sysdeps/generic/strtol_l.c

   1 /* Convert string representing a number to integer value, using given locale.
   2    Copyright (C) 1997, 2002, 2004 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, write to the Free
  18    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  19    02111-1307 USA.  */
  20
  21
  22 #if HAVE_CONFIG_H
  23 # include <config.h>
  24 #endif
  25
  26 #ifdef _LIBC
  27 # define USE_NUMBER_GROUPING
  28 # define STDC_HEADERS
  29 # define HAVE_LIMITS_H
  30 #endif
  31
  32 #include <ctype.h>
  33 #include <errno.h>
  34 #ifndef __set_errno
  35 # define __set_errno(Val) errno = (Val)
  36 #endif
  37
  38 #ifdef HAVE_LIMITS_H
  39 # include <limits.h>
  40 #endif
  41
  42 #include <stddef.h>
  43 #include <stdlib.h>
  44 #include <string.h>
  45 #include <locale.h>
  46 #include <xlocale.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
 172 #define DEF(TYPE, NAME)                                                    \
 173   const TYPE NAME[] attribute_hidden                                       \
 174         __attribute__((section(".gnu.linkonce.r." #NAME))) =               \
 175   {                                                                        \
 176     F(2), F(3), F(4), F(5), F(6), F(7), F(8), F(9), F(10),                 \
 177     F(11), F(12), F(13), F(14), F(15), F(16), F(17), F(18), F(19), F(20),  \
 178     F(21), F(22), F(23), F(24), F(25), F(26), F(27), F(28), F(29), F(30),  \
 179     F(31), F(32), F(33), F(34), F(35), F(36)                               \
 180   }
 181
 182 #define F(X)    ULONG_MAX / X
 183   DEF (unsigned long, __strtol_ul_max_tab);
 184 #undef F
 185 #if defined(QUAD) && __WORDSIZE == 32
 186 # define F(X)   ULONG_LONG_MAX / X
 187   DEF (unsigned long long, __strtol_ull_max_tab);
 188 # undef F
 189 # define F(X)   ULONG_LONG_MAX % X
 190   DEF (unsigned char, __strtol_ull_rem_tab);
 191 # undef F
 192 #else
 193 # define F(X)   ULONG_MAX % X
 194   DEF (unsigned char, __strtol_ul_rem_tab);
 195 # undef F
 196 #endif
 197 #undef DEF
 198
 199 /* Define some more readable aliases for these arrays which correspond
 200    to how they'll be used in the function below.  */
 201 #define jmax_tab        __strtol_ul_max_tab
 202 #if defined(QUAD) && __WORDSIZE == 32
 203 # define cutoff_tab     __strtol_ull_max_tab
 204 # define cutlim_tab     __strtol_ull_rem_tab
 205 #else
 206 # define cutoff_tab     __strtol_ul_max_tab
 207 # define cutlim_tab     __strtol_ul_rem_tab
 208 #endif
 209
 210
 211 /* Convert NPTR to an `unsigned long int' or `long int' in base BASE.
 212    If BASE is 0 the base is determined by the presence of a leading
 213    zero, indicating octal or a leading "0x" or "0X", indicating hexadecimal.
 214    If BASE is < 2 or > 36, it is reset to 10.
 215    If ENDPTR is not NULL, a pointer to the character after the last
 216    one converted is stored in *ENDPTR.  */
 217
 218 INT
 219 INTERNAL (__strtol_l) (nptr, endptr, base, group, loc)
 220      const STRING_TYPE *nptr;
 221      STRING_TYPE **endptr;
 222      int base;
 223      int group;
 224      __locale_t loc;
 225 {
 226   int negative;
 227   register unsigned LONG int cutoff;
 228   register unsigned int cutlim;
 229   register unsigned LONG int i;
 230   register const STRING_TYPE *s;
 231   register UCHAR_TYPE c;
 232   const STRING_TYPE *save, *end;
 233   int overflow;
 234 #ifndef USE_WIDE_CHAR
 235   size_t cnt;
 236 #endif
 237
 238 #ifdef USE_NUMBER_GROUPING
 239   struct locale_data *current = loc->__locales[LC_NUMERIC];
 240   /* The thousands character of the current locale.  */
 241 # ifdef USE_WIDE_CHAR
 242   wchar_t thousands = L'\0';
 243 # else
 244   const char *thousands = NULL;
 245   size_t thousands_len = 0;
 246 # endif
 247   /* The numeric grouping specification of the current locale,
 248      in the format described in <locale.h>.  */
 249   const char *grouping;
 250
 251   if (__builtin_expect (group, 0))
 252     {
 253       grouping = _NL_CURRENT (LC_NUMERIC, GROUPING);
 254       if (*grouping <= 0 || *grouping == CHAR_MAX)
 255         grouping = NULL;
 256       else
 257         {
 258           /* Figure out the thousands separator character.  */
 259 # ifdef USE_WIDE_CHAR
 260 #  ifdef _LIBC
 261           thousands = _NL_CURRENT_WORD (LC_NUMERIC,
 262                                         _NL_NUMERIC_THOUSANDS_SEP_WC);
 263 #  endif
 264           if (thousands == L'\0')
 265             grouping = NULL;
 266 # else
 267 #  ifdef _LIBC
 268           thousands = _NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP);
 269 #  endif
 270           if (*thousands == '\0')
 271             {
 272               thousands = NULL;
 273               grouping = NULL;
 274             }
 275 # endif
 276         }
 277     }
 278   else
 279     grouping = NULL;
 280 #endif
 281
 282   if (base < 0 || base == 1 || base > 36)
 283     {
 284       __set_errno (EINVAL);
 285       return 0;
 286     }
 287
 288   save = s = nptr;
 289
 290   /* Skip white space.  */
 291   while (ISSPACE (*s))
 292     ++s;
 293   if (__builtin_expect (*s == L_('\0'), 0))
 294     goto noconv;
 295
 296   /* Check for a sign.  */
 297   negative = 0;
 298   if (*s == L_('-'))
 299     {
 300       negative = 1;
 301       ++s;
 302     }
 303   else if (*s == L_('+'))
 304     ++s;
 305
 306   /* Recognize number prefix and if BASE is zero, figure it out ourselves.  */
 307   if (*s == L_('0'))
 308     {
 309       if ((base == 0 || base == 16) && TOUPPER (s[1]) == L_('X'))
 310         {
 311           s += 2;
 312           base = 16;
 313         }
 314       else if (base == 0)
 315         base = 8;
 316     }
 317   else if (base == 0)
 318     base = 10;
 319
 320   /* Save the pointer so we can check later if anything happened.  */
 321   save = s;
 322
 323 #ifdef USE_NUMBER_GROUPING
 324   if (base != 10)
 325     grouping = NULL;
 326
 327   if (__builtin_expect (grouping != NULL, 0))
 328     {
 329 # ifndef USE_WIDE_CHAR
 330       thousands_len = strlen (thousands);
 331 # endif
 332
 333       /* Find the end of the digit string and check its grouping.  */
 334       end = s;
 335       if (
 336 # ifdef USE_WIDE_CHAR
 337           *s != thousands
 338 # else
 339           ({ for (cnt = 0; cnt < thousands_len; ++cnt)
 340                if (thousands[cnt] != end[cnt])
 341                  break;
 342              cnt < thousands_len; })
 343 # endif
 344           )
 345         {
 346           for (c = *end; c != L_('\0'); c = *++end)
 347             if (((STRING_TYPE) c < L_('0') || (STRING_TYPE) c > L_('9'))
 348 # ifdef USE_WIDE_CHAR
 349                 && (wchar_t) c != thousands
 350 # else
 351                 && ({ for (cnt = 0; cnt < thousands_len; ++cnt)
 352                       if (thousands[cnt] != end[cnt])
 353                         break;
 354                       cnt < thousands_len; })
 355 # endif
 356                 && (!ISALPHA (c)
 357                     || (int) (TOUPPER (c) - L_('A') + 10) >= base))
 358               break;
 359
 360 # ifdef USE_WIDE_CHAR
 361           end = __correctly_grouped_prefixwc (s, end, thousands, grouping);
 362 # else
 363           end = __correctly_grouped_prefixmb (s, end, thousands, grouping);
 364 # endif
 365         }
 366     }
 367   else
 368 #endif
 369     end = NULL;
 370
 371   /* Avoid runtime division; lookup cutoff and limit.  */
 372   cutoff = cutoff_tab[base - 2];
 373   cutlim = cutlim_tab[base - 2];
 374
 375   overflow = 0;
 376   i = 0;
 377   c = *s;
 378   if (sizeof (long int) != sizeof (LONG int))
 379     {
 380       unsigned long int j = 0;
 381       unsigned long int jmax = jmax_tab[base - 2];
 382
 383       for (;c != L_('\0'); c = *++s)
 384         {
 385           if (s == end)
 386             break;
 387           if (c >= L_('0') && c <= L_('9'))
 388             c -= L_('0');
 389 #ifdef USE_NUMBER_GROUPING
 390 # ifdef USE_WIDE_CHAR
 391           else if (grouping && (wchar_t) c == thousands)
 392             continue;
 393 # else
 394           else if (thousands_len)
 395             {
 396               for (cnt = 0; cnt < thousands_len; ++cnt)
 397                 if (thousands[cnt] != s[cnt])
 398                   break;
 399               if (cnt == thousands_len)
 400                 {
 401                   s += thousands_len - 1;
 402                   continue;
 403                 }
 404               if (ISALPHA (c))
 405                 c = TOUPPER (c) - L_('A') + 10;
 406               else
 407                 break;
 408             }
 409 # endif
 410 #endif
 411           else if (ISALPHA (c))
 412             c = TOUPPER (c) - L_('A') + 10;
 413           else
 414             break;
 415           if ((int) c >= base)
 416             break;
 417           /* Note that we never can have an overflow.  */
 418           else if (j >= jmax)
 419             {
 420               /* We have an overflow.  Now use the long representation.  */
 421               i = (unsigned LONG int) j;
 422               goto use_long;
 423             }
 424           else
 425             j = j * (unsigned long int) base + c;
 426         }
 427
 428       i = (unsigned LONG int) j;
 429     }
 430   else
 431     for (;c != L_('\0'); c = *++s)
 432       {
 433         if (s == end)
 434           break;
 435         if (c >= L_('0') && c <= L_('9'))
 436           c -= L_('0');
 437 #ifdef USE_NUMBER_GROUPING
 438 # ifdef USE_WIDE_CHAR
 439         else if (grouping && (wchar_t) c == thousands)
 440           continue;
 441 # else
 442         else if (thousands_len)
 443           {
 444             for (cnt = 0; cnt < thousands_len; ++cnt)
 445               if (thousands[cnt] != s[cnt])
 446                 break;
 447             if (cnt == thousands_len)
 448               {
 449                 s += thousands_len - 1;
 450                 continue;
 451               }
 452             if (ISALPHA (c))
 453               c = TOUPPER (c) - L_('A') + 10;
 454             else
 455               break;
 456           }
 457 # endif
 458 #endif
 459         else if (ISALPHA (c))
 460           c = TOUPPER (c) - L_('A') + 10;
 461         else
 462           break;
 463         if ((int) c >= base)
 464           break;
 465         /* Check for overflow.  */
 466         if (i > cutoff || (i == cutoff && c > cutlim))
 467           overflow = 1;
 468         else
 469           {
 470           use_long:
 471             i *= (unsigned LONG int) base;
 472             i += c;
 473           }
 474       }
 475
 476   /* Check if anything actually happened.  */
 477   if (s == save)
 478     goto noconv;
 479
 480   /* Store in ENDPTR the address of one character
 481      past the last character we converted.  */
 482   if (endptr != NULL)
 483     *endptr = (STRING_TYPE *) s;
 484
 485 #if !UNSIGNED
 486   /* Check for a value that is within the range of
 487      `unsigned LONG int', but outside the range of `LONG int'.  */
 488   if (overflow == 0
 489       && i > (negative
 490               ? -((unsigned LONG int) (STRTOL_LONG_MIN + 1)) + 1
 491               : (unsigned LONG int) STRTOL_LONG_MAX))
 492     overflow = 1;
 493 #endif
 494
 495   if (__builtin_expect (overflow, 0))
 496     {
 497       __set_errno (ERANGE);
 498 #if UNSIGNED
 499       return STRTOL_ULONG_MAX;
 500 #else
 501       return negative ? STRTOL_LONG_MIN : STRTOL_LONG_MAX;
 502 #endif
 503     }
 504
 505   /* Return the result of the appropriate sign.  */
 506   return negative ? -i : i;
 507
 508 noconv:
 509   /* We must handle a special case here: the base is 0 or 16 and the
 510      first two characters are '0' and 'x', but the rest are no
 511      hexadecimal digits.  This is no error case.  We return 0 and
 512      ENDPTR points to the `x`.  */
 513   if (endptr != NULL)
 514     {
 515       if (save - nptr >= 2 && TOUPPER (save[-1]) == L_('X')
 516           && save[-2] == L_('0'))
 517         *endptr = (STRING_TYPE *) &save[-1];
 518       else
 519         /*  There was no number to convert.  */
 520         *endptr = (STRING_TYPE *) nptr;
 521     }
 522
 523   return 0L;
 524 }
 525 #if defined _LIBC && !defined USE_WIDE_CHAR
 526 libc_hidden_def (INTERNAL (__strtol_l))
 527 #endif
 528 \f
 529 /* External user entry point.  */
 530
 531 #if _LIBC - 0 == 0
 532 # undef PARAMS
 533 # if defined (__STDC__) && __STDC__
 534 #  define PARAMS(Args) Args
 535 # else
 536 #  define PARAMS(Args) ()
 537 # endif
 538
 539 /* Prototype.  */
 540 extern INT __strtol_l PARAMS ((const STRING_TYPE *nptr, STRING_TYPE **endptr,
 541                                int base));
 542 #endif
 543
 544
 545 INT
 546 #ifdef weak_function
 547 weak_function
 548 #endif
 549 __strtol_l (nptr, endptr, base, loc)
 550      const STRING_TYPE *nptr;
 551      STRING_TYPE **endptr;
 552      int base;
 553      __locale_t loc;
 554 {
 555   return INTERNAL (__strtol_l) (nptr, endptr, base, 0, loc);
 556 }
 557 weak_alias (__strtol_l, strtol_l)