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., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, 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 _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_ MSVCRT_EOF
  39 #define _EOF_RET MSVCRT_EOF
  40 #define _ISSPACE_(c) isspace(c)
  41 #define _ISDIGIT_(c) isdigit(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_ static int MSVCRT_vcscanf(const char *format, __ms_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_ static int MSVCRT_vswscanf(const MSVCRT_wchar_t *file, const MSVCRT_wchar_t *format, __ms_va_list ap)
  60 #else /* WIDE_SCANF */
  61 #define _FUNCTION_ static int MSVCRT_vsscanf(const char *file, const char *format, __ms_va_list ap)
  62 #endif /* WIDE_SCANF */
  63 #else /* STRING */
  64 #ifdef WIDE_SCANF
  65 #define _GETC_(file) (consumed++, MSVCRT_fgetwc(file))
  66 #define _UNGETC_(nch, file) do { MSVCRT_ungetwc(nch, file); consumed--; } while(0)
  67 #define _FUNCTION_ static int MSVCRT_vfwscanf(MSVCRT_FILE* file, const MSVCRT_wchar_t *format, __ms_va_list ap)
  68 #else /* WIDE_SCANF */
  69 #define _GETC_(file) (consumed++, MSVCRT_fgetc(file))
  70 #define _UNGETC_(nch, file) do { MSVCRT_ungetc(nch, file); consumed--; } while(0)
  71 #define _FUNCTION_ static int MSVCRT_vfscanf(MSVCRT_FILE* file, const char *format, __ms_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 = 1; break;
 135                 case 'l': l_prefix = 1; break;
 136                 case 'w': w_prefix = 1; break;
 137                 case 'L': L_prefix = 1; break;
 138                 case 'I':
 139                     if (*(format + 1) == '6' &&
 140                         *(format + 2) == '4') {
 141                         I64_prefix = 1;
 142                         format += 2;
 143                     }
 144                     break;
 145                 default:
 146                     prefix_finished = 1;
 147                 }
 148                 if (!prefix_finished) format++;
 149             }
 150             /* read type */
 151             switch(*format) {
 152             case 'p':
 153             case 'P': /* pointer. */
 154             case 'x':
 155             case 'X': /* hexadecimal integer. */
 156                 base = 16;
 157                 goto number;
 158             case 'o': /* octal integer */
 159                 base = 8;
 160                 goto number;
 161             case 'u': /* unsigned decimal integer */
 162                 base = 10;
 163                 goto number;
 164             case 'd': /* signed decimal integer */
 165                 base = 10;
 166                 goto number;
 167             case 'i': /* generic integer */
 168                 base = 0;
 169             number: {
 170                     /* read an integer */
 171                     ULONGLONG cur = 0;
 172                     int negative = 0;
 173                     int seendigit=0;
 174                     /* skip initial whitespace */
 175                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 176                         nch = _GETC_(file);
 177                     /* get sign */
 178                     if (nch == '-' || nch == '+') {
 179                         negative = (nch=='-');
 180                         nch = _GETC_(file);
 181                         if (width>0) width--;
 182                     }
 183                     /* look for leading indication of base */
 184                     if (width!=0 && nch == '0' && *format != 'p' && *format != 'P') {
 185                         nch = _GETC_(file);
 186                         if (width>0) width--;
 187                         seendigit=1;
 188                         if (width!=0 && (nch=='x' || nch=='X')) {
 189                             if (base==0)
 190                                 base=16;
 191                             if (base==16) {
 192                                 nch = _GETC_(file);
 193                                 if (width>0) width--;
 194                                 seendigit=0;
 195                             }
 196                         } else if (base==0)
 197                             base = 8;
 198                     }
 199                     /* format %i without indication of base */
 200                     if (base==0)
 201                         base = 10;
 202                     /* throw away leading zeros */
 203                     while (width!=0 && nch=='0') {
 204                         nch = _GETC_(file);
 205                         if (width>0) width--;
 206                         seendigit=1;
 207                     }
 208                     if (width!=0 && _CHAR2DIGIT_(nch, base)!=-1) {
 209                         cur = _CHAR2DIGIT_(nch, base);
 210                         nch = _GETC_(file);
 211                         if (width>0) width--;
 212                         seendigit=1;
 213                     }
 214                     /* read until no more digits */
 215                     while (width!=0 && (nch!=_EOF_) && _CHAR2DIGIT_(nch, base)!=-1) {
 216                         cur = cur*base + _CHAR2DIGIT_(nch, base);
 217                         nch = _GETC_(file);
 218                         if (width>0) width--;
 219                         seendigit=1;
 220                     }
 221                     /* okay, done! */
 222                     if (!seendigit) break; /* not a valid number */
 223                     st = 1;
 224                     if (!suppress) {
 225 #define _SET_NUMBER_(type) *va_arg(ap, type*) = negative ? -cur : cur
 226                         if (I64_prefix) _SET_NUMBER_(LONGLONG);
 227                         else if (l_prefix) _SET_NUMBER_(long int);
 228                         else if (h_prefix) _SET_NUMBER_(short int);
 229                         else _SET_NUMBER_(int);
 230                     }
 231                 }
 232                 break;
 233             case 'e':
 234             case 'E':
 235             case 'f':
 236             case 'g':
 237             case 'G': { /* read a float */
 238                     long double cur = 0;
 239                     int negative = 0;
 240                     /* skip initial whitespace */
 241                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 242                         nch = _GETC_(file);
 243                     /* get sign. */
 244                     if (nch == '-' || nch == '+') {
 245                         negative = (nch=='-');
 246                         if (width>0) width--;
 247                         if (width==0) break;
 248                         nch = _GETC_(file);
 249                     }
 250                     /* get first digit. */
 251                     if ('.' != nch) {
 252                       if (!_ISDIGIT_(nch)) break;
 253                       cur = (nch - '0');
 254                       nch = _GETC_(file);
 255                       if (width>0) width--;
 256                       /* read until no more digits */
 257                       while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 258                         cur = cur*10 + (nch - '0');
 259                         nch = _GETC_(file);
 260                         if (width>0) width--;
 261                       }
 262                     } else {
 263                       cur = 0; /* MaxPayneDemo Fix: .8 -> 0.8 */
 264                     }
 265                     /* handle decimals */
 266                     if (width!=0 && nch == '.') {
 267                         float dec = 1;
 268                         nch = _GETC_(file);
 269                         if (width>0) width--;
 270                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 271                             dec /= 10;
 272                             cur += dec * (nch - '0');
 273                             nch = _GETC_(file);
 274                             if (width>0) width--;
 275                         }
 276                     }
 277                     /* handle exponent */
 278                     if (width!=0 && (nch == 'e' || nch == 'E')) {
 279                         int exponent = 0, negexp = 0;
 280                         float expcnt;
 281                         nch = _GETC_(file);
 282                         if (width>0) width--;
 283                         /* possible sign on the exponent */
 284                         if (width!=0 && (nch=='+' || nch=='-')) {
 285                             negexp = (nch=='-');
 286                             nch = _GETC_(file);
 287                             if (width>0) width--;
 288                         }
 289                         /* exponent digits */
 290                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 291                             exponent *= 10;
 292                             exponent += (nch - '0');
 293                             nch = _GETC_(file);
 294                             if (width>0) width--;
 295                         }
 296                         /* update 'cur' with this exponent. */
 297                         expcnt =  negexp ? .1 : 10;
 298                         while (exponent!=0) {
 299                             if (exponent&1)
 300                                 cur*=expcnt;
 301                             exponent/=2;
 302                             expcnt=expcnt*expcnt;
 303                         }
 304                     }
 305                     st = 1;
 306                     if (!suppress) {
 307                         if (L_prefix) _SET_NUMBER_(long double);
 308                         else if (l_prefix) _SET_NUMBER_(double);
 309                         else _SET_NUMBER_(float);
 310                     }
 311                 }
 312                 break;
 313                 /* According to msdn,
 314                  * 's' reads a character string in a call to fscanf
 315                  * and 'S' a wide character string and vice versa in a
 316                  * call to fwscanf. The 'h', 'w' and 'l' prefixes override
 317                  * this behaviour. 'h' forces reading char * but 'l' and 'w'
 318                  * force reading WCHAR. */
 319             case 's':
 320                     if (w_prefix || l_prefix) goto widecharstring;
 321                     else if (h_prefix) goto charstring;
 322 #ifdef WIDE_SCANF
 323                     else goto widecharstring;
 324 #else /* WIDE_SCANF */
 325                     else goto charstring;
 326 #endif /* WIDE_SCANF */
 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 charstring;
 332 #else /* WIDE_SCANF */
 333                     else goto widecharstring;
 334 #endif /* WIDE_SCANF */
 335             charstring: { /* read a word into a char */
 336                     char *sptr = suppress ? NULL : va_arg(ap, char*);
 337                     /* skip initial whitespace */
 338                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 339                         nch = _GETC_(file);
 340                     /* read until whitespace */
 341                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 342                         if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 343                         st++;
 344                         nch = _GETC_(file);
 345                         if (width>0) width--;
 346                     }
 347                     /* terminate */
 348                     if (!suppress) *sptr = 0;
 349                 }
 350                 break;
 351             widecharstring: { /* read a word into a wchar_t* */
 352                     MSVCRT_wchar_t *sptr = suppress ? NULL : va_arg(ap, MSVCRT_wchar_t*);
 353                     /* skip initial whitespace */
 354                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 355                         nch = _GETC_(file);
 356                     /* read until whitespace */
 357                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 358                         if (!suppress) *sptr++ = _WIDE2SUPPORTED_(nch);
 359                         st++;
 360                         nch = _GETC_(file);
 361                         if (width>0) width--;
 362                     }
 363                     /* terminate */
 364                     if (!suppress) *sptr = 0;
 365                 }
 366                 break;
 367             /* 'c' and 'C work analogously to 's' and 'S' as described
 368              * above */
 369             case 'c':
 370                     if (w_prefix || l_prefix) goto widecharacter;
 371                     else if (h_prefix) goto character;
 372 #ifdef WIDE_SCANF
 373                     else goto widecharacter;
 374 #else /* WIDE_SCANF */
 375                     else goto character;
 376 #endif /* WIDE_SCANF */
 377             case 'C':
 378                     if (w_prefix || l_prefix) goto widecharacter;
 379                     else if (h_prefix) goto character;
 380 #ifdef WIDE_SCANF
 381                     else goto character;
 382 #else /* WIDE_SCANF */
 383                     else goto widecharacter;
 384 #endif /* WIDE_SCANF */
 385           character: { /* read single character into char */
 386                     char *str = suppress ? NULL : va_arg(ap, char*);
 387                     if (width == -1) width = 1;
 388                     while ((width != 0) && (nch != _EOF_))
 389                     {
 390                         if (!suppress) *str++ = _CHAR2SUPPORTED_(nch);
 391                         st++;
 392                         width--;
 393                         nch = _GETC_(file);
 394                     }
 395                 }
 396                 break;
 397           widecharacter: { /* read single character into a wchar_t */
 398                     MSVCRT_wchar_t *str = suppress ? NULL : va_arg(ap, MSVCRT_wchar_t*);
 399                     if (width == -1) width = 1;
 400                     while ((width != 0) && (nch != _EOF_))
 401                     {
 402                         if (!suppress) *str++ = _WIDE2SUPPORTED_(nch);
 403                         st++;
 404                         width--;
 405                         nch = _GETC_(file);
 406                     }
 407                 }
 408                 break;
 409             case 'n': {
 410                     if (!suppress) {
 411                         int*n = va_arg(ap, int*);
 412                         *n = consumed - 1;
 413                     }
 414                     /* This is an odd one: according to the standard,
 415                      * "Execution of a %n directive does not increment the
 416                      * assignment count returned at the completion of
 417                      * execution" even if it wasn't suppressed with the
 418                      * '*' flag.  The Corrigendum to the standard seems
 419                      * to contradict this (comment out the assignment to
 420                      * suppress below if you want to implement these
 421                      * alternate semantics) but the windows program I'm
 422                      * looking at expects the behavior I've coded here
 423                      * (which happens to be what glibc does as well).
 424                      */
 425                     suppress = 1;
 426                     st = 1;
 427                 }
 428                 break;
 429             case '[': {
 430                     _CHAR_ *str = suppress ? NULL : va_arg(ap, _CHAR_*);
 431                     _CHAR_ *sptr = str;
 432                     RTL_BITMAP bitMask;
 433                     ULONG *Mask;
 434                     int invert = 0; /* Set if we are NOT to find the chars */
 435
 436                     /* Init our bitmap */
 437                     Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, _BITMAPSIZE_/8);
 438                     RtlInitializeBitMap(&bitMask, Mask, _BITMAPSIZE_);
 439
 440                     /* Read the format */
 441                     format++;
 442                     if(*format == '^') {
 443                         invert = 1;
 444                         format++;
 445                     }
 446                     if(*format == ']') {
 447                         RtlSetBits(&bitMask, ']', 1);
 448                         format++;
 449                     }
 450                     while(*format && (*format != ']')) {
 451                         /* According to msdn:
 452                          * "Note that %[a-z] and %[z-a] are interpreted as equivalent to %[abcde...z]." */
 453                         if((*format == '-') && (*(format + 1) != ']')) {
 454                             if ((*(format - 1)) < *(format + 1))
 455                                 RtlSetBits(&bitMask, *(format - 1) +1 , *(format + 1) - *(format - 1));
 456                             else
 457                                 RtlSetBits(&bitMask, *(format + 1)    , *(format - 1) - *(format + 1));
 458                             format++;
 459                         } else
 460                             RtlSetBits(&bitMask, *format, 1);
 461                         format++;
 462                     }
 463                     /* read until char is not suitable */
 464                     while ((width != 0) && (nch != _EOF_)) {
 465                         if(!invert) {
 466                             if(RtlAreBitsSet(&bitMask, nch, 1)) {
 467                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 468                             } else
 469                                 break;
 470                         } else {
 471                             if(RtlAreBitsClear(&bitMask, nch, 1)) {
 472                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 473                             } else
 474                                 break;
 475                         }
 476                         st++;
 477                         nch = _GETC_(file);
 478                         if (width>0) width--;
 479                     }
 480                     /* terminate */
 481                     if (!suppress) *sptr = 0;
 482                     HeapFree(GetProcessHeap(), 0, Mask);
 483                 }
 484                 break;
 485             default:
 486                 /* From spec: "if a percent sign is followed by a character
 487                  * that has no meaning as a format-control character, that
 488                  * character and the following characters are treated as
 489                  * an ordinary sequence of characters, that is, a sequence
 490                  * of characters that must match the input.  For example,
 491                  * to specify that a percent-sign character is to be input,
 492                  * use %%." */
 493                 while ((nch!=_EOF_) && _ISSPACE_(nch))
 494                     nch = _GETC_(file);
 495                 if (nch==*format) {
 496                     suppress = 1; /* whoops no field to be read */
 497                     st = 1; /* but we got what we expected */
 498                     nch = _GETC_(file);
 499                 }
 500                 break;
 501             }
 502             if (st && !suppress) rd++;
 503             else if (!st) break;
 504         }
 505         /* a non-white-space character causes scanf to read, but not store,
 506          * a matching non-white-space character. */
 507         else {
 508             /* check for character match */
 509             if (nch == *format) {
 510                 nch = _GETC_(file);
 511             } else break;
 512         }
 513         format++;
 514     }
 515     if (nch!=_EOF_) {
 516         _UNGETC_(nch, file);
 517     }
 518     TRACE("returning %d\n", rd);
 519     return rd;
 520 }
 521
 522 #undef _CHAR_
 523 #undef _EOF_
 524 #undef _EOF_RET
 525 #undef _ISSPACE_
 526 #undef _ISDIGIT_
 527 #undef _CHAR2SUPPORTED_
 528 #undef _WIDE2SUPPORTED_
 529 #undef _CHAR2DIGIT_
 530 #undef _GETC_
 531 #undef _UNGETC_
 532 #undef _FUNCTION_
 533 #undef _BITMAPSIZE_