zstandard: vendor python-zstandard 0.12

The upstream source distribution from PyPI was extracted. Unwanted
files were removed.

The clang-format ignore list was updated to reflect the new source
of files.

test-repo-compengines.t was updated to reflect a change in behavior
of the zstd library.

The project contains a vendored copy of zstandard 1.4.3. The old
version was 1.3.8. This should result in some minor performance wins.

# no-check-commit because 3rd party code has different style guidelines

Differential Revision: https://phab.mercurial-scm.org/D6858

contrib/python-zstandard/zstd/compress/zstd_compress_literals.c

@@ -0,0 +1,149 b''
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /*-*************************************
+ *  Dependencies
+ ***************************************/
+#include "zstd_compress_literals.h"
+
+size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    BYTE* const ostart = (BYTE* const)dst;
+    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
+
+    RETURN_ERROR_IF(srcSize + flSize > dstCapacity, dstSize_tooSmall);
+
+    switch(flSize)
+    {
+        case 1: /* 2 - 1 - 5 */
+            ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3));
+            break;
+        case 2: /* 2 - 2 - 12 */
+            MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4)));
+            break;
+        case 3: /* 2 - 2 - 20 */
+            MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4)));
+            break;
+        default:   /* not necessary : flSize is {1,2,3} */
+            assert(0);
+    }
+
+    memcpy(ostart + flSize, src, srcSize);
+    return srcSize + flSize;
+}
+
+size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
+{
+    BYTE* const ostart = (BYTE* const)dst;
+    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
+
+    (void)dstCapacity;  /* dstCapacity already guaranteed to be >=4, hence large enough */
+
+    switch(flSize)
+    {
+        case 1: /* 2 - 1 - 5 */
+            ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3));
+            break;
+        case 2: /* 2 - 2 - 12 */
+            MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4)));
+            break;
+        case 3: /* 2 - 2 - 20 */
+            MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4)));
+            break;
+        default:   /* not necessary : flSize is {1,2,3} */
+            assert(0);
+    }
+
+    ostart[flSize] = *(const BYTE*)src;
+    return flSize+1;
+}
+
+size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
+                              ZSTD_hufCTables_t* nextHuf,
+                              ZSTD_strategy strategy, int disableLiteralCompression,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize,
+                              void* workspace, size_t wkspSize,
+                        const int bmi2)
+{
+    size_t const minGain = ZSTD_minGain(srcSize, strategy);
+    size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
+    BYTE*  const ostart = (BYTE*)dst;
+    U32 singleStream = srcSize < 256;
+    symbolEncodingType_e hType = set_compressed;
+    size_t cLitSize;
+
+    DEBUGLOG(5,"ZSTD_compressLiterals (disableLiteralCompression=%i)",
+                disableLiteralCompression);
+
+    /* Prepare nextEntropy assuming reusing the existing table */
+    memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+
+    if (disableLiteralCompression)
+        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+
+    /* small ? don't even attempt compression (speed opt) */
+#   define COMPRESS_LITERALS_SIZE_MIN 63
+    {   size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
+        if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+    }
+
+    RETURN_ERROR_IF(dstCapacity < lhSize+1, dstSize_tooSmall, "not enough space for compression");
+    {   HUF_repeat repeat = prevHuf->repeatMode;
+        int const preferRepeat = 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,
+                                      workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2)
+                                : HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,
+                                      workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);
+        if (repeat != HUF_repeat_none) {
+            /* reused the existing table */
+            hType = set_repeat;
+        }
+    }
+
+    if ((cLitSize==0) | (cLitSize >= srcSize - minGain) | ERR_isError(cLitSize)) {
+        memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
+    }
+    if (cLitSize==1) {
+        memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
+        return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
+    }
+
+    if (hType == set_compressed) {
+        /* using a newly constructed table */
+        nextHuf->repeatMode = HUF_repeat_check;
+    }
+
+    /* Build header */
+    switch(lhSize)
+    {
+    case 3: /* 2 - 2 - 10 - 10 */
+        {   U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);
+            MEM_writeLE24(ostart, lhc);
+            break;
+        }
+    case 4: /* 2 - 2 - 14 - 14 */
+        {   U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);
+            MEM_writeLE32(ostart, lhc);
+            break;
+        }
+    case 5: /* 2 - 2 - 18 - 18 */
+        {   U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);
+            MEM_writeLE32(ostart, lhc);
+            ostart[4] = (BYTE)(cLitSize >> 10);
+            break;
+        }
+    default:  /* not possible : lhSize is {3,4,5} */
+        assert(0);
+    }
+    return lhSize+cLitSize;
+}

contrib/python-zstandard/zstd/compress/zstd_compress_literals.h

@@ -0,0 +1,29 b''
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPRESS_LITERALS_H
+#define ZSTD_COMPRESS_LITERALS_H
+
+#include "zstd_compress_internal.h" /* ZSTD_hufCTables_t, ZSTD_minGain() */
+
+
+size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize);
+
+size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
+                              ZSTD_hufCTables_t* nextHuf,
+                              ZSTD_strategy strategy, int disableLiteralCompression,
+                              void* dst, size_t dstCapacity,
+                        const void* src, size_t srcSize,
+                              void* workspace, size_t wkspSize,
+                        const int bmi2);
+
+#endif /* ZSTD_COMPRESS_LITERALS_H */

contrib/python-zstandard/zstd/compress/zstd_compress_sequences.c

@@ -0,0 +1,415 b''
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+ /*-*************************************
+ *  Dependencies
+ ***************************************/
+#include "zstd_compress_sequences.h"
+
+/**
+ * -log2(x / 256) lookup table for x in [0, 256).
+ * If x == 0: Return 0
+ * Else: Return floor(-log2(x / 256) * 256)
+ */
+static unsigned const kInverseProbabilityLog256[256] = {
+    0,    2048, 1792, 1642, 1536, 1453, 1386, 1329, 1280, 1236, 1197, 1162,
+    1130, 1100, 1073, 1047, 1024, 1001, 980,  960,  941,  923,  906,  889,
+    874,  859,  844,  830,  817,  804,  791,  779,  768,  756,  745,  734,
+    724,  714,  704,  694,  685,  676,  667,  658,  650,  642,  633,  626,
+    618,  610,  603,  595,  588,  581,  574,  567,  561,  554,  548,  542,
+    535,  529,  523,  517,  512,  506,  500,  495,  489,  484,  478,  473,
+    468,  463,  458,  453,  448,  443,  438,  434,  429,  424,  420,  415,
+    411,  407,  402,  398,  394,  390,  386,  382,  377,  373,  370,  366,
+    362,  358,  354,  350,  347,  343,  339,  336,  332,  329,  325,  322,
+    318,  315,  311,  308,  305,  302,  298,  295,  292,  289,  286,  282,
+    279,  276,  273,  270,  267,  264,  261,  258,  256,  253,  250,  247,
+    244,  241,  239,  236,  233,  230,  228,  225,  222,  220,  217,  215,
+    212,  209,  207,  204,  202,  199,  197,  194,  192,  190,  187,  185,
+    182,  180,  178,  175,  173,  171,  168,  166,  164,  162,  159,  157,
+    155,  153,  151,  149,  146,  144,  142,  140,  138,  136,  134,  132,
+    130,  128,  126,  123,  121,  119,  117,  115,  114,  112,  110,  108,
+    106,  104,  102,  100,  98,   96,   94,   93,   91,   89,   87,   85,
+    83,   82,   80,   78,   76,   74,   73,   71,   69,   67,   66,   64,
+    62,   61,   59,   57,   55,   54,   52,   50,   49,   47,   46,   44,
+    42,   41,   39,   37,   36,   34,   33,   31,   30,   28,   26,   25,
+    23,   22,   20,   19,   17,   16,   14,   13,   11,   10,   8,    7,
+    5,    4,    2,    1,
+};
+
+static unsigned ZSTD_getFSEMaxSymbolValue(FSE_CTable const* ctable) {
+  void const* ptr = ctable;
+  U16 const* u16ptr = (U16 const*)ptr;
+  U32 const maxSymbolValue = MEM_read16(u16ptr + 1);
+  return maxSymbolValue;
+}
+
+/**
+ * Returns the cost in bytes of encoding the normalized count header.
+ * Returns an error if any of the helper functions return an error.
+ */
+static size_t ZSTD_NCountCost(unsigned const* count, unsigned const max,
+                              size_t const nbSeq, unsigned const FSELog)
+{
+    BYTE wksp[FSE_NCOUNTBOUND];
+    S16 norm[MaxSeq + 1];
+    const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
+    FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq, max));
+    return FSE_writeNCount(wksp, sizeof(wksp), norm, max, tableLog);
+}
+
+/**
+ * Returns the cost in bits of encoding the distribution described by count
+ * using the entropy bound.
+ */
+static size_t ZSTD_entropyCost(unsigned const* count, unsigned const max, size_t const total)
+{
+    unsigned cost = 0;
+    unsigned s;
+    for (s = 0; s <= max; ++s) {
+        unsigned norm = (unsigned)((256 * count[s]) / total);
+        if (count[s] != 0 && norm == 0)
+            norm = 1;
+        assert(count[s] < total);
+        cost += count[s] * kInverseProbabilityLog256[norm];
+    }
+    return cost >> 8;
+}
+
+/**
+ * Returns the cost in bits of encoding the distribution in count using ctable.
+ * Returns an error if ctable cannot represent all the symbols in count.
+ */
+static size_t ZSTD_fseBitCost(
+    FSE_CTable const* ctable,
+    unsigned const* count,
+    unsigned const max)
+{
+    unsigned const kAccuracyLog = 8;
+    size_t cost = 0;
+    unsigned s;
+    FSE_CState_t cstate;
+    FSE_initCState(&cstate, ctable);
+    RETURN_ERROR_IF(ZSTD_getFSEMaxSymbolValue(ctable) < max, GENERIC,
+                    "Repeat FSE_CTable has maxSymbolValue %u < %u",
+                    ZSTD_getFSEMaxSymbolValue(ctable), max);
+    for (s = 0; s <= max; ++s) {
+        unsigned const tableLog = cstate.stateLog;
+        unsigned const badCost = (tableLog + 1) << kAccuracyLog;
+        unsigned const bitCost = FSE_bitCost(cstate.symbolTT, tableLog, s, kAccuracyLog);
+        if (count[s] == 0)
+            continue;
+        RETURN_ERROR_IF(bitCost >= badCost, GENERIC,
+                        "Repeat FSE_CTable has Prob[%u] == 0", s);
+        cost += count[s] * bitCost;
+    }
+    return cost >> kAccuracyLog;
+}
+
+/**
+ * Returns the cost in bits of encoding the distribution in count using the
+ * table described by norm. The max symbol support by norm is assumed >= max.
+ * norm must be valid for every symbol with non-zero probability in count.
+ */
+static size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,
+                                    unsigned const* count, unsigned const max)
+{
+    unsigned const shift = 8 - accuracyLog;
+    size_t cost = 0;
+    unsigned s;
+    assert(accuracyLog <= 8);
+    for (s = 0; s <= max; ++s) {
+        unsigned const normAcc = norm[s] != -1 ? norm[s] : 1;
+        unsigned const norm256 = normAcc << shift;
+        assert(norm256 > 0);
+        assert(norm256 < 256);
+        cost += count[s] * kInverseProbabilityLog256[norm256];
+    }
+    return cost >> 8;
+}
+
+symbolEncodingType_e
+ZSTD_selectEncodingType(
+        FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
+        size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
+        FSE_CTable const* prevCTable,
+        short const* defaultNorm, U32 defaultNormLog,
+        ZSTD_defaultPolicy_e const isDefaultAllowed,
+        ZSTD_strategy const strategy)
+{
+    ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);
+    if (mostFrequent == nbSeq) {
+        *repeatMode = FSE_repeat_none;
+        if (isDefaultAllowed && nbSeq <= 2) {
+            /* Prefer set_basic over set_rle when there are 2 or less symbols,
+             * since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol.
+             * If basic encoding isn't possible, always choose RLE.
+             */
+            DEBUGLOG(5, "Selected set_basic");
+            return set_basic;
+        }
+        DEBUGLOG(5, "Selected set_rle");
+        return set_rle;
+    }
+    if (strategy < ZSTD_lazy) {
+        if (isDefaultAllowed) {
+            size_t const staticFse_nbSeq_max = 1000;
+            size_t const mult = 10 - strategy;
+            size_t const baseLog = 3;
+            size_t const dynamicFse_nbSeq_min = (((size_t)1 << defaultNormLog) * mult) >> baseLog;  /* 28-36 for offset, 56-72 for lengths */
+            assert(defaultNormLog >= 5 && defaultNormLog <= 6);  /* xx_DEFAULTNORMLOG */
+            assert(mult <= 9 && mult >= 7);
+            if ( (*repeatMode == FSE_repeat_valid)
+              && (nbSeq < staticFse_nbSeq_max) ) {
+                DEBUGLOG(5, "Selected set_repeat");
+                return set_repeat;
+            }
+            if ( (nbSeq < dynamicFse_nbSeq_min)
+              || (mostFrequent < (nbSeq >> (defaultNormLog-1))) ) {
+                DEBUGLOG(5, "Selected set_basic");
+                /* The format allows default tables to be repeated, but it isn't useful.
+                 * When using simple heuristics to select encoding type, we don't want
+                 * to confuse these tables with dictionaries. When running more careful
+                 * analysis, we don't need to waste time checking both repeating tables
+                 * and default tables.
+                 */
+                *repeatMode = FSE_repeat_none;
+                return set_basic;
+            }
+        }
+    } else {
+        size_t const basicCost = isDefaultAllowed ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, count, max) : ERROR(GENERIC);
+        size_t const repeatCost = *repeatMode != FSE_repeat_none ? ZSTD_fseBitCost(prevCTable, count, max) : ERROR(GENERIC);
+        size_t const NCountCost = ZSTD_NCountCost(count, max, nbSeq, FSELog);
+        size_t const compressedCost = (NCountCost << 3) + ZSTD_entropyCost(count, max, nbSeq);
+
+        if (isDefaultAllowed) {
+            assert(!ZSTD_isError(basicCost));
+            assert(!(*repeatMode == FSE_repeat_valid && ZSTD_isError(repeatCost)));
+        }
+        assert(!ZSTD_isError(NCountCost));
+        assert(compressedCost < ERROR(maxCode));
+        DEBUGLOG(5, "Estimated bit costs: basic=%u\trepeat=%u\tcompressed=%u",
+                    (unsigned)basicCost, (unsigned)repeatCost, (unsigned)compressedCost);
+        if (basicCost <= repeatCost && basicCost <= compressedCost) {
+            DEBUGLOG(5, "Selected set_basic");
+            assert(isDefaultAllowed);
+            *repeatMode = FSE_repeat_none;
+            return set_basic;
+        }
+        if (repeatCost <= compressedCost) {
+            DEBUGLOG(5, "Selected set_repeat");
+            assert(!ZSTD_isError(repeatCost));
+            return set_repeat;
+        }
+        assert(compressedCost < basicCost && compressedCost < repeatCost);
+    }
+    DEBUGLOG(5, "Selected set_compressed");
+    *repeatMode = FSE_repeat_check;
+    return set_compressed;
+}
+
+size_t
+ZSTD_buildCTable(void* dst, size_t dstCapacity,
+                FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
+                unsigned* count, U32 max,
+                const BYTE* codeTable, size_t nbSeq,
+                const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
+                const FSE_CTable* prevCTable, size_t prevCTableSize,
+                void* workspace, size_t workspaceSize)
+{
+    BYTE* op = (BYTE*)dst;
+    const BYTE* const oend = op + dstCapacity;
+    DEBUGLOG(6, "ZSTD_buildCTable (dstCapacity=%u)", (unsigned)dstCapacity);
+
+    switch (type) {
+    case set_rle:
+        FORWARD_IF_ERROR(FSE_buildCTable_rle(nextCTable, (BYTE)max));
+        RETURN_ERROR_IF(dstCapacity==0, dstSize_tooSmall);
+        *op = codeTable[0];
+        return 1;
+    case set_repeat:
+        memcpy(nextCTable, prevCTable, prevCTableSize);
+        return 0;
+    case set_basic:
+        FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, workspace, workspaceSize));  /* note : could be pre-calculated */
+        return 0;
+    case set_compressed: {
+        S16 norm[MaxSeq + 1];
+        size_t nbSeq_1 = nbSeq;
+        const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
+        if (count[codeTable[nbSeq-1]] > 1) {
+            count[codeTable[nbSeq-1]]--;
+            nbSeq_1--;
+        }
+        assert(nbSeq_1 > 1);
+        FORWARD_IF_ERROR(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max));
+        {   size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog);   /* overflow protected */
+            FORWARD_IF_ERROR(NCountSize);
+            FORWARD_IF_ERROR(FSE_buildCTable_wksp(nextCTable, norm, max, tableLog, workspace, workspaceSize));
+            return NCountSize;
+        }
+    }
+    default: assert(0); RETURN_ERROR(GENERIC);
+    }
+}
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_encodeSequences_body(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    BIT_CStream_t blockStream;
+    FSE_CState_t  stateMatchLength;
+    FSE_CState_t  stateOffsetBits;
+    FSE_CState_t  stateLitLength;
+
+    RETURN_ERROR_IF(
+        ERR_isError(BIT_initCStream(&blockStream, dst, dstCapacity)),
+        dstSize_tooSmall, "not enough space remaining");
+    DEBUGLOG(6, "available space for bitstream : %i  (dstCapacity=%u)",
+                (int)(blockStream.endPtr - blockStream.startPtr),
+                (unsigned)dstCapacity);
+
+    /* first symbols */
+    FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]);
+    FSE_initCState2(&stateOffsetBits,  CTable_OffsetBits,  ofCodeTable[nbSeq-1]);
+    FSE_initCState2(&stateLitLength,   CTable_LitLength,   llCodeTable[nbSeq-1]);
+    BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]);
+    if (MEM_32bits()) BIT_flushBits(&blockStream);
+    BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);
+    if (MEM_32bits()) BIT_flushBits(&blockStream);
+    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;
+        for (n=nbSeq-2 ; n<nbSeq ; n--) {      /* intentional underflow */
+            BYTE const llCode = llCodeTable[n];
+            BYTE const ofCode = ofCodeTable[n];
+            BYTE const mlCode = mlCodeTable[n];
+            U32  const llBits = LL_bits[llCode];
+            U32  const ofBits = ofCode;
+            U32  const mlBits = ML_bits[mlCode];
+            DEBUGLOG(6, "encoding: litlen:%2u - matchlen:%2u - offCode:%7u",
+                        (unsigned)sequences[n].litLength,
+                        (unsigned)sequences[n].matchLength + MINMATCH,
+                        (unsigned)sequences[n].offset);
+                                                                            /* 32b*/  /* 64b*/
+                                                                            /* (7)*/  /* (7)*/
+            FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode);       /* 15 */  /* 15 */
+            FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode);      /* 24 */  /* 24 */
+            if (MEM_32bits()) BIT_flushBits(&blockStream);                  /* (7)*/
+            FSE_encodeSymbol(&blockStream, &stateLitLength, llCode);        /* 16 */  /* 33 */
+            if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog)))
+                BIT_flushBits(&blockStream);                                /* (7)*/
+            BIT_addBits(&blockStream, sequences[n].litLength, llBits);
+            if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);
+            BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
+            if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream);
+            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)*/
+            DEBUGLOG(7, "remaining space : %i", (int)(blockStream.endPtr - blockStream.ptr));
+    }   }
+
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing ML state with %u bits", stateMatchLength.stateLog);
+    FSE_flushCState(&blockStream, &stateMatchLength);
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing Off state with %u bits", stateOffsetBits.stateLog);
+    FSE_flushCState(&blockStream, &stateOffsetBits);
+    DEBUGLOG(6, "ZSTD_encodeSequences: flushing LL state with %u bits", stateLitLength.stateLog);
+    FSE_flushCState(&blockStream, &stateLitLength);
+
+    {   size_t const streamSize = BIT_closeCStream(&blockStream);
+        RETURN_ERROR_IF(streamSize==0, dstSize_tooSmall, "not enough space");
+        return streamSize;
+    }
+}
+
+static size_t
+ZSTD_encodeSequences_default(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    return ZSTD_encodeSequences_body(dst, dstCapacity,
+                                    CTable_MatchLength, mlCodeTable,
+                                    CTable_OffsetBits, ofCodeTable,
+                                    CTable_LitLength, llCodeTable,
+                                    sequences, nbSeq, longOffsets);
+}
+
+
+#if DYNAMIC_BMI2
+
+static TARGET_ATTRIBUTE("bmi2") size_t
+ZSTD_encodeSequences_bmi2(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets)
+{
+    return ZSTD_encodeSequences_body(dst, dstCapacity,
+                                    CTable_MatchLength, mlCodeTable,
+                                    CTable_OffsetBits, ofCodeTable,
+                                    CTable_LitLength, llCodeTable,
+                                    sequences, nbSeq, longOffsets);
+}
+
+#endif
+
+size_t ZSTD_encodeSequences(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)
+{
+    DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity);
+#if DYNAMIC_BMI2
+    if (bmi2) {
+        return ZSTD_encodeSequences_bmi2(dst, dstCapacity,
+                                         CTable_MatchLength, mlCodeTable,
+                                         CTable_OffsetBits, ofCodeTable,
+                                         CTable_LitLength, llCodeTable,
+                                         sequences, nbSeq, longOffsets);
+    }
+#endif
+    (void)bmi2;
+    return ZSTD_encodeSequences_default(dst, dstCapacity,
+                                        CTable_MatchLength, mlCodeTable,
+                                        CTable_OffsetBits, ofCodeTable,
+                                        CTable_LitLength, llCodeTable,
+                                        sequences, nbSeq, longOffsets);
+}

contrib/python-zstandard/zstd/compress/zstd_compress_sequences.h

@@ -0,0 +1,47 b''
+/*
+ * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
+ * All rights reserved.
+ *
+ * This source code is licensed under both the BSD-style license (found in the
+ * LICENSE file in the root directory of this source tree) and the GPLv2 (found
+ * in the COPYING file in the root directory of this source tree).
+ * You may select, at your option, one of the above-listed licenses.
+ */
+
+#ifndef ZSTD_COMPRESS_SEQUENCES_H
+#define ZSTD_COMPRESS_SEQUENCES_H
+
+#include "fse.h" /* FSE_repeat, FSE_CTable */
+#include "zstd_internal.h" /* symbolEncodingType_e, ZSTD_strategy */
+
+typedef enum {
+    ZSTD_defaultDisallowed = 0,
+    ZSTD_defaultAllowed = 1
+} ZSTD_defaultPolicy_e;
+
+symbolEncodingType_e
+ZSTD_selectEncodingType(
+        FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
+        size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
+        FSE_CTable const* prevCTable,
+        short const* defaultNorm, U32 defaultNormLog,
+        ZSTD_defaultPolicy_e const isDefaultAllowed,
+        ZSTD_strategy const strategy);
+
+size_t
+ZSTD_buildCTable(void* dst, size_t dstCapacity,
+                FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
+                unsigned* count, U32 max,
+                const BYTE* codeTable, size_t nbSeq,
+                const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
+                const FSE_CTable* prevCTable, size_t prevCTableSize,
+                void* workspace, size_t workspaceSize);
+
+size_t ZSTD_encodeSequences(
+            void* dst, size_t dstCapacity,
+            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
+            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
+            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
+            seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2);
+
+#endif /* ZSTD_COMPRESS_SEQUENCES_H */

contrib/clang-format-ignorelist

@@ -49,6 +49,10 b' contrib/python-zstandard/zstd/compress/h'
 contrib/python-zstandard/zstd/compress/huf_compress.c
 contrib/python-zstandard/zstd/compress/zstd_compress.c
 contrib/python-zstandard/zstd/compress/zstd_compress_internal.h
+contrib/python-zstandard/zstd/compress/zstd_compress_literals.c
+contrib/python-zstandard/zstd/compress/zstd_compress_literals.h
+contrib/python-zstandard/zstd/compress/zstd_compress_sequences.c
+contrib/python-zstandard/zstd/compress/zstd_compress_sequences.h
 contrib/python-zstandard/zstd/compress/zstd_double_fast.c
 contrib/python-zstandard/zstd/compress/zstd_double_fast.h
 contrib/python-zstandard/zstd/compress/zstd_fast.c

contrib/python-zstandard/NEWS.rst

@@ -44,6 +44,7 b' Actions Blocking Release'
   zstd API.
 * Expose ``ZSTD_CLEVEL_DEFAULT`` constant.
 * Support ``ZSTD_p_forceAttachDict`` compression parameter.
+* Support ``ZSTD_c_literalCompressionMode `` compression parameter.
 * Use ``ZSTD_CCtx_getParameter()``/``ZSTD_CCtxParam_getParameter()`` for retrieving
   compression parameters.
 * Consider exposing ``ZSTDMT_toFlushNow()``.
@@ -66,10 +67,39 b' Other Actions Not Blocking Release'
 * API for ensuring max memory ceiling isn't exceeded.
 * Move off nose for testing.
 
+0.12.0 (released 2019-09-15)
+============================
+
+Backwards Compatibility Notes
+-----------------------------
+
+* Support for Python 3.4 has been dropped since Python 3.4 is no longer
+  a supported Python version upstream. (But it will likely continue to
+  work until Python 2.7 support is dropped and we port to Python 3.5+
+  APIs.)
+
+Bug Fixes
+---------
+
+* Fix ``ZstdDecompressor.__init__`` on 64-bit big-endian systems (#91).
+* Fix memory leak in ``ZstdDecompressionReader.seek()`` (#82).
+
+Changes
+-------
+
+* CI transitioned to Azure Pipelines (from AppVeyor and Travis CI).
+* Switched to ``pytest`` for running tests (from ``nose``).
+* Bundled zstandard library upgraded from 1.3.8 to 1.4.3.
+
+0.11.1 (released 2019-05-14)
+============================
+
+* Fix memory leak in ``ZstdDecompressionReader.seek()`` (#82).
+
 0.11.0 (released 2019-02-24)
 ============================
 
-Backwards Compatibility Nodes
+Backwards Compatibility Notes
 -----------------------------
 
 * ``ZstdDecompressor.read()`` now allows reading sizes of ``-1`` or ``0``

contrib/python-zstandard/README.rst

@@ -15,7 +15,7 b' The canonical home for this project live'
 the author. For convenience, that repository is frequently synchronized to
 https://github.com/indygreg/python-zstandard.
 
-|  |ci-status| |win-ci-status|
+|  |ci-status|
 
 Requirements
 ============
@@ -1598,9 +1598,5 b' their work, consider donating some money'
     :target: https://www.paypal.com/cgi-bin/webscr?cmd=_donations&business=gregory%2eszorc%40gmail%2ecom&lc=US&item_name=python%2dzstandard&currency_code=USD&bn=PP%2dDonationsBF%3abtn_donate_LG%2egif%3aNonHosted
     :alt: Donate via PayPal
 
-.. |ci-status| image:: https://travis-ci.org/indygreg/python-zstandard.svg?branch=master
-    :target: https://travis-ci.org/indygreg/python-zstandard
-
-.. |win-ci-status| image:: https://ci.appveyor.com/api/projects/status/github/indygreg/python-zstandard?svg=true
-    :target: https://ci.appveyor.com/project/indygreg/python-zstandard
-    :alt: Windows build status
+.. |ci-status| image:: https://dev.azure.com/gregoryszorc/python-zstandard/_apis/build/status/indygreg.python-zstandard?branchName=master
+    :target: https://dev.azure.com/gregoryszorc/python-zstandard/_apis/build/status/indygreg.python-zstandard?branchName=master

contrib/python-zstandard/c-ext/compressionparams.c

@@ -11,7 +11,7 b''
 extern PyObject* ZstdError;
 
 int set_parameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value) {
-	size_t zresult = ZSTD_CCtxParam_setParameter(params, param, value);
+	size_t zresult = ZSTD_CCtxParams_setParameter(params, param, value);
 	if (ZSTD_isError(zresult)) {
 		PyErr_Format(ZstdError, "unable to set compression context parameter: %s",
 			ZSTD_getErrorName(zresult));
@@ -25,11 +25,11 b' int set_parameter(ZSTD_CCtx_params* para'
 
 #define TRY_COPY_PARAMETER(source, dest, param) { \
 	int result; \
-	size_t zresult = ZSTD_CCtxParam_getParameter(source, param, &result); \
+	size_t zresult = ZSTD_CCtxParams_getParameter(source, param, &result); \
 	if (ZSTD_isError(zresult)) { \
 		return 1; \
 	} \
-	zresult = ZSTD_CCtxParam_setParameter(dest, param, result); \
+	zresult = ZSTD_CCtxParams_setParameter(dest, param, result); \
 	if (ZSTD_isError(zresult)) { \
 		return 1; \
 	} \
@@ -78,7 +78,7 b' int reset_params(ZstdCompressionParamete'
 }
 
 #define TRY_GET_PARAMETER(params, param, value) { \
-    size_t zresult = ZSTD_CCtxParam_getParameter(params, param, value); \
+    size_t zresult = ZSTD_CCtxParams_getParameter(params, param, value); \
     if (ZSTD_isError(zresult)) { \
         PyErr_Format(ZstdError, "unable to retrieve parameter: %s", ZSTD_getErrorName(zresult)); \
         return 1; \
@@ -436,7 +436,7 b' static void ZstdCompressionParameters_de'
     int result; \
     size_t zresult; \
     ZstdCompressionParametersObject* p = (ZstdCompressionParametersObject*)(self); \
-    zresult = ZSTD_CCtxParam_getParameter(p->params, param, &result); \
+    zresult = ZSTD_CCtxParams_getParameter(p->params, param, &result); \
     if (ZSTD_isError(zresult)) { \
         PyErr_Format(ZstdError, "unable to get compression parameter: %s", \
             ZSTD_getErrorName(zresult)); \

contrib/python-zstandard/c-ext/decompressionreader.c

@@ -653,6 +653,8 b' static PyObject* reader_seek(ZstdDecompr'
 
 		readSize = PyBytes_GET_SIZE(readResult);
 
+		Py_CLEAR(readResult);
+
 		/* Empty read means EOF. */
 		if (!readSize) {
 			break;

contrib/python-zstandard/c-ext/python-zstandard.h

@@ -16,7 +16,7 b''
 #include <zdict.h>
 
 /* Remember to change the string in zstandard/__init__ as well */
-#define PYTHON_ZSTANDARD_VERSION "0.11.0"
+#define PYTHON_ZSTANDARD_VERSION "0.12.0"
 
 typedef enum {
 	compressorobj_flush_finish,

contrib/python-zstandard/make_cffi.py

@@ -29,6 +29,8 b" SOURCES = ['zstd/%s' % p for p in ("
     'compress/hist.c',
     'compress/huf_compress.c',
     'compress/zstd_compress.c',
+    'compress/zstd_compress_literals.c',
+    'compress/zstd_compress_sequences.c',
     'compress/zstd_double_fast.c',
     'compress/zstd_fast.c',
     'compress/zstd_lazy.c',
@@ -119,7 +121,11 b' def preprocess(path):'
     os.close(fd)
 
     try:
-        process = subprocess.Popen(args + [input_file], stdout=subprocess.PIPE)
+        env = dict(os.environ)
+        if getattr(compiler, '_paths', None):
+            env['PATH'] = compiler._paths
+        process = subprocess.Popen(args + [input_file], stdout=subprocess.PIPE,
+                                   env=env)
         output = process.communicate()[0]
         ret = process.poll()
         if ret:

contrib/python-zstandard/setup.py

@@ -100,7 +100,6 b' setup('
         'License :: OSI Approved :: BSD License',
         'Programming Language :: C',
         'Programming Language :: Python :: 2.7',
-        'Programming Language :: Python :: 3.4',
         'Programming Language :: Python :: 3.5',
         'Programming Language :: Python :: 3.6',
         'Programming Language :: Python :: 3.7',

contrib/python-zstandard/setup_zstd.py

@@ -22,6 +22,8 b" zstd_sources = ['zstd/%s' % p for p in ("
     'compress/fse_compress.c',
     'compress/hist.c',
     'compress/huf_compress.c',
+    'compress/zstd_compress_literals.c',
+    'compress/zstd_compress_sequences.c',
     'compress/zstd_compress.c',
     'compress/zstd_double_fast.c',
     'compress/zstd_fast.c',

contrib/python-zstandard/tests/test_compressor.py

@@ -1038,7 +1038,7 b' class TestCompressor_stream_writer(unitt'
         d = zstd.train_dictionary(8192, samples)
 
         h = hashlib.sha1(d.as_bytes()).hexdigest()
-        self.assertEqual(h, '88ca0d38332aff379d4ced166a51c280a7679aad')
+        self.assertEqual(h, '7a2e59a876db958f74257141045af8f912e00d4e')
 
         buffer = NonClosingBytesIO()
         cctx = zstd.ZstdCompressor(level=9, dict_data=d)
@@ -1056,7 +1056,7 b' class TestCompressor_stream_writer(unitt'
         self.assertFalse(params.has_checksum)
 
         h = hashlib.sha1(compressed).hexdigest()
-        self.assertEqual(h, '8703b4316f274d26697ea5dd480f29c08e85d940')
+        self.assertEqual(h, '0a7c05635061f58039727cdbe76388c6f4cfef06')
 
         source = b'foo' + b'bar' + (b'foo' * 16384)
 
@@ -1091,7 +1091,7 b' class TestCompressor_stream_writer(unitt'
         self.assertFalse(params.has_checksum)
 
         h = hashlib.sha1(compressed).hexdigest()
-        self.assertEqual(h, '2a8111d72eb5004cdcecbdac37da9f26720d30ef')
+        self.assertEqual(h, 'dd4bb7d37c1a0235b38a2f6b462814376843ef0b')
 
     def test_write_checksum(self):
         no_checksum = NonClosingBytesIO()

contrib/python-zstandard/tests/test_data_structures.py

@@ -100,7 +100,7 b' class TestCompressionParameters(unittest'
                                            strategy=zstd.STRATEGY_DFAST)
 
         # 32-bit has slightly different values from 64-bit.
-        self.assertAlmostEqual(p.estimated_compression_context_size(), 1294072,
+        self.assertAlmostEqual(p.estimated_compression_context_size(), 1294144,
                                delta=250)
 
     def test_strategy(self):

contrib/python-zstandard/tests/test_module_attributes.py

@@ -12,9 +12,9 b' from . common import ('
 @make_cffi
 class TestModuleAttributes(unittest.TestCase):
     def test_version(self):
-        self.assertEqual(zstd.ZSTD_VERSION, (1, 3, 8))
+        self.assertEqual(zstd.ZSTD_VERSION, (1, 4, 3))
 
-        self.assertEqual(zstd.__version__, '0.11.0')
+        self.assertEqual(zstd.__version__, '0.12.0')
 
     def test_constants(self):
         self.assertEqual(zstd.MAX_COMPRESSION_LEVEL, 22)

contrib/python-zstandard/tests/test_train_dictionary.py

@@ -7,6 +7,7 b' import zstandard as zstd'
 from . common import (
     generate_samples,
     make_cffi,
+    random_input_data,
 )
 
 if sys.version_info[0] >= 3:
@@ -29,7 +30,7 b' class TestTrainDictionary(unittest.TestC'
             zstd.train_dictionary(8192, [u'foo'])
 
     def test_no_params(self):
-        d = zstd.train_dictionary(8192, generate_samples())
+        d = zstd.train_dictionary(8192, random_input_data())
         self.assertIsInstance(d.dict_id(), int_type)
 
         # The dictionary ID may be different across platforms.

contrib/python-zstandard/zstandard/__init__.py

@@ -62,4 +62,4 b' else:'
                       'cext, or cffi' % _module_policy)
 
 # Keep this in sync with python-zstandard.h.
-__version__ = '0.11.0'
+__version__ = '0.12.0'

contrib/python-zstandard/zstandard/cffi.py

@@ -416,7 +416,7 b' def estimate_decompression_context_size('
 
 
 def _set_compression_parameter(params, param, value):
-    zresult = lib.ZSTD_CCtxParam_setParameter(params, param, value)
+    zresult = lib.ZSTD_CCtxParams_setParameter(params, param, value)
     if lib.ZSTD_isError(zresult):
         raise ZstdError('unable to set compression context parameter: %s' %
                         _zstd_error(zresult))
@@ -425,7 +425,7 b' def _set_compression_parameter(params, p'
 def _get_compression_parameter(params, param):
     result = ffi.new('int *')
 
-    zresult = lib.ZSTD_CCtxParam_getParameter(params, param, result)
+    zresult = lib.ZSTD_CCtxParams_getParameter(params, param, result)
     if lib.ZSTD_isError(zresult):
         raise ZstdError('unable to get compression context parameter: %s' %
                         _zstd_error(zresult))

contrib/python-zstandard/zstd.c

@@ -210,7 +210,7 b' void zstd_module_init(PyObject* m) {'
 	   We detect this mismatch here and refuse to load the module if this
 	   scenario is detected.
 	*/
-	if (ZSTD_VERSION_NUMBER != 10308 || ZSTD_versionNumber() != 10308) {
+	if (ZSTD_VERSION_NUMBER != 10403 || ZSTD_versionNumber() != 10403) {
 		PyErr_SetString(PyExc_ImportError, "zstd C API mismatch; Python bindings not compiled against expected zstd version");
 		return;
 	}

contrib/python-zstandard/zstd/common/bitstream.h

@@ -57,6 +57,8 b' extern "C" {'
 =========================================*/
 #if defined(__BMI__) && defined(__GNUC__)
 #  include <immintrin.h>   /* support for bextr (experimental) */
+#elif defined(__ICCARM__)
+#  include <intrinsics.h>
 #endif
 
 #define STREAM_ACCUMULATOR_MIN_32  25
@@ -163,6 +165,8 b' MEM_STATIC unsigned BIT_highbit32 (U32 v'
         return (unsigned) r;
 #   elif defined(__GNUC__) && (__GNUC__ >= 3)   /* Use GCC Intrinsic */
         return 31 - __builtin_clz (val);
+#   elif defined(__ICCARM__)    /* IAR Intrinsic */
+        return 31 - __CLZ(val);
 #   else   /* Software version */
         static const unsigned DeBruijnClz[32] = { 0,  9,  1, 10, 13, 21,  2, 29,
                                                  11, 14, 16, 18, 22, 25,  3, 30,

contrib/python-zstandard/zstd/common/compiler.h

@@ -23,7 +23,7 b''
 #  define INLINE_KEYWORD
 #endif
 
-#if defined(__GNUC__)
+#if defined(__GNUC__) || defined(__ICCARM__)
 #  define FORCE_INLINE_ATTR __attribute__((always_inline))
 #elif defined(_MSC_VER)
 #  define FORCE_INLINE_ATTR __forceinline
@@ -40,7 +40,7 b''
 
 /**
  * FORCE_INLINE_TEMPLATE is used to define C "templates", which take constant
- * parameters. They must be inlined for the compiler to elimininate the constant
+ * parameters. They must be inlined for the compiler to eliminate the constant
  * branches.
  */
 #define FORCE_INLINE_TEMPLATE static INLINE_KEYWORD FORCE_INLINE_ATTR
@@ -65,7 +65,7 b''
 #ifdef _MSC_VER
 #  define FORCE_NOINLINE static __declspec(noinline)
 #else
-#  ifdef __GNUC__
+#  if defined(__GNUC__) || defined(__ICCARM__)
 #    define FORCE_NOINLINE static __attribute__((__noinline__))
 #  else
 #    define FORCE_NOINLINE static
@@ -76,7 +76,7 b''
 #ifndef __has_attribute
   #define __has_attribute(x) 0  /* Compatibility with non-clang compilers. */
 #endif
-#if defined(__GNUC__)
+#if defined(__GNUC__) || defined(__ICCARM__)
 #  define TARGET_ATTRIBUTE(target) __attribute__((__target__(target)))
 #else
 #  define TARGET_ATTRIBUTE(target)
@@ -127,6 +127,13 b''
     }                                     \
 }
 
+/* vectorization */
+#if !defined(__clang__) && defined(__GNUC__)
+#  define DONT_VECTORIZE __attribute__((optimize("no-tree-vectorize")))
+#else
+#  define DONT_VECTORIZE
+#endif
+
 /* disable warnings */
 #ifdef _MSC_VER    /* Visual Studio */
 #  include <intrin.h>                    /* For Visual 2005 */

contrib/python-zstandard/zstd/common/fse.h

@@ -358,7 +358,7 b' size_t FSE_decompress_wksp(void* dst, si'
 typedef enum {
    FSE_repeat_none,  /**< Cannot use the previous table */
    FSE_repeat_check, /**< Can use the previous table but it must be checked */
-   FSE_repeat_valid  /**< Can use the previous table and it is asumed to be valid */
+   FSE_repeat_valid  /**< Can use the previous table and it is assumed to be valid */
  } FSE_repeat;
 
 /* *****************************************

contrib/python-zstandard/zstd/common/mem.h

@@ -102,7 +102,7 b' MEM_STATIC void MEM_check(void) { MEM_ST'
 #ifndef MEM_FORCE_MEMORY_ACCESS   /* can be defined externally, on command line for example */
 #  if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) )
 #    define MEM_FORCE_MEMORY_ACCESS 2
-#  elif defined(__INTEL_COMPILER) || defined(__GNUC__)
+#  elif defined(__INTEL_COMPILER) || defined(__GNUC__) || defined(__ICCARM__)
 #    define MEM_FORCE_MEMORY_ACCESS 1
 #  endif
 #endif

contrib/python-zstandard/zstd/common/threading.c

@@ -14,8 +14,8 b''
  * This file will hold wrapper for systems, which do not support pthreads
  */
 
-/* create fake symbol to avoid empty trnaslation unit warning */
-int g_ZSTD_threading_useles_symbol;
+/* create fake symbol to avoid empty translation unit warning */
+int g_ZSTD_threading_useless_symbol;
 
 #if defined(ZSTD_MULTITHREAD) && defined(_WIN32)
 

contrib/python-zstandard/zstd/common/xxhash.c

@@ -53,7 +53,8 b''
 #  if defined(__GNUC__) && ( defined(__ARM_ARCH_6__) || defined(__ARM_ARCH_6J__) || defined(__ARM_ARCH_6K__) || defined(__ARM_ARCH_6Z__) || defined(__ARM_ARCH_6ZK__) || defined(__ARM_ARCH_6T2__) )
 #    define XXH_FORCE_MEMORY_ACCESS 2
 #  elif (defined(__INTEL_COMPILER) && !defined(WIN32)) || \
-  (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) ))
+  (defined(__GNUC__) && ( defined(__ARM_ARCH_7__) || defined(__ARM_ARCH_7A__) || defined(__ARM_ARCH_7R__) || defined(__ARM_ARCH_7M__) || defined(__ARM_ARCH_7S__) )) || \
+  defined(__ICCARM__)
 #    define XXH_FORCE_MEMORY_ACCESS 1
 #  endif
 #endif
@@ -66,10 +67,10 b''
 /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */
 
 /*!XXH_FORCE_NATIVE_FORMAT :
- * By default, xxHash library provides endian-independant Hash values, based on little-endian convention.
+ * By default, xxHash library provides endian-independent Hash values, based on little-endian convention.
  * Results are therefore identical for little-endian and big-endian CPU.
  * This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format.
- * Should endian-independance be of no importance for your application, you may set the #define below to 1,
+ * Should endian-independence be of no importance for your application, you may set the #define below to 1,
  * to improve speed for Big-endian CPU.
  * This option has no impact on Little_Endian CPU.
  */
@@ -120,7 +121,7 b' static void* XXH_memcpy(void* dest, cons'
 #  define INLINE_KEYWORD
 #endif
 
-#if defined(__GNUC__)
+#if defined(__GNUC__) || defined(__ICCARM__)
 #  define FORCE_INLINE_ATTR __attribute__((always_inline))
 #elif defined(_MSC_VER)
 #  define FORCE_INLINE_ATTR __forceinline
@@ -206,7 +207,12 b' static U64 XXH_read64(const void* memPtr'
 #  define XXH_rotl32(x,r) _rotl(x,r)
 #  define XXH_rotl64(x,r) _rotl64(x,r)
 #else
+#if defined(__ICCARM__)
+#  include <intrinsics.h>
+#  define XXH_rotl32(x,r) __ROR(x,(32 - r))
+#else
 #  define XXH_rotl32(x,r) ((x << r) | (x >> (32 - r)))
+#endif
 #  define XXH_rotl64(x,r) ((x << r) | (x >> (64 - r)))
 #endif
 

contrib/python-zstandard/zstd/common/zstd_internal.h

@@ -34,7 +34,6 b''
 #endif
 #include "xxhash.h"                /* XXH_reset, update, digest */
 
-
 #if defined (__cplusplus)
 extern "C" {
 #endif
@@ -53,8 +52,50 b' extern "C" {'
 #undef MAX
 #define MIN(a,b) ((a)<(b) ? (a) : (b))
 #define MAX(a,b) ((a)>(b) ? (a) : (b))
-#define CHECK_F(f) { size_t const errcod = f; if (ERR_isError(errcod)) return errcod; }  /* check and Forward error code */
-#define CHECK_E(f, e) { size_t const errcod = f; if (ERR_isError(errcod)) return ERROR(e); }  /* check and send Error code */
+
+/**
+ * Return the specified error if the condition evaluates to true.
+ *
+ * In debug modes, prints additional information.
+ * In order to do that (particularly, printing the conditional that failed),
+ * this can't just wrap RETURN_ERROR().
+ */
+#define RETURN_ERROR_IF(cond, err, ...) \
+  if (cond) { \
+    RAWLOG(3, "%s:%d: ERROR!: check %s failed, returning %s", __FILE__, __LINE__, ZSTD_QUOTE(cond), ZSTD_QUOTE(ERROR(err))); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  }
+
+/**
+ * Unconditionally return the specified error.
+ *
+ * In debug modes, prints additional information.
+ */
+#define RETURN_ERROR(err, ...) \
+  do { \
+    RAWLOG(3, "%s:%d: ERROR!: unconditional check failed, returning %s", __FILE__, __LINE__, ZSTD_QUOTE(ERROR(err))); \
+    RAWLOG(3, ": " __VA_ARGS__); \
+    RAWLOG(3, "\n"); \
+    return ERROR(err); \
+  } while(0);
+
+/**
+ * If the provided expression evaluates to an error code, returns that error code.
+ *
+ * In debug modes, prints additional information.
+ */
+#define FORWARD_IF_ERROR(err, ...) \
+  do { \
+    size_t const err_code = (err); \
+    if (ERR_isError(err_code)) { \
+      RAWLOG(3, "%s:%d: ERROR!: forwarding error in %s: %s", __FILE__, __LINE__, ZSTD_QUOTE(err), ERR_getErrorName(err_code)); \
+      RAWLOG(3, ": " __VA_ARGS__); \
+      RAWLOG(3, "\n"); \
+      return err_code; \
+    } \
+  } while(0);
 
 
 /*-*************************************
@@ -151,19 +192,72 b' static const U32 OF_defaultNormLog = OF_'
 *  Shared functions to include for inlining
 *********************************************/
 static void ZSTD_copy8(void* dst, const void* src) { memcpy(dst, src, 8); }
+
 #define COPY8(d,s) { ZSTD_copy8(d,s); d+=8; s+=8; }
+static void ZSTD_copy16(void* dst, const void* src) { memcpy(dst, src, 16); }
+#define COPY16(d,s) { ZSTD_copy16(d,s); d+=16; s+=16; }
+
+#define WILDCOPY_OVERLENGTH 8
+#define VECLEN 16
+
+typedef enum {
+    ZSTD_no_overlap,
+    ZSTD_overlap_src_before_dst,
+    /*  ZSTD_overlap_dst_before_src, */
+} ZSTD_overlap_e;
 
 /*! ZSTD_wildcopy() :
  *  custom version of memcpy(), can overwrite up to WILDCOPY_OVERLENGTH bytes (if length==0) */
-#define WILDCOPY_OVERLENGTH 8
-MEM_STATIC void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length)
+MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE
+void ZSTD_wildcopy(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype)
 {
+    ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
     const BYTE* ip = (const BYTE*)src;
     BYTE* op = (BYTE*)dst;
     BYTE* const oend = op + length;
-    do
-        COPY8(op, ip)
-    while (op < oend);
+
+    assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8));
+    if (length < VECLEN || (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN)) {
+      do
+          COPY8(op, ip)
+      while (op < oend);
+    }
+    else {
+      if ((length & 8) == 0)
+        COPY8(op, ip);
+      do {
+        COPY16(op, ip);
+      }
+      while (op < oend);
+    }
+}
+
+/*! ZSTD_wildcopy_16min() :
+ *  same semantics as ZSTD_wilcopy() except guaranteed to be able to copy 16 bytes at the start */
+MEM_STATIC FORCE_INLINE_ATTR DONT_VECTORIZE
+void ZSTD_wildcopy_16min(void* dst, const void* src, ptrdiff_t length, ZSTD_overlap_e ovtype)
+{
+    ptrdiff_t diff = (BYTE*)dst - (const BYTE*)src;
+    const BYTE* ip = (const BYTE*)src;
+    BYTE* op = (BYTE*)dst;
+    BYTE* const oend = op + length;
+
+    assert(length >= 8);
+    assert(diff >= 8 || (ovtype == ZSTD_no_overlap && diff < -8));
+
+    if (ovtype == ZSTD_overlap_src_before_dst && diff < VECLEN) {
+      do
+          COPY8(op, ip)
+      while (op < oend);
+    }
+    else {
+      if ((length & 8) == 0)
+        COPY8(op, ip);
+      do {
+        COPY16(op, ip);
+      }
+      while (op < oend);
+    }
 }
 
 MEM_STATIC void ZSTD_wildcopy_e(void* dst, const void* src, void* dstEnd)   /* should be faster for decoding, but strangely, not verified on all platform */
@@ -200,6 +294,17 b' typedef struct {'
     U32   longLengthPos;
 } seqStore_t;
 
+/**
+ * Contains the compressed frame size and an upper-bound for the decompressed frame size.
+ * Note: before using `compressedSize`, check for errors using ZSTD_isError().
+ *       similarly, before using `decompressedBound`, check for errors using:
+ *          `decompressedBound != ZSTD_CONTENTSIZE_ERROR`
+ */
+typedef struct {
+    size_t compressedSize;
+    unsigned long long decompressedBound;
+} ZSTD_frameSizeInfo;   /* decompress & legacy */
+
 const seqStore_t* ZSTD_getSeqStore(const ZSTD_CCtx* ctx);   /* compress & dictBuilder */
 void ZSTD_seqToCodes(const seqStore_t* seqStorePtr);   /* compress, dictBuilder, decodeCorpus (shouldn't get its definition from here) */
 
@@ -219,6 +324,8 b' MEM_STATIC U32 ZSTD_highbit32(U32 val) '
         return (unsigned)r;
 #   elif defined(__GNUC__) && (__GNUC__ >= 3)   /* GCC Intrinsic */
         return 31 - __builtin_clz(val);
+#   elif defined(__ICCARM__)    /* IAR Intrinsic */
+        return 31 - __CLZ(val);
 #   else   /* Software version */
         static const U32 DeBruijnClz[32] = { 0, 9, 1, 10, 13, 21, 2, 29, 11, 14, 16, 18, 22, 25, 3, 30, 8, 12, 20, 28, 15, 17, 24, 7, 19, 27, 23, 6, 26, 5, 4, 31 };
         U32 v = val;

contrib/python-zstandard/zstd/compress/fse_compress.c

@@ -129,9 +129,9 b' size_t FSE_buildCTable_wksp(FSE_CTable* '
     {   U32 position = 0;
         U32 symbol;
         for (symbol=0; symbol<=maxSymbolValue; symbol++) {
-            int nbOccurences;
+            int nbOccurrences;
             int const freq = normalizedCounter[symbol];
-            for (nbOccurences=0; nbOccurences<freq; nbOccurences++) {
+            for (nbOccurrences=0; nbOccurrences<freq; nbOccurrences++) {
                 tableSymbol[position] = (FSE_FUNCTION_TYPE)symbol;
                 position = (position + step) & tableMask;
                 while (position > highThreshold)

contrib/python-zstandard/zstd/compress/zstd_compress.c

@@ -21,6 +21,8 b''
 #define HUF_STATIC_LINKING_ONLY
 #include "huf.h"
 #include "zstd_compress_internal.h"
+#include "zstd_compress_sequences.h"
+#include "zstd_compress_literals.h"
 #include "zstd_fast.h"
 #include "zstd_double_fast.h"
 #include "zstd_lazy.h"
@@ -103,12 +105,31 b' ZSTD_CCtx* ZSTD_initStaticCCtx(void *wor'
     return cctx;
 }
 
+/**
+ * Clears and frees all of the dictionaries in the CCtx.
+ */
+static void ZSTD_clearAllDicts(ZSTD_CCtx* cctx)
+{
+    ZSTD_free(cctx->localDict.dictBuffer, cctx->customMem);
+    ZSTD_freeCDict(cctx->localDict.cdict);
+    memset(&cctx->localDict, 0, sizeof(cctx->localDict));
+    memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));
+    cctx->cdict = NULL;
+}
+
+static size_t ZSTD_sizeof_localDict(ZSTD_localDict dict)
+{
+    size_t const bufferSize = dict.dictBuffer != NULL ? dict.dictSize : 0;
+    size_t const cdictSize = ZSTD_sizeof_CDict(dict.cdict);
+    return bufferSize + cdictSize;
+}
+
 static void ZSTD_freeCCtxContent(ZSTD_CCtx* cctx)
 {
     assert(cctx != NULL);
     assert(cctx->staticSize == 0);
     ZSTD_free(cctx->workSpace, cctx->customMem); cctx->workSpace = NULL;
-    ZSTD_freeCDict(cctx->cdictLocal); cctx->cdictLocal = NULL;
+    ZSTD_clearAllDicts(cctx);
 #ifdef ZSTD_MULTITHREAD
     ZSTDMT_freeCCtx(cctx->mtctx); cctx->mtctx = NULL;
 #endif
@@ -117,7 +138,8 b' static void ZSTD_freeCCtxContent(ZSTD_CC'
 size_t ZSTD_freeCCtx(ZSTD_CCtx* cctx)
 {
     if (cctx==NULL) return 0;   /* support free on NULL */
-    if (cctx->staticSize) return ERROR(memory_allocation);   /* not compatible with static CCtx */
+    RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
+                    "not compatible with static CCtx");
     ZSTD_freeCCtxContent(cctx);
     ZSTD_free(cctx, cctx->customMem);
     return 0;
@@ -139,7 +161,7 b' size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx*'
 {
     if (cctx==NULL) return 0;   /* support sizeof on NULL */
     return sizeof(*cctx) + cctx->workSpaceSize
-           + ZSTD_sizeof_CDict(cctx->cdictLocal)
+           + ZSTD_sizeof_localDict(cctx->localDict)
            + ZSTD_sizeof_mtctx(cctx);
 }
 
@@ -195,7 +217,7 b' size_t ZSTD_CCtxParams_reset(ZSTD_CCtx_p'
 }
 
 size_t ZSTD_CCtxParams_init(ZSTD_CCtx_params* cctxParams, int compressionLevel) {
-    if (!cctxParams) { return ERROR(GENERIC); }
+    RETURN_ERROR_IF(!cctxParams, GENERIC);
     memset(cctxParams, 0, sizeof(*cctxParams));
     cctxParams->compressionLevel = compressionLevel;
     cctxParams->fParams.contentSizeFlag = 1;
@@ -204,8 +226,8 b' size_t ZSTD_CCtxParams_init(ZSTD_CCtx_pa'
 
 size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params)
 {
-    if (!cctxParams) { return ERROR(GENERIC); }
-    CHECK_F( ZSTD_checkCParams(params.cParams) );
+    RETURN_ERROR_IF(!cctxParams, GENERIC);
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );
     memset(cctxParams, 0, sizeof(*cctxParams));
     cctxParams->cParams = params.cParams;
     cctxParams->fParams = params.fParams;
@@ -359,6 +381,17 b' ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_c'
         bounds.upperBound = ZSTD_dictForceCopy;       /* note : how to ensure at compile time that this is the highest value enum ? */
         return bounds;
 
+    case ZSTD_c_literalCompressionMode:
+        ZSTD_STATIC_ASSERT(ZSTD_lcm_auto < ZSTD_lcm_huffman && ZSTD_lcm_huffman < ZSTD_lcm_uncompressed);
+        bounds.lowerBound = ZSTD_lcm_auto;
+        bounds.upperBound = ZSTD_lcm_uncompressed;
+        return bounds;
+
+    case ZSTD_c_targetCBlockSize:
+        bounds.lowerBound = ZSTD_TARGETCBLOCKSIZE_MIN;
+        bounds.upperBound = ZSTD_TARGETCBLOCKSIZE_MAX;
+        return bounds;
+
     default:
         {   ZSTD_bounds const boundError = { ERROR(parameter_unsupported), 0, 0 };
             return boundError;
@@ -366,22 +399,22 b' ZSTD_bounds ZSTD_cParam_getBounds(ZSTD_c'
     }
 }
 
-/* ZSTD_cParam_withinBounds:
- * @return 1 if value is within cParam bounds,
- * 0 otherwise */
-static int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value)
+/* ZSTD_cParam_clampBounds:
+ * Clamps the value into the bounded range.
+ */
+static size_t ZSTD_cParam_clampBounds(ZSTD_cParameter cParam, int* value)
 {
     ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);
-    if (ZSTD_isError(bounds.error)) return 0;
-    if (value < bounds.lowerBound) return 0;
-    if (value > bounds.upperBound) return 0;
-    return 1;
+    if (ZSTD_isError(bounds.error)) return bounds.error;
+    if (*value < bounds.lowerBound) *value = bounds.lowerBound;
+    if (*value > bounds.upperBound) *value = bounds.upperBound;
+    return 0;
 }
 
-#define BOUNDCHECK(cParam, val) {                  \
-    if (!ZSTD_cParam_withinBounds(cParam,val)) {   \
-        return ERROR(parameter_outOfBound);        \
-}   }
+#define BOUNDCHECK(cParam, val) { \
+    RETURN_ERROR_IF(!ZSTD_cParam_withinBounds(cParam,val), \
+                    parameter_outOfBound); \
+}
 
 
 static int ZSTD_isUpdateAuthorized(ZSTD_cParameter param)
@@ -413,6 +446,8 b' static int ZSTD_isUpdateAuthorized(ZSTD_'
     case ZSTD_c_ldmBucketSizeLog:
     case ZSTD_c_ldmHashRateLog:
     case ZSTD_c_forceAttachDict:
+    case ZSTD_c_literalCompressionMode:
+    case ZSTD_c_targetCBlockSize:
     default:
         return 0;
     }
@@ -425,18 +460,17 b' size_t ZSTD_CCtx_setParameter(ZSTD_CCtx*'
         if (ZSTD_isUpdateAuthorized(param)) {
             cctx->cParamsChanged = 1;
         } else {
-            return ERROR(stage_wrong);
+            RETURN_ERROR(stage_wrong);
     }   }
 
     switch(param)
     {
-    case ZSTD_c_format :
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
+    case ZSTD_c_nbWorkers:
+        RETURN_ERROR_IF((value!=0) && cctx->staticSize, parameter_unsupported,
+                        "MT not compatible with static alloc");
+        break;
 
     case ZSTD_c_compressionLevel:
-        if (cctx->cdict) return ERROR(stage_wrong);
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
     case ZSTD_c_windowLog:
     case ZSTD_c_hashLog:
     case ZSTD_c_chainLog:
@@ -444,49 +478,33 b' size_t ZSTD_CCtx_setParameter(ZSTD_CCtx*'
     case ZSTD_c_minMatch:
     case ZSTD_c_targetLength:
     case ZSTD_c_strategy:
-        if (cctx->cdict) return ERROR(stage_wrong);
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
+    case ZSTD_c_ldmHashRateLog:
+    case ZSTD_c_format:
     case ZSTD_c_contentSizeFlag:
     case ZSTD_c_checksumFlag:
     case ZSTD_c_dictIDFlag:
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
-    case ZSTD_c_forceMaxWindow :  /* Force back-references to remain < windowSize,
-                                   * even when referencing into Dictionary content.
-                                   * default : 0 when using a CDict, 1 when using a Prefix */
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
+    case ZSTD_c_forceMaxWindow:
     case ZSTD_c_forceAttachDict:
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
-    case ZSTD_c_nbWorkers:
-        if ((value!=0) && cctx->staticSize) {
-            return ERROR(parameter_unsupported);  /* MT not compatible with static alloc */
-        }
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
+    case ZSTD_c_literalCompressionMode:
     case ZSTD_c_jobSize:
     case ZSTD_c_overlapLog:
     case ZSTD_c_rsyncable:
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
     case ZSTD_c_enableLongDistanceMatching:
     case ZSTD_c_ldmHashLog:
     case ZSTD_c_ldmMinMatch:
     case ZSTD_c_ldmBucketSizeLog:
-    case ZSTD_c_ldmHashRateLog:
-        if (cctx->cdict) return ERROR(stage_wrong);
-        return ZSTD_CCtxParam_setParameter(&cctx->requestedParams, param, value);
-
-    default: return ERROR(parameter_unsupported);
+    case ZSTD_c_targetCBlockSize:
+        break;
+
+    default: RETURN_ERROR(parameter_unsupported);
     }
+    return ZSTD_CCtxParams_setParameter(&cctx->requestedParams, param, value);
 }
 
-size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* CCtxParams,
-                                   ZSTD_cParameter param, int value)
+size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* CCtxParams,
+                                    ZSTD_cParameter param, int value)
 {
-    DEBUGLOG(4, "ZSTD_CCtxParam_setParameter (%i, %i)", (int)param, value);
+    DEBUGLOG(4, "ZSTD_CCtxParams_setParameter (%i, %i)", (int)param, value);
     switch(param)
     {
     case ZSTD_c_format :
@@ -495,11 +513,9 b' size_t ZSTD_CCtxParam_setParameter(ZSTD_'
         return (size_t)CCtxParams->format;
 
     case ZSTD_c_compressionLevel : {
-        int cLevel = value;
-        if (cLevel > ZSTD_maxCLevel()) cLevel = ZSTD_maxCLevel();
-        if (cLevel < ZSTD_minCLevel()) cLevel = ZSTD_minCLevel();
-        if (cLevel) {  /* 0 : does not change current level */
-            CCtxParams->compressionLevel = cLevel;
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));
+        if (value) {  /* 0 : does not change current level */
+            CCtxParams->compressionLevel = value;
         }
         if (CCtxParams->compressionLevel >= 0) return CCtxParams->compressionLevel;
         return 0;  /* return type (size_t) cannot represent negative values */
@@ -573,33 +589,55 b' size_t ZSTD_CCtxParam_setParameter(ZSTD_'
         return CCtxParams->attachDictPref;
     }
 
+    case ZSTD_c_literalCompressionMode : {
+        const ZSTD_literalCompressionMode_e lcm = (ZSTD_literalCompressionMode_e)value;
+        BOUNDCHECK(ZSTD_c_literalCompressionMode, lcm);
+        CCtxParams->literalCompressionMode = lcm;
+        return CCtxParams->literalCompressionMode;
+    }
+
     case ZSTD_c_nbWorkers :
 #ifndef ZSTD_MULTITHREAD
-        if (value!=0) return ERROR(parameter_unsupported);
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
         return 0;
 #else
-        return ZSTDMT_CCtxParam_setNbWorkers(CCtxParams, value);
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));
+        CCtxParams->nbWorkers = value;
+        return CCtxParams->nbWorkers;
 #endif
 
     case ZSTD_c_jobSize :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
 #else
-        return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_jobSize, value);
+        /* Adjust to the minimum non-default value. */
+        if (value != 0 && value < ZSTDMT_JOBSIZE_MIN)
+            value = ZSTDMT_JOBSIZE_MIN;
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(param, &value));
+        assert(value >= 0);
+        CCtxParams->jobSize = value;
+        return CCtxParams->jobSize;
 #endif
 
     case ZSTD_c_overlapLog :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
 #else
-        return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_overlapLog, value);
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));
+        CCtxParams->overlapLog = value;
+        return CCtxParams->overlapLog;
 #endif
 
     case ZSTD_c_rsyncable :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR_IF(value!=0, parameter_unsupported, "not compiled with multithreading");
+        return 0;
 #else
-        return ZSTDMT_CCtxParam_setMTCtxParameter(CCtxParams, ZSTDMT_p_rsyncable, value);
+        FORWARD_IF_ERROR(ZSTD_cParam_clampBounds(ZSTD_c_overlapLog, &value));
+        CCtxParams->rsyncable = value;
+        return CCtxParams->rsyncable;
 #endif
 
     case ZSTD_c_enableLongDistanceMatching :
@@ -625,21 +663,27 b' size_t ZSTD_CCtxParam_setParameter(ZSTD_'
         return CCtxParams->ldmParams.bucketSizeLog;
 
     case ZSTD_c_ldmHashRateLog :
-        if (value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
-            return ERROR(parameter_outOfBound);
+        RETURN_ERROR_IF(value > ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN,
+                        parameter_outOfBound);
         CCtxParams->ldmParams.hashRateLog = value;
         return CCtxParams->ldmParams.hashRateLog;
 
-    default: return ERROR(parameter_unsupported);
+    case ZSTD_c_targetCBlockSize :
+        if (value!=0)   /* 0 ==> default */
+            BOUNDCHECK(ZSTD_c_targetCBlockSize, value);
+        CCtxParams->targetCBlockSize = value;
+        return CCtxParams->targetCBlockSize;
+
+    default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
     }
 }
 
 size_t ZSTD_CCtx_getParameter(ZSTD_CCtx* cctx, ZSTD_cParameter param, int* value)
 {
-    return ZSTD_CCtxParam_getParameter(&cctx->requestedParams, param, value);
+    return ZSTD_CCtxParams_getParameter(&cctx->requestedParams, param, value);
 }
 
-size_t ZSTD_CCtxParam_getParameter(
+size_t ZSTD_CCtxParams_getParameter(
         ZSTD_CCtx_params* CCtxParams, ZSTD_cParameter param, int* value)
 {
     switch(param)
@@ -651,13 +695,13 b' size_t ZSTD_CCtxParam_getParameter('
         *value = CCtxParams->compressionLevel;
         break;
     case ZSTD_c_windowLog :
-        *value = CCtxParams->cParams.windowLog;
+        *value = (int)CCtxParams->cParams.windowLog;
         break;
     case ZSTD_c_hashLog :
-        *value = CCtxParams->cParams.hashLog;
+        *value = (int)CCtxParams->cParams.hashLog;
         break;
     case ZSTD_c_chainLog :
-        *value = CCtxParams->cParams.chainLog;
+        *value = (int)CCtxParams->cParams.chainLog;
         break;
     case ZSTD_c_searchLog :
         *value = CCtxParams->cParams.searchLog;
@@ -686,6 +730,9 b' size_t ZSTD_CCtxParam_getParameter('
     case ZSTD_c_forceAttachDict :
         *value = CCtxParams->attachDictPref;
         break;
+    case ZSTD_c_literalCompressionMode :
+        *value = CCtxParams->literalCompressionMode;
+        break;
     case ZSTD_c_nbWorkers :
 #ifndef ZSTD_MULTITHREAD
         assert(CCtxParams->nbWorkers == 0);
@@ -694,7 +741,7 b' size_t ZSTD_CCtxParam_getParameter('
         break;
     case ZSTD_c_jobSize :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
 #else
         assert(CCtxParams->jobSize <= INT_MAX);
         *value = (int)CCtxParams->jobSize;
@@ -702,14 +749,14 b' size_t ZSTD_CCtxParam_getParameter('
 #endif
     case ZSTD_c_overlapLog :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
 #else
         *value = CCtxParams->overlapLog;
         break;
 #endif
     case ZSTD_c_rsyncable :
 #ifndef ZSTD_MULTITHREAD
-        return ERROR(parameter_unsupported);
+        RETURN_ERROR(parameter_unsupported, "not compiled with multithreading");
 #else
         *value = CCtxParams->rsyncable;
         break;
@@ -729,7 +776,10 b' size_t ZSTD_CCtxParam_getParameter('
     case ZSTD_c_ldmHashRateLog :
         *value = CCtxParams->ldmParams.hashRateLog;
         break;
-    default: return ERROR(parameter_unsupported);
+    case ZSTD_c_targetCBlockSize :
+        *value = (int)CCtxParams->targetCBlockSize;
+        break;
+    default: RETURN_ERROR(parameter_unsupported, "unknown parameter");
     }
     return 0;
 }
@@ -745,8 +795,8 b' size_t ZSTD_CCtx_setParametersUsingCCtxP'
         ZSTD_CCtx* cctx, const ZSTD_CCtx_params* params)
 {
     DEBUGLOG(4, "ZSTD_CCtx_setParametersUsingCCtxParams");
-    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
-    if (cctx->cdict) return ERROR(stage_wrong);
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
+    RETURN_ERROR_IF(cctx->cdict, stage_wrong);
 
     cctx->requestedParams = *params;
     return 0;
@@ -755,33 +805,71 b' size_t ZSTD_CCtx_setParametersUsingCCtxP'
 ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTD_CCtx_setPledgedSrcSize to %u bytes", (U32)pledgedSrcSize);
-    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
     cctx->pledgedSrcSizePlusOne = pledgedSrcSize+1;
     return 0;
 }
 
+/**
+ * Initializes the local dict using the requested parameters.
+ * NOTE: This does not use the pledged src size, because it may be used for more
+ * than one compression.
+ */
+static size_t ZSTD_initLocalDict(ZSTD_CCtx* cctx)
+{
+    ZSTD_localDict* const dl = &cctx->localDict;
+    ZSTD_compressionParameters const cParams = ZSTD_getCParamsFromCCtxParams(
+            &cctx->requestedParams, 0, dl->dictSize);
+    if (dl->dict == NULL) {
+        /* No local dictionary. */
+        assert(dl->dictBuffer == NULL);
+        assert(dl->cdict == NULL);
+        assert(dl->dictSize == 0);
+        return 0;
+    }
+    if (dl->cdict != NULL) {
+        assert(cctx->cdict == dl->cdict);
+        /* Local dictionary already initialized. */
+        return 0;
+    }
+    assert(dl->dictSize > 0);
+    assert(cctx->cdict == NULL);
+    assert(cctx->prefixDict.dict == NULL);
+
+    dl->cdict = ZSTD_createCDict_advanced(
+            dl->dict,
+            dl->dictSize,
+            ZSTD_dlm_byRef,
+            dl->dictContentType,
+            cParams,
+            cctx->customMem);
+    RETURN_ERROR_IF(!dl->cdict, memory_allocation);
+    cctx->cdict = dl->cdict;
+    return 0;
+}
+
 size_t ZSTD_CCtx_loadDictionary_advanced(
         ZSTD_CCtx* cctx, const void* dict, size_t dictSize,
         ZSTD_dictLoadMethod_e dictLoadMethod, ZSTD_dictContentType_e dictContentType)
 {
-    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
-    if (cctx->staticSize) return ERROR(memory_allocation);  /* no malloc for static CCtx */
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
+    RETURN_ERROR_IF(cctx->staticSize, memory_allocation,
+                    "no malloc for static CCtx");
     DEBUGLOG(4, "ZSTD_CCtx_loadDictionary_advanced (size: %u)", (U32)dictSize);
-    ZSTD_freeCDict(cctx->cdictLocal);  /* in case one already exists */
-    if (dict==NULL || dictSize==0) {   /* no dictionary mode */
-        cctx->cdictLocal = NULL;
-        cctx->cdict = NULL;
+    ZSTD_clearAllDicts(cctx);  /* in case one already exists */
+    if (dict == NULL || dictSize == 0)  /* no dictionary mode */
+        return 0;
+    if (dictLoadMethod == ZSTD_dlm_byRef) {
+        cctx->localDict.dict = dict;
     } else {
-        ZSTD_compressionParameters const cParams =
-                ZSTD_getCParamsFromCCtxParams(&cctx->requestedParams, cctx->pledgedSrcSizePlusOne-1, dictSize);
-        cctx->cdictLocal = ZSTD_createCDict_advanced(
-                                dict, dictSize,
-                                dictLoadMethod, dictContentType,
-                                cParams, cctx->customMem);
-        cctx->cdict = cctx->cdictLocal;
-        if (cctx->cdictLocal == NULL)
-            return ERROR(memory_allocation);
+        void* dictBuffer = ZSTD_malloc(dictSize, cctx->customMem);
+        RETURN_ERROR_IF(!dictBuffer, memory_allocation);
+        memcpy(dictBuffer, dict, dictSize);
+        cctx->localDict.dictBuffer = dictBuffer;
+        cctx->localDict.dict = dictBuffer;
     }
+    cctx->localDict.dictSize = dictSize;
+    cctx->localDict.dictContentType = dictContentType;
     return 0;
 }
 
@@ -801,9 +889,10 b' ZSTDLIB_API size_t ZSTD_CCtx_loadDiction'
 
 size_t ZSTD_CCtx_refCDict(ZSTD_CCtx* cctx, const ZSTD_CDict* cdict)
 {
-    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
+    /* Free the existing local cdict (if any) to save memory. */
+    ZSTD_clearAllDicts(cctx);
     cctx->cdict = cdict;
-    memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));  /* exclusive */
     return 0;
 }
 
@@ -815,8 +904,8 b' size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cc'
 size_t ZSTD_CCtx_refPrefix_advanced(
         ZSTD_CCtx* cctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
 {
-    if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
-    cctx->cdict = NULL;   /* prefix discards any prior cdict */
+    RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
+    ZSTD_clearAllDicts(cctx);
     cctx->prefixDict.dict = prefix;
     cctx->prefixDict.dictSize = prefixSize;
     cctx->prefixDict.dictContentType = dictContentType;
@@ -834,8 +923,8 b' size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, '
     }
     if ( (reset == ZSTD_reset_parameters)
       || (reset == ZSTD_reset_session_and_parameters) ) {
-        if (cctx->streamStage != zcss_init) return ERROR(stage_wrong);
-        cctx->cdict = NULL;
+        RETURN_ERROR_IF(cctx->streamStage != zcss_init, stage_wrong);
+        ZSTD_clearAllDicts(cctx);
         return ZSTD_CCtxParams_reset(&cctx->requestedParams);
     }
     return 0;
@@ -847,12 +936,12 b' size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, '
     @return : 0, or an error code if one value is beyond authorized range */
 size_t ZSTD_checkCParams(ZSTD_compressionParameters cParams)
 {
-    BOUNDCHECK(ZSTD_c_windowLog, cParams.windowLog);
-    BOUNDCHECK(ZSTD_c_chainLog,  cParams.chainLog);
-    BOUNDCHECK(ZSTD_c_hashLog,   cParams.hashLog);
-    BOUNDCHECK(ZSTD_c_searchLog, cParams.searchLog);
-    BOUNDCHECK(ZSTD_c_minMatch,  cParams.minMatch);
-    BOUNDCHECK(ZSTD_c_targetLength,cParams.targetLength);
+    BOUNDCHECK(ZSTD_c_windowLog, (int)cParams.windowLog);
+    BOUNDCHECK(ZSTD_c_chainLog,  (int)cParams.chainLog);
+    BOUNDCHECK(ZSTD_c_hashLog,   (int)cParams.hashLog);
+    BOUNDCHECK(ZSTD_c_searchLog, (int)cParams.searchLog);
+    BOUNDCHECK(ZSTD_c_minMatch,  (int)cParams.minMatch);
+    BOUNDCHECK(ZSTD_c_targetLength,(int)cParams.targetLength);
     BOUNDCHECK(ZSTD_c_strategy,  cParams.strategy);
     return 0;
 }
@@ -868,7 +957,7 b' ZSTD_clampCParams(ZSTD_compressionParame'
         if ((int)val<bounds.lowerBound) val=(type)bounds.lowerBound;      \
         else if ((int)val>bounds.upperBound) val=(type)bounds.upperBound; \
     }
-#   define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, int)
+#   define CLAMP(cParam, val) CLAMP_TYPE(cParam, val, unsigned)
     CLAMP(ZSTD_c_windowLog, cParams.windowLog);
     CLAMP(ZSTD_c_chainLog,  cParams.chainLog);
     CLAMP(ZSTD_c_hashLog,   cParams.hashLog);
@@ -888,10 +977,11 b' static U32 ZSTD_cycleLog(U32 hashLog, ZS'
 }
 
 /** ZSTD_adjustCParams_internal() :
-    optimize `cPar` for a given input (`srcSize` and `dictSize`).
-    mostly downsizing to reduce memory consumption and initialization latency.
-    Both `srcSize` and `dictSize` are optional (use 0 if unknown).
-    Note : cPar is assumed validated. Use ZSTD_checkCParams() to ensure this condition. */
+ *  optimize `cPar` for a specified input (`srcSize` and `dictSize`).
+ *  mostly downsize to reduce memory consumption and initialization latency.
+ * `srcSize` can be ZSTD_CONTENTSIZE_UNKNOWN when not known.
+ *  note : for the time being, `srcSize==0` means "unknown" too, for compatibility with older convention.
+ *  condition : cPar is presumed validated (can be checked using ZSTD_checkCParams()). */
 static ZSTD_compressionParameters
 ZSTD_adjustCParams_internal(ZSTD_compressionParameters cPar,
                             unsigned long long srcSize,
@@ -901,7 +991,7 b' ZSTD_adjustCParams_internal(ZSTD_compres'
     static const U64 maxWindowResize = 1ULL << (ZSTD_WINDOWLOG_MAX-1);
     assert(ZSTD_checkCParams(cPar)==0);
 
-    if (dictSize && (srcSize+1<2) /* srcSize unknown */ )
+    if (dictSize && (srcSize+1<2) /* ZSTD_CONTENTSIZE_UNKNOWN and 0 mean "unknown" */ )
         srcSize = minSrcSize;  /* presumed small when there is a dictionary */
     else if (srcSize == 0)
         srcSize = ZSTD_CONTENTSIZE_UNKNOWN;  /* 0 == unknown : presumed large */
@@ -922,7 +1012,7 b' ZSTD_adjustCParams_internal(ZSTD_compres'
     }
 
     if (cPar.windowLog < ZSTD_WINDOWLOG_ABSOLUTEMIN)
-        cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* required for frame header */
+        cPar.windowLog = ZSTD_WINDOWLOG_ABSOLUTEMIN;  /* minimum wlog required for valid frame header */
 
     return cPar;
 }
@@ -932,7 +1022,7 b' ZSTD_adjustCParams(ZSTD_compressionParam'
                    unsigned long long srcSize,
                    size_t dictSize)
 {
-    cPar = ZSTD_clampCParams(cPar);
+    cPar = ZSTD_clampCParams(cPar);   /* resulting cPar is necessarily valid (all parameters within range) */
     return ZSTD_adjustCParams_internal(cPar, srcSize, dictSize);
 }
 
@@ -973,8 +1063,7 b' ZSTD_sizeof_matchState(const ZSTD_compre'
 
 size_t ZSTD_estimateCCtxSize_usingCCtxParams(const ZSTD_CCtx_params* params)
 {
-    /* Estimate CCtx size is supported for single-threaded compression only. */
-    if (params->nbWorkers > 0) { return ERROR(GENERIC); }
+    RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
     {   ZSTD_compressionParameters const cParams =
                 ZSTD_getCParamsFromCCtxParams(params, 0, 0);
         size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
@@ -1022,10 +1111,12 b' size_t ZSTD_estimateCCtxSize(int compres'
 
 size_t ZSTD_estimateCStreamSize_usingCCtxParams(const ZSTD_CCtx_params* params)
 {
-    if (params->nbWorkers > 0) { return ERROR(GENERIC); }
-    {   size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params);
-        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << params->cParams.windowLog);
-        size_t const inBuffSize = ((size_t)1 << params->cParams.windowLog) + blockSize;
+    RETURN_ERROR_IF(params->nbWorkers > 0, GENERIC, "Estimate CCtx size is supported for single-threaded compression only.");
+    {   ZSTD_compressionParameters const cParams =
+                ZSTD_getCParamsFromCCtxParams(params, 0, 0);
+        size_t const CCtxSize = ZSTD_estimateCCtxSize_usingCCtxParams(params);
+        size_t const blockSize = MIN(ZSTD_BLOCKSIZE_MAX, (size_t)1 << cParams.windowLog);
+        size_t const inBuffSize = ((size_t)1 << cParams.windowLog) + blockSize;
         size_t const outBuffSize = ZSTD_compressBound(blockSize) + 1;
         size_t const streamingSize = inBuffSize + outBuffSize;
 
@@ -1197,15 +1288,14 b' static void ZSTD_reset_compressedBlockSt'
 }
 
 /*! ZSTD_invalidateMatchState()
- * Invalidate all the matches in the match finder tables.
- * Requires nextSrc and base to be set (can be NULL).
+ *  Invalidate all the matches in the match finder tables.
+ *  Requires nextSrc and base to be set (can be NULL).
  */
 static void ZSTD_invalidateMatchState(ZSTD_matchState_t* ms)
 {
     ZSTD_window_clear(&ms->window);
 
     ms->nextToUpdate = ms->window.dictLimit;
-    ms->nextToUpdate3 = ms->window.dictLimit;
     ms->loadedDictEnd = 0;
     ms->opt.litLengthSum = 0;  /* force reset of btopt stats */
     ms->dictMatchState = NULL;
@@ -1242,15 +1332,17 b' static size_t ZSTD_continueCCtx(ZSTD_CCt'
 
 typedef enum { ZSTDcrp_continue, ZSTDcrp_noMemset } ZSTD_compResetPolicy_e;
 
+typedef enum { ZSTD_resetTarget_CDict, ZSTD_resetTarget_CCtx } ZSTD_resetTarget_e;
+
 static void*
 ZSTD_reset_matchState(ZSTD_matchState_t* ms,
                       void* ptr,
                 const ZSTD_compressionParameters* cParams,
-                      ZSTD_compResetPolicy_e const crp, U32 const forCCtx)
+                      ZSTD_compResetPolicy_e const crp, ZSTD_resetTarget_e const forWho)
 {
     size_t const chainSize = (cParams->strategy == ZSTD_fast) ? 0 : ((size_t)1 << cParams->chainLog);
     size_t const hSize = ((size_t)1) << cParams->hashLog;
-    U32    const hashLog3 = (forCCtx && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;
+    U32    const hashLog3 = ((forWho == ZSTD_resetTarget_CCtx) && cParams->minMatch==3) ? MIN(ZSTD_HASHLOG3_MAX, cParams->windowLog) : 0;
     size_t const h3Size = ((size_t)1) << hashLog3;
     size_t const tableSpace = (chainSize + hSize + h3Size) * sizeof(U32);
 
@@ -1264,7 +1356,7 b' ZSTD_reset_matchState(ZSTD_matchState_t*'
     ZSTD_invalidateMatchState(ms);
 
     /* opt parser space */
-    if (forCCtx && (cParams->strategy >= ZSTD_btopt)) {
+    if ((forWho == ZSTD_resetTarget_CCtx) && (cParams->strategy >= ZSTD_btopt)) {
         DEBUGLOG(4, "reserving optimal parser space");
         ms->opt.litFreq = (unsigned*)ptr;
         ms->opt.litLengthFreq = ms->opt.litFreq + (1<<Litbits);
@@ -1292,6 +1384,19 b' ZSTD_reset_matchState(ZSTD_matchState_t*'
     return ptr;
 }
 
+/* ZSTD_indexTooCloseToMax() :
+ * minor optimization : prefer memset() rather than reduceIndex()
+ * which is measurably slow in some circumstances (reported for Visual Studio).
+ * Works when re-using a context for a lot of smallish inputs :
+ * if all inputs are smaller than ZSTD_INDEXOVERFLOW_MARGIN,
+ * memset() will be triggered before reduceIndex().
+ */
+#define ZSTD_INDEXOVERFLOW_MARGIN (16 MB)
+static int ZSTD_indexTooCloseToMax(ZSTD_window_t w)
+{
+    return (size_t)(w.nextSrc - w.base) > (ZSTD_CURRENT_MAX - ZSTD_INDEXOVERFLOW_MARGIN);
+}
+
 #define ZSTD_WORKSPACETOOLARGE_FACTOR 3 /* define "workspace is too large" as this number of times larger than needed */
 #define ZSTD_WORKSPACETOOLARGE_MAXDURATION 128  /* when workspace is continuously too large
                                          * during at least this number of times,
@@ -1303,7 +1408,7 b' ZSTD_reset_matchState(ZSTD_matchState_t*'
     note : `params` are assumed fully validated at this stage */
 static size_t ZSTD_resetCCtx_internal(ZSTD_CCtx* zc,
                                       ZSTD_CCtx_params params,
-                                      U64 pledgedSrcSize,
+                                      U64 const pledgedSrcSize,
                                       ZSTD_compResetPolicy_e const crp,
                                       ZSTD_buffered_policy_e const zbuff)
 {
@@ -1315,13 +1420,21 b' static size_t ZSTD_resetCCtx_internal(ZS'
         if (ZSTD_equivalentParams(zc->appliedParams, params,
                                   zc->inBuffSize,
                                   zc->seqStore.maxNbSeq, zc->seqStore.maxNbLit,
-                                  zbuff, pledgedSrcSize)) {
-            DEBUGLOG(4, "ZSTD_equivalentParams()==1 -> continue mode (wLog1=%u, blockSize1=%zu)",
-                        zc->appliedParams.cParams.windowLog, zc->blockSize);
+                                  zbuff, pledgedSrcSize) ) {
+            DEBUGLOG(4, "ZSTD_equivalentParams()==1 -> consider continue mode");
             zc->workSpaceOversizedDuration += (zc->workSpaceOversizedDuration > 0);   /* if it was too large, it still is */
-            if (zc->workSpaceOversizedDuration <= ZSTD_WORKSPACETOOLARGE_MAXDURATION)
+            if (zc->workSpaceOversizedDuration <= ZSTD_WORKSPACETOOLARGE_MAXDURATION) {
+                DEBUGLOG(4, "continue mode confirmed (wLog1=%u, blockSize1=%zu)",
+                            zc->appliedParams.cParams.windowLog, zc->blockSize);
+                if (ZSTD_indexTooCloseToMax(zc->blockState.matchState.window)) {
+                    /* prefer a reset, faster than a rescale */
+                    ZSTD_reset_matchState(&zc->blockState.matchState,
+                                           zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32,
+                                          &params.cParams,
+                                           crp, ZSTD_resetTarget_CCtx);
+                }
                 return ZSTD_continueCCtx(zc, params, pledgedSrcSize);
-    }   }
+    }   }   }
     DEBUGLOG(4, "ZSTD_equivalentParams()==0 -> reset CCtx");
 
     if (params.ldmParams.enableLdm) {
@@ -1364,16 +1477,16 b' static size_t ZSTD_resetCCtx_internal(ZS'
             DEBUGLOG(4, "windowSize: %zu - blockSize: %zu", windowSize, blockSize);
 
             if (workSpaceTooSmall || workSpaceWasteful) {
-                DEBUGLOG(4, "Need to resize workSpaceSize from %zuKB to %zuKB",
+                DEBUGLOG(4, "Resize workSpaceSize from %zuKB to %zuKB",
                             zc->workSpaceSize >> 10,
                             neededSpace >> 10);
-                /* static cctx : no resize, error out */
-                if (zc->staticSize) return ERROR(memory_allocation);
+
+                RETURN_ERROR_IF(zc->staticSize, memory_allocation, "static cctx : no resize");
 
                 zc->workSpaceSize = 0;
                 ZSTD_free(zc->workSpace, zc->customMem);
                 zc->workSpace = ZSTD_malloc(neededSpace, zc->customMem);
-                if (zc->workSpace == NULL) return ERROR(memory_allocation);
+                RETURN_ERROR_IF(zc->workSpace == NULL, memory_allocation);
                 zc->workSpaceSize = neededSpace;
                 zc->workSpaceOversizedDuration = 0;
 
@@ -1406,7 +1519,10 b' static size_t ZSTD_resetCCtx_internal(ZS'
 
         ZSTD_reset_compressedBlockState(zc->blockState.prevCBlock);
 
-        ptr = zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32;
+        ptr = ZSTD_reset_matchState(&zc->blockState.matchState,
+                                     zc->entropyWorkspace + HUF_WORKSPACE_SIZE_U32,
+                                    &params.cParams,
+                                     crp, ZSTD_resetTarget_CCtx);
 
         /* ldm hash table */
         /* initialize bucketOffsets table later for pointer alignment */
@@ -1424,8 +1540,6 b' static size_t ZSTD_resetCCtx_internal(ZS'
         }
         assert(((size_t)ptr & 3) == 0); /* ensure ptr is properly aligned */
 
-        ptr = ZSTD_reset_matchState(&zc->blockState.matchState, ptr, &params.cParams, crp, /* forCCtx */ 1);
-
         /* sequences storage */
         zc->seqStore.maxNbSeq = maxNbSeq;
         zc->seqStore.sequencesStart = (seqDef*)ptr;
@@ -1502,15 +1616,14 b' static int ZSTD_shouldAttachDict(const Z'
                                  * handled in _enforceMaxDist */
 }
 
-static size_t ZSTD_resetCCtx_byAttachingCDict(
-    ZSTD_CCtx* cctx,
-    const ZSTD_CDict* cdict,
-    ZSTD_CCtx_params params,
-    U64 pledgedSrcSize,
-    ZSTD_buffered_policy_e zbuff)
+static size_t
+ZSTD_resetCCtx_byAttachingCDict(ZSTD_CCtx* cctx,
+                        const ZSTD_CDict* cdict,
+                        ZSTD_CCtx_params params,
+                        U64 pledgedSrcSize,
+                        ZSTD_buffered_policy_e zbuff)
 {
-    {
-        const ZSTD_compressionParameters *cdict_cParams = &cdict->matchState.cParams;
+    {   const ZSTD_compressionParameters* const cdict_cParams = &cdict->matchState.cParams;
         unsigned const windowLog = params.cParams.windowLog;
         assert(windowLog != 0);
         /* Resize working context table params for input only, since the dict
@@ -1522,8 +1635,7 b' static size_t ZSTD_resetCCtx_byAttaching'
         assert(cctx->appliedParams.cParams.strategy == cdict_cParams->strategy);
     }
 
-    {
-        const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc
+    {   const U32 cdictEnd = (U32)( cdict->matchState.window.nextSrc
                                   - cdict->matchState.window.base);
         const U32 cdictLen = cdictEnd - cdict->matchState.window.dictLimit;
         if (cdictLen == 0) {
@@ -1540,9 +1652,9 b' static size_t ZSTD_resetCCtx_byAttaching'
                     cctx->blockState.matchState.window.base + cdictEnd;
                 ZSTD_window_clear(&cctx->blockState.matchState.window);
             }
+            /* loadedDictEnd is expressed within the referential of the active context */
             cctx->blockState.matchState.loadedDictEnd = cctx->blockState.matchState.window.dictLimit;
-        }
-    }
+    }   }
 
     cctx->dictID = cdict->dictID;
 
@@ -1596,7 +1708,6 b' static size_t ZSTD_resetCCtx_byCopyingCD'
         ZSTD_matchState_t* dstMatchState = &cctx->blockState.matchState;
         dstMatchState->window       = srcMatchState->window;
         dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
-        dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;
         dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
     }
 
@@ -1644,7 +1755,7 b' static size_t ZSTD_copyCCtx_internal(ZST'
                             ZSTD_buffered_policy_e zbuff)
 {
     DEBUGLOG(5, "ZSTD_copyCCtx_internal");
-    if (srcCCtx->stage!=ZSTDcs_init) return ERROR(stage_wrong);
+    RETURN_ERROR_IF(srcCCtx->stage!=ZSTDcs_init, stage_wrong);
 
     memcpy(&dstCCtx->customMem, &srcCCtx->customMem, sizeof(ZSTD_customMem));
     {   ZSTD_CCtx_params params = dstCCtx->requestedParams;
@@ -1676,7 +1787,6 b' static size_t ZSTD_copyCCtx_internal(ZST'
         ZSTD_matchState_t* dstMatchState = &dstCCtx->blockState.matchState;
         dstMatchState->window       = srcMatchState->window;
         dstMatchState->nextToUpdate = srcMatchState->nextToUpdate;
-        dstMatchState->nextToUpdate3= srcMatchState->nextToUpdate3;
         dstMatchState->loadedDictEnd= srcMatchState->loadedDictEnd;
     }
     dstCCtx->dictID = srcCCtx->dictID;
@@ -1746,16 +1856,15 b' static void ZSTD_reduceTable_btlazy2(U32'
 
 /*! ZSTD_reduceIndex() :
 *   rescale all indexes to avoid future overflow (indexes are U32) */
-static void ZSTD_reduceIndex (ZSTD_CCtx* zc, const U32 reducerValue)
+static void ZSTD_reduceIndex (ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, const U32 reducerValue)
 {
-    ZSTD_matchState_t* const ms = &zc->blockState.matchState;
-    {   U32 const hSize = (U32)1 << zc->appliedParams.cParams.hashLog;
+    {   U32 const hSize = (U32)1 << params->cParams.hashLog;
         ZSTD_reduceTable(ms->hashTable, hSize, reducerValue);
     }
 
-    if (zc->appliedParams.cParams.strategy != ZSTD_fast) {
-        U32 const chainSize = (U32)1 << zc->appliedParams.cParams.chainLog;
-        if (zc->appliedParams.cParams.strategy == ZSTD_btlazy2)
+    if (params->cParams.strategy != ZSTD_fast) {
+        U32 const chainSize = (U32)1 << params->cParams.chainLog;
+        if (params->cParams.strategy == ZSTD_btlazy2)
             ZSTD_reduceTable_btlazy2(ms->chainTable, chainSize, reducerValue);
         else
             ZSTD_reduceTable(ms->chainTable, chainSize, reducerValue);
@@ -1777,161 +1886,13 b' static void ZSTD_reduceIndex (ZSTD_CCtx*'
 static size_t ZSTD_noCompressBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize, U32 lastBlock)
 {
     U32 const cBlockHeader24 = lastBlock + (((U32)bt_raw)<<1) + (U32)(srcSize << 3);
-    if (srcSize + ZSTD_blockHeaderSize > dstCapacity) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF(srcSize + ZSTD_blockHeaderSize > dstCapacity,
+                    dstSize_tooSmall);
     MEM_writeLE24(dst, cBlockHeader24);
     memcpy((BYTE*)dst + ZSTD_blockHeaderSize, src, srcSize);
     return ZSTD_blockHeaderSize + srcSize;
 }
 
-static size_t ZSTD_noCompressLiterals (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
-{
-    BYTE* const ostart = (BYTE* const)dst;
-    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
-
-    if (srcSize + flSize > dstCapacity) return ERROR(dstSize_tooSmall);
-
-    switch(flSize)
-    {
-        case 1: /* 2 - 1 - 5 */
-            ostart[0] = (BYTE)((U32)set_basic + (srcSize<<3));
-            break;
-        case 2: /* 2 - 2 - 12 */
-            MEM_writeLE16(ostart, (U16)((U32)set_basic + (1<<2) + (srcSize<<4)));
-            break;
-        case 3: /* 2 - 2 - 20 */
-            MEM_writeLE32(ostart, (U32)((U32)set_basic + (3<<2) + (srcSize<<4)));
-            break;
-        default:   /* not necessary : flSize is {1,2,3} */
-            assert(0);
-    }
-
-    memcpy(ostart + flSize, src, srcSize);
-    return srcSize + flSize;
-}
-
-static size_t ZSTD_compressRleLiteralsBlock (void* dst, size_t dstCapacity, const void* src, size_t srcSize)
-{
-    BYTE* const ostart = (BYTE* const)dst;
-    U32   const flSize = 1 + (srcSize>31) + (srcSize>4095);
-
-    (void)dstCapacity;  /* dstCapacity already guaranteed to be >=4, hence large enough */
-
-    switch(flSize)
-    {
-        case 1: /* 2 - 1 - 5 */
-            ostart[0] = (BYTE)((U32)set_rle + (srcSize<<3));
-            break;
-        case 2: /* 2 - 2 - 12 */
-            MEM_writeLE16(ostart, (U16)((U32)set_rle + (1<<2) + (srcSize<<4)));
-            break;
-        case 3: /* 2 - 2 - 20 */
-            MEM_writeLE32(ostart, (U32)((U32)set_rle + (3<<2) + (srcSize<<4)));
-            break;
-        default:   /* not necessary : flSize is {1,2,3} */
-            assert(0);
-    }
-
-    ostart[flSize] = *(const BYTE*)src;
-    return flSize+1;
-}
-
-
-/* ZSTD_minGain() :
- * minimum compression required
- * to generate a compress block or a compressed literals section.
- * note : use same formula for both situations */
-static size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat)
-{
-    U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6;
-    ZSTD_STATIC_ASSERT(ZSTD_btultra == 8);
-    assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));
-    return (srcSize >> minlog) + 2;
-}
-
-static size_t ZSTD_compressLiterals (ZSTD_hufCTables_t const* prevHuf,
-                                     ZSTD_hufCTables_t* nextHuf,
-                                     ZSTD_strategy strategy, int disableLiteralCompression,
-                                     void* dst, size_t dstCapacity,
-                               const void* src, size_t srcSize,
-                                     void* workspace, size_t wkspSize,
-                               const int bmi2)
-{
-    size_t const minGain = ZSTD_minGain(srcSize, strategy);
-    size_t const lhSize = 3 + (srcSize >= 1 KB) + (srcSize >= 16 KB);
-    BYTE*  const ostart = (BYTE*)dst;
-    U32 singleStream = srcSize < 256;
-    symbolEncodingType_e hType = set_compressed;
-    size_t cLitSize;
-
-    DEBUGLOG(5,"ZSTD_compressLiterals (disableLiteralCompression=%i)",
-                disableLiteralCompression);
-
-    /* Prepare nextEntropy assuming reusing the existing table */
-    memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
-
-    if (disableLiteralCompression)
-        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-
-    /* small ? don't even attempt compression (speed opt) */
-#   define COMPRESS_LITERALS_SIZE_MIN 63
-    {   size_t const minLitSize = (prevHuf->repeatMode == HUF_repeat_valid) ? 6 : COMPRESS_LITERALS_SIZE_MIN;
-        if (srcSize <= minLitSize) return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-    }
-
-    if (dstCapacity < lhSize+1) return ERROR(dstSize_tooSmall);   /* not enough space for compression */
-    {   HUF_repeat repeat = prevHuf->repeatMode;
-        int const preferRepeat = 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,
-                                      workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2)
-                                : HUF_compress4X_repeat(ostart+lhSize, dstCapacity-lhSize, src, srcSize, 255, 11,
-                                      workspace, wkspSize, (HUF_CElt*)nextHuf->CTable, &repeat, preferRepeat, bmi2);
-        if (repeat != HUF_repeat_none) {
-            /* reused the existing table */
-            hType = set_repeat;
-        }
-    }
-
-    if ((cLitSize==0) | (cLitSize >= srcSize - minGain) | ERR_isError(cLitSize)) {
-        memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
-        return ZSTD_noCompressLiterals(dst, dstCapacity, src, srcSize);
-    }
-    if (cLitSize==1) {
-        memcpy(nextHuf, prevHuf, sizeof(*prevHuf));
-        return ZSTD_compressRleLiteralsBlock(dst, dstCapacity, src, srcSize);
-    }
-
-    if (hType == set_compressed) {
-        /* using a newly constructed table */
-        nextHuf->repeatMode = HUF_repeat_check;
-    }
-
-    /* Build header */
-    switch(lhSize)
-    {
-    case 3: /* 2 - 2 - 10 - 10 */
-        {   U32 const lhc = hType + ((!singleStream) << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<14);
-            MEM_writeLE24(ostart, lhc);
-            break;
-        }
-    case 4: /* 2 - 2 - 14 - 14 */
-        {   U32 const lhc = hType + (2 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<18);
-            MEM_writeLE32(ostart, lhc);
-            break;
-        }
-    case 5: /* 2 - 2 - 18 - 18 */
-        {   U32 const lhc = hType + (3 << 2) + ((U32)srcSize<<4) + ((U32)cLitSize<<22);
-            MEM_writeLE32(ostart, lhc);
-            ostart[4] = (BYTE)(cLitSize >> 10);
-            break;
-        }
-    default:  /* not possible : lhSize is {3,4,5} */
-        assert(0);
-    }
-    return lhSize+cLitSize;
-}
-
-
 void ZSTD_seqToCodes(const seqStore_t* seqStorePtr)
 {
     const seqDef* const sequences = seqStorePtr->sequencesStart;
@@ -1954,418 +1915,19 b' void ZSTD_seqToCodes(const seqStore_t* s'
         mlCodeTable[seqStorePtr->longLengthPos] = MaxML;
 }
 
-
-/**
- * -log2(x / 256) lookup table for x in [0, 256).
- * If x == 0: Return 0
- * Else: Return floor(-log2(x / 256) * 256)
- */
-static unsigned const kInverseProbabiltyLog256[256] = {
-    0,    2048, 1792, 1642, 1536, 1453, 1386, 1329, 1280, 1236, 1197, 1162,
-    1130, 1100, 1073, 1047, 1024, 1001, 980,  960,  941,  923,  906,  889,
-    874,  859,  844,  830,  817,  804,  791,  779,  768,  756,  745,  734,
-    724,  714,  704,  694,  685,  676,  667,  658,  650,  642,  633,  626,
-    618,  610,  603,  595,  588,  581,  574,  567,  561,  554,  548,  542,
-    535,  529,  523,  517,  512,  506,  500,  495,  489,  484,  478,  473,
-    468,  463,  458,  453,  448,  443,  438,  434,  429,  424,  420,  415,
-    411,  407,  402,  398,  394,  390,  386,  382,  377,  373,  370,  366,
-    362,  358,  354,  350,  347,  343,  339,  336,  332,  329,  325,  322,
-    318,  315,  311,  308,  305,  302,  298,  295,  292,  289,  286,  282,
-    279,  276,  273,  270,  267,  264,  261,  258,  256,  253,  250,  247,
-    244,  241,  239,  236,  233,  230,  228,  225,  222,  220,  217,  215,
-    212,  209,  207,  204,  202,  199,  197,  194,  192,  190,  187,  185,
-    182,  180,  178,  175,  173,  171,  168,  166,  164,  162,  159,  157,
-    155,  153,  151,  149,  146,  144,  142,  140,  138,  136,  134,  132,
-    130,  128,  126,  123,  121,  119,  117,  115,  114,  112,  110,  108,
-    106,  104,  102,  100,  98,   96,   94,   93,   91,   89,   87,   85,
-    83,   82,   80,   78,   76,   74,   73,   71,   69,   67,   66,   64,
-    62,   61,   59,   57,   55,   54,   52,   50,   49,   47,   46,   44,
-    42,   41,   39,   37,   36,   34,   33,   31,   30,   28,   26,   25,
-    23,   22,   20,   19,   17,   16,   14,   13,   11,   10,   8,    7,
-    5,    4,    2,    1,
-};
-
-
-/**
- * Returns the cost in bits of encoding the distribution described by count
- * using the entropy bound.
- */
-static size_t ZSTD_entropyCost(unsigned const* count, unsigned const max, size_t const total)
-{
-    unsigned cost = 0;
-    unsigned s;
-    for (s = 0; s <= max; ++s) {
-        unsigned norm = (unsigned)((256 * count[s]) / total);
-        if (count[s] != 0 && norm == 0)
-            norm = 1;
-        assert(count[s] < total);
-        cost += count[s] * kInverseProbabiltyLog256[norm];
-    }
-    return cost >> 8;
-}
-
-
-/**
- * Returns the cost in bits of encoding the distribution in count using the
- * table described by norm. The max symbol support by norm is assumed >= max.
- * norm must be valid for every symbol with non-zero probability in count.
- */
-static size_t ZSTD_crossEntropyCost(short const* norm, unsigned accuracyLog,
-                                    unsigned const* count, unsigned const max)
+static int ZSTD_disableLiteralsCompression(const ZSTD_CCtx_params* cctxParams)
 {
-    unsigned const shift = 8 - accuracyLog;
-    size_t cost = 0;
-    unsigned s;
-    assert(accuracyLog <= 8);
-    for (s = 0; s <= max; ++s) {
-        unsigned const normAcc = norm[s] != -1 ? norm[s] : 1;
-        unsigned const norm256 = normAcc << shift;
-        assert(norm256 > 0);
-        assert(norm256 < 256);
-        cost += count[s] * kInverseProbabiltyLog256[norm256];
-    }
-    return cost >> 8;
-}
-
-
-static unsigned ZSTD_getFSEMaxSymbolValue(FSE_CTable const* ctable) {
-  void const* ptr = ctable;
-  U16 const* u16ptr = (U16 const*)ptr;
-  U32 const maxSymbolValue = MEM_read16(u16ptr + 1);
-  return maxSymbolValue;
-}
-
-
-/**
- * Returns the cost in bits of encoding the distribution in count using ctable.
- * Returns an error if ctable cannot represent all the symbols in count.
- */
-static size_t ZSTD_fseBitCost(
-    FSE_CTable const* ctable,
-    unsigned const* count,
-    unsigned const max)
-{
-    unsigned const kAccuracyLog = 8;
-    size_t cost = 0;
-    unsigned s;
-    FSE_CState_t cstate;
-    FSE_initCState(&cstate, ctable);
-    if (ZSTD_getFSEMaxSymbolValue(ctable) < max) {
-        DEBUGLOG(5, "Repeat FSE_CTable has maxSymbolValue %u < %u",
-                    ZSTD_getFSEMaxSymbolValue(ctable), max);
-        return ERROR(GENERIC);
-    }
-    for (s = 0; s <= max; ++s) {
-        unsigned const tableLog = cstate.stateLog;
-        unsigned const badCost = (tableLog + 1) << kAccuracyLog;
-        unsigned const bitCost = FSE_bitCost(cstate.symbolTT, tableLog, s, kAccuracyLog);
-        if (count[s] == 0)
-            continue;
-        if (bitCost >= badCost) {
-            DEBUGLOG(5, "Repeat FSE_CTable has Prob[%u] == 0", s);
-            return ERROR(GENERIC);
-        }
-        cost += count[s] * bitCost;
-    }
-    return cost >> kAccuracyLog;
-}
-
-/**
- * Returns the cost in bytes of encoding the normalized count header.
- * Returns an error if any of the helper functions return an error.
- */
-static size_t ZSTD_NCountCost(unsigned const* count, unsigned const max,
-                              size_t const nbSeq, unsigned const FSELog)
-{
-    BYTE wksp[FSE_NCOUNTBOUND];
-    S16 norm[MaxSeq + 1];
-    const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
-    CHECK_F(FSE_normalizeCount(norm, tableLog, count, nbSeq, max));
-    return FSE_writeNCount(wksp, sizeof(wksp), norm, max, tableLog);
-}
-
-
-typedef enum {
-    ZSTD_defaultDisallowed = 0,
-    ZSTD_defaultAllowed = 1
-} ZSTD_defaultPolicy_e;
-
-MEM_STATIC symbolEncodingType_e
-ZSTD_selectEncodingType(
-        FSE_repeat* repeatMode, unsigned const* count, unsigned const max,
-        size_t const mostFrequent, size_t nbSeq, unsigned const FSELog,
-        FSE_CTable const* prevCTable,
-        short const* defaultNorm, U32 defaultNormLog,
-        ZSTD_defaultPolicy_e const isDefaultAllowed,
-        ZSTD_strategy const strategy)
-{
-    ZSTD_STATIC_ASSERT(ZSTD_defaultDisallowed == 0 && ZSTD_defaultAllowed != 0);
-    if (mostFrequent == nbSeq) {
-        *repeatMode = FSE_repeat_none;
-        if (isDefaultAllowed && nbSeq <= 2) {
-            /* Prefer set_basic over set_rle when there are 2 or less symbols,
-             * since RLE uses 1 byte, but set_basic uses 5-6 bits per symbol.
-             * If basic encoding isn't possible, always choose RLE.
-             */
-            DEBUGLOG(5, "Selected set_basic");
-            return set_basic;
-        }
-        DEBUGLOG(5, "Selected set_rle");
-        return set_rle;
+    switch (cctxParams->literalCompressionMode) {
+    case ZSTD_lcm_huffman:
+        return 0;
+    case ZSTD_lcm_uncompressed:
+        return 1;
+    default:
+        assert(0 /* impossible: pre-validated */);
+        /* fall-through */
+    case ZSTD_lcm_auto:
+        return (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);
     }
-    if (strategy < ZSTD_lazy) {
-        if (isDefaultAllowed) {
-            size_t const staticFse_nbSeq_max = 1000;
-            size_t const mult = 10 - strategy;
-            size_t const baseLog = 3;
-            size_t const dynamicFse_nbSeq_min = (((size_t)1 << defaultNormLog) * mult) >> baseLog;  /* 28-36 for offset, 56-72 for lengths */
-            assert(defaultNormLog >= 5 && defaultNormLog <= 6);  /* xx_DEFAULTNORMLOG */
-            assert(mult <= 9 && mult >= 7);
-            if ( (*repeatMode == FSE_repeat_valid)
-              && (nbSeq < staticFse_nbSeq_max) ) {
-                DEBUGLOG(5, "Selected set_repeat");
-                return set_repeat;
-            }
-            if ( (nbSeq < dynamicFse_nbSeq_min)
-              || (mostFrequent < (nbSeq >> (defaultNormLog-1))) ) {
-                DEBUGLOG(5, "Selected set_basic");
-                /* The format allows default tables to be repeated, but it isn't useful.
-                 * When using simple heuristics to select encoding type, we don't want
-                 * to confuse these tables with dictionaries. When running more careful
-                 * analysis, we don't need to waste time checking both repeating tables
-                 * and default tables.
-                 */
-                *repeatMode = FSE_repeat_none;
-                return set_basic;
-            }
-        }
-    } else {
-        size_t const basicCost = isDefaultAllowed ? ZSTD_crossEntropyCost(defaultNorm, defaultNormLog, count, max) : ERROR(GENERIC);
-        size_t const repeatCost = *repeatMode != FSE_repeat_none ? ZSTD_fseBitCost(prevCTable, count, max) : ERROR(GENERIC);
-        size_t const NCountCost = ZSTD_NCountCost(count, max, nbSeq, FSELog);
-        size_t const compressedCost = (NCountCost << 3) + ZSTD_entropyCost(count, max, nbSeq);
-
-        if (isDefaultAllowed) {
-            assert(!ZSTD_isError(basicCost));
-            assert(!(*repeatMode == FSE_repeat_valid && ZSTD_isError(repeatCost)));
-        }
-        assert(!ZSTD_isError(NCountCost));
-        assert(compressedCost < ERROR(maxCode));
-        DEBUGLOG(5, "Estimated bit costs: basic=%u\trepeat=%u\tcompressed=%u",
-                    (unsigned)basicCost, (unsigned)repeatCost, (unsigned)compressedCost);
-        if (basicCost <= repeatCost && basicCost <= compressedCost) {
-            DEBUGLOG(5, "Selected set_basic");
-            assert(isDefaultAllowed);
-            *repeatMode = FSE_repeat_none;
-            return set_basic;
-        }
-        if (repeatCost <= compressedCost) {
-            DEBUGLOG(5, "Selected set_repeat");
-            assert(!ZSTD_isError(repeatCost));
-            return set_repeat;
-        }
-        assert(compressedCost < basicCost && compressedCost < repeatCost);
-    }
-    DEBUGLOG(5, "Selected set_compressed");
-    *repeatMode = FSE_repeat_check;
-    return set_compressed;
-}
-
-MEM_STATIC size_t
-ZSTD_buildCTable(void* dst, size_t dstCapacity,
-                FSE_CTable* nextCTable, U32 FSELog, symbolEncodingType_e type,
-                unsigned* count, U32 max,
-                const BYTE* codeTable, size_t nbSeq,
-                const S16* defaultNorm, U32 defaultNormLog, U32 defaultMax,
-                const FSE_CTable* prevCTable, size_t prevCTableSize,
-                void* workspace, size_t workspaceSize)
-{
-    BYTE* op = (BYTE*)dst;
-    const BYTE* const oend = op + dstCapacity;
-    DEBUGLOG(6, "ZSTD_buildCTable (dstCapacity=%u)", (unsigned)dstCapacity);
-
-    switch (type) {
-    case set_rle:
-        CHECK_F(FSE_buildCTable_rle(nextCTable, (BYTE)max));
-        if (dstCapacity==0) return ERROR(dstSize_tooSmall);
-        *op = codeTable[0];
-        return 1;
-    case set_repeat:
-        memcpy(nextCTable, prevCTable, prevCTableSize);
-        return 0;
-    case set_basic:
-        CHECK_F(FSE_buildCTable_wksp(nextCTable, defaultNorm, defaultMax, defaultNormLog, workspace, workspaceSize));  /* note : could be pre-calculated */
-        return 0;
-    case set_compressed: {
-        S16 norm[MaxSeq + 1];
-        size_t nbSeq_1 = nbSeq;
-        const U32 tableLog = FSE_optimalTableLog(FSELog, nbSeq, max);
-        if (count[codeTable[nbSeq-1]] > 1) {
-            count[codeTable[nbSeq-1]]--;
-            nbSeq_1--;
-        }
-        assert(nbSeq_1 > 1);
-        CHECK_F(FSE_normalizeCount(norm, tableLog, count, nbSeq_1, max));
-        {   size_t const NCountSize = FSE_writeNCount(op, oend - op, norm, max, tableLog);   /* overflow protected */
-            if (FSE_isError(NCountSize)) return NCountSize;
-            CHECK_F(FSE_buildCTable_wksp(nextCTable, norm, max, tableLog, workspace, workspaceSize));
-            return NCountSize;
-        }
-    }
-    default: return assert(0), ERROR(GENERIC);
-    }
-}
-
-FORCE_INLINE_TEMPLATE size_t
-ZSTD_encodeSequences_body(
-            void* dst, size_t dstCapacity,
-            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
-            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
-            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
-            seqDef const* sequences, size_t nbSeq, int longOffsets)
-{
-    BIT_CStream_t blockStream;
-    FSE_CState_t  stateMatchLength;
-    FSE_CState_t  stateOffsetBits;
-    FSE_CState_t  stateLitLength;
-
-    CHECK_E(BIT_initCStream(&blockStream, dst, dstCapacity), dstSize_tooSmall); /* not enough space remaining */
-    DEBUGLOG(6, "available space for bitstream : %i  (dstCapacity=%u)",
-                (int)(blockStream.endPtr - blockStream.startPtr),
-                (unsigned)dstCapacity);
-
-    /* first symbols */
-    FSE_initCState2(&stateMatchLength, CTable_MatchLength, mlCodeTable[nbSeq-1]);
-    FSE_initCState2(&stateOffsetBits,  CTable_OffsetBits,  ofCodeTable[nbSeq-1]);
-    FSE_initCState2(&stateLitLength,   CTable_LitLength,   llCodeTable[nbSeq-1]);
-    BIT_addBits(&blockStream, sequences[nbSeq-1].litLength, LL_bits[llCodeTable[nbSeq-1]]);
-    if (MEM_32bits()) BIT_flushBits(&blockStream);
-    BIT_addBits(&blockStream, sequences[nbSeq-1].matchLength, ML_bits[mlCodeTable[nbSeq-1]]);
-    if (MEM_32bits()) BIT_flushBits(&blockStream);
-    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;
-        for (n=nbSeq-2 ; n<nbSeq ; n--) {      /* intentional underflow */
-            BYTE const llCode = llCodeTable[n];
-            BYTE const ofCode = ofCodeTable[n];
-            BYTE const mlCode = mlCodeTable[n];
-            U32  const llBits = LL_bits[llCode];
-            U32  const ofBits = ofCode;
-            U32  const mlBits = ML_bits[mlCode];
-            DEBUGLOG(6, "encoding: litlen:%2u - matchlen:%2u - offCode:%7u",
-                        (unsigned)sequences[n].litLength,
-                        (unsigned)sequences[n].matchLength + MINMATCH,
-                        (unsigned)sequences[n].offset);
-                                                                            /* 32b*/  /* 64b*/
-                                                                            /* (7)*/  /* (7)*/
-            FSE_encodeSymbol(&blockStream, &stateOffsetBits, ofCode);       /* 15 */  /* 15 */
-            FSE_encodeSymbol(&blockStream, &stateMatchLength, mlCode);      /* 24 */  /* 24 */
-            if (MEM_32bits()) BIT_flushBits(&blockStream);                  /* (7)*/
-            FSE_encodeSymbol(&blockStream, &stateLitLength, llCode);        /* 16 */  /* 33 */
-            if (MEM_32bits() || (ofBits+mlBits+llBits >= 64-7-(LLFSELog+MLFSELog+OffFSELog)))
-                BIT_flushBits(&blockStream);                                /* (7)*/
-            BIT_addBits(&blockStream, sequences[n].litLength, llBits);
-            if (MEM_32bits() && ((llBits+mlBits)>24)) BIT_flushBits(&blockStream);
-            BIT_addBits(&blockStream, sequences[n].matchLength, mlBits);
-            if (MEM_32bits() || (ofBits+mlBits+llBits > 56)) BIT_flushBits(&blockStream);
-            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)*/
-            DEBUGLOG(7, "remaining space : %i", (int)(blockStream.endPtr - blockStream.ptr));
-    }   }
-
-    DEBUGLOG(6, "ZSTD_encodeSequences: flushing ML state with %u bits", stateMatchLength.stateLog);
-    FSE_flushCState(&blockStream, &stateMatchLength);
-    DEBUGLOG(6, "ZSTD_encodeSequences: flushing Off state with %u bits", stateOffsetBits.stateLog);
-    FSE_flushCState(&blockStream, &stateOffsetBits);
-    DEBUGLOG(6, "ZSTD_encodeSequences: flushing LL state with %u bits", stateLitLength.stateLog);
-    FSE_flushCState(&blockStream, &stateLitLength);
-
-    {   size_t const streamSize = BIT_closeCStream(&blockStream);
-        if (streamSize==0) return ERROR(dstSize_tooSmall);   /* not enough space */
-        return streamSize;
-    }
-}
-
-static size_t
-ZSTD_encodeSequences_default(
-            void* dst, size_t dstCapacity,
-            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
-            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
-            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
-            seqDef const* sequences, size_t nbSeq, int longOffsets)
-{
-    return ZSTD_encodeSequences_body(dst, dstCapacity,
-                                    CTable_MatchLength, mlCodeTable,
-                                    CTable_OffsetBits, ofCodeTable,
-                                    CTable_LitLength, llCodeTable,
-                                    sequences, nbSeq, longOffsets);
-}
-
-
-#if DYNAMIC_BMI2
-
-static TARGET_ATTRIBUTE("bmi2") size_t
-ZSTD_encodeSequences_bmi2(
-            void* dst, size_t dstCapacity,
-            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
-            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
-            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
-            seqDef const* sequences, size_t nbSeq, int longOffsets)
-{
-    return ZSTD_encodeSequences_body(dst, dstCapacity,
-                                    CTable_MatchLength, mlCodeTable,
-                                    CTable_OffsetBits, ofCodeTable,
-                                    CTable_LitLength, llCodeTable,
-                                    sequences, nbSeq, longOffsets);
-}
-
-#endif
-
-static size_t ZSTD_encodeSequences(
-            void* dst, size_t dstCapacity,
-            FSE_CTable const* CTable_MatchLength, BYTE const* mlCodeTable,
-            FSE_CTable const* CTable_OffsetBits, BYTE const* ofCodeTable,
-            FSE_CTable const* CTable_LitLength, BYTE const* llCodeTable,
-            seqDef const* sequences, size_t nbSeq, int longOffsets, int bmi2)
-{
-    DEBUGLOG(5, "ZSTD_encodeSequences: dstCapacity = %u", (unsigned)dstCapacity);
-#if DYNAMIC_BMI2
-    if (bmi2) {
-        return ZSTD_encodeSequences_bmi2(dst, dstCapacity,
-                                         CTable_MatchLength, mlCodeTable,
-                                         CTable_OffsetBits, ofCodeTable,
-                                         CTable_LitLength, llCodeTable,
-                                         sequences, nbSeq, longOffsets);
-    }
-#endif
-    (void)bmi2;
-    return ZSTD_encodeSequences_default(dst, dstCapacity,
-                                        CTable_MatchLength, mlCodeTable,
-                                        CTable_OffsetBits, ofCodeTable,
-                                        CTable_LitLength, llCodeTable,
-                                        sequences, nbSeq, longOffsets);
 }
 
 /* ZSTD_compressSequences_internal():
@@ -2393,46 +1955,48 b' ZSTD_compressSequences_internal(seqStore'
     BYTE* const ostart = (BYTE*)dst;
     BYTE* const oend = ostart + dstCapacity;
     BYTE* op = ostart;
-    size_t const nbSeq = seqStorePtr->sequences - seqStorePtr->sequencesStart;
+    size_t const nbSeq = (size_t)(seqStorePtr->sequences - seqStorePtr->sequencesStart);
     BYTE* seqHead;
     BYTE* lastNCount = NULL;
 
+    DEBUGLOG(5, "ZSTD_compressSequences_internal (nbSeq=%zu)", nbSeq);
     ZSTD_STATIC_ASSERT(HUF_WORKSPACE_SIZE >= (1<<MAX(MLFSELog,LLFSELog)));
-    DEBUGLOG(5, "ZSTD_compressSequences_internal");
 
     /* Compress literals */
     {   const BYTE* const literals = seqStorePtr->litStart;
-        size_t const litSize = seqStorePtr->lit - literals;
-        int const disableLiteralCompression = (cctxParams->cParams.strategy == ZSTD_fast) && (cctxParams->cParams.targetLength > 0);
+        size_t const litSize = (size_t)(seqStorePtr->lit - literals);
         size_t const cSize = ZSTD_compressLiterals(
                                     &prevEntropy->huf, &nextEntropy->huf,
-                                    cctxParams->cParams.strategy, disableLiteralCompression,
+                                    cctxParams->cParams.strategy,
+                                    ZSTD_disableLiteralsCompression(cctxParams),
                                     op, dstCapacity,
                                     literals, litSize,
                                     workspace, wkspSize,
                                     bmi2);
-        if (ZSTD_isError(cSize))
-          return cSize;
+        FORWARD_IF_ERROR(cSize);
         assert(cSize <= dstCapacity);
         op += cSize;
     }
 
     /* Sequences Header */
-    if ((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF((oend-op) < 3 /*max nbSeq Size*/ + 1 /*seqHead*/,
+                    dstSize_tooSmall);
     if (nbSeq < 0x7F)
         *op++ = (BYTE)nbSeq;
     else if (nbSeq < LONGNBSEQ)
         op[0] = (BYTE)((nbSeq>>8) + 0x80), op[1] = (BYTE)nbSeq, op+=2;
     else
         op[0]=0xFF, MEM_writeLE16(op+1, (U16)(nbSeq - LONGNBSEQ)), op+=3;
+    assert(op <= oend);
     if (nbSeq==0) {
         /* Copy the old tables over as if we repeated them */
         memcpy(&nextEntropy->fse, &prevEntropy->fse, sizeof(prevEntropy->fse));
-        return op - ostart;
+        return (size_t)(op - ostart);
     }
 
     /* seqHead : flags for FSE encoding type */
     seqHead = op++;
+    assert(op <= oend);
 
     /* convert length/distances into codes */
     ZSTD_seqToCodes(seqStorePtr);
@@ -2448,14 +2012,15 b' ZSTD_compressSequences_internal(seqStore'
                                         ZSTD_defaultAllowed, strategy);
         assert(set_basic < set_compressed && set_rle < set_compressed);
         assert(!(LLtype < set_compressed && nextEntropy->fse.litlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
-        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
+        {   size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_LitLength, LLFSELog, (symbolEncodingType_e)LLtype,
                                                     count, max, llCodeTable, nbSeq, LL_defaultNorm, LL_defaultNormLog, MaxLL,
                                                     prevEntropy->fse.litlengthCTable, sizeof(prevEntropy->fse.litlengthCTable),
                                                     workspace, wkspSize);
-            if (ZSTD_isError(countSize)) return countSize;
+            FORWARD_IF_ERROR(countSize);
             if (LLtype == set_compressed)
                 lastNCount = op;
             op += countSize;
+            assert(op <= oend);
     }   }
     /* build CTable for Offsets */
     {   unsigned max = MaxOff;
@@ -2470,14 +2035,15 b' ZSTD_compressSequences_internal(seqStore'
                                         OF_defaultNorm, OF_defaultNormLog,
                                         defaultPolicy, strategy);
         assert(!(Offtype < set_compressed && nextEntropy->fse.offcode_repeatMode != FSE_repeat_none)); /* We don't copy tables */
-        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
+        {   size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_OffsetBits, OffFSELog, (symbolEncodingType_e)Offtype,
                                                     count, max, ofCodeTable, nbSeq, OF_defaultNorm, OF_defaultNormLog, DefaultMaxOff,
                                                     prevEntropy->fse.offcodeCTable, sizeof(prevEntropy->fse.offcodeCTable),
                                                     workspace, wkspSize);
-            if (ZSTD_isError(countSize)) return countSize;
+            FORWARD_IF_ERROR(countSize);
             if (Offtype == set_compressed)
                 lastNCount = op;
             op += countSize;
+            assert(op <= oend);
     }   }
     /* build CTable for MatchLengths */
     {   unsigned max = MaxML;
@@ -2490,29 +2056,31 b' ZSTD_compressSequences_internal(seqStore'
                                         ML_defaultNorm, ML_defaultNormLog,
                                         ZSTD_defaultAllowed, strategy);
         assert(!(MLtype < set_compressed && nextEntropy->fse.matchlength_repeatMode != FSE_repeat_none)); /* We don't copy tables */
-        {   size_t const countSize = ZSTD_buildCTable(op, oend - op, CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
+        {   size_t const countSize = ZSTD_buildCTable(op, (size_t)(oend - op), CTable_MatchLength, MLFSELog, (symbolEncodingType_e)MLtype,
                                                     count, max, mlCodeTable, nbSeq, ML_defaultNorm, ML_defaultNormLog, MaxML,
                                                     prevEntropy->fse.matchlengthCTable, sizeof(prevEntropy->fse.matchlengthCTable),
                                                     workspace, wkspSize);
-            if (ZSTD_isError(countSize)) return countSize;
+            FORWARD_IF_ERROR(countSize);
             if (MLtype == set_compressed)
                 lastNCount = op;
             op += countSize;
+            assert(op <= oend);
     }   }
 
     *seqHead = (BYTE)((LLtype<<6) + (Offtype<<4) + (MLtype<<2));
 
     {   size_t const bitstreamSize = ZSTD_encodeSequences(
-                                        op, oend - op,
+                                        op, (size_t)(oend - op),
                                         CTable_MatchLength, mlCodeTable,
                                         CTable_OffsetBits, ofCodeTable,
                                         CTable_LitLength, llCodeTable,
                                         sequences, nbSeq,
                                         longOffsets, bmi2);
-        if (ZSTD_isError(bitstreamSize)) return bitstreamSize;
+        FORWARD_IF_ERROR(bitstreamSize);
         op += bitstreamSize;
+        assert(op <= oend);
         /* zstd versions <= 1.3.4 mistakenly report corruption when
-         * FSE_readNCount() recieves a buffer < 4 bytes.
+         * FSE_readNCount() receives a buffer < 4 bytes.
          * Fixed by https://github.com/facebook/zstd/pull/1146.
          * This can happen when the last set_compressed table present is 2
          * bytes and the bitstream is only one byte.
@@ -2529,7 +2097,7 b' ZSTD_compressSequences_internal(seqStore'
     }
 
     DEBUGLOG(5, "compressed block size : %u", (unsigned)(op - ostart));
-    return op - ostart;
+    return (size_t)(op - ostart);
 }
 
 MEM_STATIC size_t
@@ -2552,7 +2120,7 b' ZSTD_compressSequences(seqStore_t* seqSt'
      */
     if ((cSize == ERROR(dstSize_tooSmall)) & (srcSize <= dstCapacity))
         return 0;  /* block not compressed */
-    if (ZSTD_isError(cSize)) return cSize;
+    FORWARD_IF_ERROR(cSize);
 
     /* Check compressibility */
     {   size_t const maxCSize = srcSize - ZSTD_minGain(srcSize, cctxParams->cParams.strategy);
@@ -2622,27 +2190,24 b' void ZSTD_resetSeqStore(seqStore_t* ssPt'
     ssPtr->longLengthID = 0;
 }
 
-static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc,
-                                        void* dst, size_t dstCapacity,
-                                        const void* src, size_t srcSize)
+typedef enum { ZSTDbss_compress, ZSTDbss_noCompress } ZSTD_buildSeqStore_e;
+
+static size_t ZSTD_buildSeqStore(ZSTD_CCtx* zc, const void* src, size_t srcSize)
 {
     ZSTD_matchState_t* const ms = &zc->blockState.matchState;
-    size_t cSize;
-    DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
-                (unsigned)dstCapacity, (unsigned)ms->window.dictLimit, (unsigned)ms->nextToUpdate);
+    DEBUGLOG(5, "ZSTD_buildSeqStore (srcSize=%zu)", srcSize);
     assert(srcSize <= ZSTD_BLOCKSIZE_MAX);
-
     /* Assert that we have correctly flushed the ctx params into the ms's copy */
     ZSTD_assertEqualCParams(zc->appliedParams.cParams, ms->cParams);
-
     if (srcSize < MIN_CBLOCK_SIZE+ZSTD_blockHeaderSize+1) {
         ZSTD_ldm_skipSequences(&zc->externSeqStore, srcSize, zc->appliedParams.cParams.minMatch);
-        cSize = 0;
-        goto out;  /* don't even attempt compression below a certain srcSize */
+        return ZSTDbss_noCompress; /* don't even attempt compression below a certain srcSize */
     }
     ZSTD_resetSeqStore(&(zc->seqStore));
-    ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy;   /* required for optimal parser to read stats from dictionary */
-
+    /* required for optimal parser to read stats from dictionary */
+    ms->opt.symbolCosts = &zc->blockState.prevCBlock->entropy;
+    /* tell the optimal parser how we expect to compress literals */
+    ms->opt.literalCompressionMode = zc->appliedParams.literalCompressionMode;
     /* a gap between an attached dict and the current window is not safe,
      * they must remain adjacent,
      * and when that stops being the case, the dict must be unset */
@@ -2679,7 +2244,7 b' static size_t ZSTD_compressBlock_interna'
             ldmSeqStore.seq = zc->ldmSequences;
             ldmSeqStore.capacity = zc->maxNbLdmSequences;
             /* Updates ldmSeqStore.size */
-            CHECK_F(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore,
+            FORWARD_IF_ERROR(ZSTD_ldm_generateSequences(&zc->ldmState, &ldmSeqStore,
                                                &zc->appliedParams.ldmParams,
                                                src, srcSize));
             /* Updates ldmSeqStore.pos */
@@ -2696,6 +2261,22 b' static size_t ZSTD_compressBlock_interna'
         {   const BYTE* const lastLiterals = (const BYTE*)src + srcSize - lastLLSize;
             ZSTD_storeLastLiterals(&zc->seqStore, lastLiterals, lastLLSize);
     }   }
+    return ZSTDbss_compress;
+}
+
+static size_t ZSTD_compressBlock_internal(ZSTD_CCtx* zc,
+                                        void* dst, size_t dstCapacity,
+                                        const void* src, size_t srcSize)
+{
+    size_t cSize;
+    DEBUGLOG(5, "ZSTD_compressBlock_internal (dstCapacity=%u, dictLimit=%u, nextToUpdate=%u)",
+                (unsigned)dstCapacity, (unsigned)zc->blockState.matchState.window.dictLimit,
+                (unsigned)zc->blockState.matchState.nextToUpdate);
+
+    {   const size_t bss = ZSTD_buildSeqStore(zc, src, srcSize);
+        FORWARD_IF_ERROR(bss);
+        if (bss == ZSTDbss_noCompress) { cSize = 0; goto out; }
+    }
 
     /* encode sequences and literals */
     cSize = ZSTD_compressSequences(&zc->seqStore,
@@ -2724,6 +2305,25 b' out:'
 }
 
 
+static void ZSTD_overflowCorrectIfNeeded(ZSTD_matchState_t* ms, ZSTD_CCtx_params const* params, void const* ip, void const* iend)
+{
+    if (ZSTD_window_needOverflowCorrection(ms->window, iend)) {
+        U32 const maxDist = (U32)1 << params->cParams.windowLog;
+        U32 const cycleLog = ZSTD_cycleLog(params->cParams.chainLog, params->cParams.strategy);
+        U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip);
+        ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);
+        ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);
+        ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);
+        ZSTD_reduceIndex(ms, params, correction);
+        if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;
+        else ms->nextToUpdate -= correction;
+        /* invalidate dictionaries on overflow correction */
+        ms->loadedDictEnd = 0;
+        ms->dictMatchState = NULL;
+    }
+}
+
+
 /*! ZSTD_compress_frameChunk() :
 *   Compress a chunk of data into one or multiple blocks.
 *   All blocks will be terminated, all input will be consumed.
@@ -2742,7 +2342,7 b' static size_t ZSTD_compress_frameChunk ('
     BYTE* const ostart = (BYTE*)dst;
     BYTE* op = ostart;
     U32 const maxDist = (U32)1 << cctx->appliedParams.cParams.windowLog;
-    assert(cctx->appliedParams.cParams.windowLog <= 31);
+    assert(cctx->appliedParams.cParams.windowLog <= ZSTD_WINDOWLOG_MAX);
 
     DEBUGLOG(5, "ZSTD_compress_frameChunk (blockSize=%u)", (unsigned)blockSize);
     if (cctx->appliedParams.fParams.checksumFlag && srcSize)
@@ -2752,33 +2352,25 b' static size_t ZSTD_compress_frameChunk ('
         ZSTD_matchState_t* const ms = &cctx->blockState.matchState;
         U32 const lastBlock = lastFrameChunk & (blockSize >= remaining);
 
-        if (dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE)
-            return ERROR(dstSize_tooSmall);   /* not enough space to store compressed block */
+        RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize + MIN_CBLOCK_SIZE,
+                        dstSize_tooSmall,
+                        "not enough space to store compressed block");
         if (remaining < blockSize) blockSize = remaining;
 
-        if (ZSTD_window_needOverflowCorrection(ms->window, ip + blockSize)) {
-            U32 const cycleLog = ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy);
-            U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, maxDist, ip);
-            ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);
-            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);
-            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);
-            ZSTD_reduceIndex(cctx, correction);
-            if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;
-            else ms->nextToUpdate -= correction;
-            ms->loadedDictEnd = 0;
-            ms->dictMatchState = NULL;
-        }
-        ZSTD_window_enforceMaxDist(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState);
+        ZSTD_overflowCorrectIfNeeded(ms, &cctx->appliedParams, ip, ip + blockSize);
+        ZSTD_checkDictValidity(&ms->window, ip + blockSize, maxDist, &ms->loadedDictEnd, &ms->dictMatchState);
+
+        /* Ensure hash/chain table insertion resumes no sooner than lowlimit */
         if (ms->nextToUpdate < ms->window.lowLimit) ms->nextToUpdate = ms->window.lowLimit;
 
         {   size_t cSize = ZSTD_compressBlock_internal(cctx,
                                 op+ZSTD_blockHeaderSize, dstCapacity-ZSTD_blockHeaderSize,
                                 ip, blockSize);
-            if (ZSTD_isError(cSize)) return cSize;
+            FORWARD_IF_ERROR(cSize);
 
             if (cSize == 0) {  /* block is not compressible */
                 cSize = ZSTD_noCompressBlock(op, dstCapacity, ip, blockSize, lastBlock);
-                if (ZSTD_isError(cSize)) return cSize;
+                FORWARD_IF_ERROR(cSize);
             } else {
                 U32 const cBlockHeader24 = lastBlock + (((U32)bt_compressed)<<1) + (U32)(cSize << 3);
                 MEM_writeLE24(op, cBlockHeader24);
@@ -2796,7 +2388,7 b' static size_t ZSTD_compress_frameChunk ('
     }   }
 
     if (lastFrameChunk && (op>ostart)) cctx->stage = ZSTDcs_ending;
-    return op-ostart;
+    return (size_t)(op-ostart);
 }
 
 
@@ -2811,11 +2403,11 b' static size_t ZSTD_writeFrameHeader(void'
     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;  /* 0-3 */
-    BYTE  const frameHeaderDecriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );
+    BYTE  const frameHeaderDescriptionByte = (BYTE)(dictIDSizeCode + (checksumFlag<<2) + (singleSegment<<5) + (fcsCode<<6) );
     size_t pos=0;
 
     assert(!(params.fParams.contentSizeFlag && pledgedSrcSize == ZSTD_CONTENTSIZE_UNKNOWN));
-    if (dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF(dstCapacity < ZSTD_FRAMEHEADERSIZE_MAX, dstSize_tooSmall);
     DEBUGLOG(4, "ZSTD_writeFrameHeader : dictIDFlag : %u ; dictID : %u ; dictIDSizeCode : %u",
                 !params.fParams.noDictIDFlag, (unsigned)dictID, (unsigned)dictIDSizeCode);
 
@@ -2823,7 +2415,7 b' static size_t ZSTD_writeFrameHeader(void'
         MEM_writeLE32(dst, ZSTD_MAGICNUMBER);
         pos = 4;
     }
-    op[pos++] = frameHeaderDecriptionByte;
+    op[pos++] = frameHeaderDescriptionByte;
     if (!singleSegment) op[pos++] = windowLogByte;
     switch(dictIDSizeCode)
     {
@@ -2847,11 +2439,11 b' static size_t ZSTD_writeFrameHeader(void'
 /* ZSTD_writeLastEmptyBlock() :
  * output an empty Block with end-of-frame mark to complete a frame
  * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
- *           or an error code if `dstCapcity` is too small (<ZSTD_blockHeaderSize)
+ *           or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
  */
 size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity)
 {
-    if (dstCapacity < ZSTD_blockHeaderSize) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF(dstCapacity < ZSTD_blockHeaderSize, dstSize_tooSmall);
     {   U32 const cBlockHeader24 = 1 /*lastBlock*/ + (((U32)bt_raw)<<1);  /* 0 size */
         MEM_writeLE24(dst, cBlockHeader24);
         return ZSTD_blockHeaderSize;
@@ -2860,10 +2452,9 b' size_t ZSTD_writeLastEmptyBlock(void* ds'
 
 size_t ZSTD_referenceExternalSequences(ZSTD_CCtx* cctx, rawSeq* seq, size_t nbSeq)
 {
-    if (cctx->stage != ZSTDcs_init)
-        return ERROR(stage_wrong);
-    if (cctx->appliedParams.ldmParams.enableLdm)
-        return ERROR(parameter_unsupported);
+    RETURN_ERROR_IF(cctx->stage != ZSTDcs_init, stage_wrong);
+    RETURN_ERROR_IF(cctx->appliedParams.ldmParams.enableLdm,
+                    parameter_unsupported);
     cctx->externSeqStore.seq = seq;
     cctx->externSeqStore.size = nbSeq;
     cctx->externSeqStore.capacity = nbSeq;
@@ -2882,12 +2473,14 b' static size_t ZSTD_compressContinue_inte'
 
     DEBUGLOG(5, "ZSTD_compressContinue_internal, stage: %u, srcSize: %u",
                 cctx->stage, (unsigned)srcSize);
-    if (cctx->stage==ZSTDcs_created) return ERROR(stage_wrong);   /* missing init (ZSTD_compressBegin) */
+    RETURN_ERROR_IF(cctx->stage==ZSTDcs_created, stage_wrong,
+                    "missing init (ZSTD_compressBegin)");
 
     if (frame && (cctx->stage==ZSTDcs_init)) {
         fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams,
                                        cctx->pledgedSrcSizePlusOne-1, cctx->dictID);
-        if (ZSTD_isError(fhSize)) return fhSize;
+        FORWARD_IF_ERROR(fhSize);
+        assert(fhSize <= dstCapacity);
         dstCapacity -= fhSize;
         dst = (char*)dst + fhSize;
         cctx->stage = ZSTDcs_ongoing;
@@ -2904,35 +2497,25 b' static size_t ZSTD_compressContinue_inte'
 
     if (!frame) {
         /* overflow check and correction for block mode */
-        if (ZSTD_window_needOverflowCorrection(ms->window, (const char*)src + srcSize)) {
-            U32 const cycleLog = ZSTD_cycleLog(cctx->appliedParams.cParams.chainLog, cctx->appliedParams.cParams.strategy);
-            U32 const correction = ZSTD_window_correctOverflow(&ms->window, cycleLog, 1 << cctx->appliedParams.cParams.windowLog, src);
-            ZSTD_STATIC_ASSERT(ZSTD_CHAINLOG_MAX <= 30);
-            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX_32 <= 30);
-            ZSTD_STATIC_ASSERT(ZSTD_WINDOWLOG_MAX <= 31);
-            ZSTD_reduceIndex(cctx, correction);
-            if (ms->nextToUpdate < correction) ms->nextToUpdate = 0;
-            else ms->nextToUpdate -= correction;
-            ms->loadedDictEnd = 0;
-            ms->dictMatchState = NULL;
-        }
+        ZSTD_overflowCorrectIfNeeded(ms, &cctx->appliedParams, src, (BYTE const*)src + srcSize);
     }
 
     DEBUGLOG(5, "ZSTD_compressContinue_internal (blockSize=%u)", (unsigned)cctx->blockSize);
     {   size_t const cSize = frame ?
                              ZSTD_compress_frameChunk (cctx, dst, dstCapacity, src, srcSize, lastFrameChunk) :
                              ZSTD_compressBlock_internal (cctx, dst, dstCapacity, src, srcSize);
-        if (ZSTD_isError(cSize)) return cSize;
+        FORWARD_IF_ERROR(cSize);
         cctx->consumedSrcSize += srcSize;
         cctx->producedCSize += (cSize + fhSize);
         assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
         if (cctx->pledgedSrcSizePlusOne != 0) {  /* control src size */
             ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
-            if (cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne) {
-                DEBUGLOG(4, "error : pledgedSrcSize = %u, while realSrcSize >= %u",
-                    (unsigned)cctx->pledgedSrcSizePlusOne-1, (unsigned)cctx->consumedSrcSize);
-                return ERROR(srcSize_wrong);
-            }
+            RETURN_ERROR_IF(
+                cctx->consumedSrcSize+1 > cctx->pledgedSrcSizePlusOne,
+                srcSize_wrong,
+                "error : pledgedSrcSize = %u, while realSrcSize >= %u",
+                (unsigned)cctx->pledgedSrcSizePlusOne-1,
+                (unsigned)cctx->consumedSrcSize);
         }
         return cSize + fhSize;
     }
@@ -2956,8 +2539,9 b' size_t ZSTD_getBlockSize(const ZSTD_CCtx'
 
 size_t ZSTD_compressBlock(ZSTD_CCtx* cctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
 {
-    size_t const blockSizeMax = ZSTD_getBlockSize(cctx);
-    if (srcSize > blockSizeMax) return ERROR(srcSize_wrong);
+    DEBUGLOG(5, "ZSTD_compressBlock: srcSize = %u", (unsigned)srcSize);
+    { size_t const blockSizeMax = ZSTD_getBlockSize(cctx);
+      RETURN_ERROR_IF(srcSize > blockSizeMax, srcSize_wrong); }
 
     return ZSTD_compressContinue_internal(cctx, dst, dstCapacity, src, srcSize, 0 /* frame mode */, 0 /* last chunk */);
 }
@@ -2970,7 +2554,7 b' static size_t ZSTD_loadDictionaryContent'
                                          const void* src, size_t srcSize,
                                          ZSTD_dictTableLoadMethod_e dtlm)
 {
-    const BYTE* const ip = (const BYTE*) src;
+    const BYTE* ip = (const BYTE*) src;
     const BYTE* const iend = ip + srcSize;
 
     ZSTD_window_update(&ms->window, src, srcSize);
@@ -2981,32 +2565,42 b' static size_t ZSTD_loadDictionaryContent'
 
     if (srcSize <= HASH_READ_SIZE) return 0;
 
-    switch(params->cParams.strategy)
-    {
-    case ZSTD_fast:
-        ZSTD_fillHashTable(ms, iend, dtlm);
-        break;
-    case ZSTD_dfast:
-        ZSTD_fillDoubleHashTable(ms, iend, dtlm);
-        break;
-
-    case ZSTD_greedy:
-    case ZSTD_lazy:
-    case ZSTD_lazy2:
-        if (srcSize >= HASH_READ_SIZE)
-            ZSTD_insertAndFindFirstIndex(ms, iend-HASH_READ_SIZE);
-        break;
-
-    case ZSTD_btlazy2:   /* we want the dictionary table fully sorted */
-    case ZSTD_btopt:
-    case ZSTD_btultra:
-    case ZSTD_btultra2:
-        if (srcSize >= HASH_READ_SIZE)
-            ZSTD_updateTree(ms, iend-HASH_READ_SIZE, iend);
-        break;
-
-    default:
-        assert(0);  /* not possible : not a valid strategy id */
+    while (iend - ip > HASH_READ_SIZE) {
+        size_t const remaining = (size_t)(iend - ip);
+        size_t const chunk = MIN(remaining, ZSTD_CHUNKSIZE_MAX);
+        const BYTE* const ichunk = ip + chunk;
+
+        ZSTD_overflowCorrectIfNeeded(ms, params, ip, ichunk);
+
+        switch(params->cParams.strategy)
+        {
+        case ZSTD_fast:
+            ZSTD_fillHashTable(ms, ichunk, dtlm);
+            break;
+        case ZSTD_dfast:
+            ZSTD_fillDoubleHashTable(ms, ichunk, dtlm);
+            break;
+
+        case ZSTD_greedy:
+        case ZSTD_lazy:
+        case ZSTD_lazy2:
+            if (chunk >= HASH_READ_SIZE)
+                ZSTD_insertAndFindFirstIndex(ms, ichunk-HASH_READ_SIZE);
+            break;
+
+        case ZSTD_btlazy2:   /* we want the dictionary table fully sorted */
+        case ZSTD_btopt:
+        case ZSTD_btultra:
+        case ZSTD_btultra2:
+            if (chunk >= HASH_READ_SIZE)
+                ZSTD_updateTree(ms, ichunk-HASH_READ_SIZE, ichunk);
+            break;
+
+        default:
+            assert(0);  /* not possible : not a valid strategy id */
+        }
+
+        ip = ichunk;
     }
 
     ms->nextToUpdate = (U32)(iend - ms->window.base);
@@ -3020,9 +2614,9 b' static size_t ZSTD_loadDictionaryContent'
    NOTE: This behavior is not standard and could be improved in the future. */
 static size_t ZSTD_checkDictNCount(short* normalizedCounter, unsigned dictMaxSymbolValue, unsigned maxSymbolValue) {
     U32 s;
-    if (dictMaxSymbolValue < maxSymbolValue) return ERROR(dictionary_corrupted);
+    RETURN_ERROR_IF(dictMaxSymbolValue < maxSymbolValue, dictionary_corrupted);
     for (s = 0; s <= maxSymbolValue; ++s) {
-        if (normalizedCounter[s] == 0) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(normalizedCounter[s] == 0, dictionary_corrupted);
     }
     return 0;
 }
@@ -3060,53 +2654,56 b' static size_t ZSTD_loadZstdDictionary(ZS'
 
     {   unsigned maxSymbolValue = 255;
         size_t const hufHeaderSize = HUF_readCTable((HUF_CElt*)bs->entropy.huf.CTable, &maxSymbolValue, dictPtr, dictEnd-dictPtr);
-        if (HUF_isError(hufHeaderSize)) return ERROR(dictionary_corrupted);
-        if (maxSymbolValue < 255) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(HUF_isError(hufHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(maxSymbolValue < 255, dictionary_corrupted);
         dictPtr += hufHeaderSize;
     }
 
     {   unsigned offcodeLog;
         size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted);
-        if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);
         /* Defer checking offcodeMaxValue because we need to know the size of the dictionary content */
         /* fill all offset symbols to avoid garbage at end of table */
-        CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.offcodeCTable,
-                                    offcodeNCount, MaxOff, offcodeLog,
-                                    workspace, HUF_WORKSPACE_SIZE),
-                 dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.offcodeCTable,
+                offcodeNCount, MaxOff, offcodeLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted);
         dictPtr += offcodeHeaderSize;
     }
 
     {   short matchlengthNCount[MaxML+1];
         unsigned matchlengthMaxValue = MaxML, matchlengthLog;
         size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted);
-        if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);
         /* Every match length code must have non-zero probability */
-        CHECK_F( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
-        CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.matchlengthCTable,
-                                    matchlengthNCount, matchlengthMaxValue, matchlengthLog,
-                                    workspace, HUF_WORKSPACE_SIZE),
-                 dictionary_corrupted);
+        FORWARD_IF_ERROR( ZSTD_checkDictNCount(matchlengthNCount, matchlengthMaxValue, MaxML));
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.matchlengthCTable,
+                matchlengthNCount, matchlengthMaxValue, matchlengthLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted);
         dictPtr += matchlengthHeaderSize;
     }
 
     {   short litlengthNCount[MaxLL+1];
         unsigned litlengthMaxValue = MaxLL, litlengthLog;
         size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted);
-        if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);
         /* Every literal length code must have non-zero probability */
-        CHECK_F( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
-        CHECK_E( FSE_buildCTable_wksp(bs->entropy.fse.litlengthCTable,
-                                    litlengthNCount, litlengthMaxValue, litlengthLog,
-                                    workspace, HUF_WORKSPACE_SIZE),
-                 dictionary_corrupted);
+        FORWARD_IF_ERROR( ZSTD_checkDictNCount(litlengthNCount, litlengthMaxValue, MaxLL));
+        RETURN_ERROR_IF(FSE_isError(FSE_buildCTable_wksp(
+                bs->entropy.fse.litlengthCTable,
+                litlengthNCount, litlengthMaxValue, litlengthLog,
+                workspace, HUF_WORKSPACE_SIZE)),
+            dictionary_corrupted);
         dictPtr += litlengthHeaderSize;
     }
 
-    if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted);
+    RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);
     bs->rep[0] = MEM_readLE32(dictPtr+0);
     bs->rep[1] = MEM_readLE32(dictPtr+4);
     bs->rep[2] = MEM_readLE32(dictPtr+8);
@@ -3119,19 +2716,19 b' static size_t ZSTD_loadZstdDictionary(ZS'
             offcodeMax = ZSTD_highbit32(maxOffset); /* Calculate minimum offset code required to represent maxOffset */
         }
         /* All offset values <= dictContentSize + 128 KB must be representable */
-        CHECK_F (ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));
+        FORWARD_IF_ERROR(ZSTD_checkDictNCount(offcodeNCount, offcodeMaxValue, MIN(offcodeMax, MaxOff)));
         /* All repCodes must be <= dictContentSize and != 0*/
         {   U32 u;
             for (u=0; u<3; u++) {
-                if (bs->rep[u] == 0) return ERROR(dictionary_corrupted);
-                if (bs->rep[u] > dictContentSize) return ERROR(dictionary_corrupted);
+                RETURN_ERROR_IF(bs->rep[u] == 0, dictionary_corrupted);
+                RETURN_ERROR_IF(bs->rep[u] > dictContentSize, dictionary_corrupted);
         }   }
 
         bs->entropy.huf.repeatMode = HUF_repeat_valid;
         bs->entropy.fse.offcode_repeatMode = FSE_repeat_valid;
         bs->entropy.fse.matchlength_repeatMode = FSE_repeat_valid;
         bs->entropy.fse.litlength_repeatMode = FSE_repeat_valid;
-        CHECK_F(ZSTD_loadDictionaryContent(ms, params, dictPtr, dictContentSize, dtlm));
+        FORWARD_IF_ERROR(ZSTD_loadDictionaryContent(ms, params, dictPtr, dictContentSize, dtlm));
         return dictID;
     }
 }
@@ -3161,8 +2758,7 b' ZSTD_compress_insertDictionary(ZSTD_comp'
             DEBUGLOG(4, "raw content dictionary detected");
             return ZSTD_loadDictionaryContent(ms, params, dict, dictSize, dtlm);
         }
-        if (dictContentType == ZSTD_dct_fullDict)
-            return ERROR(dictionary_wrong);
+        RETURN_ERROR_IF(dictContentType == ZSTD_dct_fullDict, dictionary_wrong);
         assert(0);   /* impossible */
     }
 
@@ -3189,14 +2785,13 b' static size_t ZSTD_compressBegin_interna'
         return ZSTD_resetCCtx_usingCDict(cctx, cdict, params, pledgedSrcSize, zbuff);
     }
 
-    CHECK_F( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
+    FORWARD_IF_ERROR( ZSTD_resetCCtx_internal(cctx, params, pledgedSrcSize,
                                      ZSTDcrp_continue, zbuff) );
-    {
-        size_t const dictID = ZSTD_compress_insertDictionary(
+    {   size_t const dictID = ZSTD_compress_insertDictionary(
                 cctx->blockState.prevCBlock, &cctx->blockState.matchState,
                 &params, dict, dictSize, dictContentType, dtlm, cctx->entropyWorkspace);
-        if (ZSTD_isError(dictID)) return dictID;
-        assert(dictID <= (size_t)(U32)-1);
+        FORWARD_IF_ERROR(dictID);
+        assert(dictID <= UINT_MAX);
         cctx->dictID = (U32)dictID;
     }
     return 0;
@@ -3212,7 +2807,7 b' size_t ZSTD_compressBegin_advanced_inter'
 {
     DEBUGLOG(4, "ZSTD_compressBegin_advanced_internal: wlog=%u", params.cParams.windowLog);
     /* compression parameters verification and optimization */
-    CHECK_F( ZSTD_checkCParams(params.cParams) );
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );
     return ZSTD_compressBegin_internal(cctx,
                                        dict, dictSize, dictContentType, dtlm,
                                        cdict,
@@ -3260,12 +2855,12 b' static size_t ZSTD_writeEpilogue(ZSTD_CC'
     size_t fhSize = 0;
 
     DEBUGLOG(4, "ZSTD_writeEpilogue");
-    if (cctx->stage == ZSTDcs_created) return ERROR(stage_wrong);  /* init missing */
+    RETURN_ERROR_IF(cctx->stage == ZSTDcs_created, stage_wrong, "init missing");
 
     /* special case : empty frame */
     if (cctx->stage == ZSTDcs_init) {
         fhSize = ZSTD_writeFrameHeader(dst, dstCapacity, cctx->appliedParams, 0, 0);
-        if (ZSTD_isError(fhSize)) return fhSize;
+        FORWARD_IF_ERROR(fhSize);
         dstCapacity -= fhSize;
         op += fhSize;
         cctx->stage = ZSTDcs_ongoing;
@@ -3274,7 +2869,7 b' static size_t ZSTD_writeEpilogue(ZSTD_CC'
     if (cctx->stage != ZSTDcs_ending) {
         /* write one last empty block, make it the "last" block */
         U32 const cBlockHeader24 = 1 /* last block */ + (((U32)bt_raw)<<1) + 0;
-        if (dstCapacity<4) return ERROR(dstSize_tooSmall);
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);
         MEM_writeLE32(op, cBlockHeader24);
         op += ZSTD_blockHeaderSize;
         dstCapacity -= ZSTD_blockHeaderSize;
@@ -3282,7 +2877,7 b' static size_t ZSTD_writeEpilogue(ZSTD_CC'
 
     if (cctx->appliedParams.fParams.checksumFlag) {
         U32 const checksum = (U32) XXH64_digest(&cctx->xxhState);
-        if (dstCapacity<4) return ERROR(dstSize_tooSmall);
+        RETURN_ERROR_IF(dstCapacity<4, dstSize_tooSmall);
         DEBUGLOG(4, "ZSTD_writeEpilogue: write checksum : %08X", (unsigned)checksum);
         MEM_writeLE32(op, checksum);
         op += 4;
@@ -3300,18 +2895,20 b' size_t ZSTD_compressEnd (ZSTD_CCtx* cctx'
     size_t const cSize = ZSTD_compressContinue_internal(cctx,
                                 dst, dstCapacity, src, srcSize,
                                 1 /* frame mode */, 1 /* last chunk */);
-    if (ZSTD_isError(cSize)) return cSize;
+    FORWARD_IF_ERROR(cSize);
     endResult = ZSTD_writeEpilogue(cctx, (char*)dst + cSize, dstCapacity-cSize);
-    if (ZSTD_isError(endResult)) return endResult;
+    FORWARD_IF_ERROR(endResult);
     assert(!(cctx->appliedParams.fParams.contentSizeFlag && cctx->pledgedSrcSizePlusOne == 0));
     if (cctx->pledgedSrcSizePlusOne != 0) {  /* control src size */
         ZSTD_STATIC_ASSERT(ZSTD_CONTENTSIZE_UNKNOWN == (unsigned long long)-1);
         DEBUGLOG(4, "end of frame : controlling src size");
-        if (cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1) {
-            DEBUGLOG(4, "error : pledgedSrcSize = %u, while realSrcSize = %u",
-                (unsigned)cctx->pledgedSrcSizePlusOne-1, (unsigned)cctx->consumedSrcSize);
-            return ERROR(srcSize_wrong);
-    }   }
+        RETURN_ERROR_IF(
+            cctx->pledgedSrcSizePlusOne != cctx->consumedSrcSize+1,
+            srcSize_wrong,
+             "error : pledgedSrcSize = %u, while realSrcSize = %u",
+            (unsigned)cctx->pledgedSrcSizePlusOne-1,
+            (unsigned)cctx->consumedSrcSize);
+    }
     return cSize + endResult;
 }
 
@@ -3339,7 +2936,7 b' size_t ZSTD_compress_advanced (ZSTD_CCtx'
                                ZSTD_parameters params)
 {
     DEBUGLOG(4, "ZSTD_compress_advanced");
-    CHECK_F(ZSTD_checkCParams(params.cParams));
+    FORWARD_IF_ERROR(ZSTD_checkCParams(params.cParams));
     return ZSTD_compress_internal(cctx,
                                   dst, dstCapacity,
                                   src, srcSize,
@@ -3356,7 +2953,7 b' size_t ZSTD_compress_advanced_internal('
         ZSTD_CCtx_params params)
 {
     DEBUGLOG(4, "ZSTD_compress_advanced_internal (srcSize:%u)", (unsigned)srcSize);
-    CHECK_F( ZSTD_compressBegin_internal(cctx,
+    FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
                          dict, dictSize, ZSTD_dct_auto, ZSTD_dtlm_fast, NULL,
                          params, srcSize, ZSTDb_not_buffered) );
     return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
@@ -3440,17 +3037,17 b' static size_t ZSTD_initCDict_internal('
         void* const internalBuffer = ZSTD_malloc(dictSize, cdict->customMem);
         cdict->dictBuffer = internalBuffer;
         cdict->dictContent = internalBuffer;
-        if (!internalBuffer) return ERROR(memory_allocation);
+        RETURN_ERROR_IF(!internalBuffer, memory_allocation);
         memcpy(internalBuffer, dictBuffer, dictSize);
     }
     cdict->dictContentSize = dictSize;
 
     /* Reset the state to no dictionary */
     ZSTD_reset_compressedBlockState(&cdict->cBlockState);
-    {   void* const end = ZSTD_reset_matchState(
-                &cdict->matchState,
-                (U32*)cdict->workspace + HUF_WORKSPACE_SIZE_U32,
-                &cParams, ZSTDcrp_continue, /* forCCtx */ 0);
+    {   void* const end = ZSTD_reset_matchState(&cdict->matchState,
+                            (U32*)cdict->workspace + HUF_WORKSPACE_SIZE_U32,
+                            &cParams,
+                             ZSTDcrp_continue, ZSTD_resetTarget_CDict);
         assert(end == (char*)cdict->workspace + cdict->workspaceSize);
         (void)end;
     }
@@ -3466,7 +3063,7 b' static size_t ZSTD_initCDict_internal('
                     &cdict->cBlockState, &cdict->matchState, &params,
                     cdict->dictContent, cdict->dictContentSize,
                     dictContentType, ZSTD_dtlm_full, cdict->workspace);
-            if (ZSTD_isError(dictID)) return dictID;
+            FORWARD_IF_ERROR(dictID);
             assert(dictID <= (size_t)(U32)-1);
             cdict->dictID = (U32)dictID;
         }
@@ -3596,7 +3193,7 b' size_t ZSTD_compressBegin_usingCDict_adv'
     ZSTD_frameParameters const fParams, unsigned long long const pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTD_compressBegin_usingCDict_advanced");
-    if (cdict==NULL) return ERROR(dictionary_wrong);
+    RETURN_ERROR_IF(cdict==NULL, dictionary_wrong);
     {   ZSTD_CCtx_params params = cctx->requestedParams;
         params.cParams = ZSTD_getCParamsFromCDict(cdict);
         /* Increase window log to fit the entire dictionary and source if the
@@ -3632,7 +3229,7 b' size_t ZSTD_compress_usingCDict_advanced'
                                 const void* src, size_t srcSize,
                                 const ZSTD_CDict* cdict, ZSTD_frameParameters fParams)
 {
-    CHECK_F (ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize));   /* will check if cdict != NULL */
+    FORWARD_IF_ERROR(ZSTD_compressBegin_usingCDict_advanced(cctx, cdict, fParams, srcSize));   /* will check if cdict != NULL */
     return ZSTD_compressEnd(cctx, dst, dstCapacity, src, srcSize);
 }
 
@@ -3700,7 +3297,7 b' static size_t ZSTD_resetCStream_internal'
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
     assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
 
-    CHECK_F( ZSTD_compressBegin_internal(cctx,
+    FORWARD_IF_ERROR( ZSTD_compressBegin_internal(cctx,
                                          dict, dictSize, dictContentType, ZSTD_dtlm_fast,
                                          cdict,
                                          params, pledgedSrcSize,
@@ -3718,13 +3315,17 b' static size_t ZSTD_resetCStream_internal'
 
 /* ZSTD_resetCStream():
  * pledgedSrcSize == 0 means "unknown" */
-size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pledgedSrcSize)
+size_t ZSTD_resetCStream(ZSTD_CStream* zcs, unsigned long long pss)
 {
-    ZSTD_CCtx_params params = zcs->requestedParams;
+    /* temporary : 0 interpreted as "unknown" during transition period.
+     * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.
+     * 0 will be interpreted as "empty" in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
     DEBUGLOG(4, "ZSTD_resetCStream: pledgedSrcSize = %u", (unsigned)pledgedSrcSize);
-    if (pledgedSrcSize==0) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;
-    params.fParams.contentSizeFlag = 1;
-    return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dct_auto, zcs->cdict, params, pledgedSrcSize);
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
+    return 0;
 }
 
 /*! ZSTD_initCStream_internal() :
@@ -3736,32 +3337,18 b' size_t ZSTD_initCStream_internal(ZSTD_CS'
                     ZSTD_CCtx_params params, unsigned long long pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTD_initCStream_internal");
-    params.cParams = ZSTD_getCParamsFromCCtxParams(&params, pledgedSrcSize, dictSize);
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
     assert(!ZSTD_isError(ZSTD_checkCParams(params.cParams)));
+    zcs->requestedParams = params;
     assert(!((dict) && (cdict)));  /* either dict or cdict, not both */
-
-    if (dict && dictSize >= 8) {
-        DEBUGLOG(4, "loading dictionary of size %u", (unsigned)dictSize);
-        if (zcs->staticSize) {   /* static CCtx : never uses malloc */
-            /* incompatible with internal cdict creation */
-            return ERROR(memory_allocation);
-        }
-        ZSTD_freeCDict(zcs->cdictLocal);
-        zcs->cdictLocal = ZSTD_createCDict_advanced(dict, dictSize,
-                                            ZSTD_dlm_byCopy, ZSTD_dct_auto,
-                                            params.cParams, zcs->customMem);
-        zcs->cdict = zcs->cdictLocal;
-        if (zcs->cdictLocal == NULL) return ERROR(memory_allocation);
+    if (dict) {
+        FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );
     } else {
-        if (cdict) {
-            params.cParams = ZSTD_getCParamsFromCDict(cdict);  /* cParams are enforced from cdict; it includes windowLog */
-        }
-        ZSTD_freeCDict(zcs->cdictLocal);
-        zcs->cdictLocal = NULL;
-        zcs->cdict = cdict;
+        /* Dictionary is cleared if !cdict */
+        FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
     }
-
-    return ZSTD_resetCStream_internal(zcs, NULL, 0, ZSTD_dct_auto, zcs->cdict, params, pledgedSrcSize);
+    return 0;
 }
 
 /* ZSTD_initCStream_usingCDict_advanced() :
@@ -3772,22 +3359,20 b' size_t ZSTD_initCStream_usingCDict_advan'
                                             unsigned long long pledgedSrcSize)
 {
     DEBUGLOG(4, "ZSTD_initCStream_usingCDict_advanced");
-    if (!cdict) return ERROR(dictionary_wrong); /* cannot handle NULL cdict (does not know what to do) */
-    {   ZSTD_CCtx_params params = zcs->requestedParams;
-        params.cParams = ZSTD_getCParamsFromCDict(cdict);
-        params.fParams = fParams;
-        return ZSTD_initCStream_internal(zcs,
-                                NULL, 0, cdict,
-                                params, pledgedSrcSize);
-    }
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
+    zcs->requestedParams.fParams = fParams;
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
+    return 0;
 }
 
 /* note : cdict must outlive compression session */
 size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict)
 {
-    ZSTD_frameParameters const fParams = { 0 /* contentSizeFlag */, 0 /* checksum */, 0 /* hideDictID */ };
     DEBUGLOG(4, "ZSTD_initCStream_usingCDict");
-    return ZSTD_initCStream_usingCDict_advanced(zcs, cdict, fParams, ZSTD_CONTENTSIZE_UNKNOWN);  /* note : will check that cdict != NULL */
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, cdict) );
+    return 0;
 }
 
 
@@ -3797,33 +3382,53 b' size_t ZSTD_initCStream_usingCDict(ZSTD_'
  * dict is loaded with default parameters ZSTD_dm_auto and ZSTD_dlm_byCopy. */
 size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs,
                                  const void* dict, size_t dictSize,
-                                 ZSTD_parameters params, unsigned long long pledgedSrcSize)
+                                 ZSTD_parameters params, unsigned long long pss)
 {
-    DEBUGLOG(4, "ZSTD_initCStream_advanced: pledgedSrcSize=%u, flag=%u",
-                (unsigned)pledgedSrcSize, params.fParams.contentSizeFlag);
-    CHECK_F( ZSTD_checkCParams(params.cParams) );
-    if ((pledgedSrcSize==0) && (params.fParams.contentSizeFlag==0)) pledgedSrcSize = ZSTD_CONTENTSIZE_UNKNOWN;  /* for compatibility with older programs relying on this behavior. Users should now specify ZSTD_CONTENTSIZE_UNKNOWN. This line will be removed in the future. */
+    /* for compatibility with older programs relying on this behavior.
+     * Users should now specify ZSTD_CONTENTSIZE_UNKNOWN.
+     * This line will be removed in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0 && params.fParams.contentSizeFlag==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
+    DEBUGLOG(4, "ZSTD_initCStream_advanced");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
+    FORWARD_IF_ERROR( ZSTD_checkCParams(params.cParams) );
     zcs->requestedParams = ZSTD_assignParamsToCCtxParams(zcs->requestedParams, params);
-    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL /*cdict*/, zcs->requestedParams, pledgedSrcSize);
+    FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );
+    return 0;
 }
 
 size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel)
 {
-    ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel);
-    return ZSTD_initCStream_internal(zcs, dict, dictSize, NULL, zcs->requestedParams, ZSTD_CONTENTSIZE_UNKNOWN);
+    DEBUGLOG(4, "ZSTD_initCStream_usingDict");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_loadDictionary(zcs, dict, dictSize) );
+    return 0;
 }
 
 size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pss)
 {
-    U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;  /* temporary : 0 interpreted as "unknown" during transition period. Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN. `0` will be interpreted as "empty" in the future */
-    ZSTD_CCtxParams_init(&zcs->requestedParams, compressionLevel);
-    return ZSTD_initCStream_internal(zcs, NULL, 0, NULL, zcs->requestedParams, pledgedSrcSize);
+    /* temporary : 0 interpreted as "unknown" during transition period.
+     * Users willing to specify "unknown" **must** use ZSTD_CONTENTSIZE_UNKNOWN.
+     * 0 will be interpreted as "empty" in the future.
+     */
+    U64 const pledgedSrcSize = (pss==0) ? ZSTD_CONTENTSIZE_UNKNOWN : pss;
+    DEBUGLOG(4, "ZSTD_initCStream_srcSize");
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize) );
+    return 0;
 }
 
 size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel)
 {
     DEBUGLOG(4, "ZSTD_initCStream");
-    return ZSTD_initCStream_srcSize(zcs, compressionLevel, ZSTD_CONTENTSIZE_UNKNOWN);
+    FORWARD_IF_ERROR( ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_refCDict(zcs, NULL) );
+    FORWARD_IF_ERROR( ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel) );
+    return 0;
 }
 
 /*======   Compression   ======*/
@@ -3847,10 +3452,10 b' static size_t ZSTD_limitCopy(void* dst, '
  *  internal function for all *compressStream*() variants
  *  non-static, because can be called from zstdmt_compress.c
  * @return : hint size for next input */
-size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
-                                   ZSTD_outBuffer* output,
-                                   ZSTD_inBuffer* input,
-                                   ZSTD_EndDirective const flushMode)
+static size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
+                                          ZSTD_outBuffer* output,
+                                          ZSTD_inBuffer* input,
+                                          ZSTD_EndDirective const flushMode)
 {
     const char* const istart = (const char*)input->src;
     const char* const iend = istart + input->size;
@@ -3873,8 +3478,7 b' size_t ZSTD_compressStream_generic(ZSTD_'
         switch(zcs->streamStage)
         {
         case zcss_init:
-            /* call ZSTD_initCStream() first ! */
-            return ERROR(init_missing);
+            RETURN_ERROR(init_missing, "call ZSTD_initCStream() first!");
 
         case zcss_load:
             if ( (flushMode == ZSTD_e_end)
@@ -3884,7 +3488,7 b' size_t ZSTD_compressStream_generic(ZSTD_'
                 size_t const cSize = ZSTD_compressEnd(zcs,
                                                 op, oend-op, ip, iend-ip);
                 DEBUGLOG(4, "ZSTD_compressEnd : cSize=%u", (unsigned)cSize);
-                if (ZSTD_isError(cSize)) return cSize;
+                FORWARD_IF_ERROR(cSize);
                 ip = iend;
                 op += cSize;
                 zcs->frameEnded = 1;
@@ -3925,7 +3529,7 b' size_t ZSTD_compressStream_generic(ZSTD_'
                                     zcs->inBuff + zcs->inToCompress, iSize) :
                         ZSTD_compressContinue(zcs, cDst, oSize,
                                     zcs->inBuff + zcs->inToCompress, iSize);
-                if (ZSTD_isError(cSize)) return cSize;
+                FORWARD_IF_ERROR(cSize);
                 zcs->frameEnded = lastBlock;
                 /* prepare next block */
                 zcs->inBuffTarget = zcs->inBuffPos + zcs->blockSize;
@@ -3953,7 +3557,7 b' size_t ZSTD_compressStream_generic(ZSTD_'
         case zcss_flush:
             DEBUGLOG(5, "flush stage");
             {   size_t const toFlush = zcs->outBuffContentSize - zcs->outBuffFlushedSize;
-                size_t const flushed = ZSTD_limitCopy(op, oend-op,
+                size_t const flushed = ZSTD_limitCopy(op, (size_t)(oend-op),
                             zcs->outBuff + zcs->outBuffFlushedSize, toFlush);
                 DEBUGLOG(5, "toFlush: %u into %u ==> flushed: %u",
                             (unsigned)toFlush, (unsigned)(oend-op), (unsigned)flushed);
@@ -4001,7 +3605,7 b' static size_t ZSTD_nextInputSizeHint_MTo'
 
 size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
 {
-    CHECK_F( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) );
+    FORWARD_IF_ERROR( ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue) );
     return ZSTD_nextInputSizeHint_MTorST(zcs);
 }
 
@@ -4013,14 +3617,15 b' size_t ZSTD_compressStream2( ZSTD_CCtx* '
 {
     DEBUGLOG(5, "ZSTD_compressStream2, endOp=%u ", (unsigned)endOp);
     /* check conditions */
-    if (output->pos > output->size) return ERROR(GENERIC);
-    if (input->pos  > input->size)  return ERROR(GENERIC);
+    RETURN_ERROR_IF(output->pos > output->size, GENERIC);
+    RETURN_ERROR_IF(input->pos  > input->size, GENERIC);
     assert(cctx!=NULL);
 
     /* transparent initialization stage */
     if (cctx->streamStage == zcss_init) {
         ZSTD_CCtx_params params = cctx->requestedParams;
         ZSTD_prefixDict const prefixDict = cctx->prefixDict;
+        FORWARD_IF_ERROR( ZSTD_initLocalDict(cctx) ); /* Init the local dict if present. */
         memset(&cctx->prefixDict, 0, sizeof(cctx->prefixDict));   /* single usage */
         assert(prefixDict.dict==NULL || cctx->cdict==NULL);    /* only one can be set */
         DEBUGLOG(4, "ZSTD_compressStream2 : transparent init stage");
@@ -4039,11 +3644,11 b' size_t ZSTD_compressStream2( ZSTD_CCtx* '
                 DEBUGLOG(4, "ZSTD_compressStream2: creating new mtctx for nbWorkers=%u",
                             params.nbWorkers);
                 cctx->mtctx = ZSTDMT_createCCtx_advanced(params.nbWorkers, cctx->customMem);
-                if (cctx->mtctx == NULL) return ERROR(memory_allocation);
+                RETURN_ERROR_IF(cctx->mtctx == NULL, memory_allocation);
             }
             /* mt compression */
             DEBUGLOG(4, "call ZSTDMT_initCStream_internal as nbWorkers=%u", params.nbWorkers);
-            CHECK_F( ZSTDMT_initCStream_internal(
+            FORWARD_IF_ERROR( ZSTDMT_initCStream_internal(
                         cctx->mtctx,
                         prefixDict.dict, prefixDict.dictSize, ZSTD_dct_rawContent,
                         cctx->cdict, params, cctx->pledgedSrcSizePlusOne-1) );
@@ -4051,7 +3656,7 b' size_t ZSTD_compressStream2( ZSTD_CCtx* '
             cctx->appliedParams.nbWorkers = params.nbWorkers;
         } else
 #endif
-        {   CHECK_F( ZSTD_resetCStream_internal(cctx,
+        {   FORWARD_IF_ERROR( ZSTD_resetCStream_internal(cctx,
                             prefixDict.dict, prefixDict.dictSize, prefixDict.dictContentType,
                             cctx->cdict,
                             params, cctx->pledgedSrcSizePlusOne-1) );
@@ -4063,20 +3668,30 b' size_t ZSTD_compressStream2( ZSTD_CCtx* '
     /* compression stage */
 #ifdef ZSTD_MULTITHREAD
     if (cctx->appliedParams.nbWorkers > 0) {
+        int const forceMaxProgress = (endOp == ZSTD_e_flush || endOp == ZSTD_e_end);
+        size_t flushMin;
+        assert(forceMaxProgress || endOp == ZSTD_e_continue /* Protection for a new flush type */);
         if (cctx->cParamsChanged) {
             ZSTDMT_updateCParams_whileCompressing(cctx->mtctx, &cctx->requestedParams);
             cctx->cParamsChanged = 0;
         }
-        {   size_t const flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp);
+        do {
+            flushMin = ZSTDMT_compressStream_generic(cctx->mtctx, output, input, endOp);
             if ( ZSTD_isError(flushMin)
               || (endOp == ZSTD_e_end && flushMin == 0) ) { /* compression completed */
                 ZSTD_CCtx_reset(cctx, ZSTD_reset_session_only);
             }
-            DEBUGLOG(5, "completed ZSTD_compressStream2 delegating to ZSTDMT_compressStream_generic");
-            return flushMin;
-    }   }
+            FORWARD_IF_ERROR(flushMin);
+        } while (forceMaxProgress && flushMin != 0 && output->pos < output->size);
+        DEBUGLOG(5, "completed ZSTD_compressStream2 delegating to ZSTDMT_compressStream_generic");
+        /* Either we don't require maximum forward progress, we've finished the
+         * flush, or we are out of output space.
+         */
+        assert(!forceMaxProgress || flushMin == 0 || output->pos == output->size);
+        return flushMin;
+    }
 #endif
-    CHECK_F( ZSTD_compressStream_generic(cctx, output, input, endOp) );
+    FORWARD_IF_ERROR( ZSTD_compressStream_generic(cctx, output, input, endOp) );
     DEBUGLOG(5, "completed ZSTD_compressStream2");
     return cctx->outBuffContentSize - cctx->outBuffFlushedSize; /* remaining to flush */
 }
@@ -4107,10 +3722,10 b' size_t ZSTD_compress2(ZSTD_CCtx* cctx,'
                                         dst, dstCapacity, &oPos,
                                         src, srcSize, &iPos,
                                         ZSTD_e_end);
-        if (ZSTD_isError(result)) return result;
+        FORWARD_IF_ERROR(result);
         if (result != 0) {  /* compression not completed, due to lack of output space */
             assert(oPos == dstCapacity);
-            return ERROR(dstSize_tooSmall);
+            RETURN_ERROR(dstSize_tooSmall);
         }
         assert(iPos == srcSize);   /* all input is expected consumed */
         return oPos;
@@ -4132,11 +3747,11 b' size_t ZSTD_endStream(ZSTD_CStream* zcs,'
 {
     ZSTD_inBuffer input = { NULL, 0, 0 };
     size_t const remainingToFlush = ZSTD_compressStream2(zcs, output, &input, ZSTD_e_end);
-    CHECK_F( remainingToFlush );
+    FORWARD_IF_ERROR( remainingToFlush );
     if (zcs->appliedParams.nbWorkers > 0) return remainingToFlush;   /* minimal estimation */
     /* single thread mode : attempt to calculate remaining to flush more precisely */
     {   size_t const lastBlockSize = zcs->frameEnded ? 0 : ZSTD_BLOCKHEADERSIZE;
-        size_t const checksumSize = zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4;
+        size_t const checksumSize = (size_t)(zcs->frameEnded ? 0 : zcs->appliedParams.fParams.checksumFlag * 4);
         size_t const toFlush = remainingToFlush + lastBlockSize + checksumSize;
         DEBUGLOG(4, "ZSTD_endStream : remaining to flush : %u", (unsigned)toFlush);
         return toFlush;
@@ -4151,7 +3766,7 b' int ZSTD_maxCLevel(void) { return ZSTD_M'
 int ZSTD_minCLevel(void) { return (int)-ZSTD_TARGETLENGTH_MAX; }
 
 static const ZSTD_compressionParameters ZSTD_defaultCParameters[4][ZSTD_MAX_CLEVEL+1] = {
-{   /* "default" - guarantees a monotonically increasing memory budget */
+{   /* "default" - for any srcSize > 256 KB */
     /* W,  C,  H,  S,  L, TL, strat */
     { 19, 12, 13,  1,  6,  1, ZSTD_fast    },  /* base for negative levels */
     { 19, 13, 14,  1,  7,  0, ZSTD_fast    },  /* level  1 */
@@ -4258,13 +3873,13 b' static const ZSTD_compressionParameters '
 };
 
 /*! ZSTD_getCParams() :
-*  @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.
-*   Size values are optional, provide 0 if not known or unused */
+ * @return ZSTD_compressionParameters structure for a selected compression level, srcSize and dictSize.
+ *  Size values are optional, provide 0 if not known or unused */
 ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize)
 {
     size_t const addedSize = srcSizeHint ? 0 : 500;
-    U64 const rSize = srcSizeHint+dictSize ? srcSizeHint+dictSize+addedSize : (U64)-1;
-    U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB);   /* intentional underflow for srcSizeHint == 0 */
+    U64 const rSize = srcSizeHint+dictSize ? srcSizeHint+dictSize+addedSize : ZSTD_CONTENTSIZE_UNKNOWN;  /* intentional overflow for srcSizeHint == ZSTD_CONTENTSIZE_UNKNOWN */
+    U32 const tableID = (rSize <= 256 KB) + (rSize <= 128 KB) + (rSize <= 16 KB);
     int row = compressionLevel;
     DEBUGLOG(5, "ZSTD_getCParams (cLevel=%i)", compressionLevel);
     if (compressionLevel == 0) row = ZSTD_CLEVEL_DEFAULT;   /* 0 == default */
@@ -4272,13 +3887,14 b' ZSTD_compressionParameters ZSTD_getCPara'
     if (compressionLevel > ZSTD_MAX_CLEVEL) row = ZSTD_MAX_CLEVEL;
     {   ZSTD_compressionParameters cp = ZSTD_defaultCParameters[tableID][row];
         if (compressionLevel < 0) cp.targetLength = (unsigned)(-compressionLevel);   /* acceleration factor */
-        return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize);
+        return ZSTD_adjustCParams_internal(cp, srcSizeHint, dictSize);               /* refine parameters based on srcSize & dictSize */
     }
 }
 
 /*! ZSTD_getParams() :
-*   same as ZSTD_getCParams(), but @return a `ZSTD_parameters` object (instead of `ZSTD_compressionParameters`).
-*   All fields of `ZSTD_frameParameters` are set to default (0) */
+ *  same idea as ZSTD_getCParams()
+ * @return a `ZSTD_parameters` structure (instead of `ZSTD_compressionParameters`).
+ *  Fields of `ZSTD_frameParameters` are set to default values */
 ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long srcSizeHint, size_t dictSize) {
     ZSTD_parameters params;
     ZSTD_compressionParameters const cParams = ZSTD_getCParams(compressionLevel, srcSizeHint, dictSize);

contrib/python-zstandard/zstd/compress/zstd_compress_internal.h

@@ -33,13 +33,13 b' extern "C" {'
 ***************************************/
 #define kSearchStrength      8
 #define HASH_READ_SIZE       8
-#define ZSTD_DUBT_UNSORTED_MARK 1   /* For btlazy2 strategy, index 1 now means "unsorted".
+#define ZSTD_DUBT_UNSORTED_MARK 1   /* For btlazy2 strategy, index ZSTD_DUBT_UNSORTED_MARK==1 means "unsorted".
                                        It could be confused for a real successor at index "1", if sorted as larger than its predecessor.
                                        It's not a big deal though : candidate will just be sorted again.
-                                       Additionnally, candidate position 1 will be lost.
+                                       Additionally, candidate position 1 will be lost.
                                        But candidate 1 cannot hide a large tree of candidates, so it's a minimal loss.
-                                       The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be misdhandled after table re-use with a different strategy
-                                       Constant required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
+                                       The benefit is that ZSTD_DUBT_UNSORTED_MARK cannot be mishandled after table re-use with a different strategy.
+                                       This constant is required by ZSTD_compressBlock_btlazy2() and ZSTD_reduceTable_internal() */
 
 
 /*-*************************************
@@ -55,6 +55,14 b' typedef struct ZSTD_prefixDict_s {'
 } ZSTD_prefixDict;
 
 typedef struct {
+    void* dictBuffer;
+    void const* dict;
+    size_t dictSize;
+    ZSTD_dictContentType_e dictContentType;
+    ZSTD_CDict* cdict;
+} ZSTD_localDict;
+
+typedef struct {
     U32 CTable[HUF_CTABLE_SIZE_U32(255)];
     HUF_repeat repeatMode;
 } ZSTD_hufCTables_t;
@@ -107,6 +115,7 b' typedef struct {'
     U32  offCodeSumBasePrice;    /* to compare to log2(offreq)  */
     ZSTD_OptPrice_e priceType;   /* prices can be determined dynamically, or follow a pre-defined cost structure */
     const ZSTD_entropyCTables_t* symbolCosts;  /* pre-calculated dictionary statistics */
+    ZSTD_literalCompressionMode_e literalCompressionMode;
 } optState_t;
 
 typedef struct {
@@ -119,21 +128,26 b' typedef struct {'
     BYTE const* base;       /* All regular indexes relative to this position */
     BYTE const* dictBase;   /* extDict indexes relative to this position */
     U32 dictLimit;          /* below that point, need extDict */
-    U32 lowLimit;           /* below that point, no more data */
+    U32 lowLimit;           /* below that point, no more valid data */
 } ZSTD_window_t;
 
 typedef struct ZSTD_matchState_t ZSTD_matchState_t;
 struct ZSTD_matchState_t {
     ZSTD_window_t window;   /* State for window round buffer management */
-    U32 loadedDictEnd;      /* index of end of dictionary */
+    U32 loadedDictEnd;      /* index of end of dictionary, within context's referential.
+                             * When loadedDictEnd != 0, a dictionary is in use, and still valid.
+                             * This relies on a mechanism to set loadedDictEnd=0 when dictionary is no longer within distance.
+                             * Such mechanism is provided within ZSTD_window_enforceMaxDist() and ZSTD_checkDictValidity().
+                             * When dict referential is copied into active context (i.e. not attached),
+                             * loadedDictEnd == dictSize, since referential starts from zero.
+                             */
     U32 nextToUpdate;       /* index from which to continue table update */
-    U32 nextToUpdate3;      /* index from which to continue table update */
-    U32 hashLog3;           /* dispatch table : larger == faster, more memory */
+    U32 hashLog3;           /* dispatch table for matches of len==3 : larger == faster, more memory */
     U32* hashTable;
     U32* hashTable3;
     U32* chainTable;
     optState_t opt;         /* optimal parser state */
-    const ZSTD_matchState_t * dictMatchState;
+    const ZSTD_matchState_t* dictMatchState;
     ZSTD_compressionParameters cParams;
 };
 
@@ -186,8 +200,12 b' struct ZSTD_CCtx_params_s {'
     int compressionLevel;
     int forceWindow;           /* force back-references to respect limit of
                                 * 1<<wLog, even for dictionary */
+    size_t targetCBlockSize;   /* Tries to fit compressed block size to be around targetCBlockSize.
+                                * No target when targetCBlockSize == 0.
+                                * There is no guarantee on compressed block size */
 
     ZSTD_dictAttachPref_e attachDictPref;
+    ZSTD_literalCompressionMode_e literalCompressionMode;
 
     /* Multithreading: used to pass parameters to mtctx */
     int nbWorkers;
@@ -243,7 +261,7 b' struct ZSTD_CCtx_s {'
     U32    frameEnded;
 
     /* Dictionary */
-    ZSTD_CDict* cdictLocal;
+    ZSTD_localDict localDict;
     const ZSTD_CDict* cdict;
     ZSTD_prefixDict prefixDict;   /* single-usage dictionary */
 
@@ -295,6 +313,30 b' MEM_STATIC U32 ZSTD_MLcode(U32 mlBase)'
     return (mlBase > 127) ? ZSTD_highbit32(mlBase) + ML_deltaCode : ML_Code[mlBase];
 }
 
+/* ZSTD_cParam_withinBounds:
+ * @return 1 if value is within cParam bounds,
+ * 0 otherwise */
+MEM_STATIC int ZSTD_cParam_withinBounds(ZSTD_cParameter cParam, int value)
+{
+    ZSTD_bounds const bounds = ZSTD_cParam_getBounds(cParam);
+    if (ZSTD_isError(bounds.error)) return 0;
+    if (value < bounds.lowerBound) return 0;
+    if (value > bounds.upperBound) return 0;
+    return 1;
+}
+
+/* ZSTD_minGain() :
+ * minimum compression required
+ * to generate a compress block or a compressed literals section.
+ * note : use same formula for both situations */
+MEM_STATIC size_t ZSTD_minGain(size_t srcSize, ZSTD_strategy strat)
+{
+    U32 const minlog = (strat>=ZSTD_btultra) ? (U32)(strat) - 1 : 6;
+    ZSTD_STATIC_ASSERT(ZSTD_btultra == 8);
+    assert(ZSTD_cParam_withinBounds(ZSTD_c_strategy, strat));
+    return (srcSize >> minlog) + 2;
+}
+
 /*! ZSTD_storeSeq() :
  *  Store a sequence (literal length, literals, offset code and match length code) into seqStore_t.
  *  `offsetCode` : distance to match + 3 (values 1-3 are repCodes).
@@ -314,7 +356,7 b' MEM_STATIC void ZSTD_storeSeq(seqStore_t'
     /* copy Literals */
     assert(seqStorePtr->maxNbLit <= 128 KB);
     assert(seqStorePtr->lit + litLength <= seqStorePtr->litStart + seqStorePtr->maxNbLit);
-    ZSTD_wildcopy(seqStorePtr->lit, literals, litLength);
+    ZSTD_wildcopy(seqStorePtr->lit, literals, (ptrdiff_t)litLength, ZSTD_no_overlap);
     seqStorePtr->lit += litLength;
 
     /* literal Length */
@@ -554,6 +596,9 b' MEM_STATIC U64 ZSTD_rollingHash_rotate(U'
 /*-*************************************
 *  Round buffer management
 ***************************************/
+#if (ZSTD_WINDOWLOG_MAX_64 > 31)
+# error "ZSTD_WINDOWLOG_MAX is too large : would overflow ZSTD_CURRENT_MAX"
+#endif
 /* Max current allowed */
 #define ZSTD_CURRENT_MAX ((3U << 29) + (1U << ZSTD_WINDOWLOG_MAX))
 /* Maximum chunk size before overflow correction needs to be called again */
@@ -665,31 +710,49 b' MEM_STATIC U32 ZSTD_window_correctOverfl'
  * Updates lowLimit so that:
  *    (srcEnd - base) - lowLimit == maxDist + loadedDictEnd
  *
- * This allows a simple check that index >= lowLimit to see if index is valid.
- * This must be called before a block compression call, with srcEnd as the block
- * source end.
+ * It ensures index is valid as long as index >= lowLimit.
+ * This must be called before a block compression call.
+ *
+ * loadedDictEnd is only defined if a dictionary is in use for current compression.
+ * As the name implies, loadedDictEnd represents the index at end of dictionary.
+ * The value lies within context's referential, it can be directly compared to blockEndIdx.
  *
- * If loadedDictEndPtr is not NULL, we set it to zero once we update lowLimit.
- * This is because dictionaries are allowed to be referenced as long as the last
- * byte of the dictionary is in the window, but once they are out of range,
- * they cannot be referenced. If loadedDictEndPtr is NULL, we use
- * loadedDictEnd == 0.
+ * If loadedDictEndPtr is NULL, no dictionary is in use, and we use loadedDictEnd == 0.
+ * If loadedDictEndPtr is not NULL, we set it to zero after updating lowLimit.
+ * This is because dictionaries are allowed to be referenced fully
+ * as long as the last byte of the dictionary is in the window.
+ * Once input has progressed beyond window size, dictionary cannot be referenced anymore.
  *
- * In normal dict mode, the dict is between lowLimit and dictLimit. In
- * dictMatchState mode, lowLimit and dictLimit are the same, and the dictionary
- * is below them. forceWindow and dictMatchState are therefore incompatible.
+ * In normal dict mode, the dictionary lies between lowLimit and dictLimit.
+ * In dictMatchState mode, lowLimit and dictLimit are the same,
+ * and the dictionary is below them.
+ * forceWindow and dictMatchState are therefore incompatible.
  */
 MEM_STATIC void
 ZSTD_window_enforceMaxDist(ZSTD_window_t* window,
-                           void const* srcEnd,
-                           U32 maxDist,
-                           U32* loadedDictEndPtr,
+                     const void* blockEnd,
+                           U32   maxDist,
+                           U32*  loadedDictEndPtr,
                      const ZSTD_matchState_t** dictMatchStatePtr)
 {
-    U32 const blockEndIdx = (U32)((BYTE const*)srcEnd - window->base);
-    U32 loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;
-    DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u",
-                (unsigned)blockEndIdx, (unsigned)maxDist);
+    U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
+    U32 const loadedDictEnd = (loadedDictEndPtr != NULL) ? *loadedDictEndPtr : 0;
+    DEBUGLOG(5, "ZSTD_window_enforceMaxDist: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
+                (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
+
+    /* - When there is no dictionary : loadedDictEnd == 0.
+         In which case, the test (blockEndIdx > maxDist) is merely to avoid
+         overflowing next operation `newLowLimit = blockEndIdx - maxDist`.
+       - When there is a standard dictionary :
+         Index referential is copied from the dictionary,
+         which means it starts from 0.
+         In which case, loadedDictEnd == dictSize,
+         and it makes sense to compare `blockEndIdx > maxDist + dictSize`
+         since `blockEndIdx` also starts from zero.
+       - When there is an attached dictionary :
+         loadedDictEnd is expressed within the referential of the context,
+         so it can be directly compared against blockEndIdx.
+    */
     if (blockEndIdx > maxDist + loadedDictEnd) {
         U32 const newLowLimit = blockEndIdx - maxDist;
         if (window->lowLimit < newLowLimit) window->lowLimit = newLowLimit;
@@ -698,11 +761,45 b' ZSTD_window_enforceMaxDist(ZSTD_window_t'
                         (unsigned)window->dictLimit, (unsigned)window->lowLimit);
             window->dictLimit = window->lowLimit;
         }
-        if (loadedDictEndPtr)
+        /* On reaching window size, dictionaries are invalidated */
+        if (loadedDictEndPtr) *loadedDictEndPtr = 0;
+        if (dictMatchStatePtr) *dictMatchStatePtr = NULL;
+    }
+}
+
+/* Similar to ZSTD_window_enforceMaxDist(),
+ * but only invalidates dictionary
+ * when input progresses beyond window size.
+ * assumption : loadedDictEndPtr and dictMatchStatePtr are valid (non NULL)
+ *              loadedDictEnd uses same referential as window->base
+ *              maxDist is the window size */
+MEM_STATIC void
+ZSTD_checkDictValidity(const ZSTD_window_t* window,
+                       const void* blockEnd,
+                             U32   maxDist,
+                             U32*  loadedDictEndPtr,
+                       const ZSTD_matchState_t** dictMatchStatePtr)
+{
+    assert(loadedDictEndPtr != NULL);
+    assert(dictMatchStatePtr != NULL);
+    {   U32 const blockEndIdx = (U32)((BYTE const*)blockEnd - window->base);
+        U32 const loadedDictEnd = *loadedDictEndPtr;
+        DEBUGLOG(5, "ZSTD_checkDictValidity: blockEndIdx=%u, maxDist=%u, loadedDictEnd=%u",
+                    (unsigned)blockEndIdx, (unsigned)maxDist, (unsigned)loadedDictEnd);
+        assert(blockEndIdx >= loadedDictEnd);
+
+        if (blockEndIdx > loadedDictEnd + maxDist) {
+            /* On reaching window size, dictionaries are invalidated.
+             * For simplification, if window size is reached anywhere within next block,
+             * the dictionary is invalidated for the full block.
+             */
+            DEBUGLOG(6, "invalidating dictionary for current block (distance > windowSize)");
             *loadedDictEndPtr = 0;
-        if (dictMatchStatePtr)
             *dictMatchStatePtr = NULL;
-    }
+        } else {
+            if (*loadedDictEndPtr != 0) {
+                DEBUGLOG(6, "dictionary considered valid for current block");
+    }   }   }
 }
 
 /**
@@ -744,6 +841,17 b' MEM_STATIC U32 ZSTD_window_update(ZSTD_w'
     return contiguous;
 }
 
+MEM_STATIC U32 ZSTD_getLowestMatchIndex(const ZSTD_matchState_t* ms, U32 current, unsigned windowLog)
+{
+    U32    const maxDistance = 1U << windowLog;
+    U32    const lowestValid = ms->window.lowLimit;
+    U32    const withinWindow = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid;
+    U32    const isDictionary = (ms->loadedDictEnd != 0);
+    U32    const matchLowest = isDictionary ? lowestValid : withinWindow;
+    return matchLowest;
+}
+
+
 
 /* debug functions */
 #if (DEBUGLEVEL>=2)
@@ -806,13 +914,6 b' size_t ZSTD_initCStream_internal(ZSTD_CS'
 
 void ZSTD_resetSeqStore(seqStore_t* ssPtr);
 
-/*! ZSTD_compressStream_generic() :
- *  Private use only. To be called from zstdmt_compress.c in single-thread mode. */
-size_t ZSTD_compressStream_generic(ZSTD_CStream* zcs,
-                                   ZSTD_outBuffer* output,
-                                   ZSTD_inBuffer* input,
-                                   ZSTD_EndDirective const flushMode);
-
 /*! ZSTD_getCParamsFromCDict() :
  *  as the name implies */
 ZSTD_compressionParameters ZSTD_getCParamsFromCDict(const ZSTD_CDict* cdict);
@@ -839,7 +940,7 b' size_t ZSTD_compress_advanced_internal(Z'
 /* ZSTD_writeLastEmptyBlock() :
  * output an empty Block with end-of-frame mark to complete a frame
  * @return : size of data written into `dst` (== ZSTD_blockHeaderSize (defined in zstd_internal.h))
- *           or an error code if `dstCapcity` is too small (<ZSTD_blockHeaderSize)
+ *           or an error code if `dstCapacity` is too small (<ZSTD_blockHeaderSize)
  */
 size_t ZSTD_writeLastEmptyBlock(void* dst, size_t dstCapacity);
 

contrib/python-zstandard/zstd/compress/zstd_double_fast.c

@@ -43,8 +43,7 b' void ZSTD_fillDoubleHashTable(ZSTD_match'
             /* Only load extra positions for ZSTD_dtlm_full */
             if (dtlm == ZSTD_dtlm_fast)
                 break;
-        }
-    }
+    }   }
 }
 
 
@@ -63,7 +62,11 b' size_t ZSTD_compressBlock_doubleFast_gen'
     const BYTE* const istart = (const BYTE*)src;
     const BYTE* ip = istart;
     const BYTE* anchor = istart;
-    const U32 prefixLowestIndex = ms->window.dictLimit;
+    const U32 endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32 lowestValid = ms->window.dictLimit;
+    const U32 maxDistance = 1U << cParams->windowLog;
+    /* presumes that, if there is a dictionary, it must be using Attach mode */
+    const U32 prefixLowestIndex = (endIndex - lowestValid > maxDistance) ? endIndex - maxDistance : lowestValid;
     const BYTE* const prefixLowest = base + prefixLowestIndex;
     const BYTE* const iend = istart + srcSize;
     const BYTE* const ilimit = iend - HASH_READ_SIZE;
@@ -95,8 +98,15 b' size_t ZSTD_compressBlock_doubleFast_gen'
                                      dictCParams->chainLog : hBitsS;
     const U32 dictAndPrefixLength  = (U32)(ip - prefixLowest + dictEnd - dictStart);
 
+    DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_generic");
+
     assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
 
+    /* if a dictionary is attached, it must be within window range */
+    if (dictMode == ZSTD_dictMatchState) {
+        assert(lowestValid + maxDistance >= endIndex);
+    }
+
     /* init */
     ip += (dictAndPrefixLength == 0);
     if (dictMode == ZSTD_noDict) {
@@ -138,7 +148,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
             const BYTE* repMatchEnd = repIndex < prefixLowestIndex ? dictEnd : iend;
             mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixLowest) + 4;
             ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
             goto _match_stored;
         }
 
@@ -147,7 +157,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
           && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {
             mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
             ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
             goto _match_stored;
         }
 
@@ -170,8 +180,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
                 offset = (U32)(current - dictMatchIndexL - dictIndexDelta);
                 while (((ip>anchor) & (dictMatchL>dictStart)) && (ip[-1] == dictMatchL[-1])) { ip--; dictMatchL--; mLength++; } /* catch up */
                 goto _match_found;
-            }
-        }
+        }   }
 
         if (matchIndexS > prefixLowestIndex) {
             /* check prefix short match */
@@ -186,16 +195,14 b' size_t ZSTD_compressBlock_doubleFast_gen'
 
             if (match > dictStart && MEM_read32(match) == MEM_read32(ip)) {
                 goto _search_next_long;
-            }
-        }
+        }   }
 
         ip += ((ip-anchor) >> kSearchStrength) + 1;
         continue;
 
 _search_next_long:
 
-        {
-            size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
+        {   size_t const hl3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
             size_t const dictHLNext = ZSTD_hashPtr(ip+1, dictHBitsL, 8);
             U32 const matchIndexL3 = hashLong[hl3];
             const BYTE* matchL3 = base + matchIndexL3;
@@ -221,9 +228,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
                     offset = (U32)(current + 1 - dictMatchIndexL3 - dictIndexDelta);
                     while (((ip>anchor) & (dictMatchL3>dictStart)) && (ip[-1] == dictMatchL3[-1])) { ip--; dictMatchL3--; mLength++; } /* catch up */
                     goto _match_found;
-                }
-            }
-        }
+        }   }   }
 
         /* if no long +1 match, explore the short match we found */
         if (dictMode == ZSTD_dictMatchState && matchIndexS < prefixLowestIndex) {
@@ -242,7 +247,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
         offset_2 = offset_1;
         offset_1 = offset;
 
-        ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+        ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
 _match_stored:
         /* match found */
@@ -250,11 +255,14 b' size_t ZSTD_compressBlock_doubleFast_gen'
         anchor = ip;
 
         if (ip <= ilimit) {
-            /* Fill Table */
-            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] =
-                hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;  /* here because current+2 could be > iend-8 */
-            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] =
-                hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
+            /* Complementary insertion */
+            /* done after iLimit test, as candidates could be > iend-8 */
+            {   U32 const indexToInsert = current+2;
+                hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
+                hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+                hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
+                hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
+            }
 
             /* check immediate repcode */
             if (dictMode == ZSTD_dictMatchState) {
@@ -278,8 +286,7 b' size_t ZSTD_compressBlock_doubleFast_gen'
                         continue;
                     }
                     break;
-                }
-            }
+            }   }
 
             if (dictMode == ZSTD_noDict) {
                 while ( (ip <= ilimit)
@@ -294,14 +301,15 b' size_t ZSTD_compressBlock_doubleFast_gen'
                     ip += rLength;
                     anchor = ip;
                     continue;   /* faster when present ... (?) */
-    }   }   }   }
+        }   }   }
+    }   /* while (ip < ilimit) */
 
     /* save reps for next block */
     rep[0] = offset_1 ? offset_1 : offsetSaved;
     rep[1] = offset_2 ? offset_2 : offsetSaved;
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 
@@ -360,10 +368,13 b' static size_t ZSTD_compressBlock_doubleF'
     const BYTE* anchor = istart;
     const BYTE* const iend = istart + srcSize;
     const BYTE* const ilimit = iend - 8;
-    const U32   prefixStartIndex = ms->window.dictLimit;
     const BYTE* const base = ms->window.base;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
+    const U32   dictStartIndex = lowLimit;
+    const U32   dictLimit = ms->window.dictLimit;
+    const U32   prefixStartIndex = (dictLimit > lowLimit) ? dictLimit : lowLimit;
     const BYTE* const prefixStart = base + prefixStartIndex;
-    const U32   dictStartIndex = ms->window.lowLimit;
     const BYTE* const dictBase = ms->window.dictBase;
     const BYTE* const dictStart = dictBase + dictStartIndex;
     const BYTE* const dictEnd = dictBase + prefixStartIndex;
@@ -371,6 +382,10 b' static size_t ZSTD_compressBlock_doubleF'
 
     DEBUGLOG(5, "ZSTD_compressBlock_doubleFast_extDict_generic (srcSize=%zu)", srcSize);
 
+    /* if extDict is invalidated due to maxDistance, switch to "regular" variant */
+    if (prefixStartIndex == dictStartIndex)
+        return ZSTD_compressBlock_doubleFast_generic(ms, seqStore, rep, src, srcSize, mls, ZSTD_noDict);
+
     /* Search Loop */
     while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
         const size_t hSmall = ZSTD_hashPtr(ip, hBitsS, mls);
@@ -396,7 +411,7 b' static size_t ZSTD_compressBlock_doubleF'
             const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
             mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
             ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
         } else {
             if ((matchLongIndex > dictStartIndex) && (MEM_read64(matchLong) == MEM_read64(ip))) {
                 const BYTE* const matchEnd = matchLongIndex < prefixStartIndex ? dictEnd : iend;
@@ -407,7 +422,7 b' static size_t ZSTD_compressBlock_doubleF'
                 while (((ip>anchor) & (matchLong>lowMatchPtr)) && (ip[-1] == matchLong[-1])) { ip--; matchLong--; mLength++; }   /* catch up */
                 offset_2 = offset_1;
                 offset_1 = offset;
-                ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
             } else if ((matchIndex > dictStartIndex) && (MEM_read32(match) == MEM_read32(ip))) {
                 size_t const h3 = ZSTD_hashPtr(ip+1, hBitsL, 8);
@@ -432,23 +447,27 b' static size_t ZSTD_compressBlock_doubleF'
                 }
                 offset_2 = offset_1;
                 offset_1 = offset;
-                ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
 
             } else {
                 ip += ((ip-anchor) >> kSearchStrength) + 1;
                 continue;
         }   }
 
-        /* found a match : store it */
+        /* move to next sequence start */
         ip += mLength;
         anchor = ip;
 
         if (ip <= ilimit) {
-            /* Fill Table */
-            hashSmall[ZSTD_hashPtr(base+current+2, hBitsS, mls)] = current+2;
-            hashLong[ZSTD_hashPtr(base+current+2, hBitsL, 8)] = current+2;
-            hashSmall[ZSTD_hashPtr(ip-2, hBitsS, mls)] = (U32)(ip-2-base);
-            hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+            /* Complementary insertion */
+            /* done after iLimit test, as candidates could be > iend-8 */
+            {   U32 const indexToInsert = current+2;
+                hashLong[ZSTD_hashPtr(base+indexToInsert, hBitsL, 8)] = indexToInsert;
+                hashLong[ZSTD_hashPtr(ip-2, hBitsL, 8)] = (U32)(ip-2-base);
+                hashSmall[ZSTD_hashPtr(base+indexToInsert, hBitsS, mls)] = indexToInsert;
+                hashSmall[ZSTD_hashPtr(ip-1, hBitsS, mls)] = (U32)(ip-1-base);
+            }
+
             /* check immediate repcode */
             while (ip <= ilimit) {
                 U32 const current2 = (U32)(ip-base);
@@ -475,7 +494,7 b' static size_t ZSTD_compressBlock_doubleF'
     rep[1] = offset_2;
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 

contrib/python-zstandard/zstd/compress/zstd_fast.c

@@ -13,7 +13,8 b''
 
 
 void ZSTD_fillHashTable(ZSTD_matchState_t* ms,
-                        void const* end, ZSTD_dictTableLoadMethod_e dtlm)
+                        const void* const end,
+                        ZSTD_dictTableLoadMethod_e dtlm)
 {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
     U32* const hashTable = ms->hashTable;
@@ -41,11 +42,164 b' void ZSTD_fillHashTable(ZSTD_matchState_'
     }   }   }   }
 }
 
+
 FORCE_INLINE_TEMPLATE
 size_t ZSTD_compressBlock_fast_generic(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize,
-        U32 const mls, ZSTD_dictMode_e const dictMode)
+        U32 const mls)
+{
+    const ZSTD_compressionParameters* const cParams = &ms->cParams;
+    U32* const hashTable = ms->hashTable;
+    U32 const hlog = cParams->hashLog;
+    /* support stepSize of 0 */
+    size_t const stepSize = cParams->targetLength + !(cParams->targetLength) + 1;
+    const BYTE* const base = ms->window.base;
+    const BYTE* const istart = (const BYTE*)src;
+    /* We check ip0 (ip + 0) and ip1 (ip + 1) each loop */
+    const BYTE* ip0 = istart;
+    const BYTE* ip1;
+    const BYTE* anchor = istart;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   maxDistance = 1U << cParams->windowLog;
+    const U32   validStartIndex = ms->window.dictLimit;
+    const U32   prefixStartIndex = (endIndex - validStartIndex > maxDistance) ? endIndex - maxDistance : validStartIndex;
+    const BYTE* const prefixStart = base + prefixStartIndex;
+    const BYTE* const iend = istart + srcSize;
+    const BYTE* const ilimit = iend - HASH_READ_SIZE;
+    U32 offset_1=rep[0], offset_2=rep[1];
+    U32 offsetSaved = 0;
+
+    /* init */
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_generic");
+    ip0 += (ip0 == prefixStart);
+    ip1 = ip0 + 1;
+    {
+        U32 const maxRep = (U32)(ip0 - prefixStart);
+        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
+        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
+    }
+
+    /* Main Search Loop */
+    while (ip1 < ilimit) {   /* < instead of <=, because check at ip0+2 */
+        size_t mLength;
+        BYTE const* ip2 = ip0 + 2;
+        size_t const h0 = ZSTD_hashPtr(ip0, hlog, mls);
+        U32 const val0 = MEM_read32(ip0);
+        size_t const h1 = ZSTD_hashPtr(ip1, hlog, mls);
+        U32 const val1 = MEM_read32(ip1);
+        U32 const current0 = (U32)(ip0-base);
+        U32 const current1 = (U32)(ip1-base);
+        U32 const matchIndex0 = hashTable[h0];
+        U32 const matchIndex1 = hashTable[h1];
+        BYTE const* repMatch = ip2-offset_1;
+        const BYTE* match0 = base + matchIndex0;
+        const BYTE* match1 = base + matchIndex1;
+        U32 offcode;
+        hashTable[h0] = current0;   /* update hash table */
+        hashTable[h1] = current1;   /* update hash table */
+
+        assert(ip0 + 1 == ip1);
+
+        if ((offset_1 > 0) & (MEM_read32(repMatch) == MEM_read32(ip2))) {
+            mLength = ip2[-1] == repMatch[-1] ? 1 : 0;
+            ip0 = ip2 - mLength;
+            match0 = repMatch - mLength;
+            offcode = 0;
+            goto _match;
+        }
+        if ((matchIndex0 > prefixStartIndex) && MEM_read32(match0) == val0) {
+            /* found a regular match */
+            goto _offset;
+        }
+        if ((matchIndex1 > prefixStartIndex) && MEM_read32(match1) == val1) {
+            /* found a regular match after one literal */
+            ip0 = ip1;
+            match0 = match1;
+            goto _offset;
+        }
+        {
+            size_t const step = ((ip0-anchor) >> (kSearchStrength - 1)) + stepSize;
+            assert(step >= 2);
+            ip0 += step;
+            ip1 += step;
+            continue;
+        }
+_offset: /* Requires: ip0, match0 */
+        /* Compute the offset code */
+        offset_2 = offset_1;
+        offset_1 = (U32)(ip0-match0);
+        offcode = offset_1 + ZSTD_REP_MOVE;
+        mLength = 0;
+        /* Count the backwards match length */
+        while (((ip0>anchor) & (match0>prefixStart))
+             && (ip0[-1] == match0[-1])) { ip0--; match0--; mLength++; } /* catch up */
+
+_match: /* Requires: ip0, match0, offcode */
+        /* Count the forward length */
+        mLength += ZSTD_count(ip0+mLength+4, match0+mLength+4, iend) + 4;
+        ZSTD_storeSeq(seqStore, ip0-anchor, anchor, offcode, mLength-MINMATCH);
+        /* match found */
+        ip0 += mLength;
+        anchor = ip0;
+        ip1 = ip0 + 1;
+
+        if (ip0 <= ilimit) {
+            /* Fill Table */
+            assert(base+current0+2 > istart);  /* check base overflow */
+            hashTable[ZSTD_hashPtr(base+current0+2, hlog, mls)] = current0+2;  /* here because current+2 could be > iend-8 */
+            hashTable[ZSTD_hashPtr(ip0-2, hlog, mls)] = (U32)(ip0-2-base);
+
+            while ( (ip0 <= ilimit)
+                 && ( (offset_2>0)
+                    & (MEM_read32(ip0) == MEM_read32(ip0 - offset_2)) )) {
+                /* store sequence */
+                size_t const rLength = ZSTD_count(ip0+4, ip0+4-offset_2, iend) + 4;
+                U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff;  /* swap offset_2 <=> offset_1 */
+                hashTable[ZSTD_hashPtr(ip0, hlog, mls)] = (U32)(ip0-base);
+                ip0 += rLength;
+                ip1 = ip0 + 1;
+                ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
+                anchor = ip0;
+                continue;   /* faster when present (confirmed on gcc-8) ... (?) */
+            }
+        }
+    }
+
+    /* save reps for next block */
+    rep[0] = offset_1 ? offset_1 : offsetSaved;
+    rep[1] = offset_2 ? offset_2 : offsetSaved;
+
+    /* Return the last literals size */
+    return (size_t)(iend - anchor);
+}
+
+
+size_t ZSTD_compressBlock_fast(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize)
+{
+    ZSTD_compressionParameters const* cParams = &ms->cParams;
+    U32 const mls = cParams->minMatch;
+    assert(ms->dictMatchState == NULL);
+    switch(mls)
+    {
+    default: /* includes case 3 */
+    case 4 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4);
+    case 5 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5);
+    case 6 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6);
+    case 7 :
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7);
+    }
+}
+
+FORCE_INLINE_TEMPLATE
+size_t ZSTD_compressBlock_fast_dictMatchState_generic(
+        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
+        void const* src, size_t srcSize, U32 const mls)
 {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
     U32* const hashTable = ms->hashTable;
@@ -64,46 +218,34 b' size_t ZSTD_compressBlock_fast_generic('
     U32 offsetSaved = 0;
 
     const ZSTD_matchState_t* const dms = ms->dictMatchState;
-    const ZSTD_compressionParameters* const dictCParams =
-                                     dictMode == ZSTD_dictMatchState ?
-                                     &dms->cParams : NULL;
-    const U32* const dictHashTable = dictMode == ZSTD_dictMatchState ?
-                                     dms->hashTable : NULL;
-    const U32 dictStartIndex       = dictMode == ZSTD_dictMatchState ?
-                                     dms->window.dictLimit : 0;
-    const BYTE* const dictBase     = dictMode == ZSTD_dictMatchState ?
-                                     dms->window.base : NULL;
-    const BYTE* const dictStart    = dictMode == ZSTD_dictMatchState ?
-                                     dictBase + dictStartIndex : NULL;
-    const BYTE* const dictEnd      = dictMode == ZSTD_dictMatchState ?
-                                     dms->window.nextSrc : NULL;
-    const U32 dictIndexDelta       = dictMode == ZSTD_dictMatchState ?
-                                     prefixStartIndex - (U32)(dictEnd - dictBase) :
-                                     0;
+    const ZSTD_compressionParameters* const dictCParams = &dms->cParams ;
+    const U32* const dictHashTable = dms->hashTable;
+    const U32 dictStartIndex       = dms->window.dictLimit;
+    const BYTE* const dictBase     = dms->window.base;
+    const BYTE* const dictStart    = dictBase + dictStartIndex;
+    const BYTE* const dictEnd      = dms->window.nextSrc;
+    const U32 dictIndexDelta       = prefixStartIndex - (U32)(dictEnd - dictBase);
     const U32 dictAndPrefixLength  = (U32)(ip - prefixStart + dictEnd - dictStart);
-    const U32 dictHLog             = dictMode == ZSTD_dictMatchState ?
-                                     dictCParams->hashLog : hlog;
-
-    assert(dictMode == ZSTD_noDict || dictMode == ZSTD_dictMatchState);
+    const U32 dictHLog             = dictCParams->hashLog;
 
-    /* otherwise, we would get index underflow when translating a dict index
-     * into a local index */
-    assert(dictMode != ZSTD_dictMatchState
-        || prefixStartIndex >= (U32)(dictEnd - dictBase));
+    /* if a dictionary is still attached, it necessarily means that
+     * it is within window size. So we just check it. */
+    const U32 maxDistance = 1U << cParams->windowLog;
+    const U32 endIndex = (U32)((size_t)(ip - base) + srcSize);
+    assert(endIndex - prefixStartIndex <= maxDistance);
+    (void)maxDistance; (void)endIndex;   /* these variables are not used when assert() is disabled */
+
+    /* ensure there will be no no underflow
+     * when translating a dict index into a local index */
+    assert(prefixStartIndex >= (U32)(dictEnd - dictBase));
 
     /* init */
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_dictMatchState_generic");
     ip += (dictAndPrefixLength == 0);
-    if (dictMode == ZSTD_noDict) {
-        U32 const maxRep = (U32)(ip - prefixStart);
-        if (offset_2 > maxRep) offsetSaved = offset_2, offset_2 = 0;
-        if (offset_1 > maxRep) offsetSaved = offset_1, offset_1 = 0;
-    }
-    if (dictMode == ZSTD_dictMatchState) {
-        /* dictMatchState repCode checks don't currently handle repCode == 0
-         * disabling. */
-        assert(offset_1 <= dictAndPrefixLength);
-        assert(offset_2 <= dictAndPrefixLength);
-    }
+    /* dictMatchState repCode checks don't currently handle repCode == 0
+     * disabling. */
+    assert(offset_1 <= dictAndPrefixLength);
+    assert(offset_2 <= dictAndPrefixLength);
 
     /* Main Search Loop */
     while (ip < ilimit) {   /* < instead of <=, because repcode check at (ip+1) */
@@ -113,50 +255,37 b' size_t ZSTD_compressBlock_fast_generic('
         U32 const matchIndex = hashTable[h];
         const BYTE* match = base + matchIndex;
         const U32 repIndex = current + 1 - offset_1;
-        const BYTE* repMatch = (dictMode == ZSTD_dictMatchState
-                            && repIndex < prefixStartIndex) ?
+        const BYTE* repMatch = (repIndex < prefixStartIndex) ?
                                dictBase + (repIndex - dictIndexDelta) :
                                base + repIndex;
         hashTable[h] = current;   /* update hash table */
 
-        if ( (dictMode == ZSTD_dictMatchState)
-          && ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
+        if ( ((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow : ensure repIndex isn't overlapping dict + prefix */
           && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
             const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
             mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
             ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
-        } else if ( dictMode == ZSTD_noDict
-                 && ((offset_1 > 0) & (MEM_read32(ip+1-offset_1) == MEM_read32(ip+1)))) {
-            mLength = ZSTD_count(ip+1+4, ip+1+4-offset_1, iend) + 4;
-            ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
         } else if ( (matchIndex <= prefixStartIndex) ) {
-            if (dictMode == ZSTD_dictMatchState) {
-                size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
-                U32 const dictMatchIndex = dictHashTable[dictHash];
-                const BYTE* dictMatch = dictBase + dictMatchIndex;
-                if (dictMatchIndex <= dictStartIndex ||
-                    MEM_read32(dictMatch) != MEM_read32(ip)) {
-                    assert(stepSize >= 1);
-                    ip += ((ip-anchor) >> kSearchStrength) + stepSize;
-                    continue;
-                } else {
-                    /* found a dict match */
-                    U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
-                    mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
-                    while (((ip>anchor) & (dictMatch>dictStart))
-                         && (ip[-1] == dictMatch[-1])) {
-                        ip--; dictMatch--; mLength++;
-                    } /* catch up */
-                    offset_2 = offset_1;
-                    offset_1 = offset;
-                    ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
-                }
-            } else {
+            size_t const dictHash = ZSTD_hashPtr(ip, dictHLog, mls);
+            U32 const dictMatchIndex = dictHashTable[dictHash];
+            const BYTE* dictMatch = dictBase + dictMatchIndex;
+            if (dictMatchIndex <= dictStartIndex ||
+                MEM_read32(dictMatch) != MEM_read32(ip)) {
                 assert(stepSize >= 1);
                 ip += ((ip-anchor) >> kSearchStrength) + stepSize;
                 continue;
+            } else {
+                /* found a dict match */
+                U32 const offset = (U32)(current-dictMatchIndex-dictIndexDelta);
+                mLength = ZSTD_count_2segments(ip+4, dictMatch+4, iend, dictEnd, prefixStart) + 4;
+                while (((ip>anchor) & (dictMatch>dictStart))
+                     && (ip[-1] == dictMatch[-1])) {
+                    ip--; dictMatch--; mLength++;
+                } /* catch up */
+                offset_2 = offset_1;
+                offset_1 = offset;
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
             }
         } else if (MEM_read32(match) != MEM_read32(ip)) {
             /* it's not a match, and we're not going to check the dictionary */
@@ -171,7 +300,7 b' size_t ZSTD_compressBlock_fast_generic('
                  && (ip[-1] == match[-1])) { ip--; match--; mLength++; } /* catch up */
             offset_2 = offset_1;
             offset_1 = offset;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
         }
 
         /* match found */
@@ -185,70 +314,34 b' size_t ZSTD_compressBlock_fast_generic('
             hashTable[ZSTD_hashPtr(ip-2, hlog, mls)] = (U32)(ip-2-base);
 
             /* check immediate repcode */
-            if (dictMode == ZSTD_dictMatchState) {
-                while (ip <= ilimit) {
-                    U32 const current2 = (U32)(ip-base);
-                    U32 const repIndex2 = current2 - offset_2;
-                    const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
-                            dictBase - dictIndexDelta + repIndex2 :
-                            base + repIndex2;
-                    if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
-                       && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
-                        const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
-                        size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
-                        U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
-                        ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
-                        hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
-                        ip += repLength2;
-                        anchor = ip;
-                        continue;
-                    }
-                    break;
+            while (ip <= ilimit) {
+                U32 const current2 = (U32)(ip-base);
+                U32 const repIndex2 = current2 - offset_2;
+                const BYTE* repMatch2 = repIndex2 < prefixStartIndex ?
+                        dictBase - dictIndexDelta + repIndex2 :
+                        base + repIndex2;
+                if ( ((U32)((prefixStartIndex-1) - (U32)repIndex2) >= 3 /* intentional overflow */)
+                   && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
+                    const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
+                    size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
+                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
+                    hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
+                    ip += repLength2;
+                    anchor = ip;
+                    continue;
                 }
+                break;
             }
-
-            if (dictMode == ZSTD_noDict) {
-                while ( (ip <= ilimit)
-                     && ( (offset_2>0)
-                        & (MEM_read32(ip) == MEM_read32(ip - offset_2)) )) {
-                    /* store sequence */
-                    size_t const rLength = ZSTD_count(ip+4, ip+4-offset_2, iend) + 4;
-                    U32 const tmpOff = offset_2; offset_2 = offset_1; offset_1 = tmpOff;  /* swap offset_2 <=> offset_1 */
-                    hashTable[ZSTD_hashPtr(ip, hlog, mls)] = (U32)(ip-base);
-                    ZSTD_storeSeq(seqStore, 0, anchor, 0, rLength-MINMATCH);
-                    ip += rLength;
-                    anchor = ip;
-                    continue;   /* faster when present ... (?) */
-    }   }   }   }
+        }
+    }
 
     /* save reps for next block */
     rep[0] = offset_1 ? offset_1 : offsetSaved;
     rep[1] = offset_2 ? offset_2 : offsetSaved;
 
     /* Return the last literals size */
-    return iend - anchor;
-}
-
-
-size_t ZSTD_compressBlock_fast(
-        ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
-        void const* src, size_t srcSize)
-{
-    ZSTD_compressionParameters const* cParams = &ms->cParams;
-    U32 const mls = cParams->minMatch;
-    assert(ms->dictMatchState == NULL);
-    switch(mls)
-    {
-    default: /* includes case 3 */
-    case 4 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_noDict);
-    case 5 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_noDict);
-    case 6 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_noDict);
-    case 7 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_noDict);
-    }
+    return (size_t)(iend - anchor);
 }
 
 size_t ZSTD_compressBlock_fast_dictMatchState(
@@ -262,13 +355,13 b' size_t ZSTD_compressBlock_fast_dictMatch'
     {
     default: /* includes case 3 */
     case 4 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 4, ZSTD_dictMatchState);
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 4);
     case 5 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 5, ZSTD_dictMatchState);
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 5);
     case 6 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 6, ZSTD_dictMatchState);
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 6);
     case 7 :
-        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, 7, ZSTD_dictMatchState);
+        return ZSTD_compressBlock_fast_dictMatchState_generic(ms, seqStore, rep, src, srcSize, 7);
     }
 }
 
@@ -287,15 +380,24 b' static size_t ZSTD_compressBlock_fast_ex'
     const BYTE* const istart = (const BYTE*)src;
     const BYTE* ip = istart;
     const BYTE* anchor = istart;
-    const U32   dictStartIndex = ms->window.lowLimit;
+    const U32   endIndex = (U32)((size_t)(istart - base) + srcSize);
+    const U32   lowLimit = ZSTD_getLowestMatchIndex(ms, endIndex, cParams->windowLog);
+    const U32   dictStartIndex = lowLimit;
     const BYTE* const dictStart = dictBase + dictStartIndex;
-    const U32   prefixStartIndex = ms->window.dictLimit;
+    const U32   dictLimit = ms->window.dictLimit;
+    const U32   prefixStartIndex = dictLimit < lowLimit ? lowLimit : dictLimit;
     const BYTE* const prefixStart = base + prefixStartIndex;
     const BYTE* const dictEnd = dictBase + prefixStartIndex;
     const BYTE* const iend = istart + srcSize;
     const BYTE* const ilimit = iend - 8;
     U32 offset_1=rep[0], offset_2=rep[1];
 
+    DEBUGLOG(5, "ZSTD_compressBlock_fast_extDict_generic");
+
+    /* switch to "regular" variant if extDict is invalidated due to maxDistance */
+    if (prefixStartIndex == dictStartIndex)
+        return ZSTD_compressBlock_fast_generic(ms, seqStore, rep, src, srcSize, mls);
+
     /* Search Loop */
     while (ip < ilimit) {  /* < instead of <=, because (ip+1) */
         const size_t h = ZSTD_hashPtr(ip, hlog, mls);
@@ -312,10 +414,10 b' static size_t ZSTD_compressBlock_fast_ex'
 
         if ( (((U32)((prefixStartIndex-1) - repIndex) >= 3) /* intentional underflow */ & (repIndex > dictStartIndex))
            && (MEM_read32(repMatch) == MEM_read32(ip+1)) ) {
-            const BYTE* repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
-            mLength = ZSTD_count_2segments(ip+1+4, repMatch+4, iend, repMatchEnd, prefixStart) + 4;
+            const BYTE* const repMatchEnd = repIndex < prefixStartIndex ? dictEnd : iend;
+            mLength = ZSTD_count_2segments(ip+1 +4, repMatch +4, iend, repMatchEnd, prefixStart) + 4;
             ip++;
-            ZSTD_storeSeq(seqStore, ip-anchor, anchor, 0, mLength-MINMATCH);
+            ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, 0, mLength-MINMATCH);
         } else {
             if ( (matchIndex < dictStartIndex) ||
                  (MEM_read32(match) != MEM_read32(ip)) ) {
@@ -323,15 +425,15 b' static size_t ZSTD_compressBlock_fast_ex'
                 ip += ((ip-anchor) >> kSearchStrength) + stepSize;
                 continue;
             }
-            {   const BYTE* matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
-                const BYTE* lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
+            {   const BYTE* const matchEnd = matchIndex < prefixStartIndex ? dictEnd : iend;
+                const BYTE* const lowMatchPtr = matchIndex < prefixStartIndex ? dictStart : prefixStart;
                 U32 offset;
                 mLength = ZSTD_count_2segments(ip+4, match+4, iend, matchEnd, prefixStart) + 4;
                 while (((ip>anchor) & (match>lowMatchPtr)) && (ip[-1] == match[-1])) { ip--; match--; mLength++; }   /* catch up */
                 offset = current - matchIndex;
                 offset_2 = offset_1;
                 offset_1 = offset;
-                ZSTD_storeSeq(seqStore, ip-anchor, anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
+                ZSTD_storeSeq(seqStore, (size_t)(ip-anchor), anchor, offset + ZSTD_REP_MOVE, mLength-MINMATCH);
         }   }
 
         /* found a match : store it */
@@ -351,7 +453,7 b' static size_t ZSTD_compressBlock_fast_ex'
                    && (MEM_read32(repMatch2) == MEM_read32(ip)) ) {
                     const BYTE* const repEnd2 = repIndex2 < prefixStartIndex ? dictEnd : iend;
                     size_t const repLength2 = ZSTD_count_2segments(ip+4, repMatch2+4, iend, repEnd2, prefixStart) + 4;
-                    U32 tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
+                    U32 const tmpOffset = offset_2; offset_2 = offset_1; offset_1 = tmpOffset;   /* swap offset_2 <=> offset_1 */
                     ZSTD_storeSeq(seqStore, 0, anchor, 0, repLength2-MINMATCH);
                     hashTable[ZSTD_hashPtr(ip, hlog, mls)] = current2;
                     ip += repLength2;
@@ -366,7 +468,7 b' static size_t ZSTD_compressBlock_fast_ex'
     rep[1] = offset_2;
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 

contrib/python-zstandard/zstd/compress/zstd_lazy.c

@@ -83,7 +83,10 b' ZSTD_insertDUBT1(ZSTD_matchState_t* ms,'
     U32* largerPtr  = smallerPtr + 1;
     U32 matchIndex = *smallerPtr;   /* this candidate is unsorted : next sorted candidate is reached through *smallerPtr, while *largerPtr contains previous unsorted candidate (which is already saved and can be overwritten) */
     U32 dummy32;   /* to be nullified at the end */
-    U32 const windowLow = ms->window.lowLimit;
+    U32 const windowValid = ms->window.lowLimit;
+    U32 const maxDistance = 1U << cParams->windowLog;
+    U32 const windowLow = (current - windowValid > maxDistance) ? current - maxDistance : windowValid;
+
 
     DEBUGLOG(8, "ZSTD_insertDUBT1(%u) (dictLimit=%u, lowLimit=%u)",
                 current, dictLimit, windowLow);
@@ -239,7 +242,7 b' ZSTD_DUBT_findBestMatch(ZSTD_matchState_'
 
     const BYTE* const base = ms->window.base;
     U32    const current = (U32)(ip-base);
-    U32    const windowLow = ms->window.lowLimit;
+    U32    const windowLow = ZSTD_getLowestMatchIndex(ms, current, cParams->windowLog);
 
     U32*   const bt = ms->chainTable;
     U32    const btLog  = cParams->chainLog - 1;
@@ -490,8 +493,12 b' size_t ZSTD_HcFindBestMatch_generic ('
     const U32 dictLimit = ms->window.dictLimit;
     const BYTE* const prefixStart = base + dictLimit;
     const BYTE* const dictEnd = dictBase + dictLimit;
-    const U32 lowLimit = ms->window.lowLimit;
     const U32 current = (U32)(ip-base);
+    const U32 maxDistance = 1U << cParams->windowLog;
+    const U32 lowestValid = ms->window.lowLimit;
+    const U32 withinMaxDistance = (current - lowestValid > maxDistance) ? current - maxDistance : lowestValid;
+    const U32 isDictionary = (ms->loadedDictEnd != 0);
+    const U32 lowLimit = isDictionary ? lowestValid : withinMaxDistance;
     const U32 minChain = current > chainSize ? current - chainSize : 0;
     U32 nbAttempts = 1U << cParams->searchLog;
     size_t ml=4-1;
@@ -612,12 +619,14 b' FORCE_INLINE_TEMPLATE size_t ZSTD_HcFind'
 /* *******************************
 *  Common parser - lazy strategy
 *********************************/
-FORCE_INLINE_TEMPLATE
-size_t ZSTD_compressBlock_lazy_generic(
+typedef enum { search_hashChain, search_binaryTree } searchMethod_e;
+
+FORCE_INLINE_TEMPLATE size_t
+ZSTD_compressBlock_lazy_generic(
                         ZSTD_matchState_t* ms, seqStore_t* seqStore,
                         U32 rep[ZSTD_REP_NUM],
                         const void* src, size_t srcSize,
-                        const U32 searchMethod, const U32 depth,
+                        const searchMethod_e searchMethod, const U32 depth,
                         ZSTD_dictMode_e const dictMode)
 {
     const BYTE* const istart = (const BYTE*)src;
@@ -633,8 +642,10 b' size_t ZSTD_compressBlock_lazy_generic('
                         ZSTD_matchState_t* ms,
                         const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
     searchMax_f const searchMax = dictMode == ZSTD_dictMatchState ?
-        (searchMethod ? ZSTD_BtFindBestMatch_dictMatchState_selectMLS : ZSTD_HcFindBestMatch_dictMatchState_selectMLS) :
-        (searchMethod ? ZSTD_BtFindBestMatch_selectMLS : ZSTD_HcFindBestMatch_selectMLS);
+        (searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_dictMatchState_selectMLS
+                                         : ZSTD_HcFindBestMatch_dictMatchState_selectMLS) :
+        (searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_selectMLS
+                                         : ZSTD_HcFindBestMatch_selectMLS);
     U32 offset_1 = rep[0], offset_2 = rep[1], savedOffset=0;
 
     const ZSTD_matchState_t* const dms = ms->dictMatchState;
@@ -653,7 +664,6 b' size_t ZSTD_compressBlock_lazy_generic('
 
     /* init */
     ip += (dictAndPrefixLength == 0);
-    ms->nextToUpdate3 = ms->nextToUpdate;
     if (dictMode == ZSTD_noDict) {
         U32 const maxRep = (U32)(ip - prefixLowest);
         if (offset_2 > maxRep) savedOffset = offset_2, offset_2 = 0;
@@ -844,7 +854,7 b' size_t ZSTD_compressBlock_lazy_generic('
     rep[1] = offset_2 ? offset_2 : savedOffset;
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 
@@ -852,56 +862,56 b' size_t ZSTD_compressBlock_btlazy2('
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_noDict);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_noDict);
 }
 
 size_t ZSTD_compressBlock_lazy2(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_noDict);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_noDict);
 }
 
 size_t ZSTD_compressBlock_lazy(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_noDict);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_noDict);
 }
 
 size_t ZSTD_compressBlock_greedy(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_noDict);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_noDict);
 }
 
 size_t ZSTD_compressBlock_btlazy2_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 1, 2, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2, ZSTD_dictMatchState);
 }
 
 size_t ZSTD_compressBlock_lazy2_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 2, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2, ZSTD_dictMatchState);
 }
 
 size_t ZSTD_compressBlock_lazy_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 1, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1, ZSTD_dictMatchState);
 }
 
 size_t ZSTD_compressBlock_greedy_dictMatchState(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, 0, 0, ZSTD_dictMatchState);
+    return ZSTD_compressBlock_lazy_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0, ZSTD_dictMatchState);
 }
 
 
@@ -910,7 +920,7 b' size_t ZSTD_compressBlock_lazy_extDict_g'
                         ZSTD_matchState_t* ms, seqStore_t* seqStore,
                         U32 rep[ZSTD_REP_NUM],
                         const void* src, size_t srcSize,
-                        const U32 searchMethod, const U32 depth)
+                        const searchMethod_e searchMethod, const U32 depth)
 {
     const BYTE* const istart = (const BYTE*)src;
     const BYTE* ip = istart;
@@ -928,12 +938,11 b' size_t ZSTD_compressBlock_lazy_extDict_g'
     typedef size_t (*searchMax_f)(
                         ZSTD_matchState_t* ms,
                         const BYTE* ip, const BYTE* iLimit, size_t* offsetPtr);
-    searchMax_f searchMax = searchMethod ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS;
+    searchMax_f searchMax = searchMethod==search_binaryTree ? ZSTD_BtFindBestMatch_extDict_selectMLS : ZSTD_HcFindBestMatch_extDict_selectMLS;
 
     U32 offset_1 = rep[0], offset_2 = rep[1];
 
     /* init */
-    ms->nextToUpdate3 = ms->nextToUpdate;
     ip += (ip == prefixStart);
 
     /* Match Loop */
@@ -1070,7 +1079,7 b' size_t ZSTD_compressBlock_lazy_extDict_g'
     rep[1] = offset_2;
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 
@@ -1078,7 +1087,7 b' size_t ZSTD_compressBlock_greedy_extDict'
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],
         void const* src, size_t srcSize)
 {
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 0);
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 0);
 }
 
 size_t ZSTD_compressBlock_lazy_extDict(
@@ -1086,7 +1095,7 b' size_t ZSTD_compressBlock_lazy_extDict('
         void const* src, size_t srcSize)
 
 {
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 1);
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 1);
 }
 
 size_t ZSTD_compressBlock_lazy2_extDict(
@@ -1094,7 +1103,7 b' size_t ZSTD_compressBlock_lazy2_extDict('
         void const* src, size_t srcSize)
 
 {
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 0, 2);
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_hashChain, 2);
 }
 
 size_t ZSTD_compressBlock_btlazy2_extDict(
@@ -1102,5 +1111,5 b' size_t ZSTD_compressBlock_btlazy2_extDic'
         void const* src, size_t srcSize)
 
 {
-    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, 1, 2);
+    return ZSTD_compressBlock_lazy_extDict_generic(ms, seqStore, rep, src, srcSize, search_binaryTree, 2);
 }

contrib/python-zstandard/zstd/compress/zstd_lazy.h

@@ -19,7 +19,7 b' extern "C" {'
 
 U32 ZSTD_insertAndFindFirstIndex(ZSTD_matchState_t* ms, const BYTE* ip);
 
-void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue);  /*! used in ZSTD_reduceIndex(). pre-emptively increase value of ZSTD_DUBT_UNSORTED_MARK */
+void ZSTD_preserveUnsortedMark (U32* const table, U32 const size, U32 const reducerValue);  /*! used in ZSTD_reduceIndex(). preemptively increase value of ZSTD_DUBT_UNSORTED_MARK */
 
 size_t ZSTD_compressBlock_btlazy2(
         ZSTD_matchState_t* ms, seqStore_t* seqStore, U32 rep[ZSTD_REP_NUM],

contrib/python-zstandard/zstd/compress/zstd_ldm.c

@@ -429,7 +429,7 b' size_t ZSTD_ldm_generateSequences('
      */
     assert(ldmState->window.nextSrc >= (BYTE const*)src + srcSize);
     /* The input could be very large (in zstdmt), so it must be broken up into
-     * chunks to enforce the maximmum distance and handle overflow correction.
+     * chunks to enforce the maximum distance and handle overflow correction.
      */
     assert(sequences->pos <= sequences->size);
     assert(sequences->size <= sequences->capacity);
@@ -447,7 +447,7 b' size_t ZSTD_ldm_generateSequences('
         if (ZSTD_window_needOverflowCorrection(ldmState->window, chunkEnd)) {
             U32 const ldmHSize = 1U << params->hashLog;
             U32 const correction = ZSTD_window_correctOverflow(
-                &ldmState->window, /* cycleLog */ 0, maxDist, src);
+                &ldmState->window, /* cycleLog */ 0, maxDist, chunkStart);
             ZSTD_ldm_reduceTable(ldmState->hashTable, ldmHSize, correction);
         }
         /* 2. We enforce the maximum offset allowed.

contrib/python-zstandard/zstd/compress/zstd_opt.c

@@ -64,9 +64,15 b' MEM_STATIC double ZSTD_fCost(U32 price)'
 }
 #endif
 
+static int ZSTD_compressedLiterals(optState_t const* const optPtr)
+{
+    return optPtr->literalCompressionMode != ZSTD_lcm_uncompressed;
+}
+
 static void ZSTD_setBasePrices(optState_t* optPtr, int optLevel)
 {
-    optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel);
+    if (ZSTD_compressedLiterals(optPtr))
+        optPtr->litSumBasePrice = WEIGHT(optPtr->litSum, optLevel);
     optPtr->litLengthSumBasePrice = WEIGHT(optPtr->litLengthSum, optLevel);
     optPtr->matchLengthSumBasePrice = WEIGHT(optPtr->matchLengthSum, optLevel);
     optPtr->offCodeSumBasePrice = WEIGHT(optPtr->offCodeSum, optLevel);
@@ -99,6 +105,7 b' ZSTD_rescaleFreqs(optState_t* const optP'
             const BYTE* const src, size_t const srcSize,
                   int const optLevel)
 {
+    int const compressedLiterals = ZSTD_compressedLiterals(optPtr);
     DEBUGLOG(5, "ZSTD_rescaleFreqs (srcSize=%u)", (unsigned)srcSize);
     optPtr->priceType = zop_dynamic;
 
@@ -113,9 +120,10 b' ZSTD_rescaleFreqs(optState_t* const optP'
             /* huffman table presumed generated by dictionary */
             optPtr->priceType = zop_dynamic;
 
-            assert(optPtr->litFreq != NULL);
-            optPtr->litSum = 0;
-            {   unsigned lit;
+            if (compressedLiterals) {
+                unsigned lit;
+                assert(optPtr->litFreq != NULL);
+                optPtr->litSum = 0;
                 for (lit=0; lit<=MaxLit; lit++) {
                     U32 const scaleLog = 11;   /* scale to 2K */
                     U32 const bitCost = HUF_getNbBits(optPtr->symbolCosts->huf.CTable, lit);
@@ -163,10 +171,11 b' ZSTD_rescaleFreqs(optState_t* const optP'
         } else {  /* not a dictionary */
 
             assert(optPtr->litFreq != NULL);
-            {   unsigned lit = MaxLit;
+            if (compressedLiterals) {
+                unsigned lit = MaxLit;
                 HIST_count_simple(optPtr->litFreq, &lit, src, srcSize);   /* use raw first block to init statistics */
+                optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
             }
-            optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
 
             {   unsigned ll;
                 for (ll=0; ll<=MaxLL; ll++)
@@ -190,7 +199,8 b' ZSTD_rescaleFreqs(optState_t* const optP'
 
     } else {   /* new block : re-use previous statistics, scaled down */
 
-        optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
+        if (compressedLiterals)
+            optPtr->litSum = ZSTD_downscaleStat(optPtr->litFreq, MaxLit, 1);
         optPtr->litLengthSum = ZSTD_downscaleStat(optPtr->litLengthFreq, MaxLL, 0);
         optPtr->matchLengthSum = ZSTD_downscaleStat(optPtr->matchLengthFreq, MaxML, 0);
         optPtr->offCodeSum = ZSTD_downscaleStat(optPtr->offCodeFreq, MaxOff, 0);
@@ -207,6 +217,10 b' static U32 ZSTD_rawLiteralsCost(const BY'
                                 int optLevel)
 {
     if (litLength == 0) return 0;
+
+    if (!ZSTD_compressedLiterals(optPtr))
+        return (litLength << 3) * BITCOST_MULTIPLIER;  /* Uncompressed - 8 bytes per literal. */
+
     if (optPtr->priceType == zop_predef)
         return (litLength*6) * BITCOST_MULTIPLIER;  /* 6 bit per literal - no statistic used */
 
@@ -241,13 +255,13 b' static U32 ZSTD_litLengthPrice(U32 const'
  * to provide a cost which is directly comparable to a match ending at same position */
 static int ZSTD_litLengthContribution(U32 const litLength, const optState_t* const optPtr, int optLevel)
 {
-    if (optPtr->priceType >= zop_predef) return WEIGHT(litLength, optLevel);
+    if (optPtr->priceType >= zop_predef) return (int)WEIGHT(litLength, optLevel);
 
     /* dynamic statistics */
     {   U32 const llCode = ZSTD_LLcode(litLength);
-        int const contribution = (LL_bits[llCode] * BITCOST_MULTIPLIER)
-                               + WEIGHT(optPtr->litLengthFreq[0], optLevel)   /* note: log2litLengthSum cancel out */
-                               - WEIGHT(optPtr->litLengthFreq[llCode], optLevel);
+        int const contribution = (int)(LL_bits[llCode] * BITCOST_MULTIPLIER)
+                               + (int)WEIGHT(optPtr->litLengthFreq[0], optLevel)   /* note: log2litLengthSum cancel out */
+                               - (int)WEIGHT(optPtr->litLengthFreq[llCode], optLevel);
 #if 1
         return contribution;
 #else
@@ -264,7 +278,7 b' static int ZSTD_literalsContribution(con'
                                      const optState_t* const optPtr,
                                      int optLevel)
 {
-    int const contribution = ZSTD_rawLiteralsCost(literals, litLength, optPtr, optLevel)
+    int const contribution = (int)ZSTD_rawLiteralsCost(literals, litLength, optPtr, optLevel)
                            + ZSTD_litLengthContribution(litLength, optPtr, optLevel);
     return contribution;
 }
@@ -310,7 +324,8 b' static void ZSTD_updateStats(optState_t*'
                              U32 offsetCode, U32 matchLength)
 {
     /* literals */
-    {   U32 u;
+    if (ZSTD_compressedLiterals(optPtr)) {
+        U32 u;
         for (u=0; u < litLength; u++)
             optPtr->litFreq[literals[u]] += ZSTD_LITFREQ_ADD;
         optPtr->litSum += litLength*ZSTD_LITFREQ_ADD;
@@ -357,13 +372,15 b' MEM_STATIC U32 ZSTD_readMINMATCH(const v'
 
 /* Update hashTable3 up to ip (excluded)
    Assumption : always within prefix (i.e. not within extDict) */
-static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms, const BYTE* const ip)
+static U32 ZSTD_insertAndFindFirstIndexHash3 (ZSTD_matchState_t* ms,
+                                              U32* nextToUpdate3,
+                                              const BYTE* const ip)
 {
     U32* const hashTable3 = ms->hashTable3;
     U32 const hashLog3 = ms->hashLog3;
     const BYTE* const base = ms->window.base;
-    U32 idx = ms->nextToUpdate3;
-    U32 const target = ms->nextToUpdate3 = (U32)(ip - base);
+    U32 idx = *nextToUpdate3;
+    U32 const target = (U32)(ip - base);
     size_t const hash3 = ZSTD_hash3Ptr(ip, hashLog3);
     assert(hashLog3 > 0);
 
@@ -372,6 +389,7 b' static U32 ZSTD_insertAndFindFirstIndexH'
         idx++;
     }
 
+    *nextToUpdate3 = target;
     return hashTable3[hash3];
 }
 
@@ -488,9 +506,11 b' static U32 ZSTD_insertBt1('
     }   }
 
     *smallerPtr = *largerPtr = 0;
-    if (bestLength > 384) return MIN(192, (U32)(bestLength - 384));   /* speed optimization */
-    assert(matchEndIdx > current + 8);
-    return matchEndIdx - (current + 8);
+    {   U32 positions = 0;
+        if (bestLength > 384) positions = MIN(192, (U32)(bestLength - 384));   /* speed optimization */
+        assert(matchEndIdx > current + 8);
+        return MAX(positions, matchEndIdx - (current + 8));
+    }
 }
 
 FORCE_INLINE_TEMPLATE
@@ -505,8 +525,13 b' void ZSTD_updateTree_internal('
     DEBUGLOG(6, "ZSTD_updateTree_internal, from %u to %u  (dictMode:%u)",
                 idx, target, dictMode);
 
-    while(idx < target)
-        idx += ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict);
+    while(idx < target) {
+        U32 const forward = ZSTD_insertBt1(ms, base+idx, iend, mls, dictMode == ZSTD_extDict);
+        assert(idx < (U32)(idx + forward));
+        idx += forward;
+    }
+    assert((size_t)(ip - base) <= (size_t)(U32)(-1));
+    assert((size_t)(iend - base) <= (size_t)(U32)(-1));
     ms->nextToUpdate = target;
 }
 
@@ -516,11 +541,12 b' void ZSTD_updateTree(ZSTD_matchState_t* '
 
 FORCE_INLINE_TEMPLATE
 U32 ZSTD_insertBtAndGetAllMatches (
+                    ZSTD_match_t* matches,   /* store result (found matches) in this table (presumed large enough) */
                     ZSTD_matchState_t* ms,
+                    U32* nextToUpdate3,
                     const BYTE* const ip, const BYTE* const iLimit, const ZSTD_dictMode_e dictMode,
-                    U32 rep[ZSTD_REP_NUM],
+                    const U32 rep[ZSTD_REP_NUM],
                     U32 const ll0,   /* tells if associated literal length is 0 or not. This value must be 0 or 1 */
-                    ZSTD_match_t* matches,
                     const U32 lengthToBeat,
                     U32 const mls /* template */)
 {
@@ -541,8 +567,8 b' U32 ZSTD_insertBtAndGetAllMatches ('
     U32 const dictLimit = ms->window.dictLimit;
     const BYTE* const dictEnd = dictBase + dictLimit;
     const BYTE* const prefixStart = base + dictLimit;
-    U32 const btLow = btMask >= current ? 0 : current - btMask;
-    U32 const windowLow = ms->window.lowLimit;
+    U32 const btLow = (btMask >= current) ? 0 : current - btMask;
+    U32 const windowLow = ZSTD_getLowestMatchIndex(ms, current, cParams->windowLog);
     U32 const matchLow = windowLow ? windowLow : 1;
     U32* smallerPtr = bt + 2*(current&btMask);
     U32* largerPtr  = bt + 2*(current&btMask) + 1;
@@ -612,7 +638,7 b' U32 ZSTD_insertBtAndGetAllMatches ('
 
     /* HC3 match finder */
     if ((mls == 3) /*static*/ && (bestLength < mls)) {
-        U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, ip);
+        U32 const matchIndex3 = ZSTD_insertAndFindFirstIndexHash3(ms, nextToUpdate3, ip);
         if ((matchIndex3 >= matchLow)
           & (current - matchIndex3 < (1<<18)) /*heuristic : longer distance likely too expensive*/ ) {
             size_t mlen;
@@ -638,9 +664,7 b' U32 ZSTD_insertBtAndGetAllMatches ('
                      (ip+mlen == iLimit) ) {  /* best possible length */
                     ms->nextToUpdate = current+1;  /* skip insertion */
                     return 1;
-                }
-            }
-        }
+        }   }   }
         /* no dictMatchState lookup: dicts don't have a populated HC3 table */
     }
 
@@ -648,19 +672,21 b' U32 ZSTD_insertBtAndGetAllMatches ('
 
     while (nbCompares-- && (matchIndex >= matchLow)) {
         U32* const nextPtr = bt + 2*(matchIndex & btMask);
+        const BYTE* match;
         size_t matchLength = MIN(commonLengthSmaller, commonLengthLarger);   /* guaranteed minimum nb of common bytes */
-        const BYTE* match;
         assert(current > matchIndex);
 
         if ((dictMode == ZSTD_noDict) || (dictMode == ZSTD_dictMatchState) || (matchIndex+matchLength >= dictLimit)) {
             assert(matchIndex+matchLength >= dictLimit);  /* ensure the condition is correct when !extDict */
             match = base + matchIndex;
+            if (matchIndex >= dictLimit) assert(memcmp(match, ip, matchLength) == 0);  /* ensure early section of match is equal as expected */
             matchLength += ZSTD_count(ip+matchLength, match+matchLength, iLimit);
         } else {
             match = dictBase + matchIndex;
+            assert(memcmp(match, ip, matchLength) == 0);  /* ensure early section of match is equal as expected */
             matchLength += ZSTD_count_2segments(ip+matchLength, match+matchLength, iLimit, dictEnd, prefixStart);
             if (matchIndex+matchLength >= dictLimit)
-                match = base + matchIndex;   /* prepare for match[matchLength] */
+                match = base + matchIndex;   /* prepare for match[matchLength] read */
         }
 
         if (matchLength > bestLength) {
@@ -745,10 +771,13 b' U32 ZSTD_insertBtAndGetAllMatches ('
 
 
 FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllMatches (
+                        ZSTD_match_t* matches,   /* store result (match found, increasing size) in this table */
                         ZSTD_matchState_t* ms,
+                        U32* nextToUpdate3,
                         const BYTE* ip, const BYTE* const iHighLimit, const ZSTD_dictMode_e dictMode,
-                        U32 rep[ZSTD_REP_NUM], U32 const ll0,
-                        ZSTD_match_t* matches, U32 const lengthToBeat)
+                        const U32 rep[ZSTD_REP_NUM],
+                        U32 const ll0,
+                        U32 const lengthToBeat)
 {
     const ZSTD_compressionParameters* const cParams = &ms->cParams;
     U32 const matchLengthSearch = cParams->minMatch;
@@ -757,12 +786,12 b' FORCE_INLINE_TEMPLATE U32 ZSTD_BtGetAllM'
     ZSTD_updateTree_internal(ms, ip, iHighLimit, matchLengthSearch, dictMode);
     switch(matchLengthSearch)
     {
-    case 3 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 3);
+    case 3 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 3);
     default :
-    case 4 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 4);
-    case 5 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 5);
+    case 4 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 4);
+    case 5 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 5);
     case 7 :
-    case 6 : return ZSTD_insertBtAndGetAllMatches(ms, ip, iHighLimit, dictMode, rep, ll0, matches, lengthToBeat, 6);
+    case 6 : return ZSTD_insertBtAndGetAllMatches(matches, ms, nextToUpdate3, ip, iHighLimit, dictMode, rep, ll0, lengthToBeat, 6);
     }
 }
 
@@ -838,6 +867,7 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
 
     U32 const sufficient_len = MIN(cParams->targetLength, ZSTD_OPT_NUM -1);
     U32 const minMatch = (cParams->minMatch == 3) ? 3 : 4;
+    U32 nextToUpdate3 = ms->nextToUpdate;
 
     ZSTD_optimal_t* const opt = optStatePtr->priceTable;
     ZSTD_match_t* const matches = optStatePtr->matchTable;
@@ -847,7 +877,6 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
     DEBUGLOG(5, "ZSTD_compressBlock_opt_generic: current=%u, prefix=%u, nextToUpdate=%u",
                 (U32)(ip - base), ms->window.dictLimit, ms->nextToUpdate);
     assert(optLevel <= 2);
-    ms->nextToUpdate3 = ms->nextToUpdate;
     ZSTD_rescaleFreqs(optStatePtr, (const BYTE*)src, srcSize, optLevel);
     ip += (ip==prefixStart);
 
@@ -858,7 +887,7 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
         /* find first match */
         {   U32 const litlen = (U32)(ip - anchor);
             U32 const ll0 = !litlen;
-            U32 const nbMatches = ZSTD_BtGetAllMatches(ms, ip, iend, dictMode, rep, ll0, matches, minMatch);
+            U32 const nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, ip, iend, dictMode, rep, ll0, minMatch);
             if (!nbMatches) { ip++; continue; }
 
             /* initialize opt[0] */
@@ -870,7 +899,7 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
             /* large match -> immediate encoding */
             {   U32 const maxML = matches[nbMatches-1].len;
                 U32 const maxOffset = matches[nbMatches-1].off;
-                DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new serie",
+                DEBUGLOG(6, "found %u matches of maxLength=%u and maxOffCode=%u at cPos=%u => start new series",
                             nbMatches, maxML, maxOffset, (U32)(ip-prefixStart));
 
                 if (maxML > sufficient_len) {
@@ -955,7 +984,7 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
                 U32 const litlen = (opt[cur].mlen == 0) ? opt[cur].litlen : 0;
                 U32 const previousPrice = opt[cur].price;
                 U32 const basePrice = previousPrice + ZSTD_litLengthPrice(0, optStatePtr, optLevel);
-                U32 const nbMatches = ZSTD_BtGetAllMatches(ms, inr, iend, dictMode, opt[cur].rep, ll0, matches, minMatch);
+                U32 const nbMatches = ZSTD_BtGetAllMatches(matches, ms, &nextToUpdate3, inr, iend, dictMode, opt[cur].rep, ll0, minMatch);
                 U32 matchNb;
                 if (!nbMatches) {
                     DEBUGLOG(7, "rPos:%u : no match found", cur);
@@ -1079,7 +1108,7 b' ZSTD_compressBlock_opt_generic(ZSTD_matc'
     }   /* while (ip < ilimit) */
 
     /* Return the last literals size */
-    return iend - anchor;
+    return (size_t)(iend - anchor);
 }
 
 
@@ -1108,7 +1137,8 b' static U32 ZSTD_upscaleStat(unsigned* ta'
 /* used in 2-pass strategy */
 MEM_STATIC void ZSTD_upscaleStats(optState_t* optPtr)
 {
-    optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0);
+    if (ZSTD_compressedLiterals(optPtr))
+        optPtr->litSum = ZSTD_upscaleStat(optPtr->litFreq, MaxLit, 0);
     optPtr->litLengthSum = ZSTD_upscaleStat(optPtr->litLengthFreq, MaxLL, 0);
     optPtr->matchLengthSum = ZSTD_upscaleStat(optPtr->matchLengthFreq, MaxML, 0);
     optPtr->offCodeSum = ZSTD_upscaleStat(optPtr->offCodeFreq, MaxOff, 0);
@@ -1117,7 +1147,7 b' MEM_STATIC void ZSTD_upscaleStats(optSta'
 /* ZSTD_initStats_ultra():
  * make a first compression pass, just to seed stats with more accurate starting values.
  * only works on first block, with no dictionary and no ldm.
- * this function cannot error, hence its constract must be respected.
+ * this function cannot error, hence its contract must be respected.
  */
 static void
 ZSTD_initStats_ultra(ZSTD_matchState_t* ms,
@@ -1142,7 +1172,6 b' ZSTD_initStats_ultra(ZSTD_matchState_t* '
     ms->window.dictLimit += (U32)srcSize;
     ms->window.lowLimit = ms->window.dictLimit;
     ms->nextToUpdate = ms->window.dictLimit;
-    ms->nextToUpdate3 = ms->window.dictLimit;
 
     /* re-inforce weight of collected statistics */
     ZSTD_upscaleStats(&ms->opt);

contrib/python-zstandard/zstd/compress/zstdmt_compress.c

@@ -22,6 +22,7 b''
 /* ======   Dependencies   ====== */
 #include <string.h>      /* memcpy, memset */
 #include <limits.h>      /* INT_MAX, UINT_MAX */
+#include "mem.h"         /* MEM_STATIC */
 #include "pool.h"        /* threadpool */
 #include "threading.h"   /* mutex */
 #include "zstd_compress_internal.h"  /* MIN, ERROR, ZSTD_*, ZSTD_highbit32 */
@@ -456,7 +457,7 b' typedef struct {'
      * Must be acquired after the main mutex when acquiring both.
      */
     ZSTD_pthread_mutex_t ldmWindowMutex;
-    ZSTD_pthread_cond_t ldmWindowCond;  /* Signaled when ldmWindow is udpated */
+    ZSTD_pthread_cond_t ldmWindowCond;  /* Signaled when ldmWindow is updated */
     ZSTD_window_t ldmWindow;  /* A thread-safe copy of ldmState.window */
 } serialState_t;
 
@@ -647,7 +648,7 b' static void ZSTDMT_compressionJob(void* '
     buffer_t dstBuff = job->dstBuff;
     size_t lastCBlockSize = 0;
 
-    /* ressources */
+    /* resources */
     if (cctx==NULL) JOB_ERROR(ERROR(memory_allocation));
     if (dstBuff.start == NULL) {   /* streaming job : doesn't provide a dstBuffer */
         dstBuff = ZSTDMT_getBuffer(job->bufPool);
@@ -672,7 +673,7 b' static void ZSTDMT_compressionJob(void* '
         if (ZSTD_isError(initError)) JOB_ERROR(initError);
     } else {  /* srcStart points at reloaded section */
         U64 const pledgedSrcSize = job->firstJob ? job->fullFrameSize : job->src.size;
-        {   size_t const forceWindowError = ZSTD_CCtxParam_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob);
+        {   size_t const forceWindowError = ZSTD_CCtxParams_setParameter(&jobParams, ZSTD_c_forceMaxWindow, !job->firstJob);
             if (ZSTD_isError(forceWindowError)) JOB_ERROR(forceWindowError);
         }
         {   size_t const initError = ZSTD_compressBegin_advanced_internal(cctx,
@@ -864,14 +865,10 b' static size_t ZSTDMT_expandJobsTable (ZS'
  * Internal use only */
 size_t ZSTDMT_CCtxParam_setNbWorkers(ZSTD_CCtx_params* params, unsigned nbWorkers)
 {
-    if (nbWorkers > ZSTDMT_NBWORKERS_MAX) nbWorkers = ZSTDMT_NBWORKERS_MAX;
-    params->nbWorkers = nbWorkers;
-    params->overlapLog = ZSTDMT_OVERLAPLOG_DEFAULT;
-    params->jobSize = 0;
-    return nbWorkers;
+    return ZSTD_CCtxParams_setParameter(params, ZSTD_c_nbWorkers, (int)nbWorkers);
 }
 
-ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers, ZSTD_customMem cMem)
+MEM_STATIC ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced_internal(unsigned nbWorkers, ZSTD_customMem cMem)
 {
     ZSTDMT_CCtx* mtctx;
     U32 nbJobs = nbWorkers + 2;
@@ -906,6 +903,17 b' ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced('
     return mtctx;
 }
 
+ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers, ZSTD_customMem cMem)
+{
+#ifdef ZSTD_MULTITHREAD
+    return ZSTDMT_createCCtx_advanced_internal(nbWorkers, cMem);
+#else
+    (void)nbWorkers;
+    (void)cMem;
+    return NULL;
+#endif
+}
+
 ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers)
 {
     return ZSTDMT_createCCtx_advanced(nbWorkers, ZSTD_defaultCMem);
@@ -986,26 +994,13 b' ZSTDMT_CCtxParam_setMTCtxParameter(ZSTD_'
     {
     case ZSTDMT_p_jobSize :
         DEBUGLOG(4, "ZSTDMT_CCtxParam_setMTCtxParameter : set jobSize to %i", value);
-        if ( value != 0  /* default */
-          && value < ZSTDMT_JOBSIZE_MIN)
-            value = ZSTDMT_JOBSIZE_MIN;
-        assert(value >= 0);
-        if (value > ZSTDMT_JOBSIZE_MAX) value = ZSTDMT_JOBSIZE_MAX;
-        params->jobSize = value;
-        return value;
-
+        return ZSTD_CCtxParams_setParameter(params, ZSTD_c_jobSize, value);
     case ZSTDMT_p_overlapLog :
         DEBUGLOG(4, "ZSTDMT_p_overlapLog : %i", value);
-        if (value < ZSTD_OVERLAPLOG_MIN) value = ZSTD_OVERLAPLOG_MIN;
-        if (value > ZSTD_OVERLAPLOG_MAX) value = ZSTD_OVERLAPLOG_MAX;
-        params->overlapLog = value;
-        return value;
-
+        return ZSTD_CCtxParams_setParameter(params, ZSTD_c_overlapLog, value);
     case ZSTDMT_p_rsyncable :
-        value = (value != 0);
-        params->rsyncable = value;
-        return value;
-
+        DEBUGLOG(4, "ZSTD_p_rsyncable : %i", value);
+        return ZSTD_CCtxParams_setParameter(params, ZSTD_c_rsyncable, value);
     default :
         return ERROR(parameter_unsupported);
     }
@@ -1021,32 +1016,29 b' size_t ZSTDMT_getMTCtxParameter(ZSTDMT_C'
 {
     switch (parameter) {
     case ZSTDMT_p_jobSize:
-        assert(mtctx->params.jobSize <= INT_MAX);
-        *value = (int)(mtctx->params.jobSize);
-        break;
+        return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_jobSize, value);
     case ZSTDMT_p_overlapLog:
-        *value = mtctx->params.overlapLog;
-        break;
+        return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_overlapLog, value);
     case ZSTDMT_p_rsyncable:
-        *value = mtctx->params.rsyncable;
-        break;
+        return ZSTD_CCtxParams_getParameter(&mtctx->params, ZSTD_c_rsyncable, value);
     default:
         return ERROR(parameter_unsupported);
     }
-    return 0;
 }
 
 /* Sets parameters relevant to the compression job,
  * initializing others to default values. */
 static ZSTD_CCtx_params ZSTDMT_initJobCCtxParams(ZSTD_CCtx_params const params)
 {
-    ZSTD_CCtx_params jobParams;
-    memset(&jobParams, 0, sizeof(jobParams));
-
-    jobParams.cParams = params.cParams;
-    jobParams.fParams = params.fParams;
-    jobParams.compressionLevel = params.compressionLevel;
-
+    ZSTD_CCtx_params jobParams = params;
+    /* Clear parameters related to multithreading */
+    jobParams.forceWindow = 0;
+    jobParams.nbWorkers = 0;
+    jobParams.jobSize = 0;
+    jobParams.overlapLog = 0;
+    jobParams.rsyncable = 0;
+    memset(&jobParams.ldmParams, 0, sizeof(ldmParams_t));
+    memset(&jobParams.customMem, 0, sizeof(ZSTD_customMem));
     return jobParams;
 }
 
@@ -1056,7 +1048,7 b' static ZSTD_CCtx_params ZSTDMT_initJobCC'
 static size_t ZSTDMT_resize(ZSTDMT_CCtx* mtctx, unsigned nbWorkers)
 {
     if (POOL_resize(mtctx->factory, nbWorkers)) return ERROR(memory_allocation);
-    CHECK_F( ZSTDMT_expandJobsTable(mtctx, nbWorkers) );
+    FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbWorkers) );
     mtctx->bufPool = ZSTDMT_expandBufferPool(mtctx->bufPool, nbWorkers);
     if (mtctx->bufPool == NULL) return ERROR(memory_allocation);
     mtctx->cctxPool = ZSTDMT_expandCCtxPool(mtctx->cctxPool, nbWorkers);
@@ -1137,9 +1129,14 b' size_t ZSTDMT_toFlushNow(ZSTDMT_CCtx* mt'
             size_t const produced = ZSTD_isError(cResult) ? 0 : cResult;
             size_t const flushed = ZSTD_isError(cResult) ? 0 : jobPtr->dstFlushed;
             assert(flushed <= produced);
+            assert(jobPtr->consumed <= jobPtr->src.size);
             toFlush = produced - flushed;
-            if (toFlush==0 && (jobPtr->consumed >= jobPtr->src.size)) {
-                /* doneJobID is not-fully-flushed, but toFlush==0 : doneJobID should be compressing some more data */
+            /* if toFlush==0, nothing is available to flush.
+             * However, jobID is expected to still be active:
+             * if jobID was already completed and fully flushed,
+             * ZSTDMT_flushProduced() should have already moved onto next job.
+             * Therefore, some input has not yet been consumed. */
+            if (toFlush==0) {
                 assert(jobPtr->consumed < jobPtr->src.size);
             }
         }
@@ -1156,12 +1153,16 b' size_t ZSTDMT_toFlushNow(ZSTDMT_CCtx* mt'
 
 static unsigned ZSTDMT_computeTargetJobLog(ZSTD_CCtx_params const params)
 {
-    if (params.ldmParams.enableLdm)
+    unsigned jobLog;
+    if (params.ldmParams.enableLdm) {
         /* In Long Range Mode, the windowLog is typically oversized.
          * In which case, it's preferable to determine the jobSize
          * based on chainLog instead. */
-        return MAX(21, params.cParams.chainLog + 4);
-    return MAX(20, params.cParams.windowLog + 2);
+        jobLog = MAX(21, params.cParams.chainLog + 4);
+    } else {
+        jobLog = MAX(20, params.cParams.windowLog + 2);
+    }
+    return MIN(jobLog, (unsigned)ZSTDMT_JOBLOG_MAX);
 }
 
 static int ZSTDMT_overlapLog_default(ZSTD_strategy strat)
@@ -1205,7 +1206,7 b' static size_t ZSTDMT_computeOverlapSize('
         ovLog = MIN(params.cParams.windowLog, ZSTDMT_computeTargetJobLog(params) - 2)
                 - overlapRLog;
     }
-    assert(0 <= ovLog && ovLog <= 30);
+    assert(0 <= ovLog && ovLog <= ZSTD_WINDOWLOG_MAX);
     DEBUGLOG(4, "overlapLog : %i", params.overlapLog);
     DEBUGLOG(4, "overlap size : %i", 1 << ovLog);
     return (ovLog==0) ? 0 : (size_t)1 << ovLog;
@@ -1263,7 +1264,7 b' static size_t ZSTDMT_compress_advanced_i'
     if (ZSTDMT_serialState_reset(&mtctx->serial, mtctx->seqPool, params, avgJobSize))
         return ERROR(memory_allocation);
 
-    CHECK_F( ZSTDMT_expandJobsTable(mtctx, nbJobs) );  /* only expands if necessary */
+    FORWARD_IF_ERROR( ZSTDMT_expandJobsTable(mtctx, nbJobs) );  /* only expands if necessary */
 
     {   unsigned u;
         for (u=0; u<nbJobs; u++) {
@@ -1396,10 +1397,10 b' size_t ZSTDMT_initCStream_internal('
 
     /* init */
     if (params.nbWorkers != mtctx->params.nbWorkers)
-        CHECK_F( ZSTDMT_resize(mtctx, params.nbWorkers) );
+        FORWARD_IF_ERROR( ZSTDMT_resize(mtctx, params.nbWorkers) );
 
     if (params.jobSize != 0 && params.jobSize < ZSTDMT_JOBSIZE_MIN) params.jobSize = ZSTDMT_JOBSIZE_MIN;
-    if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = ZSTDMT_JOBSIZE_MAX;
+    if (params.jobSize > (size_t)ZSTDMT_JOBSIZE_MAX) params.jobSize = (size_t)ZSTDMT_JOBSIZE_MAX;
 
     mtctx->singleBlockingThread = (pledgedSrcSize <= ZSTDMT_JOBSIZE_MIN);  /* do not trigger multi-threading when srcSize is too small */
     if (mtctx->singleBlockingThread) {
@@ -1440,6 +1441,8 b' size_t ZSTDMT_initCStream_internal('
     if (mtctx->targetSectionSize == 0) {
         mtctx->targetSectionSize = 1ULL << ZSTDMT_computeTargetJobLog(params);
     }
+    assert(mtctx->targetSectionSize <= (size_t)ZSTDMT_JOBSIZE_MAX);
+
     if (params.rsyncable) {
         /* Aim for the targetsectionSize as the average job size. */
         U32 const jobSizeMB = (U32)(mtctx->targetSectionSize >> 20);
@@ -1547,7 +1550,7 b' size_t ZSTDMT_initCStream(ZSTDMT_CCtx* m'
 /* ZSTDMT_writeLastEmptyBlock()
  * Write a single empty block with an end-of-frame to finish a frame.
  * Job must be created from streaming variant.
- * This function is always successfull if expected conditions are fulfilled.
+ * This function is always successful if expected conditions are fulfilled.
  */
 static void ZSTDMT_writeLastEmptyBlock(ZSTDMT_jobDescription* job)
 {
@@ -1987,7 +1990,7 b' size_t ZSTDMT_compressStream_generic(ZST'
     assert(input->pos  <= input->size);
 
     if (mtctx->singleBlockingThread) {  /* delegate to single-thread (synchronous) */
-        return ZSTD_compressStream_generic(mtctx->cctxPool->cctx[0], output, input, endOp);
+        return ZSTD_compressStream2(mtctx->cctxPool->cctx[0], output, input, endOp);
     }
 
     if ((mtctx->frameEnded) && (endOp==ZSTD_e_continue)) {
@@ -2051,7 +2054,7 b' size_t ZSTDMT_compressStream_generic(ZST'
       || ((endOp == ZSTD_e_end) && (!mtctx->frameEnded)) ) {   /* must finish the frame with a zero-size block */
         size_t const jobSize = mtctx->inBuff.filled;
         assert(mtctx->inBuff.filled <= mtctx->targetSectionSize);
-        CHECK_F( ZSTDMT_createCompressionJob(mtctx, jobSize, endOp) );
+        FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, jobSize, endOp) );
     }
 
     /* check for potential compressed data ready to be flushed */
@@ -2065,7 +2068,7 b' size_t ZSTDMT_compressStream_generic(ZST'
 
 size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input)
 {
-    CHECK_F( ZSTDMT_compressStream_generic(mtctx, output, input, ZSTD_e_continue) );
+    FORWARD_IF_ERROR( ZSTDMT_compressStream_generic(mtctx, output, input, ZSTD_e_continue) );
 
     /* recommended next input size : fill current input buffer */
     return mtctx->targetSectionSize - mtctx->inBuff.filled;   /* note : could be zero when input buffer is fully filled and no more availability to create new job */
@@ -2082,7 +2085,7 b' static size_t ZSTDMT_flushStream_interna'
       || ((endFrame==ZSTD_e_end) && !mtctx->frameEnded)) {  /* need a last 0-size block to end frame */
            DEBUGLOG(5, "ZSTDMT_flushStream_internal : create a new job (%u bytes, end:%u)",
                         (U32)srcSize, (U32)endFrame);
-        CHECK_F( ZSTDMT_createCompressionJob(mtctx, srcSize, endFrame) );
+        FORWARD_IF_ERROR( ZSTDMT_createCompressionJob(mtctx, srcSize, endFrame) );
     }
 
     /* check if there is any data available to flush */

contrib/python-zstandard/zstd/compress/zstdmt_compress.h

@@ -17,10 +17,25 b''
 
 
 /* Note : This is an internal API.
- *        Some methods are still exposed (ZSTDLIB_API),
+ *        These APIs used to be exposed with ZSTDLIB_API,
  *        because it used to be the only way to invoke MT compression.
- *        Now, it's recommended to use ZSTD_compress_generic() instead.
- *        These methods will stop being exposed in a future version */
+ *        Now, it's recommended to use ZSTD_compress2 and ZSTD_compressStream2()
+ *        instead.
+ *
+ *        If you depend on these APIs and can't switch, then define
+ *        ZSTD_LEGACY_MULTITHREADED_API when making the dynamic library.
+ *        However, we may completely remove these functions in a future
+ *        release, so please switch soon.
+ *
+ *        This API requires ZSTD_MULTITHREAD to be defined during compilation,
+ *        otherwise ZSTDMT_createCCtx*() will fail.
+ */
+
+#ifdef ZSTD_LEGACY_MULTITHREADED_API
+#  define ZSTDMT_API ZSTDLIB_API
+#else
+#  define ZSTDMT_API
+#endif
 
 /* ===   Dependencies   === */
 #include <stddef.h>                /* size_t */
@@ -35,22 +50,25 b''
 #ifndef ZSTDMT_JOBSIZE_MIN
 #  define ZSTDMT_JOBSIZE_MIN (1 MB)
 #endif
+#define ZSTDMT_JOBLOG_MAX   (MEM_32bits() ? 29 : 30)
 #define ZSTDMT_JOBSIZE_MAX  (MEM_32bits() ? (512 MB) : (1024 MB))
 
 
 /* ===   Memory management   === */
 typedef struct ZSTDMT_CCtx_s ZSTDMT_CCtx;
-ZSTDLIB_API ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers);
-ZSTDLIB_API ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers,
+/* Requires ZSTD_MULTITHREAD to be defined during compilation, otherwise it will return NULL. */
+ZSTDMT_API ZSTDMT_CCtx* ZSTDMT_createCCtx(unsigned nbWorkers);
+/* Requires ZSTD_MULTITHREAD to be defined during compilation, otherwise it will return NULL. */
+ZSTDMT_API ZSTDMT_CCtx* ZSTDMT_createCCtx_advanced(unsigned nbWorkers,
                                                     ZSTD_customMem cMem);
-ZSTDLIB_API size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx);
+ZSTDMT_API size_t ZSTDMT_freeCCtx(ZSTDMT_CCtx* mtctx);
 
-ZSTDLIB_API size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx);
+ZSTDMT_API size_t ZSTDMT_sizeof_CCtx(ZSTDMT_CCtx* mtctx);
 
 
 /* ===   Simple one-pass compression function   === */
 
-ZSTDLIB_API size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,
+ZSTDMT_API size_t ZSTDMT_compressCCtx(ZSTDMT_CCtx* mtctx,
                                        void* dst, size_t dstCapacity,
                                  const void* src, size_t srcSize,
                                        int compressionLevel);
@@ -59,31 +77,31 b' ZSTDLIB_API size_t ZSTDMT_compressCCtx(Z'
 
 /* ===   Streaming functions   === */
 
-ZSTDLIB_API size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel);
-ZSTDLIB_API size_t ZSTDMT_resetCStream(ZSTDMT_CCtx* mtctx, unsigned long long pledgedSrcSize);  /**< if srcSize is not known at reset time, use ZSTD_CONTENTSIZE_UNKNOWN. Note: for compatibility with older programs, 0 means the same as ZSTD_CONTENTSIZE_UNKNOWN, but it will change in the future to mean "empty" */
+ZSTDMT_API size_t ZSTDMT_initCStream(ZSTDMT_CCtx* mtctx, int compressionLevel);
+ZSTDMT_API size_t ZSTDMT_resetCStream(ZSTDMT_CCtx* mtctx, unsigned long long pledgedSrcSize);  /**< if srcSize is not known at reset time, use ZSTD_CONTENTSIZE_UNKNOWN. Note: for compatibility with older programs, 0 means the same as ZSTD_CONTENTSIZE_UNKNOWN, but it will change in the future to mean "empty" */
 
-ZSTDLIB_API size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx);
-ZSTDLIB_API size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
+ZSTDMT_API size_t ZSTDMT_nextInputSizeHint(const ZSTDMT_CCtx* mtctx);
+ZSTDMT_API size_t ZSTDMT_compressStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
 
-ZSTDLIB_API size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output);   /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */
-ZSTDLIB_API size_t ZSTDMT_endStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output);     /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */
+ZSTDMT_API size_t ZSTDMT_flushStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output);   /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */
+ZSTDMT_API size_t ZSTDMT_endStream(ZSTDMT_CCtx* mtctx, ZSTD_outBuffer* output);     /**< @return : 0 == all flushed; >0 : still some data to be flushed; or an error code (ZSTD_isError()) */
 
 
 /* ===   Advanced functions and parameters  === */
 
-ZSTDLIB_API size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
-                                           void* dst, size_t dstCapacity,
-                                     const void* src, size_t srcSize,
-                                     const ZSTD_CDict* cdict,
-                                           ZSTD_parameters params,
-                                           int overlapLog);
+ZSTDMT_API size_t ZSTDMT_compress_advanced(ZSTDMT_CCtx* mtctx,
+                                          void* dst, size_t dstCapacity,
+                                    const void* src, size_t srcSize,
+                                    const ZSTD_CDict* cdict,
+                                          ZSTD_parameters params,
+                                          int overlapLog);
 
-ZSTDLIB_API size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx,
+ZSTDMT_API size_t ZSTDMT_initCStream_advanced(ZSTDMT_CCtx* mtctx,
                                         const void* dict, size_t dictSize,   /* dict can be released after init, a local copy is preserved within zcs */
                                         ZSTD_parameters params,
                                         unsigned long long pledgedSrcSize);  /* pledgedSrcSize is optional and can be zero == unknown */
 
-ZSTDLIB_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
+ZSTDMT_API size_t ZSTDMT_initCStream_usingCDict(ZSTDMT_CCtx* mtctx,
                                         const ZSTD_CDict* cdict,
                                         ZSTD_frameParameters fparams,
                                         unsigned long long pledgedSrcSize);  /* note : zero means empty */
@@ -92,7 +110,7 b' ZSTDLIB_API size_t ZSTDMT_initCStream_us'
  * List of parameters that can be set using ZSTDMT_setMTCtxParameter() */
 typedef enum {
     ZSTDMT_p_jobSize,     /* Each job is compressed in parallel. By default, this value is dynamically determined depending on compression parameters. Can be set explicitly here. */
-    ZSTDMT_p_overlapLog,  /* Each job may reload a part of previous job to enhance compressionr ratio; 0 == no overlap, 6(default) == use 1/8th of window, >=9 == use full window. This is a "sticky" parameter : its value will be re-used on next compression job */
+    ZSTDMT_p_overlapLog,  /* Each job may reload a part of previous job to enhance compression ratio; 0 == no overlap, 6(default) == use 1/8th of window, >=9 == use full window. This is a "sticky" parameter : its value will be re-used on next compression job */
     ZSTDMT_p_rsyncable    /* Enables rsyncable mode. */
 } ZSTDMT_parameter;
 
@@ -101,12 +119,12 b' typedef enum {'
  * The function must be called typically after ZSTD_createCCtx() but __before ZSTDMT_init*() !__
  * Parameters not explicitly reset by ZSTDMT_init*() remain the same in consecutive compression sessions.
  * @return : 0, or an error code (which can be tested using ZSTD_isError()) */
-ZSTDLIB_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value);
+ZSTDMT_API size_t ZSTDMT_setMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int value);
 
 /* ZSTDMT_getMTCtxParameter() :
  * Query the ZSTDMT_CCtx for a parameter value.
  * @return : 0, or an error code (which can be tested using ZSTD_isError()) */
-ZSTDLIB_API size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value);
+ZSTDMT_API size_t ZSTDMT_getMTCtxParameter(ZSTDMT_CCtx* mtctx, ZSTDMT_parameter parameter, int* value);
 
 
 /*! ZSTDMT_compressStream_generic() :
@@ -116,7 +134,7 b' ZSTDLIB_API size_t ZSTDMT_getMTCtxParame'
  *           0 if fully flushed
  *           or an error code
  *  note : needs to be init using any ZSTD_initCStream*() variant */
-ZSTDLIB_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
+ZSTDMT_API size_t ZSTDMT_compressStream_generic(ZSTDMT_CCtx* mtctx,
                                                 ZSTD_outBuffer* output,
                                                 ZSTD_inBuffer* input,
                                                 ZSTD_EndDirective endOp);

contrib/python-zstandard/zstd/decompress/zstd_ddict.c

@@ -105,9 +105,9 b' ZSTD_loadEntropy_intoDDict(ZSTD_DDict* d'
     ddict->dictID = MEM_readLE32((const char*)ddict->dictContent + ZSTD_FRAMEIDSIZE);
 
     /* load entropy tables */
-    CHECK_E( ZSTD_loadDEntropy(&ddict->entropy,
-                                ddict->dictContent, ddict->dictSize),
-             dictionary_corrupted );
+    RETURN_ERROR_IF(ZSTD_isError(ZSTD_loadDEntropy(
+            &ddict->entropy, ddict->dictContent, ddict->dictSize)),
+        dictionary_corrupted);
     ddict->entropyPresent = 1;
     return 0;
 }
@@ -133,7 +133,7 b' static size_t ZSTD_initDDict_internal(ZS'
     ddict->entropy.hufTable[0] = (HUF_DTable)((HufLog)*0x1000001);  /* cover both little and big endian */
 
     /* parse dictionary content */
-    CHECK_F( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) );
+    FORWARD_IF_ERROR( ZSTD_loadEntropy_intoDDict(ddict, dictContentType) );
 
     return 0;
 }

contrib/python-zstandard/zstd/decompress/zstd_decompress.c

@@ -106,6 +106,7 b' static void ZSTD_initDCtx_internal(ZSTD_'
     dctx->ddictLocal  = NULL;
     dctx->dictEnd     = NULL;
     dctx->ddictIsCold = 0;
+    dctx->dictUses = ZSTD_dont_use;
     dctx->inBuff      = NULL;
     dctx->inBuffSize  = 0;
     dctx->outBuffSize = 0;
@@ -147,13 +148,20 b' ZSTD_DCtx* ZSTD_createDCtx(void)'
     return ZSTD_createDCtx_advanced(ZSTD_defaultCMem);
 }
 
+static void ZSTD_clearDict(ZSTD_DCtx* dctx)
+{
+    ZSTD_freeDDict(dctx->ddictLocal);
+    dctx->ddictLocal = NULL;
+    dctx->ddict = NULL;
+    dctx->dictUses = ZSTD_dont_use;
+}
+
 size_t ZSTD_freeDCtx(ZSTD_DCtx* dctx)
 {
     if (dctx==NULL) return 0;   /* support free on NULL */
-    if (dctx->staticSize) return ERROR(memory_allocation);   /* not compatible with static DCtx */
+    RETURN_ERROR_IF(dctx->staticSize, memory_allocation, "not compatible with static DCtx");
     {   ZSTD_customMem const cMem = dctx->customMem;
-        ZSTD_freeDDict(dctx->ddictLocal);
-        dctx->ddictLocal = NULL;
+        ZSTD_clearDict(dctx);
         ZSTD_free(dctx->inBuff, cMem);
         dctx->inBuff = NULL;
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
@@ -203,7 +211,7 b' unsigned ZSTD_isFrame(const void* buffer'
 static size_t ZSTD_frameHeaderSize_internal(const void* src, size_t srcSize, ZSTD_format_e format)
 {
     size_t const minInputSize = ZSTD_startingInputLength(format);
-    if (srcSize < minInputSize) return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(srcSize < minInputSize, srcSize_wrong);
 
     {   BYTE const fhd = ((const BYTE*)src)[minInputSize-1];
         U32 const dictID= fhd & 3;
@@ -238,7 +246,7 b' size_t ZSTD_getFrameHeader_advanced(ZSTD'
 
     memset(zfhPtr, 0, sizeof(*zfhPtr));   /* not strictly necessary, but static analyzer do not understand that zfhPtr is only going to be read only if return value is zero, since they are 2 different signals */
     if (srcSize < minInputSize) return minInputSize;
-    if (src==NULL) return ERROR(GENERIC);   /* invalid parameter */
+    RETURN_ERROR_IF(src==NULL, GENERIC, "invalid parameter");
 
     if ( (format != ZSTD_f_zstd1_magicless)
       && (MEM_readLE32(src) != ZSTD_MAGICNUMBER) ) {
@@ -251,7 +259,7 b' size_t ZSTD_getFrameHeader_advanced(ZSTD'
             zfhPtr->frameType = ZSTD_skippableFrame;
             return 0;
         }
-        return ERROR(prefix_unknown);
+        RETURN_ERROR(prefix_unknown);
     }
 
     /* ensure there is enough `srcSize` to fully read/decode frame header */
@@ -269,14 +277,13 b' size_t ZSTD_getFrameHeader_advanced(ZSTD'
         U64 windowSize = 0;
         U32 dictID = 0;
         U64 frameContentSize = ZSTD_CONTENTSIZE_UNKNOWN;
-        if ((fhdByte & 0x08) != 0)
-            return ERROR(frameParameter_unsupported); /* reserved bits, must be zero */
+        RETURN_ERROR_IF((fhdByte & 0x08) != 0, frameParameter_unsupported,
+                        "reserved bits, must be zero");
 
         if (!singleSegment) {
             BYTE const wlByte = ip[pos++];
             U32 const windowLog = (wlByte >> 3) + ZSTD_WINDOWLOG_ABSOLUTEMIN;
-            if (windowLog > ZSTD_WINDOWLOG_MAX)
-                return ERROR(frameParameter_windowTooLarge);
+            RETURN_ERROR_IF(windowLog > ZSTD_WINDOWLOG_MAX, frameParameter_windowTooLarge);
             windowSize = (1ULL << windowLog);
             windowSize += (windowSize >> 3) * (wlByte&7);
         }
@@ -348,14 +355,16 b' static size_t readSkippableFrameSize(voi'
     size_t const skippableHeaderSize = ZSTD_SKIPPABLEHEADERSIZE;
     U32 sizeU32;
 
-    if (srcSize < ZSTD_SKIPPABLEHEADERSIZE)
-        return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(srcSize < ZSTD_SKIPPABLEHEADERSIZE, srcSize_wrong);
 
     sizeU32 = MEM_readLE32((BYTE const*)src + ZSTD_FRAMEIDSIZE);
-    if ((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32)
-        return ERROR(frameParameter_unsupported);
-
-    return skippableHeaderSize + sizeU32;
+    RETURN_ERROR_IF((U32)(sizeU32 + ZSTD_SKIPPABLEHEADERSIZE) < sizeU32,
+                    frameParameter_unsupported);
+    {
+        size_t const skippableSize = skippableHeaderSize + sizeU32;
+        RETURN_ERROR_IF(skippableSize > srcSize, srcSize_wrong);
+        return skippableSize;
+    }
 }
 
 /** ZSTD_findDecompressedSize() :
@@ -372,11 +381,10 b' unsigned long long ZSTD_findDecompressed'
 
         if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
             size_t const skippableSize = readSkippableFrameSize(src, srcSize);
-            if (ZSTD_isError(skippableSize))
-                return skippableSize;
-            if (srcSize < skippableSize) {
+            if (ZSTD_isError(skippableSize)) {
                 return ZSTD_CONTENTSIZE_ERROR;
             }
+            assert(skippableSize <= srcSize);
 
             src = (const BYTE *)src + skippableSize;
             srcSize -= skippableSize;
@@ -428,13 +436,91 b' static size_t ZSTD_decodeFrameHeader(ZST'
 {
     size_t const result = ZSTD_getFrameHeader_advanced(&(dctx->fParams), src, headerSize, dctx->format);
     if (ZSTD_isError(result)) return result;    /* invalid header */
-    if (result>0) return ERROR(srcSize_wrong);  /* headerSize too small */
-    if (dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID))
-        return ERROR(dictionary_wrong);
+    RETURN_ERROR_IF(result>0, srcSize_wrong, "headerSize too small");
+#ifndef FUZZING_BUILD_MODE_UNSAFE_FOR_PRODUCTION
+    /* Skip the dictID check in fuzzing mode, because it makes the search
+     * harder.
+     */
+    RETURN_ERROR_IF(dctx->fParams.dictID && (dctx->dictID != dctx->fParams.dictID),
+                    dictionary_wrong);
+#endif
     if (dctx->fParams.checksumFlag) XXH64_reset(&dctx->xxhState, 0);
     return 0;
 }
 
+static ZSTD_frameSizeInfo ZSTD_errorFrameSizeInfo(size_t ret)
+{
+    ZSTD_frameSizeInfo frameSizeInfo;
+    frameSizeInfo.compressedSize = ret;
+    frameSizeInfo.decompressedBound = ZSTD_CONTENTSIZE_ERROR;
+    return frameSizeInfo;
+}
+
+static ZSTD_frameSizeInfo ZSTD_findFrameSizeInfo(const void* src, size_t srcSize)
+{
+    ZSTD_frameSizeInfo frameSizeInfo;
+    memset(&frameSizeInfo, 0, sizeof(ZSTD_frameSizeInfo));
+
+#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
+    if (ZSTD_isLegacy(src, srcSize))
+        return ZSTD_findFrameSizeInfoLegacy(src, srcSize);
+#endif
+
+    if ((srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
+        && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
+        frameSizeInfo.compressedSize = readSkippableFrameSize(src, srcSize);
+        assert(ZSTD_isError(frameSizeInfo.compressedSize) ||
+               frameSizeInfo.compressedSize <= srcSize);
+        return frameSizeInfo;
+    } else {
+        const BYTE* ip = (const BYTE*)src;
+        const BYTE* const ipstart = ip;
+        size_t remainingSize = srcSize;
+        size_t nbBlocks = 0;
+        ZSTD_frameHeader zfh;
+
+        /* Extract Frame Header */
+        {   size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
+            if (ZSTD_isError(ret))
+                return ZSTD_errorFrameSizeInfo(ret);
+            if (ret > 0)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+        }
+
+        ip += zfh.headerSize;
+        remainingSize -= zfh.headerSize;
+
+        /* Iterate over each block */
+        while (1) {
+            blockProperties_t blockProperties;
+            size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
+            if (ZSTD_isError(cBlockSize))
+                return ZSTD_errorFrameSizeInfo(cBlockSize);
+
+            if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+
+            ip += ZSTD_blockHeaderSize + cBlockSize;
+            remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
+            nbBlocks++;
+
+            if (blockProperties.lastBlock) break;
+        }
+
+        /* Final frame content checksum */
+        if (zfh.checksumFlag) {
+            if (remainingSize < 4)
+                return ZSTD_errorFrameSizeInfo(ERROR(srcSize_wrong));
+            ip += 4;
+        }
+
+        frameSizeInfo.compressedSize = ip - ipstart;
+        frameSizeInfo.decompressedBound = (zfh.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN)
+                                        ? zfh.frameContentSize
+                                        : nbBlocks * zfh.blockSizeMax;
+        return frameSizeInfo;
+    }
+}
 
 /** ZSTD_findFrameCompressedSize() :
  *  compatible with legacy mode
@@ -443,52 +529,33 b' static size_t ZSTD_decodeFrameHeader(ZST'
  *  @return : the compressed size of the frame starting at `src` */
 size_t ZSTD_findFrameCompressedSize(const void *src, size_t srcSize)
 {
-#if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT >= 1)
-    if (ZSTD_isLegacy(src, srcSize))
-        return ZSTD_findFrameCompressedSizeLegacy(src, srcSize);
-#endif
-    if ( (srcSize >= ZSTD_SKIPPABLEHEADERSIZE)
-      && (MEM_readLE32(src) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START ) {
-        return readSkippableFrameSize(src, srcSize);
-    } else {
-        const BYTE* ip = (const BYTE*)src;
-        const BYTE* const ipstart = ip;
-        size_t remainingSize = srcSize;
-        ZSTD_frameHeader zfh;
-
-        /* Extract Frame Header */
-        {   size_t const ret = ZSTD_getFrameHeader(&zfh, src, srcSize);
-            if (ZSTD_isError(ret)) return ret;
-            if (ret > 0) return ERROR(srcSize_wrong);
-        }
-
-        ip += zfh.headerSize;
-        remainingSize -= zfh.headerSize;
-
-        /* Loop on each block */
-        while (1) {
-            blockProperties_t blockProperties;
-            size_t const cBlockSize = ZSTD_getcBlockSize(ip, remainingSize, &blockProperties);
-            if (ZSTD_isError(cBlockSize)) return cBlockSize;
-
-            if (ZSTD_blockHeaderSize + cBlockSize > remainingSize)
-                return ERROR(srcSize_wrong);
-
-            ip += ZSTD_blockHeaderSize + cBlockSize;
-            remainingSize -= ZSTD_blockHeaderSize + cBlockSize;
-
-            if (blockProperties.lastBlock) break;
-        }
-
-        if (zfh.checksumFlag) {   /* Final frame content checksum */
-            if (remainingSize < 4) return ERROR(srcSize_wrong);
-            ip += 4;
-        }
-
-        return ip - ipstart;
-    }
+    ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+    return frameSizeInfo.compressedSize;
 }
 
+/** ZSTD_decompressBound() :
+ *  compatible with legacy mode
+ *  `src` must point to the start of a ZSTD frame or a skippeable frame
+ *  `srcSize` must be at least as large as the frame contained
+ *  @return : the maximum decompressed size of the compressed source
+ */
+unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize)
+{
+    unsigned long long bound = 0;
+    /* Iterate over each frame */
+    while (srcSize > 0) {
+        ZSTD_frameSizeInfo const frameSizeInfo = ZSTD_findFrameSizeInfo(src, srcSize);
+        size_t const compressedSize = frameSizeInfo.compressedSize;
+        unsigned long long const decompressedBound = frameSizeInfo.decompressedBound;
+        if (ZSTD_isError(compressedSize) || decompressedBound == ZSTD_CONTENTSIZE_ERROR)
+            return ZSTD_CONTENTSIZE_ERROR;
+        assert(srcSize >= compressedSize);
+        src = (const BYTE*)src + compressedSize;
+        srcSize -= compressedSize;
+        bound += decompressedBound;
+    }
+    return bound;
+}
 
 
 /*-*************************************************************
@@ -507,9 +574,10 b' void ZSTD_checkContinuity(ZSTD_DCtx* dct'
 }
 
 /** ZSTD_insertBlock() :
-    insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
+ *  insert `src` block into `dctx` history. Useful to track uncompressed blocks. */
 size_t ZSTD_insertBlock(ZSTD_DCtx* dctx, const void* blockStart, size_t blockSize)
 {
+    DEBUGLOG(5, "ZSTD_insertBlock: %u bytes", (unsigned)blockSize);
     ZSTD_checkContinuity(dctx, blockStart);
     dctx->previousDstEnd = (const char*)blockStart + blockSize;
     return blockSize;
@@ -522,9 +590,9 b' static size_t ZSTD_copyRawBlock(void* ds'
     DEBUGLOG(5, "ZSTD_copyRawBlock");
     if (dst == NULL) {
         if (srcSize == 0) return 0;
-        return ERROR(dstBuffer_null);
+        RETURN_ERROR(dstBuffer_null);
     }
-    if (srcSize > dstCapacity) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF(srcSize > dstCapacity, dstSize_tooSmall);
     memcpy(dst, src, srcSize);
     return srcSize;
 }
@@ -535,9 +603,9 b' static size_t ZSTD_setRleBlock(void* dst'
 {
     if (dst == NULL) {
         if (regenSize == 0) return 0;
-        return ERROR(dstBuffer_null);
+        RETURN_ERROR(dstBuffer_null);
     }
-    if (regenSize > dstCapacity) return ERROR(dstSize_tooSmall);
+    RETURN_ERROR_IF(regenSize > dstCapacity, dstSize_tooSmall);
     memset(dst, b, regenSize);
     return regenSize;
 }
@@ -560,15 +628,16 b' static size_t ZSTD_decompressFrame(ZSTD_'
     DEBUGLOG(4, "ZSTD_decompressFrame (srcSize:%i)", (int)*srcSizePtr);
 
     /* check */
-    if (remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN+ZSTD_blockHeaderSize)
-        return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(
+        remainingSrcSize < ZSTD_FRAMEHEADERSIZE_MIN+ZSTD_blockHeaderSize,
+        srcSize_wrong);
 
     /* Frame Header */
     {   size_t const frameHeaderSize = ZSTD_frameHeaderSize(ip, ZSTD_FRAMEHEADERSIZE_PREFIX);
         if (ZSTD_isError(frameHeaderSize)) return frameHeaderSize;
-        if (remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize)
-            return ERROR(srcSize_wrong);
-        CHECK_F( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) );
+        RETURN_ERROR_IF(remainingSrcSize < frameHeaderSize+ZSTD_blockHeaderSize,
+                        srcSize_wrong);
+        FORWARD_IF_ERROR( ZSTD_decodeFrameHeader(dctx, ip, frameHeaderSize) );
         ip += frameHeaderSize; remainingSrcSize -= frameHeaderSize;
     }
 
@@ -581,7 +650,7 b' static size_t ZSTD_decompressFrame(ZSTD_'
 
         ip += ZSTD_blockHeaderSize;
         remainingSrcSize -= ZSTD_blockHeaderSize;
-        if (cBlockSize > remainingSrcSize) return ERROR(srcSize_wrong);
+        RETURN_ERROR_IF(cBlockSize > remainingSrcSize, srcSize_wrong);
 
         switch(blockProperties.blockType)
         {
@@ -596,7 +665,7 b' static size_t ZSTD_decompressFrame(ZSTD_'
             break;
         case bt_reserved :
         default:
-            return ERROR(corruption_detected);
+            RETURN_ERROR(corruption_detected);
         }
 
         if (ZSTD_isError(decodedSize)) return decodedSize;
@@ -609,15 +678,15 b' static size_t ZSTD_decompressFrame(ZSTD_'
     }
 
     if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
-        if ((U64)(op-ostart) != dctx->fParams.frameContentSize) {
-            return ERROR(corruption_detected);
-    }   }
+        RETURN_ERROR_IF((U64)(op-ostart) != dctx->fParams.frameContentSize,
+                        corruption_detected);
+    }
     if (dctx->fParams.checksumFlag) { /* Frame content checksum verification */
         U32 const checkCalc = (U32)XXH64_digest(&dctx->xxhState);
         U32 checkRead;
-        if (remainingSrcSize<4) return ERROR(checksum_wrong);
+        RETURN_ERROR_IF(remainingSrcSize<4, checksum_wrong);
         checkRead = MEM_readLE32(ip);
-        if (checkRead != checkCalc) return ERROR(checksum_wrong);
+        RETURN_ERROR_IF(checkRead != checkCalc, checksum_wrong);
         ip += 4;
         remainingSrcSize -= 4;
     }
@@ -652,8 +721,8 b' static size_t ZSTD_decompressMultiFrame('
             size_t decodedSize;
             size_t const frameSize = ZSTD_findFrameCompressedSizeLegacy(src, srcSize);
             if (ZSTD_isError(frameSize)) return frameSize;
-            /* legacy support is not compatible with static dctx */
-            if (dctx->staticSize) return ERROR(memory_allocation);
+            RETURN_ERROR_IF(dctx->staticSize, memory_allocation,
+                "legacy support is not compatible with static dctx");
 
             decodedSize = ZSTD_decompressLegacy(dst, dstCapacity, src, frameSize, dict, dictSize);
             if (ZSTD_isError(decodedSize)) return decodedSize;
@@ -674,9 +743,8 b' static size_t ZSTD_decompressMultiFrame('
                         (unsigned)magicNumber, ZSTD_MAGICNUMBER);
             if ((magicNumber & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {
                 size_t const skippableSize = readSkippableFrameSize(src, srcSize);
-                if (ZSTD_isError(skippableSize))
-                    return skippableSize;
-                if (srcSize < skippableSize) return ERROR(srcSize_wrong);
+                FORWARD_IF_ERROR(skippableSize);
+                assert(skippableSize <= srcSize);
 
                 src = (const BYTE *)src + skippableSize;
                 srcSize -= skippableSize;
@@ -685,29 +753,29 b' static size_t ZSTD_decompressMultiFrame('
 
         if (ddict) {
             /* we were called from ZSTD_decompress_usingDDict */
-            CHECK_F(ZSTD_decompressBegin_usingDDict(dctx, ddict));
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(dctx, ddict));
         } else {
             /* this will initialize correctly with no dict if dict == NULL, so
              * use this in all cases but ddict */
-            CHECK_F(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDict(dctx, dict, dictSize));
         }
         ZSTD_checkContinuity(dctx, dst);
 
         {   const size_t res = ZSTD_decompressFrame(dctx, dst, dstCapacity,
                                                     &src, &srcSize);
-            if ( (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown)
-              && (moreThan1Frame==1) ) {
-                /* at least one frame successfully completed,
-                 * but following bytes are garbage :
-                 * it's more likely to be a srcSize error,
-                 * specifying more bytes than compressed size of frame(s).
-                 * This error message replaces ERROR(prefix_unknown),
-                 * which would be confusing, as the first header is actually correct.
-                 * Note that one could be unlucky, it might be a corruption error instead,
-                 * happening right at the place where we expect zstd magic bytes.
-                 * But this is _much_ less likely than a srcSize field error. */
-                return ERROR(srcSize_wrong);
-            }
+            RETURN_ERROR_IF(
+                (ZSTD_getErrorCode(res) == ZSTD_error_prefix_unknown)
+             && (moreThan1Frame==1),
+                srcSize_wrong,
+                "at least one frame successfully completed, but following "
+                "bytes are garbage: it's more likely to be a srcSize error, "
+                "specifying more bytes than compressed size of frame(s). This "
+                "error message replaces ERROR(prefix_unknown), which would be "
+                "confusing, as the first header is actually correct. Note that "
+                "one could be unlucky, it might be a corruption error instead, "
+                "happening right at the place where we expect zstd magic "
+                "bytes. But this is _much_ less likely than a srcSize field "
+                "error.");
             if (ZSTD_isError(res)) return res;
             assert(res <= dstCapacity);
             dst = (BYTE*)dst + res;
@@ -716,7 +784,7 b' static size_t ZSTD_decompressMultiFrame('
         moreThan1Frame = 1;
     }  /* while (srcSize >= ZSTD_frameHeaderSize_prefix) */
 
-    if (srcSize) return ERROR(srcSize_wrong); /* input not entirely consumed */
+    RETURN_ERROR_IF(srcSize, srcSize_wrong, "input not entirely consumed");
 
     return (BYTE*)dst - (BYTE*)dststart;
 }
@@ -730,9 +798,26 b' size_t ZSTD_decompress_usingDict(ZSTD_DC'
 }
 
 
+static ZSTD_DDict const* ZSTD_getDDict(ZSTD_DCtx* dctx)
+{
+    switch (dctx->dictUses) {
+    default:
+        assert(0 /* Impossible */);
+        /* fall-through */
+    case ZSTD_dont_use:
+        ZSTD_clearDict(dctx);
+        return NULL;
+    case ZSTD_use_indefinitely:
+        return dctx->ddict;
+    case ZSTD_use_once:
+        dctx->dictUses = ZSTD_dont_use;
+        return dctx->ddict;
+    }
+}
+
 size_t ZSTD_decompressDCtx(ZSTD_DCtx* dctx, void* dst, size_t dstCapacity, const void* src, size_t srcSize)
 {
-    return ZSTD_decompress_usingDict(dctx, dst, dstCapacity, src, srcSize, NULL, 0);
+    return ZSTD_decompress_usingDDict(dctx, dst, dstCapacity, src, srcSize, ZSTD_getDDict(dctx));
 }
 
 
@@ -741,7 +826,7 b' size_t ZSTD_decompress(void* dst, size_t'
 #if defined(ZSTD_HEAPMODE) && (ZSTD_HEAPMODE>=1)
     size_t regenSize;
     ZSTD_DCtx* const dctx = ZSTD_createDCtx();
-    if (dctx==NULL) return ERROR(memory_allocation);
+    RETURN_ERROR_IF(dctx==NULL, memory_allocation);
     regenSize = ZSTD_decompressDCtx(dctx, dst, dstCapacity, src, srcSize);
     ZSTD_freeDCtx(dctx);
     return regenSize;
@@ -791,8 +876,7 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
 {
     DEBUGLOG(5, "ZSTD_decompressContinue (srcSize:%u)", (unsigned)srcSize);
     /* Sanity check */
-    if (srcSize != dctx->expected)
-        return ERROR(srcSize_wrong);  /* not allowed */
+    RETURN_ERROR_IF(srcSize != dctx->expected, srcSize_wrong, "not allowed");
     if (dstCapacity) ZSTD_checkContinuity(dctx, dst);
 
     switch (dctx->stage)
@@ -817,7 +901,7 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
     case ZSTDds_decodeFrameHeader:
         assert(src != NULL);
         memcpy(dctx->headerBuffer + (dctx->headerSize - srcSize), src, srcSize);
-        CHECK_F(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
+        FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(dctx, dctx->headerBuffer, dctx->headerSize));
         dctx->expected = ZSTD_blockHeaderSize;
         dctx->stage = ZSTDds_decodeBlockHeader;
         return 0;
@@ -826,6 +910,7 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
         {   blockProperties_t bp;
             size_t const cBlockSize = ZSTD_getcBlockSize(src, ZSTD_blockHeaderSize, &bp);
             if (ZSTD_isError(cBlockSize)) return cBlockSize;
+            RETURN_ERROR_IF(cBlockSize > dctx->fParams.blockSizeMax, corruption_detected, "Block Size Exceeds Maximum");
             dctx->expected = cBlockSize;
             dctx->bType = bp.blockType;
             dctx->rleSize = bp.origSize;
@@ -867,19 +952,20 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
                 break;
             case bt_reserved :   /* should never happen */
             default:
-                return ERROR(corruption_detected);
+                RETURN_ERROR(corruption_detected);
             }
             if (ZSTD_isError(rSize)) return rSize;
+            RETURN_ERROR_IF(rSize > dctx->fParams.blockSizeMax, corruption_detected, "Decompressed Block Size Exceeds Maximum");
             DEBUGLOG(5, "ZSTD_decompressContinue: decoded size from block : %u", (unsigned)rSize);
             dctx->decodedSize += rSize;
             if (dctx->fParams.checksumFlag) XXH64_update(&dctx->xxhState, dst, rSize);
 
             if (dctx->stage == ZSTDds_decompressLastBlock) {   /* end of frame */
                 DEBUGLOG(4, "ZSTD_decompressContinue: decoded size from frame : %u", (unsigned)dctx->decodedSize);
-                if (dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN) {
-                    if (dctx->decodedSize != dctx->fParams.frameContentSize) {
-                        return ERROR(corruption_detected);
-                }   }
+                RETURN_ERROR_IF(
+                    dctx->fParams.frameContentSize != ZSTD_CONTENTSIZE_UNKNOWN
+                 && dctx->decodedSize != dctx->fParams.frameContentSize,
+                    corruption_detected);
                 if (dctx->fParams.checksumFlag) {  /* another round for frame checksum */
                     dctx->expected = 4;
                     dctx->stage = ZSTDds_checkChecksum;
@@ -900,7 +986,7 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
         {   U32 const h32 = (U32)XXH64_digest(&dctx->xxhState);
             U32 const check32 = MEM_readLE32(src);
             DEBUGLOG(4, "ZSTD_decompressContinue: checksum : calculated %08X :: %08X read", (unsigned)h32, (unsigned)check32);
-            if (check32 != h32) return ERROR(checksum_wrong);
+            RETURN_ERROR_IF(check32 != h32, checksum_wrong);
             dctx->expected = 0;
             dctx->stage = ZSTDds_getFrameHeaderSize;
             return 0;
@@ -921,7 +1007,7 b' size_t ZSTD_decompressContinue(ZSTD_DCtx'
 
     default:
         assert(0);   /* impossible */
-        return ERROR(GENERIC);   /* some compiler require default to do something */
+        RETURN_ERROR(GENERIC);   /* some compiler require default to do something */
     }
 }
 
@@ -945,7 +1031,7 b' ZSTD_loadDEntropy(ZSTD_entropyDTables_t*'
     const BYTE* dictPtr = (const BYTE*)dict;
     const BYTE* const dictEnd = dictPtr + dictSize;
 
-    if (dictSize <= 8) return ERROR(dictionary_corrupted);
+    RETURN_ERROR_IF(dictSize <= 8, dictionary_corrupted);
     assert(MEM_readLE32(dict) == ZSTD_MAGIC_DICTIONARY);   /* dict must be valid */
     dictPtr += 8;   /* skip header = magic + dictID */
 
@@ -964,16 +1050,16 b' ZSTD_loadDEntropy(ZSTD_entropyDTables_t*'
                                                 dictPtr, dictEnd - dictPtr,
                                                 workspace, workspaceSize);
 #endif
-        if (HUF_isError(hSize)) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(HUF_isError(hSize), dictionary_corrupted);
         dictPtr += hSize;
     }
 
     {   short offcodeNCount[MaxOff+1];
         unsigned offcodeMaxValue = MaxOff, offcodeLog;
         size_t const offcodeHeaderSize = FSE_readNCount(offcodeNCount, &offcodeMaxValue, &offcodeLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(offcodeHeaderSize)) return ERROR(dictionary_corrupted);
-        if (offcodeMaxValue > MaxOff) return ERROR(dictionary_corrupted);
-        if (offcodeLog > OffFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(offcodeHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(offcodeMaxValue > MaxOff, dictionary_corrupted);
+        RETURN_ERROR_IF(offcodeLog > OffFSELog, dictionary_corrupted);
         ZSTD_buildFSETable( entropy->OFTable,
                             offcodeNCount, offcodeMaxValue,
                             OF_base, OF_bits,
@@ -984,9 +1070,9 b' ZSTD_loadDEntropy(ZSTD_entropyDTables_t*'
     {   short matchlengthNCount[MaxML+1];
         unsigned matchlengthMaxValue = MaxML, matchlengthLog;
         size_t const matchlengthHeaderSize = FSE_readNCount(matchlengthNCount, &matchlengthMaxValue, &matchlengthLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(matchlengthHeaderSize)) return ERROR(dictionary_corrupted);
-        if (matchlengthMaxValue > MaxML) return ERROR(dictionary_corrupted);
-        if (matchlengthLog > MLFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(matchlengthHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(matchlengthMaxValue > MaxML, dictionary_corrupted);
+        RETURN_ERROR_IF(matchlengthLog > MLFSELog, dictionary_corrupted);
         ZSTD_buildFSETable( entropy->MLTable,
                             matchlengthNCount, matchlengthMaxValue,
                             ML_base, ML_bits,
@@ -997,9 +1083,9 b' ZSTD_loadDEntropy(ZSTD_entropyDTables_t*'
     {   short litlengthNCount[MaxLL+1];
         unsigned litlengthMaxValue = MaxLL, litlengthLog;
         size_t const litlengthHeaderSize = FSE_readNCount(litlengthNCount, &litlengthMaxValue, &litlengthLog, dictPtr, dictEnd-dictPtr);
-        if (FSE_isError(litlengthHeaderSize)) return ERROR(dictionary_corrupted);
-        if (litlengthMaxValue > MaxLL) return ERROR(dictionary_corrupted);
-        if (litlengthLog > LLFSELog) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(FSE_isError(litlengthHeaderSize), dictionary_corrupted);
+        RETURN_ERROR_IF(litlengthMaxValue > MaxLL, dictionary_corrupted);
+        RETURN_ERROR_IF(litlengthLog > LLFSELog, dictionary_corrupted);
         ZSTD_buildFSETable( entropy->LLTable,
                             litlengthNCount, litlengthMaxValue,
                             LL_base, LL_bits,
@@ -1007,12 +1093,13 b' ZSTD_loadDEntropy(ZSTD_entropyDTables_t*'
         dictPtr += litlengthHeaderSize;
     }
 
-    if (dictPtr+12 > dictEnd) return ERROR(dictionary_corrupted);
+    RETURN_ERROR_IF(dictPtr+12 > dictEnd, dictionary_corrupted);
     {   int i;
         size_t const dictContentSize = (size_t)(dictEnd - (dictPtr+12));
         for (i=0; i<3; i++) {
             U32 const rep = MEM_readLE32(dictPtr); dictPtr += 4;
-            if (rep==0 || rep >= dictContentSize) return ERROR(dictionary_corrupted);
+            RETURN_ERROR_IF(rep==0 || rep >= dictContentSize,
+                            dictionary_corrupted);
             entropy->rep[i] = rep;
     }   }
 
@@ -1030,7 +1117,7 b' static size_t ZSTD_decompress_insertDict'
 
     /* load entropy tables */
     {   size_t const eSize = ZSTD_loadDEntropy(&dctx->entropy, dict, dictSize);
-        if (ZSTD_isError(eSize)) return ERROR(dictionary_corrupted);
+        RETURN_ERROR_IF(ZSTD_isError(eSize), dictionary_corrupted);
         dict = (const char*)dict + eSize;
         dictSize -= eSize;
     }
@@ -1064,9 +1151,11 b' size_t ZSTD_decompressBegin(ZSTD_DCtx* d'
 
 size_t ZSTD_decompressBegin_usingDict(ZSTD_DCtx* dctx, const void* dict, size_t dictSize)
 {
-    CHECK_F( ZSTD_decompressBegin(dctx) );
+    FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );
     if (dict && dictSize)
-        CHECK_E(ZSTD_decompress_insertDictionary(dctx, dict, dictSize), dictionary_corrupted);
+        RETURN_ERROR_IF(
+            ZSTD_isError(ZSTD_decompress_insertDictionary(dctx, dict, dictSize)),
+            dictionary_corrupted);
     return 0;
 }
 
@@ -1085,7 +1174,7 b' size_t ZSTD_decompressBegin_usingDDict(Z'
         DEBUGLOG(4, "DDict is %s",
                     dctx->ddictIsCold ? "~cold~" : "hot!");
     }
-    CHECK_F( ZSTD_decompressBegin(dctx) );
+    FORWARD_IF_ERROR( ZSTD_decompressBegin(dctx) );
     if (ddict) {   /* NULL ddict is equivalent to no dictionary */
         ZSTD_copyDDictParameters(dctx, ddict);
     }
@@ -1104,7 +1193,7 b' unsigned ZSTD_getDictID_fromDict(const v'
 }
 
 /*! ZSTD_getDictID_fromFrame() :
- *  Provides the dictID required to decompresse frame stored within `src`.
+ *  Provides the dictID required to decompress frame stored within `src`.
  *  If @return == 0, the dictID could not be decoded.
  *  This could for one of the following reasons :
  *  - The frame does not require a dictionary (most common case).
@@ -1176,15 +1265,14 b' size_t ZSTD_DCtx_loadDictionary_advanced'
                                          ZSTD_dictLoadMethod_e dictLoadMethod,
                                          ZSTD_dictContentType_e dictContentType)
 {
-    if (dctx->streamStage != zdss_init) return ERROR(stage_wrong);
-    ZSTD_freeDDict(dctx->ddictLocal);
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
+    ZSTD_clearDict(dctx);
     if (dict && dictSize >= 8) {
         dctx->ddictLocal = ZSTD_createDDict_advanced(dict, dictSize, dictLoadMethod, dictContentType, dctx->customMem);
-        if (dctx->ddictLocal == NULL) return ERROR(memory_allocation);
-    } else {
-        dctx->ddictLocal = NULL;
+        RETURN_ERROR_IF(dctx->ddictLocal == NULL, memory_allocation);
+        dctx->ddict = dctx->ddictLocal;
+        dctx->dictUses = ZSTD_use_indefinitely;
     }
-    dctx->ddict = dctx->ddictLocal;
     return 0;
 }
 
@@ -1200,7 +1288,9 b' size_t ZSTD_DCtx_loadDictionary(ZSTD_DCt'
 
 size_t ZSTD_DCtx_refPrefix_advanced(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize, ZSTD_dictContentType_e dictContentType)
 {
-    return ZSTD_DCtx_loadDictionary_advanced(dctx, prefix, prefixSize, ZSTD_dlm_byRef, dictContentType);
+    FORWARD_IF_ERROR(ZSTD_DCtx_loadDictionary_advanced(dctx, prefix, prefixSize, ZSTD_dlm_byRef, dictContentType));
+    dctx->dictUses = ZSTD_use_once;
+    return 0;
 }
 
 size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx, const void* prefix, size_t prefixSize)
@@ -1215,9 +1305,8 b' size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dc'
 size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize)
 {
     DEBUGLOG(4, "ZSTD_initDStream_usingDict");
-    zds->streamStage = zdss_init;
-    zds->noForwardProgress = 0;
-    CHECK_F( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) );
+    FORWARD_IF_ERROR( ZSTD_DCtx_reset(zds, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_DCtx_loadDictionary(zds, dict, dictSize) );
     return ZSTD_FRAMEHEADERSIZE_PREFIX;
 }
 
@@ -1225,7 +1314,7 b' size_t ZSTD_initDStream_usingDict(ZSTD_D'
 size_t ZSTD_initDStream(ZSTD_DStream* zds)
 {
     DEBUGLOG(4, "ZSTD_initDStream");
-    return ZSTD_initDStream_usingDict(zds, NULL, 0);
+    return ZSTD_initDStream_usingDDict(zds, NULL);
 }
 
 /* ZSTD_initDStream_usingDDict() :
@@ -1233,9 +1322,9 b' size_t ZSTD_initDStream(ZSTD_DStream* zd'
  * this function cannot fail */
 size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* dctx, const ZSTD_DDict* ddict)
 {
-    size_t const initResult = ZSTD_initDStream(dctx);
-    dctx->ddict = ddict;
-    return initResult;
+    FORWARD_IF_ERROR( ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only) );
+    FORWARD_IF_ERROR( ZSTD_DCtx_refDDict(dctx, ddict) );
+    return ZSTD_FRAMEHEADERSIZE_PREFIX;
 }
 
 /* ZSTD_resetDStream() :
@@ -1243,19 +1332,19 b' size_t ZSTD_initDStream_usingDDict(ZSTD_'
  * this function cannot fail */
 size_t ZSTD_resetDStream(ZSTD_DStream* dctx)
 {
-    DEBUGLOG(4, "ZSTD_resetDStream");
-    dctx->streamStage = zdss_loadHeader;
-    dctx->lhSize = dctx->inPos = dctx->outStart = dctx->outEnd = 0;
-    dctx->legacyVersion = 0;
-    dctx->hostageByte = 0;
+    FORWARD_IF_ERROR(ZSTD_DCtx_reset(dctx, ZSTD_reset_session_only));
     return ZSTD_FRAMEHEADERSIZE_PREFIX;
 }
 
 
 size_t ZSTD_DCtx_refDDict(ZSTD_DCtx* dctx, const ZSTD_DDict* ddict)
 {
-    if (dctx->streamStage != zdss_init) return ERROR(stage_wrong);
-    dctx->ddict = ddict;
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
+    ZSTD_clearDict(dctx);
+    if (ddict) {
+        dctx->ddict = ddict;
+        dctx->dictUses = ZSTD_use_indefinitely;
+    }
     return 0;
 }
 
@@ -1267,9 +1356,9 b' size_t ZSTD_DCtx_setMaxWindowSize(ZSTD_D'
     ZSTD_bounds const bounds = ZSTD_dParam_getBounds(ZSTD_d_windowLogMax);
     size_t const min = (size_t)1 << bounds.lowerBound;
     size_t const max = (size_t)1 << bounds.upperBound;
-    if (dctx->streamStage != zdss_init) return ERROR(stage_wrong);
-    if (maxWindowSize < min) return ERROR(parameter_outOfBound);
-    if (maxWindowSize > max) return ERROR(parameter_outOfBound);
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
+    RETURN_ERROR_IF(maxWindowSize < min, parameter_outOfBound);
+    RETURN_ERROR_IF(maxWindowSize > max, parameter_outOfBound);
     dctx->maxWindowSize = maxWindowSize;
     return 0;
 }
@@ -1311,15 +1400,15 b' static int ZSTD_dParam_withinBounds(ZSTD'
 }
 
 #define CHECK_DBOUNDS(p,v) {                \
-    if (!ZSTD_dParam_withinBounds(p, v))    \
-        return ERROR(parameter_outOfBound); \
+    RETURN_ERROR_IF(!ZSTD_dParam_withinBounds(p, v), parameter_outOfBound); \
 }
 
 size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter dParam, int value)
 {
-    if (dctx->streamStage != zdss_init) return ERROR(stage_wrong);
+    RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
     switch(dParam) {
         case ZSTD_d_windowLogMax:
+            if (value == 0) value = ZSTD_WINDOWLOG_LIMIT_DEFAULT;
             CHECK_DBOUNDS(ZSTD_d_windowLogMax, value);
             dctx->maxWindowSize = ((size_t)1) << value;
             return 0;
@@ -1329,19 +1418,20 b' size_t ZSTD_DCtx_setParameter(ZSTD_DCtx*'
             return 0;
         default:;
     }
-    return ERROR(parameter_unsupported);
+    RETURN_ERROR(parameter_unsupported);
 }
 
 size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset)
 {
     if ( (reset == ZSTD_reset_session_only)
       || (reset == ZSTD_reset_session_and_parameters) ) {
-        (void)ZSTD_initDStream(dctx);
+        dctx->streamStage = zdss_init;
+        dctx->noForwardProgress = 0;
     }
     if ( (reset == ZSTD_reset_parameters)
       || (reset == ZSTD_reset_session_and_parameters) ) {
-        if (dctx->streamStage != zdss_init)
-            return ERROR(stage_wrong);
+        RETURN_ERROR_IF(dctx->streamStage != zdss_init, stage_wrong);
+        ZSTD_clearDict(dctx);
         dctx->format = ZSTD_f_zstd1;
         dctx->maxWindowSize = ZSTD_MAXWINDOWSIZE_DEFAULT;
     }
@@ -1360,7 +1450,8 b' size_t ZSTD_decodingBufferSize_min(unsig'
     unsigned long long const neededRBSize = windowSize + blockSize + (WILDCOPY_OVERLENGTH * 2);
     unsigned long long const neededSize = MIN(frameContentSize, neededRBSize);
     size_t const minRBSize = (size_t) neededSize;
-    if ((unsigned long long)minRBSize != neededSize) return ERROR(frameParameter_windowTooLarge);
+    RETURN_ERROR_IF((unsigned long long)minRBSize != neededSize,
+                    frameParameter_windowTooLarge);
     return minRBSize;
 }
 
@@ -1378,9 +1469,9 b' size_t ZSTD_estimateDStreamSize_fromFram'
     ZSTD_frameHeader zfh;
     size_t const err = ZSTD_getFrameHeader(&zfh, src, srcSize);
     if (ZSTD_isError(err)) return err;
-    if (err>0) return ERROR(srcSize_wrong);
-    if (zfh.windowSize > windowSizeMax)
-        return ERROR(frameParameter_windowTooLarge);
+    RETURN_ERROR_IF(err>0, srcSize_wrong);
+    RETURN_ERROR_IF(zfh.windowSize > windowSizeMax,
+                    frameParameter_windowTooLarge);
     return ZSTD_estimateDStreamSize((size_t)zfh.windowSize);
 }
 
@@ -1406,16 +1497,16 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
     U32 someMoreWork = 1;
 
     DEBUGLOG(5, "ZSTD_decompressStream");
-    if (input->pos > input->size) {  /* forbidden */
-        DEBUGLOG(5, "in: pos: %u   vs size: %u",
-                    (U32)input->pos, (U32)input->size);
-        return ERROR(srcSize_wrong);
-    }
-    if (output->pos > output->size) {  /* forbidden */
-        DEBUGLOG(5, "out: pos: %u   vs size: %u",
-                    (U32)output->pos, (U32)output->size);
-        return ERROR(dstSize_tooSmall);
-    }
+    RETURN_ERROR_IF(
+        input->pos > input->size,
+        srcSize_wrong,
+        "forbidden. in: pos: %u   vs size: %u",
+        (U32)input->pos, (U32)input->size);
+    RETURN_ERROR_IF(
+        output->pos > output->size,
+        dstSize_tooSmall,
+        "forbidden. out: pos: %u   vs size: %u",
+        (U32)output->pos, (U32)output->size);
     DEBUGLOG(5, "input size : %u", (U32)(input->size - input->pos));
 
     while (someMoreWork) {
@@ -1423,15 +1514,18 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
         {
         case zdss_init :
             DEBUGLOG(5, "stage zdss_init => transparent reset ");
-            ZSTD_resetDStream(zds);   /* transparent reset on starting decoding a new frame */
+            zds->streamStage = zdss_loadHeader;
+            zds->lhSize = zds->inPos = zds->outStart = zds->outEnd = 0;
+            zds->legacyVersion = 0;
+            zds->hostageByte = 0;
             /* fall-through */
 
         case zdss_loadHeader :
             DEBUGLOG(5, "stage zdss_loadHeader (srcSize : %u)", (U32)(iend - ip));
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
             if (zds->legacyVersion) {
-                /* legacy support is incompatible with static dctx */
-                if (zds->staticSize) return ERROR(memory_allocation);
+                RETURN_ERROR_IF(zds->staticSize, memory_allocation,
+                    "legacy support is incompatible with static dctx");
                 {   size_t const hint = ZSTD_decompressLegacyStream(zds->legacyContext, zds->legacyVersion, output, input);
                     if (hint==0) zds->streamStage = zdss_init;
                     return hint;
@@ -1443,12 +1537,13 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>=1)
                     U32 const legacyVersion = ZSTD_isLegacy(istart, iend-istart);
                     if (legacyVersion) {
-                        const void* const dict = zds->ddict ? ZSTD_DDict_dictContent(zds->ddict) : NULL;
-                        size_t const dictSize = zds->ddict ? ZSTD_DDict_dictSize(zds->ddict) : 0;
+                        ZSTD_DDict const* const ddict = ZSTD_getDDict(zds);
+                        const void* const dict = ddict ? ZSTD_DDict_dictContent(ddict) : NULL;
+                        size_t const dictSize = ddict ? ZSTD_DDict_dictSize(ddict) : 0;
                         DEBUGLOG(5, "ZSTD_decompressStream: detected legacy version v0.%u", legacyVersion);
-                        /* legacy support is incompatible with static dctx */
-                        if (zds->staticSize) return ERROR(memory_allocation);
-                        CHECK_F(ZSTD_initLegacyStream(&zds->legacyContext,
+                        RETURN_ERROR_IF(zds->staticSize, memory_allocation,
+                            "legacy support is incompatible with static dctx");
+                        FORWARD_IF_ERROR(ZSTD_initLegacyStream(&zds->legacyContext,
                                     zds->previousLegacyVersion, legacyVersion,
                                     dict, dictSize));
                         zds->legacyVersion = zds->previousLegacyVersion = legacyVersion;
@@ -1482,7 +1577,7 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
                 size_t const cSize = ZSTD_findFrameCompressedSize(istart, iend-istart);
                 if (cSize <= (size_t)(iend-istart)) {
                     /* shortcut : using single-pass mode */
-                    size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, oend-op, istart, cSize, zds->ddict);
+                    size_t const decompressedSize = ZSTD_decompress_usingDDict(zds, op, oend-op, istart, cSize, ZSTD_getDDict(zds));
                     if (ZSTD_isError(decompressedSize)) return decompressedSize;
                     DEBUGLOG(4, "shortcut to single-pass ZSTD_decompress_usingDDict()")
                     ip = istart + cSize;
@@ -1495,13 +1590,13 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
 
             /* Consume header (see ZSTDds_decodeFrameHeader) */
             DEBUGLOG(4, "Consume header");
-            CHECK_F(ZSTD_decompressBegin_usingDDict(zds, zds->ddict));
+            FORWARD_IF_ERROR(ZSTD_decompressBegin_usingDDict(zds, ZSTD_getDDict(zds)));
 
             if ((MEM_readLE32(zds->headerBuffer) & ZSTD_MAGIC_SKIPPABLE_MASK) == ZSTD_MAGIC_SKIPPABLE_START) {  /* skippable frame */
                 zds->expected = MEM_readLE32(zds->headerBuffer + ZSTD_FRAMEIDSIZE);
                 zds->stage = ZSTDds_skipFrame;
             } else {
-                CHECK_F(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize));
+                FORWARD_IF_ERROR(ZSTD_decodeFrameHeader(zds, zds->headerBuffer, zds->lhSize));
                 zds->expected = ZSTD_blockHeaderSize;
                 zds->stage = ZSTDds_decodeBlockHeader;
             }
@@ -1511,7 +1606,8 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
                         (U32)(zds->fParams.windowSize >>10),
                         (U32)(zds->maxWindowSize >> 10) );
             zds->fParams.windowSize = MAX(zds->fParams.windowSize, 1U << ZSTD_WINDOWLOG_ABSOLUTEMIN);
-            if (zds->fParams.windowSize > zds->maxWindowSize) return ERROR(frameParameter_windowTooLarge);
+            RETURN_ERROR_IF(zds->fParams.windowSize > zds->maxWindowSize,
+                            frameParameter_windowTooLarge);
 
             /* Adapt buffer sizes to frame header instructions */
             {   size_t const neededInBuffSize = MAX(zds->fParams.blockSizeMax, 4 /* frame checksum */);
@@ -1525,14 +1621,15 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
                     if (zds->staticSize) {  /* static DCtx */
                         DEBUGLOG(4, "staticSize : %u", (U32)zds->staticSize);
                         assert(zds->staticSize >= sizeof(ZSTD_DCtx));  /* controlled at init */
-                        if (bufferSize > zds->staticSize - sizeof(ZSTD_DCtx))
-                            return ERROR(memory_allocation);
+                        RETURN_ERROR_IF(
+                            bufferSize > zds->staticSize - sizeof(ZSTD_DCtx),
+                            memory_allocation);
                     } else {
                         ZSTD_free(zds->inBuff, zds->customMem);
                         zds->inBuffSize = 0;
                         zds->outBuffSize = 0;
                         zds->inBuff = (char*)ZSTD_malloc(bufferSize, zds->customMem);
-                        if (zds->inBuff == NULL) return ERROR(memory_allocation);
+                        RETURN_ERROR_IF(zds->inBuff == NULL, memory_allocation);
                     }
                     zds->inBuffSize = neededInBuffSize;
                     zds->outBuff = zds->inBuff + zds->inBuffSize;
@@ -1574,7 +1671,9 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
                 if (isSkipFrame) {
                     loadedSize = MIN(toLoad, (size_t)(iend-ip));
                 } else {
-                    if (toLoad > zds->inBuffSize - zds->inPos) return ERROR(corruption_detected);   /* should never happen */
+                    RETURN_ERROR_IF(toLoad > zds->inBuffSize - zds->inPos,
+                                    corruption_detected,
+                                    "should never happen");
                     loadedSize = ZSTD_limitCopy(zds->inBuff + zds->inPos, toLoad, ip, iend-ip);
                 }
                 ip += loadedSize;
@@ -1615,7 +1714,7 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
 
         default:
             assert(0);    /* impossible */
-            return ERROR(GENERIC);   /* some compiler require default to do something */
+            RETURN_ERROR(GENERIC);   /* some compiler require default to do something */
     }   }
 
     /* result */
@@ -1624,8 +1723,8 b' size_t ZSTD_decompressStream(ZSTD_DStrea'
     if ((ip==istart) && (op==ostart)) {  /* no forward progress */
         zds->noForwardProgress ++;
         if (zds->noForwardProgress >= ZSTD_NO_FORWARD_PROGRESS_MAX) {
-            if (op==oend) return ERROR(dstSize_tooSmall);
-            if (ip==iend) return ERROR(srcSize_wrong);
+            RETURN_ERROR_IF(op==oend, dstSize_tooSmall);
+            RETURN_ERROR_IF(ip==iend, srcSize_wrong);
             assert(0);
         }
     } else {

contrib/python-zstandard/zstd/decompress/zstd_decompress_block.c

@@ -56,14 +56,15 b' static void ZSTD_copy4(void* dst, const '
 size_t ZSTD_getcBlockSize(const void* src, size_t srcSize,
                           blockProperties_t* bpPtr)
 {
-    if (srcSize < ZSTD_blockHeaderSize) return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(srcSize < ZSTD_blockHeaderSize, srcSize_wrong);
+
     {   U32 const cBlockHeader = MEM_readLE24(src);
         U32 const cSize = cBlockHeader >> 3;
         bpPtr->lastBlock = cBlockHeader & 1;
         bpPtr->blockType = (blockType_e)((cBlockHeader >> 1) & 3);
         bpPtr->origSize = cSize;   /* only useful for RLE */
         if (bpPtr->blockType == bt_rle) return 1;
-        if (bpPtr->blockType == bt_reserved) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(bpPtr->blockType == bt_reserved, corruption_detected);
         return cSize;
     }
 }
@@ -78,7 +79,8 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
 size_t ZSTD_decodeLiteralsBlock(ZSTD_DCtx* dctx,
                           const void* src, size_t srcSize)   /* note : srcSize < BLOCKSIZE */
 {
-    if (srcSize < MIN_CBLOCK_SIZE) return ERROR(corruption_detected);
+    DEBUGLOG(5, "ZSTD_decodeLiteralsBlock");
+    RETURN_ERROR_IF(srcSize < MIN_CBLOCK_SIZE, corruption_detected);
 
     {   const BYTE* const istart = (const BYTE*) src;
         symbolEncodingType_e const litEncType = (symbolEncodingType_e)(istart[0] & 3);
@@ -86,11 +88,12 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
         switch(litEncType)
         {
         case set_repeat:
-            if (dctx->litEntropy==0) return ERROR(dictionary_corrupted);
+            DEBUGLOG(5, "set_repeat flag : re-using stats from previous compressed literals block");
+            RETURN_ERROR_IF(dctx->litEntropy==0, dictionary_corrupted);
             /* fall-through */
 
         case set_compressed:
-            if (srcSize < 5) return ERROR(corruption_detected);   /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3 */
+            RETURN_ERROR_IF(srcSize < 5, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need up to 5 for case 3");
             {   size_t lhSize, litSize, litCSize;
                 U32 singleStream=0;
                 U32 const lhlCode = (istart[0] >> 2) & 3;
@@ -115,11 +118,11 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
                     /* 2 - 2 - 18 - 18 */
                     lhSize = 5;
                     litSize  = (lhc >> 4) & 0x3FFFF;
-                    litCSize = (lhc >> 22) + (istart[4] << 10);
+                    litCSize = (lhc >> 22) + ((size_t)istart[4] << 10);
                     break;
                 }
-                if (litSize > ZSTD_BLOCKSIZE_MAX) return ERROR(corruption_detected);
-                if (litCSize + lhSize > srcSize) return ERROR(corruption_detected);
+                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);
+                RETURN_ERROR_IF(litCSize + lhSize > srcSize, corruption_detected);
 
                 /* prefetch huffman table if cold */
                 if (dctx->ddictIsCold && (litSize > 768 /* heuristic */)) {
@@ -157,7 +160,7 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
                     }
                 }
 
-                if (HUF_isError(hufSuccess)) return ERROR(corruption_detected);
+                RETURN_ERROR_IF(HUF_isError(hufSuccess), corruption_detected);
 
                 dctx->litPtr = dctx->litBuffer;
                 dctx->litSize = litSize;
@@ -187,7 +190,7 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
                 }
 
                 if (lhSize+litSize+WILDCOPY_OVERLENGTH > srcSize) {  /* risk reading beyond src buffer with wildcopy */
-                    if (litSize+lhSize > srcSize) return ERROR(corruption_detected);
+                    RETURN_ERROR_IF(litSize+lhSize > srcSize, corruption_detected);
                     memcpy(dctx->litBuffer, istart+lhSize, litSize);
                     dctx->litPtr = dctx->litBuffer;
                     dctx->litSize = litSize;
@@ -216,17 +219,17 b' size_t ZSTD_decodeLiteralsBlock(ZSTD_DCt'
                 case 3:
                     lhSize = 3;
                     litSize = MEM_readLE24(istart) >> 4;
-                    if (srcSize<4) return ERROR(corruption_detected);   /* srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4 */
+                    RETURN_ERROR_IF(srcSize<4, corruption_detected, "srcSize >= MIN_CBLOCK_SIZE == 3; here we need lhSize+1 = 4");
                     break;
                 }
-                if (litSize > ZSTD_BLOCKSIZE_MAX) return ERROR(corruption_detected);
+                RETURN_ERROR_IF(litSize > ZSTD_BLOCKSIZE_MAX, corruption_detected);
                 memset(dctx->litBuffer, istart[lhSize], litSize + WILDCOPY_OVERLENGTH);
                 dctx->litPtr = dctx->litBuffer;
                 dctx->litSize = litSize;
                 return lhSize+1;
             }
         default:
-            return ERROR(corruption_detected);   /* impossible */
+            RETURN_ERROR(corruption_detected, "impossible");
         }
     }
 }
@@ -390,7 +393,8 b' ZSTD_buildFSETable(ZSTD_seqSymbol* dt,'
                     symbolNext[s] = 1;
                 } else {
                     if (normalizedCounter[s] >= largeLimit) DTableH.fastMode=0;
-                    symbolNext[s] = normalizedCounter[s];
+                    assert(normalizedCounter[s]>=0);
+                    symbolNext[s] = (U16)normalizedCounter[s];
         }   }   }
         memcpy(dt, &DTableH, sizeof(DTableH));
     }
@@ -436,8 +440,8 b' static size_t ZSTD_buildSeqTable(ZSTD_se'
     switch(type)
     {
     case set_rle :
-        if (!srcSize) return ERROR(srcSize_wrong);
-        if ( (*(const BYTE*)src) > max) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(!srcSize, srcSize_wrong);
+        RETURN_ERROR_IF((*(const BYTE*)src) > max, corruption_detected);
         {   U32 const symbol = *(const BYTE*)src;
             U32 const baseline = baseValue[symbol];
             U32 const nbBits = nbAdditionalBits[symbol];
@@ -449,7 +453,7 b' static size_t ZSTD_buildSeqTable(ZSTD_se'
         *DTablePtr = defaultTable;
         return 0;
     case set_repeat:
-        if (!flagRepeatTable) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(!flagRepeatTable, corruption_detected);
         /* prefetch FSE table if used */
         if (ddictIsCold && (nbSeq > 24 /* heuristic */)) {
             const void* const pStart = *DTablePtr;
@@ -461,15 +465,15 b' static size_t ZSTD_buildSeqTable(ZSTD_se'
         {   unsigned tableLog;
             S16 norm[MaxSeq+1];
             size_t const headerSize = FSE_readNCount(norm, &max, &tableLog, src, srcSize);
-            if (FSE_isError(headerSize)) return ERROR(corruption_detected);
-            if (tableLog > maxLog) return ERROR(corruption_detected);
+            RETURN_ERROR_IF(FSE_isError(headerSize), corruption_detected);
+            RETURN_ERROR_IF(tableLog > maxLog, corruption_detected);
             ZSTD_buildFSETable(DTableSpace, norm, max, baseValue, nbAdditionalBits, tableLog);
             *DTablePtr = DTableSpace;
             return headerSize;
         }
-    default :   /* impossible */
+    default :
         assert(0);
-        return ERROR(GENERIC);
+        RETURN_ERROR(GENERIC, "impossible");
     }
 }
 
@@ -483,28 +487,28 b' size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* '
     DEBUGLOG(5, "ZSTD_decodeSeqHeaders");
 
     /* check */
-    if (srcSize < MIN_SEQUENCES_SIZE) return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(srcSize < MIN_SEQUENCES_SIZE, srcSize_wrong);
 
     /* SeqHead */
     nbSeq = *ip++;
     if (!nbSeq) {
         *nbSeqPtr=0;
-        if (srcSize != 1) return ERROR(srcSize_wrong);
+        RETURN_ERROR_IF(srcSize != 1, srcSize_wrong);
         return 1;
     }
     if (nbSeq > 0x7F) {
         if (nbSeq == 0xFF) {
-            if (ip+2 > iend) return ERROR(srcSize_wrong);
+            RETURN_ERROR_IF(ip+2 > iend, srcSize_wrong);
             nbSeq = MEM_readLE16(ip) + LONGNBSEQ, ip+=2;
         } else {
-            if (ip >= iend) return ERROR(srcSize_wrong);
+            RETURN_ERROR_IF(ip >= iend, srcSize_wrong);
             nbSeq = ((nbSeq-0x80)<<8) + *ip++;
         }
     }
     *nbSeqPtr = nbSeq;
 
     /* FSE table descriptors */
-    if (ip+4 > iend) return ERROR(srcSize_wrong); /* minimum possible size */
+    RETURN_ERROR_IF(ip+1 > iend, srcSize_wrong); /* minimum possible size: 1 byte for symbol encoding types */
     {   symbolEncodingType_e const LLtype = (symbolEncodingType_e)(*ip >> 6);
         symbolEncodingType_e const OFtype = (symbolEncodingType_e)((*ip >> 4) & 3);
         symbolEncodingType_e const MLtype = (symbolEncodingType_e)((*ip >> 2) & 3);
@@ -517,7 +521,7 b' size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* '
                                                       LL_base, LL_bits,
                                                       LL_defaultDTable, dctx->fseEntropy,
                                                       dctx->ddictIsCold, nbSeq);
-            if (ZSTD_isError(llhSize)) return ERROR(corruption_detected);
+            RETURN_ERROR_IF(ZSTD_isError(llhSize), corruption_detected);
             ip += llhSize;
         }
 
@@ -527,7 +531,7 b' size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* '
                                                       OF_base, OF_bits,
                                                       OF_defaultDTable, dctx->fseEntropy,
                                                       dctx->ddictIsCold, nbSeq);
-            if (ZSTD_isError(ofhSize)) return ERROR(corruption_detected);
+            RETURN_ERROR_IF(ZSTD_isError(ofhSize), corruption_detected);
             ip += ofhSize;
         }
 
@@ -537,7 +541,7 b' size_t ZSTD_decodeSeqHeaders(ZSTD_DCtx* '
                                                       ML_base, ML_bits,
                                                       ML_defaultDTable, dctx->fseEntropy,
                                                       dctx->ddictIsCold, nbSeq);
-            if (ZSTD_isError(mlhSize)) return ERROR(corruption_detected);
+            RETURN_ERROR_IF(ZSTD_isError(mlhSize), corruption_detected);
             ip += mlhSize;
         }
     }
@@ -590,8 +594,8 b' size_t ZSTD_execSequenceLast7(BYTE* op,'
     const BYTE* match = oLitEnd - sequence.offset;
 
     /* check */
-    if (oMatchEnd>oend) return ERROR(dstSize_tooSmall);   /* last match must fit within dstBuffer */
-    if (iLitEnd > litLimit) return ERROR(corruption_detected);   /* try to read beyond literal buffer */
+    RETURN_ERROR_IF(oMatchEnd>oend, dstSize_tooSmall, "last match must fit within dstBuffer");
+    RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "try to read beyond literal buffer");
 
     /* copy literals */
     while (op < oLitEnd) *op++ = *(*litPtr)++;
@@ -599,7 +603,7 b' size_t ZSTD_execSequenceLast7(BYTE* op,'
     /* copy Match */
     if (sequence.offset > (size_t)(oLitEnd - base)) {
         /* offset beyond prefix */
-        if (sequence.offset > (size_t)(oLitEnd - vBase)) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - vBase),corruption_detected);
         match = dictEnd - (base-match);
         if (match + sequence.matchLength <= dictEnd) {
             memmove(oLitEnd, match, sequence.matchLength);
@@ -631,22 +635,22 b' size_t ZSTD_execSequence(BYTE* op,'
     const BYTE* match = oLitEnd - sequence.offset;
 
     /* check */
-    if (oMatchEnd>oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-    if (iLitEnd > litLimit) return ERROR(corruption_detected);   /* over-read beyond lit buffer */
+    RETURN_ERROR_IF(oMatchEnd>oend, dstSize_tooSmall, "last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend");
+    RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "over-read beyond lit buffer");
     if (oLitEnd>oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, virtualStart, dictEnd);
 
     /* copy Literals */
-    ZSTD_copy8(op, *litPtr);
     if (sequence.litLength > 8)
-        ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8);   /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
+        ZSTD_wildcopy_16min(op, (*litPtr), sequence.litLength, ZSTD_no_overlap);   /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
+    else
+        ZSTD_copy8(op, *litPtr);
     op = oLitEnd;
     *litPtr = iLitEnd;   /* update for next sequence */
 
     /* copy Match */
     if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
         /* offset beyond prefix -> go into extDict */
-        if (sequence.offset > (size_t)(oLitEnd - virtualStart))
-            return ERROR(corruption_detected);
+        RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - virtualStart), corruption_detected);
         match = dictEnd + (match - prefixStart);
         if (match + sequence.matchLength <= dictEnd) {
             memmove(oLitEnd, match, sequence.matchLength);
@@ -686,13 +690,13 b' size_t ZSTD_execSequence(BYTE* op,'
 
     if (oMatchEnd > oend-(16-MINMATCH)) {
         if (op < oend_w) {
-            ZSTD_wildcopy(op, match, oend_w - op);
+            ZSTD_wildcopy(op, match, oend_w - op, ZSTD_overlap_src_before_dst);
             match += oend_w - op;
             op = oend_w;
         }
         while (op < oMatchEnd) *op++ = *match++;
     } else {
-        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8);   /* works even if matchLength < 8 */
+        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8, ZSTD_overlap_src_before_dst);   /* works even if matchLength < 8 */
     }
     return sequenceLength;
 }
@@ -712,21 +716,23 b' size_t ZSTD_execSequenceLong(BYTE* op,'
     const BYTE* match = sequence.match;
 
     /* check */
-    if (oMatchEnd > oend) return ERROR(dstSize_tooSmall); /* last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend */
-    if (iLitEnd > litLimit) return ERROR(corruption_detected);   /* over-read beyond lit buffer */
+    RETURN_ERROR_IF(oMatchEnd > oend, dstSize_tooSmall, "last match must start at a minimum distance of WILDCOPY_OVERLENGTH from oend");
+    RETURN_ERROR_IF(iLitEnd > litLimit, corruption_detected, "over-read beyond lit buffer");
     if (oLitEnd > oend_w) return ZSTD_execSequenceLast7(op, oend, sequence, litPtr, litLimit, prefixStart, dictStart, dictEnd);
 
     /* copy Literals */
-    ZSTD_copy8(op, *litPtr);  /* note : op <= oLitEnd <= oend_w == oend - 8 */
     if (sequence.litLength > 8)
-        ZSTD_wildcopy(op+8, (*litPtr)+8, sequence.litLength - 8);   /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
+        ZSTD_wildcopy_16min(op, *litPtr, sequence.litLength, ZSTD_no_overlap);   /* note : since oLitEnd <= oend-WILDCOPY_OVERLENGTH, no risk of overwrite beyond oend */
+    else
+        ZSTD_copy8(op, *litPtr);  /* note : op <= oLitEnd <= oend_w == oend - 8 */
+
     op = oLitEnd;
     *litPtr = iLitEnd;   /* update for next sequence */
 
     /* copy Match */
     if (sequence.offset > (size_t)(oLitEnd - prefixStart)) {
         /* offset beyond prefix */
-        if (sequence.offset > (size_t)(oLitEnd - dictStart)) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(sequence.offset > (size_t)(oLitEnd - dictStart), corruption_detected);
         if (match + sequence.matchLength <= dictEnd) {
             memmove(oLitEnd, match, sequence.matchLength);
             return sequenceLength;
@@ -766,13 +772,13 b' size_t ZSTD_execSequenceLong(BYTE* op,'
 
     if (oMatchEnd > oend-(16-MINMATCH)) {
         if (op < oend_w) {
-            ZSTD_wildcopy(op, match, oend_w - op);
+            ZSTD_wildcopy(op, match, oend_w - op, ZSTD_overlap_src_before_dst);
             match += oend_w - op;
             op = oend_w;
         }
         while (op < oMatchEnd) *op++ = *match++;
     } else {
-        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8);   /* works even if matchLength < 8 */
+        ZSTD_wildcopy(op, match, (ptrdiff_t)sequence.matchLength-8, ZSTD_overlap_src_before_dst);   /* works even if matchLength < 8 */
     }
     return sequenceLength;
 }
@@ -801,7 +807,7 b' ZSTD_updateFseState(ZSTD_fseState* DStat'
 /* We need to add at most (ZSTD_WINDOWLOG_MAX_32 - 1) bits to read the maximum
  * offset bits. But we can only read at most (STREAM_ACCUMULATOR_MIN_32 - 1)
  * bits before reloading. This value is the maximum number of bytes we read
- * after reloading when we are decoding long offets.
+ * after reloading when we are decoding long offsets.
  */
 #define LONG_OFFSETS_MAX_EXTRA_BITS_32                       \
     (ZSTD_WINDOWLOG_MAX_32 > STREAM_ACCUMULATOR_MIN_32       \
@@ -889,6 +895,7 b' ZSTD_decodeSequence(seqState_t* seqState'
 }
 
 FORCE_INLINE_TEMPLATE size_t
+DONT_VECTORIZE
 ZSTD_decompressSequences_body( ZSTD_DCtx* dctx,
                                void* dst, size_t maxDstSize,
                          const void* seqStart, size_t seqSize, int nbSeq,
@@ -911,11 +918,18 b' ZSTD_decompressSequences_body( ZSTD_DCtx'
         seqState_t seqState;
         dctx->fseEntropy = 1;
         { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) seqState.prevOffset[i] = dctx->entropy.rep[i]; }
-        CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
+        RETURN_ERROR_IF(
+            ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),
+            corruption_detected);
         ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
         ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
         ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
 
+        ZSTD_STATIC_ASSERT(
+                BIT_DStream_unfinished < BIT_DStream_completed &&
+                BIT_DStream_endOfBuffer < BIT_DStream_completed &&
+                BIT_DStream_completed < BIT_DStream_overflow);
+
         for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && nbSeq ; ) {
             nbSeq--;
             {   seq_t const sequence = ZSTD_decodeSequence(&seqState, isLongOffset);
@@ -927,14 +941,15 b' ZSTD_decompressSequences_body( ZSTD_DCtx'
 
         /* check if reached exact end */
         DEBUGLOG(5, "ZSTD_decompressSequences_body: after decode loop, remaining nbSeq : %i", nbSeq);
-        if (nbSeq) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(nbSeq, corruption_detected);
+        RETURN_ERROR_IF(BIT_reloadDStream(&seqState.DStream) < BIT_DStream_completed, corruption_detected);
         /* save reps for next block */
         { U32 i; for (i=0; i<ZSTD_REP_NUM; i++) dctx->entropy.rep[i] = (U32)(seqState.prevOffset[i]); }
     }
 
     /* last literal segment */
     {   size_t const lastLLSize = litEnd - litPtr;
-        if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
+        RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);
         memcpy(op, litPtr, lastLLSize);
         op += lastLLSize;
     }
@@ -1066,7 +1081,9 b' ZSTD_decompressSequencesLong_body('
         seqState.pos = (size_t)(op-prefixStart);
         seqState.dictEnd = dictEnd;
         assert(iend >= ip);
-        CHECK_E(BIT_initDStream(&seqState.DStream, ip, iend-ip), corruption_detected);
+        RETURN_ERROR_IF(
+            ERR_isError(BIT_initDStream(&seqState.DStream, ip, iend-ip)),
+            corruption_detected);
         ZSTD_initFseState(&seqState.stateLL, &seqState.DStream, dctx->LLTptr);
         ZSTD_initFseState(&seqState.stateOffb, &seqState.DStream, dctx->OFTptr);
         ZSTD_initFseState(&seqState.stateML, &seqState.DStream, dctx->MLTptr);
@@ -1076,7 +1093,7 b' ZSTD_decompressSequencesLong_body('
             sequences[seqNb] = ZSTD_decodeSequenceLong(&seqState, isLongOffset);
             PREFETCH_L1(sequences[seqNb].match); PREFETCH_L1(sequences[seqNb].match + sequences[seqNb].matchLength - 1); /* note : it's safe to invoke PREFETCH() on any memory address, including invalid ones */
         }
-        if (seqNb<seqAdvance) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(seqNb<seqAdvance, corruption_detected);
 
         /* decode and decompress */
         for ( ; (BIT_reloadDStream(&(seqState.DStream)) <= BIT_DStream_completed) && (seqNb<nbSeq) ; seqNb++) {
@@ -1087,7 +1104,7 b' ZSTD_decompressSequencesLong_body('
             sequences[seqNb & STORED_SEQS_MASK] = sequence;
             op += oneSeqSize;
         }
-        if (seqNb<nbSeq) return ERROR(corruption_detected);
+        RETURN_ERROR_IF(seqNb<nbSeq, corruption_detected);
 
         /* finish queue */
         seqNb -= seqAdvance;
@@ -1103,7 +1120,7 b' ZSTD_decompressSequencesLong_body('
 
     /* last literal segment */
     {   size_t const lastLLSize = litEnd - litPtr;
-        if (lastLLSize > (size_t)(oend-op)) return ERROR(dstSize_tooSmall);
+        RETURN_ERROR_IF(lastLLSize > (size_t)(oend-op), dstSize_tooSmall);
         memcpy(op, litPtr, lastLLSize);
         op += lastLLSize;
     }
@@ -1127,6 +1144,7 b' ZSTD_decompressSequencesLong_default(ZST'
 
 #ifndef ZSTD_FORCE_DECOMPRESS_SEQUENCES_LONG
 static TARGET_ATTRIBUTE("bmi2") size_t
+DONT_VECTORIZE
 ZSTD_decompressSequences_bmi2(ZSTD_DCtx* dctx,
                                  void* dst, size_t maxDstSize,
                            const void* seqStart, size_t seqSize, int nbSeq,
@@ -1176,7 +1194,7 b' ZSTD_decompressSequences(ZSTD_DCtx* dctx'
 /* ZSTD_decompressSequencesLong() :
  * decompression function triggered when a minimum share of offsets is considered "long",
  * aka out of cache.
- * note : "long" definition seems overloaded here, sometimes meaning "wider than bitstream register", and sometimes mearning "farther than memory cache distance".
+ * note : "long" definition seems overloaded here, sometimes meaning "wider than bitstream register", and sometimes meaning "farther than memory cache distance".
  * This function will try to mitigate main memory latency through the use of prefetching */
 static size_t
 ZSTD_decompressSequencesLong(ZSTD_DCtx* dctx,
@@ -1240,7 +1258,7 b' ZSTD_decompressBlock_internal(ZSTD_DCtx*'
     ZSTD_longOffset_e const isLongOffset = (ZSTD_longOffset_e)(MEM_32bits() && (!frame || (dctx->fParams.windowSize > (1ULL << STREAM_ACCUMULATOR_MIN))));
     DEBUGLOG(5, "ZSTD_decompressBlock_internal (size : %u)", (U32)srcSize);
 
-    if (srcSize >= ZSTD_BLOCKSIZE_MAX) return ERROR(srcSize_wrong);
+    RETURN_ERROR_IF(srcSize >= ZSTD_BLOCKSIZE_MAX, srcSize_wrong);
 
     /* Decode literals section */
     {   size_t const litCSize = ZSTD_decodeLiteralsBlock(dctx, src, srcSize);

contrib/python-zstandard/zstd/decompress/zstd_decompress_internal.h

@@ -89,6 +89,12 b' typedef enum { ZSTDds_getFrameHeaderSize'
 typedef enum { zdss_init=0, zdss_loadHeader,
                zdss_read, zdss_load, zdss_flush } ZSTD_dStreamStage;
 
+typedef enum {
+    ZSTD_use_indefinitely = -1,  /* Use the dictionary indefinitely */
+    ZSTD_dont_use = 0,           /* Do not use the dictionary (if one exists free it) */
+    ZSTD_use_once = 1            /* Use the dictionary once and set to ZSTD_dont_use */
+} ZSTD_dictUses_e;
+
 struct ZSTD_DCtx_s
 {
     const ZSTD_seqSymbol* LLTptr;
@@ -123,6 +129,7 b' struct ZSTD_DCtx_s'
     const ZSTD_DDict* ddict;     /* set by ZSTD_initDStream_usingDDict(), or ZSTD_DCtx_refDDict() */
     U32 dictID;
     int ddictIsCold;             /* if == 1 : dictionary is "new" for working context, and presumed "cold" (not in cpu cache) */
+    ZSTD_dictUses_e dictUses;
 
     /* streaming */
     ZSTD_dStreamStage streamStage;

contrib/python-zstandard/zstd/dictBuilder/cover.c

@@ -391,7 +391,7 b' static void COVER_group(COVER_ctx_t *ctx'
  *
  *     Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})
  *
- * Once the dmer d is in the dictionay we set F(d) = 0.
+ * Once the dmer d is in the dictionary we set F(d) = 0.
  */
 static COVER_segment_t COVER_selectSegment(const COVER_ctx_t *ctx, U32 *freqs,
                                            COVER_map_t *activeDmers, U32 begin,
@@ -435,7 +435,7 b' static COVER_segment_t COVER_selectSegme'
       U32 *delDmerOcc = COVER_map_at(activeDmers, delDmer);
       activeSegment.begin += 1;
       *delDmerOcc -= 1;
-      /* If this is the last occurence of the dmer, subtract its score */
+      /* If this is the last occurrence of the dmer, subtract its score */
       if (*delDmerOcc == 0) {
         COVER_map_remove(activeDmers, delDmer);
         activeSegment.score -= freqs[delDmer];
@@ -526,10 +526,10 b' static void COVER_ctx_destroy(COVER_ctx_'
  * Prepare a context for dictionary building.
  * The context is only dependent on the parameter `d` and can used multiple
  * times.
- * Returns 1 on success or zero on error.
+ * Returns 0 on success or error code on error.
  * The context must be destroyed with `COVER_ctx_destroy()`.
  */
-static int COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
+static size_t COVER_ctx_init(COVER_ctx_t *ctx, const void *samplesBuffer,
                           const size_t *samplesSizes, unsigned nbSamples,
                           unsigned d, double splitPoint) {
   const BYTE *const samples = (const BYTE *)samplesBuffer;
@@ -544,17 +544,17 b' static int COVER_ctx_init(COVER_ctx_t *c'
       totalSamplesSize >= (size_t)COVER_MAX_SAMPLES_SIZE) {
     DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",
                  (unsigned)(totalSamplesSize>>20), (COVER_MAX_SAMPLES_SIZE >> 20));
-    return 0;
+    return ERROR(srcSize_wrong);
   }
   /* Check if there are at least 5 training samples */
   if (nbTrainSamples < 5) {
     DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid.", nbTrainSamples);
-    return 0;
+    return ERROR(srcSize_wrong);
   }
   /* Check if there's testing sample */
   if (nbTestSamples < 1) {
     DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.", nbTestSamples);
-    return 0;
+    return ERROR(srcSize_wrong);
   }
   /* Zero the context */
   memset(ctx, 0, sizeof(*ctx));
@@ -577,7 +577,7 b' static int COVER_ctx_init(COVER_ctx_t *c'
   if (!ctx->suffix || !ctx->dmerAt || !ctx->offsets) {
     DISPLAYLEVEL(1, "Failed to allocate scratch buffers\n");
     COVER_ctx_destroy(ctx);
-    return 0;
+    return ERROR(memory_allocation);
   }
   ctx->freqs = NULL;
   ctx->d = d;
@@ -624,7 +624,40 b' static int COVER_ctx_init(COVER_ctx_t *c'
                 (ctx->d <= 8 ? &COVER_cmp8 : &COVER_cmp), &COVER_group);
   ctx->freqs = ctx->suffix;
   ctx->suffix = NULL;
-  return 1;
+  return 0;
+}
+
+void COVER_warnOnSmallCorpus(size_t maxDictSize, size_t nbDmers, int displayLevel)
+{
+  const double ratio = (double)nbDmers / maxDictSize;
+  if (ratio >= 10) {
+      return;
+  }
+  LOCALDISPLAYLEVEL(displayLevel, 1,
+                    "WARNING: The maximum dictionary size %u is too large "
+                    "compared to the source size %u! "
+                    "size(source)/size(dictionary) = %f, but it should be >= "
+                    "10! This may lead to a subpar dictionary! We recommend "
+                    "training on sources at least 10x, and up to 100x the "
+                    "size of the dictionary!\n", (U32)maxDictSize,
+                    (U32)nbDmers, ratio);
+}
+
+COVER_epoch_info_t COVER_computeEpochs(U32 maxDictSize,
+                                       U32 nbDmers, U32 k, U32 passes)
+{
+  const U32 minEpochSize = k * 10;
+  COVER_epoch_info_t epochs;
+  epochs.num = MAX(1, maxDictSize / k / passes);
+  epochs.size = nbDmers / epochs.num;
+  if (epochs.size >= minEpochSize) {
+      assert(epochs.size * epochs.num <= nbDmers);
+      return epochs;
+  }
+  epochs.size = MIN(minEpochSize, nbDmers);
+  epochs.num = nbDmers / epochs.size;
+  assert(epochs.size * epochs.num <= nbDmers);
+  return epochs;
 }
 
 /**
@@ -636,28 +669,34 b' static size_t COVER_buildDictionary(cons'
                                     ZDICT_cover_params_t parameters) {
   BYTE *const dict = (BYTE *)dictBuffer;
   size_t tail = dictBufferCapacity;
-  /* Divide the data up into epochs of equal size.
-   * We will select at least one segment from each epoch.
-   */
-  const unsigned epochs = MAX(1, (U32)(dictBufferCapacity / parameters.k / 4));
-  const unsigned epochSize = (U32)(ctx->suffixSize / epochs);
+  /* Divide the data into epochs. We will select one segment from each epoch. */
+  const COVER_epoch_info_t epochs = COVER_computeEpochs(
+      (U32)dictBufferCapacity, (U32)ctx->suffixSize, parameters.k, 4);
+  const size_t maxZeroScoreRun = MAX(10, MIN(100, epochs.num >> 3));
+  size_t zeroScoreRun = 0;
   size_t epoch;
   DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",
-                epochs, epochSize);
+                (U32)epochs.num, (U32)epochs.size);
   /* Loop through the epochs until there are no more segments or the dictionary
    * is full.
    */
-  for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) {
-    const U32 epochBegin = (U32)(epoch * epochSize);
-    const U32 epochEnd = epochBegin + epochSize;
+  for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {
+    const U32 epochBegin = (U32)(epoch * epochs.size);
+    const U32 epochEnd = epochBegin + epochs.size;
     size_t segmentSize;
     /* Select a segment */
     COVER_segment_t segment = COVER_selectSegment(
         ctx, freqs, activeDmers, epochBegin, epochEnd, parameters);
-    /* If the segment covers no dmers, then we are out of content */
+    /* If the segment covers no dmers, then we are out of content.
+     * There may be new content in other epochs, for continue for some time.
+     */
     if (segment.score == 0) {
-      break;
+      if (++zeroScoreRun >= maxZeroScoreRun) {
+          break;
+      }
+      continue;
     }
+    zeroScoreRun = 0;
     /* Trim the segment if necessary and if it is too small then we are done */
     segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
     if (segmentSize < parameters.d) {
@@ -690,11 +729,11 b' ZDICTLIB_API size_t ZDICT_trainFromBuffe'
   /* Checks */
   if (!COVER_checkParameters(parameters, dictBufferCapacity)) {
     DISPLAYLEVEL(1, "Cover parameters incorrect\n");
-    return ERROR(GENERIC);
+    return ERROR(parameter_outOfBound);
   }
   if (nbSamples == 0) {
     DISPLAYLEVEL(1, "Cover must have at least one input file\n");
-    return ERROR(GENERIC);
+    return ERROR(srcSize_wrong);
   }
   if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
     DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
@@ -702,14 +741,18 b' ZDICTLIB_API size_t ZDICT_trainFromBuffe'
     return ERROR(dstSize_tooSmall);
   }
   /* Initialize context and activeDmers */
-  if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
-                      parameters.d, parameters.splitPoint)) {
-    return ERROR(GENERIC);
+  {
+    size_t const initVal = COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
+                      parameters.d, parameters.splitPoint);
+    if (ZSTD_isError(initVal)) {
+      return initVal;
+    }
   }
+  COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, g_displayLevel);
   if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
     DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
     COVER_ctx_destroy(&ctx);
-    return ERROR(GENERIC);
+    return ERROR(memory_allocation);
   }
 
   DISPLAYLEVEL(2, "Building dictionary\n");
@@ -770,7 +813,7 b' size_t COVER_checkTotalCompressedSize(co'
         cctx, dst, dstCapacity, samples + offsets[i],
         samplesSizes[i], cdict);
     if (ZSTD_isError(size)) {
-      totalCompressedSize = ERROR(GENERIC);
+      totalCompressedSize = size;
       goto _compressCleanup;
     }
     totalCompressedSize += size;
@@ -846,9 +889,11 b' void COVER_best_start(COVER_best_t *best'
  * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
  * If this dictionary is the best so far save it and its parameters.
  */
-void COVER_best_finish(COVER_best_t *best, size_t compressedSize,
-                              ZDICT_cover_params_t parameters, void *dict,
-                              size_t dictSize) {
+void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
+                              COVER_dictSelection_t selection) {
+  void* dict = selection.dictContent;
+  size_t compressedSize = selection.totalCompressedSize;
+  size_t dictSize = selection.dictSize;
   if (!best) {
     return;
   }
@@ -874,10 +919,12 b' void COVER_best_finish(COVER_best_t *bes'
         }
       }
       /* Save the dictionary, parameters, and size */
-      memcpy(best->dict, dict, dictSize);
-      best->dictSize = dictSize;
-      best->parameters = parameters;
-      best->compressedSize = compressedSize;
+      if (dict) {
+        memcpy(best->dict, dict, dictSize);
+        best->dictSize = dictSize;
+        best->parameters = parameters;
+        best->compressedSize = compressedSize;
+      }
     }
     if (liveJobs == 0) {
       ZSTD_pthread_cond_broadcast(&best->cond);
@@ -886,6 +933,111 b' void COVER_best_finish(COVER_best_t *bes'
   }
 }
 
+COVER_dictSelection_t COVER_dictSelectionError(size_t error) {
+    COVER_dictSelection_t selection = { NULL, 0, error };
+    return selection;
+}
+
+unsigned COVER_dictSelectionIsError(COVER_dictSelection_t selection) {
+  return (ZSTD_isError(selection.totalCompressedSize) || !selection.dictContent);
+}
+
+void COVER_dictSelectionFree(COVER_dictSelection_t selection){
+  free(selection.dictContent);
+}
+
+COVER_dictSelection_t COVER_selectDict(BYTE* customDictContent,
+        size_t dictContentSize, const BYTE* samplesBuffer, const size_t* samplesSizes, unsigned nbFinalizeSamples,
+        size_t nbCheckSamples, size_t nbSamples, ZDICT_cover_params_t params, size_t* offsets, size_t totalCompressedSize) {
+
+  size_t largestDict = 0;
+  size_t largestCompressed = 0;
+  BYTE* customDictContentEnd = customDictContent + dictContentSize;
+
+  BYTE * largestDictbuffer = (BYTE *)malloc(dictContentSize);
+  BYTE * candidateDictBuffer = (BYTE *)malloc(dictContentSize);
+  double regressionTolerance = ((double)params.shrinkDictMaxRegression / 100.0) + 1.00;
+
+  if (!largestDictbuffer || !candidateDictBuffer) {
+    free(largestDictbuffer);
+    free(candidateDictBuffer);
+    return COVER_dictSelectionError(dictContentSize);
+  }
+
+  /* Initial dictionary size and compressed size */
+  memcpy(largestDictbuffer, customDictContent, dictContentSize);
+  dictContentSize = ZDICT_finalizeDictionary(
+    largestDictbuffer, dictContentSize, customDictContent, dictContentSize,
+    samplesBuffer, samplesSizes, nbFinalizeSamples, params.zParams);
+
+  if (ZDICT_isError(dictContentSize)) {
+    free(largestDictbuffer);
+    free(candidateDictBuffer);
+    return COVER_dictSelectionError(dictContentSize);
+  }
+
+  totalCompressedSize = COVER_checkTotalCompressedSize(params, samplesSizes,
+                                                       samplesBuffer, offsets,
+                                                       nbCheckSamples, nbSamples,
+                                                       largestDictbuffer, dictContentSize);
+
+  if (ZSTD_isError(totalCompressedSize)) {
+    free(largestDictbuffer);
+    free(candidateDictBuffer);
+    return COVER_dictSelectionError(totalCompressedSize);
+  }
+
+  if (params.shrinkDict == 0) {
+    COVER_dictSelection_t selection = { largestDictbuffer, dictContentSize, totalCompressedSize };
+    free(candidateDictBuffer);
+    return selection;
+  }
+
+  largestDict = dictContentSize;
+  largestCompressed = totalCompressedSize;
+  dictContentSize = ZDICT_DICTSIZE_MIN;
+
+  /* Largest dict is initially at least ZDICT_DICTSIZE_MIN */
+  while (dictContentSize < largestDict) {
+    memcpy(candidateDictBuffer, largestDictbuffer, largestDict);
+    dictContentSize = ZDICT_finalizeDictionary(
+      candidateDictBuffer, dictContentSize, customDictContentEnd - dictContentSize, dictContentSize,
+      samplesBuffer, samplesSizes, nbFinalizeSamples, params.zParams);
+
+    if (ZDICT_isError(dictContentSize)) {
+      free(largestDictbuffer);
+      free(candidateDictBuffer);
+      return COVER_dictSelectionError(dictContentSize);
+
+    }
+
+    totalCompressedSize = COVER_checkTotalCompressedSize(params, samplesSizes,
+                                                         samplesBuffer, offsets,
+                                                         nbCheckSamples, nbSamples,
+                                                         candidateDictBuffer, dictContentSize);
+
+    if (ZSTD_isError(totalCompressedSize)) {
+      free(largestDictbuffer);
+      free(candidateDictBuffer);
+      return COVER_dictSelectionError(totalCompressedSize);
+    }
+
+    if (totalCompressedSize <= largestCompressed * regressionTolerance) {
+      COVER_dictSelection_t selection = { candidateDictBuffer, dictContentSize, totalCompressedSize };
+      free(largestDictbuffer);
+      return selection;
+    }
+    dictContentSize *= 2;
+  }
+  dictContentSize = largestDict;
+  totalCompressedSize = largestCompressed;
+  {
+    COVER_dictSelection_t selection = { largestDictbuffer, dictContentSize, totalCompressedSize };
+    free(candidateDictBuffer);
+    return selection;
+  }
+}
+
 /**
  * Parameters for COVER_tryParameters().
  */
@@ -911,6 +1063,7 b' static void COVER_tryParameters(void *op'
   /* Allocate space for hash table, dict, and freqs */
   COVER_map_t activeDmers;
   BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
+  COVER_dictSelection_t selection = COVER_dictSelectionError(ERROR(GENERIC));
   U32 *freqs = (U32 *)malloc(ctx->suffixSize * sizeof(U32));
   if (!COVER_map_init(&activeDmers, parameters.k - parameters.d + 1)) {
     DISPLAYLEVEL(1, "Failed to allocate dmer map: out of memory\n");
@@ -926,29 +1079,21 b' static void COVER_tryParameters(void *op'
   {
     const size_t tail = COVER_buildDictionary(ctx, freqs, &activeDmers, dict,
                                               dictBufferCapacity, parameters);
-    dictBufferCapacity = ZDICT_finalizeDictionary(
-        dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
-        ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples,
-        parameters.zParams);
-    if (ZDICT_isError(dictBufferCapacity)) {
-      DISPLAYLEVEL(1, "Failed to finalize dictionary\n");
+    selection = COVER_selectDict(dict + tail, dictBufferCapacity - tail,
+        ctx->samples, ctx->samplesSizes, (unsigned)ctx->nbTrainSamples, ctx->nbTrainSamples, ctx->nbSamples, parameters, ctx->offsets,
+        totalCompressedSize);
+
+    if (COVER_dictSelectionIsError(selection)) {
+      DISPLAYLEVEL(1, "Failed to select dictionary\n");
       goto _cleanup;
     }
   }
-  /* Check total compressed size */
-  totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,
-                                                       ctx->samples, ctx->offsets,
-                                                       ctx->nbTrainSamples, ctx->nbSamples,
-                                                       dict, dictBufferCapacity);
-
 _cleanup:
-  COVER_best_finish(data->best, totalCompressedSize, parameters, dict,
-                    dictBufferCapacity);
+  free(dict);
+  COVER_best_finish(data->best, parameters, selection);
   free(data);
   COVER_map_destroy(&activeDmers);
-  if (dict) {
-    free(dict);
-  }
+  COVER_dictSelectionFree(selection);
   if (freqs) {
     free(freqs);
   }
@@ -970,6 +1115,7 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
   const unsigned kStepSize = MAX((kMaxK - kMinK) / kSteps, 1);
   const unsigned kIterations =
       (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
+  const unsigned shrinkDict = 0;
   /* Local variables */
   const int displayLevel = parameters->zParams.notificationLevel;
   unsigned iteration = 1;
@@ -977,19 +1123,20 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
   unsigned k;
   COVER_best_t best;
   POOL_ctx *pool = NULL;
+  int warned = 0;
 
   /* Checks */
   if (splitPoint <= 0 || splitPoint > 1) {
     LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
-    return ERROR(GENERIC);
+    return ERROR(parameter_outOfBound);
   }
   if (kMinK < kMaxD || kMaxK < kMinK) {
     LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect parameters\n");
-    return ERROR(GENERIC);
+    return ERROR(parameter_outOfBound);
   }
   if (nbSamples == 0) {
     DISPLAYLEVEL(1, "Cover must have at least one input file\n");
-    return ERROR(GENERIC);
+    return ERROR(srcSize_wrong);
   }
   if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
     DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
@@ -1013,11 +1160,18 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
     /* Initialize the context for this value of d */
     COVER_ctx_t ctx;
     LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
-    if (!COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint)) {
-      LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
-      COVER_best_destroy(&best);
-      POOL_free(pool);
-      return ERROR(GENERIC);
+    {
+      const size_t initVal = COVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint);
+      if (ZSTD_isError(initVal)) {
+        LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
+        COVER_best_destroy(&best);
+        POOL_free(pool);
+        return initVal;
+      }
+    }
+    if (!warned) {
+      COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.suffixSize, displayLevel);
+      warned = 1;
     }
     /* Loop through k reusing the same context */
     for (k = kMinK; k <= kMaxK; k += kStepSize) {
@@ -1030,7 +1184,7 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
         COVER_best_destroy(&best);
         COVER_ctx_destroy(&ctx);
         POOL_free(pool);
-        return ERROR(GENERIC);
+        return ERROR(memory_allocation);
       }
       data->ctx = &ctx;
       data->best = &best;
@@ -1040,6 +1194,7 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
       data->parameters.d = d;
       data->parameters.splitPoint = splitPoint;
       data->parameters.steps = kSteps;
+      data->parameters.shrinkDict = shrinkDict;
       data->parameters.zParams.notificationLevel = g_displayLevel;
       /* Check the parameters */
       if (!COVER_checkParameters(data->parameters, dictBufferCapacity)) {

contrib/python-zstandard/zstd/dictBuilder/cover.h

@@ -39,6 +39,44 b' typedef struct {'
 } COVER_segment_t;
 
 /**
+ *Number of epochs and size of each epoch.
+ */
+typedef struct {
+  U32 num;
+  U32 size;
+} COVER_epoch_info_t;
+
+/**
+ * Struct used for the dictionary selection function.
+ */
+typedef struct COVER_dictSelection {
+  BYTE* dictContent;
+  size_t dictSize;
+  size_t totalCompressedSize;
+} COVER_dictSelection_t;
+
+/**
+ * Computes the number of epochs and the size of each epoch.
+ * We will make sure that each epoch gets at least 10 * k bytes.
+ *
+ * The COVER algorithms divide the data up into epochs of equal size and
+ * select one segment from each epoch.
+ *
+ * @param maxDictSize The maximum allowed dictionary size.
+ * @param nbDmers     The number of dmers we are training on.
+ * @param k           The parameter k (segment size).
+ * @param passes      The target number of passes over the dmer corpus.
+ *                    More passes means a better dictionary.
+ */
+COVER_epoch_info_t COVER_computeEpochs(U32 maxDictSize, U32 nbDmers,
+                                       U32 k, U32 passes);
+
+/**
+ * Warns the user when their corpus is too small.
+ */
+void COVER_warnOnSmallCorpus(size_t maxDictSize, size_t nbDmers, int displayLevel);
+
+/**
  *  Checks total compressed size of a dictionary
  */
 size_t COVER_checkTotalCompressedSize(const ZDICT_cover_params_t parameters,
@@ -78,6 +116,32 b' void COVER_best_start(COVER_best_t *best'
  * Decrements liveJobs and signals any waiting threads if liveJobs == 0.
  * If this dictionary is the best so far save it and its parameters.
  */
-void COVER_best_finish(COVER_best_t *best, size_t compressedSize,
-                       ZDICT_cover_params_t parameters, void *dict,
-                       size_t dictSize);
+void COVER_best_finish(COVER_best_t *best, ZDICT_cover_params_t parameters,
+                       COVER_dictSelection_t selection);
+/**
+ * Error function for COVER_selectDict function. Checks if the return
+ * value is an error.
+ */
+unsigned COVER_dictSelectionIsError(COVER_dictSelection_t selection);
+
+ /**
+  * Error function for COVER_selectDict function. Returns a struct where
+  * return.totalCompressedSize is a ZSTD error.
+  */
+COVER_dictSelection_t COVER_dictSelectionError(size_t error);
+
+/**
+ * Always call after selectDict is called to free up used memory from
+ * newly created dictionary.
+ */
+void COVER_dictSelectionFree(COVER_dictSelection_t selection);
+
+/**
+ * Called to finalize the dictionary and select one based on whether or not
+ * the shrink-dict flag was enabled. If enabled the dictionary used is the
+ * smallest dictionary within a specified regression of the compressed size
+ * from the largest dictionary.
+ */
+ COVER_dictSelection_t COVER_selectDict(BYTE* customDictContent,
+                       size_t dictContentSize, const BYTE* samplesBuffer, const size_t* samplesSizes, unsigned nbFinalizeSamples,
+                       size_t nbCheckSamples, size_t nbSamples, ZDICT_cover_params_t params, size_t* offsets, size_t totalCompressedSize);

contrib/python-zstandard/zstd/dictBuilder/fastcover.c

@@ -132,7 +132,7 b' typedef struct {'
  *
  *     Score(S) = F(S_1) + F(S_2) + ... + F(S_{k-d+1})
  *
- * Once the dmer with hash value d is in the dictionay we set F(d) = 0.
+ * Once the dmer with hash value d is in the dictionary we set F(d) = 0.
  */
 static COVER_segment_t FASTCOVER_selectSegment(const FASTCOVER_ctx_t *ctx,
                                               U32 *freqs, U32 begin, U32 end,
@@ -161,7 +161,7 b' static COVER_segment_t FASTCOVER_selectS'
     /* Get hash value of current dmer */
     const size_t idx = FASTCOVER_hashPtrToIndex(ctx->samples + activeSegment.end, f, d);
 
-    /* Add frequency of this index to score if this is the first occurence of index in active segment */
+    /* Add frequency of this index to score if this is the first occurrence of index in active segment */
     if (segmentFreqs[idx] == 0) {
       activeSegment.score += freqs[idx];
     }
@@ -287,10 +287,10 b' FASTCOVER_computeFrequency(U32* freqs, c'
  * Prepare a context for dictionary building.
  * The context is only dependent on the parameter `d` and can used multiple
  * times.
- * Returns 1 on success or zero on error.
+ * Returns 0 on success or error code on error.
  * The context must be destroyed with `FASTCOVER_ctx_destroy()`.
  */
-static int
+static size_t
 FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,
                    const void* samplesBuffer,
                    const size_t* samplesSizes, unsigned nbSamples,
@@ -310,19 +310,19 b' FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,'
         totalSamplesSize >= (size_t)FASTCOVER_MAX_SAMPLES_SIZE) {
         DISPLAYLEVEL(1, "Total samples size is too large (%u MB), maximum size is %u MB\n",
                     (unsigned)(totalSamplesSize >> 20), (FASTCOVER_MAX_SAMPLES_SIZE >> 20));
-        return 0;
+        return ERROR(srcSize_wrong);
     }
 
     /* Check if there are at least 5 training samples */
     if (nbTrainSamples < 5) {
         DISPLAYLEVEL(1, "Total number of training samples is %u and is invalid\n", nbTrainSamples);
-        return 0;
+        return ERROR(srcSize_wrong);
     }
 
     /* Check if there's testing sample */
     if (nbTestSamples < 1) {
         DISPLAYLEVEL(1, "Total number of testing samples is %u and is invalid.\n", nbTestSamples);
-        return 0;
+        return ERROR(srcSize_wrong);
     }
 
     /* Zero the context */
@@ -347,7 +347,7 b' FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,'
     if (ctx->offsets == NULL) {
         DISPLAYLEVEL(1, "Failed to allocate scratch buffers \n");
         FASTCOVER_ctx_destroy(ctx);
-        return 0;
+        return ERROR(memory_allocation);
     }
 
     /* Fill offsets from the samplesSizes */
@@ -364,13 +364,13 b' FASTCOVER_ctx_init(FASTCOVER_ctx_t* ctx,'
     if (ctx->freqs == NULL) {
         DISPLAYLEVEL(1, "Failed to allocate frequency table \n");
         FASTCOVER_ctx_destroy(ctx);
-        return 0;
+        return ERROR(memory_allocation);
     }
 
     DISPLAYLEVEL(2, "Computing frequencies\n");
     FASTCOVER_computeFrequency(ctx->freqs, ctx);
 
-    return 1;
+    return 0;
 }
 
 
@@ -386,29 +386,35 b' FASTCOVER_buildDictionary(const FASTCOVE'
 {
   BYTE *const dict = (BYTE *)dictBuffer;
   size_t tail = dictBufferCapacity;
-  /* Divide the data up into epochs of equal size.
-   * We will select at least one segment from each epoch.
-   */
-  const unsigned epochs = MAX(1, (U32)(dictBufferCapacity / parameters.k));
-  const unsigned epochSize = (U32)(ctx->nbDmers / epochs);
+  /* Divide the data into epochs. We will select one segment from each epoch. */
+  const COVER_epoch_info_t epochs = COVER_computeEpochs(
+      (U32)dictBufferCapacity, (U32)ctx->nbDmers, parameters.k, 1);
+  const size_t maxZeroScoreRun = 10;
+  size_t zeroScoreRun = 0;
   size_t epoch;
   DISPLAYLEVEL(2, "Breaking content into %u epochs of size %u\n",
-                epochs, epochSize);
+                (U32)epochs.num, (U32)epochs.size);
   /* Loop through the epochs until there are no more segments or the dictionary
    * is full.
    */
-  for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs) {
-    const U32 epochBegin = (U32)(epoch * epochSize);
-    const U32 epochEnd = epochBegin + epochSize;
+  for (epoch = 0; tail > 0; epoch = (epoch + 1) % epochs.num) {
+    const U32 epochBegin = (U32)(epoch * epochs.size);
+    const U32 epochEnd = epochBegin + epochs.size;
     size_t segmentSize;
     /* Select a segment */
     COVER_segment_t segment = FASTCOVER_selectSegment(
         ctx, freqs, epochBegin, epochEnd, parameters, segmentFreqs);
 
-    /* If the segment covers no dmers, then we are out of content */
+    /* If the segment covers no dmers, then we are out of content.
+     * There may be new content in other epochs, for continue for some time.
+     */
     if (segment.score == 0) {
-      break;
+      if (++zeroScoreRun >= maxZeroScoreRun) {
+          break;
+      }
+      continue;
     }
+    zeroScoreRun = 0;
 
     /* Trim the segment if necessary and if it is too small then we are done */
     segmentSize = MIN(segment.end - segment.begin + parameters.d - 1, tail);
@@ -429,7 +435,6 b' FASTCOVER_buildDictionary(const FASTCOVE'
   return tail;
 }
 
-
 /**
  * Parameters for FASTCOVER_tryParameters().
  */
@@ -458,6 +463,7 b' static void FASTCOVER_tryParameters(void'
   U16* segmentFreqs = (U16 *)calloc(((U64)1 << ctx->f), sizeof(U16));
   /* Allocate space for hash table, dict, and freqs */
   BYTE *const dict = (BYTE * const)malloc(dictBufferCapacity);
+  COVER_dictSelection_t selection = COVER_dictSelectionError(ERROR(GENERIC));
   U32 *freqs = (U32*) malloc(((U64)1 << ctx->f) * sizeof(U32));
   if (!segmentFreqs || !dict || !freqs) {
     DISPLAYLEVEL(1, "Failed to allocate buffers: out of memory\n");
@@ -467,27 +473,24 b' static void FASTCOVER_tryParameters(void'
   memcpy(freqs, ctx->freqs, ((U64)1 << ctx->f) * sizeof(U32));
   /* Build the dictionary */
   { const size_t tail = FASTCOVER_buildDictionary(ctx, freqs, dict, dictBufferCapacity,
-                                                  parameters, segmentFreqs);
+                                                    parameters, segmentFreqs);
+
     const unsigned nbFinalizeSamples = (unsigned)(ctx->nbTrainSamples * ctx->accelParams.finalize / 100);
-    dictBufferCapacity = ZDICT_finalizeDictionary(
-        dict, dictBufferCapacity, dict + tail, dictBufferCapacity - tail,
-        ctx->samples, ctx->samplesSizes, nbFinalizeSamples, parameters.zParams);
-    if (ZDICT_isError(dictBufferCapacity)) {
-      DISPLAYLEVEL(1, "Failed to finalize dictionary\n");
+    selection = COVER_selectDict(dict + tail, dictBufferCapacity - tail,
+         ctx->samples, ctx->samplesSizes, nbFinalizeSamples, ctx->nbTrainSamples, ctx->nbSamples, parameters, ctx->offsets,
+         totalCompressedSize);
+
+    if (COVER_dictSelectionIsError(selection)) {
+      DISPLAYLEVEL(1, "Failed to select dictionary\n");
       goto _cleanup;
     }
   }
-  /* Check total compressed size */
-  totalCompressedSize = COVER_checkTotalCompressedSize(parameters, ctx->samplesSizes,
-                                                       ctx->samples, ctx->offsets,
-                                                       ctx->nbTrainSamples, ctx->nbSamples,
-                                                       dict, dictBufferCapacity);
 _cleanup:
-  COVER_best_finish(data->best, totalCompressedSize, parameters, dict,
-                    dictBufferCapacity);
+  free(dict);
+  COVER_best_finish(data->best, parameters, selection);
   free(data);
   free(segmentFreqs);
-  free(dict);
+  COVER_dictSelectionFree(selection);
   free(freqs);
 }
 
@@ -502,6 +505,7 b' FASTCOVER_convertToCoverParams(ZDICT_fas'
     coverParams->nbThreads = fastCoverParams.nbThreads;
     coverParams->splitPoint = fastCoverParams.splitPoint;
     coverParams->zParams = fastCoverParams.zParams;
+    coverParams->shrinkDict = fastCoverParams.shrinkDict;
 }
 
 
@@ -518,6 +522,7 b' FASTCOVER_convertToFastCoverParams(ZDICT'
     fastCoverParams->f = f;
     fastCoverParams->accel = accel;
     fastCoverParams->zParams = coverParams.zParams;
+    fastCoverParams->shrinkDict = coverParams.shrinkDict;
 }
 
 
@@ -544,11 +549,11 b' ZDICT_trainFromBuffer_fastCover(void* di'
     if (!FASTCOVER_checkParameters(coverParams, dictBufferCapacity, parameters.f,
                                    parameters.accel)) {
       DISPLAYLEVEL(1, "FASTCOVER parameters incorrect\n");
-      return ERROR(GENERIC);
+      return ERROR(parameter_outOfBound);
     }
     if (nbSamples == 0) {
       DISPLAYLEVEL(1, "FASTCOVER must have at least one input file\n");
-      return ERROR(GENERIC);
+      return ERROR(srcSize_wrong);
     }
     if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
       DISPLAYLEVEL(1, "dictBufferCapacity must be at least %u\n",
@@ -558,12 +563,16 b' ZDICT_trainFromBuffer_fastCover(void* di'
     /* Assign corresponding FASTCOVER_accel_t to accelParams*/
     accelParams = FASTCOVER_defaultAccelParameters[parameters.accel];
     /* Initialize context */
-    if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
+    {
+      size_t const initVal = FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples,
                             coverParams.d, parameters.splitPoint, parameters.f,
-                            accelParams)) {
-      DISPLAYLEVEL(1, "Failed to initialize context\n");
-      return ERROR(GENERIC);
+                            accelParams);
+      if (ZSTD_isError(initVal)) {
+        DISPLAYLEVEL(1, "Failed to initialize context\n");
+        return initVal;
+      }
     }
+    COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, g_displayLevel);
     /* Build the dictionary */
     DISPLAYLEVEL(2, "Building dictionary\n");
     {
@@ -609,6 +618,7 b' ZDICT_optimizeTrainFromBuffer_fastCover('
         (1 + (kMaxD - kMinD) / 2) * (1 + (kMaxK - kMinK) / kStepSize);
     const unsigned f = parameters->f == 0 ? DEFAULT_F : parameters->f;
     const unsigned accel = parameters->accel == 0 ? DEFAULT_ACCEL : parameters->accel;
+    const unsigned shrinkDict = 0;
     /* Local variables */
     const int displayLevel = parameters->zParams.notificationLevel;
     unsigned iteration = 1;
@@ -616,22 +626,23 b' ZDICT_optimizeTrainFromBuffer_fastCover('
     unsigned k;
     COVER_best_t best;
     POOL_ctx *pool = NULL;
+    int warned = 0;
     /* Checks */
     if (splitPoint <= 0 || splitPoint > 1) {
       LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect splitPoint\n");
-      return ERROR(GENERIC);
+      return ERROR(parameter_outOfBound);
     }
     if (accel == 0 || accel > FASTCOVER_MAX_ACCEL) {
       LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect accel\n");
-      return ERROR(GENERIC);
+      return ERROR(parameter_outOfBound);
     }
     if (kMinK < kMaxD || kMaxK < kMinK) {
       LOCALDISPLAYLEVEL(displayLevel, 1, "Incorrect k\n");
-      return ERROR(GENERIC);
+      return ERROR(parameter_outOfBound);
     }
     if (nbSamples == 0) {
       LOCALDISPLAYLEVEL(displayLevel, 1, "FASTCOVER must have at least one input file\n");
-      return ERROR(GENERIC);
+      return ERROR(srcSize_wrong);
     }
     if (dictBufferCapacity < ZDICT_DICTSIZE_MIN) {
       LOCALDISPLAYLEVEL(displayLevel, 1, "dictBufferCapacity must be at least %u\n",
@@ -658,11 +669,18 b' ZDICT_optimizeTrainFromBuffer_fastCover('
       /* Initialize the context for this value of d */
       FASTCOVER_ctx_t ctx;
       LOCALDISPLAYLEVEL(displayLevel, 3, "d=%u\n", d);
-      if (!FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f, accelParams)) {
-        LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
-        COVER_best_destroy(&best);
-        POOL_free(pool);
-        return ERROR(GENERIC);
+      {
+        size_t const initVal = FASTCOVER_ctx_init(&ctx, samplesBuffer, samplesSizes, nbSamples, d, splitPoint, f, accelParams);
+        if (ZSTD_isError(initVal)) {
+          LOCALDISPLAYLEVEL(displayLevel, 1, "Failed to initialize context\n");
+          COVER_best_destroy(&best);
+          POOL_free(pool);
+          return initVal;
+        }
+      }
+      if (!warned) {
+        COVER_warnOnSmallCorpus(dictBufferCapacity, ctx.nbDmers, displayLevel);
+        warned = 1;
       }
       /* Loop through k reusing the same context */
       for (k = kMinK; k <= kMaxK; k += kStepSize) {
@@ -675,7 +693,7 b' ZDICT_optimizeTrainFromBuffer_fastCover('
           COVER_best_destroy(&best);
           FASTCOVER_ctx_destroy(&ctx);
           POOL_free(pool);
-          return ERROR(GENERIC);
+          return ERROR(memory_allocation);
         }
         data->ctx = &ctx;
         data->best = &best;
@@ -685,6 +703,7 b' ZDICT_optimizeTrainFromBuffer_fastCover('
         data->parameters.d = d;
         data->parameters.splitPoint = splitPoint;
         data->parameters.steps = kSteps;
+        data->parameters.shrinkDict = shrinkDict;
         data->parameters.zParams.notificationLevel = g_displayLevel;
         /* Check the parameters */
         if (!FASTCOVER_checkParameters(data->parameters, dictBufferCapacity,

contrib/python-zstandard/zstd/dictBuilder/zdict.c

@@ -741,7 +741,7 b' static size_t ZDICT_analyzeEntropy(void*'
     /* analyze, build stats, starting with literals */
     {   size_t maxNbBits = HUF_buildCTable (hufTable, countLit, 255, huffLog);
         if (HUF_isError(maxNbBits)) {
-            eSize = ERROR(GENERIC);
+            eSize = maxNbBits;
             DISPLAYLEVEL(1, " HUF_buildCTable error \n");
             goto _cleanup;
         }
@@ -764,7 +764,7 b' static size_t ZDICT_analyzeEntropy(void*'
     total=0; for (u=0; u<=offcodeMax; u++) total+=offcodeCount[u];
     errorCode = FSE_normalizeCount(offcodeNCount, Offlog, offcodeCount, total, offcodeMax);
     if (FSE_isError(errorCode)) {
-        eSize = ERROR(GENERIC);
+        eSize = errorCode;
         DISPLAYLEVEL(1, "FSE_normalizeCount error with offcodeCount \n");
         goto _cleanup;
     }
@@ -773,7 +773,7 b' static size_t ZDICT_analyzeEntropy(void*'
     total=0; for (u=0; u<=MaxML; u++) total+=matchLengthCount[u];
     errorCode = FSE_normalizeCount(matchLengthNCount, mlLog, matchLengthCount, total, MaxML);
     if (FSE_isError(errorCode)) {
-        eSize = ERROR(GENERIC);
+        eSize = errorCode;
         DISPLAYLEVEL(1, "FSE_normalizeCount error with matchLengthCount \n");
         goto _cleanup;
     }
@@ -782,7 +782,7 b' static size_t ZDICT_analyzeEntropy(void*'
     total=0; for (u=0; u<=MaxLL; u++) total+=litLengthCount[u];
     errorCode = FSE_normalizeCount(litLengthNCount, llLog, litLengthCount, total, MaxLL);
     if (FSE_isError(errorCode)) {
-        eSize = ERROR(GENERIC);
+        eSize = errorCode;
         DISPLAYLEVEL(1, "FSE_normalizeCount error with litLengthCount \n");
         goto _cleanup;
     }
@@ -791,7 +791,7 b' static size_t ZDICT_analyzeEntropy(void*'
     /* write result to buffer */
     {   size_t const hhSize = HUF_writeCTable(dstPtr, maxDstSize, hufTable, 255, huffLog);
         if (HUF_isError(hhSize)) {
-            eSize = ERROR(GENERIC);
+            eSize = hhSize;
             DISPLAYLEVEL(1, "HUF_writeCTable error \n");
             goto _cleanup;
         }
@@ -802,7 +802,7 b' static size_t ZDICT_analyzeEntropy(void*'
 
     {   size_t const ohSize = FSE_writeNCount(dstPtr, maxDstSize, offcodeNCount, OFFCODE_MAX, Offlog);
         if (FSE_isError(ohSize)) {
-            eSize = ERROR(GENERIC);
+            eSize = ohSize;
             DISPLAYLEVEL(1, "FSE_writeNCount error with offcodeNCount \n");
             goto _cleanup;
         }
@@ -813,7 +813,7 b' static size_t ZDICT_analyzeEntropy(void*'
 
     {   size_t const mhSize = FSE_writeNCount(dstPtr, maxDstSize, matchLengthNCount, MaxML, mlLog);
         if (FSE_isError(mhSize)) {
-            eSize = ERROR(GENERIC);
+            eSize = mhSize;
             DISPLAYLEVEL(1, "FSE_writeNCount error with matchLengthNCount \n");
             goto _cleanup;
         }
@@ -824,7 +824,7 b' static size_t ZDICT_analyzeEntropy(void*'
 
     {   size_t const lhSize = FSE_writeNCount(dstPtr, maxDstSize, litLengthNCount, MaxLL, llLog);
         if (FSE_isError(lhSize)) {
-            eSize = ERROR(GENERIC);
+            eSize = lhSize;
             DISPLAYLEVEL(1, "FSE_writeNCount error with litlengthNCount \n");
             goto _cleanup;
         }
@@ -834,7 +834,7 b' static size_t ZDICT_analyzeEntropy(void*'
     }
 
     if (maxDstSize<12) {
-        eSize = ERROR(GENERIC);
+        eSize = ERROR(dstSize_tooSmall);
         DISPLAYLEVEL(1, "not enough space to write RepOffsets \n");
         goto _cleanup;
     }

contrib/python-zstandard/zstd/dictBuilder/zdict.h

@@ -46,7 +46,12 b' extern "C" {'
  *  The resulting dictionary will be saved into `dictBuffer`.
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
  *          or an error code, which can be tested with ZDICT_isError().
- *  Note: ZDICT_trainFromBuffer() requires about 9 bytes of memory for each input byte.
+ *  Note:  Dictionary training will fail if there are not enough samples to construct a
+ *         dictionary, or if most of the samples are too small (< 8 bytes being the lower limit).
+ *         If dictionary training fails, you should use zstd without a dictionary, as the dictionary
+ *         would've been ineffective anyways. If you believe your samples would benefit from a dictionary
+ *         please open an issue with details, and we can look into it.
+ *  Note: ZDICT_trainFromBuffer()'s memory usage is about 6 MB.
  *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
  *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
  *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
@@ -89,6 +94,8 b' typedef struct {'
     unsigned steps;              /* Number of steps : Only used for optimization : 0 means default (40) : Higher means more parameters checked */
     unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
     double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (1.0), 1.0 when all samples are used for both training and testing */
+    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
+    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
     ZDICT_params_t zParams;
 } ZDICT_cover_params_t;
 
@@ -100,6 +107,9 b' typedef struct {'
     unsigned nbThreads;          /* Number of threads : constraint: 0 < nbThreads : 1 means single-threaded : Only used for optimization : Ignored if ZSTD_MULTITHREAD is not defined */
     double splitPoint;           /* Percentage of samples used for training: Only used for optimization : the first nbSamples * splitPoint samples will be used to training, the last nbSamples * (1 - splitPoint) samples will be used for testing, 0 means default (0.75), 1.0 when all samples are used for both training and testing */
     unsigned accel;              /* Acceleration level: constraint: 0 < accel <= 10, higher means faster and less accurate, 0 means default(1) */
+    unsigned shrinkDict;         /* Train dictionaries to shrink in size starting from the minimum size and selects the smallest dictionary that is shrinkDictMaxRegression% worse than the largest dictionary. 0 means no shrinking and 1 means shrinking  */
+    unsigned shrinkDictMaxRegression; /* Sets shrinkDictMaxRegression so that a smaller dictionary can be at worse shrinkDictMaxRegression% worse than the max dict size dictionary. */
+
     ZDICT_params_t zParams;
 } ZDICT_fastCover_params_t;
 
@@ -110,6 +120,7 b' typedef struct {'
  *  The resulting dictionary will be saved into `dictBuffer`.
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
  *          or an error code, which can be tested with ZDICT_isError().
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
  *  Note: ZDICT_trainFromBuffer_cover() requires about 9 bytes of memory for each input byte.
  *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
  *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
@@ -133,8 +144,9 b' ZDICTLIB_API size_t ZDICT_trainFromBuffe'
  * If k is non-zero then we don't check multiple values of k, otherwise we check steps values in [50, 2000].
  *
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
- *           or an error code, which can be tested with ZDICT_isError().
- *           On success `*parameters` contains the parameters selected.
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          On success `*parameters` contains the parameters selected.
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
  * Note: ZDICT_optimizeTrainFromBuffer_cover() requires about 8 bytes of memory for each input byte and additionally another 5 bytes of memory for each byte of memory for each thread.
  */
 ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_cover(
@@ -151,7 +163,8 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
  *  The resulting dictionary will be saved into `dictBuffer`.
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
  *          or an error code, which can be tested with ZDICT_isError().
- *  Note: ZDICT_trainFromBuffer_fastCover() requires about 1 bytes of memory for each input byte and additionally another 6 * 2^f bytes of memory .
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ *  Note: ZDICT_trainFromBuffer_fastCover() requires 6 * 2^f bytes of memory.
  *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
  *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
  *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.
@@ -175,9 +188,10 b' ZDICTLIB_API size_t ZDICT_trainFromBuffe'
  * If accel is zero, default value of 1 is used.
  *
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
- *           or an error code, which can be tested with ZDICT_isError().
- *           On success `*parameters` contains the parameters selected.
- * Note: ZDICT_optimizeTrainFromBuffer_fastCover() requires about 1 byte of memory for each input byte and additionally another 6 * 2^f bytes of memory for each thread.
+ *          or an error code, which can be tested with ZDICT_isError().
+ *          On success `*parameters` contains the parameters selected.
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
+ * Note: ZDICT_optimizeTrainFromBuffer_fastCover() requires about 6 * 2^f bytes of memory for each thread.
  */
 ZDICTLIB_API size_t ZDICT_optimizeTrainFromBuffer_fastCover(void* dictBuffer,
                     size_t dictBufferCapacity, const void* samplesBuffer,
@@ -195,7 +209,7 b' ZDICTLIB_API size_t ZDICT_optimizeTrainF'
  * maxDictSize must be >= dictContentSize, and must be >= ZDICT_DICTSIZE_MIN bytes.
  *
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`),
- *           or an error code, which can be tested by ZDICT_isError().
+ *          or an error code, which can be tested by ZDICT_isError().
  * Note: ZDICT_finalizeDictionary() will push notifications into stderr if instructed to, using notificationLevel>0.
  * Note 2: dictBuffer and dictContent can overlap
  */
@@ -219,6 +233,7 b' typedef struct {'
  * `parameters` is optional and can be provided with values set to 0 to mean "default".
  * @return: size of dictionary stored into `dictBuffer` (<= `dictBufferCapacity`)
  *          or an error code, which can be tested with ZDICT_isError().
+ *          See ZDICT_trainFromBuffer() for details on failure modes.
  *  Tips: In general, a reasonable dictionary has a size of ~ 100 KB.
  *        It's possible to select smaller or larger size, just by specifying `dictBufferCapacity`.
  *        In general, it's recommended to provide a few thousands samples, though this can vary a lot.

contrib/python-zstandard/zstd/zstd.h

@@ -70,8 +70,8 b' extern "C" {'
 
 /*------   Version   ------*/
 #define ZSTD_VERSION_MAJOR    1
-#define ZSTD_VERSION_MINOR    3
-#define ZSTD_VERSION_RELEASE  8
+#define ZSTD_VERSION_MINOR    4
+#define ZSTD_VERSION_RELEASE  3
 
 #define ZSTD_VERSION_NUMBER  (ZSTD_VERSION_MAJOR *100*100 + ZSTD_VERSION_MINOR *100 + ZSTD_VERSION_RELEASE)
 ZSTDLIB_API unsigned ZSTD_versionNumber(void);   /**< to check runtime library version */
@@ -82,13 +82,28 b' ZSTDLIB_API unsigned ZSTD_versionNumber('
 #define ZSTD_VERSION_STRING ZSTD_EXPAND_AND_QUOTE(ZSTD_LIB_VERSION)
 ZSTDLIB_API const char* ZSTD_versionString(void);   /* requires v1.3.0+ */
 
-/***************************************
-*  Default constant
-***************************************/
+/* *************************************
+ *  Default constant
+ ***************************************/
 #ifndef ZSTD_CLEVEL_DEFAULT
 #  define ZSTD_CLEVEL_DEFAULT 3
 #endif
 
+/* *************************************
+ *  Constants
+ ***************************************/
+
+/* All magic numbers are supposed read/written to/from files/memory using little-endian convention */
+#define ZSTD_MAGICNUMBER            0xFD2FB528    /* valid since v0.8.0 */
+#define ZSTD_MAGIC_DICTIONARY       0xEC30A437    /* valid since v0.7.0 */
+#define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50    /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */
+#define ZSTD_MAGIC_SKIPPABLE_MASK   0xFFFFFFF0
+
+#define ZSTD_BLOCKSIZELOG_MAX  17
+#define ZSTD_BLOCKSIZE_MAX     (1<<ZSTD_BLOCKSIZELOG_MAX)
+
+
+
 /***************************************
 *  Simple API
 ***************************************/
@@ -145,12 +160,21 b' ZSTDLIB_API unsigned long long ZSTD_getF'
  * @return : decompressed size of `src` frame content _if known and not empty_, 0 otherwise. */
 ZSTDLIB_API unsigned long long ZSTD_getDecompressedSize(const void* src, size_t srcSize);
 
+/*! ZSTD_findFrameCompressedSize() :
+ * `src` should point to the start of a ZSTD frame or skippable frame.
+ * `srcSize` must be >= first frame size
+ * @return : the compressed size of the first frame starting at `src`,
+ *           suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
+ *        or an error code if input is invalid */
+ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
+
 
 /*======  Helper functions  ======*/
 #define ZSTD_COMPRESSBOUND(srcSize)   ((srcSize) + ((srcSize)>>8) + (((srcSize) < (128<<10)) ? (((128<<10) - (srcSize)) >> 11) /* margin, from 64 to 0 */ : 0))  /* this formula ensures that bound(A) + bound(B) <= bound(A+B) as long as A and B >= 128 KB */
 ZSTDLIB_API size_t      ZSTD_compressBound(size_t srcSize); /*!< maximum compressed size in worst case single-pass scenario */
 ZSTDLIB_API unsigned    ZSTD_isError(size_t code);          /*!< tells if a `size_t` function result is an error code */
 ZSTDLIB_API const char* ZSTD_getErrorName(size_t code);     /*!< provides readable string from an error code */
+ZSTDLIB_API int         ZSTD_minCLevel(void);               /*!< minimum negative compression level allowed */
 ZSTDLIB_API int         ZSTD_maxCLevel(void);               /*!< maximum compression level available */
 
 
@@ -159,9 +183,14 b' ZSTDLIB_API int ZSTD_maxCLevel(v'
 ***************************************/
 /*= Compression context
  *  When compressing many times,
- *  it is recommended to allocate a context just once, and re-use it for each successive compression operation.
+ *  it is recommended to allocate a context just once,
+ *  and re-use it for each successive compression operation.
  *  This will make workload friendlier for system's memory.
- *  Use one context per thread for parallel execution in multi-threaded environments. */
+ *  Note : re-using context is just a speed / resource optimization.
+ *         It doesn't change the compression ratio, which remains identical.
+ *  Note 2 : In multi-threaded environments,
+ *         use one different context per thread for parallel execution.
+ */
 typedef struct ZSTD_CCtx_s ZSTD_CCtx;
 ZSTDLIB_API ZSTD_CCtx* ZSTD_createCCtx(void);
 ZSTDLIB_API size_t     ZSTD_freeCCtx(ZSTD_CCtx* cctx);
@@ -195,279 +224,6 b' ZSTDLIB_API size_t ZSTD_decompressDCtx(Z'
                                  const void* src, size_t srcSize);
 
 
-/**************************
-*  Simple dictionary API
-***************************/
-/*! ZSTD_compress_usingDict() :
- *  Compression at an explicit compression level using a Dictionary.
- *  A dictionary can be any arbitrary data segment (also called a prefix),
- *  or a buffer with specified information (see dictBuilder/zdict.h).
- *  Note : This function loads the dictionary, resulting in significant startup delay.
- *         It's intended for a dictionary used only once.
- *  Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */
-ZSTDLIB_API 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_decompress_usingDict() :
- *  Decompression using a known Dictionary.
- *  Dictionary must be identical to the one used during compression.
- *  Note : This function loads the dictionary, resulting in significant startup delay.
- *         It's intended for a dictionary used only once.
- *  Note : When `dict == NULL || dictSize < 8` no dictionary is used. */
-ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
-                                             void* dst, size_t dstCapacity,
-                                       const void* src, size_t srcSize,
-                                       const void* dict,size_t dictSize);
-
-
-/***********************************
- *  Bulk processing dictionary API
- **********************************/
-typedef struct ZSTD_CDict_s ZSTD_CDict;
-
-/*! ZSTD_createCDict() :
- *  When compressing multiple messages / blocks using the same dictionary, it's recommended to load it only once.
- *  ZSTD_createCDict() will create a digested dictionary, ready to start future compression operations without startup cost.
- *  ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
- * `dictBuffer` can be released after ZSTD_CDict creation, because its content is copied within CDict.
- *  Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate `dictBuffer` content.
- *  Note : A ZSTD_CDict can be created from an empty dictBuffer, but it is inefficient when used to compress small data. */
-ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
-                                         int compressionLevel);
-
-/*! ZSTD_freeCDict() :
- *  Function frees memory allocated by ZSTD_createCDict(). */
-ZSTDLIB_API size_t      ZSTD_freeCDict(ZSTD_CDict* CDict);
-
-/*! ZSTD_compress_usingCDict() :
- *  Compression using a digested Dictionary.
- *  Recommended when same dictionary is used multiple times.
- *  Note : compression level is _decided at dictionary creation time_,
- *     and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */
-ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
-                                            void* dst, size_t dstCapacity,
-                                      const void* src, size_t srcSize,
-                                      const ZSTD_CDict* cdict);
-
-
-typedef struct ZSTD_DDict_s ZSTD_DDict;
-
-/*! ZSTD_createDDict() :
- *  Create a digested dictionary, ready to start decompression operation without startup delay.
- *  dictBuffer can be released after DDict creation, as its content is copied inside DDict. */
-ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
-
-/*! ZSTD_freeDDict() :
- *  Function frees memory allocated with ZSTD_createDDict() */
-ZSTDLIB_API size_t      ZSTD_freeDDict(ZSTD_DDict* ddict);
-
-/*! ZSTD_decompress_usingDDict() :
- *  Decompression using a digested Dictionary.
- *  Recommended when same dictionary is used multiple times. */
-ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
-                                              void* dst, size_t dstCapacity,
-                                        const void* src, size_t srcSize,
-                                        const ZSTD_DDict* ddict);
-
-
-/****************************
-*  Streaming
-****************************/
-
-typedef struct ZSTD_inBuffer_s {
-  const void* src;    /**< start of input buffer */
-  size_t size;        /**< size of input buffer */
-  size_t pos;         /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */
-} ZSTD_inBuffer;
-
-typedef struct ZSTD_outBuffer_s {
-  void*  dst;         /**< start of output buffer */
-  size_t size;        /**< size of output buffer */
-  size_t pos;         /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */
-} ZSTD_outBuffer;
-
-
-
-/*-***********************************************************************
-*  Streaming compression - HowTo
-*
-*  A ZSTD_CStream object is required to track streaming operation.
-*  Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
-*  ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
-*  It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
-*
-*  For parallel execution, use one separate ZSTD_CStream per thread.
-*
-*  note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
-*
-*  Parameters are sticky : when starting a new compression on the same context,
-*  it will re-use the same sticky parameters as previous compression session.
-*  When in doubt, it's recommended to fully initialize the context before usage.
-*  Use ZSTD_initCStream() to set the parameter to a selected compression level.
-*  Use advanced API (ZSTD_CCtx_setParameter(), etc.) to set more specific parameters.
-*
-*  Use ZSTD_compressStream() as many times as necessary to consume input stream.
-*  The function will automatically update both `pos` fields within `input` and `output`.
-*  Note that the function may not consume the entire input,
-*  for example, because the output buffer is already full,
-*  in which case `input.pos < input.size`.
-*  The caller must check if input has been entirely consumed.
-*  If not, the caller must make some room to receive more compressed data,
-*  and then present again remaining input data.
-* @return : a size hint, preferred nb of bytes to use as input for next function call
-*           or an error code, which can be tested using ZSTD_isError().
-*           Note 1 : it's just a hint, to help latency a little, any value will work fine.
-*           Note 2 : size hint is guaranteed to be <= ZSTD_CStreamInSize()
-*
-*  At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
-*  using ZSTD_flushStream(). `output->pos` will be updated.
-*  Note that, if `output->size` is too small, a single invocation of ZSTD_flushStream() might not be enough (return code > 0).
-*  In which case, make some room to receive more compressed data, and call again ZSTD_flushStream().
-*  @return : 0 if internal buffers are entirely flushed,
-*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
-*            or an error code, which can be tested using ZSTD_isError().
-*
-*  ZSTD_endStream() instructs to finish a frame.
-*  It will perform a flush and write frame epilogue.
-*  The epilogue is required for decoders to consider a frame completed.
-*  flush() operation is the same, and follows same rules as ZSTD_flushStream().
-*  @return : 0 if frame fully completed and fully flushed,
-*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
-*            or an error code, which can be tested using ZSTD_isError().
-*
-* *******************************************************************/
-
-typedef ZSTD_CCtx ZSTD_CStream;  /**< CCtx and CStream are now effectively same object (>= v1.3.0) */
-                                 /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */
-/*===== ZSTD_CStream management functions =====*/
-ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);
-ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
-
-/*===== Streaming compression functions =====*/
-ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
-ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
-ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
-ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
-
-ZSTDLIB_API size_t ZSTD_CStreamInSize(void);    /**< recommended size for input buffer */
-ZSTDLIB_API size_t ZSTD_CStreamOutSize(void);   /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block in all circumstances. */
-
-
-
-/*-***************************************************************************
-*  Streaming decompression - HowTo
-*
-*  A ZSTD_DStream object is required to track streaming operations.
-*  Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
-*  ZSTD_DStream objects can be re-used multiple times.
-*
-*  Use ZSTD_initDStream() to start a new decompression operation.
-* @return : recommended first input size
-*  Alternatively, use advanced API to set specific properties.
-*
-*  Use ZSTD_decompressStream() repetitively to consume your input.
-*  The function will update both `pos` fields.
-*  If `input.pos < input.size`, some input has not been consumed.
-*  It's up to the caller to present again remaining data.
-*  The function tries to flush all data decoded immediately, respecting output buffer size.
-*  If `output.pos < output.size`, decoder has flushed everything it could.
-*  But if `output.pos == output.size`, there might be some data left within internal buffers.,
-*  In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.
-*  Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.
-* @return : 0 when a frame is completely decoded and fully flushed,
-*        or an error code, which can be tested using ZSTD_isError(),
-*        or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :
-*                                the return value is a suggested next input size (just a hint for better latency)
-*                                that will never request more than the remaining frame size.
-* *******************************************************************************/
-
-typedef ZSTD_DCtx ZSTD_DStream;  /**< DCtx and DStream are now effectively same object (>= v1.3.0) */
-                                 /* For compatibility with versions <= v1.2.0, prefer differentiating them. */
-/*===== ZSTD_DStream management functions =====*/
-ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);
-ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);
-
-/*===== Streaming decompression functions =====*/
-ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);
-ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
-
-ZSTDLIB_API size_t ZSTD_DStreamInSize(void);    /*!< recommended size for input buffer */
-ZSTDLIB_API size_t ZSTD_DStreamOutSize(void);   /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */
-
-#endif  /* ZSTD_H_235446 */
-
-
-
-
-/****************************************************************************************
- *   ADVANCED AND EXPERIMENTAL FUNCTIONS
- ****************************************************************************************
- * The definitions in the following section are considered experimental.
- * They are provided for advanced scenarios.
- * They should never be used with a dynamic library, as prototypes may change in the future.
- * Use them only in association with static linking.
- * ***************************************************************************************/
-
-#if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)
-#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY
-
-
-/****************************************************************************************
- *   Candidate API for promotion to stable status
- ****************************************************************************************
- * The following symbols and constants form the "staging area" :
- * they are considered to join "stable API" by v1.4.0.
- * The proposal is written so that it can be made stable "as is",
- * though it's still possible to suggest improvements.
- * Staging is in fact last chance for changes,
- * the API is locked once reaching "stable" status.
- * ***************************************************************************************/
-
-
-/* ===  Constants   === */
-
-/* all magic numbers are supposed read/written to/from files/memory using little-endian convention */
-#define ZSTD_MAGICNUMBER            0xFD2FB528    /* valid since v0.8.0 */
-#define ZSTD_MAGIC_DICTIONARY       0xEC30A437    /* valid since v0.7.0 */
-#define ZSTD_MAGIC_SKIPPABLE_START  0x184D2A50    /* all 16 values, from 0x184D2A50 to 0x184D2A5F, signal the beginning of a skippable frame */
-#define ZSTD_MAGIC_SKIPPABLE_MASK   0xFFFFFFF0
-
-#define ZSTD_BLOCKSIZELOG_MAX  17
-#define ZSTD_BLOCKSIZE_MAX     (1<<ZSTD_BLOCKSIZELOG_MAX)
-
-
-/* ===   query limits   === */
-
-ZSTDLIB_API int ZSTD_minCLevel(void);  /*!< minimum negative compression level allowed */
-
-
-/* ===   frame size   === */
-
-/*! ZSTD_findFrameCompressedSize() :
- * `src` should point to the start of a ZSTD frame or skippable frame.
- * `srcSize` must be >= first frame size
- * @return : the compressed size of the first frame starting at `src`,
- *           suitable to pass as `srcSize` to `ZSTD_decompress` or similar,
- *        or an error code if input is invalid */
-ZSTDLIB_API size_t ZSTD_findFrameCompressedSize(const void* src, size_t srcSize);
-
-
-/* ===   Memory management   === */
-
-/*! ZSTD_sizeof_*() :
- *  These functions give the _current_ memory usage of selected object.
- *  Note that object memory usage can evolve (increase or decrease) over time. */
-ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
-ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
-ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
-ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
-ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
-ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
-
-
 /***************************************
 *  Advanced compression API
 ***************************************/
@@ -503,7 +259,10 b' typedef enum { ZSTD_fast=1,'
 
 typedef enum {
 
-    /* compression parameters */
+    /* compression parameters
+     * Note: When compressing with a ZSTD_CDict these parameters are superseded
+     * by the parameters used to construct the ZSTD_CDict. See ZSTD_CCtx_refCDict()
+     * for more info (superseded-by-cdict). */
     ZSTD_c_compressionLevel=100, /* Update all compression parameters according to pre-defined cLevel table
                               * Default level is ZSTD_CLEVEL_DEFAULT==3.
                               * Special: value 0 means default, which is controlled by ZSTD_CLEVEL_DEFAULT.
@@ -625,6 +384,8 b' typedef enum {'
      * ZSTD_c_format
      * ZSTD_c_forceMaxWindow
      * ZSTD_c_forceAttachDict
+     * ZSTD_c_literalCompressionMode
+     * ZSTD_c_targetCBlockSize
      * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
      * note : never ever use experimentalParam? names directly;
      *        also, the enums values themselves are unstable and can still change.
@@ -632,10 +393,11 b' typedef enum {'
      ZSTD_c_experimentalParam1=500,
      ZSTD_c_experimentalParam2=10,
      ZSTD_c_experimentalParam3=1000,
-     ZSTD_c_experimentalParam4=1001
+     ZSTD_c_experimentalParam4=1001,
+     ZSTD_c_experimentalParam5=1002,
+     ZSTD_c_experimentalParam6=1003,
 } ZSTD_cParameter;
 
-
 typedef struct {
     size_t error;
     int lowerBound;
@@ -677,10 +439,443 b' ZSTDLIB_API size_t ZSTD_CCtx_setParamete'
  *  Note 3 : Whenever all input data is provided and consumed in a single round,
  *           for example with ZSTD_compress2(),
  *           or invoking immediately ZSTD_compressStream2(,,,ZSTD_e_end),
- *           this value is automatically overriden by srcSize instead.
+ *           this value is automatically overridden by srcSize instead.
  */
 ZSTDLIB_API size_t ZSTD_CCtx_setPledgedSrcSize(ZSTD_CCtx* cctx, unsigned long long pledgedSrcSize);
 
+typedef enum {
+    ZSTD_reset_session_only = 1,
+    ZSTD_reset_parameters = 2,
+    ZSTD_reset_session_and_parameters = 3
+} ZSTD_ResetDirective;
+
+/*! ZSTD_CCtx_reset() :
+ *  There are 2 different things that can be reset, independently or jointly :
+ *  - The session : will stop compressing current frame, and make CCtx ready to start a new one.
+ *                  Useful after an error, or to interrupt any ongoing compression.
+ *                  Any internal data not yet flushed is cancelled.
+ *                  Compression parameters and dictionary remain unchanged.
+ *                  They will be used to compress next frame.
+ *                  Resetting session never fails.
+ *  - The parameters : changes all parameters back to "default".
+ *                  This removes any reference to any dictionary too.
+ *                  Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)
+ *                  otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())
+ *  - Both : similar to resetting the session, followed by resetting parameters.
+ */
+ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
+
+/*! ZSTD_compress2() :
+ *  Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
+ *  ZSTD_compress2() always starts a new frame.
+ *  Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
+ *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
+ *  - The function is always blocking, returns when compression is completed.
+ *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
+ * @return : compressed size written into `dst` (<= `dstCapacity),
+ *           or an error code if it fails (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx,
+                                   void* dst, size_t dstCapacity,
+                             const void* src, size_t srcSize);
+
+
+/***************************************
+*  Advanced decompression API
+***************************************/
+
+/* The advanced API pushes parameters one by one into an existing DCtx context.
+ * Parameters are sticky, and remain valid for all following frames
+ * using the same DCtx context.
+ * It's possible to reset parameters to default values using ZSTD_DCtx_reset().
+ * Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream().
+ *        Therefore, no new decompression function is necessary.
+ */
+
+typedef enum {
+
+    ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which
+                              * the streaming API will refuse to allocate memory buffer
+                              * in order to protect the host from unreasonable memory requirements.
+                              * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
+                              * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT).
+                              * Special: value 0 means "use default maximum windowLog". */
+
+    /* note : additional experimental parameters are also available
+     * within the experimental section of the API.
+     * At the time of this writing, they include :
+     * ZSTD_c_format
+     * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
+     * note : never ever use experimentalParam? names directly
+     */
+     ZSTD_d_experimentalParam1=1000
+
+} ZSTD_dParameter;
+
+/*! ZSTD_dParam_getBounds() :
+ *  All parameters must belong to an interval with lower and upper bounds,
+ *  otherwise they will either trigger an error or be automatically clamped.
+ * @return : a structure, ZSTD_bounds, which contains
+ *         - an error status field, which must be tested using ZSTD_isError()
+ *         - both lower and upper bounds, inclusive
+ */
+ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);
+
+/*! ZSTD_DCtx_setParameter() :
+ *  Set one compression parameter, selected by enum ZSTD_dParameter.
+ *  All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().
+ *  Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
+ *  Setting a parameter is only possible during frame initialization (before starting decompression).
+ * @return : 0, or an error code (which can be tested using ZSTD_isError()).
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);
+
+/*! ZSTD_DCtx_reset() :
+ *  Return a DCtx to clean state.
+ *  Session and parameters can be reset jointly or separately.
+ *  Parameters can only be reset when no active frame is being decompressed.
+ * @return : 0, or an error code, which can be tested with ZSTD_isError()
+ */
+ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);
+
+
+/****************************
+*  Streaming
+****************************/
+
+typedef struct ZSTD_inBuffer_s {
+  const void* src;    /**< start of input buffer */
+  size_t size;        /**< size of input buffer */
+  size_t pos;         /**< position where reading stopped. Will be updated. Necessarily 0 <= pos <= size */
+} ZSTD_inBuffer;
+
+typedef struct ZSTD_outBuffer_s {
+  void*  dst;         /**< start of output buffer */
+  size_t size;        /**< size of output buffer */
+  size_t pos;         /**< position where writing stopped. Will be updated. Necessarily 0 <= pos <= size */
+} ZSTD_outBuffer;
+
+
+
+/*-***********************************************************************
+*  Streaming compression - HowTo
+*
+*  A ZSTD_CStream object is required to track streaming operation.
+*  Use ZSTD_createCStream() and ZSTD_freeCStream() to create/release resources.
+*  ZSTD_CStream objects can be reused multiple times on consecutive compression operations.
+*  It is recommended to re-use ZSTD_CStream since it will play nicer with system's memory, by re-using already allocated memory.
+*
+*  For parallel execution, use one separate ZSTD_CStream per thread.
+*
+*  note : since v1.3.0, ZSTD_CStream and ZSTD_CCtx are the same thing.
+*
+*  Parameters are sticky : when starting a new compression on the same context,
+*  it will re-use the same sticky parameters as previous compression session.
+*  When in doubt, it's recommended to fully initialize the context before usage.
+*  Use ZSTD_CCtx_reset() to reset the context and ZSTD_CCtx_setParameter(),
+*  ZSTD_CCtx_setPledgedSrcSize(), or ZSTD_CCtx_loadDictionary() and friends to
+*  set more specific parameters, the pledged source size, or load a dictionary.
+*
+*  Use ZSTD_compressStream2() with ZSTD_e_continue as many times as necessary to
+*  consume input stream. The function will automatically update both `pos`
+*  fields within `input` and `output`.
+*  Note that the function may not consume the entire input, for example, because
+*  the output buffer is already full, in which case `input.pos < input.size`.
+*  The caller must check if input has been entirely consumed.
+*  If not, the caller must make some room to receive more compressed data,
+*  and then present again remaining input data.
+*  note: ZSTD_e_continue is guaranteed to make some forward progress when called,
+*        but doesn't guarantee maximal forward progress. This is especially relevant
+*        when compressing with multiple threads. The call won't block if it can
+*        consume some input, but if it can't it will wait for some, but not all,
+*        output to be flushed.
+* @return : provides a minimum amount of data remaining to be flushed from internal buffers
+*           or an error code, which can be tested using ZSTD_isError().
+*
+*  At any moment, it's possible to flush whatever data might remain stuck within internal buffer,
+*  using ZSTD_compressStream2() with ZSTD_e_flush. `output->pos` will be updated.
+*  Note that, if `output->size` is too small, a single invocation with ZSTD_e_flush might not be enough (return code > 0).
+*  In which case, make some room to receive more compressed data, and call again ZSTD_compressStream2() with ZSTD_e_flush.
+*  You must continue calling ZSTD_compressStream2() with ZSTD_e_flush until it returns 0, at which point you can change the
+*  operation.
+*  note: ZSTD_e_flush will flush as much output as possible, meaning when compressing with multiple threads, it will
+*        block until the flush is complete or the output buffer is full.
+*  @return : 0 if internal buffers are entirely flushed,
+*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
+*            or an error code, which can be tested using ZSTD_isError().
+*
+*  Calling ZSTD_compressStream2() with ZSTD_e_end instructs to finish a frame.
+*  It will perform a flush and write frame epilogue.
+*  The epilogue is required for decoders to consider a frame completed.
+*  flush operation is the same, and follows same rules as calling ZSTD_compressStream2() with ZSTD_e_flush.
+*  You must continue calling ZSTD_compressStream2() with ZSTD_e_end until it returns 0, at which point you are free to
+*  start a new frame.
+*  note: ZSTD_e_end will flush as much output as possible, meaning when compressing with multiple threads, it will
+*        block until the flush is complete or the output buffer is full.
+*  @return : 0 if frame fully completed and fully flushed,
+*            >0 if some data still present within internal buffer (the value is minimal estimation of remaining size),
+*            or an error code, which can be tested using ZSTD_isError().
+*
+* *******************************************************************/
+
+typedef ZSTD_CCtx ZSTD_CStream;  /**< CCtx and CStream are now effectively same object (>= v1.3.0) */
+                                 /* Continue to distinguish them for compatibility with older versions <= v1.2.0 */
+/*===== ZSTD_CStream management functions =====*/
+ZSTDLIB_API ZSTD_CStream* ZSTD_createCStream(void);
+ZSTDLIB_API size_t ZSTD_freeCStream(ZSTD_CStream* zcs);
+
+/*===== Streaming compression functions =====*/
+typedef enum {
+    ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */
+    ZSTD_e_flush=1,    /* flush any data provided so far,
+                        * it creates (at least) one new block, that can be decoded immediately on reception;
+                        * frame will continue: any future data can still reference previously compressed data, improving compression.
+                        * note : multithreaded compression will block to flush as much output as possible. */
+    ZSTD_e_end=2       /* flush any remaining data _and_ close current frame.
+                        * note that frame is only closed after compressed data is fully flushed (return value == 0).
+                        * After that point, any additional data starts a new frame.
+                        * note : each frame is independent (does not reference any content from previous frame).
+                        : note : multithreaded compression will block to flush as much output as possible. */
+} ZSTD_EndDirective;
+
+/*! ZSTD_compressStream2() :
+ *  Behaves about the same as ZSTD_compressStream, with additional control on end directive.
+ *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
+ *  - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)
+ *  - output->pos must be <= dstCapacity, input->pos must be <= srcSize
+ *  - output->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
+ *  - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.
+ *  - When nbWorkers>=1, function is non-blocking : it just acquires a copy of input, and distributes jobs to internal worker threads, flush whatever is available,
+ *                                                  and then immediately returns, just indicating that there is some data remaining to be flushed.
+ *                                                  The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.
+ *  - Exception : if the first call requests a ZSTD_e_end directive and provides enough dstCapacity, the function delegates to ZSTD_compress2() which is always blocking.
+ *  - @return provides a minimum amount of data remaining to be flushed from internal buffers
+ *            or an error code, which can be tested using ZSTD_isError().
+ *            if @return != 0, flush is not fully completed, there is still some data left within internal buffers.
+ *            This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.
+ *            For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.
+ *  - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),
+ *            only ZSTD_e_end or ZSTD_e_flush operations are allowed.
+ *            Before starting a new compression job, or changing compression parameters,
+ *            it is required to fully flush internal buffers.
+ */
+ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
+                                         ZSTD_outBuffer* output,
+                                         ZSTD_inBuffer* input,
+                                         ZSTD_EndDirective endOp);
+
+
+/* These buffer sizes are softly recommended.
+ * They are not required : ZSTD_compressStream*() happily accepts any buffer size, for both input and output.
+ * Respecting the recommended size just makes it a bit easier for ZSTD_compressStream*(),
+ * reducing the amount of memory shuffling and buffering, resulting in minor performance savings.
+ *
+ * However, note that these recommendations are from the perspective of a C caller program.
+ * If the streaming interface is invoked from some other language,
+ * especially managed ones such as Java or Go, through a foreign function interface such as jni or cgo,
+ * a major performance rule is to reduce crossing such interface to an absolute minimum.
+ * It's not rare that performance ends being spent more into the interface, rather than compression itself.
+ * In which cases, prefer using large buffers, as large as practical,
+ * for both input and output, to reduce the nb of roundtrips.
+ */
+ZSTDLIB_API size_t ZSTD_CStreamInSize(void);    /**< recommended size for input buffer */
+ZSTDLIB_API size_t ZSTD_CStreamOutSize(void);   /**< recommended size for output buffer. Guarantee to successfully flush at least one complete compressed block. */
+
+
+/* *****************************************************************************
+ * This following is a legacy streaming API.
+ * It can be replaced by ZSTD_CCtx_reset() and ZSTD_compressStream2().
+ * It is redundant, but remains fully supported.
+ * Advanced parameters and dictionary compression can only be used through the
+ * new API.
+ ******************************************************************************/
+
+/*!
+ * Equivalent to:
+ *
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ */
+ZSTDLIB_API size_t ZSTD_initCStream(ZSTD_CStream* zcs, int compressionLevel);
+/*!
+ * Alternative for ZSTD_compressStream2(zcs, output, input, ZSTD_e_continue).
+ * NOTE: The return value is different. ZSTD_compressStream() returns a hint for
+ * the next read size (if non-zero and not an error). ZSTD_compressStream2()
+ * returns the minimum nb of bytes left to flush (if non-zero and not an error).
+ */
+ZSTDLIB_API size_t ZSTD_compressStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
+/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_flush). */
+ZSTDLIB_API size_t ZSTD_flushStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
+/*! Equivalent to ZSTD_compressStream2(zcs, output, &emptyInput, ZSTD_e_end). */
+ZSTDLIB_API size_t ZSTD_endStream(ZSTD_CStream* zcs, ZSTD_outBuffer* output);
+
+
+/*-***************************************************************************
+*  Streaming decompression - HowTo
+*
+*  A ZSTD_DStream object is required to track streaming operations.
+*  Use ZSTD_createDStream() and ZSTD_freeDStream() to create/release resources.
+*  ZSTD_DStream objects can be re-used multiple times.
+*
+*  Use ZSTD_initDStream() to start a new decompression operation.
+* @return : recommended first input size
+*  Alternatively, use advanced API to set specific properties.
+*
+*  Use ZSTD_decompressStream() repetitively to consume your input.
+*  The function will update both `pos` fields.
+*  If `input.pos < input.size`, some input has not been consumed.
+*  It's up to the caller to present again remaining data.
+*  The function tries to flush all data decoded immediately, respecting output buffer size.
+*  If `output.pos < output.size`, decoder has flushed everything it could.
+*  But if `output.pos == output.size`, there might be some data left within internal buffers.,
+*  In which case, call ZSTD_decompressStream() again to flush whatever remains in the buffer.
+*  Note : with no additional input provided, amount of data flushed is necessarily <= ZSTD_BLOCKSIZE_MAX.
+* @return : 0 when a frame is completely decoded and fully flushed,
+*        or an error code, which can be tested using ZSTD_isError(),
+*        or any other value > 0, which means there is still some decoding or flushing to do to complete current frame :
+*                                the return value is a suggested next input size (just a hint for better latency)
+*                                that will never request more than the remaining frame size.
+* *******************************************************************************/
+
+typedef ZSTD_DCtx ZSTD_DStream;  /**< DCtx and DStream are now effectively same object (>= v1.3.0) */
+                                 /* For compatibility with versions <= v1.2.0, prefer differentiating them. */
+/*===== ZSTD_DStream management functions =====*/
+ZSTDLIB_API ZSTD_DStream* ZSTD_createDStream(void);
+ZSTDLIB_API size_t ZSTD_freeDStream(ZSTD_DStream* zds);
+
+/*===== Streaming decompression functions =====*/
+
+/* This function is redundant with the advanced API and equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds);
+ *     ZSTD_DCtx_refDDict(zds, NULL);
+ */
+ZSTDLIB_API size_t ZSTD_initDStream(ZSTD_DStream* zds);
+
+ZSTDLIB_API size_t ZSTD_decompressStream(ZSTD_DStream* zds, ZSTD_outBuffer* output, ZSTD_inBuffer* input);
+
+ZSTDLIB_API size_t ZSTD_DStreamInSize(void);    /*!< recommended size for input buffer */
+ZSTDLIB_API size_t ZSTD_DStreamOutSize(void);   /*!< recommended size for output buffer. Guarantee to successfully flush at least one complete block in all circumstances. */
+
+
+/**************************
+*  Simple dictionary API
+***************************/
+/*! ZSTD_compress_usingDict() :
+ *  Compression at an explicit compression level using a Dictionary.
+ *  A dictionary can be any arbitrary data segment (also called a prefix),
+ *  or a buffer with specified information (see dictBuilder/zdict.h).
+ *  Note : This function loads the dictionary, resulting in significant startup delay.
+ *         It's intended for a dictionary used only once.
+ *  Note 2 : When `dict == NULL || dictSize < 8` no dictionary is used. */
+ZSTDLIB_API 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_decompress_usingDict() :
+ *  Decompression using a known Dictionary.
+ *  Dictionary must be identical to the one used during compression.
+ *  Note : This function loads the dictionary, resulting in significant startup delay.
+ *         It's intended for a dictionary used only once.
+ *  Note : When `dict == NULL || dictSize < 8` no dictionary is used. */
+ZSTDLIB_API size_t ZSTD_decompress_usingDict(ZSTD_DCtx* dctx,
+                                             void* dst, size_t dstCapacity,
+                                       const void* src, size_t srcSize,
+                                       const void* dict,size_t dictSize);
+
+
+/***********************************
+ *  Bulk processing dictionary API
+ **********************************/
+typedef struct ZSTD_CDict_s ZSTD_CDict;
+
+/*! ZSTD_createCDict() :
+ *  When compressing multiple messages / blocks using the same dictionary, it's recommended to load it only once.
+ *  ZSTD_createCDict() will create a digested dictionary, ready to start future compression operations without startup cost.
+ *  ZSTD_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
+ * `dictBuffer` can be released after ZSTD_CDict creation, because its content is copied within CDict.
+ *  Consider experimental function `ZSTD_createCDict_byReference()` if you prefer to not duplicate `dictBuffer` content.
+ *  Note : A ZSTD_CDict can be created from an empty dictBuffer, but it is inefficient when used to compress small data. */
+ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict(const void* dictBuffer, size_t dictSize,
+                                         int compressionLevel);
+
+/*! ZSTD_freeCDict() :
+ *  Function frees memory allocated by ZSTD_createCDict(). */
+ZSTDLIB_API size_t      ZSTD_freeCDict(ZSTD_CDict* CDict);
+
+/*! ZSTD_compress_usingCDict() :
+ *  Compression using a digested Dictionary.
+ *  Recommended when same dictionary is used multiple times.
+ *  Note : compression level is _decided at dictionary creation time_,
+ *     and frame parameters are hardcoded (dictID=yes, contentSize=yes, checksum=no) */
+ZSTDLIB_API size_t ZSTD_compress_usingCDict(ZSTD_CCtx* cctx,
+                                            void* dst, size_t dstCapacity,
+                                      const void* src, size_t srcSize,
+                                      const ZSTD_CDict* cdict);
+
+
+typedef struct ZSTD_DDict_s ZSTD_DDict;
+
+/*! ZSTD_createDDict() :
+ *  Create a digested dictionary, ready to start decompression operation without startup delay.
+ *  dictBuffer can be released after DDict creation, as its content is copied inside DDict. */
+ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict(const void* dictBuffer, size_t dictSize);
+
+/*! ZSTD_freeDDict() :
+ *  Function frees memory allocated with ZSTD_createDDict() */
+ZSTDLIB_API size_t      ZSTD_freeDDict(ZSTD_DDict* ddict);
+
+/*! ZSTD_decompress_usingDDict() :
+ *  Decompression using a digested Dictionary.
+ *  Recommended when same dictionary is used multiple times. */
+ZSTDLIB_API size_t ZSTD_decompress_usingDDict(ZSTD_DCtx* dctx,
+                                              void* dst, size_t dstCapacity,
+                                        const void* src, size_t srcSize,
+                                        const ZSTD_DDict* ddict);
+
+
+/********************************
+ *  Dictionary helper functions
+ *******************************/
+
+/*! ZSTD_getDictID_fromDict() :
+ *  Provides the dictID stored within dictionary.
+ *  if @return == 0, the dictionary is not conformant with Zstandard specification.
+ *  It can still be loaded, but as a content-only dictionary. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
+
+/*! ZSTD_getDictID_fromDDict() :
+ *  Provides the dictID of the dictionary loaded into `ddict`.
+ *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
+ *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
+
+/*! ZSTD_getDictID_fromFrame() :
+ *  Provides the dictID required to decompressed the frame stored within `src`.
+ *  If @return == 0, the dictID could not be decoded.
+ *  This could for one of the following reasons :
+ *  - The frame does not require a dictionary to be decoded (most common case).
+ *  - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
+ *    Note : this use case also happens when using a non-conformant dictionary.
+ *  - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
+ *  - This is not a Zstandard frame.
+ *  When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */
+ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
+
+
+/*******************************************************************************
+ * Advanced dictionary and prefix API
+ *
+ * This API allows dictionaries to be used with ZSTD_compress2(),
+ * ZSTD_compressStream2(), and ZSTD_decompress(). Dictionaries are sticky, and
+ * only reset with the context is reset with ZSTD_reset_parameters or
+ * ZSTD_reset_session_and_parameters. Prefixes are single-use.
+ ******************************************************************************/
+
+
 /*! ZSTD_CCtx_loadDictionary() :
  *  Create an internal CDict from `dict` buffer.
  *  Decompression will have to use same dictionary.
@@ -703,7 +898,9 b' ZSTDLIB_API size_t ZSTD_CCtx_loadDiction'
 /*! ZSTD_CCtx_refCDict() :
  *  Reference a prepared dictionary, to be used for all next compressed frames.
  *  Note that compression parameters are enforced from within CDict,
- *  and supercede any compression parameter previously set within CCtx.
+ *  and supersede any compression parameter previously set within CCtx.
+ *  The parameters ignored are labled as "superseded-by-cdict" in the ZSTD_cParameter enum docs.
+ *  The ignored parameters will be used again if the CCtx is returned to no-dictionary mode.
  *  The dictionary will remain valid for future compressed frames using same CCtx.
  * @result : 0, or an error code (which can be tested with ZSTD_isError()).
  *  Special : Referencing a NULL CDict means "return to no-dictionary mode".
@@ -733,136 +930,6 b' ZSTDLIB_API size_t ZSTD_CCtx_refCDict(ZS'
 ZSTDLIB_API size_t ZSTD_CCtx_refPrefix(ZSTD_CCtx* cctx,
                                  const void* prefix, size_t prefixSize);
 
-
-typedef enum {
-    ZSTD_reset_session_only = 1,
-    ZSTD_reset_parameters = 2,
-    ZSTD_reset_session_and_parameters = 3
-} ZSTD_ResetDirective;
-
-/*! ZSTD_CCtx_reset() :
- *  There are 2 different things that can be reset, independently or jointly :
- *  - The session : will stop compressing current frame, and make CCtx ready to start a new one.
- *                  Useful after an error, or to interrupt any ongoing compression.
- *                  Any internal data not yet flushed is cancelled.
- *                  Compression parameters and dictionary remain unchanged.
- *                  They will be used to compress next frame.
- *                  Resetting session never fails.
- *  - The parameters : changes all parameters back to "default".
- *                  This removes any reference to any dictionary too.
- *                  Parameters can only be changed between 2 sessions (i.e. no compression is currently ongoing)
- *                  otherwise the reset fails, and function returns an error value (which can be tested using ZSTD_isError())
- *  - Both : similar to resetting the session, followed by resetting parameters.
- */
-ZSTDLIB_API size_t ZSTD_CCtx_reset(ZSTD_CCtx* cctx, ZSTD_ResetDirective reset);
-
-
-
-/*! ZSTD_compress2() :
- *  Behave the same as ZSTD_compressCCtx(), but compression parameters are set using the advanced API.
- *  ZSTD_compress2() always starts a new frame.
- *  Should cctx hold data from a previously unfinished frame, everything about it is forgotten.
- *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
- *  - The function is always blocking, returns when compression is completed.
- *  Hint : compression runs faster if `dstCapacity` >=  `ZSTD_compressBound(srcSize)`.
- * @return : compressed size written into `dst` (<= `dstCapacity),
- *           or an error code if it fails (which can be tested using ZSTD_isError()).
- */
-ZSTDLIB_API size_t ZSTD_compress2( ZSTD_CCtx* cctx,
-                                   void* dst, size_t dstCapacity,
-                             const void* src, size_t srcSize);
-
-typedef enum {
-    ZSTD_e_continue=0, /* collect more data, encoder decides when to output compressed result, for optimal compression ratio */
-    ZSTD_e_flush=1,    /* flush any data provided so far,
-                        * it creates (at least) one new block, that can be decoded immediately on reception;
-                        * frame will continue: any future data can still reference previously compressed data, improving compression. */
-    ZSTD_e_end=2       /* flush any remaining data _and_ close current frame.
-                        * note that frame is only closed after compressed data is fully flushed (return value == 0).
-                        * After that point, any additional data starts a new frame.
-                        * note : each frame is independent (does not reference any content from previous frame). */
-} ZSTD_EndDirective;
-
-/*! ZSTD_compressStream2() :
- *  Behaves about the same as ZSTD_compressStream, with additional control on end directive.
- *  - Compression parameters are pushed into CCtx before starting compression, using ZSTD_CCtx_set*()
- *  - Compression parameters cannot be changed once compression is started (save a list of exceptions in multi-threading mode)
- *  - outpot->pos must be <= dstCapacity, input->pos must be <= srcSize
- *  - outpot->pos and input->pos will be updated. They are guaranteed to remain below their respective limit.
- *  - When nbWorkers==0 (default), function is blocking : it completes its job before returning to caller.
- *  - When nbWorkers>=1, function is non-blocking : it just acquires a copy of input, and distributes jobs to internal worker threads, flush whatever is available,
- *                                                  and then immediately returns, just indicating that there is some data remaining to be flushed.
- *                                                  The function nonetheless guarantees forward progress : it will return only after it reads or write at least 1+ byte.
- *  - Exception : if the first call requests a ZSTD_e_end directive and provides enough dstCapacity, the function delegates to ZSTD_compress2() which is always blocking.
- *  - @return provides a minimum amount of data remaining to be flushed from internal buffers
- *            or an error code, which can be tested using ZSTD_isError().
- *            if @return != 0, flush is not fully completed, there is still some data left within internal buffers.
- *            This is useful for ZSTD_e_flush, since in this case more flushes are necessary to empty all buffers.
- *            For ZSTD_e_end, @return == 0 when internal buffers are fully flushed and frame is completed.
- *  - after a ZSTD_e_end directive, if internal buffer is not fully flushed (@return != 0),
- *            only ZSTD_e_end or ZSTD_e_flush operations are allowed.
- *            Before starting a new compression job, or changing compression parameters,
- *            it is required to fully flush internal buffers.
- */
-ZSTDLIB_API size_t ZSTD_compressStream2( ZSTD_CCtx* cctx,
-                                         ZSTD_outBuffer* output,
-                                         ZSTD_inBuffer* input,
-                                         ZSTD_EndDirective endOp);
-
-
-
-/* ============================== */
-/*   Advanced decompression API   */
-/* ============================== */
-
-/* The advanced API pushes parameters one by one into an existing DCtx context.
- * Parameters are sticky, and remain valid for all following frames
- * using the same DCtx context.
- * It's possible to reset parameters to default values using ZSTD_DCtx_reset().
- * Note : This API is compatible with existing ZSTD_decompressDCtx() and ZSTD_decompressStream().
- *        Therefore, no new decompression function is necessary.
- */
-
-
-typedef enum {
-
-    ZSTD_d_windowLogMax=100, /* Select a size limit (in power of 2) beyond which
-                              * the streaming API will refuse to allocate memory buffer
-                              * in order to protect the host from unreasonable memory requirements.
-                              * This parameter is only useful in streaming mode, since no internal buffer is allocated in single-pass mode.
-                              * By default, a decompression context accepts window sizes <= (1 << ZSTD_WINDOWLOG_LIMIT_DEFAULT) */
-
-    /* note : additional experimental parameters are also available
-     * within the experimental section of the API.
-     * At the time of this writing, they include :
-     * ZSTD_c_format
-     * Because they are not stable, it's necessary to define ZSTD_STATIC_LINKING_ONLY to access them.
-     * note : never ever use experimentalParam? names directly
-     */
-     ZSTD_d_experimentalParam1=1000
-
-} ZSTD_dParameter;
-
-
-/*! ZSTD_dParam_getBounds() :
- *  All parameters must belong to an interval with lower and upper bounds,
- *  otherwise they will either trigger an error or be automatically clamped.
- * @return : a structure, ZSTD_bounds, which contains
- *         - an error status field, which must be tested using ZSTD_isError()
- *         - both lower and upper bounds, inclusive
- */
-ZSTDLIB_API ZSTD_bounds ZSTD_dParam_getBounds(ZSTD_dParameter dParam);
-
-/*! ZSTD_DCtx_setParameter() :
- *  Set one compression parameter, selected by enum ZSTD_dParameter.
- *  All parameters have valid bounds. Bounds can be queried using ZSTD_dParam_getBounds().
- *  Providing a value beyond bound will either clamp it, or trigger an error (depending on parameter).
- *  Setting a parameter is only possible during frame initialization (before starting decompression).
- * @return : 0, or an error code (which can be tested using ZSTD_isError()).
- */
-ZSTDLIB_API size_t ZSTD_DCtx_setParameter(ZSTD_DCtx* dctx, ZSTD_dParameter param, int value);
-
-
 /*! ZSTD_DCtx_loadDictionary() :
  *  Create an internal DDict from dict buffer,
  *  to be used to decompress next frames.
@@ -910,15 +977,32 b' ZSTDLIB_API size_t ZSTD_DCtx_refDDict(ZS'
 ZSTDLIB_API size_t ZSTD_DCtx_refPrefix(ZSTD_DCtx* dctx,
                                  const void* prefix, size_t prefixSize);
 
-/*! ZSTD_DCtx_reset() :
- *  Return a DCtx to clean state.
- *  Session and parameters can be reset jointly or separately.
- *  Parameters can only be reset when no active frame is being decompressed.
- * @return : 0, or an error code, which can be tested with ZSTD_isError()
- */
-ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_DCtx* dctx, ZSTD_ResetDirective reset);
+/* ===   Memory management   === */
+
+/*! ZSTD_sizeof_*() :
+ *  These functions give the _current_ memory usage of selected object.
+ *  Note that object memory usage can evolve (increase or decrease) over time. */
+ZSTDLIB_API size_t ZSTD_sizeof_CCtx(const ZSTD_CCtx* cctx);
+ZSTDLIB_API size_t ZSTD_sizeof_DCtx(const ZSTD_DCtx* dctx);
+ZSTDLIB_API size_t ZSTD_sizeof_CStream(const ZSTD_CStream* zcs);
+ZSTDLIB_API size_t ZSTD_sizeof_DStream(const ZSTD_DStream* zds);
+ZSTDLIB_API size_t ZSTD_sizeof_CDict(const ZSTD_CDict* cdict);
+ZSTDLIB_API size_t ZSTD_sizeof_DDict(const ZSTD_DDict* ddict);
+
+#endif  /* ZSTD_H_235446 */
 
 
+/* **************************************************************************************
+ *   ADVANCED AND EXPERIMENTAL FUNCTIONS
+ ****************************************************************************************
+ * The definitions in the following section are considered experimental.
+ * They are provided for advanced scenarios.
+ * They should never be used with a dynamic library, as prototypes may change in the future.
+ * Use them only in association with static linking.
+ * ***************************************************************************************/
+
+#if defined(ZSTD_STATIC_LINKING_ONLY) && !defined(ZSTD_H_ZSTD_STATIC_LINKING_ONLY)
+#define ZSTD_H_ZSTD_STATIC_LINKING_ONLY
 
 /****************************************************************************************
  *   experimental API (static linking only)
@@ -962,7 +1046,7 b' ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_'
 #define ZSTD_WINDOWLOG_LIMIT_DEFAULT 27   /* by default, the streaming decoder will refuse any frame
                                            * requiring larger than (1<<ZSTD_WINDOWLOG_LIMIT_DEFAULT) window size,
                                            * to preserve host's memory from unreasonable requirements.
-                                           * This limit can be overriden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,).
+                                           * This limit can be overridden using ZSTD_DCtx_setParameter(,ZSTD_d_windowLogMax,).
                                            * The limit does not apply for one-pass decoders (such as ZSTD_decompress()), since no additional memory is allocated */
 
 
@@ -976,6 +1060,10 b' ZSTDLIB_API size_t ZSTD_DCtx_reset(ZSTD_'
 #define ZSTD_LDM_HASHRATELOG_MIN     0
 #define ZSTD_LDM_HASHRATELOG_MAX (ZSTD_WINDOWLOG_MAX - ZSTD_HASHLOG_MIN)
 
+/* Advanced parameter bounds */
+#define ZSTD_TARGETCBLOCKSIZE_MIN   64
+#define ZSTD_TARGETCBLOCKSIZE_MAX   ZSTD_BLOCKSIZE_MAX
+
 /* internal */
 #define ZSTD_HASHLOG3_MAX           17
 
@@ -1064,15 +1152,24 b' typedef enum {'
     ZSTD_dictForceCopy     = 2, /* Always copy the dictionary. */
 } ZSTD_dictAttachPref_e;
 
+typedef enum {
+  ZSTD_lcm_auto = 0,          /**< Automatically determine the compression mode based on the compression level.
+                               *   Negative compression levels will be uncompressed, and positive compression
+                               *   levels will be compressed. */
+  ZSTD_lcm_huffman = 1,       /**< Always attempt Huffman compression. Uncompressed literals will still be
+                               *   emitted if Huffman compression is not profitable. */
+  ZSTD_lcm_uncompressed = 2,  /**< Always emit uncompressed literals. */
+} ZSTD_literalCompressionMode_e;
+
 
 /***************************************
 *  Frame size functions
 ***************************************/
 
 /*! ZSTD_findDecompressedSize() :
- *  `src` should point the start of a series of ZSTD encoded and/or skippable frames
+ *  `src` should point to the start of a series of ZSTD encoded and/or skippable frames
  *  `srcSize` must be the _exact_ size of this series
- *       (i.e. there should be a frame boundary exactly at `srcSize` bytes after `src`)
+ *       (i.e. there should be a frame boundary at `src + srcSize`)
  *  @return : - decompressed size of all data in all successive frames
  *            - if the decompressed size cannot be determined: ZSTD_CONTENTSIZE_UNKNOWN
  *            - if an error occurred: ZSTD_CONTENTSIZE_ERROR
@@ -1092,6 +1189,21 b' typedef enum {'
  *            however it does mean that all frame data must be present and valid. */
 ZSTDLIB_API unsigned long long ZSTD_findDecompressedSize(const void* src, size_t srcSize);
 
+/*! ZSTD_decompressBound() :
+ *  `src` should point to the start of a series of ZSTD encoded and/or skippable frames
+ *  `srcSize` must be the _exact_ size of this series
+ *       (i.e. there should be a frame boundary at `src + srcSize`)
+ *  @return : - upper-bound for the decompressed size of all data in all successive frames
+ *            - if an error occured: ZSTD_CONTENTSIZE_ERROR
+ *
+ *  note 1  : an error can occur if `src` contains an invalid or incorrectly formatted frame.
+ *  note 2  : the upper-bound is exact when the decompressed size field is available in every ZSTD encoded frame of `src`.
+ *            in this case, `ZSTD_findDecompressedSize` and `ZSTD_decompressBound` return the same value.
+ *  note 3  : when the decompressed size field isn't available, the upper-bound for that frame is calculated by:
+ *              upper-bound = # blocks * min(128 KB, Window_Size)
+ */
+ZSTDLIB_API unsigned long long ZSTD_decompressBound(const void* src, size_t srcSize);
+
 /*! ZSTD_frameHeaderSize() :
  *  srcSize must be >= ZSTD_FRAMEHEADERSIZE_PREFIX.
  * @return : size of the Frame Header,
@@ -1110,7 +1222,7 b' ZSTDLIB_API size_t ZSTD_frameHeaderSize('
  *  It will also consider src size to be arbitrarily "large", which is worst case.
  *  If srcSize is known to always be small, ZSTD_estimateCCtxSize_usingCParams() can provide a tighter estimation.
  *  ZSTD_estimateCCtxSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
- *  ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
+ *  ZSTD_estimateCCtxSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
  *  Note : CCtx size estimation is only correct for single-threaded compression. */
 ZSTDLIB_API size_t ZSTD_estimateCCtxSize(int compressionLevel);
 ZSTDLIB_API size_t ZSTD_estimateCCtxSize_usingCParams(ZSTD_compressionParameters cParams);
@@ -1122,7 +1234,7 b' ZSTDLIB_API size_t ZSTD_estimateDCtxSize'
  *  It will also consider src size to be arbitrarily "large", which is worst case.
  *  If srcSize is known to always be small, ZSTD_estimateCStreamSize_usingCParams() can provide a tighter estimation.
  *  ZSTD_estimateCStreamSize_usingCParams() can be used in tandem with ZSTD_getCParams() to create cParams from compressionLevel.
- *  ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParam_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
+ *  ZSTD_estimateCStreamSize_usingCCtxParams() can be used in tandem with ZSTD_CCtxParams_setParameter(). Only single-threaded compression is supported. This function will return an error code if ZSTD_c_nbWorkers is >= 1.
  *  Note : CStream size estimation is only correct for single-threaded compression.
  *  ZSTD_DStream memory budget depends on window Size.
  *  This information can be passed manually, using ZSTD_estimateDStreamSize,
@@ -1226,22 +1338,26 b' ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict'
 ZSTDLIB_API ZSTD_CDict* ZSTD_createCDict_byReference(const void* dictBuffer, size_t dictSize, int compressionLevel);
 
 /*! ZSTD_getCParams() :
-*   @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
-*   `estimatedSrcSize` value is optional, select 0 if not known */
+ * @return ZSTD_compressionParameters structure for a selected compression level and estimated srcSize.
+ * `estimatedSrcSize` value is optional, select 0 if not known */
 ZSTDLIB_API ZSTD_compressionParameters ZSTD_getCParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
 
 /*! ZSTD_getParams() :
-*   same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.
-*   All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */
+ *  same as ZSTD_getCParams(), but @return a full `ZSTD_parameters` object instead of sub-component `ZSTD_compressionParameters`.
+ *  All fields of `ZSTD_frameParameters` are set to default : contentSize=1, checksum=0, noDictID=0 */
 ZSTDLIB_API ZSTD_parameters ZSTD_getParams(int compressionLevel, unsigned long long estimatedSrcSize, size_t dictSize);
 
 /*! ZSTD_checkCParams() :
-*   Ensure param values remain within authorized range */
+ *  Ensure param values remain within authorized range.
+ * @return 0 on success, or an error code (can be checked with ZSTD_isError()) */
 ZSTDLIB_API size_t ZSTD_checkCParams(ZSTD_compressionParameters params);
 
 /*! ZSTD_adjustCParams() :
  *  optimize params for a given `srcSize` and `dictSize`.
- *  both values are optional, select `0` if unknown. */
+ * `srcSize` can be unknown, in which case use ZSTD_CONTENTSIZE_UNKNOWN.
+ * `dictSize` must be `0` when there is no dictionary.
+ *  cPar can be invalid : all parameters will be clamped within valid range in the @return struct.
+ *  This function never fails (wide contract) */
 ZSTDLIB_API ZSTD_compressionParameters ZSTD_adjustCParams(ZSTD_compressionParameters cPar, unsigned long long srcSize, size_t dictSize);
 
 /*! ZSTD_compress_advanced() :
@@ -1314,6 +1430,17 b' ZSTDLIB_API size_t ZSTD_CCtx_refPrefix_a'
  * See the comments on that enum for an explanation of the feature. */
 #define ZSTD_c_forceAttachDict ZSTD_c_experimentalParam4
 
+/* Controls how the literals are compressed (default is auto).
+ * The value must be of type ZSTD_literalCompressionMode_e.
+ * See ZSTD_literalCompressionMode_t enum definition for details.
+ */
+#define ZSTD_c_literalCompressionMode ZSTD_c_experimentalParam5
+
+/* Tries to fit compressed block size to be around targetCBlockSize.
+ * No target when targetCBlockSize == 0.
+ * There is no guarantee on compressed block size (default:0) */
+#define ZSTD_c_targetCBlockSize ZSTD_c_experimentalParam6
+
 /*! ZSTD_CCtx_getParameter() :
  *  Get the requested compression parameter value, selected by enum ZSTD_cParameter,
  *  and store it into int* value.
@@ -1325,10 +1452,10 b' ZSTDLIB_API size_t ZSTD_CCtx_getParamete'
 /*! ZSTD_CCtx_params :
  *  Quick howto :
  *  - ZSTD_createCCtxParams() : Create a ZSTD_CCtx_params structure
- *  - ZSTD_CCtxParam_setParameter() : Push parameters one by one into
- *                                    an existing ZSTD_CCtx_params structure.
- *                                    This is similar to
- *                                    ZSTD_CCtx_setParameter().
+ *  - ZSTD_CCtxParams_setParameter() : Push parameters one by one into
+ *                                     an existing ZSTD_CCtx_params structure.
+ *                                     This is similar to
+ *                                     ZSTD_CCtx_setParameter().
  *  - ZSTD_CCtx_setParametersUsingCCtxParams() : Apply parameters to
  *                                    an existing CCtx.
  *                                    These parameters will be applied to
@@ -1359,20 +1486,20 b' ZSTDLIB_API size_t ZSTD_CCtxParams_init('
  */
 ZSTDLIB_API size_t ZSTD_CCtxParams_init_advanced(ZSTD_CCtx_params* cctxParams, ZSTD_parameters params);
 
-/*! ZSTD_CCtxParam_setParameter() :
+/*! ZSTD_CCtxParams_setParameter() :
  *  Similar to ZSTD_CCtx_setParameter.
  *  Set one compression parameter, selected by enum ZSTD_cParameter.
  *  Parameters must be applied to a ZSTD_CCtx using ZSTD_CCtx_setParametersUsingCCtxParams().
  * @result : 0, or an error code (which can be tested with ZSTD_isError()).
  */
-ZSTDLIB_API size_t ZSTD_CCtxParam_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
+ZSTDLIB_API size_t ZSTD_CCtxParams_setParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int value);
 
-/*! ZSTD_CCtxParam_getParameter() :
+/*! ZSTD_CCtxParams_getParameter() :
  * Similar to ZSTD_CCtx_getParameter.
  * Get the requested value of one compression parameter, selected by enum ZSTD_cParameter.
  * @result : 0, or an error code (which can be tested with ZSTD_isError()).
  */
-ZSTDLIB_API size_t ZSTD_CCtxParam_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
+ZSTDLIB_API size_t ZSTD_CCtxParams_getParameter(ZSTD_CCtx_params* params, ZSTD_cParameter param, int* value);
 
 /*! ZSTD_CCtx_setParametersUsingCCtxParams() :
  *  Apply a set of ZSTD_CCtx_params to the compression context.
@@ -1415,31 +1542,6 b' ZSTDLIB_API unsigned ZSTD_isFrame(const '
  *  it must remain read accessible throughout the lifetime of DDict */
 ZSTDLIB_API ZSTD_DDict* ZSTD_createDDict_byReference(const void* dictBuffer, size_t dictSize);
 
-
-/*! ZSTD_getDictID_fromDict() :
- *  Provides the dictID stored within dictionary.
- *  if @return == 0, the dictionary is not conformant with Zstandard specification.
- *  It can still be loaded, but as a content-only dictionary. */
-ZSTDLIB_API unsigned ZSTD_getDictID_fromDict(const void* dict, size_t dictSize);
-
-/*! ZSTD_getDictID_fromDDict() :
- *  Provides the dictID of the dictionary loaded into `ddict`.
- *  If @return == 0, the dictionary is not conformant to Zstandard specification, or empty.
- *  Non-conformant dictionaries can still be loaded, but as content-only dictionaries. */
-ZSTDLIB_API unsigned ZSTD_getDictID_fromDDict(const ZSTD_DDict* ddict);
-
-/*! ZSTD_getDictID_fromFrame() :
- *  Provides the dictID required to decompressed the frame stored within `src`.
- *  If @return == 0, the dictID could not be decoded.
- *  This could for one of the following reasons :
- *  - The frame does not require a dictionary to be decoded (most common case).
- *  - The frame was built with dictID intentionally removed. Whatever dictionary is necessary is a hidden information.
- *    Note : this use case also happens when using a non-conformant dictionary.
- *  - `srcSize` is too small, and as a result, the frame header could not be decoded (only possible if `srcSize < ZSTD_FRAMEHEADERSIZE_MAX`).
- *  - This is not a Zstandard frame.
- *  When identifying the exact failure cause, it's possible to use ZSTD_getFrameHeader(), which will provide a more precise error code. */
-ZSTDLIB_API unsigned ZSTD_getDictID_fromFrame(const void* src, size_t srcSize);
-
 /*! ZSTD_DCtx_loadDictionary_byReference() :
  *  Same as ZSTD_DCtx_loadDictionary(),
  *  but references `dict` content instead of copying it into `dctx`.
@@ -1501,14 +1603,68 b' ZSTDLIB_API size_t ZSTD_decompressStream'
 ********************************************************************/
 
 /*=====   Advanced Streaming compression functions  =====*/
-ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize);   /**< pledgedSrcSize must be correct. If it is not known at init time, use ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs, "0" also disables frame content size field. It may be enabled in the future. */
-ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel); /**< creates of an internal CDict (incompatible with static CCtx), except if dict == NULL or dictSize < 8, in which case no dict is used. Note: dict is loaded with ZSTD_dm_auto (treated as a full zstd dictionary if it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.*/
+/**! ZSTD_initCStream_srcSize() :
+ * This function is deprecated, and equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, NULL); // clear the dictionary (if any)
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *
+ * pledgedSrcSize must be correct. If it is not known at init time, use
+ * ZSTD_CONTENTSIZE_UNKNOWN. Note that, for compatibility with older programs,
+ * "0" also disables frame content size field. It may be enabled in the future.
+ */
+ZSTDLIB_API size_t ZSTD_initCStream_srcSize(ZSTD_CStream* zcs, int compressionLevel, unsigned long long pledgedSrcSize);
+/**! ZSTD_initCStream_usingDict() :
+ * This function is deprecated, and is equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setParameter(zcs, ZSTD_c_compressionLevel, compressionLevel);
+ *     ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
+ *
+ * Creates of an internal CDict (incompatible with static CCtx), except if
+ * dict == NULL or dictSize < 8, in which case no dict is used.
+ * Note: dict is loaded with ZSTD_dm_auto (treated as a full zstd dictionary if
+ * it begins with ZSTD_MAGIC_DICTIONARY, else as raw content) and ZSTD_dlm_byCopy.
+ */
+ZSTDLIB_API size_t ZSTD_initCStream_usingDict(ZSTD_CStream* zcs, const void* dict, size_t dictSize, int compressionLevel);
+/**! ZSTD_initCStream_advanced() :
+ * This function is deprecated, and is approximately equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setZstdParams(zcs, params); // Set the zstd params and leave the rest as-is
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *     ZSTD_CCtx_loadDictionary(zcs, dict, dictSize);
+ *
+ * pledgedSrcSize must be correct. If srcSize is not known at init time, use
+ * value ZSTD_CONTENTSIZE_UNKNOWN. dict is loaded with ZSTD_dm_auto and ZSTD_dlm_byCopy.
+ */
 ZSTDLIB_API size_t ZSTD_initCStream_advanced(ZSTD_CStream* zcs, const void* dict, size_t dictSize,
-                                             ZSTD_parameters params, unsigned long long pledgedSrcSize);  /**< pledgedSrcSize must be correct. If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN. dict is loaded with ZSTD_dm_auto and ZSTD_dlm_byCopy. */
-ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);  /**< note : cdict will just be referenced, and must outlive compression session */
-ZSTDLIB_API size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize);  /**< same as ZSTD_initCStream_usingCDict(), with control over frame parameters. pledgedSrcSize must be correct. If srcSize is not known at init time, use value ZSTD_CONTENTSIZE_UNKNOWN. */
+                                             ZSTD_parameters params, unsigned long long pledgedSrcSize);
+/**! ZSTD_initCStream_usingCDict() :
+ * This function is deprecated, and equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_refCDict(zcs, cdict);
+ *
+ * note : cdict will just be referenced, and must outlive compression session
+ */
+ZSTDLIB_API size_t ZSTD_initCStream_usingCDict(ZSTD_CStream* zcs, const ZSTD_CDict* cdict);
+/**! ZSTD_initCStream_usingCDict_advanced() :
+ * This function is deprecated, and is approximately equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setZstdFrameParams(zcs, fParams); // Set the zstd frame params and leave the rest as-is
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *     ZSTD_CCtx_refCDict(zcs, cdict);
+ *
+ * same as ZSTD_initCStream_usingCDict(), with control over frame parameters.
+ * pledgedSrcSize must be correct. If srcSize is not known at init time, use
+ * value ZSTD_CONTENTSIZE_UNKNOWN.
+ */
+ZSTDLIB_API size_t ZSTD_initCStream_usingCDict_advanced(ZSTD_CStream* zcs, const ZSTD_CDict* cdict, ZSTD_frameParameters fParams, unsigned long long pledgedSrcSize);
 
 /*! ZSTD_resetCStream() :
+ * This function is deprecated, and is equivalent to:
+ *     ZSTD_CCtx_reset(zcs, ZSTD_reset_session_only);
+ *     ZSTD_CCtx_setPledgedSrcSize(zcs, pledgedSrcSize);
+ *
  *  start a new frame, using same parameters from previous frame.
  *  This is typically useful to skip dictionary loading stage, since it will re-use it in-place.
  *  Note that zcs must be init at least once before using ZSTD_resetCStream().
@@ -1555,9 +1711,32 b' ZSTDLIB_API size_t ZSTD_toFlushNow(ZSTD_'
 
 
 /*=====   Advanced Streaming decompression functions  =====*/
-ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize); /**< note: no dictionary will be used if dict == NULL or dictSize < 8 */
-ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);  /**< note : ddict is referenced, it must outlive decompression session */
-ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);  /**< re-use decompression parameters from previous init; saves dictionary loading */
+/**
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *     ZSTD_DCtx_loadDictionary(zds, dict, dictSize);
+ *
+ * note: no dictionary will be used if dict == NULL or dictSize < 8
+ */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDict(ZSTD_DStream* zds, const void* dict, size_t dictSize);
+/**
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *     ZSTD_DCtx_refDDict(zds, ddict);
+ *
+ * note : ddict is referenced, it must outlive decompression session
+ */
+ZSTDLIB_API size_t ZSTD_initDStream_usingDDict(ZSTD_DStream* zds, const ZSTD_DDict* ddict);
+/**
+ * This function is deprecated, and is equivalent to:
+ *
+ *     ZSTD_DCtx_reset(zds, ZSTD_reset_session_only);
+ *
+ * re-use decompression parameters from previous init; saves dictionary loading
+ */
+ZSTDLIB_API size_t ZSTD_resetDStream(ZSTD_DStream* zds);
 
 
 /*********************************************************************
@@ -1696,7 +1875,7 b' typedef struct {'
     unsigned checksumFlag;
 } ZSTD_frameHeader;
 
-/** ZSTD_getFrameHeader() :
+/*! ZSTD_getFrameHeader() :
  *  decode Frame Header, or requires larger `srcSize`.
  * @return : 0, `zfhPtr` is correctly filled,
  *          >0, `srcSize` is too small, value is wanted `srcSize` amount,
@@ -1730,7 +1909,7 b' ZSTDLIB_API ZSTD_nextInputType_e ZSTD_ne'
 /*!
     Block functions produce and decode raw zstd blocks, without frame metadata.
     Frame metadata cost is typically ~18 bytes, which can be non-negligible for very small blocks (< 100 bytes).
-    User will have to take in charge required information to regenerate data, such as compressed and content sizes.
+    But users will have to take in charge needed metadata to regenerate data, such as compressed and content sizes.
 
     A few rules to respect :
     - Compressing and decompressing require a context structure
@@ -1741,12 +1920,14 b' ZSTDLIB_API ZSTD_nextInputType_e ZSTD_ne'
       + copyCCtx() and copyDCtx() can be used too
     - Block size is limited, it must be <= ZSTD_getBlockSize() <= ZSTD_BLOCKSIZE_MAX == 128 KB
       + If input is larger than a block size, it's necessary to split input data into multiple blocks
-      + For inputs larger than a single block, really consider using regular ZSTD_compress() instead.
-        Frame metadata is not that costly, and quickly becomes negligible as source size grows larger.
-    - When a block is considered not compressible enough, ZSTD_compressBlock() result will be zero.
-      In which case, nothing is produced into `dst` !
-      + User must test for such outcome and deal directly with uncompressed data
-      + ZSTD_decompressBlock() doesn't accept uncompressed data as input !!!
+      + For inputs larger than a single block, consider using regular ZSTD_compress() instead.
+        Frame metadata is not that costly, and quickly becomes negligible as source size grows larger than a block.
+    - When a block is considered not compressible enough, ZSTD_compressBlock() result will be 0 (zero) !
+      ===> In which case, nothing is produced into `dst` !
+      + User __must__ test for such outcome and deal directly with uncompressed data
+      + A block cannot be declared incompressible if ZSTD_compressBlock() return value was != 0.
+        Doing so would mess up with statistics history, leading to potential data corruption.
+      + ZSTD_decompressBlock() _doesn't accept uncompressed data as input_ !!
       + In case of multiple successive blocks, should some of them be uncompressed,
         decoder must be informed of their existence in order to follow proper history.
         Use ZSTD_insertBlock() for such a case.

tests/test-repo-compengines.t

@@ -169,7 +169,7 b' checking zstd options'
   > done
 
   $ $RUNTESTDIR/f -s zstd-*/.hg/store/data/*
-  zstd-level-1/.hg/store/data/a.i: size=4097
+  zstd-level-1/.hg/store/data/a.i: size=4114
   zstd-level-22/.hg/store/data/a.i: size=4091
   zstd-level-default/\.hg/store/data/a\.i: size=(4094|4102) (re)