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>
96 #if defined STDC_HEADERS || defined HAVE_STDLIB_H
107 # ifndef _D_EXACT_NAMLEN
108 # define _D_EXACT_NAMLEN(dp) strlen ((dp)->d_name)
113 # define CLOSEDIR(d) (closedir(d), 0)
115 # define CLOSEDIR(d) closedir(d)
119 # include <utimens.h>
122 #define RW_USER (S_IRUSR | S_IWUSR) /* creation mode for open() */
125 # define MAX_PATH_LEN 1024 /* max pathname length */
137 off_t lseek OF((int fd, off_t offset, int whence));
141 #define OFF_T_MIN (~ (off_t) 0 << (sizeof (off_t) * CHAR_BIT - 1))
145 #define OFF_T_MAX (~ (off_t) 0 - OFF_T_MIN)
148 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
151 # define SA_NOCLDSTOP 0
152 # define sigprocmask(how, set, oset) /* empty */
153 # define sigset_t int
154 # if ! HAVE_SIGINTERRUPT
155 # define siginterrupt(sig, flag) /* empty */
159 #ifndef HAVE_WORKING_O_NOFOLLOW
160 # define HAVE_WORKING_O_NOFOLLOW 0
164 # define ELOOP EINVAL
167 /* Separator for file name parts (see shorten_name()) */
168 #ifdef NO_MULTIPLE_DOTS
169 # define PART_SEP "-"
171 # define PART_SEP "."
176 DECLARE(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
177 DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
178 DECLARE(ush, d_buf, DIST_BUFSIZE);
179 DECLARE(uch, window, 2L*WSIZE);
181 DECLARE(ush, tab_prefix, 1L<<BITS);
183 DECLARE(ush, tab_prefix0, 1L<<(BITS-1));
184 DECLARE(ush, tab_prefix1, 1L<<(BITS-1));
187 /* local variables */
189 int ascii = 0; /* convert end-of-lines to local OS conventions */
190 int to_stdout = 0; /* output to stdout (-c) */
191 int decompress = 0; /* decompress (-d) */
192 int force = 0; /* don't ask questions, compress links (-f) */
193 int no_name = -1; /* don't save or restore the original file name */
194 int no_time = -1; /* don't save or restore the original file time */
195 int recursive = 0; /* recurse through directories (-r) */
196 int list = 0; /* list the file contents (-l) */
197 int verbose = 0; /* be verbose (-v) */
198 int quiet = 0; /* be very quiet (-q) */
199 int do_lzw = 0; /* generate output compatible with old compress (-Z) */
200 int test = 0; /* test .gz file integrity */
201 int foreground; /* set if program run in foreground */
202 char *program_name; /* program name */
203 int maxbits = BITS; /* max bits per code for LZW */
204 int method = DEFLATED;/* compression method */
205 int level = 6; /* compression level */
206 int exit_code = OK; /* program exit code */
207 int save_orig_name; /* set if original name must be saved */
208 int last_member; /* set for .zip and .Z files */
209 int part_nb; /* number of parts in .gz file */
210 struct timespec time_stamp; /* original time stamp (modification time) */
211 off_t ifile_size; /* input file size, -1 for devices (debug only) */
212 char *env; /* contents of GZIP env variable */
213 char **args = NULL; /* argv pointer if GZIP env variable defined */
214 char *z_suffix; /* default suffix (can be set with --suffix) */
215 size_t z_len; /* strlen(z_suffix) */
217 /* The set of signals that are caught. */
218 static sigset_t caught_signals;
220 /* If nonzero then exit with status 1, rather than with the usual
221 signal status, on receipt of a signal with this value. This
222 suppresses a "Broken Pipe" message with some shells. */
223 static int volatile exiting_signal;
225 /* If nonnegative, close this file descriptor and unlink ofname on error. */
226 static int volatile remove_ofname_fd = -1;
228 off_t bytes_in; /* number of input bytes */
229 off_t bytes_out; /* number of output bytes */
230 off_t total_in; /* input bytes for all files */
231 off_t total_out; /* output bytes for all files */
232 char ifname[MAX_PATH_LEN]; /* input file name */
233 char ofname[MAX_PATH_LEN]; /* output file name */
234 struct stat istat; /* status for input file */
235 int ifd; /* input file descriptor */
236 int ofd; /* output file descriptor */
237 unsigned insize; /* valid bytes in inbuf */
238 unsigned inptr; /* index of next byte to be processed in inbuf */
239 unsigned outcnt; /* bytes in output buffer */
241 struct option longopts[] =
243 /* { name has_arg *flag val } */
244 {"ascii", 0, 0, 'a'}, /* ascii text mode */
245 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
246 {"stdout", 0, 0, 'c'}, /* write output on standard output */
247 {"decompress", 0, 0, 'd'}, /* decompress */
248 {"uncompress", 0, 0, 'd'}, /* decompress */
249 /* {"encrypt", 0, 0, 'e'}, encrypt */
250 {"force", 0, 0, 'f'}, /* force overwrite of output file */
251 {"help", 0, 0, 'h'}, /* give help */
252 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
253 {"list", 0, 0, 'l'}, /* list .gz file contents */
254 {"license", 0, 0, 'L'}, /* display software license */
255 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
256 {"name", 0, 0, 'N'}, /* save or restore original name & time */
257 {"quiet", 0, 0, 'q'}, /* quiet mode */
258 {"silent", 0, 0, 'q'}, /* quiet mode */
259 {"recursive", 0, 0, 'r'}, /* recurse through directories */
260 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
261 {"test", 0, 0, 't'}, /* test compressed file integrity */
262 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
263 {"verbose", 0, 0, 'v'}, /* verbose mode */
264 {"version", 0, 0, 'V'}, /* display version number */
265 {"fast", 0, 0, '1'}, /* compress faster */
266 {"best", 0, 0, '9'}, /* compress better */
267 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
268 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
272 /* local functions */
274 local void try_help OF((void)) ATTRIBUTE_NORETURN;
275 local void help OF((void));
276 local void license OF((void));
277 local void version OF((void));
278 local int input_eof OF((void));
279 local void treat_stdin OF((void));
280 local void treat_file OF((char *iname));
281 local int create_outfile OF((void));
282 local char *get_suffix OF((char *name));
283 local int open_input_file OF((char *iname, struct stat *sbuf));
284 local int make_ofname OF((void));
285 local void shorten_name OF((char *name));
286 local int get_method OF((int in));
287 local void do_list OF((int ifd, int method));
288 local int check_ofname OF((void));
289 local void copy_stat OF((struct stat *ifstat));
290 local void install_signal_handlers OF((void));
291 local void remove_output_file OF((void));
292 local RETSIGTYPE abort_gzip_signal OF((int));
293 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
294 int main OF((int argc, char **argv));
295 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
298 local void treat_dir OF((int fd, char *dir));
301 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
306 fprintf (stderr, "Try `%s --help' for more information.\n",
311 /* ======================================================================== */
314 static char *help_msg[] = {
315 "Compress or uncompress FILEs (by default, compress FILES in-place).",
317 "Mandatory arguments to long options are mandatory for short options too.",
320 " -a, --ascii ascii text; convert end-of-line using local conventions",
322 " -c, --stdout write on standard output, keep original files unchanged",
323 " -d, --decompress decompress",
324 /* -e, --encrypt encrypt */
325 " -f, --force force overwrite of output file and compress links",
326 " -h, --help give this help",
327 /* -k, --pkzip force output in pkzip format */
328 " -l, --list list compressed file contents",
329 " -L, --license display software license",
331 " -m, --no-time do not save or restore the original modification time",
332 " -M, --time save or restore the original modification time",
334 " -n, --no-name do not save or restore the original name and time stamp",
335 " -N, --name save or restore the original name and time stamp",
336 " -q, --quiet suppress all warnings",
338 " -r, --recursive operate recursively on directories",
340 " -S, --suffix=SUF use suffix SUF on compressed files",
341 " -t, --test test compressed file integrity",
342 " -v, --verbose verbose mode",
343 " -V, --version display version number",
344 " -1, --fast compress faster",
345 " -9, --best compress better",
347 " -Z, --lzw produce output compatible with old compress",
348 " -b, --bits=BITS max number of bits per code (implies -Z)",
351 "With no FILE, or when FILE is -, read standard input.",
353 "Report bugs to <bug-gzip@gnu.org>.",
357 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
358 while (*p) printf ("%s\n", *p++);
361 /* ======================================================================== */
364 char **p = license_msg;
366 printf ("%s %s\n", program_name, VERSION);
367 while (*p) printf ("%s\n", *p++);
370 /* ======================================================================== */
375 printf ("Written by Jean-loup Gailly.\n");
378 local void progerror (string)
382 fprintf (stderr, "%s: ", program_name);
388 /* ======================================================================== */
389 int main (argc, argv)
393 int file_count; /* number of files to process */
394 size_t proglen; /* length of program_name */
395 int optc; /* current option */
397 EXPAND(argc, argv); /* wild card expansion if necessary */
399 program_name = gzip_base_name (argv[0]);
400 proglen = strlen (program_name);
402 /* Suppress .exe for MSDOS, OS/2 and VMS: */
403 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
404 program_name[proglen - 4] = '\0';
406 /* Add options in GZIP environment variable if there is one */
407 env = add_envopt(&argc, &argv, OPTIONS_VAR);
408 if (env != NULL) args = argv;
411 /* For compatibility with old compress, use program name as an option.
412 * If you compile with -DGNU_STANDARD, this program will behave as
413 * gzip even if it is invoked under the name gunzip or zcat.
415 * Systems which do not support links can still use -d or -dc.
416 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
418 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
419 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
421 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
422 || strequ (program_name, "gzcat")) /* gzcat */
423 decompress = to_stdout = 1;
427 z_len = strlen(z_suffix);
429 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
430 longopts, (int *)0)) != -1) {
435 maxbits = atoi(optarg);
436 for (; *optarg; optarg++)
437 if (! ('0' <= *optarg && *optarg <= '9'))
439 fprintf (stderr, "%s: -b operand is not an integer\n",
445 to_stdout = 1; break;
447 decompress = 1; break;
451 help(); do_exit(OK); break;
453 list = decompress = to_stdout = 1; break;
455 license(); do_exit(OK); break;
456 case 'm': /* undocumented, may change later */
458 case 'M': /* undocumented, may change later */
461 no_name = no_time = 1; break;
463 no_name = no_time = 0; break;
465 quiet = 1; verbose = 0; break;
468 fprintf (stderr, "%s: -r not supported on this system\n",
476 #ifdef NO_MULTIPLE_DOTS
477 if (*optarg == '.') optarg++;
479 z_len = strlen(optarg);
483 test = decompress = to_stdout = 1;
486 verbose++; quiet = 0; break;
488 version(); do_exit(OK); break;
493 fprintf(stderr, "%s: -Z not supported in this version\n",
498 case '1': case '2': case '3': case '4':
499 case '5': case '6': case '7': case '8': case '9':
503 /* Error message already emitted by getopt_long. */
506 } /* loop on all arguments */
508 /* By default, save name and timestamp on compression but do not
509 * restore them on decompression.
511 if (no_time < 0) no_time = decompress;
512 if (no_name < 0) no_name = decompress;
514 file_count = argc - optind;
518 if (ascii && !quiet) {
519 fprintf(stderr, "%s: option --ascii ignored on this system\n",
523 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
524 fprintf(stderr, "%s: incorrect suffix '%s'\n",
525 program_name, z_suffix);
528 if (do_lzw && !decompress) work = lzw;
530 /* Allocate all global buffers (for DYN_ALLOC option) */
531 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
532 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
533 ALLOC(ush, d_buf, DIST_BUFSIZE);
534 ALLOC(uch, window, 2L*WSIZE);
536 ALLOC(ush, tab_prefix, 1L<<BITS);
538 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
539 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
542 exiting_signal = quiet ? SIGPIPE : 0;
543 install_signal_handlers ();
545 /* And get to work */
546 if (file_count != 0) {
547 if (to_stdout && !test && !list && (!decompress || !ascii)) {
548 SET_BINARY_MODE(fileno(stdout));
550 while (optind < argc) {
551 treat_file(argv[optind++]);
553 } else { /* Standard input */
556 if (list && !quiet && file_count > 1) {
557 do_list(-1, -1); /* print totals */
560 return exit_code; /* just to avoid lint warning */
563 /* Return nonzero when at end of file on input. */
567 if (!decompress || last_member)
572 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
575 /* Unget the char that fill_inbuf got. */
582 /* ========================================================================
583 * Compress or decompress stdin
585 local void treat_stdin()
587 if (!force && !list &&
588 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
589 /* Do not send compressed data to the terminal or read it from
590 * the terminal. We get here when user invoked the program
591 * without parameters, so be helpful. According to the GNU standards:
593 * If there is one behavior you think is most useful when the output
594 * is to a terminal, and another that you think is most useful when
595 * the output is a file or a pipe, then it is usually best to make
596 * the default behavior the one that is useful with output to a
597 * terminal, and have an option for the other behavior.
599 * Here we use the --force option to get the other behavior.
602 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
603 program_name, decompress ? "read from" : "written to",
604 decompress ? "de" : "");
605 fprintf (stderr, "For help, type: %s -h\n", program_name);
609 if (decompress || !ascii) {
610 SET_BINARY_MODE(fileno(stdin));
612 if (!test && !list && (!decompress || !ascii)) {
613 SET_BINARY_MODE(fileno(stdout));
615 strcpy(ifname, "stdin");
616 strcpy(ofname, "stdout");
618 /* Get the time stamp on the input file. */
619 time_stamp.tv_nsec = -1; /* The time is unknown by default. */
621 #ifndef NO_STDIN_FSTAT
622 if (list || !no_time) {
623 if (fstat(fileno(stdin), &istat) != 0) {
624 progerror("standard input");
627 # ifdef NO_PIPE_TIMESTAMP
628 if (S_ISREG(istat.st_mode))
630 time_stamp = get_stat_mtime (&istat);
631 #endif /* NO_STDIN_FSTAT */
633 ifile_size = -1L; /* convention for unknown size */
635 clear_bufs(); /* clear input and output buffers */
640 method = get_method(ifd);
642 do_exit(exit_code); /* error message already emitted */
646 do_list(ifd, method);
650 /* Actually do the compression/decompression. Loop over zipped members.
653 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
658 method = get_method(ifd);
659 if (method < 0) return; /* error message already emitted */
660 bytes_out = 0; /* required for length check */
665 fprintf(stderr, " OK\n");
667 } else if (!decompress) {
668 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
669 fprintf(stderr, "\n");
670 #ifdef DISPLAY_STDIN_RATIO
672 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
673 fprintf(stderr, "\n");
679 /* ========================================================================
680 * Compress or decompress the given file
682 local void treat_file(iname)
685 /* Accept "-" as synonym for stdin */
686 if (strequ(iname, "-")) {
687 int cflag = to_stdout;
693 /* Check if the input file is present, set ifname and istat: */
694 ifd = open_input_file (iname, &istat);
698 /* If the input name is that of a directory, recurse or ignore: */
699 if (S_ISDIR(istat.st_mode)) {
702 treat_dir (ifd, iname);
703 /* Warning: ifname is now garbage */
708 WARN ((stderr, "%s: %s is a directory -- ignored\n",
709 program_name, ifname));
712 if (!S_ISREG(istat.st_mode)) {
714 "%s: %s is not a directory or a regular file - ignored\n",
715 program_name, ifname));
720 if (istat.st_mode & S_ISUID)
722 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
723 program_name, ifname));
727 if (istat.st_mode & S_ISGID)
729 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
730 program_name, ifname));
734 if (istat.st_mode & S_ISVTX)
736 WARN ((stderr, "%s: %s has the sticky bit set - file ignored\n",
737 program_name, ifname));
742 if (istat.st_nlink > 1 && !to_stdout && !force) {
743 WARN((stderr, "%s: %s has %lu other link%c -- unchanged\n",
744 program_name, ifname, (unsigned long) istat.st_nlink - 1,
745 istat.st_nlink > 2 ? 's' : ' '));
750 ifile_size = istat.st_size;
751 if (no_time && !list)
752 time_stamp.tv_nsec = -1;
754 time_stamp = get_stat_mtime (&istat);
756 /* Generate output file name. For -r and (-t or -l), skip files
757 * without a valid gzip suffix (check done in make_ofname).
759 if (to_stdout && !list && !test) {
760 strcpy(ofname, "stdout");
762 } else if (make_ofname() != OK) {
767 clear_bufs(); /* clear input and output buffers */
771 method = get_method(ifd); /* updates ofname if original given */
774 return; /* error message already emitted */
778 do_list(ifd, method);
779 if (close (ifd) != 0)
784 /* If compressing to a file, check if ofname is not ambiguous
785 * because the operating system truncates names. Otherwise, generate
786 * a new ofname and save the original name in the compressed file.
789 ofd = fileno(stdout);
790 /* Keep remove_ofname_fd negative. */
792 if (create_outfile() != OK) return;
794 if (!decompress && save_orig_name && !verbose && !quiet) {
795 fprintf(stderr, "%s: %s compressed to %s\n",
796 program_name, ifname, ofname);
799 /* Keep the name even if not truncated except with --no-name: */
800 if (!save_orig_name) save_orig_name = !no_name;
803 fprintf(stderr, "%s:\t", ifname);
806 /* Actually do the compression/decompression. Loop over zipped members.
809 if ((*work)(ifd, ofd) != OK) {
810 method = -1; /* force cleanup */
817 method = get_method(ifd);
818 if (method < 0) break; /* error message already emitted */
819 bytes_out = 0; /* required for length check */
822 if (close (ifd) != 0)
831 if (close (ofd) != 0)
834 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
835 remove_ofname_fd = -1;
836 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
837 sigprocmask (SIG_SETMASK, &oldset, NULL);
841 WARN ((stderr, "%s: ", program_name));
844 errno = unlink_errno;
852 remove_output_file ();
856 /* Display statistics */
859 fprintf(stderr, " OK");
860 } else if (decompress) {
861 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
863 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
865 if (!test && !to_stdout) {
866 fprintf(stderr, " -- replaced with %s", ofname);
868 fprintf(stderr, "\n");
872 /* ========================================================================
873 * Create the output file. Return OK or ERROR.
874 * Try several times if necessary to avoid truncating the z_suffix. For
875 * example, do not create a compressed file of name "1234567890123."
876 * Sets save_orig_name to true if the file name has been truncated.
877 * IN assertions: the input file has already been open (ifd is set) and
878 * ofname has already been updated if there was an original name.
879 * OUT assertions: ifd and ofd are closed in case of error.
881 local int create_outfile()
883 int name_shortened = 0;
884 int flags = (O_WRONLY | O_CREAT | O_EXCL
885 | (ascii && decompress ? 0 : O_BINARY));
892 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
893 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
895 sigprocmask (SIG_SETMASK, &oldset, NULL);
904 shorten_name (ofname);
910 if (check_ofname () != OK)
924 if (name_shortened && decompress)
926 /* name might be too long if an original name was saved */
927 WARN ((stderr, "%s: %s: warning, name truncated\n",
928 program_name, ofname));
934 /* ========================================================================
935 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
936 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
937 * accepted suffixes, in addition to the value of the --suffix option.
938 * ".tgz" is a useful convention for tar.z files on systems limited
939 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
940 * also accepted suffixes. For Unix, we do not want to accept any
941 * .??z suffix as indicating a compressed file; some people use .xyz
942 * to denote volume data.
943 * On systems allowing multiple versions of the same file (such as VMS),
944 * this function removes any version suffix in the given name.
946 local char *get_suffix(name)
950 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
951 static char *known_suffixes[] =
952 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
957 char **suf = known_suffixes;
960 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
963 /* strip a version number from the file name */
965 char *v = strrchr(name, SUFFIX_SEP);
966 if (v != NULL) *v = '\0';
970 if (nlen <= MAX_SUFFIX+2) {
971 strcpy(suffix, name);
973 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
976 slen = strlen(suffix);
978 int s = strlen(*suf);
979 if (slen > s && suffix[slen-s-1] != PATH_SEP
980 && strequ(suffix + slen - s, *suf)) {
983 } while (*++suf != NULL);
989 /* Open file NAME with the given flags and mode and store its status
990 into *ST. Return a file descriptor to the newly opened file, or -1
991 (setting errno) on failure. */
993 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
997 /* Refuse to follow symbolic links unless -c or -f. */
998 if (!to_stdout && !force)
1000 if (HAVE_WORKING_O_NOFOLLOW)
1001 flags |= O_NOFOLLOW;
1004 #if HAVE_LSTAT || defined lstat
1005 if (lstat (name, st) != 0)
1007 else if (S_ISLNK (st->st_mode))
1016 fd = OPEN (name, flags, mode);
1017 if (0 <= fd && fstat (fd, st) != 0)
1028 /* ========================================================================
1029 * Set ifname to the input file name (with a suffix appended if necessary)
1030 * and istat to its stats. For decompression, if no file exists with the
1031 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1032 * For MSDOS, we try only z_suffix and z.
1033 * Return an open file descriptor or -1.
1036 open_input_file (iname, sbuf)
1040 int ilen; /* strlen(ifname) */
1041 int z_suffix_errno = 0;
1042 static char *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1043 char **suf = suffixes;
1045 #ifdef NO_MULTIPLE_DOTS
1046 char *dot; /* pointer to ifname extension, or NULL */
1049 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1050 | (ascii && !decompress ? 0 : O_BINARY));
1054 if (sizeof ifname - 1 <= strlen (iname))
1057 strcpy(ifname, iname);
1059 /* If input file exists, return OK. */
1060 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1064 if (!decompress || errno != ENOENT) {
1068 /* file.ext doesn't exist, try adding a suffix (after removing any
1069 * version number for VMS).
1071 s = get_suffix(ifname);
1073 progerror(ifname); /* ifname already has z suffix and does not exist */
1076 #ifdef NO_MULTIPLE_DOTS
1077 dot = strrchr(ifname, '.');
1079 strcat(ifname, ".");
1080 dot = strrchr(ifname, '.');
1083 ilen = strlen(ifname);
1084 if (strequ(z_suffix, ".gz")) suf++;
1086 /* Search for all suffixes */
1088 char *s0 = s = *suf;
1089 strcpy (ifname, iname);
1090 #ifdef NO_MULTIPLE_DOTS
1092 if (*dot == '\0') strcpy (dot, ".");
1094 #ifdef MAX_EXT_CHARS
1095 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1096 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1098 if (sizeof ifname <= ilen + strlen (s))
1101 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1104 if (errno != ENOENT)
1109 if (strequ (s0, z_suffix))
1110 z_suffix_errno = errno;
1111 } while (*++suf != NULL);
1113 /* No suffix found, complain using z_suffix: */
1114 strcpy(ifname, iname);
1115 #ifdef NO_MULTIPLE_DOTS
1116 if (*dot == '\0') strcpy(dot, ".");
1118 #ifdef MAX_EXT_CHARS
1119 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1120 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1122 strcat(ifname, z_suffix);
1123 errno = z_suffix_errno;
1128 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1133 /* ========================================================================
1134 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1135 * Sets save_orig_name to true if the file name has been truncated.
1137 local int make_ofname()
1139 char *suff; /* ofname z suffix */
1141 strcpy(ofname, ifname);
1142 /* strip a version number if any and get the gzip suffix if present: */
1143 suff = get_suffix(ofname);
1147 /* With -t or -l, try all files (even without .gz suffix)
1148 * except with -r (behave as with just -dr).
1150 if (!recursive && (list || test)) return OK;
1152 /* Avoid annoying messages with -r */
1153 if (verbose || (!recursive && !quiet)) {
1154 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1155 program_name, ifname));
1159 /* Make a special case for .tgz and .taz: */
1161 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1162 strcpy(suff, ".tar");
1164 *suff = '\0'; /* strip the z suffix */
1166 /* ofname might be changed later if infile contains an original name */
1168 } else if (suff != NULL) {
1169 /* Avoid annoying messages with -r (see treat_dir()) */
1170 if (verbose || (!recursive && !quiet)) {
1171 /* Don't use WARN, as it affects exit status. */
1172 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1173 program_name, ifname, suff);
1179 #ifdef NO_MULTIPLE_DOTS
1180 suff = strrchr(ofname, '.');
1182 if (sizeof ofname <= strlen (ofname) + 1)
1184 strcat(ofname, ".");
1185 # ifdef MAX_EXT_CHARS
1186 if (strequ(z_suffix, "z")) {
1187 if (sizeof ofname <= strlen (ofname) + 2)
1189 strcat(ofname, "gz"); /* enough room */
1192 /* On the Atari and some versions of MSDOS,
1193 * ENAMETOOLONG does not work correctly. So we
1194 * must truncate here.
1196 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1197 suff[MAX_SUFFIX+1-z_len] = '\0';
1201 #endif /* NO_MULTIPLE_DOTS */
1202 if (sizeof ofname <= strlen (ofname) + z_len)
1204 strcat(ofname, z_suffix);
1206 } /* decompress ? */
1210 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1215 /* ========================================================================
1216 * Check the magic number of the input file and update ofname if an
1217 * original name was given and to_stdout is not set.
1218 * Return the compression method, -1 for error, -2 for warning.
1219 * Set inptr to the offset of the next byte to be processed.
1220 * Updates time_stamp if there is one and --no-time is not used.
1221 * This function may be called repeatedly for an input file consisting
1222 * of several contiguous gzip'ed members.
1223 * IN assertions: there is at least one remaining compressed member.
1224 * If the member is a zip file, it must be the only one.
1226 local int get_method(in)
1227 int in; /* input file descriptor */
1229 uch flags; /* compression flags */
1230 char magic[2]; /* magic header */
1231 int imagic1; /* like magic[1], but can represent EOF */
1232 ulg stamp; /* time stamp */
1234 /* If --force and --stdout, zcat == cat, so do not complain about
1235 * premature end of file: use try_byte instead of get_byte.
1237 if (force && to_stdout) {
1238 magic[0] = (char)try_byte();
1239 imagic1 = try_byte ();
1240 magic[1] = (char) imagic1;
1241 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1243 magic[0] = (char)get_byte();
1244 magic[1] = (char)get_byte();
1245 imagic1 = 0; /* avoid lint warning */
1247 method = -1; /* unknown yet */
1248 part_nb++; /* number of parts in gzip file */
1250 last_member = RECORD_IO;
1251 /* assume multiple members in gzip file except for record oriented I/O */
1253 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1254 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1256 method = (int)get_byte();
1257 if (method != DEFLATED) {
1259 "%s: %s: unknown method %d -- not supported\n",
1260 program_name, ifname, method);
1265 flags = (uch)get_byte();
1267 if ((flags & ENCRYPTED) != 0) {
1269 "%s: %s is encrypted -- not supported\n",
1270 program_name, ifname);
1274 if ((flags & CONTINUATION) != 0) {
1276 "%s: %s is a a multi-part gzip file -- not supported\n",
1277 program_name, ifname);
1279 if (force <= 1) return -1;
1281 if ((flags & RESERVED) != 0) {
1283 "%s: %s has flags 0x%x -- not supported\n",
1284 program_name, ifname, flags);
1286 if (force <= 1) return -1;
1288 stamp = (ulg)get_byte();
1289 stamp |= ((ulg)get_byte()) << 8;
1290 stamp |= ((ulg)get_byte()) << 16;
1291 stamp |= ((ulg)get_byte()) << 24;
1292 if (stamp != 0 && !no_time)
1294 time_stamp.tv_sec = stamp;
1295 time_stamp.tv_nsec = 0;
1298 (void)get_byte(); /* Ignore extra flags for the moment */
1299 (void)get_byte(); /* Ignore OS type for the moment */
1301 if ((flags & CONTINUATION) != 0) {
1302 unsigned part = (unsigned)get_byte();
1303 part |= ((unsigned)get_byte())<<8;
1305 fprintf(stderr,"%s: %s: part number %u\n",
1306 program_name, ifname, part);
1309 if ((flags & EXTRA_FIELD) != 0) {
1310 unsigned len = (unsigned)get_byte();
1311 len |= ((unsigned)get_byte())<<8;
1313 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1314 program_name, ifname, len);
1316 while (len--) (void)get_byte();
1319 /* Get original file name if it was truncated */
1320 if ((flags & ORIG_NAME) != 0) {
1321 if (no_name || (to_stdout && !list) || part_nb > 1) {
1322 /* Discard the old name */
1323 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1324 do {c=get_byte();} while (c != 0);
1326 /* Copy the base name. Keep a directory prefix intact. */
1327 char *p = gzip_base_name (ofname);
1330 *p = (char)get_char();
1331 if (*p++ == '\0') break;
1332 if (p >= ofname+sizeof(ofname)) {
1333 gzip_error ("corrupted input -- file name too large");
1336 p = gzip_base_name (base);
1337 memmove (base, p, strlen (p) + 1);
1338 /* If necessary, adapt the name to local OS conventions: */
1340 MAKE_LEGAL_NAME(base);
1341 if (base) list=0; /* avoid warning about unused variable */
1343 } /* no_name || to_stdout */
1346 /* Discard file comment if any */
1347 if ((flags & COMMENT) != 0) {
1348 while (get_char() != 0) /* null */ ;
1351 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1354 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1355 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1356 /* To simplify the code, we support a zip file when alone only.
1357 * We are thus guaranteed that the entire local header fits in inbuf.
1361 if (check_zipfile(in) != OK) return -1;
1362 /* check_zipfile may get ofname from the local header */
1365 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1369 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1371 method = COMPRESSED;
1374 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1379 } else if (force && to_stdout && !list) { /* pass input unchanged */
1385 if (method >= 0) return method;
1388 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1389 program_name, ifname);
1396 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1401 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1402 program_name, ifname));
1407 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1408 program_name, ifname));
1413 /* ========================================================================
1414 * Display the characteristics of the compressed file.
1415 * If the given method is < 0, display the accumulated totals.
1416 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1418 local void do_list(ifd, method)
1419 int ifd; /* input file descriptor */
1420 int method; /* compression method */
1422 ulg crc; /* original crc */
1423 static int first_time = 1;
1424 static char* methods[MAX_METHODS] = {
1429 "", "", "", "", /* 4 to 7 reserved */
1431 int positive_off_t_width = 1;
1434 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1435 positive_off_t_width++;
1438 if (first_time && method >= 0) {
1441 printf("method crc date time ");
1444 printf("%*.*s %*.*s ratio uncompressed_name\n",
1445 positive_off_t_width, positive_off_t_width, "compressed",
1446 positive_off_t_width, positive_off_t_width, "uncompressed");
1448 } else if (method < 0) {
1449 if (total_in <= 0 || total_out <= 0) return;
1453 if (verbose || !quiet) {
1454 fprint_off(stdout, total_in, positive_off_t_width);
1456 fprint_off(stdout, total_out, positive_off_t_width);
1459 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1460 /* header_bytes is not meaningful but used to ensure the same
1461 * ratio if there is a single file.
1463 printf(" (totals)\n");
1466 crc = (ulg)~0; /* unknown */
1468 bytes_in = ifile_size;
1471 if (method == DEFLATED && !last_member) {
1472 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1473 * If the lseek fails, we could use read() to get to the end, but
1474 * --list is used to get quick results.
1475 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1476 * you are not concerned about speed.
1478 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1479 if (bytes_in != -1L) {
1482 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1486 bytes_out = LG(buf+4);
1489 #endif /* RECORD_IO */
1492 struct tm *tm = localtime (&time_stamp.tv_sec);
1493 printf ("%5s %08lx ", methods[method], crc);
1495 printf ("%s%3d %02d:%02d ",
1496 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1498 tm->tm_mday, tm->tm_hour, tm->tm_min);
1500 printf ("??? ?? ??:?? ");
1502 fprint_off(stdout, bytes_in, positive_off_t_width);
1504 fprint_off(stdout, bytes_out, positive_off_t_width);
1506 if (bytes_in == -1L) {
1508 bytes_in = bytes_out = header_bytes = 0;
1509 } else if (total_in >= 0) {
1510 total_in += bytes_in;
1512 if (bytes_out == -1L) {
1514 bytes_in = bytes_out = header_bytes = 0;
1515 } else if (total_out >= 0) {
1516 total_out += bytes_out;
1518 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1519 printf(" %s\n", ofname);
1522 /* ========================================================================
1523 * Shorten the given name by one character, or replace a .tar extension
1524 * with .tgz. Truncate the last part of the name which is longer than
1525 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1526 * has only parts shorter than MIN_PART truncate the longest part.
1527 * For decompression, just remove the last character of the name.
1529 * IN assertion: for compression, the suffix of the given name is z_suffix.
1531 local void shorten_name(name)
1534 int len; /* length of name without z_suffix */
1535 char *trunc = NULL; /* character to be truncated */
1536 int plen; /* current part length */
1537 int min_part = MIN_PART; /* current minimum part length */
1543 gzip_error ("name too short");
1547 p = get_suffix(name);
1549 gzip_error ("can't recover suffix\n");
1553 /* compress 1234567890.tar to 1234567890.tgz */
1554 if (len > 4 && strequ(p-4, ".tar")) {
1555 strcpy(p-4, ".tgz");
1558 /* Try keeping short extensions intact:
1559 * 1234.678.012.gz -> 123.678.012.gz
1562 p = strrchr(name, PATH_SEP);
1565 plen = strcspn(p, PART_SEP);
1567 if (plen > min_part) trunc = p-1;
1570 } while (trunc == NULL && --min_part != 0);
1572 if (trunc != NULL) {
1574 trunc[0] = trunc[1];
1578 trunc = strrchr(name, PART_SEP[0]);
1580 gzip_error ("internal error in shorten_name");
1581 if (trunc[1] == '\0') trunc--; /* force truncation */
1583 strcpy(trunc, z_suffix);
1586 /* ========================================================================
1587 * The compressed file already exists, so ask for confirmation.
1588 * Return ERROR if the file must be skipped.
1590 local int check_ofname()
1592 /* Ask permission to overwrite the existing file */
1595 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1596 if (foreground && isatty(fileno(stdin))) {
1597 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1602 fprintf(stderr, "\tnot overwritten\n");
1603 if (exit_code == OK) exit_code = WARNING;
1607 if (xunlink (ofname)) {
1615 /* ========================================================================
1616 * Copy modes, times, ownership from input file to output file.
1617 * IN assertion: to_stdout is false.
1619 local void copy_stat(ifstat)
1620 struct stat *ifstat;
1622 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1626 struct timespec timespec[2];
1627 timespec[0] = get_stat_atime (ifstat);
1628 timespec[1] = get_stat_mtime (ifstat);
1630 if (decompress && 0 <= time_stamp.tv_nsec
1631 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1632 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1634 timespec[1] = time_stamp;
1636 fprintf(stderr, "%s: time stamp restored\n", ofname);
1640 if (futimens (ofd, ofname, timespec) != 0)
1643 WARN ((stderr, "%s: ", program_name));
1654 fchown (ofd, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1656 chown(ofname, ifstat->st_uid, ifstat->st_gid); /* Copy ownership */
1660 /* Copy the protection modes */
1662 r = fchmod (ofd, mode);
1664 r = chmod (ofname, mode);
1668 WARN ((stderr, "%s: ", program_name));
1678 /* ========================================================================
1679 * Recurse through the given directory. This code is taken from ncompress.
1681 local void treat_dir (fd, dir)
1687 char nbuf[MAX_PATH_LEN];
1691 dirp = fdopendir (fd);
1694 dirp = opendir(dir);
1705 ** WARNING: the following algorithm could occasionally cause
1706 ** compress to produce error warnings of the form "<filename>.gz
1707 ** already has .gz suffix - ignored". This occurs when the
1708 ** .gz output file is inserted into the directory below
1709 ** readdir's current pointer.
1710 ** These warnings are harmless but annoying, so they are suppressed
1711 ** with option -r (except when -v is on). An alternative
1712 ** to allowing this would be to store the entire directory
1713 ** list in memory, then compress the entries in the stored
1714 ** list. Given the depth-first recursive algorithm used here,
1715 ** this could use up a tremendous amount of memory. I don't
1716 ** think it's worth it. -- Dave Mack
1717 ** (An other alternative might be two passes to avoid depth-first.)
1720 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1722 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1726 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1728 if (len != 0 /* dir = "" means current dir on Amiga */
1730 && dir[len-1] != PATH_SEP2
1733 && dir[len-1] != PATH_SEP3
1736 nbuf[len++] = PATH_SEP;
1738 strcpy(nbuf+len, dp->d_name);
1741 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1742 program_name, dir, dp->d_name);
1748 if (CLOSEDIR(dirp) != 0)
1751 #endif /* ! NO_DIR */
1753 /* Make sure signals get handled properly. */
1756 install_signal_handlers ()
1760 /* SIGINT must be first, as 'foreground' depends on it. */
1781 int nsigs = sizeof sig / sizeof sig[0];
1785 struct sigaction act;
1787 sigemptyset (&caught_signals);
1788 for (i = 0; i < nsigs; i++)
1790 sigaction (sig[i], NULL, &act);
1791 if (act.sa_handler != SIG_IGN)
1792 sigaddset (&caught_signals, sig[i]);
1795 act.sa_handler = abort_gzip_signal;
1796 act.sa_mask = caught_signals;
1799 for (i = 0; i < nsigs; i++)
1800 if (sigismember (&caught_signals, sig[i]))
1804 sigaction (sig[i], &act, NULL);
1807 for (i = 0; i < nsigs; i++)
1808 if (signal (sig[i], SIG_IGN) != SIG_IGN)
1812 signal (sig[i], abort_gzip_signal);
1813 siginterrupt (sig[i], 1);
1818 /* ========================================================================
1819 * Free all dynamically allocated variables and exit with the given code.
1821 local void do_exit(exitcode)
1824 static int in_exit = 0;
1826 if (in_exit) exit(exitcode);
1828 if (env != NULL) free(env), env = NULL;
1829 if (args != NULL) free((char*)args), args = NULL;
1843 /* ========================================================================
1844 * Close and unlink the output file.
1847 remove_output_file ()
1852 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1853 fd = remove_ofname_fd;
1856 remove_ofname_fd = -1;
1860 sigprocmask (SIG_SETMASK, &oldset, NULL);
1863 /* ========================================================================
1869 remove_output_file ();
1873 /* ========================================================================
1877 abort_gzip_signal (sig)
1881 signal (sig, SIG_IGN);
1882 remove_output_file ();
1883 if (sig == exiting_signal)
1885 signal (sig, SIG_DFL);