1 /* 2 ** 2001 September 15 3 ** 4 ** The author disclaims copyright to this source code. In place of 5 ** a legal notice, here is a blessing: 6 ** 7 ** May you do good and not evil. 8 ** May you find forgiveness for yourself and forgive others. 9 ** May you share freely, never taking more than you give. 10 ** 11 ************************************************************************* 12 ** Utility functions used throughout sqlite. 13 ** 14 ** This file contains functions for allocating memory, comparing 15 ** strings, and stuff like that. 16 ** 17 ** $Id: util.c,v 1.74.2.1 2004/07/15 13:08:41 drh Exp $ 18 */ 19 #include "sqliteInt.h" 20 #include <stdarg.h> 21 #include <ctype.h> 22 23 /* 24 ** If malloc() ever fails, this global variable gets set to 1. 25 ** This causes the library to abort and never again function. 26 */ 27 int sqlite_malloc_failed = 0; 28 29 /* 30 ** If MEMORY_DEBUG is defined, then use versions of malloc() and 31 ** free() that track memory usage and check for buffer overruns. 32 */ 33 #ifdef MEMORY_DEBUG 34 35 /* 36 ** For keeping track of the number of mallocs and frees. This 37 ** is used to check for memory leaks. 38 */ 39 int sqlite_nMalloc; /* Number of sqliteMalloc() calls */ 40 int sqlite_nFree; /* Number of sqliteFree() calls */ 41 int sqlite_iMallocFail; /* Fail sqliteMalloc() after this many calls */ 42 #if MEMORY_DEBUG>1 43 static int memcnt = 0; 44 #endif 45 46 /* 47 ** Number of 32-bit guard words 48 */ 49 #define N_GUARD 1 50 51 /* 52 ** Allocate new memory and set it to zero. Return NULL if 53 ** no memory is available. 54 */ 55 void *sqliteMalloc_(int n, int bZero, char *zFile, int line){ 56 void *p; 57 int *pi; 58 int i, k; 59 if( sqlite_iMallocFail>=0 ){ 60 sqlite_iMallocFail--; 61 if( sqlite_iMallocFail==0 ){ 62 sqlite_malloc_failed++; 63 #if MEMORY_DEBUG>1 64 fprintf(stderr,"**** failed to allocate %d bytes at %s:%d\n", 65 n, zFile,line); 66 #endif 67 sqlite_iMallocFail--; 68 return 0; 69 } 70 } 71 if( n==0 ) return 0; 72 k = (n+sizeof(int)-1)/sizeof(int); 73 pi = malloc( (N_GUARD*2+1+k)*sizeof(int)); 74 if( pi==0 ){ 75 sqlite_malloc_failed++; 76 return 0; 77 } 78 sqlite_nMalloc++; 79 for(i=0; i<N_GUARD; i++) pi[i] = 0xdead1122; 80 pi[N_GUARD] = n; 81 for(i=0; i<N_GUARD; i++) pi[k+1+N_GUARD+i] = 0xdead3344; 82 p = &pi[N_GUARD+1]; 83 memset(p, bZero==0, n); 84 #if MEMORY_DEBUG>1 85 fprintf(stderr,"%06d malloc %d bytes at 0x%x from %s:%d\n", 86 ++memcnt, n, (int)p, zFile,line); 87 #endif 88 return p; 89 } 90 91 /* 92 ** Check to see if the given pointer was obtained from sqliteMalloc() 93 ** and is able to hold at least N bytes. Raise an exception if this 94 ** is not the case. 95 ** 96 ** This routine is used for testing purposes only. 97 */ 98 void sqliteCheckMemory(void *p, int N){ 99 int *pi = p; 100 int n, i, k; 101 pi -= N_GUARD+1; 102 for(i=0; i<N_GUARD; i++){ 103 assert( pi[i]==0xdead1122 ); 104 } 105 n = pi[N_GUARD]; 106 assert( N>=0 && N<n ); 107 k = (n+sizeof(int)-1)/sizeof(int); 108 for(i=0; i<N_GUARD; i++){ 109 assert( pi[k+N_GUARD+1+i]==0xdead3344 ); 110 } 111 } 112 113 /* 114 ** Free memory previously obtained from sqliteMalloc() 115 */ 116 void sqliteFree_(void *p, char *zFile, int line){ 117 if( p ){ 118 int *pi, i, k, n; 119 pi = p; 120 pi -= N_GUARD+1; 121 sqlite_nFree++; 122 for(i=0; i<N_GUARD; i++){ 123 if( pi[i]!=0xdead1122 ){ 124 fprintf(stderr,"Low-end memory corruption at 0x%x\n", (int)p); 125 return; 126 } 127 } 128 n = pi[N_GUARD]; 129 k = (n+sizeof(int)-1)/sizeof(int); 130 for(i=0; i<N_GUARD; i++){ 131 if( pi[k+N_GUARD+1+i]!=0xdead3344 ){ 132 fprintf(stderr,"High-end memory corruption at 0x%x\n", (int)p); 133 return; 134 } 135 } 136 memset(pi, 0xff, (k+N_GUARD*2+1)*sizeof(int)); 137 #if MEMORY_DEBUG>1 138 fprintf(stderr,"%06d free %d bytes at 0x%x from %s:%d\n", 139 ++memcnt, n, (int)p, zFile,line); 140 #endif 141 free(pi); 142 } 143 } 144 145 /* 146 ** Resize a prior allocation. If p==0, then this routine 147 ** works just like sqliteMalloc(). If n==0, then this routine 148 ** works just like sqliteFree(). 149 */ 150 void *sqliteRealloc_(void *oldP, int n, char *zFile, int line){ 151 int *oldPi, *pi, i, k, oldN, oldK; 152 void *p; 153 if( oldP==0 ){ 154 return sqliteMalloc_(n,1,zFile,line); 155 } 156 if( n==0 ){ 157 sqliteFree_(oldP,zFile,line); 158 return 0; 159 } 160 oldPi = oldP; 161 oldPi -= N_GUARD+1; 162 if( oldPi[0]!=0xdead1122 ){ 163 fprintf(stderr,"Low-end memory corruption in realloc at 0x%x\n", (int)oldP); 164 return 0; 165 } 166 oldN = oldPi[N_GUARD]; 167 oldK = (oldN+sizeof(int)-1)/sizeof(int); 168 for(i=0; i<N_GUARD; i++){ 169 if( oldPi[oldK+N_GUARD+1+i]!=0xdead3344 ){ 170 fprintf(stderr,"High-end memory corruption in realloc at 0x%x\n", 171 (int)oldP); 172 return 0; 173 } 174 } 175 k = (n + sizeof(int) - 1)/sizeof(int); 176 pi = malloc( (k+N_GUARD*2+1)*sizeof(int) ); 177 if( pi==0 ){ 178 sqlite_malloc_failed++; 179 return 0; 180 } 181 for(i=0; i<N_GUARD; i++) pi[i] = 0xdead1122; 182 pi[N_GUARD] = n; 183 for(i=0; i<N_GUARD; i++) pi[k+N_GUARD+1+i] = 0xdead3344; 184 p = &pi[N_GUARD+1]; 185 memcpy(p, oldP, n>oldN ? oldN : n); 186 if( n>oldN ){ 187 memset(&((char*)p)[oldN], 0, n-oldN); 188 } 189 memset(oldPi, 0xab, (oldK+N_GUARD+2)*sizeof(int)); 190 free(oldPi); 191 #if MEMORY_DEBUG>1 192 fprintf(stderr,"%06d realloc %d to %d bytes at 0x%x to 0x%x at %s:%d\n", 193 ++memcnt, oldN, n, (int)oldP, (int)p, zFile, line); 194 #endif 195 return p; 196 } 197 198 /* 199 ** Make a duplicate of a string into memory obtained from malloc() 200 ** Free the original string using sqliteFree(). 201 ** 202 ** This routine is called on all strings that are passed outside of 203 ** the SQLite library. That way clients can free the string using free() 204 ** rather than having to call sqliteFree(). 205 */ 206 void sqliteStrRealloc(char **pz){ 207 char *zNew; 208 if( pz==0 || *pz==0 ) return; 209 zNew = malloc( strlen(*pz) + 1 ); 210 if( zNew==0 ){ 211 sqlite_malloc_failed++; 212 sqliteFree(*pz); 213 *pz = 0; 214 } 215 strcpy(zNew, *pz); 216 sqliteFree(*pz); 217 *pz = zNew; 218 } 219 220 /* 221 ** Make a copy of a string in memory obtained from sqliteMalloc() 222 */ 223 char *sqliteStrDup_(const char *z, char *zFile, int line){ 224 char *zNew; 225 if( z==0 ) return 0; 226 zNew = sqliteMalloc_(strlen(z)+1, 0, zFile, line); 227 if( zNew ) strcpy(zNew, z); 228 return zNew; 229 } 230 char *sqliteStrNDup_(const char *z, int n, char *zFile, int line){ 231 char *zNew; 232 if( z==0 ) return 0; 233 zNew = sqliteMalloc_(n+1, 0, zFile, line); 234 if( zNew ){ 235 memcpy(zNew, z, n); 236 zNew[n] = 0; 237 } 238 return zNew; 239 } 240 #endif /* MEMORY_DEBUG */ 241 242 /* 243 ** The following versions of malloc() and free() are for use in a 244 ** normal build. 245 */ 246 #if !defined(MEMORY_DEBUG) 247 248 /* 249 ** Allocate new memory and set it to zero. Return NULL if 250 ** no memory is available. See also sqliteMallocRaw(). 251 */ 252 void *sqliteMalloc(int n){ 253 void *p; 254 if( (p = malloc(n))==0 ){ 255 if( n>0 ) sqlite_malloc_failed++; 256 }else{ 257 memset(p, 0, n); 258 } 259 return p; 260 } 261 262 /* 263 ** Allocate new memory but do not set it to zero. Return NULL if 264 ** no memory is available. See also sqliteMalloc(). 265 */ 266 void *sqliteMallocRaw(int n){ 267 void *p; 268 if( (p = malloc(n))==0 ){ 269 if( n>0 ) sqlite_malloc_failed++; 270 } 271 return p; 272 } 273 274 /* 275 ** Free memory previously obtained from sqliteMalloc() 276 */ 277 void sqliteFree(void *p){ 278 if( p ){ 279 free(p); 280 } 281 } 282 283 /* 284 ** Resize a prior allocation. If p==0, then this routine 285 ** works just like sqliteMalloc(). If n==0, then this routine 286 ** works just like sqliteFree(). 287 */ 288 void *sqliteRealloc(void *p, int n){ 289 void *p2; 290 if( p==0 ){ 291 return sqliteMalloc(n); 292 } 293 if( n==0 ){ 294 sqliteFree(p); 295 return 0; 296 } 297 p2 = realloc(p, n); 298 if( p2==0 ){ 299 sqlite_malloc_failed++; 300 } 301 return p2; 302 } 303 304 /* 305 ** Make a copy of a string in memory obtained from sqliteMalloc() 306 */ 307 char *sqliteStrDup(const char *z){ 308 char *zNew; 309 if( z==0 ) return 0; 310 zNew = sqliteMallocRaw(strlen(z)+1); 311 if( zNew ) strcpy(zNew, z); 312 return zNew; 313 } 314 char *sqliteStrNDup(const char *z, int n){ 315 char *zNew; 316 if( z==0 ) return 0; 317 zNew = sqliteMallocRaw(n+1); 318 if( zNew ){ 319 memcpy(zNew, z, n); 320 zNew[n] = 0; 321 } 322 return zNew; 323 } 324 #endif /* !defined(MEMORY_DEBUG) */ 325 326 /* 327 ** Create a string from the 2nd and subsequent arguments (up to the 328 ** first NULL argument), store the string in memory obtained from 329 ** sqliteMalloc() and make the pointer indicated by the 1st argument 330 ** point to that string. The 1st argument must either be NULL or 331 ** point to memory obtained from sqliteMalloc(). 332 */ 333 void sqliteSetString(char **pz, const char *zFirst, ...){ 334 va_list ap; 335 int nByte; 336 const char *z; 337 char *zResult; 338 339 if( pz==0 ) return; 340 nByte = strlen(zFirst) + 1; 341 va_start(ap, zFirst); 342 while( (z = va_arg(ap, const char*))!=0 ){ 343 nByte += strlen(z); 344 } 345 va_end(ap); 346 sqliteFree(*pz); 347 *pz = zResult = sqliteMallocRaw( nByte ); 348 if( zResult==0 ){ 349 return; 350 } 351 strcpy(zResult, zFirst); 352 zResult += strlen(zResult); 353 va_start(ap, zFirst); 354 while( (z = va_arg(ap, const char*))!=0 ){ 355 strcpy(zResult, z); 356 zResult += strlen(zResult); 357 } 358 va_end(ap); 359 #ifdef MEMORY_DEBUG 360 #if MEMORY_DEBUG>1 361 fprintf(stderr,"string at 0x%x is %s\n", (int)*pz, *pz); 362 #endif 363 #endif 364 } 365 366 /* 367 ** Works like sqliteSetString, but each string is now followed by 368 ** a length integer which specifies how much of the source string 369 ** to copy (in bytes). -1 means use the whole string. The 1st 370 ** argument must either be NULL or point to memory obtained from 371 ** sqliteMalloc(). 372 */ 373 void sqliteSetNString(char **pz, ...){ 374 va_list ap; 375 int nByte; 376 const char *z; 377 char *zResult; 378 int n; 379 380 if( pz==0 ) return; 381 nByte = 0; 382 va_start(ap, pz); 383 while( (z = va_arg(ap, const char*))!=0 ){ 384 n = va_arg(ap, int); 385 if( n<=0 ) n = strlen(z); 386 nByte += n; 387 } 388 va_end(ap); 389 sqliteFree(*pz); 390 *pz = zResult = sqliteMallocRaw( nByte + 1 ); 391 if( zResult==0 ) return; 392 va_start(ap, pz); 393 while( (z = va_arg(ap, const char*))!=0 ){ 394 n = va_arg(ap, int); 395 if( n<=0 ) n = strlen(z); 396 strncpy(zResult, z, n); 397 zResult += n; 398 } 399 *zResult = 0; 400 #ifdef MEMORY_DEBUG 401 #if MEMORY_DEBUG>1 402 fprintf(stderr,"string at 0x%x is %s\n", (int)*pz, *pz); 403 #endif 404 #endif 405 va_end(ap); 406 } 407 408 /* 409 ** Add an error message to pParse->zErrMsg and increment pParse->nErr. 410 ** The following formatting characters are allowed: 411 ** 412 ** %s Insert a string 413 ** %z A string that should be freed after use 414 ** %d Insert an integer 415 ** %T Insert a token 416 ** %S Insert the first element of a SrcList 417 */ 418 void sqliteErrorMsg(Parse *pParse, const char *zFormat, ...){ 419 va_list ap; 420 pParse->nErr++; 421 sqliteFree(pParse->zErrMsg); 422 va_start(ap, zFormat); 423 pParse->zErrMsg = sqliteVMPrintf(zFormat, ap); 424 va_end(ap); 425 } 426 427 /* 428 ** Convert an SQL-style quoted string into a normal string by removing 429 ** the quote characters. The conversion is done in-place. If the 430 ** input does not begin with a quote character, then this routine 431 ** is a no-op. 432 ** 433 ** 2002-Feb-14: This routine is extended to remove MS-Access style 434 ** brackets from around identifers. For example: "[a-b-c]" becomes 435 ** "a-b-c". 436 */ 437 void sqliteDequote(char *z){ 438 int quote; 439 int i, j; 440 if( z==0 ) return; 441 quote = z[0]; 442 switch( quote ){ 443 case '\'': break; 444 case '"': break; 445 case '[': quote = ']'; break; 446 default: return; 447 } 448 for(i=1, j=0; z[i]; i++){ 449 if( z[i]==quote ){ 450 if( z[i+1]==quote ){ 451 z[j++] = quote; 452 i++; 453 }else{ 454 z[j++] = 0; 455 break; 456 } 457 }else{ 458 z[j++] = z[i]; 459 } 460 } 461 } 462 463 /* An array to map all upper-case characters into their corresponding 464 ** lower-case character. 465 */ 466 static unsigned char UpperToLower[] = { 467 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 468 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 469 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 470 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 97, 98, 99,100,101,102,103, 471 104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121, 472 122, 91, 92, 93, 94, 95, 96, 97, 98, 99,100,101,102,103,104,105,106,107, 473 108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125, 474 126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143, 475 144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161, 476 162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179, 477 180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197, 478 198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215, 479 216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233, 480 234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251, 481 252,253,254,255 482 }; 483 484 /* 485 ** This function computes a hash on the name of a keyword. 486 ** Case is not significant. 487 */ 488 int sqliteHashNoCase(const char *z, int n){ 489 int h = 0; 490 if( n<=0 ) n = strlen(z); 491 while( n > 0 ){ 492 h = (h<<3) ^ h ^ UpperToLower[(unsigned char)*z++]; 493 n--; 494 } 495 return h & 0x7fffffff; 496 } 497 498 /* 499 ** Some systems have stricmp(). Others have strcasecmp(). Because 500 ** there is no consistency, we will define our own. 501 */ 502 int sqliteStrICmp(const char *zLeft, const char *zRight){ 503 register unsigned char *a, *b; 504 a = (unsigned char *)zLeft; 505 b = (unsigned char *)zRight; 506 while( *a!=0 && UpperToLower[*a]==UpperToLower[*b]){ a++; b++; } 507 return UpperToLower[*a] - UpperToLower[*b]; 508 } 509 int sqliteStrNICmp(const char *zLeft, const char *zRight, int N){ 510 register unsigned char *a, *b; 511 a = (unsigned char *)zLeft; 512 b = (unsigned char *)zRight; 513 while( N-- > 0 && *a!=0 && UpperToLower[*a]==UpperToLower[*b]){ a++; b++; } 514 return N<0 ? 0 : UpperToLower[*a] - UpperToLower[*b]; 515 } 516 517 /* 518 ** Return TRUE if z is a pure numeric string. Return FALSE if the 519 ** string contains any character which is not part of a number. 520 ** 521 ** Am empty string is considered non-numeric. 522 */ 523 int sqliteIsNumber(const char *z){ 524 if( *z=='-' || *z=='+' ) z++; 525 if( !isdigit(*z) ){ 526 return 0; 527 } 528 z++; 529 while( isdigit(*z) ){ z++; } 530 if( *z=='.' ){ 531 z++; 532 if( !isdigit(*z) ) return 0; 533 while( isdigit(*z) ){ z++; } 534 } 535 if( *z=='e' || *z=='E' ){ 536 z++; 537 if( *z=='+' || *z=='-' ) z++; 538 if( !isdigit(*z) ) return 0; 539 while( isdigit(*z) ){ z++; } 540 } 541 return *z==0; 542 } 543 544 /* 545 ** The string z[] is an ascii representation of a real number. 546 ** Convert this string to a double. 547 ** 548 ** This routine assumes that z[] really is a valid number. If it 549 ** is not, the result is undefined. 550 ** 551 ** This routine is used instead of the library atof() function because 552 ** the library atof() might want to use "," as the decimal point instead 553 ** of "." depending on how locale is set. But that would cause problems 554 ** for SQL. So this routine always uses "." regardless of locale. 555 */ 556 double sqliteAtoF(const char *z, const char **pzEnd){ 557 int sign = 1; 558 LONGDOUBLE_TYPE v1 = 0.0; 559 if( *z=='-' ){ 560 sign = -1; 561 z++; 562 }else if( *z=='+' ){ 563 z++; 564 } 565 while( isdigit(*z) ){ 566 v1 = v1*10.0 + (*z - '0'); 567 z++; 568 } 569 if( *z=='.' ){ 570 LONGDOUBLE_TYPE divisor = 1.0; 571 z++; 572 while( isdigit(*z) ){ 573 v1 = v1*10.0 + (*z - '0'); 574 divisor *= 10.0; 575 z++; 576 } 577 v1 /= divisor; 578 } 579 if( *z=='e' || *z=='E' ){ 580 int esign = 1; 581 int eval = 0; 582 LONGDOUBLE_TYPE scale = 1.0; 583 z++; 584 if( *z=='-' ){ 585 esign = -1; 586 z++; 587 }else if( *z=='+' ){ 588 z++; 589 } 590 while( isdigit(*z) ){ 591 eval = eval*10 + *z - '0'; 592 z++; 593 } 594 while( eval>=64 ){ scale *= 1.0e+64; eval -= 64; } 595 while( eval>=16 ){ scale *= 1.0e+16; eval -= 16; } 596 while( eval>=4 ){ scale *= 1.0e+4; eval -= 4; } 597 while( eval>=1 ){ scale *= 1.0e+1; eval -= 1; } 598 if( esign<0 ){ 599 v1 /= scale; 600 }else{ 601 v1 *= scale; 602 } 603 } 604 if( pzEnd ) *pzEnd = z; 605 return sign<0 ? -v1 : v1; 606 } 607 608 /* 609 ** The string zNum represents an integer. There might be some other 610 ** information following the integer too, but that part is ignored. 611 ** If the integer that the prefix of zNum represents will fit in a 612 ** 32-bit signed integer, return TRUE. Otherwise return FALSE. 613 ** 614 ** This routine returns FALSE for the string -2147483648 even that 615 ** that number will, in theory fit in a 32-bit integer. But positive 616 ** 2147483648 will not fit in 32 bits. So it seems safer to return 617 ** false. 618 */ 619 int sqliteFitsIn32Bits(const char *zNum){ 620 int i, c; 621 if( *zNum=='-' || *zNum=='+' ) zNum++; 622 for(i=0; (c=zNum[i])>='0' && c<='9'; i++){} 623 return i<10 || (i==10 && memcmp(zNum,"2147483647",10)<=0); 624 } 625 626 /* This comparison routine is what we use for comparison operations 627 ** between numeric values in an SQL expression. "Numeric" is a little 628 ** bit misleading here. What we mean is that the strings have a 629 ** type of "numeric" from the point of view of SQL. The strings 630 ** do not necessarily contain numbers. They could contain text. 631 ** 632 ** If the input strings both look like actual numbers then they 633 ** compare in numerical order. Numerical strings are always less 634 ** than non-numeric strings so if one input string looks like a 635 ** number and the other does not, then the one that looks like 636 ** a number is the smaller. Non-numeric strings compare in 637 ** lexigraphical order (the same order as strcmp()). 638 */ 639 int sqliteCompare(const char *atext, const char *btext){ 640 int result; 641 int isNumA, isNumB; 642 if( atext==0 ){ 643 return -1; 644 }else if( btext==0 ){ 645 return 1; 646 } 647 isNumA = sqliteIsNumber(atext); 648 isNumB = sqliteIsNumber(btext); 649 if( isNumA ){ 650 if( !isNumB ){ 651 result = -1; 652 }else{ 653 double rA, rB; 654 rA = sqliteAtoF(atext, 0); 655 rB = sqliteAtoF(btext, 0); 656 if( rA<rB ){ 657 result = -1; 658 }else if( rA>rB ){ 659 result = +1; 660 }else{ 661 result = 0; 662 } 663 } 664 }else if( isNumB ){ 665 result = +1; 666 }else { 667 result = strcmp(atext, btext); 668 } 669 return result; 670 } 671 672 /* 673 ** This routine is used for sorting. Each key is a list of one or more 674 ** null-terminated elements. The list is terminated by two nulls in 675 ** a row. For example, the following text is a key with three elements 676 ** 677 ** Aone\000Dtwo\000Athree\000\000 678 ** 679 ** All elements begin with one of the characters "+-AD" and end with "\000" 680 ** with zero or more text elements in between. Except, NULL elements 681 ** consist of the special two-character sequence "N\000". 682 ** 683 ** Both arguments will have the same number of elements. This routine 684 ** returns negative, zero, or positive if the first argument is less 685 ** than, equal to, or greater than the first. (Result is a-b). 686 ** 687 ** Each element begins with one of the characters "+", "-", "A", "D". 688 ** This character determines the sort order and collating sequence: 689 ** 690 ** + Sort numerically in ascending order 691 ** - Sort numerically in descending order 692 ** A Sort as strings in ascending order 693 ** D Sort as strings in descending order. 694 ** 695 ** For the "+" and "-" sorting, pure numeric strings (strings for which the 696 ** isNum() function above returns TRUE) always compare less than strings 697 ** that are not pure numerics. Non-numeric strings compare in memcmp() 698 ** order. This is the same sort order as the sqliteCompare() function 699 ** above generates. 700 ** 701 ** The last point is a change from version 2.6.3 to version 2.7.0. In 702 ** version 2.6.3 and earlier, substrings of digits compare in numerical 703 ** and case was used only to break a tie. 704 ** 705 ** Elements that begin with 'A' or 'D' compare in memcmp() order regardless 706 ** of whether or not they look like a number. 707 ** 708 ** Note that the sort order imposed by the rules above is the same 709 ** from the ordering defined by the "<", "<=", ">", and ">=" operators 710 ** of expressions and for indices. This was not the case for version 711 ** 2.6.3 and earlier. 712 */ 713 int sqliteSortCompare(const char *a, const char *b){ 714 int res = 0; 715 int isNumA, isNumB; 716 int dir = 0; 717 718 while( res==0 && *a && *b ){ 719 if( a[0]=='N' || b[0]=='N' ){ 720 if( a[0]==b[0] ){ 721 a += 2; 722 b += 2; 723 continue; 724 } 725 if( a[0]=='N' ){ 726 dir = b[0]; 727 res = -1; 728 }else{ 729 dir = a[0]; 730 res = +1; 731 } 732 break; 733 } 734 assert( a[0]==b[0] ); 735 if( (dir=a[0])=='A' || a[0]=='D' ){ 736 res = strcmp(&a[1],&b[1]); 737 if( res ) break; 738 }else{ 739 isNumA = sqliteIsNumber(&a[1]); 740 isNumB = sqliteIsNumber(&b[1]); 741 if( isNumA ){ 742 double rA, rB; 743 if( !isNumB ){ 744 res = -1; 745 break; 746 } 747 rA = sqliteAtoF(&a[1], 0); 748 rB = sqliteAtoF(&b[1], 0); 749 if( rA<rB ){ 750 res = -1; 751 break; 752 } 753 if( rA>rB ){ 754 res = +1; 755 break; 756 } 757 }else if( isNumB ){ 758 res = +1; 759 break; 760 }else{ 761 res = strcmp(&a[1],&b[1]); 762 if( res ) break; 763 } 764 } 765 a += strlen(&a[1]) + 2; 766 b += strlen(&b[1]) + 2; 767 } 768 if( dir=='-' || dir=='D' ) res = -res; 769 return res; 770 } 771 772 /* 773 ** Some powers of 64. These constants are needed in the 774 ** sqliteRealToSortable() routine below. 775 */ 776 #define _64e3 (64.0 * 64.0 * 64.0) 777 #define _64e4 (64.0 * 64.0 * 64.0 * 64.0) 778 #define _64e15 (_64e3 * _64e4 * _64e4 * _64e4) 779 #define _64e16 (_64e4 * _64e4 * _64e4 * _64e4) 780 #define _64e63 (_64e15 * _64e16 * _64e16 * _64e16) 781 #define _64e64 (_64e16 * _64e16 * _64e16 * _64e16) 782 783 /* 784 ** The following procedure converts a double-precision floating point 785 ** number into a string. The resulting string has the property that 786 ** two such strings comparied using strcmp() or memcmp() will give the 787 ** same results as a numeric comparison of the original floating point 788 ** numbers. 789 ** 790 ** This routine is used to generate database keys from floating point 791 ** numbers such that the keys sort in the same order as the original 792 ** floating point numbers even though the keys are compared using 793 ** memcmp(). 794 ** 795 ** The calling function should have allocated at least 14 characters 796 ** of space for the buffer z[]. 797 */ 798 void sqliteRealToSortable(double r, char *z){ 799 int neg; 800 int exp; 801 int cnt = 0; 802 803 /* This array maps integers between 0 and 63 into base-64 digits. 804 ** The digits must be chosen such at their ASCII codes are increasing. 805 ** This means we can not use the traditional base-64 digit set. */ 806 static const char zDigit[] = 807 "0123456789" 808 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 809 "abcdefghijklmnopqrstuvwxyz" 810 "|~"; 811 if( r<0.0 ){ 812 neg = 1; 813 r = -r; 814 *z++ = '-'; 815 } else { 816 neg = 0; 817 *z++ = '0'; 818 } 819 exp = 0; 820 821 if( r==0.0 ){ 822 exp = -1024; 823 }else if( r<(0.5/64.0) ){ 824 while( r < 0.5/_64e64 && exp > -961 ){ r *= _64e64; exp -= 64; } 825 while( r < 0.5/_64e16 && exp > -1009 ){ r *= _64e16; exp -= 16; } 826 while( r < 0.5/_64e4 && exp > -1021 ){ r *= _64e4; exp -= 4; } 827 while( r < 0.5/64.0 && exp > -1024 ){ r *= 64.0; exp -= 1; } 828 }else if( r>=0.5 ){ 829 while( r >= 0.5*_64e63 && exp < 960 ){ r *= 1.0/_64e64; exp += 64; } 830 while( r >= 0.5*_64e15 && exp < 1008 ){ r *= 1.0/_64e16; exp += 16; } 831 while( r >= 0.5*_64e3 && exp < 1020 ){ r *= 1.0/_64e4; exp += 4; } 832 while( r >= 0.5 && exp < 1023 ){ r *= 1.0/64.0; exp += 1; } 833 } 834 if( neg ){ 835 exp = -exp; 836 r = -r; 837 } 838 exp += 1024; 839 r += 0.5; 840 if( exp<0 ) return; 841 if( exp>=2048 || r>=1.0 ){ 842 strcpy(z, "~~~~~~~~~~~~"); 843 return; 844 } 845 *z++ = zDigit[(exp>>6)&0x3f]; 846 *z++ = zDigit[exp & 0x3f]; 847 while( r>0.0 && cnt<10 ){ 848 int digit; 849 r *= 64.0; 850 digit = (int)r; 851 assert( digit>=0 && digit<64 ); 852 *z++ = zDigit[digit & 0x3f]; 853 r -= digit; 854 cnt++; 855 } 856 *z = 0; 857 } 858 859 #ifdef SQLITE_UTF8 860 /* 861 ** X is a pointer to the first byte of a UTF-8 character. Increment 862 ** X so that it points to the next character. This only works right 863 ** if X points to a well-formed UTF-8 string. 864 */ 865 #define sqliteNextChar(X) while( (0xc0&*++(X))==0x80 ){} 866 #define sqliteCharVal(X) sqlite_utf8_to_int(X) 867 868 #else /* !defined(SQLITE_UTF8) */ 869 /* 870 ** For iso8859 encoding, the next character is just the next byte. 871 */ 872 #define sqliteNextChar(X) (++(X)); 873 #define sqliteCharVal(X) ((int)*(X)) 874 875 #endif /* defined(SQLITE_UTF8) */ 876 877 878 #ifdef SQLITE_UTF8 879 /* 880 ** Convert the UTF-8 character to which z points into a 31-bit 881 ** UCS character. This only works right if z points to a well-formed 882 ** UTF-8 string. 883 */ 884 static int sqlite_utf8_to_int(const unsigned char *z){ 885 int c; 886 static const int initVal[] = { 887 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 888 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 889 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 890 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 891 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 892 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 893 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101, 102, 103, 104, 894 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 895 120, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133, 134, 896 135, 136, 137, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 897 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 160, 161, 162, 163, 164, 898 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 899 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 0, 1, 2, 900 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 901 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 0, 902 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 903 0, 1, 2, 3, 4, 5, 6, 7, 0, 1, 2, 3, 0, 1, 254, 904 255, 905 }; 906 c = initVal[*(z++)]; 907 while( (0xc0&*z)==0x80 ){ 908 c = (c<<6) | (0x3f&*(z++)); 909 } 910 return c; 911 } 912 #endif 913 914 /* 915 ** Compare two UTF-8 strings for equality where the first string can 916 ** potentially be a "glob" expression. Return true (1) if they 917 ** are the same and false (0) if they are different. 918 ** 919 ** Globbing rules: 920 ** 921 ** '*' Matches any sequence of zero or more characters. 922 ** 923 ** '?' Matches exactly one character. 924 ** 925 ** [...] Matches one character from the enclosed list of 926 ** characters. 927 ** 928 ** [^...] Matches one character not in the enclosed list. 929 ** 930 ** With the [...] and [^...] matching, a ']' character can be included 931 ** in the list by making it the first character after '[' or '^'. A 932 ** range of characters can be specified using '-'. Example: 933 ** "[a-z]" matches any single lower-case letter. To match a '-', make 934 ** it the last character in the list. 935 ** 936 ** This routine is usually quick, but can be N**2 in the worst case. 937 ** 938 ** Hints: to match '*' or '?', put them in "[]". Like this: 939 ** 940 ** abc[*]xyz Matches "abc*xyz" only 941 */ 942 int 943 sqliteGlobCompare(const unsigned char *zPattern, const unsigned char *zString){ 944 register int c; 945 int invert; 946 int seen; 947 int c2; 948 949 while( (c = *zPattern)!=0 ){ 950 switch( c ){ 951 case '*': 952 while( (c=zPattern[1]) == '*' || c == '?' ){ 953 if( c=='?' ){ 954 if( *zString==0 ) return 0; 955 sqliteNextChar(zString); 956 } 957 zPattern++; 958 } 959 if( c==0 ) return 1; 960 if( c=='[' ){ 961 while( *zString && sqliteGlobCompare(&zPattern[1],zString)==0 ){ 962 sqliteNextChar(zString); 963 } 964 return *zString!=0; 965 }else{ 966 while( (c2 = *zString)!=0 ){ 967 while( c2 != 0 && c2 != c ){ c2 = *++zString; } 968 if( c2==0 ) return 0; 969 if( sqliteGlobCompare(&zPattern[1],zString) ) return 1; 970 sqliteNextChar(zString); 971 } 972 return 0; 973 } 974 case '?': { 975 if( *zString==0 ) return 0; 976 sqliteNextChar(zString); 977 zPattern++; 978 break; 979 } 980 case '[': { 981 int prior_c = 0; 982 seen = 0; 983 invert = 0; 984 c = sqliteCharVal(zString); 985 if( c==0 ) return 0; 986 c2 = *++zPattern; 987 if( c2=='^' ){ invert = 1; c2 = *++zPattern; } 988 if( c2==']' ){ 989 if( c==']' ) seen = 1; 990 c2 = *++zPattern; 991 } 992 while( (c2 = sqliteCharVal(zPattern))!=0 && c2!=']' ){ 993 if( c2=='-' && zPattern[1]!=']' && zPattern[1]!=0 && prior_c>0 ){ 994 zPattern++; 995 c2 = sqliteCharVal(zPattern); 996 if( c>=prior_c && c<=c2 ) seen = 1; 997 prior_c = 0; 998 }else if( c==c2 ){ 999 seen = 1; 1000 prior_c = c2; 1001 }else{ 1002 prior_c = c2; 1003 } 1004 sqliteNextChar(zPattern); 1005 } 1006 if( c2==0 || (seen ^ invert)==0 ) return 0; 1007 sqliteNextChar(zString); 1008 zPattern++; 1009 break; 1010 } 1011 default: { 1012 if( c != *zString ) return 0; 1013 zPattern++; 1014 zString++; 1015 break; 1016 } 1017 } 1018 } 1019 return *zString==0; 1020 } 1021 1022 /* 1023 ** Compare two UTF-8 strings for equality using the "LIKE" operator of 1024 ** SQL. The '%' character matches any sequence of 0 or more 1025 ** characters and '_' matches any single character. Case is 1026 ** not significant. 1027 ** 1028 ** This routine is just an adaptation of the sqliteGlobCompare() 1029 ** routine above. 1030 */ 1031 int 1032 sqliteLikeCompare(const unsigned char *zPattern, const unsigned char *zString){ 1033 register int c; 1034 int c2; 1035 1036 while( (c = UpperToLower[*zPattern])!=0 ){ 1037 switch( c ){ 1038 case '%': { 1039 while( (c=zPattern[1]) == '%' || c == '_' ){ 1040 if( c=='_' ){ 1041 if( *zString==0 ) return 0; 1042 sqliteNextChar(zString); 1043 } 1044 zPattern++; 1045 } 1046 if( c==0 ) return 1; 1047 c = UpperToLower[c]; 1048 while( (c2=UpperToLower[*zString])!=0 ){ 1049 while( c2 != 0 && c2 != c ){ c2 = UpperToLower[*++zString]; } 1050 if( c2==0 ) return 0; 1051 if( sqliteLikeCompare(&zPattern[1],zString) ) return 1; 1052 sqliteNextChar(zString); 1053 } 1054 return 0; 1055 } 1056 case '_': { 1057 if( *zString==0 ) return 0; 1058 sqliteNextChar(zString); 1059 zPattern++; 1060 break; 1061 } 1062 default: { 1063 if( c != UpperToLower[*zString] ) return 0; 1064 zPattern++; 1065 zString++; 1066 break; 1067 } 1068 } 1069 } 1070 return *zString==0; 1071 } 1072 1073 /* 1074 ** Change the sqlite.magic from SQLITE_MAGIC_OPEN to SQLITE_MAGIC_BUSY. 1075 ** Return an error (non-zero) if the magic was not SQLITE_MAGIC_OPEN 1076 ** when this routine is called. 1077 ** 1078 ** This routine is a attempt to detect if two threads use the 1079 ** same sqlite* pointer at the same time. There is a race 1080 ** condition so it is possible that the error is not detected. 1081 ** But usually the problem will be seen. The result will be an 1082 ** error which can be used to debug the application that is 1083 ** using SQLite incorrectly. 1084 ** 1085 ** Ticket #202: If db->magic is not a valid open value, take care not 1086 ** to modify the db structure at all. It could be that db is a stale 1087 ** pointer. In other words, it could be that there has been a prior 1088 ** call to sqlite_close(db) and db has been deallocated. And we do 1089 ** not want to write into deallocated memory. 1090 */ 1091 int sqliteSafetyOn(sqlite *db){ 1092 if( db->magic==SQLITE_MAGIC_OPEN ){ 1093 db->magic = SQLITE_MAGIC_BUSY; 1094 return 0; 1095 }else if( db->magic==SQLITE_MAGIC_BUSY || db->magic==SQLITE_MAGIC_ERROR 1096 || db->want_to_close ){ 1097 db->magic = SQLITE_MAGIC_ERROR; 1098 db->flags |= SQLITE_Interrupt; 1099 } 1100 return 1; 1101 } 1102 1103 /* 1104 ** Change the magic from SQLITE_MAGIC_BUSY to SQLITE_MAGIC_OPEN. 1105 ** Return an error (non-zero) if the magic was not SQLITE_MAGIC_BUSY 1106 ** when this routine is called. 1107 */ 1108 int sqliteSafetyOff(sqlite *db){ 1109 if( db->magic==SQLITE_MAGIC_BUSY ){ 1110 db->magic = SQLITE_MAGIC_OPEN; 1111 return 0; 1112 }else if( db->magic==SQLITE_MAGIC_OPEN || db->magic==SQLITE_MAGIC_ERROR 1113 || db->want_to_close ){ 1114 db->magic = SQLITE_MAGIC_ERROR; 1115 db->flags |= SQLITE_Interrupt; 1116 } 1117 return 1; 1118 } 1119 1120 /* 1121 ** Check to make sure we are not currently executing an sqlite_exec(). 1122 ** If we are currently in an sqlite_exec(), return true and set 1123 ** sqlite.magic to SQLITE_MAGIC_ERROR. This will cause a complete 1124 ** shutdown of the database. 1125 ** 1126 ** This routine is used to try to detect when API routines are called 1127 ** at the wrong time or in the wrong sequence. 1128 */ 1129 int sqliteSafetyCheck(sqlite *db){ 1130 if( db->pVdbe!=0 ){ 1131 db->magic = SQLITE_MAGIC_ERROR; 1132 return 1; 1133 } 1134 return 0; 1135 } 1136