1 /* vi: set sw=4 ts=4: */
3 * Mini unzip implementation for busybox
5 * Copyright (C) 2004 by Ed Clark
7 * Loosely based on original busybox unzip applet by Laurence Anderson.
8 * All options and features should work in this version.
10 * Licensed under the GPL v2 or later, see the file LICENSE in this tarball.
14 * http://www.pkware.com/company/standards/appnote/
15 * http://www.info-zip.org/pub/infozip/doc/appnote-iz-latest.zip
19 * Zip64 + other methods
23 #include "unarchive.h"
27 ZIP_FILEHEADER_MAGIC = 0x504b0304,
28 ZIP_CDS_MAGIC = 0x504b0102,
29 ZIP_CDE_MAGIC = 0x504b0506,
30 ZIP_DD_MAGIC = 0x504b0708,
32 ZIP_FILEHEADER_MAGIC = 0x04034b50,
33 ZIP_CDS_MAGIC = 0x02014b50,
34 ZIP_CDE_MAGIC = 0x06054b50,
35 ZIP_DD_MAGIC = 0x08074b50,
39 #define ZIP_HEADER_LEN 26
42 uint8_t raw[ZIP_HEADER_LEN];
44 uint16_t version; /* 0-1 */
45 uint16_t flags; /* 2-3 */
46 uint16_t method; /* 4-5 */
47 uint16_t modtime; /* 6-7 */
48 uint16_t moddate; /* 8-9 */
49 uint32_t crc32 PACKED; /* 10-13 */
50 uint32_t cmpsize PACKED; /* 14-17 */
51 uint32_t ucmpsize PACKED; /* 18-21 */
52 uint16_t filename_len; /* 22-23 */
53 uint16_t extra_len; /* 24-25 */
55 } zip_header_t; /* PACKED - gcc 4.2.1 doesn't like it (spews warning) */
57 /* Check the offset of the last element, not the length. This leniency
58 * allows for poor packing, whereby the overall struct may be too long,
59 * even though the elements are all in the right place.
61 struct BUG_zip_header_must_be_26_bytes {
62 char BUG_zip_header_must_be_26_bytes[
63 offsetof(zip_header_t, formatted.extra_len) + 2
64 == ZIP_HEADER_LEN ? 1 : -1];
67 #define FIX_ENDIANNESS_ZIP(zip_header) do { \
68 (zip_header).formatted.version = SWAP_LE16((zip_header).formatted.version ); \
69 (zip_header).formatted.flags = SWAP_LE16((zip_header).formatted.flags ); \
70 (zip_header).formatted.method = SWAP_LE16((zip_header).formatted.method ); \
71 (zip_header).formatted.modtime = SWAP_LE16((zip_header).formatted.modtime ); \
72 (zip_header).formatted.moddate = SWAP_LE16((zip_header).formatted.moddate ); \
73 (zip_header).formatted.crc32 = SWAP_LE32((zip_header).formatted.crc32 ); \
74 (zip_header).formatted.cmpsize = SWAP_LE32((zip_header).formatted.cmpsize ); \
75 (zip_header).formatted.ucmpsize = SWAP_LE32((zip_header).formatted.ucmpsize ); \
76 (zip_header).formatted.filename_len = SWAP_LE16((zip_header).formatted.filename_len); \
77 (zip_header).formatted.extra_len = SWAP_LE16((zip_header).formatted.extra_len ); \
80 #define CDS_HEADER_LEN 42
83 uint8_t raw[CDS_HEADER_LEN];
85 /* uint32_t signature; 50 4b 01 02 */
86 uint16_t version_made_by; /* 0-1 */
87 uint16_t version_needed; /* 2-3 */
88 uint16_t cds_flags; /* 4-5 */
89 uint16_t method; /* 6-7 */
90 uint16_t mtime; /* 8-9 */
91 uint16_t mdate; /* 10-11 */
92 uint32_t crc32; /* 12-15 */
93 uint32_t cmpsize; /* 16-19 */
94 uint32_t ucmpsize; /* 20-23 */
95 uint16_t file_name_length; /* 24-25 */
96 uint16_t extra_field_length; /* 26-27 */
97 uint16_t file_comment_length; /* 28-29 */
98 uint16_t disk_number_start; /* 30-31 */
99 uint16_t internal_file_attributes; /* 32-33 */
100 uint32_t external_file_attributes PACKED; /* 34-37 */
101 uint32_t relative_offset_of_local_header PACKED; /* 38-41 */
105 struct BUG_cds_header_must_be_42_bytes {
106 char BUG_cds_header_must_be_42_bytes[
107 offsetof(cds_header_t, formatted.relative_offset_of_local_header) + 4
108 == CDS_HEADER_LEN ? 1 : -1];
111 #define FIX_ENDIANNESS_CDS(cds_header) do { \
112 (cds_header).formatted.crc32 = SWAP_LE32((cds_header).formatted.crc32 ); \
113 (cds_header).formatted.cmpsize = SWAP_LE32((cds_header).formatted.cmpsize ); \
114 (cds_header).formatted.ucmpsize = SWAP_LE32((cds_header).formatted.ucmpsize ); \
115 (cds_header).formatted.file_name_length = SWAP_LE16((cds_header).formatted.file_name_length); \
116 (cds_header).formatted.extra_field_length = SWAP_LE16((cds_header).formatted.extra_field_length); \
117 (cds_header).formatted.file_comment_length = SWAP_LE16((cds_header).formatted.file_comment_length); \
120 #define CDE_HEADER_LEN 16
123 uint8_t raw[CDE_HEADER_LEN];
125 /* uint32_t signature; 50 4b 05 06 */
126 uint16_t this_disk_no;
127 uint16_t disk_with_cds_no;
128 uint16_t cds_entries_on_this_disk;
129 uint16_t cds_entries_total;
132 /* uint16_t file_comment_length; */
133 /* .ZIP file comment (variable size) */
137 struct BUG_cde_header_must_be_16_bytes {
138 char BUG_cde_header_must_be_16_bytes[
139 sizeof(cde_header_t) == CDE_HEADER_LEN ? 1 : -1];
142 #define FIX_ENDIANNESS_CDE(cde_header) do { \
143 (cde_header).formatted.cds_offset = SWAP_LE32((cde_header).formatted.cds_offset); \
150 /* NB: does not preserve file position! */
151 static uint32_t find_cds_offset(void)
153 unsigned char buf[1024];
154 cde_header_t cde_header;
158 end = xlseek(zip_fd, 0, SEEK_END);
162 xlseek(zip_fd, end, SEEK_SET);
163 full_read(zip_fd, buf, 1024);
166 while (p <= buf + 1024 - CDE_HEADER_LEN - 4) {
178 memcpy(cde_header.raw, p + 1, CDE_HEADER_LEN);
179 FIX_ENDIANNESS_CDE(cde_header);
180 return cde_header.formatted.cds_offset;
182 bb_error_msg_and_die("can't find file table");
185 static uint32_t read_next_cds(int count_m1, uint32_t cds_offset, cds_header_t *cds_ptr)
189 org = xlseek(zip_fd, 0, SEEK_CUR);
192 cds_offset = find_cds_offset();
194 while (count_m1-- >= 0) {
195 xlseek(zip_fd, cds_offset + 4, SEEK_SET);
196 xread(zip_fd, cds_ptr->raw, CDS_HEADER_LEN);
197 FIX_ENDIANNESS_CDS(*cds_ptr);
198 cds_offset += 4 + CDS_HEADER_LEN
199 + cds_ptr->formatted.file_name_length
200 + cds_ptr->formatted.extra_field_length
201 + cds_ptr->formatted.file_comment_length;
204 xlseek(zip_fd, org, SEEK_SET);
209 static void unzip_skip(off_t skip)
211 if (lseek(zip_fd, skip, SEEK_CUR) == (off_t)-1)
212 bb_copyfd_exact_size(zip_fd, -1, skip);
215 static void unzip_create_leading_dirs(const char *fn)
217 /* Create all leading directories */
218 char *name = xstrdup(fn);
219 if (bb_make_directory(dirname(name), 0777, FILEUTILS_RECUR)) {
220 bb_error_msg_and_die("exiting"); /* bb_make_directory is noisy */
225 static void unzip_extract(zip_header_t *zip_header, int dst_fd)
227 if (zip_header->formatted.method == 0) {
228 /* Method 0 - stored (not compressed) */
229 off_t size = zip_header->formatted.ucmpsize;
231 bb_copyfd_exact_size(zip_fd, dst_fd, size);
233 /* Method 8 - inflate */
234 inflate_unzip_result res;
235 if (inflate_unzip(&res, zip_header->formatted.cmpsize, zip_fd, dst_fd) < 0)
236 bb_error_msg_and_die("inflate error");
237 /* Validate decompression - crc */
238 if (zip_header->formatted.crc32 != (res.crc ^ 0xffffffffL)) {
239 bb_error_msg_and_die("crc error");
241 /* Validate decompression - size */
242 if (zip_header->formatted.ucmpsize != res.bytes_out) {
243 /* Don't die. Who knows, maybe len calculation
244 * was botched somewhere. After all, crc matched! */
245 bb_error_msg("bad length");
250 int unzip_main(int argc, char **argv) MAIN_EXTERNALLY_VISIBLE;
251 int unzip_main(int argc, char **argv)
253 enum { O_PROMPT, O_NEVER, O_ALWAYS };
255 zip_header_t zip_header;
257 IF_NOT_DESKTOP(const) smallint verbose = 0;
258 smallint listing = 0;
259 smallint overwrite = O_PROMPT;
262 unsigned cds_entries;
264 unsigned long total_usize;
265 unsigned long total_size;
266 unsigned total_entries;
270 llist_t *zaccept = NULL;
271 llist_t *zreject = NULL;
272 char *base_dir = NULL;
276 struct stat stat_buf;
278 /* -q, -l and -v: UnZip 5.52 of 28 February 2005, by Info-ZIP:
280 * # /usr/bin/unzip -qq -v decompress_unlzma.i.zip
281 * 204372 Defl:N 35278 83% 09-06-09 14:23 0d056252 decompress_unlzma.i
282 * # /usr/bin/unzip -q -v decompress_unlzma.i.zip
283 * Length Method Size Ratio Date Time CRC-32 Name
284 * -------- ------ ------- ----- ---- ---- ------ ----
285 * 204372 Defl:N 35278 83% 09-06-09 14:23 0d056252 decompress_unlzma.i
286 * -------- ------- --- -------
287 * 204372 35278 83% 1 file
288 * # /usr/bin/unzip -v decompress_unlzma.i.zip
289 * Archive: decompress_unlzma.i.zip
290 * Length Method Size Ratio Date Time CRC-32 Name
291 * -------- ------ ------- ----- ---- ---- ------ ----
292 * 204372 Defl:N 35278 83% 09-06-09 14:23 0d056252 decompress_unlzma.i
293 * -------- ------- --- -------
294 * 204372 35278 83% 1 file
295 * # unzip -v decompress_unlzma.i.zip
296 * Archive: decompress_unlzma.i.zip
297 * Length Date Time Name
298 * -------- ---- ---- ----
299 * 204372 09-06-09 14:23 decompress_unlzma.i
302 * # /usr/bin/unzip -l -qq decompress_unlzma.i.zip
303 * 204372 09-06-09 14:23 decompress_unlzma.i
304 * # /usr/bin/unzip -l -q decompress_unlzma.i.zip
305 * Length Date Time Name
306 * -------- ---- ---- ----
307 * 204372 09-06-09 14:23 decompress_unlzma.i
310 * # /usr/bin/unzip -l decompress_unlzma.i.zip
311 * Archive: decompress_unlzma.i.zip
312 * Length Date Time Name
313 * -------- ---- ---- ----
314 * 204372 09-06-09 14:23 decompress_unlzma.i
319 /* '-' makes getopt return 1 for non-options */
320 while ((opt = getopt(argc, argv, "-d:lnopqxv")) != -1) {
322 case 0: /* Options */
328 case 'n': /* Never overwrite existing files */
332 case 'o': /* Always overwrite existing files */
333 overwrite = O_ALWAYS;
336 case 'p': /* Extract files to stdout and fall through to set verbosity */
337 dst_fd = STDOUT_FILENO;
339 case 'q': /* Be quiet */
343 case 'v': /* Verbose list */
344 IF_DESKTOP(verbose++;)
348 case 1: /* The zip file */
349 /* +5: space for ".zip" and NUL */
350 src_fn = xmalloc(strlen(optarg) + 5);
351 strcpy(src_fn, optarg);
361 case 1: /* Include files */
363 llist_add_to(&zaccept, optarg);
377 case 2 : /* Exclude files */
379 llist_add_to(&zreject, optarg);
382 if (opt == 'd') { /* Extract to base directory */
394 if (src_fn == NULL) {
398 /* Open input file */
399 if (LONE_DASH(src_fn)) {
400 xdup2(STDIN_FILENO, zip_fd);
401 /* Cannot use prompt mode since zip data is arriving on STDIN */
402 if (overwrite == O_PROMPT)
405 static const char extn[][5] = {"", ".zip", ".ZIP"};
406 int orig_src_fn_len = strlen(src_fn);
409 for (i = 0; (i < 3) && (src_fd == -1); i++) {
410 strcpy(src_fn + orig_src_fn_len, extn[i]);
411 src_fd = open(src_fn, O_RDONLY);
414 src_fn[orig_src_fn_len] = '\0';
415 bb_error_msg_and_die("can't open %s, %s.zip, %s.ZIP", src_fn, src_fn, src_fn);
417 xmove_fd(src_fd, zip_fd);
420 /* Change dir if necessary */
424 if (quiet <= 1) { /* not -qq */
426 printf("Archive: %s\n", src_fn);
429 " Length Method Size Ratio Date Time CRC-32 Name\n"
430 "-------- ------ ------- ----- ---- ---- ------ ----"
432 " Length Date Time Name\n"
433 " -------- ---- ---- ----"
448 /* Check magic number */
449 xread(zip_fd, &magic, 4);
450 /* Central directory? It's at the end, so exit */
451 if (magic == ZIP_CDS_MAGIC)
454 /* Data descriptor? It was a streaming file, go on */
455 if (magic == ZIP_DD_MAGIC) {
456 /* skip over duplicate crc32, cmpsize and ucmpsize */
461 if (magic != ZIP_FILEHEADER_MAGIC)
462 bb_error_msg_and_die("invalid zip magic %08X", (int)magic);
464 /* Read the file header */
465 xread(zip_fd, zip_header.raw, ZIP_HEADER_LEN);
466 FIX_ENDIANNESS_ZIP(zip_header);
467 if ((zip_header.formatted.method != 0) && (zip_header.formatted.method != 8)) {
468 bb_error_msg_and_die("unsupported method %d", zip_header.formatted.method);
471 if (zip_header.formatted.flags & 0x0009) {
472 bb_error_msg_and_die("zip flags 1 and 8 are not supported");
475 if (zip_header.formatted.flags & 0x0001) {
476 /* 0x0001 - encrypted */
477 bb_error_msg_and_die("zip flag 1 (encryption) is not supported");
479 if (zip_header.formatted.flags & 0x0008) {
480 cds_header_t cds_header;
481 /* 0x0008 - streaming. [u]cmpsize can be reliably gotten
482 * only from Central Directory. See unzip_doc.txt */
483 cds_offset = read_next_cds(total_entries - cds_entries, cds_offset, &cds_header);
484 cds_entries = total_entries + 1;
485 zip_header.formatted.crc32 = cds_header.formatted.crc32;
486 zip_header.formatted.cmpsize = cds_header.formatted.cmpsize;
487 zip_header.formatted.ucmpsize = cds_header.formatted.ucmpsize;
493 dst_fn = xzalloc(zip_header.formatted.filename_len + 1);
494 xread(zip_fd, dst_fn, zip_header.formatted.filename_len);
496 /* Skip extra header bytes */
497 unzip_skip(zip_header.formatted.extra_len);
499 /* Filter zip entries */
500 if (find_list_entry(zreject, dst_fn)
501 || (zaccept && !find_list_entry(zaccept, dst_fn))
505 } else { /* Extract entry */
506 if (listing) { /* List entry */
507 unsigned dostime = zip_header.formatted.modtime | (zip_header.formatted.moddate << 16);
509 // " Length Date Time Name\n"
510 // " -------- ---- ---- ----"
511 printf( "%9u %02u-%02u-%02u %02u:%02u %s\n",
512 (unsigned)zip_header.formatted.ucmpsize,
513 (dostime & 0x01e00000) >> 21,
514 (dostime & 0x001f0000) >> 16,
515 (((dostime & 0xfe000000) >> 25) + 1980) % 100,
516 (dostime & 0x0000f800) >> 11,
517 (dostime & 0x000007e0) >> 5,
519 total_usize += zip_header.formatted.ucmpsize;
521 unsigned long percents = zip_header.formatted.ucmpsize - zip_header.formatted.cmpsize;
522 percents = percents * 100;
523 if (zip_header.formatted.ucmpsize)
524 percents /= zip_header.formatted.ucmpsize;
525 // " Length Method Size Ratio Date Time CRC-32 Name\n"
526 // "-------- ------ ------- ----- ---- ---- ------ ----"
527 printf( "%8u Defl:N" "%9u%4u%% %02u-%02u-%02u %02u:%02u %08x %s\n",
528 (unsigned)zip_header.formatted.ucmpsize,
529 (unsigned)zip_header.formatted.cmpsize,
531 (dostime & 0x01e00000) >> 21,
532 (dostime & 0x001f0000) >> 16,
533 (((dostime & 0xfe000000) >> 25) + 1980) % 100,
534 (dostime & 0x0000f800) >> 11,
535 (dostime & 0x000007e0) >> 5,
536 zip_header.formatted.crc32,
538 total_usize += zip_header.formatted.ucmpsize;
539 total_size += zip_header.formatted.cmpsize;
542 } else if (dst_fd == STDOUT_FILENO) { /* Extracting to STDOUT */
544 } else if (last_char_is(dst_fn, '/')) { /* Extract directory */
545 if (stat(dst_fn, &stat_buf) == -1) {
546 if (errno != ENOENT) {
547 bb_perror_msg_and_die("can't stat '%s'", dst_fn);
550 printf(" creating: %s\n", dst_fn);
552 unzip_create_leading_dirs(dst_fn);
553 if (bb_make_directory(dst_fn, 0777, 0)) {
554 bb_error_msg_and_die("exiting");
557 if (!S_ISDIR(stat_buf.st_mode)) {
558 bb_error_msg_and_die("'%s' exists but is not directory", dst_fn);
563 } else { /* Extract file */
565 if (stat(dst_fn, &stat_buf) == -1) { /* File does not exist */
566 if (errno != ENOENT) {
567 bb_perror_msg_and_die("can't stat '%s'", dst_fn);
570 } else { /* File already exists */
571 if (overwrite == O_NEVER) {
573 } else if (S_ISREG(stat_buf.st_mode)) { /* File is regular file */
574 if (overwrite == O_ALWAYS) {
577 printf("replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ", dst_fn);
578 if (!fgets(key_buf, sizeof(key_buf), stdin)) {
579 bb_perror_msg_and_die("can't read input");
583 } else { /* File is not regular file */
584 bb_error_msg_and_die("'%s' exists but is not regular file", dst_fn);
592 overwrite = O_ALWAYS;
593 case 'y': /* Open file and fall into unzip */
594 unzip_create_leading_dirs(dst_fn);
595 dst_fd = xopen(dst_fn, O_WRONLY | O_CREAT | O_TRUNC);
598 printf(" inflating: %s\n", dst_fn);
600 unzip_extract(&zip_header, dst_fd);
601 if (dst_fd != STDOUT_FILENO) {
602 /* closing STDOUT is potentially bad for future business */
610 /* Skip entry data */
611 unzip_skip(zip_header.formatted.cmpsize);
615 /* Prompt for new name */
616 printf("new name: ");
617 if (!fgets(key_buf, sizeof(key_buf), stdin)) {
618 bb_perror_msg_and_die("can't read input");
621 dst_fn = xstrdup(key_buf);
626 printf("error: invalid response [%c]\n", (char)i);
633 if (listing && quiet <= 1) {
635 // " Length Date Time Name\n"
636 // " -------- ---- ---- ----"
637 printf( " -------- -------\n"
638 "%9lu" " %u files\n",
639 total_usize, total_entries);
641 unsigned long percents = total_usize - total_size;
642 percents = percents * 100;
644 percents /= total_usize;
645 // " Length Method Size Ratio Date Time CRC-32 Name\n"
646 // "-------- ------ ------- ----- ---- ---- ------ ----"
647 printf( "-------- ------- --- -------\n"
648 "%8lu" "%17lu%4u%% %u files\n",
649 total_usize, total_size, (unsigned)percents,