Wed, 31 Dec 2014 06:09:35 +0100
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
michael@0 | 1 | /* |
michael@0 | 2 | ****************************************************************************** |
michael@0 | 3 | * Copyright (C) 1997-2013, International Business Machines |
michael@0 | 4 | * Corporation and others. All Rights Reserved. |
michael@0 | 5 | ****************************************************************************** |
michael@0 | 6 | */ |
michael@0 | 7 | |
michael@0 | 8 | /** |
michael@0 | 9 | * \file |
michael@0 | 10 | * \brief C++ API: Collation Element Iterator. |
michael@0 | 11 | */ |
michael@0 | 12 | |
michael@0 | 13 | /** |
michael@0 | 14 | * File coleitr.h |
michael@0 | 15 | * |
michael@0 | 16 | * |
michael@0 | 17 | * |
michael@0 | 18 | * Created by: Helena Shih |
michael@0 | 19 | * |
michael@0 | 20 | * Modification History: |
michael@0 | 21 | * |
michael@0 | 22 | * Date Name Description |
michael@0 | 23 | * |
michael@0 | 24 | * 8/18/97 helena Added internal API documentation. |
michael@0 | 25 | * 08/03/98 erm Synched with 1.2 version CollationElementIterator.java |
michael@0 | 26 | * 12/10/99 aliu Ported Thai collation support from Java. |
michael@0 | 27 | * 01/25/01 swquek Modified into a C++ wrapper calling C APIs (ucoliter.h) |
michael@0 | 28 | * 02/19/01 swquek Removed CollationElementsIterator() since it is |
michael@0 | 29 | * private constructor and no calls are made to it |
michael@0 | 30 | */ |
michael@0 | 31 | |
michael@0 | 32 | #ifndef COLEITR_H |
michael@0 | 33 | #define COLEITR_H |
michael@0 | 34 | |
michael@0 | 35 | #include "unicode/utypes.h" |
michael@0 | 36 | |
michael@0 | 37 | |
michael@0 | 38 | #if !UCONFIG_NO_COLLATION |
michael@0 | 39 | |
michael@0 | 40 | #include "unicode/uobject.h" |
michael@0 | 41 | #include "unicode/tblcoll.h" |
michael@0 | 42 | #include "unicode/ucoleitr.h" |
michael@0 | 43 | |
michael@0 | 44 | /** |
michael@0 | 45 | * The UCollationElements struct. |
michael@0 | 46 | * For usage in C programs. |
michael@0 | 47 | * @stable ICU 2.0 |
michael@0 | 48 | */ |
michael@0 | 49 | typedef struct UCollationElements UCollationElements; |
michael@0 | 50 | |
michael@0 | 51 | U_NAMESPACE_BEGIN |
michael@0 | 52 | |
michael@0 | 53 | /** |
michael@0 | 54 | * The CollationElementIterator class is used as an iterator to walk through |
michael@0 | 55 | * each character of an international string. Use the iterator to return the |
michael@0 | 56 | * ordering priority of the positioned character. The ordering priority of a |
michael@0 | 57 | * character, which we refer to as a key, defines how a character is collated in |
michael@0 | 58 | * the given collation object. |
michael@0 | 59 | * For example, consider the following in Spanish: |
michael@0 | 60 | * <pre> |
michael@0 | 61 | * "ca" -> the first key is key('c') and second key is key('a'). |
michael@0 | 62 | * "cha" -> the first key is key('ch') and second key is key('a').</pre> |
michael@0 | 63 | * And in German, |
michael@0 | 64 | * <pre> \htmlonly "æb"-> the first key is key('a'), the second key is key('e'), and |
michael@0 | 65 | * the third key is key('b'). \endhtmlonly </pre> |
michael@0 | 66 | * The key of a character, is an integer composed of primary order(short), |
michael@0 | 67 | * secondary order(char), and tertiary order(char). Java strictly defines the |
michael@0 | 68 | * size and signedness of its primitive data types. Therefore, the static |
michael@0 | 69 | * functions primaryOrder(), secondaryOrder(), and tertiaryOrder() return |
michael@0 | 70 | * int32_t to ensure the correctness of the key value. |
michael@0 | 71 | * <p>Example of the iterator usage: (without error checking) |
michael@0 | 72 | * <pre> |
michael@0 | 73 | * \code |
michael@0 | 74 | * void CollationElementIterator_Example() |
michael@0 | 75 | * { |
michael@0 | 76 | * UnicodeString str = "This is a test"; |
michael@0 | 77 | * UErrorCode success = U_ZERO_ERROR; |
michael@0 | 78 | * RuleBasedCollator* rbc = |
michael@0 | 79 | * (RuleBasedCollator*) RuleBasedCollator::createInstance(success); |
michael@0 | 80 | * CollationElementIterator* c = |
michael@0 | 81 | * rbc->createCollationElementIterator( str ); |
michael@0 | 82 | * int32_t order = c->next(success); |
michael@0 | 83 | * c->reset(); |
michael@0 | 84 | * order = c->previous(success); |
michael@0 | 85 | * delete c; |
michael@0 | 86 | * delete rbc; |
michael@0 | 87 | * } |
michael@0 | 88 | * \endcode |
michael@0 | 89 | * </pre> |
michael@0 | 90 | * <p> |
michael@0 | 91 | * The method next() returns the collation order of the next character based on |
michael@0 | 92 | * the comparison level of the collator. The method previous() returns the |
michael@0 | 93 | * collation order of the previous character based on the comparison level of |
michael@0 | 94 | * the collator. The Collation Element Iterator moves only in one direction |
michael@0 | 95 | * between calls to reset(), setOffset(), or setText(). That is, next() |
michael@0 | 96 | * and previous() can not be inter-used. Whenever previous() is to be called after |
michael@0 | 97 | * next() or vice versa, reset(), setOffset() or setText() has to be called first |
michael@0 | 98 | * to reset the status, shifting pointers to either the end or the start of |
michael@0 | 99 | * the string (reset() or setText()), or the specified position (setOffset()). |
michael@0 | 100 | * Hence at the next call of next() or previous(), the first or last collation order, |
michael@0 | 101 | * or collation order at the spefcifieid position will be returned. If a change of |
michael@0 | 102 | * direction is done without one of these calls, the result is undefined. |
michael@0 | 103 | * <p> |
michael@0 | 104 | * The result of a forward iterate (next()) and reversed result of the backward |
michael@0 | 105 | * iterate (previous()) on the same string are equivalent, if collation orders |
michael@0 | 106 | * with the value UCOL_IGNORABLE are ignored. |
michael@0 | 107 | * Character based on the comparison level of the collator. A collation order |
michael@0 | 108 | * consists of primary order, secondary order and tertiary order. The data |
michael@0 | 109 | * type of the collation order is <strong>t_int32</strong>. |
michael@0 | 110 | * |
michael@0 | 111 | * Note, CollationElementIterator should not be subclassed. |
michael@0 | 112 | * @see Collator |
michael@0 | 113 | * @see RuleBasedCollator |
michael@0 | 114 | * @version 1.8 Jan 16 2001 |
michael@0 | 115 | */ |
michael@0 | 116 | class U_I18N_API CollationElementIterator : public UObject { |
michael@0 | 117 | public: |
michael@0 | 118 | |
michael@0 | 119 | // CollationElementIterator public data member ------------------------------ |
michael@0 | 120 | |
michael@0 | 121 | enum { |
michael@0 | 122 | /** |
michael@0 | 123 | * NULLORDER indicates that an error has occured while processing |
michael@0 | 124 | * @stable ICU 2.0 |
michael@0 | 125 | */ |
michael@0 | 126 | NULLORDER = (int32_t)0xffffffff |
michael@0 | 127 | }; |
michael@0 | 128 | |
michael@0 | 129 | // CollationElementIterator public constructor/destructor ------------------- |
michael@0 | 130 | |
michael@0 | 131 | /** |
michael@0 | 132 | * Copy constructor. |
michael@0 | 133 | * |
michael@0 | 134 | * @param other the object to be copied from |
michael@0 | 135 | * @stable ICU 2.0 |
michael@0 | 136 | */ |
michael@0 | 137 | CollationElementIterator(const CollationElementIterator& other); |
michael@0 | 138 | |
michael@0 | 139 | /** |
michael@0 | 140 | * Destructor |
michael@0 | 141 | * @stable ICU 2.0 |
michael@0 | 142 | */ |
michael@0 | 143 | virtual ~CollationElementIterator(); |
michael@0 | 144 | |
michael@0 | 145 | // CollationElementIterator public methods ---------------------------------- |
michael@0 | 146 | |
michael@0 | 147 | /** |
michael@0 | 148 | * Returns true if "other" is the same as "this" |
michael@0 | 149 | * |
michael@0 | 150 | * @param other the object to be compared |
michael@0 | 151 | * @return true if "other" is the same as "this" |
michael@0 | 152 | * @stable ICU 2.0 |
michael@0 | 153 | */ |
michael@0 | 154 | UBool operator==(const CollationElementIterator& other) const; |
michael@0 | 155 | |
michael@0 | 156 | /** |
michael@0 | 157 | * Returns true if "other" is not the same as "this". |
michael@0 | 158 | * |
michael@0 | 159 | * @param other the object to be compared |
michael@0 | 160 | * @return true if "other" is not the same as "this" |
michael@0 | 161 | * @stable ICU 2.0 |
michael@0 | 162 | */ |
michael@0 | 163 | UBool operator!=(const CollationElementIterator& other) const; |
michael@0 | 164 | |
michael@0 | 165 | /** |
michael@0 | 166 | * Resets the cursor to the beginning of the string. |
michael@0 | 167 | * @stable ICU 2.0 |
michael@0 | 168 | */ |
michael@0 | 169 | void reset(void); |
michael@0 | 170 | |
michael@0 | 171 | /** |
michael@0 | 172 | * Gets the ordering priority of the next character in the string. |
michael@0 | 173 | * @param status the error code status. |
michael@0 | 174 | * @return the next character's ordering. otherwise returns NULLORDER if an |
michael@0 | 175 | * error has occured or if the end of string has been reached |
michael@0 | 176 | * @stable ICU 2.0 |
michael@0 | 177 | */ |
michael@0 | 178 | int32_t next(UErrorCode& status); |
michael@0 | 179 | |
michael@0 | 180 | /** |
michael@0 | 181 | * Get the ordering priority of the previous collation element in the string. |
michael@0 | 182 | * @param status the error code status. |
michael@0 | 183 | * @return the previous element's ordering. otherwise returns NULLORDER if an |
michael@0 | 184 | * error has occured or if the start of string has been reached |
michael@0 | 185 | * @stable ICU 2.0 |
michael@0 | 186 | */ |
michael@0 | 187 | int32_t previous(UErrorCode& status); |
michael@0 | 188 | |
michael@0 | 189 | /** |
michael@0 | 190 | * Gets the primary order of a collation order. |
michael@0 | 191 | * @param order the collation order |
michael@0 | 192 | * @return the primary order of a collation order. |
michael@0 | 193 | * @stable ICU 2.0 |
michael@0 | 194 | */ |
michael@0 | 195 | static inline int32_t primaryOrder(int32_t order); |
michael@0 | 196 | |
michael@0 | 197 | /** |
michael@0 | 198 | * Gets the secondary order of a collation order. |
michael@0 | 199 | * @param order the collation order |
michael@0 | 200 | * @return the secondary order of a collation order. |
michael@0 | 201 | * @stable ICU 2.0 |
michael@0 | 202 | */ |
michael@0 | 203 | static inline int32_t secondaryOrder(int32_t order); |
michael@0 | 204 | |
michael@0 | 205 | /** |
michael@0 | 206 | * Gets the tertiary order of a collation order. |
michael@0 | 207 | * @param order the collation order |
michael@0 | 208 | * @return the tertiary order of a collation order. |
michael@0 | 209 | * @stable ICU 2.0 |
michael@0 | 210 | */ |
michael@0 | 211 | static inline int32_t tertiaryOrder(int32_t order); |
michael@0 | 212 | |
michael@0 | 213 | /** |
michael@0 | 214 | * Return the maximum length of any expansion sequences that end with the |
michael@0 | 215 | * specified comparison order. |
michael@0 | 216 | * @param order a collation order returned by previous or next. |
michael@0 | 217 | * @return maximum size of the expansion sequences ending with the collation |
michael@0 | 218 | * element or 1 if collation element does not occur at the end of any |
michael@0 | 219 | * expansion sequence |
michael@0 | 220 | * @stable ICU 2.0 |
michael@0 | 221 | */ |
michael@0 | 222 | int32_t getMaxExpansion(int32_t order) const; |
michael@0 | 223 | |
michael@0 | 224 | /** |
michael@0 | 225 | * Gets the comparison order in the desired strength. Ignore the other |
michael@0 | 226 | * differences. |
michael@0 | 227 | * @param order The order value |
michael@0 | 228 | * @stable ICU 2.0 |
michael@0 | 229 | */ |
michael@0 | 230 | int32_t strengthOrder(int32_t order) const; |
michael@0 | 231 | |
michael@0 | 232 | /** |
michael@0 | 233 | * Sets the source string. |
michael@0 | 234 | * @param str the source string. |
michael@0 | 235 | * @param status the error code status. |
michael@0 | 236 | * @stable ICU 2.0 |
michael@0 | 237 | */ |
michael@0 | 238 | void setText(const UnicodeString& str, UErrorCode& status); |
michael@0 | 239 | |
michael@0 | 240 | /** |
michael@0 | 241 | * Sets the source string. |
michael@0 | 242 | * @param str the source character iterator. |
michael@0 | 243 | * @param status the error code status. |
michael@0 | 244 | * @stable ICU 2.0 |
michael@0 | 245 | */ |
michael@0 | 246 | void setText(CharacterIterator& str, UErrorCode& status); |
michael@0 | 247 | |
michael@0 | 248 | /** |
michael@0 | 249 | * Checks if a comparison order is ignorable. |
michael@0 | 250 | * @param order the collation order. |
michael@0 | 251 | * @return TRUE if a character is ignorable, FALSE otherwise. |
michael@0 | 252 | * @stable ICU 2.0 |
michael@0 | 253 | */ |
michael@0 | 254 | static inline UBool isIgnorable(int32_t order); |
michael@0 | 255 | |
michael@0 | 256 | /** |
michael@0 | 257 | * Gets the offset of the currently processed character in the source string. |
michael@0 | 258 | * @return the offset of the character. |
michael@0 | 259 | * @stable ICU 2.0 |
michael@0 | 260 | */ |
michael@0 | 261 | int32_t getOffset(void) const; |
michael@0 | 262 | |
michael@0 | 263 | /** |
michael@0 | 264 | * Sets the offset of the currently processed character in the source string. |
michael@0 | 265 | * @param newOffset the new offset. |
michael@0 | 266 | * @param status the error code status. |
michael@0 | 267 | * @return the offset of the character. |
michael@0 | 268 | * @stable ICU 2.0 |
michael@0 | 269 | */ |
michael@0 | 270 | void setOffset(int32_t newOffset, UErrorCode& status); |
michael@0 | 271 | |
michael@0 | 272 | /** |
michael@0 | 273 | * ICU "poor man's RTTI", returns a UClassID for the actual class. |
michael@0 | 274 | * |
michael@0 | 275 | * @stable ICU 2.2 |
michael@0 | 276 | */ |
michael@0 | 277 | virtual UClassID getDynamicClassID() const; |
michael@0 | 278 | |
michael@0 | 279 | /** |
michael@0 | 280 | * ICU "poor man's RTTI", returns a UClassID for this class. |
michael@0 | 281 | * |
michael@0 | 282 | * @stable ICU 2.2 |
michael@0 | 283 | */ |
michael@0 | 284 | static UClassID U_EXPORT2 getStaticClassID(); |
michael@0 | 285 | |
michael@0 | 286 | private: |
michael@0 | 287 | friend class RuleBasedCollator; |
michael@0 | 288 | |
michael@0 | 289 | /** |
michael@0 | 290 | * CollationElementIterator constructor. This takes the source string and the |
michael@0 | 291 | * collation object. The cursor will walk thru the source string based on the |
michael@0 | 292 | * predefined collation rules. If the source string is empty, NULLORDER will |
michael@0 | 293 | * be returned on the calls to next(). |
michael@0 | 294 | * @param sourceText the source string. |
michael@0 | 295 | * @param order the collation object. |
michael@0 | 296 | * @param status the error code status. |
michael@0 | 297 | */ |
michael@0 | 298 | CollationElementIterator(const UnicodeString& sourceText, |
michael@0 | 299 | const RuleBasedCollator* order, UErrorCode& status); |
michael@0 | 300 | |
michael@0 | 301 | /** |
michael@0 | 302 | * CollationElementIterator constructor. This takes the source string and the |
michael@0 | 303 | * collation object. The cursor will walk thru the source string based on the |
michael@0 | 304 | * predefined collation rules. If the source string is empty, NULLORDER will |
michael@0 | 305 | * be returned on the calls to next(). |
michael@0 | 306 | * @param sourceText the source string. |
michael@0 | 307 | * @param order the collation object. |
michael@0 | 308 | * @param status the error code status. |
michael@0 | 309 | */ |
michael@0 | 310 | CollationElementIterator(const CharacterIterator& sourceText, |
michael@0 | 311 | const RuleBasedCollator* order, UErrorCode& status); |
michael@0 | 312 | |
michael@0 | 313 | /** |
michael@0 | 314 | * Assignment operator |
michael@0 | 315 | * |
michael@0 | 316 | * @param other the object to be copied |
michael@0 | 317 | */ |
michael@0 | 318 | const CollationElementIterator& |
michael@0 | 319 | operator=(const CollationElementIterator& other); |
michael@0 | 320 | |
michael@0 | 321 | CollationElementIterator(); // default constructor not implemented |
michael@0 | 322 | |
michael@0 | 323 | // CollationElementIterator private data members ---------------------------- |
michael@0 | 324 | |
michael@0 | 325 | /** |
michael@0 | 326 | * Data wrapper for collation elements |
michael@0 | 327 | */ |
michael@0 | 328 | UCollationElements *m_data_; |
michael@0 | 329 | |
michael@0 | 330 | /** |
michael@0 | 331 | * Indicates if m_data_ belongs to this object. |
michael@0 | 332 | */ |
michael@0 | 333 | UBool isDataOwned_; |
michael@0 | 334 | }; |
michael@0 | 335 | |
michael@0 | 336 | // CollationElementIterator inline method defination -------------------------- |
michael@0 | 337 | |
michael@0 | 338 | /** |
michael@0 | 339 | * Get the primary order of a collation order. |
michael@0 | 340 | * @param order the collation order |
michael@0 | 341 | * @return the primary order of a collation order. |
michael@0 | 342 | */ |
michael@0 | 343 | inline int32_t CollationElementIterator::primaryOrder(int32_t order) |
michael@0 | 344 | { |
michael@0 | 345 | order &= RuleBasedCollator::PRIMARYORDERMASK; |
michael@0 | 346 | return (order >> RuleBasedCollator::PRIMARYORDERSHIFT); |
michael@0 | 347 | } |
michael@0 | 348 | |
michael@0 | 349 | /** |
michael@0 | 350 | * Get the secondary order of a collation order. |
michael@0 | 351 | * @param order the collation order |
michael@0 | 352 | * @return the secondary order of a collation order. |
michael@0 | 353 | */ |
michael@0 | 354 | inline int32_t CollationElementIterator::secondaryOrder(int32_t order) |
michael@0 | 355 | { |
michael@0 | 356 | order = order & RuleBasedCollator::SECONDARYORDERMASK; |
michael@0 | 357 | return (order >> RuleBasedCollator::SECONDARYORDERSHIFT); |
michael@0 | 358 | } |
michael@0 | 359 | |
michael@0 | 360 | /** |
michael@0 | 361 | * Get the tertiary order of a collation order. |
michael@0 | 362 | * @param order the collation order |
michael@0 | 363 | * @return the tertiary order of a collation order. |
michael@0 | 364 | */ |
michael@0 | 365 | inline int32_t CollationElementIterator::tertiaryOrder(int32_t order) |
michael@0 | 366 | { |
michael@0 | 367 | return (order &= RuleBasedCollator::TERTIARYORDERMASK); |
michael@0 | 368 | } |
michael@0 | 369 | |
michael@0 | 370 | inline int32_t CollationElementIterator::getMaxExpansion(int32_t order) const |
michael@0 | 371 | { |
michael@0 | 372 | return ucol_getMaxExpansion(m_data_, (uint32_t)order); |
michael@0 | 373 | } |
michael@0 | 374 | |
michael@0 | 375 | inline UBool CollationElementIterator::isIgnorable(int32_t order) |
michael@0 | 376 | { |
michael@0 | 377 | return (primaryOrder(order) == RuleBasedCollator::PRIMIGNORABLE); |
michael@0 | 378 | } |
michael@0 | 379 | |
michael@0 | 380 | U_NAMESPACE_END |
michael@0 | 381 | |
michael@0 | 382 | #endif /* #if !UCONFIG_NO_COLLATION */ |
michael@0 | 383 | |
michael@0 | 384 | #endif |