1 /*
2 LZ4io.c - LZ4 File/Stream Interface
3 Copyright (C) Yann Collet 2011-2017
4
5 GPL v2 License
6
7 This program is free software; you can redistribute it and/or modify
8 it under the terms of the GNU General Public License as published by
9 the Free Software Foundation; either version 2 of the License, or
10 (at your option) any later version.
11
12 This program is distributed in the hope that it will be useful,
13 but WITHOUT ANY WARRANTY; without even the implied warranty of
14 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 GNU General Public License for more details.
16
17 You should have received a copy of the GNU General Public License along
18 with this program; if not, write to the Free Software Foundation, Inc.,
19 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
20
21 You can contact the author at :
22 - LZ4 source repository : https://github.com/lz4/lz4
23 - LZ4 public forum : https://groups.google.com/forum/#!forum/lz4c
24 */
25 /*
26 Note : this is stand-alone program.
27 It is not part of LZ4 compression library, it is a user code of the LZ4 library.
28 - The license of LZ4 library is BSD.
29 - The license of xxHash library is BSD.
30 - The license of this source file is GPLv2.
31 */
32
33
34 /*-************************************
35 * Compiler options
36 **************************************/
37 #ifdef _MSC_VER /* Visual Studio */
38 # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */
39 #endif
40 #if defined(__MINGW32__) && !defined(_POSIX_SOURCE)
41 # define _POSIX_SOURCE 1 /* disable %llu warnings with MinGW on Windows */
42 #endif
43
44
45 /*****************************
46 * Includes
47 *****************************/
48 #include "platform.h" /* Large File Support, SET_BINARY_MODE, SET_SPARSE_FILE_MODE, PLATFORM_POSIX_VERSION, __64BIT__ */
49 #include "util.h" /* UTIL_getFileStat, UTIL_setFileStat */
50 #include <stdio.h> /* fprintf, fopen, fread, stdin, stdout, fflush, getchar */
51 #include <stdlib.h> /* malloc, free */
52 #include <string.h> /* strerror, strcmp, strlen */
53 #include <time.h> /* clock */
54 #include <sys/types.h> /* stat64 */
55 #include <sys/stat.h> /* stat64 */
56 #include "lz4.h" /* still required for legacy format */
57 #include "lz4hc.h" /* still required for legacy format */
58 #define LZ4F_STATIC_LINKING_ONLY
59 #include "lz4frame.h"
60 #include "lz4io.h"
61
62
63 /*****************************
64 * Constants
65 *****************************/
66 #define KB *(1 <<10)
67 #define MB *(1 <<20)
68 #define GB *(1U<<30)
69
70 #define _1BIT 0x01
71 #define _2BITS 0x03
72 #define _3BITS 0x07
73 #define _4BITS 0x0F
74 #define _8BITS 0xFF
75
76 #define MAGICNUMBER_SIZE 4
77 #define LZ4IO_MAGICNUMBER 0x184D2204
78 #define LZ4IO_SKIPPABLE0 0x184D2A50
79 #define LZ4IO_SKIPPABLEMASK 0xFFFFFFF0
80 #define LEGACY_MAGICNUMBER 0x184C2102
81
82 #define CACHELINE 64
83 #define LEGACY_BLOCKSIZE (8 MB)
84 #define MIN_STREAM_BUFSIZE (192 KB)
85 #define LZ4IO_BLOCKSIZEID_DEFAULT 7
86 #define LZ4_MAX_DICT_SIZE (64 KB)
87
88
89 /**************************************
90 * Macros
91 **************************************/
92 #define DISPLAY(...) fprintf(stderr, __VA_ARGS__)
93 #define DISPLAYOUT(...) fprintf(stdout, __VA_ARGS__)
94 #define DISPLAYLEVEL(l, ...) if (g_displayLevel>=l) { DISPLAY(__VA_ARGS__); }
95 static int g_displayLevel = 0; /* 0 : no display ; 1: errors ; 2 : + result + interaction + warnings ; 3 : + progression; 4 : + information */
96
97 #define DISPLAYUPDATE(l, ...) if (g_displayLevel>=l) { \
98 if ( ((clock() - g_time) > refreshRate) \
99 || (g_displayLevel>=4) ) { \
100 g_time = clock(); \
101 DISPLAY(__VA_ARGS__); \
102 if (g_displayLevel>=4) fflush(stderr); \
103 } }
104 static const clock_t refreshRate = CLOCKS_PER_SEC / 6;
105 static clock_t g_time = 0;
106 #define LZ4IO_STATIC_ASSERT(c) { enum { LZ4IO_static_assert = 1/(int)(!!(c)) }; } /* use after variable declarations */
107
108
109 /**************************************
110 * Local Parameters
111 **************************************/
112
113 struct LZ4IO_prefs_s {
114 int passThrough;
115 int overwrite;
116 int testMode;
117 int blockSizeId;
118 size_t blockSize;
119 int blockChecksum;
120 int streamChecksum;
121 int blockIndependence;
122 int sparseFileSupport;
123 int contentSizeFlag;
124 int useDictionary;
125 unsigned favorDecSpeed;
126 const char* dictionaryFilename;
127 int removeSrcFile;
128 };
129
130 /**************************************
131 * Exceptions
132 ***************************************/
133 #ifndef DEBUG
134 # define DEBUG 0
135 #endif
136 #define DEBUGOUTPUT(...) if (DEBUG) DISPLAY(__VA_ARGS__);
137 #define EXM_THROW(error, ...) \
138 { \
139 DEBUGOUTPUT("Error defined at %s, line %i : \n", __FILE__, __LINE__); \
140 DISPLAYLEVEL(1, "Error %i : ", error); \
141 DISPLAYLEVEL(1, __VA_ARGS__); \
142 DISPLAYLEVEL(1, " \n"); \
143 exit(error); \
144 }
145
146
147 /**************************************
148 * Version modifiers
149 **************************************/
150 #define EXTENDED_ARGUMENTS
151 #define EXTENDED_HELP
152 #define EXTENDED_FORMAT
153 #define DEFAULT_DECOMPRESSOR LZ4IO_decompressLZ4F
154
155
156 /* ************************************************** */
157 /* ****************** Parameters ******************** */
158 /* ************************************************** */
159
LZ4IO_defaultPreferences(void)160 LZ4IO_prefs_t* LZ4IO_defaultPreferences(void)
161 {
162 LZ4IO_prefs_t* const ret = (LZ4IO_prefs_t*)malloc(sizeof(*ret));
163 if (!ret) EXM_THROW(21, "Allocation error : not enough memory");
164 ret->passThrough = 0;
165 ret->overwrite = 1;
166 ret->testMode = 0;
167 ret->blockSizeId = LZ4IO_BLOCKSIZEID_DEFAULT;
168 ret->blockSize = 0;
169 ret->blockChecksum = 0;
170 ret->streamChecksum = 1;
171 ret->blockIndependence = 1;
172 ret->sparseFileSupport = 1;
173 ret->contentSizeFlag = 0;
174 ret->useDictionary = 0;
175 ret->favorDecSpeed = 0;
176 ret->dictionaryFilename = NULL;
177 ret->removeSrcFile = 0;
178 return ret;
179 }
180
LZ4IO_freePreferences(LZ4IO_prefs_t * prefs)181 void LZ4IO_freePreferences(LZ4IO_prefs_t* prefs)
182 {
183 free(prefs);
184 }
185
186
LZ4IO_setDictionaryFilename(LZ4IO_prefs_t * const prefs,const char * dictionaryFilename)187 int LZ4IO_setDictionaryFilename(LZ4IO_prefs_t* const prefs, const char* dictionaryFilename)
188 {
189 prefs->dictionaryFilename = dictionaryFilename;
190 prefs->useDictionary = dictionaryFilename != NULL;
191 return prefs->useDictionary;
192 }
193
194 /* Default setting : passThrough = 0; return : passThrough mode (0/1) */
LZ4IO_setPassThrough(LZ4IO_prefs_t * const prefs,int yes)195 int LZ4IO_setPassThrough(LZ4IO_prefs_t* const prefs, int yes)
196 {
197 prefs->passThrough = (yes!=0);
198 return prefs->passThrough;
199 }
200
201
202 /* Default setting : overwrite = 1; return : overwrite mode (0/1) */
LZ4IO_setOverwrite(LZ4IO_prefs_t * const prefs,int yes)203 int LZ4IO_setOverwrite(LZ4IO_prefs_t* const prefs, int yes)
204 {
205 prefs->overwrite = (yes!=0);
206 return prefs->overwrite;
207 }
208
209 /* Default setting : testMode = 0; return : testMode (0/1) */
LZ4IO_setTestMode(LZ4IO_prefs_t * const prefs,int yes)210 int LZ4IO_setTestMode(LZ4IO_prefs_t* const prefs, int yes)
211 {
212 prefs->testMode = (yes!=0);
213 return prefs->testMode;
214 }
215
216 /* blockSizeID : valid values : 4-5-6-7 */
LZ4IO_setBlockSizeID(LZ4IO_prefs_t * const prefs,unsigned bsid)217 size_t LZ4IO_setBlockSizeID(LZ4IO_prefs_t* const prefs, unsigned bsid)
218 {
219 static const size_t blockSizeTable[] = { 64 KB, 256 KB, 1 MB, 4 MB };
220 static const unsigned minBlockSizeID = 4;
221 static const unsigned maxBlockSizeID = 7;
222 if ((bsid < minBlockSizeID) || (bsid > maxBlockSizeID)) return 0;
223 prefs->blockSizeId = (int)bsid;
224 prefs->blockSize = blockSizeTable[(unsigned)prefs->blockSizeId-minBlockSizeID];
225 return prefs->blockSize;
226 }
227
LZ4IO_setBlockSize(LZ4IO_prefs_t * const prefs,size_t blockSize)228 size_t LZ4IO_setBlockSize(LZ4IO_prefs_t* const prefs, size_t blockSize)
229 {
230 static const size_t minBlockSize = 32;
231 static const size_t maxBlockSize = 4 MB;
232 unsigned bsid = 0;
233 if (blockSize < minBlockSize) blockSize = minBlockSize;
234 if (blockSize > maxBlockSize) blockSize = maxBlockSize;
235 prefs->blockSize = blockSize;
236 blockSize--;
237 /* find which of { 64k, 256k, 1MB, 4MB } is closest to blockSize */
238 while (blockSize >>= 2)
239 bsid++;
240 if (bsid < 7) bsid = 7;
241 prefs->blockSizeId = (int)(bsid-3);
242 return prefs->blockSize;
243 }
244
245 /* Default setting : 1 == independent blocks */
LZ4IO_setBlockMode(LZ4IO_prefs_t * const prefs,LZ4IO_blockMode_t blockMode)246 int LZ4IO_setBlockMode(LZ4IO_prefs_t* const prefs, LZ4IO_blockMode_t blockMode)
247 {
248 prefs->blockIndependence = (blockMode == LZ4IO_blockIndependent);
249 return prefs->blockIndependence;
250 }
251
252 /* Default setting : 0 == no block checksum */
LZ4IO_setBlockChecksumMode(LZ4IO_prefs_t * const prefs,int enable)253 int LZ4IO_setBlockChecksumMode(LZ4IO_prefs_t* const prefs, int enable)
254 {
255 prefs->blockChecksum = (enable != 0);
256 return prefs->blockChecksum;
257 }
258
259 /* Default setting : 1 == checksum enabled */
LZ4IO_setStreamChecksumMode(LZ4IO_prefs_t * const prefs,int enable)260 int LZ4IO_setStreamChecksumMode(LZ4IO_prefs_t* const prefs, int enable)
261 {
262 prefs->streamChecksum = (enable != 0);
263 return prefs->streamChecksum;
264 }
265
266 /* Default setting : 0 (no notification) */
LZ4IO_setNotificationLevel(int level)267 int LZ4IO_setNotificationLevel(int level)
268 {
269 g_displayLevel = level;
270 return g_displayLevel;
271 }
272
273 /* Default setting : 1 (auto: enabled on file, disabled on stdout) */
LZ4IO_setSparseFile(LZ4IO_prefs_t * const prefs,int enable)274 int LZ4IO_setSparseFile(LZ4IO_prefs_t* const prefs, int enable)
275 {
276 prefs->sparseFileSupport = 2*(enable!=0); /* 2==force enable */
277 return prefs->sparseFileSupport;
278 }
279
280 /* Default setting : 0 (disabled) */
LZ4IO_setContentSize(LZ4IO_prefs_t * const prefs,int enable)281 int LZ4IO_setContentSize(LZ4IO_prefs_t* const prefs, int enable)
282 {
283 prefs->contentSizeFlag = (enable!=0);
284 return prefs->contentSizeFlag;
285 }
286
287 /* Default setting : 0 (disabled) */
LZ4IO_favorDecSpeed(LZ4IO_prefs_t * const prefs,int favor)288 void LZ4IO_favorDecSpeed(LZ4IO_prefs_t* const prefs, int favor)
289 {
290 prefs->favorDecSpeed = (favor!=0);
291 }
292
LZ4IO_setRemoveSrcFile(LZ4IO_prefs_t * const prefs,unsigned flag)293 void LZ4IO_setRemoveSrcFile(LZ4IO_prefs_t* const prefs, unsigned flag)
294 {
295 prefs->removeSrcFile = (flag>0);
296 }
297
298
299
300 /* ************************************************************************ **
301 ** ********************** LZ4 File / Pipe compression ********************* **
302 ** ************************************************************************ */
303
LZ4IO_isSkippableMagicNumber(unsigned int magic)304 static int LZ4IO_isSkippableMagicNumber(unsigned int magic) {
305 return (magic & LZ4IO_SKIPPABLEMASK) == LZ4IO_SKIPPABLE0;
306 }
307
308
309 /** LZ4IO_openSrcFile() :
310 * condition : `srcFileName` must be non-NULL.
311 * @result : FILE* to `dstFileName`, or NULL if it fails */
LZ4IO_openSrcFile(const char * srcFileName)312 static FILE* LZ4IO_openSrcFile(const char* srcFileName)
313 {
314 FILE* f;
315
316 if (!strcmp (srcFileName, stdinmark)) {
317 DISPLAYLEVEL(4,"Using stdin for input\n");
318 f = stdin;
319 SET_BINARY_MODE(stdin);
320 } else {
321 f = fopen(srcFileName, "rb");
322 if ( f==NULL ) DISPLAYLEVEL(1, "%s: %s \n", srcFileName, strerror(errno));
323 }
324
325 return f;
326 }
327
328 /** FIO_openDstFile() :
329 * prefs is writable, because sparseFileSupport might be updated.
330 * condition : `dstFileName` must be non-NULL.
331 * @result : FILE* to `dstFileName`, or NULL if it fails */
LZ4IO_openDstFile(const char * dstFileName,const LZ4IO_prefs_t * const prefs)332 static FILE* LZ4IO_openDstFile(const char* dstFileName, const LZ4IO_prefs_t* const prefs)
333 {
334 FILE* f;
335 assert(dstFileName != NULL);
336
337 if (!strcmp (dstFileName, stdoutmark)) {
338 DISPLAYLEVEL(4, "Using stdout for output \n");
339 f = stdout;
340 SET_BINARY_MODE(stdout);
341 if (prefs->sparseFileSupport==1) {
342 DISPLAYLEVEL(4, "Sparse File Support automatically disabled on stdout ;"
343 " to force-enable it, add --sparse command \n");
344 }
345 } else {
346 if (!prefs->overwrite && strcmp (dstFileName, nulmark)) { /* Check if destination file already exists */
347 FILE* const testf = fopen( dstFileName, "rb" );
348 if (testf != NULL) { /* dest exists, prompt for overwrite authorization */
349 fclose(testf);
350 if (g_displayLevel <= 1) { /* No interaction possible */
351 DISPLAY("%s already exists; not overwritten \n", dstFileName);
352 return NULL;
353 }
354 DISPLAY("%s already exists; do you wish to overwrite (y/N) ? ", dstFileName);
355 { int ch = getchar();
356 if ((ch!='Y') && (ch!='y')) {
357 DISPLAY(" not overwritten \n");
358 return NULL;
359 }
360 while ((ch!=EOF) && (ch!='\n')) ch = getchar(); /* flush rest of input line */
361 } } }
362 f = fopen( dstFileName, "wb" );
363 if (f==NULL) DISPLAYLEVEL(1, "%s: %s\n", dstFileName, strerror(errno));
364 }
365
366 /* sparse file */
367 { int const sparseMode = (prefs->sparseFileSupport - (f==stdout)) > 0;
368 if (f && sparseMode) { SET_SPARSE_FILE_MODE(f); }
369 }
370
371 return f;
372 }
373
374
375
376 /***************************************
377 * Legacy Compression
378 ***************************************/
379
380 /* unoptimized version; solves endianess & alignment issues */
LZ4IO_writeLE32(void * p,unsigned value32)381 static void LZ4IO_writeLE32 (void* p, unsigned value32)
382 {
383 unsigned char* const dstPtr = (unsigned char*)p;
384 dstPtr[0] = (unsigned char)value32;
385 dstPtr[1] = (unsigned char)(value32 >> 8);
386 dstPtr[2] = (unsigned char)(value32 >> 16);
387 dstPtr[3] = (unsigned char)(value32 >> 24);
388 }
389
LZ4IO_LZ4_compress(const char * src,char * dst,int srcSize,int dstSize,int cLevel)390 static int LZ4IO_LZ4_compress(const char* src, char* dst, int srcSize, int dstSize, int cLevel)
391 {
392 (void)cLevel;
393 return LZ4_compress_fast(src, dst, srcSize, dstSize, 1);
394 }
395
396 /* LZ4IO_compressFilename_Legacy :
397 * This function is intentionally "hidden" (not published in .h)
398 * It generates compressed streams using the old 'legacy' format */
LZ4IO_compressFilename_Legacy(const char * input_filename,const char * output_filename,int compressionlevel,const LZ4IO_prefs_t * prefs)399 int LZ4IO_compressFilename_Legacy(const char* input_filename, const char* output_filename,
400 int compressionlevel, const LZ4IO_prefs_t* prefs)
401 {
402 typedef int (*compress_f)(const char* src, char* dst, int srcSize, int dstSize, int cLevel);
403 compress_f const compressionFunction = (compressionlevel < 3) ? LZ4IO_LZ4_compress : LZ4_compress_HC;
404 unsigned long long filesize = 0;
405 unsigned long long compressedfilesize = MAGICNUMBER_SIZE;
406 char* in_buff;
407 char* out_buff;
408 const int outBuffSize = LZ4_compressBound(LEGACY_BLOCKSIZE);
409 FILE* const finput = LZ4IO_openSrcFile(input_filename);
410 FILE* foutput;
411 clock_t clockEnd;
412
413 /* Init */
414 clock_t const clockStart = clock();
415 if (finput == NULL)
416 EXM_THROW(20, "%s : open file error ", input_filename);
417
418 foutput = LZ4IO_openDstFile(output_filename, prefs);
419 if (foutput == NULL) {
420 fclose(finput);
421 EXM_THROW(20, "%s : open file error ", input_filename);
422 }
423
424 /* Allocate Memory */
425 in_buff = (char*)malloc(LEGACY_BLOCKSIZE);
426 out_buff = (char*)malloc((size_t)outBuffSize + 4);
427 if (!in_buff || !out_buff)
428 EXM_THROW(21, "Allocation error : not enough memory");
429
430 /* Write Archive Header */
431 LZ4IO_writeLE32(out_buff, LEGACY_MAGICNUMBER);
432 if (fwrite(out_buff, 1, MAGICNUMBER_SIZE, foutput) != MAGICNUMBER_SIZE)
433 EXM_THROW(22, "Write error : cannot write header");
434
435 /* Main Loop */
436 while (1) {
437 int outSize;
438 /* Read Block */
439 size_t const inSize = fread(in_buff, (size_t)1, (size_t)LEGACY_BLOCKSIZE, finput);
440 if (inSize == 0) break;
441 assert(inSize <= LEGACY_BLOCKSIZE);
442 filesize += inSize;
443
444 /* Compress Block */
445 outSize = compressionFunction(in_buff, out_buff+4, (int)inSize, outBuffSize, compressionlevel);
446 assert(outSize >= 0);
447 compressedfilesize += (unsigned long long)outSize+4;
448 DISPLAYUPDATE(2, "\rRead : %i MB ==> %.2f%% ",
449 (int)(filesize>>20), (double)compressedfilesize/filesize*100);
450
451 /* Write Block */
452 assert(outSize > 0);
453 assert(outSize < outBuffSize);
454 LZ4IO_writeLE32(out_buff, (unsigned)outSize);
455 if (fwrite(out_buff, 1, (size_t)outSize+4, foutput) != (size_t)(outSize+4)) {
456 EXM_THROW(24, "Write error : cannot write compressed block");
457 } }
458 if (ferror(finput)) EXM_THROW(25, "Error while reading %s ", input_filename);
459
460 /* Status */
461 clockEnd = clock();
462 if (clockEnd==clockStart) clockEnd+=1; /* avoid division by zero (speed) */
463 filesize += !filesize; /* avoid division by zero (ratio) */
464 DISPLAYLEVEL(2, "\r%79s\r", ""); /* blank line */
465 DISPLAYLEVEL(2,"Compressed %llu bytes into %llu bytes ==> %.2f%%\n",
466 filesize, compressedfilesize, (double)compressedfilesize / filesize * 100);
467 { double const seconds = (double)(clockEnd - clockStart) / CLOCKS_PER_SEC;
468 DISPLAYLEVEL(4,"Done in %.2f s ==> %.2f MB/s\n", seconds,
469 (double)filesize / seconds / 1024 / 1024);
470 }
471
472 /* Close & Free */
473 free(in_buff);
474 free(out_buff);
475 fclose(finput);
476 if (strcmp(output_filename,stdoutmark)) fclose(foutput); /* do not close stdout */
477
478 return 0;
479 }
480
481 #define FNSPACE 30
482 /* LZ4IO_compressMultipleFilenames_Legacy :
483 * This function is intentionally "hidden" (not published in .h)
484 * It generates multiple compressed streams using the old 'legacy' format */
LZ4IO_compressMultipleFilenames_Legacy(const char ** inFileNamesTable,int ifntSize,const char * suffix,int compressionLevel,const LZ4IO_prefs_t * prefs)485 int LZ4IO_compressMultipleFilenames_Legacy(
486 const char** inFileNamesTable, int ifntSize,
487 const char* suffix,
488 int compressionLevel, const LZ4IO_prefs_t* prefs)
489 {
490 int i;
491 int missed_files = 0;
492 char* dstFileName = (char*)malloc(FNSPACE);
493 size_t ofnSize = FNSPACE;
494 const size_t suffixSize = strlen(suffix);
495
496 if (dstFileName == NULL) return ifntSize; /* not enough memory */
497
498 /* loop on each file */
499 for (i=0; i<ifntSize; i++) {
500 size_t const ifnSize = strlen(inFileNamesTable[i]);
501 if (!strcmp(suffix, stdoutmark)) {
502 missed_files += LZ4IO_compressFilename_Legacy(
503 inFileNamesTable[i], stdoutmark,
504 compressionLevel, prefs);
505 continue;
506 }
507
508 if (ofnSize <= ifnSize+suffixSize+1) {
509 free(dstFileName);
510 ofnSize = ifnSize + 20;
511 dstFileName = (char*)malloc(ofnSize);
512 if (dstFileName==NULL) {
513 return ifntSize;
514 } }
515 strcpy(dstFileName, inFileNamesTable[i]);
516 strcat(dstFileName, suffix);
517
518 missed_files += LZ4IO_compressFilename_Legacy(
519 inFileNamesTable[i], dstFileName,
520 compressionLevel, prefs);
521 }
522
523 /* Close & Free */
524 free(dstFileName);
525
526 return missed_files;
527 }
528
529
530 /*********************************************
531 * Compression using Frame format
532 *********************************************/
533
534 typedef struct {
535 void* srcBuffer;
536 size_t srcBufferSize;
537 void* dstBuffer;
538 size_t dstBufferSize;
539 LZ4F_compressionContext_t ctx;
540 LZ4F_CDict* cdict;
541 } cRess_t;
542
LZ4IO_createDict(size_t * dictSize,const char * const dictFilename)543 static void* LZ4IO_createDict(size_t* dictSize, const char* const dictFilename)
544 {
545 size_t readSize;
546 size_t dictEnd = 0;
547 size_t dictLen = 0;
548 size_t dictStart;
549 size_t circularBufSize = LZ4_MAX_DICT_SIZE;
550 char* circularBuf = (char*)malloc(circularBufSize);
551 char* dictBuf;
552 FILE* dictFile;
553
554 if (!circularBuf) EXM_THROW(25, "Allocation error : not enough memory for circular buffer");
555 if (!dictFilename) EXM_THROW(25, "Dictionary error : no filename provided");
556
557 dictFile = LZ4IO_openSrcFile(dictFilename);
558 if (!dictFile) EXM_THROW(25, "Dictionary error : could not open dictionary file");
559
560 /* opportunistically seek to the part of the file we care about. If this */
561 /* fails it's not a problem since we'll just read everything anyways. */
562 if (strcmp(dictFilename, stdinmark)) {
563 (void)UTIL_fseek(dictFile, -LZ4_MAX_DICT_SIZE, SEEK_END);
564 }
565
566 do {
567 readSize = fread(circularBuf + dictEnd, 1, circularBufSize - dictEnd, dictFile);
568 dictEnd = (dictEnd + readSize) % circularBufSize;
569 dictLen += readSize;
570 } while (readSize>0);
571
572 if (dictLen > LZ4_MAX_DICT_SIZE) {
573 dictLen = LZ4_MAX_DICT_SIZE;
574 }
575
576 *dictSize = dictLen;
577
578 dictStart = (circularBufSize + dictEnd - dictLen) % circularBufSize;
579
580 if (dictStart == 0) {
581 /* We're in the simple case where the dict starts at the beginning of our circular buffer. */
582 dictBuf = circularBuf;
583 circularBuf = NULL;
584 } else {
585 /* Otherwise, we will alloc a new buffer and copy our dict into that. */
586 dictBuf = (char *)malloc(dictLen ? dictLen : 1);
587 if (!dictBuf) EXM_THROW(25, "Allocation error : not enough memory");
588
589 memcpy(dictBuf, circularBuf + dictStart, circularBufSize - dictStart);
590 memcpy(dictBuf + circularBufSize - dictStart, circularBuf, dictLen - (circularBufSize - dictStart));
591 }
592
593 fclose(dictFile);
594 free(circularBuf);
595
596 return dictBuf;
597 }
598
LZ4IO_createCDict(const LZ4IO_prefs_t * const prefs)599 static LZ4F_CDict* LZ4IO_createCDict(const LZ4IO_prefs_t* const prefs)
600 {
601 size_t dictionarySize;
602 void* dictionaryBuffer;
603 LZ4F_CDict* cdict;
604 if (!prefs->useDictionary) return NULL;
605 dictionaryBuffer = LZ4IO_createDict(&dictionarySize, prefs->dictionaryFilename);
606 if (!dictionaryBuffer) EXM_THROW(25, "Dictionary error : could not create dictionary");
607 cdict = LZ4F_createCDict(dictionaryBuffer, dictionarySize);
608 free(dictionaryBuffer);
609 return cdict;
610 }
611
LZ4IO_createCResources(const LZ4IO_prefs_t * const prefs)612 static cRess_t LZ4IO_createCResources(const LZ4IO_prefs_t* const prefs)
613 {
614 const size_t blockSize = prefs->blockSize;
615 cRess_t ress;
616
617 LZ4F_errorCode_t const errorCode = LZ4F_createCompressionContext(&(ress.ctx), LZ4F_VERSION);
618 if (LZ4F_isError(errorCode)) EXM_THROW(30, "Allocation error : can't create LZ4F context : %s", LZ4F_getErrorName(errorCode));
619
620 /* Allocate Memory */
621 ress.srcBuffer = malloc(blockSize);
622 ress.srcBufferSize = blockSize;
623 ress.dstBufferSize = LZ4F_compressFrameBound(blockSize, NULL); /* cover worst case */
624 ress.dstBuffer = malloc(ress.dstBufferSize);
625 if (!ress.srcBuffer || !ress.dstBuffer) EXM_THROW(31, "Allocation error : not enough memory");
626
627 ress.cdict = LZ4IO_createCDict(prefs);
628
629 return ress;
630 }
631
LZ4IO_freeCResources(cRess_t ress)632 static void LZ4IO_freeCResources(cRess_t ress)
633 {
634 free(ress.srcBuffer);
635 free(ress.dstBuffer);
636
637 LZ4F_freeCDict(ress.cdict);
638 ress.cdict = NULL;
639
640 { LZ4F_errorCode_t const errorCode = LZ4F_freeCompressionContext(ress.ctx);
641 if (LZ4F_isError(errorCode)) EXM_THROW(38, "Error : can't free LZ4F context resource : %s", LZ4F_getErrorName(errorCode)); }
642 }
643
644 /*
645 * LZ4IO_compressFilename_extRess()
646 * result : 0 : compression completed correctly
647 * 1 : missing or pb opening srcFileName
648 */
649 static int
LZ4IO_compressFilename_extRess(cRess_t ress,const char * srcFileName,const char * dstFileName,int compressionLevel,const LZ4IO_prefs_t * const io_prefs)650 LZ4IO_compressFilename_extRess(cRess_t ress,
651 const char* srcFileName, const char* dstFileName,
652 int compressionLevel, const LZ4IO_prefs_t* const io_prefs)
653 {
654 unsigned long long filesize = 0;
655 unsigned long long compressedfilesize = 0;
656 FILE* dstFile;
657 void* const srcBuffer = ress.srcBuffer;
658 void* const dstBuffer = ress.dstBuffer;
659 const size_t dstBufferSize = ress.dstBufferSize;
660 const size_t blockSize = io_prefs->blockSize;
661 size_t readSize;
662 LZ4F_compressionContext_t ctx = ress.ctx; /* just a pointer */
663 LZ4F_preferences_t prefs;
664
665 /* Init */
666 FILE* const srcFile = LZ4IO_openSrcFile(srcFileName);
667 if (srcFile == NULL) return 1;
668 dstFile = LZ4IO_openDstFile(dstFileName, io_prefs);
669 if (dstFile == NULL) { fclose(srcFile); return 1; }
670 memset(&prefs, 0, sizeof(prefs));
671
672 /* Set compression parameters */
673 prefs.autoFlush = 1;
674 prefs.compressionLevel = compressionLevel;
675 prefs.frameInfo.blockMode = (LZ4F_blockMode_t)io_prefs->blockIndependence;
676 prefs.frameInfo.blockSizeID = (LZ4F_blockSizeID_t)io_prefs->blockSizeId;
677 prefs.frameInfo.blockChecksumFlag = (LZ4F_blockChecksum_t)io_prefs->blockChecksum;
678 prefs.frameInfo.contentChecksumFlag = (LZ4F_contentChecksum_t)io_prefs->streamChecksum;
679 prefs.favorDecSpeed = io_prefs->favorDecSpeed;
680 if (io_prefs->contentSizeFlag) {
681 U64 const fileSize = UTIL_getOpenFileSize(srcFile);
682 prefs.frameInfo.contentSize = fileSize; /* == 0 if input == stdin */
683 if (fileSize==0)
684 DISPLAYLEVEL(3, "Warning : cannot determine input content size \n");
685 }
686
687 /* read first block */
688 readSize = fread(srcBuffer, (size_t)1, blockSize, srcFile);
689 if (ferror(srcFile)) EXM_THROW(30, "Error reading %s ", srcFileName);
690 filesize += readSize;
691
692 /* single-block file */
693 if (readSize < blockSize) {
694 /* Compress in single pass */
695 size_t const cSize = LZ4F_compressFrame_usingCDict(ctx, dstBuffer, dstBufferSize, srcBuffer, readSize, ress.cdict, &prefs);
696 if (LZ4F_isError(cSize))
697 EXM_THROW(31, "Compression failed : %s", LZ4F_getErrorName(cSize));
698 compressedfilesize = cSize;
699 DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%% ",
700 (unsigned)(filesize>>20), (double)compressedfilesize/(filesize+!filesize)*100); /* avoid division by zero */
701
702 /* Write Block */
703 if (fwrite(dstBuffer, 1, cSize, dstFile) != cSize) {
704 EXM_THROW(32, "Write error : failed writing single-block compressed frame");
705 } }
706
707 else
708
709 /* multiple-blocks file */
710 {
711 /* Write Frame Header */
712 size_t const headerSize = LZ4F_compressBegin_usingCDict(ctx, dstBuffer, dstBufferSize, ress.cdict, &prefs);
713 if (LZ4F_isError(headerSize)) EXM_THROW(33, "File header generation failed : %s", LZ4F_getErrorName(headerSize));
714 if (fwrite(dstBuffer, 1, headerSize, dstFile) != headerSize)
715 EXM_THROW(34, "Write error : cannot write header");
716 compressedfilesize += headerSize;
717
718 /* Main Loop - one block at a time */
719 while (readSize>0) {
720 size_t const outSize = LZ4F_compressUpdate(ctx, dstBuffer, dstBufferSize, srcBuffer, readSize, NULL);
721 if (LZ4F_isError(outSize))
722 EXM_THROW(35, "Compression failed : %s", LZ4F_getErrorName(outSize));
723 compressedfilesize += outSize;
724 DISPLAYUPDATE(2, "\rRead : %u MB ==> %.2f%% ",
725 (unsigned)(filesize>>20), (double)compressedfilesize/filesize*100);
726
727 /* Write Block */
728 if (fwrite(dstBuffer, 1, outSize, dstFile) != outSize)
729 EXM_THROW(36, "Write error : cannot write compressed block");
730
731 /* Read next block */
732 readSize = fread(srcBuffer, (size_t)1, (size_t)blockSize, srcFile);
733 filesize += readSize;
734 }
735 if (ferror(srcFile)) EXM_THROW(37, "Error reading %s ", srcFileName);
736
737 /* End of Frame mark */
738 { size_t const endSize = LZ4F_compressEnd(ctx, dstBuffer, dstBufferSize, NULL);
739 if (LZ4F_isError(endSize))
740 EXM_THROW(38, "End of frame error : %s", LZ4F_getErrorName(endSize));
741 if (fwrite(dstBuffer, 1, endSize, dstFile) != endSize)
742 EXM_THROW(39, "Write error : cannot write end of frame");
743 compressedfilesize += endSize;
744 } }
745
746 /* Release file handlers */
747 fclose (srcFile);
748 if (strcmp(dstFileName,stdoutmark)) fclose (dstFile); /* do not close stdout */
749
750 /* Copy owner, file permissions and modification time */
751 { stat_t statbuf;
752 if (strcmp (srcFileName, stdinmark)
753 && strcmp (dstFileName, stdoutmark)
754 && strcmp (dstFileName, nulmark)
755 && UTIL_getFileStat(srcFileName, &statbuf)) {
756 UTIL_setFileStat(dstFileName, &statbuf);
757 } }
758
759 if (io_prefs->removeSrcFile) { /* remove source file : --rm */
760 if (remove(srcFileName))
761 EXM_THROW(40, "Remove error : %s: %s", srcFileName, strerror(errno));
762 }
763
764 /* Final Status */
765 DISPLAYLEVEL(2, "\r%79s\r", "");
766 DISPLAYLEVEL(2, "Compressed %llu bytes into %llu bytes ==> %.2f%%\n",
767 filesize, compressedfilesize,
768 (double)compressedfilesize / (filesize + !filesize /* avoid division by zero */ ) * 100);
769
770 return 0;
771 }
772
773
LZ4IO_compressFilename(const char * srcFileName,const char * dstFileName,int compressionLevel,const LZ4IO_prefs_t * prefs)774 int LZ4IO_compressFilename(const char* srcFileName, const char* dstFileName, int compressionLevel, const LZ4IO_prefs_t* prefs)
775 {
776 UTIL_time_t const timeStart = UTIL_getTime();
777 clock_t const cpuStart = clock();
778 cRess_t const ress = LZ4IO_createCResources(prefs);
779
780 int const result = LZ4IO_compressFilename_extRess(ress, srcFileName, dstFileName, compressionLevel, prefs);
781
782 /* Free resources */
783 LZ4IO_freeCResources(ress);
784
785 /* Final Status */
786 { clock_t const cpuEnd = clock();
787 double const cpuLoad_s = (double)(cpuEnd - cpuStart) / CLOCKS_PER_SEC;
788 U64 const timeLength_ns = UTIL_clockSpanNano(timeStart);
789 double const timeLength_s = (double)timeLength_ns / 1000000000;
790 DISPLAYLEVEL(4, "Completed in %.2f sec (cpu load : %.0f%%)\n",
791 timeLength_s, (cpuLoad_s / timeLength_s) * 100);
792 }
793
794 return result;
795 }
796
797
LZ4IO_compressMultipleFilenames(const char ** inFileNamesTable,int ifntSize,const char * suffix,int compressionLevel,const LZ4IO_prefs_t * prefs)798 int LZ4IO_compressMultipleFilenames(
799 const char** inFileNamesTable, int ifntSize,
800 const char* suffix,
801 int compressionLevel,
802 const LZ4IO_prefs_t* prefs)
803 {
804 int i;
805 int missed_files = 0;
806 char* dstFileName = (char*)malloc(FNSPACE);
807 size_t ofnSize = FNSPACE;
808 const size_t suffixSize = strlen(suffix);
809 cRess_t ress;
810
811 if (dstFileName == NULL) return ifntSize; /* not enough memory */
812 ress = LZ4IO_createCResources(prefs);
813
814 /* loop on each file */
815 for (i=0; i<ifntSize; i++) {
816 size_t const ifnSize = strlen(inFileNamesTable[i]);
817 if (!strcmp(suffix, stdoutmark)) {
818 missed_files += LZ4IO_compressFilename_extRess(ress,
819 inFileNamesTable[i], stdoutmark,
820 compressionLevel, prefs);
821 continue;
822 }
823 if (ofnSize <= ifnSize+suffixSize+1) {
824 free(dstFileName);
825 ofnSize = ifnSize + 20;
826 dstFileName = (char*)malloc(ofnSize);
827 if (dstFileName==NULL) {
828 LZ4IO_freeCResources(ress);
829 return ifntSize;
830 } }
831 strcpy(dstFileName, inFileNamesTable[i]);
832 strcat(dstFileName, suffix);
833
834 missed_files += LZ4IO_compressFilename_extRess(ress,
835 inFileNamesTable[i], dstFileName,
836 compressionLevel, prefs);
837 }
838
839 /* Close & Free */
840 LZ4IO_freeCResources(ress);
841 free(dstFileName);
842
843 return missed_files;
844 }
845
846
847 /* ********************************************************************* */
848 /* ********************** LZ4 file-stream Decompression **************** */
849 /* ********************************************************************* */
850
851 /* It's presumed that s points to a memory space of size >= 4 */
LZ4IO_readLE32(const void * s)852 static unsigned LZ4IO_readLE32 (const void* s)
853 {
854 const unsigned char* const srcPtr = (const unsigned char*)s;
855 unsigned value32 = srcPtr[0];
856 value32 += (unsigned)srcPtr[1] << 8;
857 value32 += (unsigned)srcPtr[2] << 16;
858 value32 += (unsigned)srcPtr[3] << 24;
859 return value32;
860 }
861
862
863 static unsigned
LZ4IO_fwriteSparse(FILE * file,const void * buffer,size_t bufferSize,int sparseFileSupport,unsigned storedSkips)864 LZ4IO_fwriteSparse(FILE* file,
865 const void* buffer, size_t bufferSize,
866 int sparseFileSupport,
867 unsigned storedSkips)
868 {
869 const size_t sizeT = sizeof(size_t);
870 const size_t maskT = sizeT -1 ;
871 const size_t* const bufferT = (const size_t*)buffer; /* Buffer is supposed malloc'ed, hence aligned on size_t */
872 const size_t* ptrT = bufferT;
873 size_t bufferSizeT = bufferSize / sizeT;
874 const size_t* const bufferTEnd = bufferT + bufferSizeT;
875 const size_t segmentSizeT = (32 KB) / sizeT;
876 int const sparseMode = (sparseFileSupport - (file==stdout)) > 0;
877
878 if (!sparseMode) { /* normal write */
879 size_t const sizeCheck = fwrite(buffer, 1, bufferSize, file);
880 if (sizeCheck != bufferSize) EXM_THROW(70, "Write error : cannot write decoded block");
881 return 0;
882 }
883
884 /* avoid int overflow */
885 if (storedSkips > 1 GB) {
886 int const seekResult = UTIL_fseek(file, 1 GB, SEEK_CUR);
887 if (seekResult != 0) EXM_THROW(71, "1 GB skip error (sparse file support)");
888 storedSkips -= 1 GB;
889 }
890
891 while (ptrT < bufferTEnd) {
892 size_t seg0SizeT = segmentSizeT;
893 size_t nb0T;
894
895 /* count leading zeros */
896 if (seg0SizeT > bufferSizeT) seg0SizeT = bufferSizeT;
897 bufferSizeT -= seg0SizeT;
898 for (nb0T=0; (nb0T < seg0SizeT) && (ptrT[nb0T] == 0); nb0T++) ;
899 storedSkips += (unsigned)(nb0T * sizeT);
900
901 if (nb0T != seg0SizeT) { /* not all 0s */
902 errno = 0;
903 { int const seekResult = UTIL_fseek(file, storedSkips, SEEK_CUR);
904 if (seekResult) EXM_THROW(72, "Sparse skip error(%d): %s ; try --no-sparse", (int)errno, strerror(errno));
905 }
906 storedSkips = 0;
907 seg0SizeT -= nb0T;
908 ptrT += nb0T;
909 { size_t const sizeCheck = fwrite(ptrT, sizeT, seg0SizeT, file);
910 if (sizeCheck != seg0SizeT) EXM_THROW(73, "Write error : cannot write decoded block");
911 } }
912 ptrT += seg0SizeT;
913 }
914
915 if (bufferSize & maskT) { /* size not multiple of sizeT : implies end of block */
916 const char* const restStart = (const char*)bufferTEnd;
917 const char* restPtr = restStart;
918 size_t const restSize = bufferSize & maskT;
919 const char* const restEnd = restStart + restSize;
920 for (; (restPtr < restEnd) && (*restPtr == 0); restPtr++) ;
921 storedSkips += (unsigned) (restPtr - restStart);
922 if (restPtr != restEnd) {
923 int const seekResult = UTIL_fseek(file, storedSkips, SEEK_CUR);
924 if (seekResult) EXM_THROW(74, "Sparse skip error ; try --no-sparse");
925 storedSkips = 0;
926 { size_t const sizeCheck = fwrite(restPtr, 1, (size_t)(restEnd - restPtr), file);
927 if (sizeCheck != (size_t)(restEnd - restPtr)) EXM_THROW(75, "Write error : cannot write decoded end of block");
928 } }
929 }
930
931 return storedSkips;
932 }
933
LZ4IO_fwriteSparseEnd(FILE * file,unsigned storedSkips)934 static void LZ4IO_fwriteSparseEnd(FILE* file, unsigned storedSkips)
935 {
936 if (storedSkips>0) { /* implies sparseFileSupport>0 */
937 const char lastZeroByte[1] = { 0 };
938 if (UTIL_fseek(file, storedSkips-1, SEEK_CUR) != 0)
939 EXM_THROW(69, "Final skip error (sparse file)\n");
940 if (fwrite(lastZeroByte, 1, 1, file) != 1)
941 EXM_THROW(69, "Write error : cannot write last zero\n");
942 }
943 }
944
945
946 static unsigned g_magicRead = 0; /* out-parameter of LZ4IO_decodeLegacyStream() */
LZ4IO_decodeLegacyStream(FILE * finput,FILE * foutput,const LZ4IO_prefs_t * prefs)947 static unsigned long long LZ4IO_decodeLegacyStream(FILE* finput, FILE* foutput, const LZ4IO_prefs_t* prefs)
948 {
949 unsigned long long streamSize = 0;
950 unsigned storedSkips = 0;
951
952 /* Allocate Memory */
953 char* const in_buff = (char*)malloc((size_t)LZ4_compressBound(LEGACY_BLOCKSIZE));
954 char* const out_buff = (char*)malloc(LEGACY_BLOCKSIZE);
955 if (!in_buff || !out_buff) EXM_THROW(51, "Allocation error : not enough memory");
956
957 /* Main Loop */
958 while (1) {
959 unsigned int blockSize;
960
961 /* Block Size */
962 { size_t const sizeCheck = fread(in_buff, 1, 4, finput);
963 if (sizeCheck == 0) break; /* Nothing to read : file read is completed */
964 if (sizeCheck != 4) EXM_THROW(52, "Read error : cannot access block size "); }
965 blockSize = LZ4IO_readLE32(in_buff); /* Convert to Little Endian */
966 if (blockSize > LZ4_COMPRESSBOUND(LEGACY_BLOCKSIZE)) {
967 /* Cannot read next block : maybe new stream ? */
968 g_magicRead = blockSize;
969 break;
970 }
971
972 /* Read Block */
973 { size_t const sizeCheck = fread(in_buff, 1, blockSize, finput);
974 if (sizeCheck!=blockSize) EXM_THROW(52, "Read error : cannot access compressed block !"); }
975
976 /* Decode Block */
977 { int const decodeSize = LZ4_decompress_safe(in_buff, out_buff, (int)blockSize, LEGACY_BLOCKSIZE);
978 if (decodeSize < 0) EXM_THROW(53, "Decoding Failed ! Corrupted input detected !");
979 streamSize += (unsigned long long)decodeSize;
980 /* Write Block */
981 storedSkips = LZ4IO_fwriteSparse(foutput, out_buff, (size_t)decodeSize, prefs->sparseFileSupport, storedSkips); /* success or die */
982 } }
983 if (ferror(finput)) EXM_THROW(54, "Read error : ferror");
984
985 LZ4IO_fwriteSparseEnd(foutput, storedSkips);
986
987 /* Free */
988 free(in_buff);
989 free(out_buff);
990
991 return streamSize;
992 }
993
994
995
996 typedef struct {
997 void* srcBuffer;
998 size_t srcBufferSize;
999 void* dstBuffer;
1000 size_t dstBufferSize;
1001 FILE* dstFile;
1002 LZ4F_decompressionContext_t dCtx;
1003 void* dictBuffer;
1004 size_t dictBufferSize;
1005 } dRess_t;
1006
LZ4IO_loadDDict(dRess_t * ress,const LZ4IO_prefs_t * const prefs)1007 static void LZ4IO_loadDDict(dRess_t* ress, const LZ4IO_prefs_t* const prefs)
1008 {
1009 if (!prefs->useDictionary) {
1010 ress->dictBuffer = NULL;
1011 ress->dictBufferSize = 0;
1012 return;
1013 }
1014
1015 ress->dictBuffer = LZ4IO_createDict(&ress->dictBufferSize, prefs->dictionaryFilename);
1016 if (!ress->dictBuffer) EXM_THROW(25, "Dictionary error : could not create dictionary");
1017 }
1018
1019 static const size_t LZ4IO_dBufferSize = 64 KB;
LZ4IO_createDResources(const LZ4IO_prefs_t * const prefs)1020 static dRess_t LZ4IO_createDResources(const LZ4IO_prefs_t* const prefs)
1021 {
1022 dRess_t ress;
1023
1024 /* init */
1025 LZ4F_errorCode_t const errorCode = LZ4F_createDecompressionContext(&ress.dCtx, LZ4F_VERSION);
1026 if (LZ4F_isError(errorCode)) EXM_THROW(60, "Can't create LZ4F context : %s", LZ4F_getErrorName(errorCode));
1027
1028 /* Allocate Memory */
1029 ress.srcBufferSize = LZ4IO_dBufferSize;
1030 ress.srcBuffer = malloc(ress.srcBufferSize);
1031 ress.dstBufferSize = LZ4IO_dBufferSize;
1032 ress.dstBuffer = malloc(ress.dstBufferSize);
1033 if (!ress.srcBuffer || !ress.dstBuffer) EXM_THROW(61, "Allocation error : not enough memory");
1034
1035 LZ4IO_loadDDict(&ress, prefs);
1036
1037 ress.dstFile = NULL;
1038 return ress;
1039 }
1040
LZ4IO_freeDResources(dRess_t ress)1041 static void LZ4IO_freeDResources(dRess_t ress)
1042 {
1043 LZ4F_errorCode_t errorCode = LZ4F_freeDecompressionContext(ress.dCtx);
1044 if (LZ4F_isError(errorCode)) EXM_THROW(69, "Error : can't free LZ4F context resource : %s", LZ4F_getErrorName(errorCode));
1045 free(ress.srcBuffer);
1046 free(ress.dstBuffer);
1047 free(ress.dictBuffer);
1048 }
1049
1050
1051 static unsigned long long
LZ4IO_decompressLZ4F(dRess_t ress,FILE * const srcFile,FILE * const dstFile,const LZ4IO_prefs_t * const prefs)1052 LZ4IO_decompressLZ4F(dRess_t ress,
1053 FILE* const srcFile, FILE* const dstFile,
1054 const LZ4IO_prefs_t* const prefs)
1055 {
1056 unsigned long long filesize = 0;
1057 LZ4F_errorCode_t nextToLoad;
1058 unsigned storedSkips = 0;
1059
1060 /* Init feed with magic number (already consumed from FILE* sFile) */
1061 { size_t inSize = MAGICNUMBER_SIZE;
1062 size_t outSize= 0;
1063 LZ4IO_writeLE32(ress.srcBuffer, LZ4IO_MAGICNUMBER);
1064 nextToLoad = LZ4F_decompress_usingDict(ress.dCtx, ress.dstBuffer, &outSize, ress.srcBuffer, &inSize, ress.dictBuffer, ress.dictBufferSize, NULL);
1065 if (LZ4F_isError(nextToLoad)) EXM_THROW(62, "Header error : %s", LZ4F_getErrorName(nextToLoad));
1066 }
1067
1068 /* Main Loop */
1069 for (;nextToLoad;) {
1070 size_t readSize;
1071 size_t pos = 0;
1072 size_t decodedBytes = ress.dstBufferSize;
1073
1074 /* Read input */
1075 if (nextToLoad > ress.srcBufferSize) nextToLoad = ress.srcBufferSize;
1076 readSize = fread(ress.srcBuffer, 1, nextToLoad, srcFile);
1077 if (!readSize) break; /* reached end of file or stream */
1078
1079 while ((pos < readSize) || (decodedBytes == ress.dstBufferSize)) { /* still to read, or still to flush */
1080 /* Decode Input (at least partially) */
1081 size_t remaining = readSize - pos;
1082 decodedBytes = ress.dstBufferSize;
1083 nextToLoad = LZ4F_decompress_usingDict(ress.dCtx, ress.dstBuffer, &decodedBytes, (char*)(ress.srcBuffer)+pos, &remaining, ress.dictBuffer, ress.dictBufferSize, NULL);
1084 if (LZ4F_isError(nextToLoad)) EXM_THROW(66, "Decompression error : %s", LZ4F_getErrorName(nextToLoad));
1085 pos += remaining;
1086
1087 /* Write Block */
1088 if (decodedBytes) {
1089 if (!prefs->testMode)
1090 storedSkips = LZ4IO_fwriteSparse(dstFile, ress.dstBuffer, decodedBytes, prefs->sparseFileSupport, storedSkips);
1091 filesize += decodedBytes;
1092 DISPLAYUPDATE(2, "\rDecompressed : %u MB ", (unsigned)(filesize>>20));
1093 }
1094
1095 if (!nextToLoad) break;
1096 }
1097 }
1098 /* can be out because readSize == 0, which could be an fread() error */
1099 if (ferror(srcFile)) EXM_THROW(67, "Read error");
1100
1101 if (!prefs->testMode) LZ4IO_fwriteSparseEnd(dstFile, storedSkips);
1102 if (nextToLoad!=0) EXM_THROW(68, "Unfinished stream");
1103
1104 return filesize;
1105 }
1106
1107
1108 /* LZ4IO_passThrough:
1109 * just output the same content as input, no decoding.
1110 * This is a capability of zcat, and by extension lz4cat
1111 * MNstore : contain the first MAGICNUMBER_SIZE bytes already read from finput
1112 */
1113 #define PTSIZE (64 KB)
1114 #define PTSIZET (PTSIZE / sizeof(size_t))
1115 static unsigned long long
LZ4IO_passThrough(FILE * finput,FILE * foutput,unsigned char MNstore[MAGICNUMBER_SIZE],int sparseFileSupport)1116 LZ4IO_passThrough(FILE* finput, FILE* foutput,
1117 unsigned char MNstore[MAGICNUMBER_SIZE],
1118 int sparseFileSupport)
1119 {
1120 size_t buffer[PTSIZET];
1121 size_t readBytes = 1;
1122 unsigned long long total = MAGICNUMBER_SIZE;
1123 unsigned storedSkips = 0;
1124
1125 if (fwrite(MNstore, 1, MAGICNUMBER_SIZE, foutput) != MAGICNUMBER_SIZE) {
1126 EXM_THROW(50, "Pass-through write error");
1127 }
1128 while (readBytes) {
1129 readBytes = fread(buffer, 1, sizeof(buffer), finput);
1130 total += readBytes;
1131 storedSkips = LZ4IO_fwriteSparse(foutput, buffer, readBytes, sparseFileSupport, storedSkips);
1132 }
1133 if (ferror(finput)) EXM_THROW(51, "Read Error");
1134
1135 LZ4IO_fwriteSparseEnd(foutput, storedSkips);
1136 return total;
1137 }
1138
1139
1140 /** Safely handle cases when (unsigned)offset > LONG_MAX */
fseek_u32(FILE * fp,unsigned offset,int where)1141 static int fseek_u32(FILE *fp, unsigned offset, int where)
1142 {
1143 const unsigned stepMax = 1U << 30;
1144 int errorNb = 0;
1145
1146 if (where != SEEK_CUR) return -1; /* Only allows SEEK_CUR */
1147 while (offset > 0) {
1148 unsigned s = offset;
1149 if (s > stepMax) s = stepMax;
1150 errorNb = UTIL_fseek(fp, (long) s, SEEK_CUR);
1151 if (errorNb != 0) break;
1152 offset -= s;
1153 }
1154 return errorNb;
1155 }
1156
1157 #define ENDOFSTREAM ((unsigned long long)-1)
1158 static unsigned long long
selectDecoder(dRess_t ress,FILE * finput,FILE * foutput,const LZ4IO_prefs_t * const prefs)1159 selectDecoder(dRess_t ress,
1160 FILE* finput, FILE* foutput,
1161 const LZ4IO_prefs_t* const prefs)
1162 {
1163 unsigned char MNstore[MAGICNUMBER_SIZE];
1164 unsigned magicNumber;
1165 static unsigned nbFrames = 0;
1166
1167 /* init */
1168 nbFrames++;
1169
1170 /* Check Archive Header */
1171 if (g_magicRead) { /* magic number already read from finput (see legacy frame)*/
1172 magicNumber = g_magicRead;
1173 g_magicRead = 0;
1174 } else {
1175 size_t const nbReadBytes = fread(MNstore, 1, MAGICNUMBER_SIZE, finput);
1176 if (nbReadBytes==0) { nbFrames = 0; return ENDOFSTREAM; } /* EOF */
1177 if (nbReadBytes != MAGICNUMBER_SIZE)
1178 EXM_THROW(40, "Unrecognized header : Magic Number unreadable");
1179 magicNumber = LZ4IO_readLE32(MNstore); /* Little Endian format */
1180 }
1181 if (LZ4IO_isSkippableMagicNumber(magicNumber))
1182 magicNumber = LZ4IO_SKIPPABLE0; /* fold skippable magic numbers */
1183
1184 switch(magicNumber)
1185 {
1186 case LZ4IO_MAGICNUMBER:
1187 return LZ4IO_decompressLZ4F(ress, finput, foutput, prefs);
1188 case LEGACY_MAGICNUMBER:
1189 DISPLAYLEVEL(4, "Detected : Legacy format \n");
1190 return LZ4IO_decodeLegacyStream(finput, foutput, prefs);
1191 case LZ4IO_SKIPPABLE0:
1192 DISPLAYLEVEL(4, "Skipping detected skippable area \n");
1193 { size_t const nbReadBytes = fread(MNstore, 1, 4, finput);
1194 if (nbReadBytes != 4)
1195 EXM_THROW(42, "Stream error : skippable size unreadable");
1196 }
1197 { unsigned const size = LZ4IO_readLE32(MNstore);
1198 int const errorNb = fseek_u32(finput, size, SEEK_CUR);
1199 if (errorNb != 0)
1200 EXM_THROW(43, "Stream error : cannot skip skippable area");
1201 }
1202 return 0;
1203 EXTENDED_FORMAT; /* macro extension for custom formats */
1204 default:
1205 if (nbFrames == 1) { /* just started */
1206 /* Wrong magic number at the beginning of 1st stream */
1207 if (!prefs->testMode && prefs->overwrite && prefs->passThrough) {
1208 nbFrames = 0;
1209 return LZ4IO_passThrough(finput, foutput, MNstore, prefs->sparseFileSupport);
1210 }
1211 EXM_THROW(44,"Unrecognized header : file cannot be decoded");
1212 }
1213 { long int const position = ftell(finput); /* only works for files < 2 GB */
1214 DISPLAYLEVEL(2, "Stream followed by undecodable data ");
1215 if (position != -1L)
1216 DISPLAYLEVEL(2, "at position %i ", (int)position);
1217 DISPLAYLEVEL(2, "\n");
1218 }
1219 return ENDOFSTREAM;
1220 }
1221 }
1222
1223
1224 static int
LZ4IO_decompressSrcFile(dRess_t ress,const char * input_filename,const char * output_filename,const LZ4IO_prefs_t * const prefs)1225 LZ4IO_decompressSrcFile(dRess_t ress,
1226 const char* input_filename, const char* output_filename,
1227 const LZ4IO_prefs_t* const prefs)
1228 {
1229 FILE* const foutput = ress.dstFile;
1230 unsigned long long filesize = 0;
1231
1232 /* Init */
1233 FILE* const finput = LZ4IO_openSrcFile(input_filename);
1234 if (finput==NULL) return 1;
1235 assert(foutput != NULL);
1236
1237 /* Loop over multiple streams */
1238 for ( ; ; ) { /* endless loop, see break condition */
1239 unsigned long long const decodedSize =
1240 selectDecoder(ress, finput, foutput, prefs);
1241 if (decodedSize == ENDOFSTREAM) break;
1242 filesize += decodedSize;
1243 }
1244
1245 /* Close input */
1246 fclose(finput);
1247 if (prefs->removeSrcFile) { /* --rm */
1248 if (remove(input_filename))
1249 EXM_THROW(45, "Remove error : %s: %s", input_filename, strerror(errno));
1250 }
1251
1252 /* Final Status */
1253 DISPLAYLEVEL(2, "\r%79s\r", "");
1254 DISPLAYLEVEL(2, "%-20.20s : decoded %llu bytes \n", input_filename, filesize);
1255 (void)output_filename;
1256
1257 return 0;
1258 }
1259
1260
1261 static int
LZ4IO_decompressDstFile(dRess_t ress,const char * input_filename,const char * output_filename,const LZ4IO_prefs_t * const prefs)1262 LZ4IO_decompressDstFile(dRess_t ress,
1263 const char* input_filename, const char* output_filename,
1264 const LZ4IO_prefs_t* const prefs)
1265 {
1266 stat_t statbuf;
1267 int stat_result = 0;
1268 FILE* const foutput = LZ4IO_openDstFile(output_filename, prefs);
1269 if (foutput==NULL) return 1; /* failure */
1270
1271 if ( strcmp(input_filename, stdinmark)
1272 && UTIL_getFileStat(input_filename, &statbuf))
1273 stat_result = 1;
1274
1275 ress.dstFile = foutput;
1276 LZ4IO_decompressSrcFile(ress, input_filename, output_filename, prefs);
1277
1278 fclose(foutput);
1279
1280 /* Copy owner, file permissions and modification time */
1281 if ( stat_result != 0
1282 && strcmp (output_filename, stdoutmark)
1283 && strcmp (output_filename, nulmark)) {
1284 UTIL_setFileStat(output_filename, &statbuf);
1285 /* should return value be read ? or is silent fail good enough ? */
1286 }
1287
1288 return 0;
1289 }
1290
1291
LZ4IO_decompressFilename(const char * input_filename,const char * output_filename,const LZ4IO_prefs_t * prefs)1292 int LZ4IO_decompressFilename(const char* input_filename, const char* output_filename, const LZ4IO_prefs_t* prefs)
1293 {
1294 dRess_t const ress = LZ4IO_createDResources(prefs);
1295 clock_t const start = clock();
1296
1297 int const missingFiles = LZ4IO_decompressDstFile(ress, input_filename, output_filename, prefs);
1298
1299 clock_t const end = clock();
1300 double const seconds = (double)(end - start) / CLOCKS_PER_SEC;
1301 DISPLAYLEVEL(4, "Done in %.2f sec \n", seconds);
1302
1303 LZ4IO_freeDResources(ress);
1304 return missingFiles;
1305 }
1306
1307
LZ4IO_decompressMultipleFilenames(const char ** inFileNamesTable,int ifntSize,const char * suffix,const LZ4IO_prefs_t * prefs)1308 int LZ4IO_decompressMultipleFilenames(
1309 const char** inFileNamesTable, int ifntSize,
1310 const char* suffix,
1311 const LZ4IO_prefs_t* prefs)
1312 {
1313 int i;
1314 int skippedFiles = 0;
1315 int missingFiles = 0;
1316 char* outFileName = (char*)malloc(FNSPACE);
1317 size_t ofnSize = FNSPACE;
1318 size_t const suffixSize = strlen(suffix);
1319 dRess_t ress = LZ4IO_createDResources(prefs);
1320
1321 if (outFileName==NULL) EXM_THROW(70, "Memory allocation error");
1322 ress.dstFile = LZ4IO_openDstFile(stdoutmark, prefs);
1323
1324 for (i=0; i<ifntSize; i++) {
1325 size_t const ifnSize = strlen(inFileNamesTable[i]);
1326 const char* const suffixPtr = inFileNamesTable[i] + ifnSize - suffixSize;
1327 if (!strcmp(suffix, stdoutmark)) {
1328 missingFiles += LZ4IO_decompressSrcFile(ress, inFileNamesTable[i], stdoutmark, prefs);
1329 continue;
1330 }
1331 if (ofnSize <= ifnSize-suffixSize+1) {
1332 free(outFileName);
1333 ofnSize = ifnSize + 20;
1334 outFileName = (char*)malloc(ofnSize);
1335 if (outFileName==NULL) EXM_THROW(71, "Memory allocation error");
1336 }
1337 if (ifnSize <= suffixSize || strcmp(suffixPtr, suffix) != 0) {
1338 DISPLAYLEVEL(1, "File extension doesn't match expected LZ4_EXTENSION (%4s); will not process file: %s\n", suffix, inFileNamesTable[i]);
1339 skippedFiles++;
1340 continue;
1341 }
1342 memcpy(outFileName, inFileNamesTable[i], ifnSize - suffixSize);
1343 outFileName[ifnSize-suffixSize] = '\0';
1344 missingFiles += LZ4IO_decompressDstFile(ress, inFileNamesTable[i], outFileName, prefs);
1345 }
1346
1347 LZ4IO_freeDResources(ress);
1348 free(outFileName);
1349 return missingFiles + skippedFiles;
1350 }
1351
1352
1353 /* ********************************************************************* */
1354 /* ********************** LZ4 --list command *********************** */
1355 /* ********************************************************************* */
1356
1357 typedef enum
1358 {
1359 lz4Frame = 0,
1360 legacyFrame,
1361 skippableFrame
1362 } LZ4IO_frameType_t;
1363
1364 typedef struct {
1365 LZ4F_frameInfo_t lz4FrameInfo;
1366 LZ4IO_frameType_t frameType;
1367 } LZ4IO_frameInfo_t;
1368
1369 #define LZ4IO_INIT_FRAMEINFO { LZ4F_INIT_FRAMEINFO, lz4Frame }
1370
1371 typedef struct {
1372 const char* fileName;
1373 unsigned long long fileSize;
1374 unsigned long long frameCount;
1375 LZ4IO_frameInfo_t frameSummary;
1376 unsigned short eqFrameTypes;
1377 unsigned short eqBlockTypes;
1378 unsigned short allContentSize;
1379 } LZ4IO_cFileInfo_t;
1380
1381 #define LZ4IO_INIT_CFILEINFO { NULL, 0ULL, 0, LZ4IO_INIT_FRAMEINFO, 1, 1, 1 }
1382
1383 typedef enum { LZ4IO_LZ4F_OK, LZ4IO_format_not_known, LZ4IO_not_a_file } LZ4IO_infoResult;
1384
1385 static const char * LZ4IO_frameTypeNames[] = {"LZ4Frame", "LegacyFrame", "SkippableFrame" };
1386
1387 /* Read block headers and skip block data
1388 Return total blocks size for this frame including block headers,
1389 block checksums and content checksums.
1390 returns 0 in case it can't succesfully skip block data.
1391 Assumes SEEK_CUR after frame header.
1392 */
1393 static unsigned long long
LZ4IO_skipBlocksData(FILE * finput,const LZ4F_blockChecksum_t blockChecksumFlag,const LZ4F_contentChecksum_t contentChecksumFlag)1394 LZ4IO_skipBlocksData(FILE* finput,
1395 const LZ4F_blockChecksum_t blockChecksumFlag,
1396 const LZ4F_contentChecksum_t contentChecksumFlag)
1397 {
1398 unsigned char blockInfo[LZ4F_BLOCK_HEADER_SIZE];
1399 unsigned long long totalBlocksSize = 0;
1400 for (;;) {
1401 if (!fread(blockInfo, 1, LZ4F_BLOCK_HEADER_SIZE, finput)) {
1402 if (feof(finput)) return totalBlocksSize;
1403 return 0;
1404 }
1405 totalBlocksSize += LZ4F_BLOCK_HEADER_SIZE;
1406 { const unsigned long nextCBlockSize = LZ4IO_readLE32(&blockInfo) & 0x7FFFFFFFU;
1407 const unsigned long nextBlock = nextCBlockSize + (blockChecksumFlag * LZ4F_BLOCK_CHECKSUM_SIZE);
1408 if (nextCBlockSize == 0) {
1409 /* Reached EndMark */
1410 if (contentChecksumFlag) {
1411 /* Skip content checksum */
1412 if (UTIL_fseek(finput, LZ4F_CONTENT_CHECKSUM_SIZE, SEEK_CUR) != 0) {
1413 return 0;
1414 }
1415 totalBlocksSize += LZ4F_CONTENT_CHECKSUM_SIZE;
1416 }
1417 break;
1418 }
1419 totalBlocksSize += nextBlock;
1420 /* skip to the next block */
1421 assert(nextBlock < LONG_MAX);
1422 if (UTIL_fseek(finput, (long)nextBlock, SEEK_CUR) != 0) return 0;
1423 } }
1424 return totalBlocksSize;
1425 }
1426
1427 /* For legacy frames only.
1428 Read block headers and skip block data.
1429 Return total blocks size for this frame including block headers.
1430 or 0 in case it can't succesfully skip block data.
1431 This works as long as legacy block header size = magic number size.
1432 Assumes SEEK_CUR after frame header.
1433 */
LZ4IO_skipLegacyBlocksData(FILE * finput)1434 static unsigned long long LZ4IO_skipLegacyBlocksData(FILE* finput)
1435 {
1436 unsigned char blockInfo[LZIO_LEGACY_BLOCK_HEADER_SIZE];
1437 unsigned long long totalBlocksSize = 0;
1438 LZ4IO_STATIC_ASSERT(LZIO_LEGACY_BLOCK_HEADER_SIZE == MAGICNUMBER_SIZE);
1439 for (;;) {
1440 if (!fread(blockInfo, 1, LZIO_LEGACY_BLOCK_HEADER_SIZE, finput)) {
1441 if (feof(finput)) return totalBlocksSize;
1442 return 0;
1443 }
1444 { const unsigned int nextCBlockSize = LZ4IO_readLE32(&blockInfo);
1445 if ( nextCBlockSize == LEGACY_MAGICNUMBER ||
1446 nextCBlockSize == LZ4IO_MAGICNUMBER ||
1447 LZ4IO_isSkippableMagicNumber(nextCBlockSize)) {
1448 /* Rewind back. we want cursor at the begining of next frame.*/
1449 if (fseek(finput, -LZIO_LEGACY_BLOCK_HEADER_SIZE, SEEK_CUR) != 0) {
1450 return 0;
1451 }
1452 break;
1453 }
1454 totalBlocksSize += LZIO_LEGACY_BLOCK_HEADER_SIZE + nextCBlockSize;
1455 /* skip to the next block */
1456 if (UTIL_fseek(finput, nextCBlockSize, SEEK_CUR) != 0) {
1457 return 0;
1458 } } }
1459 return totalBlocksSize;
1460 }
1461
1462 /* LZ4IO_blockTypeID:
1463 * return human-readable block type, following command line convention
1464 * buffer : must be a valid memory area of at least 4 bytes */
LZ4IO_blockTypeID(LZ4F_blockSizeID_t sizeID,LZ4F_blockMode_t blockMode,char buffer[4])1465 const char* LZ4IO_blockTypeID(LZ4F_blockSizeID_t sizeID, LZ4F_blockMode_t blockMode, char buffer[4])
1466 {
1467 buffer[0] = 'B';
1468 assert(sizeID >= 4); assert(sizeID <= 7);
1469 buffer[1] = (char)(sizeID + '0');
1470 buffer[2] = (blockMode == LZ4F_blockIndependent) ? 'I' : 'D';
1471 buffer[3] = 0;
1472 return buffer;
1473 }
1474
1475 /* buffer : must be valid memory area of at least 10 bytes */
LZ4IO_toHuman(long double size,char * buf)1476 static const char* LZ4IO_toHuman(long double size, char *buf)
1477 {
1478 const char units[] = {"\0KMGTPEZY"};
1479 size_t i = 0;
1480 for (; size >= 1024; i++) size /= 1024;
1481 sprintf(buf, "%.2Lf%c", size, units[i]);
1482 return buf;
1483 }
1484
1485 /* Get filename without path prefix */
LZ4IO_baseName(const char * input_filename)1486 static const char* LZ4IO_baseName(const char* input_filename)
1487 {
1488 const char* b = strrchr(input_filename, '/');
1489 if (!b) b = strrchr(input_filename, '\\');
1490 if (!b) return input_filename;
1491 return b + 1;
1492 }
1493
1494 /* Report frame/s information (--list) in verbose mode (-v).
1495 * Will populate file info with fileName and frameSummary where applicable.
1496 * - TODO :
1497 * + report nb of blocks, hence max. possible decompressed size (when not reported in header)
1498 */
1499 static LZ4IO_infoResult
LZ4IO_getCompressedFileInfo(LZ4IO_cFileInfo_t * cfinfo,const char * input_filename)1500 LZ4IO_getCompressedFileInfo(LZ4IO_cFileInfo_t* cfinfo, const char* input_filename)
1501 {
1502 LZ4IO_infoResult result = LZ4IO_format_not_known; /* default result (error) */
1503 unsigned char buffer[LZ4F_HEADER_SIZE_MAX];
1504 FILE* const finput = LZ4IO_openSrcFile(input_filename);
1505
1506 if (finput == NULL) return LZ4IO_not_a_file;
1507 cfinfo->fileSize = UTIL_getOpenFileSize(finput);
1508
1509 while (!feof(finput)) {
1510 LZ4IO_frameInfo_t frameInfo = LZ4IO_INIT_FRAMEINFO;
1511 unsigned magicNumber;
1512 /* Get MagicNumber */
1513 { size_t const nbReadBytes = fread(buffer, 1, MAGICNUMBER_SIZE, finput);
1514 if (nbReadBytes == 0) { break; } /* EOF */
1515 result = LZ4IO_format_not_known; /* default result (error) */
1516 if (nbReadBytes != MAGICNUMBER_SIZE) {
1517 EXM_THROW(40, "Unrecognized header : Magic Number unreadable");
1518 } }
1519 magicNumber = LZ4IO_readLE32(buffer); /* Little Endian format */
1520 if (LZ4IO_isSkippableMagicNumber(magicNumber))
1521 magicNumber = LZ4IO_SKIPPABLE0; /* fold skippable magic numbers */
1522
1523 switch (magicNumber) {
1524 case LZ4IO_MAGICNUMBER:
1525 if (cfinfo->frameSummary.frameType != lz4Frame) cfinfo->eqFrameTypes = 0;
1526 /* Get frame info */
1527 { const size_t readBytes = fread(buffer + MAGICNUMBER_SIZE, 1, LZ4F_HEADER_SIZE_MIN - MAGICNUMBER_SIZE, finput);
1528 if (!readBytes || ferror(finput)) EXM_THROW(71, "Error reading %s", input_filename);
1529 }
1530 { size_t hSize = LZ4F_headerSize(&buffer, LZ4F_HEADER_SIZE_MIN);
1531 if (LZ4F_isError(hSize)) break;
1532 if (hSize > (LZ4F_HEADER_SIZE_MIN + MAGICNUMBER_SIZE)) {
1533 /* We've already read LZ4F_HEADER_SIZE_MIN so read any extra until hSize*/
1534 const size_t readBytes = fread(buffer + LZ4F_HEADER_SIZE_MIN, 1, hSize - LZ4F_HEADER_SIZE_MIN, finput);
1535 if (!readBytes || ferror(finput)) EXM_THROW(72, "Error reading %s", input_filename);
1536 }
1537 /* Create decompression context */
1538 { LZ4F_dctx* dctx;
1539 if ( LZ4F_isError(LZ4F_createDecompressionContext(&dctx, LZ4F_VERSION)) ) break;
1540 { unsigned const frameInfoError = LZ4F_isError(LZ4F_getFrameInfo(dctx, &frameInfo.lz4FrameInfo, buffer, &hSize));
1541 LZ4F_freeDecompressionContext(dctx);
1542 if (frameInfoError) break;
1543 if ((cfinfo->frameSummary.lz4FrameInfo.blockSizeID != frameInfo.lz4FrameInfo.blockSizeID ||
1544 cfinfo->frameSummary.lz4FrameInfo.blockMode != frameInfo.lz4FrameInfo.blockMode)
1545 && cfinfo->frameCount != 0)
1546 cfinfo->eqBlockTypes = 0;
1547 { const unsigned long long totalBlocksSize = LZ4IO_skipBlocksData(finput,
1548 frameInfo.lz4FrameInfo.blockChecksumFlag,
1549 frameInfo.lz4FrameInfo.contentChecksumFlag);
1550 if (totalBlocksSize) {
1551 char bTypeBuffer[5];
1552 LZ4IO_blockTypeID(frameInfo.lz4FrameInfo.blockSizeID, frameInfo.lz4FrameInfo.blockMode, bTypeBuffer);
1553 DISPLAYLEVEL(3, " %6llu %14s %5s %8s",
1554 cfinfo->frameCount + 1,
1555 LZ4IO_frameTypeNames[frameInfo.frameType],
1556 bTypeBuffer,
1557 frameInfo.lz4FrameInfo.contentChecksumFlag ? "XXH32" : "-");
1558 if (frameInfo.lz4FrameInfo.contentSize) {
1559 { double const ratio = (double)(totalBlocksSize + hSize) / frameInfo.lz4FrameInfo.contentSize * 100;
1560 DISPLAYLEVEL(3, " %20llu %20llu %9.2f%%\n",
1561 totalBlocksSize + hSize,
1562 frameInfo.lz4FrameInfo.contentSize,
1563 ratio);
1564 }
1565 /* Now we've consumed frameInfo we can use it to store the total contentSize */
1566 frameInfo.lz4FrameInfo.contentSize += cfinfo->frameSummary.lz4FrameInfo.contentSize;
1567 }
1568 else {
1569 DISPLAYLEVEL(3, " %20llu %20s %9s \n", totalBlocksSize + hSize, "-", "-");
1570 cfinfo->allContentSize = 0;
1571 }
1572 result = LZ4IO_LZ4F_OK;
1573 } } } } }
1574 break;
1575 case LEGACY_MAGICNUMBER:
1576 frameInfo.frameType = legacyFrame;
1577 if (cfinfo->frameSummary.frameType != legacyFrame && cfinfo->frameCount != 0) cfinfo->eqFrameTypes = 0;
1578 cfinfo->eqBlockTypes = 0;
1579 cfinfo->allContentSize = 0;
1580 { const unsigned long long totalBlocksSize = LZ4IO_skipLegacyBlocksData(finput);
1581 if (totalBlocksSize) {
1582 DISPLAYLEVEL(3, " %6llu %14s %5s %8s %20llu %20s %9s\n",
1583 cfinfo->frameCount + 1,
1584 LZ4IO_frameTypeNames[frameInfo.frameType],
1585 "-", "-",
1586 totalBlocksSize + 4,
1587 "-", "-");
1588 result = LZ4IO_LZ4F_OK;
1589 } }
1590 break;
1591 case LZ4IO_SKIPPABLE0:
1592 frameInfo.frameType = skippableFrame;
1593 if (cfinfo->frameSummary.frameType != skippableFrame && cfinfo->frameCount != 0) cfinfo->eqFrameTypes = 0;
1594 cfinfo->eqBlockTypes = 0;
1595 cfinfo->allContentSize = 0;
1596 { size_t const nbReadBytes = fread(buffer, 1, 4, finput);
1597 if (nbReadBytes != 4)
1598 EXM_THROW(42, "Stream error : skippable size unreadable");
1599 }
1600 { unsigned const size = LZ4IO_readLE32(buffer);
1601 int const errorNb = fseek_u32(finput, size, SEEK_CUR);
1602 if (errorNb != 0)
1603 EXM_THROW(43, "Stream error : cannot skip skippable area");
1604 DISPLAYLEVEL(3, " %6llu %14s %5s %8s %20u %20s %9s\n",
1605 cfinfo->frameCount + 1,
1606 "SkippableFrame",
1607 "-", "-", size + 8, "-", "-");
1608
1609 result = LZ4IO_LZ4F_OK;
1610 }
1611 break;
1612 default:
1613 { long int const position = ftell(finput); /* only works for files < 2 GB */
1614 DISPLAYLEVEL(3, "Stream followed by undecodable data ");
1615 if (position != -1L)
1616 DISPLAYLEVEL(3, "at position %i ", (int)position);
1617 DISPLAYLEVEL(3, "\n");
1618 }
1619 break;
1620 }
1621 if (result != LZ4IO_LZ4F_OK) break;
1622 cfinfo->frameSummary = frameInfo;
1623 cfinfo->frameCount++;
1624 } /* while (!feof(finput)) */
1625 fclose(finput);
1626 return result;
1627 }
1628
1629
LZ4IO_displayCompressedFilesInfo(const char ** inFileNames,size_t ifnIdx)1630 int LZ4IO_displayCompressedFilesInfo(const char** inFileNames, size_t ifnIdx)
1631 {
1632 int result = 0;
1633 size_t idx = 0;
1634 if (g_displayLevel < 3) {
1635 DISPLAYOUT("%10s %14s %5s %11s %13s %9s %s\n",
1636 "Frames", "Type", "Block", "Compressed", "Uncompressed", "Ratio", "Filename");
1637 }
1638 for (; idx < ifnIdx; idx++) {
1639 /* Get file info */
1640 LZ4IO_cFileInfo_t cfinfo = LZ4IO_INIT_CFILEINFO;
1641 cfinfo.fileName = LZ4IO_baseName(inFileNames[idx]);
1642 if (!UTIL_isRegFile(inFileNames[idx])) {
1643 DISPLAYLEVEL(1, "lz4: %s is not a regular file \n", inFileNames[idx]);
1644 return 0;
1645 }
1646 DISPLAYLEVEL(3, "%s(%llu/%llu)\n", cfinfo.fileName, (unsigned long long)idx + 1, (unsigned long long)ifnIdx);
1647 DISPLAYLEVEL(3, " %6s %14s %5s %8s %20s %20s %9s\n",
1648 "Frame", "Type", "Block", "Checksum", "Compressed", "Uncompressed", "Ratio")
1649 { LZ4IO_infoResult const op_result = LZ4IO_getCompressedFileInfo(&cfinfo, inFileNames[idx]);
1650 if (op_result != LZ4IO_LZ4F_OK) {
1651 assert(op_result == LZ4IO_format_not_known);
1652 DISPLAYLEVEL(1, "lz4: %s: File format not recognized \n", inFileNames[idx]);
1653 return 0;
1654 } }
1655 DISPLAYLEVEL(3, "\n");
1656 if (g_displayLevel < 3) {
1657 /* Display Summary */
1658 { char buffers[3][10];
1659 DISPLAYOUT("%10llu %14s %5s %11s %13s ",
1660 cfinfo.frameCount,
1661 cfinfo.eqFrameTypes ? LZ4IO_frameTypeNames[cfinfo.frameSummary.frameType] : "-" ,
1662 cfinfo.eqBlockTypes ? LZ4IO_blockTypeID(cfinfo.frameSummary.lz4FrameInfo.blockSizeID,
1663 cfinfo.frameSummary.lz4FrameInfo.blockMode, buffers[0]) : "-",
1664 LZ4IO_toHuman((long double)cfinfo.fileSize, buffers[1]),
1665 cfinfo.allContentSize ? LZ4IO_toHuman((long double)cfinfo.frameSummary.lz4FrameInfo.contentSize, buffers[2]) : "-");
1666 if (cfinfo.allContentSize) {
1667 double const ratio = (double)cfinfo.fileSize / cfinfo.frameSummary.lz4FrameInfo.contentSize * 100;
1668 DISPLAYOUT("%9.2f%% %s \n", ratio, cfinfo.fileName);
1669 } else {
1670 DISPLAYOUT("%9s %s\n",
1671 "-",
1672 cfinfo.fileName);
1673 } } } /* if (g_displayLevel < 3) */
1674 } /* for (; idx < ifnIdx; idx++) */
1675
1676 return result;
1677 }
1678