lib/strtol.c

   1 /* Convert string representation of a number into an integer value.
   2
   3    Copyright (C) 1991-1992, 1994-1999, 2003, 2005-2007, 2009-2011 Free Software
   4    Foundation, Inc.
   5
   6    NOTE: The canonical source of this file is maintained with the GNU C
   7    Library.  Bugs can be reported to bug-glibc@gnu.org.
   8
   9    This program is free software: you can redistribute it and/or modify it
  10    under the terms of the GNU General Public License as published by the
  11    Free Software Foundation; either version 3 of the License, or any
  12    later version.
  13
  14    This program is distributed in the hope that it will be useful,
  15    but WITHOUT ANY WARRANTY; without even the implied warranty of
  16    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17    GNU General Public License for more details.
  18
  19    You should have received a copy of the GNU General Public License
  20    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
  21
  22 #ifdef _LIBC
  23 # define USE_NUMBER_GROUPING
  24 #else
  25 # include <config.h>
  26 #endif
  27
  28 #include <ctype.h>
  29 #include <errno.h>
  30 #ifndef __set_errno
  31 # define __set_errno(Val) errno = (Val)
  32 #endif
  33
  34 #include <limits.h>
  35 #include <stddef.h>
  36 #include <stdlib.h>
  37 #include <string.h>
  38
  39 #ifdef USE_NUMBER_GROUPING
  40 # include "../locale/localeinfo.h"
  41 #endif
  42
  43 /* Nonzero if we are defining `strtoul' or `strtoull', operating on
  44    unsigned integers.  */
  45 #ifndef UNSIGNED
  46 # define UNSIGNED 0
  47 # define INT LONG int
  48 #else
  49 # define INT unsigned LONG int
  50 #endif
  51
  52 /* Determine the name.  */
  53 #ifdef USE_IN_EXTENDED_LOCALE_MODEL
  54 # if UNSIGNED
  55 #  ifdef USE_WIDE_CHAR
  56 #   ifdef QUAD
  57 #    define strtol __wcstoull_l
  58 #   else
  59 #    define strtol __wcstoul_l
  60 #   endif
  61 #  else
  62 #   ifdef QUAD
  63 #    define strtol __strtoull_l
  64 #   else
  65 #    define strtol __strtoul_l
  66 #   endif
  67 #  endif
  68 # else
  69 #  ifdef USE_WIDE_CHAR
  70 #   ifdef QUAD
  71 #    define strtol __wcstoll_l
  72 #   else
  73 #    define strtol __wcstol_l
  74 #   endif
  75 #  else
  76 #   ifdef QUAD
  77 #    define strtol __strtoll_l
  78 #   else
  79 #    define strtol __strtol_l
  80 #   endif
  81 #  endif
  82 # endif
  83 #else
  84 # if UNSIGNED
  85 #  ifdef USE_WIDE_CHAR
  86 #   ifdef QUAD
  87 #    define strtol wcstoull
  88 #   else
  89 #    define strtol wcstoul
  90 #   endif
  91 #  else
  92 #   ifdef QUAD
  93 #    define strtol strtoull
  94 #   else
  95 #    define strtol strtoul
  96 #   endif
  97 #  endif
  98 # else
  99 #  ifdef USE_WIDE_CHAR
 100 #   ifdef QUAD
 101 #    define strtol wcstoll
 102 #   else
 103 #    define strtol wcstol
 104 #   endif
 105 #  else
 106 #   ifdef QUAD
 107 #    define strtol strtoll
 108 #   endif
 109 #  endif
 110 # endif
 111 #endif
 112
 113 /* If QUAD is defined, we are defining `strtoll' or `strtoull',
 114    operating on `long long int's.  */
 115 #ifdef QUAD
 116 # define LONG long long
 117 # define STRTOL_LONG_MIN LLONG_MIN
 118 # define STRTOL_LONG_MAX LLONG_MAX
 119 # define STRTOL_ULONG_MAX ULLONG_MAX
 120
 121 /* The extra casts in the following macros work around compiler bugs,
 122    e.g., in Cray C 5.0.3.0.  */
 123
 124 /* True if negative values of the signed integer type T use two's
 125    complement, ones' complement, or signed magnitude representation,
 126    respectively.  Much GNU code assumes two's complement, but some
 127    people like to be portable to all possible C hosts.  */
 128 # define TYPE_TWOS_COMPLEMENT(t) ((t) ~ (t) 0 == (t) -1)
 129 # define TYPE_ONES_COMPLEMENT(t) ((t) ~ (t) 0 == 0)
 130 # define TYPE_SIGNED_MAGNITUDE(t) ((t) ~ (t) 0 < (t) -1)
 131
 132 /* True if the arithmetic type T is signed.  */
 133 # define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
 134
 135 /* The maximum and minimum values for the integer type T.  These
 136    macros have undefined behavior if T is signed and has padding bits.
 137    If this is a problem for you, please let us know how to fix it for
 138    your host.  */
 139 # define TYPE_MINIMUM(t) \
 140    ((t) (! TYPE_SIGNED (t) \
 141          ? (t) 0 \
 142          : TYPE_SIGNED_MAGNITUDE (t) \
 143          ? ~ (t) 0 \
 144          : ~ TYPE_MAXIMUM (t)))
 145 # define TYPE_MAXIMUM(t) \
 146    ((t) (! TYPE_SIGNED (t) \
 147          ? (t) -1 \
 148          : ((((t) 1 << (sizeof (t) * CHAR_BIT - 2)) - 1) * 2 + 1)))
 149
 150 # ifndef ULLONG_MAX
 151 #  define ULLONG_MAX TYPE_MAXIMUM (unsigned long long)
 152 # endif
 153 # ifndef LLONG_MAX
 154 #  define LLONG_MAX TYPE_MAXIMUM (long long int)
 155 # endif
 156 # ifndef LLONG_MIN
 157 #  define LLONG_MIN TYPE_MINIMUM (long long int)
 158 # endif
 159
 160 # if __GNUC__ == 2 && __GNUC_MINOR__ < 7
 161    /* Work around gcc bug with using this constant.  */
 162    static const unsigned long long int maxquad = ULLONG_MAX;
 163 #  undef STRTOL_ULONG_MAX
 164 #  define STRTOL_ULONG_MAX maxquad
 165 # endif
 166 #else
 167 # define LONG long
 168 # define STRTOL_LONG_MIN LONG_MIN
 169 # define STRTOL_LONG_MAX LONG_MAX
 170 # define STRTOL_ULONG_MAX ULONG_MAX
 171 #endif
 172
 173
 174 /* We use this code also for the extended locale handling where the
 175    function gets as an additional argument the locale which has to be
 176    used.  To access the values we have to redefine the _NL_CURRENT
 177    macro.  */
 178 #ifdef USE_IN_EXTENDED_LOCALE_MODEL
 179 # undef _NL_CURRENT
 180 # define _NL_CURRENT(category, item) \
 181   (current->values[_NL_ITEM_INDEX (item)].string)
 182 # define LOCALE_PARAM , loc
 183 # define LOCALE_PARAM_PROTO , __locale_t loc
 184 #else
 185 # define LOCALE_PARAM
 186 # define LOCALE_PARAM_PROTO
 187 #endif
 188
 189 #ifdef USE_WIDE_CHAR
 190 # include <wchar.h>
 191 # include <wctype.h>
 192 # define L_(Ch) L##Ch
 193 # define UCHAR_TYPE wint_t
 194 # define STRING_TYPE wchar_t
 195 # ifdef USE_IN_EXTENDED_LOCALE_MODEL
 196 #  define ISSPACE(Ch) __iswspace_l ((Ch), loc)
 197 #  define ISALPHA(Ch) __iswalpha_l ((Ch), loc)
 198 #  define TOUPPER(Ch) __towupper_l ((Ch), loc)
 199 # else
 200 #  define ISSPACE(Ch) iswspace (Ch)
 201 #  define ISALPHA(Ch) iswalpha (Ch)
 202 #  define TOUPPER(Ch) towupper (Ch)
 203 # endif
 204 #else
 205 # define L_(Ch) Ch
 206 # define UCHAR_TYPE unsigned char
 207 # define STRING_TYPE char
 208 # ifdef USE_IN_EXTENDED_LOCALE_MODEL
 209 #  define ISSPACE(Ch) __isspace_l ((Ch), loc)
 210 #  define ISALPHA(Ch) __isalpha_l ((Ch), loc)
 211 #  define TOUPPER(Ch) __toupper_l ((Ch), loc)
 212 # else
 213 #  define ISSPACE(Ch) isspace (Ch)
 214 #  define ISALPHA(Ch) isalpha (Ch)
 215 #  define TOUPPER(Ch) toupper (Ch)
 216 # endif
 217 #endif
 218
 219 #define INTERNAL(X) INTERNAL1(X)
 220 #define INTERNAL1(X) __##X##_internal
 221 #define WEAKNAME(X) WEAKNAME1(X)
 222
 223 #ifdef USE_NUMBER_GROUPING
 224 /* This file defines a function to check for correct grouping.  */
 225 # include "grouping.h"
 226 #endif
 227
 228
 229
 230 /* Convert NPTR to an `unsigned long int' or `long int' in base BASE.
 231    If BASE is 0 the base is determined by the presence of a leading
 232    zero, indicating octal or a leading "0x" or "0X", indicating hexadecimal.
 233    If BASE is < 2 or > 36, it is reset to 10.
 234    If ENDPTR is not NULL, a pointer to the character after the last
 235    one converted is stored in *ENDPTR.  */
 236
 237 INT
 238 INTERNAL (strtol) (const STRING_TYPE *nptr, STRING_TYPE **endptr,
 239                    int base, int group LOCALE_PARAM_PROTO)
 240 {
 241   int negative;
 242   register unsigned LONG int cutoff;
 243   register unsigned int cutlim;
 244   register unsigned LONG int i;
 245   register const STRING_TYPE *s;
 246   register UCHAR_TYPE c;
 247   const STRING_TYPE *save, *end;
 248   int overflow;
 249
 250 #ifdef USE_NUMBER_GROUPING
 251 # ifdef USE_IN_EXTENDED_LOCALE_MODEL
 252   struct locale_data *current = loc->__locales[LC_NUMERIC];
 253 # endif
 254   /* The thousands character of the current locale.  */
 255   wchar_t thousands = L'\0';
 256   /* The numeric grouping specification of the current locale,
 257      in the format described in <locale.h>.  */
 258   const char *grouping;
 259
 260   if (group)
 261     {
 262       grouping = _NL_CURRENT (LC_NUMERIC, GROUPING);
 263       if (*grouping <= 0 || *grouping == CHAR_MAX)
 264         grouping = NULL;
 265       else
 266         {
 267           /* Figure out the thousands separator character.  */
 268 # if defined _LIBC || defined _HAVE_BTOWC
 269           thousands = __btowc (*_NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP));
 270           if (thousands == WEOF)
 271             thousands = L'\0';
 272 # endif
 273           if (thousands == L'\0')
 274             grouping = NULL;
 275         }
 276     }
 277   else
 278     grouping = NULL;
 279 #endif
 280
 281   if (base < 0 || base == 1 || base > 36)
 282     {
 283       __set_errno (EINVAL);
 284       return 0;
 285     }
 286
 287   save = s = nptr;
 288
 289   /* Skip white space.  */
 290   while (ISSPACE (*s))
 291     ++s;
 292   if (*s == L_('\0'))
 293     goto noconv;
 294
 295   /* Check for a sign.  */
 296   if (*s == L_('-'))
 297     {
 298       negative = 1;
 299       ++s;
 300     }
 301   else if (*s == L_('+'))
 302     {
 303       negative = 0;
 304       ++s;
 305     }
 306   else
 307     negative = 0;
 308
 309   /* Recognize number prefix and if BASE is zero, figure it out ourselves.  */
 310   if (*s == L_('0'))
 311     {
 312       if ((base == 0 || base == 16) && TOUPPER (s[1]) == L_('X'))
 313         {
 314           s += 2;
 315           base = 16;
 316         }
 317       else if (base == 0)
 318         base = 8;
 319     }
 320   else if (base == 0)
 321     base = 10;
 322
 323   /* Save the pointer so we can check later if anything happened.  */
 324   save = s;
 325
 326 #ifdef USE_NUMBER_GROUPING
 327   if (group)
 328     {
 329       /* Find the end of the digit string and check its grouping.  */
 330       end = s;
 331       for (c = *end; c != L_('\0'); c = *++end)
 332         if ((wchar_t) c != thousands
 333             && ((wchar_t) c < L_('0') || (wchar_t) c > L_('9'))
 334             && (!ISALPHA (c) || (int) (TOUPPER (c) - L_('A') + 10) >= base))
 335           break;
 336       if (*s == thousands)
 337         end = s;
 338       else
 339         end = correctly_grouped_prefix (s, end, thousands, grouping);
 340     }
 341   else
 342 #endif
 343     end = NULL;
 344
 345   cutoff = STRTOL_ULONG_MAX / (unsigned LONG int) base;
 346   cutlim = STRTOL_ULONG_MAX % (unsigned LONG int) base;
 347
 348   overflow = 0;
 349   i = 0;
 350   for (c = *s; c != L_('\0'); c = *++s)
 351     {
 352       if (s == end)
 353         break;
 354       if (c >= L_('0') && c <= L_('9'))
 355         c -= L_('0');
 356       else if (ISALPHA (c))
 357         c = TOUPPER (c) - L_('A') + 10;
 358       else
 359         break;
 360       if ((int) c >= base)
 361         break;
 362       /* Check for overflow.  */
 363       if (i > cutoff || (i == cutoff && c > cutlim))
 364         overflow = 1;
 365       else
 366         {
 367           i *= (unsigned LONG int) base;
 368           i += c;
 369         }
 370     }
 371
 372   /* Check if anything actually happened.  */
 373   if (s == save)
 374     goto noconv;
 375
 376   /* Store in ENDPTR the address of one character
 377      past the last character we converted.  */
 378   if (endptr != NULL)
 379     *endptr = (STRING_TYPE *) s;
 380
 381 #if !UNSIGNED
 382   /* Check for a value that is within the range of
 383      `unsigned LONG int', but outside the range of `LONG int'.  */
 384   if (overflow == 0
 385       && i > (negative
 386               ? -((unsigned LONG int) (STRTOL_LONG_MIN + 1)) + 1
 387               : (unsigned LONG int) STRTOL_LONG_MAX))
 388     overflow = 1;
 389 #endif
 390
 391   if (overflow)
 392     {
 393       __set_errno (ERANGE);
 394 #if UNSIGNED
 395       return STRTOL_ULONG_MAX;
 396 #else
 397       return negative ? STRTOL_LONG_MIN : STRTOL_LONG_MAX;
 398 #endif
 399     }
 400
 401   /* Return the result of the appropriate sign.  */
 402   return negative ? -i : i;
 403
 404 noconv:
 405   /* We must handle a special case here: the base is 0 or 16 and the
 406      first two characters are '0' and 'x', but the rest are no
 407      hexadecimal digits.  This is no error case.  We return 0 and
 408      ENDPTR points to the `x`.  */
 409   if (endptr != NULL)
 410     {
 411       if (save - nptr >= 2 && TOUPPER (save[-1]) == L_('X')
 412           && save[-2] == L_('0'))
 413         *endptr = (STRING_TYPE *) &save[-1];
 414       else
 415         /*  There was no number to convert.  */
 416         *endptr = (STRING_TYPE *) nptr;
 417     }
 418
 419   return 0L;
 420 }
 421 \f
 422 /* External user entry point.  */
 423
 424
 425 INT
 426 #ifdef weak_function
 427 weak_function
 428 #endif
 429 strtol (const STRING_TYPE *nptr, STRING_TYPE **endptr,
 430         int base LOCALE_PARAM_PROTO)
 431 {
 432   return INTERNAL (strtol) (nptr, endptr, base, 0 LOCALE_PARAM);
 433 }