1 // © 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 3 /* 4 ******************************************************************************* 5 * 6 * Copyright (C) 2004-2012, International Business Machines 7 * Corporation and others. All Rights Reserved. 8 * 9 ******************************************************************************* 10 * file name: ucase.h 11 * encoding: UTF-8 12 * tab size: 8 (not used) 13 * indentation:4 14 * 15 * created on: 2004aug30 16 * created by: Markus W. Scherer 17 * 18 * Low-level Unicode character/string case mapping code. 19 */ 20 21 #ifndef __UCASE_H__ 22 #define __UCASE_H__ 23 24 #include "unicode/utypes.h" 25 #include "unicode/uset.h" 26 #include "putilimp.h" 27 #include "uset_imp.h" 28 #include "udataswp.h" 29 #include "utrie2.h" 30 31 #ifdef __cplusplus 32 U_NAMESPACE_BEGIN 33 34 class UnicodeString; 35 36 U_NAMESPACE_END 37 #endif 38 39 /* library API -------------------------------------------------------------- */ 40 41 U_CFUNC void U_EXPORT2 42 ucase_addPropertyStarts(const USetAdder *sa, UErrorCode *pErrorCode); 43 44 /** 45 * Requires non-NULL locale ID but otherwise does the equivalent of 46 * checking for language codes as if uloc_getLanguage() were called: 47 * Accepts both 2- and 3-letter codes and accepts case variants. 48 */ 49 U_CFUNC int32_t 50 ucase_getCaseLocale(const char *locale); 51 52 /* Casing locale types for ucase_getCaseLocale */ 53 enum { 54 UCASE_LOC_UNKNOWN, 55 UCASE_LOC_ROOT, 56 UCASE_LOC_TURKISH, 57 UCASE_LOC_LITHUANIAN, 58 UCASE_LOC_GREEK, 59 UCASE_LOC_DUTCH, 60 UCASE_LOC_ARMENIAN 61 }; 62 63 /** 64 * Bit mask for getting just the options from a string compare options word 65 * that are relevant for case-insensitive string comparison. 66 * See stringoptions.h. Also include _STRNCMP_STYLE and U_COMPARE_CODE_POINT_ORDER. 67 * @internal 68 */ 69 #define _STRCASECMP_OPTIONS_MASK 0xffff 70 71 /** 72 * Bit mask for getting just the options from a string compare options word 73 * that are relevant for case folding (of a single string or code point). 74 * 75 * Currently only bit 0 for U_FOLD_CASE_EXCLUDE_SPECIAL_I. 76 * It is conceivable that at some point we might use one more bit for using uppercase sharp s. 77 * It is conceivable that at some point we might want the option to use only simple case foldings 78 * when operating on strings. 79 * 80 * See stringoptions.h. 81 * @internal 82 */ 83 #define _FOLD_CASE_OPTIONS_MASK 7 84 85 /* single-code point functions */ 86 87 U_CAPI UChar32 U_EXPORT2 88 ucase_tolower(UChar32 c); 89 90 U_CAPI UChar32 U_EXPORT2 91 ucase_toupper(UChar32 c); 92 93 U_CAPI UChar32 U_EXPORT2 94 ucase_totitle(UChar32 c); 95 96 U_CAPI UChar32 U_EXPORT2 97 ucase_fold(UChar32 c, uint32_t options); 98 99 /** 100 * Adds all simple case mappings and the full case folding for c to sa, 101 * and also adds special case closure mappings. 102 * c itself is not added. 103 * For example, the mappings 104 * - for s include long s 105 * - for sharp s include ss 106 * - for k include the Kelvin sign 107 */ 108 U_CFUNC void U_EXPORT2 109 ucase_addCaseClosure(UChar32 c, const USetAdder *sa); 110 111 /** 112 * Maps the string to single code points and adds the associated case closure 113 * mappings. 114 * The string is mapped to code points if it is their full case folding string. 115 * In other words, this performs a reverse full case folding and then 116 * adds the case closure items of the resulting code points. 117 * If the string is found and its closure applied, then 118 * the string itself is added as well as part of its code points' closure. 119 * It must be length>=0. 120 * 121 * @return true if the string was found 122 */ 123 U_CFUNC UBool U_EXPORT2 124 ucase_addStringCaseClosure(const UChar *s, int32_t length, const USetAdder *sa); 125 126 #ifdef __cplusplus 127 U_NAMESPACE_BEGIN 128 129 /** 130 * Iterator over characters with more than one code point in the full default Case_Folding. 131 */ 132 class U_COMMON_API FullCaseFoldingIterator { 133 public: 134 /** Constructor. */ 135 FullCaseFoldingIterator(); 136 /** 137 * Returns the next (cp, full) pair where "full" is cp's full default Case_Folding. 138 * Returns a negative cp value at the end of the iteration. 139 */ 140 UChar32 next(UnicodeString &full); 141 private: 142 FullCaseFoldingIterator(const FullCaseFoldingIterator &); // no copy 143 FullCaseFoldingIterator &operator=(const FullCaseFoldingIterator &); // no assignment 144 145 const UChar *unfold; 146 int32_t unfoldRows; 147 int32_t unfoldRowWidth; 148 int32_t unfoldStringWidth; 149 int32_t currentRow; 150 int32_t rowCpIndex; 151 }; 152 153 /** 154 * Fast case mapping data for ASCII/Latin. 155 * Linear arrays of delta bytes: 0=no mapping; EXC=exception. 156 * Deltas must not cross the ASCII boundary, or else they cannot be easily used 157 * in simple UTF-8 code. 158 */ 159 namespace LatinCase { 160 161 /** Case mapping/folding data for code points up to U+017F. */ 162 constexpr UChar LIMIT = 0x180; 163 /** U+017F case-folds and uppercases crossing the ASCII boundary. */ 164 constexpr UChar LONG_S = 0x17f; 165 /** Exception: Complex mapping, or too-large delta. */ 166 constexpr int8_t EXC = -0x80; 167 168 /** Deltas for lowercasing for most locales, and default case folding. */ 169 extern const int8_t TO_LOWER_NORMAL[LIMIT]; 170 /** Deltas for lowercasing for tr/az/lt, and Turkic case folding. */ 171 extern const int8_t TO_LOWER_TR_LT[LIMIT]; 172 173 /** Deltas for uppercasing for most locales. */ 174 extern const int8_t TO_UPPER_NORMAL[LIMIT]; 175 /** Deltas for uppercasing for tr/az. */ 176 extern const int8_t TO_UPPER_TR[LIMIT]; 177 178 } // namespace LatinCase 179 180 U_NAMESPACE_END 181 #endif 182 183 /** @return UCASE_NONE, UCASE_LOWER, UCASE_UPPER, UCASE_TITLE */ 184 U_CAPI int32_t U_EXPORT2 185 ucase_getType(UChar32 c); 186 187 /** @return like ucase_getType() but also sets UCASE_IGNORABLE if c is case-ignorable */ 188 U_CAPI int32_t U_EXPORT2 189 ucase_getTypeOrIgnorable(UChar32 c); 190 191 U_CAPI UBool U_EXPORT2 192 ucase_isSoftDotted(UChar32 c); 193 194 U_CAPI UBool U_EXPORT2 195 ucase_isCaseSensitive(UChar32 c); 196 197 /* string case mapping functions */ 198 199 U_CDECL_BEGIN 200 201 /** 202 * Iterator function for string case mappings, which need to look at the 203 * context (surrounding text) of a given character for conditional mappings. 204 * 205 * The iterator only needs to go backward or forward away from the 206 * character in question. It does not use any indexes on this interface. 207 * It does not support random access or an arbitrary change of 208 * iteration direction. 209 * 210 * The code point being case-mapped itself is never returned by 211 * this iterator. 212 * 213 * @param context A pointer to the iterator's working data. 214 * @param dir If <0 then start iterating backward from the character; 215 * if >0 then start iterating forward from the character; 216 * if 0 then continue iterating in the current direction. 217 * @return Next code point, or <0 when the iteration is done. 218 */ 219 typedef UChar32 U_CALLCONV 220 UCaseContextIterator(void *context, int8_t dir); 221 222 /** 223 * Sample struct which may be used by some implementations of 224 * UCaseContextIterator. 225 */ 226 struct UCaseContext { 227 void *p; 228 int32_t start, index, limit; 229 int32_t cpStart, cpLimit; 230 int8_t dir; 231 int8_t b1, b2, b3; 232 }; 233 typedef struct UCaseContext UCaseContext; 234 235 U_CDECL_END 236 237 #define UCASECONTEXT_INITIALIZER { NULL, 0, 0, 0, 0, 0, 0, 0, 0, 0 } 238 239 enum { 240 /** 241 * For string case mappings, a single character (a code point) is mapped 242 * either to itself (in which case in-place mapping functions do nothing), 243 * or to another single code point, or to a string. 244 * Aside from the string contents, these are indicated with a single int32_t 245 * value as follows: 246 * 247 * Mapping to self: Negative values (~self instead of -self to support U+0000) 248 * 249 * Mapping to another code point: Positive values >UCASE_MAX_STRING_LENGTH 250 * 251 * Mapping to a string: The string length (0..UCASE_MAX_STRING_LENGTH) is 252 * returned. Note that the string result may indeed have zero length. 253 */ 254 UCASE_MAX_STRING_LENGTH=0x1f 255 }; 256 257 /** 258 * Get the full lowercase mapping for c. 259 * 260 * @param csp Case mapping properties. 261 * @param c Character to be mapped. 262 * @param iter Character iterator, used for context-sensitive mappings. 263 * See UCaseContextIterator for details. 264 * If iter==NULL then a context-independent result is returned. 265 * @param context Pointer to be passed into iter. 266 * @param pString If the mapping result is a string, then the pointer is 267 * written to *pString. 268 * @param caseLocale Case locale value from ucase_getCaseLocale(). 269 * @return Output code point or string length, see UCASE_MAX_STRING_LENGTH. 270 * 271 * @see UCaseContextIterator 272 * @see UCASE_MAX_STRING_LENGTH 273 * @internal 274 */ 275 U_CAPI int32_t U_EXPORT2 276 ucase_toFullLower(UChar32 c, 277 UCaseContextIterator *iter, void *context, 278 const UChar **pString, 279 int32_t caseLocale); 280 281 U_CAPI int32_t U_EXPORT2 282 ucase_toFullUpper(UChar32 c, 283 UCaseContextIterator *iter, void *context, 284 const UChar **pString, 285 int32_t caseLocale); 286 287 U_CAPI int32_t U_EXPORT2 288 ucase_toFullTitle(UChar32 c, 289 UCaseContextIterator *iter, void *context, 290 const UChar **pString, 291 int32_t caseLocale); 292 293 U_CAPI int32_t U_EXPORT2 294 ucase_toFullFolding(UChar32 c, 295 const UChar **pString, 296 uint32_t options); 297 298 U_CFUNC int32_t U_EXPORT2 299 ucase_hasBinaryProperty(UChar32 c, UProperty which); 300 301 302 U_CDECL_BEGIN 303 304 /** 305 * @internal 306 */ 307 typedef int32_t U_CALLCONV 308 UCaseMapFull(UChar32 c, 309 UCaseContextIterator *iter, void *context, 310 const UChar **pString, 311 int32_t caseLocale); 312 313 U_CDECL_END 314 315 /* file definitions --------------------------------------------------------- */ 316 317 #define UCASE_DATA_NAME "ucase" 318 #define UCASE_DATA_TYPE "icu" 319 320 /* format "cAsE" */ 321 #define UCASE_FMT_0 0x63 322 #define UCASE_FMT_1 0x41 323 #define UCASE_FMT_2 0x53 324 #define UCASE_FMT_3 0x45 325 326 /* indexes into indexes[] */ 327 enum { 328 UCASE_IX_INDEX_TOP, 329 UCASE_IX_LENGTH, 330 UCASE_IX_TRIE_SIZE, 331 UCASE_IX_EXC_LENGTH, 332 UCASE_IX_UNFOLD_LENGTH, 333 334 UCASE_IX_MAX_FULL_LENGTH=15, 335 UCASE_IX_TOP=16 336 }; 337 338 /* definitions for 16-bit case properties word ------------------------------ */ 339 340 U_CFUNC const UTrie2 * U_EXPORT2 341 ucase_getTrie(); 342 343 /* 2-bit constants for types of cased characters */ 344 #define UCASE_TYPE_MASK 3 345 enum { 346 UCASE_NONE, 347 UCASE_LOWER, 348 UCASE_UPPER, 349 UCASE_TITLE 350 }; 351 352 #define UCASE_GET_TYPE(props) ((props)&UCASE_TYPE_MASK) 353 #define UCASE_GET_TYPE_AND_IGNORABLE(props) ((props)&7) 354 355 #define UCASE_IS_UPPER_OR_TITLE(props) ((props)&2) 356 357 #define UCASE_IGNORABLE 4 358 #define UCASE_EXCEPTION 8 359 #define UCASE_SENSITIVE 0x10 360 361 #define UCASE_HAS_EXCEPTION(props) ((props)&UCASE_EXCEPTION) 362 363 #define UCASE_DOT_MASK 0x60 364 enum { 365 UCASE_NO_DOT=0, /* normal characters with cc=0 */ 366 UCASE_SOFT_DOTTED=0x20, /* soft-dotted characters with cc=0 */ 367 UCASE_ABOVE=0x40, /* "above" accents with cc=230 */ 368 UCASE_OTHER_ACCENT=0x60 /* other accent character (0<cc!=230) */ 369 }; 370 371 /* no exception: bits 15..7 are a 9-bit signed case mapping delta */ 372 #define UCASE_DELTA_SHIFT 7 373 #define UCASE_DELTA_MASK 0xff80 374 #define UCASE_MAX_DELTA 0xff 375 #define UCASE_MIN_DELTA (-UCASE_MAX_DELTA-1) 376 377 #if U_SIGNED_RIGHT_SHIFT_IS_ARITHMETIC 378 # define UCASE_GET_DELTA(props) ((int16_t)(props)>>UCASE_DELTA_SHIFT) 379 #else 380 # define UCASE_GET_DELTA(props) (int16_t)(((props)&0x8000) ? (((props)>>UCASE_DELTA_SHIFT)|0xfe00) : ((uint16_t)(props)>>UCASE_DELTA_SHIFT)) 381 #endif 382 383 /* exception: bits 15..4 are an unsigned 12-bit index into the exceptions array */ 384 #define UCASE_EXC_SHIFT 4 385 #define UCASE_EXC_MASK 0xfff0 386 #define UCASE_MAX_EXCEPTIONS ((UCASE_EXC_MASK>>UCASE_EXC_SHIFT)+1) 387 388 /* definitions for 16-bit main exceptions word ------------------------------ */ 389 390 /* first 8 bits indicate values in optional slots */ 391 enum { 392 UCASE_EXC_LOWER, 393 UCASE_EXC_FOLD, 394 UCASE_EXC_UPPER, 395 UCASE_EXC_TITLE, 396 UCASE_EXC_DELTA, 397 UCASE_EXC_5, /* reserved */ 398 UCASE_EXC_CLOSURE, 399 UCASE_EXC_FULL_MAPPINGS, 400 UCASE_EXC_ALL_SLOTS /* one past the last slot */ 401 }; 402 403 /* each slot is 2 uint16_t instead of 1 */ 404 #define UCASE_EXC_DOUBLE_SLOTS 0x100 405 406 enum { 407 UCASE_EXC_NO_SIMPLE_CASE_FOLDING=0x200, 408 UCASE_EXC_DELTA_IS_NEGATIVE=0x400, 409 UCASE_EXC_SENSITIVE=0x800 410 }; 411 412 /* UCASE_EXC_DOT_MASK=UCASE_DOT_MASK<<UCASE_EXC_DOT_SHIFT */ 413 #define UCASE_EXC_DOT_SHIFT 7 414 415 /* normally stored in the main word, but pushed out for larger exception indexes */ 416 #define UCASE_EXC_DOT_MASK 0x3000 417 enum { 418 UCASE_EXC_NO_DOT=0, 419 UCASE_EXC_SOFT_DOTTED=0x1000, 420 UCASE_EXC_ABOVE=0x2000, /* "above" accents with cc=230 */ 421 UCASE_EXC_OTHER_ACCENT=0x3000 /* other character (0<cc!=230) */ 422 }; 423 424 /* complex/conditional mappings */ 425 #define UCASE_EXC_CONDITIONAL_SPECIAL 0x4000 426 #define UCASE_EXC_CONDITIONAL_FOLD 0x8000 427 428 /* definitions for lengths word for full case mappings */ 429 #define UCASE_FULL_LOWER 0xf 430 #define UCASE_FULL_FOLDING 0xf0 431 #define UCASE_FULL_UPPER 0xf00 432 #define UCASE_FULL_TITLE 0xf000 433 434 /* maximum lengths */ 435 #define UCASE_FULL_MAPPINGS_MAX_LENGTH (4*0xf) 436 #define UCASE_CLOSURE_MAX_LENGTH 0xf 437 438 /* constants for reverse case folding ("unfold") data */ 439 enum { 440 UCASE_UNFOLD_ROWS, 441 UCASE_UNFOLD_ROW_WIDTH, 442 UCASE_UNFOLD_STRING_WIDTH 443 }; 444 445 #endif 446