/* pucrunch.c: Pucrunch compression code Pucrunch 1997-2005 by Pasi 'Albert' Ojala, a1bert@iki.fi Copyright (C) 2000 Thomas Nussbaumer Copyright (C) 2007 Kevin Kofler 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, 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, write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. */ #include #include #include #include #include #include #include "../generic.h" #include "../formats/packhead.h" // compressed header definition #include "pucrunch.h" #define VERBOSE_OUT stdout #ifndef min #define min(a,b) ((a 1..127 */ #define MAXLZLEN (2< 1..127 */ #define DEFAULT_LZLEN LRANGE unsigned short *rle, *elr, *lzlen, *lzpos; int *length, inlen; unsigned char *indata, *mode, *newesc; unsigned short *backSkip; enum MODE { LITERAL = 0, LZ77 = 1, RLE = 2, MMARK = 4 }; int lzopt = 0; int maxGamma = 7; int reservedBytes = 2; int escBits = 2; int escMask = 0xc0; int extraLZPosBits = 0; int rleUsed = 31; unsigned char rleLen[256]; int lenValue[256]; int lrange, maxlzlen, maxrlelen; int gainedEscaped = 0; int gainedRle = 0, gainedSRle = 0, gainedLRle = 0; int gainedLz = 0, gainedRlecode = 0; int timesEscaped = 0, timesNormal = 0; int timesRle = 0, timesSRle = 0, timesLRle = 0; int timesLz = 0; int lenStat[8][4]; unsigned char rleValues[32] = {1,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0}; int rleHist[256]; #define OUT_SIZE 65536 unsigned char outBuffer[OUT_SIZE]; int outPointer = 0; int bitMask = 0x80; void TTPackInit(void) { int i; rleValues[0] = 1; for (i=1;i<32;i++) rleValues[i] = 0; for (i=0;i<256;i++) { rleLen[i] = 0; lenValue[i] = 0; rleHist[i] = 0; } for (i=0;i> 8); cth.magic1 = MAGIC_CHAR1; cth.magic2 = MAGIC_CHAR2; cth.compsize_lo = (size + rleUsed + sizeof(PackedHeader)) & 0xff; cth.compsize_hi = (size + rleUsed + sizeof(PackedHeader)) >> 8; cth.esc1 = (escape >> (8-escBits)); cth.notused3 = 0; // just to make sure it has a defined value cth.notused4 = 0; // just to make sure it has a defined value cth.esc2 = escBits; cth.gamma1 = maxGamma + 1; cth.gamma2 = (1 << maxGamma); cth.extralz = extraLZPosBits; cth.notused1 = 0; // just to make sure it has a defined value cth.notused2 = 0; // just to make sure it has a defined value cth.rleentries = rleUsed; for(i=0; i>= 1; if (!bitMask) { bitMask = 0x80; outPointer++; } } //============================================================================= // //============================================================================= void PutValue(int value) { int bits = 0, count = 0; while (value>1) { bits = (bits<<1) | (value & 1); /* is reversed compared to value */ value >>= 1; count++; PutBit(1); } if (count same as value */ bits >>= 1; } } //-------------------------------------------- // why not initializing value lenValue[0] ???? //-------------------------------------------- //============================================================================= // //============================================================================= void InitValueLen() { int i; // could be heavily optimized, but isn't necessary for (i=1; i<256; i++) { int count = 0; if (i<2) count = 0; /* 1 */ else if (i<4) count = 1; /* 2-3 */ else if (i<8) count = 2; /* 4-7 */ else if (i<16) count = 3; /* 8-15 */ else if (i<32) count = 4; /* 16-31 */ else if (i<64) count = 5; /* 32-63 */ else if (i<128) count = 6; /* 64-127 */ else if (i<256) count = 7; /* 128-255 */ lenValue[i] = 2*count; if (count>(8-escBits)), escBits); /* escBits>=0 */ PutBit(0); PutBit(1); PutBit(0); *esc = newesc; PutNBits((*esc>>(8-escBits)), escBits); /* escBits>=0 */ PutNBits(data[0], 8-escBits); gainedEscaped += escBits + 3; timesEscaped++; return 1; } PutNBits(data[0], 8); return 0; } //============================================================================= // //============================================================================= void OutputEof(int *esc) { /* EOF marker */ PutNBits((*esc>>(8-escBits)), escBits); /* escBits>=0 */ PutValue(3); /* >2 */ PutValue((2<>3)); gainedRlecode -= LenValue(32+(data>>3)) + 3; PutNBits(data, 3); lenStat[5][3]++; } //============================================================================= // //============================================================================= void InitRleLen() { int i; for (i=0; i<256; i++) rleLen[i] = LenValue(32 + 0) + 3; for (i=1; i<32; i++) rleLen[rleValues[i]] = LenValue(i); } #define LenRleByte(d) (rleLen[d]) //============================================================================= // //============================================================================= int LenRle(int len, int data) { int out = 0; do { if (len == 1) { out += escBits + 3 + 8; len = 0; } else if (len <= (1<>8)+1) + LenRleByte(data); len -= tmp; } } while (len); return out; } //============================================================================= // //============================================================================= int OutputRle(int *esc, unsigned char *data, int rlelen) { int len = rlelen, tmp; while (len) { if (len >= 2 && len <= (1<>(8-escBits)), escBits); /* escBits>=0 */ PutBit(0); PutBit(1); PutBit(1); PutValue(len-1); PutRleByte(*data); tmp = 8*len -escBits -3 -LenValue(len-1) -LenRleByte(*data); gainedRle += tmp; gainedSRle += tmp; timesRle++; timesSRle++; return 0; } if (len<3) { while (len--) OutputNormal(esc, data, *esc); return 0; } if (len <= maxrlelen) { /* Run-length encoding */ PutNBits((*esc>>(8-escBits)), escBits); /* escBits>=0 */ PutBit(0); PutBit(1); PutBit(1); PutValue((1<>(8-maxGamma))); PutNBits((len-1), 8-maxGamma); PutValue(((len-1)>>8) + 1); PutRleByte(*data); tmp = 8*len -escBits -3 -maxGamma -8 -LenValue(((len-1)>>8)+1) -LenRleByte(*data); gainedRle += tmp; gainedLRle += tmp; timesRle++; timesLRle++; return 0; } /* Run-length encoding */ PutNBits((*esc>>(8-escBits)), escBits); /* escBits>=0 */ PutBit(0); PutBit(1); PutBit(1); PutValue((1<>(8-maxGamma))); PutNBits((maxrlelen-1) & 0xff, 8-maxGamma); PutValue(((maxrlelen-1)>>8)+1); PutRleByte(*data); tmp = 8*maxrlelen -escBits -3 -maxGamma -8 -LenValue(((maxrlelen-1)>>8)+1) -LenRleByte(*data); gainedRle += tmp; gainedLRle += tmp; timesRle++; timesLRle++; len -= maxrlelen; data += maxrlelen; } return 0; } //============================================================================= // //============================================================================= int LenLz(int lzlen, int lzpos) { if (lzlen==2) { if (lzpos <= 256) return escBits + 2 + 8; else return 100000; } return escBits + 8 + extraLZPosBits + LenValue(((lzpos-1)>>(8+extraLZPosBits))+1) + LenValue(lzlen-1); } //============================================================================= // //============================================================================= int OutputLz(int *esc, int lzlen, int lzpos, int curpos) { if (lzlen==2) lenStat[0][1]++; else if (lzlen<=4) lenStat[1][1]++; else if (lzlen<=8) lenStat[2][1]++; else if (lzlen<=16) lenStat[3][1]++; else if (lzlen<=32) lenStat[4][1]++; else if (lzlen<=64) lenStat[5][1]++; else if (lzlen<=128) lenStat[6][1]++; else if (lzlen<=256) lenStat[7][1]++; if (lzlen >= 2 && lzlen <= maxlzlen) { int tmp; PutNBits((*esc>>(8-escBits)), escBits); /* escBits>=0 */ tmp = ((lzpos-1)>>(8+extraLZPosBits))+2; if (tmp==2) lenStat[0][0]++; else if (tmp<=4) lenStat[1][0]++; else if (tmp<=8) lenStat[2][0]++; else if (tmp<=16) lenStat[3][0]++; else if (tmp<=32) lenStat[4][0]++; else if (tmp<=64) lenStat[5][0]++; else if (tmp<=128) lenStat[6][0]++; else if (tmp<=256) lenStat[6][0]++; if (lzlen==2) { PutValue(lzlen-1); PutBit(0); if (lzpos > 256) fprintf(stderr,"ERROR at %d: lzpos too long (%d) for lzlen==2\n",curpos, lzpos); } else { PutValue(lzlen-1); PutValue( ((lzpos-1) >> (8+extraLZPosBits)) +1); PutNBits( ((lzpos-1) >> 8), extraLZPosBits); } PutNBits(((lzpos-1) & 0xff) ^ 0xff, 8); gainedLz += 8*lzlen -LenLz(lzlen, lzpos); timesLz++; return 3; } fprintf(stderr, "ERROR: lzlen too short/long (%d)\n", lzlen); return lzlen; } /* Non-recursive version */ /* NOTE! IMPORTANT! the "length" array length must be inlen+1 */ //============================================================================= // //============================================================================= int OptimizeLength(int optimize) { int i; length[inlen] = 0; /* one off the end, our 'target' */ for (i=inlen-1; i>=0; i--) { int r1 = 8 + length[i+1], r2, r3; if (!lzlen[i] && !rle[i]) { length[i] = r1; mode[i] = LITERAL; continue; } /* If rle>maxlzlen, skip to the start of the rle-maxlzlen.. */ if (rle[i] > maxlzlen && elr[i] > 1) { int z = elr[i]; i -= elr[i]; r2 = LenRle(rle[i], indata[i]) + length[i+ rle[i]]; if (optimize) { int ii, mini = rle[i], minv = r2; int bot = rle[i] - (1<=bot; ii--) { int v = LenRle(ii, indata[i]) + length[i + ii]; if (v < minv) { minv = v; mini = ii; } } if (minv != r2) { lzopt += r2 - minv; rle[i] = mini; r2 = minv; } } length[i] = r2; mode[i] = RLE; for (; z>=0; z--) { length[i+z] = r2; mode[i+z] = RLE; } continue; } r3 = r2 = r1 + 1000; /* r3 >= r2 > r1 */ if (rle[i]) { r2 = LenRle(rle[i], indata[i]) + length[i+ rle[i]]; if (optimize) { int ii, mini = rle[i], minv = r2; /* Check only the original length and all shorter lengths that are power of two. Does not really miss many 'minimums' this way, at least not globally.. Makes the assumption that the Elias Gamma Code is used, i.e. values of the form 2^n are 'optimal' */ ii = 2; while (rle[i] > ii) { int v = LenRle(ii, indata[i]) + length[i + ii]; if (v < minv) { minv = v; mini = ii; } ii <<= 1; } if (minv != r2) { lzopt += r2 - minv; rle[i] = mini; r2 = minv; } } } if (lzlen[i]) { r3 = LenLz(lzlen[i], lzpos[i]) + length[i + lzlen[i]]; if (optimize && lzlen[i]>2) { int ii, mini = lzlen[i], minv = r3; int topLen = LenLz(lzlen[i], lzpos[i]) - LenValue(lzlen[i]-1); /* Check only the original length and all shorter lengths that are power of two. Does not really miss many 'minimums' this way, at least not globally.. Makes the assumption that the Elias Gamma Code is used, i.e. values of the form 2^n are 'optimal' */ ii = 4; while (lzlen[i] > ii) { int v = topLen + LenValue(ii-1) + length[i + ii]; if (v < minv) { minv = v; mini = ii; } ii <<= 1; } /* Note: 2-byte optimization checks are no longer done with the rest, because the equation gives too long code lengths for 2-byte matches if extraLzPosBits>0. */ /* Two-byte rescan/check */ if (backSkip[i] && backSkip[i] <= 256) { /* There are previous occurrances (near enough) */ int v = LenLz(2, (int)backSkip[i]) + length[i + 2]; if (v < minv) { minv = v; mini = 2; lzlen[i] = mini; r3 = minv; lzpos[i] = (int)backSkip[i]; } } if (minv != r3 && minv < r2) { lzopt += r3 - minv; lzlen[i] = mini; r3 = minv; } } } if (r2 <= r1) { if (r2 <= r3) { length[i] = r2; mode[i] = RLE; } else { length[i] = r3; mode[i] = LZ77; } } else { if (r3 <= r1) { length[i] = r3; mode[i] = LZ77; } else { length[i] = r1; mode[i] = LITERAL; } } } return length[0]; } /* The algorithm in the OptimizeEscape() works as follows: 1) Only unpacked bytes are processed, they are marked with MMARK. We proceed from the end to the beginning. Variable A (old/new length) is updated. 2) At each unpacked byte, one and only one possible escape matches. A new escape code must be selected for this case. The optimal selection is the one which provides the shortest number of escapes to the end of the file, i.e. A[esc] = 1+min(A[0], A[1], .. A[states-1]). For other states A[esc] = A[esc]; If we change escape in this byte, the new escape is the one with the smallest value in A. 3) The starting escape is selected from the possibilities and mode 0 is restored to all mode 3 locations. */ //============================================================================= // //============================================================================= int OptimizeEscape(int *startEscape, int *nonNormal) { int i, j, states = (1<256) { fprintf(stderr, "Escape optimize: only 256 states (%d)!\n",states); return 0; } /* Mark those bytes that are actually outputted */ for (i=0; i=0; i--) { /* Using a table to skip non-normal bytes does not help.. */ if (mode[i] == MMARK) { int k = (indata[i] >> esc8); /* Change the tag values back to normal */ mode[i] = LITERAL; /* k are the matching bytes, minv is the minimum value, minp is the minimum index */ newesc[i] = (minp << esc8); a[k] = minv + 1; b[k] = b[minp] + 1; if (k==minp) { /* Minimum changed -> need to find a new minimum */ /* a[k] may still be the minimum */ minv++; for (k=states-1; k>=0; k--) { if (a[k] < minv) { minv = a[k]; minp = k; /* There may be others, but the first one that is smaller than the old minimum is equal to any other new minimum. */ break; } } } } } /* Select the best value for the initial escape */ if (startEscape) { i = inlen; /* make it big enough */ for (j=states-1; j>=0; j--) { if (a[j] <= i) { *startEscape = (j << esc8); i = a[j]; } } } if (nonNormal) *nonNormal = other; return b[startEscape ? (*startEscape>>esc8) : 0]; } //============================================================================= // Initialize the RLE byte code table according to all RLE's found so far O(n) //============================================================================= void InitRle(void) { int p, mr, mv, i; for (i=1; i<32; i++) { mr = -1; mv = 0; for (p=0; p<256; p++) { if (rleHist[p] > mv) { mv = rleHist[p]; mr = p; } } if (mv>0) { rleValues[i] = mr; rleHist[mr] = -1; } else break; } InitRleLen(); } //============================================================================= // Initialize the RLE byte code table according to RLE's actually used O(n) //============================================================================= void OptimizeRle(int flags) { int p, mr, mv, i; if (flags & F_STATS) fprintf(VERBOSE_OUT, "RLE Byte Code Re-Tune, RLE Ranks:\n"); for (p=0; p<256; p++) rleHist[p] = 0; for (p=0; p mv) { mv = rleHist[p]; mr = p; } } if (mv>0) { rleValues[i] = mr; if (flags & F_STATS) { fprintf(VERBOSE_OUT, " %2d.0x%02x %-3d ", i, mr, mv); if (!((i - 1) % 6)) fprintf(VERBOSE_OUT, "\n"); } rleHist[mr] = -1; } else { break; } } rleUsed = i-1; if (flags & F_STATS) if (((i - 1) % 6)!=1) fprintf(VERBOSE_OUT, "\n"); InitRleLen(); } //============================================================================= // //============================================================================= int PackLz77(int lzsz, int flags, int *startEscape) { int i, j, outlen, p, headerSize; int escape; unsigned char *hashValue; unsigned char *a; int k; unsigned short *lastPair; unsigned short err_occured = 0; int rescan = 0; if (lzsz < 0 || lzsz > lrange) { fprintf(stderr, "LZ range must be from 0 to %d (was %d). Set to %d.\n", lrange, lzsz, lrange); lzsz = lrange; } if (lzsz > 65535) { fprintf(stderr, "LZ range must be from 0 to 65535 (was %d). Set to 65535.\n", lzsz); lzsz = 65535; } if (!lzsz) fprintf(stderr, "Warning: zero LZ range. Only RLE packing used.\n"); InitRleLen(); length = (int *)calloc(sizeof(int), inlen + 1); mode = (unsigned char *)calloc(sizeof(unsigned char), inlen); rle = (unsigned short *)calloc(sizeof(unsigned short), inlen); elr = (unsigned short *)calloc(sizeof(unsigned short), inlen); lzlen = (unsigned short *)calloc(sizeof(unsigned short), inlen); lzpos = (unsigned short *)calloc(sizeof(unsigned short), inlen); newesc = (unsigned char *)calloc(sizeof(unsigned char), inlen); backSkip = (unsigned short *)calloc(sizeof(unsigned short), inlen); hashValue = (unsigned char *)malloc(inlen); lastPair = (unsigned short *)calloc(sizeof(unsigned short), 256*256); /* error checking */ if (!length || !mode || !rle || !elr || !lzlen || !lzpos || !newesc || !lastPair || !backSkip || !hashValue) { fprintf(stderr, "ERROR: Memory allocation failed!\n"); err_occured = 1; goto errorexit; } i = 0; j = 0; a = indata + inlen; for (p=inlen-1; p>=0; p--) { k = j; j = i; i = *--a; /* Only one read per position */ hashValue[p] = i*3 + j*5 + k*7; /* 7.95 % */ } /* Detect all RLE and LZ77 jump possibilities */ for (p=0; p=2) { rleHist[indata[p]]++; for (i=rlelen-1; i>=0; i--) { rle[p+i] = rlelen-i; elr[p+i] = i; /* For RLE backward skipping */ } } } /* check LZ77 code */ if (p+rle[p]+1=0 && i>=bot) { /* Got a 2-byte match at least */ maxval = 2; maxpos = p-i; /* A..AB rlep # of A's, B is something else.. Search for bytes that are in p + (rlep-1), i.e. the last rle byte ('A') and the non-matching one ('B'). When found, check if the rle in the compare position (i) is long enough (i.e. the same number of A's at p and i-rlep+1). There are dramatically less matches for AB than for AA, so we get a huge speedup with this approach. We are still guaranteed to find the most recent longest match there is. */ i = (int)lastPair[(indata[p+(rlep-1)]<<8) | indata[p+rlep]] -1; while (i>=bot /* && i>=rlep-1 */) { /* bot>=rlep-1, i>=bot ==> i>=rlep-1 */ /* Equal number of A's ? */ if (!(rlep-1) || rle[i-(rlep-1)]==rlep) { /* 'head' matches */ /* rlep==1 ==> (rlep-1)==0 */ /* ivanova.run: 443517 rlep==1, 709846 rle[i+1-rlep]==rlep */ /* Check the hash values corresponding to the last two bytes of the currently longest match and the first new matching(?) byte. If the hash values don't match, don't bother to check the data itself. */ if ( hashValue[i+maxval-rlep-1] == hashCompare ) { unsigned char *a = indata + i+2; /* match */ unsigned char *b = indata + p+rlep-1+2;/* curpos */ int topindex = inlen-(p+rlep-1); /* the 2 first bytes ARE the same.. */ j = 2; while (j < topindex && *a++==*b++) j++; if (j + rlep-1 > maxval) { int tmplen = j+rlep-1, tmppos = p-i+rlep-1; if (tmplen > maxlzlen) tmplen = maxlzlen; /* Accept only versions that really are shorter */ if (tmplen*8 - LenLz(tmplen, tmppos) > maxval*8 - LenLz(maxval, maxpos)) { maxval = tmplen; maxpos = tmppos; hashCompare = hashValue[p+maxval-2]; } if (maxval == maxlzlen) break; } } } if (!backSkip[i]) break; /* No previous occurrances (near enough) */ i -= (int)backSkip[i]; } /* If there is 'A' in the previous position also, RLE-like LZ77 is possible, although rarely shorter than real RLE. */ if (p && rle[p-1] > maxval) { maxval = rle[p-1] - 1; maxpos = 1; } /* Last, try to find as long as possible match for the RLE part only. */ if (maxval < maxlzlen && rlep > maxval) { bot = p - lzsz; if (bot < 0) bot = 0; /* Note: indata[p] == indata[p+1] */ i = (int)lastPair[indata[p]*257] -1; while (/* i>= rlep-2 &&*/ i>=bot) { if (elr[i] + 2 > maxval) { maxval = min(elr[i] + 2, rlep); maxpos = p - i + (maxval-2); if(maxval == rlep) break; /* Got enough */ } i -= elr[i]; if (!backSkip[i]) break; /* No previous occurrances (near enough) */ i -= (int)backSkip[i]; } } if (p+maxval > inlen) { fprintf(stderr,"Error @ %d, lzlen %d, pos %d - exceeds inlen\n",p, maxval, maxpos); maxval = inlen - p; } if (maxpos<=256 || maxval > 2) { if (maxpos < 0) fprintf(stderr, "Error @ %d, lzlen %d, pos %d\n",p, maxval, maxpos); lzlen[p] = (maxval p || ptr > 0xffff) ptr = 0; backSkip[p] = ptr; lastPair[index] = p+1; } } if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT, "\rChecked: %d \n", p); fflush(VERBOSE_OUT); } /* Initialize the RLE selections */ InitRle(); /* Check the normal bytes / all ratio */ { int mb, mv; if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT, "Selecting the number of escape bits.. "); fflush(VERBOSE_OUT); } /* Absolute maximum number of escaped bytes with the escape optimize is 2^-n, where n is the number of escape bits used. This worst case happens only on equal- distributed normal bytes (01230123..). This is why the typical values are so much smaller. */ mb = 0; mv = 8*OUT_SIZE; for (escBits=1; escBits<9; escBits++) { int escaped, other = 0, c; escMask = (0xff00>>escBits) & 0xff; /* Find the optimum path for selected escape bits (no optimize) */ OptimizeLength(0); /* Optimize the escape selections for this path & escBits */ escaped = OptimizeEscape(&escape, &other); /* Compare value: bits lost for escaping -- bits lost for prefix */ c = (escBits+3)*escaped + other*escBits; if (flags & F_STATS) { fprintf(VERBOSE_OUT, " %d:%d", escBits, c); fflush(VERBOSE_OUT); /* for SAS/C */ } if (c < mv) { mb = escBits; mv = c; } else { /* minimum found */ break; } if (escBits==4 && (flags & F_STATS)) fprintf(VERBOSE_OUT, "\n"); } if (mb==1) { /* Minimum was 1, check 0 */ int escaped; escBits = 0; escMask = 0; /* Find the optimum path for selected escape bits (no optimize) */ OptimizeLength(0); /* Optimize the escape selections for this path & escBits */ escaped = OptimizeEscape(&escape, NULL); if ((flags & F_STATS)) { fprintf(VERBOSE_OUT, " %d:%d", escBits, 3*escaped); fflush(VERBOSE_OUT); /* for SAS/C */ } if (3*escaped < mv) { mb = 0; /* mv = 3*escaped; */ } } if ((flags & F_STATS)) fprintf(VERBOSE_OUT, "\n"); if (flags & F_VERBOSE) fprintf(VERBOSE_OUT, "Selected %d-bit escapes\n", mb); escBits = mb; escMask = (0xff00>>escBits) & 0xff; } if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT, "Optimizing LZ77 and RLE lengths..."); fflush(VERBOSE_OUT); } /* Find the optimum path (optimize) */ OptimizeLength(1); if (flags & F_STATS) { fprintf(VERBOSE_OUT, " gained %d units.\n", lzopt/8); } else { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT, "\n"); } { long lzstat[5] = {0,0,0,0,0}, i, cur = 0, old = extraLZPosBits; if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT, "Selecting LZPOS LO length.. "); fflush(VERBOSE_OUT); } for (p=0; p> 8)+1 > (1< (1< (1<<(maxGamma-1))) { if (rle[p] <= (1< 10) { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"Note: Using option -m%d you may get better results.\n",maxGamma+1); } if (maxGamma > 5 && stat[0] + stat[1] + stat[3] < 4) { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"Note: Using option -m%d you may get better results.\n",maxGamma-1); } } /* Optimize the escape selections */ OptimizeEscape(&escape, NULL); if (startEscape) *startEscape = escape; OptimizeRle(flags); /* Retune the RLE selections */ if (flags & F_VERBOSE) { int oldEscape = escape; if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"normal RLE LZLEN LZPOS(absolute)\n\n"); for (p=0; p"); else fprintf(VERBOSE_OUT," "); } if (j==p) { if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT,"*001* %03d %03d %04x(%04x) %02x %s %02x", rle[p], lzlen[p], lzpos[p], p-lzpos[p], indata[p], (mode[p] & MMARK)?"#":" ", newesc[p]); } if ((indata[p] & escMask) == escape) { escape = newesc[p]; if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"<<"); } if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"\n"); j += 1; } else { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"*001* %03d %03d %04x(%04x) %02x %s %02x\n", rle[p], lzlen[p], lzpos[p], p-lzpos[p], indata[p], (mode[p] & MMARK)?"#":" ", newesc[p]); } break; case MMARK | LZ77: case LZ77: if (j==p) { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,">"); j += lzlen[p]; } else if (flags & F_VERBOSE) fprintf(VERBOSE_OUT," "); if (flags & F_VERBOSE) fprintf(VERBOSE_OUT," 001 %03d *%03d* %04x(%04x) %02x %s\n", rle[p], lzlen[p], lzpos[p], p-lzpos[p], indata[p], (mode[p] & MMARK)?"#":" "); break; case MMARK | RLE: case RLE: if (j==p) { if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,">"); j += rle[p]; } else if (flags & F_VERBOSE) fprintf(VERBOSE_OUT," "); if (flags & F_VERBOSE) fprintf(VERBOSE_OUT," 001 *%03d* %03d %04x(%04x) %02x %s\n", rle[p], lzlen[p], lzpos[p], p-lzpos[p], indata[p], (mode[p] & MMARK)?"#":" "); break; default: j++; break; } mode[p] &= ~MMARK; } escape = oldEscape; } for (p=0; p 2 && lzlen[p] > rle[p]) { int bot = p - lzpos[p] + 1, i; unsigned short rlep = rle[p]; if (!rlep) rlep = 1; if (bot < 0) bot = 0; bot += (rlep-1); i = p - (int)backSkip[p]; while (i>=bot /* && i>=rlep-1 */) { /* Equal number of A's ? */ if (rlep==1 || rle[i-rlep+1]==rlep) { /* 'head' matches */ unsigned char *a = indata + i+1; /* match */ unsigned char *b = indata + p+rlep-1+1; /* curpos */ int topindex = inlen-(p+rlep-1); j = 1; while (j < topindex && *a++==*b++) j++; if (j + rlep-1 >= lzlen[p]) { int tmppos = p-i+rlep-1; rescan += LenLz(lzlen[p], lzpos[p]) - LenLz(lzlen[p], tmppos); lzpos[p] = tmppos; break; } } if (!backSkip[i]) break; /* No previous occurrances (near enough) */ i -= (int)backSkip[i]; } } for (i=0; i 0xffff) { fprintf(stderr,"ERROR: File is too large to handle (>64936 Bytes)"); if (indata) free(indata); return 1; } if (flags & F_VERBOSE) { fprintf(VERBOSE_OUT, "Load address 0x%04x=%d, Last byte 0x%04x=%d\n", startAddr, startAddr, startAddr+inlen-1, startAddr+inlen-1); fprintf(VERBOSE_OUT, "New load address 0x%04x=%d\n", memStart, memStart); } n = PackLz77(lzlen, flags, &startEscape); if (!n) { int endAddr = startAddr + inlen; /* end for uncompressed data */ if (endAddr - ((outPointer + 255) & ~255) < memStart + 3) { /* would overwrite the decompressor, move a bit upwards */ if (flags & F_VERBOSE) fprintf(VERBOSE_OUT,"$%x < $%x, decompressor overwrite possible, moving upwards\n", endAddr - ((outPointer + 255) & ~255), memStart + 3); endAddr = memStart + 3 + ((outPointer + 255) & ~255); } /* 3 bytes reserved for EOF */ /* bytes reserved for temporary data expansion (escaped chars) */ endAddr += 3 + reservedBytes; // outBuffer ... static global array (65536 Bytes) SavePack(outBuffer, outPointer, out_file, startEscape, rleValues, extraLZPosBits); timeused = clock()-timeused; if (!timeused) timeused++; if (flags & F_VERBOSE) fprintf(VERBOSE_OUT, "Compressed %d bytes in %4.2f seconds (%4.2f kB/sec)\n", inlen, (double)timeused/CLOCKS_PER_SEC, (double)CLOCKS_PER_SEC*inlen/timeused/1024.0); return 0; } if (indata) free(indata); return n; }