archiver_zip.c
1
2
3
/*
* ZIP support routines for PhysicsFS.
*
4
* Please see the file LICENSE.txt in the source's root directory.
6
7
* This file written by Ryan C. Gordon, with some peeking at "unzip.c"
* by Gilles Vollant.
10
11
12
13
#define __PHYSICSFS_INTERNAL__
#include "physfs_internal.h"
#if PHYSFS_SUPPORTS_ZIP
15
#include <errno.h>
16
#include <time.h>
18
19
#include "physfs_miniz.h"
21
22
* A buffer of ZIP_READBUFSIZE is allocated for each compressed file opened,
* and is freed when you close the file; compressed data is read into
23
24
25
26
27
28
29
30
31
32
33
* this buffer, and then is decompressed into the buffer passed to
* PHYSFS_read().
*
* Uncompressed entries in a zipfile do not allocate this buffer; they just
* read data directly into the buffer passed to PHYSFS_read().
*
* Depending on your speed and memory requirements, you should tweak this
* value.
*/
#define ZIP_READBUFSIZE (16 * 1024)
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
/*
* Entries are "unresolved" until they are first opened. At that time,
* local file headers parsed/validated, data offsets will be updated to look
* at the actual file data instead of the header, and symlinks will be
* followed and optimized. This means that we don't seek and read around the
* archive until forced to do so, and after the first time, we had to do
* less reading and parsing, which is very CD-ROM friendly.
*/
typedef enum
{
ZIP_UNRESOLVED_FILE,
ZIP_UNRESOLVED_SYMLINK,
ZIP_RESOLVING,
ZIP_RESOLVED,
49
ZIP_DIRECTORY,
50
ZIP_BROKEN_FILE,
51
ZIP_BROKEN_SYMLINK
52
53
54
} ZipResolveType;
55
56
57
/*
* One ZIPentry is kept for each file in an open ZIP archive.
*/
58
typedef struct _ZIPentry
60
61
62
char *name; /* Name of file in archive */
struct _ZIPentry *symlink; /* NULL or file we symlink to */
ZipResolveType resolved; /* Have we resolved file/symlink? */
63
PHYSFS_uint64 offset; /* offset of data in archive */
64
65
PHYSFS_uint16 version; /* version made by */
PHYSFS_uint16 version_needed; /* version needed to extract */
66
PHYSFS_uint16 general_bits; /* general purpose bits */
67
68
PHYSFS_uint16 compression_method; /* compression method */
PHYSFS_uint32 crc; /* crc-32 */
69
70
PHYSFS_uint64 compressed_size; /* compressed size */
PHYSFS_uint64 uncompressed_size; /* uncompressed size */
71
PHYSFS_sint64 last_mod_time; /* last file mod time */
72
PHYSFS_uint32 dos_mod_time; /* original MS-DOS style mod time */
73
74
75
struct _ZIPentry *hashnext; /* next item in this hash bucket */
struct _ZIPentry *children; /* linked list of kids, if dir */
struct _ZIPentry *sibling; /* next item in same dir */
78
79
80
/*
* One ZIPinfo is kept for each open ZIP archive.
*/
81
82
typedef struct
{
83
84
85
86
87
88
PHYSFS_Io *io; /* the i/o interface for this archive. */
ZIPentry root; /* root of directory tree. */
ZIPentry **hash; /* all entries hashed for fast lookup. */
size_t hashBuckets; /* number of buckets in hash. */
int zip64; /* non-zero if this is a Zip64 archive. */
int has_crypto; /* non-zero if any entry uses encryption. */
91
92
93
/*
* One ZIPfileinfo is kept for each open file in a ZIP archive.
*/
94
95
typedef struct
{
96
ZIPentry *entry; /* Info on file. */
97
PHYSFS_Io *io; /* physical file handle. */
98
99
100
PHYSFS_uint32 compressed_position; /* offset in compressed data. */
PHYSFS_uint32 uncompressed_position; /* tell() position. */
PHYSFS_uint8 *buffer; /* decompression buffer. */
101
102
PHYSFS_uint32 crypto_keys[3]; /* for "traditional" crypto. */
PHYSFS_uint32 initial_crypto_keys[3]; /* for "traditional" crypto. */
103
z_stream stream; /* zlib stream state. */
104
105
106
} ZIPfileinfo;
107
/* Magic numbers... */
108
109
110
111
112
113
#define ZIP_LOCAL_FILE_SIG 0x04034b50
#define ZIP_CENTRAL_DIR_SIG 0x02014b50
#define ZIP_END_OF_CENTRAL_DIR_SIG 0x06054b50
#define ZIP64_END_OF_CENTRAL_DIR_SIG 0x06064b50
#define ZIP64_END_OF_CENTRAL_DIRECTORY_LOCATOR_SIG 0x07064b50
#define ZIP64_EXTENDED_INFO_EXTRA_FIELD_SIG 0x0001
114
115
116
117
118
119
/* compression methods... */
#define COMPMETH_NONE 0
/* ...and others... */
120
121
122
#define UNIX_FILETYPE_MASK 0170000
#define UNIX_FILETYPE_SYMLINK 0120000
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
#define ZIP_GENERAL_BITS_TRADITIONAL_CRYPTO (1 << 0)
#define ZIP_GENERAL_BITS_IGNORE_LOCAL_HEADER (1 << 3)
/* support for "traditional" PKWARE encryption. */
static int zip_entry_is_tradional_crypto(const ZIPentry *entry)
{
return (entry->general_bits & ZIP_GENERAL_BITS_TRADITIONAL_CRYPTO) != 0;
} /* zip_entry_is_traditional_crypto */
static int zip_entry_ignore_local_header(const ZIPentry *entry)
{
return (entry->general_bits & ZIP_GENERAL_BITS_IGNORE_LOCAL_HEADER) != 0;
} /* zip_entry_is_traditional_crypto */
static PHYSFS_uint32 zip_crypto_crc32(const PHYSFS_uint32 crc, const PHYSFS_uint8 val)
{
int i;
PHYSFS_uint32 xorval = (crc ^ ((PHYSFS_uint32) val)) & 0xFF;
for (i = 0; i < 8; i++)
xorval = ((xorval & 1) ? (0xEDB88320 ^ (xorval >> 1)) : (xorval >> 1));
return xorval ^ (crc >> 8);
} /* zip_crc32 */
static void zip_update_crypto_keys(PHYSFS_uint32 *keys, const PHYSFS_uint8 val)
{
keys[0] = zip_crypto_crc32(keys[0], val);
keys[1] = keys[1] + (keys[0] & 0x000000FF);
keys[1] = (keys[1] * 134775813) + 1;
keys[2] = zip_crypto_crc32(keys[2], (PHYSFS_uint8) ((keys[1] >> 24) & 0xFF));
} /* zip_update_crypto_keys */
static PHYSFS_uint8 zip_decrypt_byte(const PHYSFS_uint32 *keys)
{
const PHYSFS_uint16 tmp = keys[2] | 2;
return (PHYSFS_uint8) ((tmp * (tmp ^ 1)) >> 8);
} /* zip_decrypt_byte */
static PHYSFS_sint64 zip_read_decrypt(ZIPfileinfo *finfo, void *buf, PHYSFS_uint64 len)
{
PHYSFS_Io *io = finfo->io;
const PHYSFS_sint64 br = io->read(io, buf, len);
/* Decompression the new data if necessary. */
if (zip_entry_is_tradional_crypto(finfo->entry) && (br > 0))
{
PHYSFS_uint32 *keys = finfo->crypto_keys;
PHYSFS_uint8 *ptr = (PHYSFS_uint8 *) buf;
PHYSFS_sint64 i;
for (i = 0; i < br; i++, ptr++)
{
const PHYSFS_uint8 ch = *ptr ^ zip_decrypt_byte(keys);
zip_update_crypto_keys(keys, ch);
*ptr = ch;
} /* for */
} /* if */
return br;
} /* zip_read_decrypt */
static int zip_prep_crypto_keys(ZIPfileinfo *finfo, const PHYSFS_uint8 *crypto_header, const PHYSFS_uint8 *password)
{
/* It doesn't appear to be documented in PKWare's APPNOTE.TXT, but you
need to use a different byte in the header to verify the password
if general purpose bit 3 is set. Discovered this from Info-Zip.
That's what the (verifier) value is doing, below. */
PHYSFS_uint32 *keys = finfo->crypto_keys;
const ZIPentry *entry = finfo->entry;
const int usedate = zip_entry_ignore_local_header(entry);
const PHYSFS_uint8 verifier = (PHYSFS_uint8) ((usedate ? (entry->dos_mod_time >> 8) : (entry->crc >> 24)) & 0xFF);
PHYSFS_uint8 finalbyte = 0;
int i = 0;
/* initialize vector with defaults, then password, then header. */
keys[0] = 305419896;
keys[1] = 591751049;
keys[2] = 878082192;
while (*password)
zip_update_crypto_keys(keys, *(password++));
for (i = 0; i < 12; i++)
{
const PHYSFS_uint8 c = crypto_header[i] ^ zip_decrypt_byte(keys);
zip_update_crypto_keys(keys, c);
finalbyte = c;
} /* for */
/* you have a 1/256 chance of passing this test incorrectly. :/ */
if (finalbyte != verifier)
213
BAIL(PHYSFS_ERR_BAD_PASSWORD, 0);
214
215
216
217
218
219
/* save the initial vector for seeking purposes. Not secure!! */
memcpy(finfo->initial_crypto_keys, finfo->crypto_keys, 12);
return 1;
} /* zip_prep_crypto_keys */
221
222
223
224
225
/*
* Bridge physfs allocation functions to zlib's format...
*/
static voidpf zlibPhysfsAlloc(voidpf opaque, uInt items, uInt size)
{
226
return ((PHYSFS_Allocator *) opaque)->Malloc(items * size);
227
228
229
230
231
232
233
} /* zlibPhysfsAlloc */
/*
* Bridge physfs allocation functions to zlib's format...
*/
static void zlibPhysfsFree(voidpf opaque, voidpf address)
{
234
((PHYSFS_Allocator *) opaque)->Free(address);
235
236
237
238
239
240
241
242
243
244
245
} /* zlibPhysfsFree */
/*
* Construct a new z_stream to a sane state.
*/
static void initializeZStream(z_stream *pstr)
{
memset(pstr, '\0', sizeof (z_stream));
pstr->zalloc = zlibPhysfsAlloc;
pstr->zfree = zlibPhysfsFree;
246
pstr->opaque = &allocator;
247
248
249
} /* initializeZStream */
250
static PHYSFS_ErrorCode zlib_error_code(int rc)
251
252
253
{
switch (rc)
{
254
255
256
257
258
case Z_OK: return PHYSFS_ERR_OK; /* not an error. */
case Z_STREAM_END: return PHYSFS_ERR_OK; /* not an error. */
case Z_ERRNO: return PHYSFS_ERR_IO;
case Z_MEM_ERROR: return PHYSFS_ERR_OUT_OF_MEMORY;
default: return PHYSFS_ERR_CORRUPT;
259
} /* switch */
260
261
} /* zlib_error_string */
263
264
265
/*
* Wrap all zlib calls in this, so the physfs error state is set appropriately.
*/
266
static int zlib_err(const int rc)
268
PHYSFS_setErrorCode(zlib_error_code(rc));
270
271
} /* zlib_err */
272
273
274
275
276
277
278
/*
* Hash a string for lookup an a ZIPinfo hashtable.
*/
static inline PHYSFS_uint32 zip_hash_string(const ZIPinfo *info, const char *s)
{
return __PHYSFS_hashString(s, strlen(s)) % info->hashBuckets;
} /* zip_hash_string */
280
281
282
283
284
285
/*
* Read an unsigned 64-bit int and swap to native byte order.
*/
static int readui64(PHYSFS_Io *io, PHYSFS_uint64 *val)
{
PHYSFS_uint64 v;
286
BAIL_IF_ERRPASS(!__PHYSFS_readAll(io, &v, sizeof (v)), 0);
287
288
289
290
*val = PHYSFS_swapULE64(v);
return 1;
} /* readui64 */
291
292
293
/*
* Read an unsigned 32-bit int and swap to native byte order.
*/
294
static int readui32(PHYSFS_Io *io, PHYSFS_uint32 *val)
295
296
{
PHYSFS_uint32 v;
297
BAIL_IF_ERRPASS(!__PHYSFS_readAll(io, &v, sizeof (v)), 0);
298
*val = PHYSFS_swapULE32(v);
300
301
302
303
304
305
} /* readui32 */
/*
* Read an unsigned 16-bit int and swap to native byte order.
*/
306
static int readui16(PHYSFS_Io *io, PHYSFS_uint16 *val)
307
308
{
PHYSFS_uint16 v;
309
BAIL_IF_ERRPASS(!__PHYSFS_readAll(io, &v, sizeof (v)), 0);
310
*val = PHYSFS_swapULE16(v);
312
313
314
} /* readui16 */
315
static PHYSFS_sint64 ZIP_read(PHYSFS_Io *_io, void *buf, PHYSFS_uint64 len)
317
ZIPfileinfo *finfo = (ZIPfileinfo *) _io->opaque;
318
319
ZIPentry *entry = finfo->entry;
PHYSFS_sint64 retval = 0;
320
PHYSFS_sint64 maxread = (PHYSFS_sint64) len;
321
322
323
324
PHYSFS_sint64 avail = entry->uncompressed_size -
finfo->uncompressed_position;
if (avail < maxread)
325
maxread = avail;
327
BAIL_IF_ERRPASS(maxread == 0, 0); /* quick rejection. */
329
if (entry->compression_method == COMPMETH_NONE)
330
retval = zip_read_decrypt(finfo, buf, maxread);
331
332
333
else
{
finfo->stream.next_out = buf;
334
finfo->stream.avail_out = (uInt) maxread;
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
while (retval < maxread)
{
PHYSFS_uint32 before = finfo->stream.total_out;
int rc;
if (finfo->stream.avail_in == 0)
{
PHYSFS_sint64 br;
br = entry->compressed_size - finfo->compressed_position;
if (br > 0)
{
if (br > ZIP_READBUFSIZE)
br = ZIP_READBUFSIZE;
351
br = zip_read_decrypt(finfo, finfo->buffer, (PHYSFS_uint64) br);
352
353
354
if (br <= 0)
break;
355
finfo->compressed_position += (PHYSFS_uint32) br;
356
finfo->stream.next_in = finfo->buffer;
357
finfo->stream.avail_in = (PHYSFS_uint32) br;
358
359
} /* if */
} /* if */
361
362
rc = zlib_err(inflate(&finfo->stream, Z_SYNC_FLUSH));
retval += (finfo->stream.total_out - before);
364
365
366
367
368
369
if (rc != Z_OK)
break;
} /* while */
} /* else */
if (retval > 0)
370
finfo->uncompressed_position += (PHYSFS_uint32) retval;
372
return retval;
373
374
375
} /* ZIP_read */
376
static PHYSFS_sint64 ZIP_write(PHYSFS_Io *io, const void *b, PHYSFS_uint64 len)
378
BAIL(PHYSFS_ERR_READ_ONLY, -1);
379
380
381
} /* ZIP_write */
382
static PHYSFS_sint64 ZIP_tell(PHYSFS_Io *io)
384
return ((ZIPfileinfo *) io->opaque)->uncompressed_position;
385
386
387
} /* ZIP_tell */
388
static int ZIP_seek(PHYSFS_Io *_io, PHYSFS_uint64 offset)
390
ZIPfileinfo *finfo = (ZIPfileinfo *) _io->opaque;
391
ZIPentry *entry = finfo->entry;
392
PHYSFS_Io *io = finfo->io;
393
const int encrypted = zip_entry_is_tradional_crypto(entry);
395
BAIL_IF(offset > entry->uncompressed_size, PHYSFS_ERR_PAST_EOF, 0);
397
if (!encrypted && (entry->compression_method == COMPMETH_NONE))
399
PHYSFS_sint64 newpos = offset + entry->offset;
400
BAIL_IF_ERRPASS(!io->seek(io, newpos), 0);
401
finfo->uncompressed_position = (PHYSFS_uint32) offset;
402
} /* if */
404
else
406
407
408
409
/*
* If seeking backwards, we need to redecode the file
* from the start and throw away the compressed bits until we hit
* the offset we need. If seeking forward, we still need to
410
* decode, but we don't rewind first.
411
412
413
414
415
*/
if (offset < finfo->uncompressed_position)
{
/* we do a copy so state is sane if inflateInit2() fails. */
z_stream str;
416
initializeZStream(&str);
417
if (zlib_err(inflateInit2(&str, -MAX_WBITS)) != Z_OK)
420
if (!io->seek(io, entry->offset + (encrypted ? 12 : 0)))
423
424
425
inflateEnd(&finfo->stream);
memcpy(&finfo->stream, &str, sizeof (z_stream));
finfo->uncompressed_position = finfo->compressed_position = 0;
426
427
428
if (encrypted)
memcpy(finfo->crypto_keys, finfo->initial_crypto_keys, 12);
429
430
431
432
433
} /* if */
while (finfo->uncompressed_position != offset)
{
PHYSFS_uint8 buf[512];
434
435
436
PHYSFS_uint32 maxread;
maxread = (PHYSFS_uint32) (offset - finfo->uncompressed_position);
437
438
439
if (maxread > sizeof (buf))
maxread = sizeof (buf);
440
if (ZIP_read(_io, buf, maxread) != maxread)
442
443
444
} /* while */
} /* else */
446
447
448
} /* ZIP_seek */
449
static PHYSFS_sint64 ZIP_length(PHYSFS_Io *io)
451
const ZIPfileinfo *finfo = (ZIPfileinfo *) io->opaque;
452
return (PHYSFS_sint64) finfo->entry->uncompressed_size;
453
} /* ZIP_length */
456
457
458
static PHYSFS_Io *zip_get_io(PHYSFS_Io *io, ZIPinfo *inf, ZIPentry *entry);
static PHYSFS_Io *ZIP_duplicate(PHYSFS_Io *io)
460
461
462
ZIPfileinfo *origfinfo = (ZIPfileinfo *) io->opaque;
PHYSFS_Io *retval = (PHYSFS_Io *) allocator.Malloc(sizeof (PHYSFS_Io));
ZIPfileinfo *finfo = (ZIPfileinfo *) allocator.Malloc(sizeof (ZIPfileinfo));
463
464
GOTO_IF(!retval, PHYSFS_ERR_OUT_OF_MEMORY, failed);
GOTO_IF(!finfo, PHYSFS_ERR_OUT_OF_MEMORY, failed);
465
466
467
468
memset(finfo, '\0', sizeof (*finfo));
finfo->entry = origfinfo->entry;
finfo->io = zip_get_io(origfinfo->io, NULL, finfo->entry);
469
GOTO_IF_ERRPASS(!finfo->io, failed);
470
471
472
473
if (finfo->entry->compression_method != COMPMETH_NONE)
{
finfo->buffer = (PHYSFS_uint8 *) allocator.Malloc(ZIP_READBUFSIZE);
474
GOTO_IF(!finfo->buffer, PHYSFS_ERR_OUT_OF_MEMORY, failed);
475
476
if (zlib_err(inflateInit2(&finfo->stream, -MAX_WBITS)) != Z_OK)
goto failed;
477
478
479
480
481
482
} /* if */
memcpy(retval, io, sizeof (PHYSFS_Io));
retval->opaque = finfo;
return retval;
483
failed:
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
if (finfo != NULL)
{
if (finfo->io != NULL)
finfo->io->destroy(finfo->io);
if (finfo->buffer != NULL)
{
allocator.Free(finfo->buffer);
inflateEnd(&finfo->stream);
} /* if */
allocator.Free(finfo);
} /* if */
if (retval != NULL)
allocator.Free(retval);
return NULL;
} /* ZIP_duplicate */
static int ZIP_flush(PHYSFS_Io *io) { return 1; /* no write support. */ }
static void ZIP_destroy(PHYSFS_Io *io)
{
ZIPfileinfo *finfo = (ZIPfileinfo *) io->opaque;
finfo->io->destroy(finfo->io);
510
511
512
513
514
if (finfo->entry->compression_method != COMPMETH_NONE)
inflateEnd(&finfo->stream);
if (finfo->buffer != NULL)
515
allocator.Free(finfo->buffer);
517
allocator.Free(finfo);
518
519
520
521
522
523
allocator.Free(io);
} /* ZIP_destroy */
static const PHYSFS_Io ZIP_Io =
{
524
CURRENT_PHYSFS_IO_API_VERSION, NULL,
525
526
527
528
529
530
531
ZIP_read,
ZIP_write,
ZIP_seek,
ZIP_tell,
ZIP_length,
ZIP_duplicate,
ZIP_flush,
532
ZIP_destroy
537
static PHYSFS_sint64 zip_find_end_of_central_dir(PHYSFS_Io *io, PHYSFS_sint64 *len)
539
PHYSFS_uint8 buf[256];
540
PHYSFS_uint8 extra[4] = { 0, 0, 0, 0 };
541
PHYSFS_sint32 i = 0;
542
543
544
PHYSFS_sint64 filelen;
PHYSFS_sint64 filepos;
PHYSFS_sint32 maxread;
545
546
PHYSFS_sint32 totalread = 0;
int found = 0;
548
filelen = io->length(io);
549
BAIL_IF_ERRPASS(filelen == -1, -1);
550
551
552
553
554
555
556
557
558
/*
* Jump to the end of the file and start reading backwards.
* The last thing in the file is the zipfile comment, which is variable
* length, and the field that specifies its size is before it in the
* file (argh!)...this means that we need to scan backwards until we
* hit the end-of-central-dir signature. We can then sanity check that
* the comment was as big as it should be to make sure we're in the
* right place. The comment length field is 16 bits, so we can stop
559
560
* searching for that signature after a little more than 64k at most,
* and call it a corrupted zipfile.
561
562
563
564
565
566
567
568
569
570
*/
if (sizeof (buf) < filelen)
{
filepos = filelen - sizeof (buf);
maxread = sizeof (buf);
} /* if */
else
{
filepos = 0;
571
maxread = (PHYSFS_uint32) filelen;
572
573
} /* else */
574
while ((totalread < filelen) && (totalread < 65557))
576
BAIL_IF_ERRPASS(!io->seek(io, filepos), -1);
577
578
579
/* make sure we catch a signature between buffers. */
if (totalread != 0)
581
if (!__PHYSFS_readAll(io, buf, maxread - 4))
583
memcpy(&buf[maxread - 4], &extra, sizeof (extra));
584
totalread += maxread - 4;
585
} /* if */
588
if (!__PHYSFS_readAll(io, buf, maxread))
590
591
592
totalread += maxread;
} /* else */
593
memcpy(&extra, buf, sizeof (extra));
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
for (i = maxread - 4; i > 0; i--)
{
if ((buf[i + 0] == 0x50) &&
(buf[i + 1] == 0x4B) &&
(buf[i + 2] == 0x05) &&
(buf[i + 3] == 0x06) )
{
found = 1; /* that's the signature! */
break;
} /* if */
} /* for */
if (found)
break;
filepos -= (maxread - 4);
611
612
if (filepos < 0)
filepos = 0;
613
614
} /* while */
615
BAIL_IF(!found, PHYSFS_ERR_UNSUPPORTED, -1);
616
617
618
619
if (len != NULL)
*len = filelen;
620
return (filepos + i);
621
} /* zip_find_end_of_central_dir */
622
623
624
static int isZip(PHYSFS_Io *io)
626
PHYSFS_uint32 sig = 0;
627
int retval = 0;
628
629
630
631
632
/*
* The first thing in a zip file might be the signature of the
* first local file record, so it makes for a quick determination.
*/
633
if (readui32(io, &sig))
635
636
637
638
639
640
641
642
retval = (sig == ZIP_LOCAL_FILE_SIG);
if (!retval)
{
/*
* No sig...might be a ZIP with data at the start
* (a self-extracting executable, etc), so we'll have to do
* it the hard way...
*/
643
retval = (zip_find_end_of_central_dir(io, NULL) != -1);
644
} /* if */
647
return retval;
648
} /* isZip */
651
652
/* Find the ZIPentry for a path in platform-independent notation. */
static ZIPentry *zip_find_entry(ZIPinfo *info, const char *path)
654
655
656
PHYSFS_uint32 hashval;
ZIPentry *prev = NULL;
ZIPentry *retval;
658
659
if (*path == '\0')
return &info->root;
661
662
hashval = zip_hash_string(info, path);
for (retval = info->hash[hashval]; retval; retval = retval->hashnext)
664
if (strcmp(retval->name, path) == 0)
666
if (prev != NULL) /* move this to the front of the list */
668
669
670
prev->hashnext = retval->hashnext;
retval->hashnext = info->hash[hashval];
info->hash[hashval] = retval;
671
672
} /* if */
673
return retval;
674
} /* if */
676
677
prev = retval;
} /* for */
679
BAIL(PHYSFS_ERR_NOT_FOUND, NULL);
680
} /* zip_find_entry */
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
/* Convert paths from old, buggy DOS zippers... */
static void zip_convert_dos_path(ZIPentry *entry, char *path)
{
PHYSFS_uint8 hosttype = (PHYSFS_uint8) ((entry->version >> 8) & 0xFF);
if (hosttype == 0) /* FS_FAT_ */
{
while (*path)
{
if (*path == '\\')
*path = '/';
path++;
} /* while */
} /* if */
} /* zip_convert_dos_path */
697
698
699
static void zip_expand_symlink_path(char *path)
701
702
char *ptr = path;
char *prevptr = path;
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
ptr = strchr(ptr, '/');
if (ptr == NULL)
break;
if (*(ptr + 1) == '.')
{
if (*(ptr + 2) == '/')
{
/* current dir in middle of string: ditch it. */
memmove(ptr, ptr + 2, strlen(ptr + 2) + 1);
} /* else if */
else if (*(ptr + 2) == '\0')
{
/* current dir at end of string: ditch it. */
*ptr = '\0';
} /* else if */
else if (*(ptr + 2) == '.')
{
if (*(ptr + 3) == '/')
{
/* parent dir in middle: move back one, if possible. */
memmove(prevptr, ptr + 4, strlen(ptr + 4) + 1);
ptr = prevptr;
while (prevptr != path)
{
prevptr--;
if (*prevptr == '/')
{
prevptr++;
break;
} /* if */
} /* while */
} /* if */
742
743
744
745
746
747
748
749
750
751
if (*(ptr + 3) == '\0')
{
/* parent dir at end: move back one, if possible. */
*prevptr = '\0';
} /* if */
} /* if */
} /* if */
else
{
prevptr = ptr;
753
754
755
} /* else */
} /* while */
} /* zip_expand_symlink_path */
757
/* (forward reference: zip_follow_symlink and zip_resolve call each other.) */
758
static int zip_resolve(PHYSFS_Io *io, ZIPinfo *info, ZIPentry *entry);
761
762
763
* Look for the entry named by (path). If it exists, resolve it, and return
* a pointer to that entry. If it's another symlink, keep resolving until you
* hit a real file and then return a pointer to the final non-symlink entry.
764
* If there's a problem, return NULL.
766
static ZIPentry *zip_follow_symlink(PHYSFS_Io *io, ZIPinfo *info, char *path)
768
769
770
ZIPentry *entry;
zip_expand_symlink_path(path);
771
entry = zip_find_entry(info, path);
772
773
if (entry != NULL)
{
774
if (!zip_resolve(io, info, entry)) /* recursive! */
775
776
777
778
779
780
781
782
entry = NULL;
else
{
if (entry->symlink != NULL)
entry = entry->symlink;
} /* else */
} /* if */
784
} /* zip_follow_symlink */
787
static int zip_resolve_symlink(PHYSFS_Io *io, ZIPinfo *info, ZIPentry *entry)
789
const PHYSFS_uint64 size = entry->uncompressed_size;
790
char *path = NULL;
791
792
int rc = 0;
793
794
795
796
797
798
/*
* We've already parsed the local file header of the symlink at this
* point. Now we need to read the actual link from the file data and
* follow it.
*/
799
BAIL_IF_ERRPASS(!io->seek(io, entry->offset), 0);
801
path = (char *) __PHYSFS_smallAlloc(size + 1);
802
BAIL_IF(!path, PHYSFS_ERR_OUT_OF_MEMORY, 0);
803
804
if (entry->compression_method == COMPMETH_NONE)
805
rc = __PHYSFS_readAll(io, path, size);
806
807
808
809
else /* symlink target path is compressed... */
{
z_stream stream;
810
const PHYSFS_uint64 complen = entry->compressed_size;
811
PHYSFS_uint8 *compressed = (PHYSFS_uint8*) __PHYSFS_smallAlloc(complen);
812
813
if (compressed != NULL)
{
814
if (__PHYSFS_readAll(io, compressed, complen))
816
initializeZStream(&stream);
817
stream.next_in = compressed;
818
stream.avail_in = complen;
819
stream.next_out = (unsigned char *) path;
820
821
822
stream.avail_out = size;
if (zlib_err(inflateInit2(&stream, -MAX_WBITS)) == Z_OK)
{
823
rc = zlib_err(inflate(&stream, Z_FINISH));
824
inflateEnd(&stream);
825
826
827
/* both are acceptable outcomes... */
rc = ((rc == Z_OK) || (rc == Z_STREAM_END));
828
829
} /* if */
} /* if */
830
__PHYSFS_smallFree(compressed);
831
832
} /* if */
} /* else */
834
if (rc)
835
836
837
{
path[entry->uncompressed_size] = '\0'; /* null-terminate it. */
zip_convert_dos_path(entry, path);
838
entry->symlink = zip_follow_symlink(io, info, path);
839
840
} /* else */
841
842
__PHYSFS_smallFree(path);
843
return (entry->symlink != NULL);
844
845
846
847
848
849
} /* zip_resolve_symlink */
/*
* Parse the local file header of an entry, and update entry->offset.
*/
850
static int zip_parse_local(PHYSFS_Io *io, ZIPentry *entry)
851
852
853
854
855
856
{
PHYSFS_uint32 ui32;
PHYSFS_uint16 ui16;
PHYSFS_uint16 fnamelen;
PHYSFS_uint16 extralen;
857
858
859
860
861
/*
* crc and (un)compressed_size are always zero if this is a "JAR"
* archive created with Sun's Java tools, apparently. We only
* consider this archive corrupted if those entries don't match and
* aren't zero. That seems to work well.
862
863
* We also ignore a mismatch if the value is 0xFFFFFFFF here, since it's
* possible that's a Zip64 thing.
866
867
868
869
/* !!! FIXME: apparently these are zero if general purpose bit 3 is set,
!!! FIXME: which is probably true for Jar files, fwiw, but we don't
!!! FIXME: care about these values anyhow. */
870
871
872
873
874
875
876
877
878
879
880
881
882
883
BAIL_IF_ERRPASS(!io->seek(io, entry->offset), 0);
BAIL_IF_ERRPASS(!readui32(io, &ui32), 0);
BAIL_IF(ui32 != ZIP_LOCAL_FILE_SIG, PHYSFS_ERR_CORRUPT, 0);
BAIL_IF_ERRPASS(!readui16(io, &ui16), 0);
BAIL_IF(ui16 != entry->version_needed, PHYSFS_ERR_CORRUPT, 0);
BAIL_IF_ERRPASS(!readui16(io, &ui16), 0); /* general bits. */
BAIL_IF_ERRPASS(!readui16(io, &ui16), 0);
BAIL_IF(ui16 != entry->compression_method, PHYSFS_ERR_CORRUPT, 0);
BAIL_IF_ERRPASS(!readui32(io, &ui32), 0); /* date/time */
BAIL_IF_ERRPASS(!readui32(io, &ui32), 0);
BAIL_IF(ui32 && (ui32 != entry->crc), PHYSFS_ERR_CORRUPT, 0);
BAIL_IF_ERRPASS(!readui32(io, &ui32), 0);
BAIL_IF(ui32 && (ui32 != 0xFFFFFFFF) &&
884
885
(ui32 != entry->compressed_size), PHYSFS_ERR_CORRUPT, 0);
886
887
BAIL_IF_ERRPASS(!readui32(io, &ui32), 0);
BAIL_IF(ui32 && (ui32 != 0xFFFFFFFF) &&
888
889
(ui32 != entry->uncompressed_size), PHYSFS_ERR_CORRUPT, 0);
890
891
BAIL_IF_ERRPASS(!readui16(io, &fnamelen), 0);
BAIL_IF_ERRPASS(!readui16(io, &extralen), 0);
892
893
entry->offset += fnamelen + extralen + 30;
895
896
897
} /* zip_parse_local */
898
static int zip_resolve(PHYSFS_Io *io, ZIPinfo *info, ZIPentry *entry)
899
900
{
int retval = 1;
901
902
903
904
const ZipResolveType resolve_type = entry->resolved;
if (resolve_type == ZIP_DIRECTORY)
return 1; /* we're good. */
905
906
/* Don't bother if we've failed to resolve this entry before. */
907
908
BAIL_IF(resolve_type == ZIP_BROKEN_FILE, PHYSFS_ERR_CORRUPT, 0);
BAIL_IF(resolve_type == ZIP_BROKEN_SYMLINK, PHYSFS_ERR_CORRUPT, 0);
909
910
/* uhoh...infinite symlink loop! */
911
BAIL_IF(resolve_type == ZIP_RESOLVING, PHYSFS_ERR_SYMLINK_LOOP, 0);
912
913
914
915
916
917
918
919
920
921
922
923
/*
* We fix up the offset to point to the actual data on the
* first open, since we don't want to seek across the whole file on
* archive open (can be SLOW on large, CD-stored files), but we
* need to check the local file header...not just for corruption,
* but since it stores offset info the central directory does not.
*/
if (resolve_type != ZIP_RESOLVED)
{
entry->resolved = ZIP_RESOLVING;
924
retval = zip_parse_local(io, entry);
925
926
927
928
929
930
931
932
if (retval)
{
/*
* If it's a symlink, find the original file. This will cause
* resolution of other entries (other symlinks and, eventually,
* the real file) if all goes well.
*/
if (resolve_type == ZIP_UNRESOLVED_SYMLINK)
933
retval = zip_resolve_symlink(io, info, entry);
934
935
936
937
938
939
} /* if */
if (resolve_type == ZIP_UNRESOLVED_SYMLINK)
entry->resolved = ((retval) ? ZIP_RESOLVED : ZIP_BROKEN_SYMLINK);
else if (resolve_type == ZIP_UNRESOLVED_FILE)
entry->resolved = ((retval) ? ZIP_RESOLVED : ZIP_BROKEN_FILE);
940
} /* if */
942
return retval;
943
} /* zip_resolve */
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
static int zip_hash_entry(ZIPinfo *info, ZIPentry *entry);
/* Fill in missing parent directories. */
static ZIPentry *zip_hash_ancestors(ZIPinfo *info, char *name)
{
ZIPentry *retval = &info->root;
char *sep = strrchr(name, '/');
if (sep)
{
const size_t namelen = (sep - name) + 1;
*sep = '\0'; /* chop off last piece. */
retval = zip_find_entry(info, name);
*sep = '/';
if (retval != NULL)
{
if (retval->resolved != ZIP_DIRECTORY)
965
BAIL(PHYSFS_ERR_CORRUPT, NULL);
966
967
968
969
970
return retval; /* already hashed. */
} /* if */
/* okay, this is a new dir. Build and hash us. */
retval = (ZIPentry *) allocator.Malloc(sizeof (ZIPentry) + namelen);
971
BAIL_IF(!retval, PHYSFS_ERR_OUT_OF_MEMORY, NULL);
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
memset(retval, '\0', sizeof (*retval));
retval->name = ((char *) retval) + sizeof (ZIPentry);
memcpy(retval->name, name, namelen);
retval->name[namelen] = '\0';
retval->resolved = ZIP_DIRECTORY;
if (!zip_hash_entry(info, retval))
{
allocator.Free(retval);
return NULL;
} /* if */
} /* else */
return retval;
} /* zip_hash_ancestors */
static int zip_hash_entry(ZIPinfo *info, ZIPentry *entry)
{
PHYSFS_uint32 hashval;
ZIPentry *parent;
assert(!zip_find_entry(info, entry->name)); /* checked elsewhere */
parent = zip_hash_ancestors(info, entry->name);
if (!parent)
return 0;
hashval = zip_hash_string(info, entry->name);
entry->hashnext = info->hash[hashval];