123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333 |
- #include "BellTar.h"
- #include <sys/stat.h> // for mkdir
- using namespace bell::BellTar;
- #include <algorithm> // for min
- #include <cassert> // for assert
- #include <cstdint> // for uint8_t
- #include <cstdio> // for sprintf, size_t, sscanf, EOF, NULL
- #include <cstdlib> // for rand
- #include <cstring> // for memset, strlen
- #include <ctime> // for time
- #include <fstream> // for ofstream
- #include <vector> // for vector
- #ifdef _WIN32
- #include <direct.h>
- #endif
- #ifdef ENABLE_LOGGING
- #define LOG printf
- #else
- #ifdef _WIN32
- #define LOG(fmt, ...) ((void)0)
- #else
- #define LOG(fmt, args...) ((void)0)
- #endif
- #endif
- const char FILL_CHAR = '\0';
- const int FILE_NAME_LENGTH = 100;
- // From http://en.wikipedia.org/wiki/Tar_(computing)#UStar_format
- typedef enum tar_file_type {
- tar_file_type_normal = '0',
- tar_file_type_hard_link = '1',
- tar_file_type_soft_link = '2',
- tar_file_type_directory = '5'
- } tar_file_type_t;
- struct tar_header {
- char name[FILE_NAME_LENGTH]; // file name
- char mode[8]; // file mode
- char uid[8]; // Owner's numeric user ID
- char gid[8]; // Group's numeric user ID
- char size[12]; // File size in bytes (octal base)
- char mtime[12]; // Last modification time in
- // numeric Unix time format (octal)
- char checksum[8]; // Checksum for header record
- char typeflag[1]; // file type, see tar_file_type_t
- char linkname[100]; // Name of linked file
- char magic[6]; // UStar indicator "ustar"
- char version[2]; // UStar version "00"
- char uname[32]; // Owner user name
- char gname[32]; // Owner group name
- char devmajor[8]; // Device major number
- char devminor[8]; // Device minor number
- char prefix[155]; // Filename prefix
- char pad[12]; // padding
- };
- void header_set_metadata(tar_header* header) {
- std::memset(header, 0, sizeof(tar_header));
- std::sprintf(header->magic, "ustar");
- std::sprintf(header->mtime, "%011lo", (unsigned long)std::time(NULL));
- std::sprintf(header->mode, "%07o", 0644);
- std::sprintf(header->uname, "unkown"); // ... a bit random
- std::sprintf(header->gname, "users");
- header->typeflag[0] = 0; // always just a normal file
- }
- /* From Wikipedia: The checksum is calculated by taking the sum of the
- * unsigned byte values of the header record with the eight checksum
- * bytes taken to be ascii spaces. */
- void header_set_checksum(tar_header* header) {
- unsigned int sum = 0;
- char* pointer = (char*)header;
- char* end = pointer + sizeof(tar_header);
- // Iterate over header struct until we are at checksum field.
- while (pointer < header->checksum) {
- sum += *pointer & 0xff;
- pointer++;
- }
- // ... then add eight 'ascii spaces' ...
- sum += ' ' * 8;
- pointer += 8;
- // ... and go until the end.
- while (pointer < end) {
- sum += *pointer & 0xff;
- pointer++;
- }
- std::sprintf(header->checksum, "%06o", sum);
- }
- void header_set_filetype(tar_header* header, tar_file_type_t file_type) {
- header->typeflag[0] = file_type;
- }
- tar_file_type_t header_get_filetype(tar_header* header) {
- return tar_file_type_t(header->typeflag[0]);
- }
- void header_set_filesize(tar_header* header, file_size_t file_size) {
- std::sprintf(header->size, "%011llo", file_size);
- }
- file_size_t header_get_filesize(tar_header* header) {
- file_size_t file_size;
- std::sscanf(header->size, "%011llo", &file_size);
- return file_size;
- }
- void header_set_filename(tar_header* header, const char* file_name) {
- size_t len = std::strlen(file_name);
- // len > 0 also ensures that the header does not start with \0
- if (len == 0 || len >= FILE_NAME_LENGTH) {
- LOG("Invalid file name for tar: %s\n", file_name);
- std::sprintf(header->name, "INVALID_%d", std::rand());
- } else {
- std::sprintf(header->name, "%s", file_name);
- }
- }
- std::string header_get_filename(tar_header* header) {
- return std::string(header->name);
- }
- ////////////////////////////////////////
- /* Every file in a tar file starts with the tar header */
- void _write_header(std::ostream& dst, const char* file_name,
- file_size_t file_size,
- tar_file_type_t file_type = tar_file_type_normal) {
- tar_header header;
- header_set_metadata(&header);
- header_set_filename(&header, file_name);
- header_set_filesize(&header, file_size);
- header_set_filetype(&header, file_type);
- header_set_checksum(&header);
- dst.write((const char*)&header, sizeof(tar_header));
- }
- void _read_header(std::istream& inp, tar_header* header) {
- inp.read((char*)header, sizeof(tar_header));
- }
- /* The length of the data after the header must be rounded up to a
- multiple of 512 bytes, the length of the header. */
- void _fill(std::ostream& dst, unsigned long file_size) {
- while (file_size % sizeof(tar_header) != 0) {
- dst.put(FILL_CHAR);
- file_size++;
- }
- }
- bool _check_if_header_is_next(std::istream& inp) {
- if (inp.eof() || inp.peek() == EOF) {
- LOG("Can not read next file info, istream at EOF.\n");
- return false;
- }
- if (inp.peek() == FILL_CHAR) {
- LOG("Can not read next file info, istream is pointing "
- "to %d, which a tar header can not start with.\n",
- FILL_CHAR);
- return false;
- }
- return true;
- }
- void _seek_to_next_header(std::istream& inp) {
- // Advance to start of next header or to end of file
- // Works because
- // - header never starts with FILL_CHAR
- // - at end of file, peek() returns EOF.
- // - FILL_CHAR != EOF
- while (inp.peek() == FILL_CHAR)
- inp.get();
- }
- ////////////////////////////////////////
- // writer Implementation
- ////////////////////////////////////////
- void writer::put(std::string path_in_tar, char const* const data,
- const file_size_t data_size) {
- _write_header(_dst, path_in_tar.c_str(), data_size);
- _dst.write(data, data_size);
- _fill(_dst, data_size);
- }
- void writer::put_directory(std::string path_in_tar) {
- _write_header(_dst, path_in_tar.c_str(), 0, tar_file_type_directory);
- }
- /* The end of an tar is marked by at least two consecutive zero-filled
- * records, a record having the size of the header. */
- void writer::finish() {
- unsigned long i = 0;
- while (i < 2 * sizeof(tar_header)) {
- _dst.put(FILL_CHAR);
- i++;
- }
- }
- ////////////////////////////////////////
- // reader Implementation
- ////////////////////////////////////////
- bool reader::contains_another_file() {
- return _check_if_header_is_next(_inp);
- }
- void reader::_cache_header() {
- if (_cached_header_data_valid)
- return;
- assert(contains_another_file());
- tar_header h;
- _read_header(_inp, &h);
- _cached_header_data.file_name = header_get_filename(&h);
- _cached_header_data.file_size = header_get_filesize(&h);
- _cached_header_data.file_type = h.typeflag[0];
- _cached_header_data_valid = true;
- }
- std::string reader::get_next_file_name() {
- _cache_header();
- return _cached_header_data.file_name;
- }
- file_size_t reader::get_next_file_size() {
- _cache_header();
- return _cached_header_data.file_size;
- }
- void reader::read_next_file(char* const data) {
- _inp.read(data, get_next_file_size());
- _cached_header_data_valid = false;
- _seek_to_next_header(_inp);
- }
- void reader::skip_next_file() {
- _inp.seekg(get_next_file_size(), std::ios::cur);
- _cached_header_data_valid = false;
- _seek_to_next_header(_inp);
- }
- char reader::get_next_file_type() {
- _cache_header();
- return _cached_header_data.file_type;
- }
- int reader::number_of_files() {
- if (_number_of_files == -1) {
- std::streampos current_position = _inp.tellg();
- _inp.seekg(0, std::ios::beg);
- _number_of_files = 0;
- while (contains_another_file()) {
- _number_of_files++;
- skip_next_file();
- }
- _inp.seekg(current_position);
- }
- return _number_of_files;
- }
- void reader::extract_all_files(std::string dest_directory) {
- std::vector<uint8_t> scratch_buffer(1024);
- while (contains_another_file()) {
- char fileType = get_next_file_type();
- auto fileName = get_next_file_name();
- // 0 is the normal file type, skip apple's ._ files
- #if __cplusplus >= 202002L
- if (fileType == '0' && !fileName.starts_with("._")) {
- #else
- if (fileType == '0' && fileName.find("._") != 0) {
- #endif
- std::string path = dest_directory + "/" + fileName;
- size_t pos = 0;
- while ((pos = path.find('/', pos)) != std::string::npos) {
- std::string dir = path.substr(0, pos);
- // Create the directory if it doesn't exist
- #ifdef _WIN32
- mkdir(dir.c_str());
- #else
- mkdir(dir.c_str(), 0777);
- #endif
- pos++;
- }
- std::ofstream out(path, std::ios::binary);
- size_t read_size = 0;
- size_t file_size = get_next_file_size();
- while (read_size < file_size) {
- size_t to_read = std::min(file_size - read_size, scratch_buffer.size());
- _inp.read((char*)scratch_buffer.data(), to_read);
- // Move the read size forward
- read_size += _inp.gcount();
- // Write the data to the destination file
- out.write((char*)scratch_buffer.data(), _inp.gcount());
- }
- _cached_header_data_valid = false;
- _seek_to_next_header(_inp);
- } else {
- skip_next_file();
- }
- }
- }
|