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