1 /*
2 * Copyright (c) 2016-present, Yann Collet, Facebook, Inc.
3 * All rights reserved.
4 *
5 * This source code is licensed under both the BSD-style license (found in the
6 * LICENSE file in the root directory of this source tree) and the GPLv2 (found
7 * in the COPYING file in the root directory of this source tree).
8 * You may select, at your option, one of the above-listed licenses.
9 */
10
11
12 /*-************************************
13 * Tuning parameters
14 **************************************/
15 #ifndef ZSTDCLI_CLEVEL_DEFAULT
16 # define ZSTDCLI_CLEVEL_DEFAULT 3
17 #endif
18
19 #ifndef ZSTDCLI_CLEVEL_MAX
20 # define ZSTDCLI_CLEVEL_MAX 19 /* without using --ultra */
21 #endif
22
23
24
25 /*-************************************
26 * Dependencies
27 **************************************/
28 #include "platform.h" /* IS_CONSOLE, PLATFORM_POSIX_VERSION */
29 #include "util.h" /* UTIL_HAS_CREATEFILELIST, UTIL_createFileList */
30 #include <stdio.h> /* fprintf(), stdin, stdout, stderr */
31 #include <stdlib.h> /* getenv */
32 #include <string.h> /* strcmp, strlen */
33 #include <errno.h> /* errno */
34 #include "fileio.h" /* stdinmark, stdoutmark, ZSTD_EXTENSION */
35 #ifndef ZSTD_NOBENCH
36 # include "benchzstd.h" /* BMK_benchFiles */
37 #endif
38 #ifndef ZSTD_NODICT
39 # include "dibio.h" /* ZDICT_cover_params_t, DiB_trainFromFiles() */
40 #endif
41 #define ZSTD_STATIC_LINKING_ONLY /* ZSTD_minCLevel */
42 #include "zstd.h" /* ZSTD_VERSION_STRING, ZSTD_maxCLevel */
43
44
45 /*-************************************
46 * Constants
47 **************************************/
48 #define COMPRESSOR_NAME "zstd command line interface"
49 #ifndef ZSTD_VERSION
50 # define ZSTD_VERSION "v" ZSTD_VERSION_STRING
51 #endif
52 #define AUTHOR "Yann Collet"
53 #define WELCOME_MESSAGE "*** %s %i-bits %s, by %s ***\n", COMPRESSOR_NAME, (int)(sizeof(size_t)*8), ZSTD_VERSION, AUTHOR
54
55 #define ZSTD_ZSTDMT "zstdmt"
56 #define ZSTD_UNZSTD "unzstd"
57 #define ZSTD_CAT "zstdcat"
58 #define ZSTD_ZCAT "zcat"
59 #define ZSTD_GZ "gzip"
60 #define ZSTD_GUNZIP "gunzip"
61 #define ZSTD_GZCAT "gzcat"
62 #define ZSTD_LZMA "lzma"
63 #define ZSTD_UNLZMA "unlzma"
64 #define ZSTD_XZ "xz"
65 #define ZSTD_UNXZ "unxz"
66 #define ZSTD_LZ4 "lz4"
67 #define ZSTD_UNLZ4 "unlz4"
68
69 #define KB *(1 <<10)
70 #define MB *(1 <<20)
71 #define GB *(1U<<30)
72
73 #define DISPLAY_LEVEL_DEFAULT 2
74
75 static const char* g_defaultDictName = "dictionary";
76 static const unsigned g_defaultMaxDictSize = 110 KB;
77 static const int g_defaultDictCLevel = 3;
78 static const unsigned g_defaultSelectivityLevel = 9;
79 static const unsigned g_defaultMaxWindowLog = 27;
80 #define OVERLAP_LOG_DEFAULT 9999
81 #define LDM_PARAM_DEFAULT 9999 /* Default for parameters where 0 is valid */
82 static U32 g_overlapLog = OVERLAP_LOG_DEFAULT;
83 static U32 g_ldmHashLog = 0;
84 static U32 g_ldmMinMatch = 0;
85 static U32 g_ldmHashRateLog = LDM_PARAM_DEFAULT;
86 static U32 g_ldmBucketSizeLog = LDM_PARAM_DEFAULT;
87
88
89 #define DEFAULT_ACCEL 1
90
91 typedef enum { cover, fastCover, legacy } dictType;
92
93 /*-************************************
94 * Display Macros
95 **************************************/
96 #define DISPLAY(...) fprintf(g_displayOut, __VA_ARGS__)
97 #define DISPLAYLEVEL(l, ...) { if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); } }
98 static int g_displayLevel = DISPLAY_LEVEL_DEFAULT; /* 0 : no display, 1: errors, 2 : + result + interaction + warnings, 3 : + progression, 4 : + information */
99 static FILE* g_displayOut;
100
101
102 /*-************************************
103 * Command Line
104 **************************************/
usage(const char * programName)105 static int usage(const char* programName)
106 {
107 DISPLAY( "Usage : \n");
108 DISPLAY( " %s [args] [FILE(s)] [-o file] \n", programName);
109 DISPLAY( "\n");
110 DISPLAY( "FILE : a filename \n");
111 DISPLAY( " with no FILE, or when FILE is - , read standard input\n");
112 DISPLAY( "Arguments : \n");
113 #ifndef ZSTD_NOCOMPRESS
114 DISPLAY( " -# : # compression level (1-%d, default: %d) \n", ZSTDCLI_CLEVEL_MAX, ZSTDCLI_CLEVEL_DEFAULT);
115 #endif
116 #ifndef ZSTD_NODECOMPRESS
117 DISPLAY( " -d : decompression \n");
118 #endif
119 DISPLAY( " -D file: use `file` as Dictionary \n");
120 DISPLAY( " -o file: result stored into `file` (only if 1 input file) \n");
121 DISPLAY( " -f : overwrite output without prompting and (de)compress links \n");
122 DISPLAY( "--rm : remove source file(s) after successful de/compression \n");
123 DISPLAY( " -k : preserve source file(s) (default) \n");
124 DISPLAY( " -h/-H : display help/long help and exit \n");
125 return 0;
126 }
127
usage_advanced(const char * programName)128 static int usage_advanced(const char* programName)
129 {
130 DISPLAY(WELCOME_MESSAGE);
131 usage(programName);
132 DISPLAY( "\n");
133 DISPLAY( "Advanced arguments : \n");
134 DISPLAY( " -V : display Version number and exit \n");
135 DISPLAY( " -v : verbose mode; specify multiple times to increase verbosity\n");
136 DISPLAY( " -q : suppress warnings; specify twice to suppress errors too\n");
137 DISPLAY( " -c : force write to standard output, even if it is the console\n");
138 DISPLAY( " -l : print information about zstd compressed files \n");
139 DISPLAY( "--exclude-compressed: only compress files that are not previously compressed \n");
140 #ifndef ZSTD_NOCOMPRESS
141 DISPLAY( "--ultra : enable levels beyond %i, up to %i (requires more memory)\n", ZSTDCLI_CLEVEL_MAX, ZSTD_maxCLevel());
142 DISPLAY( "--long[=#]: enable long distance matching with given window log (default: %u)\n", g_defaultMaxWindowLog);
143 DISPLAY( "--fast[=#]: switch to very fast compression levels (default: %u)\n", 1);
144 DISPLAY( "--adapt : dynamically adapt compression level to I/O conditions \n");
145 DISPLAY( "--stream-size=# : optimize compression parameters for streaming input of given number of bytes \n");
146 DISPLAY( "--size-hint=# optimize compression parameters for streaming input of approximately this size\n");
147 DISPLAY( "--target-compressed-block-size=# : make compressed block near targeted size \n");
148 #ifdef ZSTD_MULTITHREAD
149 DISPLAY( " -T# : spawns # compression threads (default: 1, 0==# cores) \n");
150 DISPLAY( " -B# : select size of each job (default: 0==automatic) \n");
151 DISPLAY( "--rsyncable : compress using a rsync-friendly method (-B sets block size) \n");
152 #endif
153 DISPLAY( "--no-dictID : don't write dictID into header (dictionary compression)\n");
154 DISPLAY( "--[no-]check : integrity check (default: enabled) \n");
155 DISPLAY( "--[no-]compress-literals : force (un)compressed literals \n");
156 #endif
157 #ifdef UTIL_HAS_CREATEFILELIST
158 DISPLAY( " -r : operate recursively on directories \n");
159 DISPLAY( "--output-dir-flat[=directory]: all resulting files stored into `directory`. \n");
160 #endif
161 DISPLAY( "--format=zstd : compress files to the .zst format (default) \n");
162 #ifdef ZSTD_GZCOMPRESS
163 DISPLAY( "--format=gzip : compress files to the .gz format \n");
164 #endif
165 #ifdef ZSTD_LZMACOMPRESS
166 DISPLAY( "--format=xz : compress files to the .xz format \n");
167 DISPLAY( "--format=lzma : compress files to the .lzma format \n");
168 #endif
169 #ifdef ZSTD_LZ4COMPRESS
170 DISPLAY( "--format=lz4 : compress files to the .lz4 format \n");
171 #endif
172 #ifndef ZSTD_NODECOMPRESS
173 DISPLAY( "--test : test compressed file integrity \n");
174 #if ZSTD_SPARSE_DEFAULT
175 DISPLAY( "--[no-]sparse : sparse mode (default: enabled on file, disabled on stdout)\n");
176 #else
177 DISPLAY( "--[no-]sparse : sparse mode (default: disabled)\n");
178 #endif
179 #endif
180 DISPLAY( " -M# : Set a memory usage limit for decompression \n");
181 DISPLAY( "--no-progress : do not display the progress bar \n");
182 DISPLAY( "-- : All arguments after \"--\" are treated as files \n");
183 #ifndef ZSTD_NODICT
184 DISPLAY( "\n");
185 DISPLAY( "Dictionary builder : \n");
186 DISPLAY( "--train ## : create a dictionary from a training set of files \n");
187 DISPLAY( "--train-cover[=k=#,d=#,steps=#,split=#,shrink[=#]] : use the cover algorithm with optional args\n");
188 DISPLAY( "--train-fastcover[=k=#,d=#,f=#,steps=#,split=#,accel=#,shrink[=#]] : use the fast cover algorithm with optional args\n");
189 DISPLAY( "--train-legacy[=s=#] : use the legacy algorithm with selectivity (default: %u)\n", g_defaultSelectivityLevel);
190 DISPLAY( " -o file : `file` is dictionary name (default: %s) \n", g_defaultDictName);
191 DISPLAY( "--maxdict=# : limit dictionary to specified size (default: %u) \n", g_defaultMaxDictSize);
192 DISPLAY( "--dictID=# : force dictionary ID to specified value (default: random)\n");
193 #endif
194 #ifndef ZSTD_NOBENCH
195 DISPLAY( "\n");
196 DISPLAY( "Benchmark arguments : \n");
197 DISPLAY( " -b# : benchmark file(s), using # compression level (default: %d) \n", ZSTDCLI_CLEVEL_DEFAULT);
198 DISPLAY( " -e# : test all compression levels from -bX to # (default: 1)\n");
199 DISPLAY( " -i# : minimum evaluation time in seconds (default: 3s) \n");
200 DISPLAY( " -B# : cut file into independent blocks of size # (default: no block)\n");
201 DISPLAY( "--priority=rt : set process priority to real-time \n");
202 #endif
203 return 0;
204 }
205
badusage(const char * programName)206 static int badusage(const char* programName)
207 {
208 DISPLAYLEVEL(1, "Incorrect parameters\n");
209 if (g_displayLevel >= 2) usage(programName);
210 return 1;
211 }
212
waitEnter(void)213 static void waitEnter(void)
214 {
215 int unused;
216 DISPLAY("Press enter to continue...\n");
217 unused = getchar();
218 (void)unused;
219 }
220
lastNameFromPath(const char * path)221 static const char* lastNameFromPath(const char* path)
222 {
223 const char* name = path;
224 if (strrchr(name, '/')) name = strrchr(name, '/') + 1;
225 if (strrchr(name, '\\')) name = strrchr(name, '\\') + 1; /* windows */
226 return name;
227 }
228
229 /*! exeNameMatch() :
230 @return : a non-zero value if exeName matches test, excluding the extension
231 */
exeNameMatch(const char * exeName,const char * test)232 static int exeNameMatch(const char* exeName, const char* test)
233 {
234 return !strncmp(exeName, test, strlen(test)) &&
235 (exeName[strlen(test)] == '\0' || exeName[strlen(test)] == '.');
236 }
237
errorOut(const char * msg)238 static void errorOut(const char* msg)
239 {
240 DISPLAY("%s \n", msg); exit(1);
241 }
242
243 /*! readU32FromCharChecked() :
244 * @return 0 if success, and store the result in *value.
245 * allows and interprets K, KB, KiB, M, MB and MiB suffix.
246 * Will also modify `*stringPtr`, advancing it to position where it stopped reading.
247 * @return 1 if an overflow error occurs */
readU32FromCharChecked(const char ** stringPtr,unsigned * value)248 static int readU32FromCharChecked(const char** stringPtr, unsigned* value)
249 {
250 unsigned result = 0;
251 while ((**stringPtr >='0') && (**stringPtr <='9')) {
252 unsigned const max = (((unsigned)(-1)) / 10) - 1;
253 if (result > max) return 1; /* overflow error */
254 result *= 10;
255 result += (unsigned)(**stringPtr - '0');
256 (*stringPtr)++ ;
257 }
258 if ((**stringPtr=='K') || (**stringPtr=='M')) {
259 unsigned const maxK = ((unsigned)(-1)) >> 10;
260 if (result > maxK) return 1; /* overflow error */
261 result <<= 10;
262 if (**stringPtr=='M') {
263 if (result > maxK) return 1; /* overflow error */
264 result <<= 10;
265 }
266 (*stringPtr)++; /* skip `K` or `M` */
267 if (**stringPtr=='i') (*stringPtr)++;
268 if (**stringPtr=='B') (*stringPtr)++;
269 }
270 *value = result;
271 return 0;
272 }
273
274 /*! readU32FromChar() :
275 * @return : unsigned integer value read from input in `char` format.
276 * allows and interprets K, KB, KiB, M, MB and MiB suffix.
277 * Will also modify `*stringPtr`, advancing it to position where it stopped reading.
278 * Note : function will exit() program if digit sequence overflows */
readU32FromChar(const char ** stringPtr)279 static unsigned readU32FromChar(const char** stringPtr) {
280 static const char errorMsg[] = "error: numeric value too large";
281 unsigned result;
282 if (readU32FromCharChecked(stringPtr, &result)) { errorOut(errorMsg); }
283 return result;
284 }
285
286 /** longCommandWArg() :
287 * check if *stringPtr is the same as longCommand.
288 * If yes, @return 1 and advances *stringPtr to the position which immediately follows longCommand.
289 * @return 0 and doesn't modify *stringPtr otherwise.
290 */
longCommandWArg(const char ** stringPtr,const char * longCommand)291 static int longCommandWArg(const char** stringPtr, const char* longCommand)
292 {
293 size_t const comSize = strlen(longCommand);
294 int const result = !strncmp(*stringPtr, longCommand, comSize);
295 if (result) *stringPtr += comSize;
296 return result;
297 }
298
299
300 #ifndef ZSTD_NODICT
301
302 static const unsigned kDefaultRegression = 1;
303 /**
304 * parseCoverParameters() :
305 * reads cover parameters from *stringPtr (e.g. "--train-cover=k=48,d=8,steps=32") into *params
306 * @return 1 means that cover parameters were correct
307 * @return 0 in case of malformed parameters
308 */
parseCoverParameters(const char * stringPtr,ZDICT_cover_params_t * params)309 static unsigned parseCoverParameters(const char* stringPtr, ZDICT_cover_params_t* params)
310 {
311 memset(params, 0, sizeof(*params));
312 for (; ;) {
313 if (longCommandWArg(&stringPtr, "k=")) { params->k = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
314 if (longCommandWArg(&stringPtr, "d=")) { params->d = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
315 if (longCommandWArg(&stringPtr, "steps=")) { params->steps = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
316 if (longCommandWArg(&stringPtr, "split=")) {
317 unsigned splitPercentage = readU32FromChar(&stringPtr);
318 params->splitPoint = (double)splitPercentage / 100.0;
319 if (stringPtr[0]==',') { stringPtr++; continue; } else break;
320 }
321 if (longCommandWArg(&stringPtr, "shrink")) {
322 params->shrinkDictMaxRegression = kDefaultRegression;
323 params->shrinkDict = 1;
324 if (stringPtr[0]=='=') {
325 stringPtr++;
326 params->shrinkDictMaxRegression = readU32FromChar(&stringPtr);
327 }
328 if (stringPtr[0]==',') {
329 stringPtr++;
330 continue;
331 }
332 else break;
333 }
334 return 0;
335 }
336 if (stringPtr[0] != 0) return 0;
337 DISPLAYLEVEL(4, "cover: k=%u\nd=%u\nsteps=%u\nsplit=%u\nshrink%u\n", params->k, params->d, params->steps, (unsigned)(params->splitPoint * 100), params->shrinkDictMaxRegression);
338 return 1;
339 }
340
341 /**
342 * parseFastCoverParameters() :
343 * reads fastcover parameters from *stringPtr (e.g. "--train-fastcover=k=48,d=8,f=20,steps=32,accel=2") into *params
344 * @return 1 means that fastcover parameters were correct
345 * @return 0 in case of malformed parameters
346 */
parseFastCoverParameters(const char * stringPtr,ZDICT_fastCover_params_t * params)347 static unsigned parseFastCoverParameters(const char* stringPtr, ZDICT_fastCover_params_t* params)
348 {
349 memset(params, 0, sizeof(*params));
350 for (; ;) {
351 if (longCommandWArg(&stringPtr, "k=")) { params->k = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
352 if (longCommandWArg(&stringPtr, "d=")) { params->d = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
353 if (longCommandWArg(&stringPtr, "f=")) { params->f = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
354 if (longCommandWArg(&stringPtr, "steps=")) { params->steps = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
355 if (longCommandWArg(&stringPtr, "accel=")) { params->accel = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
356 if (longCommandWArg(&stringPtr, "split=")) {
357 unsigned splitPercentage = readU32FromChar(&stringPtr);
358 params->splitPoint = (double)splitPercentage / 100.0;
359 if (stringPtr[0]==',') { stringPtr++; continue; } else break;
360 }
361 if (longCommandWArg(&stringPtr, "shrink")) {
362 params->shrinkDictMaxRegression = kDefaultRegression;
363 params->shrinkDict = 1;
364 if (stringPtr[0]=='=') {
365 stringPtr++;
366 params->shrinkDictMaxRegression = readU32FromChar(&stringPtr);
367 }
368 if (stringPtr[0]==',') {
369 stringPtr++;
370 continue;
371 }
372 else break;
373 }
374 return 0;
375 }
376 if (stringPtr[0] != 0) return 0;
377 DISPLAYLEVEL(4, "cover: k=%u\nd=%u\nf=%u\nsteps=%u\nsplit=%u\naccel=%u\nshrink=%u\n", params->k, params->d, params->f, params->steps, (unsigned)(params->splitPoint * 100), params->accel, params->shrinkDictMaxRegression);
378 return 1;
379 }
380
381 /**
382 * parseLegacyParameters() :
383 * reads legacy dictionary builder parameters from *stringPtr (e.g. "--train-legacy=selectivity=8") into *selectivity
384 * @return 1 means that legacy dictionary builder parameters were correct
385 * @return 0 in case of malformed parameters
386 */
parseLegacyParameters(const char * stringPtr,unsigned * selectivity)387 static unsigned parseLegacyParameters(const char* stringPtr, unsigned* selectivity)
388 {
389 if (!longCommandWArg(&stringPtr, "s=") && !longCommandWArg(&stringPtr, "selectivity=")) { return 0; }
390 *selectivity = readU32FromChar(&stringPtr);
391 if (stringPtr[0] != 0) return 0;
392 DISPLAYLEVEL(4, "legacy: selectivity=%u\n", *selectivity);
393 return 1;
394 }
395
defaultCoverParams(void)396 static ZDICT_cover_params_t defaultCoverParams(void)
397 {
398 ZDICT_cover_params_t params;
399 memset(¶ms, 0, sizeof(params));
400 params.d = 8;
401 params.steps = 4;
402 params.splitPoint = 1.0;
403 params.shrinkDict = 0;
404 params.shrinkDictMaxRegression = kDefaultRegression;
405 return params;
406 }
407
defaultFastCoverParams(void)408 static ZDICT_fastCover_params_t defaultFastCoverParams(void)
409 {
410 ZDICT_fastCover_params_t params;
411 memset(¶ms, 0, sizeof(params));
412 params.d = 8;
413 params.f = 20;
414 params.steps = 4;
415 params.splitPoint = 0.75; /* different from default splitPoint of cover */
416 params.accel = DEFAULT_ACCEL;
417 params.shrinkDict = 0;
418 params.shrinkDictMaxRegression = kDefaultRegression;
419 return params;
420 }
421 #endif
422
423
424 /** parseAdaptParameters() :
425 * reads adapt parameters from *stringPtr (e.g. "--zstd=min=1,max=19) and store them into adaptMinPtr and adaptMaxPtr.
426 * Both adaptMinPtr and adaptMaxPtr must be already allocated and correctly initialized.
427 * There is no guarantee that any of these values will be updated.
428 * @return 1 means that parsing was successful,
429 * @return 0 in case of malformed parameters
430 */
parseAdaptParameters(const char * stringPtr,int * adaptMinPtr,int * adaptMaxPtr)431 static unsigned parseAdaptParameters(const char* stringPtr, int* adaptMinPtr, int* adaptMaxPtr)
432 {
433 for ( ; ;) {
434 if (longCommandWArg(&stringPtr, "min=")) { *adaptMinPtr = (int)readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
435 if (longCommandWArg(&stringPtr, "max=")) { *adaptMaxPtr = (int)readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
436 DISPLAYLEVEL(4, "invalid compression parameter \n");
437 return 0;
438 }
439 if (stringPtr[0] != 0) return 0; /* check the end of string */
440 if (*adaptMinPtr > *adaptMaxPtr) {
441 DISPLAYLEVEL(4, "incoherent adaptation limits \n");
442 return 0;
443 }
444 return 1;
445 }
446
447
448 /** parseCompressionParameters() :
449 * reads compression parameters from *stringPtr (e.g. "--zstd=wlog=23,clog=23,hlog=22,slog=6,mml=3,tlen=48,strat=6") into *params
450 * @return 1 means that compression parameters were correct
451 * @return 0 in case of malformed parameters
452 */
parseCompressionParameters(const char * stringPtr,ZSTD_compressionParameters * params)453 static unsigned parseCompressionParameters(const char* stringPtr, ZSTD_compressionParameters* params)
454 {
455 for ( ; ;) {
456 if (longCommandWArg(&stringPtr, "windowLog=") || longCommandWArg(&stringPtr, "wlog=")) { params->windowLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
457 if (longCommandWArg(&stringPtr, "chainLog=") || longCommandWArg(&stringPtr, "clog=")) { params->chainLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
458 if (longCommandWArg(&stringPtr, "hashLog=") || longCommandWArg(&stringPtr, "hlog=")) { params->hashLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
459 if (longCommandWArg(&stringPtr, "searchLog=") || longCommandWArg(&stringPtr, "slog=")) { params->searchLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
460 if (longCommandWArg(&stringPtr, "minMatch=") || longCommandWArg(&stringPtr, "mml=")) { params->minMatch = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
461 if (longCommandWArg(&stringPtr, "targetLength=") || longCommandWArg(&stringPtr, "tlen=")) { params->targetLength = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
462 if (longCommandWArg(&stringPtr, "strategy=") || longCommandWArg(&stringPtr, "strat=")) { params->strategy = (ZSTD_strategy)(readU32FromChar(&stringPtr)); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
463 if (longCommandWArg(&stringPtr, "overlapLog=") || longCommandWArg(&stringPtr, "ovlog=")) { g_overlapLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
464 if (longCommandWArg(&stringPtr, "ldmHashLog=") || longCommandWArg(&stringPtr, "lhlog=")) { g_ldmHashLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
465 if (longCommandWArg(&stringPtr, "ldmMinMatch=") || longCommandWArg(&stringPtr, "lmml=")) { g_ldmMinMatch = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
466 if (longCommandWArg(&stringPtr, "ldmBucketSizeLog=") || longCommandWArg(&stringPtr, "lblog=")) { g_ldmBucketSizeLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
467 if (longCommandWArg(&stringPtr, "ldmHashRateLog=") || longCommandWArg(&stringPtr, "lhrlog=")) { g_ldmHashRateLog = readU32FromChar(&stringPtr); if (stringPtr[0]==',') { stringPtr++; continue; } else break; }
468 DISPLAYLEVEL(4, "invalid compression parameter \n");
469 return 0;
470 }
471
472 DISPLAYLEVEL(4, "windowLog=%d, chainLog=%d, hashLog=%d, searchLog=%d \n", params->windowLog, params->chainLog, params->hashLog, params->searchLog);
473 DISPLAYLEVEL(4, "minMatch=%d, targetLength=%d, strategy=%d \n", params->minMatch, params->targetLength, params->strategy);
474 if (stringPtr[0] != 0) return 0; /* check the end of string */
475 return 1;
476 }
477
printVersion(void)478 static void printVersion(void)
479 {
480 DISPLAY(WELCOME_MESSAGE);
481 /* format support */
482 DISPLAYLEVEL(3, "*** supports: zstd");
483 #if defined(ZSTD_LEGACY_SUPPORT) && (ZSTD_LEGACY_SUPPORT>0) && (ZSTD_LEGACY_SUPPORT<8)
484 DISPLAYLEVEL(3, ", zstd legacy v0.%d+", ZSTD_LEGACY_SUPPORT);
485 #endif
486 #ifdef ZSTD_GZCOMPRESS
487 DISPLAYLEVEL(3, ", gzip");
488 #endif
489 #ifdef ZSTD_LZ4COMPRESS
490 DISPLAYLEVEL(3, ", lz4");
491 #endif
492 #ifdef ZSTD_LZMACOMPRESS
493 DISPLAYLEVEL(3, ", lzma, xz ");
494 #endif
495 DISPLAYLEVEL(3, "\n");
496 /* posix support */
497 #ifdef _POSIX_C_SOURCE
498 DISPLAYLEVEL(4, "_POSIX_C_SOURCE defined: %ldL\n", (long) _POSIX_C_SOURCE);
499 #endif
500 #ifdef _POSIX_VERSION
501 DISPLAYLEVEL(4, "_POSIX_VERSION defined: %ldL \n", (long) _POSIX_VERSION);
502 #endif
503 #ifdef PLATFORM_POSIX_VERSION
504 DISPLAYLEVEL(4, "PLATFORM_POSIX_VERSION defined: %ldL\n", (long) PLATFORM_POSIX_VERSION);
505 #endif
506 }
507
508 /* Environment variables for parameter setting */
509 #define ENV_CLEVEL "ZSTD_CLEVEL"
510
511 /* functions that pick up environment variables */
init_cLevel(void)512 static int init_cLevel(void) {
513 const char* const env = getenv(ENV_CLEVEL);
514 if (env) {
515 const char *ptr = env;
516 int sign = 1;
517 if (*ptr == '-') {
518 sign = -1;
519 ptr++;
520 } else if (*ptr == '+') {
521 ptr++;
522 }
523
524 if ((*ptr>='0') && (*ptr<='9')) {
525 unsigned absLevel;
526 if (readU32FromCharChecked(&ptr, &absLevel)) {
527 DISPLAYLEVEL(2, "Ignore environment variable setting %s=%s: numeric value too large\n", ENV_CLEVEL, env);
528 return ZSTDCLI_CLEVEL_DEFAULT;
529 } else if (*ptr == 0) {
530 return sign * (int)absLevel;
531 }
532 }
533
534 DISPLAYLEVEL(2, "Ignore environment variable setting %s=%s: not a valid integer value\n", ENV_CLEVEL, env);
535 }
536
537 return ZSTDCLI_CLEVEL_DEFAULT;
538 }
539
540 typedef enum { zom_compress, zom_decompress, zom_test, zom_bench, zom_train, zom_list } zstd_operation_mode;
541
542 #define CLEAN_RETURN(i) { operationResult = (i); goto _end; }
543
544 #ifdef ZSTD_NOCOMPRESS
545 /* symbols from compression library are not defined and should not be invoked */
546 # define MINCLEVEL -50
547 # define MAXCLEVEL 22
548 #else
549 # define MINCLEVEL ZSTD_minCLevel()
550 # define MAXCLEVEL ZSTD_maxCLevel()
551 #endif
552
main(int argCount,const char * argv[])553 int main(int argCount, const char* argv[])
554 {
555 int argNb,
556 followLinks = 0,
557 forceStdout = 0,
558 lastCommand = 0,
559 ldmFlag = 0,
560 main_pause = 0,
561 nbWorkers = 0,
562 adapt = 0,
563 adaptMin = MINCLEVEL,
564 adaptMax = MAXCLEVEL,
565 rsyncable = 0,
566 nextArgumentIsOutFileName = 0,
567 nextArgumentIsOutDirName = 0,
568 nextArgumentIsMaxDict = 0,
569 nextArgumentIsDictID = 0,
570 nextArgumentsAreFiles = 0,
571 nextEntryIsDictionary = 0,
572 operationResult = 0,
573 separateFiles = 0,
574 setRealTimePrio = 0,
575 singleThread = 0,
576 ultra=0;
577 double compressibility = 0.5;
578 unsigned bench_nbSeconds = 3; /* would be better if this value was synchronized from bench */
579 size_t blockSize = 0;
580
581 FIO_prefs_t* const prefs = FIO_createPreferences();
582 zstd_operation_mode operation = zom_compress;
583 ZSTD_compressionParameters compressionParams;
584 int cLevel;
585 int cLevelLast = -1000000000;
586 unsigned recursive = 0;
587 unsigned memLimit = 0;
588 const char** filenameTable = (const char**)malloc((size_t)argCount * sizeof(const char*)); /* argCount >= 1 */
589 unsigned filenameIdx = 0;
590 const char* programName = argv[0];
591 const char* outFileName = NULL;
592 const char* outDirName = NULL;
593 const char* dictFileName = NULL;
594 const char* suffix = ZSTD_EXTENSION;
595 unsigned maxDictSize = g_defaultMaxDictSize;
596 unsigned dictID = 0;
597 size_t streamSrcSize = 0;
598 size_t targetCBlockSize = 0;
599 size_t srcSizeHint = 0;
600 int dictCLevel = g_defaultDictCLevel;
601 unsigned dictSelect = g_defaultSelectivityLevel;
602 #ifdef UTIL_HAS_CREATEFILELIST
603 const char** extendedFileList = NULL;
604 char* fileNamesBuf = NULL;
605 unsigned fileNamesNb;
606 #endif
607 #ifndef ZSTD_NODICT
608 ZDICT_cover_params_t coverParams = defaultCoverParams();
609 ZDICT_fastCover_params_t fastCoverParams = defaultFastCoverParams();
610 dictType dict = fastCover;
611 #endif
612 #ifndef ZSTD_NOBENCH
613 BMK_advancedParams_t benchParams = BMK_initAdvancedParams();
614 #endif
615 ZSTD_literalCompressionMode_e literalCompressionMode = ZSTD_lcm_auto;
616
617
618 /* init */
619 (void)recursive; (void)cLevelLast; /* not used when ZSTD_NOBENCH set */
620 (void)memLimit; /* not used when ZSTD_NODECOMPRESS set */
621 if (filenameTable==NULL) { DISPLAY("zstd: %s \n", strerror(errno)); exit(1); }
622 filenameTable[0] = stdinmark;
623 g_displayOut = stderr;
624 cLevel = init_cLevel();
625 programName = lastNameFromPath(programName);
626 #ifdef ZSTD_MULTITHREAD
627 nbWorkers = 1;
628 #endif
629
630 /* preset behaviors */
631 if (exeNameMatch(programName, ZSTD_ZSTDMT)) nbWorkers=0, singleThread=0;
632 if (exeNameMatch(programName, ZSTD_UNZSTD)) operation=zom_decompress;
633 if (exeNameMatch(programName, ZSTD_CAT)) { operation=zom_decompress; FIO_overwriteMode(prefs); forceStdout=1; followLinks=1; outFileName=stdoutmark; g_displayLevel=1; } /* supports multiple formats */
634 if (exeNameMatch(programName, ZSTD_ZCAT)) { operation=zom_decompress; FIO_overwriteMode(prefs); forceStdout=1; followLinks=1; outFileName=stdoutmark; g_displayLevel=1; } /* behave like zcat, also supports multiple formats */
635 if (exeNameMatch(programName, ZSTD_GZ)) { suffix = GZ_EXTENSION; FIO_setCompressionType(prefs, FIO_gzipCompression); FIO_setRemoveSrcFile(prefs, 1); } /* behave like gzip */
636 if (exeNameMatch(programName, ZSTD_GUNZIP)) { operation=zom_decompress; FIO_setRemoveSrcFile(prefs, 1); } /* behave like gunzip, also supports multiple formats */
637 if (exeNameMatch(programName, ZSTD_GZCAT)) { operation=zom_decompress; FIO_overwriteMode(prefs); forceStdout=1; followLinks=1; outFileName=stdoutmark; g_displayLevel=1; } /* behave like gzcat, also supports multiple formats */
638 if (exeNameMatch(programName, ZSTD_LZMA)) { suffix = LZMA_EXTENSION; FIO_setCompressionType(prefs, FIO_lzmaCompression); FIO_setRemoveSrcFile(prefs, 1); } /* behave like lzma */
639 if (exeNameMatch(programName, ZSTD_UNLZMA)) { operation=zom_decompress; FIO_setCompressionType(prefs, FIO_lzmaCompression); FIO_setRemoveSrcFile(prefs, 1); } /* behave like unlzma, also supports multiple formats */
640 if (exeNameMatch(programName, ZSTD_XZ)) { suffix = XZ_EXTENSION; FIO_setCompressionType(prefs, FIO_xzCompression); FIO_setRemoveSrcFile(prefs, 1); } /* behave like xz */
641 if (exeNameMatch(programName, ZSTD_UNXZ)) { operation=zom_decompress; FIO_setCompressionType(prefs, FIO_xzCompression); FIO_setRemoveSrcFile(prefs, 1); } /* behave like unxz, also supports multiple formats */
642 if (exeNameMatch(programName, ZSTD_LZ4)) { suffix = LZ4_EXTENSION; FIO_setCompressionType(prefs, FIO_lz4Compression); } /* behave like lz4 */
643 if (exeNameMatch(programName, ZSTD_UNLZ4)) { operation=zom_decompress; FIO_setCompressionType(prefs, FIO_lz4Compression); } /* behave like unlz4, also supports multiple formats */
644 memset(&compressionParams, 0, sizeof(compressionParams));
645
646 /* init crash handler */
647 FIO_addAbortHandler();
648
649 /* command switches */
650 for (argNb=1; argNb<argCount; argNb++) {
651 const char* argument = argv[argNb];
652 if(!argument) continue; /* Protection if argument empty */
653
654 if (nextArgumentsAreFiles==0) {
655 /* "-" means stdin/stdout */
656 if (!strcmp(argument, "-")){
657 if (!filenameIdx) {
658 filenameIdx=1, filenameTable[0]=stdinmark;
659 outFileName=stdoutmark;
660 g_displayLevel-=(g_displayLevel==2);
661 continue;
662 } }
663
664 /* Decode commands (note : aggregated commands are allowed) */
665 if (argument[0]=='-') {
666
667 if (argument[1]=='-') {
668 /* long commands (--long-word) */
669 if (!strcmp(argument, "--")) { nextArgumentsAreFiles=1; continue; } /* only file names allowed from now on */
670 if (!strcmp(argument, "--list")) { operation=zom_list; continue; }
671 if (!strcmp(argument, "--compress")) { operation=zom_compress; continue; }
672 if (!strcmp(argument, "--decompress")) { operation=zom_decompress; continue; }
673 if (!strcmp(argument, "--uncompress")) { operation=zom_decompress; continue; }
674 if (!strcmp(argument, "--force")) { FIO_overwriteMode(prefs); forceStdout=1; followLinks=1; continue; }
675 if (!strcmp(argument, "--version")) { g_displayOut=stdout; DISPLAY(WELCOME_MESSAGE); CLEAN_RETURN(0); }
676 if (!strcmp(argument, "--help")) { g_displayOut=stdout; CLEAN_RETURN(usage_advanced(programName)); }
677 if (!strcmp(argument, "--verbose")) { g_displayLevel++; continue; }
678 if (!strcmp(argument, "--quiet")) { g_displayLevel--; continue; }
679 if (!strcmp(argument, "--stdout")) { forceStdout=1; outFileName=stdoutmark; g_displayLevel-=(g_displayLevel==2); continue; }
680 if (!strcmp(argument, "--ultra")) { ultra=1; continue; }
681 if (!strcmp(argument, "--check")) { FIO_setChecksumFlag(prefs, 2); continue; }
682 if (!strcmp(argument, "--no-check")) { FIO_setChecksumFlag(prefs, 0); continue; }
683 if (!strcmp(argument, "--sparse")) { FIO_setSparseWrite(prefs, 2); continue; }
684 if (!strcmp(argument, "--no-sparse")) { FIO_setSparseWrite(prefs, 0); continue; }
685 if (!strcmp(argument, "--test")) { operation=zom_test; continue; }
686 if (!strcmp(argument, "--train")) { operation=zom_train; if (outFileName==NULL) outFileName=g_defaultDictName; continue; }
687 if (!strcmp(argument, "--maxdict")) { nextArgumentIsMaxDict=1; lastCommand=1; continue; } /* kept available for compatibility with old syntax ; will be removed one day */
688 if (!strcmp(argument, "--dictID")) { nextArgumentIsDictID=1; lastCommand=1; continue; } /* kept available for compatibility with old syntax ; will be removed one day */
689 if (!strcmp(argument, "--no-dictID")) { FIO_setDictIDFlag(prefs, 0); continue; }
690 if (!strcmp(argument, "--keep")) { FIO_setRemoveSrcFile(prefs, 0); continue; }
691 if (!strcmp(argument, "--rm")) { FIO_setRemoveSrcFile(prefs, 1); continue; }
692 if (!strcmp(argument, "--priority=rt")) { setRealTimePrio = 1; continue; }
693 if (!strcmp(argument, "--output-dir-flat")) {nextArgumentIsOutDirName=1; lastCommand=1; continue; }
694 if (!strcmp(argument, "--adapt")) { adapt = 1; continue; }
695 if (longCommandWArg(&argument, "--adapt=")) { adapt = 1; if (!parseAdaptParameters(argument, &adaptMin, &adaptMax)) CLEAN_RETURN(badusage(programName)); continue; }
696 if (!strcmp(argument, "--single-thread")) { nbWorkers = 0; singleThread = 1; continue; }
697 if (!strcmp(argument, "--format=zstd")) { suffix = ZSTD_EXTENSION; FIO_setCompressionType(prefs, FIO_zstdCompression); continue; }
698 #ifdef ZSTD_GZCOMPRESS
699 if (!strcmp(argument, "--format=gzip")) { suffix = GZ_EXTENSION; FIO_setCompressionType(prefs, FIO_gzipCompression); continue; }
700 #endif
701 #ifdef ZSTD_LZMACOMPRESS
702 if (!strcmp(argument, "--format=lzma")) { suffix = LZMA_EXTENSION; FIO_setCompressionType(prefs, FIO_lzmaCompression); continue; }
703 if (!strcmp(argument, "--format=xz")) { suffix = XZ_EXTENSION; FIO_setCompressionType(prefs, FIO_xzCompression); continue; }
704 #endif
705 #ifdef ZSTD_LZ4COMPRESS
706 if (!strcmp(argument, "--format=lz4")) { suffix = LZ4_EXTENSION; FIO_setCompressionType(prefs, FIO_lz4Compression); continue; }
707 #endif
708 if (!strcmp(argument, "--rsyncable")) { rsyncable = 1; continue; }
709 if (!strcmp(argument, "--compress-literals")) { literalCompressionMode = ZSTD_lcm_huffman; continue; }
710 if (!strcmp(argument, "--no-compress-literals")) { literalCompressionMode = ZSTD_lcm_uncompressed; continue; }
711 if (!strcmp(argument, "--no-progress")) { FIO_setNoProgress(1); continue; }
712 if (!strcmp(argument, "--exclude-compressed")) { FIO_setExcludeCompressedFile(prefs, 1); continue; }
713 /* long commands with arguments */
714 #ifndef ZSTD_NODICT
715 if (longCommandWArg(&argument, "--train-cover")) {
716 operation = zom_train;
717 if (outFileName == NULL)
718 outFileName = g_defaultDictName;
719 dict = cover;
720 /* Allow optional arguments following an = */
721 if (*argument == 0) { memset(&coverParams, 0, sizeof(coverParams)); }
722 else if (*argument++ != '=') { CLEAN_RETURN(badusage(programName)); }
723 else if (!parseCoverParameters(argument, &coverParams)) { CLEAN_RETURN(badusage(programName)); }
724 continue;
725 }
726 if (longCommandWArg(&argument, "--train-fastcover")) {
727 operation = zom_train;
728 if (outFileName == NULL)
729 outFileName = g_defaultDictName;
730 dict = fastCover;
731 /* Allow optional arguments following an = */
732 if (*argument == 0) { memset(&fastCoverParams, 0, sizeof(fastCoverParams)); }
733 else if (*argument++ != '=') { CLEAN_RETURN(badusage(programName)); }
734 else if (!parseFastCoverParameters(argument, &fastCoverParams)) { CLEAN_RETURN(badusage(programName)); }
735 continue;
736 }
737 if (longCommandWArg(&argument, "--train-legacy")) {
738 operation = zom_train;
739 if (outFileName == NULL)
740 outFileName = g_defaultDictName;
741 dict = legacy;
742 /* Allow optional arguments following an = */
743 if (*argument == 0) { continue; }
744 else if (*argument++ != '=') { CLEAN_RETURN(badusage(programName)); }
745 else if (!parseLegacyParameters(argument, &dictSelect)) { CLEAN_RETURN(badusage(programName)); }
746 continue;
747 }
748 #endif
749 if (longCommandWArg(&argument, "--threads=")) { nbWorkers = (int)readU32FromChar(&argument); continue; }
750 if (longCommandWArg(&argument, "--memlimit=")) { memLimit = readU32FromChar(&argument); continue; }
751 if (longCommandWArg(&argument, "--memory=")) { memLimit = readU32FromChar(&argument); continue; }
752 if (longCommandWArg(&argument, "--memlimit-decompress=")) { memLimit = readU32FromChar(&argument); continue; }
753 if (longCommandWArg(&argument, "--block-size=")) { blockSize = readU32FromChar(&argument); continue; }
754 if (longCommandWArg(&argument, "--maxdict=")) { maxDictSize = readU32FromChar(&argument); continue; }
755 if (longCommandWArg(&argument, "--dictID=")) { dictID = readU32FromChar(&argument); continue; }
756 if (longCommandWArg(&argument, "--zstd=")) { if (!parseCompressionParameters(argument, &compressionParams)) CLEAN_RETURN(badusage(programName)); continue; }
757 if (longCommandWArg(&argument, "--stream-size=")) { streamSrcSize = readU32FromChar(&argument); continue; }
758 if (longCommandWArg(&argument, "--target-compressed-block-size=")) { targetCBlockSize = readU32FromChar(&argument); continue; }
759 if (longCommandWArg(&argument, "--size-hint=")) { srcSizeHint = readU32FromChar(&argument); continue; }
760 if (longCommandWArg(&argument, "--output-dir-flat=")) { outDirName = argument; continue; }
761 if (longCommandWArg(&argument, "--long")) {
762 unsigned ldmWindowLog = 0;
763 ldmFlag = 1;
764 /* Parse optional window log */
765 if (*argument == '=') {
766 ++argument;
767 ldmWindowLog = readU32FromChar(&argument);
768 } else if (*argument != 0) {
769 /* Invalid character following --long */
770 CLEAN_RETURN(badusage(programName));
771 }
772 /* Only set windowLog if not already set by --zstd */
773 if (compressionParams.windowLog == 0)
774 compressionParams.windowLog = ldmWindowLog;
775 continue;
776 }
777 #ifndef ZSTD_NOCOMPRESS /* linking ZSTD_minCLevel() requires compression support */
778 if (longCommandWArg(&argument, "--fast")) {
779 /* Parse optional acceleration factor */
780 if (*argument == '=') {
781 U32 const maxFast = (U32)-ZSTD_minCLevel();
782 U32 fastLevel;
783 ++argument;
784 fastLevel = readU32FromChar(&argument);
785 if (fastLevel > maxFast) fastLevel = maxFast;
786 if (fastLevel) {
787 dictCLevel = cLevel = -(int)fastLevel;
788 } else {
789 CLEAN_RETURN(badusage(programName));
790 }
791 } else if (*argument != 0) {
792 /* Invalid character following --fast */
793 CLEAN_RETURN(badusage(programName));
794 } else {
795 cLevel = -1; /* default for --fast */
796 }
797 continue;
798 }
799 #endif
800 /* fall-through, will trigger bad_usage() later on */
801 }
802
803 argument++;
804 while (argument[0]!=0) {
805 if (lastCommand) {
806 DISPLAY("error : command must be followed by argument \n");
807 CLEAN_RETURN(1);
808 }
809 #ifndef ZSTD_NOCOMPRESS
810 /* compression Level */
811 if ((*argument>='0') && (*argument<='9')) {
812 dictCLevel = cLevel = (int)readU32FromChar(&argument);
813 continue;
814 }
815 #endif
816
817 switch(argument[0])
818 {
819 /* Display help */
820 case 'V': g_displayOut=stdout; printVersion(); CLEAN_RETURN(0); /* Version Only */
821 case 'H':
822 case 'h': g_displayOut=stdout; CLEAN_RETURN(usage_advanced(programName));
823
824 /* Compress */
825 case 'z': operation=zom_compress; argument++; break;
826
827 /* Decoding */
828 case 'd':
829 #ifndef ZSTD_NOBENCH
830 benchParams.mode = BMK_decodeOnly;
831 if (operation==zom_bench) { argument++; break; } /* benchmark decode (hidden option) */
832 #endif
833 operation=zom_decompress; argument++; break;
834
835 /* Force stdout, even if stdout==console */
836 case 'c': forceStdout=1; outFileName=stdoutmark; argument++; break;
837
838 /* Use file content as dictionary */
839 case 'D': nextEntryIsDictionary = 1; lastCommand = 1; argument++; break;
840
841 /* Overwrite */
842 case 'f': FIO_overwriteMode(prefs); forceStdout=1; followLinks=1; argument++; break;
843
844 /* Verbose mode */
845 case 'v': g_displayLevel++; argument++; break;
846
847 /* Quiet mode */
848 case 'q': g_displayLevel--; argument++; break;
849
850 /* keep source file (default) */
851 case 'k': FIO_setRemoveSrcFile(prefs, 0); argument++; break;
852
853 /* Checksum */
854 case 'C': FIO_setChecksumFlag(prefs, 2); argument++; break;
855
856 /* test compressed file */
857 case 't': operation=zom_test; argument++; break;
858
859 /* destination file name */
860 case 'o': nextArgumentIsOutFileName=1; lastCommand=1; argument++; break;
861
862 /* limit decompression memory */
863 case 'M':
864 argument++;
865 memLimit = readU32FromChar(&argument);
866 break;
867 case 'l': operation=zom_list; argument++; break;
868 #ifdef UTIL_HAS_CREATEFILELIST
869 /* recursive */
870 case 'r': recursive=1; argument++; break;
871 #endif
872
873 #ifndef ZSTD_NOBENCH
874 /* Benchmark */
875 case 'b':
876 operation=zom_bench;
877 argument++;
878 break;
879
880 /* range bench (benchmark only) */
881 case 'e':
882 /* compression Level */
883 argument++;
884 cLevelLast = (int)readU32FromChar(&argument);
885 break;
886
887 /* Modify Nb Iterations (benchmark only) */
888 case 'i':
889 argument++;
890 bench_nbSeconds = readU32FromChar(&argument);
891 break;
892
893 /* cut input into blocks (benchmark only) */
894 case 'B':
895 argument++;
896 blockSize = readU32FromChar(&argument);
897 break;
898
899 /* benchmark files separately (hidden option) */
900 case 'S':
901 argument++;
902 separateFiles = 1;
903 break;
904
905 #endif /* ZSTD_NOBENCH */
906
907 /* nb of threads (hidden option) */
908 case 'T':
909 argument++;
910 nbWorkers = (int)readU32FromChar(&argument);
911 break;
912
913 /* Dictionary Selection level */
914 case 's':
915 argument++;
916 dictSelect = readU32FromChar(&argument);
917 break;
918
919 /* Pause at the end (-p) or set an additional param (-p#) (hidden option) */
920 case 'p': argument++;
921 #ifndef ZSTD_NOBENCH
922 if ((*argument>='0') && (*argument<='9')) {
923 benchParams.additionalParam = (int)readU32FromChar(&argument);
924 } else
925 #endif
926 main_pause=1;
927 break;
928
929 /* Select compressibility of synthetic sample */
930 case 'P':
931 { argument++;
932 compressibility = (double)readU32FromChar(&argument) / 100;
933 }
934 break;
935
936 /* unknown command */
937 default : CLEAN_RETURN(badusage(programName));
938 }
939 }
940 continue;
941 } /* if (argument[0]=='-') */
942
943 if (nextArgumentIsMaxDict) { /* kept available for compatibility with old syntax ; will be removed one day */
944 nextArgumentIsMaxDict = 0;
945 lastCommand = 0;
946 maxDictSize = readU32FromChar(&argument);
947 continue;
948 }
949
950 if (nextArgumentIsDictID) { /* kept available for compatibility with old syntax ; will be removed one day */
951 nextArgumentIsDictID = 0;
952 lastCommand = 0;
953 dictID = readU32FromChar(&argument);
954 continue;
955 }
956
957 } /* if (nextArgumentIsAFile==0) */
958
959 if (nextEntryIsDictionary) {
960 nextEntryIsDictionary = 0;
961 lastCommand = 0;
962 dictFileName = argument;
963 continue;
964 }
965
966 if (nextArgumentIsOutFileName) {
967 nextArgumentIsOutFileName = 0;
968 lastCommand = 0;
969 outFileName = argument;
970 if (!strcmp(outFileName, "-")) outFileName = stdoutmark;
971 continue;
972 }
973
974 if (nextArgumentIsOutDirName) {
975 nextArgumentIsOutDirName = 0;
976 lastCommand = 0;
977 outDirName = argument;
978 continue;
979 }
980
981 /* add filename to list */
982 filenameTable[filenameIdx++] = argument;
983 }
984
985 if (lastCommand) { /* forgotten argument */
986 DISPLAY("error : command must be followed by argument \n");
987 CLEAN_RETURN(1);
988 }
989
990 /* Welcome message (if verbose) */
991 DISPLAYLEVEL(3, WELCOME_MESSAGE);
992
993 #ifdef ZSTD_MULTITHREAD
994 if ((nbWorkers==0) && (!singleThread)) {
995 /* automatically set # workers based on # of reported cpus */
996 nbWorkers = UTIL_countPhysicalCores();
997 DISPLAYLEVEL(3, "Note: %d physical core(s) detected \n", nbWorkers);
998 }
999 #else
1000 (void)singleThread; (void)nbWorkers;
1001 #endif
1002
1003 #ifdef UTIL_HAS_CREATEFILELIST
1004 g_utilDisplayLevel = g_displayLevel;
1005 if (!followLinks) {
1006 unsigned u;
1007 for (u=0, fileNamesNb=0; u<filenameIdx; u++) {
1008 if (UTIL_isLink(filenameTable[u])
1009 #ifndef _MSC_VER
1010 && !UTIL_isFIFO(filenameTable[u])
1011 #endif /* _MSC_VER */
1012 ) {
1013 DISPLAYLEVEL(2, "Warning : %s is a symbolic link, ignoring\n", filenameTable[u]);
1014 } else {
1015 filenameTable[fileNamesNb++] = filenameTable[u];
1016 }
1017 }
1018 if (fileNamesNb == 0 && filenameIdx > 0)
1019 CLEAN_RETURN(1);
1020 filenameIdx = fileNamesNb;
1021 }
1022 if (recursive) { /* at this stage, filenameTable is a list of paths, which can contain both files and directories */
1023 extendedFileList = UTIL_createFileList(filenameTable, filenameIdx, &fileNamesBuf, &fileNamesNb, followLinks);
1024 if (extendedFileList) {
1025 unsigned u;
1026 for (u=0; u<fileNamesNb; u++) DISPLAYLEVEL(4, "%u %s\n", u, extendedFileList[u]);
1027 free((void*)filenameTable);
1028 filenameTable = extendedFileList;
1029 filenameIdx = fileNamesNb;
1030 }
1031 }
1032 #else
1033 (void)followLinks;
1034 #endif
1035
1036 if (operation == zom_list) {
1037 #ifndef ZSTD_NODECOMPRESS
1038 int const ret = FIO_listMultipleFiles(filenameIdx, filenameTable, g_displayLevel);
1039 CLEAN_RETURN(ret);
1040 #else
1041 DISPLAY("file information is not supported \n");
1042 CLEAN_RETURN(1);
1043 #endif
1044 }
1045
1046 /* Check if benchmark is selected */
1047 if (operation==zom_bench) {
1048 #ifndef ZSTD_NOBENCH
1049 benchParams.blockSize = blockSize;
1050 benchParams.nbWorkers = nbWorkers;
1051 benchParams.realTime = (unsigned)setRealTimePrio;
1052 benchParams.nbSeconds = bench_nbSeconds;
1053 benchParams.ldmFlag = ldmFlag;
1054 benchParams.ldmMinMatch = (int)g_ldmMinMatch;
1055 benchParams.ldmHashLog = (int)g_ldmHashLog;
1056 if (g_ldmBucketSizeLog != LDM_PARAM_DEFAULT) {
1057 benchParams.ldmBucketSizeLog = (int)g_ldmBucketSizeLog;
1058 }
1059 if (g_ldmHashRateLog != LDM_PARAM_DEFAULT) {
1060 benchParams.ldmHashRateLog = (int)g_ldmHashRateLog;
1061 }
1062 benchParams.literalCompressionMode = literalCompressionMode;
1063
1064 if (cLevel > ZSTD_maxCLevel()) cLevel = ZSTD_maxCLevel();
1065 if (cLevelLast > ZSTD_maxCLevel()) cLevelLast = ZSTD_maxCLevel();
1066 if (cLevelLast < cLevel) cLevelLast = cLevel;
1067 if (cLevelLast > cLevel)
1068 DISPLAYLEVEL(3, "Benchmarking levels from %d to %d\n", cLevel, cLevelLast);
1069 if(filenameIdx) {
1070 if(separateFiles) {
1071 unsigned i;
1072 for(i = 0; i < filenameIdx; i++) {
1073 int c;
1074 DISPLAYLEVEL(3, "Benchmarking %s \n", filenameTable[i]);
1075 for(c = cLevel; c <= cLevelLast; c++) {
1076 BMK_benchFilesAdvanced(&filenameTable[i], 1, dictFileName, c, &compressionParams, g_displayLevel, &benchParams);
1077 }
1078 }
1079 } else {
1080 for(; cLevel <= cLevelLast; cLevel++) {
1081 BMK_benchFilesAdvanced(filenameTable, filenameIdx, dictFileName, cLevel, &compressionParams, g_displayLevel, &benchParams);
1082 }
1083 }
1084 } else {
1085 for(; cLevel <= cLevelLast; cLevel++) {
1086 BMK_syntheticTest(cLevel, compressibility, &compressionParams, g_displayLevel, &benchParams);
1087 }
1088 }
1089
1090 #else
1091 (void)bench_nbSeconds; (void)blockSize; (void)setRealTimePrio; (void)separateFiles; (void)compressibility;
1092 #endif
1093 goto _end;
1094 }
1095
1096 /* Check if dictionary builder is selected */
1097 if (operation==zom_train) {
1098 #ifndef ZSTD_NODICT
1099 ZDICT_params_t zParams;
1100 zParams.compressionLevel = dictCLevel;
1101 zParams.notificationLevel = (unsigned)g_displayLevel;
1102 zParams.dictID = dictID;
1103 if (dict == cover) {
1104 int const optimize = !coverParams.k || !coverParams.d;
1105 coverParams.nbThreads = (unsigned)nbWorkers;
1106 coverParams.zParams = zParams;
1107 operationResult = DiB_trainFromFiles(outFileName, maxDictSize, filenameTable, filenameIdx, blockSize, NULL, &coverParams, NULL, optimize);
1108 } else if (dict == fastCover) {
1109 int const optimize = !fastCoverParams.k || !fastCoverParams.d;
1110 fastCoverParams.nbThreads = (unsigned)nbWorkers;
1111 fastCoverParams.zParams = zParams;
1112 operationResult = DiB_trainFromFiles(outFileName, maxDictSize, filenameTable, filenameIdx, blockSize, NULL, NULL, &fastCoverParams, optimize);
1113 } else {
1114 ZDICT_legacy_params_t dictParams;
1115 memset(&dictParams, 0, sizeof(dictParams));
1116 dictParams.selectivityLevel = dictSelect;
1117 dictParams.zParams = zParams;
1118 operationResult = DiB_trainFromFiles(outFileName, maxDictSize, filenameTable, filenameIdx, blockSize, &dictParams, NULL, NULL, 0);
1119 }
1120 #else
1121 (void)dictCLevel; (void)dictSelect; (void)dictID; (void)maxDictSize; /* not used when ZSTD_NODICT set */
1122 DISPLAYLEVEL(1, "training mode not available \n");
1123 operationResult = 1;
1124 #endif
1125 goto _end;
1126 }
1127
1128 #ifndef ZSTD_NODECOMPRESS
1129 if (operation==zom_test) { FIO_setTestMode(prefs, 1); outFileName=nulmark; FIO_setRemoveSrcFile(prefs, 0); } /* test mode */
1130 #endif
1131
1132 /* No input filename ==> use stdin and stdout */
1133 filenameIdx += !filenameIdx; /* filenameTable[0] is stdin by default */
1134 if (!strcmp(filenameTable[0], stdinmark) && !outFileName)
1135 outFileName = stdoutmark; /* when input is stdin, default output is stdout */
1136
1137 /* Check if input/output defined as console; trigger an error in this case */
1138 if (!strcmp(filenameTable[0], stdinmark) && IS_CONSOLE(stdin) )
1139 CLEAN_RETURN(badusage(programName));
1140 if ( outFileName && !strcmp(outFileName, stdoutmark)
1141 && IS_CONSOLE(stdout)
1142 && !strcmp(filenameTable[0], stdinmark)
1143 && !forceStdout
1144 && operation!=zom_decompress )
1145 CLEAN_RETURN(badusage(programName));
1146
1147 #ifndef ZSTD_NOCOMPRESS
1148 /* check compression level limits */
1149 { int const maxCLevel = ultra ? ZSTD_maxCLevel() : ZSTDCLI_CLEVEL_MAX;
1150 if (cLevel > maxCLevel) {
1151 DISPLAYLEVEL(2, "Warning : compression level higher than max, reduced to %i \n", maxCLevel);
1152 cLevel = maxCLevel;
1153 } }
1154 #endif
1155
1156 /* No status message in pipe mode (stdin - stdout) or multi-files mode */
1157 if (!strcmp(filenameTable[0], stdinmark) && outFileName && !strcmp(outFileName,stdoutmark) && (g_displayLevel==2)) g_displayLevel=1;
1158 if ((filenameIdx>1) & (g_displayLevel==2)) g_displayLevel=1;
1159
1160 /* IO Stream/File */
1161 FIO_setNotificationLevel(g_displayLevel);
1162 if (operation==zom_compress) {
1163 #ifndef ZSTD_NOCOMPRESS
1164 FIO_setNbWorkers(prefs, nbWorkers);
1165 FIO_setBlockSize(prefs, (int)blockSize);
1166 if (g_overlapLog!=OVERLAP_LOG_DEFAULT) FIO_setOverlapLog(prefs, (int)g_overlapLog);
1167 FIO_setLdmFlag(prefs, (unsigned)ldmFlag);
1168 FIO_setLdmHashLog(prefs, (int)g_ldmHashLog);
1169 FIO_setLdmMinMatch(prefs, (int)g_ldmMinMatch);
1170 if (g_ldmBucketSizeLog != LDM_PARAM_DEFAULT) FIO_setLdmBucketSizeLog(prefs, (int)g_ldmBucketSizeLog);
1171 if (g_ldmHashRateLog != LDM_PARAM_DEFAULT) FIO_setLdmHashRateLog(prefs, (int)g_ldmHashRateLog);
1172 FIO_setAdaptiveMode(prefs, (unsigned)adapt);
1173 FIO_setAdaptMin(prefs, adaptMin);
1174 FIO_setAdaptMax(prefs, adaptMax);
1175 FIO_setRsyncable(prefs, rsyncable);
1176 FIO_setStreamSrcSize(prefs, streamSrcSize);
1177 FIO_setTargetCBlockSize(prefs, targetCBlockSize);
1178 FIO_setSrcSizeHint(prefs, srcSizeHint);
1179 FIO_setLiteralCompressionMode(prefs, literalCompressionMode);
1180 if (adaptMin > cLevel) cLevel = adaptMin;
1181 if (adaptMax < cLevel) cLevel = adaptMax;
1182
1183 if ((filenameIdx==1) && outFileName)
1184 operationResult = FIO_compressFilename(prefs, outFileName, filenameTable[0], dictFileName, cLevel, compressionParams);
1185 else
1186 operationResult = FIO_compressMultipleFilenames(prefs, filenameTable, filenameIdx, outDirName, outFileName, suffix, dictFileName, cLevel, compressionParams);
1187 #else
1188 (void)suffix; (void)adapt; (void)rsyncable; (void)ultra; (void)cLevel; (void)ldmFlag; (void)literalCompressionMode; (void)targetCBlockSize; (void)streamSrcSize; (void)srcSizeHint; /* not used when ZSTD_NOCOMPRESS set */
1189 DISPLAY("Compression not supported \n");
1190 #endif
1191 } else { /* decompression or test */
1192 #ifndef ZSTD_NODECOMPRESS
1193 if (memLimit == 0) {
1194 if (compressionParams.windowLog == 0)
1195 memLimit = (U32)1 << g_defaultMaxWindowLog;
1196 else {
1197 memLimit = (U32)1 << (compressionParams.windowLog & 31);
1198 }
1199 }
1200 FIO_setMemLimit(prefs, memLimit);
1201 if (filenameIdx==1 && outFileName)
1202 operationResult = FIO_decompressFilename(prefs, outFileName, filenameTable[0], dictFileName);
1203 else
1204 operationResult = FIO_decompressMultipleFilenames(prefs, filenameTable, filenameIdx, outDirName, outFileName, dictFileName);
1205 #else
1206 DISPLAY("Decompression not supported \n");
1207 #endif
1208 }
1209
1210 _end:
1211 FIO_freePreferences(prefs);
1212
1213 if (main_pause) waitEnter();
1214 #ifdef UTIL_HAS_CREATEFILELIST
1215 if (extendedFileList)
1216 UTIL_freeFileList(extendedFileList, fileNamesBuf);
1217 else
1218 #endif
1219 free((void*)filenameTable);
1220 return operationResult;
1221 }
1222