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 */
222 static int handled_sig[] =
224 /* SIGINT must be first, as 'foreground' depends on it. */
246 struct option longopts[] =
248 /* { name has_arg *flag val } */
249 {"ascii", 0, 0, 'a'}, /* ascii text mode */
250 {"to-stdout", 0, 0, 'c'}, /* write output on standard output */
251 {"stdout", 0, 0, 'c'}, /* write output on standard output */
252 {"decompress", 0, 0, 'd'}, /* decompress */
253 {"uncompress", 0, 0, 'd'}, /* decompress */
254 /* {"encrypt", 0, 0, 'e'}, encrypt */
255 {"force", 0, 0, 'f'}, /* force overwrite of output file */
256 {"help", 0, 0, 'h'}, /* give help */
257 /* {"pkzip", 0, 0, 'k'}, force output in pkzip format */
258 {"list", 0, 0, 'l'}, /* list .gz file contents */
259 {"license", 0, 0, 'L'}, /* display software license */
260 {"no-name", 0, 0, 'n'}, /* don't save or restore original name & time */
261 {"name", 0, 0, 'N'}, /* save or restore original name & time */
262 {"quiet", 0, 0, 'q'}, /* quiet mode */
263 {"silent", 0, 0, 'q'}, /* quiet mode */
264 {"recursive", 0, 0, 'r'}, /* recurse through directories */
265 {"suffix", 1, 0, 'S'}, /* use given suffix instead of .gz */
266 {"test", 0, 0, 't'}, /* test compressed file integrity */
267 {"no-time", 0, 0, 'T'}, /* don't save or restore the time stamp */
268 {"verbose", 0, 0, 'v'}, /* verbose mode */
269 {"version", 0, 0, 'V'}, /* display version number */
270 {"fast", 0, 0, '1'}, /* compress faster */
271 {"best", 0, 0, '9'}, /* compress better */
272 {"lzw", 0, 0, 'Z'}, /* make output compatible with old compress */
273 {"bits", 1, 0, 'b'}, /* max number of bits per code (implies -Z) */
277 /* local functions */
279 local void try_help OF((void)) ATTRIBUTE_NORETURN;
280 local void help OF((void));
281 local void license OF((void));
282 local void version OF((void));
283 local int input_eof OF((void));
284 local void treat_stdin OF((void));
285 local void treat_file OF((char *iname));
286 local int create_outfile OF((void));
287 local char *get_suffix OF((char *name));
288 local int open_input_file OF((char *iname, struct stat *sbuf));
289 local int make_ofname OF((void));
290 local void shorten_name OF((char *name));
291 local int get_method OF((int in));
292 local void do_list OF((int ifd, int method));
293 local int check_ofname OF((void));
294 local void copy_stat OF((struct stat *ifstat));
295 local void install_signal_handlers OF((void));
296 local void remove_output_file OF((void));
297 local RETSIGTYPE abort_gzip_signal OF((int));
298 local void do_exit OF((int exitcode)) ATTRIBUTE_NORETURN;
299 int main OF((int argc, char **argv));
300 int (*work) OF((int infile, int outfile)) = zip; /* function to call */
303 local void treat_dir OF((int fd, char *dir));
306 #define strequ(s1, s2) (strcmp((s1),(s2)) == 0)
311 fprintf (stderr, "Try `%s --help' for more information.\n",
316 /* ======================================================================== */
319 static char const* const help_msg[] = {
320 "Compress or uncompress FILEs (by default, compress FILES in-place).",
322 "Mandatory arguments to long options are mandatory for short options too.",
325 " -a, --ascii ascii text; convert end-of-line using local conventions",
327 " -c, --stdout write on standard output, keep original files unchanged",
328 " -d, --decompress decompress",
329 /* -e, --encrypt encrypt */
330 " -f, --force force overwrite of output file and compress links",
331 " -h, --help give this help",
332 /* -k, --pkzip force output in pkzip format */
333 " -l, --list list compressed file contents",
334 " -L, --license display software license",
336 " -m, --no-time do not save or restore the original modification time",
337 " -M, --time save or restore the original modification time",
339 " -n, --no-name do not save or restore the original name and time stamp",
340 " -N, --name save or restore the original name and time stamp",
341 " -q, --quiet suppress all warnings",
343 " -r, --recursive operate recursively on directories",
345 " -S, --suffix=SUF use suffix SUF on compressed files",
346 " -t, --test test compressed file integrity",
347 " -v, --verbose verbose mode",
348 " -V, --version display version number",
349 " -1, --fast compress faster",
350 " -9, --best compress better",
352 " -Z, --lzw produce output compatible with old compress",
353 " -b, --bits=BITS max number of bits per code (implies -Z)",
356 "With no FILE, or when FILE is -, read standard input.",
358 "Report bugs to <bug-gzip@gnu.org>.",
360 char const *const *p = help_msg;
362 printf ("Usage: %s [OPTION]... [FILE]...\n", program_name);
363 while (*p) printf ("%s\n", *p++);
366 /* ======================================================================== */
369 char const *const *p = license_msg;
371 printf ("%s %s\n", program_name, VERSION);
372 while (*p) printf ("%s\n", *p++);
375 /* ======================================================================== */
380 printf ("Written by Jean-loup Gailly.\n");
383 local void progerror (char const *string)
386 fprintf (stderr, "%s: ", program_name);
392 /* ======================================================================== */
393 int main (int argc, char **argv)
395 int file_count; /* number of files to process */
396 size_t proglen; /* length of program_name */
397 int optc; /* current option */
399 EXPAND(argc, argv); /* wild card expansion if necessary */
401 program_name = gzip_base_name (argv[0]);
402 proglen = strlen (program_name);
404 atexit (close_stdin);
406 /* Suppress .exe for MSDOS, OS/2 and VMS: */
407 if (4 < proglen && strequ (program_name + proglen - 4, ".exe"))
408 program_name[proglen - 4] = '\0';
410 /* Add options in GZIP environment variable if there is one */
411 env = add_envopt(&argc, &argv, OPTIONS_VAR);
412 if (env != NULL) args = argv;
415 # define GNU_STANDARD 1
418 /* For compatibility with old compress, use program name as an option.
419 * Unless you compile with -DGNU_STANDARD=0, this program will behave as
420 * gzip even if it is invoked under the name gunzip or zcat.
422 * Systems which do not support links can still use -d or -dc.
423 * Ignore an .exe extension for MSDOS, OS/2 and VMS.
425 if (strncmp (program_name, "un", 2) == 0 /* ungzip, uncompress */
426 || strncmp (program_name, "gun", 3) == 0) /* gunzip */
428 else if (strequ (program_name + 1, "cat") /* zcat, pcat, gcat */
429 || strequ (program_name, "gzcat")) /* gzcat */
430 decompress = to_stdout = 1;
434 z_len = strlen(z_suffix);
436 while ((optc = getopt_long (argc, argv, "ab:cdfhH?lLmMnNqrS:tvVZ123456789",
437 longopts, (int *)0)) != -1) {
442 maxbits = atoi(optarg);
443 for (; *optarg; optarg++)
444 if (! ('0' <= *optarg && *optarg <= '9'))
446 fprintf (stderr, "%s: -b operand is not an integer\n",
452 to_stdout = 1; break;
454 decompress = 1; break;
458 help(); do_exit(OK); break;
460 list = decompress = to_stdout = 1; break;
462 license(); do_exit(OK); break;
463 case 'm': /* undocumented, may change later */
465 case 'M': /* undocumented, may change later */
468 no_name = no_time = 1; break;
470 no_name = no_time = 0; break;
472 quiet = 1; verbose = 0; break;
475 fprintf (stderr, "%s: -r not supported on this system\n",
483 #ifdef NO_MULTIPLE_DOTS
484 if (*optarg == '.') optarg++;
486 z_len = strlen(optarg);
490 test = decompress = to_stdout = 1;
493 verbose++; quiet = 0; break;
495 version(); do_exit(OK); break;
500 fprintf(stderr, "%s: -Z not supported in this version\n",
505 case '1': case '2': case '3': case '4':
506 case '5': case '6': case '7': case '8': case '9':
510 /* Error message already emitted by getopt_long. */
513 } /* loop on all arguments */
515 /* By default, save name and timestamp on compression but do not
516 * restore them on decompression.
518 if (no_time < 0) no_time = decompress;
519 if (no_name < 0) no_name = decompress;
521 file_count = argc - optind;
525 if (ascii && !quiet) {
526 fprintf(stderr, "%s: option --ascii ignored on this system\n",
530 if ((z_len == 0 && !decompress) || z_len > MAX_SUFFIX) {
531 fprintf(stderr, "%s: incorrect suffix '%s'\n",
532 program_name, z_suffix);
535 if (do_lzw && !decompress) work = lzw;
537 /* Allocate all global buffers (for DYN_ALLOC option) */
538 ALLOC(uch, inbuf, INBUFSIZ +INBUF_EXTRA);
539 ALLOC(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
540 ALLOC(ush, d_buf, DIST_BUFSIZE);
541 ALLOC(uch, window, 2L*WSIZE);
543 ALLOC(ush, tab_prefix, 1L<<BITS);
545 ALLOC(ush, tab_prefix0, 1L<<(BITS-1));
546 ALLOC(ush, tab_prefix1, 1L<<(BITS-1));
549 exiting_signal = quiet ? SIGPIPE : 0;
550 install_signal_handlers ();
552 /* And get to work */
553 if (file_count != 0) {
554 if (to_stdout && !test && !list && (!decompress || !ascii)) {
555 SET_BINARY_MODE(fileno(stdout));
557 while (optind < argc) {
558 treat_file(argv[optind++]);
560 } else { /* Standard input */
563 if (list && !quiet && file_count > 1) {
564 do_list(-1, -1); /* print totals */
567 return exit_code; /* just to avoid lint warning */
570 /* Return nonzero when at end of file on input. */
574 if (!decompress || last_member)
579 if (insize != INBUFSIZ || fill_inbuf (1) == EOF)
582 /* Unget the char that fill_inbuf got. */
589 /* ========================================================================
590 * Compress or decompress stdin
592 local void treat_stdin()
594 if (!force && !list &&
595 isatty(fileno((FILE *)(decompress ? stdin : stdout)))) {
596 /* Do not send compressed data to the terminal or read it from
597 * the terminal. We get here when user invoked the program
598 * without parameters, so be helpful. According to the GNU standards:
600 * If there is one behavior you think is most useful when the output
601 * is to a terminal, and another that you think is most useful when
602 * the output is a file or a pipe, then it is usually best to make
603 * the default behavior the one that is useful with output to a
604 * terminal, and have an option for the other behavior.
606 * Here we use the --force option to get the other behavior.
609 "%s: compressed data not %s a terminal. Use -f to force %scompression.\n",
610 program_name, decompress ? "read from" : "written to",
611 decompress ? "de" : "");
612 fprintf (stderr, "For help, type: %s -h\n", program_name);
616 if (decompress || !ascii) {
617 SET_BINARY_MODE(fileno(stdin));
619 if (!test && !list && (!decompress || !ascii)) {
620 SET_BINARY_MODE(fileno(stdout));
622 strcpy(ifname, "stdin");
623 strcpy(ofname, "stdout");
625 /* Get the file's time stamp and size. */
626 if (fstat (fileno (stdin), &istat) != 0)
628 progerror ("standard input");
631 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
632 time_stamp.tv_nsec = -1;
633 if (!no_time || list)
634 time_stamp = get_stat_mtime (&istat);
636 clear_bufs(); /* clear input and output buffers */
642 method = get_method(ifd);
644 do_exit(exit_code); /* error message already emitted */
648 do_list(ifd, method);
652 /* Actually do the compression/decompression. Loop over zipped members.
655 if ((*work)(fileno(stdin), fileno(stdout)) != OK) return;
660 method = get_method(ifd);
661 if (method < 0) return; /* error message already emitted */
662 bytes_out = 0; /* required for length check */
667 fprintf(stderr, " OK\n");
669 } else if (!decompress) {
670 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
671 fprintf(stderr, "\n");
672 #ifdef DISPLAY_STDIN_RATIO
674 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
675 fprintf(stderr, "\n");
681 /* ========================================================================
682 * Compress or decompress the given file
684 local void treat_file(iname)
687 /* Accept "-" as synonym for stdin */
688 if (strequ(iname, "-")) {
689 int cflag = to_stdout;
695 /* Check if the input file is present, set ifname and istat: */
696 ifd = open_input_file (iname, &istat);
700 /* If the input name is that of a directory, recurse or ignore: */
701 if (S_ISDIR(istat.st_mode)) {
704 treat_dir (ifd, iname);
705 /* Warning: ifname is now garbage */
710 WARN ((stderr, "%s: %s is a directory -- ignored\n",
711 program_name, ifname));
717 if (! S_ISREG (istat.st_mode))
720 "%s: %s is not a directory or a regular file - ignored\n",
721 program_name, ifname));
725 if (istat.st_mode & S_ISUID)
727 WARN ((stderr, "%s: %s is set-user-ID on execution - ignored\n",
728 program_name, ifname));
732 if (istat.st_mode & S_ISGID)
734 WARN ((stderr, "%s: %s is set-group-ID on execution - ignored\n",
735 program_name, ifname));
742 if (istat.st_mode & S_ISVTX)
745 "%s: %s has the sticky bit set - file ignored\n",
746 program_name, ifname));
750 if (2 <= istat.st_nlink)
752 WARN ((stderr, "%s: %s has %lu other link%c -- unchanged\n",
753 program_name, ifname,
754 (unsigned long int) istat.st_nlink - 1,
755 istat.st_nlink == 2 ? ' ' : 's'));
762 ifile_size = S_ISREG (istat.st_mode) ? istat.st_size : -1;
763 time_stamp.tv_nsec = -1;
764 if (!no_time || list)
765 time_stamp = get_stat_mtime (&istat);
767 /* Generate output file name. For -r and (-t or -l), skip files
768 * without a valid gzip suffix (check done in make_ofname).
770 if (to_stdout && !list && !test) {
771 strcpy(ofname, "stdout");
773 } else if (make_ofname() != OK) {
778 clear_bufs(); /* clear input and output buffers */
782 method = get_method(ifd); /* updates ofname if original given */
785 return; /* error message already emitted */
789 do_list(ifd, method);
790 if (close (ifd) != 0)
795 /* If compressing to a file, check if ofname is not ambiguous
796 * because the operating system truncates names. Otherwise, generate
797 * a new ofname and save the original name in the compressed file.
800 ofd = fileno(stdout);
801 /* Keep remove_ofname_fd negative. */
803 if (create_outfile() != OK) return;
805 if (!decompress && save_orig_name && !verbose && !quiet) {
806 fprintf(stderr, "%s: %s compressed to %s\n",
807 program_name, ifname, ofname);
810 /* Keep the name even if not truncated except with --no-name: */
811 if (!save_orig_name) save_orig_name = !no_name;
814 fprintf(stderr, "%s:\t", ifname);
817 /* Actually do the compression/decompression. Loop over zipped members.
820 if ((*work)(ifd, ofd) != OK) {
821 method = -1; /* force cleanup */
828 method = get_method(ifd);
829 if (method < 0) break; /* error message already emitted */
830 bytes_out = 0; /* required for length check */
833 if (close (ifd) != 0)
842 if (close (ofd) != 0)
845 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
846 remove_ofname_fd = -1;
847 unlink_errno = xunlink (ifname) == 0 ? 0 : errno;
848 sigprocmask (SIG_SETMASK, &oldset, NULL);
852 WARN ((stderr, "%s: ", program_name));
855 errno = unlink_errno;
863 remove_output_file ();
867 /* Display statistics */
870 fprintf(stderr, " OK");
871 } else if (decompress) {
872 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out,stderr);
874 display_ratio(bytes_in-(bytes_out-header_bytes), bytes_in, stderr);
876 if (!test && !to_stdout) {
877 fprintf(stderr, " -- replaced with %s", ofname);
879 fprintf(stderr, "\n");
883 /* ========================================================================
884 * Create the output file. Return OK or ERROR.
885 * Try several times if necessary to avoid truncating the z_suffix. For
886 * example, do not create a compressed file of name "1234567890123."
887 * Sets save_orig_name to true if the file name has been truncated.
888 * IN assertions: the input file has already been open (ifd is set) and
889 * ofname has already been updated if there was an original name.
890 * OUT assertions: ifd and ofd are closed in case of error.
892 local int create_outfile()
894 int name_shortened = 0;
895 int flags = (O_WRONLY | O_CREAT | O_EXCL
896 | (ascii && decompress ? 0 : O_BINARY));
903 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
904 remove_ofname_fd = ofd = OPEN (ofname, flags, RW_USER);
906 sigprocmask (SIG_SETMASK, &oldset, NULL);
915 shorten_name (ofname);
921 if (check_ofname () != OK)
935 if (name_shortened && decompress)
937 /* name might be too long if an original name was saved */
938 WARN ((stderr, "%s: %s: warning, name truncated\n",
939 program_name, ofname));
945 /* ========================================================================
946 * Return a pointer to the 'z' suffix of a file name, or NULL. For all
947 * systems, ".gz", ".z", ".Z", ".taz", ".tgz", "-gz", "-z" and "_z" are
948 * accepted suffixes, in addition to the value of the --suffix option.
949 * ".tgz" is a useful convention for tar.z files on systems limited
950 * to 3 characters extensions. On such systems, ".?z" and ".??z" are
951 * also accepted suffixes. For Unix, we do not want to accept any
952 * .??z suffix as indicating a compressed file; some people use .xyz
953 * to denote volume data.
954 * On systems allowing multiple versions of the same file (such as VMS),
955 * this function removes any version suffix in the given name.
957 local char *get_suffix(name)
961 char suffix[MAX_SUFFIX+3]; /* last chars of name, forced to lower case */
962 static char const *known_suffixes[] =
963 {NULL, ".gz", ".z", ".taz", ".tgz", "-gz", "-z", "_z",
968 char const **suf = known_suffixes;
971 if (strequ(z_suffix, "z")) suf++; /* check long suffixes first */
974 /* strip a version number from the file name */
976 char *v = strrchr(name, SUFFIX_SEP);
977 if (v != NULL) *v = '\0';
981 if (nlen <= MAX_SUFFIX+2) {
982 strcpy(suffix, name);
984 strcpy(suffix, name+nlen-MAX_SUFFIX-2);
987 slen = strlen(suffix);
989 int s = strlen(*suf);
990 if (slen > s && suffix[slen-s-1] != PATH_SEP
991 && strequ(suffix + slen - s, *suf)) {
994 } while (*++suf != NULL);
1000 /* Open file NAME with the given flags and mode and store its status
1001 into *ST. Return a file descriptor to the newly opened file, or -1
1002 (setting errno) on failure. */
1004 open_and_stat (char *name, int flags, mode_t mode, struct stat *st)
1008 /* Refuse to follow symbolic links unless -c or -f. */
1009 if (!to_stdout && !force)
1011 if (HAVE_WORKING_O_NOFOLLOW)
1012 flags |= O_NOFOLLOW;
1015 #if HAVE_LSTAT || defined lstat
1016 if (lstat (name, st) != 0)
1018 else if (S_ISLNK (st->st_mode))
1027 fd = OPEN (name, flags, mode);
1028 if (0 <= fd && fstat (fd, st) != 0)
1039 /* ========================================================================
1040 * Set ifname to the input file name (with a suffix appended if necessary)
1041 * and istat to its stats. For decompression, if no file exists with the
1042 * original name, try adding successively z_suffix, .gz, .z, -z and .Z.
1043 * For MSDOS, we try only z_suffix and z.
1044 * Return an open file descriptor or -1.
1047 open_input_file (iname, sbuf)
1051 int ilen; /* strlen(ifname) */
1052 int z_suffix_errno = 0;
1053 static char const *suffixes[] = {NULL, ".gz", ".z", "-z", ".Z", NULL};
1054 char const **suf = suffixes;
1056 #ifdef NO_MULTIPLE_DOTS
1057 char *dot; /* pointer to ifname extension, or NULL */
1060 int open_flags = (O_RDONLY | O_NONBLOCK | O_NOCTTY
1061 | (ascii && !decompress ? 0 : O_BINARY));
1065 if (sizeof ifname - 1 <= strlen (iname))
1068 strcpy(ifname, iname);
1070 /* If input file exists, return OK. */
1071 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1075 if (!decompress || errno != ENOENT) {
1079 /* file.ext doesn't exist, try adding a suffix (after removing any
1080 * version number for VMS).
1082 s = get_suffix(ifname);
1084 progerror(ifname); /* ifname already has z suffix and does not exist */
1087 #ifdef NO_MULTIPLE_DOTS
1088 dot = strrchr(ifname, '.');
1090 strcat(ifname, ".");
1091 dot = strrchr(ifname, '.');
1094 ilen = strlen(ifname);
1095 if (strequ(z_suffix, ".gz")) suf++;
1097 /* Search for all suffixes */
1099 char const *s0 = s = *suf;
1100 strcpy (ifname, iname);
1101 #ifdef NO_MULTIPLE_DOTS
1103 if (*dot == '\0') strcpy (dot, ".");
1105 #ifdef MAX_EXT_CHARS
1106 if (MAX_EXT_CHARS < strlen (s) + strlen (dot + 1))
1107 dot[MAX_EXT_CHARS + 1 - strlen (s)] = '\0';
1109 if (sizeof ifname <= ilen + strlen (s))
1112 fd = open_and_stat (ifname, open_flags, RW_USER, sbuf);
1115 if (errno != ENOENT)
1120 if (strequ (s0, z_suffix))
1121 z_suffix_errno = errno;
1122 } while (*++suf != NULL);
1124 /* No suffix found, complain using z_suffix: */
1125 strcpy(ifname, iname);
1126 #ifdef NO_MULTIPLE_DOTS
1127 if (*dot == '\0') strcpy(dot, ".");
1129 #ifdef MAX_EXT_CHARS
1130 if (MAX_EXT_CHARS < z_len + strlen (dot + 1))
1131 dot[MAX_EXT_CHARS + 1 - z_len] = '\0';
1133 strcat(ifname, z_suffix);
1134 errno = z_suffix_errno;
1139 fprintf (stderr, "%s: %s: file name too long\n", program_name, iname);
1144 /* ========================================================================
1145 * Generate ofname given ifname. Return OK, or WARNING if file must be skipped.
1146 * Sets save_orig_name to true if the file name has been truncated.
1148 local int make_ofname()
1150 char *suff; /* ofname z suffix */
1152 strcpy(ofname, ifname);
1153 /* strip a version number if any and get the gzip suffix if present: */
1154 suff = get_suffix(ofname);
1158 /* With -t or -l, try all files (even without .gz suffix)
1159 * except with -r (behave as with just -dr).
1161 if (!recursive && (list || test)) return OK;
1163 /* Avoid annoying messages with -r */
1164 if (verbose || (!recursive && !quiet)) {
1165 WARN((stderr,"%s: %s: unknown suffix -- ignored\n",
1166 program_name, ifname));
1170 /* Make a special case for .tgz and .taz: */
1172 if (strequ(suff, ".tgz") || strequ(suff, ".taz")) {
1173 strcpy(suff, ".tar");
1175 *suff = '\0'; /* strip the z suffix */
1177 /* ofname might be changed later if infile contains an original name */
1179 } else if (suff && ! force) {
1180 /* Avoid annoying messages with -r (see treat_dir()) */
1181 if (verbose || (!recursive && !quiet)) {
1182 /* Don't use WARN, as it affects exit status. */
1183 fprintf (stderr, "%s: %s already has %s suffix -- unchanged\n",
1184 program_name, ifname, suff);
1190 #ifdef NO_MULTIPLE_DOTS
1191 suff = strrchr(ofname, '.');
1193 if (sizeof ofname <= strlen (ofname) + 1)
1195 strcat(ofname, ".");
1196 # ifdef MAX_EXT_CHARS
1197 if (strequ(z_suffix, "z")) {
1198 if (sizeof ofname <= strlen (ofname) + 2)
1200 strcat(ofname, "gz"); /* enough room */
1203 /* On the Atari and some versions of MSDOS,
1204 * ENAMETOOLONG does not work correctly. So we
1205 * must truncate here.
1207 } else if (strlen(suff)-1 + z_len > MAX_SUFFIX) {
1208 suff[MAX_SUFFIX+1-z_len] = '\0';
1212 #endif /* NO_MULTIPLE_DOTS */
1213 if (sizeof ofname <= strlen (ofname) + z_len)
1215 strcat(ofname, z_suffix);
1217 } /* decompress ? */
1221 WARN ((stderr, "%s: %s: file name too long\n", program_name, ifname));
1226 /* ========================================================================
1227 * Check the magic number of the input file and update ofname if an
1228 * original name was given and to_stdout is not set.
1229 * Return the compression method, -1 for error, -2 for warning.
1230 * Set inptr to the offset of the next byte to be processed.
1231 * Updates time_stamp if there is one and --no-time is not used.
1232 * This function may be called repeatedly for an input file consisting
1233 * of several contiguous gzip'ed members.
1234 * IN assertions: there is at least one remaining compressed member.
1235 * If the member is a zip file, it must be the only one.
1237 local int get_method(in)
1238 int in; /* input file descriptor */
1240 uch flags; /* compression flags */
1241 char magic[2]; /* magic header */
1242 int imagic1; /* like magic[1], but can represent EOF */
1243 ulg stamp; /* time stamp */
1245 /* If --force and --stdout, zcat == cat, so do not complain about
1246 * premature end of file: use try_byte instead of get_byte.
1248 if (force && to_stdout) {
1249 magic[0] = (char)try_byte();
1250 imagic1 = try_byte ();
1251 magic[1] = (char) imagic1;
1252 /* If try_byte returned EOF, magic[1] == (char) EOF. */
1254 magic[0] = (char)get_byte();
1256 magic[1] = (char)get_byte();
1257 imagic1 = 0; /* avoid lint warning */
1259 imagic1 = try_byte ();
1260 magic[1] = (char) imagic1;
1263 method = -1; /* unknown yet */
1264 part_nb++; /* number of parts in gzip file */
1266 last_member = RECORD_IO;
1267 /* assume multiple members in gzip file except for record oriented I/O */
1269 if (memcmp(magic, GZIP_MAGIC, 2) == 0
1270 || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
1272 method = (int)get_byte();
1273 if (method != DEFLATED) {
1275 "%s: %s: unknown method %d -- not supported\n",
1276 program_name, ifname, method);
1281 flags = (uch)get_byte();
1283 if ((flags & ENCRYPTED) != 0) {
1285 "%s: %s is encrypted -- not supported\n",
1286 program_name, ifname);
1290 if ((flags & CONTINUATION) != 0) {
1292 "%s: %s is a multi-part gzip file -- not supported\n",
1293 program_name, ifname);
1295 if (force <= 1) return -1;
1297 if ((flags & RESERVED) != 0) {
1299 "%s: %s has flags 0x%x -- not supported\n",
1300 program_name, ifname, flags);
1302 if (force <= 1) return -1;
1304 stamp = (ulg)get_byte();
1305 stamp |= ((ulg)get_byte()) << 8;
1306 stamp |= ((ulg)get_byte()) << 16;
1307 stamp |= ((ulg)get_byte()) << 24;
1308 if (stamp != 0 && !no_time)
1310 time_stamp.tv_sec = stamp;
1311 time_stamp.tv_nsec = 0;
1314 (void)get_byte(); /* Ignore extra flags for the moment */
1315 (void)get_byte(); /* Ignore OS type for the moment */
1317 if ((flags & CONTINUATION) != 0) {
1318 unsigned part = (unsigned)get_byte();
1319 part |= ((unsigned)get_byte())<<8;
1321 fprintf(stderr,"%s: %s: part number %u\n",
1322 program_name, ifname, part);
1325 if ((flags & EXTRA_FIELD) != 0) {
1326 unsigned len = (unsigned)get_byte();
1327 len |= ((unsigned)get_byte())<<8;
1329 fprintf(stderr,"%s: %s: extra field of %u bytes ignored\n",
1330 program_name, ifname, len);
1332 while (len--) (void)get_byte();
1335 /* Get original file name if it was truncated */
1336 if ((flags & ORIG_NAME) != 0) {
1337 if (no_name || (to_stdout && !list) || part_nb > 1) {
1338 /* Discard the old name */
1339 char c; /* dummy used for NeXTstep 3.0 cc optimizer bug */
1340 do {c=get_byte();} while (c != 0);
1342 /* Copy the base name. Keep a directory prefix intact. */
1343 char *p = gzip_base_name (ofname);
1346 *p = (char)get_char();
1347 if (*p++ == '\0') break;
1348 if (p >= ofname+sizeof(ofname)) {
1349 gzip_error ("corrupted input -- file name too large");
1352 p = gzip_base_name (base);
1353 memmove (base, p, strlen (p) + 1);
1354 /* If necessary, adapt the name to local OS conventions: */
1356 MAKE_LEGAL_NAME(base);
1357 if (base) list=0; /* avoid warning about unused variable */
1359 } /* no_name || to_stdout */
1362 /* Discard file comment if any */
1363 if ((flags & COMMENT) != 0) {
1364 while (get_char() != 0) /* null */ ;
1367 header_bytes = inptr + 2*sizeof(long); /* include crc and size */
1370 } else if (memcmp(magic, PKZIP_MAGIC, 2) == 0 && inptr == 2
1371 && memcmp((char*)inbuf, PKZIP_MAGIC, 4) == 0) {
1372 /* To simplify the code, we support a zip file when alone only.
1373 * We are thus guaranteed that the entire local header fits in inbuf.
1377 if (check_zipfile(in) != OK) return -1;
1378 /* check_zipfile may get ofname from the local header */
1381 } else if (memcmp(magic, PACK_MAGIC, 2) == 0) {
1385 } else if (memcmp(magic, LZW_MAGIC, 2) == 0) {
1387 method = COMPRESSED;
1390 } else if (memcmp(magic, LZH_MAGIC, 2) == 0) {
1395 } else if (force && to_stdout && !list) { /* pass input unchanged */
1401 if (method >= 0) return method;
1404 fprintf (stderr, "\n%s: %s: not in gzip format\n",
1405 program_name, ifname);
1412 for (inbyte = imagic1; inbyte == 0; inbyte = try_byte ())
1417 WARN ((stderr, "\n%s: %s: decompression OK, trailing zero bytes ignored\n",
1418 program_name, ifname));
1423 WARN((stderr, "\n%s: %s: decompression OK, trailing garbage ignored\n",
1424 program_name, ifname));
1429 /* ========================================================================
1430 * Display the characteristics of the compressed file.
1431 * If the given method is < 0, display the accumulated totals.
1432 * IN assertions: time_stamp, header_bytes and ifile_size are initialized.
1434 local void do_list(ifd, method)
1435 int ifd; /* input file descriptor */
1436 int method; /* compression method */
1438 ulg crc; /* original crc */
1439 static int first_time = 1;
1440 static char const *const methods[MAX_METHODS] = {
1445 "", "", "", "", /* 4 to 7 reserved */
1447 int positive_off_t_width = 1;
1450 for (o = OFF_T_MAX; 9 < o; o /= 10) {
1451 positive_off_t_width++;
1454 if (first_time && method >= 0) {
1457 printf("method crc date time ");
1460 printf("%*.*s %*.*s ratio uncompressed_name\n",
1461 positive_off_t_width, positive_off_t_width, "compressed",
1462 positive_off_t_width, positive_off_t_width, "uncompressed");
1464 } else if (method < 0) {
1465 if (total_in <= 0 || total_out <= 0) return;
1469 if (verbose || !quiet) {
1470 fprint_off(stdout, total_in, positive_off_t_width);
1472 fprint_off(stdout, total_out, positive_off_t_width);
1475 display_ratio(total_out-(total_in-header_bytes), total_out, stdout);
1476 /* header_bytes is not meaningful but used to ensure the same
1477 * ratio if there is a single file.
1479 printf(" (totals)\n");
1482 crc = (ulg)~0; /* unknown */
1484 bytes_in = ifile_size;
1487 if (method == DEFLATED && !last_member) {
1488 /* Get the crc and uncompressed size for gzip'ed (not zip'ed) files.
1489 * If the lseek fails, we could use read() to get to the end, but
1490 * --list is used to get quick results.
1491 * Use "gunzip < foo.gz | wc -c" to get the uncompressed size if
1492 * you are not concerned about speed.
1494 bytes_in = lseek(ifd, (off_t)(-8), SEEK_END);
1495 if (bytes_in != -1L) {
1498 if (read(ifd, (char*)buf, sizeof(buf)) != sizeof(buf)) {
1502 bytes_out = LG(buf+4);
1505 #endif /* RECORD_IO */
1508 struct tm *tm = localtime (&time_stamp.tv_sec);
1509 printf ("%5s %08lx ", methods[method], crc);
1511 printf ("%s%3d %02d:%02d ",
1512 ("Jan\0Feb\0Mar\0Apr\0May\0Jun\0Jul\0Aug\0Sep\0Oct\0Nov\0Dec"
1514 tm->tm_mday, tm->tm_hour, tm->tm_min);
1516 printf ("??? ?? ??:?? ");
1518 fprint_off(stdout, bytes_in, positive_off_t_width);
1520 fprint_off(stdout, bytes_out, positive_off_t_width);
1522 if (bytes_in == -1L) {
1524 bytes_in = bytes_out = header_bytes = 0;
1525 } else if (total_in >= 0) {
1526 total_in += bytes_in;
1528 if (bytes_out == -1L) {
1530 bytes_in = bytes_out = header_bytes = 0;
1531 } else if (total_out >= 0) {
1532 total_out += bytes_out;
1534 display_ratio(bytes_out-(bytes_in-header_bytes), bytes_out, stdout);
1535 printf(" %s\n", ofname);
1538 /* ========================================================================
1539 * Shorten the given name by one character, or replace a .tar extension
1540 * with .tgz. Truncate the last part of the name which is longer than
1541 * MIN_PART characters: 1234.678.012.gz -> 123.678.012.gz. If the name
1542 * has only parts shorter than MIN_PART truncate the longest part.
1543 * For decompression, just remove the last character of the name.
1545 * IN assertion: for compression, the suffix of the given name is z_suffix.
1547 local void shorten_name(name)
1550 int len; /* length of name without z_suffix */
1551 char *trunc = NULL; /* character to be truncated */
1552 int plen; /* current part length */
1553 int min_part = MIN_PART; /* current minimum part length */
1559 gzip_error ("name too short");
1563 p = get_suffix(name);
1565 gzip_error ("can't recover suffix\n");
1569 /* compress 1234567890.tar to 1234567890.tgz */
1570 if (len > 4 && strequ(p-4, ".tar")) {
1571 strcpy(p-4, ".tgz");
1574 /* Try keeping short extensions intact:
1575 * 1234.678.012.gz -> 123.678.012.gz
1578 p = strrchr(name, PATH_SEP);
1581 plen = strcspn(p, PART_SEP);
1583 if (plen > min_part) trunc = p-1;
1586 } while (trunc == NULL && --min_part != 0);
1588 if (trunc != NULL) {
1590 trunc[0] = trunc[1];
1594 trunc = strrchr(name, PART_SEP[0]);
1596 gzip_error ("internal error in shorten_name");
1597 if (trunc[1] == '\0') trunc--; /* force truncation */
1599 strcpy(trunc, z_suffix);
1602 /* ========================================================================
1603 * The compressed file already exists, so ask for confirmation.
1604 * Return ERROR if the file must be skipped.
1606 local int check_ofname()
1608 /* Ask permission to overwrite the existing file */
1611 fprintf (stderr, "%s: %s already exists;", program_name, ofname);
1612 if (foreground && isatty(fileno(stdin))) {
1613 fprintf(stderr, " do you wish to overwrite (y or n)? ");
1618 fprintf(stderr, "\tnot overwritten\n");
1619 if (exit_code == OK) exit_code = WARNING;
1623 if (xunlink (ofname)) {
1631 /* ========================================================================
1632 * Copy modes, times, ownership from input file to output file.
1633 * IN assertion: to_stdout is false.
1635 local void copy_stat(ifstat)
1636 struct stat *ifstat;
1638 mode_t mode = ifstat->st_mode & S_IRWXUGO;
1642 struct timespec timespec[2];
1643 timespec[0] = get_stat_atime (ifstat);
1644 timespec[1] = get_stat_mtime (ifstat);
1646 if (decompress && 0 <= time_stamp.tv_nsec
1647 && ! (timespec[1].tv_sec == time_stamp.tv_sec
1648 && timespec[1].tv_nsec == time_stamp.tv_nsec))
1650 timespec[1] = time_stamp;
1652 fprintf(stderr, "%s: time stamp restored\n", ofname);
1656 if (gl_futimens (ofd, ofname, timespec) != 0)
1659 WARN ((stderr, "%s: ", program_name));
1669 /* Copy ownership */
1671 ignore_value (fchown (ofd, ifstat->st_uid, ifstat->st_gid));
1673 ignore_value (chown (ofname, ifstat->st_uid, ifstat->st_gid));
1677 /* Copy the protection modes */
1679 r = fchmod (ofd, mode);
1681 r = chmod (ofname, mode);
1685 WARN ((stderr, "%s: ", program_name));
1695 /* ========================================================================
1696 * Recurse through the given directory. This code is taken from ncompress.
1698 local void treat_dir (fd, dir)
1704 char nbuf[MAX_PATH_LEN];
1707 dirp = fdopendir (fd);
1715 ** WARNING: the following algorithm could occasionally cause
1716 ** compress to produce error warnings of the form "<filename>.gz
1717 ** already has .gz suffix - ignored". This occurs when the
1718 ** .gz output file is inserted into the directory below
1719 ** readdir's current pointer.
1720 ** These warnings are harmless but annoying, so they are suppressed
1721 ** with option -r (except when -v is on). An alternative
1722 ** to allowing this would be to store the entire directory
1723 ** list in memory, then compress the entries in the stored
1724 ** list. Given the depth-first recursive algorithm used here,
1725 ** this could use up a tremendous amount of memory. I don't
1726 ** think it's worth it. -- Dave Mack
1727 ** (An other alternative might be two passes to avoid depth-first.)
1730 while ((errno = 0, dp = readdir(dirp)) != NULL) {
1732 if (strequ(dp->d_name,".") || strequ(dp->d_name,"..")) {
1736 if (len + _D_EXACT_NAMLEN (dp) + 1 < MAX_PATH_LEN - 1) {
1738 if (len != 0 /* dir = "" means current dir on Amiga */
1740 && dir[len-1] != PATH_SEP2
1743 && dir[len-1] != PATH_SEP3
1746 nbuf[len++] = PATH_SEP;
1748 strcpy(nbuf+len, dp->d_name);
1751 fprintf(stderr,"%s: %s/%s: pathname too long\n",
1752 program_name, dir, dp->d_name);
1758 if (CLOSEDIR(dirp) != 0)
1761 #endif /* ! NO_DIR */
1763 /* Make sure signals get handled properly. */
1766 install_signal_handlers ()
1768 int nsigs = sizeof handled_sig / sizeof handled_sig[0];
1772 struct sigaction act;
1774 sigemptyset (&caught_signals);
1775 for (i = 0; i < nsigs; i++)
1777 sigaction (handled_sig[i], NULL, &act);
1778 if (act.sa_handler != SIG_IGN)
1779 sigaddset (&caught_signals, handled_sig[i]);
1782 act.sa_handler = abort_gzip_signal;
1783 act.sa_mask = caught_signals;
1786 for (i = 0; i < nsigs; i++)
1787 if (sigismember (&caught_signals, handled_sig[i]))
1791 sigaction (handled_sig[i], &act, NULL);
1794 for (i = 0; i < nsigs; i++)
1795 if (signal (handled_sig[i], SIG_IGN) != SIG_IGN)
1799 signal (handled_sig[i], abort_gzip_signal);
1800 siginterrupt (handled_sig[i], 1);
1805 /* ========================================================================
1806 * Free all dynamically allocated variables and exit with the given code.
1808 local void do_exit(exitcode)
1811 static int in_exit = 0;
1813 if (in_exit) exit(exitcode);
1832 /* ========================================================================
1833 * Close and unlink the output file.
1836 remove_output_file ()
1841 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
1842 fd = remove_ofname_fd;
1845 remove_ofname_fd = -1;
1849 sigprocmask (SIG_SETMASK, &oldset, NULL);
1852 /* ========================================================================
1858 remove_output_file ();
1862 /* ========================================================================
1866 abort_gzip_signal (sig)
1870 signal (sig, SIG_IGN);
1871 remove_output_file ();
1872 if (sig == exiting_signal)
1874 signal (sig, SIG_DFL);