Mercurial > repos > rliterman > csp2

annotate CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/include/unicode/edits.h @ 69:33d812a61356

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 17:55:14 -0400
parents
children
rev   line source
jpayne@69 1 // © 2016 and later: Unicode, Inc. and others.
jpayne@69 2 // License & terms of use: http://www.unicode.org/copyright.html
jpayne@69 3
jpayne@69 4 // edits.h
jpayne@69 5 // created: 2016dec30 Markus W. Scherer
jpayne@69 6
jpayne@69 7 #ifndef __EDITS_H__
jpayne@69 8 #define __EDITS_H__
jpayne@69 9
jpayne@69 10 #include "unicode/utypes.h"
jpayne@69 11
jpayne@69 12 #if U_SHOW_CPLUSPLUS_API
jpayne@69 13
jpayne@69 14 #include "unicode/uobject.h"
jpayne@69 15
jpayne@69 16 /**
jpayne@69 17 * \file
jpayne@69 18 * \brief C++ API: C++ class Edits for low-level string transformations on styled text.
jpayne@69 19 */
jpayne@69 20
jpayne@69 21 U_NAMESPACE_BEGIN
jpayne@69 22
jpayne@69 23 class UnicodeString;
jpayne@69 24
jpayne@69 25 /**
jpayne@69 26 * Records lengths of string edits but not replacement text. Supports replacements, insertions, deletions
jpayne@69 27 * in linear progression. Does not support moving/reordering of text.
jpayne@69 28 *
jpayne@69 29 * There are two types of edits: <em>change edits</em> and <em>no-change edits</em>. Add edits to
jpayne@69 30 * instances of this class using {@link #addReplace(int32_t, int32_t)} (for change edits) and
jpayne@69 31 * {@link #addUnchanged(int32_t)} (for no-change edits). Change edits are retained with full granularity,
jpayne@69 32 * whereas adjacent no-change edits are always merged together. In no-change edits, there is a one-to-one
jpayne@69 33 * mapping between code points in the source and destination strings.
jpayne@69 34 *
jpayne@69 35 * After all edits have been added, instances of this class should be considered immutable, and an
jpayne@69 36 * {@link Edits::Iterator} can be used for queries.
jpayne@69 37 *
jpayne@69 38 * There are four flavors of Edits::Iterator:
jpayne@69 39 *
jpayne@69 40 * <ul>
jpayne@69 41 * <li>{@link #getFineIterator()} retains full granularity of change edits.
jpayne@69 42 * <li>{@link #getFineChangesIterator()} retains full granularity of change edits, and when calling
jpayne@69 43 * next() on the iterator, skips over no-change edits (unchanged regions).
jpayne@69 44 * <li>{@link #getCoarseIterator()} treats adjacent change edits as a single edit. (Adjacent no-change
jpayne@69 45 * edits are automatically merged during the construction phase.)
jpayne@69 46 * <li>{@link #getCoarseChangesIterator()} treats adjacent change edits as a single edit, and when
jpayne@69 47 * calling next() on the iterator, skips over no-change edits (unchanged regions).
jpayne@69 48 * </ul>
jpayne@69 49 *
jpayne@69 50 * For example, consider the string "abcßDeF", which case-folds to "abcssdef". This string has the
jpayne@69 51 * following fine edits:
jpayne@69 52 * <ul>
jpayne@69 53 * <li>abc ⇨ abc (no-change)
jpayne@69 54 * <li>ß ⇨ ss (change)
jpayne@69 55 * <li>D ⇨ d (change)
jpayne@69 56 * <li>e ⇨ e (no-change)
jpayne@69 57 * <li>F ⇨ f (change)
jpayne@69 58 * </ul>
jpayne@69 59 * and the following coarse edits (note how adjacent change edits get merged together):
jpayne@69 60 * <ul>
jpayne@69 61 * <li>abc ⇨ abc (no-change)
jpayne@69 62 * <li>ßD ⇨ ssd (change)
jpayne@69 63 * <li>e ⇨ e (no-change)
jpayne@69 64 * <li>F ⇨ f (change)
jpayne@69 65 * </ul>
jpayne@69 66 *
jpayne@69 67 * The "fine changes" and "coarse changes" iterators will step through only the change edits when their
jpayne@69 68 * `Edits::Iterator::next()` methods are called. They are identical to the non-change iterators when
jpayne@69 69 * their `Edits::Iterator::findSourceIndex()` or `Edits::Iterator::findDestinationIndex()`
jpayne@69 70 * methods are used to walk through the string.
jpayne@69 71 *
jpayne@69 72 * For examples of how to use this class, see the test `TestCaseMapEditsIteratorDocs` in
jpayne@69 73 * UCharacterCaseTest.java.
jpayne@69 74 *
jpayne@69 75 * An Edits object tracks a separate UErrorCode, but ICU string transformation functions
jpayne@69 76 * (e.g., case mapping functions) merge any such errors into their API's UErrorCode.
jpayne@69 77 *
jpayne@69 78 * @stable ICU 59
jpayne@69 79 */
jpayne@69 80 class U_COMMON_API Edits U_FINAL : public UMemory {
jpayne@69 81 public:
jpayne@69 82 /**
jpayne@69 83 * Constructs an empty object.
jpayne@69 84 * @stable ICU 59
jpayne@69 85 */
jpayne@69 86 Edits() :
jpayne@69 87 array(stackArray), capacity(STACK_CAPACITY), length(0), delta(0), numChanges(0),
jpayne@69 88 errorCode_(U_ZERO_ERROR) {}
jpayne@69 89 /**
jpayne@69 90 * Copy constructor.
jpayne@69 91 * @param other source edits
jpayne@69 92 * @stable ICU 60
jpayne@69 93 */
jpayne@69 94 Edits(const Edits &other) :
jpayne@69 95 array(stackArray), capacity(STACK_CAPACITY), length(other.length),
jpayne@69 96 delta(other.delta), numChanges(other.numChanges),
jpayne@69 97 errorCode_(other.errorCode_) {
jpayne@69 98 copyArray(other);
jpayne@69 99 }
jpayne@69 100 /**
jpayne@69 101 * Move constructor, might leave src empty.
jpayne@69 102 * This object will have the same contents that the source object had.
jpayne@69 103 * @param src source edits
jpayne@69 104 * @stable ICU 60
jpayne@69 105 */
jpayne@69 106 Edits(Edits &&src) U_NOEXCEPT :
jpayne@69 107 array(stackArray), capacity(STACK_CAPACITY), length(src.length),
jpayne@69 108 delta(src.delta), numChanges(src.numChanges),
jpayne@69 109 errorCode_(src.errorCode_) {
jpayne@69 110 moveArray(src);
jpayne@69 111 }
jpayne@69 112
jpayne@69 113 /**
jpayne@69 114 * Destructor.
jpayne@69 115 * @stable ICU 59
jpayne@69 116 */
jpayne@69 117 ~Edits();
jpayne@69 118
jpayne@69 119 /**
jpayne@69 120 * Assignment operator.
jpayne@69 121 * @param other source edits
jpayne@69 122 * @return *this
jpayne@69 123 * @stable ICU 60
jpayne@69 124 */
jpayne@69 125 Edits &operator=(const Edits &other);
jpayne@69 126
jpayne@69 127 /**
jpayne@69 128 * Move assignment operator, might leave src empty.
jpayne@69 129 * This object will have the same contents that the source object had.
jpayne@69 130 * The behavior is undefined if *this and src are the same object.
jpayne@69 131 * @param src source edits
jpayne@69 132 * @return *this
jpayne@69 133 * @stable ICU 60
jpayne@69 134 */
jpayne@69 135 Edits &operator=(Edits &&src) U_NOEXCEPT;
jpayne@69 136
jpayne@69 137 /**
jpayne@69 138 * Resets the data but may not release memory.
jpayne@69 139 * @stable ICU 59
jpayne@69 140 */
jpayne@69 141 void reset() U_NOEXCEPT;
jpayne@69 142
jpayne@69 143 /**
jpayne@69 144 * Adds a no-change edit: a record for an unchanged segment of text.
jpayne@69 145 * Normally called from inside ICU string transformation functions, not user code.
jpayne@69 146 * @stable ICU 59
jpayne@69 147 */
jpayne@69 148 void addUnchanged(int32_t unchangedLength);
jpayne@69 149 /**
jpayne@69 150 * Adds a change edit: a record for a text replacement/insertion/deletion.
jpayne@69 151 * Normally called from inside ICU string transformation functions, not user code.
jpayne@69 152 * @stable ICU 59
jpayne@69 153 */
jpayne@69 154 void addReplace(int32_t oldLength, int32_t newLength);
jpayne@69 155 /**
jpayne@69 156 * Sets the UErrorCode if an error occurred while recording edits.
jpayne@69 157 * Preserves older error codes in the outErrorCode.
jpayne@69 158 * Normally called from inside ICU string transformation functions, not user code.
jpayne@69 159 * @param outErrorCode Set to an error code if it does not contain one already
jpayne@69 160 * and an error occurred while recording edits.
jpayne@69 161 * Otherwise unchanged.
jpayne@69 162 * @return TRUE if U_FAILURE(outErrorCode)
jpayne@69 163 * @stable ICU 59
jpayne@69 164 */
jpayne@69 165 UBool copyErrorTo(UErrorCode &outErrorCode) const;
jpayne@69 166
jpayne@69 167 /**
jpayne@69 168 * How much longer is the new text compared with the old text?
jpayne@69 169 * @return new length minus old length
jpayne@69 170 * @stable ICU 59
jpayne@69 171 */
jpayne@69 172 int32_t lengthDelta() const { return delta; }
jpayne@69 173 /**
jpayne@69 174 * @return TRUE if there are any change edits
jpayne@69 175 * @stable ICU 59
jpayne@69 176 */
jpayne@69 177 UBool hasChanges() const { return numChanges != 0; }
jpayne@69 178
jpayne@69 179 /**
jpayne@69 180 * @return the number of change edits
jpayne@69 181 * @stable ICU 60
jpayne@69 182 */
jpayne@69 183 int32_t numberOfChanges() const { return numChanges; }
jpayne@69 184
jpayne@69 185 /**
jpayne@69 186 * Access to the list of edits.
jpayne@69 187 *
jpayne@69 188 * At any moment in time, an instance of this class points to a single edit: a "window" into a span
jpayne@69 189 * of the source string and the corresponding span of the destination string. The source string span
jpayne@69 190 * starts at {@link #sourceIndex()} and runs for {@link #oldLength()} chars; the destination string
jpayne@69 191 * span starts at {@link #destinationIndex()} and runs for {@link #newLength()} chars.
jpayne@69 192 *
jpayne@69 193 * The iterator can be moved between edits using the `next()`, `findSourceIndex(int32_t, UErrorCode &)`,
jpayne@69 194 * and `findDestinationIndex(int32_t, UErrorCode &)` methods.
jpayne@69 195 * Calling any of these methods mutates the iterator to make it point to the corresponding edit.
jpayne@69 196 *
jpayne@69 197 * For more information, see the documentation for {@link Edits}.
jpayne@69 198 *
jpayne@69 199 * @see getCoarseIterator
jpayne@69 200 * @see getFineIterator
jpayne@69 201 * @stable ICU 59
jpayne@69 202 */
jpayne@69 203 struct U_COMMON_API Iterator U_FINAL : public UMemory {
jpayne@69 204 /**
jpayne@69 205 * Default constructor, empty iterator.
jpayne@69 206 * @stable ICU 60
jpayne@69 207 */
jpayne@69 208 Iterator() :
jpayne@69 209 array(nullptr), index(0), length(0),
jpayne@69 210 remaining(0), onlyChanges_(FALSE), coarse(FALSE),
jpayne@69 211 dir(0), changed(FALSE), oldLength_(0), newLength_(0),
jpayne@69 212 srcIndex(0), replIndex(0), destIndex(0) {}
jpayne@69 213 /**
jpayne@69 214 * Copy constructor.
jpayne@69 215 * @stable ICU 59
jpayne@69 216 */
jpayne@69 217 Iterator(const Iterator &other) = default;
jpayne@69 218 /**
jpayne@69 219 * Assignment operator.
jpayne@69 220 * @stable ICU 59
jpayne@69 221 */
jpayne@69 222 Iterator &operator=(const Iterator &other) = default;
jpayne@69 223
jpayne@69 224 /**
jpayne@69 225 * Advances the iterator to the next edit.
jpayne@69 226 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 227 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 228 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 229 * @return TRUE if there is another edit
jpayne@69 230 * @stable ICU 59
jpayne@69 231 */
jpayne@69 232 UBool next(UErrorCode &errorCode) { return next(onlyChanges_, errorCode); }
jpayne@69 233
jpayne@69 234 /**
jpayne@69 235 * Moves the iterator to the edit that contains the source index.
jpayne@69 236 * The source index may be found in a no-change edit
jpayne@69 237 * even if normal iteration would skip no-change edits.
jpayne@69 238 * Normal iteration can continue from a found edit.
jpayne@69 239 *
jpayne@69 240 * The iterator state before this search logically does not matter.
jpayne@69 241 * (It may affect the performance of the search.)
jpayne@69 242 *
jpayne@69 243 * The iterator state after this search is undefined
jpayne@69 244 * if the source index is out of bounds for the source string.
jpayne@69 245 *
jpayne@69 246 * @param i source index
jpayne@69 247 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 248 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 249 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 250 * @return TRUE if the edit for the source index was found
jpayne@69 251 * @stable ICU 59
jpayne@69 252 */
jpayne@69 253 UBool findSourceIndex(int32_t i, UErrorCode &errorCode) {
jpayne@69 254 return findIndex(i, TRUE, errorCode) == 0;
jpayne@69 255 }
jpayne@69 256
jpayne@69 257 /**
jpayne@69 258 * Moves the iterator to the edit that contains the destination index.
jpayne@69 259 * The destination index may be found in a no-change edit
jpayne@69 260 * even if normal iteration would skip no-change edits.
jpayne@69 261 * Normal iteration can continue from a found edit.
jpayne@69 262 *
jpayne@69 263 * The iterator state before this search logically does not matter.
jpayne@69 264 * (It may affect the performance of the search.)
jpayne@69 265 *
jpayne@69 266 * The iterator state after this search is undefined
jpayne@69 267 * if the source index is out of bounds for the source string.
jpayne@69 268 *
jpayne@69 269 * @param i destination index
jpayne@69 270 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 271 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 272 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 273 * @return TRUE if the edit for the destination index was found
jpayne@69 274 * @stable ICU 60
jpayne@69 275 */
jpayne@69 276 UBool findDestinationIndex(int32_t i, UErrorCode &errorCode) {
jpayne@69 277 return findIndex(i, FALSE, errorCode) == 0;
jpayne@69 278 }
jpayne@69 279
jpayne@69 280 /**
jpayne@69 281 * Computes the destination index corresponding to the given source index.
jpayne@69 282 * If the source index is inside a change edit (not at its start),
jpayne@69 283 * then the destination index at the end of that edit is returned,
jpayne@69 284 * since there is no information about index mapping inside a change edit.
jpayne@69 285 *
jpayne@69 286 * (This means that indexes to the start and middle of an edit,
jpayne@69 287 * for example around a grapheme cluster, are mapped to indexes
jpayne@69 288 * encompassing the entire edit.
jpayne@69 289 * The alternative, mapping an interior index to the start,
jpayne@69 290 * would map such an interval to an empty one.)
jpayne@69 291 *
jpayne@69 292 * This operation will usually but not always modify this object.
jpayne@69 293 * The iterator state after this search is undefined.
jpayne@69 294 *
jpayne@69 295 * @param i source index
jpayne@69 296 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 297 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 298 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 299 * @return destination index; undefined if i is not 0..string length
jpayne@69 300 * @stable ICU 60
jpayne@69 301 */
jpayne@69 302 int32_t destinationIndexFromSourceIndex(int32_t i, UErrorCode &errorCode);
jpayne@69 303
jpayne@69 304 /**
jpayne@69 305 * Computes the source index corresponding to the given destination index.
jpayne@69 306 * If the destination index is inside a change edit (not at its start),
jpayne@69 307 * then the source index at the end of that edit is returned,
jpayne@69 308 * since there is no information about index mapping inside a change edit.
jpayne@69 309 *
jpayne@69 310 * (This means that indexes to the start and middle of an edit,
jpayne@69 311 * for example around a grapheme cluster, are mapped to indexes
jpayne@69 312 * encompassing the entire edit.
jpayne@69 313 * The alternative, mapping an interior index to the start,
jpayne@69 314 * would map such an interval to an empty one.)
jpayne@69 315 *
jpayne@69 316 * This operation will usually but not always modify this object.
jpayne@69 317 * The iterator state after this search is undefined.
jpayne@69 318 *
jpayne@69 319 * @param i destination index
jpayne@69 320 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 321 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 322 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 323 * @return source index; undefined if i is not 0..string length
jpayne@69 324 * @stable ICU 60
jpayne@69 325 */
jpayne@69 326 int32_t sourceIndexFromDestinationIndex(int32_t i, UErrorCode &errorCode);
jpayne@69 327
jpayne@69 328 /**
jpayne@69 329 * Returns whether the edit currently represented by the iterator is a change edit.
jpayne@69 330 *
jpayne@69 331 * @return TRUE if this edit replaces oldLength() units with newLength() different ones.
jpayne@69 332 * FALSE if oldLength units remain unchanged.
jpayne@69 333 * @stable ICU 59
jpayne@69 334 */
jpayne@69 335 UBool hasChange() const { return changed; }
jpayne@69 336
jpayne@69 337 /**
jpayne@69 338 * The length of the current span in the source string, which starts at {@link #sourceIndex}.
jpayne@69 339 *
jpayne@69 340 * @return the number of units in the original string which are replaced or remain unchanged.
jpayne@69 341 * @stable ICU 59
jpayne@69 342 */
jpayne@69 343 int32_t oldLength() const { return oldLength_; }
jpayne@69 344
jpayne@69 345 /**
jpayne@69 346 * The length of the current span in the destination string, which starts at
jpayne@69 347 * {@link #destinationIndex}, or in the replacement string, which starts at
jpayne@69 348 * {@link #replacementIndex}.
jpayne@69 349 *
jpayne@69 350 * @return the number of units in the modified string, if hasChange() is TRUE.
jpayne@69 351 * Same as oldLength if hasChange() is FALSE.
jpayne@69 352 * @stable ICU 59
jpayne@69 353 */
jpayne@69 354 int32_t newLength() const { return newLength_; }
jpayne@69 355
jpayne@69 356 /**
jpayne@69 357 * The start index of the current span in the source string; the span has length
jpayne@69 358 * {@link #oldLength}.
jpayne@69 359 *
jpayne@69 360 * @return the current index into the source string
jpayne@69 361 * @stable ICU 59
jpayne@69 362 */
jpayne@69 363 int32_t sourceIndex() const { return srcIndex; }
jpayne@69 364
jpayne@69 365 /**
jpayne@69 366 * The start index of the current span in the replacement string; the span has length
jpayne@69 367 * {@link #newLength}. Well-defined only if the current edit is a change edit.
jpayne@69 368 *
jpayne@69 369 * The *replacement string* is the concatenation of all substrings of the destination
jpayne@69 370 * string corresponding to change edits.
jpayne@69 371 *
jpayne@69 372 * This method is intended to be used together with operations that write only replacement
jpayne@69 373 * characters (e.g. operations specifying the \ref U_OMIT_UNCHANGED_TEXT option).
jpayne@69 374 * The source string can then be modified in-place.
jpayne@69 375 *
jpayne@69 376 * @return the current index into the replacement-characters-only string,
jpayne@69 377 * not counting unchanged spans
jpayne@69 378 * @stable ICU 59
jpayne@69 379 */
jpayne@69 380 int32_t replacementIndex() const {
jpayne@69 381 // TODO: Throw an exception if we aren't in a change edit?
jpayne@69 382 return replIndex;
jpayne@69 383 }
jpayne@69 384
jpayne@69 385 /**
jpayne@69 386 * The start index of the current span in the destination string; the span has length
jpayne@69 387 * {@link #newLength}.
jpayne@69 388 *
jpayne@69 389 * @return the current index into the full destination string
jpayne@69 390 * @stable ICU 59
jpayne@69 391 */
jpayne@69 392 int32_t destinationIndex() const { return destIndex; }
jpayne@69 393
jpayne@69 394 #ifndef U_HIDE_INTERNAL_API
jpayne@69 395 /**
jpayne@69 396 * A string representation of the current edit represented by the iterator for debugging. You
jpayne@69 397 * should not depend on the contents of the return string.
jpayne@69 398 * @internal
jpayne@69 399 */
jpayne@69 400 UnicodeString& toString(UnicodeString& appendTo) const;
jpayne@69 401 #endif // U_HIDE_INTERNAL_API
jpayne@69 402
jpayne@69 403 private:
jpayne@69 404 friend class Edits;
jpayne@69 405
jpayne@69 406 Iterator(const uint16_t *a, int32_t len, UBool oc, UBool crs);
jpayne@69 407
jpayne@69 408 int32_t readLength(int32_t head);
jpayne@69 409 void updateNextIndexes();
jpayne@69 410 void updatePreviousIndexes();
jpayne@69 411 UBool noNext();
jpayne@69 412 UBool next(UBool onlyChanges, UErrorCode &errorCode);
jpayne@69 413 UBool previous(UErrorCode &errorCode);
jpayne@69 414 /** @return -1: error or i<0; 0: found; 1: i>=string length */
jpayne@69 415 int32_t findIndex(int32_t i, UBool findSource, UErrorCode &errorCode);
jpayne@69 416
jpayne@69 417 const uint16_t *array;
jpayne@69 418 int32_t index, length;
jpayne@69 419 // 0 if we are not within compressed equal-length changes.
jpayne@69 420 // Otherwise the number of remaining changes, including the current one.
jpayne@69 421 int32_t remaining;
jpayne@69 422 UBool onlyChanges_, coarse;
jpayne@69 423
jpayne@69 424 int8_t dir; // iteration direction: back(<0), initial(0), forward(>0)
jpayne@69 425 UBool changed;
jpayne@69 426 int32_t oldLength_, newLength_;
jpayne@69 427 int32_t srcIndex, replIndex, destIndex;
jpayne@69 428 };
jpayne@69 429
jpayne@69 430 /**
jpayne@69 431 * Returns an Iterator for coarse-grained change edits
jpayne@69 432 * (adjacent change edits are treated as one).
jpayne@69 433 * Can be used to perform simple string updates.
jpayne@69 434 * Skips no-change edits.
jpayne@69 435 * @return an Iterator that merges adjacent changes.
jpayne@69 436 * @stable ICU 59
jpayne@69 437 */
jpayne@69 438 Iterator getCoarseChangesIterator() const {
jpayne@69 439 return Iterator(array, length, TRUE, TRUE);
jpayne@69 440 }
jpayne@69 441
jpayne@69 442 /**
jpayne@69 443 * Returns an Iterator for coarse-grained change and no-change edits
jpayne@69 444 * (adjacent change edits are treated as one).
jpayne@69 445 * Can be used to perform simple string updates.
jpayne@69 446 * Adjacent change edits are treated as one edit.
jpayne@69 447 * @return an Iterator that merges adjacent changes.
jpayne@69 448 * @stable ICU 59
jpayne@69 449 */
jpayne@69 450 Iterator getCoarseIterator() const {
jpayne@69 451 return Iterator(array, length, FALSE, TRUE);
jpayne@69 452 }
jpayne@69 453
jpayne@69 454 /**
jpayne@69 455 * Returns an Iterator for fine-grained change edits
jpayne@69 456 * (full granularity of change edits is retained).
jpayne@69 457 * Can be used for modifying styled text.
jpayne@69 458 * Skips no-change edits.
jpayne@69 459 * @return an Iterator that separates adjacent changes.
jpayne@69 460 * @stable ICU 59
jpayne@69 461 */
jpayne@69 462 Iterator getFineChangesIterator() const {
jpayne@69 463 return Iterator(array, length, TRUE, FALSE);
jpayne@69 464 }
jpayne@69 465
jpayne@69 466 /**
jpayne@69 467 * Returns an Iterator for fine-grained change and no-change edits
jpayne@69 468 * (full granularity of change edits is retained).
jpayne@69 469 * Can be used for modifying styled text.
jpayne@69 470 * @return an Iterator that separates adjacent changes.
jpayne@69 471 * @stable ICU 59
jpayne@69 472 */
jpayne@69 473 Iterator getFineIterator() const {
jpayne@69 474 return Iterator(array, length, FALSE, FALSE);
jpayne@69 475 }
jpayne@69 476
jpayne@69 477 /**
jpayne@69 478 * Merges the two input Edits and appends the result to this object.
jpayne@69 479 *
jpayne@69 480 * Consider two string transformations (for example, normalization and case mapping)
jpayne@69 481 * where each records Edits in addition to writing an output string.<br>
jpayne@69 482 * Edits ab reflect how substrings of input string a
jpayne@69 483 * map to substrings of intermediate string b.<br>
jpayne@69 484 * Edits bc reflect how substrings of intermediate string b
jpayne@69 485 * map to substrings of output string c.<br>
jpayne@69 486 * This function merges ab and bc such that the additional edits
jpayne@69 487 * recorded in this object reflect how substrings of input string a
jpayne@69 488 * map to substrings of output string c.
jpayne@69 489 *
jpayne@69 490 * If unrelated Edits are passed in where the output string of the first
jpayne@69 491 * has a different length than the input string of the second,
jpayne@69 492 * then a U_ILLEGAL_ARGUMENT_ERROR is reported.
jpayne@69 493 *
jpayne@69 494 * @param ab reflects how substrings of input string a
jpayne@69 495 * map to substrings of intermediate string b.
jpayne@69 496 * @param bc reflects how substrings of intermediate string b
jpayne@69 497 * map to substrings of output string c.
jpayne@69 498 * @param errorCode ICU error code. Its input value must pass the U_SUCCESS() test,
jpayne@69 499 * or else the function returns immediately. Check for U_FAILURE()
jpayne@69 500 * on output or use with function chaining. (See User Guide for details.)
jpayne@69 501 * @return *this, with the merged edits appended
jpayne@69 502 * @stable ICU 60
jpayne@69 503 */
jpayne@69 504 Edits &mergeAndAppend(const Edits &ab, const Edits &bc, UErrorCode &errorCode);
jpayne@69 505
jpayne@69 506 private:
jpayne@69 507 void releaseArray() U_NOEXCEPT;
jpayne@69 508 Edits &copyArray(const Edits &other);
jpayne@69 509 Edits &moveArray(Edits &src) U_NOEXCEPT;
jpayne@69 510
jpayne@69 511 void setLastUnit(int32_t last) { array[length - 1] = (uint16_t)last; }
jpayne@69 512 int32_t lastUnit() const { return length > 0 ? array[length - 1] : 0xffff; }
jpayne@69 513
jpayne@69 514 void append(int32_t r);
jpayne@69 515 UBool growArray();
jpayne@69 516
jpayne@69 517 static const int32_t STACK_CAPACITY = 100;
jpayne@69 518 uint16_t *array;
jpayne@69 519 int32_t capacity;
jpayne@69 520 int32_t length;
jpayne@69 521 int32_t delta;
jpayne@69 522 int32_t numChanges;
jpayne@69 523 UErrorCode errorCode_;
jpayne@69 524 uint16_t stackArray[STACK_CAPACITY];
jpayne@69 525 };
jpayne@69 526
jpayne@69 527 U_NAMESPACE_END
jpayne@69 528
jpayne@69 529 #endif /* U_SHOW_CPLUSPLUS_API */
jpayne@69 530
jpayne@69 531 #endif // __EDITS_H__