off64_t zip_find_end_of_central_directory(ar_stream *stream)
{
uint8_t data[512];
off64_t filesize;
int fromend = 0;
int count, i;
if (!ar_seek(stream, 0, SEEK_END))
return -1;
filesize = ar_tell(stream);
while (fromend < UINT16_MAX + ZIP_END_OF_CENTRAL_DIR_SIZE && fromend < filesize) {
count = (int)(filesize - fromend < sizeof(data) ? filesize - fromend : sizeof(data));
fromend += count;
if (count < ZIP_END_OF_CENTRAL_DIR_SIZE)
return -1;
if (!ar_seek(stream, -fromend, SEEK_END))
return -1;
if (ar_read(stream, data, count) != (size_t)count)
return -1;
for (i = count - ZIP_END_OF_CENTRAL_DIR_SIZE; i >= 0; i--) {
if (uint32le(data + i) == SIG_END_OF_CENTRAL_DIRECTORY)
return filesize - fromend + i;
}
fromend -= ZIP_END_OF_CENTRAL_DIR_SIZE - 1;
}
return -1;
}
const char *zip_get_name(ar_archive *ar)
{
ar_archive_zip *zip = (ar_archive_zip *)ar;
if (!zip->entry.name) {
struct zip_entry entry;
char *name;
if (zip->dir.end_offset >= 0) {
if (!ar_seek(ar->stream, ar->entry_offset, SEEK_SET))
return NULL;
if (!zip_parse_directory_entry(zip, &entry))
return NULL;
if (!ar_seek(ar->stream, ar->entry_offset + ZIP_DIR_ENTRY_FIXED_SIZE, SEEK_SET))
return NULL;
}
else {
if (!ar_seek(ar->stream, zip->entry.offset, SEEK_SET))
return NULL;
if (!zip_parse_local_file_entry(zip, &entry))
return NULL;
if (!ar_seek(ar->stream, ar->entry_offset + ZIP_LOCAL_ENTRY_FIXED_SIZE, SEEK_SET))
return NULL;
}
name = malloc(entry.namelen + 1);
if (!name || ar_read(ar->stream, name, entry.namelen) != entry.namelen) {
free(name);
return NULL;
}
name[entry.namelen] = '\0';
if ((entry.flags & (1 << 11))) {
zip->entry.name = name;
}
else {
zip->entry.name = ar_conv_dos_to_utf8(name);
free(name);
}
/* normalize path separators */
if (zip->entry.name) {
char *p = zip->entry.name;
while ((p = strchr(p, '\\')) != NULL) {
*p = '/';
}
}
}
return zip->entry.name;
}
ar_archive *ar_open_7z_archive(ar_stream *stream)
{
ar_archive *ar;
ar_archive_7z *_7z;
SRes res;
if (!ar_seek(stream, 0, SEEK_SET))
return NULL;
ar = ar_open_archive(stream, sizeof(ar_archive_7z), _7z_close, _7z_parse_entry, _7z_get_name, _7z_uncompress, NULL, 0);
if (!ar)
return NULL;
_7z = (ar_archive_7z *)ar;
CSeekStream_CreateVTable(&_7z->in_stream, stream);
LookToRead_CreateVTable(&_7z->look_stream, False);
_7z->look_stream.realStream = &_7z->in_stream.super;
LookToRead_Init(&_7z->look_stream);
#ifdef USE_7Z_CRC32
CrcGenerateTable();
#endif
SzArEx_Init(&_7z->data);
res = SzArEx_Open(&_7z->data, &_7z->look_stream.s, &gSzAlloc, &gSzAlloc);
if (res != SZ_OK) {
if (res != SZ_ERROR_NO_ARCHIVE)
warn("Invalid 7z archive (failed with error %d)", res);
free(ar);
return NULL;
}
return ar;
}
bool ar_parse_entry_at(ar_archive *ar, size_t offset)
{
if (!ar_seek(ar->stream, offset, SEEK_SET))
return false;
ar->entry_offset = 0;
ar->entry_size_block = offset;
return ar->parse_entry(ar);
}
static SRes CSeekStream_Seek(void *p, Int64 *pos, ESzSeek origin)
{
struct CSeekStream *stm = p;
if (!ar_seek(stm->stream, *pos, (int)origin))
return SZ_ERROR_FAIL;
*pos = ar_tell(stm->stream);
return SZ_OK;
}
bool zip_seek_to_compressed_data(ar_archive_zip *zip)
{
struct zip_entry entry;
if (!ar_seek(zip->super.stream, zip->entry.offset, SEEK_SET))
return false;
if (!zip_parse_local_file_entry(zip, &entry))
return false;
if (zip->entry.method != entry.method) {
warn("Compression methods don't match: %d != %d", zip->entry.method, entry.method);
if (!zip->entry.method)
zip->entry.method = entry.method;
}
if (zip->entry.dosdate != entry.dosdate) {
warn("Timestamps don't match");
if (!zip->entry.dosdate) {
zip->entry.dosdate = entry.dosdate;
zip->super.entry_filetime = ar_conv_dosdate_to_filetime(zip->entry.dosdate);
}
}
return ar_seek(zip->super.stream, zip->entry.offset + ZIP_LOCAL_ENTRY_FIXED_SIZE + entry.namelen + entry.extralen, SEEK_SET);
}
off64_t zip_find_end_of_last_directory_entry(ar_stream *stream, struct zip_eocd64 *eocd)
{
uint8_t data[ZIP_DIR_ENTRY_FIXED_SIZE];
uint64_t i;
if (!ar_seek(stream, eocd->dir_offset, SEEK_SET))
return -1;
for (i = 0; i < eocd->numentries; i++) {
if (ar_read(stream, data, sizeof(data)) != sizeof(data))
return -1;
if (uint32le(data + 0) != SIG_CENTRAL_DIRECTORY)
return -1;
if (!ar_skip(stream, uint16le(data + 28) + uint16le(data + 30) + uint16le(data + 32)))
return -1;
}
return ar_tell(stream);
}
ar_archive *ar_open_rar_archive(ar_stream *stream)
{
char signature[FILE_SIGNATURE_SIZE];
if (!ar_seek(stream, 0, SEEK_SET))
return NULL;
if (ar_read(stream, signature, sizeof(signature)) != sizeof(signature))
return NULL;
if (memcmp(signature, "Rar!\x1A\x07\x00", sizeof(signature)) != 0) {
if (memcmp(signature, "Rar!\x1A\x07\x01", sizeof(signature)) == 0)
warn("RAR 5 format isn't supported");
else if (memcmp(signature, "RE~^", 4) == 0)
warn("Ancient RAR format isn't supported");
else if (memcmp(signature, "MZ", 2) == 0 || memcmp(signature, "\x7F\x45LF", 4) == 0)
warn("SFX archives aren't supported");
return NULL;
}
return ar_open_archive(stream, sizeof(ar_archive_rar), rar_close, rar_parse_entry, rar_get_name, rar_uncompress, NULL, FILE_SIGNATURE_SIZE);
}
ar_archive *ar_open_tar_archive(ar_stream *stream)
{
ar_archive *ar;
ar_archive_tar *tar;
if (!ar_seek(stream, 0, SEEK_SET))
return NULL;
ar = ar_open_archive(stream, sizeof(ar_archive_tar), tar_close, tar_parse_entry, tar_get_name, tar_uncompress, NULL, 0);
if (!ar)
return NULL;
tar = (ar_archive_tar *)ar;
if (!tar_parse_header(tar) || !tar->entry.checksum) {
free(ar);
return NULL;
}
return ar;
}
off64_t zip_find_next_local_file_entry(ar_stream *stream, off64_t offset)
{
uint8_t data[512];
int count, i;
if (!ar_seek(stream, offset, SEEK_SET))
return false;
count = (int)ar_read(stream, data, sizeof(data));
while (count >= ZIP_LOCAL_ENTRY_FIXED_SIZE) {
for (i = 0; i < count - 4; i++) {
if (uint32le(data + i) == SIG_LOCAL_FILE_HEADER)
return offset + i;
}
memmove(data, data + count - 4, count);
offset += count - 4;
count = (int)ar_read(stream, data + 4, sizeof(data) - 4) + 4;
}
return -1;
}
bool zip_parse_end_of_central_directory(ar_stream *stream, struct zip_eocd64 *eocd)
{
uint8_t data[56];
if (ar_read(stream, data, ZIP_END_OF_CENTRAL_DIR_SIZE) != ZIP_END_OF_CENTRAL_DIR_SIZE)
return false;
eocd->signature = uint32le(data + 0);
eocd->diskno = uint16le(data + 4);
eocd->diskno_dir = uint16le(data + 6);
eocd->numentries_disk = uint16le(data + 8);
eocd->numentries = uint16le(data + 10);
eocd->dir_size = uint32le(data + 12);
eocd->dir_offset = uint32le(data + 16);
eocd->commentlen = uint16le(data + 20);
if (eocd->signature != SIG_END_OF_CENTRAL_DIRECTORY)
return false;
/* try to locate the ZIP64 end of central directory */
if (!ar_skip(stream, -42))
return eocd->dir_size < 20;
if (ar_read(stream, data, 20) != 20)
return false;
if (uint32le(data + 0) != SIG_END_OF_CENTRAL_DIRECTORY_64_LOCATOR)
return true;
if ((eocd->diskno != UINT16_MAX && uint32le(data + 4) != eocd->diskno) || uint32le(data + 16) != 1) {
warn("Archive spanning isn't supported");
return false;
}
if (!ar_seek(stream, (off64_t)uint64le(data + 8), SEEK_SET))
return false;
if (ar_read(stream, data, 56) != 56)
return false;
/* use data from ZIP64 end of central directory (when necessary) */
eocd->signature = uint32le(data + 0);
eocd->version = uint16le(data + 12);
eocd->min_version = uint16le(data + 14);
if (eocd->diskno == UINT16_MAX)
eocd->diskno = uint32le(data + 16);
if (eocd->diskno_dir == UINT16_MAX)
eocd->diskno_dir = uint32le(data + 20);
if (eocd->numentries_disk == UINT16_MAX)
eocd->numentries_disk = uint64le(data + 24);
if (eocd->numentries == UINT16_MAX)
eocd->numentries = uint64le(data + 32);
if (eocd->dir_size == UINT32_MAX)
eocd->dir_size = uint64le(data + 40);
if (eocd->dir_offset == UINT32_MAX)
eocd->dir_offset = (off64_t)uint64le(data + 48);
if (eocd->signature != SIG_END_OF_CENTRAL_DIRECTORY_64)
return false;
if (eocd->diskno != eocd->diskno_dir || eocd->numentries != eocd->numentries_disk) {
warn("Archive spanning isn't supported");
return false;
}
if (uint64le(data + 4) > 44)
log("ZIP64 extensible data sector present @" PRIi64, ar_tell(stream));
return true;
}
static bool tar_parse_entry(ar_archive *ar, off64_t offset)
{
ar_archive_tar *tar = (ar_archive_tar *)ar;
char *longname;
if (!ar_seek(ar->stream, offset, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIi64, offset);
return false;
}
if (!tar_parse_header(tar)) {
warn("Invalid tar header data @%" PRIi64, offset);
return false;
}
if (!tar->entry.checksum) {
ar->at_eof = true;
return false;
}
ar->entry_offset = offset;
ar->entry_offset_next = offset + TAR_BLOCK_SIZE + (tar->entry.filesize + TAR_BLOCK_SIZE - 1) / TAR_BLOCK_SIZE * TAR_BLOCK_SIZE;
ar->entry_size_uncompressed = tar->entry.filesize;
ar->entry_filetime = tar->entry.mtime;
tar->bytes_done = 0;
switch (tar->entry.filetype) {
case TYPE_FILE:
case TYPE_FILE_OLD:
return true;
case TYPE_DIRECTORY:
log("Skipping directory entry \"%s\"", tar_get_name(ar));
return tar_parse_entry(ar, ar->entry_offset_next);
case TYPE_LONGNAME:
longname = malloc(tar->entry.filesize + 1);
if (!longname) {
log("Falling back to the short filename on OOM");
return tar_parse_entry(ar, ar->entry_offset_next);
}
if (!ar_entry_uncompress(ar, longname, tar->entry.filesize)) {
free(longname);
return false;
}
longname[tar->entry.filesize] = '\0';
if (!tar_parse_entry(ar, ar->entry_offset_next)) {
free(longname);
return false;
}
ar->entry_offset = offset;
/* name could be in any encoding, assume UTF-8 or whatever (DOS) */
if (ar_is_valid_utf8(longname)) {
tar->entry.name = longname;
}
else {
tar->entry.name = ar_conv_dos_to_utf8(longname);
free(longname);
}
return true;
default:
warn("Unknown entry type '%c'", tar->entry.filetype);
return true;
}
}
static bool rar_parse_entry(ar_archive *ar)
{
ar_archive_rar *rar = (ar_archive_rar *)ar;
struct rar_header header;
struct rar_entry entry;
/* without solid data, most/all previous files have to be decompressed again */
bool has_solid_data = rar->super.entry_offset != 0 && rar->uncomp.initialized && rar->progr.data_left == 0;
if (rar->super.entry_offset != 0) {
if (!ar_seek(rar->super.stream, rar->super.entry_offset + rar->super.entry_size_block, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIuPTR, rar->super.entry_offset + rar->super.entry_size_block);
return false;
}
}
for (;;) {
rar->super.entry_offset = ar_tell(rar->super.stream);
rar->super.entry_size_block = 0;
rar->super.entry_size_uncompressed = 0;
if (!rar_parse_header(&rar->super, &header))
return false;
switch (header.type) {
case TYPE_MAIN_HEADER:
if ((header.flags & MHD_PASSWORD)) {
warn("Encrypted archives aren't supported");
return false;
}
ar_skip(rar->super.stream, 6 /* reserved data */);
if ((header.flags & MHD_ENCRYPTVER)) {
log("MHD_ENCRYPTVER is set");
ar_skip(rar->super.stream, 1);
}
if ((header.flags & MHD_COMMENT))
log("MHD_COMMENT is set");
if (ar_tell(rar->super.stream) - rar->super.entry_offset > header.size) {
warn("Invalid RAR header size: %" PRIuPTR, header.size);
return false;
}
rar->archive_flags = header.flags;
break;
case TYPE_FILE_ENTRY:
if (!rar_parse_header_entry(rar, &header, &entry))
return false;
if ((header.flags & LHD_PASSWORD))
warn("Encrypted entries will fail to uncompress");
if ((header.flags & LHD_DIRECTORY) == LHD_DIRECTORY) {
log("Skipping directory entry \"%s\"", rar_get_name(&rar->super));
break;
}
if ((header.flags & (LHD_SPLIT_BEFORE | LHD_SPLIT_AFTER)))
warn("Splitting files isn't really supported");
// TODO: handle multi-part files (only needed for split files)?
rar->super.entry_size_block = header.size + (size_t)header.datasize;
rar->super.entry_size_uncompressed = (size_t)entry.size;
if (rar->super.entry_size_block < rar->entry.header_size) {
warn("Integer overflow due to overly large data size");
return false;
}
if (!has_solid_data || !rar->entry.restart_solid || rar->entry.method == METHOD_STORE)
rar_clear_uncompress(&rar->uncomp);
else
rar->entry.restart_solid = false;
#ifdef DEBUG
// TODO: CRC checks don't always hold (claim in XADRARParser.m @readBlockHeader)
if (!rar_check_header_crc(&rar->super))
warn("Invalid header checksum @%" PRIuPTR, rar->super.entry_offset);
#endif
if (!ar_seek(rar->super.stream, rar->super.entry_offset + rar->entry.header_size, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIuPTR, rar->super.entry_offset + rar->entry.header_size);
return false;
}
return true;
case TYPE_NEWSUB:
log("Skipping newsub header @%" PRIuPTR, rar->super.entry_offset);
break;
case TYPE_END_OF_ARCHIVE:
rar->super.at_eof = true;
return false;
default:
log("Unknown RAR header type %02x", header.type);
break;
}
#ifdef DEBUG
// TODO: CRC checks don't always hold (claim in XADRARParser.m @readBlockHeader)
if (!rar_check_header_crc(&rar->super))
warn("Invalid header checksum @%" PRIuPTR, rar->super.entry_offset);
#endif
if (!ar_seek(rar->super.stream, rar->super.entry_offset + header.size + (ptrdiff_t)header.datasize, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIu64, rar->super.entry_offset + header.size + header.datasize);
return false;
}
if (ar_tell(rar->super.stream) <= rar->super.entry_offset) {
warn("Integer overflow due to overly large data size");
return false;
}
}
}
static bool rar_parse_entry(ar_archive *ar, off64_t offset)
{
ar_archive_rar *rar = (ar_archive_rar *)ar;
struct rar_header header;
struct rar_entry entry;
bool out_of_order = offset != ar->entry_offset_next;
if (!ar_seek(ar->stream, offset, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIi64, offset);
return false;
}
for (;;) {
ar->entry_offset = ar_tell(ar->stream);
ar->entry_size_uncompressed = 0;
if (!rar_parse_header(ar, &header))
return false;
ar->entry_offset_next = ar->entry_offset + header.size + header.datasize;
if (ar->entry_offset_next < ar->entry_offset + header.size) {
warn("Integer overflow due to overly large data size");
return false;
}
switch (header.type) {
case TYPE_MAIN_HEADER:
if ((header.flags & MHD_PASSWORD)) {
warn("Encrypted archives aren't supported");
return false;
}
ar_skip(ar->stream, 6 /* reserved data */);
if ((header.flags & MHD_ENCRYPTVER)) {
log("MHD_ENCRYPTVER is set");
ar_skip(ar->stream, 1);
}
if ((header.flags & MHD_COMMENT))
log("MHD_COMMENT is set");
if (ar_tell(ar->stream) - ar->entry_offset > header.size) {
warn("Invalid RAR header size: %d", header.size);
return false;
}
rar->archive_flags = header.flags;
break;
case TYPE_FILE_ENTRY:
if (!rar_parse_header_entry(rar, &header, &entry))
return false;
if ((header.flags & LHD_PASSWORD))
warn("Encrypted entries will fail to uncompress");
if ((header.flags & LHD_DIRECTORY) == LHD_DIRECTORY) {
if (header.datasize == 0) {
log("Skipping directory entry \"%s\"", rar_get_name(ar));
break;
}
warn("Can't skip directory entries containing data");
}
if ((header.flags & (LHD_SPLIT_BEFORE | LHD_SPLIT_AFTER)))
warn("Splitting files isn't really supported");
ar->entry_size_uncompressed = (size_t)entry.size;
ar->entry_filetime = ar_conv_dosdate_to_filetime(entry.dosdate);
if (!rar->entry.solid || rar->entry.method == METHOD_STORE || out_of_order) {
rar_clear_uncompress(&rar->uncomp);
memset(&rar->solid, 0, sizeof(rar->solid));
}
else {
br_clear_leftover_bits(&rar->uncomp);
}
rar->solid.restart = rar->entry.solid && (out_of_order || !rar->solid.part_done);
rar->solid.part_done = !ar->entry_size_uncompressed;
rar->progress.data_left = (size_t)header.datasize;
rar->progress.bytes_done = 0;
rar->progress.crc = 0;
/* TODO: CRC checks don't always hold (claim in XADRARParser.m @readBlockHeader) */
if (!rar_check_header_crc(ar))
warn("Invalid header checksum @%" PRIi64, ar->entry_offset);
if (ar_tell(ar->stream) != ar->entry_offset + rar->entry.header_size) {
warn("Couldn't seek to offset %" PRIi64, ar->entry_offset + rar->entry.header_size);
return false;
}
return true;
case TYPE_NEWSUB:
log("Skipping newsub header @%" PRIi64, ar->entry_offset);
break;
case TYPE_END_OF_ARCHIVE:
ar->at_eof = true;
return false;
default:
log("Unknown RAR header type %02x", header.type);
break;
}
/* TODO: CRC checks don't always hold (claim in XADRARParser.m @readBlockHeader) */
if (!rar_check_header_crc(ar))
warn("Invalid header checksum @%" PRIi64, ar->entry_offset);
if (!ar_seek(ar->stream, ar->entry_offset_next, SEEK_SET)) {
warn("Couldn't seek to offset %" PRIi64, ar->entry_offset_next);
return false;
}
}
}
