Merge pull request #670 from lz4/dev

v1.9.0

11 files changed, 2089 insertions, 1288 deletions

diff --git a/lib/Makefile b/lib/Makefile
index 88d9b4f..cb1571c 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -71,7 +71,11 @@ else
 	SHARED_EXT_VER = $(SHARED_EXT).$(LIBVER)
 endif
 
+ifneq (,$(filter Windows%,$(OS)))
+LIBLZ4 = liblz4-$(LIBVER_MAJOR)
+else
 LIBLZ4 = liblz4.$(SHARED_EXT_VER)
+endif
 
 .PHONY: default
 default: lib-release
@@ -103,8 +107,7 @@ $(LIBLZ4): $(SRCFILES)
 ifeq ($(BUILD_SHARED),yes)  # can be disabled on command line
 	@echo compiling dynamic library $(LIBVER)
 ifneq (,$(filter Windows%,$(OS)))
-	$(Q)$(CC) $(FLAGS) -DLZ4_DLL_EXPORT=1 -shared $^ -o dll\$@.dll
-	dlltool -D dll\liblz4.dll -d dll\liblz4.def -l dll\liblz4.lib
+	$(Q)$(CC) $(FLAGS) -DLZ4_DLL_EXPORT=1 -shared $^ -o dll/$@.dll -Wl,--out-implib,dll/liblz4.lib
 else
 	$(Q)$(CC) $(FLAGS) -shared $^ -fPIC -fvisibility=hidden $(SONAME_FLAGS) -o $@
 	@echo creating versioned links
@@ -124,7 +127,7 @@ clean:
 #-----------------------------------------------------------------------------
 # make install is validated only for Linux, OSX, BSD, Hurd and Solaris targets
 #-----------------------------------------------------------------------------
-ifneq (,$(filter $(shell uname),Linux Darwin GNU/kFreeBSD GNU OpenBSD FreeBSD NetBSD DragonFly SunOS Haiku MidnightBSD))
+ifneq (,$(filter $(shell uname),Linux Darwin GNU/kFreeBSD GNU OpenBSD FreeBSD NetBSD DragonFly SunOS Haiku MidnightBSD MINGW32_NT-6.1 MINGW64_NT-6.1 MINGW32_NT-10.0 MINGW64_NT-10.0))
 
 .PHONY: listL120
 listL120:  # extract lines >= 120 characters in *.{c,h}, by Takayuki Matsuoka (note : $$, for Makefile compatibility)
@@ -138,12 +141,14 @@ PREFIX      ?= /usr/local
 prefix      ?= $(PREFIX)
 EXEC_PREFIX ?= $(prefix)
 exec_prefix ?= $(EXEC_PREFIX)
+BINDIR      ?= $(exec_prefix)/bin
+bindir      ?= $(BINDIR)
 LIBDIR      ?= $(exec_prefix)/lib
 libdir      ?= $(LIBDIR)
 INCLUDEDIR  ?= $(prefix)/include
 includedir  ?= $(INCLUDEDIR)
 
-ifneq (,$(filter $(OS),OpenBSD FreeBSD NetBSD DragonFly))
+ifneq (,$(filter $(OS),OpenBSD FreeBSD NetBSD DragonFly MidnightBSD))
 PKGCONFIGDIR ?= $(prefix)/libdata/pkgconfig
 else
 PKGCONFIGDIR ?= $(libdir)/pkgconfig
@@ -168,7 +173,7 @@ liblz4.pc: liblz4.pc.in Makefile
           $< >$@
 
 install: lib liblz4.pc
-	$(Q)$(INSTALL) -d -m 755 $(DESTDIR)$(pkgconfigdir)/ $(DESTDIR)$(includedir)/ $(DESTDIR)$(libdir)/
+	$(Q)$(INSTALL) -d -m 755 $(DESTDIR)$(pkgconfigdir)/ $(DESTDIR)$(includedir)/ $(DESTDIR)$(libdir)/ $(DESTDIR)$(bindir)/
 	$(Q)$(INSTALL_DATA) liblz4.pc $(DESTDIR)$(pkgconfigdir)/
 	@echo Installing libraries
 ifeq ($(BUILD_STATIC),yes)
@@ -176,10 +181,18 @@ ifeq ($(BUILD_STATIC),yes)
 	$(Q)$(INSTALL_DATA) lz4frame_static.h $(DESTDIR)$(includedir)/lz4frame_static.h
 endif
 ifeq ($(BUILD_SHARED),yes)
+# Traditionnally, one installs the DLLs in the bin directory as programs
+# search them first in their directory. This allows to not pollute system
+# directories (like c:/windows/system32), nor modify the PATH variable.
+ifneq (,$(filter Windows%,$(OS)))
+	$(Q)$(INSTALL_PROGRAM) dll/$(LIBLZ4).dll $(DESTDIR)$(bindir)
+	$(Q)$(INSTALL_PROGRAM) dll/liblz4.lib $(DESTDIR)$(libdir)
+else
 	$(Q)$(INSTALL_PROGRAM) liblz4.$(SHARED_EXT_VER) $(DESTDIR)$(libdir)
 	$(Q)ln -sf liblz4.$(SHARED_EXT_VER) $(DESTDIR)$(libdir)/liblz4.$(SHARED_EXT_MAJOR)
 	$(Q)ln -sf liblz4.$(SHARED_EXT_VER) $(DESTDIR)$(libdir)/liblz4.$(SHARED_EXT)
 endif
+endif
 	@echo Installing headers in $(includedir)
 	$(Q)$(INSTALL_DATA) lz4.h $(DESTDIR)$(includedir)/lz4.h
 	$(Q)$(INSTALL_DATA) lz4hc.h $(DESTDIR)$(includedir)/lz4hc.h
@@ -188,9 +201,14 @@ endif
 
 uninstall:
 	$(Q)$(RM) $(DESTDIR)$(pkgconfigdir)/liblz4.pc
+ifneq (,$(filter Windows%,$(OS)))
+	$(Q)$(RM) $(DESTDIR)$(bindir)/$(LIBLZ4).dll
+	$(Q)$(RM) $(DESTDIR)$(libdir)/liblz4.lib
+else
 	$(Q)$(RM) $(DESTDIR)$(libdir)/liblz4.$(SHARED_EXT)
 	$(Q)$(RM) $(DESTDIR)$(libdir)/liblz4.$(SHARED_EXT_MAJOR)
 	$(Q)$(RM) $(DESTDIR)$(libdir)/liblz4.$(SHARED_EXT_VER)
+endif
 	$(Q)$(RM) $(DESTDIR)$(libdir)/liblz4.a
 	$(Q)$(RM) $(DESTDIR)$(includedir)/lz4.h
 	$(Q)$(RM) $(DESTDIR)$(includedir)/lz4hc.h
diff --git a/lib/README.md b/lib/README.md
index 7082fe3..b753195 100644
--- a/lib/README.md
+++ b/lib/README.md
@@ -7,8 +7,8 @@ not all of them are necessary.
 #### Minimal LZ4 build
 
 The minimum required is **`lz4.c`** and **`lz4.h`**,
-which provides the fast compression and decompression algorithm.
-They generate and decode data using [LZ4 block format].
+which provides the fast compression and decompression algorithms.
+They generate and decode data using the [LZ4 block format].
 
 
 #### High Compression variant
@@ -16,13 +16,14 @@ They generate and decode data using [LZ4 block format].
 For more compression ratio at the cost of compression speed,
 the High Compression variant called **lz4hc** is available.
 Add files **`lz4hc.c`** and **`lz4hc.h`**.
-The variant still depends on regular `lib/lz4.*` source files.
+This variant also compresses data using the [LZ4 block format],
+and depends on regular `lib/lz4.*` source files.
 
 
-#### Frame variant, for interoperability
+#### Frame support, for interoperability
 
 In order to produce compressed data compatible with `lz4` command line utility,
-it's necessary to encode lz4-compressed blocks using the [official interoperable frame format].
+it's necessary to use the [official interoperable frame format].
 This format is generated and decoded automatically by the **lz4frame** library.
 Its public API is described in `lib/lz4frame.h`.
 In order to work properly, lz4frame needs all other modules present in `/lib`,
@@ -32,15 +33,63 @@ So it's necessary to include all `*.c` and `*.h` files present in `/lib`.
 
 #### Advanced / Experimental API
 
-A complex API defined in `lz4frame_static.h` contains definitions
-which are not guaranteed to remain stable in future versions.
-As a consequence, it must be used with static linking ***only***.
+Definitions which are not guaranteed to remain stable in future versions,
+are protected behind macros, such as `LZ4_STATIC_LINKING_ONLY`.
+As the name implies, these definitions can only be invoked
+in the context of static linking ***only***.
+Otherwise, dependent application may fail on API or ABI break in the future.
+The associated symbols are also not present in dynamic library by default.
+Should they be nonetheless needed, it's possible to force their publication
+by using build macro `LZ4_PUBLISH_STATIC_FUNCTIONS`.
+
+
+#### Build macros
+
+The following build macro can be selected at compilation time :
+
+- `LZ4_FAST_DEC_LOOP` : this triggers the optimized decompression loop.
+  This loops works great on x86/x64 cpus, and is automatically enabled on this platform.
+  It's possible to enable or disable it manually, by passing `LZ4_FAST_DEC_LOOP=1` or `0` to the preprocessor.
+  For example, with `gcc` : `-DLZ4_FAST_DEC_LOOP=1`,
+  and with `make` : `CPPFLAGS+=-DLZ4_FAST_DEC_LOOP=1 make lz4`.
+
+- `LZ4_DISTANCE_MAX` : control the maximum offset that the compressor will allow.
+  Set to 65535 by default, which is the maximum value supported by lz4 format.
+  Reducing maximum distance will reduce opportunities for LZ4 to find matches,
+  hence will produce worse the compression ratio.
+  However, a smaller max distance may allow compatibility with specific decoders using limited memory budget.
+  This build macro only influences the compressed output of the compressor.
+
+- `LZ4_DISABLE_DEPRECATE_WARNINGS` : invoking a deprecated function will make the compiler generate a warning.
+  This is meant to invite users to update their source code.
+  Should this be a problem, it's generally possible to make the compiler ignore these warnings,
+  for example with `-Wno-deprecated-declarations` on `gcc`,
+  or `_CRT_SECURE_NO_WARNINGS` for Visual Studio.
+  Another method is to define `LZ4_DISABLE_DEPRECATE_WARNINGS`
+  before including the LZ4 header files.
+
+
+#### Amalgamation
+
+lz4 source code can be amalgamated into a single file.
+One can combine all source code into `lz4_all.c` by using following command:
+```
+cat lz4.c > lz4_all.c
+cat lz4hc.c >> lz4_all.c
+cat lz4frame.c >> lz4_all.c
+```
+(`cat` file order is important) then compile `lz4_all.c`.
+All `*.h` files present in `/lib` remain necessary to compile `lz4_all.c`.
 
 
 #### Windows : using MinGW+MSYS to create DLL
 
 DLL can be created using MinGW+MSYS with the `make liblz4` command.
 This command creates `dll\liblz4.dll` and the import library `dll\liblz4.lib`.
+To override the `dlltool` command  when cross-compiling on Linux, just set the `DLLTOOL` variable. Example of cross compilation on Linux with mingw-w64 64 bits:
+```
+make BUILD_STATIC=no CC=x86_64-w64-mingw32-gcc DLLTOOL=x86_64-w64-mingw32-dlltool OS=Windows_NT
+```
 The import library is only required with Visual C++.
 The header files `lz4.h`, `lz4hc.h`, `lz4frame.h` and the dynamic library
 `dll\liblz4.dll` are required to compile a project using gcc/MinGW.
@@ -48,7 +97,7 @@ The dynamic library has to be added to linking options.
 It means that if a project that uses LZ4 consists of a single `test-dll.c`
 file it should be linked with `dll\liblz4.dll`. For example:
 ```
-    gcc $(CFLAGS) -Iinclude/ test-dll.c -o test-dll dll\liblz4.dll
+    $(CC) $(CFLAGS) -Iinclude/ test-dll.c -o test-dll dll\liblz4.dll
 ```
 The compiled executable will require LZ4 DLL which is available at `dll\liblz4.dll`.
 
diff --git a/lib/dll/liblz4.def b/lib/dll/liblz4.def
deleted file mode 100644
index 0ace223..0000000
--- a/lib/dll/liblz4.def
+++ /dev/null
@@ -1,62 +0,0 @@
-LIBRARY liblz4.dll
-EXPORTS
-    LZ4F_compressBegin
-    LZ4F_compressBound
-    LZ4F_compressEnd
-    LZ4F_compressFrame
-    LZ4F_compressFrameBound
-    LZ4F_compressUpdate
-    LZ4F_createCompressionContext
-    LZ4F_createDecompressionContext
-    LZ4F_decompress
-    LZ4F_flush
-    LZ4F_freeCompressionContext
-    LZ4F_freeDecompressionContext
-    LZ4F_getErrorName
-    LZ4F_getFrameInfo
-    LZ4F_getVersion
-    LZ4F_isError
-    LZ4_compress
-    LZ4_compressBound
-    LZ4_compressHC
-    LZ4_compressHC_continue
-    LZ4_compressHC_limitedOutput
-    LZ4_compressHC_limitedOutput_continue
-    LZ4_compressHC_limitedOutput_withStateHC
-    LZ4_compressHC_withStateHC
-    LZ4_compress_HC
-    LZ4_compress_HC_continue
-    LZ4_compress_HC_extStateHC
-    LZ4_compress_continue
-    LZ4_compress_default
-    LZ4_compress_destSize
-    LZ4_compress_fast
-    LZ4_compress_fast_continue
-    LZ4_compress_fast_extState
-    LZ4_compress_limitedOutput
-    LZ4_compress_limitedOutput_continue
-    LZ4_compress_limitedOutput_withState
-    LZ4_compress_withState
-    LZ4_createStream
-    LZ4_createStreamDecode
-    LZ4_createStreamHC
-    LZ4_decompress_fast
-    LZ4_decompress_fast_continue
-    LZ4_decompress_fast_usingDict
-    LZ4_decompress_safe
-    LZ4_decompress_safe_continue
-    LZ4_decompress_safe_partial
-    LZ4_decompress_safe_usingDict
-    LZ4_freeStream
-    LZ4_freeStreamDecode
-    LZ4_freeStreamHC
-    LZ4_loadDict
-    LZ4_loadDictHC
-    LZ4_resetStream
-    LZ4_resetStreamHC
-    LZ4_saveDict
-    LZ4_saveDictHC
-    LZ4_setStreamDecode
-    LZ4_sizeofState
-    LZ4_sizeofStateHC
-    LZ4_versionNumber
diff --git a/lib/lz4.c b/lib/lz4.c
index 4046102..693bdaf 100644
--- a/lib/lz4.c
+++ b/lib/lz4.c
@@ -32,7 +32,6 @@
     - LZ4 source repository : https://github.com/lz4/lz4
 */
 
-
 /*-************************************
 *  Tuning parameters
 **************************************/
@@ -91,6 +90,14 @@
 /*-************************************
 *  Dependency
 **************************************/
+/*
+ * LZ4_SRC_INCLUDED:
+ * Amalgamation flag, whether lz4.c is included
+ */
+#ifndef LZ4_SRC_INCLUDED
+#  define LZ4_SRC_INCLUDED 1
+#endif
+
 #define LZ4_STATIC_LINKING_ONLY
 #define LZ4_DISABLE_DEPRECATE_WARNINGS /* due to LZ4_decompress_safe_withPrefix64k */
 #include "lz4.h"
@@ -136,7 +143,7 @@
  * and also LZ4_wildCopy is forcibly inlined, so that the O2 attribute
  * of LZ4_wildCopy does not affect the compression speed.
  */
-#if defined(__PPC64__) && defined(__LITTLE_ENDIAN__) && defined(__GNUC__)
+#if defined(__PPC64__) && defined(__LITTLE_ENDIAN__) && defined(__GNUC__) && !defined(__clang__)
 #  define LZ4_FORCE_O2_GCC_PPC64LE __attribute__((optimize("O2")))
 #  define LZ4_FORCE_O2_INLINE_GCC_PPC64LE __attribute__((optimize("O2"))) LZ4_FORCE_INLINE
 #else
@@ -170,7 +177,7 @@
 
 
 /*-************************************
-*  Basic Types
+*  Types
 **************************************/
 #if defined(__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
 # include <stdint.h>
@@ -195,6 +202,13 @@
   typedef size_t reg_t;   /* 32-bits in x32 mode */
 #endif
 
+typedef enum {
+    notLimited = 0,
+    limitedOutput = 1,
+    fillOutput = 2
+} limitedOutput_directive;
+
+
 /*-************************************
 *  Reading and writing into memory
 **************************************/
@@ -228,7 +242,7 @@ static reg_t LZ4_read_ARCH(const void* ptr) { return ((const unalign*)ptr)->uArc
 static void LZ4_write16(void* memPtr, U16 value) { ((unalign*)memPtr)->u16 = value; }
 static void LZ4_write32(void* memPtr, U32 value) { ((unalign*)memPtr)->u32 = value; }
 
-#else  /* safe and portable access through memcpy() */
+#else  /* safe and portable access using memcpy() */
 
 static U16 LZ4_read16(const void* memPtr)
 {
@@ -290,6 +304,86 @@ void LZ4_wildCopy(void* dstPtr, const void* srcPtr, void* dstEnd)
     do { memcpy(d,s,8); d+=8; s+=8; } while (d<e);
 }
 
+static const unsigned inc32table[8] = {0, 1, 2,  1,  0,  4, 4, 4};
+static const int      dec64table[8] = {0, 0, 0, -1, -4,  1, 2, 3};
+
+
+#ifndef LZ4_FAST_DEC_LOOP
+#  if defined(__i386__) || defined(__x86_64__)
+#    define LZ4_FAST_DEC_LOOP 1
+#  else
+#    define LZ4_FAST_DEC_LOOP 0
+#  endif
+#endif
+
+#if LZ4_FAST_DEC_LOOP
+
+LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_memcpy_using_offset_base(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const size_t offset)
+{
+    if (offset < 8) {
+        dstPtr[0] = srcPtr[0];
+        dstPtr[1] = srcPtr[1];
+        dstPtr[2] = srcPtr[2];
+        dstPtr[3] = srcPtr[3];
+        srcPtr += inc32table[offset];
+        memcpy(dstPtr+4, srcPtr, 4);
+        srcPtr -= dec64table[offset];
+        dstPtr += 8;
+    } else {
+        memcpy(dstPtr, srcPtr, 8);
+        dstPtr += 8;
+        srcPtr += 8;
+    }
+
+    LZ4_wildCopy(dstPtr, srcPtr, dstEnd);
+}
+
+/* customized variant of memcpy, which can overwrite up to 32 bytes beyond dstEnd
+ * this version copies two times 16 bytes (instead of one time 32 bytes)
+ * because it must be compatible with offsets >= 16. */
+LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_wildCopy32(void* dstPtr, const void* srcPtr, void* dstEnd)
+{
+    BYTE* d = (BYTE*)dstPtr;
+    const BYTE* s = (const BYTE*)srcPtr;
+    BYTE* const e = (BYTE*)dstEnd;
+
+    do { memcpy(d,s,16); memcpy(d+16,s+16,16); d+=32; s+=32; } while (d<e);
+}
+
+LZ4_FORCE_O2_INLINE_GCC_PPC64LE void
+LZ4_memcpy_using_offset(BYTE* dstPtr, const BYTE* srcPtr, BYTE* dstEnd, const size_t offset)
+{
+    BYTE v[8];
+    switch(offset) {
+    case 1:
+        memset(v, *srcPtr, 8);
+        goto copy_loop;
+    case 2:
+        memcpy(v, srcPtr, 2);
+        memcpy(&v[2], srcPtr, 2);
+        memcpy(&v[4], &v[0], 4);
+        goto copy_loop;
+    case 4:
+        memcpy(v, srcPtr, 4);
+        memcpy(&v[4], srcPtr, 4);
+        goto copy_loop;
+    default:
+        LZ4_memcpy_using_offset_base(dstPtr, srcPtr, dstEnd, offset);
+        return;
+    }
+
+ copy_loop:
+    memcpy(dstPtr, v, 8);
+    dstPtr += 8;
+    while (dstPtr < dstEnd) {
+        memcpy(dstPtr, v, 8);
+        dstPtr += 8;
+    }
+}
+#endif
+
 
 /*-************************************
 *  Common Constants
@@ -300,14 +394,20 @@ void LZ4_wildCopy(void* dstPtr, const void* srcPtr, void* dstEnd)
 #define LASTLITERALS   5   /* see ../doc/lz4_Block_format.md#parsing-restrictions */
 #define MFLIMIT       12   /* see ../doc/lz4_Block_format.md#parsing-restrictions */
 #define MATCH_SAFEGUARD_DISTANCE  ((2*WILDCOPYLENGTH) - MINMATCH)   /* ensure it's possible to write 2 x wildcopyLength without overflowing output buffer */
+#define FASTLOOP_SAFE_DISTANCE 64
 static const int LZ4_minLength = (MFLIMIT+1);
 
 #define KB *(1 <<10)
 #define MB *(1 <<20)
 #define GB *(1U<<30)
 
-#define MAXD_LOG 16
-#define MAX_DISTANCE ((1 << MAXD_LOG) - 1)
+#ifndef LZ4_DISTANCE_MAX   /* can be user - defined at compile time */
+#  define LZ4_DISTANCE_MAX 65535
+#endif
+
+#if (LZ4_DISTANCE_MAX > 65535)   /* max supported by LZ4 format */
+#  error "LZ4_DISTANCE_MAX is too big : must be <= 65535"
+#endif
 
 #define ML_BITS  4
 #define ML_MASK  ((1U<<ML_BITS)-1)
@@ -455,7 +555,6 @@ static const U32 LZ4_skipTrigger = 6;  /* Increase this value ==> compression ru
 /*-************************************
 *  Local Structures and types
 **************************************/
-typedef enum { notLimited = 0, limitedOutput = 1, fillOutput = 2 } limitedOutput_directive;
 typedef enum { clearedTable = 0, byPtr, byU32, byU16 } tableType_t;
 
 /**
@@ -522,13 +621,14 @@ static U32 LZ4_hash4(U32 sequence, tableType_t const tableType)
 
 static U32 LZ4_hash5(U64 sequence, tableType_t const tableType)
 {
-    static const U64 prime5bytes = 889523592379ULL;
-    static const U64 prime8bytes = 11400714785074694791ULL;
     const U32 hashLog = (tableType == byU16) ? LZ4_HASHLOG+1 : LZ4_HASHLOG;
-    if (LZ4_isLittleEndian())
+    if (LZ4_isLittleEndian()) {
+        const U64 prime5bytes = 889523592379ULL;
         return (U32)(((sequence << 24) * prime5bytes) >> (64 - hashLog));
-    else
+    } else {
+        const U64 prime8bytes = 11400714785074694791ULL;
         return (U32)(((sequence >> 24) * prime8bytes) >> (64 - hashLog));
+    }
 }
 
 LZ4_FORCE_INLINE U32 LZ4_hashPosition(const void* const p, tableType_t const tableType)
@@ -609,6 +709,15 @@ LZ4_FORCE_INLINE void LZ4_prepareTable(
         LZ4_stream_t_internal* const cctx,
         const int inputSize,
         const tableType_t tableType) {
+    /* If compression failed during the previous step, then the context
+     * is marked as dirty, therefore, it has to be fully reset.
+     */
+    if (cctx->dirty) {
+        DEBUGLOG(5, "LZ4_prepareTable: Full reset for %p", cctx);
+        MEM_INIT(cctx, 0, sizeof(LZ4_stream_t_internal));
+        return;
+    }
+
     /* If the table hasn't been used, it's guaranteed to be zeroed out, and is
      * therefore safe to use no matter what mode we're in. Otherwise, we figure
      * out if it's safe to leave as is or whether it needs to be reset.
@@ -629,7 +738,7 @@ LZ4_FORCE_INLINE void LZ4_prepareTable(
         }
     }
 
-    /* Adding a gap, so all previous entries are > MAX_DISTANCE back, is faster
+    /* Adding a gap, so all previous entries are > LZ4_DISTANCE_MAX back, is faster
      * than compressing without a gap. However, compressing with
      * currentOffset == 0 is faster still, so we preserve that case.
      */
@@ -651,14 +760,15 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
                  const char* const source,
                  char* const dest,
                  const int inputSize,
-                 int *inputConsumed, /* only written when outputLimited == fillOutput */
+                 int *inputConsumed, /* only written when outputDirective == fillOutput */
                  const int maxOutputSize,
-                 const limitedOutput_directive outputLimited,
+                 const limitedOutput_directive outputDirective,
                  const tableType_t tableType,
                  const dict_directive dictDirective,
                  const dictIssue_directive dictIssue,
-                 const U32 acceleration)
+                 const int acceleration)
 {
+    int result;
     const BYTE* ip = (const BYTE*) source;
 
     U32 const startIndex = cctx->currentOffset;
@@ -693,9 +803,10 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
     U32 forwardH;
 
     DEBUGLOG(5, "LZ4_compress_generic: srcSize=%i, tableType=%u", inputSize, tableType);
-    /* Init conditions */
-    if (outputLimited == fillOutput && maxOutputSize < 1) return 0; /* Impossible to store anything */
-    if ((U32)inputSize > (U32)LZ4_MAX_INPUT_SIZE) return 0;   /* Unsupported inputSize, too large (or negative) */
+    /* If init conditions are not met, we don't have to mark stream
+     * as having dirty context, since no action was taken yet */
+    if (outputDirective == fillOutput && maxOutputSize < 1) return 0;   /* Impossible to store anything */
+    if ((U32)inputSize > (U32)LZ4_MAX_INPUT_SIZE) return 0;           /* Unsupported inputSize, too large (or negative) */
     if ((tableType == byU16) && (inputSize>=LZ4_64Klimit)) return 0;  /* Size too large (not within 64K limit) */
     if (tableType==byPtr) assert(dictDirective==noDict);      /* only supported use case with byPtr */
     assert(acceleration >= 1);
@@ -728,8 +839,8 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
         /* Find a match */
         if (tableType == byPtr) {
             const BYTE* forwardIp = ip;
-            unsigned step = 1;
-            unsigned searchMatchNb = acceleration << LZ4_skipTrigger;
+            int step = 1;
+            int searchMatchNb = acceleration << LZ4_skipTrigger;
             do {
                 U32 const h = forwardH;
                 ip = forwardIp;
@@ -743,14 +854,14 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
                 forwardH = LZ4_hashPosition(forwardIp, tableType);
                 LZ4_putPositionOnHash(ip, h, cctx->hashTable, tableType, base);
 
-            } while ( (match+MAX_DISTANCE < ip)
+            } while ( (match+LZ4_DISTANCE_MAX < ip)
                    || (LZ4_read32(match) != LZ4_read32(ip)) );
 
         } else {   /* byU32, byU16 */
 
             const BYTE* forwardIp = ip;
-            unsigned step = 1;
-            unsigned searchMatchNb = acceleration << LZ4_skipTrigger;
+            int step = 1;
+            int searchMatchNb = acceleration << LZ4_skipTrigger;
             do {
                 U32 const h = forwardH;
                 U32 const current = (U32)(forwardIp - base);
@@ -794,8 +905,8 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
 
                 if ((dictIssue == dictSmall) && (matchIndex < prefixIdxLimit)) continue;    /* match outside of valid area */
                 assert(matchIndex < current);
-                if ((tableType != byU16) && (matchIndex+MAX_DISTANCE < current)) continue;  /* too far */
-                if (tableType == byU16) assert((current - matchIndex) <= MAX_DISTANCE);     /* too_far presumed impossible with byU16 */
+                if ((tableType != byU16) && (matchIndex+LZ4_DISTANCE_MAX < current)) continue;  /* too far */
+                if (tableType == byU16) assert((current - matchIndex) <= LZ4_DISTANCE_MAX);     /* too_far presumed impossible with byU16 */
 
                 if (LZ4_read32(match) == LZ4_read32(ip)) {
                     if (maybe_extMem) offset = current - matchIndex;
@@ -811,16 +922,17 @@ LZ4_FORCE_INLINE int LZ4_compress_generic(
         /* Encode Literals */
         {   unsigned const litLength = (unsigned)(ip - anchor);
             token = op++;
-            if ((outputLimited == limitedOutput) &&  /* Check output buffer overflow */
-                (unlikely(op + litLength + (2 + 1 + LASTLITERALS) + (litLength/255) > olimit)))
-                return 0;
-            if ((outputLimited == fillOutput) &&
+            if ((outputDirective == limitedOutput) &&  /* Check output buffer overflow */
+                (unlikely(op + litLength + (2 + 1 + LASTLITERALS) + (litLength/255) > olimit)) )
+                return 0;   /* cannot compress within `dst` budget. Stored indexes in hash table are nonetheless fine */
+
+            if ((outputDirective == fillOutput) &&
                 (unlikely(op + (litLength+240)/255 /* litlen */ + litLength /* literals */ + 2 /* offset */ + 1 /* token */ + MFLIMIT - MINMATCH /* min last literals so last match is <= end - MFLIMIT */ > olimit))) {
                 op--;
                 goto _last_literals;
             }
             if (litLength >= RUN_MASK) {
-                int len = (int)litLength-RUN_MASK;
+                int len = (int)(litLength - RUN_MASK);
                 *token = (RUN_MASK<<ML_BITS);
                 for(; len >= 255 ; len-=255) *op++ = 255;
                 *op++ = (BYTE)len;
@@ -843,7 +955,7 @@ _next_match:
          * - token and *token : position to write 4-bits for match length; higher 4-bits for literal length supposed already written
          */
 
-        if ((outputLimited == fillOutput) &&
+        if ((outputDirective == fillOutput) &&
             (op + 2 /* offset */ + 1 /* token */ + MFLIMIT - MINMATCH /* min last literals so last match is <= end - MFLIMIT */ > olimit)) {
             /* the match was too close to the end, rewind and go to last literals */
             op = token;
@@ -853,11 +965,11 @@ _next_match:
         /* Encode Offset */
         if (maybe_extMem) {   /* static test */
             DEBUGLOG(6, "             with offset=%u  (ext if > %i)", offset, (int)(ip - (const BYTE*)source));
-            assert(offset <= MAX_DISTANCE && offset > 0);
+            assert(offset <= LZ4_DISTANCE_MAX && offset > 0);
             LZ4_writeLE16(op, (U16)offset); op+=2;
         } else  {
             DEBUGLOG(6, "             with offset=%u  (same segment)", (U32)(ip - match));
-            assert(ip-match <= MAX_DISTANCE);
+            assert(ip-match <= LZ4_DISTANCE_MAX);
             LZ4_writeLE16(op, (U16)(ip - match)); op+=2;
         }
 
@@ -870,7 +982,7 @@ _next_match:
                 assert(dictEnd > match);
                 if (limit > matchlimit) limit = matchlimit;
                 matchCode = LZ4_count(ip+MINMATCH, match+MINMATCH, limit);
-                ip += MINMATCH + matchCode;
+                ip += (size_t)matchCode + MINMATCH;
                 if (ip==limit) {
                     unsigned const more = LZ4_count(limit, (const BYTE*)source, matchlimit);
                     matchCode += more;
@@ -879,19 +991,20 @@ _next_match:
                 DEBUGLOG(6, "             with matchLength=%u starting in extDict", matchCode+MINMATCH);
             } else {
                 matchCode = LZ4_count(ip+MINMATCH, match+MINMATCH, matchlimit);
-                ip += MINMATCH + matchCode;
+                ip += (size_t)matchCode + MINMATCH;
                 DEBUGLOG(6, "             with matchLength=%u", matchCode+MINMATCH);
             }
 
-            if ((outputLimited) &&    /* Check output buffer overflow */
+            if ((outputDirective) &&    /* Check output buffer overflow */
                 (unlikely(op + (1 + LASTLITERALS) + (matchCode>>8) > olimit)) ) {
-                if (outputLimited == limitedOutput)
-                  return 0;
-                if (outputLimited == fillOutput) {
+                if (outputDirective == fillOutput) {
                     /* Match description too long : reduce it */
                     U32 newMatchCode = 15 /* in token */ - 1 /* to avoid needing a zero byte */ + ((U32)(olimit - op) - 2 - 1 - LASTLITERALS) * 255;
                     ip -= matchCode - newMatchCode;
                     matchCode = newMatchCode;
+                } else {
+                    assert(outputDirective == limitedOutput);
+                    return 0;   /* cannot compress within `dst` budget. Stored indexes in hash table are nonetheless fine */
                 }
             }
             if (matchCode >= ML_MASK) {
@@ -922,7 +1035,7 @@ _next_match:
 
             match = LZ4_getPosition(ip, cctx->hashTable, tableType, base);
             LZ4_putPosition(ip, cctx->hashTable, tableType, base);
-            if ( (match+MAX_DISTANCE >= ip)
+            if ( (match+LZ4_DISTANCE_MAX >= ip)
               && (LZ4_read32(match) == LZ4_read32(ip)) )
             { token=op++; *token=0; goto _next_match; }
 
@@ -957,7 +1070,7 @@ _next_match:
             LZ4_putIndexOnHash(current, h, cctx->hashTable, tableType);
             assert(matchIndex < current);
             if ( ((dictIssue==dictSmall) ? (matchIndex >= prefixIdxLimit) : 1)
-              && ((tableType==byU16) ? 1 : (matchIndex+MAX_DISTANCE >= current))
+              && ((tableType==byU16) ? 1 : (matchIndex+LZ4_DISTANCE_MAX >= current))
               && (LZ4_read32(match) == LZ4_read32(ip)) ) {
                 token=op++;
                 *token=0;
@@ -976,15 +1089,17 @@ _next_match:
 _last_literals:
     /* Encode Last Literals */
     {   size_t lastRun = (size_t)(iend - anchor);
-        if ( (outputLimited) &&  /* Check output buffer overflow */
+        if ( (outputDirective) &&  /* Check output buffer overflow */
             (op + lastRun + 1 + ((lastRun+255-RUN_MASK)/255) > olimit)) {
-            if (outputLimited == fillOutput) {
+            if (outputDirective == fillOutput) {
                 /* adapt lastRun to fill 'dst' */
-                lastRun  = (olimit-op) - 1;
+                assert(olimit >= op);
+                lastRun  = (size_t)(olimit-op) - 1;
                 lastRun -= (lastRun+240)/255;
+            } else {
+                assert(outputDirective == limitedOutput);
+                return 0;   /* cannot compress within `dst` budget. Stored indexes in hash table are nonetheless fine */
             }
-            if (outputLimited == limitedOutput)
-                return 0;
         }
         if (lastRun >= RUN_MASK) {
             size_t accumulator = lastRun - RUN_MASK;
@@ -999,31 +1114,33 @@ _last_literals:
         op += lastRun;
     }
 
-    if (outputLimited == fillOutput) {
+    if (outputDirective == fillOutput) {
         *inputConsumed = (int) (((const char*)ip)-source);
     }
     DEBUGLOG(5, "LZ4_compress_generic: compressed %i bytes into %i bytes", inputSize, (int)(((char*)op) - dest));
-    return (int)(((char*)op) - dest);
+    result = (int)(((char*)op) - dest);
+    assert(result > 0);
+    return result;
 }
 
 
 int LZ4_compress_fast_extState(void* state, const char* source, char* dest, int inputSize, int maxOutputSize, int acceleration)
 {
-    LZ4_stream_t_internal* ctx = &((LZ4_stream_t*)state)->internal_donotuse;
+    LZ4_stream_t_internal* const ctx = & LZ4_initStream(state, sizeof(LZ4_stream_t)) -> internal_donotuse;
+    assert(ctx != NULL);
     if (acceleration < 1) acceleration = ACCELERATION_DEFAULT;
-    LZ4_resetStream((LZ4_stream_t*)state);
     if (maxOutputSize >= LZ4_compressBound(inputSize)) {
         if (inputSize < LZ4_64Klimit) {
             return LZ4_compress_generic(ctx, source, dest, inputSize, NULL, 0, notLimited, byU16, noDict, noDictIssue, acceleration);
         } else {
-            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)source > MAX_DISTANCE)) ? byPtr : byU32;
+            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)source > LZ4_DISTANCE_MAX)) ? byPtr : byU32;
             return LZ4_compress_generic(ctx, source, dest, inputSize, NULL, 0, notLimited, tableType, noDict, noDictIssue, acceleration);
         }
     } else {
         if (inputSize < LZ4_64Klimit) {;
             return LZ4_compress_generic(ctx, source, dest, inputSize, NULL, maxOutputSize, limitedOutput, byU16, noDict, noDictIssue, acceleration);
         } else {
-            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)source > MAX_DISTANCE)) ? byPtr : byU32;
+            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)source > LZ4_DISTANCE_MAX)) ? byPtr : byU32;
             return LZ4_compress_generic(ctx, source, dest, inputSize, NULL, maxOutputSize, limitedOutput, tableType, noDict, noDictIssue, acceleration);
         }
     }
@@ -1053,7 +1170,7 @@ int LZ4_compress_fast_extState_fastReset(void* state, const char* src, char* dst
                 return LZ4_compress_generic(ctx, src, dst, srcSize, NULL, 0, notLimited, tableType, noDict, noDictIssue, acceleration);
             }
         } else {
-            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)src > MAX_DISTANCE)) ? byPtr : byU32;
+            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)src > LZ4_DISTANCE_MAX)) ? byPtr : byU32;
             LZ4_prepareTable(ctx, srcSize, tableType);
             return LZ4_compress_generic(ctx, src, dst, srcSize, NULL, 0, notLimited, tableType, noDict, noDictIssue, acceleration);
         }
@@ -1067,7 +1184,7 @@ int LZ4_compress_fast_extState_fastReset(void* state, const char* src, char* dst
                 return LZ4_compress_generic(ctx, src, dst, srcSize, NULL, dstCapacity, limitedOutput, tableType, noDict, noDictIssue, acceleration);
             }
         } else {
-            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)src > MAX_DISTANCE)) ? byPtr : byU32;
+            const tableType_t tableType = ((sizeof(void*)==4) && ((uptrval)src > LZ4_DISTANCE_MAX)) ? byPtr : byU32;
             LZ4_prepareTable(ctx, srcSize, tableType);
             return LZ4_compress_generic(ctx, src, dst, srcSize, NULL, dstCapacity, limitedOutput, tableType, noDict, noDictIssue, acceleration);
         }
@@ -1094,23 +1211,25 @@ int LZ4_compress_fast(const char* source, char* dest, int inputSize, int maxOutp
 }
 
 
-int LZ4_compress_default(const char* source, char* dest, int inputSize, int maxOutputSize)
+int LZ4_compress_default(const char* src, char* dst, int srcSize, int maxOutputSize)
 {
-    return LZ4_compress_fast(source, dest, inputSize, maxOutputSize, 1);
+    return LZ4_compress_fast(src, dst, srcSize, maxOutputSize, 1);
 }
 
 
 /* hidden debug function */
 /* strangely enough, gcc generates faster code when this function is uncommented, even if unused */
-int LZ4_compress_fast_force(const char* source, char* dest, int inputSize, int maxOutputSize, int acceleration)
+int LZ4_compress_fast_force(const char* src, char* dst, int srcSize, int dstCapacity, int acceleration)
 {
     LZ4_stream_t ctx;
-    LZ4_resetStream(&ctx);
+    LZ4_initStream(&ctx, sizeof(ctx));
 
-    if (inputSize < LZ4_64Klimit)
-        return LZ4_compress_generic(&ctx.internal_donotuse, source, dest, inputSize, NULL, maxOutputSize, limitedOutput, byU16,                        noDict, noDictIssue, acceleration);
-    else
-        return LZ4_compress_generic(&ctx.internal_donotuse, source, dest, inputSize, NULL, maxOutputSize, limitedOutput, sizeof(void*)==8 ? byU32 : byPtr, noDict, noDictIssue, acceleration);
+    if (srcSize < LZ4_64Klimit) {
+        return LZ4_compress_generic(&ctx.internal_donotuse, src, dst, srcSize, NULL, dstCapacity, limitedOutput, byU16,    noDict, noDictIssue, acceleration);
+    } else {
+        tableType_t const addrMode = (sizeof(void*) > 4) ? byU32 : byPtr;
+        return LZ4_compress_generic(&ctx.internal_donotuse, src, dst, srcSize, NULL, dstCapacity, limitedOutput, addrMode, noDict, noDictIssue, acceleration);
+    }
 }
 
 
@@ -1119,7 +1238,8 @@ int LZ4_compress_fast_force(const char* source, char* dest, int inputSize, int m
  * _continue() call without resetting it. */
 static int LZ4_compress_destSize_extState (LZ4_stream_t* state, const char* src, char* dst, int* srcSizePtr, int targetDstSize)
 {
-    LZ4_resetStream(state);
+    void* const s = LZ4_initStream(state, sizeof (*state));
+    assert(s != NULL); (void)s;
 
     if (targetDstSize >= LZ4_compressBound(*srcSizePtr)) {  /* compression success is guaranteed */
         return LZ4_compress_fast_extState(state, src, dst, *srcSizePtr, targetDstSize, 1);
@@ -1127,8 +1247,8 @@ static int LZ4_compress_destSize_extState (LZ4_stream_t* state, const char* src,
         if (*srcSizePtr < LZ4_64Klimit) {
             return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, byU16, noDict, noDictIssue, 1);
         } else {
-            tableType_t const tableType = ((sizeof(void*)==4) && ((uptrval)src > MAX_DISTANCE)) ? byPtr : byU32;
-            return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, tableType, noDict, noDictIssue, 1);
+            tableType_t const addrMode = ((sizeof(void*)==4) && ((uptrval)src > LZ4_DISTANCE_MAX)) ? byPtr : byU32;
+            return LZ4_compress_generic(&state->internal_donotuse, src, dst, *srcSizePtr, srcSizePtr, targetDstSize, fillOutput, addrMode, noDict, noDictIssue, 1);
     }   }
 }
 
@@ -1159,14 +1279,40 @@ int LZ4_compress_destSize(const char* src, char* dst, int* srcSizePtr, int targe
 
 LZ4_stream_t* LZ4_createStream(void)
 {
-    LZ4_stream_t* lz4s = (LZ4_stream_t*)ALLOC(sizeof(LZ4_stream_t));
+    LZ4_stream_t* const lz4s = (LZ4_stream_t*)ALLOC(sizeof(LZ4_stream_t));
     LZ4_STATIC_ASSERT(LZ4_STREAMSIZE >= sizeof(LZ4_stream_t_internal));    /* A compilation error here means LZ4_STREAMSIZE is not large enough */
     DEBUGLOG(4, "LZ4_createStream %p", lz4s);
     if (lz4s == NULL) return NULL;
-    LZ4_resetStream(lz4s);
+    LZ4_initStream(lz4s, sizeof(*lz4s));
     return lz4s;
 }
 
+#ifndef _MSC_VER  /* for some reason, Visual fails the aligment test on 32-bit x86 :
+                     it reports an aligment of 8-bytes,
+                     while actually aligning LZ4_stream_t on 4 bytes. */
+static size_t LZ4_stream_t_alignment(void)
+{
+    struct { char c; LZ4_stream_t t; } t_a;
+    return sizeof(t_a) - sizeof(t_a.t);
+}
+#endif
+
+LZ4_stream_t* LZ4_initStream (void* buffer, size_t size)
+{
+    DEBUGLOG(5, "LZ4_initStream");
+    if (buffer == NULL) return NULL;
+    if (size < sizeof(LZ4_stream_t)) return NULL;
+#ifndef _MSC_VER  /* for some reason, Visual fails the aligment test on 32-bit x86 :
+                     it reports an aligment of 8-bytes,
+                     while actually aligning LZ4_stream_t on 4 bytes. */
+    if (((size_t)buffer) & (LZ4_stream_t_alignment() - 1)) return NULL;  /* alignment check */
+#endif
+    MEM_INIT(buffer, 0, sizeof(LZ4_stream_t));
+    return (LZ4_stream_t*)buffer;
+}
+
+/* resetStream is now deprecated,
+ * prefer initStream() which is more general */
 void LZ4_resetStream (LZ4_stream_t* LZ4_stream)
 {
     DEBUGLOG(5, "LZ4_resetStream (ctx:%p)", LZ4_stream);
@@ -1226,10 +1372,16 @@ int LZ4_loadDict (LZ4_stream_t* LZ4_dict, const char* dictionary, int dictSize)
         p+=3;
     }
 
-    return dict->dictSize;
+    return (int)dict->dictSize;
 }
 
 void LZ4_attach_dictionary(LZ4_stream_t *working_stream, const LZ4_stream_t *dictionary_stream) {
+    /* Calling LZ4_resetStream_fast() here makes sure that changes will not be
+     * erased by subsequent calls to LZ4_resetStream_fast() in case stream was
+     * marked as having dirty context, e.g. requiring full reset.
+     */
+    LZ4_resetStream_fast(working_stream);
+
     if (dictionary_stream != NULL) {
         /* If the current offset is zero, we will never look in the
          * external dictionary context, since there is no value a table
@@ -1248,7 +1400,8 @@ void LZ4_attach_dictionary(LZ4_stream_t *working_stream, const LZ4_stream_t *dic
 
 static void LZ4_renormDictT(LZ4_stream_t_internal* LZ4_dict, int nextSize)
 {
-    if (LZ4_dict->currentOffset + nextSize > 0x80000000) {   /* potential ptrdiff_t overflow (32-bits mode) */
+    assert(nextSize >= 0);
+    if (LZ4_dict->currentOffset + (unsigned)nextSize > 0x80000000) {   /* potential ptrdiff_t overflow (32-bits mode) */
         /* rescale hash table */
         U32 const delta = LZ4_dict->currentOffset - 64 KB;
         const BYTE* dictEnd = LZ4_dict->dictionary + LZ4_dict->dictSize;
@@ -1265,7 +1418,10 @@ static void LZ4_renormDictT(LZ4_stream_t_internal* LZ4_dict, int nextSize)
 }
 
 
-int LZ4_compress_fast_continue (LZ4_stream_t* LZ4_stream, const char* source, char* dest, int inputSize, int maxOutputSize, int acceleration)
+int LZ4_compress_fast_continue (LZ4_stream_t* LZ4_stream,
+                                const char* source, char* dest,
+                                int inputSize, int maxOutputSize,
+                                int acceleration)
 {
     const tableType_t tableType = byU32;
     LZ4_stream_t_internal* streamPtr = &LZ4_stream->internal_donotuse;
@@ -1273,12 +1429,12 @@ int LZ4_compress_fast_continue (LZ4_stream_t* LZ4_stream, const char* source, ch
 
     DEBUGLOG(5, "LZ4_compress_fast_continue (inputSize=%i)", inputSize);
 
-    if (streamPtr->initCheck) return 0;   /* Uninitialized structure detected */
+    if (streamPtr->dirty) return 0;   /* Uninitialized structure detected */
     LZ4_renormDictT(streamPtr, inputSize);   /* avoid index overflow */
     if (acceleration < 1) acceleration = ACCELERATION_DEFAULT;
 
     /* invalidate tiny dictionaries */
-    if ( (streamPtr->dictSize-1 < 4)   /* intentional underflow */
+    if ( (streamPtr->dictSize-1 < 4-1)   /* intentional underflow */
       && (dictEnd != (const BYTE*)source) ) {
         DEBUGLOG(5, "LZ4_compress_fast_continue: dictSize(%u) at addr:%p is too small", streamPtr->dictSize, streamPtr->dictionary);
         streamPtr->dictSize = 0;
@@ -1371,7 +1527,7 @@ int LZ4_saveDict (LZ4_stream_t* LZ4_dict, char* safeBuffer, int dictSize)
     const BYTE* const previousDictEnd = dict->dictionary + dict->dictSize;
 
     if ((U32)dictSize > 64 KB) dictSize = 64 KB;   /* useless to define a dictionary > 64 KB */
-    if ((U32)dictSize > dict->dictSize) dictSize = dict->dictSize;
+    if ((U32)dictSize > dict->dictSize) dictSize = (int)dict->dictSize;
 
     memmove(safeBuffer, previousDictEnd - dictSize, dictSize);
 
@@ -1393,6 +1549,37 @@ typedef enum { decode_full_block = 0, partial_decode = 1 } earlyEnd_directive;
 #undef MIN
 #define MIN(a,b)    ( (a) < (b) ? (a) : (b) )
 
+/* Read the variable-length literal or match length.
+ *
+ * ip - pointer to use as input.
+ * lencheck - end ip.  Return an error if ip advances >= lencheck.
+ * loop_check - check ip >= lencheck in body of loop.  Returns loop_error if so.
+ * initial_check - check ip >= lencheck before start of loop.  Returns initial_error if so.
+ * error (output) - error code.  Should be set to 0 before call.
+ */
+typedef enum { loop_error = -2, initial_error = -1, ok = 0 } variable_length_error;
+LZ4_FORCE_INLINE unsigned
+read_variable_length(const BYTE**ip, const BYTE* lencheck, int loop_check, int initial_check, variable_length_error* error)
+{
+  unsigned length = 0;
+  unsigned s;
+  if (initial_check && unlikely((*ip) >= lencheck)) {    /* overflow detection */
+    *error = initial_error;
+    return length;
+  }
+  do {
+    s = **ip;
+    (*ip)++;
+    length += s;
+    if (loop_check && unlikely((*ip) >= lencheck)) {    /* overflow detection */
+      *error = loop_error;
+      return length;
+    }
+  } while (s==255);
+
+  return length;
+}
+
 /*! LZ4_decompress_generic() :
  *  This generic decompression function covers all use cases.
  *  It shall be instantiated several times, using different sets of directives.
@@ -1414,234 +1601,360 @@ LZ4_decompress_generic(
                  const size_t dictSize         /* note : = 0 if noDict */
                  )
 {
-    const BYTE* ip = (const BYTE*) src;
-    const BYTE* const iend = ip + srcSize;
+    if (src == NULL) return -1;
 
-    BYTE* op = (BYTE*) dst;
-    BYTE* const oend = op + outputSize;
-    BYTE* cpy;
+    {   const BYTE* ip = (const BYTE*) src;
+        const BYTE* const iend = ip + srcSize;
 
-    const BYTE* const dictEnd = (const BYTE*)dictStart + dictSize;
-    const unsigned inc32table[8] = {0, 1, 2,  1,  0,  4, 4, 4};
-    const int      dec64table[8] = {0, 0, 0, -1, -4,  1, 2, 3};
+        BYTE* op = (BYTE*) dst;
+        BYTE* const oend = op + outputSize;
+        BYTE* cpy;
 
-    const int safeDecode = (endOnInput==endOnInputSize);
-    const int checkOffset = ((safeDecode) && (dictSize < (int)(64 KB)));
+        const BYTE* const dictEnd = (dictStart == NULL) ? NULL : dictStart + dictSize;
 
-    /* Set up the "end" pointers for the shortcut. */
-    const BYTE* const shortiend = iend - (endOnInput ? 14 : 8) /*maxLL*/ - 2 /*offset*/;
-    const BYTE* const shortoend = oend - (endOnInput ? 14 : 8) /*maxLL*/ - 18 /*maxML*/;
+        const int safeDecode = (endOnInput==endOnInputSize);
+        const int checkOffset = ((safeDecode) && (dictSize < (int)(64 KB)));
 
-    DEBUGLOG(5, "LZ4_decompress_generic (srcSize:%i, dstSize:%i)", srcSize, outputSize);
 
-    /* Special cases */
-    assert(lowPrefix <= op);
-    assert(src != NULL);
-    if ((endOnInput) && (unlikely(outputSize==0))) return ((srcSize==1) && (*ip==0)) ? 0 : -1;  /* Empty output buffer */
-    if ((!endOnInput) && (unlikely(outputSize==0))) return (*ip==0 ? 1 : -1);
-    if ((endOnInput) && unlikely(srcSize==0)) return -1;
+        /* Set up the "end" pointers for the shortcut. */
+        const BYTE* const shortiend = iend - (endOnInput ? 14 : 8) /*maxLL*/ - 2 /*offset*/;
+        const BYTE* const shortoend = oend - (endOnInput ? 14 : 8) /*maxLL*/ - 18 /*maxML*/;
 
-    /* Main Loop : decode sequences */
-    while (1) {
         const BYTE* match;
         size_t offset;
+        unsigned token;
+        size_t length;
+
+
+        DEBUGLOG(5, "LZ4_decompress_generic (srcSize:%i, dstSize:%i)", srcSize, outputSize);
+
+        /* Special cases */
+        assert(lowPrefix <= op);
+        if ((endOnInput) && (unlikely(outputSize==0))) return ((srcSize==1) && (*ip==0)) ? 0 : -1;  /* Empty output buffer */
+        if ((!endOnInput) && (unlikely(outputSize==0))) return (*ip==0 ? 1 : -1);
+        if ((endOnInput) && unlikely(srcSize==0)) return -1;
+
+	/* Currently the fast loop shows a regression on qualcomm arm chips. */
+#if LZ4_FAST_DEC_LOOP
+        if ((oend - op) < FASTLOOP_SAFE_DISTANCE)
+            goto safe_decode;
+
+        /* Fast loop : decode sequences as long as output < iend-FASTLOOP_SAFE_DISTANCE */
+        while (1) {
+            /* Main fastloop assertion: We can always wildcopy FASTLOOP_SAFE_DISTANCE */
+            assert(oend - op >= FASTLOOP_SAFE_DISTANCE);
+
+            token = *ip++;
+            length = token >> ML_BITS;  /* literal length */
+
+            assert(!endOnInput || ip <= iend); /* ip < iend before the increment */
+
+            /* decode literal length */
+            if (length == RUN_MASK) {
+                variable_length_error error = ok;
+                length += read_variable_length(&ip, iend-RUN_MASK, endOnInput, endOnInput, &error);
+                if (error == initial_error) goto _output_error;
+                if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)(op))) goto _output_error;   /* overflow detection */
+                if ((safeDecode) && unlikely((uptrval)(ip)+length<(uptrval)(ip))) goto _output_error;   /* overflow detection */
+
+                /* copy literals */
+                cpy = op+length;
+                LZ4_STATIC_ASSERT(MFLIMIT >= WILDCOPYLENGTH);
+                if ( ((endOnInput) && ((cpy>oend-FASTLOOP_SAFE_DISTANCE) || (ip+length>iend-(2+1+LASTLITERALS))) )
+                     || ((!endOnInput) && (cpy>oend-FASTLOOP_SAFE_DISTANCE)) )
+                    {
+                        goto safe_literal_copy;
+                    }
+                LZ4_wildCopy32(op, ip, cpy);
+                ip += length; op = cpy;
+            } else {
+                cpy = op+length;
+                /* We don't need to check oend, since we check it once for each loop below */
+                if ( ((endOnInput) && (ip+16>iend-(2+1+LASTLITERALS))))
+                    {
+                        goto safe_literal_copy;
+                    }
+                /* Literals can only be 14, but hope compilers optimize if we copy by a register size */
+                memcpy(op, ip, 16);
+                ip += length; op = cpy;
+            }
 
-        unsigned const token = *ip++;
-        size_t length = token >> ML_BITS;  /* literal length */
+            /* get offset */
+            offset = LZ4_readLE16(ip); ip+=2;
+            match = op - offset;
 
-        assert(!endOnInput || ip <= iend); /* ip < iend before the increment */
+            /* get matchlength */
+            length = token & ML_MASK;
 
-        /* A two-stage shortcut for the most common case:
-         * 1) If the literal length is 0..14, and there is enough space,
-         * enter the shortcut and copy 16 bytes on behalf of the literals
-         * (in the fast mode, only 8 bytes can be safely copied this way).
-         * 2) Further if the match length is 4..18, copy 18 bytes in a similar
-         * manner; but we ensure that there's enough space in the output for
-         * those 18 bytes earlier, upon entering the shortcut (in other words,
-         * there is a combined check for both stages).
-         */
-        if ( (endOnInput ? length != RUN_MASK : length <= 8)
-            /* strictly "less than" on input, to re-enter the loop with at least one byte */
-          && likely((endOnInput ? ip < shortiend : 1) & (op <= shortoend)) ) {
-            /* Copy the literals */
-            memcpy(op, ip, endOnInput ? 16 : 8);
-            op += length; ip += length;
-
-            /* The second stage: prepare for match copying, decode full info.
-             * If it doesn't work out, the info won't be wasted. */
-            length = token & ML_MASK; /* match length */
-            offset = LZ4_readLE16(ip); ip += 2;
-            match = op - offset;
-            assert(match <= op); /* check overflow */
-
-            /* Do not deal with overlapping matches. */
-            if ( (length != ML_MASK)
-              && (offset >= 8)
-              && (dict==withPrefix64k || match >= lowPrefix) ) {
-                /* Copy the match. */
-                memcpy(op + 0, match + 0, 8);
-                memcpy(op + 8, match + 8, 8);
-                memcpy(op +16, match +16, 2);
-                op += length + MINMATCH;
-                /* Both stages worked, load the next token. */
+            if ((checkOffset) && (unlikely(match + dictSize < lowPrefix))) goto _output_error;   /* Error : offset outside buffers */
+
+            if (length == ML_MASK) {
+              variable_length_error error = ok;
+              length += read_variable_length(&ip, iend - LASTLITERALS + 1, endOnInput, 0, &error);
+              if (error != ok) goto _output_error;
+                if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)op)) goto _output_error;   /* overflow detection */
+                length += MINMATCH;
+                if (op + length >= oend - FASTLOOP_SAFE_DISTANCE) {
+                    goto safe_match_copy;
+                }
+            } else {
+                length += MINMATCH;
+                if (op + length >= oend - FASTLOOP_SAFE_DISTANCE) {
+                    goto safe_match_copy;
+                }
+
+                /* Fastpath check: Avoids a branch in LZ4_wildCopy32 if true */
+                if (!(dict == usingExtDict) || (match >= lowPrefix)) {
+                    if (offset >= 8) {
+                        memcpy(op, match, 8);
+                        memcpy(op+8, match+8, 8);
+                        memcpy(op+16, match+16, 2);
+                        op += length;
+                        continue;
+            }   }   }
+
+            /* match starting within external dictionary */
+            if ((dict==usingExtDict) && (match < lowPrefix)) {
+                if (unlikely(op+length > oend-LASTLITERALS)) {
+                    if (partialDecoding) length = MIN(length, (size_t)(oend-op));
+                    else goto _output_error;   /* doesn't respect parsing restriction */
+                }
+
+                if (length <= (size_t)(lowPrefix-match)) {
+                    /* match fits entirely within external dictionary : just copy */
+                    memmove(op, dictEnd - (lowPrefix-match), length);
+                    op += length;
+                } else {
+                    /* match stretches into both external dictionary and current block */
+                    size_t const copySize = (size_t)(lowPrefix - match);
+                    size_t const restSize = length - copySize;
+                    memcpy(op, dictEnd - copySize, copySize);
+                    op += copySize;
+                    if (restSize > (size_t)(op - lowPrefix)) {  /* overlap copy */
+                        BYTE* const endOfMatch = op + restSize;
+                        const BYTE* copyFrom = lowPrefix;
+                        while (op < endOfMatch) *op++ = *copyFrom++;
+                    } else {
+                        memcpy(op, lowPrefix, restSize);
+                        op += restSize;
+                }   }
                 continue;
             }
 
-            /* The second stage didn't work out, but the info is ready.
-             * Propel it right to the point of match copying. */
-            goto _copy_match;
-        }
+            /* copy match within block */
+            cpy = op + length;
 
-        /* decode literal length */
-        if (length == RUN_MASK) {
-            unsigned s;
-            if (unlikely(endOnInput ? ip >= iend-RUN_MASK : 0)) goto _output_error;   /* overflow detection */
-            do {
-                s = *ip++;
-                length += s;
-            } while ( likely(endOnInput ? ip<iend-RUN_MASK : 1) & (s==255) );
-            if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)(op))) goto _output_error;   /* overflow detection */
-            if ((safeDecode) && unlikely((uptrval)(ip)+length<(uptrval)(ip))) goto _output_error;   /* overflow detection */
+            assert((op <= oend) && (oend-op >= 32));
+            if (unlikely(offset<16)) {
+                LZ4_memcpy_using_offset(op, match, cpy, offset);
+            } else {
+                LZ4_wildCopy32(op, match, cpy);
+            }
+
+            op = cpy;   /* wildcopy correction */
         }
+    safe_decode:
+#endif
 
-        /* copy literals */
-        cpy = op+length;
-        LZ4_STATIC_ASSERT(MFLIMIT >= WILDCOPYLENGTH);
-        if ( ((endOnInput) && ((cpy>oend-MFLIMIT) || (ip+length>iend-(2+1+LASTLITERALS))) )
-          || ((!endOnInput) && (cpy>oend-WILDCOPYLENGTH)) )
-        {
-            if (partialDecoding) {
-                if (cpy > oend) { cpy = oend; length = oend-op; }             /* Partial decoding : stop in the middle of literal segment */
-                if ((endOnInput) && (ip+length > iend)) goto _output_error;   /* Error : read attempt beyond end of input buffer */
-            } else {
-                if ((!endOnInput) && (cpy != oend)) goto _output_error;       /* Error : block decoding must stop exactly there */
-                if ((endOnInput) && ((ip+length != iend) || (cpy > oend))) goto _output_error;   /* Error : input must be consumed */
+        /* Main Loop : decode remaining sequences where output < FASTLOOP_SAFE_DISTANCE */
+        while (1) {
+            token = *ip++;
+            length = token >> ML_BITS;  /* literal length */
+
+            assert(!endOnInput || ip <= iend); /* ip < iend before the increment */
+
+            /* A two-stage shortcut for the most common case:
+             * 1) If the literal length is 0..14, and there is enough space,
+             * enter the shortcut and copy 16 bytes on behalf of the literals
+             * (in the fast mode, only 8 bytes can be safely copied this way).
+             * 2) Further if the match length is 4..18, copy 18 bytes in a similar
+             * manner; but we ensure that there's enough space in the output for
+             * those 18 bytes earlier, upon entering the shortcut (in other words,
+             * there is a combined check for both stages).
+             */
+            if ( (endOnInput ? length != RUN_MASK : length <= 8)
+                /* strictly "less than" on input, to re-enter the loop with at least one byte */
+              && likely((endOnInput ? ip < shortiend : 1) & (op <= shortoend)) ) {
+                /* Copy the literals */
+                memcpy(op, ip, endOnInput ? 16 : 8);
+                op += length; ip += length;
+
+                /* The second stage: prepare for match copying, decode full info.
+                 * If it doesn't work out, the info won't be wasted. */
+                length = token & ML_MASK; /* match length */
+                offset = LZ4_readLE16(ip); ip += 2;
+                match = op - offset;
+                assert(match <= op); /* check overflow */
+
+                /* Do not deal with overlapping matches. */
+                if ( (length != ML_MASK)
+                  && (offset >= 8)
+                  && (dict==withPrefix64k || match >= lowPrefix) ) {
+                    /* Copy the match. */
+                    memcpy(op + 0, match + 0, 8);
+                    memcpy(op + 8, match + 8, 8);
+                    memcpy(op +16, match +16, 2);
+                    op += length + MINMATCH;
+                    /* Both stages worked, load the next token. */
+                    continue;
+                }
+
+                /* The second stage didn't work out, but the info is ready.
+                 * Propel it right to the point of match copying. */
+                goto _copy_match;
             }
-            memcpy(op, ip, length);
-            ip += length;
-            op += length;
-            if (!partialDecoding || (cpy == oend)) {
-                /* Necessarily EOF, due to parsing restrictions */
-                break;
+
+            /* decode literal length */
+            if (length == RUN_MASK) {
+              variable_length_error error = ok;
+              length += read_variable_length(&ip, iend-RUN_MASK, endOnInput, endOnInput, &error);
+              if (error == initial_error) goto _output_error;
+                if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)(op))) goto _output_error;   /* overflow detection */
+                if ((safeDecode) && unlikely((uptrval)(ip)+length<(uptrval)(ip))) goto _output_error;   /* overflow detection */
             }
 
-        } else {
-            LZ4_wildCopy(op, ip, cpy);   /* may overwrite up to WILDCOPYLENGTH beyond cpy */
-            ip += length; op = cpy;
-        }
+            /* copy literals */
+            cpy = op+length;
+#if LZ4_FAST_DEC_LOOP
+        safe_literal_copy:
+#endif
+            LZ4_STATIC_ASSERT(MFLIMIT >= WILDCOPYLENGTH);
+            if ( ((endOnInput) && ((cpy>oend-MFLIMIT) || (ip+length>iend-(2+1+LASTLITERALS))) )
+              || ((!endOnInput) && (cpy>oend-WILDCOPYLENGTH)) )
+            {
+                if (partialDecoding) {
+                    if (cpy > oend) { cpy = oend; assert(op<=oend); length = (size_t)(oend-op); }  /* Partial decoding : stop in the middle of literal segment */
+                    if ((endOnInput) && (ip+length > iend)) goto _output_error;   /* Error : read attempt beyond end of input buffer */
+                } else {
+                    if ((!endOnInput) && (cpy != oend)) goto _output_error;       /* Error : block decoding must stop exactly there */
+                    if ((endOnInput) && ((ip+length != iend) || (cpy > oend))) goto _output_error;   /* Error : input must be consumed */
+                }
+                memcpy(op, ip, length);
+                ip += length;
+                op += length;
+                if (!partialDecoding || (cpy == oend)) {
+                    /* Necessarily EOF, due to parsing restrictions */
+                    break;
+                }
 
-        /* get offset */
-        offset = LZ4_readLE16(ip); ip+=2;
-        match = op - offset;
+            } else {
+                LZ4_wildCopy(op, ip, cpy);   /* may overwrite up to WILDCOPYLENGTH beyond cpy */
+                ip += length; op = cpy;
+            }
 
-        /* get matchlength */
-        length = token & ML_MASK;
+            /* get offset */
+            offset = LZ4_readLE16(ip); ip+=2;
+            match = op - offset;
 
-_copy_match:
-        if ((checkOffset) && (unlikely(match + dictSize < lowPrefix))) goto _output_error;   /* Error : offset outside buffers */
-        if (!partialDecoding) {
-            assert(oend > op);
-            assert(oend - op >= 4);
-            LZ4_write32(op, 0);   /* silence an msan warning when offset==0; costs <1%; */
-        }   /* note : when partialDecoding, there is no guarantee that at least 4 bytes remain available in output buffer */
+            /* get matchlength */
+            length = token & ML_MASK;
+
+    _copy_match:
+            if ((checkOffset) && (unlikely(match + dictSize < lowPrefix))) goto _output_error;   /* Error : offset outside buffers */
+            if (!partialDecoding) {
+                assert(oend > op);
+                assert(oend - op >= 4);
+                LZ4_write32(op, 0);   /* silence an msan warning when offset==0; costs <1%; */
+            }   /* note : when partialDecoding, there is no guarantee that at least 4 bytes remain available in output buffer */
+
+            if (length == ML_MASK) {
+              variable_length_error error = ok;
+              length += read_variable_length(&ip, iend - LASTLITERALS + 1, endOnInput, 0, &error);
+              if (error != ok) goto _output_error;
+                if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)op)) goto _output_error;   /* overflow detection */
+            }
+            length += MINMATCH;
 
-        if (length == ML_MASK) {
-            unsigned s;
-            do {
-                s = *ip++;
-                if ((endOnInput) && (ip > iend-LASTLITERALS)) goto _output_error;
-                length += s;
-            } while (s==255);
-            if ((safeDecode) && unlikely((uptrval)(op)+length<(uptrval)op)) goto _output_error;   /* overflow detection */
-        }
-        length += MINMATCH;
+#if LZ4_FAST_DEC_LOOP
+        safe_match_copy:
+#endif
+            /* match starting within external dictionary */
+            if ((dict==usingExtDict) && (match < lowPrefix)) {
+                if (unlikely(op+length > oend-LASTLITERALS)) {
+                    if (partialDecoding) length = MIN(length, (size_t)(oend-op));
+                    else goto _output_error;   /* doesn't respect parsing restriction */
+                }
 
-        /* match starting within external dictionary */
-        if ((dict==usingExtDict) && (match < lowPrefix)) {
-            if (unlikely(op+length > oend-LASTLITERALS)) {
-                if (partialDecoding) length = MIN(length, (size_t)(oend-op));
-                else goto _output_error;   /* doesn't respect parsing restriction */
+                if (length <= (size_t)(lowPrefix-match)) {
+                    /* match fits entirely within external dictionary : just copy */
+                    memmove(op, dictEnd - (lowPrefix-match), length);
+                    op += length;
+                } else {
+                    /* match stretches into both external dictionary and current block */
+                    size_t const copySize = (size_t)(lowPrefix - match);
+                    size_t const restSize = length - copySize;
+                    memcpy(op, dictEnd - copySize, copySize);
+                    op += copySize;
+                    if (restSize > (size_t)(op - lowPrefix)) {  /* overlap copy */
+                        BYTE* const endOfMatch = op + restSize;
+                        const BYTE* copyFrom = lowPrefix;
+                        while (op < endOfMatch) *op++ = *copyFrom++;
+                    } else {
+                        memcpy(op, lowPrefix, restSize);
+                        op += restSize;
+                }   }
+                continue;
             }
 
-            if (length <= (size_t)(lowPrefix-match)) {
-                /* match fits entirely within external dictionary : just copy */
-                memmove(op, dictEnd - (lowPrefix-match), length);
-                op += length;
-            } else {
-                /* match stretches into both external dictionary and current block */
-                size_t const copySize = (size_t)(lowPrefix - match);
-                size_t const restSize = length - copySize;
-                memcpy(op, dictEnd - copySize, copySize);
-                op += copySize;
-                if (restSize > (size_t)(op - lowPrefix)) {  /* overlap copy */
-                    BYTE* const endOfMatch = op + restSize;
-                    const BYTE* copyFrom = lowPrefix;
-                    while (op < endOfMatch) *op++ = *copyFrom++;
+            /* copy match within block */
+            cpy = op + length;
+
+            /* partialDecoding : may end anywhere within the block */
+            assert(op<=oend);
+            if (partialDecoding && (cpy > oend-MATCH_SAFEGUARD_DISTANCE)) {
+                size_t const mlen = MIN(length, (size_t)(oend-op));
+                const BYTE* const matchEnd = match + mlen;
+                BYTE* const copyEnd = op + mlen;
+                if (matchEnd > op) {   /* overlap copy */
+                    while (op < copyEnd) *op++ = *match++;
                 } else {
-                    memcpy(op, lowPrefix, restSize);
-                    op += restSize;
-            }   }
-            continue;
-        }
+                    memcpy(op, match, mlen);
+                }
+                op = copyEnd;
+                if (op==oend) break;
+                continue;
+            }
 
-        /* copy match within block */
-        cpy = op + length;
-
-        /* partialDecoding : may not respect endBlock parsing restrictions */
-        assert(op<=oend);
-        if (partialDecoding && (cpy > oend-MATCH_SAFEGUARD_DISTANCE)) {
-            size_t const mlen = MIN(length, (size_t)(oend-op));
-            const BYTE* const matchEnd = match + mlen;
-            BYTE* const copyEnd = op + mlen;
-            if (matchEnd > op) {   /* overlap copy */
-                while (op < copyEnd) *op++ = *match++;
+            if (unlikely(offset<8)) {
+                op[0] = match[0];
+                op[1] = match[1];
+                op[2] = match[2];
+                op[3] = match[3];
+                match += inc32table[offset];
+                memcpy(op+4, match, 4);
+                match -= dec64table[offset];
             } else {
-                memcpy(op, match, mlen);
+                memcpy(op, match, 8);
+                match += 8;
             }
-            op = copyEnd;
-            if (op==oend) break;
-            continue;
-        }
-
-        if (unlikely(offset<8)) {
-            op[0] = match[0];
-            op[1] = match[1];
-            op[2] = match[2];
-            op[3] = match[3];
-            match += inc32table[offset];
-            memcpy(op+4, match, 4);
-            match -= dec64table[offset];
-        } else {
-            memcpy(op, match, 8);
-            match += 8;
-        }
-        op += 8;
-
-        if (unlikely(cpy > oend-MATCH_SAFEGUARD_DISTANCE)) {
-            BYTE* const oCopyLimit = oend - (WILDCOPYLENGTH-1);
-            if (cpy > oend-LASTLITERALS) goto _output_error;    /* Error : last LASTLITERALS bytes must be literals (uncompressed) */
-            if (op < oCopyLimit) {
-                LZ4_wildCopy(op, match, oCopyLimit);
-                match += oCopyLimit - op;
-                op = oCopyLimit;
+            op += 8;
+
+            if (unlikely(cpy > oend-MATCH_SAFEGUARD_DISTANCE)) {
+                BYTE* const oCopyLimit = oend - (WILDCOPYLENGTH-1);
+                if (cpy > oend-LASTLITERALS) goto _output_error;    /* Error : last LASTLITERALS bytes must be literals (uncompressed) */
+                if (op < oCopyLimit) {
+                    LZ4_wildCopy(op, match, oCopyLimit);
+                    match += oCopyLimit - op;
+                    op = oCopyLimit;
+                }
+                while (op < cpy) *op++ = *match++;
+            } else {
+                memcpy(op, match, 8);
+                if (length > 16) LZ4_wildCopy(op+8, match+8, cpy);
             }
-            while (op < cpy) *op++ = *match++;
-        } else {
-            memcpy(op, match, 8);
-            if (length > 16) LZ4_wildCopy(op+8, match+8, cpy);
+            op = cpy;   /* wildcopy correction */
         }
-        op = cpy;   /* wildcopy correction */
-    }
 
-    /* end of decoding */
-    if (endOnInput)
-       return (int) (((char*)op)-dst);     /* Nb of output bytes decoded */
-    else
-       return (int) (((const char*)ip)-src);   /* Nb of input bytes read */
+        /* end of decoding */
+        if (endOnInput)
+           return (int) (((char*)op)-dst);     /* Nb of output bytes decoded */
+        else
+           return (int) (((const char*)ip)-src);   /* Nb of input bytes read */
 
-    /* Overflow error detected */
-_output_error:
-    return (int) (-(((const char*)ip)-src))-1;
+        /* Overflow error detected */
+    _output_error:
+        return (int) (-(((const char*)ip)-src))-1;
+    }
 }
 
 
@@ -1745,12 +2058,13 @@ int LZ4_decompress_fast_doubleDict(const char* source, char* dest, int originalS
 LZ4_streamDecode_t* LZ4_createStreamDecode(void)
 {
     LZ4_streamDecode_t* lz4s = (LZ4_streamDecode_t*) ALLOC_AND_ZERO(sizeof(LZ4_streamDecode_t));
+    LZ4_STATIC_ASSERT(LZ4_STREAMDECODESIZE >= sizeof(LZ4_streamDecode_t_internal));    /* A compilation error here means LZ4_STREAMDECODESIZE is not large enough */
     return lz4s;
 }
 
 int LZ4_freeStreamDecode (LZ4_streamDecode_t* LZ4_stream)
 {
-    if (!LZ4_stream) return 0;   /* support free on NULL */
+    if (LZ4_stream == NULL) return 0;   /* support free on NULL */
     FREEMEM(LZ4_stream);
     return 0;
 }
@@ -1808,7 +2122,7 @@ int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
         assert(lz4sd->extDictSize == 0);
         result = LZ4_decompress_safe(source, dest, compressedSize, maxOutputSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize = result;
+        lz4sd->prefixSize = (size_t)result;
         lz4sd->prefixEnd = (BYTE*)dest + result;
     } else if (lz4sd->prefixEnd == (BYTE*)dest) {
         /* They're rolling the current segment. */
@@ -1821,7 +2135,7 @@ int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
             result = LZ4_decompress_safe_doubleDict(source, dest, compressedSize, maxOutputSize,
                                                     lz4sd->prefixSize, lz4sd->externalDict, lz4sd->extDictSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize += result;
+        lz4sd->prefixSize += (size_t)result;
         lz4sd->prefixEnd  += result;
     } else {
         /* The buffer wraps around, or they're switching to another buffer. */
@@ -1830,7 +2144,7 @@ int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
         result = LZ4_decompress_safe_forceExtDict(source, dest, compressedSize, maxOutputSize,
                                                   lz4sd->externalDict, lz4sd->extDictSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize = result;
+        lz4sd->prefixSize = (size_t)result;
         lz4sd->prefixEnd  = (BYTE*)dest + result;
     }
 
@@ -1842,12 +2156,13 @@ int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
 {
     LZ4_streamDecode_t_internal* lz4sd = &LZ4_streamDecode->internal_donotuse;
     int result;
+    assert(originalSize >= 0);
 
     if (lz4sd->prefixSize == 0) {
         assert(lz4sd->extDictSize == 0);
         result = LZ4_decompress_fast(source, dest, originalSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize = originalSize;
+        lz4sd->prefixSize = (size_t)originalSize;
         lz4sd->prefixEnd = (BYTE*)dest + originalSize;
     } else if (lz4sd->prefixEnd == (BYTE*)dest) {
         if (lz4sd->prefixSize >= 64 KB - 1 || lz4sd->extDictSize == 0)
@@ -1856,7 +2171,7 @@ int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
             result = LZ4_decompress_fast_doubleDict(source, dest, originalSize,
                                                     lz4sd->prefixSize, lz4sd->externalDict, lz4sd->extDictSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize += originalSize;
+        lz4sd->prefixSize += (size_t)originalSize;
         lz4sd->prefixEnd  += originalSize;
     } else {
         lz4sd->extDictSize = lz4sd->prefixSize;
@@ -1864,7 +2179,7 @@ int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const ch
         result = LZ4_decompress_fast_extDict(source, dest, originalSize,
                                              lz4sd->externalDict, lz4sd->extDictSize);
         if (result <= 0) return result;
-        lz4sd->prefixSize = originalSize;
+        lz4sd->prefixSize = (size_t)originalSize;
         lz4sd->prefixEnd  = (BYTE*)dest + originalSize;
     }
 
diff --git a/lib/lz4.h b/lib/lz4.h
index 059ef7c..1589be9 100644
--- a/lib/lz4.h
+++ b/lib/lz4.h
@@ -51,19 +51,23 @@ extern "C" {
   multiple GB/s per core, typically reaching RAM speed limits on multi-core systems.
 
   The LZ4 compression library provides in-memory compression and decompression functions.
+  It gives full buffer control to user.
   Compression can be done in:
     - a single step (described as Simple Functions)
     - a single step, reusing a context (described in Advanced Functions)
     - unbounded multiple steps (described as Streaming compression)
 
-  lz4.h provides block compression functions. It gives full buffer control to user.
-  Decompressing an lz4-compressed block also requires metadata (such as compressed size).
-  Each application is free to encode such metadata in whichever way it wants.
+  lz4.h generates and decodes LZ4-compressed blocks (doc/lz4_Block_format.md).
+  Decompressing a block requires additional metadata, such as its compressed size.
+  Each application is free to encode and pass such metadata in whichever way it wants.
 
-  An additional format, called LZ4 frame specification (doc/lz4_Frame_format.md),
-  take care of encoding standard metadata alongside LZ4-compressed blocks.
-  Frame format is required for interoperability.
-  It is delivered through a companion API, declared in lz4frame.h.
+  lz4.h only handle blocks, it can not generate Frames.
+
+  Blocks are different from Frames (doc/lz4_Frame_format.md).
+  Frames bundle both blocks and metadata in a specified manner.
+  This are required for compressed data to be self-contained and portable.
+  Frame format is delivered through a companion API, declared in lz4frame.h.
+  Note that the `lz4` CLI can only manage frames.
 */
 
 /*^***************************************************************
@@ -92,8 +96,8 @@ extern "C" {
 
 /*------   Version   ------*/
 #define LZ4_VERSION_MAJOR    1    /* for breaking interface changes  */
-#define LZ4_VERSION_MINOR    8    /* for new (non-breaking) interface capabilities */
-#define LZ4_VERSION_RELEASE  3    /* for tweaks, bug-fixes, or development */
+#define LZ4_VERSION_MINOR    9    /* for new (non-breaking) interface capabilities */
+#define LZ4_VERSION_RELEASE  0    /* for tweaks, bug-fixes, or development */
 
 #define LZ4_VERSION_NUMBER (LZ4_VERSION_MAJOR *100*100 + LZ4_VERSION_MINOR *100 + LZ4_VERSION_RELEASE)
 
@@ -103,7 +107,7 @@ extern "C" {
 #define LZ4_VERSION_STRING LZ4_EXPAND_AND_QUOTE(LZ4_LIB_VERSION)
 
 LZ4LIB_API int LZ4_versionNumber (void);  /**< library version number; useful to check dll version */
-LZ4LIB_API const char* LZ4_versionString (void);   /**< library version string; unseful to check dll version */
+LZ4LIB_API const char* LZ4_versionString (void);   /**< library version string; useful to check dll version */
 
 
 /*-************************************
@@ -112,14 +116,15 @@ LZ4LIB_API const char* LZ4_versionString (void);   /**< library version string;
 /*!
  * LZ4_MEMORY_USAGE :
  * Memory usage formula : N->2^N Bytes (examples : 10 -> 1KB; 12 -> 4KB ; 16 -> 64KB; 20 -> 1MB; etc.)
- * Increasing memory usage improves compression ratio
- * Reduced memory usage may improve speed, thanks to cache effect
+ * Increasing memory usage improves compression ratio.
+ * Reduced memory usage may improve speed, thanks to better cache locality.
  * Default value is 14, for 16KB, which nicely fits into Intel x86 L1 cache
  */
 #ifndef LZ4_MEMORY_USAGE
 # define LZ4_MEMORY_USAGE 14
 #endif
 
+
 /*-************************************
 *  Simple Functions
 **************************************/
@@ -130,21 +135,22 @@ LZ4LIB_API const char* LZ4_versionString (void);   /**< library version string;
     It also runs faster, so it's a recommended setting.
     If the function cannot compress 'src' into a more limited 'dst' budget,
     compression stops *immediately*, and the function result is zero.
-    Note : as a consequence, 'dst' content is not valid.
-    Note 2 : This function is protected against buffer overflow scenarios (never writes outside 'dst' buffer, nor read outside 'source' buffer).
+    In which case, 'dst' content is undefined (invalid).
         srcSize : max supported value is LZ4_MAX_INPUT_SIZE.
         dstCapacity : size of buffer 'dst' (which must be already allocated)
-        return  : the number of bytes written into buffer 'dst' (necessarily <= dstCapacity)
-                  or 0 if compression fails */
+       @return  : the number of bytes written into buffer 'dst' (necessarily <= dstCapacity)
+                  or 0 if compression fails
+    Note : This function is protected against buffer overflow scenarios (never writes outside 'dst' buffer, nor read outside 'source' buffer).
+*/
 LZ4LIB_API int LZ4_compress_default(const char* src, char* dst, int srcSize, int dstCapacity);
 
 /*! LZ4_decompress_safe() :
     compressedSize : is the exact complete size of the compressed block.
     dstCapacity : is the size of destination buffer, which must be already allocated.
-    return : the number of bytes decompressed into destination buffer (necessarily <= dstCapacity)
+   @return : the number of bytes decompressed into destination buffer (necessarily <= dstCapacity)
              If destination buffer is not large enough, decoding will stop and output an error code (negative value).
              If the source stream is detected malformed, the function will stop decoding and return a negative result.
-             This function is protected against malicious data packets.
+    Note : This function is protected against malicious data packets (never writes outside 'dst' buffer, nor read outside 'source' buffer).
 */
 LZ4LIB_API int LZ4_decompress_safe (const char* src, char* dst, int compressedSize, int dstCapacity);
 
@@ -155,8 +161,7 @@ LZ4LIB_API int LZ4_decompress_safe (const char* src, char* dst, int compressedSi
 #define LZ4_MAX_INPUT_SIZE        0x7E000000   /* 2 113 929 216 bytes */
 #define LZ4_COMPRESSBOUND(isize)  ((unsigned)(isize) > (unsigned)LZ4_MAX_INPUT_SIZE ? 0 : (isize) + ((isize)/255) + 16)
 
-/*!
-LZ4_compressBound() :
+/*! LZ4_compressBound() :
     Provides the maximum size that LZ4 compression may output in a "worst case" scenario (input data not compressible)
     This function is primarily useful for memory allocation purposes (destination buffer size).
     Macro LZ4_COMPRESSBOUND() is also provided for compilation-time evaluation (stack memory allocation for example).
@@ -167,8 +172,7 @@ LZ4_compressBound() :
 */
 LZ4LIB_API int LZ4_compressBound(int inputSize);
 
-/*!
-LZ4_compress_fast() :
+/*! LZ4_compress_fast() :
     Same as LZ4_compress_default(), but allows selection of "acceleration" factor.
     The larger the acceleration value, the faster the algorithm, but also the lesser the compression.
     It's a trade-off. It can be fine tuned, with each successive value providing roughly +~3% to speed.
@@ -178,13 +182,12 @@ LZ4_compress_fast() :
 LZ4LIB_API int LZ4_compress_fast (const char* src, char* dst, int srcSize, int dstCapacity, int acceleration);
 
 
-/*!
-LZ4_compress_fast_extState() :
-    Same compression function, just using an externally allocated memory space to store compression state.
-    Use LZ4_sizeofState() to know how much memory must be allocated,
-    and allocate it on 8-bytes boundaries (using malloc() typically).
-    Then, provide this buffer as 'void* state' to compression function.
-*/
+/*! LZ4_compress_fast_extState() :
+ *  Same as LZ4_compress_fast(), using an externally allocated memory space for its state.
+ *  Use LZ4_sizeofState() to know how much memory must be allocated,
+ *  and allocate it on 8-bytes boundaries (using `malloc()` typically).
+ *  Then, provide this buffer as `void* state` to compression function.
+ */
 LZ4LIB_API int LZ4_sizeofState(void);
 LZ4LIB_API int LZ4_compress_fast_extState (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration);
 
@@ -204,27 +207,6 @@ LZ4LIB_API int LZ4_compress_fast_extState (void* state, const char* src, char* d
 LZ4LIB_API int LZ4_compress_destSize (const char* src, char* dst, int* srcSizePtr, int targetDstSize);
 
 
-/*! LZ4_decompress_fast() : **unsafe!**
- *  This function used to be a bit faster than LZ4_decompress_safe(),
- *  though situation has changed in recent versions,
- *  and now `LZ4_decompress_safe()` can be as fast and sometimes faster than `LZ4_decompress_fast()`.
- *  Moreover, LZ4_decompress_fast() is not protected vs malformed input, as it doesn't perform full validation of compressed data.
- *  As a consequence, this function is no longer recommended, and may be deprecated in future versions.
- *  It's only remaining specificity is that it can decompress data without knowing its compressed size.
- *
- *  originalSize : is the uncompressed size to regenerate.
- *                 `dst` must be already allocated, its size must be >= 'originalSize' bytes.
- * @return : number of bytes read from source buffer (== compressed size).
- *           If the source stream is detected malformed, the function stops decoding and returns a negative result.
- *  note : This function requires uncompressed originalSize to be known in advance.
- *         The function never writes past the output buffer.
- *         However, since it doesn't know its 'src' size, it may read past the intended input.
- *         Also, because match offsets are not validated during decoding,
- *         reads from 'src' may underflow.
- *         Use this function in trusted environment **only**.
- */
-LZ4LIB_API int LZ4_decompress_fast (const char* src, char* dst, int originalSize);
-
 /*! LZ4_decompress_safe_partial() :
  *  Decompress an LZ4 compressed block, of size 'srcSize' at position 'src',
  *  into destination buffer 'dst' of size 'dstCapacity'.
@@ -257,30 +239,49 @@ LZ4LIB_API int LZ4_decompress_safe_partial (const char* src, char* dst, int srcS
 ***********************************************/
 typedef union LZ4_stream_u LZ4_stream_t;  /* incomplete type (defined later) */
 
-/*! LZ4_createStream() and LZ4_freeStream() :
- *  LZ4_createStream() will allocate and initialize an `LZ4_stream_t` structure.
- *  LZ4_freeStream() releases its memory.
- */
 LZ4LIB_API LZ4_stream_t* LZ4_createStream(void);
 LZ4LIB_API int           LZ4_freeStream (LZ4_stream_t* streamPtr);
 
-/*! LZ4_resetStream() :
- *  An LZ4_stream_t structure can be allocated once and re-used multiple times.
- *  Use this function to start compressing a new stream.
+/*! LZ4_resetStream_fast() : v1.9.0+
+ *  Use this to prepare an LZ4_stream_t for a new chain of dependent blocks
+ *  (e.g., LZ4_compress_fast_continue()).
+ *
+ *  An LZ4_stream_t must be initialized once before usage.
+ *  This is automatically done when created by LZ4_createStream().
+ *  However, should the LZ4_stream_t be simply declared on stack (for example),
+ *  it's necessary to initialize it first, using LZ4_initStream().
+ *
+ *  After init, start any new stream with LZ4_resetStream_fast().
+ *  A same LZ4_stream_t can be re-used multiple times consecutively
+ *  and compress multiple streams,
+ *  provided that it starts each new stream with LZ4_resetStream_fast().
+ *
+ *  LZ4_resetStream_fast() is much faster than LZ4_initStream(),
+ *  but is not compatible with memory regions containing garbage data.
+ *
+ *  Note: it's only useful to call LZ4_resetStream_fast()
+ *        in the context of streaming compression.
+ *        The *extState* functions perform their own resets.
+ *        Invoking LZ4_resetStream_fast() before is redundant, and even counterproductive.
  */
-LZ4LIB_API void LZ4_resetStream (LZ4_stream_t* streamPtr);
+LZ4LIB_API void LZ4_resetStream_fast (LZ4_stream_t* streamPtr);
 
 /*! LZ4_loadDict() :
- *  Use this function to load a static dictionary into LZ4_stream_t.
- *  Any previous data will be forgotten, only 'dictionary' will remain in memory.
+ *  Use this function to reference a static dictionary into LZ4_stream_t.
+ *  The dictionary must remain available during compression.
+ *  LZ4_loadDict() triggers a reset, so any previous data will be forgotten.
+ *  The same dictionary will have to be loaded on decompression side for successful decoding.
+ *  Dictionary are useful for better compression of small data (KB range).
+ *  While LZ4 accept any input as dictionary,
+ *  results are generally better when using Zstandard's Dictionary Builder.
  *  Loading a size of 0 is allowed, and is the same as reset.
- * @return : dictionary size, in bytes (necessarily <= 64 KB)
+ * @return : loaded dictionary size, in bytes (necessarily <= 64 KB)
  */
 LZ4LIB_API int LZ4_loadDict (LZ4_stream_t* streamPtr, const char* dictionary, int dictSize);
 
 /*! LZ4_compress_fast_continue() :
  *  Compress 'src' content using data from previously compressed blocks, for better compression ratio.
- *  'dst' buffer must be already allocated.
+ * 'dst' buffer must be already allocated.
  *  If dstCapacity >= LZ4_compressBound(srcSize), compression is guaranteed to succeed, and runs faster.
  *
  * @return : size of compressed block
@@ -288,10 +289,10 @@ LZ4LIB_API int LZ4_loadDict (LZ4_stream_t* streamPtr, const char* dictionary, in
  *
  *  Note 1 : Each invocation to LZ4_compress_fast_continue() generates a new block.
  *           Each block has precise boundaries.
+ *           Each block must be decompressed separately, calling LZ4_decompress_*() with relevant metadata.
  *           It's not possible to append blocks together and expect a single invocation of LZ4_decompress_*() to decompress them together.
- *           Each block must be decompressed separately, calling LZ4_decompress_*() with associated metadata.
  *
- *  Note 2 : The previous 64KB of source data is __assumed__ to remain present, unmodified, at same address in memory!
+ *  Note 2 : The previous 64KB of source data is __assumed__ to remain present, unmodified, at same address in memory !
  *
  *  Note 3 : When input is structured as a double-buffer, each buffer can have any size, including < 64 KB.
  *           Make sure that buffers are separated, by at least one byte.
@@ -299,7 +300,7 @@ LZ4LIB_API int LZ4_loadDict (LZ4_stream_t* streamPtr, const char* dictionary, in
  *
  *  Note 4 : If input buffer is a ring-buffer, it can have any size, including < 64 KB.
  *
- *  Note 5 : After an error, the stream status is invalid, it can only be reset or freed.
+ *  Note 5 : After an error, the stream status is undefined (invalid), it can only be reset or freed.
  */
 LZ4LIB_API int LZ4_compress_fast_continue (LZ4_stream_t* streamPtr, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration);
 
@@ -335,7 +336,7 @@ LZ4LIB_API int                 LZ4_freeStreamDecode (LZ4_streamDecode_t* LZ4_str
  */
 LZ4LIB_API int LZ4_setStreamDecode (LZ4_streamDecode_t* LZ4_streamDecode, const char* dictionary, int dictSize);
 
-/*! LZ4_decoderRingBufferSize() : v1.8.2
+/*! LZ4_decoderRingBufferSize() : v1.8.2+
  *  Note : in a ring buffer scenario (optional),
  *  blocks are presumed decompressed next to each other
  *  up to the moment there is not enough remaining space for next block (remainingSize < maxBlockSize),
@@ -347,7 +348,7 @@ LZ4LIB_API int LZ4_setStreamDecode (LZ4_streamDecode_t* LZ4_streamDecode, const
  *           or 0 if there is an error (invalid maxBlockSize).
  */
 LZ4LIB_API int LZ4_decoderRingBufferSize(int maxBlockSize);
-#define LZ4_DECODER_RING_BUFFER_SIZE(mbs) (65536 + 14 + (mbs))  /* for static allocation; mbs presumed valid */
+#define LZ4_DECODER_RING_BUFFER_SIZE(maxBlockSize) (65536 + 14 + (maxBlockSize))  /* for static allocation; maxBlockSize presumed valid */
 
 /*! LZ4_decompress_*_continue() :
  *  These decoding functions allow decompression of consecutive blocks in "streaming" mode.
@@ -375,83 +376,67 @@ LZ4LIB_API int LZ4_decoderRingBufferSize(int maxBlockSize);
  *  then indicate where this data is saved using LZ4_setStreamDecode(), before decompressing next block.
 */
 LZ4LIB_API int LZ4_decompress_safe_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* src, char* dst, int srcSize, int dstCapacity);
-LZ4LIB_API int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* src, char* dst, int originalSize);
 
 
 /*! LZ4_decompress_*_usingDict() :
  *  These decoding functions work the same as
  *  a combination of LZ4_setStreamDecode() followed by LZ4_decompress_*_continue()
  *  They are stand-alone, and don't need an LZ4_streamDecode_t structure.
- *  Dictionary is presumed stable : it must remain accessible and unmodified during next decompression.
+ *  Dictionary is presumed stable : it must remain accessible and unmodified during decompression.
+ *  Performance tip : Decompression speed can be substantially increased
+ *                    when dst == dictStart + dictSize.
  */
 LZ4LIB_API int LZ4_decompress_safe_usingDict (const char* src, char* dst, int srcSize, int dstCapcity, const char* dictStart, int dictSize);
-LZ4LIB_API int LZ4_decompress_fast_usingDict (const char* src, char* dst, int originalSize, const char* dictStart, int dictSize);
 
 
-/*^**********************************************
+/*^*************************************
  * !!!!!!   STATIC LINKING ONLY   !!!!!!
- ***********************************************/
+ ***************************************/
 
-/*-************************************
- *  Unstable declarations
- **************************************
- * Declarations in this section should be considered unstable.
- * Use at your own peril, etc., etc.
- * They may be removed in the future.
- * Their signatures may change.
- **************************************/
+/*-****************************************************************************
+ * Experimental section
+ *
+ * Symbols declared in this section must be considered unstable. Their
+ * signatures or semantics may change, or they may be removed altogether in the
+ * future. They are therefore only safe to depend on when the caller is
+ * statically linked against the library.
+ *
+ * To protect against unsafe usage, not only are the declarations guarded,
+ * the definitions are hidden by default
+ * when building LZ4 as a shared/dynamic library.
+ *
+ * In order to access these declarations,
+ * define LZ4_STATIC_LINKING_ONLY in your application
+ * before including LZ4's headers.
+ *
+ * In order to make their implementations accessible dynamically, you must
+ * define LZ4_PUBLISH_STATIC_FUNCTIONS when building the LZ4 library.
+ ******************************************************************************/
+
+#ifdef LZ4_PUBLISH_STATIC_FUNCTIONS
+#define LZ4LIB_STATIC_API LZ4LIB_API
+#else
+#define LZ4LIB_STATIC_API
+#endif
 
 #ifdef LZ4_STATIC_LINKING_ONLY
 
-/*! LZ4_resetStream_fast() :
- *  Use this, like LZ4_resetStream(), to prepare a context for a new chain of
- *  calls to a streaming API (e.g., LZ4_compress_fast_continue()).
- *
- *  Note:
- *  Using this in advance of a non- streaming-compression function is redundant,
- *  and potentially bad for performance, since they all perform their own custom
- *  reset internally.
- *
- *  Differences from LZ4_resetStream():
- *  When an LZ4_stream_t is known to be in a internally coherent state,
- *  it can often be prepared for a new compression with almost no work, only
- *  sometimes falling back to the full, expensive reset that is always required
- *  when the stream is in an indeterminate state (i.e., the reset performed by
- *  LZ4_resetStream()).
- *
- *  LZ4_streams are guaranteed to be in a valid state when:
- *  - returned from LZ4_createStream()
- *  - reset by LZ4_resetStream()
- *  - memset(stream, 0, sizeof(LZ4_stream_t)), though this is discouraged
- *  - the stream was in a valid state and was reset by LZ4_resetStream_fast()
- *  - the stream was in a valid state and was then used in any compression call
- *    that returned success
- *  - the stream was in an indeterminate state and was used in a compression
- *    call that fully reset the state (e.g., LZ4_compress_fast_extState()) and
- *    that returned success
- *
- *  When a stream isn't known to be in a valid state, it is not safe to pass to
- *  any fastReset or streaming function. It must first be cleansed by the full
- *  LZ4_resetStream().
- */
-LZ4LIB_API void LZ4_resetStream_fast (LZ4_stream_t* streamPtr);
 
 /*! LZ4_compress_fast_extState_fastReset() :
  *  A variant of LZ4_compress_fast_extState().
  *
- *  Using this variant avoids an expensive initialization step. It is only safe
- *  to call if the state buffer is known to be correctly initialized already
- *  (see above comment on LZ4_resetStream_fast() for a definition of "correctly
- *  initialized"). From a high level, the difference is that this function
- *  initializes the provided state with a call to something like
- *  LZ4_resetStream_fast() while LZ4_compress_fast_extState() starts with a
- *  call to LZ4_resetStream().
+ *  Using this variant avoids an expensive initialization step.
+ *  It is only safe to call if the state buffer is known to be correctly initialized already
+ *  (see above comment on LZ4_resetStream_fast() for a definition of "correctly initialized").
+ *  From a high level, the difference is that
+ *  this function initializes the provided state with a call to something like LZ4_resetStream_fast()
+ *  while LZ4_compress_fast_extState() starts with a call to LZ4_resetStream().
  */
-LZ4LIB_API int LZ4_compress_fast_extState_fastReset (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration);
+LZ4LIB_STATIC_API int LZ4_compress_fast_extState_fastReset (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int acceleration);
 
 /*! LZ4_attach_dictionary() :
- *  This is an experimental API that allows for the efficient use of a
- *  static dictionary many times.
+ *  This is an experimental API that allows
+ *  efficient use of a static dictionary many times.
  *
  *  Rather than re-loading the dictionary buffer into a working context before
  *  each compression, or copying a pre-loaded dictionary's LZ4_stream_t into a
@@ -462,8 +447,8 @@ LZ4LIB_API int LZ4_compress_fast_extState_fastReset (void* state, const char* sr
  *  Currently, only streams which have been prepared by LZ4_loadDict() should
  *  be expected to work.
  *
- *  Alternatively, the provided dictionary stream pointer may be NULL, in which
- *  case any existing dictionary stream is unset.
+ *  Alternatively, the provided dictionaryStream may be NULL,
+ *  in which case any existing dictionary stream is unset.
  *
  *  If a dictionary is provided, it replaces any pre-existing stream history.
  *  The dictionary contents are the only history that can be referenced and
@@ -475,17 +460,18 @@ LZ4LIB_API int LZ4_compress_fast_extState_fastReset (void* state, const char* sr
  *  stream (and source buffer) must remain in-place / accessible / unchanged
  *  through the completion of the first compression call on the stream.
  */
-LZ4LIB_API void LZ4_attach_dictionary(LZ4_stream_t *working_stream, const LZ4_stream_t *dictionary_stream);
+LZ4LIB_STATIC_API void LZ4_attach_dictionary(LZ4_stream_t* workingStream, const LZ4_stream_t* dictionaryStream);
 
 #endif
 
-/*-************************************
- *  Private definitions
- **************************************
- * Do not use these definitions.
- * They are exposed to allow static allocation of `LZ4_stream_t` and `LZ4_streamDecode_t`.
- * Using these definitions will expose code to API and/or ABI break in future versions of the library.
- **************************************/
+
+/*-************************************************************
+ *  PRIVATE DEFINITIONS
+ **************************************************************
+ * Do not use these definitions directly.
+ * They are only exposed to allow static allocation of `LZ4_stream_t` and `LZ4_streamDecode_t`.
+ * Accessing members will expose code to API and/or ABI break in future versions of the library.
+ **************************************************************/
 #define LZ4_HASHLOG   (LZ4_MEMORY_USAGE-2)
 #define LZ4_HASHTABLESIZE (1 << LZ4_MEMORY_USAGE)
 #define LZ4_HASH_SIZE_U32 (1 << LZ4_HASHLOG)       /* required as macro for static allocation */
@@ -497,7 +483,7 @@ typedef struct LZ4_stream_t_internal LZ4_stream_t_internal;
 struct LZ4_stream_t_internal {
     uint32_t hashTable[LZ4_HASH_SIZE_U32];
     uint32_t currentOffset;
-    uint16_t initCheck;
+    uint16_t dirty;
     uint16_t tableType;
     const uint8_t* dictionary;
     const LZ4_stream_t_internal* dictCtx;
@@ -517,7 +503,7 @@ typedef struct LZ4_stream_t_internal LZ4_stream_t_internal;
 struct LZ4_stream_t_internal {
     unsigned int hashTable[LZ4_HASH_SIZE_U32];
     unsigned int currentOffset;
-    unsigned short initCheck;
+    unsigned short dirty;
     unsigned short tableType;
     const unsigned char* dictionary;
     const LZ4_stream_t_internal* dictCtx;
@@ -526,38 +512,54 @@ struct LZ4_stream_t_internal {
 
 typedef struct {
     const unsigned char* externalDict;
-    size_t extDictSize;
     const unsigned char* prefixEnd;
+    size_t extDictSize;
     size_t prefixSize;
 } LZ4_streamDecode_t_internal;
 
 #endif
 
-/*!
- * LZ4_stream_t :
- * information structure to track an LZ4 stream.
- * init this structure before first use.
- * note : only use in association with static linking !
- *        this definition is not API/ABI safe,
- *        it may change in a future version !
+/*! LZ4_stream_t :
+ *  information structure to track an LZ4 stream.
+ *  LZ4_stream_t can also be created using LZ4_createStream(), which is recommended.
+ *  The structure definition can be convenient for static allocation
+ *  (on stack, or as part of larger structure).
+ *  Init this structure with LZ4_initStream() before first use.
+ *  note : only use this definition in association with static linking !
+ *    this definition is not API/ABI safe, and may change in a future version.
  */
-#define LZ4_STREAMSIZE_U64 ((1 << (LZ4_MEMORY_USAGE-3)) + 4)
+#define LZ4_STREAMSIZE_U64 ((1 << (LZ4_MEMORY_USAGE-3)) + 4 + ((sizeof(void*)==16) ? 4 : 0) /*AS-400*/ )
 #define LZ4_STREAMSIZE     (LZ4_STREAMSIZE_U64 * sizeof(unsigned long long))
 union LZ4_stream_u {
     unsigned long long table[LZ4_STREAMSIZE_U64];
     LZ4_stream_t_internal internal_donotuse;
 } ;  /* previously typedef'd to LZ4_stream_t */
 
+/*! LZ4_initStream() : v1.9.0+
+ *  An LZ4_stream_t structure must be initialized at least once.
+ *  This is automatically done when invoking LZ4_createStream(),
+ *  but it's not when the structure is simply declared on stack (for example).
+ *
+ *  Use LZ4_initStream() to properly initialize a newly declared LZ4_stream_t.
+ *  It can also initialize any arbitrary buffer of sufficient size,
+ *  and will @return a pointer of proper type upon initialization.
+ *
+ *  Note : initialization fails if size and alignment conditions are not respected.
+ *         In which case, the function will @return NULL.
+ *  Note2: An LZ4_stream_t structure guarantees correct alignment and size.
+ *  Note3: Before v1.9.0, use LZ4_resetStream() instead
+ */
+LZ4LIB_API LZ4_stream_t* LZ4_initStream (void* buffer, size_t size);
 
-/*!
- * LZ4_streamDecode_t :
- * information structure to track an LZ4 stream during decompression.
- * init this structure  using LZ4_setStreamDecode (or memset()) before first use
- * note : only use in association with static linking !
- *        this definition is not API/ABI safe,
- *        and may change in a future version !
+
+/*! LZ4_streamDecode_t :
+ *  information structure to track an LZ4 stream during decompression.
+ *  init this structure  using LZ4_setStreamDecode() before first use.
+ *  note : only use in association with static linking !
+ *         this definition is not API/ABI safe,
+ *         and may change in a future version !
  */
-#define LZ4_STREAMDECODESIZE_U64  4
+#define LZ4_STREAMDECODESIZE_U64 (4 + ((sizeof(void*)==16) ? 2 : 0) /*AS-400*/ )
 #define LZ4_STREAMDECODESIZE     (LZ4_STREAMDECODESIZE_U64 * sizeof(unsigned long long))
 union LZ4_streamDecode_u {
     unsigned long long table[LZ4_STREAMDECODESIZE_U64];
@@ -570,11 +572,16 @@ union LZ4_streamDecode_u {
 **************************************/
 
 /*! Deprecation warnings
-   Should deprecation warnings be a problem,
-   it is generally possible to disable them,
-   typically with -Wno-deprecated-declarations for gcc
-   or _CRT_SECURE_NO_WARNINGS in Visual.
-   Otherwise, it's also possible to define LZ4_DISABLE_DEPRECATE_WARNINGS */
+ *
+ *  Deprecated functions make the compiler generate a warning when invoked.
+ *  This is meant to invite users to update their source code.
+ *  Should deprecation warnings be a problem, it is generally possible to disable them,
+ *  typically with -Wno-deprecated-declarations for gcc
+ *  or _CRT_SECURE_NO_WARNINGS in Visual.
+ *
+ *  Another method is to define LZ4_DISABLE_DEPRECATE_WARNINGS
+ *  before including the header file.
+ */
 #ifdef LZ4_DISABLE_DEPRECATE_WARNINGS
 #  define LZ4_DEPRECATED(message)   /* disable deprecation warnings */
 #else
@@ -594,8 +601,8 @@ union LZ4_streamDecode_u {
 #endif /* LZ4_DISABLE_DEPRECATE_WARNINGS */
 
 /* Obsolete compression functions */
-LZ4_DEPRECATED("use LZ4_compress_default() instead") LZ4LIB_API int LZ4_compress               (const char* source, char* dest, int sourceSize);
-LZ4_DEPRECATED("use LZ4_compress_default() instead") LZ4LIB_API int LZ4_compress_limitedOutput (const char* source, char* dest, int sourceSize, int maxOutputSize);
+LZ4_DEPRECATED("use LZ4_compress_default() instead")       LZ4LIB_API int LZ4_compress               (const char* source, char* dest, int sourceSize);
+LZ4_DEPRECATED("use LZ4_compress_default() instead")       LZ4LIB_API int LZ4_compress_limitedOutput (const char* source, char* dest, int sourceSize, int maxOutputSize);
 LZ4_DEPRECATED("use LZ4_compress_fast_extState() instead") LZ4LIB_API int LZ4_compress_withState               (void* state, const char* source, char* dest, int inputSize);
 LZ4_DEPRECATED("use LZ4_compress_fast_extState() instead") LZ4LIB_API int LZ4_compress_limitedOutput_withState (void* state, const char* source, char* dest, int inputSize, int maxOutputSize);
 LZ4_DEPRECATED("use LZ4_compress_fast_continue() instead") LZ4LIB_API int LZ4_compress_continue                (LZ4_stream_t* LZ4_streamPtr, const char* source, char* dest, int inputSize);
@@ -616,13 +623,54 @@ LZ4_DEPRECATED("use LZ4_decompress_safe() instead") LZ4LIB_API int LZ4_uncompres
  */
 LZ4_DEPRECATED("Use LZ4_createStream() instead") LZ4LIB_API void* LZ4_create (char* inputBuffer);
 LZ4_DEPRECATED("Use LZ4_createStream() instead") LZ4LIB_API int   LZ4_sizeofStreamState(void);
-LZ4_DEPRECATED("Use LZ4_resetStream() instead") LZ4LIB_API  int   LZ4_resetStreamState(void* state, char* inputBuffer);
-LZ4_DEPRECATED("Use LZ4_saveDict() instead") LZ4LIB_API     char* LZ4_slideInputBuffer (void* state);
+LZ4_DEPRECATED("Use LZ4_resetStream() instead")  LZ4LIB_API int   LZ4_resetStreamState(void* state, char* inputBuffer);
+LZ4_DEPRECATED("Use LZ4_saveDict() instead")     LZ4LIB_API char* LZ4_slideInputBuffer (void* state);
 
 /* Obsolete streaming decoding functions */
 LZ4_DEPRECATED("use LZ4_decompress_safe_usingDict() instead") LZ4LIB_API int LZ4_decompress_safe_withPrefix64k (const char* src, char* dst, int compressedSize, int maxDstSize);
 LZ4_DEPRECATED("use LZ4_decompress_fast_usingDict() instead") LZ4LIB_API int LZ4_decompress_fast_withPrefix64k (const char* src, char* dst, int originalSize);
 
+/*! LZ4_decompress_fast() : **unsafe!**
+ *  These functions are generally slightly faster than LZ4_decompress_safe(),
+ *  though the difference is small (generally ~5%).
+ *  However, the real cost is a risk :  LZ4_decompress_safe() is protected vs malformed input, while `LZ4_decompress_fast()` is not, making it a security liability.
+ *  As a consequence, LZ4_decompress_fast() is strongly discouraged, and deprecated.
+ *  These functions will generate a deprecation warning in the future.
+ *
+ *  Last LZ4_decompress_fast() specificity is that it can decompress a block without knowing its compressed size.
+ *  Note that even that functionality could be achieved in a more secure manner if need be,
+ *  though it would require new prototypes, and adaptation of the implementation to this new use case.
+ *
+ *  Parameters:
+ *  originalSize : is the uncompressed size to regenerate.
+ *                 `dst` must be already allocated, its size must be >= 'originalSize' bytes.
+ * @return : number of bytes read from source buffer (== compressed size).
+ *           The function expects to finish at block's end exactly.
+ *           If the source stream is detected malformed, the function stops decoding and returns a negative result.
+ *  note : LZ4_decompress_fast*() requires originalSize. Thanks to this information, it never writes past the output buffer.
+ *         However, since it doesn't know its 'src' size, it may read an unknown amount of input, past input buffer bounds.
+ *         Also, since match offsets are not validated, match reads from 'src' may underflow too.
+ *         These issues never happen if input (compressed) data is correct.
+ *         But they may happen if input data is invalid (error or intentional tampering).
+ *         As a consequence, use these functions in trusted environments with trusted data **only**.
+ */
+
+/* LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe() instead")  */
+LZ4LIB_API int LZ4_decompress_fast (const char* src, char* dst, int originalSize);
+/* LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_continue() instead") */
+LZ4LIB_API int LZ4_decompress_fast_continue (LZ4_streamDecode_t* LZ4_streamDecode, const char* src, char* dst, int originalSize);
+/* LZ4_DEPRECATED("This function is deprecated and unsafe. Consider using LZ4_decompress_safe_usingDict() instead") */
+LZ4LIB_API int LZ4_decompress_fast_usingDict (const char* src, char* dst, int originalSize, const char* dictStart, int dictSize);
+
+/*! LZ4_resetStream() :
+ *  An LZ4_stream_t structure must be initialized at least once.
+ *  This is done with LZ4_initStream(), or LZ4_resetStream().
+ *  Consider switching to LZ4_initStream(),
+ *  invoking LZ4_resetStream() will trigger deprecation warnings in the future.
+ */
+LZ4LIB_API void LZ4_resetStream (LZ4_stream_t* streamPtr);
+
+
 #endif /* LZ4_H_2983827168210 */
 
 
diff --git a/lib/lz4frame.c b/lib/lz4frame.c
index 08bf0fa..19efd0b 100644
--- a/lib/lz4frame.c
+++ b/lib/lz4frame.c
@@ -1,41 +1,44 @@
 /*
-LZ4 auto-framing library
-Copyright (C) 2011-2016, Yann Collet.
-
-BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
-
-Redistribution and use in source and binary forms, with or without
-modification, are permitted provided that the following conditions are
-met:
-
-* Redistributions of source code must retain the above copyright
-notice, this list of conditions and the following disclaimer.
-* Redistributions in binary form must reproduce the above
-copyright notice, this list of conditions and the following disclaimer
-in the documentation and/or other materials provided with the
-distribution.
-
-THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
-"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
-LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
-A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
-OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
-SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
-LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
-DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
-THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
-(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
-OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
-
-You can contact the author at :
-- LZ4 homepage : http://www.lz4.org
-- LZ4 source repository : https://github.com/lz4/lz4
-*/
+ * LZ4 auto-framing library
+ * Copyright (C) 2011-2016, Yann Collet.
+ *
+ * BSD 2-Clause License (http://www.opensource.org/licenses/bsd-license.php)
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are
+ * met:
+ *
+ * - Redistributions of source code must retain the above copyright
+ *   notice, this list of conditions and the following disclaimer.
+ * - Redistributions in binary form must reproduce the above
+ *   copyright notice, this list of conditions and the following disclaimer
+ *   in the documentation and/or other materials provided with the
+ *   distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+ * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
+ * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
+ * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * You can contact the author at :
+ * - LZ4 homepage : http://www.lz4.org
+ * - LZ4 source repository : https://github.com/lz4/lz4
+ */
 
 /* LZ4F is a stand-alone API to create LZ4-compressed Frames
-*  in full conformance with specification v1.5.0
-*  All related operations, including memory management, are handled by the library.
-* */
+ * in full conformance with specification v1.6.1 .
+ * This library rely upon memory management capabilities (malloc, free)
+ * provided either by <stdlib.h>,
+ * or redirected towards another library of user's choice
+ * (see Memory Routines below).
+ */
 
 
 /*-************************************
@@ -62,16 +65,27 @@ You can contact the author at :
 /*-************************************
 *  Memory routines
 **************************************/
+/*
+ * User may redirect invocations of
+ * malloc(), calloc() and free()
+ * towards another library or solution of their choice
+ * by modifying below section.
+ */
 #include <stdlib.h>   /* malloc, calloc, free */
-#define ALLOC(s)   malloc(s)
-#define ALLOC_AND_ZERO(s)   calloc(1,s)
-#define FREEMEM        free
+#define ALLOC(s)       malloc(s)
+#ifndef LZ4_SRC_INCLUDED   /* avoid redefinition when sources are coalesced */
+#  define ALLOC_AND_ZERO(s)  calloc(1,(s))
+#endif
+#define FREEMEM(p)     free(p)
+
 #include <string.h>   /* memset, memcpy, memmove */
-#define MEM_INIT       memset
+#ifndef LZ4_SRC_INCLUDED  /* avoid redefinition when sources are coalesced */
+#  define MEM_INIT       memset
+#endif
 
 
 /*-************************************
-*  Includes
+*  Library declarations
 **************************************/
 #define LZ4F_STATIC_LINKING_ONLY
 #include "lz4frame.h"
@@ -134,8 +148,8 @@ static U32 LZ4F_readLE32 (const void* src)
 {
     const BYTE* const srcPtr = (const BYTE*)src;
     U32 value32 = srcPtr[0];
-    value32 += (srcPtr[1]<<8);
-    value32 += (srcPtr[2]<<16);
+    value32 += ((U32)srcPtr[1])<< 8;
+    value32 += ((U32)srcPtr[2])<<16;
     value32 += ((U32)srcPtr[3])<<24;
     return value32;
 }
@@ -180,9 +194,11 @@ static void LZ4F_writeLE64 (void* dst, U64 value64)
 /*-************************************
 *  Constants
 **************************************/
-#define KB *(1<<10)
-#define MB *(1<<20)
-#define GB *(1<<30)
+#ifndef LZ4_SRC_INCLUDED   /* avoid double definition */
+#  define KB *(1<<10)
+#  define MB *(1<<20)
+#  define GB *(1<<30)
+#endif
 
 #define _1BIT  0x01
 #define _2BITS 0x03
@@ -195,9 +211,10 @@ static void LZ4F_writeLE64 (void* dst, U64 value64)
 #define LZ4F_BLOCKUNCOMPRESSED_FLAG 0x80000000U
 #define LZ4F_BLOCKSIZEID_DEFAULT LZ4F_max64KB
 
-static const size_t minFHSize = 7;
+static const size_t minFHSize = LZ4F_HEADER_SIZE_MIN;   /*  7 */
 static const size_t maxFHSize = LZ4F_HEADER_SIZE_MAX;   /* 19 */
-static const size_t BHSize = 4;
+static const size_t BHSize = 4;  /* block header : size, and compress flag */
+static const size_t BFSize = 4;  /* block footer : checksum (optional) */
 
 
 /*-************************************
@@ -258,22 +275,22 @@ unsigned LZ4F_getVersion(void) { return LZ4F_VERSION; }
 
 int LZ4F_compressionLevel_max(void) { return LZ4HC_CLEVEL_MAX; }
 
-
-/*-************************************
-*  Private functions
-**************************************/
-#define MIN(a,b)   ( (a) < (b) ? (a) : (b) )
-
-static size_t LZ4F_getBlockSize(unsigned blockSizeID)
+size_t LZ4F_getBlockSize(unsigned blockSizeID)
 {
     static const size_t blockSizes[4] = { 64 KB, 256 KB, 1 MB, 4 MB };
 
     if (blockSizeID == 0) blockSizeID = LZ4F_BLOCKSIZEID_DEFAULT;
-    blockSizeID -= 4;
-    if (blockSizeID > 3) return err0r(LZ4F_ERROR_maxBlockSize_invalid);
+    if (blockSizeID < LZ4F_max64KB || blockSizeID > LZ4F_max4MB)
+        return err0r(LZ4F_ERROR_maxBlockSize_invalid);
+    blockSizeID -= LZ4F_max64KB;
     return blockSizes[blockSizeID];
 }
 
+/*-************************************
+*  Private functions
+**************************************/
+#define MIN(a,b)   ( (a) < (b) ? (a) : (b) )
+
 static BYTE LZ4F_headerChecksum (const void* header, size_t length)
 {
     U32 const xxh = XXH32(header, length, 0);
@@ -323,11 +340,10 @@ static size_t LZ4F_compressBound_internal(size_t srcSize,
         size_t const lastBlockSize = flush ? partialBlockSize : 0;
         unsigned const nbBlocks = nbFullBlocks + (lastBlockSize>0);
 
-        size_t const blockHeaderSize = 4;
-        size_t const blockCRCSize = 4 * prefsPtr->frameInfo.blockChecksumFlag;
-        size_t const frameEnd = 4 + (prefsPtr->frameInfo.contentChecksumFlag*4);
+        size_t const blockCRCSize = BFSize * prefsPtr->frameInfo.blockChecksumFlag;
+        size_t const frameEnd = BHSize + (prefsPtr->frameInfo.contentChecksumFlag*BFSize);
 
-        return ((blockHeaderSize + blockCRCSize) * nbBlocks) +
+        return ((BHSize + blockCRCSize) * nbBlocks) +
                (blockSize * nbFullBlocks) + lastBlockSize + frameEnd;
     }
 }
@@ -388,15 +404,18 @@ size_t LZ4F_compressFrame_usingCDict(LZ4F_cctx* cctx,
       if (LZ4F_isError(headerSize)) return headerSize;
       dstPtr += headerSize;   /* header size */ }
 
-    { size_t const cSize = LZ4F_compressUpdate(cctx, dstPtr, dstEnd-dstPtr, srcBuffer, srcSize, &options);
+    assert(dstEnd >= dstPtr);
+    { size_t const cSize = LZ4F_compressUpdate(cctx, dstPtr, (size_t)(dstEnd-dstPtr), srcBuffer, srcSize, &options);
       if (LZ4F_isError(cSize)) return cSize;
       dstPtr += cSize; }
 
-    { size_t const tailSize = LZ4F_compressEnd(cctx, dstPtr, dstEnd-dstPtr, &options);   /* flush last block, and generate suffix */
+    assert(dstEnd >= dstPtr);
+    { size_t const tailSize = LZ4F_compressEnd(cctx, dstPtr, (size_t)(dstEnd-dstPtr), &options);   /* flush last block, and generate suffix */
       if (LZ4F_isError(tailSize)) return tailSize;
       dstPtr += tailSize; }
 
-    return (dstPtr - dstStart);
+    assert(dstEnd >= dstStart);
+    return (size_t)(dstPtr - dstStart);
 }
 
 
@@ -428,7 +447,7 @@ size_t LZ4F_compressFrame(void* dstBuffer, size_t dstCapacity,
     if (preferencesPtr == NULL ||
         preferencesPtr->compressionLevel < LZ4HC_CLEVEL_MIN)
     {
-        LZ4_resetStream(&lz4ctx);
+        LZ4_initStream(&lz4ctx, sizeof(lz4ctx));
         cctxPtr->lz4CtxPtr = &lz4ctx;
         cctxPtr->lz4CtxAlloc = 1;
         cctxPtr->lz4CtxState = 1;
@@ -598,20 +617,22 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
         if (cctxPtr->lz4CtxAlloc < ctxTypeID) {
             FREEMEM(cctxPtr->lz4CtxPtr);
             if (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) {
-                cctxPtr->lz4CtxPtr = (void*)LZ4_createStream();
+                cctxPtr->lz4CtxPtr = LZ4_createStream();
             } else {
-                cctxPtr->lz4CtxPtr = (void*)LZ4_createStreamHC();
+                cctxPtr->lz4CtxPtr = LZ4_createStreamHC();
             }
-            if (cctxPtr->lz4CtxPtr == NULL) return err0r(LZ4F_ERROR_allocation_failed);
+            if (cctxPtr->lz4CtxPtr == NULL)
+                return err0r(LZ4F_ERROR_allocation_failed);
             cctxPtr->lz4CtxAlloc = ctxTypeID;
             cctxPtr->lz4CtxState = ctxTypeID;
         } else if (cctxPtr->lz4CtxState != ctxTypeID) {
             /* otherwise, a sufficient buffer is allocated, but we need to
              * reset it to the correct context type */
             if (cctxPtr->prefs.compressionLevel < LZ4HC_CLEVEL_MIN) {
-                LZ4_resetStream((LZ4_stream_t *) cctxPtr->lz4CtxPtr);
+                LZ4_initStream((LZ4_stream_t *) cctxPtr->lz4CtxPtr, sizeof (LZ4_stream_t));
             } else {
-                LZ4_resetStreamHC((LZ4_streamHC_t *) cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel);
+                LZ4_initStreamHC((LZ4_streamHC_t *) cctxPtr->lz4CtxPtr, sizeof(LZ4_streamHC_t));
+                LZ4_setCompressionLevel((LZ4_streamHC_t *) cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel);
             }
             cctxPtr->lz4CtxState = ctxTypeID;
         }
@@ -623,8 +644,8 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
     cctxPtr->maxBlockSize = LZ4F_getBlockSize(cctxPtr->prefs.frameInfo.blockSizeID);
 
     {   size_t const requiredBuffSize = preferencesPtr->autoFlush ?
-                (cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) * 64 KB :  /* only needs windows size */
-                cctxPtr->maxBlockSize + ((cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) * 128 KB);
+                ((cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 64 KB : 0) :  /* only needs past data up to window size */
+                cctxPtr->maxBlockSize + ((cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked) ? 128 KB : 0);
 
         if (cctxPtr->maxBufferSize < requiredBuffSize) {
             cctxPtr->maxBufferSize = 0;
@@ -635,7 +656,7 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
     }   }
     cctxPtr->tmpIn = cctxPtr->tmpBuff;
     cctxPtr->tmpInSize = 0;
-    XXH32_reset(&(cctxPtr->xxh), 0);
+    (void)XXH32_reset(&(cctxPtr->xxh), 0);
 
     /* context init */
     cctxPtr->cdict = cdict;
@@ -644,7 +665,7 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
         LZ4F_initStream(cctxPtr->lz4CtxPtr, cdict, cctxPtr->prefs.compressionLevel, LZ4F_blockLinked);
     }
     if (preferencesPtr->compressionLevel >= LZ4HC_CLEVEL_MIN) {
-          LZ4_favorDecompressionSpeed((LZ4_streamHC_t*)cctxPtr->lz4CtxPtr, (int)preferencesPtr->favorDecSpeed);
+        LZ4_favorDecompressionSpeed((LZ4_streamHC_t*)cctxPtr->lz4CtxPtr, (int)preferencesPtr->favorDecSpeed);
     }
 
     /* Magic Number */
@@ -656,7 +677,7 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
     *dstPtr++ = (BYTE)(((1 & _2BITS) << 6)    /* Version('01') */
         + ((cctxPtr->prefs.frameInfo.blockMode & _1BIT ) << 5)
         + ((cctxPtr->prefs.frameInfo.blockChecksumFlag & _1BIT ) << 4)
-        + ((cctxPtr->prefs.frameInfo.contentSize > 0) << 3)
+        + ((unsigned)(cctxPtr->prefs.frameInfo.contentSize > 0) << 3)
         + ((cctxPtr->prefs.frameInfo.contentChecksumFlag & _1BIT ) << 2)
         +  (cctxPtr->prefs.frameInfo.dictID > 0) );
     /* BD Byte */
@@ -673,11 +694,11 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
         dstPtr += 4;
     }
     /* Header CRC Byte */
-    *dstPtr = LZ4F_headerChecksum(headerStart, dstPtr - headerStart);
+    *dstPtr = LZ4F_headerChecksum(headerStart, (size_t)(dstPtr - headerStart));
     dstPtr++;
 
     cctxPtr->cStage = 1;   /* header written, now request input data block */
-    return (dstPtr - dstStart);
+    return (size_t)(dstPtr - dstStart);
 }
 
 
@@ -686,7 +707,7 @@ size_t LZ4F_compressBegin_usingCDict(LZ4F_cctx* cctxPtr,
  *  dstBuffer must be >= LZ4F_HEADER_SIZE_MAX bytes.
  *  preferencesPtr can be NULL, in which case default parameters are selected.
  * @return : number of bytes written into dstBuffer for the header
- *           or an error code (can be tested using LZ4F_isError())
+ *        or an error code (can be tested using LZ4F_isError())
  */
 size_t LZ4F_compressBegin(LZ4F_cctx* cctxPtr,
                           void* dstBuffer, size_t dstCapacity,
@@ -712,27 +733,31 @@ typedef int (*compressFunc_t)(void* ctx, const char* src, char* dst, int srcSize
 
 
 /*! LZ4F_makeBlock():
- *  compress a single block, add header and checksum
- *  assumption : dst buffer capacity is >= srcSize */
-static size_t LZ4F_makeBlock(void* dst, const void* src, size_t srcSize,
+ *  compress a single block, add header and optional checksum.
+ *  assumption : dst buffer capacity is >= BHSize + srcSize + crcSize
+ */
+static size_t LZ4F_makeBlock(void* dst,
+                       const void* src, size_t srcSize,
                              compressFunc_t compress, void* lz4ctx, int level,
-                             const LZ4F_CDict* cdict, LZ4F_blockChecksum_t crcFlag)
+                       const LZ4F_CDict* cdict,
+                             LZ4F_blockChecksum_t crcFlag)
 {
     BYTE* const cSizePtr = (BYTE*)dst;
-    U32 cSize = (U32)compress(lz4ctx, (const char*)src, (char*)(cSizePtr+4),
+    U32 cSize = (U32)compress(lz4ctx, (const char*)src, (char*)(cSizePtr+BHSize),
                                       (int)(srcSize), (int)(srcSize-1),
                                       level, cdict);
-    LZ4F_writeLE32(cSizePtr, cSize);
     if (cSize == 0) {  /* compression failed */
         cSize = (U32)srcSize;
         LZ4F_writeLE32(cSizePtr, cSize | LZ4F_BLOCKUNCOMPRESSED_FLAG);
-        memcpy(cSizePtr+4, src, srcSize);
+        memcpy(cSizePtr+BHSize, src, srcSize);
+    } else {
+        LZ4F_writeLE32(cSizePtr, cSize);
     }
     if (crcFlag) {
-        U32 const crc32 = XXH32(cSizePtr+4, cSize, 0);  /* checksum of compressed data */
-        LZ4F_writeLE32(cSizePtr+4+cSize, crc32);
+        U32 const crc32 = XXH32(cSizePtr+BHSize, cSize, 0);  /* checksum of compressed data */
+        LZ4F_writeLE32(cSizePtr+BHSize+cSize, crc32);
     }
-    return 4 + cSize + ((U32)crcFlag)*4;
+    return BHSize + cSize + ((U32)crcFlag)*BFSize;
 }
 
 
@@ -832,9 +857,11 @@ size_t LZ4F_compressUpdate(LZ4F_cctx* cctxPtr,
             memcpy(cctxPtr->tmpIn + cctxPtr->tmpInSize, srcBuffer, sizeToCopy);
             srcPtr += sizeToCopy;
 
-            dstPtr += LZ4F_makeBlock(dstPtr, cctxPtr->tmpIn, blockSize,
+            dstPtr += LZ4F_makeBlock(dstPtr,
+                                     cctxPtr->tmpIn, blockSize,
                                      compress, cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel,
-                                     cctxPtr->cdict, cctxPtr->prefs.frameInfo.blockChecksumFlag);
+                                     cctxPtr->cdict,
+                                     cctxPtr->prefs.frameInfo.blockChecksumFlag);
 
             if (cctxPtr->prefs.frameInfo.blockMode==LZ4F_blockLinked) cctxPtr->tmpIn += blockSize;
             cctxPtr->tmpInSize = 0;
@@ -844,18 +871,22 @@ size_t LZ4F_compressUpdate(LZ4F_cctx* cctxPtr,
     while ((size_t)(srcEnd - srcPtr) >= blockSize) {
         /* compress full blocks */
         lastBlockCompressed = fromSrcBuffer;
-        dstPtr += LZ4F_makeBlock(dstPtr, srcPtr, blockSize,
+        dstPtr += LZ4F_makeBlock(dstPtr,
+                                 srcPtr, blockSize,
                                  compress, cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel,
-                                 cctxPtr->cdict, cctxPtr->prefs.frameInfo.blockChecksumFlag);
+                                 cctxPtr->cdict,
+                                 cctxPtr->prefs.frameInfo.blockChecksumFlag);
         srcPtr += blockSize;
     }
 
     if ((cctxPtr->prefs.autoFlush) && (srcPtr < srcEnd)) {
         /* compress remaining input < blockSize */
         lastBlockCompressed = fromSrcBuffer;
-        dstPtr += LZ4F_makeBlock(dstPtr, srcPtr, srcEnd - srcPtr,
+        dstPtr += LZ4F_makeBlock(dstPtr,
+                                 srcPtr, (size_t)(srcEnd - srcPtr),
                                  compress, cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel,
-                                 cctxPtr->cdict, cctxPtr->prefs.frameInfo.blockChecksumFlag);
+                                 cctxPtr->cdict,
+                                 cctxPtr->prefs.frameInfo.blockChecksumFlag);
         srcPtr  = srcEnd;
     }
 
@@ -881,28 +912,30 @@ size_t LZ4F_compressUpdate(LZ4F_cctx* cctxPtr,
     /* some input data left, necessarily < blockSize */
     if (srcPtr < srcEnd) {
         /* fill tmp buffer */
-        size_t const sizeToCopy = srcEnd - srcPtr;
+        size_t const sizeToCopy = (size_t)(srcEnd - srcPtr);
         memcpy(cctxPtr->tmpIn, srcPtr, sizeToCopy);
         cctxPtr->tmpInSize = sizeToCopy;
     }
 
     if (cctxPtr->prefs.frameInfo.contentChecksumFlag == LZ4F_contentChecksumEnabled)
-        XXH32_update(&(cctxPtr->xxh), srcBuffer, srcSize);
+        (void)XXH32_update(&(cctxPtr->xxh), srcBuffer, srcSize);
 
     cctxPtr->totalInSize += srcSize;
-    return dstPtr - dstStart;
+    return (size_t)(dstPtr - dstStart);
 }
 
 
 /*! LZ4F_flush() :
- *  Should you need to create compressed data immediately, without waiting for a block to be filled,
- *  you can call LZ4_flush(), which will immediately compress any remaining data stored within compressionContext.
- *  The result of the function is the number of bytes written into dstBuffer
- *  (it can be zero, this means there was no data left within compressionContext)
+ *  When compressed data must be sent immediately, without waiting for a block to be filled,
+ *  invoke LZ4_flush(), which will immediately compress any remaining data stored within LZ4F_cctx.
+ *  The result of the function is the number of bytes written into dstBuffer.
+ *  It can be zero, this means there was no data left within LZ4F_cctx.
  *  The function outputs an error code if it fails (can be tested using LZ4F_isError())
- *  The LZ4F_compressOptions_t structure is optional : you can provide NULL as argument.
+ *  LZ4F_compressOptions_t* is optional. NULL is a valid argument.
  */
-size_t LZ4F_flush(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstCapacity, const LZ4F_compressOptions_t* compressOptionsPtr)
+size_t LZ4F_flush(LZ4F_cctx* cctxPtr,
+                  void* dstBuffer, size_t dstCapacity,
+            const LZ4F_compressOptions_t* compressOptionsPtr)
 {
     BYTE* const dstStart = (BYTE*)dstBuffer;
     BYTE* dstPtr = dstStart;
@@ -910,52 +943,65 @@ size_t LZ4F_flush(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstCapacity, const
 
     if (cctxPtr->tmpInSize == 0) return 0;   /* nothing to flush */
     if (cctxPtr->cStage != 1) return err0r(LZ4F_ERROR_GENERIC);
-    if (dstCapacity < (cctxPtr->tmpInSize + 4)) return err0r(LZ4F_ERROR_dstMaxSize_tooSmall);   /* +4 : block header(4)  */
+    if (dstCapacity < (cctxPtr->tmpInSize + BHSize + BFSize))
+        return err0r(LZ4F_ERROR_dstMaxSize_tooSmall);
     (void)compressOptionsPtr;   /* not yet useful */
 
     /* select compression function */
     compress = LZ4F_selectCompression(cctxPtr->prefs.frameInfo.blockMode, cctxPtr->prefs.compressionLevel);
 
     /* compress tmp buffer */
-    dstPtr += LZ4F_makeBlock(dstPtr, cctxPtr->tmpIn, cctxPtr->tmpInSize,
+    dstPtr += LZ4F_makeBlock(dstPtr,
+                             cctxPtr->tmpIn, cctxPtr->tmpInSize,
                              compress, cctxPtr->lz4CtxPtr, cctxPtr->prefs.compressionLevel,
-                             cctxPtr->cdict, cctxPtr->prefs.frameInfo.blockChecksumFlag);
-    if (cctxPtr->prefs.frameInfo.blockMode==LZ4F_blockLinked) cctxPtr->tmpIn += cctxPtr->tmpInSize;
+                             cctxPtr->cdict,
+                             cctxPtr->prefs.frameInfo.blockChecksumFlag);
+    assert(((void)"flush overflows dstBuffer!", (size_t)(dstPtr - dstStart) <= dstCapacity));
+
+    if (cctxPtr->prefs.frameInfo.blockMode == LZ4F_blockLinked)
+        cctxPtr->tmpIn += cctxPtr->tmpInSize;
     cctxPtr->tmpInSize = 0;
 
     /* keep tmpIn within limits */
     if ((cctxPtr->tmpIn + cctxPtr->maxBlockSize) > (cctxPtr->tmpBuff + cctxPtr->maxBufferSize)) {  /* necessarily LZ4F_blockLinked */
-        int realDictSize = LZ4F_localSaveDict(cctxPtr);
+        int const realDictSize = LZ4F_localSaveDict(cctxPtr);
         cctxPtr->tmpIn = cctxPtr->tmpBuff + realDictSize;
     }
 
-    return dstPtr - dstStart;
+    return (size_t)(dstPtr - dstStart);
 }
 
 
 /*! LZ4F_compressEnd() :
- * When you want to properly finish the compressed frame, just call LZ4F_compressEnd().
- * It will flush whatever data remained within compressionContext (like LZ4_flush())
- * but also properly finalize the frame, with an endMark and a checksum.
- * The result of the function is the number of bytes written into dstBuffer (necessarily >= 4 (endMark size))
- * The function outputs an error code if it fails (can be tested using LZ4F_isError())
- * The LZ4F_compressOptions_t structure is optional : you can provide NULL as argument.
- * compressionContext can then be used again, starting with LZ4F_compressBegin(). The preferences will remain the same.
+ *  When you want to properly finish the compressed frame, just call LZ4F_compressEnd().
+ *  It will flush whatever data remained within compressionContext (like LZ4_flush())
+ *  but also properly finalize the frame, with an endMark and an (optional) checksum.
+ *  LZ4F_compressOptions_t structure is optional : you can provide NULL as argument.
+ * @return: the number of bytes written into dstBuffer (necessarily >= 4 (endMark size))
+ *       or an error code if it fails (can be tested using LZ4F_isError())
+ *  The context can then be used again to compress a new frame, starting with LZ4F_compressBegin().
  */
-size_t LZ4F_compressEnd(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstMaxSize, const LZ4F_compressOptions_t* compressOptionsPtr)
+size_t LZ4F_compressEnd(LZ4F_cctx* cctxPtr,
+                        void* dstBuffer, size_t dstCapacity,
+                  const LZ4F_compressOptions_t* compressOptionsPtr)
 {
     BYTE* const dstStart = (BYTE*)dstBuffer;
     BYTE* dstPtr = dstStart;
 
-    size_t const flushSize = LZ4F_flush(cctxPtr, dstBuffer, dstMaxSize, compressOptionsPtr);
+    size_t const flushSize = LZ4F_flush(cctxPtr, dstBuffer, dstCapacity, compressOptionsPtr);
     if (LZ4F_isError(flushSize)) return flushSize;
     dstPtr += flushSize;
 
+    assert(flushSize <= dstCapacity);
+    dstCapacity -= flushSize;
+
+    if (dstCapacity < 4) return err0r(LZ4F_ERROR_dstMaxSize_tooSmall);
     LZ4F_writeLE32(dstPtr, 0);
-    dstPtr+=4;   /* endMark */
+    dstPtr += 4;   /* endMark */
 
     if (cctxPtr->prefs.frameInfo.contentChecksumFlag == LZ4F_contentChecksumEnabled) {
         U32 const xxh = XXH32_digest(&(cctxPtr->xxh));
+        if (dstCapacity < 8) return err0r(LZ4F_ERROR_dstMaxSize_tooSmall);
         LZ4F_writeLE32(dstPtr, xxh);
         dstPtr+=4;   /* content Checksum */
     }
@@ -968,7 +1014,7 @@ size_t LZ4F_compressEnd(LZ4F_cctx* cctxPtr, void* dstBuffer, size_t dstMaxSize,
             return err0r(LZ4F_ERROR_frameSize_wrong);
     }
 
-    return dstPtr - dstStart;
+    return (size_t)(dstPtr - dstStart);
 }
 
 
@@ -1049,31 +1095,6 @@ void LZ4F_resetDecompressionContext(LZ4F_dctx* dctx)
 }
 
 
-/*! LZ4F_headerSize() :
- *   @return : size of frame header
- *             or an error code, which can be tested using LZ4F_isError()
- */
-static size_t LZ4F_headerSize(const void* src, size_t srcSize)
-{
-    /* minimal srcSize to determine header size */
-    if (srcSize < 5) return err0r(LZ4F_ERROR_frameHeader_incomplete);
-
-    /* special case : skippable frames */
-    if ((LZ4F_readLE32(src) & 0xFFFFFFF0U) == LZ4F_MAGIC_SKIPPABLE_START) return 8;
-
-    /* control magic number */
-    if (LZ4F_readLE32(src) != LZ4F_MAGICNUMBER)
-        return err0r(LZ4F_ERROR_frameType_unknown);
-
-    /* Frame Header Size */
-    {   BYTE const FLG = ((const BYTE*)src)[4];
-        U32 const contentSizeFlag = (FLG>>3) & _1BIT;
-        U32 const dictIDFlag = FLG & _1BIT;
-        return minFHSize + (contentSizeFlag*8) + (dictIDFlag*4);
-    }
-}
-
-
 /*! LZ4F_decodeHeader() :
  *  input   : `src` points at the **beginning of the frame**
  *  output  : set internal values of dctx, such as
@@ -1125,7 +1146,7 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize
     }
 
     /* Frame Header Size */
-    frameHeaderSize = minFHSize + (contentSizeFlag*8) + (dictIDFlag*4);
+    frameHeaderSize = minFHSize + (contentSizeFlag?8:0) + (dictIDFlag?4:0);
 
     if (srcSize < frameHeaderSize) {
         /* not enough input to fully decode frame header */
@@ -1146,6 +1167,7 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize
     }
 
     /* check header */
+    assert(frameHeaderSize > 5);
     {   BYTE const HC = LZ4F_headerChecksum(srcPtr+4, frameHeaderSize-5);
         if (HC != srcPtr[frameHeaderSize-1])
             return err0r(LZ4F_ERROR_headerChecksum_invalid);
@@ -1169,6 +1191,34 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize
 }
 
 
+/*! LZ4F_headerSize() :
+ * @return : size of frame header
+ *           or an error code, which can be tested using LZ4F_isError()
+ */
+size_t LZ4F_headerSize(const void* src, size_t srcSize)
+{
+    if (src == NULL) return err0r(LZ4F_ERROR_srcPtr_wrong);
+
+    /* minimal srcSize to determine header size */
+    if (srcSize < LZ4F_MIN_SIZE_TO_KNOW_HEADER_LENGTH)
+        return err0r(LZ4F_ERROR_frameHeader_incomplete);
+
+    /* special case : skippable frames */
+    if ((LZ4F_readLE32(src) & 0xFFFFFFF0U) == LZ4F_MAGIC_SKIPPABLE_START)
+        return 8;
+
+    /* control magic number */
+    if (LZ4F_readLE32(src) != LZ4F_MAGICNUMBER)
+        return err0r(LZ4F_ERROR_frameType_unknown);
+
+    /* Frame Header Size */
+    {   BYTE const FLG = ((const BYTE*)src)[4];
+        U32 const contentSizeFlag = (FLG>>3) & _1BIT;
+        U32 const dictIDFlag = FLG & _1BIT;
+        return minFHSize + (contentSizeFlag?8:0) + (dictIDFlag?4:0);
+    }
+}
+
 /*! LZ4F_getFrameInfo() :
  *  This function extracts frame parameters (max blockSize, frame checksum, etc.).
  *  Usage is optional. Objective is to provide relevant information for allocation purposes.
@@ -1184,10 +1234,12 @@ static size_t LZ4F_decodeHeader(LZ4F_dctx* dctx, const void* src, size_t srcSize
  *  note 1 : in case of error, dctx is not modified. Decoding operations can resume from where they stopped.
  *  note 2 : frame parameters are *copied into* an already allocated LZ4F_frameInfo_t structure.
  */
-LZ4F_errorCode_t LZ4F_getFrameInfo(LZ4F_dctx* dctx, LZ4F_frameInfo_t* frameInfoPtr,
-                                   const void* srcBuffer, size_t* srcSizePtr)
+LZ4F_errorCode_t LZ4F_getFrameInfo(LZ4F_dctx* dctx,
+                                   LZ4F_frameInfo_t* frameInfoPtr,
+                             const void* srcBuffer, size_t* srcSizePtr)
 {
-    if (dctx->dStage > dstage_storeFrameHeader) {  /* assumption :  dstage_* header enum at beginning of range */
+    LZ4F_STATIC_ASSERT(dstage_getFrameHeader < dstage_storeFrameHeader);
+    if (dctx->dStage > dstage_storeFrameHeader) {
         /* frameInfo already decoded */
         size_t o=0, i=0;
         *srcSizePtr = 0;
@@ -1200,7 +1252,6 @@ LZ4F_errorCode_t LZ4F_getFrameInfo(LZ4F_dctx* dctx, LZ4F_frameInfo_t* frameInfoP
             *srcSizePtr = 0;
             return err0r(LZ4F_ERROR_frameDecoding_alreadyStarted);
         } else {
-            size_t decodeResult;
             size_t const hSize = LZ4F_headerSize(srcBuffer, *srcSizePtr);
             if (LZ4F_isError(hSize)) { *srcSizePtr=0; return hSize; }
             if (*srcSizePtr < hSize) {
@@ -1208,16 +1259,16 @@ LZ4F_errorCode_t LZ4F_getFrameInfo(LZ4F_dctx* dctx, LZ4F_frameInfo_t* frameInfoP
                 return err0r(LZ4F_ERROR_frameHeader_incomplete);
             }
 
-            decodeResult = LZ4F_decodeHeader(dctx, srcBuffer, hSize);
-            if (LZ4F_isError(decodeResult)) {
-                *srcSizePtr = 0;
-            } else {
-                *srcSizePtr = decodeResult;
-                decodeResult = BHSize;   /* block header size */
-            }
-            *frameInfoPtr = dctx->frameInfo;
-            return decodeResult;
-    }   }
+            {   size_t decodeResult = LZ4F_decodeHeader(dctx, srcBuffer, hSize);
+                if (LZ4F_isError(decodeResult)) {
+                    *srcSizePtr = 0;
+                } else {
+                    *srcSizePtr = decodeResult;
+                    decodeResult = BHSize;   /* block header size */
+                }
+                *frameInfoPtr = dctx->frameInfo;
+                return decodeResult;
+    }   }   }
 }
 
 
@@ -1235,9 +1286,10 @@ static void LZ4F_updateDict(LZ4F_dctx* dctx,
         return;
     }
 
-    if (dstPtr - dstBufferStart + dstSize >= 64 KB) {  /* history in dstBuffer becomes large enough to become dictionary */
+    assert(dstPtr >= dstBufferStart);
+    if ((size_t)(dstPtr - dstBufferStart) + dstSize >= 64 KB) {  /* history in dstBuffer becomes large enough to become dictionary */
         dctx->dict = (const BYTE*)dstBufferStart;
-        dctx->dictSize = dstPtr - dstBufferStart + dstSize;
+        dctx->dictSize = (size_t)(dstPtr - dstBufferStart) + dstSize;
         return;
     }
 
@@ -1253,7 +1305,7 @@ static void LZ4F_updateDict(LZ4F_dctx* dctx,
     }
 
     if (withinTmp) { /* copy relevant dict portion in front of tmpOut within tmpOutBuffer */
-        size_t const preserveSize = dctx->tmpOut - dctx->tmpOutBuffer;
+        size_t const preserveSize = (size_t)(dctx->tmpOut - dctx->tmpOutBuffer);
         size_t copySize = 64 KB - dctx->tmpOutSize;
         const BYTE* const oldDictEnd = dctx->dict + dctx->dictSize - dctx->tmpOutStart;
         if (dctx->tmpOutSize > 64 KB) copySize = 0;
@@ -1338,7 +1390,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
 
         case dstage_getFrameHeader:
             if ((size_t)(srcEnd-srcPtr) >= maxFHSize) {  /* enough to decode - shortcut */
-                size_t const hSize = LZ4F_decodeHeader(dctx, srcPtr, srcEnd-srcPtr);  /* will update dStage appropriately */
+                size_t const hSize = LZ4F_decodeHeader(dctx, srcPtr, (size_t)(srcEnd-srcPtr));  /* will update dStage appropriately */
                 if (LZ4F_isError(hSize)) return hSize;
                 srcPtr += hSize;
                 break;
@@ -1366,14 +1418,14 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
             break;
 
         case dstage_init:
-            if (dctx->frameInfo.contentChecksumFlag) XXH32_reset(&(dctx->xxh), 0);
+            if (dctx->frameInfo.contentChecksumFlag) (void)XXH32_reset(&(dctx->xxh), 0);
             /* internal buffers allocation */
             {   size_t const bufferNeeded = dctx->maxBlockSize
-                    + ((dctx->frameInfo.blockMode==LZ4F_blockLinked) * 128 KB);
+                    + ((dctx->frameInfo.blockMode==LZ4F_blockLinked) ? 128 KB : 0);
                 if (bufferNeeded > dctx->maxBufferSize) {   /* tmp buffers too small */
                     dctx->maxBufferSize = 0;   /* ensure allocation will be re-attempted on next entry*/
                     FREEMEM(dctx->tmpIn);
-                    dctx->tmpIn = (BYTE*)ALLOC(dctx->maxBlockSize + 4 /* block checksum */);
+                    dctx->tmpIn = (BYTE*)ALLOC(dctx->maxBlockSize + BFSize /* block checksum */);
                     if (dctx->tmpIn == NULL)
                         return err0r(LZ4F_ERROR_allocation_failed);
                     FREEMEM(dctx->tmpOutBuffer);
@@ -1420,7 +1472,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
 
         /* decode block header */
             {   size_t const nextCBlockSize = LZ4F_readLE32(selectedIn) & 0x7FFFFFFFU;
-                size_t const crcSize = dctx->frameInfo.blockChecksumFlag * 4;
+                size_t const crcSize = dctx->frameInfo.blockChecksumFlag * BFSize;
                 if (nextCBlockSize==0) {  /* frameEnd signal, no more block */
                     dctx->dStage = dstage_getSuffix;
                     break;
@@ -1431,7 +1483,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                     /* next block is uncompressed */
                     dctx->tmpInTarget = nextCBlockSize;
                     if (dctx->frameInfo.blockChecksumFlag) {
-                        XXH32_reset(&dctx->blockChecksum, 0);
+                        (void)XXH32_reset(&dctx->blockChecksum, 0);
                     }
                     dctx->dStage = dstage_copyDirect;
                     break;
@@ -1440,7 +1492,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                 dctx->tmpInTarget = nextCBlockSize + crcSize;
                 dctx->dStage = dstage_getCBlock;
                 if (dstPtr==dstEnd) {
-                    nextSrcSizeHint = nextCBlockSize + crcSize + BHSize;
+                    nextSrcSizeHint = BHSize + nextCBlockSize + crcSize;
                     doAnotherStage = 0;
                 }
                 break;
@@ -1451,10 +1503,10 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                 size_t const sizeToCopy = MIN(dctx->tmpInTarget, minBuffSize);
                 memcpy(dstPtr, srcPtr, sizeToCopy);
                 if (dctx->frameInfo.blockChecksumFlag) {
-                    XXH32_update(&dctx->blockChecksum, srcPtr, sizeToCopy);
+                    (void)XXH32_update(&dctx->blockChecksum, srcPtr, sizeToCopy);
                 }
                 if (dctx->frameInfo.contentChecksumFlag)
-                    XXH32_update(&dctx->xxh, srcPtr, sizeToCopy);
+                    (void)XXH32_update(&dctx->xxh, srcPtr, sizeToCopy);
                 if (dctx->frameInfo.contentSize)
                     dctx->frameRemainingSize -= sizeToCopy;
 
@@ -1474,7 +1526,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                 }
                 dctx->tmpInTarget -= sizeToCopy;  /* need to copy more */
                 nextSrcSizeHint = dctx->tmpInTarget +
-                                + dctx->frameInfo.contentChecksumFlag * 4  /* block checksum */
+                                +(dctx->frameInfo.blockChecksumFlag ? BFSize : 0)
                                 + BHSize /* next header size */;
                 doAnotherStage = 0;
                 break;
@@ -1525,8 +1577,10 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                 dctx->tmpInSize += sizeToCopy;
                 srcPtr += sizeToCopy;
                 if (dctx->tmpInSize < dctx->tmpInTarget) { /* need more input */
-                    nextSrcSizeHint = (dctx->tmpInTarget - dctx->tmpInSize) + BHSize;
-                    doAnotherStage=0;
+                    nextSrcSizeHint = (dctx->tmpInTarget - dctx->tmpInSize)
+                                    + (dctx->frameInfo.blockChecksumFlag ? BFSize : 0)
+                                    + BHSize /* next header size */;
+                    doAnotherStage = 0;
                     break;
                 }
                 selectedIn = dctx->tmpIn;
@@ -1558,13 +1612,13 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                         dict, (int)dictSize);
                 if (decodedSize < 0) return err0r(LZ4F_ERROR_GENERIC);   /* decompression failed */
                 if (dctx->frameInfo.contentChecksumFlag)
-                    XXH32_update(&(dctx->xxh), dstPtr, decodedSize);
+                    XXH32_update(&(dctx->xxh), dstPtr, (size_t)decodedSize);
                 if (dctx->frameInfo.contentSize)
-                    dctx->frameRemainingSize -= decodedSize;
+                    dctx->frameRemainingSize -= (size_t)decodedSize;
 
                 /* dictionary management */
                 if (dctx->frameInfo.blockMode==LZ4F_blockLinked)
-                    LZ4F_updateDict(dctx, dstPtr, decodedSize, dstStart, 0);
+                    LZ4F_updateDict(dctx, dstPtr, (size_t)decodedSize, dstStart, 0);
 
                 dstPtr += decodedSize;
                 dctx->dStage = dstage_getBlockHeader;
@@ -1601,10 +1655,10 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
                 if (decodedSize < 0)  /* decompression failed */
                     return err0r(LZ4F_ERROR_decompressionFailed);
                 if (dctx->frameInfo.contentChecksumFlag)
-                    XXH32_update(&(dctx->xxh), dctx->tmpOut, decodedSize);
+                    XXH32_update(&(dctx->xxh), dctx->tmpOut, (size_t)decodedSize);
                 if (dctx->frameInfo.contentSize)
-                    dctx->frameRemainingSize -= decodedSize;
-                dctx->tmpOutSize = decodedSize;
+                    dctx->frameRemainingSize -= (size_t)decodedSize;
+                dctx->tmpOutSize = (size_t)decodedSize;
                 dctx->tmpOutStart = 0;
                 dctx->dStage = dstage_flushOut;
             }
@@ -1732,7 +1786,7 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
       && ((unsigned)(dctx->dStage)-2 < (unsigned)(dstage_getSuffix)-2) )  /* valid stages : [init ... getSuffix[ */
     {
         if (dctx->dStage == dstage_flushOut) {
-            size_t const preserveSize = dctx->tmpOut - dctx->tmpOutBuffer;
+            size_t const preserveSize = (size_t)(dctx->tmpOut - dctx->tmpOutBuffer);
             size_t copySize = 64 KB - dctx->tmpOutSize;
             const BYTE* oldDictEnd = dctx->dict + dctx->dictSize - dctx->tmpOutStart;
             if (dctx->tmpOutSize > 64 KB) copySize = 0;
@@ -1756,8 +1810,8 @@ size_t LZ4F_decompress(LZ4F_dctx* dctx,
         }
     }
 
-    *srcSizePtr = (srcPtr - srcStart);
-    *dstSizePtr = (dstPtr - dstStart);
+    *srcSizePtr = (size_t)(srcPtr - srcStart);
+    *dstSizePtr = (size_t)(dstPtr - dstStart);
     return nextSrcSizeHint;
 }
 
diff --git a/lib/lz4frame.h b/lib/lz4frame.h
index 75f1fd9..ca20dc9 100644
--- a/lib/lz4frame.h
+++ b/lib/lz4frame.h
@@ -32,11 +32,14 @@
    - LZ4 public forum : https://groups.google.com/forum/#!forum/lz4c
 */
 
-/* LZ4F is a stand-alone API to create LZ4-compressed frames
- * conformant with specification v1.6.1.
- * It also offers streaming capabilities.
+/* LZ4F is a stand-alone API able to create and decode LZ4 frames
+ * conformant with specification v1.6.1 in doc/lz4_Frame_format.md .
+ * Generated frames are compatible with `lz4` CLI.
+ *
+ * LZ4F also offers streaming capabilities.
+ *
  * lz4.h is not required when using lz4frame.h,
- * except to get constant such as LZ4_VERSION_NUMBER.
+ * except to extract common constant such as LZ4_VERSION_NUMBER.
  * */
 
 #ifndef LZ4F_H_09782039843
@@ -195,7 +198,7 @@ typedef struct {
 *  Simple compression function
 ***********************************/
 
-LZ4FLIB_API int LZ4F_compressionLevel_max(void);
+LZ4FLIB_API int LZ4F_compressionLevel_max(void);   /* v1.8.0+ */
 
 /*! LZ4F_compressFrameBound() :
  *  Returns the maximum possible compressed size with LZ4F_compressFrame() given srcSize and preferences.
@@ -247,7 +250,9 @@ LZ4FLIB_API LZ4F_errorCode_t LZ4F_freeCompressionContext(LZ4F_cctx* cctx);
 
 /*----    Compression    ----*/
 
-#define LZ4F_HEADER_SIZE_MAX 19   /* LZ4 Frame header size can vary from 7 to 19 bytes */
+#define LZ4F_HEADER_SIZE_MIN  7   /* LZ4 Frame header size can vary, depending on selected paramaters */
+#define LZ4F_HEADER_SIZE_MAX 19
+
 /*! LZ4F_compressBegin() :
  *  will write the frame header into dstBuffer.
  *  dstCapacity must be >= LZ4F_HEADER_SIZE_MAX bytes.
@@ -260,15 +265,19 @@ LZ4FLIB_API size_t LZ4F_compressBegin(LZ4F_cctx* cctx,
                                       const LZ4F_preferences_t* prefsPtr);
 
 /*! LZ4F_compressBound() :
- *  Provides minimum dstCapacity required to guarantee compression success
- *  given a srcSize and preferences, covering worst case scenario.
+ *  Provides minimum dstCapacity required to guarantee success of
+ *  LZ4F_compressUpdate(), given a srcSize and preferences, for a worst case scenario.
+ *  When srcSize==0, LZ4F_compressBound() provides an upper bound for LZ4F_flush() and LZ4F_compressEnd() instead.
+ *  Note that the result is only valid for a single invocation of LZ4F_compressUpdate().
+ *  When invoking LZ4F_compressUpdate() multiple times,
+ *  if the output buffer is gradually filled up instead of emptied and re-used from its start,
+ *  one must check if there is enough remaining capacity before each invocation, using LZ4F_compressBound().
+ * @return is always the same for a srcSize and prefsPtr.
  *  prefsPtr is optional : when NULL is provided, preferences will be set to cover worst case scenario.
- *  Estimation is valid for either LZ4F_compressUpdate(), LZ4F_flush() or LZ4F_compressEnd(),
- *  Estimation includes the possibility that internal buffer might already be filled by up to (blockSize-1) bytes.
- *  It also includes frame footer (ending + checksum), which would have to be generated by LZ4F_compressEnd().
- *  Estimation doesn't include frame header, as it was already generated by LZ4F_compressBegin().
- *  Result is always the same for a srcSize and prefsPtr, so it can be trusted to size reusable buffers.
- *  When srcSize==0, LZ4F_compressBound() provides an upper bound for LZ4F_flush() and LZ4F_compressEnd() operations.
+ *  tech details :
+ * @return includes the possibility that internal buffer might already be filled by up to (blockSize-1) bytes.
+ *  It also includes frame footer (ending + checksum), since it might be generated by LZ4F_compressEnd().
+ * @return doesn't include frame header, as it was already generated by LZ4F_compressBegin().
  */
 LZ4FLIB_API size_t LZ4F_compressBound(size_t srcSize, const LZ4F_preferences_t* prefsPtr);
 
@@ -295,6 +304,7 @@ LZ4FLIB_API size_t LZ4F_compressUpdate(LZ4F_cctx* cctx,
  * `cOptPtr` is optional : it's possible to provide NULL, all options will be set to default.
  * @return : nb of bytes written into dstBuffer (can be zero, when there is no data stored within cctx)
  *           or an error code if it fails (which can be tested using LZ4F_isError())
+ *  Note : LZ4F_flush() is guaranteed to be successful when dstCapacity >= LZ4F_compressBound(0, prefsPtr).
  */
 LZ4FLIB_API size_t LZ4F_flush(LZ4F_cctx* cctx,
                               void* dstBuffer, size_t dstCapacity,
@@ -307,6 +317,7 @@ LZ4FLIB_API size_t LZ4F_flush(LZ4F_cctx* cctx,
  * `cOptPtr` is optional : NULL can be provided, in which case all options will be set to default.
  * @return : nb of bytes written into dstBuffer, necessarily >= 4 (endMark),
  *           or an error code if it fails (which can be tested using LZ4F_isError())
+ *  Note : LZ4F_compressEnd() is guaranteed to be successful when dstCapacity >= LZ4F_compressBound(0, prefsPtr).
  *  A successful call to LZ4F_compressEnd() makes `cctx` available again for another compression task.
  */
 LZ4FLIB_API size_t LZ4F_compressEnd(LZ4F_cctx* cctx,
@@ -345,23 +356,58 @@ LZ4FLIB_API LZ4F_errorCode_t LZ4F_freeDecompressionContext(LZ4F_dctx* dctx);
 *  Streaming decompression functions
 *************************************/
 
+#define LZ4F_MIN_SIZE_TO_KNOW_HEADER_LENGTH 5
+
+/*! LZ4F_headerSize() : v1.9.0+
+ *  Provide the header size of a frame starting at `src`.
+ * `srcSize` must be >= LZ4F_MIN_SIZE_TO_KNOW_HEADER_LENGTH,
+ *  which is enough to decode the header length.
+ * @return : size of frame header
+ *           or an error code, which can be tested using LZ4F_isError()
+ *  note : Frame header size is variable, but is guaranteed to be
+ *         >= LZ4F_HEADER_SIZE_MIN bytes, and <= LZ4F_HEADER_SIZE_MAX bytes.
+ */
+size_t LZ4F_headerSize(const void* src, size_t srcSize);
+
 /*! LZ4F_getFrameInfo() :
  *  This function extracts frame parameters (max blockSize, dictID, etc.).
- *  Its usage is optional.
- *  Extracted information is typically useful for allocation and dictionary.
- *  This function works in 2 situations :
- *   - At the beginning of a new frame, in which case
- *     it will decode information from `srcBuffer`, starting the decoding process.
- *     Input size must be large enough to successfully decode the entire frame header.
- *     Frame header size is variable, but is guaranteed to be <= LZ4F_HEADER_SIZE_MAX bytes.
- *     It's allowed to provide more input data than this minimum.
- *   - After decoding has been started.
- *     In which case, no input is read, frame parameters are extracted from dctx.
- *   - If decoding has barely started, but not yet extracted information from header,
+ *  Its usage is optional: user can call LZ4F_decompress() directly.
+ *
+ *  Extracted information will fill an existing LZ4F_frameInfo_t structure.
+ *  This can be useful for allocation and dictionary identification purposes.
+ *
+ *  LZ4F_getFrameInfo() can work in the following situations :
+ *
+ *  1) At the beginning of a new frame, before any invocation of LZ4F_decompress().
+ *     It will decode header from `srcBuffer`,
+ *     consuming the header and starting the decoding process.
+ *
+ *     Input size must be large enough to contain the full frame header.
+ *     Frame header size can be known beforehand by LZ4F_headerSize().
+ *     Frame header size is variable, but is guaranteed to be >= LZ4F_HEADER_SIZE_MIN bytes,
+ *     and not more than <= LZ4F_HEADER_SIZE_MAX bytes.
+ *     Hence, blindly providing LZ4F_HEADER_SIZE_MAX bytes or more will always work.
+ *     It's allowed to provide more input data than the header size,
+ *     LZ4F_getFrameInfo() will only consume the header.
+ *
+ *     If input size is not large enough,
+ *     aka if it's smaller than header size,
+ *     function will fail and return an error code.
+ *
+ *  2) After decoding has been started,
+ *     it's possible to invoke LZ4F_getFrameInfo() anytime
+ *     to extract already decoded frame parameters stored within dctx.
+ *
+ *     Note that, if decoding has barely started,
+ *     and not yet read enough information to decode the header,
  *     LZ4F_getFrameInfo() will fail.
- *  The number of bytes consumed from srcBuffer will be updated within *srcSizePtr (necessarily <= original value).
- *  Decompression must resume from (srcBuffer + *srcSizePtr).
- * @return : an hint about how many srcSize bytes LZ4F_decompress() expects for next call,
+ *
+ *  The number of bytes consumed from srcBuffer will be updated in *srcSizePtr (necessarily <= original value).
+ *  LZ4F_getFrameInfo() only consumes bytes when decoding has not yet started,
+ *  and when decoding the header has been successful.
+ *  Decompression must then resume from (srcBuffer + *srcSizePtr).
+ *
+ * @return : a hint about how many srcSize bytes LZ4F_decompress() expects for next call,
  *           or an error code which can be tested using LZ4F_isError().
  *  note 1 : in case of error, dctx is not modified. Decoding operation can resume from beginning safely.
  *  note 2 : frame parameters are *copied into* an already allocated LZ4F_frameInfo_t structure.
@@ -427,15 +473,15 @@ LZ4FLIB_API void LZ4F_resetDecompressionContext(LZ4F_dctx* dctx);   /* always su
 extern "C" {
 #endif
 
-/* These declarations are not stable and may change in the future. They are
- * therefore only safe to depend on when the caller is statically linked
- * against the library. To access their declarations, define
- * LZ4F_STATIC_LINKING_ONLY.
+/* These declarations are not stable and may change in the future.
+ * They are therefore only safe to depend on
+ * when the caller is statically linked against the library.
+ * To access their declarations, define LZ4F_STATIC_LINKING_ONLY.
  *
- * There is a further protection mechanism where these symbols aren't published
- * into shared/dynamic libraries. You can override this behavior and force
- * them to be published by defining LZ4F_PUBLISH_STATIC_FUNCTIONS. Use at
- * your own risk.
+ * By default, these symbols aren't published into shared/dynamic libraries.
+ * You can override this behavior and force them to be published
+ * by defining LZ4F_PUBLISH_STATIC_FUNCTIONS.
+ * Use at your own risk.
  */
 #ifdef LZ4F_PUBLISH_STATIC_FUNCTIONS
 #define LZ4FLIB_STATIC_API LZ4FLIB_API
@@ -471,19 +517,38 @@ extern "C" {
 #define LZ4F_GENERATE_ENUM(ENUM) LZ4F_##ENUM,
 
 /* enum list is exposed, to handle specific errors */
-typedef enum { LZ4F_LIST_ERRORS(LZ4F_GENERATE_ENUM) } LZ4F_errorCodes;
+typedef enum { LZ4F_LIST_ERRORS(LZ4F_GENERATE_ENUM)
+              _LZ4F_dummy_error_enum_for_c89_never_used } LZ4F_errorCodes;
 
 LZ4FLIB_STATIC_API LZ4F_errorCodes LZ4F_getErrorCode(size_t functionResult);
 
-
+LZ4FLIB_STATIC_API size_t LZ4F_getBlockSize(unsigned);
 
 /**********************************
  *  Bulk processing dictionary API
  *********************************/
+
+/* A Dictionary is useful for the compression of small messages (KB range).
+ * It dramatically improves compression efficiency.
+ *
+ * LZ4 can ingest any input as dictionary, though only the last 64 KB are useful.
+ * Best results are generally achieved by using Zstandard's Dictionary Builder
+ * to generate a high-quality dictionary from a set of samples.
+ *
+ * Loading a dictionary has a cost, since it involves construction of tables.
+ * The Bulk processing dictionary API makes it possible to share this cost
+ * over an arbitrary number of compression jobs, even concurrently,
+ * markedly improving compression latency for these cases.
+ *
+ * The same dictionary will have to be used on the decompression side
+ * for decoding to be successful.
+ * To help identify the correct dictionary at decoding stage,
+ * the frame header allows optional embedding of a dictID field.
+ */
 typedef struct LZ4F_CDict_s LZ4F_CDict;
 
 /*! LZ4_createCDict() :
- *  When compressing multiple messages / blocks with the same dictionary, it's recommended to load it just once.
+ *  When compressing multiple messages / blocks using the same dictionary, it's recommended to load it just once.
  *  LZ4_createCDict() will create a digested dictionary, ready to start future compression operations without startup delay.
  *  LZ4_CDict can be created once and shared by multiple threads concurrently, since its usage is read-only.
  * `dictBuffer` can be released after LZ4_CDict creation, since its content is copied within CDict */
diff --git a/lib/lz4hc.c b/lib/lz4hc.c
index e913ee7..f6ed779 100644
--- a/lib/lz4hc.c
+++ b/lib/lz4hc.c
@@ -61,9 +61,14 @@
 #  pragma clang diagnostic ignored "-Wunused-function"
 #endif
 
+/*===   Enums   ===*/
+typedef enum { noDictCtx, usingDictCtxHc } dictCtx_directive;
+
+
 #define LZ4_COMMONDEFS_ONLY
+#ifndef LZ4_SRC_INCLUDED
 #include "lz4.c"   /* LZ4_count, constants, mem */
-
+#endif
 
 /*===   Constants   ===*/
 #define OPTIMAL_ML (int)((ML_MASK-1)+MINMATCH)
@@ -76,12 +81,11 @@
 #define HASH_FUNCTION(i)         (((i) * 2654435761U) >> ((MINMATCH*8)-LZ4HC_HASH_LOG))
 #define DELTANEXTMAXD(p)         chainTable[(p) & LZ4HC_MAXD_MASK]    /* flexible, LZ4HC_MAXD dependent */
 #define DELTANEXTU16(table, pos) table[(U16)(pos)]   /* faster */
+/* Make fields passed to, and updated by LZ4HC_encodeSequence explicit */
+#define UPDATABLE(ip, op, anchor) &ip, &op, &anchor
 
 static U32 LZ4HC_hashPtr(const void* ptr) { return HASH_FUNCTION(LZ4_read32(ptr)); }
 
-/*===   Enums   ===*/
-typedef enum { noDictCtx, usingDictCtx } dictCtx_directive;
-
 
 /**************************************
 *  HC Compression
@@ -92,9 +96,9 @@ static void LZ4HC_clearTables (LZ4HC_CCtx_internal* hc4)
     MEM_INIT(hc4->chainTable, 0xFF, sizeof(hc4->chainTable));
 }
 
-static void LZ4HC_init (LZ4HC_CCtx_internal* hc4, const BYTE* start)
+static void LZ4HC_init_internal (LZ4HC_CCtx_internal* hc4, const BYTE* start)
 {
-    uptrval startingOffset = hc4->end - hc4->base;
+    uptrval startingOffset = (uptrval)(hc4->end - hc4->base);
     if (startingOffset > 1 GB) {
         LZ4HC_clearTables(hc4);
         startingOffset = 0;
@@ -121,7 +125,7 @@ LZ4_FORCE_INLINE void LZ4HC_Insert (LZ4HC_CCtx_internal* hc4, const BYTE* ip)
     while (idx < target) {
         U32 const h = LZ4HC_hashPtr(base+idx);
         size_t delta = idx - hashTable[h];
-        if (delta>MAX_DISTANCE) delta = MAX_DISTANCE;
+        if (delta>LZ4_DISTANCE_MAX) delta = LZ4_DISTANCE_MAX;
         DELTANEXTU16(chainTable, idx) = (U16)delta;
         hashTable[h] = idx;
         idx++;
@@ -224,14 +228,13 @@ LZ4HC_InsertAndGetWiderMatch (
     const U32 dictLimit = hc4->dictLimit;
     const BYTE* const lowPrefixPtr = base + dictLimit;
     const U32 ipIndex = (U32)(ip - base);
-    const U32 lowestMatchIndex = (hc4->lowLimit + 64 KB > ipIndex) ? hc4->lowLimit : ipIndex - MAX_DISTANCE;
+    const U32 lowestMatchIndex = (hc4->lowLimit + 64 KB > ipIndex) ? hc4->lowLimit : ipIndex - LZ4_DISTANCE_MAX;
     const BYTE* const dictBase = hc4->dictBase;
     int const lookBackLength = (int)(ip-iLowLimit);
     int nbAttempts = maxNbAttempts;
-    int matchChainPos = 0;
+    U32 matchChainPos = 0;
     U32 const pattern = LZ4_read32(ip);
     U32 matchIndex;
-    U32 dictMatchIndex;
     repeat_state_e repeat = rep_untested;
     size_t srcPatternLength = 0;
 
@@ -256,7 +259,7 @@ LZ4HC_InsertAndGetWiderMatch (
             if (LZ4_read16(iLowLimit + longest - 1) == LZ4_read16(matchPtr - lookBackLength + longest - 1)) {
                 if (LZ4_read32(matchPtr) == pattern) {
                     int const back = lookBackLength ? LZ4HC_countBack(ip, matchPtr, iLowLimit, lowPrefixPtr) : 0;
-                    matchLength = MINMATCH + LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, iHighLimit);
+                    matchLength = MINMATCH + (int)LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, iHighLimit);
                     matchLength -= back;
                     if (matchLength > longest) {
                         longest = matchLength;
@@ -270,7 +273,7 @@ LZ4HC_InsertAndGetWiderMatch (
                 int back = 0;
                 const BYTE* vLimit = ip + (dictLimit - matchIndex);
                 if (vLimit > iHighLimit) vLimit = iHighLimit;
-                matchLength = LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, vLimit) + MINMATCH;
+                matchLength = (int)LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, vLimit) + MINMATCH;
                 if ((ip+matchLength == vLimit) && (vLimit < iHighLimit))
                     matchLength += LZ4_count(ip+matchLength, lowPrefixPtr, iHighLimit);
                 back = lookBackLength ? LZ4HC_countBack(ip, matchPtr, iLowLimit, dictStart) : 0;
@@ -283,14 +286,14 @@ LZ4HC_InsertAndGetWiderMatch (
 
         if (chainSwap && matchLength==longest) {    /* better match => select a better chain */
             assert(lookBackLength==0);   /* search forward only */
-            if (matchIndex + longest <= ipIndex) {
+            if (matchIndex + (U32)longest <= ipIndex) {
                 U32 distanceToNextMatch = 1;
                 int pos;
                 for (pos = 0; pos <= longest - MINMATCH; pos++) {
-                    U32 const candidateDist = DELTANEXTU16(chainTable, matchIndex + pos);
+                    U32 const candidateDist = DELTANEXTU16(chainTable, matchIndex + (U32)pos);
                     if (candidateDist > distanceToNextMatch) {
                         distanceToNextMatch = candidateDist;
-                        matchChainPos = pos;
+                        matchChainPos = (U32)pos;
                 }   }
                 if (distanceToNextMatch > 1) {
                     if (distanceToNextMatch > matchIndex) break;   /* avoid overflow */
@@ -315,7 +318,7 @@ LZ4HC_InsertAndGetWiderMatch (
                     const BYTE* const matchPtr = base + matchCandidateIdx;
                     if (LZ4_read32(matchPtr) == pattern) {  /* good candidate */
                         size_t const forwardPatternLength = LZ4HC_countPattern(matchPtr+sizeof(pattern), iHighLimit, pattern) + sizeof(pattern);
-                        const BYTE* const lowestMatchPtr = (lowPrefixPtr + MAX_DISTANCE >= ip) ? lowPrefixPtr : ip - MAX_DISTANCE;
+                        const BYTE* const lowestMatchPtr = (lowPrefixPtr + LZ4_DISTANCE_MAX >= ip) ? lowPrefixPtr : ip - LZ4_DISTANCE_MAX;
                         size_t const backLength = LZ4HC_reverseCountPattern(matchPtr, lowestMatchPtr, pattern);
                         size_t const currentSegmentLength = backLength + forwardPatternLength;
 
@@ -328,7 +331,7 @@ LZ4HC_InsertAndGetWiderMatch (
                                 size_t const maxML = MIN(currentSegmentLength, srcPatternLength);
                                 if ((size_t)longest < maxML) {
                                     assert(base + matchIndex < ip);
-                                    if (ip - (base+matchIndex) > MAX_DISTANCE) break;
+                                    if (ip - (base+matchIndex) > LZ4_DISTANCE_MAX) break;
                                     assert(maxML < 2 GB);
                                     longest = (int)maxML;
                                     *matchpos = base + matchIndex;   /* virtual pos, relative to ip, to retrieve offset */
@@ -343,16 +346,18 @@ LZ4HC_InsertAndGetWiderMatch (
         }   }   /* PA optimization */
 
         /* follow current chain */
-        matchIndex -= DELTANEXTU16(chainTable, matchIndex+matchChainPos);
+        matchIndex -= DELTANEXTU16(chainTable, matchIndex + matchChainPos);
 
     }  /* while ((matchIndex>=lowestMatchIndex) && (nbAttempts)) */
 
-    if (dict == usingDictCtx && nbAttempts && ipIndex - lowestMatchIndex < MAX_DISTANCE) {
-        size_t const dictEndOffset = dictCtx->end - dictCtx->base;
+    if ( dict == usingDictCtxHc
+      && nbAttempts
+      && ipIndex - lowestMatchIndex < LZ4_DISTANCE_MAX) {
+        size_t const dictEndOffset = (size_t)(dictCtx->end - dictCtx->base);
+        U32 dictMatchIndex = dictCtx->hashTable[LZ4HC_hashPtr(ip)];
         assert(dictEndOffset <= 1 GB);
-        dictMatchIndex = dictCtx->hashTable[LZ4HC_hashPtr(ip)];
         matchIndex = dictMatchIndex + lowestMatchIndex - (U32)dictEndOffset;
-        while (ipIndex - matchIndex <= MAX_DISTANCE && nbAttempts--) {
+        while (ipIndex - matchIndex <= LZ4_DISTANCE_MAX && nbAttempts--) {
             const BYTE* const matchPtr = dictCtx->base + dictMatchIndex;
 
             if (LZ4_read32(matchPtr) == pattern) {
@@ -360,22 +365,19 @@ LZ4HC_InsertAndGetWiderMatch (
                 int back = 0;
                 const BYTE* vLimit = ip + (dictEndOffset - dictMatchIndex);
                 if (vLimit > iHighLimit) vLimit = iHighLimit;
-                mlt = LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, vLimit) + MINMATCH;
+                mlt = (int)LZ4_count(ip+MINMATCH, matchPtr+MINMATCH, vLimit) + MINMATCH;
                 back = lookBackLength ? LZ4HC_countBack(ip, matchPtr, iLowLimit, dictCtx->base + dictCtx->dictLimit) : 0;
                 mlt -= back;
                 if (mlt > longest) {
                     longest = mlt;
                     *matchpos = base + matchIndex + back;
                     *startpos = ip + back;
-                }
-            }
+            }   }
 
             {   U32 const nextOffset = DELTANEXTU16(dictCtx->chainTable, dictMatchIndex);
                 dictMatchIndex -= nextOffset;
                 matchIndex -= nextOffset;
-            }
-        }
-    }
+    }   }   }
 
     return longest;
 }
@@ -395,14 +397,6 @@ int LZ4HC_InsertAndFindBestMatch(LZ4HC_CCtx_internal* const hc4,   /* Index tabl
     return LZ4HC_InsertAndGetWiderMatch(hc4, ip, ip, iLimit, MINMATCH-1, matchpos, &uselessPtr, maxNbAttempts, patternAnalysis, 0 /*chainSwap*/, dict, favorCompressionRatio);
 }
 
-
-
-typedef enum {
-    noLimit = 0,
-    limitedOutput = 1,
-    limitedDestSize = 2,
-} limitedOutput_directive;
-
 /* LZ4HC_encodeSequence() :
  * @return : 0 if ok,
  *           1 if buffer issue detected */
@@ -437,7 +431,7 @@ LZ4_FORCE_INLINE int LZ4HC_encodeSequence (
 
     /* Encode Literal length */
     length = (size_t)(*ip - *anchor);
-    if ((limit) && ((*op + (length >> 8) + length + (2 + 1 + LASTLITERALS)) > oend)) return 1;   /* Check output limit */
+    if ((limit) && ((*op + (length / 255) + length + (2 + 1 + LASTLITERALS)) > oend)) return 1;   /* Check output limit */
     if (length >= RUN_MASK) {
         size_t len = length - RUN_MASK;
         *token = (RUN_MASK << ML_BITS);
@@ -452,13 +446,13 @@ LZ4_FORCE_INLINE int LZ4HC_encodeSequence (
     *op += length;
 
     /* Encode Offset */
-    assert( (*ip - match) <= MAX_DISTANCE );   /* note : consider providing offset as a value, rather than as a pointer difference */
+    assert( (*ip - match) <= LZ4_DISTANCE_MAX );   /* note : consider providing offset as a value, rather than as a pointer difference */
     LZ4_writeLE16(*op, (U16)(*ip-match)); *op += 2;
 
     /* Encode MatchLength */
     assert(matchLength >= MINMATCH);
-    length = (size_t)(matchLength - MINMATCH);
-    if ((limit) && (*op + (length >> 8) + (1 + LASTLITERALS) > oend)) return 1;   /* Check output limit */
+    length = (size_t)matchLength - MINMATCH;
+    if ((limit) && (*op + (length / 255) + (1 + LASTLITERALS) > oend)) return 1;   /* Check output limit */
     if (length >= ML_MASK) {
         *token += ML_MASK;
         length -= ML_MASK;
@@ -511,12 +505,12 @@ LZ4_FORCE_INLINE int LZ4HC_compress_hashChain (
 
     /* init */
     *srcSizePtr = 0;
-    if (limit == limitedDestSize) oend -= LASTLITERALS;                  /* Hack for support LZ4 format restriction */
+    if (limit == fillOutput) oend -= LASTLITERALS;                  /* Hack for support LZ4 format restriction */
     if (inputSize < LZ4_minLength) goto _last_literals;                  /* Input too small, no compression (all literals) */
 
     /* Main Loop */
     while (ip <= mflimit) {
-        ml = LZ4HC_InsertAndFindBestMatch (ctx, ip, matchlimit, &ref, maxNbAttempts, patternAnalysis, dict);
+        ml = LZ4HC_InsertAndFindBestMatch(ctx, ip, matchlimit, &ref, maxNbAttempts, patternAnalysis, dict);
         if (ml<MINMATCH) { ip++; continue; }
 
         /* saved, in case we would skip too much */
@@ -533,7 +527,7 @@ _Search2:
 
         if (ml2 == ml) { /* No better match => encode ML1 */
             optr = op;
-            if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) goto _dest_overflow;
+            if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml, ref, limit, oend)) goto _dest_overflow;
             continue;
         }
 
@@ -581,10 +575,10 @@ _Search3:
             if (start2 < ip+ml)  ml = (int)(start2 - ip);
             /* Now, encode 2 sequences */
             optr = op;
-            if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) goto _dest_overflow;
+            if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml, ref, limit, oend)) goto _dest_overflow;
             ip = start2;
             optr = op;
-            if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml2, ref2, limit, oend)) goto _dest_overflow;
+            if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml2, ref2, limit, oend)) goto _dest_overflow;
             continue;
         }
 
@@ -603,7 +597,7 @@ _Search3:
                 }
 
                 optr = op;
-                if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) goto _dest_overflow;
+                if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml, ref, limit, oend)) goto _dest_overflow;
                 ip  = start3;
                 ref = ref3;
                 ml  = ml3;
@@ -641,7 +635,7 @@ _Search3:
             }
         }
         optr = op;
-        if (LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ref, limit, oend)) goto _dest_overflow;
+        if (LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml, ref, limit, oend)) goto _dest_overflow;
 
         /* ML2 becomes ML1 */
         ip = start2; ref = ref2; ml = ml2;
@@ -658,7 +652,7 @@ _last_literals:
     {   size_t lastRunSize = (size_t)(iend - anchor);  /* literals */
         size_t litLength = (lastRunSize + 255 - RUN_MASK) / 255;
         size_t const totalSize = 1 + litLength + lastRunSize;
-        if (limit == limitedDestSize) oend += LASTLITERALS;  /* restore correct value */
+        if (limit == fillOutput) oend += LASTLITERALS;  /* restore correct value */
         if (limit && (op + totalSize > oend)) {
             if (limit == limitedOutput) return 0;  /* Check output limit */
             /* adapt lastRunSize to fill 'dest' */
@@ -685,7 +679,7 @@ _last_literals:
     return (int) (((char*)op)-dest);
 
 _dest_overflow:
-    if (limit == limitedDestSize) {
+    if (limit == fillOutput) {
         op = optr;  /* restore correct out pointer */
         goto _last_literals;
     }
@@ -735,56 +729,64 @@ LZ4_FORCE_INLINE int LZ4HC_compress_generic_internal (
         { lz4opt,16384,LZ4_OPT_NUM },  /* 12==LZ4HC_CLEVEL_MAX */
     };
 
-    DEBUGLOG(4, "LZ4HC_compress_generic(%p, %p, %d)", ctx, src, *srcSizePtr);
+    DEBUGLOG(4, "LZ4HC_compress_generic(ctx=%p, src=%p, srcSize=%d)", ctx, src, *srcSizePtr);
 
-    if (limit == limitedDestSize && dstCapacity < 1) return 0;         /* Impossible to store anything */
-    if ((U32)*srcSizePtr > (U32)LZ4_MAX_INPUT_SIZE) return 0;          /* Unsupported input size (too large or negative) */
+    if (limit == fillOutput && dstCapacity < 1) return 0;   /* Impossible to store anything */
+    if ((U32)*srcSizePtr > (U32)LZ4_MAX_INPUT_SIZE) return 0;    /* Unsupported input size (too large or negative) */
 
     ctx->end += *srcSizePtr;
     if (cLevel < 1) cLevel = LZ4HC_CLEVEL_DEFAULT;   /* note : convention is different from lz4frame, maybe something to review */
     cLevel = MIN(LZ4HC_CLEVEL_MAX, cLevel);
     {   cParams_t const cParam = clTable[cLevel];
         HCfavor_e const favor = ctx->favorDecSpeed ? favorDecompressionSpeed : favorCompressionRatio;
-        if (cParam.strat == lz4hc)
-            return LZ4HC_compress_hashChain(ctx,
+        int result;
+
+        if (cParam.strat == lz4hc) {
+            result = LZ4HC_compress_hashChain(ctx,
                                 src, dst, srcSizePtr, dstCapacity,
                                 cParam.nbSearches, limit, dict);
-        assert(cParam.strat == lz4opt);
-        return LZ4HC_compress_optimal(ctx,
-                            src, dst, srcSizePtr, dstCapacity,
-                            cParam.nbSearches, cParam.targetLength, limit,
-                            cLevel == LZ4HC_CLEVEL_MAX,   /* ultra mode */
-                            dict, favor);
+        } else {
+            assert(cParam.strat == lz4opt);
+            result = LZ4HC_compress_optimal(ctx,
+                                src, dst, srcSizePtr, dstCapacity,
+                                (int)cParam.nbSearches, cParam.targetLength, limit,
+                                cLevel == LZ4HC_CLEVEL_MAX,   /* ultra mode */
+                                dict, favor);
+        }
+        if (result <= 0) ctx->dirty = 1;
+        return result;
     }
 }
 
 static void LZ4HC_setExternalDict(LZ4HC_CCtx_internal* ctxPtr, const BYTE* newBlock);
 
-static int LZ4HC_compress_generic_noDictCtx (
-    LZ4HC_CCtx_internal* const ctx,
-    const char* const src,
-    char* const dst,
-    int* const srcSizePtr,
-    int const dstCapacity,
-    int cLevel,
-    limitedOutput_directive limit
-    )
+static int
+LZ4HC_compress_generic_noDictCtx (
+        LZ4HC_CCtx_internal* const ctx,
+        const char* const src,
+        char* const dst,
+        int* const srcSizePtr,
+        int const dstCapacity,
+        int cLevel,
+        limitedOutput_directive limit
+        )
 {
     assert(ctx->dictCtx == NULL);
     return LZ4HC_compress_generic_internal(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit, noDictCtx);
 }
 
-static int LZ4HC_compress_generic_dictCtx (
-    LZ4HC_CCtx_internal* const ctx,
-    const char* const src,
-    char* const dst,
-    int* const srcSizePtr,
-    int const dstCapacity,
-    int cLevel,
-    limitedOutput_directive limit
-    )
+static int
+LZ4HC_compress_generic_dictCtx (
+        LZ4HC_CCtx_internal* const ctx,
+        const char* const src,
+        char* const dst,
+        int* const srcSizePtr,
+        int const dstCapacity,
+        int cLevel,
+        limitedOutput_directive limit
+        )
 {
-    const size_t position = ctx->end - ctx->base - ctx->lowLimit;
+    const size_t position = (size_t)(ctx->end - ctx->base) - ctx->lowLimit;
     assert(ctx->dictCtx != NULL);
     if (position >= 64 KB) {
         ctx->dictCtx = NULL;
@@ -795,19 +797,20 @@ static int LZ4HC_compress_generic_dictCtx (
         ctx->compressionLevel = (short)cLevel;
         return LZ4HC_compress_generic_noDictCtx(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit);
     } else {
-        return LZ4HC_compress_generic_internal(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit, usingDictCtx);
+        return LZ4HC_compress_generic_internal(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit, usingDictCtxHc);
     }
 }
 
-static int LZ4HC_compress_generic (
-    LZ4HC_CCtx_internal* const ctx,
-    const char* const src,
-    char* const dst,
-    int* const srcSizePtr,
-    int const dstCapacity,
-    int cLevel,
-    limitedOutput_directive limit
-    )
+static int
+LZ4HC_compress_generic (
+        LZ4HC_CCtx_internal* const ctx,
+        const char* const src,
+        char* const dst,
+        int* const srcSizePtr,
+        int const dstCapacity,
+        int cLevel,
+        limitedOutput_directive limit
+        )
 {
     if (ctx->dictCtx == NULL) {
         return LZ4HC_compress_generic_noDictCtx(ctx, src, dst, srcSizePtr, dstCapacity, cLevel, limit);
@@ -817,24 +820,41 @@ static int LZ4HC_compress_generic (
 }
 
 
-int LZ4_sizeofStateHC(void) { return sizeof(LZ4_streamHC_t); }
+int LZ4_sizeofStateHC(void) { return (int)sizeof(LZ4_streamHC_t); }
 
+#ifndef _MSC_VER  /* for some reason, Visual fails the aligment test on 32-bit x86 :
+                   * it reports an aligment of 8-bytes,
+                   * while actually aligning LZ4_streamHC_t on 4 bytes. */
+static size_t LZ4_streamHC_t_alignment(void)
+{
+    struct { char c; LZ4_streamHC_t t; } t_a;
+    return sizeof(t_a) - sizeof(t_a.t);
+}
+#endif
+
+/* state is presumed correctly initialized,
+ * in which case its size and alignment have already been validate */
 int LZ4_compress_HC_extStateHC_fastReset (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int compressionLevel)
 {
     LZ4HC_CCtx_internal* const ctx = &((LZ4_streamHC_t*)state)->internal_donotuse;
+#ifndef _MSC_VER  /* for some reason, Visual fails the aligment test on 32-bit x86 :
+                   * it reports an aligment of 8-bytes,
+                   * while actually aligning LZ4_streamHC_t on 4 bytes. */
+    assert(((size_t)state & (LZ4_streamHC_t_alignment() - 1)) == 0);  /* check alignment */
+#endif
     if (((size_t)(state)&(sizeof(void*)-1)) != 0) return 0;   /* Error : state is not aligned for pointers (32 or 64 bits) */
     LZ4_resetStreamHC_fast((LZ4_streamHC_t*)state, compressionLevel);
-    LZ4HC_init (ctx, (const BYTE*)src);
+    LZ4HC_init_internal (ctx, (const BYTE*)src);
     if (dstCapacity < LZ4_compressBound(srcSize))
         return LZ4HC_compress_generic (ctx, src, dst, &srcSize, dstCapacity, compressionLevel, limitedOutput);
     else
-        return LZ4HC_compress_generic (ctx, src, dst, &srcSize, dstCapacity, compressionLevel, noLimit);
+        return LZ4HC_compress_generic (ctx, src, dst, &srcSize, dstCapacity, compressionLevel, notLimited);
 }
 
 int LZ4_compress_HC_extStateHC (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int compressionLevel)
 {
-    if (((size_t)(state)&(sizeof(void*)-1)) != 0) return 0;   /* Error : state is not aligned for pointers (32 or 64 bits) */
-    LZ4_resetStreamHC ((LZ4_streamHC_t*)state, compressionLevel);
+    LZ4_streamHC_t* const ctx = LZ4_initStreamHC(state, sizeof(*ctx));
+    if (ctx==NULL) return 0;   /* init failure */
     return LZ4_compress_HC_extStateHC_fastReset(state, src, dst, srcSize, dstCapacity, compressionLevel);
 }
 
@@ -853,14 +873,14 @@ int LZ4_compress_HC(const char* src, char* dst, int srcSize, int dstCapacity, in
     return cSize;
 }
 
-/* LZ4_compress_HC_destSize() :
- * only compatible with regular HC parser */
-int LZ4_compress_HC_destSize(void* LZ4HC_Data, const char* source, char* dest, int* sourceSizePtr, int targetDestSize, int cLevel)
+/* state is presumed sized correctly (>= sizeof(LZ4_streamHC_t)) */
+int LZ4_compress_HC_destSize(void* state, const char* source, char* dest, int* sourceSizePtr, int targetDestSize, int cLevel)
 {
-    LZ4HC_CCtx_internal* const ctx = &((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse;
-    LZ4_resetStreamHC((LZ4_streamHC_t*)LZ4HC_Data, cLevel);
-    LZ4HC_init(ctx, (const BYTE*) source);
-    return LZ4HC_compress_generic(ctx, source, dest, sourceSizePtr, targetDestSize, cLevel, limitedDestSize);
+    LZ4_streamHC_t* const ctx = LZ4_initStreamHC(state, sizeof(*ctx));
+    if (ctx==NULL) return 0;   /* init failure */
+    LZ4HC_init_internal(&ctx->internal_donotuse, (const BYTE*) source);
+    LZ4_setCompressionLevel(ctx, cLevel);
+    return LZ4HC_compress_generic(&ctx->internal_donotuse, source, dest, sourceSizePtr, targetDestSize, cLevel, fillOutput);
 }
 
 
@@ -869,14 +889,16 @@ int LZ4_compress_HC_destSize(void* LZ4HC_Data, const char* source, char* dest, i
 *  Streaming Functions
 **************************************/
 /* allocation */
-LZ4_streamHC_t* LZ4_createStreamHC(void) {
+LZ4_streamHC_t* LZ4_createStreamHC(void)
+{
     LZ4_streamHC_t* const LZ4_streamHCPtr = (LZ4_streamHC_t*)ALLOC(sizeof(LZ4_streamHC_t));
     if (LZ4_streamHCPtr==NULL) return NULL;
-    LZ4_resetStreamHC(LZ4_streamHCPtr, LZ4HC_CLEVEL_DEFAULT);
+    LZ4_initStreamHC(LZ4_streamHCPtr, sizeof(*LZ4_streamHCPtr));  /* full initialization, malloc'ed buffer can be full of garbage */
     return LZ4_streamHCPtr;
 }
 
-int LZ4_freeStreamHC (LZ4_streamHC_t* LZ4_streamHCPtr) {
+int LZ4_freeStreamHC (LZ4_streamHC_t* LZ4_streamHCPtr)
+{
     DEBUGLOG(4, "LZ4_freeStreamHC(%p)", LZ4_streamHCPtr);
     if (!LZ4_streamHCPtr) return 0;  /* support free on NULL */
     free(LZ4_streamHCPtr);
@@ -884,29 +906,53 @@ int LZ4_freeStreamHC (LZ4_streamHC_t* LZ4_streamHCPtr) {
 }
 
 
-/* initialization */
-void LZ4_resetStreamHC (LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel)
+LZ4_streamHC_t* LZ4_initStreamHC (void* buffer, size_t size)
 {
-    LZ4_STATIC_ASSERT(sizeof(LZ4HC_CCtx_internal) <= sizeof(size_t) * LZ4_STREAMHCSIZE_SIZET);   /* if compilation fails here, LZ4_STREAMHCSIZE must be increased */
-    DEBUGLOG(4, "LZ4_resetStreamHC(%p, %d)", LZ4_streamHCPtr, compressionLevel);
+    LZ4_streamHC_t* const LZ4_streamHCPtr = (LZ4_streamHC_t*)buffer;
+    if (buffer == NULL) return NULL;
+    if (size < sizeof(LZ4_streamHC_t)) return NULL;
+#ifndef _MSC_VER  /* for some reason, Visual fails the aligment test on 32-bit x86 :
+                   * it reports an aligment of 8-bytes,
+                   * while actually aligning LZ4_streamHC_t on 4 bytes. */
+    if (((size_t)buffer) & (LZ4_streamHC_t_alignment() - 1)) return NULL;  /* alignment check */
+#endif
+    /* if compilation fails here, LZ4_STREAMHCSIZE must be increased */
+    LZ4_STATIC_ASSERT(sizeof(LZ4HC_CCtx_internal) <= LZ4_STREAMHCSIZE);
+    DEBUGLOG(4, "LZ4_initStreamHC(%p, %u)", LZ4_streamHCPtr, (unsigned)size);
+    /* end-base will trigger a clearTable on starting compression */
     LZ4_streamHCPtr->internal_donotuse.end = (const BYTE *)(ptrdiff_t)-1;
     LZ4_streamHCPtr->internal_donotuse.base = NULL;
     LZ4_streamHCPtr->internal_donotuse.dictCtx = NULL;
     LZ4_streamHCPtr->internal_donotuse.favorDecSpeed = 0;
+    LZ4_streamHCPtr->internal_donotuse.dirty = 0;
+    LZ4_setCompressionLevel(LZ4_streamHCPtr, LZ4HC_CLEVEL_DEFAULT);
+    return LZ4_streamHCPtr;
+}
+
+/* just a stub */
+void LZ4_resetStreamHC (LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel)
+{
+    LZ4_initStreamHC(LZ4_streamHCPtr, sizeof(*LZ4_streamHCPtr));
     LZ4_setCompressionLevel(LZ4_streamHCPtr, compressionLevel);
 }
 
 void LZ4_resetStreamHC_fast (LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel)
 {
     DEBUGLOG(4, "LZ4_resetStreamHC_fast(%p, %d)", LZ4_streamHCPtr, compressionLevel);
-    LZ4_streamHCPtr->internal_donotuse.end -= (uptrval)LZ4_streamHCPtr->internal_donotuse.base;
-    LZ4_streamHCPtr->internal_donotuse.base = NULL;
-    LZ4_streamHCPtr->internal_donotuse.dictCtx = NULL;
+    if (LZ4_streamHCPtr->internal_donotuse.dirty) {
+        LZ4_initStreamHC(LZ4_streamHCPtr, sizeof(*LZ4_streamHCPtr));
+    } else {
+        /* preserve end - base : can trigger clearTable's threshold */
+        LZ4_streamHCPtr->internal_donotuse.end -= (uptrval)LZ4_streamHCPtr->internal_donotuse.base;
+        LZ4_streamHCPtr->internal_donotuse.base = NULL;
+        LZ4_streamHCPtr->internal_donotuse.dictCtx = NULL;
+    }
     LZ4_setCompressionLevel(LZ4_streamHCPtr, compressionLevel);
 }
 
 void LZ4_setCompressionLevel(LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel)
 {
+    DEBUGLOG(5, "LZ4_setCompressionLevel(%p, %d)", LZ4_streamHCPtr, compressionLevel);
     if (compressionLevel < 1) compressionLevel = LZ4HC_CLEVEL_DEFAULT;
     if (compressionLevel > LZ4HC_CLEVEL_MAX) compressionLevel = LZ4HC_CLEVEL_MAX;
     LZ4_streamHCPtr->internal_donotuse.compressionLevel = (short)compressionLevel;
@@ -917,16 +963,24 @@ void LZ4_favorDecompressionSpeed(LZ4_streamHC_t* LZ4_streamHCPtr, int favor)
     LZ4_streamHCPtr->internal_donotuse.favorDecSpeed = (favor!=0);
 }
 
-int LZ4_loadDictHC (LZ4_streamHC_t* LZ4_streamHCPtr, const char* dictionary, int dictSize)
+/* LZ4_loadDictHC() :
+ * LZ4_streamHCPtr is presumed properly initialized */
+int LZ4_loadDictHC (LZ4_streamHC_t* LZ4_streamHCPtr,
+              const char* dictionary, int dictSize)
 {
     LZ4HC_CCtx_internal* const ctxPtr = &LZ4_streamHCPtr->internal_donotuse;
     DEBUGLOG(4, "LZ4_loadDictHC(%p, %p, %d)", LZ4_streamHCPtr, dictionary, dictSize);
+    assert(LZ4_streamHCPtr != NULL);
     if (dictSize > 64 KB) {
-        dictionary += dictSize - 64 KB;
+        dictionary += (size_t)dictSize - 64 KB;
         dictSize = 64 KB;
     }
-    LZ4_resetStreamHC(LZ4_streamHCPtr, ctxPtr->compressionLevel);
-    LZ4HC_init (ctxPtr, (const BYTE*)dictionary);
+    /* need a full initialization, there are bad side-effects when using resetFast() */
+    {   int const cLevel = ctxPtr->compressionLevel;
+        LZ4_initStreamHC(LZ4_streamHCPtr, sizeof(*LZ4_streamHCPtr));
+        LZ4_setCompressionLevel(LZ4_streamHCPtr, cLevel);
+    }
+    LZ4HC_init_internal (ctxPtr, (const BYTE*)dictionary);
     ctxPtr->end = (const BYTE*)dictionary + dictSize;
     if (dictSize >= 4) LZ4HC_Insert (ctxPtr, ctxPtr->end-3);
     return dictSize;
@@ -959,9 +1013,11 @@ static int LZ4_compressHC_continue_generic (LZ4_streamHC_t* LZ4_streamHCPtr,
                                             limitedOutput_directive limit)
 {
     LZ4HC_CCtx_internal* const ctxPtr = &LZ4_streamHCPtr->internal_donotuse;
-    DEBUGLOG(4, "LZ4_compressHC_continue_generic(%p, %p, %d)", LZ4_streamHCPtr, src, *srcSizePtr);
+    DEBUGLOG(4, "LZ4_compressHC_continue_generic(ctx=%p, src=%p, srcSize=%d)",
+                LZ4_streamHCPtr, src, *srcSizePtr);
+    assert(ctxPtr != NULL);
     /* auto-init if forgotten */
-    if (ctxPtr->base == NULL) LZ4HC_init (ctxPtr, (const BYTE*) src);
+    if (ctxPtr->base == NULL) LZ4HC_init_internal (ctxPtr, (const BYTE*) src);
 
     /* Check overflow */
     if ((size_t)(ctxPtr->end - ctxPtr->base) > 2 GB) {
@@ -971,7 +1027,8 @@ static int LZ4_compressHC_continue_generic (LZ4_streamHC_t* LZ4_streamHCPtr,
     }
 
     /* Check if blocks follow each other */
-    if ((const BYTE*)src != ctxPtr->end) LZ4HC_setExternalDict(ctxPtr, (const BYTE*)src);
+    if ((const BYTE*)src != ctxPtr->end)
+        LZ4HC_setExternalDict(ctxPtr, (const BYTE*)src);
 
     /* Check overlapping input/dictionary space */
     {   const BYTE* sourceEnd = (const BYTE*) src + *srcSizePtr;
@@ -992,12 +1049,12 @@ int LZ4_compress_HC_continue (LZ4_streamHC_t* LZ4_streamHCPtr, const char* src,
     if (dstCapacity < LZ4_compressBound(srcSize))
         return LZ4_compressHC_continue_generic (LZ4_streamHCPtr, src, dst, &srcSize, dstCapacity, limitedOutput);
     else
-        return LZ4_compressHC_continue_generic (LZ4_streamHCPtr, src, dst, &srcSize, dstCapacity, noLimit);
+        return LZ4_compressHC_continue_generic (LZ4_streamHCPtr, src, dst, &srcSize, dstCapacity, notLimited);
 }
 
 int LZ4_compress_HC_continue_destSize (LZ4_streamHC_t* LZ4_streamHCPtr, const char* src, char* dst, int* srcSizePtr, int targetDestSize)
 {
-    return LZ4_compressHC_continue_generic(LZ4_streamHCPtr, src, dst, srcSizePtr, targetDestSize, limitedDestSize);
+    return LZ4_compressHC_continue_generic(LZ4_streamHCPtr, src, dst, srcSizePtr, targetDestSize, fillOutput);
 }
 
 
@@ -1016,19 +1073,21 @@ int LZ4_saveDictHC (LZ4_streamHC_t* LZ4_streamHCPtr, char* safeBuffer, int dictS
     {   U32 const endIndex = (U32)(streamPtr->end - streamPtr->base);
         streamPtr->end = (const BYTE*)safeBuffer + dictSize;
         streamPtr->base = streamPtr->end - endIndex;
-        streamPtr->dictLimit = endIndex - dictSize;
-        streamPtr->lowLimit = endIndex - dictSize;
+        streamPtr->dictLimit = endIndex - (U32)dictSize;
+        streamPtr->lowLimit = endIndex - (U32)dictSize;
         if (streamPtr->nextToUpdate < streamPtr->dictLimit) streamPtr->nextToUpdate = streamPtr->dictLimit;
     }
     return dictSize;
 }
 
 
-/***********************************
+/***************************************************
 *  Deprecated Functions
-***********************************/
+***************************************************/
+
 /* These functions currently generate deprecation warnings */
-/* Deprecated compression functions */
+
+/* Wrappers for deprecated compression functions */
 int LZ4_compressHC(const char* src, char* dst, int srcSize) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), 0); }
 int LZ4_compressHC_limitedOutput(const char* src, char* dst, int srcSize, int maxDstSize) { return LZ4_compress_HC(src, dst, srcSize, maxDstSize, 0); }
 int LZ4_compressHC2(const char* src, char* dst, int srcSize, int cLevel) { return LZ4_compress_HC (src, dst, srcSize, LZ4_compressBound(srcSize), cLevel); }
@@ -1044,25 +1103,26 @@ int LZ4_compressHC_limitedOutput_continue (LZ4_streamHC_t* ctx, const char* src,
 /* Deprecated streaming functions */
 int LZ4_sizeofStreamStateHC(void) { return LZ4_STREAMHCSIZE; }
 
+/* state is presumed correctly sized, aka >= sizeof(LZ4_streamHC_t)
+ * @return : 0 on success, !=0 if error */
 int LZ4_resetStreamStateHC(void* state, char* inputBuffer)
 {
-    LZ4HC_CCtx_internal *ctx = &((LZ4_streamHC_t*)state)->internal_donotuse;
-    if ((((size_t)state) & (sizeof(void*)-1)) != 0) return 1;   /* Error : pointer is not aligned for pointer (32 or 64 bits) */
-    LZ4_resetStreamHC((LZ4_streamHC_t*)state, ((LZ4_streamHC_t*)state)->internal_donotuse.compressionLevel);
-    LZ4HC_init(ctx, (const BYTE*)inputBuffer);
+    LZ4_streamHC_t* const hc4 = LZ4_initStreamHC(state, sizeof(*hc4));
+    if (hc4 == NULL) return 1;   /* init failed */
+    LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer);
     return 0;
 }
 
 void* LZ4_createHC (const char* inputBuffer)
 {
-    LZ4_streamHC_t* hc4 = (LZ4_streamHC_t*)ALLOC(sizeof(LZ4_streamHC_t));
+    LZ4_streamHC_t* const hc4 = LZ4_createStreamHC();
     if (hc4 == NULL) return NULL;   /* not enough memory */
-    LZ4_resetStreamHC(hc4, 0 /* compressionLevel */);
-    LZ4HC_init (&hc4->internal_donotuse, (const BYTE*)inputBuffer);
+    LZ4HC_init_internal (&hc4->internal_donotuse, (const BYTE*)inputBuffer);
     return hc4;
 }
 
-int LZ4_freeHC (void* LZ4HC_Data) {
+int LZ4_freeHC (void* LZ4HC_Data)
+{
     if (!LZ4HC_Data) return 0;  /* support free on NULL */
     FREEMEM(LZ4HC_Data);
     return 0;
@@ -1070,7 +1130,7 @@ int LZ4_freeHC (void* LZ4HC_Data) {
 
 int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int cLevel)
 {
-    return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, 0, cLevel, noLimit);
+    return LZ4HC_compress_generic (&((LZ4_streamHC_t*)LZ4HC_Data)->internal_donotuse, src, dst, &srcSize, 0, cLevel, notLimited);
 }
 
 int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* src, char* dst, int srcSize, int dstCapacity, int cLevel)
@@ -1089,7 +1149,7 @@ char* LZ4_slideInputBufferHC(void* LZ4HC_Data)
 
 
 /* ================================================
- * LZ4 Optimal parser (levels 10-12)
+ *  LZ4 Optimal parser (levels [LZ4HC_CLEVEL_OPT_MIN - LZ4HC_CLEVEL_MAX])
  * ===============================================*/
 typedef struct {
     int price;
@@ -1102,8 +1162,9 @@ typedef struct {
 LZ4_FORCE_INLINE int LZ4HC_literalsPrice(int const litlen)
 {
     int price = litlen;
+    assert(litlen >= 0);
     if (litlen >= (int)RUN_MASK)
-        price += 1 + (litlen-RUN_MASK)/255;
+        price += 1 + ((litlen-(int)RUN_MASK) / 255);
     return price;
 }
 
@@ -1112,11 +1173,13 @@ LZ4_FORCE_INLINE int LZ4HC_literalsPrice(int const litlen)
 LZ4_FORCE_INLINE int LZ4HC_sequencePrice(int litlen, int mlen)
 {
     int price = 1 + 2 ; /* token + 16-bit offset */
+    assert(litlen >= 0);
+    assert(mlen >= MINMATCH);
 
     price += LZ4HC_literalsPrice(litlen);
 
     if (mlen >= (int)(ML_MASK+MINMATCH))
-        price += 1 + (mlen-(ML_MASK+MINMATCH))/255;
+        price += 1 + ((mlen-(int)(ML_MASK+MINMATCH)) / 255);
 
     return price;
 }
@@ -1175,9 +1238,9 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx,
     BYTE* oend = op + dstCapacity;
 
     /* init */
-    DEBUGLOG(5, "LZ4HC_compress_optimal");
+    DEBUGLOG(5, "LZ4HC_compress_optimal(dst=%p, dstCapa=%u)", dst, (unsigned)dstCapacity);
     *srcSizePtr = 0;
-    if (limit == limitedDestSize) oend -= LASTLITERALS;   /* Hack for support LZ4 format restriction */
+    if (limit == fillOutput) oend -= LASTLITERALS;   /* Hack for support LZ4 format restriction */
     if (sufficient_len >= LZ4_OPT_NUM) sufficient_len = LZ4_OPT_NUM-1;
 
     /* Main Loop */
@@ -1195,7 +1258,7 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx,
              int const firstML = firstMatch.len;
              const BYTE* const matchPos = ip - firstMatch.off;
              opSaved = op;
-             if ( LZ4HC_encodeSequence(&ip, &op, &anchor, firstML, matchPos, limit, oend) )   /* updates ip, op and anchor */
+             if ( LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), firstML, matchPos, limit, oend) )   /* updates ip, op and anchor */
                  goto _dest_overflow;
              continue;
          }
@@ -1365,9 +1428,9 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx,
                  if (ml == 1) { ip++; rPos++; continue; }  /* literal; note: can end up with several literals, in which case, skip them */
                  rPos += ml;
                  assert(ml >= MINMATCH);
-                 assert((offset >= 1) && (offset <= MAX_DISTANCE));
+                 assert((offset >= 1) && (offset <= LZ4_DISTANCE_MAX));
                  opSaved = op;
-                 if ( LZ4HC_encodeSequence(&ip, &op, &anchor, ml, ip - offset, limit, oend) )   /* updates ip, op and anchor */
+                 if ( LZ4HC_encodeSequence(UPDATABLE(ip, op, anchor), ml, ip - offset, limit, oend) )   /* updates ip, op and anchor */
                      goto _dest_overflow;
          }   }
      }  /* while (ip <= mflimit) */
@@ -1377,7 +1440,7 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx,
      {   size_t lastRunSize = (size_t)(iend - anchor);  /* literals */
          size_t litLength = (lastRunSize + 255 - RUN_MASK) / 255;
          size_t const totalSize = 1 + litLength + lastRunSize;
-         if (limit == limitedDestSize) oend += LASTLITERALS;  /* restore correct value */
+         if (limit == fillOutput) oend += LASTLITERALS;  /* restore correct value */
          if (limit && (op + totalSize > oend)) {
              if (limit == limitedOutput) return 0;  /* Check output limit */
              /* adapt lastRunSize to fill 'dst' */
@@ -1404,7 +1467,7 @@ static int LZ4HC_compress_optimal ( LZ4HC_CCtx_internal* ctx,
      return (int) ((char*)op-dst);
 
  _dest_overflow:
-     if (limit == limitedDestSize) {
+     if (limit == fillOutput) {
          op = opSaved;  /* restore correct out pointer */
          goto _last_literals;
      }
diff --git a/lib/lz4hc.h b/lib/lz4hc.h
index 970fa39..cdc6d89 100644
--- a/lib/lz4hc.h
+++ b/lib/lz4hc.h
@@ -54,7 +54,7 @@ extern "C" {
  *  Block Compression
  **************************************/
 /*! LZ4_compress_HC() :
- *  Compress data from `src` into `dst`, using the more powerful but slower "HC" algorithm.
+ *  Compress data from `src` into `dst`, using the powerful but slower "HC" algorithm.
  * `dst` must be already allocated.
  *  Compression is guaranteed to succeed if `dstCapacity >= LZ4_compressBound(srcSize)` (see "lz4.h")
  *  Max supported `srcSize` value is LZ4_MAX_INPUT_SIZE (see "lz4.h")
@@ -77,7 +77,21 @@ LZ4LIB_API int LZ4_compress_HC (const char* src, char* dst, int srcSize, int dst
  *  Memory segment must be aligned on 8-bytes boundaries (which a normal malloc() should do properly).
  */
 LZ4LIB_API int LZ4_sizeofStateHC(void);
-LZ4LIB_API int LZ4_compress_HC_extStateHC(void* state, const char* src, char* dst, int srcSize, int maxDstSize, int compressionLevel);
+LZ4LIB_API int LZ4_compress_HC_extStateHC(void* stateHC, const char* src, char* dst, int srcSize, int maxDstSize, int compressionLevel);
+
+
+/*! LZ4_compress_HC_destSize() : v1.9.0+
+ *  Will compress as much data as possible from `src`
+ *  to fit into `targetDstSize` budget.
+ *  Result is provided in 2 parts :
+ * @return : the number of bytes written into 'dst' (necessarily <= targetDstSize)
+ *           or 0 if compression fails.
+ * `srcSizePtr` : on success, *srcSizePtr is updated to indicate how much bytes were read from `src`
+ */
+LZ4LIB_API int LZ4_compress_HC_destSize(void* stateHC,
+                                  const char* src, char* dst,
+                                        int* srcSizePtr, int targetDstSize,
+                                        int compressionLevel);
 
 
 /*-************************************
@@ -89,46 +103,92 @@ LZ4LIB_API int LZ4_compress_HC_extStateHC(void* state, const char* src, char* ds
 /*! LZ4_createStreamHC() and LZ4_freeStreamHC() :
  *  These functions create and release memory for LZ4 HC streaming state.
  *  Newly created states are automatically initialized.
- *  Existing states can be re-used several times, using LZ4_resetStreamHC().
- *  These methods are API and ABI stable, they can be used in combination with a DLL.
+ *  A same state can be used multiple times consecutively,
+ *  starting with LZ4_resetStreamHC_fast() to start a new stream of blocks.
  */
 LZ4LIB_API LZ4_streamHC_t* LZ4_createStreamHC(void);
 LZ4LIB_API int             LZ4_freeStreamHC (LZ4_streamHC_t* streamHCPtr);
 
-LZ4LIB_API void LZ4_resetStreamHC (LZ4_streamHC_t* streamHCPtr, int compressionLevel);
-LZ4LIB_API int  LZ4_loadDictHC (LZ4_streamHC_t* streamHCPtr, const char* dictionary, int dictSize);
-
-LZ4LIB_API int LZ4_compress_HC_continue (LZ4_streamHC_t* streamHCPtr, const char* src, char* dst, int srcSize, int maxDstSize);
-
-LZ4LIB_API int LZ4_saveDictHC (LZ4_streamHC_t* streamHCPtr, char* safeBuffer, int maxDictSize);
-
 /*
-  These functions compress data in successive blocks of any size, using previous blocks as dictionary.
+  These functions compress data in successive blocks of any size,
+  using previous blocks as dictionary, to improve compression ratio.
   One key assumption is that previous blocks (up to 64 KB) remain read-accessible while compressing next blocks.
   There is an exception for ring buffers, which can be smaller than 64 KB.
-  Ring buffers scenario is automatically detected and handled by LZ4_compress_HC_continue().
+  Ring-buffer scenario is automatically detected and handled within LZ4_compress_HC_continue().
+
+  Before starting compression, state must be allocated and properly initialized.
+  LZ4_createStreamHC() does both, though compression level is set to LZ4HC_CLEVEL_DEFAULT.
+
+  Selecting the compression level can be done with LZ4_resetStreamHC_fast() (starts a new stream)
+  or LZ4_setCompressionLevel() (anytime, between blocks in the same stream) (experimental).
+  LZ4_resetStreamHC_fast() only works on states which have been properly initialized at least once,
+  which is automatically the case when state is created using LZ4_createStreamHC().
+
+  After reset, a first "fictional block" can be designated as initial dictionary,
+  using LZ4_loadDictHC() (Optional).
+
+  Invoke LZ4_compress_HC_continue() to compress each successive block.
+  The number of blocks is unlimited.
+  Previous input blocks, including initial dictionary when present,
+  must remain accessible and unmodified during compression.
+
+  It's allowed to update compression level anytime between blocks,
+  using LZ4_setCompressionLevel() (experimental).
+
+  'dst' buffer should be sized to handle worst case scenarios
+  (see LZ4_compressBound(), it ensures compression success).
+  In case of failure, the API does not guarantee recovery,
+  so the state _must_ be reset.
+  To ensure compression success
+  whenever `dst` buffer size cannot be made >= LZ4_compressBound(),
+  consider using LZ4_compress_HC_continue_destSize().
+
+  Whenever previous input blocks can't be preserved unmodified in-place during compression of next blocks,
+  it's possible to copy the last blocks into a more stable memory space, using LZ4_saveDictHC().
+  Return value of LZ4_saveDictHC() is the size of dictionary effectively saved into 'safeBuffer' (<= 64 KB)
+
+  After completing a streaming compression,
+  it's possible to start a new stream of blocks, using the same LZ4_streamHC_t state,
+  just by resetting it, using LZ4_resetStreamHC_fast().
+*/
 
-  Before starting compression, state must be properly initialized, using LZ4_resetStreamHC().
-  A first "fictional block" can then be designated as initial dictionary, using LZ4_loadDictHC() (Optional).
+LZ4LIB_API void LZ4_resetStreamHC_fast(LZ4_streamHC_t* streamHCPtr, int compressionLevel);   /* v1.9.0+ */
+LZ4LIB_API int  LZ4_loadDictHC (LZ4_streamHC_t* streamHCPtr, const char* dictionary, int dictSize);
+
+LZ4LIB_API int LZ4_compress_HC_continue (LZ4_streamHC_t* streamHCPtr,
+                                   const char* src, char* dst,
+                                         int srcSize, int maxDstSize);
+
+/*! LZ4_compress_HC_continue_destSize() : v1.9.0+
+ *  Similar to LZ4_compress_HC_continue(),
+ *  but will read as much data as possible from `src`
+ *  to fit into `targetDstSize` budget.
+ *  Result is provided into 2 parts :
+ * @return : the number of bytes written into 'dst' (necessarily <= targetDstSize)
+ *           or 0 if compression fails.
+ * `srcSizePtr` : on success, *srcSizePtr will be updated to indicate how much bytes were read from `src`.
+ *           Note that this function may not consume the entire input.
+ */
+LZ4LIB_API int LZ4_compress_HC_continue_destSize(LZ4_streamHC_t* LZ4_streamHCPtr,
+                                           const char* src, char* dst,
+                                                 int* srcSizePtr, int targetDstSize);
+
+LZ4LIB_API int LZ4_saveDictHC (LZ4_streamHC_t* streamHCPtr, char* safeBuffer, int maxDictSize);
 
-  Then, use LZ4_compress_HC_continue() to compress each successive block.
-  Previous memory blocks (including initial dictionary when present) must remain accessible and unmodified during compression.
-  'dst' buffer should be sized to handle worst case scenarios (see LZ4_compressBound()), to ensure operation success.
-  Because in case of failure, the API does not guarantee context recovery, and context will have to be reset.
-  If `dst` buffer budget cannot be >= LZ4_compressBound(), consider using LZ4_compress_HC_continue_destSize() instead.
 
-  If, for any reason, previous data block can't be preserved unmodified in memory for next compression block,
-  you can save it to a more stable memory space, using LZ4_saveDictHC().
-  Return value of LZ4_saveDictHC() is the size of dictionary effectively saved into 'safeBuffer'.
-*/
 
+/*^**********************************************
+ * !!!!!!   STATIC LINKING ONLY   !!!!!!
+ ***********************************************/
 
-/*-**************************************************************
+/*-******************************************************************
  * PRIVATE DEFINITIONS :
- * Do not use these definitions.
- * They are exposed to allow static allocation of `LZ4_streamHC_t`.
- * Using these definitions makes the code vulnerable to potential API break when upgrading LZ4
- ****************************************************************/
+ * Do not use these definitions directly.
+ * They are merely exposed to allow static allocation of `LZ4_streamHC_t`.
+ * Declare an `LZ4_streamHC_t` directly, rather than any type below.
+ * Even then, only do so in the context of static linking, as definitions may change between versions.
+ ********************************************************************/
+
 #define LZ4HC_DICTIONARY_LOGSIZE 16
 #define LZ4HC_MAXD (1<<LZ4HC_DICTIONARY_LOGSIZE)
 #define LZ4HC_MAXD_MASK (LZ4HC_MAXD - 1)
@@ -153,7 +213,9 @@ struct LZ4HC_CCtx_internal
     uint32_t   lowLimit;        /* below that point, no more dict */
     uint32_t   nextToUpdate;    /* index from which to continue dictionary update */
     short      compressionLevel;
-    short      favorDecSpeed;
+    int8_t     favorDecSpeed;   /* favor decompression speed if this flag set,
+                                   otherwise, favor compression ratio */
+    int8_t     dirty;           /* stream has to be fully reset if this flag is set */
     const LZ4HC_CCtx_internal* dictCtx;
 };
 
@@ -171,26 +233,43 @@ struct LZ4HC_CCtx_internal
     unsigned int   lowLimit;         /* below that point, no more dict */
     unsigned int   nextToUpdate;     /* index from which to continue dictionary update */
     short          compressionLevel;
-    short          favorDecSpeed;
+    char           favorDecSpeed;    /* favor decompression speed if this flag set,
+                                        otherwise, favor compression ratio */
+    char           dirty;            /* stream has to be fully reset if this flag is set */
     const LZ4HC_CCtx_internal* dictCtx;
 };
 
 #endif
 
-#define LZ4_STREAMHCSIZE       (4*LZ4HC_HASHTABLESIZE + 2*LZ4HC_MAXD + 56) /* 262200 */
+
+/* Do not use these definitions directly !
+ * Declare or allocate an LZ4_streamHC_t instead.
+ */
+#define LZ4_STREAMHCSIZE       (4*LZ4HC_HASHTABLESIZE + 2*LZ4HC_MAXD + 56 + ((sizeof(void*)==16) ? 56 : 0) /* AS400*/ ) /* 262200 or 262256*/
 #define LZ4_STREAMHCSIZE_SIZET (LZ4_STREAMHCSIZE / sizeof(size_t))
 union LZ4_streamHC_u {
     size_t table[LZ4_STREAMHCSIZE_SIZET];
     LZ4HC_CCtx_internal internal_donotuse;
-};   /* previously typedef'd to LZ4_streamHC_t */
-/*
-  LZ4_streamHC_t :
-  This structure allows static allocation of LZ4 HC streaming state.
-  State must be initialized using LZ4_resetStreamHC() before first use.
+}; /* previously typedef'd to LZ4_streamHC_t */
 
-  Static allocation shall only be used in combination with static linking.
-  When invoking LZ4 from a DLL, use create/free functions instead, which are API and ABI stable.
-*/
+/* LZ4_streamHC_t :
+ * This structure allows static allocation of LZ4 HC streaming state.
+ * This can be used to allocate statically, on state, or as part of a larger structure.
+ *
+ * Such state **must** be initialized using LZ4_initStreamHC() before first use.
+ *
+ * Note that invoking LZ4_initStreamHC() is not required when
+ * the state was created using LZ4_createStreamHC() (which is recommended).
+ * Using the normal builder, a newly created state is automatically initialized.
+ *
+ * Static allocation shall only be used in combination with static linking.
+ */
+
+/* LZ4_initStreamHC() : v1.9.0+
+ * Required before first use of a statically allocated LZ4_streamHC_t.
+ * Before v1.9.0 : use LZ4_resetStreamHC() instead
+ */
+LZ4LIB_API LZ4_streamHC_t* LZ4_initStreamHC (void* buffer, size_t size);
 
 
 /*-************************************
@@ -201,11 +280,11 @@ union LZ4_streamHC_u {
 /* deprecated compression functions */
 LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC               (const char* source, char* dest, int inputSize);
 LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC_limitedOutput (const char* source, char* dest, int inputSize, int maxOutputSize);
-LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC2 (const char* source, char* dest, int inputSize, int compressionLevel);
-LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC2_limitedOutput (const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
+LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC2              (const char* source, char* dest, int inputSize, int compressionLevel);
+LZ4_DEPRECATED("use LZ4_compress_HC() instead") LZ4LIB_API int LZ4_compressHC2_limitedOutput(const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
 LZ4_DEPRECATED("use LZ4_compress_HC_extStateHC() instead") LZ4LIB_API int LZ4_compressHC_withStateHC               (void* state, const char* source, char* dest, int inputSize);
 LZ4_DEPRECATED("use LZ4_compress_HC_extStateHC() instead") LZ4LIB_API int LZ4_compressHC_limitedOutput_withStateHC (void* state, const char* source, char* dest, int inputSize, int maxOutputSize);
-LZ4_DEPRECATED("use LZ4_compress_HC_extStateHC() instead") LZ4LIB_API int LZ4_compressHC2_withStateHC (void* state, const char* source, char* dest, int inputSize, int compressionLevel);
+LZ4_DEPRECATED("use LZ4_compress_HC_extStateHC() instead") LZ4LIB_API int LZ4_compressHC2_withStateHC              (void* state, const char* source, char* dest, int inputSize, int compressionLevel);
 LZ4_DEPRECATED("use LZ4_compress_HC_extStateHC() instead") LZ4LIB_API int LZ4_compressHC2_limitedOutput_withStateHC(void* state, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
 LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_compressHC_continue               (LZ4_streamHC_t* LZ4_streamHCPtr, const char* source, char* dest, int inputSize);
 LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_compressHC_limitedOutput_continue (LZ4_streamHC_t* LZ4_streamHCPtr, const char* source, char* dest, int inputSize, int maxOutputSize);
@@ -221,10 +300,21 @@ LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_comp
 LZ4_DEPRECATED("use LZ4_createStreamHC() instead") LZ4LIB_API void* LZ4_createHC (const char* inputBuffer);
 LZ4_DEPRECATED("use LZ4_saveDictHC() instead") LZ4LIB_API     char* LZ4_slideInputBufferHC (void* LZ4HC_Data);
 LZ4_DEPRECATED("use LZ4_freeStreamHC() instead") LZ4LIB_API   int   LZ4_freeHC (void* LZ4HC_Data);
-LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_compressHC2_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int compressionLevel);
+LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_compressHC2_continue               (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int compressionLevel);
 LZ4_DEPRECATED("use LZ4_compress_HC_continue() instead") LZ4LIB_API int LZ4_compressHC2_limitedOutput_continue (void* LZ4HC_Data, const char* source, char* dest, int inputSize, int maxOutputSize, int compressionLevel);
 LZ4_DEPRECATED("use LZ4_createStreamHC() instead") LZ4LIB_API int   LZ4_sizeofStreamStateHC(void);
-LZ4_DEPRECATED("use LZ4_resetStreamHC() instead") LZ4LIB_API  int   LZ4_resetStreamStateHC(void* state, char* inputBuffer);
+LZ4_DEPRECATED("use LZ4_initStreamHC() instead") LZ4LIB_API  int   LZ4_resetStreamStateHC(void* state, char* inputBuffer);
+
+
+/* LZ4_resetStreamHC() is now replaced by LZ4_initStreamHC().
+ * The intention is to emphasize the difference with LZ4_resetStreamHC_fast(),
+ * which is now the recommended function to start a new stream of blocks,
+ * but cannot be used to initialize a memory segment containing arbitrary garbage data.
+ *
+ * It is recommended to switch to LZ4_initStreamHC().
+ * LZ4_resetStreamHC() will generate deprecation warnings in a future version.
+ */
+LZ4LIB_API void LZ4_resetStreamHC (LZ4_streamHC_t* streamHCPtr, int compressionLevel);
 
 
 #if defined (__cplusplus)
@@ -250,44 +340,22 @@ LZ4_DEPRECATED("use LZ4_resetStreamHC() instead") LZ4LIB_API  int   LZ4_resetStr
 extern "C" {
 #endif
 
-/*! LZ4_compress_HC_destSize() : v1.8.0 (experimental)
- *  Will try to compress as much data from `src` as possible
- *  that can fit into `targetDstSize` budget.
- *  Result is provided in 2 parts :
- * @return : the number of bytes written into 'dst'
- *           or 0 if compression fails.
- * `srcSizePtr` : value will be updated to indicate how much bytes were read from `src`
- */
-int LZ4_compress_HC_destSize(void* LZ4HC_Data,
-                             const char* src, char* dst,
-                             int* srcSizePtr, int targetDstSize,
-                             int compressionLevel);
-
-/*! LZ4_compress_HC_continue_destSize() : v1.8.0 (experimental)
- *  Similar as LZ4_compress_HC_continue(),
- *  but will read a variable nb of bytes from `src`
- *  to fit into `targetDstSize` budget.
- *  Result is provided in 2 parts :
- * @return : the number of bytes written into 'dst'
- *           or 0 if compression fails.
- * `srcSizePtr` : value will be updated to indicate how much bytes were read from `src`.
- */
-int LZ4_compress_HC_continue_destSize(LZ4_streamHC_t* LZ4_streamHCPtr,
-                            const char* src, char* dst,
-                            int* srcSizePtr, int targetDstSize);
-
-/*! LZ4_setCompressionLevel() : v1.8.0 (experimental)
- *  It's possible to change compression level between 2 invocations of LZ4_compress_HC_continue*()
+/*! LZ4_setCompressionLevel() : v1.8.0+ (experimental)
+ *  It's possible to change compression level
+ *  between successive invocations of LZ4_compress_HC_continue*()
+ *  for dynamic adaptation.
  */
-void LZ4_setCompressionLevel(LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel);
+LZ4LIB_STATIC_API void LZ4_setCompressionLevel(
+    LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel);
 
-/*! LZ4_favorDecompressionSpeed() : v1.8.2 (experimental)
- *  Parser will select decisions favoring decompression over compression ratio.
- *  Only work at highest compression settings (level >= LZ4HC_CLEVEL_OPT_MIN)
+/*! LZ4_favorDecompressionSpeed() : v1.8.2+ (experimental)
+ *  Opt. Parser will favor decompression speed over compression ratio.
+ *  Only applicable to levels >= LZ4HC_CLEVEL_OPT_MIN.
  */
-void LZ4_favorDecompressionSpeed(LZ4_streamHC_t* LZ4_streamHCPtr, int favor);
+LZ4LIB_STATIC_API void LZ4_favorDecompressionSpeed(
+    LZ4_streamHC_t* LZ4_streamHCPtr, int favor);
 
-/*! LZ4_resetStreamHC_fast() :
+/*! LZ4_resetStreamHC_fast() : v1.9.0+
  *  When an LZ4_streamHC_t is known to be in a internally coherent state,
  *  it can often be prepared for a new compression with almost no work, only
  *  sometimes falling back to the full, expensive reset that is always required
@@ -304,8 +372,14 @@ void LZ4_favorDecompressionSpeed(LZ4_streamHC_t* LZ4_streamHCPtr, int favor);
  *  - the stream was in an indeterminate state and was used in a compression
  *    call that fully reset the state (LZ4_compress_HC_extStateHC()) and that
  *    returned success
+ *
+ *  Note:
+ *  A stream that was last used in a compression call that returned an error
+ *  may be passed to this function. However, it will be fully reset, which will
+ *  clear any existing history and settings from the context.
  */
-void LZ4_resetStreamHC_fast(LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel);
+LZ4LIB_STATIC_API void LZ4_resetStreamHC_fast(
+    LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLevel);
 
 /*! LZ4_compress_HC_extStateHC_fastReset() :
  *  A variant of LZ4_compress_HC_extStateHC().
@@ -318,7 +392,11 @@ void LZ4_resetStreamHC_fast(LZ4_streamHC_t* LZ4_streamHCPtr, int compressionLeve
  *  LZ4_resetStreamHC_fast() while LZ4_compress_HC_extStateHC() starts with a
  *  call to LZ4_resetStreamHC().
  */
-int LZ4_compress_HC_extStateHC_fastReset (void* state, const char* src, char* dst, int srcSize, int dstCapacity, int compressionLevel);
+LZ4LIB_STATIC_API int LZ4_compress_HC_extStateHC_fastReset (
+    void* state,
+    const char* src, char* dst,
+    int srcSize, int dstCapacity,
+    int compressionLevel);
 
 /*! LZ4_attach_HC_dictionary() :
  *  This is an experimental API that allows for the efficient use of a
@@ -345,7 +423,9 @@ int LZ4_compress_HC_extStateHC_fastReset (void* state, const char* src, char* ds
  *  stream (and source buffer) must remain in-place / accessible / unchanged
  *  through the lifetime of the stream session.
  */
-LZ4LIB_API void LZ4_attach_HC_dictionary(LZ4_streamHC_t *working_stream, const LZ4_streamHC_t *dictionary_stream);
+LZ4LIB_STATIC_API void LZ4_attach_HC_dictionary(
+          LZ4_streamHC_t *working_stream,
+    const LZ4_streamHC_t *dictionary_stream);
 
 #if defined (__cplusplus)
 }
diff --git a/lib/xxhash.c b/lib/xxhash.c
index 3fc97fd..ff28749 100644
--- a/lib/xxhash.c
+++ b/lib/xxhash.c
@@ -50,20 +50,26 @@
  * Prefer these methods in priority order (0 > 1 > 2)
  */
 #ifndef XXH_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__) )
+#  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__) ))
 #    define XXH_FORCE_MEMORY_ACCESS 1
 #  endif
 #endif
 
 /*!XXH_ACCEPT_NULL_INPUT_POINTER :
- * If the input pointer is a null pointer, xxHash default behavior is to trigger a memory access error, since it is a bad pointer.
- * When this option is enabled, xxHash output for null input pointers will be the same as a null-length input.
- * By default, this option is disabled. To enable it, uncomment below define :
+ * If input pointer is NULL, xxHash default behavior is to dereference it, triggering a segfault.
+ * When this macro is enabled, xxHash actively checks input for null pointer.
+ * It it is, result for null input pointers is the same as a null-length input.
  */
-/* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */
+#ifndef XXH_ACCEPT_NULL_INPUT_POINTER   /* can be defined externally */
+#  define XXH_ACCEPT_NULL_INPUT_POINTER 0
+#endif
 
 /*!XXH_FORCE_NATIVE_FORMAT :
  * By default, xxHash library provides endian-independent Hash values, based on little-endian convention.
@@ -80,8 +86,9 @@
 /*!XXH_FORCE_ALIGN_CHECK :
  * This is a minor performance trick, only useful with lots of very small keys.
  * It means : check for aligned/unaligned input.
- * The check costs one initial branch per hash; set to 0 when the input data
- * is guaranteed to be aligned.
+ * The check costs one initial branch per hash;
+ * set it to 0 when the input is guaranteed to be aligned,
+ * or when alignment doesn't matter for performance.
  */
 #ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */
 #  if defined(__i386) || defined(_M_IX86) || defined(__x86_64__) || defined(_M_X64)
@@ -104,6 +111,8 @@ static void  XXH_free  (void* p)  { free(p); }
 #include <string.h>
 static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); }
 
+#include <assert.h>   /* assert */
+
 #define XXH_STATIC_LINKING_ONLY
 #include "xxhash.h"
 
@@ -113,40 +122,35 @@ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcp
 ***************************************/
 #ifdef _MSC_VER    /* Visual Studio */
 #  pragma warning(disable : 4127)      /* disable: C4127: conditional expression is constant */
-#endif
-
-#ifndef XXH_FORCE_INLINE
-#  ifdef _MSC_VER    /* Visual Studio */
-#    define XXH_FORCE_INLINE static __forceinline
-#  else
-#    if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
-#      ifdef __GNUC__
-#        define XXH_FORCE_INLINE static inline __attribute__((always_inline))
-#      else
-#        define XXH_FORCE_INLINE static inline
-#      endif
+#  define FORCE_INLINE static __forceinline
+#else
+#  if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
+#    ifdef __GNUC__
+#      define FORCE_INLINE static inline __attribute__((always_inline))
 #    else
-#      define XXH_FORCE_INLINE static
-#    endif /* __STDC_VERSION__ */
-#  endif  /* _MSC_VER */
-#endif /* XXH_FORCE_INLINE */
+#      define FORCE_INLINE static inline
+#    endif
+#  else
+#    define FORCE_INLINE static
+#  endif /* __STDC_VERSION__ */
+#endif
 
 
 /* *************************************
 *  Basic Types
 ***************************************/
 #ifndef MEM_MODULE
-# if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
+# if !defined (__VMS) \
+  && (defined (__cplusplus) \
+  || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
 #   include <stdint.h>
     typedef uint8_t  BYTE;
     typedef uint16_t U16;
     typedef uint32_t U32;
-    typedef  int32_t S32;
 # else
     typedef unsigned char      BYTE;
     typedef unsigned short     U16;
     typedef unsigned int       U32;
-    typedef   signed int       S32;
 # endif
 #endif
 
@@ -213,8 +217,12 @@ typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
 
 /* XXH_CPU_LITTLE_ENDIAN can be defined externally, for example on the compiler command line */
 #ifndef XXH_CPU_LITTLE_ENDIAN
-    static const int g_one = 1;
-#   define XXH_CPU_LITTLE_ENDIAN   (*(const char*)(&g_one))
+static int XXH_isLittleEndian(void)
+{
+    const union { U32 u; BYTE c[4]; } one = { 1 };   /* don't use static : performance detrimental  */
+    return one.c[0];
+}
+#   define XXH_CPU_LITTLE_ENDIAN   XXH_isLittleEndian()
 #endif
 
 
@@ -223,7 +231,7 @@ typedef enum { XXH_bigEndian=0, XXH_littleEndian=1 } XXH_endianess;
 *****************************/
 typedef enum { XXH_aligned, XXH_unaligned } XXH_alignment;
 
-XXH_FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
+FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
 {
     if (align==XXH_unaligned)
         return endian==XXH_littleEndian ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr));
@@ -231,7 +239,7 @@ XXH_FORCE_INLINE U32 XXH_readLE32_align(const void* ptr, XXH_endianess endian, X
         return endian==XXH_littleEndian ? *(const U32*)ptr : XXH_swap32(*(const U32*)ptr);
 }
 
-XXH_FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianess endian)
+FORCE_INLINE U32 XXH_readLE32(const void* ptr, XXH_endianess endian)
 {
     return XXH_readLE32_align(ptr, endian, XXH_unaligned);
 }
@@ -245,12 +253,12 @@ static U32 XXH_readBE32(const void* ptr)
 /* *************************************
 *  Macros
 ***************************************/
-#define XXH_STATIC_ASSERT(c)   { enum { XXH_static_assert = 1/(int)(!!(c)) }; }    /* use only *after* variable declarations */
+#define XXH_STATIC_ASSERT(c)  { enum { XXH_sa = 1/(int)(!!(c)) }; }  /* use after variable declarations */
 XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; }
 
 
 /* *******************************************************************
-*  32-bits hash functions
+*  32-bit hash functions
 *********************************************************************/
 static const U32 PRIME32_1 = 2654435761U;
 static const U32 PRIME32_2 = 2246822519U;
@@ -266,14 +274,89 @@ static U32 XXH32_round(U32 seed, U32 input)
     return seed;
 }
 
-XXH_FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed, XXH_endianess endian, XXH_alignment align)
+/* mix all bits */
+static U32 XXH32_avalanche(U32 h32)
+{
+    h32 ^= h32 >> 15;
+    h32 *= PRIME32_2;
+    h32 ^= h32 >> 13;
+    h32 *= PRIME32_3;
+    h32 ^= h32 >> 16;
+    return(h32);
+}
+
+#define XXH_get32bits(p) XXH_readLE32_align(p, endian, align)
+
+static U32
+XXH32_finalize(U32 h32, const void* ptr, size_t len,
+                XXH_endianess endian, XXH_alignment align)
+
+{
+    const BYTE* p = (const BYTE*)ptr;
+
+#define PROCESS1               \
+    h32 += (*p++) * PRIME32_5; \
+    h32 = XXH_rotl32(h32, 11) * PRIME32_1 ;
+
+#define PROCESS4                         \
+    h32 += XXH_get32bits(p) * PRIME32_3; \
+    p+=4;                                \
+    h32  = XXH_rotl32(h32, 17) * PRIME32_4 ;
+
+    switch(len&15)  /* or switch(bEnd - p) */
+    {
+      case 12:      PROCESS4;
+                    /* fallthrough */
+      case 8:       PROCESS4;
+                    /* fallthrough */
+      case 4:       PROCESS4;
+                    return XXH32_avalanche(h32);
+
+      case 13:      PROCESS4;
+                    /* fallthrough */
+      case 9:       PROCESS4;
+                    /* fallthrough */
+      case 5:       PROCESS4;
+                    PROCESS1;
+                    return XXH32_avalanche(h32);
+
+      case 14:      PROCESS4;
+                    /* fallthrough */
+      case 10:      PROCESS4;
+                    /* fallthrough */
+      case 6:       PROCESS4;
+                    PROCESS1;
+                    PROCESS1;
+                    return XXH32_avalanche(h32);
+
+      case 15:      PROCESS4;
+                    /* fallthrough */
+      case 11:      PROCESS4;
+                    /* fallthrough */
+      case 7:       PROCESS4;
+                    /* fallthrough */
+      case 3:       PROCESS1;
+                    /* fallthrough */
+      case 2:       PROCESS1;
+                    /* fallthrough */
+      case 1:       PROCESS1;
+                    /* fallthrough */
+      case 0:       return XXH32_avalanche(h32);
+    }
+    assert(0);
+    return h32;   /* reaching this point is deemed impossible */
+}
+
+
+FORCE_INLINE U32
+XXH32_endian_align(const void* input, size_t len, U32 seed,
+                    XXH_endianess endian, XXH_alignment align)
 {
     const BYTE* p = (const BYTE*)input;
     const BYTE* bEnd = p + len;
     U32 h32;
-#define XXH_get32bits(p) XXH_readLE32_align(p, endian, align)
 
-#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
+#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
     if (p==NULL) {
         len=0;
         bEnd=p=(const BYTE*)(size_t)16;
@@ -281,7 +364,7 @@ XXH_FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed,
 #endif
 
     if (len>=16) {
-        const BYTE* const limit = bEnd - 16;
+        const BYTE* const limit = bEnd - 15;
         U32 v1 = seed + PRIME32_1 + PRIME32_2;
         U32 v2 = seed + PRIME32_2;
         U32 v3 = seed + 0;
@@ -292,34 +375,17 @@ XXH_FORCE_INLINE U32 XXH32_endian_align(const void* input, size_t len, U32 seed,
             v2 = XXH32_round(v2, XXH_get32bits(p)); p+=4;
             v3 = XXH32_round(v3, XXH_get32bits(p)); p+=4;
             v4 = XXH32_round(v4, XXH_get32bits(p)); p+=4;
-        } while (p<=limit);
+        } while (p < limit);
 
-        h32 = XXH_rotl32(v1, 1) + XXH_rotl32(v2, 7) + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
+        h32 = XXH_rotl32(v1, 1)  + XXH_rotl32(v2, 7)
+            + XXH_rotl32(v3, 12) + XXH_rotl32(v4, 18);
     } else {
         h32  = seed + PRIME32_5;
     }
 
-    h32 += (U32) len;
-
-    while (p+4<=bEnd) {
-        h32 += XXH_get32bits(p) * PRIME32_3;
-        h32  = XXH_rotl32(h32, 17) * PRIME32_4 ;
-        p+=4;
-    }
-
-    while (p<bEnd) {
-        h32 += (*p) * PRIME32_5;
-        h32 = XXH_rotl32(h32, 11) * PRIME32_1 ;
-        p++;
-    }
-
-    h32 ^= h32 >> 15;
-    h32 *= PRIME32_2;
-    h32 ^= h32 >> 13;
-    h32 *= PRIME32_3;
-    h32 ^= h32 >> 16;
+    h32 += (U32)len;
 
-    return h32;
+    return XXH32_finalize(h32, p, len&15, endian, align);
 }
 
 
@@ -371,74 +437,81 @@ XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t
 XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, unsigned int seed)
 {
     XXH32_state_t state;   /* using a local state to memcpy() in order to avoid strict-aliasing warnings */
-    memset(&state, 0, sizeof(state)-4);   /* do not write into reserved, for future removal */
+    memset(&state, 0, sizeof(state));
     state.v1 = seed + PRIME32_1 + PRIME32_2;
     state.v2 = seed + PRIME32_2;
     state.v3 = seed + 0;
     state.v4 = seed - PRIME32_1;
-    memcpy(statePtr, &state, sizeof(state));
+    /* do not write into reserved, planned to be removed in a future version */
+    memcpy(statePtr, &state, sizeof(state) - sizeof(state.reserved));
     return XXH_OK;
 }
 
 
-XXH_FORCE_INLINE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian)
+FORCE_INLINE XXH_errorcode
+XXH32_update_endian(XXH32_state_t* state, const void* input, size_t len, XXH_endianess endian)
 {
-    const BYTE* p = (const BYTE*)input;
-    const BYTE* const bEnd = p + len;
-
-#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
-    if (input==NULL) return XXH_ERROR;
+    if (input==NULL)
+#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
+        return XXH_OK;
+#else
+        return XXH_ERROR;
 #endif
 
-    state->total_len_32 += (unsigned)len;
-    state->large_len |= (len>=16) | (state->total_len_32>=16);
+    {   const BYTE* p = (const BYTE*)input;
+        const BYTE* const bEnd = p + len;
 
-    if (state->memsize + len < 16)  {   /* fill in tmp buffer */
-        XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len);
-        state->memsize += (unsigned)len;
-        return XXH_OK;
-    }
+        state->total_len_32 += (unsigned)len;
+        state->large_len |= (len>=16) | (state->total_len_32>=16);
 
-    if (state->memsize) {   /* some data left from previous update */
-        XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, 16-state->memsize);
-        {   const U32* p32 = state->mem32;
-            state->v1 = XXH32_round(state->v1, XXH_readLE32(p32, endian)); p32++;
-            state->v2 = XXH32_round(state->v2, XXH_readLE32(p32, endian)); p32++;
-            state->v3 = XXH32_round(state->v3, XXH_readLE32(p32, endian)); p32++;
-            state->v4 = XXH32_round(state->v4, XXH_readLE32(p32, endian)); p32++;
+        if (state->memsize + len < 16)  {   /* fill in tmp buffer */
+            XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len);
+            state->memsize += (unsigned)len;
+            return XXH_OK;
         }
-        p += 16-state->memsize;
-        state->memsize = 0;
-    }
-
-    if (p <= bEnd-16) {
-        const BYTE* const limit = bEnd - 16;
-        U32 v1 = state->v1;
-        U32 v2 = state->v2;
-        U32 v3 = state->v3;
-        U32 v4 = state->v4;
 
-        do {
-            v1 = XXH32_round(v1, XXH_readLE32(p, endian)); p+=4;
-            v2 = XXH32_round(v2, XXH_readLE32(p, endian)); p+=4;
-            v3 = XXH32_round(v3, XXH_readLE32(p, endian)); p+=4;
-            v4 = XXH32_round(v4, XXH_readLE32(p, endian)); p+=4;
-        } while (p<=limit);
+        if (state->memsize) {   /* some data left from previous update */
+            XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, 16-state->memsize);
+            {   const U32* p32 = state->mem32;
+                state->v1 = XXH32_round(state->v1, XXH_readLE32(p32, endian)); p32++;
+                state->v2 = XXH32_round(state->v2, XXH_readLE32(p32, endian)); p32++;
+                state->v3 = XXH32_round(state->v3, XXH_readLE32(p32, endian)); p32++;
+                state->v4 = XXH32_round(state->v4, XXH_readLE32(p32, endian));
+            }
+            p += 16-state->memsize;
+            state->memsize = 0;
+        }
 
-        state->v1 = v1;
-        state->v2 = v2;
-        state->v3 = v3;
-        state->v4 = v4;
-    }
+        if (p <= bEnd-16) {
+            const BYTE* const limit = bEnd - 16;
+            U32 v1 = state->v1;
+            U32 v2 = state->v2;
+            U32 v3 = state->v3;
+            U32 v4 = state->v4;
+
+            do {
+                v1 = XXH32_round(v1, XXH_readLE32(p, endian)); p+=4;
+                v2 = XXH32_round(v2, XXH_readLE32(p, endian)); p+=4;
+                v3 = XXH32_round(v3, XXH_readLE32(p, endian)); p+=4;
+                v4 = XXH32_round(v4, XXH_readLE32(p, endian)); p+=4;
+            } while (p<=limit);
+
+            state->v1 = v1;
+            state->v2 = v2;
+            state->v3 = v3;
+            state->v4 = v4;
+        }
 
-    if (p < bEnd) {
-        XXH_memcpy(state->mem32, p, (size_t)(bEnd-p));
-        state->memsize = (unsigned)(bEnd-p);
+        if (p < bEnd) {
+            XXH_memcpy(state->mem32, p, (size_t)(bEnd-p));
+            state->memsize = (unsigned)(bEnd-p);
+        }
     }
 
     return XXH_OK;
 }
 
+
 XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void* input, size_t len)
 {
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
@@ -450,40 +523,23 @@ XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* state_in, const void*
 }
 
 
-
-XXH_FORCE_INLINE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian)
+FORCE_INLINE U32
+XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess endian)
 {
-    const BYTE * p = (const BYTE*)state->mem32;
-    const BYTE* const bEnd = (const BYTE*)(state->mem32) + state->memsize;
     U32 h32;
 
     if (state->large_len) {
-        h32 = XXH_rotl32(state->v1, 1) + XXH_rotl32(state->v2, 7) + XXH_rotl32(state->v3, 12) + XXH_rotl32(state->v4, 18);
+        h32 = XXH_rotl32(state->v1, 1)
+            + XXH_rotl32(state->v2, 7)
+            + XXH_rotl32(state->v3, 12)
+            + XXH_rotl32(state->v4, 18);
     } else {
         h32 = state->v3 /* == seed */ + PRIME32_5;
     }
 
     h32 += state->total_len_32;
 
-    while (p+4<=bEnd) {
-        h32 += XXH_readLE32(p, endian) * PRIME32_3;
-        h32  = XXH_rotl32(h32, 17) * PRIME32_4;
-        p+=4;
-    }
-
-    while (p<bEnd) {
-        h32 += (*p) * PRIME32_5;
-        h32  = XXH_rotl32(h32, 11) * PRIME32_1;
-        p++;
-    }
-
-    h32 ^= h32 >> 15;
-    h32 *= PRIME32_2;
-    h32 ^= h32 >> 13;
-    h32 *= PRIME32_3;
-    h32 ^= h32 >> 16;
-
-    return h32;
+    return XXH32_finalize(h32, state->mem32, state->memsize, endian, XXH_aligned);
 }
 
 
@@ -503,7 +559,7 @@ XXH_PUBLIC_API unsigned int XXH32_digest (const XXH32_state_t* state_in)
 /*! Default XXH result types are basic unsigned 32 and 64 bits.
 *   The canonical representation follows human-readable write convention, aka big-endian (large digits first).
 *   These functions allow transformation of hash result into and from its canonical format.
-*   This way, hash values can be written into a file or buffer, and remain comparable across different systems and programs.
+*   This way, hash values can be written into a file or buffer, remaining comparable across different systems.
 */
 
 XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash)
@@ -522,18 +578,21 @@ XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src
 #ifndef XXH_NO_LONG_LONG
 
 /* *******************************************************************
-*  64-bits hash functions
+*  64-bit hash functions
 *********************************************************************/
 
 /*======   Memory access   ======*/
 
 #ifndef MEM_MODULE
 # define MEM_MODULE
-# if !defined (__VMS) && (defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
+# if !defined (__VMS) \
+  && (defined (__cplusplus) \
+  || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
 #   include <stdint.h>
     typedef uint64_t U64;
 # else
-    typedef unsigned long long U64;   /* if your compiler doesn't support unsigned long long, replace by another 64-bit type here. Note that xxhash.h will also need to be updated. */
+    /* if compiler doesn't support unsigned long long, replace by another 64-bit type */
+    typedef unsigned long long U64;
 # endif
 #endif
 
@@ -583,7 +642,7 @@ static U64 XXH_swap64 (U64 x)
 }
 #endif
 
-XXH_FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
+FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, XXH_alignment align)
 {
     if (align==XXH_unaligned)
         return endian==XXH_littleEndian ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr));
@@ -591,7 +650,7 @@ XXH_FORCE_INLINE U64 XXH_readLE64_align(const void* ptr, XXH_endianess endian, X
         return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr);
 }
 
-XXH_FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)
+FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)
 {
     return XXH_readLE64_align(ptr, endian, XXH_unaligned);
 }
@@ -626,14 +685,137 @@ static U64 XXH64_mergeRound(U64 acc, U64 val)
     return acc;
 }
 
-XXH_FORCE_INLINE U64 XXH64_endian_align(const void* input, size_t len, U64 seed, XXH_endianess endian, XXH_alignment align)
+static U64 XXH64_avalanche(U64 h64)
+{
+    h64 ^= h64 >> 33;
+    h64 *= PRIME64_2;
+    h64 ^= h64 >> 29;
+    h64 *= PRIME64_3;
+    h64 ^= h64 >> 32;
+    return h64;
+}
+
+
+#define XXH_get64bits(p) XXH_readLE64_align(p, endian, align)
+
+static U64
+XXH64_finalize(U64 h64, const void* ptr, size_t len,
+               XXH_endianess endian, XXH_alignment align)
+{
+    const BYTE* p = (const BYTE*)ptr;
+
+#define PROCESS1_64            \
+    h64 ^= (*p++) * PRIME64_5; \
+    h64 = XXH_rotl64(h64, 11) * PRIME64_1;
+
+#define PROCESS4_64          \
+    h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1; \
+    p+=4;                    \
+    h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
+
+#define PROCESS8_64 {        \
+    U64 const k1 = XXH64_round(0, XXH_get64bits(p)); \
+    p+=8;                    \
+    h64 ^= k1;               \
+    h64  = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4; \
+}
+
+    switch(len&31) {
+      case 24: PROCESS8_64;
+                    /* fallthrough */
+      case 16: PROCESS8_64;
+                    /* fallthrough */
+      case  8: PROCESS8_64;
+               return XXH64_avalanche(h64);
+
+      case 28: PROCESS8_64;
+                    /* fallthrough */
+      case 20: PROCESS8_64;
+                    /* fallthrough */
+      case 12: PROCESS8_64;
+                    /* fallthrough */
+      case  4: PROCESS4_64;
+               return XXH64_avalanche(h64);
+
+      case 25: PROCESS8_64;
+                    /* fallthrough */
+      case 17: PROCESS8_64;
+                    /* fallthrough */
+      case  9: PROCESS8_64;
+               PROCESS1_64;
+               return XXH64_avalanche(h64);
+
+      case 29: PROCESS8_64;
+                    /* fallthrough */
+      case 21: PROCESS8_64;
+                    /* fallthrough */
+      case 13: PROCESS8_64;
+                    /* fallthrough */
+      case  5: PROCESS4_64;
+               PROCESS1_64;
+               return XXH64_avalanche(h64);
+
+      case 26: PROCESS8_64;
+                    /* fallthrough */
+      case 18: PROCESS8_64;
+                    /* fallthrough */
+      case 10: PROCESS8_64;
+               PROCESS1_64;
+               PROCESS1_64;
+               return XXH64_avalanche(h64);
+
+      case 30: PROCESS8_64;
+                    /* fallthrough */
+      case 22: PROCESS8_64;
+                    /* fallthrough */
+      case 14: PROCESS8_64;
+                    /* fallthrough */
+      case  6: PROCESS4_64;
+               PROCESS1_64;
+               PROCESS1_64;
+               return XXH64_avalanche(h64);
+
+      case 27: PROCESS8_64;
+                    /* fallthrough */
+      case 19: PROCESS8_64;
+                    /* fallthrough */
+      case 11: PROCESS8_64;
+               PROCESS1_64;
+               PROCESS1_64;
+               PROCESS1_64;
+               return XXH64_avalanche(h64);
+
+      case 31: PROCESS8_64;
+                    /* fallthrough */
+      case 23: PROCESS8_64;
+                    /* fallthrough */
+      case 15: PROCESS8_64;
+                    /* fallthrough */
+      case  7: PROCESS4_64;
+                    /* fallthrough */
+      case  3: PROCESS1_64;
+                    /* fallthrough */
+      case  2: PROCESS1_64;
+                    /* fallthrough */
+      case  1: PROCESS1_64;
+                    /* fallthrough */
+      case  0: return XXH64_avalanche(h64);
+    }
+
+    /* impossible to reach */
+    assert(0);
+    return 0;  /* unreachable, but some compilers complain without it */
+}
+
+FORCE_INLINE U64
+XXH64_endian_align(const void* input, size_t len, U64 seed,
+                XXH_endianess endian, XXH_alignment align)
 {
     const BYTE* p = (const BYTE*)input;
-    const BYTE* const bEnd = p + len;
+    const BYTE* bEnd = p + len;
     U64 h64;
-#define XXH_get64bits(p) XXH_readLE64_align(p, endian, align)
 
-#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
+#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
     if (p==NULL) {
         len=0;
         bEnd=p=(const BYTE*)(size_t)32;
@@ -666,32 +848,7 @@ XXH_FORCE_INLINE U64 XXH64_endian_align(const void* input, size_t len, U64 seed,
 
     h64 += (U64) len;
 
-    while (p+8<=bEnd) {
-        U64 const k1 = XXH64_round(0, XXH_get64bits(p));
-        h64 ^= k1;
-        h64  = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
-        p+=8;
-    }
-
-    if (p+4<=bEnd) {
-        h64 ^= (U64)(XXH_get32bits(p)) * PRIME64_1;
-        h64 = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
-        p+=4;
-    }
-
-    while (p<bEnd) {
-        h64 ^= (*p) * PRIME64_5;
-        h64 = XXH_rotl64(h64, 11) * PRIME64_1;
-        p++;
-    }
-
-    h64 ^= h64 >> 33;
-    h64 *= PRIME64_2;
-    h64 ^= h64 >> 29;
-    h64 *= PRIME64_3;
-    h64 ^= h64 >> 32;
-
-    return h64;
+    return XXH64_finalize(h64, p, len, endian, align);
 }
 
 
@@ -741,65 +898,71 @@ XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* dstState, const XXH64_state_t
 XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH64_state_t* statePtr, unsigned long long seed)
 {
     XXH64_state_t state;   /* using a local state to memcpy() in order to avoid strict-aliasing warnings */
-    memset(&state, 0, sizeof(state)-8);   /* do not write into reserved, for future removal */
+    memset(&state, 0, sizeof(state));
     state.v1 = seed + PRIME64_1 + PRIME64_2;
     state.v2 = seed + PRIME64_2;
     state.v3 = seed + 0;
     state.v4 = seed - PRIME64_1;
-    memcpy(statePtr, &state, sizeof(state));
+     /* do not write into reserved, planned to be removed in a future version */
+    memcpy(statePtr, &state, sizeof(state) - sizeof(state.reserved));
     return XXH_OK;
 }
 
-XXH_FORCE_INLINE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian)
+FORCE_INLINE XXH_errorcode
+XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian)
 {
-    const BYTE* p = (const BYTE*)input;
-    const BYTE* const bEnd = p + len;
-
-#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
-    if (input==NULL) return XXH_ERROR;
+    if (input==NULL)
+#if defined(XXH_ACCEPT_NULL_INPUT_POINTER) && (XXH_ACCEPT_NULL_INPUT_POINTER>=1)
+        return XXH_OK;
+#else
+        return XXH_ERROR;
 #endif
 
-    state->total_len += len;
-
-    if (state->memsize + len < 32) {  /* fill in tmp buffer */
-        XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, len);
-        state->memsize += (U32)len;
-        return XXH_OK;
-    }
+    {   const BYTE* p = (const BYTE*)input;
+        const BYTE* const bEnd = p + len;
 
-    if (state->memsize) {   /* tmp buffer is full */
-        XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, 32-state->memsize);
-        state->v1 = XXH64_round(state->v1, XXH_readLE64(state->mem64+0, endian));
-        state->v2 = XXH64_round(state->v2, XXH_readLE64(state->mem64+1, endian));
-        state->v3 = XXH64_round(state->v3, XXH_readLE64(state->mem64+2, endian));
-        state->v4 = XXH64_round(state->v4, XXH_readLE64(state->mem64+3, endian));
-        p += 32-state->memsize;
-        state->memsize = 0;
-    }
+        state->total_len += len;
 
-    if (p+32 <= bEnd) {
-        const BYTE* const limit = bEnd - 32;
-        U64 v1 = state->v1;
-        U64 v2 = state->v2;
-        U64 v3 = state->v3;
-        U64 v4 = state->v4;
+        if (state->memsize + len < 32) {  /* fill in tmp buffer */
+            XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, len);
+            state->memsize += (U32)len;
+            return XXH_OK;
+        }
 
-        do {
-            v1 = XXH64_round(v1, XXH_readLE64(p, endian)); p+=8;
-            v2 = XXH64_round(v2, XXH_readLE64(p, endian)); p+=8;
-            v3 = XXH64_round(v3, XXH_readLE64(p, endian)); p+=8;
-            v4 = XXH64_round(v4, XXH_readLE64(p, endian)); p+=8;
-        } while (p<=limit);
+        if (state->memsize) {   /* tmp buffer is full */
+            XXH_memcpy(((BYTE*)state->mem64) + state->memsize, input, 32-state->memsize);
+            state->v1 = XXH64_round(state->v1, XXH_readLE64(state->mem64+0, endian));
+            state->v2 = XXH64_round(state->v2, XXH_readLE64(state->mem64+1, endian));
+            state->v3 = XXH64_round(state->v3, XXH_readLE64(state->mem64+2, endian));
+            state->v4 = XXH64_round(state->v4, XXH_readLE64(state->mem64+3, endian));
+            p += 32-state->memsize;
+            state->memsize = 0;
+        }
 
-        state->v1 = v1;
-        state->v2 = v2;
-        state->v3 = v3;
-        state->v4 = v4;
-    }
+        if (p+32 <= bEnd) {
+            const BYTE* const limit = bEnd - 32;
+            U64 v1 = state->v1;
+            U64 v2 = state->v2;
+            U64 v3 = state->v3;
+            U64 v4 = state->v4;
+
+            do {
+                v1 = XXH64_round(v1, XXH_readLE64(p, endian)); p+=8;
+                v2 = XXH64_round(v2, XXH_readLE64(p, endian)); p+=8;
+                v3 = XXH64_round(v3, XXH_readLE64(p, endian)); p+=8;
+                v4 = XXH64_round(v4, XXH_readLE64(p, endian)); p+=8;
+            } while (p<=limit);
+
+            state->v1 = v1;
+            state->v2 = v2;
+            state->v3 = v3;
+            state->v4 = v4;
+        }
 
-    if (p < bEnd) {
-        XXH_memcpy(state->mem64, p, (size_t)(bEnd-p));
-        state->memsize = (unsigned)(bEnd-p);
+        if (p < bEnd) {
+            XXH_memcpy(state->mem64, p, (size_t)(bEnd-p));
+            state->memsize = (unsigned)(bEnd-p);
+        }
     }
 
     return XXH_OK;
@@ -815,10 +978,8 @@ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void*
         return XXH64_update_endian(state_in, input, len, XXH_bigEndian);
 }
 
-XXH_FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian)
+FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian)
 {
-    const BYTE * p = (const BYTE*)state->mem64;
-    const BYTE* const bEnd = (const BYTE*)state->mem64 + state->memsize;
     U64 h64;
 
     if (state->total_len >= 32) {
@@ -833,37 +994,12 @@ XXH_FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endian
         h64 = XXH64_mergeRound(h64, v3);
         h64 = XXH64_mergeRound(h64, v4);
     } else {
-        h64  = state->v3 + PRIME64_5;
+        h64  = state->v3 /*seed*/ + PRIME64_5;
     }
 
     h64 += (U64) state->total_len;
 
-    while (p+8<=bEnd) {
-        U64 const k1 = XXH64_round(0, XXH_readLE64(p, endian));
-        h64 ^= k1;
-        h64  = XXH_rotl64(h64,27) * PRIME64_1 + PRIME64_4;
-        p+=8;
-    }
-
-    if (p+4<=bEnd) {
-        h64 ^= (U64)(XXH_readLE32(p, endian)) * PRIME64_1;
-        h64  = XXH_rotl64(h64, 23) * PRIME64_2 + PRIME64_3;
-        p+=4;
-    }
-
-    while (p<bEnd) {
-        h64 ^= (*p) * PRIME64_5;
-        h64  = XXH_rotl64(h64, 11) * PRIME64_1;
-        p++;
-    }
-
-    h64 ^= h64 >> 33;
-    h64 *= PRIME64_2;
-    h64 ^= h64 >> 29;
-    h64 *= PRIME64_3;
-    h64 ^= h64 >> 32;
-
-    return h64;
+    return XXH64_finalize(h64, state->mem64, (size_t)state->total_len, endian, XXH_aligned);
 }
 
 XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in)
diff --git a/lib/xxhash.h b/lib/xxhash.h
index 870a6d9..d6bad94 100644
--- a/lib/xxhash.h
+++ b/lib/xxhash.h
@@ -57,8 +57,8 @@ Q.Score is a measure of quality of the hash function.
 It depends on successfully passing SMHasher test set.
 10 is a perfect score.
 
-A 64-bits version, named XXH64, is available since r35.
-It offers much better speed, but for 64-bits applications only.
+A 64-bit version, named XXH64, is available since r35.
+It offers much better speed, but for 64-bit applications only.
 Name     Speed on 64 bits    Speed on 32 bits
 XXH64       13.8 GB/s            1.9 GB/s
 XXH32        6.8 GB/s            6.0 GB/s
@@ -80,18 +80,19 @@ typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
 
 
 /* ****************************
-*  API modifier
-******************************/
-/** XXH_PRIVATE_API
-*   This is useful to include xxhash functions in `static` mode
-*   in order to inline them, and remove their symbol from the public list.
-*   Methodology :
-*     #define XXH_PRIVATE_API
-*     #include "xxhash.h"
-*   `xxhash.c` is automatically included.
-*   It's not useful to compile and link it as a separate module.
-*/
-#ifdef XXH_PRIVATE_API
+ *  API modifier
+ ******************************/
+/** XXH_INLINE_ALL (and XXH_PRIVATE_API)
+ *  This is useful to include xxhash functions in `static` mode
+ *  in order to inline them, and remove their symbol from the public list.
+ *  Inlining can offer dramatic performance improvement on small keys.
+ *  Methodology :
+ *     #define XXH_INLINE_ALL
+ *     #include "xxhash.h"
+ * `xxhash.c` is automatically included.
+ *  It's not useful to compile and link it as a separate module.
+ */
+#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)
 #  ifndef XXH_STATIC_LINKING_ONLY
 #    define XXH_STATIC_LINKING_ONLY
 #  endif
@@ -102,23 +103,24 @@ typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
 #  elif defined(_MSC_VER)
 #    define XXH_PUBLIC_API static __inline
 #  else
-#    define XXH_PUBLIC_API static   /* this version may generate warnings for unused static functions; disable the relevant warning */
+     /* this version may generate warnings for unused static functions */
+#    define XXH_PUBLIC_API static
 #  endif
 #else
 #  define XXH_PUBLIC_API   /* do nothing */
-#endif /* XXH_PRIVATE_API */
-
-/*!XXH_NAMESPACE, aka Namespace Emulation :
-
-If you want to include _and expose_ xxHash functions from within your own library,
-but also want to avoid symbol collisions with other libraries which may also include xxHash,
-
-you can use XXH_NAMESPACE, to automatically prefix any public symbol from xxhash library
-with the value of XXH_NAMESPACE (therefore, avoid NULL and numeric values).
-
-Note that no change is required within the calling program as long as it includes `xxhash.h` :
-regular symbol name will be automatically translated by this header.
-*/
+#endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */
+
+/*! XXH_NAMESPACE, aka Namespace Emulation :
+ *
+ * If you want to include _and expose_ xxHash functions from within your own library,
+ * but also want to avoid symbol collisions with other libraries which may also include xxHash,
+ *
+ * you can use XXH_NAMESPACE, to automatically prefix any public symbol from xxhash library
+ * with the value of XXH_NAMESPACE (therefore, avoid NULL and numeric values).
+ *
+ * Note that no change is required within the calling program as long as it includes `xxhash.h` :
+ * regular symbol name will be automatically translated by this header.
+ */
 #ifdef XXH_NAMESPACE
 #  define XXH_CAT(A,B) A##B
 #  define XXH_NAME2(A,B) XXH_CAT(A,B)
@@ -149,18 +151,18 @@ regular symbol name will be automatically translated by this header.
 ***************************************/
 #define XXH_VERSION_MAJOR    0
 #define XXH_VERSION_MINOR    6
-#define XXH_VERSION_RELEASE  2
+#define XXH_VERSION_RELEASE  5
 #define XXH_VERSION_NUMBER  (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE)
 XXH_PUBLIC_API unsigned XXH_versionNumber (void);
 
 
 /*-**********************************************************************
-*  32-bits hash
+*  32-bit hash
 ************************************************************************/
-typedef unsigned int       XXH32_hash_t;
+typedef unsigned int XXH32_hash_t;
 
 /*! XXH32() :
-    Calculate the 32-bits hash of sequence "length" bytes stored at memory address "input".
+    Calculate the 32-bit hash of sequence "length" bytes stored at memory address "input".
     The memory between input & input+length must be valid (allocated and read-accessible).
     "seed" can be used to alter the result predictably.
     Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s */
@@ -177,26 +179,25 @@ XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void*
 XXH_PUBLIC_API XXH32_hash_t  XXH32_digest (const XXH32_state_t* statePtr);
 
 /*
-These functions generate the xxHash of an input provided in multiple segments.
-Note that, for small input, they are slower than single-call functions, due to state management.
-For small input, prefer `XXH32()` and `XXH64()` .
-
-XXH state must first be allocated, using XXH*_createState() .
-
-Start a new hash by initializing state with a seed, using XXH*_reset().
-
-Then, feed the hash state by calling XXH*_update() as many times as necessary.
-Obviously, input must be allocated and read accessible.
-The function returns an error code, with 0 meaning OK, and any other value meaning there is an error.
-
-Finally, a hash value can be produced anytime, by using XXH*_digest().
-This function returns the nn-bits hash as an int or long long.
-
-It's still possible to continue inserting input into the hash state after a digest,
-and generate some new hashes later on, by calling again XXH*_digest().
-
-When done, free XXH state space if it was allocated dynamically.
-*/
+ * Streaming functions generate the xxHash of an input provided in multiple segments.
+ * Note that, for small input, they are slower than single-call functions, due to state management.
+ * For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized.
+ *
+ * XXH state must first be allocated, using XXH*_createState() .
+ *
+ * Start a new hash by initializing state with a seed, using XXH*_reset().
+ *
+ * Then, feed the hash state by calling XXH*_update() as many times as necessary.
+ * The function returns an error code, with 0 meaning OK, and any other value meaning there is an error.
+ *
+ * Finally, a hash value can be produced anytime, by using XXH*_digest().
+ * This function returns the nn-bits hash as an int or long long.
+ *
+ * It's still possible to continue inserting input into the hash state after a digest,
+ * and generate some new hashes later on, by calling again XXH*_digest().
+ *
+ * When done, free XXH state space if it was allocated dynamically.
+ */
 
 /*======   Canonical representation   ======*/
 
@@ -205,22 +206,22 @@ XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t
 XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src);
 
 /* Default result type for XXH functions are primitive unsigned 32 and 64 bits.
-*  The canonical representation uses human-readable write convention, aka big-endian (large digits first).
-*  These functions allow transformation of hash result into and from its canonical format.
-*  This way, hash values can be written into a file / memory, and remain comparable on different systems and programs.
-*/
+ * The canonical representation uses human-readable write convention, aka big-endian (large digits first).
+ * These functions allow transformation of hash result into and from its canonical format.
+ * This way, hash values can be written into a file / memory, and remain comparable on different systems and programs.
+ */
 
 
 #ifndef XXH_NO_LONG_LONG
 /*-**********************************************************************
-*  64-bits hash
+*  64-bit hash
 ************************************************************************/
 typedef unsigned long long XXH64_hash_t;
 
 /*! XXH64() :
-    Calculate the 64-bits hash of sequence of length "len" stored at memory address "input".
+    Calculate the 64-bit hash of sequence of length "len" stored at memory address "input".
     "seed" can be used to alter the result predictably.
-    This function runs faster on 64-bits systems, but slower on 32-bits systems (see benchmark).
+    This function runs faster on 64-bit systems, but slower on 32-bit systems (see benchmark).
 */
 XXH_PUBLIC_API XXH64_hash_t XXH64 (const void* input, size_t length, unsigned long long seed);
 
@@ -241,48 +242,82 @@ XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src
 #endif  /* XXH_NO_LONG_LONG */
 
 
+
 #ifdef XXH_STATIC_LINKING_ONLY
 
 /* ================================================================================================
-   This section contains definitions which are not guaranteed to remain stable.
+   This section contains declarations which are not guaranteed to remain stable.
    They may change in future versions, becoming incompatible with a different version of the library.
-   They shall only be used with static linking.
-   Never use these definitions in association with dynamic linking !
+   These declarations should only be used with static linking.
+   Never use them in association with dynamic linking !
 =================================================================================================== */
 
-/* These definitions are only meant to allow allocation of XXH state
-   statically, on stack, or in a struct for example.
-   Do not use members directly. */
-
-   struct XXH32_state_s {
-       unsigned total_len_32;
-       unsigned large_len;
-       unsigned v1;
-       unsigned v2;
-       unsigned v3;
-       unsigned v4;
-       unsigned mem32[4];   /* buffer defined as U32 for alignment */
-       unsigned memsize;
-       unsigned reserved;   /* never read nor write, will be removed in a future version */
-   };   /* typedef'd to XXH32_state_t */
-
-#ifndef XXH_NO_LONG_LONG
-   struct XXH64_state_s {
-       unsigned long long total_len;
-       unsigned long long v1;
-       unsigned long long v2;
-       unsigned long long v3;
-       unsigned long long v4;
-       unsigned long long mem64[4];   /* buffer defined as U64 for alignment */
-       unsigned memsize;
-       unsigned reserved[2];          /* never read nor write, will be removed in a future version */
-   };   /* typedef'd to XXH64_state_t */
+/* These definitions are only present to allow
+ * static allocation of XXH state, on stack or in a struct for example.
+ * Never **ever** use members directly. */
+
+#if !defined (__VMS) \
+  && (defined (__cplusplus) \
+  || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) )
+#   include <stdint.h>
+
+struct XXH32_state_s {
+   uint32_t total_len_32;
+   uint32_t large_len;
+   uint32_t v1;
+   uint32_t v2;
+   uint32_t v3;
+   uint32_t v4;
+   uint32_t mem32[4];
+   uint32_t memsize;
+   uint32_t reserved;   /* never read nor write, might be removed in a future version */
+};   /* typedef'd to XXH32_state_t */
+
+struct XXH64_state_s {
+   uint64_t total_len;
+   uint64_t v1;
+   uint64_t v2;
+   uint64_t v3;
+   uint64_t v4;
+   uint64_t mem64[4];
+   uint32_t memsize;
+   uint32_t reserved[2];          /* never read nor write, might be removed in a future version */
+};   /* typedef'd to XXH64_state_t */
+
+# else
+
+struct XXH32_state_s {
+   unsigned total_len_32;
+   unsigned large_len;
+   unsigned v1;
+   unsigned v2;
+   unsigned v3;
+   unsigned v4;
+   unsigned mem32[4];
+   unsigned memsize;
+   unsigned reserved;   /* never read nor write, might be removed in a future version */
+};   /* typedef'd to XXH32_state_t */
+
+#   ifndef XXH_NO_LONG_LONG  /* remove 64-bit support */
+struct XXH64_state_s {
+   unsigned long long total_len;
+   unsigned long long v1;
+   unsigned long long v2;
+   unsigned long long v3;
+   unsigned long long v4;
+   unsigned long long mem64[4];
+   unsigned memsize;
+   unsigned reserved[2];     /* never read nor write, might be removed in a future version */
+};   /* typedef'd to XXH64_state_t */
+#    endif
+
+# endif
+
+
+#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)
+#  include "xxhash.c"   /* include xxhash function bodies as `static`, for inlining */
 #endif
 
-#  ifdef XXH_PRIVATE_API
-#    include "xxhash.c"   /* include xxhash function bodies as `static`, for inlining */
-#  endif
-
 #endif /* XXH_STATIC_LINKING_ONLY */