1: /*
2: * Copyright (C) 1999-2001, 2005, 2011, 2016 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, see <https://www.gnu.org/licenses/>.
18: */
19:
20: /*
21: * GB18030 two-byte extension
22: */
23:
24: static const unsigned short gb18030ext_2uni_pagea9[13] = {
25: /* 0xa9 */
26: 0x303e, 0x2ff0, 0x2ff1, 0x2ff2, 0x2ff3, 0x2ff4, 0x2ff5, 0x2ff6,
27: 0x2ff7, 0x2ff8, 0x2ff9, 0x2ffa, 0x2ffb,
28: };
29: static const unsigned int gb18030ext_2uni_pagefe[96] = {
30: /* 0xfe */
31: 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
32: 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd,
33: 0x2e81, 0x20087, 0x20089, 0x200cc, 0x2e84, 0x3473, 0x3447, 0x2e88,
34: 0x2e8b, 0x9fb4, 0x359e, 0x361a, 0x360e, 0x2e8c, 0x2e97, 0x396e,
35: 0x3918, 0x9fb5, 0x39cf, 0x39df, 0x3a73, 0x39d0, 0x9fb6, 0x9fb7,
36: 0x3b4e, 0x3c6e, 0x3ce0, 0x2ea7, 0x215d7, 0x9fb8, 0x2eaa, 0x4056,
37: 0x415f, 0x2eae, 0x4337, 0x2eb3, 0x2eb6, 0x2eb7, 0x2298f, 0x43b1,
38: 0x43ac, 0x2ebb, 0x43dd, 0x44d6, 0x4661, 0x464c, 0x9fb9, 0x4723,
39: 0x4729, 0x477c, 0x478d, 0x2eca, 0x4947, 0x497a, 0x497d, 0x4982,
40: 0x4983, 0x4985, 0x4986, 0x499f, 0x499b, 0x49b7, 0x49b6, 0x9fba,
41: 0x241fe, 0x4ca3, 0x4c9f, 0x4ca0, 0x4ca1, 0x4c77, 0x4ca2, 0x4d13,
42: 0x4d14, 0x4d15, 0x4d16, 0x4d17, 0x4d18, 0x4d19, 0x4dae, 0x9fbb,
43: };
44:
45: static int
46: gb18030ext_mbtowc (conv_t conv, ucs4_t *pwc, const unsigned char *s, size_t n)
47: {
48: unsigned char c1 = s[0];
49: if ((c1 == 0xa2) || (c1 >= 0xa4 && c1 <= 0xa9) || (c1 == 0xd7) || (c1 == 0xfe)) {
50: if (n >= 2) {
51: unsigned char c2 = s[1];
52: if ((c2 >= 0x40 && c2 < 0x7f) || (c2 >= 0x80 && c2 < 0xff)) {
53: unsigned int i = 190 * (c1 - 0x81) + (c2 - (c2 >= 0x80 ? 0x41 : 0x40));
54: unsigned int wc = 0xfffd;
55: switch (c1) {
56: case 0xa2:
57: if (i >= 6376 && i <= 6381) /* 0xA2AB..0xA2B0 */
58: wc = 0xe766 + (i - 6376);
59: else if (i == 6432) /* 0xA2E3 */
60: wc = 0x20ac;
61: else if (i == 6433) /* 0xA2E4 */
62: wc = 0xe76d;
63: else if (i >= 6444 && i <= 6445) /* 0xA2EF..0xA2F0 */
64: wc = 0xe76e + (i - 6444);
65: else if (i >= 6458 && i <= 6459) /* 0xA2FD..0xA2FE */
66: wc = 0xe770 + (i - 6458);
67: break;
68: case 0xa4:
69: if (i >= 6829 && i <= 6839) /* 0xA4F4..0xA4FE */
70: wc = 0xe772 + (i - 6829);
71: break;
72: case 0xa5:
73: if (i >= 7022 && i <= 7029) /* 0xA5F7..0xA5FE */
74: wc = 0xe77d + (i - 7022);
75: break;
76: case 0xa6:
77: if (i >= 7150 && i <= 7157) /* 0xA6B9..0xA6C0 */
78: wc = 0xe785 + (i - 7150);
79: else if (i >= 7183 && i <= 7184) /* 0xA6DA..0xA6DB */
80: wc = 0xfe12 - (i - 7183);
81: else if (i >= 7182 && i <= 7190) /* 0xA6D9..0xA6DF */
82: wc = 0xfe10 + (i - 7182);
83: else if (i >= 7201 && i <= 7202) /* 0xA6EC..0xA6ED */
84: wc = 0xfe17 + (i - 7201);
85: else if (i == 7208) /* 0xA6F3 */
86: wc = 0xfe19;
87: else if (i >= 7211 && i <= 7219) /* 0xA6F6..0xA6FE */
88: wc = 0xe797 + (i - 7211);
89: break;
90: case 0xa7:
91: if (i >= 7349 && i <= 7363) /* 0xA7C2..0xA7D0 */
92: wc = 0xe7a0 + (i - 7349);
93: else if (i >= 7397 && i <= 7409) /* 0xA7F2..0xA7FE */
94: wc = 0xe7af + (i - 7397);
95: break;
96: case 0xa8:
97: if (i >= 7495 && i <= 7505) /* 0xA896..0xA8A0 */
98: wc = 0xe7bc + (i - 7495);
99: else if (i == 7533) /* 0xA8BC */
100: wc = 0x1e3f;
101: else if (i == 7536) /* 0xA8BF */
102: wc = 0x01f9;
103: else if (i >= 7538 && i <= 7541) /* 0xA8C1..0xA8C4 */
104: wc = 0xe7c9 + (i - 7538);
105: else if (i >= 7579 && i <= 7599) /* 0xA8EA..0xA8FE */
106: wc = 0xe7cd + (i - 7579);
107: break;
108: case 0xa9:
109: if (i == 7624) /* 0xA958 */
110: wc = 0xe7e2;
111: else if (i == 7627) /* 0xA95B */
112: wc = 0xe7e3;
113: else if (i >= 7629 && i <= 7631) /* 0xA95D..0xA95F */
114: wc = 0xe7e4 + (i - 7629);
115: else if (i >= 7672 && i < 7685) /* 0xA989..0xA995 */
116: wc = gb18030ext_2uni_pagea9[i-7672];
117: else if (i >= 7686 && i <= 7698) /* 0xA997..0xA9A3 */
118: wc = 0xe7f4 + (i - 7686);
119: else if (i >= 7775 && i <= 7789) /* 0xA9F0..0xA9FE */
120: wc = 0xe801 + (i - 7775);
121: break;
122: case 0xd7:
123: if (i >= 16525 && i <= 16529) /* 0xD7FA..0xD7FE */
124: wc = 0xe810 + (i - 16525);
125: break;
126: case 0xfe:
127: if (i < 23846)
128: wc = gb18030ext_2uni_pagefe[i-23750];
129: break;
130: default:
131: break;
132: }
133: if (wc != 0xfffd) {
134: *pwc = (ucs4_t) wc;
135: return 2;
136: }
137: }
138: return RET_ILSEQ;
139: }
140: return RET_TOOFEW(0);
141: }
142: return RET_ILSEQ;
143: }
144:
145: static const unsigned short gb18030ext_page2e[80] = {
146: 0x0000, 0xfe50, 0x0000, 0x0000, 0xfe54, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
147: 0xfe57, 0x0000, 0x0000, 0xfe58, 0xfe5d, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
148: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5e, /*0x90-0x97*/
149: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x98-0x9f*/
150: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe6b, /*0xa0-0xa7*/
151: 0x0000, 0x0000, 0xfe6e, 0x0000, 0x0000, 0x0000, 0xfe71, 0x0000, /*0xa8-0xaf*/
152: 0x0000, 0x0000, 0x0000, 0xfe73, 0x0000, 0x0000, 0xfe74, 0xfe75, /*0xb0-0xb7*/
153: 0x0000, 0x0000, 0x0000, 0xfe79, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
154: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
155: 0x0000, 0x0000, 0xfe84, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
156: };
157: static const unsigned short gb18030ext_page2f[16] = {
158: 0xa98a, 0xa98b, 0xa98c, 0xa98d, 0xa98e, 0xa98f, 0xa990, 0xa991, /*0xf0-0xf7*/
159: 0xa992, 0xa993, 0xa994, 0xa995, 0x0000, 0x0000, 0x0000, 0x0000, /*0xf8-0xff*/
160: };
161: static const unsigned short gb18030ext_page34[56] = {
162: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe56, /*0x40-0x47*/
163: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
164: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
165: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
166: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
167: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
168: 0x0000, 0x0000, 0x0000, 0xfe55, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
169: };
170: static const unsigned short gb18030ext_page36[24] = {
171: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe5c, 0x0000, /*0x08-0x0f*/
172: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x10-0x17*/
173: 0x0000, 0x0000, 0xfe5b, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
174: };
175: static const unsigned short gb18030ext_page39[24] = {
176: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe62, /*0xc8-0xcf*/
177: 0xfe65, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
178: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe63, /*0xd8-0xdf*/
179: };
180: static const unsigned short gb18030ext_page43[56] = {
181: 0x0000, 0x0000, 0x0000, 0x0000, 0xfe78, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
182: 0x0000, 0xfe77, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb0-0xb7*/
183: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
184: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc0-0xc7*/
185: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xc8-0xcf*/
186: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xd0-0xd7*/
187: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7a, 0x0000, 0x0000, /*0xd8-0xdf*/
188: };
189: static const unsigned short gb18030ext_page46[32] = {
190: 0x0000, 0x0000, 0x0000, 0x0000, 0xfe7d, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
191: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
192: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
193: 0x0000, 0xfe7c, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
194: };
195: static const unsigned short gb18030ext_page47_1[16] = {
196: 0x0000, 0x0000, 0x0000, 0xfe80, 0x0000, 0x0000, 0x0000, 0x0000, /*0x20-0x27*/
197: 0x0000, 0xfe81, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x28-0x2f*/
198: };
199: static const unsigned short gb18030ext_page47_2[24] = {
200: 0x0000, 0x0000, 0x0000, 0x0000, 0xfe82, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
201: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
202: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe83, 0x0000, 0x0000, /*0x88-0x8f*/
203: };
204: static const unsigned short gb18030ext_page49[120] = {
205: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe85, /*0x40-0x47*/
206: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x48-0x4f*/
207: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x50-0x57*/
208: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x58-0x5f*/
209: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x60-0x67*/
210: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x68-0x6f*/
211: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x70-0x77*/
212: 0x0000, 0x0000, 0xfe86, 0x0000, 0x0000, 0xfe87, 0x0000, 0x0000, /*0x78-0x7f*/
213: 0x0000, 0x0000, 0xfe88, 0xfe89, 0x0000, 0xfe8a, 0xfe8b, 0x0000, /*0x80-0x87*/
214: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
215: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
216: 0x0000, 0x0000, 0x0000, 0xfe8d, 0x0000, 0x0000, 0x0000, 0xfe8c, /*0x98-0x9f*/
217: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
218: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa8-0xaf*/
219: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe8f, 0xfe8e, /*0xb0-0xb7*/
220: };
221: static const unsigned short gb18030ext_page4c[56] = {
222: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe96, /*0x70-0x77*/
223: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x78-0x7f*/
224: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x80-0x87*/
225: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x88-0x8f*/
226: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x90-0x97*/
227: 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0xfe93, /*0x98-0x9f*/
228: 0xfe94, 0xfe95, 0xfe97, 0xfe92, 0x0000, 0x0000, 0x0000, 0x0000, /*0xa0-0xa7*/
229: };
230: static const unsigned short gb18030ext_page4d[16] = {
231: 0x0000, 0x0000, 0x0000, 0xfe98, 0xfe99, 0xfe9a, 0xfe9b, 0xfe9c, /*0x10-0x17*/
232: 0xfe9d, 0xfe9e, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
233: };
234: static const unsigned short gb18030ext_page9f[16] = {
235: 0x0000, 0x0000, 0x0000, 0x0000, 0xfe59, 0xfe61, 0xfe66, 0xfe67, /*0xb0-0xb7*/
236: 0xfe6d, 0xfe7e, 0xfe90, 0xfea0, 0x0000, 0x0000, 0x0000, 0x0000, /*0xb8-0xbf*/
237: };
238: static const unsigned short gb18030ext_pagefe[16] = {
239: 0xa6d9, 0xa6db, 0xa6da, 0xa6dc, 0xa6dd, 0xa6de, 0xa6df, 0xa6ec, /*0x10-0x17*/
240: 0xa6ed, 0xa6f3, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, /*0x18-0x1f*/
241: };
242:
243: static int
244: gb18030ext_wctomb (conv_t conv, unsigned char *r, ucs4_t wc, size_t n)
245: {
246: if (n >= 2) {
247: unsigned short c = 0;
248: if (wc == 0x01f9)
249: c = 0xa8bf;
250: else if (wc == 0x1e3f)
251: c = 0xa8bc;
252: else if (wc == 0x20ac)
253: c = 0xa2e3;
254: else if (wc >= 0x2e80 && wc < 0x2ed0)
255: c = gb18030ext_page2e[wc-0x2e80];
256: else if (wc >= 0x2ff0 && wc < 0x3000)
257: c = gb18030ext_page2f[wc-0x2ff0];
258: else if (wc == 0x303e)
259: c = 0xa989;
260: else if (wc >= 0x3440 && wc < 0x3478)
261: c = gb18030ext_page34[wc-0x3440];
262: else if (wc == 0x359e)
263: c = 0xfe5a;
264: else if (wc >= 0x3608 && wc < 0x3620)
265: c = gb18030ext_page36[wc-0x3608];
266: else if (wc == 0x3918)
267: c = 0xfe60;
268: else if (wc == 0x396e)
269: c = 0xfe5f;
270: else if (wc >= 0x39c8 && wc < 0x39e0)
271: c = gb18030ext_page39[wc-0x39c8];
272: else if (wc == 0x3a73)
273: c = 0xfe64;
274: else if (wc == 0x3b4e)
275: c = 0xfe68;
276: else if (wc == 0x3c6e)
277: c = 0xfe69;
278: else if (wc == 0x3ce0)
279: c = 0xfe6a;
280: else if (wc == 0x4056)
281: c = 0xfe6f;
282: else if (wc == 0x415f)
283: c = 0xfe70;
284: else if (wc == 0x4337)
285: c = 0xfe72;
286: else if (wc >= 0x43a8 && wc < 0x43e0)
287: c = gb18030ext_page43[wc-0x43a8];
288: else if (wc == 0x44d6)
289: c = 0xfe7b;
290: else if (wc >= 0x4648 && wc < 0x4668)
291: c = gb18030ext_page46[wc-0x4648];
292: else if (wc >= 0x4720 && wc < 0x4730)
293: c = gb18030ext_page47_1[wc-0x4720];
294: else if (wc >= 0x4778 && wc < 0x4790)
295: c = gb18030ext_page47_2[wc-0x4778];
296: else if (wc >= 0x4940 && wc < 0x49b8)
297: c = gb18030ext_page49[wc-0x4940];
298: else if (wc >= 0x4c70 && wc < 0x4ca8)
299: c = gb18030ext_page4c[wc-0x4c70];
300: else if (wc >= 0x4d10 && wc < 0x4d20)
301: c = gb18030ext_page4d[wc-0x4d10];
302: else if (wc == 0x4dae)
303: c = 0xfe9f;
304: else if (wc >= 0x9fb4 && wc < 0x9fbc)
305: c = gb18030ext_page9f[wc-0x9fb0];
306: else if (wc >= 0xfe10 && wc < 0xfe1a)
307: c = gb18030ext_pagefe[wc-0xfe10];
308: else if (wc == 0x20087)
309: c = 0xfe51;
310: else if (wc == 0x20089)
311: c = 0xfe52;
312: else if (wc == 0x200cc)
313: c = 0xfe53;
314: else if (wc == 0x215d7)
315: c = 0xfe6c;
316: else if (wc == 0x2298f)
317: c = 0xfe76;
318: else if (wc == 0x241fe)
319: c = 0xfe91;
320: if (c != 0) {
321: r[0] = (c >> 8); r[1] = (c & 0xff);
322: return 2;
323: }
324: return RET_ILUNI;
325: }
326: return RET_TOOSMALL;
327: }
FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>
![[BACK]](/icons/cvsweb/back.gif){kind=icon}
Return to gb18030ext.h CVS log ![[TXT]](/icons/cvsweb/text.gif){kind=icon}
![[DIR]](/icons/cvsweb/dir.gif){kind=icon}
Up to [ELWIX - Embedded LightWeight unIX -] / embedaddon / libiconv / lib