lib/sdk/crt/string/scanf.h

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