diff options
Diffstat (limited to 'lzip.h')
| -rw-r--r-- | lzip.h | 148 |
1 files changed, 118 insertions, 30 deletions
@@ -1,5 +1,5 @@ /* Lziprecover - Data recovery tool for the lzip format - Copyright (C) 2009-2018 Antonio Diaz Diaz. + Copyright (C) 2009-2019 Antonio Diaz Diaz. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -121,7 +121,7 @@ public: { const std::string & s = filenames[i]; const unsigned len = ( s == "-" ) ? stdin_name_len : s.size(); - if( len > longest_name ) longest_name = len; + if( longest_name < len ) longest_name = len; } if( longest_name == 0 ) longest_name = stdin_name_len; } @@ -140,7 +140,7 @@ public: if( filename.size() && filename != "-" ) name_ = filename; else name_ = stdin_name; padded_name = " "; padded_name += name_; padded_name += ": "; - if( name_.size() < longest_name ) + if( longest_name > name_.size() ) padded_name.append( longest_name - name_.size(), ' ' ); first_post = true; } @@ -198,30 +198,30 @@ inline int real_bits( unsigned value ) } -const uint8_t magic_string[4] = { 0x4C, 0x5A, 0x49, 0x50 }; // "LZIP" +const uint8_t lzip_magic[4] = { 0x4C, 0x5A, 0x49, 0x50 }; // "LZIP" -struct File_header +struct Lzip_header { uint8_t data[6]; // 0-3 magic bytes // 4 version // 5 coded_dict_size enum { size = 6 }; - void set_magic() { std::memcpy( data, magic_string, 4 ); data[4] = 1; } + void set_magic() { std::memcpy( data, lzip_magic, 4 ); data[4] = 1; } bool verify_magic() const - { return ( std::memcmp( data, magic_string, 4 ) == 0 ); } + { return ( std::memcmp( data, lzip_magic, 4 ) == 0 ); } bool verify_prefix( const int sz ) const // detect (truncated) header { for( int i = 0; i < sz && i < 4; ++i ) - if( data[i] != magic_string[i] ) return false; + if( data[i] != lzip_magic[i] ) return false; return ( sz > 0 ); } bool verify_corrupt() const // detect corrupt header { int matches = 0; for( int i = 0; i < 4; ++i ) - if( data[i] == magic_string[i] ) ++matches; + if( data[i] == lzip_magic[i] ) ++matches; return ( matches > 1 && matches < 4 ); } @@ -253,12 +253,11 @@ struct File_header }; -struct File_trailer +struct Lzip_trailer { uint8_t data[20]; // 0-3 CRC32 of the uncompressed data // 4-11 size of the uncompressed data // 12-19 member size including header and trailer - enum { size = 20 }; unsigned data_crc() const @@ -290,6 +289,20 @@ struct File_trailer void member_size( unsigned long long sz ) { for( int i = 12; i <= 19; ++i ) { data[i] = (uint8_t)sz; sz >>= 8; } } + + bool verify_consistency() const // check internal consistency + { + const unsigned crc = data_crc(); + const unsigned long long dsize = data_size(); + if( ( crc == 0 ) != ( dsize == 0 ) ) return false; + const unsigned long long msize = member_size(); + if( msize < min_member_size ) return false; + const unsigned long long mlimit = ( 9 * dsize + 7 ) / 8 + min_member_size; + if( mlimit > dsize && msize > mlimit ) return false; + const unsigned long long dlimit = 7090 * ( msize - 26 ) - 1; + if( dlimit > msize && dsize > dlimit ) return false; + return true; + } }; @@ -310,6 +323,72 @@ struct Bad_byte }; +#ifndef INT64_MAX +#define INT64_MAX 0x7FFFFFFFFFFFFFFFLL +#endif + +class Block + { + long long pos_, size_; // pos + size <= INT64_MAX + +public: + Block( const long long p, const long long s ) : pos_( p ), size_( s ) {} + + long long pos() const { return pos_; } + long long size() const { return size_; } + long long end() const { return pos_ + size_; } + + void pos( const long long p ) { pos_ = p; } + void size( const long long s ) { size_ = s; } + + bool operator==( const Block & b ) const + { return pos_ == b.pos_ && size_ == b.size_; } + bool operator!=( const Block & b ) const + { return pos_ != b.pos_ || size_ != b.size_; } + + bool operator<( const Block & b ) const { return pos_ < b.pos_; } + + bool includes( const long long pos ) const + { return ( pos_ <= pos && end() > pos ); } + bool overlaps( const Block & b ) const + { return ( pos_ < b.end() && b.pos_ < end() ); } + + void shift( Block & b ) { ++size_; ++b.pos_; --b.size_; } + Block split( const long long pos ); + }; + + +struct Member_list // members/gaps/tdata to be dumped/removed/stripped + { + bool damaged; + bool tdata; + bool in, rin; + std::vector< Block > range_vector, rrange_vector; + + Member_list() : damaged( false ), tdata( false ), in( true ), rin( true ) {} + void parse( const char * p ); + + bool range() const { return range_vector.size() || rrange_vector.size(); } + + // blocks is the sum of members + gaps, excluding trailing data + bool includes( const long i, const long blocks ) const + { + for( unsigned j = 0; j < range_vector.size(); ++j ) + { + if( range_vector[j].pos() > i ) break; + if( range_vector[j].end() > i ) return in; + } + if( i >= 0 && i < blocks ) + for( unsigned j = 0; j < rrange_vector.size(); ++j ) + { + if( rrange_vector[j].pos() > blocks - i - 1 ) break; + if( rrange_vector[j].end() > blocks - i - 1 ) return rin; + } + return !in || !rin; + } + }; + + struct Error { const char * const msg; @@ -334,14 +413,25 @@ int alone_to_lz( const int infd, const Pretty_print & pp ); long readblock( const int fd, uint8_t * const buf, const long size ); long writeblock( const int fd, const uint8_t * const buf, const long size ); -// defined in file_index.cc -int seek_read( const int fd, uint8_t * const buf, const int size, - const long long pos ); +// defined in dump_remove.cc +int dump_members( const std::vector< std::string > & filenames, + const std::string & default_output_filename, + const Member_list & member_list, const bool force, + bool ignore_errors, bool ignore_trailing, + const bool loose_trailing, const bool strip ); +int remove_members( const std::vector< std::string > & filenames, + const Member_list & member_list, bool ignore_errors, + bool ignore_trailing, const bool loose_trailing ); // defined in list.cc int list_files( const std::vector< std::string > & filenames, + const bool ignore_errors, const bool ignore_trailing, const bool loose_trailing ); +// defined in lzip_index.cc +int seek_read( const int fd, uint8_t * const buf, const int size, + const long long pos ); + // defined in main.cc extern std::string output_filename; // global vars for output file extern int outfd; @@ -357,6 +447,7 @@ bool open_outstream( const bool force, const bool from_stdin, const bool rw = false, const bool skipping = true ); bool file_exists( const std::string & filename ); void cleanup_and_fail( const int retval ); +void set_signal_handler(); int close_outstream( const struct stat * const in_statsp ); std::string insert_fixed( std::string name ); void show_error( const char * const msg, const int errcode = 0, @@ -364,9 +455,7 @@ void show_error( const char * const msg, const int errcode = 0, void show_file_error( const char * const filename, const char * const msg, const int errcode = 0 ); void internal_error( const char * const msg ); -void show_error2( const char * const msg1, const char * const name, - const char * const msg2 ); -void show_error4( const char * const msg1, const char * const name1, +void show_2file_error( const char * const msg1, const char * const name1, const char * const name2, const char * const msg2 ); class Range_decoder; void show_dprogress( const unsigned long long cfile_size = 0, @@ -377,32 +466,31 @@ void show_dprogress( const unsigned long long cfile_size = 0, // defined in merge.cc bool copy_file( const int infd, const int outfd, const long long max_size = -1 ); -bool test_member_from_file( const int infd, const unsigned long long msize, - long long * const failure_posp = 0 ); +int test_member_from_file( const int infd, const unsigned long long msize, + long long * const failure_posp = 0 ); int merge_files( const std::vector< std::string > & filenames, const std::string & default_output_filename, - const bool force ); + const bool force, const char terminator ); // defined in range_dec.cc bool safe_seek( const int fd, const long long pos ); +int range_decompress( const std::string & input_filename, + const std::string & default_output_filename, + Block range, const bool force, const bool ignore_errors, + const bool ignore_trailing, const bool loose_trailing, + const bool to_stdout ); // defined in repair.cc int repair_file( const std::string & input_filename, const std::string & default_output_filename, - const bool force ); + const bool force, const char terminator ); +int debug_delay( const std::string & input_filename, Block range, + const char terminator ); int debug_repair( const std::string & input_filename, - const Bad_byte & bad_byte ); + const Bad_byte & bad_byte, const char terminator ); int debug_decompress( const std::string & input_filename, const Bad_byte & bad_byte, const bool show_packets ); // defined in split.cc -bool verify_header( const File_header & header, const Pretty_print & pp ); int split_file( const std::string & input_filename, const std::string & default_output_filename, const bool force ); - -// defined in trailing_data.cc -int dump_tdata( const std::vector< std::string > & filenames, - const std::string & default_output_filename, const bool force, - const bool strip, const bool loose_trailing ); -int remove_tdata( const std::vector< std::string > & filenames, - const bool loose_trailing ); |