path: root/encoder.c

/* Clzip - LZMA lossless data compressor
   Copyright (C) 2010-2024 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 2 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/>.
*/

#define _FILE_OFFSET_BITS 64

#include <errno.h>
#include <stdbool.h>
#include <stdint.h>
#include <stdlib.h>
#include <string.h>

#include "lzip.h"
#include "encoder_base.h"
#include "encoder.h"


CRC32 crc32;


int LZe_get_match_pairs( struct LZ_encoder * const e, struct Pair * pairs )
  {
  int len_limit = e->match_len_limit;
  if( len_limit > Mb_available_bytes( &e->eb.mb ) )
    {
    len_limit = Mb_available_bytes( &e->eb.mb );
    if( len_limit < 4 ) return 0;
    }

  int maxlen = 3;			/* only used if pairs != 0 */
  int num_pairs = 0;
  const int min_pos = ( e->eb.mb.pos > e->eb.mb.dictionary_size ) ?
                        e->eb.mb.pos - e->eb.mb.dictionary_size : 0;
  const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb );

  unsigned tmp = crc32[data[0]] ^ data[1];
  const int key2 = tmp & ( num_prev_positions2 - 1 );
  tmp ^= (unsigned)data[2] << 8;
  const int key3 = num_prev_positions2 + ( tmp & ( num_prev_positions3 - 1 ) );
  const int key4 = num_prev_positions2 + num_prev_positions3 +
                   ( ( tmp ^ ( crc32[data[3]] << 5 ) ) & e->eb.mb.key4_mask );

  if( pairs )
    {
    const int np2 = e->eb.mb.prev_positions[key2];
    const int np3 = e->eb.mb.prev_positions[key3];
    if( np2 > min_pos && e->eb.mb.buffer[np2-1] == data[0] )
      {
      pairs[0].dis = e->eb.mb.pos - np2;
      pairs[0].len = maxlen = 2 + ( np2 == np3 );
      num_pairs = 1;
      }
    if( np2 != np3 && np3 > min_pos && e->eb.mb.buffer[np3-1] == data[0] )
      {
      maxlen = 3;
      pairs[num_pairs++].dis = e->eb.mb.pos - np3;
      }
    if( num_pairs > 0 )
      {
      const int delta = pairs[num_pairs-1].dis + 1;
      while( maxlen < len_limit && data[maxlen-delta] == data[maxlen] )
        ++maxlen;
      pairs[num_pairs-1].len = maxlen;
      if( maxlen < 3 ) maxlen = 3;
      if( maxlen >= len_limit ) pairs = 0;	/* done. now just skip */
      }
    }

  const int pos1 = e->eb.mb.pos + 1;
  e->eb.mb.prev_positions[key2] = pos1;
  e->eb.mb.prev_positions[key3] = pos1;
  int newpos1 = e->eb.mb.prev_positions[key4];
  e->eb.mb.prev_positions[key4] = pos1;

  int32_t * ptr0 = e->eb.mb.pos_array + ( e->eb.mb.cyclic_pos << 1 );
  int32_t * ptr1 = ptr0 + 1;
  int len = 0, len0 = 0, len1 = 0;

  int count;
  for( count = e->cycles; ; )
    {
    if( newpos1 <= min_pos || --count < 0 ) { *ptr0 = *ptr1 = 0; break; }

    const int delta = pos1 - newpos1;
    int32_t * const newptr = e->eb.mb.pos_array +
      ( ( e->eb.mb.cyclic_pos - delta +
          ( (e->eb.mb.cyclic_pos >= delta) ? 0 : e->eb.mb.dictionary_size + 1 ) ) << 1 );
    if( data[len-delta] == data[len] )
      {
      while( ++len < len_limit && data[len-delta] == data[len] ) {}
      if( pairs && maxlen < len )
        {
        pairs[num_pairs].dis = delta - 1;
        pairs[num_pairs].len = maxlen = len;
        ++num_pairs;
        }
      if( len >= len_limit )
        {
        *ptr0 = newptr[0];
        *ptr1 = newptr[1];
        break;
        }
      }
    if( data[len-delta] < data[len] )
      {
      *ptr0 = newpos1;
      ptr0 = newptr + 1;
      newpos1 = *ptr0;
      len0 = len; if( len1 < len ) len = len1;
      }
    else
      {
      *ptr1 = newpos1;
      ptr1 = newptr;
      newpos1 = *ptr1;
      len1 = len; if( len0 < len ) len = len0;
      }
    }
  return num_pairs;
  }


static void LZe_update_distance_prices( struct LZ_encoder * const e )
  {
  int dis, len_state;
  for( dis = start_dis_model; dis < modeled_distances; ++dis )
    {
    const int dis_slot = dis_slots[dis];
    const int direct_bits = ( dis_slot >> 1 ) - 1;
    const int base = ( 2 | ( dis_slot & 1 ) ) << direct_bits;
    const int price = price_symbol_reversed( e->eb.bm_dis + ( base - dis_slot ),
                                             dis - base, direct_bits );
    for( len_state = 0; len_state < len_states; ++len_state )
      e->dis_prices[len_state][dis] = price;
    }

  for( len_state = 0; len_state < len_states; ++len_state )
    {
    int * const dsp = e->dis_slot_prices[len_state];
    const Bit_model * const bmds = e->eb.bm_dis_slot[len_state];
    int slot = 0;
    for( ; slot < end_dis_model; ++slot )
      dsp[slot] = price_symbol6( bmds, slot );
    for( ; slot < e->num_dis_slots; ++slot )
      dsp[slot] = price_symbol6( bmds, slot ) +
                  (((( slot >> 1 ) - 1 ) - dis_align_bits ) << price_shift_bits );

    int * const dp = e->dis_prices[len_state];
    for( dis = 0; dis < start_dis_model; ++dis )
      dp[dis] = dsp[dis];
    for( ; dis < modeled_distances; ++dis )
      dp[dis] += dsp[dis_slots[dis]];
    }
  }


/* Return the number of bytes advanced (ahead).
   trials[0]..trials[ahead-1] contain the steps to encode.
   ( trials[0].dis4 == -1 ) means literal.
   A match/rep longer or equal than match_len_limit finishes the sequence.
*/
static int LZe_sequence_optimizer( struct LZ_encoder * const e,
                                   const int reps[num_rep_distances],
                                   const State state )
  {
  int num_pairs, num_trials;
  int i, rep, len;

  if( e->pending_num_pairs > 0 )		/* from previous call */
    {
    num_pairs = e->pending_num_pairs;
    e->pending_num_pairs = 0;
    }
  else
    num_pairs = LZe_read_match_distances( e );
  const int main_len = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0;

  int replens[num_rep_distances];
  int rep_index = 0;
  for( i = 0; i < num_rep_distances; ++i )
    {
    replens[i] = Mb_true_match_len( &e->eb.mb, 0, reps[i] + 1 );
    if( replens[i] > replens[rep_index] ) rep_index = i;
    }
  if( replens[rep_index] >= e->match_len_limit )
    {
    e->trials[0].price = replens[rep_index];
    e->trials[0].dis4 = rep_index;
    LZe_move_and_update( e, replens[rep_index] );
    return replens[rep_index];
    }

  if( main_len >= e->match_len_limit )
    {
    e->trials[0].price = main_len;
    e->trials[0].dis4 = e->pairs[num_pairs-1].dis + num_rep_distances;
    LZe_move_and_update( e, main_len );
    return main_len;
    }

  const int pos_state = Mb_data_position( &e->eb.mb ) & pos_state_mask;
  const uint8_t prev_byte = Mb_peek( &e->eb.mb, 1 );
  const uint8_t cur_byte = Mb_peek( &e->eb.mb, 0 );
  const uint8_t match_byte = Mb_peek( &e->eb.mb, reps[0] + 1 );

  e->trials[1].price = price0( e->eb.bm_match[state][pos_state] );
  if( St_is_char( state ) )
    e->trials[1].price += LZeb_price_literal( &e->eb, prev_byte, cur_byte );
  else
    e->trials[1].price += LZeb_price_matched( &e->eb, prev_byte, cur_byte, match_byte );
  e->trials[1].dis4 = -1;				/* literal */

  const int match_price = price1( e->eb.bm_match[state][pos_state] );
  const int rep_match_price = match_price + price1( e->eb.bm_rep[state] );

  if( match_byte == cur_byte )
    Tr_update( &e->trials[1], rep_match_price +
               LZeb_price_shortrep( &e->eb, state, pos_state ), 0, 0 );

  num_trials = max( main_len, replens[rep_index] );

  if( num_trials < min_match_len )
    {
    e->trials[0].price = 1;
    e->trials[0].dis4 = e->trials[1].dis4;
    Mb_move_pos( &e->eb.mb );
    return 1;
    }

  e->trials[0].state = state;
  for( i = 0; i < num_rep_distances; ++i )
    e->trials[0].reps[i] = reps[i];

  for( len = min_match_len; len <= num_trials; ++len )
    e->trials[len].price = infinite_price;

  for( rep = 0; rep < num_rep_distances; ++rep )
    {
    if( replens[rep] < min_match_len ) continue;
    const int price = rep_match_price + LZeb_price_rep( &e->eb, rep, state, pos_state );
    for( len = min_match_len; len <= replens[rep]; ++len )
      Tr_update( &e->trials[len], price +
                 Lp_price( &e->rep_len_prices, len, pos_state ), rep, 0 );
    }

  if( main_len > replens[0] )
    {
    const int normal_match_price = match_price + price0( e->eb.bm_rep[state] );
    int i = 0, len = max( replens[0] + 1, min_match_len );
    while( len > e->pairs[i].len ) ++i;
    while( true )
      {
      const int dis = e->pairs[i].dis;
      Tr_update( &e->trials[len], normal_match_price +
                 LZe_price_pair( e, dis, len, pos_state ),
                 dis + num_rep_distances, 0 );
      if( ++len > e->pairs[i].len && ++i >= num_pairs ) break;
      }
    }

  int cur = 0;
  while( true )				/* price optimization loop */
    {
    Mb_move_pos( &e->eb.mb );
    if( ++cur >= num_trials )		/* no more initialized trials */
      {
      LZe_backward( e, cur );
      return cur;
      }

    const int num_pairs = LZe_read_match_distances( e );
    const int newlen = ( num_pairs > 0 ) ? e->pairs[num_pairs-1].len : 0;
    if( newlen >= e->match_len_limit )
      {
      e->pending_num_pairs = num_pairs;
      LZe_backward( e, cur );
      return cur;
      }

    /* give final values to current trial */
    struct Trial * cur_trial = &e->trials[cur];
    State cur_state;
    {
    const int dis4 = cur_trial->dis4;
    int prev_index = cur_trial->prev_index;
    const int prev_index2 = cur_trial->prev_index2;

    if( prev_index2 == single_step_trial )
      {
      cur_state = e->trials[prev_index].state;
      if( prev_index + 1 == cur )			/* len == 1 */
        {
        if( dis4 == 0 ) cur_state = St_set_short_rep( cur_state );
        else cur_state = St_set_char( cur_state );	/* literal */
        }
      else if( dis4 < num_rep_distances ) cur_state = St_set_rep( cur_state );
      else cur_state = St_set_match( cur_state );
      }
    else
      {
      if( prev_index2 == dual_step_trial )	/* dis4 == 0 (rep0) */
        --prev_index;
      else					/* prev_index2 >= 0 */
        prev_index = prev_index2;
      cur_state = St_set_char_rep();
      }
    cur_trial->state = cur_state;
    for( i = 0; i < num_rep_distances; ++i )
      cur_trial->reps[i] = e->trials[prev_index].reps[i];
    mtf_reps( dis4, cur_trial->reps );		/* literal is ignored */
    }

    const int pos_state = Mb_data_position( &e->eb.mb ) & pos_state_mask;
    const uint8_t prev_byte = Mb_peek( &e->eb.mb, 1 );
    const uint8_t cur_byte = Mb_peek( &e->eb.mb, 0 );
    const uint8_t match_byte = Mb_peek( &e->eb.mb, cur_trial->reps[0] + 1 );

    int next_price = cur_trial->price +
                     price0( e->eb.bm_match[cur_state][pos_state] );
    if( St_is_char( cur_state ) )
      next_price += LZeb_price_literal( &e->eb, prev_byte, cur_byte );
    else
      next_price += LZeb_price_matched( &e->eb, prev_byte, cur_byte, match_byte );

    /* try last updates to next trial */
    struct Trial * next_trial = &e->trials[cur+1];

    Tr_update( next_trial, next_price, -1, cur );	/* literal */

    const int match_price = cur_trial->price + price1( e->eb.bm_match[cur_state][pos_state] );
    const int rep_match_price = match_price + price1( e->eb.bm_rep[cur_state] );

    if( match_byte == cur_byte && next_trial->dis4 != 0 &&
        next_trial->prev_index2 == single_step_trial )
      {
      const int price = rep_match_price +
                        LZeb_price_shortrep( &e->eb, cur_state, pos_state );
      if( price <= next_trial->price )
        {
        next_trial->price = price;
        next_trial->dis4 = 0;				/* rep0 */
        next_trial->prev_index = cur;
        }
      }

    const int triable_bytes =
      min( Mb_available_bytes( &e->eb.mb ), max_num_trials - 1 - cur );
    if( triable_bytes < min_match_len ) continue;

    const int len_limit = min( e->match_len_limit, triable_bytes );

    /* try literal + rep0 */
    if( match_byte != cur_byte && next_trial->prev_index != cur )
      {
      const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb );
      const int dis = cur_trial->reps[0] + 1;
      const int limit = min( e->match_len_limit + 1, triable_bytes );
      int len = 1;
      while( len < limit && data[len-dis] == data[len] ) ++len;
      if( --len >= min_match_len )
        {
        const int pos_state2 = ( pos_state + 1 ) & pos_state_mask;
        const State state2 = St_set_char( cur_state );
        const int price = next_price +
                          price1( e->eb.bm_match[state2][pos_state2] ) +
                          price1( e->eb.bm_rep[state2] ) +
                          LZe_price_rep0_len( e, len, state2, pos_state2 );
        while( num_trials < cur + 1 + len )
          e->trials[++num_trials].price = infinite_price;
        Tr_update2( &e->trials[cur+1+len], price, cur + 1 );
        }
      }

    int start_len = min_match_len;

    /* try rep distances */
    for( rep = 0; rep < num_rep_distances; ++rep )
      {
      const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb );
      const int dis = cur_trial->reps[rep] + 1;

      if( data[0-dis] != data[0] || data[1-dis] != data[1] ) continue;
      for( len = min_match_len; len < len_limit; ++len )
        if( data[len-dis] != data[len] ) break;
      while( num_trials < cur + len )
        e->trials[++num_trials].price = infinite_price;
      int price = rep_match_price + LZeb_price_rep( &e->eb, rep, cur_state, pos_state );
      for( i = min_match_len; i <= len; ++i )
        Tr_update( &e->trials[cur+i], price +
                   Lp_price( &e->rep_len_prices, i, pos_state ), rep, cur );

      if( rep == 0 ) start_len = len + 1;	/* discard shorter matches */

      /* try rep + literal + rep0 */
      int len2 = len + 1;
      const int limit = min( e->match_len_limit + len2, triable_bytes );
      while( len2 < limit && data[len2-dis] == data[len2] ) ++len2;
      len2 -= len + 1;
      if( len2 < min_match_len ) continue;

      int pos_state2 = ( pos_state + len ) & pos_state_mask;
      State state2 = St_set_rep( cur_state );
      price += Lp_price( &e->rep_len_prices, len, pos_state ) +
               price0( e->eb.bm_match[state2][pos_state2] ) +
               LZeb_price_matched( &e->eb, data[len-1], data[len], data[len-dis] );
      pos_state2 = ( pos_state2 + 1 ) & pos_state_mask;
      state2 = St_set_char( state2 );
      price += price1( e->eb.bm_match[state2][pos_state2] ) +
               price1( e->eb.bm_rep[state2] ) +
               LZe_price_rep0_len( e, len2, state2, pos_state2 );
      while( num_trials < cur + len + 1 + len2 )
        e->trials[++num_trials].price = infinite_price;
      Tr_update3( &e->trials[cur+len+1+len2], price, rep, cur + len + 1, cur );
      }

    /* try matches */
    if( newlen >= start_len && newlen <= len_limit )
      {
      const int normal_match_price = match_price +
                                     price0( e->eb.bm_rep[cur_state] );

      while( num_trials < cur + newlen )
        e->trials[++num_trials].price = infinite_price;

      int i = 0;
      while( e->pairs[i].len < start_len ) ++i;
      int dis = e->pairs[i].dis;
      for( len = start_len; ; ++len )
        {
        int price = normal_match_price + LZe_price_pair( e, dis, len, pos_state );
        Tr_update( &e->trials[cur+len], price, dis + num_rep_distances, cur );

        /* try match + literal + rep0 */
        if( len == e->pairs[i].len )
          {
          const uint8_t * const data = Mb_ptr_to_current_pos( &e->eb.mb );
          const int dis2 = dis + 1;
          int len2 = len + 1;
          const int limit = min( e->match_len_limit + len2, triable_bytes );
          while( len2 < limit && data[len2-dis2] == data[len2] ) ++len2;
          len2 -= len + 1;
          if( len2 >= min_match_len )
            {
            int pos_state2 = ( pos_state + len ) & pos_state_mask;
            State state2 = St_set_match( cur_state );
            price += price0( e->eb.bm_match[state2][pos_state2] ) +
                     LZeb_price_matched( &e->eb, data[len-1], data[len], data[len-dis2] );
            pos_state2 = ( pos_state2 + 1 ) & pos_state_mask;
            state2 = St_set_char( state2 );
            price += price1( e->eb.bm_match[state2][pos_state2] ) +
                     price1( e->eb.bm_rep[state2] ) +
                     LZe_price_rep0_len( e, len2, state2, pos_state2 );

            while( num_trials < cur + len + 1 + len2 )
              e->trials[++num_trials].price = infinite_price;
            Tr_update3( &e->trials[cur+len+1+len2], price,
                        dis + num_rep_distances, cur + len + 1, cur );
            }
          if( ++i >= num_pairs ) break;
          dis = e->pairs[i].dis;
          }
        }
      }
    }
  }


bool LZe_encode_member( struct LZ_encoder * const e,
                        const unsigned long long member_size )
  {
  const unsigned long long member_size_limit =
    member_size - Lt_size - max_marker_size;
  const bool best = ( e->match_len_limit > 12 );
  const int dis_price_count = best ? 1 : 512;
  const int align_price_count = best ? 1 : dis_align_size;
  const int price_count = ( e->match_len_limit > 36 ) ? 1013 : 4093;
  int price_counter = 0;		/* counters may decrement below 0 */
  int dis_price_counter = 0;
  int align_price_counter = 0;
  int i;
  int reps[num_rep_distances];
  State state = 0;
  for( i = 0; i < num_rep_distances; ++i ) reps[i] = 0;

  if( Mb_data_position( &e->eb.mb ) != 0 ||
      Re_member_position( &e->eb.renc ) != Lh_size )
    return false;				/* can be called only once */

  if( !Mb_data_finished( &e->eb.mb ) )		/* encode first byte */
    {
    const uint8_t prev_byte = 0;
    const uint8_t cur_byte = Mb_peek( &e->eb.mb, 0 );
    Re_encode_bit( &e->eb.renc, &e->eb.bm_match[state][0], 0 );
    LZeb_encode_literal( &e->eb, prev_byte, cur_byte );
    CRC32_update_byte( &e->eb.crc, cur_byte );
    LZe_get_match_pairs( e, 0 );
    Mb_move_pos( &e->eb.mb );
    }

  while( !Mb_data_finished( &e->eb.mb ) )
    {
    if( price_counter <= 0 && e->pending_num_pairs == 0 )
      {
      price_counter = price_count;	/* recalculate prices every these bytes */
      if( dis_price_counter <= 0 )
        { dis_price_counter = dis_price_count; LZe_update_distance_prices( e ); }
      if( align_price_counter <= 0 )
        {
        align_price_counter = align_price_count;
        for( i = 0; i < dis_align_size; ++i )
          e->align_prices[i] = price_symbol_reversed( e->eb.bm_align, i, dis_align_bits );
        }
      Lp_update_prices( &e->match_len_prices );
      Lp_update_prices( &e->rep_len_prices );
      }

    int ahead = LZe_sequence_optimizer( e, reps, state );
    price_counter -= ahead;

    for( i = 0; ahead > 0; )
      {
      const int pos_state =
        ( Mb_data_position( &e->eb.mb ) - ahead ) & pos_state_mask;
      const int len = e->trials[i].price;
      int dis = e->trials[i].dis4;

      bool bit = ( dis < 0 );
      Re_encode_bit( &e->eb.renc, &e->eb.bm_match[state][pos_state], !bit );
      if( bit )					/* literal byte */
        {
        const uint8_t prev_byte = Mb_peek( &e->eb.mb, ahead + 1 );
        const uint8_t cur_byte = Mb_peek( &e->eb.mb, ahead );
        CRC32_update_byte( &e->eb.crc, cur_byte );
        if( ( state = St_set_char( state ) ) < 4 )
          LZeb_encode_literal( &e->eb, prev_byte, cur_byte );
        else
          {
          const uint8_t match_byte = Mb_peek( &e->eb.mb, ahead + reps[0] + 1 );
          LZeb_encode_matched( &e->eb, prev_byte, cur_byte, match_byte );
          }
        }
      else					/* match or repeated match */
        {
        CRC32_update_buf( &e->eb.crc, Mb_ptr_to_current_pos( &e->eb.mb ) - ahead, len );
        mtf_reps( dis, reps );
        bit = ( dis < num_rep_distances );
        Re_encode_bit( &e->eb.renc, &e->eb.bm_rep[state], bit );
        if( bit )				/* repeated match */
          {
          bit = ( dis == 0 );
          Re_encode_bit( &e->eb.renc, &e->eb.bm_rep0[state], !bit );
          if( bit )
            Re_encode_bit( &e->eb.renc, &e->eb.bm_len[state][pos_state], len > 1 );
          else
            {
            Re_encode_bit( &e->eb.renc, &e->eb.bm_rep1[state], dis > 1 );
            if( dis > 1 )
              Re_encode_bit( &e->eb.renc, &e->eb.bm_rep2[state], dis > 2 );
            }
          if( len == 1 ) state = St_set_short_rep( state );
          else
            {
            Re_encode_len( &e->eb.renc, &e->eb.rep_len_model, len, pos_state );
            Lp_decrement_counter( &e->rep_len_prices, pos_state );
            state = St_set_rep( state );
            }
          }
        else					/* match */
          {
          dis -= num_rep_distances;
          LZeb_encode_pair( &e->eb, dis, len, pos_state );
          if( dis >= modeled_distances ) --align_price_counter;
          --dis_price_counter;
          Lp_decrement_counter( &e->match_len_prices, pos_state );
          state = St_set_match( state );
          }
        }
      ahead -= len; i += len;
      if( Re_member_position( &e->eb.renc ) >= member_size_limit )
        {
        if( !Mb_dec_pos( &e->eb.mb, ahead ) ) return false;
        LZeb_full_flush( &e->eb, state );
        return true;
        }
      }
    }
  LZeb_full_flush( &e->eb, state );
  return true;
  }