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_ja.c
26 * by moriyoshi koizumi <moriyoshi@php.net> on 4 dec 2002.
27 *
28 */
29
30 #ifdef HAVE_CONFIG_H
31 #include "config.h"
32 #endif
33
34 #include "mbfilter.h"
35 #include "mbfilter_cp932.h"
36
37 #include "unicode_table_cp932_ext.h"
38 #include "unicode_table_jis.h"
39
40 static int mbfl_filt_ident_cp932(int c, mbfl_identify_filter *filter);
41
42 static const unsigned char mblen_table_sjis[] = { /* 0x80-0x9f,0xE0-0xFF */
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, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
47 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
48 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
49 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
50 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
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 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
54 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
55 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
56 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
57 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
58 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2
59 };
60
61 static const char *mbfl_encoding_cp932_aliases[] = {"MS932", "Windows-31J", "MS_Kanji", NULL};
62
63 const mbfl_encoding mbfl_encoding_cp932 = {
64 mbfl_no_encoding_cp932,
65 "CP932",
66 "Shift_JIS",
67 (const char *(*)[])&mbfl_encoding_cp932_aliases,
68 mblen_table_sjis,
69 MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE
70 };
71
72 const struct mbfl_identify_vtbl vtbl_identify_cp932 = {
73 mbfl_no_encoding_cp932,
74 mbfl_filt_ident_common_ctor,
75 mbfl_filt_ident_common_dtor,
76 mbfl_filt_ident_cp932
77 };
78
79 const struct mbfl_convert_vtbl vtbl_cp932_wchar = {
80 mbfl_no_encoding_cp932,
81 mbfl_no_encoding_wchar,
82 mbfl_filt_conv_common_ctor,
83 mbfl_filt_conv_common_dtor,
84 mbfl_filt_conv_cp932_wchar,
85 mbfl_filt_conv_common_flush
86 };
87
88 const struct mbfl_convert_vtbl vtbl_wchar_cp932 = {
89 mbfl_no_encoding_wchar,
90 mbfl_no_encoding_cp932,
91 mbfl_filt_conv_common_ctor,
92 mbfl_filt_conv_common_dtor,
93 mbfl_filt_conv_wchar_cp932,
94 mbfl_filt_conv_common_flush
95 };
96
97 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
98
99 #define SJIS_ENCODE(c1,c2,s1,s2) \
100 do { \
101 s1 = c1; \
102 s1--; \
103 s1 >>= 1; \
104 if ((c1) < 0x5f) { \
105 s1 += 0x71; \
106 } else { \
107 s1 += 0xb1; \
108 } \
109 s2 = c2; \
110 if ((c1) & 1) { \
111 if ((c2) < 0x60) { \
112 s2--; \
113 } \
114 s2 += 0x20; \
115 } else { \
116 s2 += 0x7e; \
117 } \
118 } while (0)
119
120 #define SJIS_DECODE(c1,c2,s1,s2) \
121 do { \
122 s1 = c1; \
123 if (s1 < 0xa0) { \
124 s1 -= 0x81; \
125 } else { \
126 s1 -= 0xc1; \
127 } \
128 s1 <<= 1; \
129 s1 += 0x21; \
130 s2 = c2; \
131 if (s2 < 0x9f) { \
132 if (s2 < 0x7f) { \
133 s2++; \
134 } \
135 s2 -= 0x20; \
136 } else { \
137 s1++; \
138 s2 -= 0x7e; \
139 } \
140 } while (0)
141
142
143 /*
144 * SJIS-win => wchar
145 */
146 int
147 mbfl_filt_conv_cp932_wchar(int c, mbfl_convert_filter *filter)
148 {
149 int c1, s, s1, s2, w;
150
151 switch (filter->status) {
152 case 0:
153 if (c >= 0 && c < 0x80) { /* latin */
154 CK((*filter->output_function)(c, filter->data));
155 } else if (c > 0xa0 && c < 0xe0) { /* kana */
156 CK((*filter->output_function)(0xfec0 + c, filter->data));
157 } else if (c > 0x80 && c < 0xfd && c != 0xa0) { /* kanji first char */
158 filter->status = 1;
159 filter->cache = c;
160 } else {
161 w = c & MBFL_WCSGROUP_MASK;
162 w |= MBFL_WCSGROUP_THROUGH;
163 CK((*filter->output_function)(w, filter->data));
164 }
165 break;
166
167 case 1: /* kanji second char */
168 filter->status = 0;
169 c1 = filter->cache;
170 if (c >= 0x40 && c <= 0xfc && c != 0x7f) {
171 w = 0;
172 SJIS_DECODE(c1, c, s1, s2);
173 s = (s1 - 0x21)*94 + s2 - 0x21;
174 if (s <= 137) {
175 if (s == 31) {
176 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
177 } else if (s == 32) {
178 w = 0xff5e; /* FULLWIDTH TILDE */
179 } else if (s == 33) {
180 w = 0x2225; /* PARALLEL TO */
181 } else if (s == 60) {
182 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
183 } else if (s == 80) {
184 w = 0xffe0; /* FULLWIDTH CENT SIGN */
185 } else if (s == 81) {
186 w = 0xffe1; /* FULLWIDTH POUND SIGN */
187 } else if (s == 137) {
188 w = 0xffe2; /* FULLWIDTH NOT SIGN */
189 }
190 }
191 if (w == 0) {
192 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
193 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
194 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
195 w = jisx0208_ucs_table[s];
196 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
197 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
198 } else if (s >= cp932ext3_ucs_table_min && s < cp932ext3_ucs_table_max) { /* vendor ext3 (115ku - 119ku) */
199 w = cp932ext3_ucs_table[s - cp932ext3_ucs_table_min];
200 } else if (s >= (94*94) && s < (114*94)) { /* user (95ku - 114ku) */
201 w = s - (94*94) + 0xe000;
202 }
203 }
204 if (w <= 0) {
205 w = (s1 << 8) | s2;
206 w &= MBFL_WCSPLANE_MASK;
207 w |= MBFL_WCSPLANE_WINCP932;
208 }
209 CK((*filter->output_function)(w, filter->data));
210 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
211 CK((*filter->output_function)(c, filter->data));
212 } else {
213 w = (c1 << 8) | c;
214 w &= MBFL_WCSGROUP_MASK;
215 w |= MBFL_WCSGROUP_THROUGH;
216 CK((*filter->output_function)(w, filter->data));
217 }
218 break;
219
220 default:
221 filter->status = 0;
222 break;
223 }
224
225 return c;
226 }
227
228 /*
229 * wchar => SJIS-win
230 */
231 int
232 mbfl_filt_conv_wchar_cp932(int c, mbfl_convert_filter *filter)
233 {
234 int c1, c2, s1, s2;
235
236 s1 = 0;
237 s2 = 0;
238 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
239 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
240 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
241 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
242 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
243 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
244 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
245 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
246 } else if (c >= 0xe000 && c < (0xe000 + 20*94)) { /* user (95ku - 114ku) */
247 s1 = c - 0xe000;
248 c1 = s1/94 + 0x7f;
249 c2 = s1%94 + 0x21;
250 s1 = (c1 << 8) | c2;
251 s2 = 1;
252 }
253 if (s1 <= 0) {
254 c1 = c & ~MBFL_WCSPLANE_MASK;
255 if (c1 == MBFL_WCSPLANE_WINCP932) {
256 s1 = c & MBFL_WCSPLANE_MASK;
257 s2 = 1;
258 } else if (c1 == MBFL_WCSPLANE_JIS0208) {
259 s1 = c & MBFL_WCSPLANE_MASK;
260 } else if (c1 == MBFL_WCSPLANE_JIS0212) {
261 s1 = c & MBFL_WCSPLANE_MASK;
262 s1 |= 0x8080;
263 } else if (c == 0xa5) { /* YEN SIGN */
264 s1 = 0x005c; /* YEN SIGN */
265 } else if (c == 0x203e) { /* OVER LINE */
266 s1 = 0x007e; /* FULLWIDTH MACRON */
267 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
268 s1 = 0x2140;
269 } else if (c == 0xff5e) { /* FULLWIDTH TILDE */
270 s1 = 0x2141;
271 } else if (c == 0x2225) { /* PARALLEL TO */
272 s1 = 0x2142;
273 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
274 s1 = 0x215d;
275 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
276 s1 = 0x2171;
277 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
278 s1 = 0x2172;
279 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
280 s1 = 0x224c;
281 }
282 }
283 if ((s1 <= 0) || (s1 >= 0x8080 && s2 == 0)) { /* not found or X 0212 */
284 s1 = -1;
285 c1 = 0;
286 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
287 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
288 if (c == cp932ext1_ucs_table[c1]) {
289 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
290 break;
291 }
292 c1++;
293 }
294 if (s1 <= 0) {
295 c1 = 0;
296 c2 = cp932ext3_ucs_table_max - cp932ext3_ucs_table_min;
297 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
298 if (c == cp932ext3_ucs_table[c1]) {
299 s1 = ((c1/94 + 0x93) << 8) + (c1%94 + 0x21);
300 break;
301 }
302 c1++;
303 }
304 }
305 if (c == 0) {
306 s1 = 0;
307 } else if (s1 <= 0) {
308 s1 = -1;
309 }
310 }
311 if (s1 >= 0) {
312 if (s1 < 0x100) { /* latin or kana */
313 CK((*filter->output_function)(s1, filter->data));
314 } else { /* kanji */
315 c1 = (s1 >> 8) & 0xff;
316 c2 = s1 & 0xff;
317 SJIS_ENCODE(c1, c2, s1, s2);
318 CK((*filter->output_function)(s1, filter->data));
319 CK((*filter->output_function)(s2, filter->data));
320 }
321 } else {
322 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
323 CK(mbfl_filt_conv_illegal_output(c, filter));
324 }
325 }
326
327 return c;
328 }
329
330 static int mbfl_filt_ident_cp932(int c, mbfl_identify_filter *filter)
331 {
332 if (filter->status) { /* kanji second char */
333 if (c < 0x40 || c > 0xfc || c == 0x7f) { /* bad */
334 filter->flag = 1;
335 }
336 filter->status = 0;
337 } else if (c >= 0 && c < 0x80) { /* latin ok */
338 ;
339 } else if (c > 0xa0 && c < 0xe0) { /* kana ok */
340 ;
341 } else if (c > 0x80 && c < 0xfd && c != 0xa0) { /* kanji first char */
342 filter->status = 1;
343 } else { /* bad */
344 filter->flag = 1;
345 }
346
347 return c;
348 }
349
350