diff options
Diffstat (limited to '')
| -rw-r--r-- | decoder.h | 98 |
1 files changed, 40 insertions, 58 deletions
@@ -15,74 +15,57 @@ along with this program. If not, see <http://www.gnu.org/licenses/>. */ -class Input_buffer +class Range_decoder { - enum { buffer_size = 65536 }; - uint8_t * const buffer; + enum { buffer_size = 16384 }; + long long partial_member_pos; + uint8_t * const buffer; // input buffer int pos; int stream_pos; // when reached, a new block must be read - const int ides_; // input file descriptor + uint32_t code; + uint32_t range; + const int infd_; // input file descriptor bool at_stream_end; bool read_block(); public: - Input_buffer( const int ides ) + Range_decoder( const int infd ) : + partial_member_pos( 0 ), buffer( new uint8_t[buffer_size] ), pos( 0 ), stream_pos( 0 ), - ides_( ides ), + code( 0 ), + range( 0xFFFFFFFFU ), + infd_( infd ), at_stream_end( false ) {} - ~Input_buffer() { delete[] buffer; } + ~Range_decoder() { delete[] buffer; } + bool code_is_zero() const throw() { return ( code == 0 ); } bool finished() const throw() { return at_stream_end && pos >= stream_pos; } + long long member_position() const throw() + { return partial_member_pos + pos; } + void reset_member_position() throw() + { partial_member_pos = -pos; } uint8_t get_byte() { if( pos >= stream_pos && !read_block() ) return 0; return buffer[pos++]; } - }; - - -class Range_decoder - { - mutable long long member_pos; - uint32_t code; - uint32_t range; - Input_buffer & ibuf; - -public: - Range_decoder( const int header_size, Input_buffer & buf ) - : - member_pos( header_size ), - code( 0 ), - range( 0xFFFFFFFF ), - ibuf( buf ) - { for( int i = 0; i < 5; ++i ) code = (code << 8) | get_byte(); } - - bool code_is_zero() const throw() { return ( code == 0 ); } - bool finished() const throw() { return ibuf.finished(); } - long long member_position() const throw() { return member_pos; } - - uint8_t get_byte() const - { - ++member_pos; - return ibuf.get_byte(); - } - void reload() throw() + void load() throw() { code = 0; - range = 0xFFFFFFFF; + range = 0xFFFFFFFFU; for( int i = 0; i < 5; ++i ) code = (code << 8) | get_byte(); } void normalize() { - if( range <= 0x00FFFFFF ) + if( range <= 0x00FFFFFFU ) { range <<= 8; code = (code << 8) | get_byte(); } } @@ -92,7 +75,7 @@ public: for( int i = num_bits; i > 0; --i ) { symbol <<= 1; - if( range <= 0x00FFFFFF ) + if( range <= 0x00FFFFFFU ) { range <<= 7; code = (code << 8) | get_byte(); if( code >= range ) { code -= range; symbol |= 1; } @@ -148,16 +131,16 @@ public: int decode_matched( Bit_model bm[], const int match_byte ) { - Bit_model *bm1 = bm + 0x100; + Bit_model * const bm1 = bm + 0x100; int symbol = 1; - for( int i = 1; i <= 8; ++i ) + for( int i = 7; i >= 0; --i ) { - const int match_bit = ( match_byte << i ) & 0x100; - const int bit = decode_bit( bm1[match_bit+symbol] ); + const int match_bit = ( match_byte >> i ) & 1; + const int bit = decode_bit( bm1[(match_bit<<8)+symbol] ); symbol = ( symbol << 1 ) | bit; - if( ( match_bit && !bit ) || ( !match_bit && bit ) ) + if( match_bit != bit ) { - while( ++i <= 8 ) + while( --i >= 0 ) symbol = ( symbol << 1 ) | decode_bit( bm[symbol] ); break; } @@ -193,30 +176,30 @@ class Literal_decoder { Bit_model bm_literal[1<<literal_context_bits][0x300]; - int state( const int prev_byte ) const throw() + int lstate( const int prev_byte ) const throw() { return ( prev_byte >> ( 8 - literal_context_bits ) ); } public: uint8_t decode( Range_decoder & range_decoder, const uint8_t prev_byte ) - { return range_decoder.decode_tree( bm_literal[state(prev_byte)], 8 ); } + { return range_decoder.decode_tree( bm_literal[lstate(prev_byte)], 8 ); } uint8_t decode_matched( Range_decoder & range_decoder, const uint8_t prev_byte, const uint8_t match_byte ) - { return range_decoder.decode_matched( bm_literal[state(prev_byte)], match_byte ); } + { return range_decoder.decode_matched( bm_literal[lstate(prev_byte)], match_byte ); } }; class LZ_decoder { long long partial_data_pos; - const int format_version; + const int member_version; const int dictionary_size; const int buffer_size; uint8_t * const buffer; int pos; int stream_pos; // first byte not yet written to file uint32_t crc_; - const int odes_; // output file descriptor + const int outfd_; // output file descriptor Bit_model bm_match[State::states][pos_states]; Bit_model bm_rep[State::states]; @@ -228,7 +211,7 @@ class LZ_decoder Bit_model bm_dis[modeled_distances-end_dis_model]; Bit_model bm_align[dis_align_size]; - Range_decoder range_decoder; + Range_decoder & range_decoder; Len_decoder len_decoder; Len_decoder rep_match_len_decoder; Literal_decoder literal_decoder; @@ -267,24 +250,23 @@ class LZ_decoder bool verify_trailer( const Pretty_print & pp ) const; public: - LZ_decoder( const File_header & header, Input_buffer & ibuf, const int odes ) + LZ_decoder( const File_header & header, Range_decoder & rdec, const int outfd ) : partial_data_pos( 0 ), - format_version( header.version ), + member_version( header.version() ), dictionary_size( header.dictionary_size() ), buffer_size( std::max( 65536, dictionary_size ) ), buffer( new uint8_t[buffer_size] ), pos( 0 ), stream_pos( 0 ), - crc_( 0xFFFFFFFF ), - odes_( odes ), - range_decoder( sizeof header, ibuf ), - literal_decoder() + crc_( 0xFFFFFFFFU ), + outfd_( outfd ), + range_decoder( rdec ) { buffer[buffer_size-1] = 0; } // prev_byte of first_byte ~LZ_decoder() { delete[] buffer; } - uint32_t crc() const throw() { return crc_ ^ 0xFFFFFFFF; } + uint32_t crc() const throw() { return crc_ ^ 0xFFFFFFFFU; } int decode_member( const Pretty_print & pp ); long long member_position() const throw() |