mbfl_encoding.c 8.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329
  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. #ifdef HAVE_CONFIG_H
  31. #include "config.h"
  32. #endif
  33. #ifdef HAVE_STDDEF_H
  34. #include <stddef.h>
  35. #endif
  36. #ifdef HAVE_STRING_H
  37. #include <string.h>
  38. #endif
  39. #ifdef HAVE_STRINGS_H
  40. #include <strings.h>
  41. #endif
  42. #include "mbfl_encoding.h"
  43. #include "mbfilter_pass.h"
  44. #include "mbfilter_8bit.h"
  45. #include "mbfilter_wchar.h"
  46. #include "filters/mbfilter_euc_cn.h"
  47. #include "filters/mbfilter_hz.h"
  48. #include "filters/mbfilter_euc_tw.h"
  49. #include "filters/mbfilter_big5.h"
  50. #include "filters/mbfilter_uhc.h"
  51. #include "filters/mbfilter_euc_kr.h"
  52. #include "filters/mbfilter_iso2022_kr.h"
  53. #include "filters/mbfilter_sjis.h"
  54. #include "filters/mbfilter_sjis_open.h"
  55. #include "filters/mbfilter_sjis_mobile.h"
  56. #include "filters/mbfilter_sjis_mac.h"
  57. #include "filters/mbfilter_sjis_2004.h"
  58. #include "filters/mbfilter_cp51932.h"
  59. #include "filters/mbfilter_jis.h"
  60. #include "filters/mbfilter_iso2022_jp_ms.h"
  61. #include "filters/mbfilter_iso2022jp_2004.h"
  62. #include "filters/mbfilter_iso2022jp_mobile.h"
  63. #include "filters/mbfilter_euc_jp.h"
  64. #include "filters/mbfilter_euc_jp_win.h"
  65. #include "filters/mbfilter_euc_jp_2004.h"
  66. #include "filters/mbfilter_gb18030.h"
  67. #include "filters/mbfilter_ascii.h"
  68. #include "filters/mbfilter_koi8r.h"
  69. #include "filters/mbfilter_koi8u.h"
  70. #include "filters/mbfilter_cp866.h"
  71. #include "filters/mbfilter_cp932.h"
  72. #include "filters/mbfilter_cp936.h"
  73. #include "filters/mbfilter_cp1251.h"
  74. #include "filters/mbfilter_cp1252.h"
  75. #include "filters/mbfilter_cp1254.h"
  76. #include "filters/mbfilter_cp5022x.h"
  77. #include "filters/mbfilter_iso8859_1.h"
  78. #include "filters/mbfilter_iso8859_2.h"
  79. #include "filters/mbfilter_iso8859_3.h"
  80. #include "filters/mbfilter_iso8859_4.h"
  81. #include "filters/mbfilter_iso8859_5.h"
  82. #include "filters/mbfilter_iso8859_6.h"
  83. #include "filters/mbfilter_iso8859_7.h"
  84. #include "filters/mbfilter_iso8859_8.h"
  85. #include "filters/mbfilter_iso8859_9.h"
  86. #include "filters/mbfilter_iso8859_10.h"
  87. #include "filters/mbfilter_iso8859_13.h"
  88. #include "filters/mbfilter_iso8859_14.h"
  89. #include "filters/mbfilter_iso8859_15.h"
  90. #include "filters/mbfilter_iso8859_16.h"
  91. #include "filters/mbfilter_base64.h"
  92. #include "filters/mbfilter_qprint.h"
  93. #include "filters/mbfilter_uuencode.h"
  94. #include "filters/mbfilter_7bit.h"
  95. #include "filters/mbfilter_utf7.h"
  96. #include "filters/mbfilter_utf7imap.h"
  97. #include "filters/mbfilter_utf8.h"
  98. #include "filters/mbfilter_utf8_mobile.h"
  99. #include "filters/mbfilter_utf16.h"
  100. #include "filters/mbfilter_utf32.h"
  101. #include "filters/mbfilter_byte2.h"
  102. #include "filters/mbfilter_byte4.h"
  103. #include "filters/mbfilter_ucs4.h"
  104. #include "filters/mbfilter_ucs2.h"
  105. #include "filters/mbfilter_htmlent.h"
  106. #include "filters/mbfilter_armscii8.h"
  107. #include "filters/mbfilter_cp850.h"
  108. #ifndef HAVE_STRCASECMP
  109. #ifdef HAVE_STRICMP
  110. #define strcasecmp stricmp
  111. #endif
  112. #endif
  113. static const mbfl_encoding *mbfl_encoding_ptr_list[] = {
  114. &mbfl_encoding_pass,
  115. &mbfl_encoding_wchar,
  116. &mbfl_encoding_byte2be,
  117. &mbfl_encoding_byte2le,
  118. &mbfl_encoding_byte4be,
  119. &mbfl_encoding_byte4le,
  120. &mbfl_encoding_base64,
  121. &mbfl_encoding_uuencode,
  122. &mbfl_encoding_html_ent,
  123. &mbfl_encoding_qprint,
  124. &mbfl_encoding_7bit,
  125. &mbfl_encoding_8bit,
  126. &mbfl_encoding_ucs4,
  127. &mbfl_encoding_ucs4be,
  128. &mbfl_encoding_ucs4le,
  129. &mbfl_encoding_ucs2,
  130. &mbfl_encoding_ucs2be,
  131. &mbfl_encoding_ucs2le,
  132. &mbfl_encoding_utf32,
  133. &mbfl_encoding_utf32be,
  134. &mbfl_encoding_utf32le,
  135. &mbfl_encoding_utf16,
  136. &mbfl_encoding_utf16be,
  137. &mbfl_encoding_utf16le,
  138. &mbfl_encoding_utf8,
  139. &mbfl_encoding_utf7,
  140. &mbfl_encoding_utf7imap,
  141. &mbfl_encoding_ascii,
  142. &mbfl_encoding_euc_jp,
  143. &mbfl_encoding_sjis,
  144. &mbfl_encoding_eucjp_win,
  145. &mbfl_encoding_eucjp2004,
  146. &mbfl_encoding_sjis_open,
  147. &mbfl_encoding_sjis_docomo,
  148. &mbfl_encoding_sjis_kddi,
  149. &mbfl_encoding_sjis_sb,
  150. &mbfl_encoding_sjis_mac,
  151. &mbfl_encoding_sjis2004,
  152. &mbfl_encoding_utf8_docomo,
  153. &mbfl_encoding_utf8_kddi_a,
  154. &mbfl_encoding_utf8_kddi_b,
  155. &mbfl_encoding_utf8_sb,
  156. &mbfl_encoding_cp932,
  157. &mbfl_encoding_cp51932,
  158. &mbfl_encoding_jis,
  159. &mbfl_encoding_2022jp,
  160. &mbfl_encoding_2022jpms,
  161. &mbfl_encoding_gb18030,
  162. &mbfl_encoding_cp1252,
  163. &mbfl_encoding_cp1254,
  164. &mbfl_encoding_8859_1,
  165. &mbfl_encoding_8859_2,
  166. &mbfl_encoding_8859_3,
  167. &mbfl_encoding_8859_4,
  168. &mbfl_encoding_8859_5,
  169. &mbfl_encoding_8859_6,
  170. &mbfl_encoding_8859_7,
  171. &mbfl_encoding_8859_8,
  172. &mbfl_encoding_8859_9,
  173. &mbfl_encoding_8859_10,
  174. &mbfl_encoding_8859_13,
  175. &mbfl_encoding_8859_14,
  176. &mbfl_encoding_8859_15,
  177. &mbfl_encoding_8859_16,
  178. &mbfl_encoding_euc_cn,
  179. &mbfl_encoding_cp936,
  180. &mbfl_encoding_hz,
  181. &mbfl_encoding_euc_tw,
  182. &mbfl_encoding_big5,
  183. &mbfl_encoding_cp950,
  184. &mbfl_encoding_euc_kr,
  185. &mbfl_encoding_uhc,
  186. &mbfl_encoding_2022kr,
  187. &mbfl_encoding_cp1251,
  188. &mbfl_encoding_cp866,
  189. &mbfl_encoding_koi8r,
  190. &mbfl_encoding_koi8u,
  191. &mbfl_encoding_armscii8,
  192. &mbfl_encoding_cp850,
  193. &mbfl_encoding_jis_ms,
  194. &mbfl_encoding_2022jp_2004,
  195. &mbfl_encoding_2022jp_kddi,
  196. &mbfl_encoding_cp50220,
  197. &mbfl_encoding_cp50220raw,
  198. &mbfl_encoding_cp50221,
  199. &mbfl_encoding_cp50222,
  200. NULL
  201. };
  202. /* encoding resolver */
  203. const mbfl_encoding *
  204. mbfl_name2encoding(const char *name)
  205. {
  206. const mbfl_encoding *encoding;
  207. int i, j;
  208. if (name == NULL) {
  209. return NULL;
  210. }
  211. i = 0;
  212. while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){
  213. if (strcasecmp(encoding->name, name) == 0) {
  214. return encoding;
  215. }
  216. }
  217. /* serch MIME charset name */
  218. i = 0;
  219. while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) {
  220. if (encoding->mime_name != NULL) {
  221. if (strcasecmp(encoding->mime_name, name) == 0) {
  222. return encoding;
  223. }
  224. }
  225. }
  226. /* serch aliases */
  227. i = 0;
  228. while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL) {
  229. if (encoding->aliases != NULL) {
  230. j = 0;
  231. while ((*encoding->aliases)[j] != NULL) {
  232. if (strcasecmp((*encoding->aliases)[j], name) == 0) {
  233. return encoding;
  234. }
  235. j++;
  236. }
  237. }
  238. }
  239. return NULL;
  240. }
  241. const mbfl_encoding *
  242. mbfl_no2encoding(enum mbfl_no_encoding no_encoding)
  243. {
  244. const mbfl_encoding *encoding;
  245. int i;
  246. i = 0;
  247. while ((encoding = mbfl_encoding_ptr_list[i++]) != NULL){
  248. if (encoding->no_encoding == no_encoding) {
  249. return encoding;
  250. }
  251. }
  252. return NULL;
  253. }
  254. enum mbfl_no_encoding
  255. mbfl_name2no_encoding(const char *name)
  256. {
  257. const mbfl_encoding *encoding;
  258. encoding = mbfl_name2encoding(name);
  259. if (encoding == NULL) {
  260. return mbfl_no_encoding_invalid;
  261. } else {
  262. return encoding->no_encoding;
  263. }
  264. }
  265. const char *
  266. mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding)
  267. {
  268. const mbfl_encoding *encoding;
  269. encoding = mbfl_no2encoding(no_encoding);
  270. if (encoding == NULL) {
  271. return "";
  272. } else {
  273. return encoding->name;
  274. }
  275. }
  276. const mbfl_encoding **
  277. mbfl_get_supported_encodings(void)
  278. {
  279. return mbfl_encoding_ptr_list;
  280. }
  281. const char *
  282. mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding)
  283. {
  284. const mbfl_encoding *encoding;
  285. encoding = mbfl_no2encoding(no_encoding);
  286. if (encoding != NULL && encoding->mime_name != NULL && encoding->mime_name[0] != '\0') {
  287. return encoding->mime_name;
  288. } else {
  289. return NULL;
  290. }
  291. }
  292. int
  293. mbfl_is_support_encoding(const char *name)
  294. {
  295. const mbfl_encoding *encoding;
  296. encoding = mbfl_name2encoding(name);
  297. if (encoding == NULL) {
  298. return 0;
  299. } else {
  300. return 1;
  301. }
  302. }