_output_file ();
883 /* Display statistics */
886 fprintf(stderr, " OK");
887 } else if (decompress) {
888 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
890 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
892 if (!test && !to_stdout) {
893 fprintf(stderr, " -- replaced with %s", ofname);
895 fprintf(stderr, "\n");
899 /* ========================================================================
900 * Create the output file. Return OK or ERROR.
901 * Try several times if necessary to avoid truncating the z_suffix. For
902 * example, do not create a compressed file of name "1234567890123."
903 * Sets save_orig_name to true if the file name has been truncated.
904 * IN assertions: the input file has already been open (ifd is set) and
905 * ofname has already been updated if there was an original name.
906 * OUT assertions: ifd and ofd are closed in case of error.
908 local int create_outfile()
910 int name_shortened = 0;
911 int flags = (O_WRONLY | O_CREAT | O_EXCL
912 | (ascii && decompress ? 0 : O_BINARY));
919 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
920 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
922 sigprocmask (SIG_SETMASK, &oldset, NULL);
931 shorten_name (ofname);
937 if (check_ofname () != OK)
951 if (name_shortened && decompress)
953 /* name might be too long if an original name was saved */
954 WARN ((stderr, "%s: %s: warning, name truncated\n",
955 program_name, ofname));
961 /* ========================================================================
962 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
963 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
964 * accepted suffixes, in addition to the value of the --suffix option.
965 * ".tgz" is a useful convention for tar.z files on systems limited
966 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
967 * also accepted suffixes. For Unix, we do not want to accept any
968 * .??z suffix as indicating a compressed file; some people use .xyz
969 * to denote volume data.
970 * On systems allowing multiple versions of the same file (such as VMS),
971 * this function removes any version suffix in the given name.
973 local char *get_suffix(name)
977 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
978 static char const *known_suffixes[] =
979 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
984 char const **suf = known_suffixes;
987 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
990 /* strip a version number from the file name */
992 char *v = strrchr(name, SUFFIX_SEP);
993 if (v != NULL) *v = '\0';
997 if (nlen <= MAX_SUFFIX+2) {
998 strcpy(suffix, name);
1000 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
1003 slen = strlen(suffix);
1005 int s = strlen(*suf);
1006 if (slen > s && suffix[slen-s-1] != PATH_SEP
1007 && strequ(suffix + slen - s, *suf)) {
1010 } while (*++suf != NULL);
1016 /* Open file NAME with the given flags and mode and store its status
1017 into *ST. Return a file descriptor to the newly opened file, or -1
1018 (setting errno) on failure. */
1020 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1024 /* Refuse to follow symbolic links unless -c or -f. */
1025 if (!to_stdout && !force)
1027 if (HAVE_WORKING_O_NOFOLLOW)
1028 flags |= O_NOFOLLOW;
1031 #if HAVE_LSTAT || defined lstat
1032 if (lstat (name, st) != 0)
1034 else if (S_ISLNK (st->st_mode))
1043 fd = OPEN (name, flags, mode);
1044 if (0 <= fd && fstat (fd, st) != 0)
1055 /* ========================================================================
1056 * Set ifname to the input file name (with a suffix appended if necessary)
1057 * and istat to its stats. For decompression, if no file exists with the
1058 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1059 * For MSDOS, we try only z_suffix and z.
1060 * Return an open file descriptor or -1.
1063 open_input_file (iname, sbuf)
1067 int ilen; /* strlen(ifname) */
1068 int z_suffix_errno = 0;
1069 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1070 char const **suf = suffixes;
1072 #ifdef NO_MULTIPLE_DOTS
1073 char *dot; /* pointer to ifname extension, or NULL */
1076 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1077 | (ascii && !decompress ? 0 : O_BINARY));
1081 if (sizeof ifname - 1 <= strlen (iname))
1084 strcpy(ifname, iname);
1086 /* If input file exists, return OK. */
1087 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1091 if (!decompress || errno != ENOENT) {
1095 /* file.ext doesn't exist, try adding a suffix (after removing any
1096 * version number for VMS).
1098 s = get_suffix(ifname);
1100 progerror(ifname); /* ifname already has z suffix and does not exist */
1103 #ifdef NO_MULTIPLE_DOTS
1104 dot = strrchr(ifname, '.');
1106 strcat(ifname, ".");
1107 dot = strrchr(ifname, '.');
1110 ilen = strlen(ifname);
1111 if (strequ(z_suffix, ".gz")) suf++;
1113 /* Search for all suffixes */
1115 char const *s0 = s = *suf;
1116 strcpy (ifname, iname);
1117 #ifdef NO_MULTIPLE_DOTS
1119 if (*dot == '\0') strcpy (dot, ".");
1121 #ifdef MAX_EXT_CHARS
1122 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1123 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1125 if (sizeof ifname <= ilen + strlen (s))
1128 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1131 if (errno != ENOENT)
1136 if (strequ (s0, z_suffix))
1137 z_suffix_errno = errno;
1138 } while (*++suf != NULL);
1140 /* No suffix found, complain using z_suffix: */
1141 strcpy(ifname, iname);
1142 #ifdef NO_MULTIPLE_DOTS
1143 if (*dot == '\0') strcpy(dot, ".");
1145 #ifdef MAX_EXT_CHARS
1146 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1147 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1149 strcat(ifname, z_suffix);
1150 errno = z_suffix_errno;
1155 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1160 /* ========================================================================
1161 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1162 * Sets save_orig_name to true if the file name has been truncated.
1164 local int make_ofname()
1166 char *suff; /* ofname z suffix */
1168 strcpy(ofname, ifname);
1169 /* strip a version number if any and get the gzip suffix if present: */
1170 suff = get_suffix(ofname);
1174 /* With -t or -l, try all files (even without .gz suffix)
1175 * except with -r (behave as with just -dr).
1177 if (!recursive && (list || test)) return OK;
1179 /* Avoid annoying messages with -r */
1180 if (verbose || (!recursive && !quiet)) {
1181 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1182 program_name, ifname));
1186 /* Make a special case for .tgz and .taz: */
1188 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1189 strcpy(suff, ".tar");
1191 *suff = '\0'; /* strip the z suffix */
1193 /* ofname might be changed later if infile contains an original name */
1195 } else if (suff && ! force) {
1196 /* Avoid annoying messages with -r (see treat_dir()) */
1197 if (verbose || (!recursive && !quiet)) {
1198 /* Don't use WARN, as it affects exit status. */
1199 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1200 program_name, ifname, suff);
1206 #ifdef NO_MULTIPLE_DOTS
1207 suff = strrchr(ofname, '.');
1209 if (sizeof ofname <= strlen (ofname) + 1)
1211 strcat(ofname, ".");
1212 # ifdef MAX_EXT_CHARS
1213 if (strequ(z_suffix, "z")) {
1214 if (sizeof ofname <= strlen (ofname) + 2)
1216 strcat(ofname, "gz"); /* enough room */
1219 /* On the Atari and some versions of MSDOS,
1220 * ENAMETOOLONG does not work correctly. So we
1221 * must truncate here.
1223 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1224 suff[MAX_SUFFIX+1-z_len] = '\0';
1228 #endif /* NO_MULTIPLE_DOTS */
1229 if (sizeof ofname <= strlen (ofname) + z_len)
1231 strcat(ofname, z_suffix);
1233 } /* decompress ? */
1237 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1241 /* Discard NBYTES input bytes from the input, or up through the next
1242 zero byte if NBYTES == (size_t) -1. If FLAGS say that the header
1243 CRC should be computed, update the CRC accordingly. */
1245 discard_input_bytes (nbytes, flags)
1251 uch c = get_byte ();
1252 if (flags & HEADER_CRC)
1254 if (nbytes != (size_t) -1)
1261 /* ========================================================================
1262 * Check the magic number of the input file and update ofname if an
1263 * original name was given and to_stdout is not set.
1264 * Return the compression method, -1 for error, -2 for warning.
1265 * Set inptr to the offset of the next byte to be processed.
1266 * Updates time_stamp if there is one and --no-time is not used.
1267 * This function may be called repeatedly for an input file consisting
1268 * of several contiguous gzip'ed members.
1269 * IN assertions: there is at least one remaining compressed member.
1270 * If the member is a zip file, it must be the only one.
1272 local int get_method(in)
1273 int in; /* input file descriptor */
1275 uch flags; /* compression flags */
1276 uch magic[10]; /* magic header */
1277 int imagic0; /* first magic byte or EOF */
1278 int imagic1; /* like magic[1], but can represent EOF */
1279 ulg stamp; /* time stamp */
1281 /* If --force and --stdout, zcat == cat, so do not complain about
1282 * premature end of file: use try_byte instead of get_byte.
1284 if (force && to_stdout) {
1285 imagic0 = try_byte();
1287 imagic1 = try_byte ();
1289 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1291 magic[0] = get_byte ();
1294 magic[1] = get_byte ();
1295 imagic1 = 0; /* avoid lint warning */
1297 imagic1 = try_byte ();
1301 method = -1; /* unknown yet */
1302 part_nb++; /* number of parts in gzip file */
1304 last_member = RECORD_IO;
1305 /* assume multiple members in gzip file except for record oriented I/O */
1307 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1308 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1310 method = (int)get_byte();
1311 if (method != DEFLATED) {
1313 "%s: %s: unknown method %d -- not supported\n",
1314 program_name, ifname, method);
1319 flags = (uch)get_byte();
1321 if ((flags & ENCRYPTED) != 0) {
1323 "%s: %s is encrypted -- not supported\n",
1324 program_name, ifname);
1328 if ((flags & RESERVED) != 0) {
1330 "%s: %s has flags 0x%x -- not supported\n",
1331 program_name, ifname, flags);
1333 if (force <= 1) return -1;
1335 stamp = (ulg)get_byte();
1336 stamp |= ((ulg)get_byte()) << 8;
1337 stamp |= ((ulg)get_byte()) << 16;
1338 stamp |= ((ulg)get_byte()) << 24;
1339 if (stamp != 0 && !no_time)
1341 time_stamp.tv_sec = stamp;
1342 time_stamp.tv_nsec = 0;
1345 magic[8] = get_byte (); /* Ignore extra flags. */
1346 magic[9] = get_byte (); /* Ignore OS type. */
1348 if (flags & HEADER_CRC)
1350 magic[2] = DEFLATED;
1352 magic[4] = stamp & 0xff;
1353 magic[5] = (stamp >> 8) & 0xff;
1354 magic[6] = (stamp >> 16) & 0xff;
1355 magic[7] = stamp >> 24;
1360 if ((flags & EXTRA_FIELD) != 0) {
1362 unsigned int len = lenbuf[0] = get_byte ();
1363 len |= (lenbuf[1] = get_byte ()) << 8;
1365 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1366 program_name, ifname, len);
1368 if (flags & HEADER_CRC)
1370 discard_input_bytes (len, flags);
1373 /* Get original file name if it was truncated */
1374 if ((flags & ORIG_NAME) != 0) {
1375 if (no_name || (to_stdout && !list) || part_nb > 1) {
1376 /* Discard the old name */
1377 discard_input_bytes (-1, flags);
1379 /* Copy the base name. Keep a directory prefix intact. */
1380 char *p = gzip_base_name (ofname);
1383 *p = (char) get_byte ();
1384 if (*p++ == '\0') break;
1385 if (p >= ofname+sizeof(ofname)) {
1386 gzip_error ("corrupted input -- file name too large");
1389 if (flags & HEADER_CRC)
1390 updcrc ((uch *) base, p - base);
1391 p = gzip_base_name (base);
1392 memmove (base, p, strlen (p) + 1);
1393 /* If necessary, adapt the name to local OS conventions: */
1395 MAKE_LEGAL_NAME(base);
1396 if (base) list=0; /* avoid warning about unused variable */
1398 } /* no_name || to_stdout */
1401 /* Discard file comment if any */
1402 if ((flags & COMMENT) != 0) {
1403 discard_input_bytes (-1, flags);
1406 if (flags & HEADER_CRC)
1408 unsigned int crc16 = updcrc (magic, 0) & 0xffff;
1409 unsigned int header16 = get_byte ();
1410 header16 |= ((unsigned int) get_byte ()) << 8;
1411 if (header16 != crc16)
1414 "%s: %s: header checksum 0x%04x != computed checksum 0x%04x\n",
1415 program_name, ifname, header16, crc16);
1423 header_bytes = inptr + 2*4; /* include crc and size */
1426 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1427 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1428 /* To simplify the code, we support a zip file when alone only.
1429 * We are thus guaranteed that the entire local header fits in inbuf.
1433 if (check_zipfile(in) != OK) return -1;
1434 /* check_zipfile may get ofname from the local header */
1437 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1441 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1443 method = COMPRESSED;
1446 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1451 } else if (force && to_stdout && !list) { /* pass input unchanged */
1457 if (imagic0 != EOF) {
1458 write_buf(fileno(stdout), magic, 1);
1462 if (method >= 0) return method;
1465 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1466 program_name, ifname);
1473 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1478 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1479 program_name, ifname));
1484 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1485 program_name, ifname));
1490 /* ========================================================================
1491 * Display the characteristics of the compressed file.
1492 * If the given method is < 0, display the accumulated totals.
1493 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1495 local void do_list(ifd, method)
1496 int ifd; /* input file descriptor */
1497 int method; /* compression method */
1499 ulg crc; /* original crc */
1500 static int first_time = 1;
1501 static char const *const methods[MAX_METHODS] = {
1506 "", "", "", "", /* 4 to 7 reserved */
1508 int positive_off_t_width = 1;
1511 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1512 positive_off_t_width++;
1515 if (first_time && method >= 0) {
1518 printf("method crc date time ");
1521 printf("%*.*s %*.*s ratio uncompressed_name\n",
1522 positive_off_t_width, positive_off_t_width, "compressed",
1523 positive_off_t_width, positive_off_t_width, "uncompressed");
1525 } else if (method < 0) {
1526 if (total_in <= 0 || total_out <= 0) return;
1530 if (verbose || !quiet) {
1531 fprint_off(stdout, total_in, positive_off_t_width);
1533 fprint_off(stdout, total_out, positive_off_t_width);
1536 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1537 /* header_bytes is not meaningful but used to ensure the same
1538 * ratio if there is a single file.
1540 printf(" (totals)\n");
1543 crc = (ulg)~0; /* unknown */
1545 bytes_in = ifile_size;
1547 if (!RECORD_IO && method == DEFLATED && !last_member) {
1548 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1549 * If the lseek fails, we could use read() to get to the end, but
1550 * --list is used to get quick results.
1551 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1552 * you are not concerned about speed.
1554 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1555 if (bytes_in != -1L) {
1558 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1562 bytes_out = LG(buf+4);
1568 struct tm *tm = localtime (&time_stamp.tv_sec);
1569 printf ("%5s %08lx ", methods[method], crc);
1571 printf ("%s%3d %02d:%02d ",
1572 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1574 tm->tm_mday, tm->tm_hour, tm->tm_min);
1576 printf ("??? ?? ??:?? ");
1578 fprint_off(stdout, bytes_in, positive_off_t_width);
1580 fprint_off(stdout, bytes_out, positive_off_t_width);
1582 if (bytes_in == -1L) {
1584 bytes_in = bytes_out = header_bytes = 0;
1585 } else if (total_in >= 0) {
1586 total_in += bytes_in;
1588 if (bytes_out == -1L) {
1590 bytes_in = bytes_out = header_bytes = 0;
1591 } else if (total_out >= 0) {
1592 total_out += bytes_out;
1594 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1595 printf(" %s\n", ofname);
1598 /* ========================================================================
1599 * Shorten the given name by one character, or replace a .tar extension
1600 * with .tgz. Truncate the last part of the name which is longer than
1601 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1602 * has only parts shorter than MIN_PART truncate the longest part.
1603 * For decompression, just remove the last character of the name.
1605 * IN assertion: for compression, the suffix of the given name is z_suffix.
1607 local void shorten_name(name)
1610 int len; /* length of name without z_suffix */
1611 char *trunc = NULL; /* character to be truncated */
1612 int plen; /* current part length */
1613 int min_part = MIN_PART; /* current minimum part length */
1619 gzip_error ("name too short");
1623 p = get_suffix(name);
1625 gzip_error ("can't recover suffix\n");
1629 /* compress 1234567890.tar to 1234567890.tgz */
1630 if (len > 4 && strequ(p-4, ".tar")) {
1631 strcpy(p-4, ".tgz");
1634 /* Try keeping short extensions intact:
1635 * 1234.678.012.gz -> 123.678.012.gz
1638 p = strrchr(name, PATH_SEP);
1641 plen = strcspn(p, PART_SEP);
1643 if (plen > min_part) trunc = p-1;
1646 } while (trunc == NULL && --min_part != 0);
1648 if (trunc != NULL) {
1650 trunc[0] = trunc[1];
1654 trunc = strrchr(name, PART_SEP[0]);
1656 gzip_error ("internal error in shorten_name");
1657 if (trunc[1] == '\0') trunc--; /* force truncation */
1659 strcpy(trunc, z_suffix);
1662 /* ========================================================================
1663 * The compressed file already exists, so ask for confirmation.
1664 * Return ERROR if the file must be skipped.
1666 local int check_ofname()
1668 /* Ask permission to overwrite the existing file */
1671 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1672 if (foreground && (presume_input_tty || isatty(fileno(stdin)))) {
1673 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1678 fprintf(stderr, "\tnot overwritten\n");
1679 if (exit_code == OK) exit_code = WARNING;
1683 if (xunlink (ofname)) {
1691 /* ========================================================================
1692 * Copy modes, times, ownership from input file to output file.
1693 * IN assertion: to_stdout is false.
1695 local void copy_stat(ifstat)
1696 struct stat *ifstat;
1698 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1702 struct timespec timespec[2];
1703 timespec[0] = get_stat_atime (ifstat);
1704 timespec[1] = get_stat_mtime (ifstat);
1706 if (decompress && 0 <= time_stamp.tv_nsec
1707 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1708 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1710 timespec[1] = time_stamp;
1712 fprintf(stderr, "%s: time stamp restored\n", ofname);
1716 if (fdutimens (ofd, ofname, timespec) != 0)
1719 WARN ((stderr, "%s: ", program_name));
1729 /* Copy ownership */
1731 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1733 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1737 /* Copy the protection modes */
1739 r = fchmod (ofd, mode);
1741 r = chmod (ofname, mode);
1745 WARN ((stderr, "%s: ", program_name));
1755 /* ========================================================================
1756 * Recurse through the given directory.
1758 local void treat_dir (fd, dir)
1763 char nbuf[MAX_PATH_LEN];
1768 dirp = fdopendir (fd);
1776 entries = streamsavedir (dirp);
1779 if (closedir (dirp) != 0)
1784 for (entry = entries; *entry; entry += entrylen + 1) {
1785 size_t len = strlen (dir);
1786 entrylen = strlen (entry);
1787 if (strequ (entry, ".") || strequ (entry, ".."))
1789 if (len + entrylen < MAX_PATH_LEN - 2) {
1791 if (len != 0 /* dir = "" means current dir on Amiga */
1793 && dir[len-1] != PATH_SEP2
1796 && dir[len-1] != PATH_SEP3
1799 nbuf[len++] = PATH_SEP;
1801 strcpy (nbuf + len, entry);
1804 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1805 program_name, dir, entry);
1811 #endif /* ! NO_DIR */
1813 /* Make sure signals get handled properly. */
1816 install_signal_handlers ()
1818 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1822 struct sigaction act;
1824 sigemptyset (&caught_signals);
1825 for (i = 0; i < nsigs; i++)
1827 sigaction (handled_sig[i], NULL, &act);
1828 if (act.sa_handler != SIG_IGN)
1829 sigaddset (&caught_signals, handled_sig[i]);
1832 act.sa_handler = abort_gzip_signal;
1833 act.sa_mask = caught_signals;
1836 for (i = 0; i < nsigs; i++)
1837 if (sigismember (&caught_signals, handled_sig[i]))
1841 sigaction (handled_sig[i], &act, NULL);
1844 for (i = 0; i < nsigs; i++)
1845 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1849 signal (handled_sig[i], abort_gzip_signal);
1850 siginterrupt (handled_sig[i], 1);
1855 /* ========================================================================
1856 * Free all dynamically allocated variables and exit with the given code.
1858 local void do_exit(exitcode)
1861 static int in_exit = 0;
1863 if (in_exit) exit(exitcode);
1882 /* ========================================================================
1883 * Close and unlink the output file.
1886 remove_output_file ()
1891 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1892 fd = remove_ofname_fd;
1895 remove_ofname_fd = -1;
1899 sigprocmask (SIG_SETMASK, &oldset, NULL);
1902 /* ========================================================================
1908 remove_output_file ();
1912 /* ========================================================================
1916 abort_gzip_signal (sig)
1920 signal (sig, SIG_IGN);
1921 remove_output_file ();
1922 if (sig == exiting_signal)
1924 signal (sig, SIG_DFL);