1 /* Convert multibyte character to wide character.
2 Copyright (C) 1999-2002, 2005-2011 Free Software Foundation, Inc.
3 Written by Bruno Haible <bruno@clisp.org>, 2008.
5 This program is free software: you can redistribute it and/or modify
6 it under the terms of the GNU General Public License as published by
7 the Free Software Foundation; either version 3 of the License, or
8 (at your option) any later version.
10 This program is distributed in the hope that it will be useful,
11 but WITHOUT ANY WARRANTY; without even the implied warranty of
12 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 GNU General Public License for more details.
15 You should have received a copy of the GNU General Public License
16 along with this program. If not, see <http://www.gnu.org/licenses/>. */
23 #if GNULIB_defined_mbstate_t
24 /* Implement mbrtowc() on top of mbtowc(). */
29 # include "localcharset.h"
34 verify (sizeof (mbstate_t) >= 4);
36 static char internal_state[4];
39 mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
41 char *pstate = (char *)ps;
56 pstate = internal_state;
59 size_t nstate = pstate[0];
95 # if __GLIBC__ || defined __UCLIBC__
96 /* Work around bug <http://sourceware.org/bugzilla/show_bug.cgi?id=9674> */
97 mbtowc (NULL, NULL, 0);
100 int res = mbtowc (pwc, p, m);
104 if (pwc != NULL && ((*pwc == 0) != (res == 0)))
106 if (nstate >= (res > 0 ? res : 1))
113 /* mbtowc does not distinguish between invalid and incomplete multibyte
114 sequences. But mbrtowc needs to make this distinction.
115 There are two possible approaches:
116 - Use iconv() and its return value.
117 - Use built-in knowledge about the possible encodings.
118 Given the low quality of implementation of iconv() on the systems that
119 lack mbrtowc(), we use the second approach.
120 The possible encodings are:
122 - EUC-JP, EUC-KR, GB2312, EUC-TW, BIG5, GB18030, SJIS,
124 Use specialized code for each. */
125 if (m >= 4 || m >= MB_CUR_MAX)
127 /* Here MB_CUR_MAX > 1 and 0 < m < 4. */
129 const char *encoding = locale_charset ();
131 if (STREQ (encoding, "UTF-8", 'U', 'T', 'F', '-', '8', 0, 0, 0, 0))
133 /* Cf. unistr/u8-mblen.c. */
134 unsigned char c = (unsigned char) p[0];
149 unsigned char c2 = (unsigned char) p[1];
151 if ((c2 ^ 0x80) < 0x40
152 && (c >= 0xe1 || c2 >= 0xa0)
153 && (c != 0xed || c2 < 0xa0))
161 else /* m == 2 || m == 3 */
163 unsigned char c2 = (unsigned char) p[1];
165 if ((c2 ^ 0x80) < 0x40
166 && (c >= 0xf1 || c2 >= 0x90)
167 && (c < 0xf4 || (c == 0xf4 && c2 < 0x90)))
173 unsigned char c3 = (unsigned char) p[2];
175 if ((c3 ^ 0x80) < 0x40)
185 /* As a reference for this code, you can use the GNU libiconv
186 implementation. Look for uses of the RET_TOOFEW macro. */
188 if (STREQ (encoding, "EUC-JP", 'E', 'U', 'C', '-', 'J', 'P', 0, 0, 0))
192 unsigned char c = (unsigned char) p[0];
194 if ((c >= 0xa1 && c < 0xff) || c == 0x8e || c == 0x8f)
199 unsigned char c = (unsigned char) p[0];
203 unsigned char c2 = (unsigned char) p[1];
205 if (c2 >= 0xa1 && c2 < 0xff)
211 if (STREQ (encoding, "EUC-KR", 'E', 'U', 'C', '-', 'K', 'R', 0, 0, 0)
212 || STREQ (encoding, "GB2312", 'G', 'B', '2', '3', '1', '2', 0, 0, 0)
213 || STREQ (encoding, "BIG5", 'B', 'I', 'G', '5', 0, 0, 0, 0, 0))
217 unsigned char c = (unsigned char) p[0];
219 if (c >= 0xa1 && c < 0xff)
224 if (STREQ (encoding, "EUC-TW", 'E', 'U', 'C', '-', 'T', 'W', 0, 0, 0))
228 unsigned char c = (unsigned char) p[0];
230 if ((c >= 0xa1 && c < 0xff) || c == 0x8e)
233 else /* m == 2 || m == 3 */
235 unsigned char c = (unsigned char) p[0];
242 if (STREQ (encoding, "GB18030", 'G', 'B', '1', '8', '0', '3', '0', 0, 0))
246 unsigned char c = (unsigned char) p[0];
248 if ((c >= 0x90 && c <= 0xe3) || (c >= 0xf8 && c <= 0xfe))
251 else /* m == 2 || m == 3 */
253 unsigned char c = (unsigned char) p[0];
255 if (c >= 0x90 && c <= 0xe3)
257 unsigned char c2 = (unsigned char) p[1];
259 if (c2 >= 0x30 && c2 <= 0x39)
265 unsigned char c3 = (unsigned char) p[2];
267 if (c3 >= 0x81 && c3 <= 0xfe)
275 if (STREQ (encoding, "SJIS", 'S', 'J', 'I', 'S', 0, 0, 0, 0, 0))
279 unsigned char c = (unsigned char) p[0];
281 if ((c >= 0x81 && c <= 0x9f) || (c >= 0xe0 && c <= 0xea)
282 || (c >= 0xf0 && c <= 0xf9))
288 /* An unknown multibyte encoding. */
295 /* Here 0 <= k < m < 4. */
311 /* The conversion state is undefined, says POSIX. */
318 /* Override the system's mbrtowc() function. */
323 rpl_mbrtowc (wchar_t *pwc, const char *s, size_t n, mbstate_t *ps)
325 # if MBRTOWC_NULL_ARG2_BUG || MBRTOWC_RETVAL_BUG
334 # if MBRTOWC_RETVAL_BUG
336 static mbstate_t internal_state;
338 /* Override mbrtowc's internal state. We cannot call mbsinit() on the
339 hidden internal state, but we can call it on our variable. */
341 ps = &internal_state;
345 /* Parse the rest of the multibyte character byte for byte. */
347 for (; n > 0; s++, n--)
350 size_t ret = mbrtowc (&wc, s, 1, ps);
352 if (ret == (size_t)(-1))
355 if (ret != (size_t)(-2))
357 /* The multibyte character has been completed. */
360 return (wc == 0 ? 0 : count);
368 # if MBRTOWC_NUL_RETVAL_BUG
371 size_t ret = mbrtowc (&wc, s, n, ps);
373 if (ret != (size_t)(-1) && ret != (size_t)(-2))
384 # if MBRTOWC_NULL_ARG1_BUG
391 return mbrtowc (pwc, s, n, ps);