lib/unictype.in.h

   1 /* Unicode character classification and properties.
   2    Copyright (C) 2002, 2005-2020 Free Software Foundation, Inc.
   3
   4    This program is free software: you can redistribute it and/or modify it
   5    under the terms of the GNU Lesser General Public License as published
   6    by 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 GNU
  12    Lesser General Public License for more details.
  13
  14    You should have received a copy of the GNU Lesser General Public License
  15    along with this program.  If not, see <https://www.gnu.org/licenses/>.  */
  16
  17 #ifndef _UNICTYPE_H
  18 #define _UNICTYPE_H
  19
  20 #include "unitypes.h"
  21
  22 /* Get bool.  */
  23 #include <stdbool.h>
  24
  25 /* Get size_t.  */
  26 #include <stddef.h>
  27
  28 #ifdef __cplusplus
  29 extern "C" {
  30 #endif
  31
  32 /* ========================================================================= */
  33
  34 /* Field 1 of Unicode Character Database: Character name.
  35    See "uniname.h".  */
  36
  37 /* ========================================================================= */
  38
  39 /* Field 2 of Unicode Character Database: General category.  */
  40
  41 /* Data type denoting a General category value.  This is not just a bitmask,
  42    but rather a bitmask and a pointer to the lookup table, so that programs
  43    that use only the predefined bitmasks (i.e. don't combine bitmasks with &
  44    and |) don't have a link-time dependency towards the big general table.  */
  45 typedef struct
  46 {
  47   uint32_t bitmask : 31;
  48   /*bool*/ unsigned int generic : 1;
  49   union
  50   {
  51     const void *table;                               /* when generic is 0 */
  52     bool (*lookup_fn) (ucs4_t uc, uint32_t bitmask); /* when generic is 1 */
  53   } lookup;
  54 }
  55 uc_general_category_t;
  56
  57 /* Bits and bit masks denoting General category values.  UnicodeData-3.2.0.html
  58    says a 32-bit integer will always suffice to represent them.
  59    These bit masks can only be used with the uc_is_general_category_withtable
  60    function.  */
  61 enum
  62 {
  63   UC_CATEGORY_MASK_L  = 0x0000001f,
  64   UC_CATEGORY_MASK_LC = 0x00000007,
  65   UC_CATEGORY_MASK_Lu = 0x00000001,
  66   UC_CATEGORY_MASK_Ll = 0x00000002,
  67   UC_CATEGORY_MASK_Lt = 0x00000004,
  68   UC_CATEGORY_MASK_Lm = 0x00000008,
  69   UC_CATEGORY_MASK_Lo = 0x00000010,
  70   UC_CATEGORY_MASK_M  = 0x000000e0,
  71   UC_CATEGORY_MASK_Mn = 0x00000020,
  72   UC_CATEGORY_MASK_Mc = 0x00000040,
  73   UC_CATEGORY_MASK_Me = 0x00000080,
  74   UC_CATEGORY_MASK_N  = 0x00000700,
  75   UC_CATEGORY_MASK_Nd = 0x00000100,
  76   UC_CATEGORY_MASK_Nl = 0x00000200,
  77   UC_CATEGORY_MASK_No = 0x00000400,
  78   UC_CATEGORY_MASK_P  = 0x0003f800,
  79   UC_CATEGORY_MASK_Pc = 0x00000800,
  80   UC_CATEGORY_MASK_Pd = 0x00001000,
  81   UC_CATEGORY_MASK_Ps = 0x00002000,
  82   UC_CATEGORY_MASK_Pe = 0x00004000,
  83   UC_CATEGORY_MASK_Pi = 0x00008000,
  84   UC_CATEGORY_MASK_Pf = 0x00010000,
  85   UC_CATEGORY_MASK_Po = 0x00020000,
  86   UC_CATEGORY_MASK_S  = 0x003c0000,
  87   UC_CATEGORY_MASK_Sm = 0x00040000,
  88   UC_CATEGORY_MASK_Sc = 0x00080000,
  89   UC_CATEGORY_MASK_Sk = 0x00100000,
  90   UC_CATEGORY_MASK_So = 0x00200000,
  91   UC_CATEGORY_MASK_Z  = 0x01c00000,
  92   UC_CATEGORY_MASK_Zs = 0x00400000,
  93   UC_CATEGORY_MASK_Zl = 0x00800000,
  94   UC_CATEGORY_MASK_Zp = 0x01000000,
  95   UC_CATEGORY_MASK_C  = 0x3e000000,
  96   UC_CATEGORY_MASK_Cc = 0x02000000,
  97   UC_CATEGORY_MASK_Cf = 0x04000000,
  98   UC_CATEGORY_MASK_Cs = 0x08000000,
  99   UC_CATEGORY_MASK_Co = 0x10000000,
 100   UC_CATEGORY_MASK_Cn = 0x20000000
 101 };
 102
 103 /* Predefined General category values.  */
 104 extern const uc_general_category_t UC_CATEGORY_L;
 105 extern const uc_general_category_t UC_CATEGORY_LC;
 106 extern const uc_general_category_t UC_CATEGORY_Lu;
 107 extern const uc_general_category_t UC_CATEGORY_Ll;
 108 extern const uc_general_category_t UC_CATEGORY_Lt;
 109 extern const uc_general_category_t UC_CATEGORY_Lm;
 110 extern const uc_general_category_t UC_CATEGORY_Lo;
 111 extern const uc_general_category_t UC_CATEGORY_M;
 112 extern const uc_general_category_t UC_CATEGORY_Mn;
 113 extern const uc_general_category_t UC_CATEGORY_Mc;
 114 extern const uc_general_category_t UC_CATEGORY_Me;
 115 extern const uc_general_category_t UC_CATEGORY_N;
 116 extern const uc_general_category_t UC_CATEGORY_Nd;
 117 extern const uc_general_category_t UC_CATEGORY_Nl;
 118 extern const uc_general_category_t UC_CATEGORY_No;
 119 extern const uc_general_category_t UC_CATEGORY_P;
 120 extern const uc_general_category_t UC_CATEGORY_Pc;
 121 extern const uc_general_category_t UC_CATEGORY_Pd;
 122 extern const uc_general_category_t UC_CATEGORY_Ps;
 123 extern const uc_general_category_t UC_CATEGORY_Pe;
 124 extern const uc_general_category_t UC_CATEGORY_Pi;
 125 extern const uc_general_category_t UC_CATEGORY_Pf;
 126 extern const uc_general_category_t UC_CATEGORY_Po;
 127 extern const uc_general_category_t UC_CATEGORY_S;
 128 extern const uc_general_category_t UC_CATEGORY_Sm;
 129 extern const uc_general_category_t UC_CATEGORY_Sc;
 130 extern const uc_general_category_t UC_CATEGORY_Sk;
 131 extern const uc_general_category_t UC_CATEGORY_So;
 132 extern const uc_general_category_t UC_CATEGORY_Z;
 133 extern const uc_general_category_t UC_CATEGORY_Zs;
 134 extern const uc_general_category_t UC_CATEGORY_Zl;
 135 extern const uc_general_category_t UC_CATEGORY_Zp;
 136 extern const uc_general_category_t UC_CATEGORY_C;
 137 extern const uc_general_category_t UC_CATEGORY_Cc;
 138 extern const uc_general_category_t UC_CATEGORY_Cf;
 139 extern const uc_general_category_t UC_CATEGORY_Cs;
 140 extern const uc_general_category_t UC_CATEGORY_Co;
 141 extern const uc_general_category_t UC_CATEGORY_Cn;
 142 /* Non-public.  */
 143 extern const uc_general_category_t _UC_CATEGORY_NONE;
 144
 145 /* Alias names for predefined General category values.  */
 146 #define UC_LETTER                    UC_CATEGORY_L
 147 #define UC_CASED_LETTER              UC_CATEGORY_LC
 148 #define UC_UPPERCASE_LETTER          UC_CATEGORY_Lu
 149 #define UC_LOWERCASE_LETTER          UC_CATEGORY_Ll
 150 #define UC_TITLECASE_LETTER          UC_CATEGORY_Lt
 151 #define UC_MODIFIER_LETTER           UC_CATEGORY_Lm
 152 #define UC_OTHER_LETTER              UC_CATEGORY_Lo
 153 #define UC_MARK                      UC_CATEGORY_M
 154 #define UC_NON_SPACING_MARK          UC_CATEGORY_Mn
 155 #define UC_COMBINING_SPACING_MARK    UC_CATEGORY_Mc
 156 #define UC_ENCLOSING_MARK            UC_CATEGORY_Me
 157 #define UC_NUMBER                    UC_CATEGORY_N
 158 #define UC_DECIMAL_DIGIT_NUMBER      UC_CATEGORY_Nd
 159 #define UC_LETTER_NUMBER             UC_CATEGORY_Nl
 160 #define UC_OTHER_NUMBER              UC_CATEGORY_No
 161 #define UC_PUNCTUATION               UC_CATEGORY_P
 162 #define UC_CONNECTOR_PUNCTUATION     UC_CATEGORY_Pc
 163 #define UC_DASH_PUNCTUATION          UC_CATEGORY_Pd
 164 #define UC_OPEN_PUNCTUATION          UC_CATEGORY_Ps /* a.k.a. UC_START_PUNCTUATION */
 165 #define UC_CLOSE_PUNCTUATION         UC_CATEGORY_Pe /* a.k.a. UC_END_PUNCTUATION */
 166 #define UC_INITIAL_QUOTE_PUNCTUATION UC_CATEGORY_Pi
 167 #define UC_FINAL_QUOTE_PUNCTUATION   UC_CATEGORY_Pf
 168 #define UC_OTHER_PUNCTUATION         UC_CATEGORY_Po
 169 #define UC_SYMBOL                    UC_CATEGORY_S
 170 #define UC_MATH_SYMBOL               UC_CATEGORY_Sm
 171 #define UC_CURRENCY_SYMBOL           UC_CATEGORY_Sc
 172 #define UC_MODIFIER_SYMBOL           UC_CATEGORY_Sk
 173 #define UC_OTHER_SYMBOL              UC_CATEGORY_So
 174 #define UC_SEPARATOR                 UC_CATEGORY_Z
 175 #define UC_SPACE_SEPARATOR           UC_CATEGORY_Zs
 176 #define UC_LINE_SEPARATOR            UC_CATEGORY_Zl
 177 #define UC_PARAGRAPH_SEPARATOR       UC_CATEGORY_Zp
 178 #define UC_OTHER                     UC_CATEGORY_C
 179 #define UC_CONTROL                   UC_CATEGORY_Cc
 180 #define UC_FORMAT                    UC_CATEGORY_Cf
 181 #define UC_SURROGATE                 UC_CATEGORY_Cs /* all of them are invalid characters */
 182 #define UC_PRIVATE_USE               UC_CATEGORY_Co
 183 #define UC_UNASSIGNED                UC_CATEGORY_Cn /* some of them are invalid characters */
 184
 185 /* Return the union of two general categories.
 186    This corresponds to the unions of the two sets of characters.  */
 187 extern uc_general_category_t
 188        uc_general_category_or (uc_general_category_t category1,
 189                                uc_general_category_t category2);
 190
 191 /* Return the intersection of two general categories as bit masks.
 192    This *does*not* correspond to the intersection of the two sets of
 193    characters.  */
 194 extern uc_general_category_t
 195        uc_general_category_and (uc_general_category_t category1,
 196                                 uc_general_category_t category2);
 197
 198 /* Return the intersection of a general category with the complement of a
 199    second general category, as bit masks.
 200    This *does*not* correspond to the intersection with complement, when
 201    viewing the categories as sets of characters.  */
 202 extern uc_general_category_t
 203        uc_general_category_and_not (uc_general_category_t category1,
 204                                     uc_general_category_t category2);
 205
 206 /* Return the name of a general category.  */
 207 extern const char *
 208        uc_general_category_name (uc_general_category_t category)
 209        _UC_ATTRIBUTE_PURE;
 210
 211 /* Return the long name of a general category.  */
 212 extern const char *
 213        uc_general_category_long_name (uc_general_category_t category)
 214        _UC_ATTRIBUTE_PURE;
 215
 216 /* Return the general category given by name, e.g. "Lu", or by long name,
 217    e.g. "Uppercase Letter".  */
 218 extern uc_general_category_t
 219        uc_general_category_byname (const char *category_name)
 220        _UC_ATTRIBUTE_PURE;
 221
 222 /* Return the general category of a Unicode character.  */
 223 extern uc_general_category_t
 224        uc_general_category (ucs4_t uc)
 225        _UC_ATTRIBUTE_PURE;
 226
 227 /* Test whether a Unicode character belongs to a given category.
 228    The CATEGORY argument can be the combination of several predefined
 229    general categories.  */
 230 extern bool
 231        uc_is_general_category (ucs4_t uc, uc_general_category_t category)
 232        _UC_ATTRIBUTE_PURE;
 233 /* Likewise.  This function uses a big table comprising all categories.  */
 234 extern bool
 235        uc_is_general_category_withtable (ucs4_t uc, uint32_t bitmask)
 236        _UC_ATTRIBUTE_CONST;
 237
 238 /* ========================================================================= */
 239
 240 /* Field 3 of Unicode Character Database: Canonical combining class.  */
 241
 242 /* The possible results of uc_combining_class (0..255) are described in
 243    UCD.html.  The list here is not definitive; more values can be added
 244    in future versions.  */
 245 enum
 246 {
 247   UC_CCC_NR   =   0, /* Not Reordered */
 248   UC_CCC_OV   =   1, /* Overlay */
 249   UC_CCC_NK   =   7, /* Nukta */
 250   UC_CCC_KV   =   8, /* Kana Voicing */
 251   UC_CCC_VR   =   9, /* Virama */
 252   UC_CCC_ATBL = 200, /* Attached Below Left */
 253   UC_CCC_ATB  = 202, /* Attached Below */
 254   UC_CCC_ATA  = 214, /* Attached Above */
 255   UC_CCC_ATAR = 216, /* Attached Above Right */
 256   UC_CCC_BL   = 218, /* Below Left */
 257   UC_CCC_B    = 220, /* Below */
 258   UC_CCC_BR   = 222, /* Below Right */
 259   UC_CCC_L    = 224, /* Left */
 260   UC_CCC_R    = 226, /* Right */
 261   UC_CCC_AL   = 228, /* Above Left */
 262   UC_CCC_A    = 230, /* Above */
 263   UC_CCC_AR   = 232, /* Above Right */
 264   UC_CCC_DB   = 233, /* Double Below */
 265   UC_CCC_DA   = 234, /* Double Above */
 266   UC_CCC_IS   = 240  /* Iota Subscript */
 267 };
 268
 269 /* Return the canonical combining class of a Unicode character.  */
 270 extern int
 271        uc_combining_class (ucs4_t uc)
 272        _UC_ATTRIBUTE_CONST;
 273
 274 /* Return the name of a canonical combining class.  */
 275 extern const char *
 276        uc_combining_class_name (int ccc)
 277        _UC_ATTRIBUTE_CONST;
 278
 279 /* Return the long name of a canonical combining class.  */
 280 extern const char *
 281        uc_combining_class_long_name (int ccc)
 282        _UC_ATTRIBUTE_CONST;
 283
 284 /* Return the canonical combining class given by name, e.g. "BL", or by long
 285    name, e.g. "Below Left".  */
 286 extern int
 287        uc_combining_class_byname (const char *ccc_name)
 288        _UC_ATTRIBUTE_PURE;
 289
 290 /* ========================================================================= */
 291
 292 /* Field 4 of Unicode Character Database: Bidi class.
 293    Before Unicode 4.0, this field was called "Bidirectional category".  */
 294
 295 enum
 296 {
 297   UC_BIDI_L,   /* Left-to-Right */
 298   UC_BIDI_LRE, /* Left-to-Right Embedding */
 299   UC_BIDI_LRO, /* Left-to-Right Override */
 300   UC_BIDI_R,   /* Right-to-Left */
 301   UC_BIDI_AL,  /* Right-to-Left Arabic */
 302   UC_BIDI_RLE, /* Right-to-Left Embedding */
 303   UC_BIDI_RLO, /* Right-to-Left Override */
 304   UC_BIDI_PDF, /* Pop Directional Format */
 305   UC_BIDI_EN,  /* European Number */
 306   UC_BIDI_ES,  /* European Number Separator */
 307   UC_BIDI_ET,  /* European Number Terminator */
 308   UC_BIDI_AN,  /* Arabic Number */
 309   UC_BIDI_CS,  /* Common Number Separator */
 310   UC_BIDI_NSM, /* Non-Spacing Mark */
 311   UC_BIDI_BN,  /* Boundary Neutral */
 312   UC_BIDI_B,   /* Paragraph Separator */
 313   UC_BIDI_S,   /* Segment Separator */
 314   UC_BIDI_WS,  /* Whitespace */
 315   UC_BIDI_ON,  /* Other Neutral */
 316   UC_BIDI_LRI, /* Left-to-Right Isolate */
 317   UC_BIDI_RLI, /* Right-to-Left Isolate */
 318   UC_BIDI_FSI, /* First Strong Isolate */
 319   UC_BIDI_PDI  /* Pop Directional Isolate */
 320 };
 321
 322 /* Return the name of a bidi class.  */
 323 extern const char *
 324        uc_bidi_class_name (int bidi_class)
 325        _UC_ATTRIBUTE_CONST;
 326 /* Same; obsolete function name.  */
 327 extern const char *
 328        uc_bidi_category_name (int category)
 329        _UC_ATTRIBUTE_CONST;
 330
 331 /* Return the long name of a bidi class.  */
 332 extern const char *
 333        uc_bidi_class_long_name (int bidi_class)
 334        _UC_ATTRIBUTE_CONST;
 335
 336 /* Return the bidi class given by name, e.g. "LRE", or by long name, e.g.
 337    "Left-to-Right Embedding".  */
 338 extern int
 339        uc_bidi_class_byname (const char *bidi_class_name)
 340        _UC_ATTRIBUTE_PURE;
 341 /* Same; obsolete function name.  */
 342 extern int
 343        uc_bidi_category_byname (const char *category_name)
 344        _UC_ATTRIBUTE_PURE;
 345
 346 /* Return the bidi class of a Unicode character.  */
 347 extern int
 348        uc_bidi_class (ucs4_t uc)
 349        _UC_ATTRIBUTE_CONST;
 350 /* Same; obsolete function name.  */
 351 extern int
 352        uc_bidi_category (ucs4_t uc)
 353        _UC_ATTRIBUTE_CONST;
 354
 355 /* Test whether a Unicode character belongs to a given bidi class.  */
 356 extern bool
 357        uc_is_bidi_class (ucs4_t uc, int bidi_class)
 358        _UC_ATTRIBUTE_CONST;
 359 /* Same; obsolete function name.  */
 360 extern bool
 361        uc_is_bidi_category (ucs4_t uc, int category)
 362        _UC_ATTRIBUTE_CONST;
 363
 364 /* ========================================================================= */
 365
 366 /* Field 5 of Unicode Character Database: Character decomposition mapping.
 367    See "uninorm.h".  */
 368
 369 /* ========================================================================= */
 370
 371 /* Field 6 of Unicode Character Database: Decimal digit value.  */
 372
 373 /* Return the decimal digit value of a Unicode character.  */
 374 extern int
 375        uc_decimal_value (ucs4_t uc)
 376        _UC_ATTRIBUTE_CONST;
 377
 378 /* ========================================================================= */
 379
 380 /* Field 7 of Unicode Character Database: Digit value.  */
 381
 382 /* Return the digit value of a Unicode character.  */
 383 extern int
 384        uc_digit_value (ucs4_t uc)
 385        _UC_ATTRIBUTE_CONST;
 386
 387 /* ========================================================================= */
 388
 389 /* Field 8 of Unicode Character Database: Numeric value.  */
 390
 391 /* Return the numeric value of a Unicode character.  */
 392 typedef struct
 393 {
 394   int numerator;
 395   int denominator;
 396 }
 397 uc_fraction_t;
 398 extern uc_fraction_t
 399        uc_numeric_value (ucs4_t uc)
 400        _UC_ATTRIBUTE_CONST;
 401
 402 /* ========================================================================= */
 403
 404 /* Field 9 of Unicode Character Database: Mirrored.  */
 405
 406 /* Return the mirrored character of a Unicode character UC in *PUC.  */
 407 extern bool
 408        uc_mirror_char (ucs4_t uc, ucs4_t *puc);
 409
 410 /* ========================================================================= */
 411
 412 /* Field 10 of Unicode Character Database: Unicode 1.0 Name.
 413    Not available in this library.  */
 414
 415 /* ========================================================================= */
 416
 417 /* Field 11 of Unicode Character Database: ISO 10646 comment.
 418    Not available in this library.  */
 419
 420 /* ========================================================================= */
 421
 422 /* Field 12, 13, 14 of Unicode Character Database: Uppercase mapping,
 423    lowercase mapping, titlecase mapping.  See "unicase.h".  */
 424
 425 /* ========================================================================= */
 426
 427 /* Field 2 of the file ArabicShaping.txt in the Unicode Character Database.  */
 428
 429 /* Possible joining types.  */
 430 enum
 431 {
 432   UC_JOINING_TYPE_U, /* Non_Joining */
 433   UC_JOINING_TYPE_T, /* Transparent */
 434   UC_JOINING_TYPE_C, /* Join_Causing */
 435   UC_JOINING_TYPE_L, /* Left_Joining */
 436   UC_JOINING_TYPE_R, /* Right_Joining */
 437   UC_JOINING_TYPE_D  /* Dual_Joining */
 438 };
 439
 440 /* Return the name of a joining type.  */
 441 extern const char *
 442        uc_joining_type_name (int joining_type)
 443        _UC_ATTRIBUTE_CONST;
 444
 445 /* Return the long name of a joining type.  */
 446 extern const char *
 447        uc_joining_type_long_name (int joining_type)
 448        _UC_ATTRIBUTE_CONST;
 449
 450 /* Return the joining type given by name, e.g. "D", or by long name, e.g.
 451    "Dual Joining".  */
 452 extern int
 453        uc_joining_type_byname (const char *joining_type_name)
 454        _UC_ATTRIBUTE_PURE;
 455
 456 /* Return the joining type of a Unicode character.  */
 457 extern int
 458        uc_joining_type (ucs4_t uc)
 459        _UC_ATTRIBUTE_CONST;
 460
 461 /* ========================================================================= */
 462
 463 /* Field 3 of the file ArabicShaping.txt in the Unicode Character Database.  */
 464
 465 /* Possible joining groups.
 466    This enumeration may be extended in the future.  */
 467 enum
 468 {
 469   UC_JOINING_GROUP_NONE,                  /* No_Joining_Group */
 470   UC_JOINING_GROUP_AIN,                   /* Ain */
 471   UC_JOINING_GROUP_ALAPH,                 /* Alaph */
 472   UC_JOINING_GROUP_ALEF,                  /* Alef */
 473   UC_JOINING_GROUP_BEH,                   /* Beh */
 474   UC_JOINING_GROUP_BETH,                  /* Beth */
 475   UC_JOINING_GROUP_BURUSHASKI_YEH_BARREE, /* Burushaski_Yeh_Barree */
 476   UC_JOINING_GROUP_DAL,                   /* Dal */
 477   UC_JOINING_GROUP_DALATH_RISH,           /* Dalath_Rish */
 478   UC_JOINING_GROUP_E,                     /* E */
 479   UC_JOINING_GROUP_FARSI_YEH,             /* Farsi_Yeh */
 480   UC_JOINING_GROUP_FE,                    /* Fe */
 481   UC_JOINING_GROUP_FEH,                   /* Feh */
 482   UC_JOINING_GROUP_FINAL_SEMKATH,         /* Final_Semkath */
 483   UC_JOINING_GROUP_GAF,                   /* Gaf */
 484   UC_JOINING_GROUP_GAMAL,                 /* Gamal */
 485   UC_JOINING_GROUP_HAH,                   /* Hah */
 486   UC_JOINING_GROUP_HE,                    /* He */
 487   UC_JOINING_GROUP_HEH,                   /* Heh */
 488   UC_JOINING_GROUP_HEH_GOAL,              /* Heh_Goal */
 489   UC_JOINING_GROUP_HETH,                  /* Heth */
 490   UC_JOINING_GROUP_KAF,                   /* Kaf */
 491   UC_JOINING_GROUP_KAPH,                  /* Kaph */
 492   UC_JOINING_GROUP_KHAPH,                 /* Khaph */
 493   UC_JOINING_GROUP_KNOTTED_HEH,           /* Knotted_Heh */
 494   UC_JOINING_GROUP_LAM,                   /* Lam */
 495   UC_JOINING_GROUP_LAMADH,                /* Lamadh */
 496   UC_JOINING_GROUP_MEEM,                  /* Meem */
 497   UC_JOINING_GROUP_MIM,                   /* Mim */
 498   UC_JOINING_GROUP_NOON,                  /* Noon */
 499   UC_JOINING_GROUP_NUN,                   /* Nun */
 500   UC_JOINING_GROUP_NYA,                   /* Nya */
 501   UC_JOINING_GROUP_PE,                    /* Pe */
 502   UC_JOINING_GROUP_QAF,                   /* Qaf */
 503   UC_JOINING_GROUP_QAPH,                  /* Qaph */
 504   UC_JOINING_GROUP_REH,                   /* Reh */
 505   UC_JOINING_GROUP_REVERSED_PE,           /* Reversed_Pe */
 506   UC_JOINING_GROUP_SAD,                   /* Sad */
 507   UC_JOINING_GROUP_SADHE,                 /* Sadhe */
 508   UC_JOINING_GROUP_SEEN,                  /* Seen */
 509   UC_JOINING_GROUP_SEMKATH,               /* Semkath */
 510   UC_JOINING_GROUP_SHIN,                  /* Shin */
 511   UC_JOINING_GROUP_SWASH_KAF,             /* Swash_Kaf */
 512   UC_JOINING_GROUP_SYRIAC_WAW,            /* Syriac_Waw */
 513   UC_JOINING_GROUP_TAH,                   /* Tah */
 514   UC_JOINING_GROUP_TAW,                   /* Taw */
 515   UC_JOINING_GROUP_TEH_MARBUTA,           /* Teh_Marbuta */
 516   UC_JOINING_GROUP_TEH_MARBUTA_GOAL,      /* Teh_Marbuta_Goal */
 517   UC_JOINING_GROUP_TETH,                  /* Teth */
 518   UC_JOINING_GROUP_WAW,                   /* Waw */
 519   UC_JOINING_GROUP_YEH,                   /* Yeh */
 520   UC_JOINING_GROUP_YEH_BARREE,            /* Yeh_Barree */
 521   UC_JOINING_GROUP_YEH_WITH_TAIL,         /* Yeh_With_Tail */
 522   UC_JOINING_GROUP_YUDH,                  /* Yudh */
 523   UC_JOINING_GROUP_YUDH_HE,               /* Yudh_He */
 524   UC_JOINING_GROUP_ZAIN,                  /* Zain */
 525   UC_JOINING_GROUP_ZHAIN,                 /* Zhain */
 526   UC_JOINING_GROUP_ROHINGYA_YEH,          /* Rohingya_Yeh */
 527   UC_JOINING_GROUP_STRAIGHT_WAW,          /* Straight_Waw */
 528   UC_JOINING_GROUP_MANICHAEAN_ALEPH,      /* Manichaean_Aleph */
 529   UC_JOINING_GROUP_MANICHAEAN_BETH,       /* Manichaean_Beth */
 530   UC_JOINING_GROUP_MANICHAEAN_GIMEL,      /* Manichaean_Gimel */
 531   UC_JOINING_GROUP_MANICHAEAN_DALETH,     /* Manichaean_Daleth */
 532   UC_JOINING_GROUP_MANICHAEAN_WAW,        /* Manichaean_Waw */
 533   UC_JOINING_GROUP_MANICHAEAN_ZAYIN,      /* Manichaean_Zayin */
 534   UC_JOINING_GROUP_MANICHAEAN_HETH,       /* Manichaean_Heth */
 535   UC_JOINING_GROUP_MANICHAEAN_TETH,       /* Manichaean_Teth */
 536   UC_JOINING_GROUP_MANICHAEAN_YODH,       /* Manichaean_Yodh */
 537   UC_JOINING_GROUP_MANICHAEAN_KAPH,       /* Manichaean_Kaph */
 538   UC_JOINING_GROUP_MANICHAEAN_LAMEDH,     /* Manichaean_Lamedh */
 539   UC_JOINING_GROUP_MANICHAEAN_DHAMEDH,    /* Manichaean_Dhamedh */
 540   UC_JOINING_GROUP_MANICHAEAN_THAMEDH,    /* Manichaean_Thamedh */
 541   UC_JOINING_GROUP_MANICHAEAN_MEM,        /* Manichaean_Mem */
 542   UC_JOINING_GROUP_MANICHAEAN_NUN,        /* Manichaean_Nun */
 543   UC_JOINING_GROUP_MANICHAEAN_SAMEKH,     /* Manichaean_Aleph */
 544   UC_JOINING_GROUP_MANICHAEAN_AYIN,       /* Manichaean_Ayin */
 545   UC_JOINING_GROUP_MANICHAEAN_PE,         /* Manichaean_Pe */
 546   UC_JOINING_GROUP_MANICHAEAN_SADHE,      /* Manichaean_Sadhe */
 547   UC_JOINING_GROUP_MANICHAEAN_QOPH,       /* Manichaean_Qoph */
 548   UC_JOINING_GROUP_MANICHAEAN_RESH,       /* Manichaean_Resh */
 549   UC_JOINING_GROUP_MANICHAEAN_TAW,        /* Manichaean_Taw */
 550   UC_JOINING_GROUP_MANICHAEAN_ONE,        /* Manichaean_One */
 551   UC_JOINING_GROUP_MANICHAEAN_FIVE,       /* Manichaean_Five */
 552   UC_JOINING_GROUP_MANICHAEAN_TEN,        /* Manichaean_Ten */
 553   UC_JOINING_GROUP_MANICHAEAN_TWENTY,     /* Manichaean_Twenty */
 554   UC_JOINING_GROUP_MANICHAEAN_HUNDRED,    /* Manichaean_Hundred */
 555   UC_JOINING_GROUP_AFRICAN_FEH,           /* African_Feh */
 556   UC_JOINING_GROUP_AFRICAN_QAF,           /* African_Qaf */
 557   UC_JOINING_GROUP_AFRICAN_NOON           /* African_Noon */
 558 };
 559
 560 /* Return the name of a joining group.  */
 561 extern const char *
 562        uc_joining_group_name (int joining_group)
 563        _UC_ATTRIBUTE_CONST;
 564
 565 /* Return the joining group given by name, e.g. "Teh_Marbuta".  */
 566 extern int
 567        uc_joining_group_byname (const char *joining_group_name)
 568        _UC_ATTRIBUTE_PURE;
 569
 570 /* Return the joining group of a Unicode character.  */
 571 extern int
 572        uc_joining_group (ucs4_t uc)
 573        _UC_ATTRIBUTE_CONST;
 574
 575 /* ========================================================================= */
 576
 577 /* Common API for properties.  */
 578
 579 /* Data type denoting a property.  This is not just a number, but rather a
 580    pointer to the test functions, so that programs that use only few of the
 581    properties don't have a link-time dependency towards all the tables.  */
 582 typedef struct
 583 {
 584   bool (*test_fn) (ucs4_t uc);
 585 }
 586 uc_property_t;
 587
 588 /* Predefined properties.  */
 589 /* General.  */
 590 extern const uc_property_t UC_PROPERTY_WHITE_SPACE;
 591 extern const uc_property_t UC_PROPERTY_ALPHABETIC;
 592 extern const uc_property_t UC_PROPERTY_OTHER_ALPHABETIC;
 593 extern const uc_property_t UC_PROPERTY_NOT_A_CHARACTER;
 594 extern const uc_property_t UC_PROPERTY_DEFAULT_IGNORABLE_CODE_POINT;
 595 extern const uc_property_t UC_PROPERTY_OTHER_DEFAULT_IGNORABLE_CODE_POINT;
 596 extern const uc_property_t UC_PROPERTY_DEPRECATED;
 597 extern const uc_property_t UC_PROPERTY_LOGICAL_ORDER_EXCEPTION;
 598 extern const uc_property_t UC_PROPERTY_VARIATION_SELECTOR;
 599 extern const uc_property_t UC_PROPERTY_PRIVATE_USE;
 600 extern const uc_property_t UC_PROPERTY_UNASSIGNED_CODE_VALUE;
 601 /* Case.  */
 602 extern const uc_property_t UC_PROPERTY_UPPERCASE;
 603 extern const uc_property_t UC_PROPERTY_OTHER_UPPERCASE;
 604 extern const uc_property_t UC_PROPERTY_LOWERCASE;
 605 extern const uc_property_t UC_PROPERTY_OTHER_LOWERCASE;
 606 extern const uc_property_t UC_PROPERTY_TITLECASE;
 607 extern const uc_property_t UC_PROPERTY_CASED;
 608 extern const uc_property_t UC_PROPERTY_CASE_IGNORABLE;
 609 extern const uc_property_t UC_PROPERTY_CHANGES_WHEN_LOWERCASED;
 610 extern const uc_property_t UC_PROPERTY_CHANGES_WHEN_UPPERCASED;
 611 extern const uc_property_t UC_PROPERTY_CHANGES_WHEN_TITLECASED;
 612 extern const uc_property_t UC_PROPERTY_CHANGES_WHEN_CASEFOLDED;
 613 extern const uc_property_t UC_PROPERTY_CHANGES_WHEN_CASEMAPPED;
 614 extern const uc_property_t UC_PROPERTY_SOFT_DOTTED;
 615 /* Identifiers.  */
 616 extern const uc_property_t UC_PROPERTY_ID_START;
 617 extern const uc_property_t UC_PROPERTY_OTHER_ID_START;
 618 extern const uc_property_t UC_PROPERTY_ID_CONTINUE;
 619 extern const uc_property_t UC_PROPERTY_OTHER_ID_CONTINUE;
 620 extern const uc_property_t UC_PROPERTY_XID_START;
 621 extern const uc_property_t UC_PROPERTY_XID_CONTINUE;
 622 extern const uc_property_t UC_PROPERTY_PATTERN_WHITE_SPACE;
 623 extern const uc_property_t UC_PROPERTY_PATTERN_SYNTAX;
 624 /* Shaping and rendering.  */
 625 extern const uc_property_t UC_PROPERTY_JOIN_CONTROL;
 626 extern const uc_property_t UC_PROPERTY_GRAPHEME_BASE;
 627 extern const uc_property_t UC_PROPERTY_GRAPHEME_EXTEND;
 628 extern const uc_property_t UC_PROPERTY_OTHER_GRAPHEME_EXTEND;
 629 extern const uc_property_t UC_PROPERTY_GRAPHEME_LINK;
 630 /* Bidi.  */
 631 extern const uc_property_t UC_PROPERTY_BIDI_CONTROL;
 632 extern const uc_property_t UC_PROPERTY_BIDI_LEFT_TO_RIGHT;
 633 extern const uc_property_t UC_PROPERTY_BIDI_HEBREW_RIGHT_TO_LEFT;
 634 extern const uc_property_t UC_PROPERTY_BIDI_ARABIC_RIGHT_TO_LEFT;
 635 extern const uc_property_t UC_PROPERTY_BIDI_EUROPEAN_DIGIT;
 636 extern const uc_property_t UC_PROPERTY_BIDI_EUR_NUM_SEPARATOR;
 637 extern const uc_property_t UC_PROPERTY_BIDI_EUR_NUM_TERMINATOR;
 638 extern const uc_property_t UC_PROPERTY_BIDI_ARABIC_DIGIT;
 639 extern const uc_property_t UC_PROPERTY_BIDI_COMMON_SEPARATOR;
 640 extern const uc_property_t UC_PROPERTY_BIDI_BLOCK_SEPARATOR;
 641 extern const uc_property_t UC_PROPERTY_BIDI_SEGMENT_SEPARATOR;
 642 extern const uc_property_t UC_PROPERTY_BIDI_WHITESPACE;
 643 extern const uc_property_t UC_PROPERTY_BIDI_NON_SPACING_MARK;
 644 extern const uc_property_t UC_PROPERTY_BIDI_BOUNDARY_NEUTRAL;
 645 extern const uc_property_t UC_PROPERTY_BIDI_PDF;
 646 extern const uc_property_t UC_PROPERTY_BIDI_EMBEDDING_OR_OVERRIDE;
 647 extern const uc_property_t UC_PROPERTY_BIDI_OTHER_NEUTRAL;
 648 /* Numeric.  */
 649 extern const uc_property_t UC_PROPERTY_HEX_DIGIT;
 650 extern const uc_property_t UC_PROPERTY_ASCII_HEX_DIGIT;
 651 /* CJK.  */
 652 extern const uc_property_t UC_PROPERTY_IDEOGRAPHIC;
 653 extern const uc_property_t UC_PROPERTY_UNIFIED_IDEOGRAPH;
 654 extern const uc_property_t UC_PROPERTY_RADICAL;
 655 extern const uc_property_t UC_PROPERTY_IDS_BINARY_OPERATOR;
 656 extern const uc_property_t UC_PROPERTY_IDS_TRINARY_OPERATOR;
 657 /* Misc.  */
 658 extern const uc_property_t UC_PROPERTY_ZERO_WIDTH;
 659 extern const uc_property_t UC_PROPERTY_SPACE;
 660 extern const uc_property_t UC_PROPERTY_NON_BREAK;
 661 extern const uc_property_t UC_PROPERTY_ISO_CONTROL;
 662 extern const uc_property_t UC_PROPERTY_FORMAT_CONTROL;
 663 extern const uc_property_t UC_PROPERTY_DASH;
 664 extern const uc_property_t UC_PROPERTY_HYPHEN;
 665 extern const uc_property_t UC_PROPERTY_PUNCTUATION;
 666 extern const uc_property_t UC_PROPERTY_LINE_SEPARATOR;
 667 extern const uc_property_t UC_PROPERTY_PARAGRAPH_SEPARATOR;
 668 extern const uc_property_t UC_PROPERTY_QUOTATION_MARK;
 669 extern const uc_property_t UC_PROPERTY_SENTENCE_TERMINAL;
 670 extern const uc_property_t UC_PROPERTY_TERMINAL_PUNCTUATION;
 671 extern const uc_property_t UC_PROPERTY_CURRENCY_SYMBOL;
 672 extern const uc_property_t UC_PROPERTY_MATH;
 673 extern const uc_property_t UC_PROPERTY_OTHER_MATH;
 674 extern const uc_property_t UC_PROPERTY_PAIRED_PUNCTUATION;
 675 extern const uc_property_t UC_PROPERTY_LEFT_OF_PAIR;
 676 extern const uc_property_t UC_PROPERTY_COMBINING;
 677 extern const uc_property_t UC_PROPERTY_COMPOSITE;
 678 extern const uc_property_t UC_PROPERTY_DECIMAL_DIGIT;
 679 extern const uc_property_t UC_PROPERTY_NUMERIC;
 680 extern const uc_property_t UC_PROPERTY_DIACRITIC;
 681 extern const uc_property_t UC_PROPERTY_EXTENDER;
 682 extern const uc_property_t UC_PROPERTY_IGNORABLE_CONTROL;
 683
 684 /* Return the property given by name, e.g. "White space".  */
 685 extern uc_property_t
 686        uc_property_byname (const char *property_name);
 687
 688 /* Test whether a property is valid.  */
 689 #define uc_property_is_valid(property) ((property).test_fn != NULL)
 690
 691 /* Test whether a Unicode character has a given property.  */
 692 extern bool
 693        uc_is_property (ucs4_t uc, uc_property_t property);
 694 extern bool uc_is_property_white_space (ucs4_t uc)
 695        _UC_ATTRIBUTE_CONST;
 696 extern bool uc_is_property_alphabetic (ucs4_t uc)
 697        _UC_ATTRIBUTE_CONST;
 698 extern bool uc_is_property_other_alphabetic (ucs4_t uc)
 699        _UC_ATTRIBUTE_CONST;
 700 extern bool uc_is_property_not_a_character (ucs4_t uc)
 701        _UC_ATTRIBUTE_CONST;
 702 extern bool uc_is_property_default_ignorable_code_point (ucs4_t uc)
 703        _UC_ATTRIBUTE_CONST;
 704 extern bool uc_is_property_other_default_ignorable_code_point (ucs4_t uc)
 705        _UC_ATTRIBUTE_CONST;
 706 extern bool uc_is_property_deprecated (ucs4_t uc)
 707        _UC_ATTRIBUTE_CONST;
 708 extern bool uc_is_property_logical_order_exception (ucs4_t uc)
 709        _UC_ATTRIBUTE_CONST;
 710 extern bool uc_is_property_variation_selector (ucs4_t uc)
 711        _UC_ATTRIBUTE_CONST;
 712 extern bool uc_is_property_private_use (ucs4_t uc)
 713        _UC_ATTRIBUTE_CONST;
 714 extern bool uc_is_property_unassigned_code_value (ucs4_t uc)
 715        _UC_ATTRIBUTE_CONST;
 716 extern bool uc_is_property_uppercase (ucs4_t uc)
 717        _UC_ATTRIBUTE_CONST;
 718 extern bool uc_is_property_other_uppercase (ucs4_t uc)
 719        _UC_ATTRIBUTE_CONST;
 720 extern bool uc_is_property_lowercase (ucs4_t uc)
 721        _UC_ATTRIBUTE_CONST;
 722 extern bool uc_is_property_other_lowercase (ucs4_t uc)
 723        _UC_ATTRIBUTE_CONST;
 724 extern bool uc_is_property_titlecase (ucs4_t uc)
 725        _UC_ATTRIBUTE_CONST;
 726 extern bool uc_is_property_cased (ucs4_t uc)
 727        _UC_ATTRIBUTE_CONST;
 728 extern bool uc_is_property_case_ignorable (ucs4_t uc)
 729        _UC_ATTRIBUTE_CONST;
 730 extern bool uc_is_property_changes_when_lowercased (ucs4_t uc)
 731        _UC_ATTRIBUTE_CONST;
 732 extern bool uc_is_property_changes_when_uppercased (ucs4_t uc)
 733        _UC_ATTRIBUTE_CONST;
 734 extern bool uc_is_property_changes_when_titlecased (ucs4_t uc)
 735        _UC_ATTRIBUTE_CONST;
 736 extern bool uc_is_property_changes_when_casefolded (ucs4_t uc)
 737        _UC_ATTRIBUTE_CONST;
 738 extern bool uc_is_property_changes_when_casemapped (ucs4_t uc)
 739        _UC_ATTRIBUTE_CONST;
 740 extern bool uc_is_property_soft_dotted (ucs4_t uc)
 741        _UC_ATTRIBUTE_CONST;
 742 extern bool uc_is_property_id_start (ucs4_t uc)
 743        _UC_ATTRIBUTE_CONST;
 744 extern bool uc_is_property_other_id_start (ucs4_t uc)
 745        _UC_ATTRIBUTE_CONST;
 746 extern bool uc_is_property_id_continue (ucs4_t uc)
 747        _UC_ATTRIBUTE_CONST;
 748 extern bool uc_is_property_other_id_continue (ucs4_t uc)
 749        _UC_ATTRIBUTE_CONST;
 750 extern bool uc_is_property_xid_start (ucs4_t uc)
 751        _UC_ATTRIBUTE_CONST;
 752 extern bool uc_is_property_xid_continue (ucs4_t uc)
 753        _UC_ATTRIBUTE_CONST;
 754 extern bool uc_is_property_pattern_white_space (ucs4_t uc)
 755        _UC_ATTRIBUTE_CONST;
 756 extern bool uc_is_property_pattern_syntax (ucs4_t uc)
 757        _UC_ATTRIBUTE_CONST;
 758 extern bool uc_is_property_join_control (ucs4_t uc)
 759        _UC_ATTRIBUTE_CONST;
 760 extern bool uc_is_property_grapheme_base (ucs4_t uc)
 761        _UC_ATTRIBUTE_CONST;
 762 extern bool uc_is_property_grapheme_extend (ucs4_t uc)
 763        _UC_ATTRIBUTE_CONST;
 764 extern bool uc_is_property_other_grapheme_extend (ucs4_t uc)
 765        _UC_ATTRIBUTE_CONST;
 766 extern bool uc_is_property_grapheme_link (ucs4_t uc)
 767        _UC_ATTRIBUTE_CONST;
 768 extern bool uc_is_property_bidi_control (ucs4_t uc)
 769        _UC_ATTRIBUTE_CONST;
 770 extern bool uc_is_property_bidi_left_to_right (ucs4_t uc)
 771        _UC_ATTRIBUTE_CONST;
 772 extern bool uc_is_property_bidi_hebrew_right_to_left (ucs4_t uc)
 773        _UC_ATTRIBUTE_CONST;
 774 extern bool uc_is_property_bidi_arabic_right_to_left (ucs4_t uc)
 775        _UC_ATTRIBUTE_CONST;
 776 extern bool uc_is_property_bidi_european_digit (ucs4_t uc)
 777        _UC_ATTRIBUTE_CONST;
 778 extern bool uc_is_property_bidi_eur_num_separator (ucs4_t uc)
 779        _UC_ATTRIBUTE_CONST;
 780 extern bool uc_is_property_bidi_eur_num_terminator (ucs4_t uc)
 781        _UC_ATTRIBUTE_CONST;
 782 extern bool uc_is_property_bidi_arabic_digit (ucs4_t uc)
 783        _UC_ATTRIBUTE_CONST;
 784 extern bool uc_is_property_bidi_common_separator (ucs4_t uc)
 785        _UC_ATTRIBUTE_CONST;
 786 extern bool uc_is_property_bidi_block_separator (ucs4_t uc)
 787        _UC_ATTRIBUTE_CONST;
 788 extern bool uc_is_property_bidi_segment_separator (ucs4_t uc)
 789        _UC_ATTRIBUTE_CONST;
 790 extern bool uc_is_property_bidi_whitespace (ucs4_t uc)
 791        _UC_ATTRIBUTE_CONST;
 792 extern bool uc_is_property_bidi_non_spacing_mark (ucs4_t uc)
 793        _UC_ATTRIBUTE_CONST;
 794 extern bool uc_is_property_bidi_boundary_neutral (ucs4_t uc)
 795        _UC_ATTRIBUTE_CONST;
 796 extern bool uc_is_property_bidi_pdf (ucs4_t uc)
 797        _UC_ATTRIBUTE_CONST;
 798 extern bool uc_is_property_bidi_embedding_or_override (ucs4_t uc)
 799        _UC_ATTRIBUTE_CONST;
 800 extern bool uc_is_property_bidi_other_neutral (ucs4_t uc)
 801        _UC_ATTRIBUTE_CONST;
 802 extern bool uc_is_property_hex_digit (ucs4_t uc)
 803        _UC_ATTRIBUTE_CONST;
 804 extern bool uc_is_property_ascii_hex_digit (ucs4_t uc)
 805        _UC_ATTRIBUTE_CONST;
 806 extern bool uc_is_property_ideographic (ucs4_t uc)
 807        _UC_ATTRIBUTE_CONST;
 808 extern bool uc_is_property_unified_ideograph (ucs4_t uc)
 809        _UC_ATTRIBUTE_CONST;
 810 extern bool uc_is_property_radical (ucs4_t uc)
 811        _UC_ATTRIBUTE_CONST;
 812 extern bool uc_is_property_ids_binary_operator (ucs4_t uc)
 813        _UC_ATTRIBUTE_CONST;
 814 extern bool uc_is_property_ids_trinary_operator (ucs4_t uc)
 815        _UC_ATTRIBUTE_CONST;
 816 extern bool uc_is_property_zero_width (ucs4_t uc)
 817        _UC_ATTRIBUTE_CONST;
 818 extern bool uc_is_property_space (ucs4_t uc)
 819        _UC_ATTRIBUTE_CONST;
 820 extern bool uc_is_property_non_break (ucs4_t uc)
 821        _UC_ATTRIBUTE_CONST;
 822 extern bool uc_is_property_iso_control (ucs4_t uc)
 823        _UC_ATTRIBUTE_CONST;
 824 extern bool uc_is_property_format_control (ucs4_t uc)
 825        _UC_ATTRIBUTE_CONST;
 826 extern bool uc_is_property_dash (ucs4_t uc)
 827        _UC_ATTRIBUTE_CONST;
 828 extern bool uc_is_property_hyphen (ucs4_t uc)
 829        _UC_ATTRIBUTE_CONST;
 830 extern bool uc_is_property_punctuation (ucs4_t uc)
 831        _UC_ATTRIBUTE_CONST;
 832 extern bool uc_is_property_line_separator (ucs4_t uc)
 833        _UC_ATTRIBUTE_CONST;
 834 extern bool uc_is_property_paragraph_separator (ucs4_t uc)
 835        _UC_ATTRIBUTE_CONST;
 836 extern bool uc_is_property_quotation_mark (ucs4_t uc)
 837        _UC_ATTRIBUTE_CONST;
 838 extern bool uc_is_property_sentence_terminal (ucs4_t uc)
 839        _UC_ATTRIBUTE_CONST;
 840 extern bool uc_is_property_terminal_punctuation (ucs4_t uc)
 841        _UC_ATTRIBUTE_CONST;
 842 extern bool uc_is_property_currency_symbol (ucs4_t uc)
 843        _UC_ATTRIBUTE_CONST;
 844 extern bool uc_is_property_math (ucs4_t uc)
 845        _UC_ATTRIBUTE_CONST;
 846 extern bool uc_is_property_other_math (ucs4_t uc)
 847        _UC_ATTRIBUTE_CONST;
 848 extern bool uc_is_property_paired_punctuation (ucs4_t uc)
 849        _UC_ATTRIBUTE_CONST;
 850 extern bool uc_is_property_left_of_pair (ucs4_t uc)
 851        _UC_ATTRIBUTE_CONST;
 852 extern bool uc_is_property_combining (ucs4_t uc)
 853        _UC_ATTRIBUTE_CONST;
 854 extern bool uc_is_property_composite (ucs4_t uc)
 855        _UC_ATTRIBUTE_CONST;
 856 extern bool uc_is_property_decimal_digit (ucs4_t uc)
 857        _UC_ATTRIBUTE_CONST;
 858 extern bool uc_is_property_numeric (ucs4_t uc)
 859        _UC_ATTRIBUTE_CONST;
 860 extern bool uc_is_property_diacritic (ucs4_t uc)
 861        _UC_ATTRIBUTE_CONST;
 862 extern bool uc_is_property_extender (ucs4_t uc)
 863        _UC_ATTRIBUTE_CONST;
 864 extern bool uc_is_property_ignorable_control (ucs4_t uc)
 865        _UC_ATTRIBUTE_CONST;
 866
 867 /* ========================================================================= */
 868
 869 /* Subdivision of the Unicode characters into scripts.  */
 870
 871 typedef struct
 872 {
 873   unsigned int code : 21;
 874   unsigned int start : 1;
 875   unsigned int end : 1;
 876 }
 877 uc_interval_t;
 878 typedef struct
 879 {
 880   unsigned int nintervals;
 881   const uc_interval_t *intervals;
 882   const char *name;
 883 }
 884 uc_script_t;
 885
 886 /* Return the script of a Unicode character.  */
 887 extern const uc_script_t *
 888        uc_script (ucs4_t uc)
 889        _UC_ATTRIBUTE_CONST;
 890
 891 /* Return the script given by name, e.g. "HAN".  */
 892 extern const uc_script_t *
 893        uc_script_byname (const char *script_name)
 894        _UC_ATTRIBUTE_PURE;
 895
 896 /* Test whether a Unicode character belongs to a given script.  */
 897 extern bool
 898        uc_is_script (ucs4_t uc, const uc_script_t *script)
 899        _UC_ATTRIBUTE_PURE;
 900
 901 /* Get the list of all scripts.  */
 902 extern void
 903        uc_all_scripts (const uc_script_t **scripts, size_t *count);
 904
 905 /* ========================================================================= */
 906
 907 /* Subdivision of the Unicode character range into blocks.  */
 908
 909 typedef struct
 910 {
 911   ucs4_t start;
 912   ucs4_t end;
 913   const char *name;
 914 }
 915 uc_block_t;
 916
 917 /* Return the block a character belongs to.  */
 918 extern const uc_block_t *
 919        uc_block (ucs4_t uc)
 920        _UC_ATTRIBUTE_CONST;
 921
 922 /* Test whether a Unicode character belongs to a given block.  */
 923 extern bool
 924        uc_is_block (ucs4_t uc, const uc_block_t *block)
 925        _UC_ATTRIBUTE_PURE;
 926
 927 /* Get the list of all blocks.  */
 928 extern void
 929        uc_all_blocks (const uc_block_t **blocks, size_t *count);
 930
 931 /* ========================================================================= */
 932
 933 /* Properties taken from language standards.  */
 934
 935 /* Test whether a Unicode character is considered whitespace in ISO C 99.  */
 936 extern bool
 937        uc_is_c_whitespace (ucs4_t uc)
 938        _UC_ATTRIBUTE_CONST;
 939
 940 /* Test whether a Unicode character is considered whitespace in Java.  */
 941 extern bool
 942        uc_is_java_whitespace (ucs4_t uc)
 943        _UC_ATTRIBUTE_CONST;
 944
 945 enum
 946 {
 947   UC_IDENTIFIER_START,    /* valid as first or subsequent character */
 948   UC_IDENTIFIER_VALID,    /* valid as subsequent character only */
 949   UC_IDENTIFIER_INVALID,  /* not valid */
 950   UC_IDENTIFIER_IGNORABLE /* ignorable (Java only) */
 951 };
 952
 953 /* Return the categorization of a Unicode character w.r.t. the ISO C 99
 954    identifier syntax.  */
 955 extern int
 956        uc_c_ident_category (ucs4_t uc)
 957        _UC_ATTRIBUTE_CONST;
 958
 959 /* Return the categorization of a Unicode character w.r.t. the Java
 960    identifier syntax.  */
 961 extern int
 962        uc_java_ident_category (ucs4_t uc)
 963        _UC_ATTRIBUTE_CONST;
 964
 965 /* ========================================================================= */
 966
 967 /* Like ISO C <ctype.h> and <wctype.h>.  These functions are deprecated,
 968    because this set of functions was designed with ASCII in mind and cannot
 969    reflect the more diverse reality of the Unicode character set.  But they
 970    can be a quick-and-dirty porting aid when migrating from wchar_t APIs
 971    to Unicode strings.  */
 972
 973 /* Test for any character for which 'uc_is_alpha' or 'uc_is_digit' is true.  */
 974 extern bool
 975        uc_is_alnum (ucs4_t uc)
 976        _UC_ATTRIBUTE_CONST;
 977
 978 /* Test for any character for which 'uc_is_upper' or 'uc_is_lower' is true,
 979    or any character that is one of a locale-specific set of characters for
 980    which none of 'uc_is_cntrl', 'uc_is_digit', 'uc_is_punct', or 'uc_is_space'
 981    is true.  */
 982 extern bool
 983        uc_is_alpha (ucs4_t uc)
 984        _UC_ATTRIBUTE_CONST;
 985
 986 /* Test for any control character.  */
 987 extern bool
 988        uc_is_cntrl (ucs4_t uc)
 989        _UC_ATTRIBUTE_CONST;
 990
 991 /* Test for any character that corresponds to a decimal-digit character.  */
 992 extern bool
 993        uc_is_digit (ucs4_t uc)
 994        _UC_ATTRIBUTE_CONST;
 995
 996 /* Test for any character for which 'uc_is_print' is true and 'uc_is_space'
 997    is false.  */
 998 extern bool
 999        uc_is_graph (ucs4_t uc)
1000        _UC_ATTRIBUTE_CONST;
1001
1002 /* Test for any character that corresponds to a lowercase letter or is one
1003    of a locale-specific set of characters for which none of 'uc_is_cntrl',
1004    'uc_is_digit', 'uc_is_punct', or 'uc_is_space' is true.  */
1005 extern bool
1006        uc_is_lower (ucs4_t uc)
1007        _UC_ATTRIBUTE_CONST;
1008
1009 /* Test for any printing character.  */
1010 extern bool
1011        uc_is_print (ucs4_t uc)
1012        _UC_ATTRIBUTE_CONST;
1013
1014 /* Test for any printing character that is one of a locale-specific set of
1015    characters for which neither 'uc_is_space' nor 'uc_is_alnum' is true.  */
1016 extern bool
1017        uc_is_punct (ucs4_t uc)
1018        _UC_ATTRIBUTE_CONST;
1019
1020 /* Test for any character that corresponds to a locale-specific set of
1021    characters for which none of 'uc_is_alnum', 'uc_is_graph', or 'uc_is_punct'
1022    is true.  */
1023 extern bool
1024        uc_is_space (ucs4_t uc)
1025        _UC_ATTRIBUTE_CONST;
1026
1027 /* Test for any character that corresponds to an uppercase letter or is one
1028    of a locale-specific set of character for which none of 'uc_is_cntrl',
1029    'uc_is_digit', 'uc_is_punct', or 'uc_is_space' is true.  */
1030 extern bool
1031        uc_is_upper (ucs4_t uc)
1032        _UC_ATTRIBUTE_CONST;
1033
1034 /* Test for any character that corresponds to a hexadecimal-digit
1035    character.  */
1036 extern bool
1037        uc_is_xdigit (ucs4_t uc)
1038        _UC_ATTRIBUTE_CONST;
1039
1040 /* GNU extension. */
1041 /* Test for any character that corresponds to a standard blank character or
1042    a locale-specific set of characters for which 'uc_is_alnum' is false.  */
1043 extern bool
1044        uc_is_blank (ucs4_t uc)
1045        _UC_ATTRIBUTE_CONST;
1046
1047 /* ========================================================================= */
1048
1049 #ifdef __cplusplus
1050 }
1051 #endif
1052
1053 #endif /* _UNICTYPE_H */