2 * general implementation of scanf used by scanf, sscanf, fscanf,
3 * _cscanf, wscanf, swscanf and fwscanf
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
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.
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.
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
27 #define _CHAR_ MSVCRT_wchar_t
28 #define _EOF_ MSVCRT_WEOF
29 #define _EOF_RET (short)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 */
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 */
49 #define _GETC_(file) (consumed++, _getch())
50 #define _UNGETC_(nch, file) do { _ungetch(nch); consumed--; } while(0)
51 #define _LOCK_FILE_(file) MSVCRT__lock_file(MSVCRT_stdin)
52 #define _UNLOCK_FILE_(file) MSVCRT__unlock_file(MSVCRT_stdin)
55 #define _FUNCTION_ static int MSVCRT_vcwscanf_s_l(const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
57 #define _FUNCTION_ static int MSVCRT_vcwscanf_l(const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
59 #else /* WIDE_SCANF */
61 #define _FUNCTION_ static int MSVCRT_vcscanf_s_l(const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
63 #define _FUNCTION_ static int MSVCRT_vcscanf_l(const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
65 #endif /* WIDE_SCANF */
71 #define _GETC_(file) (consumed==length ? '\0' : (consumed++, *file++))
72 #define _UNGETC_(nch, file) do { file--; consumed--; } while(0)
73 #define _LOCK_FILE_(file) do {} while(0)
74 #define _UNLOCK_FILE_(file) do {} while(0)
77 #define _FUNCTION_ static int MSVCRT_vsnwscanf_s_l(const MSVCRT_wchar_t *file, MSVCRT_size_t length, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
79 #define _FUNCTION_ static int MSVCRT_vsnwscanf_l(const MSVCRT_wchar_t *file, MSVCRT_size_t length, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
81 #else /* WIDE_SCANF */
83 #define _FUNCTION_ static int MSVCRT_vsnscanf_s_l(const char *file, MSVCRT_size_t length, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
85 #define _FUNCTION_ static int MSVCRT_vsnscanf_l(const char *file, MSVCRT_size_t length, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
87 #endif /* WIDE_SCANF */
88 #else /* STRING_LEN */
89 #define _GETC_(file) (consumed++, *file++)
90 #define _UNGETC_(nch, file) do { file--; consumed--; } while(0)
91 #define _LOCK_FILE_(file) do {} while(0)
92 #define _UNLOCK_FILE_(file) do {} while(0)
95 #define _FUNCTION_ static int MSVCRT_vswscanf_s_l(const MSVCRT_wchar_t *file, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
97 #define _FUNCTION_ static int MSVCRT_vswscanf_l(const MSVCRT_wchar_t *file, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
99 #else /* WIDE_SCANF */
101 #define _FUNCTION_ static int MSVCRT_vsscanf_s_l(const char *file, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
103 #define _FUNCTION_ static int MSVCRT_vsscanf_l(const char *file, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
105 #endif /* WIDE_SCANF */
106 #endif /* STRING_LEN */
109 #define _GETC_(file) (consumed++, MSVCRT_fgetwc(file))
110 #define _UNGETC_(nch, file) do { MSVCRT_ungetwc(nch, file); consumed--; } while(0)
111 #define _LOCK_FILE_(file) MSVCRT__lock_file(file)
112 #define _UNLOCK_FILE_(file) MSVCRT__unlock_file(file)
114 #define _FUNCTION_ static int MSVCRT_vfwscanf_s_l(MSVCRT_FILE* file, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
116 #define _FUNCTION_ static int MSVCRT_vfwscanf_l(MSVCRT_FILE* file, const MSVCRT_wchar_t *format, MSVCRT__locale_t locale, __ms_va_list ap)
118 #else /* WIDE_SCANF */
119 #define _GETC_(file) (consumed++, MSVCRT_fgetc(file))
120 #define _UNGETC_(nch, file) do { MSVCRT_ungetc(nch, file); consumed--; } while(0)
121 #define _LOCK_FILE_(file) MSVCRT__lock_file(file)
122 #define _UNLOCK_FILE_(file) MSVCRT__unlock_file(file)
124 #define _FUNCTION_ static int MSVCRT_vfscanf_s_l(MSVCRT_FILE* file, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
126 #define _FUNCTION_ static int MSVCRT_vfscanf_l(MSVCRT_FILE* file, const char *format, MSVCRT__locale_t locale, __ms_va_list ap)
128 #endif /* WIDE_SCANF */
133 MSVCRT_pthreadlocinfo locinfo
;
134 int rd
= 0, consumed
= 0;
136 if (!*format
) return 0;
139 TRACE("(%s):\n", debugstr_a(format
));
142 TRACE("%s (%s)\n", debugstr_a(file
), debugstr_a(format
));
144 TRACE("%p (%s)\n", file
, debugstr_a(format
));
147 #endif /* WIDE_SCANF */
157 locinfo
= get_locinfo();
159 locinfo
= locale
->locinfo
;
162 /* a whitespace character in the format string causes scanf to read,
163 * but not store, all consecutive white-space characters in the input
164 * up to the next non-white-space character. One white space character
165 * in the input matches any number (including zero) and combination of
166 * white-space characters in the input. */
167 if (_ISSPACE_(*format
)) {
168 /* skip whitespace */
169 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
172 /* a format specification causes scanf to read and convert characters
173 * in the input into values of a specified type. The value is assigned
174 * to an argument in the argument list. Format specifications have
175 * the form %[*][width][{h | l | I64 | L}]type */
176 else if (*format
== '%') {
177 int st
= 0; int suppress
= 0; int width
= 0;
183 int prefix_finished
= 0;
186 /* look for leading asterisk, which means 'suppress assignment of
192 /* look for width specification */
193 while (_ISDIGIT_(*format
)) {
195 width
+=*format
++ - '0';
197 if (width
==0) width
=-1; /* no width spec seen */
198 /* read prefix (if any) */
199 while (!prefix_finished
) {
201 case 'h': h_prefix
++; break;
203 if(*(format
+1) == 'l') {
209 case 'w': w_prefix
= 1; break;
210 case 'L': L_prefix
= 1; break;
212 if (*(format
+ 1) == '6' &&
213 *(format
+ 2) == '4') {
221 if (!prefix_finished
) format
++;
226 case 'P': /* pointer. */
227 if (sizeof(void *) == sizeof(LONGLONG
)) I64_prefix
= 1;
230 case 'X': /* hexadecimal integer. */
233 case 'o': /* octal integer */
236 case 'u': /* unsigned decimal integer */
239 case 'd': /* signed decimal integer */
242 case 'i': /* generic integer */
245 /* read an integer */
249 /* skip initial whitespace */
250 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
253 if (nch
== '-' || nch
== '+') {
254 negative
= (nch
=='-');
256 if (width
>0) width
--;
258 /* look for leading indication of base */
259 if (width
!=0 && nch
== '0' && *format
!= 'p' && *format
!= 'P') {
261 if (width
>0) width
--;
263 if (width
!=0 && (nch
=='x' || nch
=='X')) {
268 if (width
>0) width
--;
274 /* format %i without indication of base */
277 /* throw away leading zeros */
278 while (width
!=0 && nch
=='0') {
280 if (width
>0) width
--;
283 if (width
!=0 && _CHAR2DIGIT_(nch
, base
)!=-1) {
284 cur
= _CHAR2DIGIT_(nch
, base
);
286 if (width
>0) width
--;
289 /* read until no more digits */
290 while (width
!=0 && (nch
!=_EOF_
) && _CHAR2DIGIT_(nch
, base
)!=-1) {
291 cur
= cur
*base
+ _CHAR2DIGIT_(nch
, base
);
293 if (width
>0) width
--;
297 if (!seendigit
) break; /* not a valid number */
300 #define _SET_NUMBER_(type) *va_arg(ap, type*) = negative ? -cur : cur
301 if (I64_prefix
) _SET_NUMBER_(LONGLONG
);
302 else if (l_prefix
) _SET_NUMBER_(LONG
);
303 else if (h_prefix
== 1) _SET_NUMBER_(short int);
304 else _SET_NUMBER_(int);
312 case 'G': { /* read a float */
315 /* skip initial whitespace */
316 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
319 if (nch
== '-' || nch
== '+') {
320 negative
= (nch
=='-');
321 if (width
>0) width
--;
325 /* get first digit. */
326 if (*locinfo
->lconv
->decimal_point
!= nch
) {
327 if (!_ISDIGIT_(nch
)) break;
330 if (width
>0) width
--;
331 /* read until no more digits */
332 while (width
!=0 && (nch
!=_EOF_
) && _ISDIGIT_(nch
)) {
333 cur
= cur
*10 + (nch
- '0');
335 if (width
>0) width
--;
338 cur
= 0; /* Fix: .8 -> 0.8 */
340 /* handle decimals */
341 if (width
!=0 && nch
== *locinfo
->lconv
->decimal_point
) {
344 if (width
>0) width
--;
345 while (width
!=0 && (nch
!=_EOF_
) && _ISDIGIT_(nch
)) {
347 cur
+= dec
* (nch
- '0');
349 if (width
>0) width
--;
352 /* handle exponent */
353 if (width
!=0 && (nch
== 'e' || nch
== 'E')) {
354 int exponent
= 0, negexp
= 0;
357 if (width
>0) width
--;
358 /* possible sign on the exponent */
359 if (width
!=0 && (nch
=='+' || nch
=='-')) {
362 if (width
>0) width
--;
364 /* exponent digits */
365 while (width
!=0 && (nch
!=_EOF_
) && _ISDIGIT_(nch
)) {
367 exponent
+= (nch
- '0');
369 if (width
>0) width
--;
371 /* update 'cur' with this exponent. */
372 expcnt
= negexp
? .1 : 10;
373 while (exponent
!=0) {
377 expcnt
=expcnt
*expcnt
;
382 if (L_prefix
) _SET_NUMBER_(long double);
383 else if (l_prefix
) _SET_NUMBER_(double);
384 else _SET_NUMBER_(float);
388 /* According to msdn,
389 * 's' reads a character string in a call to fscanf
390 * and 'S' a wide character string and vice versa in a
391 * call to fwscanf. The 'h', 'w' and 'l' prefixes override
392 * this behaviour. 'h' forces reading char * but 'l' and 'w'
393 * force reading WCHAR. */
395 if (w_prefix
|| l_prefix
) goto widecharstring
;
396 else if (h_prefix
) goto charstring
;
398 else goto widecharstring
;
399 #else /* WIDE_SCANF */
400 else goto charstring
;
401 #endif /* WIDE_SCANF */
403 if (w_prefix
|| l_prefix
) goto widecharstring
;
404 else if (h_prefix
) goto charstring
;
406 else goto charstring
;
407 #else /* WIDE_SCANF */
408 else goto widecharstring
;
409 #endif /* WIDE_SCANF */
410 charstring
: { /* read a word into a char */
411 char *sptr
= suppress
? NULL
: va_arg(ap
, char*);
412 char *sptr_beg
= sptr
;
414 unsigned size
= suppress
? UINT_MAX
: va_arg(ap
, unsigned);
416 unsigned size
= UINT_MAX
;
418 /* skip initial whitespace */
419 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
421 /* read until whitespace */
422 while (width
!=0 && (nch
!=_EOF_
) && !_ISSPACE_(nch
)) {
424 *sptr
++ = _CHAR2SUPPORTED_(nch
);
434 if (width
>0) width
--;
437 if (!suppress
) *sptr
= 0;
440 widecharstring
: { /* read a word into a wchar_t* */
441 MSVCRT_wchar_t
*sptr
= suppress
? NULL
: va_arg(ap
, MSVCRT_wchar_t
*);
442 MSVCRT_wchar_t
*sptr_beg
= sptr
;
444 unsigned size
= suppress
? UINT_MAX
: va_arg(ap
, unsigned);
446 unsigned size
= UINT_MAX
;
448 /* skip initial whitespace */
449 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
451 /* read until whitespace */
452 while (width
!=0 && (nch
!=_EOF_
) && !_ISSPACE_(nch
)) {
454 *sptr
++ = _WIDE2SUPPORTED_(nch
);
464 if (width
>0) width
--;
467 if (!suppress
) *sptr
= 0;
470 /* 'c' and 'C work analogously to 's' and 'S' as described
473 if (w_prefix
|| l_prefix
) goto widecharacter
;
474 else if (h_prefix
) goto character
;
476 else goto widecharacter
;
477 #else /* WIDE_SCANF */
479 #endif /* WIDE_SCANF */
481 if (w_prefix
|| l_prefix
) goto widecharacter
;
482 else if (h_prefix
) goto character
;
485 #else /* WIDE_SCANF */
486 else goto widecharacter
;
487 #endif /* WIDE_SCANF */
488 character
: { /* read single character into char */
489 char *str
= suppress
? NULL
: va_arg(ap
, char*);
492 unsigned size
= suppress
? UINT_MAX
: va_arg(ap
, unsigned)/sizeof(char);
494 unsigned size
= UINT_MAX
;
496 if (width
== -1) width
= 1;
497 while (width
&& (nch
!= _EOF_
))
500 *str
++ = _CHAR2SUPPORTED_(nch
);
514 widecharacter
: { /* read single character into a wchar_t */
515 MSVCRT_wchar_t
*str
= suppress
? NULL
: va_arg(ap
, MSVCRT_wchar_t
*);
516 MSVCRT_wchar_t
*pstr
= str
;
518 unsigned size
= suppress
? UINT_MAX
: va_arg(ap
, unsigned)/sizeof(MSVCRT_wchar_t
);
520 unsigned size
= UINT_MAX
;
522 if (width
== -1) width
= 1;
523 while (width
&& (nch
!= _EOF_
))
526 *str
++ = _WIDE2SUPPORTED_(nch
);
542 int*n
= va_arg(ap
, int*);
545 /* This is an odd one: according to the standard,
546 * "Execution of a %n directive does not increment the
547 * assignment count returned at the completion of
548 * execution" even if it wasn't suppressed with the
549 * '*' flag. The Corrigendum to the standard seems
550 * to contradict this (comment out the assignment to
551 * suppress below if you want to implement these
552 * alternate semantics) but the windows program I'm
553 * looking at expects the behavior I've coded here
554 * (which happens to be what glibc does as well).
561 _CHAR_
*str
= suppress
? NULL
: va_arg(ap
, _CHAR_
*);
565 int invert
= 0; /* Set if we are NOT to find the chars */
567 unsigned size
= suppress
? UINT_MAX
: va_arg(ap
, unsigned)/sizeof(_CHAR_
);
569 unsigned size
= UINT_MAX
;
572 /* Init our bitmap */
573 Mask
= HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY
, _BITMAPSIZE_
/8);
574 RtlInitializeBitMap(&bitMask
, Mask
, _BITMAPSIZE_
);
576 /* Read the format */
583 RtlSetBits(&bitMask
, ']', 1);
586 while(*format
&& (*format
!= ']')) {
587 /* According to msdn:
588 * "Note that %[a-z] and %[z-a] are interpreted as equivalent to %[abcde...z]." */
589 if((*format
== '-') && (*(format
+ 1) != ']')) {
590 if ((*(format
- 1)) < *(format
+ 1))
591 RtlSetBits(&bitMask
, *(format
- 1) +1 , *(format
+ 1) - *(format
- 1));
593 RtlSetBits(&bitMask
, *(format
+ 1) , *(format
- 1) - *(format
+ 1));
596 RtlSetBits(&bitMask
, *format
, 1);
599 /* read until char is not suitable */
600 while ((width
!= 0) && (nch
!= _EOF_
)) {
602 if(RtlAreBitsSet(&bitMask
, nch
, 1)) {
603 if (!suppress
) *sptr
++ = _CHAR2SUPPORTED_(nch
);
607 if(RtlAreBitsClear(&bitMask
, nch
, 1)) {
608 if (!suppress
) *sptr
++ = _CHAR2SUPPORTED_(nch
);
614 if (width
>0) width
--;
623 if (!suppress
) *sptr
= 0;
624 HeapFree(GetProcessHeap(), 0, Mask
);
628 /* From spec: "if a percent sign is followed by a character
629 * that has no meaning as a format-control character, that
630 * character and the following characters are treated as
631 * an ordinary sequence of characters, that is, a sequence
632 * of characters that must match the input. For example,
633 * to specify that a percent-sign character is to be input,
635 while ((nch
!=_EOF_
) && _ISSPACE_(nch
))
638 suppress
= 1; /* whoops no field to be read */
639 st
= 1; /* but we got what we expected */
644 if (st
&& !suppress
) rd
++;
647 /* a non-white-space character causes scanf to read, but not store,
648 * a matching non-white-space character. */
650 /* check for character match */
651 if (nch
== *format
) {
661 TRACE("returning %d\n", rd
);
671 #undef _CHAR2SUPPORTED_
672 #undef _WIDE2SUPPORTED_