gunicode.h 40 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929
  1. /* gunicode.h - Unicode manipulation functions
  2. *
  3. * Copyright (C) 1999, 2000 Tom Tromey
  4. * Copyright 2000, 2005 Red Hat, Inc.
  5. *
  6. * This library is free software; you can redistribute it and/or
  7. * modify it under the terms of the GNU Lesser General Public
  8. * License as published by the Free Software Foundation; either
  9. * version 2.1 of the License, or (at your option) any later version.
  10. *
  11. * This library 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 GNU
  14. * Lesser General Public License for more details.
  15. *
  16. * You should have received a copy of the GNU Lesser General Public License
  17. * along with this library; if not, see <http://www.gnu.org/licenses/>.
  18. */
  19. #ifndef __G_UNICODE_H__
  20. #define __G_UNICODE_H__
  21. #if !defined (__GLIB_H_INSIDE__) && !defined (GLIB_COMPILATION)
  22. #error "Only <glib.h> can be included directly."
  23. #endif
  24. #include <glib/gerror.h>
  25. #include <glib/gtypes.h>
  26. G_BEGIN_DECLS
  27. /**
  28. * gunichar:
  29. *
  30. * A type which can hold any UTF-32 or UCS-4 character code,
  31. * also known as a Unicode code point.
  32. *
  33. * If you want to produce the UTF-8 representation of a #gunichar,
  34. * use g_ucs4_to_utf8(). See also g_utf8_to_ucs4() for the reverse
  35. * process.
  36. *
  37. * To print/scan values of this type as integer, use
  38. * %G_GINT32_MODIFIER and/or %G_GUINT32_FORMAT.
  39. *
  40. * The notation to express a Unicode code point in running text is
  41. * as a hexadecimal number with four to six digits and uppercase
  42. * letters, prefixed by the string "U+". Leading zeros are omitted,
  43. * unless the code point would have fewer than four hexadecimal digits.
  44. * For example, "U+0041 LATIN CAPITAL LETTER A". To print a code point
  45. * in the U+-notation, use the format string "U+\%04"G_GINT32_FORMAT"X".
  46. * To scan, use the format string "U+\%06"G_GINT32_FORMAT"X".
  47. *
  48. * |[
  49. * gunichar c;
  50. * sscanf ("U+0041", "U+%06"G_GINT32_FORMAT"X", &amp;c)
  51. * g_print ("Read U+%04"G_GINT32_FORMAT"X", c);
  52. * ]|
  53. */
  54. typedef guint32 gunichar;
  55. /**
  56. * gunichar2:
  57. *
  58. * A type which can hold any UTF-16 code
  59. * point<footnote id="utf16_surrogate_pairs">UTF-16 also has so called
  60. * <firstterm>surrogate pairs</firstterm> to encode characters beyond
  61. * the BMP as pairs of 16bit numbers. Surrogate pairs cannot be stored
  62. * in a single gunichar2 field, but all GLib functions accepting gunichar2
  63. * arrays will correctly interpret surrogate pairs.</footnote>.
  64. *
  65. * To print/scan values of this type to/from text you need to convert
  66. * to/from UTF-8, using g_utf16_to_utf8()/g_utf8_to_utf16().
  67. *
  68. * To print/scan values of this type as integer, use
  69. * %G_GINT16_MODIFIER and/or %G_GUINT16_FORMAT.
  70. */
  71. typedef guint16 gunichar2;
  72. /**
  73. * GUnicodeType:
  74. * @G_UNICODE_CONTROL: General category "Other, Control" (Cc)
  75. * @G_UNICODE_FORMAT: General category "Other, Format" (Cf)
  76. * @G_UNICODE_UNASSIGNED: General category "Other, Not Assigned" (Cn)
  77. * @G_UNICODE_PRIVATE_USE: General category "Other, Private Use" (Co)
  78. * @G_UNICODE_SURROGATE: General category "Other, Surrogate" (Cs)
  79. * @G_UNICODE_LOWERCASE_LETTER: General category "Letter, Lowercase" (Ll)
  80. * @G_UNICODE_MODIFIER_LETTER: General category "Letter, Modifier" (Lm)
  81. * @G_UNICODE_OTHER_LETTER: General category "Letter, Other" (Lo)
  82. * @G_UNICODE_TITLECASE_LETTER: General category "Letter, Titlecase" (Lt)
  83. * @G_UNICODE_UPPERCASE_LETTER: General category "Letter, Uppercase" (Lu)
  84. * @G_UNICODE_SPACING_MARK: General category "Mark, Spacing" (Mc)
  85. * @G_UNICODE_ENCLOSING_MARK: General category "Mark, Enclosing" (Me)
  86. * @G_UNICODE_NON_SPACING_MARK: General category "Mark, Nonspacing" (Mn)
  87. * @G_UNICODE_DECIMAL_NUMBER: General category "Number, Decimal Digit" (Nd)
  88. * @G_UNICODE_LETTER_NUMBER: General category "Number, Letter" (Nl)
  89. * @G_UNICODE_OTHER_NUMBER: General category "Number, Other" (No)
  90. * @G_UNICODE_CONNECT_PUNCTUATION: General category "Punctuation, Connector" (Pc)
  91. * @G_UNICODE_DASH_PUNCTUATION: General category "Punctuation, Dash" (Pd)
  92. * @G_UNICODE_CLOSE_PUNCTUATION: General category "Punctuation, Close" (Pe)
  93. * @G_UNICODE_FINAL_PUNCTUATION: General category "Punctuation, Final quote" (Pf)
  94. * @G_UNICODE_INITIAL_PUNCTUATION: General category "Punctuation, Initial quote" (Pi)
  95. * @G_UNICODE_OTHER_PUNCTUATION: General category "Punctuation, Other" (Po)
  96. * @G_UNICODE_OPEN_PUNCTUATION: General category "Punctuation, Open" (Ps)
  97. * @G_UNICODE_CURRENCY_SYMBOL: General category "Symbol, Currency" (Sc)
  98. * @G_UNICODE_MODIFIER_SYMBOL: General category "Symbol, Modifier" (Sk)
  99. * @G_UNICODE_MATH_SYMBOL: General category "Symbol, Math" (Sm)
  100. * @G_UNICODE_OTHER_SYMBOL: General category "Symbol, Other" (So)
  101. * @G_UNICODE_LINE_SEPARATOR: General category "Separator, Line" (Zl)
  102. * @G_UNICODE_PARAGRAPH_SEPARATOR: General category "Separator, Paragraph" (Zp)
  103. * @G_UNICODE_SPACE_SEPARATOR: General category "Separator, Space" (Zs)
  104. *
  105. * These are the possible character classifications from the
  106. * Unicode specification.
  107. * See [Unicode Character Database](http://www.unicode.org/reports/tr44/#General_Category_Values).
  108. */
  109. typedef enum
  110. {
  111. G_UNICODE_CONTROL,
  112. G_UNICODE_FORMAT,
  113. G_UNICODE_UNASSIGNED,
  114. G_UNICODE_PRIVATE_USE,
  115. G_UNICODE_SURROGATE,
  116. G_UNICODE_LOWERCASE_LETTER,
  117. G_UNICODE_MODIFIER_LETTER,
  118. G_UNICODE_OTHER_LETTER,
  119. G_UNICODE_TITLECASE_LETTER,
  120. G_UNICODE_UPPERCASE_LETTER,
  121. G_UNICODE_SPACING_MARK,
  122. G_UNICODE_ENCLOSING_MARK,
  123. G_UNICODE_NON_SPACING_MARK,
  124. G_UNICODE_DECIMAL_NUMBER,
  125. G_UNICODE_LETTER_NUMBER,
  126. G_UNICODE_OTHER_NUMBER,
  127. G_UNICODE_CONNECT_PUNCTUATION,
  128. G_UNICODE_DASH_PUNCTUATION,
  129. G_UNICODE_CLOSE_PUNCTUATION,
  130. G_UNICODE_FINAL_PUNCTUATION,
  131. G_UNICODE_INITIAL_PUNCTUATION,
  132. G_UNICODE_OTHER_PUNCTUATION,
  133. G_UNICODE_OPEN_PUNCTUATION,
  134. G_UNICODE_CURRENCY_SYMBOL,
  135. G_UNICODE_MODIFIER_SYMBOL,
  136. G_UNICODE_MATH_SYMBOL,
  137. G_UNICODE_OTHER_SYMBOL,
  138. G_UNICODE_LINE_SEPARATOR,
  139. G_UNICODE_PARAGRAPH_SEPARATOR,
  140. G_UNICODE_SPACE_SEPARATOR
  141. } GUnicodeType;
  142. /**
  143. * G_UNICODE_COMBINING_MARK:
  144. *
  145. * Older name for %G_UNICODE_SPACING_MARK.
  146. *
  147. * Deprecated: 2.30: Use %G_UNICODE_SPACING_MARK.
  148. */
  149. #define G_UNICODE_COMBINING_MARK G_UNICODE_SPACING_MARK GLIB_DEPRECATED_MACRO_IN_2_30_FOR(G_UNICODE_SPACING_MARK)
  150. /**
  151. * GUnicodeBreakType:
  152. * @G_UNICODE_BREAK_MANDATORY: Mandatory Break (BK)
  153. * @G_UNICODE_BREAK_CARRIAGE_RETURN: Carriage Return (CR)
  154. * @G_UNICODE_BREAK_LINE_FEED: Line Feed (LF)
  155. * @G_UNICODE_BREAK_COMBINING_MARK: Attached Characters and Combining Marks (CM)
  156. * @G_UNICODE_BREAK_SURROGATE: Surrogates (SG)
  157. * @G_UNICODE_BREAK_ZERO_WIDTH_SPACE: Zero Width Space (ZW)
  158. * @G_UNICODE_BREAK_INSEPARABLE: Inseparable (IN)
  159. * @G_UNICODE_BREAK_NON_BREAKING_GLUE: Non-breaking ("Glue") (GL)
  160. * @G_UNICODE_BREAK_CONTINGENT: Contingent Break Opportunity (CB)
  161. * @G_UNICODE_BREAK_SPACE: Space (SP)
  162. * @G_UNICODE_BREAK_AFTER: Break Opportunity After (BA)
  163. * @G_UNICODE_BREAK_BEFORE: Break Opportunity Before (BB)
  164. * @G_UNICODE_BREAK_BEFORE_AND_AFTER: Break Opportunity Before and After (B2)
  165. * @G_UNICODE_BREAK_HYPHEN: Hyphen (HY)
  166. * @G_UNICODE_BREAK_NON_STARTER: Nonstarter (NS)
  167. * @G_UNICODE_BREAK_OPEN_PUNCTUATION: Opening Punctuation (OP)
  168. * @G_UNICODE_BREAK_CLOSE_PUNCTUATION: Closing Punctuation (CL)
  169. * @G_UNICODE_BREAK_QUOTATION: Ambiguous Quotation (QU)
  170. * @G_UNICODE_BREAK_EXCLAMATION: Exclamation/Interrogation (EX)
  171. * @G_UNICODE_BREAK_IDEOGRAPHIC: Ideographic (ID)
  172. * @G_UNICODE_BREAK_NUMERIC: Numeric (NU)
  173. * @G_UNICODE_BREAK_INFIX_SEPARATOR: Infix Separator (Numeric) (IS)
  174. * @G_UNICODE_BREAK_SYMBOL: Symbols Allowing Break After (SY)
  175. * @G_UNICODE_BREAK_ALPHABETIC: Ordinary Alphabetic and Symbol Characters (AL)
  176. * @G_UNICODE_BREAK_PREFIX: Prefix (Numeric) (PR)
  177. * @G_UNICODE_BREAK_POSTFIX: Postfix (Numeric) (PO)
  178. * @G_UNICODE_BREAK_COMPLEX_CONTEXT: Complex Content Dependent (South East Asian) (SA)
  179. * @G_UNICODE_BREAK_AMBIGUOUS: Ambiguous (Alphabetic or Ideographic) (AI)
  180. * @G_UNICODE_BREAK_UNKNOWN: Unknown (XX)
  181. * @G_UNICODE_BREAK_NEXT_LINE: Next Line (NL)
  182. * @G_UNICODE_BREAK_WORD_JOINER: Word Joiner (WJ)
  183. * @G_UNICODE_BREAK_HANGUL_L_JAMO: Hangul L Jamo (JL)
  184. * @G_UNICODE_BREAK_HANGUL_V_JAMO: Hangul V Jamo (JV)
  185. * @G_UNICODE_BREAK_HANGUL_T_JAMO: Hangul T Jamo (JT)
  186. * @G_UNICODE_BREAK_HANGUL_LV_SYLLABLE: Hangul LV Syllable (H2)
  187. * @G_UNICODE_BREAK_HANGUL_LVT_SYLLABLE: Hangul LVT Syllable (H3)
  188. * @G_UNICODE_BREAK_CLOSE_PARANTHESIS: Closing Parenthesis (CP). Since 2.28
  189. * @G_UNICODE_BREAK_CONDITIONAL_JAPANESE_STARTER: Conditional Japanese Starter (CJ). Since: 2.32
  190. * @G_UNICODE_BREAK_HEBREW_LETTER: Hebrew Letter (HL). Since: 2.32
  191. * @G_UNICODE_BREAK_REGIONAL_INDICATOR: Regional Indicator (RI). Since: 2.36
  192. * @G_UNICODE_BREAK_EMOJI_BASE: Emoji Base (EB). Since: 2.50
  193. * @G_UNICODE_BREAK_EMOJI_MODIFIER: Emoji Modifier (EM). Since: 2.50
  194. * @G_UNICODE_BREAK_ZERO_WIDTH_JOINER: Zero Width Joiner (ZWJ). Since: 2.50
  195. *
  196. * These are the possible line break classifications.
  197. *
  198. * Since new unicode versions may add new types here, applications should be ready
  199. * to handle unknown values. They may be regarded as %G_UNICODE_BREAK_UNKNOWN.
  200. *
  201. * See [Unicode Line Breaking Algorithm](http://www.unicode.org/unicode/reports/tr14/).
  202. */
  203. typedef enum
  204. {
  205. G_UNICODE_BREAK_MANDATORY,
  206. G_UNICODE_BREAK_CARRIAGE_RETURN,
  207. G_UNICODE_BREAK_LINE_FEED,
  208. G_UNICODE_BREAK_COMBINING_MARK,
  209. G_UNICODE_BREAK_SURROGATE,
  210. G_UNICODE_BREAK_ZERO_WIDTH_SPACE,
  211. G_UNICODE_BREAK_INSEPARABLE,
  212. G_UNICODE_BREAK_NON_BREAKING_GLUE,
  213. G_UNICODE_BREAK_CONTINGENT,
  214. G_UNICODE_BREAK_SPACE,
  215. G_UNICODE_BREAK_AFTER,
  216. G_UNICODE_BREAK_BEFORE,
  217. G_UNICODE_BREAK_BEFORE_AND_AFTER,
  218. G_UNICODE_BREAK_HYPHEN,
  219. G_UNICODE_BREAK_NON_STARTER,
  220. G_UNICODE_BREAK_OPEN_PUNCTUATION,
  221. G_UNICODE_BREAK_CLOSE_PUNCTUATION,
  222. G_UNICODE_BREAK_QUOTATION,
  223. G_UNICODE_BREAK_EXCLAMATION,
  224. G_UNICODE_BREAK_IDEOGRAPHIC,
  225. G_UNICODE_BREAK_NUMERIC,
  226. G_UNICODE_BREAK_INFIX_SEPARATOR,
  227. G_UNICODE_BREAK_SYMBOL,
  228. G_UNICODE_BREAK_ALPHABETIC,
  229. G_UNICODE_BREAK_PREFIX,
  230. G_UNICODE_BREAK_POSTFIX,
  231. G_UNICODE_BREAK_COMPLEX_CONTEXT,
  232. G_UNICODE_BREAK_AMBIGUOUS,
  233. G_UNICODE_BREAK_UNKNOWN,
  234. G_UNICODE_BREAK_NEXT_LINE,
  235. G_UNICODE_BREAK_WORD_JOINER,
  236. G_UNICODE_BREAK_HANGUL_L_JAMO,
  237. G_UNICODE_BREAK_HANGUL_V_JAMO,
  238. G_UNICODE_BREAK_HANGUL_T_JAMO,
  239. G_UNICODE_BREAK_HANGUL_LV_SYLLABLE,
  240. G_UNICODE_BREAK_HANGUL_LVT_SYLLABLE,
  241. G_UNICODE_BREAK_CLOSE_PARANTHESIS,
  242. G_UNICODE_BREAK_CONDITIONAL_JAPANESE_STARTER,
  243. G_UNICODE_BREAK_HEBREW_LETTER,
  244. G_UNICODE_BREAK_REGIONAL_INDICATOR,
  245. G_UNICODE_BREAK_EMOJI_BASE,
  246. G_UNICODE_BREAK_EMOJI_MODIFIER,
  247. G_UNICODE_BREAK_ZERO_WIDTH_JOINER
  248. } GUnicodeBreakType;
  249. /**
  250. * GUnicodeScript:
  251. * @G_UNICODE_SCRIPT_INVALID_CODE:
  252. * a value never returned from g_unichar_get_script()
  253. * @G_UNICODE_SCRIPT_COMMON: a character used by multiple different scripts
  254. * @G_UNICODE_SCRIPT_INHERITED: a mark glyph that takes its script from the
  255. * base glyph to which it is attached
  256. * @G_UNICODE_SCRIPT_ARABIC: Arabic
  257. * @G_UNICODE_SCRIPT_ARMENIAN: Armenian
  258. * @G_UNICODE_SCRIPT_BENGALI: Bengali
  259. * @G_UNICODE_SCRIPT_BOPOMOFO: Bopomofo
  260. * @G_UNICODE_SCRIPT_CHEROKEE: Cherokee
  261. * @G_UNICODE_SCRIPT_COPTIC: Coptic
  262. * @G_UNICODE_SCRIPT_CYRILLIC: Cyrillic
  263. * @G_UNICODE_SCRIPT_DESERET: Deseret
  264. * @G_UNICODE_SCRIPT_DEVANAGARI: Devanagari
  265. * @G_UNICODE_SCRIPT_ETHIOPIC: Ethiopic
  266. * @G_UNICODE_SCRIPT_GEORGIAN: Georgian
  267. * @G_UNICODE_SCRIPT_GOTHIC: Gothic
  268. * @G_UNICODE_SCRIPT_GREEK: Greek
  269. * @G_UNICODE_SCRIPT_GUJARATI: Gujarati
  270. * @G_UNICODE_SCRIPT_GURMUKHI: Gurmukhi
  271. * @G_UNICODE_SCRIPT_HAN: Han
  272. * @G_UNICODE_SCRIPT_HANGUL: Hangul
  273. * @G_UNICODE_SCRIPT_HEBREW: Hebrew
  274. * @G_UNICODE_SCRIPT_HIRAGANA: Hiragana
  275. * @G_UNICODE_SCRIPT_KANNADA: Kannada
  276. * @G_UNICODE_SCRIPT_KATAKANA: Katakana
  277. * @G_UNICODE_SCRIPT_KHMER: Khmer
  278. * @G_UNICODE_SCRIPT_LAO: Lao
  279. * @G_UNICODE_SCRIPT_LATIN: Latin
  280. * @G_UNICODE_SCRIPT_MALAYALAM: Malayalam
  281. * @G_UNICODE_SCRIPT_MONGOLIAN: Mongolian
  282. * @G_UNICODE_SCRIPT_MYANMAR: Myanmar
  283. * @G_UNICODE_SCRIPT_OGHAM: Ogham
  284. * @G_UNICODE_SCRIPT_OLD_ITALIC: Old Italic
  285. * @G_UNICODE_SCRIPT_ORIYA: Oriya
  286. * @G_UNICODE_SCRIPT_RUNIC: Runic
  287. * @G_UNICODE_SCRIPT_SINHALA: Sinhala
  288. * @G_UNICODE_SCRIPT_SYRIAC: Syriac
  289. * @G_UNICODE_SCRIPT_TAMIL: Tamil
  290. * @G_UNICODE_SCRIPT_TELUGU: Telugu
  291. * @G_UNICODE_SCRIPT_THAANA: Thaana
  292. * @G_UNICODE_SCRIPT_THAI: Thai
  293. * @G_UNICODE_SCRIPT_TIBETAN: Tibetan
  294. * @G_UNICODE_SCRIPT_CANADIAN_ABORIGINAL:
  295. * Canadian Aboriginal
  296. * @G_UNICODE_SCRIPT_YI: Yi
  297. * @G_UNICODE_SCRIPT_TAGALOG: Tagalog
  298. * @G_UNICODE_SCRIPT_HANUNOO: Hanunoo
  299. * @G_UNICODE_SCRIPT_BUHID: Buhid
  300. * @G_UNICODE_SCRIPT_TAGBANWA: Tagbanwa
  301. * @G_UNICODE_SCRIPT_BRAILLE: Braille
  302. * @G_UNICODE_SCRIPT_CYPRIOT: Cypriot
  303. * @G_UNICODE_SCRIPT_LIMBU: Limbu
  304. * @G_UNICODE_SCRIPT_OSMANYA: Osmanya
  305. * @G_UNICODE_SCRIPT_SHAVIAN: Shavian
  306. * @G_UNICODE_SCRIPT_LINEAR_B: Linear B
  307. * @G_UNICODE_SCRIPT_TAI_LE: Tai Le
  308. * @G_UNICODE_SCRIPT_UGARITIC: Ugaritic
  309. * @G_UNICODE_SCRIPT_NEW_TAI_LUE:
  310. * New Tai Lue
  311. * @G_UNICODE_SCRIPT_BUGINESE: Buginese
  312. * @G_UNICODE_SCRIPT_GLAGOLITIC: Glagolitic
  313. * @G_UNICODE_SCRIPT_TIFINAGH: Tifinagh
  314. * @G_UNICODE_SCRIPT_SYLOTI_NAGRI:
  315. * Syloti Nagri
  316. * @G_UNICODE_SCRIPT_OLD_PERSIAN:
  317. * Old Persian
  318. * @G_UNICODE_SCRIPT_KHAROSHTHI: Kharoshthi
  319. * @G_UNICODE_SCRIPT_UNKNOWN: an unassigned code point
  320. * @G_UNICODE_SCRIPT_BALINESE: Balinese
  321. * @G_UNICODE_SCRIPT_CUNEIFORM: Cuneiform
  322. * @G_UNICODE_SCRIPT_PHOENICIAN: Phoenician
  323. * @G_UNICODE_SCRIPT_PHAGS_PA: Phags-pa
  324. * @G_UNICODE_SCRIPT_NKO: N'Ko
  325. * @G_UNICODE_SCRIPT_KAYAH_LI: Kayah Li. Since 2.16.3
  326. * @G_UNICODE_SCRIPT_LEPCHA: Lepcha. Since 2.16.3
  327. * @G_UNICODE_SCRIPT_REJANG: Rejang. Since 2.16.3
  328. * @G_UNICODE_SCRIPT_SUNDANESE: Sundanese. Since 2.16.3
  329. * @G_UNICODE_SCRIPT_SAURASHTRA: Saurashtra. Since 2.16.3
  330. * @G_UNICODE_SCRIPT_CHAM: Cham. Since 2.16.3
  331. * @G_UNICODE_SCRIPT_OL_CHIKI: Ol Chiki. Since 2.16.3
  332. * @G_UNICODE_SCRIPT_VAI: Vai. Since 2.16.3
  333. * @G_UNICODE_SCRIPT_CARIAN: Carian. Since 2.16.3
  334. * @G_UNICODE_SCRIPT_LYCIAN: Lycian. Since 2.16.3
  335. * @G_UNICODE_SCRIPT_LYDIAN: Lydian. Since 2.16.3
  336. * @G_UNICODE_SCRIPT_AVESTAN: Avestan. Since 2.26
  337. * @G_UNICODE_SCRIPT_BAMUM: Bamum. Since 2.26
  338. * @G_UNICODE_SCRIPT_EGYPTIAN_HIEROGLYPHS:
  339. * Egyptian Hieroglpyhs. Since 2.26
  340. * @G_UNICODE_SCRIPT_IMPERIAL_ARAMAIC:
  341. * Imperial Aramaic. Since 2.26
  342. * @G_UNICODE_SCRIPT_INSCRIPTIONAL_PAHLAVI:
  343. * Inscriptional Pahlavi. Since 2.26
  344. * @G_UNICODE_SCRIPT_INSCRIPTIONAL_PARTHIAN:
  345. * Inscriptional Parthian. Since 2.26
  346. * @G_UNICODE_SCRIPT_JAVANESE: Javanese. Since 2.26
  347. * @G_UNICODE_SCRIPT_KAITHI: Kaithi. Since 2.26
  348. * @G_UNICODE_SCRIPT_LISU: Lisu. Since 2.26
  349. * @G_UNICODE_SCRIPT_MEETEI_MAYEK:
  350. * Meetei Mayek. Since 2.26
  351. * @G_UNICODE_SCRIPT_OLD_SOUTH_ARABIAN:
  352. * Old South Arabian. Since 2.26
  353. * @G_UNICODE_SCRIPT_OLD_TURKIC: Old Turkic. Since 2.28
  354. * @G_UNICODE_SCRIPT_SAMARITAN: Samaritan. Since 2.26
  355. * @G_UNICODE_SCRIPT_TAI_THAM: Tai Tham. Since 2.26
  356. * @G_UNICODE_SCRIPT_TAI_VIET: Tai Viet. Since 2.26
  357. * @G_UNICODE_SCRIPT_BATAK: Batak. Since 2.28
  358. * @G_UNICODE_SCRIPT_BRAHMI: Brahmi. Since 2.28
  359. * @G_UNICODE_SCRIPT_MANDAIC: Mandaic. Since 2.28
  360. * @G_UNICODE_SCRIPT_CHAKMA: Chakma. Since: 2.32
  361. * @G_UNICODE_SCRIPT_MEROITIC_CURSIVE: Meroitic Cursive. Since: 2.32
  362. * @G_UNICODE_SCRIPT_MEROITIC_HIEROGLYPHS: Meroitic Hieroglyphs. Since: 2.32
  363. * @G_UNICODE_SCRIPT_MIAO: Miao. Since: 2.32
  364. * @G_UNICODE_SCRIPT_SHARADA: Sharada. Since: 2.32
  365. * @G_UNICODE_SCRIPT_SORA_SOMPENG: Sora Sompeng. Since: 2.32
  366. * @G_UNICODE_SCRIPT_TAKRI: Takri. Since: 2.32
  367. * @G_UNICODE_SCRIPT_BASSA_VAH: Bassa. Since: 2.42
  368. * @G_UNICODE_SCRIPT_CAUCASIAN_ALBANIAN: Caucasian Albanian. Since: 2.42
  369. * @G_UNICODE_SCRIPT_DUPLOYAN: Duployan. Since: 2.42
  370. * @G_UNICODE_SCRIPT_ELBASAN: Elbasan. Since: 2.42
  371. * @G_UNICODE_SCRIPT_GRANTHA: Grantha. Since: 2.42
  372. * @G_UNICODE_SCRIPT_KHOJKI: Kjohki. Since: 2.42
  373. * @G_UNICODE_SCRIPT_KHUDAWADI: Khudawadi, Sindhi. Since: 2.42
  374. * @G_UNICODE_SCRIPT_LINEAR_A: Linear A. Since: 2.42
  375. * @G_UNICODE_SCRIPT_MAHAJANI: Mahajani. Since: 2.42
  376. * @G_UNICODE_SCRIPT_MANICHAEAN: Manichaean. Since: 2.42
  377. * @G_UNICODE_SCRIPT_MENDE_KIKAKUI: Mende Kikakui. Since: 2.42
  378. * @G_UNICODE_SCRIPT_MODI: Modi. Since: 2.42
  379. * @G_UNICODE_SCRIPT_MRO: Mro. Since: 2.42
  380. * @G_UNICODE_SCRIPT_NABATAEAN: Nabataean. Since: 2.42
  381. * @G_UNICODE_SCRIPT_OLD_NORTH_ARABIAN: Old North Arabian. Since: 2.42
  382. * @G_UNICODE_SCRIPT_OLD_PERMIC: Old Permic. Since: 2.42
  383. * @G_UNICODE_SCRIPT_PAHAWH_HMONG: Pahawh Hmong. Since: 2.42
  384. * @G_UNICODE_SCRIPT_PALMYRENE: Palmyrene. Since: 2.42
  385. * @G_UNICODE_SCRIPT_PAU_CIN_HAU: Pau Cin Hau. Since: 2.42
  386. * @G_UNICODE_SCRIPT_PSALTER_PAHLAVI: Psalter Pahlavi. Since: 2.42
  387. * @G_UNICODE_SCRIPT_SIDDHAM: Siddham. Since: 2.42
  388. * @G_UNICODE_SCRIPT_TIRHUTA: Tirhuta. Since: 2.42
  389. * @G_UNICODE_SCRIPT_WARANG_CITI: Warang Citi. Since: 2.42
  390. * @G_UNICODE_SCRIPT_AHOM: Ahom. Since: 2.48
  391. * @G_UNICODE_SCRIPT_ANATOLIAN_HIEROGLYPHS: Anatolian Hieroglyphs. Since: 2.48
  392. * @G_UNICODE_SCRIPT_HATRAN: Hatran. Since: 2.48
  393. * @G_UNICODE_SCRIPT_MULTANI: Multani. Since: 2.48
  394. * @G_UNICODE_SCRIPT_OLD_HUNGARIAN: Old Hungarian. Since: 2.48
  395. * @G_UNICODE_SCRIPT_SIGNWRITING: Signwriting. Since: 2.48
  396. * @G_UNICODE_SCRIPT_ADLAM: Adlam. Since: 2.50
  397. * @G_UNICODE_SCRIPT_BHAIKSUKI: Bhaiksuki. Since: 2.50
  398. * @G_UNICODE_SCRIPT_MARCHEN: Marchen. Since: 2.50
  399. * @G_UNICODE_SCRIPT_NEWA: Newa. Since: 2.50
  400. * @G_UNICODE_SCRIPT_OSAGE: Osage. Since: 2.50
  401. * @G_UNICODE_SCRIPT_TANGUT: Tangut. Since: 2.50
  402. * @G_UNICODE_SCRIPT_MASARAM_GONDI: Masaram Gondi. Since: 2.54
  403. * @G_UNICODE_SCRIPT_NUSHU: Nushu. Since: 2.54
  404. * @G_UNICODE_SCRIPT_SOYOMBO: Soyombo. Since: 2.54
  405. * @G_UNICODE_SCRIPT_ZANABAZAR_SQUARE: Zanabazar Square. Since: 2.54
  406. * @G_UNICODE_SCRIPT_DOGRA: Dogra. Since: 2.58
  407. * @G_UNICODE_SCRIPT_GUNJALA_GONDI: Gunjala Gondi. Since: 2.58
  408. * @G_UNICODE_SCRIPT_HANIFI_ROHINGYA: Hanifi Rohingya. Since: 2.58
  409. * @G_UNICODE_SCRIPT_MAKASAR: Makasar. Since: 2.58
  410. * @G_UNICODE_SCRIPT_MEDEFAIDRIN: Medefaidrin. Since: 2.58
  411. * @G_UNICODE_SCRIPT_OLD_SOGDIAN: Old Sogdian. Since: 2.58
  412. * @G_UNICODE_SCRIPT_SOGDIAN: Sogdian. Since: 2.58
  413. * @G_UNICODE_SCRIPT_ELYMAIC: Elym. Since: 2.62
  414. * @G_UNICODE_SCRIPT_NANDINAGARI: Nand. Since: 2.62
  415. * @G_UNICODE_SCRIPT_NYIAKENG_PUACHUE_HMONG: Rohg. Since: 2.62
  416. * @G_UNICODE_SCRIPT_WANCHO: Wcho. Since: 2.62
  417. *
  418. * The #GUnicodeScript enumeration identifies different writing
  419. * systems. The values correspond to the names as defined in the
  420. * Unicode standard. The enumeration has been added in GLib 2.14,
  421. * and is interchangeable with #PangoScript.
  422. *
  423. * Note that new types may be added in the future. Applications
  424. * should be ready to handle unknown values.
  425. * See [Unicode Standard Annex #24: Script names](http://www.unicode.org/reports/tr24/).
  426. */
  427. typedef enum
  428. { /* ISO 15924 code */
  429. G_UNICODE_SCRIPT_INVALID_CODE = -1,
  430. G_UNICODE_SCRIPT_COMMON = 0, /* Zyyy */
  431. G_UNICODE_SCRIPT_INHERITED, /* Zinh (Qaai) */
  432. G_UNICODE_SCRIPT_ARABIC, /* Arab */
  433. G_UNICODE_SCRIPT_ARMENIAN, /* Armn */
  434. G_UNICODE_SCRIPT_BENGALI, /* Beng */
  435. G_UNICODE_SCRIPT_BOPOMOFO, /* Bopo */
  436. G_UNICODE_SCRIPT_CHEROKEE, /* Cher */
  437. G_UNICODE_SCRIPT_COPTIC, /* Copt (Qaac) */
  438. G_UNICODE_SCRIPT_CYRILLIC, /* Cyrl (Cyrs) */
  439. G_UNICODE_SCRIPT_DESERET, /* Dsrt */
  440. G_UNICODE_SCRIPT_DEVANAGARI, /* Deva */
  441. G_UNICODE_SCRIPT_ETHIOPIC, /* Ethi */
  442. G_UNICODE_SCRIPT_GEORGIAN, /* Geor (Geon, Geoa) */
  443. G_UNICODE_SCRIPT_GOTHIC, /* Goth */
  444. G_UNICODE_SCRIPT_GREEK, /* Grek */
  445. G_UNICODE_SCRIPT_GUJARATI, /* Gujr */
  446. G_UNICODE_SCRIPT_GURMUKHI, /* Guru */
  447. G_UNICODE_SCRIPT_HAN, /* Hani */
  448. G_UNICODE_SCRIPT_HANGUL, /* Hang */
  449. G_UNICODE_SCRIPT_HEBREW, /* Hebr */
  450. G_UNICODE_SCRIPT_HIRAGANA, /* Hira */
  451. G_UNICODE_SCRIPT_KANNADA, /* Knda */
  452. G_UNICODE_SCRIPT_KATAKANA, /* Kana */
  453. G_UNICODE_SCRIPT_KHMER, /* Khmr */
  454. G_UNICODE_SCRIPT_LAO, /* Laoo */
  455. G_UNICODE_SCRIPT_LATIN, /* Latn (Latf, Latg) */
  456. G_UNICODE_SCRIPT_MALAYALAM, /* Mlym */
  457. G_UNICODE_SCRIPT_MONGOLIAN, /* Mong */
  458. G_UNICODE_SCRIPT_MYANMAR, /* Mymr */
  459. G_UNICODE_SCRIPT_OGHAM, /* Ogam */
  460. G_UNICODE_SCRIPT_OLD_ITALIC, /* Ital */
  461. G_UNICODE_SCRIPT_ORIYA, /* Orya */
  462. G_UNICODE_SCRIPT_RUNIC, /* Runr */
  463. G_UNICODE_SCRIPT_SINHALA, /* Sinh */
  464. G_UNICODE_SCRIPT_SYRIAC, /* Syrc (Syrj, Syrn, Syre) */
  465. G_UNICODE_SCRIPT_TAMIL, /* Taml */
  466. G_UNICODE_SCRIPT_TELUGU, /* Telu */
  467. G_UNICODE_SCRIPT_THAANA, /* Thaa */
  468. G_UNICODE_SCRIPT_THAI, /* Thai */
  469. G_UNICODE_SCRIPT_TIBETAN, /* Tibt */
  470. G_UNICODE_SCRIPT_CANADIAN_ABORIGINAL, /* Cans */
  471. G_UNICODE_SCRIPT_YI, /* Yiii */
  472. G_UNICODE_SCRIPT_TAGALOG, /* Tglg */
  473. G_UNICODE_SCRIPT_HANUNOO, /* Hano */
  474. G_UNICODE_SCRIPT_BUHID, /* Buhd */
  475. G_UNICODE_SCRIPT_TAGBANWA, /* Tagb */
  476. /* Unicode-4.0 additions */
  477. G_UNICODE_SCRIPT_BRAILLE, /* Brai */
  478. G_UNICODE_SCRIPT_CYPRIOT, /* Cprt */
  479. G_UNICODE_SCRIPT_LIMBU, /* Limb */
  480. G_UNICODE_SCRIPT_OSMANYA, /* Osma */
  481. G_UNICODE_SCRIPT_SHAVIAN, /* Shaw */
  482. G_UNICODE_SCRIPT_LINEAR_B, /* Linb */
  483. G_UNICODE_SCRIPT_TAI_LE, /* Tale */
  484. G_UNICODE_SCRIPT_UGARITIC, /* Ugar */
  485. /* Unicode-4.1 additions */
  486. G_UNICODE_SCRIPT_NEW_TAI_LUE, /* Talu */
  487. G_UNICODE_SCRIPT_BUGINESE, /* Bugi */
  488. G_UNICODE_SCRIPT_GLAGOLITIC, /* Glag */
  489. G_UNICODE_SCRIPT_TIFINAGH, /* Tfng */
  490. G_UNICODE_SCRIPT_SYLOTI_NAGRI, /* Sylo */
  491. G_UNICODE_SCRIPT_OLD_PERSIAN, /* Xpeo */
  492. G_UNICODE_SCRIPT_KHAROSHTHI, /* Khar */
  493. /* Unicode-5.0 additions */
  494. G_UNICODE_SCRIPT_UNKNOWN, /* Zzzz */
  495. G_UNICODE_SCRIPT_BALINESE, /* Bali */
  496. G_UNICODE_SCRIPT_CUNEIFORM, /* Xsux */
  497. G_UNICODE_SCRIPT_PHOENICIAN, /* Phnx */
  498. G_UNICODE_SCRIPT_PHAGS_PA, /* Phag */
  499. G_UNICODE_SCRIPT_NKO, /* Nkoo */
  500. /* Unicode-5.1 additions */
  501. G_UNICODE_SCRIPT_KAYAH_LI, /* Kali */
  502. G_UNICODE_SCRIPT_LEPCHA, /* Lepc */
  503. G_UNICODE_SCRIPT_REJANG, /* Rjng */
  504. G_UNICODE_SCRIPT_SUNDANESE, /* Sund */
  505. G_UNICODE_SCRIPT_SAURASHTRA, /* Saur */
  506. G_UNICODE_SCRIPT_CHAM, /* Cham */
  507. G_UNICODE_SCRIPT_OL_CHIKI, /* Olck */
  508. G_UNICODE_SCRIPT_VAI, /* Vaii */
  509. G_UNICODE_SCRIPT_CARIAN, /* Cari */
  510. G_UNICODE_SCRIPT_LYCIAN, /* Lyci */
  511. G_UNICODE_SCRIPT_LYDIAN, /* Lydi */
  512. /* Unicode-5.2 additions */
  513. G_UNICODE_SCRIPT_AVESTAN, /* Avst */
  514. G_UNICODE_SCRIPT_BAMUM, /* Bamu */
  515. G_UNICODE_SCRIPT_EGYPTIAN_HIEROGLYPHS, /* Egyp */
  516. G_UNICODE_SCRIPT_IMPERIAL_ARAMAIC, /* Armi */
  517. G_UNICODE_SCRIPT_INSCRIPTIONAL_PAHLAVI, /* Phli */
  518. G_UNICODE_SCRIPT_INSCRIPTIONAL_PARTHIAN, /* Prti */
  519. G_UNICODE_SCRIPT_JAVANESE, /* Java */
  520. G_UNICODE_SCRIPT_KAITHI, /* Kthi */
  521. G_UNICODE_SCRIPT_LISU, /* Lisu */
  522. G_UNICODE_SCRIPT_MEETEI_MAYEK, /* Mtei */
  523. G_UNICODE_SCRIPT_OLD_SOUTH_ARABIAN, /* Sarb */
  524. G_UNICODE_SCRIPT_OLD_TURKIC, /* Orkh */
  525. G_UNICODE_SCRIPT_SAMARITAN, /* Samr */
  526. G_UNICODE_SCRIPT_TAI_THAM, /* Lana */
  527. G_UNICODE_SCRIPT_TAI_VIET, /* Tavt */
  528. /* Unicode-6.0 additions */
  529. G_UNICODE_SCRIPT_BATAK, /* Batk */
  530. G_UNICODE_SCRIPT_BRAHMI, /* Brah */
  531. G_UNICODE_SCRIPT_MANDAIC, /* Mand */
  532. /* Unicode-6.1 additions */
  533. G_UNICODE_SCRIPT_CHAKMA, /* Cakm */
  534. G_UNICODE_SCRIPT_MEROITIC_CURSIVE, /* Merc */
  535. G_UNICODE_SCRIPT_MEROITIC_HIEROGLYPHS, /* Mero */
  536. G_UNICODE_SCRIPT_MIAO, /* Plrd */
  537. G_UNICODE_SCRIPT_SHARADA, /* Shrd */
  538. G_UNICODE_SCRIPT_SORA_SOMPENG, /* Sora */
  539. G_UNICODE_SCRIPT_TAKRI, /* Takr */
  540. /* Unicode 7.0 additions */
  541. G_UNICODE_SCRIPT_BASSA_VAH, /* Bass */
  542. G_UNICODE_SCRIPT_CAUCASIAN_ALBANIAN, /* Aghb */
  543. G_UNICODE_SCRIPT_DUPLOYAN, /* Dupl */
  544. G_UNICODE_SCRIPT_ELBASAN, /* Elba */
  545. G_UNICODE_SCRIPT_GRANTHA, /* Gran */
  546. G_UNICODE_SCRIPT_KHOJKI, /* Khoj */
  547. G_UNICODE_SCRIPT_KHUDAWADI, /* Sind */
  548. G_UNICODE_SCRIPT_LINEAR_A, /* Lina */
  549. G_UNICODE_SCRIPT_MAHAJANI, /* Mahj */
  550. G_UNICODE_SCRIPT_MANICHAEAN, /* Mani */
  551. G_UNICODE_SCRIPT_MENDE_KIKAKUI, /* Mend */
  552. G_UNICODE_SCRIPT_MODI, /* Modi */
  553. G_UNICODE_SCRIPT_MRO, /* Mroo */
  554. G_UNICODE_SCRIPT_NABATAEAN, /* Nbat */
  555. G_UNICODE_SCRIPT_OLD_NORTH_ARABIAN, /* Narb */
  556. G_UNICODE_SCRIPT_OLD_PERMIC, /* Perm */
  557. G_UNICODE_SCRIPT_PAHAWH_HMONG, /* Hmng */
  558. G_UNICODE_SCRIPT_PALMYRENE, /* Palm */
  559. G_UNICODE_SCRIPT_PAU_CIN_HAU, /* Pauc */
  560. G_UNICODE_SCRIPT_PSALTER_PAHLAVI, /* Phlp */
  561. G_UNICODE_SCRIPT_SIDDHAM, /* Sidd */
  562. G_UNICODE_SCRIPT_TIRHUTA, /* Tirh */
  563. G_UNICODE_SCRIPT_WARANG_CITI, /* Wara */
  564. /* Unicode 8.0 additions */
  565. G_UNICODE_SCRIPT_AHOM, /* Ahom */
  566. G_UNICODE_SCRIPT_ANATOLIAN_HIEROGLYPHS, /* Hluw */
  567. G_UNICODE_SCRIPT_HATRAN, /* Hatr */
  568. G_UNICODE_SCRIPT_MULTANI, /* Mult */
  569. G_UNICODE_SCRIPT_OLD_HUNGARIAN, /* Hung */
  570. G_UNICODE_SCRIPT_SIGNWRITING, /* Sgnw */
  571. /* Unicode 9.0 additions */
  572. G_UNICODE_SCRIPT_ADLAM, /* Adlm */
  573. G_UNICODE_SCRIPT_BHAIKSUKI, /* Bhks */
  574. G_UNICODE_SCRIPT_MARCHEN, /* Marc */
  575. G_UNICODE_SCRIPT_NEWA, /* Newa */
  576. G_UNICODE_SCRIPT_OSAGE, /* Osge */
  577. G_UNICODE_SCRIPT_TANGUT, /* Tang */
  578. /* Unicode 10.0 additions */
  579. G_UNICODE_SCRIPT_MASARAM_GONDI, /* Gonm */
  580. G_UNICODE_SCRIPT_NUSHU, /* Nshu */
  581. G_UNICODE_SCRIPT_SOYOMBO, /* Soyo */
  582. G_UNICODE_SCRIPT_ZANABAZAR_SQUARE, /* Zanb */
  583. /* Unicode 11.0 additions */
  584. G_UNICODE_SCRIPT_DOGRA, /* Dogr */
  585. G_UNICODE_SCRIPT_GUNJALA_GONDI, /* Gong */
  586. G_UNICODE_SCRIPT_HANIFI_ROHINGYA, /* Rohg */
  587. G_UNICODE_SCRIPT_MAKASAR, /* Maka */
  588. G_UNICODE_SCRIPT_MEDEFAIDRIN, /* Medf */
  589. G_UNICODE_SCRIPT_OLD_SOGDIAN, /* Sogo */
  590. G_UNICODE_SCRIPT_SOGDIAN, /* Sogd */
  591. /* Unicode 12.0 additions */
  592. G_UNICODE_SCRIPT_ELYMAIC, /* Elym */
  593. G_UNICODE_SCRIPT_NANDINAGARI, /* Nand */
  594. G_UNICODE_SCRIPT_NYIAKENG_PUACHUE_HMONG, /* Rohg */
  595. G_UNICODE_SCRIPT_WANCHO /* Wcho */
  596. } GUnicodeScript;
  597. GLIB_AVAILABLE_IN_ALL
  598. guint32 g_unicode_script_to_iso15924 (GUnicodeScript script);
  599. GLIB_AVAILABLE_IN_ALL
  600. GUnicodeScript g_unicode_script_from_iso15924 (guint32 iso15924);
  601. /* These are all analogs of the <ctype.h> functions.
  602. */
  603. GLIB_AVAILABLE_IN_ALL
  604. gboolean g_unichar_isalnum (gunichar c) G_GNUC_CONST;
  605. GLIB_AVAILABLE_IN_ALL
  606. gboolean g_unichar_isalpha (gunichar c) G_GNUC_CONST;
  607. GLIB_AVAILABLE_IN_ALL
  608. gboolean g_unichar_iscntrl (gunichar c) G_GNUC_CONST;
  609. GLIB_AVAILABLE_IN_ALL
  610. gboolean g_unichar_isdigit (gunichar c) G_GNUC_CONST;
  611. GLIB_AVAILABLE_IN_ALL
  612. gboolean g_unichar_isgraph (gunichar c) G_GNUC_CONST;
  613. GLIB_AVAILABLE_IN_ALL
  614. gboolean g_unichar_islower (gunichar c) G_GNUC_CONST;
  615. GLIB_AVAILABLE_IN_ALL
  616. gboolean g_unichar_isprint (gunichar c) G_GNUC_CONST;
  617. GLIB_AVAILABLE_IN_ALL
  618. gboolean g_unichar_ispunct (gunichar c) G_GNUC_CONST;
  619. GLIB_AVAILABLE_IN_ALL
  620. gboolean g_unichar_isspace (gunichar c) G_GNUC_CONST;
  621. GLIB_AVAILABLE_IN_ALL
  622. gboolean g_unichar_isupper (gunichar c) G_GNUC_CONST;
  623. GLIB_AVAILABLE_IN_ALL
  624. gboolean g_unichar_isxdigit (gunichar c) G_GNUC_CONST;
  625. GLIB_AVAILABLE_IN_ALL
  626. gboolean g_unichar_istitle (gunichar c) G_GNUC_CONST;
  627. GLIB_AVAILABLE_IN_ALL
  628. gboolean g_unichar_isdefined (gunichar c) G_GNUC_CONST;
  629. GLIB_AVAILABLE_IN_ALL
  630. gboolean g_unichar_iswide (gunichar c) G_GNUC_CONST;
  631. GLIB_AVAILABLE_IN_ALL
  632. gboolean g_unichar_iswide_cjk(gunichar c) G_GNUC_CONST;
  633. GLIB_AVAILABLE_IN_ALL
  634. gboolean g_unichar_iszerowidth(gunichar c) G_GNUC_CONST;
  635. GLIB_AVAILABLE_IN_ALL
  636. gboolean g_unichar_ismark (gunichar c) G_GNUC_CONST;
  637. /* More <ctype.h> functions. These convert between the three cases.
  638. * See the Unicode book to understand title case. */
  639. GLIB_AVAILABLE_IN_ALL
  640. gunichar g_unichar_toupper (gunichar c) G_GNUC_CONST;
  641. GLIB_AVAILABLE_IN_ALL
  642. gunichar g_unichar_tolower (gunichar c) G_GNUC_CONST;
  643. GLIB_AVAILABLE_IN_ALL
  644. gunichar g_unichar_totitle (gunichar c) G_GNUC_CONST;
  645. /* If C is a digit (according to 'g_unichar_isdigit'), then return its
  646. numeric value. Otherwise return -1. */
  647. GLIB_AVAILABLE_IN_ALL
  648. gint g_unichar_digit_value (gunichar c) G_GNUC_CONST;
  649. GLIB_AVAILABLE_IN_ALL
  650. gint g_unichar_xdigit_value (gunichar c) G_GNUC_CONST;
  651. /* Return the Unicode character type of a given character. */
  652. GLIB_AVAILABLE_IN_ALL
  653. GUnicodeType g_unichar_type (gunichar c) G_GNUC_CONST;
  654. /* Return the line break property for a given character */
  655. GLIB_AVAILABLE_IN_ALL
  656. GUnicodeBreakType g_unichar_break_type (gunichar c) G_GNUC_CONST;
  657. /* Returns the combining class for a given character */
  658. GLIB_AVAILABLE_IN_ALL
  659. gint g_unichar_combining_class (gunichar uc) G_GNUC_CONST;
  660. GLIB_AVAILABLE_IN_ALL
  661. gboolean g_unichar_get_mirror_char (gunichar ch,
  662. gunichar *mirrored_ch);
  663. GLIB_AVAILABLE_IN_ALL
  664. GUnicodeScript g_unichar_get_script (gunichar ch) G_GNUC_CONST;
  665. /* Validate a Unicode character */
  666. GLIB_AVAILABLE_IN_ALL
  667. gboolean g_unichar_validate (gunichar ch) G_GNUC_CONST;
  668. /* Pairwise canonical compose/decompose */
  669. GLIB_AVAILABLE_IN_ALL
  670. gboolean g_unichar_compose (gunichar a,
  671. gunichar b,
  672. gunichar *ch);
  673. GLIB_AVAILABLE_IN_ALL
  674. gboolean g_unichar_decompose (gunichar ch,
  675. gunichar *a,
  676. gunichar *b);
  677. GLIB_AVAILABLE_IN_ALL
  678. gsize g_unichar_fully_decompose (gunichar ch,
  679. gboolean compat,
  680. gunichar *result,
  681. gsize result_len);
  682. /**
  683. * G_UNICHAR_MAX_DECOMPOSITION_LENGTH:
  684. *
  685. * The maximum length (in codepoints) of a compatibility or canonical
  686. * decomposition of a single Unicode character.
  687. *
  688. * This is as defined by Unicode 6.1.
  689. *
  690. * Since: 2.32
  691. */
  692. #define G_UNICHAR_MAX_DECOMPOSITION_LENGTH 18 /* codepoints */
  693. /* Compute canonical ordering of a string in-place. This rearranges
  694. decomposed characters in the string according to their combining
  695. classes. See the Unicode manual for more information. */
  696. GLIB_AVAILABLE_IN_ALL
  697. void g_unicode_canonical_ordering (gunichar *string,
  698. gsize len);
  699. GLIB_DEPRECATED_IN_2_30
  700. gunichar *g_unicode_canonical_decomposition (gunichar ch,
  701. gsize *result_len) G_GNUC_MALLOC;
  702. /* Array of skip-bytes-per-initial character.
  703. */
  704. GLIB_VAR const gchar * const g_utf8_skip;
  705. /**
  706. * g_utf8_next_char:
  707. * @p: Pointer to the start of a valid UTF-8 character
  708. *
  709. * Skips to the next character in a UTF-8 string. The string must be
  710. * valid; this macro is as fast as possible, and has no error-checking.
  711. * You would use this macro to iterate over a string character by
  712. * character. The macro returns the start of the next UTF-8 character.
  713. * Before using this macro, use g_utf8_validate() to validate strings
  714. * that may contain invalid UTF-8.
  715. */
  716. #define g_utf8_next_char(p) (char *)((p) + g_utf8_skip[*(const guchar *)(p)])
  717. GLIB_AVAILABLE_IN_ALL
  718. gunichar g_utf8_get_char (const gchar *p) G_GNUC_PURE;
  719. GLIB_AVAILABLE_IN_ALL
  720. gunichar g_utf8_get_char_validated (const gchar *p,
  721. gssize max_len) G_GNUC_PURE;
  722. GLIB_AVAILABLE_IN_ALL
  723. gchar* g_utf8_offset_to_pointer (const gchar *str,
  724. glong offset) G_GNUC_PURE;
  725. GLIB_AVAILABLE_IN_ALL
  726. glong g_utf8_pointer_to_offset (const gchar *str,
  727. const gchar *pos) G_GNUC_PURE;
  728. GLIB_AVAILABLE_IN_ALL
  729. gchar* g_utf8_prev_char (const gchar *p) G_GNUC_PURE;
  730. GLIB_AVAILABLE_IN_ALL
  731. gchar* g_utf8_find_next_char (const gchar *p,
  732. const gchar *end) G_GNUC_PURE;
  733. GLIB_AVAILABLE_IN_ALL
  734. gchar* g_utf8_find_prev_char (const gchar *str,
  735. const gchar *p) G_GNUC_PURE;
  736. GLIB_AVAILABLE_IN_ALL
  737. glong g_utf8_strlen (const gchar *p,
  738. gssize max) G_GNUC_PURE;
  739. GLIB_AVAILABLE_IN_2_30
  740. gchar *g_utf8_substring (const gchar *str,
  741. glong start_pos,
  742. glong end_pos) G_GNUC_MALLOC;
  743. GLIB_AVAILABLE_IN_ALL
  744. gchar *g_utf8_strncpy (gchar *dest,
  745. const gchar *src,
  746. gsize n);
  747. /* Find the UTF-8 character corresponding to ch, in string p. These
  748. functions are equivalants to strchr and strrchr */
  749. GLIB_AVAILABLE_IN_ALL
  750. gchar* g_utf8_strchr (const gchar *p,
  751. gssize len,
  752. gunichar c);
  753. GLIB_AVAILABLE_IN_ALL
  754. gchar* g_utf8_strrchr (const gchar *p,
  755. gssize len,
  756. gunichar c);
  757. GLIB_AVAILABLE_IN_ALL
  758. gchar* g_utf8_strreverse (const gchar *str,
  759. gssize len);
  760. GLIB_AVAILABLE_IN_ALL
  761. gunichar2 *g_utf8_to_utf16 (const gchar *str,
  762. glong len,
  763. glong *items_read,
  764. glong *items_written,
  765. GError **error) G_GNUC_MALLOC;
  766. GLIB_AVAILABLE_IN_ALL
  767. gunichar * g_utf8_to_ucs4 (const gchar *str,
  768. glong len,
  769. glong *items_read,
  770. glong *items_written,
  771. GError **error) G_GNUC_MALLOC;
  772. GLIB_AVAILABLE_IN_ALL
  773. gunichar * g_utf8_to_ucs4_fast (const gchar *str,
  774. glong len,
  775. glong *items_written) G_GNUC_MALLOC;
  776. GLIB_AVAILABLE_IN_ALL
  777. gunichar * g_utf16_to_ucs4 (const gunichar2 *str,
  778. glong len,
  779. glong *items_read,
  780. glong *items_written,
  781. GError **error) G_GNUC_MALLOC;
  782. GLIB_AVAILABLE_IN_ALL
  783. gchar* g_utf16_to_utf8 (const gunichar2 *str,
  784. glong len,
  785. glong *items_read,
  786. glong *items_written,
  787. GError **error) G_GNUC_MALLOC;
  788. GLIB_AVAILABLE_IN_ALL
  789. gunichar2 *g_ucs4_to_utf16 (const gunichar *str,
  790. glong len,
  791. glong *items_read,
  792. glong *items_written,
  793. GError **error) G_GNUC_MALLOC;
  794. GLIB_AVAILABLE_IN_ALL
  795. gchar* g_ucs4_to_utf8 (const gunichar *str,
  796. glong len,
  797. glong *items_read,
  798. glong *items_written,
  799. GError **error) G_GNUC_MALLOC;
  800. GLIB_AVAILABLE_IN_ALL
  801. gint g_unichar_to_utf8 (gunichar c,
  802. gchar *outbuf);
  803. GLIB_AVAILABLE_IN_ALL
  804. gboolean g_utf8_validate (const gchar *str,
  805. gssize max_len,
  806. const gchar **end);
  807. GLIB_AVAILABLE_IN_2_60
  808. gboolean g_utf8_validate_len (const gchar *str,
  809. gsize max_len,
  810. const gchar **end);
  811. GLIB_AVAILABLE_IN_ALL
  812. gchar *g_utf8_strup (const gchar *str,
  813. gssize len) G_GNUC_MALLOC;
  814. GLIB_AVAILABLE_IN_ALL
  815. gchar *g_utf8_strdown (const gchar *str,
  816. gssize len) G_GNUC_MALLOC;
  817. GLIB_AVAILABLE_IN_ALL
  818. gchar *g_utf8_casefold (const gchar *str,
  819. gssize len) G_GNUC_MALLOC;
  820. /**
  821. * GNormalizeMode:
  822. * @G_NORMALIZE_DEFAULT: standardize differences that do not affect the
  823. * text content, such as the above-mentioned accent representation
  824. * @G_NORMALIZE_NFD: another name for %G_NORMALIZE_DEFAULT
  825. * @G_NORMALIZE_DEFAULT_COMPOSE: like %G_NORMALIZE_DEFAULT, but with
  826. * composed forms rather than a maximally decomposed form
  827. * @G_NORMALIZE_NFC: another name for %G_NORMALIZE_DEFAULT_COMPOSE
  828. * @G_NORMALIZE_ALL: beyond %G_NORMALIZE_DEFAULT also standardize the
  829. * "compatibility" characters in Unicode, such as SUPERSCRIPT THREE
  830. * to the standard forms (in this case DIGIT THREE). Formatting
  831. * information may be lost but for most text operations such
  832. * characters should be considered the same
  833. * @G_NORMALIZE_NFKD: another name for %G_NORMALIZE_ALL
  834. * @G_NORMALIZE_ALL_COMPOSE: like %G_NORMALIZE_ALL, but with composed
  835. * forms rather than a maximally decomposed form
  836. * @G_NORMALIZE_NFKC: another name for %G_NORMALIZE_ALL_COMPOSE
  837. *
  838. * Defines how a Unicode string is transformed in a canonical
  839. * form, standardizing such issues as whether a character with
  840. * an accent is represented as a base character and combining
  841. * accent or as a single precomposed character. Unicode strings
  842. * should generally be normalized before comparing them.
  843. */
  844. typedef enum {
  845. G_NORMALIZE_DEFAULT,
  846. G_NORMALIZE_NFD = G_NORMALIZE_DEFAULT,
  847. G_NORMALIZE_DEFAULT_COMPOSE,
  848. G_NORMALIZE_NFC = G_NORMALIZE_DEFAULT_COMPOSE,
  849. G_NORMALIZE_ALL,
  850. G_NORMALIZE_NFKD = G_NORMALIZE_ALL,
  851. G_NORMALIZE_ALL_COMPOSE,
  852. G_NORMALIZE_NFKC = G_NORMALIZE_ALL_COMPOSE
  853. } GNormalizeMode;
  854. GLIB_AVAILABLE_IN_ALL
  855. gchar *g_utf8_normalize (const gchar *str,
  856. gssize len,
  857. GNormalizeMode mode) G_GNUC_MALLOC;
  858. GLIB_AVAILABLE_IN_ALL
  859. gint g_utf8_collate (const gchar *str1,
  860. const gchar *str2) G_GNUC_PURE;
  861. GLIB_AVAILABLE_IN_ALL
  862. gchar *g_utf8_collate_key (const gchar *str,
  863. gssize len) G_GNUC_MALLOC;
  864. GLIB_AVAILABLE_IN_ALL
  865. gchar *g_utf8_collate_key_for_filename (const gchar *str,
  866. gssize len) G_GNUC_MALLOC;
  867. GLIB_AVAILABLE_IN_2_52
  868. gchar *g_utf8_make_valid (const gchar *str,
  869. gssize len) G_GNUC_MALLOC;
  870. G_END_DECLS
  871. #endif /* __G_UNICODE_H__ */