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 /* CP932 is Microsoft's version of Shift-JIS.
31 *
32 * What we call "SJIS-win" is a variant of CP932 which maps U+00A5
33 * and U+203E the same way as eucJP-win; namely, instead of mapping
34 * U+00A5 (YEN SIGN) to 0x5C and U+203E (OVERLINE) to 0x7E,
35 * these codepoints are mapped to appropriate JIS X 0208 characters.
36 *
37 * When converting from Shift-JIS to Unicode, there is no difference
38 * between CP932 and "SJIS-win".
39 *
40 * Additional facts:
41 *
42 * • In the libmbfl library which formed the base for mbstring, "CP932" and
43 * "SJIS-win" were originally aliases. The differing mappings were added in
44 * December 2002. The libmbfl author later stated that this was done so that
45 * "CP932" would comply with a certain specification, while "SJIS-win" would
46 * maintain the existing mappings. He does not remember which specification
47 * it was.
48 * • The WHATWG specification for "Shift_JIS" (followed by web browsers)
49 * agrees with our mappings for "CP932".
50 * • Microsoft Windows' "best-fit" mappings for CP932 (via the
51 * WideCharToMultiByte API) convert U+00A5 to 0x5C, which also agrees with
52 * our mappings for "CP932".
53 * • glibc's iconv converts U+203E to CP932 0x7E, which again agrees with
54 * our mappings for "CP932".
55 * • When converting Shift-JIS to CP932, the conversion goes through Unicode.
56 * Shift-JIS 0x7E converts to U+203E, so mapping U+203E to 0x7E means that
57 * 0x7E will go to 0x7E when converting Shift-JIS to CP932.
58 */
59
60 #include "mbfilter.h"
61 #include "mbfilter_cp932.h"
62
63 #include "unicode_table_cp932_ext.h"
64 #include "unicode_table_jis.h"
65
66 static int mbfl_filt_conv_cp932_wchar_flush(mbfl_convert_filter *filter);
67
68 static const unsigned char mblen_table_sjis[] = { /* 0x81-0x9f,0xE0-0xFF */
69 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
70 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
71 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
72 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
73 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
74 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
75 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
76 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
77 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
78 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
79 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
80 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
81 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
82 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
83 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2,
84 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2
85 };
86
87 static const char *mbfl_encoding_cp932_aliases[] = {"MS932", "Windows-31J", "MS_Kanji", NULL};
88 static const char *mbfl_encoding_sjiswin_aliases[] = {"SJIS-ms", "SJIS-open", NULL};
89
90 const mbfl_encoding mbfl_encoding_cp932 = {
91 mbfl_no_encoding_cp932,
92 "CP932",
93 "Shift_JIS",
94 mbfl_encoding_cp932_aliases,
95 mblen_table_sjis,
96 MBFL_ENCTYPE_GL_UNSAFE,
97 &vtbl_cp932_wchar,
98 &vtbl_wchar_cp932,
99 NULL
100 };
101
102 const struct mbfl_convert_vtbl vtbl_cp932_wchar = {
103 mbfl_no_encoding_cp932,
104 mbfl_no_encoding_wchar,
105 mbfl_filt_conv_common_ctor,
106 NULL,
107 mbfl_filt_conv_cp932_wchar,
108 mbfl_filt_conv_cp932_wchar_flush,
109 NULL,
110 };
111
112 const struct mbfl_convert_vtbl vtbl_wchar_cp932 = {
113 mbfl_no_encoding_wchar,
114 mbfl_no_encoding_cp932,
115 mbfl_filt_conv_common_ctor,
116 NULL,
117 mbfl_filt_conv_wchar_cp932,
118 mbfl_filt_conv_common_flush,
119 NULL,
120 };
121
122 const mbfl_encoding mbfl_encoding_sjiswin = {
123 mbfl_no_encoding_sjiswin,
124 "SJIS-win",
125 "Shift_JIS",
126 mbfl_encoding_sjiswin_aliases,
127 mblen_table_sjis,
128 MBFL_ENCTYPE_GL_UNSAFE,
129 &vtbl_sjiswin_wchar,
130 &vtbl_wchar_sjiswin,
131 NULL
132 };
133
134 const struct mbfl_convert_vtbl vtbl_sjiswin_wchar = {
135 mbfl_no_encoding_sjiswin,
136 mbfl_no_encoding_wchar,
137 mbfl_filt_conv_common_ctor,
138 NULL,
139 mbfl_filt_conv_cp932_wchar,
140 mbfl_filt_conv_cp932_wchar_flush,
141 NULL,
142 };
143
144 const struct mbfl_convert_vtbl vtbl_wchar_sjiswin = {
145 mbfl_no_encoding_wchar,
146 mbfl_no_encoding_sjiswin,
147 mbfl_filt_conv_common_ctor,
148 NULL,
149 mbfl_filt_conv_wchar_sjiswin,
150 mbfl_filt_conv_common_flush,
151 NULL,
152 };
153
154 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
155
156 #define SJIS_ENCODE(c1,c2,s1,s2) \
157 do { \
158 s1 = c1; \
159 s1--; \
160 s1 >>= 1; \
161 if ((c1) < 0x5f) { \
162 s1 += 0x71; \
163 } else { \
164 s1 += 0xb1; \
165 } \
166 s2 = c2; \
167 if ((c1) & 1) { \
168 if ((c2) < 0x60) { \
169 s2--; \
170 } \
171 s2 += 0x20; \
172 } else { \
173 s2 += 0x7e; \
174 } \
175 } while (0)
176
177 #define SJIS_DECODE(c1,c2,s1,s2) \
178 do { \
179 s1 = c1; \
180 if (s1 < 0xa0) { \
181 s1 -= 0x81; \
182 } else { \
183 s1 -= 0xc1; \
184 } \
185 s1 <<= 1; \
186 s1 += 0x21; \
187 s2 = c2; \
188 if (s2 < 0x9f) { \
189 if (s2 < 0x7f) { \
190 s2++; \
191 } \
192 s2 -= 0x20; \
193 } else { \
194 s1++; \
195 s2 -= 0x7e; \
196 } \
197 } while (0)
198
mbfl_filt_conv_cp932_wchar(int c,mbfl_convert_filter * filter)199 int mbfl_filt_conv_cp932_wchar(int c, mbfl_convert_filter *filter)
200 {
201 int c1, s, s1, s2, w;
202
203 switch (filter->status) {
204 case 0:
205 if (c >= 0 && c < 0x80) { /* latin */
206 CK((*filter->output_function)(c, filter->data));
207 } else if (c > 0xa0 && c < 0xe0) { /* kana */
208 CK((*filter->output_function)(0xfec0 + c, filter->data));
209 } else if (c > 0x80 && c < 0xfd && c != 0xa0) { /* kanji first char */
210 filter->status = 1;
211 filter->cache = c;
212 } else {
213 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
214 }
215 break;
216
217 case 1: /* kanji second char */
218 filter->status = 0;
219 c1 = filter->cache;
220 if (c >= 0x40 && c <= 0xfc && c != 0x7f) {
221 w = 0;
222 SJIS_DECODE(c1, c, s1, s2);
223 s = (s1 - 0x21)*94 + s2 - 0x21;
224 if (s <= 137) {
225 if (s == 31) {
226 w = 0xff3c; /* FULLWIDTH REVERSE SOLIDUS */
227 } else if (s == 32) {
228 w = 0xff5e; /* FULLWIDTH TILDE */
229 } else if (s == 33) {
230 w = 0x2225; /* PARALLEL TO */
231 } else if (s == 60) {
232 w = 0xff0d; /* FULLWIDTH HYPHEN-MINUS */
233 } else if (s == 80) {
234 w = 0xffe0; /* FULLWIDTH CENT SIGN */
235 } else if (s == 81) {
236 w = 0xffe1; /* FULLWIDTH POUND SIGN */
237 } else if (s == 137) {
238 w = 0xffe2; /* FULLWIDTH NOT SIGN */
239 }
240 }
241 if (w == 0) {
242 if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) { /* vendor ext1 (13ku) */
243 w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
244 } else if (s >= 0 && s < jisx0208_ucs_table_size) { /* X 0208 */
245 w = jisx0208_ucs_table[s];
246 } else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) { /* vendor ext2 (89ku - 92ku) */
247 w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
248 } else if (s >= cp932ext3_ucs_table_min && s < cp932ext3_ucs_table_max) { /* vendor ext3 (115ku - 119ku) */
249 w = cp932ext3_ucs_table[s - cp932ext3_ucs_table_min];
250 } else if (s >= (94*94) && s < (114*94)) { /* user (95ku - 114ku) */
251 w = s - (94*94) + 0xe000;
252 }
253 }
254
255 if (w <= 0) {
256 w = MBFL_BAD_INPUT;
257 }
258
259 CK((*filter->output_function)(w, filter->data));
260 } else {
261 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
262 }
263 break;
264
265 default:
266 filter->status = 0;
267 break;
268 }
269
270 return 0;
271 }
272
mbfl_filt_conv_cp932_wchar_flush(mbfl_convert_filter * filter)273 static int mbfl_filt_conv_cp932_wchar_flush(mbfl_convert_filter *filter)
274 {
275 if (filter->status) {
276 (*filter->output_function)(MBFL_BAD_INPUT, filter->data);
277 filter->status = 0;
278 }
279
280 if (filter->flush_function) {
281 (*filter->flush_function)(filter->data);
282 }
283
284 return 0;
285 }
286
mbfl_filt_conv_wchar_cp932(int c,mbfl_convert_filter * filter)287 int mbfl_filt_conv_wchar_cp932(int c, mbfl_convert_filter *filter)
288 {
289 int c1, c2, s1, s2;
290
291 s1 = 0;
292 s2 = 0;
293 if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
294 s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
295 } else if (c == 0x203E) {
296 s1 = 0x7E;
297 } else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
298 s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
299 } else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
300 s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
301 } else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
302 s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
303 } else if (c >= 0xe000 && c < (0xe000 + 20*94)) { /* user (95ku - 114ku) */
304 s1 = c - 0xe000;
305 c1 = s1/94 + 0x7f;
306 c2 = s1%94 + 0x21;
307 s1 = (c1 << 8) | c2;
308 s2 = 1;
309 }
310 if (s1 <= 0) {
311 if (c == 0xa5) { /* YEN SIGN */
312 s1 = 0x5C;
313 } else if (c == 0xff3c) { /* FULLWIDTH REVERSE SOLIDUS */
314 s1 = 0x2140;
315 } else if (c == 0x2225) { /* PARALLEL TO */
316 s1 = 0x2142;
317 } else if (c == 0xff0d) { /* FULLWIDTH HYPHEN-MINUS */
318 s1 = 0x215d;
319 } else if (c == 0xffe0) { /* FULLWIDTH CENT SIGN */
320 s1 = 0x2171;
321 } else if (c == 0xffe1) { /* FULLWIDTH POUND SIGN */
322 s1 = 0x2172;
323 } else if (c == 0xffe2) { /* FULLWIDTH NOT SIGN */
324 s1 = 0x224c;
325 }
326 }
327 if ((s1 <= 0) || (s1 >= 0x8080 && s2 == 0)) { /* not found or X 0212 */
328 s1 = -1;
329 c1 = 0;
330 c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
331 while (c1 < c2) { /* CP932 vendor ext1 (13ku) */
332 if (c == cp932ext1_ucs_table[c1]) {
333 s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
334 break;
335 }
336 c1++;
337 }
338 if (s1 <= 0) {
339 c1 = 0;
340 c2 = cp932ext3_ucs_table_max - cp932ext3_ucs_table_min;
341 while (c1 < c2) { /* CP932 vendor ext3 (115ku - 119ku) */
342 if (c == cp932ext3_ucs_table[c1]) {
343 s1 = ((c1/94 + 0x93) << 8) + (c1%94 + 0x21);
344 break;
345 }
346 c1++;
347 }
348 }
349 if (c == 0) {
350 s1 = 0;
351 } else if (s1 <= 0) {
352 s1 = -1;
353 }
354 }
355 if (s1 >= 0) {
356 if (s1 < 0x100) { /* latin or kana */
357 CK((*filter->output_function)(s1, filter->data));
358 } else { /* kanji */
359 c1 = (s1 >> 8) & 0xff;
360 c2 = s1 & 0xff;
361 SJIS_ENCODE(c1, c2, s1, s2);
362 CK((*filter->output_function)(s1, filter->data));
363 CK((*filter->output_function)(s2, filter->data));
364 }
365 } else {
366 CK(mbfl_filt_conv_illegal_output(c, filter));
367 }
368
369 return 0;
370 }
371
mbfl_filt_conv_wchar_sjiswin(int c,mbfl_convert_filter * filter)372 int mbfl_filt_conv_wchar_sjiswin(int c, mbfl_convert_filter *filter)
373 {
374 if (c == 0xA5) {
375 CK((*filter->output_function)(0x81, filter->data));
376 CK((*filter->output_function)(0x8F, filter->data));
377 } else if (c == 0x203E) {
378 CK((*filter->output_function)(0x81, filter->data));
379 CK((*filter->output_function)(0x50, filter->data));
380 } else {
381 return mbfl_filt_conv_wchar_cp932(c, filter);
382 }
383 return 0;
384 }
385
