1 /*
2 * "streamable kanji code filter and converter"
3 * Copyright (c) 1998-2002 HappySize, Inc. All rights reserved.
4 *
5 * LICENSE NOTICES
6 *
7 * This file is part of "streamable kanji code filter and converter",
8 * which is distributed under the terms of GNU Lesser General Public
9 * License (version 2) as published by the Free Software Foundation.
10 *
11 * This software is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with "streamable kanji code filter and converter";
18 * if not, write to the Free Software Foundation, Inc., 59 Temple Place,
19 * Suite 330, Boston, MA 02111-1307 USA
20 *
21 * The author of this file:
22 *
23 */
24 /*
25 * the source code included in this files was separated from mbfilter_cn.c
26 * by moriyoshi koizumi <moriyoshi@php.net> on 4 dec 2002.
27 *
28 */
29
30 #include "mbfilter.h"
31 #include "mbfilter_cp936.h"
32 #define UNICODE_TABLE_CP936_DEF
33 #include "unicode_table_cp936.h"
34
35 static int mbfl_filt_ident_cp936(int c, mbfl_identify_filter *filter);
36
37 static const unsigned char mblen_table_cp936[] = { /* 0x81-0xFE */
38 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
39 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
40 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
41 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
42 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
43 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
44 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
45 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
46 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
47 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
48 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
49 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
50 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
51 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
52 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
53 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1
54 };
55
56 static const char *mbfl_encoding_cp936_aliases[] = {"CP-936", "GBK", NULL};
57
58 const mbfl_encoding mbfl_encoding_cp936 = {
59 mbfl_no_encoding_cp936,
60 "CP936",
61 "CP936",
62 (const char *(*)[])&mbfl_encoding_cp936_aliases,
63 mblen_table_cp936,
64 MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE,
65 &vtbl_cp936_wchar,
66 &vtbl_wchar_cp936
67 };
68
69 const struct mbfl_identify_vtbl vtbl_identify_cp936 = {
70 mbfl_no_encoding_cp936,
71 mbfl_filt_ident_common_ctor,
72 mbfl_filt_ident_cp936
73 };
74
75 const struct mbfl_convert_vtbl vtbl_cp936_wchar = {
76 mbfl_no_encoding_cp936,
77 mbfl_no_encoding_wchar,
78 mbfl_filt_conv_common_ctor,
79 NULL,
80 mbfl_filt_conv_cp936_wchar,
81 mbfl_filt_conv_common_flush,
82 NULL,
83 };
84
85 const struct mbfl_convert_vtbl vtbl_wchar_cp936 = {
86 mbfl_no_encoding_wchar,
87 mbfl_no_encoding_cp936,
88 mbfl_filt_conv_common_ctor,
89 NULL,
90 mbfl_filt_conv_wchar_cp936,
91 mbfl_filt_conv_common_flush,
92 NULL,
93 };
94
95
96 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
97
98 /*
99 * CP936 => wchar
100 */
101 int
mbfl_filt_conv_cp936_wchar(int c,mbfl_convert_filter * filter)102 mbfl_filt_conv_cp936_wchar(int c, mbfl_convert_filter *filter)
103 {
104 int k;
105 int c1, c2, w = -1;
106
107 switch (filter->status) {
108 case 0:
109 if (c >= 0 && c < 0x80) { /* latin */
110 CK((*filter->output_function)(c, filter->data));
111 } else if (c == 0x80) { /* euro sign */
112 CK((*filter->output_function)(0x20ac, filter->data));
113 } else if (c < 0xff) { /* dbcs lead byte */
114 filter->status = 1;
115 filter->cache = c;
116 } else { /* 0xff */
117 CK((*filter->output_function)(0xf8f5, filter->data));
118 }
119 break;
120
121 case 1: /* dbcs second byte */
122 filter->status = 0;
123 c1 = filter->cache;
124
125 if (((c1 >= 0xaa && c1 <= 0xaf) || (c1 >= 0xf8 && c1 <= 0xfe)) &&
126 (c >= 0xa1 && c <= 0xfe)) {
127 /* UDA part1,2: U+E000-U+E4C5 */
128 w = 94*(c1 >= 0xf8 ? c1 - 0xf2 : c1 - 0xaa) + (c - 0xa1) + 0xe000;
129 CK((*filter->output_function)(w, filter->data));
130 } else if (c1 >= 0xa1 && c1 <= 0xa7 && c >= 0x40 && c < 0xa1 && c != 0x7f) {
131 /* UDA part3 : U+E4C6-U+E765*/
132 w = 96*(c1 - 0xa1) + c - (c >= 0x80 ? 0x41 : 0x40) + 0xe4c6;
133 CK((*filter->output_function)(w, filter->data));
134 }
135
136 c2 = (c1 << 8) | c;
137
138 if (w <= 0 &&
139 ((c2 >= 0xa2ab && c2 <= 0xa9f0 + (0xe80f-0xe801)) ||
140 (c2 >= 0xd7fa && c2 <= 0xd7fa + (0xe814-0xe810)) ||
141 (c2 >= 0xfe50 && c2 <= 0xfe80 + (0xe864-0xe844)))) {
142 for (k = 0; k < mbfl_cp936_pua_tbl_max; k++) {
143 if (c2 >= mbfl_cp936_pua_tbl[k][2] &&
144 c2 <= mbfl_cp936_pua_tbl[k][2] +
145 mbfl_cp936_pua_tbl[k][1] - mbfl_cp936_pua_tbl[k][0]) {
146 w = c2 - mbfl_cp936_pua_tbl[k][2] + mbfl_cp936_pua_tbl[k][0];
147 CK((*filter->output_function)(w, filter->data));
148 break;
149 }
150 }
151 }
152
153 if (w <= 0) {
154 if (c1 < 0xff && c1 > 0x80 && c > 0x39 && c < 0xff && c != 0x7f) {
155 w = (c1 - 0x81)*192 + (c - 0x40);
156 if (w >= 0 && w < cp936_ucs_table_size) {
157 w = cp936_ucs_table[w];
158 } else {
159 w = 0;
160 }
161 if (w <= 0) {
162 w = (c1 << 8) | c;
163 w &= MBFL_WCSPLANE_MASK;
164 w |= MBFL_WCSPLANE_WINCP936;
165 }
166 CK((*filter->output_function)(w, filter->data));
167 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
168 CK((*filter->output_function)(c, filter->data));
169 } else {
170 w = (c1 << 8) | c;
171 w &= MBFL_WCSGROUP_MASK;
172 w |= MBFL_WCSGROUP_THROUGH;
173 CK((*filter->output_function)(w, filter->data));
174 }
175 }
176 break;
177
178 default:
179 filter->status = 0;
180 break;
181 }
182
183 return c;
184 }
185
186 /*
187 * wchar => CP936
188 */
189 int
mbfl_filt_conv_wchar_cp936(int c,mbfl_convert_filter * filter)190 mbfl_filt_conv_wchar_cp936(int c, mbfl_convert_filter *filter)
191 {
192 int k, k1, k2;
193 int c1, s = 0;
194
195 if (c >= ucs_a1_cp936_table_min && c < ucs_a1_cp936_table_max) {
196 /* U+0000 - U+0451 */
197 s = ucs_a1_cp936_table[c - ucs_a1_cp936_table_min];
198 } else if (c >= ucs_a2_cp936_table_min && c < ucs_a2_cp936_table_max) {
199 /* U+2000 - U+26FF */
200 if (c == 0x203e) {
201 s = 0xa3fe;
202 } else if (c == 0x2218) {
203 s = 0xa1e3;
204 } else if (c == 0x223c) {
205 s = 0xa1ab;
206 } else {
207 s = ucs_a2_cp936_table[c - ucs_a2_cp936_table_min];
208 }
209 } else if (c >= ucs_a3_cp936_table_min && c < ucs_a3_cp936_table_max) {
210 /* U+2F00 - U+33FF */
211 s = ucs_a3_cp936_table[c - ucs_a3_cp936_table_min];
212 } else if (c >= ucs_i_cp936_table_min && c < ucs_i_cp936_table_max) {
213 /* U+4D00-9FFF CJK Unified Ideographs (+ Extension A) */
214 s = ucs_i_cp936_table[c - ucs_i_cp936_table_min];
215 } else if (c >= 0xe000 && c <= 0xe864) { /* PUA */
216 if (c < 0xe766) {
217 if (c < 0xe4c6) {
218 c1 = c - 0xe000;
219 s = (c1 % 94) + 0xa1; c1 /= 94;
220 s |= (c1 < 0x06 ? c1 + 0xaa : c1 + 0xf2) << 8;
221 } else {
222 c1 = c - 0xe4c6;
223 s = ((c1 / 96) + 0xa1) << 8; c1 %= 96;
224 s |= c1 + (c1 >= 0x3f ? 0x41 : 0x40);
225 }
226 } else {
227 /* U+E766..U+E864 */
228 k1 = 0; k2 = mbfl_cp936_pua_tbl_max;
229 while (k1 < k2) {
230 k = (k1 + k2) >> 1;
231 if (c < mbfl_cp936_pua_tbl[k][0]) {
232 k2 = k;
233 } else if (c > mbfl_cp936_pua_tbl[k][1]) {
234 k1 = k + 1;
235 } else {
236 s = c - mbfl_cp936_pua_tbl[k][0] + mbfl_cp936_pua_tbl[k][2];
237 break;
238 }
239 }
240 }
241 } else if (c == 0xf8f5) {
242 s = 0xff;
243 } else if (c >= ucs_ci_cp936_table_min && c < ucs_ci_cp936_table_max) {
244 /* U+F900-FA2F CJK Compatibility Ideographs */
245 s = ucs_ci_cp936_table[c - ucs_ci_cp936_table_min];
246 } else if (c >= ucs_cf_cp936_table_min && c < ucs_cf_cp936_table_max) {
247 s = ucs_cf_cp936_table[c - ucs_cf_cp936_table_min];
248 } else if (c >= ucs_sfv_cp936_table_min && c < ucs_sfv_cp936_table_max) {
249 s = ucs_sfv_cp936_table[c - ucs_sfv_cp936_table_min]; /* U+FE50-FE6F Small Form Variants */
250 } else if (c >= ucs_hff_cp936_table_min && c < ucs_hff_cp936_table_max) {
251 /* U+FF00-FFFF HW/FW Forms */
252 if (c == 0xff04) {
253 s = 0xa1e7;
254 } else if (c == 0xff5e) {
255 s = 0xa1ab;
256 } else if (c >= 0xff01 && c <= 0xff5d) {
257 s = c - 0xff01 + 0xa3a1;
258 } else if (c >= 0xffe0 && c <= 0xffe5) {
259 s = ucs_hff_s_cp936_table[c-0xffe0];
260 }
261 }
262 if (s <= 0) {
263 c1 = c & ~MBFL_WCSPLANE_MASK;
264 if (c1 == MBFL_WCSPLANE_WINCP936) {
265 s = c & MBFL_WCSPLANE_MASK;
266 }
267 if (c == 0) {
268 s = 0;
269 } else if (s <= 0) {
270 s = -1;
271 }
272 }
273 if (s >= 0) {
274 if (s <= 0x80 || s == 0xff) { /* latin */
275 CK((*filter->output_function)(s, filter->data));
276 } else {
277 CK((*filter->output_function)((s >> 8) & 0xff, filter->data));
278 CK((*filter->output_function)(s & 0xff, filter->data));
279 }
280 } else {
281 CK(mbfl_filt_conv_illegal_output(c, filter));
282 }
283
284 return c;
285 }
286
mbfl_filt_ident_cp936(int c,mbfl_identify_filter * filter)287 static int mbfl_filt_ident_cp936(int c, mbfl_identify_filter *filter)
288 {
289 if (filter->status) { /* kanji second char */
290 if (c < 0x40 || c > 0xfe || c == 0x7f) { /* bad */
291 filter->flag = 1;
292 }
293 filter->status = 0;
294 } else if (c >= 0 && c < 0x80) { /* latin ok */
295 ;
296 } else if (c > 0x80 && c < 0xff) { /* DBCS lead byte */
297 filter->status = 1;
298 } else { /* bad */
299 filter->flag = 1;
300 }
301
302 return c;
303 }
304
