| 1 | /* Convert multibyte character to wide character. |
| 2 | Copyright (C) 1999-2002, 2005-2020 Free Software Foundation, Inc. |
| 3 | |
| 4 | This program is free software: you can redistribute it and/or modify |
| 5 | it under the terms of the GNU General Public License as published by |
| 6 | the Free Software Foundation; either version 3 of the License, or |
| 7 | (at your option) any later version. |
| 8 | |
| 9 | This program is distributed in the hope that it will be useful, |
| 10 | but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 11 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 12 | GNU General Public License for more details. |
| 13 | |
| 14 | You should have received a copy of the GNU General Public License |
| 15 | along with this program. If not, see <https://www.gnu.org/licenses/>. */ |
| 16 | |
| 17 | /* Written by Bruno Haible <bruno@clisp.org>, 2008. */ |
| 18 | |
| 19 | /* This file contains the body of the mbrtowc and mbrtoc32 functions, |
| 20 | when GNULIB_defined_mbstate_t is defined. */ |
| 21 | |
| 22 | char *pstate = (char *)ps; |
| 23 | |
| 24 | if (s == NULL) |
| 25 | { |
| 26 | pwc = NULL; |
| 27 | s = ""; |
| 28 | n = 1; |
| 29 | } |
| 30 | |
| 31 | if (n == 0) |
| 32 | return (size_t)(-2); |
| 33 | |
| 34 | /* Here n > 0. */ |
| 35 | |
| 36 | if (pstate == NULL) |
| 37 | pstate = internal_state; |
| 38 | |
| 39 | { |
| 40 | size_t nstate = pstate[0]; |
| 41 | char buf[4]; |
| 42 | const char *p; |
| 43 | size_t m; |
| 44 | enc_t enc; |
| 45 | int res; |
| 46 | |
| 47 | switch (nstate) |
| 48 | { |
| 49 | case 0: |
| 50 | p = s; |
| 51 | m = n; |
| 52 | break; |
| 53 | case 3: |
| 54 | buf[2] = pstate[3]; |
| 55 | FALLTHROUGH; |
| 56 | case 2: |
| 57 | buf[1] = pstate[2]; |
| 58 | FALLTHROUGH; |
| 59 | case 1: |
| 60 | buf[0] = pstate[1]; |
| 61 | p = buf; |
| 62 | m = nstate; |
| 63 | buf[m++] = s[0]; |
| 64 | if (n >= 2 && m < 4) |
| 65 | { |
| 66 | buf[m++] = s[1]; |
| 67 | if (n >= 3 && m < 4) |
| 68 | buf[m++] = s[2]; |
| 69 | } |
| 70 | break; |
| 71 | default: |
| 72 | errno = EINVAL; |
| 73 | return (size_t)(-1); |
| 74 | } |
| 75 | |
| 76 | /* Here m > 0. */ |
| 77 | |
| 78 | enc = locale_encoding_classification (); |
| 79 | |
| 80 | if (enc == enc_utf8) /* UTF-8 */ |
| 81 | { |
| 82 | /* Achieve |
| 83 | - multi-thread safety and |
| 84 | - the ability to produce wide character values > WCHAR_MAX |
| 85 | by not calling mbtowc() at all. */ |
| 86 | #include "mbrtowc-impl-utf8.h" |
| 87 | } |
| 88 | else |
| 89 | { |
| 90 | /* The hidden internal state of mbtowc would make this function not |
| 91 | multi-thread safe. Achieve multi-thread safety through a lock. */ |
| 92 | wchar_t wc; |
| 93 | res = mbtowc_with_lock (&wc, p, m); |
| 94 | |
| 95 | if (res >= 0) |
| 96 | { |
| 97 | if ((wc == 0) != (res == 0)) |
| 98 | abort (); |
| 99 | if (pwc != NULL) |
| 100 | *pwc = wc; |
| 101 | goto success; |
| 102 | } |
| 103 | |
| 104 | /* mbtowc does not distinguish between invalid and incomplete multibyte |
| 105 | sequences. But mbrtowc needs to make this distinction. |
| 106 | There are two possible approaches: |
| 107 | - Use iconv() and its return value. |
| 108 | - Use built-in knowledge about the possible encodings. |
| 109 | Given the low quality of implementation of iconv() on the systems |
| 110 | that lack mbrtowc(), we use the second approach. |
| 111 | The possible encodings are: |
| 112 | - 8-bit encodings, |
| 113 | - EUC-JP, EUC-KR, GB2312, EUC-TW, BIG5, GB18030, SJIS, |
| 114 | - UTF-8 (already handled above). |
| 115 | Use specialized code for each. */ |
| 116 | if (m >= 4 || m >= MB_CUR_MAX) |
| 117 | goto invalid; |
| 118 | /* Here MB_CUR_MAX > 1 and 0 < m < 4. */ |
| 119 | switch (enc) |
| 120 | { |
| 121 | /* As a reference for this code, you can use the GNU libiconv |
| 122 | implementation. Look for uses of the RET_TOOFEW macro. */ |
| 123 | |
| 124 | case enc_eucjp: /* EUC-JP */ |
| 125 | { |
| 126 | if (m == 1) |
| 127 | { |
| 128 | unsigned char c = (unsigned char) p[0]; |
| 129 | |
| 130 | if ((c >= 0xa1 && c < 0xff) || c == 0x8e || c == 0x8f) |
| 131 | goto incomplete; |
| 132 | } |
| 133 | if (m == 2) |
| 134 | { |
| 135 | unsigned char c = (unsigned char) p[0]; |
| 136 | |
| 137 | if (c == 0x8f) |
| 138 | { |
| 139 | unsigned char c2 = (unsigned char) p[1]; |
| 140 | |
| 141 | if (c2 >= 0xa1 && c2 < 0xff) |
| 142 | goto incomplete; |
| 143 | } |
| 144 | } |
| 145 | goto invalid; |
| 146 | } |
| 147 | |
| 148 | case enc_94: /* EUC-KR, GB2312, BIG5 */ |
| 149 | { |
| 150 | if (m == 1) |
| 151 | { |
| 152 | unsigned char c = (unsigned char) p[0]; |
| 153 | |
| 154 | if (c >= 0xa1 && c < 0xff) |
| 155 | goto incomplete; |
| 156 | } |
| 157 | goto invalid; |
| 158 | } |
| 159 | |
| 160 | case enc_euctw: /* EUC-TW */ |
| 161 | { |
| 162 | if (m == 1) |
| 163 | { |
| 164 | unsigned char c = (unsigned char) p[0]; |
| 165 | |
| 166 | if ((c >= 0xa1 && c < 0xff) || c == 0x8e) |
| 167 | goto incomplete; |
| 168 | } |
| 169 | else /* m == 2 || m == 3 */ |
| 170 | { |
| 171 | unsigned char c = (unsigned char) p[0]; |
| 172 | |
| 173 | if (c == 0x8e) |
| 174 | goto incomplete; |
| 175 | } |
| 176 | goto invalid; |
| 177 | } |
| 178 | |
| 179 | case enc_gb18030: /* GB18030 */ |
| 180 | { |
| 181 | if (m == 1) |
| 182 | { |
| 183 | unsigned char c = (unsigned char) p[0]; |
| 184 | |
| 185 | if ((c >= 0x90 && c <= 0xe3) || (c >= 0xf8 && c <= 0xfe)) |
| 186 | goto incomplete; |
| 187 | } |
| 188 | else /* m == 2 || m == 3 */ |
| 189 | { |
| 190 | unsigned char c = (unsigned char) p[0]; |
| 191 | |
| 192 | if (c >= 0x90 && c <= 0xe3) |
| 193 | { |
| 194 | unsigned char c2 = (unsigned char) p[1]; |
| 195 | |
| 196 | if (c2 >= 0x30 && c2 <= 0x39) |
| 197 | { |
| 198 | if (m == 2) |
| 199 | goto incomplete; |
| 200 | else /* m == 3 */ |
| 201 | { |
| 202 | unsigned char c3 = (unsigned char) p[2]; |
| 203 | |
| 204 | if (c3 >= 0x81 && c3 <= 0xfe) |
| 205 | goto incomplete; |
| 206 | } |
| 207 | } |
| 208 | } |
| 209 | } |
| 210 | goto invalid; |
| 211 | } |
| 212 | |
| 213 | case enc_sjis: /* SJIS */ |
| 214 | { |
| 215 | if (m == 1) |
| 216 | { |
| 217 | unsigned char c = (unsigned char) p[0]; |
| 218 | |
| 219 | if ((c >= 0x81 && c <= 0x9f) || (c >= 0xe0 && c <= 0xea) |
| 220 | || (c >= 0xf0 && c <= 0xf9)) |
| 221 | goto incomplete; |
| 222 | } |
| 223 | goto invalid; |
| 224 | } |
| 225 | |
| 226 | default: |
| 227 | /* An unknown multibyte encoding. */ |
| 228 | goto incomplete; |
| 229 | } |
| 230 | } |
| 231 | |
| 232 | success: |
| 233 | /* res >= 0 is the corrected return value of |
| 234 | mbtowc_with_lock (&wc, p, m). */ |
| 235 | if (nstate >= (res > 0 ? res : 1)) |
| 236 | abort (); |
| 237 | res -= nstate; |
| 238 | pstate[0] = 0; |
| 239 | return res; |
| 240 | |
| 241 | incomplete: |
| 242 | { |
| 243 | size_t k = nstate; |
| 244 | /* Here 0 <= k < m < 4. */ |
| 245 | pstate[++k] = s[0]; |
| 246 | if (k < m) |
| 247 | { |
| 248 | pstate[++k] = s[1]; |
| 249 | if (k < m) |
| 250 | pstate[++k] = s[2]; |
| 251 | } |
| 252 | if (k != m) |
| 253 | abort (); |
| 254 | } |
| 255 | pstate[0] = m; |
| 256 | return (size_t)(-2); |
| 257 | |
| 258 | invalid: |
| 259 | errno = EILSEQ; |
| 260 | /* The conversion state is undefined, says POSIX. */ |
| 261 | return (size_t)(-1); |
| 262 | } |