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 #include "mbfilter.h"
31 #include "mbfilter_cp51932.h"
32
33 #include "unicode_table_cp932_ext.h"
34 #include "unicode_table_jis.h"
35 #include "cp932_table.h"
36
37 static int mbfl_filt_ident_cp51932(int c, mbfl_identify_filter *filter);
38
39 static const unsigned char mblen_table_eucjp[] = { /* 0xA1-0xFE */
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, 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, 2, 1,
49 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
50 1, 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, 2,
54 2, 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, 1
56 };
57
58
59 static const char *mbfl_encoding_cp51932_aliases[] = {"cp51932", NULL};
60
61 const struct mbfl_identify_vtbl vtbl_identify_cp51932 = {
62 mbfl_no_encoding_cp51932,
63 mbfl_filt_ident_common_ctor,
64 mbfl_filt_ident_cp51932
65 };
66
67 const mbfl_encoding mbfl_encoding_cp51932 = {
68 mbfl_no_encoding_cp51932,
69 "CP51932",
70 "CP51932",
71 (const char *(*)[])&mbfl_encoding_cp51932_aliases,
72 mblen_table_eucjp,
73 MBFL_ENCTYPE_MBCS,
74 &vtbl_cp51932_wchar,
75 &vtbl_wchar_cp51932
76 };
77
78 const struct mbfl_convert_vtbl vtbl_cp51932_wchar = {
79 mbfl_no_encoding_cp51932,
80 mbfl_no_encoding_wchar,
81 mbfl_filt_conv_common_ctor,
82 NULL,
83 mbfl_filt_conv_cp51932_wchar,
84 mbfl_filt_conv_common_flush,
85 NULL,
86 };
87
88 const struct mbfl_convert_vtbl vtbl_wchar_cp51932 = {
89 mbfl_no_encoding_wchar,
90 mbfl_no_encoding_cp51932,
91 mbfl_filt_conv_common_ctor,
92 NULL,
93 mbfl_filt_conv_wchar_cp51932,
94 mbfl_filt_conv_common_flush,
95 NULL,
96 };
97
98 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
99
100 #define sjistoidx(c1, c2) \
101 (((c1) > 0x9f) \
102 ? (((c1) - 0xc1) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)) \
103 : (((c1) - 0x81) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)))
104 #define idxtoeuc1(c) (((c) / 94) + 0xa1)
105 #define idxtoeuc2(c) (((c) % 94) + 0xa1)
106
107 /*
108 * cp51932 => wchar
109 */
110 int
mbfl_filt_conv_cp51932_wchar(int c,mbfl_convert_filter * filter)111 mbfl_filt_conv_cp51932_wchar(int c, mbfl_convert_filter *filter)
112 {
113 int c1, s, w;
114
115 switch (filter->status) {
116 case 0:
117 if (c >= 0 && c < 0x80) { /* latin */
118 CK((*filter->output_function)(c, filter->data));
119 } else if (c > 0xa0 && c < 0xff) { /* CP932 first char */
120 filter->status = 1;
121 filter->cache = c;
122 } else if (c == 0x8e) { /* kana first char */
123 filter->status = 2;
124 } else {
125 w = c & MBFL_WCSGROUP_MASK;
126 w |= MBFL_WCSGROUP_THROUGH;
127 CK((*filter->output_function)(w, filter->data));
128 }
129 break;
130
131 case 1: /* got first half */
132 filter->status = 0;
133 c1 = filter->cache;
134 if (c > 0xa0 && c < 0xff) {
135 w = 0;
136 s = (c1 - 0xa1)*94 + c - 0xa1;
137 if (s <= 137) {
138 if (s == 31) {
139 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
140 } else if (s == 32) {
141 w = 0xff5e; /* FULLWIDTH TILDE */
142 } else if (s == 33) {
143 w = 0x2225; /* PARALLEL TO */
144 } else if (s == 60) {
145 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
146 } else if (s == 80) {
147 w = 0xffe0; /* FULLWIDTH CENT SIGN */
148 } else if (s == 81) {
149 w = 0xffe1; /* FULLWIDTH POUND SIGN */
150 } else if (s == 137) {
151 w = 0xffe2; /* FULLWIDTH NOT SIGN */
152 }
153 }
154 if (w == 0) {
155 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
156 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
157 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
158 w = jisx0208_ucs_table[s];
159 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
160 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
161 }
162 }
163 if (w <= 0) {
164 w = ((c1 & 0x7f) << 8) | (c & 0x7f);
165 w &= MBFL_WCSPLANE_MASK;
166 w |= MBFL_WCSPLANE_WINCP932;
167 }
168 CK((*filter->output_function)(w, filter->data));
169 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
170 CK((*filter->output_function)(c, filter->data));
171 } else {
172 w = (c1 << 8) | c;
173 w &= MBFL_WCSGROUP_MASK;
174 w |= MBFL_WCSGROUP_THROUGH;
175 CK((*filter->output_function)(w, filter->data));
176 }
177 break;
178
179 case 2: /* got 0x8e, X0201 kana */
180 filter->status = 0;
181 if (c > 0xa0 && c < 0xe0) {
182 w = 0xfec0 + c;
183 CK((*filter->output_function)(w, filter->data));
184 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
185 CK((*filter->output_function)(c, filter->data));
186 } else {
187 w = 0x8e00 | c;
188 w &= MBFL_WCSGROUP_MASK;
189 w |= MBFL_WCSGROUP_THROUGH;
190 CK((*filter->output_function)(w, filter->data));
191 }
192 break;
193
194 default:
195 filter->status = 0;
196 break;
197 }
198
199 return c;
200 }
201
202 /*
203 * wchar => cp51932
204 */
205 int
mbfl_filt_conv_wchar_cp51932(int c,mbfl_convert_filter * filter)206 mbfl_filt_conv_wchar_cp51932(int c, mbfl_convert_filter *filter)
207 {
208 int c1, c2, s1;
209
210 s1 = 0;
211 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
212 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
213 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
214 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
215 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
216 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
217 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
218 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
219 }
220 if (s1 >= 0x8080) s1 = -1; /* we don't support JIS X0213 */
221 if (s1 <= 0) {
222 c1 = c & ~MBFL_WCSPLANE_MASK;
223 if (c1 == MBFL_WCSPLANE_WINCP932) {
224 s1 = c & MBFL_WCSPLANE_MASK;
225 if (s1 >= ((85 + 0x20) << 8)) { /* 85ku - 120ku */
226 s1 = -1;
227 }
228 } else if (c1 == MBFL_WCSPLANE_JIS0208) {
229 s1 = c & MBFL_WCSPLANE_MASK;
230 if ((s1 >= ((85 + 0x20) << 8) && /* 85ku - 94ku */
231 s1 <= ((88 + 0x20) << 8)) ||/* IBM extension */
232 (s1 >= ((93 + 0x20) << 8) && /* 89ku - 92ku */
233 s1 <= ((94 + 0x20) << 8))) {
234 s1 = -1;
235 }
236 } else if (c == 0xa5) { /* YEN SIGN */
237 s1 = 0x005c; /* YEN SIGN */
238 } else if (c == 0x203e) { /* OVER LINE */
239 s1 = 0x007e; /* FULLWIDTH MACRON */
240 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
241 s1 = 0x2140;
242 } else if (c == 0xff5e) { /* FULLWIDTH TILDE */
243 s1 = 0x2141;
244 } else if (c == 0x2225) { /* PARALLEL TO */
245 s1 = 0x2142;
246 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
247 s1 = 0x215d;
248 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
249 s1 = 0x2171;
250 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
251 s1 = 0x2172;
252 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
253 s1 = 0x224c;
254 } else {
255 s1 = -1;
256 c1 = 0;
257 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
258 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
259 if (c == cp932ext1_ucs_table[c1]) {
260 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
261 break;
262 }
263 c1++;
264 }
265 if (s1 < 0) {
266 c1 = 0;
267 c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
268 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
269 if (c == cp932ext2_ucs_table[c1]) {
270 s1 = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
271 break;
272 }
273 c1++;
274 }
275 }
276 }
277 if (c == 0) {
278 s1 = 0;
279 } else if (s1 <= 0) {
280 s1 = -1;
281 }
282 }
283
284 if (s1 >= 0) {
285 if (s1 < 0x80) { /* latin */
286 CK((*filter->output_function)(s1, filter->data));
287 } else if (s1 < 0x100) { /* kana */
288 CK((*filter->output_function)(0x8e, filter->data));
289 CK((*filter->output_function)(s1, filter->data));
290 } else if (s1 < 0x8080) { /* X 0208 */
291 CK((*filter->output_function)(((s1 >> 8) & 0xff) | 0x80, filter->data));
292 CK((*filter->output_function)((s1 & 0xff) | 0x80, filter->data));
293 } else {
294 CK(mbfl_filt_conv_illegal_output(c, filter));
295 }
296 } else {
297 CK(mbfl_filt_conv_illegal_output(c, filter));
298 }
299
300 return c;
301 }
302
mbfl_filt_ident_cp51932(int c,mbfl_identify_filter * filter)303 static int mbfl_filt_ident_cp51932(int c, mbfl_identify_filter *filter)
304 {
305 switch (filter->status) {
306 case 0: /* latin */
307 if (c >= 0 && c < 0x80) { /* ok */
308 ;
309 } else if (c > 0xa0 && c < 0xff) { /* kanji first char */
310 filter->status = 1;
311 } else if (c == 0x8e) { /* kana first char */
312 filter->status = 2;
313 } else { /* bad */
314 filter->flag = 1;
315 }
316 break;
317
318 case 1: /* got first half */
319 if (c < 0xa1 || c > 0xfe) { /* bad */
320 filter->flag = 1;
321 }
322 filter->status = 0;
323 break;
324
325 case 2: /* got 0x8e */
326 if (c < 0xa1 || c > 0xdf) { /* bad */
327 filter->flag = 1;
328 }
329 filter->status = 0;
330 break;
331
332 default:
333 filter->status = 0;
334 break;
335 }
336
337 return c;
338 }
339
