?  search:
summary | log | graphiclog1 | graphiclog2 | commit | commitdiff | tree | refs | edit | fork
blame | history | raw | HEAD
Fix a heap corruption in MIX_Init ( sizeof() != strlen() ).
[wine/hacks.git] / dlls / msvcrt / scanf.h
blobdbd08fd7c7c3f10e92f1cf6620259d92f1e78c7d
1 /*
2 * general implementation of scanf used by scanf, sscanf, fscanf,
3 * _cscanf, wscanf, swscanf and fwscanf
4 *
5 * Copyright 1996,1998 Marcus Meissner
6 * Copyright 1996 Jukka Iivonen
7 * Copyright 1997,2000, 2003 Uwe Bonnes
8 * Copyright 2000 Jon Griffiths
9 * Copyright 2002 Daniel Gudbjartsson
10 *
11 * This library is free software; you can redistribute it and/or
12 * modify it under the terms of the GNU Lesser General Public
13 * License as published by the Free Software Foundation; either
14 * version 2.1 of the License, or (at your option) any later version.
15 *
16 * This library is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
19 * Lesser General Public License for more details.
20 *
21 * You should have received a copy of the GNU Lesser General Public
22 * License along with this library; if not, write to the Free Software
23 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
24 */
25
26 #ifdef WIDE_SCANF
27 #define _CHAR_ MSVCRT_wchar_t
28 #define _EOF_ MSVCRT_WEOF
29 #define _EOF_RET MSVCRT_WEOF
30 #define _ISSPACE_(c) MSVCRT_iswspace(c)
31 #define _ISDIGIT_(c) MSVCRT_iswdigit(c)
32 #define _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, number_signed;
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; number_signed = 0;
161 goto number;
162 case 'o': /* octal integer */
163 base = 8; number_signed = 0;
164 goto number;
165 case 'u': /* unsigned decimal integer */
166 base = 10; number_signed = 0;
167 goto number;
168 case 'd': /* signed decimal integer */
169 base = 10; number_signed = 1;
170 goto number;
171 case 'i': /* generic integer */
172 base = 10; number_signed = 1;
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 (number_signed && (nch == '-' ||
183 nch == '+')) {
184 negative = (nch=='-');
185 nch = _GETC_(file);
186 if (width>0) width--;
187 }
188 /* look for leading indication of base */
189 if (width!=0 && nch == '0') {
190 nch = _GETC_(file);
191 if (width>0) width--;
192 seendigit=1;
193 if (width!=0 && (nch=='x' || nch=='X')) {
194 if (base==0)
195 base=16;
196 if (base==16) {
197 nch = _GETC_(file);
198 if (width>0) width--;
199 seendigit=0;
200 }
201 } else if (base==0)
202 base = 8;
203 }
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 (number_signed) {
229 if (I64_prefix) _SET_NUMBER_(LONGLONG);
230 else if (l_prefix) _SET_NUMBER_(long int);
231 else if (h_prefix) _SET_NUMBER_(short int);
232 else _SET_NUMBER_(int);
233 } else {
234 if (negative) {
235 WARN("Dropping sign in reading a negative number into an unsigned value");
236 negative = 0;
237 }
238 if (I64_prefix) _SET_NUMBER_(ULONGLONG);
239 else if (l_prefix) _SET_NUMBER_(unsigned long int);
240 else if (h_prefix)
241 _SET_NUMBER_(unsigned short int);
242 else _SET_NUMBER_(unsigned int);
243 }
244 }
245 }
246 break;
247 case 'e':
248 case 'E':
249 case 'f':
250 case 'g':
251 case 'G': { /* read a float */
252 long double cur = 0;
253 int negative = 0;
254 /* skip initial whitespace */
255 while ((nch!=_EOF_) && _ISSPACE_(nch))
256 nch = _GETC_(file);
257 /* get sign. */
258 if (nch == '-' || nch == '+') {
259 negative = (nch=='-');
260 if (width>0) width--;
261 if (width==0) break;
262 nch = _GETC_(file);
263 }
264 /* get first digit. */
265 if ('.' != nch) {
266 if (!_ISDIGIT_(nch)) break;
267 cur = (nch - '0');
268 nch = _GETC_(file);
269 if (width>0) width--;
270 /* read until no more digits */
271 while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
272 cur = cur*10 + (nch - '0');
273 nch = _GETC_(file);
274 if (width>0) width--;
275 }
276 } else {
277 cur = 0; /* MaxPayneDemo Fix: .8 -> 0.8 */
278 }
279 /* handle decimals */
280 if (width!=0 && nch == '.') {
281 float dec = 1;
282 nch = _GETC_(file);
283 if (width>0) width--;
284 while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
285 dec /= 10;
286 cur += dec * (nch - '0');
287 nch = _GETC_(file);
288 if (width>0) width--;
289 }
290 }
291 /* handle exponent */
292 if (width!=0 && (nch == 'e' || nch == 'E')) {
293 int exponent = 0, negexp = 0;
294 float expcnt;
295 nch = _GETC_(file);
296 if (width>0) width--;
297 /* possible sign on the exponent */
298 if (width!=0 && (nch=='+' || nch=='-')) {
299 negexp = (nch=='-');
300 nch = _GETC_(file);
301 if (width>0) width--;
302 }
303 /* exponent digits */
304 while (width!=0 && (nch!=_EOF_) && _ISDIGIT_(nch)) {
305 exponent *= 10;
306 exponent += (nch - '0');
307 nch = _GETC_(file);
308 if (width>0) width--;
309 }
310 /* update 'cur' with this exponent. */
311 expcnt = negexp ? .1 : 10;
312 while (exponent!=0) {
313 if (exponent&1)
314 cur*=expcnt;
315 exponent/=2;
316 expcnt=expcnt*expcnt;
317 }
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
328 * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vclib/html/_crt_scanf_type_field_characters.asp
329 * 's' reads a character string in a call to fscanf
330 * and 'S' a wide character string and vice versa in a
331 * call to fwscanf. The 'h', 'w' and 'l' prefixes override
332 * this behaviour. 'h' forces reading char * but 'l' and 'w'
333 * force reading WCHAR. */
334 case 's':
335 if (w_prefix || l_prefix) goto widecharstring;
336 else if (h_prefix) goto charstring;
337 #ifdef WIDE_SCANF
338 else goto widecharstring;
339 #else /* WIDE_SCANF */
340 else goto charstring;
341 #endif /* WIDE_SCANF */
342 case 'S':
343 if (w_prefix || l_prefix) goto widecharstring;
344 else if (h_prefix) goto charstring;
345 #ifdef WIDE_SCANF
346 else goto charstring;
347 #else /* WIDE_SCANF */
348 else goto widecharstring;
349 #endif /* WIDE_SCANF */
350 charstring: { /* read a word into a char */
351 char*str = suppress ? NULL : va_arg(ap, char*);
352 char*sptr = str;
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++ = _CHAR2SUPPORTED_(nch);
359 st++;
360 nch = _GETC_(file);
361 if (width>0) width--;
362 }
363 /* terminate */
364 if (!suppress) *sptr = 0;
365 }
366 break;
367 widecharstring: { /* read a word into a wchar_t* */
368 MSVCRT_wchar_t*str =
369 suppress ? NULL : va_arg(ap, MSVCRT_wchar_t*);
370 MSVCRT_wchar_t*sptr = str;
371 /* skip initial whitespace */
372 while ((nch!=_EOF_) && _ISSPACE_(nch))
373 nch = _GETC_(file);
374 /* read until whitespace */
375 while (width!=0 && (nch!=_EOF_) && !_ISSPACE_(nch)) {
376 if (!suppress) *sptr++ = _WIDE2SUPPORTED_(nch);
377 st++;
378 nch = _GETC_(file);
379 if (width>0) width--;
380 }
381 /* terminate */
382 if (!suppress) *sptr = 0;
383 }
384 break;
385 /* 'c' and 'C work analogously to 's' and 'S' as described
386 * above */
387 case 'c':
388 if (w_prefix || l_prefix) goto widecharacter;
389 else if (h_prefix) goto character;
390 #ifdef WIDE_SCANF
391 else goto widecharacter;
392 #else /* WIDE_SCANF */
393 else goto character;
394 #endif /* WIDE_SCANF */
395 case 'C':
396 if (w_prefix || l_prefix) goto widecharacter;
397 else if (h_prefix) goto character;
398 #ifdef WIDE_SCANF
399 else goto character;
400 #else /* WIDE_SCANF */
401 else goto widecharacter;
402 #endif /* WIDE_SCANF */
403 character: { /* read single character into char */
404 if (!suppress) {
405 char*c = va_arg(ap, char*);
406 *c = _CHAR2SUPPORTED_(nch);
407 }
408 st = 1;
409 nch = _GETC_(file);
410 }
411 break;
412 widecharacter: {
413 if (!suppress) { /* read single character into a wchar_t */
414 MSVCRT_wchar_t*c = va_arg(ap, MSVCRT_wchar_t*);
415 *c = _WIDE2SUPPORTED_(nch);
416 }
417 nch = _GETC_(file);
418 st = 1;
419 }
420 break;
421 case 'n': {
422 if (!suppress) {
423 int*n = va_arg(ap, int*);
424 *n = consumed - (nch!=_EOF_);
425 }
426 }
427 break;
428 case '[': {
429 _CHAR_ *str = suppress ? NULL : va_arg(ap, _CHAR_*);
430 _CHAR_ *sptr = str;
431 RTL_BITMAP bitMask;
432 ULONG *Mask;
433 int invert = 0; /* Set if we are NOT to find the chars */
434
435 /* Init our bitmap */
436 Mask = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, _BITMAPSIZE_/8);
437 RtlInitializeBitMap(&bitMask, Mask, _BITMAPSIZE_);
438
439 /* Read the format */
440 format++;
441 if(*format == '^') {
442 invert = 1;
443 format++;
444 }
445 if(*format == ']') {
446 RtlSetBits(&bitMask, ']', 1);
447 format++;
448 }
449 while(*format && (*format != ']')) {
450 /* According to:
451 * http://msdn.microsoft.com/library/default.asp?url=/library/en-us/vccore98/html/_crt_scanf_width_specification.asp
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_
useful hacks and other patches not (yet) suitable for mainline