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_conv_cp51932_wchar_flush(mbfl_convert_filter *filter);
38 static size_t mb_cp51932_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state);
39 static void mb_wchar_to_cp51932(uint32_t *in, size_t len, mb_convert_buf *buf, bool end);
40
41 static const unsigned char mblen_table_eucjp[] = { /* 0xA1-0xFE */
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, 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, 2, 1,
51 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
52 1, 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, 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, 1
58 };
59
60 static const char *mbfl_encoding_cp51932_aliases[] = {"cp51932", NULL};
61
62 const mbfl_encoding mbfl_encoding_cp51932 = {
63 mbfl_no_encoding_cp51932,
64 "CP51932",
65 "CP51932",
66 mbfl_encoding_cp51932_aliases,
67 mblen_table_eucjp,
68 0,
69 &vtbl_cp51932_wchar,
70 &vtbl_wchar_cp51932,
71 mb_cp51932_to_wchar,
72 mb_wchar_to_cp51932,
73 NULL
74 };
75
76 const struct mbfl_convert_vtbl vtbl_cp51932_wchar = {
77 mbfl_no_encoding_cp51932,
78 mbfl_no_encoding_wchar,
79 mbfl_filt_conv_common_ctor,
80 NULL,
81 mbfl_filt_conv_cp51932_wchar,
82 mbfl_filt_conv_cp51932_wchar_flush,
83 NULL,
84 };
85
86 const struct mbfl_convert_vtbl vtbl_wchar_cp51932 = {
87 mbfl_no_encoding_wchar,
88 mbfl_no_encoding_cp51932,
89 mbfl_filt_conv_common_ctor,
90 NULL,
91 mbfl_filt_conv_wchar_cp51932,
92 mbfl_filt_conv_common_flush,
93 NULL,
94 };
95
96 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
97
98 /*
99 * cp51932 => wchar
100 */
101 int
mbfl_filt_conv_cp51932_wchar(int c,mbfl_convert_filter * filter)102 mbfl_filt_conv_cp51932_wchar(int c, mbfl_convert_filter *filter)
103 {
104 int c1, s, w;
105
106 switch (filter->status) {
107 case 0:
108 if (c >= 0 && c < 0x80) { /* latin */
109 CK((*filter->output_function)(c, filter->data));
110 } else if (c >= 0xA1 && c <= 0xFE) { /* CP932, first byte */
111 filter->status = 1;
112 filter->cache = c;
113 } else if (c == 0x8e) { /* kana first char */
114 filter->status = 2;
115 } else {
116 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
117 }
118 break;
119
120 case 1: /* got first half */
121 filter->status = 0;
122 c1 = filter->cache;
123 if (c > 0xa0 && c < 0xff) {
124 w = 0;
125 s = (c1 - 0xa1)*94 + c - 0xa1;
126 if (s <= 137) {
127 if (s == 31) {
128 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
129 } else if (s == 32) {
130 w = 0xff5e; /* FULLWIDTH TILDE */
131 } else if (s == 33) {
132 w = 0x2225; /* PARALLEL TO */
133 } else if (s == 60) {
134 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
135 } else if (s == 80) {
136 w = 0xffe0; /* FULLWIDTH CENT SIGN */
137 } else if (s == 81) {
138 w = 0xffe1; /* FULLWIDTH POUND SIGN */
139 } else if (s == 137) {
140 w = 0xffe2; /* FULLWIDTH NOT SIGN */
141 }
142 }
143 if (w == 0) {
144 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
145 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
146 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
147 w = jisx0208_ucs_table[s];
148 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
149 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
150 }
151 }
152 if (w <= 0) {
153 w = MBFL_BAD_INPUT;
154 }
155 CK((*filter->output_function)(w, filter->data));
156 } else {
157 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
158 }
159 break;
160
161 case 2: /* got 0x8e, X0201 kana */
162 filter->status = 0;
163 if (c > 0xa0 && c < 0xe0) {
164 w = 0xfec0 + c;
165 CK((*filter->output_function)(w, filter->data));
166 } else {
167 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
168 }
169 break;
170
171 EMPTY_SWITCH_DEFAULT_CASE();
172 }
173
174 return 0;
175 }
176
mbfl_filt_conv_cp51932_wchar_flush(mbfl_convert_filter * filter)177 static int mbfl_filt_conv_cp51932_wchar_flush(mbfl_convert_filter *filter)
178 {
179 if (filter->status) {
180 /* Input string was truncated */
181 (*filter->output_function)(MBFL_BAD_INPUT, filter->data);
182 filter->status = 0;
183 }
184
185 if (filter->flush_function) {
186 (*filter->flush_function)(filter->data);
187 }
188
189 return 0;
190 }
191
192 /*
193 * wchar => cp51932
194 */
195 int
mbfl_filt_conv_wchar_cp51932(int c,mbfl_convert_filter * filter)196 mbfl_filt_conv_wchar_cp51932(int c, mbfl_convert_filter *filter)
197 {
198 int c1, c2, s1;
199
200 s1 = 0;
201 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
202 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
203 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
204 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
205 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
206 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
207 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
208 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
209 }
210 if (s1 >= 0x8080) s1 = -1; /* we don't support JIS X0213 */
211 if (s1 <= 0) {
212 if (c == 0xa5) { /* YEN SIGN */
213 s1 = 0x216F; /* FULLWIDTH YEN SIGN */
214 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
215 s1 = 0x2140;
216 } else if (c == 0x2225) { /* PARALLEL TO */
217 s1 = 0x2142;
218 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
219 s1 = 0x215d;
220 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
221 s1 = 0x2171;
222 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
223 s1 = 0x2172;
224 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
225 s1 = 0x224c;
226 } else {
227 s1 = -1;
228 c1 = 0;
229 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
230 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
231 if (c == cp932ext1_ucs_table[c1]) {
232 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
233 break;
234 }
235 c1++;
236 }
237 if (s1 < 0) {
238 c1 = 0;
239 c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
240 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
241 if (c == cp932ext2_ucs_table[c1]) {
242 s1 = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
243 break;
244 }
245 c1++;
246 }
247 }
248 }
249 if (c == 0) {
250 s1 = 0;
251 } else if (s1 <= 0) {
252 s1 = -1;
253 }
254 }
255
256 if (s1 >= 0) {
257 if (s1 < 0x80) { /* latin */
258 CK((*filter->output_function)(s1, filter->data));
259 } else if (s1 < 0x100) { /* kana */
260 CK((*filter->output_function)(0x8e, filter->data));
261 CK((*filter->output_function)(s1, filter->data));
262 } else if (s1 < 0x8080) { /* X 0208 */
263 CK((*filter->output_function)(((s1 >> 8) & 0xff) | 0x80, filter->data));
264 CK((*filter->output_function)((s1 & 0xff) | 0x80, filter->data));
265 } else {
266 CK(mbfl_filt_conv_illegal_output(c, filter));
267 }
268 } else {
269 CK(mbfl_filt_conv_illegal_output(c, filter));
270 }
271
272 return 0;
273 }
274
mb_cp51932_to_wchar(unsigned char ** in,size_t * in_len,uint32_t * buf,size_t bufsize,unsigned int * state)275 static size_t mb_cp51932_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state)
276 {
277 unsigned char *p = *in, *e = p + *in_len;
278 uint32_t *out = buf, *limit = buf + bufsize;
279
280 while (p < e && out < limit) {
281 unsigned char c = *p++;
282
283 if (c < 0x80) {
284 *out++ = c;
285 } else if (c >= 0xA1 && c <= 0xFE && p < e) {
286 unsigned char c2 = *p++;
287 if (c2 >= 0xA1 && c2 <= 0xFE) {
288 unsigned int s = (c - 0xA1)*94 + c2 - 0xA1, w = 0;
289
290 if (s <= 137) {
291 if (s == 31) {
292 w = 0xFF3C; /* FULLWIDTH REVERSE SOLIDUS */
293 } else if (s == 32) {
294 w = 0xFF5E; /* FULLWIDTH TILDE */
295 } else if (s == 33) {
296 w = 0x2225; /* PARALLEL TO */
297 } else if (s == 60) {
298 w = 0xFF0D; /* FULLWIDTH HYPHEN-MINUS */
299 } else if (s == 80) {
300 w = 0xFFE0; /* FULLWIDTH CENT SIGN */
301 } else if (s == 81) {
302 w = 0xFFE1; /* FULLWIDTH POUND SIGN */
303 } else if (s == 137) {
304 w = 0xFFE2; /* FULLWIDTH NOT SIGN */
305 }
306 }
307
308 if (w == 0) {
309 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) {
310 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
311 } else if (s < jisx0208_ucs_table_size) {
312 w = jisx0208_ucs_table[s];
313 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) {
314 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
315 }
316 }
317
318 if (!w)
319 w = MBFL_BAD_INPUT;
320 *out++ = w;
321 } else {
322 *out++ = MBFL_BAD_INPUT;
323 }
324 } else if (c == 0x8E && p < e) {
325 unsigned char c2 = *p++;
326 if (c2 >= 0xA1 && c2 <= 0xDF) {
327 *out++ = 0xFEC0 + c2;
328 } else {
329 *out++ = MBFL_BAD_INPUT;
330 }
331 } else {
332 *out++ = MBFL_BAD_INPUT;
333 }
334 }
335
336 *in_len = e - p;
337 *in = p;
338 return out - buf;
339 }
340
mb_wchar_to_cp51932(uint32_t * in,size_t len,mb_convert_buf * buf,bool end)341 static void mb_wchar_to_cp51932(uint32_t *in, size_t len, mb_convert_buf *buf, bool end)
342 {
343 unsigned char *out, *limit;
344 MB_CONVERT_BUF_LOAD(buf, out, limit);
345 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 2);
346
347 while (len--) {
348 uint32_t w = *in++;
349 unsigned int s = 0;
350
351 if (w == 0) {
352 out = mb_convert_buf_add(out, 0);
353 continue;
354 } else if (w >= ucs_a1_jis_table_min && w < ucs_a1_jis_table_max) {
355 s = ucs_a1_jis_table[w - ucs_a1_jis_table_min];
356 } else if (w >= ucs_a2_jis_table_min && w < ucs_a2_jis_table_max) {
357 s = ucs_a2_jis_table[w - ucs_a2_jis_table_min];
358 } else if (w >= ucs_i_jis_table_min && w < ucs_i_jis_table_max) {
359 s = ucs_i_jis_table[w - ucs_i_jis_table_min];
360 } else if (w >= ucs_r_jis_table_min && w < ucs_r_jis_table_max) {
361 s = ucs_r_jis_table[w - ucs_r_jis_table_min];
362 }
363
364 if (s >= 0x8080) s = 0; /* We don't support JIS X0213 */
365
366 if (s == 0) {
367 if (w == 0xA5) { /* YEN SIGN */
368 s = 0x216F; /* FULLWIDTH YEN SIGN */
369 } else if (w == 0xFF3C) { /* FULLWIDTH REVERSE SOLIDUS */
370 s = 0x2140;
371 } else if (w == 0x2225) { /* PARALLEL TO */
372 s = 0x2142;
373 } else if (w == 0xFF0D) { /* FULLWIDTH HYPHEN-MINUS */
374 s = 0x215D;
375 } else if (w == 0xFFE0) { /* FULLWIDTH CENT SIGN */
376 s = 0x2171;
377 } else if (w == 0xFFE1) { /* FULLWIDTH POUND SIGN */
378 s = 0x2172;
379 } else if (w == 0xFFE2) { /* FULLWIDTH NOT SIGN */
380 s = 0x224C;
381 } else {
382 for (int i = 0; i < cp932ext1_ucs_table_max - cp932ext1_ucs_table_min; i++) {
383 if (cp932ext1_ucs_table[i] == w) {
384 s = ((i/94 + 0x2D) << 8) + (i%94) + 0x21;
385 goto found_it;
386 }
387 }
388
389 for (int i = 0; i < cp932ext2_ucs_table_max - cp932ext2_ucs_table_min; i++) {
390 if (cp932ext2_ucs_table[i] == w) {
391 s = ((i/94 + 0x79) << 8) + (i%94) + 0x21;
392 goto found_it;
393 }
394 }
395 }
396 found_it: ;
397 }
398
399 if (!s || s >= 0x8080) {
400 MB_CONVERT_ERROR(buf, out, limit, w, mb_wchar_to_cp51932);
401 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 2);
402 } else if (s < 0x80) {
403 out = mb_convert_buf_add(out, s);
404 } else if (s < 0x100) {
405 out = mb_convert_buf_add2(out, 0x8E, s);
406 } else {
407 out = mb_convert_buf_add2(out, ((s >> 8) & 0xFF) | 0x80, (s & 0xFF) | 0x80);
408 }
409 }
410
411 MB_CONVERT_BUF_STORE(buf, out, limit);
412 }
413