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, 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, va_list ap)
  60 #else /* WIDE_SCANF */
  61 #define _FUNCTION_ static int MSVCRT_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++, 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, 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, va_list ap)
  72 #endif /* WIDE_SCANF */
  73 #endif /* STRING */
  74 #endif /* CONSOLE */
  75
  76 /*********************************************************************
  77  * Implemented based on
  78  * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/html/_crt_format_specification_fields_.2d_.scanf_and_wscanf_functions.asp
  79  * Extended by C. Scott Ananian <cananian@alumni.princeton.edu> to handle
  80  * more types of format spec.
  81  */
  82 _FUNCTION_ {
  83     int rd = 0, consumed = 0;
  84     int nch;
  85     if (!*format) return 0;
  86 #ifndef WIDE_SCANF
  87 #ifdef CONSOLE
  88     TRACE("(%s): \n", debugstr_a(format));
  89 #else /* CONSOLE */
  90 #ifdef STRING
  91     TRACE("%s (%s)\n", file, debugstr_a(format));
  92 #else /* STRING */
  93     TRACE("%p (%s)\n", file, debugstr_a(format));
  94 #endif /* STRING */
  95 #endif /* CONSOLE */
  96 #endif /* WIDE_SCANF */
  97     nch = _GETC_(file);
  98     if (nch == _EOF_) return _EOF_RET;
  99
 100     while (*format) {
 101         /* a whitespace character in the format string causes scanf to read,
 102          * but not store, all consecutive white-space characters in the input
 103          * up to the next non-white-space character.  One white space character
 104          * in the input matches any number (including zero) and combination of
 105          * white-space characters in the input. */
 106         if (_ISSPACE_(*format)) {
 107             /* skip whitespace */
 108             while ((nch!=_EOF_) && _ISSPACE_(nch))
 109                 nch = _GETC_(file);
 110         }
 111         /* a format specification causes scanf to read and convert characters
 112          * in the input into values of a specified type.  The value is assigned
 113          * to an argument in the argument list.  Format specifications have
 114          * the form %[*][width][{h | l | I64 | L}]type */
 115         else if (*format == '%') {
 116             int st = 0; int suppress = 0; int width = 0;
 117             int base;
 118             int h_prefix = 0;
 119             int l_prefix = 0;
 120             int L_prefix = 0;
 121             int w_prefix = 0;
 122             int prefix_finished = 0;
 123             int I64_prefix = 0;
 124             format++;
 125             /* look for leading asterisk, which means 'suppress assignment of
 126              * this field'. */
 127             if (*format=='*') {
 128                 format++;
 129                 suppress=1;
 130             }
 131             /* look for width specification */
 132             while (_ISDIGIT_(*format)) {
 133                 width*=10;
 134                 width+=*format++ - '0';
 135             }
 136             if (width==0) width=-1; /* no width spec seen */
 137             /* read prefix (if any) */
 138             while (!prefix_finished) {
 139                 switch(*format) {
 140                 case 'h': h_prefix = 1; break;
 141                 case 'l': l_prefix = 1; 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 'x':
 159             case 'X': /* hexadecimal integer. */
 160                 base = 16;
 161                 goto number;
 162             case 'o': /* octal integer */
 163                 base = 8;
 164                 goto number;
 165             case 'u': /* unsigned decimal integer */
 166                 base = 10;
 167                 goto number;
 168             case 'd': /* signed decimal integer */
 169                 base = 10;
 170                 goto number;
 171             case 'i': /* generic integer */
 172                 base = 10;
 173             number: {
 174                     /* read an integer */
 175                     ULONGLONG cur = 0;
 176                     int negative = 0;
 177                     int seendigit=0;
 178                     /* skip initial whitespace */
 179                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 180                         nch = _GETC_(file);
 181                     /* get sign */
 182                     if (nch == '-' || 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 (I64_prefix) _SET_NUMBER_(LONGLONG);
 228                         else if (l_prefix) _SET_NUMBER_(long int);
 229                         else if (h_prefix) _SET_NUMBER_(short int);
 230                         else _SET_NUMBER_(int);
 231                     }
 232                 }
 233                 break;
 234             case 'e':
 235             case 'E':
 236             case 'f':
 237             case 'g':
 238             case 'G': { /* read a float */
 239                     long double cur = 0;
 240                     int negative = 0;
 241                     /* skip initial whitespace */
 242                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 243                         nch = _GETC_(file);
 244                     /* get sign. */
 245                     if (nch == '-' || nch == '+') {
 246                         negative = (nch=='-');
 247                         if (width>0) width--;
 248                         if (width==0) break;
 249                         nch = _GETC_(file);
 250                     }
 251                     /* get first digit. */
 252                     if ('.' != nch) {
 253                       if (!_ISDIGIT_(nch)) break;
 254                       cur = (nch - '0');
 255                       nch = _GETC_(file);
 256                       if (width>0) width--;
 257                       /* read until no more digits */
 258                       while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 259                         cur = cur*10 + (nch - '0');
 260                         nch = _GETC_(file);
 261                         if (width>0) width--;
 262                       }
 263                     } else {
 264                       cur = 0; /* MaxPayneDemo Fix: .8 -> 0.8 */
 265                     }
 266                     /* handle decimals */
 267                     if (width!=0 && nch == '.') {
 268                         float dec = 1;
 269                         nch = _GETC_(file);
 270                         if (width>0) width--;
 271                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 272                             dec /= 10;
 273                             cur += dec * (nch - '0');
 274                             nch = _GETC_(file);
 275                             if (width>0) width--;
 276                         }
 277                     }
 278                     /* handle exponent */
 279                     if (width!=0 && (nch == 'e' || nch == 'E')) {
 280                         int exponent = 0, negexp = 0;
 281                         float expcnt;
 282                         nch = _GETC_(file);
 283                         if (width>0) width--;
 284                         /* possible sign on the exponent */
 285                         if (width!=0 && (nch=='+' || nch=='-')) {
 286                             negexp = (nch=='-');
 287                             nch = _GETC_(file);
 288                             if (width>0) width--;
 289                         }
 290                         /* exponent digits */
 291                         while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
 292                             exponent *= 10;
 293                             exponent += (nch - '0');
 294                             nch = _GETC_(file);
 295                             if (width>0) width--;
 296                         }
 297                         /* update 'cur' with this exponent. */
 298                         expcnt =  negexp ? .1 : 10;
 299                         while (exponent!=0) {
 300                             if (exponent&1)
 301                                 cur*=expcnt;
 302                             exponent/=2;
 303                             expcnt=expcnt*expcnt;
 304                         }
 305                     }
 306                     st = 1;
 307                     if (!suppress) {
 308                         if (L_prefix) _SET_NUMBER_(long double);
 309                         else if (l_prefix) _SET_NUMBER_(double);
 310                         else _SET_NUMBER_(float);
 311                     }
 312                 }
 313                 break;
 314                 /* According to
 315                  * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclib/html/_crt_scanf_type_field_characters.asp
 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*str = suppress ? NULL : va_arg(ap, char*);
 339                     char*sptr = str;
 340                     /* skip initial whitespace */
 341                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 342                         nch = _GETC_(file);
 343                     /* read until whitespace */
 344                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 345                         if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 346                         st++;
 347                         nch = _GETC_(file);
 348                         if (width>0) width--;
 349                     }
 350                     /* terminate */
 351                     if (!suppress) *sptr = 0;
 352                 }
 353                 break;
 354             widecharstring: { /* read a word into a wchar_t* */
 355                     MSVCRT_wchar_t*str =
 356                         suppress ? NULL : va_arg(ap, MSVCRT_wchar_t*);
 357                     MSVCRT_wchar_t*sptr = str;
 358                     /* skip initial whitespace */
 359                     while ((nch!=_EOF_) && _ISSPACE_(nch))
 360                         nch = _GETC_(file);
 361                     /* read until whitespace */
 362                     while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
 363                         if (!suppress) *sptr++ = _WIDE2SUPPORTED_(nch);
 364                         st++;
 365                         nch = _GETC_(file);
 366                         if (width>0) width--;
 367                     }
 368                     /* terminate */
 369                     if (!suppress) *sptr = 0;
 370                 }
 371                 break;
 372             /* 'c' and 'C work analogously to 's' and 'S' as described
 373              * above */
 374             case 'c':
 375                     if (w_prefix || l_prefix) goto widecharacter;
 376                     else if (h_prefix) goto character;
 377 #ifdef WIDE_SCANF
 378                     else goto widecharacter;
 379 #else /* WIDE_SCANF */
 380                     else goto character;
 381 #endif /* WIDE_SCANF */
 382             case 'C':
 383                     if (w_prefix || l_prefix) goto widecharacter;
 384                     else if (h_prefix) goto character;
 385 #ifdef WIDE_SCANF
 386                     else goto character;
 387 #else /* WIDE_SCANF */
 388                     else goto widecharacter;
 389 #endif /* WIDE_SCANF */
 390           character: { /* read single character into char */
 391                     if (nch!=_EOF_) {
 392                         if (!suppress) {
 393                             char*c = va_arg(ap, char*);
 394                             *c = _CHAR2SUPPORTED_(nch);
 395                         }
 396                         st = 1;
 397                         nch = _GETC_(file);
 398                     }
 399                 }
 400                 break;
 401           widecharacter: { /* read single character into a wchar_t */
 402                     if (nch!=_EOF_) {
 403                         if (!suppress) {
 404                             MSVCRT_wchar_t*c = va_arg(ap, MSVCRT_wchar_t*);
 405                             *c = _WIDE2SUPPORTED_(nch);
 406                         }
 407                         nch = _GETC_(file);
 408                         st = 1;
 409                     }
 410                 }
 411                 break;
 412             case 'n': {
 413                     if (!suppress) {
 414                         int*n = va_arg(ap, int*);
 415                         *n = consumed - (nch!=_EOF_);
 416                     }
 417                     /* This is an odd one: according to the standard,
 418                      * "Execution of a %n directive does not increment the
 419                      * assignment count returned at the completion of
 420                      * execution" even if it wasn't suppressed with the
 421                      * '*' flag.  The Corrigendum to the standard seems
 422                      * to contradict this (comment out the assignment to
 423                      * suppress below if you want to implement these
 424                      * alternate semantics) but the windows program I'm
 425                      * looking at expects the behavior I've coded here
 426                      * (which happens to be what glibc does as well).
 427                      */
 428                     suppress = 1;
 429                     st = 1;
 430                 }
 431                 break;
 432             case '[': {
 433                     _CHAR_ *str = suppress ? NULL : va_arg(ap, _CHAR_*);
 434                     _CHAR_ *sptr = str;
 435                     RTL_BITMAP bitMask;
 436                     ULONG *Mask;
 437                     int invert = 0; /* Set if we are NOT to find the chars */
 438
 439                     /* Init our bitmap */
 440                     Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, _BITMAPSIZE_/8);
 441                     RtlInitializeBitMap(&bitMask, Mask, _BITMAPSIZE_);
 442
 443                     /* Read the format */
 444                     format++;
 445                     if(*format == '^') {
 446                         invert = 1;
 447                         format++;
 448                     }
 449                     if(*format == ']') {
 450                         RtlSetBits(&bitMask, ']', 1);
 451                         format++;
 452                     }
 453                     while(*format && (*format != ']')) {
 454                         /* According to:
 455                          * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/html/_crt_scanf_width_specification.asp
 456                          * "Note that %[a-z] and %[z-a] are interpreted as equivalent to %[abcde...z]." */
 457                         if((*format == '-') && (*(format + 1) != ']')) {
 458                             if ((*(format - 1)) < *(format + 1))
 459                                 RtlSetBits(&bitMask, *(format - 1) +1 , *(format + 1) - *(format - 1));
 460                             else
 461                                 RtlSetBits(&bitMask, *(format + 1)    , *(format - 1) - *(format + 1));
 462                             format++;
 463                         } else
 464                             RtlSetBits(&bitMask, *format, 1);
 465                         format++;
 466                     }
 467                     /* read until char is not suitable */
 468                     while ((width != 0) && (nch != _EOF_)) {
 469                         if(!invert) {
 470                             if(RtlAreBitsSet(&bitMask, nch, 1)) {
 471                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 472                             } else
 473                                 break;
 474                         } else {
 475                             if(RtlAreBitsClear(&bitMask, nch, 1)) {
 476                                 if (!suppress) *sptr++ = _CHAR2SUPPORTED_(nch);
 477                             } else
 478                                 break;
 479                         }
 480                         st++;
 481                         nch = _GETC_(file);
 482                         if (width>0) width--;
 483                     }
 484                     /* terminate */
 485                     if (!suppress) *sptr = 0;
 486                     HeapFree(GetProcessHeap(), 0, Mask);
 487                 }
 488                 break;
 489             default:
 490                 /* From spec: "if a percent sign is followed by a character
 491                  * that has no meaning as a format-control character, that
 492                  * character and the following characters are treated as
 493                  * an ordinary sequence of characters, that is, a sequence
 494                  * of characters that must match the input.  For example,
 495                  * to specify that a percent-sign character is to be input,
 496                  * use %%." */
 497                 while ((nch!=_EOF_) && _ISSPACE_(nch))
 498                     nch = _GETC_(file);
 499                 if (nch==*format) {
 500                     suppress = 1; /* whoops no field to be read */
 501                     st = 1; /* but we got what we expected */
 502                     nch = _GETC_(file);
 503                 }
 504                 break;
 505             }
 506             if (st && !suppress) rd++;
 507             else if (!st) break;
 508         }
 509         /* a non-white-space character causes scanf to read, but not store,
 510          * a matching non-white-space character. */
 511         else {
 512             /* check for character match */
 513             if (nch == *format) {
 514                 nch = _GETC_(file);
 515             } else break;
 516         }
 517         format++;
 518     }
 519     if (nch!=_EOF_) {
 520         _UNGETC_(nch, file);
 521     }
 522     TRACE("returning %d\n", rd);
 523     return rd;
 524 }
 525
 526 #undef _CHAR_
 527 #undef _EOF_
 528 #undef _EOF_RET
 529 #undef _ISSPACE_
 530 #undef _ISDIGIT_
 531 #undef _CHAR2SUPPORTED_
 532 #undef _WIDE2SUPPORTED_
 533 #undef _CHAR2DIGIT_
 534 #undef _GETC_
 535 #undef _UNGETC_
 536 #undef _FUNCTION_
 537 #undef _BITMAPSIZE_