xref: /Universal-ctags/gnulib/mbrtowc.c (revision 820c1a8d46849a90376d8eb15b319ac05439f656)
1*820c1a8dSHiroo HAYASHI /* Convert multibyte character to wide character.
2*820c1a8dSHiroo HAYASHI    Copyright (C) 1999-2002, 2005-2021 Free Software Foundation, Inc.
3*820c1a8dSHiroo HAYASHI    Written by Bruno Haible <bruno@clisp.org>, 2008.
4*820c1a8dSHiroo HAYASHI 
5*820c1a8dSHiroo HAYASHI    This file is free software: you can redistribute it and/or modify
6*820c1a8dSHiroo HAYASHI    it under the terms of the GNU Lesser General Public License as
7*820c1a8dSHiroo HAYASHI    published by the Free Software Foundation; either version 2.1 of the
8*820c1a8dSHiroo HAYASHI    License, or (at your option) any later version.
9*820c1a8dSHiroo HAYASHI 
10*820c1a8dSHiroo HAYASHI    This file is distributed in the hope that it will be useful,
11*820c1a8dSHiroo HAYASHI    but WITHOUT ANY WARRANTY; without even the implied warranty of
12*820c1a8dSHiroo HAYASHI    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
13*820c1a8dSHiroo HAYASHI    GNU Lesser General Public License for more details.
14*820c1a8dSHiroo HAYASHI 
15*820c1a8dSHiroo HAYASHI    You should have received a copy of the GNU Lesser General Public License
16*820c1a8dSHiroo HAYASHI    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
17*820c1a8dSHiroo HAYASHI 
18*820c1a8dSHiroo HAYASHI #include <config.h>
19*820c1a8dSHiroo HAYASHI 
20*820c1a8dSHiroo HAYASHI /* Specification.  */
21*820c1a8dSHiroo HAYASHI #include <wchar.h>
22*820c1a8dSHiroo HAYASHI 
23*820c1a8dSHiroo HAYASHI #if GNULIB_defined_mbstate_t
24*820c1a8dSHiroo HAYASHI /* Implement mbrtowc() on top of mbtowc() for the non-UTF-8 locales
25*820c1a8dSHiroo HAYASHI    and directly for the UTF-8 locales.  */
26*820c1a8dSHiroo HAYASHI 
27*820c1a8dSHiroo HAYASHI # include <errno.h>
28*820c1a8dSHiroo HAYASHI # include <stdint.h>
29*820c1a8dSHiroo HAYASHI # include <stdlib.h>
30*820c1a8dSHiroo HAYASHI 
31*820c1a8dSHiroo HAYASHI # if defined _WIN32 && !defined __CYGWIN__
32*820c1a8dSHiroo HAYASHI 
33*820c1a8dSHiroo HAYASHI #  define WIN32_LEAN_AND_MEAN  /* avoid including junk */
34*820c1a8dSHiroo HAYASHI #  include <windows.h>
35*820c1a8dSHiroo HAYASHI 
36*820c1a8dSHiroo HAYASHI # elif HAVE_PTHREAD_API
37*820c1a8dSHiroo HAYASHI 
38*820c1a8dSHiroo HAYASHI #  include <pthread.h>
39*820c1a8dSHiroo HAYASHI #  if HAVE_THREADS_H && HAVE_WEAK_SYMBOLS
40*820c1a8dSHiroo HAYASHI #   include <threads.h>
41*820c1a8dSHiroo HAYASHI #   pragma weak thrd_exit
42*820c1a8dSHiroo HAYASHI #   define c11_threads_in_use() (thrd_exit != NULL)
43*820c1a8dSHiroo HAYASHI #  else
44*820c1a8dSHiroo HAYASHI #   define c11_threads_in_use() 0
45*820c1a8dSHiroo HAYASHI #  endif
46*820c1a8dSHiroo HAYASHI 
47*820c1a8dSHiroo HAYASHI # elif HAVE_THREADS_H
48*820c1a8dSHiroo HAYASHI 
49*820c1a8dSHiroo HAYASHI #  include <threads.h>
50*820c1a8dSHiroo HAYASHI 
51*820c1a8dSHiroo HAYASHI # endif
52*820c1a8dSHiroo HAYASHI 
53*820c1a8dSHiroo HAYASHI # include "attribute.h"
54*820c1a8dSHiroo HAYASHI # include "verify.h"
55*820c1a8dSHiroo HAYASHI # include "lc-charset-dispatch.h"
56*820c1a8dSHiroo HAYASHI # include "mbtowc-lock.h"
57*820c1a8dSHiroo HAYASHI 
58*820c1a8dSHiroo HAYASHI verify (sizeof (mbstate_t) >= 4);
59*820c1a8dSHiroo HAYASHI static char internal_state[4];
60*820c1a8dSHiroo HAYASHI 
61*820c1a8dSHiroo HAYASHI size_t
mbrtowc(wchar_t * pwc,const char * s,size_t n,mbstate_t * ps)62*820c1a8dSHiroo HAYASHI mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
63*820c1a8dSHiroo HAYASHI {
64*820c1a8dSHiroo HAYASHI # define FITS_IN_CHAR_TYPE(wc)  ((wc) <= WCHAR_MAX)
65*820c1a8dSHiroo HAYASHI # include "mbrtowc-impl.h"
66*820c1a8dSHiroo HAYASHI }
67*820c1a8dSHiroo HAYASHI 
68*820c1a8dSHiroo HAYASHI #else
69*820c1a8dSHiroo HAYASHI /* Override the system's mbrtowc() function.  */
70*820c1a8dSHiroo HAYASHI 
71*820c1a8dSHiroo HAYASHI # if MBRTOWC_IN_C_LOCALE_MAYBE_EILSEQ
72*820c1a8dSHiroo HAYASHI #  include "hard-locale.h"
73*820c1a8dSHiroo HAYASHI #  include <locale.h>
74*820c1a8dSHiroo HAYASHI # endif
75*820c1a8dSHiroo HAYASHI 
76*820c1a8dSHiroo HAYASHI # undef mbrtowc
77*820c1a8dSHiroo HAYASHI 
78*820c1a8dSHiroo HAYASHI size_t
rpl_mbrtowc(wchar_t * pwc,const char * s,size_t n,mbstate_t * ps)79*820c1a8dSHiroo HAYASHI rpl_mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
80*820c1a8dSHiroo HAYASHI {
81*820c1a8dSHiroo HAYASHI   size_t ret;
82*820c1a8dSHiroo HAYASHI   wchar_t wc;
83*820c1a8dSHiroo HAYASHI 
84*820c1a8dSHiroo HAYASHI # if MBRTOWC_NULL_ARG2_BUG || MBRTOWC_RETVAL_BUG || MBRTOWC_EMPTY_INPUT_BUG
85*820c1a8dSHiroo HAYASHI   if (s == NULL)
86*820c1a8dSHiroo HAYASHI     {
87*820c1a8dSHiroo HAYASHI       pwc = NULL;
88*820c1a8dSHiroo HAYASHI       s = "";
89*820c1a8dSHiroo HAYASHI       n = 1;
90*820c1a8dSHiroo HAYASHI     }
91*820c1a8dSHiroo HAYASHI # endif
92*820c1a8dSHiroo HAYASHI 
93*820c1a8dSHiroo HAYASHI # if MBRTOWC_EMPTY_INPUT_BUG
94*820c1a8dSHiroo HAYASHI   if (n == 0)
95*820c1a8dSHiroo HAYASHI     return (size_t) -2;
96*820c1a8dSHiroo HAYASHI # endif
97*820c1a8dSHiroo HAYASHI 
98*820c1a8dSHiroo HAYASHI   if (! pwc)
99*820c1a8dSHiroo HAYASHI     pwc = &wc;
100*820c1a8dSHiroo HAYASHI 
101*820c1a8dSHiroo HAYASHI # if MBRTOWC_RETVAL_BUG
102*820c1a8dSHiroo HAYASHI   {
103*820c1a8dSHiroo HAYASHI     static mbstate_t internal_state;
104*820c1a8dSHiroo HAYASHI 
105*820c1a8dSHiroo HAYASHI     /* Override mbrtowc's internal state.  We cannot call mbsinit() on the
106*820c1a8dSHiroo HAYASHI        hidden internal state, but we can call it on our variable.  */
107*820c1a8dSHiroo HAYASHI     if (ps == NULL)
108*820c1a8dSHiroo HAYASHI       ps = &internal_state;
109*820c1a8dSHiroo HAYASHI 
110*820c1a8dSHiroo HAYASHI     if (!mbsinit (ps))
111*820c1a8dSHiroo HAYASHI       {
112*820c1a8dSHiroo HAYASHI         /* Parse the rest of the multibyte character byte for byte.  */
113*820c1a8dSHiroo HAYASHI         size_t count = 0;
114*820c1a8dSHiroo HAYASHI         for (; n > 0; s++, n--)
115*820c1a8dSHiroo HAYASHI           {
116*820c1a8dSHiroo HAYASHI             ret = mbrtowc (&wc, s, 1, ps);
117*820c1a8dSHiroo HAYASHI 
118*820c1a8dSHiroo HAYASHI             if (ret == (size_t)(-1))
119*820c1a8dSHiroo HAYASHI               return (size_t)(-1);
120*820c1a8dSHiroo HAYASHI             count++;
121*820c1a8dSHiroo HAYASHI             if (ret != (size_t)(-2))
122*820c1a8dSHiroo HAYASHI               {
123*820c1a8dSHiroo HAYASHI                 /* The multibyte character has been completed.  */
124*820c1a8dSHiroo HAYASHI                 *pwc = wc;
125*820c1a8dSHiroo HAYASHI                 return (wc == 0 ? 0 : count);
126*820c1a8dSHiroo HAYASHI               }
127*820c1a8dSHiroo HAYASHI           }
128*820c1a8dSHiroo HAYASHI         return (size_t)(-2);
129*820c1a8dSHiroo HAYASHI       }
130*820c1a8dSHiroo HAYASHI   }
131*820c1a8dSHiroo HAYASHI # endif
132*820c1a8dSHiroo HAYASHI 
133*820c1a8dSHiroo HAYASHI # if MBRTOWC_STORES_INCOMPLETE_BUG
134*820c1a8dSHiroo HAYASHI   ret = mbrtowc (&wc, s, n, ps);
135*820c1a8dSHiroo HAYASHI   if (ret < (size_t) -2 && pwc != NULL)
136*820c1a8dSHiroo HAYASHI     *pwc = wc;
137*820c1a8dSHiroo HAYASHI # else
138*820c1a8dSHiroo HAYASHI   ret = mbrtowc (pwc, s, n, ps);
139*820c1a8dSHiroo HAYASHI # endif
140*820c1a8dSHiroo HAYASHI 
141*820c1a8dSHiroo HAYASHI # if MBRTOWC_NUL_RETVAL_BUG
142*820c1a8dSHiroo HAYASHI   if (ret < (size_t) -2 && !*pwc)
143*820c1a8dSHiroo HAYASHI     return 0;
144*820c1a8dSHiroo HAYASHI # endif
145*820c1a8dSHiroo HAYASHI 
146*820c1a8dSHiroo HAYASHI # if MBRTOWC_IN_C_LOCALE_MAYBE_EILSEQ
147*820c1a8dSHiroo HAYASHI   if ((size_t) -2 <= ret && n != 0 && ! hard_locale (LC_CTYPE))
148*820c1a8dSHiroo HAYASHI     {
149*820c1a8dSHiroo HAYASHI       unsigned char uc = *s;
150*820c1a8dSHiroo HAYASHI       *pwc = uc;
151*820c1a8dSHiroo HAYASHI       return 1;
152*820c1a8dSHiroo HAYASHI     }
153*820c1a8dSHiroo HAYASHI # endif
154*820c1a8dSHiroo HAYASHI 
155*820c1a8dSHiroo HAYASHI   return ret;
156*820c1a8dSHiroo HAYASHI }
157*820c1a8dSHiroo HAYASHI 
158*820c1a8dSHiroo HAYASHI #endif
159