Merge remote-tracking branch 'refs/remotes/lz4/dev' into dev

17 files changed, 813 insertions, 382 deletions

diff --git a/NEWS b/NEWS
index 3257fd5..ebf8e1c 100644
--- a/NEWS
+++ b/NEWS
@@ -9,11 +9,14 @@ Improved cmake build script, by Evan Nemerson
 New liblz4-dll project, by Przemyslaw Skibinki
 Makefile: Generates object files (*.o) for faster (re)compilation on low power systems
 cli : new : --rm command
-cli : new : file attributes are preserved, by Przemyslaw Skibinki
+cli : new : preserved file attributes, by Przemyslaw Skibinki
 cli : fix : crash on some invalid inputs
 cli : fix : -t correctly validates lz4-compressed files, by Nick Terrell
 cli : fix : detects and reports fread() errors, thanks to Hiroshi Fujishima report #243
 cli : bench : new : -r recursive mode
+lz4cat : can cat multiple files in a single command line (#184)
+Added : doc/lz4_manual.html, by Przemyslaw Skibinski
+Added : dictionary compression example, by Nick Terrell
 Added : Debianization, by Evgeniy Polyakov
 
 r131
diff --git a/contrib/gen_manual/Makefile b/contrib/gen_manual/Makefile
index 49616ee..adbcca2 100644
--- a/contrib/gen_manual/Makefile
+++ b/contrib/gen_manual/Makefile
@@ -1,11 +1,33 @@
-# ##########################################################################
-# Copyright (c) 2016-present, Przemyslaw Skibinski
+# ################################################################
+# Copyright (C) Przemyslaw Skibinski 2016-present
 # All rights reserved.
 #
-# This source code is licensed under the BSD-style license found in the
-# LICENSE file in the root directory of this source tree. An additional grant
-# of patent rights can be found in the PATENTS file in the same directory.
-# ##########################################################################
+# BSD license
+# 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 HOLDER 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 source repository : https://github.com/Cyan4973/lz4
+#  - LZ4 forum froup : https://groups.google.com/forum/#!forum/lz4c
+# ################################################################
 
 
 CFLAGS ?= -O3
diff --git a/contrib/gen_manual/README.md b/contrib/gen_manual/README.md
index 3d146d4..7664ac6 100644
--- a/contrib/gen_manual/README.md
+++ b/contrib/gen_manual/README.md
@@ -1,5 +1,5 @@
-gen_manual - a program for automatic generation of zstd manual 
-============================================================
+gen_manual - a program for automatic generation of manual from source code
+==========================================================================
 
 #### Introduction
 
@@ -24,8 +24,8 @@ The program requires 3 parameters:
 gen_manual [lz4_version] [input_file] [output_html]
 ```
 
-To compile program and generate lz4 manual we have used: 
+To compile program and generate lz4 manual we have used:
 ```
 make
-./gen_manual.exe 1.7.3 ../../lib/lz4.h zstd_manual.html
+./gen_manual.exe 1.7.3 ../../lib/lz4.h lz4_manual.html
 ```
diff --git a/contrib/gen_manual/gen_manual.cpp b/contrib/gen_manual/gen_manual.cpp
index caf03fc..2df081d 100644
--- a/contrib/gen_manual/gen_manual.cpp
+++ b/contrib/gen_manual/gen_manual.cpp
@@ -1,11 +1,36 @@
 /*
- * Copyright (c) 2016-present, Przemyslaw Skibinski
- * All rights reserved.
- *
- * This source code is licensed under the BSD-style license found in the
- * LICENSE file in the root directory of this source tree. An additional grant
- * of patent rights can be found in the PATENTS file in the same directory.
- */
+Copyright (c) 2016-present, Przemyslaw Skibinski
+All rights reserved.
+
+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
+*/
 
 #include <iostream>
 #include <fstream>
@@ -213,4 +238,4 @@ int main(int argc, char *argv[]) {
     ostream << "</html>" << endl << "</body>" << endl;
 
     return 0;
-}
\ No newline at end of file
+}
diff --git a/examples/.gitignore b/examples/.gitignore
index 4893866..3ceb90d 100644
--- a/examples/.gitignore
+++ b/examples/.gitignore
@@ -1,6 +1,7 @@
 /Makefile.lz4*
 /printVersion
 /doubleBuffer
+/dictionaryRandomAccess
 /ringBuffer
 /ringBufferHC
 /lineCompress
diff --git a/examples/Makefile b/examples/Makefile
index c8caf24..aad713b 100644
--- a/examples/Makefile
+++ b/examples/Makefile
@@ -48,7 +48,7 @@ endif
 
 default: all
 
-all: printVersion doubleBuffer ringBuffer ringBufferHC lineCompress frameCompress
+all: printVersion doubleBuffer dictionaryRandomAccess ringBuffer ringBufferHC lineCompress frameCompress
 
 printVersion: $(LZ4DIR)/lz4.c printVersion.c
 	$(CC)      $(FLAGS) $^ -o $@$(EXT)
@@ -56,6 +56,9 @@ printVersion: $(LZ4DIR)/lz4.c printVersion.c
 doubleBuffer: $(LZ4DIR)/lz4.c blockStreaming_doubleBuffer.c
 	$(CC)      $(FLAGS) $^ -o $@$(EXT)
 
+dictionaryRandomAccess: $(LZ4DIR)/lz4.c dictionaryRandomAccess.c
+	$(CC)      $(FLAGS) $^ -o $@$(EXT)
+
 ringBuffer  : $(LZ4DIR)/lz4.c blockStreaming_ringBuffer.c
 	$(CC)      $(FLAGS) $^ -o $@$(EXT)
 
@@ -66,7 +69,7 @@ lineCompress: $(LZ4DIR)/lz4.c blockStreaming_lineByLine.c
 	$(CC)      $(FLAGS) $^ -o $@$(EXT)
 
 frameCompress: frameCompress.c
-	$(CC)      $(FLAGS) $^ -o $@$(EXT) -L$(LZ4DIR) -llz4
+	$(CC)      $(FLAGS) $^ -o $@$(EXT) $(LZ4DIR)/liblz4.a
 
 compressFunctions: $(LZ4DIR)/lz4.c compress_functions.c
 	$(CC)      $(FLAGS) $^ -o $@$(EXT) -lrt
@@ -77,15 +80,16 @@ simpleBuffer: $(LZ4DIR)/lz4.c simple_buffer.c
 test : all
 	./printVersion$(EXT)
 	./doubleBuffer$(EXT) $(TESTFILE)
+	./dictionaryRandomAccess$(EXT) $(TESTFILE) $(TESTFILE) 1100 1400
 	./ringBuffer$(EXT)   $(TESTFILE)
 	./ringBufferHC$(EXT) $(TESTFILE)
 	./lineCompress$(EXT) $(TESTFILE)
-	LD_LIBRARY_PATH=$(LZ4DIR) ./frameCompress$(EXT) $(TESTFILE)
+	./frameCompress$(EXT) $(TESTFILE)
 	$(LZ4) -vt $(TESTFILE).lz4
 
 clean:
 	@rm -f core *.o *.dec *-0 *-9 *-8192 *.lz4s *.lz4 \
-	 printVersion$(EXT) doubleBuffer$(EXT) ringBuffer$(EXT) ringBufferHC$(EXT) \
-	 lineCompress$(EXT) frameCompress$(EXT) compressFunctions$(EXT) simpleBuffer$(EXT)
+	 printVersion$(EXT) doubleBuffer$(EXT) dictionaryRandomAccess$(EXT) \
+	 ringBuffer$(EXT) ringBufferHC$(EXT) lineCompress$(EXT) frameCompress$(EXT) \
+	 compressFunctions$(EXT) simpleBuffer$(EXT)
 	@echo Cleaning completed
-
diff --git a/examples/README.md b/examples/README.md
index 74527d4..e6839e9 100644
--- a/examples/README.md
+++ b/examples/README.md
@@ -8,3 +8,4 @@ All examples are GPL-v2 licensed.
  - Examples
      - [Double Buffer](blockStreaming_doubleBuffer.md)
      - [Line by Line Text Compression](blockStreaming_lineByLine.md)
+     - [Dictionary Random Access](dictionaryRandomAccess.md)
diff --git a/examples/dictionaryRandomAccess.c b/examples/dictionaryRandomAccess.c
new file mode 100644
index 0000000..6acf99b
--- /dev/null
+++ b/examples/dictionaryRandomAccess.c
@@ -0,0 +1,280 @@
+// LZ4 API example : Dictionary Random Access
+
+#ifdef _MSC_VER    /* Visual Studio */
+#  define _CRT_SECURE_NO_WARNINGS
+#  define snprintf sprintf_s
+#endif
+#include "lz4.h"
+
+#include <stdio.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define MIN(x, y) (x) < (y) ? (x) : (y)
+
+enum {
+    BLOCK_BYTES = 1024,  /* 1 KiB of uncompressed data in a block */
+    DICTIONARY_BYTES = 1024, /* Load a 1 KiB dictionary */
+    MAX_BLOCKS = 1024 /* For simplicity of implementation */
+};
+
+/**
+ * Magic bytes for this test case.
+ * This is not a great magic number because it is a common word in ASCII.
+ * However, it is important to have some versioning system in your format.
+ */
+const char kTestMagic[] = { 'T', 'E', 'S', 'T' };
+
+
+void write_int(FILE* fp, int i) {
+    size_t written = fwrite(&i, sizeof(i), 1, fp);
+    if (written != 1) { exit(10); }
+}
+
+void write_bin(FILE* fp, const void* array, size_t arrayBytes) {
+    size_t written = fwrite(array, 1, arrayBytes, fp);
+    if (written != arrayBytes) { exit(11); }
+}
+
+void read_int(FILE* fp, int* i) {
+    size_t read = fread(i, sizeof(*i), 1, fp);
+    if (read != 1) { exit(12); }
+}
+
+size_t read_bin(FILE* fp, void* array, size_t arrayBytes) {
+    size_t read = fread(array, 1, arrayBytes, fp);
+    if (ferror(fp)) { exit(12); }
+    return read;
+}
+
+void seek_bin(FILE* fp, long offset, int origin) {
+    if (fseek(fp, offset, origin)) { exit(14); }
+}
+
+
+void test_compress(FILE* outFp, FILE* inpFp, void *dict, int dictSize)
+{
+    LZ4_stream_t lz4Stream_body;
+    LZ4_stream_t* lz4Stream = &lz4Stream_body;
+
+    char inpBuf[BLOCK_BYTES];
+    int offsets[MAX_BLOCKS];
+    int *offsetsEnd = offsets;
+
+
+    LZ4_resetStream(lz4Stream);
+
+    /* Write header magic */
+    write_bin(outFp, kTestMagic, sizeof(kTestMagic));
+
+    *offsetsEnd++ = sizeof(kTestMagic);
+    /* Write compressed data blocks.  Each block contains BLOCK_BYTES of plain
+       data except possibly the last. */
+    for(;;) {
+        const int inpBytes = (int) read_bin(inpFp, inpBuf, BLOCK_BYTES);
+        if(0 == inpBytes) {
+            break;
+        }
+
+        /* Forget previously compressed data and load the dictionary */
+        LZ4_loadDict(lz4Stream, dict, dictSize);
+        {
+            char cmpBuf[LZ4_COMPRESSBOUND(BLOCK_BYTES)];
+            const int cmpBytes = LZ4_compress_fast_continue(
+                lz4Stream, inpBuf, cmpBuf, inpBytes, sizeof(cmpBuf), 1);
+            if(cmpBytes <= 0) { exit(1); }
+            write_bin(outFp, cmpBuf, (size_t)cmpBytes);
+            /* Keep track of the offsets */
+            *offsetsEnd = *(offsetsEnd - 1) + cmpBytes;
+            ++offsetsEnd;
+        }
+        if (offsetsEnd - offsets > MAX_BLOCKS) { exit(2); }
+    }
+    /* Write the tailing jump table */
+    {
+        int *ptr = offsets;
+        while (ptr != offsetsEnd) {
+            write_int(outFp, *ptr++);
+        }
+        write_int(outFp, offsetsEnd - offsets);
+    }
+}
+
+
+void test_decompress(FILE* outFp, FILE* inpFp, void *dict, int dictSize, int offset, int length)
+{
+    LZ4_streamDecode_t lz4StreamDecode_body;
+    LZ4_streamDecode_t* lz4StreamDecode = &lz4StreamDecode_body;
+
+    /* The blocks [currentBlock, endBlock) contain the data we want */
+    int currentBlock = offset / BLOCK_BYTES;
+    int endBlock = ((offset + length - 1) / BLOCK_BYTES) + 1;
+
+    char decBuf[BLOCK_BYTES];
+    int offsets[MAX_BLOCKS];
+
+    /* Special cases */
+    if (length == 0) { return; }
+
+    /* Read the magic bytes */
+    {
+        char magic[sizeof(kTestMagic)];
+        size_t read = read_bin(inpFp, magic, sizeof(magic));
+        if (read != sizeof(magic)) { exit(1); }
+        if (memcmp(kTestMagic, magic, sizeof(magic))) { exit(2); }
+    }
+
+    /* Read the offsets tail */
+    {
+        int numOffsets;
+        int block;
+        int *offsetsPtr = offsets;
+        seek_bin(inpFp, -4, SEEK_END);
+        read_int(inpFp, &numOffsets);
+        if (numOffsets <= endBlock) { exit(3); }
+        seek_bin(inpFp, -4 * (numOffsets + 1), SEEK_END);
+        for (block = 0; block <= endBlock; ++block) {
+            read_int(inpFp, offsetsPtr++);
+        }
+    }
+    /* Seek to the first block to read */
+    seek_bin(inpFp, offsets[currentBlock], SEEK_SET);
+    offset = offset % BLOCK_BYTES;
+
+    /* Start decoding */
+    for(; currentBlock < endBlock; ++currentBlock) {
+        char cmpBuf[LZ4_COMPRESSBOUND(BLOCK_BYTES)];
+        /* The difference in offsets is the size of the block */
+        int  cmpBytes = offsets[currentBlock + 1] - offsets[currentBlock];
+        {
+            const size_t read = read_bin(inpFp, cmpBuf, (size_t)cmpBytes);
+            if(read != (size_t)cmpBytes) { exit(4); }
+        }
+
+        /* Load the dictionary */
+        LZ4_setStreamDecode(lz4StreamDecode, dict, dictSize);
+        {
+            const int decBytes = LZ4_decompress_safe_continue(
+                lz4StreamDecode, cmpBuf, decBuf, cmpBytes, BLOCK_BYTES);
+            if(decBytes <= 0) { exit(5); }
+            {
+                /* Write out the part of the data we care about */
+                int blockLength = MIN(length, (decBytes - offset));
+                write_bin(outFp, decBuf + offset, (size_t)blockLength);
+                offset = 0;
+                length -= blockLength;
+            }
+        }
+    }
+}
+
+
+int compare(FILE* fp0, FILE* fp1, int length)
+{
+    int result = 0;
+
+    while(0 == result) {
+        char b0[4096];
+        char b1[4096];
+        const size_t r0 = read_bin(fp0, b0, MIN(length, (int)sizeof(b0)));
+        const size_t r1 = read_bin(fp1, b1, MIN(length, (int)sizeof(b1)));
+
+        result = (int) r0 - (int) r1;
+
+        if(0 == r0 || 0 == r1) {
+            break;
+        }
+        if(0 == result) {
+            result = memcmp(b0, b1, r0);
+        }
+        length -= r0;
+    }
+
+    return result;
+}
+
+
+int main(int argc, char* argv[])
+{
+    char inpFilename[256] = { 0 };
+    char lz4Filename[256] = { 0 };
+    char decFilename[256] = { 0 };
+    char dictFilename[256] = { 0 };
+    int offset;
+    int length;
+    char dict[DICTIONARY_BYTES];
+    int dictSize;
+
+    if(argc < 5) {
+        printf("Usage: %s input dictionary offset length", argv[0]);
+        return 0;
+    }
+
+    snprintf(inpFilename, 256, "%s", argv[1]);
+    snprintf(lz4Filename, 256, "%s.lz4s-%d", argv[1], BLOCK_BYTES);
+    snprintf(decFilename, 256, "%s.lz4s-%d.dec", argv[1], BLOCK_BYTES);
+    snprintf(dictFilename, 256, "%s", argv[2]);
+    offset = atoi(argv[3]);
+    length = atoi(argv[4]);
+
+    printf("inp    = [%s]\n", inpFilename);
+    printf("lz4    = [%s]\n", lz4Filename);
+    printf("dec    = [%s]\n", decFilename);
+    printf("dict   = [%s]\n", dictFilename);
+    printf("offset = [%d]\n", offset);
+    printf("length = [%d]\n", length);
+
+    /* Load dictionary */
+    {
+        FILE* dictFp = fopen(dictFilename, "rb");
+        dictSize = (int)read_bin(dictFp, dict, DICTIONARY_BYTES);
+        fclose(dictFp);
+    }
+
+    /* compress */
+    {
+        FILE* inpFp = fopen(inpFilename, "rb");
+        FILE* outFp = fopen(lz4Filename, "wb");
+
+        printf("compress : %s -> %s\n", inpFilename, lz4Filename);
+        test_compress(outFp, inpFp, dict, dictSize);
+        printf("compress : done\n");
+
+        fclose(outFp);
+        fclose(inpFp);
+    }
+
+    /* decompress */
+    {
+        FILE* inpFp = fopen(lz4Filename, "rb");
+        FILE* outFp = fopen(decFilename, "wb");
+
+        printf("decompress : %s -> %s\n", lz4Filename, decFilename);
+        test_decompress(outFp, inpFp, dict, DICTIONARY_BYTES, offset, length);
+        printf("decompress : done\n");
+
+        fclose(outFp);
+        fclose(inpFp);
+    }
+
+    /* verify */
+    {
+        FILE* inpFp = fopen(inpFilename, "rb");
+        FILE* decFp = fopen(decFilename, "rb");
+        seek_bin(inpFp, offset, SEEK_SET);
+
+        printf("verify : %s <-> %s\n", inpFilename, decFilename);
+        const int cmp = compare(inpFp, decFp, length);
+        if(0 == cmp) {
+            printf("verify : OK\n");
+        } else {
+            printf("verify : NG\n");
+        }
+
+        fclose(decFp);
+        fclose(inpFp);
+    }
+
+    return 0;
+}
diff --git a/examples/dictionaryRandomAccess.md b/examples/dictionaryRandomAccess.md
new file mode 100644
index 0000000..53d825d
--- /dev/null
+++ b/examples/dictionaryRandomAccess.md
@@ -0,0 +1,67 @@
+# LZ4 API Example : Dictionary Random Access
+
+`dictionaryRandomAccess.c` is LZ4 API example which implements dictionary compression and random access decompression.
+
+Please note that the output file is not compatible with lz4frame and is platform dependent.
+
+
+## What's the point of this example ?
+
+ - Dictionary based compression for homogenous files.
+ - Random access to compressed blocks.
+
+
+## How the compression works
+
+Reads the dictionary from a file, and uses it as the history for each block.
+This allows each block to be independent, but maintains compression ratio.
+
+```
+    Dictionary
+         +
+         |
+         v
+    +---------+
+    | Block#1 |
+    +----+----+
+         |
+         v
+      {Out#1}
+
+
+    Dictionary
+         +
+         |
+         v
+    +---------+
+    | Block#2 |
+    +----+----+
+         |
+         v
+      {Out#2}
+```
+
+After writing the magic bytes `TEST` and then the compressed blocks, write out the jump table.
+The last 4 bytes is an integer containing the number of blocks in the stream.
+If there are `N` blocks, then just before the last 4 bytes is `N + 1` 4 byte integers containing the offsets at the beginning and end of each block.
+Let `Offset#K` be the total number of bytes written after writing out `Block#K` *including* the magic bytes for simplicity.
+
+```
++------+---------+     +---------+---+----------+     +----------+-----+
+| TEST | Block#1 | ... | Block#N | 4 | Offset#1 | ... | Offset#N | N+1 |
++------+---------+     +---------+---+----------+     +----------+-----+
+```
+
+## How the decompression works
+
+Decompression will do reverse order.
+
+ - Seek to the last 4 bytes of the file and read the number of offsets.
+ - Read each offset into an array.
+ - Seek to the first block containing data we want to read.
+   We know where to look because we know each block contains a fixed amount of uncompressed data, except possibly the last.
+ - Decompress it and write what data we need from it to the file.
+ - Read the next block.
+ - Decompress it and write that page to the file.
+
+Continue these procedure until all the required data has been read.
diff --git a/lib/README.md b/lib/README.md
index 051ef54..3ca2776 100644
--- a/lib/README.md
+++ b/lib/README.md
@@ -10,7 +10,7 @@ The minimum required is **`lz4.c`** and **`lz4.h`**,
 which will provide the fast compression and decompression algorithm.
 
 
-#### The High Compression variant od LZ4
+#### The High Compression variant of LZ4
 
 For more compression at the cost of compression speed,
 the High Compression variant **lz4hc** is available.
diff --git a/lib/lz4.c b/lib/lz4.c
index 53b503c..f86e6f8 100644
--- a/lib/lz4.c
+++ b/lib/lz4.c
@@ -1066,10 +1066,17 @@ int LZ4_compress_forceExtDict (LZ4_stream_t* LZ4_dict, const char* source, char*
 }
 
 
+/*! LZ4_saveDict() :
+ *  If previously compressed data block is not guaranteed to remain available at its memory location,
+ *  save it into a safer place (char* safeBuffer).
+ *  Note : you don't need to call LZ4_loadDict() afterwards,
+ *         dictionary is immediately usable, you can therefore call LZ4_compress_fast_continue().
+ *  Return : saved dictionary size in bytes (necessarily <= dictSize), or 0 if error.
+ */
 int LZ4_saveDict (LZ4_stream_t* LZ4_dict, char* safeBuffer, int dictSize)
 {
-    LZ4_stream_t_internal* dict = (LZ4_stream_t_internal*) LZ4_dict;
-    const BYTE* previousDictEnd = dict->dictionary + dict->dictSize;
+    LZ4_stream_t_internal* const dict = (LZ4_stream_t_internal*) LZ4_dict;
+    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;
diff --git a/lib/xxhash.c b/lib/xxhash.c
index 156b484..e9ff2d4 100644
--- a/lib/xxhash.c
+++ b/lib/xxhash.c
@@ -46,7 +46,7 @@
  * Method 2 : direct access. This method doesn't depend on compiler but violate C standard.
  *            It can generate buggy code on targets which do not support unaligned memory accesses.
  *            But in some circumstances, it's the only known way to get the most performance (ie GCC + ARMv6)
- * See https://stackoverflow.com/a/32095106/646947 for details.
+ * See http://stackoverflow.com/a/32095106/646947 for details.
  * Prefer these methods in priority order (0 > 1 > 2)
  */
 #ifndef XXH_FORCE_MEMORY_ACCESS   /* can be defined externally, on command line for example */
@@ -66,10 +66,10 @@
 /* #define XXH_ACCEPT_NULL_INPUT_POINTER 1 */
 
 /*!XXH_FORCE_NATIVE_FORMAT :
- * By default, xxHash library provides endian-independant Hash values, based on little-endian convention.
+ * By default, xxHash library provides endian-independent Hash values, based on little-endian convention.
  * Results are therefore identical for little-endian and big-endian CPU.
  * This comes at a performance cost for big-endian CPU, since some swapping is required to emulate little-endian format.
- * Should endian-independance be of no importance for your application, you may set the #define below to 1,
+ * Should endian-independence be of no importance for your application, you may set the #define below to 1,
  * to improve speed for Big-endian CPU.
  * This option has no impact on Little_Endian CPU.
  */
@@ -95,12 +95,12 @@
 /* *************************************
 *  Includes & Memory related functions
 ***************************************/
-/* Modify the local functions below should you wish to use some other memory routines */
-/* for malloc(), free() */
+/*! Modify the local functions below should you wish to use some other memory routines
+*   for malloc(), free() */
 #include <stdlib.h>
 static void* XXH_malloc(size_t s) { return malloc(s); }
 static void  XXH_free  (void* p)  { free(p); }
-/* for memcpy() */
+/*! and for memcpy() */
 #include <string.h>
 static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); }
 
@@ -115,7 +115,7 @@ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcp
 #  pragma warning(disable : 4127)      /* disable: C4127: conditional expression is constant */
 #  define FORCE_INLINE static __forceinline
 #else
-#  if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
+#  if defined (__cplusplus) || defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L   /* C99 */
 #    ifdef __GNUC__
 #      define FORCE_INLINE static inline __attribute__((always_inline))
 #    else
@@ -131,45 +131,37 @@ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcp
 *  Basic Types
 ***************************************/
 #ifndef MEM_MODULE
-# define MEM_MODULE
-# if 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;
-    typedef uint64_t U64;
-#  else
+# else
     typedef unsigned char      BYTE;
     typedef unsigned short     U16;
     typedef unsigned int       U32;
     typedef   signed int       S32;
-    typedef unsigned long long U64;
-#  endif
+# endif
 #endif
 
-
 #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2))
 
 /* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */
 static U32 XXH_read32(const void* memPtr) { return *(const U32*) memPtr; }
-static U64 XXH_read64(const void* memPtr) { return *(const U64*) memPtr; }
 
 #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1))
 
 /* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
 /* currently only defined for gcc and icc */
-typedef union { U32 u32; U64 u64; } __attribute__((packed)) unalign;
-
+typedef union { U32 u32; } __attribute__((packed)) unalign;
 static U32 XXH_read32(const void* ptr) { return ((const unalign*)ptr)->u32; }
-static U64 XXH_read64(const void* ptr) { return ((const unalign*)ptr)->u64; }
 
 #else
 
 /* portable and safe solution. Generally efficient.
- * see : https://stackoverflow.com/a/32095106/646947
+ * see : http://stackoverflow.com/a/32095106/646947
  */
-
 static U32 XXH_read32(const void* memPtr)
 {
     U32 val;
@@ -177,20 +169,13 @@ static U32 XXH_read32(const void* memPtr)
     return val;
 }
 
-static U64 XXH_read64(const void* memPtr)
-{
-    U64 val;
-    memcpy(&val, memPtr, sizeof(val));
-    return val;
-}
-
 #endif   /* XXH_FORCE_DIRECT_MEMORY_ACCESS */
 
 
 /* ****************************************
 *  Compiler-specific Functions and Macros
 ******************************************/
-#define GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
+#define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
 
 /* Note : although _rotl exists for minGW (GCC under windows), performance seems poor */
 #if defined(_MSC_VER)
@@ -203,10 +188,8 @@ static U64 XXH_read64(const void* memPtr)
 
 #if defined(_MSC_VER)     /* Visual Studio */
 #  define XXH_swap32 _byteswap_ulong
-#  define XXH_swap64 _byteswap_uint64
-#elif GCC_VERSION >= 403
+#elif XXH_GCC_VERSION >= 403
 #  define XXH_swap32 __builtin_bswap32
-#  define XXH_swap64 __builtin_bswap64
 #else
 static U32 XXH_swap32 (U32 x)
 {
@@ -215,17 +198,6 @@ static U32 XXH_swap32 (U32 x)
             ((x >>  8) & 0x0000ff00 ) |
             ((x >> 24) & 0x000000ff );
 }
-static U64 XXH_swap64 (U64 x)
-{
-    return  ((x << 56) & 0xff00000000000000ULL) |
-            ((x << 40) & 0x00ff000000000000ULL) |
-            ((x << 24) & 0x0000ff0000000000ULL) |
-            ((x << 8)  & 0x000000ff00000000ULL) |
-            ((x >> 8)  & 0x00000000ff000000ULL) |
-            ((x >> 24) & 0x0000000000ff0000ULL) |
-            ((x >> 40) & 0x000000000000ff00ULL) |
-            ((x >> 56) & 0x00000000000000ffULL);
-}
 #endif
 
 
@@ -264,67 +236,23 @@ static U32 XXH_readBE32(const void* ptr)
     return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr);
 }
 
-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));
-    else
-        return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr);
-}
-
-FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)
-{
-    return XXH_readLE64_align(ptr, endian, XXH_unaligned);
-}
-
-static U64 XXH_readBE64(const void* ptr)
-{
-    return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr);
-}
-
 
 /* *************************************
 *  Macros
 ***************************************/
 #define XXH_STATIC_ASSERT(c)   { enum { XXH_static_assert = 1/(int)(!!(c)) }; }    /* use only *after* variable declarations */
+XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; }
 
 
-/* *************************************
-*  Constants
-***************************************/
+/* *******************************************************************
+*  32-bits hash functions
+*********************************************************************/
 static const U32 PRIME32_1 = 2654435761U;
 static const U32 PRIME32_2 = 2246822519U;
 static const U32 PRIME32_3 = 3266489917U;
 static const U32 PRIME32_4 =  668265263U;
 static const U32 PRIME32_5 =  374761393U;
 
-static const U64 PRIME64_1 = 11400714785074694791ULL;
-static const U64 PRIME64_2 = 14029467366897019727ULL;
-static const U64 PRIME64_3 =  1609587929392839161ULL;
-static const U64 PRIME64_4 =  9650029242287828579ULL;
-static const U64 PRIME64_5 =  2870177450012600261ULL;
-
-XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; }
-
-
-/* **************************
-*  Utils
-****************************/
-XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* restrict dstState, const XXH32_state_t* restrict srcState)
-{
-    memcpy(dstState, srcState, sizeof(*dstState));
-}
-
-XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* restrict dstState, const XXH64_state_t* restrict srcState)
-{
-    memcpy(dstState, srcState, sizeof(*dstState));
-}
-
-
-/* ***************************
-*  Simple Hash Functions
-*****************************/
-
 static U32 XXH32_round(U32 seed, U32 input)
 {
     seed += input * PRIME32_2;
@@ -394,10 +322,10 @@ XXH_PUBLIC_API unsigned int XXH32 (const void* input, size_t len, unsigned int s
 {
 #if 0
     /* Simple version, good for code maintenance, but unfortunately slow for small inputs */
-    XXH32_CREATESTATE_STATIC(state);
-    XXH32_reset(state, seed);
-    XXH32_update(state, input, len);
-    return XXH32_digest(state);
+    XXH32_state_t state;
+    XXH32_reset(&state, seed);
+    XXH32_update(&state, input, len);
+    return XXH32_digest(&state);
 #else
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
 
@@ -417,121 +345,8 @@ XXH_PUBLIC_API unsigned int XXH32 (const void* input, size_t len, unsigned int s
 }
 
 
-static U64 XXH64_round(U64 acc, U64 input)
-{
-    acc += input * PRIME64_2;
-    acc  = XXH_rotl64(acc, 31);
-    acc *= PRIME64_1;
-    return acc;
-}
-
-static U64 XXH64_mergeRound(U64 acc, U64 val)
-{
-    val  = XXH64_round(0, val);
-    acc ^= val;
-    acc  = acc * PRIME64_1 + PRIME64_4;
-    return acc;
-}
-
-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;
-    U64 h64;
-#define XXH_get64bits(p) XXH_readLE64_align(p, endian, align)
-
-#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
-    if (p==NULL) {
-        len=0;
-        bEnd=p=(const BYTE*)(size_t)32;
-    }
-#endif
-
-    if (len>=32) {
-        const BYTE* const limit = bEnd - 32;
-        U64 v1 = seed + PRIME64_1 + PRIME64_2;
-        U64 v2 = seed + PRIME64_2;
-        U64 v3 = seed + 0;
-        U64 v4 = seed - PRIME64_1;
-
-        do {
-            v1 = XXH64_round(v1, XXH_get64bits(p)); p+=8;
-            v2 = XXH64_round(v2, XXH_get64bits(p)); p+=8;
-            v3 = XXH64_round(v3, XXH_get64bits(p)); p+=8;
-            v4 = XXH64_round(v4, XXH_get64bits(p)); p+=8;
-        } while (p<=limit);
-
-        h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
-        h64 = XXH64_mergeRound(h64, v1);
-        h64 = XXH64_mergeRound(h64, v2);
-        h64 = XXH64_mergeRound(h64, v3);
-        h64 = XXH64_mergeRound(h64, v4);
-
-    } else {
-        h64  = seed + PRIME64_5;
-    }
-
-    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;
-}
-
-
-XXH_PUBLIC_API unsigned long long XXH64 (const void* input, size_t len, unsigned long long seed)
-{
-#if 0
-    /* Simple version, good for code maintenance, but unfortunately slow for small inputs */
-    XXH64_CREATESTATE_STATIC(state);
-    XXH64_reset(state, seed);
-    XXH64_update(state, input, len);
-    return XXH64_digest(state);
-#else
-    XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
 
-    if (XXH_FORCE_ALIGN_CHECK) {
-        if ((((size_t)input) & 7)==0) {  /* Input is aligned, let's leverage the speed advantage */
-            if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
-                return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
-            else
-                return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned);
-    }   }
-
-    if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
-        return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned);
-    else
-        return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned);
-#endif
-}
-
-
-/* **************************************************
-*  Advanced Hash Functions
-****************************************************/
+/*======   Hash streaming   ======*/
 
 XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void)
 {
@@ -543,24 +358,15 @@ XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr)
     return XXH_OK;
 }
 
-XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void)
+XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState)
 {
-    return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t));
-}
-XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr)
-{
-    XXH_free(statePtr);
-    return XXH_OK;
+    memcpy(dstState, srcState, sizeof(*dstState));
 }
 
-
-/*** Hash feed ***/
-
 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));
-    state.seed = seed;
+    memset(&state, 0, sizeof(state)-4);   /* do not write into reserved, for future removal */
     state.v1 = seed + PRIME32_1 + PRIME32_2;
     state.v2 = seed + PRIME32_2;
     state.v3 = seed + 0;
@@ -570,20 +376,6 @@ XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, unsigned int s
 }
 
 
-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));
-    state.seed = seed;
-    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));
-    return XXH_OK;
-}
-
-
 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;
@@ -593,11 +385,12 @@ FORCE_INLINE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void
     if (input==NULL) return XXH_ERROR;
 #endif
 
-    state->total_len += len;
+    state->total_len_32 += (unsigned)len;
+    state->large_len |= (len>=16) | (state->total_len_32>=16);
 
     if (state->memsize + len < 16)  {   /* fill in tmp buffer */
         XXH_memcpy((BYTE*)(state->mem32) + state->memsize, input, len);
-        state->memsize += (U32)len;
+        state->memsize += (unsigned)len;
         return XXH_OK;
     }
 
@@ -634,8 +427,8 @@ FORCE_INLINE XXH_errorcode XXH32_update_endian (XXH32_state_t* state, const void
     }
 
     if (p < bEnd) {
-        XXH_memcpy(state->mem32, p, bEnd-p);
-        state->memsize = (int)(bEnd-p);
+        XXH_memcpy(state->mem32, p, (size_t)(bEnd-p));
+        state->memsize = (unsigned)(bEnd-p);
     }
 
     return XXH_OK;
@@ -659,13 +452,13 @@ FORCE_INLINE U32 XXH32_digest_endian (const XXH32_state_t* state, XXH_endianess
     const BYTE* const bEnd = (const BYTE*)(state->mem32) + state->memsize;
     U32 h32;
 
-    if (state->total_len >= 16) {
+    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);
     } else {
-        h32 = state->seed + PRIME32_5;
+        h32 = state->v3 /* == seed */ + PRIME32_5;
     }
 
-    h32 += (U32) state->total_len;
+    h32 += state->total_len_32;
 
     while (p+4<=bEnd) {
         h32 += XXH_readLE32(p, endian) * PRIME32_3;
@@ -700,8 +493,257 @@ XXH_PUBLIC_API unsigned int XXH32_digest (const XXH32_state_t* state_in)
 }
 
 
+/*======   Canonical representation   ======*/
+
+/*! 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.
+*/
+
+XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash)
+{
+    XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t));
+    if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash);
+    memcpy(dst, &hash, sizeof(*dst));
+}
+
+XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src)
+{
+    return XXH_readBE32(src);
+}
+
+
+#ifndef XXH_NO_LONG_LONG
 
-/* **** XXH64 **** */
+/* *******************************************************************
+*  64-bits hash functions
+*********************************************************************/
+
+/*======   Memory access   ======*/
+
+#ifndef MEM_MODULE
+# define MEM_MODULE
+# 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. */
+# endif
+#endif
+
+
+#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2))
+
+/* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */
+static U64 XXH_read64(const void* memPtr) { return *(const U64*) memPtr; }
+
+#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1))
+
+/* __pack instructions are safer, but compiler specific, hence potentially problematic for some compilers */
+/* currently only defined for gcc and icc */
+typedef union { U32 u32; U64 u64; } __attribute__((packed)) unalign64;
+static U64 XXH_read64(const void* ptr) { return ((const unalign64*)ptr)->u64; }
+
+#else
+
+/* portable and safe solution. Generally efficient.
+ * see : http://stackoverflow.com/a/32095106/646947
+ */
+
+static U64 XXH_read64(const void* memPtr)
+{
+    U64 val;
+    memcpy(&val, memPtr, sizeof(val));
+    return val;
+}
+
+#endif   /* XXH_FORCE_DIRECT_MEMORY_ACCESS */
+
+#if defined(_MSC_VER)     /* Visual Studio */
+#  define XXH_swap64 _byteswap_uint64
+#elif XXH_GCC_VERSION >= 403
+#  define XXH_swap64 __builtin_bswap64
+#else
+static U64 XXH_swap64 (U64 x)
+{
+    return  ((x << 56) & 0xff00000000000000ULL) |
+            ((x << 40) & 0x00ff000000000000ULL) |
+            ((x << 24) & 0x0000ff0000000000ULL) |
+            ((x << 8)  & 0x000000ff00000000ULL) |
+            ((x >> 8)  & 0x00000000ff000000ULL) |
+            ((x >> 24) & 0x0000000000ff0000ULL) |
+            ((x >> 40) & 0x000000000000ff00ULL) |
+            ((x >> 56) & 0x00000000000000ffULL);
+}
+#endif
+
+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));
+    else
+        return endian==XXH_littleEndian ? *(const U64*)ptr : XXH_swap64(*(const U64*)ptr);
+}
+
+FORCE_INLINE U64 XXH_readLE64(const void* ptr, XXH_endianess endian)
+{
+    return XXH_readLE64_align(ptr, endian, XXH_unaligned);
+}
+
+static U64 XXH_readBE64(const void* ptr)
+{
+    return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr);
+}
+
+
+/*======   xxh64   ======*/
+
+static const U64 PRIME64_1 = 11400714785074694791ULL;
+static const U64 PRIME64_2 = 14029467366897019727ULL;
+static const U64 PRIME64_3 =  1609587929392839161ULL;
+static const U64 PRIME64_4 =  9650029242287828579ULL;
+static const U64 PRIME64_5 =  2870177450012600261ULL;
+
+static U64 XXH64_round(U64 acc, U64 input)
+{
+    acc += input * PRIME64_2;
+    acc  = XXH_rotl64(acc, 31);
+    acc *= PRIME64_1;
+    return acc;
+}
+
+static U64 XXH64_mergeRound(U64 acc, U64 val)
+{
+    val  = XXH64_round(0, val);
+    acc ^= val;
+    acc  = acc * PRIME64_1 + PRIME64_4;
+    return acc;
+}
+
+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;
+    U64 h64;
+#define XXH_get64bits(p) XXH_readLE64_align(p, endian, align)
+
+#ifdef XXH_ACCEPT_NULL_INPUT_POINTER
+    if (p==NULL) {
+        len=0;
+        bEnd=p=(const BYTE*)(size_t)32;
+    }
+#endif
+
+    if (len>=32) {
+        const BYTE* const limit = bEnd - 32;
+        U64 v1 = seed + PRIME64_1 + PRIME64_2;
+        U64 v2 = seed + PRIME64_2;
+        U64 v3 = seed + 0;
+        U64 v4 = seed - PRIME64_1;
+
+        do {
+            v1 = XXH64_round(v1, XXH_get64bits(p)); p+=8;
+            v2 = XXH64_round(v2, XXH_get64bits(p)); p+=8;
+            v3 = XXH64_round(v3, XXH_get64bits(p)); p+=8;
+            v4 = XXH64_round(v4, XXH_get64bits(p)); p+=8;
+        } while (p<=limit);
+
+        h64 = XXH_rotl64(v1, 1) + XXH_rotl64(v2, 7) + XXH_rotl64(v3, 12) + XXH_rotl64(v4, 18);
+        h64 = XXH64_mergeRound(h64, v1);
+        h64 = XXH64_mergeRound(h64, v2);
+        h64 = XXH64_mergeRound(h64, v3);
+        h64 = XXH64_mergeRound(h64, v4);
+
+    } else {
+        h64  = seed + PRIME64_5;
+    }
+
+    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;
+}
+
+
+XXH_PUBLIC_API unsigned long long XXH64 (const void* input, size_t len, unsigned long long seed)
+{
+#if 0
+    /* Simple version, good for code maintenance, but unfortunately slow for small inputs */
+    XXH64_state_t state;
+    XXH64_reset(&state, seed);
+    XXH64_update(&state, input, len);
+    return XXH64_digest(&state);
+#else
+    XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
+
+    if (XXH_FORCE_ALIGN_CHECK) {
+        if ((((size_t)input) & 7)==0) {  /* Input is aligned, let's leverage the speed advantage */
+            if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+                return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_aligned);
+            else
+                return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_aligned);
+    }   }
+
+    if ((endian_detected==XXH_littleEndian) || XXH_FORCE_NATIVE_FORMAT)
+        return XXH64_endian_align(input, len, seed, XXH_littleEndian, XXH_unaligned);
+    else
+        return XXH64_endian_align(input, len, seed, XXH_bigEndian, XXH_unaligned);
+#endif
+}
+
+/*======   Hash Streaming   ======*/
+
+XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void)
+{
+    return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t));
+}
+XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr)
+{
+    XXH_free(statePtr);
+    return XXH_OK;
+}
+
+XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* dstState, const XXH64_state_t* srcState)
+{
+    memcpy(dstState, srcState, sizeof(*dstState));
+}
+
+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 */
+    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));
+    return XXH_OK;
+}
 
 FORCE_INLINE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void* input, size_t len, XXH_endianess endian)
 {
@@ -751,8 +793,8 @@ FORCE_INLINE XXH_errorcode XXH64_update_endian (XXH64_state_t* state, const void
     }
 
     if (p < bEnd) {
-        XXH_memcpy(state->mem64, p, bEnd-p);
-        state->memsize = (int)(bEnd-p);
+        XXH_memcpy(state->mem64, p, (size_t)(bEnd-p));
+        state->memsize = (unsigned)(bEnd-p);
     }
 
     return XXH_OK;
@@ -768,8 +810,6 @@ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* state_in, const void*
         return XXH64_update_endian(state_in, input, len, XXH_bigEndian);
 }
 
-
-
 FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess endian)
 {
     const BYTE * p = (const BYTE*)state->mem64;
@@ -788,7 +828,7 @@ FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess
         h64 = XXH64_mergeRound(h64, v3);
         h64 = XXH64_mergeRound(h64, v4);
     } else {
-        h64  = state->seed + PRIME64_5;
+        h64  = state->v3 + PRIME64_5;
     }
 
     h64 += (U64) state->total_len;
@@ -821,7 +861,6 @@ FORCE_INLINE U64 XXH64_digest_endian (const XXH64_state_t* state, XXH_endianess
     return h64;
 }
 
-
 XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in)
 {
     XXH_endianess endian_detected = (XXH_endianess)XXH_CPU_LITTLE_ENDIAN;
@@ -833,22 +872,7 @@ XXH_PUBLIC_API unsigned long long XXH64_digest (const XXH64_state_t* state_in)
 }
 
 
-/* **************************
-*  Canonical representation
-****************************/
-
-/*! 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.
-*/
-
-XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash)
-{
-    XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t));
-    if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash);
-    memcpy(dst, &hash, sizeof(*dst));
-}
+/*====== Canonical representation   ======*/
 
 XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash)
 {
@@ -857,12 +881,9 @@ XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t
     memcpy(dst, &hash, sizeof(*dst));
 }
 
-XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src)
-{
-    return XXH_readBE32(src);
-}
-
 XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src)
 {
     return XXH_readBE64(src);
 }
+
+#endif  /* XXH_NO_LONG_LONG */
diff --git a/lib/xxhash.h b/lib/xxhash.h
index 4d7feff..870a6d9 100644
--- a/lib/xxhash.h
+++ b/lib/xxhash.h
@@ -83,20 +83,20 @@ typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
 *  API modifier
 ******************************/
 /** XXH_PRIVATE_API
-*   This is useful if you want to include xxhash functions in `static` mode
+*   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, so the file is still needed,
-*   but it's not useful to compile and link it anymore.
+*   `xxhash.c` is automatically included.
+*   It's not useful to compile and link it as a separate module.
 */
 #ifdef XXH_PRIVATE_API
 #  ifndef XXH_STATIC_LINKING_ONLY
 #    define XXH_STATIC_LINKING_ONLY
 #  endif
 #  if defined(__GNUC__)
-#    define XXH_PUBLIC_API static __attribute__((unused))
+#    define XXH_PUBLIC_API static __inline __attribute__((unused))
 #  elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */)
 #    define XXH_PUBLIC_API static inline
 #  elif defined(_MSC_VER)
@@ -111,10 +111,10 @@ typedef enum { XXH_OK=0, XXH_ERROR } XXH_errorcode;
 /*!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 another library which also includes xxHash,
+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 (so avoid to keep it NULL and avoid numeric values).
+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.
@@ -122,21 +122,25 @@ 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)
-#  define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32)
-#  define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64)
 #  define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber)
+#  define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32)
 #  define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState)
-#  define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState)
 #  define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState)
-#  define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState)
 #  define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset)
-#  define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset)
 #  define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update)
-#  define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update)
 #  define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest)
-#  define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest)
 #  define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState)
+#  define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash)
+#  define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical)
+#  define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64)
+#  define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState)
+#  define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState)
+#  define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset)
+#  define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update)
+#  define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest)
 #  define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState)
+#  define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash)
+#  define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical)
 #endif
 
 
@@ -145,58 +149,33 @@ regular symbol name will be automatically translated by this header.
 ***************************************/
 #define XXH_VERSION_MAJOR    0
 #define XXH_VERSION_MINOR    6
-#define XXH_VERSION_RELEASE  1
+#define XXH_VERSION_RELEASE  2
 #define XXH_VERSION_NUMBER  (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE)
 XXH_PUBLIC_API unsigned XXH_versionNumber (void);
 
 
-/* ****************************
-*  Simple Hash Functions
-******************************/
+/*-**********************************************************************
+*  32-bits hash
+************************************************************************/
 typedef unsigned int       XXH32_hash_t;
-typedef unsigned long long XXH64_hash_t;
-
-XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t length, unsigned int seed);
-XXH_PUBLIC_API XXH64_hash_t XXH64 (const void* input, size_t length, unsigned long long seed);
 
-/*!
-XXH32() :
+/*! XXH32() :
     Calculate the 32-bits 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
-XXH64() :
-    Calculate the 64-bits hash of sequence of length "len" stored at memory address "input".
-    "seed" can be used to alter the result predictably.
-    This function runs 2x faster on 64-bits systems, but slower on 32-bits systems (see benchmark).
-*/
-
+    Speed on Core 2 Duo @ 3 GHz (single thread, SMHasher benchmark) : 5.4 GB/s */
+XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t length, unsigned int seed);
 
-/* ****************************
-*  Streaming Hash Functions
-******************************/
+/*======   Streaming   ======*/
 typedef struct XXH32_state_s XXH32_state_t;   /* incomplete type */
-typedef struct XXH64_state_s XXH64_state_t;   /* incomplete type */
-
-/*! State allocation, compatible with dynamic libraries */
-
 XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void);
 XXH_PUBLIC_API XXH_errorcode  XXH32_freeState(XXH32_state_t* statePtr);
-
-XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void);
-XXH_PUBLIC_API XXH_errorcode  XXH64_freeState(XXH64_state_t* statePtr);
-
-
-/* hash streaming */
+XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state);
 
 XXH_PUBLIC_API XXH_errorcode XXH32_reset  (XXH32_state_t* statePtr, unsigned int seed);
 XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length);
 XXH_PUBLIC_API XXH32_hash_t  XXH32_digest (const XXH32_state_t* statePtr);
 
-XXH_PUBLIC_API XXH_errorcode XXH64_reset  (XXH64_state_t* statePtr, unsigned long long seed);
-XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* statePtr, const void* input, size_t length);
-XXH_PUBLIC_API XXH64_hash_t  XXH64_digest (const XXH64_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.
@@ -219,29 +198,11 @@ 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   ======*/
 
-/* **************************
-*  Utils
-****************************/
-#if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L))   /* ! C99 */
-#  define restrict   /* disable restrict */
-#endif
-
-XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* restrict dst_state, const XXH32_state_t* restrict src_state);
-XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* restrict dst_state, const XXH64_state_t* restrict src_state);
-
-
-/* **************************
-*  Canonical representation
-****************************/
 typedef struct { unsigned char digest[4]; } XXH32_canonical_t;
-typedef struct { unsigned char digest[8]; } XXH64_canonical_t;
-
 XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash);
-XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash);
-
 XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src);
-XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_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).
@@ -250,41 +211,76 @@ XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(const XXH64_canonical_t* src
 */
 
 
+#ifndef XXH_NO_LONG_LONG
+/*-**********************************************************************
+*  64-bits hash
+************************************************************************/
+typedef unsigned long long XXH64_hash_t;
+
+/*! XXH64() :
+    Calculate the 64-bits 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).
+*/
+XXH_PUBLIC_API XXH64_hash_t XXH64 (const void* input, size_t length, unsigned long long seed);
+
+/*======   Streaming   ======*/
+typedef struct XXH64_state_s XXH64_state_t;   /* incomplete type */
+XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void);
+XXH_PUBLIC_API XXH_errorcode  XXH64_freeState(XXH64_state_t* statePtr);
+XXH_PUBLIC_API void XXH64_copyState(XXH64_state_t* dst_state, const XXH64_state_t* src_state);
+
+XXH_PUBLIC_API XXH_errorcode XXH64_reset  (XXH64_state_t* statePtr, unsigned long long seed);
+XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH64_state_t* statePtr, const void* input, size_t length);
+XXH_PUBLIC_API XXH64_hash_t  XXH64_digest (const XXH64_state_t* statePtr);
+
+/*======   Canonical representation   ======*/
+typedef struct { unsigned char digest[8]; } XXH64_canonical_t;
+XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH64_canonical_t* dst, XXH64_hash_t hash);
+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.
-   They could change in a future version, becoming incompatible with a different version of the library.
+   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 definitions allow allocating XXH state statically (on stack) */
+/* 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 long long total_len;
-       unsigned seed;
+       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 seed;
        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 */
-
+#endif
 
 #  ifdef XXH_PRIVATE_API
-#    include "xxhash.c"   /* include xxhash functions as `static`, for inlining */
+#    include "xxhash.c"   /* include xxhash function bodies as `static`, for inlining */
 #  endif
 
 #endif /* XXH_STATIC_LINKING_ONLY */
diff --git a/programs/lz4.1 b/programs/lz4.1
index fcc7980..1d9238d 100644
--- a/programs/lz4.1
+++ b/programs/lz4.1
@@ -141,11 +141,13 @@ Benchmark mode, using # compression level.
 .
 .SS "Operation modifiers"
 .TP
-.B \-1
- fast compression (default)
-.TP
-.B \-9
- high compression
+.B \-#
+ compression level, with # being any value from 1 to 16.
+ Higher values trade compression speed for compression ratio.
+ Values above 16 are considered the same as 16.
+ Recommended values are 1 for fast compression (default), and 9 for high compression.
+ Speed/compression trade-off will vary depending on data to compress.
+ Decompression speed remains fast at all settings.
 
 .TP
 .BR \-f ", " --[no-]force
@@ -180,7 +182,7 @@ for files that have not been compressed with
  By default, the second filename is used as the destination filename for the compressed file.
  With
 .B -m
-, you can specify any number of input filenames. Each of them will be compressed
+, it is possible to specify any number of input filenames. Each of them will be compressed
 independently, and the resulting name of each compressed file will be
 .B filename.lz4
 .
diff --git a/programs/lz4io.c b/programs/lz4io.c
index 725d3a8..00576c2 100644
--- a/programs/lz4io.c
+++ b/programs/lz4io.c
@@ -258,7 +258,7 @@ static FILE* LZ4IO_openSrcFile(const char* srcFileName)
         SET_BINARY_MODE(stdin);
     } else {
         f = fopen(srcFileName, "rb");
-        if ( f==NULL ) DISPLAYLEVEL(1, "zstd: %s: %s \n", srcFileName, strerror(errno));
+        if ( f==NULL ) DISPLAYLEVEL(1, "%s: %s \n", srcFileName, strerror(errno));
     }
 
     return f;
@@ -285,10 +285,10 @@ static FILE* LZ4IO_openDstFile(const char* dstFileName)
             if (f != NULL) {  /* dest exists, prompt for overwrite authorization */
                 fclose(f);
                 if (g_displayLevel <= 1) {  /* No interaction possible */
-                    DISPLAY("zstd: %s already exists; not overwritten  \n", dstFileName);
+                    DISPLAY("%s already exists; not overwritten  \n", dstFileName);
                     return NULL;
                 }
-                DISPLAY("zstd: %s already exists; do you wish to overwrite (y/N) ? ", dstFileName);
+                DISPLAY("%s already exists; do you wish to overwrite (y/N) ? ", dstFileName);
                 {   int ch = getchar();
                     if ((ch!='Y') && (ch!='y')) {
                         DISPLAY("    not overwritten  \n");
@@ -297,7 +297,7 @@ static FILE* LZ4IO_openDstFile(const char* dstFileName)
                     while ((ch!=EOF) && (ch!='\n')) ch = getchar();  /* flush rest of input line */
         }   }   }
         f = fopen( dstFileName, "wb" );
-        if (f==NULL) DISPLAYLEVEL(1, "zstd: %s: %s\n", dstFileName, strerror(errno));
+        if (f==NULL) DISPLAYLEVEL(1, "%s: %s\n", dstFileName, strerror(errno));
     }
 
     /* sparse file */
diff --git a/programs/util.h b/programs/util.h
index d76ce6a..1ad61bc 100644
--- a/programs/util.h
+++ b/programs/util.h
@@ -427,7 +427,7 @@ UTIL_STATIC int UTIL_prepareFileList(const char *dirName, char** bufStart, size_
 UTIL_STATIC int UTIL_prepareFileList(const char *dirName, char** bufStart, size_t* pos, char** bufEnd)
 {
     (void)bufStart; (void)bufEnd; (void)pos;
-    fprintf(stderr, "Directory %s ignored (zstd compiled without _WIN32 or _POSIX_C_SOURCE)\n", dirName);
+    fprintf(stderr, "Directory %s ignored (lz4 compiled without _WIN32 or _POSIX_C_SOURCE)\n", dirName);
     return 0;
 }
 
diff --git a/tests/fullbench.c b/tests/fullbench.c
index 975cf8a..fa37fa6 100644
--- a/tests/fullbench.c
+++ b/tests/fullbench.c
@@ -386,7 +386,7 @@ static int local_LZ4_decompress_safe_partial(const char* in, char* out, int inSi
 /* frame functions */
 static int local_LZ4F_compressFrame(const char* in, char* out, int inSize)
 {
-    return (int)LZ4F_compressFrame(out, 2*inSize + 16, in, inSize, NULL);
+    return (int)LZ4F_compressFrame(out, LZ4F_compressFrameBound(inSize, NULL), in, inSize, NULL);
 }
 
 static LZ4F_decompressionContext_t g_dCtx;
@@ -532,9 +532,11 @@ int fullSpeedBench(const char** fileNamesTable, int nbFiles)
                         chunkP[0].origSize = (int)benchedSize; nbChunks=1;
                         break;
             case 40: compressionFunction = local_LZ4_saveDict; compressorName = "LZ4_saveDict";
+                        if (chunkP[0].origSize < 8) { DISPLAY(" cannot bench %s with less then 8 bytes \n", compressorName); continue; }
                         LZ4_loadDict(&LZ4_stream, chunkP[0].origBuffer, chunkP[0].origSize);
                         break;
             case 41: compressionFunction = local_LZ4_saveDictHC; compressorName = "LZ4_saveDictHC";
+                        if (chunkP[0].origSize < 8) { DISPLAY(" cannot bench %s with less then 8 bytes \n", compressorName); continue; }
                         LZ4_loadDictHC(&LZ4_streamHC, chunkP[0].origBuffer, chunkP[0].origSize);
                         break;
             case 60: DISPLAY("Obsolete compression functions : \n"); continue;