reactos/lib/sdk/crt/string/scanf.h

   1 /*
   2  * general implementation of scanf used by scanf, sscanf, fscanf,
   3  * _cscanf, wscanf, swscanf and fwscanf
   4  *
   5  * Copyright 1996,1998 Marcus Meissner
   6  * Copyright 1996 Jukka Iivonen
   7  * Copyright 1997,2000, 2003 Uwe Bonnes
   8  * Copyright 2000 Jon Griffiths
   9  * Copyright 2002 Daniel Gudbjartsson
  10  *
  11  * This library is free software; you can redistribute it and/or
  12  * modify it under the terms of the GNU Lesser General Public
  13  * License as published by the Free Software Foundation; either
  14  * version 2.1 of the License, or (at your option) any later version.
  15  *
  16  * This library is distributed in the hope that it will be useful,
  17  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  18  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  19  * Lesser General Public License for more details.
  20  *
  21  * You should have received a copy of the GNU Lesser General Public
  22  * License along with this library; if not, write to the Free Software
  23  * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
  24  */
  25
  26 #ifdef WIDE_SCANF
  27 #define _CHAR_ wchar_t
  28 #define _EOF_ WEOF
  29 #define _EOF_RET WEOF
  30 #define _ISSPACE_(c) iswspace(c)
  31 #define _ISDIGIT_(c) iswdigit(c)
  32 #define _WIDE2SUPPORTED_(c) c /* No conversion needed (wide to wide) */
  33 #define _CHAR2SUPPORTED_(c) c /* FIXME: convert char to wide char */
  34 #define _CHAR2DIGIT_(c, base) wchar2digit((c), (base))
  35 #define _BITMAPSIZE_ 256*256
  36 #else /* WIDE_SCANF */
  37 #define _CHAR_ char
  38 #define _EOF_ EOF
  39 #define _EOF_RET EOF
  40 #define _ISSPACE_(c) isspace((unsigned char)(c))
  41 #define _ISDIGIT_(c) isdigit((unsigned char)(c))
  42 #define _WIDE2SUPPORTED_(c) c /* FIXME: convert wide char to char */
  43 #define _CHAR2SUPPORTED_(c) c /* No conversion needed (char to char) */
  44 #define _CHAR2DIGIT_(c, base) char2digit((c), (base))
  45 #define _BITMAPSIZE_ 256
  46 #endif /* WIDE_SCANF */
  47
  48 #ifdef CONSOLE
  49 #define _GETC_(file) (consumed++, _getch())
  50 #define _UNGETC_(nch, file) do { _ungetch(nch); consumed--; } while(0)
  51 #define _FUNCTION_ int vcscanf(const char *format, va_list ap)
  52 #else
  53 #ifdef STRING
  54 #undef _EOF_
  55 #define _EOF_ 0
  56 #define _GETC_(file) (consumed++, *file++)
  57 #define _UNGETC_(nch, file) do { file--; consumed--; } while(0)
  58 #ifdef WIDE_SCANF
  59 #define _FUNCTION_ int vswscanf(const wchar_t *file, const wchar_t *format, va_list ap)
  60 #else /* WIDE_SCANF */
  61 #define _FUNCTION_ int vsscanf(const char *file, const char *format, va_list ap)
  62 #endif /* WIDE_SCANF */
  63 #else /* STRING */
  64 #ifdef WIDE_SCANF
  65 #define _GETC_(file) (consumed++, fgetwc(file))
  66 #define _UNGETC_(nch, file) do { ungetwc(nch, file); consumed--; } while(0)
  67 #define _FUNCTION_ int vfwscanf(FILE* file, const wchar_t *format, va_list ap)
  68 #else /* WIDE_SCANF */
  69 #define _GETC_(file) (consumed++, fgetc(file))
  70 #define _UNGETC_(nch, file) do { ungetc(nch, file); consumed--; } while(0)
  71 #define _FUNCTION_ int vfscanf(FILE* file, const char *format, va_list ap)
  72 #endif /* WIDE_SCANF */
  73 #endif /* STRING */
  74 #endif /* CONSOLE */
  75
  76 _FUNCTION_ {
  77     int rd = 0, consumed = 0;
  78     int nch;
  79     if (!*format) return 0;
  80 #ifndef WIDE_SCANF
  81 #ifdef CONSOLE
  82     TRACE("(%s):\n", debugstr_a(format));
  83 #else /* CONSOLE */
  84 #ifdef STRING
  85     TRACE("%s (%s)\n", file, debugstr_a(format));
  86 #else /* STRING */
  87     TRACE("%p (%s)\n", file, debugstr_a(format));
  88 #endif /* STRING */
  89 #endif /* CONSOLE */
  90 #endif /* WIDE_SCANF */
  91     nch = _GETC_(file);
  92     if (nch == _EOF_) return _EOF_RET;
  93
  94     while (*format) {
  95         /* a whitespace character in the format string causes scanf to read,
  96          * but not store, all consecutive white-space characters in the input
  97          * up to the next non-white-space character.  One white space character
  98          * in the input matches any number (including zero) and combination of
  99          * white-space characters in the input. */
 100         if (_ISSPACE_(*format)) {
 101             /* skip whitespace */
 102             while ((nch!=_EOF_) && _ISSPACE_(nch))
 103                 nch = _GETC_(file);
 104         }
 105         /* a format specification causes scanf to read and convert characters
 106          * in the input into values of a specified type.  The value is assigned
 107          * to an argument in the argument list.  Format specifications have
 108          * the form %[*][width][{h | l | I64 | L}]type */
 109         else if (*format == '%') {
 110             int st = 0; int suppress = 0; int width = 0;
 111             int base;
 112             int h_prefix = 0;
 113             int l_prefix = 0;
 114             int L_prefix = 0;
 115             int w_prefix = 0;
 116             int prefix_finished = 0;
 117             int I64_prefix = 0;
 118             format++;
 119             /* look for leading asterisk, which means 'suppress assignment of
 120              * this field'. */
 121             if (*format=='*') {
 122                 format++;
 123                 suppress=1;
 124             }
 125             /* look for width specification */
 126             while (_ISDIGIT_(*format)) {
 127                 width*=10;
 128                 width+=*format++ - '0';
 129             }
 130             if (width==0) width=-1; /* no width spec seen */
 131             /* read prefix (if any) */
 132             while (!prefix_finished) {
 133                 switch(*format) {
 134                 case 'h': h_prefix++; break;
 135                 case 'l':
 136                     if(*(format+1) == 'l') {
 137                         I64_prefix = 1;
 138                         format++;
 139                     }
 140                     l_prefix = 1;
 141                     break;
 142                 case 'w': w_prefix = 1; break;
 143                 case 'L': L_prefix = 1; break;
 144                 case 'I':
 145                     if (*(format + 1) == '6' &&
 146                         *(format + 2) == '4') {
 147                         I64_prefix = 1;
 148                         format += 2;
 149                     }
 150                     break;
 151                 default:
 152                     prefix_finished = 1;
 153                 }
 154                 if (!prefix_finished) format++;
 155             }
 156             /* read type */
 157             switch(*format) {
 158             case 'p':
 159             case 'P': /* pointer. */
 160                 if (sizeof(void *) == sizeof(LONGLONG)) I64_prefix = 1;
 161                 /* fall through */
 162             case 'x':
 163             case 'X': /* hexadecimal integer. */
 164                 base = 16;
 165                 goto number;
 166             case 'o': /* octal integer */
 167                 base = 8;
 168                 goto number;
 169             case 'u': /* unsigned decimal integer */
 170                 base = 10;
 171                 goto number;
 172             case 'd': /* signed decimal integer */
 173                 base = 10;
 174                 goto number;
 175             case 'i': /* generic integer */
 176                 base = 0;
 177             number: {
 178                     /* read an integer */
 179                     __int64 cur = 0;
 180                     int negative = 0;
 181                     int seendigit=0;
 182                     /* skip initial whitespace */
 183                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 184                         nch = _GETC_(file);
 185                     /* get sign */
 186                     if (nch == '-' || nch == '+') {
 187                         negative = (nch=='-');
 188                         nch = _GETC_(file);
 189                         if (width>0) width--;
 190                     }
 191                     /* look for leading indication of base */
 192                     if (width!=0 && nch == '0' && *format != 'p' && *format != 'P') {
 193                         nch = _GETC_(file);
 194                         if (width>0) width--;
 195                         seendigit=1;
 196                         if (width!=0 && (nch=='x' || nch=='X')) {
 197                             if (base==0)
 198                                 base=16;
 199                             if (base==16) {
 200                                 nch = _GETC_(file);
 201                                 if (width>0) width--;
 202                                 seendigit=0;
 203                             }
 204                         } else if (base==0)
 205                             base = 8;
 206                     }
 207                     /* format %i without indication of base */
 208                     if (base==0)
 209                         base = 10;
 210                     /* throw away leading zeros */
 211                     while (width!=0 && nch=='0') {
 212                         nch = _GETC_(file);
 213                         if (width>0) width--;
 214                         seendigit=1;
 215                     }
 216                     if (width!=0 && _CHAR2DIGIT_(nch, base)!=-1) {
 217                         cur = _CHAR2DIGIT_(nch, base);
 218                         nch = _GETC_(file);
 219                         if (width>0) width--;
 220                         seendigit=1;
 221                     }
 222                     /* read until no more digits */
 223                     while (width!=0 && (nch!=_EOF_) && _CHAR2DIGIT_(nch, base)!=-1) {
 224                         cur = cur*base + _CHAR2DIGIT_(nch, base);
 225                         nch = _GETC_(file);
 226                         if (width>0) width--;
 227                         seendigit=1;
 228                     }
 229                     /* okay, done! */
 230                     if (!seendigit) break; /* not a valid number */
 231                     st = 1;
 232                     if (!suppress) {
 233 #define _SET_NUMBER_(type) *va_arg(ap, type*) = (type)(negative ? -cur : cur)
 234                         if (I64_prefix) _SET_NUMBER_(LONGLONG);
 235                         else if (l_prefix) _SET_NUMBER_(LONG);
 236                         else if (h_prefix == 1) _SET_NUMBER_(short int);
 237                         else _SET_NUMBER_(int);
 238                     }
 239                 }
 240                 break;
 241             case 'e':
 242             case 'E':
 243             case 'f':
 244             case 'g':
 245             case 'G': { /* read a float */
 246                     long double cur = 0;
 247                     int negative = 0;
 248                     /* skip initial whitespace */
 249                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 250                         nch = _GETC_(file);
 251                     /* get sign. */
 252                     if (nch == '-' || nch == '+') {
 253                         negative = (nch=='-');
 254                         if (width>0) width--;
 255                         if (width==0) break;
 256                         nch = _GETC_(file);
 257                     }
 258                     /* get first digit. */
 259                     if ('.' != nch) {
 260                       if (!_ISDIGIT_(nch)) break;
 261                       cur = (nch - '0');
 262                       nch = _GETC_(file);
 263                       if (width>0) width--;
 264                       /* read until no more digits */
 265                       while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 266                         cur = cur*10 + (nch - '0');
 267                         nch = _GETC_(file);
 268                         if (width>0) width--;
 269                       }
 270                     } else {
 271                       cur = 0; /* MaxPayneDemo Fix: .8 -> 0.8 */
 272                     }
 273                     /* handle decimals */
 274                     if (width!=0 && nch == '.') {
 275                         long double dec = 1;
 276                         nch = _GETC_(file);
 277                         if (width>0) width--;
 278                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 279                             dec /= 10;
 280                             cur += dec * (nch - '0');
 281                             nch = _GETC_(file);
 282                             if (width>0) width--;
 283                         }
 284                     }
 285                     /* handle exponent */
 286                     if (width!=0 && (nch == 'e' || nch == 'E')) {
 287                         int exponent = 0, negexp = 0;
 288                         double expcnt, shift;
 289                         nch = _GETC_(file);
 290                         if (width>0) width--;
 291                         /* possible sign on the exponent */
 292                         if (width!=0 && (nch=='+' || nch=='-')) {
 293                             negexp = (nch=='-');
 294                             nch = _GETC_(file);
 295                             if (width>0) width--;
 296                         }
 297                         /* exponent digits */
 298                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 299                             exponent *= 10;
 300                             exponent += (nch - '0');
 301                             nch = _GETC_(file);
 302                             if (width>0) width--;
 303                         }
 304                         /* update 'cur' with this exponent. */
 305                         expcnt = 10;
 306                         shift = 1.0;
 307                         while (exponent!=0) {
 308                             if (exponent&1)
 309                                 shift *= expcnt;
 310                             exponent/=2;
 311                             expcnt=expcnt*expcnt;
 312                         }
 313                         cur = (negexp ? cur / shift : cur * shift);
 314                     }
 315                     st = 1;
 316                     if (!suppress) {
 317                         if (L_prefix) _SET_NUMBER_(long double);
 318                         else if (l_prefix) _SET_NUMBER_(double);
 319                         else _SET_NUMBER_(float);
 320                     }
 321                 }
 322                 break;
 323                 /* According to msdn,
 324                  * 's' reads a character string in a call to fscanf
 325                  * and 'S' a wide character string and vice versa in a
 326                  * call to fwscanf. The 'h', 'w' and 'l' prefixes override
 327                  * this behaviour. 'h' forces reading char * but 'l' and 'w'
 328                  * force reading WCHAR. */
 329             case 's':
 330                     if (w_prefix || l_prefix) goto widecharstring;
 331                     else if (h_prefix) goto charstring;
 332 #ifdef WIDE_SCANF
 333                     else goto widecharstring;
 334 #else /* WIDE_SCANF */
 335                     else goto charstring;
 336 #endif /* WIDE_SCANF */
 337             case 'S':
 338                     if (w_prefix || l_prefix) goto widecharstring;
 339                     else if (h_prefix) goto charstring;
 340 #ifdef WIDE_SCANF
 341                     else goto charstring;
 342 #else /* WIDE_SCANF */
 343                     else goto widecharstring;
 344 #endif /* WIDE_SCANF */
 345             charstring: { /* read a word into a char */
 346                     char *sptr = suppress ? NULL : va_arg(ap, char*);
 347                     /* skip initial whitespace */
 348                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 349                         nch = _GETC_(file);
 350                     /* read until whitespace */
 351                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 352                         if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 353                         st++;
 354                         nch = _GETC_(file);
 355                         if (width>0) width--;
 356                     }
 357                     /* terminate */
 358                     if (!suppress) *sptr = 0;
 359                 }
 360                 break;
 361             widecharstring: { /* read a word into a wchar_t* */
 362                     wchar_t *sptr = suppress ? NULL : va_arg(ap, wchar_t*);
 363                     /* skip initial whitespace */
 364                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 365                         nch = _GETC_(file);
 366                     /* read until whitespace */
 367                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 368                         if (!suppress) *sptr++ = _WIDE2SUPPORTED_(nch);
 369                         st++;
 370                         nch = _GETC_(file);
 371                         if (width>0) width--;
 372                     }
 373                     /* terminate */
 374                     if (!suppress) *sptr = 0;
 375                 }
 376                 break;
 377             /* 'c' and 'C work analogously to 's' and 'S' as described
 378              * above */
 379             case 'c':
 380                     if (w_prefix || l_prefix) goto widecharacter;
 381                     else if (h_prefix) goto character;
 382 #ifdef WIDE_SCANF
 383                     else goto widecharacter;
 384 #else /* WIDE_SCANF */
 385                     else goto character;
 386 #endif /* WIDE_SCANF */
 387             case 'C':
 388                     if (w_prefix || l_prefix) goto widecharacter;
 389                     else if (h_prefix) goto character;
 390 #ifdef WIDE_SCANF
 391                     else goto character;
 392 #else /* WIDE_SCANF */
 393                     else goto widecharacter;
 394 #endif /* WIDE_SCANF */
 395           character: { /* read single character into char */
 396                     char *str = suppress ? NULL : va_arg(ap, char*);
 397                     if (width == -1) width = 1;
 398                     while ((width != 0) && (nch != _EOF_))
 399                     {
 400                         if (!suppress) *str++ = _CHAR2SUPPORTED_(nch);
 401                         st++;
 402                         width--;
 403                         nch = _GETC_(file);
 404                     }
 405                 }
 406                 break;
 407           widecharacter: { /* read single character into a wchar_t */
 408                     wchar_t *str = suppress ? NULL : va_arg(ap, wchar_t*);
 409                     if (width == -1) width = 1;
 410                     while ((width != 0) && (nch != _EOF_))
 411                     {
 412                         if (!suppress) *str++ = _WIDE2SUPPORTED_(nch);
 413                         st++;
 414                         width--;
 415                         nch = _GETC_(file);
 416                     }
 417                 }
 418                 break;
 419             case 'n': {
 420                     if (!suppress) {
 421                         int*n = va_arg(ap, int*);
 422                         *n = consumed - 1;
 423                     }
 424                     /* This is an odd one: according to the standard,
 425                      * "Execution of a %n directive does not increment the
 426                      * assignment count returned at the completion of
 427                      * execution" even if it wasn't suppressed with the
 428                      * '*' flag.  The Corrigendum to the standard seems
 429                      * to contradict this (comment out the assignment to
 430                      * suppress below if you want to implement these
 431                      * alternate semantics) but the windows program I'm
 432                      * looking at expects the behavior I've coded here
 433                      * (which happens to be what glibc does as well).
 434                      */
 435                     suppress = 1;
 436                     st = 1;
 437                 }
 438                 break;
 439             case '[': {
 440                     _CHAR_ *str = suppress ? NULL : va_arg(ap, _CHAR_*);
 441                     _CHAR_ *sptr = str;
 442                     RTL_BITMAP bitMask;
 443                     ULONG *Mask;
 444                     int invert = 0; /* Set if we are NOT to find the chars */
 445
 446             /* Init our bitmap */
 447 #ifdef _LIBCNT_
 448             Mask = RtlAllocateHeap(RtlGetProcessHeap(), HEAP_ZERO_MEMORY, _BITMAPSIZE_/8);
 449 #else
 450             Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, _BITMAPSIZE_/8);
 451 #endif
 452             RtlInitializeBitMap(&bitMask, Mask, _BITMAPSIZE_);
 453
 454                     /* Read the format */
 455                     format++;
 456                     if(*format == '^') {
 457                         invert = 1;
 458                         format++;
 459                     }
 460                     if(*format == ']') {
 461                         RtlSetBits(&bitMask, ']', 1);
 462                         format++;
 463                     }
 464                     while(*format && (*format != ']')) {
 465                         /* According to msdn:
 466                          * "Note that %[a-z] and %[z-a] are interpreted as equivalent to %[abcde...z]." */
 467                         if((*format == '-') && (*(format + 1) != ']')) {
 468                             if ((*(format - 1)) < *(format + 1))
 469                                 RtlSetBits(&bitMask, *(format - 1) +1 , *(format + 1) - *(format - 1));
 470                             else
 471                                 RtlSetBits(&bitMask, *(format + 1)    , *(format - 1) - *(format + 1));
 472                             format++;
 473                         } else
 474                             RtlSetBits(&bitMask, *format, 1);
 475                         format++;
 476                     }
 477                     /* read until char is not suitable */
 478                     while ((width != 0) && (nch != _EOF_)) {
 479                         if(!invert) {
 480                             if(RtlAreBitsSet(&bitMask, nch, 1)) {
 481                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 482                             } else
 483                                 break;
 484                         } else {
 485                             if(RtlAreBitsClear(&bitMask, nch, 1)) {
 486                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 487                             } else
 488                                 break;
 489                         }
 490                         st++;
 491                         nch = _GETC_(file);
 492                         if (width>0) width--;
 493                     }
 494                     /* terminate */
 495                     if (!suppress) *sptr = 0;
 496 #ifdef _LIBCNT_
 497             RtlFreeHeap(RtlGetProcessHeap(), 0, Mask);
 498 #else
 499             HeapFree(GetProcessHeap(), 0, Mask);
 500 #endif
 501                 }
 502                 break;
 503             default:
 504                 /* From spec: "if a percent sign is followed by a character
 505                  * that has no meaning as a format-control character, that
 506                  * character and the following characters are treated as
 507                  * an ordinary sequence of characters, that is, a sequence
 508                  * of characters that must match the input.  For example,
 509                  * to specify that a percent-sign character is to be input,
 510                  * use %%." */
 511                 while ((nch!=_EOF_) && _ISSPACE_(nch))
 512                     nch = _GETC_(file);
 513                 if (nch==*format) {
 514                     suppress = 1; /* whoops no field to be read */
 515                     st = 1; /* but we got what we expected */
 516                     nch = _GETC_(file);
 517                 }
 518                 break;
 519             }
 520             if (st && !suppress) rd++;
 521             else if (!st) break;
 522         }
 523         /* a non-white-space character causes scanf to read, but not store,
 524          * a matching non-white-space character. */
 525         else {
 526             /* check for character match */
 527             if (nch == *format) {
 528                 nch = _GETC_(file);
 529             } else break;
 530         }
 531         format++;
 532     }
 533     if (nch!=_EOF_) {
 534         _UNGETC_(nch, file);
 535     }
 536
 537     TRACE("returning %d\n", rd);
 538     return rd;
 539 }
 540
 541 #undef _CHAR_
 542 #undef _EOF_
 543 #undef _EOF_RET
 544 #undef _ISSPACE_
 545 #undef _ISDIGIT_
 546 #undef _CHAR2SUPPORTED_
 547 #undef _WIDE2SUPPORTED_
 548 #undef _CHAR2DIGIT_
 549 #undef _GETC_
 550 #undef _UNGETC_
 551 #undef _FUNCTION_
 552 #undef _BITMAPSIZE_