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_cp51932.h"
36
37 #include "unicode_table_cp932_ext.h"
38 #include "unicode_table_jis.h"
39 #include "cp932_table.h"
40
41 static int mbfl_filt_ident_cp51932(int c, mbfl_identify_filter *filter);
42
43 static const unsigned char mblen_table_eucjp[] = { /* 0xA1-0xFE */
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 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
52 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1,
53 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
54 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
55 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
56 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
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 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1
60 };
61
62
63 static const char *mbfl_encoding_cp51932_aliases[] = {"cp51932", NULL};
64
65 const struct mbfl_identify_vtbl vtbl_identify_cp51932 = {
66 mbfl_no_encoding_cp51932,
67 mbfl_filt_ident_common_ctor,
68 mbfl_filt_ident_common_dtor,
69 mbfl_filt_ident_cp51932
70 };
71
72 const mbfl_encoding mbfl_encoding_cp51932 = {
73 mbfl_no_encoding_cp51932,
74 "CP51932",
75 "CP51932",
76 (const char *(*)[])&mbfl_encoding_cp51932_aliases,
77 mblen_table_eucjp,
78 MBFL_ENCTYPE_MBCS,
79 &vtbl_cp51932_wchar,
80 &vtbl_wchar_cp51932
81 };
82
83 const struct mbfl_convert_vtbl vtbl_cp51932_wchar = {
84 mbfl_no_encoding_cp51932,
85 mbfl_no_encoding_wchar,
86 mbfl_filt_conv_common_ctor,
87 mbfl_filt_conv_common_dtor,
88 mbfl_filt_conv_cp51932_wchar,
89 mbfl_filt_conv_common_flush
90 };
91
92 const struct mbfl_convert_vtbl vtbl_wchar_cp51932 = {
93 mbfl_no_encoding_wchar,
94 mbfl_no_encoding_cp51932,
95 mbfl_filt_conv_common_ctor,
96 mbfl_filt_conv_common_dtor,
97 mbfl_filt_conv_wchar_cp51932,
98 mbfl_filt_conv_common_flush
99 };
100
101 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
102
103 #define sjistoidx(c1, c2) \
104 (((c1) > 0x9f) \
105 ? (((c1) - 0xc1) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)) \
106 : (((c1) - 0x81) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)))
107 #define idxtoeuc1(c) (((c) / 94) + 0xa1)
108 #define idxtoeuc2(c) (((c) % 94) + 0xa1)
109
110 /*
111 * cp51932 => wchar
112 */
113 int
mbfl_filt_conv_cp51932_wchar(int c,mbfl_convert_filter * filter)114 mbfl_filt_conv_cp51932_wchar(int c, mbfl_convert_filter *filter)
115 {
116 int c1, s, w;
117
118 switch (filter->status) {
119 case 0:
120 if (c >= 0 && c < 0x80) { /* latin */
121 CK((*filter->output_function)(c, filter->data));
122 } else if (c > 0xa0 && c < 0xff) { /* CP932 first char */
123 filter->status = 1;
124 filter->cache = c;
125 } else if (c == 0x8e) { /* kana first char */
126 filter->status = 2;
127 } else {
128 w = c & MBFL_WCSGROUP_MASK;
129 w |= MBFL_WCSGROUP_THROUGH;
130 CK((*filter->output_function)(w, filter->data));
131 }
132 break;
133
134 case 1: /* got first half */
135 filter->status = 0;
136 c1 = filter->cache;
137 if (c > 0xa0 && c < 0xff) {
138 w = 0;
139 s = (c1 - 0xa1)*94 + c - 0xa1;
140 if (s <= 137) {
141 if (s == 31) {
142 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
143 } else if (s == 32) {
144 w = 0xff5e; /* FULLWIDTH TILDE */
145 } else if (s == 33) {
146 w = 0x2225; /* PARALLEL TO */
147 } else if (s == 60) {
148 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
149 } else if (s == 80) {
150 w = 0xffe0; /* FULLWIDTH CENT SIGN */
151 } else if (s == 81) {
152 w = 0xffe1; /* FULLWIDTH POUND SIGN */
153 } else if (s == 137) {
154 w = 0xffe2; /* FULLWIDTH NOT SIGN */
155 }
156 }
157 if (w == 0) {
158 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
159 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
160 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
161 w = jisx0208_ucs_table[s];
162 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
163 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
164 }
165 }
166 if (w <= 0) {
167 w = ((c1 & 0x7f) << 8) | (c & 0x7f);
168 w &= MBFL_WCSPLANE_MASK;
169 w |= MBFL_WCSPLANE_WINCP932;
170 }
171 CK((*filter->output_function)(w, filter->data));
172 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
173 CK((*filter->output_function)(c, filter->data));
174 } else {
175 w = (c1 << 8) | c;
176 w &= MBFL_WCSGROUP_MASK;
177 w |= MBFL_WCSGROUP_THROUGH;
178 CK((*filter->output_function)(w, filter->data));
179 }
180 break;
181
182 case 2: /* got 0x8e, X0201 kana */
183 filter->status = 0;
184 if (c > 0xa0 && c < 0xe0) {
185 w = 0xfec0 + c;
186 CK((*filter->output_function)(w, filter->data));
187 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
188 CK((*filter->output_function)(c, filter->data));
189 } else {
190 w = 0x8e00 | c;
191 w &= MBFL_WCSGROUP_MASK;
192 w |= MBFL_WCSGROUP_THROUGH;
193 CK((*filter->output_function)(w, filter->data));
194 }
195 break;
196
197 default:
198 filter->status = 0;
199 break;
200 }
201
202 return c;
203 }
204
205 /*
206 * wchar => cp51932
207 */
208 int
mbfl_filt_conv_wchar_cp51932(int c,mbfl_convert_filter * filter)209 mbfl_filt_conv_wchar_cp51932(int c, mbfl_convert_filter *filter)
210 {
211 int c1, c2, s1;
212
213 s1 = 0;
214 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
215 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
216 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
217 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
218 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
219 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
220 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
221 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
222 }
223 if (s1 >= 0x8080) s1 = -1; /* we don't support JIS X0213 */
224 if (s1 <= 0) {
225 c1 = c & ~MBFL_WCSPLANE_MASK;
226 if (c1 == MBFL_WCSPLANE_WINCP932) {
227 s1 = c & MBFL_WCSPLANE_MASK;
228 if (s1 >= ((85 + 0x20) << 8)) { /* 85ku - 120ku */
229 s1 = -1;
230 }
231 } else if (c1 == MBFL_WCSPLANE_JIS0208) {
232 s1 = c & MBFL_WCSPLANE_MASK;
233 if ((s1 >= ((85 + 0x20) << 8) && /* 85ku - 94ku */
234 s1 <= ((88 + 0x20) << 8)) ||/* IBM extension */
235 (s1 >= ((93 + 0x20) << 8) && /* 89ku - 92ku */
236 s1 <= ((94 + 0x20) << 8))) {
237 s1 = -1;
238 }
239 } else if (c == 0xa5) { /* YEN SIGN */
240 s1 = 0x005c; /* YEN SIGN */
241 } else if (c == 0x203e) { /* OVER LINE */
242 s1 = 0x007e; /* FULLWIDTH MACRON */
243 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
244 s1 = 0x2140;
245 } else if (c == 0xff5e) { /* FULLWIDTH TILDE */
246 s1 = 0x2141;
247 } else if (c == 0x2225) { /* PARALLEL TO */
248 s1 = 0x2142;
249 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
250 s1 = 0x215d;
251 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
252 s1 = 0x2171;
253 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
254 s1 = 0x2172;
255 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
256 s1 = 0x224c;
257 } else {
258 s1 = -1;
259 c1 = 0;
260 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
261 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
262 if (c == cp932ext1_ucs_table[c1]) {
263 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
264 break;
265 }
266 c1++;
267 }
268 if (s1 < 0) {
269 c1 = 0;
270 c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
271 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
272 if (c == cp932ext2_ucs_table[c1]) {
273 s1 = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
274 break;
275 }
276 c1++;
277 }
278 }
279 }
280 if (c == 0) {
281 s1 = 0;
282 } else if (s1 <= 0) {
283 s1 = -1;
284 }
285 }
286
287 if (s1 >= 0) {
288 if (s1 < 0x80) { /* latin */
289 CK((*filter->output_function)(s1, filter->data));
290 } else if (s1 < 0x100) { /* kana */
291 CK((*filter->output_function)(0x8e, filter->data));
292 CK((*filter->output_function)(s1, filter->data));
293 } else if (s1 < 0x8080) { /* X 0208 */
294 CK((*filter->output_function)(((s1 >> 8) & 0xff) | 0x80, filter->data));
295 CK((*filter->output_function)((s1 & 0xff) | 0x80, filter->data));
296 } else {
297 CK(mbfl_filt_conv_illegal_output(c, filter));
298 }
299 } else {
300 CK(mbfl_filt_conv_illegal_output(c, filter));
301 }
302
303 return c;
304 }
305
mbfl_filt_ident_cp51932(int c,mbfl_identify_filter * filter)306 static int mbfl_filt_ident_cp51932(int c, mbfl_identify_filter *filter)
307 {
308 switch (filter->status) {
309 case 0: /* latin */
310 if (c >= 0 && c < 0x80) { /* ok */
311 ;
312 } else if (c > 0xa0 && c < 0xff) { /* kanji first char */
313 filter->status = 1;
314 } else if (c == 0x8e) { /* kana first char */
315 filter->status = 2;
316 } else { /* bad */
317 filter->flag = 1;
318 }
319 break;
320
321 case 1: /* got first half */
322 if (c < 0xa1 || c > 0xfe) { /* bad */
323 filter->flag = 1;
324 }
325 filter->status = 0;
326 break;
327
328 case 2: /* got 0x8e */
329 if (c < 0xa1 || c > 0xdf) { /* bad */
330 filter->flag = 1;
331 }
332 filter->status = 0;
333 break;
334
335 default:
336 filter->status = 0;
337 break;
338 }
339
340 return c;
341 }
342
