dlls/msvcrt/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., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
  24  */
  25
  26 #ifdef WIDE_SCANF
  27 #define _CHAR_ MSVCRT_wchar_t
  28 #define _EOF_ MSVCRT_WEOF
  29 #define _EOF_RET MSVCRT_WEOF
  30 #define _ISSPACE_(c) MSVCRT_iswspace(c)
  31 #define _ISDIGIT_(c) MSVCRT_iswdigit(c)
  32 #define _CONVERT_(c) c /*** FIXME ***/
  33 #define _CHAR2DIGIT_(c, base) wchar2digit((c), (base))
  34 #else /* WIDE_SCANF */
  35 #define _CHAR_ char
  36 #define _EOF_ MSVCRT_EOF
  37 #define _EOF_RET MSVCRT_EOF
  38 #define _ISSPACE_(c) isspace(c)
  39 #define _ISDIGIT_(c) isdigit(c)
  40 #define _CONVERT_(c) c /*** FIXME ***/
  41 #define _CHAR2DIGIT_(c, base) char2digit((c), (base))
  42 #endif /* WIDE_SCANF */
  43
  44 #ifdef CONSOLE
  45 #define _GETC_(file) (consumed++, _getch())
  46 #define _UNGETC_(nch, file) do { _ungetch(nch); consumed--; } while(0)
  47 #define _FUNCTION_ _cscanf(const _CHAR_ *format, ...)
  48 #else
  49 #ifdef STRING
  50 #undef _EOF_
  51 #define _EOF_ 0
  52 #define _GETC_(file) (consumed++, *file++)
  53 #define _UNGETC_(nch, file) do { file--; consumed--; } while(0)
  54 #ifdef WIDE_SCANF
  55 #define _FUNCTION_ MSVCRT_swscanf(const MSVCRT_wchar_t *file, const MSVCRT_wchar_t *format, ...)
  56 #else /* WIDE_SCANF */
  57 #define _FUNCTION_ MSVCRT_sscanf(const char *file, const char *format, ...)
  58 #endif /* WIDE_SCANF */
  59 #else /* STRING */
  60 #ifdef WIDE_SCANF
  61 #define _GETC_(file) (consumed++, MSVCRT_fgetwc(file))
  62 #define _UNGETC_(nch, file) do { MSVCRT_ungetwc(nch, file); consumed--; } while(0)
  63 #define _FUNCTION_ MSVCRT_fwscanf(MSVCRT_FILE* file, const MSVCRT_wchar_t *format, ...)
  64 #else /* WIDE_SCANF */
  65 #define _GETC_(file) (consumed++, MSVCRT_fgetc(file))
  66 #define _UNGETC_(nch, file) do { MSVCRT_ungetc(nch, file); consumed--; } while(0)
  67 #define _FUNCTION_ MSVCRT_fscanf(MSVCRT_FILE* file, const char *format, ...)
  68 #endif /* WIDE_SCANF */
  69 #endif /* STRING */
  70 #endif /* CONSOLE */
  71
  72 /*********************************************************************
  73  * Implemented based on
  74  * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/html/_crt_format_specification_fields_.2d_.scanf_and_wscanf_functions.asp
  75  * Extended by C. Scott Ananian <cananian@alumni.princeton.edu> to handle
  76  * more types of format spec.
  77  */
  78 int _FUNCTION_ {
  79     int rd = 0, consumed = 0;
  80     int nch;
  81     va_list ap;
  82     if (!*format) return 0;
  83 #ifndef WIDE_SCANF
  84 #ifdef CONSOLE
  85     WARN("(%s): semi-stub\n", debugstr_a(format));
  86 #else /* CONSOLE */
  87 #ifdef STRING
  88     WARN("%s (%s): semi-stub\n", file, debugstr_a(format));
  89 #else /* STRING */
  90     WARN("%p (%s): semi-stub\n", file, debugstr_a(format));
  91 #endif /* STRING */
  92 #endif /* CONSOLE */
  93 #endif /* WIDE_SCANF */
  94     nch = _GETC_(file);
  95     if (nch == _EOF_) return _EOF_RET;
  96
  97     va_start(ap, format);
  98     while (*format) {
  99         /* a whitespace character in the format string causes scanf to read,
 100          * but not store, all consecutive white-space characters in the input
 101          * up to the next non-white-space character.  One white space character
 102          * in the input matches any number (including zero) and combination of
 103          * white-space characters in the input. */
 104         if (_ISSPACE_(*format)) {
 105             /* skip whitespace */
 106             while ((nch!=_EOF_) && _ISSPACE_(nch))
 107                 nch = _GETC_(file);
 108         }
 109         /* a format specification causes scanf to read and convert characters
 110          * in the input into values of a specified type.  The value is assigned
 111          * to an argument in the argument list.  Format specifications have
 112          * the form %[*][width][{h | l | I64 | L}]type */
 113         else if (*format == '%') {
 114             int st = 0; int suppress = 0; int width = 0;
 115             int base, number_signed;
 116             int h_prefix = 0;
 117             int l_prefix = 0;
 118             int L_prefix = 0;
 119             int w_prefix = 0;
 120             int prefix_finished = 0;
 121             /* int I64_prefix = 0; */
 122             format++;
 123             /* look for leading asterisk, which means 'suppress assignment of
 124              * this field'. */
 125             if (*format=='*') {
 126                 format++;
 127                 suppress=1;
 128             }
 129             /* look for width specification */
 130             while (_ISDIGIT_(*format)) {
 131                 width*=10;
 132                 width+=*format++ - '0';
 133             }
 134             if (width==0) width=-1; /* no width spec seen */
 135             /* read prefix (if any) */
 136             while (!prefix_finished) {
 137                 switch(*format) {
 138                 case 'h': h_prefix = 1; break;
 139                 case 'l': l_prefix = 1; break;
 140                 case 'w': w_prefix = 1; break;
 141                 case 'L': L_prefix = 1; break;
 142                 case 'I':
 143                     if (*(format + 1) == '6' &&
 144                         *(format + 2) == '4') {
 145                         /* I64_prefix = 1; */
 146                         format += 2;
 147                         FIXME("I64 prefix currently not implemented in fscanf/fwscanf");
 148                     }
 149                     break;
 150                 default:
 151                     prefix_finished = 1;
 152                 }
 153                 if (!prefix_finished) format++;
 154             }
 155             /* read type */
 156             switch(*format) {
 157             case 'x':
 158             case 'X': /* hexadecimal integer. */
 159                 base = 16; number_signed = 0;
 160                 goto number;
 161             case 'o': /* octal integer */
 162                 base = 8; number_signed = 0;
 163                 goto number;
 164             case 'u': /* unsigned decimal integer */
 165                 base = 10; number_signed = 0;
 166                 goto number;
 167             case 'd': /* signed decimal integer */
 168                 base = 10; number_signed = 1;
 169                 goto number;
 170             case 'i': /* generic integer */
 171                 base = 0; number_signed = 1;
 172             number: {
 173                     /* read an integer */
 174                     long unsigned int cur = 0;
 175                     int negative = 0;
 176                     int seendigit=0;
 177                     /* skip initial whitespace */
 178                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 179                         nch = _GETC_(file);
 180                     /* get sign */
 181                     if (number_signed && (nch == '-' ||
 182                                           nch == '+')) {
 183                         negative = (nch=='-');
 184                         nch = _GETC_(file);
 185                         if (width>0) width--;
 186                     }
 187                     /* look for leading indication of base */
 188                     if (width!=0 && nch == '0') {
 189                         nch = _GETC_(file);
 190                         if (width>0) width--;
 191                         seendigit=1;
 192                         if (width!=0 && (nch=='x' || nch=='X')) {
 193                             if (base==0)
 194                                 base=16;
 195                             if (base==16) {
 196                                 nch = _GETC_(file);
 197                                 if (width>0) width--;
 198                                 seendigit=0;
 199                             }
 200                         } else if (base==0)
 201                             base = 8;
 202                     }
 203                     /* throw away leading zeros */
 204                     while (width!=0 && nch=='0') {
 205                         nch = _GETC_(file);
 206                         if (width>0) width--;
 207                         seendigit=1;
 208                     }
 209                     if (width!=0 && _CHAR2DIGIT_(nch, base)!=-1) {
 210                         cur = _CHAR2DIGIT_(nch, base);
 211                         nch = _GETC_(file);
 212                         if (width>0) width--;
 213                         seendigit=1;
 214                     }
 215                     /* read until no more digits */
 216                     while (width!=0 && (nch!=_EOF_) && _CHAR2DIGIT_(nch, base)!=-1) {
 217                         cur = cur*base + _CHAR2DIGIT_(nch, base);
 218                         nch = _GETC_(file);
 219                         if (width>0) width--;
 220                         seendigit=1;
 221                     }
 222                     /* okay, done! */
 223                     if (!seendigit) break; /* not a valid number */
 224                     st = 1;
 225                     if (!suppress) {
 226 #define _SET_NUMBER_(type) *va_arg(ap, type*) = negative ? -cur : cur
 227                         if (number_signed) {
 228                             if (l_prefix) _SET_NUMBER_(long int);
 229                             else if (h_prefix) _SET_NUMBER_(short int);
 230                             else _SET_NUMBER_(int);
 231                         } else {
 232                             if (negative) {
 233                                 WARN("Dropping sign in reading a negative number into an unsigned value");
 234                                 negative = 0;
 235                             }
 236                             if (l_prefix) _SET_NUMBER_(unsigned long int);
 237                             else if (h_prefix)
 238                                 _SET_NUMBER_(unsigned short int);
 239                             else _SET_NUMBER_(unsigned int);
 240                         }
 241                     }
 242                 }
 243                 break;
 244             case 'e':
 245             case 'E':
 246             case 'f':
 247             case 'g':
 248             case 'G': { /* read a float */
 249                     long double cur = 0;
 250                     int negative = 0;
 251                     /* skip initial whitespace */
 252                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 253                         nch = _GETC_(file);
 254                     /* get sign. */
 255                     if (nch == '-' || nch == '+') {
 256                         negative = (nch=='-');
 257                         if (width>0) width--;
 258                         if (width==0) break;
 259                         nch = _GETC_(file);
 260                     }
 261                     /* get first digit. */
 262                     if (!_ISDIGIT_(nch)) break;
 263                     cur = (nch - '0');
 264                     nch = _GETC_(file);
 265                     if (width>0) width--;
 266                     /* read until no more digits */
 267                     while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 268                         cur = cur*10 + (nch - '0');
 269                         nch = _GETC_(file);
 270                         if (width>0) width--;
 271                     }
 272                     /* handle decimals */
 273                     if (width!=0 && nch == '.') {
 274                         float dec = 1;
 275                         nch = _GETC_(file);
 276                         if (width>0) width--;
 277                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 278                             dec /= 10;
 279                             cur += dec * (nch - '0');
 280                             nch = _GETC_(file);
 281                             if (width>0) width--;
 282                         }
 283                     }
 284                     /* handle exponent */
 285                     if (width!=0 && (nch == 'e' || nch == 'E')) {
 286                         int exponent = 0, negexp = 0;
 287                         float expcnt;
 288                         nch = _GETC_(file);
 289                         if (width>0) width--;
 290                         /* possible sign on the exponent */
 291                         if (width!=0 && (nch=='+' || nch=='-')) {
 292                             negexp = (nch=='-');
 293                             nch = _GETC_(file);
 294                             if (width>0) width--;
 295                         }
 296                         /* exponent digits */
 297                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 298                             exponent *= 10;
 299                             exponent += (nch - '0');
 300                             nch = _GETC_(file);
 301                             if (width>0) width--;
 302                         }
 303                         /* update 'cur' with this exponent. */
 304                         expcnt =  negexp ? .1 : 10;
 305                         while (exponent!=0) {
 306                             if (exponent&1)
 307                                 cur*=expcnt;
 308                             exponent/=2;
 309                             expcnt=expcnt*expcnt;
 310                         }
 311                     }
 312                     st = 1;
 313                     if (!suppress) {
 314                         if (L_prefix) _SET_NUMBER_(long double);
 315                         else if (l_prefix) _SET_NUMBER_(double);
 316                         else _SET_NUMBER_(float);
 317                     }
 318                 }
 319                 break;
 320                 /* According to
 321                  * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclib/html/_crt_scanf_type_field_characters.asp
 322                  * 's' reads a character string in a call to fscanf
 323                  * and 'S' a wide character string and vice versa in a
 324                  * call to fwscanf. The 'h', 'w' and 'l' prefixes override
 325                  * this behaviour. 'h' forces reading char * but 'l' and 'w'
 326                  * force reading WCHAR. */
 327             case 's':
 328                     if (w_prefix || l_prefix) goto widecharstring;
 329                     else if (h_prefix) goto charstring;
 330 #ifdef WIDE_SCANF
 331                     else goto widecharstring;
 332 #else /* WIDE_SCANF */
 333                     else goto charstring;
 334 #endif /* WIDE_SCANF */
 335             case 'S':
 336                     if (w_prefix || l_prefix) goto widecharstring;
 337                     else if (h_prefix) goto charstring;
 338 #ifdef WIDE_SCANF
 339                     else goto charstring;
 340 #else /* WIDE_SCANF */
 341                     else goto widecharstring;
 342 #endif /* WIDE_SCANF */
 343             charstring: { /* read a word into a char */
 344                     char*str = suppress ? NULL : va_arg(ap, char*);
 345                     char*sptr = str;
 346                     /* skip initial whitespace */
 347                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 348                         nch = _GETC_(file);
 349                     /* read until whitespace */
 350                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 351 #ifdef WIDE_SCANF
 352                         if (!suppress) *sptr++ = _CONVERT_(nch);
 353 #else /* WIDE_SCANF */
 354                         if (!suppress) *sptr++ = nch;
 355 #endif /* WIDE_SCANF */
 356                         st++;
 357                         nch = _GETC_(file);
 358                         if (width>0) width--;
 359                     }
 360                     /* terminate */
 361                     if (!suppress) *sptr = 0;
 362                 }
 363                 break;
 364             widecharstring: { /* read a word into a wchar_t* */
 365                     MSVCRT_wchar_t*str =
 366                         suppress ? NULL : va_arg(ap, MSVCRT_wchar_t*);
 367                     MSVCRT_wchar_t*sptr = str;
 368                     /* skip initial whitespace */
 369                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 370                         nch = _GETC_(file);
 371                     /* read until whitespace */
 372                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 373 #ifdef WIDE_SCANF
 374                         if (!suppress) *sptr++ = nch;
 375 #else /* WIDE_SCANF */
 376                         if (!suppress) *sptr++ = _CONVERT_(nch);
 377 #endif /* WIDE_SCANF */
 378                         st++;
 379                         nch = _GETC_(file);
 380                         if (width>0) width--;
 381                     }
 382                     /* terminate */
 383                     if (!suppress) *sptr = 0;
 384                 }
 385                 break;
 386             /* 'c' and 'C work analogously to 's' and 'S' as described
 387              * above */
 388             case 'c':
 389                     if (w_prefix || l_prefix) goto widecharacter;
 390                     else if (h_prefix) goto character;
 391 #ifdef WIDE_SCANF
 392                     else goto widecharacter;
 393 #else /* WIDE_SCANF */
 394                     else goto character;
 395 #endif /* WIDE_SCANF */
 396             case 'C':
 397                     if (w_prefix || l_prefix) goto widecharacter;
 398                     else if (h_prefix) goto character;
 399 #ifdef WIDE_SCANF
 400                     else goto character;
 401 #else /* WIDE_SCANF */
 402                     else goto widecharacter;
 403 #endif /* WIDE_SCANF */
 404           character: { /* read single character into char */
 405                     if (!suppress) {
 406                         char*c = va_arg(ap, char*);
 407 #ifdef WIDE_SCANF
 408                         *c = _CONVERT_(nch);
 409 #else /* WIDE_SCANF */
 410                         *c = nch;
 411 #endif /* WIDE_SCANF */
 412                     }
 413                     st = 1;
 414                     nch = _GETC_(file);
 415                 }
 416                 break;
 417           widecharacter: {
 418                     if (!suppress) { /* read single character into a wchar_t */
 419                         MSVCRT_wchar_t*c = va_arg(ap, MSVCRT_wchar_t*);
 420 #ifdef WIDE_SCANF
 421                         *c = nch;
 422 #else /* WIDE_SCANF */
 423                         *c = _CONVERT_(nch);
 424 #endif /* WIDE_SCANF */
 425                     }
 426                     nch = _GETC_(file);
 427                     st = 1;
 428                 }
 429                 break;
 430             case 'n': {
 431                     if (!suppress) {
 432                         int*n = va_arg(ap, int*);
 433                         *n = consumed - (nch!=_EOF_);
 434                     }
 435                 }
 436                 break;
 437             case '[': {
 438                     _CHAR_ *str = suppress ? NULL : va_arg(ap, _CHAR_*);
 439                     _CHAR_ *sptr = str;
 440                     RTL_BITMAP bitMask;
 441                     LPBYTE Mask;
 442                     int invert = 0; /* Set if we are NOT to find the chars */
 443
 444                     /* Init our bitmap */
 445 #ifdef WIDE_SCANF
 446                     Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, 65536/8);
 447                     RtlInitializeBitMap(&bitMask, Mask, 65536);
 448 #else /* WIDE_SCANF */
 449                     Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, 256/8);
 450                     RtlInitializeBitMap(&bitMask, Mask, 256);
 451 #endif /* WIDE_SCANF */
 452
 453                     /* Read the format */
 454                     format++;
 455                     if(*format == '^') {
 456                         invert = 1;
 457                         format++;
 458                     }
 459                     if(*format == ']') {
 460                         RtlSetBits(&bitMask, ']', 1);
 461                         format++;
 462                     }
 463                     while(*format && (*format != ']')) {
 464                         if((*format == '-') && (*(format + 1) != ']')) {
 465                             int n = 0;
 466                             for(;(n + *(format - 1)) < *(format + 1); n++)
 467                                 RtlSetBits(&bitMask, n + *(format - 1), 1);
 468                             format++;
 469                         }
 470                         RtlSetBits(&bitMask, *format, 1);
 471                         format++;
 472                     }
 473                     /* read until char is not suitable */
 474                     while ((width != 0) && (nch != _EOF_)) {
 475                         if(!invert) {
 476                             if(RtlAreBitsSet(&bitMask, nch, 1)) {
 477 #ifdef WIDE_SCANF
 478                                 if (!suppress) *sptr++ = _CONVERT_(nch);
 479 #else /* WIDE_SCANF */
 480                                 if (!suppress) *sptr++ = nch;
 481 #endif /* WIDE_SCANF */
 482                             } else
 483                                 break;
 484                         } else {
 485                             if(RtlAreBitsClear(&bitMask, nch, 1)) {
 486 #ifdef WIDE_SCANF
 487                                 if (!suppress) *sptr++ = _CONVERT_(nch);
 488 #else /* WIDE_SCANF */
 489                                 if (!suppress) *sptr++ = nch;
 490 #endif /* WIDE_SCANF */
 491                             } else
 492                                 break;
 493                         }
 494                         st++;
 495                         nch = _GETC_(file);
 496                         if (width>0) width--;
 497                     }
 498                     /* terminate */
 499                     if (!suppress) *sptr = 0;
 500                     HeapFree(GetProcessHeap(), 0, Mask);
 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     va_end(ap);
 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 _CONVERT_
 547 #undef _CHAR2DIGIT_
 548 #undef _GETC_
 549 #undef _UNGETC_
 550 #undef _FUNCTION_