1 /*
2 * Copyright (C) 1999-2001, 2005 Free Software Foundation, Inc.
3 * This file is part of the GNU LIBICONV Library.
4 *
5 * The GNU LIBICONV Library is free software; you can redistribute it
6 * and/or modify it under the terms of the GNU Library General Public
7 * License as published by the Free Software Foundation; either version 2
8 * of the License, or (at your option) any later version.
9 *
10 * The GNU LIBICONV Library is distributed in the hope that it will be
11 * useful, but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * Library General Public License for more details.
14 *
15 * You should have received a copy of the GNU Library General Public
16 * License along with the GNU LIBICONV Library; see the file COPYING.LIB.
17 * If not, write to the Free Software Foundation, Inc., 51 Franklin Street,
18 * Fifth Floor, Boston, MA 02110-1301, USA.
19 */
20
21 /*
22 * GB18030 two-byte extension
23 */
24
25 static const unsigned short gb18030ext_2uni_pagea9[13] = {
26 /* 0xa9 */
27 0x303e, 0x2ff0, 0x2ff1, 0x2ff2, 0x2ff3, 0x2ff4, 0x2ff5, 0x2ff6,
28 0x2ff7, 0x2ff8, 0x2ff9, 0x2ffa, 0x2ffb,
29 };
30 static const unsigned short gb18030ext_2uni_pagefe[96] = {
31 /* 0xfe */
32 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
33 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
34 0x2e81, 0xe816, 0xe817, 0xe818, 0x2e84, 0x3473, 0x3447, 0x2e88,
35 0x2e8b, 0xe81e, 0x359e, 0x361a, 0x360e, 0x2e8c, 0x2e97, 0x396e,
36 0x3918, 0xe826, 0x39cf, 0x39df, 0x3a73, 0x39d0, 0xe82b, 0xe82c,
37 0x3b4e, 0x3c6e, 0x3ce0, 0x2ea7, 0xe831, 0xe832, 0x2eaa, 0x4056,
38 0x415f, 0x2eae, 0x4337, 0x2eb3, 0x2eb6, 0x2eb7, 0xe83b, 0x43b1,
39 0x43ac, 0x2ebb, 0x43dd, 0x44d6, 0x4661, 0x464c, 0xe843, 0x4723,
40 0x4729, 0x477c, 0x478d, 0x2eca, 0x4947, 0x497a, 0x497d, 0x4982,
41 0x4983, 0x4985, 0x4986, 0x499f, 0x499b, 0x49b7, 0x49b6, 0xe854,
42 0xe855, 0x4ca3, 0x4c9f, 0x4ca0, 0x4ca1, 0x4c77, 0x4ca2, 0x4d13,
43 0x4d14, 0x4d15, 0x4d16, 0x4d17, 0x4d18, 0x4d19, 0x4dae, 0xe864,
44 };
45
46 static int
gb18030ext_mbtowc(conv_t conv,ucs4_t * pwc,const unsigned char * s,int n)47 gb18030ext_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, int n)
48 {
49 unsigned char c1 = s[0];
50 if ((c1 == 0xa2) || (c1 >= 0xa4 && c1 <= 0xa9) || (c1 == 0xd7) || (c1 == 0xfe)) {
51 if (n >= 2) {
52 unsigned char c2 = s[1];
53 if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0x80 && c2 < 0xff)) {
54 unsigned int i = 190 * (c1 - 0x81) + (c2 - (c2 >= 0x80 ? 0x41 : 0x40));
55 unsigned short wc = 0xfffd;
56 switch (c1) {
57 case 0xa2:
58 if (i >= 6376 && i <= 6381) /* 0xA2AB..0xA2B0 */
59 wc = 0xe766 + (i - 6376);
60 else if (i == 6432) /* 0xA2E3 */
61 wc = 0x20ac;
62 else if (i == 6433) /* 0xA2E4 */
63 wc = 0xe76d;
64 else if (i >= 6444 && i <= 6445) /* 0xA2EF..0xA2F0 */
65 wc = 0xe76e + (i - 6444);
66 else if (i >= 6458 && i <= 6459) /* 0xA2FD..0xA2FE */
67 wc = 0xe770 + (i - 6458);
68 break;
69 case 0xa4:
70 if (i >= 6829 && i <= 6839) /* 0xA4F4..0xA4FE */
71 wc = 0xe772 + (i - 6829);
72 break;
73 case 0xa5:
74 if (i >= 7022 && i <= 7029) /* 0xA5F7..0xA5FE */
75 wc = 0xe77d + (i - 7022);
76 break;
77 case 0xa6:
78 if (i >= 7150 && i <= 7157) /* 0xA6B9..0xA6C0 */
79 wc = 0xe785 + (i - 7150);
80 else if (i >= 7182 && i <= 7190) /* 0xA6D9..0xA6DF */
81 wc = 0xe78d + (i - 7182);
82 else if (i >= 7201 && i <= 7202) /* 0xA6EC..0xA6ED */
83 wc = 0xe794 + (i - 7201);
84 else if (i == 7208) /* 0xA6F3 */
85 wc = 0xe796;
86 else if (i >= 7211 && i <= 7219) /* 0xA6F6..0xA6FE */
87 wc = 0xe797 + (i - 7211);
88 break;
89 case 0xa7:
90 if (i >= 7349 && i <= 7363) /* 0xA7C2..0xA7D0 */
91 wc = 0xe7a0 + (i - 7349);
92 else if (i >= 7397 && i <= 7409) /* 0xA7F2..0xA7FE */
93 wc = 0xe7af + (i - 7397);
94 break;
95 case 0xa8:
96 if (i >= 7495 && i <= 7505) /* 0xA896..0xA8A0 */
97 wc = 0xe7bc + (i - 7495);
98 else if (i == 7533) /* 0xA8BC */
99 wc = 0xe7c7;
100 else if (i == 7536) /* 0xA8BF */
101 wc = 0x01f9;
102 else if (i >= 7538 && i <= 7541) /* 0xA8C1..0xA8C4 */
103 wc = 0xe7c9 + (i - 7538);
104 else if (i >= 7579 && i <= 7599) /* 0xA8EA..0xA8FE */
105 wc = 0xe7cd + (i - 7579);
106 break;
107 case 0xa9:
108 if (i == 7624) /* 0xA958 */
109 wc = 0xe7e2;
110 else if (i == 7627) /* 0xA95B */
111 wc = 0xe7e3;
112 else if (i >= 7629 && i <= 7631) /* 0xA95D..0xA95F */
113 wc = 0xe7e4 + (i - 7629);
114 else if (i >= 7672 && i < 7685) /* 0xA989..0xA995 */
115 wc = gb18030ext_2uni_pagea9[i-7672];
116 else if (i >= 7686 && i <= 7698) /* 0xA997..0xA9A3 */
117 wc = 0xe7f4 + (i - 7686);
118 else if (i >= 7775 && i <= 7789) /* 0xA9F0..0xA9FE */
119 wc = 0xe801 + (i - 7775);
120 break;
121 case 0xd7:
122 if (i >= 16525 && i <= 16529) /* 0xD7FA..0xD7FE */
123 wc = 0xe810 + (i - 16525);
124 break;
125 case 0xfe:
126 if (i < 23846)
127 wc = gb18030ext_2uni_pagefe[i-23750];
128 break;
129 default:
130 break;
131 }
132 if (wc != 0xfffd) {
133 *pwc = (ucs4_t) wc;
134 return 2;
135 }
136 }
137 return RET_ILSEQ;
138 }
139 return RET_TOOFEW(0);
140 }
141 return RET_ILSEQ;
142 }
143
144 static const unsigned short gb18030ext_page2e[80] = {
145 0x0000, 0xfe50, 0x0000, 0x0000, 0xfe54, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
146 0xfe57, 0x0000, 0x0000, 0xfe58, 0xfe5d, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
147 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5e, /*0x90-0x97*/
148 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x98-0x9f*/
149 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe6b, /*0xa0-0xa7*/
150 0x0000, 0x0000, 0xfe6e, 0x0000, 0x0000, 0x0000, 0xfe71, 0x0000, /*0xa8-0xaf*/
151 0x0000, 0x0000, 0x0000, 0xfe73, 0x0000, 0x0000, 0xfe74, 0xfe75, /*0xb0-0xb7*/
152 0x0000, 0x0000, 0x0000, 0xfe79, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
153 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
154 0x0000, 0x0000, 0xfe84, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
155 };
156 static const unsigned short gb18030ext_page2f[16] = {
157 0xa98a, 0xa98b, 0xa98c, 0xa98d, 0xa98e, 0xa98f, 0xa990, 0xa991, /*0xf0-0xf7*/
158 0xa992, 0xa993, 0xa994, 0xa995, 0x0000, 0x0000, 0x0000, 0x0000, /*0xf8-0xff*/
159 };
160 static const unsigned short gb18030ext_page34[56] = {
161 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe56, /*0x40-0x47*/
162 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
163 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
164 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
165 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
166 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
167 0x0000, 0x0000, 0x0000, 0xfe55, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
168 };
169 static const unsigned short gb18030ext_page36[24] = {
170 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5c, 0x0000, /*0x08-0x0f*/
171 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x10-0x17*/
172 0x0000, 0x0000, 0xfe5b, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
173 };
174 static const unsigned short gb18030ext_page39[24] = {
175 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe62, /*0xc8-0xcf*/
176 0xfe65, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
177 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe63, /*0xd8-0xdf*/
178 };
179 static const unsigned short gb18030ext_page43[56] = {
180 0x0000, 0x0000, 0x0000, 0x0000, 0xfe78, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
181 0x0000, 0xfe77, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb0-0xb7*/
182 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
183 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
184 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
185 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
186 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7a, 0x0000, 0x0000, /*0xd8-0xdf*/
187 };
188 static const unsigned short gb18030ext_page46[32] = {
189 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7d, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
190 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
191 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
192 0x0000, 0xfe7c, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
193 };
194 static const unsigned short gb18030ext_page47_1[16] = {
195 0x0000, 0x0000, 0x0000, 0xfe80, 0x0000, 0x0000, 0x0000, 0x0000, /*0x20-0x27*/
196 0x0000, 0xfe81, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x28-0x2f*/
197 };
198 static const unsigned short gb18030ext_page47_2[24] = {
199 0x0000, 0x0000, 0x0000, 0x0000, 0xfe82, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
200 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
201 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe83, 0x0000, 0x0000, /*0x88-0x8f*/
202 };
203 static const unsigned short gb18030ext_page49[120] = {
204 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe85, /*0x40-0x47*/
205 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
206 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
207 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
208 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
209 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
210 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
211 0x0000, 0x0000, 0xfe86, 0x0000, 0x0000, 0xfe87, 0x0000, 0x0000, /*0x78-0x7f*/
212 0x0000, 0x0000, 0xfe88, 0xfe89, 0x0000, 0xfe8a, 0xfe8b, 0x0000, /*0x80-0x87*/
213 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
214 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
215 0x0000, 0x0000, 0x0000, 0xfe8d, 0x0000, 0x0000, 0x0000, 0xfe8c, /*0x98-0x9f*/
216 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
217 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
218 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe8f, 0xfe8e, /*0xb0-0xb7*/
219 };
220 static const unsigned short gb18030ext_page4c[56] = {
221 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe96, /*0x70-0x77*/
222 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
223 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
224 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
225 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
226 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe93, /*0x98-0x9f*/
227 0xfe94, 0xfe95, 0xfe97, 0xfe92, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
228 };
229 static const unsigned short gb18030ext_page4d[16] = {
230 0x0000, 0x0000, 0x0000, 0xfe98, 0xfe99, 0xfe9a, 0xfe9b, 0xfe9c, /*0x10-0x17*/
231 0xfe9d, 0xfe9e, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
232 };
233
234 static int
gb18030ext_wctomb(conv_t conv,unsigned char * r,ucs4_t wc,int n)235 gb18030ext_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, int n)
236 {
237 if (n >= 2) {
238 unsigned short c = 0;
239 if (wc == 0x01f9)
240 c = 0xa8bf;
241 else if (wc == 0x20ac)
242 c = 0xa2e3;
243 else if (wc >= 0x2e80 && wc < 0x2ed0)
244 c = gb18030ext_page2e[wc-0x2e80];
245 else if (wc >= 0x2ff0 && wc < 0x3000)
246 c = gb18030ext_page2f[wc-0x2ff0];
247 else if (wc == 0x303e)
248 c = 0xa989;
249 else if (wc >= 0x3440 && wc < 0x3478)
250 c = gb18030ext_page34[wc-0x3440];
251 else if (wc == 0x359e)
252 c = 0xfe5a;
253 else if (wc >= 0x3608 && wc < 0x3620)
254 c = gb18030ext_page36[wc-0x3608];
255 else if (wc == 0x3918)
256 c = 0xfe60;
257 else if (wc == 0x396e)
258 c = 0xfe5f;
259 else if (wc >= 0x39c8 && wc < 0x39e0)
260 c = gb18030ext_page39[wc-0x39c8];
261 else if (wc == 0x3a73)
262 c = 0xfe64;
263 else if (wc == 0x3b4e)
264 c = 0xfe68;
265 else if (wc == 0x3c6e)
266 c = 0xfe69;
267 else if (wc == 0x3ce0)
268 c = 0xfe6a;
269 else if (wc == 0x4056)
270 c = 0xfe6f;
271 else if (wc == 0x415f)
272 c = 0xfe70;
273 else if (wc == 0x4337)
274 c = 0xfe72;
275 else if (wc >= 0x43a8 && wc < 0x43e0)
276 c = gb18030ext_page43[wc-0x43a8];
277 else if (wc == 0x44d6)
278 c = 0xfe7b;
279 else if (wc >= 0x4648 && wc < 0x4668)
280 c = gb18030ext_page46[wc-0x4648];
281 else if (wc >= 0x4720 && wc < 0x4730)
282 c = gb18030ext_page47_1[wc-0x4720];
283 else if (wc >= 0x4778 && wc < 0x4790)
284 c = gb18030ext_page47_2[wc-0x4778];
285 else if (wc >= 0x4940 && wc < 0x49b8)
286 c = gb18030ext_page49[wc-0x4940];
287 else if (wc >= 0x4c70 && wc < 0x4ca8)
288 c = gb18030ext_page4c[wc-0x4c70];
289 else if (wc >= 0x4d10 && wc < 0x4d20)
290 c = gb18030ext_page4d[wc-0x4d10];
291 else if (wc == 0x4dae)
292 c = 0xfe9f;
293 if (c != 0) {
294 r[0] = (c >> 8); r[1] = (c & 0xff);
295 return 2;
296 }
297 return RET_ILUNI;
298 }
299 return RET_TOOSMALL;
300 }
301