/*-------------------------------------------------------------*//*--- Compression machinery (not incl block sorting)        ---*//*---                                            compress.c ---*//*-------------------------------------------------------------*//*--  This file is a part of bzip2 and/or libbzip2, a program and  library for lossless, block-sorting data compression.  Copyright (C) 1996-2000 Julian R Seward.  All rights reserved.  Redistribution and use in source and binary forms, with or without  modification, are permitted provided that the following conditions  are met:  1. Redistributions of source code must retain the above copyright     notice, this list of conditions and the following disclaimer.  2. The origin of this software must not be misrepresented; you must      not claim that you wrote the original software.  If you use this      software in a product, an acknowledgment in the product      documentation would be appreciated but is not required.  3. Altered source versions must be plainly marked as such, and must     not be misrepresented as being the original software.  4. The name of the author may not be used to endorse or promote      products derived from this software without specific prior written      permission.  THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS  OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED  WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE  ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY  DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL  DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE  GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY,  WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING  NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS  SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.  Julian Seward, Cambridge, UK.  jseward@acm.org  bzip2/libbzip2 version 1.0 of 21 March 2000  This program is based on (at least) the work of:     Mike Burrows     David Wheeler     Peter Fenwick     Alistair Moffat     Radford Neal     Ian H. Witten     Robert Sedgewick     Jon L. Bentley  For more information on these sources, see the manual.--*//*--   CHANGES   ~~~~~~~   0.9.0 -- original version.   0.9.0a/b -- no changes in this file.   0.9.0c      * changed setting of nGroups in sendMTFValues() so as to         do a bit better on small files--*/#include "bzlib_private.h"/*---------------------------------------------------*//*--- Bit stream I/O                              ---*//*---------------------------------------------------*//*---------------------------------------------------*/void BZ2_bsInitWrite ( EState* s ){   s->bsLive = 0;   s->bsBuff = 0;}/*---------------------------------------------------*/staticvoid bsFinishWrite ( EState* s ){   while (s->bsLive > 0) {      s->zbits[s->numZ] = (UChar)(s->bsBuff >> 24);      s->numZ++;      s->bsBuff <<= 8;      s->bsLive -= 8;   }}/*---------------------------------------------------*/#define bsNEEDW(nz)                           \{                                             \   while (s->bsLive >= 8) {                   \      s->zbits[s->numZ]                       \         = (UChar)(s->bsBuff >> 24);          \      s->numZ++;                              \      s->bsBuff <<= 8;                        \      s->bsLive -= 8;                         \   }                                          \}/*---------------------------------------------------*/static__inline__void bsW ( EState* s, Int32 n, UInt32 v ){   bsNEEDW ( n );   s->bsBuff |= (v << (32 - s->bsLive - n));   s->bsLive += n;}/*---------------------------------------------------*/staticvoid bsPutUInt32 ( EState* s, UInt32 u ){   bsW ( s, 8, (u >> 24) & 0xffL );   bsW ( s, 8, (u >> 16) & 0xffL );   bsW ( s, 8, (u >>  8) & 0xffL );   bsW ( s, 8,  u        & 0xffL );}/*---------------------------------------------------*/staticvoid bsPutUChar ( EState* s, UChar c ){   bsW( s, 8, (UInt32)c );}/*---------------------------------------------------*//*--- The back end proper                         ---*//*---------------------------------------------------*//*---------------------------------------------------*/staticvoid makeMaps_e ( EState* s ){   Int32 i;   s->nInUse = 0;   for (i = 0; i < 256; i++)      if (s->inUse[i]) {         s->unseqToSeq[i] = s->nInUse;         s->nInUse++;      }}/*---------------------------------------------------*/staticvoid generateMTFValues ( EState* s ){   UChar   yy[256];   Int32   i, j;   Int32   zPend;   Int32   wr;   Int32   EOB;   /*       After sorting (eg, here),         s->arr1 [ 0 .. s->nblock-1 ] holds sorted order,         and         ((UChar*)s->arr2) [ 0 .. s->nblock-1 ]          holds the original block data.      The first thing to do is generate the MTF values,      and put them in         ((UInt16*)s->arr1) [ 0 .. s->nblock-1 ].      Because there are strictly fewer or equal MTF values      than block values, ptr values in this area are overwritten      with MTF values only when they are no longer needed.      The final compressed bitstream is generated into the      area starting at         (UChar*) (&((UChar*)s->arr2)[s->nblock])      These storage aliases are set up in bzCompressInit(),      except for the last one, which is arranged in       compressBlock().   */   UInt32* ptr   = s->ptr;   UChar* block  = s->block;   UInt16* mtfv  = s->mtfv;   makeMaps_e ( s );   EOB = s->nInUse+1;   for (i = 0; i <= EOB; i++) s->mtfFreq[i] = 0;   wr = 0;   zPend = 0;   for (i = 0; i < s->nInUse; i++) yy[i] = (UChar) i;   for (i = 0; i < s->nblock; i++) {      UChar ll_i;      AssertD ( wr <= i, "generateMTFValues(1)" );      j = ptr[i]-1; if (j < 0) j += s->nblock;      ll_i = s->unseqToSeq[block[j]];      AssertD ( ll_i < s->nInUse, "generateMTFValues(2a)" );      if (yy[0] == ll_i) {          zPend++;      } else {         if (zPend > 0) {            zPend--;            while (True) {               if (zPend & 1) {                  mtfv[wr] = BZ_RUNB; wr++;                   s->mtfFreq[BZ_RUNB]++;                } else {                  mtfv[wr] = BZ_RUNA; wr++;                   s->mtfFreq[BZ_RUNA]++;                }               if (zPend < 2) break;               zPend = (zPend - 2) / 2;            };            zPend = 0;         }         {            register UChar  rtmp;            register UChar* ryy_j;            register UChar  rll_i;            rtmp  = yy[1];            yy[1] = yy[0];            ryy_j = &(yy[1]);            rll_i = ll_i;            while ( rll_i != rtmp ) {               register UChar rtmp2;               ryy_j++;               rtmp2  = rtmp;               rtmp   = *ryy_j;               *ryy_j = rtmp2;            };            yy[0] = rtmp;            j = ryy_j - &(yy[0]);            mtfv[wr] = j+1; wr++; s->mtfFreq[j+1]++;         }      }   }   if (zPend > 0) {      zPend--;      while (True) {         if (zPend & 1) {            mtfv[wr] = BZ_RUNB; wr++;             s->mtfFreq[BZ_RUNB]++;          } else {            mtfv[wr] = BZ_RUNA; wr++;             s->mtfFreq[BZ_RUNA]++;          }         if (zPend < 2) break;         zPend = (zPend - 2) / 2;      };      zPend = 0;   }   mtfv[wr] = EOB; wr++; s->mtfFreq[EOB]++;   s->nMTF = wr;}/*---------------------------------------------------*/#define BZ_LESSER_ICOST  0#define BZ_GREATER_ICOST 15staticvoid sendMTFValues ( EState* s ){   Int32 v, t, i, j, gs, ge, totc, bt, bc, iter;   Int32 nSelectors, alphaSize, minLen, maxLen, selCtr;   Int32 nGroups, nBytes;   /*--   UChar  len [BZ_N_GROUPS][BZ_MAX_ALPHA_SIZE];   is a global since the decoder also needs it.   Int32  code[BZ_N_GROUPS][BZ_MAX_ALPHA_SIZE];   Int32  rfreq[BZ_N_GROUPS][BZ_MAX_ALPHA_SIZE];   are also globals only used in this proc.   Made global to keep stack frame size small.   --*/   UInt16 cost[BZ_N_GROUPS];   Int32  fave[BZ_N_GROUPS];   UInt16* mtfv = s->mtfv;   if (s->verbosity >= 3)      VPrintf3( "      %d in block, %d after MTF & 1-2 coding, "                "%d+2 syms in use\n",                 s->nblock, s->nMTF, s->nInUse );   alphaSize = s->nInUse+2;   for (t = 0; t < BZ_N_GROUPS; t++)      for (v = 0; v < alphaSize; v++)         s->len[t][v] = BZ_GREATER_ICOST;   /*--- Decide how many coding tables to use ---*/   AssertH ( s->nMTF > 0, 3001 );   if (s->nMTF < 200)  nGroups = 2; else   if (s->nMTF < 600)  nGroups = 3; else   if (s->nMTF < 1200) nGroups = 4; else   if (s->nMTF < 2400) nGroups = 5; else                       nGroups = 6;   /*--- Generate an initial set of coding tables ---*/   {       Int32 nPart, remF, tFreq, aFreq;      nPart = nGroups;      remF  = s->nMTF;      gs = 0;      while (nPart > 0) {         tFreq = remF / nPart;         ge = gs-1;         aFreq = 0;         while (aFreq < tFreq && ge < alphaSize-1) {            ge++;            aFreq += s->mtfFreq[ge];         }         if (ge > gs              && nPart != nGroups && nPart != 1              && ((nGroups-nPart) % 2 == 1)) {            aFreq -= s->mtfFreq[ge];            ge--;         }         if (s->verbosity >= 3)            VPrintf5( "      initial group %d, [%d .. %d], "                      "has %d syms (%4.1f%%)\n",                      nPart, gs, ge, aFreq,                       (100.0 * (float)aFreq) / (float)(s->nMTF) );          for (v = 0; v < alphaSize; v++)            if (v >= gs && v <= ge)                s->len[nPart-1][v] = BZ_LESSER_ICOST; else               s->len[nPart-1][v] = BZ_GREATER_ICOST;          nPart--;         gs = ge+1;