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_sjis_open.h"
36
37 #include "unicode_table_cp932_ext.h"
38 #include "unicode_table_jis.h"
39
40 static int mbfl_filt_ident_sjis_open(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_sjis_open_aliases[] = {"SJIS-open", "SJIS-ms", NULL};
62
63 const mbfl_encoding mbfl_encoding_sjis_open = {
64 mbfl_no_encoding_sjis_open,
65 "SJIS-win",
66 "Shift_JIS",
67 (const char *(*)[])&mbfl_encoding_sjis_open_aliases,
68 mblen_table_sjis,
69 MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_GL_UNSAFE,
70 &vtbl_sjis_open_wchar,
71 &vtbl_wchar_sjis_open
72 };
73
74 const struct mbfl_identify_vtbl vtbl_identify_sjis_open = {
75 mbfl_no_encoding_sjis_open,
76 mbfl_filt_ident_common_ctor,
77 mbfl_filt_ident_common_dtor,
78 mbfl_filt_ident_sjis_open
79 };
80
81 const struct mbfl_convert_vtbl vtbl_sjis_open_wchar = {
82 mbfl_no_encoding_sjis_open,
83 mbfl_no_encoding_wchar,
84 mbfl_filt_conv_common_ctor,
85 mbfl_filt_conv_common_dtor,
86 mbfl_filt_conv_sjis_open_wchar,
87 mbfl_filt_conv_common_flush
88 };
89
90 const struct mbfl_convert_vtbl vtbl_wchar_sjis_open = {
91 mbfl_no_encoding_wchar,
92 mbfl_no_encoding_sjis_open,
93 mbfl_filt_conv_common_ctor,
94 mbfl_filt_conv_common_dtor,
95 mbfl_filt_conv_wchar_sjis_open,
96 mbfl_filt_conv_common_flush
97 };
98
99 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
100
101 #define SJIS_ENCODE(c1,c2,s1,s2) \
102 do { \
103 s1 = c1; \
104 s1--; \
105 s1 >>= 1; \
106 if ((c1) < 0x5f) { \
107 s1 += 0x71; \
108 } else { \
109 s1 += 0xb1; \
110 } \
111 s2 = c2; \
112 if ((c1) & 1) { \
113 if ((c2) < 0x60) { \
114 s2--; \
115 } \
116 s2 += 0x20; \
117 } else { \
118 s2 += 0x7e; \
119 } \
120 } while (0)
121
122 #define SJIS_DECODE(c1,c2,s1,s2) \
123 do { \
124 s1 = c1; \
125 if (s1 < 0xa0) { \
126 s1 -= 0x81; \
127 } else { \
128 s1 -= 0xc1; \
129 } \
130 s1 <<= 1; \
131 s1 += 0x21; \
132 s2 = c2; \
133 if (s2 < 0x9f) { \
134 if (s2 < 0x7f) { \
135 s2++; \
136 } \
137 s2 -= 0x20; \
138 } else { \
139 s1++; \
140 s2 -= 0x7e; \
141 } \
142 } while (0)
143
144 /*
145 * SJIS-win => wchar
146 */
147 int
mbfl_filt_conv_sjis_open_wchar(int c,mbfl_convert_filter * filter)148 mbfl_filt_conv_sjis_open_wchar(int c, mbfl_convert_filter *filter)
149 {
150 int c1, s, s1, s2, w;
151
152 switch (filter->status) {
153 case 0:
154 if (c >= 0 && c < 0x80) { /* latin */
155 CK((*filter->output_function)(c, filter->data));
156 } else if (c > 0xa0 && c < 0xe0) { /* kana */
157 CK((*filter->output_function)(0xfec0 + c, filter->data));
158 } else if (c > 0x80 && c < 0xfd && c != 0xa0) { /* kanji first char */
159 filter->status = 1;
160 filter->cache = c;
161 } else {
162 w = c & MBFL_WCSGROUP_MASK;
163 w |= MBFL_WCSGROUP_THROUGH;
164 CK((*filter->output_function)(w, filter->data));
165 }
166 break;
167
168 case 1: /* kanji second char */
169 filter->status = 0;
170 c1 = filter->cache;
171 if (c >= 0x40 && c <= 0xfc && c != 0x7f) {
172 w = 0;
173 SJIS_DECODE(c1, c, s1, s2);
174 s = (s1 - 0x21)*94 + s2 - 0x21;
175 if (s <= 137) {
176 if (s == 31) {
177 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
178 } else if (s == 32) {
179 w = 0xff5e; /* FULLWIDTH TILDE */
180 } else if (s == 33) {
181 w = 0x2225; /* PARALLEL TO */
182 } else if (s == 60) {
183 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
184 } else if (s == 80) {
185 w = 0xffe0; /* FULLWIDTH CENT SIGN */
186 } else if (s == 81) {
187 w = 0xffe1; /* FULLWIDTH POUND SIGN */
188 } else if (s == 137) {
189 w = 0xffe2; /* FULLWIDTH NOT SIGN */
190 }
191 }
192 if (w == 0) {
193 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
194 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
195 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
196 w = jisx0208_ucs_table[s];
197 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
198 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
199 } else if (s >= cp932ext3_ucs_table_min && s < cp932ext3_ucs_table_max) { /* vendor ext3 (115ku - 119ku) */
200 w = cp932ext3_ucs_table[s - cp932ext3_ucs_table_min];
201 } else if (s >= (94*94) && s < (114*94)) { /* user (95ku - 114ku) */
202 w = s - (94*94) + 0xe000;
203 }
204 }
205 if (w <= 0) {
206 w = (s1 << 8) | s2;
207 w &= MBFL_WCSPLANE_MASK;
208 w |= MBFL_WCSPLANE_WINCP932;
209 }
210 CK((*filter->output_function)(w, filter->data));
211 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
212 CK((*filter->output_function)(c, filter->data));
213 } else {
214 w = (c1 << 8) | c;
215 w &= MBFL_WCSGROUP_MASK;
216 w |= MBFL_WCSGROUP_THROUGH;
217 CK((*filter->output_function)(w, filter->data));
218 }
219 break;
220
221 default:
222 filter->status = 0;
223 break;
224 }
225
226 return c;
227 }
228
229 /*
230 * wchar => SJIS-win
231 */
232 int
mbfl_filt_conv_wchar_sjis_open(int c,mbfl_convert_filter * filter)233 mbfl_filt_conv_wchar_sjis_open(int c, mbfl_convert_filter *filter)
234 {
235 int c1, c2, s1, s2;
236
237 s1 = 0;
238 s2 = 0;
239 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
240 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
241 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
242 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
243 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
244 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
245 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
246 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
247 } else if (c >= 0xe000 && c < (0xe000 + 20*94)) { /* user (95ku - 114ku) */
248 s1 = c - 0xe000;
249 c1 = s1/94 + 0x7f;
250 c2 = s1%94 + 0x21;
251 s1 = (c1 << 8) | c2;
252 s2 = 1;
253 }
254 if (s1 <= 0) {
255 c1 = c & ~MBFL_WCSPLANE_MASK;
256 if (c1 == MBFL_WCSPLANE_WINCP932) {
257 s1 = c & MBFL_WCSPLANE_MASK;
258 s2 = 1;
259 } else if (c1 == MBFL_WCSPLANE_JIS0208) {
260 s1 = c & MBFL_WCSPLANE_MASK;
261 } else if (c1 == MBFL_WCSPLANE_JIS0212) {
262 s1 = c & MBFL_WCSPLANE_MASK;
263 s1 |= 0x8080;
264 } else if (c == 0xa5) { /* YEN SIGN */
265 s1 = 0x216f; /* FULLWIDTH YEN SIGN */
266 } else if (c == 0x203e) { /* OVER LINE */
267 s1 = 0x2131; /* FULLWIDTH MACRON */
268 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
269 s1 = 0x2140;
270 } else if (c == 0xff5e) { /* FULLWIDTH TILDE */
271 s1 = 0x2141;
272 } else if (c == 0x2225) { /* PARALLEL TO */
273 s1 = 0x2142;
274 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
275 s1 = 0x215d;
276 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
277 s1 = 0x2171;
278 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
279 s1 = 0x2172;
280 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
281 s1 = 0x224c;
282 }
283 }
284 if ((s1 <= 0) || (s1 >= 0x8080 && s2 == 0)) { /* not found or X 0212 */
285 s1 = -1;
286 c1 = 0;
287 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
288 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
289 if (c == cp932ext1_ucs_table[c1]) {
290 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
291 break;
292 }
293 c1++;
294 }
295 if (s1 <= 0) {
296 c1 = 0;
297 c2 = cp932ext3_ucs_table_max - cp932ext3_ucs_table_min;
298 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
299 if (c == cp932ext3_ucs_table[c1]) {
300 s1 = ((c1/94 + 0x93) << 8) + (c1%94 + 0x21);
301 break;
302 }
303 c1++;
304 }
305 }
306 if (c == 0) {
307 s1 = 0;
308 } else if (s1 <= 0) {
309 s1 = -1;
310 }
311 }
312
313 if (s1 >= 0) {
314 if (s1 < 0x100) { /* latin or kana */
315 CK((*filter->output_function)(s1, filter->data));
316 } else { /* kanji */
317 c1 = (s1 >> 8) & 0xff;
318 c2 = s1 & 0xff;
319 SJIS_ENCODE(c1, c2, s1, s2);
320 CK((*filter->output_function)(s1, filter->data));
321 CK((*filter->output_function)(s2, filter->data));
322 }
323 } else {
324 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
325 CK(mbfl_filt_conv_illegal_output(c, filter));
326 }
327 }
328
329 return c;
330 }
331
mbfl_filt_ident_sjis_open(int c,mbfl_identify_filter * filter)332 static int mbfl_filt_ident_sjis_open(int c, mbfl_identify_filter *filter)
333 {
334 if (filter->status) { /* kanji second char */
335 if (c < 0x40 || c > 0xfc || c == 0x7f) { /* bad */
336 filter->flag = 1;
337 }
338 filter->status = 0;
339 } else if (c >= 0 && c < 0x80) { /* latin ok */
340 ;
341 } else if (c > 0xa0 && c < 0xe0) { /* kana ok */
342 ;
343 } else if (c > 0x80 && c < 0xfd && c != 0xa0) { /* kanji first char */
344 filter->status = 1;
345 } else { /* bad */
346 filter->flag = 1;
347 }
348
349 return c;
350 }
351