extzstd 0.1 → 0.1.1

data/contrib/zstd/{compress → lib/compress}/zstd_compress.c

@@ -13,8 +13,6 @@
 ***************************************/
 #include <string.h>         /* memset */
 #include "mem.h"
-#define XXH_STATIC_LINKING_ONLY   /* XXH64_state_t */
-#include "xxhash.h"               /* XXH_reset, update, digest */
 #define FSE_STATIC_LINKING_ONLY   /* FSE_encodeSymbol */
 #include "fse.h"
 #define HUF_STATIC_LINKING_ONLY
@@ -33,6 +31,7 @@ typedef enum { ZSTDcs_created=0, ZSTDcs_init, ZSTDcs_ongoing, ZSTDcs_ending } ZS
 /*-*************************************
 *  Helper functions
 ***************************************/
+#define ZSTD_STATIC_ASSERT(c) { enum { ZSTD_static_assert = 1/(int)(!!(c)) }; }
 size_t ZSTD_compressBound(size_t srcSize) { return FSE_compressBound(srcSize) + 12; }
 
 
@@ -50,8 +49,7 @@ static void ZSTD_resetSeqStore(seqStore_t* ssPtr)
 /*-*************************************
 *  Context memory management
 ***************************************/
-struct ZSTD_CCtx_s
-{
+struct ZSTD_CCtx_s {
     const BYTE* nextSrc;    /* next block here to continue on current prefix */
     const BYTE* base;       /* All regular indexes relative to this position */
     const BYTE* dictBase;   /* extDict indexes relative to this position */
@@ -60,10 +58,12 @@ struct ZSTD_CCtx_s
     U32   nextToUpdate;     /* index from which to continue dictionary update */
     U32   nextToUpdate3;    /* index from which to continue dictionary update */
     U32   hashLog3;         /* dispatch table : larger == faster, more memory */
-    U32   loadedDictEnd;
+    U32   loadedDictEnd;    /* index of end of dictionary */
+    U32   forceWindow;      /* force back-references to respect limit of 1<<wLog, even for dictionary */
+    U32   forceRawDict;     /* Force loading dictionary in "content-only" mode (no header analysis) */
     ZSTD_compressionStage_e stage;
     U32   rep[ZSTD_REP_NUM];
-    U32   savedRep[ZSTD_REP_NUM];
+    U32   repToConfirm[ZSTD_REP_NUM];
     U32   dictID;
     ZSTD_parameters params;
     void* workSpace;
@@ -79,9 +79,11 @@ struct ZSTD_CCtx_s
     U32* chainTable;
     HUF_CElt* hufTable;
     U32 flagStaticTables;
+    HUF_repeat flagStaticHufTable;
     FSE_CTable offcodeCTable  [FSE_CTABLE_SIZE_U32(OffFSELog, MaxOff)];
     FSE_CTable matchlengthCTable[FSE_CTABLE_SIZE_U32(MLFSELog, MaxML)];
     FSE_CTable litlengthCTable  [FSE_CTABLE_SIZE_U32(LLFSELog, MaxLL)];
+    unsigned tmpCounters[HUF_WORKSPACE_SIZE_U32];
 };
 
 ZSTD_CCtx* ZSTD_createCCtx(void)
@@ -99,7 +101,7 @@ ZSTD_CCtx* ZSTD_createCCtx_advanced(ZSTD_customMem customMem)
     cctx = (ZSTD_CCtx*) ZSTD_malloc(sizeof(ZSTD_CCtx), customMem);
     if (!cctx) return NULL;
     memset(cctx, 0, sizeof(ZSTD_CCtx));
-    memcpy(&(cctx->customMem), &customMem, sizeof(customMem));
+    cctx->customMem = customMem;
     return cctx;
 }
 
@@ -117,6 +119,16 @@ size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx)
     return sizeof(*cctx) + cctx->workSpaceSize;
 }
 
+size_t ZSTD_setCCtxParameter(ZSTD_CCtx* cctx, ZSTD_CCtxParameter param, unsigned value)
+{
+    switch(param)
+    {
+    case ZSTD_p_forceWindow : cctx->forceWindow = value>0; cctx->loadedDictEnd = 0; return 0;
+    case ZSTD_p_forceRawDict : cctx->forceRawDict = value>0; return 0;
+    default: return ERROR(parameter_unknown);
+    }
+}
+
 const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx)   /* hidden interface */
 {
     return &(ctx->seqStore);
@@ -147,6 +159,14 @@ size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
 }
 
 
+/** ZSTD_cycleLog() :
+ *  condition for correct operation : hashLog > 1 */
+static U32 ZSTD_cycleLog(U32 hashLog, ZSTD_strategy strat)
+{
+    U32 const btScale = ((U32)strat >= (U32)ZSTD_btlazy2);
+    return hashLog - btScale;
+}
+
 /** ZSTD_adjustCParams() :
     optimize `cPar` for a given input (`srcSize` and `dictSize`).
     mostly downsizing to reduce memory consumption and initialization.
@@ -165,9 +185,9 @@ ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, u
             if (cPar.windowLog > srcLog) cPar.windowLog = srcLog;
     }   }
     if (cPar.hashLog > cPar.windowLog) cPar.hashLog = cPar.windowLog;
-    {   U32 const btPlus = (cPar.strategy == ZSTD_btlazy2) | (cPar.strategy == ZSTD_btopt) | (cPar.strategy == ZSTD_btopt2);
-        U32 const maxChainLog = cPar.windowLog+btPlus;
-        if (cPar.chainLog > maxChainLog) cPar.chainLog = maxChainLog; }   /* <= ZSTD_CHAINLOG_MAX */
+    {   U32 const cycleLog = ZSTD_cycleLog(cPar.chainLog, cPar.strategy);
+        if (cycleLog > cPar.windowLog) cPar.chainLog -= (cycleLog - cPar.windowLog);
+    }
 
     if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN) cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* required for frame header */
 
@@ -227,14 +247,17 @@ static size_t ZSTD_continueCCtx(ZSTD_CCtx* cctx, ZSTD_parameters params, U64 fra
 typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset, ZSTDcrp_fullReset } ZSTD_compResetPolicy_e;
 
 /*! ZSTD_resetCCtx_advanced() :
-    note : 'params' must be validated */
+    note : `params` must be validated */
 static size_t ZSTD_resetCCtx_advanced (ZSTD_CCtx* zc,
                                        ZSTD_parameters params, U64 frameContentSize,
                                        ZSTD_compResetPolicy_e const crp)
 {
     if (crp == ZSTDcrp_continue)
-        if (ZSTD_equivalentParams(params, zc->params))
+        if (ZSTD_equivalentParams(params, zc->params)) {
+            zc->flagStaticTables = 0;
+            zc->flagStaticHufTable = HUF_repeat_none;
             return ZSTD_continueCCtx(zc, params, frameContentSize);
+        }
 
     {   size_t const blockSize = MIN(ZSTD_BLOCKSIZE_ABSOLUTEMAX, (size_t)1 << params.cParams.windowLog);
         U32    const divider = (params.cParams.searchLength==3) ? 3 : 4;
@@ -268,6 +291,7 @@ static size_t ZSTD_resetCCtx_advanced (ZSTD_CCtx* zc,
         ptr = zc->hashTable3 + h3Size;
         zc->hufTable = (HUF_CElt*)ptr;
         zc->flagStaticTables = 0;
+        zc->flagStaticHufTable = HUF_repeat_none;
         ptr = ((U32*)ptr) + 256;  /* note : HUF_CElt* is incomplete type, size is simulated using U32 */
 
         zc->nextToUpdate = 1;
@@ -308,6 +332,14 @@ static size_t ZSTD_resetCCtx_advanced (ZSTD_CCtx* zc,
     }
 }
 
+/* ZSTD_invalidateRepCodes() :
+ * ensures next compression will not use repcodes from previous block.
+ * Note : only works with regular variant;
+ *        do not use with extDict variant ! */
+void ZSTD_invalidateRepCodes(ZSTD_CCtx* cctx) {
+    int i;
+    for (i=0; i<ZSTD_REP_NUM; i++) cctx->rep[i] = 0;
+}
 
 /*! ZSTD_copyCCtx() :
 *   Duplicate an existing context `srcCCtx` into another one `dstCCtx`.
@@ -317,8 +349,12 @@ size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long
 {
     if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong);
 
+
     memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
-    ZSTD_resetCCtx_advanced(dstCCtx, srcCCtx->params, pledgedSrcSize, ZSTDcrp_noMemset);
+    {   ZSTD_parameters params = srcCCtx->params;
+        params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
+        ZSTD_resetCCtx_advanced(dstCCtx, params, pledgedSrcSize, ZSTDcrp_noMemset);
+    }
 
     /* copy tables */
     {   size_t const chainSize = (srcCCtx->params.cParams.strategy == ZSTD_fast) ? 0 : (1 << srcCCtx->params.cParams.chainLog);
@@ -341,12 +377,15 @@ size_t ZSTD_copyCCtx(ZSTD_CCtx* dstCCtx, const ZSTD_CCtx* srcCCtx, unsigned long
 
     /* copy entropy tables */
     dstCCtx->flagStaticTables = srcCCtx->flagStaticTables;
+    dstCCtx->flagStaticHufTable = srcCCtx->flagStaticHufTable;
     if (srcCCtx->flagStaticTables) {
-        memcpy(dstCCtx->hufTable, srcCCtx->hufTable, 256*4);
         memcpy(dstCCtx->litlengthCTable, srcCCtx->litlengthCTable, sizeof(dstCCtx->litlengthCTable));
         memcpy(dstCCtx->matchlengthCTable, srcCCtx->matchlengthCTable, sizeof(dstCCtx->matchlengthCTable));
         memcpy(dstCCtx->offcodeCTable, srcCCtx->offcodeCTable, sizeof(dstCCtx->offcodeCTable));
     }
+    if (srcCCtx->flagStaticHufTable) {
+        memcpy(dstCCtx->hufTable, srcCCtx->hufTable, 256*4);
+    }
 
     return 0;
 }
@@ -460,24 +499,28 @@ static size_t ZSTD_compressLiterals (ZSTD_CCtx* zc,
 
     /* small ? don't even attempt compression (speed opt) */
 #   define LITERAL_NOENTROPY 63
-    {   size_t const minLitSize = zc->flagStaticTables ? 6 : LITERAL_NOENTROPY;
+    {   size_t const minLitSize = zc->flagStaticHufTable == HUF_repeat_valid ? 6 : LITERAL_NOENTROPY;
         if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
     }
 
     if (dstCapacity < lhSize+1) return ERROR(dstSize_tooSmall);   /* not enough space for compression */
-    if (zc->flagStaticTables && (lhSize==3)) {
-        hType = set_repeat;
-        singleStream = 1;
-        cLitSize = HUF_compress1X_usingCTable(ostart+lhSize, dstCapacity-lhSize, src, srcSize, zc->hufTable);
-    } else {
-        cLitSize = singleStream ? HUF_compress1X(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11)
-                                : HUF_compress2 (ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11);
+    {   HUF_repeat repeat = zc->flagStaticHufTable;
+        int const preferRepeat = zc->params.cParams.strategy < ZSTD_lazy ? srcSize <= 1024 : 0;
+        if (repeat == HUF_repeat_valid && lhSize == 3) singleStream = 1;
+        cLitSize = singleStream ? HUF_compress1X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, zc->tmpCounters, sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat)
+                                : HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11, zc->tmpCounters, sizeof(zc->tmpCounters), zc->hufTable, &repeat, preferRepeat);
+        if (repeat != HUF_repeat_none) { hType = set_repeat; }    /* reused the existing table */
+        else { zc->flagStaticHufTable = HUF_repeat_check; }       /* now have a table to reuse */
     }
 
-    if ((cLitSize==0) | (cLitSize >= srcSize - minGain))
+    if ((cLitSize==0) | (cLitSize >= srcSize - minGain)) {
+        zc->flagStaticHufTable = HUF_repeat_none;
         return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-    if (cLitSize==1)
+    }
+    if (cLitSize==1) {
+        zc->flagStaticHufTable = HUF_repeat_none;
         return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
+    }
 
     /* Build header */
     switch(lhSize)
@@ -545,11 +588,11 @@ void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
         mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
 }
 
-
-size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
+MEM_STATIC size_t ZSTD_compressSequences (ZSTD_CCtx* zc,
                               void* dst, size_t dstCapacity,
                               size_t srcSize)
 {
+    const int longOffsets = zc->params.cParams.windowLog > STREAM_ACCUMULATOR_MIN;
     const seqStore_t* seqStorePtr = &(zc->seqStore);
     U32 count[MaxSeq+1];
     S16 norm[MaxSeq+1];
@@ -566,6 +609,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
     BYTE* op = ostart;
     size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
     BYTE* seqHead;
+    BYTE scratchBuffer[1<<MAX(MLFSELog,LLFSELog)];
 
     /* Compress literals */
     {   const BYTE* const literals = seqStorePtr->litStart;
@@ -593,7 +637,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
 
     /* CTable for Literal Lengths */
     {   U32 max = MaxLL;
-        size_t const mostFrequent = FSE_countFast(count, &max, llCodeTable, nbSeq);
+        size_t const mostFrequent = FSE_countFast_wksp(count, &max, llCodeTable, nbSeq, zc->tmpCounters);
         if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
             *op++ = llCodeTable[0];
             FSE_buildCTable_rle(CTable_LitLength, (BYTE)max);
@@ -601,7 +645,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
         } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
             LLtype = set_repeat;
         } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (LL_defaultNormLog-1)))) {
-            FSE_buildCTable(CTable_LitLength, LL_defaultNorm, MaxLL, LL_defaultNormLog);
+            FSE_buildCTable_wksp(CTable_LitLength, LL_defaultNorm, MaxLL, LL_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));
             LLtype = set_basic;
         } else {
             size_t nbSeq_1 = nbSeq;
@@ -611,13 +655,13 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
             { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog);   /* overflow protected */
               if (FSE_isError(NCountSize)) return ERROR(GENERIC);
               op += NCountSize; }
-            FSE_buildCTable(CTable_LitLength, norm, max, tableLog);
+            FSE_buildCTable_wksp(CTable_LitLength, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));
             LLtype = set_compressed;
     }   }
 
     /* CTable for Offsets */
     {   U32 max = MaxOff;
-        size_t const mostFrequent = FSE_countFast(count, &max, ofCodeTable, nbSeq);
+        size_t const mostFrequent = FSE_countFast_wksp(count, &max, ofCodeTable, nbSeq, zc->tmpCounters);
         if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
             *op++ = ofCodeTable[0];
             FSE_buildCTable_rle(CTable_OffsetBits, (BYTE)max);
@@ -625,7 +669,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
         } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
             Offtype = set_repeat;
         } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (OF_defaultNormLog-1)))) {
-            FSE_buildCTable(CTable_OffsetBits, OF_defaultNorm, MaxOff, OF_defaultNormLog);
+            FSE_buildCTable_wksp(CTable_OffsetBits, OF_defaultNorm, MaxOff, OF_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));
             Offtype = set_basic;
         } else {
             size_t nbSeq_1 = nbSeq;
@@ -635,13 +679,13 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
             { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog);   /* overflow protected */
               if (FSE_isError(NCountSize)) return ERROR(GENERIC);
               op += NCountSize; }
-            FSE_buildCTable(CTable_OffsetBits, norm, max, tableLog);
+            FSE_buildCTable_wksp(CTable_OffsetBits, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));
             Offtype = set_compressed;
     }   }
 
     /* CTable for MatchLengths */
     {   U32 max = MaxML;
-        size_t const mostFrequent = FSE_countFast(count, &max, mlCodeTable, nbSeq);
+        size_t const mostFrequent = FSE_countFast_wksp(count, &max, mlCodeTable, nbSeq, zc->tmpCounters);
         if ((mostFrequent == nbSeq) && (nbSeq > 2)) {
             *op++ = *mlCodeTable;
             FSE_buildCTable_rle(CTable_MatchLength, (BYTE)max);
@@ -649,7 +693,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
         } else if ((zc->flagStaticTables) && (nbSeq < MAX_SEQ_FOR_STATIC_FSE)) {
             MLtype = set_repeat;
         } else if ((nbSeq < MIN_SEQ_FOR_DYNAMIC_FSE) || (mostFrequent < (nbSeq >> (ML_defaultNormLog-1)))) {
-            FSE_buildCTable(CTable_MatchLength, ML_defaultNorm, MaxML, ML_defaultNormLog);
+            FSE_buildCTable_wksp(CTable_MatchLength, ML_defaultNorm, MaxML, ML_defaultNormLog, scratchBuffer, sizeof(scratchBuffer));
             MLtype = set_basic;
         } else {
             size_t nbSeq_1 = nbSeq;
@@ -659,7 +703,7 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
             { size_t const NCountSize = FSE_writeNCount(op, oend-op, norm, max, tableLog);   /* overflow protected */
               if (FSE_isError(NCountSize)) return ERROR(GENERIC);
               op += NCountSize; }
-            FSE_buildCTable(CTable_MatchLength, norm, max, tableLog);
+            FSE_buildCTable_wksp(CTable_MatchLength, norm, max, tableLog, scratchBuffer, sizeof(scratchBuffer));
             MLtype = set_compressed;
     }   }
 
@@ -682,7 +726,18 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
         if (MEM_32bits()) BIT_flushBits(&blockStream);
         BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);
         if (MEM_32bits()) BIT_flushBits(&blockStream);
-        BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]);
+        if (longOffsets) {
+            U32 const ofBits = ofCodeTable[nbSeq-1];
+            int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);
+            if (extraBits) {
+                BIT_addBits(&blockStream, sequences[nbSeq-1].offset, extraBits);
+                BIT_flushBits(&blockStream);
+            }
+            BIT_addBits(&blockStream, sequences[nbSeq-1].offset >> extraBits,
+                        ofBits - extraBits);
+        } else {
+            BIT_addBits(&blockStream, sequences[nbSeq-1].offset, ofCodeTable[nbSeq-1]);
+        }
         BIT_flushBits(&blockStream);
 
         {   size_t n;
@@ -704,7 +759,17 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
                 if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);
                 BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
                 if (MEM_32bits()) BIT_flushBits(&blockStream);                  /* (7)*/
-                BIT_addBits(&blockStream, sequences[n].offset, ofBits);         /* 31 */
+                if (longOffsets) {
+                    int const extraBits = ofBits - MIN(ofBits, STREAM_ACCUMULATOR_MIN-1);
+                    if (extraBits) {
+                        BIT_addBits(&blockStream, sequences[n].offset, extraBits);
+                        BIT_flushBits(&blockStream);                            /* (7)*/
+                    }
+                    BIT_addBits(&blockStream, sequences[n].offset >> extraBits,
+                                ofBits - extraBits);                            /* 31 */
+                } else {
+                    BIT_addBits(&blockStream, sequences[n].offset, ofBits);     /* 31 */
+                }
                 BIT_flushBits(&blockStream);                                    /* (7)*/
         }   }
 
@@ -719,16 +784,25 @@ size_t ZSTD_compressSequences(ZSTD_CCtx* zc,
 
     /* check compressibility */
 _check_compressibility:
-    { size_t const minGain = ZSTD_minGain(srcSize);
-      size_t const maxCSize = srcSize - minGain;
-      if ((size_t)(op-ostart) >= maxCSize) return 0; }
+    {   size_t const minGain = ZSTD_minGain(srcSize);
+        size_t const maxCSize = srcSize - minGain;
+        if ((size_t)(op-ostart) >= maxCSize) {
+            zc->flagStaticHufTable = HUF_repeat_none;
+            return 0;
+    }   }
 
     /* confirm repcodes */
-    { int i; for (i=0; i<ZSTD_REP_NUM; i++) zc->rep[i] = zc->savedRep[i]; }
+    { int i; for (i=0; i<ZSTD_REP_NUM; i++) zc->rep[i] = zc->repToConfirm[i]; }
 
     return op - ostart;
 }
 
+#if 0 /* for debug */
+#  define STORESEQ_DEBUG
+#include <stdio.h>   /* fprintf */
+U32 g_startDebug = 0;
+const BYTE* g_start = NULL;
+#endif
 
 /*! ZSTD_storeSeq() :
     Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
@@ -737,13 +811,14 @@ _check_compressibility:
 */
 MEM_STATIC void ZSTD_storeSeq(seqStore_t* seqStorePtr, size_t litLength, const void* literals, U32 offsetCode, size_t matchCode)
 {
-#if 0  /* for debug */
-    static const BYTE* g_start = NULL;
-    const U32 pos = (U32)(literals - g_start);
-    if (g_start==NULL) g_start = literals;
-    //if ((pos > 1) && (pos < 50000))
-        printf("Cpos %6u :%5u literals & match %3u bytes at distance %6u \n",
-               pos, (U32)litLength, (U32)matchCode+MINMATCH, (U32)offsetCode);
+#ifdef STORESEQ_DEBUG
+    if (g_startDebug) {
+        const U32 pos = (U32)((const BYTE*)literals - g_start);
+        if (g_start==NULL) g_start = (const BYTE*)literals;
+        if ((pos > 1895000) && (pos < 1895300))
+            fprintf(stderr, "Cpos %6u :%5u literals & match %3u bytes at distance %6u \n",
+                   pos, (U32)litLength, (U32)matchCode+MINMATCH, (U32)offsetCode);
+    }
 #endif
     /* copy Literals */
     ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
@@ -993,8 +1068,8 @@ void ZSTD_compressBlock_fast_generic(ZSTD_CCtx* cctx,
     }   }   }
 
     /* save reps for next block */
-    cctx->savedRep[0] = offset_1 ? offset_1 : offsetSaved;
-    cctx->savedRep[1] = offset_2 ? offset_2 : offsetSaved;
+    cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
+    cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -1108,7 +1183,7 @@ static void ZSTD_compressBlock_fast_extDict_generic(ZSTD_CCtx* ctx,
     }   }   }
 
     /* save reps for next block */
-    ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;
+    ctx->repToConfirm[0] = offset_1; ctx->repToConfirm[1] = offset_2;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -1262,8 +1337,8 @@ void ZSTD_compressBlock_doubleFast_generic(ZSTD_CCtx* cctx,
     }   }   }
 
     /* save reps for next block */
-    cctx->savedRep[0] = offset_1 ? offset_1 : offsetSaved;
-    cctx->savedRep[1] = offset_2 ? offset_2 : offsetSaved;
+    cctx->repToConfirm[0] = offset_1 ? offset_1 : offsetSaved;
+    cctx->repToConfirm[1] = offset_2 ? offset_2 : offsetSaved;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -1412,7 +1487,7 @@ static void ZSTD_compressBlock_doubleFast_extDict_generic(ZSTD_CCtx* ctx,
     }   }   }
 
     /* save reps for next block */
-    ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;
+    ctx->repToConfirm[0] = offset_1; ctx->repToConfirm[1] = offset_2;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -1482,8 +1557,9 @@ static U32 ZSTD_insertBt1(ZSTD_CCtx* zc, const BYTE* const ip, const U32 mls, co
     hashTable[h] = current;   /* Update Hash Table */
 
     while (nbCompares-- && (matchIndex > windowLow)) {
-        U32* nextPtr = bt + 2*(matchIndex & btMask);
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
         size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
+
 #ifdef ZSTD_C_PREDICT   /* note : can create issues when hlog small <= 11 */
         const U32* predictPtr = bt + 2*((matchIndex-1) & btMask);   /* written this way, as bt is a roll buffer */
         if (matchIndex == predictedSmall) {
@@ -1579,7 +1655,7 @@ static size_t ZSTD_insertBtAndFindBestMatch (
     hashTable[h] = current;   /* Update Hash Table */
 
     while (nbCompares-- && (matchIndex > windowLow)) {
-        U32* nextPtr = bt + 2*(matchIndex & btMask);
+        U32* const nextPtr = bt + 2*(matchIndex & btMask);
         size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
         const BYTE* match;
 
@@ -1711,7 +1787,7 @@ static size_t ZSTD_BtFindBestMatch_selectMLS_extDict (
 #define NEXT_IN_CHAIN(d, mask)   chainTable[(d) & mask]
 
 /* Update chains up to ip (excluded)
-   Assumption : always within prefix (ie. not within extDict) */
+   Assumption : always within prefix (i.e. not within extDict) */
 FORCE_INLINE
 U32 ZSTD_insertAndFindFirstIndex (ZSTD_CCtx* zc, const BYTE* ip, U32 mls)
 {
@@ -1943,8 +2019,8 @@ _storeSequence:
     }   }
 
     /* Save reps for next block */
-    ctx->savedRep[0] = offset_1 ? offset_1 : savedOffset;
-    ctx->savedRep[1] = offset_2 ? offset_2 : savedOffset;
+    ctx->repToConfirm[0] = offset_1 ? offset_1 : savedOffset;
+    ctx->repToConfirm[1] = offset_2 ? offset_2 : savedOffset;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -2138,7 +2214,7 @@ _storeSequence:
     }   }
 
     /* Save reps for next block */
-    ctx->savedRep[0] = offset_1; ctx->savedRep[1] = offset_2;
+    ctx->repToConfirm[0] = offset_1; ctx->repToConfirm[1] = offset_2;
 
     /* Last Literals */
     {   size_t const lastLLSize = iend - anchor;
@@ -2271,16 +2347,16 @@ static size_t ZSTD_compress_generic (ZSTD_CCtx* cctx,
         if (remaining < blockSize) blockSize = remaining;
 
         /* preemptive overflow correction */
-        if (cctx->lowLimit > (1<<30)) {
-            U32 const btplus = (cctx->params.cParams.strategy == ZSTD_btlazy2) | (cctx->params.cParams.strategy == ZSTD_btopt) | (cctx->params.cParams.strategy == ZSTD_btopt2);
-            U32 const chainMask = (1 << (cctx->params.cParams.chainLog - btplus)) - 1;
-            U32 const supLog = MAX(cctx->params.cParams.chainLog, 17 /* blockSize */);
-            U32 const newLowLimit = (cctx->lowLimit & chainMask) + (1 << supLog);   /* preserve position % chainSize, ensure current-repcode doesn't underflow */
-            U32 const correction = cctx->lowLimit - newLowLimit;
+        if (cctx->lowLimit > (3U<<29)) {
+            U32 const cycleMask = (1 << ZSTD_cycleLog(cctx->params.cParams.hashLog, cctx->params.cParams.strategy)) - 1;
+            U32 const current = (U32)(ip - cctx->base);
+            U32 const newCurrent = (current & cycleMask) + (1 << cctx->params.cParams.windowLog);
+            U32 const correction = current - newCurrent;
+            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_64 <= 30);
             ZSTD_reduceIndex(cctx, correction);
             cctx->base += correction;
             cctx->dictBase += correction;
-            cctx->lowLimit = newLowLimit;
+            cctx->lowLimit -= correction;
             cctx->dictLimit -= correction;
             if (cctx->nextToUpdate < correction) cctx->nextToUpdate = 0;
             else cctx->nextToUpdate -= correction;
@@ -2325,7 +2401,7 @@ static size_t ZSTD_writeFrameHeader(void* dst, size_t dstCapacity,
     U32   const dictIDSizeCode = (dictID>0) + (dictID>=256) + (dictID>=65536);   /* 0-3 */
     U32   const checksumFlag = params.fParams.checksumFlag>0;
     U32   const windowSize = 1U << params.cParams.windowLog;
-    U32   const singleSegment = params.fParams.contentSizeFlag && (windowSize > (pledgedSrcSize-1));
+    U32   const singleSegment = params.fParams.contentSizeFlag && (windowSize >= pledgedSrcSize);
     BYTE  const windowLogByte = (BYTE)((params.cParams.windowLog - ZSTD_WINDOWLOG_ABSOLUTEMIN) << 3);
     U32   const fcsCode = params.fParams.contentSizeFlag ?
                      (pledgedSrcSize>=256) + (pledgedSrcSize>=65536+256) + (pledgedSrcSize>=0xFFFFFFFFU) :   /* 0-3 */
@@ -2397,12 +2473,14 @@ static size_t ZSTD_compressContinue_internal (ZSTD_CCtx* cctx,
 
     cctx->nextSrc = ip + srcSize;
 
-    {   size_t const cSize = frame ?
+    if (srcSize) {
+        size_t const cSize = frame ?
                              ZSTD_compress_generic (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :
                              ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize);
         if (ZSTD_isError(cSize)) return cSize;
         return cSize + fhSize;
-    }
+    } else
+        return fhSize;
 }
 
 
@@ -2438,7 +2516,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t
     zc->dictBase = zc->base;
     zc->base += ip - zc->nextSrc;
     zc->nextToUpdate = zc->dictLimit;
-    zc->loadedDictEnd = (U32)(iend - zc->base);
+    zc->loadedDictEnd = zc->forceWindow ? 0 : (U32)(iend - zc->base);
 
     zc->nextSrc = iend;
     if (srcSize <= HASH_READ_SIZE) return 0;
@@ -2469,7 +2547,7 @@ static size_t ZSTD_loadDictionaryContent(ZSTD_CCtx* zc, const void* src, size_t
         return ERROR(GENERIC);   /* strategy doesn't exist; impossible */
     }
 
-    zc->nextToUpdate = zc->loadedDictEnd;
+    zc->nextToUpdate = (U32)(iend - zc->base);
     return 0;
 }
 
@@ -2506,6 +2584,7 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_
     const BYTE* const dictEnd = dictPtr + dictSize;
     short offcodeNCount[MaxOff+1];
     unsigned offcodeMaxValue = MaxOff;
+    BYTE scratchBuffer[1<<MAX(MLFSELog,LLFSELog)];
 
     {   size_t const hufHeaderSize = HUF_readCTable(cctx->hufTable, 255, dict, dictSize);
         if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted);
@@ -2517,7 +2596,7 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_
         if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted);
         if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted);
         /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */
-        CHECK_E (FSE_buildCTable(cctx->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog), dictionary_corrupted);
+        CHECK_E (FSE_buildCTable_wksp(cctx->offcodeCTable, offcodeNCount, offcodeMaxValue, offcodeLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);
         dictPtr += offcodeHeaderSize;
     }
 
@@ -2528,7 +2607,7 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_
         if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted);
         /* Every match length code must have non-zero probability */
         CHECK_F (ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
-        CHECK_E (FSE_buildCTable(cctx->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog), dictionary_corrupted);
+        CHECK_E (FSE_buildCTable_wksp(cctx->matchlengthCTable, matchlengthNCount, matchlengthMaxValue, matchlengthLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);
         dictPtr += matchlengthHeaderSize;
     }
 
@@ -2539,14 +2618,14 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_
         if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted);
         /* Every literal length code must have non-zero probability */
         CHECK_F (ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
-        CHECK_E(FSE_buildCTable(cctx->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog), dictionary_corrupted);
+        CHECK_E(FSE_buildCTable_wksp(cctx->litlengthCTable, litlengthNCount, litlengthMaxValue, litlengthLog, scratchBuffer, sizeof(scratchBuffer)), dictionary_corrupted);
         dictPtr += litlengthHeaderSize;
     }
 
     if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted);
-    cctx->rep[0] = MEM_readLE32(dictPtr+0); if (cctx->rep[0] >= dictSize) return ERROR(dictionary_corrupted);
-    cctx->rep[1] = MEM_readLE32(dictPtr+4); if (cctx->rep[1] >= dictSize) return ERROR(dictionary_corrupted);
-    cctx->rep[2] = MEM_readLE32(dictPtr+8); if (cctx->rep[2] >= dictSize) return ERROR(dictionary_corrupted);
+    cctx->rep[0] = MEM_readLE32(dictPtr+0); if (cctx->rep[0] == 0 || cctx->rep[0] >= dictSize) return ERROR(dictionary_corrupted);
+    cctx->rep[1] = MEM_readLE32(dictPtr+4); if (cctx->rep[1] == 0 || cctx->rep[1] >= dictSize) return ERROR(dictionary_corrupted);
+    cctx->rep[2] = MEM_readLE32(dictPtr+8); if (cctx->rep[2] == 0 || cctx->rep[2] >= dictSize) return ERROR(dictionary_corrupted);
     dictPtr += 12;
 
     {   U32 offcodeMax = MaxOff;
@@ -2560,6 +2639,7 @@ static size_t ZSTD_loadDictEntropyStats(ZSTD_CCtx* cctx, const void* dict, size_
     }
 
     cctx->flagStaticTables = 1;
+    cctx->flagStaticHufTable = HUF_repeat_valid;
     return dictPtr - (const BYTE*)dict;
 }
 
@@ -2569,8 +2649,9 @@ static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx* zc, const void* dict, si
 {
     if ((dict==NULL) || (dictSize<=8)) return 0;
 
-    /* default : dict is pure content */
-    if (MEM_readLE32(dict) != ZSTD_DICT_MAGIC) return ZSTD_loadDictionaryContent(zc, dict, dictSize);
+    /* dict as pure content */
+    if ((MEM_readLE32(dict) != ZSTD_DICT_MAGIC) || (zc->forceRawDict))
+        return ZSTD_loadDictionaryContent(zc, dict, dictSize);
     zc->dictID = zc->params.fParams.noDictIDFlag ? 0 :  MEM_readLE32((const char*)dict+4);
 
     /* known magic number : dict is parsed for entropy stats and content */
@@ -2581,7 +2662,6 @@ static size_t ZSTD_compress_insertDictionary(ZSTD_CCtx* zc, const void* dict, si
     }
 }
 
-
 /*! ZSTD_compressBegin_internal() :
 *   @return : 0, or an error code */
 static size_t ZSTD_compressBegin_internal(ZSTD_CCtx* cctx,
@@ -2613,9 +2693,9 @@ size_t ZSTD_compressBegin_usingDict(ZSTD_CCtx* cctx, const void* dict, size_t di
 }
 
 
-size_t ZSTD_compressBegin(ZSTD_CCtx* zc, int compressionLevel)
+size_t ZSTD_compressBegin(ZSTD_CCtx* cctx, int compressionLevel)
 {
-    return ZSTD_compressBegin_usingDict(zc, NULL, 0, compressionLevel);
+    return ZSTD_compressBegin_usingDict(cctx, NULL, 0, compressionLevel);
 }
 
 
@@ -2695,7 +2775,7 @@ size_t ZSTD_compress_advanced (ZSTD_CCtx* ctx,
 
 size_t ZSTD_compress_usingDict(ZSTD_CCtx* ctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize, const void* dict, size_t dictSize, int compressionLevel)
 {
-    ZSTD_parameters params = ZSTD_getParams(compressionLevel, srcSize, dictSize);
+    ZSTD_parameters params = ZSTD_getParams(compressionLevel, srcSize, dict ? dictSize : 0);
     params.fParams.contentSizeFlag = 1;
     return ZSTD_compress_internal(ctx, dst, dstCapacity, src, srcSize, dict, dictSize, params);
 }
@@ -2720,7 +2800,8 @@ size_t ZSTD_compress(void* dst, size_t dstCapacity, const void* src, size_t srcS
 /* =====  Dictionary API  ===== */
 
 struct ZSTD_CDict_s {
-    void* dictContent;
+    void* dictBuffer;
+    const void* dictContent;
     size_t dictContentSize;
     ZSTD_CCtx* refContext;
 };  /* typedef'd tp ZSTD_CDict within "zstd.h" */
@@ -2728,39 +2809,45 @@ struct ZSTD_CDict_s {
 size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict)
 {
     if (cdict==NULL) return 0;   /* support sizeof on NULL */
-    return ZSTD_sizeof_CCtx(cdict->refContext) + cdict->dictContentSize;
+    return ZSTD_sizeof_CCtx(cdict->refContext) + (cdict->dictBuffer ? cdict->dictContentSize : 0) + sizeof(*cdict);
 }
 
-ZSTD_CDict* ZSTD_createCDict_advanced(const void* dict, size_t dictSize, ZSTD_parameters params, ZSTD_customMem customMem)
+ZSTD_CDict* ZSTD_createCDict_advanced(const void* dictBuffer, size_t dictSize, unsigned byReference,
+                                      ZSTD_parameters params, ZSTD_customMem customMem)
 {
     if (!customMem.customAlloc && !customMem.customFree) customMem = defaultCustomMem;
     if (!customMem.customAlloc || !customMem.customFree) return NULL;
 
     {   ZSTD_CDict* const cdict = (ZSTD_CDict*) ZSTD_malloc(sizeof(ZSTD_CDict), customMem);
-        void* const dictContent = ZSTD_malloc(dictSize, customMem);
         ZSTD_CCtx* const cctx = ZSTD_createCCtx_advanced(customMem);
 
-        if (!dictContent || !cdict || !cctx) {
-            ZSTD_free(dictContent, customMem);
+        if (!cdict || !cctx) {
             ZSTD_free(cdict, customMem);
-            ZSTD_free(cctx, customMem);
+            ZSTD_freeCCtx(cctx);
             return NULL;
         }
 
-        if (dictSize) {
-            memcpy(dictContent, dict, dictSize);
+        if ((byReference) || (!dictBuffer) || (!dictSize)) {
+            cdict->dictBuffer = NULL;
+            cdict->dictContent = dictBuffer;
+        } else {
+            void* const internalBuffer = ZSTD_malloc(dictSize, customMem);
+            if (!internalBuffer) { ZSTD_free(cctx, customMem); ZSTD_free(cdict, customMem); return NULL; }
+            memcpy(internalBuffer, dictBuffer, dictSize);
+            cdict->dictBuffer = internalBuffer;
+            cdict->dictContent = internalBuffer;
         }
-        {   size_t const errorCode = ZSTD_compressBegin_advanced(cctx, dictContent, dictSize, params, 0);
+
+        {   size_t const errorCode = ZSTD_compressBegin_advanced(cctx, cdict->dictContent, dictSize, params, 0);
             if (ZSTD_isError(errorCode)) {
-                ZSTD_free(dictContent, customMem);
+                ZSTD_free(cdict->dictBuffer, customMem);
                 ZSTD_free(cdict, customMem);
-                ZSTD_free(cctx, customMem);
+                ZSTD_freeCCtx(cctx);
                 return NULL;
         }   }
 
-        cdict->dictContent = dictContent;
-        cdict->dictContentSize = dictSize;
         cdict->refContext = cctx;
+        cdict->dictContentSize = dictSize;
         return cdict;
     }
 }
@@ -2770,7 +2857,15 @@ ZSTD_CDict* ZSTD_createCDict(const void* dict, size_t dictSize, int compressionL
     ZSTD_customMem const allocator = { NULL, NULL, NULL };
     ZSTD_parameters params = ZSTD_getParams(compressionLevel, 0, dictSize);
     params.fParams.contentSizeFlag = 1;
-    return ZSTD_createCDict_advanced(dict, dictSize, params, allocator);
+    return ZSTD_createCDict_advanced(dict, dictSize, 0, params, allocator);
+}
+
+ZSTD_CDict* ZSTD_createCDict_byReference(const void* dict, size_t dictSize, int compressionLevel)
+{
+    ZSTD_customMem const allocator = { NULL, NULL, NULL };
+    ZSTD_parameters params = ZSTD_getParams(compressionLevel, 0, dictSize);
+    params.fParams.contentSizeFlag = 1;
+    return ZSTD_createCDict_advanced(dict, dictSize, 1, params, allocator);
 }
 
 size_t ZSTD_freeCDict(ZSTD_CDict* cdict)
@@ -2778,7 +2873,7 @@ size_t ZSTD_freeCDict(ZSTD_CDict* cdict)
     if (cdict==NULL) return 0;   /* support free on NULL */
     {   ZSTD_customMem const cMem = cdict->refContext->customMem;
         ZSTD_freeCCtx(cdict->refContext);
-        ZSTD_free(cdict->dictContent, cMem);
+        ZSTD_free(cdict->dictBuffer, cMem);
         ZSTD_free(cdict, cMem);
         return 0;
     }
@@ -2788,10 +2883,14 @@ static ZSTD_parameters ZSTD_getParamsFromCDict(const ZSTD_CDict* cdict) {
     return ZSTD_getParamsFromCCtx(cdict->refContext);
 }
 
-size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict, U64 pledgedSrcSize)
+size_t ZSTD_compressBegin_usingCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict, unsigned long long pledgedSrcSize)
 {
     if (cdict->dictContentSize) CHECK_F(ZSTD_copyCCtx(cctx, cdict->refContext, pledgedSrcSize))
-    else CHECK_F(ZSTD_compressBegin_advanced(cctx, NULL, 0, cdict->refContext->params, pledgedSrcSize));
+    else {
+        ZSTD_parameters params = cdict->refContext->params;
+        params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
+        CHECK_F(ZSTD_compressBegin_advanced(cctx, NULL, 0, params, pledgedSrcSize));
+    }
     return 0;
 }
 
@@ -2839,6 +2938,8 @@ struct ZSTD_CStream_s {
     ZSTD_cStreamStage stage;
     U32    checksum;
     U32    frameEnded;
+    U64    pledgedSrcSize;
+    U64    inputProcessed;
     ZSTD_parameters params;
     ZSTD_customMem customMem;
 };   /* typedef'd to ZSTD_CStream within "zstd.h" */
@@ -2883,9 +2984,9 @@ size_t ZSTD_freeCStream(ZSTD_CStream* zcs)
 size_t ZSTD_CStreamInSize(void)  { return ZSTD_BLOCKSIZE_ABSOLUTEMAX; }
 size_t ZSTD_CStreamOutSize(void) { return ZSTD_compressBound(ZSTD_BLOCKSIZE_ABSOLUTEMAX) + ZSTD_blockHeaderSize + 4 /* 32-bits hash */ ; }
 
-size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
+static size_t ZSTD_resetCStream_internal(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
 {
-    if (zcs->inBuffSize==0) return ERROR(stage_wrong);   /* zcs has not been init at least once */
+    if (zcs->inBuffSize==0) return ERROR(stage_wrong);   /* zcs has not been init at least once => can't reset */
 
     if (zcs->cdict) CHECK_F(ZSTD_compressBegin_usingCDict(zcs->cctx, zcs->cdict, pledgedSrcSize))
     else CHECK_F(ZSTD_compressBegin_advanced(zcs->cctx, NULL, 0, zcs->params, pledgedSrcSize));
@@ -2896,9 +2997,19 @@ size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
     zcs->outBuffContentSize = zcs->outBuffFlushedSize = 0;
     zcs->stage = zcss_load;
     zcs->frameEnded = 0;
+    zcs->pledgedSrcSize = pledgedSrcSize;
+    zcs->inputProcessed = 0;
     return 0;   /* ready to go */
 }
 
+size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
+{
+
+    zcs->params.fParams.contentSizeFlag = (pledgedSrcSize > 0);
+
+    return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
+}
+
 size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
                                  const void* dict, size_t dictSize,
                                  ZSTD_parameters params, unsigned long long pledgedSrcSize)
@@ -2920,9 +3031,9 @@ size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
         if (zcs->outBuff == NULL) return ERROR(memory_allocation);
     }
 
-    if (dict) {
+    if (dict && dictSize >= 8) {
         ZSTD_freeCDict(zcs->cdictLocal);
-        zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, params, zcs->customMem);
+        zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize, 0, params, zcs->customMem);
         if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);
         zcs->cdict = zcs->cdictLocal;
     } else zcs->cdict = NULL;
@@ -2930,7 +3041,7 @@ size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
     zcs->checksum = params.fParams.checksumFlag > 0;
     zcs->params = params;
 
-    return ZSTD_resetCStream(zcs, pledgedSrcSize);
+    return ZSTD_resetCStream_internal(zcs, pledgedSrcSize);
 }
 
 /* note : cdict must outlive compression session */
@@ -2939,6 +3050,7 @@ size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)
     ZSTD_parameters const params = ZSTD_getParamsFromCDict(cdict);
     size_t const initError =  ZSTD_initCStream_advanced(zcs, NULL, 0, params, 0);
     zcs->cdict = cdict;
+    zcs->cctx->dictID = params.fParams.noDictIDFlag ? 0 : cdict->refContext->dictID;
     return initError;
 }
 
@@ -2948,6 +3060,13 @@ size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t di
     return ZSTD_initCStream_advanced(zcs, dict, dictSize, params, 0);
 }
 
+size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize)
+{
+    ZSTD_parameters params = ZSTD_getParams(compressionLevel, pledgedSrcSize, 0);
+    if (pledgedSrcSize) params.fParams.contentSizeFlag = 1;
+    return ZSTD_initCStream_advanced(zcs, NULL, 0, params, pledgedSrcSize);
+}
+
 size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
 {
     return ZSTD_initCStream_usingDict(zcs, NULL, 0, compressionLevel);
@@ -2956,7 +3075,7 @@ size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
 size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs)
 {
     if (zcs==NULL) return 0;   /* support sizeof on NULL */
-    return sizeof(zcs) + ZSTD_sizeof_CCtx(zcs->cctx) + ZSTD_sizeof_CDict(zcs->cdictLocal) + zcs->outBuffSize + zcs->inBuffSize;
+    return sizeof(*zcs) + ZSTD_sizeof_CCtx(zcs->cctx) + ZSTD_sizeof_CDict(zcs->cdictLocal) + zcs->outBuffSize + zcs->inBuffSize;
 }
 
 /*======   Compression   ======*/
@@ -3044,6 +3163,7 @@ static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
 
     *srcSizePtr = ip - istart;
     *dstCapacityPtr = op - ostart;
+    zcs->inputProcessed += *srcSizePtr;
     if (zcs->frameEnded) return 0;
     {   size_t hintInSize = zcs->inBuffTarget - zcs->inBuffPos;
         if (hintInSize==0) hintInSize = zcs->blockSize;
@@ -3088,6 +3208,9 @@ size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output)
     BYTE* const oend = (BYTE*)(output->dst) + output->size;
     BYTE* op = ostart;
 
+    if ((zcs->pledgedSrcSize) && (zcs->inputProcessed != zcs->pledgedSrcSize))
+        return ERROR(srcSize_wrong);   /* pledgedSrcSize not respected */
+
     if (zcs->stage != zcss_final) {
         /* flush whatever remains */
         size_t srcSize = 0;