diff --git a/lib/dictBuilder.c b/lib/dictBuilder.c
index fe40bea464906bdfa9711fb7bc8cdeefe86f55f3..50b67d58d78c65750cb50e31a2951662ba82ad16 100644
--- a/lib/dictBuilder.c
+++ b/lib/dictBuilder.c
@@ -28,7 +28,7 @@
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
You can contact the author at :
- - Zstd source repository : https://www.zstd.net
+ - Zstd homepage : https://www.zstd.net
*/
/*-**************************************
@@ -124,23 +124,6 @@ void ZDICT_printHex(U32 dlevel, const void* ptr, size_t length)
}
-/*-*************************************
-* Exceptions
-***************************************/
-#ifndef DEBUG
-# define DEBUG 0
-#endif
-#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
-#define EXM_THROW(error, ...) \
-{ \
- DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
- DISPLAYLEVEL(1, "Error %i : ", error); \
- DISPLAYLEVEL(1, __VA_ARGS__); \
- DISPLAYLEVEL(1, "\n"); \
- exit(error); \
-}
-
-
/*-********************************************************
* Helper functions
**********************************************************/
@@ -516,7 +499,7 @@ static U32 ZDICT_dictSize(const dictItem* dictList)
}
-static void ZDICT_trainBuffer(dictItem* dictList, U32 dictListSize,
+static size_t ZDICT_trainBuffer(dictItem* dictList, U32 dictListSize,
const void* const buffer, const size_t bufferSize, /* buffer must end with noisy guard band */
const size_t* fileSizes, unsigned nbFiles,
U32 shiftRatio, unsigned maxDictSize)
@@ -528,18 +511,21 @@ static void ZDICT_trainBuffer(dictItem* dictList, U32 dictListSize,
U32* filePos = (U32*)malloc(nbFiles * sizeof(*filePos));
U32 minRatio = nbFiles >> shiftRatio;
int divSuftSortResult;
+ size_t result = 0;
/* init */
DISPLAYLEVEL(2, "\r%70s\r", ""); /* clean display line */
- if (!suffix0 || !reverseSuffix || !doneMarks || !filePos)
- EXM_THROW(1, "not enough memory for ZDICT_trainBuffer");
+ if (!suffix0 || !reverseSuffix || !doneMarks || !filePos) {
+ result = ERROR(memory_allocation);
+ goto _cleanup;
+ }
if (minRatio < MINRATIO) minRatio = MINRATIO;
memset(doneMarks, 0, bufferSize+16);
/* sort */
DISPLAYLEVEL(2, "sorting %u files of total size %u MB ...\n", nbFiles, (U32)(bufferSize>>20));
divSuftSortResult = divsufsort((const unsigned char*)buffer, suffix, (int)bufferSize, 0);
- if (divSuftSortResult != 0) EXM_THROW(2, "sort failed");
+ if (divSuftSortResult != 0) { result = ERROR(GENERIC); goto _cleanup; }
suffix[bufferSize] = (int)bufferSize; /* leads into noise */
suffix0[0] = (int)bufferSize; /* leads into noise */
{
@@ -578,10 +564,12 @@ static void ZDICT_trainBuffer(dictItem* dictList, U32 dictListSize,
dictList->pos = n;
}
+_cleanup:
free(suffix0);
free(reverseSuffix);
free(doneMarks);
free(filePos);
+ return result;
}
@@ -661,7 +649,11 @@ static size_t ZDICT_analyzeEntropy(void* dstBuffer, size_t maxDstSize,
esr.ref = ZSTD_createCCtx();
esr.zc = ZSTD_createCCtx();
esr.workPlace = malloc(BLOCKSIZE);
- if (!esr.ref || !esr.zc || !esr.workPlace) EXM_THROW(30, "Not enough memory");
+ if (!esr.ref || !esr.zc || !esr.workPlace) {
+ eSize = ERROR(memory_allocation);
+ DISPLAYLEVEL(1, "Not enough memory");
+ goto _cleanup;
+ }
if (compressionLevel==0) compressionLevel=g_compressionLevel_default;
params = ZSTD_getParams(compressionLevel, dictBufferSize + 15 KB);
params.strategy = ZSTD_greedy;
@@ -677,50 +669,82 @@ static size_t ZDICT_analyzeEntropy(void* dstBuffer, size_t maxDstSize,
/* analyze */
errorCode = HUF_buildCTable (hufTable, countLit, 255, huffLog);
- if (HUF_isError(errorCode)) EXM_THROW(31, "HUF_buildCTable error");
+ if (HUF_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "HUF_buildCTable error");
+ goto _cleanup;
+ }
huffLog = (U32)errorCode;
total=0; for (u=0; u<=OFFCODE_MAX; u++) total+=offcodeCount[u];
errorCode = FSE_normalizeCount(offcodeNCount, Offlog, offcodeCount, total, OFFCODE_MAX);
- if (FSE_isError(errorCode)) EXM_THROW(32, "FSE_normalizeCount error with offcodeCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_normalizeCount error with offcodeCount");
+ goto _cleanup;
+ }
Offlog = (U32)errorCode;
total=0; for (u=0; u<=MaxML; u++) total+=matchLengthCount[u];
errorCode = FSE_normalizeCount(matchLengthNCount, mlLog, matchLengthCount, total, MaxML);
- if (FSE_isError(errorCode)) EXM_THROW(33, "FSE_normalizeCount error with matchLengthCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_normalizeCount error with matchLengthCount");
+ goto _cleanup;
+ }
mlLog = (U32)errorCode;
total=0; for (u=0; u<=MaxLL; u++) total+=litlengthCount[u];
errorCode = FSE_normalizeCount(litlengthNCount, llLog, litlengthCount, total, MaxLL);
- if (FSE_isError(errorCode)) EXM_THROW(34, "FSE_normalizeCount error with litlengthCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_normalizeCount error with litlengthCount");
+ goto _cleanup;
+ }
llLog = (U32)errorCode;
/* write result to buffer */
errorCode = HUF_writeCTable(dstBuffer, maxDstSize, hufTable, 255, huffLog);
- if (HUF_isError(errorCode)) EXM_THROW(41, "HUF_writeCTable error");
+ if (HUF_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "HUF_writeCTable error");
+ goto _cleanup;
+ }
dstBuffer = (char*)dstBuffer + errorCode;
maxDstSize -= errorCode;
eSize += errorCode;
errorCode = FSE_writeNCount(dstBuffer, maxDstSize, offcodeNCount, OFFCODE_MAX, Offlog);
- if (FSE_isError(errorCode)) EXM_THROW(42, "FSE_writeNCount error with offcodeNCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_writeNCount error with offcodeNCount");
+ goto _cleanup;
+ }
dstBuffer = (char*)dstBuffer + errorCode;
maxDstSize -= errorCode;
eSize += errorCode;
errorCode = FSE_writeNCount(dstBuffer, maxDstSize, matchLengthNCount, MaxML, mlLog);
- if (FSE_isError(errorCode)) EXM_THROW(43, "FSE_writeNCount error with matchLengthNCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_writeNCount error with matchLengthNCount");
+ goto _cleanup;
+ }
dstBuffer = (char*)dstBuffer + errorCode;
maxDstSize -= errorCode;
eSize += errorCode;
errorCode = FSE_writeNCount(dstBuffer, maxDstSize, litlengthNCount, MaxLL, llLog);
- if (FSE_isError(errorCode)) EXM_THROW(43, "FSE_writeNCount error with litlengthNCount");
+ if (FSE_isError(errorCode)) {
+ eSize = ERROR(GENERIC);
+ DISPLAYLEVEL(1, "FSE_writeNCount error with litlengthNCount");
+ goto _cleanup;
+ }
dstBuffer = (char*)dstBuffer + errorCode;
maxDstSize -= errorCode;
eSize += errorCode;
- /* clean */
+_cleanup:
ZSTD_freeCCtx(esr.ref);
ZSTD_freeCCtx(esr.zc);
free(esr.workPlace);
@@ -730,12 +754,12 @@ static size_t ZDICT_analyzeEntropy(void* dstBuffer, size_t maxDstSize,
#define DIB_FASTSEGMENTSIZE 64
-/*! ZDICT_fastSampling (based on an idea by Giuseppe Ottaviano)
- Fill `dictBuffer` with stripes of size DIB_FASTSEGMENTSIZE from `samplesBuffer`
+/*! ZDICT_fastSampling() (based on an idea proposed by Giuseppe Ottaviano) :
+ Fill `dictBuffer` with stripes of size DIB_FASTSEGMENTSIZE from `samplesBuffer`,
up to `dictSize`.
Filling starts from the end of `dictBuffer`, down to maximum possible.
if `dictSize` is not a multiply of DIB_FASTSEGMENTSIZE, some bytes at beginning of `dictBuffer` won't be used.
- @return : amount of data written into `dictBuffer`
+ @return : amount of data written into `dictBuffer`,
or an error code
*/
static size_t ZDICT_fastSampling(void* dictBuffer, size_t dictSize,
diff --git a/programs/Makefile b/programs/Makefile
index a23b2f7f1e88689402f720b60e1e86532a20c677..97da43997759dd61ee8c7911f42d597639a422f6 100644
--- a/programs/Makefile
+++ b/programs/Makefile
@@ -53,7 +53,7 @@ BINDIR = $(PREFIX)/bin
MANDIR = $(PREFIX)/share/man/man1
ZSTDDIR = ../lib
-ZSTD_FILES := $(ZSTDDIR)/zstd_compress.c $(ZSTDDIR)/zstd_decompress.c $(ZSTDDIR)/fse.c $(ZSTDDIR)/huff0.c
+ZSTD_FILES := $(ZSTDDIR)/zstd_compress.c $(ZSTDDIR)/zstd_decompress.c $(ZSTDDIR)/fse.c $(ZSTDDIR)/huff0.c
ZSTD_LEGACY:= $(ZSTDDIR)/legacy/zstd_v01.c $(ZSTDDIR)/legacy/zstd_v02.c $(ZSTDDIR)/legacy/zstd_v03.c $(ZSTDDIR)/legacy/zstd_v04.c
ifeq ($(ZSTD_LEGACY_SUPPORT), 0)
@@ -84,12 +84,12 @@ default: zstd
all: zstd zstd32 fullbench fullbench32 fuzzer fuzzer32 zbufftest zbufftest32 paramgrill datagen
-zstd : $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c \
- zstdcli.c fileio.c $(ZSTD_FILEIO_LEGACY) bench.c xxhash.c datagen.c
+zstd : $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c $(ZSTDDIR)/dictBuilder.c $(ZSTDDIR)/divsufsort.c \
+ zstdcli.c fileio.c $(ZSTD_FILEIO_LEGACY) bench.c xxhash.c datagen.c dibio.c
$(CC) $(FLAGS) $^ -o $@$(EXT)
-zstd32: $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c \
- zstdcli.c fileio.c $(ZSTD_FILEIO_LEGACY) bench.c xxhash.c datagen.c
+zstd32: $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c $(ZSTDDIR)/dictBuilder.c $(ZSTDDIR)/divsufsort.c \
+ zstdcli.c fileio.c $(ZSTD_FILEIO_LEGACY) bench.c xxhash.c datagen.c dibio.c
$(CC) -m32 $(FLAGS) $^ -o $@$(EXT)
zstd_nolegacy :
@@ -101,12 +101,12 @@ zstd-pgo : clean zstd
rm zstd
$(MAKE) zstd MOREFLAGS=-fprofile-use
-zstd-noBench: $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c \
+zstd-frugal: $(ZSTD_FILES) $(ZSTDDIR)/zstd_buffered.c \
zstdcli.c fileio.c $(ZSTD_FILEIO_LEGACY)
- $(CC) $(FLAGS) -DZSTD_NOBENCH $^ -o zstd$(EXT)
+ $(CC) $(FLAGS) -DZSTD_NOBENCH -DZSTD_NODICT $^ -o zstd$(EXT)
-zstd-frugal: clean
- $(MAKE) zstd-noBench ZSTD_LEGACY_SUPPORT=0
+zstd-small: clean
+ CFLAGS=-Os $(MAKE) zstd-frugal ZSTD_LEGACY_SUPPORT=0
fullbench : $(ZSTD_FILES) \
datagen.c fullbench.c
@@ -140,7 +140,7 @@ datagen : datagen.c datagencli.c
$(CC) $(FLAGS) $^ -o $@$(EXT)
clean:
- @rm -f core *.o tmp* result* *.gcda \
+ @rm -f core *.o tmp* result* *.gcda dictionary *.zst \
zstd$(EXT) zstd32$(EXT) \
fullbench$(EXT) fullbench32$(EXT) \
fuzzer$(EXT) fuzzer32$(EXT) zbufftest$(EXT) zbufftest32$(EXT) \
diff --git a/programs/dibio.c b/programs/dibio.c
new file mode 100644
index 0000000000000000000000000000000000000000..a471d5450af6cf3662d621bd3c1b85657dc438f3
--- /dev/null
+++ b/programs/dibio.c
@@ -0,0 +1,275 @@
+/*
+ dibio - I/O API for dictionary builder
+ Copyright (C) Yann Collet 2016
+
+ GPL v2 License
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+ You can contact the author at :
+ - zstd homepage : http://www.zstd.net/
+*/
+
+/*-**************************************
+* Compiler Options
+****************************************/
+/* Disable some Visual warning messages */
+#ifdef _MSC_VER
+# define _CRT_SECURE_NO_WARNINGS /* fopen */
+# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
+#endif
+
+/* Unix Large Files support (>4GB) */
+#define _FILE_OFFSET_BITS 64
+#if (defined(__sun__) && (!defined(__LP64__))) /* Sun Solaris 32-bits requires specific definitions */
+# define _LARGEFILE_SOURCE
+#elif ! defined(__LP64__) /* No point defining Large file for 64 bit */
+# define _LARGEFILE64_SOURCE
+#endif
+
+
+/*-*************************************
+* Includes
+***************************************/
+#include <stdlib.h> /* malloc, free */
+#include <string.h> /* memset */
+#include <stdio.h> /* fprintf, fopen, ftello64 */
+#include <sys/types.h> /* stat64 */
+#include <sys/stat.h> /* stat64 */
+#include <time.h> /* clock */
+
+#include "mem.h" /* read */
+#include "error_private.h"
+#include "dictBuilder_static.h"
+
+/*-*************************************
+* Compiler specifics
+***************************************/
+#if !defined(S_ISREG)
+# define S_ISREG(x) (((x) & S_IFMT) == S_IFREG)
+#endif
+
+
+/*-*************************************
+* Constants
+***************************************/
+#define KB *(1 <<10)
+#define MB *(1 <<20)
+#define GB *(1U<<30)
+
+#define DICTLISTSIZE 10000
+#define MEMMULT 11
+static const size_t maxMemory = (sizeof(size_t) == 4) ? (2 GB - 64 MB) : ((size_t)(512 MB) << sizeof(size_t));
+
+#define NOISELENGTH 32
+#define PRIME1 2654435761U
+#define PRIME2 2246822519U
+
+
+/*-*************************************
+* Console display
+***************************************/
+#define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
+#define DISPLAYLEVEL(l, ...) if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); }
+static unsigned g_displayLevel = 0; /* 0 : no display; 1: errors; 2: default; 4: full information */
+void DiB_setNotificationLevel(unsigned l) { g_displayLevel=l; ZDICT_setNotificationLevel(l); }
+
+void DiB_printHex(U32 dlevel, const void* ptr, size_t length)
+{
+ const BYTE* const b = (const BYTE*)ptr;
+ size_t u;
+ for (u=0; u<length; u++) {
+ BYTE c = b[u];
+ if (c<32 || c>126) c = '.'; /* non-printable char */
+ DISPLAYLEVEL(dlevel, "%c", c);
+ }
+}
+
+
+/*-*************************************
+* Exceptions
+***************************************/
+#ifndef DEBUG
+# define DEBUG 0
+#endif
+#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
+#define EXM_THROW(error, ...) \
+{ \
+ DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
+ DISPLAYLEVEL(1, "Error %i : ", error); \
+ DISPLAYLEVEL(1, __VA_ARGS__); \
+ DISPLAYLEVEL(1, "\n"); \
+ exit(error); \
+}
+
+
+/* ********************************************************
+* Helper functions
+**********************************************************/
+unsigned DiB_isError(size_t errorCode) { return ERR_isError(errorCode); }
+
+const char* DiB_getErrorName(size_t errorCode) { return ERR_getErrorName(errorCode); }
+
+
+/* ********************************************************
+* File related operations
+**********************************************************/
+static unsigned long long DiB_getFileSize(const char* infilename)
+{
+ int r;
+#if defined(_MSC_VER)
+ struct _stat64 statbuf;
+ r = _stat64(infilename, &statbuf);
+#else
+ struct stat statbuf;
+ r = stat(infilename, &statbuf);
+#endif
+ if (r || !S_ISREG(statbuf.st_mode)) return 0; /* No good... */
+ return (unsigned long long)statbuf.st_size;
+}
+
+
+static unsigned long long DiB_getTotalFileSize(const char** fileNamesTable, unsigned nbFiles)
+{
+ unsigned long long total = 0;
+ unsigned n;
+ for (n=0; n<nbFiles; n++)
+ total += DiB_getFileSize(fileNamesTable[n]);
+ return total;
+}
+
+
+static void DiB_loadFiles(void* buffer, size_t bufferSize,
+ size_t* fileSizes,
+ const char** fileNamesTable, unsigned nbFiles)
+{
+ char* buff = (char*)buffer;
+ size_t pos = 0;
+ unsigned n;
+
+ for (n=0; n<nbFiles; n++) {
+ size_t readSize;
+ unsigned long long fileSize = DiB_getFileSize(fileNamesTable[n]);
+ FILE* f = fopen(fileNamesTable[n], "rb");
+ if (f==NULL) EXM_THROW(10, "impossible to open file %s", fileNamesTable[n]);
+ DISPLAYLEVEL(2, "Loading %s... \r", fileNamesTable[n]);
+ if (fileSize > bufferSize-pos) fileSize = 0; /* stop there, not enough memory to load all files */
+ readSize = fread(buff+pos, 1, (size_t)fileSize, f);
+ if (readSize != (size_t)fileSize) EXM_THROW(11, "could not read %s", fileNamesTable[n]);
+ pos += readSize;
+ fileSizes[n] = (size_t)fileSize;
+ fclose(f);
+ }
+}
+
+
+/*-********************************************************
+* Dictionary training functions
+**********************************************************/
+static size_t DiB_findMaxMem(unsigned long long requiredMem)
+{
+ size_t step = 8 MB;
+ void* testmem = NULL;
+
+ requiredMem = (((requiredMem >> 23) + 1) << 23);
+ requiredMem += 2 * step;
+ if (requiredMem > maxMemory) requiredMem = maxMemory;
+
+ while (!testmem) {
+ requiredMem -= step;
+ testmem = malloc((size_t)requiredMem);
+ }
+
+ free(testmem);
+ return (size_t)(requiredMem - step);
+}
+
+
+static void DiB_fillNoise(void* buffer, size_t length)
+{
+ unsigned acc = PRIME1;
+ size_t p=0;;
+
+ for (p=0; p<length; p++) {
+ acc *= PRIME2;
+ ((unsigned char*)buffer)[p] = (unsigned char)(acc >> 21);
+ }
+}
+
+
+static void DiB_saveDict(const char* dictFileName,
+ const void* buff, size_t buffSize)
+{
+ FILE* f;
+ size_t n;
+
+ f = fopen(dictFileName, "wb");
+ if (f==NULL) EXM_THROW(3, "cannot open %s ", dictFileName);
+
+ n = fwrite(buff, 1, buffSize, f);
+ if (n!=buffSize) EXM_THROW(4, "%s : write error", dictFileName)
+
+ n = (size_t)fclose(f);
+ if (n!=0) EXM_THROW(5, "%s : flush error", dictFileName)
+}
+
+
+int DiB_trainFromFiles(const char* dictFileName, unsigned maxDictSize,
+ const char** fileNamesTable, unsigned nbFiles,
+ ZDICT_params_t params)
+{
+ void* srcBuffer;
+ size_t benchedSize;
+ size_t* fileSizes = (size_t*)malloc(nbFiles * sizeof(size_t));
+ unsigned long long totalSizeToLoad = DiB_getTotalFileSize(fileNamesTable, nbFiles);
+ void* dictBuffer = malloc(maxDictSize);
+ size_t dictSize;
+ int result = 0;
+
+ /* init */
+ benchedSize = DiB_findMaxMem(totalSizeToLoad * MEMMULT) / MEMMULT;
+ if ((unsigned long long)benchedSize > totalSizeToLoad) benchedSize = (size_t)totalSizeToLoad;
+ if (benchedSize < totalSizeToLoad)
+ DISPLAYLEVEL(1, "Not enough memory; training on %u MB only...\n", (unsigned)(benchedSize >> 20));
+
+ /* Memory allocation & restrictions */
+ srcBuffer = malloc(benchedSize+NOISELENGTH); /* + noise */
+ if ((!fileSizes) || (!srcBuffer) || (!dictBuffer)) EXM_THROW(12, "not enough memory for DiB_trainFiles"); /* should not happen */
+
+ /* Load input buffer */
+ DiB_loadFiles(srcBuffer, benchedSize, fileSizes, fileNamesTable, nbFiles);
+ DiB_fillNoise((char*)srcBuffer + benchedSize, NOISELENGTH); /* guard band, for end of buffer condition */
+
+ /* call buffer version */
+ dictSize = ZDICT_trainFromBuffer_unsafe(dictBuffer, maxDictSize,
+ srcBuffer, fileSizes, nbFiles,
+ params);
+ if (ZDICT_isError(dictSize)) {
+ DISPLAYLEVEL(1, "dictionary training failed : %s", ZDICT_getErrorName(dictSize)); /* should not happen */
+ result = 1;
+ goto _cleanup;
+ }
+
+ /* save dict */
+ DISPLAYLEVEL(2, "Save dictionary of size %u into file %s \n", (U32)dictSize, dictFileName);
+ DiB_saveDict(dictFileName, dictBuffer, dictSize);
+
+ /* clean up */
+_cleanup:
+ free(srcBuffer);
+ free(dictBuffer);
+ free(fileSizes);
+ return result;
+}
diff --git a/programs/dibio.h b/programs/dibio.h
new file mode 100644
index 0000000000000000000000000000000000000000..9b8deb54b1ac4111288d5ef5aad9254a281edefb
--- /dev/null
+++ b/programs/dibio.h
@@ -0,0 +1,63 @@
+/*
+ dibio.h - I/O API for dictionary builder
+ Copyright (C) Yann Collet 2016
+
+ GPL v2 License
+
+ This program is free software; you can redistribute it and/or modify
+ it under the terms of the GNU General Public License as published by
+ the Free Software Foundation; either version 2 of the License, or
+ (at your option) any later version.
+
+ This program is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ GNU General Public License for more details.
+
+ You should have received a copy of the GNU General Public License along
+ with this program; if not, write to the Free Software Foundation, Inc.,
+ 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+
+ You can contact the author at :
+ - zstd homepage : http://www.zstd.net/
+*/
+
+/* This library is designed for a single-threaded console application.
+* It exit() and printf() into stderr when it encounters an error condition. */
+
+#ifndef DIBIO_H_003
+#define DIBIO_H_003
+
+
+/*-*************************************
+* Dependencies
+***************************************/
+#include "dictBuilder_static.h" /* ZDICT_params_t */
+
+
+/*-*************************************
+* Public functions
+***************************************/
+/*! DiB_trainFromFiles() :
+ Train a dictionary from a set of files provided by `fileNamesTable`.
+ Resulting dictionary is written into file `dictFileName`.
+ `parameters` is optional and can be provided with values set to 0, meaning "default".
+ @return : 0 == ok. Any other : error.
+*/
+int DiB_trainFromFiles(const char* dictFileName, unsigned maxDictSize,
+ const char** fileNamesTable, unsigned nbFiles,
+ ZDICT_params_t parameters);
+
+
+/*-*************************************
+* Helper functions
+***************************************/
+/*! DiB_setNotificationLevel
+ Set amount of notification to be displayed on the console.
+ default initial value : 0 = no console notification.
+ Note : not thread-safe (use a global constant)
+*/
+void DiB_setNotificationLevel(unsigned l);
+
+
+#endif
diff --git a/programs/zstdcli.c b/programs/zstdcli.c
index 3f3c5cd51e33393787bfbac092eef629698f69d1..21ec9ecd9bf359134898a23679e8fa1ea86a7ad8 100644
--- a/programs/zstdcli.c
+++ b/programs/zstdcli.c
@@ -22,10 +22,9 @@
- zstd homepage : http://www.zstd.net/
*/
/*
- Note : this is user program.
- It is not part of zstd compression library.
- The license of this compression CLI program is GPLv2.
- The license of zstd library is BSD.
+ Note : this is user program, not part of libzstd.
+ The license of this command line program is GPLv2.
+ The license of libzstd is BSD.
*/
@@ -46,7 +45,10 @@
#ifndef ZSTD_NOBENCH
# include "bench.h" /* BMK_benchFiles, BMK_SetNbIterations */
#endif
-#include "zstd.h" /* ZSTD version numbers */
+#include "zstd_static.h" /* ZSTD_maxCLevel, ZSTD version numbers */
+#ifndef ZSTD_NODICT
+# include "dibio.h" /* BMK_benchFiles, BMK_SetNbIterations */
+#endif
/*-************************************
@@ -55,9 +57,6 @@
#if defined(MSDOS) || defined(OS2) || defined(WIN32) || defined(_WIN32) || defined(__CYGWIN__)
# include <fcntl.h> /* _O_BINARY */
# include <io.h> /* _setmode, _isatty */
-# ifdef __MINGW32__
- /* int _fileno(FILE *stream); // seems no longer useful // MINGW somehow forgets to include this windows declaration into <stdio.h> */
-# endif
# define SET_BINARY_MODE(file) _setmode(_fileno(file), _O_BINARY)
# define IS_CONSOLE(stdStream) _isatty(_fileno(stdStream))
#else
@@ -78,6 +77,7 @@
#endif
#define AUTHOR "Yann Collet"
#define WELCOME_MESSAGE "*** %s %i-bits %s, by %s ***\n", COMPRESSOR_NAME, (int)(sizeof(void*)*8), ZSTD_VERSION, AUTHOR
+
#define ZSTD_EXTENSION ".zst"
#define ZSTD_CAT "zstdcat"
#define ZSTD_UNZSTD "unzstd"
@@ -86,6 +86,11 @@
#define MB *(1 <<20)
#define GB *(1U<<30)
+static const char* g_defaultDictName = "dictionary";
+static const unsigned g_defaultMaxDictSize = 110 KB;
+static const unsigned g_defaultDictCLevel = 5;
+static const unsigned g_defaultSelectivityLevel = 9;
+
/*-************************************
* Display Macros
@@ -96,35 +101,21 @@ static FILE* displayOut;
static unsigned displayLevel = 2; /* 0 : no display, 1: errors, 2 : + result + interaction + warnings, 3 : + progression, 4 : + information */
-/*-************************************
-* Exceptions
-**************************************/
-#define DEBUG 0
-#define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
-#define EXM_THROW(error, ...) \
-{ \
- DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
- DISPLAYLEVEL(1, "Error %i : ", error); \
- DISPLAYLEVEL(1, __VA_ARGS__); \
- DISPLAYLEVEL(1, "\n"); \
- exit(error); \
-}
-
-
/*-************************************
* Command Line
**************************************/
static int usage(const char* programName)
{
DISPLAY( "Usage :\n");
- DISPLAY( " %s [arg] [input] [output]\n", programName);
+ DISPLAY( " %s [args] [FILE(s)] [-o file]\n", programName);
DISPLAY( "\n");
- DISPLAY( "input : a filename\n");
+ DISPLAY( "FILE : a filename\n");
DISPLAY( " with no FILE, or when FILE is - , read standard input\n");
DISPLAY( "Arguments :\n");
- DISPLAY( " -# : # compression level (1-19, default:1) \n");
+ DISPLAY( " -# : # compression level (1-%u, default:1) \n", ZSTD_maxCLevel());
DISPLAY( " -d : decompression \n");
DISPLAY( " -D file: use `file` as Dictionary \n");
+ DISPLAY( " -o file: result stored into `file` (only possible if 1 input file) \n");
DISPLAY( " -f : overwrite output without prompting \n");
DISPLAY( " -h/-H : display help/long help and exit\n");
return 0;
@@ -139,14 +130,20 @@ static int usage_advanced(const char* programName)
DISPLAY( " -V : display Version number and exit\n");
DISPLAY( " -v : verbose mode\n");
DISPLAY( " -q : suppress warnings; specify twice to suppress errors too\n");
- DISPLAY( " -m : multiple input filenames mode \n");
DISPLAY( " -c : force write to standard output, even if it is the console\n");
+#ifndef ZSTD_NODICT
+ DISPLAY( "Dictionary builder :\n");
+ DISPLAY( "--train : Create a dictionary from a set of files \n");
+ DISPLAY( " -o file: `file` is dictionary name (default: %s) \n", g_defaultDictName);
+ DISPLAY( "--maxdict:limit dictionary to specified size (default : %u) \n", g_defaultMaxDictSize);
+ DISPLAY( " -s# : dictionary selectivity level (default: %u)\n", g_defaultSelectivityLevel);
+#endif
#ifndef ZSTD_NOBENCH
DISPLAY( "Benchmark arguments :\n");
DISPLAY( " -b# : benchmark file(s), using # compression level (default : 1) \n");
- DISPLAY( " -B# : cut file into independent blocks of size # (default : no block)\n");
+ DISPLAY( " -B# : cut file into independent blocks of size # (default: no block)\n");
DISPLAY( " -i# : iteration loops [1-9](default : 3)\n");
- DISPLAY( " -r# : test all compression levels from 1 to # (default : disabled)\n");
+ DISPLAY( " -r# : test all compression levels from 1 to # (default: disabled)\n");
#endif
return 0;
}
@@ -176,8 +173,10 @@ int main(int argCount, const char** argv)
forceStdout=0,
main_pause=0,
nextEntryIsDictionary=0,
- multiple=0,
- operationResult=0;
+ operationResult=0,
+ dictBuild=0,
+ nextArgumentIsOutFileName=0,
+ nextArgumentIsMaxDict=0;
unsigned cLevel = 1;
const char** filenameTable = (const char**)malloc(argCount * sizeof(const char*)); /* argCount >= 1 */
unsigned filenameIdx = 0;
@@ -185,11 +184,13 @@ int main(int argCount, const char** argv)
const char* outFileName = NULL;
const char* dictFileName = NULL;
char* dynNameSpace = NULL;
- const char extension[] = ZSTD_EXTENSION;
int rangeBench = 1;
+ unsigned maxDictSize = g_defaultMaxDictSize;
+ unsigned dictCLevel = g_defaultDictCLevel;
+ unsigned dictSelect = g_defaultSelectivityLevel;
/* init */
- (void)rangeBench; /* not used when ZSTD_NOBENCH set */
+ (void)rangeBench; (void)dictCLevel; /* not used when ZSTD_NOBENCH / ZSTD_NODICT set */
if (filenameTable==NULL) { DISPLAY("not enough memory\n"); exit(1); }
displayOut = stderr;
/* Pick out program name from path. Don't rely on stdlib because of conflicting behavior */
@@ -208,32 +209,32 @@ int main(int argCount, const char** argv)
/* long commands (--long-word) */
if (!strcmp(argument, "--version")) { displayOut=stdout; DISPLAY(WELCOME_MESSAGE); return 0; }
if (!strcmp(argument, "--help")) { displayOut=stdout; return usage_advanced(programName); }
- if (!strcmp(argument, "--multiple")) { multiple=1; continue; }
if (!strcmp(argument, "--verbose")) { displayLevel=4; continue; }
if (!strcmp(argument, "--quiet")) { displayLevel--; continue; }
+ if (!strcmp(argument, "--train")) { dictBuild=1; outFileName=g_defaultDictName; continue; }
+ if (!strcmp(argument, "--maxdict")) { nextArgumentIsMaxDict=1; continue; }
+
+ /* '-' means stdin/stdout */
+ if (!strcmp(argument, "-")){
+ if (!filenameIdx) { filenameIdx=1, filenameTable[0]=stdinmark; continue; }
+ outFileName=stdoutmark; continue;
+ }
/* Decode commands (note : aggregated commands are allowed) */
if (argument[0]=='-') {
- /* '-' means stdin/stdout */
- if (argument[1]==0) {
- if (!filenameIdx) { filenameIdx=1, filenameTable[0]=stdinmark; continue; }
- outFileName=stdoutmark; continue;
- }
-
argument++;
- while (argument[0]!=0)
- {
+ while (argument[0]!=0) {
+
/* compression Level */
- if ((*argument>='0') && (*argument<='9'))
- {
+ if ((*argument>='0') && (*argument<='9')) {
cLevel = 0;
- while ((*argument >= '0') && (*argument <= '9'))
- {
+ while ((*argument >= '0') && (*argument <= '9')) {
cLevel *= 10;
cLevel += *argument - '0';
argument++;
}
+ dictCLevel = cLevel;
continue;
}
@@ -247,9 +248,6 @@ int main(int argCount, const char** argv)
/* Decoding */
case 'd': decode=1; argument++; break;
- /* Multiple input files */
- case 'm': multiple=1; argument++; break;
-
/* Force stdout, even if stdout==console */
case 'c': forceStdout=1; outFileName=stdoutmark; displayLevel=1; argument++; break;
@@ -268,6 +266,9 @@ int main(int argCount, const char** argv)
/* keep source file (default anyway, so useless; for gzip/xz compatibility) */
case 'k': argument++; break;
+ /* dictionary name */
+ case 'o': nextArgumentIsOutFileName=1; argument++; break;
+
#ifndef ZSTD_NOBENCH
/* Benchmark */
case 'b': bench=1; argument++; break;
@@ -304,6 +305,13 @@ int main(int argCount, const char** argv)
break;
#endif /* ZSTD_NOBENCH */
+ /* Selection level */
+ case 's': argument++;
+ dictSelect = 0;
+ while ((*argument >= '0') && (*argument <= '9'))
+ dictSelect *= 10, dictSelect += *argument++ - '0';
+ break;
+
/* Pause at the end (hidden option) */
case 'p': main_pause=1; argument++; break;
@@ -314,13 +322,29 @@ int main(int argCount, const char** argv)
continue;
}
- /* dictionary */
if (nextEntryIsDictionary) {
nextEntryIsDictionary = 0;
dictFileName = argument;
continue;
}
+ if (nextArgumentIsOutFileName) {
+ nextArgumentIsOutFileName = 0;
+ outFileName = argument;
+ if (!strcmp(outFileName, "-")) outFileName = stdoutmark;
+ continue;
+ }
+
+ if (nextArgumentIsMaxDict) {
+ nextArgumentIsMaxDict = 0;
+ maxDictSize = 0;
+ while ((*argument>='0') && (*argument<='9'))
+ maxDictSize = maxDictSize * 10 + (*argument - '0'), argument++;
+ if (*argument=='k' || *argument=='K')
+ maxDictSize <<= 10;
+ continue;
+ }
+
/* add filename to list */
filenameTable[filenameIdx++] = argument;
}
@@ -336,65 +360,47 @@ int main(int argCount, const char** argv)
goto _end;
}
+ /* Check if dictionary builder is selected */
+ if (dictBuild) {
+#ifndef ZSTD_NODICT
+ ZDICT_params_t dictParams;
+ dictParams.compressionLevel = dictCLevel;
+ dictParams.selectivityLevel = dictSelect;
+ DiB_setNotificationLevel(displayLevel);
+ DiB_trainFromFiles(outFileName, maxDictSize, filenameTable, filenameIdx, dictParams);
+#endif
+ goto _end;
+ }
+
/* No input filename ==> use stdin */
if(!filenameIdx) filenameIdx=1, filenameTable[0]=stdinmark;
- /* Check if input defined as console; trigger an error in this case */
+ /* Check if input/output defined as console; trigger an error in this case */
if (!strcmp(filenameTable[0], stdinmark) && IS_CONSOLE(stdin) ) return badusage(programName);
+ if (outFileName && !strcmp(outFileName, stdoutmark) && IS_CONSOLE(stdout) && !forceStdout) return badusage(programName);
- /* No output filename ==> try to select one automatically (when possible) */
- if (filenameIdx>=2) outFileName = filenameTable[1];
- while (!outFileName) { /* while : just to allow break statement */
- if (!IS_CONSOLE(stdout)) { outFileName=stdoutmark; break; } /* Default to stdout whenever possible (i.e. not a console) */
- if (!decode) { /* compression to file */
- size_t l = strlen(filenameTable[0]);
- dynNameSpace = (char*)calloc(1,l+5);
- if (dynNameSpace==NULL) { DISPLAY("not enough memory\n"); exit(1); }
- strcpy(dynNameSpace, filenameTable[0]);
- strcpy(dynNameSpace+l, ZSTD_EXTENSION);
- outFileName = dynNameSpace;
- DISPLAYLEVEL(2, "Compressed filename will be : %s \n", outFileName);
- break;
- }
- /* decompression to file (automatic name will work only if input filename has correct format extension) */
- {
- size_t filenameSize = strlen(filenameTable[0]);
- if (strcmp(filenameTable[0] + (filenameSize-4), extension)) {
- DISPLAYLEVEL(1, "unknown suffix - cannot determine destination filename\n");
- return badusage(programName);
- }
- dynNameSpace = (char*)calloc(1,filenameSize+1);
- if (dynNameSpace==NULL) { DISPLAY("not enough memory\n"); exit(1); }
- outFileName = dynNameSpace;
- strcpy(dynNameSpace, filenameTable[0]);
- dynNameSpace[filenameSize-4]=0;
- DISPLAYLEVEL(2, "Decoding file %s \n", outFileName);
- } }
-
- /* Check if output is defined as console; trigger an error in this case */
- if (!strcmp(outFileName,stdoutmark) && IS_CONSOLE(stdout) && !forceStdout) return badusage(programName);
-
- /* No warning message in pure pipe mode (stdin + stdout) or multiple mode */
+ /* No warning message in pipe mode (stdin + stdout) or multiple mode */
if (!strcmp(filenameTable[0], stdinmark) && !strcmp(outFileName,stdoutmark) && (displayLevel==2)) displayLevel=1;
- if (multiple && (displayLevel==2)) displayLevel=1;
+ if ((filenameIdx>1) && (displayLevel==2)) displayLevel=1;
- if ((!multiple) && (filenameIdx>2)) {
- DISPLAY("Too many files on the command line (%u > 2). Do you mean -m ? \n", filenameIdx);
+ /* user-selected output filename only possible with a single file */
+ if ((outFileName) && (filenameIdx>1)) {
+ DISPLAY("Too many files (%u) on the command line. \n", filenameIdx);
return filenameIdx;
}
/* IO Stream/File */
FIO_setNotificationLevel(displayLevel);
if (decode) {
- if (multiple)
- operationResult = FIO_decompressMultipleFilenames(filenameTable, filenameIdx, ZSTD_EXTENSION, dictFileName);
- else
+ if (outFileName)
operationResult = FIO_decompressFilename(outFileName, filenameTable[0], dictFileName);
+ else
+ operationResult = FIO_decompressMultipleFilenames(filenameTable, filenameIdx, ZSTD_EXTENSION, dictFileName);
} else { /* compression */
- if (multiple)
- operationResult = FIO_compressMultipleFilenames(filenameTable, filenameIdx, ZSTD_EXTENSION, dictFileName, cLevel);
- else
+ if (outFileName)
operationResult = FIO_compressFilename(outFileName, filenameTable[0], dictFileName, cLevel);
+ else
+ operationResult = FIO_compressMultipleFilenames(filenameTable, filenameIdx, ZSTD_EXTENSION, dictFileName, cLevel);
}
_end: