935 lines
37 KiB
C
935 lines
37 KiB
C
#include <stdio.h>
|
|
|
|
typedef struct
|
|
{
|
|
int off;
|
|
int len;
|
|
int back;
|
|
} LZ5HC_match_t;
|
|
|
|
typedef struct
|
|
{
|
|
int price;
|
|
int off;
|
|
int mlen;
|
|
int litlen;
|
|
int rep;
|
|
} LZ5HC_optimal_t;
|
|
|
|
#if 1
|
|
#define LZ5_LOG_PARSER(fmt, args...) ;// printf(fmt, ##args)
|
|
#define LZ5_LOG_PRICE(fmt, args...) ;//printf(fmt, ##args)
|
|
#define LZ5_LOG_ENCODE(fmt, args...) ;//printf(fmt, ##args)
|
|
#else
|
|
#define LZ5_LOG_PARSER(fmt, args...) printf(fmt, ##args)
|
|
#define LZ5_LOG_PRICE(fmt, args...) printf(fmt, ##args)
|
|
#define LZ5_LOG_ENCODE(fmt, args...) printf(fmt, ##args)
|
|
#endif
|
|
|
|
#define LZ5_OPT_NUM (1<<12)
|
|
|
|
|
|
/*
|
|
I assume that you are using 4 entropy-coder tables:
|
|
litLength = FSE_decodeSymbol(&(seqState->stateLL), &(seqState->DStream));
|
|
offsetCode = FSE_decodeSymbol(&(seqState->stateOffb), &(seqState->DStream));
|
|
matchLength = FSE_decodeSymbol(&(seqState->stateML), &(seqState->DStream));
|
|
HUF_decompress(dst, litSize, ip+5, litCSize)
|
|
|
|
With the following max values:
|
|
#define MLbits 7
|
|
#define LLbits 6
|
|
#define Offbits 5
|
|
#define MaxML ((1<<MLbits) - 1)
|
|
#define MaxLL ((1<<LLbits) - 1)
|
|
#define MaxOff ((1<<Offbits)- 1)
|
|
When value >= maxValue then use additional 1 byte (if value <
|
|
maxValue+255) or 3 bytes (to encode value up to MaxValue+255+65535).
|
|
|
|
For offsets only the numbers of bits == log2(offset) are
|
|
entropy encoded and the offset is put into a binary stream.
|
|
*/
|
|
|
|
#define LZ5_LIT_ONLY_COST(len) (((len)<<3)+1+0)
|
|
|
|
#define LZ5_LIT_COST(len) (((len)<<3)+0)
|
|
|
|
FORCE_INLINE U32 LZ5HC_get_price(U32 litlen, U32 offset, U32 mlen)
|
|
{
|
|
size_t lit_cost = (litlen<<3)+0;
|
|
size_t match_cost = /*MLbits +*/ ZSTD_highbit((U32)mlen+1) + Offbits + ZSTD_highbit((U32)offset+1);
|
|
return lit_cost + match_cost;
|
|
}
|
|
|
|
|
|
#define SET_PRICE(pos, mlen, offset, litlen, price) \
|
|
{ \
|
|
while (last_pos < pos) { opt[last_pos+1].price = 1<<30; last_pos++; } \
|
|
opt[pos].mlen = mlen; \
|
|
opt[pos].off = offset; \
|
|
opt[pos].litlen = litlen; \
|
|
opt[pos].price = price; \
|
|
LZ5_LOG_PARSER("%d: SET price[%d/%d]=%d litlen=%d len=%d off=%d\n", (int)(inr-base), (int)pos, (int)last_pos, opt[pos].price, opt[pos].litlen, opt[pos].mlen, opt[pos].off); \
|
|
if (mlen > 1 && mlen < MINMATCH) { printf("%d: ERROR SET price[%d/%d]=%d litlen=%d len=%d off=%d\n", (int)(inr-base), (int)pos, (int)last_pos, opt[pos].price, opt[pos].litlen, opt[pos].mlen, opt[pos].off); exit(0); }; \
|
|
}
|
|
|
|
|
|
|
|
FORCE_INLINE /* inlining is important to hardwire a hot branch (template emulation) */
|
|
size_t ZSTD_insertBtAndGetAllMatches (
|
|
ZSTD_CCtx* zc,
|
|
const BYTE* const ip, const BYTE* const iend,
|
|
U32 nbCompares, const U32 mls,
|
|
U32 extDict, LZ5HC_match_t* matches)
|
|
{
|
|
U32* const hashTable = zc->hashTable;
|
|
const U32 hashLog = zc->params.hashLog;
|
|
const size_t h = ZSTD_hashPtr(ip, hashLog, mls);
|
|
U32* const bt = zc->contentTable;
|
|
const U32 btLog = zc->params.contentLog - 1;
|
|
const U32 btMask= (1 << btLog) - 1;
|
|
U32 matchIndex = hashTable[h];
|
|
size_t commonLengthSmaller=0, commonLengthLarger=0;
|
|
const BYTE* const base = zc->base;
|
|
const BYTE* const dictBase = zc->dictBase;
|
|
const U32 dictLimit = zc->dictLimit;
|
|
const BYTE* const dictEnd = dictBase + dictLimit;
|
|
const BYTE* const prefixStart = base + dictLimit;
|
|
const U32 current = (U32)(ip-base);
|
|
const U32 btLow = btMask >= current ? 0 : current - btMask;
|
|
const U32 windowLow = zc->lowLimit;
|
|
U32* smallerPtr = bt + 2*(current&btMask);
|
|
U32* largerPtr = bt + 2*(current&btMask) + 1;
|
|
size_t bestLength = 0;
|
|
U32 matchEndIdx = current+8;
|
|
U32 dummy32; /* to be nullified at the end */
|
|
size_t mnum = 0;
|
|
|
|
hashTable[h] = current; /* Update Hash Table */
|
|
|
|
while (nbCompares-- && (matchIndex > windowLow)) {
|
|
U32* nextPtr = bt + 2*(matchIndex & btMask);
|
|
size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger); /* guaranteed minimum nb of common bytes */
|
|
const BYTE* match;
|
|
|
|
if ((!extDict) || (matchIndex+matchLength >= dictLimit)) {
|
|
match = base + matchIndex;
|
|
if (match[matchLength] == ip[matchLength])
|
|
matchLength += ZSTD_count(ip+matchLength+1, match+matchLength+1, iend) +1;
|
|
} else {
|
|
match = dictBase + matchIndex;
|
|
matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iend, dictEnd, prefixStart);
|
|
if (matchIndex+matchLength >= dictLimit)
|
|
match = base + matchIndex; /* to prepare for next usage of match[matchLength] */
|
|
}
|
|
|
|
if (matchLength > bestLength) {
|
|
if (matchLength > matchEndIdx - matchIndex)
|
|
matchEndIdx = matchIndex + (U32)matchLength;
|
|
#if 0
|
|
if ( (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit(current-matchIndex+1) - ZSTD_highbit((U32)offsetPtr[0]+1)) )
|
|
bestLength = matchLength, *offsetPtr = current - matchIndex;
|
|
#else
|
|
if (mnum == 0 || (4*(int)(matchLength-bestLength)) > (int)(ZSTD_highbit(current-matchIndex+1) - ZSTD_highbit((U32)matches[mnum-1].off+1)) )
|
|
{
|
|
if (matchLength >= MINMATCH)
|
|
{
|
|
bestLength = matchLength;
|
|
matches[mnum].off = current - matchIndex;
|
|
matches[mnum].len = matchLength;
|
|
matches[mnum].back = 0;
|
|
mnum++;
|
|
}
|
|
if (matchLength > LZ5_OPT_NUM) break;
|
|
}
|
|
#endif
|
|
if (ip+matchLength == iend) /* equal : no way to know if inf or sup */
|
|
break; /* drop, to guarantee consistency (miss a little bit of compression) */
|
|
}
|
|
|
|
if (match[matchLength] < ip[matchLength]) {
|
|
/* match is smaller than current */
|
|
*smallerPtr = matchIndex; /* update smaller idx */
|
|
commonLengthSmaller = matchLength; /* all smaller will now have at least this guaranteed common length */
|
|
if (matchIndex <= btLow) { smallerPtr=&dummy32; break; } /* beyond tree size, stop the search */
|
|
smallerPtr = nextPtr+1; /* new "smaller" => larger of match */
|
|
matchIndex = nextPtr[1]; /* new matchIndex larger than previous (closer to current) */
|
|
} else {
|
|
/* match is larger than current */
|
|
*largerPtr = matchIndex;
|
|
commonLengthLarger = matchLength;
|
|
if (matchIndex <= btLow) { largerPtr=&dummy32; break; } /* beyond tree size, stop the search */
|
|
largerPtr = nextPtr;
|
|
matchIndex = nextPtr[0];
|
|
}
|
|
}
|
|
|
|
*smallerPtr = *largerPtr = 0;
|
|
|
|
zc->nextToUpdate = (matchEndIdx > current + 8) ? matchEndIdx - 8 : current+1;
|
|
return mnum;
|
|
}
|
|
|
|
|
|
/** Tree updater, providing best match */
|
|
FORCE_INLINE /* inlining is important to hardwire a hot branch (template emulation) */
|
|
size_t ZSTD_BtGetAllMatches (
|
|
ZSTD_CCtx* zc,
|
|
const BYTE* const ip, const BYTE* const iLimit,
|
|
const U32 maxNbAttempts, const U32 mls, LZ5HC_match_t* matches)
|
|
{
|
|
if (ip < zc->base + zc->nextToUpdate) return 0; /* skipped area */
|
|
ZSTD_updateTree(zc, ip, iLimit, maxNbAttempts, mls);
|
|
return ZSTD_insertBtAndGetAllMatches(zc, ip, iLimit, maxNbAttempts, mls, 0, matches);
|
|
}
|
|
|
|
|
|
FORCE_INLINE size_t ZSTD_BtGetAllMatches_selectMLS (
|
|
ZSTD_CCtx* zc, /* Index table will be updated */
|
|
const BYTE* ip, const BYTE* const iLimit,
|
|
const U32 maxNbAttempts, const U32 matchLengthSearch, LZ5HC_match_t* matches)
|
|
{
|
|
switch(matchLengthSearch)
|
|
{
|
|
default :
|
|
case 4 : return ZSTD_BtGetAllMatches(zc, ip, iLimit, maxNbAttempts, 4, matches);
|
|
case 5 : return ZSTD_BtGetAllMatches(zc, ip, iLimit, maxNbAttempts, 5, matches);
|
|
case 6 : return ZSTD_BtGetAllMatches(zc, ip, iLimit, maxNbAttempts, 6, matches);
|
|
}
|
|
}
|
|
|
|
|
|
FORCE_INLINE /* inlining is important to hardwire a hot branch (template emulation) */
|
|
size_t ZSTD_HcGetAllMatches_generic (
|
|
ZSTD_CCtx* zc, /* Index table will be updated */
|
|
const BYTE* const ip, const BYTE* const iLimit,
|
|
const U32 maxNbAttempts, const U32 mls, const U32 extDict, LZ5HC_match_t* matches)
|
|
{
|
|
U32* const chainTable = zc->contentTable;
|
|
const U32 chainSize = (1 << zc->params.contentLog);
|
|
const U32 chainMask = chainSize-1;
|
|
const BYTE* const base = zc->base;
|
|
const BYTE* const dictBase = zc->dictBase;
|
|
const U32 dictLimit = zc->dictLimit;
|
|
const BYTE* const prefixStart = base + dictLimit;
|
|
const BYTE* const dictEnd = dictBase + dictLimit;
|
|
const U32 lowLimit = zc->lowLimit;
|
|
const U32 current = (U32)(ip-base);
|
|
const U32 minChain = current > chainSize ? current - chainSize : 0;
|
|
U32 matchIndex;
|
|
const BYTE* match;
|
|
int nbAttempts=maxNbAttempts;
|
|
size_t ml=MINMATCH-1;
|
|
size_t mnum = 0;
|
|
|
|
/* HC4 match finder */
|
|
matchIndex = ZSTD_insertAndFindFirstIndex (zc, ip, mls);
|
|
|
|
while ((matchIndex>lowLimit) && (nbAttempts)) {
|
|
size_t currentMl=0;
|
|
nbAttempts--;
|
|
if ((!extDict) || matchIndex >= dictLimit) {
|
|
match = base + matchIndex;
|
|
if (match[ml] == ip[ml]) /* potentially better */
|
|
currentMl = ZSTD_count(ip, match, iLimit);
|
|
} else {
|
|
match = dictBase + matchIndex;
|
|
if (MEM_read32(match) == MEM_read32(ip)) /* assumption : matchIndex <= dictLimit-4 (by table construction) */
|
|
currentMl = ZSTD_count_2segments(ip+MINMATCH, match+MINMATCH, iLimit, dictEnd, prefixStart) + MINMATCH;
|
|
}
|
|
|
|
/* save best solution */
|
|
if (currentMl > ml) {
|
|
ml = currentMl;
|
|
matches[mnum].off = current - matchIndex;
|
|
matches[mnum].len = currentMl;
|
|
matches[mnum].back = 0;
|
|
mnum++;
|
|
if (currentMl > LZ5_OPT_NUM) break;
|
|
if (ip+currentMl == iLimit) break; /* best possible, and avoid read overflow*/
|
|
}
|
|
|
|
if (matchIndex <= minChain) break;
|
|
matchIndex = NEXT_IN_CHAIN(matchIndex, chainMask);
|
|
}
|
|
|
|
return mnum;
|
|
}
|
|
|
|
|
|
FORCE_INLINE size_t ZSTD_HcGetAllMatches_selectMLS (
|
|
ZSTD_CCtx* zc,
|
|
const BYTE* ip, const BYTE* const iLimit,
|
|
const U32 maxNbAttempts, const U32 matchLengthSearch, LZ5HC_match_t* matches)
|
|
{
|
|
switch(matchLengthSearch)
|
|
{
|
|
default :
|
|
case 4 : return ZSTD_HcGetAllMatches_generic(zc, ip, iLimit, maxNbAttempts, 4, 0, matches);
|
|
case 5 : return ZSTD_HcGetAllMatches_generic(zc, ip, iLimit, maxNbAttempts, 5, 0, matches);
|
|
case 6 : return ZSTD_HcGetAllMatches_generic(zc, ip, iLimit, maxNbAttempts, 6, 0, matches);
|
|
}
|
|
}
|
|
|
|
|
|
void print_hex_text(uint8_t* buf, int bufsize, int endline)
|
|
{
|
|
int i, j;
|
|
for (i=0; i<bufsize; i+=16)
|
|
{
|
|
printf("%02d:", i);
|
|
for (j=0; j<16; j++)
|
|
if (i+j<bufsize)
|
|
printf("%02x,",buf[i+j]);
|
|
else
|
|
printf(" ");
|
|
printf(" ");
|
|
for (j=0; i+j<bufsize && j<16; j++)
|
|
printf("%c",buf[i+j]>32?buf[i+j]:'.');
|
|
printf("\n");
|
|
}
|
|
if (endline) printf("\n");
|
|
}
|
|
|
|
|
|
/* *******************************
|
|
* Optimal parser OLD
|
|
*********************************/
|
|
FORCE_INLINE
|
|
void ZSTD_compressBlock_opt2_generic(ZSTD_CCtx* ctx,
|
|
const void* src, size_t srcSize,
|
|
const U32 searchMethod, const U32 depth)
|
|
{
|
|
seqStore_t* seqStorePtr = &(ctx->seqStore);
|
|
const BYTE* const istart = (const BYTE*)src;
|
|
const BYTE* ip = istart;
|
|
const BYTE* anchor = istart;
|
|
const BYTE* const iend = istart + srcSize;
|
|
const BYTE* const ilimit = iend - 8;
|
|
const BYTE* const base = ctx->base + ctx->dictLimit;
|
|
|
|
size_t offset_2=REPCODE_STARTVALUE, offset_1=REPCODE_STARTVALUE;
|
|
const U32 maxSearches = 1 << ctx->params.searchLog;
|
|
const U32 mls = ctx->params.searchLength;
|
|
|
|
typedef size_t (*searchMax_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
|
|
size_t* offsetPtr,
|
|
U32 maxNbAttempts, U32 matchLengthSearch);
|
|
searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS;
|
|
|
|
#if 0
|
|
typedef size_t (*getAllMatches_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
|
|
U32 maxNbAttempts, U32 matchLengthSearch, LZ5HC_match_t* matches);
|
|
getAllMatches_f getAllMatches = searchMethod ? ZSTD_BtGetAllMatches_selectMLS : ZSTD_HcGetAllMatches_selectMLS;
|
|
|
|
LZ5HC_match_t matches[LZ5_OPT_NUM+1];
|
|
#endif
|
|
|
|
/* init */
|
|
ZSTD_resetSeqStore(seqStorePtr);
|
|
if ((ip-base) < REPCODE_STARTVALUE) ip = base + REPCODE_STARTVALUE;
|
|
|
|
/* Match Loop */
|
|
while (ip < ilimit) {
|
|
size_t matchLength=0;
|
|
size_t offset=0;
|
|
const BYTE* start=ip+1;
|
|
|
|
#define ZSTD_USE_REP
|
|
#ifdef ZSTD_USE_REP
|
|
/* check repCode */
|
|
if (MEM_read32(start) == MEM_read32(start - offset_1)) {
|
|
/* repcode : we take it */
|
|
matchLength = ZSTD_count(start+MINMATCH, start+MINMATCH-offset_1, iend) + MINMATCH;
|
|
if (depth==0) goto _storeSequence;
|
|
}
|
|
#endif
|
|
|
|
{
|
|
/* first search (depth 0) */
|
|
#if 1
|
|
size_t offsetFound = 99999999;
|
|
size_t ml2 = searchMax(ctx, ip, iend, &offsetFound, maxSearches, mls);
|
|
if (ml2 > matchLength)
|
|
start=ip, matchLength = ml2, offset=offsetFound;
|
|
#else
|
|
size_t mnum = getAllMatches(ctx, ip, iend, maxSearches, mls, matches);
|
|
if (mnum > 0) {
|
|
if (matches[mnum-1].len > matchLength)
|
|
start=ip, matchLength = matches[mnum-1].len, offset=matches[mnum-1].off;
|
|
}
|
|
#endif
|
|
}
|
|
|
|
if (matchLength < MINMATCH) {
|
|
// ip += ((ip-anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
|
|
ip++;
|
|
continue;
|
|
}
|
|
|
|
#if 1
|
|
/* let's try to find a better solution */
|
|
if (depth>=1)
|
|
while (ip<ilimit) {
|
|
ip ++;
|
|
#ifdef ZSTD_USE_REP
|
|
if ((offset) && (MEM_read32(ip) == MEM_read32(ip - offset_1))) {
|
|
size_t mlRep = ZSTD_count(ip+MINMATCH, ip+MINMATCH-offset_1, iend) + MINMATCH;
|
|
int gain2 = (int)(mlRep * 3);
|
|
int gain1 = (int)(matchLength*3 - ZSTD_highbit((U32)offset+1) + 1);
|
|
if ((mlRep >= MINMATCH) && (gain2 > gain1))
|
|
matchLength = mlRep, offset = 0, start = ip;
|
|
}
|
|
#endif
|
|
{
|
|
size_t offset2=999999;
|
|
size_t ml2 = searchMax(ctx, ip, iend, &offset2, maxSearches, mls);
|
|
int gain2 = (int)(ml2*4 - ZSTD_highbit((U32)offset2+1)); /* raw approx */
|
|
int gain1 = (int)(matchLength*4 - ZSTD_highbit((U32)offset+1) + 4);
|
|
if ((ml2 >= MINMATCH) && (gain2 > gain1)) {
|
|
matchLength = ml2, offset = offset2, start = ip;
|
|
continue; /* search a better one */
|
|
} }
|
|
|
|
break; /* nothing found : store previous solution */
|
|
}
|
|
#endif
|
|
|
|
/* store sequence */
|
|
_storeSequence:
|
|
|
|
/* catch up */
|
|
if (offset) {
|
|
while ((start>anchor) && (start>base+offset) && (start[-1] == start[-1-offset])) /* only search for offset within prefix */
|
|
{ start--; matchLength++; }
|
|
offset_2 = offset_1; offset_1 = offset;
|
|
}
|
|
|
|
{
|
|
size_t litLength = start - anchor;
|
|
LZ5_LOG_ENCODE("%d/%d: ENCODE literals=%d off=%d mlen=%d\n", (int)(ip-base), (int)(iend-base), (int)(litLength), (int)(offset), (int)matchLength);
|
|
ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, matchLength-MINMATCH);
|
|
anchor = ip = start + matchLength;
|
|
}
|
|
|
|
#ifdef ZSTD_USE_REP /* check immediate repcode */
|
|
while ( (ip <= ilimit)
|
|
&& (MEM_read32(ip) == MEM_read32(ip - offset_2)) ) {
|
|
/* store sequence */
|
|
matchLength = ZSTD_count(ip+MINMATCH, ip+MINMATCH-offset_2, iend);
|
|
offset = offset_2;
|
|
offset_2 = offset_1;
|
|
offset_1 = offset;
|
|
ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, matchLength);
|
|
ip += matchLength+MINMATCH;
|
|
anchor = ip;
|
|
continue; /* faster when present ... (?) */
|
|
}
|
|
#endif
|
|
}
|
|
|
|
/* Last Literals */
|
|
{
|
|
size_t lastLLSize = iend - anchor;
|
|
LZ5_LOG_ENCODE("%d/%d: ENCODE lastLLSize=%d\n", (int)(ip-base), (int)(iend-base), (int)(lastLLSize));
|
|
memcpy(seqStorePtr->lit, anchor, lastLLSize);
|
|
seqStorePtr->lit += lastLLSize;
|
|
}
|
|
}
|
|
|
|
|
|
|
|
|
|
/* *******************************
|
|
* Optimal parser
|
|
*********************************/
|
|
FORCE_INLINE
|
|
void ZSTD_compressBlock_opt_generic(ZSTD_CCtx* ctx,
|
|
const void* src, size_t srcSize,
|
|
const U32 searchMethod, const U32 depth)
|
|
{
|
|
seqStore_t* seqStorePtr = &(ctx->seqStore);
|
|
const BYTE* const istart = (const BYTE*)src;
|
|
const BYTE* ip = istart;
|
|
const BYTE* anchor = istart;
|
|
const BYTE* const iend = istart + srcSize;
|
|
const BYTE* const ilimit = iend - 8;
|
|
const BYTE* const base = ctx->base + ctx->dictLimit;
|
|
|
|
size_t rep_2=REPCODE_STARTVALUE, rep_1=REPCODE_STARTVALUE;
|
|
const U32 maxSearches = 1 << ctx->params.searchLog;
|
|
const U32 mls = ctx->params.searchLength;
|
|
|
|
typedef size_t (*getAllMatches_f)(ZSTD_CCtx* zc, const BYTE* ip, const BYTE* iLimit,
|
|
U32 maxNbAttempts, U32 matchLengthSearch, LZ5HC_match_t* matches);
|
|
getAllMatches_f getAllMatches = searchMethod ? ZSTD_BtGetAllMatches_selectMLS : ZSTD_HcGetAllMatches_selectMLS;
|
|
|
|
LZ5HC_optimal_t opt[LZ5_OPT_NUM+4];
|
|
LZ5HC_match_t matches[LZ5_OPT_NUM+1];
|
|
const uint8_t *inr;
|
|
int cur, cur2, cur_min, skip_num = 0;
|
|
int llen, litlen, price, match_num, last_pos;
|
|
|
|
const int sufficient_len = 128; //ctx->params.sufficientLength;
|
|
const int faster_get_matches = (ctx->params.strategy == ZSTD_opt);
|
|
|
|
|
|
// printf("orig_file="); print_hex_text(ip, srcSize, 0);
|
|
|
|
/* init */
|
|
ZSTD_resetSeqStore(seqStorePtr);
|
|
if ((ip-base) < REPCODE_STARTVALUE) ip = base + REPCODE_STARTVALUE;
|
|
|
|
|
|
/* Match Loop */
|
|
while (ip < ilimit) {
|
|
int mlen=0;
|
|
int best_mlen=0;
|
|
int best_off=0;
|
|
memset(opt, 0, sizeof(LZ5HC_optimal_t));
|
|
last_pos = 0;
|
|
llen = ip - anchor;
|
|
inr = ip;
|
|
|
|
#if 1
|
|
cur = 1;
|
|
/* check repCode */
|
|
if (MEM_read32(ip+cur) == MEM_read32(ip+cur - rep_1)) {
|
|
/* repcode : we take it */
|
|
mlen = ZSTD_count(ip+cur+MINMATCH, ip+cur+MINMATCH-rep_1, iend) + MINMATCH;
|
|
|
|
LZ5_LOG_PARSER("%d: start try REP rep=%d mlen=%d\n", (int)(ip-base), (int)rep_1, (int)mlen);
|
|
if (depth==0 || mlen > sufficient_len || mlen >= LZ5_OPT_NUM) {
|
|
ip+=cur; best_mlen = mlen; best_off = 0; cur = 0; last_pos = 1;
|
|
goto _storeSequence;
|
|
}
|
|
|
|
do
|
|
{
|
|
litlen = 0;
|
|
price = LZ5HC_get_price(llen + cur, 0, mlen - MINMATCH) - LZ5_LIT_COST(llen + cur);
|
|
if (mlen + cur > last_pos || price < opt[mlen + cur].price)
|
|
SET_PRICE(mlen + cur, mlen, 0, litlen, price);
|
|
mlen--;
|
|
}
|
|
while (mlen >= MINMATCH);
|
|
}
|
|
#endif
|
|
|
|
best_mlen = (last_pos) ? last_pos : MINMATCH;
|
|
|
|
if (faster_get_matches && last_pos)
|
|
match_num = 0;
|
|
else
|
|
{
|
|
/* first search (depth 0) */
|
|
match_num = getAllMatches(ctx, ip, iend, maxSearches, mls, matches);
|
|
}
|
|
|
|
LZ5_LOG_PARSER("%d: match_num=%d last_pos=%d\n", (int)(ip-base), match_num, last_pos);
|
|
if (!last_pos && !match_num) { ip++; continue; }
|
|
|
|
if (match_num && matches[match_num-1].len > sufficient_len)
|
|
{
|
|
best_mlen = matches[match_num-1].len;
|
|
best_off = matches[match_num-1].off;
|
|
cur = 0;
|
|
last_pos = 1;
|
|
goto _storeSequence;
|
|
}
|
|
|
|
// set prices using matches at position = 0
|
|
for (int i = 0; i < match_num; i++)
|
|
{
|
|
mlen = (i>0) ? matches[i-1].len+1 : best_mlen;
|
|
best_mlen = (matches[i].len < LZ5_OPT_NUM) ? matches[i].len : LZ5_OPT_NUM;
|
|
LZ5_LOG_PARSER("%d: start Found mlen=%d off=%d best_mlen=%d last_pos=%d\n", (int)(ip-base), matches[i].len, matches[i].off, (int)best_mlen, (int)last_pos);
|
|
while (mlen <= best_mlen)
|
|
{
|
|
litlen = 0;
|
|
price = LZ5HC_get_price(llen + litlen, matches[i].off, mlen - MINMATCH) - LZ5_LIT_COST(llen);
|
|
if (mlen > last_pos || price < opt[mlen].price)
|
|
SET_PRICE(mlen, mlen, matches[i].off, litlen, price);
|
|
mlen++;
|
|
}
|
|
}
|
|
|
|
if (last_pos < MINMATCH) {
|
|
// ip += ((ip-anchor) >> g_searchStrength) + 1; /* jump faster over incompressible sections */
|
|
ip++; continue;
|
|
}
|
|
|
|
// printf("%d: last_pos=%d\n", (int)(ip - base), (int)last_pos);
|
|
|
|
// opt[0].rep = opt[1].rep = rep_1;
|
|
// opt[0].mlen = opt[1].mlen = 1;
|
|
opt[0].rep = rep_1;
|
|
opt[0].mlen = 1;
|
|
|
|
|
|
// check further positions
|
|
for (skip_num = 0, cur = 1; cur <= last_pos; cur++)
|
|
{
|
|
inr = ip + cur;
|
|
|
|
if (opt[cur-1].mlen == 1)
|
|
{
|
|
litlen = opt[cur-1].litlen + 1;
|
|
|
|
if (cur != litlen)
|
|
{
|
|
price = opt[cur - litlen].price + LZ5_LIT_ONLY_COST(litlen);
|
|
LZ5_LOG_PRICE("%d: TRY1 opt[%d].price=%d price=%d cur=%d litlen=%d\n", (int)(inr-base), cur - litlen, opt[cur - litlen].price, price, cur, litlen);
|
|
}
|
|
else
|
|
{
|
|
price = LZ5_LIT_ONLY_COST(llen + litlen) - llen;
|
|
LZ5_LOG_PRICE("%d: TRY2 price=%d cur=%d litlen=%d llen=%d\n", (int)(inr-base), price, cur, litlen, llen);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
litlen = 1;
|
|
price = opt[cur - 1].price + LZ5_LIT_ONLY_COST(litlen);
|
|
LZ5_LOG_PRICE("%d: TRY3 price=%d cur=%d litlen=%d litonly=%d\n", (int)(inr-base), price, cur, litlen, LZ5_LIT_ONLY_COST(litlen));
|
|
}
|
|
|
|
mlen = 1;
|
|
best_mlen = 0;
|
|
LZ5_LOG_PRICE("%d: TRY4 price=%d opt[%d].price=%d\n", (int)(inr-base), price, cur, opt[cur].price);
|
|
|
|
if (cur > last_pos || price <= opt[cur].price) // || ((price == opt[cur].price) && (opt[cur-1].mlen == 1) && (cur != litlen)))
|
|
SET_PRICE(cur, mlen, best_mlen, litlen, price);
|
|
|
|
if (cur == last_pos) break;
|
|
|
|
if (opt[cur].mlen > 1)
|
|
{
|
|
mlen = opt[cur].mlen;
|
|
if (opt[cur].off < 1)
|
|
{
|
|
opt[cur].rep = opt[cur-mlen].rep;
|
|
LZ5_LOG_PARSER("%d: COPYREP1 cur=%d mlen=%d rep=%d\n", (int)(inr-base), cur, mlen, opt[cur-mlen].rep);
|
|
}
|
|
else
|
|
{
|
|
opt[cur].rep = 0;
|
|
LZ5_LOG_PARSER("%d: COPYREP2 cur=%d offset=%d rep=%d\n", (int)(inr-base), cur, 0, opt[cur].rep);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
opt[cur].rep = opt[cur-1].rep; // copy rep
|
|
}
|
|
|
|
|
|
LZ5_LOG_PARSER("%d: CURRENT price[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(inr-base), cur, last_pos, opt[cur].price, opt[cur].off, opt[cur].mlen, opt[cur].litlen, opt[cur].rep);
|
|
|
|
#if 0
|
|
// check rep
|
|
// best_mlen = 0;
|
|
mlen = ZSTD_count(inr, inr - opt[cur].rep, iend);
|
|
if (mlen >= MINMATCH && mlen > best_mlen)
|
|
{
|
|
LZ5_LOG_PARSER("%d: try REP rep=%d mlen=%d\n", (int)(inr-base), opt[cur].rep, mlen);
|
|
LZ5_LOG_PARSER("%d: Found REP mlen=%d off=%d rep=%d opt[%d].off=%d\n", (int)(inr-base), mlen, 0, opt[cur].rep, cur, opt[cur].off);
|
|
|
|
if (mlen > sufficient_len || cur + mlen >= LZ5_OPT_NUM)
|
|
{
|
|
best_mlen = mlen;
|
|
best_off = 0;
|
|
LZ5_LOG_PARSER("%d: REP sufficient_len=%d best_mlen=%d best_off=%d last_pos=%d\n", (int)(inr-base), sufficient_len, best_mlen, best_off, last_pos);
|
|
last_pos = cur + 1;
|
|
goto _storeSequence;
|
|
}
|
|
|
|
if (opt[cur].mlen == 1)
|
|
{
|
|
litlen = opt[cur].litlen;
|
|
|
|
if (cur != litlen)
|
|
{
|
|
price = opt[cur - litlen].price + LZ5HC_get_price(litlen, 0, mlen - MINMATCH);
|
|
LZ5_LOG_PRICE("%d: TRY1 opt[%d].price=%d price=%d cur=%d litlen=%d\n", (int)(inr-base), cur - litlen, opt[cur - litlen].price, price, cur, litlen);
|
|
}
|
|
else
|
|
{
|
|
price = LZ5HC_get_price(llen + litlen, 0, mlen - MINMATCH) - LZ5_LIT_COST(llen);
|
|
LZ5_LOG_PRICE("%d: TRY2 price=%d cur=%d litlen=%d llen=%d\n", (int)(inr-base), price, cur, litlen, llen);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
litlen = 0;
|
|
price = opt[cur].price + LZ5HC_get_price(litlen, 0, mlen - MINMATCH);
|
|
LZ5_LOG_PRICE("%d: TRY3 price=%d cur=%d litlen=%d getprice=%d\n", (int)(inr-base), price, cur, litlen, LZ5HC_get_price(litlen, 0, mlen - MINMATCH));
|
|
}
|
|
|
|
best_mlen = mlen;
|
|
if (faster_get_matches)
|
|
skip_num = best_mlen;
|
|
|
|
LZ5_LOG_PARSER("%d: Found REP mlen=%d off=%d price=%d litlen=%d price[%d]=%d\n", (int)(inr-base), mlen, 0, price, litlen, cur - litlen, opt[cur - litlen].price);
|
|
|
|
do
|
|
{
|
|
if (cur + mlen > last_pos || price <= opt[cur + mlen].price) // || ((price == opt[cur + mlen].price) && (opt[cur].mlen == 1) && (cur != litlen))) // at equal price prefer REP instead of MATCH
|
|
SET_PRICE(cur + mlen, mlen, 0, litlen, price);
|
|
mlen--;
|
|
}
|
|
while (mlen >= MINMATCH);
|
|
}
|
|
#endif
|
|
|
|
if (faster_get_matches && skip_num > 0)
|
|
{
|
|
skip_num--;
|
|
continue;
|
|
}
|
|
|
|
|
|
best_mlen = (best_mlen > MINMATCH) ? best_mlen : MINMATCH;
|
|
|
|
match_num = getAllMatches(ctx, inr, iend, maxSearches, mls, matches);
|
|
// match_num = LZ5HC_GetAllMatches(ctx, inr, ip, matchlimit, best_mlen, matches);
|
|
LZ5_LOG_PARSER("%d: LZ5HC_GetAllMatches match_num=%d\n", (int)(inr-base), match_num);
|
|
|
|
|
|
if (match_num > 0 && matches[match_num-1].len > sufficient_len)
|
|
{
|
|
cur -= matches[match_num-1].back;
|
|
best_mlen = matches[match_num-1].len;
|
|
best_off = matches[match_num-1].off;
|
|
last_pos = cur + 1;
|
|
goto _storeSequence;
|
|
}
|
|
|
|
cur_min = cur;
|
|
|
|
// set prices using matches at position = cur
|
|
for (int i = 0; i < match_num; i++)
|
|
{
|
|
mlen = (i>0) ? matches[i-1].len+1 : best_mlen;
|
|
cur2 = cur - matches[i].back;
|
|
best_mlen = (cur2 + matches[i].len < LZ5_OPT_NUM) ? matches[i].len : LZ5_OPT_NUM - cur2;
|
|
|
|
#if 0
|
|
if (mlen < MINMATCH)
|
|
{
|
|
printf("i=%d match_num=%d matches[i-1].len=%d\n", i, match_num, matches[i-1].len);
|
|
printf("%d: ERROR mlen=%d Found1 cur=%d cur2=%d mlen=%d off=%d best_mlen=%d last_pos=%d\n", (int)(inr-base), mlen, cur, cur2, matches[i].len, matches[i].off, best_mlen, last_pos), exit(0);
|
|
}
|
|
#endif
|
|
|
|
LZ5_LOG_PARSER("%d: Found1 cur=%d cur2=%d mlen=%d off=%d best_mlen=%d last_pos=%d\n", (int)(inr-base), cur, cur2, matches[i].len, matches[i].off, best_mlen, last_pos);
|
|
|
|
while (mlen <= best_mlen)
|
|
{
|
|
if (opt[cur2].mlen == 1)
|
|
{
|
|
litlen = opt[cur2].litlen;
|
|
|
|
if (cur2 != litlen)
|
|
price = opt[cur2 - litlen].price + LZ5HC_get_price(litlen, matches[i].off, mlen - MINMATCH);
|
|
else
|
|
price = LZ5HC_get_price(llen + litlen, matches[i].off, mlen - MINMATCH) - LZ5_LIT_COST(llen);
|
|
}
|
|
else
|
|
{
|
|
litlen = 0;
|
|
price = opt[cur2].price + LZ5HC_get_price(litlen, matches[i].off, mlen - MINMATCH);
|
|
}
|
|
|
|
LZ5_LOG_PARSER("%d: Found2 pred=%d mlen=%d best_mlen=%d off=%d price=%d litlen=%d price[%d]=%d\n", (int)(inr-base), matches[i].back, mlen, best_mlen, matches[i].off, price, litlen, cur - litlen, opt[cur - litlen].price);
|
|
// LZ5_LOG_PRICE("%d: TRY5 price=%d opt[%d].price=%d\n", (int)(inr-base), price, cur2 + mlen, opt[cur2 + mlen].price);
|
|
if (cur2 + mlen > last_pos || price < opt[cur2 + mlen].price)
|
|
{
|
|
SET_PRICE(cur2 + mlen, mlen, matches[i].off, litlen, price);
|
|
|
|
opt[cur2 + mlen].rep = matches[i].off; // update reps
|
|
if (cur2 < cur_min) cur_min = cur2;
|
|
}
|
|
|
|
mlen++;
|
|
}
|
|
}
|
|
|
|
if (cur_min < cur)
|
|
{
|
|
for (int i=cur_min-1; i<=last_pos; i++)
|
|
{
|
|
LZ5_LOG_PARSER("%d: BEFORE price[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(ip-base+i), i, last_pos, opt[i].price, opt[i].off, opt[i].mlen, opt[i].litlen, opt[i].rep);
|
|
}
|
|
|
|
for (int i=cur_min+1; i<=last_pos; i++)
|
|
if (opt[i].price < (1<<30) && (opt[i].off) < 1 && i - opt[i].mlen > cur_min) // invalidate reps
|
|
{
|
|
if (opt[i-1].mlen == 1)
|
|
{
|
|
litlen = opt[i-1].litlen + 1;
|
|
|
|
if (i != litlen)
|
|
{
|
|
price = opt[i - litlen].price + LZ5_LIT_ONLY_COST(litlen);
|
|
// LZ5_LOG_PRICE("%d: TRY1 opt[%d].price=%d price=%d cur=%d litlen=%d\n", (int)(inr-base), i - litlen, opt[i - litlen].price, price, i, litlen);
|
|
}
|
|
else
|
|
{
|
|
price = LZ5_LIT_ONLY_COST(llen + litlen) - llen;
|
|
// LZ5_LOG_PRICE("%d: TRY2 price=%d cur=%d litlen=%d llen=%d\n", (int)(inr-base), price, i, litlen, llen);
|
|
}
|
|
}
|
|
else
|
|
{
|
|
litlen = 1;
|
|
price = opt[i - 1].price + LZ5_LIT_ONLY_COST(litlen);
|
|
// LZ5_LOG_PRICE("%d: TRY3 price=%d cur=%d litlen=%d\n", (int)(inr-base), price, i, litlen);
|
|
}
|
|
|
|
mlen = 1;
|
|
best_mlen = 0;
|
|
LZ5_LOG_PRICE("%d: TRY6 price=%d opt[%d].price=%d\n", (int)(inr-base), price, i + mlen, opt[i + mlen].price);
|
|
SET_PRICE(i, mlen, best_mlen, litlen, price);
|
|
|
|
opt[i].rep = opt[i-1].rep; // copy reps
|
|
|
|
LZ5_LOG_PARSER("%d: INVALIDATE pred=%d price[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(inr-base), cur-cur_min, i, last_pos, opt[i].price, opt[i].off, opt[i].mlen, opt[i].litlen, opt[i].rep);
|
|
}
|
|
|
|
for (int i=cur_min-1; i<=last_pos; i++)
|
|
{
|
|
LZ5_LOG_PARSER("%d: AFTER price[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(ip-base+i), i, last_pos, opt[i].price, opt[i].off, opt[i].mlen, opt[i].litlen, opt[i].rep);
|
|
}
|
|
|
|
}
|
|
} // for (skip_num = 0, cur = 1; cur <= last_pos; cur++)
|
|
|
|
|
|
best_mlen = opt[last_pos].mlen;
|
|
best_off = opt[last_pos].off;
|
|
cur = last_pos - best_mlen;
|
|
// printf("%d: start=%d best_mlen=%d best_off=%d cur=%d\n", (int)(ip - base), (int)(start - ip), (int)best_mlen, (int)best_off, cur);
|
|
|
|
/* store sequence */
|
|
_storeSequence: // cur, last_pos, best_mlen, best_off have to be set
|
|
for (int i = 1; i <= last_pos; i++)
|
|
LZ5_LOG_PARSER("%d: price[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(ip-base+i), i, last_pos, opt[i].price, opt[i].off, opt[i].mlen, opt[i].litlen, opt[i].rep);
|
|
LZ5_LOG_PARSER("%d: cur=%d/%d best_mlen=%d best_off=%d rep=%d\n", (int)(ip-base+cur), (int)cur, (int)last_pos, (int)best_mlen, (int)best_off, opt[cur].rep);
|
|
|
|
opt[0].mlen = 1;
|
|
size_t offset;
|
|
|
|
while (cur >= 0)
|
|
{
|
|
mlen = opt[cur].mlen;
|
|
offset = opt[cur].off;
|
|
opt[cur].mlen = best_mlen;
|
|
opt[cur].off = best_off;
|
|
best_mlen = mlen;
|
|
best_off = offset;
|
|
cur -= mlen;
|
|
}
|
|
|
|
for (int i = 0; i <= last_pos;)
|
|
{
|
|
LZ5_LOG_PARSER("%d: price2[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(ip-base+i), i, last_pos, opt[i].price, opt[i].off, opt[i].mlen, opt[i].litlen, opt[i].rep);
|
|
i += opt[i].mlen;
|
|
}
|
|
|
|
cur = 0;
|
|
|
|
while (cur < last_pos)
|
|
{
|
|
LZ5_LOG_PARSER("%d: price3[%d/%d]=%d off=%d mlen=%d litlen=%d rep=%d\n", (int)(ip-base+cur), cur, last_pos, opt[cur].price, opt[cur].off, opt[cur].mlen, opt[cur].litlen, opt[cur].rep);
|
|
mlen = opt[cur].mlen;
|
|
if (mlen == 1) { ip++; cur++; continue; }
|
|
offset = opt[cur].off;
|
|
cur += mlen;
|
|
|
|
#if 1
|
|
if (offset)
|
|
{
|
|
size_t ml2 = ZSTD_count(ip, ip-offset, iend);
|
|
if (ml2 < mlen && ml2 < MINMATCH)
|
|
{
|
|
printf("ERROR %d: iend=%d mlen=%d offset=%d ml2=%d\n", (int)(ip - base), (int)(iend - ip), (int)mlen, (int)offset, (int)ml2);
|
|
exit(0);
|
|
}
|
|
}
|
|
// else
|
|
// printf("ERROR %d: mlen=%d offset=%d\n", (int)(ip - base), (int)mlen, (int)offset);
|
|
|
|
if (ip < anchor)
|
|
{
|
|
printf("ERROR %d: ip < anchor iend=%d mlen=%d offset=%d\n", (int)(ip - base), (int)(iend - ip), (int)mlen, (int)offset);
|
|
exit(0);
|
|
}
|
|
if (ip - offset < base)
|
|
{
|
|
printf("ERROR %d: ip - offset < base iend=%d mlen=%d offset=%d\n", (int)(ip - base), (int)(iend - ip), (int)mlen, (int)offset);
|
|
exit(0);
|
|
}
|
|
if (mlen < MINMATCH)
|
|
{
|
|
printf("ERROR %d: mlen < MINMATCH iend=%d mlen=%d offset=%d\n", (int)(ip - base), (int)(iend - ip), (int)mlen, (int)offset);
|
|
exit(0);
|
|
}
|
|
if (ip + mlen > iend)
|
|
{
|
|
printf("ERROR %d: ip + mlen >= iend iend=%d mlen=%d offset=%d\n", (int)(ip - base), (int)(iend - ip), (int)mlen, (int)offset);
|
|
exit(0);
|
|
}
|
|
#endif
|
|
|
|
size_t litLength = ip - anchor;
|
|
LZ5_LOG_ENCODE("%d/%d: ENCODE literals=%d off=%d mlen=%d\n", (int)(ip-base), (int)(iend-base), (int)(litLength), (int)(offset), (int)mlen);
|
|
// printf("orig="); print_hex_text(ip, mlen, 0);
|
|
// printf("match="); print_hex_text(ip-offset, mlen, 0);
|
|
ZSTD_storeSeq(seqStorePtr, litLength, anchor, offset, mlen-MINMATCH);
|
|
anchor = ip = ip + mlen;
|
|
|
|
if (offset)
|
|
{
|
|
rep_2 = rep_1;
|
|
rep_1 = offset;
|
|
}
|
|
else
|
|
{
|
|
/* best_off = rep_2;
|
|
rep_2 = rep_1;
|
|
rep_1 = best_off;*/
|
|
}
|
|
LZ5_LOG_PARSER("%d: offset=%d rep=%d\n", (int)(ip-base), (int)offset, (int)rep_1);
|
|
}
|
|
|
|
|
|
#if 0
|
|
// check immediate repcode
|
|
while ( (ip <= ilimit)
|
|
&& (MEM_read32(ip) == MEM_read32(ip - rep_2)) ) {
|
|
/* store sequence */
|
|
best_mlen = ZSTD_count(ip+MINMATCH, ip+MINMATCH-rep_2, iend);
|
|
best_off = rep_2;
|
|
rep_2 = rep_1;
|
|
rep_1 = best_off;
|
|
ZSTD_storeSeq(seqStorePtr, 0, anchor, 0, best_mlen);
|
|
ip += best_mlen+MINMATCH;
|
|
anchor = ip;
|
|
continue; // faster when present ... (?)
|
|
}
|
|
#endif
|
|
|
|
}
|
|
|
|
/* Last Literals */
|
|
{
|
|
size_t lastLLSize = iend - anchor;
|
|
LZ5_LOG_ENCODE("%d: lastLLSize literals=%d\n", (int)(ip-base), (int)(lastLLSize));
|
|
memcpy(seqStorePtr->lit, anchor, lastLLSize);
|
|
seqStorePtr->lit += lastLLSize;
|
|
}
|
|
}
|
|
|
|
|
|
|