icuSources/common/ucase.h

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