godot/thirdparty/icu4c/common/unicode/edits.h

// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html

// edits.h
// created: 2016dec30 Markus W. Scherer

#ifndef __EDITS_H__
#define __EDITS_H__

#include "unicode/utypes.h"

#if U_SHOW_CPLUSPLUS_API

#include "unicode/uobject.h"

/**
 * \file
 * \brief C++ API: C++ class Edits for low-level string transformations on styled text.
 */

U_NAMESPACE_BEGIN

class UnicodeString;

/**
 * Records lengths of string edits but not replacement text. Supports replacements, insertions, deletions
 * in linear progression. Does not support moving/reordering of text.
 *
 * There are two types of edits: <em>change edits</em> and <em>no-change edits</em>. Add edits to
 * instances of this class using {@link #addReplace(int32_t, int32_t)} (for change edits) and
 * {@link #addUnchanged(int32_t)} (for no-change edits). Change edits are retained with full granularity,
 * whereas adjacent no-change edits are always merged together. In no-change edits, there is a one-to-one
 * mapping between code points in the source and destination strings.
 *
 * After all edits have been added, instances of this class should be considered immutable, and an
 * {@link Edits::Iterator} can be used for queries.
 *
 * There are four flavors of Edits::Iterator:
 *
 * <ul>
 * <li>{@link #getFineIterator()} retains full granularity of change edits.
 * <li>{@link #getFineChangesIterator()} retains full granularity of change edits, and when calling
 * next() on the iterator, skips over no-change edits (unchanged regions).
 * <li>{@link #getCoarseIterator()} treats adjacent change edits as a single edit. (Adjacent no-change
 * edits are automatically merged during the construction phase.)
 * <li>{@link #getCoarseChangesIterator()} treats adjacent change edits as a single edit, and when
 * calling next() on the iterator, skips over no-change edits (unchanged regions).
 * </ul>
 *
 * For example, consider the string "abcßDeF", which case-folds to "abcssdef". This string has the
 * following fine edits:
 * <ul>
 * <li>abc ⇨ abc (no-change)
 * <li>ß ⇨ ss (change)
 * <li>D ⇨ d (change)
 * <li>e ⇨ e (no-change)
 * <li>F ⇨ f (change)
 * </ul>
 * and the following coarse edits (note how adjacent change edits get merged together):
 * <ul>
 * <li>abc ⇨ abc (no-change)
 * <li>ßD ⇨ ssd (change)
 * <li>e ⇨ e (no-change)
 * <li>F ⇨ f (change)
 * </ul>
 *
 * The "fine changes" and "coarse changes" iterators will step through only the change edits when their
 * `Edits::Iterator::next()` methods are called. They are identical to the non-change iterators when
 * their `Edits::Iterator::findSourceIndex()` or `Edits::Iterator::findDestinationIndex()`
 * methods are used to walk through the string.
 *
 * For examples of how to use this class, see the test `TestCaseMapEditsIteratorDocs` in
 * UCharacterCaseTest.java.
 *
 * An Edits object tracks a separate UErrorCode, but ICU string transformation functions
 * (e.g., case mapping functions) merge any such errors into their API's UErrorCode.
 *
 * @stable ICU 59
 */
class U_COMMON_API Edits final : public UMemory { … };

U_NAMESPACE_END

#endif /* U_SHOW_CPLUSPLUS_API */

#endif  // __EDITS_H__