1/* Convert string representation of a number into an integer value.
2
3   Copyright (C) 1991, 1992, 1994, 1995, 1996, 1997, 1998, 1999, 2003, 2005
4   Free Software 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 2, or (at your option) 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, write to the Free Software Foundation,
21   Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.  */
22
23#ifdef HAVE_CONFIG_H
24# include <config.h>
25#endif
26
27#ifdef _LIBC
28# define USE_NUMBER_GROUPING
29#endif
30
31#include <ctype.h>
32#include <errno.h>
33#ifndef errno
34extern int errno;
35#endif
36#ifndef __set_errno
37# define __set_errno(Val) errno = (Val)
38#endif
39
40#include <limits.h>
41#include <stddef.h>
42#include <stdlib.h>
43#include <string.h>
44
45#ifdef USE_NUMBER_GROUPING
46# include "../locale/localeinfo.h"
47#endif
48
49/* Nonzero if we are defining `strtoul' or `strtoull', operating on
50   unsigned integers.  */
51#ifndef UNSIGNED
52# define UNSIGNED 0
53# define INT LONG int
54#else
55# define INT unsigned LONG int
56#endif
57
58/* Determine the name.  */
59#ifdef USE_IN_EXTENDED_LOCALE_MODEL
60# if UNSIGNED
61#  ifdef USE_WIDE_CHAR
62#   ifdef QUAD
63#    define strtol __wcstoull_l
64#   else
65#    define strtol __wcstoul_l
66#   endif
67#  else
68#   ifdef QUAD
69#    define strtol __strtoull_l
70#   else
71#    define strtol __strtoul_l
72#   endif
73#  endif
74# else
75#  ifdef USE_WIDE_CHAR
76#   ifdef QUAD
77#    define strtol __wcstoll_l
78#   else
79#    define strtol __wcstol_l
80#   endif
81#  else
82#   ifdef QUAD
83#    define strtol __strtoll_l
84#   else
85#    define strtol __strtol_l
86#   endif
87#  endif
88# endif
89#else
90# if UNSIGNED
91#  ifdef USE_WIDE_CHAR
92#   ifdef QUAD
93#    define strtol wcstoull
94#   else
95#    define strtol wcstoul
96#   endif
97#  else
98#   ifdef QUAD
99#    define strtol strtoull
100#   else
101#    define strtol strtoul
102#   endif
103#  endif
104# else
105#  ifdef USE_WIDE_CHAR
106#   ifdef QUAD
107#    define strtol wcstoll
108#   else
109#    define strtol wcstol
110#   endif
111#  else
112#   ifdef QUAD
113#    define strtol strtoll
114#   endif
115#  endif
116# endif
117#endif
118
119/* If QUAD is defined, we are defining `strtoll' or `strtoull',
120   operating on `long long int's.  */
121#ifdef QUAD
122# define LONG long long
123# define STRTOL_LONG_MIN LONG_LONG_MIN
124# define STRTOL_LONG_MAX LONG_LONG_MAX
125# define STRTOL_ULONG_MAX ULONG_LONG_MAX
126
127/* The extra casts in the following macros work around compiler bugs,
128   e.g., in Cray C 5.0.3.0.  */
129
130/* True if negative values of the signed integer type T use two's
131   complement, ones' complement, or signed magnitude representation,
132   respectively.  Much GNU code assumes two's complement, but some
133   people like to be portable to all possible C hosts.  */
134# define TYPE_TWOS_COMPLEMENT(t) ((t) ~ (t) 0 == (t) -1)
135# define TYPE_ONES_COMPLEMENT(t) ((t) ~ (t) 0 == 0)
136# define TYPE_SIGNED_MAGNITUDE(t) ((t) ~ (t) 0 < (t) -1)
137
138/* True if the arithmetic type T is signed.  */
139# define TYPE_SIGNED(t) (! ((t) 0 < (t) -1))
140
141/* The maximum and minimum values for the integer type T.  These
142   macros have undefined behavior if T is signed and has padding bits.
143   If this is a problem for you, please let us know how to fix it for
144   your host.  */
145# define TYPE_MINIMUM(t) \
146   ((t) (! TYPE_SIGNED (t) \
147	 ? (t) 0 \
148	 : TYPE_SIGNED_MAGNITUDE (t) \
149	 ? ~ (t) 0 \
150	 : ~ (t) 0 << (sizeof (t) * CHAR_BIT - 1)))
151# define TYPE_MAXIMUM(t) \
152   ((t) (! TYPE_SIGNED (t) \
153	 ? (t) -1 \
154	 : ~ (~ (t) 0 << (sizeof (t) * CHAR_BIT - 1))))
155
156# ifndef ULONG_LONG_MAX
157#  define ULONG_LONG_MAX TYPE_MAXIMUM (unsigned long long)
158# endif
159# ifndef LONG_LONG_MAX
160#  define LONG_LONG_MAX TYPE_MAXIMUM (long long int)
161# endif
162# ifndef LONG_LONG_MIN
163#  define LONG_LONG_MIN TYPE_MINIMUM (long long int)
164# endif
165
166# if __GNUC__ == 2 && __GNUC_MINOR__ < 7
167   /* Work around gcc bug with using this constant.  */
168   static const unsigned long long int maxquad = ULONG_LONG_MAX;
169#  undef STRTOL_ULONG_MAX
170#  define STRTOL_ULONG_MAX maxquad
171# endif
172#else
173# define LONG long
174# define STRTOL_LONG_MIN LONG_MIN
175# define STRTOL_LONG_MAX LONG_MAX
176# define STRTOL_ULONG_MAX ULONG_MAX
177#endif
178
179
180/* We use this code also for the extended locale handling where the
181   function gets as an additional argument the locale which has to be
182   used.  To access the values we have to redefine the _NL_CURRENT
183   macro.  */
184#ifdef USE_IN_EXTENDED_LOCALE_MODEL
185# undef _NL_CURRENT
186# define _NL_CURRENT(category, item) \
187  (current->values[_NL_ITEM_INDEX (item)].string)
188# define LOCALE_PARAM , loc
189# define LOCALE_PARAM_PROTO , __locale_t loc
190#else
191# define LOCALE_PARAM
192# define LOCALE_PARAM_PROTO
193#endif
194
195#if defined _LIBC || defined HAVE_WCHAR_H
196# include <wchar.h>
197#endif
198
199#ifdef USE_WIDE_CHAR
200# include <wctype.h>
201# define L_(Ch) L##Ch
202# define UCHAR_TYPE wint_t
203# define STRING_TYPE wchar_t
204# ifdef USE_IN_EXTENDED_LOCALE_MODEL
205#  define ISSPACE(Ch) __iswspace_l ((Ch), loc)
206#  define ISALPHA(Ch) __iswalpha_l ((Ch), loc)
207#  define TOUPPER(Ch) __towupper_l ((Ch), loc)
208# else
209#  define ISSPACE(Ch) iswspace (Ch)
210#  define ISALPHA(Ch) iswalpha (Ch)
211#  define TOUPPER(Ch) towupper (Ch)
212# endif
213#else
214# if defined STDC_HEADERS || (!defined isascii && !defined HAVE_ISASCII)
215#  define IN_CTYPE_DOMAIN(c) 1
216# else
217#  define IN_CTYPE_DOMAIN(c) isascii(c)
218# endif
219# define L_(Ch) Ch
220# define UCHAR_TYPE unsigned char
221# define STRING_TYPE char
222# ifdef USE_IN_EXTENDED_LOCALE_MODEL
223#  define ISSPACE(Ch) __isspace_l ((Ch), loc)
224#  define ISALPHA(Ch) __isalpha_l ((Ch), loc)
225#  define TOUPPER(Ch) __toupper_l ((Ch), loc)
226# else
227#  define ISSPACE(Ch) (IN_CTYPE_DOMAIN (Ch) && isspace (Ch))
228#  define ISALPHA(Ch) (IN_CTYPE_DOMAIN (Ch) && isalpha (Ch))
229#  define TOUPPER(Ch) (IN_CTYPE_DOMAIN (Ch) ? toupper (Ch) : (Ch))
230# endif
231#endif
232
233#define INTERNAL(X) INTERNAL1(X)
234#define INTERNAL1(X) __##X##_internal
235#define WEAKNAME(X) WEAKNAME1(X)
236
237#ifdef USE_NUMBER_GROUPING
238/* This file defines a function to check for correct grouping.  */
239# include "grouping.h"
240#endif
241
242
243
244/* Convert NPTR to an `unsigned long int' or `long int' in base BASE.
245   If BASE is 0 the base is determined by the presence of a leading
246   zero, indicating octal or a leading "0x" or "0X", indicating hexadecimal.
247   If BASE is < 2 or > 36, it is reset to 10.
248   If ENDPTR is not NULL, a pointer to the character after the last
249   one converted is stored in *ENDPTR.  */
250
251INT
252INTERNAL (strtol) (const STRING_TYPE *nptr, STRING_TYPE **endptr,
253		   int base, int group LOCALE_PARAM_PROTO)
254{
255  int negative;
256  register unsigned LONG int cutoff;
257  register unsigned int cutlim;
258  register unsigned LONG int i;
259  register const STRING_TYPE *s;
260  register UCHAR_TYPE c;
261  const STRING_TYPE *save, *end;
262  int overflow;
263
264#ifdef USE_NUMBER_GROUPING
265# ifdef USE_IN_EXTENDED_LOCALE_MODEL
266  struct locale_data *current = loc->__locales[LC_NUMERIC];
267# endif
268  /* The thousands character of the current locale.  */
269  wchar_t thousands = L'\0';
270  /* The numeric grouping specification of the current locale,
271     in the format described in <locale.h>.  */
272  const char *grouping;
273
274  if (group)
275    {
276      grouping = _NL_CURRENT (LC_NUMERIC, GROUPING);
277      if (*grouping <= 0 || *grouping == CHAR_MAX)
278	grouping = NULL;
279      else
280	{
281	  /* Figure out the thousands separator character.  */
282# if defined _LIBC || defined _HAVE_BTOWC
283	  thousands = __btowc (*_NL_CURRENT (LC_NUMERIC, THOUSANDS_SEP));
284	  if (thousands == WEOF)
285	    thousands = L'\0';
286# endif
287	  if (thousands == L'\0')
288	    grouping = NULL;
289	}
290    }
291  else
292    grouping = NULL;
293#endif
294
295  if (base < 0 || base == 1 || base > 36)
296    {
297      __set_errno (EINVAL);
298      return 0;
299    }
300
301  save = s = nptr;
302
303  /* Skip white space.  */
304  while (ISSPACE (*s))
305    ++s;
306  if (*s == L_('\0'))
307    goto noconv;
308
309  /* Check for a sign.  */
310  if (*s == L_('-'))
311    {
312      negative = 1;
313      ++s;
314    }
315  else if (*s == L_('+'))
316    {
317      negative = 0;
318      ++s;
319    }
320  else
321    negative = 0;
322
323  /* Recognize number prefix and if BASE is zero, figure it out ourselves.  */
324  if (*s == L_('0'))
325    {
326      if ((base == 0 || base == 16) && TOUPPER (s[1]) == L_('X'))
327	{
328	  s += 2;
329	  base = 16;
330	}
331      else if (base == 0)
332	base = 8;
333    }
334  else if (base == 0)
335    base = 10;
336
337  /* Save the pointer so we can check later if anything happened.  */
338  save = s;
339
340#ifdef USE_NUMBER_GROUPING
341  if (group)
342    {
343      /* Find the end of the digit string and check its grouping.  */
344      end = s;
345      for (c = *end; c != L_('\0'); c = *++end)
346	if ((wchar_t) c != thousands
347	    && ((wchar_t) c < L_('0') || (wchar_t) c > L_('9'))
348	    && (!ISALPHA (c) || (int) (TOUPPER (c) - L_('A') + 10) >= base))
349	  break;
350      if (*s == thousands)
351	end = s;
352      else
353	end = correctly_grouped_prefix (s, end, thousands, grouping);
354    }
355  else
356#endif
357    end = NULL;
358
359  cutoff = STRTOL_ULONG_MAX / (unsigned LONG int) base;
360  cutlim = STRTOL_ULONG_MAX % (unsigned LONG int) base;
361
362  overflow = 0;
363  i = 0;
364  for (c = *s; c != L_('\0'); c = *++s)
365    {
366      if (s == end)
367	break;
368      if (c >= L_('0') && c <= L_('9'))
369	c -= L_('0');
370      else if (ISALPHA (c))
371	c = TOUPPER (c) - L_('A') + 10;
372      else
373	break;
374      if ((int) c >= base)
375	break;
376      /* Check for overflow.  */
377      if (i > cutoff || (i == cutoff && c > cutlim))
378	overflow = 1;
379      else
380	{
381	  i *= (unsigned LONG int) base;
382	  i += c;
383	}
384    }
385
386  /* Check if anything actually happened.  */
387  if (s == save)
388    goto noconv;
389
390  /* Store in ENDPTR the address of one character
391     past the last character we converted.  */
392  if (endptr != NULL)
393    *endptr = (STRING_TYPE *) s;
394
395#if !UNSIGNED
396  /* Check for a value that is within the range of
397     `unsigned LONG int', but outside the range of `LONG int'.  */
398  if (overflow == 0
399      && i > (negative
400	      ? -((unsigned LONG int) (STRTOL_LONG_MIN + 1)) + 1
401	      : (unsigned LONG int) STRTOL_LONG_MAX))
402    overflow = 1;
403#endif
404
405  if (overflow)
406    {
407      __set_errno (ERANGE);
408#if UNSIGNED
409      return STRTOL_ULONG_MAX;
410#else
411      return negative ? STRTOL_LONG_MIN : STRTOL_LONG_MAX;
412#endif
413    }
414
415  /* Return the result of the appropriate sign.  */
416  return negative ? -i : i;
417
418noconv:
419  /* We must handle a special case here: the base is 0 or 16 and the
420     first two characters are '0' and 'x', but the rest are no
421     hexadecimal digits.  This is no error case.  We return 0 and
422     ENDPTR points to the `x`.  */
423  if (endptr != NULL)
424    {
425      if (save - nptr >= 2 && TOUPPER (save[-1]) == L_('X')
426	  && save[-2] == L_('0'))
427	*endptr = (STRING_TYPE *) &save[-1];
428      else
429	/*  There was no number to convert.  */
430	*endptr = (STRING_TYPE *) nptr;
431    }
432
433  return 0L;
434}
435
436/* External user entry point.  */
437
438
439INT
440#ifdef weak_function
441weak_function
442#endif
443strtol (const STRING_TYPE *nptr, STRING_TYPE **endptr,
444	int base LOCALE_PARAM_PROTO)
445{
446  return INTERNAL (strtol) (nptr, endptr, base, 0 LOCALE_PARAM);
447}
448