1 /* Copyright (C) 1991, 92, 93, 94, 95, 96 Free Software Foundation, Inc.
2 This file is part of the GNU C Library.
4 The GNU C Library is free software; you can redistribute it and/or
5 modify it under the terms of the GNU Library General Public License as
6 published by the Free Software Foundation; either version 2 of the
7 License, or (at your option) any later version.
9 The GNU C Library is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 Library General Public License for more details.
14 You should have received a copy of the GNU Library General Public
15 License along with the GNU C Library; see the file COPYING.LIB. If not,
16 write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
17 Boston, MA 02111-1307, USA. */
19 #include "../locale/localeinfo.h"
28 #include <libc-lock.h>
32 #define LONGLONG long long
37 /* Those are flags in the conversion format. */
38 # define LONG 0x001 /* l: long or double */
39 # define LONGDBL 0x002 /* L: long long or long double */
40 # define SHORT 0x004 /* h: short */
41 # define SUPPRESS 0x008 /* *: suppress assignment */
42 # define POINTER 0x010 /* weird %p pointer (`fake hex') */
43 # define NOSKIP 0x020 /* do not skip blanks */
44 # define WIDTH 0x040 /* width was given */
45 # define GROUP 0x080 /* ': group numbers */
46 # define MALLOC 0x100 /* a: malloc strings */
48 # define TYPEMOD (LONG|LONGDBL|SHORT)
56 # define va_list _IO_va_list
57 # define ungetc(c, s) (--read_in, _IO_ungetc (c, s))
58 # define inchar() ((c = _IO_getc_unlocked (s)), (void) ++read_in, c)
59 # define encode_error() do { \
60 if (errp != NULL) *errp |= 4; \
61 _IO_funlockfile (s); \
62 __set_errno (EILSEQ); \
65 # define conv_error() do { \
66 if (errp != NULL) *errp |= 2; \
67 _IO_funlockfile (s); \
70 # define input_error() do { \
71 _IO_funlockfile (s); \
72 if (errp != NULL) *errp |= 1; \
75 # define memory_error() do { \
76 _IO_funlockfile (s); \
77 __set_errno (ENOMEM); \
80 # define ARGCHECK(s, format) \
83 /* Check file argument for consistence. */ \
84 CHECK_FILE (s, EOF); \
85 if (s->_flags & _IO_NO_READS) \
87 __set_errno (EBADF); \
90 else if (format == NULL) \
96 # define LOCK_STREAM(S) \
97 __libc_cleanup_region_start ((void (*) (void *)) &_IO_funlockfile, (S)); \
99 # define UNLOCK_STREAM __libc_cleanup_region_end (1)
101 # define ungetc(c, s) (--read_in, ungetc (c, s))
102 # define inchar() ((c = getc (s)), (void) ++read_in, c)
103 # define encode_error() do { \
105 __set_errno (EILSEQ); \
108 # define conv_error() do { \
112 # define input_error() do { \
114 return done ?: EOF; \
116 # define memory_error() do { \
118 __set_errno (ENOMEM); \
121 # define ARGCHECK(s, format) \
124 /* Check file argument for consistence. */ \
125 if (!__validfp (s) || !s->__mode.__read) \
127 __set_errno (EBADF); \
130 else if (format == NULL) \
132 __set_errno (EINVAL); \
137 /* XXX For now !!! */
138 # define flockfile(S) /* nothing */
139 # define funlockfile(S) /* nothing */
140 # define LOCK_STREAM(S)
141 # define UNLOCK_STREAM
143 # define LOCK_STREAM(S) \
144 __libc_cleanup_region_start (&__funlockfile, (S)); \
146 # define UNLOCK_STREAM __libc_cleanup_region_end (1)
151 /* Read formatted input from S according to the format string
152 FORMAT, using the argument list in ARG.
153 Return the number of assignments made, or -1 for an input error. */
156 _IO_vfscanf (s, format, argptr, errp)
163 __vfscanf (FILE *s, const char *format, va_list argptr)
166 va_list arg = (va_list) argptr;
168 register const char *f = format;
169 register unsigned char fc; /* Current character of the format. */
170 register size_t done = 0; /* Assignments done. */
171 register size_t read_in = 0; /* Chars read in. */
172 register int c = 0; /* Last char read. */
173 register int width; /* Maximum field width. */
174 register int flags; /* Modifiers for current format element. */
176 /* Status for reading F-P nums. */
177 char got_dot, got_e, negative;
178 /* If a [...] is a [^...]. */
180 /* Base for integral numbers. */
182 /* Signedness for integral numbers. */
184 /* Decimal point character. */
186 /* The thousands character of the current locale. */
188 /* Integral holding variables. */
192 unsigned long long int uq;
194 unsigned long int ul;
196 /* Character-buffer pointer. */
198 wchar_t *wstr = NULL;
199 char **strptr = NULL;
201 /* We must not react on white spaces immediately because they can
202 possibly be matched even if in the input stream no character is
203 available anymore. */
206 char *tw; /* Temporary pointer. */
207 char *wp = NULL; /* Workspace. */
208 size_t wpmax = 0; /* Maximal size of workspace. */
209 size_t wpsize; /* Currently used bytes in workspace. */
213 if (wpsize == wpmax) \
216 wpmax = UCHAR_MAX > 2 * wpmax ? UCHAR_MAX : 2 * wpmax; \
217 wp = (char *) alloca (wpmax); \
219 memcpy (wp, old, wpsize); \
221 wp[wpsize++] = (Ch); \
225 ARGCHECK (s, format);
227 /* Figure out the decimal point character. */
228 if (mbtowc (&decimal, _NL_CURRENT (LC_NUMERIC, DECIMAL_POINT),
229 strlen (_NL_CURRENT (LC_NUMERIC, DECIMAL_POINT))) <= 0)
230 decimal = (wchar_t) *_NL_CURRENT (LC_NUMERIC, DECIMAL_POINT);
231 /* Figure out the thousands separator character. */
232 if (mbtowc (&thousands, _NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP),
233 strlen (_NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP))) <= 0)
234 thousands = (wchar_t) *_NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP);
236 /* Lock the stream. */
239 /* Run through the format string. */
243 /* Extract the next argument, which is of type TYPE.
244 For a %N$... spec, this is the Nth argument from the beginning;
245 otherwise it is the next argument after the state now in ARG. */
247 /* XXX Possible optimization. */
248 # define ARG(type) (argpos == 0 ? va_arg (arg, type) : \
249 ({ va_list arg = (va_list) argptr; \
250 arg = (va_list) ((char *) arg \
252 * __va_rounded_size (void *)); \
253 va_arg (arg, type); \
256 # define ARG(type) (argpos == 0 ? va_arg (arg, type) : \
257 ({ unsigned int pos = argpos; \
258 va_list arg = (va_list) argptr; \
260 (void) va_arg (arg, void *); \
261 va_arg (arg, type); \
267 /* Non-ASCII, may be a multibyte. */
268 int len = mblen (f, strlen (f));
290 /* Remember to skip spaces. */
297 /* Read a character. */
300 /* Characters other than format specs must just match. */
304 /* We saw white space char as the last character in the format
305 string. Now it's time to skip all leading white space. */
309 if (inchar () == EOF && errno == EINTR)
323 /* This is the start of the conversion string. */
326 /* Initialize state of modifiers. */
329 /* Prepare temporary buffer. */
332 /* Check for a positional parameter specification. */
337 argpos = argpos * 10 + (*f++ - '0');
342 /* Oops; that was actually the field width. */
350 /* Check for the assignment-suppressing and the number grouping flag. */
351 while (*f == '*' || *f == '\'')
362 /* We have seen width. */
366 /* Find the maximum field width. */
377 /* Check for type modifiers. */
378 while (*f == 'h' || *f == 'l' || *f == 'L' || *f == 'a' || *f == 'q')
382 /* int's are short int's. */
384 /* Signal illegal format element. */
389 if (flags & (SHORT|LONGDBL))
391 else if (flags & LONG)
393 /* A double `l' is equivalent to an `L'. */
398 /* int's are long int's. */
403 /* double's are long double's, and int's are long long int's. */
405 /* Signal illegal format element. */
411 /* Signal illegal format element. */
413 /* String conversions (%s, %[) take a `char **'
414 arg and fill it in with a malloc'd pointer. */
419 /* End of the format string? */
423 /* We must take care for EINTR errors. */
424 if (c == EOF && errno == EINTR)
427 /* Find the conversion specifier. */
429 if (skip_space || (fc != '[' && fc != 'c' && fc != 'C' && fc != 'n'))
431 /* Eat whitespace. */
433 if (inchar () == EOF && errno == EINTR)
442 case '%': /* Must match a literal '%'. */
451 case 'n': /* Answer number of assignments done. */
452 /* Corrigendum 1 to ISO C 1990 describes the allowed flags
453 with the 'n' conversion specifier. */
454 if (!(flags & SUPPRESS))
455 /* Don't count the read-ahead. */
457 *ARG (long long int *) = read_in;
458 else if (flags & LONG)
459 *ARG (long int *) = read_in;
460 else if (flags & SHORT)
461 *ARG (short int *) = read_in;
463 *ARG (int *) = read_in;
466 case 'c': /* Match characters. */
467 if ((flags & LONG) == 0)
469 if (!(flags & SUPPRESS))
483 if (!(flags & SUPPRESS))
487 while (--width > 0 && inchar () != EOF);
490 while (--width > 0 && inchar () != EOF);
493 /* I.e., EOF was read. */
496 if (!(flags & SUPPRESS))
503 /* Get UTF-8 encoded wide character. Here we assume (as in
504 other parts of the libc) that we only have to handle
511 if (!(flags & SUPPRESS))
513 wstr = ARG (wchar_t *);
520 #define NEXT_WIDE_CHAR(First) \
523 /* EOF is only an error for the first character. */ \
534 if ((c & 0xc0) == 0x80 || (c & 0xfe) == 0xfe) \
536 if ((c & 0xe0) == 0xc0) \
538 /* We expect two bytes. */ \
542 else if ((c & 0xf0) == 0xe0) \
544 /* We expect three bytes. */ \
548 else if ((c & 0xf8) == 0xf0) \
550 /* We expect four bytes. */ \
554 else if ((c & 0xfc) == 0xf8) \
556 /* We expect five bytes. */ \
562 /* We expect six bytes. */ \
571 || (c & 0xc0) == 0x80 || (c & 0xfe) == 0xfe) \
579 if (!(flags & SUPPRESS)) \
583 NEXT_WIDE_CHAR (first);
588 /* I.e., EOF was read. */
591 if (!(flags & SUPPRESS))
596 case 's': /* Read a string. */
598 /* We have to process a wide character string. */
599 goto wide_char_string;
601 #define STRING_ARG(Str, Type) \
602 if (!(flags & SUPPRESS)) \
604 if (flags & MALLOC) \
606 /* The string is to be stored in a malloc'd buffer. */ \
607 strptr = ARG (char **); \
608 if (strptr == NULL) \
610 /* Allocate an initial buffer. */ \
612 *strptr = malloc (strsize * sizeof (Type)); \
613 Str = (Type *) *strptr; \
616 Str = ARG (Type *); \
620 STRING_ARG (str, char);
633 #define STRING_ADD_CHAR(Str, c, Type) \
634 if (!(flags & SUPPRESS)) \
637 if ((flags & MALLOC) && (char *) Str == *strptr + strsize) \
639 /* Enlarge the buffer. */ \
640 Str = realloc (*strptr, strsize * 2 * sizeof (Type)); \
643 /* Can't allocate that much. Last-ditch effort. */\
644 Str = realloc (*strptr, \
645 (strsize + 1) * sizeof (Type)); \
648 /* We lose. Oh well. \
649 Terminate the string and stop converting, \
650 so at least we don't skip any input. */ \
651 ((Type *) (*strptr))[strsize] = '\0'; \
657 *strptr = (char *) Str; \
658 Str = ((Type *) *strptr) + strsize; \
664 *strptr = (char *) Str; \
665 Str = ((Type *) *strptr) + strsize; \
670 STRING_ADD_CHAR (str, c, char);
671 } while ((width <= 0 || --width > 0) && inchar () != EOF);
676 if (!(flags & SUPPRESS))
684 /* Wide character string. */
689 STRING_ARG (wstr, wchar_t);
694 NEXT_WIDE_CHAR (first);
698 /* XXX We would have to push back the whole wide char
699 with possibly many bytes. But since scanf does
700 not make a difference for white space characters
701 we can simply push back a simple <SP> which is
702 guaranteed to be in the [:space:] class. */
707 STRING_ADD_CHAR (wstr, val, wchar_t);
710 while (width <= 0 || --width > 0);
712 if (!(flags & SUPPRESS))
720 case 'x': /* Hexadecimal integer. */
721 case 'X': /* Ditto. */
726 case 'o': /* Octal integer. */
731 case 'u': /* Unsigned decimal integer. */
736 case 'd': /* Signed decimal integer. */
741 case 'i': /* Generic number. */
750 /* Check for a sign. */
751 if (c == '-' || c == '+')
759 /* Look for a leading indication of base. */
760 if (width != 0 && c == '0')
768 if (width != 0 && tolower (c) == 'x')
786 /* Read the number into workspace. */
787 while (c != EOF && width != 0)
789 if (base == 16 ? !isxdigit (c) :
790 ((!isdigit (c) || c - '0' >= base) &&
791 !((flags & GROUP) && base == 10 && c == thousands)))
800 /* The just read character is not part of the number anymore. */
804 (wpsize == 1 && (wp[0] == '+' || wp[0] == '-')))
805 /* There was no number. */
808 /* Convert the number. */
813 num.q = __strtoq_internal (wp, &tw, base, flags & GROUP);
815 num.uq = __strtouq_internal (wp, &tw, base, flags & GROUP);
820 num.l = __strtol_internal (wp, &tw, base, flags & GROUP);
822 num.ul = __strtoul_internal (wp, &tw, base, flags & GROUP);
827 if (!(flags & SUPPRESS))
832 *ARG (unsigned LONGLONG int *) = num.uq;
833 else if (flags & LONG)
834 *ARG (unsigned long int *) = num.ul;
835 else if (flags & SHORT)
836 *ARG (unsigned short int *)
837 = (unsigned short int) num.ul;
839 *ARG (unsigned int *) = (unsigned int) num.ul;
844 *ARG (LONGLONG int *) = num.q;
845 else if (flags & LONG)
846 *ARG (long int *) = num.l;
847 else if (flags & SHORT)
848 *ARG (short int *) = (short int) num.l;
850 *ARG (int *) = (int) num.l;
856 case 'e': /* Floating-point numbers. */
865 /* Check for a sign. */
866 if (c == '-' || c == '+')
869 if (inchar () == EOF)
870 /* EOF is only an input error before we read any chars. */
883 else if (got_e && wp[wpsize - 1] == 'e'
884 && (c == '-' || c == '+'))
886 else if (wpsize > 0 && !got_e && tolower (c) == 'e')
891 else if (c == decimal && !got_dot)
896 else if ((flags & GROUP) && c == thousands && !got_dot)
903 while (inchar () != EOF && width != 0);
905 /* The last read character is not part of the number anymore. */
911 /* Convert the number. */
915 long double d = __strtold_internal (wp, &tw, flags & GROUP);
916 if (!(flags & SUPPRESS) && tw != wp)
917 *ARG (long double *) = negative ? -d : d;
919 else if (flags & LONG)
921 double d = __strtod_internal (wp, &tw, flags & GROUP);
922 if (!(flags & SUPPRESS) && tw != wp)
923 *ARG (double *) = negative ? -d : d;
927 float d = __strtof_internal (wp, &tw, flags & GROUP);
928 if (!(flags & SUPPRESS) && tw != wp)
929 *ARG (float *) = negative ? -d : d;
935 if (!(flags & SUPPRESS))
939 case '[': /* Character class. */
942 STRING_ARG (wstr, wchar_t);
943 c = '\0'; /* This is to keep gcc quiet. */
947 STRING_ARG (str, char);
962 /* Fill WP with byte flags indexed by character.
963 We will use this flag map for matching input characters. */
964 if (wpmax < UCHAR_MAX)
967 wp = (char *) alloca (wpmax);
969 memset (wp, 0, UCHAR_MAX);
972 if (fc == ']' || fc == '-')
974 /* If ] or - appears before any char in the set, it is not
975 the terminator or separator, but the first char in the
981 while ((fc = *f++) != '\0' && fc != ']')
983 if (fc == '-' && *f != '\0' && *f != ']' &&
984 (unsigned char) f[-2] <= (unsigned char) *f)
986 /* Add all characters from the one before the '-'
987 up to (but not including) the next format char. */
988 for (fc = f[-2]; fc < *f; ++fc)
992 /* Add the character to the flag map. */
1010 NEXT_WIDE_CHAR (first);
1011 if (val > 255 || wp[val] == not_in)
1013 /* XXX We have a problem here. We read a wide
1014 character and this possibly took several
1015 bytes. But we can only push back one single
1016 character. To be sure we don't create wrong
1017 input we push it back only in case it is
1018 representable within one byte. */
1023 STRING_ADD_CHAR (wstr, val, wchar_t);
1033 if (!(flags & SUPPRESS))
1041 num.ul = read_in - 1; /* -1 because we already read one char. */
1044 if (wp[c] == not_in)
1049 STRING_ADD_CHAR (str, c, char);
1053 while (width != 0 && inchar () != EOF);
1055 if (read_in == num.ul)
1058 if (!(flags & SUPPRESS))
1066 case 'p': /* Generic pointer. */
1068 /* A PTR must be the same size as a `long int'. */
1069 flags &= ~(SHORT|LONGDBL);
1076 /* The last thing we saw int the format string was a white space.
1077 Consume the last white spaces. */
1082 while (isspace (c));
1086 /* Unlock stream. */
1094 __vfscanf (FILE *s, const char *format, va_list argptr)
1096 return _IO_vfscanf (s, format, argptr, NULL);
1100 weak_alias (__vfscanf, vfscanf)