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 };
80 
81 const struct mbfl_convert_vtbl vtbl_cp51932_wchar = {
82 	mbfl_no_encoding_cp51932,
83 	mbfl_no_encoding_wchar,
84 	mbfl_filt_conv_common_ctor,
85 	mbfl_filt_conv_common_dtor,
86 	mbfl_filt_conv_cp51932_wchar,
87 	mbfl_filt_conv_common_flush
88 };
89 
90 const struct mbfl_convert_vtbl vtbl_wchar_cp51932 = {
91 	mbfl_no_encoding_wchar,
92 	mbfl_no_encoding_cp51932,
93 	mbfl_filt_conv_common_ctor,
94 	mbfl_filt_conv_common_dtor,
95 	mbfl_filt_conv_wchar_cp51932,
96 	mbfl_filt_conv_common_flush
97 };
98 
99 #define CK(statement)	do { if ((statement) < 0) return (-1); } while (0)
100 
101 #define sjistoidx(c1, c2) \
102         (((c1) > 0x9f) \
103         ? (((c1) - 0xc1) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)) \
104         : (((c1) - 0x81) * 188 + (c2) - (((c2) > 0x7e) ? 0x41 : 0x40)))
105 #define idxtoeuc1(c) (((c) / 94) + 0xa1)
106 #define idxtoeuc2(c) (((c) % 94) + 0xa1)
107 
108 /*
109  * cp51932 => wchar
110  */
111 int
mbfl_filt_conv_cp51932_wchar(int c,mbfl_convert_filter * filter)112 mbfl_filt_conv_cp51932_wchar(int c, mbfl_convert_filter *filter)
113 {
114 	int c1, s, w;
115 
116 	switch (filter->status) {
117 	case 0:
118 		if (c >= 0 && c < 0x80) {	/* latin */
119 			CK((*filter->output_function)(c, filter->data));
120 		} else if (c > 0xa0 && c < 0xff) {	/* CP932 first char */
121 			filter->status = 1;
122 			filter->cache = c;
123 		} else if (c == 0x8e) {	/* kana first char */
124 			filter->status = 2;
125 		} else {
126 			w = c & MBFL_WCSGROUP_MASK;
127 			w |= MBFL_WCSGROUP_THROUGH;
128 			CK((*filter->output_function)(w, filter->data));
129 		}
130 		break;
131 
132 	case 1:	/* got first half */
133 		filter->status = 0;
134 		c1 = filter->cache;
135 		if (c > 0xa0 && c < 0xff) {
136 			w = 0;
137 			s = (c1 - 0xa1)*94 + c - 0xa1;
138 			if (s <= 137) {
139 				if (s == 31) {
140 					w = 0xff3c;			/* FULLWIDTH REVERSE SOLIDUS */
141 				} else if (s == 32) {
142 					w = 0xff5e;			/* FULLWIDTH TILDE */
143 				} else if (s == 33) {
144 					w = 0x2225;			/* PARALLEL TO */
145 				} else if (s == 60) {
146 					w = 0xff0d;			/* FULLWIDTH HYPHEN-MINUS */
147 				} else if (s == 80) {
148 					w = 0xffe0;			/* FULLWIDTH CENT SIGN */
149 				} else if (s == 81) {
150 					w = 0xffe1;			/* FULLWIDTH POUND SIGN */
151 				} else if (s == 137) {
152 					w = 0xffe2;			/* FULLWIDTH NOT SIGN */
153 				}
154 			}
155 			if (w == 0) {
156 				if (s >= cp932ext1_ucs_table_min && s < cp932ext1_ucs_table_max) {		/* vendor ext1 (13ku) */
157 					w = cp932ext1_ucs_table[s - cp932ext1_ucs_table_min];
158 				} else if (s >= 0 && s < jisx0208_ucs_table_size) {		/* X 0208 */
159 					w = jisx0208_ucs_table[s];
160 				} else if (s >= cp932ext2_ucs_table_min && s < cp932ext2_ucs_table_max) {		/* vendor ext2 (89ku - 92ku) */
161 					w = cp932ext2_ucs_table[s - cp932ext2_ucs_table_min];
162 				}
163 			}
164 			if (w <= 0) {
165 				w = ((c1 & 0x7f) << 8) | (c & 0x7f);
166 				w &= MBFL_WCSPLANE_MASK;
167 				w |= MBFL_WCSPLANE_WINCP932;
168 			}
169 			CK((*filter->output_function)(w, filter->data));
170 		} else if ((c >= 0 && c < 0x21) || c == 0x7f) {		/* CTLs */
171 			CK((*filter->output_function)(c, filter->data));
172 		} else {
173 			w = (c1 << 8) | c;
174 			w &= MBFL_WCSGROUP_MASK;
175 			w |= MBFL_WCSGROUP_THROUGH;
176 			CK((*filter->output_function)(w, filter->data));
177 		}
178 		break;
179 
180 	case 2:	/* got 0x8e, X0201 kana */
181 		filter->status = 0;
182 		if (c > 0xa0 && c < 0xe0) {
183 			w = 0xfec0 + c;
184 			CK((*filter->output_function)(w, filter->data));
185 		} else if ((c >= 0 && c < 0x21) || c == 0x7f) {		/* CTLs */
186 			CK((*filter->output_function)(c, filter->data));
187 		} else {
188 			w = 0x8e00 | c;
189 			w &= MBFL_WCSGROUP_MASK;
190 			w |= MBFL_WCSGROUP_THROUGH;
191 			CK((*filter->output_function)(w, filter->data));
192 		}
193 		break;
194 
195 	default:
196 		filter->status = 0;
197 		break;
198 	}
199 
200 	return c;
201 }
202 
203 /*
204  * wchar => cp51932
205  */
206 int
mbfl_filt_conv_wchar_cp51932(int c,mbfl_convert_filter * filter)207 mbfl_filt_conv_wchar_cp51932(int c, mbfl_convert_filter *filter)
208 {
209 	int c1, c2, s1;
210 
211 	s1 = 0;
212 	if (c >= ucs_a1_jis_table_min && c < ucs_a1_jis_table_max) {
213 		s1 = ucs_a1_jis_table[c - ucs_a1_jis_table_min];
214 	} else if (c >= ucs_a2_jis_table_min && c < ucs_a2_jis_table_max) {
215 		s1 = ucs_a2_jis_table[c - ucs_a2_jis_table_min];
216 	} else if (c >= ucs_i_jis_table_min && c < ucs_i_jis_table_max) {
217 		s1 = ucs_i_jis_table[c - ucs_i_jis_table_min];
218 	} else if (c >= ucs_r_jis_table_min && c < ucs_r_jis_table_max) {
219 		s1 = ucs_r_jis_table[c - ucs_r_jis_table_min];
220 	}
221 	if (s1 >= 0x8080) s1 = -1; /* we don't support JIS X0213 */
222 	if (s1 <= 0) {
223 		c1 = c & ~MBFL_WCSPLANE_MASK;
224 		if (c1 == MBFL_WCSPLANE_WINCP932) {
225 			s1 = c & MBFL_WCSPLANE_MASK;
226 			if (s1 >= ((85 + 0x20) << 8)) {	/* 85ku - 120ku */
227 				s1 = -1;
228 			}
229 		} else if (c1 == MBFL_WCSPLANE_JIS0208) {
230 			s1 = c & MBFL_WCSPLANE_MASK;
231 			if ((s1 >= ((85 + 0x20) << 8) &&  /* 85ku - 94ku */
232                              s1 <= ((88 + 0x20) << 8)) ||/* IBM extension */
233 			    (s1 >= ((93 + 0x20) << 8) && /* 89ku - 92ku */
234                              s1 <= ((94 + 0x20) << 8))) {
235 				s1 = -1;
236 			}
237 		} else if (c == 0xa5) {		/* YEN SIGN */
238 			s1 = 0x005c;			/* YEN SIGN */
239 		} else if (c == 0x203e) {	/* OVER LINE */
240 			s1 = 0x007e;			/* FULLWIDTH MACRON */
241 		} else if (c == 0xff3c) {	/* FULLWIDTH REVERSE SOLIDUS */
242 			s1 = 0x2140;
243 		} else if (c == 0xff5e) {	/* FULLWIDTH TILDE */
244 			s1 = 0x2141;
245 		} else if (c == 0x2225) {	/* PARALLEL TO */
246 			s1 = 0x2142;
247 		} else if (c == 0xff0d) {	/* FULLWIDTH HYPHEN-MINUS */
248 			s1 = 0x215d;
249 		} else if (c == 0xffe0) {	/* FULLWIDTH CENT SIGN */
250 			s1 = 0x2171;
251 		} else if (c == 0xffe1) {	/* FULLWIDTH POUND SIGN */
252 			s1 = 0x2172;
253 		} else if (c == 0xffe2) {	/* FULLWIDTH NOT SIGN */
254 			s1 = 0x224c;
255 		} else {
256 			s1 = -1;
257 			c1 = 0;
258 			c2 = cp932ext1_ucs_table_max - cp932ext1_ucs_table_min;
259 			while (c1 < c2) {		/* CP932 vendor ext1 (13ku) */
260 				if (c == cp932ext1_ucs_table[c1]) {
261 					s1 = ((c1/94 + 0x2d) << 8) + (c1%94 + 0x21);
262 					break;
263 				}
264 				c1++;
265 			}
266 			if (s1 < 0) {
267 				c1 = 0;
268 				c2 = cp932ext2_ucs_table_max - cp932ext2_ucs_table_min;
269 				while (c1 < c2) {		/* CP932 vendor ext3 (115ku - 119ku) */
270 					if (c == cp932ext2_ucs_table[c1]) {
271 					  s1 = ((c1/94 + 0x79) << 8) +(c1%94 + 0x21);
272 					  break;
273 					}
274 					c1++;
275 				}
276 			}
277 		}
278 		if (c == 0) {
279 			s1 = 0;
280 		} else if (s1 <= 0) {
281 			s1 = -1;
282 		}
283 	}
284 
285 	if (s1 >= 0) {
286 		if (s1 < 0x80) {	/* latin */
287 			CK((*filter->output_function)(s1, filter->data));
288 		} else if (s1 < 0x100) {	/* kana */
289 			CK((*filter->output_function)(0x8e, filter->data));
290 			CK((*filter->output_function)(s1, filter->data));
291 		} else if (s1 < 0x8080)  {	/* X 0208 */
292 			CK((*filter->output_function)(((s1 >> 8) & 0xff) | 0x80, filter->data));
293 			CK((*filter->output_function)((s1 & 0xff) | 0x80, filter->data));
294 		} else {
295 		  if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
296 		    CK(mbfl_filt_conv_illegal_output(c, filter));
297 		  }
298 		}
299 	} else {
300 		if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
301 			CK(mbfl_filt_conv_illegal_output(c, filter));
302 		}
303 	}
304 
305 	return c;
306 }
307 
mbfl_filt_ident_cp51932(int c,mbfl_identify_filter * filter)308 static int mbfl_filt_ident_cp51932(int c, mbfl_identify_filter *filter)
309 {
310 	switch (filter->status) {
311 	case  0:	/* latin */
312 		if (c >= 0 && c < 0x80) {	/* ok */
313 			;
314 		} else if (c > 0xa0 && c < 0xff) {	/* kanji first char */
315 			filter->status = 1;
316 		} else if (c == 0x8e) {				/* kana first char */
317 			filter->status = 2;
318 		} else {							/* bad */
319 			filter->flag = 1;
320 		}
321 		break;
322 
323 	case  1:	/* got first half */
324 		if (c < 0xa1 || c > 0xfe) {		/* bad */
325 			filter->flag = 1;
326 		}
327 		filter->status = 0;
328 		break;
329 
330 	case  2:	/* got 0x8e */
331 		if (c < 0xa1 || c > 0xdf) {		/* bad */
332 			filter->flag = 1;
333 		}
334 		filter->status = 0;
335 		break;
336 
337 	default:
338 		filter->status = 0;
339 		break;
340 	}
341 
342 	return c;
343 }
344