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.c
26 * by Moriyoshi Koizumi <moriyoshi@php.net> on 20 Dec 2002. The file
27 * mbfilter.c is included in this package .
28 *
29 */
30
31 #ifdef HAVE_CONFIG_H
32 #include "config.h"
33 #endif
34
35 #ifdef HAVE_STDDEF_H
36 #include <stddef.h>
37 #endif
38
39 #include "mbfl_ident.h"
40 #include "mbfl_allocators.h"
41 #include "mbfilter_pass.h"
42 #include "mbfilter_8bit.h"
43 #include "mbfilter_wchar.h"
44
45 #include "filters/mbfilter_euc_cn.h"
46 #include "filters/mbfilter_hz.h"
47 #include "filters/mbfilter_euc_tw.h"
48 #include "filters/mbfilter_big5.h"
49 #include "filters/mbfilter_uhc.h"
50 #include "filters/mbfilter_euc_kr.h"
51 #include "filters/mbfilter_iso2022_kr.h"
52 #include "filters/mbfilter_sjis.h"
53 #include "filters/mbfilter_sjis_open.h"
54 #include "filters/mbfilter_jis.h"
55 #include "filters/mbfilter_iso2022_jp_ms.h"
56 #include "filters/mbfilter_euc_jp.h"
57 #include "filters/mbfilter_euc_jp_win.h"
58 #include "filters/mbfilter_ascii.h"
59 #include "filters/mbfilter_koi8r.h"
60 #include "filters/mbfilter_koi8u.h"
61 #include "filters/mbfilter_cp866.h"
62 #include "filters/mbfilter_cp932.h"
63 #include "filters/mbfilter_cp936.h"
64 #include "filters/mbfilter_cp1251.h"
65 #include "filters/mbfilter_cp1252.h"
66 #include "filters/mbfilter_cp1254.h"
67 #include "filters/mbfilter_cp51932.h"
68 #include "filters/mbfilter_cp5022x.h"
69 #include "filters/mbfilter_iso8859_1.h"
70 #include "filters/mbfilter_iso8859_2.h"
71 #include "filters/mbfilter_iso8859_3.h"
72 #include "filters/mbfilter_iso8859_4.h"
73 #include "filters/mbfilter_iso8859_5.h"
74 #include "filters/mbfilter_iso8859_6.h"
75 #include "filters/mbfilter_iso8859_7.h"
76 #include "filters/mbfilter_iso8859_8.h"
77 #include "filters/mbfilter_iso8859_9.h"
78 #include "filters/mbfilter_iso8859_10.h"
79 #include "filters/mbfilter_iso8859_13.h"
80 #include "filters/mbfilter_iso8859_14.h"
81 #include "filters/mbfilter_iso8859_15.h"
82 #include "filters/mbfilter_base64.h"
83 #include "filters/mbfilter_qprint.h"
84 #include "filters/mbfilter_uuencode.h"
85 #include "filters/mbfilter_7bit.h"
86 #include "filters/mbfilter_utf7.h"
87 #include "filters/mbfilter_utf7imap.h"
88 #include "filters/mbfilter_utf8.h"
89 #include "filters/mbfilter_utf16.h"
90 #include "filters/mbfilter_utf32.h"
91 #include "filters/mbfilter_byte2.h"
92 #include "filters/mbfilter_byte4.h"
93 #include "filters/mbfilter_ucs4.h"
94 #include "filters/mbfilter_ucs2.h"
95 #include "filters/mbfilter_htmlent.h"
96 #include "filters/mbfilter_armscii8.h"
97 #include "filters/mbfilter_cp850.h"
98
99 static const struct mbfl_identify_vtbl vtbl_identify_false = {
100 mbfl_no_encoding_pass,
101 mbfl_filt_ident_false_ctor,
102 mbfl_filt_ident_common_dtor,
103 mbfl_filt_ident_false };
104
105
106 static const struct mbfl_identify_vtbl *mbfl_identify_filter_list[] = {
107 &vtbl_identify_utf8,
108 &vtbl_identify_utf7,
109 &vtbl_identify_ascii,
110 &vtbl_identify_eucjp,
111 &vtbl_identify_sjis,
112 &vtbl_identify_sjis_open,
113 &vtbl_identify_eucjpwin,
114 &vtbl_identify_cp932,
115 &vtbl_identify_jis,
116 &vtbl_identify_2022jp,
117 &vtbl_identify_2022jpms,
118 &vtbl_identify_cp51932,
119 &vtbl_identify_euccn,
120 &vtbl_identify_cp936,
121 &vtbl_identify_hz,
122 &vtbl_identify_euctw,
123 &vtbl_identify_big5,
124 &vtbl_identify_euckr,
125 &vtbl_identify_uhc,
126 &vtbl_identify_2022kr,
127 &vtbl_identify_cp1251,
128 &vtbl_identify_cp866,
129 &vtbl_identify_koi8r,
130 &vtbl_identify_koi8u,
131 &vtbl_identify_cp1252,
132 &vtbl_identify_cp1254,
133 &vtbl_identify_8859_1,
134 &vtbl_identify_8859_2,
135 &vtbl_identify_8859_3,
136 &vtbl_identify_8859_4,
137 &vtbl_identify_8859_5,
138 &vtbl_identify_8859_6,
139 &vtbl_identify_8859_7,
140 &vtbl_identify_8859_8,
141 &vtbl_identify_8859_9,
142 &vtbl_identify_8859_10,
143 &vtbl_identify_8859_13,
144 &vtbl_identify_8859_14,
145 &vtbl_identify_8859_15,
146 &vtbl_identify_armscii8,
147 &vtbl_identify_cp850,
148 &vtbl_identify_jis_ms,
149 &vtbl_identify_cp50220,
150 &vtbl_identify_cp50221,
151 &vtbl_identify_cp50222,
152 &vtbl_identify_false,
153 NULL
154 };
155
156
157
158 /*
159 * identify filter
160 */
mbfl_identify_filter_get_vtbl(enum mbfl_no_encoding encoding)161 const struct mbfl_identify_vtbl * mbfl_identify_filter_get_vtbl(enum mbfl_no_encoding encoding)
162 {
163 const struct mbfl_identify_vtbl * vtbl;
164 int i;
165
166 i = 0;
167 while ((vtbl = mbfl_identify_filter_list[i++]) != NULL) {
168 if (vtbl->encoding == encoding) {
169 break;
170 }
171 }
172
173 return vtbl;
174 }
175
mbfl_identify_filter_new(enum mbfl_no_encoding encoding)176 mbfl_identify_filter *mbfl_identify_filter_new(enum mbfl_no_encoding encoding)
177 {
178 mbfl_identify_filter *filter;
179
180 /* allocate */
181 filter = (mbfl_identify_filter *)mbfl_malloc(sizeof(mbfl_identify_filter));
182 if (filter == NULL) {
183 return NULL;
184 }
185
186 if (mbfl_identify_filter_init(filter, encoding)) {
187 mbfl_free(filter);
188 return NULL;
189 }
190
191 return filter;
192 }
193
mbfl_identify_filter_init(mbfl_identify_filter * filter,enum mbfl_no_encoding encoding)194 int mbfl_identify_filter_init(mbfl_identify_filter *filter, enum mbfl_no_encoding encoding)
195 {
196 const struct mbfl_identify_vtbl *vtbl;
197
198 /* encoding structure */
199 filter->encoding = mbfl_no2encoding(encoding);
200 if (filter->encoding == NULL) {
201 filter->encoding = &mbfl_encoding_pass;
202 }
203
204 filter->status = 0;
205 filter->flag = 0;
206 filter->score = 0;
207
208 /* setup the function table */
209 vtbl = mbfl_identify_filter_get_vtbl(filter->encoding->no_encoding);
210 if (vtbl == NULL) {
211 vtbl = &vtbl_identify_false;
212 }
213 filter->filter_ctor = vtbl->filter_ctor;
214 filter->filter_dtor = vtbl->filter_dtor;
215 filter->filter_function = vtbl->filter_function;
216
217 /* constructor */
218 (*filter->filter_ctor)(filter);
219
220 return 0;
221 }
222
mbfl_identify_filter_delete(mbfl_identify_filter * filter)223 void mbfl_identify_filter_delete(mbfl_identify_filter *filter)
224 {
225 if (filter == NULL) {
226 return;
227 }
228
229 mbfl_identify_filter_cleanup(filter);
230 mbfl_free((void*)filter);
231 }
232
mbfl_identify_filter_cleanup(mbfl_identify_filter * filter)233 void mbfl_identify_filter_cleanup(mbfl_identify_filter *filter)
234 {
235 (*filter->filter_dtor)(filter);
236 }
237
mbfl_filt_ident_common_ctor(mbfl_identify_filter * filter)238 void mbfl_filt_ident_common_ctor(mbfl_identify_filter *filter)
239 {
240 filter->status = 0;
241 filter->flag = 0;
242 }
243
mbfl_filt_ident_common_dtor(mbfl_identify_filter * filter)244 void mbfl_filt_ident_common_dtor(mbfl_identify_filter *filter)
245 {
246 filter->status = 0;
247 }
248
mbfl_filt_ident_false(int c,mbfl_identify_filter * filter)249 int mbfl_filt_ident_false(int c, mbfl_identify_filter *filter)
250 {
251 filter->flag = 1; /* bad */
252 return c;
253 }
254
mbfl_filt_ident_false_ctor(mbfl_identify_filter * filter)255 void mbfl_filt_ident_false_ctor(mbfl_identify_filter *filter)
256 {
257 filter->status = 0;
258 filter->flag = 1;
259 }
260
mbfl_filt_ident_true(int c,mbfl_identify_filter * filter)261 int mbfl_filt_ident_true(int c, mbfl_identify_filter *filter)
262 {
263 return c;
264 }
265
