stdlib/strtol_l.c

   1 /* Convert string representing a number to integer value, using given locale.
   2    Copyright (C) 1997-2014 Free Software Foundation, Inc.
   3    This file is part of the GNU C Library.
   4    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1997.
   5
   6    The GNU C Library is free software; you can redistribute it and/or
   7    modify it under the terms of the GNU Lesser General Public
   8    License as published by the Free Software Foundation; either
   9    version 2.1 of the License, or (at your option) any later version.
  10
  11    The GNU C Library is distributed in the hope that it will be useful,
  12    but WITHOUT ANY WARRANTY; without even the implied warranty of
  13    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  14    Lesser General Public License for more details.
  15
  16    You should have received a copy of the GNU Lesser General Public
  17    License along with the GNU C Library; if not, see
  18    <http://www.gnu.org/licenses/>.  */
  19
  20
  21 #if HAVE_CONFIG_H
  22 # include <config.h>
  23 #endif
  24
  25 #ifdef _LIBC
  26 # define USE_NUMBER_GROUPING
  27 # define STDC_HEADERS
  28 # define HAVE_LIMITS_H
  29 #endif
  30
  31 #include <ctype.h>
  32 #include <errno.h>
  33 #ifndef __set_errno
  34 # define __set_errno(Val) errno = (Val)
  35 #endif
  36
  37 #ifdef HAVE_LIMITS_H
  38 # include <limits.h>
  39 #endif
  40
  41 #include <stddef.h>
  42 #include <stdlib.h>
  43 #include <string.h>
  44 #include <locale.h>
  45 #include <xlocale.h>
  46 #include <stdint.h>
  47 #include <bits/wordsize.h>
  48
  49 #ifdef USE_NUMBER_GROUPING
  50 # include "../locale/localeinfo.h"
  51 #endif
  52
  53 /* Nonzero if we are defining `strtoul' or `strtoull', operating on
  54    unsigned integers.  */
  55 #ifndef UNSIGNED
  56 # define UNSIGNED 0
  57 # define INT LONG int
  58 #else
  59 # define INT unsigned LONG int
  60 #endif
  61
  62 /* Determine the name.  */
  63 #if UNSIGNED
  64 # ifdef USE_WIDE_CHAR
  65 #  ifdef QUAD
  66 #   define strtol_l wcstoull_l
  67 #  else
  68 #   define strtol_l wcstoul_l
  69 #  endif
  70 # else
  71 #  ifdef QUAD
  72 #   define strtol_l strtoull_l
  73 #  else
  74 #   define strtol_l strtoul_l
  75 #  endif
  76 # endif
  77 #else
  78 # ifdef USE_WIDE_CHAR
  79 #  ifdef QUAD
  80 #   define strtol_l wcstoll_l
  81 #  else
  82 #   define strtol_l wcstol_l
  83 #  endif
  84 # else
  85 #  ifdef QUAD
  86 #   define strtol_l strtoll_l
  87 #  else
  88 #   define strtol_l strtol_l
  89 #  endif
  90 # endif
  91 #endif
  92
  93 #define __strtol_l __strtol_l2(strtol_l)
  94 #define __strtol_l2(name) __strtol_l3(name)
  95 #define __strtol_l3(name) __##name
  96
  97
  98 /* If QUAD is defined, we are defining `strtoll' or `strtoull',
  99    operating on `long long int's.  */
 100 #ifdef QUAD
 101 # define LONG long long
 102 # define STRTOL_LONG_MIN LONG_LONG_MIN
 103 # define STRTOL_LONG_MAX LONG_LONG_MAX
 104 # define STRTOL_ULONG_MAX ULONG_LONG_MAX
 105 #else
 106 # define LONG long
 107
 108 # ifndef ULONG_MAX
 109 #  define ULONG_MAX ((unsigned long int) ~(unsigned long int) 0)
 110 # endif
 111 # ifndef LONG_MAX
 112 #  define LONG_MAX ((long int) (ULONG_MAX >> 1))
 113 # endif
 114 # define STRTOL_LONG_MIN LONG_MIN
 115 # define STRTOL_LONG_MAX LONG_MAX
 116 # define STRTOL_ULONG_MAX ULONG_MAX
 117 #endif
 118
 119
 120 /* We use this code for the extended locale handling where the
 121    function gets as an additional argument the locale which has to be
 122    used.  To access the values we have to redefine the _NL_CURRENT and
 123    _NL_CURRENT_WORD macros.  */
 124 #undef _NL_CURRENT
 125 #define _NL_CURRENT(category, item) \
 126   (current->values[_NL_ITEM_INDEX (item)].string)
 127 #undef _NL_CURRENT_WORD
 128 #define _NL_CURRENT_WORD(category, item) \
 129   ((uint32_t) current->values[_NL_ITEM_INDEX (item)].word)
 130
 131 #if defined _LIBC || defined HAVE_WCHAR_H
 132 # include <wchar.h>
 133 #endif
 134
 135 #ifdef USE_WIDE_CHAR
 136 # include <wctype.h>
 137 # define L_(Ch) L##Ch
 138 # define UCHAR_TYPE wint_t
 139 # define STRING_TYPE wchar_t
 140 # define ISSPACE(Ch) __iswspace_l ((Ch), loc)
 141 # define ISALPHA(Ch) __iswalpha_l ((Ch), loc)
 142 # define TOUPPER(Ch) __towupper_l ((Ch), loc)
 143 #else
 144 # if defined _LIBC \
 145    || defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
 146 #  define IN_CTYPE_DOMAIN(c) 1
 147 # else
 148 #  define IN_CTYPE_DOMAIN(c) isascii(c)
 149 # endif
 150 # define L_(Ch) Ch
 151 # define UCHAR_TYPE unsigned char
 152 # define STRING_TYPE char
 153 # define ISSPACE(Ch) __isspace_l ((Ch), loc)
 154 # define ISALPHA(Ch) __isalpha_l ((Ch), loc)
 155 # define TOUPPER(Ch) __toupper_l ((Ch), loc)
 156 #endif
 157
 158 #define INTERNAL(X) INTERNAL1(X)
 159 #define INTERNAL1(X) __##X##_internal
 160 #define WEAKNAME(X) WEAKNAME1(X)
 161
 162 #ifdef USE_NUMBER_GROUPING
 163 /* This file defines a function to check for correct grouping.  */
 164 # include "grouping.h"
 165 #endif
 166
 167
 168 /* Define tables of maximum values and remainders in order to detect
 169    overflow.  Do this at compile-time in order to avoid the runtime
 170    overhead of the division.  */
 171 extern const unsigned long __strtol_ul_max_tab[] attribute_hidden;
 172 extern const unsigned char __strtol_ul_rem_tab[] attribute_hidden;
 173 #if defined(QUAD) && __WORDSIZE == 32
 174 extern const unsigned long long __strtol_ull_max_tab[] attribute_hidden;
 175 extern const unsigned char __strtol_ull_rem_tab[] attribute_hidden;
 176 #endif
 177
 178 #define DEF(TYPE, NAME)                                                    \
 179   const TYPE NAME[] attribute_hidden =                                     \
 180   {                                                                        \
 181     F(2), F(3), F(4), F(5), F(6), F(7), F(8), F(9), F(10),                 \
 182     F(11), F(12), F(13), F(14), F(15), F(16), F(17), F(18), F(19), F(20),  \
 183     F(21), F(22), F(23), F(24), F(25), F(26), F(27), F(28), F(29), F(30),  \
 184     F(31), F(32), F(33), F(34), F(35), F(36)                               \
 185   }
 186
 187 #if !UNSIGNED && !defined (USE_WIDE_CHAR) && !defined (QUAD)
 188 # define F(X)   ULONG_MAX / X
 189   DEF (unsigned long, __strtol_ul_max_tab);
 190 # undef F
 191 # define F(X)   ULONG_MAX % X
 192   DEF (unsigned char, __strtol_ul_rem_tab);
 193 # undef F
 194 #endif
 195 #if !UNSIGNED && !defined (USE_WIDE_CHAR) && defined (QUAD) \
 196     && __WORDSIZE == 32
 197 # define F(X)   ULONG_LONG_MAX / X
 198   DEF (unsigned long long, __strtol_ull_max_tab);
 199 # undef F
 200 # define F(X)   ULONG_LONG_MAX % X
 201   DEF (unsigned char, __strtol_ull_rem_tab);
 202 # undef F
 203 #endif
 204 #undef DEF
 205
 206 /* Define some more readable aliases for these arrays which correspond
 207    to how they'll be used in the function below.  */
 208 #define jmax_tab        __strtol_ul_max_tab
 209 #if defined(QUAD) && __WORDSIZE == 32
 210 # define cutoff_tab     __strtol_ull_max_tab
 211 # define cutlim_tab     __strtol_ull_rem_tab
 212 #else
 213 # define cutoff_tab     __strtol_ul_max_tab
 214 # define cutlim_tab     __strtol_ul_rem_tab
 215 #endif
 216
 217
 218 /* Convert NPTR to an `unsigned long int' or `long int' in base BASE.
 219    If BASE is 0 the base is determined by the presence of a leading
 220    zero, indicating octal or a leading "0x" or "0X", indicating hexadecimal.
 221    If BASE is < 2 or > 36, it is reset to 10.
 222    If ENDPTR is not NULL, a pointer to the character after the last
 223    one converted is stored in *ENDPTR.  */
 224
 225 INT
 226 INTERNAL (__strtol_l) (nptr, endptr, base, group, loc)
 227      const STRING_TYPE *nptr;
 228      STRING_TYPE **endptr;
 229      int base;
 230      int group;
 231      __locale_t loc;
 232 {
 233   int negative;
 234   unsigned LONG int cutoff;
 235   unsigned int cutlim;
 236   unsigned LONG int i;
 237   const STRING_TYPE *s;
 238   UCHAR_TYPE c;
 239   const STRING_TYPE *save, *end;
 240   int overflow;
 241 #ifndef USE_WIDE_CHAR
 242   size_t cnt;
 243 #endif
 244
 245 #ifdef USE_NUMBER_GROUPING
 246   struct __locale_data *current = loc->__locales[LC_NUMERIC];
 247   /* The thousands character of the current locale.  */
 248 # ifdef USE_WIDE_CHAR
 249   wchar_t thousands = L'\0';
 250 # else
 251   const char *thousands = NULL;
 252   size_t thousands_len = 0;
 253 # endif
 254   /* The numeric grouping specification of the current locale,
 255      in the format described in <locale.h>.  */
 256   const char *grouping;
 257
 258   if (__builtin_expect (group, 0))
 259     {
 260       grouping = _NL_CURRENT (LC_NUMERIC, GROUPING);
 261       if (*grouping <= 0 || *grouping == CHAR_MAX)
 262         grouping = NULL;
 263       else
 264         {
 265           /* Figure out the thousands separator character.  */
 266 # ifdef USE_WIDE_CHAR
 267 #  ifdef _LIBC
 268           thousands = _NL_CURRENT_WORD (LC_NUMERIC,
 269                                         _NL_NUMERIC_THOUSANDS_SEP_WC);
 270 #  endif
 271           if (thousands == L'\0')
 272             grouping = NULL;
 273 # else
 274 #  ifdef _LIBC
 275           thousands = _NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP);
 276 #  endif
 277           if (*thousands == '\0')
 278             {
 279               thousands = NULL;
 280               grouping = NULL;
 281             }
 282 # endif
 283         }
 284     }
 285   else
 286     grouping = NULL;
 287 #endif
 288
 289   if (base < 0 || base == 1 || base > 36)
 290     {
 291       __set_errno (EINVAL);
 292       return 0;
 293     }
 294
 295   save = s = nptr;
 296
 297   /* Skip white space.  */
 298   while (ISSPACE (*s))
 299     ++s;
 300   if (__builtin_expect (*s == L_('\0'), 0))
 301     goto noconv;
 302
 303   /* Check for a sign.  */
 304   negative = 0;
 305   if (*s == L_('-'))
 306     {
 307       negative = 1;
 308       ++s;
 309     }
 310   else if (*s == L_('+'))
 311     ++s;
 312
 313   /* Recognize number prefix and if BASE is zero, figure it out ourselves.  */
 314   if (*s == L_('0'))
 315     {
 316       if ((base == 0 || base == 16) && TOUPPER (s[1]) == L_('X'))
 317         {
 318           s += 2;
 319           base = 16;
 320         }
 321       else if (base == 0)
 322         base = 8;
 323     }
 324   else if (base == 0)
 325     base = 10;
 326
 327   /* Save the pointer so we can check later if anything happened.  */
 328   save = s;
 329
 330 #ifdef USE_NUMBER_GROUPING
 331   if (base != 10)
 332     grouping = NULL;
 333
 334   if (__builtin_expect (grouping != NULL, 0))
 335     {
 336 # ifndef USE_WIDE_CHAR
 337       thousands_len = strlen (thousands);
 338 # endif
 339
 340       /* Find the end of the digit string and check its grouping.  */
 341       end = s;
 342       if (
 343 # ifdef USE_WIDE_CHAR
 344           *s != thousands
 345 # else
 346           ({ for (cnt = 0; cnt < thousands_len; ++cnt)
 347                if (thousands[cnt] != end[cnt])
 348                  break;
 349              cnt < thousands_len; })
 350 # endif
 351           )
 352         {
 353           for (c = *end; c != L_('\0'); c = *++end)
 354             if (((STRING_TYPE) c < L_('0') || (STRING_TYPE) c > L_('9'))
 355 # ifdef USE_WIDE_CHAR
 356                 && (wchar_t) c != thousands
 357 # else
 358                 && ({ for (cnt = 0; cnt < thousands_len; ++cnt)
 359                       if (thousands[cnt] != end[cnt])
 360                         break;
 361                       cnt < thousands_len; })
 362 # endif
 363                 && (!ISALPHA (c)
 364                     || (int) (TOUPPER (c) - L_('A') + 10) >= base))
 365               break;
 366
 367 # ifdef USE_WIDE_CHAR
 368           end = __correctly_grouped_prefixwc (s, end, thousands, grouping);
 369 # else
 370           end = __correctly_grouped_prefixmb (s, end, thousands, grouping);
 371 # endif
 372         }
 373     }
 374   else
 375 #endif
 376     end = NULL;
 377
 378   /* Avoid runtime division; lookup cutoff and limit.  */
 379   cutoff = cutoff_tab[base - 2];
 380   cutlim = cutlim_tab[base - 2];
 381
 382   overflow = 0;
 383   i = 0;
 384   c = *s;
 385   if (sizeof (long int) != sizeof (LONG int))
 386     {
 387       unsigned long int j = 0;
 388       unsigned long int jmax = jmax_tab[base - 2];
 389
 390       for (;c != L_('\0'); c = *++s)
 391         {
 392           if (s == end)
 393             break;
 394           if (c >= L_('0') && c <= L_('9'))
 395             c -= L_('0');
 396 #ifdef USE_NUMBER_GROUPING
 397 # ifdef USE_WIDE_CHAR
 398           else if (grouping && (wchar_t) c == thousands)
 399             continue;
 400 # else
 401           else if (thousands_len)
 402             {
 403               for (cnt = 0; cnt < thousands_len; ++cnt)
 404                 if (thousands[cnt] != s[cnt])
 405                   break;
 406               if (cnt == thousands_len)
 407                 {
 408                   s += thousands_len - 1;
 409                   continue;
 410                 }
 411               if (ISALPHA (c))
 412                 c = TOUPPER (c) - L_('A') + 10;
 413               else
 414                 break;
 415             }
 416 # endif
 417 #endif
 418           else if (ISALPHA (c))
 419             c = TOUPPER (c) - L_('A') + 10;
 420           else
 421             break;
 422           if ((int) c >= base)
 423             break;
 424           /* Note that we never can have an overflow.  */
 425           else if (j >= jmax)
 426             {
 427               /* We have an overflow.  Now use the long representation.  */
 428               i = (unsigned LONG int) j;
 429               goto use_long;
 430             }
 431           else
 432             j = j * (unsigned long int) base + c;
 433         }
 434
 435       i = (unsigned LONG int) j;
 436     }
 437   else
 438     for (;c != L_('\0'); c = *++s)
 439       {
 440         if (s == end)
 441           break;
 442         if (c >= L_('0') && c <= L_('9'))
 443           c -= L_('0');
 444 #ifdef USE_NUMBER_GROUPING
 445 # ifdef USE_WIDE_CHAR
 446         else if (grouping && (wchar_t) c == thousands)
 447           continue;
 448 # else
 449         else if (thousands_len)
 450           {
 451             for (cnt = 0; cnt < thousands_len; ++cnt)
 452               if (thousands[cnt] != s[cnt])
 453                 break;
 454             if (cnt == thousands_len)
 455               {
 456                 s += thousands_len - 1;
 457                 continue;
 458               }
 459             if (ISALPHA (c))
 460               c = TOUPPER (c) - L_('A') + 10;
 461             else
 462               break;
 463           }
 464 # endif
 465 #endif
 466         else if (ISALPHA (c))
 467           c = TOUPPER (c) - L_('A') + 10;
 468         else
 469           break;
 470         if ((int) c >= base)
 471           break;
 472         /* Check for overflow.  */
 473         if (i > cutoff || (i == cutoff && c > cutlim))
 474           overflow = 1;
 475         else
 476           {
 477           use_long:
 478             i *= (unsigned LONG int) base;
 479             i += c;
 480           }
 481       }
 482
 483   /* Check if anything actually happened.  */
 484   if (s == save)
 485     goto noconv;
 486
 487   /* Store in ENDPTR the address of one character
 488      past the last character we converted.  */
 489   if (endptr != NULL)
 490     *endptr = (STRING_TYPE *) s;
 491
 492 #if !UNSIGNED
 493   /* Check for a value that is within the range of
 494      `unsigned LONG int', but outside the range of `LONG int'.  */
 495   if (overflow == 0
 496       && i > (negative
 497               ? -((unsigned LONG int) (STRTOL_LONG_MIN + 1)) + 1
 498               : (unsigned LONG int) STRTOL_LONG_MAX))
 499     overflow = 1;
 500 #endif
 501
 502   if (__builtin_expect (overflow, 0))
 503     {
 504       __set_errno (ERANGE);
 505 #if UNSIGNED
 506       return STRTOL_ULONG_MAX;
 507 #else
 508       return negative ? STRTOL_LONG_MIN : STRTOL_LONG_MAX;
 509 #endif
 510     }
 511
 512   /* Return the result of the appropriate sign.  */
 513   return negative ? -i : i;
 514
 515 noconv:
 516   /* We must handle a special case here: the base is 0 or 16 and the
 517      first two characters are '0' and 'x', but the rest are no
 518      hexadecimal digits.  This is no error case.  We return 0 and
 519      ENDPTR points to the `x`.  */
 520   if (endptr != NULL)
 521     {
 522       if (save - nptr >= 2 && TOUPPER (save[-1]) == L_('X')
 523           && save[-2] == L_('0'))
 524         *endptr = (STRING_TYPE *) &save[-1];
 525       else
 526         /*  There was no number to convert.  */
 527         *endptr = (STRING_TYPE *) nptr;
 528     }
 529
 530   return 0L;
 531 }
 532 #if defined _LIBC && !defined USE_WIDE_CHAR
 533 libc_hidden_def (INTERNAL (__strtol_l))
 534 #endif
 535 \f
 536 /* External user entry point.  */
 537
 538 #if _LIBC - 0 == 0
 539
 540 /* Prototype.  */
 541 extern INT __strtol_l (const STRING_TYPE *nptr, STRING_TYPE **endptr,
 542                        int base);
 543 #endif
 544
 545
 546 INT
 547 #ifdef weak_function
 548 weak_function
 549 #endif
 550 __strtol_l (nptr, endptr, base, loc)
 551      const STRING_TYPE *nptr;
 552      STRING_TYPE **endptr;
 553      int base;
 554      __locale_t loc;
 555 {
 556   return INTERNAL (__strtol_l) (nptr, endptr, base, 0, loc);
 557 }
 558 libc_hidden_def (__strtol_l)
 559 weak_alias (__strtol_l, strtol_l)