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 bb_copyfd_exact_size(zip_fd
, -1, skip
);
214 static void unzip_create_leading_dirs(const char *fn
)
216 /* Create all leading directories */
217 char *name
= xstrdup(fn
);
218 if (bb_make_directory(dirname(name
), 0777, FILEUTILS_RECUR
)) {
219 bb_error_msg_and_die("exiting"); /* bb_make_directory is noisy */
224 static void unzip_extract(zip_header_t
*zip_header
, int dst_fd
)
226 if (zip_header
->formatted
.method
== 0) {
227 /* Method 0 - stored (not compressed) */
228 off_t size
= zip_header
->formatted
.ucmpsize
;
230 bb_copyfd_exact_size(zip_fd
, dst_fd
, size
);
232 /* Method 8 - inflate */
233 inflate_unzip_result res
;
234 if (inflate_unzip(&res
, zip_header
->formatted
.cmpsize
, zip_fd
, dst_fd
) < 0)
235 bb_error_msg_and_die("inflate error");
236 /* Validate decompression - crc */
237 if (zip_header
->formatted
.crc32
!= (res
.crc
^ 0xffffffffL
)) {
238 bb_error_msg_and_die("crc error");
240 /* Validate decompression - size */
241 if (zip_header
->formatted
.ucmpsize
!= res
.bytes_out
) {
242 /* Don't die. Who knows, maybe len calculation
243 * was botched somewhere. After all, crc matched! */
244 bb_error_msg("bad length");
249 int unzip_main(int argc
, char **argv
) MAIN_EXTERNALLY_VISIBLE
;
250 int unzip_main(int argc
, char **argv
)
252 enum { O_PROMPT
, O_NEVER
, O_ALWAYS
};
254 zip_header_t zip_header
;
255 smallint verbose
= 1;
256 smallint listing
= 0;
257 smallint overwrite
= O_PROMPT
;
260 unsigned cds_entries
;
263 unsigned total_entries
;
267 llist_t
*zaccept
= NULL
;
268 llist_t
*zreject
= NULL
;
269 char *base_dir
= NULL
;
273 struct stat stat_buf
;
275 /* '-' makes getopt return 1 for non-options */
276 while ((opt
= getopt(argc
, argv
, "-d:lnopqx")) != -1) {
278 case 0: /* Options */
284 case 'n': /* Never overwrite existing files */
288 case 'o': /* Always overwrite existing files */
289 overwrite
= O_ALWAYS
;
292 case 'p': /* Extract files to stdout and fall through to set verbosity */
293 dst_fd
= STDOUT_FILENO
;
295 case 'q': /* Be quiet */
299 case 1: /* The zip file */
300 /* +5: space for ".zip" and NUL */
301 src_fn
= xmalloc(strlen(optarg
) + 5);
302 strcpy(src_fn
, optarg
);
312 case 1: /* Include files */
314 llist_add_to(&zaccept
, optarg
);
328 case 2 : /* Exclude files */
330 llist_add_to(&zreject
, optarg
);
333 if (opt
== 'd') { /* Extract to base directory */
345 if (src_fn
== NULL
) {
349 /* Open input file */
350 if (LONE_DASH(src_fn
)) {
351 xdup2(STDIN_FILENO
, zip_fd
);
352 /* Cannot use prompt mode since zip data is arriving on STDIN */
353 if (overwrite
== O_PROMPT
)
356 static const char extn
[][5] = {"", ".zip", ".ZIP"};
357 int orig_src_fn_len
= strlen(src_fn
);
360 for (i
= 0; (i
< 3) && (src_fd
== -1); i
++) {
361 strcpy(src_fn
+ orig_src_fn_len
, extn
[i
]);
362 src_fd
= open(src_fn
, O_RDONLY
);
365 src_fn
[orig_src_fn_len
] = '\0';
366 bb_error_msg_and_die("can't open %s, %s.zip, %s.ZIP", src_fn
, src_fn
, src_fn
);
368 xmove_fd(src_fd
, zip_fd
);
371 /* Change dir if necessary */
376 printf("Archive: %s\n", src_fn
);
378 puts(" Length Date Time Name\n"
379 " -------- ---- ---- ----");
392 /* Check magic number */
393 xread(zip_fd
, &magic
, 4);
394 /* Central directory? It's at the end, so exit */
395 if (magic
== ZIP_CDS_MAGIC
)
398 /* Data descriptor? It was a streaming file, go on */
399 if (magic
== ZIP_DD_MAGIC
) {
400 /* skip over duplicate crc32, cmpsize and ucmpsize */
405 if (magic
!= ZIP_FILEHEADER_MAGIC
)
406 bb_error_msg_and_die("invalid zip magic %08X", (int)magic
);
408 /* Read the file header */
409 xread(zip_fd
, zip_header
.raw
, ZIP_HEADER_LEN
);
410 FIX_ENDIANNESS_ZIP(zip_header
);
411 if ((zip_header
.formatted
.method
!= 0) && (zip_header
.formatted
.method
!= 8)) {
412 bb_error_msg_and_die("unsupported method %d", zip_header
.formatted
.method
);
415 if (zip_header
.formatted
.flags
& 0x0009) {
416 bb_error_msg_and_die("zip flags 1 and 8 are not supported");
419 if (zip_header
.formatted
.flags
& 0x0001) {
420 /* 0x0001 - encrypted */
421 bb_error_msg_and_die("zip flag 1 (encryption) is not supported");
423 if (zip_header
.formatted
.flags
& 0x0008) {
424 cds_header_t cds_header
;
425 /* 0x0008 - streaming. [u]cmpsize can be reliably gotten
426 * only from Central Directory. See unzip_doc.txt */
427 cds_offset
= read_next_cds(total_entries
- cds_entries
, cds_offset
, &cds_header
);
428 cds_entries
= total_entries
+ 1;
429 zip_header
.formatted
.crc32
= cds_header
.formatted
.crc32
;
430 zip_header
.formatted
.cmpsize
= cds_header
.formatted
.cmpsize
;
431 zip_header
.formatted
.ucmpsize
= cds_header
.formatted
.ucmpsize
;
437 dst_fn
= xzalloc(zip_header
.formatted
.filename_len
+ 1);
438 xread(zip_fd
, dst_fn
, zip_header
.formatted
.filename_len
);
440 /* Skip extra header bytes */
441 unzip_skip(zip_header
.formatted
.extra_len
);
443 /* Filter zip entries */
444 if (find_list_entry(zreject
, dst_fn
)
445 || (zaccept
&& !find_list_entry(zaccept
, dst_fn
))
449 } else { /* Extract entry */
450 if (listing
) { /* List entry */
452 unsigned dostime
= zip_header
.formatted
.modtime
| (zip_header
.formatted
.moddate
<< 16);
453 printf("%9u %02u-%02u-%02u %02u:%02u %s\n",
454 zip_header
.formatted
.ucmpsize
,
455 (dostime
& 0x01e00000) >> 21,
456 (dostime
& 0x001f0000) >> 16,
457 (((dostime
& 0xfe000000) >> 25) + 1980) % 100,
458 (dostime
& 0x0000f800) >> 11,
459 (dostime
& 0x000007e0) >> 5,
461 total_size
+= zip_header
.formatted
.ucmpsize
;
463 /* short listing -- filenames only */
467 } else if (dst_fd
== STDOUT_FILENO
) { /* Extracting to STDOUT */
469 } else if (last_char_is(dst_fn
, '/')) { /* Extract directory */
470 if (stat(dst_fn
, &stat_buf
) == -1) {
471 if (errno
!= ENOENT
) {
472 bb_perror_msg_and_die("can't stat '%s'", dst_fn
);
475 printf(" creating: %s\n", dst_fn
);
477 unzip_create_leading_dirs(dst_fn
);
478 if (bb_make_directory(dst_fn
, 0777, 0)) {
479 bb_error_msg_and_die("exiting");
482 if (!S_ISDIR(stat_buf
.st_mode
)) {
483 bb_error_msg_and_die("'%s' exists but is not directory", dst_fn
);
488 } else { /* Extract file */
490 if (stat(dst_fn
, &stat_buf
) == -1) { /* File does not exist */
491 if (errno
!= ENOENT
) {
492 bb_perror_msg_and_die("can't stat '%s'", dst_fn
);
495 } else { /* File already exists */
496 if (overwrite
== O_NEVER
) {
498 } else if (S_ISREG(stat_buf
.st_mode
)) { /* File is regular file */
499 if (overwrite
== O_ALWAYS
) {
502 printf("replace %s? [y]es, [n]o, [A]ll, [N]one, [r]ename: ", dst_fn
);
503 if (!fgets(key_buf
, sizeof(key_buf
), stdin
)) {
504 bb_perror_msg_and_die("can't read input");
508 } else { /* File is not regular file */
509 bb_error_msg_and_die("'%s' exists but is not regular file", dst_fn
);
517 overwrite
= O_ALWAYS
;
518 case 'y': /* Open file and fall into unzip */
519 unzip_create_leading_dirs(dst_fn
);
520 dst_fd
= xopen(dst_fn
, O_WRONLY
| O_CREAT
| O_TRUNC
);
523 printf(" inflating: %s\n", dst_fn
);
525 unzip_extract(&zip_header
, dst_fd
);
526 if (dst_fd
!= STDOUT_FILENO
) {
527 /* closing STDOUT is potentially bad for future business */
535 /* Skip entry data */
536 unzip_skip(zip_header
.formatted
.cmpsize
);
540 /* Prompt for new name */
541 printf("new name: ");
542 if (!fgets(key_buf
, sizeof(key_buf
), stdin
)) {
543 bb_perror_msg_and_die("can't read input");
546 dst_fn
= xstrdup(key_buf
);
551 printf("error: invalid response [%c]\n",(char)i
);
558 if (listing
&& verbose
) {
559 printf(" -------- -------\n"
561 total_size
, total_entries
);