icuSources/common/ucase.h

   1 /*
   2 *******************************************************************************
   3 *
   4 *   Copyright (C) 2004-2012, International Business Machines
   5 *   Corporation and others.  All Rights Reserved.
   6 *
   7 *******************************************************************************
   8 *   file name:  ucase.h
   9 *   encoding:   US-ASCII
  10 *   tab size:   8 (not used)
  11 *   indentation:4
  12 *
  13 *   created on: 2004aug30
  14 *   created by: Markus W. Scherer
  15 *
  16 *   Low-level Unicode character/string case mapping code.
  17 */
  18
  19 #ifndef __UCASE_H__
  20 #define __UCASE_H__
  21
  22 #include "unicode/utypes.h"
  23 #include "unicode/uset.h"
  24 #include "putilimp.h"
  25 #include "uset_imp.h"
  26 #include "udataswp.h"
  27
  28 #ifdef __cplusplus
  29 U_NAMESPACE_BEGIN
  30
  31 class UnicodeString;
  32
  33 U_NAMESPACE_END
  34 #endif
  35
  36 /* library API -------------------------------------------------------------- */
  37
  38 U_CDECL_BEGIN
  39
  40 struct UCaseProps;
  41 typedef struct UCaseProps UCaseProps;
  42
  43 U_CDECL_END
  44
  45 U_CAPI const UCaseProps * U_EXPORT2
  46 ucase_getSingleton(void);
  47
  48 U_CFUNC void U_EXPORT2
  49 ucase_addPropertyStarts(const UCaseProps *csp, const USetAdder *sa, UErrorCode *pErrorCode);
  50
  51 /**
  52  * Requires non-NULL locale ID but otherwise does the equivalent of
  53  * checking for language codes as if uloc_getLanguage() were called:
  54  * Accepts both 2- and 3-letter codes and accepts case variants.
  55  */
  56 U_CFUNC int32_t
  57 ucase_getCaseLocale(const char *locale, int32_t *locCache);
  58
  59 /* Casing locale types for ucase_getCaseLocale */
  60 enum {
  61     UCASE_LOC_UNKNOWN,
  62     UCASE_LOC_ROOT,
  63     UCASE_LOC_TURKISH,
  64     UCASE_LOC_LITHUANIAN,
  65     UCASE_LOC_DUTCH
  66 };
  67
  68 /**
  69  * Bit mask for getting just the options from a string compare options word
  70  * that are relevant for case-insensitive string comparison.
  71  * See uchar.h. Also include _STRNCMP_STYLE and U_COMPARE_CODE_POINT_ORDER.
  72  * @internal
  73  */
  74 #define _STRCASECMP_OPTIONS_MASK 0xffff
  75
  76 /**
  77  * Bit mask for getting just the options from a string compare options word
  78  * that are relevant for case folding (of a single string or code point).
  79  * See uchar.h.
  80  * @internal
  81  */
  82 #define _FOLD_CASE_OPTIONS_MASK 0xff
  83
  84 /* single-code point functions */
  85
  86 U_CAPI UChar32 U_EXPORT2
  87 ucase_tolower(const UCaseProps *csp, UChar32 c);
  88
  89 U_CAPI UChar32 U_EXPORT2
  90 ucase_toupper(const UCaseProps *csp, UChar32 c);
  91
  92 U_CAPI UChar32 U_EXPORT2
  93 ucase_totitle(const UCaseProps *csp, UChar32 c);
  94
  95 U_CAPI UChar32 U_EXPORT2
  96 ucase_fold(const UCaseProps *csp, 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(const UCaseProps *csp, 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 UCaseProps *csp, 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 U_NAMESPACE_END
 153 #endif
 154
 155 /** @return UCASE_NONE, UCASE_LOWER, UCASE_UPPER, UCASE_TITLE */
 156 U_CAPI int32_t U_EXPORT2
 157 ucase_getType(const UCaseProps *csp, UChar32 c);
 158
 159 /** @return same as ucase_getType(), or <0 if c is case-ignorable */
 160 U_CAPI int32_t U_EXPORT2
 161 ucase_getTypeOrIgnorable(const UCaseProps *csp, UChar32 c);
 162
 163 U_CAPI UBool U_EXPORT2
 164 ucase_isSoftDotted(const UCaseProps *csp, UChar32 c);
 165
 166 U_CAPI UBool U_EXPORT2
 167 ucase_isCaseSensitive(const UCaseProps *csp, UChar32 c);
 168
 169 /* string case mapping functions */
 170
 171 U_CDECL_BEGIN
 172
 173 /**
 174  * Iterator function for string case mappings, which need to look at the
 175  * context (surrounding text) of a given character for conditional mappings.
 176  *
 177  * The iterator only needs to go backward or forward away from the
 178  * character in question. It does not use any indexes on this interface.
 179  * It does not support random access or an arbitrary change of
 180  * iteration direction.
 181  *
 182  * The code point being case-mapped itself is never returned by
 183  * this iterator.
 184  *
 185  * @param context A pointer to the iterator's working data.
 186  * @param dir If <0 then start iterating backward from the character;
 187  *            if >0 then start iterating forward from the character;
 188  *            if 0 then continue iterating in the current direction.
 189  * @return Next code point, or <0 when the iteration is done.
 190  */
 191 typedef UChar32 U_CALLCONV
 192 UCaseContextIterator(void *context, int8_t dir);
 193
 194 /**
 195  * Sample struct which may be used by some implementations of
 196  * UCaseContextIterator.
 197  */
 198 struct UCaseContext {
 199     void *p;
 200     int32_t start, index, limit;
 201     int32_t cpStart, cpLimit;
 202     int8_t dir;
 203     int8_t b1, b2, b3;
 204 };
 205 typedef struct UCaseContext UCaseContext;
 206
 207 U_CDECL_END
 208
 209 #define UCASECONTEXT_INITIALIZER { NULL,  0, 0, 0,  0, 0,  0,  0, 0, 0 }
 210
 211 enum {
 212     /**
 213      * For string case mappings, a single character (a code point) is mapped
 214      * either to itself (in which case in-place mapping functions do nothing),
 215      * or to another single code point, or to a string.
 216      * Aside from the string contents, these are indicated with a single int32_t
 217      * value as follows:
 218      *
 219      * Mapping to self: Negative values (~self instead of -self to support U+0000)
 220      *
 221      * Mapping to another code point: Positive values >UCASE_MAX_STRING_LENGTH
 222      *
 223      * Mapping to a string: The string length (0..UCASE_MAX_STRING_LENGTH) is
 224      * returned. Note that the string result may indeed have zero length.
 225      */
 226     UCASE_MAX_STRING_LENGTH=0x1f
 227 };
 228
 229 /**
 230  * Get the full lowercase mapping for c.
 231  *
 232  * @param csp Case mapping properties.
 233  * @param c Character to be mapped.
 234  * @param iter Character iterator, used for context-sensitive mappings.
 235  *             See UCaseContextIterator for details.
 236  *             If iter==NULL then a context-independent result is returned.
 237  * @param context Pointer to be passed into iter.
 238  * @param pString If the mapping result is a string, then the pointer is
 239  *                written to *pString.
 240  * @param locale Locale ID for locale-dependent mappings.
 241  * @param locCache Initialize to 0; may be used to cache the result of parsing
 242  *                 the locale ID for subsequent calls.
 243  *                 Can be NULL.
 244  * @return Output code point or string length, see UCASE_MAX_STRING_LENGTH.
 245  *
 246  * @see UCaseContextIterator
 247  * @see UCASE_MAX_STRING_LENGTH
 248  * @internal
 249  */
 250 U_CAPI int32_t U_EXPORT2
 251 ucase_toFullLower(const UCaseProps *csp, UChar32 c,
 252                   UCaseContextIterator *iter, void *context,
 253                   const UChar **pString,
 254                   const char *locale, int32_t *locCache);
 255
 256 U_CAPI int32_t U_EXPORT2
 257 ucase_toFullUpper(const UCaseProps *csp, UChar32 c,
 258                   UCaseContextIterator *iter, void *context,
 259                   const UChar **pString,
 260                   const char *locale, int32_t *locCache);
 261
 262 U_CAPI int32_t U_EXPORT2
 263 ucase_toFullTitle(const UCaseProps *csp, UChar32 c,
 264                   UCaseContextIterator *iter, void *context,
 265                   const UChar **pString,
 266                   const char *locale, int32_t *locCache);
 267
 268 U_CAPI int32_t U_EXPORT2
 269 ucase_toFullFolding(const UCaseProps *csp, UChar32 c,
 270                     const UChar **pString,
 271                     uint32_t options);
 272
 273 U_CFUNC int32_t U_EXPORT2
 274 ucase_hasBinaryProperty(UChar32 c, UProperty which);
 275
 276
 277 U_CDECL_BEGIN
 278
 279 /**
 280  * @internal
 281  */
 282 typedef int32_t U_CALLCONV
 283 UCaseMapFull(const UCaseProps *csp, UChar32 c,
 284              UCaseContextIterator *iter, void *context,
 285              const UChar **pString,
 286              const char *locale, int32_t *locCache);
 287
 288 U_CDECL_END
 289
 290 /* file definitions --------------------------------------------------------- */
 291
 292 #define UCASE_DATA_NAME "ucase"
 293 #define UCASE_DATA_TYPE "icu"
 294
 295 /* format "cAsE" */
 296 #define UCASE_FMT_0 0x63
 297 #define UCASE_FMT_1 0x41
 298 #define UCASE_FMT_2 0x53
 299 #define UCASE_FMT_3 0x45
 300
 301 /* indexes into indexes[] */
 302 enum {
 303     UCASE_IX_INDEX_TOP,
 304     UCASE_IX_LENGTH,
 305     UCASE_IX_TRIE_SIZE,
 306     UCASE_IX_EXC_LENGTH,
 307     UCASE_IX_UNFOLD_LENGTH,
 308
 309     UCASE_IX_MAX_FULL_LENGTH=15,
 310     UCASE_IX_TOP=16
 311 };
 312
 313 /* definitions for 16-bit case properties word ------------------------------ */
 314
 315 /* 2-bit constants for types of cased characters */
 316 #define UCASE_TYPE_MASK     3
 317 enum {
 318     UCASE_NONE,
 319     UCASE_LOWER,
 320     UCASE_UPPER,
 321     UCASE_TITLE
 322 };
 323
 324 #define UCASE_GET_TYPE(props) ((props)&UCASE_TYPE_MASK)
 325 #define UCASE_GET_TYPE_AND_IGNORABLE(props) ((props)&7)
 326
 327 #define UCASE_IGNORABLE         4
 328 #define UCASE_SENSITIVE         8
 329 #define UCASE_EXCEPTION         0x10
 330
 331 #define UCASE_DOT_MASK      0x60
 332 enum {
 333     UCASE_NO_DOT=0,         /* normal characters with cc=0 */
 334     UCASE_SOFT_DOTTED=0x20, /* soft-dotted characters with cc=0 */
 335     UCASE_ABOVE=0x40,       /* "above" accents with cc=230 */
 336     UCASE_OTHER_ACCENT=0x60 /* other accent character (0<cc!=230) */
 337 };
 338
 339 /* no exception: bits 15..7 are a 9-bit signed case mapping delta */
 340 #define UCASE_DELTA_SHIFT   7
 341 #define UCASE_DELTA_MASK    0xff80
 342 #define UCASE_MAX_DELTA     0xff
 343 #define UCASE_MIN_DELTA     (-UCASE_MAX_DELTA-1)
 344
 345 #if U_SIGNED_RIGHT_SHIFT_IS_ARITHMETIC
 346 #   define UCASE_GET_DELTA(props) ((int16_t)(props)>>UCASE_DELTA_SHIFT)
 347 #else
 348 #   define UCASE_GET_DELTA(props) (int16_t)(((props)&0x8000) ? (((props)>>UCASE_DELTA_SHIFT)|0xfe00) : ((uint16_t)(props)>>UCASE_DELTA_SHIFT))
 349 #endif
 350
 351 /* exception: bits 15..5 are an unsigned 11-bit index into the exceptions array */
 352 #define UCASE_EXC_SHIFT     5
 353 #define UCASE_EXC_MASK      0xffe0
 354 #define UCASE_MAX_EXCEPTIONS ((UCASE_EXC_MASK>>UCASE_EXC_SHIFT)+1)
 355
 356 /* definitions for 16-bit main exceptions word ------------------------------ */
 357
 358 /* first 8 bits indicate values in optional slots */
 359 enum {
 360     UCASE_EXC_LOWER,
 361     UCASE_EXC_FOLD,
 362     UCASE_EXC_UPPER,
 363     UCASE_EXC_TITLE,
 364     UCASE_EXC_4,            /* reserved */
 365     UCASE_EXC_5,            /* reserved */
 366     UCASE_EXC_CLOSURE,
 367     UCASE_EXC_FULL_MAPPINGS,
 368     UCASE_EXC_ALL_SLOTS     /* one past the last slot */
 369 };
 370
 371 /* each slot is 2 uint16_t instead of 1 */
 372 #define UCASE_EXC_DOUBLE_SLOTS      0x100
 373
 374 /* reserved: exception bits 11..9 */
 375
 376 /* UCASE_EXC_DOT_MASK=UCASE_DOT_MASK<<UCASE_EXC_DOT_SHIFT */
 377 #define UCASE_EXC_DOT_SHIFT     7
 378
 379 /* normally stored in the main word, but pushed out for larger exception indexes */
 380 #define UCASE_EXC_DOT_MASK      0x3000
 381 enum {
 382     UCASE_EXC_NO_DOT=0,
 383     UCASE_EXC_SOFT_DOTTED=0x1000,
 384     UCASE_EXC_ABOVE=0x2000,         /* "above" accents with cc=230 */
 385     UCASE_EXC_OTHER_ACCENT=0x3000   /* other character (0<cc!=230) */
 386 };
 387
 388 /* complex/conditional mappings */
 389 #define UCASE_EXC_CONDITIONAL_SPECIAL   0x4000
 390 #define UCASE_EXC_CONDITIONAL_FOLD      0x8000
 391
 392 /* definitions for lengths word for full case mappings */
 393 #define UCASE_FULL_LOWER    0xf
 394 #define UCASE_FULL_FOLDING  0xf0
 395 #define UCASE_FULL_UPPER    0xf00
 396 #define UCASE_FULL_TITLE    0xf000
 397
 398 /* maximum lengths */
 399 #define UCASE_FULL_MAPPINGS_MAX_LENGTH (4*0xf)
 400 #define UCASE_CLOSURE_MAX_LENGTH 0xf
 401
 402 /* constants for reverse case folding ("unfold") data */
 403 enum {
 404     UCASE_UNFOLD_ROWS,
 405     UCASE_UNFOLD_ROW_WIDTH,
 406     UCASE_UNFOLD_STRING_WIDTH
 407 };
 408
 409 #endif