summaryrefslogtreecommitdiffstats
path: root/decoder.h
diff options
context:
space:
mode:
Diffstat (limited to 'decoder.h')
-rw-r--r--decoder.h393
1 files changed, 393 insertions, 0 deletions
diff --git a/decoder.h b/decoder.h
new file mode 100644
index 0000000..7df68b9
--- /dev/null
+++ b/decoder.h
@@ -0,0 +1,393 @@
+/* Clzip - A data compressor based on the LZMA algorithm
+ Copyright (C) 2010 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
+ the Free Software Foundation, either version 3 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License
+ along with this program. If not, see <http://www.gnu.org/licenses/>.
+*/
+
+enum { Ib_buffer_size = 65536 };
+
+struct Input_buffer
+ {
+ uint8_t * buffer;
+ int pos;
+ int stream_pos; // when reached, a new block must be read
+ int infd_; // input file descriptor
+ bool at_stream_end;
+ };
+
+bool Ib_read_block( struct Input_buffer * const ibuf );
+
+static inline void Ib_init( struct Input_buffer * const ibuf, const int infd )
+ {
+ ibuf->buffer = (uint8_t *)malloc( Ib_buffer_size );
+ if( !ibuf->buffer )
+ {
+ show_error( "not enough memory. Find a machine with more memory", 0, false );
+ cleanup_and_fail( 1 );
+ }
+ ibuf->pos = 0;
+ ibuf->stream_pos = 0;
+ ibuf->infd_ = infd;
+ ibuf->at_stream_end = false;
+ }
+
+static inline void Ib_free( struct Input_buffer * const ibuf )
+ { free( ibuf->buffer ); ibuf->buffer = 0; }
+
+static inline bool Ib_finished( struct Input_buffer * const ibuf )
+ { return ibuf->at_stream_end && ibuf->pos >= ibuf->stream_pos; }
+
+static inline uint8_t Ib_get_byte( struct Input_buffer * const ibuf )
+ {
+ if( ibuf->pos >= ibuf->stream_pos && !Ib_read_block( ibuf ) )
+ return 0;
+ return ibuf->buffer[ibuf->pos++];
+ }
+
+
+struct Range_decoder
+ {
+ long long member_pos;
+ uint32_t code;
+ uint32_t range;
+ struct Input_buffer * ibuf;
+ };
+
+static inline uint8_t Rd_get_byte( struct Range_decoder * const range_decoder )
+ {
+ ++range_decoder->member_pos;
+ return Ib_get_byte( range_decoder->ibuf );
+ }
+
+static inline void Rd_init( struct Range_decoder * const range_decoder,
+ struct Input_buffer * const buf )
+ {
+ range_decoder->member_pos = sizeof (File_header);
+ range_decoder->code = 0;
+ range_decoder->range = 0xFFFFFFFF;
+ range_decoder->ibuf = buf;
+ for( int i = 0; i < 5; ++i )
+ range_decoder->code = (range_decoder->code << 8) |
+ Rd_get_byte( range_decoder );
+ }
+
+static inline bool Rd_code_is_zero( struct Range_decoder * const range_decoder )
+ { return ( range_decoder->code == 0 ); }
+static inline bool Rd_finished( struct Range_decoder * const range_decoder )
+ { return Ib_finished( range_decoder->ibuf ); }
+static inline long long Rd_member_position( struct Range_decoder * const range_decoder )
+ { return range_decoder->member_pos; }
+
+static inline void Rd_reload( struct Range_decoder * const range_decoder )
+ {
+ range_decoder->code = 0;
+ range_decoder->range = 0xFFFFFFFF;
+ for( int i = 0; i < 5; ++i )
+ range_decoder->code = (range_decoder->code << 8) |
+ Rd_get_byte( range_decoder );
+ }
+
+static inline void Rd_normalize( struct Range_decoder * const range_decoder )
+ {
+ if( range_decoder->range <= 0x00FFFFFF )
+ {
+ range_decoder->range <<= 8;
+ range_decoder->code = (range_decoder->code << 8) |
+ Rd_get_byte( range_decoder );
+ }
+ }
+
+static inline int Rd_decode( struct Range_decoder * const range_decoder,
+ const int num_bits )
+ {
+ int symbol = 0;
+ for( int i = num_bits; i > 0; --i )
+ {
+ symbol <<= 1;
+ if( range_decoder->range <= 0x00FFFFFF )
+ {
+ range_decoder->range <<= 7;
+ range_decoder->code = (range_decoder->code << 8) |
+ Rd_get_byte( range_decoder );
+ if( range_decoder->code >= range_decoder->range )
+ { range_decoder->code -= range_decoder->range; symbol |= 1; }
+ }
+ else
+ {
+ range_decoder->range >>= 1;
+ if( range_decoder->code >= range_decoder->range )
+ { range_decoder->code -= range_decoder->range; symbol |= 1; }
+ }
+ }
+ return symbol;
+ }
+
+static inline int Rd_decode_bit( struct Range_decoder * const range_decoder,
+ Bit_model * const probability )
+ {
+ Rd_normalize( range_decoder );
+ const uint32_t bound = ( range_decoder->range >> bit_model_total_bits ) *
+ *probability;
+ if( range_decoder->code < bound )
+ {
+ range_decoder->range = bound;
+ *probability += (bit_model_total - *probability) >> bit_model_move_bits;
+ return 0;
+ }
+ else
+ {
+ range_decoder->range -= bound;
+ range_decoder->code -= bound;
+ *probability -= *probability >> bit_model_move_bits;
+ return 1;
+ }
+ }
+
+static inline int Rd_decode_tree( struct Range_decoder * const range_decoder,
+ Bit_model bm[], const int num_bits )
+ {
+ int model = 1;
+ for( int i = num_bits; i > 0; --i )
+ model = ( model << 1 ) | Rd_decode_bit( range_decoder, &bm[model] );
+ return model - (1 << num_bits);
+ }
+
+static inline int Rd_decode_tree_reversed( struct Range_decoder * const range_decoder,
+ Bit_model bm[], const int num_bits )
+ {
+ int model = 1;
+ int symbol = 0;
+ for( int i = 0; i < num_bits; ++i )
+ {
+ const int bit = Rd_decode_bit( range_decoder, &bm[model] );
+ model <<= 1;
+ if( bit ) { model |= 1; symbol |= (1 << i); }
+ }
+ return symbol;
+ }
+
+static inline int Rd_decode_matched( struct Range_decoder * const range_decoder,
+ Bit_model bm[], const int match_byte )
+ {
+ Bit_model *bm1 = bm + 0x100;
+ int symbol = 1;
+ for( int i = 1; i <= 8; ++i )
+ {
+ const int match_bit = ( match_byte << i ) & 0x100;
+ const int bit = Rd_decode_bit( range_decoder, &bm1[match_bit+symbol] );
+ symbol = ( symbol << 1 ) | bit;
+ if( ( match_bit && !bit ) || ( !match_bit && bit ) )
+ {
+ while( ++i <= 8 )
+ symbol = ( symbol << 1 ) | Rd_decode_bit( range_decoder, &bm[symbol] );
+ break;
+ }
+ }
+ return symbol & 0xFF;
+ }
+
+
+struct Len_decoder
+ {
+ Bit_model choice1;
+ Bit_model choice2;
+ Bit_model bm_low[pos_states][len_low_symbols];
+ Bit_model bm_mid[pos_states][len_mid_symbols];
+ Bit_model bm_high[len_high_symbols];
+ };
+
+static inline void Led_init( struct Len_decoder * const len_decoder )
+ {
+ Bm_init( &len_decoder->choice1 );
+ Bm_init( &len_decoder->choice2 );
+ for( int i = 0; i < pos_states; ++i )
+ for( int j = 0; j < len_low_symbols; ++j )
+ Bm_init( &len_decoder->bm_low[i][j] );
+ for( int i = 0; i < pos_states; ++i )
+ for( int j = 0; j < len_mid_symbols; ++j )
+ Bm_init( &len_decoder->bm_mid[i][j] );
+ for( int i = 0; i < len_high_symbols; ++i )
+ Bm_init( &len_decoder->bm_high[i] );
+ }
+
+static inline int Led_decode( struct Len_decoder * const len_decoder,
+ struct Range_decoder * const range_decoder,
+ const int pos_state )
+ {
+ if( Rd_decode_bit( range_decoder, &len_decoder->choice1 ) == 0 )
+ return Rd_decode_tree( range_decoder, len_decoder->bm_low[pos_state],
+ len_low_bits );
+ if( Rd_decode_bit( range_decoder, &len_decoder->choice2 ) == 0 )
+ return len_low_symbols +
+ Rd_decode_tree( range_decoder, len_decoder->bm_mid[pos_state],
+ len_mid_bits );
+ return len_low_symbols + len_mid_symbols +
+ Rd_decode_tree( range_decoder, len_decoder->bm_high, len_high_bits );
+ }
+
+
+struct Literal_decoder
+ {
+ Bit_model bm_literal[1<<literal_context_bits][0x300];
+ };
+
+static inline void Lid_init( struct Literal_decoder * const literal_decoder )
+ {
+ for( int i = 0; i < 1<<literal_context_bits; ++i )
+ for( int j = 0; j < 0x300; ++j )
+ Bm_init( &literal_decoder->bm_literal[i][j] );
+ }
+
+static inline int Lid_state( const int prev_byte )
+ { return ( prev_byte >> ( 8 - literal_context_bits ) ); }
+
+static inline uint8_t Lid_decode( struct Literal_decoder * const literal_decoder,
+ struct Range_decoder * const range_decoder,
+ const uint8_t prev_byte )
+ { return Rd_decode_tree( range_decoder, literal_decoder->bm_literal[Lid_state(prev_byte)], 8 ); }
+
+static inline uint8_t Lid_decode_matched( struct Literal_decoder * const literal_decoder,
+ struct Range_decoder * const range_decoder,
+ const uint8_t prev_byte,
+ const uint8_t match_byte )
+ { return Rd_decode_matched( range_decoder, literal_decoder->bm_literal[Lid_state(prev_byte)], match_byte ); }
+
+
+struct LZ_decoder
+ {
+ long long partial_data_pos;
+ int format_version;
+ int dictionary_size;
+ int buffer_size;
+ uint8_t * buffer;
+ int pos;
+ int stream_pos; // first byte not yet written to file
+ uint32_t crc_;
+ int outfd_; // output file descriptor
+
+ Bit_model bm_match[St_states][pos_states];
+ Bit_model bm_rep[St_states];
+ Bit_model bm_rep0[St_states];
+ Bit_model bm_rep1[St_states];
+ Bit_model bm_rep2[St_states];
+ Bit_model bm_len[St_states][pos_states];
+ Bit_model bm_dis_slot[max_dis_states][1<<dis_slot_bits];
+ Bit_model bm_dis[modeled_distances-end_dis_model];
+ Bit_model bm_align[dis_align_size];
+
+ struct Range_decoder range_decoder;
+ struct Len_decoder len_decoder;
+ struct Len_decoder rep_match_len_decoder;
+ struct Literal_decoder literal_decoder;
+ };
+
+void LZd_flush_data( struct LZ_decoder * const decoder );
+
+static inline uint8_t LZd_get_byte( struct LZ_decoder * const decoder,
+ const int distance )
+ {
+ int i = decoder->pos - distance - 1;
+ if( i < 0 ) i += decoder->buffer_size;
+ return decoder->buffer[i];
+ }
+
+static inline void LZd_put_byte( struct LZ_decoder * const decoder,
+ const uint8_t b )
+ {
+ decoder->buffer[decoder->pos] = b;
+ if( ++decoder->pos >= decoder->buffer_size )
+ LZd_flush_data( decoder );
+ }
+
+static inline void LZd_copy_block( struct LZ_decoder * const decoder,
+ const int distance, int len )
+ {
+ int i = decoder->pos - distance - 1;
+ if( i < 0 ) i += decoder->buffer_size;
+ if( len < decoder->buffer_size - max( decoder->pos, i ) &&
+ len <= abs( decoder->pos - i ) )
+ {
+ memcpy( decoder->buffer + decoder->pos, decoder->buffer + i, len );
+ decoder->pos += len;
+ }
+ else for( ; len > 0 ; --len )
+ {
+ decoder->buffer[decoder->pos] = decoder->buffer[i];
+ if( ++decoder->pos >= decoder->buffer_size )
+ LZd_flush_data( decoder );
+ if( ++i >= decoder->buffer_size ) i = 0;
+ }
+ }
+
+bool LZd_verify_trailer( struct LZ_decoder * const decoder,
+ struct Pretty_print * const pp );
+
+static inline void LZd_init( struct LZ_decoder * const decoder,
+ const File_header header,
+ struct Input_buffer * const ibuf, const int outfd )
+ {
+ decoder->partial_data_pos = 0;
+ decoder->format_version = Fh_version( header );
+ decoder->dictionary_size = Fh_get_dictionary_size( header );
+ decoder->buffer_size = max( 65536, decoder->dictionary_size );
+ decoder->buffer = (uint8_t *)malloc( decoder->buffer_size );
+ if( !decoder->buffer )
+ {
+ show_error( "not enough memory. Find a machine with more memory", 0, false );
+ cleanup_and_fail( 1 );
+ }
+ decoder->pos = 0;
+ decoder->stream_pos = 0;
+ decoder->crc_ = 0xFFFFFFFF;
+ decoder->outfd_ = outfd;
+
+ for( int i = 0; i < St_states; ++i )
+ {
+ for( int j = 0; j < pos_states; ++j )
+ {
+ Bm_init( &decoder->bm_match[i][j] );
+ Bm_init( &decoder->bm_len[i][j] );
+ }
+ Bm_init( &decoder->bm_rep[i] );
+ Bm_init( &decoder->bm_rep0[i] );
+ Bm_init( &decoder->bm_rep1[i] );
+ Bm_init( &decoder->bm_rep2[i] );
+ }
+ for( int i = 0; i < max_dis_states; ++i )
+ for( int j = 0; j < 1<<dis_slot_bits; ++j )
+ Bm_init( &decoder->bm_dis_slot[i][j] );
+ for( int i = 0; i < modeled_distances-end_dis_model; ++i )
+ Bm_init( &decoder->bm_dis[i] );
+ for( int i = 0; i < dis_align_size; ++i )
+ Bm_init( &decoder->bm_align[i] );
+
+ Rd_init( &decoder->range_decoder, ibuf );
+ Led_init( &decoder->len_decoder );
+ Led_init( &decoder->rep_match_len_decoder );
+ Lid_init( &decoder->literal_decoder );
+ decoder->buffer[decoder->buffer_size-1] = 0; // prev_byte of first_byte
+ }
+
+static inline void LZd_free( struct LZ_decoder * const decoder )
+ { free( decoder->buffer ); decoder->buffer = 0; }
+
+static inline uint32_t LZd_crc( struct LZ_decoder * const decoder )
+ { return decoder->crc_ ^ 0xFFFFFFFF; }
+int LZd_decode_member( struct LZ_decoder * const decoder,
+ struct Pretty_print * const pp );
+
+static inline long long LZd_member_position( struct LZ_decoder * const decoder )
+ { return Rd_member_position( &decoder->range_decoder ); }
+static inline long long LZd_data_position( struct LZ_decoder * const decoder )
+ { return decoder->partial_data_pos + decoder->pos; }