* Zip64 + other methods
*/
//config:config UNZIP
-//config: bool "unzip"
+//config: bool "unzip (24 kb)"
//config: default y
//config: help
-//config: unzip will list or extract files from a ZIP archive,
-//config: commonly found on DOS/WIN systems. The default behavior
-//config: (with no options) is to extract the archive into the
-//config: current directory.
+//config: unzip will list or extract files from a ZIP archive,
+//config: commonly found on DOS/WIN systems. The default behavior
+//config: (with no options) is to extract the archive into the
+//config: current directory.
//config:
//config:config FEATURE_UNZIP_CDF
//config: bool "Read and use Central Directory data"
//config: default y
//config: depends on UNZIP
//config: help
-//config: If you know that you only need to deal with simple
-//config: ZIP files without deleted/updated files, SFX archives etc,
-//config: you can reduce code size by unselecting this option.
-//config: To support less trivial ZIPs, say Y.
+//config: If you know that you only need to deal with simple
+//config: ZIP files without deleted/updated files, SFX archives etc,
+//config: you can reduce code size by unselecting this option.
+//config: To support less trivial ZIPs, say Y.
+//config:
+//config:config FEATURE_UNZIP_BZIP2
+//config: bool "Support compression method 12 (bzip2)"
+//config: default y
+//config: depends on FEATURE_UNZIP_CDF && DESKTOP
+// FEATURE_UNZIP_CDF is needed, otherwise we can't find start of next file
+// DESKTOP is needed to get back uncompressed length
+//config:
+//config:config FEATURE_UNZIP_LZMA
+//config: bool "Support compression method 14 (lzma)"
+//config: default y
+//config: depends on FEATURE_UNZIP_CDF && DESKTOP
+//config:
+//config:config FEATURE_UNZIP_XZ
+//config: bool "Support compression method 95 (xz)"
+//config: default y
+//config: depends on FEATURE_UNZIP_CDF && DESKTOP
//applet:IF_UNZIP(APPLET(unzip, BB_DIR_USR_BIN, BB_SUID_DROP))
//kbuild:lib-$(CONFIG_UNZIP) += unzip.o
//usage: "\n -l List contents (with -q for short form)"
//usage: "\n -n Never overwrite files (default: ask)"
//usage: "\n -o Overwrite"
+//usage: "\n -j Do not restore paths"
//usage: "\n -p Print to stdout"
//usage: "\n -q Quiet"
//usage: "\n -x FILE Exclude FILEs"
#define FIX_ENDIANNESS_ZIP(zip) \
do { if (BB_BIG_ENDIAN) { \
+ (zip).fmt.method = SWAP_LE16((zip).fmt.method ); \
(zip).fmt.crc32 = SWAP_LE32((zip).fmt.crc32 ); \
(zip).fmt.cmpsize = SWAP_LE32((zip).fmt.cmpsize ); \
(zip).fmt.ucmpsize = SWAP_LE32((zip).fmt.ucmpsize ); \
end = 0;
dbg("Looking for cdf_offset starting from 0x%"OFF_FMT"x", end);
- xlseek(zip_fd, end, SEEK_SET);
+ xlseek(zip_fd, end, SEEK_SET);
buf = xzalloc(PEEK_FROM_END);
full_read(zip_fd, buf, PEEK_FROM_END);
};
#endif
+static void die_if_bad_fnamesize(unsigned sz)
+{
+ if (sz > 0xfff) /* more than 4k?! no funny business please */
+ bb_error_msg_and_die("bad archive");
+}
+
static void unzip_skip(off_t skip)
{
if (skip != 0)
free(name);
}
+#if ENABLE_FEATURE_UNZIP_CDF
+static void unzip_extract_symlink(zip_header_t *zip, const char *dst_fn)
+{
+ char *target;
+
+ die_if_bad_fnamesize(zip->fmt.ucmpsize);
+
+ if (zip->fmt.method == 0) {
+ /* Method 0 - stored (not compressed) */
+ target = xzalloc(zip->fmt.ucmpsize + 1);
+ xread(zip_fd, target, zip->fmt.ucmpsize);
+ } else {
+#if 1
+ bb_error_msg_and_die("compressed symlink is not supported");
+#else
+ transformer_state_t xstate;
+ init_transformer_state(&xstate);
+ xstate.mem_output_size_max = zip->fmt.ucmpsize;
+ /* ...unpack... */
+ if (!xstate.mem_output_buf)
+ WTF();
+ target = xstate.mem_output_buf;
+ target = xrealloc(target, xstate.mem_output_size + 1);
+ target[xstate.mem_output_size] = '\0';
+#endif
+ }
+//TODO: libbb candidate
+ if (symlink(target, dst_fn))
+ bb_perror_msg_and_die("can't create symlink '%s'", dst_fn);
+ free(target);
+}
+#endif
+
static void unzip_extract(zip_header_t *zip, int dst_fd)
{
+ transformer_state_t xstate;
+
if (zip->fmt.method == 0) {
/* Method 0 - stored (not compressed) */
off_t size = zip->fmt.ucmpsize;
if (size)
bb_copyfd_exact_size(zip_fd, dst_fd, size);
- } else {
+ return;
+ }
+
+ init_transformer_state(&xstate);
+ xstate.bytes_in = zip->fmt.cmpsize;
+ xstate.src_fd = zip_fd;
+ xstate.dst_fd = dst_fd;
+ if (zip->fmt.method == 8) {
/* Method 8 - inflate */
- transformer_state_t xstate;
- init_transformer_state(&xstate);
- xstate.bytes_in = zip->fmt.cmpsize;
- xstate.src_fd = zip_fd;
- xstate.dst_fd = dst_fd;
if (inflate_unzip(&xstate) < 0)
bb_error_msg_and_die("inflate error");
/* Validate decompression - crc */
if (zip->fmt.crc32 != (xstate.crc32 ^ 0xffffffffL)) {
bb_error_msg_and_die("crc error");
}
- /* Validate decompression - size */
- if (zip->fmt.ucmpsize != xstate.bytes_out) {
- /* Don't die. Who knows, maybe len calculation
- * was botched somewhere. After all, crc matched! */
- bb_error_msg("bad length");
- }
}
- /* TODO? method 12: bzip2, method 14: LZMA */
+#if ENABLE_FEATURE_UNZIP_BZIP2
+ else if (zip->fmt.method == 12) {
+ /* Tested. Unpacker reads too much, but we use CDF
+ * and will seek to the correct beginning of next file.
+ */
+ xstate.bytes_out = unpack_bz2_stream(&xstate);
+ if (xstate.bytes_out < 0)
+ bb_error_msg_and_die("inflate error");
+ }
+#endif
+#if ENABLE_FEATURE_UNZIP_LZMA
+ else if (zip->fmt.method == 14) {
+ /* Not tested yet */
+ xstate.bytes_out = unpack_lzma_stream(&xstate);
+ if (xstate.bytes_out < 0)
+ bb_error_msg_and_die("inflate error");
+ }
+#endif
+#if ENABLE_FEATURE_UNZIP_XZ
+ else if (zip->fmt.method == 95) {
+ /* Not tested yet */
+ xstate.bytes_out = unpack_xz_stream(&xstate);
+ if (xstate.bytes_out < 0)
+ bb_error_msg_and_die("inflate error");
+ }
+#endif
+ else {
+ bb_error_msg_and_die("unsupported method %u", zip->fmt.method);
+ }
+
+ /* Validate decompression - size */
+ if (zip->fmt.ucmpsize != xstate.bytes_out) {
+ /* Don't die. Who knows, maybe len calculation
+ * was botched somewhere. After all, crc matched! */
+ bb_error_msg("bad length");
+ }
}
static void my_fgets80(char *buf80)
}
}
+static int get_lstat_mode(const char *dst_fn)
+{
+ struct stat stat_buf;
+ if (lstat(dst_fn, &stat_buf) == -1) {
+ if (errno != ENOENT) {
+ bb_perror_msg_and_die("can't stat '%s'", dst_fn);
+ }
+ /* File does not exist */
+ return -1;
+ }
+ return stat_buf.st_mode;
+}
+
int unzip_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
int unzip_main(int argc, char **argv)
{
- enum { O_PROMPT, O_NEVER, O_ALWAYS };
-
+ enum {
+ OPT_l = (1 << 0),
+ OPT_x = (1 << 1),
+ OPT_j = (1 << 2),
+ };
+ unsigned opts;
smallint quiet = 0;
IF_NOT_FEATURE_UNZIP_CDF(const) smallint verbose = 0;
- smallint listing = 0;
+ enum { O_PROMPT, O_NEVER, O_ALWAYS };
smallint overwrite = O_PROMPT;
- smallint x_opt_seen;
uint32_t cdf_offset;
unsigned long total_usize;
unsigned long total_size;
llist_t *zaccept = NULL;
llist_t *zreject = NULL;
char *base_dir = NULL;
- int i, opt;
+ int i;
char key_buf[80]; /* must match size used by my_fgets80 */
- struct stat stat_buf;
/* -q, -l and -v: UnZip 5.52 of 28 February 2005, by Info-ZIP:
*
* 204372 1 file
*/
- x_opt_seen = 0;
+ opts = 0;
/* '-' makes getopt return 1 for non-options */
- while ((opt = getopt(argc, argv, "-d:lnopqxv")) != -1) {
- switch (opt) {
+ while ((i = getopt(argc, argv, "-d:lnopqxjv")) != -1) {
+ switch (i) {
case 'd': /* Extract to base directory */
base_dir = optarg;
break;
case 'l': /* List */
- listing = 1;
+ opts |= OPT_l;
break;
case 'n': /* Never overwrite existing files */
case 'v': /* Verbose list */
IF_FEATURE_UNZIP_CDF(verbose++;)
- listing = 1;
+ opts |= OPT_l;
break;
case 'x':
- x_opt_seen = 1;
+ opts |= OPT_x;
+ break;
+
+ case 'j':
+ opts |= OPT_j;
break;
case 1:
/* +5: space for ".zip" and NUL */
src_fn = xmalloc(strlen(optarg) + 5);
strcpy(src_fn, optarg);
- } else if (!x_opt_seen) {
+ } else if (!(opts & OPT_x)) {
/* Include files */
llist_add_to(&zaccept, optarg);
} else {
if (quiet <= 1) { /* not -qq */
if (quiet == 0)
printf("Archive: %s\n", src_fn);
- if (listing) {
+ if (opts & OPT_l) {
puts(verbose ?
" Length Method Size Cmpr Date Time CRC-32 Name\n"
"-------- ------ ------- ---- ---------- ----- -------- ----"
xread(zip_fd, zip.raw, ZIP_HEADER_LEN);
FIX_ENDIANNESS_ZIP(zip);
- if ((zip.fmt.method != 0)
- && (zip.fmt.method != 8)
- ) {
- /* TODO? method 12: bzip2, method 14: LZMA */
- bb_error_msg_and_die("unsupported method %d", zip.fmt.method);
- }
- if (zip.fmt.zip_flags & SWAP_LE16(0x0009)) {
- bb_error_msg_and_die("zip flags 1 and 8 are not supported");
+ if (zip.fmt.zip_flags & SWAP_LE16(0x0008)) {
+ bb_error_msg_and_die("zip flag %s is not supported",
+ "8 (streaming)");
}
}
#if ENABLE_FEATURE_UNZIP_CDF
zip.fmt.cmpsize = cdf.fmt.cmpsize;
zip.fmt.ucmpsize = cdf.fmt.ucmpsize;
}
+// Seen in some zipfiles: central directory 9 byte extra field contains
+// a subfield with ID 0x5455 and 5 data bytes, which is a Unix-style UTC mtime.
+// Local header version:
+// u16 0x5455 ("UT")
+// u16 size (1 + 4 * n)
+// u8 flags: bit 0:mtime is present, bit 1:atime is present, bit 2:ctime is present
+// u32 mtime
+// u32 atime
+// u32 ctime
+// Central header version:
+// u16 0x5455 ("UT")
+// u16 size (5 (or 1?))
+// u8 flags: bit 0:mtime is present, bit 1:atime is present, bit 2:ctime is present
+// u32 mtime (CDF does not store atime/ctime)
# else
/* CDF has the same data as local header, no need to read the latter...
* ...not really. An archive was seen with cdf.extra_len == 6 but
if (zip.fmt.zip_flags & SWAP_LE16(0x0001)) {
/* 0x0001 - encrypted */
- bb_error_msg_and_die("zip flag 1 (encryption) is not supported");
+ bb_error_msg_and_die("zip flag %s is not supported",
+ "1 (encryption)");
}
dbg("File cmpsize:0x%x extra_len:0x%x ucmpsize:0x%x",
(unsigned)zip.fmt.cmpsize,
/* Read filename */
free(dst_fn);
+ die_if_bad_fnamesize(zip.fmt.filename_len);
dst_fn = xzalloc(zip.fmt.filename_len + 1);
xread(zip_fd, dst_fn, zip.fmt.filename_len);
-
/* Skip extra header bytes */
unzip_skip(zip.fmt.extra_len);
/* Guard against "/abspath", "/../" and similar attacks */
overlapping_strcpy(dst_fn, strip_unsafe_prefix(dst_fn));
+ if (opts & OPT_j) /* Strip paths? */
+ overlapping_strcpy(dst_fn, bb_basename(dst_fn));
+
+ /* Did this strip everything ("DIR/" case)? Then skip */
+ if (!dst_fn[0])
+ goto skip_cmpsize;
+
/* Filter zip entries */
if (find_list_entry(zreject, dst_fn)
|| (zaccept && !find_list_entry(zaccept, dst_fn))
) { /* Skip entry */
- i = 'n';
- } else {
- if (listing) {
- /* List entry */
- char dtbuf[sizeof("mm-dd-yyyy hh:mm")];
- sprintf(dtbuf, "%02u-%02u-%04u %02u:%02u",
- (zip.fmt.moddate >> 5) & 0xf, // mm: 0x01e0
- (zip.fmt.moddate) & 0x1f, // dd: 0x001f
- (zip.fmt.moddate >> 9) + 1980, // yy: 0xfe00
- (zip.fmt.modtime >> 11), // hh: 0xf800
- (zip.fmt.modtime >> 5) & 0x3f // mm: 0x07e0
- // seconds/2 not shown, encoded in -- 0x001f
- );
- if (!verbose) {
- // " Length Date Time Name\n"
- // "--------- ---------- ----- ----"
- printf( "%9u " "%s " "%s\n",
- (unsigned)zip.fmt.ucmpsize,
- dtbuf,
- dst_fn);
- } else {
- unsigned long percents = zip.fmt.ucmpsize - zip.fmt.cmpsize;
- if ((int32_t)percents < 0)
- percents = 0; /* happens if ucmpsize < cmpsize */
- percents = percents * 100;
- if (zip.fmt.ucmpsize)
- percents /= zip.fmt.ucmpsize;
- // " Length Method Size Cmpr Date Time CRC-32 Name\n"
- // "-------- ------ ------- ---- ---------- ----- -------- ----"
- printf( "%8u %s" "%9u%4u%% " "%s " "%08x " "%s\n",
- (unsigned)zip.fmt.ucmpsize,
- zip.fmt.method == 0 ? "Stored" : "Defl:N", /* Defl is method 8 */
-/* TODO: show other methods?
- * 1 - Shrunk
- * 2 - Reduced with compression factor 1
- * 3 - Reduced with compression factor 2
- * 4 - Reduced with compression factor 3
- * 5 - Reduced with compression factor 4
- * 6 - Imploded
- * 7 - Reserved for Tokenizing compression algorithm
- * 9 - Deflate64
- * 10 - PKWARE Data Compression Library Imploding
- * 11 - Reserved by PKWARE
- * 12 - BZIP2
- * 14 - LZMA
- */
- (unsigned)zip.fmt.cmpsize,
- (unsigned)percents,
- dtbuf,
- zip.fmt.crc32,
- dst_fn);
- total_size += zip.fmt.cmpsize;
+ goto skip_cmpsize;
+ }
+
+ if (opts & OPT_l) {
+ /* List entry */
+ char dtbuf[sizeof("mm-dd-yyyy hh:mm")];
+ sprintf(dtbuf, "%02u-%02u-%04u %02u:%02u",
+ (zip.fmt.moddate >> 5) & 0xf, // mm: 0x01e0
+ (zip.fmt.moddate) & 0x1f, // dd: 0x001f
+ (zip.fmt.moddate >> 9) + 1980, // yy: 0xfe00
+ (zip.fmt.modtime >> 11), // hh: 0xf800
+ (zip.fmt.modtime >> 5) & 0x3f // mm: 0x07e0
+ // seconds/2 not shown, encoded in -- 0x001f
+ );
+ if (!verbose) {
+ // " Length Date Time Name\n"
+ // "--------- ---------- ----- ----"
+ printf( "%9u " "%s " "%s\n",
+ (unsigned)zip.fmt.ucmpsize,
+ dtbuf,
+ dst_fn);
+ } else {
+ char method6[7];
+ unsigned long percents;
+
+ sprintf(method6, "%6u", zip.fmt.method);
+ if (zip.fmt.method == 0) {
+ strcpy(method6, "Stored");
}
- total_usize += zip.fmt.ucmpsize;
- i = 'n';
- } else if (dst_fd == STDOUT_FILENO) {
- /* Extracting to STDOUT */
- i = -1;
- } else if (last_char_is(dst_fn, '/')) {
- /* Extract directory */
- if (stat(dst_fn, &stat_buf) == -1) {
- if (errno != ENOENT) {
- bb_perror_msg_and_die("can't stat '%s'", dst_fn);
- }
- if (!quiet) {
- printf(" creating: %s\n", dst_fn);
- }
- unzip_create_leading_dirs(dst_fn);
- if (bb_make_directory(dst_fn, dir_mode, FILEUTILS_IGNORE_CHMOD_ERR)) {
- xfunc_die();
- }
- } else {
- if (!S_ISDIR(stat_buf.st_mode)) {
- bb_error_msg_and_die("'%s' exists but is not a %s",
- dst_fn, "directory");
- }
+ if (zip.fmt.method == 8) {
+ strcpy(method6, "Defl:N");
+ /* normal, maximum, fast, superfast */
+ IF_DESKTOP(method6[5] = "NXFS"[(zip.fmt.zip_flags >> 1) & 3];)
+ }
+ percents = zip.fmt.ucmpsize - zip.fmt.cmpsize;
+ if ((int32_t)percents < 0)
+ percents = 0; /* happens if ucmpsize < cmpsize */
+ percents = percents * 100;
+ if (zip.fmt.ucmpsize)
+ percents /= zip.fmt.ucmpsize;
+ // " Length Method Size Cmpr Date Time CRC-32 Name\n"
+ // "-------- ------ ------- ---- ---------- ----- -------- ----"
+ printf( "%8u %s" "%9u%4u%% " "%s " "%08x " "%s\n",
+ (unsigned)zip.fmt.ucmpsize,
+ method6,
+ (unsigned)zip.fmt.cmpsize,
+ (unsigned)percents,
+ dtbuf,
+ zip.fmt.crc32,
+ dst_fn);
+ total_size += zip.fmt.cmpsize;
+ }
+ total_usize += zip.fmt.ucmpsize;
+ goto skip_cmpsize;
+ }
+
+ if (dst_fd == STDOUT_FILENO) {
+ /* Extracting to STDOUT */
+ goto do_extract;
+ }
+ if (last_char_is(dst_fn, '/')) {
+ int mode;
+
+ /* Extract directory */
+ mode = get_lstat_mode(dst_fn);
+ if (mode == -1) { /* ENOENT */
+ if (!quiet) {
+ printf(" creating: %s\n", dst_fn);
+ }
+ unzip_create_leading_dirs(dst_fn);
+ if (bb_make_directory(dst_fn, dir_mode, FILEUTILS_IGNORE_CHMOD_ERR)) {
+ xfunc_die();
}
- i = 'n';
} else {
- /* Extract file */
- check_file:
- if (stat(dst_fn, &stat_buf) == -1) {
- /* File does not exist */
- if (errno != ENOENT) {
- bb_perror_msg_and_die("can't stat '%s'", dst_fn);
- }
- i = 'y';
- } else {
- /* File already exists */
- if (overwrite == O_NEVER) {
- i = 'n';
- } else if (S_ISREG(stat_buf.st_mode)) {
- /* File is regular file */
- if (overwrite == O_ALWAYS) {
- i = 'y';
- } else {
- printf("replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ", dst_fn);
- my_fgets80(key_buf);
- i = key_buf[0];
- }
- } else {
- /* File is not regular file */
- bb_error_msg_and_die("'%s' exists but is not a %s",
- dst_fn, "regular file");
- }
+ if (!S_ISDIR(mode)) {
+ bb_error_msg_and_die("'%s' exists but is not a %s",
+ dst_fn, "directory");
}
}
+ goto skip_cmpsize;
+ }
+ check_file:
+ /* Does target file already exist? */
+ {
+ int mode = get_lstat_mode(dst_fn);
+ if (mode == -1) {
+ /* ENOENT: does not exist */
+ goto do_open_and_extract;
+ }
+ if (overwrite == O_NEVER) {
+ goto skip_cmpsize;
+ }
+ if (!S_ISREG(mode)) {
+ fishy:
+ bb_error_msg_and_die("'%s' exists but is not a %s",
+ dst_fn, "regular file");
+ }
+ if (overwrite == O_ALWAYS) {
+ goto do_open_and_extract;
+ }
+ printf("replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ", dst_fn);
+ my_fgets80(key_buf);
+ /* User input could take a long time. Is it still a regular file? */
+ mode = get_lstat_mode(dst_fn);
+ if (!S_ISREG(mode))
+ goto fishy;
}
- switch (i) {
+ /* Extract (or skip) it */
+ switch (key_buf[0]) {
case 'A':
overwrite = O_ALWAYS;
case 'y': /* Open file and fall into unzip */
+ do_open_and_extract:
unzip_create_leading_dirs(dst_fn);
#if ENABLE_FEATURE_UNZIP_CDF
- dst_fd = xopen3(dst_fn, O_WRONLY | O_CREAT | O_TRUNC, file_mode);
+ dst_fd = -1;
+ if (!S_ISLNK(file_mode)) {
+ dst_fd = xopen3(dst_fn,
+ O_WRONLY | O_CREAT | O_TRUNC | O_NOFOLLOW,
+ file_mode);
+ }
#else
- dst_fd = xopen(dst_fn, O_WRONLY | O_CREAT | O_TRUNC);
+ /* O_NOFOLLOW defends against symlink attacks */
+ dst_fd = xopen(dst_fn, O_WRONLY | O_CREAT | O_TRUNC | O_NOFOLLOW);
#endif
- case -1: /* Unzip */
+ do_extract:
if (!quiet) {
printf(/* zip.fmt.method == 0
? " extracting: %s\n"
: */ " inflating: %s\n", dst_fn);
}
- unzip_extract(&zip, dst_fd);
- if (dst_fd != STDOUT_FILENO) {
- /* closing STDOUT is potentially bad for future business */
- close(dst_fd);
+#if ENABLE_FEATURE_UNZIP_CDF
+ if (S_ISLNK(file_mode)) {
+ if (dst_fd != STDOUT_FILENO) /* not -p? */
+ unzip_extract_symlink(&zip, dst_fn);
+ } else
+#endif
+ {
+ unzip_extract(&zip, dst_fd);
+ if (dst_fd != STDOUT_FILENO) {
+ /* closing STDOUT is potentially bad for future business */
+ close(dst_fd);
+ }
}
break;
case 'N':
overwrite = O_NEVER;
- case 'n':
- /* Skip entry data */
+ case 'n': /* Skip entry data */
+ skip_cmpsize:
unzip_skip(zip.fmt.cmpsize);
break;
goto check_file;
default:
- printf("error: invalid response [%c]\n", (char)i);
+ printf("error: invalid response [%c]\n", (char)key_buf[0]);
goto check_file;
}
total_entries++;
}
- if (listing && quiet <= 1) {
+ if ((opts & OPT_l) && quiet <= 1) {
if (!verbose) {
// " Length Date Time Name\n"
// "--------- ---------- ----- ----"
projects / oweals / busybox.git / blobdiff