diff options
Diffstat (limited to 'merge.cc')
| -rw-r--r-- | merge.cc | 166 |
1 files changed, 89 insertions, 77 deletions
@@ -1,5 +1,5 @@ /* Lziprecover - Data recovery tool for the lzip format - Copyright (C) 2009-2015 Antonio Diaz Diaz. + Copyright (C) 2009-2016 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 @@ -17,6 +17,7 @@ #define _FILE_OFFSET_BITS 64 +#include <algorithm> #include <cerrno> #include <climits> #include <cstdio> @@ -36,6 +37,28 @@ namespace { +bool file_crc( uint32_t & crc, const int infd ) + { + const int buffer_size = 65536; + crc = 0xFFFFFFFFU; + uint8_t * const buffer = new uint8_t[buffer_size]; + bool error = false; + + while( true ) + { + const int rd = readblock( infd, buffer, buffer_size ); + if( rd != buffer_size && errno ) + { show_error( "Error reading input file", errno ); error = true; break; } + if( rd > 0 ) + crc32.update_buf( crc, buffer, rd ); + if( rd < buffer_size ) break; // EOF + } + delete[] buffer; + crc ^= 0xFFFFFFFFU; + return !error; + } + + // Add 'bv' to 'block_vector' splitting blocks as needed to keep all the // edges (pos and end of every block). // 'block_vector' contains the result. 'bv' is destroyed. @@ -171,37 +194,43 @@ long ipow( const unsigned base, const unsigned exponent ) int open_input_files( const std::vector< std::string > & filenames, std::vector< int > & infd_vector, - File_index & file_index, const int verbosity ) + File_index & file_index, struct stat * const in_statsp, + const int verbosity ) { const int files = filenames.size(); - bool identical = false; - for( int i = 1; i < files; ++i ) - if( filenames[0] == filenames[i] ) - { identical = true; break; } - if( !identical ) - for( int i = 0; i < files; ++i ) - { - struct stat in_stats; - ino_t st_ino0 = 0; - dev_t st_dev0 = 0; - infd_vector[i] = open_instream( filenames[i].c_str(), &in_stats, true, true ); - if( infd_vector[i] < 0 ) return 1; - if( i == 0 ) { st_ino0 = in_stats.st_ino; st_dev0 = in_stats.st_dev; } - else if( st_ino0 == in_stats.st_ino && st_dev0 == in_stats.st_dev ) - { identical = true; break; } - } - if( identical ) { show_error( "Two input files are the same." ); return 2; } + for( int i = 0; i + 1 < files; ++i ) + for( int j = i + 1; j < files; ++j ) + if( filenames[i] == filenames[j] ) + { show_error2( "Input file", filenames[i].c_str(), "given twice." ); + return 2; } + { + std::vector< uint32_t > crc_vector( files ); + for( int i = 0; i < files; ++i ) + { + struct stat in_stats; // not used + infd_vector[i] = open_instream( filenames[i].c_str(), + ( i == 0 ) ? in_statsp : &in_stats, true, true ); + if( infd_vector[i] < 0 ) return 1; + if( !file_crc( crc_vector[i], infd_vector[i] ) ) return 1; + for( int j = 0; j < i; ++j ) + if( crc_vector[i] == crc_vector[j] ) + { show_error4( "Input files", filenames[j].c_str(), + filenames[i].c_str(), "are identical." ); return 2; } + } + } long long isize = 0; + int good_fi = -1; for( int i = 0; i < files; ++i ) { long long tmp; const File_index fi( infd_vector[i] ); if( fi.retval() == 0 ) // file format is intact { - if( file_index.retval() != 0 ) file_index = fi; + if( good_fi < 0 ) { good_fi = i; file_index = fi; } else if( file_index != fi ) - { show_error( "Input files are different." ); return 2; } + { show_error4( "Input files", filenames[good_fi].c_str(), + filenames[i].c_str(), "are different." ); return 2; } tmp = file_index.file_size(); } else // file format is damaged @@ -213,15 +242,13 @@ int open_input_files( const std::vector< std::string > & filenames, return 1; } } - if( i == 0 ) - { - isize = tmp; - if( isize < min_member_size ) - { show_error2( "Input file", filenames[i].c_str(), "is too short." ); - return 2; } - } + if( tmp < min_member_size ) + { show_error2( "Input file", filenames[i].c_str(), "is too short." ); + return 2; } + if( i == 0 ) isize = tmp; else if( isize != tmp ) - { show_error( "Sizes of input files are different." ); return 2; } + { show_error4( "Sizes of input files", filenames[0].c_str(), + filenames[i].c_str(), "are different." ); return 2; } } if( file_index.retval() != 0 ) @@ -242,7 +269,7 @@ int open_input_files( const std::vector< std::string > & filenames, const long long mpos = file_index.mblock( j ).pos(); const long long msize = file_index.mblock( j ).size(); if( !safe_seek( infd, mpos ) ) return 1; - if( !try_decompress_member( infd, msize ) ) { error = true; break; } + if( !test_member_from_file( infd, msize ) ) { error = true; break; } } if( !error ) { @@ -261,8 +288,7 @@ bool try_merge_member( const long long mpos, const long long msize, const std::vector< Block > & block_vector, const std::vector< int > & color_vector, const std::vector< int > & infd_vector, - const std::string & output_filename, - const int outfd, const int verbosity ) + const int verbosity ) { const int blocks = block_vector.size(); const int files = infd_vector.size(); @@ -273,7 +299,7 @@ bool try_merge_member( const long long mpos, const long long msize, show_error( "Too many damaged blocks. Try merging fewer files." ); else show_error( "Too many damaged blocks. Merging is not possible." ); - cleanup_and_fail( output_filename, outfd, 2 ); + cleanup_and_fail( 2 ); } int bi = 0; // block index std::vector< int > file_idx( blocks, 0 ); // file to read each block from @@ -294,13 +320,13 @@ bool try_merge_member( const long long mpos, const long long msize, if( !safe_seek( infd, block_vector[bi].pos() ) || !safe_seek( outfd, block_vector[bi].pos() ) || !copy_file( infd, outfd, block_vector[bi].size() ) ) - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); ++bi; } if( !safe_seek( outfd, mpos ) ) - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); long long failure_pos = 0; - if( try_decompress_member( outfd, msize, &failure_pos ) ) return true; + if( test_member_from_file( outfd, msize, &failure_pos ) ) return true; while( bi > 0 && mpos + failure_pos < block_vector[bi-1].pos() ) --bi; while( --bi >= 0 ) { @@ -325,8 +351,7 @@ bool try_merge_member1( const long long mpos, const long long msize, const std::vector< Block > & block_vector, const std::vector< int > & color_vector, const std::vector< int > & infd_vector, - const std::string & output_filename, - const int outfd, const int verbosity ) + const int verbosity ) { if( block_vector.size() != 1 || block_vector[0].size() <= 1 ) return false; const long long pos = block_vector[0].pos(); @@ -344,7 +369,7 @@ bool try_merge_member1( const long long mpos, const long long msize, !safe_seek( infd_vector[i2], pos ) || !safe_seek( outfd, pos ) || !copy_file( infd_vector[i2], outfd, size ) ) - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); const int var = ( i1 * ( files - 1 ) ) + i2 - ( i2 > i1 ) + 1; for( long long i = 0; i < size; ++i ) { @@ -358,9 +383,9 @@ bool try_merge_member1( const long long mpos, const long long msize, readblock( infd, &byte, 1 ) != 1 || writeblock( outfd, &byte, 1 ) != 1 || !safe_seek( outfd, mpos ) ) - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); long long failure_pos = 0; - if( try_decompress_member( outfd, msize, &failure_pos ) ) return true; + if( test_member_from_file( outfd, msize, &failure_pos ) ) return true; if( mpos + failure_pos <= pos + i ) break; } } @@ -370,16 +395,6 @@ bool try_merge_member1( const long long mpos, const long long msize, } // end namespace -void cleanup_and_fail( const std::string & output_filename, - const int outfd, const int retval ) - { - if( outfd >= 0 ) close( outfd ); - if( std::remove( output_filename.c_str() ) != 0 && errno != ENOENT ) - show_error( "WARNING: deletion of output file (apparently) failed." ); - std::exit( retval ); - } - - // max_size < 0 means no size limit. bool copy_file( const int infd, const int outfd, const long long max_size ) { @@ -410,44 +425,45 @@ bool copy_file( const int infd, const int outfd, const long long max_size ) } -bool try_decompress_member( const int fd, const unsigned long long msize, - long long * failure_posp ) +bool test_member_from_file( const int infd, const unsigned long long msize, + long long * const failure_posp ) { - Range_decoder rdec( fd ); + Range_decoder rdec( infd ); File_header header; rdec.read_data( header.data, File_header::size ); - if( !rdec.finished() && // End Of File - header.verify_magic() && header.verify_version() && - header.dictionary_size() >= min_dictionary_size && - header.dictionary_size() <= max_dictionary_size ) + const unsigned dictionary_size = header.dictionary_size(); + if( !rdec.finished() && header.verify_magic() && + header.verify_version() && isvalid_ds( dictionary_size ) ) { - LZ_decoder decoder( header, rdec, -1 ); + LZ_decoder decoder( rdec, dictionary_size, -1 ); Pretty_print dummy( "", -1 ); if( decoder.decode_member( dummy ) == 0 && rdec.member_position() == msize ) return true; - if( failure_posp ) *failure_posp = rdec.member_position(); } + if( failure_posp ) *failure_posp = rdec.member_position(); return false; } int merge_files( const std::vector< std::string > & filenames, - const std::string & output_filename, const int verbosity, - const bool force ) + const std::string & default_output_filename, + const int verbosity, const bool force ) { const int files = filenames.size(); std::vector< int > infd_vector( files ); File_index file_index; + struct stat in_stats; const int retval = - open_input_files( filenames, infd_vector, file_index, verbosity ); + open_input_files( filenames, infd_vector, file_index, &in_stats, verbosity ); if( retval >= 0 ) return retval; if( !safe_seek( infd_vector[0], 0 ) ) return 1; - const int outfd = open_outstream_rw( output_filename, force ); - if( outfd < 0 ) return 1; + output_filename = default_output_filename.empty() ? + insert_fixed( filenames[0] ) : default_output_filename; + if( !open_outstream( force, false, true, false ) ) return 1; if( !copy_file( infd_vector[0], outfd ) ) // copy whole file - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); for( long j = 0; j < file_index.members(); ++j ) { @@ -458,14 +474,14 @@ int merge_files( const std::vector< std::string > & filenames, std::vector< int > color_vector( files, 0 ); if( !diff_member( mpos, msize, infd_vector, block_vector, color_vector ) || !safe_seek( outfd, mpos ) ) - cleanup_and_fail( output_filename, outfd, 1 ); + cleanup_and_fail( 1 ); if( block_vector.empty() ) { - if( file_index.members() > 1 && try_decompress_member( outfd, msize ) ) + if( file_index.members() > 1 && test_member_from_file( outfd, msize ) ) continue; show_error( "Input files are (partially) identical. Merging is not possible." ); - cleanup_and_fail( output_filename, outfd, 2 ); + cleanup_and_fail( 2 ); } if( verbosity >= 1 && file_index.members() > 1 ) @@ -479,12 +495,12 @@ int merge_files( const std::vector< std::string > & filenames, if( file_index.members() > 1 || block_vector.size() > 1 ) { done = try_merge_member( mpos, msize, block_vector, color_vector, - infd_vector, output_filename, outfd, verbosity ); + infd_vector, verbosity ); if( !done && verbosity >= 1 ) std::fputc( '\n', stdout ); } if( !done ) done = try_merge_member1( mpos, msize, block_vector, color_vector, - infd_vector, output_filename, outfd, verbosity ); + infd_vector, verbosity ); if( verbosity >= 1 ) std::fputc( '\n', stdout ); if( !done ) { @@ -493,15 +509,11 @@ int merge_files( const std::vector< std::string > & filenames, std::fprintf( stderr, "area %2d from position %6lld to %6lld\n", i + 1, block_vector[i].pos(), block_vector[i].end() - 1 ); show_error( "Some error areas overlap. Can't recover input file." ); - cleanup_and_fail( output_filename, outfd, 2 ); + cleanup_and_fail( 2 ); } } - if( close( outfd ) != 0 ) - { - show_error( "Error closing output file", errno ); - cleanup_and_fail( output_filename, -1, 1 ); - } + if( close_outstream( &in_stats ) != 0 ) return 1; if( verbosity >= 1 ) std::fputs( "Input files merged successfully.\n", stdout ); return 0; |