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