eSpeak NG is an open source speech synthesizer that supports more than hundred languages and accents.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

encoding.c 30KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592
  1. /*
  2. * Copyright (C) 2017 Reece H. Dunn
  3. *
  4. * This program is free software; you can redistribute it and/or modify
  5. * it under the terms of the GNU General Public License as published by
  6. * the Free Software Foundation; either version 3 of the License, or
  7. * (at your option) any later version.
  8. *
  9. * This program is distributed in the hope that it will be useful,
  10. * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11. * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  12. * GNU General Public License for more details.
  13. *
  14. * You should have received a copy of the GNU General Public License
  15. * along with this program; if not, see: <http://www.gnu.org/licenses/>.
  16. */
  17. #include "config.h"
  18. #include <stdint.h>
  19. #include <stdlib.h>
  20. #include <espeak-ng/espeak_ng.h>
  21. #include "speech.h"
  22. #include "encoding.h"
  23. // http://www.iana.org/assignments/character-sets/character-sets.xhtml
  24. MNEM_TAB mnem_encoding[] = {
  25. { "ANSI_X3.4-1968", ESPEAKNG_ENCODING_US_ASCII },
  26. { "ANSI_X3.4-1986", ESPEAKNG_ENCODING_US_ASCII },
  27. { "ASMO-708", ESPEAKNG_ENCODING_ISO_8859_6 },
  28. { "ECMA-114", ESPEAKNG_ENCODING_ISO_8859_6 },
  29. { "ECMA-118", ESPEAKNG_ENCODING_ISO_8859_7 },
  30. { "ELOT_928", ESPEAKNG_ENCODING_ISO_8859_7 },
  31. { "IBM367", ESPEAKNG_ENCODING_US_ASCII },
  32. { "IBM819", ESPEAKNG_ENCODING_ISO_8859_1 },
  33. { "ISCII", ESPEAKNG_ENCODING_ISCII },
  34. { "ISO_646.irv:1991", ESPEAKNG_ENCODING_US_ASCII },
  35. { "ISO_8859-1", ESPEAKNG_ENCODING_ISO_8859_1 },
  36. { "ISO_8859-1:1987", ESPEAKNG_ENCODING_ISO_8859_1 },
  37. { "ISO_8859-2", ESPEAKNG_ENCODING_ISO_8859_2 },
  38. { "ISO_8859-2:1987", ESPEAKNG_ENCODING_ISO_8859_2 },
  39. { "ISO_8859-3", ESPEAKNG_ENCODING_ISO_8859_3 },
  40. { "ISO_8859-3:1988", ESPEAKNG_ENCODING_ISO_8859_3 },
  41. { "ISO_8859-4", ESPEAKNG_ENCODING_ISO_8859_4 },
  42. { "ISO_8859-4:1988", ESPEAKNG_ENCODING_ISO_8859_4 },
  43. { "ISO_8859-5", ESPEAKNG_ENCODING_ISO_8859_5 },
  44. { "ISO_8859-5:1988", ESPEAKNG_ENCODING_ISO_8859_5 },
  45. { "ISO_8859-6", ESPEAKNG_ENCODING_ISO_8859_6 },
  46. { "ISO_8859-6:1987", ESPEAKNG_ENCODING_ISO_8859_6 },
  47. { "ISO_8859-7", ESPEAKNG_ENCODING_ISO_8859_7 },
  48. { "ISO_8859-7:1987", ESPEAKNG_ENCODING_ISO_8859_7 },
  49. { "ISO_8859-8", ESPEAKNG_ENCODING_ISO_8859_8 },
  50. { "ISO_8859-8:1988", ESPEAKNG_ENCODING_ISO_8859_8 },
  51. { "ISO_8859-9", ESPEAKNG_ENCODING_ISO_8859_9 },
  52. { "ISO_8859-9:1989", ESPEAKNG_ENCODING_ISO_8859_9 },
  53. { "ISO_8859-10", ESPEAKNG_ENCODING_ISO_8859_10 },
  54. { "ISO_8859-10:1992", ESPEAKNG_ENCODING_ISO_8859_10 },
  55. { "ISO_8859-14", ESPEAKNG_ENCODING_ISO_8859_14 },
  56. { "ISO_8859-14:1998", ESPEAKNG_ENCODING_ISO_8859_14 },
  57. { "ISO_8859-15", ESPEAKNG_ENCODING_ISO_8859_15 },
  58. { "ISO_8859-16", ESPEAKNG_ENCODING_ISO_8859_16 },
  59. { "ISO_8859-16:2001", ESPEAKNG_ENCODING_ISO_8859_16 },
  60. { "ISO646-US", ESPEAKNG_ENCODING_US_ASCII },
  61. { "ISO-8859-1", ESPEAKNG_ENCODING_ISO_8859_1 },
  62. { "ISO-8859-2", ESPEAKNG_ENCODING_ISO_8859_2 },
  63. { "ISO-8859-3", ESPEAKNG_ENCODING_ISO_8859_3 },
  64. { "ISO-8859-4", ESPEAKNG_ENCODING_ISO_8859_4 },
  65. { "ISO-8859-5", ESPEAKNG_ENCODING_ISO_8859_5 },
  66. { "ISO-8859-6", ESPEAKNG_ENCODING_ISO_8859_6 },
  67. { "ISO-8859-7", ESPEAKNG_ENCODING_ISO_8859_7 },
  68. { "ISO-8859-8", ESPEAKNG_ENCODING_ISO_8859_8 },
  69. { "ISO-8859-9", ESPEAKNG_ENCODING_ISO_8859_9 },
  70. { "ISO-8859-10", ESPEAKNG_ENCODING_ISO_8859_10 },
  71. { "ISO-8859-11", ESPEAKNG_ENCODING_ISO_8859_11 },
  72. { "ISO-8859-13", ESPEAKNG_ENCODING_ISO_8859_13 },
  73. { "ISO-8859-14", ESPEAKNG_ENCODING_ISO_8859_14 },
  74. { "ISO-8859-15", ESPEAKNG_ENCODING_ISO_8859_15 },
  75. { "ISO-8859-16", ESPEAKNG_ENCODING_ISO_8859_16 },
  76. { "KOI8-R", ESPEAKNG_ENCODING_KOI8_R },
  77. { "Latin-9", ESPEAKNG_ENCODING_ISO_8859_15 },
  78. { "TIS-620", ESPEAKNG_ENCODING_ISO_8859_11 },
  79. { "US-ASCII", ESPEAKNG_ENCODING_US_ASCII },
  80. { "cp367", ESPEAKNG_ENCODING_US_ASCII },
  81. { "cp819", ESPEAKNG_ENCODING_ISO_8859_1 },
  82. { "csASCII", ESPEAKNG_ENCODING_US_ASCII },
  83. { "csISO885913", ESPEAKNG_ENCODING_ISO_8859_13 },
  84. { "csISO885914", ESPEAKNG_ENCODING_ISO_8859_14 },
  85. { "csISO885915", ESPEAKNG_ENCODING_ISO_8859_15 },
  86. { "csISO885916", ESPEAKNG_ENCODING_ISO_8859_16 },
  87. { "csISOLatin1", ESPEAKNG_ENCODING_ISO_8859_1 },
  88. { "csISOLatin2", ESPEAKNG_ENCODING_ISO_8859_2 },
  89. { "csISOLatin3", ESPEAKNG_ENCODING_ISO_8859_3 },
  90. { "csISOLatin4", ESPEAKNG_ENCODING_ISO_8859_4 },
  91. { "csISOLatin5", ESPEAKNG_ENCODING_ISO_8859_9 },
  92. { "csISOLatin6", ESPEAKNG_ENCODING_ISO_8859_10 },
  93. { "csISOLatinArabic", ESPEAKNG_ENCODING_ISO_8859_6 },
  94. { "csISOLatinCyrillic",ESPEAKNG_ENCODING_ISO_8859_5 },
  95. { "csISOLatinGreek", ESPEAKNG_ENCODING_ISO_8859_7 },
  96. { "csISOLatinHebrew", ESPEAKNG_ENCODING_ISO_8859_8 },
  97. { "csKOI8R", ESPEAKNG_ENCODING_KOI8_R },
  98. { "csTIS620", ESPEAKNG_ENCODING_ISO_8859_11 },
  99. { "arabic", ESPEAKNG_ENCODING_ISO_8859_6 },
  100. { "cyrillic", ESPEAKNG_ENCODING_ISO_8859_5 },
  101. { "greek", ESPEAKNG_ENCODING_ISO_8859_7 },
  102. { "greek8", ESPEAKNG_ENCODING_ISO_8859_7 },
  103. { "hebrew", ESPEAKNG_ENCODING_ISO_8859_8 },
  104. { "iso-celtic", ESPEAKNG_ENCODING_ISO_8859_14 },
  105. { "iso-ir-6", ESPEAKNG_ENCODING_US_ASCII },
  106. { "iso-ir-100", ESPEAKNG_ENCODING_ISO_8859_1 },
  107. { "iso-ir-101", ESPEAKNG_ENCODING_ISO_8859_2 },
  108. { "iso-ir-109", ESPEAKNG_ENCODING_ISO_8859_3 },
  109. { "iso-ir-110", ESPEAKNG_ENCODING_ISO_8859_4 },
  110. { "iso-ir-126", ESPEAKNG_ENCODING_ISO_8859_7 },
  111. { "iso-ir-127", ESPEAKNG_ENCODING_ISO_8859_6 },
  112. { "iso-ir-138", ESPEAKNG_ENCODING_ISO_8859_8 },
  113. { "iso-ir-144", ESPEAKNG_ENCODING_ISO_8859_5 },
  114. { "iso-ir-148", ESPEAKNG_ENCODING_ISO_8859_9 },
  115. { "iso-ir-157", ESPEAKNG_ENCODING_ISO_8859_10 },
  116. { "iso-ir-199", ESPEAKNG_ENCODING_ISO_8859_14 },
  117. { "iso-ir-226", ESPEAKNG_ENCODING_ISO_8859_16 },
  118. { "latin1", ESPEAKNG_ENCODING_ISO_8859_1 },
  119. { "latin2", ESPEAKNG_ENCODING_ISO_8859_2 },
  120. { "latin3", ESPEAKNG_ENCODING_ISO_8859_3 },
  121. { "latin4", ESPEAKNG_ENCODING_ISO_8859_4 },
  122. { "latin5", ESPEAKNG_ENCODING_ISO_8859_9 },
  123. { "latin6", ESPEAKNG_ENCODING_ISO_8859_10 },
  124. { "latin8", ESPEAKNG_ENCODING_ISO_8859_14 },
  125. { "latin10", ESPEAKNG_ENCODING_ISO_8859_16 },
  126. { "l1", ESPEAKNG_ENCODING_ISO_8859_1 },
  127. { "l2", ESPEAKNG_ENCODING_ISO_8859_2 },
  128. { "l3", ESPEAKNG_ENCODING_ISO_8859_3 },
  129. { "l4", ESPEAKNG_ENCODING_ISO_8859_4 },
  130. { "l5", ESPEAKNG_ENCODING_ISO_8859_9 },
  131. { "l6", ESPEAKNG_ENCODING_ISO_8859_10 },
  132. { "l8", ESPEAKNG_ENCODING_ISO_8859_14 },
  133. { "l10", ESPEAKNG_ENCODING_ISO_8859_16 },
  134. { "us", ESPEAKNG_ENCODING_US_ASCII },
  135. { NULL, ESPEAKNG_ENCODING_UNKNOWN }
  136. };
  137. #pragma GCC visibility push(default)
  138. espeak_ng_ENCODING
  139. espeak_ng_EncodingFromName(const char *encoding)
  140. {
  141. return LookupMnem(mnem_encoding, encoding);
  142. }
  143. #pragma GCC visibility pop
  144. struct espeak_ng_TEXT_DECODER_
  145. {
  146. const char *current;
  147. const char *end;
  148. uint32_t (*get)(espeak_ng_TEXT_DECODER *decoder);
  149. const uint32_t *codepage;
  150. };
  151. // Reference: http://www.iana.org/go/rfc1345
  152. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-2.TXT
  153. static const uint32_t ISO_8859_2[0x80] = {
  154. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  155. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  156. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  157. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  158. 0x00a0, 0x0104, 0x02d8, 0x0141, 0x00a4, 0x013d, 0x015a, 0x00a7, // a0
  159. 0x00a8, 0x0160, 0x015e, 0x0164, 0x0179, 0x00ad, 0x017d, 0x017b, // a8
  160. 0x00b0, 0x0105, 0x02db, 0x0142, 0x00b4, 0x013e, 0x015b, 0x02c7, // b0
  161. 0x00b8, 0x0161, 0x015f, 0x0165, 0x017a, 0x02dd, 0x017e, 0x017c, // b8
  162. 0x0154, 0x00c1, 0x00c2, 0x0102, 0x00c4, 0x0139, 0x0106, 0x00c7, // c0
  163. 0x010c, 0x00c9, 0x0118, 0x00cb, 0x011a, 0x00cd, 0x00ce, 0x010e, // c8
  164. 0x0110, 0x0143, 0x0147, 0x00d3, 0x00d4, 0x0150, 0x00d6, 0x00d7, // d0
  165. 0x0158, 0x016e, 0x00da, 0x0170, 0x00dc, 0x00dd, 0x0162, 0x00df, // d8
  166. 0x0155, 0x00e1, 0x00e2, 0x0103, 0x00e4, 0x013a, 0x0107, 0x00e7, // e0
  167. 0x010d, 0x00e9, 0x0119, 0x00eb, 0x011b, 0x00ed, 0x00ee, 0x010f, // e8
  168. 0x0111, 0x0144, 0x0148, 0x00f3, 0x00f4, 0x0151, 0x00f6, 0x00f7, // f0
  169. 0x0159, 0x016f, 0x00fa, 0x0171, 0x00fc, 0x00fd, 0x0163, 0x02d9, // f8
  170. };
  171. // Reference: http://www.iana.org/go/rfc1345
  172. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-3.TXT
  173. static const uint32_t ISO_8859_3[0x80] = {
  174. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  175. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  176. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  177. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  178. 0x00a0, 0x0126, 0x02d8, 0x00a3, 0x00a4, 0xfffd, 0x0124, 0x00a7, // a0
  179. 0x00a8, 0x0130, 0x015e, 0x011e, 0x0134, 0x00ad, 0xfffd, 0x017b, // a8
  180. 0x00b0, 0x0127, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x0125, 0x00b7, // b0
  181. 0x00b8, 0x0131, 0x015f, 0x011f, 0x0135, 0x00bd, 0xfffd, 0x017c, // b8
  182. 0x00c0, 0x00c1, 0x00c2, 0xfffd, 0x00c4, 0x010a, 0x0108, 0x00c7, // c0
  183. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  184. 0xfffd, 0x00d1, 0x00d2, 0x00d3, 0x00d4, 0x0120, 0x00d6, 0x00d7, // d0
  185. 0x011c, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x016c, 0x015c, 0x00df, // d8
  186. 0x00e0, 0x00e1, 0x00e2, 0xfffd, 0x00e4, 0x010b, 0x0109, 0x00e7, // e0
  187. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  188. 0xfffd, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x0121, 0x00f6, 0x00f7, // f0
  189. 0x011d, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x016d, 0x015d, 0x02d9, // f8
  190. };
  191. // Reference: http://www.iana.org/go/rfc1345
  192. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-4.TXT
  193. static const uint32_t ISO_8859_4[0x80] = {
  194. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  195. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  196. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  197. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  198. 0x00a0, 0x0104, 0x0138, 0x0156, 0x00a4, 0x0128, 0x013b, 0x00a7, // a0
  199. 0x00a8, 0x0160, 0x0112, 0x0122, 0x0166, 0x00ad, 0x017d, 0x00af, // a8
  200. 0x00b0, 0x0105, 0x02db, 0x0157, 0x00b4, 0x0129, 0x013c, 0x02c7, // b0
  201. 0x00b8, 0x0161, 0x0113, 0x0123, 0x0167, 0x014a, 0x017e, 0x014b, // b8
  202. 0x0100, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x012e, // c0
  203. 0x010c, 0x00c9, 0x0118, 0x00cb, 0x0116, 0x00cd, 0x00ce, 0x012a, // c8
  204. 0x0110, 0x0145, 0x014c, 0x0136, 0x00d4, 0x00d5, 0x00d6, 0x00d7, // d0
  205. 0x00d8, 0x0172, 0x00da, 0x00db, 0x00dc, 0x0168, 0x016a, 0x00df, // d8
  206. 0x0101, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x012f, // e0
  207. 0x010d, 0x00e9, 0x0119, 0x00eb, 0x0117, 0x00ed, 0x00ee, 0x012b, // e8
  208. 0x0111, 0x0146, 0x014d, 0x0137, 0x00f4, 0x00f5, 0x00f6, 0x00f7, // f0
  209. 0x00f8, 0x0173, 0x00fa, 0x00fb, 0x00fc, 0x0169, 0x016b, 0x02d9, // f8
  210. };
  211. // Reference: http://www.iana.org/go/rfc1345
  212. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-5.TXT
  213. static const uint32_t ISO_8859_5[0x80] = {
  214. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  215. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  216. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  217. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  218. 0x00a0, 0x0401, 0x0402, 0x0403, 0x0404, 0x0405, 0x0406, 0x0407, // a0
  219. 0x0408, 0x0409, 0x040a, 0x040b, 0x040c, 0x00ad, 0x040e, 0x040f, // a8
  220. 0x0410, 0x0411, 0x0412, 0x0413, 0x0414, 0x0415, 0x0416, 0x0417, // b0
  221. 0x0418, 0x0419, 0x041a, 0x041b, 0x041c, 0x041d, 0x041e, 0x041f, // b8
  222. 0x0420, 0x0421, 0x0422, 0x0423, 0x0424, 0x0425, 0x0426, 0x0427, // c0
  223. 0x0428, 0x0429, 0x042a, 0x042b, 0x042c, 0x042d, 0x042e, 0x042f, // c8
  224. 0x0430, 0x0431, 0x0432, 0x0433, 0x0434, 0x0435, 0x0436, 0x0437, // d0
  225. 0x0438, 0x0439, 0x043a, 0x043b, 0x043c, 0x043d, 0x043e, 0x043f, // d8
  226. 0x0440, 0x0441, 0x0442, 0x0443, 0x0444, 0x0445, 0x0446, 0x0447, // e0
  227. 0x0448, 0x0449, 0x044a, 0x044b, 0x044c, 0x044d, 0x044e, 0x044f, // e8
  228. 0x2116, 0x0451, 0x0452, 0x0453, 0x0454, 0x0455, 0x0456, 0x0457, // f0
  229. 0x0458, 0x0459, 0x045a, 0x045b, 0x045c, 0x00a7, 0x045e, 0x045f, // f8
  230. };
  231. // Reference: http://www.iana.org/go/rfc1345
  232. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-6.TXT
  233. static const uint32_t ISO_8859_6[0x80] = {
  234. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  235. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  236. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  237. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  238. 0x00a0, 0xfffd, 0xfffd, 0xfffd, 0x00a4, 0xfffd, 0xfffd, 0xfffd, // a0
  239. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x060c, 0x00ad, 0xfffd, 0xfffd, // a8
  240. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // b0
  241. 0xfffd, 0xfffd, 0xfffd, 0x061b, 0xfffd, 0xfffd, 0xfffd, 0x061f, // b8
  242. 0xfffd, 0x0621, 0x0622, 0x0623, 0x0624, 0x0625, 0x0626, 0x0627, // c0
  243. 0x0628, 0x0629, 0x062a, 0x062b, 0x062c, 0x062d, 0x062e, 0x062f, // c8
  244. 0x0630, 0x0631, 0x0632, 0x0633, 0x0634, 0x0635, 0x0636, 0x0637, // d0
  245. 0x0638, 0x0639, 0x063a, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // d8
  246. 0x0640, 0x0641, 0x0642, 0x0643, 0x0644, 0x0645, 0x0646, 0x0647, // e0
  247. 0x0648, 0x0649, 0x064a, 0x064b, 0x064c, 0x064d, 0x064e, 0x064f, // e8
  248. 0x0650, 0x0651, 0x0652, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // f0
  249. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // f8
  250. };
  251. // Reference: http://www.iana.org/go/rfc1345
  252. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-7.TXT
  253. static const uint32_t ISO_8859_7[0x80] = {
  254. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  255. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  256. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  257. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  258. 0x00a0, 0x2018, 0x2019, 0x00a3, 0x20ac, 0x20af, 0x00a6, 0x00a7, // a0
  259. 0x00a8, 0x00a9, 0x037a, 0x00ab, 0x00ac, 0x00ad, 0xfffd, 0x2015, // a8
  260. 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x0384, 0x0385, 0x0386, 0x00b7, // b0
  261. 0x0388, 0x0389, 0x038a, 0x00bb, 0x038c, 0x00bd, 0x038e, 0x038f, // b8
  262. 0x0390, 0x0391, 0x0392, 0x0393, 0x0394, 0x0395, 0x0396, 0x0397, // c0
  263. 0x0398, 0x0399, 0x039a, 0x039b, 0x039c, 0x039d, 0x039e, 0x039f, // c8
  264. 0x03a0, 0x03a1, 0xfffd, 0x03a3, 0x03a4, 0x03a5, 0x03a6, 0x03a7, // d0
  265. 0x03a8, 0x03a9, 0x03aa, 0x03ab, 0x03ac, 0x03ad, 0x03ae, 0x03af, // d8
  266. 0x03b0, 0x03b1, 0x03b2, 0x03b3, 0x03b4, 0x03b5, 0x03b6, 0x03b7, // e0
  267. 0x03b8, 0x03b9, 0x03ba, 0x03bb, 0x03bc, 0x03bd, 0x03be, 0x03bf, // e8
  268. 0x03c0, 0x03c1, 0x03c2, 0x03c3, 0x03c4, 0x03c5, 0x03c6, 0x03c7, // f0
  269. 0x03c8, 0x03c9, 0x03ca, 0x03cb, 0x03cc, 0x03cd, 0x03ce, 0xfffd, // f8
  270. };
  271. // Reference: http://www.iana.org/go/rfc1345
  272. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-8.TXT
  273. static const uint32_t ISO_8859_8[0x80] = {
  274. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  275. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  276. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  277. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  278. 0x00a0, 0xfffd, 0x00a2, 0x00a3, 0x00a4, 0x00a5, 0x00a6, 0x00a7, // a0
  279. 0x00a8, 0x00a9, 0x00d7, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, // a8
  280. 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7, // b0
  281. 0x00b8, 0x00b9, 0x00f7, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0xfffd, // b8
  282. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // c0
  283. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // c8
  284. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // d0
  285. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x2017, // d8
  286. 0x05d0, 0x05d1, 0x05d2, 0x05d3, 0x05d4, 0x05d5, 0x05d6, 0x05d7, // e0
  287. 0x05d8, 0x05d9, 0x05da, 0x05db, 0x05dc, 0x05dd, 0x05de, 0x05df, // e8
  288. 0x05e0, 0x05e1, 0x05e2, 0x05e3, 0x05e4, 0x05e5, 0x05e6, 0x05e7, // f0
  289. 0x05e8, 0x05e9, 0x05ea, 0xfffd, 0xfffd, 0x200e, 0x200f, 0xfffd, // f8
  290. };
  291. // Reference: http://www.iana.org/go/rfc1345
  292. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-9.TXT
  293. static const uint32_t ISO_8859_9[0x80] = {
  294. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  295. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  296. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  297. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  298. 0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x00a4, 0x00a5, 0x00a6, 0x00a7, // a0
  299. 0x00a8, 0x00a9, 0x00aa, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, // a8
  300. 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7, // b0
  301. 0x00b8, 0x00b9, 0x00ba, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf, // b8
  302. 0x00c0, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x00c7, // c0
  303. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  304. 0x011e, 0x00d1, 0x00d2, 0x00d3, 0x00d4, 0x00d5, 0x00d6, 0x00d7, // d0
  305. 0x00d8, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x0130, 0x015e, 0x00df, // d8
  306. 0x00e0, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x00e7, // e0
  307. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  308. 0x011f, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x00f7, // f0
  309. 0x00f8, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x0131, 0x015f, 0x00ff, // f8
  310. };
  311. // Reference: http://www.iana.org/go/rfc1345
  312. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-10.TXT
  313. static const uint32_t ISO_8859_10[0x80] = {
  314. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  315. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  316. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  317. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  318. 0x00a0, 0x0104, 0x0112, 0x0122, 0x012a, 0x0128, 0x0136, 0x00a7, // a0
  319. 0x013b, 0x0110, 0x0160, 0x0166, 0x017d, 0x00ad, 0x016a, 0x014a, // a8
  320. 0x00b0, 0x0105, 0x0113, 0x0123, 0x012b, 0x0129, 0x0137, 0x00b7, // b0
  321. 0x013c, 0x0111, 0x0161, 0x0167, 0x017e, 0x2015, 0x016b, 0x014b, // b8
  322. 0x0100, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x012e, // c0
  323. 0x010c, 0x00c9, 0x0118, 0x00cb, 0x0116, 0x00cd, 0x00ce, 0x00cf, // c8
  324. 0x00d0, 0x0145, 0x014c, 0x00d3, 0x00d4, 0x00d5, 0x00d6, 0x0168, // d0
  325. 0x00d8, 0x0172, 0x00da, 0x00db, 0x00dc, 0x00dd, 0x00de, 0x00df, // d8
  326. 0x0101, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x012f, // e0
  327. 0x010d, 0x00e9, 0x0119, 0x00eb, 0x0117, 0x00ed, 0x00ee, 0x00ef, // e8
  328. 0x00f0, 0x0146, 0x014d, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x0169, // f0
  329. 0x00f8, 0x0173, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x00fe, 0x0138, // f8
  330. };
  331. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-11.TXT
  332. static const uint32_t ISO_8859_11[0x80] = {
  333. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  334. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  335. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  336. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  337. 0x00a0, 0x0e01, 0x0e02, 0x0e03, 0x0e04, 0x0e05, 0x0e06, 0x0e07, // a0
  338. 0x0e08, 0x0e09, 0x0e0a, 0x0e0b, 0x0e0c, 0x0e0d, 0x0e0e, 0x0e0f, // a8
  339. 0x0e10, 0x0e11, 0x0e12, 0x0e13, 0x0e14, 0x0e15, 0x0e16, 0x0e17, // b0
  340. 0x0e18, 0x0e19, 0x0e1a, 0x0e1b, 0x0e1c, 0x0e1d, 0x0e1e, 0x0e1f, // b8
  341. 0x0e20, 0x0e21, 0x0e22, 0x0e23, 0x0e24, 0x0e25, 0x0e26, 0x0e27, // c0
  342. 0x0e28, 0x0e29, 0x0e2a, 0x0e2b, 0x0e2c, 0x0e2d, 0x0e2e, 0x0e2f, // c8
  343. 0x0e30, 0x0e31, 0x0e32, 0x0e33, 0x0e34, 0x0e35, 0x0e36, 0x0e37, // d0
  344. 0x0e38, 0x0e39, 0x0e3a, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0x0e3f, // d8
  345. 0x0e40, 0x0e41, 0x0e42, 0x0e43, 0x0e44, 0x0e45, 0x0e46, 0x0e47, // e0
  346. 0x0e48, 0x0e49, 0x0e4a, 0x0e4b, 0x0e4c, 0x0e4d, 0x0e4e, 0x0e4f, // e8
  347. 0x0e50, 0x0e51, 0x0e52, 0x0e53, 0x0e54, 0x0e55, 0x0e56, 0x0e57, // f0
  348. 0x0e58, 0x0e59, 0x0e5a, 0x0e5b, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // f8
  349. };
  350. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-13.TXT
  351. static const uint32_t ISO_8859_13[0x80] = {
  352. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  353. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  354. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  355. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  356. 0x00a0, 0x201d, 0x00a2, 0x00a3, 0x00a4, 0x201e, 0x00a6, 0x00a7, // a0
  357. 0x00d8, 0x00a9, 0x0156, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00c6, // a8
  358. 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x201c, 0x00b5, 0x00b6, 0x00b7, // b0
  359. 0x00f8, 0x00b9, 0x0157, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00e6, // b8
  360. 0x0104, 0x012e, 0x0100, 0x0106, 0x00c4, 0x00c5, 0x0118, 0x0112, // c0
  361. 0x010c, 0x00c9, 0x0179, 0x0116, 0x0122, 0x0136, 0x012a, 0x013b, // c8
  362. 0x0160, 0x0143, 0x0145, 0x00d3, 0x014c, 0x00d5, 0x00d6, 0x00d7, // d0
  363. 0x0172, 0x0141, 0x015a, 0x016a, 0x00dc, 0x017b, 0x017d, 0x00df, // d8
  364. 0x0105, 0x012f, 0x0101, 0x0107, 0x00e4, 0x00e5, 0x0119, 0x0113, // e0
  365. 0x010d, 0x00e9, 0x017a, 0x0117, 0x0123, 0x0137, 0x012b, 0x013c, // e8
  366. 0x0161, 0x0144, 0x0146, 0x00f3, 0x014d, 0x00f5, 0x00f6, 0x00f7, // f0
  367. 0x0173, 0x0142, 0x015b, 0x016b, 0x00fc, 0x017c, 0x017e, 0x2019, // f8
  368. };
  369. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-14.TXT
  370. static const uint32_t ISO_8859_14[0x80] = {
  371. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  372. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  373. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  374. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  375. 0x00a0, 0x1e02, 0x1e03, 0x00a3, 0x010a, 0x010b, 0x1e0a, 0x00a7, // a0
  376. 0x1e80, 0x00a9, 0x1e82, 0x1e0b, 0x1ef2, 0x00ad, 0x00ae, 0x0178, // a8
  377. 0x1e1e, 0x1e1f, 0x0120, 0x0121, 0x1e40, 0x1e41, 0x00b6, 0x1e56, // b0
  378. 0x1e81, 0x1e57, 0x1e83, 0x1e60, 0x1ef3, 0x1e84, 0x1e85, 0x1e61, // b8
  379. 0x00c0, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x00c7, // c0
  380. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  381. 0x0174, 0x00d1, 0x00d2, 0x00d3, 0x00d4, 0x00d5, 0x00d6, 0x1e6a, // d0
  382. 0x00d8, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x00dd, 0x0176, 0x00df, // d8
  383. 0x00e0, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x00e7, // e0
  384. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  385. 0x0175, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x1e6b, // f0
  386. 0x00f8, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x0177, 0x00ff, // f8
  387. };
  388. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-15.TXT
  389. static const uint32_t ISO_8859_15[0x80] = {
  390. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  391. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  392. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  393. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  394. 0x00a0, 0x00a1, 0x00a2, 0x00a3, 0x20ac, 0x00a5, 0x0160, 0x00a7, // a0
  395. 0x0161, 0x00a9, 0x00aa, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, // a8
  396. 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x017d, 0x00b5, 0x00b6, 0x00b7, // b0
  397. 0x017e, 0x00b9, 0x00ba, 0x00bb, 0x0152, 0x0153, 0x0178, 0x00bf, // b8
  398. 0x00c0, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x00c7, // c0
  399. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  400. 0x00d0, 0x00d1, 0x00d2, 0x00d3, 0x00d4, 0x00d5, 0x00d6, 0x00d7, // d0
  401. 0x00d8, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x00dd, 0x00de, 0x00df, // d8
  402. 0x00e0, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x00e7, // e0
  403. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  404. 0x00f0, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x00f7, // f0
  405. 0x00f8, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x00fe, 0x00ff, // f8
  406. };
  407. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-16.TXT
  408. static const uint32_t ISO_8859_16[0x80] = {
  409. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  410. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  411. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  412. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  413. 0x00a0, 0x0104, 0x0105, 0x0141, 0x20ac, 0x201e, 0x0160, 0x00a7, // a0
  414. 0x0161, 0x00a9, 0x0218, 0x00ab, 0x0179, 0x00ad, 0x017a, 0x017b, // a8
  415. 0x00b0, 0x00b1, 0x010c, 0x0142, 0x017d, 0x201d, 0x00b6, 0x00b7, // b0
  416. 0x017e, 0x010d, 0x0219, 0x00bb, 0x0152, 0x0153, 0x0178, 0x017c, // b8
  417. 0x00c0, 0x00c1, 0x00c2, 0x0102, 0x00c4, 0x0106, 0x00c6, 0x00c7, // c0
  418. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  419. 0x0110, 0x0143, 0x00d2, 0x00d3, 0x00d4, 0x0150, 0x00d6, 0x015a, // d0
  420. 0x0170, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x0118, 0x021a, 0x00df, // d8
  421. 0x00e0, 0x00e1, 0x00e2, 0x0103, 0x00e4, 0x0107, 0x00e6, 0x00e7, // e0
  422. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  423. 0x0111, 0x0144, 0x00f2, 0x00f3, 0x00f4, 0x0151, 0x00f6, 0x015b, // f0
  424. 0x0171, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x0119, 0x021b, 0x00ff, // f8
  425. };
  426. // Reference: http://www.iana.org/go/rfc1489
  427. // Reference: http://www.unicode.org/Public/MAPPINGS/VENDORS/MISC/KOI8-R.TXT
  428. static const uint32_t KOI8_R[0x80] = {
  429. 0x0080, 0x0081, 0x0082, 0x0083, 0x0084, 0x0085, 0x0086, 0x0087, // 80
  430. 0x0088, 0x0089, 0x008a, 0x008b, 0x008c, 0x008d, 0x008e, 0x008f, // 88
  431. 0x0090, 0x0091, 0x0092, 0x0093, 0x0094, 0x0095, 0x0096, 0x0097, // 90
  432. 0x0098, 0x0099, 0x009a, 0x009b, 0x009c, 0x009d, 0x009e, 0x009f, // 98
  433. 0x00a0, 0x0104, 0x0105, 0x0141, 0x20ac, 0x201e, 0x0160, 0x00a7, // a0
  434. 0x0161, 0x00a9, 0x0218, 0x00ab, 0x0179, 0x00ad, 0x017a, 0x017b, // a8
  435. 0x00b0, 0x00b1, 0x010c, 0x0142, 0x017d, 0x201d, 0x00b6, 0x00b7, // b0
  436. 0x017e, 0x010d, 0x0219, 0x00bb, 0x0152, 0x0153, 0x0178, 0x017c, // b8
  437. 0x00c0, 0x00c1, 0x00c2, 0x0102, 0x00c4, 0x0106, 0x00c6, 0x00c7, // c0
  438. 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, // c8
  439. 0x0110, 0x0143, 0x00d2, 0x00d3, 0x00d4, 0x0150, 0x00d6, 0x015a, // d0
  440. 0x0170, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x0118, 0x021a, 0x00df, // d8
  441. 0x00e0, 0x00e1, 0x00e2, 0x0103, 0x00e4, 0x0107, 0x00e6, 0x00e7, // e0
  442. 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, // e8
  443. 0x0111, 0x0144, 0x00f2, 0x00f3, 0x00f4, 0x0151, 0x00f6, 0x015b, // f0
  444. 0x0171, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x0119, 0x021b, 0x00ff, // f8
  445. };
  446. // Reference: http://varamozhi.sourceforge.net/iscii91.pdf
  447. static const uint32_t ISCII[0x80] = {
  448. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // 80
  449. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // 88
  450. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // 90
  451. 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // 98
  452. 0xfffd, 0x0901, 0x0902, 0x0903, 0x0905, 0x0906, 0x0907, 0x0908, // a0
  453. 0x0909, 0x090a, 0x090b, 0x090e, 0x090f, 0x0910, 0x090d, 0x0912, // a8
  454. 0x0913, 0x0914, 0x0911, 0x0915, 0x0916, 0x0917, 0x0918, 0x0919, // b0
  455. 0x091a, 0x091b, 0x091c, 0x091d, 0x091e, 0x091f, 0x0920, 0x0921, // b8
  456. 0x0922, 0x0923, 0x0924, 0x0925, 0x0926, 0x0927, 0x0928, 0x0929, // c0
  457. 0x092a, 0x092b, 0x092c, 0x092d, 0x092e, 0x092f, 0x095f, 0x0930, // c8
  458. 0x0931, 0x0932, 0x0933, 0x0934, 0x0935, 0x0936, 0x0937, 0x0938, // d0
  459. 0x0939, 0x0020, 0x093e, 0x093f, 0x0940, 0x0941, 0x0942, 0x0943, // d8
  460. 0x0946, 0x0947, 0x0948, 0x0945, 0x094a, 0x094b, 0x094c, 0x0949, // e0
  461. 0x094d, 0x093c, 0x0964, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // e8
  462. 0x0020, 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, // f0
  463. 0x0037, 0x0038, 0x0039, 0xfffd, 0xfffd, 0xfffd, 0xfffd, 0xfffd, // f8
  464. };
  465. static uint32_t
  466. string_decoder_getc_us_ascii(espeak_ng_TEXT_DECODER *decoder)
  467. {
  468. uint8_t c = *decoder->current++ & 0xFF;
  469. return (c >= 0x80) ? 0xFFFD : c;
  470. }
  471. // Reference: http://www.iana.org/go/rfc1345
  472. // Reference: http://www.unicode.org/Public/MAPPINGS/ISO8859/8859-1.TXT
  473. static uint32_t
  474. string_decoder_getc_iso_8859_1(espeak_ng_TEXT_DECODER *decoder)
  475. {
  476. return *decoder->current++ & 0xFF;
  477. }
  478. static uint32_t
  479. string_decoder_getc_codepage(espeak_ng_TEXT_DECODER *decoder)
  480. {
  481. uint8_t c = *decoder->current++ & 0xFF;
  482. return (c >= 0x80) ? decoder->codepage[c - 0x80] : c;
  483. }
  484. typedef struct
  485. {
  486. uint32_t (*get)(espeak_ng_TEXT_DECODER *decoder);
  487. const uint32_t *codepage;
  488. } encoding_t;
  489. static const encoding_t string_decoders[] = {
  490. { NULL, NULL },
  491. { string_decoder_getc_us_ascii, NULL },
  492. { string_decoder_getc_iso_8859_1, NULL },
  493. { string_decoder_getc_codepage, ISO_8859_2 },
  494. { string_decoder_getc_codepage, ISO_8859_3 },
  495. { string_decoder_getc_codepage, ISO_8859_4 },
  496. { string_decoder_getc_codepage, ISO_8859_5 },
  497. { string_decoder_getc_codepage, ISO_8859_6 },
  498. { string_decoder_getc_codepage, ISO_8859_7 },
  499. { string_decoder_getc_codepage, ISO_8859_8 },
  500. { string_decoder_getc_codepage, ISO_8859_9 },
  501. { string_decoder_getc_codepage, ISO_8859_10 },
  502. { string_decoder_getc_codepage, ISO_8859_11 },
  503. // ISO-8859-12 is not a valid encoding.
  504. { string_decoder_getc_codepage, ISO_8859_13 },
  505. { string_decoder_getc_codepage, ISO_8859_14 },
  506. { string_decoder_getc_codepage, ISO_8859_15 },
  507. { string_decoder_getc_codepage, ISO_8859_16 },
  508. { string_decoder_getc_codepage, KOI8_R },
  509. { string_decoder_getc_codepage, ISCII },
  510. };
  511. espeak_ng_TEXT_DECODER *
  512. create_text_decoder(void)
  513. {
  514. espeak_ng_TEXT_DECODER *decoder = malloc(sizeof(espeak_ng_TEXT_DECODER));
  515. if (!decoder) return NULL;
  516. decoder->current = NULL;
  517. decoder->end = NULL;
  518. decoder->get = NULL;
  519. decoder->codepage = NULL;
  520. return decoder;
  521. }
  522. void
  523. destroy_text_decoder(espeak_ng_TEXT_DECODER *decoder)
  524. {
  525. if (decoder) free(decoder);
  526. }
  527. espeak_ng_STATUS
  528. text_decoder_decode_string(espeak_ng_TEXT_DECODER *decoder,
  529. const char *string,
  530. int length,
  531. espeak_ng_ENCODING encoding)
  532. {
  533. if (encoding > ESPEAKNG_ENCODING_ISCII)
  534. return ENS_UNKNOWN_TEXT_ENCODING;
  535. const encoding_t *enc = string_decoders + encoding;
  536. if (enc->get == NULL)
  537. return ENS_UNKNOWN_TEXT_ENCODING;
  538. decoder->get = enc->get;
  539. decoder->codepage = enc->codepage;
  540. decoder->current = string;
  541. decoder->end = string + length;
  542. return ENS_OK;
  543. }
  544. int
  545. text_decoder_eof(espeak_ng_TEXT_DECODER *decoder)
  546. {
  547. return decoder->current == decoder->end;
  548. }
  549. uint32_t
  550. text_decoder_getc(espeak_ng_TEXT_DECODER *decoder)
  551. {
  552. return decoder->get(decoder);
  553. }