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 : #ifdef HAVE_STRING_H
40 : #include <string.h>
41 : #endif
42 :
43 : #ifdef HAVE_STRINGS_H
44 : #include <strings.h>
45 : #endif
46 :
47 : #include "mbfl_encoding.h"
48 : #include "mbfilter_pass.h"
49 : #include "mbfilter_8bit.h"
50 : #include "mbfilter_wchar.h"
51 :
52 : #include "filters/mbfilter_euc_cn.h"
53 : #include "filters/mbfilter_hz.h"
54 : #include "filters/mbfilter_euc_tw.h"
55 : #include "filters/mbfilter_big5.h"
56 : #include "filters/mbfilter_uhc.h"
57 : #include "filters/mbfilter_euc_kr.h"
58 : #include "filters/mbfilter_iso2022_kr.h"
59 : #include "filters/mbfilter_sjis.h"
60 : #include "filters/mbfilter_cp51932.h"
61 : #include "filters/mbfilter_jis.h"
62 : #include "filters/mbfilter_iso2022_jp_ms.h"
63 : #include "filters/mbfilter_euc_jp.h"
64 : #include "filters/mbfilter_euc_jp_win.h"
65 : #include "filters/mbfilter_ascii.h"
66 : #include "filters/mbfilter_koi8r.h"
67 : #include "filters/mbfilter_cp866.h"
68 : #include "filters/mbfilter_cp932.h"
69 : #include "filters/mbfilter_cp936.h"
70 : #include "filters/mbfilter_cp1251.h"
71 : #include "filters/mbfilter_cp1252.h"
72 : #include "filters/mbfilter_iso8859_1.h"
73 : #include "filters/mbfilter_iso8859_2.h"
74 : #include "filters/mbfilter_iso8859_3.h"
75 : #include "filters/mbfilter_iso8859_4.h"
76 : #include "filters/mbfilter_iso8859_5.h"
77 : #include "filters/mbfilter_iso8859_6.h"
78 : #include "filters/mbfilter_iso8859_7.h"
79 : #include "filters/mbfilter_iso8859_8.h"
80 : #include "filters/mbfilter_iso8859_9.h"
81 : #include "filters/mbfilter_iso8859_10.h"
82 : #include "filters/mbfilter_iso8859_13.h"
83 : #include "filters/mbfilter_iso8859_14.h"
84 : #include "filters/mbfilter_iso8859_15.h"
85 : #include "filters/mbfilter_iso8859_16.h"
86 : #include "filters/mbfilter_base64.h"
87 : #include "filters/mbfilter_qprint.h"
88 : #include "filters/mbfilter_uuencode.h"
89 : #include "filters/mbfilter_7bit.h"
90 : #include "filters/mbfilter_utf7.h"
91 : #include "filters/mbfilter_utf7imap.h"
92 : #include "filters/mbfilter_utf8.h"
93 : #include "filters/mbfilter_utf16.h"
94 : #include "filters/mbfilter_utf32.h"
95 : #include "filters/mbfilter_byte2.h"
96 : #include "filters/mbfilter_byte4.h"
97 : #include "filters/mbfilter_ucs4.h"
98 : #include "filters/mbfilter_ucs2.h"
99 : #include "filters/mbfilter_htmlent.h"
100 : #include "filters/mbfilter_armscii8.h"
101 :
102 : #ifndef HAVE_STRCASECMP
103 : #ifdef HAVE_STRICMP
104 : #define strcasecmp stricmp
105 : #endif
106 : #endif
107 :
108 :
109 : static const char *mbfl_encoding_auto_aliases[] = {"unknown", NULL};
110 :
111 : static const mbfl_encoding mbfl_encoding_auto = {
112 : mbfl_no_encoding_auto,
113 : "auto",
114 : NULL,
115 : (const char *(*)[])&mbfl_encoding_auto_aliases,
116 : NULL,
117 : 0
118 : };
119 :
120 : static const mbfl_encoding *mbfl_encoding_ptr_list[] = {
121 : &mbfl_encoding_pass,
122 : &mbfl_encoding_auto,
123 : &mbfl_encoding_wchar,
124 : &mbfl_encoding_byte2be,
125 : &mbfl_encoding_byte2le,
126 : &mbfl_encoding_byte4be,
127 : &mbfl_encoding_byte4le,
128 : &mbfl_encoding_base64,
129 : &mbfl_encoding_uuencode,
130 : &mbfl_encoding_html_ent,
131 : &mbfl_encoding_qprint,
132 : &mbfl_encoding_7bit,
133 : &mbfl_encoding_8bit,
134 : &mbfl_encoding_ucs4,
135 : &mbfl_encoding_ucs4be,
136 : &mbfl_encoding_ucs4le,
137 : &mbfl_encoding_ucs2,
138 : &mbfl_encoding_ucs2be,
139 : &mbfl_encoding_ucs2le,
140 : &mbfl_encoding_utf32,
141 : &mbfl_encoding_utf32be,
142 : &mbfl_encoding_utf32le,
143 : &mbfl_encoding_utf16,
144 : &mbfl_encoding_utf16be,
145 : &mbfl_encoding_utf16le,
146 : &mbfl_encoding_utf8,
147 : &mbfl_encoding_utf7,
148 : &mbfl_encoding_utf7imap,
149 : &mbfl_encoding_ascii,
150 : &mbfl_encoding_euc_jp,
151 : &mbfl_encoding_sjis,
152 : &mbfl_encoding_eucjp_win,
153 : &mbfl_encoding_sjis_win,
154 : &mbfl_encoding_cp51932,
155 : &mbfl_encoding_jis,
156 : &mbfl_encoding_2022jp,
157 : &mbfl_encoding_2022jpms,
158 : &mbfl_encoding_cp1252,
159 : &mbfl_encoding_8859_1,
160 : &mbfl_encoding_8859_2,
161 : &mbfl_encoding_8859_3,
162 : &mbfl_encoding_8859_4,
163 : &mbfl_encoding_8859_5,
164 : &mbfl_encoding_8859_6,
165 : &mbfl_encoding_8859_7,
166 : &mbfl_encoding_8859_8,
167 : &mbfl_encoding_8859_9,
168 : &mbfl_encoding_8859_10,
169 : &mbfl_encoding_8859_13,
170 : &mbfl_encoding_8859_14,
171 : &mbfl_encoding_8859_15,
172 : &mbfl_encoding_8859_16,
173 : &mbfl_encoding_euc_cn,
174 : &mbfl_encoding_cp936,
175 : &mbfl_encoding_hz,
176 : &mbfl_encoding_euc_tw,
177 : &mbfl_encoding_big5,
178 : &mbfl_encoding_euc_kr,
179 : &mbfl_encoding_uhc,
180 : &mbfl_encoding_2022kr,
181 : &mbfl_encoding_cp1251,
182 : &mbfl_encoding_cp866,
183 : &mbfl_encoding_koi8r,
184 : &mbfl_encoding_armscii8,
185 : NULL
186 : };
187 :
188 : /* encoding resolver */
189 : const mbfl_encoding *
190 : mbfl_name2encoding(const char *name)
191 153262 : {
192 : const mbfl_encoding *encoding;
193 : int i, j;
194 :
195 153262 : if (name == NULL) {
196 8 : return NULL;
197 : }
198 :
199 153254 : i = 0;
200 4293854 : while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){
201 4139670 : if (strcasecmp(encoding->name, name) == 0) {
202 152324 : return encoding;
203 : }
204 : }
205 :
206 : /* serch MIME charset name */
207 930 : i = 0;
208 50382 : while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) {
209 48847 : if (encoding->mime_name != NULL) {
210 41407 : if (strcasecmp(encoding->mime_name, name) == 0) {
211 325 : return encoding;
212 : }
213 : }
214 : }
215 :
216 : /* serch aliases */
217 605 : i = 0;
218 31181 : while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) {
219 30249 : if (encoding->aliases != NULL) {
220 18658 : j = 0;
221 78113 : while ((*encoding->aliases)[j] != NULL) {
222 41075 : if (strcasecmp((*encoding->aliases)[j], name) == 0) {
223 278 : return encoding;
224 : }
225 40797 : j++;
226 : }
227 : }
228 : }
229 :
230 327 : return NULL;
231 : }
232 :
233 : const mbfl_encoding *
234 : mbfl_no2encoding(enum mbfl_no_encoding no_encoding)
235 44813 : {
236 : const mbfl_encoding *encoding;
237 : int i;
238 :
239 44813 : i = 0;
240 701038 : while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){
241 656225 : if (encoding->no_encoding == no_encoding) {
242 44813 : return encoding;
243 : }
244 : }
245 :
246 0 : return NULL;
247 : }
248 :
249 : enum mbfl_no_encoding
250 : mbfl_name2no_encoding(const char *name)
251 153119 : {
252 : const mbfl_encoding *encoding;
253 :
254 153119 : encoding = mbfl_name2encoding(name);
255 153119 : if (encoding == NULL) {
256 335 : return mbfl_no_encoding_invalid;
257 : } else {
258 152784 : return encoding->no_encoding;
259 : }
260 : }
261 :
262 : const char *
263 : mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding)
264 190 : {
265 : const mbfl_encoding *encoding;
266 :
267 190 : encoding = mbfl_no2encoding(no_encoding);
268 190 : if (encoding == NULL) {
269 0 : return "";
270 : } else {
271 190 : return encoding->name;
272 : }
273 : }
274 :
275 : const mbfl_encoding **
276 : mbfl_get_supported_encodings(void)
277 0 : {
278 0 : return mbfl_encoding_ptr_list;
279 : }
280 :
281 : const char *
282 : mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding)
283 913 : {
284 : const mbfl_encoding *encoding;
285 :
286 913 : encoding = mbfl_no2encoding(no_encoding);
287 913 : if (encoding != NULL && encoding->mime_name != NULL && encoding->mime_name[0] != '\0') {
288 913 : return encoding->mime_name;
289 : } else {
290 0 : return NULL;
291 : }
292 : }
293 :
294 : int
295 : mbfl_is_support_encoding(const char *name)
296 0 : {
297 : const mbfl_encoding *encoding;
298 :
299 0 : encoding = mbfl_name2encoding(name);
300 0 : if (encoding == NULL) {
301 0 : return 0;
302 : } else {
303 0 : return 1;
304 : }
305 : }
|