mbfl_encoding.c 6.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243
  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. #include "libmbfl/config.h"
  31. #ifdef HAVE_STRINGS_H
  32. /* For strcasecmp */
  33. #include <strings.h>
  34. #endif
  35. #include "mbfl_encoding.h"
  36. #include "mbfilter_pass.h"
  37. #include "mbfilter_8bit.h"
  38. #include "filters/mbfilter_euc_cn.h"
  39. #include "filters/mbfilter_hz.h"
  40. #include "filters/mbfilter_euc_tw.h"
  41. #include "filters/mbfilter_big5.h"
  42. #include "filters/mbfilter_uhc.h"
  43. #include "filters/mbfilter_euc_kr.h"
  44. #include "filters/mbfilter_iso2022_kr.h"
  45. #include "filters/mbfilter_sjis.h"
  46. #include "filters/mbfilter_sjis_mobile.h"
  47. #include "filters/mbfilter_sjis_mac.h"
  48. #include "filters/mbfilter_sjis_2004.h"
  49. #include "filters/mbfilter_cp51932.h"
  50. #include "filters/mbfilter_jis.h"
  51. #include "filters/mbfilter_iso2022_jp_ms.h"
  52. #include "filters/mbfilter_iso2022jp_2004.h"
  53. #include "filters/mbfilter_iso2022jp_mobile.h"
  54. #include "filters/mbfilter_euc_jp.h"
  55. #include "filters/mbfilter_euc_jp_win.h"
  56. #include "filters/mbfilter_euc_jp_2004.h"
  57. #include "filters/mbfilter_gb18030.h"
  58. #include "filters/mbfilter_cp932.h"
  59. #include "filters/mbfilter_cp936.h"
  60. #include "filters/mbfilter_cp5022x.h"
  61. #include "filters/mbfilter_base64.h"
  62. #include "filters/mbfilter_qprint.h"
  63. #include "filters/mbfilter_uuencode.h"
  64. #include "filters/mbfilter_7bit.h"
  65. #include "filters/mbfilter_utf7.h"
  66. #include "filters/mbfilter_utf7imap.h"
  67. #include "filters/mbfilter_utf8.h"
  68. #include "filters/mbfilter_utf8_mobile.h"
  69. #include "filters/mbfilter_utf16.h"
  70. #include "filters/mbfilter_utf32.h"
  71. #include "filters/mbfilter_ucs4.h"
  72. #include "filters/mbfilter_ucs2.h"
  73. #include "filters/mbfilter_htmlent.h"
  74. #include "filters/mbfilter_singlebyte.h"
  75. #ifndef HAVE_STRCASECMP
  76. #ifdef HAVE_STRICMP
  77. #define strcasecmp stricmp
  78. #endif
  79. #endif
  80. static const mbfl_encoding *mbfl_encoding_ptr_list[] = {
  81. &mbfl_encoding_base64,
  82. &mbfl_encoding_uuencode,
  83. &mbfl_encoding_html_ent,
  84. &mbfl_encoding_qprint,
  85. &mbfl_encoding_7bit,
  86. &mbfl_encoding_8bit,
  87. &mbfl_encoding_ucs4,
  88. &mbfl_encoding_ucs4be,
  89. &mbfl_encoding_ucs4le,
  90. &mbfl_encoding_ucs2,
  91. &mbfl_encoding_ucs2be,
  92. &mbfl_encoding_ucs2le,
  93. &mbfl_encoding_utf32,
  94. &mbfl_encoding_utf32be,
  95. &mbfl_encoding_utf32le,
  96. &mbfl_encoding_utf16,
  97. &mbfl_encoding_utf16be,
  98. &mbfl_encoding_utf16le,
  99. &mbfl_encoding_utf8,
  100. &mbfl_encoding_utf7,
  101. &mbfl_encoding_utf7imap,
  102. &mbfl_encoding_ascii,
  103. &mbfl_encoding_euc_jp,
  104. &mbfl_encoding_sjis,
  105. &mbfl_encoding_eucjp_win,
  106. &mbfl_encoding_eucjp2004,
  107. &mbfl_encoding_sjis_docomo,
  108. &mbfl_encoding_sjis_kddi,
  109. &mbfl_encoding_sjis_sb,
  110. &mbfl_encoding_sjis_mac,
  111. &mbfl_encoding_sjis2004,
  112. &mbfl_encoding_utf8_docomo,
  113. &mbfl_encoding_utf8_kddi_a,
  114. &mbfl_encoding_utf8_kddi_b,
  115. &mbfl_encoding_utf8_sb,
  116. &mbfl_encoding_cp932,
  117. &mbfl_encoding_sjiswin,
  118. &mbfl_encoding_cp51932,
  119. &mbfl_encoding_jis,
  120. &mbfl_encoding_2022jp,
  121. &mbfl_encoding_2022jpms,
  122. &mbfl_encoding_gb18030,
  123. &mbfl_encoding_cp1252,
  124. &mbfl_encoding_cp1254,
  125. &mbfl_encoding_8859_1,
  126. &mbfl_encoding_8859_2,
  127. &mbfl_encoding_8859_3,
  128. &mbfl_encoding_8859_4,
  129. &mbfl_encoding_8859_5,
  130. &mbfl_encoding_8859_6,
  131. &mbfl_encoding_8859_7,
  132. &mbfl_encoding_8859_8,
  133. &mbfl_encoding_8859_9,
  134. &mbfl_encoding_8859_10,
  135. &mbfl_encoding_8859_13,
  136. &mbfl_encoding_8859_14,
  137. &mbfl_encoding_8859_15,
  138. &mbfl_encoding_8859_16,
  139. &mbfl_encoding_euc_cn,
  140. &mbfl_encoding_cp936,
  141. &mbfl_encoding_hz,
  142. &mbfl_encoding_euc_tw,
  143. &mbfl_encoding_big5,
  144. &mbfl_encoding_cp950,
  145. &mbfl_encoding_euc_kr,
  146. &mbfl_encoding_uhc,
  147. &mbfl_encoding_2022kr,
  148. &mbfl_encoding_cp1251,
  149. &mbfl_encoding_cp866,
  150. &mbfl_encoding_koi8r,
  151. &mbfl_encoding_koi8u,
  152. &mbfl_encoding_armscii8,
  153. &mbfl_encoding_cp850,
  154. &mbfl_encoding_2022jp_2004,
  155. &mbfl_encoding_2022jp_kddi,
  156. &mbfl_encoding_cp50220,
  157. &mbfl_encoding_cp50221,
  158. &mbfl_encoding_cp50222,
  159. NULL
  160. };
  161. const mbfl_encoding *mbfl_name2encoding(const char *name)
  162. {
  163. const mbfl_encoding **encoding;
  164. for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) {
  165. if (strcasecmp((*encoding)->name, name) == 0) {
  166. return *encoding;
  167. }
  168. }
  169. /* search MIME charset name */
  170. for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) {
  171. if ((*encoding)->mime_name) {
  172. if (strcasecmp((*encoding)->mime_name, name) == 0) {
  173. return *encoding;
  174. }
  175. }
  176. }
  177. /* search aliases */
  178. for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) {
  179. if ((*encoding)->aliases) {
  180. for (const char **alias = (*encoding)->aliases; *alias; alias++) {
  181. if (strcasecmp(*alias, name) == 0) {
  182. return *encoding;
  183. }
  184. }
  185. }
  186. }
  187. return NULL;
  188. }
  189. const mbfl_encoding *mbfl_no2encoding(enum mbfl_no_encoding no_encoding)
  190. {
  191. const mbfl_encoding **encoding;
  192. for (encoding = mbfl_encoding_ptr_list; *encoding; encoding++) {
  193. if ((*encoding)->no_encoding == no_encoding) {
  194. return *encoding;
  195. }
  196. }
  197. return NULL;
  198. }
  199. enum mbfl_no_encoding mbfl_name2no_encoding(const char *name)
  200. {
  201. const mbfl_encoding *encoding = mbfl_name2encoding(name);
  202. return encoding ? encoding->no_encoding : mbfl_no_encoding_invalid;
  203. }
  204. const char *mbfl_no_encoding2name(enum mbfl_no_encoding no_encoding)
  205. {
  206. const mbfl_encoding *encoding = mbfl_no2encoding(no_encoding);
  207. return encoding ? encoding->name : "";
  208. }
  209. const mbfl_encoding **mbfl_get_supported_encodings(void)
  210. {
  211. return mbfl_encoding_ptr_list;
  212. }
  213. const char *mbfl_no2preferred_mime_name(enum mbfl_no_encoding no_encoding)
  214. {
  215. return mbfl_encoding_preferred_mime_name(mbfl_no2encoding(no_encoding));
  216. }
  217. const char *mbfl_encoding_preferred_mime_name(const mbfl_encoding *encoding)
  218. {
  219. if (encoding->mime_name && encoding->mime_name[0] != '\0') {
  220. return encoding->mime_name;
  221. }
  222. return NULL;
  223. }